diff options
Diffstat (limited to 'fs')
244 files changed, 13988 insertions, 7924 deletions
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 8d7f3e69ae29..7f6c67703195 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -814,7 +814,6 @@ int v9fs_vfs_unlink(struct inode *i, struct dentry *d) int v9fs_vfs_rmdir(struct inode *i, struct dentry *d) { - dentry_unhash(d); return v9fs_remove(i, d, 1); } @@ -840,9 +839,6 @@ v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct p9_fid *newdirfid; struct p9_wstat wstat; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - P9_DPRINTK(P9_DEBUG_VFS, "\n"); retval = 0; old_inode = old_dentry->d_inode; diff --git a/fs/affs/namei.c b/fs/affs/namei.c index 03330e2e390c..e3e9efc1fdd8 100644 --- a/fs/affs/namei.c +++ b/fs/affs/namei.c @@ -320,8 +320,6 @@ affs_rmdir(struct inode *dir, struct dentry *dentry) dentry->d_inode->i_ino, (int)dentry->d_name.len, dentry->d_name.name); - dentry_unhash(dentry); - return affs_remove_header(dentry); } @@ -419,9 +417,6 @@ affs_rename(struct inode *old_dir, struct dentry *old_dentry, struct buffer_head *bh = NULL; int retval; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - pr_debug("AFFS: rename(old=%u,\"%*s\" to new=%u,\"%*s\")\n", (u32)old_dir->i_ino, (int)old_dentry->d_name.len, old_dentry->d_name.name, (u32)new_dir->i_ino, (int)new_dentry->d_name.len, new_dentry->d_name.name); diff --git a/fs/afs/dir.c b/fs/afs/dir.c index 2c4e05160042..1b0b19550015 100644 --- a/fs/afs/dir.c +++ b/fs/afs/dir.c @@ -584,11 +584,11 @@ static struct dentry *afs_lookup(struct inode *dir, struct dentry *dentry, success: d_add(dentry, inode); - _leave(" = 0 { vn=%u u=%u } -> { ino=%lu v=%llu }", + _leave(" = 0 { vn=%u u=%u } -> { ino=%lu v=%u }", fid.vnode, fid.unique, dentry->d_inode->i_ino, - (unsigned long long)dentry->d_inode->i_version); + dentry->d_inode->i_generation); return NULL; } @@ -671,10 +671,10 @@ static int afs_d_revalidate(struct dentry *dentry, struct nameidata *nd) * been deleted and replaced, and the original vnode ID has * been reused */ if (fid.unique != vnode->fid.unique) { - _debug("%s: file deleted (uq %u -> %u I:%llu)", + _debug("%s: file deleted (uq %u -> %u I:%u)", dentry->d_name.name, fid.unique, vnode->fid.unique, - (unsigned long long)dentry->d_inode->i_version); + dentry->d_inode->i_generation); spin_lock(&vnode->lock); set_bit(AFS_VNODE_DELETED, &vnode->flags); spin_unlock(&vnode->lock); @@ -845,8 +845,6 @@ static int afs_rmdir(struct inode *dir, struct dentry *dentry) _enter("{%x:%u},{%s}", dvnode->fid.vid, dvnode->fid.vnode, dentry->d_name.name); - dentry_unhash(dentry); - ret = -ENAMETOOLONG; if (dentry->d_name.len >= AFSNAMEMAX) goto error; @@ -1148,9 +1146,6 @@ static int afs_rename(struct inode *old_dir, struct dentry *old_dentry, struct key *key; int ret; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - vnode = AFS_FS_I(old_dentry->d_inode); orig_dvnode = AFS_FS_I(old_dir); new_dvnode = AFS_FS_I(new_dir); diff --git a/fs/afs/fsclient.c b/fs/afs/fsclient.c index 4bd0218473a9..346e3289abd7 100644 --- a/fs/afs/fsclient.c +++ b/fs/afs/fsclient.c @@ -89,7 +89,7 @@ static void xdr_decode_AFSFetchStatus(const __be32 **_bp, i_size_write(&vnode->vfs_inode, size); vnode->vfs_inode.i_uid = status->owner; vnode->vfs_inode.i_gid = status->group; - vnode->vfs_inode.i_version = vnode->fid.unique; + vnode->vfs_inode.i_generation = vnode->fid.unique; vnode->vfs_inode.i_nlink = status->nlink; mode = vnode->vfs_inode.i_mode; @@ -102,6 +102,7 @@ static void xdr_decode_AFSFetchStatus(const __be32 **_bp, vnode->vfs_inode.i_ctime.tv_sec = status->mtime_server; vnode->vfs_inode.i_mtime = vnode->vfs_inode.i_ctime; vnode->vfs_inode.i_atime = vnode->vfs_inode.i_ctime; + vnode->vfs_inode.i_version = data_version; } expected_version = status->data_version; diff --git a/fs/afs/inode.c b/fs/afs/inode.c index db66c5201474..0fdab6e03d87 100644 --- a/fs/afs/inode.c +++ b/fs/afs/inode.c @@ -75,7 +75,8 @@ static int afs_inode_map_status(struct afs_vnode *vnode, struct key *key) inode->i_ctime.tv_nsec = 0; inode->i_atime = inode->i_mtime = inode->i_ctime; inode->i_blocks = 0; - inode->i_version = vnode->fid.unique; + inode->i_generation = vnode->fid.unique; + inode->i_version = vnode->status.data_version; inode->i_mapping->a_ops = &afs_fs_aops; /* check to see whether a symbolic link is really a mountpoint */ @@ -100,7 +101,7 @@ static int afs_iget5_test(struct inode *inode, void *opaque) struct afs_iget_data *data = opaque; return inode->i_ino == data->fid.vnode && - inode->i_version == data->fid.unique; + inode->i_generation == data->fid.unique; } /* @@ -122,7 +123,7 @@ static int afs_iget5_set(struct inode *inode, void *opaque) struct afs_vnode *vnode = AFS_FS_I(inode); inode->i_ino = data->fid.vnode; - inode->i_version = data->fid.unique; + inode->i_generation = data->fid.unique; vnode->fid = data->fid; vnode->volume = data->volume; @@ -380,8 +381,7 @@ int afs_getattr(struct vfsmount *mnt, struct dentry *dentry, inode = dentry->d_inode; - _enter("{ ino=%lu v=%llu }", inode->i_ino, - (unsigned long long)inode->i_version); + _enter("{ ino=%lu v=%u }", inode->i_ino, inode->i_generation); generic_fillattr(inode, stat); return 0; diff --git a/fs/afs/super.c b/fs/afs/super.c index fb240e8766d6..356dcf0929e8 100644 --- a/fs/afs/super.c +++ b/fs/afs/super.c @@ -31,8 +31,8 @@ static void afs_i_init_once(void *foo); static struct dentry *afs_mount(struct file_system_type *fs_type, int flags, const char *dev_name, void *data); +static void afs_kill_super(struct super_block *sb); static struct inode *afs_alloc_inode(struct super_block *sb); -static void afs_put_super(struct super_block *sb); static void afs_destroy_inode(struct inode *inode); static int afs_statfs(struct dentry *dentry, struct kstatfs *buf); @@ -40,7 +40,7 @@ struct file_system_type afs_fs_type = { .owner = THIS_MODULE, .name = "afs", .mount = afs_mount, - .kill_sb = kill_anon_super, + .kill_sb = afs_kill_super, .fs_flags = 0, }; @@ -50,7 +50,6 @@ static const struct super_operations afs_super_ops = { .drop_inode = afs_drop_inode, .destroy_inode = afs_destroy_inode, .evict_inode = afs_evict_inode, - .put_super = afs_put_super, .show_options = generic_show_options, }; @@ -282,19 +281,25 @@ static int afs_parse_device_name(struct afs_mount_params *params, */ static int afs_test_super(struct super_block *sb, void *data) { - struct afs_mount_params *params = data; + struct afs_super_info *as1 = data; struct afs_super_info *as = sb->s_fs_info; - return as->volume == params->volume; + return as->volume == as1->volume; +} + +static int afs_set_super(struct super_block *sb, void *data) +{ + sb->s_fs_info = data; + return set_anon_super(sb, NULL); } /* * fill in the superblock */ -static int afs_fill_super(struct super_block *sb, void *data) +static int afs_fill_super(struct super_block *sb, + struct afs_mount_params *params) { - struct afs_mount_params *params = data; - struct afs_super_info *as = NULL; + struct afs_super_info *as = sb->s_fs_info; struct afs_fid fid; struct dentry *root = NULL; struct inode *inode = NULL; @@ -302,23 +307,13 @@ static int afs_fill_super(struct super_block *sb, void *data) _enter(""); - /* allocate a superblock info record */ - as = kzalloc(sizeof(struct afs_super_info), GFP_KERNEL); - if (!as) { - _leave(" = -ENOMEM"); - return -ENOMEM; - } - - afs_get_volume(params->volume); - as->volume = params->volume; - /* fill in the superblock */ sb->s_blocksize = PAGE_CACHE_SIZE; sb->s_blocksize_bits = PAGE_CACHE_SHIFT; sb->s_magic = AFS_FS_MAGIC; sb->s_op = &afs_super_ops; - sb->s_fs_info = as; sb->s_bdi = &as->volume->bdi; + strlcpy(sb->s_id, as->volume->vlocation->vldb.name, sizeof(sb->s_id)); /* allocate the root inode and dentry */ fid.vid = as->volume->vid; @@ -326,7 +321,7 @@ static int afs_fill_super(struct super_block *sb, void *data) fid.unique = 1; inode = afs_iget(sb, params->key, &fid, NULL, NULL); if (IS_ERR(inode)) - goto error_inode; + return PTR_ERR(inode); if (params->autocell) set_bit(AFS_VNODE_AUTOCELL, &AFS_FS_I(inode)->flags); @@ -342,16 +337,8 @@ static int afs_fill_super(struct super_block *sb, void *data) _leave(" = 0"); return 0; -error_inode: - ret = PTR_ERR(inode); - inode = NULL; error: iput(inode); - afs_put_volume(as->volume); - kfree(as); - - sb->s_fs_info = NULL; - _leave(" = %d", ret); return ret; } @@ -367,6 +354,7 @@ static struct dentry *afs_mount(struct file_system_type *fs_type, struct afs_volume *vol; struct key *key; char *new_opts = kstrdup(options, GFP_KERNEL); + struct afs_super_info *as; int ret; _enter(",,%s,%p", dev_name, options); @@ -399,12 +387,22 @@ static struct dentry *afs_mount(struct file_system_type *fs_type, ret = PTR_ERR(vol); goto error; } - params.volume = vol; + + /* allocate a superblock info record */ + as = kzalloc(sizeof(struct afs_super_info), GFP_KERNEL); + if (!as) { + ret = -ENOMEM; + afs_put_volume(vol); + goto error; + } + as->volume = vol; /* allocate a deviceless superblock */ - sb = sget(fs_type, afs_test_super, set_anon_super, ¶ms); + sb = sget(fs_type, afs_test_super, afs_set_super, as); if (IS_ERR(sb)) { ret = PTR_ERR(sb); + afs_put_volume(vol); + kfree(as); goto error; } @@ -422,16 +420,16 @@ static struct dentry *afs_mount(struct file_system_type *fs_type, } else { _debug("reuse"); ASSERTCMP(sb->s_flags, &, MS_ACTIVE); + afs_put_volume(vol); + kfree(as); } - afs_put_volume(params.volume); afs_put_cell(params.cell); kfree(new_opts); _leave(" = 0 [%p]", sb); return dget(sb->s_root); error: - afs_put_volume(params.volume); afs_put_cell(params.cell); key_put(params.key); kfree(new_opts); @@ -439,18 +437,12 @@ error: return ERR_PTR(ret); } -/* - * finish the unmounting process on the superblock - */ -static void afs_put_super(struct super_block *sb) +static void afs_kill_super(struct super_block *sb) { struct afs_super_info *as = sb->s_fs_info; - - _enter(""); - + kill_anon_super(sb); afs_put_volume(as->volume); - - _leave(""); + kfree(as); } /* diff --git a/fs/afs/write.c b/fs/afs/write.c index 789b3afb3423..b806285ff853 100644 --- a/fs/afs/write.c +++ b/fs/afs/write.c @@ -84,23 +84,21 @@ void afs_put_writeback(struct afs_writeback *wb) * partly or wholly fill a page that's under preparation for writing */ static int afs_fill_page(struct afs_vnode *vnode, struct key *key, - loff_t pos, unsigned len, struct page *page) + loff_t pos, struct page *page) { loff_t i_size; - unsigned eof; int ret; + int len; - _enter(",,%llu,%u", (unsigned long long)pos, len); - - ASSERTCMP(len, <=, PAGE_CACHE_SIZE); + _enter(",,%llu", (unsigned long long)pos); i_size = i_size_read(&vnode->vfs_inode); - if (pos + len > i_size) - eof = i_size; + if (pos + PAGE_CACHE_SIZE > i_size) + len = i_size - pos; else - eof = PAGE_CACHE_SIZE; + len = PAGE_CACHE_SIZE; - ret = afs_vnode_fetch_data(vnode, key, 0, eof, page); + ret = afs_vnode_fetch_data(vnode, key, pos, len, page); if (ret < 0) { if (ret == -ENOENT) { _debug("got NOENT from server" @@ -153,9 +151,8 @@ int afs_write_begin(struct file *file, struct address_space *mapping, *pagep = page; /* page won't leak in error case: it eventually gets cleaned off LRU */ - if (!PageUptodate(page)) { - _debug("not up to date"); - ret = afs_fill_page(vnode, key, pos, len, page); + if (!PageUptodate(page) && len != PAGE_CACHE_SIZE) { + ret = afs_fill_page(vnode, key, index << PAGE_CACHE_SHIFT, page); if (ret < 0) { kfree(candidate); _leave(" = %d [prep]", ret); diff --git a/fs/attr.c b/fs/attr.c index 91dbe2a107f2..caf2aa521e2b 100644 --- a/fs/attr.c +++ b/fs/attr.c @@ -175,6 +175,13 @@ int notify_change(struct dentry * dentry, struct iattr * attr) return -EPERM; } + if ((ia_valid & ATTR_MODE)) { + mode_t amode = attr->ia_mode; + /* Flag setting protected by i_mutex */ + if (is_sxid(amode)) + inode->i_flags &= ~S_NOSEC; + } + now = current_fs_time(inode->i_sb); attr->ia_ctime = now; diff --git a/fs/autofs4/root.c b/fs/autofs4/root.c index 87d95a8cddbc..f55ae23b137e 100644 --- a/fs/autofs4/root.c +++ b/fs/autofs4/root.c @@ -583,8 +583,6 @@ static int autofs4_dir_unlink(struct inode *dir, struct dentry *dentry) if (!autofs4_oz_mode(sbi) && !capable(CAP_SYS_ADMIN)) return -EACCES; - dentry_unhash(dentry); - if (atomic_dec_and_test(&ino->count)) { p_ino = autofs4_dentry_ino(dentry->d_parent); if (p_ino && dentry->d_parent != dentry) diff --git a/fs/bad_inode.c b/fs/bad_inode.c index 9ad2369d9e35..bfcb18feb1df 100644 --- a/fs/bad_inode.c +++ b/fs/bad_inode.c @@ -231,9 +231,6 @@ static int bad_inode_readlink(struct dentry *dentry, char __user *buffer, static int bad_inode_permission(struct inode *inode, int mask, unsigned int flags) { - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - return -EIO; } diff --git a/fs/bfs/dir.c b/fs/bfs/dir.c index c7d1d06b0483..b14cebfd9047 100644 --- a/fs/bfs/dir.c +++ b/fs/bfs/dir.c @@ -224,9 +224,6 @@ static int bfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct bfs_sb_info *info; int error = -ENOENT; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - old_bh = new_bh = NULL; old_inode = old_dentry->d_inode; if (S_ISDIR(old_inode->i_mode)) @@ -638,10 +638,11 @@ static int __bio_add_page(struct request_queue *q, struct bio *bio, struct page * @offset: vec entry offset * * Attempt to add a page to the bio_vec maplist. This can fail for a - * number of reasons, such as the bio being full or target block - * device limitations. The target block device must allow bio's - * smaller than PAGE_SIZE, so it is always possible to add a single - * page to an empty bio. This should only be used by REQ_PC bios. + * number of reasons, such as the bio being full or target block device + * limitations. The target block device must allow bio's up to PAGE_SIZE, + * so it is always possible to add a single page to an empty bio. + * + * This should only be used by REQ_PC bios. */ int bio_add_pc_page(struct request_queue *q, struct bio *bio, struct page *page, unsigned int len, unsigned int offset) @@ -659,10 +660,9 @@ EXPORT_SYMBOL(bio_add_pc_page); * @offset: vec entry offset * * Attempt to add a page to the bio_vec maplist. This can fail for a - * number of reasons, such as the bio being full or target block - * device limitations. The target block device must allow bio's - * smaller than PAGE_SIZE, so it is always possible to add a single - * page to an empty bio. + * number of reasons, such as the bio being full or target block device + * limitations. The target block device must allow bio's up to PAGE_SIZE, + * so it is always possible to add a single page to an empty bio. */ int bio_add_page(struct bio *bio, struct page *page, unsigned int len, unsigned int offset) diff --git a/fs/block_dev.c b/fs/block_dev.c index 1f2b19978333..610e8e0b04b8 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -762,7 +762,19 @@ static struct block_device *bd_start_claiming(struct block_device *bdev, if (!disk) return ERR_PTR(-ENXIO); - whole = bdget_disk(disk, 0); + /* + * Normally, @bdev should equal what's returned from bdget_disk() + * if partno is 0; however, some drivers (floppy) use multiple + * bdev's for the same physical device and @bdev may be one of the + * aliases. Keep @bdev if partno is 0. This means claimer + * tracking is broken for those devices but it has always been that + * way. + */ + if (partno) + whole = bdget_disk(disk, 0); + else + whole = bdgrab(bdev); + module_put(disk->fops->owner); put_disk(disk); if (!whole) @@ -1272,8 +1284,8 @@ int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder) * individual writeable reference is too fragile given the * way @mode is used in blkdev_get/put(). */ - if ((disk->flags & GENHD_FL_BLOCK_EVENTS_ON_EXCL_WRITE) && - !res && (mode & FMODE_WRITE) && !bdev->bd_write_holder) { + if (!res && (mode & FMODE_WRITE) && !bdev->bd_write_holder && + (disk->flags & GENHD_FL_BLOCK_EVENTS_ON_EXCL_WRITE)) { bdev->bd_write_holder = true; disk_block_events(disk); } diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile index 31610ea73aec..9b72dcf1cd25 100644 --- a/fs/btrfs/Makefile +++ b/fs/btrfs/Makefile @@ -7,4 +7,4 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \ extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \ extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \ export.o tree-log.o acl.o free-space-cache.o zlib.o lzo.o \ - compression.o delayed-ref.o relocation.o + compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c index 44ea5b92e1ba..f66fc9959733 100644 --- a/fs/btrfs/acl.c +++ b/fs/btrfs/acl.c @@ -288,7 +288,7 @@ int btrfs_acl_chmod(struct inode *inode) return 0; acl = btrfs_get_acl(inode, ACL_TYPE_ACCESS); - if (IS_ERR(acl) || !acl) + if (IS_ERR_OR_NULL(acl)) return PTR_ERR(acl); clone = posix_acl_clone(acl, GFP_KERNEL); diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h index 57c3bb2884ce..52d7eca8c7bf 100644 --- a/fs/btrfs/btrfs_inode.h +++ b/fs/btrfs/btrfs_inode.h @@ -22,6 +22,7 @@ #include "extent_map.h" #include "extent_io.h" #include "ordered-data.h" +#include "delayed-inode.h" /* in memory btrfs inode */ struct btrfs_inode { @@ -120,9 +121,6 @@ struct btrfs_inode { */ u64 index_cnt; - /* the start of block group preferred for allocations. */ - u64 block_group; - /* the fsync log has some corner cases that mean we have to check * directories to see if any unlinks have been done before * the directory was logged. See tree-log.c for all the @@ -152,20 +150,34 @@ struct btrfs_inode { unsigned ordered_data_close:1; unsigned orphan_meta_reserved:1; unsigned dummy_inode:1; + unsigned in_defrag:1; /* * always compress this one file */ unsigned force_compress:4; + struct btrfs_delayed_node *delayed_node; + struct inode vfs_inode; }; +extern unsigned char btrfs_filetype_table[]; + static inline struct btrfs_inode *BTRFS_I(struct inode *inode) { return container_of(inode, struct btrfs_inode, vfs_inode); } +static inline u64 btrfs_ino(struct inode *inode) +{ + u64 ino = BTRFS_I(inode)->location.objectid; + + if (ino <= BTRFS_FIRST_FREE_OBJECTID) + ino = inode->i_ino; + return ino; +} + static inline void btrfs_i_size_write(struct inode *inode, u64 size) { i_size_write(inode, size); diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 41d1d7c70e29..bfe42b03eaf9 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -125,9 +125,10 @@ static int check_compressed_csum(struct inode *inode, kunmap_atomic(kaddr, KM_USER0); if (csum != *cb_sum) { - printk(KERN_INFO "btrfs csum failed ino %lu " + printk(KERN_INFO "btrfs csum failed ino %llu " "extent %llu csum %u " - "wanted %u mirror %d\n", inode->i_ino, + "wanted %u mirror %d\n", + (unsigned long long)btrfs_ino(inode), (unsigned long long)disk_start, csum, *cb_sum, cb->mirror_num); ret = -EIO; @@ -332,7 +333,7 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, struct compressed_bio *cb; unsigned long bytes_left; struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; - int page_index = 0; + int pg_index = 0; struct page *page; u64 first_byte = disk_start; struct block_device *bdev; @@ -366,8 +367,8 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, /* create and submit bios for the compressed pages */ bytes_left = compressed_len; - for (page_index = 0; page_index < cb->nr_pages; page_index++) { - page = compressed_pages[page_index]; + for (pg_index = 0; pg_index < cb->nr_pages; pg_index++) { + page = compressed_pages[pg_index]; page->mapping = inode->i_mapping; if (bio->bi_size) ret = io_tree->ops->merge_bio_hook(page, 0, @@ -432,7 +433,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, struct compressed_bio *cb) { unsigned long end_index; - unsigned long page_index; + unsigned long pg_index; u64 last_offset; u64 isize = i_size_read(inode); int ret; @@ -456,13 +457,13 @@ static noinline int add_ra_bio_pages(struct inode *inode, end_index = (i_size_read(inode) - 1) >> PAGE_CACHE_SHIFT; while (last_offset < compressed_end) { - page_index = last_offset >> PAGE_CACHE_SHIFT; + pg_index = last_offset >> PAGE_CACHE_SHIFT; - if (page_index > end_index) + if (pg_index > end_index) break; rcu_read_lock(); - page = radix_tree_lookup(&mapping->page_tree, page_index); + page = radix_tree_lookup(&mapping->page_tree, pg_index); rcu_read_unlock(); if (page) { misses++; @@ -476,7 +477,7 @@ static noinline int add_ra_bio_pages(struct inode *inode, if (!page) break; - if (add_to_page_cache_lru(page, mapping, page_index, + if (add_to_page_cache_lru(page, mapping, pg_index, GFP_NOFS)) { page_cache_release(page); goto next; @@ -560,7 +561,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, unsigned long uncompressed_len = bio->bi_vcnt * PAGE_CACHE_SIZE; unsigned long compressed_len; unsigned long nr_pages; - unsigned long page_index; + unsigned long pg_index; struct page *page; struct block_device *bdev; struct bio *comp_bio; @@ -613,10 +614,10 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev; - for (page_index = 0; page_index < nr_pages; page_index++) { - cb->compressed_pages[page_index] = alloc_page(GFP_NOFS | + for (pg_index = 0; pg_index < nr_pages; pg_index++) { + cb->compressed_pages[pg_index] = alloc_page(GFP_NOFS | __GFP_HIGHMEM); - if (!cb->compressed_pages[page_index]) + if (!cb->compressed_pages[pg_index]) goto fail2; } cb->nr_pages = nr_pages; @@ -634,8 +635,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, comp_bio->bi_end_io = end_compressed_bio_read; atomic_inc(&cb->pending_bios); - for (page_index = 0; page_index < nr_pages; page_index++) { - page = cb->compressed_pages[page_index]; + for (pg_index = 0; pg_index < nr_pages; pg_index++) { + page = cb->compressed_pages[pg_index]; page->mapping = inode->i_mapping; page->index = em_start >> PAGE_CACHE_SHIFT; @@ -702,8 +703,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, return 0; fail2: - for (page_index = 0; page_index < nr_pages; page_index++) - free_page((unsigned long)cb->compressed_pages[page_index]); + for (pg_index = 0; pg_index < nr_pages; pg_index++) + free_page((unsigned long)cb->compressed_pages[pg_index]); kfree(cb->compressed_pages); fail1: @@ -945,7 +946,7 @@ void btrfs_exit_compress(void) int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, unsigned long total_out, u64 disk_start, struct bio_vec *bvec, int vcnt, - unsigned long *page_index, + unsigned long *pg_index, unsigned long *pg_offset) { unsigned long buf_offset; @@ -954,7 +955,7 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, unsigned long working_bytes = total_out - buf_start; unsigned long bytes; char *kaddr; - struct page *page_out = bvec[*page_index].bv_page; + struct page *page_out = bvec[*pg_index].bv_page; /* * start byte is the first byte of the page we're currently @@ -995,11 +996,11 @@ int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, /* check if we need to pick another page */ if (*pg_offset == PAGE_CACHE_SIZE) { - (*page_index)++; - if (*page_index >= vcnt) + (*pg_index)++; + if (*pg_index >= vcnt) return 0; - page_out = bvec[*page_index].bv_page; + page_out = bvec[*pg_index].bv_page; *pg_offset = 0; start_byte = page_offset(page_out) - disk_start; diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h index 51000174b9d7..a12059f4f0fd 100644 --- a/fs/btrfs/compression.h +++ b/fs/btrfs/compression.h @@ -37,7 +37,7 @@ int btrfs_decompress(int type, unsigned char *data_in, struct page *dest_page, int btrfs_decompress_buf2page(char *buf, unsigned long buf_start, unsigned long total_out, u64 disk_start, struct bio_vec *bvec, int vcnt, - unsigned long *page_index, + unsigned long *pg_index, unsigned long *pg_offset); int btrfs_submit_compressed_write(struct inode *inode, u64 start, diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 84d7ca1fe0ba..2e667868e0d2 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -38,18 +38,11 @@ static int balance_node_right(struct btrfs_trans_handle *trans, struct extent_buffer *src_buf); static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, int level, int slot); -static int setup_items_for_insert(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - u32 total_data, u32 total_size, int nr); - struct btrfs_path *btrfs_alloc_path(void) { struct btrfs_path *path; path = kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS); - if (path) - path->reada = 1; return path; } @@ -107,7 +100,7 @@ void btrfs_free_path(struct btrfs_path *p) { if (!p) return; - btrfs_release_path(NULL, p); + btrfs_release_path(p); kmem_cache_free(btrfs_path_cachep, p); } @@ -117,7 +110,7 @@ void btrfs_free_path(struct btrfs_path *p) * * It is safe to call this on paths that no locks or extent buffers held. */ -noinline void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p) +noinline void btrfs_release_path(struct btrfs_path *p) { int i; @@ -1229,11 +1222,13 @@ static void reada_for_search(struct btrfs_root *root, u64 search; u64 target; u64 nread = 0; + u64 gen; int direction = path->reada; struct extent_buffer *eb; u32 nr; u32 blocksize; u32 nscan = 0; + bool map = true; if (level != 1) return; @@ -1255,7 +1250,19 @@ static void reada_for_search(struct btrfs_root *root, nritems = btrfs_header_nritems(node); nr = slot; + if (node->map_token || path->skip_locking) + map = false; + while (1) { + if (map && !node->map_token) { + unsigned long offset = btrfs_node_key_ptr_offset(nr); + map_private_extent_buffer(node, offset, + sizeof(struct btrfs_key_ptr), + &node->map_token, + &node->kaddr, + &node->map_start, + &node->map_len, KM_USER1); + } if (direction < 0) { if (nr == 0) break; @@ -1273,14 +1280,23 @@ static void reada_for_search(struct btrfs_root *root, search = btrfs_node_blockptr(node, nr); if ((search <= target && target - search <= 65536) || (search > target && search - target <= 65536)) { - readahead_tree_block(root, search, blocksize, - btrfs_node_ptr_generation(node, nr)); + gen = btrfs_node_ptr_generation(node, nr); + if (map && node->map_token) { + unmap_extent_buffer(node, node->map_token, + KM_USER1); + node->map_token = NULL; + } + readahead_tree_block(root, search, blocksize, gen); nread += blocksize; } nscan++; if ((nread > 65536 || nscan > 32)) break; } + if (map && node->map_token) { + unmap_extent_buffer(node, node->map_token, KM_USER1); + node->map_token = NULL; + } } /* @@ -1328,7 +1344,7 @@ static noinline int reada_for_balance(struct btrfs_root *root, ret = -EAGAIN; /* release the whole path */ - btrfs_release_path(root, path); + btrfs_release_path(path); /* read the blocks */ if (block1) @@ -1475,7 +1491,7 @@ read_block_for_search(struct btrfs_trans_handle *trans, return 0; } free_extent_buffer(tmp); - btrfs_release_path(NULL, p); + btrfs_release_path(p); return -EIO; } } @@ -1494,7 +1510,7 @@ read_block_for_search(struct btrfs_trans_handle *trans, if (p->reada) reada_for_search(root, p, level, slot, key->objectid); - btrfs_release_path(NULL, p); + btrfs_release_path(p); ret = -EAGAIN; tmp = read_tree_block(root, blocknr, blocksize, 0); @@ -1563,7 +1579,7 @@ setup_nodes_for_search(struct btrfs_trans_handle *trans, } b = p->nodes[level]; if (!b) { - btrfs_release_path(NULL, p); + btrfs_release_path(p); goto again; } BUG_ON(btrfs_header_nritems(b) == 1); @@ -1653,9 +1669,6 @@ again: } cow_done: BUG_ON(!cow && ins_len); - if (level != btrfs_header_level(b)) - WARN_ON(1); - level = btrfs_header_level(b); p->nodes[level] = b; if (!p->skip_locking) @@ -1753,7 +1766,7 @@ done: if (!p->leave_spinning) btrfs_set_path_blocking(p); if (ret < 0) - btrfs_release_path(root, p); + btrfs_release_path(p); return ret; } @@ -3026,7 +3039,7 @@ static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans, struct btrfs_file_extent_item); extent_len = btrfs_file_extent_num_bytes(leaf, fi); } - btrfs_release_path(root, path); + btrfs_release_path(path); path->keep_locks = 1; path->search_for_split = 1; @@ -3216,7 +3229,6 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, struct btrfs_path *path, u32 new_size, int from_end) { - int ret = 0; int slot; struct extent_buffer *leaf; struct btrfs_item *item; @@ -3314,12 +3326,11 @@ int btrfs_truncate_item(struct btrfs_trans_handle *trans, btrfs_set_item_size(leaf, item, new_size); btrfs_mark_buffer_dirty(leaf); - ret = 0; if (btrfs_leaf_free_space(root, leaf) < 0) { btrfs_print_leaf(root, leaf); BUG(); } - return ret; + return 0; } /* @@ -3329,7 +3340,6 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, u32 data_size) { - int ret = 0; int slot; struct extent_buffer *leaf; struct btrfs_item *item; @@ -3394,12 +3404,11 @@ int btrfs_extend_item(struct btrfs_trans_handle *trans, btrfs_set_item_size(leaf, item, old_size + data_size); btrfs_mark_buffer_dirty(leaf); - ret = 0; if (btrfs_leaf_free_space(root, leaf) < 0) { btrfs_print_leaf(root, leaf); BUG(); } - return ret; + return 0; } /* @@ -3559,11 +3568,10 @@ out: * to save stack depth by doing the bulk of the work in a function * that doesn't call btrfs_search_slot */ -static noinline_for_stack int -setup_items_for_insert(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - u32 total_data, u32 total_size, int nr) +int setup_items_for_insert(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *cpu_key, u32 *data_size, + u32 total_data, u32 total_size, int nr) { struct btrfs_item *item; int i; @@ -3647,7 +3655,6 @@ setup_items_for_insert(struct btrfs_trans_handle *trans, ret = 0; if (slot == 0) { - struct btrfs_disk_key disk_key; btrfs_cpu_key_to_disk(&disk_key, cpu_key); ret = fixup_low_keys(trans, root, path, &disk_key, 1); } @@ -3949,7 +3956,7 @@ int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path) else return 1; - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) return ret; @@ -4073,7 +4080,7 @@ find_next_key: sret = btrfs_find_next_key(root, path, min_key, level, cache_only, min_trans); if (sret == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } else { goto out; @@ -4152,7 +4159,7 @@ next: btrfs_node_key_to_cpu(c, &cur_key, slot); orig_lowest = path->lowest_level; - btrfs_release_path(root, path); + btrfs_release_path(path); path->lowest_level = level; ret = btrfs_search_slot(NULL, root, &cur_key, path, 0, 0); @@ -4229,7 +4236,7 @@ int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) again: level = 1; next = NULL; - btrfs_release_path(root, path); + btrfs_release_path(path); path->keep_locks = 1; @@ -4285,7 +4292,7 @@ again: goto again; if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto done; } @@ -4324,7 +4331,7 @@ again: goto again; if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto done; } diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 8f4b81de3ae2..f30ac05dbda7 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -19,10 +19,10 @@ #ifndef __BTRFS_CTREE__ #define __BTRFS_CTREE__ -#include <linux/version.h> #include <linux/mm.h> #include <linux/highmem.h> #include <linux/fs.h> +#include <linux/rwsem.h> #include <linux/completion.h> #include <linux/backing-dev.h> #include <linux/wait.h> @@ -33,6 +33,7 @@ #include "extent_io.h" #include "extent_map.h" #include "async-thread.h" +#include "ioctl.h" struct btrfs_trans_handle; struct btrfs_transaction; @@ -105,6 +106,12 @@ struct btrfs_ordered_sum; /* For storing free space cache */ #define BTRFS_FREE_SPACE_OBJECTID -11ULL +/* + * The inode number assigned to the special inode for sotring + * free ino cache + */ +#define BTRFS_FREE_INO_OBJECTID -12ULL + /* dummy objectid represents multiple objectids */ #define BTRFS_MULTIPLE_OBJECTIDS -255ULL @@ -187,7 +194,6 @@ struct btrfs_mapping_tree { struct extent_map_tree map_tree; }; -#define BTRFS_UUID_SIZE 16 struct btrfs_dev_item { /* the internal btrfs device id */ __le64 devid; @@ -294,7 +300,6 @@ static inline unsigned long btrfs_chunk_item_size(int num_stripes) sizeof(struct btrfs_stripe) * (num_stripes - 1); } -#define BTRFS_FSID_SIZE 16 #define BTRFS_HEADER_FLAG_WRITTEN (1ULL << 0) #define BTRFS_HEADER_FLAG_RELOC (1ULL << 1) @@ -510,6 +515,12 @@ struct btrfs_extent_item_v0 { /* use full backrefs for extent pointers in the block */ #define BTRFS_BLOCK_FLAG_FULL_BACKREF (1ULL << 8) +/* + * this flag is only used internally by scrub and may be changed at any time + * it is only declared here to avoid collisions + */ +#define BTRFS_EXTENT_FLAG_SUPER (1ULL << 48) + struct btrfs_tree_block_info { struct btrfs_disk_key key; u8 level; @@ -740,12 +751,12 @@ struct btrfs_space_info { */ unsigned long reservation_progress; - int full:1; /* indicates that we cannot allocate any more + unsigned int full:1; /* indicates that we cannot allocate any more chunks for this space */ - int chunk_alloc:1; /* set if we are allocating a chunk */ + unsigned int chunk_alloc:1; /* set if we are allocating a chunk */ - int force_alloc; /* set if we need to force a chunk alloc for - this space */ + unsigned int force_alloc; /* set if we need to force a chunk + alloc for this space */ struct list_head list; @@ -830,9 +841,6 @@ struct btrfs_block_group_cache { u64 bytes_super; u64 flags; u64 sectorsize; - int extents_thresh; - int free_extents; - int total_bitmaps; unsigned int ro:1; unsigned int dirty:1; unsigned int iref:1; @@ -847,9 +855,7 @@ struct btrfs_block_group_cache { struct btrfs_space_info *space_info; /* free space cache stuff */ - spinlock_t tree_lock; - struct rb_root free_space_offset; - u64 free_space; + struct btrfs_free_space_ctl *free_space_ctl; /* block group cache stuff */ struct rb_node cache_node; @@ -869,6 +875,7 @@ struct btrfs_block_group_cache { struct reloc_control; struct btrfs_device; struct btrfs_fs_devices; +struct btrfs_delayed_root; struct btrfs_fs_info { u8 fsid[BTRFS_FSID_SIZE]; u8 chunk_tree_uuid[BTRFS_UUID_SIZE]; @@ -895,7 +902,10 @@ struct btrfs_fs_info { /* logical->physical extent mapping */ struct btrfs_mapping_tree mapping_tree; - /* block reservation for extent, checksum and root tree */ + /* + * block reservation for extent, checksum, root tree and + * delayed dir index item + */ struct btrfs_block_rsv global_block_rsv; /* block reservation for delay allocation */ struct btrfs_block_rsv delalloc_block_rsv; @@ -919,7 +929,6 @@ struct btrfs_fs_info { * is required instead of the faster short fsync log commits */ u64 last_trans_log_full_commit; - u64 open_ioctl_trans; unsigned long mount_opt:20; unsigned long compress_type:4; u64 max_inline; @@ -936,7 +945,6 @@ struct btrfs_fs_info { struct super_block *sb; struct inode *btree_inode; struct backing_dev_info bdi; - struct mutex trans_mutex; struct mutex tree_log_mutex; struct mutex transaction_kthread_mutex; struct mutex cleaner_mutex; @@ -957,6 +965,13 @@ struct btrfs_fs_info { struct rw_semaphore subvol_sem; struct srcu_struct subvol_srcu; + spinlock_t trans_lock; + /* + * the reloc mutex goes with the trans lock, it is taken + * during commit to protect us from the relocation code + */ + struct mutex reloc_mutex; + struct list_head trans_list; struct list_head hashers; struct list_head dead_roots; @@ -969,6 +984,7 @@ struct btrfs_fs_info { atomic_t async_submit_draining; atomic_t nr_async_bios; atomic_t async_delalloc_pages; + atomic_t open_ioctl_trans; /* * this is used by the balancing code to wait for all the pending @@ -1022,6 +1038,7 @@ struct btrfs_fs_info { * for the sys_munmap function call path */ struct btrfs_workers fixup_workers; + struct btrfs_workers delayed_workers; struct task_struct *transaction_kthread; struct task_struct *cleaner_kthread; int thread_pool_size; @@ -1032,6 +1049,7 @@ struct btrfs_fs_info { int closing; int log_root_recovering; int enospc_unlink; + int trans_no_join; u64 total_pinned; @@ -1053,7 +1071,6 @@ struct btrfs_fs_info { struct reloc_control *reloc_ctl; spinlock_t delalloc_lock; - spinlock_t new_trans_lock; u64 delalloc_bytes; /* data_alloc_cluster is only used in ssd mode */ @@ -1062,6 +1079,11 @@ struct btrfs_fs_info { /* all metadata allocations go through this cluster */ struct btrfs_free_cluster meta_alloc_cluster; + /* auto defrag inodes go here */ + spinlock_t defrag_inodes_lock; + struct rb_root defrag_inodes; + atomic_t defrag_running; + spinlock_t ref_cache_lock; u64 total_ref_cache_size; @@ -1077,8 +1099,21 @@ struct btrfs_fs_info { void *bdev_holder; + /* private scrub information */ + struct mutex scrub_lock; + atomic_t scrubs_running; + atomic_t scrub_pause_req; + atomic_t scrubs_paused; + atomic_t scrub_cancel_req; + wait_queue_head_t scrub_pause_wait; + struct rw_semaphore scrub_super_lock; + int scrub_workers_refcnt; + struct btrfs_workers scrub_workers; + /* filesystem state */ u64 fs_state; + + struct btrfs_delayed_root *delayed_root; }; /* @@ -1088,9 +1123,6 @@ struct btrfs_fs_info { struct btrfs_root { struct extent_buffer *node; - /* the node lock is held while changing the node pointer */ - spinlock_t node_lock; - struct extent_buffer *commit_root; struct btrfs_root *log_root; struct btrfs_root *reloc_root; @@ -1107,6 +1139,16 @@ struct btrfs_root { spinlock_t accounting_lock; struct btrfs_block_rsv *block_rsv; + /* free ino cache stuff */ + struct mutex fs_commit_mutex; + struct btrfs_free_space_ctl *free_ino_ctl; + enum btrfs_caching_type cached; + spinlock_t cache_lock; + wait_queue_head_t cache_wait; + struct btrfs_free_space_ctl *free_ino_pinned; + u64 cache_progress; + struct inode *cache_inode; + struct mutex log_mutex; wait_queue_head_t log_writer_wait; wait_queue_head_t log_commit_wait[2]; @@ -1135,6 +1177,14 @@ struct btrfs_root { u32 type; u64 highest_objectid; + + /* btrfs_record_root_in_trans is a multi-step process, + * and it can race with the balancing code. But the + * race is very small, and only the first time the root + * is added to each transaction. So in_trans_setup + * is used to tell us when more checks are required + */ + unsigned long in_trans_setup; int ref_cows; int track_dirty; int in_radix; @@ -1144,7 +1194,6 @@ struct btrfs_root { struct btrfs_key defrag_max; int defrag_running; char *name; - int in_sysfs; /* the dirty list is only used by non-reference counted roots */ struct list_head dirty_list; @@ -1162,12 +1211,49 @@ struct btrfs_root { struct rb_root inode_tree; /* + * radix tree that keeps track of delayed nodes of every inode, + * protected by inode_lock + */ + struct radix_tree_root delayed_nodes_tree; + /* * right now this just gets used so that a root has its own devid * for stat. It may be used for more later */ struct super_block anon_super; }; +struct btrfs_ioctl_defrag_range_args { + /* start of the defrag operation */ + __u64 start; + + /* number of bytes to defrag, use (u64)-1 to say all */ + __u64 len; + + /* + * flags for the operation, which can include turning + * on compression for this one defrag + */ + __u64 flags; + + /* + * any extent bigger than this will be considered + * already defragged. Use 0 to take the kernel default + * Use 1 to say every single extent must be rewritten + */ + __u32 extent_thresh; + + /* + * which compression method to use if turning on compression + * for this defrag operation. If unspecified, zlib will + * be used + */ + __u32 compress_type; + + /* spare for later */ + __u32 unused[4]; +}; + + /* * inode items have the data typically returned from stat and store other * info about object characteristics. There is one for every file and dir in @@ -1265,6 +1351,8 @@ struct btrfs_root { #define BTRFS_MOUNT_CLEAR_CACHE (1 << 13) #define BTRFS_MOUNT_USER_SUBVOL_RM_ALLOWED (1 << 14) #define BTRFS_MOUNT_ENOSPC_DEBUG (1 << 15) +#define BTRFS_MOUNT_AUTO_DEFRAG (1 << 16) +#define BTRFS_MOUNT_INODE_MAP_CACHE (1 << 17) #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt) #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt) @@ -1440,26 +1528,12 @@ static inline u64 btrfs_stripe_offset_nr(struct extent_buffer *eb, return btrfs_stripe_offset(eb, btrfs_stripe_nr(c, nr)); } -static inline void btrfs_set_stripe_offset_nr(struct extent_buffer *eb, - struct btrfs_chunk *c, int nr, - u64 val) -{ - btrfs_set_stripe_offset(eb, btrfs_stripe_nr(c, nr), val); -} - static inline u64 btrfs_stripe_devid_nr(struct extent_buffer *eb, struct btrfs_chunk *c, int nr) { return btrfs_stripe_devid(eb, btrfs_stripe_nr(c, nr)); } -static inline void btrfs_set_stripe_devid_nr(struct extent_buffer *eb, - struct btrfs_chunk *c, int nr, - u64 val) -{ - btrfs_set_stripe_devid(eb, btrfs_stripe_nr(c, nr), val); -} - /* struct btrfs_block_group_item */ BTRFS_SETGET_STACK_FUNCS(block_group_used, struct btrfs_block_group_item, used, 64); @@ -1517,14 +1591,6 @@ btrfs_inode_ctime(struct btrfs_inode_item *inode_item) return (struct btrfs_timespec *)ptr; } -static inline struct btrfs_timespec * -btrfs_inode_otime(struct btrfs_inode_item *inode_item) -{ - unsigned long ptr = (unsigned long)inode_item; - ptr += offsetof(struct btrfs_inode_item, otime); - return (struct btrfs_timespec *)ptr; -} - BTRFS_SETGET_FUNCS(timespec_sec, struct btrfs_timespec, sec, 64); BTRFS_SETGET_FUNCS(timespec_nsec, struct btrfs_timespec, nsec, 32); @@ -1875,33 +1941,6 @@ static inline u8 *btrfs_header_chunk_tree_uuid(struct extent_buffer *eb) return (u8 *)ptr; } -static inline u8 *btrfs_super_fsid(struct extent_buffer *eb) -{ - unsigned long ptr = offsetof(struct btrfs_super_block, fsid); - return (u8 *)ptr; -} - -static inline u8 *btrfs_header_csum(struct extent_buffer *eb) -{ - unsigned long ptr = offsetof(struct btrfs_header, csum); - return (u8 *)ptr; -} - -static inline struct btrfs_node *btrfs_buffer_node(struct extent_buffer *eb) -{ - return NULL; -} - -static inline struct btrfs_leaf *btrfs_buffer_leaf(struct extent_buffer *eb) -{ - return NULL; -} - -static inline struct btrfs_header *btrfs_buffer_header(struct extent_buffer *eb) -{ - return NULL; -} - static inline int btrfs_is_leaf(struct extent_buffer *eb) { return btrfs_header_level(eb) == 0; @@ -2055,22 +2094,6 @@ static inline struct btrfs_root *btrfs_sb(struct super_block *sb) return sb->s_fs_info; } -static inline int btrfs_set_root_name(struct btrfs_root *root, - const char *name, int len) -{ - /* if we already have a name just free it */ - kfree(root->name); - - root->name = kmalloc(len+1, GFP_KERNEL); - if (!root->name) - return -ENOMEM; - - memcpy(root->name, name, len); - root->name[len] = '\0'; - - return 0; -} - static inline u32 btrfs_level_size(struct btrfs_root *root, int level) { if (level == 0) @@ -2099,6 +2122,13 @@ static inline bool btrfs_mixed_space_info(struct btrfs_space_info *space_info) } /* extent-tree.c */ +static inline u64 btrfs_calc_trans_metadata_size(struct btrfs_root *root, + int num_items) +{ + return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * + 3 * num_items; +} + void btrfs_put_block_group(struct btrfs_block_group_cache *cache); int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, struct btrfs_root *root, unsigned long count); @@ -2108,12 +2138,9 @@ int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans, u64 num_bytes, u64 *refs, u64 *flags); int btrfs_pin_extent(struct btrfs_root *root, u64 bytenr, u64 num, int reserved); -int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, struct extent_buffer *leaf); int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 objectid, u64 offset, u64 bytenr); -int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy); struct btrfs_block_group_cache *btrfs_lookup_block_group( struct btrfs_fs_info *info, u64 bytenr); @@ -2224,6 +2251,9 @@ int btrfs_block_rsv_migrate(struct btrfs_block_rsv *src_rsv, void btrfs_block_rsv_release(struct btrfs_root *root, struct btrfs_block_rsv *block_rsv, u64 num_bytes); +int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_block_rsv *rsv); int btrfs_set_block_group_ro(struct btrfs_root *root, struct btrfs_block_group_cache *cache); int btrfs_set_block_group_rw(struct btrfs_root *root, @@ -2290,10 +2320,12 @@ int btrfs_realloc_node(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *parent, int start_slot, int cache_only, u64 *last_ret, struct btrfs_key *progress); -void btrfs_release_path(struct btrfs_root *root, struct btrfs_path *p); +void btrfs_release_path(struct btrfs_path *p); struct btrfs_path *btrfs_alloc_path(void); void btrfs_free_path(struct btrfs_path *p); void btrfs_set_path_blocking(struct btrfs_path *p); +void btrfs_clear_path_blocking(struct btrfs_path *p, + struct extent_buffer *held); void btrfs_unlock_up_safe(struct btrfs_path *p, int level); int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2305,13 +2337,12 @@ static inline int btrfs_del_item(struct btrfs_trans_handle *trans, return btrfs_del_items(trans, root, path, path->slots[0], 1); } +int setup_items_for_insert(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct btrfs_path *path, + struct btrfs_key *cpu_key, u32 *data_size, + u32 total_data, u32 total_size, int nr); int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_key *key, void *data, u32 data_size); -int btrfs_insert_some_items(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *cpu_key, u32 *data_size, - int nr); int btrfs_insert_empty_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, @@ -2335,6 +2366,15 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *node, struct extent_buffer *parent); +static inline int btrfs_fs_closing(struct btrfs_fs_info *fs_info) +{ + /* + * Get synced with close_ctree() + */ + smp_mb(); + return fs_info->closing; +} + /* root-item.c */ int btrfs_find_root_ref(struct btrfs_root *tree_root, struct btrfs_path *path, @@ -2357,8 +2397,6 @@ int btrfs_update_root(struct btrfs_trans_handle *trans, struct btrfs_root *item); int btrfs_find_last_root(struct btrfs_root *root, u64 objectid, struct btrfs_root_item *item, struct btrfs_key *key); -int btrfs_search_root(struct btrfs_root *root, u64 search_start, - u64 *found_objectid); int btrfs_find_dead_roots(struct btrfs_root *root, u64 objectid); int btrfs_find_orphan_roots(struct btrfs_root *tree_root); int btrfs_set_root_node(struct btrfs_root_item *item, @@ -2368,7 +2406,7 @@ void btrfs_check_and_init_root_item(struct btrfs_root_item *item); /* dir-item.c */ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, - int name_len, u64 dir, + int name_len, struct inode *dir, struct btrfs_key *location, u8 type, u64 index); struct btrfs_dir_item *btrfs_lookup_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2413,12 +2451,6 @@ int btrfs_del_orphan_item(struct btrfs_trans_handle *trans, struct btrfs_root *root, u64 offset); int btrfs_find_orphan_item(struct btrfs_root *root, u64 offset); -/* inode-map.c */ -int btrfs_find_free_objectid(struct btrfs_trans_handle *trans, - struct btrfs_root *fs_root, - u64 dirid, u64 *objectid); -int btrfs_find_highest_inode(struct btrfs_root *fs_root, u64 *objectid); - /* inode-item.c */ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2463,8 +2495,6 @@ int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, struct btrfs_ordered_sum *sums); int btrfs_csum_one_bio(struct btrfs_root *root, struct inode *inode, struct bio *bio, u64 file_start, int contig); -int btrfs_csum_file_bytes(struct btrfs_root *root, struct inode *inode, - u64 start, unsigned long len); struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, @@ -2472,8 +2502,8 @@ struct btrfs_csum_item *btrfs_lookup_csum(struct btrfs_trans_handle *trans, int btrfs_csum_truncate(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_path *path, u64 isize); -int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, - u64 end, struct list_head *list); +int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, + struct list_head *list, int search_commit); /* inode.c */ /* RHEL and EL kernels have a patch that renames PG_checked to FsMisc */ @@ -2502,15 +2532,12 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, u32 min_type); int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput); -int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput, - int sync); int btrfs_set_extent_delalloc(struct inode *inode, u64 start, u64 end, struct extent_state **cached_state); int btrfs_writepages(struct address_space *mapping, struct writeback_control *wbc); int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, - struct btrfs_root *new_root, - u64 new_dirid, u64 alloc_hint); + struct btrfs_root *new_root, u64 new_dirid); int btrfs_merge_bio_hook(struct page *page, unsigned long offset, size_t size, struct bio *bio, unsigned long bio_flags); @@ -2520,9 +2547,8 @@ unsigned long btrfs_force_ra(struct address_space *mapping, int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf); int btrfs_readpage(struct file *file, struct page *page); void btrfs_evict_inode(struct inode *inode); -void btrfs_put_inode(struct inode *inode); int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc); -void btrfs_dirty_inode(struct inode *inode); +void btrfs_dirty_inode(struct inode *inode, int flags); struct inode *btrfs_alloc_inode(struct super_block *sb); void btrfs_destroy_inode(struct inode *inode); int btrfs_drop_inode(struct inode *inode); @@ -2531,10 +2557,8 @@ void btrfs_destroy_cachep(void); long btrfs_ioctl_trans_end(struct file *file); struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, struct btrfs_root *root, int *was_new); -int btrfs_commit_write(struct file *file, struct page *page, - unsigned from, unsigned to); struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page, - size_t page_offset, u64 start, u64 end, + size_t pg_offset, u64 start, u64 end, int create); int btrfs_update_inode(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -2566,12 +2590,16 @@ extern const struct dentry_operations btrfs_dentry_operations; long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg); void btrfs_update_iflags(struct inode *inode); void btrfs_inherit_iflags(struct inode *inode, struct inode *dir); - +int btrfs_defrag_file(struct inode *inode, struct file *file, + struct btrfs_ioctl_defrag_range_args *range, + u64 newer_than, unsigned long max_pages); /* file.c */ +int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, + struct inode *inode); +int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info); int btrfs_sync_file(struct file *file, int datasync); int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, int skip_pinned); -int btrfs_check_file(struct btrfs_root *root, struct inode *inode); extern const struct file_operations btrfs_file_operations; int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, u64 start, u64 end, u64 *hint_byte, int drop_cache); @@ -2591,10 +2619,6 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, /* sysfs.c */ int btrfs_init_sysfs(void); void btrfs_exit_sysfs(void); -int btrfs_sysfs_add_super(struct btrfs_fs_info *fs); -int btrfs_sysfs_add_root(struct btrfs_root *root); -void btrfs_sysfs_del_root(struct btrfs_root *root); -void btrfs_sysfs_del_super(struct btrfs_fs_info *root); /* xattr.c */ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size); @@ -2637,4 +2661,18 @@ void btrfs_reloc_pre_snapshot(struct btrfs_trans_handle *trans, u64 *bytes_to_reserve); void btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending); + +/* scrub.c */ +int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, + struct btrfs_scrub_progress *progress, int readonly); +int btrfs_scrub_pause(struct btrfs_root *root); +int btrfs_scrub_pause_super(struct btrfs_root *root); +int btrfs_scrub_continue(struct btrfs_root *root); +int btrfs_scrub_continue_super(struct btrfs_root *root); +int btrfs_scrub_cancel(struct btrfs_root *root); +int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev); +int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid); +int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, + struct btrfs_scrub_progress *progress); + #endif diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c new file mode 100644 index 000000000000..98c68e658a9b --- /dev/null +++ b/fs/btrfs/delayed-inode.c @@ -0,0 +1,1773 @@ +/* + * Copyright (C) 2011 Fujitsu. All rights reserved. + * Written by Miao Xie <miaox@cn.fujitsu.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/slab.h> +#include "delayed-inode.h" +#include "disk-io.h" +#include "transaction.h" + +#define BTRFS_DELAYED_WRITEBACK 400 +#define BTRFS_DELAYED_BACKGROUND 100 + +static struct kmem_cache *delayed_node_cache; + +int __init btrfs_delayed_inode_init(void) +{ + delayed_node_cache = kmem_cache_create("delayed_node", + sizeof(struct btrfs_delayed_node), + 0, + SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, + NULL); + if (!delayed_node_cache) + return -ENOMEM; + return 0; +} + +void btrfs_delayed_inode_exit(void) +{ + if (delayed_node_cache) + kmem_cache_destroy(delayed_node_cache); +} + +static inline void btrfs_init_delayed_node( + struct btrfs_delayed_node *delayed_node, + struct btrfs_root *root, u64 inode_id) +{ + delayed_node->root = root; + delayed_node->inode_id = inode_id; + atomic_set(&delayed_node->refs, 0); + delayed_node->count = 0; + delayed_node->in_list = 0; + delayed_node->inode_dirty = 0; + delayed_node->ins_root = RB_ROOT; + delayed_node->del_root = RB_ROOT; + mutex_init(&delayed_node->mutex); + delayed_node->index_cnt = 0; + INIT_LIST_HEAD(&delayed_node->n_list); + INIT_LIST_HEAD(&delayed_node->p_list); + delayed_node->bytes_reserved = 0; +} + +static inline int btrfs_is_continuous_delayed_item( + struct btrfs_delayed_item *item1, + struct btrfs_delayed_item *item2) +{ + if (item1->key.type == BTRFS_DIR_INDEX_KEY && + item1->key.objectid == item2->key.objectid && + item1->key.type == item2->key.type && + item1->key.offset + 1 == item2->key.offset) + return 1; + return 0; +} + +static inline struct btrfs_delayed_root *btrfs_get_delayed_root( + struct btrfs_root *root) +{ + return root->fs_info->delayed_root; +} + +static struct btrfs_delayed_node *btrfs_get_delayed_node(struct inode *inode) +{ + struct btrfs_inode *btrfs_inode = BTRFS_I(inode); + struct btrfs_root *root = btrfs_inode->root; + u64 ino = btrfs_ino(inode); + struct btrfs_delayed_node *node; + + node = ACCESS_ONCE(btrfs_inode->delayed_node); + if (node) { + atomic_inc(&node->refs); + return node; + } + + spin_lock(&root->inode_lock); + node = radix_tree_lookup(&root->delayed_nodes_tree, ino); + if (node) { + if (btrfs_inode->delayed_node) { + atomic_inc(&node->refs); /* can be accessed */ + BUG_ON(btrfs_inode->delayed_node != node); + spin_unlock(&root->inode_lock); + return node; + } + btrfs_inode->delayed_node = node; + atomic_inc(&node->refs); /* can be accessed */ + atomic_inc(&node->refs); /* cached in the inode */ + spin_unlock(&root->inode_lock); + return node; + } + spin_unlock(&root->inode_lock); + + return NULL; +} + +static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node( + struct inode *inode) +{ + struct btrfs_delayed_node *node; + struct btrfs_inode *btrfs_inode = BTRFS_I(inode); + struct btrfs_root *root = btrfs_inode->root; + u64 ino = btrfs_ino(inode); + int ret; + +again: + node = btrfs_get_delayed_node(inode); + if (node) + return node; + + node = kmem_cache_alloc(delayed_node_cache, GFP_NOFS); + if (!node) + return ERR_PTR(-ENOMEM); + btrfs_init_delayed_node(node, root, ino); + + atomic_inc(&node->refs); /* cached in the btrfs inode */ + atomic_inc(&node->refs); /* can be accessed */ + + ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM); + if (ret) { + kmem_cache_free(delayed_node_cache, node); + return ERR_PTR(ret); + } + + spin_lock(&root->inode_lock); + ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node); + if (ret == -EEXIST) { + kmem_cache_free(delayed_node_cache, node); + spin_unlock(&root->inode_lock); + radix_tree_preload_end(); + goto again; + } + btrfs_inode->delayed_node = node; + spin_unlock(&root->inode_lock); + radix_tree_preload_end(); + + return node; +} + +/* + * Call it when holding delayed_node->mutex + * + * If mod = 1, add this node into the prepared list. + */ +static void btrfs_queue_delayed_node(struct btrfs_delayed_root *root, + struct btrfs_delayed_node *node, + int mod) +{ + spin_lock(&root->lock); + if (node->in_list) { + if (!list_empty(&node->p_list)) + list_move_tail(&node->p_list, &root->prepare_list); + else if (mod) + list_add_tail(&node->p_list, &root->prepare_list); + } else { + list_add_tail(&node->n_list, &root->node_list); + list_add_tail(&node->p_list, &root->prepare_list); + atomic_inc(&node->refs); /* inserted into list */ + root->nodes++; + node->in_list = 1; + } + spin_unlock(&root->lock); +} + +/* Call it when holding delayed_node->mutex */ +static void btrfs_dequeue_delayed_node(struct btrfs_delayed_root *root, + struct btrfs_delayed_node *node) +{ + spin_lock(&root->lock); + if (node->in_list) { + root->nodes--; + atomic_dec(&node->refs); /* not in the list */ + list_del_init(&node->n_list); + if (!list_empty(&node->p_list)) + list_del_init(&node->p_list); + node->in_list = 0; + } + spin_unlock(&root->lock); +} + +struct btrfs_delayed_node *btrfs_first_delayed_node( + struct btrfs_delayed_root *delayed_root) +{ + struct list_head *p; + struct btrfs_delayed_node *node = NULL; + + spin_lock(&delayed_root->lock); + if (list_empty(&delayed_root->node_list)) + goto out; + + p = delayed_root->node_list.next; + node = list_entry(p, struct btrfs_delayed_node, n_list); + atomic_inc(&node->refs); +out: + spin_unlock(&delayed_root->lock); + + return node; +} + +struct btrfs_delayed_node *btrfs_next_delayed_node( + struct btrfs_delayed_node *node) +{ + struct btrfs_delayed_root *delayed_root; + struct list_head *p; + struct btrfs_delayed_node *next = NULL; + + delayed_root = node->root->fs_info->delayed_root; + spin_lock(&delayed_root->lock); + if (!node->in_list) { /* not in the list */ + if (list_empty(&delayed_root->node_list)) + goto out; + p = delayed_root->node_list.next; + } else if (list_is_last(&node->n_list, &delayed_root->node_list)) + goto out; + else + p = node->n_list.next; + + next = list_entry(p, struct btrfs_delayed_node, n_list); + atomic_inc(&next->refs); +out: + spin_unlock(&delayed_root->lock); + + return next; +} + +static void __btrfs_release_delayed_node( + struct btrfs_delayed_node *delayed_node, + int mod) +{ + struct btrfs_delayed_root *delayed_root; + + if (!delayed_node) + return; + + delayed_root = delayed_node->root->fs_info->delayed_root; + + mutex_lock(&delayed_node->mutex); + if (delayed_node->count) + btrfs_queue_delayed_node(delayed_root, delayed_node, mod); + else + btrfs_dequeue_delayed_node(delayed_root, delayed_node); + mutex_unlock(&delayed_node->mutex); + + if (atomic_dec_and_test(&delayed_node->refs)) { + struct btrfs_root *root = delayed_node->root; + spin_lock(&root->inode_lock); + if (atomic_read(&delayed_node->refs) == 0) { + radix_tree_delete(&root->delayed_nodes_tree, + delayed_node->inode_id); + kmem_cache_free(delayed_node_cache, delayed_node); + } + spin_unlock(&root->inode_lock); + } +} + +static inline void btrfs_release_delayed_node(struct btrfs_delayed_node *node) +{ + __btrfs_release_delayed_node(node, 0); +} + +struct btrfs_delayed_node *btrfs_first_prepared_delayed_node( + struct btrfs_delayed_root *delayed_root) +{ + struct list_head *p; + struct btrfs_delayed_node *node = NULL; + + spin_lock(&delayed_root->lock); + if (list_empty(&delayed_root->prepare_list)) + goto out; + + p = delayed_root->prepare_list.next; + list_del_init(p); + node = list_entry(p, struct btrfs_delayed_node, p_list); + atomic_inc(&node->refs); +out: + spin_unlock(&delayed_root->lock); + + return node; +} + +static inline void btrfs_release_prepared_delayed_node( + struct btrfs_delayed_node *node) +{ + __btrfs_release_delayed_node(node, 1); +} + +struct btrfs_delayed_item *btrfs_alloc_delayed_item(u32 data_len) +{ + struct btrfs_delayed_item *item; + item = kmalloc(sizeof(*item) + data_len, GFP_NOFS); + if (item) { + item->data_len = data_len; + item->ins_or_del = 0; + item->bytes_reserved = 0; + item->delayed_node = NULL; + atomic_set(&item->refs, 1); + } + return item; +} + +/* + * __btrfs_lookup_delayed_item - look up the delayed item by key + * @delayed_node: pointer to the delayed node + * @key: the key to look up + * @prev: used to store the prev item if the right item isn't found + * @next: used to store the next item if the right item isn't found + * + * Note: if we don't find the right item, we will return the prev item and + * the next item. + */ +static struct btrfs_delayed_item *__btrfs_lookup_delayed_item( + struct rb_root *root, + struct btrfs_key *key, + struct btrfs_delayed_item **prev, + struct btrfs_delayed_item **next) +{ + struct rb_node *node, *prev_node = NULL; + struct btrfs_delayed_item *delayed_item = NULL; + int ret = 0; + + node = root->rb_node; + + while (node) { + delayed_item = rb_entry(node, struct btrfs_delayed_item, + rb_node); + prev_node = node; + ret = btrfs_comp_cpu_keys(&delayed_item->key, key); + if (ret < 0) + node = node->rb_right; + else if (ret > 0) + node = node->rb_left; + else + return delayed_item; + } + + if (prev) { + if (!prev_node) + *prev = NULL; + else if (ret < 0) + *prev = delayed_item; + else if ((node = rb_prev(prev_node)) != NULL) { + *prev = rb_entry(node, struct btrfs_delayed_item, + rb_node); + } else + *prev = NULL; + } + + if (next) { + if (!prev_node) + *next = NULL; + else if (ret > 0) + *next = delayed_item; + else if ((node = rb_next(prev_node)) != NULL) { + *next = rb_entry(node, struct btrfs_delayed_item, + rb_node); + } else + *next = NULL; + } + return NULL; +} + +struct btrfs_delayed_item *__btrfs_lookup_delayed_insertion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item; + + item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key, + NULL, NULL); + return item; +} + +struct btrfs_delayed_item *__btrfs_lookup_delayed_deletion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item; + + item = __btrfs_lookup_delayed_item(&delayed_node->del_root, key, + NULL, NULL); + return item; +} + +struct btrfs_delayed_item *__btrfs_search_delayed_insertion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item, *next; + + item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key, + NULL, &next); + if (!item) + item = next; + + return item; +} + +struct btrfs_delayed_item *__btrfs_search_delayed_deletion_item( + struct btrfs_delayed_node *delayed_node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item, *next; + + item = __btrfs_lookup_delayed_item(&delayed_node->del_root, key, + NULL, &next); + if (!item) + item = next; + + return item; +} + +static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node, + struct btrfs_delayed_item *ins, + int action) +{ + struct rb_node **p, *node; + struct rb_node *parent_node = NULL; + struct rb_root *root; + struct btrfs_delayed_item *item; + int cmp; + + if (action == BTRFS_DELAYED_INSERTION_ITEM) + root = &delayed_node->ins_root; + else if (action == BTRFS_DELAYED_DELETION_ITEM) + root = &delayed_node->del_root; + else + BUG(); + p = &root->rb_node; + node = &ins->rb_node; + + while (*p) { + parent_node = *p; + item = rb_entry(parent_node, struct btrfs_delayed_item, + rb_node); + + cmp = btrfs_comp_cpu_keys(&item->key, &ins->key); + if (cmp < 0) + p = &(*p)->rb_right; + else if (cmp > 0) + p = &(*p)->rb_left; + else + return -EEXIST; + } + + rb_link_node(node, parent_node, p); + rb_insert_color(node, root); + ins->delayed_node = delayed_node; + ins->ins_or_del = action; + + if (ins->key.type == BTRFS_DIR_INDEX_KEY && + action == BTRFS_DELAYED_INSERTION_ITEM && + ins->key.offset >= delayed_node->index_cnt) + delayed_node->index_cnt = ins->key.offset + 1; + + delayed_node->count++; + atomic_inc(&delayed_node->root->fs_info->delayed_root->items); + return 0; +} + +static int __btrfs_add_delayed_insertion_item(struct btrfs_delayed_node *node, + struct btrfs_delayed_item *item) +{ + return __btrfs_add_delayed_item(node, item, + BTRFS_DELAYED_INSERTION_ITEM); +} + +static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node, + struct btrfs_delayed_item *item) +{ + return __btrfs_add_delayed_item(node, item, + BTRFS_DELAYED_DELETION_ITEM); +} + +static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item) +{ + struct rb_root *root; + struct btrfs_delayed_root *delayed_root; + + delayed_root = delayed_item->delayed_node->root->fs_info->delayed_root; + + BUG_ON(!delayed_root); + BUG_ON(delayed_item->ins_or_del != BTRFS_DELAYED_DELETION_ITEM && + delayed_item->ins_or_del != BTRFS_DELAYED_INSERTION_ITEM); + + if (delayed_item->ins_or_del == BTRFS_DELAYED_INSERTION_ITEM) + root = &delayed_item->delayed_node->ins_root; + else + root = &delayed_item->delayed_node->del_root; + + rb_erase(&delayed_item->rb_node, root); + delayed_item->delayed_node->count--; + atomic_dec(&delayed_root->items); + if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND && + waitqueue_active(&delayed_root->wait)) + wake_up(&delayed_root->wait); +} + +static void btrfs_release_delayed_item(struct btrfs_delayed_item *item) +{ + if (item) { + __btrfs_remove_delayed_item(item); + if (atomic_dec_and_test(&item->refs)) + kfree(item); + } +} + +struct btrfs_delayed_item *__btrfs_first_delayed_insertion_item( + struct btrfs_delayed_node *delayed_node) +{ + struct rb_node *p; + struct btrfs_delayed_item *item = NULL; + + p = rb_first(&delayed_node->ins_root); + if (p) + item = rb_entry(p, struct btrfs_delayed_item, rb_node); + + return item; +} + +struct btrfs_delayed_item *__btrfs_first_delayed_deletion_item( + struct btrfs_delayed_node *delayed_node) +{ + struct rb_node *p; + struct btrfs_delayed_item *item = NULL; + + p = rb_first(&delayed_node->del_root); + if (p) + item = rb_entry(p, struct btrfs_delayed_item, rb_node); + + return item; +} + +struct btrfs_delayed_item *__btrfs_next_delayed_item( + struct btrfs_delayed_item *item) +{ + struct rb_node *p; + struct btrfs_delayed_item *next = NULL; + + p = rb_next(&item->rb_node); + if (p) + next = rb_entry(p, struct btrfs_delayed_item, rb_node); + + return next; +} + +static inline struct btrfs_root *btrfs_get_fs_root(struct btrfs_root *root, + u64 root_id) +{ + struct btrfs_key root_key; + + if (root->objectid == root_id) + return root; + + root_key.objectid = root_id; + root_key.type = BTRFS_ROOT_ITEM_KEY; + root_key.offset = (u64)-1; + return btrfs_read_fs_root_no_name(root->fs_info, &root_key); +} + +static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_delayed_item *item) +{ + struct btrfs_block_rsv *src_rsv; + struct btrfs_block_rsv *dst_rsv; + u64 num_bytes; + int ret; + + if (!trans->bytes_reserved) + return 0; + + src_rsv = trans->block_rsv; + dst_rsv = &root->fs_info->global_block_rsv; + + num_bytes = btrfs_calc_trans_metadata_size(root, 1); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + if (!ret) + item->bytes_reserved = num_bytes; + + return ret; +} + +static void btrfs_delayed_item_release_metadata(struct btrfs_root *root, + struct btrfs_delayed_item *item) +{ + struct btrfs_block_rsv *rsv; + + if (!item->bytes_reserved) + return; + + rsv = &root->fs_info->global_block_rsv; + btrfs_block_rsv_release(root, rsv, + item->bytes_reserved); +} + +static int btrfs_delayed_inode_reserve_metadata( + struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_block_rsv *src_rsv; + struct btrfs_block_rsv *dst_rsv; + u64 num_bytes; + int ret; + + if (!trans->bytes_reserved) + return 0; + + src_rsv = trans->block_rsv; + dst_rsv = &root->fs_info->global_block_rsv; + + num_bytes = btrfs_calc_trans_metadata_size(root, 1); + ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes); + if (!ret) + node->bytes_reserved = num_bytes; + + return ret; +} + +static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_block_rsv *rsv; + + if (!node->bytes_reserved) + return; + + rsv = &root->fs_info->global_block_rsv; + btrfs_block_rsv_release(root, rsv, + node->bytes_reserved); + node->bytes_reserved = 0; +} + +/* + * This helper will insert some continuous items into the same leaf according + * to the free space of the leaf. + */ +static int btrfs_batch_insert_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_item *item) +{ + struct btrfs_delayed_item *curr, *next; + int free_space; + int total_data_size = 0, total_size = 0; + struct extent_buffer *leaf; + char *data_ptr; + struct btrfs_key *keys; + u32 *data_size; + struct list_head head; + int slot; + int nitems; + int i; + int ret = 0; + + BUG_ON(!path->nodes[0]); + + leaf = path->nodes[0]; + free_space = btrfs_leaf_free_space(root, leaf); + INIT_LIST_HEAD(&head); + + next = item; + nitems = 0; + + /* + * count the number of the continuous items that we can insert in batch + */ + while (total_size + next->data_len + sizeof(struct btrfs_item) <= + free_space) { + total_data_size += next->data_len; + total_size += next->data_len + sizeof(struct btrfs_item); + list_add_tail(&next->tree_list, &head); + nitems++; + + curr = next; + next = __btrfs_next_delayed_item(curr); + if (!next) + break; + + if (!btrfs_is_continuous_delayed_item(curr, next)) + break; + } + + if (!nitems) { + ret = 0; + goto out; + } + + /* + * we need allocate some memory space, but it might cause the task + * to sleep, so we set all locked nodes in the path to blocking locks + * first. + */ + btrfs_set_path_blocking(path); + + keys = kmalloc(sizeof(struct btrfs_key) * nitems, GFP_NOFS); + if (!keys) { + ret = -ENOMEM; + goto out; + } + + data_size = kmalloc(sizeof(u32) * nitems, GFP_NOFS); + if (!data_size) { + ret = -ENOMEM; + goto error; + } + + /* get keys of all the delayed items */ + i = 0; + list_for_each_entry(next, &head, tree_list) { + keys[i] = next->key; + data_size[i] = next->data_len; + i++; + } + + /* reset all the locked nodes in the patch to spinning locks. */ + btrfs_clear_path_blocking(path, NULL); + + /* insert the keys of the items */ + ret = setup_items_for_insert(trans, root, path, keys, data_size, + total_data_size, total_size, nitems); + if (ret) + goto error; + + /* insert the dir index items */ + slot = path->slots[0]; + list_for_each_entry_safe(curr, next, &head, tree_list) { + data_ptr = btrfs_item_ptr(leaf, slot, char); + write_extent_buffer(leaf, &curr->data, + (unsigned long)data_ptr, + curr->data_len); + slot++; + + btrfs_delayed_item_release_metadata(root, curr); + + list_del(&curr->tree_list); + btrfs_release_delayed_item(curr); + } + +error: + kfree(data_size); + kfree(keys); +out: + return ret; +} + +/* + * This helper can just do simple insertion that needn't extend item for new + * data, such as directory name index insertion, inode insertion. + */ +static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_item *delayed_item) +{ + struct extent_buffer *leaf; + struct btrfs_item *item; + char *ptr; + int ret; + + ret = btrfs_insert_empty_item(trans, root, path, &delayed_item->key, + delayed_item->data_len); + if (ret < 0 && ret != -EEXIST) + return ret; + + leaf = path->nodes[0]; + + item = btrfs_item_nr(leaf, path->slots[0]); + ptr = btrfs_item_ptr(leaf, path->slots[0], char); + + write_extent_buffer(leaf, delayed_item->data, (unsigned long)ptr, + delayed_item->data_len); + btrfs_mark_buffer_dirty(leaf); + + btrfs_delayed_item_release_metadata(root, delayed_item); + return 0; +} + +/* + * we insert an item first, then if there are some continuous items, we try + * to insert those items into the same leaf. + */ +static int btrfs_insert_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_path *path, + struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_delayed_item *curr, *prev; + int ret = 0; + +do_again: + mutex_lock(&node->mutex); + curr = __btrfs_first_delayed_insertion_item(node); + if (!curr) + goto insert_end; + + ret = btrfs_insert_delayed_item(trans, root, path, curr); + if (ret < 0) { + btrfs_release_path(path); + goto insert_end; + } + + prev = curr; + curr = __btrfs_next_delayed_item(prev); + if (curr && btrfs_is_continuous_delayed_item(prev, curr)) { + /* insert the continuous items into the same leaf */ + path->slots[0]++; + btrfs_batch_insert_items(trans, root, path, curr); + } + btrfs_release_delayed_item(prev); + btrfs_mark_buffer_dirty(path->nodes[0]); + + btrfs_release_path(path); + mutex_unlock(&node->mutex); + goto do_again; + +insert_end: + mutex_unlock(&node->mutex); + return ret; +} + +static int btrfs_batch_delete_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_item *item) +{ + struct btrfs_delayed_item *curr, *next; + struct extent_buffer *leaf; + struct btrfs_key key; + struct list_head head; + int nitems, i, last_item; + int ret = 0; + + BUG_ON(!path->nodes[0]); + + leaf = path->nodes[0]; + + i = path->slots[0]; + last_item = btrfs_header_nritems(leaf) - 1; + if (i > last_item) + return -ENOENT; /* FIXME: Is errno suitable? */ + + next = item; + INIT_LIST_HEAD(&head); + btrfs_item_key_to_cpu(leaf, &key, i); + nitems = 0; + /* + * count the number of the dir index items that we can delete in batch + */ + while (btrfs_comp_cpu_keys(&next->key, &key) == 0) { + list_add_tail(&next->tree_list, &head); + nitems++; + + curr = next; + next = __btrfs_next_delayed_item(curr); + if (!next) + break; + + if (!btrfs_is_continuous_delayed_item(curr, next)) + break; + + i++; + if (i > last_item) + break; + btrfs_item_key_to_cpu(leaf, &key, i); + } + + if (!nitems) + return 0; + + ret = btrfs_del_items(trans, root, path, path->slots[0], nitems); + if (ret) + goto out; + + list_for_each_entry_safe(curr, next, &head, tree_list) { + btrfs_delayed_item_release_metadata(root, curr); + list_del(&curr->tree_list); + btrfs_release_delayed_item(curr); + } + +out: + return ret; +} + +static int btrfs_delete_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_path *path, + struct btrfs_root *root, + struct btrfs_delayed_node *node) +{ + struct btrfs_delayed_item *curr, *prev; + int ret = 0; + +do_again: + mutex_lock(&node->mutex); + curr = __btrfs_first_delayed_deletion_item(node); + if (!curr) + goto delete_fail; + + ret = btrfs_search_slot(trans, root, &curr->key, path, -1, 1); + if (ret < 0) + goto delete_fail; + else if (ret > 0) { + /* + * can't find the item which the node points to, so this node + * is invalid, just drop it. + */ + prev = curr; + curr = __btrfs_next_delayed_item(prev); + btrfs_release_delayed_item(prev); + ret = 0; + btrfs_release_path(path); + if (curr) + goto do_again; + else + goto delete_fail; + } + + btrfs_batch_delete_items(trans, root, path, curr); + btrfs_release_path(path); + mutex_unlock(&node->mutex); + goto do_again; + +delete_fail: + btrfs_release_path(path); + mutex_unlock(&node->mutex); + return ret; +} + +static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node) +{ + struct btrfs_delayed_root *delayed_root; + + if (delayed_node && delayed_node->inode_dirty) { + BUG_ON(!delayed_node->root); + delayed_node->inode_dirty = 0; + delayed_node->count--; + + delayed_root = delayed_node->root->fs_info->delayed_root; + atomic_dec(&delayed_root->items); + if (atomic_read(&delayed_root->items) < + BTRFS_DELAYED_BACKGROUND && + waitqueue_active(&delayed_root->wait)) + wake_up(&delayed_root->wait); + } +} + +static int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_path *path, + struct btrfs_delayed_node *node) +{ + struct btrfs_key key; + struct btrfs_inode_item *inode_item; + struct extent_buffer *leaf; + int ret; + + mutex_lock(&node->mutex); + if (!node->inode_dirty) { + mutex_unlock(&node->mutex); + return 0; + } + + key.objectid = node->inode_id; + btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); + key.offset = 0; + ret = btrfs_lookup_inode(trans, root, path, &key, 1); + if (ret > 0) { + btrfs_release_path(path); + mutex_unlock(&node->mutex); + return -ENOENT; + } else if (ret < 0) { + mutex_unlock(&node->mutex); + return ret; + } + + btrfs_unlock_up_safe(path, 1); + leaf = path->nodes[0]; + inode_item = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_inode_item); + write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item, + sizeof(struct btrfs_inode_item)); + btrfs_mark_buffer_dirty(leaf); + btrfs_release_path(path); + + btrfs_delayed_inode_release_metadata(root, node); + btrfs_release_delayed_inode(node); + mutex_unlock(&node->mutex); + + return 0; +} + +/* Called when committing the transaction. */ +int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + struct btrfs_delayed_node *curr_node, *prev_node; + struct btrfs_path *path; + struct btrfs_block_rsv *block_rsv; + int ret = 0; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->leave_spinning = 1; + + block_rsv = trans->block_rsv; + trans->block_rsv = &root->fs_info->global_block_rsv; + + delayed_root = btrfs_get_delayed_root(root); + + curr_node = btrfs_first_delayed_node(delayed_root); + while (curr_node) { + root = curr_node->root; + ret = btrfs_insert_delayed_items(trans, path, root, + curr_node); + if (!ret) + ret = btrfs_delete_delayed_items(trans, path, root, + curr_node); + if (!ret) + ret = btrfs_update_delayed_inode(trans, root, path, + curr_node); + if (ret) { + btrfs_release_delayed_node(curr_node); + break; + } + + prev_node = curr_node; + curr_node = btrfs_next_delayed_node(curr_node); + btrfs_release_delayed_node(prev_node); + } + + btrfs_free_path(path); + trans->block_rsv = block_rsv; + return ret; +} + +static int __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_delayed_node *node) +{ + struct btrfs_path *path; + struct btrfs_block_rsv *block_rsv; + int ret; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + path->leave_spinning = 1; + + block_rsv = trans->block_rsv; + trans->block_rsv = &node->root->fs_info->global_block_rsv; + + ret = btrfs_insert_delayed_items(trans, path, node->root, node); + if (!ret) + ret = btrfs_delete_delayed_items(trans, path, node->root, node); + if (!ret) + ret = btrfs_update_delayed_inode(trans, node->root, path, node); + btrfs_free_path(path); + + trans->block_rsv = block_rsv; + return ret; +} + +int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, + struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); + int ret; + + if (!delayed_node) + return 0; + + mutex_lock(&delayed_node->mutex); + if (!delayed_node->count) { + mutex_unlock(&delayed_node->mutex); + btrfs_release_delayed_node(delayed_node); + return 0; + } + mutex_unlock(&delayed_node->mutex); + + ret = __btrfs_commit_inode_delayed_items(trans, delayed_node); + btrfs_release_delayed_node(delayed_node); + return ret; +} + +void btrfs_remove_delayed_node(struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node; + + delayed_node = ACCESS_ONCE(BTRFS_I(inode)->delayed_node); + if (!delayed_node) + return; + + BTRFS_I(inode)->delayed_node = NULL; + btrfs_release_delayed_node(delayed_node); +} + +struct btrfs_async_delayed_node { + struct btrfs_root *root; + struct btrfs_delayed_node *delayed_node; + struct btrfs_work work; +}; + +static void btrfs_async_run_delayed_node_done(struct btrfs_work *work) +{ + struct btrfs_async_delayed_node *async_node; + struct btrfs_trans_handle *trans; + struct btrfs_path *path; + struct btrfs_delayed_node *delayed_node = NULL; + struct btrfs_root *root; + struct btrfs_block_rsv *block_rsv; + unsigned long nr = 0; + int need_requeue = 0; + int ret; + + async_node = container_of(work, struct btrfs_async_delayed_node, work); + + path = btrfs_alloc_path(); + if (!path) + goto out; + path->leave_spinning = 1; + + delayed_node = async_node->delayed_node; + root = delayed_node->root; + + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) + goto free_path; + + block_rsv = trans->block_rsv; + trans->block_rsv = &root->fs_info->global_block_rsv; + + ret = btrfs_insert_delayed_items(trans, path, root, delayed_node); + if (!ret) + ret = btrfs_delete_delayed_items(trans, path, root, + delayed_node); + + if (!ret) + btrfs_update_delayed_inode(trans, root, path, delayed_node); + + /* + * Maybe new delayed items have been inserted, so we need requeue + * the work. Besides that, we must dequeue the empty delayed nodes + * to avoid the race between delayed items balance and the worker. + * The race like this: + * Task1 Worker thread + * count == 0, needn't requeue + * also needn't insert the + * delayed node into prepare + * list again. + * add lots of delayed items + * queue the delayed node + * already in the list, + * and not in the prepare + * list, it means the delayed + * node is being dealt with + * by the worker. + * do delayed items balance + * the delayed node is being + * dealt with by the worker + * now, just wait. + * the worker goto idle. + * Task1 will sleep until the transaction is commited. + */ + mutex_lock(&delayed_node->mutex); + if (delayed_node->count) + need_requeue = 1; + else + btrfs_dequeue_delayed_node(root->fs_info->delayed_root, + delayed_node); + mutex_unlock(&delayed_node->mutex); + + nr = trans->blocks_used; + + trans->block_rsv = block_rsv; + btrfs_end_transaction_dmeta(trans, root); + __btrfs_btree_balance_dirty(root, nr); +free_path: + btrfs_free_path(path); +out: + if (need_requeue) + btrfs_requeue_work(&async_node->work); + else { + btrfs_release_prepared_delayed_node(delayed_node); + kfree(async_node); + } +} + +static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root, + struct btrfs_root *root, int all) +{ + struct btrfs_async_delayed_node *async_node; + struct btrfs_delayed_node *curr; + int count = 0; + +again: + curr = btrfs_first_prepared_delayed_node(delayed_root); + if (!curr) + return 0; + + async_node = kmalloc(sizeof(*async_node), GFP_NOFS); + if (!async_node) { + btrfs_release_prepared_delayed_node(curr); + return -ENOMEM; + } + + async_node->root = root; + async_node->delayed_node = curr; + + async_node->work.func = btrfs_async_run_delayed_node_done; + async_node->work.flags = 0; + + btrfs_queue_worker(&root->fs_info->delayed_workers, &async_node->work); + count++; + + if (all || count < 4) + goto again; + + return 0; +} + +void btrfs_assert_delayed_root_empty(struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + delayed_root = btrfs_get_delayed_root(root); + WARN_ON(btrfs_first_delayed_node(delayed_root)); +} + +void btrfs_balance_delayed_items(struct btrfs_root *root) +{ + struct btrfs_delayed_root *delayed_root; + + delayed_root = btrfs_get_delayed_root(root); + + if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND) + return; + + if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) { + int ret; + ret = btrfs_wq_run_delayed_node(delayed_root, root, 1); + if (ret) + return; + + wait_event_interruptible_timeout( + delayed_root->wait, + (atomic_read(&delayed_root->items) < + BTRFS_DELAYED_BACKGROUND), + HZ); + return; + } + + btrfs_wq_run_delayed_node(delayed_root, root, 0); +} + +int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, const char *name, + int name_len, struct inode *dir, + struct btrfs_disk_key *disk_key, u8 type, + u64 index) +{ + struct btrfs_delayed_node *delayed_node; + struct btrfs_delayed_item *delayed_item; + struct btrfs_dir_item *dir_item; + int ret; + + delayed_node = btrfs_get_or_create_delayed_node(dir); + if (IS_ERR(delayed_node)) + return PTR_ERR(delayed_node); + + delayed_item = btrfs_alloc_delayed_item(sizeof(*dir_item) + name_len); + if (!delayed_item) { + ret = -ENOMEM; + goto release_node; + } + + ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item); + /* + * we have reserved enough space when we start a new transaction, + * so reserving metadata failure is impossible + */ + BUG_ON(ret); + + delayed_item->key.objectid = btrfs_ino(dir); + btrfs_set_key_type(&delayed_item->key, BTRFS_DIR_INDEX_KEY); + delayed_item->key.offset = index; + + dir_item = (struct btrfs_dir_item *)delayed_item->data; + dir_item->location = *disk_key; + dir_item->transid = cpu_to_le64(trans->transid); + dir_item->data_len = 0; + dir_item->name_len = cpu_to_le16(name_len); + dir_item->type = type; + memcpy((char *)(dir_item + 1), name, name_len); + + mutex_lock(&delayed_node->mutex); + ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item); + if (unlikely(ret)) { + printk(KERN_ERR "err add delayed dir index item(name: %s) into " + "the insertion tree of the delayed node" + "(root id: %llu, inode id: %llu, errno: %d)\n", + name, + (unsigned long long)delayed_node->root->objectid, + (unsigned long long)delayed_node->inode_id, + ret); + BUG(); + } + mutex_unlock(&delayed_node->mutex); + +release_node: + btrfs_release_delayed_node(delayed_node); + return ret; +} + +static int btrfs_delete_delayed_insertion_item(struct btrfs_root *root, + struct btrfs_delayed_node *node, + struct btrfs_key *key) +{ + struct btrfs_delayed_item *item; + + mutex_lock(&node->mutex); + item = __btrfs_lookup_delayed_insertion_item(node, key); + if (!item) { + mutex_unlock(&node->mutex); + return 1; + } + + btrfs_delayed_item_release_metadata(root, item); + btrfs_release_delayed_item(item); + mutex_unlock(&node->mutex); + return 0; +} + +int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *dir, + u64 index) +{ + struct btrfs_delayed_node *node; + struct btrfs_delayed_item *item; + struct btrfs_key item_key; + int ret; + + node = btrfs_get_or_create_delayed_node(dir); + if (IS_ERR(node)) + return PTR_ERR(node); + + item_key.objectid = btrfs_ino(dir); + btrfs_set_key_type(&item_key, BTRFS_DIR_INDEX_KEY); + item_key.offset = index; + + ret = btrfs_delete_delayed_insertion_item(root, node, &item_key); + if (!ret) + goto end; + + item = btrfs_alloc_delayed_item(0); + if (!item) { + ret = -ENOMEM; + goto end; + } + + item->key = item_key; + + ret = btrfs_delayed_item_reserve_metadata(trans, root, item); + /* + * we have reserved enough space when we start a new transaction, + * so reserving metadata failure is impossible. + */ + BUG_ON(ret); + + mutex_lock(&node->mutex); + ret = __btrfs_add_delayed_deletion_item(node, item); + if (unlikely(ret)) { + printk(KERN_ERR "err add delayed dir index item(index: %llu) " + "into the deletion tree of the delayed node" + "(root id: %llu, inode id: %llu, errno: %d)\n", + (unsigned long long)index, + (unsigned long long)node->root->objectid, + (unsigned long long)node->inode_id, + ret); + BUG(); + } + mutex_unlock(&node->mutex); +end: + btrfs_release_delayed_node(node); + return ret; +} + +int btrfs_inode_delayed_dir_index_count(struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode); + + if (!delayed_node) + return -ENOENT; + + /* + * Since we have held i_mutex of this directory, it is impossible that + * a new directory index is added into the delayed node and index_cnt + * is updated now. So we needn't lock the delayed node. + */ + if (!delayed_node->index_cnt) { + btrfs_release_delayed_node(delayed_node); + return -EINVAL; + } + + BTRFS_I(inode)->index_cnt = delayed_node->index_cnt; + btrfs_release_delayed_node(delayed_node); + return 0; +} + +void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, + struct list_head *del_list) +{ + struct btrfs_delayed_node *delayed_node; + struct btrfs_delayed_item *item; + + delayed_node = btrfs_get_delayed_node(inode); + if (!delayed_node) + return; + + mutex_lock(&delayed_node->mutex); + item = __btrfs_first_delayed_insertion_item(delayed_node); + while (item) { + atomic_inc(&item->refs); + list_add_tail(&item->readdir_list, ins_list); + item = __btrfs_next_delayed_item(item); + } + + item = __btrfs_first_delayed_deletion_item(delayed_node); + while (item) { + atomic_inc(&item->refs); + list_add_tail(&item->readdir_list, del_list); + item = __btrfs_next_delayed_item(item); + } + mutex_unlock(&delayed_node->mutex); + /* + * This delayed node is still cached in the btrfs inode, so refs + * must be > 1 now, and we needn't check it is going to be freed + * or not. + * + * Besides that, this function is used to read dir, we do not + * insert/delete delayed items in this period. So we also needn't + * requeue or dequeue this delayed node. + */ + atomic_dec(&delayed_node->refs); +} + +void btrfs_put_delayed_items(struct list_head *ins_list, + struct list_head *del_list) +{ + struct btrfs_delayed_item *curr, *next; + + list_for_each_entry_safe(curr, next, ins_list, readdir_list) { + list_del(&curr->readdir_list); + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + } + + list_for_each_entry_safe(curr, next, del_list, readdir_list) { + list_del(&curr->readdir_list); + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + } +} + +int btrfs_should_delete_dir_index(struct list_head *del_list, + u64 index) +{ + struct btrfs_delayed_item *curr, *next; + int ret; + + if (list_empty(del_list)) + return 0; + + list_for_each_entry_safe(curr, next, del_list, readdir_list) { + if (curr->key.offset > index) + break; + + list_del(&curr->readdir_list); + ret = (curr->key.offset == index); + + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + + if (ret) + return 1; + else + continue; + } + return 0; +} + +/* + * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree + * + */ +int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent, + filldir_t filldir, + struct list_head *ins_list) +{ + struct btrfs_dir_item *di; + struct btrfs_delayed_item *curr, *next; + struct btrfs_key location; + char *name; + int name_len; + int over = 0; + unsigned char d_type; + + if (list_empty(ins_list)) + return 0; + + /* + * Changing the data of the delayed item is impossible. So + * we needn't lock them. And we have held i_mutex of the + * directory, nobody can delete any directory indexes now. + */ + list_for_each_entry_safe(curr, next, ins_list, readdir_list) { + list_del(&curr->readdir_list); + + if (curr->key.offset < filp->f_pos) { + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + continue; + } + + filp->f_pos = curr->key.offset; + + di = (struct btrfs_dir_item *)curr->data; + name = (char *)(di + 1); + name_len = le16_to_cpu(di->name_len); + + d_type = btrfs_filetype_table[di->type]; + btrfs_disk_key_to_cpu(&location, &di->location); + + over = filldir(dirent, name, name_len, curr->key.offset, + location.objectid, d_type); + + if (atomic_dec_and_test(&curr->refs)) + kfree(curr); + + if (over) + return 1; + } + return 0; +} + +BTRFS_SETGET_STACK_FUNCS(stack_inode_generation, struct btrfs_inode_item, + generation, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_sequence, struct btrfs_inode_item, + sequence, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_transid, struct btrfs_inode_item, + transid, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_size, struct btrfs_inode_item, size, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_nbytes, struct btrfs_inode_item, + nbytes, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_block_group, struct btrfs_inode_item, + block_group, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_nlink, struct btrfs_inode_item, nlink, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_uid, struct btrfs_inode_item, uid, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_gid, struct btrfs_inode_item, gid, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_mode, struct btrfs_inode_item, mode, 32); +BTRFS_SETGET_STACK_FUNCS(stack_inode_rdev, struct btrfs_inode_item, rdev, 64); +BTRFS_SETGET_STACK_FUNCS(stack_inode_flags, struct btrfs_inode_item, flags, 64); + +BTRFS_SETGET_STACK_FUNCS(stack_timespec_sec, struct btrfs_timespec, sec, 64); +BTRFS_SETGET_STACK_FUNCS(stack_timespec_nsec, struct btrfs_timespec, nsec, 32); + +static void fill_stack_inode_item(struct btrfs_trans_handle *trans, + struct btrfs_inode_item *inode_item, + struct inode *inode) +{ + btrfs_set_stack_inode_uid(inode_item, inode->i_uid); + btrfs_set_stack_inode_gid(inode_item, inode->i_gid); + btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size); + btrfs_set_stack_inode_mode(inode_item, inode->i_mode); + btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink); + btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode)); + btrfs_set_stack_inode_generation(inode_item, + BTRFS_I(inode)->generation); + btrfs_set_stack_inode_sequence(inode_item, BTRFS_I(inode)->sequence); + btrfs_set_stack_inode_transid(inode_item, trans->transid); + btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev); + btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags); + btrfs_set_stack_inode_block_group(inode_item, 0); + + btrfs_set_stack_timespec_sec(btrfs_inode_atime(inode_item), + inode->i_atime.tv_sec); + btrfs_set_stack_timespec_nsec(btrfs_inode_atime(inode_item), + inode->i_atime.tv_nsec); + + btrfs_set_stack_timespec_sec(btrfs_inode_mtime(inode_item), + inode->i_mtime.tv_sec); + btrfs_set_stack_timespec_nsec(btrfs_inode_mtime(inode_item), + inode->i_mtime.tv_nsec); + + btrfs_set_stack_timespec_sec(btrfs_inode_ctime(inode_item), + inode->i_ctime.tv_sec); + btrfs_set_stack_timespec_nsec(btrfs_inode_ctime(inode_item), + inode->i_ctime.tv_nsec); +} + +int btrfs_fill_inode(struct inode *inode, u32 *rdev) +{ + struct btrfs_delayed_node *delayed_node; + struct btrfs_inode_item *inode_item; + struct btrfs_timespec *tspec; + + delayed_node = btrfs_get_delayed_node(inode); + if (!delayed_node) + return -ENOENT; + + mutex_lock(&delayed_node->mutex); + if (!delayed_node->inode_dirty) { + mutex_unlock(&delayed_node->mutex); + btrfs_release_delayed_node(delayed_node); + return -ENOENT; + } + + inode_item = &delayed_node->inode_item; + + inode->i_uid = btrfs_stack_inode_uid(inode_item); + inode->i_gid = btrfs_stack_inode_gid(inode_item); + btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item)); + inode->i_mode = btrfs_stack_inode_mode(inode_item); + inode->i_nlink = btrfs_stack_inode_nlink(inode_item); + inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item)); + BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item); + BTRFS_I(inode)->sequence = btrfs_stack_inode_sequence(inode_item); + inode->i_rdev = 0; + *rdev = btrfs_stack_inode_rdev(inode_item); + BTRFS_I(inode)->flags = btrfs_stack_inode_flags(inode_item); + + tspec = btrfs_inode_atime(inode_item); + inode->i_atime.tv_sec = btrfs_stack_timespec_sec(tspec); + inode->i_atime.tv_nsec = btrfs_stack_timespec_nsec(tspec); + + tspec = btrfs_inode_mtime(inode_item); + inode->i_mtime.tv_sec = btrfs_stack_timespec_sec(tspec); + inode->i_mtime.tv_nsec = btrfs_stack_timespec_nsec(tspec); + + tspec = btrfs_inode_ctime(inode_item); + inode->i_ctime.tv_sec = btrfs_stack_timespec_sec(tspec); + inode->i_ctime.tv_nsec = btrfs_stack_timespec_nsec(tspec); + + inode->i_generation = BTRFS_I(inode)->generation; + BTRFS_I(inode)->index_cnt = (u64)-1; + + mutex_unlock(&delayed_node->mutex); + btrfs_release_delayed_node(delayed_node); + return 0; +} + +int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node; + int ret = 0; + + delayed_node = btrfs_get_or_create_delayed_node(inode); + if (IS_ERR(delayed_node)) + return PTR_ERR(delayed_node); + + mutex_lock(&delayed_node->mutex); + if (delayed_node->inode_dirty) { + fill_stack_inode_item(trans, &delayed_node->inode_item, inode); + goto release_node; + } + + ret = btrfs_delayed_inode_reserve_metadata(trans, root, delayed_node); + /* + * we must reserve enough space when we start a new transaction, + * so reserving metadata failure is impossible + */ + BUG_ON(ret); + + fill_stack_inode_item(trans, &delayed_node->inode_item, inode); + delayed_node->inode_dirty = 1; + delayed_node->count++; + atomic_inc(&root->fs_info->delayed_root->items); +release_node: + mutex_unlock(&delayed_node->mutex); + btrfs_release_delayed_node(delayed_node); + return ret; +} + +static void __btrfs_kill_delayed_node(struct btrfs_delayed_node *delayed_node) +{ + struct btrfs_root *root = delayed_node->root; + struct btrfs_delayed_item *curr_item, *prev_item; + + mutex_lock(&delayed_node->mutex); + curr_item = __btrfs_first_delayed_insertion_item(delayed_node); + while (curr_item) { + btrfs_delayed_item_release_metadata(root, curr_item); + prev_item = curr_item; + curr_item = __btrfs_next_delayed_item(prev_item); + btrfs_release_delayed_item(prev_item); + } + + curr_item = __btrfs_first_delayed_deletion_item(delayed_node); + while (curr_item) { + btrfs_delayed_item_release_metadata(root, curr_item); + prev_item = curr_item; + curr_item = __btrfs_next_delayed_item(prev_item); + btrfs_release_delayed_item(prev_item); + } + + if (delayed_node->inode_dirty) { + btrfs_delayed_inode_release_metadata(root, delayed_node); + btrfs_release_delayed_inode(delayed_node); + } + mutex_unlock(&delayed_node->mutex); +} + +void btrfs_kill_delayed_inode_items(struct inode *inode) +{ + struct btrfs_delayed_node *delayed_node; + + delayed_node = btrfs_get_delayed_node(inode); + if (!delayed_node) + return; + + __btrfs_kill_delayed_node(delayed_node); + btrfs_release_delayed_node(delayed_node); +} + +void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) +{ + u64 inode_id = 0; + struct btrfs_delayed_node *delayed_nodes[8]; + int i, n; + + while (1) { + spin_lock(&root->inode_lock); + n = radix_tree_gang_lookup(&root->delayed_nodes_tree, + (void **)delayed_nodes, inode_id, + ARRAY_SIZE(delayed_nodes)); + if (!n) { + spin_unlock(&root->inode_lock); + break; + } + + inode_id = delayed_nodes[n - 1]->inode_id + 1; + + for (i = 0; i < n; i++) + atomic_inc(&delayed_nodes[i]->refs); + spin_unlock(&root->inode_lock); + + for (i = 0; i < n; i++) { + __btrfs_kill_delayed_node(delayed_nodes[i]); + btrfs_release_delayed_node(delayed_nodes[i]); + } + } +} diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h new file mode 100644 index 000000000000..8d27af4bd8b9 --- /dev/null +++ b/fs/btrfs/delayed-inode.h @@ -0,0 +1,145 @@ +/* + * Copyright (C) 2011 Fujitsu. All rights reserved. + * Written by Miao Xie <miaox@cn.fujitsu.com> + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#ifndef __DELAYED_TREE_OPERATION_H +#define __DELAYED_TREE_OPERATION_H + +#include <linux/rbtree.h> +#include <linux/spinlock.h> +#include <linux/mutex.h> +#include <linux/list.h> +#include <linux/wait.h> +#include <asm/atomic.h> + +#include "ctree.h" + +/* types of the delayed item */ +#define BTRFS_DELAYED_INSERTION_ITEM 1 +#define BTRFS_DELAYED_DELETION_ITEM 2 + +struct btrfs_delayed_root { + spinlock_t lock; + struct list_head node_list; + /* + * Used for delayed nodes which is waiting to be dealt with by the + * worker. If the delayed node is inserted into the work queue, we + * drop it from this list. + */ + struct list_head prepare_list; + atomic_t items; /* for delayed items */ + int nodes; /* for delayed nodes */ + wait_queue_head_t wait; +}; + +struct btrfs_delayed_node { + u64 inode_id; + u64 bytes_reserved; + struct btrfs_root *root; + /* Used to add the node into the delayed root's node list. */ + struct list_head n_list; + /* + * Used to add the node into the prepare list, the nodes in this list + * is waiting to be dealt with by the async worker. + */ + struct list_head p_list; + struct rb_root ins_root; + struct rb_root del_root; + struct mutex mutex; + struct btrfs_inode_item inode_item; + atomic_t refs; + u64 index_cnt; + bool in_list; + bool inode_dirty; + int count; +}; + +struct btrfs_delayed_item { + struct rb_node rb_node; + struct btrfs_key key; + struct list_head tree_list; /* used for batch insert/delete items */ + struct list_head readdir_list; /* used for readdir items */ + u64 bytes_reserved; + struct btrfs_delayed_node *delayed_node; + atomic_t refs; + int ins_or_del; + u32 data_len; + char data[0]; +}; + +static inline void btrfs_init_delayed_root( + struct btrfs_delayed_root *delayed_root) +{ + atomic_set(&delayed_root->items, 0); + delayed_root->nodes = 0; + spin_lock_init(&delayed_root->lock); + init_waitqueue_head(&delayed_root->wait); + INIT_LIST_HEAD(&delayed_root->node_list); + INIT_LIST_HEAD(&delayed_root->prepare_list); +} + +int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, const char *name, + int name_len, struct inode *dir, + struct btrfs_disk_key *disk_key, u8 type, + u64 index); + +int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *dir, + u64 index); + +int btrfs_inode_delayed_dir_index_count(struct inode *inode); + +int btrfs_run_delayed_items(struct btrfs_trans_handle *trans, + struct btrfs_root *root); + +void btrfs_balance_delayed_items(struct btrfs_root *root); + +int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans, + struct inode *inode); +/* Used for evicting the inode. */ +void btrfs_remove_delayed_node(struct inode *inode); +void btrfs_kill_delayed_inode_items(struct inode *inode); + + +int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, struct inode *inode); +int btrfs_fill_inode(struct inode *inode, u32 *rdev); + +/* Used for drop dead root */ +void btrfs_kill_all_delayed_nodes(struct btrfs_root *root); + +/* Used for readdir() */ +void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list, + struct list_head *del_list); +void btrfs_put_delayed_items(struct list_head *ins_list, + struct list_head *del_list); +int btrfs_should_delete_dir_index(struct list_head *del_list, + u64 index); +int btrfs_readdir_delayed_dir_index(struct file *filp, void *dirent, + filldir_t filldir, + struct list_head *ins_list); + +/* for init */ +int __init btrfs_delayed_inode_init(void); +void btrfs_delayed_inode_exit(void); + +/* for debugging */ +void btrfs_assert_delayed_root_empty(struct btrfs_root *root); + +#endif diff --git a/fs/btrfs/delayed-ref.c b/fs/btrfs/delayed-ref.c index bce28f653899..125cf76fcd08 100644 --- a/fs/btrfs/delayed-ref.c +++ b/fs/btrfs/delayed-ref.c @@ -281,44 +281,6 @@ again: } /* - * This checks to see if there are any delayed refs in the - * btree for a given bytenr. It returns one if it finds any - * and zero otherwise. - * - * If it only finds a head node, it returns 0. - * - * The idea is to use this when deciding if you can safely delete an - * extent from the extent allocation tree. There may be a pending - * ref in the rbtree that adds or removes references, so as long as this - * returns one you need to leave the BTRFS_EXTENT_ITEM in the extent - * allocation tree. - */ -int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr) -{ - struct btrfs_delayed_ref_node *ref; - struct btrfs_delayed_ref_root *delayed_refs; - struct rb_node *prev_node; - int ret = 0; - - delayed_refs = &trans->transaction->delayed_refs; - spin_lock(&delayed_refs->lock); - - ref = find_ref_head(&delayed_refs->root, bytenr, NULL); - if (ref) { - prev_node = rb_prev(&ref->rb_node); - if (!prev_node) - goto out; - ref = rb_entry(prev_node, struct btrfs_delayed_ref_node, - rb_node); - if (ref->bytenr == bytenr) - ret = 1; - } -out: - spin_unlock(&delayed_refs->lock); - return ret; -} - -/* * helper function to update an extent delayed ref in the * rbtree. existing and update must both have the same * bytenr and parent @@ -747,79 +709,3 @@ btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr) return btrfs_delayed_node_to_head(ref); return NULL; } - -/* - * add a delayed ref to the tree. This does all of the accounting required - * to make sure the delayed ref is eventually processed before this - * transaction commits. - * - * The main point of this call is to add and remove a backreference in a single - * shot, taking the lock only once, and only searching for the head node once. - * - * It is the same as doing a ref add and delete in two separate calls. - */ -#if 0 -int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans, - u64 bytenr, u64 num_bytes, u64 orig_parent, - u64 parent, u64 orig_ref_root, u64 ref_root, - u64 orig_ref_generation, u64 ref_generation, - u64 owner_objectid, int pin) -{ - struct btrfs_delayed_ref *ref; - struct btrfs_delayed_ref *old_ref; - struct btrfs_delayed_ref_head *head_ref; - struct btrfs_delayed_ref_root *delayed_refs; - int ret; - - ref = kmalloc(sizeof(*ref), GFP_NOFS); - if (!ref) - return -ENOMEM; - - old_ref = kmalloc(sizeof(*old_ref), GFP_NOFS); - if (!old_ref) { - kfree(ref); - return -ENOMEM; - } - - /* - * the parent = 0 case comes from cases where we don't actually - * know the parent yet. It will get updated later via a add/drop - * pair. - */ - if (parent == 0) - parent = bytenr; - if (orig_parent == 0) - orig_parent = bytenr; - - head_ref = kmalloc(sizeof(*head_ref), GFP_NOFS); - if (!head_ref) { - kfree(ref); - kfree(old_ref); - return -ENOMEM; - } - delayed_refs = &trans->transaction->delayed_refs; - spin_lock(&delayed_refs->lock); - - /* - * insert both the head node and the new ref without dropping - * the spin lock - */ - ret = __btrfs_add_delayed_ref(trans, &head_ref->node, bytenr, num_bytes, - (u64)-1, 0, 0, 0, - BTRFS_UPDATE_DELAYED_HEAD, 0); - BUG_ON(ret); - - ret = __btrfs_add_delayed_ref(trans, &ref->node, bytenr, num_bytes, - parent, ref_root, ref_generation, - owner_objectid, BTRFS_ADD_DELAYED_REF, 0); - BUG_ON(ret); - - ret = __btrfs_add_delayed_ref(trans, &old_ref->node, bytenr, num_bytes, - orig_parent, orig_ref_root, - orig_ref_generation, owner_objectid, - BTRFS_DROP_DELAYED_REF, pin); - BUG_ON(ret); - spin_unlock(&delayed_refs->lock); - return 0; -} -#endif diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h index 50e3cf92fbda..e287e3b0eab0 100644 --- a/fs/btrfs/delayed-ref.h +++ b/fs/btrfs/delayed-ref.h @@ -166,12 +166,6 @@ int btrfs_add_delayed_extent_op(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_head * btrfs_find_delayed_ref_head(struct btrfs_trans_handle *trans, u64 bytenr); -int btrfs_delayed_ref_pending(struct btrfs_trans_handle *trans, u64 bytenr); -int btrfs_update_delayed_ref(struct btrfs_trans_handle *trans, - u64 bytenr, u64 num_bytes, u64 orig_parent, - u64 parent, u64 orig_ref_root, u64 ref_root, - u64 orig_ref_generation, u64 ref_generation, - u64 owner_objectid, int pin); int btrfs_delayed_ref_lock(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_head *head); int btrfs_find_ref_cluster(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/dir-item.c b/fs/btrfs/dir-item.c index c62f02f6ae69..685f2593c4f0 100644 --- a/fs/btrfs/dir-item.c +++ b/fs/btrfs/dir-item.c @@ -50,7 +50,6 @@ static struct btrfs_dir_item *insert_with_overflow(struct btrfs_trans_handle if (di) return ERR_PTR(-EEXIST); ret = btrfs_extend_item(trans, root, path, data_size); - WARN_ON(ret > 0); } if (ret < 0) return ERR_PTR(ret); @@ -124,8 +123,9 @@ int btrfs_insert_xattr_item(struct btrfs_trans_handle *trans, * to use for the second index (if one is created). */ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root - *root, const char *name, int name_len, u64 dir, - struct btrfs_key *location, u8 type, u64 index) + *root, const char *name, int name_len, + struct inode *dir, struct btrfs_key *location, + u8 type, u64 index) { int ret = 0; int ret2 = 0; @@ -137,13 +137,17 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root struct btrfs_disk_key disk_key; u32 data_size; - key.objectid = dir; + key.objectid = btrfs_ino(dir); btrfs_set_key_type(&key, BTRFS_DIR_ITEM_KEY); key.offset = btrfs_name_hash(name, name_len); path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; path->leave_spinning = 1; + btrfs_cpu_key_to_disk(&disk_key, location); + data_size = sizeof(*dir_item) + name_len; dir_item = insert_with_overflow(trans, root, path, &key, data_size, name, name_len); @@ -155,7 +159,6 @@ int btrfs_insert_dir_item(struct btrfs_trans_handle *trans, struct btrfs_root } leaf = path->nodes[0]; - btrfs_cpu_key_to_disk(&disk_key, location); btrfs_set_dir_item_key(leaf, dir_item, &disk_key); btrfs_set_dir_type(leaf, dir_item, type); btrfs_set_dir_data_len(leaf, dir_item, 0); @@ -172,29 +175,11 @@ second_insert: ret = 0; goto out_free; } - btrfs_release_path(root, path); - - btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY); - key.offset = index; - dir_item = insert_with_overflow(trans, root, path, &key, data_size, - name, name_len); - if (IS_ERR(dir_item)) { - ret2 = PTR_ERR(dir_item); - goto out_free; - } - leaf = path->nodes[0]; - btrfs_cpu_key_to_disk(&disk_key, location); - btrfs_set_dir_item_key(leaf, dir_item, &disk_key); - btrfs_set_dir_type(leaf, dir_item, type); - btrfs_set_dir_data_len(leaf, dir_item, 0); - btrfs_set_dir_name_len(leaf, dir_item, name_len); - btrfs_set_dir_transid(leaf, dir_item, trans->transid); - name_ptr = (unsigned long)(dir_item + 1); - write_extent_buffer(leaf, name, name_ptr, name_len); - btrfs_mark_buffer_dirty(leaf); + btrfs_release_path(path); + ret2 = btrfs_insert_delayed_dir_index(trans, root, name, name_len, dir, + &disk_key, type, index); out_free: - btrfs_free_path(path); if (ret) return ret; @@ -452,7 +437,7 @@ int verify_dir_item(struct btrfs_root *root, namelen = XATTR_NAME_MAX; if (btrfs_dir_name_len(leaf, dir_item) > namelen) { - printk(KERN_CRIT "btrfS: invalid dir item name len: %u\n", + printk(KERN_CRIT "btrfs: invalid dir item name len: %u\n", (unsigned)btrfs_dir_data_len(leaf, dir_item)); return 1; } diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 228cf36ece83..1ac8db5dc0a3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -29,6 +29,7 @@ #include <linux/crc32c.h> #include <linux/slab.h> #include <linux/migrate.h> +#include <linux/ratelimit.h> #include <asm/unaligned.h> #include "compat.h" #include "ctree.h" @@ -41,6 +42,7 @@ #include "locking.h" #include "tree-log.h" #include "free-space-cache.h" +#include "inode-map.h" static struct extent_io_ops btree_extent_io_ops; static void end_workqueue_fn(struct btrfs_work *work); @@ -137,7 +139,7 @@ static const char *btrfs_eb_name[BTRFS_MAX_LEVEL + 1] = { * that covers the entire device */ static struct extent_map *btree_get_extent(struct inode *inode, - struct page *page, size_t page_offset, u64 start, u64 len, + struct page *page, size_t pg_offset, u64 start, u64 len, int create) { struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; @@ -154,7 +156,7 @@ static struct extent_map *btree_get_extent(struct inode *inode, } read_unlock(&em_tree->lock); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { em = ERR_PTR(-ENOMEM); goto out; @@ -254,14 +256,12 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf, memcpy(&found, result, csum_size); read_extent_buffer(buf, &val, 0, csum_size); - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs: %s checksum verify " + printk_ratelimited(KERN_INFO "btrfs: %s checksum verify " "failed on %llu wanted %X found %X " "level %d\n", root->fs_info->sb->s_id, (unsigned long long)buf->start, val, found, btrfs_header_level(buf)); - } if (result != (char *)&inline_result) kfree(result); return 1; @@ -296,13 +296,11 @@ static int verify_parent_transid(struct extent_io_tree *io_tree, ret = 0; goto out; } - if (printk_ratelimit()) { - printk("parent transid verify failed on %llu wanted %llu " + printk_ratelimited("parent transid verify failed on %llu wanted %llu " "found %llu\n", (unsigned long long)eb->start, (unsigned long long)parent_transid, (unsigned long long)btrfs_header_generation(eb)); - } ret = 1; clear_extent_buffer_uptodate(io_tree, eb, &cached_state); out: @@ -380,7 +378,7 @@ static int csum_dirty_buffer(struct btrfs_root *root, struct page *page) len = page->private >> 2; WARN_ON(len == 0); - eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS); + eb = alloc_extent_buffer(tree, start, len, page); if (eb == NULL) { WARN_ON(1); goto out; @@ -525,7 +523,7 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, len = page->private >> 2; WARN_ON(len == 0); - eb = alloc_extent_buffer(tree, start, len, page, GFP_NOFS); + eb = alloc_extent_buffer(tree, start, len, page); if (eb == NULL) { ret = -EIO; goto out; @@ -533,12 +531,10 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, found_start = btrfs_header_bytenr(eb); if (found_start != start) { - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs bad tree block start " + printk_ratelimited(KERN_INFO "btrfs bad tree block start " "%llu %llu\n", (unsigned long long)found_start, (unsigned long long)eb->start); - } ret = -EIO; goto err; } @@ -550,10 +546,8 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, goto err; } if (check_tree_block_fsid(root, eb)) { - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs bad fsid on block %llu\n", + printk_ratelimited(KERN_INFO "btrfs bad fsid on block %llu\n", (unsigned long long)eb->start); - } ret = -EIO; goto err; } @@ -650,12 +644,6 @@ unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info) return 256 * limit; } -int btrfs_congested_async(struct btrfs_fs_info *info, int iodone) -{ - return atomic_read(&info->nr_async_bios) > - btrfs_async_submit_limit(info); -} - static void run_one_async_start(struct btrfs_work *work) { struct async_submit_bio *async; @@ -963,7 +951,7 @@ struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, struct inode *btree_inode = root->fs_info->btree_inode; struct extent_buffer *eb; eb = find_extent_buffer(&BTRFS_I(btree_inode)->io_tree, - bytenr, blocksize, GFP_NOFS); + bytenr, blocksize); return eb; } @@ -974,7 +962,7 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, struct extent_buffer *eb; eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree, - bytenr, blocksize, NULL, GFP_NOFS); + bytenr, blocksize, NULL); return eb; } @@ -1056,15 +1044,14 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->last_trans = 0; root->highest_objectid = 0; root->name = NULL; - root->in_sysfs = 0; root->inode_tree = RB_ROOT; + INIT_RADIX_TREE(&root->delayed_nodes_tree, GFP_ATOMIC); root->block_rsv = NULL; root->orphan_block_rsv = NULL; INIT_LIST_HEAD(&root->dirty_list); INIT_LIST_HEAD(&root->orphan_list); INIT_LIST_HEAD(&root->root_list); - spin_lock_init(&root->node_lock); spin_lock_init(&root->orphan_lock); spin_lock_init(&root->inode_lock); spin_lock_init(&root->accounting_lock); @@ -1080,7 +1067,7 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->log_transid = 0; root->last_log_commit = 0; extent_io_tree_init(&root->dirty_log_pages, - fs_info->btree_inode->i_mapping, GFP_NOFS); + fs_info->btree_inode->i_mapping); memset(&root->root_key, 0, sizeof(root->root_key)); memset(&root->root_item, 0, sizeof(root->root_item)); @@ -1283,21 +1270,6 @@ out: return root; } -struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info, - u64 root_objectid) -{ - struct btrfs_root *root; - - if (root_objectid == BTRFS_ROOT_TREE_OBJECTID) - return fs_info->tree_root; - if (root_objectid == BTRFS_EXTENT_TREE_OBJECTID) - return fs_info->extent_root; - - root = radix_tree_lookup(&fs_info->fs_roots_radix, - (unsigned long)root_objectid); - return root; -} - struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, struct btrfs_key *location) { @@ -1326,7 +1298,22 @@ again: if (IS_ERR(root)) return root; - set_anon_super(&root->anon_super, NULL); + root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS); + root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned), + GFP_NOFS); + if (!root->free_ino_pinned || !root->free_ino_ctl) { + ret = -ENOMEM; + goto fail; + } + + btrfs_init_free_ino_ctl(root); + mutex_init(&root->fs_commit_mutex); + spin_lock_init(&root->cache_lock); + init_waitqueue_head(&root->cache_wait); + + ret = set_anon_super(&root->anon_super, NULL); + if (ret) + goto fail; if (btrfs_root_refs(&root->root_item) == 0) { ret = -ENOENT; @@ -1369,41 +1356,6 @@ fail: return ERR_PTR(ret); } -struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info, - struct btrfs_key *location, - const char *name, int namelen) -{ - return btrfs_read_fs_root_no_name(fs_info, location); -#if 0 - struct btrfs_root *root; - int ret; - - root = btrfs_read_fs_root_no_name(fs_info, location); - if (!root) - return NULL; - - if (root->in_sysfs) - return root; - - ret = btrfs_set_root_name(root, name, namelen); - if (ret) { - free_extent_buffer(root->node); - kfree(root); - return ERR_PTR(ret); - } - - ret = btrfs_sysfs_add_root(root); - if (ret) { - free_extent_buffer(root->node); - kfree(root->name); - kfree(root); - return ERR_PTR(ret); - } - root->in_sysfs = 1; - return root; -#endif -} - static int btrfs_congested_fn(void *congested_data, int bdi_bits) { struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data; @@ -1411,7 +1363,8 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) struct btrfs_device *device; struct backing_dev_info *bdi; - list_for_each_entry(device, &info->fs_devices->devices, dev_list) { + rcu_read_lock(); + list_for_each_entry_rcu(device, &info->fs_devices->devices, dev_list) { if (!device->bdev) continue; bdi = blk_get_backing_dev_info(device->bdev); @@ -1420,6 +1373,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits) break; } } + rcu_read_unlock(); return ret; } @@ -1522,6 +1476,7 @@ static int cleaner_kthread(void *arg) btrfs_run_delayed_iputs(root); btrfs_clean_old_snapshots(root); mutex_unlock(&root->fs_info->cleaner_mutex); + btrfs_run_defrag_inodes(root->fs_info); } if (freezing(current)) { @@ -1551,24 +1506,24 @@ static int transaction_kthread(void *arg) vfs_check_frozen(root->fs_info->sb, SB_FREEZE_WRITE); mutex_lock(&root->fs_info->transaction_kthread_mutex); - spin_lock(&root->fs_info->new_trans_lock); + spin_lock(&root->fs_info->trans_lock); cur = root->fs_info->running_transaction; if (!cur) { - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); goto sleep; } now = get_seconds(); if (!cur->blocked && (now < cur->start_time || now - cur->start_time < 30)) { - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); delay = HZ * 5; goto sleep; } transid = cur->transid; - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); if (transid == trans->transid) { ret = btrfs_commit_transaction(trans, root); @@ -1611,7 +1566,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, struct btrfs_root *csum_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); struct btrfs_root *tree_root = btrfs_sb(sb); - struct btrfs_fs_info *fs_info = tree_root->fs_info; + struct btrfs_fs_info *fs_info = NULL; struct btrfs_root *chunk_root = kzalloc(sizeof(struct btrfs_root), GFP_NOFS); struct btrfs_root *dev_root = kzalloc(sizeof(struct btrfs_root), @@ -1623,11 +1578,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, struct btrfs_super_block *disk_super; - if (!extent_root || !tree_root || !fs_info || + if (!extent_root || !tree_root || !tree_root->fs_info || !chunk_root || !dev_root || !csum_root) { err = -ENOMEM; goto fail; } + fs_info = tree_root->fs_info; ret = init_srcu_struct(&fs_info->subvol_srcu); if (ret) { @@ -1658,10 +1614,12 @@ struct btrfs_root *open_ctree(struct super_block *sb, INIT_LIST_HEAD(&fs_info->ordered_operations); INIT_LIST_HEAD(&fs_info->caching_block_groups); spin_lock_init(&fs_info->delalloc_lock); - spin_lock_init(&fs_info->new_trans_lock); + spin_lock_init(&fs_info->trans_lock); spin_lock_init(&fs_info->ref_cache_lock); spin_lock_init(&fs_info->fs_roots_radix_lock); spin_lock_init(&fs_info->delayed_iput_lock); + spin_lock_init(&fs_info->defrag_inodes_lock); + mutex_init(&fs_info->reloc_mutex); init_completion(&fs_info->kobj_unregister); fs_info->tree_root = tree_root; @@ -1684,15 +1642,34 @@ struct btrfs_root *open_ctree(struct super_block *sb, atomic_set(&fs_info->async_delalloc_pages, 0); atomic_set(&fs_info->async_submit_draining, 0); atomic_set(&fs_info->nr_async_bios, 0); + atomic_set(&fs_info->defrag_running, 0); fs_info->sb = sb; fs_info->max_inline = 8192 * 1024; fs_info->metadata_ratio = 0; + fs_info->defrag_inodes = RB_ROOT; + fs_info->trans_no_join = 0; fs_info->thread_pool_size = min_t(unsigned long, num_online_cpus() + 2, 8); INIT_LIST_HEAD(&fs_info->ordered_extents); spin_lock_init(&fs_info->ordered_extent_lock); + fs_info->delayed_root = kmalloc(sizeof(struct btrfs_delayed_root), + GFP_NOFS); + if (!fs_info->delayed_root) { + err = -ENOMEM; + goto fail_iput; + } + btrfs_init_delayed_root(fs_info->delayed_root); + + mutex_init(&fs_info->scrub_lock); + atomic_set(&fs_info->scrubs_running, 0); + atomic_set(&fs_info->scrub_pause_req, 0); + atomic_set(&fs_info->scrubs_paused, 0); + atomic_set(&fs_info->scrub_cancel_req, 0); + init_waitqueue_head(&fs_info->scrub_pause_wait); + init_rwsem(&fs_info->scrub_super_lock); + fs_info->scrub_workers_refcnt = 0; sb->s_blocksize = 4096; sb->s_blocksize_bits = blksize_bits(4096); @@ -1711,10 +1688,8 @@ struct btrfs_root *open_ctree(struct super_block *sb, RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node); extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree, - fs_info->btree_inode->i_mapping, - GFP_NOFS); - extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree, - GFP_NOFS); + fs_info->btree_inode->i_mapping); + extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree); BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops; @@ -1728,14 +1703,13 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->block_group_cache_tree = RB_ROOT; extent_io_tree_init(&fs_info->freed_extents[0], - fs_info->btree_inode->i_mapping, GFP_NOFS); + fs_info->btree_inode->i_mapping); extent_io_tree_init(&fs_info->freed_extents[1], - fs_info->btree_inode->i_mapping, GFP_NOFS); + fs_info->btree_inode->i_mapping); fs_info->pinned_extents = &fs_info->freed_extents[0]; fs_info->do_barriers = 1; - mutex_init(&fs_info->trans_mutex); mutex_init(&fs_info->ordered_operations_mutex); mutex_init(&fs_info->tree_log_mutex); mutex_init(&fs_info->chunk_mutex); @@ -1760,7 +1734,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, bh = btrfs_read_dev_super(fs_devices->latest_bdev); if (!bh) { err = -EINVAL; - goto fail_iput; + goto fail_alloc; } memcpy(&fs_info->super_copy, bh->b_data, sizeof(fs_info->super_copy)); @@ -1772,7 +1746,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, disk_super = &fs_info->super_copy; if (!btrfs_super_root(disk_super)) - goto fail_iput; + goto fail_alloc; /* check FS state, whether FS is broken. */ fs_info->fs_state |= btrfs_super_flags(disk_super); @@ -1788,7 +1762,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, ret = btrfs_parse_options(tree_root, options); if (ret) { err = ret; - goto fail_iput; + goto fail_alloc; } features = btrfs_super_incompat_flags(disk_super) & @@ -1798,7 +1772,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, "unsupported optional features (%Lx).\n", (unsigned long long)features); err = -EINVAL; - goto fail_iput; + goto fail_alloc; } features = btrfs_super_incompat_flags(disk_super); @@ -1814,7 +1788,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, "unsupported option features (%Lx).\n", (unsigned long long)features); err = -EINVAL; - goto fail_iput; + goto fail_alloc; } btrfs_init_workers(&fs_info->generic_worker, @@ -1861,6 +1835,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, &fs_info->generic_worker); btrfs_init_workers(&fs_info->endio_freespace_worker, "freespace-write", 1, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta", + fs_info->thread_pool_size, + &fs_info->generic_worker); /* * endios are largely parallel and should have a very @@ -1882,6 +1859,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_start_workers(&fs_info->endio_meta_write_workers, 1); btrfs_start_workers(&fs_info->endio_write_workers, 1); btrfs_start_workers(&fs_info->endio_freespace_worker, 1); + btrfs_start_workers(&fs_info->delayed_workers, 1); fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super); fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages, @@ -2138,6 +2116,9 @@ fail_sb_buffer: btrfs_stop_workers(&fs_info->endio_write_workers); btrfs_stop_workers(&fs_info->endio_freespace_worker); btrfs_stop_workers(&fs_info->submit_workers); + btrfs_stop_workers(&fs_info->delayed_workers); +fail_alloc: + kfree(fs_info->delayed_root); fail_iput: invalidate_inode_pages2(fs_info->btree_inode->i_mapping); iput(fs_info->btree_inode); @@ -2165,11 +2146,9 @@ static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate) if (uptodate) { set_buffer_uptodate(bh); } else { - if (printk_ratelimit()) { - printk(KERN_WARNING "lost page write due to " + printk_ratelimited(KERN_WARNING "lost page write due to " "I/O error on %s\n", bdevname(bh->b_bdev, b)); - } /* note, we dont' set_buffer_write_io_error because we have * our own ways of dealing with the IO errors */ @@ -2333,7 +2312,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) mutex_lock(&root->fs_info->fs_devices->device_list_mutex); head = &root->fs_info->fs_devices->devices; - list_for_each_entry(dev, head, dev_list) { + list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) { total_errors++; continue; @@ -2366,7 +2345,7 @@ int write_all_supers(struct btrfs_root *root, int max_mirrors) } total_errors = 0; - list_for_each_entry(dev, head, dev_list) { + list_for_each_entry_rcu(dev, head, dev_list) { if (!dev->bdev) continue; if (!dev->in_fs_metadata || !dev->writeable) @@ -2404,12 +2383,15 @@ int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root) if (btrfs_root_refs(&root->root_item) == 0) synchronize_srcu(&fs_info->subvol_srcu); + __btrfs_remove_free_space_cache(root->free_ino_pinned); + __btrfs_remove_free_space_cache(root->free_ino_ctl); free_fs_root(root); return 0; } static void free_fs_root(struct btrfs_root *root) { + iput(root->cache_inode); WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree)); if (root->anon_super.s_dev) { down_write(&root->anon_super.s_umount); @@ -2417,6 +2399,8 @@ static void free_fs_root(struct btrfs_root *root) } free_extent_buffer(root->node); free_extent_buffer(root->commit_root); + kfree(root->free_ino_ctl); + kfree(root->free_ino_pinned); kfree(root->name); kfree(root); } @@ -2495,13 +2479,13 @@ int btrfs_commit_super(struct btrfs_root *root) down_write(&root->fs_info->cleanup_work_sem); up_write(&root->fs_info->cleanup_work_sem); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); ret = btrfs_commit_transaction(trans, root); BUG_ON(ret); /* run commit again to drop the original snapshot */ - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); btrfs_commit_transaction(trans, root); @@ -2520,6 +2504,15 @@ int close_ctree(struct btrfs_root *root) fs_info->closing = 1; smp_mb(); + btrfs_scrub_cancel(root); + + /* wait for any defraggers to finish */ + wait_event(fs_info->transaction_wait, + (atomic_read(&fs_info->defrag_running) == 0)); + + /* clear out the rbtree of defraggable inodes */ + btrfs_run_defrag_inodes(root->fs_info); + btrfs_put_block_group_cache(fs_info); /* @@ -2578,6 +2571,7 @@ int close_ctree(struct btrfs_root *root) del_fs_roots(fs_info); iput(fs_info->btree_inode); + kfree(fs_info->delayed_root); btrfs_stop_workers(&fs_info->generic_worker); btrfs_stop_workers(&fs_info->fixup_workers); @@ -2589,6 +2583,7 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->endio_write_workers); btrfs_stop_workers(&fs_info->endio_freespace_worker); btrfs_stop_workers(&fs_info->submit_workers); + btrfs_stop_workers(&fs_info->delayed_workers); btrfs_close_devices(fs_info->fs_devices); btrfs_mapping_tree_free(&fs_info->mapping_tree); @@ -2665,6 +2660,29 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) if (current->flags & PF_MEMALLOC) return; + btrfs_balance_delayed_items(root); + + num_dirty = root->fs_info->dirty_metadata_bytes; + + if (num_dirty > thresh) { + balance_dirty_pages_ratelimited_nr( + root->fs_info->btree_inode->i_mapping, 1); + } + return; +} + +void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) +{ + /* + * looks as though older kernels can get into trouble with + * this code, they end up stuck in balance_dirty_pages forever + */ + u64 num_dirty; + unsigned long thresh = 32 * 1024 * 1024; + + if (current->flags & PF_MEMALLOC) + return; + num_dirty = root->fs_info->dirty_metadata_bytes; if (num_dirty > thresh) { @@ -2697,7 +2715,7 @@ int btree_lock_page_hook(struct page *page) goto out; len = page->private >> 2; - eb = find_extent_buffer(io_tree, bytenr, len, GFP_NOFS); + eb = find_extent_buffer(io_tree, bytenr, len); if (!eb) goto out; @@ -2893,9 +2911,8 @@ static int btrfs_destroy_delalloc_inodes(struct btrfs_root *root) INIT_LIST_HEAD(&splice); - list_splice_init(&root->fs_info->delalloc_inodes, &splice); - spin_lock(&root->fs_info->delalloc_lock); + list_splice_init(&root->fs_info->delalloc_inodes, &splice); while (!list_empty(&splice)) { btrfs_inode = list_entry(splice.next, struct btrfs_inode, @@ -3006,10 +3023,13 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) WARN_ON(1); - mutex_lock(&root->fs_info->trans_mutex); mutex_lock(&root->fs_info->transaction_kthread_mutex); + spin_lock(&root->fs_info->trans_lock); list_splice_init(&root->fs_info->trans_list, &list); + root->fs_info->trans_no_join = 1; + spin_unlock(&root->fs_info->trans_lock); + while (!list_empty(&list)) { t = list_entry(list.next, struct btrfs_transaction, list); if (!t) @@ -3034,23 +3054,18 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) t->blocked = 0; if (waitqueue_active(&root->fs_info->transaction_wait)) wake_up(&root->fs_info->transaction_wait); - mutex_unlock(&root->fs_info->trans_mutex); - mutex_lock(&root->fs_info->trans_mutex); t->commit_done = 1; if (waitqueue_active(&t->commit_wait)) wake_up(&t->commit_wait); - mutex_unlock(&root->fs_info->trans_mutex); - - mutex_lock(&root->fs_info->trans_mutex); btrfs_destroy_pending_snapshots(t); btrfs_destroy_delalloc_inodes(root); - spin_lock(&root->fs_info->new_trans_lock); + spin_lock(&root->fs_info->trans_lock); root->fs_info->running_transaction = NULL; - spin_unlock(&root->fs_info->new_trans_lock); + spin_unlock(&root->fs_info->trans_lock); btrfs_destroy_marked_extents(root, &t->dirty_pages, EXTENT_DIRTY); @@ -3064,8 +3079,10 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) kmem_cache_free(btrfs_transaction_cachep, t); } + spin_lock(&root->fs_info->trans_lock); + root->fs_info->trans_no_join = 0; + spin_unlock(&root->fs_info->trans_lock); mutex_unlock(&root->fs_info->transaction_kthread_mutex); - mutex_unlock(&root->fs_info->trans_mutex); return 0; } diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index 07b20dc2fd95..a0b610a67aae 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -55,35 +55,20 @@ int btrfs_commit_super(struct btrfs_root *root); int btrfs_error_commit_super(struct btrfs_root *root); struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize); -struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info, - u64 root_objectid); -struct btrfs_root *btrfs_read_fs_root(struct btrfs_fs_info *fs_info, - struct btrfs_key *location, - const char *name, int namelen); struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, struct btrfs_key *location); struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, struct btrfs_key *location); int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info); -int btrfs_insert_dev_radix(struct btrfs_root *root, - struct block_device *bdev, - u64 device_id, - u64 block_start, - u64 num_blocks); void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); +void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root); void btrfs_mark_buffer_dirty(struct extent_buffer *buf); -void btrfs_mark_buffer_dirty_nonblocking(struct extent_buffer *buf); int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid); int btrfs_set_buffer_uptodate(struct extent_buffer *buf); -int wait_on_tree_block_writeback(struct btrfs_root *root, - struct extent_buffer *buf); int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid); u32 btrfs_csum_data(struct btrfs_root *root, char *data, u32 seed, size_t len); void btrfs_csum_final(u32 crc, char *result); -int btrfs_open_device(struct btrfs_device *dev); -int btrfs_verify_block_csum(struct btrfs_root *root, - struct extent_buffer *buf); int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio, int metadata); int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode, @@ -91,8 +76,6 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode, unsigned long bio_flags, u64 bio_offset, extent_submit_bio_hook_t *submit_bio_start, extent_submit_bio_hook_t *submit_bio_done); - -int btrfs_congested_async(struct btrfs_fs_info *info, int iodone); unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info); int btrfs_write_tree_block(struct extent_buffer *buf); int btrfs_wait_tree_block_writeback(struct extent_buffer *buf); diff --git a/fs/btrfs/export.c b/fs/btrfs/export.c index b4ffad859adb..1b8dc33778f9 100644 --- a/fs/btrfs/export.c +++ b/fs/btrfs/export.c @@ -32,7 +32,7 @@ static int btrfs_encode_fh(struct dentry *dentry, u32 *fh, int *max_len, len = BTRFS_FID_SIZE_NON_CONNECTABLE; type = FILEID_BTRFS_WITHOUT_PARENT; - fid->objectid = inode->i_ino; + fid->objectid = btrfs_ino(inode); fid->root_objectid = BTRFS_I(inode)->root->objectid; fid->gen = inode->i_generation; @@ -178,13 +178,13 @@ static struct dentry *btrfs_get_parent(struct dentry *child) if (!path) return ERR_PTR(-ENOMEM); - if (dir->i_ino == BTRFS_FIRST_FREE_OBJECTID) { + if (btrfs_ino(dir) == BTRFS_FIRST_FREE_OBJECTID) { key.objectid = root->root_key.objectid; key.type = BTRFS_ROOT_BACKREF_KEY; key.offset = (u64)-1; root = root->fs_info->tree_root; } else { - key.objectid = dir->i_ino; + key.objectid = btrfs_ino(dir); key.type = BTRFS_INODE_REF_KEY; key.offset = (u64)-1; } @@ -244,6 +244,7 @@ static int btrfs_get_name(struct dentry *parent, char *name, struct btrfs_key key; int name_len; int ret; + u64 ino; if (!dir || !inode) return -EINVAL; @@ -251,19 +252,21 @@ static int btrfs_get_name(struct dentry *parent, char *name, if (!S_ISDIR(dir->i_mode)) return -EINVAL; + ino = btrfs_ino(inode); + path = btrfs_alloc_path(); if (!path) return -ENOMEM; path->leave_spinning = 1; - if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) { + if (ino == BTRFS_FIRST_FREE_OBJECTID) { key.objectid = BTRFS_I(inode)->root->root_key.objectid; key.type = BTRFS_ROOT_BACKREF_KEY; key.offset = (u64)-1; root = root->fs_info->tree_root; } else { - key.objectid = inode->i_ino; - key.offset = dir->i_ino; + key.objectid = ino; + key.offset = btrfs_ino(dir); key.type = BTRFS_INODE_REF_KEY; } @@ -272,7 +275,7 @@ static int btrfs_get_name(struct dentry *parent, char *name, btrfs_free_path(path); return ret; } else if (ret > 0) { - if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) { + if (ino == BTRFS_FIRST_FREE_OBJECTID) { path->slots[0]--; } else { btrfs_free_path(path); @@ -281,11 +284,11 @@ static int btrfs_get_name(struct dentry *parent, char *name, } leaf = path->nodes[0]; - if (inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) { - rref = btrfs_item_ptr(leaf, path->slots[0], + if (ino == BTRFS_FIRST_FREE_OBJECTID) { + rref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); - name_ptr = (unsigned long)(rref + 1); - name_len = btrfs_root_ref_name_len(leaf, rref); + name_ptr = (unsigned long)(rref + 1); + name_len = btrfs_root_ref_name_len(leaf, rref); } else { iref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_ref); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 9ee6bd55e16c..71cd456fdb60 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -94,7 +94,7 @@ static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits) return (cache->flags & bits) == bits; } -void btrfs_get_block_group(struct btrfs_block_group_cache *cache) +static void btrfs_get_block_group(struct btrfs_block_group_cache *cache) { atomic_inc(&cache->count); } @@ -105,6 +105,7 @@ void btrfs_put_block_group(struct btrfs_block_group_cache *cache) WARN_ON(cache->pinned > 0); WARN_ON(cache->reserved > 0); WARN_ON(cache->reserved_pinned > 0); + kfree(cache->free_space_ctl); kfree(cache); } } @@ -347,7 +348,7 @@ static int caching_kthread(void *data) */ path->skip_locking = 1; path->search_commit_root = 1; - path->reada = 2; + path->reada = 1; key.objectid = last; key.offset = 0; @@ -365,8 +366,7 @@ again: nritems = btrfs_header_nritems(leaf); while (1) { - smp_mb(); - if (fs_info->closing > 1) { + if (btrfs_fs_closing(fs_info) > 1) { last = (u64)-1; break; } @@ -378,15 +378,18 @@ again: if (ret) break; - caching_ctl->progress = last; - btrfs_release_path(extent_root, path); - up_read(&fs_info->extent_commit_sem); - mutex_unlock(&caching_ctl->mutex); - if (btrfs_transaction_in_commit(fs_info)) - schedule_timeout(1); - else + if (need_resched() || + btrfs_next_leaf(extent_root, path)) { + caching_ctl->progress = last; + btrfs_release_path(path); + up_read(&fs_info->extent_commit_sem); + mutex_unlock(&caching_ctl->mutex); cond_resched(); - goto again; + goto again; + } + leaf = path->nodes[0]; + nritems = btrfs_header_nritems(leaf); + continue; } if (key.objectid < block_group->key.objectid) { @@ -754,8 +757,12 @@ again: atomic_inc(&head->node.refs); spin_unlock(&delayed_refs->lock); - btrfs_release_path(root->fs_info->extent_root, path); + btrfs_release_path(path); + /* + * Mutex was contended, block until it's released and try + * again + */ mutex_lock(&head->mutex); mutex_unlock(&head->mutex); btrfs_put_delayed_ref(&head->node); @@ -934,7 +941,7 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, break; } } - btrfs_release_path(root, path); + btrfs_release_path(path); if (owner < BTRFS_FIRST_FREE_OBJECTID) new_size += sizeof(*bi); @@ -947,7 +954,6 @@ static int convert_extent_item_v0(struct btrfs_trans_handle *trans, BUG_ON(ret); ret = btrfs_extend_item(trans, root, path, new_size); - BUG_ON(ret); leaf = path->nodes[0]; item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); @@ -1042,7 +1048,7 @@ again: return 0; #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 key.type = BTRFS_EXTENT_REF_V0_KEY; - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret < 0) { err = ret; @@ -1080,7 +1086,7 @@ again: if (match_extent_data_ref(leaf, ref, root_objectid, owner, offset)) { if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } err = 0; @@ -1141,7 +1147,7 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, if (match_extent_data_ref(leaf, ref, root_objectid, owner, offset)) break; - btrfs_release_path(root, path); + btrfs_release_path(path); key.offset++; ret = btrfs_insert_empty_item(trans, root, path, &key, size); @@ -1167,7 +1173,7 @@ static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(leaf); ret = 0; fail: - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } @@ -1293,7 +1299,7 @@ static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans, ret = -ENOENT; #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 if (ret == -ENOENT && parent) { - btrfs_release_path(root, path); + btrfs_release_path(path); key.type = BTRFS_EXTENT_REF_V0_KEY; ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret > 0) @@ -1322,7 +1328,7 @@ static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans, } ret = btrfs_insert_empty_item(trans, root, path, &key, 0); - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } @@ -1555,7 +1561,6 @@ int setup_inline_extent_backref(struct btrfs_trans_handle *trans, size = btrfs_extent_inline_ref_size(type); ret = btrfs_extend_item(trans, root, path, size); - BUG_ON(ret); ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item); refs = btrfs_extent_refs(leaf, ei); @@ -1608,7 +1613,7 @@ static int lookup_extent_backref(struct btrfs_trans_handle *trans, if (ret != -ENOENT) return ret; - btrfs_release_path(root, path); + btrfs_release_path(path); *ref_ret = NULL; if (owner < BTRFS_FIRST_FREE_OBJECTID) { @@ -1684,7 +1689,6 @@ int update_inline_extent_backref(struct btrfs_trans_handle *trans, end - ptr - size); item_size -= size; ret = btrfs_truncate_item(trans, root, path, item_size, 1); - BUG_ON(ret); } btrfs_mark_buffer_dirty(leaf); return 0; @@ -1862,7 +1866,7 @@ static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans, __run_delayed_extent_op(extent_op, leaf, item); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root->fs_info->extent_root, path); + btrfs_release_path(path); path->reada = 1; path->leave_spinning = 1; @@ -2297,6 +2301,10 @@ again: atomic_inc(&ref->refs); spin_unlock(&delayed_refs->lock); + /* + * Mutex was contended, block until it's + * released and try again + */ mutex_lock(&head->mutex); mutex_unlock(&head->mutex); @@ -2361,8 +2369,12 @@ static noinline int check_delayed_ref(struct btrfs_trans_handle *trans, atomic_inc(&head->node.refs); spin_unlock(&delayed_refs->lock); - btrfs_release_path(root->fs_info->extent_root, path); + btrfs_release_path(path); + /* + * Mutex was contended, block until it's released and let + * caller try again + */ mutex_lock(&head->mutex); mutex_unlock(&head->mutex); btrfs_put_delayed_ref(&head->node); @@ -2510,126 +2522,6 @@ out: return ret; } -#if 0 -int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, - struct extent_buffer *buf, u32 nr_extents) -{ - struct btrfs_key key; - struct btrfs_file_extent_item *fi; - u64 root_gen; - u32 nritems; - int i; - int level; - int ret = 0; - int shared = 0; - - if (!root->ref_cows) - return 0; - - if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) { - shared = 0; - root_gen = root->root_key.offset; - } else { - shared = 1; - root_gen = trans->transid - 1; - } - - level = btrfs_header_level(buf); - nritems = btrfs_header_nritems(buf); - - if (level == 0) { - struct btrfs_leaf_ref *ref; - struct btrfs_extent_info *info; - - ref = btrfs_alloc_leaf_ref(root, nr_extents); - if (!ref) { - ret = -ENOMEM; - goto out; - } - - ref->root_gen = root_gen; - ref->bytenr = buf->start; - ref->owner = btrfs_header_owner(buf); - ref->generation = btrfs_header_generation(buf); - ref->nritems = nr_extents; - info = ref->extents; - - for (i = 0; nr_extents > 0 && i < nritems; i++) { - u64 disk_bytenr; - btrfs_item_key_to_cpu(buf, &key, i); - if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY) - continue; - fi = btrfs_item_ptr(buf, i, - struct btrfs_file_extent_item); - if (btrfs_file_extent_type(buf, fi) == - BTRFS_FILE_EXTENT_INLINE) - continue; - disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi); - if (disk_bytenr == 0) - continue; - - info->bytenr = disk_bytenr; - info->num_bytes = - btrfs_file_extent_disk_num_bytes(buf, fi); - info->objectid = key.objectid; - info->offset = key.offset; - info++; - } - - ret = btrfs_add_leaf_ref(root, ref, shared); - if (ret == -EEXIST && shared) { - struct btrfs_leaf_ref *old; - old = btrfs_lookup_leaf_ref(root, ref->bytenr); - BUG_ON(!old); - btrfs_remove_leaf_ref(root, old); - btrfs_free_leaf_ref(root, old); - ret = btrfs_add_leaf_ref(root, ref, shared); - } - WARN_ON(ret); - btrfs_free_leaf_ref(root, ref); - } -out: - return ret; -} - -/* when a block goes through cow, we update the reference counts of - * everything that block points to. The internal pointers of the block - * can be in just about any order, and it is likely to have clusters of - * things that are close together and clusters of things that are not. - * - * To help reduce the seeks that come with updating all of these reference - * counts, sort them by byte number before actual updates are done. - * - * struct refsort is used to match byte number to slot in the btree block. - * we sort based on the byte number and then use the slot to actually - * find the item. - * - * struct refsort is smaller than strcut btrfs_item and smaller than - * struct btrfs_key_ptr. Since we're currently limited to the page size - * for a btree block, there's no way for a kmalloc of refsorts for a - * single node to be bigger than a page. - */ -struct refsort { - u64 bytenr; - u32 slot; -}; - -/* - * for passing into sort() - */ -static int refsort_cmp(const void *a_void, const void *b_void) -{ - const struct refsort *a = a_void; - const struct refsort *b = b_void; - - if (a->bytenr < b->bytenr) - return -1; - if (a->bytenr > b->bytenr) - return 1; - return 0; -} -#endif - static int __btrfs_mod_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct extent_buffer *buf, @@ -2732,7 +2624,7 @@ static int write_one_cache_group(struct btrfs_trans_handle *trans, bi = btrfs_item_ptr_offset(leaf, path->slots[0]); write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item)); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); fail: if (ret) return ret; @@ -2785,7 +2677,7 @@ again: inode = lookup_free_space_inode(root, block_group, path); if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { ret = PTR_ERR(inode); - btrfs_release_path(root, path); + btrfs_release_path(path); goto out; } @@ -2854,7 +2746,7 @@ again: out_put: iput(inode); out_free: - btrfs_release_path(root, path); + btrfs_release_path(path); out: spin_lock(&block_group->lock); block_group->disk_cache_state = dcs; @@ -3144,7 +3036,8 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes) /* make sure bytes are sectorsize aligned */ bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1); - if (root == root->fs_info->tree_root) { + if (root == root->fs_info->tree_root || + BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) { alloc_chunk = 0; committed = 1; } @@ -3174,7 +3067,7 @@ again: spin_unlock(&data_sinfo->lock); alloc: alloc_target = btrfs_get_alloc_profile(root, 1); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); @@ -3196,13 +3089,21 @@ alloc: } goto again; } + + /* + * If we have less pinned bytes than we want to allocate then + * don't bother committing the transaction, it won't help us. + */ + if (data_sinfo->bytes_pinned < bytes) + committed = 1; spin_unlock(&data_sinfo->lock); /* commit the current transaction and try again */ commit_trans: - if (!committed && !root->fs_info->open_ioctl_trans) { + if (!committed && + !atomic_read(&root->fs_info->open_ioctl_trans)) { committed = 1; - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); ret = btrfs_commit_transaction(trans, root); @@ -3211,18 +3112,6 @@ commit_trans: goto again; } -#if 0 /* I hope we never need this code again, just in case */ - printk(KERN_ERR "no space left, need %llu, %llu bytes_used, " - "%llu bytes_reserved, " "%llu bytes_pinned, " - "%llu bytes_readonly, %llu may use %llu total\n", - (unsigned long long)bytes, - (unsigned long long)data_sinfo->bytes_used, - (unsigned long long)data_sinfo->bytes_reserved, - (unsigned long long)data_sinfo->bytes_pinned, - (unsigned long long)data_sinfo->bytes_readonly, - (unsigned long long)data_sinfo->bytes_may_use, - (unsigned long long)data_sinfo->total_bytes); -#endif return -ENOSPC; } data_sinfo->bytes_may_use += bytes; @@ -3589,7 +3478,7 @@ again: goto out; ret = -ENOSPC; - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) goto out; ret = btrfs_commit_transaction(trans, root); @@ -3651,8 +3540,8 @@ static void block_rsv_add_bytes(struct btrfs_block_rsv *block_rsv, spin_unlock(&block_rsv->lock); } -void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, - struct btrfs_block_rsv *dest, u64 num_bytes) +static void block_rsv_release_bytes(struct btrfs_block_rsv *block_rsv, + struct btrfs_block_rsv *dest, u64 num_bytes) { struct btrfs_space_info *space_info = block_rsv->space_info; @@ -3816,7 +3705,7 @@ int btrfs_block_rsv_check(struct btrfs_trans_handle *trans, if (trans) return -EAGAIN; - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); ret = btrfs_commit_transaction(trans, root); return 0; @@ -3855,23 +3744,7 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info) u64 meta_used; u64 data_used; int csum_size = btrfs_super_csum_size(&fs_info->super_copy); -#if 0 - /* - * per tree used space accounting can be inaccuracy, so we - * can't rely on it. - */ - spin_lock(&fs_info->extent_root->accounting_lock); - num_bytes = btrfs_root_used(&fs_info->extent_root->root_item); - spin_unlock(&fs_info->extent_root->accounting_lock); - spin_lock(&fs_info->csum_root->accounting_lock); - num_bytes += btrfs_root_used(&fs_info->csum_root->root_item); - spin_unlock(&fs_info->csum_root->accounting_lock); - - spin_lock(&fs_info->tree_root->accounting_lock); - num_bytes += btrfs_root_used(&fs_info->tree_root->root_item); - spin_unlock(&fs_info->tree_root->accounting_lock); -#endif sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA); spin_lock(&sinfo->lock); data_used = sinfo->bytes_used; @@ -3924,10 +3797,7 @@ static void update_global_block_rsv(struct btrfs_fs_info *fs_info) block_rsv->reserved = block_rsv->size; block_rsv->full = 1; } -#if 0 - printk(KERN_INFO"global block rsv size %llu reserved %llu\n", - block_rsv->size, block_rsv->reserved); -#endif + spin_unlock(&sinfo->lock); spin_unlock(&block_rsv->lock); } @@ -3973,10 +3843,35 @@ static void release_global_block_rsv(struct btrfs_fs_info *fs_info) WARN_ON(fs_info->chunk_block_rsv.reserved > 0); } -static u64 calc_trans_metadata_size(struct btrfs_root *root, int num_items) +int btrfs_truncate_reserve_metadata(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct btrfs_block_rsv *rsv) { - return (root->leafsize + root->nodesize * (BTRFS_MAX_LEVEL - 1)) * - 3 * num_items; + struct btrfs_block_rsv *trans_rsv = &root->fs_info->trans_block_rsv; + u64 num_bytes; + int ret; + + /* + * Truncate should be freeing data, but give us 2 items just in case it + * needs to use some space. We may want to be smarter about this in the + * future. + */ + num_bytes = btrfs_calc_trans_metadata_size(root, 2); + + /* We already have enough bytes, just return */ + if (rsv->reserved >= num_bytes) + return 0; + + num_bytes -= rsv->reserved; + + /* + * You should have reserved enough space before hand to do this, so this + * should not fail. + */ + ret = block_rsv_migrate_bytes(trans_rsv, rsv, num_bytes); + BUG_ON(ret); + + return 0; } int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans, @@ -3989,7 +3884,7 @@ int btrfs_trans_reserve_metadata(struct btrfs_trans_handle *trans, if (num_items == 0 || root->fs_info->chunk_root == root) return 0; - num_bytes = calc_trans_metadata_size(root, num_items); + num_bytes = btrfs_calc_trans_metadata_size(root, num_items); ret = btrfs_block_rsv_add(trans, root, &root->fs_info->trans_block_rsv, num_bytes); if (!ret) { @@ -4019,23 +3914,18 @@ int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans, struct btrfs_block_rsv *dst_rsv = root->orphan_block_rsv; /* - * one for deleting orphan item, one for updating inode and - * two for calling btrfs_truncate_inode_items. - * - * btrfs_truncate_inode_items is a delete operation, it frees - * more space than it uses in most cases. So two units of - * metadata space should be enough for calling it many times. - * If all of the metadata space is used, we can commit - * transaction and use space it freed. + * We need to hold space in order to delete our orphan item once we've + * added it, so this takes the reservation so we can release it later + * when we are truly done with the orphan item. */ - u64 num_bytes = calc_trans_metadata_size(root, 4); + u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } void btrfs_orphan_release_metadata(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; - u64 num_bytes = calc_trans_metadata_size(root, 4); + u64 num_bytes = btrfs_calc_trans_metadata_size(root, 1); btrfs_block_rsv_release(root, root->orphan_block_rsv, num_bytes); } @@ -4049,7 +3939,7 @@ int btrfs_snap_reserve_metadata(struct btrfs_trans_handle *trans, * two for root back/forward refs, two for directory entries * and one for root of the snapshot. */ - u64 num_bytes = calc_trans_metadata_size(root, 5); + u64 num_bytes = btrfs_calc_trans_metadata_size(root, 5); dst_rsv->space_info = src_rsv->space_info; return block_rsv_migrate_bytes(src_rsv, dst_rsv, num_bytes); } @@ -4078,7 +3968,7 @@ int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes) if (nr_extents > reserved_extents) { nr_extents -= reserved_extents; - to_reserve = calc_trans_metadata_size(root, nr_extents); + to_reserve = btrfs_calc_trans_metadata_size(root, nr_extents); } else { nr_extents = 0; to_reserve = 0; @@ -4132,7 +4022,7 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes) to_free = calc_csum_metadata_size(inode, num_bytes); if (nr_extents > 0) - to_free += calc_trans_metadata_size(root, nr_extents); + to_free += btrfs_calc_trans_metadata_size(root, nr_extents); btrfs_block_rsv_release(root, &root->fs_info->delalloc_block_rsv, to_free); @@ -4541,7 +4431,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, NULL, refs_to_drop, is_data); BUG_ON(ret); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); path->leave_spinning = 1; key.objectid = bytenr; @@ -4580,7 +4470,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, owner_objectid, 0); BUG_ON(ret < 0); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); path->leave_spinning = 1; key.objectid = bytenr; @@ -4650,7 +4540,7 @@ static int __btrfs_free_extent(struct btrfs_trans_handle *trans, ret = btrfs_del_items(trans, extent_root, path, path->slots[0], num_to_del); BUG_ON(ret); - btrfs_release_path(extent_root, path); + btrfs_release_path(path); if (is_data) { ret = btrfs_del_csums(trans, root, bytenr, num_bytes); @@ -4893,7 +4783,7 @@ wait_block_group_cache_progress(struct btrfs_block_group_cache *cache, return 0; wait_event(caching_ctl->wait, block_group_cache_done(cache) || - (cache->free_space >= num_bytes)); + (cache->free_space_ctl->free_space >= num_bytes)); put_caching_control(caching_ctl); return 0; @@ -4952,7 +4842,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, u64 num_bytes, u64 empty_size, u64 search_start, u64 search_end, u64 hint_byte, struct btrfs_key *ins, - int data) + u64 data) { int ret = 0; struct btrfs_root *root = orig_root->fs_info->extent_root; @@ -4979,7 +4869,7 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans, space_info = __find_space_info(root->fs_info, data); if (!space_info) { - printk(KERN_ERR "No space info for %d\n", data); + printk(KERN_ERR "No space info for %llu\n", data); return -ENOSPC; } @@ -5129,6 +5019,15 @@ have_block_group: if (unlikely(block_group->ro)) goto loop; + spin_lock(&block_group->free_space_ctl->tree_lock); + if (cached && + block_group->free_space_ctl->free_space < + num_bytes + empty_size) { + spin_unlock(&block_group->free_space_ctl->tree_lock); + goto loop; + } + spin_unlock(&block_group->free_space_ctl->tree_lock); + /* * Ok we want to try and use the cluster allocator, so lets look * there, unless we are on LOOP_NO_EMPTY_SIZE, since we will @@ -5292,6 +5191,7 @@ checks: btrfs_add_free_space(block_group, offset, search_start - offset); BUG_ON(offset > search_start); + btrfs_put_block_group(block_group); break; loop: failed_cluster_refill = false; @@ -5314,9 +5214,7 @@ loop: * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try * again */ - if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE && - (found_uncached_bg || empty_size || empty_cluster || - allowed_chunk_alloc)) { + if (!ins->objectid && loop < LOOP_NO_EMPTY_SIZE) { index = 0; if (loop == LOOP_FIND_IDEAL && found_uncached_bg) { found_uncached_bg = false; @@ -5356,42 +5254,39 @@ loop: goto search; } - if (loop < LOOP_CACHING_WAIT) { - loop++; - goto search; - } + loop++; if (loop == LOOP_ALLOC_CHUNK) { - empty_size = 0; - empty_cluster = 0; - } + if (allowed_chunk_alloc) { + ret = do_chunk_alloc(trans, root, num_bytes + + 2 * 1024 * 1024, data, + CHUNK_ALLOC_LIMITED); + allowed_chunk_alloc = 0; + if (ret == 1) + done_chunk_alloc = 1; + } else if (!done_chunk_alloc && + space_info->force_alloc == + CHUNK_ALLOC_NO_FORCE) { + space_info->force_alloc = CHUNK_ALLOC_LIMITED; + } - if (allowed_chunk_alloc) { - ret = do_chunk_alloc(trans, root, num_bytes + - 2 * 1024 * 1024, data, - CHUNK_ALLOC_LIMITED); - allowed_chunk_alloc = 0; - done_chunk_alloc = 1; - } else if (!done_chunk_alloc && - space_info->force_alloc == CHUNK_ALLOC_NO_FORCE) { - space_info->force_alloc = CHUNK_ALLOC_LIMITED; + /* + * We didn't allocate a chunk, go ahead and drop the + * empty size and loop again. + */ + if (!done_chunk_alloc) + loop = LOOP_NO_EMPTY_SIZE; } - if (loop < LOOP_NO_EMPTY_SIZE) { - loop++; - goto search; + if (loop == LOOP_NO_EMPTY_SIZE) { + empty_size = 0; + empty_cluster = 0; } - ret = -ENOSPC; + + goto search; } else if (!ins->objectid) { ret = -ENOSPC; - } - - /* we found what we needed */ - if (ins->objectid) { - if (!(data & BTRFS_BLOCK_GROUP_DATA)) - trans->block_group = block_group->key.objectid; - - btrfs_put_block_group(block_group); + } else if (ins->objectid) { ret = 0; } @@ -6480,7 +6375,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, trans->block_rsv = block_rsv; } } - btrfs_release_path(root, path); + btrfs_release_path(path); BUG_ON(err); ret = btrfs_del_root(trans, tree_root, &root->root_key); @@ -6584,1514 +6479,6 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, return ret; } -#if 0 -static unsigned long calc_ra(unsigned long start, unsigned long last, - unsigned long nr) -{ - return min(last, start + nr - 1); -} - -static noinline int relocate_inode_pages(struct inode *inode, u64 start, - u64 len) -{ - u64 page_start; - u64 page_end; - unsigned long first_index; - unsigned long last_index; - unsigned long i; - struct page *page; - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; - struct file_ra_state *ra; - struct btrfs_ordered_extent *ordered; - unsigned int total_read = 0; - unsigned int total_dirty = 0; - int ret = 0; - - ra = kzalloc(sizeof(*ra), GFP_NOFS); - if (!ra) - return -ENOMEM; - - mutex_lock(&inode->i_mutex); - first_index = start >> PAGE_CACHE_SHIFT; - last_index = (start + len - 1) >> PAGE_CACHE_SHIFT; - - /* make sure the dirty trick played by the caller work */ - ret = invalidate_inode_pages2_range(inode->i_mapping, - first_index, last_index); - if (ret) - goto out_unlock; - - file_ra_state_init(ra, inode->i_mapping); - - for (i = first_index ; i <= last_index; i++) { - if (total_read % ra->ra_pages == 0) { - btrfs_force_ra(inode->i_mapping, ra, NULL, i, - calc_ra(i, last_index, ra->ra_pages)); - } - total_read++; -again: - if (((u64)i << PAGE_CACHE_SHIFT) > i_size_read(inode)) - BUG_ON(1); - page = grab_cache_page(inode->i_mapping, i); - if (!page) { - ret = -ENOMEM; - goto out_unlock; - } - if (!PageUptodate(page)) { - btrfs_readpage(NULL, page); - lock_page(page); - if (!PageUptodate(page)) { - unlock_page(page); - page_cache_release(page); - ret = -EIO; - goto out_unlock; - } - } - wait_on_page_writeback(page); - - page_start = (u64)page->index << PAGE_CACHE_SHIFT; - page_end = page_start + PAGE_CACHE_SIZE - 1; - lock_extent(io_tree, page_start, page_end, GFP_NOFS); - - ordered = btrfs_lookup_ordered_extent(inode, page_start); - if (ordered) { - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - unlock_page(page); - page_cache_release(page); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - goto again; - } - set_page_extent_mapped(page); - - if (i == first_index) - set_extent_bits(io_tree, page_start, page_end, - EXTENT_BOUNDARY, GFP_NOFS); - btrfs_set_extent_delalloc(inode, page_start, page_end); - - set_page_dirty(page); - total_dirty++; - - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - unlock_page(page); - page_cache_release(page); - } - -out_unlock: - kfree(ra); - mutex_unlock(&inode->i_mutex); - balance_dirty_pages_ratelimited_nr(inode->i_mapping, total_dirty); - return ret; -} - -static noinline int relocate_data_extent(struct inode *reloc_inode, - struct btrfs_key *extent_key, - u64 offset) -{ - struct btrfs_root *root = BTRFS_I(reloc_inode)->root; - struct extent_map_tree *em_tree = &BTRFS_I(reloc_inode)->extent_tree; - struct extent_map *em; - u64 start = extent_key->objectid - offset; - u64 end = start + extent_key->offset - 1; - - em = alloc_extent_map(GFP_NOFS); - BUG_ON(!em); - - em->start = start; - em->len = extent_key->offset; - em->block_len = extent_key->offset; - em->block_start = extent_key->objectid; - em->bdev = root->fs_info->fs_devices->latest_bdev; - set_bit(EXTENT_FLAG_PINNED, &em->flags); - - /* setup extent map to cheat btrfs_readpage */ - lock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS); - while (1) { - int ret; - write_lock(&em_tree->lock); - ret = add_extent_mapping(em_tree, em); - write_unlock(&em_tree->lock); - if (ret != -EEXIST) { - free_extent_map(em); - break; - } - btrfs_drop_extent_cache(reloc_inode, start, end, 0); - } - unlock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS); - - return relocate_inode_pages(reloc_inode, start, extent_key->offset); -} - -struct btrfs_ref_path { - u64 extent_start; - u64 nodes[BTRFS_MAX_LEVEL]; - u64 root_objectid; - u64 root_generation; - u64 owner_objectid; - u32 num_refs; - int lowest_level; - int current_level; - int shared_level; - - struct btrfs_key node_keys[BTRFS_MAX_LEVEL]; - u64 new_nodes[BTRFS_MAX_LEVEL]; -}; - -struct disk_extent { - u64 ram_bytes; - u64 disk_bytenr; - u64 disk_num_bytes; - u64 offset; - u64 num_bytes; - u8 compression; - u8 encryption; - u16 other_encoding; -}; - -static int is_cowonly_root(u64 root_objectid) -{ - if (root_objectid == BTRFS_ROOT_TREE_OBJECTID || - root_objectid == BTRFS_EXTENT_TREE_OBJECTID || - root_objectid == BTRFS_CHUNK_TREE_OBJECTID || - root_objectid == BTRFS_DEV_TREE_OBJECTID || - root_objectid == BTRFS_TREE_LOG_OBJECTID || - root_objectid == BTRFS_CSUM_TREE_OBJECTID) - return 1; - return 0; -} - -static noinline int __next_ref_path(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_ref_path *ref_path, - int first_time) -{ - struct extent_buffer *leaf; - struct btrfs_path *path; - struct btrfs_extent_ref *ref; - struct btrfs_key key; - struct btrfs_key found_key; - u64 bytenr; - u32 nritems; - int level; - int ret = 1; - - path = btrfs_alloc_path(); - if (!path) - return -ENOMEM; - - if (first_time) { - ref_path->lowest_level = -1; - ref_path->current_level = -1; - ref_path->shared_level = -1; - goto walk_up; - } -walk_down: - level = ref_path->current_level - 1; - while (level >= -1) { - u64 parent; - if (level < ref_path->lowest_level) - break; - - if (level >= 0) - bytenr = ref_path->nodes[level]; - else - bytenr = ref_path->extent_start; - BUG_ON(bytenr == 0); - - parent = ref_path->nodes[level + 1]; - ref_path->nodes[level + 1] = 0; - ref_path->current_level = level; - BUG_ON(parent == 0); - - key.objectid = bytenr; - key.offset = parent + 1; - key.type = BTRFS_EXTENT_REF_KEY; - - ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0); - if (ret < 0) - goto out; - BUG_ON(ret == 0); - - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(extent_root, path); - if (ret < 0) - goto out; - if (ret > 0) - goto next; - leaf = path->nodes[0]; - } - - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid == bytenr && - found_key.type == BTRFS_EXTENT_REF_KEY) { - if (level < ref_path->shared_level) - ref_path->shared_level = level; - goto found; - } -next: - level--; - btrfs_release_path(extent_root, path); - cond_resched(); - } - /* reached lowest level */ - ret = 1; - goto out; -walk_up: - level = ref_path->current_level; - while (level < BTRFS_MAX_LEVEL - 1) { - u64 ref_objectid; - - if (level >= 0) - bytenr = ref_path->nodes[level]; - else - bytenr = ref_path->extent_start; - - BUG_ON(bytenr == 0); - - key.objectid = bytenr; - key.offset = 0; - key.type = BTRFS_EXTENT_REF_KEY; - - ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0); - if (ret < 0) - goto out; - - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(extent_root, path); - if (ret < 0) - goto out; - if (ret > 0) { - /* the extent was freed by someone */ - if (ref_path->lowest_level == level) - goto out; - btrfs_release_path(extent_root, path); - goto walk_down; - } - leaf = path->nodes[0]; - } - - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid != bytenr || - found_key.type != BTRFS_EXTENT_REF_KEY) { - /* the extent was freed by someone */ - if (ref_path->lowest_level == level) { - ret = 1; - goto out; - } - btrfs_release_path(extent_root, path); - goto walk_down; - } -found: - ref = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_extent_ref); - ref_objectid = btrfs_ref_objectid(leaf, ref); - if (ref_objectid < BTRFS_FIRST_FREE_OBJECTID) { - if (first_time) { - level = (int)ref_objectid; - BUG_ON(level >= BTRFS_MAX_LEVEL); - ref_path->lowest_level = level; - ref_path->current_level = level; - ref_path->nodes[level] = bytenr; - } else { - WARN_ON(ref_objectid != level); - } - } else { - WARN_ON(level != -1); - } - first_time = 0; - - if (ref_path->lowest_level == level) { - ref_path->owner_objectid = ref_objectid; - ref_path->num_refs = btrfs_ref_num_refs(leaf, ref); - } - - /* - * the block is tree root or the block isn't in reference - * counted tree. - */ - if (found_key.objectid == found_key.offset || - is_cowonly_root(btrfs_ref_root(leaf, ref))) { - ref_path->root_objectid = btrfs_ref_root(leaf, ref); - ref_path->root_generation = - btrfs_ref_generation(leaf, ref); - if (level < 0) { - /* special reference from the tree log */ - ref_path->nodes[0] = found_key.offset; - ref_path->current_level = 0; - } - ret = 0; - goto out; - } - - level++; - BUG_ON(ref_path->nodes[level] != 0); - ref_path->nodes[level] = found_key.offset; - ref_path->current_level = level; - - /* - * the reference was created in the running transaction, - * no need to continue walking up. - */ - if (btrfs_ref_generation(leaf, ref) == trans->transid) { - ref_path->root_objectid = btrfs_ref_root(leaf, ref); - ref_path->root_generation = - btrfs_ref_generation(leaf, ref); - ret = 0; - goto out; - } - - btrfs_release_path(extent_root, path); - cond_resched(); - } - /* reached max tree level, but no tree root found. */ - BUG(); -out: - btrfs_free_path(path); - return ret; -} - -static int btrfs_first_ref_path(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_ref_path *ref_path, - u64 extent_start) -{ - memset(ref_path, 0, sizeof(*ref_path)); - ref_path->extent_start = extent_start; - - return __next_ref_path(trans, extent_root, ref_path, 1); -} - -static int btrfs_next_ref_path(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_ref_path *ref_path) -{ - return __next_ref_path(trans, extent_root, ref_path, 0); -} - -static noinline int get_new_locations(struct inode *reloc_inode, - struct btrfs_key *extent_key, - u64 offset, int no_fragment, - struct disk_extent **extents, - int *nr_extents) -{ - struct btrfs_root *root = BTRFS_I(reloc_inode)->root; - struct btrfs_path *path; - struct btrfs_file_extent_item *fi; - struct extent_buffer *leaf; - struct disk_extent *exts = *extents; - struct btrfs_key found_key; - u64 cur_pos; - u64 last_byte; - u32 nritems; - int nr = 0; - int max = *nr_extents; - int ret; - - WARN_ON(!no_fragment && *extents); - if (!exts) { - max = 1; - exts = kmalloc(sizeof(*exts) * max, GFP_NOFS); - if (!exts) - return -ENOMEM; - } - - path = btrfs_alloc_path(); - if (!path) { - if (exts != *extents) - kfree(exts); - return -ENOMEM; - } - - cur_pos = extent_key->objectid - offset; - last_byte = extent_key->objectid + extent_key->offset; - ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino, - cur_pos, 0); - if (ret < 0) - goto out; - if (ret > 0) { - ret = -ENOENT; - goto out; - } - - while (1) { - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - if (path->slots[0] >= nritems) { - ret = btrfs_next_leaf(root, path); - if (ret < 0) - goto out; - if (ret > 0) - break; - leaf = path->nodes[0]; - } - - btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.offset != cur_pos || - found_key.type != BTRFS_EXTENT_DATA_KEY || - found_key.objectid != reloc_inode->i_ino) - break; - - fi = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_file_extent_item); - if (btrfs_file_extent_type(leaf, fi) != - BTRFS_FILE_EXTENT_REG || - btrfs_file_extent_disk_bytenr(leaf, fi) == 0) - break; - - if (nr == max) { - struct disk_extent *old = exts; - max *= 2; - exts = kzalloc(sizeof(*exts) * max, GFP_NOFS); - if (!exts) { - ret = -ENOMEM; - goto out; - } - memcpy(exts, old, sizeof(*exts) * nr); - if (old != *extents) - kfree(old); - } - - exts[nr].disk_bytenr = - btrfs_file_extent_disk_bytenr(leaf, fi); - exts[nr].disk_num_bytes = - btrfs_file_extent_disk_num_bytes(leaf, fi); - exts[nr].offset = btrfs_file_extent_offset(leaf, fi); - exts[nr].num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - exts[nr].ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi); - exts[nr].compression = btrfs_file_extent_compression(leaf, fi); - exts[nr].encryption = btrfs_file_extent_encryption(leaf, fi); - exts[nr].other_encoding = btrfs_file_extent_other_encoding(leaf, - fi); - BUG_ON(exts[nr].offset > 0); - BUG_ON(exts[nr].compression || exts[nr].encryption); - BUG_ON(exts[nr].num_bytes != exts[nr].disk_num_bytes); - - cur_pos += exts[nr].num_bytes; - nr++; - - if (cur_pos + offset >= last_byte) - break; - - if (no_fragment) { - ret = 1; - goto out; - } - path->slots[0]++; - } - - BUG_ON(cur_pos + offset > last_byte); - if (cur_pos + offset < last_byte) { - ret = -ENOENT; - goto out; - } - ret = 0; -out: - btrfs_free_path(path); - if (ret) { - if (exts != *extents) - kfree(exts); - } else { - *extents = exts; - *nr_extents = nr; - } - return ret; -} - -static noinline int replace_one_extent(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *extent_key, - struct btrfs_key *leaf_key, - struct btrfs_ref_path *ref_path, - struct disk_extent *new_extents, - int nr_extents) -{ - struct extent_buffer *leaf; - struct btrfs_file_extent_item *fi; - struct inode *inode = NULL; - struct btrfs_key key; - u64 lock_start = 0; - u64 lock_end = 0; - u64 num_bytes; - u64 ext_offset; - u64 search_end = (u64)-1; - u32 nritems; - int nr_scaned = 0; - int extent_locked = 0; - int extent_type; - int ret; - - memcpy(&key, leaf_key, sizeof(key)); - if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) { - if (key.objectid < ref_path->owner_objectid || - (key.objectid == ref_path->owner_objectid && - key.type < BTRFS_EXTENT_DATA_KEY)) { - key.objectid = ref_path->owner_objectid; - key.type = BTRFS_EXTENT_DATA_KEY; - key.offset = 0; - } - } - - while (1) { - ret = btrfs_search_slot(trans, root, &key, path, 0, 1); - if (ret < 0) - goto out; - - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); -next: - if (extent_locked && ret > 0) { - /* - * the file extent item was modified by someone - * before the extent got locked. - */ - unlock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - extent_locked = 0; - } - - if (path->slots[0] >= nritems) { - if (++nr_scaned > 2) - break; - - BUG_ON(extent_locked); - ret = btrfs_next_leaf(root, path); - if (ret < 0) - goto out; - if (ret > 0) - break; - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - } - - btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - - if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) { - if ((key.objectid > ref_path->owner_objectid) || - (key.objectid == ref_path->owner_objectid && - key.type > BTRFS_EXTENT_DATA_KEY) || - key.offset >= search_end) - break; - } - - if (inode && key.objectid != inode->i_ino) { - BUG_ON(extent_locked); - btrfs_release_path(root, path); - mutex_unlock(&inode->i_mutex); - iput(inode); - inode = NULL; - continue; - } - - if (key.type != BTRFS_EXTENT_DATA_KEY) { - path->slots[0]++; - ret = 1; - goto next; - } - fi = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_file_extent_item); - extent_type = btrfs_file_extent_type(leaf, fi); - if ((extent_type != BTRFS_FILE_EXTENT_REG && - extent_type != BTRFS_FILE_EXTENT_PREALLOC) || - (btrfs_file_extent_disk_bytenr(leaf, fi) != - extent_key->objectid)) { - path->slots[0]++; - ret = 1; - goto next; - } - - num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - ext_offset = btrfs_file_extent_offset(leaf, fi); - - if (search_end == (u64)-1) { - search_end = key.offset - ext_offset + - btrfs_file_extent_ram_bytes(leaf, fi); - } - - if (!extent_locked) { - lock_start = key.offset; - lock_end = lock_start + num_bytes - 1; - } else { - if (lock_start > key.offset || - lock_end + 1 < key.offset + num_bytes) { - unlock_extent(&BTRFS_I(inode)->io_tree, - lock_start, lock_end, GFP_NOFS); - extent_locked = 0; - } - } - - if (!inode) { - btrfs_release_path(root, path); - - inode = btrfs_iget_locked(root->fs_info->sb, - key.objectid, root); - if (inode->i_state & I_NEW) { - BTRFS_I(inode)->root = root; - BTRFS_I(inode)->location.objectid = - key.objectid; - BTRFS_I(inode)->location.type = - BTRFS_INODE_ITEM_KEY; - BTRFS_I(inode)->location.offset = 0; - btrfs_read_locked_inode(inode); - unlock_new_inode(inode); - } - /* - * some code call btrfs_commit_transaction while - * holding the i_mutex, so we can't use mutex_lock - * here. - */ - if (is_bad_inode(inode) || - !mutex_trylock(&inode->i_mutex)) { - iput(inode); - inode = NULL; - key.offset = (u64)-1; - goto skip; - } - } - - if (!extent_locked) { - struct btrfs_ordered_extent *ordered; - - btrfs_release_path(root, path); - - lock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - ordered = btrfs_lookup_first_ordered_extent(inode, - lock_end); - if (ordered && - ordered->file_offset <= lock_end && - ordered->file_offset + ordered->len > lock_start) { - unlock_extent(&BTRFS_I(inode)->io_tree, - lock_start, lock_end, GFP_NOFS); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - key.offset += num_bytes; - goto skip; - } - if (ordered) - btrfs_put_ordered_extent(ordered); - - extent_locked = 1; - continue; - } - - if (nr_extents == 1) { - /* update extent pointer in place */ - btrfs_set_file_extent_disk_bytenr(leaf, fi, - new_extents[0].disk_bytenr); - btrfs_set_file_extent_disk_num_bytes(leaf, fi, - new_extents[0].disk_num_bytes); - btrfs_mark_buffer_dirty(leaf); - - btrfs_drop_extent_cache(inode, key.offset, - key.offset + num_bytes - 1, 0); - - ret = btrfs_inc_extent_ref(trans, root, - new_extents[0].disk_bytenr, - new_extents[0].disk_num_bytes, - leaf->start, - root->root_key.objectid, - trans->transid, - key.objectid); - BUG_ON(ret); - - ret = btrfs_free_extent(trans, root, - extent_key->objectid, - extent_key->offset, - leaf->start, - btrfs_header_owner(leaf), - btrfs_header_generation(leaf), - key.objectid, 0); - BUG_ON(ret); - - btrfs_release_path(root, path); - key.offset += num_bytes; - } else { - BUG_ON(1); -#if 0 - u64 alloc_hint; - u64 extent_len; - int i; - /* - * drop old extent pointer at first, then insert the - * new pointers one bye one - */ - btrfs_release_path(root, path); - ret = btrfs_drop_extents(trans, root, inode, key.offset, - key.offset + num_bytes, - key.offset, &alloc_hint); - BUG_ON(ret); - - for (i = 0; i < nr_extents; i++) { - if (ext_offset >= new_extents[i].num_bytes) { - ext_offset -= new_extents[i].num_bytes; - continue; - } - extent_len = min(new_extents[i].num_bytes - - ext_offset, num_bytes); - - ret = btrfs_insert_empty_item(trans, root, - path, &key, - sizeof(*fi)); - BUG_ON(ret); - - leaf = path->nodes[0]; - fi = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_file_extent_item); - btrfs_set_file_extent_generation(leaf, fi, - trans->transid); - btrfs_set_file_extent_type(leaf, fi, - BTRFS_FILE_EXTENT_REG); - btrfs_set_file_extent_disk_bytenr(leaf, fi, - new_extents[i].disk_bytenr); - btrfs_set_file_extent_disk_num_bytes(leaf, fi, - new_extents[i].disk_num_bytes); - btrfs_set_file_extent_ram_bytes(leaf, fi, - new_extents[i].ram_bytes); - - btrfs_set_file_extent_compression(leaf, fi, - new_extents[i].compression); - btrfs_set_file_extent_encryption(leaf, fi, - new_extents[i].encryption); - btrfs_set_file_extent_other_encoding(leaf, fi, - new_extents[i].other_encoding); - - btrfs_set_file_extent_num_bytes(leaf, fi, - extent_len); - ext_offset += new_extents[i].offset; - btrfs_set_file_extent_offset(leaf, fi, - ext_offset); - btrfs_mark_buffer_dirty(leaf); - - btrfs_drop_extent_cache(inode, key.offset, - key.offset + extent_len - 1, 0); - - ret = btrfs_inc_extent_ref(trans, root, - new_extents[i].disk_bytenr, - new_extents[i].disk_num_bytes, - leaf->start, - root->root_key.objectid, - trans->transid, key.objectid); - BUG_ON(ret); - btrfs_release_path(root, path); - - inode_add_bytes(inode, extent_len); - - ext_offset = 0; - num_bytes -= extent_len; - key.offset += extent_len; - - if (num_bytes == 0) - break; - } - BUG_ON(i >= nr_extents); -#endif - } - - if (extent_locked) { - unlock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - extent_locked = 0; - } -skip: - if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS && - key.offset >= search_end) - break; - - cond_resched(); - } - ret = 0; -out: - btrfs_release_path(root, path); - if (inode) { - mutex_unlock(&inode->i_mutex); - if (extent_locked) { - unlock_extent(&BTRFS_I(inode)->io_tree, lock_start, - lock_end, GFP_NOFS); - } - iput(inode); - } - return ret; -} - -int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct extent_buffer *buf, u64 orig_start) -{ - int level; - int ret; - - BUG_ON(btrfs_header_generation(buf) != trans->transid); - BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID); - - level = btrfs_header_level(buf); - if (level == 0) { - struct btrfs_leaf_ref *ref; - struct btrfs_leaf_ref *orig_ref; - - orig_ref = btrfs_lookup_leaf_ref(root, orig_start); - if (!orig_ref) - return -ENOENT; - - ref = btrfs_alloc_leaf_ref(root, orig_ref->nritems); - if (!ref) { - btrfs_free_leaf_ref(root, orig_ref); - return -ENOMEM; - } - - ref->nritems = orig_ref->nritems; - memcpy(ref->extents, orig_ref->extents, - sizeof(ref->extents[0]) * ref->nritems); - - btrfs_free_leaf_ref(root, orig_ref); - - ref->root_gen = trans->transid; - ref->bytenr = buf->start; - ref->owner = btrfs_header_owner(buf); - ref->generation = btrfs_header_generation(buf); - - ret = btrfs_add_leaf_ref(root, ref, 0); - WARN_ON(ret); - btrfs_free_leaf_ref(root, ref); - } - return 0; -} - -static noinline int invalidate_extent_cache(struct btrfs_root *root, - struct extent_buffer *leaf, - struct btrfs_block_group_cache *group, - struct btrfs_root *target_root) -{ - struct btrfs_key key; - struct inode *inode = NULL; - struct btrfs_file_extent_item *fi; - struct extent_state *cached_state = NULL; - u64 num_bytes; - u64 skip_objectid = 0; - u32 nritems; - u32 i; - - nritems = btrfs_header_nritems(leaf); - for (i = 0; i < nritems; i++) { - btrfs_item_key_to_cpu(leaf, &key, i); - if (key.objectid == skip_objectid || - key.type != BTRFS_EXTENT_DATA_KEY) - continue; - fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item); - if (btrfs_file_extent_type(leaf, fi) == - BTRFS_FILE_EXTENT_INLINE) - continue; - if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0) - continue; - if (!inode || inode->i_ino != key.objectid) { - iput(inode); - inode = btrfs_ilookup(target_root->fs_info->sb, - key.objectid, target_root, 1); - } - if (!inode) { - skip_objectid = key.objectid; - continue; - } - num_bytes = btrfs_file_extent_num_bytes(leaf, fi); - - lock_extent_bits(&BTRFS_I(inode)->io_tree, key.offset, - key.offset + num_bytes - 1, 0, &cached_state, - GFP_NOFS); - btrfs_drop_extent_cache(inode, key.offset, - key.offset + num_bytes - 1, 1); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, key.offset, - key.offset + num_bytes - 1, &cached_state, - GFP_NOFS); - cond_resched(); - } - iput(inode); - return 0; -} - -static noinline int replace_extents_in_leaf(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct extent_buffer *leaf, - struct btrfs_block_group_cache *group, - struct inode *reloc_inode) -{ - struct btrfs_key key; - struct btrfs_key extent_key; - struct btrfs_file_extent_item *fi; - struct btrfs_leaf_ref *ref; - struct disk_extent *new_extent; - u64 bytenr; - u64 num_bytes; - u32 nritems; - u32 i; - int ext_index; - int nr_extent; - int ret; - - new_extent = kmalloc(sizeof(*new_extent), GFP_NOFS); - if (!new_extent) - return -ENOMEM; - - ref = btrfs_lookup_leaf_ref(root, leaf->start); - BUG_ON(!ref); - - ext_index = -1; - nritems = btrfs_header_nritems(leaf); - for (i = 0; i < nritems; i++) { - btrfs_item_key_to_cpu(leaf, &key, i); - if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY) - continue; - fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item); - if (btrfs_file_extent_type(leaf, fi) == - BTRFS_FILE_EXTENT_INLINE) - continue; - bytenr = btrfs_file_extent_disk_bytenr(leaf, fi); - num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi); - if (bytenr == 0) - continue; - - ext_index++; - if (bytenr >= group->key.objectid + group->key.offset || - bytenr + num_bytes <= group->key.objectid) - continue; - - extent_key.objectid = bytenr; - extent_key.offset = num_bytes; - extent_key.type = BTRFS_EXTENT_ITEM_KEY; - nr_extent = 1; - ret = get_new_locations(reloc_inode, &extent_key, - group->key.objectid, 1, - &new_extent, &nr_extent); - if (ret > 0) - continue; - BUG_ON(ret < 0); - - BUG_ON(ref->extents[ext_index].bytenr != bytenr); - BUG_ON(ref->extents[ext_index].num_bytes != num_bytes); - ref->extents[ext_index].bytenr = new_extent->disk_bytenr; - ref->extents[ext_index].num_bytes = new_extent->disk_num_bytes; - - btrfs_set_file_extent_disk_bytenr(leaf, fi, - new_extent->disk_bytenr); - btrfs_set_file_extent_disk_num_bytes(leaf, fi, - new_extent->disk_num_bytes); - btrfs_mark_buffer_dirty(leaf); - - ret = btrfs_inc_extent_ref(trans, root, - new_extent->disk_bytenr, - new_extent->disk_num_bytes, - leaf->start, - root->root_key.objectid, - trans->transid, key.objectid); - BUG_ON(ret); - - ret = btrfs_free_extent(trans, root, - bytenr, num_bytes, leaf->start, - btrfs_header_owner(leaf), - btrfs_header_generation(leaf), - key.objectid, 0); - BUG_ON(ret); - cond_resched(); - } - kfree(new_extent); - BUG_ON(ext_index + 1 != ref->nritems); - btrfs_free_leaf_ref(root, ref); - return 0; -} - -int btrfs_free_reloc_root(struct btrfs_trans_handle *trans, - struct btrfs_root *root) -{ - struct btrfs_root *reloc_root; - int ret; - - if (root->reloc_root) { - reloc_root = root->reloc_root; - root->reloc_root = NULL; - list_add(&reloc_root->dead_list, - &root->fs_info->dead_reloc_roots); - - btrfs_set_root_bytenr(&reloc_root->root_item, - reloc_root->node->start); - btrfs_set_root_level(&root->root_item, - btrfs_header_level(reloc_root->node)); - memset(&reloc_root->root_item.drop_progress, 0, - sizeof(struct btrfs_disk_key)); - reloc_root->root_item.drop_level = 0; - - ret = btrfs_update_root(trans, root->fs_info->tree_root, - &reloc_root->root_key, - &reloc_root->root_item); - BUG_ON(ret); - } - return 0; -} - -int btrfs_drop_dead_reloc_roots(struct btrfs_root *root) -{ - struct btrfs_trans_handle *trans; - struct btrfs_root *reloc_root; - struct btrfs_root *prev_root = NULL; - struct list_head dead_roots; - int ret; - unsigned long nr; - - INIT_LIST_HEAD(&dead_roots); - list_splice_init(&root->fs_info->dead_reloc_roots, &dead_roots); - - while (!list_empty(&dead_roots)) { - reloc_root = list_entry(dead_roots.prev, - struct btrfs_root, dead_list); - list_del_init(&reloc_root->dead_list); - - BUG_ON(reloc_root->commit_root != NULL); - while (1) { - trans = btrfs_join_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - - mutex_lock(&root->fs_info->drop_mutex); - ret = btrfs_drop_snapshot(trans, reloc_root); - if (ret != -EAGAIN) - break; - mutex_unlock(&root->fs_info->drop_mutex); - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, root); - BUG_ON(ret); - btrfs_btree_balance_dirty(root, nr); - } - - free_extent_buffer(reloc_root->node); - - ret = btrfs_del_root(trans, root->fs_info->tree_root, - &reloc_root->root_key); - BUG_ON(ret); - mutex_unlock(&root->fs_info->drop_mutex); - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, root); - BUG_ON(ret); - btrfs_btree_balance_dirty(root, nr); - - kfree(prev_root); - prev_root = reloc_root; - } - if (prev_root) { - btrfs_remove_leaf_refs(prev_root, (u64)-1, 0); - kfree(prev_root); - } - return 0; -} - -int btrfs_add_dead_reloc_root(struct btrfs_root *root) -{ - list_add(&root->dead_list, &root->fs_info->dead_reloc_roots); - return 0; -} - -int btrfs_cleanup_reloc_trees(struct btrfs_root *root) -{ - struct btrfs_root *reloc_root; - struct btrfs_trans_handle *trans; - struct btrfs_key location; - int found; - int ret; - - mutex_lock(&root->fs_info->tree_reloc_mutex); - ret = btrfs_find_dead_roots(root, BTRFS_TREE_RELOC_OBJECTID, NULL); - BUG_ON(ret); - found = !list_empty(&root->fs_info->dead_reloc_roots); - mutex_unlock(&root->fs_info->tree_reloc_mutex); - - if (found) { - trans = btrfs_start_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); - } - - location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID; - location.offset = (u64)-1; - location.type = BTRFS_ROOT_ITEM_KEY; - - reloc_root = btrfs_read_fs_root_no_name(root->fs_info, &location); - BUG_ON(!reloc_root); - ret = btrfs_orphan_cleanup(reloc_root); - BUG_ON(ret); - return 0; -} - -static noinline int init_reloc_tree(struct btrfs_trans_handle *trans, - struct btrfs_root *root) -{ - struct btrfs_root *reloc_root; - struct extent_buffer *eb; - struct btrfs_root_item *root_item; - struct btrfs_key root_key; - int ret; - - BUG_ON(!root->ref_cows); - if (root->reloc_root) - return 0; - - root_item = kmalloc(sizeof(*root_item), GFP_NOFS); - if (!root_item) - return -ENOMEM; - - ret = btrfs_copy_root(trans, root, root->commit_root, - &eb, BTRFS_TREE_RELOC_OBJECTID); - BUG_ON(ret); - - root_key.objectid = BTRFS_TREE_RELOC_OBJECTID; - root_key.offset = root->root_key.objectid; - root_key.type = BTRFS_ROOT_ITEM_KEY; - - memcpy(root_item, &root->root_item, sizeof(root_item)); - btrfs_set_root_refs(root_item, 0); - btrfs_set_root_bytenr(root_item, eb->start); - btrfs_set_root_level(root_item, btrfs_header_level(eb)); - btrfs_set_root_generation(root_item, trans->transid); - - btrfs_tree_unlock(eb); - free_extent_buffer(eb); - - ret = btrfs_insert_root(trans, root->fs_info->tree_root, - &root_key, root_item); - BUG_ON(ret); - kfree(root_item); - - reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root, - &root_key); - BUG_ON(IS_ERR(reloc_root)); - reloc_root->last_trans = trans->transid; - reloc_root->commit_root = NULL; - reloc_root->ref_tree = &root->fs_info->reloc_ref_tree; - - root->reloc_root = reloc_root; - return 0; -} - -/* - * Core function of space balance. - * - * The idea is using reloc trees to relocate tree blocks in reference - * counted roots. There is one reloc tree for each subvol, and all - * reloc trees share same root key objectid. Reloc trees are snapshots - * of the latest committed roots of subvols (root->commit_root). - * - * To relocate a tree block referenced by a subvol, there are two steps. - * COW the block through subvol's reloc tree, then update block pointer - * in the subvol to point to the new block. Since all reloc trees share - * same root key objectid, doing special handing for tree blocks owned - * by them is easy. Once a tree block has been COWed in one reloc tree, - * we can use the resulting new block directly when the same block is - * required to COW again through other reloc trees. By this way, relocated - * tree blocks are shared between reloc trees, so they are also shared - * between subvols. - */ -static noinline int relocate_one_path(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *first_key, - struct btrfs_ref_path *ref_path, - struct btrfs_block_group_cache *group, - struct inode *reloc_inode) -{ - struct btrfs_root *reloc_root; - struct extent_buffer *eb = NULL; - struct btrfs_key *keys; - u64 *nodes; - int level; - int shared_level; - int lowest_level = 0; - int ret; - - if (ref_path->owner_objectid < BTRFS_FIRST_FREE_OBJECTID) - lowest_level = ref_path->owner_objectid; - - if (!root->ref_cows) { - path->lowest_level = lowest_level; - ret = btrfs_search_slot(trans, root, first_key, path, 0, 1); - BUG_ON(ret < 0); - path->lowest_level = 0; - btrfs_release_path(root, path); - return 0; - } - - mutex_lock(&root->fs_info->tree_reloc_mutex); - ret = init_reloc_tree(trans, root); - BUG_ON(ret); - reloc_root = root->reloc_root; - - shared_level = ref_path->shared_level; - ref_path->shared_level = BTRFS_MAX_LEVEL - 1; - - keys = ref_path->node_keys; - nodes = ref_path->new_nodes; - memset(&keys[shared_level + 1], 0, - sizeof(*keys) * (BTRFS_MAX_LEVEL - shared_level - 1)); - memset(&nodes[shared_level + 1], 0, - sizeof(*nodes) * (BTRFS_MAX_LEVEL - shared_level - 1)); - - if (nodes[lowest_level] == 0) { - path->lowest_level = lowest_level; - ret = btrfs_search_slot(trans, reloc_root, first_key, path, - 0, 1); - BUG_ON(ret); - for (level = lowest_level; level < BTRFS_MAX_LEVEL; level++) { - eb = path->nodes[level]; - if (!eb || eb == reloc_root->node) - break; - nodes[level] = eb->start; - if (level == 0) - btrfs_item_key_to_cpu(eb, &keys[level], 0); - else - btrfs_node_key_to_cpu(eb, &keys[level], 0); - } - if (nodes[0] && - ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - eb = path->nodes[0]; - ret = replace_extents_in_leaf(trans, reloc_root, eb, - group, reloc_inode); - BUG_ON(ret); - } - btrfs_release_path(reloc_root, path); - } else { - ret = btrfs_merge_path(trans, reloc_root, keys, nodes, - lowest_level); - BUG_ON(ret); - } - - /* - * replace tree blocks in the fs tree with tree blocks in - * the reloc tree. - */ - ret = btrfs_merge_path(trans, root, keys, nodes, lowest_level); - BUG_ON(ret < 0); - - if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - ret = btrfs_search_slot(trans, reloc_root, first_key, path, - 0, 0); - BUG_ON(ret); - extent_buffer_get(path->nodes[0]); - eb = path->nodes[0]; - btrfs_release_path(reloc_root, path); - ret = invalidate_extent_cache(reloc_root, eb, group, root); - BUG_ON(ret); - free_extent_buffer(eb); - } - - mutex_unlock(&root->fs_info->tree_reloc_mutex); - path->lowest_level = 0; - return 0; -} - -static noinline int relocate_tree_block(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct btrfs_key *first_key, - struct btrfs_ref_path *ref_path) -{ - int ret; - - ret = relocate_one_path(trans, root, path, first_key, - ref_path, NULL, NULL); - BUG_ON(ret); - - return 0; -} - -static noinline int del_extent_zero(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct btrfs_path *path, - struct btrfs_key *extent_key) -{ - int ret; - - ret = btrfs_search_slot(trans, extent_root, extent_key, path, -1, 1); - if (ret) - goto out; - ret = btrfs_del_item(trans, extent_root, path); -out: - btrfs_release_path(extent_root, path); - return ret; -} - -static noinline struct btrfs_root *read_ref_root(struct btrfs_fs_info *fs_info, - struct btrfs_ref_path *ref_path) -{ - struct btrfs_key root_key; - - root_key.objectid = ref_path->root_objectid; - root_key.type = BTRFS_ROOT_ITEM_KEY; - if (is_cowonly_root(ref_path->root_objectid)) - root_key.offset = 0; - else - root_key.offset = (u64)-1; - - return btrfs_read_fs_root_no_name(fs_info, &root_key); -} - -static noinline int relocate_one_extent(struct btrfs_root *extent_root, - struct btrfs_path *path, - struct btrfs_key *extent_key, - struct btrfs_block_group_cache *group, - struct inode *reloc_inode, int pass) -{ - struct btrfs_trans_handle *trans; - struct btrfs_root *found_root; - struct btrfs_ref_path *ref_path = NULL; - struct disk_extent *new_extents = NULL; - int nr_extents = 0; - int loops; - int ret; - int level; - struct btrfs_key first_key; - u64 prev_block = 0; - - - trans = btrfs_start_transaction(extent_root, 1); - BUG_ON(IS_ERR(trans)); - - if (extent_key->objectid == 0) { - ret = del_extent_zero(trans, extent_root, path, extent_key); - goto out; - } - - ref_path = kmalloc(sizeof(*ref_path), GFP_NOFS); - if (!ref_path) { - ret = -ENOMEM; - goto out; - } - - for (loops = 0; ; loops++) { - if (loops == 0) { - ret = btrfs_first_ref_path(trans, extent_root, ref_path, - extent_key->objectid); - } else { - ret = btrfs_next_ref_path(trans, extent_root, ref_path); - } - if (ret < 0) - goto out; - if (ret > 0) - break; - - if (ref_path->root_objectid == BTRFS_TREE_LOG_OBJECTID || - ref_path->root_objectid == BTRFS_TREE_RELOC_OBJECTID) - continue; - - found_root = read_ref_root(extent_root->fs_info, ref_path); - BUG_ON(!found_root); - /* - * for reference counted tree, only process reference paths - * rooted at the latest committed root. - */ - if (found_root->ref_cows && - ref_path->root_generation != found_root->root_key.offset) - continue; - - if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - if (pass == 0) { - /* - * copy data extents to new locations - */ - u64 group_start = group->key.objectid; - ret = relocate_data_extent(reloc_inode, - extent_key, - group_start); - if (ret < 0) - goto out; - break; - } - level = 0; - } else { - level = ref_path->owner_objectid; - } - - if (prev_block != ref_path->nodes[level]) { - struct extent_buffer *eb; - u64 block_start = ref_path->nodes[level]; - u64 block_size = btrfs_level_size(found_root, level); - - eb = read_tree_block(found_root, block_start, - block_size, 0); - if (!eb) { - ret = -EIO; - goto out; - } - btrfs_tree_lock(eb); - BUG_ON(level != btrfs_header_level(eb)); - - if (level == 0) - btrfs_item_key_to_cpu(eb, &first_key, 0); - else - btrfs_node_key_to_cpu(eb, &first_key, 0); - - btrfs_tree_unlock(eb); - free_extent_buffer(eb); - prev_block = block_start; - } - - mutex_lock(&extent_root->fs_info->trans_mutex); - btrfs_record_root_in_trans(found_root); - mutex_unlock(&extent_root->fs_info->trans_mutex); - if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) { - /* - * try to update data extent references while - * keeping metadata shared between snapshots. - */ - if (pass == 1) { - ret = relocate_one_path(trans, found_root, - path, &first_key, ref_path, - group, reloc_inode); - if (ret < 0) - goto out; - continue; - } - /* - * use fallback method to process the remaining - * references. - */ - if (!new_extents) { - u64 group_start = group->key.objectid; - new_extents = kmalloc(sizeof(*new_extents), - GFP_NOFS); - if (!new_extents) { - ret = -ENOMEM; - goto out; - } - nr_extents = 1; - ret = get_new_locations(reloc_inode, - extent_key, - group_start, 1, - &new_extents, - &nr_extents); - if (ret) - goto out; - } - ret = replace_one_extent(trans, found_root, - path, extent_key, - &first_key, ref_path, - new_extents, nr_extents); - } else { - ret = relocate_tree_block(trans, found_root, path, - &first_key, ref_path); - } - if (ret < 0) - goto out; - } - ret = 0; -out: - btrfs_end_transaction(trans, extent_root); - kfree(new_extents); - kfree(ref_path); - return ret; -} -#endif - static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) { u64 num_devices; @@ -8176,7 +6563,7 @@ int btrfs_set_block_group_ro(struct btrfs_root *root, BUG_ON(cache->ro); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); alloc_flags = update_block_group_flags(root, cache->flags); @@ -8532,6 +6919,7 @@ int btrfs_read_block_groups(struct btrfs_root *root) path = btrfs_alloc_path(); if (!path) return -ENOMEM; + path->reada = 1; cache_gen = btrfs_super_cache_generation(&root->fs_info->super_copy); if (cache_gen != 0 && @@ -8555,10 +6943,16 @@ int btrfs_read_block_groups(struct btrfs_root *root) ret = -ENOMEM; goto error; } + cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl), + GFP_NOFS); + if (!cache->free_space_ctl) { + kfree(cache); + ret = -ENOMEM; + goto error; + } atomic_set(&cache->count, 1); spin_lock_init(&cache->lock); - spin_lock_init(&cache->tree_lock); cache->fs_info = info; INIT_LIST_HEAD(&cache->list); INIT_LIST_HEAD(&cache->cluster_list); @@ -8566,24 +6960,18 @@ int btrfs_read_block_groups(struct btrfs_root *root) if (need_clear) cache->disk_cache_state = BTRFS_DC_CLEAR; - /* - * we only want to have 32k of ram per block group for keeping - * track of free space, and if we pass 1/2 of that we want to - * start converting things over to using bitmaps - */ - cache->extents_thresh = ((1024 * 32) / 2) / - sizeof(struct btrfs_free_space); - read_extent_buffer(leaf, &cache->item, btrfs_item_ptr_offset(leaf, path->slots[0]), sizeof(cache->item)); memcpy(&cache->key, &found_key, sizeof(found_key)); key.objectid = found_key.objectid + found_key.offset; - btrfs_release_path(root, path); + btrfs_release_path(path); cache->flags = btrfs_block_group_flags(&cache->item); cache->sectorsize = root->sectorsize; + btrfs_init_free_space_ctl(cache); + /* * We need to exclude the super stripes now so that the space * info has super bytes accounted for, otherwise we'll think @@ -8670,6 +7058,12 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, cache = kzalloc(sizeof(*cache), GFP_NOFS); if (!cache) return -ENOMEM; + cache->free_space_ctl = kzalloc(sizeof(*cache->free_space_ctl), + GFP_NOFS); + if (!cache->free_space_ctl) { + kfree(cache); + return -ENOMEM; + } cache->key.objectid = chunk_offset; cache->key.offset = size; @@ -8677,19 +7071,13 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, cache->sectorsize = root->sectorsize; cache->fs_info = root->fs_info; - /* - * we only want to have 32k of ram per block group for keeping track - * of free space, and if we pass 1/2 of that we want to start - * converting things over to using bitmaps - */ - cache->extents_thresh = ((1024 * 32) / 2) / - sizeof(struct btrfs_free_space); atomic_set(&cache->count, 1); spin_lock_init(&cache->lock); - spin_lock_init(&cache->tree_lock); INIT_LIST_HEAD(&cache->list); INIT_LIST_HEAD(&cache->cluster_list); + btrfs_init_free_space_ctl(cache); + btrfs_set_block_group_used(&cache->item, bytes_used); btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid); cache->flags = type; @@ -8802,12 +7190,12 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, if (ret < 0) goto out; if (ret > 0) - btrfs_release_path(tree_root, path); + btrfs_release_path(path); if (ret == 0) { ret = btrfs_del_item(trans, tree_root, path); if (ret) goto out; - btrfs_release_path(tree_root, path); + btrfs_release_path(path); } spin_lock(&root->fs_info->block_group_cache_lock); diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 4f9893243dae..7055d11c1efd 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -103,7 +103,7 @@ void extent_io_exit(void) } void extent_io_tree_init(struct extent_io_tree *tree, - struct address_space *mapping, gfp_t mask) + struct address_space *mapping) { tree->state = RB_ROOT; INIT_RADIX_TREE(&tree->buffer, GFP_ATOMIC); @@ -441,6 +441,15 @@ static int clear_state_bit(struct extent_io_tree *tree, return ret; } +static struct extent_state * +alloc_extent_state_atomic(struct extent_state *prealloc) +{ + if (!prealloc) + prealloc = alloc_extent_state(GFP_ATOMIC); + + return prealloc; +} + /* * clear some bits on a range in the tree. This may require splitting * or inserting elements in the tree, so the gfp mask is used to @@ -531,8 +540,8 @@ hit_next: */ if (state->start < start) { - if (!prealloc) - prealloc = alloc_extent_state(GFP_ATOMIC); + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, start); BUG_ON(err == -EEXIST); prealloc = NULL; @@ -553,8 +562,8 @@ hit_next: * on the first half */ if (state->start <= end && state->end > end) { - if (!prealloc) - prealloc = alloc_extent_state(GFP_ATOMIC); + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, end + 1); BUG_ON(err == -EEXIST); if (wake) @@ -727,8 +736,7 @@ int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, again: if (!prealloc && (mask & __GFP_WAIT)) { prealloc = alloc_extent_state(mask); - if (!prealloc) - return -ENOMEM; + BUG_ON(!prealloc); } spin_lock(&tree->lock); @@ -745,6 +753,8 @@ again: */ node = tree_search(tree, start); if (!node) { + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = insert_state(tree, prealloc, start, end, &bits); prealloc = NULL; BUG_ON(err == -EEXIST); @@ -773,20 +783,18 @@ hit_next: if (err) goto out; + next_node = rb_next(node); cache_state(state, cached_state); merge_state(tree, state); if (last_end == (u64)-1) goto out; start = last_end + 1; - if (start < end && prealloc && !need_resched()) { - next_node = rb_next(node); - if (next_node) { - state = rb_entry(next_node, struct extent_state, - rb_node); - if (state->start == start) - goto hit_next; - } + if (next_node && start < end && prealloc && !need_resched()) { + state = rb_entry(next_node, struct extent_state, + rb_node); + if (state->start == start) + goto hit_next; } goto search_again; } @@ -813,6 +821,9 @@ hit_next: err = -EEXIST; goto out; } + + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, start); BUG_ON(err == -EEXIST); prealloc = NULL; @@ -843,14 +854,25 @@ hit_next: this_end = end; else this_end = last_start - 1; + + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); + + /* + * Avoid to free 'prealloc' if it can be merged with + * the later extent. + */ + atomic_inc(&prealloc->refs); err = insert_state(tree, prealloc, start, this_end, &bits); BUG_ON(err == -EEXIST); if (err) { + free_extent_state(prealloc); prealloc = NULL; goto out; } cache_state(prealloc, cached_state); + free_extent_state(prealloc); prealloc = NULL; start = this_end + 1; goto search_again; @@ -867,6 +889,9 @@ hit_next: err = -EEXIST; goto out; } + + prealloc = alloc_extent_state_atomic(prealloc); + BUG_ON(!prealloc); err = split_state(tree, state, prealloc, end + 1); BUG_ON(err == -EEXIST); @@ -943,13 +968,6 @@ int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end, NULL, mask); } -static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask) -{ - return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, - NULL, mask); -} - int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask) { @@ -965,11 +983,6 @@ static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, cached_state, mask); } -int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end) -{ - return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK); -} - /* * either insert or lock state struct between start and end use mask to tell * us if waiting is desired. @@ -1030,25 +1043,6 @@ int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask) } /* - * helper function to set pages and extents in the tree dirty - */ -int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end) -{ - unsigned long index = start >> PAGE_CACHE_SHIFT; - unsigned long end_index = end >> PAGE_CACHE_SHIFT; - struct page *page; - - while (index <= end_index) { - page = find_get_page(tree->mapping, index); - BUG_ON(!page); - __set_page_dirty_nobuffers(page); - page_cache_release(page); - index++; - } - return 0; -} - -/* * helper function to set both pages and extents in the tree writeback */ static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end) @@ -1482,7 +1476,7 @@ u64 count_range_bits(struct extent_io_tree *tree, if (total_bytes >= max_bytes) break; if (!found) { - *start = state->start; + *start = max(cur_start, state->start); found = 1; } last = state->end; @@ -1821,46 +1815,6 @@ static void end_bio_extent_readpage(struct bio *bio, int err) bio_put(bio); } -/* - * IO done from prepare_write is pretty simple, we just unlock - * the structs in the extent tree when done, and set the uptodate bits - * as appropriate. - */ -static void end_bio_extent_preparewrite(struct bio *bio, int err) -{ - const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); - struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1; - struct extent_io_tree *tree; - u64 start; - u64 end; - - do { - struct page *page = bvec->bv_page; - struct extent_state *cached = NULL; - tree = &BTRFS_I(page->mapping->host)->io_tree; - - start = ((u64)page->index << PAGE_CACHE_SHIFT) + - bvec->bv_offset; - end = start + bvec->bv_len - 1; - - if (--bvec >= bio->bi_io_vec) - prefetchw(&bvec->bv_page->flags); - - if (uptodate) { - set_extent_uptodate(tree, start, end, &cached, - GFP_ATOMIC); - } else { - ClearPageUptodate(page); - SetPageError(page); - } - - unlock_extent_cached(tree, start, end, &cached, GFP_ATOMIC); - - } while (bvec >= bio->bi_io_vec); - - bio_put(bio); -} - struct bio * btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, gfp_t gfp_flags) @@ -2009,7 +1963,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, struct btrfs_ordered_extent *ordered; int ret; int nr = 0; - size_t page_offset = 0; + size_t pg_offset = 0; size_t iosize; size_t disk_io_size; size_t blocksize = inode->i_sb->s_blocksize; @@ -2052,9 +2006,9 @@ static int __extent_read_full_page(struct extent_io_tree *tree, char *userpage; struct extent_state *cached = NULL; - iosize = PAGE_CACHE_SIZE - page_offset; + iosize = PAGE_CACHE_SIZE - pg_offset; userpage = kmap_atomic(page, KM_USER0); - memset(userpage + page_offset, 0, iosize); + memset(userpage + pg_offset, 0, iosize); flush_dcache_page(page); kunmap_atomic(userpage, KM_USER0); set_extent_uptodate(tree, cur, cur + iosize - 1, @@ -2063,9 +2017,9 @@ static int __extent_read_full_page(struct extent_io_tree *tree, &cached, GFP_NOFS); break; } - em = get_extent(inode, page, page_offset, cur, + em = get_extent(inode, page, pg_offset, cur, end - cur + 1, 0); - if (IS_ERR(em) || !em) { + if (IS_ERR_OR_NULL(em)) { SetPageError(page); unlock_extent(tree, cur, end, GFP_NOFS); break; @@ -2103,7 +2057,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, struct extent_state *cached = NULL; userpage = kmap_atomic(page, KM_USER0); - memset(userpage + page_offset, 0, iosize); + memset(userpage + pg_offset, 0, iosize); flush_dcache_page(page); kunmap_atomic(userpage, KM_USER0); @@ -2112,7 +2066,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, unlock_extent_cached(tree, cur, cur + iosize - 1, &cached, GFP_NOFS); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; continue; } /* the get_extent function already copied into the page */ @@ -2121,7 +2075,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, check_page_uptodate(tree, page); unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; continue; } /* we have an inline extent but it didn't get marked up @@ -2131,7 +2085,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, SetPageError(page); unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; continue; } @@ -2144,7 +2098,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1; pnr -= page->index; ret = submit_extent_page(READ, tree, page, - sector, disk_io_size, page_offset, + sector, disk_io_size, pg_offset, bdev, bio, pnr, end_bio_extent_readpage, mirror_num, *bio_flags, @@ -2155,7 +2109,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree, if (ret) SetPageError(page); cur = cur + iosize; - page_offset += iosize; + pg_offset += iosize; } out: if (!nr) { @@ -2351,7 +2305,7 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc, } em = epd->get_extent(inode, page, pg_offset, cur, end - cur + 1, 1); - if (IS_ERR(em) || !em) { + if (IS_ERR_OR_NULL(em)) { SetPageError(page); break; } @@ -2730,128 +2684,6 @@ int extent_invalidatepage(struct extent_io_tree *tree, } /* - * simple commit_write call, set_range_dirty is used to mark both - * the pages and the extent records as dirty - */ -int extent_commit_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to) -{ - loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to; - - set_page_extent_mapped(page); - set_page_dirty(page); - - if (pos > inode->i_size) { - i_size_write(inode, pos); - mark_inode_dirty(inode); - } - return 0; -} - -int extent_prepare_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to, get_extent_t *get_extent) -{ - u64 page_start = (u64)page->index << PAGE_CACHE_SHIFT; - u64 page_end = page_start + PAGE_CACHE_SIZE - 1; - u64 block_start; - u64 orig_block_start; - u64 block_end; - u64 cur_end; - struct extent_map *em; - unsigned blocksize = 1 << inode->i_blkbits; - size_t page_offset = 0; - size_t block_off_start; - size_t block_off_end; - int err = 0; - int iocount = 0; - int ret = 0; - int isnew; - - set_page_extent_mapped(page); - - block_start = (page_start + from) & ~((u64)blocksize - 1); - block_end = (page_start + to - 1) | (blocksize - 1); - orig_block_start = block_start; - - lock_extent(tree, page_start, page_end, GFP_NOFS); - while (block_start <= block_end) { - em = get_extent(inode, page, page_offset, block_start, - block_end - block_start + 1, 1); - if (IS_ERR(em) || !em) - goto err; - - cur_end = min(block_end, extent_map_end(em) - 1); - block_off_start = block_start & (PAGE_CACHE_SIZE - 1); - block_off_end = block_off_start + blocksize; - isnew = clear_extent_new(tree, block_start, cur_end, GFP_NOFS); - - if (!PageUptodate(page) && isnew && - (block_off_end > to || block_off_start < from)) { - void *kaddr; - - kaddr = kmap_atomic(page, KM_USER0); - if (block_off_end > to) - memset(kaddr + to, 0, block_off_end - to); - if (block_off_start < from) - memset(kaddr + block_off_start, 0, - from - block_off_start); - flush_dcache_page(page); - kunmap_atomic(kaddr, KM_USER0); - } - if ((em->block_start != EXTENT_MAP_HOLE && - em->block_start != EXTENT_MAP_INLINE) && - !isnew && !PageUptodate(page) && - (block_off_end > to || block_off_start < from) && - !test_range_bit(tree, block_start, cur_end, - EXTENT_UPTODATE, 1, NULL)) { - u64 sector; - u64 extent_offset = block_start - em->start; - size_t iosize; - sector = (em->block_start + extent_offset) >> 9; - iosize = (cur_end - block_start + blocksize) & - ~((u64)blocksize - 1); - /* - * we've already got the extent locked, but we - * need to split the state such that our end_bio - * handler can clear the lock. - */ - set_extent_bit(tree, block_start, - block_start + iosize - 1, - EXTENT_LOCKED, 0, NULL, NULL, GFP_NOFS); - ret = submit_extent_page(READ, tree, page, - sector, iosize, page_offset, em->bdev, - NULL, 1, - end_bio_extent_preparewrite, 0, - 0, 0); - if (ret && !err) - err = ret; - iocount++; - block_start = block_start + iosize; - } else { - struct extent_state *cached = NULL; - - set_extent_uptodate(tree, block_start, cur_end, &cached, - GFP_NOFS); - unlock_extent_cached(tree, block_start, cur_end, - &cached, GFP_NOFS); - block_start = cur_end + 1; - } - page_offset = block_start & (PAGE_CACHE_SIZE - 1); - free_extent_map(em); - } - if (iocount) { - wait_extent_bit(tree, orig_block_start, - block_end, EXTENT_LOCKED); - } - check_page_uptodate(tree, page); -err: - /* FIXME, zero out newly allocated blocks on error */ - return err; -} - -/* * a helper for releasepage, this tests for areas of the page that * are locked or under IO and drops the related state bits if it is safe * to drop the page. @@ -2909,7 +2741,7 @@ int try_release_extent_mapping(struct extent_map_tree *map, len = end - start + 1; write_lock(&map->lock); em = lookup_extent_mapping(map, start, len); - if (!em || IS_ERR(em)) { + if (IS_ERR_OR_NULL(em)) { write_unlock(&map->lock); break; } @@ -2937,33 +2769,6 @@ int try_release_extent_mapping(struct extent_map_tree *map, return try_release_extent_state(map, tree, page, mask); } -sector_t extent_bmap(struct address_space *mapping, sector_t iblock, - get_extent_t *get_extent) -{ - struct inode *inode = mapping->host; - struct extent_state *cached_state = NULL; - u64 start = iblock << inode->i_blkbits; - sector_t sector = 0; - size_t blksize = (1 << inode->i_blkbits); - struct extent_map *em; - - lock_extent_bits(&BTRFS_I(inode)->io_tree, start, start + blksize - 1, - 0, &cached_state, GFP_NOFS); - em = get_extent(inode, NULL, 0, start, blksize, 0); - unlock_extent_cached(&BTRFS_I(inode)->io_tree, start, - start + blksize - 1, &cached_state, GFP_NOFS); - if (!em || IS_ERR(em)) - return 0; - - if (em->block_start > EXTENT_MAP_LAST_BYTE) - goto out; - - sector = (em->block_start + start - em->start) >> inode->i_blkbits; -out: - free_extent_map(em); - return sector; -} - /* * helper function for fiemap, which doesn't want to see any holes. * This maps until we find something past 'last' @@ -2986,7 +2791,7 @@ static struct extent_map *get_extent_skip_holes(struct inode *inode, break; len = (len + sectorsize - 1) & ~(sectorsize - 1); em = get_extent(inode, NULL, 0, offset, len, 0); - if (!em || IS_ERR(em)) + if (IS_ERR_OR_NULL(em)) return em; /* if this isn't a hole return it */ @@ -3040,7 +2845,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, * because there might be preallocation past i_size */ ret = btrfs_lookup_file_extent(NULL, BTRFS_I(inode)->root, - path, inode->i_ino, -1, 0); + path, btrfs_ino(inode), -1, 0); if (ret < 0) { btrfs_free_path(path); return ret; @@ -3053,7 +2858,7 @@ int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, found_type = btrfs_key_type(&found_key); /* No extents, but there might be delalloc bits */ - if (found_key.objectid != inode->i_ino || + if (found_key.objectid != btrfs_ino(inode) || found_type != BTRFS_EXTENT_DATA_KEY) { /* have to trust i_size as the end */ last = (u64)-1; @@ -3276,8 +3081,7 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb) struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len, - struct page *page0, - gfp_t mask) + struct page *page0) { unsigned long num_pages = num_extent_pages(start, len); unsigned long i; @@ -3298,7 +3102,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, } rcu_read_unlock(); - eb = __alloc_extent_buffer(tree, start, len, mask); + eb = __alloc_extent_buffer(tree, start, len, GFP_NOFS); if (!eb) return NULL; @@ -3315,7 +3119,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, i = 0; } for (; i < num_pages; i++, index++) { - p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM); + p = find_or_create_page(mapping, index, GFP_NOFS | __GFP_HIGHMEM); if (!p) { WARN_ON(1); goto free_eb; @@ -3387,8 +3191,7 @@ free_eb: } struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len, - gfp_t mask) + u64 start, unsigned long len) { struct extent_buffer *eb; @@ -3449,13 +3252,6 @@ int clear_extent_buffer_dirty(struct extent_io_tree *tree, return 0; } -int wait_on_extent_buffer_writeback(struct extent_io_tree *tree, - struct extent_buffer *eb) -{ - return wait_on_extent_writeback(tree, eb->start, - eb->start + eb->len - 1); -} - int set_extent_buffer_dirty(struct extent_io_tree *tree, struct extent_buffer *eb) { diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index af2d7179c372..a11a92ee2d30 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -126,9 +126,9 @@ struct extent_buffer { unsigned long map_len; struct page *first_page; unsigned long bflags; - atomic_t refs; struct list_head leak_list; struct rcu_head rcu_head; + atomic_t refs; /* the spinlock is used to protect most operations */ spinlock_t lock; @@ -153,23 +153,14 @@ static inline int extent_compress_type(unsigned long bio_flags) struct extent_map_tree; -static inline struct extent_state *extent_state_next(struct extent_state *state) -{ - struct rb_node *node; - node = rb_next(&state->rb_node); - if (!node) - return NULL; - return rb_entry(node, struct extent_state, rb_node); -} - typedef struct extent_map *(get_extent_t)(struct inode *inode, struct page *page, - size_t page_offset, + size_t pg_offset, u64 start, u64 len, int create); void extent_io_tree_init(struct extent_io_tree *tree, - struct address_space *mapping, gfp_t mask); + struct address_space *mapping); int try_release_extent_mapping(struct extent_map_tree *map, struct extent_io_tree *tree, struct page *page, gfp_t mask); @@ -215,14 +206,8 @@ int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask); -int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask); -int clear_extent_ordered_metadata(struct extent_io_tree *tree, u64 start, - u64 end, gfp_t mask); int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end, struct extent_state **cached_state, gfp_t mask); -int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end, - gfp_t mask); int find_first_extent_bit(struct extent_io_tree *tree, u64 start, u64 *start_ret, u64 *end_ret, int bits); struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree, @@ -243,28 +228,17 @@ int extent_readpages(struct extent_io_tree *tree, struct address_space *mapping, struct list_head *pages, unsigned nr_pages, get_extent_t get_extent); -int extent_prepare_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to, get_extent_t *get_extent); -int extent_commit_write(struct extent_io_tree *tree, - struct inode *inode, struct page *page, - unsigned from, unsigned to); -sector_t extent_bmap(struct address_space *mapping, sector_t iblock, - get_extent_t *get_extent); int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, __u64 start, __u64 len, get_extent_t *get_extent); -int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end); int set_state_private(struct extent_io_tree *tree, u64 start, u64 private); int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private); void set_page_extent_mapped(struct page *page); struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree, u64 start, unsigned long len, - struct page *page0, - gfp_t mask); + struct page *page0); struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree, - u64 start, unsigned long len, - gfp_t mask); + u64 start, unsigned long len); void free_extent_buffer(struct extent_buffer *eb); int read_extent_buffer_pages(struct extent_io_tree *tree, struct extent_buffer *eb, u64 start, int wait, @@ -292,16 +266,11 @@ void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset, unsigned long src_offset, unsigned long len); void memset_extent_buffer(struct extent_buffer *eb, char c, unsigned long start, unsigned long len); -int wait_on_extent_buffer_writeback(struct extent_io_tree *tree, - struct extent_buffer *eb); -int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end); int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits); int clear_extent_buffer_dirty(struct extent_io_tree *tree, struct extent_buffer *eb); int set_extent_buffer_dirty(struct extent_io_tree *tree, struct extent_buffer *eb); -int test_extent_buffer_dirty(struct extent_io_tree *tree, - struct extent_buffer *eb); int set_extent_buffer_uptodate(struct extent_io_tree *tree, struct extent_buffer *eb); int clear_extent_buffer_uptodate(struct extent_io_tree *tree, @@ -319,7 +288,6 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long offset, unsigned long *map_start, unsigned long *map_len, int km); void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km); -int release_extent_buffer_tail_pages(struct extent_buffer *eb); int extent_range_uptodate(struct extent_io_tree *tree, u64 start, u64 end); int extent_clear_unlock_delalloc(struct inode *inode, diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c index a24a3f2fa13e..2d0410344ea3 100644 --- a/fs/btrfs/extent_map.c +++ b/fs/btrfs/extent_map.c @@ -28,12 +28,11 @@ void extent_map_exit(void) /** * extent_map_tree_init - initialize extent map tree * @tree: tree to initialize - * @mask: flags for memory allocations during tree operations * * Initialize the extent tree @tree. Should be called for each new inode * or other user of the extent_map interface. */ -void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask) +void extent_map_tree_init(struct extent_map_tree *tree) { tree->map = RB_ROOT; rwlock_init(&tree->lock); @@ -41,16 +40,15 @@ void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask) /** * alloc_extent_map - allocate new extent map structure - * @mask: memory allocation flags * * Allocate a new extent_map structure. The new structure is * returned with a reference count of one and needs to be * freed using free_extent_map() */ -struct extent_map *alloc_extent_map(gfp_t mask) +struct extent_map *alloc_extent_map(void) { struct extent_map *em; - em = kmem_cache_alloc(extent_map_cache, mask); + em = kmem_cache_alloc(extent_map_cache, GFP_NOFS); if (!em) return NULL; em->in_tree = 0; diff --git a/fs/btrfs/extent_map.h b/fs/btrfs/extent_map.h index 28b44dbd1e35..33a7890b1f40 100644 --- a/fs/btrfs/extent_map.h +++ b/fs/btrfs/extent_map.h @@ -49,14 +49,14 @@ static inline u64 extent_map_block_end(struct extent_map *em) return em->block_start + em->block_len; } -void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask); +void extent_map_tree_init(struct extent_map_tree *tree); struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree, u64 start, u64 len); int add_extent_mapping(struct extent_map_tree *tree, struct extent_map *em); int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em); -struct extent_map *alloc_extent_map(gfp_t mask); +struct extent_map *alloc_extent_map(void); void free_extent_map(struct extent_map *em); int __init extent_map_init(void); void extent_map_exit(void); diff --git a/fs/btrfs/file-item.c b/fs/btrfs/file-item.c index a6a9d4e8b491..90d4ee52cd45 100644 --- a/fs/btrfs/file-item.c +++ b/fs/btrfs/file-item.c @@ -193,7 +193,7 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, u32 item_size; if (item) - btrfs_release_path(root, path); + btrfs_release_path(path); item = btrfs_lookup_csum(NULL, root->fs_info->csum_root, path, disk_bytenr, 0); if (IS_ERR(item)) { @@ -208,12 +208,13 @@ static int __btrfs_lookup_bio_sums(struct btrfs_root *root, EXTENT_NODATASUM, GFP_NOFS); } else { printk(KERN_INFO "btrfs no csum found " - "for inode %lu start %llu\n", - inode->i_ino, + "for inode %llu start %llu\n", + (unsigned long long) + btrfs_ino(inode), (unsigned long long)offset); } item = NULL; - btrfs_release_path(root, path); + btrfs_release_path(path); goto found; } btrfs_item_key_to_cpu(path->nodes[0], &found_key, @@ -266,7 +267,7 @@ int btrfs_lookup_bio_sums_dio(struct btrfs_root *root, struct inode *inode, } int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, - struct list_head *list) + struct list_head *list, int search_commit) { struct btrfs_key key; struct btrfs_path *path; @@ -283,6 +284,12 @@ int btrfs_lookup_csums_range(struct btrfs_root *root, u64 start, u64 end, path = btrfs_alloc_path(); BUG_ON(!path); + if (search_commit) { + path->skip_locking = 1; + path->reada = 2; + path->search_commit_root = 1; + } + key.objectid = BTRFS_EXTENT_CSUM_OBJECTID; key.offset = start; key.type = BTRFS_EXTENT_CSUM_KEY; @@ -495,7 +502,6 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, u32 new_size = (bytenr - key->offset) >> blocksize_bits; new_size *= csum_size; ret = btrfs_truncate_item(trans, root, path, new_size, 1); - BUG_ON(ret); } else if (key->offset >= bytenr && csum_end > end_byte && end_byte > key->offset) { /* @@ -508,7 +514,6 @@ static noinline int truncate_one_csum(struct btrfs_trans_handle *trans, new_size *= csum_size; ret = btrfs_truncate_item(trans, root, path, new_size, 0); - BUG_ON(ret); key->offset = end_byte; ret = btrfs_set_item_key_safe(trans, root, path, key); @@ -551,10 +556,10 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, ret = btrfs_search_slot(trans, root, &key, path, -1, 1); if (ret > 0) { if (path->slots[0] == 0) - goto out; + break; path->slots[0]--; } else if (ret < 0) { - goto out; + break; } leaf = path->nodes[0]; @@ -579,7 +584,8 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, /* delete the entire item, it is inside our range */ if (key.offset >= bytenr && csum_end <= end_byte) { ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); + if (ret) + goto out; if (key.offset == bytenr) break; } else if (key.offset < bytenr && csum_end > end_byte) { @@ -631,11 +637,12 @@ int btrfs_del_csums(struct btrfs_trans_handle *trans, if (key.offset < bytenr) break; } - btrfs_release_path(root, path); + btrfs_release_path(path); } + ret = 0; out: btrfs_free_path(path); - return 0; + return ret; } int btrfs_csum_file_blocks(struct btrfs_trans_handle *trans, @@ -722,7 +729,7 @@ again: * at this point, we know the tree has an item, but it isn't big * enough yet to put our csum in. Grow it */ - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(trans, root, &file_key, path, csum_size, 1); if (ret < 0) @@ -761,12 +768,11 @@ again: goto insert; ret = btrfs_extend_item(trans, root, path, diff); - BUG_ON(ret); goto csum; } insert: - btrfs_release_path(root, path); + btrfs_release_path(path); csum_offset = 0; if (found_next) { u64 tmp = total_bytes + root->sectorsize; @@ -850,7 +856,7 @@ next_sector: } btrfs_mark_buffer_dirty(path->nodes[0]); if (total_bytes < sums->len) { - btrfs_release_path(root, path); + btrfs_release_path(path); cond_resched(); goto again; } diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 75899a01dded..fa4ef18b66b1 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -40,6 +40,263 @@ #include "locking.h" #include "compat.h" +/* + * when auto defrag is enabled we + * queue up these defrag structs to remember which + * inodes need defragging passes + */ +struct inode_defrag { + struct rb_node rb_node; + /* objectid */ + u64 ino; + /* + * transid where the defrag was added, we search for + * extents newer than this + */ + u64 transid; + + /* root objectid */ + u64 root; + + /* last offset we were able to defrag */ + u64 last_offset; + + /* if we've wrapped around back to zero once already */ + int cycled; +}; + +/* pop a record for an inode into the defrag tree. The lock + * must be held already + * + * If you're inserting a record for an older transid than an + * existing record, the transid already in the tree is lowered + * + * If an existing record is found the defrag item you + * pass in is freed + */ +static int __btrfs_add_inode_defrag(struct inode *inode, + struct inode_defrag *defrag) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + struct inode_defrag *entry; + struct rb_node **p; + struct rb_node *parent = NULL; + + p = &root->fs_info->defrag_inodes.rb_node; + while (*p) { + parent = *p; + entry = rb_entry(parent, struct inode_defrag, rb_node); + + if (defrag->ino < entry->ino) + p = &parent->rb_left; + else if (defrag->ino > entry->ino) + p = &parent->rb_right; + else { + /* if we're reinserting an entry for + * an old defrag run, make sure to + * lower the transid of our existing record + */ + if (defrag->transid < entry->transid) + entry->transid = defrag->transid; + if (defrag->last_offset > entry->last_offset) + entry->last_offset = defrag->last_offset; + goto exists; + } + } + BTRFS_I(inode)->in_defrag = 1; + rb_link_node(&defrag->rb_node, parent, p); + rb_insert_color(&defrag->rb_node, &root->fs_info->defrag_inodes); + return 0; + +exists: + kfree(defrag); + return 0; + +} + +/* + * insert a defrag record for this inode if auto defrag is + * enabled + */ +int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, + struct inode *inode) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; + struct inode_defrag *defrag; + int ret = 0; + u64 transid; + + if (!btrfs_test_opt(root, AUTO_DEFRAG)) + return 0; + + if (btrfs_fs_closing(root->fs_info)) + return 0; + + if (BTRFS_I(inode)->in_defrag) + return 0; + + if (trans) + transid = trans->transid; + else + transid = BTRFS_I(inode)->root->last_trans; + + defrag = kzalloc(sizeof(*defrag), GFP_NOFS); + if (!defrag) + return -ENOMEM; + + defrag->ino = btrfs_ino(inode); + defrag->transid = transid; + defrag->root = root->root_key.objectid; + + spin_lock(&root->fs_info->defrag_inodes_lock); + if (!BTRFS_I(inode)->in_defrag) + ret = __btrfs_add_inode_defrag(inode, defrag); + spin_unlock(&root->fs_info->defrag_inodes_lock); + return ret; +} + +/* + * must be called with the defrag_inodes lock held + */ +struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, u64 ino, + struct rb_node **next) +{ + struct inode_defrag *entry = NULL; + struct rb_node *p; + struct rb_node *parent = NULL; + + p = info->defrag_inodes.rb_node; + while (p) { + parent = p; + entry = rb_entry(parent, struct inode_defrag, rb_node); + + if (ino < entry->ino) + p = parent->rb_left; + else if (ino > entry->ino) + p = parent->rb_right; + else + return entry; + } + + if (next) { + while (parent && ino > entry->ino) { + parent = rb_next(parent); + entry = rb_entry(parent, struct inode_defrag, rb_node); + } + *next = parent; + } + return NULL; +} + +/* + * run through the list of inodes in the FS that need + * defragging + */ +int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) +{ + struct inode_defrag *defrag; + struct btrfs_root *inode_root; + struct inode *inode; + struct rb_node *n; + struct btrfs_key key; + struct btrfs_ioctl_defrag_range_args range; + u64 first_ino = 0; + int num_defrag; + int defrag_batch = 1024; + + memset(&range, 0, sizeof(range)); + range.len = (u64)-1; + + atomic_inc(&fs_info->defrag_running); + spin_lock(&fs_info->defrag_inodes_lock); + while(1) { + n = NULL; + + /* find an inode to defrag */ + defrag = btrfs_find_defrag_inode(fs_info, first_ino, &n); + if (!defrag) { + if (n) + defrag = rb_entry(n, struct inode_defrag, rb_node); + else if (first_ino) { + first_ino = 0; + continue; + } else { + break; + } + } + + /* remove it from the rbtree */ + first_ino = defrag->ino + 1; + rb_erase(&defrag->rb_node, &fs_info->defrag_inodes); + + if (btrfs_fs_closing(fs_info)) + goto next_free; + + spin_unlock(&fs_info->defrag_inodes_lock); + + /* get the inode */ + key.objectid = defrag->root; + btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY); + key.offset = (u64)-1; + inode_root = btrfs_read_fs_root_no_name(fs_info, &key); + if (IS_ERR(inode_root)) + goto next; + + key.objectid = defrag->ino; + btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); + key.offset = 0; + + inode = btrfs_iget(fs_info->sb, &key, inode_root, NULL); + if (IS_ERR(inode)) + goto next; + + /* do a chunk of defrag */ + BTRFS_I(inode)->in_defrag = 0; + range.start = defrag->last_offset; + num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid, + defrag_batch); + /* + * if we filled the whole defrag batch, there + * must be more work to do. Queue this defrag + * again + */ + if (num_defrag == defrag_batch) { + defrag->last_offset = range.start; + __btrfs_add_inode_defrag(inode, defrag); + /* + * we don't want to kfree defrag, we added it back to + * the rbtree + */ + defrag = NULL; + } else if (defrag->last_offset && !defrag->cycled) { + /* + * we didn't fill our defrag batch, but + * we didn't start at zero. Make sure we loop + * around to the start of the file. + */ + defrag->last_offset = 0; + defrag->cycled = 1; + __btrfs_add_inode_defrag(inode, defrag); + defrag = NULL; + } + + iput(inode); +next: + spin_lock(&fs_info->defrag_inodes_lock); +next_free: + kfree(defrag); + } + spin_unlock(&fs_info->defrag_inodes_lock); + + atomic_dec(&fs_info->defrag_running); + + /* + * during unmount, we use the transaction_wait queue to + * wait for the defragger to stop + */ + wake_up(&fs_info->transaction_wait); + return 0; +} /* simple helper to fault in pages and copy. This should go away * and be replaced with calls into generic code. @@ -191,9 +448,9 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, } while (1) { if (!split) - split = alloc_extent_map(GFP_NOFS); + split = alloc_extent_map(); if (!split2) - split2 = alloc_extent_map(GFP_NOFS); + split2 = alloc_extent_map(); BUG_ON(!split || !split2); write_lock(&em_tree->lock); @@ -298,6 +555,7 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, struct btrfs_path *path; struct btrfs_key key; struct btrfs_key new_key; + u64 ino = btrfs_ino(inode); u64 search_start = start; u64 disk_bytenr = 0; u64 num_bytes = 0; @@ -318,14 +576,14 @@ int btrfs_drop_extents(struct btrfs_trans_handle *trans, struct inode *inode, while (1) { recow = 0; - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, ino, search_start, -1); if (ret < 0) break; if (ret > 0 && path->slots[0] > 0 && search_start == start) { leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0] - 1); - if (key.objectid == inode->i_ino && + if (key.objectid == ino && key.type == BTRFS_EXTENT_DATA_KEY) path->slots[0]--; } @@ -346,7 +604,7 @@ next_slot: } btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - if (key.objectid > inode->i_ino || + if (key.objectid > ino || key.type > BTRFS_EXTENT_DATA_KEY || key.offset >= end) break; @@ -376,7 +634,7 @@ next_slot: search_start = max(key.offset, start); if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } @@ -393,7 +651,7 @@ next_slot: ret = btrfs_duplicate_item(trans, root, path, &new_key); if (ret == -EAGAIN) { - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } if (ret < 0) @@ -516,7 +774,7 @@ next_slot: del_nr = 0; del_slot = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } @@ -592,6 +850,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, int del_slot = 0; int recow; int ret; + u64 ino = btrfs_ino(inode); btrfs_drop_extent_cache(inode, start, end - 1, 0); @@ -600,7 +859,7 @@ int btrfs_mark_extent_written(struct btrfs_trans_handle *trans, again: recow = 0; split = start; - key.objectid = inode->i_ino; + key.objectid = ino; key.type = BTRFS_EXTENT_DATA_KEY; key.offset = split; @@ -612,8 +871,7 @@ again: leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - BUG_ON(key.objectid != inode->i_ino || - key.type != BTRFS_EXTENT_DATA_KEY); + BUG_ON(key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY); fi = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); BUG_ON(btrfs_file_extent_type(leaf, fi) != @@ -630,7 +888,7 @@ again: other_start = 0; other_end = start; if (extent_mergeable(leaf, path->slots[0] - 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { new_key.offset = end; btrfs_set_item_key_safe(trans, root, path, &new_key); @@ -653,7 +911,7 @@ again: other_start = end; other_end = 0; if (extent_mergeable(leaf, path->slots[0] + 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { fi = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_file_extent_item); @@ -681,7 +939,7 @@ again: new_key.offset = split; ret = btrfs_duplicate_item(trans, root, path, &new_key); if (ret == -EAGAIN) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } BUG_ON(ret < 0); @@ -702,7 +960,7 @@ again: ret = btrfs_inc_extent_ref(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - inode->i_ino, orig_offset); + ino, orig_offset); BUG_ON(ret); if (split == start) { @@ -718,10 +976,10 @@ again: other_start = end; other_end = 0; if (extent_mergeable(leaf, path->slots[0] + 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } extent_end = other_end; @@ -729,16 +987,16 @@ again: del_nr++; ret = btrfs_free_extent(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - inode->i_ino, orig_offset); + ino, orig_offset); BUG_ON(ret); } other_start = 0; other_end = start; if (extent_mergeable(leaf, path->slots[0] - 1, - inode->i_ino, bytenr, orig_offset, + ino, bytenr, orig_offset, &other_start, &other_end)) { if (recow) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } key.offset = other_start; @@ -746,7 +1004,7 @@ again: del_nr++; ret = btrfs_free_extent(trans, root, bytenr, num_bytes, 0, root->root_key.objectid, - inode->i_ino, orig_offset); + ino, orig_offset); BUG_ON(ret); } if (del_nr == 0) { @@ -1222,14 +1480,12 @@ int btrfs_sync_file(struct file *file, int datasync) * the current transaction, we can bail out now without any * syncing */ - mutex_lock(&root->fs_info->trans_mutex); + smp_mb(); if (BTRFS_I(inode)->last_trans <= root->fs_info->last_trans_committed) { BTRFS_I(inode)->last_trans = 0; - mutex_unlock(&root->fs_info->trans_mutex); goto out; } - mutex_unlock(&root->fs_info->trans_mutex); /* * ok we haven't committed the transaction yet, lets do a commit @@ -1375,7 +1631,7 @@ static long btrfs_fallocate(struct file *file, int mode, while (1) { em = btrfs_get_extent(inode, NULL, 0, cur_offset, alloc_end - cur_offset, 0); - BUG_ON(IS_ERR(em) || !em); + BUG_ON(IS_ERR_OR_NULL(em)); last_byte = min(extent_map_end(em), alloc_end); last_byte = (last_byte + mask) & ~mask; if (em->block_start == EXTENT_MAP_HOLE || diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index 63731a1fb0a1..bf0d61567f3d 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -25,18 +25,17 @@ #include "transaction.h" #include "disk-io.h" #include "extent_io.h" +#include "inode-map.h" #define BITS_PER_BITMAP (PAGE_CACHE_SIZE * 8) #define MAX_CACHE_BYTES_PER_GIG (32 * 1024) -static void recalculate_thresholds(struct btrfs_block_group_cache - *block_group); -static int link_free_space(struct btrfs_block_group_cache *block_group, +static int link_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info); -struct inode *lookup_free_space_inode(struct btrfs_root *root, - struct btrfs_block_group_cache - *block_group, struct btrfs_path *path) +static struct inode *__lookup_free_space_inode(struct btrfs_root *root, + struct btrfs_path *path, + u64 offset) { struct btrfs_key key; struct btrfs_key location; @@ -46,22 +45,15 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, struct inode *inode = NULL; int ret; - spin_lock(&block_group->lock); - if (block_group->inode) - inode = igrab(block_group->inode); - spin_unlock(&block_group->lock); - if (inode) - return inode; - key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) return ERR_PTR(ret); if (ret > 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return ERR_PTR(-ENOENT); } @@ -70,7 +62,7 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, struct btrfs_free_space_header); btrfs_free_space_key(leaf, header, &disk_key); btrfs_disk_key_to_cpu(&location, &disk_key); - btrfs_release_path(root, path); + btrfs_release_path(path); inode = btrfs_iget(root->fs_info->sb, &location, root, NULL); if (!inode) @@ -84,8 +76,29 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, inode->i_mapping->flags &= ~__GFP_FS; + return inode; +} + +struct inode *lookup_free_space_inode(struct btrfs_root *root, + struct btrfs_block_group_cache + *block_group, struct btrfs_path *path) +{ + struct inode *inode = NULL; + spin_lock(&block_group->lock); - if (!root->fs_info->closing) { + if (block_group->inode) + inode = igrab(block_group->inode); + spin_unlock(&block_group->lock); + if (inode) + return inode; + + inode = __lookup_free_space_inode(root, path, + block_group->key.objectid); + if (IS_ERR(inode)) + return inode; + + spin_lock(&block_group->lock); + if (!btrfs_fs_closing(root->fs_info)) { block_group->inode = igrab(inode); block_group->iref = 1; } @@ -94,24 +107,18 @@ struct inode *lookup_free_space_inode(struct btrfs_root *root, return inode; } -int create_free_space_inode(struct btrfs_root *root, - struct btrfs_trans_handle *trans, - struct btrfs_block_group_cache *block_group, - struct btrfs_path *path) +int __create_free_space_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path, u64 ino, u64 offset) { struct btrfs_key key; struct btrfs_disk_key disk_key; struct btrfs_free_space_header *header; struct btrfs_inode_item *inode_item; struct extent_buffer *leaf; - u64 objectid; int ret; - ret = btrfs_find_free_objectid(trans, root, 0, &objectid); - if (ret < 0) - return ret; - - ret = btrfs_insert_empty_inode(trans, root, path, objectid); + ret = btrfs_insert_empty_inode(trans, root, path, ino); if (ret) return ret; @@ -131,19 +138,18 @@ int create_free_space_inode(struct btrfs_root *root, BTRFS_INODE_PREALLOC | BTRFS_INODE_NODATASUM); btrfs_set_inode_nlink(leaf, inode_item, 1); btrfs_set_inode_transid(leaf, inode_item, trans->transid); - btrfs_set_inode_block_group(leaf, inode_item, - block_group->key.objectid); + btrfs_set_inode_block_group(leaf, inode_item, offset); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_insert_empty_item(trans, root, path, &key, sizeof(struct btrfs_free_space_header)); if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } leaf = path->nodes[0]; @@ -152,11 +158,27 @@ int create_free_space_inode(struct btrfs_root *root, memset_extent_buffer(leaf, 0, (unsigned long)header, sizeof(*header)); btrfs_set_free_space_key(leaf, header, &disk_key); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } +int create_free_space_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_block_group_cache *block_group, + struct btrfs_path *path) +{ + int ret; + u64 ino; + + ret = btrfs_find_free_objectid(root, &ino); + if (ret < 0) + return ret; + + return __create_free_space_inode(root, trans, path, ino, + block_group->key.objectid); +} + int btrfs_truncate_free_space_cache(struct btrfs_root *root, struct btrfs_trans_handle *trans, struct btrfs_path *path, @@ -187,7 +209,8 @@ int btrfs_truncate_free_space_cache(struct btrfs_root *root, return ret; } - return btrfs_update_inode(trans, root, inode); + ret = btrfs_update_inode(trans, root, inode); + return ret; } static int readahead_cache(struct inode *inode) @@ -209,15 +232,13 @@ static int readahead_cache(struct inode *inode) return 0; } -int load_free_space_cache(struct btrfs_fs_info *fs_info, - struct btrfs_block_group_cache *block_group) +int __load_free_space_cache(struct btrfs_root *root, struct inode *inode, + struct btrfs_free_space_ctl *ctl, + struct btrfs_path *path, u64 offset) { - struct btrfs_root *root = fs_info->tree_root; - struct inode *inode; struct btrfs_free_space_header *header; struct extent_buffer *leaf; struct page *page; - struct btrfs_path *path; u32 *checksums = NULL, *crc; char *disk_crcs = NULL; struct btrfs_key key; @@ -225,76 +246,47 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, u64 num_entries; u64 num_bitmaps; u64 generation; - u64 used = btrfs_block_group_used(&block_group->item); u32 cur_crc = ~(u32)0; pgoff_t index = 0; unsigned long first_page_offset; int num_checksums; int ret = 0; - /* - * If we're unmounting then just return, since this does a search on the - * normal root and not the commit root and we could deadlock. - */ - smp_mb(); - if (fs_info->closing) - return 0; - - /* - * If this block group has been marked to be cleared for one reason or - * another then we can't trust the on disk cache, so just return. - */ - spin_lock(&block_group->lock); - if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { - spin_unlock(&block_group->lock); - return 0; - } - spin_unlock(&block_group->lock); - INIT_LIST_HEAD(&bitmaps); - path = btrfs_alloc_path(); - if (!path) - return 0; - - inode = lookup_free_space_inode(root, block_group, path); - if (IS_ERR(inode)) { - btrfs_free_path(path); - return 0; - } - /* Nothing in the space cache, goodbye */ - if (!i_size_read(inode)) { - btrfs_free_path(path); + if (!i_size_read(inode)) goto out; - } key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret) { - btrfs_free_path(path); + if (ret < 0) + goto out; + else if (ret > 0) { + btrfs_release_path(path); + ret = 0; goto out; } + ret = -1; + leaf = path->nodes[0]; header = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_free_space_header); num_entries = btrfs_free_space_entries(leaf, header); num_bitmaps = btrfs_free_space_bitmaps(leaf, header); generation = btrfs_free_space_generation(leaf, header); - btrfs_free_path(path); + btrfs_release_path(path); if (BTRFS_I(inode)->generation != generation) { printk(KERN_ERR "btrfs: free space inode generation (%llu) did" - " not match free space cache generation (%llu) for " - "block group %llu\n", + " not match free space cache generation (%llu)\n", (unsigned long long)BTRFS_I(inode)->generation, - (unsigned long long)generation, - (unsigned long long)block_group->key.objectid); - goto free_cache; + (unsigned long long)generation); + goto out; } if (!num_entries) @@ -311,10 +303,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, goto out; ret = readahead_cache(inode); - if (ret) { - ret = 0; + if (ret) goto out; - } while (1) { struct btrfs_free_space_entry *entry; @@ -333,10 +323,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, } page = grab_cache_page(inode->i_mapping, index); - if (!page) { - ret = 0; + if (!page) goto free_cache; - } if (!PageUptodate(page)) { btrfs_readpage(NULL, page); @@ -345,9 +333,7 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, unlock_page(page); page_cache_release(page); printk(KERN_ERR "btrfs: error reading free " - "space cache: %llu\n", - (unsigned long long) - block_group->key.objectid); + "space cache\n"); goto free_cache; } } @@ -360,13 +346,10 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, gen = addr + (sizeof(u32) * num_checksums); if (*gen != BTRFS_I(inode)->generation) { printk(KERN_ERR "btrfs: space cache generation" - " (%llu) does not match inode (%llu) " - "for block group %llu\n", + " (%llu) does not match inode (%llu)\n", (unsigned long long)*gen, (unsigned long long) - BTRFS_I(inode)->generation, - (unsigned long long) - block_group->key.objectid); + BTRFS_I(inode)->generation); kunmap(page); unlock_page(page); page_cache_release(page); @@ -382,9 +365,8 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, PAGE_CACHE_SIZE - start_offset); btrfs_csum_final(cur_crc, (char *)&cur_crc); if (cur_crc != *crc) { - printk(KERN_ERR "btrfs: crc mismatch for page %lu in " - "block group %llu\n", index, - (unsigned long long)block_group->key.objectid); + printk(KERN_ERR "btrfs: crc mismatch for page %lu\n", + index); kunmap(page); unlock_page(page); page_cache_release(page); @@ -417,10 +399,17 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, } if (entry->type == BTRFS_FREE_SPACE_EXTENT) { - spin_lock(&block_group->tree_lock); - ret = link_free_space(block_group, e); - spin_unlock(&block_group->tree_lock); - BUG_ON(ret); + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + spin_unlock(&ctl->tree_lock); + if (ret) { + printk(KERN_ERR "Duplicate entries in " + "free space cache, dumping\n"); + kunmap(page); + unlock_page(page); + page_cache_release(page); + goto free_cache; + } } else { e->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); if (!e->bitmap) { @@ -431,11 +420,19 @@ int load_free_space_cache(struct btrfs_fs_info *fs_info, page_cache_release(page); goto free_cache; } - spin_lock(&block_group->tree_lock); - ret = link_free_space(block_group, e); - block_group->total_bitmaps++; - recalculate_thresholds(block_group); - spin_unlock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); + ret = link_free_space(ctl, e); + ctl->total_bitmaps++; + ctl->op->recalc_thresholds(ctl); + spin_unlock(&ctl->tree_lock); + if (ret) { + printk(KERN_ERR "Duplicate entries in " + "free space cache, dumping\n"); + kunmap(page); + unlock_page(page); + page_cache_release(page); + goto free_cache; + } list_add_tail(&e->list, &bitmaps); } @@ -471,41 +468,96 @@ next: index++; } - spin_lock(&block_group->tree_lock); - if (block_group->free_space != (block_group->key.offset - used - - block_group->bytes_super)) { - spin_unlock(&block_group->tree_lock); - printk(KERN_ERR "block group %llu has an wrong amount of free " - "space\n", block_group->key.objectid); - ret = 0; - goto free_cache; - } - spin_unlock(&block_group->tree_lock); - ret = 1; out: kfree(checksums); kfree(disk_crcs); - iput(inode); return ret; - free_cache: - /* This cache is bogus, make sure it gets cleared */ + __btrfs_remove_free_space_cache(ctl); + goto out; +} + +int load_free_space_cache(struct btrfs_fs_info *fs_info, + struct btrfs_block_group_cache *block_group) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct btrfs_root *root = fs_info->tree_root; + struct inode *inode; + struct btrfs_path *path; + int ret; + bool matched; + u64 used = btrfs_block_group_used(&block_group->item); + + /* + * If we're unmounting then just return, since this does a search on the + * normal root and not the commit root and we could deadlock. + */ + if (btrfs_fs_closing(fs_info)) + return 0; + + /* + * If this block group has been marked to be cleared for one reason or + * another then we can't trust the on disk cache, so just return. + */ spin_lock(&block_group->lock); - block_group->disk_cache_state = BTRFS_DC_CLEAR; + if (block_group->disk_cache_state != BTRFS_DC_WRITTEN) { + spin_unlock(&block_group->lock); + return 0; + } spin_unlock(&block_group->lock); - btrfs_remove_free_space_cache(block_group); - goto out; + + path = btrfs_alloc_path(); + if (!path) + return 0; + + inode = lookup_free_space_inode(root, block_group, path); + if (IS_ERR(inode)) { + btrfs_free_path(path); + return 0; + } + + ret = __load_free_space_cache(fs_info->tree_root, inode, ctl, + path, block_group->key.objectid); + btrfs_free_path(path); + if (ret <= 0) + goto out; + + spin_lock(&ctl->tree_lock); + matched = (ctl->free_space == (block_group->key.offset - used - + block_group->bytes_super)); + spin_unlock(&ctl->tree_lock); + + if (!matched) { + __btrfs_remove_free_space_cache(ctl); + printk(KERN_ERR "block group %llu has an wrong amount of free " + "space\n", block_group->key.objectid); + ret = -1; + } +out: + if (ret < 0) { + /* This cache is bogus, make sure it gets cleared */ + spin_lock(&block_group->lock); + block_group->disk_cache_state = BTRFS_DC_CLEAR; + spin_unlock(&block_group->lock); + ret = 0; + + printk(KERN_ERR "btrfs: failed to load free space cache " + "for block group %llu\n", block_group->key.objectid); + } + + iput(inode); + return ret; } -int btrfs_write_out_cache(struct btrfs_root *root, - struct btrfs_trans_handle *trans, - struct btrfs_block_group_cache *block_group, - struct btrfs_path *path) +int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode, + struct btrfs_free_space_ctl *ctl, + struct btrfs_block_group_cache *block_group, + struct btrfs_trans_handle *trans, + struct btrfs_path *path, u64 offset) { struct btrfs_free_space_header *header; struct extent_buffer *leaf; - struct inode *inode; struct rb_node *node; struct list_head *pos, *n; struct page **pages; @@ -522,64 +574,53 @@ int btrfs_write_out_cache(struct btrfs_root *root, int index = 0, num_pages = 0; int entries = 0; int bitmaps = 0; - int ret = 0; + int ret = -1; bool next_page = false; bool out_of_space = false; - root = root->fs_info->tree_root; - INIT_LIST_HEAD(&bitmap_list); - spin_lock(&block_group->lock); - if (block_group->disk_cache_state < BTRFS_DC_SETUP) { - spin_unlock(&block_group->lock); - return 0; - } - spin_unlock(&block_group->lock); - - inode = lookup_free_space_inode(root, block_group, path); - if (IS_ERR(inode)) - return 0; - - if (!i_size_read(inode)) { - iput(inode); + node = rb_first(&ctl->free_space_offset); + if (!node) return 0; - } - node = rb_first(&block_group->free_space_offset); - if (!node) { - iput(inode); - return 0; - } + if (!i_size_read(inode)) + return -1; num_pages = (i_size_read(inode) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; + + /* Since the first page has all of our checksums and our generation we + * need to calculate the offset into the page that we can start writing + * our entries. + */ + first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64); + filemap_write_and_wait(inode->i_mapping); btrfs_wait_ordered_range(inode, inode->i_size & ~(root->sectorsize - 1), (u64)-1); + /* make sure we don't overflow that first page */ + if (first_page_offset + sizeof(struct btrfs_free_space_entry) >= PAGE_CACHE_SIZE) { + /* this is really the same as running out of space, where we also return 0 */ + printk(KERN_CRIT "Btrfs: free space cache was too big for the crc page\n"); + ret = 0; + goto out_update; + } + /* We need a checksum per page. */ crc = checksums = kzalloc(sizeof(u32) * num_pages, GFP_NOFS); - if (!crc) { - iput(inode); - return 0; - } + if (!crc) + return -1; pages = kzalloc(sizeof(struct page *) * num_pages, GFP_NOFS); if (!pages) { kfree(crc); - iput(inode); - return 0; + return -1; } - /* Since the first page has all of our checksums and our generation we - * need to calculate the offset into the page that we can start writing - * our entries. - */ - first_page_offset = (sizeof(u32) * num_pages) + sizeof(u64); - /* Get the cluster for this block_group if it exists */ - if (!list_empty(&block_group->cluster_list)) + if (block_group && !list_empty(&block_group->cluster_list)) cluster = list_entry(block_group->cluster_list.next, struct btrfs_free_cluster, block_group_list); @@ -621,7 +662,8 @@ int btrfs_write_out_cache(struct btrfs_root *root, * When searching for pinned extents, we need to start at our start * offset. */ - start = block_group->key.objectid; + if (block_group) + start = block_group->key.objectid; /* Write out the extent entries */ do { @@ -679,8 +721,9 @@ int btrfs_write_out_cache(struct btrfs_root *root, * We want to add any pinned extents to our free space cache * so we don't leak the space */ - while (!next_page && (start < block_group->key.objectid + - block_group->key.offset)) { + while (block_group && !next_page && + (start < block_group->key.objectid + + block_group->key.offset)) { ret = find_first_extent_bit(unpin, start, &start, &end, EXTENT_DIRTY); if (ret) { @@ -798,12 +841,12 @@ int btrfs_write_out_cache(struct btrfs_root *root, filemap_write_and_wait(inode->i_mapping); key.objectid = BTRFS_FREE_SPACE_OBJECTID; - key.offset = block_group->key.objectid; + key.offset = offset; key.type = 0; ret = btrfs_search_slot(trans, root, &key, path, 1, 1); if (ret < 0) { - ret = 0; + ret = -1; clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); @@ -816,13 +859,13 @@ int btrfs_write_out_cache(struct btrfs_root *root, path->slots[0]--; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); if (found_key.objectid != BTRFS_FREE_SPACE_OBJECTID || - found_key.offset != block_group->key.objectid) { - ret = 0; + found_key.offset != offset) { + ret = -1; clear_extent_bit(&BTRFS_I(inode)->io_tree, 0, bytes - 1, EXTENT_DIRTY | EXTENT_DELALLOC | EXTENT_DO_ACCOUNTING, 0, 0, NULL, GFP_NOFS); - btrfs_release_path(root, path); + btrfs_release_path(path); goto out_free; } } @@ -832,49 +875,85 @@ int btrfs_write_out_cache(struct btrfs_root *root, btrfs_set_free_space_bitmaps(leaf, header, bitmaps); btrfs_set_free_space_generation(leaf, header, trans->transid); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = 1; out_free: - if (ret == 0) { + kfree(checksums); + kfree(pages); + +out_update: + if (ret != 1) { invalidate_inode_pages2_range(inode->i_mapping, 0, index); + BTRFS_I(inode)->generation = 0; + } + btrfs_update_inode(trans, root, inode); + return ret; +} + +int btrfs_write_out_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_block_group_cache *block_group, + struct btrfs_path *path) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct inode *inode; + int ret = 0; + + root = root->fs_info->tree_root; + + spin_lock(&block_group->lock); + if (block_group->disk_cache_state < BTRFS_DC_SETUP) { + spin_unlock(&block_group->lock); + return 0; + } + spin_unlock(&block_group->lock); + + inode = lookup_free_space_inode(root, block_group, path); + if (IS_ERR(inode)) + return 0; + + ret = __btrfs_write_out_cache(root, inode, ctl, block_group, trans, + path, block_group->key.objectid); + if (ret < 0) { spin_lock(&block_group->lock); block_group->disk_cache_state = BTRFS_DC_ERROR; spin_unlock(&block_group->lock); - BTRFS_I(inode)->generation = 0; + ret = 0; + + printk(KERN_ERR "btrfs: failed to write free space cace " + "for block group %llu\n", block_group->key.objectid); } - kfree(checksums); - kfree(pages); - btrfs_update_inode(trans, root, inode); + iput(inode); return ret; } -static inline unsigned long offset_to_bit(u64 bitmap_start, u64 sectorsize, +static inline unsigned long offset_to_bit(u64 bitmap_start, u32 unit, u64 offset) { BUG_ON(offset < bitmap_start); offset -= bitmap_start; - return (unsigned long)(div64_u64(offset, sectorsize)); + return (unsigned long)(div_u64(offset, unit)); } -static inline unsigned long bytes_to_bits(u64 bytes, u64 sectorsize) +static inline unsigned long bytes_to_bits(u64 bytes, u32 unit) { - return (unsigned long)(div64_u64(bytes, sectorsize)); + return (unsigned long)(div_u64(bytes, unit)); } -static inline u64 offset_to_bitmap(struct btrfs_block_group_cache *block_group, +static inline u64 offset_to_bitmap(struct btrfs_free_space_ctl *ctl, u64 offset) { u64 bitmap_start; u64 bytes_per_bitmap; - bytes_per_bitmap = BITS_PER_BITMAP * block_group->sectorsize; - bitmap_start = offset - block_group->key.objectid; + bytes_per_bitmap = BITS_PER_BITMAP * ctl->unit; + bitmap_start = offset - ctl->start; bitmap_start = div64_u64(bitmap_start, bytes_per_bitmap); bitmap_start *= bytes_per_bitmap; - bitmap_start += block_group->key.objectid; + bitmap_start += ctl->start; return bitmap_start; } @@ -909,10 +988,16 @@ static int tree_insert_offset(struct rb_root *root, u64 offset, * logically. */ if (bitmap) { - WARN_ON(info->bitmap); + if (info->bitmap) { + WARN_ON_ONCE(1); + return -EEXIST; + } p = &(*p)->rb_right; } else { - WARN_ON(!info->bitmap); + if (!info->bitmap) { + WARN_ON_ONCE(1); + return -EEXIST; + } p = &(*p)->rb_left; } } @@ -932,10 +1017,10 @@ static int tree_insert_offset(struct rb_root *root, u64 offset, * offset. */ static struct btrfs_free_space * -tree_search_offset(struct btrfs_block_group_cache *block_group, +tree_search_offset(struct btrfs_free_space_ctl *ctl, u64 offset, int bitmap_only, int fuzzy) { - struct rb_node *n = block_group->free_space_offset.rb_node; + struct rb_node *n = ctl->free_space_offset.rb_node; struct btrfs_free_space *entry, *prev = NULL; /* find entry that is closest to the 'offset' */ @@ -1031,8 +1116,7 @@ tree_search_offset(struct btrfs_block_group_cache *block_group, break; } } - if (entry->offset + BITS_PER_BITMAP * - block_group->sectorsize > offset) + if (entry->offset + BITS_PER_BITMAP * ctl->unit > offset) return entry; } else if (entry->offset + entry->bytes > offset) return entry; @@ -1043,7 +1127,7 @@ tree_search_offset(struct btrfs_block_group_cache *block_group, while (1) { if (entry->bitmap) { if (entry->offset + BITS_PER_BITMAP * - block_group->sectorsize > offset) + ctl->unit > offset) break; } else { if (entry->offset + entry->bytes > offset) @@ -1059,42 +1143,47 @@ tree_search_offset(struct btrfs_block_group_cache *block_group, } static inline void -__unlink_free_space(struct btrfs_block_group_cache *block_group, +__unlink_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info) { - rb_erase(&info->offset_index, &block_group->free_space_offset); - block_group->free_extents--; + rb_erase(&info->offset_index, &ctl->free_space_offset); + ctl->free_extents--; } -static void unlink_free_space(struct btrfs_block_group_cache *block_group, +static void unlink_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info) { - __unlink_free_space(block_group, info); - block_group->free_space -= info->bytes; + __unlink_free_space(ctl, info); + ctl->free_space -= info->bytes; } -static int link_free_space(struct btrfs_block_group_cache *block_group, +static int link_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info) { int ret = 0; BUG_ON(!info->bitmap && !info->bytes); - ret = tree_insert_offset(&block_group->free_space_offset, info->offset, + ret = tree_insert_offset(&ctl->free_space_offset, info->offset, &info->offset_index, (info->bitmap != NULL)); if (ret) return ret; - block_group->free_space += info->bytes; - block_group->free_extents++; + ctl->free_space += info->bytes; + ctl->free_extents++; return ret; } -static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) +static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) { + struct btrfs_block_group_cache *block_group = ctl->private; u64 max_bytes; u64 bitmap_bytes; u64 extent_bytes; u64 size = block_group->key.offset; + u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize; + int max_bitmaps = div64_u64(size + bytes_per_bg - 1, bytes_per_bg); + + BUG_ON(ctl->total_bitmaps > max_bitmaps); /* * The goal is to keep the total amount of memory used per 1gb of space @@ -1112,10 +1201,10 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) * sure we don't go over our overall goal of MAX_CACHE_BYTES_PER_GIG as * we add more bitmaps. */ - bitmap_bytes = (block_group->total_bitmaps + 1) * PAGE_CACHE_SIZE; + bitmap_bytes = (ctl->total_bitmaps + 1) * PAGE_CACHE_SIZE; if (bitmap_bytes >= max_bytes) { - block_group->extents_thresh = 0; + ctl->extents_thresh = 0; return; } @@ -1126,47 +1215,43 @@ static void recalculate_thresholds(struct btrfs_block_group_cache *block_group) extent_bytes = max_bytes - bitmap_bytes; extent_bytes = min_t(u64, extent_bytes, div64_u64(max_bytes, 2)); - block_group->extents_thresh = + ctl->extents_thresh = div64_u64(extent_bytes, (sizeof(struct btrfs_free_space))); } -static void bitmap_clear_bits(struct btrfs_block_group_cache *block_group, +static void bitmap_clear_bits(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, u64 offset, u64 bytes) { - unsigned long start, end; - unsigned long i; + unsigned long start, count; - start = offset_to_bit(info->offset, block_group->sectorsize, offset); - end = start + bytes_to_bits(bytes, block_group->sectorsize); - BUG_ON(end > BITS_PER_BITMAP); + start = offset_to_bit(info->offset, ctl->unit, offset); + count = bytes_to_bits(bytes, ctl->unit); + BUG_ON(start + count > BITS_PER_BITMAP); - for (i = start; i < end; i++) - clear_bit(i, info->bitmap); + bitmap_clear(info->bitmap, start, count); info->bytes -= bytes; - block_group->free_space -= bytes; + ctl->free_space -= bytes; } -static void bitmap_set_bits(struct btrfs_block_group_cache *block_group, +static void bitmap_set_bits(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, u64 offset, u64 bytes) { - unsigned long start, end; - unsigned long i; + unsigned long start, count; - start = offset_to_bit(info->offset, block_group->sectorsize, offset); - end = start + bytes_to_bits(bytes, block_group->sectorsize); - BUG_ON(end > BITS_PER_BITMAP); + start = offset_to_bit(info->offset, ctl->unit, offset); + count = bytes_to_bits(bytes, ctl->unit); + BUG_ON(start + count > BITS_PER_BITMAP); - for (i = start; i < end; i++) - set_bit(i, info->bitmap); + bitmap_set(info->bitmap, start, count); info->bytes += bytes; - block_group->free_space += bytes; + ctl->free_space += bytes; } -static int search_bitmap(struct btrfs_block_group_cache *block_group, +static int search_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *bitmap_info, u64 *offset, u64 *bytes) { @@ -1174,9 +1259,9 @@ static int search_bitmap(struct btrfs_block_group_cache *block_group, unsigned long bits, i; unsigned long next_zero; - i = offset_to_bit(bitmap_info->offset, block_group->sectorsize, + i = offset_to_bit(bitmap_info->offset, ctl->unit, max_t(u64, *offset, bitmap_info->offset)); - bits = bytes_to_bits(*bytes, block_group->sectorsize); + bits = bytes_to_bits(*bytes, ctl->unit); for (i = find_next_bit(bitmap_info->bitmap, BITS_PER_BITMAP, i); i < BITS_PER_BITMAP; @@ -1191,29 +1276,25 @@ static int search_bitmap(struct btrfs_block_group_cache *block_group, } if (found_bits) { - *offset = (u64)(i * block_group->sectorsize) + - bitmap_info->offset; - *bytes = (u64)(found_bits) * block_group->sectorsize; + *offset = (u64)(i * ctl->unit) + bitmap_info->offset; + *bytes = (u64)(found_bits) * ctl->unit; return 0; } return -1; } -static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache - *block_group, u64 *offset, - u64 *bytes, int debug) +static struct btrfs_free_space * +find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes) { struct btrfs_free_space *entry; struct rb_node *node; int ret; - if (!block_group->free_space_offset.rb_node) + if (!ctl->free_space_offset.rb_node) return NULL; - entry = tree_search_offset(block_group, - offset_to_bitmap(block_group, *offset), - 0, 1); + entry = tree_search_offset(ctl, offset_to_bitmap(ctl, *offset), 0, 1); if (!entry) return NULL; @@ -1223,7 +1304,7 @@ static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache continue; if (entry->bitmap) { - ret = search_bitmap(block_group, entry, offset, bytes); + ret = search_bitmap(ctl, entry, offset, bytes); if (!ret) return entry; continue; @@ -1237,33 +1318,28 @@ static struct btrfs_free_space *find_free_space(struct btrfs_block_group_cache return NULL; } -static void add_new_bitmap(struct btrfs_block_group_cache *block_group, +static void add_new_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, u64 offset) { - u64 bytes_per_bg = BITS_PER_BITMAP * block_group->sectorsize; - int max_bitmaps = (int)div64_u64(block_group->key.offset + - bytes_per_bg - 1, bytes_per_bg); - BUG_ON(block_group->total_bitmaps >= max_bitmaps); - - info->offset = offset_to_bitmap(block_group, offset); + info->offset = offset_to_bitmap(ctl, offset); info->bytes = 0; - link_free_space(block_group, info); - block_group->total_bitmaps++; + link_free_space(ctl, info); + ctl->total_bitmaps++; - recalculate_thresholds(block_group); + ctl->op->recalc_thresholds(ctl); } -static void free_bitmap(struct btrfs_block_group_cache *block_group, +static void free_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *bitmap_info) { - unlink_free_space(block_group, bitmap_info); + unlink_free_space(ctl, bitmap_info); kfree(bitmap_info->bitmap); kmem_cache_free(btrfs_free_space_cachep, bitmap_info); - block_group->total_bitmaps--; - recalculate_thresholds(block_group); + ctl->total_bitmaps--; + ctl->op->recalc_thresholds(ctl); } -static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_group, +static noinline int remove_from_bitmap(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *bitmap_info, u64 *offset, u64 *bytes) { @@ -1272,8 +1348,7 @@ static noinline int remove_from_bitmap(struct btrfs_block_group_cache *block_gro int ret; again: - end = bitmap_info->offset + - (u64)(BITS_PER_BITMAP * block_group->sectorsize) - 1; + end = bitmap_info->offset + (u64)(BITS_PER_BITMAP * ctl->unit) - 1; /* * XXX - this can go away after a few releases. @@ -1288,24 +1363,22 @@ again: search_start = *offset; search_bytes = *bytes; search_bytes = min(search_bytes, end - search_start + 1); - ret = search_bitmap(block_group, bitmap_info, &search_start, - &search_bytes); + ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); BUG_ON(ret < 0 || search_start != *offset); if (*offset > bitmap_info->offset && *offset + *bytes > end) { - bitmap_clear_bits(block_group, bitmap_info, *offset, - end - *offset + 1); + bitmap_clear_bits(ctl, bitmap_info, *offset, end - *offset + 1); *bytes -= end - *offset + 1; *offset = end + 1; } else if (*offset >= bitmap_info->offset && *offset + *bytes <= end) { - bitmap_clear_bits(block_group, bitmap_info, *offset, *bytes); + bitmap_clear_bits(ctl, bitmap_info, *offset, *bytes); *bytes = 0; } if (*bytes) { struct rb_node *next = rb_next(&bitmap_info->offset_index); if (!bitmap_info->bytes) - free_bitmap(block_group, bitmap_info); + free_bitmap(ctl, bitmap_info); /* * no entry after this bitmap, but we still have bytes to @@ -1332,31 +1405,45 @@ again: */ search_start = *offset; search_bytes = *bytes; - ret = search_bitmap(block_group, bitmap_info, &search_start, + ret = search_bitmap(ctl, bitmap_info, &search_start, &search_bytes); if (ret < 0 || search_start != *offset) return -EAGAIN; goto again; } else if (!bitmap_info->bytes) - free_bitmap(block_group, bitmap_info); + free_bitmap(ctl, bitmap_info); return 0; } -static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, - struct btrfs_free_space *info) +static u64 add_bytes_to_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info, u64 offset, + u64 bytes) { - struct btrfs_free_space *bitmap_info; - int added = 0; - u64 bytes, offset, end; - int ret; + u64 bytes_to_set = 0; + u64 end; + + end = info->offset + (u64)(BITS_PER_BITMAP * ctl->unit); + + bytes_to_set = min(end - offset, bytes); + + bitmap_set_bits(ctl, info, offset, bytes_to_set); + + return bytes_to_set; + +} + +static bool use_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + struct btrfs_block_group_cache *block_group = ctl->private; /* * If we are below the extents threshold then we can add this as an * extent, and don't have to deal with the bitmap */ - if (block_group->free_extents < block_group->extents_thresh) { + if (ctl->free_extents < ctl->extents_thresh) { /* * If this block group has some small extents we don't want to * use up all of our free slots in the cache with them, we want @@ -1365,11 +1452,10 @@ static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, * the overhead of a bitmap if we don't have to. */ if (info->bytes <= block_group->sectorsize * 4) { - if (block_group->free_extents * 2 <= - block_group->extents_thresh) - return 0; + if (ctl->free_extents * 2 <= ctl->extents_thresh) + return false; } else { - return 0; + return false; } } @@ -1379,35 +1465,85 @@ static int insert_into_bitmap(struct btrfs_block_group_cache *block_group, */ if (BITS_PER_BITMAP * block_group->sectorsize > block_group->key.offset) - return 0; + return false; + + return true; +} + +static struct btrfs_free_space_op free_space_op = { + .recalc_thresholds = recalculate_thresholds, + .use_bitmap = use_bitmap, +}; + +static int insert_into_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + struct btrfs_free_space *bitmap_info; + struct btrfs_block_group_cache *block_group = NULL; + int added = 0; + u64 bytes, offset, bytes_added; + int ret; bytes = info->bytes; offset = info->offset; + if (!ctl->op->use_bitmap(ctl, info)) + return 0; + + if (ctl->op == &free_space_op) + block_group = ctl->private; again: - bitmap_info = tree_search_offset(block_group, - offset_to_bitmap(block_group, offset), + /* + * Since we link bitmaps right into the cluster we need to see if we + * have a cluster here, and if so and it has our bitmap we need to add + * the free space to that bitmap. + */ + if (block_group && !list_empty(&block_group->cluster_list)) { + struct btrfs_free_cluster *cluster; + struct rb_node *node; + struct btrfs_free_space *entry; + + cluster = list_entry(block_group->cluster_list.next, + struct btrfs_free_cluster, + block_group_list); + spin_lock(&cluster->lock); + node = rb_first(&cluster->root); + if (!node) { + spin_unlock(&cluster->lock); + goto no_cluster_bitmap; + } + + entry = rb_entry(node, struct btrfs_free_space, offset_index); + if (!entry->bitmap) { + spin_unlock(&cluster->lock); + goto no_cluster_bitmap; + } + + if (entry->offset == offset_to_bitmap(ctl, offset)) { + bytes_added = add_bytes_to_bitmap(ctl, entry, + offset, bytes); + bytes -= bytes_added; + offset += bytes_added; + } + spin_unlock(&cluster->lock); + if (!bytes) { + ret = 1; + goto out; + } + } + +no_cluster_bitmap: + bitmap_info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 1, 0); if (!bitmap_info) { BUG_ON(added); goto new_bitmap; } - end = bitmap_info->offset + - (u64)(BITS_PER_BITMAP * block_group->sectorsize); - - if (offset >= bitmap_info->offset && offset + bytes > end) { - bitmap_set_bits(block_group, bitmap_info, offset, - end - offset); - bytes -= end - offset; - offset = end; - added = 0; - } else if (offset >= bitmap_info->offset && offset + bytes <= end) { - bitmap_set_bits(block_group, bitmap_info, offset, bytes); - bytes = 0; - } else { - BUG(); - } + bytes_added = add_bytes_to_bitmap(ctl, bitmap_info, offset, bytes); + bytes -= bytes_added; + offset += bytes_added; + added = 0; if (!bytes) { ret = 1; @@ -1417,19 +1553,19 @@ again: new_bitmap: if (info && info->bitmap) { - add_new_bitmap(block_group, info, offset); + add_new_bitmap(ctl, info, offset); added = 1; info = NULL; goto again; } else { - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); /* no pre-allocated info, allocate a new one */ if (!info) { info = kmem_cache_zalloc(btrfs_free_space_cachep, GFP_NOFS); if (!info) { - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); ret = -ENOMEM; goto out; } @@ -1437,7 +1573,7 @@ new_bitmap: /* allocate the bitmap */ info->bitmap = kzalloc(PAGE_CACHE_SIZE, GFP_NOFS); - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); if (!info->bitmap) { ret = -ENOMEM; goto out; @@ -1455,7 +1591,7 @@ out: return ret; } -bool try_merge_free_space(struct btrfs_block_group_cache *block_group, +static bool try_merge_free_space(struct btrfs_free_space_ctl *ctl, struct btrfs_free_space *info, bool update_stat) { struct btrfs_free_space *left_info; @@ -1469,18 +1605,18 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group, * are adding, if there is remove that struct and add a new one to * cover the entire range */ - right_info = tree_search_offset(block_group, offset + bytes, 0, 0); + right_info = tree_search_offset(ctl, offset + bytes, 0, 0); if (right_info && rb_prev(&right_info->offset_index)) left_info = rb_entry(rb_prev(&right_info->offset_index), struct btrfs_free_space, offset_index); else - left_info = tree_search_offset(block_group, offset - 1, 0, 0); + left_info = tree_search_offset(ctl, offset - 1, 0, 0); if (right_info && !right_info->bitmap) { if (update_stat) - unlink_free_space(block_group, right_info); + unlink_free_space(ctl, right_info); else - __unlink_free_space(block_group, right_info); + __unlink_free_space(ctl, right_info); info->bytes += right_info->bytes; kmem_cache_free(btrfs_free_space_cachep, right_info); merged = true; @@ -1489,9 +1625,9 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group, if (left_info && !left_info->bitmap && left_info->offset + left_info->bytes == offset) { if (update_stat) - unlink_free_space(block_group, left_info); + unlink_free_space(ctl, left_info); else - __unlink_free_space(block_group, left_info); + __unlink_free_space(ctl, left_info); info->offset = left_info->offset; info->bytes += left_info->bytes; kmem_cache_free(btrfs_free_space_cachep, left_info); @@ -1501,8 +1637,8 @@ bool try_merge_free_space(struct btrfs_block_group_cache *block_group, return merged; } -int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, - u64 offset, u64 bytes) +int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl, + u64 offset, u64 bytes) { struct btrfs_free_space *info; int ret = 0; @@ -1514,9 +1650,9 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, info->offset = offset; info->bytes = bytes; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); - if (try_merge_free_space(block_group, info, true)) + if (try_merge_free_space(ctl, info, true)) goto link; /* @@ -1524,7 +1660,7 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, * extent then we know we're going to have to allocate a new extent, so * before we do that see if we need to drop this into a bitmap */ - ret = insert_into_bitmap(block_group, info); + ret = insert_into_bitmap(ctl, info); if (ret < 0) { goto out; } else if (ret) { @@ -1532,11 +1668,11 @@ int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, goto out; } link: - ret = link_free_space(block_group, info); + ret = link_free_space(ctl, info); if (ret) kmem_cache_free(btrfs_free_space_cachep, info); out: - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); if (ret) { printk(KERN_CRIT "btrfs: unable to add free space :%d\n", ret); @@ -1549,21 +1685,21 @@ out: int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, u64 offset, u64 bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *info; struct btrfs_free_space *next_info = NULL; int ret = 0; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); again: - info = tree_search_offset(block_group, offset, 0, 0); + info = tree_search_offset(ctl, offset, 0, 0); if (!info) { /* * oops didn't find an extent that matched the space we wanted * to remove, look for a bitmap instead */ - info = tree_search_offset(block_group, - offset_to_bitmap(block_group, offset), + info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 1, 0); if (!info) { WARN_ON(1); @@ -1578,8 +1714,8 @@ again: offset_index); if (next_info->bitmap) - end = next_info->offset + BITS_PER_BITMAP * - block_group->sectorsize - 1; + end = next_info->offset + + BITS_PER_BITMAP * ctl->unit - 1; else end = next_info->offset + next_info->bytes; @@ -1599,20 +1735,20 @@ again: } if (info->bytes == bytes) { - unlink_free_space(block_group, info); + unlink_free_space(ctl, info); if (info->bitmap) { kfree(info->bitmap); - block_group->total_bitmaps--; + ctl->total_bitmaps--; } kmem_cache_free(btrfs_free_space_cachep, info); goto out_lock; } if (!info->bitmap && info->offset == offset) { - unlink_free_space(block_group, info); + unlink_free_space(ctl, info); info->offset += bytes; info->bytes -= bytes; - link_free_space(block_group, info); + link_free_space(ctl, info); goto out_lock; } @@ -1626,13 +1762,13 @@ again: * first unlink the old info and then * insert it again after the hole we're creating */ - unlink_free_space(block_group, info); + unlink_free_space(ctl, info); if (offset + bytes < info->offset + info->bytes) { u64 old_end = info->offset + info->bytes; info->offset = offset + bytes; info->bytes = old_end - info->offset; - ret = link_free_space(block_group, info); + ret = link_free_space(ctl, info); WARN_ON(ret); if (ret) goto out_lock; @@ -1642,7 +1778,7 @@ again: */ kmem_cache_free(btrfs_free_space_cachep, info); } - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); /* step two, insert a new info struct to cover * anything before the hole @@ -1653,12 +1789,12 @@ again: goto out; } - ret = remove_from_bitmap(block_group, info, &offset, &bytes); + ret = remove_from_bitmap(ctl, info, &offset, &bytes); if (ret == -EAGAIN) goto again; BUG_ON(ret); out_lock: - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); out: return ret; } @@ -1666,11 +1802,12 @@ out: void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, u64 bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *info; struct rb_node *n; int count = 0; - for (n = rb_first(&block_group->free_space_offset); n; n = rb_next(n)) { + for (n = rb_first(&ctl->free_space_offset); n; n = rb_next(n)) { info = rb_entry(n, struct btrfs_free_space, offset_index); if (info->bytes >= bytes) count++; @@ -1685,19 +1822,23 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, "\n", count); } -u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group) +void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group) { - struct btrfs_free_space *info; - struct rb_node *n; - u64 ret = 0; + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; - for (n = rb_first(&block_group->free_space_offset); n; - n = rb_next(n)) { - info = rb_entry(n, struct btrfs_free_space, offset_index); - ret += info->bytes; - } + spin_lock_init(&ctl->tree_lock); + ctl->unit = block_group->sectorsize; + ctl->start = block_group->key.objectid; + ctl->private = block_group; + ctl->op = &free_space_op; - return ret; + /* + * we only want to have 32k of ram per block group for keeping + * track of free space, and if we pass 1/2 of that we want to + * start converting things over to using bitmaps + */ + ctl->extents_thresh = ((1024 * 32) / 2) / + sizeof(struct btrfs_free_space); } /* @@ -1711,6 +1852,7 @@ __btrfs_return_cluster_to_free_space( struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry; struct rb_node *node; @@ -1732,8 +1874,8 @@ __btrfs_return_cluster_to_free_space( bitmap = (entry->bitmap != NULL); if (!bitmap) - try_merge_free_space(block_group, entry, false); - tree_insert_offset(&block_group->free_space_offset, + try_merge_free_space(ctl, entry, false); + tree_insert_offset(&ctl->free_space_offset, entry->offset, &entry->offset_index, bitmap); } cluster->root = RB_ROOT; @@ -1744,14 +1886,41 @@ out: return 0; } -void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) +void __btrfs_remove_free_space_cache_locked(struct btrfs_free_space_ctl *ctl) { struct btrfs_free_space *info; struct rb_node *node; + + while ((node = rb_last(&ctl->free_space_offset)) != NULL) { + info = rb_entry(node, struct btrfs_free_space, offset_index); + if (!info->bitmap) { + unlink_free_space(ctl, info); + kmem_cache_free(btrfs_free_space_cachep, info); + } else { + free_bitmap(ctl, info); + } + if (need_resched()) { + spin_unlock(&ctl->tree_lock); + cond_resched(); + spin_lock(&ctl->tree_lock); + } + } +} + +void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl) +{ + spin_lock(&ctl->tree_lock); + __btrfs_remove_free_space_cache_locked(ctl); + spin_unlock(&ctl->tree_lock); +} + +void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) +{ + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_cluster *cluster; struct list_head *head; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); while ((head = block_group->cluster_list.next) != &block_group->cluster_list) { cluster = list_entry(head, struct btrfs_free_cluster, @@ -1760,60 +1929,46 @@ void btrfs_remove_free_space_cache(struct btrfs_block_group_cache *block_group) WARN_ON(cluster->block_group != block_group); __btrfs_return_cluster_to_free_space(block_group, cluster); if (need_resched()) { - spin_unlock(&block_group->tree_lock); - cond_resched(); - spin_lock(&block_group->tree_lock); - } - } - - while ((node = rb_last(&block_group->free_space_offset)) != NULL) { - info = rb_entry(node, struct btrfs_free_space, offset_index); - if (!info->bitmap) { - unlink_free_space(block_group, info); - kmem_cache_free(btrfs_free_space_cachep, info); - } else { - free_bitmap(block_group, info); - } - - if (need_resched()) { - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); cond_resched(); - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); } } + __btrfs_remove_free_space_cache_locked(ctl); + spin_unlock(&ctl->tree_lock); - spin_unlock(&block_group->tree_lock); } u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, u64 offset, u64 bytes, u64 empty_size) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry = NULL; u64 bytes_search = bytes + empty_size; u64 ret = 0; - spin_lock(&block_group->tree_lock); - entry = find_free_space(block_group, &offset, &bytes_search, 0); + spin_lock(&ctl->tree_lock); + entry = find_free_space(ctl, &offset, &bytes_search); if (!entry) goto out; ret = offset; if (entry->bitmap) { - bitmap_clear_bits(block_group, entry, offset, bytes); + bitmap_clear_bits(ctl, entry, offset, bytes); if (!entry->bytes) - free_bitmap(block_group, entry); + free_bitmap(ctl, entry); } else { - unlink_free_space(block_group, entry); + unlink_free_space(ctl, entry); entry->offset += bytes; entry->bytes -= bytes; if (!entry->bytes) kmem_cache_free(btrfs_free_space_cachep, entry); else - link_free_space(block_group, entry); + link_free_space(ctl, entry); } out: - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); return ret; } @@ -1830,6 +1985,7 @@ int btrfs_return_cluster_to_free_space( struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster) { + struct btrfs_free_space_ctl *ctl; int ret; /* first, get a safe pointer to the block group */ @@ -1848,10 +2004,12 @@ int btrfs_return_cluster_to_free_space( atomic_inc(&block_group->count); spin_unlock(&cluster->lock); + ctl = block_group->free_space_ctl; + /* now return any extents the cluster had on it */ - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); ret = __btrfs_return_cluster_to_free_space(block_group, cluster); - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); /* finally drop our ref */ btrfs_put_block_group(block_group); @@ -1863,6 +2021,7 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, struct btrfs_free_space *entry, u64 bytes, u64 min_start) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; int err; u64 search_start = cluster->window_start; u64 search_bytes = bytes; @@ -1871,13 +2030,12 @@ static u64 btrfs_alloc_from_bitmap(struct btrfs_block_group_cache *block_group, search_start = min_start; search_bytes = bytes; - err = search_bitmap(block_group, entry, &search_start, - &search_bytes); + err = search_bitmap(ctl, entry, &search_start, &search_bytes); if (err) return 0; ret = search_start; - bitmap_clear_bits(block_group, entry, ret, bytes); + bitmap_clear_bits(ctl, entry, ret, bytes); return ret; } @@ -1891,6 +2049,7 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, u64 bytes, u64 min_start) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry = NULL; struct rb_node *node; u64 ret = 0; @@ -1910,8 +2069,6 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, while(1) { if (entry->bytes < bytes || (!entry->bitmap && entry->offset < min_start)) { - struct rb_node *node; - node = rb_next(&entry->offset_index); if (!node) break; @@ -1925,7 +2082,6 @@ u64 btrfs_alloc_from_cluster(struct btrfs_block_group_cache *block_group, cluster, entry, bytes, min_start); if (ret == 0) { - struct rb_node *node; node = rb_next(&entry->offset_index); if (!node) break; @@ -1951,20 +2107,20 @@ out: if (!ret) return 0; - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); - block_group->free_space -= bytes; + ctl->free_space -= bytes; if (entry->bytes == 0) { - block_group->free_extents--; + ctl->free_extents--; if (entry->bitmap) { kfree(entry->bitmap); - block_group->total_bitmaps--; - recalculate_thresholds(block_group); + ctl->total_bitmaps--; + ctl->op->recalc_thresholds(ctl); } kmem_cache_free(btrfs_free_space_cachep, entry); } - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); return ret; } @@ -1974,6 +2130,7 @@ static int btrfs_bitmap_cluster(struct btrfs_block_group_cache *block_group, struct btrfs_free_cluster *cluster, u64 offset, u64 bytes, u64 min_bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; unsigned long next_zero; unsigned long i; unsigned long search_bits; @@ -2028,7 +2185,7 @@ again: cluster->window_start = start * block_group->sectorsize + entry->offset; - rb_erase(&entry->offset_index, &block_group->free_space_offset); + rb_erase(&entry->offset_index, &ctl->free_space_offset); ret = tree_insert_offset(&cluster->root, entry->offset, &entry->offset_index, 1); BUG_ON(ret); @@ -2039,10 +2196,13 @@ again: /* * This searches the block group for just extents to fill the cluster with. */ -static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, - struct btrfs_free_cluster *cluster, - u64 offset, u64 bytes, u64 min_bytes) +static noinline int +setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + struct list_head *bitmaps, u64 offset, u64 bytes, + u64 min_bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *first = NULL; struct btrfs_free_space *entry = NULL; struct btrfs_free_space *prev = NULL; @@ -2053,7 +2213,7 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, u64 max_extent; u64 max_gap = 128 * 1024; - entry = tree_search_offset(block_group, offset, 0, 1); + entry = tree_search_offset(ctl, offset, 0, 1); if (!entry) return -ENOSPC; @@ -2062,6 +2222,8 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, * extent entry. */ while (entry->bitmap) { + if (list_empty(&entry->list)) + list_add_tail(&entry->list, bitmaps); node = rb_next(&entry->offset_index); if (!node) return -ENOSPC; @@ -2081,8 +2243,12 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, return -ENOSPC; entry = rb_entry(node, struct btrfs_free_space, offset_index); - if (entry->bitmap) + if (entry->bitmap) { + if (list_empty(&entry->list)) + list_add_tail(&entry->list, bitmaps); continue; + } + /* * we haven't filled the empty size and the window is * very large. reset and try again @@ -2119,7 +2285,7 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, if (entry->bitmap) continue; - rb_erase(&entry->offset_index, &block_group->free_space_offset); + rb_erase(&entry->offset_index, &ctl->free_space_offset); ret = tree_insert_offset(&cluster->root, entry->offset, &entry->offset_index, 0); BUG_ON(ret); @@ -2134,23 +2300,53 @@ static int setup_cluster_no_bitmap(struct btrfs_block_group_cache *block_group, * This specifically looks for bitmaps that may work in the cluster, we assume * that we have already failed to find extents that will work. */ -static int setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, - struct btrfs_free_cluster *cluster, - u64 offset, u64 bytes, u64 min_bytes) +static noinline int +setup_cluster_bitmap(struct btrfs_block_group_cache *block_group, + struct btrfs_free_cluster *cluster, + struct list_head *bitmaps, u64 offset, u64 bytes, + u64 min_bytes) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry; struct rb_node *node; int ret = -ENOSPC; - if (block_group->total_bitmaps == 0) + if (ctl->total_bitmaps == 0) return -ENOSPC; - entry = tree_search_offset(block_group, - offset_to_bitmap(block_group, offset), - 0, 1); + /* + * First check our cached list of bitmaps and see if there is an entry + * here that will work. + */ + list_for_each_entry(entry, bitmaps, list) { + if (entry->bytes < min_bytes) + continue; + ret = btrfs_bitmap_cluster(block_group, entry, cluster, offset, + bytes, min_bytes); + if (!ret) + return 0; + } + + /* + * If we do have entries on our list and we are here then we didn't find + * anything, so go ahead and get the next entry after the last entry in + * this list and start the search from there. + */ + if (!list_empty(bitmaps)) { + entry = list_entry(bitmaps->prev, struct btrfs_free_space, + list); + node = rb_next(&entry->offset_index); + if (!node) + return -ENOSPC; + entry = rb_entry(node, struct btrfs_free_space, offset_index); + goto search; + } + + entry = tree_search_offset(ctl, offset_to_bitmap(ctl, offset), 0, 1); if (!entry) return -ENOSPC; +search: node = &entry->offset_index; do { entry = rb_entry(node, struct btrfs_free_space, offset_index); @@ -2180,6 +2376,9 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, struct btrfs_free_cluster *cluster, u64 offset, u64 bytes, u64 empty_size) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; + struct list_head bitmaps; + struct btrfs_free_space *entry, *tmp; u64 min_bytes; int ret; @@ -2199,14 +2398,14 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, } else min_bytes = max(bytes, (bytes + empty_size) >> 2); - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); /* * If we know we don't have enough space to make a cluster don't even * bother doing all the work to try and find one. */ - if (block_group->free_space < min_bytes) { - spin_unlock(&block_group->tree_lock); + if (ctl->free_space < min_bytes) { + spin_unlock(&ctl->tree_lock); return -ENOSPC; } @@ -2218,11 +2417,16 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, goto out; } - ret = setup_cluster_no_bitmap(block_group, cluster, offset, bytes, - min_bytes); + INIT_LIST_HEAD(&bitmaps); + ret = setup_cluster_no_bitmap(block_group, cluster, &bitmaps, offset, + bytes, min_bytes); if (ret) - ret = setup_cluster_bitmap(block_group, cluster, offset, - bytes, min_bytes); + ret = setup_cluster_bitmap(block_group, cluster, &bitmaps, + offset, bytes, min_bytes); + + /* Clear our temporary list */ + list_for_each_entry_safe(entry, tmp, &bitmaps, list) + list_del_init(&entry->list); if (!ret) { atomic_inc(&block_group->count); @@ -2232,7 +2436,7 @@ int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, } out: spin_unlock(&cluster->lock); - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); return ret; } @@ -2253,6 +2457,7 @@ void btrfs_init_free_cluster(struct btrfs_free_cluster *cluster) int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, u64 *trimmed, u64 start, u64 end, u64 minlen) { + struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl; struct btrfs_free_space *entry = NULL; struct btrfs_fs_info *fs_info = block_group->fs_info; u64 bytes = 0; @@ -2262,52 +2467,50 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, *trimmed = 0; while (start < end) { - spin_lock(&block_group->tree_lock); + spin_lock(&ctl->tree_lock); - if (block_group->free_space < minlen) { - spin_unlock(&block_group->tree_lock); + if (ctl->free_space < minlen) { + spin_unlock(&ctl->tree_lock); break; } - entry = tree_search_offset(block_group, start, 0, 1); + entry = tree_search_offset(ctl, start, 0, 1); if (!entry) - entry = tree_search_offset(block_group, - offset_to_bitmap(block_group, - start), + entry = tree_search_offset(ctl, + offset_to_bitmap(ctl, start), 1, 1); if (!entry || entry->offset >= end) { - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); break; } if (entry->bitmap) { - ret = search_bitmap(block_group, entry, &start, &bytes); + ret = search_bitmap(ctl, entry, &start, &bytes); if (!ret) { if (start >= end) { - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); break; } bytes = min(bytes, end - start); - bitmap_clear_bits(block_group, entry, - start, bytes); + bitmap_clear_bits(ctl, entry, start, bytes); if (entry->bytes == 0) - free_bitmap(block_group, entry); + free_bitmap(ctl, entry); } else { start = entry->offset + BITS_PER_BITMAP * block_group->sectorsize; - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); ret = 0; continue; } } else { start = entry->offset; bytes = min(entry->bytes, end - start); - unlink_free_space(block_group, entry); + unlink_free_space(ctl, entry); kmem_cache_free(btrfs_free_space_cachep, entry); } - spin_unlock(&block_group->tree_lock); + spin_unlock(&ctl->tree_lock); if (bytes >= minlen) { int update_ret; @@ -2319,8 +2522,7 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, bytes, &actually_trimmed); - btrfs_add_free_space(block_group, - start, bytes); + btrfs_add_free_space(block_group, start, bytes); if (!update_ret) btrfs_update_reserved_bytes(block_group, bytes, 0, 1); @@ -2342,3 +2544,150 @@ int btrfs_trim_block_group(struct btrfs_block_group_cache *block_group, return ret; } + +/* + * Find the left-most item in the cache tree, and then return the + * smallest inode number in the item. + * + * Note: the returned inode number may not be the smallest one in + * the tree, if the left-most item is a bitmap. + */ +u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root) +{ + struct btrfs_free_space_ctl *ctl = fs_root->free_ino_ctl; + struct btrfs_free_space *entry = NULL; + u64 ino = 0; + + spin_lock(&ctl->tree_lock); + + if (RB_EMPTY_ROOT(&ctl->free_space_offset)) + goto out; + + entry = rb_entry(rb_first(&ctl->free_space_offset), + struct btrfs_free_space, offset_index); + + if (!entry->bitmap) { + ino = entry->offset; + + unlink_free_space(ctl, entry); + entry->offset++; + entry->bytes--; + if (!entry->bytes) + kmem_cache_free(btrfs_free_space_cachep, entry); + else + link_free_space(ctl, entry); + } else { + u64 offset = 0; + u64 count = 1; + int ret; + + ret = search_bitmap(ctl, entry, &offset, &count); + BUG_ON(ret); + + ino = offset; + bitmap_clear_bits(ctl, entry, offset, 1); + if (entry->bytes == 0) + free_bitmap(ctl, entry); + } +out: + spin_unlock(&ctl->tree_lock); + + return ino; +} + +struct inode *lookup_free_ino_inode(struct btrfs_root *root, + struct btrfs_path *path) +{ + struct inode *inode = NULL; + + spin_lock(&root->cache_lock); + if (root->cache_inode) + inode = igrab(root->cache_inode); + spin_unlock(&root->cache_lock); + if (inode) + return inode; + + inode = __lookup_free_space_inode(root, path, 0); + if (IS_ERR(inode)) + return inode; + + spin_lock(&root->cache_lock); + if (!btrfs_fs_closing(root->fs_info)) + root->cache_inode = igrab(inode); + spin_unlock(&root->cache_lock); + + return inode; +} + +int create_free_ino_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path) +{ + return __create_free_space_inode(root, trans, path, + BTRFS_FREE_INO_OBJECTID, 0); +} + +int load_free_ino_cache(struct btrfs_fs_info *fs_info, struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_path *path; + struct inode *inode; + int ret = 0; + u64 root_gen = btrfs_root_generation(&root->root_item); + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return 0; + + /* + * If we're unmounting then just return, since this does a search on the + * normal root and not the commit root and we could deadlock. + */ + if (btrfs_fs_closing(fs_info)) + return 0; + + path = btrfs_alloc_path(); + if (!path) + return 0; + + inode = lookup_free_ino_inode(root, path); + if (IS_ERR(inode)) + goto out; + + if (root_gen != BTRFS_I(inode)->generation) + goto out_put; + + ret = __load_free_space_cache(root, inode, ctl, path, 0); + + if (ret < 0) + printk(KERN_ERR "btrfs: failed to load free ino cache for " + "root %llu\n", root->root_key.objectid); +out_put: + iput(inode); +out: + btrfs_free_path(path); + return ret; +} + +int btrfs_write_out_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct inode *inode; + int ret; + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return 0; + + inode = lookup_free_ino_inode(root, path); + if (IS_ERR(inode)) + return 0; + + ret = __btrfs_write_out_cache(root, inode, ctl, NULL, trans, path, 0); + if (ret < 0) + printk(KERN_ERR "btrfs: failed to write free ino cache " + "for root %llu\n", root->root_key.objectid); + + iput(inode); + return ret; +} diff --git a/fs/btrfs/free-space-cache.h b/fs/btrfs/free-space-cache.h index 65c3b935289f..8f2613f779ed 100644 --- a/fs/btrfs/free-space-cache.h +++ b/fs/btrfs/free-space-cache.h @@ -27,6 +27,25 @@ struct btrfs_free_space { struct list_head list; }; +struct btrfs_free_space_ctl { + spinlock_t tree_lock; + struct rb_root free_space_offset; + u64 free_space; + int extents_thresh; + int free_extents; + int total_bitmaps; + int unit; + u64 start; + struct btrfs_free_space_op *op; + void *private; +}; + +struct btrfs_free_space_op { + void (*recalc_thresholds)(struct btrfs_free_space_ctl *ctl); + bool (*use_bitmap)(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info); +}; + struct inode *lookup_free_space_inode(struct btrfs_root *root, struct btrfs_block_group_cache *block_group, struct btrfs_path *path); @@ -45,17 +64,38 @@ int btrfs_write_out_cache(struct btrfs_root *root, struct btrfs_trans_handle *trans, struct btrfs_block_group_cache *block_group, struct btrfs_path *path); -int btrfs_add_free_space(struct btrfs_block_group_cache *block_group, - u64 bytenr, u64 size); + +struct inode *lookup_free_ino_inode(struct btrfs_root *root, + struct btrfs_path *path); +int create_free_ino_inode(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path); +int load_free_ino_cache(struct btrfs_fs_info *fs_info, + struct btrfs_root *root); +int btrfs_write_out_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans, + struct btrfs_path *path); + +void btrfs_init_free_space_ctl(struct btrfs_block_group_cache *block_group); +int __btrfs_add_free_space(struct btrfs_free_space_ctl *ctl, + u64 bytenr, u64 size); +static inline int +btrfs_add_free_space(struct btrfs_block_group_cache *block_group, + u64 bytenr, u64 size) +{ + return __btrfs_add_free_space(block_group->free_space_ctl, + bytenr, size); +} int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group, u64 bytenr, u64 size); +void __btrfs_remove_free_space_cache(struct btrfs_free_space_ctl *ctl); void btrfs_remove_free_space_cache(struct btrfs_block_group_cache - *block_group); + *block_group); u64 btrfs_find_space_for_alloc(struct btrfs_block_group_cache *block_group, u64 offset, u64 bytes, u64 empty_size); +u64 btrfs_find_ino_for_alloc(struct btrfs_root *fs_root); void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group, u64 bytes); -u64 btrfs_block_group_free_space(struct btrfs_block_group_cache *block_group); int btrfs_find_space_cluster(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct btrfs_block_group_cache *block_group, diff --git a/fs/btrfs/inode-item.c b/fs/btrfs/inode-item.c index 64f1150bb48d..baa74f3db691 100644 --- a/fs/btrfs/inode-item.c +++ b/fs/btrfs/inode-item.c @@ -130,7 +130,6 @@ int btrfs_del_inode_ref(struct btrfs_trans_handle *trans, item_size - (ptr + sub_item_len - item_start)); ret = btrfs_truncate_item(trans, root, path, item_size - sub_item_len, 1); - BUG_ON(ret); out: btrfs_free_path(path); return ret; @@ -167,7 +166,6 @@ int btrfs_insert_inode_ref(struct btrfs_trans_handle *trans, old_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]); ret = btrfs_extend_item(trans, root, path, ins_len); - BUG_ON(ret); ref = btrfs_item_ptr(path->nodes[0], path->slots[0], struct btrfs_inode_ref); ref = (struct btrfs_inode_ref *)((unsigned long)ref + old_size); diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c index c05a08f4c411..b4087e0fa871 100644 --- a/fs/btrfs/inode-map.c +++ b/fs/btrfs/inode-map.c @@ -16,11 +16,476 @@ * Boston, MA 021110-1307, USA. */ +#include <linux/delay.h> +#include <linux/kthread.h> +#include <linux/pagemap.h> + #include "ctree.h" #include "disk-io.h" +#include "free-space-cache.h" +#include "inode-map.h" #include "transaction.h" -int btrfs_find_highest_inode(struct btrfs_root *root, u64 *objectid) +static int caching_kthread(void *data) +{ + struct btrfs_root *root = data; + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_key key; + struct btrfs_path *path; + struct extent_buffer *leaf; + u64 last = (u64)-1; + int slot; + int ret; + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return 0; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + /* Since the commit root is read-only, we can safely skip locking. */ + path->skip_locking = 1; + path->search_commit_root = 1; + path->reada = 2; + + key.objectid = BTRFS_FIRST_FREE_OBJECTID; + key.offset = 0; + key.type = BTRFS_INODE_ITEM_KEY; +again: + /* need to make sure the commit_root doesn't disappear */ + mutex_lock(&root->fs_commit_mutex); + + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + + while (1) { + if (btrfs_fs_closing(fs_info)) + goto out; + + leaf = path->nodes[0]; + slot = path->slots[0]; + if (slot >= btrfs_header_nritems(leaf)) { + ret = btrfs_next_leaf(root, path); + if (ret < 0) + goto out; + else if (ret > 0) + break; + + if (need_resched() || + btrfs_transaction_in_commit(fs_info)) { + leaf = path->nodes[0]; + + if (btrfs_header_nritems(leaf) == 0) { + WARN_ON(1); + break; + } + + /* + * Save the key so we can advances forward + * in the next search. + */ + btrfs_item_key_to_cpu(leaf, &key, 0); + btrfs_release_path(path); + root->cache_progress = last; + mutex_unlock(&root->fs_commit_mutex); + schedule_timeout(1); + goto again; + } else + continue; + } + + btrfs_item_key_to_cpu(leaf, &key, slot); + + if (key.type != BTRFS_INODE_ITEM_KEY) + goto next; + + if (key.objectid >= root->highest_objectid) + break; + + if (last != (u64)-1 && last + 1 != key.objectid) { + __btrfs_add_free_space(ctl, last + 1, + key.objectid - last - 1); + wake_up(&root->cache_wait); + } + + last = key.objectid; +next: + path->slots[0]++; + } + + if (last < root->highest_objectid - 1) { + __btrfs_add_free_space(ctl, last + 1, + root->highest_objectid - last - 1); + } + + spin_lock(&root->cache_lock); + root->cached = BTRFS_CACHE_FINISHED; + spin_unlock(&root->cache_lock); + + root->cache_progress = (u64)-1; + btrfs_unpin_free_ino(root); +out: + wake_up(&root->cache_wait); + mutex_unlock(&root->fs_commit_mutex); + + btrfs_free_path(path); + + return ret; +} + +static void start_caching(struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct task_struct *tsk; + int ret; + u64 objectid; + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return; + + spin_lock(&root->cache_lock); + if (root->cached != BTRFS_CACHE_NO) { + spin_unlock(&root->cache_lock); + return; + } + + root->cached = BTRFS_CACHE_STARTED; + spin_unlock(&root->cache_lock); + + ret = load_free_ino_cache(root->fs_info, root); + if (ret == 1) { + spin_lock(&root->cache_lock); + root->cached = BTRFS_CACHE_FINISHED; + spin_unlock(&root->cache_lock); + return; + } + + /* + * It can be quite time-consuming to fill the cache by searching + * through the extent tree, and this can keep ino allocation path + * waiting. Therefore at start we quickly find out the highest + * inode number and we know we can use inode numbers which fall in + * [highest_ino + 1, BTRFS_LAST_FREE_OBJECTID]. + */ + ret = btrfs_find_free_objectid(root, &objectid); + if (!ret && objectid <= BTRFS_LAST_FREE_OBJECTID) { + __btrfs_add_free_space(ctl, objectid, + BTRFS_LAST_FREE_OBJECTID - objectid + 1); + } + + tsk = kthread_run(caching_kthread, root, "btrfs-ino-cache-%llu\n", + root->root_key.objectid); + BUG_ON(IS_ERR(tsk)); +} + +int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid) +{ + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return btrfs_find_free_objectid(root, objectid); + +again: + *objectid = btrfs_find_ino_for_alloc(root); + + if (*objectid != 0) + return 0; + + start_caching(root); + + wait_event(root->cache_wait, + root->cached == BTRFS_CACHE_FINISHED || + root->free_ino_ctl->free_space > 0); + + if (root->cached == BTRFS_CACHE_FINISHED && + root->free_ino_ctl->free_space == 0) + return -ENOSPC; + else + goto again; +} + +void btrfs_return_ino(struct btrfs_root *root, u64 objectid) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_free_space_ctl *pinned = root->free_ino_pinned; + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return; + +again: + if (root->cached == BTRFS_CACHE_FINISHED) { + __btrfs_add_free_space(ctl, objectid, 1); + } else { + /* + * If we are in the process of caching free ino chunks, + * to avoid adding the same inode number to the free_ino + * tree twice due to cross transaction, we'll leave it + * in the pinned tree until a transaction is committed + * or the caching work is done. + */ + + mutex_lock(&root->fs_commit_mutex); + spin_lock(&root->cache_lock); + if (root->cached == BTRFS_CACHE_FINISHED) { + spin_unlock(&root->cache_lock); + mutex_unlock(&root->fs_commit_mutex); + goto again; + } + spin_unlock(&root->cache_lock); + + start_caching(root); + + if (objectid <= root->cache_progress || + objectid > root->highest_objectid) + __btrfs_add_free_space(ctl, objectid, 1); + else + __btrfs_add_free_space(pinned, objectid, 1); + + mutex_unlock(&root->fs_commit_mutex); + } +} + +/* + * When a transaction is committed, we'll move those inode numbers which + * are smaller than root->cache_progress from pinned tree to free_ino tree, + * and others will just be dropped, because the commit root we were + * searching has changed. + * + * Must be called with root->fs_commit_mutex held + */ +void btrfs_unpin_free_ino(struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct rb_root *rbroot = &root->free_ino_pinned->free_space_offset; + struct btrfs_free_space *info; + struct rb_node *n; + u64 count; + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return; + + while (1) { + n = rb_first(rbroot); + if (!n) + break; + + info = rb_entry(n, struct btrfs_free_space, offset_index); + BUG_ON(info->bitmap); + + if (info->offset > root->cache_progress) + goto free; + else if (info->offset + info->bytes > root->cache_progress) + count = root->cache_progress - info->offset + 1; + else + count = info->bytes; + + __btrfs_add_free_space(ctl, info->offset, count); +free: + rb_erase(&info->offset_index, rbroot); + kfree(info); + } +} + +#define INIT_THRESHOLD (((1024 * 32) / 2) / sizeof(struct btrfs_free_space)) +#define INODES_PER_BITMAP (PAGE_CACHE_SIZE * 8) + +/* + * The goal is to keep the memory used by the free_ino tree won't + * exceed the memory if we use bitmaps only. + */ +static void recalculate_thresholds(struct btrfs_free_space_ctl *ctl) +{ + struct btrfs_free_space *info; + struct rb_node *n; + int max_ino; + int max_bitmaps; + + n = rb_last(&ctl->free_space_offset); + if (!n) { + ctl->extents_thresh = INIT_THRESHOLD; + return; + } + info = rb_entry(n, struct btrfs_free_space, offset_index); + + /* + * Find the maximum inode number in the filesystem. Note we + * ignore the fact that this can be a bitmap, because we are + * not doing precise calculation. + */ + max_ino = info->bytes - 1; + + max_bitmaps = ALIGN(max_ino, INODES_PER_BITMAP) / INODES_PER_BITMAP; + if (max_bitmaps <= ctl->total_bitmaps) { + ctl->extents_thresh = 0; + return; + } + + ctl->extents_thresh = (max_bitmaps - ctl->total_bitmaps) * + PAGE_CACHE_SIZE / sizeof(*info); +} + +/* + * We don't fall back to bitmap, if we are below the extents threshold + * or this chunk of inode numbers is a big one. + */ +static bool use_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + if (ctl->free_extents < ctl->extents_thresh || + info->bytes > INODES_PER_BITMAP / 10) + return false; + + return true; +} + +static struct btrfs_free_space_op free_ino_op = { + .recalc_thresholds = recalculate_thresholds, + .use_bitmap = use_bitmap, +}; + +static void pinned_recalc_thresholds(struct btrfs_free_space_ctl *ctl) +{ +} + +static bool pinned_use_bitmap(struct btrfs_free_space_ctl *ctl, + struct btrfs_free_space *info) +{ + /* + * We always use extents for two reasons: + * + * - The pinned tree is only used during the process of caching + * work. + * - Make code simpler. See btrfs_unpin_free_ino(). + */ + return false; +} + +static struct btrfs_free_space_op pinned_free_ino_op = { + .recalc_thresholds = pinned_recalc_thresholds, + .use_bitmap = pinned_use_bitmap, +}; + +void btrfs_init_free_ino_ctl(struct btrfs_root *root) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_free_space_ctl *pinned = root->free_ino_pinned; + + spin_lock_init(&ctl->tree_lock); + ctl->unit = 1; + ctl->start = 0; + ctl->private = NULL; + ctl->op = &free_ino_op; + + /* + * Initially we allow to use 16K of ram to cache chunks of + * inode numbers before we resort to bitmaps. This is somewhat + * arbitrary, but it will be adjusted in runtime. + */ + ctl->extents_thresh = INIT_THRESHOLD; + + spin_lock_init(&pinned->tree_lock); + pinned->unit = 1; + pinned->start = 0; + pinned->private = NULL; + pinned->extents_thresh = 0; + pinned->op = &pinned_free_ino_op; +} + +int btrfs_save_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans) +{ + struct btrfs_free_space_ctl *ctl = root->free_ino_ctl; + struct btrfs_path *path; + struct inode *inode; + u64 alloc_hint = 0; + int ret; + int prealloc; + bool retry = false; + + /* only fs tree and subvol/snap needs ino cache */ + if (root->root_key.objectid != BTRFS_FS_TREE_OBJECTID && + (root->root_key.objectid < BTRFS_FIRST_FREE_OBJECTID || + root->root_key.objectid > BTRFS_LAST_FREE_OBJECTID)) + return 0; + + /* Don't save inode cache if we are deleting this root */ + if (btrfs_root_refs(&root->root_item) == 0 && + root != root->fs_info->tree_root) + return 0; + + if (!btrfs_test_opt(root, INODE_MAP_CACHE)) + return 0; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + +again: + inode = lookup_free_ino_inode(root, path); + if (IS_ERR(inode) && PTR_ERR(inode) != -ENOENT) { + ret = PTR_ERR(inode); + goto out; + } + + if (IS_ERR(inode)) { + BUG_ON(retry); + retry = true; + + ret = create_free_ino_inode(root, trans, path); + if (ret) + goto out; + goto again; + } + + BTRFS_I(inode)->generation = 0; + ret = btrfs_update_inode(trans, root, inode); + WARN_ON(ret); + + if (i_size_read(inode) > 0) { + ret = btrfs_truncate_free_space_cache(root, trans, path, inode); + if (ret) + goto out_put; + } + + spin_lock(&root->cache_lock); + if (root->cached != BTRFS_CACHE_FINISHED) { + ret = -1; + spin_unlock(&root->cache_lock); + goto out_put; + } + spin_unlock(&root->cache_lock); + + spin_lock(&ctl->tree_lock); + prealloc = sizeof(struct btrfs_free_space) * ctl->free_extents; + prealloc = ALIGN(prealloc, PAGE_CACHE_SIZE); + prealloc += ctl->total_bitmaps * PAGE_CACHE_SIZE; + spin_unlock(&ctl->tree_lock); + + /* Just to make sure we have enough space */ + prealloc += 8 * PAGE_CACHE_SIZE; + + ret = btrfs_check_data_free_space(inode, prealloc); + if (ret) + goto out_put; + + ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc, + prealloc, prealloc, &alloc_hint); + if (ret) + goto out_put; + btrfs_free_reserved_data_space(inode, prealloc); + +out_put: + iput(inode); +out: + if (ret == 0) + ret = btrfs_write_out_ino_cache(root, trans, path); + + btrfs_free_path(path); + return ret; +} + +static int btrfs_find_highest_objectid(struct btrfs_root *root, u64 *objectid) { struct btrfs_path *path; int ret; @@ -55,15 +520,14 @@ error: return ret; } -int btrfs_find_free_objectid(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - u64 dirid, u64 *objectid) +int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid) { int ret; mutex_lock(&root->objectid_mutex); if (unlikely(root->highest_objectid < BTRFS_FIRST_FREE_OBJECTID)) { - ret = btrfs_find_highest_inode(root, &root->highest_objectid); + ret = btrfs_find_highest_objectid(root, + &root->highest_objectid); if (ret) goto out; } diff --git a/fs/btrfs/inode-map.h b/fs/btrfs/inode-map.h new file mode 100644 index 000000000000..ddb347bfee23 --- /dev/null +++ b/fs/btrfs/inode-map.h @@ -0,0 +1,13 @@ +#ifndef __BTRFS_INODE_MAP +#define __BTRFS_INODE_MAP + +void btrfs_init_free_ino_ctl(struct btrfs_root *root); +void btrfs_unpin_free_ino(struct btrfs_root *root); +void btrfs_return_ino(struct btrfs_root *root, u64 objectid); +int btrfs_find_free_ino(struct btrfs_root *root, u64 *objectid); +int btrfs_save_ino_cache(struct btrfs_root *root, + struct btrfs_trans_handle *trans); + +int btrfs_find_free_objectid(struct btrfs_root *root, u64 *objectid); + +#endif diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 7cd8ab0ef04d..d340f63d8f07 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -37,6 +37,7 @@ #include <linux/posix_acl.h> #include <linux/falloc.h> #include <linux/slab.h> +#include <linux/ratelimit.h> #include "compat.h" #include "ctree.h" #include "disk-io.h" @@ -51,6 +52,7 @@ #include "compression.h" #include "locking.h" #include "free-space-cache.h" +#include "inode-map.h" struct btrfs_iget_args { u64 ino; @@ -136,9 +138,8 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, return -ENOMEM; path->leave_spinning = 1; - btrfs_set_trans_block_group(trans, inode); - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); key.offset = start; btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY); datasize = btrfs_file_extent_calc_inline_size(cur_size); @@ -340,6 +341,10 @@ static noinline int compress_file_range(struct inode *inode, int will_compress; int compress_type = root->fs_info->compress_type; + /* if this is a small write inside eof, kick off a defragbot */ + if (end <= BTRFS_I(inode)->disk_i_size && (end - start + 1) < 16 * 1024) + btrfs_add_inode_defrag(NULL, inode); + actual_end = min_t(u64, isize, end + 1); again: will_compress = 0; @@ -420,9 +425,8 @@ again: } } if (start == 0) { - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); trans->block_rsv = &root->fs_info->delalloc_block_rsv; /* lets try to make an inline extent */ @@ -617,8 +621,9 @@ retry: async_extent->start + async_extent->ram_size - 1, GFP_NOFS); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); + trans->block_rsv = &root->fs_info->delalloc_block_rsv; ret = btrfs_reserve_extent(trans, root, async_extent->compressed_size, async_extent->compressed_size, @@ -649,7 +654,7 @@ retry: async_extent->start + async_extent->ram_size - 1, 0); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); BUG_ON(!em); em->start = async_extent->start; em->len = async_extent->ram_size; @@ -745,6 +750,15 @@ static u64 get_extent_allocation_hint(struct inode *inode, u64 start, return alloc_hint; } +static inline bool is_free_space_inode(struct btrfs_root *root, + struct inode *inode) +{ + if (root == root->fs_info->tree_root || + BTRFS_I(inode)->location.objectid == BTRFS_FREE_INO_OBJECTID) + return true; + return false; +} + /* * when extent_io.c finds a delayed allocation range in the file, * the call backs end up in this code. The basic idea is to @@ -777,10 +791,9 @@ static noinline int cow_file_range(struct inode *inode, struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; int ret = 0; - BUG_ON(root == root->fs_info->tree_root); - trans = btrfs_join_transaction(root, 1); + BUG_ON(is_free_space_inode(root, inode)); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); trans->block_rsv = &root->fs_info->delalloc_block_rsv; num_bytes = (end - start + blocksize) & ~(blocksize - 1); @@ -788,6 +801,10 @@ static noinline int cow_file_range(struct inode *inode, disk_num_bytes = num_bytes; ret = 0; + /* if this is a small write inside eof, kick off defrag */ + if (end <= BTRFS_I(inode)->disk_i_size && num_bytes < 64 * 1024) + btrfs_add_inode_defrag(trans, inode); + if (start == 0) { /* lets try to make an inline extent */ ret = cow_file_range_inline(trans, root, inode, @@ -826,7 +843,7 @@ static noinline int cow_file_range(struct inode *inode, (u64)-1, &ins, 1); BUG_ON(ret); - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); BUG_ON(!em); em->start = start; em->orig_start = em->start; @@ -1008,7 +1025,7 @@ static noinline int csum_exist_in_range(struct btrfs_root *root, LIST_HEAD(list); ret = btrfs_lookup_csums_range(root->fs_info->csum_root, bytenr, - bytenr + num_bytes - 1, &list); + bytenr + num_bytes - 1, &list, 0); if (ret == 0 && list_empty(&list)) return 0; @@ -1049,29 +1066,33 @@ static noinline int run_delalloc_nocow(struct inode *inode, int type; int nocow; int check_prev = 1; - bool nolock = false; + bool nolock; + u64 ino = btrfs_ino(inode); path = btrfs_alloc_path(); BUG_ON(!path); - if (root == root->fs_info->tree_root) { - nolock = true; - trans = btrfs_join_transaction_nolock(root, 1); - } else { - trans = btrfs_join_transaction(root, 1); - } + + nolock = is_free_space_inode(root, inode); + + if (nolock) + trans = btrfs_join_transaction_nolock(root); + else + trans = btrfs_join_transaction(root); + BUG_ON(IS_ERR(trans)); + trans->block_rsv = &root->fs_info->delalloc_block_rsv; cow_start = (u64)-1; cur_offset = start; while (1) { - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, ino, cur_offset, 0); BUG_ON(ret < 0); if (ret > 0 && path->slots[0] > 0 && check_prev) { leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0] - 1); - if (found_key.objectid == inode->i_ino && + if (found_key.objectid == ino && found_key.type == BTRFS_EXTENT_DATA_KEY) path->slots[0]--; } @@ -1092,7 +1113,7 @@ next_slot: num_bytes = 0; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid > inode->i_ino || + if (found_key.objectid > ino || found_key.type > BTRFS_EXTENT_DATA_KEY || found_key.offset > end) break; @@ -1127,7 +1148,7 @@ next_slot: goto out_check; if (btrfs_extent_readonly(root, disk_bytenr)) goto out_check; - if (btrfs_cross_ref_exist(trans, root, inode->i_ino, + if (btrfs_cross_ref_exist(trans, root, ino, found_key.offset - extent_offset, disk_bytenr)) goto out_check; @@ -1164,7 +1185,7 @@ out_check: goto next_slot; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (cow_start != (u64)-1) { ret = cow_file_range(inode, locked_page, cow_start, found_key.offset - 1, page_started, @@ -1177,7 +1198,7 @@ out_check: struct extent_map *em; struct extent_map_tree *em_tree; em_tree = &BTRFS_I(inode)->extent_tree; - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); BUG_ON(!em); em->start = cur_offset; em->orig_start = em->start; @@ -1222,7 +1243,7 @@ out_check: if (cur_offset > end) break; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (cur_offset <= end && cow_start == (u64)-1) cow_start = cur_offset; @@ -1310,14 +1331,13 @@ static int btrfs_set_bit_hook(struct inode *inode, /* * set_bit and clear bit hooks normally require _irqsave/restore - * but in this case, we are only testeing for the DELALLOC + * but in this case, we are only testing for the DELALLOC * bit, which is only set or cleared with irqs on */ if (!(state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - int do_list = (root->root_key.objectid != - BTRFS_ROOT_TREE_OBJECTID); + bool do_list = !is_free_space_inode(root, inode); if (*bits & EXTENT_FIRST_DELALLOC) *bits &= ~EXTENT_FIRST_DELALLOC; @@ -1344,14 +1364,13 @@ static int btrfs_clear_bit_hook(struct inode *inode, { /* * set_bit and clear bit hooks normally require _irqsave/restore - * but in this case, we are only testeing for the DELALLOC + * but in this case, we are only testing for the DELALLOC * bit, which is only set or cleared with irqs on */ if ((state->state & EXTENT_DELALLOC) && (*bits & EXTENT_DELALLOC)) { struct btrfs_root *root = BTRFS_I(inode)->root; u64 len = state->end + 1 - state->start; - int do_list = (root->root_key.objectid != - BTRFS_ROOT_TREE_OBJECTID); + bool do_list = !is_free_space_inode(root, inode); if (*bits & EXTENT_FIRST_DELALLOC) *bits &= ~EXTENT_FIRST_DELALLOC; @@ -1458,7 +1477,7 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM; - if (root == root->fs_info->tree_root) + if (is_free_space_inode(root, inode)) ret = btrfs_bio_wq_end_io(root->fs_info, bio, 2); else ret = btrfs_bio_wq_end_io(root->fs_info, bio, 0); @@ -1500,8 +1519,6 @@ static noinline int add_pending_csums(struct btrfs_trans_handle *trans, { struct btrfs_ordered_sum *sum; - btrfs_set_trans_block_group(trans, inode); - list_for_each_entry(sum, list, list) { btrfs_csum_file_blocks(trans, BTRFS_I(inode)->root->fs_info->csum_root, sum); @@ -1644,7 +1661,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, &hint, 0); BUG_ON(ret); - ins.objectid = inode->i_ino; + ins.objectid = btrfs_ino(inode); ins.offset = file_pos; ins.type = BTRFS_EXTENT_DATA_KEY; ret = btrfs_insert_empty_item(trans, root, path, &ins, sizeof(*fi)); @@ -1675,7 +1692,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans, ins.type = BTRFS_EXTENT_ITEM_KEY; ret = btrfs_alloc_reserved_file_extent(trans, root, root->root_key.objectid, - inode->i_ino, file_pos, &ins); + btrfs_ino(inode), file_pos, &ins); BUG_ON(ret); btrfs_free_path(path); @@ -1701,7 +1718,7 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) struct extent_state *cached_state = NULL; int compress_type = 0; int ret; - bool nolock = false; + bool nolock; ret = btrfs_dec_test_ordered_pending(inode, &ordered_extent, start, end - start + 1); @@ -1709,18 +1726,17 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) return 0; BUG_ON(!ordered_extent); - nolock = (root == root->fs_info->tree_root); + nolock = is_free_space_inode(root, inode); if (test_bit(BTRFS_ORDERED_NOCOW, &ordered_extent->flags)) { BUG_ON(!list_empty(&ordered_extent->list)); ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); if (!ret) { if (nolock) - trans = btrfs_join_transaction_nolock(root, 1); + trans = btrfs_join_transaction_nolock(root); else - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); trans->block_rsv = &root->fs_info->delalloc_block_rsv; ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); @@ -1733,11 +1749,10 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) 0, &cached_state, GFP_NOFS); if (nolock) - trans = btrfs_join_transaction_nolock(root, 1); + trans = btrfs_join_transaction_nolock(root); else - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); trans->block_rsv = &root->fs_info->delalloc_block_rsv; if (test_bit(BTRFS_ORDERED_COMPRESSED, &ordered_extent->flags)) @@ -1855,7 +1870,7 @@ static int btrfs_io_failed_hook(struct bio *failed_bio, } read_unlock(&em_tree->lock); - if (!em || IS_ERR(em)) { + if (IS_ERR_OR_NULL(em)) { kfree(failrec); return -EIO; } @@ -1971,7 +1986,7 @@ static int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end, } if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM) - return 0; + goto good; if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID && test_range_bit(io_tree, start, end, EXTENT_NODATASUM, 1, NULL)) { @@ -2004,12 +2019,11 @@ good: return 0; zeroit: - if (printk_ratelimit()) { - printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u " - "private %llu\n", page->mapping->host->i_ino, + printk_ratelimited(KERN_INFO "btrfs csum failed ino %llu off %llu csum %u " + "private %llu\n", + (unsigned long long)btrfs_ino(page->mapping->host), (unsigned long long)start, csum, (unsigned long long)private); - } memset(kaddr + offset, 1, end - start + 1); flush_dcache_page(page); kunmap_atomic(kaddr, KM_USER0); @@ -2244,7 +2258,7 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) /* insert an orphan item to track this unlinked/truncated file */ if (insert >= 1) { - ret = btrfs_insert_orphan_item(trans, root, inode->i_ino); + ret = btrfs_insert_orphan_item(trans, root, btrfs_ino(inode)); BUG_ON(ret); } @@ -2281,7 +2295,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode) spin_unlock(&root->orphan_lock); if (trans && delete_item) { - ret = btrfs_del_orphan_item(trans, root, inode->i_ino); + ret = btrfs_del_orphan_item(trans, root, btrfs_ino(inode)); BUG_ON(ret); } @@ -2346,7 +2360,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) break; /* release the path since we're done with it */ - btrfs_release_path(root, path); + btrfs_release_path(path); /* * this is where we are basically btrfs_lookup, without the @@ -2413,7 +2427,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) (u64)-1); if (root->orphan_block_rsv || root->orphan_item_inserted) { - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (!IS_ERR(trans)) btrfs_end_transaction(trans, root); } @@ -2493,12 +2507,17 @@ static void btrfs_read_locked_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_key location; int maybe_acls; - u64 alloc_group_block; u32 rdev; int ret; + bool filled = false; + + ret = btrfs_fill_inode(inode, &rdev); + if (!ret) + filled = true; path = btrfs_alloc_path(); BUG_ON(!path); + path->leave_spinning = 1; memcpy(&location, &BTRFS_I(inode)->location, sizeof(location)); ret = btrfs_lookup_inode(NULL, root, path, &location, 0); @@ -2506,8 +2525,18 @@ static void btrfs_read_locked_inode(struct inode *inode) goto make_bad; leaf = path->nodes[0]; + + if (filled) + goto cache_acl; + inode_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item); + if (!leaf->map_token) + map_private_extent_buffer(leaf, (unsigned long)inode_item, + sizeof(struct btrfs_inode_item), + &leaf->map_token, &leaf->kaddr, + &leaf->map_start, &leaf->map_len, + KM_USER1); inode->i_mode = btrfs_inode_mode(leaf, inode_item); inode->i_nlink = btrfs_inode_nlink(leaf, inode_item); @@ -2536,21 +2565,22 @@ static void btrfs_read_locked_inode(struct inode *inode) BTRFS_I(inode)->index_cnt = (u64)-1; BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item); - - alloc_group_block = btrfs_inode_block_group(leaf, inode_item); - +cache_acl: /* * try to precache a NULL acl entry for files that don't have * any xattrs or acls */ - maybe_acls = acls_after_inode_item(leaf, path->slots[0], inode->i_ino); + maybe_acls = acls_after_inode_item(leaf, path->slots[0], + btrfs_ino(inode)); if (!maybe_acls) cache_no_acl(inode); - BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0, - alloc_group_block, 0); + if (leaf->map_token) { + unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); + leaf->map_token = NULL; + } + btrfs_free_path(path); - inode_item = NULL; switch (inode->i_mode & S_IFMT) { case S_IFREG: @@ -2628,7 +2658,7 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, btrfs_set_inode_transid(leaf, item, trans->transid); btrfs_set_inode_rdev(leaf, item, inode->i_rdev); btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags); - btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group); + btrfs_set_inode_block_group(leaf, item, 0); if (leaf->map_token) { unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); @@ -2647,11 +2677,26 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, struct extent_buffer *leaf; int ret; + /* + * If root is tree root, it means this inode is used to + * store free space information. And these inodes are updated + * when committing the transaction, so they needn't delaye to + * be updated, or deadlock will occured. + */ + if (!is_free_space_inode(root, inode)) { + ret = btrfs_delayed_update_inode(trans, root, inode); + if (!ret) + btrfs_set_inode_last_trans(trans, inode); + return ret; + } + path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) + return -ENOMEM; + path->leave_spinning = 1; - ret = btrfs_lookup_inode(trans, root, path, - &BTRFS_I(inode)->location, 1); + ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location, + 1); if (ret) { if (ret > 0) ret = -ENOENT; @@ -2661,7 +2706,7 @@ noinline int btrfs_update_inode(struct btrfs_trans_handle *trans, btrfs_unlock_up_safe(path, 1); leaf = path->nodes[0]; inode_item = btrfs_item_ptr(leaf, path->slots[0], - struct btrfs_inode_item); + struct btrfs_inode_item); fill_inode_item(trans, leaf, inode_item, inode); btrfs_mark_buffer_dirty(leaf); @@ -2672,7 +2717,6 @@ failed: return ret; } - /* * unlink helper that gets used here in inode.c and in the tree logging * recovery code. It remove a link in a directory with a given name, and @@ -2689,6 +2733,8 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, struct btrfs_dir_item *di; struct btrfs_key key; u64 index; + u64 ino = btrfs_ino(inode); + u64 dir_ino = btrfs_ino(dir); path = btrfs_alloc_path(); if (!path) { @@ -2697,7 +2743,7 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, } path->leave_spinning = 1; - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, root, path, dir_ino, name, name_len, -1); if (IS_ERR(di)) { ret = PTR_ERR(di); @@ -2712,33 +2758,23 @@ static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, ret = btrfs_delete_one_dir_name(trans, root, path, di); if (ret) goto err; - btrfs_release_path(root, path); + btrfs_release_path(path); - ret = btrfs_del_inode_ref(trans, root, name, name_len, - inode->i_ino, - dir->i_ino, &index); + ret = btrfs_del_inode_ref(trans, root, name, name_len, ino, + dir_ino, &index); if (ret) { printk(KERN_INFO "btrfs failed to delete reference to %.*s, " - "inode %lu parent %lu\n", name_len, name, - inode->i_ino, dir->i_ino); + "inode %llu parent %llu\n", name_len, name, + (unsigned long long)ino, (unsigned long long)dir_ino); goto err; } - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, - index, name, name_len, -1); - if (IS_ERR(di)) { - ret = PTR_ERR(di); - goto err; - } - if (!di) { - ret = -ENOENT; + ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); + if (ret) goto err; - } - ret = btrfs_delete_one_dir_name(trans, root, path, di); - btrfs_release_path(root, path); ret = btrfs_del_inode_ref_in_log(trans, root, name, name_len, - inode, dir->i_ino); + inode, dir_ino); BUG_ON(ret != 0 && ret != -ENOENT); ret = btrfs_del_dir_entries_in_log(trans, root, name, name_len, @@ -2816,12 +2852,14 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, int check_link = 1; int err = -ENOSPC; int ret; + u64 ino = btrfs_ino(inode); + u64 dir_ino = btrfs_ino(dir); trans = btrfs_start_transaction(root, 10); if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) return trans; - if (inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) + if (ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) return ERR_PTR(-ENOSPC); /* check if there is someone else holds reference */ @@ -2862,7 +2900,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, } else { check_link = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_lookup_inode(trans, root, path, &BTRFS_I(inode)->location, 0); @@ -2876,11 +2914,11 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, } else { check_link = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret == 0 && S_ISREG(inode->i_mode)) { ret = btrfs_lookup_file_extent(trans, root, path, - inode->i_ino, (u64)-1, 0); + ino, (u64)-1, 0); if (ret < 0) { err = ret; goto out; @@ -2888,7 +2926,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, BUG_ON(ret == 0); if (check_path_shared(root, path)) goto out; - btrfs_release_path(root, path); + btrfs_release_path(path); } if (!check_link) { @@ -2896,7 +2934,7 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, goto out; } - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, root, path, dir_ino, dentry->d_name.name, dentry->d_name.len, 0); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2909,11 +2947,11 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, err = 0; goto out; } - btrfs_release_path(root, path); + btrfs_release_path(path); ref = btrfs_lookup_inode_ref(trans, root, path, dentry->d_name.name, dentry->d_name.len, - inode->i_ino, dir->i_ino, 0); + ino, dir_ino, 0); if (IS_ERR(ref)) { err = PTR_ERR(ref); goto out; @@ -2922,9 +2960,17 @@ static struct btrfs_trans_handle *__unlink_start_trans(struct inode *dir, if (check_path_shared(root, path)) goto out; index = btrfs_inode_ref_index(path->nodes[0], ref); - btrfs_release_path(root, path); + btrfs_release_path(path); - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, index, + /* + * This is a commit root search, if we can lookup inode item and other + * relative items in the commit root, it means the transaction of + * dir/file creation has been committed, and the dir index item that we + * delay to insert has also been inserted into the commit root. So + * we needn't worry about the delayed insertion of the dir index item + * here. + */ + di = btrfs_lookup_dir_index_item(trans, root, path, dir_ino, index, dentry->d_name.name, dentry->d_name.len, 0); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2969,8 +3015,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry) if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, dir); - btrfs_record_unlink_dir(trans, dir, dentry->d_inode, 0); ret = btrfs_unlink_inode(trans, root, dir, dentry->d_inode, @@ -2999,47 +3043,41 @@ int btrfs_unlink_subvol(struct btrfs_trans_handle *trans, struct btrfs_key key; u64 index; int ret; + u64 dir_ino = btrfs_ino(dir); path = btrfs_alloc_path(); if (!path) return -ENOMEM; - di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, root, path, dir_ino, name, name_len, -1); - BUG_ON(!di || IS_ERR(di)); + BUG_ON(IS_ERR_OR_NULL(di)); leaf = path->nodes[0]; btrfs_dir_item_key_to_cpu(leaf, di, &key); WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); ret = btrfs_delete_one_dir_name(trans, root, path, di); BUG_ON(ret); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_del_root_ref(trans, root->fs_info->tree_root, objectid, root->root_key.objectid, - dir->i_ino, &index, name, name_len); + dir_ino, &index, name, name_len); if (ret < 0) { BUG_ON(ret != -ENOENT); - di = btrfs_search_dir_index_item(root, path, dir->i_ino, + di = btrfs_search_dir_index_item(root, path, dir_ino, name, name_len); - BUG_ON(!di || IS_ERR(di)); + BUG_ON(IS_ERR_OR_NULL(di)); leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(root, path); + btrfs_release_path(path); index = key.offset; } + btrfs_release_path(path); - di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino, - index, name, name_len, -1); - BUG_ON(!di || IS_ERR(di)); - - leaf = path->nodes[0]; - btrfs_dir_item_key_to_cpu(leaf, di, &key); - WARN_ON(key.type != BTRFS_ROOT_ITEM_KEY || key.objectid != objectid); - ret = btrfs_delete_one_dir_name(trans, root, path, di); + ret = btrfs_delete_delayed_dir_index(trans, root, dir, index); BUG_ON(ret); - btrfs_release_path(root, path); btrfs_i_size_write(dir, dir->i_size - name_len * 2); dir->i_mtime = dir->i_ctime = CURRENT_TIME; @@ -3059,16 +3097,14 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) unsigned long nr = 0; if (inode->i_size > BTRFS_EMPTY_DIR_SIZE || - inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + btrfs_ino(inode) == BTRFS_FIRST_FREE_OBJECTID) return -ENOTEMPTY; trans = __unlink_start_trans(dir, dentry); if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, dir); - - if (unlikely(inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { + if (unlikely(btrfs_ino(inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { err = btrfs_unlink_subvol(trans, root, dir, BTRFS_I(inode)->location.objectid, dentry->d_name.name, @@ -3093,178 +3129,6 @@ out: return err; } -#if 0 -/* - * when truncating bytes in a file, it is possible to avoid reading - * the leaves that contain only checksum items. This can be the - * majority of the IO required to delete a large file, but it must - * be done carefully. - * - * The keys in the level just above the leaves are checked to make sure - * the lowest key in a given leaf is a csum key, and starts at an offset - * after the new size. - * - * Then the key for the next leaf is checked to make sure it also has - * a checksum item for the same file. If it does, we know our target leaf - * contains only checksum items, and it can be safely freed without reading - * it. - * - * This is just an optimization targeted at large files. It may do - * nothing. It will return 0 unless things went badly. - */ -static noinline int drop_csum_leaves(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct btrfs_path *path, - struct inode *inode, u64 new_size) -{ - struct btrfs_key key; - int ret; - int nritems; - struct btrfs_key found_key; - struct btrfs_key other_key; - struct btrfs_leaf_ref *ref; - u64 leaf_gen; - u64 leaf_start; - - path->lowest_level = 1; - key.objectid = inode->i_ino; - key.type = BTRFS_CSUM_ITEM_KEY; - key.offset = new_size; -again: - ret = btrfs_search_slot(trans, root, &key, path, -1, 1); - if (ret < 0) - goto out; - - if (path->nodes[1] == NULL) { - ret = 0; - goto out; - } - ret = 0; - btrfs_node_key_to_cpu(path->nodes[1], &found_key, path->slots[1]); - nritems = btrfs_header_nritems(path->nodes[1]); - - if (!nritems) - goto out; - - if (path->slots[1] >= nritems) - goto next_node; - - /* did we find a key greater than anything we want to delete? */ - if (found_key.objectid > inode->i_ino || - (found_key.objectid == inode->i_ino && found_key.type > key.type)) - goto out; - - /* we check the next key in the node to make sure the leave contains - * only checksum items. This comparison doesn't work if our - * leaf is the last one in the node - */ - if (path->slots[1] + 1 >= nritems) { -next_node: - /* search forward from the last key in the node, this - * will bring us into the next node in the tree - */ - btrfs_node_key_to_cpu(path->nodes[1], &found_key, nritems - 1); - - /* unlikely, but we inc below, so check to be safe */ - if (found_key.offset == (u64)-1) - goto out; - - /* search_forward needs a path with locks held, do the - * search again for the original key. It is possible - * this will race with a balance and return a path that - * we could modify, but this drop is just an optimization - * and is allowed to miss some leaves. - */ - btrfs_release_path(root, path); - found_key.offset++; - - /* setup a max key for search_forward */ - other_key.offset = (u64)-1; - other_key.type = key.type; - other_key.objectid = key.objectid; - - path->keep_locks = 1; - ret = btrfs_search_forward(root, &found_key, &other_key, - path, 0, 0); - path->keep_locks = 0; - if (ret || found_key.objectid != key.objectid || - found_key.type != key.type) { - ret = 0; - goto out; - } - - key.offset = found_key.offset; - btrfs_release_path(root, path); - cond_resched(); - goto again; - } - - /* we know there's one more slot after us in the tree, - * read that key so we can verify it is also a checksum item - */ - btrfs_node_key_to_cpu(path->nodes[1], &other_key, path->slots[1] + 1); - - if (found_key.objectid < inode->i_ino) - goto next_key; - - if (found_key.type != key.type || found_key.offset < new_size) - goto next_key; - - /* - * if the key for the next leaf isn't a csum key from this objectid, - * we can't be sure there aren't good items inside this leaf. - * Bail out - */ - if (other_key.objectid != inode->i_ino || other_key.type != key.type) - goto out; - - leaf_start = btrfs_node_blockptr(path->nodes[1], path->slots[1]); - leaf_gen = btrfs_node_ptr_generation(path->nodes[1], path->slots[1]); - /* - * it is safe to delete this leaf, it contains only - * csum items from this inode at an offset >= new_size - */ - ret = btrfs_del_leaf(trans, root, path, leaf_start); - BUG_ON(ret); - - if (root->ref_cows && leaf_gen < trans->transid) { - ref = btrfs_alloc_leaf_ref(root, 0); - if (ref) { - ref->root_gen = root->root_key.offset; - ref->bytenr = leaf_start; - ref->owner = 0; - ref->generation = leaf_gen; - ref->nritems = 0; - - btrfs_sort_leaf_ref(ref); - - ret = btrfs_add_leaf_ref(root, ref, 0); - WARN_ON(ret); - btrfs_free_leaf_ref(root, ref); - } else { - WARN_ON(1); - } - } -next_key: - btrfs_release_path(root, path); - - if (other_key.objectid == inode->i_ino && - other_key.type == key.type && other_key.offset > key.offset) { - key.offset = other_key.offset; - cond_resched(); - goto again; - } - ret = 0; -out: - /* fixup any changes we've made to the path */ - path->lowest_level = 0; - path->keep_locks = 0; - btrfs_release_path(root, path); - return ret; -} - -#endif - /* * this can truncate away extent items, csum items and directory items. * It starts at a high offset and removes keys until it can't find @@ -3300,17 +3164,27 @@ int btrfs_truncate_inode_items(struct btrfs_trans_handle *trans, int encoding; int ret; int err = 0; + u64 ino = btrfs_ino(inode); BUG_ON(new_size > 0 && min_type != BTRFS_EXTENT_DATA_KEY); if (root->ref_cows || root == root->fs_info->tree_root) btrfs_drop_extent_cache(inode, new_size & (~mask), (u64)-1, 0); + /* + * This function is also used to drop the items in the log tree before + * we relog the inode, so if root != BTRFS_I(inode)->root, it means + * it is used to drop the loged items. So we shouldn't kill the delayed + * items. + */ + if (min_type == 0 && root == BTRFS_I(inode)->root) + btrfs_kill_delayed_inode_items(inode); + path = btrfs_alloc_path(); BUG_ON(!path); path->reada = -1; - key.objectid = inode->i_ino; + key.objectid = ino; key.offset = (u64)-1; key.type = (u8)-1; @@ -3338,7 +3212,7 @@ search_again: found_type = btrfs_key_type(&found_key); encoding = 0; - if (found_key.objectid != inode->i_ino) + if (found_key.objectid != ino) break; if (found_type < min_type) @@ -3428,7 +3302,6 @@ search_again: btrfs_file_extent_calc_inline_size(size); ret = btrfs_truncate_item(trans, root, path, size, 1); - BUG_ON(ret); } else if (root->ref_cows) { inode_sub_bytes(inode, item_end + 1 - found_key.offset); @@ -3457,7 +3330,7 @@ delete: ret = btrfs_free_extent(trans, root, extent_start, extent_num_bytes, 0, btrfs_header_owner(leaf), - inode->i_ino, extent_offset); + ino, extent_offset); BUG_ON(ret); } @@ -3466,7 +3339,9 @@ delete: if (path->slots[0] == 0 || path->slots[0] != pending_del_slot) { - if (root->ref_cows) { + if (root->ref_cows && + BTRFS_I(inode)->location.objectid != + BTRFS_FREE_INO_OBJECTID) { err = -EAGAIN; goto out; } @@ -3477,7 +3352,7 @@ delete: BUG_ON(ret); pending_del_nr = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); goto search_again; } else { path->slots[0]--; @@ -3635,7 +3510,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) while (1) { em = btrfs_get_extent(inode, NULL, 0, cur_offset, block_end - cur_offset, 0); - BUG_ON(IS_ERR(em) || !em); + BUG_ON(IS_ERR_OR_NULL(em)); last_byte = min(extent_map_end(em), block_end); last_byte = (last_byte + mask) & ~mask; if (!test_bit(EXTENT_FLAG_PREALLOC, &em->flags)) { @@ -3647,7 +3522,6 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) err = PTR_ERR(trans); break; } - btrfs_set_trans_block_group(trans, inode); err = btrfs_drop_extents(trans, inode, cur_offset, cur_offset + hole_size, @@ -3656,7 +3530,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) break; err = btrfs_insert_file_extent(trans, root, - inode->i_ino, cur_offset, 0, + btrfs_ino(inode), cur_offset, 0, 0, hole_size, 0, hole_size, 0, 0, 0); if (err) @@ -3758,7 +3632,7 @@ void btrfs_evict_inode(struct inode *inode) truncate_inode_pages(&inode->i_data, 0); if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 || - root == root->fs_info->tree_root)) + is_free_space_inode(root, inode))) goto no_delete; if (is_bad_inode(inode)) { @@ -3781,9 +3655,8 @@ void btrfs_evict_inode(struct inode *inode) btrfs_i_size_write(inode, 0); while (1) { - trans = btrfs_start_transaction(root, 0); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; ret = btrfs_block_rsv_check(trans, root, @@ -3811,6 +3684,10 @@ void btrfs_evict_inode(struct inode *inode) BUG_ON(ret); } + if (!(root == root->fs_info->tree_root || + root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) + btrfs_return_ino(root, btrfs_ino(inode)); + nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -3836,12 +3713,12 @@ static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry, path = btrfs_alloc_path(); BUG_ON(!path); - di = btrfs_lookup_dir_item(NULL, root, path, dir->i_ino, name, + di = btrfs_lookup_dir_item(NULL, root, path, btrfs_ino(dir), name, namelen, 0); if (IS_ERR(di)) ret = PTR_ERR(di); - if (!di || IS_ERR(di)) + if (IS_ERR_OR_NULL(di)) goto out_err; btrfs_dir_item_key_to_cpu(path->nodes[0], di, location); @@ -3889,7 +3766,7 @@ static int fixup_tree_root_location(struct btrfs_root *root, leaf = path->nodes[0]; ref = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_root_ref); - if (btrfs_root_ref_dirid(leaf, ref) != dir->i_ino || + if (btrfs_root_ref_dirid(leaf, ref) != btrfs_ino(dir) || btrfs_root_ref_name_len(leaf, ref) != dentry->d_name.len) goto out; @@ -3899,7 +3776,7 @@ static int fixup_tree_root_location(struct btrfs_root *root, if (ret) goto out; - btrfs_release_path(root->fs_info->tree_root, path); + btrfs_release_path(path); new_root = btrfs_read_fs_root_no_name(root->fs_info, location); if (IS_ERR(new_root)) { @@ -3928,6 +3805,7 @@ static void inode_tree_add(struct inode *inode) struct btrfs_inode *entry; struct rb_node **p; struct rb_node *parent; + u64 ino = btrfs_ino(inode); again: p = &root->inode_tree.rb_node; parent = NULL; @@ -3940,9 +3818,9 @@ again: parent = *p; entry = rb_entry(parent, struct btrfs_inode, rb_node); - if (inode->i_ino < entry->vfs_inode.i_ino) + if (ino < btrfs_ino(&entry->vfs_inode)) p = &parent->rb_left; - else if (inode->i_ino > entry->vfs_inode.i_ino) + else if (ino > btrfs_ino(&entry->vfs_inode)) p = &parent->rb_right; else { WARN_ON(!(entry->vfs_inode.i_state & @@ -4006,9 +3884,9 @@ again: prev = node; entry = rb_entry(node, struct btrfs_inode, rb_node); - if (objectid < entry->vfs_inode.i_ino) + if (objectid < btrfs_ino(&entry->vfs_inode)) node = node->rb_left; - else if (objectid > entry->vfs_inode.i_ino) + else if (objectid > btrfs_ino(&entry->vfs_inode)) node = node->rb_right; else break; @@ -4016,7 +3894,7 @@ again: if (!node) { while (prev) { entry = rb_entry(prev, struct btrfs_inode, rb_node); - if (objectid <= entry->vfs_inode.i_ino) { + if (objectid <= btrfs_ino(&entry->vfs_inode)) { node = prev; break; } @@ -4025,7 +3903,7 @@ again: } while (node) { entry = rb_entry(node, struct btrfs_inode, rb_node); - objectid = entry->vfs_inode.i_ino + 1; + objectid = btrfs_ino(&entry->vfs_inode) + 1; inode = igrab(&entry->vfs_inode); if (inode) { spin_unlock(&root->inode_lock); @@ -4063,7 +3941,7 @@ static int btrfs_init_locked_inode(struct inode *inode, void *p) static int btrfs_find_actor(struct inode *inode, void *opaque) { struct btrfs_iget_args *args = opaque; - return args->ino == inode->i_ino && + return args->ino == btrfs_ino(inode) && args->root == BTRFS_I(inode)->root; } @@ -4208,7 +4086,7 @@ static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry, return d_splice_alias(inode, dentry); } -static unsigned char btrfs_filetype_table[] = { +unsigned char btrfs_filetype_table[] = { DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK }; @@ -4222,6 +4100,8 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, struct btrfs_key key; struct btrfs_key found_key; struct btrfs_path *path; + struct list_head ins_list; + struct list_head del_list; int ret; struct extent_buffer *leaf; int slot; @@ -4234,6 +4114,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, char tmp_name[32]; char *name_ptr; int name_len; + int is_curr = 0; /* filp->f_pos points to the current index? */ /* FIXME, use a real flag for deciding about the key type */ if (root->fs_info->tree_root == root) @@ -4241,9 +4122,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, /* special case for "." */ if (filp->f_pos == 0) { - over = filldir(dirent, ".", 1, - 1, inode->i_ino, - DT_DIR); + over = filldir(dirent, ".", 1, 1, btrfs_ino(inode), DT_DIR); if (over) return 0; filp->f_pos = 1; @@ -4258,11 +4137,20 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, filp->f_pos = 2; } path = btrfs_alloc_path(); - path->reada = 2; + if (!path) + return -ENOMEM; + + path->reada = 1; + + if (key_type == BTRFS_DIR_INDEX_KEY) { + INIT_LIST_HEAD(&ins_list); + INIT_LIST_HEAD(&del_list); + btrfs_get_delayed_items(inode, &ins_list, &del_list); + } btrfs_set_key_type(&key, key_type); key.offset = filp->f_pos; - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) @@ -4289,8 +4177,13 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, break; if (found_key.offset < filp->f_pos) goto next; + if (key_type == BTRFS_DIR_INDEX_KEY && + btrfs_should_delete_dir_index(&del_list, + found_key.offset)) + goto next; filp->f_pos = found_key.offset; + is_curr = 1; di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item); di_cur = 0; @@ -4345,6 +4238,15 @@ next: path->slots[0]++; } + if (key_type == BTRFS_DIR_INDEX_KEY) { + if (is_curr) + filp->f_pos++; + ret = btrfs_readdir_delayed_dir_index(filp, dirent, filldir, + &ins_list); + if (ret) + goto nopos; + } + /* Reached end of directory/root. Bump pos past the last item. */ if (key_type == BTRFS_DIR_INDEX_KEY) /* @@ -4357,6 +4259,8 @@ next: nopos: ret = 0; err: + if (key_type == BTRFS_DIR_INDEX_KEY) + btrfs_put_delayed_items(&ins_list, &del_list); btrfs_free_path(path); return ret; } @@ -4371,17 +4275,16 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) if (BTRFS_I(inode)->dummy_inode) return 0; - smp_mb(); - nolock = (root->fs_info->closing && root == root->fs_info->tree_root); + if (btrfs_fs_closing(root->fs_info) && is_free_space_inode(root, inode)) + nolock = true; if (wbc->sync_mode == WB_SYNC_ALL) { if (nolock) - trans = btrfs_join_transaction_nolock(root, 1); + trans = btrfs_join_transaction_nolock(root); else - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, inode); if (nolock) ret = btrfs_end_transaction_nolock(trans, root); else @@ -4396,7 +4299,7 @@ int btrfs_write_inode(struct inode *inode, struct writeback_control *wbc) * FIXME, needs more benchmarking...there are no reasons other than performance * to keep or drop this code. */ -void btrfs_dirty_inode(struct inode *inode) +void btrfs_dirty_inode(struct inode *inode, int flags) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; @@ -4405,9 +4308,8 @@ void btrfs_dirty_inode(struct inode *inode) if (BTRFS_I(inode)->dummy_inode) return; - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); - btrfs_set_trans_block_group(trans, inode); ret = btrfs_update_inode(trans, root, inode); if (ret && ret == -ENOSPC) { @@ -4415,25 +4317,24 @@ void btrfs_dirty_inode(struct inode *inode) btrfs_end_transaction(trans, root); trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) { - if (printk_ratelimit()) { - printk(KERN_ERR "btrfs: fail to " - "dirty inode %lu error %ld\n", - inode->i_ino, PTR_ERR(trans)); - } + printk_ratelimited(KERN_ERR "btrfs: fail to " + "dirty inode %llu error %ld\n", + (unsigned long long)btrfs_ino(inode), + PTR_ERR(trans)); return; } - btrfs_set_trans_block_group(trans, inode); ret = btrfs_update_inode(trans, root, inode); if (ret) { - if (printk_ratelimit()) { - printk(KERN_ERR "btrfs: fail to " - "dirty inode %lu error %d\n", - inode->i_ino, ret); - } + printk_ratelimited(KERN_ERR "btrfs: fail to " + "dirty inode %llu error %d\n", + (unsigned long long)btrfs_ino(inode), + ret); } } btrfs_end_transaction(trans, root); + if (BTRFS_I(inode)->delayed_node) + btrfs_balance_delayed_items(root); } /* @@ -4449,7 +4350,7 @@ static int btrfs_set_inode_index_count(struct inode *inode) struct extent_buffer *leaf; int ret; - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); btrfs_set_key_type(&key, BTRFS_DIR_INDEX_KEY); key.offset = (u64)-1; @@ -4481,7 +4382,7 @@ static int btrfs_set_inode_index_count(struct inode *inode) leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); - if (found_key.objectid != inode->i_ino || + if (found_key.objectid != btrfs_ino(inode) || btrfs_key_type(&found_key) != BTRFS_DIR_INDEX_KEY) { BTRFS_I(inode)->index_cnt = 2; goto out; @@ -4502,9 +4403,12 @@ int btrfs_set_inode_index(struct inode *dir, u64 *index) int ret = 0; if (BTRFS_I(dir)->index_cnt == (u64)-1) { - ret = btrfs_set_inode_index_count(dir); - if (ret) - return ret; + ret = btrfs_inode_delayed_dir_index_count(dir); + if (ret) { + ret = btrfs_set_inode_index_count(dir); + if (ret) + return ret; + } } *index = BTRFS_I(dir)->index_cnt; @@ -4517,8 +4421,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *dir, const char *name, int name_len, - u64 ref_objectid, u64 objectid, - u64 alloc_hint, int mode, u64 *index) + u64 ref_objectid, u64 objectid, int mode, + u64 *index) { struct inode *inode; struct btrfs_inode_item *inode_item; @@ -4540,6 +4444,12 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, return ERR_PTR(-ENOMEM); } + /* + * we have to initialize this early, so we can reclaim the inode + * number if we fail afterwards in this function. + */ + inode->i_ino = objectid; + if (dir) { trace_btrfs_inode_request(dir); @@ -4565,8 +4475,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, owner = 0; else owner = 1; - BTRFS_I(inode)->block_group = - btrfs_find_block_group(root, 0, alloc_hint, owner); key[0].objectid = objectid; btrfs_set_key_type(&key[0], BTRFS_INODE_ITEM_KEY); @@ -4585,7 +4493,6 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, goto fail; inode_init_owner(inode, dir, mode); - inode->i_ino = objectid; inode_set_bytes(inode, 0); inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0], @@ -4621,6 +4528,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, inode_tree_add(inode); trace_btrfs_inode_new(inode); + btrfs_set_inode_last_trans(trans, inode); return inode; fail: @@ -4649,29 +4557,29 @@ int btrfs_add_link(struct btrfs_trans_handle *trans, int ret = 0; struct btrfs_key key; struct btrfs_root *root = BTRFS_I(parent_inode)->root; + u64 ino = btrfs_ino(inode); + u64 parent_ino = btrfs_ino(parent_inode); - if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { memcpy(&key, &BTRFS_I(inode)->root->root_key, sizeof(key)); } else { - key.objectid = inode->i_ino; + key.objectid = ino; btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); key.offset = 0; } - if (unlikely(inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(ino == BTRFS_FIRST_FREE_OBJECTID)) { ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, key.objectid, root->root_key.objectid, - parent_inode->i_ino, - index, name, name_len); + parent_ino, index, name, name_len); } else if (add_backref) { - ret = btrfs_insert_inode_ref(trans, root, - name, name_len, inode->i_ino, - parent_inode->i_ino, index); + ret = btrfs_insert_inode_ref(trans, root, name, name_len, ino, + parent_ino, index); } if (ret == 0) { ret = btrfs_insert_dir_item(trans, root, name, name_len, - parent_inode->i_ino, &key, + parent_inode, &key, btrfs_inode_type(inode), index); BUG_ON(ret); @@ -4714,10 +4622,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, if (!new_valid_dev(rdev)) return -EINVAL; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; - /* * 2 for inode item and ref * 2 for dir items @@ -4727,11 +4631,13 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_unlock; inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, - BTRFS_I(dir)->block_group, mode, &index); + dentry->d_name.len, btrfs_ino(dir), objectid, + mode, &index); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto out_unlock; @@ -4743,7 +4649,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, goto out_unlock; } - btrfs_set_trans_block_group(trans, inode); err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; @@ -4752,8 +4657,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, init_special_inode(inode, inode->i_mode, rdev); btrfs_update_inode(trans, root, inode); } - btrfs_update_inode_block_group(trans, inode); - btrfs_update_inode_block_group(trans, dir); out_unlock: nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, root); @@ -4777,9 +4680,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, u64 objectid; u64 index = 0; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; /* * 2 for inode item and ref * 2 for dir items @@ -4789,11 +4689,13 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_unlock; inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, - BTRFS_I(dir)->block_group, mode, &index); + dentry->d_name.len, btrfs_ino(dir), objectid, + mode, &index); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto out_unlock; @@ -4805,7 +4707,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, goto out_unlock; } - btrfs_set_trans_block_group(trans, inode); err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; @@ -4816,8 +4717,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } - btrfs_update_inode_block_group(trans, inode); - btrfs_update_inode_block_group(trans, dir); out_unlock: nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, root); @@ -4864,8 +4763,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, btrfs_inc_nlink(inode); inode->i_ctime = CURRENT_TIME; - - btrfs_set_trans_block_group(trans, dir); ihold(inode); err = btrfs_add_nondir(trans, dir, dentry, inode, 1, index); @@ -4874,7 +4771,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, drop_inode = 1; } else { struct dentry *parent = dget_parent(dentry); - btrfs_update_inode_block_group(trans, dir); err = btrfs_update_inode(trans, root, inode); BUG_ON(err); btrfs_log_new_name(trans, inode, NULL, parent); @@ -4903,10 +4799,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) u64 index = 0; unsigned long nr = 1; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; - /* * 2 items for inode and ref * 2 items for dir items @@ -4915,12 +4807,14 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) trans = btrfs_start_transaction(root, 5); if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, dir); + + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_fail; inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, - BTRFS_I(dir)->block_group, S_IFDIR | mode, - &index); + dentry->d_name.len, btrfs_ino(dir), objectid, + S_IFDIR | mode, &index); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto out_fail; @@ -4934,7 +4828,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) inode->i_op = &btrfs_dir_inode_operations; inode->i_fop = &btrfs_dir_file_operations; - btrfs_set_trans_block_group(trans, inode); btrfs_i_size_write(inode, 0); err = btrfs_update_inode(trans, root, inode); @@ -4948,8 +4841,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) d_instantiate(dentry, inode); drop_on_err = 0; - btrfs_update_inode_block_group(trans, inode); - btrfs_update_inode_block_group(trans, dir); out_fail: nr = trans->blocks_used; @@ -5041,7 +4932,7 @@ struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page, u64 bytenr; u64 extent_start = 0; u64 extent_end = 0; - u64 objectid = inode->i_ino; + u64 objectid = btrfs_ino(inode); u32 found_type; struct btrfs_path *path = NULL; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -5069,7 +4960,7 @@ again: else goto out; } - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { err = -ENOMEM; goto out; @@ -5082,7 +4973,15 @@ again: if (!path) { path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + err = -ENOMEM; + goto out; + } + /* + * Chances are we'll be called again, so go ahead and do + * readahead + */ + path->reada = 1; } ret = btrfs_lookup_file_extent(trans, root, path, @@ -5223,8 +5122,10 @@ again: kunmap(page); free_extent_map(em); em = NULL; - btrfs_release_path(root, path); - trans = btrfs_join_transaction(root, 1); + + btrfs_release_path(path); + trans = btrfs_join_transaction(root); + if (IS_ERR(trans)) return ERR_CAST(trans); goto again; @@ -5249,7 +5150,7 @@ not_found_em: em->block_start = EXTENT_MAP_HOLE; set_bit(EXTENT_FLAG_VACANCY, &em->flags); insert: - btrfs_release_path(root, path); + btrfs_release_path(path); if (em->start > start || extent_map_end(em) <= start) { printk(KERN_ERR "Btrfs: bad extent! em: [%llu %llu] passed " "[%llu %llu]\n", (unsigned long long)em->start, @@ -5382,7 +5283,7 @@ struct extent_map *btrfs_get_extent_fiemap(struct inode *inode, struct page *pag u64 hole_start = start; u64 hole_len = len; - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { err = -ENOMEM; goto out; @@ -5468,10 +5369,13 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, btrfs_drop_extent_cache(inode, start, start + len - 1, 0); } - trans = btrfs_join_transaction(root, 0); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) return ERR_CAST(trans); + if (start <= BTRFS_I(inode)->disk_i_size && len < 64 * 1024) + btrfs_add_inode_defrag(trans, inode); + trans->block_rsv = &root->fs_info->delalloc_block_rsv; alloc_hint = get_extent_allocation_hint(inode, start, len); @@ -5483,7 +5387,7 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, } if (!em) { - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) { em = ERR_PTR(-ENOMEM); goto out; @@ -5549,7 +5453,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans, if (!path) return -ENOMEM; - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode), offset, 0); if (ret < 0) goto out; @@ -5566,7 +5470,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans, ret = 0; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, slot); - if (key.objectid != inode->i_ino || + if (key.objectid != btrfs_ino(inode) || key.type != BTRFS_EXTENT_DATA_KEY) { /* not our file or wrong item type, must cow */ goto out; @@ -5600,7 +5504,7 @@ static noinline int can_nocow_odirect(struct btrfs_trans_handle *trans, * look for other files referencing this extent, if we * find any we must cow */ - if (btrfs_cross_ref_exist(trans, root, inode->i_ino, + if (btrfs_cross_ref_exist(trans, root, btrfs_ino(inode), key.offset - backref_offset, disk_bytenr)) goto out; @@ -5701,7 +5605,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock, * to make sure the current transaction stays open * while we look for nocow cross refs */ - trans = btrfs_join_transaction(root, 0); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) goto must_cow; @@ -5790,9 +5694,10 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) flush_dcache_page(bvec->bv_page); if (csum != *private) { - printk(KERN_ERR "btrfs csum failed ino %lu off" + printk(KERN_ERR "btrfs csum failed ino %llu off" " %llu csum %u private %u\n", - inode->i_ino, (unsigned long long)start, + (unsigned long long)btrfs_ino(inode), + (unsigned long long)start, csum, *private); err = -EIO; } @@ -5839,7 +5744,7 @@ again: BUG_ON(!ordered); - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) { err = -ENOMEM; goto out; @@ -5939,9 +5844,9 @@ static void btrfs_end_dio_bio(struct bio *bio, int err) struct btrfs_dio_private *dip = bio->bi_private; if (err) { - printk(KERN_ERR "btrfs direct IO failed ino %lu rw %lu " + printk(KERN_ERR "btrfs direct IO failed ino %llu rw %lu " "sector %#Lx len %u err no %d\n", - dip->inode->i_ino, bio->bi_rw, + (unsigned long long)btrfs_ino(dip->inode), bio->bi_rw, (unsigned long long)bio->bi_sector, bio->bi_size, err); dip->errors = 1; @@ -6589,6 +6494,7 @@ out: static int btrfs_truncate(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; + struct btrfs_block_rsv *rsv; int ret; int err = 0; struct btrfs_trans_handle *trans; @@ -6602,28 +6508,80 @@ static int btrfs_truncate(struct inode *inode) btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1); btrfs_ordered_update_i_size(inode, inode->i_size, NULL); - trans = btrfs_start_transaction(root, 5); - if (IS_ERR(trans)) - return PTR_ERR(trans); + /* + * Yes ladies and gentelment, this is indeed ugly. The fact is we have + * 3 things going on here + * + * 1) We need to reserve space for our orphan item and the space to + * delete our orphan item. Lord knows we don't want to have a dangling + * orphan item because we didn't reserve space to remove it. + * + * 2) We need to reserve space to update our inode. + * + * 3) We need to have something to cache all the space that is going to + * be free'd up by the truncate operation, but also have some slack + * space reserved in case it uses space during the truncate (thank you + * very much snapshotting). + * + * And we need these to all be seperate. The fact is we can use alot of + * space doing the truncate, and we have no earthly idea how much space + * we will use, so we need the truncate reservation to be seperate so it + * doesn't end up using space reserved for updating the inode or + * removing the orphan item. We also need to be able to stop the + * transaction and start a new one, which means we need to be able to + * update the inode several times, and we have no idea of knowing how + * many times that will be, so we can't just reserve 1 item for the + * entirety of the opration, so that has to be done seperately as well. + * Then there is the orphan item, which does indeed need to be held on + * to for the whole operation, and we need nobody to touch this reserved + * space except the orphan code. + * + * So that leaves us with + * + * 1) root->orphan_block_rsv - for the orphan deletion. + * 2) rsv - for the truncate reservation, which we will steal from the + * transaction reservation. + * 3) fs_info->trans_block_rsv - this will have 1 items worth left for + * updating the inode. + */ + rsv = btrfs_alloc_block_rsv(root); + if (!rsv) + return -ENOMEM; + btrfs_add_durable_block_rsv(root->fs_info, rsv); - btrfs_set_trans_block_group(trans, inode); + trans = btrfs_start_transaction(root, 4); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); + goto out; + } + + /* + * Reserve space for the truncate process. Truncate should be adding + * space, but if there are snapshots it may end up using space. + */ + ret = btrfs_truncate_reserve_metadata(trans, root, rsv); + BUG_ON(ret); ret = btrfs_orphan_add(trans, inode); if (ret) { btrfs_end_transaction(trans, root); - return ret; + goto out; } nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); - /* Now start a transaction for the truncate */ - trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) - return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, inode); - trans->block_rsv = root->orphan_block_rsv; + /* + * Ok so we've already migrated our bytes over for the truncate, so here + * just reserve the one slot we need for updating the inode. + */ + trans = btrfs_start_transaction(root, 1); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); + goto out; + } + trans->block_rsv = rsv; /* * setattr is responsible for setting the ordered_data_close flag, @@ -6647,24 +6605,17 @@ static int btrfs_truncate(struct inode *inode) while (1) { if (!trans) { - trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) - return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, inode); - trans->block_rsv = root->orphan_block_rsv; - } + trans = btrfs_start_transaction(root, 3); + if (IS_ERR(trans)) { + err = PTR_ERR(trans); + goto out; + } - ret = btrfs_block_rsv_check(trans, root, - root->orphan_block_rsv, 0, 5); - if (ret == -EAGAIN) { - ret = btrfs_commit_transaction(trans, root); - if (ret) - return ret; - trans = NULL; - continue; - } else if (ret) { - err = ret; - break; + ret = btrfs_truncate_reserve_metadata(trans, root, + rsv); + BUG_ON(ret); + + trans->block_rsv = rsv; } ret = btrfs_truncate_inode_items(trans, root, inode, @@ -6675,6 +6626,7 @@ static int btrfs_truncate(struct inode *inode) break; } + trans->block_rsv = &root->fs_info->trans_block_rsv; ret = btrfs_update_inode(trans, root, inode); if (ret) { err = ret; @@ -6688,6 +6640,7 @@ static int btrfs_truncate(struct inode *inode) } if (ret == 0 && inode->i_nlink > 0) { + trans->block_rsv = root->orphan_block_rsv; ret = btrfs_orphan_del(trans, inode); if (ret) err = ret; @@ -6699,15 +6652,20 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_orphan_del(NULL, inode); } + trans->block_rsv = &root->fs_info->trans_block_rsv; ret = btrfs_update_inode(trans, root, inode); if (ret && !err) err = ret; nr = trans->blocks_used; ret = btrfs_end_transaction_throttle(trans, root); + btrfs_btree_balance_dirty(root, nr); + +out: + btrfs_free_block_rsv(root, rsv); + if (ret && !err) err = ret; - btrfs_btree_balance_dirty(root, nr); return err; } @@ -6716,15 +6674,14 @@ static int btrfs_truncate(struct inode *inode) * create a new subvolume directory/inode (helper for the ioctl). */ int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, - struct btrfs_root *new_root, - u64 new_dirid, u64 alloc_hint) + struct btrfs_root *new_root, u64 new_dirid) { struct inode *inode; int err; u64 index = 0; inode = btrfs_new_inode(trans, new_root, NULL, "..", 2, new_dirid, - new_dirid, alloc_hint, S_IFDIR | 0700, &index); + new_dirid, S_IFDIR | 0700, &index); if (IS_ERR(inode)) return PTR_ERR(inode); inode->i_op = &btrfs_dir_inode_operations; @@ -6782,12 +6739,15 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) ei->ordered_data_close = 0; ei->orphan_meta_reserved = 0; ei->dummy_inode = 0; + ei->in_defrag = 0; ei->force_compress = BTRFS_COMPRESS_NONE; + ei->delayed_node = NULL; + inode = &ei->vfs_inode; - extent_map_tree_init(&ei->extent_tree, GFP_NOFS); - extent_io_tree_init(&ei->io_tree, &inode->i_data, GFP_NOFS); - extent_io_tree_init(&ei->io_failure_tree, &inode->i_data, GFP_NOFS); + extent_map_tree_init(&ei->extent_tree); + extent_io_tree_init(&ei->io_tree, &inode->i_data); + extent_io_tree_init(&ei->io_failure_tree, &inode->i_data); mutex_init(&ei->log_mutex); btrfs_ordered_inode_tree_init(&ei->ordered_tree); INIT_LIST_HEAD(&ei->i_orphan); @@ -6834,25 +6794,10 @@ void btrfs_destroy_inode(struct inode *inode) spin_unlock(&root->fs_info->ordered_extent_lock); } - if (root == root->fs_info->tree_root) { - struct btrfs_block_group_cache *block_group; - - block_group = btrfs_lookup_block_group(root->fs_info, - BTRFS_I(inode)->block_group); - if (block_group && block_group->inode == inode) { - spin_lock(&block_group->lock); - block_group->inode = NULL; - spin_unlock(&block_group->lock); - btrfs_put_block_group(block_group); - } else if (block_group) { - btrfs_put_block_group(block_group); - } - } - spin_lock(&root->orphan_lock); if (!list_empty(&BTRFS_I(inode)->i_orphan)) { - printk(KERN_INFO "BTRFS: inode %lu still on the orphan list\n", - inode->i_ino); + printk(KERN_INFO "BTRFS: inode %llu still on the orphan list\n", + (unsigned long long)btrfs_ino(inode)); list_del_init(&BTRFS_I(inode)->i_orphan); } spin_unlock(&root->orphan_lock); @@ -6874,6 +6819,7 @@ void btrfs_destroy_inode(struct inode *inode) inode_tree_del(inode); btrfs_drop_extent_cache(inode, 0, (u64)-1, 0); free: + btrfs_remove_delayed_node(inode); call_rcu(&inode->i_rcu, btrfs_i_callback); } @@ -6882,7 +6828,7 @@ int btrfs_drop_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; if (btrfs_root_refs(&root->root_item) == 0 && - root != root->fs_info->tree_root) + !is_free_space_inode(root, inode)) return 1; else return generic_drop_inode(inode); @@ -6991,16 +6937,17 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, u64 index = 0; u64 root_objectid; int ret; + u64 old_ino = btrfs_ino(old_inode); - if (new_dir->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) + if (btrfs_ino(new_dir) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID) return -EPERM; /* we only allow rename subvolume link between subvolumes */ - if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) + if (old_ino != BTRFS_FIRST_FREE_OBJECTID && root != dest) return -EXDEV; - if (old_inode->i_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID || - (new_inode && new_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) + if (old_ino == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID || + (new_inode && btrfs_ino(new_inode) == BTRFS_FIRST_FREE_OBJECTID)) return -ENOTEMPTY; if (S_ISDIR(old_inode->i_mode) && new_inode && @@ -7016,7 +6963,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, filemap_flush(old_inode->i_mapping); /* close the racy window with snapshot create/destroy ioctl */ - if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) down_read(&root->fs_info->subvol_sem); /* * We want to reserve the absolute worst case amount of items. So if @@ -7032,8 +6979,6 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, goto out_notrans; } - btrfs_set_trans_block_group(trans, new_dir); - if (dest != root) btrfs_record_root_in_trans(trans, dest); @@ -7041,15 +6986,15 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (ret) goto out_fail; - if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { /* force full log commit if subvolume involved. */ root->fs_info->last_trans_log_full_commit = trans->transid; } else { ret = btrfs_insert_inode_ref(trans, dest, new_dentry->d_name.name, new_dentry->d_name.len, - old_inode->i_ino, - new_dir->i_ino, index); + old_ino, + btrfs_ino(new_dir), index); if (ret) goto out_fail; /* @@ -7065,10 +7010,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * make sure the inode gets flushed if it is replacing * something. */ - if (new_inode && new_inode->i_size && - old_inode && S_ISREG(old_inode->i_mode)) { + if (new_inode && new_inode->i_size && S_ISREG(old_inode->i_mode)) btrfs_add_ordered_operation(trans, root, old_inode); - } old_dir->i_ctime = old_dir->i_mtime = ctime; new_dir->i_ctime = new_dir->i_mtime = ctime; @@ -7077,7 +7020,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (old_dentry->d_parent != new_dentry->d_parent) btrfs_record_unlink_dir(trans, old_dir, old_inode, 1); - if (unlikely(old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID)) { + if (unlikely(old_ino == BTRFS_FIRST_FREE_OBJECTID)) { root_objectid = BTRFS_I(old_inode)->root->root_key.objectid; ret = btrfs_unlink_subvol(trans, root, old_dir, root_objectid, old_dentry->d_name.name, @@ -7094,7 +7037,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, if (new_inode) { new_inode->i_ctime = CURRENT_TIME; - if (unlikely(new_inode->i_ino == + if (unlikely(btrfs_ino(new_inode) == BTRFS_EMPTY_SUBVOL_DIR_OBJECTID)) { root_objectid = BTRFS_I(new_inode)->location.objectid; ret = btrfs_unlink_subvol(trans, dest, new_dir, @@ -7122,7 +7065,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, new_dentry->d_name.len, 0, index); BUG_ON(ret); - if (old_inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (old_ino != BTRFS_FIRST_FREE_OBJECTID) { struct dentry *parent = dget_parent(new_dentry); btrfs_log_new_name(trans, old_inode, old_dir, parent); dput(parent); @@ -7131,7 +7074,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, out_fail: btrfs_end_transaction_throttle(trans, root); out_notrans: - if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) + if (old_ino == BTRFS_FIRST_FREE_OBJECTID) up_read(&root->fs_info->subvol_sem); return ret; @@ -7185,58 +7128,6 @@ int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput) return 0; } -int btrfs_start_one_delalloc_inode(struct btrfs_root *root, int delay_iput, - int sync) -{ - struct btrfs_inode *binode; - struct inode *inode = NULL; - - spin_lock(&root->fs_info->delalloc_lock); - while (!list_empty(&root->fs_info->delalloc_inodes)) { - binode = list_entry(root->fs_info->delalloc_inodes.next, - struct btrfs_inode, delalloc_inodes); - inode = igrab(&binode->vfs_inode); - if (inode) { - list_move_tail(&binode->delalloc_inodes, - &root->fs_info->delalloc_inodes); - break; - } - - list_del_init(&binode->delalloc_inodes); - cond_resched_lock(&root->fs_info->delalloc_lock); - } - spin_unlock(&root->fs_info->delalloc_lock); - - if (inode) { - if (sync) { - filemap_write_and_wait(inode->i_mapping); - /* - * We have to do this because compression doesn't - * actually set PG_writeback until it submits the pages - * for IO, which happens in an async thread, so we could - * race and not actually wait for any writeback pages - * because they've not been submitted yet. Technically - * this could still be the case for the ordered stuff - * since the async thread may not have started to do its - * work yet. If this becomes the case then we need to - * figure out a way to make sure that in writepage we - * wait for any async pages to be submitted before - * returning so that fdatawait does what its supposed to - * do. - */ - btrfs_wait_ordered_range(inode, 0, (u64)-1); - } else { - filemap_flush(inode->i_mapping); - } - if (delay_iput) - btrfs_add_delayed_iput(inode); - else - iput(inode); - return 1; - } - return 0; -} - static int btrfs_symlink(struct inode *dir, struct dentry *dentry, const char *symname) { @@ -7260,9 +7151,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root)) return -ENAMETOOLONG; - err = btrfs_find_free_objectid(NULL, root, dir->i_ino, &objectid); - if (err) - return err; /* * 2 items for inode item and ref * 2 items for dir items @@ -7272,12 +7160,13 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, dir); + err = btrfs_find_free_ino(root, &objectid); + if (err) + goto out_unlock; inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, - dentry->d_name.len, dir->i_ino, objectid, - BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO, - &index); + dentry->d_name.len, btrfs_ino(dir), objectid, + S_IFLNK|S_IRWXUGO, &index); if (IS_ERR(inode)) { err = PTR_ERR(inode); goto out_unlock; @@ -7289,7 +7178,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, goto out_unlock; } - btrfs_set_trans_block_group(trans, inode); err = btrfs_add_nondir(trans, dir, dentry, inode, 0, index); if (err) drop_inode = 1; @@ -7300,14 +7188,12 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, inode->i_op = &btrfs_file_inode_operations; BTRFS_I(inode)->io_tree.ops = &btrfs_extent_io_ops; } - btrfs_update_inode_block_group(trans, inode); - btrfs_update_inode_block_group(trans, dir); if (drop_inode) goto out_unlock; path = btrfs_alloc_path(); BUG_ON(!path); - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); key.offset = 0; btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY); datasize = btrfs_file_extent_calc_inline_size(name_len); @@ -7315,6 +7201,7 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, datasize); if (err) { drop_inode = 1; + btrfs_free_path(path); goto out_unlock; } leaf = path->nodes[0]; diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 2616f7ed4799..a3c4751e07db 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -50,6 +50,7 @@ #include "print-tree.h" #include "volumes.h" #include "locking.h" +#include "inode-map.h" /* Mask out flags that are inappropriate for the given type of inode. */ static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags) @@ -242,7 +243,7 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg) ip->flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS); } - trans = btrfs_join_transaction(root, 1); + trans = btrfs_join_transaction(root); BUG_ON(IS_ERR(trans)); ret = btrfs_update_inode(trans, root, inode); @@ -281,8 +282,9 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) if (!capable(CAP_SYS_ADMIN)) return -EPERM; - mutex_lock(&fs_info->fs_devices->device_list_mutex); - list_for_each_entry(device, &fs_info->fs_devices->devices, dev_list) { + rcu_read_lock(); + list_for_each_entry_rcu(device, &fs_info->fs_devices->devices, + dev_list) { if (!device->bdev) continue; q = bdev_get_queue(device->bdev); @@ -292,7 +294,7 @@ static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg) minlen); } } - mutex_unlock(&fs_info->fs_devices->device_list_mutex); + rcu_read_unlock(); if (!num_devices) return -EOPNOTSUPP; @@ -329,8 +331,7 @@ static noinline int create_subvol(struct btrfs_root *root, u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID; u64 index = 0; - ret = btrfs_find_free_objectid(NULL, root->fs_info->tree_root, - 0, &objectid); + ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid); if (ret) { dput(parent); return ret; @@ -413,8 +414,7 @@ static noinline int create_subvol(struct btrfs_root *root, btrfs_record_root_in_trans(trans, new_root); - ret = btrfs_create_subvol_root(trans, new_root, new_dirid, - BTRFS_I(dir)->block_group); + ret = btrfs_create_subvol_root(trans, new_root, new_dirid); /* * insert the directory item */ @@ -422,7 +422,7 @@ static noinline int create_subvol(struct btrfs_root *root, BUG_ON(ret); ret = btrfs_insert_dir_item(trans, root, - name, namelen, dir->i_ino, &key, + name, namelen, dir, &key, BTRFS_FT_DIR, index); if (ret) goto fail; @@ -433,7 +433,7 @@ static noinline int create_subvol(struct btrfs_root *root, ret = btrfs_add_root_ref(trans, root->fs_info->tree_root, objectid, root->root_key.objectid, - dir->i_ino, index, name, namelen); + btrfs_ino(dir), index, name, namelen); BUG_ON(ret); @@ -482,8 +482,10 @@ static int create_snapshot(struct btrfs_root *root, struct dentry *dentry, ret = btrfs_snap_reserve_metadata(trans, pending_snapshot); BUG_ON(ret); + spin_lock(&root->fs_info->trans_lock); list_add(&pending_snapshot->list, &trans->transaction->pending_snapshots); + spin_unlock(&root->fs_info->trans_lock); if (async_transid) { *async_transid = trans->transid; ret = btrfs_commit_transaction_async(trans, @@ -655,6 +657,107 @@ out_unlock: return error; } +/* + * When we're defragging a range, we don't want to kick it off again + * if it is really just waiting for delalloc to send it down. + * If we find a nice big extent or delalloc range for the bytes in the + * file you want to defrag, we return 0 to let you know to skip this + * part of the file + */ +static int check_defrag_in_cache(struct inode *inode, u64 offset, int thresh) +{ + struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; + struct extent_map *em = NULL; + struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; + u64 end; + + read_lock(&em_tree->lock); + em = lookup_extent_mapping(em_tree, offset, PAGE_CACHE_SIZE); + read_unlock(&em_tree->lock); + + if (em) { + end = extent_map_end(em); + free_extent_map(em); + if (end - offset > thresh) + return 0; + } + /* if we already have a nice delalloc here, just stop */ + thresh /= 2; + end = count_range_bits(io_tree, &offset, offset + thresh, + thresh, EXTENT_DELALLOC, 1); + if (end >= thresh) + return 0; + return 1; +} + +/* + * helper function to walk through a file and find extents + * newer than a specific transid, and smaller than thresh. + * + * This is used by the defragging code to find new and small + * extents + */ +static int find_new_extents(struct btrfs_root *root, + struct inode *inode, u64 newer_than, + u64 *off, int thresh) +{ + struct btrfs_path *path; + struct btrfs_key min_key; + struct btrfs_key max_key; + struct extent_buffer *leaf; + struct btrfs_file_extent_item *extent; + int type; + int ret; + u64 ino = btrfs_ino(inode); + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + min_key.objectid = ino; + min_key.type = BTRFS_EXTENT_DATA_KEY; + min_key.offset = *off; + + max_key.objectid = ino; + max_key.type = (u8)-1; + max_key.offset = (u64)-1; + + path->keep_locks = 1; + + while(1) { + ret = btrfs_search_forward(root, &min_key, &max_key, + path, 0, newer_than); + if (ret != 0) + goto none; + if (min_key.objectid != ino) + goto none; + if (min_key.type != BTRFS_EXTENT_DATA_KEY) + goto none; + + leaf = path->nodes[0]; + extent = btrfs_item_ptr(leaf, path->slots[0], + struct btrfs_file_extent_item); + + type = btrfs_file_extent_type(leaf, extent); + if (type == BTRFS_FILE_EXTENT_REG && + btrfs_file_extent_num_bytes(leaf, extent) < thresh && + check_defrag_in_cache(inode, min_key.offset, thresh)) { + *off = min_key.offset; + btrfs_free_path(path); + return 0; + } + + if (min_key.offset == (u64)-1) + goto none; + + min_key.offset++; + btrfs_release_path(path); + } +none: + btrfs_free_path(path); + return -ENOENT; +} + static int should_defrag_range(struct inode *inode, u64 start, u64 len, int thresh, u64 *last_len, u64 *skip, u64 *defrag_end) @@ -664,10 +767,6 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; int ret = 1; - - if (thresh == 0) - thresh = 256 * 1024; - /* * make sure that once we start defragging and extent, we keep on * defragging it @@ -726,27 +825,176 @@ static int should_defrag_range(struct inode *inode, u64 start, u64 len, return ret; } -static int btrfs_defrag_file(struct file *file, - struct btrfs_ioctl_defrag_range_args *range) +/* + * it doesn't do much good to defrag one or two pages + * at a time. This pulls in a nice chunk of pages + * to COW and defrag. + * + * It also makes sure the delalloc code has enough + * dirty data to avoid making new small extents as part + * of the defrag + * + * It's a good idea to start RA on this range + * before calling this. + */ +static int cluster_pages_for_defrag(struct inode *inode, + struct page **pages, + unsigned long start_index, + int num_pages) { - struct inode *inode = fdentry(file)->d_inode; - struct btrfs_root *root = BTRFS_I(inode)->root; - struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree; + unsigned long file_end; + u64 isize = i_size_read(inode); + u64 page_start; + u64 page_end; + int ret; + int i; + int i_done; struct btrfs_ordered_extent *ordered; - struct page *page; + struct extent_state *cached_state = NULL; + + if (isize == 0) + return 0; + file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + + ret = btrfs_delalloc_reserve_space(inode, + num_pages << PAGE_CACHE_SHIFT); + if (ret) + return ret; +again: + ret = 0; + i_done = 0; + + /* step one, lock all the pages */ + for (i = 0; i < num_pages; i++) { + struct page *page; + page = grab_cache_page(inode->i_mapping, + start_index + i); + if (!page) + break; + + if (!PageUptodate(page)) { + btrfs_readpage(NULL, page); + lock_page(page); + if (!PageUptodate(page)) { + unlock_page(page); + page_cache_release(page); + ret = -EIO; + break; + } + } + isize = i_size_read(inode); + file_end = (isize - 1) >> PAGE_CACHE_SHIFT; + if (!isize || page->index > file_end || + page->mapping != inode->i_mapping) { + /* whoops, we blew past eof, skip this page */ + unlock_page(page); + page_cache_release(page); + break; + } + pages[i] = page; + i_done++; + } + if (!i_done || ret) + goto out; + + if (!(inode->i_sb->s_flags & MS_ACTIVE)) + goto out; + + /* + * so now we have a nice long stream of locked + * and up to date pages, lets wait on them + */ + for (i = 0; i < i_done; i++) + wait_on_page_writeback(pages[i]); + + page_start = page_offset(pages[0]); + page_end = page_offset(pages[i_done - 1]) + PAGE_CACHE_SIZE; + + lock_extent_bits(&BTRFS_I(inode)->io_tree, + page_start, page_end - 1, 0, &cached_state, + GFP_NOFS); + ordered = btrfs_lookup_first_ordered_extent(inode, page_end - 1); + if (ordered && + ordered->file_offset + ordered->len > page_start && + ordered->file_offset < page_end) { + btrfs_put_ordered_extent(ordered); + unlock_extent_cached(&BTRFS_I(inode)->io_tree, + page_start, page_end - 1, + &cached_state, GFP_NOFS); + for (i = 0; i < i_done; i++) { + unlock_page(pages[i]); + page_cache_release(pages[i]); + } + btrfs_wait_ordered_range(inode, page_start, + page_end - page_start); + goto again; + } + if (ordered) + btrfs_put_ordered_extent(ordered); + + clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start, + page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC | + EXTENT_DO_ACCOUNTING, 0, 0, &cached_state, + GFP_NOFS); + + if (i_done != num_pages) { + atomic_inc(&BTRFS_I(inode)->outstanding_extents); + btrfs_delalloc_release_space(inode, + (num_pages - i_done) << PAGE_CACHE_SHIFT); + } + + + btrfs_set_extent_delalloc(inode, page_start, page_end - 1, + &cached_state); + + unlock_extent_cached(&BTRFS_I(inode)->io_tree, + page_start, page_end - 1, &cached_state, + GFP_NOFS); + + for (i = 0; i < i_done; i++) { + clear_page_dirty_for_io(pages[i]); + ClearPageChecked(pages[i]); + set_page_extent_mapped(pages[i]); + set_page_dirty(pages[i]); + unlock_page(pages[i]); + page_cache_release(pages[i]); + } + return i_done; +out: + for (i = 0; i < i_done; i++) { + unlock_page(pages[i]); + page_cache_release(pages[i]); + } + btrfs_delalloc_release_space(inode, num_pages << PAGE_CACHE_SHIFT); + return ret; + +} + +int btrfs_defrag_file(struct inode *inode, struct file *file, + struct btrfs_ioctl_defrag_range_args *range, + u64 newer_than, unsigned long max_to_defrag) +{ + struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_super_block *disk_super; + struct file_ra_state *ra = NULL; unsigned long last_index; - unsigned long ra_pages = root->fs_info->bdi.ra_pages; - unsigned long total_read = 0; u64 features; - u64 page_start; - u64 page_end; u64 last_len = 0; u64 skip = 0; u64 defrag_end = 0; + u64 newer_off = range->start; + int newer_left = 0; unsigned long i; int ret; + int defrag_count = 0; int compress_type = BTRFS_COMPRESS_ZLIB; + int extent_thresh = range->extent_thresh; + int newer_cluster = (256 * 1024) >> PAGE_CACHE_SHIFT; + u64 new_align = ~((u64)128 * 1024 - 1); + struct page **pages = NULL; + + if (extent_thresh == 0) + extent_thresh = 256 * 1024; if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) { if (range->compress_type > BTRFS_COMPRESS_TYPES) @@ -758,6 +1006,27 @@ static int btrfs_defrag_file(struct file *file, if (inode->i_size == 0) return 0; + /* + * if we were not given a file, allocate a readahead + * context + */ + if (!file) { + ra = kzalloc(sizeof(*ra), GFP_NOFS); + if (!ra) + return -ENOMEM; + file_ra_state_init(ra, inode->i_mapping); + } else { + ra = &file->f_ra; + } + + pages = kmalloc(sizeof(struct page *) * newer_cluster, + GFP_NOFS); + if (!pages) { + ret = -ENOMEM; + goto out_ra; + } + + /* find the last page to defrag */ if (range->start + range->len > range->start) { last_index = min_t(u64, inode->i_size - 1, range->start + range->len - 1) >> PAGE_CACHE_SHIFT; @@ -765,11 +1034,37 @@ static int btrfs_defrag_file(struct file *file, last_index = (inode->i_size - 1) >> PAGE_CACHE_SHIFT; } - i = range->start >> PAGE_CACHE_SHIFT; - while (i <= last_index) { - if (!should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT, + if (newer_than) { + ret = find_new_extents(root, inode, newer_than, + &newer_off, 64 * 1024); + if (!ret) { + range->start = newer_off; + /* + * we always align our defrag to help keep + * the extents in the file evenly spaced + */ + i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; + newer_left = newer_cluster; + } else + goto out_ra; + } else { + i = range->start >> PAGE_CACHE_SHIFT; + } + if (!max_to_defrag) + max_to_defrag = last_index - 1; + + while (i <= last_index && defrag_count < max_to_defrag) { + /* + * make sure we stop running if someone unmounts + * the FS + */ + if (!(inode->i_sb->s_flags & MS_ACTIVE)) + break; + + if (!newer_than && + !should_defrag_range(inode, (u64)i << PAGE_CACHE_SHIFT, PAGE_CACHE_SIZE, - range->extent_thresh, + extent_thresh, &last_len, &skip, &defrag_end)) { unsigned long next; @@ -781,92 +1076,39 @@ static int btrfs_defrag_file(struct file *file, i = max(i + 1, next); continue; } - - if (total_read % ra_pages == 0) { - btrfs_force_ra(inode->i_mapping, &file->f_ra, file, i, - min(last_index, i + ra_pages - 1)); - } - total_read++; - mutex_lock(&inode->i_mutex); if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) BTRFS_I(inode)->force_compress = compress_type; - ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE); - if (ret) - goto err_unlock; -again: - if (inode->i_size == 0 || - i > ((inode->i_size - 1) >> PAGE_CACHE_SHIFT)) { - ret = 0; - goto err_reservations; - } - - page = grab_cache_page(inode->i_mapping, i); - if (!page) { - ret = -ENOMEM; - goto err_reservations; - } - - if (!PageUptodate(page)) { - btrfs_readpage(NULL, page); - lock_page(page); - if (!PageUptodate(page)) { - unlock_page(page); - page_cache_release(page); - ret = -EIO; - goto err_reservations; - } - } - - if (page->mapping != inode->i_mapping) { - unlock_page(page); - page_cache_release(page); - goto again; - } + btrfs_force_ra(inode->i_mapping, ra, file, i, newer_cluster); - wait_on_page_writeback(page); + ret = cluster_pages_for_defrag(inode, pages, i, newer_cluster); + if (ret < 0) + goto out_ra; - if (PageDirty(page)) { - btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); - goto loop_unlock; - } + defrag_count += ret; + balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret); + i += ret; - page_start = (u64)page->index << PAGE_CACHE_SHIFT; - page_end = page_start + PAGE_CACHE_SIZE - 1; - lock_extent(io_tree, page_start, page_end, GFP_NOFS); + if (newer_than) { + if (newer_off == (u64)-1) + break; - ordered = btrfs_lookup_ordered_extent(inode, page_start); - if (ordered) { - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - unlock_page(page); - page_cache_release(page); - btrfs_start_ordered_extent(inode, ordered, 1); - btrfs_put_ordered_extent(ordered); - goto again; + newer_off = max(newer_off + 1, + (u64)i << PAGE_CACHE_SHIFT); + + ret = find_new_extents(root, inode, + newer_than, &newer_off, + 64 * 1024); + if (!ret) { + range->start = newer_off; + i = (newer_off & new_align) >> PAGE_CACHE_SHIFT; + newer_left = newer_cluster; + } else { + break; + } + } else { + i++; } - set_page_extent_mapped(page); - - /* - * this makes sure page_mkwrite is called on the - * page if it is dirtied again later - */ - clear_page_dirty_for_io(page); - clear_extent_bits(&BTRFS_I(inode)->io_tree, page_start, - page_end, EXTENT_DIRTY | EXTENT_DELALLOC | - EXTENT_DO_ACCOUNTING, GFP_NOFS); - - btrfs_set_extent_delalloc(inode, page_start, page_end, NULL); - ClearPageChecked(page); - set_page_dirty(page); - unlock_extent(io_tree, page_start, page_end, GFP_NOFS); - -loop_unlock: - unlock_page(page); - page_cache_release(page); - mutex_unlock(&inode->i_mutex); - - balance_dirty_pages_ratelimited_nr(inode->i_mapping, 1); - i++; } if ((range->flags & BTRFS_DEFRAG_RANGE_START_IO)) @@ -898,12 +1140,14 @@ loop_unlock: btrfs_set_super_incompat_flags(disk_super, features); } - return 0; + if (!file) + kfree(ra); + return defrag_count; -err_reservations: - btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE); -err_unlock: - mutex_unlock(&inode->i_mutex); +out_ra: + if (!file) + kfree(ra); + kfree(pages); return ret; } @@ -1129,7 +1373,7 @@ static noinline int btrfs_ioctl_subvol_getflags(struct file *file, int ret = 0; u64 flags = 0; - if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) return -EINVAL; down_read(&root->fs_info->subvol_sem); @@ -1156,7 +1400,7 @@ static noinline int btrfs_ioctl_subvol_setflags(struct file *file, if (root->fs_info->sb->s_flags & MS_RDONLY) return -EROFS; - if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) return -EINVAL; if (copy_from_user(&flags, arg, sizeof(flags))) @@ -1279,7 +1523,6 @@ static noinline int copy_to_sk(struct btrfs_root *root, int nritems; int i; int slot; - int found = 0; int ret = 0; leaf = path->nodes[0]; @@ -1326,7 +1569,7 @@ static noinline int copy_to_sk(struct btrfs_root *root, item_off, item_len); *sk_offset += item_len; } - found++; + (*num_found)++; if (*num_found >= sk->nr_items) break; @@ -1345,7 +1588,6 @@ advance_key: } else ret = 1; overflow: - *num_found += found; return ret; } @@ -1402,7 +1644,7 @@ static noinline int search_ioctl(struct inode *inode, } ret = copy_to_sk(root, path, &key, sk, args->buf, &sk_offset, &num_found); - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret || num_found >= sk->nr_items) break; @@ -1509,7 +1751,7 @@ static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info, if (key.offset == BTRFS_FIRST_FREE_OBJECTID) break; - btrfs_release_path(root, path); + btrfs_release_path(path); key.objectid = key.offset; key.offset = (u64)-1; dirid = key.objectid; @@ -1639,7 +1881,7 @@ static noinline int btrfs_ioctl_snap_destroy(struct file *file, goto out_dput; } - if (inode->i_ino != BTRFS_FIRST_FREE_OBJECTID) { + if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) { err = -EINVAL; goto out_dput; } @@ -1757,7 +1999,10 @@ static int btrfs_ioctl_defrag(struct file *file, void __user *argp) /* the rest are all set to zero by kzalloc */ range->len = (u64)-1; } - ret = btrfs_defrag_file(file, range); + ret = btrfs_defrag_file(fdentry(file)->d_inode, file, + range, 0, 0); + if (ret > 0) + ret = 0; kfree(range); break; default: @@ -1809,6 +2054,80 @@ static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg) return ret; } +static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_fs_info_args *fi_args; + struct btrfs_device *device; + struct btrfs_device *next; + struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; + int ret = 0; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + fi_args = kzalloc(sizeof(*fi_args), GFP_KERNEL); + if (!fi_args) + return -ENOMEM; + + fi_args->num_devices = fs_devices->num_devices; + memcpy(&fi_args->fsid, root->fs_info->fsid, sizeof(fi_args->fsid)); + + mutex_lock(&fs_devices->device_list_mutex); + list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) { + if (device->devid > fi_args->max_id) + fi_args->max_id = device->devid; + } + mutex_unlock(&fs_devices->device_list_mutex); + + if (copy_to_user(arg, fi_args, sizeof(*fi_args))) + ret = -EFAULT; + + kfree(fi_args); + return ret; +} + +static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg) +{ + struct btrfs_ioctl_dev_info_args *di_args; + struct btrfs_device *dev; + struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices; + int ret = 0; + char *s_uuid = NULL; + char empty_uuid[BTRFS_UUID_SIZE] = {0}; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + di_args = memdup_user(arg, sizeof(*di_args)); + if (IS_ERR(di_args)) + return PTR_ERR(di_args); + + if (memcmp(empty_uuid, di_args->uuid, BTRFS_UUID_SIZE) != 0) + s_uuid = di_args->uuid; + + mutex_lock(&fs_devices->device_list_mutex); + dev = btrfs_find_device(root, di_args->devid, s_uuid, NULL); + mutex_unlock(&fs_devices->device_list_mutex); + + if (!dev) { + ret = -ENODEV; + goto out; + } + + di_args->devid = dev->devid; + di_args->bytes_used = dev->bytes_used; + di_args->total_bytes = dev->total_bytes; + memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid)); + strncpy(di_args->path, dev->name, sizeof(di_args->path)); + +out: + if (ret == 0 && copy_to_user(arg, di_args, sizeof(*di_args))) + ret = -EFAULT; + + kfree(di_args); + return ret; +} + static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, u64 off, u64 olen, u64 destoff) { @@ -1925,7 +2244,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, } /* clone data */ - key.objectid = src->i_ino; + key.objectid = btrfs_ino(src); key.type = BTRFS_EXTENT_DATA_KEY; key.offset = 0; @@ -1952,7 +2271,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, btrfs_item_key_to_cpu(leaf, &key, slot); if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY || - key.objectid != src->i_ino) + key.objectid != btrfs_ino(src)) break; if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) { @@ -1988,14 +2307,14 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, datal = btrfs_file_extent_ram_bytes(leaf, extent); } - btrfs_release_path(root, path); + btrfs_release_path(path); if (key.offset + datal <= off || key.offset >= off+len) goto next; memcpy(&new_key, &key, sizeof(new_key)); - new_key.objectid = inode->i_ino; + new_key.objectid = btrfs_ino(inode); if (off <= key.offset) new_key.offset = key.offset + destoff - off; else @@ -2049,7 +2368,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, ret = btrfs_inc_extent_ref(trans, root, disko, diskl, 0, root->root_key.objectid, - inode->i_ino, + btrfs_ino(inode), new_key.offset - datao); BUG_ON(ret); } @@ -2098,7 +2417,7 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, } btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); inode->i_mtime = inode->i_ctime = CURRENT_TIME; @@ -2119,12 +2438,12 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd, btrfs_end_transaction(trans, root); } next: - btrfs_release_path(root, path); + btrfs_release_path(path); key.offset++; } ret = 0; out: - btrfs_release_path(root, path); + btrfs_release_path(path); unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS); out_unlock: mutex_unlock(&src->i_mutex); @@ -2177,12 +2496,10 @@ static long btrfs_ioctl_trans_start(struct file *file) if (ret) goto out; - mutex_lock(&root->fs_info->trans_mutex); - root->fs_info->open_ioctl_trans++; - mutex_unlock(&root->fs_info->trans_mutex); + atomic_inc(&root->fs_info->open_ioctl_trans); ret = -ENOMEM; - trans = btrfs_start_ioctl_transaction(root, 0); + trans = btrfs_start_ioctl_transaction(root); if (IS_ERR(trans)) goto out_drop; @@ -2190,9 +2507,7 @@ static long btrfs_ioctl_trans_start(struct file *file) return 0; out_drop: - mutex_lock(&root->fs_info->trans_mutex); - root->fs_info->open_ioctl_trans--; - mutex_unlock(&root->fs_info->trans_mutex); + atomic_dec(&root->fs_info->open_ioctl_trans); mnt_drop_write(file->f_path.mnt); out: return ret; @@ -2426,9 +2741,7 @@ long btrfs_ioctl_trans_end(struct file *file) btrfs_end_transaction(trans, root); - mutex_lock(&root->fs_info->trans_mutex); - root->fs_info->open_ioctl_trans--; - mutex_unlock(&root->fs_info->trans_mutex); + atomic_dec(&root->fs_info->open_ioctl_trans); mnt_drop_write(file->f_path.mnt); return 0; @@ -2471,6 +2784,58 @@ static noinline long btrfs_ioctl_wait_sync(struct file *file, void __user *argp) return btrfs_wait_for_commit(root, transid); } +static long btrfs_ioctl_scrub(struct btrfs_root *root, void __user *arg) +{ + int ret; + struct btrfs_ioctl_scrub_args *sa; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + ret = btrfs_scrub_dev(root, sa->devid, sa->start, sa->end, + &sa->progress, sa->flags & BTRFS_SCRUB_READONLY); + + if (copy_to_user(arg, sa, sizeof(*sa))) + ret = -EFAULT; + + kfree(sa); + return ret; +} + +static long btrfs_ioctl_scrub_cancel(struct btrfs_root *root, void __user *arg) +{ + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + return btrfs_scrub_cancel(root); +} + +static long btrfs_ioctl_scrub_progress(struct btrfs_root *root, + void __user *arg) +{ + struct btrfs_ioctl_scrub_args *sa; + int ret; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + sa = memdup_user(arg, sizeof(*sa)); + if (IS_ERR(sa)) + return PTR_ERR(sa); + + ret = btrfs_scrub_progress(root, sa->devid, &sa->progress); + + if (copy_to_user(arg, sa, sizeof(*sa))) + ret = -EFAULT; + + kfree(sa); + return ret; +} + long btrfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { @@ -2510,6 +2875,10 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_add_dev(root, argp); case BTRFS_IOC_RM_DEV: return btrfs_ioctl_rm_dev(root, argp); + case BTRFS_IOC_FS_INFO: + return btrfs_ioctl_fs_info(root, argp); + case BTRFS_IOC_DEV_INFO: + return btrfs_ioctl_dev_info(root, argp); case BTRFS_IOC_BALANCE: return btrfs_balance(root->fs_info->dev_root); case BTRFS_IOC_CLONE: @@ -2533,6 +2902,12 @@ long btrfs_ioctl(struct file *file, unsigned int return btrfs_ioctl_start_sync(file, argp); case BTRFS_IOC_WAIT_SYNC: return btrfs_ioctl_wait_sync(file, argp); + case BTRFS_IOC_SCRUB: + return btrfs_ioctl_scrub(root, argp); + case BTRFS_IOC_SCRUB_CANCEL: + return btrfs_ioctl_scrub_cancel(root, argp); + case BTRFS_IOC_SCRUB_PROGRESS: + return btrfs_ioctl_scrub_progress(root, argp); } return -ENOTTY; diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h index 8fb382167b13..ad1ea789fcb4 100644 --- a/fs/btrfs/ioctl.h +++ b/fs/btrfs/ioctl.h @@ -32,6 +32,8 @@ struct btrfs_ioctl_vol_args { #define BTRFS_SUBVOL_CREATE_ASYNC (1ULL << 0) #define BTRFS_SUBVOL_RDONLY (1ULL << 1) +#define BTRFS_FSID_SIZE 16 +#define BTRFS_UUID_SIZE 16 #define BTRFS_SUBVOL_NAME_MAX 4039 struct btrfs_ioctl_vol_args_v2 { @@ -42,6 +44,71 @@ struct btrfs_ioctl_vol_args_v2 { char name[BTRFS_SUBVOL_NAME_MAX + 1]; }; +/* + * structure to report errors and progress to userspace, either as a + * result of a finished scrub, a canceled scrub or a progress inquiry + */ +struct btrfs_scrub_progress { + __u64 data_extents_scrubbed; /* # of data extents scrubbed */ + __u64 tree_extents_scrubbed; /* # of tree extents scrubbed */ + __u64 data_bytes_scrubbed; /* # of data bytes scrubbed */ + __u64 tree_bytes_scrubbed; /* # of tree bytes scrubbed */ + __u64 read_errors; /* # of read errors encountered (EIO) */ + __u64 csum_errors; /* # of failed csum checks */ + __u64 verify_errors; /* # of occurences, where the metadata + * of a tree block did not match the + * expected values, like generation or + * logical */ + __u64 no_csum; /* # of 4k data block for which no csum + * is present, probably the result of + * data written with nodatasum */ + __u64 csum_discards; /* # of csum for which no data was found + * in the extent tree. */ + __u64 super_errors; /* # of bad super blocks encountered */ + __u64 malloc_errors; /* # of internal kmalloc errors. These + * will likely cause an incomplete + * scrub */ + __u64 uncorrectable_errors; /* # of errors where either no intact + * copy was found or the writeback + * failed */ + __u64 corrected_errors; /* # of errors corrected */ + __u64 last_physical; /* last physical address scrubbed. In + * case a scrub was aborted, this can + * be used to restart the scrub */ + __u64 unverified_errors; /* # of occurences where a read for a + * full (64k) bio failed, but the re- + * check succeeded for each 4k piece. + * Intermittent error. */ +}; + +#define BTRFS_SCRUB_READONLY 1 +struct btrfs_ioctl_scrub_args { + __u64 devid; /* in */ + __u64 start; /* in */ + __u64 end; /* in */ + __u64 flags; /* in */ + struct btrfs_scrub_progress progress; /* out */ + /* pad to 1k */ + __u64 unused[(1024-32-sizeof(struct btrfs_scrub_progress))/8]; +}; + +#define BTRFS_DEVICE_PATH_NAME_MAX 1024 +struct btrfs_ioctl_dev_info_args { + __u64 devid; /* in/out */ + __u8 uuid[BTRFS_UUID_SIZE]; /* in/out */ + __u64 bytes_used; /* out */ + __u64 total_bytes; /* out */ + __u64 unused[379]; /* pad to 4k */ + __u8 path[BTRFS_DEVICE_PATH_NAME_MAX]; /* out */ +}; + +struct btrfs_ioctl_fs_info_args { + __u64 max_id; /* out */ + __u64 num_devices; /* out */ + __u8 fsid[BTRFS_FSID_SIZE]; /* out */ + __u64 reserved[124]; /* pad to 1k */ +}; + #define BTRFS_INO_LOOKUP_PATH_MAX 4080 struct btrfs_ioctl_ino_lookup_args { __u64 treeid; @@ -114,37 +181,6 @@ struct btrfs_ioctl_clone_range_args { #define BTRFS_DEFRAG_RANGE_COMPRESS 1 #define BTRFS_DEFRAG_RANGE_START_IO 2 -struct btrfs_ioctl_defrag_range_args { - /* start of the defrag operation */ - __u64 start; - - /* number of bytes to defrag, use (u64)-1 to say all */ - __u64 len; - - /* - * flags for the operation, which can include turning - * on compression for this one defrag - */ - __u64 flags; - - /* - * any extent bigger than this will be considered - * already defragged. Use 0 to take the kernel default - * Use 1 to say every single extent must be rewritten - */ - __u32 extent_thresh; - - /* - * which compression method to use if turning on compression - * for this defrag operation. If unspecified, zlib will - * be used - */ - __u32 compress_type; - - /* spare for later */ - __u32 unused[4]; -}; - struct btrfs_ioctl_space_info { __u64 flags; __u64 total_bytes; @@ -203,4 +239,13 @@ struct btrfs_ioctl_space_args { struct btrfs_ioctl_vol_args_v2) #define BTRFS_IOC_SUBVOL_GETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 25, __u64) #define BTRFS_IOC_SUBVOL_SETFLAGS _IOW(BTRFS_IOCTL_MAGIC, 26, __u64) +#define BTRFS_IOC_SCRUB _IOWR(BTRFS_IOCTL_MAGIC, 27, \ + struct btrfs_ioctl_scrub_args) +#define BTRFS_IOC_SCRUB_CANCEL _IO(BTRFS_IOCTL_MAGIC, 28) +#define BTRFS_IOC_SCRUB_PROGRESS _IOWR(BTRFS_IOCTL_MAGIC, 29, \ + struct btrfs_ioctl_scrub_args) +#define BTRFS_IOC_DEV_INFO _IOWR(BTRFS_IOCTL_MAGIC, 30, \ + struct btrfs_ioctl_dev_info_args) +#define BTRFS_IOC_FS_INFO _IOR(BTRFS_IOCTL_MAGIC, 31, \ + struct btrfs_ioctl_fs_info_args) #endif diff --git a/fs/btrfs/locking.c b/fs/btrfs/locking.c index 6151f2ea38bb..66fa43dc3f0f 100644 --- a/fs/btrfs/locking.c +++ b/fs/btrfs/locking.c @@ -185,31 +185,6 @@ sleep: return 0; } -/* - * Very quick trylock, this does not spin or schedule. It returns - * 1 with the spinlock held if it was able to take the lock, or it - * returns zero if it was unable to take the lock. - * - * After this call, scheduling is not safe without first calling - * btrfs_set_lock_blocking() - */ -int btrfs_try_tree_lock(struct extent_buffer *eb) -{ - if (spin_trylock(&eb->lock)) { - if (test_bit(EXTENT_BUFFER_BLOCKING, &eb->bflags)) { - /* - * we've got the spinlock, but the real owner is - * blocking. Drop the spinlock and return failure - */ - spin_unlock(&eb->lock); - return 0; - } - return 1; - } - /* someone else has the spinlock giveup */ - return 0; -} - int btrfs_tree_unlock(struct extent_buffer *eb) { /* diff --git a/fs/btrfs/locking.h b/fs/btrfs/locking.h index 6c4ce457168c..5c33a560a2f1 100644 --- a/fs/btrfs/locking.h +++ b/fs/btrfs/locking.h @@ -21,8 +21,6 @@ int btrfs_tree_lock(struct extent_buffer *eb); int btrfs_tree_unlock(struct extent_buffer *eb); - -int btrfs_try_tree_lock(struct extent_buffer *eb); int btrfs_try_spin_lock(struct extent_buffer *eb); void btrfs_set_lock_blocking(struct extent_buffer *eb); diff --git a/fs/btrfs/ref-cache.c b/fs/btrfs/ref-cache.c index a97314cf6bd6..82d569cb6267 100644 --- a/fs/btrfs/ref-cache.c +++ b/fs/btrfs/ref-cache.c @@ -23,56 +23,6 @@ #include "ref-cache.h" #include "transaction.h" -/* - * leaf refs are used to cache the information about which extents - * a given leaf has references on. This allows us to process that leaf - * in btrfs_drop_snapshot without needing to read it back from disk. - */ - -/* - * kmalloc a leaf reference struct and update the counters for the - * total ref cache size - */ -struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root, - int nr_extents) -{ - struct btrfs_leaf_ref *ref; - size_t size = btrfs_leaf_ref_size(nr_extents); - - ref = kmalloc(size, GFP_NOFS); - if (ref) { - spin_lock(&root->fs_info->ref_cache_lock); - root->fs_info->total_ref_cache_size += size; - spin_unlock(&root->fs_info->ref_cache_lock); - - memset(ref, 0, sizeof(*ref)); - atomic_set(&ref->usage, 1); - INIT_LIST_HEAD(&ref->list); - } - return ref; -} - -/* - * free a leaf reference struct and update the counters for the - * total ref cache size - */ -void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref) -{ - if (!ref) - return; - WARN_ON(atomic_read(&ref->usage) == 0); - if (atomic_dec_and_test(&ref->usage)) { - size_t size = btrfs_leaf_ref_size(ref->nritems); - - BUG_ON(ref->in_tree); - kfree(ref); - - spin_lock(&root->fs_info->ref_cache_lock); - root->fs_info->total_ref_cache_size -= size; - spin_unlock(&root->fs_info->ref_cache_lock); - } -} - static struct rb_node *tree_insert(struct rb_root *root, u64 bytenr, struct rb_node *node) { @@ -116,117 +66,3 @@ static struct rb_node *tree_search(struct rb_root *root, u64 bytenr) } return NULL; } - -int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen, - int shared) -{ - struct btrfs_leaf_ref *ref = NULL; - struct btrfs_leaf_ref_tree *tree = root->ref_tree; - - if (shared) - tree = &root->fs_info->shared_ref_tree; - if (!tree) - return 0; - - spin_lock(&tree->lock); - while (!list_empty(&tree->list)) { - ref = list_entry(tree->list.next, struct btrfs_leaf_ref, list); - BUG_ON(ref->tree != tree); - if (ref->root_gen > max_root_gen) - break; - if (!xchg(&ref->in_tree, 0)) { - cond_resched_lock(&tree->lock); - continue; - } - - rb_erase(&ref->rb_node, &tree->root); - list_del_init(&ref->list); - - spin_unlock(&tree->lock); - btrfs_free_leaf_ref(root, ref); - cond_resched(); - spin_lock(&tree->lock); - } - spin_unlock(&tree->lock); - return 0; -} - -/* - * find the leaf ref for a given extent. This returns the ref struct with - * a usage reference incremented - */ -struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root, - u64 bytenr) -{ - struct rb_node *rb; - struct btrfs_leaf_ref *ref = NULL; - struct btrfs_leaf_ref_tree *tree = root->ref_tree; -again: - if (tree) { - spin_lock(&tree->lock); - rb = tree_search(&tree->root, bytenr); - if (rb) - ref = rb_entry(rb, struct btrfs_leaf_ref, rb_node); - if (ref) - atomic_inc(&ref->usage); - spin_unlock(&tree->lock); - if (ref) - return ref; - } - if (tree != &root->fs_info->shared_ref_tree) { - tree = &root->fs_info->shared_ref_tree; - goto again; - } - return NULL; -} - -/* - * add a fully filled in leaf ref struct - * remove all the refs older than a given root generation - */ -int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref, - int shared) -{ - int ret = 0; - struct rb_node *rb; - struct btrfs_leaf_ref_tree *tree = root->ref_tree; - - if (shared) - tree = &root->fs_info->shared_ref_tree; - - spin_lock(&tree->lock); - rb = tree_insert(&tree->root, ref->bytenr, &ref->rb_node); - if (rb) { - ret = -EEXIST; - } else { - atomic_inc(&ref->usage); - ref->tree = tree; - ref->in_tree = 1; - list_add_tail(&ref->list, &tree->list); - } - spin_unlock(&tree->lock); - return ret; -} - -/* - * remove a single leaf ref from the tree. This drops the ref held by the tree - * only - */ -int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref) -{ - struct btrfs_leaf_ref_tree *tree; - - if (!xchg(&ref->in_tree, 0)) - return 0; - - tree = ref->tree; - spin_lock(&tree->lock); - - rb_erase(&ref->rb_node, &tree->root); - list_del_init(&ref->list); - - spin_unlock(&tree->lock); - - btrfs_free_leaf_ref(root, ref); - return 0; -} diff --git a/fs/btrfs/ref-cache.h b/fs/btrfs/ref-cache.h index e2a55cb2072b..24f7001f6387 100644 --- a/fs/btrfs/ref-cache.h +++ b/fs/btrfs/ref-cache.h @@ -49,28 +49,4 @@ static inline size_t btrfs_leaf_ref_size(int nr_extents) return sizeof(struct btrfs_leaf_ref) + sizeof(struct btrfs_extent_info) * nr_extents; } - -static inline void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree) -{ - tree->root = RB_ROOT; - INIT_LIST_HEAD(&tree->list); - spin_lock_init(&tree->lock); -} - -static inline int btrfs_leaf_ref_tree_empty(struct btrfs_leaf_ref_tree *tree) -{ - return RB_EMPTY_ROOT(&tree->root); -} - -void btrfs_leaf_ref_tree_init(struct btrfs_leaf_ref_tree *tree); -struct btrfs_leaf_ref *btrfs_alloc_leaf_ref(struct btrfs_root *root, - int nr_extents); -void btrfs_free_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref); -struct btrfs_leaf_ref *btrfs_lookup_leaf_ref(struct btrfs_root *root, - u64 bytenr); -int btrfs_add_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref, - int shared); -int btrfs_remove_leaf_refs(struct btrfs_root *root, u64 max_root_gen, - int shared); -int btrfs_remove_leaf_ref(struct btrfs_root *root, struct btrfs_leaf_ref *ref); #endif diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index f340f7c99d09..5e0a3dc79a45 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -30,6 +30,7 @@ #include "btrfs_inode.h" #include "async-thread.h" #include "free-space-cache.h" +#include "inode-map.h" /* * backref_node, mapping_node and tree_block start with this @@ -507,6 +508,7 @@ static int update_backref_cache(struct btrfs_trans_handle *trans, return 1; } + static int should_ignore_root(struct btrfs_root *root) { struct btrfs_root *reloc_root; @@ -529,7 +531,6 @@ static int should_ignore_root(struct btrfs_root *root) */ return 1; } - /* * find reloc tree by address of tree root */ @@ -676,6 +677,8 @@ struct backref_node *build_backref_tree(struct reloc_control *rc, err = -ENOMEM; goto out; } + path1->reada = 1; + path2->reada = 2; node = alloc_backref_node(cache); if (!node) { @@ -961,7 +964,7 @@ again: lower = upper; upper = NULL; } - btrfs_release_path(root, path2); + btrfs_release_path(path2); next: if (ptr < end) { ptr += btrfs_extent_inline_ref_size(key.type); @@ -974,7 +977,7 @@ next: if (ptr >= end) path1->slots[0]++; } - btrfs_release_path(rc->extent_root, path1); + btrfs_release_path(path1); cur->checked = 1; WARN_ON(exist); @@ -1365,7 +1368,7 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans, int ret; if (!root->reloc_root) - return 0; + goto out; reloc_root = root->reloc_root; root_item = &reloc_root->root_item; @@ -1387,6 +1390,8 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans, ret = btrfs_update_root(trans, root->fs_info->tree_root, &reloc_root->root_key, root_item); BUG_ON(ret); + +out: return 0; } @@ -1409,9 +1414,9 @@ again: prev = node; entry = rb_entry(node, struct btrfs_inode, rb_node); - if (objectid < entry->vfs_inode.i_ino) + if (objectid < btrfs_ino(&entry->vfs_inode)) node = node->rb_left; - else if (objectid > entry->vfs_inode.i_ino) + else if (objectid > btrfs_ino(&entry->vfs_inode)) node = node->rb_right; else break; @@ -1419,7 +1424,7 @@ again: if (!node) { while (prev) { entry = rb_entry(prev, struct btrfs_inode, rb_node); - if (objectid <= entry->vfs_inode.i_ino) { + if (objectid <= btrfs_ino(&entry->vfs_inode)) { node = prev; break; } @@ -1434,7 +1439,7 @@ again: return inode; } - objectid = entry->vfs_inode.i_ino + 1; + objectid = btrfs_ino(&entry->vfs_inode) + 1; if (cond_resched_lock(&root->inode_lock)) goto again; @@ -1470,7 +1475,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr, return -ENOMEM; bytenr -= BTRFS_I(reloc_inode)->index_cnt; - ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino, + ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(reloc_inode), bytenr, 0); if (ret < 0) goto out; @@ -1558,11 +1563,11 @@ int replace_file_extents(struct btrfs_trans_handle *trans, if (first) { inode = find_next_inode(root, key.objectid); first = 0; - } else if (inode && inode->i_ino < key.objectid) { + } else if (inode && btrfs_ino(inode) < key.objectid) { btrfs_add_delayed_iput(inode); inode = find_next_inode(root, key.objectid); } - if (inode && inode->i_ino == key.objectid) { + if (inode && btrfs_ino(inode) == key.objectid) { end = key.offset + btrfs_file_extent_num_bytes(leaf, fi); WARN_ON(!IS_ALIGNED(key.offset, @@ -1749,7 +1754,7 @@ again: btrfs_node_key_to_cpu(path->nodes[level], &key, path->slots[level]); - btrfs_release_path(src, path); + btrfs_release_path(path); path->lowest_level = level; ret = btrfs_search_slot(trans, src, &key, path, 0, 1); @@ -1893,6 +1898,7 @@ static int invalidate_extent_cache(struct btrfs_root *root, struct inode *inode = NULL; u64 objectid; u64 start, end; + u64 ino; objectid = min_key->objectid; while (1) { @@ -1905,17 +1911,18 @@ static int invalidate_extent_cache(struct btrfs_root *root, inode = find_next_inode(root, objectid); if (!inode) break; + ino = btrfs_ino(inode); - if (inode->i_ino > max_key->objectid) { + if (ino > max_key->objectid) { iput(inode); break; } - objectid = inode->i_ino + 1; + objectid = ino + 1; if (!S_ISREG(inode->i_mode)) continue; - if (unlikely(min_key->objectid == inode->i_ino)) { + if (unlikely(min_key->objectid == ino)) { if (min_key->type > BTRFS_EXTENT_DATA_KEY) continue; if (min_key->type < BTRFS_EXTENT_DATA_KEY) @@ -1928,7 +1935,7 @@ static int invalidate_extent_cache(struct btrfs_root *root, start = 0; } - if (unlikely(max_key->objectid == inode->i_ino)) { + if (unlikely(max_key->objectid == ino)) { if (max_key->type < BTRFS_EXTENT_DATA_KEY) continue; if (max_key->type > BTRFS_EXTENT_DATA_KEY) { @@ -1996,6 +2003,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc, path = btrfs_alloc_path(); if (!path) return -ENOMEM; + path->reada = 1; reloc_root = root->reloc_root; root_item = &reloc_root->root_item; @@ -2136,10 +2144,11 @@ int prepare_to_merge(struct reloc_control *rc, int err) u64 num_bytes = 0; int ret; - mutex_lock(&root->fs_info->trans_mutex); + mutex_lock(&root->fs_info->reloc_mutex); rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2; rc->merging_rsv_size += rc->nodes_relocated * 2; - mutex_unlock(&root->fs_info->trans_mutex); + mutex_unlock(&root->fs_info->reloc_mutex); + again: if (!err) { num_bytes = rc->merging_rsv_size; @@ -2149,7 +2158,7 @@ again: err = ret; } - trans = btrfs_join_transaction(rc->extent_root, 1); + trans = btrfs_join_transaction(rc->extent_root); if (IS_ERR(trans)) { if (!err) btrfs_block_rsv_release(rc->extent_root, @@ -2208,9 +2217,16 @@ int merge_reloc_roots(struct reloc_control *rc) int ret; again: root = rc->extent_root; - mutex_lock(&root->fs_info->trans_mutex); + + /* + * this serializes us with btrfs_record_root_in_transaction, + * we have to make sure nobody is in the middle of + * adding their roots to the list while we are + * doing this splice + */ + mutex_lock(&root->fs_info->reloc_mutex); list_splice_init(&rc->reloc_roots, &reloc_roots); - mutex_unlock(&root->fs_info->trans_mutex); + mutex_unlock(&root->fs_info->reloc_mutex); while (!list_empty(&reloc_roots)) { found = 1; @@ -2496,7 +2512,7 @@ static int do_relocation(struct btrfs_trans_handle *trans, path->locks[upper->level] = 0; slot = path->slots[upper->level]; - btrfs_release_path(NULL, path); + btrfs_release_path(path); } else { ret = btrfs_bin_search(upper->eb, key, upper->level, &slot); @@ -2737,7 +2753,7 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans, } else { path->lowest_level = node->level; ret = btrfs_search_slot(trans, root, key, path, 0, 1); - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret > 0) ret = 0; } @@ -2870,7 +2886,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end, struct extent_map *em; int ret = 0; - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) return -ENOMEM; @@ -3119,7 +3135,7 @@ static int add_tree_block(struct reloc_control *rc, #endif } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); BUG_ON(level == -1); @@ -3220,7 +3236,7 @@ static int delete_block_group_cache(struct btrfs_fs_info *fs_info, key.offset = 0; inode = btrfs_iget(fs_info->sb, &key, root, NULL); - if (!inode || IS_ERR(inode) || is_bad_inode(inode)) { + if (IS_ERR_OR_NULL(inode) || is_bad_inode(inode)) { if (inode && !IS_ERR(inode)) iput(inode); return -ENOENT; @@ -3233,7 +3249,7 @@ truncate: goto out; } - trans = btrfs_join_transaction(root, 0); + trans = btrfs_join_transaction(root); if (IS_ERR(trans)) { btrfs_free_path(path); ret = PTR_ERR(trans); @@ -3297,6 +3313,7 @@ static int find_data_references(struct reloc_control *rc, path = btrfs_alloc_path(); if (!path) return -ENOMEM; + path->reada = 1; root = read_fs_root(rc->extent_root->fs_info, ref_root); if (IS_ERR(root)) { @@ -3505,7 +3522,7 @@ int add_data_references(struct reloc_control *rc, } path->slots[0]++; } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); if (err) free_block_list(blocks); return err; @@ -3568,7 +3585,7 @@ next: EXTENT_DIRTY); if (ret == 0 && start <= key.objectid) { - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); rc->search_start = end + 1; } else { rc->search_start = key.objectid + key.offset; @@ -3576,24 +3593,26 @@ next: return 0; } } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); return ret; } static void set_reloc_control(struct reloc_control *rc) { struct btrfs_fs_info *fs_info = rc->extent_root->fs_info; - mutex_lock(&fs_info->trans_mutex); + + mutex_lock(&fs_info->reloc_mutex); fs_info->reloc_ctl = rc; - mutex_unlock(&fs_info->trans_mutex); + mutex_unlock(&fs_info->reloc_mutex); } static void unset_reloc_control(struct reloc_control *rc) { struct btrfs_fs_info *fs_info = rc->extent_root->fs_info; - mutex_lock(&fs_info->trans_mutex); + + mutex_lock(&fs_info->reloc_mutex); fs_info->reloc_ctl = NULL; - mutex_unlock(&fs_info->trans_mutex); + mutex_unlock(&fs_info->reloc_mutex); } static int check_extent_flags(u64 flags) @@ -3642,7 +3661,7 @@ int prepare_to_relocate(struct reloc_control *rc) rc->create_reloc_tree = 1; set_reloc_control(rc); - trans = btrfs_join_transaction(rc->extent_root, 1); + trans = btrfs_join_transaction(rc->extent_root); BUG_ON(IS_ERR(trans)); btrfs_commit_transaction(trans, rc->extent_root); return 0; @@ -3665,6 +3684,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) path = btrfs_alloc_path(); if (!path) return -ENOMEM; + path->reada = 1; ret = prepare_to_relocate(rc); if (ret) { @@ -3713,7 +3733,7 @@ restart: flags = BTRFS_EXTENT_FLAG_DATA; if (path_change) { - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); path->search_commit_root = 1; path->skip_locking = 1; @@ -3736,7 +3756,7 @@ restart: (flags & BTRFS_EXTENT_FLAG_DATA)) { ret = add_data_references(rc, &key, path, &blocks); } else { - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); ret = 0; } if (ret < 0) { @@ -3799,7 +3819,7 @@ restart: } } - btrfs_release_path(rc->extent_root, path); + btrfs_release_path(path); clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY, GFP_NOFS); @@ -3831,7 +3851,7 @@ restart: btrfs_block_rsv_release(rc->extent_root, rc->block_rsv, (u64)-1); /* get rid of pinned extents */ - trans = btrfs_join_transaction(rc->extent_root, 1); + trans = btrfs_join_transaction(rc->extent_root); if (IS_ERR(trans)) err = PTR_ERR(trans); else @@ -3867,7 +3887,7 @@ static int __insert_orphan_inode(struct btrfs_trans_handle *trans, btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS | BTRFS_INODE_PREALLOC); btrfs_mark_buffer_dirty(leaf); - btrfs_release_path(root, path); + btrfs_release_path(path); out: btrfs_free_path(path); return ret; @@ -3897,7 +3917,7 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info, if (IS_ERR(trans)) return ERR_CAST(trans); - err = btrfs_find_free_objectid(trans, root, objectid, &objectid); + err = btrfs_find_free_objectid(root, &objectid); if (err) goto out; @@ -3935,7 +3955,7 @@ static struct reloc_control *alloc_reloc_control(void) INIT_LIST_HEAD(&rc->reloc_roots); backref_cache_init(&rc->backref_cache); mapping_tree_init(&rc->reloc_root_tree); - extent_io_tree_init(&rc->processed_blocks, NULL, GFP_NOFS); + extent_io_tree_init(&rc->processed_blocks, NULL); return rc; } @@ -4090,6 +4110,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) path = btrfs_alloc_path(); if (!path) return -ENOMEM; + path->reada = -1; key.objectid = BTRFS_TREE_RELOC_OBJECTID; key.type = BTRFS_ROOT_ITEM_KEY; @@ -4109,7 +4130,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) } leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(root->fs_info->tree_root, path); + btrfs_release_path(path); if (key.objectid != BTRFS_TREE_RELOC_OBJECTID || key.type != BTRFS_ROOT_ITEM_KEY) @@ -4141,7 +4162,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) key.offset--; } - btrfs_release_path(root->fs_info->tree_root, path); + btrfs_release_path(path); if (list_empty(&reloc_roots)) goto out; @@ -4156,7 +4177,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) set_reloc_control(rc); - trans = btrfs_join_transaction(rc->extent_root, 1); + trans = btrfs_join_transaction(rc->extent_root); if (IS_ERR(trans)) { unset_reloc_control(rc); err = PTR_ERR(trans); @@ -4190,7 +4211,7 @@ int btrfs_recover_relocation(struct btrfs_root *root) unset_reloc_control(rc); - trans = btrfs_join_transaction(rc->extent_root, 1); + trans = btrfs_join_transaction(rc->extent_root); if (IS_ERR(trans)) err = PTR_ERR(trans); else @@ -4242,7 +4263,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len) disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt; ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr, - disk_bytenr + len - 1, &list); + disk_bytenr + len - 1, &list, 0); while (!list_empty(&list)) { sums = list_entry(list.next, struct btrfs_ordered_sum, list); diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index 6928bff62daa..ebe45443de06 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -22,53 +22,6 @@ #include "print-tree.h" /* - * search forward for a root, starting with objectid 'search_start' - * if a root key is found, the objectid we find is filled into 'found_objectid' - * and 0 is returned. < 0 is returned on error, 1 if there is nothing - * left in the tree. - */ -int btrfs_search_root(struct btrfs_root *root, u64 search_start, - u64 *found_objectid) -{ - struct btrfs_path *path; - struct btrfs_key search_key; - int ret; - - root = root->fs_info->tree_root; - search_key.objectid = search_start; - search_key.type = (u8)-1; - search_key.offset = (u64)-1; - - path = btrfs_alloc_path(); - BUG_ON(!path); -again: - ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0); - if (ret < 0) - goto out; - if (ret == 0) { - ret = 1; - goto out; - } - if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) { - ret = btrfs_next_leaf(root, path); - if (ret) - goto out; - } - btrfs_item_key_to_cpu(path->nodes[0], &search_key, path->slots[0]); - if (search_key.type != BTRFS_ROOT_ITEM_KEY) { - search_key.offset++; - btrfs_release_path(root, path); - goto again; - } - ret = 0; - *found_objectid = search_key.objectid; - -out: - btrfs_free_path(path); - return ret; -} - -/* * lookup the root with the highest offset for a given objectid. The key we do * find is copied into 'key'. If we find something return 0, otherwise 1, < 0 * on error. @@ -230,7 +183,7 @@ again: memcpy(&found_key, &key, sizeof(key)); key.offset++; - btrfs_release_path(root, path); + btrfs_release_path(path); dead_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root, &found_key); @@ -292,7 +245,7 @@ int btrfs_find_orphan_roots(struct btrfs_root *tree_root) } btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(tree_root, path); + btrfs_release_path(path); if (key.objectid != BTRFS_ORPHAN_OBJECTID || key.type != BTRFS_ORPHAN_ITEM_KEY) @@ -385,18 +338,22 @@ again: *sequence = btrfs_root_ref_sequence(leaf, ref); ret = btrfs_del_item(trans, tree_root, path); - BUG_ON(ret); + if (ret) { + err = ret; + goto out; + } } else err = -ENOENT; if (key.type == BTRFS_ROOT_BACKREF_KEY) { - btrfs_release_path(tree_root, path); + btrfs_release_path(path); key.objectid = ref_id; key.type = BTRFS_ROOT_REF_KEY; key.offset = root_id; goto again; } +out: btrfs_free_path(path); return err; } @@ -463,7 +420,7 @@ again: btrfs_mark_buffer_dirty(leaf); if (key.type == BTRFS_ROOT_BACKREF_KEY) { - btrfs_release_path(tree_root, path); + btrfs_release_path(path); key.objectid = ref_id; key.type = BTRFS_ROOT_REF_KEY; key.offset = root_id; diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c new file mode 100644 index 000000000000..a8d03d5efb5d --- /dev/null +++ b/fs/btrfs/scrub.c @@ -0,0 +1,1395 @@ +/* + * Copyright (C) 2011 STRATO. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License v2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + */ + +#include <linux/blkdev.h> +#include "ctree.h" +#include "volumes.h" +#include "disk-io.h" +#include "ordered-data.h" + +/* + * This is only the first step towards a full-features scrub. It reads all + * extent and super block and verifies the checksums. In case a bad checksum + * is found or the extent cannot be read, good data will be written back if + * any can be found. + * + * Future enhancements: + * - To enhance the performance, better read-ahead strategies for the + * extent-tree can be employed. + * - In case an unrepairable extent is encountered, track which files are + * affected and report them + * - In case of a read error on files with nodatasum, map the file and read + * the extent to trigger a writeback of the good copy + * - track and record media errors, throw out bad devices + * - add a mode to also read unallocated space + * - make the prefetch cancellable + */ + +struct scrub_bio; +struct scrub_page; +struct scrub_dev; +static void scrub_bio_end_io(struct bio *bio, int err); +static void scrub_checksum(struct btrfs_work *work); +static int scrub_checksum_data(struct scrub_dev *sdev, + struct scrub_page *spag, void *buffer); +static int scrub_checksum_tree_block(struct scrub_dev *sdev, + struct scrub_page *spag, u64 logical, + void *buffer); +static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer); +static int scrub_fixup_check(struct scrub_bio *sbio, int ix); +static void scrub_fixup_end_io(struct bio *bio, int err); +static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, + struct page *page); +static void scrub_fixup(struct scrub_bio *sbio, int ix); + +#define SCRUB_PAGES_PER_BIO 16 /* 64k per bio */ +#define SCRUB_BIOS_PER_DEV 16 /* 1 MB per device in flight */ + +struct scrub_page { + u64 flags; /* extent flags */ + u64 generation; + u64 mirror_num; + int have_csum; + u8 csum[BTRFS_CSUM_SIZE]; +}; + +struct scrub_bio { + int index; + struct scrub_dev *sdev; + struct bio *bio; + int err; + u64 logical; + u64 physical; + struct scrub_page spag[SCRUB_PAGES_PER_BIO]; + u64 count; + int next_free; + struct btrfs_work work; +}; + +struct scrub_dev { + struct scrub_bio *bios[SCRUB_BIOS_PER_DEV]; + struct btrfs_device *dev; + int first_free; + int curr; + atomic_t in_flight; + spinlock_t list_lock; + wait_queue_head_t list_wait; + u16 csum_size; + struct list_head csum_list; + atomic_t cancel_req; + int readonly; + /* + * statistics + */ + struct btrfs_scrub_progress stat; + spinlock_t stat_lock; +}; + +static void scrub_free_csums(struct scrub_dev *sdev) +{ + while (!list_empty(&sdev->csum_list)) { + struct btrfs_ordered_sum *sum; + sum = list_first_entry(&sdev->csum_list, + struct btrfs_ordered_sum, list); + list_del(&sum->list); + kfree(sum); + } +} + +static void scrub_free_bio(struct bio *bio) +{ + int i; + struct page *last_page = NULL; + + if (!bio) + return; + + for (i = 0; i < bio->bi_vcnt; ++i) { + if (bio->bi_io_vec[i].bv_page == last_page) + continue; + last_page = bio->bi_io_vec[i].bv_page; + __free_page(last_page); + } + bio_put(bio); +} + +static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev) +{ + int i; + + if (!sdev) + return; + + for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) { + struct scrub_bio *sbio = sdev->bios[i]; + + if (!sbio) + break; + + scrub_free_bio(sbio->bio); + kfree(sbio); + } + + scrub_free_csums(sdev); + kfree(sdev); +} + +static noinline_for_stack +struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev) +{ + struct scrub_dev *sdev; + int i; + struct btrfs_fs_info *fs_info = dev->dev_root->fs_info; + + sdev = kzalloc(sizeof(*sdev), GFP_NOFS); + if (!sdev) + goto nomem; + sdev->dev = dev; + for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) { + struct scrub_bio *sbio; + + sbio = kzalloc(sizeof(*sbio), GFP_NOFS); + if (!sbio) + goto nomem; + sdev->bios[i] = sbio; + + sbio->index = i; + sbio->sdev = sdev; + sbio->count = 0; + sbio->work.func = scrub_checksum; + + if (i != SCRUB_BIOS_PER_DEV-1) + sdev->bios[i]->next_free = i + 1; + else + sdev->bios[i]->next_free = -1; + } + sdev->first_free = 0; + sdev->curr = -1; + atomic_set(&sdev->in_flight, 0); + atomic_set(&sdev->cancel_req, 0); + sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy); + INIT_LIST_HEAD(&sdev->csum_list); + + spin_lock_init(&sdev->list_lock); + spin_lock_init(&sdev->stat_lock); + init_waitqueue_head(&sdev->list_wait); + return sdev; + +nomem: + scrub_free_dev(sdev); + return ERR_PTR(-ENOMEM); +} + +/* + * scrub_recheck_error gets called when either verification of the page + * failed or the bio failed to read, e.g. with EIO. In the latter case, + * recheck_error gets called for every page in the bio, even though only + * one may be bad + */ +static void scrub_recheck_error(struct scrub_bio *sbio, int ix) +{ + if (sbio->err) { + if (scrub_fixup_io(READ, sbio->sdev->dev->bdev, + (sbio->physical + ix * PAGE_SIZE) >> 9, + sbio->bio->bi_io_vec[ix].bv_page) == 0) { + if (scrub_fixup_check(sbio, ix) == 0) + return; + } + } + + scrub_fixup(sbio, ix); +} + +static int scrub_fixup_check(struct scrub_bio *sbio, int ix) +{ + int ret = 1; + struct page *page; + void *buffer; + u64 flags = sbio->spag[ix].flags; + + page = sbio->bio->bi_io_vec[ix].bv_page; + buffer = kmap_atomic(page, KM_USER0); + if (flags & BTRFS_EXTENT_FLAG_DATA) { + ret = scrub_checksum_data(sbio->sdev, + sbio->spag + ix, buffer); + } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + ret = scrub_checksum_tree_block(sbio->sdev, + sbio->spag + ix, + sbio->logical + ix * PAGE_SIZE, + buffer); + } else { + WARN_ON(1); + } + kunmap_atomic(buffer, KM_USER0); + + return ret; +} + +static void scrub_fixup_end_io(struct bio *bio, int err) +{ + complete((struct completion *)bio->bi_private); +} + +static void scrub_fixup(struct scrub_bio *sbio, int ix) +{ + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; + struct btrfs_multi_bio *multi = NULL; + u64 logical = sbio->logical + ix * PAGE_SIZE; + u64 length; + int i; + int ret; + DECLARE_COMPLETION_ONSTACK(complete); + + if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) && + (sbio->spag[ix].have_csum == 0)) { + /* + * nodatasum, don't try to fix anything + * FIXME: we can do better, open the inode and trigger a + * writeback + */ + goto uncorrectable; + } + + length = PAGE_SIZE; + ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length, + &multi, 0); + if (ret || !multi || length < PAGE_SIZE) { + printk(KERN_ERR + "scrub_fixup: btrfs_map_block failed us for %llu\n", + (unsigned long long)logical); + WARN_ON(1); + return; + } + + if (multi->num_stripes == 1) + /* there aren't any replicas */ + goto uncorrectable; + + /* + * first find a good copy + */ + for (i = 0; i < multi->num_stripes; ++i) { + if (i == sbio->spag[ix].mirror_num) + continue; + + if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev, + multi->stripes[i].physical >> 9, + sbio->bio->bi_io_vec[ix].bv_page)) { + /* I/O-error, this is not a good copy */ + continue; + } + + if (scrub_fixup_check(sbio, ix) == 0) + break; + } + if (i == multi->num_stripes) + goto uncorrectable; + + if (!sdev->readonly) { + /* + * bi_io_vec[ix].bv_page now contains good data, write it back + */ + if (scrub_fixup_io(WRITE, sdev->dev->bdev, + (sbio->physical + ix * PAGE_SIZE) >> 9, + sbio->bio->bi_io_vec[ix].bv_page)) { + /* I/O-error, writeback failed, give up */ + goto uncorrectable; + } + } + + kfree(multi); + spin_lock(&sdev->stat_lock); + ++sdev->stat.corrected_errors; + spin_unlock(&sdev->stat_lock); + + if (printk_ratelimit()) + printk(KERN_ERR "btrfs: fixed up at %llu\n", + (unsigned long long)logical); + return; + +uncorrectable: + kfree(multi); + spin_lock(&sdev->stat_lock); + ++sdev->stat.uncorrectable_errors; + spin_unlock(&sdev->stat_lock); + + if (printk_ratelimit()) + printk(KERN_ERR "btrfs: unable to fixup at %llu\n", + (unsigned long long)logical); +} + +static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector, + struct page *page) +{ + struct bio *bio = NULL; + int ret; + DECLARE_COMPLETION_ONSTACK(complete); + + bio = bio_alloc(GFP_NOFS, 1); + bio->bi_bdev = bdev; + bio->bi_sector = sector; + bio_add_page(bio, page, PAGE_SIZE, 0); + bio->bi_end_io = scrub_fixup_end_io; + bio->bi_private = &complete; + submit_bio(rw, bio); + + /* this will also unplug the queue */ + wait_for_completion(&complete); + + ret = !test_bit(BIO_UPTODATE, &bio->bi_flags); + bio_put(bio); + return ret; +} + +static void scrub_bio_end_io(struct bio *bio, int err) +{ + struct scrub_bio *sbio = bio->bi_private; + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + + sbio->err = err; + sbio->bio = bio; + + btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work); +} + +static void scrub_checksum(struct btrfs_work *work) +{ + struct scrub_bio *sbio = container_of(work, struct scrub_bio, work); + struct scrub_dev *sdev = sbio->sdev; + struct page *page; + void *buffer; + int i; + u64 flags; + u64 logical; + int ret; + + if (sbio->err) { + for (i = 0; i < sbio->count; ++i) + scrub_recheck_error(sbio, i); + + sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1); + sbio->bio->bi_flags |= 1 << BIO_UPTODATE; + sbio->bio->bi_phys_segments = 0; + sbio->bio->bi_idx = 0; + + for (i = 0; i < sbio->count; i++) { + struct bio_vec *bi; + bi = &sbio->bio->bi_io_vec[i]; + bi->bv_offset = 0; + bi->bv_len = PAGE_SIZE; + } + + spin_lock(&sdev->stat_lock); + ++sdev->stat.read_errors; + spin_unlock(&sdev->stat_lock); + goto out; + } + for (i = 0; i < sbio->count; ++i) { + page = sbio->bio->bi_io_vec[i].bv_page; + buffer = kmap_atomic(page, KM_USER0); + flags = sbio->spag[i].flags; + logical = sbio->logical + i * PAGE_SIZE; + ret = 0; + if (flags & BTRFS_EXTENT_FLAG_DATA) { + ret = scrub_checksum_data(sdev, sbio->spag + i, buffer); + } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) { + ret = scrub_checksum_tree_block(sdev, sbio->spag + i, + logical, buffer); + } else if (flags & BTRFS_EXTENT_FLAG_SUPER) { + BUG_ON(i); + (void)scrub_checksum_super(sbio, buffer); + } else { + WARN_ON(1); + } + kunmap_atomic(buffer, KM_USER0); + if (ret) + scrub_recheck_error(sbio, i); + } + +out: + scrub_free_bio(sbio->bio); + sbio->bio = NULL; + spin_lock(&sdev->list_lock); + sbio->next_free = sdev->first_free; + sdev->first_free = sbio->index; + spin_unlock(&sdev->list_lock); + atomic_dec(&sdev->in_flight); + wake_up(&sdev->list_wait); +} + +static int scrub_checksum_data(struct scrub_dev *sdev, + struct scrub_page *spag, void *buffer) +{ + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + int fail = 0; + struct btrfs_root *root = sdev->dev->dev_root; + + if (!spag->have_csum) + return 0; + + crc = btrfs_csum_data(root, buffer, crc, PAGE_SIZE); + btrfs_csum_final(crc, csum); + if (memcmp(csum, spag->csum, sdev->csum_size)) + fail = 1; + + spin_lock(&sdev->stat_lock); + ++sdev->stat.data_extents_scrubbed; + sdev->stat.data_bytes_scrubbed += PAGE_SIZE; + if (fail) + ++sdev->stat.csum_errors; + spin_unlock(&sdev->stat_lock); + + return fail; +} + +static int scrub_checksum_tree_block(struct scrub_dev *sdev, + struct scrub_page *spag, u64 logical, + void *buffer) +{ + struct btrfs_header *h; + struct btrfs_root *root = sdev->dev->dev_root; + struct btrfs_fs_info *fs_info = root->fs_info; + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + int fail = 0; + int crc_fail = 0; + + /* + * we don't use the getter functions here, as we + * a) don't have an extent buffer and + * b) the page is already kmapped + */ + h = (struct btrfs_header *)buffer; + + if (logical != le64_to_cpu(h->bytenr)) + ++fail; + + if (spag->generation != le64_to_cpu(h->generation)) + ++fail; + + if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE)) + ++fail; + + if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid, + BTRFS_UUID_SIZE)) + ++fail; + + crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc, + PAGE_SIZE - BTRFS_CSUM_SIZE); + btrfs_csum_final(crc, csum); + if (memcmp(csum, h->csum, sdev->csum_size)) + ++crc_fail; + + spin_lock(&sdev->stat_lock); + ++sdev->stat.tree_extents_scrubbed; + sdev->stat.tree_bytes_scrubbed += PAGE_SIZE; + if (crc_fail) + ++sdev->stat.csum_errors; + if (fail) + ++sdev->stat.verify_errors; + spin_unlock(&sdev->stat_lock); + + return fail || crc_fail; +} + +static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer) +{ + struct btrfs_super_block *s; + u64 logical; + struct scrub_dev *sdev = sbio->sdev; + struct btrfs_root *root = sdev->dev->dev_root; + struct btrfs_fs_info *fs_info = root->fs_info; + u8 csum[BTRFS_CSUM_SIZE]; + u32 crc = ~(u32)0; + int fail = 0; + + s = (struct btrfs_super_block *)buffer; + logical = sbio->logical; + + if (logical != le64_to_cpu(s->bytenr)) + ++fail; + + if (sbio->spag[0].generation != le64_to_cpu(s->generation)) + ++fail; + + if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE)) + ++fail; + + crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc, + PAGE_SIZE - BTRFS_CSUM_SIZE); + btrfs_csum_final(crc, csum); + if (memcmp(csum, s->csum, sbio->sdev->csum_size)) + ++fail; + + if (fail) { + /* + * if we find an error in a super block, we just report it. + * They will get written with the next transaction commit + * anyway + */ + spin_lock(&sdev->stat_lock); + ++sdev->stat.super_errors; + spin_unlock(&sdev->stat_lock); + } + + return fail; +} + +static int scrub_submit(struct scrub_dev *sdev) +{ + struct scrub_bio *sbio; + struct bio *bio; + int i; + + if (sdev->curr == -1) + return 0; + + sbio = sdev->bios[sdev->curr]; + + bio = bio_alloc(GFP_NOFS, sbio->count); + if (!bio) + goto nomem; + + bio->bi_private = sbio; + bio->bi_end_io = scrub_bio_end_io; + bio->bi_bdev = sdev->dev->bdev; + bio->bi_sector = sbio->physical >> 9; + + for (i = 0; i < sbio->count; ++i) { + struct page *page; + int ret; + + page = alloc_page(GFP_NOFS); + if (!page) + goto nomem; + + ret = bio_add_page(bio, page, PAGE_SIZE, 0); + if (!ret) { + __free_page(page); + goto nomem; + } + } + + sbio->err = 0; + sdev->curr = -1; + atomic_inc(&sdev->in_flight); + + submit_bio(READ, bio); + + return 0; + +nomem: + scrub_free_bio(bio); + + return -ENOMEM; +} + +static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len, + u64 physical, u64 flags, u64 gen, u64 mirror_num, + u8 *csum, int force) +{ + struct scrub_bio *sbio; + +again: + /* + * grab a fresh bio or wait for one to become available + */ + while (sdev->curr == -1) { + spin_lock(&sdev->list_lock); + sdev->curr = sdev->first_free; + if (sdev->curr != -1) { + sdev->first_free = sdev->bios[sdev->curr]->next_free; + sdev->bios[sdev->curr]->next_free = -1; + sdev->bios[sdev->curr]->count = 0; + spin_unlock(&sdev->list_lock); + } else { + spin_unlock(&sdev->list_lock); + wait_event(sdev->list_wait, sdev->first_free != -1); + } + } + sbio = sdev->bios[sdev->curr]; + if (sbio->count == 0) { + sbio->physical = physical; + sbio->logical = logical; + } else if (sbio->physical + sbio->count * PAGE_SIZE != physical || + sbio->logical + sbio->count * PAGE_SIZE != logical) { + int ret; + + ret = scrub_submit(sdev); + if (ret) + return ret; + goto again; + } + sbio->spag[sbio->count].flags = flags; + sbio->spag[sbio->count].generation = gen; + sbio->spag[sbio->count].have_csum = 0; + sbio->spag[sbio->count].mirror_num = mirror_num; + if (csum) { + sbio->spag[sbio->count].have_csum = 1; + memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size); + } + ++sbio->count; + if (sbio->count == SCRUB_PAGES_PER_BIO || force) { + int ret; + + ret = scrub_submit(sdev); + if (ret) + return ret; + } + + return 0; +} + +static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len, + u8 *csum) +{ + struct btrfs_ordered_sum *sum = NULL; + int ret = 0; + unsigned long i; + unsigned long num_sectors; + u32 sectorsize = sdev->dev->dev_root->sectorsize; + + while (!list_empty(&sdev->csum_list)) { + sum = list_first_entry(&sdev->csum_list, + struct btrfs_ordered_sum, list); + if (sum->bytenr > logical) + return 0; + if (sum->bytenr + sum->len > logical) + break; + + ++sdev->stat.csum_discards; + list_del(&sum->list); + kfree(sum); + sum = NULL; + } + if (!sum) + return 0; + + num_sectors = sum->len / sectorsize; + for (i = 0; i < num_sectors; ++i) { + if (sum->sums[i].bytenr == logical) { + memcpy(csum, &sum->sums[i].sum, sdev->csum_size); + ret = 1; + break; + } + } + if (ret && i == num_sectors - 1) { + list_del(&sum->list); + kfree(sum); + } + return ret; +} + +/* scrub extent tries to collect up to 64 kB for each bio */ +static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len, + u64 physical, u64 flags, u64 gen, u64 mirror_num) +{ + int ret; + u8 csum[BTRFS_CSUM_SIZE]; + + while (len) { + u64 l = min_t(u64, len, PAGE_SIZE); + int have_csum = 0; + + if (flags & BTRFS_EXTENT_FLAG_DATA) { + /* push csums to sbio */ + have_csum = scrub_find_csum(sdev, logical, l, csum); + if (have_csum == 0) + ++sdev->stat.no_csum; + } + ret = scrub_page(sdev, logical, l, physical, flags, gen, + mirror_num, have_csum ? csum : NULL, 0); + if (ret) + return ret; + len -= l; + logical += l; + physical += l; + } + return 0; +} + +static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev, + struct map_lookup *map, int num, u64 base, u64 length) +{ + struct btrfs_path *path; + struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info; + struct btrfs_root *root = fs_info->extent_root; + struct btrfs_root *csum_root = fs_info->csum_root; + struct btrfs_extent_item *extent; + struct blk_plug plug; + u64 flags; + int ret; + int slot; + int i; + u64 nstripes; + int start_stripe; + struct extent_buffer *l; + struct btrfs_key key; + u64 physical; + u64 logical; + u64 generation; + u64 mirror_num; + + u64 increment = map->stripe_len; + u64 offset; + + nstripes = length; + offset = 0; + do_div(nstripes, map->stripe_len); + if (map->type & BTRFS_BLOCK_GROUP_RAID0) { + offset = map->stripe_len * num; + increment = map->stripe_len * map->num_stripes; + mirror_num = 0; + } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) { + int factor = map->num_stripes / map->sub_stripes; + offset = map->stripe_len * (num / map->sub_stripes); + increment = map->stripe_len * factor; + mirror_num = num % map->sub_stripes; + } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { + increment = map->stripe_len; + mirror_num = num % map->num_stripes; + } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { + increment = map->stripe_len; + mirror_num = num % map->num_stripes; + } else { + increment = map->stripe_len; + mirror_num = 0; + } + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + path->reada = 2; + path->search_commit_root = 1; + path->skip_locking = 1; + + /* + * find all extents for each stripe and just read them to get + * them into the page cache + * FIXME: we can do better. build a more intelligent prefetching + */ + logical = base + offset; + physical = map->stripes[num].physical; + ret = 0; + for (i = 0; i < nstripes; ++i) { + key.objectid = logical; + key.type = BTRFS_EXTENT_ITEM_KEY; + key.offset = (u64)0; + + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out_noplug; + + /* + * we might miss half an extent here, but that doesn't matter, + * as it's only the prefetch + */ + while (1) { + l = path->nodes[0]; + slot = path->slots[0]; + if (slot >= btrfs_header_nritems(l)) { + ret = btrfs_next_leaf(root, path); + if (ret == 0) + continue; + if (ret < 0) + goto out_noplug; + + break; + } + btrfs_item_key_to_cpu(l, &key, slot); + + if (key.objectid >= logical + map->stripe_len) + break; + + path->slots[0]++; + } + btrfs_release_path(path); + logical += increment; + physical += map->stripe_len; + cond_resched(); + } + + /* + * collect all data csums for the stripe to avoid seeking during + * the scrub. This might currently (crc32) end up to be about 1MB + */ + start_stripe = 0; + blk_start_plug(&plug); +again: + logical = base + offset + start_stripe * increment; + for (i = start_stripe; i < nstripes; ++i) { + ret = btrfs_lookup_csums_range(csum_root, logical, + logical + map->stripe_len - 1, + &sdev->csum_list, 1); + if (ret) + goto out; + + logical += increment; + cond_resched(); + } + /* + * now find all extents for each stripe and scrub them + */ + logical = base + offset + start_stripe * increment; + physical = map->stripes[num].physical + start_stripe * map->stripe_len; + ret = 0; + for (i = start_stripe; i < nstripes; ++i) { + /* + * canceled? + */ + if (atomic_read(&fs_info->scrub_cancel_req) || + atomic_read(&sdev->cancel_req)) { + ret = -ECANCELED; + goto out; + } + /* + * check to see if we have to pause + */ + if (atomic_read(&fs_info->scrub_pause_req)) { + /* push queued extents */ + scrub_submit(sdev); + wait_event(sdev->list_wait, + atomic_read(&sdev->in_flight) == 0); + atomic_inc(&fs_info->scrubs_paused); + wake_up(&fs_info->scrub_pause_wait); + mutex_lock(&fs_info->scrub_lock); + while (atomic_read(&fs_info->scrub_pause_req)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrub_pause_req) == 0); + mutex_lock(&fs_info->scrub_lock); + } + atomic_dec(&fs_info->scrubs_paused); + mutex_unlock(&fs_info->scrub_lock); + wake_up(&fs_info->scrub_pause_wait); + scrub_free_csums(sdev); + start_stripe = i; + goto again; + } + + key.objectid = logical; + key.type = BTRFS_EXTENT_ITEM_KEY; + key.offset = (u64)0; + + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + goto out; + if (ret > 0) { + ret = btrfs_previous_item(root, path, 0, + BTRFS_EXTENT_ITEM_KEY); + if (ret < 0) + goto out; + if (ret > 0) { + /* there's no smaller item, so stick with the + * larger one */ + btrfs_release_path(path); + ret = btrfs_search_slot(NULL, root, &key, + path, 0, 0); + if (ret < 0) + goto out; + } + } + + while (1) { + l = path->nodes[0]; + slot = path->slots[0]; + if (slot >= btrfs_header_nritems(l)) { + ret = btrfs_next_leaf(root, path); + if (ret == 0) + continue; + if (ret < 0) + goto out; + + break; + } + btrfs_item_key_to_cpu(l, &key, slot); + + if (key.objectid + key.offset <= logical) + goto next; + + if (key.objectid >= logical + map->stripe_len) + break; + + if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY) + goto next; + + extent = btrfs_item_ptr(l, slot, + struct btrfs_extent_item); + flags = btrfs_extent_flags(l, extent); + generation = btrfs_extent_generation(l, extent); + + if (key.objectid < logical && + (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) { + printk(KERN_ERR + "btrfs scrub: tree block %llu spanning " + "stripes, ignored. logical=%llu\n", + (unsigned long long)key.objectid, + (unsigned long long)logical); + goto next; + } + + /* + * trim extent to this stripe + */ + if (key.objectid < logical) { + key.offset -= logical - key.objectid; + key.objectid = logical; + } + if (key.objectid + key.offset > + logical + map->stripe_len) { + key.offset = logical + map->stripe_len - + key.objectid; + } + + ret = scrub_extent(sdev, key.objectid, key.offset, + key.objectid - logical + physical, + flags, generation, mirror_num); + if (ret) + goto out; + +next: + path->slots[0]++; + } + btrfs_release_path(path); + logical += increment; + physical += map->stripe_len; + spin_lock(&sdev->stat_lock); + sdev->stat.last_physical = physical; + spin_unlock(&sdev->stat_lock); + } + /* push queued extents */ + scrub_submit(sdev); + +out: + blk_finish_plug(&plug); +out_noplug: + btrfs_free_path(path); + return ret < 0 ? ret : 0; +} + +static noinline_for_stack int scrub_chunk(struct scrub_dev *sdev, + u64 chunk_tree, u64 chunk_objectid, u64 chunk_offset, u64 length) +{ + struct btrfs_mapping_tree *map_tree = + &sdev->dev->dev_root->fs_info->mapping_tree; + struct map_lookup *map; + struct extent_map *em; + int i; + int ret = -EINVAL; + + read_lock(&map_tree->map_tree.lock); + em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1); + read_unlock(&map_tree->map_tree.lock); + + if (!em) + return -EINVAL; + + map = (struct map_lookup *)em->bdev; + if (em->start != chunk_offset) + goto out; + + if (em->len < length) + goto out; + + for (i = 0; i < map->num_stripes; ++i) { + if (map->stripes[i].dev == sdev->dev) { + ret = scrub_stripe(sdev, map, i, chunk_offset, length); + if (ret) + goto out; + } + } +out: + free_extent_map(em); + + return ret; +} + +static noinline_for_stack +int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end) +{ + struct btrfs_dev_extent *dev_extent = NULL; + struct btrfs_path *path; + struct btrfs_root *root = sdev->dev->dev_root; + struct btrfs_fs_info *fs_info = root->fs_info; + u64 length; + u64 chunk_tree; + u64 chunk_objectid; + u64 chunk_offset; + int ret; + int slot; + struct extent_buffer *l; + struct btrfs_key key; + struct btrfs_key found_key; + struct btrfs_block_group_cache *cache; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + path->reada = 2; + path->search_commit_root = 1; + path->skip_locking = 1; + + key.objectid = sdev->dev->devid; + key.offset = 0ull; + key.type = BTRFS_DEV_EXTENT_KEY; + + + while (1) { + ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); + if (ret < 0) + break; + if (ret > 0) { + if (path->slots[0] >= + btrfs_header_nritems(path->nodes[0])) { + ret = btrfs_next_leaf(root, path); + if (ret) + break; + } + } + + l = path->nodes[0]; + slot = path->slots[0]; + + btrfs_item_key_to_cpu(l, &found_key, slot); + + if (found_key.objectid != sdev->dev->devid) + break; + + if (btrfs_key_type(&found_key) != BTRFS_DEV_EXTENT_KEY) + break; + + if (found_key.offset >= end) + break; + + if (found_key.offset < key.offset) + break; + + dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent); + length = btrfs_dev_extent_length(l, dev_extent); + + if (found_key.offset + length <= start) { + key.offset = found_key.offset + length; + btrfs_release_path(path); + continue; + } + + chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent); + chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent); + chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent); + + /* + * get a reference on the corresponding block group to prevent + * the chunk from going away while we scrub it + */ + cache = btrfs_lookup_block_group(fs_info, chunk_offset); + if (!cache) { + ret = -ENOENT; + break; + } + ret = scrub_chunk(sdev, chunk_tree, chunk_objectid, + chunk_offset, length); + btrfs_put_block_group(cache); + if (ret) + break; + + key.offset = found_key.offset + length; + btrfs_release_path(path); + } + + btrfs_free_path(path); + + /* + * ret can still be 1 from search_slot or next_leaf, + * that's not an error + */ + return ret < 0 ? ret : 0; +} + +static noinline_for_stack int scrub_supers(struct scrub_dev *sdev) +{ + int i; + u64 bytenr; + u64 gen; + int ret; + struct btrfs_device *device = sdev->dev; + struct btrfs_root *root = device->dev_root; + + gen = root->fs_info->last_trans_committed; + + for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) { + bytenr = btrfs_sb_offset(i); + if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes) + break; + + ret = scrub_page(sdev, bytenr, PAGE_SIZE, bytenr, + BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1); + if (ret) + return ret; + } + wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0); + + return 0; +} + +/* + * get a reference count on fs_info->scrub_workers. start worker if necessary + */ +static noinline_for_stack int scrub_workers_get(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + if (fs_info->scrub_workers_refcnt == 0) { + btrfs_init_workers(&fs_info->scrub_workers, "scrub", + fs_info->thread_pool_size, &fs_info->generic_worker); + fs_info->scrub_workers.idle_thresh = 4; + btrfs_start_workers(&fs_info->scrub_workers, 1); + } + ++fs_info->scrub_workers_refcnt; + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} + +static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + if (--fs_info->scrub_workers_refcnt == 0) + btrfs_stop_workers(&fs_info->scrub_workers); + WARN_ON(fs_info->scrub_workers_refcnt < 0); + mutex_unlock(&fs_info->scrub_lock); +} + + +int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, + struct btrfs_scrub_progress *progress, int readonly) +{ + struct scrub_dev *sdev; + struct btrfs_fs_info *fs_info = root->fs_info; + int ret; + struct btrfs_device *dev; + + if (btrfs_fs_closing(root->fs_info)) + return -EINVAL; + + /* + * check some assumptions + */ + if (root->sectorsize != PAGE_SIZE || + root->sectorsize != root->leafsize || + root->sectorsize != root->nodesize) { + printk(KERN_ERR "btrfs_scrub: size assumptions fail\n"); + return -EINVAL; + } + + ret = scrub_workers_get(root); + if (ret) + return ret; + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(root, devid, NULL, NULL); + if (!dev || dev->missing) { + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return -ENODEV; + } + mutex_lock(&fs_info->scrub_lock); + + if (!dev->in_fs_metadata) { + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return -ENODEV; + } + + if (dev->scrub_device) { + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return -EINPROGRESS; + } + sdev = scrub_setup_dev(dev); + if (IS_ERR(sdev)) { + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + scrub_workers_put(root); + return PTR_ERR(sdev); + } + sdev->readonly = readonly; + dev->scrub_device = sdev; + + atomic_inc(&fs_info->scrubs_running); + mutex_unlock(&fs_info->scrub_lock); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + + down_read(&fs_info->scrub_super_lock); + ret = scrub_supers(sdev); + up_read(&fs_info->scrub_super_lock); + + if (!ret) + ret = scrub_enumerate_chunks(sdev, start, end); + + wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0); + + atomic_dec(&fs_info->scrubs_running); + wake_up(&fs_info->scrub_pause_wait); + + if (progress) + memcpy(progress, &sdev->stat, sizeof(*progress)); + + mutex_lock(&fs_info->scrub_lock); + dev->scrub_device = NULL; + mutex_unlock(&fs_info->scrub_lock); + + scrub_free_dev(sdev); + scrub_workers_put(root); + + return ret; +} + +int btrfs_scrub_pause(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + atomic_inc(&fs_info->scrub_pause_req); + while (atomic_read(&fs_info->scrubs_paused) != + atomic_read(&fs_info->scrubs_running)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrubs_paused) == + atomic_read(&fs_info->scrubs_running)); + mutex_lock(&fs_info->scrub_lock); + } + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} + +int btrfs_scrub_continue(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + atomic_dec(&fs_info->scrub_pause_req); + wake_up(&fs_info->scrub_pause_wait); + return 0; +} + +int btrfs_scrub_pause_super(struct btrfs_root *root) +{ + down_write(&root->fs_info->scrub_super_lock); + return 0; +} + +int btrfs_scrub_continue_super(struct btrfs_root *root) +{ + up_write(&root->fs_info->scrub_super_lock); + return 0; +} + +int btrfs_scrub_cancel(struct btrfs_root *root) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + + mutex_lock(&fs_info->scrub_lock); + if (!atomic_read(&fs_info->scrubs_running)) { + mutex_unlock(&fs_info->scrub_lock); + return -ENOTCONN; + } + + atomic_inc(&fs_info->scrub_cancel_req); + while (atomic_read(&fs_info->scrubs_running)) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + atomic_read(&fs_info->scrubs_running) == 0); + mutex_lock(&fs_info->scrub_lock); + } + atomic_dec(&fs_info->scrub_cancel_req); + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} + +int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct scrub_dev *sdev; + + mutex_lock(&fs_info->scrub_lock); + sdev = dev->scrub_device; + if (!sdev) { + mutex_unlock(&fs_info->scrub_lock); + return -ENOTCONN; + } + atomic_inc(&sdev->cancel_req); + while (dev->scrub_device) { + mutex_unlock(&fs_info->scrub_lock); + wait_event(fs_info->scrub_pause_wait, + dev->scrub_device == NULL); + mutex_lock(&fs_info->scrub_lock); + } + mutex_unlock(&fs_info->scrub_lock); + + return 0; +} +int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid) +{ + struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_device *dev; + int ret; + + /* + * we have to hold the device_list_mutex here so the device + * does not go away in cancel_dev. FIXME: find a better solution + */ + mutex_lock(&fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(root, devid, NULL, NULL); + if (!dev) { + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + return -ENODEV; + } + ret = btrfs_scrub_cancel_dev(root, dev); + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + + return ret; +} + +int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, + struct btrfs_scrub_progress *progress) +{ + struct btrfs_device *dev; + struct scrub_dev *sdev = NULL; + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(root, devid, NULL, NULL); + if (dev) + sdev = dev->scrub_device; + if (sdev) + memcpy(progress, &sdev->stat, sizeof(*progress)); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + + return dev ? (sdev ? 0 : -ENOTCONN) : -ENODEV; +} diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index be4ffa12f3ef..0bb4ebbb71b7 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -41,6 +41,7 @@ #include <linux/slab.h> #include <linux/cleancache.h> #include "compat.h" +#include "delayed-inode.h" #include "ctree.h" #include "disk-io.h" #include "transaction.h" @@ -160,7 +161,8 @@ enum { Opt_compress_type, Opt_compress_force, Opt_compress_force_type, Opt_notreelog, Opt_ratio, Opt_flushoncommit, Opt_discard, Opt_space_cache, Opt_clear_cache, Opt_user_subvol_rm_allowed, - Opt_enospc_debug, Opt_subvolrootid, Opt_err, + Opt_enospc_debug, Opt_subvolrootid, Opt_defrag, + Opt_inode_cache, Opt_err, }; static match_table_t tokens = { @@ -191,6 +193,8 @@ static match_table_t tokens = { {Opt_user_subvol_rm_allowed, "user_subvol_rm_allowed"}, {Opt_enospc_debug, "enospc_debug"}, {Opt_subvolrootid, "subvolrootid=%d"}, + {Opt_defrag, "autodefrag"}, + {Opt_inode_cache, "inode_cache"}, {Opt_err, NULL}, }; @@ -359,6 +363,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) printk(KERN_INFO "btrfs: enabling disk space caching\n"); btrfs_set_opt(info->mount_opt, SPACE_CACHE); break; + case Opt_inode_cache: + printk(KERN_INFO "btrfs: enabling inode map caching\n"); + btrfs_set_opt(info->mount_opt, INODE_MAP_CACHE); + break; case Opt_clear_cache: printk(KERN_INFO "btrfs: force clearing of disk cache\n"); btrfs_set_opt(info->mount_opt, CLEAR_CACHE); @@ -369,6 +377,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) case Opt_enospc_debug: btrfs_set_opt(info->mount_opt, ENOSPC_DEBUG); break; + case Opt_defrag: + printk(KERN_INFO "btrfs: enabling auto defrag"); + btrfs_set_opt(info->mount_opt, AUTO_DEFRAG); + break; case Opt_err: printk(KERN_INFO "btrfs: unrecognized mount option " "'%s'\n", p); @@ -507,8 +519,10 @@ static struct dentry *get_default_root(struct super_block *sb, */ dir_id = btrfs_super_root_dir(&root->fs_info->super_copy); di = btrfs_lookup_dir_item(NULL, root, path, dir_id, "default", 7, 0); - if (IS_ERR(di)) + if (IS_ERR(di)) { + btrfs_free_path(path); return ERR_CAST(di); + } if (!di) { /* * Ok the default dir item isn't there. This is weird since @@ -741,7 +755,7 @@ static int btrfs_set_super(struct super_block *s, void *data) * for multiple device setup. Make sure to keep it in sync. */ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, - const char *dev_name, void *data) + const char *device_name, void *data) { struct block_device *bdev = NULL; struct super_block *s; @@ -764,7 +778,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, if (error) return ERR_PTR(error); - error = btrfs_scan_one_device(dev_name, mode, fs_type, &fs_devices); + error = btrfs_scan_one_device(device_name, mode, fs_type, &fs_devices); if (error) goto error_free_subvol_name; @@ -811,7 +825,7 @@ static struct dentry *btrfs_mount(struct file_system_type *fs_type, int flags, } else { char b[BDEVNAME_SIZE]; - s->s_flags = flags; + s->s_flags = flags | MS_NOSEC; strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id)); error = btrfs_fill_super(s, fs_devices, data, flags & MS_SILENT ? 1 : 0); @@ -915,6 +929,32 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) return 0; } +/* Used to sort the devices by max_avail(descending sort) */ +static int btrfs_cmp_device_free_bytes(const void *dev_info1, + const void *dev_info2) +{ + if (((struct btrfs_device_info *)dev_info1)->max_avail > + ((struct btrfs_device_info *)dev_info2)->max_avail) + return -1; + else if (((struct btrfs_device_info *)dev_info1)->max_avail < + ((struct btrfs_device_info *)dev_info2)->max_avail) + return 1; + else + return 0; +} + +/* + * sort the devices by max_avail, in which max free extent size of each device + * is stored.(Descending Sort) + */ +static inline void btrfs_descending_sort_devices( + struct btrfs_device_info *devices, + size_t nr_devices) +{ + sort(devices, nr_devices, sizeof(struct btrfs_device_info), + btrfs_cmp_device_free_bytes, NULL); +} + /* * The helper to calc the free space on the devices that can be used to store * file data. @@ -1208,10 +1248,14 @@ static int __init init_btrfs_fs(void) if (err) goto free_extent_io; - err = btrfs_interface_init(); + err = btrfs_delayed_inode_init(); if (err) goto free_extent_map; + err = btrfs_interface_init(); + if (err) + goto free_delayed_inode; + err = register_filesystem(&btrfs_fs_type); if (err) goto unregister_ioctl; @@ -1221,6 +1265,8 @@ static int __init init_btrfs_fs(void) unregister_ioctl: btrfs_interface_exit(); +free_delayed_inode: + btrfs_delayed_inode_exit(); free_extent_map: extent_map_exit(); free_extent_io: @@ -1237,6 +1283,7 @@ free_sysfs: static void __exit exit_btrfs_fs(void) { btrfs_destroy_cachep(); + btrfs_delayed_inode_exit(); extent_map_exit(); extent_io_exit(); btrfs_interface_exit(); diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index 4ce16ef702a3..daac9ae6d731 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -28,232 +28,9 @@ #include "disk-io.h" #include "transaction.h" -static ssize_t root_blocks_used_show(struct btrfs_root *root, char *buf) -{ - return snprintf(buf, PAGE_SIZE, "%llu\n", - (unsigned long long)btrfs_root_used(&root->root_item)); -} - -static ssize_t root_block_limit_show(struct btrfs_root *root, char *buf) -{ - return snprintf(buf, PAGE_SIZE, "%llu\n", - (unsigned long long)btrfs_root_limit(&root->root_item)); -} - -static ssize_t super_blocks_used_show(struct btrfs_fs_info *fs, char *buf) -{ - - return snprintf(buf, PAGE_SIZE, "%llu\n", - (unsigned long long)btrfs_super_bytes_used(&fs->super_copy)); -} - -static ssize_t super_total_blocks_show(struct btrfs_fs_info *fs, char *buf) -{ - return snprintf(buf, PAGE_SIZE, "%llu\n", - (unsigned long long)btrfs_super_total_bytes(&fs->super_copy)); -} - -static ssize_t super_blocksize_show(struct btrfs_fs_info *fs, char *buf) -{ - return snprintf(buf, PAGE_SIZE, "%llu\n", - (unsigned long long)btrfs_super_sectorsize(&fs->super_copy)); -} - -/* this is for root attrs (subvols/snapshots) */ -struct btrfs_root_attr { - struct attribute attr; - ssize_t (*show)(struct btrfs_root *, char *); - ssize_t (*store)(struct btrfs_root *, const char *, size_t); -}; - -#define ROOT_ATTR(name, mode, show, store) \ -static struct btrfs_root_attr btrfs_root_attr_##name = __ATTR(name, mode, \ - show, store) - -ROOT_ATTR(blocks_used, 0444, root_blocks_used_show, NULL); -ROOT_ATTR(block_limit, 0644, root_block_limit_show, NULL); - -static struct attribute *btrfs_root_attrs[] = { - &btrfs_root_attr_blocks_used.attr, - &btrfs_root_attr_block_limit.attr, - NULL, -}; - -/* this is for super attrs (actual full fs) */ -struct btrfs_super_attr { - struct attribute attr; - ssize_t (*show)(struct btrfs_fs_info *, char *); - ssize_t (*store)(struct btrfs_fs_info *, const char *, size_t); -}; - -#define SUPER_ATTR(name, mode, show, store) \ -static struct btrfs_super_attr btrfs_super_attr_##name = __ATTR(name, mode, \ - show, store) - -SUPER_ATTR(blocks_used, 0444, super_blocks_used_show, NULL); -SUPER_ATTR(total_blocks, 0444, super_total_blocks_show, NULL); -SUPER_ATTR(blocksize, 0444, super_blocksize_show, NULL); - -static struct attribute *btrfs_super_attrs[] = { - &btrfs_super_attr_blocks_used.attr, - &btrfs_super_attr_total_blocks.attr, - &btrfs_super_attr_blocksize.attr, - NULL, -}; - -static ssize_t btrfs_super_attr_show(struct kobject *kobj, - struct attribute *attr, char *buf) -{ - struct btrfs_fs_info *fs = container_of(kobj, struct btrfs_fs_info, - super_kobj); - struct btrfs_super_attr *a = container_of(attr, - struct btrfs_super_attr, - attr); - - return a->show ? a->show(fs, buf) : 0; -} - -static ssize_t btrfs_super_attr_store(struct kobject *kobj, - struct attribute *attr, - const char *buf, size_t len) -{ - struct btrfs_fs_info *fs = container_of(kobj, struct btrfs_fs_info, - super_kobj); - struct btrfs_super_attr *a = container_of(attr, - struct btrfs_super_attr, - attr); - - return a->store ? a->store(fs, buf, len) : 0; -} - -static ssize_t btrfs_root_attr_show(struct kobject *kobj, - struct attribute *attr, char *buf) -{ - struct btrfs_root *root = container_of(kobj, struct btrfs_root, - root_kobj); - struct btrfs_root_attr *a = container_of(attr, - struct btrfs_root_attr, - attr); - - return a->show ? a->show(root, buf) : 0; -} - -static ssize_t btrfs_root_attr_store(struct kobject *kobj, - struct attribute *attr, - const char *buf, size_t len) -{ - struct btrfs_root *root = container_of(kobj, struct btrfs_root, - root_kobj); - struct btrfs_root_attr *a = container_of(attr, - struct btrfs_root_attr, - attr); - return a->store ? a->store(root, buf, len) : 0; -} - -static void btrfs_super_release(struct kobject *kobj) -{ - struct btrfs_fs_info *fs = container_of(kobj, struct btrfs_fs_info, - super_kobj); - complete(&fs->kobj_unregister); -} - -static void btrfs_root_release(struct kobject *kobj) -{ - struct btrfs_root *root = container_of(kobj, struct btrfs_root, - root_kobj); - complete(&root->kobj_unregister); -} - -static const struct sysfs_ops btrfs_super_attr_ops = { - .show = btrfs_super_attr_show, - .store = btrfs_super_attr_store, -}; - -static const struct sysfs_ops btrfs_root_attr_ops = { - .show = btrfs_root_attr_show, - .store = btrfs_root_attr_store, -}; - -static struct kobj_type btrfs_root_ktype = { - .default_attrs = btrfs_root_attrs, - .sysfs_ops = &btrfs_root_attr_ops, - .release = btrfs_root_release, -}; - -static struct kobj_type btrfs_super_ktype = { - .default_attrs = btrfs_super_attrs, - .sysfs_ops = &btrfs_super_attr_ops, - .release = btrfs_super_release, -}; - /* /sys/fs/btrfs/ entry */ static struct kset *btrfs_kset; -int btrfs_sysfs_add_super(struct btrfs_fs_info *fs) -{ - int error; - char *name; - char c; - int len = strlen(fs->sb->s_id) + 1; - int i; - - name = kmalloc(len, GFP_NOFS); - if (!name) { - error = -ENOMEM; - goto fail; - } - - for (i = 0; i < len; i++) { - c = fs->sb->s_id[i]; - if (c == '/' || c == '\\') - c = '!'; - name[i] = c; - } - name[len] = '\0'; - - fs->super_kobj.kset = btrfs_kset; - error = kobject_init_and_add(&fs->super_kobj, &btrfs_super_ktype, - NULL, "%s", name); - kfree(name); - if (error) - goto fail; - - return 0; - -fail: - printk(KERN_ERR "btrfs: sysfs creation for super failed\n"); - return error; -} - -int btrfs_sysfs_add_root(struct btrfs_root *root) -{ - int error; - - error = kobject_init_and_add(&root->root_kobj, &btrfs_root_ktype, - &root->fs_info->super_kobj, - "%s", root->name); - if (error) - goto fail; - - return 0; - -fail: - printk(KERN_ERR "btrfs: sysfs creation for root failed\n"); - return error; -} - -void btrfs_sysfs_del_root(struct btrfs_root *root) -{ - kobject_put(&root->root_kobj); - wait_for_completion(&root->kobj_unregister); -} - -void btrfs_sysfs_del_super(struct btrfs_fs_info *fs) -{ - kobject_put(&fs->super_kobj); - wait_for_completion(&fs->kobj_unregister); -} - int btrfs_init_sysfs(void) { btrfs_kset = kset_create_and_add("btrfs", NULL, fs_kobj); diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index c571734d5e5a..51dcec86757f 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -27,6 +27,7 @@ #include "transaction.h" #include "locking.h" #include "tree-log.h" +#include "inode-map.h" #define BTRFS_ROOT_TRANS_TAG 0 @@ -34,6 +35,7 @@ static noinline void put_transaction(struct btrfs_transaction *transaction) { WARN_ON(atomic_read(&transaction->use_count) == 0); if (atomic_dec_and_test(&transaction->use_count)) { + BUG_ON(!list_empty(&transaction->list)); memset(transaction, 0, sizeof(*transaction)); kmem_cache_free(btrfs_transaction_cachep, transaction); } @@ -48,47 +50,72 @@ static noinline void switch_commit_root(struct btrfs_root *root) /* * either allocate a new transaction or hop into the existing one */ -static noinline int join_transaction(struct btrfs_root *root) +static noinline int join_transaction(struct btrfs_root *root, int nofail) { struct btrfs_transaction *cur_trans; + + spin_lock(&root->fs_info->trans_lock); + if (root->fs_info->trans_no_join) { + if (!nofail) { + spin_unlock(&root->fs_info->trans_lock); + return -EBUSY; + } + } + cur_trans = root->fs_info->running_transaction; - if (!cur_trans) { - cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, - GFP_NOFS); - if (!cur_trans) - return -ENOMEM; - root->fs_info->generation++; - atomic_set(&cur_trans->num_writers, 1); - cur_trans->num_joined = 0; - cur_trans->transid = root->fs_info->generation; - init_waitqueue_head(&cur_trans->writer_wait); - init_waitqueue_head(&cur_trans->commit_wait); - cur_trans->in_commit = 0; - cur_trans->blocked = 0; - atomic_set(&cur_trans->use_count, 1); - cur_trans->commit_done = 0; - cur_trans->start_time = get_seconds(); - - cur_trans->delayed_refs.root = RB_ROOT; - cur_trans->delayed_refs.num_entries = 0; - cur_trans->delayed_refs.num_heads_ready = 0; - cur_trans->delayed_refs.num_heads = 0; - cur_trans->delayed_refs.flushing = 0; - cur_trans->delayed_refs.run_delayed_start = 0; - spin_lock_init(&cur_trans->delayed_refs.lock); - - INIT_LIST_HEAD(&cur_trans->pending_snapshots); - list_add_tail(&cur_trans->list, &root->fs_info->trans_list); - extent_io_tree_init(&cur_trans->dirty_pages, - root->fs_info->btree_inode->i_mapping, - GFP_NOFS); - spin_lock(&root->fs_info->new_trans_lock); - root->fs_info->running_transaction = cur_trans; - spin_unlock(&root->fs_info->new_trans_lock); - } else { + if (cur_trans) { + atomic_inc(&cur_trans->use_count); atomic_inc(&cur_trans->num_writers); cur_trans->num_joined++; + spin_unlock(&root->fs_info->trans_lock); + return 0; } + spin_unlock(&root->fs_info->trans_lock); + + cur_trans = kmem_cache_alloc(btrfs_transaction_cachep, GFP_NOFS); + if (!cur_trans) + return -ENOMEM; + spin_lock(&root->fs_info->trans_lock); + if (root->fs_info->running_transaction) { + kmem_cache_free(btrfs_transaction_cachep, cur_trans); + cur_trans = root->fs_info->running_transaction; + atomic_inc(&cur_trans->use_count); + atomic_inc(&cur_trans->num_writers); + cur_trans->num_joined++; + spin_unlock(&root->fs_info->trans_lock); + return 0; + } + atomic_set(&cur_trans->num_writers, 1); + cur_trans->num_joined = 0; + init_waitqueue_head(&cur_trans->writer_wait); + init_waitqueue_head(&cur_trans->commit_wait); + cur_trans->in_commit = 0; + cur_trans->blocked = 0; + /* + * One for this trans handle, one so it will live on until we + * commit the transaction. + */ + atomic_set(&cur_trans->use_count, 2); + cur_trans->commit_done = 0; + cur_trans->start_time = get_seconds(); + + cur_trans->delayed_refs.root = RB_ROOT; + cur_trans->delayed_refs.num_entries = 0; + cur_trans->delayed_refs.num_heads_ready = 0; + cur_trans->delayed_refs.num_heads = 0; + cur_trans->delayed_refs.flushing = 0; + cur_trans->delayed_refs.run_delayed_start = 0; + spin_lock_init(&cur_trans->commit_lock); + spin_lock_init(&cur_trans->delayed_refs.lock); + + INIT_LIST_HEAD(&cur_trans->pending_snapshots); + list_add_tail(&cur_trans->list, &root->fs_info->trans_list); + extent_io_tree_init(&cur_trans->dirty_pages, + root->fs_info->btree_inode->i_mapping); + root->fs_info->generation++; + cur_trans->transid = root->fs_info->generation; + root->fs_info->running_transaction = cur_trans; + spin_unlock(&root->fs_info->trans_lock); return 0; } @@ -99,36 +126,82 @@ static noinline int join_transaction(struct btrfs_root *root) * to make sure the old root from before we joined the transaction is deleted * when the transaction commits */ -static noinline int record_root_in_trans(struct btrfs_trans_handle *trans, - struct btrfs_root *root) +static int record_root_in_trans(struct btrfs_trans_handle *trans, + struct btrfs_root *root) { if (root->ref_cows && root->last_trans < trans->transid) { WARN_ON(root == root->fs_info->extent_root); WARN_ON(root->commit_root != root->node); + /* + * see below for in_trans_setup usage rules + * we have the reloc mutex held now, so there + * is only one writer in this function + */ + root->in_trans_setup = 1; + + /* make sure readers find in_trans_setup before + * they find our root->last_trans update + */ + smp_wmb(); + + spin_lock(&root->fs_info->fs_roots_radix_lock); + if (root->last_trans == trans->transid) { + spin_unlock(&root->fs_info->fs_roots_radix_lock); + return 0; + } radix_tree_tag_set(&root->fs_info->fs_roots_radix, (unsigned long)root->root_key.objectid, BTRFS_ROOT_TRANS_TAG); + spin_unlock(&root->fs_info->fs_roots_radix_lock); root->last_trans = trans->transid; + + /* this is pretty tricky. We don't want to + * take the relocation lock in btrfs_record_root_in_trans + * unless we're really doing the first setup for this root in + * this transaction. + * + * Normally we'd use root->last_trans as a flag to decide + * if we want to take the expensive mutex. + * + * But, we have to set root->last_trans before we + * init the relocation root, otherwise, we trip over warnings + * in ctree.c. The solution used here is to flag ourselves + * with root->in_trans_setup. When this is 1, we're still + * fixing up the reloc trees and everyone must wait. + * + * When this is zero, they can trust root->last_trans and fly + * through btrfs_record_root_in_trans without having to take the + * lock. smp_wmb() makes sure that all the writes above are + * done before we pop in the zero below + */ btrfs_init_reloc_root(trans, root); + smp_wmb(); + root->in_trans_setup = 0; } return 0; } + int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans, struct btrfs_root *root) { if (!root->ref_cows) return 0; - mutex_lock(&root->fs_info->trans_mutex); - if (root->last_trans == trans->transid) { - mutex_unlock(&root->fs_info->trans_mutex); + /* + * see record_root_in_trans for comments about in_trans_setup usage + * and barriers + */ + smp_rmb(); + if (root->last_trans == trans->transid && + !root->in_trans_setup) return 0; - } + mutex_lock(&root->fs_info->reloc_mutex); record_root_in_trans(trans, root); - mutex_unlock(&root->fs_info->trans_mutex); + mutex_unlock(&root->fs_info->reloc_mutex); + return 0; } @@ -140,21 +213,23 @@ static void wait_current_trans(struct btrfs_root *root) { struct btrfs_transaction *cur_trans; + spin_lock(&root->fs_info->trans_lock); cur_trans = root->fs_info->running_transaction; if (cur_trans && cur_trans->blocked) { DEFINE_WAIT(wait); atomic_inc(&cur_trans->use_count); + spin_unlock(&root->fs_info->trans_lock); while (1) { prepare_to_wait(&root->fs_info->transaction_wait, &wait, TASK_UNINTERRUPTIBLE); if (!cur_trans->blocked) break; - mutex_unlock(&root->fs_info->trans_mutex); schedule(); - mutex_lock(&root->fs_info->trans_mutex); } finish_wait(&root->fs_info->transaction_wait, &wait); put_transaction(cur_trans); + } else { + spin_unlock(&root->fs_info->trans_lock); } } @@ -167,10 +242,16 @@ enum btrfs_trans_type { static int may_wait_transaction(struct btrfs_root *root, int type) { - if (!root->fs_info->log_root_recovering && - ((type == TRANS_START && !root->fs_info->open_ioctl_trans) || - type == TRANS_USERSPACE)) + if (root->fs_info->log_root_recovering) + return 0; + + if (type == TRANS_USERSPACE) + return 1; + + if (type == TRANS_START && + !atomic_read(&root->fs_info->open_ioctl_trans)) return 1; + return 0; } @@ -184,36 +265,44 @@ static struct btrfs_trans_handle *start_transaction(struct btrfs_root *root, if (root->fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) return ERR_PTR(-EROFS); + + if (current->journal_info) { + WARN_ON(type != TRANS_JOIN && type != TRANS_JOIN_NOLOCK); + h = current->journal_info; + h->use_count++; + h->orig_rsv = h->block_rsv; + h->block_rsv = NULL; + goto got_it; + } again: h = kmem_cache_alloc(btrfs_trans_handle_cachep, GFP_NOFS); if (!h) return ERR_PTR(-ENOMEM); - if (type != TRANS_JOIN_NOLOCK) - mutex_lock(&root->fs_info->trans_mutex); if (may_wait_transaction(root, type)) wait_current_trans(root); - ret = join_transaction(root); + do { + ret = join_transaction(root, type == TRANS_JOIN_NOLOCK); + if (ret == -EBUSY) + wait_current_trans(root); + } while (ret == -EBUSY); + if (ret < 0) { kmem_cache_free(btrfs_trans_handle_cachep, h); - if (type != TRANS_JOIN_NOLOCK) - mutex_unlock(&root->fs_info->trans_mutex); return ERR_PTR(ret); } cur_trans = root->fs_info->running_transaction; - atomic_inc(&cur_trans->use_count); - if (type != TRANS_JOIN_NOLOCK) - mutex_unlock(&root->fs_info->trans_mutex); h->transid = cur_trans->transid; h->transaction = cur_trans; h->blocks_used = 0; - h->block_group = 0; h->bytes_reserved = 0; h->delayed_ref_updates = 0; + h->use_count = 1; h->block_rsv = NULL; + h->orig_rsv = NULL; smp_mb(); if (cur_trans->blocked && may_wait_transaction(root, type)) { @@ -241,11 +330,8 @@ again: } } - if (type != TRANS_JOIN_NOLOCK) - mutex_lock(&root->fs_info->trans_mutex); - record_root_in_trans(h, root); - if (type != TRANS_JOIN_NOLOCK) - mutex_unlock(&root->fs_info->trans_mutex); +got_it: + btrfs_record_root_in_trans(h, root); if (!current->journal_info && type != TRANS_USERSPACE) current->journal_info = h; @@ -257,22 +343,19 @@ struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root, { return start_transaction(root, num_items, TRANS_START); } -struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root, - int num_blocks) +struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root) { return start_transaction(root, 0, TRANS_JOIN); } -struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root, - int num_blocks) +struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root) { return start_transaction(root, 0, TRANS_JOIN_NOLOCK); } -struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r, - int num_blocks) +struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root) { - return start_transaction(r, 0, TRANS_USERSPACE); + return start_transaction(root, 0, TRANS_USERSPACE); } /* wait for a transaction commit to be fully complete */ @@ -280,17 +363,13 @@ static noinline int wait_for_commit(struct btrfs_root *root, struct btrfs_transaction *commit) { DEFINE_WAIT(wait); - mutex_lock(&root->fs_info->trans_mutex); while (!commit->commit_done) { prepare_to_wait(&commit->commit_wait, &wait, TASK_UNINTERRUPTIBLE); if (commit->commit_done) break; - mutex_unlock(&root->fs_info->trans_mutex); schedule(); - mutex_lock(&root->fs_info->trans_mutex); } - mutex_unlock(&root->fs_info->trans_mutex); finish_wait(&commit->commit_wait, &wait); return 0; } @@ -300,102 +379,56 @@ int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid) struct btrfs_transaction *cur_trans = NULL, *t; int ret; - mutex_lock(&root->fs_info->trans_mutex); - ret = 0; if (transid) { if (transid <= root->fs_info->last_trans_committed) - goto out_unlock; + goto out; /* find specified transaction */ + spin_lock(&root->fs_info->trans_lock); list_for_each_entry(t, &root->fs_info->trans_list, list) { if (t->transid == transid) { cur_trans = t; + atomic_inc(&cur_trans->use_count); break; } if (t->transid > transid) break; } + spin_unlock(&root->fs_info->trans_lock); ret = -EINVAL; if (!cur_trans) - goto out_unlock; /* bad transid */ + goto out; /* bad transid */ } else { /* find newest transaction that is committing | committed */ + spin_lock(&root->fs_info->trans_lock); list_for_each_entry_reverse(t, &root->fs_info->trans_list, list) { if (t->in_commit) { if (t->commit_done) - goto out_unlock; + break; cur_trans = t; + atomic_inc(&cur_trans->use_count); break; } } + spin_unlock(&root->fs_info->trans_lock); if (!cur_trans) - goto out_unlock; /* nothing committing|committed */ + goto out; /* nothing committing|committed */ } - atomic_inc(&cur_trans->use_count); - mutex_unlock(&root->fs_info->trans_mutex); - wait_for_commit(root, cur_trans); - mutex_lock(&root->fs_info->trans_mutex); put_transaction(cur_trans); ret = 0; -out_unlock: - mutex_unlock(&root->fs_info->trans_mutex); +out: return ret; } -#if 0 -/* - * rate limit against the drop_snapshot code. This helps to slow down new - * operations if the drop_snapshot code isn't able to keep up. - */ -static void throttle_on_drops(struct btrfs_root *root) -{ - struct btrfs_fs_info *info = root->fs_info; - int harder_count = 0; - -harder: - if (atomic_read(&info->throttles)) { - DEFINE_WAIT(wait); - int thr; - thr = atomic_read(&info->throttle_gen); - - do { - prepare_to_wait(&info->transaction_throttle, - &wait, TASK_UNINTERRUPTIBLE); - if (!atomic_read(&info->throttles)) { - finish_wait(&info->transaction_throttle, &wait); - break; - } - schedule(); - finish_wait(&info->transaction_throttle, &wait); - } while (thr == atomic_read(&info->throttle_gen)); - harder_count++; - - if (root->fs_info->total_ref_cache_size > 1 * 1024 * 1024 && - harder_count < 2) - goto harder; - - if (root->fs_info->total_ref_cache_size > 5 * 1024 * 1024 && - harder_count < 10) - goto harder; - - if (root->fs_info->total_ref_cache_size > 10 * 1024 * 1024 && - harder_count < 20) - goto harder; - } -} -#endif - void btrfs_throttle(struct btrfs_root *root) { - mutex_lock(&root->fs_info->trans_mutex); - if (!root->fs_info->open_ioctl_trans) + if (!atomic_read(&root->fs_info->open_ioctl_trans)) wait_current_trans(root); - mutex_unlock(&root->fs_info->trans_mutex); } static int should_end_transaction(struct btrfs_trans_handle *trans, @@ -413,6 +446,7 @@ int btrfs_should_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_transaction *cur_trans = trans->transaction; int updates; + smp_mb(); if (cur_trans->blocked || cur_trans->delayed_refs.flushing) return 1; @@ -431,6 +465,11 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_fs_info *info = root->fs_info; int count = 0; + if (--trans->use_count) { + trans->block_rsv = trans->orig_rsv; + return 0; + } + while (count < 4) { unsigned long cur = trans->delayed_ref_updates; trans->delayed_ref_updates = 0; @@ -453,9 +492,11 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, btrfs_trans_release_metadata(trans, root); - if (lock && !root->fs_info->open_ioctl_trans && - should_end_transaction(trans, root)) + if (lock && !atomic_read(&root->fs_info->open_ioctl_trans) && + should_end_transaction(trans, root)) { trans->transaction->blocked = 1; + smp_wmb(); + } if (lock && cur_trans->blocked && !cur_trans->in_commit) { if (throttle) @@ -487,19 +528,40 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, int btrfs_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - return __btrfs_end_transaction(trans, root, 0, 1); + int ret; + + ret = __btrfs_end_transaction(trans, root, 0, 1); + if (ret) + return ret; + return 0; } int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - return __btrfs_end_transaction(trans, root, 1, 1); + int ret; + + ret = __btrfs_end_transaction(trans, root, 1, 1); + if (ret) + return ret; + return 0; } int btrfs_end_transaction_nolock(struct btrfs_trans_handle *trans, struct btrfs_root *root) { - return __btrfs_end_transaction(trans, root, 0, 0); + int ret; + + ret = __btrfs_end_transaction(trans, root, 0, 0); + if (ret) + return ret; + return 0; +} + +int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans, + struct btrfs_root *root) +{ + return __btrfs_end_transaction(trans, root, 1, 1); } /* @@ -725,9 +787,9 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, */ int btrfs_add_dead_root(struct btrfs_root *root) { - mutex_lock(&root->fs_info->trans_mutex); + spin_lock(&root->fs_info->trans_lock); list_add(&root->root_list, &root->fs_info->dead_roots); - mutex_unlock(&root->fs_info->trans_mutex); + spin_unlock(&root->fs_info->trans_lock); return 0; } @@ -743,6 +805,7 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, int ret; int err = 0; + spin_lock(&fs_info->fs_roots_radix_lock); while (1) { ret = radix_tree_gang_lookup_tag(&fs_info->fs_roots_radix, (void **)gang, 0, @@ -755,13 +818,20 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, radix_tree_tag_clear(&fs_info->fs_roots_radix, (unsigned long)root->root_key.objectid, BTRFS_ROOT_TRANS_TAG); + spin_unlock(&fs_info->fs_roots_radix_lock); btrfs_free_log(trans, root); btrfs_update_reloc_root(trans, root); btrfs_orphan_commit_root(trans, root); + btrfs_save_ino_cache(root, trans); + if (root->commit_root != root->node) { + mutex_lock(&root->fs_commit_mutex); switch_commit_root(root); + btrfs_unpin_free_ino(root); + mutex_unlock(&root->fs_commit_mutex); + btrfs_set_root_node(&root->root_item, root->node); } @@ -769,10 +839,12 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, err = btrfs_update_root(trans, fs_info->tree_root, &root->root_key, &root->root_item); + spin_lock(&fs_info->fs_roots_radix_lock); if (err) break; } } + spin_unlock(&fs_info->fs_roots_radix_lock); return err; } @@ -802,104 +874,13 @@ int btrfs_defrag_root(struct btrfs_root *root, int cacheonly) btrfs_btree_balance_dirty(info->tree_root, nr); cond_resched(); - if (root->fs_info->closing || ret != -EAGAIN) + if (btrfs_fs_closing(root->fs_info) || ret != -EAGAIN) break; } root->defrag_running = 0; return ret; } -#if 0 -/* - * when dropping snapshots, we generate a ton of delayed refs, and it makes - * sense not to join the transaction while it is trying to flush the current - * queue of delayed refs out. - * - * This is used by the drop snapshot code only - */ -static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info) -{ - DEFINE_WAIT(wait); - - mutex_lock(&info->trans_mutex); - while (info->running_transaction && - info->running_transaction->delayed_refs.flushing) { - prepare_to_wait(&info->transaction_wait, &wait, - TASK_UNINTERRUPTIBLE); - mutex_unlock(&info->trans_mutex); - - schedule(); - - mutex_lock(&info->trans_mutex); - finish_wait(&info->transaction_wait, &wait); - } - mutex_unlock(&info->trans_mutex); - return 0; -} - -/* - * Given a list of roots that need to be deleted, call btrfs_drop_snapshot on - * all of them - */ -int btrfs_drop_dead_root(struct btrfs_root *root) -{ - struct btrfs_trans_handle *trans; - struct btrfs_root *tree_root = root->fs_info->tree_root; - unsigned long nr; - int ret; - - while (1) { - /* - * we don't want to jump in and create a bunch of - * delayed refs if the transaction is starting to close - */ - wait_transaction_pre_flush(tree_root->fs_info); - trans = btrfs_start_transaction(tree_root, 1); - - /* - * we've joined a transaction, make sure it isn't - * closing right now - */ - if (trans->transaction->delayed_refs.flushing) { - btrfs_end_transaction(trans, tree_root); - continue; - } - - ret = btrfs_drop_snapshot(trans, root); - if (ret != -EAGAIN) - break; - - ret = btrfs_update_root(trans, tree_root, - &root->root_key, - &root->root_item); - if (ret) - break; - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, tree_root); - BUG_ON(ret); - - btrfs_btree_balance_dirty(tree_root, nr); - cond_resched(); - } - BUG_ON(ret); - - ret = btrfs_del_root(trans, tree_root, &root->root_key); - BUG_ON(ret); - - nr = trans->blocks_used; - ret = btrfs_end_transaction(trans, tree_root); - BUG_ON(ret); - - free_extent_buffer(root->node); - free_extent_buffer(root->commit_root); - kfree(root); - - btrfs_btree_balance_dirty(tree_root, nr); - return ret; -} -#endif - /* * new snapshots need to be created at a very specific time in the * transaction commit. This does the actual creation @@ -930,7 +911,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto fail; } - ret = btrfs_find_free_objectid(trans, tree_root, 0, &objectid); + ret = btrfs_find_free_objectid(tree_root, &objectid); if (ret) { pending->error = ret; goto fail; @@ -967,7 +948,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, BUG_ON(ret); ret = btrfs_insert_dir_item(trans, parent_root, dentry->d_name.name, dentry->d_name.len, - parent_inode->i_ino, &key, + parent_inode, &key, BTRFS_FT_DIR, index); BUG_ON(ret); @@ -976,6 +957,15 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, ret = btrfs_update_inode(trans, parent_root, parent_inode); BUG_ON(ret); + /* + * pull in the delayed directory update + * and the delayed inode item + * otherwise we corrupt the FS during + * snapshot + */ + ret = btrfs_run_delayed_items(trans, root); + BUG_ON(ret); + record_root_in_trans(trans, root); btrfs_set_root_last_snapshot(&root->root_item, trans->transid); memcpy(new_root_item, &root->root_item, sizeof(*new_root_item)); @@ -1009,7 +999,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, */ ret = btrfs_add_root_ref(trans, tree_root, objectid, parent_root->root_key.objectid, - parent_inode->i_ino, index, + btrfs_ino(parent_inode), index, dentry->d_name.name, dentry->d_name.len); BUG_ON(ret); dput(parent); @@ -1066,20 +1056,20 @@ static void update_super_roots(struct btrfs_root *root) int btrfs_transaction_in_commit(struct btrfs_fs_info *info) { int ret = 0; - spin_lock(&info->new_trans_lock); + spin_lock(&info->trans_lock); if (info->running_transaction) ret = info->running_transaction->in_commit; - spin_unlock(&info->new_trans_lock); + spin_unlock(&info->trans_lock); return ret; } int btrfs_transaction_blocked(struct btrfs_fs_info *info) { int ret = 0; - spin_lock(&info->new_trans_lock); + spin_lock(&info->trans_lock); if (info->running_transaction) ret = info->running_transaction->blocked; - spin_unlock(&info->new_trans_lock); + spin_unlock(&info->trans_lock); return ret; } @@ -1103,9 +1093,7 @@ static void wait_current_trans_commit_start(struct btrfs_root *root, &wait); break; } - mutex_unlock(&root->fs_info->trans_mutex); schedule(); - mutex_lock(&root->fs_info->trans_mutex); finish_wait(&root->fs_info->transaction_blocked_wait, &wait); } } @@ -1131,9 +1119,7 @@ static void wait_current_trans_commit_start_and_unblock(struct btrfs_root *root, &wait); break; } - mutex_unlock(&root->fs_info->trans_mutex); schedule(); - mutex_lock(&root->fs_info->trans_mutex); finish_wait(&root->fs_info->transaction_wait, &wait); } @@ -1171,7 +1157,7 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, INIT_DELAYED_WORK(&ac->work, do_async_commit); ac->root = root; - ac->newtrans = btrfs_join_transaction(root, 0); + ac->newtrans = btrfs_join_transaction(root); if (IS_ERR(ac->newtrans)) { int err = PTR_ERR(ac->newtrans); kfree(ac); @@ -1179,23 +1165,22 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, } /* take transaction reference */ - mutex_lock(&root->fs_info->trans_mutex); cur_trans = trans->transaction; atomic_inc(&cur_trans->use_count); - mutex_unlock(&root->fs_info->trans_mutex); btrfs_end_transaction(trans, root); schedule_delayed_work(&ac->work, 0); /* wait for transaction to start and unblock */ - mutex_lock(&root->fs_info->trans_mutex); if (wait_for_unblock) wait_current_trans_commit_start_and_unblock(root, cur_trans); else wait_current_trans_commit_start(root, cur_trans); - put_transaction(cur_trans); - mutex_unlock(&root->fs_info->trans_mutex); + if (current->journal_info == trans) + current->journal_info = NULL; + + put_transaction(cur_trans); return 0; } @@ -1238,38 +1223,41 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, ret = btrfs_run_delayed_refs(trans, root, 0); BUG_ON(ret); - mutex_lock(&root->fs_info->trans_mutex); + spin_lock(&cur_trans->commit_lock); if (cur_trans->in_commit) { + spin_unlock(&cur_trans->commit_lock); atomic_inc(&cur_trans->use_count); - mutex_unlock(&root->fs_info->trans_mutex); btrfs_end_transaction(trans, root); ret = wait_for_commit(root, cur_trans); BUG_ON(ret); - mutex_lock(&root->fs_info->trans_mutex); put_transaction(cur_trans); - mutex_unlock(&root->fs_info->trans_mutex); return 0; } trans->transaction->in_commit = 1; trans->transaction->blocked = 1; + spin_unlock(&cur_trans->commit_lock); wake_up(&root->fs_info->transaction_blocked_wait); + spin_lock(&root->fs_info->trans_lock); if (cur_trans->list.prev != &root->fs_info->trans_list) { prev_trans = list_entry(cur_trans->list.prev, struct btrfs_transaction, list); if (!prev_trans->commit_done) { atomic_inc(&prev_trans->use_count); - mutex_unlock(&root->fs_info->trans_mutex); + spin_unlock(&root->fs_info->trans_lock); wait_for_commit(root, prev_trans); - mutex_lock(&root->fs_info->trans_mutex); put_transaction(prev_trans); + } else { + spin_unlock(&root->fs_info->trans_lock); } + } else { + spin_unlock(&root->fs_info->trans_lock); } if (now < cur_trans->start_time || now - cur_trans->start_time < 1) @@ -1277,12 +1265,12 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, do { int snap_pending = 0; + joined = cur_trans->num_joined; if (!list_empty(&trans->transaction->pending_snapshots)) snap_pending = 1; WARN_ON(cur_trans != trans->transaction); - mutex_unlock(&root->fs_info->trans_mutex); if (flush_on_commit || snap_pending) { btrfs_start_delalloc_inodes(root, 1); @@ -1290,6 +1278,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, BUG_ON(ret); } + ret = btrfs_run_delayed_items(trans, root); + BUG_ON(ret); + /* * rename don't use btrfs_join_transaction, so, once we * set the transaction to blocked above, we aren't going @@ -1302,25 +1293,51 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, prepare_to_wait(&cur_trans->writer_wait, &wait, TASK_UNINTERRUPTIBLE); - smp_mb(); if (atomic_read(&cur_trans->num_writers) > 1) schedule_timeout(MAX_SCHEDULE_TIMEOUT); else if (should_grow) schedule_timeout(1); - mutex_lock(&root->fs_info->trans_mutex); finish_wait(&cur_trans->writer_wait, &wait); } while (atomic_read(&cur_trans->num_writers) > 1 || (should_grow && cur_trans->num_joined != joined)); + /* + * Ok now we need to make sure to block out any other joins while we + * commit the transaction. We could have started a join before setting + * no_join so make sure to wait for num_writers to == 1 again. + */ + spin_lock(&root->fs_info->trans_lock); + root->fs_info->trans_no_join = 1; + spin_unlock(&root->fs_info->trans_lock); + wait_event(cur_trans->writer_wait, + atomic_read(&cur_trans->num_writers) == 1); + + /* + * the reloc mutex makes sure that we stop + * the balancing code from coming in and moving + * extents around in the middle of the commit + */ + mutex_lock(&root->fs_info->reloc_mutex); + + ret = btrfs_run_delayed_items(trans, root); + BUG_ON(ret); + ret = create_pending_snapshots(trans, root->fs_info); BUG_ON(ret); ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); BUG_ON(ret); + /* + * make sure none of the code above managed to slip in a + * delayed item + */ + btrfs_assert_delayed_root_empty(root); + WARN_ON(cur_trans != trans->transaction); + btrfs_scrub_pause(root); /* btrfs_commit_tree_roots is responsible for getting the * various roots consistent with each other. Every pointer * in the tree of tree roots has to point to the most up to date @@ -1350,9 +1367,6 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, btrfs_prepare_extent_commit(trans, root); cur_trans = root->fs_info->running_transaction; - spin_lock(&root->fs_info->new_trans_lock); - root->fs_info->running_transaction = NULL; - spin_unlock(&root->fs_info->new_trans_lock); btrfs_set_root_node(&root->fs_info->tree_root->root_item, root->fs_info->tree_root->node); @@ -1373,10 +1387,14 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, sizeof(root->fs_info->super_copy)); trans->transaction->blocked = 0; + spin_lock(&root->fs_info->trans_lock); + root->fs_info->running_transaction = NULL; + root->fs_info->trans_no_join = 0; + spin_unlock(&root->fs_info->trans_lock); + mutex_unlock(&root->fs_info->reloc_mutex); wake_up(&root->fs_info->transaction_wait); - mutex_unlock(&root->fs_info->trans_mutex); ret = btrfs_write_and_wait_transaction(trans, root); BUG_ON(ret); write_ctree_super(trans, root, 0); @@ -1389,21 +1407,22 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans, btrfs_finish_extent_commit(trans, root); - mutex_lock(&root->fs_info->trans_mutex); - cur_trans->commit_done = 1; root->fs_info->last_trans_committed = cur_trans->transid; wake_up(&cur_trans->commit_wait); + spin_lock(&root->fs_info->trans_lock); list_del_init(&cur_trans->list); + spin_unlock(&root->fs_info->trans_lock); + put_transaction(cur_trans); put_transaction(cur_trans); trace_btrfs_transaction_commit(root); - mutex_unlock(&root->fs_info->trans_mutex); + btrfs_scrub_continue(root); if (current->journal_info == trans) current->journal_info = NULL; @@ -1424,14 +1443,16 @@ int btrfs_clean_old_snapshots(struct btrfs_root *root) LIST_HEAD(list); struct btrfs_fs_info *fs_info = root->fs_info; - mutex_lock(&fs_info->trans_mutex); + spin_lock(&fs_info->trans_lock); list_splice_init(&fs_info->dead_roots, &list); - mutex_unlock(&fs_info->trans_mutex); + spin_unlock(&fs_info->trans_lock); while (!list_empty(&list)) { root = list_entry(list.next, struct btrfs_root, root_list); list_del(&root->root_list); + btrfs_kill_all_delayed_nodes(root); + if (btrfs_header_backref_rev(root->node) < BTRFS_MIXED_BACKREF_REV) btrfs_drop_snapshot(root, NULL, 0); diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h index e441acc6c584..02564e6230ac 100644 --- a/fs/btrfs/transaction.h +++ b/fs/btrfs/transaction.h @@ -28,10 +28,12 @@ struct btrfs_transaction { * transaction can end */ atomic_t num_writers; + atomic_t use_count; unsigned long num_joined; + + spinlock_t commit_lock; int in_commit; - atomic_t use_count; int commit_done; int blocked; struct list_head list; @@ -45,13 +47,14 @@ struct btrfs_transaction { struct btrfs_trans_handle { u64 transid; - u64 block_group; u64 bytes_reserved; + unsigned long use_count; unsigned long blocks_reserved; unsigned long blocks_used; unsigned long delayed_ref_updates; struct btrfs_transaction *transaction; struct btrfs_block_rsv *block_rsv; + struct btrfs_block_rsv *orig_rsv; }; struct btrfs_pending_snapshot { @@ -66,19 +69,6 @@ struct btrfs_pending_snapshot { struct list_head list; }; -static inline void btrfs_set_trans_block_group(struct btrfs_trans_handle *trans, - struct inode *inode) -{ - trans->block_group = BTRFS_I(inode)->block_group; -} - -static inline void btrfs_update_inode_block_group( - struct btrfs_trans_handle *trans, - struct inode *inode) -{ - BTRFS_I(inode)->block_group = trans->block_group; -} - static inline void btrfs_set_inode_last_trans(struct btrfs_trans_handle *trans, struct inode *inode) { @@ -92,20 +82,14 @@ int btrfs_end_transaction_nolock(struct btrfs_trans_handle *trans, struct btrfs_root *root); struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root, int num_items); -struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root, - int num_blocks); -struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root, - int num_blocks); -struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *r, - int num_blocks); +struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root); +struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root); +struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root); int btrfs_wait_for_commit(struct btrfs_root *root, u64 transid); int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root); -int btrfs_commit_tree_roots(struct btrfs_trans_handle *trans, - struct btrfs_root *root); int btrfs_add_dead_root(struct btrfs_root *root); -int btrfs_drop_dead_root(struct btrfs_root *root); int btrfs_defrag_root(struct btrfs_root *root, int cacheonly); int btrfs_clean_old_snapshots(struct btrfs_root *root); int btrfs_commit_transaction(struct btrfs_trans_handle *trans, @@ -115,6 +99,8 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, int wait_for_unblock); int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans, struct btrfs_root *root); +int btrfs_end_transaction_dmeta(struct btrfs_trans_handle *trans, + struct btrfs_root *root); int btrfs_should_end_transaction(struct btrfs_trans_handle *trans, struct btrfs_root *root); void btrfs_throttle(struct btrfs_root *root); diff --git a/fs/btrfs/tree-defrag.c b/fs/btrfs/tree-defrag.c index 992ab425599d..3b580ee8ab1d 100644 --- a/fs/btrfs/tree-defrag.c +++ b/fs/btrfs/tree-defrag.c @@ -97,7 +97,7 @@ int btrfs_defrag_leaves(struct btrfs_trans_handle *trans, ret = 0; goto out; } - btrfs_release_path(root, path); + btrfs_release_path(path); wret = btrfs_search_slot(trans, root, &key, path, 0, 1); if (wret < 0) { diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index f997ec0c1ba4..4ce8a9f41d1e 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c @@ -333,13 +333,13 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans, goto insert; if (item_size == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } dst_copy = kmalloc(item_size, GFP_NOFS); src_copy = kmalloc(item_size, GFP_NOFS); if (!dst_copy || !src_copy) { - btrfs_release_path(root, path); + btrfs_release_path(path); kfree(dst_copy); kfree(src_copy); return -ENOMEM; @@ -361,13 +361,13 @@ static noinline int overwrite_item(struct btrfs_trans_handle *trans, * sync */ if (ret == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } } insert: - btrfs_release_path(root, path); + btrfs_release_path(path); /* try to insert the key into the destination tree */ ret = btrfs_insert_empty_item(trans, root, path, key, item_size); @@ -382,7 +382,6 @@ insert: } else if (found_size < item_size) { ret = btrfs_extend_item(trans, root, path, item_size - found_size); - BUG_ON(ret); } } else if (ret) { return ret; @@ -438,7 +437,7 @@ insert: } no_copy: btrfs_mark_buffer_dirty(path->nodes[0]); - btrfs_release_path(root, path); + btrfs_release_path(path); return 0; } @@ -519,7 +518,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, * file. This must be done before the btrfs_drop_extents run * so we don't try to drop this extent. */ - ret = btrfs_lookup_file_extent(trans, root, path, inode->i_ino, + ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode), start, 0); if (ret == 0 && @@ -544,11 +543,11 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, * we don't have to do anything */ if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); goto out; } } - btrfs_release_path(root, path); + btrfs_release_path(path); saved_nbytes = inode_get_bytes(inode); /* drop any overlapping extents */ @@ -590,6 +589,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, ins.objectid, ins.offset, 0, root->root_key.objectid, key->objectid, offset); + BUG_ON(ret); } else { /* * insert the extent pointer in the extent @@ -600,7 +600,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, key->objectid, offset, &ins); BUG_ON(ret); } - btrfs_release_path(root, path); + btrfs_release_path(path); if (btrfs_file_extent_compression(eb, item)) { csum_start = ins.objectid; @@ -614,7 +614,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, ret = btrfs_lookup_csums_range(root->log_root, csum_start, csum_end - 1, - &ordered_sums); + &ordered_sums, 0); BUG_ON(ret); while (!list_empty(&ordered_sums)) { struct btrfs_ordered_sum *sums; @@ -629,7 +629,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans, kfree(sums); } } else { - btrfs_release_path(root, path); + btrfs_release_path(path); } } else if (found_type == BTRFS_FILE_EXTENT_INLINE) { /* inline extents are easy, we just overwrite them */ @@ -675,10 +675,13 @@ static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans, return -ENOMEM; read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len); - btrfs_release_path(root, path); + btrfs_release_path(path); inode = read_one_inode(root, location.objectid); - BUG_ON(!inode); + if (!inode) { + kfree(name); + return -EIO; + } ret = link_to_fixup_dir(trans, root, path, location.objectid); BUG_ON(ret); @@ -713,7 +716,7 @@ static noinline int inode_in_dir(struct btrfs_root *root, goto out; } else goto out; - btrfs_release_path(root, path); + btrfs_release_path(path); di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0); if (di && !IS_ERR(di)) { @@ -724,7 +727,7 @@ static noinline int inode_in_dir(struct btrfs_root *root, goto out; match = 1; out: - btrfs_release_path(root, path); + btrfs_release_path(path); return match; } @@ -817,7 +820,10 @@ static noinline int add_inode_ref(struct btrfs_trans_handle *trans, return -ENOENT; inode = read_one_inode(root, key->objectid); - BUG_ON(!inode); + if (!inode) { + iput(dir); + return -EIO; + } ref_ptr = btrfs_item_ptr_offset(eb, slot); ref_end = ref_ptr + btrfs_item_size_nr(eb, slot); @@ -832,7 +838,7 @@ again: read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen); /* if we already have a perfect match, we're done */ - if (inode_in_dir(root, path, dir->i_ino, inode->i_ino, + if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode), btrfs_inode_ref_index(eb, ref), name, namelen)) { goto out; @@ -884,7 +890,7 @@ again: if (!backref_in_log(log, key, victim_name, victim_name_len)) { btrfs_inc_nlink(inode); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_unlink_inode(trans, root, dir, inode, victim_name, @@ -901,7 +907,7 @@ again: */ search_done = 1; } - btrfs_release_path(root, path); + btrfs_release_path(path); insert: /* insert our name */ @@ -922,7 +928,7 @@ out: BUG_ON(ret); out_nowrite: - btrfs_release_path(root, path); + btrfs_release_path(path); iput(dir); iput(inode); return 0; @@ -960,8 +966,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, unsigned long ptr; unsigned long ptr_end; int name_len; + u64 ino = btrfs_ino(inode); - key.objectid = inode->i_ino; + key.objectid = ino; key.type = BTRFS_INODE_REF_KEY; key.offset = (u64)-1; @@ -980,7 +987,7 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, } btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]); - if (key.objectid != inode->i_ino || + if (key.objectid != ino || key.type != BTRFS_INODE_REF_KEY) break; ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]); @@ -999,9 +1006,9 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, if (key.offset == 0) break; key.offset--; - btrfs_release_path(root, path); + btrfs_release_path(path); } - btrfs_release_path(root, path); + btrfs_release_path(path); if (nlink != inode->i_nlink) { inode->i_nlink = nlink; btrfs_update_inode(trans, root, inode); @@ -1011,10 +1018,10 @@ static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans, if (inode->i_nlink == 0) { if (S_ISDIR(inode->i_mode)) { ret = replay_dir_deletes(trans, root, NULL, path, - inode->i_ino, 1); + ino, 1); BUG_ON(ret); } - ret = insert_orphan_item(trans, root, inode->i_ino); + ret = insert_orphan_item(trans, root, ino); BUG_ON(ret); } btrfs_free_path(path); @@ -1050,11 +1057,13 @@ static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans, break; ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); + if (ret) + goto out; - btrfs_release_path(root, path); + btrfs_release_path(path); inode = read_one_inode(root, key.offset); - BUG_ON(!inode); + if (!inode) + return -EIO; ret = fixup_inode_link_count(trans, root, inode); BUG_ON(ret); @@ -1068,8 +1077,10 @@ static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans, */ key.offset = (u64)-1; } - btrfs_release_path(root, path); - return 0; + ret = 0; +out: + btrfs_release_path(path); + return ret; } @@ -1088,7 +1099,8 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans, struct inode *inode; inode = read_one_inode(root, objectid); - BUG_ON(!inode); + if (!inode) + return -EIO; key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID; btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY); @@ -1096,7 +1108,7 @@ static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans, ret = btrfs_insert_empty_item(trans, root, path, &key, 0); - btrfs_release_path(root, path); + btrfs_release_path(path); if (ret == 0) { btrfs_inc_nlink(inode); btrfs_update_inode(trans, root, inode); @@ -1175,7 +1187,8 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, int ret; dir = read_one_inode(root, key->objectid); - BUG_ON(!dir); + if (!dir) + return -EIO; name_len = btrfs_dir_name_len(eb, di); name = kmalloc(name_len, GFP_NOFS); @@ -1192,7 +1205,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, exists = 1; else exists = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); if (key->type == BTRFS_DIR_ITEM_KEY) { dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid, @@ -1205,7 +1218,7 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, } else { BUG(); } - if (!dst_di || IS_ERR(dst_di)) { + if (IS_ERR_OR_NULL(dst_di)) { /* we need a sequence number to insert, so we only * do inserts for the BTRFS_DIR_INDEX_KEY types */ @@ -1236,13 +1249,13 @@ static noinline int replay_one_name(struct btrfs_trans_handle *trans, if (key->type == BTRFS_DIR_INDEX_KEY) goto insert; out: - btrfs_release_path(root, path); + btrfs_release_path(path); kfree(name); iput(dir); return 0; insert: - btrfs_release_path(root, path); + btrfs_release_path(path); ret = insert_one_name(trans, root, path, key->objectid, key->offset, name, name_len, log_type, &log_key); @@ -1363,7 +1376,7 @@ next: *end_ret = found_end; ret = 0; out: - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } @@ -1426,12 +1439,15 @@ again: dir_key->offset, name, name_len, 0); } - if (!log_di || IS_ERR(log_di)) { + if (IS_ERR_OR_NULL(log_di)) { btrfs_dir_item_key_to_cpu(eb, di, &location); - btrfs_release_path(root, path); - btrfs_release_path(log, log_path); + btrfs_release_path(path); + btrfs_release_path(log_path); inode = read_one_inode(root, location.objectid); - BUG_ON(!inode); + if (!inode) { + kfree(name); + return -EIO; + } ret = link_to_fixup_dir(trans, root, path, location.objectid); @@ -1453,7 +1469,7 @@ again: ret = 0; goto out; } - btrfs_release_path(log, log_path); + btrfs_release_path(log_path); kfree(name); ptr = (unsigned long)(di + 1); @@ -1461,8 +1477,8 @@ again: } ret = 0; out: - btrfs_release_path(root, path); - btrfs_release_path(log, log_path); + btrfs_release_path(path); + btrfs_release_path(log_path); return ret; } @@ -1550,7 +1566,7 @@ again: break; dir_key.offset = found_key.offset + 1; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (range_end == (u64)-1) break; range_start = range_end + 1; @@ -1561,11 +1577,11 @@ next_type: if (key_type == BTRFS_DIR_LOG_ITEM_KEY) { key_type = BTRFS_DIR_LOG_INDEX_KEY; dir_key.type = BTRFS_DIR_INDEX_KEY; - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } out: - btrfs_release_path(root, path); + btrfs_release_path(path); btrfs_free_path(log_path); iput(dir); return ret; @@ -2093,7 +2109,9 @@ int btrfs_sync_log(struct btrfs_trans_handle *trans, * the running transaction open, so a full commit can't hop * in and cause problems either. */ + btrfs_scrub_pause_super(root); write_ctree_super(trans, root->fs_info->tree_root, 1); + btrfs_scrub_continue_super(root); ret = 0; mutex_lock(&root->log_mutex); @@ -2197,6 +2215,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, int ret; int err = 0; int bytes_del = 0; + u64 dir_ino = btrfs_ino(dir); if (BTRFS_I(dir)->logged_trans < trans->transid) return 0; @@ -2214,7 +2233,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, goto out_unlock; } - di = btrfs_lookup_dir_item(trans, log, path, dir->i_ino, + di = btrfs_lookup_dir_item(trans, log, path, dir_ino, name, name_len, -1); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2225,8 +2244,8 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, bytes_del += name_len; BUG_ON(ret); } - btrfs_release_path(log, path); - di = btrfs_lookup_dir_index_item(trans, log, path, dir->i_ino, + btrfs_release_path(path); + di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino, index, name, name_len, -1); if (IS_ERR(di)) { err = PTR_ERR(di); @@ -2244,10 +2263,10 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, if (bytes_del) { struct btrfs_key key; - key.objectid = dir->i_ino; + key.objectid = dir_ino; key.offset = 0; key.type = BTRFS_INODE_ITEM_KEY; - btrfs_release_path(log, path); + btrfs_release_path(path); ret = btrfs_search_slot(trans, log, &key, path, 0, 1); if (ret < 0) { @@ -2269,7 +2288,7 @@ int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans, btrfs_mark_buffer_dirty(path->nodes[0]); } else ret = 0; - btrfs_release_path(log, path); + btrfs_release_path(path); } fail: btrfs_free_path(path); @@ -2303,7 +2322,7 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, log = root->log_root; mutex_lock(&BTRFS_I(inode)->log_mutex); - ret = btrfs_del_inode_ref(trans, log, name, name_len, inode->i_ino, + ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode), dirid, &index); mutex_unlock(&BTRFS_I(inode)->log_mutex); if (ret == -ENOSPC) { @@ -2344,7 +2363,7 @@ static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans, struct btrfs_dir_log_item); btrfs_set_dir_log_end(path->nodes[0], item, last_offset); btrfs_mark_buffer_dirty(path->nodes[0]); - btrfs_release_path(log, path); + btrfs_release_path(path); return 0; } @@ -2369,13 +2388,14 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, int nritems; u64 first_offset = min_offset; u64 last_offset = (u64)-1; + u64 ino = btrfs_ino(inode); log = root->log_root; - max_key.objectid = inode->i_ino; + max_key.objectid = ino; max_key.offset = (u64)-1; max_key.type = key_type; - min_key.objectid = inode->i_ino; + min_key.objectid = ino; min_key.type = key_type; min_key.offset = min_offset; @@ -2388,18 +2408,17 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, * we didn't find anything from this transaction, see if there * is anything at all */ - if (ret != 0 || min_key.objectid != inode->i_ino || - min_key.type != key_type) { - min_key.objectid = inode->i_ino; + if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) { + min_key.objectid = ino; min_key.type = key_type; min_key.offset = (u64)-1; - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0); if (ret < 0) { - btrfs_release_path(root, path); + btrfs_release_path(path); return ret; } - ret = btrfs_previous_item(root, path, inode->i_ino, key_type); + ret = btrfs_previous_item(root, path, ino, key_type); /* if ret == 0 there are items for this type, * create a range to tell us the last key of this type. @@ -2417,7 +2436,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, } /* go backward to find any previous key */ - ret = btrfs_previous_item(root, path, inode->i_ino, key_type); + ret = btrfs_previous_item(root, path, ino, key_type); if (ret == 0) { struct btrfs_key tmp; btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]); @@ -2432,7 +2451,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, } } } - btrfs_release_path(root, path); + btrfs_release_path(path); /* find the first key from this transaction again */ ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0); @@ -2452,8 +2471,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, for (i = path->slots[0]; i < nritems; i++) { btrfs_item_key_to_cpu(src, &min_key, i); - if (min_key.objectid != inode->i_ino || - min_key.type != key_type) + if (min_key.objectid != ino || min_key.type != key_type) goto done; ret = overwrite_item(trans, log, dst_path, src, i, &min_key); @@ -2474,7 +2492,7 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, goto done; } btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]); - if (tmp.objectid != inode->i_ino || tmp.type != key_type) { + if (tmp.objectid != ino || tmp.type != key_type) { last_offset = (u64)-1; goto done; } @@ -2490,8 +2508,8 @@ static noinline int log_dir_items(struct btrfs_trans_handle *trans, } } done: - btrfs_release_path(root, path); - btrfs_release_path(log, dst_path); + btrfs_release_path(path); + btrfs_release_path(dst_path); if (err == 0) { *last_offset_ret = last_offset; @@ -2500,8 +2518,7 @@ done: * is valid */ ret = insert_dir_log_key(trans, log, path, key_type, - inode->i_ino, first_offset, - last_offset); + ino, first_offset, last_offset); if (ret) err = ret; } @@ -2587,10 +2604,11 @@ static int drop_objectid_items(struct btrfs_trans_handle *trans, break; ret = btrfs_del_item(trans, log, path); - BUG_ON(ret); - btrfs_release_path(log, path); + if (ret) + break; + btrfs_release_path(path); } - btrfs_release_path(log, path); + btrfs_release_path(path); return ret; } @@ -2665,6 +2683,9 @@ static noinline int copy_items(struct btrfs_trans_handle *trans, extent = btrfs_item_ptr(src, start_slot + i, struct btrfs_file_extent_item); + if (btrfs_file_extent_generation(src, extent) < trans->transid) + continue; + found_type = btrfs_file_extent_type(src, extent); if (found_type == BTRFS_FILE_EXTENT_REG || found_type == BTRFS_FILE_EXTENT_PREALLOC) { @@ -2689,14 +2710,14 @@ static noinline int copy_items(struct btrfs_trans_handle *trans, ret = btrfs_lookup_csums_range( log->fs_info->csum_root, ds + cs, ds + cs + cl - 1, - &ordered_sums); + &ordered_sums, 0); BUG_ON(ret); } } } btrfs_mark_buffer_dirty(dst_path->nodes[0]); - btrfs_release_path(log, dst_path); + btrfs_release_path(dst_path); kfree(ins_data); /* @@ -2745,6 +2766,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, int nritems; int ins_start_slot = 0; int ins_nr; + u64 ino = btrfs_ino(inode); log = root->log_root; @@ -2757,11 +2779,11 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, return -ENOMEM; } - min_key.objectid = inode->i_ino; + min_key.objectid = ino; min_key.type = BTRFS_INODE_ITEM_KEY; min_key.offset = 0; - max_key.objectid = inode->i_ino; + max_key.objectid = ino; /* today the code can only do partial logging of directories */ if (!S_ISDIR(inode->i_mode)) @@ -2773,6 +2795,13 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, max_key.type = (u8)-1; max_key.offset = (u64)-1; + ret = btrfs_commit_inode_delayed_items(trans, inode); + if (ret) { + btrfs_free_path(path); + btrfs_free_path(dst_path); + return ret; + } + mutex_lock(&BTRFS_I(inode)->log_mutex); /* @@ -2784,8 +2813,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, if (inode_only == LOG_INODE_EXISTS) max_key_type = BTRFS_XATTR_ITEM_KEY; - ret = drop_objectid_items(trans, log, path, - inode->i_ino, max_key_type); + ret = drop_objectid_items(trans, log, path, ino, max_key_type); } else { ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0); } @@ -2803,7 +2831,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans, break; again: /* note, ins_nr might be > 0 here, cleanup outside the loop */ - if (min_key.objectid != inode->i_ino) + if (min_key.objectid != ino) break; if (min_key.type > max_key.type) break; @@ -2845,7 +2873,7 @@ next_slot: } ins_nr = 0; } - btrfs_release_path(root, path); + btrfs_release_path(path); if (min_key.offset < (u64)-1) min_key.offset++; @@ -2868,8 +2896,8 @@ next_slot: } WARN_ON(ins_nr); if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) { - btrfs_release_path(root, path); - btrfs_release_path(log, dst_path); + btrfs_release_path(path); + btrfs_release_path(dst_path); ret = log_directory_changes(trans, root, inode, path, dst_path); if (ret) { err = ret; @@ -3136,7 +3164,7 @@ again: } btrfs_item_key_to_cpu(path->nodes[0], &found_key, path->slots[0]); - btrfs_release_path(log_root_tree, path); + btrfs_release_path(path); if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID) break; @@ -3149,7 +3177,7 @@ again: tmp_key.offset = (u64)-1; wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key); - BUG_ON(!wc.replay_dest); + BUG_ON(IS_ERR_OR_NULL(wc.replay_dest)); wc.replay_dest->log_root = log; btrfs_record_root_in_trans(trans, wc.replay_dest); @@ -3171,7 +3199,7 @@ again: if (found_key.offset == 0) break; } - btrfs_release_path(log_root_tree, path); + btrfs_release_path(path); /* step one is to pin it all, step two is to replay just inodes */ if (wc.pin) { diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h index 3dfae84c8cc8..2270ac58d746 100644 --- a/fs/btrfs/tree-log.h +++ b/fs/btrfs/tree-log.h @@ -38,7 +38,6 @@ int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans, struct btrfs_root *root, const char *name, int name_len, struct inode *inode, u64 dirid); -int btrfs_join_running_log_trans(struct btrfs_root *root); int btrfs_end_log_trans(struct btrfs_root *root); int btrfs_pin_log_trans(struct btrfs_root *root); int btrfs_log_inode_parent(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/version.sh b/fs/btrfs/version.sh deleted file mode 100644 index 1ca1952fd917..000000000000 --- a/fs/btrfs/version.sh +++ /dev/null @@ -1,43 +0,0 @@ -#!/bin/bash -# -# determine-version -- report a useful version for releases -# -# Copyright 2008, Aron Griffis <agriffis@n01se.net> -# Copyright 2008, Oracle -# Released under the GNU GPLv2 - -v="v0.16" - -which git &> /dev/null -if [ $? == 0 ]; then - git branch >& /dev/null - if [ $? == 0 ]; then - if head=`git rev-parse --verify HEAD 2>/dev/null`; then - if tag=`git describe --tags 2>/dev/null`; then - v="$tag" - fi - - # Are there uncommitted changes? - git update-index --refresh --unmerged > /dev/null - if git diff-index --name-only HEAD | \ - grep -v "^scripts/package" \ - | read dummy; then - v="$v"-dirty - fi - fi - fi -fi - -echo "#ifndef __BUILD_VERSION" > .build-version.h -echo "#define __BUILD_VERSION" >> .build-version.h -echo "#define BTRFS_BUILD_VERSION \"Btrfs $v\"" >> .build-version.h -echo "#endif" >> .build-version.h - -diff -q version.h .build-version.h >& /dev/null - -if [ $? == 0 ]; then - rm .build-version.h - exit 0 -fi - -mv .build-version.h version.h diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index c7367ae5a3e6..1efa56e18f9b 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -38,22 +38,9 @@ static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_device *device); static int btrfs_relocate_sys_chunks(struct btrfs_root *root); -#define map_lookup_size(n) (sizeof(struct map_lookup) + \ - (sizeof(struct btrfs_bio_stripe) * (n))) - static DEFINE_MUTEX(uuid_mutex); static LIST_HEAD(fs_uuids); -void btrfs_lock_volumes(void) -{ - mutex_lock(&uuid_mutex); -} - -void btrfs_unlock_volumes(void) -{ - mutex_unlock(&uuid_mutex); -} - static void lock_chunks(struct btrfs_root *root) { mutex_lock(&root->fs_info->chunk_mutex); @@ -363,7 +350,7 @@ static noinline int device_list_add(const char *path, INIT_LIST_HEAD(&device->dev_alloc_list); mutex_lock(&fs_devices->device_list_mutex); - list_add(&device->dev_list, &fs_devices->devices); + list_add_rcu(&device->dev_list, &fs_devices->devices); mutex_unlock(&fs_devices->device_list_mutex); device->fs_devices = fs_devices; @@ -406,7 +393,7 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig) fs_devices->latest_trans = orig->latest_trans; memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid)); - mutex_lock(&orig->device_list_mutex); + /* We have held the volume lock, it is safe to get the devices. */ list_for_each_entry(orig_dev, &orig->devices, dev_list) { device = kzalloc(sizeof(*device), GFP_NOFS); if (!device) @@ -429,10 +416,8 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig) device->fs_devices = fs_devices; fs_devices->num_devices++; } - mutex_unlock(&orig->device_list_mutex); return fs_devices; error: - mutex_unlock(&orig->device_list_mutex); free_fs_devices(fs_devices); return ERR_PTR(-ENOMEM); } @@ -443,7 +428,7 @@ int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices) mutex_lock(&uuid_mutex); again: - mutex_lock(&fs_devices->device_list_mutex); + /* This is the initialized path, it is safe to release the devices. */ list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) { if (device->in_fs_metadata) continue; @@ -463,7 +448,6 @@ again: kfree(device->name); kfree(device); } - mutex_unlock(&fs_devices->device_list_mutex); if (fs_devices->seed) { fs_devices = fs_devices->seed; @@ -474,6 +458,29 @@ again: return 0; } +static void __free_device(struct work_struct *work) +{ + struct btrfs_device *device; + + device = container_of(work, struct btrfs_device, rcu_work); + + if (device->bdev) + blkdev_put(device->bdev, device->mode); + + kfree(device->name); + kfree(device); +} + +static void free_device(struct rcu_head *head) +{ + struct btrfs_device *device; + + device = container_of(head, struct btrfs_device, rcu); + + INIT_WORK(&device->rcu_work, __free_device); + schedule_work(&device->rcu_work); +} + static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) { struct btrfs_device *device; @@ -481,20 +488,32 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) if (--fs_devices->opened > 0) return 0; + mutex_lock(&fs_devices->device_list_mutex); list_for_each_entry(device, &fs_devices->devices, dev_list) { - if (device->bdev) { - blkdev_put(device->bdev, device->mode); + struct btrfs_device *new_device; + + if (device->bdev) fs_devices->open_devices--; - } + if (device->writeable) { list_del_init(&device->dev_alloc_list); fs_devices->rw_devices--; } - device->bdev = NULL; - device->writeable = 0; - device->in_fs_metadata = 0; + new_device = kmalloc(sizeof(*new_device), GFP_NOFS); + BUG_ON(!new_device); + memcpy(new_device, device, sizeof(*new_device)); + new_device->name = kstrdup(device->name, GFP_NOFS); + BUG_ON(device->name && !new_device->name); + new_device->bdev = NULL; + new_device->writeable = 0; + new_device->in_fs_metadata = 0; + list_replace_rcu(&device->dev_list, &new_device->dev_list); + + call_rcu(&device->rcu, free_device); } + mutex_unlock(&fs_devices->device_list_mutex); + WARN_ON(fs_devices->open_devices); WARN_ON(fs_devices->rw_devices); fs_devices->opened = 0; @@ -597,6 +616,7 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, list_add(&device->dev_alloc_list, &fs_devices->alloc_list); } + brelse(bh); continue; error_brelse: @@ -669,12 +689,8 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, transid = btrfs_super_generation(disk_super); if (disk_super->label[0]) printk(KERN_INFO "device label %s ", disk_super->label); - else { - /* FIXME, make a readl uuid parser */ - printk(KERN_INFO "device fsid %llx-%llx ", - *(unsigned long long *)disk_super->fsid, - *(unsigned long long *)(disk_super->fsid + 8)); - } + else + printk(KERN_INFO "device fsid %pU ", disk_super->fsid); printk(KERN_CONT "devid %llu transid %llu %s\n", (unsigned long long)devid, (unsigned long long)transid, path); ret = device_list_add(path, disk_super, devid, fs_devices_ret); @@ -815,10 +831,7 @@ int find_free_dev_extent(struct btrfs_trans_handle *trans, /* we don't want to overwrite the superblock on the drive, * so we make sure to start at an offset of at least 1MB */ - search_start = 1024 * 1024; - - if (root->fs_info->alloc_start + num_bytes <= search_end) - search_start = max(root->fs_info->alloc_start, search_start); + search_start = max(root->fs_info->alloc_start, 1024ull * 1024); max_hole_start = search_start; max_hole_size = 0; @@ -949,14 +962,14 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, if (ret > 0) { ret = btrfs_previous_item(root, path, key.objectid, BTRFS_DEV_EXTENT_KEY); - BUG_ON(ret); + if (ret) + goto out; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); extent = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_extent); BUG_ON(found_key.offset > start || found_key.offset + btrfs_dev_extent_length(leaf, extent) < start); - ret = 0; } else if (ret == 0) { leaf = path->nodes[0]; extent = btrfs_item_ptr(leaf, path->slots[0], @@ -967,8 +980,8 @@ static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans, if (device->bytes_used > 0) device->bytes_used -= btrfs_dev_extent_length(leaf, extent); ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); +out: btrfs_free_path(path); return ret; } @@ -1203,11 +1216,13 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) struct block_device *bdev; struct buffer_head *bh = NULL; struct btrfs_super_block *disk_super; + struct btrfs_fs_devices *cur_devices; u64 all_avail; u64 devid; u64 num_devices; u8 *dev_uuid; int ret = 0; + bool clear_super = false; mutex_lock(&uuid_mutex); mutex_lock(&root->fs_info->volume_mutex); @@ -1238,14 +1253,16 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) device = NULL; devices = &root->fs_info->fs_devices->devices; - mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + /* + * It is safe to read the devices since the volume_mutex + * is held. + */ list_for_each_entry(tmp, devices, dev_list) { if (tmp->in_fs_metadata && !tmp->bdev) { device = tmp; break; } } - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); bdev = NULL; bh = NULL; disk_super = NULL; @@ -1287,8 +1304,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) } if (device->writeable) { + lock_chunks(root); list_del_init(&device->dev_alloc_list); + unlock_chunks(root); root->fs_info->fs_devices->rw_devices--; + clear_super = true; } ret = btrfs_shrink_device(device, 0); @@ -1300,15 +1320,17 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) goto error_undo; device->in_fs_metadata = 0; + btrfs_scrub_cancel_dev(root, device); /* * the device list mutex makes sure that we don't change * the device list while someone else is writing out all * the device supers. */ + + cur_devices = device->fs_devices; mutex_lock(&root->fs_info->fs_devices->device_list_mutex); - list_del_init(&device->dev_list); - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + list_del_rcu(&device->dev_list); device->fs_devices->num_devices--; @@ -1322,34 +1344,36 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (device->bdev == root->fs_info->fs_devices->latest_bdev) root->fs_info->fs_devices->latest_bdev = next_device->bdev; - if (device->bdev) { - blkdev_put(device->bdev, device->mode); - device->bdev = NULL; + if (device->bdev) device->fs_devices->open_devices--; - } + + call_rcu(&device->rcu, free_device); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1; btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices); - if (device->fs_devices->open_devices == 0) { + if (cur_devices->open_devices == 0) { struct btrfs_fs_devices *fs_devices; fs_devices = root->fs_info->fs_devices; while (fs_devices) { - if (fs_devices->seed == device->fs_devices) + if (fs_devices->seed == cur_devices) break; fs_devices = fs_devices->seed; } - fs_devices->seed = device->fs_devices->seed; - device->fs_devices->seed = NULL; - __btrfs_close_devices(device->fs_devices); - free_fs_devices(device->fs_devices); + fs_devices->seed = cur_devices->seed; + cur_devices->seed = NULL; + lock_chunks(root); + __btrfs_close_devices(cur_devices); + unlock_chunks(root); + free_fs_devices(cur_devices); } /* * at this point, the device is zero sized. We want to * remove it from the devices list and zero out the old super */ - if (device->writeable) { + if (clear_super) { /* make sure this device isn't detected as part of * the FS anymore */ @@ -1358,8 +1382,6 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) sync_dirty_buffer(bh); } - kfree(device->name); - kfree(device); ret = 0; error_brelse: @@ -1373,8 +1395,10 @@ out: return ret; error_undo: if (device->writeable) { + lock_chunks(root); list_add(&device->dev_alloc_list, &root->fs_info->fs_devices->alloc_list); + unlock_chunks(root); root->fs_info->fs_devices->rw_devices++; } goto error_brelse; @@ -1414,7 +1438,12 @@ static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans, INIT_LIST_HEAD(&seed_devices->devices); INIT_LIST_HEAD(&seed_devices->alloc_list); mutex_init(&seed_devices->device_list_mutex); - list_splice_init(&fs_devices->devices, &seed_devices->devices); + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + list_splice_init_rcu(&fs_devices->devices, &seed_devices->devices, + synchronize_rcu); + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list); list_for_each_entry(device, &seed_devices->devices, dev_list) { device->fs_devices = seed_devices; @@ -1475,7 +1504,7 @@ next_slot: goto error; leaf = path->nodes[0]; btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); - btrfs_release_path(root, path); + btrfs_release_path(path); continue; } @@ -1611,7 +1640,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) * half setup */ mutex_lock(&root->fs_info->fs_devices->device_list_mutex); - list_add(&device->dev_list, &root->fs_info->fs_devices->devices); + list_add_rcu(&device->dev_list, &root->fs_info->fs_devices->devices); list_add(&device->dev_alloc_list, &root->fs_info->fs_devices->alloc_list); root->fs_info->fs_devices->num_devices++; @@ -1769,10 +1798,9 @@ static int btrfs_free_chunk(struct btrfs_trans_handle *trans, BUG_ON(ret); ret = btrfs_del_item(trans, root, path); - BUG_ON(ret); btrfs_free_path(path); - return 0; + return ret; } static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64 @@ -1947,7 +1975,7 @@ again: chunk = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_chunk); chunk_type = btrfs_chunk_type(leaf, chunk); - btrfs_release_path(chunk_root, path); + btrfs_release_path(path); if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) { ret = btrfs_relocate_chunk(chunk_root, chunk_tree, @@ -2065,7 +2093,7 @@ int btrfs_balance(struct btrfs_root *dev_root) if (found_key.offset == 0) break; - btrfs_release_path(chunk_root, path); + btrfs_release_path(path); ret = btrfs_relocate_chunk(chunk_root, chunk_root->root_key.objectid, found_key.objectid, @@ -2137,7 +2165,7 @@ again: goto done; if (ret) { ret = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); break; } @@ -2146,7 +2174,7 @@ again: btrfs_item_key_to_cpu(l, &key, path->slots[0]); if (key.objectid != device->devid) { - btrfs_release_path(root, path); + btrfs_release_path(path); break; } @@ -2154,14 +2182,14 @@ again: length = btrfs_dev_extent_length(l, dev_extent); if (key.offset + length <= new_size) { - btrfs_release_path(root, path); + btrfs_release_path(path); break; } chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent); chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent); chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent); - btrfs_release_path(root, path); + btrfs_release_path(path); ret = btrfs_relocate_chunk(root, chunk_tree, chunk_objectid, chunk_offset); @@ -2237,275 +2265,204 @@ static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans, return 0; } -static noinline u64 chunk_bytes_by_type(u64 type, u64 calc_size, - int num_stripes, int sub_stripes) +/* + * sort the devices in descending order by max_avail, total_avail + */ +static int btrfs_cmp_device_info(const void *a, const void *b) { - if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP)) - return calc_size; - else if (type & BTRFS_BLOCK_GROUP_RAID10) - return calc_size * (num_stripes / sub_stripes); - else - return calc_size * num_stripes; -} + const struct btrfs_device_info *di_a = a; + const struct btrfs_device_info *di_b = b; -/* Used to sort the devices by max_avail(descending sort) */ -int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2) -{ - if (((struct btrfs_device_info *)dev_info1)->max_avail > - ((struct btrfs_device_info *)dev_info2)->max_avail) + if (di_a->max_avail > di_b->max_avail) return -1; - else if (((struct btrfs_device_info *)dev_info1)->max_avail < - ((struct btrfs_device_info *)dev_info2)->max_avail) + if (di_a->max_avail < di_b->max_avail) return 1; - else - return 0; + if (di_a->total_avail > di_b->total_avail) + return -1; + if (di_a->total_avail < di_b->total_avail) + return 1; + return 0; } -static int __btrfs_calc_nstripes(struct btrfs_fs_devices *fs_devices, u64 type, - int *num_stripes, int *min_stripes, - int *sub_stripes) +static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, + struct btrfs_root *extent_root, + struct map_lookup **map_ret, + u64 *num_bytes_out, u64 *stripe_size_out, + u64 start, u64 type) { - *num_stripes = 1; - *min_stripes = 1; - *sub_stripes = 0; + struct btrfs_fs_info *info = extent_root->fs_info; + struct btrfs_fs_devices *fs_devices = info->fs_devices; + struct list_head *cur; + struct map_lookup *map = NULL; + struct extent_map_tree *em_tree; + struct extent_map *em; + struct btrfs_device_info *devices_info = NULL; + u64 total_avail; + int num_stripes; /* total number of stripes to allocate */ + int sub_stripes; /* sub_stripes info for map */ + int dev_stripes; /* stripes per dev */ + int devs_max; /* max devs to use */ + int devs_min; /* min devs needed */ + int devs_increment; /* ndevs has to be a multiple of this */ + int ncopies; /* how many copies to data has */ + int ret; + u64 max_stripe_size; + u64 max_chunk_size; + u64 stripe_size; + u64 num_bytes; + int ndevs; + int i; + int j; - if (type & (BTRFS_BLOCK_GROUP_RAID0)) { - *num_stripes = fs_devices->rw_devices; - *min_stripes = 2; - } - if (type & (BTRFS_BLOCK_GROUP_DUP)) { - *num_stripes = 2; - *min_stripes = 2; - } - if (type & (BTRFS_BLOCK_GROUP_RAID1)) { - if (fs_devices->rw_devices < 2) - return -ENOSPC; - *num_stripes = 2; - *min_stripes = 2; - } - if (type & (BTRFS_BLOCK_GROUP_RAID10)) { - *num_stripes = fs_devices->rw_devices; - if (*num_stripes < 4) - return -ENOSPC; - *num_stripes &= ~(u32)1; - *sub_stripes = 2; - *min_stripes = 4; + if ((type & BTRFS_BLOCK_GROUP_RAID1) && + (type & BTRFS_BLOCK_GROUP_DUP)) { + WARN_ON(1); + type &= ~BTRFS_BLOCK_GROUP_DUP; } - return 0; -} + if (list_empty(&fs_devices->alloc_list)) + return -ENOSPC; -static u64 __btrfs_calc_stripe_size(struct btrfs_fs_devices *fs_devices, - u64 proposed_size, u64 type, - int num_stripes, int small_stripe) -{ - int min_stripe_size = 1 * 1024 * 1024; - u64 calc_size = proposed_size; - u64 max_chunk_size = calc_size; - int ncopies = 1; + sub_stripes = 1; + dev_stripes = 1; + devs_increment = 1; + ncopies = 1; + devs_max = 0; /* 0 == as many as possible */ + devs_min = 1; - if (type & (BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_DUP | - BTRFS_BLOCK_GROUP_RAID10)) + /* + * define the properties of each RAID type. + * FIXME: move this to a global table and use it in all RAID + * calculation code + */ + if (type & (BTRFS_BLOCK_GROUP_DUP)) { + dev_stripes = 2; ncopies = 2; + devs_max = 1; + } else if (type & (BTRFS_BLOCK_GROUP_RAID0)) { + devs_min = 2; + } else if (type & (BTRFS_BLOCK_GROUP_RAID1)) { + devs_increment = 2; + ncopies = 2; + devs_max = 2; + devs_min = 2; + } else if (type & (BTRFS_BLOCK_GROUP_RAID10)) { + sub_stripes = 2; + devs_increment = 2; + ncopies = 2; + devs_min = 4; + } else { + devs_max = 1; + } if (type & BTRFS_BLOCK_GROUP_DATA) { - max_chunk_size = 10 * calc_size; - min_stripe_size = 64 * 1024 * 1024; + max_stripe_size = 1024 * 1024 * 1024; + max_chunk_size = 10 * max_stripe_size; } else if (type & BTRFS_BLOCK_GROUP_METADATA) { - max_chunk_size = 256 * 1024 * 1024; - min_stripe_size = 32 * 1024 * 1024; + max_stripe_size = 256 * 1024 * 1024; + max_chunk_size = max_stripe_size; } else if (type & BTRFS_BLOCK_GROUP_SYSTEM) { - calc_size = 8 * 1024 * 1024; - max_chunk_size = calc_size * 2; - min_stripe_size = 1 * 1024 * 1024; + max_stripe_size = 8 * 1024 * 1024; + max_chunk_size = 2 * max_stripe_size; + } else { + printk(KERN_ERR "btrfs: invalid chunk type 0x%llx requested\n", + type); + BUG_ON(1); } /* we don't want a chunk larger than 10% of writeable space */ max_chunk_size = min(div_factor(fs_devices->total_rw_bytes, 1), max_chunk_size); - if (calc_size * num_stripes > max_chunk_size * ncopies) { - calc_size = max_chunk_size * ncopies; - do_div(calc_size, num_stripes); - do_div(calc_size, BTRFS_STRIPE_LEN); - calc_size *= BTRFS_STRIPE_LEN; - } + devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices, + GFP_NOFS); + if (!devices_info) + return -ENOMEM; - /* we don't want tiny stripes */ - if (!small_stripe) - calc_size = max_t(u64, min_stripe_size, calc_size); + cur = fs_devices->alloc_list.next; /* - * we're about to do_div by the BTRFS_STRIPE_LEN so lets make sure - * we end up with something bigger than a stripe + * in the first pass through the devices list, we gather information + * about the available holes on each device. */ - calc_size = max_t(u64, calc_size, BTRFS_STRIPE_LEN); - - do_div(calc_size, BTRFS_STRIPE_LEN); - calc_size *= BTRFS_STRIPE_LEN; - - return calc_size; -} - -static struct map_lookup *__shrink_map_lookup_stripes(struct map_lookup *map, - int num_stripes) -{ - struct map_lookup *new; - size_t len = map_lookup_size(num_stripes); - - BUG_ON(map->num_stripes < num_stripes); - - if (map->num_stripes == num_stripes) - return map; - - new = kmalloc(len, GFP_NOFS); - if (!new) { - /* just change map->num_stripes */ - map->num_stripes = num_stripes; - return map; - } - - memcpy(new, map, len); - new->num_stripes = num_stripes; - kfree(map); - return new; -} + ndevs = 0; + while (cur != &fs_devices->alloc_list) { + struct btrfs_device *device; + u64 max_avail; + u64 dev_offset; -/* - * helper to allocate device space from btrfs_device_info, in which we stored - * max free space information of every device. It is used when we can not - * allocate chunks by default size. - * - * By this helper, we can allocate a new chunk as larger as possible. - */ -static int __btrfs_alloc_tiny_space(struct btrfs_trans_handle *trans, - struct btrfs_fs_devices *fs_devices, - struct btrfs_device_info *devices, - int nr_device, u64 type, - struct map_lookup **map_lookup, - int min_stripes, u64 *stripe_size) -{ - int i, index, sort_again = 0; - int min_devices = min_stripes; - u64 max_avail, min_free; - struct map_lookup *map = *map_lookup; - int ret; + device = list_entry(cur, struct btrfs_device, dev_alloc_list); - if (nr_device < min_stripes) - return -ENOSPC; + cur = cur->next; - btrfs_descending_sort_devices(devices, nr_device); + if (!device->writeable) { + printk(KERN_ERR + "btrfs: read-only device in alloc_list\n"); + WARN_ON(1); + continue; + } - max_avail = devices[0].max_avail; - if (!max_avail) - return -ENOSPC; + if (!device->in_fs_metadata) + continue; - for (i = 0; i < nr_device; i++) { - /* - * if dev_offset = 0, it means the free space of this device - * is less than what we need, and we didn't search max avail - * extent on this device, so do it now. + if (device->total_bytes > device->bytes_used) + total_avail = device->total_bytes - device->bytes_used; + else + total_avail = 0; + /* avail is off by max(alloc_start, 1MB), but that is the same + * for all devices, so it doesn't hurt the sorting later on */ - if (!devices[i].dev_offset) { - ret = find_free_dev_extent(trans, devices[i].dev, - max_avail, - &devices[i].dev_offset, - &devices[i].max_avail); - if (ret != 0 && ret != -ENOSPC) - return ret; - sort_again = 1; - } - } - /* we update the max avail free extent of each devices, sort again */ - if (sort_again) - btrfs_descending_sort_devices(devices, nr_device); - - if (type & BTRFS_BLOCK_GROUP_DUP) - min_devices = 1; + ret = find_free_dev_extent(trans, device, + max_stripe_size * dev_stripes, + &dev_offset, &max_avail); + if (ret && ret != -ENOSPC) + goto error; - if (!devices[min_devices - 1].max_avail) - return -ENOSPC; + if (ret == 0) + max_avail = max_stripe_size * dev_stripes; - max_avail = devices[min_devices - 1].max_avail; - if (type & BTRFS_BLOCK_GROUP_DUP) - do_div(max_avail, 2); + if (max_avail < BTRFS_STRIPE_LEN * dev_stripes) + continue; - max_avail = __btrfs_calc_stripe_size(fs_devices, max_avail, type, - min_stripes, 1); - if (type & BTRFS_BLOCK_GROUP_DUP) - min_free = max_avail * 2; - else - min_free = max_avail; + devices_info[ndevs].dev_offset = dev_offset; + devices_info[ndevs].max_avail = max_avail; + devices_info[ndevs].total_avail = total_avail; + devices_info[ndevs].dev = device; + ++ndevs; + } - if (min_free > devices[min_devices - 1].max_avail) - return -ENOSPC; + /* + * now sort the devices by hole size / available space + */ + sort(devices_info, ndevs, sizeof(struct btrfs_device_info), + btrfs_cmp_device_info, NULL); - map = __shrink_map_lookup_stripes(map, min_stripes); - *stripe_size = max_avail; + /* round down to number of usable stripes */ + ndevs -= ndevs % devs_increment; - index = 0; - for (i = 0; i < min_stripes; i++) { - map->stripes[i].dev = devices[index].dev; - map->stripes[i].physical = devices[index].dev_offset; - if (type & BTRFS_BLOCK_GROUP_DUP) { - i++; - map->stripes[i].dev = devices[index].dev; - map->stripes[i].physical = devices[index].dev_offset + - max_avail; - } - index++; + if (ndevs < devs_increment * sub_stripes || ndevs < devs_min) { + ret = -ENOSPC; + goto error; } - *map_lookup = map; - - return 0; -} -static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, - struct btrfs_root *extent_root, - struct map_lookup **map_ret, - u64 *num_bytes, u64 *stripe_size, - u64 start, u64 type) -{ - struct btrfs_fs_info *info = extent_root->fs_info; - struct btrfs_device *device = NULL; - struct btrfs_fs_devices *fs_devices = info->fs_devices; - struct list_head *cur; - struct map_lookup *map; - struct extent_map_tree *em_tree; - struct extent_map *em; - struct btrfs_device_info *devices_info; - struct list_head private_devs; - u64 calc_size = 1024 * 1024 * 1024; - u64 min_free; - u64 avail; - u64 dev_offset; - int num_stripes; - int min_stripes; - int sub_stripes; - int min_devices; /* the min number of devices we need */ - int i; - int ret; - int index; + if (devs_max && ndevs > devs_max) + ndevs = devs_max; + /* + * the primary goal is to maximize the number of stripes, so use as many + * devices as possible, even if the stripes are not maximum sized. + */ + stripe_size = devices_info[ndevs-1].max_avail; + num_stripes = ndevs * dev_stripes; - if ((type & BTRFS_BLOCK_GROUP_RAID1) && - (type & BTRFS_BLOCK_GROUP_DUP)) { - WARN_ON(1); - type &= ~BTRFS_BLOCK_GROUP_DUP; + if (stripe_size * num_stripes > max_chunk_size * ncopies) { + stripe_size = max_chunk_size * ncopies; + do_div(stripe_size, num_stripes); } - if (list_empty(&fs_devices->alloc_list)) - return -ENOSPC; - - ret = __btrfs_calc_nstripes(fs_devices, type, &num_stripes, - &min_stripes, &sub_stripes); - if (ret) - return ret; - devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices, - GFP_NOFS); - if (!devices_info) - return -ENOMEM; + do_div(stripe_size, dev_stripes); + do_div(stripe_size, BTRFS_STRIPE_LEN); + stripe_size *= BTRFS_STRIPE_LEN; map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS); if (!map) { @@ -2514,85 +2471,12 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, } map->num_stripes = num_stripes; - cur = fs_devices->alloc_list.next; - index = 0; - i = 0; - - calc_size = __btrfs_calc_stripe_size(fs_devices, calc_size, type, - num_stripes, 0); - - if (type & BTRFS_BLOCK_GROUP_DUP) { - min_free = calc_size * 2; - min_devices = 1; - } else { - min_free = calc_size; - min_devices = min_stripes; - } - - INIT_LIST_HEAD(&private_devs); - while (index < num_stripes) { - device = list_entry(cur, struct btrfs_device, dev_alloc_list); - BUG_ON(!device->writeable); - if (device->total_bytes > device->bytes_used) - avail = device->total_bytes - device->bytes_used; - else - avail = 0; - cur = cur->next; - - if (device->in_fs_metadata && avail >= min_free) { - ret = find_free_dev_extent(trans, device, min_free, - &devices_info[i].dev_offset, - &devices_info[i].max_avail); - if (ret == 0) { - list_move_tail(&device->dev_alloc_list, - &private_devs); - map->stripes[index].dev = device; - map->stripes[index].physical = - devices_info[i].dev_offset; - index++; - if (type & BTRFS_BLOCK_GROUP_DUP) { - map->stripes[index].dev = device; - map->stripes[index].physical = - devices_info[i].dev_offset + - calc_size; - index++; - } - } else if (ret != -ENOSPC) - goto error; - - devices_info[i].dev = device; - i++; - } else if (device->in_fs_metadata && - avail >= BTRFS_STRIPE_LEN) { - devices_info[i].dev = device; - devices_info[i].max_avail = avail; - i++; - } - - if (cur == &fs_devices->alloc_list) - break; - } - - list_splice(&private_devs, &fs_devices->alloc_list); - if (index < num_stripes) { - if (index >= min_stripes) { - num_stripes = index; - if (type & (BTRFS_BLOCK_GROUP_RAID10)) { - num_stripes /= sub_stripes; - num_stripes *= sub_stripes; - } - - map = __shrink_map_lookup_stripes(map, num_stripes); - } else if (i >= min_devices) { - ret = __btrfs_alloc_tiny_space(trans, fs_devices, - devices_info, i, type, - &map, min_stripes, - &calc_size); - if (ret) - goto error; - } else { - ret = -ENOSPC; - goto error; + for (i = 0; i < ndevs; ++i) { + for (j = 0; j < dev_stripes; ++j) { + int s = i * dev_stripes + j; + map->stripes[s].dev = devices_info[i].dev; + map->stripes[s].physical = devices_info[i].dev_offset + + j * stripe_size; } } map->sector_size = extent_root->sectorsize; @@ -2603,20 +2487,21 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, map->sub_stripes = sub_stripes; *map_ret = map; - *stripe_size = calc_size; - *num_bytes = chunk_bytes_by_type(type, calc_size, - map->num_stripes, sub_stripes); + num_bytes = stripe_size * (num_stripes / ncopies); - trace_btrfs_chunk_alloc(info->chunk_root, map, start, *num_bytes); + *stripe_size_out = stripe_size; + *num_bytes_out = num_bytes; - em = alloc_extent_map(GFP_NOFS); + trace_btrfs_chunk_alloc(info->chunk_root, map, start, num_bytes); + + em = alloc_extent_map(); if (!em) { ret = -ENOMEM; goto error; } em->bdev = (struct block_device *)map; em->start = start; - em->len = *num_bytes; + em->len = num_bytes; em->block_start = 0; em->block_len = em->len; @@ -2629,20 +2514,21 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, ret = btrfs_make_block_group(trans, extent_root, 0, type, BTRFS_FIRST_CHUNK_TREE_OBJECTID, - start, *num_bytes); + start, num_bytes); BUG_ON(ret); - index = 0; - while (index < map->num_stripes) { - device = map->stripes[index].dev; - dev_offset = map->stripes[index].physical; + for (i = 0; i < map->num_stripes; ++i) { + struct btrfs_device *device; + u64 dev_offset; + + device = map->stripes[i].dev; + dev_offset = map->stripes[i].physical; ret = btrfs_alloc_dev_extent(trans, device, info->chunk_root->root_key.objectid, BTRFS_FIRST_CHUNK_TREE_OBJECTID, - start, dev_offset, calc_size); + start, dev_offset, stripe_size); BUG_ON(ret); - index++; } kfree(devices_info); @@ -2849,7 +2735,7 @@ int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset) void btrfs_mapping_init(struct btrfs_mapping_tree *tree) { - extent_map_tree_init(&tree->map_tree, GFP_NOFS); + extent_map_tree_init(&tree->map_tree); } void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree) @@ -3499,7 +3385,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, free_extent_map(em); } - em = alloc_extent_map(GFP_NOFS); + em = alloc_extent_map(); if (!em) return -ENOMEM; num_stripes = btrfs_chunk_num_stripes(leaf, chunk); @@ -3688,15 +3574,6 @@ static int read_one_dev(struct btrfs_root *root, return ret; } -int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf) -{ - struct btrfs_dev_item *dev_item; - - dev_item = (struct btrfs_dev_item *)offsetof(struct btrfs_super_block, - dev_item); - return read_one_dev(root, buf, dev_item); -} - int btrfs_read_sys_array(struct btrfs_root *root) { struct btrfs_super_block *super_copy = &root->fs_info->super_copy; @@ -3813,7 +3690,7 @@ again: } if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) { key.objectid = 0; - btrfs_release_path(root, path); + btrfs_release_path(path); goto again; } ret = 0; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index cc2eadaf7a27..7c12d61ae7ae 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -85,7 +85,12 @@ struct btrfs_device { /* physical drive uuid (or lvm uuid) */ u8 uuid[BTRFS_UUID_SIZE]; + /* per-device scrub information */ + struct scrub_dev *scrub_device; + struct btrfs_work work; + struct rcu_head rcu; + struct work_struct rcu_work; }; struct btrfs_fs_devices { @@ -144,6 +149,7 @@ struct btrfs_device_info { struct btrfs_device *dev; u64 dev_offset; u64 max_avail; + u64 total_avail; }; struct map_lookup { @@ -157,20 +163,8 @@ struct map_lookup { struct btrfs_bio_stripe stripes[]; }; -/* Used to sort the devices by max_avail(descending sort) */ -int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2); - -/* - * sort the devices by max_avail, in which max free extent size of each device - * is stored.(Descending Sort) - */ -static inline void btrfs_descending_sort_devices( - struct btrfs_device_info *devices, - size_t nr_devices) -{ - sort(devices, nr_devices, sizeof(struct btrfs_device_info), - btrfs_cmp_device_free_bytes, NULL); -} +#define map_lookup_size(n) (sizeof(struct map_lookup) + \ + (sizeof(struct btrfs_bio_stripe) * (n))) int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start, u64 end, u64 *length); @@ -196,7 +190,6 @@ void btrfs_mapping_init(struct btrfs_mapping_tree *tree); void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree); int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, int mirror_num, int async_submit); -int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf); int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, fmode_t flags, void *holder); int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, @@ -209,8 +202,6 @@ int btrfs_add_device(struct btrfs_trans_handle *trans, int btrfs_rm_device(struct btrfs_root *root, char *device_path); int btrfs_cleanup_fs_uuids(void); int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len); -int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree, - u64 logical, struct page *page); int btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size); struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, @@ -218,8 +209,6 @@ struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, int btrfs_shrink_device(struct btrfs_device *device, u64 new_size); int btrfs_init_new_device(struct btrfs_root *root, char *path); int btrfs_balance(struct btrfs_root *dev_root); -void btrfs_unlock_volumes(void); -void btrfs_lock_volumes(void); int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset); int find_free_dev_extent(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 num_bytes, diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c index cfd660550ded..5366fe452ab0 100644 --- a/fs/btrfs/xattr.c +++ b/fs/btrfs/xattr.c @@ -44,7 +44,7 @@ ssize_t __btrfs_getxattr(struct inode *inode, const char *name, return -ENOMEM; /* lookup the xattr by name */ - di = btrfs_lookup_xattr(NULL, root, path, inode->i_ino, name, + di = btrfs_lookup_xattr(NULL, root, path, btrfs_ino(inode), name, strlen(name), 0); if (!di) { ret = -ENODATA; @@ -103,7 +103,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans, return -ENOMEM; /* first lets see if we already have this xattr */ - di = btrfs_lookup_xattr(trans, root, path, inode->i_ino, name, + di = btrfs_lookup_xattr(trans, root, path, btrfs_ino(inode), name, strlen(name), -1); if (IS_ERR(di)) { ret = PTR_ERR(di); @@ -120,13 +120,13 @@ static int do_setxattr(struct btrfs_trans_handle *trans, ret = btrfs_delete_one_dir_name(trans, root, path, di); BUG_ON(ret); - btrfs_release_path(root, path); + btrfs_release_path(path); /* if we don't have a value then we are removing the xattr */ if (!value) goto out; } else { - btrfs_release_path(root, path); + btrfs_release_path(path); if (flags & XATTR_REPLACE) { /* we couldn't find the attr to replace */ @@ -136,7 +136,7 @@ static int do_setxattr(struct btrfs_trans_handle *trans, } /* ok we have to create a completely new xattr */ - ret = btrfs_insert_xattr_item(trans, root, path, inode->i_ino, + ret = btrfs_insert_xattr_item(trans, root, path, btrfs_ino(inode), name, name_len, value, size); BUG_ON(ret); out: @@ -158,8 +158,6 @@ int __btrfs_setxattr(struct btrfs_trans_handle *trans, if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, inode); - ret = do_setxattr(trans, inode, name, value, size, flags); if (ret) goto out; @@ -190,7 +188,7 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size) * NOTE: we set key.offset = 0; because we want to start with the * first xattr that we find and walk forward */ - key.objectid = inode->i_ino; + key.objectid = btrfs_ino(inode); btrfs_set_key_type(&key, BTRFS_XATTR_ITEM_KEY); key.offset = 0; diff --git a/fs/buffer.c b/fs/buffer.c index 698c6b2cc462..1a80b048ade8 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -1902,10 +1902,8 @@ int __block_write_begin(struct page *page, loff_t pos, unsigned len, if (!buffer_uptodate(*wait_bh)) err = -EIO; } - if (unlikely(err)) { + if (unlikely(err)) page_zero_new_buffers(page, from, to); - ClearPageUptodate(page); - } return err; } EXPORT_SYMBOL(__block_write_begin); @@ -2382,6 +2380,7 @@ int __block_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf, ret = -EAGAIN; goto out_unlock; } + wait_on_page_writeback(page); return 0; out_unlock: unlock_page(page); diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 33da49dc3cc6..5a3953db8118 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -453,7 +453,7 @@ static int ceph_writepage(struct page *page, struct writeback_control *wbc) int err; struct inode *inode = page->mapping->host; BUG_ON(!inode); - igrab(inode); + ihold(inode); err = writepage_nounlock(page, wbc); unlock_page(page); iput(inode); diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c index 1f72b00447c4..f605753c8fe9 100644 --- a/fs/ceph/caps.c +++ b/fs/ceph/caps.c @@ -2940,14 +2940,12 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc) while (!list_empty(&mdsc->cap_dirty)) { ci = list_first_entry(&mdsc->cap_dirty, struct ceph_inode_info, i_dirty_item); - inode = igrab(&ci->vfs_inode); + inode = &ci->vfs_inode; + ihold(inode); dout("flush_dirty_caps %p\n", inode); spin_unlock(&mdsc->cap_dirty_lock); - if (inode) { - ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_FLUSH, - NULL); - iput(inode); - } + ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_FLUSH, NULL); + iput(inode); spin_lock(&mdsc->cap_dirty_lock); } spin_unlock(&mdsc->cap_dirty_lock); diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c index 33729e822bb9..ef8f08c343e8 100644 --- a/fs/ceph/dir.c +++ b/fs/ceph/dir.c @@ -308,7 +308,8 @@ more: req = ceph_mdsc_create_request(mdsc, op, USE_AUTH_MDS); if (IS_ERR(req)) return PTR_ERR(req); - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_dentry = dget(filp->f_dentry); /* hints to request -> mds selection code */ req->r_direct_mode = USE_AUTH_MDS; @@ -787,10 +788,12 @@ static int ceph_link(struct dentry *old_dentry, struct inode *dir, req->r_dentry_drop = CEPH_CAP_FILE_SHARED; req->r_dentry_unless = CEPH_CAP_FILE_EXCL; err = ceph_mdsc_do_request(mdsc, dir, req); - if (err) + if (err) { d_drop(dentry); - else if (!req->r_reply_info.head->is_dentry) - d_instantiate(dentry, igrab(old_dentry->d_inode)); + } else if (!req->r_reply_info.head->is_dentry) { + ihold(old_dentry->d_inode); + d_instantiate(dentry, old_dentry->d_inode); + } ceph_mdsc_put_request(req); return err; } diff --git a/fs/ceph/export.c b/fs/ceph/export.c index a610d3d67488..f67b687550de 100644 --- a/fs/ceph/export.c +++ b/fs/ceph/export.c @@ -109,7 +109,7 @@ static struct dentry *__fh_to_dentry(struct super_block *sb, err = ceph_mdsc_do_request(mdsc, NULL, req); inode = req->r_target_inode; if (inode) - igrab(inode); + ihold(inode); ceph_mdsc_put_request(req); if (!inode) return ERR_PTR(-ESTALE); @@ -167,7 +167,7 @@ static struct dentry *__cfh_to_dentry(struct super_block *sb, err = ceph_mdsc_do_request(mdsc, NULL, req); inode = req->r_target_inode; if (inode) - igrab(inode); + ihold(inode); ceph_mdsc_put_request(req); if (!inode) return ERR_PTR(err ? err : -ESTALE); diff --git a/fs/ceph/file.c b/fs/ceph/file.c index 203252d88d9f..9542f07d0b93 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -191,7 +191,8 @@ int ceph_open(struct inode *inode, struct file *file) err = PTR_ERR(req); goto out; } - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_num_caps = 1; err = ceph_mdsc_do_request(mdsc, parent_inode, req); if (!err) @@ -282,7 +283,7 @@ int ceph_release(struct inode *inode, struct file *file) static int striped_read(struct inode *inode, u64 off, u64 len, struct page **pages, int num_pages, - int *checkeof, bool align_to_pages, + int *checkeof, bool o_direct, unsigned long buf_align) { struct ceph_fs_client *fsc = ceph_inode_to_client(inode); @@ -307,7 +308,7 @@ static int striped_read(struct inode *inode, io_align = off & ~PAGE_MASK; more: - if (align_to_pages) + if (o_direct) page_align = (pos - io_align + buf_align) & ~PAGE_MASK; else page_align = pos & ~PAGE_MASK; @@ -317,10 +318,10 @@ more: ci->i_truncate_seq, ci->i_truncate_size, page_pos, pages_left, page_align); - hit_stripe = this_len < left; - was_short = ret >= 0 && ret < this_len; if (ret == -ENOENT) ret = 0; + hit_stripe = this_len < left; + was_short = ret >= 0 && ret < this_len; dout("striped_read %llu~%u (read %u) got %d%s%s\n", pos, left, read, ret, hit_stripe ? " HITSTRIPE" : "", was_short ? " SHORT" : ""); @@ -345,20 +346,22 @@ more: } if (was_short) { - /* was original extent fully inside i_size? */ - if (pos + left <= inode->i_size) { - dout("zero tail\n"); - ceph_zero_page_vector_range(page_off + read, len - read, + /* did we bounce off eof? */ + if (pos + left > inode->i_size) + *checkeof = 1; + + /* zero trailing bytes (inside i_size) */ + if (left > 0 && pos < inode->i_size) { + if (pos + left > inode->i_size) + left = inode->i_size - pos; + + dout("zero tail %d\n", left); + ceph_zero_page_vector_range(page_off + read, left, pages); - read = len; - goto out; + read += left; } - - /* check i_size */ - *checkeof = 1; } -out: if (ret >= 0) ret = read; dout("striped_read returns %d\n", ret); @@ -658,7 +661,7 @@ out: /* hit EOF or hole? */ if (statret == 0 && *ppos < inode->i_size) { - dout("aio_read sync_read hit hole, reading more\n"); + dout("aio_read sync_read hit hole, ppos %lld < size %lld, reading more\n", *ppos, inode->i_size); read += ret; base += ret; len -= ret; diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c index 70b6a4839c38..d8858e96ab18 100644 --- a/fs/ceph/inode.c +++ b/fs/ceph/inode.c @@ -1101,10 +1101,10 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, goto done; } req->r_dentry = dn; /* may have spliced */ - igrab(in); + ihold(in); } else if (ceph_ino(in) == vino.ino && ceph_snap(in) == vino.snap) { - igrab(in); + ihold(in); } else { dout(" %p links to %p %llx.%llx, not %llx.%llx\n", dn, in, ceph_ino(in), ceph_snap(in), @@ -1144,7 +1144,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req, goto done; } req->r_dentry = dn; /* may have spliced */ - igrab(in); + ihold(in); rinfo->head->is_dentry = 1; /* fool notrace handlers */ } @@ -1328,7 +1328,7 @@ void ceph_queue_writeback(struct inode *inode) if (queue_work(ceph_inode_to_client(inode)->wb_wq, &ceph_inode(inode)->i_wb_work)) { dout("ceph_queue_writeback %p\n", inode); - igrab(inode); + ihold(inode); } else { dout("ceph_queue_writeback %p failed\n", inode); } @@ -1353,7 +1353,7 @@ void ceph_queue_invalidate(struct inode *inode) if (queue_work(ceph_inode_to_client(inode)->pg_inv_wq, &ceph_inode(inode)->i_pg_inv_work)) { dout("ceph_queue_invalidate %p\n", inode); - igrab(inode); + ihold(inode); } else { dout("ceph_queue_invalidate %p failed\n", inode); } @@ -1477,7 +1477,7 @@ void ceph_queue_vmtruncate(struct inode *inode) if (queue_work(ceph_sb_to_client(inode->i_sb)->trunc_wq, &ci->i_vmtruncate_work)) { dout("ceph_queue_vmtruncate %p\n", inode); - igrab(inode); + ihold(inode); } else { dout("ceph_queue_vmtruncate %p failed, pending=%d\n", inode, ci->i_truncate_pending); @@ -1738,7 +1738,8 @@ int ceph_setattr(struct dentry *dentry, struct iattr *attr) __mark_inode_dirty(inode, inode_dirty_flags); if (mask) { - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_inode_drop = release; req->r_args.setattr.mask = cpu_to_le32(mask); req->r_num_caps = 1; @@ -1779,7 +1780,8 @@ int ceph_do_getattr(struct inode *inode, int mask) req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_GETATTR, USE_ANY_MDS); if (IS_ERR(req)) return PTR_ERR(req); - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_num_caps = 1; req->r_args.getattr.mask = cpu_to_le32(mask); err = ceph_mdsc_do_request(mdsc, NULL, req); diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c index 8888c9ba68db..ef0b5f48e13a 100644 --- a/fs/ceph/ioctl.c +++ b/fs/ceph/ioctl.c @@ -73,7 +73,8 @@ static long ceph_ioctl_set_layout(struct file *file, void __user *arg) USE_AUTH_MDS); if (IS_ERR(req)) return PTR_ERR(req); - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_inode_drop = CEPH_CAP_FILE_SHARED | CEPH_CAP_FILE_EXCL; req->r_args.setlayout.layout.fl_stripe_unit = @@ -135,7 +136,8 @@ static long ceph_ioctl_set_layout_policy (struct file *file, void __user *arg) if (IS_ERR(req)) return PTR_ERR(req); - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_args.setlayout.layout.fl_stripe_unit = cpu_to_le32(l.stripe_unit); diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c index 476b329867d4..80576d05d687 100644 --- a/fs/ceph/locks.c +++ b/fs/ceph/locks.c @@ -23,7 +23,8 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, req = ceph_mdsc_create_request(mdsc, operation, USE_AUTH_MDS); if (IS_ERR(req)) return PTR_ERR(req); - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); /* mds requires start and length rather than start and end */ if (LLONG_MAX == fl->fl_end) @@ -32,11 +33,10 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, length = fl->fl_end - fl->fl_start + 1; dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " - "length: %llu, wait: %d, type`: %d", (int)lock_type, + "length: %llu, wait: %d, type: %d", (int)lock_type, (int)operation, (u64)fl->fl_pid, fl->fl_start, length, wait, fl->fl_type); - req->r_args.filelock_change.rule = lock_type; req->r_args.filelock_change.type = cmd; req->r_args.filelock_change.pid = cpu_to_le64((u64)fl->fl_pid); @@ -70,7 +70,7 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, } ceph_mdsc_put_request(req); dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " - "length: %llu, wait: %d, type`: %d, err code %d", (int)lock_type, + "length: %llu, wait: %d, type: %d, err code %d", (int)lock_type, (int)operation, (u64)fl->fl_pid, fl->fl_start, length, wait, fl->fl_type, err); return err; @@ -109,16 +109,20 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl) dout("mds locked, locking locally"); err = posix_lock_file(file, fl, NULL); if (err && (CEPH_MDS_OP_SETFILELOCK == op)) { - /* undo! This should only happen if the kernel detects - * local deadlock. */ + /* undo! This should only happen if + * the kernel detects local + * deadlock. */ ceph_lock_message(CEPH_LOCK_FCNTL, op, file, CEPH_LOCK_UNLOCK, 0, fl); - dout("got %d on posix_lock_file, undid lock", err); + dout("got %d on posix_lock_file, undid lock", + err); } } - } else { - dout("mds returned error code %d", err); + } else if (err == -ERESTARTSYS) { + dout("undoing lock\n"); + ceph_lock_message(CEPH_LOCK_FCNTL, op, file, + CEPH_LOCK_UNLOCK, 0, fl); } return err; } @@ -155,8 +159,11 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl) file, CEPH_LOCK_UNLOCK, 0, fl); dout("got %d on flock_lock_file_wait, undid lock", err); } - } else { - dout("mds error code %d", err); + } else if (err == -ERESTARTSYS) { + dout("undoing lock\n"); + ceph_lock_message(CEPH_LOCK_FLOCK, + CEPH_MDS_OP_SETFILELOCK, + file, CEPH_LOCK_UNLOCK, 0, fl); } return err; } diff --git a/fs/ceph/snap.c b/fs/ceph/snap.c index 24067d68a554..54b14de2e729 100644 --- a/fs/ceph/snap.c +++ b/fs/ceph/snap.c @@ -722,7 +722,7 @@ static void flush_snaps(struct ceph_mds_client *mdsc) ci = list_first_entry(&mdsc->snap_flush_list, struct ceph_inode_info, i_snap_flush_item); inode = &ci->vfs_inode; - igrab(inode); + ihold(inode); spin_unlock(&mdsc->snap_flush_lock); spin_lock(&inode->i_lock); __ceph_flush_snaps(ci, &session, 0); diff --git a/fs/ceph/xattr.c b/fs/ceph/xattr.c index f2b628696180..f42d730f1b66 100644 --- a/fs/ceph/xattr.c +++ b/fs/ceph/xattr.c @@ -665,7 +665,8 @@ static int ceph_sync_setxattr(struct dentry *dentry, const char *name, err = PTR_ERR(req); goto out; } - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_inode_drop = CEPH_CAP_XATTR_SHARED; req->r_num_caps = 1; req->r_args.setxattr.flags = cpu_to_le32(flags); @@ -795,7 +796,8 @@ static int ceph_send_removexattr(struct dentry *dentry, const char *name) USE_AUTH_MDS); if (IS_ERR(req)) return PTR_ERR(req); - req->r_inode = igrab(inode); + req->r_inode = inode; + ihold(inode); req->r_inode_drop = CEPH_CAP_XATTR_SHARED; req->r_num_caps = 1; req->r_path2 = kstrdup(name, GFP_NOFS); diff --git a/fs/cifs/Kconfig b/fs/cifs/Kconfig index 75c47cd8d086..f66cc1625150 100644 --- a/fs/cifs/Kconfig +++ b/fs/cifs/Kconfig @@ -7,6 +7,7 @@ config CIFS select CRYPTO_MD5 select CRYPTO_HMAC select CRYPTO_ARC4 + select CRYPTO_ECB select CRYPTO_DES help This is the client VFS module for the Common Internet File System @@ -148,33 +149,13 @@ config CIFS_FSCACHE config CIFS_ACL bool "Provide CIFS ACL support (EXPERIMENTAL)" - depends on EXPERIMENTAL && CIFS_XATTR + depends on EXPERIMENTAL && CIFS_XATTR && KEYS help Allows to fetch CIFS/NTFS ACL from the server. The DACL blob is handed over to the application/caller. -config CIFS_SMB2 - bool "SMB2 network file system support (EXPERIMENTAL)" - depends on EXPERIMENTAL && INET && BROKEN - select NLS - select KEYS - select FSCACHE - select DNS_RESOLVER - - help - This enables experimental support for the SMB2 (Server Message Block - version 2) protocol. The SMB2 protocol is the successor to the - popular CIFS and SMB network file sharing protocols. SMB2 is the - native file sharing mechanism for recent versions of Windows - operating systems (since Vista). SMB2 enablement will eventually - allow users better performance, security and features, than would be - possible with cifs. Note that smb2 mount options also are simpler - (compared to cifs) due to protocol improvements. - - Unless you are a developer or tester, say N. - config CIFS_NFSD_EXPORT bool "Allow nfsd to export CIFS file system (EXPERIMENTAL)" - depends on CIFS && EXPERIMENTAL + depends on CIFS && EXPERIMENTAL && BROKEN help Allows NFS server to export a CIFS mounted share (nfsd over cifs) diff --git a/fs/cifs/README b/fs/cifs/README index 4a3ca0e5ca24..c5c2c5e5f0f2 100644 --- a/fs/cifs/README +++ b/fs/cifs/README @@ -457,6 +457,9 @@ A partial list of the supported mount options follows: otherwise - read from the server. All written data are stored in the cache, but if the client doesn't have Exclusive Oplock, it writes the data to the server. + rwpidforward Forward pid of a process who opened a file to any read or write + operation on that file. This prevent applications like WINE + from failing on read and write if we use mandatory brlock style. acl Allow setfacl and getfacl to manage posix ACLs if server supports them. (default) noacl Do not allow setfacl and getfacl calls on this mount diff --git a/fs/cifs/cache.c b/fs/cifs/cache.c index 53d57a3fe427..545509c3313b 100644 --- a/fs/cifs/cache.c +++ b/fs/cifs/cache.c @@ -92,7 +92,7 @@ static uint16_t cifs_server_get_key(const void *cookie_netfs_data, break; default: - cERROR(1, "CIFS: Unknown network family '%d'", sa->sa_family); + cERROR(1, "Unknown network family '%d'", sa->sa_family); key_len = 0; break; } @@ -146,13 +146,13 @@ static char *extract_sharename(const char *treename) static uint16_t cifs_super_get_key(const void *cookie_netfs_data, void *buffer, uint16_t maxbuf) { - const struct cifsTconInfo *tcon = cookie_netfs_data; + const struct cifs_tcon *tcon = cookie_netfs_data; char *sharename; uint16_t len; sharename = extract_sharename(tcon->treeName); if (IS_ERR(sharename)) { - cFYI(1, "CIFS: couldn't extract sharename\n"); + cFYI(1, "%s: couldn't extract sharename\n", __func__); sharename = NULL; return 0; } @@ -173,7 +173,7 @@ cifs_fscache_super_get_aux(const void *cookie_netfs_data, void *buffer, uint16_t maxbuf) { struct cifs_fscache_super_auxdata auxdata; - const struct cifsTconInfo *tcon = cookie_netfs_data; + const struct cifs_tcon *tcon = cookie_netfs_data; memset(&auxdata, 0, sizeof(auxdata)); auxdata.resource_id = tcon->resource_id; @@ -192,7 +192,7 @@ fscache_checkaux cifs_fscache_super_check_aux(void *cookie_netfs_data, uint16_t datalen) { struct cifs_fscache_super_auxdata auxdata; - const struct cifsTconInfo *tcon = cookie_netfs_data; + const struct cifs_tcon *tcon = cookie_netfs_data; if (datalen != sizeof(auxdata)) return FSCACHE_CHECKAUX_OBSOLETE; @@ -302,7 +302,7 @@ static void cifs_fscache_inode_now_uncached(void *cookie_netfs_data) pagevec_init(&pvec, 0); first = 0; - cFYI(1, "cifs inode 0x%p now uncached", cifsi); + cFYI(1, "%s: cifs inode 0x%p now uncached", __func__, cifsi); for (;;) { nr_pages = pagevec_lookup(&pvec, diff --git a/fs/cifs/cifs_debug.c b/fs/cifs/cifs_debug.c index 18f4272d9047..2fe3cf13b2e9 100644 --- a/fs/cifs/cifs_debug.c +++ b/fs/cifs/cifs_debug.c @@ -110,8 +110,8 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v) struct list_head *tmp1, *tmp2, *tmp3; struct mid_q_entry *mid_entry; struct TCP_Server_Info *server; - struct cifsSesInfo *ses; - struct cifsTconInfo *tcon; + struct cifs_ses *ses; + struct cifs_tcon *tcon; int i, j; __u32 dev_type; @@ -152,7 +152,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v) tcp_ses_list); i++; list_for_each(tmp2, &server->smb_ses_list) { - ses = list_entry(tmp2, struct cifsSesInfo, + ses = list_entry(tmp2, struct cifs_ses, smb_ses_list); if ((ses->serverDomain == NULL) || (ses->serverOS == NULL) || @@ -171,7 +171,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v) seq_printf(m, "TCP status: %d\n\tLocal Users To " "Server: %d SecMode: 0x%x Req On Wire: %d", server->tcpStatus, server->srv_count, - server->secMode, + server->sec_mode, atomic_read(&server->inFlight)); #ifdef CONFIG_CIFS_STATS2 @@ -183,7 +183,7 @@ static int cifs_debug_data_proc_show(struct seq_file *m, void *v) seq_puts(m, "\n\tShares:"); j = 0; list_for_each(tmp3, &ses->tcon_list) { - tcon = list_entry(tmp3, struct cifsTconInfo, + tcon = list_entry(tmp3, struct cifs_tcon, tcon_list); ++j; dev_type = le32_to_cpu(tcon->fsDevInfo.DeviceType); @@ -256,8 +256,8 @@ static ssize_t cifs_stats_proc_write(struct file *file, int rc; struct list_head *tmp1, *tmp2, *tmp3; struct TCP_Server_Info *server; - struct cifsSesInfo *ses; - struct cifsTconInfo *tcon; + struct cifs_ses *ses; + struct cifs_tcon *tcon; rc = get_user(c, buffer); if (rc) @@ -273,11 +273,11 @@ static ssize_t cifs_stats_proc_write(struct file *file, server = list_entry(tmp1, struct TCP_Server_Info, tcp_ses_list); list_for_each(tmp2, &server->smb_ses_list) { - ses = list_entry(tmp2, struct cifsSesInfo, + ses = list_entry(tmp2, struct cifs_ses, smb_ses_list); list_for_each(tmp3, &ses->tcon_list) { tcon = list_entry(tmp3, - struct cifsTconInfo, + struct cifs_tcon, tcon_list); atomic_set(&tcon->num_smbs_sent, 0); atomic_set(&tcon->num_writes, 0); @@ -312,8 +312,8 @@ static int cifs_stats_proc_show(struct seq_file *m, void *v) int i; struct list_head *tmp1, *tmp2, *tmp3; struct TCP_Server_Info *server; - struct cifsSesInfo *ses; - struct cifsTconInfo *tcon; + struct cifs_ses *ses; + struct cifs_tcon *tcon; seq_printf(m, "Resources in use\nCIFS Session: %d\n", @@ -346,11 +346,11 @@ static int cifs_stats_proc_show(struct seq_file *m, void *v) server = list_entry(tmp1, struct TCP_Server_Info, tcp_ses_list); list_for_each(tmp2, &server->smb_ses_list) { - ses = list_entry(tmp2, struct cifsSesInfo, + ses = list_entry(tmp2, struct cifs_ses, smb_ses_list); list_for_each(tmp3, &ses->tcon_list) { tcon = list_entry(tmp3, - struct cifsTconInfo, + struct cifs_tcon, tcon_list); i++; seq_printf(m, "\n%d) %s", i, tcon->treeName); diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c index 2b68ac57d97d..8d8f28c94c0f 100644 --- a/fs/cifs/cifs_dfs_ref.c +++ b/fs/cifs/cifs_dfs_ref.c @@ -272,7 +272,7 @@ static struct vfsmount *cifs_dfs_do_automount(struct dentry *mntpt) struct dfs_info3_param *referrals = NULL; unsigned int num_referrals = 0; struct cifs_sb_info *cifs_sb; - struct cifsSesInfo *ses; + struct cifs_ses *ses; char *full_path; int xid, i; int rc; diff --git a/fs/cifs/cifs_fs_sb.h b/fs/cifs/cifs_fs_sb.h index a9d5692e0c20..7260e11e21f8 100644 --- a/fs/cifs/cifs_fs_sb.h +++ b/fs/cifs/cifs_fs_sb.h @@ -41,6 +41,8 @@ #define CIFS_MOUNT_MF_SYMLINKS 0x10000 /* Minshall+French Symlinks enabled */ #define CIFS_MOUNT_MULTIUSER 0x20000 /* multiuser mount */ #define CIFS_MOUNT_STRICT_IO 0x40000 /* strict cache mode */ +#define CIFS_MOUNT_RWPIDFORWARD 0x80000 /* use pid forwarding for rw */ +#define CIFS_MOUNT_POSIXACL 0x100000 /* mirror of MS_POSIXACL in mnt_cifs_flags */ struct cifs_sb_info { struct rb_root tlink_tree; @@ -56,8 +58,6 @@ struct cifs_sb_info { mode_t mnt_file_mode; mode_t mnt_dir_mode; unsigned int mnt_cifs_flags; - int prepathlen; - char *prepath; /* relative path under the share to mount to */ char *mountdata; /* options received at mount time or via DFS refs */ struct backing_dev_info bdi; struct delayed_work prune_tlinks; diff --git a/fs/cifs/cifs_spnego.c b/fs/cifs/cifs_spnego.c index 33d221394aca..2272fd5fe5b7 100644 --- a/fs/cifs/cifs_spnego.c +++ b/fs/cifs/cifs_spnego.c @@ -95,7 +95,7 @@ struct key_type cifs_spnego_key_type = { /* get a key struct with a SPNEGO security blob, suitable for session setup */ struct key * -cifs_get_spnego_key(struct cifsSesInfo *sesInfo) +cifs_get_spnego_key(struct cifs_ses *sesInfo) { struct TCP_Server_Info *server = sesInfo->server; struct sockaddr_in *sa = (struct sockaddr_in *) &server->dstaddr; diff --git a/fs/cifs/cifs_spnego.h b/fs/cifs/cifs_spnego.h index e4041ec4d712..31bef9ee078b 100644 --- a/fs/cifs/cifs_spnego.h +++ b/fs/cifs/cifs_spnego.h @@ -41,7 +41,7 @@ struct cifs_spnego_msg { #ifdef __KERNEL__ extern struct key_type cifs_spnego_key_type; -extern struct key *cifs_get_spnego_key(struct cifsSesInfo *sesInfo); +extern struct key *cifs_get_spnego_key(struct cifs_ses *sesInfo); #endif /* KERNEL */ #endif /* _CIFS_SPNEGO_H */ diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c index f3c6fb9942ac..21de1d6d5849 100644 --- a/fs/cifs/cifsacl.c +++ b/fs/cifs/cifsacl.c @@ -38,7 +38,7 @@ static const struct cifs_sid sid_everyone = { 1, 1, {0, 0, 0, 0, 0, 1}, {0} }; /* security id for Authenticated Users system group */ static const struct cifs_sid sid_authusers = { - 1, 1, {0, 0, 0, 0, 0, 5}, {11} }; + 1, 1, {0, 0, 0, 0, 0, 5}, {__constant_cpu_to_le32(11)} }; /* group users */ static const struct cifs_sid sid_user = {1, 2 , {0, 0, 0, 0, 0, 5}, {} }; @@ -74,8 +74,9 @@ shrink_idmap_tree(struct rb_root *root, int nr_to_scan, int *nr_rem, * Run idmap cache shrinker. */ static int -cifs_idmap_shrinker(struct shrinker *shrink, int nr_to_scan, gfp_t gfp_mask) +cifs_idmap_shrinker(struct shrinker *shrink, struct shrink_control *sc) { + int nr_to_scan = sc->nr_to_scan; int nr_del = 0; int nr_rem = 0; struct rb_root *root; @@ -458,7 +459,8 @@ int compare_sids(const struct cifs_sid *ctsid, const struct cifs_sid *cwsid) if (num_subauth) { for (i = 0; i < num_subauth; ++i) { if (ctsid->sub_auth[i] != cwsid->sub_auth[i]) { - if (ctsid->sub_auth[i] > cwsid->sub_auth[i]) + if (le32_to_cpu(ctsid->sub_auth[i]) > + le32_to_cpu(cwsid->sub_auth[i])) return 1; else return -1; @@ -945,7 +947,7 @@ static struct cifs_ntsd *get_cifs_acl_by_path(struct cifs_sb_info *cifs_sb, int oplock = 0; int xid, rc; __u16 fid; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct tcon_link *tlink = cifs_sb_tlink(cifs_sb); if (IS_ERR(tlink)) @@ -1013,7 +1015,7 @@ static int set_cifs_acl_by_path(struct cifs_sb_info *cifs_sb, const char *path, int oplock = 0; int xid, rc; __u16 fid; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct tcon_link *tlink = cifs_sb_tlink(cifs_sb); if (IS_ERR(tlink)) diff --git a/fs/cifs/cifsencrypt.c b/fs/cifs/cifsencrypt.c index 45c3f78c8f81..5a0ee7f2af06 100644 --- a/fs/cifs/cifsencrypt.c +++ b/fs/cifs/cifsencrypt.c @@ -184,7 +184,7 @@ int cifs_verify_signature(struct smb_hdr *cifs_pdu, if (cifs_pdu == NULL || server == NULL) return -EINVAL; - if (cifs_pdu->Command == SMB_COM_NEGOTIATE) + if (!server->session_estab) return 0; if (cifs_pdu->Command == SMB_COM_LOCKING_ANDX) { @@ -229,7 +229,7 @@ int cifs_verify_signature(struct smb_hdr *cifs_pdu, } /* first calculate 24 bytes ntlm response and then 16 byte session key */ -int setup_ntlm_response(struct cifsSesInfo *ses) +int setup_ntlm_response(struct cifs_ses *ses) { int rc = 0; unsigned int temp_len = CIFS_SESS_KEY_SIZE + CIFS_AUTH_RESP_SIZE; @@ -312,7 +312,7 @@ int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt, * Allocate domain name which gets freed when session struct is deallocated. */ static int -build_avpair_blob(struct cifsSesInfo *ses, const struct nls_table *nls_cp) +build_avpair_blob(struct cifs_ses *ses, const struct nls_table *nls_cp) { unsigned int dlen; unsigned int wlen; @@ -400,7 +400,7 @@ build_avpair_blob(struct cifsSesInfo *ses, const struct nls_table *nls_cp) * about target string i.e. for some, just user name might suffice. */ static int -find_domain_name(struct cifsSesInfo *ses, const struct nls_table *nls_cp) +find_domain_name(struct cifs_ses *ses, const struct nls_table *nls_cp) { unsigned int attrsize; unsigned int type; @@ -445,7 +445,7 @@ find_domain_name(struct cifsSesInfo *ses, const struct nls_table *nls_cp) return 0; } -static int calc_ntlmv2_hash(struct cifsSesInfo *ses, char *ntlmv2_hash, +static int calc_ntlmv2_hash(struct cifs_ses *ses, char *ntlmv2_hash, const struct nls_table *nls_cp) { int rc = 0; @@ -527,7 +527,7 @@ calc_exit_2: } static int -CalcNTLMv2_response(const struct cifsSesInfo *ses, char *ntlmv2_hash) +CalcNTLMv2_response(const struct cifs_ses *ses, char *ntlmv2_hash) { int rc; unsigned int offset = CIFS_SESS_KEY_SIZE + 8; @@ -563,7 +563,7 @@ CalcNTLMv2_response(const struct cifsSesInfo *ses, char *ntlmv2_hash) int -setup_ntlmv2_rsp(struct cifsSesInfo *ses, const struct nls_table *nls_cp) +setup_ntlmv2_rsp(struct cifs_ses *ses, const struct nls_table *nls_cp) { int rc; int baselen; @@ -649,7 +649,7 @@ setup_ntlmv2_rsp_ret: } int -calc_seckey(struct cifsSesInfo *ses) +calc_seckey(struct cifs_ses *ses) { int rc; struct crypto_blkcipher *tfm_arc4; diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index 493b74ca5648..35f9154615fa 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -104,52 +104,24 @@ cifs_sb_deactive(struct super_block *sb) } static int -cifs_read_super(struct super_block *sb, void *data, - const char *devname, int silent) +cifs_read_super(struct super_block *sb) { struct inode *inode; struct cifs_sb_info *cifs_sb; int rc = 0; - /* BB should we make this contingent on mount parm? */ - sb->s_flags |= MS_NODIRATIME | MS_NOATIME; - sb->s_fs_info = kzalloc(sizeof(struct cifs_sb_info), GFP_KERNEL); cifs_sb = CIFS_SB(sb); - if (cifs_sb == NULL) - return -ENOMEM; - spin_lock_init(&cifs_sb->tlink_tree_lock); - cifs_sb->tlink_tree = RB_ROOT; + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIXACL) + sb->s_flags |= MS_POSIXACL; - rc = bdi_setup_and_register(&cifs_sb->bdi, "cifs", BDI_CAP_MAP_COPY); - if (rc) { - kfree(cifs_sb); - return rc; - } - cifs_sb->bdi.ra_pages = default_backing_dev_info.ra_pages; - - /* - * Copy mount params to sb for use in submounts. Better to do - * the copy here and deal with the error before cleanup gets - * complicated post-mount. - */ - if (data) { - cifs_sb->mountdata = kstrndup(data, PAGE_SIZE, GFP_KERNEL); - if (cifs_sb->mountdata == NULL) { - bdi_destroy(&cifs_sb->bdi); - kfree(sb->s_fs_info); - sb->s_fs_info = NULL; - return -ENOMEM; - } - } + if (cifs_sb_master_tcon(cifs_sb)->ses->capabilities & CAP_LARGE_FILES) + sb->s_maxbytes = MAX_LFS_FILESIZE; + else + sb->s_maxbytes = MAX_NON_LFS; - rc = cifs_mount(sb, cifs_sb, devname); - - if (rc) { - if (!silent) - cERROR(1, "cifs_mount failed w/return code = %d", rc); - goto out_mount_failed; - } + /* BB FIXME fix time_gran to be larger for LANMAN sessions */ + sb->s_time_gran = 100; sb->s_magic = CIFS_MAGIC_NUMBER; sb->s_op = &cifs_super_ops; @@ -191,45 +163,14 @@ out_no_root: if (inode) iput(inode); - cifs_umount(sb, cifs_sb); - -out_mount_failed: - if (cifs_sb) { - if (cifs_sb->mountdata) { - kfree(cifs_sb->mountdata); - cifs_sb->mountdata = NULL; - } - unload_nls(cifs_sb->local_nls); - bdi_destroy(&cifs_sb->bdi); - kfree(cifs_sb); - } return rc; } -static void -cifs_put_super(struct super_block *sb) +static void cifs_kill_sb(struct super_block *sb) { - int rc = 0; - struct cifs_sb_info *cifs_sb; - - cFYI(1, "In cifs_put_super"); - cifs_sb = CIFS_SB(sb); - if (cifs_sb == NULL) { - cFYI(1, "Empty cifs superblock info passed to unmount"); - return; - } - - rc = cifs_umount(sb, cifs_sb); - if (rc) - cERROR(1, "cifs_umount failed with return code %d", rc); - if (cifs_sb->mountdata) { - kfree(cifs_sb->mountdata); - cifs_sb->mountdata = NULL; - } - - unload_nls(cifs_sb->local_nls); - bdi_destroy(&cifs_sb->bdi); - kfree(cifs_sb); + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); + kill_anon_super(sb); + cifs_umount(cifs_sb); } static int @@ -237,7 +178,7 @@ cifs_statfs(struct dentry *dentry, struct kstatfs *buf) { struct super_block *sb = dentry->d_sb; struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); int rc = -EOPNOTSUPP; int xid; @@ -286,9 +227,6 @@ static int cifs_permission(struct inode *inode, int mask, unsigned int flags) { struct cifs_sb_info *cifs_sb; - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - cifs_sb = CIFS_SB(inode->i_sb); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_PERM) { @@ -381,6 +319,37 @@ cifs_show_address(struct seq_file *s, struct TCP_Server_Info *server) } } +static void +cifs_show_security(struct seq_file *s, struct TCP_Server_Info *server) +{ + seq_printf(s, ",sec="); + + switch (server->secType) { + case LANMAN: + seq_printf(s, "lanman"); + break; + case NTLMv2: + seq_printf(s, "ntlmv2"); + break; + case NTLM: + seq_printf(s, "ntlm"); + break; + case Kerberos: + seq_printf(s, "krb5"); + break; + case RawNTLMSSP: + seq_printf(s, "ntlmssp"); + break; + default: + /* shouldn't ever happen */ + seq_printf(s, "unknown"); + break; + } + + if (server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) + seq_printf(s, "i"); +} + /* * cifs_show_options() is for displaying mount options in /proc/mounts. * Not all settable options are displayed but most of the important @@ -390,10 +359,12 @@ static int cifs_show_options(struct seq_file *s, struct vfsmount *m) { struct cifs_sb_info *cifs_sb = CIFS_SB(m->mnt_sb); - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); struct sockaddr *srcaddr; srcaddr = (struct sockaddr *)&tcon->ses->server->srcaddr; + cifs_show_security(s, tcon->ses->server); + seq_printf(s, ",unc=%s", tcon->treeName); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_MULTIUSER) @@ -444,14 +415,20 @@ cifs_show_options(struct seq_file *s, struct vfsmount *m) seq_printf(s, ",nocase"); if (tcon->retry) seq_printf(s, ",hard"); - if (cifs_sb->prepath) - seq_printf(s, ",prepath=%s", cifs_sb->prepath); + if (tcon->unix_ext) + seq_printf(s, ",unix"); + else + seq_printf(s, ",nounix"); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) seq_printf(s, ",posixpaths"); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SET_UID) seq_printf(s, ",setuids"); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_SERVER_INUM) seq_printf(s, ",serverino"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + seq_printf(s, ",rwpidforward"); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NOPOSIXBRL) + seq_printf(s, ",forcemand"); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_DIRECT_IO) seq_printf(s, ",directio"); if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_NO_XATTR) @@ -484,7 +461,7 @@ cifs_show_options(struct seq_file *s, struct vfsmount *m) static void cifs_umount_begin(struct super_block *sb) { struct cifs_sb_info *cifs_sb = CIFS_SB(sb); - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; if (cifs_sb == NULL) return; @@ -541,7 +518,6 @@ static int cifs_drop_inode(struct inode *inode) } static const struct super_operations cifs_super_ops = { - .put_super = cifs_put_super, .statfs = cifs_statfs, .alloc_inode = cifs_alloc_inode, .destroy_inode = cifs_destroy_inode, @@ -559,29 +535,194 @@ static const struct super_operations cifs_super_ops = { #endif }; +/* + * Get root dentry from superblock according to prefix path mount option. + * Return dentry with refcount + 1 on success and NULL otherwise. + */ +static struct dentry * +cifs_get_root(struct smb_vol *vol, struct super_block *sb) +{ + int xid, rc; + struct inode *inode; + struct qstr name; + struct dentry *dparent = NULL, *dchild = NULL, *alias; + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); + unsigned int i, full_len, len; + char *full_path = NULL, *pstart; + char sep; + + full_path = cifs_build_path_to_root(vol, cifs_sb, + cifs_sb_master_tcon(cifs_sb)); + if (full_path == NULL) + return ERR_PTR(-ENOMEM); + + cFYI(1, "Get root dentry for %s", full_path); + + xid = GetXid(); + sep = CIFS_DIR_SEP(cifs_sb); + dparent = dget(sb->s_root); + full_len = strlen(full_path); + full_path[full_len] = sep; + pstart = full_path + 1; + + for (i = 1, len = 0; i <= full_len; i++) { + if (full_path[i] != sep || !len) { + len++; + continue; + } + + full_path[i] = 0; + cFYI(1, "get dentry for %s", pstart); + + name.name = pstart; + name.len = len; + name.hash = full_name_hash(pstart, len); + dchild = d_lookup(dparent, &name); + if (dchild == NULL) { + cFYI(1, "not exists"); + dchild = d_alloc(dparent, &name); + if (dchild == NULL) { + dput(dparent); + dparent = ERR_PTR(-ENOMEM); + goto out; + } + } + + cFYI(1, "get inode"); + if (dchild->d_inode == NULL) { + cFYI(1, "not exists"); + inode = NULL; + if (cifs_sb_master_tcon(CIFS_SB(sb))->unix_ext) + rc = cifs_get_inode_info_unix(&inode, full_path, + sb, xid); + else + rc = cifs_get_inode_info(&inode, full_path, + NULL, sb, xid, NULL); + if (rc) { + dput(dchild); + dput(dparent); + dparent = ERR_PTR(rc); + goto out; + } + alias = d_materialise_unique(dchild, inode); + if (alias != NULL) { + dput(dchild); + if (IS_ERR(alias)) { + dput(dparent); + dparent = ERR_PTR(-EINVAL); /* XXX */ + goto out; + } + dchild = alias; + } + } + cFYI(1, "parent %p, child %p", dparent, dchild); + + dput(dparent); + dparent = dchild; + len = 0; + pstart = full_path + i + 1; + full_path[i] = sep; + } +out: + _FreeXid(xid); + kfree(full_path); + return dparent; +} + +static int cifs_set_super(struct super_block *sb, void *data) +{ + struct cifs_mnt_data *mnt_data = data; + sb->s_fs_info = mnt_data->cifs_sb; + return set_anon_super(sb, NULL); +} + static struct dentry * cifs_do_mount(struct file_system_type *fs_type, - int flags, const char *dev_name, void *data) + int flags, const char *dev_name, void *data) { int rc; struct super_block *sb; - - sb = sget(fs_type, NULL, set_anon_super, NULL); + struct cifs_sb_info *cifs_sb; + struct smb_vol *volume_info; + struct cifs_mnt_data mnt_data; + struct dentry *root; cFYI(1, "Devname: %s flags: %d ", dev_name, flags); - if (IS_ERR(sb)) - return ERR_CAST(sb); + rc = cifs_setup_volume_info(&volume_info, (char *)data, dev_name); + if (rc) + return ERR_PTR(rc); + + cifs_sb = kzalloc(sizeof(struct cifs_sb_info), GFP_KERNEL); + if (cifs_sb == NULL) { + root = ERR_PTR(-ENOMEM); + goto out_nls; + } + + cifs_sb->mountdata = kstrndup(data, PAGE_SIZE, GFP_KERNEL); + if (cifs_sb->mountdata == NULL) { + root = ERR_PTR(-ENOMEM); + goto out_cifs_sb; + } - sb->s_flags = flags; + cifs_setup_cifs_sb(volume_info, cifs_sb); - rc = cifs_read_super(sb, data, dev_name, flags & MS_SILENT ? 1 : 0); + rc = cifs_mount(cifs_sb, volume_info); if (rc) { - deactivate_locked_super(sb); - return ERR_PTR(rc); + if (!(flags & MS_SILENT)) + cERROR(1, "cifs_mount failed w/return code = %d", rc); + root = ERR_PTR(rc); + goto out_mountdata; + } + + mnt_data.vol = volume_info; + mnt_data.cifs_sb = cifs_sb; + mnt_data.flags = flags; + + sb = sget(fs_type, cifs_match_super, cifs_set_super, &mnt_data); + if (IS_ERR(sb)) { + root = ERR_CAST(sb); + cifs_umount(cifs_sb); + goto out; + } + + if (sb->s_root) { + cFYI(1, "Use existing superblock"); + cifs_umount(cifs_sb); + } else { + sb->s_flags = flags; + /* BB should we make this contingent on mount parm? */ + sb->s_flags |= MS_NODIRATIME | MS_NOATIME; + + rc = cifs_read_super(sb); + if (rc) { + root = ERR_PTR(rc); + goto out_super; + } + + sb->s_flags |= MS_ACTIVE; } - sb->s_flags |= MS_ACTIVE; - return dget(sb->s_root); + + root = cifs_get_root(volume_info, sb); + if (IS_ERR(root)) + goto out_super; + + cFYI(1, "dentry root is: %p", root); + goto out; + +out_super: + deactivate_locked_super(sb); +out: + cifs_cleanup_volume_info(&volume_info); + return root; + +out_mountdata: + kfree(cifs_sb->mountdata); +out_cifs_sb: + kfree(cifs_sb); +out_nls: + unload_nls(volume_info->local_nls); + goto out; } static ssize_t cifs_file_aio_write(struct kiocb *iocb, const struct iovec *iov, @@ -670,7 +811,7 @@ struct file_system_type cifs_fs_type = { .owner = THIS_MODULE, .name = "cifs", .mount = cifs_do_mount, - .kill_sb = kill_anon_super, + .kill_sb = cifs_kill_sb, /* .fs_flags */ }; const struct inode_operations cifs_dir_inode_ops = { diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h index 64313f778ebf..0900e1658c96 100644 --- a/fs/cifs/cifsfs.h +++ b/fs/cifs/cifsfs.h @@ -129,5 +129,5 @@ extern long cifs_ioctl(struct file *filep, unsigned int cmd, unsigned long arg); extern const struct export_operations cifs_export_ops; #endif /* CIFS_NFSD_EXPORT */ -#define CIFS_VERSION "1.72" +#define CIFS_VERSION "1.73" #endif /* _CIFSFS_H */ diff --git a/fs/cifs/cifsglob.h b/fs/cifs/cifsglob.h index 76b4517e74b0..6255fa812c7a 100644 --- a/fs/cifs/cifsglob.h +++ b/fs/cifs/cifsglob.h @@ -155,6 +155,81 @@ struct cifs_cred { ***************************************************************** */ +struct smb_vol { + char *username; + char *password; + char *domainname; + char *UNC; + char *UNCip; + char *iocharset; /* local code page for mapping to and from Unicode */ + char source_rfc1001_name[RFC1001_NAME_LEN_WITH_NULL]; /* clnt nb name */ + char target_rfc1001_name[RFC1001_NAME_LEN_WITH_NULL]; /* srvr nb name */ + uid_t cred_uid; + uid_t linux_uid; + gid_t linux_gid; + mode_t file_mode; + mode_t dir_mode; + unsigned secFlg; + bool retry:1; + bool intr:1; + bool setuids:1; + bool override_uid:1; + bool override_gid:1; + bool dynperm:1; + bool noperm:1; + bool no_psx_acl:1; /* set if posix acl support should be disabled */ + bool cifs_acl:1; + bool no_xattr:1; /* set if xattr (EA) support should be disabled*/ + bool server_ino:1; /* use inode numbers from server ie UniqueId */ + bool direct_io:1; + bool strict_io:1; /* strict cache behavior */ + bool remap:1; /* set to remap seven reserved chars in filenames */ + bool posix_paths:1; /* unset to not ask for posix pathnames. */ + bool no_linux_ext:1; + bool sfu_emul:1; + bool nullauth:1; /* attempt to authenticate with null user */ + bool nocase:1; /* request case insensitive filenames */ + bool nobrl:1; /* disable sending byte range locks to srv */ + bool mand_lock:1; /* send mandatory not posix byte range lock reqs */ + bool seal:1; /* request transport encryption on share */ + bool nodfs:1; /* Do not request DFS, even if available */ + bool local_lease:1; /* check leases only on local system, not remote */ + bool noblocksnd:1; + bool noautotune:1; + bool nostrictsync:1; /* do not force expensive SMBflush on every sync */ + bool fsc:1; /* enable fscache */ + bool mfsymlinks:1; /* use Minshall+French Symlinks */ + bool multiuser:1; + bool rwpidforward:1; /* pid forward for read/write operations */ + unsigned int rsize; + unsigned int wsize; + bool sockopt_tcp_nodelay:1; + unsigned short int port; + unsigned long actimeo; /* attribute cache timeout (jiffies) */ + char *prepath; + struct sockaddr_storage srcaddr; /* allow binding to a local IP */ + struct nls_table *local_nls; +}; + +#define CIFS_MOUNT_MASK (CIFS_MOUNT_NO_PERM | CIFS_MOUNT_SET_UID | \ + CIFS_MOUNT_SERVER_INUM | CIFS_MOUNT_DIRECT_IO | \ + CIFS_MOUNT_NO_XATTR | CIFS_MOUNT_MAP_SPECIAL_CHR | \ + CIFS_MOUNT_UNX_EMUL | CIFS_MOUNT_NO_BRL | \ + CIFS_MOUNT_CIFS_ACL | CIFS_MOUNT_OVERR_UID | \ + CIFS_MOUNT_OVERR_GID | CIFS_MOUNT_DYNPERM | \ + CIFS_MOUNT_NOPOSIXBRL | CIFS_MOUNT_NOSSYNC | \ + CIFS_MOUNT_FSCACHE | CIFS_MOUNT_MF_SYMLINKS | \ + CIFS_MOUNT_MULTIUSER | CIFS_MOUNT_STRICT_IO) + +#define CIFS_MS_MASK (MS_RDONLY | MS_MANDLOCK | MS_NOEXEC | MS_NOSUID | \ + MS_NODEV | MS_SYNCHRONOUS) + +struct cifs_mnt_data { + struct cifs_sb_info *cifs_sb; + struct smb_vol *vol; + int flags; +}; + struct TCP_Server_Info { struct list_head tcp_ses_list; struct list_head smb_ses_list; @@ -179,7 +254,7 @@ struct TCP_Server_Info { struct mutex srv_mutex; struct task_struct *tsk; char server_GUID[16]; - char secMode; + char sec_mode; bool session_estab; /* mark when very first sess is established */ u16 dialect; /* dialect index that server chose */ enum securityEnum secType; @@ -254,7 +329,7 @@ static inline void cifs_set_net_ns(struct TCP_Server_Info *srv, struct net *net) /* * Session structure. One of these for each uid session with a particular host */ -struct cifsSesInfo { +struct cifs_ses { struct list_head smb_ses_list; struct list_head tcon_list; struct mutex session_mutex; @@ -294,11 +369,11 @@ struct cifsSesInfo { * there is one of these for each connection to a resource on a particular * session */ -struct cifsTconInfo { +struct cifs_tcon { struct list_head tcon_list; int tc_count; struct list_head openFileList; - struct cifsSesInfo *ses; /* pointer to session associated with */ + struct cifs_ses *ses; /* pointer to session associated with */ char treeName[MAX_TREE_SIZE + 1]; /* UNC name of resource in ASCII */ char *nativeFileSystem; char *password; /* for share-level security */ @@ -380,12 +455,12 @@ struct tcon_link { #define TCON_LINK_IN_TREE 2 unsigned long tl_time; atomic_t tl_count; - struct cifsTconInfo *tl_tcon; + struct cifs_tcon *tl_tcon; }; extern struct tcon_link *cifs_sb_tlink(struct cifs_sb_info *cifs_sb); -static inline struct cifsTconInfo * +static inline struct cifs_tcon * tlink_tcon(struct tcon_link *tlink) { return tlink->tl_tcon; @@ -402,7 +477,7 @@ cifs_get_tlink(struct tcon_link *tlink) } /* This function is always expected to succeed */ -extern struct cifsTconInfo *cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb); +extern struct cifs_tcon *cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb); /* * This info hangs off the cifsFileInfo structure, pointed to by llist. @@ -455,6 +530,14 @@ struct cifsFileInfo { struct work_struct oplock_break; /* work for oplock breaks */ }; +struct cifs_io_parms { + __u16 netfid; + __u32 pid; + __u64 offset; + unsigned int length; + struct cifs_tcon *tcon; +}; + /* * Take a reference on the file private data. Must be called with * cifs_file_list_lock held. @@ -509,10 +592,30 @@ static inline char CIFS_DIR_SEP(const struct cifs_sb_info *cifs_sb) return '\\'; } +static inline void +convert_delimiter(char *path, char delim) +{ + int i; + char old_delim; + + if (path == NULL) + return; + + if (delim == '/') + old_delim = '\\'; + else + old_delim = '/'; + + for (i = 0; path[i] != '\0'; i++) { + if (path[i] == old_delim) + path[i] = delim; + } +} + #ifdef CONFIG_CIFS_STATS #define cifs_stats_inc atomic_inc -static inline void cifs_stats_bytes_written(struct cifsTconInfo *tcon, +static inline void cifs_stats_bytes_written(struct cifs_tcon *tcon, unsigned int bytes) { if (bytes) { @@ -522,7 +625,7 @@ static inline void cifs_stats_bytes_written(struct cifsTconInfo *tcon, } } -static inline void cifs_stats_bytes_read(struct cifsTconInfo *tcon, +static inline void cifs_stats_bytes_read(struct cifs_tcon *tcon, unsigned int bytes) { spin_lock(&tcon->stat_lock); @@ -543,9 +646,8 @@ struct mid_q_entry; * This is the prototype for the mid callback function. When creating one, * take special care to avoid deadlocks. Things to bear in mind: * - * - it will be called by cifsd - * - the GlobalMid_Lock will be held - * - the mid will be removed from the pending_mid_q list + * - it will be called by cifsd, with no locks held + * - the mid will be removed from any lists */ typedef void (mid_callback_t)(struct mid_q_entry *mid); @@ -573,7 +675,7 @@ struct mid_q_entry { struct oplock_q_entry { struct list_head qhead; struct inode *pinode; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; __u16 netfid; }; @@ -656,6 +758,7 @@ static inline void free_dfs_info_array(struct dfs_info3_param *param, #define MID_RESPONSE_RECEIVED 4 #define MID_RETRY_NEEDED 8 /* session closed while this request out */ #define MID_RESPONSE_MALFORMED 0x10 +#define MID_SHUTDOWN 0x20 /* Types of response buffer returned from SendReceive2 */ #define CIFS_NO_BUFFER 0 /* Response buffer not returned */ diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index 6e69e06a30b3..257f312ede42 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -57,8 +57,9 @@ extern int init_cifs_idmap(void); extern void exit_cifs_idmap(void); extern void cifs_destroy_idmaptrees(void); extern char *build_path_from_dentry(struct dentry *); -extern char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb, - struct cifsTconInfo *tcon); +extern char *cifs_build_path_to_root(struct smb_vol *vol, + struct cifs_sb_info *cifs_sb, + struct cifs_tcon *tcon); extern char *build_wildcard_path_from_dentry(struct dentry *direntry); extern char *cifs_compose_mount_options(const char *sb_mountdata, const char *fullpath, const struct dfs_info3_param *ref, @@ -67,20 +68,22 @@ extern char *cifs_compose_mount_options(const char *sb_mountdata, extern struct mid_q_entry *AllocMidQEntry(const struct smb_hdr *smb_buffer, struct TCP_Server_Info *server); extern void DeleteMidQEntry(struct mid_q_entry *midEntry); -extern int cifs_call_async(struct TCP_Server_Info *server, - struct smb_hdr *in_buf, mid_callback_t *callback, - void *cbdata); -extern int SendReceive(const unsigned int /* xid */ , struct cifsSesInfo *, +extern int cifs_call_async(struct TCP_Server_Info *server, struct kvec *iov, + unsigned int nvec, mid_callback_t *callback, + void *cbdata, bool ignore_pend); +extern int SendReceive(const unsigned int /* xid */ , struct cifs_ses *, struct smb_hdr * /* input */ , struct smb_hdr * /* out */ , int * /* bytes returned */ , const int long_op); -extern int SendReceiveNoRsp(const unsigned int xid, struct cifsSesInfo *ses, +extern int SendReceiveNoRsp(const unsigned int xid, struct cifs_ses *ses, struct smb_hdr *in_buf, int flags); -extern int SendReceive2(const unsigned int /* xid */ , struct cifsSesInfo *, +extern int cifs_check_receive(struct mid_q_entry *mid, + struct TCP_Server_Info *server, bool log_error); +extern int SendReceive2(const unsigned int /* xid */ , struct cifs_ses *, struct kvec *, int /* nvec to send */, int * /* type of buf returned */ , const int flags); extern int SendReceiveBlockingLock(const unsigned int xid, - struct cifsTconInfo *ptcon, + struct cifs_tcon *ptcon, struct smb_hdr *in_buf , struct smb_hdr *out_buf, int *bytes_returned); @@ -99,14 +102,14 @@ extern int cifs_convert_address(struct sockaddr *dst, const char *src, int len); extern int cifs_set_port(struct sockaddr *addr, const unsigned short int port); extern int cifs_fill_sockaddr(struct sockaddr *dst, const char *src, int len, const unsigned short int port); -extern int map_smb_to_linux_error(struct smb_hdr *smb, int logErr); +extern int map_smb_to_linux_error(struct smb_hdr *smb, bool logErr); extern void header_assemble(struct smb_hdr *, char /* command */ , - const struct cifsTconInfo *, int /* length of + const struct cifs_tcon *, int /* length of fixed section (word count) in two byte units */); extern int small_smb_init_no_tc(const int smb_cmd, const int wct, - struct cifsSesInfo *ses, + struct cifs_ses *ses, void **request_buf); -extern int CIFS_SessSetup(unsigned int xid, struct cifsSesInfo *ses, +extern int CIFS_SessSetup(unsigned int xid, struct cifs_ses *ses, const struct nls_table *nls_cp); extern __u16 GetNextMid(struct TCP_Server_Info *server); extern struct timespec cifs_NTtimeToUnix(__le64 utc_nanoseconds_since_1601); @@ -148,102 +151,108 @@ extern struct cifs_ntsd *get_cifs_acl(struct cifs_sb_info *, struct inode *, extern int set_cifs_acl(struct cifs_ntsd *, __u32, struct inode *, const char *); -extern int cifs_mount(struct super_block *, struct cifs_sb_info *, - const char *); -extern int cifs_umount(struct super_block *, struct cifs_sb_info *); +extern void cifs_setup_cifs_sb(struct smb_vol *pvolume_info, + struct cifs_sb_info *cifs_sb); +extern int cifs_match_super(struct super_block *, void *); +extern void cifs_cleanup_volume_info(struct smb_vol **pvolume_info); +extern int cifs_setup_volume_info(struct smb_vol **pvolume_info, + char *mount_data, const char *devname); +extern int cifs_mount(struct cifs_sb_info *, struct smb_vol *); +extern void cifs_umount(struct cifs_sb_info *); extern void cifs_dfs_release_automount_timer(void); void cifs_proc_init(void); void cifs_proc_clean(void); extern int cifs_negotiate_protocol(unsigned int xid, - struct cifsSesInfo *ses); -extern int cifs_setup_session(unsigned int xid, struct cifsSesInfo *ses, + struct cifs_ses *ses); +extern int cifs_setup_session(unsigned int xid, struct cifs_ses *ses, struct nls_table *nls_info); -extern int CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses); +extern int CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses); -extern int CIFSTCon(unsigned int xid, struct cifsSesInfo *ses, - const char *tree, struct cifsTconInfo *tcon, +extern int CIFSTCon(unsigned int xid, struct cifs_ses *ses, + const char *tree, struct cifs_tcon *tcon, const struct nls_table *); -extern int CIFSFindFirst(const int xid, struct cifsTconInfo *tcon, +extern int CIFSFindFirst(const int xid, struct cifs_tcon *tcon, const char *searchName, const struct nls_table *nls_codepage, __u16 *searchHandle, struct cifs_search_info *psrch_inf, int map, const char dirsep); -extern int CIFSFindNext(const int xid, struct cifsTconInfo *tcon, +extern int CIFSFindNext(const int xid, struct cifs_tcon *tcon, __u16 searchHandle, struct cifs_search_info *psrch_inf); -extern int CIFSFindClose(const int, struct cifsTconInfo *tcon, +extern int CIFSFindClose(const int, struct cifs_tcon *tcon, const __u16 search_handle); -extern int CIFSSMBQFileInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBQFileInfo(const int xid, struct cifs_tcon *tcon, u16 netfid, FILE_ALL_INFO *pFindData); -extern int CIFSSMBQPathInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBQPathInfo(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, FILE_ALL_INFO *findData, int legacy /* whether to use old info level */, const struct nls_table *nls_codepage, int remap); -extern int SMBQueryInformation(const int xid, struct cifsTconInfo *tcon, +extern int SMBQueryInformation(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, FILE_ALL_INFO *findData, const struct nls_table *nls_codepage, int remap); -extern int CIFSSMBUnixQFileInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBUnixQFileInfo(const int xid, struct cifs_tcon *tcon, u16 netfid, FILE_UNIX_BASIC_INFO *pFindData); extern int CIFSSMBUnixQPathInfo(const int xid, - struct cifsTconInfo *tcon, + struct cifs_tcon *tcon, const unsigned char *searchName, FILE_UNIX_BASIC_INFO *pFindData, const struct nls_table *nls_codepage, int remap); -extern int CIFSGetDFSRefer(const int xid, struct cifsSesInfo *ses, +extern int CIFSGetDFSRefer(const int xid, struct cifs_ses *ses, const unsigned char *searchName, struct dfs_info3_param **target_nodes, unsigned int *number_of_nodes_in_array, const struct nls_table *nls_codepage, int remap); -extern int get_dfs_path(int xid, struct cifsSesInfo *pSesInfo, +extern int get_dfs_path(int xid, struct cifs_ses *pSesInfo, const char *old_path, const struct nls_table *nls_codepage, unsigned int *pnum_referrals, struct dfs_info3_param **preferrals, int remap); -extern void reset_cifs_unix_caps(int xid, struct cifsTconInfo *tcon, - struct super_block *sb, struct smb_vol *vol); -extern int CIFSSMBQFSInfo(const int xid, struct cifsTconInfo *tcon, +extern void reset_cifs_unix_caps(int xid, struct cifs_tcon *tcon, + struct cifs_sb_info *cifs_sb, + struct smb_vol *vol); +extern int CIFSSMBQFSInfo(const int xid, struct cifs_tcon *tcon, struct kstatfs *FSData); -extern int SMBOldQFSInfo(const int xid, struct cifsTconInfo *tcon, +extern int SMBOldQFSInfo(const int xid, struct cifs_tcon *tcon, struct kstatfs *FSData); -extern int CIFSSMBSetFSUnixInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetFSUnixInfo(const int xid, struct cifs_tcon *tcon, __u64 cap); extern int CIFSSMBQFSAttributeInfo(const int xid, - struct cifsTconInfo *tcon); -extern int CIFSSMBQFSDeviceInfo(const int xid, struct cifsTconInfo *tcon); -extern int CIFSSMBQFSUnixInfo(const int xid, struct cifsTconInfo *tcon); -extern int CIFSSMBQFSPosixInfo(const int xid, struct cifsTconInfo *tcon, + struct cifs_tcon *tcon); +extern int CIFSSMBQFSDeviceInfo(const int xid, struct cifs_tcon *tcon); +extern int CIFSSMBQFSUnixInfo(const int xid, struct cifs_tcon *tcon); +extern int CIFSSMBQFSPosixInfo(const int xid, struct cifs_tcon *tcon, struct kstatfs *FSData); -extern int CIFSSMBSetPathInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetPathInfo(const int xid, struct cifs_tcon *tcon, const char *fileName, const FILE_BASIC_INFO *data, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBSetFileInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetFileInfo(const int xid, struct cifs_tcon *tcon, const FILE_BASIC_INFO *data, __u16 fid, __u32 pid_of_opener); -extern int CIFSSMBSetFileDisposition(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetFileDisposition(const int xid, struct cifs_tcon *tcon, bool delete_file, __u16 fid, __u32 pid_of_opener); #if 0 -extern int CIFSSMBSetAttrLegacy(int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetAttrLegacy(int xid, struct cifs_tcon *tcon, char *fileName, __u16 dos_attributes, const struct nls_table *nls_codepage); #endif /* possibly unneeded function */ -extern int CIFSSMBSetEOF(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetEOF(const int xid, struct cifs_tcon *tcon, const char *fileName, __u64 size, bool setAllocationSizeFlag, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBSetFileSize(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetFileSize(const int xid, struct cifs_tcon *tcon, __u64 size, __u16 fileHandle, __u32 opener_pid, bool AllocSizeFlag); @@ -257,120 +266,116 @@ struct cifs_unix_set_info_args { dev_t device; }; -extern int CIFSSMBUnixSetFileInfo(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBUnixSetFileInfo(const int xid, struct cifs_tcon *tcon, const struct cifs_unix_set_info_args *args, u16 fid, u32 pid_of_opener); -extern int CIFSSMBUnixSetPathInfo(const int xid, struct cifsTconInfo *pTcon, +extern int CIFSSMBUnixSetPathInfo(const int xid, struct cifs_tcon *pTcon, char *fileName, const struct cifs_unix_set_info_args *args, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBMkDir(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBMkDir(const int xid, struct cifs_tcon *tcon, const char *newName, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBRmDir(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBRmDir(const int xid, struct cifs_tcon *tcon, const char *name, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSPOSIXDelFile(const int xid, struct cifsTconInfo *tcon, +extern int CIFSPOSIXDelFile(const int xid, struct cifs_tcon *tcon, const char *name, __u16 type, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBDelFile(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBDelFile(const int xid, struct cifs_tcon *tcon, const char *name, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBRename(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBRename(const int xid, struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBRenameOpenFile(const int xid, struct cifsTconInfo *pTcon, +extern int CIFSSMBRenameOpenFile(const int xid, struct cifs_tcon *pTcon, int netfid, const char *target_name, const struct nls_table *nls_codepage, int remap_special_chars); extern int CIFSCreateHardLink(const int xid, - struct cifsTconInfo *tcon, + struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap_special_chars); extern int CIFSUnixCreateHardLink(const int xid, - struct cifsTconInfo *tcon, + struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap_special_chars); extern int CIFSUnixCreateSymLink(const int xid, - struct cifsTconInfo *tcon, + struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage); extern int CIFSSMBUnixQuerySymLink(const int xid, - struct cifsTconInfo *tcon, + struct cifs_tcon *tcon, const unsigned char *searchName, char **syminfo, const struct nls_table *nls_codepage); #ifdef CONFIG_CIFS_SYMLINK_EXPERIMENTAL extern int CIFSSMBQueryReparseLinkInfo(const int xid, - struct cifsTconInfo *tcon, + struct cifs_tcon *tcon, const unsigned char *searchName, char *symlinkinfo, const int buflen, __u16 fid, const struct nls_table *nls_codepage); #endif /* temporarily unused until cifs_symlink fixed */ -extern int CIFSSMBOpen(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBOpen(const int xid, struct cifs_tcon *tcon, const char *fileName, const int disposition, const int access_flags, const int omode, __u16 *netfid, int *pOplock, FILE_ALL_INFO *, const struct nls_table *nls_codepage, int remap); -extern int SMBLegacyOpen(const int xid, struct cifsTconInfo *tcon, +extern int SMBLegacyOpen(const int xid, struct cifs_tcon *tcon, const char *fileName, const int disposition, const int access_flags, const int omode, __u16 *netfid, int *pOplock, FILE_ALL_INFO *, const struct nls_table *nls_codepage, int remap); -extern int CIFSPOSIXCreate(const int xid, struct cifsTconInfo *tcon, +extern int CIFSPOSIXCreate(const int xid, struct cifs_tcon *tcon, u32 posix_flags, __u64 mode, __u16 *netfid, FILE_UNIX_BASIC_INFO *pRetData, __u32 *pOplock, const char *name, const struct nls_table *nls_codepage, int remap); -extern int CIFSSMBClose(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBClose(const int xid, struct cifs_tcon *tcon, const int smb_file_id); -extern int CIFSSMBFlush(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBFlush(const int xid, struct cifs_tcon *tcon, const int smb_file_id); -extern int CIFSSMBRead(const int xid, struct cifsTconInfo *tcon, - const int netfid, unsigned int count, - const __u64 lseek, unsigned int *nbytes, char **buf, +extern int CIFSSMBRead(const int xid, struct cifs_io_parms *io_parms, + unsigned int *nbytes, char **buf, int *return_buf_type); -extern int CIFSSMBWrite(const int xid, struct cifsTconInfo *tcon, - const int netfid, const unsigned int count, - const __u64 lseek, unsigned int *nbytes, - const char *buf, const char __user *ubuf, +extern int CIFSSMBWrite(const int xid, struct cifs_io_parms *io_parms, + unsigned int *nbytes, const char *buf, + const char __user *ubuf, const int long_op); +extern int CIFSSMBWrite2(const int xid, struct cifs_io_parms *io_parms, + unsigned int *nbytes, struct kvec *iov, const int nvec, const int long_op); -extern int CIFSSMBWrite2(const int xid, struct cifsTconInfo *tcon, - const int netfid, const unsigned int count, - const __u64 offset, unsigned int *nbytes, - struct kvec *iov, const int nvec, const int long_op); -extern int CIFSGetSrvInodeNumber(const int xid, struct cifsTconInfo *tcon, +extern int CIFSGetSrvInodeNumber(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, __u64 *inode_number, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBLock(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBLock(const int xid, struct cifs_tcon *tcon, const __u16 netfid, const __u64 len, const __u64 offset, const __u32 numUnlock, const __u32 numLock, const __u8 lockType, const bool waitFlag, const __u8 oplock_level); -extern int CIFSSMBPosixLock(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBPosixLock(const int xid, struct cifs_tcon *tcon, const __u16 smb_file_id, const int get_flag, const __u64 len, struct file_lock *, const __u16 lock_type, const bool waitFlag); -extern int CIFSSMBTDis(const int xid, struct cifsTconInfo *tcon); +extern int CIFSSMBTDis(const int xid, struct cifs_tcon *tcon); extern int CIFSSMBEcho(struct TCP_Server_Info *server); -extern int CIFSSMBLogoff(const int xid, struct cifsSesInfo *ses); +extern int CIFSSMBLogoff(const int xid, struct cifs_ses *ses); -extern struct cifsSesInfo *sesInfoAlloc(void); -extern void sesInfoFree(struct cifsSesInfo *); -extern struct cifsTconInfo *tconInfoAlloc(void); -extern void tconInfoFree(struct cifsTconInfo *); +extern struct cifs_ses *sesInfoAlloc(void); +extern void sesInfoFree(struct cifs_ses *); +extern struct cifs_tcon *tconInfoAlloc(void); +extern void tconInfoFree(struct cifs_tcon *); extern int cifs_sign_smb(struct smb_hdr *, struct TCP_Server_Info *, __u32 *); extern int cifs_sign_smb2(struct kvec *iov, int n_vec, struct TCP_Server_Info *, @@ -379,51 +384,51 @@ extern int cifs_verify_signature(struct smb_hdr *, struct TCP_Server_Info *server, __u32 expected_sequence_number); extern int SMBNTencrypt(unsigned char *, unsigned char *, unsigned char *); -extern int setup_ntlm_response(struct cifsSesInfo *); -extern int setup_ntlmv2_rsp(struct cifsSesInfo *, const struct nls_table *); +extern int setup_ntlm_response(struct cifs_ses *); +extern int setup_ntlmv2_rsp(struct cifs_ses *, const struct nls_table *); extern int cifs_crypto_shash_allocate(struct TCP_Server_Info *); extern void cifs_crypto_shash_release(struct TCP_Server_Info *); -extern int calc_seckey(struct cifsSesInfo *); +extern int calc_seckey(struct cifs_ses *); #ifdef CONFIG_CIFS_WEAK_PW_HASH extern int calc_lanman_hash(const char *password, const char *cryptkey, bool encrypt, char *lnm_session_key); #endif /* CIFS_WEAK_PW_HASH */ #ifdef CONFIG_CIFS_DNOTIFY_EXPERIMENTAL /* unused temporarily */ -extern int CIFSSMBNotify(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBNotify(const int xid, struct cifs_tcon *tcon, const int notify_subdirs, const __u16 netfid, __u32 filter, struct file *file, int multishot, const struct nls_table *nls_codepage); #endif /* was needed for dnotify, and will be needed for inotify when VFS fix */ extern int CIFSSMBCopy(int xid, - struct cifsTconInfo *source_tcon, + struct cifs_tcon *source_tcon, const char *fromName, const __u16 target_tid, const char *toName, const int flags, const struct nls_table *nls_codepage, int remap_special_chars); -extern ssize_t CIFSSMBQAllEAs(const int xid, struct cifsTconInfo *tcon, +extern ssize_t CIFSSMBQAllEAs(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, const unsigned char *ea_name, char *EAData, size_t bufsize, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBSetEA(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetEA(const int xid, struct cifs_tcon *tcon, const char *fileName, const char *ea_name, const void *ea_value, const __u16 ea_value_len, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBGetCIFSACL(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBGetCIFSACL(const int xid, struct cifs_tcon *tcon, __u16 fid, struct cifs_ntsd **acl_inf, __u32 *buflen); -extern int CIFSSMBSetCIFSACL(const int, struct cifsTconInfo *, __u16, +extern int CIFSSMBSetCIFSACL(const int, struct cifs_tcon *, __u16, struct cifs_ntsd *, __u32); -extern int CIFSSMBGetPosixACL(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBGetPosixACL(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, char *acl_inf, const int buflen, const int acl_type, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSSMBSetPosixACL(const int xid, struct cifsTconInfo *tcon, +extern int CIFSSMBSetPosixACL(const int xid, struct cifs_tcon *tcon, const unsigned char *fileName, const char *local_acl, const int buflen, const int acl_type, const struct nls_table *nls_codepage, int remap_special_chars); -extern int CIFSGetExtAttr(const int xid, struct cifsTconInfo *tcon, +extern int CIFSGetExtAttr(const int xid, struct cifs_tcon *tcon, const int netfid, __u64 *pExtAttrBits, __u64 *pMask); extern void cifs_autodisable_serverino(struct cifs_sb_info *cifs_sb); extern bool CIFSCouldBeMFSymlink(const struct cifs_fattr *fattr); @@ -434,4 +439,22 @@ extern int mdfour(unsigned char *, unsigned char *, int); extern int E_md4hash(const unsigned char *passwd, unsigned char *p16); extern int SMBencrypt(unsigned char *passwd, const unsigned char *c8, unsigned char *p24); + +/* asynchronous write support */ +struct cifs_writedata { + struct kref refcount; + enum writeback_sync_modes sync_mode; + struct work_struct work; + struct cifsFileInfo *cfile; + __u64 offset; + unsigned int bytes; + int result; + unsigned int nr_pages; + struct page *pages[1]; +}; + +int cifs_async_writev(struct cifs_writedata *wdata); +struct cifs_writedata *cifs_writedata_alloc(unsigned int nr_pages); +void cifs_writedata_release(struct kref *refcount); + #endif /* _CIFSPROTO_H */ diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c index 83df937b814e..1a9fe7f816d1 100644 --- a/fs/cifs/cifssmb.c +++ b/fs/cifs/cifssmb.c @@ -32,6 +32,7 @@ #include <linux/vfs.h> #include <linux/slab.h> #include <linux/posix_acl_xattr.h> +#include <linux/pagemap.h> #include <asm/uaccess.h> #include "cifspdu.h" #include "cifsglob.h" @@ -84,7 +85,7 @@ static struct { /* Mark as invalid, all open files on tree connections since they were closed when session to server was lost */ -static void mark_open_files_invalid(struct cifsTconInfo *pTcon) +static void mark_open_files_invalid(struct cifs_tcon *pTcon) { struct cifsFileInfo *open_file = NULL; struct list_head *tmp; @@ -104,10 +105,10 @@ static void mark_open_files_invalid(struct cifsTconInfo *pTcon) /* reconnect the socket, tcon, and smb session if needed */ static int -cifs_reconnect_tcon(struct cifsTconInfo *tcon, int smb_command) +cifs_reconnect_tcon(struct cifs_tcon *tcon, int smb_command) { int rc = 0; - struct cifsSesInfo *ses; + struct cifs_ses *ses; struct TCP_Server_Info *server; struct nls_table *nls_codepage; @@ -226,7 +227,7 @@ out: SMB information in the SMB header. If the return code is zero, this function must have filled in request_buf pointer */ static int -small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, +small_smb_init(int smb_command, int wct, struct cifs_tcon *tcon, void **request_buf) { int rc; @@ -252,7 +253,7 @@ small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, int small_smb_init_no_tc(const int smb_command, const int wct, - struct cifsSesInfo *ses, void **request_buf) + struct cifs_ses *ses, void **request_buf) { int rc; struct smb_hdr *buffer; @@ -278,7 +279,7 @@ small_smb_init_no_tc(const int smb_command, const int wct, /* If the return code is zero, this function must fill in request_buf pointer */ static int -__smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, +__smb_init(int smb_command, int wct, struct cifs_tcon *tcon, void **request_buf, void **response_buf) { *request_buf = cifs_buf_get(); @@ -304,7 +305,7 @@ __smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, /* If the return code is zero, this function must fill in request_buf pointer */ static int -smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, +smb_init(int smb_command, int wct, struct cifs_tcon *tcon, void **request_buf, void **response_buf) { int rc; @@ -317,7 +318,7 @@ smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, } static int -smb_init_no_reconnect(int smb_command, int wct, struct cifsTconInfo *tcon, +smb_init_no_reconnect(int smb_command, int wct, struct cifs_tcon *tcon, void **request_buf, void **response_buf) { if (tcon->ses->need_reconnect || tcon->need_reconnect) @@ -366,7 +367,7 @@ static inline void inc_rfc1001_len(void *pSMB, int count) } int -CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) +CIFSSMBNegotiate(unsigned int xid, struct cifs_ses *ses) { NEGOTIATE_REQ *pSMB; NEGOTIATE_RSP *pSMBr; @@ -450,7 +451,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) rc = -EOPNOTSUPP; goto neg_err_exit; } - server->secMode = (__u8)le16_to_cpu(rsp->SecurityMode); + server->sec_mode = (__u8)le16_to_cpu(rsp->SecurityMode); server->maxReq = le16_to_cpu(rsp->MaxMpxCount); server->maxBuf = min((__u32)le16_to_cpu(rsp->MaxBufSize), (__u32)CIFSMaxBufSize + MAX_CIFS_HDR_SIZE); @@ -504,7 +505,7 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) cpu_to_le16(CIFS_CRYPTO_KEY_SIZE)) { memcpy(ses->server->cryptkey, rsp->EncryptionKey, CIFS_CRYPTO_KEY_SIZE); - } else if (server->secMode & SECMODE_PW_ENCRYPT) { + } else if (server->sec_mode & SECMODE_PW_ENCRYPT) { rc = -EIO; /* need cryptkey unless plain text */ goto neg_err_exit; } @@ -526,11 +527,11 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) goto neg_err_exit; } /* else wct == 17 NTLM */ - server->secMode = pSMBr->SecurityMode; - if ((server->secMode & SECMODE_USER) == 0) + server->sec_mode = pSMBr->SecurityMode; + if ((server->sec_mode & SECMODE_USER) == 0) cFYI(1, "share mode security"); - if ((server->secMode & SECMODE_PW_ENCRYPT) == 0) + if ((server->sec_mode & SECMODE_PW_ENCRYPT) == 0) #ifdef CONFIG_CIFS_WEAK_PW_HASH if ((secFlags & CIFSSEC_MAY_PLNTXT) == 0) #endif /* CIFS_WEAK_PW_HASH */ @@ -570,18 +571,10 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) if (pSMBr->EncryptionKeyLength == CIFS_CRYPTO_KEY_SIZE) { memcpy(ses->server->cryptkey, pSMBr->u.EncryptionKey, CIFS_CRYPTO_KEY_SIZE); - } else if ((pSMBr->hdr.Flags2 & SMBFLG2_EXT_SEC) - && (pSMBr->EncryptionKeyLength == 0)) { + } else if ((pSMBr->hdr.Flags2 & SMBFLG2_EXT_SEC || + server->capabilities & CAP_EXTENDED_SECURITY) && + (pSMBr->EncryptionKeyLength == 0)) { /* decode security blob */ - } else if (server->secMode & SECMODE_PW_ENCRYPT) { - rc = -EIO; /* no crypt key only if plain text pwd */ - goto neg_err_exit; - } - - /* BB might be helpful to save off the domain of server here */ - - if ((pSMBr->hdr.Flags2 & SMBFLG2_EXT_SEC) && - (server->capabilities & CAP_EXTENDED_SECURITY)) { count = get_bcc(&pSMBr->hdr); if (count < 16) { rc = -EIO; @@ -624,6 +617,9 @@ CIFSSMBNegotiate(unsigned int xid, struct cifsSesInfo *ses) } else rc = -EOPNOTSUPP; } + } else if (server->sec_mode & SECMODE_PW_ENCRYPT) { + rc = -EIO; /* no crypt key only if plain text pwd */ + goto neg_err_exit; } else server->capabilities &= ~CAP_EXTENDED_SECURITY; @@ -634,27 +630,27 @@ signing_check: /* MUST_SIGN already includes the MAY_SIGN FLAG so if this is zero it means that signing is disabled */ cFYI(1, "Signing disabled"); - if (server->secMode & SECMODE_SIGN_REQUIRED) { + if (server->sec_mode & SECMODE_SIGN_REQUIRED) { cERROR(1, "Server requires " "packet signing to be enabled in " "/proc/fs/cifs/SecurityFlags."); rc = -EOPNOTSUPP; } - server->secMode &= + server->sec_mode &= ~(SECMODE_SIGN_ENABLED | SECMODE_SIGN_REQUIRED); } else if ((secFlags & CIFSSEC_MUST_SIGN) == CIFSSEC_MUST_SIGN) { /* signing required */ cFYI(1, "Must sign - secFlags 0x%x", secFlags); - if ((server->secMode & + if ((server->sec_mode & (SECMODE_SIGN_ENABLED | SECMODE_SIGN_REQUIRED)) == 0) { cERROR(1, "signing required but server lacks support"); rc = -EOPNOTSUPP; } else - server->secMode |= SECMODE_SIGN_REQUIRED; + server->sec_mode |= SECMODE_SIGN_REQUIRED; } else { /* signing optional ie CIFSSEC_MAY_SIGN */ - if ((server->secMode & SECMODE_SIGN_REQUIRED) == 0) - server->secMode &= + if ((server->sec_mode & SECMODE_SIGN_REQUIRED) == 0) + server->sec_mode &= ~(SECMODE_SIGN_ENABLED | SECMODE_SIGN_REQUIRED); } @@ -666,7 +662,7 @@ neg_err_exit: } int -CIFSSMBTDis(const int xid, struct cifsTconInfo *tcon) +CIFSSMBTDis(const int xid, struct cifs_tcon *tcon) { struct smb_hdr *smb_buffer; int rc = 0; @@ -725,6 +721,7 @@ CIFSSMBEcho(struct TCP_Server_Info *server) { ECHO_REQ *smb; int rc = 0; + struct kvec iov; cFYI(1, "In echo request"); @@ -739,9 +736,10 @@ CIFSSMBEcho(struct TCP_Server_Info *server) put_bcc(1, &smb->hdr); smb->Data[0] = 'a'; inc_rfc1001_len(smb, 3); + iov.iov_base = smb; + iov.iov_len = be32_to_cpu(smb->hdr.smb_buf_length) + 4; - rc = cifs_call_async(server, (struct smb_hdr *)smb, - cifs_echo_callback, server); + rc = cifs_call_async(server, &iov, 1, cifs_echo_callback, server, true); if (rc) cFYI(1, "Echo request failed: %d", rc); @@ -751,7 +749,7 @@ CIFSSMBEcho(struct TCP_Server_Info *server) } int -CIFSSMBLogoff(const int xid, struct cifsSesInfo *ses) +CIFSSMBLogoff(const int xid, struct cifs_ses *ses) { LOGOFF_ANDX_REQ *pSMB; int rc = 0; @@ -778,7 +776,7 @@ CIFSSMBLogoff(const int xid, struct cifsSesInfo *ses) pSMB->hdr.Mid = GetNextMid(ses->server); - if (ses->server->secMode & + if (ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) pSMB->hdr.Flags2 |= SMBFLG2_SECURITY_SIGNATURE; @@ -798,7 +796,7 @@ session_already_dead: } int -CIFSPOSIXDelFile(const int xid, struct cifsTconInfo *tcon, const char *fileName, +CIFSPOSIXDelFile(const int xid, struct cifs_tcon *tcon, const char *fileName, __u16 type, const struct nls_table *nls_codepage, int remap) { TRANSACTION2_SPI_REQ *pSMB = NULL; @@ -873,7 +871,7 @@ PsxDelete: } int -CIFSSMBDelFile(const int xid, struct cifsTconInfo *tcon, const char *fileName, +CIFSSMBDelFile(const int xid, struct cifs_tcon *tcon, const char *fileName, const struct nls_table *nls_codepage, int remap) { DELETE_FILE_REQ *pSMB = NULL; @@ -918,7 +916,7 @@ DelFileRetry: } int -CIFSSMBRmDir(const int xid, struct cifsTconInfo *tcon, const char *dirName, +CIFSSMBRmDir(const int xid, struct cifs_tcon *tcon, const char *dirName, const struct nls_table *nls_codepage, int remap) { DELETE_DIRECTORY_REQ *pSMB = NULL; @@ -961,7 +959,7 @@ RmDirRetry: } int -CIFSSMBMkDir(const int xid, struct cifsTconInfo *tcon, +CIFSSMBMkDir(const int xid, struct cifs_tcon *tcon, const char *name, const struct nls_table *nls_codepage, int remap) { int rc = 0; @@ -1004,7 +1002,7 @@ MkDirRetry: } int -CIFSPOSIXCreate(const int xid, struct cifsTconInfo *tcon, __u32 posix_flags, +CIFSPOSIXCreate(const int xid, struct cifs_tcon *tcon, __u32 posix_flags, __u64 mode, __u16 *netfid, FILE_UNIX_BASIC_INFO *pRetData, __u32 *pOplock, const char *name, const struct nls_table *nls_codepage, int remap) @@ -1170,7 +1168,7 @@ access_flags_to_smbopen_mode(const int access_flags) } int -SMBLegacyOpen(const int xid, struct cifsTconInfo *tcon, +SMBLegacyOpen(const int xid, struct cifs_tcon *tcon, const char *fileName, const int openDisposition, const int access_flags, const int create_options, __u16 *netfid, int *pOplock, FILE_ALL_INFO *pfile_info, @@ -1277,7 +1275,7 @@ OldOpenRetry: } int -CIFSSMBOpen(const int xid, struct cifsTconInfo *tcon, +CIFSSMBOpen(const int xid, struct cifs_tcon *tcon, const char *fileName, const int openDisposition, const int access_flags, const int create_options, __u16 *netfid, int *pOplock, FILE_ALL_INFO *pfile_info, @@ -1379,8 +1377,7 @@ openRetry: } int -CIFSSMBRead(const int xid, struct cifsTconInfo *tcon, const int netfid, - const unsigned int count, const __u64 lseek, unsigned int *nbytes, +CIFSSMBRead(const int xid, struct cifs_io_parms *io_parms, unsigned int *nbytes, char **buf, int *pbuf_type) { int rc = -EACCES; @@ -1390,13 +1387,18 @@ CIFSSMBRead(const int xid, struct cifsTconInfo *tcon, const int netfid, int wct; int resp_buf_type = 0; struct kvec iov[1]; + __u32 pid = io_parms->pid; + __u16 netfid = io_parms->netfid; + __u64 offset = io_parms->offset; + struct cifs_tcon *tcon = io_parms->tcon; + unsigned int count = io_parms->length; cFYI(1, "Reading %d bytes on fid %d", count, netfid); if (tcon->ses->capabilities & CAP_LARGE_FILES) wct = 12; else { wct = 10; /* old style read */ - if ((lseek >> 32) > 0) { + if ((offset >> 32) > 0) { /* can not handle this big offset for old */ return -EIO; } @@ -1407,15 +1409,18 @@ CIFSSMBRead(const int xid, struct cifsTconInfo *tcon, const int netfid, if (rc) return rc; + pSMB->hdr.Pid = cpu_to_le16((__u16)pid); + pSMB->hdr.PidHigh = cpu_to_le16((__u16)(pid >> 16)); + /* tcon and ses pointer are checked in smb_init */ if (tcon->ses->server == NULL) return -ECONNABORTED; pSMB->AndXCommand = 0xFF; /* none */ pSMB->Fid = netfid; - pSMB->OffsetLow = cpu_to_le32(lseek & 0xFFFFFFFF); + pSMB->OffsetLow = cpu_to_le32(offset & 0xFFFFFFFF); if (wct == 12) - pSMB->OffsetHigh = cpu_to_le32(lseek >> 32); + pSMB->OffsetHigh = cpu_to_le32(offset >> 32); pSMB->Remaining = 0; pSMB->MaxCount = cpu_to_le16(count & 0xFFFF); @@ -1484,9 +1489,8 @@ CIFSSMBRead(const int xid, struct cifsTconInfo *tcon, const int netfid, int -CIFSSMBWrite(const int xid, struct cifsTconInfo *tcon, - const int netfid, const unsigned int count, - const __u64 offset, unsigned int *nbytes, const char *buf, +CIFSSMBWrite(const int xid, struct cifs_io_parms *io_parms, + unsigned int *nbytes, const char *buf, const char __user *ubuf, const int long_op) { int rc = -EACCES; @@ -1495,6 +1499,11 @@ CIFSSMBWrite(const int xid, struct cifsTconInfo *tcon, int bytes_returned, wct; __u32 bytes_sent; __u16 byte_count; + __u32 pid = io_parms->pid; + __u16 netfid = io_parms->netfid; + __u64 offset = io_parms->offset; + struct cifs_tcon *tcon = io_parms->tcon; + unsigned int count = io_parms->length; *nbytes = 0; @@ -1516,6 +1525,10 @@ CIFSSMBWrite(const int xid, struct cifsTconInfo *tcon, (void **) &pSMBr); if (rc) return rc; + + pSMB->hdr.Pid = cpu_to_le16((__u16)pid); + pSMB->hdr.PidHigh = cpu_to_le16((__u16)(pid >> 16)); + /* tcon and ses pointer are checked in smb_init */ if (tcon->ses->server == NULL) return -ECONNABORTED; @@ -1602,17 +1615,259 @@ CIFSSMBWrite(const int xid, struct cifsTconInfo *tcon, return rc; } +void +cifs_writedata_release(struct kref *refcount) +{ + struct cifs_writedata *wdata = container_of(refcount, + struct cifs_writedata, refcount); + + if (wdata->cfile) + cifsFileInfo_put(wdata->cfile); + + kfree(wdata); +} + +/* + * Write failed with a retryable error. Resend the write request. It's also + * possible that the page was redirtied so re-clean the page. + */ +static void +cifs_writev_requeue(struct cifs_writedata *wdata) +{ + int i, rc; + struct inode *inode = wdata->cfile->dentry->d_inode; + + for (i = 0; i < wdata->nr_pages; i++) { + lock_page(wdata->pages[i]); + clear_page_dirty_for_io(wdata->pages[i]); + } + + do { + rc = cifs_async_writev(wdata); + } while (rc == -EAGAIN); + + for (i = 0; i < wdata->nr_pages; i++) { + if (rc != 0) + SetPageError(wdata->pages[i]); + unlock_page(wdata->pages[i]); + } + + mapping_set_error(inode->i_mapping, rc); + kref_put(&wdata->refcount, cifs_writedata_release); +} + +static void +cifs_writev_complete(struct work_struct *work) +{ + struct cifs_writedata *wdata = container_of(work, + struct cifs_writedata, work); + struct inode *inode = wdata->cfile->dentry->d_inode; + int i = 0; + + if (wdata->result == 0) { + cifs_update_eof(CIFS_I(inode), wdata->offset, wdata->bytes); + cifs_stats_bytes_written(tlink_tcon(wdata->cfile->tlink), + wdata->bytes); + } else if (wdata->sync_mode == WB_SYNC_ALL && wdata->result == -EAGAIN) + return cifs_writev_requeue(wdata); + + for (i = 0; i < wdata->nr_pages; i++) { + struct page *page = wdata->pages[i]; + if (wdata->result == -EAGAIN) + __set_page_dirty_nobuffers(page); + else if (wdata->result < 0) + SetPageError(page); + end_page_writeback(page); + page_cache_release(page); + } + if (wdata->result != -EAGAIN) + mapping_set_error(inode->i_mapping, wdata->result); + kref_put(&wdata->refcount, cifs_writedata_release); +} + +struct cifs_writedata * +cifs_writedata_alloc(unsigned int nr_pages) +{ + struct cifs_writedata *wdata; + + /* this would overflow */ + if (nr_pages == 0) { + cERROR(1, "%s: called with nr_pages == 0!", __func__); + return NULL; + } + + /* writedata + number of page pointers */ + wdata = kzalloc(sizeof(*wdata) + + sizeof(struct page *) * (nr_pages - 1), GFP_NOFS); + if (wdata != NULL) { + INIT_WORK(&wdata->work, cifs_writev_complete); + kref_init(&wdata->refcount); + } + return wdata; +} + +/* + * Check the midState and signature on received buffer (if any), and queue the + * workqueue completion task. + */ +static void +cifs_writev_callback(struct mid_q_entry *mid) +{ + struct cifs_writedata *wdata = mid->callback_data; + struct cifs_tcon *tcon = tlink_tcon(wdata->cfile->tlink); + unsigned int written; + WRITE_RSP *smb = (WRITE_RSP *)mid->resp_buf; + + switch (mid->midState) { + case MID_RESPONSE_RECEIVED: + wdata->result = cifs_check_receive(mid, tcon->ses->server, 0); + if (wdata->result != 0) + break; + + written = le16_to_cpu(smb->CountHigh); + written <<= 16; + written += le16_to_cpu(smb->Count); + /* + * Mask off high 16 bits when bytes written as returned + * by the server is greater than bytes requested by the + * client. OS/2 servers are known to set incorrect + * CountHigh values. + */ + if (written > wdata->bytes) + written &= 0xFFFF; + + if (written < wdata->bytes) + wdata->result = -ENOSPC; + else + wdata->bytes = written; + break; + case MID_REQUEST_SUBMITTED: + case MID_RETRY_NEEDED: + wdata->result = -EAGAIN; + break; + default: + wdata->result = -EIO; + break; + } + + queue_work(system_nrt_wq, &wdata->work); + DeleteMidQEntry(mid); + atomic_dec(&tcon->ses->server->inFlight); + wake_up(&tcon->ses->server->request_q); +} + +/* cifs_async_writev - send an async write, and set up mid to handle result */ +int +cifs_async_writev(struct cifs_writedata *wdata) +{ + int i, rc = -EACCES; + WRITE_REQ *smb = NULL; + int wct; + struct cifs_tcon *tcon = tlink_tcon(wdata->cfile->tlink); + struct inode *inode = wdata->cfile->dentry->d_inode; + struct kvec *iov = NULL; + + if (tcon->ses->capabilities & CAP_LARGE_FILES) { + wct = 14; + } else { + wct = 12; + if (wdata->offset >> 32 > 0) { + /* can not handle big offset for old srv */ + return -EIO; + } + } + + rc = small_smb_init(SMB_COM_WRITE_ANDX, wct, tcon, (void **)&smb); + if (rc) + goto async_writev_out; + + /* 1 iov per page + 1 for header */ + iov = kzalloc((wdata->nr_pages + 1) * sizeof(*iov), GFP_NOFS); + if (iov == NULL) { + rc = -ENOMEM; + goto async_writev_out; + } + + smb->hdr.Pid = cpu_to_le16((__u16)wdata->cfile->pid); + smb->hdr.PidHigh = cpu_to_le16((__u16)(wdata->cfile->pid >> 16)); + + smb->AndXCommand = 0xFF; /* none */ + smb->Fid = wdata->cfile->netfid; + smb->OffsetLow = cpu_to_le32(wdata->offset & 0xFFFFFFFF); + if (wct == 14) + smb->OffsetHigh = cpu_to_le32(wdata->offset >> 32); + smb->Reserved = 0xFFFFFFFF; + smb->WriteMode = 0; + smb->Remaining = 0; + + smb->DataOffset = + cpu_to_le16(offsetof(struct smb_com_write_req, Data) - 4); + + /* 4 for RFC1001 length + 1 for BCC */ + iov[0].iov_len = be32_to_cpu(smb->hdr.smb_buf_length) + 4 + 1; + iov[0].iov_base = smb; + + /* marshal up the pages into iov array */ + wdata->bytes = 0; + for (i = 0; i < wdata->nr_pages; i++) { + iov[i + 1].iov_len = min(inode->i_size - + page_offset(wdata->pages[i]), + (loff_t)PAGE_CACHE_SIZE); + iov[i + 1].iov_base = kmap(wdata->pages[i]); + wdata->bytes += iov[i + 1].iov_len; + } + + cFYI(1, "async write at %llu %u bytes", wdata->offset, wdata->bytes); + + smb->DataLengthLow = cpu_to_le16(wdata->bytes & 0xFFFF); + smb->DataLengthHigh = cpu_to_le16(wdata->bytes >> 16); + + if (wct == 14) { + inc_rfc1001_len(&smb->hdr, wdata->bytes + 1); + put_bcc(wdata->bytes + 1, &smb->hdr); + } else { + /* wct == 12 */ + struct smb_com_writex_req *smbw = + (struct smb_com_writex_req *)smb; + inc_rfc1001_len(&smbw->hdr, wdata->bytes + 5); + put_bcc(wdata->bytes + 5, &smbw->hdr); + iov[0].iov_len += 4; /* pad bigger by four bytes */ + } + + kref_get(&wdata->refcount); + rc = cifs_call_async(tcon->ses->server, iov, wdata->nr_pages + 1, + cifs_writev_callback, wdata, false); + + if (rc == 0) + cifs_stats_inc(&tcon->num_writes); + else + kref_put(&wdata->refcount, cifs_writedata_release); + + /* send is done, unmap pages */ + for (i = 0; i < wdata->nr_pages; i++) + kunmap(wdata->pages[i]); + +async_writev_out: + cifs_small_buf_release(smb); + kfree(iov); + return rc; +} + int -CIFSSMBWrite2(const int xid, struct cifsTconInfo *tcon, - const int netfid, const unsigned int count, - const __u64 offset, unsigned int *nbytes, struct kvec *iov, - int n_vec, const int long_op) +CIFSSMBWrite2(const int xid, struct cifs_io_parms *io_parms, + unsigned int *nbytes, struct kvec *iov, int n_vec, + const int long_op) { int rc = -EACCES; WRITE_REQ *pSMB = NULL; int wct; int smb_hdr_len; int resp_buf_type = 0; + __u32 pid = io_parms->pid; + __u16 netfid = io_parms->netfid; + __u64 offset = io_parms->offset; + struct cifs_tcon *tcon = io_parms->tcon; + unsigned int count = io_parms->length; *nbytes = 0; @@ -1630,6 +1885,10 @@ CIFSSMBWrite2(const int xid, struct cifsTconInfo *tcon, rc = small_smb_init(SMB_COM_WRITE_ANDX, wct, tcon, (void **) &pSMB); if (rc) return rc; + + pSMB->hdr.Pid = cpu_to_le16((__u16)pid); + pSMB->hdr.PidHigh = cpu_to_le16((__u16)(pid >> 16)); + /* tcon and ses pointer are checked in smb_init */ if (tcon->ses->server == NULL) return -ECONNABORTED; @@ -1705,7 +1964,7 @@ CIFSSMBWrite2(const int xid, struct cifsTconInfo *tcon, int -CIFSSMBLock(const int xid, struct cifsTconInfo *tcon, +CIFSSMBLock(const int xid, struct cifs_tcon *tcon, const __u16 smb_file_id, const __u64 len, const __u64 offset, const __u32 numUnlock, const __u32 numLock, const __u8 lockType, @@ -1775,7 +2034,7 @@ CIFSSMBLock(const int xid, struct cifsTconInfo *tcon, } int -CIFSSMBPosixLock(const int xid, struct cifsTconInfo *tcon, +CIFSSMBPosixLock(const int xid, struct cifs_tcon *tcon, const __u16 smb_file_id, const int get_flag, const __u64 len, struct file_lock *pLockData, const __u16 lock_type, const bool waitFlag) @@ -1913,7 +2172,7 @@ plk_err_exit: int -CIFSSMBClose(const int xid, struct cifsTconInfo *tcon, int smb_file_id) +CIFSSMBClose(const int xid, struct cifs_tcon *tcon, int smb_file_id) { int rc = 0; CLOSE_REQ *pSMB = NULL; @@ -1946,7 +2205,7 @@ CIFSSMBClose(const int xid, struct cifsTconInfo *tcon, int smb_file_id) } int -CIFSSMBFlush(const int xid, struct cifsTconInfo *tcon, int smb_file_id) +CIFSSMBFlush(const int xid, struct cifs_tcon *tcon, int smb_file_id) { int rc = 0; FLUSH_REQ *pSMB = NULL; @@ -1967,7 +2226,7 @@ CIFSSMBFlush(const int xid, struct cifsTconInfo *tcon, int smb_file_id) } int -CIFSSMBRename(const int xid, struct cifsTconInfo *tcon, +CIFSSMBRename(const int xid, struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap) { @@ -2034,7 +2293,7 @@ renameRetry: return rc; } -int CIFSSMBRenameOpenFile(const int xid, struct cifsTconInfo *pTcon, +int CIFSSMBRenameOpenFile(const int xid, struct cifs_tcon *pTcon, int netfid, const char *target_name, const struct nls_table *nls_codepage, int remap) { @@ -2114,7 +2373,7 @@ int CIFSSMBRenameOpenFile(const int xid, struct cifsTconInfo *pTcon, } int -CIFSSMBCopy(const int xid, struct cifsTconInfo *tcon, const char *fromName, +CIFSSMBCopy(const int xid, struct cifs_tcon *tcon, const char *fromName, const __u16 target_tid, const char *toName, const int flags, const struct nls_table *nls_codepage, int remap) { @@ -2182,7 +2441,7 @@ copyRetry: } int -CIFSUnixCreateSymLink(const int xid, struct cifsTconInfo *tcon, +CIFSUnixCreateSymLink(const int xid, struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage) { @@ -2271,7 +2530,7 @@ createSymLinkRetry: } int -CIFSUnixCreateHardLink(const int xid, struct cifsTconInfo *tcon, +CIFSUnixCreateHardLink(const int xid, struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap) { @@ -2356,7 +2615,7 @@ createHardLinkRetry: } int -CIFSCreateHardLink(const int xid, struct cifsTconInfo *tcon, +CIFSCreateHardLink(const int xid, struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap) { @@ -2428,7 +2687,7 @@ winCreateHardLinkRetry: } int -CIFSSMBUnixQuerySymLink(const int xid, struct cifsTconInfo *tcon, +CIFSSMBUnixQuerySymLink(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, char **symlinkinfo, const struct nls_table *nls_codepage) { @@ -2533,7 +2792,7 @@ querySymLinkRetry: * it is not compiled in by default until callers fixed up and more tested. */ int -CIFSSMBQueryReparseLinkInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBQueryReparseLinkInfo(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, char *symlinkinfo, const int buflen, __u16 fid, const struct nls_table *nls_codepage) @@ -2771,7 +3030,7 @@ static __u16 ACL_to_cifs_posix(char *parm_data, const char *pACL, } int -CIFSSMBGetPosixACL(const int xid, struct cifsTconInfo *tcon, +CIFSSMBGetPosixACL(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, char *acl_inf, const int buflen, const int acl_type, const struct nls_table *nls_codepage, int remap) @@ -2859,7 +3118,7 @@ queryAclRetry: } int -CIFSSMBSetPosixACL(const int xid, struct cifsTconInfo *tcon, +CIFSSMBSetPosixACL(const int xid, struct cifs_tcon *tcon, const unsigned char *fileName, const char *local_acl, const int buflen, const int acl_type, @@ -2939,7 +3198,7 @@ setACLerrorExit: /* BB fix tabs in this function FIXME BB */ int -CIFSGetExtAttr(const int xid, struct cifsTconInfo *tcon, +CIFSGetExtAttr(const int xid, struct cifs_tcon *tcon, const int netfid, __u64 *pExtAttrBits, __u64 *pMask) { int rc = 0; @@ -3032,7 +3291,7 @@ GetExtAttrOut: */ static int smb_init_nttransact(const __u16 sub_command, const int setup_count, - const int parm_len, struct cifsTconInfo *tcon, + const int parm_len, struct cifs_tcon *tcon, void **ret_buf) { int rc; @@ -3115,7 +3374,7 @@ validate_ntransact(char *buf, char **ppparm, char **ppdata, /* Get Security Descriptor (by handle) from remote server for a file or dir */ int -CIFSSMBGetCIFSACL(const int xid, struct cifsTconInfo *tcon, __u16 fid, +CIFSSMBGetCIFSACL(const int xid, struct cifs_tcon *tcon, __u16 fid, struct cifs_ntsd **acl_inf, __u32 *pbuflen) { int rc = 0; @@ -3207,7 +3466,7 @@ qsec_out: } int -CIFSSMBSetCIFSACL(const int xid, struct cifsTconInfo *tcon, __u16 fid, +CIFSSMBSetCIFSACL(const int xid, struct cifs_tcon *tcon, __u16 fid, struct cifs_ntsd *pntsd, __u32 acllen) { __u16 byte_count, param_count, data_count, param_offset, data_offset; @@ -3273,7 +3532,7 @@ setCifsAclRetry: /* Legacy Query Path Information call for lookup to old servers such as Win9x/WinME */ -int SMBQueryInformation(const int xid, struct cifsTconInfo *tcon, +int SMBQueryInformation(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, FILE_ALL_INFO *pFinfo, const struct nls_table *nls_codepage, int remap) @@ -3341,7 +3600,7 @@ QInfRetry: } int -CIFSSMBQFileInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBQFileInfo(const int xid, struct cifs_tcon *tcon, u16 netfid, FILE_ALL_INFO *pFindData) { struct smb_t2_qfi_req *pSMB = NULL; @@ -3408,7 +3667,7 @@ QFileInfoRetry: } int -CIFSSMBQPathInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBQPathInfo(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, FILE_ALL_INFO *pFindData, int legacy /* old style infolevel */, @@ -3509,7 +3768,7 @@ QPathInfoRetry: } int -CIFSSMBUnixQFileInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBUnixQFileInfo(const int xid, struct cifs_tcon *tcon, u16 netfid, FILE_UNIX_BASIC_INFO *pFindData) { struct smb_t2_qfi_req *pSMB = NULL; @@ -3578,7 +3837,7 @@ UnixQFileInfoRetry: } int -CIFSSMBUnixQPathInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBUnixQPathInfo(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, FILE_UNIX_BASIC_INFO *pFindData, const struct nls_table *nls_codepage, int remap) @@ -3664,7 +3923,7 @@ UnixQPathInfoRetry: /* xid, tcon, searchName and codepage are input parms, rest are returned */ int -CIFSFindFirst(const int xid, struct cifsTconInfo *tcon, +CIFSFindFirst(const int xid, struct cifs_tcon *tcon, const char *searchName, const struct nls_table *nls_codepage, __u16 *pnetfid, @@ -3812,7 +4071,7 @@ findFirstRetry: return rc; } -int CIFSFindNext(const int xid, struct cifsTconInfo *tcon, +int CIFSFindNext(const int xid, struct cifs_tcon *tcon, __u16 searchHandle, struct cifs_search_info *psrch_inf) { TRANSACTION2_FNEXT_REQ *pSMB = NULL; @@ -3950,7 +4209,7 @@ FNext2_err_exit: } int -CIFSFindClose(const int xid, struct cifsTconInfo *tcon, +CIFSFindClose(const int xid, struct cifs_tcon *tcon, const __u16 searchHandle) { int rc = 0; @@ -3982,7 +4241,7 @@ CIFSFindClose(const int xid, struct cifsTconInfo *tcon, } int -CIFSGetSrvInodeNumber(const int xid, struct cifsTconInfo *tcon, +CIFSGetSrvInodeNumber(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, __u64 *inode_number, const struct nls_table *nls_codepage, int remap) @@ -4184,7 +4443,7 @@ parse_DFS_referrals_exit: } int -CIFSGetDFSRefer(const int xid, struct cifsSesInfo *ses, +CIFSGetDFSRefer(const int xid, struct cifs_ses *ses, const unsigned char *searchName, struct dfs_info3_param **target_nodes, unsigned int *num_of_nodes, @@ -4233,7 +4492,7 @@ getDFSRetry: } if (ses->server) { - if (ses->server->secMode & + if (ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) pSMB->hdr.Flags2 |= SMBFLG2_SECURITY_SIGNATURE; } @@ -4298,7 +4557,7 @@ GetDFSRefExit: /* Query File System Info such as free space to old servers such as Win 9x */ int -SMBOldQFSInfo(const int xid, struct cifsTconInfo *tcon, struct kstatfs *FSData) +SMBOldQFSInfo(const int xid, struct cifs_tcon *tcon, struct kstatfs *FSData) { /* level 0x01 SMB_QUERY_FILE_SYSTEM_INFO */ TRANSACTION2_QFSI_REQ *pSMB = NULL; @@ -4377,7 +4636,7 @@ oldQFSInfoRetry: } int -CIFSSMBQFSInfo(const int xid, struct cifsTconInfo *tcon, struct kstatfs *FSData) +CIFSSMBQFSInfo(const int xid, struct cifs_tcon *tcon, struct kstatfs *FSData) { /* level 0x103 SMB_QUERY_FILE_SYSTEM_INFO */ TRANSACTION2_QFSI_REQ *pSMB = NULL; @@ -4456,7 +4715,7 @@ QFSInfoRetry: } int -CIFSSMBQFSAttributeInfo(const int xid, struct cifsTconInfo *tcon) +CIFSSMBQFSAttributeInfo(const int xid, struct cifs_tcon *tcon) { /* level 0x105 SMB_QUERY_FILE_SYSTEM_INFO */ TRANSACTION2_QFSI_REQ *pSMB = NULL; @@ -4526,7 +4785,7 @@ QFSAttributeRetry: } int -CIFSSMBQFSDeviceInfo(const int xid, struct cifsTconInfo *tcon) +CIFSSMBQFSDeviceInfo(const int xid, struct cifs_tcon *tcon) { /* level 0x104 SMB_QUERY_FILE_SYSTEM_INFO */ TRANSACTION2_QFSI_REQ *pSMB = NULL; @@ -4597,7 +4856,7 @@ QFSDeviceRetry: } int -CIFSSMBQFSUnixInfo(const int xid, struct cifsTconInfo *tcon) +CIFSSMBQFSUnixInfo(const int xid, struct cifs_tcon *tcon) { /* level 0x200 SMB_QUERY_CIFS_UNIX_INFO */ TRANSACTION2_QFSI_REQ *pSMB = NULL; @@ -4667,7 +4926,7 @@ QFSUnixRetry: } int -CIFSSMBSetFSUnixInfo(const int xid, struct cifsTconInfo *tcon, __u64 cap) +CIFSSMBSetFSUnixInfo(const int xid, struct cifs_tcon *tcon, __u64 cap) { /* level 0x200 SMB_SET_CIFS_UNIX_INFO */ TRANSACTION2_SETFSI_REQ *pSMB = NULL; @@ -4741,7 +5000,7 @@ SETFSUnixRetry: int -CIFSSMBQFSPosixInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBQFSPosixInfo(const int xid, struct cifs_tcon *tcon, struct kstatfs *FSData) { /* level 0x201 SMB_QUERY_CIFS_POSIX_INFO */ @@ -4834,7 +5093,7 @@ QFSPosixRetry: in Samba which this routine can run into */ int -CIFSSMBSetEOF(const int xid, struct cifsTconInfo *tcon, const char *fileName, +CIFSSMBSetEOF(const int xid, struct cifs_tcon *tcon, const char *fileName, __u64 size, bool SetAllocation, const struct nls_table *nls_codepage, int remap) { @@ -4923,7 +5182,7 @@ SetEOFRetry: } int -CIFSSMBSetFileSize(const int xid, struct cifsTconInfo *tcon, __u64 size, +CIFSSMBSetFileSize(const int xid, struct cifs_tcon *tcon, __u64 size, __u16 fid, __u32 pid_of_opener, bool SetAllocation) { struct smb_com_transaction2_sfi_req *pSMB = NULL; @@ -5005,7 +5264,7 @@ CIFSSMBSetFileSize(const int xid, struct cifsTconInfo *tcon, __u64 size, time and resort to the original setpathinfo level which takes the ancient DOS time format with 2 second granularity */ int -CIFSSMBSetFileInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBSetFileInfo(const int xid, struct cifs_tcon *tcon, const FILE_BASIC_INFO *data, __u16 fid, __u32 pid_of_opener) { struct smb_com_transaction2_sfi_req *pSMB = NULL; @@ -5067,7 +5326,7 @@ CIFSSMBSetFileInfo(const int xid, struct cifsTconInfo *tcon, } int -CIFSSMBSetFileDisposition(const int xid, struct cifsTconInfo *tcon, +CIFSSMBSetFileDisposition(const int xid, struct cifs_tcon *tcon, bool delete_file, __u16 fid, __u32 pid_of_opener) { struct smb_com_transaction2_sfi_req *pSMB = NULL; @@ -5123,7 +5382,7 @@ CIFSSMBSetFileDisposition(const int xid, struct cifsTconInfo *tcon, } int -CIFSSMBSetPathInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBSetPathInfo(const int xid, struct cifs_tcon *tcon, const char *fileName, const FILE_BASIC_INFO *data, const struct nls_table *nls_codepage, int remap) { @@ -5207,7 +5466,7 @@ SetTimesRetry: handling it anyway and NT4 was what we thought it would be needed for Do not delete it until we prove whether needed for Win9x though */ int -CIFSSMBSetAttrLegacy(int xid, struct cifsTconInfo *tcon, char *fileName, +CIFSSMBSetAttrLegacy(int xid, struct cifs_tcon *tcon, char *fileName, __u16 dos_attrs, const struct nls_table *nls_codepage) { SETATTR_REQ *pSMB = NULL; @@ -5295,7 +5554,7 @@ cifs_fill_unix_set_info(FILE_UNIX_BASIC_INFO *data_offset, } int -CIFSSMBUnixSetFileInfo(const int xid, struct cifsTconInfo *tcon, +CIFSSMBUnixSetFileInfo(const int xid, struct cifs_tcon *tcon, const struct cifs_unix_set_info_args *args, u16 fid, u32 pid_of_opener) { @@ -5358,7 +5617,7 @@ CIFSSMBUnixSetFileInfo(const int xid, struct cifsTconInfo *tcon, } int -CIFSSMBUnixSetPathInfo(const int xid, struct cifsTconInfo *tcon, char *fileName, +CIFSSMBUnixSetPathInfo(const int xid, struct cifs_tcon *tcon, char *fileName, const struct cifs_unix_set_info_args *args, const struct nls_table *nls_codepage, int remap) { @@ -5445,7 +5704,7 @@ setPermsRetry: * the data isn't copied to it, but the length is returned. */ ssize_t -CIFSSMBQAllEAs(const int xid, struct cifsTconInfo *tcon, +CIFSSMBQAllEAs(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, const unsigned char *ea_name, char *EAData, size_t buf_size, const struct nls_table *nls_codepage, int remap) @@ -5626,7 +5885,7 @@ QAllEAsOut: } int -CIFSSMBSetEA(const int xid, struct cifsTconInfo *tcon, const char *fileName, +CIFSSMBSetEA(const int xid, struct cifs_tcon *tcon, const char *fileName, const char *ea_name, const void *ea_value, const __u16 ea_value_len, const struct nls_table *nls_codepage, int remap) @@ -5753,7 +6012,7 @@ SetEARetry: * incompatible for network fs clients, we could instead simply * expose this config flag by adding a future cifs (and smb2) notify ioctl. */ -int CIFSSMBNotify(const int xid, struct cifsTconInfo *tcon, +int CIFSSMBNotify(const int xid, struct cifs_tcon *tcon, const int notify_subdirs, const __u16 netfid, __u32 filter, struct file *pfile, int multishot, const struct nls_table *nls_codepage) diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index da284e3cb653..7f540df52527 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -57,62 +57,6 @@ extern mempool_t *cifs_req_poolp; -struct smb_vol { - char *username; - char *password; - char *domainname; - char *UNC; - char *UNCip; - char *iocharset; /* local code page for mapping to and from Unicode */ - char source_rfc1001_name[RFC1001_NAME_LEN_WITH_NULL]; /* clnt nb name */ - char target_rfc1001_name[RFC1001_NAME_LEN_WITH_NULL]; /* srvr nb name */ - uid_t cred_uid; - uid_t linux_uid; - gid_t linux_gid; - mode_t file_mode; - mode_t dir_mode; - unsigned secFlg; - bool retry:1; - bool intr:1; - bool setuids:1; - bool override_uid:1; - bool override_gid:1; - bool dynperm:1; - bool noperm:1; - bool no_psx_acl:1; /* set if posix acl support should be disabled */ - bool cifs_acl:1; - bool no_xattr:1; /* set if xattr (EA) support should be disabled*/ - bool server_ino:1; /* use inode numbers from server ie UniqueId */ - bool direct_io:1; - bool strict_io:1; /* strict cache behavior */ - bool remap:1; /* set to remap seven reserved chars in filenames */ - bool posix_paths:1; /* unset to not ask for posix pathnames. */ - bool no_linux_ext:1; - bool sfu_emul:1; - bool nullauth:1; /* attempt to authenticate with null user */ - bool nocase:1; /* request case insensitive filenames */ - bool nobrl:1; /* disable sending byte range locks to srv */ - bool mand_lock:1; /* send mandatory not posix byte range lock reqs */ - bool seal:1; /* request transport encryption on share */ - bool nodfs:1; /* Do not request DFS, even if available */ - bool local_lease:1; /* check leases only on local system, not remote */ - bool noblocksnd:1; - bool noautotune:1; - bool nostrictsync:1; /* do not force expensive SMBflush on every sync */ - bool fsc:1; /* enable fscache */ - bool mfsymlinks:1; /* use Minshall+French Symlinks */ - bool multiuser:1; - bool use_smb2:1; /* force smb2 use on mount instead of cifs */ - unsigned int rsize; - unsigned int wsize; - bool sockopt_tcp_nodelay:1; - unsigned short int port; - unsigned long actimeo; /* attribute cache timeout (jiffies) */ - char *prepath; - struct sockaddr_storage srcaddr; /* allow binding to a local IP */ - struct nls_table *local_nls; -}; - /* FIXME: should these be tunable? */ #define TLINK_ERROR_EXPIRE (1 * HZ) #define TLINK_IDLE_EXPIRE (600 * HZ) @@ -135,9 +79,10 @@ cifs_reconnect(struct TCP_Server_Info *server) { int rc = 0; struct list_head *tmp, *tmp2; - struct cifsSesInfo *ses; - struct cifsTconInfo *tcon; + struct cifs_ses *ses; + struct cifs_tcon *tcon; struct mid_q_entry *mid_entry; + struct list_head retry_list; spin_lock(&GlobalMid_Lock); if (server->tcpStatus == CifsExiting) { @@ -157,11 +102,11 @@ cifs_reconnect(struct TCP_Server_Info *server) cFYI(1, "%s: marking sessions and tcons for reconnect", __func__); spin_lock(&cifs_tcp_ses_lock); list_for_each(tmp, &server->smb_ses_list) { - ses = list_entry(tmp, struct cifsSesInfo, smb_ses_list); + ses = list_entry(tmp, struct cifs_ses, smb_ses_list); ses->need_reconnect = true; ses->ipc_tid = 0; list_for_each(tmp2, &ses->tcon_list) { - tcon = list_entry(tmp2, struct cifsTconInfo, tcon_list); + tcon = list_entry(tmp2, struct cifs_tcon, tcon_list); tcon->need_reconnect = true; } } @@ -189,18 +134,25 @@ cifs_reconnect(struct TCP_Server_Info *server) mutex_unlock(&server->srv_mutex); /* mark submitted MIDs for retry and issue callback */ - cFYI(1, "%s: issuing mid callbacks", __func__); + INIT_LIST_HEAD(&retry_list); + cFYI(1, "%s: moving mids to private list", __func__); spin_lock(&GlobalMid_Lock); list_for_each_safe(tmp, tmp2, &server->pending_mid_q) { mid_entry = list_entry(tmp, struct mid_q_entry, qhead); if (mid_entry->midState == MID_REQUEST_SUBMITTED) mid_entry->midState = MID_RETRY_NEEDED; + list_move(&mid_entry->qhead, &retry_list); + } + spin_unlock(&GlobalMid_Lock); + + cFYI(1, "%s: issuing mid callbacks", __func__); + list_for_each_safe(tmp, tmp2, &retry_list) { + mid_entry = list_entry(tmp, struct mid_q_entry, qhead); list_del_init(&mid_entry->qhead); mid_entry->callback(mid_entry); } - spin_unlock(&GlobalMid_Lock); - while (server->tcpStatus == CifsNeedReconnect) { + do { try_to_freeze(); /* we should try only the port we connected to before */ @@ -215,7 +167,7 @@ cifs_reconnect(struct TCP_Server_Info *server) server->tcpStatus = CifsNeedNegotiate; spin_unlock(&GlobalMid_Lock); } - } + } while (server->tcpStatus == CifsNeedReconnect); return rc; } @@ -672,12 +624,12 @@ multi_t2_fnd: mid_entry->when_received = jiffies; #endif list_del_init(&mid_entry->qhead); - mid_entry->callback(mid_entry); break; } spin_unlock(&GlobalMid_Lock); if (mid_entry != NULL) { + mid_entry->callback(mid_entry); /* Was previous buf put in mpx struct for multi-rsp? */ if (!isMultiRsp) { /* smb buffer will be freed by user thread */ @@ -741,15 +693,25 @@ multi_t2_fnd: cifs_small_buf_release(smallbuf); if (!list_empty(&server->pending_mid_q)) { + struct list_head dispose_list; + + INIT_LIST_HEAD(&dispose_list); spin_lock(&GlobalMid_Lock); list_for_each_safe(tmp, tmp2, &server->pending_mid_q) { mid_entry = list_entry(tmp, struct mid_q_entry, qhead); - cFYI(1, "Clearing Mid 0x%x - issuing callback", - mid_entry->mid); + cFYI(1, "Clearing mid 0x%x", mid_entry->mid); + mid_entry->midState = MID_SHUTDOWN; + list_move(&mid_entry->qhead, &dispose_list); + } + spin_unlock(&GlobalMid_Lock); + + /* now walk dispose list and issue callbacks */ + list_for_each_safe(tmp, tmp2, &dispose_list) { + mid_entry = list_entry(tmp, struct mid_q_entry, qhead); + cFYI(1, "Callback mid 0x%x", mid_entry->mid); list_del_init(&mid_entry->qhead); mid_entry->callback(mid_entry); } - spin_unlock(&GlobalMid_Lock); /* 1/8th of sec is more than enough time for them to exit */ msleep(125); } @@ -822,7 +784,7 @@ cifs_parse_mount_options(const char *mountdata, const char *devname, struct smb_vol *vol) { char *value, *data, *end; - char *mountdata_copy, *options; + char *mountdata_copy = NULL, *options; unsigned int temp_len, i, j; char separator[2]; short int override_uid = -1; @@ -1062,13 +1024,6 @@ cifs_parse_mount_options(const char *mountdata, const char *devname, (strnicmp(value, "1", 1) == 0)) { /* this is the default */ continue; - } else if ((strnicmp(value, "smb2", 4) == 0) || - (strnicmp(value, "2", 1) == 0)) { -#ifdef CONFIG_CIFS_SMB2 - vol->use_smb2 = true; -#else - cERROR(1, "smb2 support not enabled"); -#endif /* CONFIG_CIFS_SMB2 */ } } else if ((strnicmp(data, "unc", 3) == 0) || (strnicmp(data, "target", 6) == 0) @@ -1404,6 +1359,8 @@ cifs_parse_mount_options(const char *mountdata, const char *devname, vol->server_ino = 1; } else if (strnicmp(data, "noserverino", 9) == 0) { vol->server_ino = 0; + } else if (strnicmp(data, "rwpidforward", 4) == 0) { + vol->rwpidforward = 1; } else if (strnicmp(data, "cifsacl", 7) == 0) { vol->cifs_acl = 1; } else if (strnicmp(data, "nocifsacl", 9) == 0) { @@ -1434,7 +1391,7 @@ cifs_parse_mount_options(const char *mountdata, const char *devname, "/proc/fs/cifs/LookupCacheEnabled to 0\n"); } else if (strnicmp(data, "fsc", 3) == 0) { #ifndef CONFIG_CIFS_FSCACHE - cERROR(1, "FS-Cache support needs CONFIG_CIFS_FSCACHE" + cERROR(1, "FS-Cache support needs CONFIG_CIFS_FSCACHE " "kernel config option set"); goto cifs_parse_mount_err; #endif @@ -1640,16 +1597,35 @@ match_security(struct TCP_Server_Info *server, struct smb_vol *vol) /* now check if signing mode is acceptable */ if ((secFlags & CIFSSEC_MAY_SIGN) == 0 && - (server->secMode & SECMODE_SIGN_REQUIRED)) + (server->sec_mode & SECMODE_SIGN_REQUIRED)) return false; else if (((secFlags & CIFSSEC_MUST_SIGN) == CIFSSEC_MUST_SIGN) && - (server->secMode & + (server->sec_mode & (SECMODE_SIGN_ENABLED|SECMODE_SIGN_REQUIRED)) == 0) return false; return true; } +static int match_server(struct TCP_Server_Info *server, struct sockaddr *addr, + struct smb_vol *vol) +{ + if (!net_eq(cifs_net_ns(server), current->nsproxy->net_ns)) + return 0; + + if (!match_address(server, addr, + (struct sockaddr *)&vol->srcaddr)) + return 0; + + if (!match_port(server, addr)) + return 0; + + if (!match_security(server, vol)) + return 0; + + return 1; +} + static struct TCP_Server_Info * cifs_find_tcp_session(struct sockaddr *addr, struct smb_vol *vol) { @@ -1657,17 +1633,7 @@ cifs_find_tcp_session(struct sockaddr *addr, struct smb_vol *vol) spin_lock(&cifs_tcp_ses_lock); list_for_each_entry(server, &cifs_tcp_ses_list, tcp_ses_list) { - if (!net_eq(cifs_net_ns(server), current->nsproxy->net_ns)) - continue; - - if (!match_address(server, addr, - (struct sockaddr *)&vol->srcaddr)) - continue; - - if (!match_port(server, addr)) - continue; - - if (!match_security(server, vol)) + if (!match_server(server, addr, vol)) continue; ++server->srv_count; @@ -1861,32 +1827,39 @@ out_err: return ERR_PTR(rc); } -static struct cifsSesInfo * +static int match_session(struct cifs_ses *ses, struct smb_vol *vol) +{ + switch (ses->server->secType) { + case Kerberos: + if (vol->cred_uid != ses->cred_uid) + return 0; + break; + default: + /* anything else takes username/password */ + if (ses->user_name == NULL) + return 0; + if (strncmp(ses->user_name, vol->username, + MAX_USERNAME_SIZE)) + return 0; + if (strlen(vol->username) != 0 && + ses->password != NULL && + strncmp(ses->password, + vol->password ? vol->password : "", + MAX_PASSWORD_SIZE)) + return 0; + } + return 1; +} + +static struct cifs_ses * cifs_find_smb_ses(struct TCP_Server_Info *server, struct smb_vol *vol) { - struct cifsSesInfo *ses; + struct cifs_ses *ses; spin_lock(&cifs_tcp_ses_lock); list_for_each_entry(ses, &server->smb_ses_list, smb_ses_list) { - switch (server->secType) { - case Kerberos: - if (vol->cred_uid != ses->cred_uid) - continue; - break; - default: - /* anything else takes username/password */ - if (ses->user_name == NULL) - continue; - if (strncmp(ses->user_name, vol->username, - MAX_USERNAME_SIZE)) - continue; - if (strlen(vol->username) != 0 && - ses->password != NULL && - strncmp(ses->password, - vol->password ? vol->password : "", - MAX_PASSWORD_SIZE)) - continue; - } + if (!match_session(ses, vol)) + continue; ++ses->ses_count; spin_unlock(&cifs_tcp_ses_lock); return ses; @@ -1896,7 +1869,7 @@ cifs_find_smb_ses(struct TCP_Server_Info *server, struct smb_vol *vol) } static void -cifs_put_smb_ses(struct cifsSesInfo *ses) +cifs_put_smb_ses(struct cifs_ses *ses) { int xid; struct TCP_Server_Info *server = ses->server; @@ -1922,11 +1895,11 @@ cifs_put_smb_ses(struct cifsSesInfo *ses) static bool warned_on_ntlm; /* globals init to false automatically */ -static struct cifsSesInfo * +static struct cifs_ses * cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb_vol *volume_info) { int rc = -ENOMEM, xid; - struct cifsSesInfo *ses; + struct cifs_ses *ses; struct sockaddr_in *addr = (struct sockaddr_in *)&server->dstaddr; struct sockaddr_in6 *addr6 = (struct sockaddr_in6 *)&server->dstaddr; @@ -2003,7 +1976,7 @@ cifs_get_smb_ses(struct TCP_Server_Info *server, struct smb_vol *volume_info) warned_on_ntlm = true; cERROR(1, "default security mechanism requested. The default " "security mechanism will be upgraded from ntlm to " - "ntlmv2 in kernel release 2.6.41"); + "ntlmv2 in kernel release 3.1"); } ses->overrideSecFlg = volume_info->secFlg; @@ -2029,20 +2002,26 @@ get_ses_fail: return ERR_PTR(rc); } -static struct cifsTconInfo * -cifs_find_tcon(struct cifsSesInfo *ses, const char *unc) +static int match_tcon(struct cifs_tcon *tcon, const char *unc) +{ + if (tcon->tidStatus == CifsExiting) + return 0; + if (strncmp(tcon->treeName, unc, MAX_TREE_SIZE)) + return 0; + return 1; +} + +static struct cifs_tcon * +cifs_find_tcon(struct cifs_ses *ses, const char *unc) { struct list_head *tmp; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; spin_lock(&cifs_tcp_ses_lock); list_for_each(tmp, &ses->tcon_list) { - tcon = list_entry(tmp, struct cifsTconInfo, tcon_list); - if (tcon->tidStatus == CifsExiting) + tcon = list_entry(tmp, struct cifs_tcon, tcon_list); + if (!match_tcon(tcon, unc)) continue; - if (strncmp(tcon->treeName, unc, MAX_TREE_SIZE)) - continue; - ++tcon->tc_count; spin_unlock(&cifs_tcp_ses_lock); return tcon; @@ -2052,10 +2031,10 @@ cifs_find_tcon(struct cifsSesInfo *ses, const char *unc) } static void -cifs_put_tcon(struct cifsTconInfo *tcon) +cifs_put_tcon(struct cifs_tcon *tcon) { int xid; - struct cifsSesInfo *ses = tcon->ses; + struct cifs_ses *ses = tcon->ses; cFYI(1, "%s: tc_count=%d\n", __func__, tcon->tc_count); spin_lock(&cifs_tcp_ses_lock); @@ -2076,11 +2055,11 @@ cifs_put_tcon(struct cifsTconInfo *tcon) cifs_put_smb_ses(ses); } -static struct cifsTconInfo * -cifs_get_tcon(struct cifsSesInfo *ses, struct smb_vol *volume_info) +static struct cifs_tcon * +cifs_get_tcon(struct cifs_ses *ses, struct smb_vol *volume_info) { int rc, xid; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; tcon = cifs_find_tcon(ses, volume_info->UNC); if (tcon) { @@ -2169,8 +2148,105 @@ cifs_put_tlink(struct tcon_link *tlink) return; } +static inline struct tcon_link * +cifs_sb_master_tlink(struct cifs_sb_info *cifs_sb) +{ + return cifs_sb->master_tlink; +} + +static int +compare_mount_options(struct super_block *sb, struct cifs_mnt_data *mnt_data) +{ + struct cifs_sb_info *old = CIFS_SB(sb); + struct cifs_sb_info *new = mnt_data->cifs_sb; + + if ((sb->s_flags & CIFS_MS_MASK) != (mnt_data->flags & CIFS_MS_MASK)) + return 0; + + if ((old->mnt_cifs_flags & CIFS_MOUNT_MASK) != + (new->mnt_cifs_flags & CIFS_MOUNT_MASK)) + return 0; + + if (old->rsize != new->rsize) + return 0; + + /* + * We want to share sb only if we don't specify wsize or specified wsize + * is greater or equal than existing one. + */ + if (new->wsize && new->wsize < old->wsize) + return 0; + + if (old->mnt_uid != new->mnt_uid || old->mnt_gid != new->mnt_gid) + return 0; + + if (old->mnt_file_mode != new->mnt_file_mode || + old->mnt_dir_mode != new->mnt_dir_mode) + return 0; + + if (strcmp(old->local_nls->charset, new->local_nls->charset)) + return 0; + + if (old->actimeo != new->actimeo) + return 0; + + return 1; +} + +int +cifs_match_super(struct super_block *sb, void *data) +{ + struct cifs_mnt_data *mnt_data = (struct cifs_mnt_data *)data; + struct smb_vol *volume_info; + struct cifs_sb_info *cifs_sb; + struct TCP_Server_Info *tcp_srv; + struct cifs_ses *ses; + struct cifs_tcon *tcon; + struct tcon_link *tlink; + struct sockaddr_storage addr; + int rc = 0; + + memset(&addr, 0, sizeof(struct sockaddr_storage)); + + spin_lock(&cifs_tcp_ses_lock); + cifs_sb = CIFS_SB(sb); + tlink = cifs_get_tlink(cifs_sb_master_tlink(cifs_sb)); + if (IS_ERR(tlink)) { + spin_unlock(&cifs_tcp_ses_lock); + return rc; + } + tcon = tlink_tcon(tlink); + ses = tcon->ses; + tcp_srv = ses->server; + + volume_info = mnt_data->vol; + + if (!volume_info->UNCip || !volume_info->UNC) + goto out; + + rc = cifs_fill_sockaddr((struct sockaddr *)&addr, + volume_info->UNCip, + strlen(volume_info->UNCip), + volume_info->port); + if (!rc) + goto out; + + if (!match_server(tcp_srv, (struct sockaddr *)&addr, volume_info) || + !match_session(ses, volume_info) || + !match_tcon(tcon, volume_info->UNC)) { + rc = 0; + goto out; + } + + rc = compare_mount_options(sb, mnt_data); +out: + cifs_put_tlink(tlink); + spin_unlock(&cifs_tcp_ses_lock); + return rc; +} + int -get_dfs_path(int xid, struct cifsSesInfo *pSesInfo, const char *old_path, +get_dfs_path(int xid, struct cifs_ses *pSesInfo, const char *old_path, const struct nls_table *nls_codepage, unsigned int *pnum_referrals, struct dfs_info3_param **preferrals, int remap) { @@ -2469,8 +2545,8 @@ ip_connect(struct TCP_Server_Info *server) return generic_ip_connect(server); } -void reset_cifs_unix_caps(int xid, struct cifsTconInfo *tcon, - struct super_block *sb, struct smb_vol *vol_info) +void reset_cifs_unix_caps(int xid, struct cifs_tcon *tcon, + struct cifs_sb_info *cifs_sb, struct smb_vol *vol_info) { /* if we are reconnecting then should we check to see if * any requested capabilities changed locally e.g. via @@ -2498,7 +2574,7 @@ void reset_cifs_unix_caps(int xid, struct cifsTconInfo *tcon, if (!CIFSSMBQFSUnixInfo(xid, tcon)) { __u64 cap = le64_to_cpu(tcon->fsUnixInfo.Capability); - + cFYI(1, "unix caps which server supports %lld", cap); /* check for reconnect case in which we do not want to change the mount behavior if we can avoid it */ if (vol_info == NULL) { @@ -2516,33 +2592,31 @@ void reset_cifs_unix_caps(int xid, struct cifsTconInfo *tcon, } } + if (cap & CIFS_UNIX_TRANSPORT_ENCRYPTION_MANDATORY_CAP) + cERROR(1, "per-share encryption not supported yet"); + cap &= CIFS_UNIX_CAP_MASK; if (vol_info && vol_info->no_psx_acl) cap &= ~CIFS_UNIX_POSIX_ACL_CAP; else if (CIFS_UNIX_POSIX_ACL_CAP & cap) { cFYI(1, "negotiated posix acl support"); - if (sb) - sb->s_flags |= MS_POSIXACL; + if (cifs_sb) + cifs_sb->mnt_cifs_flags |= + CIFS_MOUNT_POSIXACL; } if (vol_info && vol_info->posix_paths == 0) cap &= ~CIFS_UNIX_POSIX_PATHNAMES_CAP; else if (cap & CIFS_UNIX_POSIX_PATHNAMES_CAP) { cFYI(1, "negotiate posix pathnames"); - if (sb) - CIFS_SB(sb)->mnt_cifs_flags |= + if (cifs_sb) + cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_POSIX_PATHS; } - /* We might be setting the path sep back to a different - form if we are reconnecting and the server switched its - posix path capability for this share */ - if (sb && (CIFS_SB(sb)->prepathlen > 0)) - CIFS_SB(sb)->prepath[0] = CIFS_DIR_SEP(CIFS_SB(sb)); - - if (sb && (CIFS_SB(sb)->rsize > 127 * 1024)) { + if (cifs_sb && (cifs_sb->rsize > 127 * 1024)) { if ((cap & CIFS_UNIX_LARGE_READ_CAP) == 0) { - CIFS_SB(sb)->rsize = 127 * 1024; + cifs_sb->rsize = 127 * 1024; cFYI(DBG2, "larger reads not supported by srv"); } } @@ -2564,6 +2638,10 @@ void reset_cifs_unix_caps(int xid, struct cifsTconInfo *tcon, cFYI(1, "very large read cap"); if (cap & CIFS_UNIX_LARGE_WRITE_CAP) cFYI(1, "very large write cap"); + if (cap & CIFS_UNIX_TRANSPORT_ENCRYPTION_CAP) + cFYI(1, "transport encryption cap"); + if (cap & CIFS_UNIX_TRANSPORT_ENCRYPTION_MANDATORY_CAP) + cFYI(1, "mandatory transport encryption cap"); #endif /* CIFS_DEBUG2 */ if (CIFSSMBSetFSUnixInfo(xid, tcon, cap)) { if (vol_info == NULL) { @@ -2580,31 +2658,14 @@ void reset_cifs_unix_caps(int xid, struct cifsTconInfo *tcon, } } -static void -convert_delimiter(char *path, char delim) -{ - int i; - char old_delim; - - if (path == NULL) - return; - - if (delim == '/') - old_delim = '\\'; - else - old_delim = '/'; - - for (i = 0; path[i] != '\0'; i++) { - if (path[i] == old_delim) - path[i] = delim; - } -} - -static void setup_cifs_sb(struct smb_vol *pvolume_info, - struct cifs_sb_info *cifs_sb) +void cifs_setup_cifs_sb(struct smb_vol *pvolume_info, + struct cifs_sb_info *cifs_sb) { INIT_DELAYED_WORK(&cifs_sb->prune_tlinks, cifs_prune_tlinks); + spin_lock_init(&cifs_sb->tlink_tree_lock); + cifs_sb->tlink_tree = RB_ROOT; + if (pvolume_info->rsize > CIFSMaxBufSize) { cERROR(1, "rsize %d too large, using MaxBufSize", pvolume_info->rsize); @@ -2615,40 +2676,19 @@ static void setup_cifs_sb(struct smb_vol *pvolume_info, else /* default */ cifs_sb->rsize = CIFSMaxBufSize; - if (pvolume_info->wsize > PAGEVEC_SIZE * PAGE_CACHE_SIZE) { - cERROR(1, "wsize %d too large, using 4096 instead", - pvolume_info->wsize); - cifs_sb->wsize = 4096; - } else if (pvolume_info->wsize) - cifs_sb->wsize = pvolume_info->wsize; - else - cifs_sb->wsize = min_t(const int, - PAGEVEC_SIZE * PAGE_CACHE_SIZE, - 127*1024); - /* old default of CIFSMaxBufSize was too small now - that SMB Write2 can send multiple pages in kvec. - RFC1001 does not describe what happens when frame - bigger than 128K is sent so use that as max in - conjunction with 52K kvec constraint on arch with 4K - page size */ - if (cifs_sb->rsize < 2048) { cifs_sb->rsize = 2048; /* Windows ME may prefer this */ cFYI(1, "readsize set to minimum: 2048"); } - /* calculate prepath */ - cifs_sb->prepath = pvolume_info->prepath; - if (cifs_sb->prepath) { - cifs_sb->prepathlen = strlen(cifs_sb->prepath); - /* we can not convert the / to \ in the path - separators in the prefixpath yet because we do not - know (until reset_cifs_unix_caps is called later) - whether POSIX PATH CAP is available. We normalize - the / to \ after reset_cifs_unix_caps is called */ - pvolume_info->prepath = NULL; - } else - cifs_sb->prepathlen = 0; + + /* + * Temporarily set wsize for matching superblock. If we end up using + * new sb then cifs_negotiate_wsize will later negotiate it downward + * if needed. + */ + cifs_sb->wsize = pvolume_info->wsize; + cifs_sb->mnt_uid = pvolume_info->linux_uid; cifs_sb->mnt_gid = pvolume_info->linux_gid; cifs_sb->mnt_file_mode = pvolume_info->file_mode; @@ -2657,6 +2697,7 @@ static void setup_cifs_sb(struct smb_vol *pvolume_info, cifs_sb->mnt_file_mode, cifs_sb->mnt_dir_mode); cifs_sb->actimeo = pvolume_info->actimeo; + cifs_sb->local_nls = pvolume_info->local_nls; if (pvolume_info->noperm) cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_NO_PERM; @@ -2676,6 +2717,8 @@ static void setup_cifs_sb(struct smb_vol *pvolume_info, cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_NOSSYNC; if (pvolume_info->mand_lock) cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_NOPOSIXBRL; + if (pvolume_info->rwpidforward) + cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_RWPIDFORWARD; if (pvolume_info->cifs_acl) cifs_sb->mnt_cifs_flags |= CIFS_MOUNT_CIFS_ACL; if (pvolume_info->override_uid) @@ -2709,8 +2752,62 @@ static void setup_cifs_sb(struct smb_vol *pvolume_info, "mount option supported"); } +/* + * When the server supports very large writes via POSIX extensions, we can + * allow up to 2^24-1, minus the size of a WRITE_AND_X header, not including + * the RFC1001 length. + * + * Note that this might make for "interesting" allocation problems during + * writeback however as we have to allocate an array of pointers for the + * pages. A 16M write means ~32kb page array with PAGE_CACHE_SIZE == 4096. + */ +#define CIFS_MAX_WSIZE ((1<<24) - 1 - sizeof(WRITE_REQ) + 4) + +/* + * When the server doesn't allow large posix writes, only allow a wsize of + * 128k minus the size of the WRITE_AND_X header. That allows for a write up + * to the maximum size described by RFC1002. + */ +#define CIFS_MAX_RFC1002_WSIZE (128 * 1024 - sizeof(WRITE_REQ) + 4) + +/* + * The default wsize is 1M. find_get_pages seems to return a maximum of 256 + * pages in a single call. With PAGE_CACHE_SIZE == 4k, this means we can fill + * a single wsize request with a single call. + */ +#define CIFS_DEFAULT_WSIZE (1024 * 1024) + +static unsigned int +cifs_negotiate_wsize(struct cifs_tcon *tcon, struct smb_vol *pvolume_info) +{ + __u64 unix_cap = le64_to_cpu(tcon->fsUnixInfo.Capability); + struct TCP_Server_Info *server = tcon->ses->server; + unsigned int wsize = pvolume_info->wsize ? pvolume_info->wsize : + CIFS_DEFAULT_WSIZE; + + /* can server support 24-bit write sizes? (via UNIX extensions) */ + if (!tcon->unix_ext || !(unix_cap & CIFS_UNIX_LARGE_WRITE_CAP)) + wsize = min_t(unsigned int, wsize, CIFS_MAX_RFC1002_WSIZE); + + /* + * no CAP_LARGE_WRITE_X or is signing enabled without CAP_UNIX set? + * Limit it to max buffer offered by the server, minus the size of the + * WRITEX header, not including the 4 byte RFC1001 length. + */ + if (!(server->capabilities & CAP_LARGE_WRITE_X) || + (!(server->capabilities & CAP_UNIX) && + (server->sec_mode & (SECMODE_SIGN_ENABLED|SECMODE_SIGN_REQUIRED)))) + wsize = min_t(unsigned int, wsize, + server->maxBuf - sizeof(WRITE_REQ) + 4); + + /* hard limit of CIFS_MAX_WSIZE */ + wsize = min_t(unsigned int, wsize, CIFS_MAX_WSIZE); + + return wsize; +} + static int -is_path_accessible(int xid, struct cifsTconInfo *tcon, +is_path_accessible(int xid, struct cifs_tcon *tcon, struct cifs_sb_info *cifs_sb, const char *full_path) { int rc; @@ -2733,8 +2830,8 @@ is_path_accessible(int xid, struct cifsTconInfo *tcon, return rc; } -static void -cleanup_volume_info(struct smb_vol **pvolume_info) +void +cifs_cleanup_volume_info(struct smb_vol **pvolume_info) { struct smb_vol *volume_info; @@ -2764,24 +2861,13 @@ build_unc_path_to_root(const struct smb_vol *volume_info, char *full_path; int unc_len = strnlen(volume_info->UNC, MAX_TREE_SIZE + 1); - full_path = kmalloc(unc_len + cifs_sb->prepathlen + 1, GFP_KERNEL); + full_path = kmalloc(unc_len + 1, GFP_KERNEL); if (full_path == NULL) return ERR_PTR(-ENOMEM); strncpy(full_path, volume_info->UNC, unc_len); - if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) { - int i; - for (i = 0; i < unc_len; i++) { - if (full_path[i] == '\\') - full_path[i] = '/'; - } - } - - if (cifs_sb->prepathlen) - strncpy(full_path + unc_len, cifs_sb->prepath, - cifs_sb->prepathlen); - - full_path[unc_len + cifs_sb->prepathlen] = 0; /* add trailing null */ + full_path[unc_len] = 0; /* add trailing null */ + convert_delimiter(full_path, CIFS_DIR_SEP(cifs_sb)); return full_path; } @@ -2796,7 +2882,7 @@ build_unc_path_to_root(const struct smb_vol *volume_info, * determine whether there were referrals. */ static int -expand_dfs_referral(int xid, struct cifsSesInfo *pSesInfo, +expand_dfs_referral(int xid, struct cifs_ses *pSesInfo, struct smb_vol *volume_info, struct cifs_sb_info *cifs_sb, int check_prefix) { @@ -2840,40 +2926,13 @@ expand_dfs_referral(int xid, struct cifsSesInfo *pSesInfo, } #endif -int -cifs_mount(struct super_block *sb, struct cifs_sb_info *cifs_sb, - const char *devname) +int cifs_setup_volume_info(struct smb_vol **pvolume_info, char *mount_data, + const char *devname) { - int rc; - int xid; struct smb_vol *volume_info; - struct cifsSesInfo *pSesInfo; - struct cifsTconInfo *tcon; - struct TCP_Server_Info *srvTcp; - char *full_path; - struct tcon_link *tlink; -#ifdef CONFIG_CIFS_DFS_UPCALL - int referral_walks_count = 0; -try_mount_again: - /* cleanup activities if we're chasing a referral */ - if (referral_walks_count) { - if (tcon) - cifs_put_tcon(tcon); - else if (pSesInfo) - cifs_put_smb_ses(pSesInfo); - - cleanup_volume_info(&volume_info); - FreeXid(xid); - } -#endif - rc = 0; - tcon = NULL; - pSesInfo = NULL; - srvTcp = NULL; - full_path = NULL; - tlink = NULL; + int rc = 0; - xid = GetXid(); + *pvolume_info = NULL; volume_info = kzalloc(sizeof(struct smb_vol), GFP_KERNEL); if (!volume_info) { @@ -2881,7 +2940,7 @@ try_mount_again: goto out; } - if (cifs_parse_mount_options(cifs_sb->mountdata, devname, + if (cifs_parse_mount_options(mount_data, devname, volume_info)) { rc = -EINVAL; goto out; @@ -2889,7 +2948,11 @@ try_mount_again: if (volume_info->nullauth) { cFYI(1, "null user"); - volume_info->username = ""; + volume_info->username = kzalloc(1, GFP_KERNEL); + if (volume_info->username == NULL) { + rc = -ENOMEM; + goto out; + } } else if (volume_info->username) { /* BB fixme parse for domain name here */ cFYI(1, "Username: %s", volume_info->username); @@ -2914,12 +2977,58 @@ try_mount_again: goto out; } } - cifs_sb->local_nls = volume_info->local_nls; + + *pvolume_info = volume_info; + return rc; +out: + cifs_cleanup_volume_info(&volume_info); + return rc; +} + +int +cifs_mount(struct cifs_sb_info *cifs_sb, struct smb_vol *volume_info) +{ + int rc = 0; + int xid; + struct cifs_ses *pSesInfo; + struct cifs_tcon *tcon; + struct TCP_Server_Info *srvTcp; + char *full_path; + struct tcon_link *tlink; +#ifdef CONFIG_CIFS_DFS_UPCALL + int referral_walks_count = 0; + + rc = bdi_setup_and_register(&cifs_sb->bdi, "cifs", BDI_CAP_MAP_COPY); + if (rc) + return rc; + + cifs_sb->bdi.ra_pages = default_backing_dev_info.ra_pages; + +try_mount_again: + /* cleanup activities if we're chasing a referral */ + if (referral_walks_count) { + if (tcon) + cifs_put_tcon(tcon); + else if (pSesInfo) + cifs_put_smb_ses(pSesInfo); + + cifs_cleanup_volume_info(&volume_info); + FreeXid(xid); + } +#endif + tcon = NULL; + pSesInfo = NULL; + srvTcp = NULL; + full_path = NULL; + tlink = NULL; + + xid = GetXid(); /* get a reference to a tcp session */ srvTcp = cifs_get_tcp_session(volume_info); if (IS_ERR(srvTcp)) { rc = PTR_ERR(srvTcp); + bdi_destroy(&cifs_sb->bdi); goto out; } @@ -2931,15 +3040,6 @@ try_mount_again: goto mount_fail_check; } - setup_cifs_sb(volume_info, cifs_sb); - if (pSesInfo->capabilities & CAP_LARGE_FILES) - sb->s_maxbytes = MAX_LFS_FILESIZE; - else - sb->s_maxbytes = MAX_NON_LFS; - - /* BB FIXME fix time_gran to be larger for LANMAN sessions */ - sb->s_time_gran = 100; - /* search for existing tcon to this server share */ tcon = cifs_get_tcon(pSesInfo, volume_info); if (IS_ERR(tcon)) { @@ -2948,35 +3048,36 @@ try_mount_again: goto remote_path_check; } - /* do not care if following two calls succeed - informational */ - if (!tcon->ipc) { - CIFSSMBQFSDeviceInfo(xid, tcon); - CIFSSMBQFSAttributeInfo(xid, tcon); - } - /* tell server which Unix caps we support */ - if (tcon->ses->capabilities & CAP_UNIX) + if (tcon->ses->capabilities & CAP_UNIX) { /* reset of caps checks mount to see if unix extensions disabled for just this mount */ - reset_cifs_unix_caps(xid, tcon, sb, volume_info); - else + reset_cifs_unix_caps(xid, tcon, cifs_sb, volume_info); + if ((tcon->ses->server->tcpStatus == CifsNeedReconnect) && + (le64_to_cpu(tcon->fsUnixInfo.Capability) & + CIFS_UNIX_TRANSPORT_ENCRYPTION_MANDATORY_CAP)) { + rc = -EACCES; + goto mount_fail_check; + } + } else tcon->unix_ext = 0; /* server does not support them */ - /* convert forward to back slashes in prepath here if needed */ - if ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) == 0) - convert_delimiter(cifs_sb->prepath, CIFS_DIR_SEP(cifs_sb)); + /* do not care if following two calls succeed - informational */ + if (!tcon->ipc) { + CIFSSMBQFSDeviceInfo(xid, tcon); + CIFSSMBQFSAttributeInfo(xid, tcon); + } if ((tcon->unix_ext == 0) && (cifs_sb->rsize > (1024 * 127))) { cifs_sb->rsize = 1024 * 127; cFYI(DBG2, "no very large read support, rsize now 127K"); } - if (!(tcon->ses->capabilities & CAP_LARGE_WRITE_X)) - cifs_sb->wsize = min(cifs_sb->wsize, - (tcon->ses->server->maxBuf - MAX_CIFS_HDR_SIZE)); if (!(tcon->ses->capabilities & CAP_LARGE_READ_X)) cifs_sb->rsize = min(cifs_sb->rsize, (tcon->ses->server->maxBuf - MAX_CIFS_HDR_SIZE)); + cifs_sb->wsize = cifs_negotiate_wsize(tcon, volume_info); + remote_path_check: #ifdef CONFIG_CIFS_DFS_UPCALL /* @@ -2996,10 +3097,10 @@ remote_path_check: } #endif - /* check if a whole path (including prepath) is not remote */ + /* check if a whole path is not remote */ if (!rc && tcon) { /* build_path_to_root works only when we have a valid tcon */ - full_path = cifs_build_path_to_root(cifs_sb, tcon); + full_path = cifs_build_path_to_root(volume_info, cifs_sb, tcon); if (full_path == NULL) { rc = -ENOMEM; goto mount_fail_check; @@ -3025,10 +3126,6 @@ remote_path_check: rc = -ELOOP; goto mount_fail_check; } - /* convert forward to back slashes in prepath here if needed */ - if ((cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) == 0) - convert_delimiter(cifs_sb->prepath, - CIFS_DIR_SEP(cifs_sb)); rc = expand_dfs_referral(xid, pSesInfo, volume_info, cifs_sb, true); @@ -3078,6 +3175,7 @@ mount_fail_check: cifs_put_smb_ses(pSesInfo); else cifs_put_tcp_session(srvTcp); + bdi_destroy(&cifs_sb->bdi); goto out; } @@ -3087,14 +3185,17 @@ mount_fail_check: password will be freed at unmount time) */ out: /* zero out password before freeing */ - cleanup_volume_info(&volume_info); FreeXid(xid); return rc; } +/* + * Issue a TREE_CONNECT request. Note that for IPC$ shares, that the tcon + * pointer may be NULL. + */ int -CIFSTCon(unsigned int xid, struct cifsSesInfo *ses, - const char *tree, struct cifsTconInfo *tcon, +CIFSTCon(unsigned int xid, struct cifs_ses *ses, + const char *tree, struct cifs_tcon *tcon, const struct nls_table *nls_codepage) { struct smb_hdr *smb_buffer; @@ -3126,7 +3227,7 @@ CIFSTCon(unsigned int xid, struct cifsSesInfo *ses, pSMB->AndXCommand = 0xFF; pSMB->Flags = cpu_to_le16(TCON_EXTENDED_SECINFO); bcc_ptr = &pSMB->Password[0]; - if ((ses->server->secMode) & SECMODE_USER) { + if (!tcon || (ses->server->sec_mode & SECMODE_USER)) { pSMB->PasswordLength = cpu_to_le16(1); /* minimum */ *bcc_ptr = 0; /* password is null byte */ bcc_ptr++; /* skip password */ @@ -3143,7 +3244,7 @@ CIFSTCon(unsigned int xid, struct cifsSesInfo *ses, if ((global_secflags & CIFSSEC_MAY_LANMAN) && (ses->server->secType == LANMAN)) calc_lanman_hash(tcon->password, ses->server->cryptkey, - ses->server->secMode & + ses->server->sec_mode & SECMODE_PW_ENCRYPT ? true : false, bcc_ptr); else @@ -3159,7 +3260,7 @@ CIFSTCon(unsigned int xid, struct cifsSesInfo *ses, } } - if (ses->server->secMode & + if (ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) smb_buffer->Flags2 |= SMBFLG2_SECURITY_SIGNATURE; @@ -3249,13 +3350,12 @@ CIFSTCon(unsigned int xid, struct cifsSesInfo *ses, return rc; } -int -cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb) +void +cifs_umount(struct cifs_sb_info *cifs_sb) { struct rb_root *root = &cifs_sb->tlink_tree; struct rb_node *node; struct tcon_link *tlink; - char *tmp; cancel_delayed_work_sync(&cifs_sb->prune_tlinks); @@ -3272,15 +3372,13 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb) } spin_unlock(&cifs_sb->tlink_tree_lock); - tmp = cifs_sb->prepath; - cifs_sb->prepathlen = 0; - cifs_sb->prepath = NULL; - kfree(tmp); - - return 0; + bdi_destroy(&cifs_sb->bdi); + kfree(cifs_sb->mountdata); + unload_nls(cifs_sb->local_nls); + kfree(cifs_sb); } -int cifs_negotiate_protocol(unsigned int xid, struct cifsSesInfo *ses) +int cifs_negotiate_protocol(unsigned int xid, struct cifs_ses *ses) { int rc = 0; struct TCP_Server_Info *server = ses->server; @@ -3298,7 +3396,7 @@ int cifs_negotiate_protocol(unsigned int xid, struct cifsSesInfo *ses) } if (rc == 0) { spin_lock(&GlobalMid_Lock); - if (server->tcpStatus != CifsExiting) + if (server->tcpStatus == CifsNeedNegotiate) server->tcpStatus = CifsGood; else rc = -EHOSTDOWN; @@ -3310,7 +3408,7 @@ int cifs_negotiate_protocol(unsigned int xid, struct cifsSesInfo *ses) } -int cifs_setup_session(unsigned int xid, struct cifsSesInfo *ses, +int cifs_setup_session(unsigned int xid, struct cifs_ses *ses, struct nls_table *nls_info) { int rc = 0; @@ -3322,7 +3420,7 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *ses, ses->capabilities &= (~CAP_UNIX); cFYI(1, "Security Mode: 0x%x Capabilities: 0x%x TimeAdjust: %d", - server->secMode, server->capabilities, server->timeAdj); + server->sec_mode, server->capabilities, server->timeAdj); rc = CIFS_SessSetup(xid, ses, nls_info); if (rc) { @@ -3354,12 +3452,12 @@ int cifs_setup_session(unsigned int xid, struct cifsSesInfo *ses, return rc; } -static struct cifsTconInfo * +static struct cifs_tcon * cifs_construct_tcon(struct cifs_sb_info *cifs_sb, uid_t fsuid) { - struct cifsTconInfo *master_tcon = cifs_sb_master_tcon(cifs_sb); - struct cifsSesInfo *ses; - struct cifsTconInfo *tcon = NULL; + struct cifs_tcon *master_tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_ses *ses; + struct cifs_tcon *tcon = NULL; struct smb_vol *vol_info; char username[28]; /* big enough for "krb50x" + hex of ULONG_MAX 6+16 */ /* We used to have this as MAX_USERNAME which is */ @@ -3392,7 +3490,7 @@ cifs_construct_tcon(struct cifs_sb_info *cifs_sb, uid_t fsuid) ses = cifs_get_smb_ses(master_tcon->ses->server, vol_info); if (IS_ERR(ses)) { - tcon = (struct cifsTconInfo *)ses; + tcon = (struct cifs_tcon *)ses; cifs_put_tcp_session(master_tcon->ses->server); goto out; } @@ -3411,13 +3509,7 @@ out: return tcon; } -static inline struct tcon_link * -cifs_sb_master_tlink(struct cifs_sb_info *cifs_sb) -{ - return cifs_sb->master_tlink; -} - -struct cifsTconInfo * +struct cifs_tcon * cifs_sb_master_tcon(struct cifs_sb_info *cifs_sb) { return tlink_tcon(cifs_sb_master_tlink(cifs_sb)); diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c index 9ea65cf36714..81914df47ef1 100644 --- a/fs/cifs/dir.c +++ b/fs/cifs/dir.c @@ -50,12 +50,11 @@ build_path_from_dentry(struct dentry *direntry) { struct dentry *temp; int namelen; - int pplen; int dfsplen; char *full_path; char dirsep; struct cifs_sb_info *cifs_sb = CIFS_SB(direntry->d_sb); - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); if (direntry == NULL) return NULL; /* not much we can do if dentry is freed and @@ -63,13 +62,12 @@ build_path_from_dentry(struct dentry *direntry) when the server crashed */ dirsep = CIFS_DIR_SEP(cifs_sb); - pplen = cifs_sb->prepathlen; if (tcon->Flags & SMB_SHARE_IS_IN_DFS) dfsplen = strnlen(tcon->treeName, MAX_TREE_SIZE + 1); else dfsplen = 0; cifs_bp_rename_retry: - namelen = pplen + dfsplen; + namelen = dfsplen; for (temp = direntry; !IS_ROOT(temp);) { namelen += (1 + temp->d_name.len); temp = temp->d_parent; @@ -100,7 +98,7 @@ cifs_bp_rename_retry: return NULL; } } - if (namelen != pplen + dfsplen) { + if (namelen != dfsplen) { cERROR(1, "did not end path lookup where expected namelen is %d", namelen); /* presumably this is only possible if racing with a rename @@ -126,7 +124,6 @@ cifs_bp_rename_retry: } } } - strncpy(full_path + dfsplen, CIFS_SB(direntry->d_sb)->prepath, pplen); return full_path; } @@ -152,7 +149,7 @@ cifs_create(struct inode *inode, struct dentry *direntry, int mode, __u16 fileHandle; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; char *full_path = NULL; FILE_ALL_INFO *buf = NULL; struct inode *newinode = NULL; @@ -356,7 +353,8 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode, int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; + struct cifs_io_parms io_parms; char *full_path = NULL; struct inode *newinode = NULL; int oplock = 0; @@ -439,16 +437,19 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode, * timestamps in, but we can reuse it safely */ pdev = (struct win_dev *)buf; + io_parms.netfid = fileHandle; + io_parms.pid = current->tgid; + io_parms.tcon = pTcon; + io_parms.offset = 0; + io_parms.length = sizeof(struct win_dev); if (S_ISCHR(mode)) { memcpy(pdev->type, "IntxCHR", 8); pdev->major = cpu_to_le64(MAJOR(device_number)); pdev->minor = cpu_to_le64(MINOR(device_number)); - rc = CIFSSMBWrite(xid, pTcon, - fileHandle, - sizeof(struct win_dev), - 0, &bytes_written, (char *)pdev, + rc = CIFSSMBWrite(xid, &io_parms, + &bytes_written, (char *)pdev, NULL, 0); } else if (S_ISBLK(mode)) { memcpy(pdev->type, "IntxBLK", 8); @@ -456,10 +457,8 @@ int cifs_mknod(struct inode *inode, struct dentry *direntry, int mode, cpu_to_le64(MAJOR(device_number)); pdev->minor = cpu_to_le64(MINOR(device_number)); - rc = CIFSSMBWrite(xid, pTcon, - fileHandle, - sizeof(struct win_dev), - 0, &bytes_written, (char *)pdev, + rc = CIFSSMBWrite(xid, &io_parms, + &bytes_written, (char *)pdev, NULL, 0); } /* else if (S_ISFIFO) */ CIFSSMBClose(xid, pTcon, fileHandle); @@ -486,7 +485,7 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry, bool posix_open = false; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct cifsFileInfo *cfile; struct inode *newInode = NULL; char *full_path = NULL; diff --git a/fs/cifs/file.c b/fs/cifs/file.c index c672afef0c09..bb71471a4d9d 100644 --- a/fs/cifs/file.c +++ b/fs/cifs/file.c @@ -114,7 +114,7 @@ int cifs_posix_open(char *full_path, struct inode **pinode, struct cifs_sb_info *cifs_sb = CIFS_SB(sb); struct cifs_fattr fattr; struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; cFYI(1, "posix open %s", full_path); @@ -168,7 +168,7 @@ posix_open_ret: static int cifs_nt_open(char *full_path, struct inode *inode, struct cifs_sb_info *cifs_sb, - struct cifsTconInfo *tcon, unsigned int f_flags, __u32 *poplock, + struct cifs_tcon *tcon, unsigned int f_flags, __u32 *poplock, __u16 *pnetfid, int xid) { int rc; @@ -285,7 +285,7 @@ cifs_new_fileinfo(__u16 fileHandle, struct file *file, void cifsFileInfo_put(struct cifsFileInfo *cifs_file) { struct inode *inode = cifs_file->dentry->d_inode; - struct cifsTconInfo *tcon = tlink_tcon(cifs_file->tlink); + struct cifs_tcon *tcon = tlink_tcon(cifs_file->tlink); struct cifsInodeInfo *cifsi = CIFS_I(inode); struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct cifsLockInfo *li, *tmp; @@ -343,7 +343,7 @@ int cifs_open(struct inode *inode, struct file *file) int xid; __u32 oplock; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct tcon_link *tlink; struct cifsFileInfo *pCifsFile = NULL; char *full_path = NULL; @@ -457,7 +457,7 @@ static int cifs_reopen_file(struct cifsFileInfo *pCifsFile, bool can_flush) int xid; __u32 oplock; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct cifsInodeInfo *pCifsInode; struct inode *inode; char *full_path = NULL; @@ -596,7 +596,7 @@ int cifs_closedir(struct inode *inode, struct file *file) xid = GetXid(); if (pCFileStruct) { - struct cifsTconInfo *pTcon = tlink_tcon(pCFileStruct->tlink); + struct cifs_tcon *pTcon = tlink_tcon(pCFileStruct->tlink); cFYI(1, "Freeing private data in close dir"); spin_lock(&cifs_file_list_lock); @@ -653,7 +653,7 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *pfLock) __u64 length; bool wait_flag = false; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; __u16 netfid; __u8 lockType = LOCKING_ANDX_LARGE_FILES; bool posix_locking = 0; @@ -725,8 +725,8 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *pfLock) else posix_lock_type = CIFS_WRLCK; rc = CIFSSMBPosixLock(xid, tcon, netfid, 1 /* get */, - length, pfLock, - posix_lock_type, wait_flag); + length, pfLock, posix_lock_type, + wait_flag); FreeXid(xid); return rc; } @@ -797,8 +797,8 @@ int cifs_lock(struct file *file, int cmd, struct file_lock *pfLock) posix_lock_type = CIFS_UNLCK; rc = CIFSSMBPosixLock(xid, tcon, netfid, 0 /* set */, - length, pfLock, - posix_lock_type, wait_flag); + length, pfLock, posix_lock_type, + wait_flag); } else { struct cifsFileInfo *fid = file->private_data; @@ -857,7 +857,7 @@ cifs_update_eof(struct cifsInodeInfo *cifsi, loff_t offset, cifsi->server_eof = end_of_write; } -static ssize_t cifs_write(struct cifsFileInfo *open_file, +static ssize_t cifs_write(struct cifsFileInfo *open_file, __u32 pid, const char *write_data, size_t write_size, loff_t *poffset) { @@ -865,10 +865,11 @@ static ssize_t cifs_write(struct cifsFileInfo *open_file, unsigned int bytes_written = 0; unsigned int total_written; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; int xid; struct dentry *dentry = open_file->dentry; struct cifsInodeInfo *cifsi = CIFS_I(dentry->d_inode); + struct cifs_io_parms io_parms; cifs_sb = CIFS_SB(dentry->d_sb); @@ -901,8 +902,13 @@ static ssize_t cifs_write(struct cifsFileInfo *open_file, /* iov[0] is reserved for smb header */ iov[1].iov_base = (char *)write_data + total_written; iov[1].iov_len = len; - rc = CIFSSMBWrite2(xid, pTcon, open_file->netfid, len, - *poffset, &bytes_written, iov, 1, 0); + io_parms.netfid = open_file->netfid; + io_parms.pid = pid; + io_parms.tcon = pTcon; + io_parms.offset = *poffset; + io_parms.length = len; + rc = CIFSSMBWrite2(xid, &io_parms, &bytes_written, iov, + 1, 0); } if (rc || (bytes_written == 0)) { if (total_written) @@ -1071,8 +1077,8 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to) open_file = find_writable_file(CIFS_I(mapping->host), false); if (open_file) { - bytes_written = cifs_write(open_file, write_data, - to - from, &offset); + bytes_written = cifs_write(open_file, open_file->pid, + write_data, to - from, &offset); cifsFileInfo_put(open_file); /* Does mm or vfs already set times? */ inode->i_atime = inode->i_mtime = current_fs_time(inode->i_sb); @@ -1092,58 +1098,20 @@ static int cifs_partialpagewrite(struct page *page, unsigned from, unsigned to) static int cifs_writepages(struct address_space *mapping, struct writeback_control *wbc) { - unsigned int bytes_to_write; - unsigned int bytes_written; - struct cifs_sb_info *cifs_sb; - int done = 0; - pgoff_t end; - pgoff_t index; - int range_whole = 0; - struct kvec *iov; - int len; - int n_iov = 0; - pgoff_t next; - int nr_pages; - __u64 offset = 0; - struct cifsFileInfo *open_file; - struct cifsTconInfo *tcon; - struct cifsInodeInfo *cifsi = CIFS_I(mapping->host); + struct cifs_sb_info *cifs_sb = CIFS_SB(mapping->host->i_sb); + bool done = false, scanned = false, range_whole = false; + pgoff_t end, index; + struct cifs_writedata *wdata; struct page *page; - struct pagevec pvec; int rc = 0; - int scanned = 0; - int xid; - - cifs_sb = CIFS_SB(mapping->host->i_sb); /* - * If wsize is smaller that the page cache size, default to writing + * If wsize is smaller than the page cache size, default to writing * one page at a time via cifs_writepage */ if (cifs_sb->wsize < PAGE_CACHE_SIZE) return generic_writepages(mapping, wbc); - iov = kmalloc(32 * sizeof(struct kvec), GFP_KERNEL); - if (iov == NULL) - return generic_writepages(mapping, wbc); - - /* - * if there's no open file, then this is likely to fail too, - * but it'll at least handle the return. Maybe it should be - * a BUG() instead? - */ - open_file = find_writable_file(CIFS_I(mapping->host), false); - if (!open_file) { - kfree(iov); - return generic_writepages(mapping, wbc); - } - - tcon = tlink_tcon(open_file->tlink); - cifsFileInfo_put(open_file); - - xid = GetXid(); - - pagevec_init(&pvec, 0); if (wbc->range_cyclic) { index = mapping->writeback_index; /* Start from prev offset */ end = -1; @@ -1151,24 +1119,49 @@ static int cifs_writepages(struct address_space *mapping, index = wbc->range_start >> PAGE_CACHE_SHIFT; end = wbc->range_end >> PAGE_CACHE_SHIFT; if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX) - range_whole = 1; - scanned = 1; + range_whole = true; + scanned = true; } retry: - while (!done && (index <= end) && - (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, - PAGECACHE_TAG_DIRTY, - min(end - index, (pgoff_t)PAGEVEC_SIZE - 1) + 1))) { - int first; - unsigned int i; - - first = -1; - next = 0; - n_iov = 0; - bytes_to_write = 0; - - for (i = 0; i < nr_pages; i++) { - page = pvec.pages[i]; + while (!done && index <= end) { + unsigned int i, nr_pages, found_pages; + pgoff_t next = 0, tofind; + struct page **pages; + + tofind = min((cifs_sb->wsize / PAGE_CACHE_SIZE) - 1, + end - index) + 1; + + wdata = cifs_writedata_alloc((unsigned int)tofind); + if (!wdata) { + rc = -ENOMEM; + break; + } + + /* + * find_get_pages_tag seems to return a max of 256 on each + * iteration, so we must call it several times in order to + * fill the array or the wsize is effectively limited to + * 256 * PAGE_CACHE_SIZE. + */ + found_pages = 0; + pages = wdata->pages; + do { + nr_pages = find_get_pages_tag(mapping, &index, + PAGECACHE_TAG_DIRTY, + tofind, pages); + found_pages += nr_pages; + tofind -= nr_pages; + pages += nr_pages; + } while (nr_pages && tofind && index <= end); + + if (found_pages == 0) { + kref_put(&wdata->refcount, cifs_writedata_release); + break; + } + + nr_pages = 0; + for (i = 0; i < found_pages; i++) { + page = wdata->pages[i]; /* * At this point we hold neither mapping->tree_lock nor * lock on the page itself: the page may be truncated or @@ -1177,7 +1170,7 @@ retry: * mapping */ - if (first < 0) + if (nr_pages == 0) lock_page(page); else if (!trylock_page(page)) break; @@ -1188,7 +1181,7 @@ retry: } if (!wbc->range_cyclic && page->index > end) { - done = 1; + done = true; unlock_page(page); break; } @@ -1215,119 +1208,89 @@ retry: set_page_writeback(page); if (page_offset(page) >= mapping->host->i_size) { - done = 1; + done = true; unlock_page(page); end_page_writeback(page); break; } - /* - * BB can we get rid of this? pages are held by pvec - */ - page_cache_get(page); + wdata->pages[i] = page; + next = page->index + 1; + ++nr_pages; + } - len = min(mapping->host->i_size - page_offset(page), - (loff_t)PAGE_CACHE_SIZE); + /* reset index to refind any pages skipped */ + if (nr_pages == 0) + index = wdata->pages[0]->index + 1; - /* reserve iov[0] for the smb header */ - n_iov++; - iov[n_iov].iov_base = kmap(page); - iov[n_iov].iov_len = len; - bytes_to_write += len; + /* put any pages we aren't going to use */ + for (i = nr_pages; i < found_pages; i++) { + page_cache_release(wdata->pages[i]); + wdata->pages[i] = NULL; + } - if (first < 0) { - first = i; - offset = page_offset(page); - } - next = page->index + 1; - if (bytes_to_write + PAGE_CACHE_SIZE > cifs_sb->wsize) - break; + /* nothing to write? */ + if (nr_pages == 0) { + kref_put(&wdata->refcount, cifs_writedata_release); + continue; } - if (n_iov) { -retry_write: - open_file = find_writable_file(CIFS_I(mapping->host), - false); - if (!open_file) { - cERROR(1, "No writable handles for inode"); - rc = -EBADF; - } else { - rc = CIFSSMBWrite2(xid, tcon, open_file->netfid, - bytes_to_write, offset, - &bytes_written, iov, n_iov, - 0); - cifsFileInfo_put(open_file); - } - cFYI(1, "Write2 rc=%d, wrote=%u", rc, bytes_written); + wdata->sync_mode = wbc->sync_mode; + wdata->nr_pages = nr_pages; + wdata->offset = page_offset(wdata->pages[0]); - /* - * For now, treat a short write as if nothing got - * written. A zero length write however indicates - * ENOSPC or EFBIG. We have no way to know which - * though, so call it ENOSPC for now. EFBIG would - * get translated to AS_EIO anyway. - * - * FIXME: make it take into account the data that did - * get written - */ - if (rc == 0) { - if (bytes_written == 0) - rc = -ENOSPC; - else if (bytes_written < bytes_to_write) - rc = -EAGAIN; + do { + if (wdata->cfile != NULL) + cifsFileInfo_put(wdata->cfile); + wdata->cfile = find_writable_file(CIFS_I(mapping->host), + false); + if (!wdata->cfile) { + cERROR(1, "No writable handles for inode"); + rc = -EBADF; + break; } + rc = cifs_async_writev(wdata); + } while (wbc->sync_mode == WB_SYNC_ALL && rc == -EAGAIN); - /* retry on data-integrity flush */ - if (wbc->sync_mode == WB_SYNC_ALL && rc == -EAGAIN) - goto retry_write; - - /* fix the stats and EOF */ - if (bytes_written > 0) { - cifs_stats_bytes_written(tcon, bytes_written); - cifs_update_eof(cifsi, offset, bytes_written); - } + for (i = 0; i < nr_pages; ++i) + unlock_page(wdata->pages[i]); - for (i = 0; i < n_iov; i++) { - page = pvec.pages[first + i]; - /* on retryable write error, redirty page */ + /* send failure -- clean up the mess */ + if (rc != 0) { + for (i = 0; i < nr_pages; ++i) { if (rc == -EAGAIN) - redirty_page_for_writepage(wbc, page); - else if (rc != 0) - SetPageError(page); - kunmap(page); - unlock_page(page); - end_page_writeback(page); - page_cache_release(page); + redirty_page_for_writepage(wbc, + wdata->pages[i]); + else + SetPageError(wdata->pages[i]); + end_page_writeback(wdata->pages[i]); + page_cache_release(wdata->pages[i]); } - if (rc != -EAGAIN) mapping_set_error(mapping, rc); - else - rc = 0; + } + kref_put(&wdata->refcount, cifs_writedata_release); - if ((wbc->nr_to_write -= n_iov) <= 0) - done = 1; - index = next; - } else - /* Need to re-find the pages we skipped */ - index = pvec.pages[0]->index + 1; + wbc->nr_to_write -= nr_pages; + if (wbc->nr_to_write <= 0) + done = true; - pagevec_release(&pvec); + index = next; } + if (!scanned && !done) { /* * We hit the last page and there is more work to be done: wrap * back to the start of the file */ - scanned = 1; + scanned = true; index = 0; goto retry; } + if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0)) mapping->writeback_index = index; - FreeXid(xid); - kfree(iov); return rc; } @@ -1383,6 +1346,14 @@ static int cifs_write_end(struct file *file, struct address_space *mapping, { int rc; struct inode *inode = mapping->host; + struct cifsFileInfo *cfile = file->private_data; + struct cifs_sb_info *cifs_sb = CIFS_SB(cfile->dentry->d_sb); + __u32 pid; + + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + pid = cfile->pid; + else + pid = current->tgid; cFYI(1, "write_end for page %p from pos %lld with %d bytes", page, pos, copied); @@ -1406,8 +1377,7 @@ static int cifs_write_end(struct file *file, struct address_space *mapping, /* BB check if anything else missing out of ppw such as updating last write time */ page_data = kmap(page); - rc = cifs_write(file->private_data, page_data + offset, - copied, &pos); + rc = cifs_write(cfile, pid, page_data + offset, copied, &pos); /* if (rc < 0) should we set writebehind rc? */ kunmap(page); @@ -1435,7 +1405,7 @@ int cifs_strict_fsync(struct file *file, int datasync) { int xid; int rc = 0; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct cifsFileInfo *smbfile = file->private_data; struct inode *inode = file->f_path.dentry->d_inode; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); @@ -1465,7 +1435,7 @@ int cifs_fsync(struct file *file, int datasync) { int xid; int rc = 0; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct cifsFileInfo *smbfile = file->private_data; struct cifs_sb_info *cifs_sb = CIFS_SB(file->f_path.dentry->d_sb); @@ -1556,9 +1526,11 @@ cifs_iovec_write(struct file *file, const struct iovec *iov, struct iov_iter it; struct inode *inode; struct cifsFileInfo *open_file; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct cifs_sb_info *cifs_sb; + struct cifs_io_parms io_parms; int xid, rc; + __u32 pid; len = iov_length(iov, nr_segs); if (!len) @@ -1590,6 +1562,12 @@ cifs_iovec_write(struct file *file, const struct iovec *iov, xid = GetXid(); open_file = file->private_data; + + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + pid = open_file->pid; + else + pid = current->tgid; + pTcon = tlink_tcon(open_file->tlink); inode = file->f_path.dentry->d_inode; @@ -1616,9 +1594,13 @@ cifs_iovec_write(struct file *file, const struct iovec *iov, if (rc != 0) break; } - rc = CIFSSMBWrite2(xid, pTcon, open_file->netfid, - cur_len, *poffset, &written, - to_send, npages, 0); + io_parms.netfid = open_file->netfid; + io_parms.pid = pid; + io_parms.tcon = pTcon; + io_parms.offset = *poffset; + io_parms.length = cur_len; + rc = CIFSSMBWrite2(xid, &io_parms, &written, to_send, + npages, 0); } while (rc == -EAGAIN); for (i = 0; i < npages; i++) @@ -1711,10 +1693,12 @@ cifs_iovec_read(struct file *file, const struct iovec *iov, size_t len, cur_len; int iov_offset = 0; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct cifsFileInfo *open_file; struct smb_com_read_rsp *pSMBr; + struct cifs_io_parms io_parms; char *read_data; + __u32 pid; if (!nr_segs) return 0; @@ -1729,6 +1713,11 @@ cifs_iovec_read(struct file *file, const struct iovec *iov, open_file = file->private_data; pTcon = tlink_tcon(open_file->tlink); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + pid = open_file->pid; + else + pid = current->tgid; + if ((file->f_flags & O_ACCMODE) == O_WRONLY) cFYI(1, "attempting read on write only file instance"); @@ -1744,8 +1733,12 @@ cifs_iovec_read(struct file *file, const struct iovec *iov, if (rc != 0) break; } - rc = CIFSSMBRead(xid, pTcon, open_file->netfid, - cur_len, *poffset, &bytes_read, + io_parms.netfid = open_file->netfid; + io_parms.pid = pid; + io_parms.tcon = pTcon; + io_parms.offset = *poffset; + io_parms.length = len; + rc = CIFSSMBRead(xid, &io_parms, &bytes_read, &read_data, &buf_type); pSMBr = (struct smb_com_read_rsp *)read_data; if (read_data) { @@ -1822,11 +1815,13 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size, unsigned int total_read; unsigned int current_read_size; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; int xid; char *current_offset; struct cifsFileInfo *open_file; + struct cifs_io_parms io_parms; int buf_type = CIFS_NO_BUFFER; + __u32 pid; xid = GetXid(); cifs_sb = CIFS_SB(file->f_path.dentry->d_sb); @@ -1839,6 +1834,11 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size, open_file = file->private_data; pTcon = tlink_tcon(open_file->tlink); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + pid = open_file->pid; + else + pid = current->tgid; + if ((file->f_flags & O_ACCMODE) == O_WRONLY) cFYI(1, "attempting read on write only file instance"); @@ -1861,11 +1861,13 @@ static ssize_t cifs_read(struct file *file, char *read_data, size_t read_size, if (rc != 0) break; } - rc = CIFSSMBRead(xid, pTcon, - open_file->netfid, - current_read_size, *poffset, - &bytes_read, ¤t_offset, - &buf_type); + io_parms.netfid = open_file->netfid; + io_parms.pid = pid; + io_parms.tcon = pTcon; + io_parms.offset = *poffset; + io_parms.length = current_read_size; + rc = CIFSSMBRead(xid, &io_parms, &bytes_read, + ¤t_offset, &buf_type); } if (rc || (bytes_read == 0)) { if (total_read) { @@ -1996,13 +1998,15 @@ static int cifs_readpages(struct file *file, struct address_space *mapping, loff_t offset; struct page *page; struct cifs_sb_info *cifs_sb; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; unsigned int bytes_read = 0; unsigned int read_size, i; char *smb_read_data = NULL; struct smb_com_read_rsp *pSMBr; struct cifsFileInfo *open_file; + struct cifs_io_parms io_parms; int buf_type = CIFS_NO_BUFFER; + __u32 pid; xid = GetXid(); if (file->private_data == NULL) { @@ -2024,6 +2028,11 @@ static int cifs_readpages(struct file *file, struct address_space *mapping, goto read_complete; cFYI(DBG2, "rpages: num pages %d", num_pages); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_RWPIDFORWARD) + pid = open_file->pid; + else + pid = current->tgid; + for (i = 0; i < num_pages; ) { unsigned contig_pages; struct page *tmp_page; @@ -2065,12 +2074,13 @@ static int cifs_readpages(struct file *file, struct address_space *mapping, if (rc != 0) break; } - - rc = CIFSSMBRead(xid, pTcon, - open_file->netfid, - read_size, offset, - &bytes_read, &smb_read_data, - &buf_type); + io_parms.netfid = open_file->netfid; + io_parms.pid = pid; + io_parms.tcon = pTcon; + io_parms.offset = offset; + io_parms.length = read_size; + rc = CIFSSMBRead(xid, &io_parms, &bytes_read, + &smb_read_data, &buf_type); /* BB more RC checks ? */ if (rc == -EAGAIN) { if (smb_read_data) { diff --git a/fs/cifs/fscache.c b/fs/cifs/fscache.c index 297a43d0ff7f..816696621ec9 100644 --- a/fs/cifs/fscache.c +++ b/fs/cifs/fscache.c @@ -28,32 +28,32 @@ void cifs_fscache_get_client_cookie(struct TCP_Server_Info *server) server->fscache = fscache_acquire_cookie(cifs_fscache_netfs.primary_index, &cifs_fscache_server_index_def, server); - cFYI(1, "CIFS: get client cookie (0x%p/0x%p)", server, - server->fscache); + cFYI(1, "%s: (0x%p/0x%p)", __func__, server, + server->fscache); } void cifs_fscache_release_client_cookie(struct TCP_Server_Info *server) { - cFYI(1, "CIFS: release client cookie (0x%p/0x%p)", server, - server->fscache); + cFYI(1, "%s: (0x%p/0x%p)", __func__, server, + server->fscache); fscache_relinquish_cookie(server->fscache, 0); server->fscache = NULL; } -void cifs_fscache_get_super_cookie(struct cifsTconInfo *tcon) +void cifs_fscache_get_super_cookie(struct cifs_tcon *tcon) { struct TCP_Server_Info *server = tcon->ses->server; tcon->fscache = fscache_acquire_cookie(server->fscache, &cifs_fscache_super_index_def, tcon); - cFYI(1, "CIFS: get superblock cookie (0x%p/0x%p)", - server->fscache, tcon->fscache); + cFYI(1, "%s: (0x%p/0x%p)", __func__, server->fscache, + tcon->fscache); } -void cifs_fscache_release_super_cookie(struct cifsTconInfo *tcon) +void cifs_fscache_release_super_cookie(struct cifs_tcon *tcon) { - cFYI(1, "CIFS: releasing superblock cookie (0x%p)", tcon->fscache); + cFYI(1, "%s: (0x%p)", __func__, tcon->fscache); fscache_relinquish_cookie(tcon->fscache, 0); tcon->fscache = NULL; } @@ -62,7 +62,7 @@ static void cifs_fscache_enable_inode_cookie(struct inode *inode) { struct cifsInodeInfo *cifsi = CIFS_I(inode); struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); if (cifsi->fscache) return; @@ -70,8 +70,8 @@ static void cifs_fscache_enable_inode_cookie(struct inode *inode) if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_FSCACHE) { cifsi->fscache = fscache_acquire_cookie(tcon->fscache, &cifs_fscache_inode_object_def, cifsi); - cFYI(1, "CIFS: got FH cookie (0x%p/0x%p)", tcon->fscache, - cifsi->fscache); + cFYI(1, "%s: got FH cookie (0x%p/0x%p)", __func__, + tcon->fscache, cifsi->fscache); } } @@ -80,8 +80,7 @@ void cifs_fscache_release_inode_cookie(struct inode *inode) struct cifsInodeInfo *cifsi = CIFS_I(inode); if (cifsi->fscache) { - cFYI(1, "CIFS releasing inode cookie (0x%p)", - cifsi->fscache); + cFYI(1, "%s: (0x%p)", __func__, cifsi->fscache); fscache_relinquish_cookie(cifsi->fscache, 0); cifsi->fscache = NULL; } @@ -92,8 +91,7 @@ static void cifs_fscache_disable_inode_cookie(struct inode *inode) struct cifsInodeInfo *cifsi = CIFS_I(inode); if (cifsi->fscache) { - cFYI(1, "CIFS disabling inode cookie (0x%p)", - cifsi->fscache); + cFYI(1, "%s: (0x%p)", __func__, cifsi->fscache); fscache_relinquish_cookie(cifsi->fscache, 1); cifsi->fscache = NULL; } @@ -121,8 +119,8 @@ void cifs_fscache_reset_inode_cookie(struct inode *inode) cifs_sb_master_tcon(cifs_sb)->fscache, &cifs_fscache_inode_object_def, cifsi); - cFYI(1, "CIFS: new cookie 0x%p oldcookie 0x%p", - cifsi->fscache, old); + cFYI(1, "%s: new cookie 0x%p oldcookie 0x%p", + __func__, cifsi->fscache, old); } } @@ -132,8 +130,8 @@ int cifs_fscache_release_page(struct page *page, gfp_t gfp) struct inode *inode = page->mapping->host; struct cifsInodeInfo *cifsi = CIFS_I(inode); - cFYI(1, "CIFS: fscache release page (0x%p/0x%p)", - page, cifsi->fscache); + cFYI(1, "%s: (0x%p/0x%p)", __func__, page, + cifsi->fscache); if (!fscache_maybe_release_page(cifsi->fscache, page, gfp)) return 0; } @@ -144,8 +142,7 @@ int cifs_fscache_release_page(struct page *page, gfp_t gfp) static void cifs_readpage_from_fscache_complete(struct page *page, void *ctx, int error) { - cFYI(1, "CFS: readpage_from_fscache_complete (0x%p/%d)", - page, error); + cFYI(1, "%s: (0x%p/%d)", __func__, page, error); if (!error) SetPageUptodate(page); unlock_page(page); @@ -158,7 +155,7 @@ int __cifs_readpage_from_fscache(struct inode *inode, struct page *page) { int ret; - cFYI(1, "CIFS: readpage_from_fscache(fsc:%p, p:%p, i:0x%p", + cFYI(1, "%s: (fsc:%p, p:%p, i:0x%p", __func__, CIFS_I(inode)->fscache, page, inode); ret = fscache_read_or_alloc_page(CIFS_I(inode)->fscache, page, cifs_readpage_from_fscache_complete, @@ -167,11 +164,11 @@ int __cifs_readpage_from_fscache(struct inode *inode, struct page *page) switch (ret) { case 0: /* page found in fscache, read submitted */ - cFYI(1, "CIFS: readpage_from_fscache: submitted"); + cFYI(1, "%s: submitted", __func__); return ret; case -ENOBUFS: /* page won't be cached */ case -ENODATA: /* page not in cache */ - cFYI(1, "CIFS: readpage_from_fscache %d", ret); + cFYI(1, "%s: %d", __func__, ret); return 1; default: @@ -190,7 +187,7 @@ int __cifs_readpages_from_fscache(struct inode *inode, { int ret; - cFYI(1, "CIFS: __cifs_readpages_from_fscache (0x%p/%u/0x%p)", + cFYI(1, "%s: (0x%p/%u/0x%p)", __func__, CIFS_I(inode)->fscache, *nr_pages, inode); ret = fscache_read_or_alloc_pages(CIFS_I(inode)->fscache, mapping, pages, nr_pages, @@ -199,12 +196,12 @@ int __cifs_readpages_from_fscache(struct inode *inode, mapping_gfp_mask(mapping)); switch (ret) { case 0: /* read submitted to the cache for all pages */ - cFYI(1, "CIFS: readpages_from_fscache: submitted"); + cFYI(1, "%s: submitted", __func__); return ret; case -ENOBUFS: /* some pages are not cached and can't be */ case -ENODATA: /* some pages are not cached */ - cFYI(1, "CIFS: readpages_from_fscache: no page"); + cFYI(1, "%s: no page", __func__); return 1; default: @@ -218,7 +215,7 @@ void __cifs_readpage_to_fscache(struct inode *inode, struct page *page) { int ret; - cFYI(1, "CIFS: readpage_to_fscache(fsc: %p, p: %p, i: %p", + cFYI(1, "%s: (fsc: %p, p: %p, i: %p)", __func__, CIFS_I(inode)->fscache, page, inode); ret = fscache_write_page(CIFS_I(inode)->fscache, page, GFP_KERNEL); if (ret != 0) @@ -230,7 +227,7 @@ void __cifs_fscache_invalidate_page(struct page *page, struct inode *inode) struct cifsInodeInfo *cifsi = CIFS_I(inode); struct fscache_cookie *cookie = cifsi->fscache; - cFYI(1, "CIFS: fscache invalidatepage (0x%p/0x%p)", page, cookie); + cFYI(1, "%s: (0x%p/0x%p)", __func__, page, cookie); fscache_wait_on_page_write(cookie, page); fscache_uncache_page(cookie, page); } diff --git a/fs/cifs/fscache.h b/fs/cifs/fscache.h index 31b88ec2341e..63539323e0b9 100644 --- a/fs/cifs/fscache.h +++ b/fs/cifs/fscache.h @@ -40,8 +40,8 @@ extern void cifs_fscache_unregister(void); */ extern void cifs_fscache_get_client_cookie(struct TCP_Server_Info *); extern void cifs_fscache_release_client_cookie(struct TCP_Server_Info *); -extern void cifs_fscache_get_super_cookie(struct cifsTconInfo *); -extern void cifs_fscache_release_super_cookie(struct cifsTconInfo *); +extern void cifs_fscache_get_super_cookie(struct cifs_tcon *); +extern void cifs_fscache_release_super_cookie(struct cifs_tcon *); extern void cifs_fscache_release_inode_cookie(struct inode *); extern void cifs_fscache_set_inode_cookie(struct inode *, struct file *); @@ -99,9 +99,9 @@ static inline void cifs_fscache_get_client_cookie(struct TCP_Server_Info *server) {} static inline void cifs_fscache_release_client_cookie(struct TCP_Server_Info *server) {} -static inline void cifs_fscache_get_super_cookie(struct cifsTconInfo *tcon) {} +static inline void cifs_fscache_get_super_cookie(struct cifs_tcon *tcon) {} static inline void -cifs_fscache_release_super_cookie(struct cifsTconInfo *tcon) {} +cifs_fscache_release_super_cookie(struct cifs_tcon *tcon) {} static inline void cifs_fscache_release_inode_cookie(struct inode *inode) {} static inline void cifs_fscache_set_inode_cookie(struct inode *inode, diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c index de02ed5e25c2..9b018c8334fa 100644 --- a/fs/cifs/inode.c +++ b/fs/cifs/inode.c @@ -295,7 +295,7 @@ int cifs_get_file_info_unix(struct file *filp) struct inode *inode = filp->f_path.dentry->d_inode; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct cifsFileInfo *cfile = filp->private_data; - struct cifsTconInfo *tcon = tlink_tcon(cfile->tlink); + struct cifs_tcon *tcon = tlink_tcon(cfile->tlink); xid = GetXid(); rc = CIFSSMBUnixQFileInfo(xid, tcon, cfile->netfid, &find_data); @@ -318,7 +318,7 @@ int cifs_get_inode_info_unix(struct inode **pinode, int rc; FILE_UNIX_BASIC_INFO find_data; struct cifs_fattr fattr; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct tcon_link *tlink; struct cifs_sb_info *cifs_sb = CIFS_SB(sb); @@ -373,7 +373,8 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path, int oplock = 0; __u16 netfid; struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; + struct cifs_io_parms io_parms; char buf[24]; unsigned int bytes_read; char *pbuf; @@ -405,9 +406,13 @@ cifs_sfu_type(struct cifs_fattr *fattr, const unsigned char *path, if (rc == 0) { int buf_type = CIFS_NO_BUFFER; /* Read header */ - rc = CIFSSMBRead(xid, tcon, netfid, - 24 /* length */, 0 /* offset */, - &bytes_read, &pbuf, &buf_type); + io_parms.netfid = netfid; + io_parms.pid = current->tgid; + io_parms.tcon = tcon; + io_parms.offset = 0; + io_parms.length = 24; + rc = CIFSSMBRead(xid, &io_parms, &bytes_read, &pbuf, + &buf_type); if ((rc == 0) && (bytes_read >= 8)) { if (memcmp("IntxBLK", pbuf, 8) == 0) { cFYI(1, "Block device"); @@ -468,7 +473,7 @@ static int cifs_sfu_mode(struct cifs_fattr *fattr, const unsigned char *path, char ea_value[4]; __u32 mode; struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; tlink = cifs_sb_tlink(cifs_sb); if (IS_ERR(tlink)) @@ -502,7 +507,7 @@ static void cifs_all_info_to_fattr(struct cifs_fattr *fattr, FILE_ALL_INFO *info, struct cifs_sb_info *cifs_sb, bool adjust_tz) { - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); memset(fattr, 0, sizeof(*fattr)); fattr->cf_cifsattrs = le32_to_cpu(info->Attributes); @@ -553,7 +558,7 @@ int cifs_get_file_info(struct file *filp) struct inode *inode = filp->f_path.dentry->d_inode; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct cifsFileInfo *cfile = filp->private_data; - struct cifsTconInfo *tcon = tlink_tcon(cfile->tlink); + struct cifs_tcon *tcon = tlink_tcon(cfile->tlink); xid = GetXid(); rc = CIFSSMBQFileInfo(xid, tcon, cfile->netfid, &find_data); @@ -590,7 +595,7 @@ int cifs_get_inode_info(struct inode **pinode, struct super_block *sb, int xid, const __u16 *pfid) { int rc = 0, tmprc; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct tcon_link *tlink; struct cifs_sb_info *cifs_sb = CIFS_SB(sb); char *buf = NULL; @@ -735,10 +740,10 @@ static const struct inode_operations cifs_ipc_inode_ops = { .lookup = cifs_lookup, }; -char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb, - struct cifsTconInfo *tcon) +char *cifs_build_path_to_root(struct smb_vol *vol, struct cifs_sb_info *cifs_sb, + struct cifs_tcon *tcon) { - int pplen = cifs_sb->prepathlen; + int pplen = vol->prepath ? strlen(vol->prepath) : 0; int dfsplen; char *full_path = NULL; @@ -772,7 +777,7 @@ char *cifs_build_path_to_root(struct cifs_sb_info *cifs_sb, } } } - strncpy(full_path + dfsplen, cifs_sb->prepath, pplen); + strncpy(full_path + dfsplen, vol->prepath, pplen); full_path[dfsplen + pplen] = 0; /* add trailing null */ return full_path; } @@ -884,19 +889,13 @@ struct inode *cifs_root_iget(struct super_block *sb) struct cifs_sb_info *cifs_sb = CIFS_SB(sb); struct inode *inode = NULL; long rc; - char *full_path; - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); - - full_path = cifs_build_path_to_root(cifs_sb, tcon); - if (full_path == NULL) - return ERR_PTR(-ENOMEM); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); xid = GetXid(); if (tcon->unix_ext) - rc = cifs_get_inode_info_unix(&inode, full_path, sb, xid); + rc = cifs_get_inode_info_unix(&inode, "", sb, xid); else - rc = cifs_get_inode_info(&inode, full_path, NULL, sb, - xid, NULL); + rc = cifs_get_inode_info(&inode, "", NULL, sb, xid, NULL); if (!inode) { inode = ERR_PTR(rc); @@ -922,7 +921,6 @@ struct inode *cifs_root_iget(struct super_block *sb) } out: - kfree(full_path); /* can not call macro FreeXid here since in a void func * TODO: This is no longer true */ @@ -943,7 +941,7 @@ cifs_set_file_info(struct inode *inode, struct iattr *attrs, int xid, struct cifsInodeInfo *cifsInode = CIFS_I(inode); struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink = NULL; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; FILE_BASIC_INFO info_buf; if (attrs == NULL) @@ -1061,7 +1059,7 @@ cifs_rename_pending_delete(char *full_path, struct dentry *dentry, int xid) struct cifsInodeInfo *cifsInode = CIFS_I(inode); struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; __u32 dosattr, origattr; FILE_BASIC_INFO *info_buf = NULL; @@ -1179,7 +1177,7 @@ int cifs_unlink(struct inode *dir, struct dentry *dentry) struct super_block *sb = dir->i_sb; struct cifs_sb_info *cifs_sb = CIFS_SB(sb); struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; struct iattr *attrs = NULL; __u32 dosattr = 0, origattr = 0; @@ -1277,7 +1275,7 @@ int cifs_mkdir(struct inode *inode, struct dentry *direntry, int mode) int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; char *full_path = NULL; struct inode *newinode = NULL; struct cifs_fattr fattr; @@ -1455,7 +1453,7 @@ int cifs_rmdir(struct inode *inode, struct dentry *direntry) int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; char *full_path = NULL; struct cifsInodeInfo *cifsInode; @@ -1512,7 +1510,7 @@ cifs_do_rename(int xid, struct dentry *from_dentry, const char *fromPath, { struct cifs_sb_info *cifs_sb = CIFS_SB(from_dentry->d_sb); struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; __u16 srcfid; int oplock, rc; @@ -1564,7 +1562,7 @@ int cifs_rename(struct inode *source_dir, struct dentry *source_dentry, char *toName = NULL; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; FILE_UNIX_BASIC_INFO *info_buf_source = NULL; FILE_UNIX_BASIC_INFO *info_buf_target; int xid, rc, tmprc; @@ -1794,7 +1792,7 @@ int cifs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) { struct cifs_sb_info *cifs_sb = CIFS_SB(dentry->d_sb); - struct cifsTconInfo *tcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *tcon = cifs_sb_master_tcon(cifs_sb); struct inode *inode = dentry->d_inode; int rc; @@ -1872,7 +1870,8 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs, struct cifsInodeInfo *cifsInode = CIFS_I(inode); struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink = NULL; - struct cifsTconInfo *pTcon = NULL; + struct cifs_tcon *pTcon = NULL; + struct cifs_io_parms io_parms; /* * To avoid spurious oplock breaks from server, in the case of @@ -1894,8 +1893,14 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs, cFYI(1, "SetFSize for attrs rc = %d", rc); if ((rc == -EINVAL) || (rc == -EOPNOTSUPP)) { unsigned int bytes_written; - rc = CIFSSMBWrite(xid, pTcon, nfid, 0, attrs->ia_size, - &bytes_written, NULL, NULL, 1); + + io_parms.netfid = nfid; + io_parms.pid = npid; + io_parms.tcon = pTcon; + io_parms.offset = 0; + io_parms.length = attrs->ia_size; + rc = CIFSSMBWrite(xid, &io_parms, &bytes_written, + NULL, NULL, 1); cFYI(1, "Wrt seteof rc %d", rc); } } else @@ -1930,10 +1935,15 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs, CIFS_MOUNT_MAP_SPECIAL_CHR); if (rc == 0) { unsigned int bytes_written; - rc = CIFSSMBWrite(xid, pTcon, netfid, 0, - attrs->ia_size, - &bytes_written, NULL, - NULL, 1); + + io_parms.netfid = netfid; + io_parms.pid = current->tgid; + io_parms.tcon = pTcon; + io_parms.offset = 0; + io_parms.length = attrs->ia_size; + rc = CIFSSMBWrite(xid, &io_parms, + &bytes_written, + NULL, NULL, 1); cFYI(1, "wrt seteof rc %d", rc); CIFSSMBClose(xid, pTcon, netfid); } @@ -1961,7 +1971,7 @@ cifs_setattr_unix(struct dentry *direntry, struct iattr *attrs) struct cifsInodeInfo *cifsInode = CIFS_I(inode); struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct cifs_unix_set_info_args *args = NULL; struct cifsFileInfo *open_file; @@ -2247,7 +2257,7 @@ cifs_setattr(struct dentry *direntry, struct iattr *attrs) { struct inode *inode = direntry->d_inode; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); - struct cifsTconInfo *pTcon = cifs_sb_master_tcon(cifs_sb); + struct cifs_tcon *pTcon = cifs_sb_master_tcon(cifs_sb); if (pTcon->unix_ext) return cifs_setattr_unix(direntry, attrs); diff --git a/fs/cifs/ioctl.c b/fs/cifs/ioctl.c index 0c98672d0122..4221b5e48a42 100644 --- a/fs/cifs/ioctl.c +++ b/fs/cifs/ioctl.c @@ -38,7 +38,7 @@ long cifs_ioctl(struct file *filep, unsigned int command, unsigned long arg) struct cifs_sb_info *cifs_sb; #ifdef CONFIG_CIFS_POSIX struct cifsFileInfo *pSMBFile = filep->private_data; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; __u64 ExtAttrBits = 0; __u64 ExtAttrMask = 0; __u64 caps; diff --git a/fs/cifs/link.c b/fs/cifs/link.c index ce417a9764a3..556b1a0b54de 100644 --- a/fs/cifs/link.c +++ b/fs/cifs/link.c @@ -175,7 +175,7 @@ CIFSFormatMFSymlink(u8 *buf, unsigned int buf_len, const char *link_str) } static int -CIFSCreateMFSymLink(const int xid, struct cifsTconInfo *tcon, +CIFSCreateMFSymLink(const int xid, struct cifs_tcon *tcon, const char *fromName, const char *toName, const struct nls_table *nls_codepage, int remap) { @@ -184,6 +184,7 @@ CIFSCreateMFSymLink(const int xid, struct cifsTconInfo *tcon, __u16 netfid = 0; u8 *buf; unsigned int bytes_written = 0; + struct cifs_io_parms io_parms; buf = kmalloc(CIFS_MF_SYMLINK_FILE_SIZE, GFP_KERNEL); if (!buf) @@ -203,10 +204,13 @@ CIFSCreateMFSymLink(const int xid, struct cifsTconInfo *tcon, return rc; } - rc = CIFSSMBWrite(xid, tcon, netfid, - CIFS_MF_SYMLINK_FILE_SIZE /* length */, - 0 /* offset */, - &bytes_written, buf, NULL, 0); + io_parms.netfid = netfid; + io_parms.pid = current->tgid; + io_parms.tcon = tcon; + io_parms.offset = 0; + io_parms.length = CIFS_MF_SYMLINK_FILE_SIZE; + + rc = CIFSSMBWrite(xid, &io_parms, &bytes_written, buf, NULL, 0); CIFSSMBClose(xid, tcon, netfid); kfree(buf); if (rc != 0) @@ -219,7 +223,7 @@ CIFSCreateMFSymLink(const int xid, struct cifsTconInfo *tcon, } static int -CIFSQueryMFSymLink(const int xid, struct cifsTconInfo *tcon, +CIFSQueryMFSymLink(const int xid, struct cifs_tcon *tcon, const unsigned char *searchName, char **symlinkinfo, const struct nls_table *nls_codepage, int remap) { @@ -231,6 +235,7 @@ CIFSQueryMFSymLink(const int xid, struct cifsTconInfo *tcon, unsigned int bytes_read = 0; int buf_type = CIFS_NO_BUFFER; unsigned int link_len = 0; + struct cifs_io_parms io_parms; FILE_ALL_INFO file_info; rc = CIFSSMBOpen(xid, tcon, searchName, FILE_OPEN, GENERIC_READ, @@ -249,11 +254,13 @@ CIFSQueryMFSymLink(const int xid, struct cifsTconInfo *tcon, if (!buf) return -ENOMEM; pbuf = buf; + io_parms.netfid = netfid; + io_parms.pid = current->tgid; + io_parms.tcon = tcon; + io_parms.offset = 0; + io_parms.length = CIFS_MF_SYMLINK_FILE_SIZE; - rc = CIFSSMBRead(xid, tcon, netfid, - CIFS_MF_SYMLINK_FILE_SIZE /* length */, - 0 /* offset */, - &bytes_read, &pbuf, &buf_type); + rc = CIFSSMBRead(xid, &io_parms, &bytes_read, &pbuf, &buf_type); CIFSSMBClose(xid, tcon, netfid); if (rc != 0) { kfree(buf); @@ -291,7 +298,8 @@ CIFSCheckMFSymlink(struct cifs_fattr *fattr, int oplock = 0; __u16 netfid = 0; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; + struct cifs_io_parms io_parms; u8 *buf; char *pbuf; unsigned int bytes_read = 0; @@ -328,11 +336,13 @@ CIFSCheckMFSymlink(struct cifs_fattr *fattr, goto out; } pbuf = buf; + io_parms.netfid = netfid; + io_parms.pid = current->tgid; + io_parms.tcon = pTcon; + io_parms.offset = 0; + io_parms.length = CIFS_MF_SYMLINK_FILE_SIZE; - rc = CIFSSMBRead(xid, pTcon, netfid, - CIFS_MF_SYMLINK_FILE_SIZE /* length */, - 0 /* offset */, - &bytes_read, &pbuf, &buf_type); + rc = CIFSSMBRead(xid, &io_parms, &bytes_read, &pbuf, &buf_type); CIFSSMBClose(xid, pTcon, netfid); if (rc != 0) { kfree(buf); @@ -370,7 +380,7 @@ cifs_hardlink(struct dentry *old_file, struct inode *inode, char *toName = NULL; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct cifsInodeInfo *cifsInode; tlink = cifs_sb_tlink(cifs_sb); @@ -445,7 +455,7 @@ cifs_follow_link(struct dentry *direntry, struct nameidata *nd) char *target_path = NULL; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink = NULL; - struct cifsTconInfo *tcon; + struct cifs_tcon *tcon; xid = GetXid(); @@ -518,7 +528,7 @@ cifs_symlink(struct inode *inode, struct dentry *direntry, const char *symname) int xid; struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb); struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; char *full_path = NULL; struct inode *newinode = NULL; diff --git a/fs/cifs/misc.c b/fs/cifs/misc.c index 907531ac5888..03a1f491d39b 100644 --- a/fs/cifs/misc.c +++ b/fs/cifs/misc.c @@ -67,12 +67,12 @@ _FreeXid(unsigned int xid) spin_unlock(&GlobalMid_Lock); } -struct cifsSesInfo * +struct cifs_ses * sesInfoAlloc(void) { - struct cifsSesInfo *ret_buf; + struct cifs_ses *ret_buf; - ret_buf = kzalloc(sizeof(struct cifsSesInfo), GFP_KERNEL); + ret_buf = kzalloc(sizeof(struct cifs_ses), GFP_KERNEL); if (ret_buf) { atomic_inc(&sesInfoAllocCount); ret_buf->status = CifsNew; @@ -85,7 +85,7 @@ sesInfoAlloc(void) } void -sesInfoFree(struct cifsSesInfo *buf_to_free) +sesInfoFree(struct cifs_ses *buf_to_free) { if (buf_to_free == NULL) { cFYI(1, "Null buffer passed to sesInfoFree"); @@ -105,11 +105,11 @@ sesInfoFree(struct cifsSesInfo *buf_to_free) kfree(buf_to_free); } -struct cifsTconInfo * +struct cifs_tcon * tconInfoAlloc(void) { - struct cifsTconInfo *ret_buf; - ret_buf = kzalloc(sizeof(struct cifsTconInfo), GFP_KERNEL); + struct cifs_tcon *ret_buf; + ret_buf = kzalloc(sizeof(struct cifs_tcon), GFP_KERNEL); if (ret_buf) { atomic_inc(&tconInfoAllocCount); ret_buf->tidStatus = CifsNew; @@ -124,7 +124,7 @@ tconInfoAlloc(void) } void -tconInfoFree(struct cifsTconInfo *buf_to_free) +tconInfoFree(struct cifs_tcon *buf_to_free) { if (buf_to_free == NULL) { cFYI(1, "Null buffer passed to tconInfoFree"); @@ -295,11 +295,11 @@ __u16 GetNextMid(struct TCP_Server_Info *server) case it is responsbility of caller to set the mid */ void header_assemble(struct smb_hdr *buffer, char smb_command /* command */ , - const struct cifsTconInfo *treeCon, int word_count + const struct cifs_tcon *treeCon, int word_count /* length of fixed section (word count) in two byte units */) { struct list_head *temp_item; - struct cifsSesInfo *ses; + struct cifs_ses *ses; char *temp = (char *) buffer; memset(temp, 0, 256); /* bigger than MAX_CIFS_HDR_SIZE */ @@ -359,7 +359,7 @@ header_assemble(struct smb_hdr *buffer, char smb_command /* command */ , "did not match tcon uid"); spin_lock(&cifs_tcp_ses_lock); list_for_each(temp_item, &treeCon->ses->server->smb_ses_list) { - ses = list_entry(temp_item, struct cifsSesInfo, smb_ses_list); + ses = list_entry(temp_item, struct cifs_ses, smb_ses_list); if (ses->linux_uid == current_fsuid()) { if (ses->server == treeCon->ses->server) { cFYI(1, "found matching uid substitute right smb_uid"); @@ -380,7 +380,7 @@ header_assemble(struct smb_hdr *buffer, char smb_command /* command */ , if (treeCon->nocase) buffer->Flags |= SMBFLG_CASELESS; if ((treeCon->ses) && (treeCon->ses->server)) - if (treeCon->ses->server->secMode & + if (treeCon->ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) buffer->Flags2 |= SMBFLG2_SECURITY_SIGNATURE; } @@ -507,8 +507,8 @@ is_valid_oplock_break(struct smb_hdr *buf, struct TCP_Server_Info *srv) { struct smb_com_lock_req *pSMB = (struct smb_com_lock_req *)buf; struct list_head *tmp, *tmp1, *tmp2; - struct cifsSesInfo *ses; - struct cifsTconInfo *tcon; + struct cifs_ses *ses; + struct cifs_tcon *tcon; struct cifsInodeInfo *pCifsInode; struct cifsFileInfo *netfile; @@ -566,9 +566,9 @@ is_valid_oplock_break(struct smb_hdr *buf, struct TCP_Server_Info *srv) /* look up tcon based on tid & uid */ spin_lock(&cifs_tcp_ses_lock); list_for_each(tmp, &srv->smb_ses_list) { - ses = list_entry(tmp, struct cifsSesInfo, smb_ses_list); + ses = list_entry(tmp, struct cifs_ses, smb_ses_list); list_for_each(tmp1, &ses->tcon_list) { - tcon = list_entry(tmp1, struct cifsTconInfo, tcon_list); + tcon = list_entry(tmp1, struct cifs_tcon, tcon_list); if (tcon->tid != buf->Tid) continue; diff --git a/fs/cifs/netmisc.c b/fs/cifs/netmisc.c index 79b71c2c7c9d..73e47e84b61a 100644 --- a/fs/cifs/netmisc.c +++ b/fs/cifs/netmisc.c @@ -836,7 +836,7 @@ ntstatus_to_dos(__u32 ntstatus, __u8 *eclass, __u16 *ecode) } int -map_smb_to_linux_error(struct smb_hdr *smb, int logErr) +map_smb_to_linux_error(struct smb_hdr *smb, bool logErr) { unsigned int i; int rc = -EIO; /* if transport error smb error may not be set */ diff --git a/fs/cifs/readdir.c b/fs/cifs/readdir.c index f8e4cd2a7912..6751e745bbc6 100644 --- a/fs/cifs/readdir.c +++ b/fs/cifs/readdir.c @@ -195,7 +195,7 @@ int get_symlink_reparse_path(char *full_path, struct cifs_sb_info *cifs_sb, int len; int oplock = 0; int rc; - struct cifsTconInfo *ptcon = cifs_sb_tcon(cifs_sb); + struct cifs_tcon *ptcon = cifs_sb_tcon(cifs_sb); char *tmpbuffer; rc = CIFSSMBOpen(xid, ptcon, full_path, FILE_OPEN, GENERIC_READ, @@ -223,7 +223,7 @@ static int initiate_cifs_search(const int xid, struct file *file) struct cifsFileInfo *cifsFile; struct cifs_sb_info *cifs_sb = CIFS_SB(file->f_path.dentry->d_sb); struct tcon_link *tlink = NULL; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; if (file->private_data == NULL) { tlink = cifs_sb_tlink(cifs_sb); @@ -496,7 +496,7 @@ static int cifs_save_resume_key(const char *current_entry, assume that they are located in the findfirst return buffer.*/ /* We start counting in the buffer with entry 2 and increment for every entry (do not increment for . or .. entry) */ -static int find_cifs_entry(const int xid, struct cifsTconInfo *pTcon, +static int find_cifs_entry(const int xid, struct cifs_tcon *pTcon, struct file *file, char **ppCurrentEntry, int *num_to_ret) { int rc = 0; @@ -764,7 +764,7 @@ int cifs_readdir(struct file *file, void *direntry, filldir_t filldir) { int rc = 0; int xid, i; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct cifsFileInfo *cifsFile = NULL; char *current_entry; int num_to_fill = 0; diff --git a/fs/cifs/sess.c b/fs/cifs/sess.c index 7dd462100378..3892ab817a36 100644 --- a/fs/cifs/sess.c +++ b/fs/cifs/sess.c @@ -37,13 +37,13 @@ * the socket has been reestablished (so we know whether to use vc 0). * Called while holding the cifs_tcp_ses_lock, so do not block */ -static bool is_first_ses_reconnect(struct cifsSesInfo *ses) +static bool is_first_ses_reconnect(struct cifs_ses *ses) { struct list_head *tmp; - struct cifsSesInfo *tmp_ses; + struct cifs_ses *tmp_ses; list_for_each(tmp, &ses->server->smb_ses_list) { - tmp_ses = list_entry(tmp, struct cifsSesInfo, + tmp_ses = list_entry(tmp, struct cifs_ses, smb_ses_list); if (tmp_ses->need_reconnect == false) return false; @@ -61,11 +61,11 @@ static bool is_first_ses_reconnect(struct cifsSesInfo *ses) * any vc but zero (some servers reset the connection on vcnum zero) * */ -static __le16 get_next_vcnum(struct cifsSesInfo *ses) +static __le16 get_next_vcnum(struct cifs_ses *ses) { __u16 vcnum = 0; struct list_head *tmp; - struct cifsSesInfo *tmp_ses; + struct cifs_ses *tmp_ses; __u16 max_vcs = ses->server->max_vcs; __u16 i; int free_vc_found = 0; @@ -87,7 +87,7 @@ static __le16 get_next_vcnum(struct cifsSesInfo *ses) free_vc_found = 1; list_for_each(tmp, &ses->server->smb_ses_list) { - tmp_ses = list_entry(tmp, struct cifsSesInfo, + tmp_ses = list_entry(tmp, struct cifs_ses, smb_ses_list); if (tmp_ses->vcnum == i) { free_vc_found = 0; @@ -114,7 +114,7 @@ get_vc_num_exit: return cpu_to_le16(vcnum); } -static __u32 cifs_ssetup_hdr(struct cifsSesInfo *ses, SESSION_SETUP_ANDX *pSMB) +static __u32 cifs_ssetup_hdr(struct cifs_ses *ses, SESSION_SETUP_ANDX *pSMB) { __u32 capabilities = 0; @@ -136,7 +136,7 @@ static __u32 cifs_ssetup_hdr(struct cifsSesInfo *ses, SESSION_SETUP_ANDX *pSMB) capabilities = CAP_LARGE_FILES | CAP_NT_SMBS | CAP_LEVEL_II_OPLOCKS | CAP_LARGE_WRITE_X | CAP_LARGE_READ_X; - if (ses->server->secMode & + if (ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) pSMB->req.hdr.Flags2 |= SMBFLG2_SECURITY_SIGNATURE; @@ -181,7 +181,7 @@ unicode_oslm_strings(char **pbcc_area, const struct nls_table *nls_cp) *pbcc_area = bcc_ptr; } -static void unicode_domain_string(char **pbcc_area, struct cifsSesInfo *ses, +static void unicode_domain_string(char **pbcc_area, struct cifs_ses *ses, const struct nls_table *nls_cp) { char *bcc_ptr = *pbcc_area; @@ -204,7 +204,7 @@ static void unicode_domain_string(char **pbcc_area, struct cifsSesInfo *ses, } -static void unicode_ssetup_strings(char **pbcc_area, struct cifsSesInfo *ses, +static void unicode_ssetup_strings(char **pbcc_area, struct cifs_ses *ses, const struct nls_table *nls_cp) { char *bcc_ptr = *pbcc_area; @@ -236,7 +236,7 @@ static void unicode_ssetup_strings(char **pbcc_area, struct cifsSesInfo *ses, *pbcc_area = bcc_ptr; } -static void ascii_ssetup_strings(char **pbcc_area, struct cifsSesInfo *ses, +static void ascii_ssetup_strings(char **pbcc_area, struct cifs_ses *ses, const struct nls_table *nls_cp) { char *bcc_ptr = *pbcc_area; @@ -276,7 +276,7 @@ static void ascii_ssetup_strings(char **pbcc_area, struct cifsSesInfo *ses, } static void -decode_unicode_ssetup(char **pbcc_area, int bleft, struct cifsSesInfo *ses, +decode_unicode_ssetup(char **pbcc_area, int bleft, struct cifs_ses *ses, const struct nls_table *nls_cp) { int len; @@ -310,7 +310,7 @@ decode_unicode_ssetup(char **pbcc_area, int bleft, struct cifsSesInfo *ses, } static int decode_ascii_ssetup(char **pbcc_area, __u16 bleft, - struct cifsSesInfo *ses, + struct cifs_ses *ses, const struct nls_table *nls_cp) { int rc = 0; @@ -364,7 +364,7 @@ static int decode_ascii_ssetup(char **pbcc_area, __u16 bleft, } static int decode_ntlmssp_challenge(char *bcc_ptr, int blob_len, - struct cifsSesInfo *ses) + struct cifs_ses *ses) { unsigned int tioffset; /* challenge message target info area */ unsigned int tilen; /* challenge message target info area length */ @@ -411,7 +411,7 @@ static int decode_ntlmssp_challenge(char *bcc_ptr, int blob_len, /* We do not malloc the blob, it is passed in pbuffer, because it is fixed size, and small, making this approach cleaner */ static void build_ntlmssp_negotiate_blob(unsigned char *pbuffer, - struct cifsSesInfo *ses) + struct cifs_ses *ses) { NEGOTIATE_MESSAGE *sec_blob = (NEGOTIATE_MESSAGE *)pbuffer; __u32 flags; @@ -424,7 +424,7 @@ static void build_ntlmssp_negotiate_blob(unsigned char *pbuffer, flags = NTLMSSP_NEGOTIATE_56 | NTLMSSP_REQUEST_TARGET | NTLMSSP_NEGOTIATE_128 | NTLMSSP_NEGOTIATE_UNICODE | NTLMSSP_NEGOTIATE_NTLM | NTLMSSP_NEGOTIATE_EXTENDED_SEC; - if (ses->server->secMode & + if (ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) { flags |= NTLMSSP_NEGOTIATE_SIGN; if (!ses->server->session_estab) @@ -449,7 +449,7 @@ static void build_ntlmssp_negotiate_blob(unsigned char *pbuffer, This function returns the length of the data in the blob */ static int build_ntlmssp_auth_blob(unsigned char *pbuffer, u16 *buflen, - struct cifsSesInfo *ses, + struct cifs_ses *ses, const struct nls_table *nls_cp) { int rc; @@ -464,10 +464,10 @@ static int build_ntlmssp_auth_blob(unsigned char *pbuffer, NTLMSSP_REQUEST_TARGET | NTLMSSP_NEGOTIATE_TARGET_INFO | NTLMSSP_NEGOTIATE_128 | NTLMSSP_NEGOTIATE_UNICODE | NTLMSSP_NEGOTIATE_NTLM | NTLMSSP_NEGOTIATE_EXTENDED_SEC; - if (ses->server->secMode & + if (ses->server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) flags |= NTLMSSP_NEGOTIATE_SIGN; - if (ses->server->secMode & SECMODE_SIGN_REQUIRED) + if (ses->server->sec_mode & SECMODE_SIGN_REQUIRED) flags |= NTLMSSP_NEGOTIATE_ALWAYS_SIGN; tmp = pbuffer + sizeof(AUTHENTICATE_MESSAGE); @@ -551,7 +551,7 @@ setup_ntlmv2_ret: } int -CIFS_SessSetup(unsigned int xid, struct cifsSesInfo *ses, +CIFS_SessSetup(unsigned int xid, struct cifs_ses *ses, const struct nls_table *nls_cp) { int rc = 0; @@ -657,7 +657,7 @@ ssetup_ntlmssp_authenticate: */ rc = calc_lanman_hash(ses->password, ses->server->cryptkey, - ses->server->secMode & SECMODE_PW_ENCRYPT ? + ses->server->sec_mode & SECMODE_PW_ENCRYPT ? true : false, lnm_session_key); ses->flags |= CIFS_SES_LANMAN; diff --git a/fs/cifs/smbencrypt.c b/fs/cifs/smbencrypt.c index 1525d5e662b6..1c5b770c3141 100644 --- a/fs/cifs/smbencrypt.c +++ b/fs/cifs/smbencrypt.c @@ -90,12 +90,10 @@ smbhash(unsigned char *out, const unsigned char *in, unsigned char *key) sg_init_one(&sgout, out, 8); rc = crypto_blkcipher_encrypt(&desc, &sgout, &sgin, 8); - if (rc) { + if (rc) cERROR(1, "could not encrypt crypt key rc: %d\n", rc); - crypto_free_blkcipher(tfm_des); - goto smbhash_err; - } + crypto_free_blkcipher(tfm_des); smbhash_err: return rc; } diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c index f2513fb8c391..147aa22c3c3a 100644 --- a/fs/cifs/transport.c +++ b/fs/cifs/transport.c @@ -295,7 +295,7 @@ static int wait_for_free_request(struct TCP_Server_Info *server, return 0; } -static int allocate_mid(struct cifsSesInfo *ses, struct smb_hdr *in_buf, +static int allocate_mid(struct cifs_ses *ses, struct smb_hdr *in_buf, struct mid_q_entry **ppmidQ) { if (ses->server->tcpStatus == CifsExiting) { @@ -342,22 +342,24 @@ wait_for_response(struct TCP_Server_Info *server, struct mid_q_entry *midQ) * the result. Caller is responsible for dealing with timeouts. */ int -cifs_call_async(struct TCP_Server_Info *server, struct smb_hdr *in_buf, - mid_callback_t *callback, void *cbdata) +cifs_call_async(struct TCP_Server_Info *server, struct kvec *iov, + unsigned int nvec, mid_callback_t *callback, void *cbdata, + bool ignore_pend) { int rc; struct mid_q_entry *mid; + struct smb_hdr *hdr = (struct smb_hdr *)iov[0].iov_base; - rc = wait_for_free_request(server, CIFS_ASYNC_OP); + rc = wait_for_free_request(server, ignore_pend ? CIFS_ASYNC_OP : 0); if (rc) return rc; /* enable signing if server requires it */ - if (server->secMode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) - in_buf->Flags2 |= SMBFLG2_SECURITY_SIGNATURE; + if (server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) + hdr->Flags2 |= SMBFLG2_SECURITY_SIGNATURE; mutex_lock(&server->srv_mutex); - mid = AllocMidQEntry(in_buf, server); + mid = AllocMidQEntry(hdr, server); if (mid == NULL) { mutex_unlock(&server->srv_mutex); return -ENOMEM; @@ -368,7 +370,7 @@ cifs_call_async(struct TCP_Server_Info *server, struct smb_hdr *in_buf, list_add_tail(&mid->qhead, &server->pending_mid_q); spin_unlock(&GlobalMid_Lock); - rc = cifs_sign_smb(in_buf, server, &mid->sequence_number); + rc = cifs_sign_smb2(iov, nvec, server, &mid->sequence_number); if (rc) { mutex_unlock(&server->srv_mutex); goto out_err; @@ -380,7 +382,7 @@ cifs_call_async(struct TCP_Server_Info *server, struct smb_hdr *in_buf, #ifdef CONFIG_CIFS_STATS2 atomic_inc(&server->inSend); #endif - rc = smb_send(server, in_buf, be32_to_cpu(in_buf->smb_buf_length)); + rc = smb_sendv(server, iov, nvec); #ifdef CONFIG_CIFS_STATS2 atomic_dec(&server->inSend); mid->when_sent = jiffies; @@ -407,7 +409,7 @@ out_err: * */ int -SendReceiveNoRsp(const unsigned int xid, struct cifsSesInfo *ses, +SendReceiveNoRsp(const unsigned int xid, struct cifs_ses *ses, struct smb_hdr *in_buf, int flags) { int rc; @@ -424,7 +426,7 @@ SendReceiveNoRsp(const unsigned int xid, struct cifsSesInfo *ses, } static int -sync_mid_result(struct mid_q_entry *mid, struct TCP_Server_Info *server) +cifs_sync_mid_result(struct mid_q_entry *mid, struct TCP_Server_Info *server) { int rc = 0; @@ -432,28 +434,21 @@ sync_mid_result(struct mid_q_entry *mid, struct TCP_Server_Info *server) mid->mid, mid->midState); spin_lock(&GlobalMid_Lock); - /* ensure that it's no longer on the pending_mid_q */ - list_del_init(&mid->qhead); - switch (mid->midState) { case MID_RESPONSE_RECEIVED: spin_unlock(&GlobalMid_Lock); return rc; - case MID_REQUEST_SUBMITTED: - /* socket is going down, reject all calls */ - if (server->tcpStatus == CifsExiting) { - cERROR(1, "%s: canceling mid=%d cmd=0x%x state=%d", - __func__, mid->mid, mid->command, mid->midState); - rc = -EHOSTDOWN; - break; - } case MID_RETRY_NEEDED: rc = -EAGAIN; break; case MID_RESPONSE_MALFORMED: rc = -EIO; break; + case MID_SHUTDOWN: + rc = -EHOSTDOWN; + break; default: + list_del_init(&mid->qhead); cERROR(1, "%s: invalid mid state mid=%d state=%d", __func__, mid->mid, mid->midState); rc = -EIO; @@ -502,13 +497,31 @@ send_nt_cancel(struct TCP_Server_Info *server, struct smb_hdr *in_buf, } int -SendReceive2(const unsigned int xid, struct cifsSesInfo *ses, +cifs_check_receive(struct mid_q_entry *mid, struct TCP_Server_Info *server, + bool log_error) +{ + dump_smb(mid->resp_buf, + min_t(u32, 92, be32_to_cpu(mid->resp_buf->smb_buf_length))); + + /* convert the length into a more usable form */ + if (server->sec_mode & (SECMODE_SIGN_REQUIRED | SECMODE_SIGN_ENABLED)) { + /* FIXME: add code to kill session */ + if (cifs_verify_signature(mid->resp_buf, server, + mid->sequence_number + 1) != 0) + cERROR(1, "Unexpected SMB signature"); + } + + /* BB special case reconnect tid and uid here? */ + return map_smb_to_linux_error(mid->resp_buf, log_error); +} + +int +SendReceive2(const unsigned int xid, struct cifs_ses *ses, struct kvec *iov, int n_vec, int *pRespBufType /* ret */, const int flags) { int rc = 0; int long_op; - unsigned int receive_len; struct mid_q_entry *midQ; struct smb_hdr *in_buf = iov[0].iov_base; @@ -598,61 +611,31 @@ SendReceive2(const unsigned int xid, struct cifsSesInfo *ses, cifs_small_buf_release(in_buf); - rc = sync_mid_result(midQ, ses->server); + rc = cifs_sync_mid_result(midQ, ses->server); if (rc != 0) { atomic_dec(&ses->server->inFlight); wake_up(&ses->server->request_q); return rc; } - receive_len = be32_to_cpu(midQ->resp_buf->smb_buf_length); - - if (receive_len > CIFSMaxBufSize + MAX_CIFS_HDR_SIZE) { - cERROR(1, "Frame too large received. Length: %d Xid: %d", - receive_len, xid); + if (!midQ->resp_buf || midQ->midState != MID_RESPONSE_RECEIVED) { rc = -EIO; + cFYI(1, "Bad MID state?"); goto out; } - /* rcvd frame is ok */ - - if (midQ->resp_buf && - (midQ->midState == MID_RESPONSE_RECEIVED)) { - - iov[0].iov_base = (char *)midQ->resp_buf; - if (midQ->largeBuf) - *pRespBufType = CIFS_LARGE_BUFFER; - else - *pRespBufType = CIFS_SMALL_BUFFER; - iov[0].iov_len = receive_len + 4; - - dump_smb(midQ->resp_buf, 80); - /* convert the length into a more usable form */ - if ((receive_len > 24) && - (ses->server->secMode & (SECMODE_SIGN_REQUIRED | - SECMODE_SIGN_ENABLED))) { - rc = cifs_verify_signature(midQ->resp_buf, - ses->server, - midQ->sequence_number+1); - if (rc) { - cERROR(1, "Unexpected SMB signature"); - /* BB FIXME add code to kill session */ - } - } - - /* BB special case reconnect tid and uid here? */ - rc = map_smb_to_linux_error(midQ->resp_buf, - flags & CIFS_LOG_ERROR); + iov[0].iov_base = (char *)midQ->resp_buf; + iov[0].iov_len = be32_to_cpu(midQ->resp_buf->smb_buf_length) + 4; + if (midQ->largeBuf) + *pRespBufType = CIFS_LARGE_BUFFER; + else + *pRespBufType = CIFS_SMALL_BUFFER; - if ((flags & CIFS_NO_RESP) == 0) - midQ->resp_buf = NULL; /* mark it so buf will - not be freed by - delete_mid */ - } else { - rc = -EIO; - cFYI(1, "Bad MID state?"); - } + rc = cifs_check_receive(midQ, ses->server, flags & CIFS_LOG_ERROR); + /* mark it so buf will not be freed by delete_mid */ + if ((flags & CIFS_NO_RESP) == 0) + midQ->resp_buf = NULL; out: delete_mid(midQ); atomic_dec(&ses->server->inFlight); @@ -662,12 +645,11 @@ out: } int -SendReceive(const unsigned int xid, struct cifsSesInfo *ses, +SendReceive(const unsigned int xid, struct cifs_ses *ses, struct smb_hdr *in_buf, struct smb_hdr *out_buf, int *pbytes_returned, const int long_op) { int rc = 0; - unsigned int receive_len; struct mid_q_entry *midQ; if (ses == NULL) { @@ -750,54 +732,23 @@ SendReceive(const unsigned int xid, struct cifsSesInfo *ses, spin_unlock(&GlobalMid_Lock); } - rc = sync_mid_result(midQ, ses->server); + rc = cifs_sync_mid_result(midQ, ses->server); if (rc != 0) { atomic_dec(&ses->server->inFlight); wake_up(&ses->server->request_q); return rc; } - receive_len = be32_to_cpu(midQ->resp_buf->smb_buf_length); - - if (receive_len > CIFSMaxBufSize + MAX_CIFS_HDR_SIZE) { - cERROR(1, "Frame too large received. Length: %d Xid: %d", - receive_len, xid); - rc = -EIO; - goto out; - } - - /* rcvd frame is ok */ - - if (midQ->resp_buf && out_buf - && (midQ->midState == MID_RESPONSE_RECEIVED)) { - out_buf->smb_buf_length = cpu_to_be32(receive_len); - memcpy((char *)out_buf + 4, - (char *)midQ->resp_buf + 4, - receive_len); - - dump_smb(out_buf, 92); - /* convert the length into a more usable form */ - if ((receive_len > 24) && - (ses->server->secMode & (SECMODE_SIGN_REQUIRED | - SECMODE_SIGN_ENABLED))) { - rc = cifs_verify_signature(out_buf, - ses->server, - midQ->sequence_number+1); - if (rc) { - cERROR(1, "Unexpected SMB signature"); - /* BB FIXME add code to kill session */ - } - } - - *pbytes_returned = be32_to_cpu(out_buf->smb_buf_length); - - /* BB special case reconnect tid and uid here? */ - rc = map_smb_to_linux_error(out_buf, 0 /* no log */ ); - } else { + if (!midQ->resp_buf || !out_buf || + midQ->midState != MID_RESPONSE_RECEIVED) { rc = -EIO; cERROR(1, "Bad MID state?"); + goto out; } + *pbytes_returned = be32_to_cpu(midQ->resp_buf->smb_buf_length); + memcpy(out_buf, midQ->resp_buf, *pbytes_returned + 4); + rc = cifs_check_receive(midQ, ses->server, 0); out: delete_mid(midQ); atomic_dec(&ses->server->inFlight); @@ -810,12 +761,12 @@ out: blocking lock to return. */ static int -send_lock_cancel(const unsigned int xid, struct cifsTconInfo *tcon, +send_lock_cancel(const unsigned int xid, struct cifs_tcon *tcon, struct smb_hdr *in_buf, struct smb_hdr *out_buf) { int bytes_returned; - struct cifsSesInfo *ses = tcon->ses; + struct cifs_ses *ses = tcon->ses; LOCK_REQ *pSMB = (LOCK_REQ *)in_buf; /* We just modify the current in_buf to change @@ -832,15 +783,14 @@ send_lock_cancel(const unsigned int xid, struct cifsTconInfo *tcon, } int -SendReceiveBlockingLock(const unsigned int xid, struct cifsTconInfo *tcon, +SendReceiveBlockingLock(const unsigned int xid, struct cifs_tcon *tcon, struct smb_hdr *in_buf, struct smb_hdr *out_buf, int *pbytes_returned) { int rc = 0; int rstart = 0; - unsigned int receive_len; struct mid_q_entry *midQ; - struct cifsSesInfo *ses; + struct cifs_ses *ses; if (tcon == NULL || tcon->ses == NULL) { cERROR(1, "Null smb session"); @@ -957,50 +907,20 @@ SendReceiveBlockingLock(const unsigned int xid, struct cifsTconInfo *tcon, rstart = 1; } - rc = sync_mid_result(midQ, ses->server); + rc = cifs_sync_mid_result(midQ, ses->server); if (rc != 0) return rc; - receive_len = be32_to_cpu(midQ->resp_buf->smb_buf_length); - if (receive_len > CIFSMaxBufSize + MAX_CIFS_HDR_SIZE) { - cERROR(1, "Frame too large received. Length: %d Xid: %d", - receive_len, xid); - rc = -EIO; - goto out; - } - /* rcvd frame is ok */ - - if ((out_buf == NULL) || (midQ->midState != MID_RESPONSE_RECEIVED)) { + if (out_buf == NULL || midQ->midState != MID_RESPONSE_RECEIVED) { rc = -EIO; cERROR(1, "Bad MID state?"); goto out; } - out_buf->smb_buf_length = cpu_to_be32(receive_len); - memcpy((char *)out_buf + 4, - (char *)midQ->resp_buf + 4, - receive_len); - - dump_smb(out_buf, 92); - /* convert the length into a more usable form */ - if ((receive_len > 24) && - (ses->server->secMode & (SECMODE_SIGN_REQUIRED | - SECMODE_SIGN_ENABLED))) { - rc = cifs_verify_signature(out_buf, - ses->server, - midQ->sequence_number+1); - if (rc) { - cERROR(1, "Unexpected SMB signature"); - /* BB FIXME add code to kill session */ - } - } - - *pbytes_returned = be32_to_cpu(out_buf->smb_buf_length); - - /* BB special case reconnect tid and uid here? */ - rc = map_smb_to_linux_error(out_buf, 0 /* no log */ ); - + *pbytes_returned = be32_to_cpu(midQ->resp_buf->smb_buf_length); + memcpy(out_buf, midQ->resp_buf, *pbytes_returned + 4); + rc = cifs_check_receive(midQ, ses->server, 0); out: delete_mid(midQ); if (rstart && rc == -EACCES) diff --git a/fs/cifs/xattr.c b/fs/cifs/xattr.c index 912995e013ec..2a22fb2989e4 100644 --- a/fs/cifs/xattr.c +++ b/fs/cifs/xattr.c @@ -49,7 +49,7 @@ int cifs_removexattr(struct dentry *direntry, const char *ea_name) int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct super_block *sb; char *full_path = NULL; @@ -109,7 +109,7 @@ int cifs_setxattr(struct dentry *direntry, const char *ea_name, int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct super_block *sb; char *full_path; struct cifs_ntsd *pacl; @@ -240,7 +240,7 @@ ssize_t cifs_getxattr(struct dentry *direntry, const char *ea_name, int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct super_block *sb; char *full_path; @@ -372,7 +372,7 @@ ssize_t cifs_listxattr(struct dentry *direntry, char *data, size_t buf_size) int xid; struct cifs_sb_info *cifs_sb; struct tcon_link *tlink; - struct cifsTconInfo *pTcon; + struct cifs_tcon *pTcon; struct super_block *sb; char *full_path; diff --git a/fs/coda/dir.c b/fs/coda/dir.c index a46126fd5735..2b8dae4d121e 100644 --- a/fs/coda/dir.c +++ b/fs/coda/dir.c @@ -336,8 +336,6 @@ static int coda_rmdir(struct inode *dir, struct dentry *de) int len = de->d_name.len; int error; - dentry_unhash(de); - error = venus_rmdir(dir->i_sb, coda_i2f(dir), name, len); if (!error) { /* VFS may delete the child */ @@ -361,9 +359,6 @@ static int coda_rename(struct inode *old_dir, struct dentry *old_dentry, int new_length = new_dentry->d_name.len; int error; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - error = venus_rename(old_dir->i_sb, coda_i2f(old_dir), coda_i2f(new_dir), old_length, new_length, (const char *) old_name, (const char *)new_name); diff --git a/fs/coda/pioctl.c b/fs/coda/pioctl.c index 6cbb3afb36dc..cb140ef293e4 100644 --- a/fs/coda/pioctl.c +++ b/fs/coda/pioctl.c @@ -43,8 +43,6 @@ const struct file_operations coda_ioctl_operations = { /* the coda pioctl inode ops */ static int coda_ioctl_permission(struct inode *inode, int mask, unsigned int flags) { - if (flags & IPERM_FLAG_RCU) - return -ECHILD; return (mask & MAY_EXEC) ? -EACCES : 0; } diff --git a/fs/configfs/dir.c b/fs/configfs/dir.c index 9d17d350abc5..9a37a9b6de3a 100644 --- a/fs/configfs/dir.c +++ b/fs/configfs/dir.c @@ -1359,8 +1359,6 @@ static int configfs_rmdir(struct inode *dir, struct dentry *dentry) struct module *subsys_owner = NULL, *dead_item_owner = NULL; int ret; - dentry_unhash(dentry); - if (dentry->d_parent == configfs_sb->s_root) return -EPERM; diff --git a/fs/dcookies.c b/fs/dcookies.c index a21cabdbd87b..dda0dc702d1b 100644 --- a/fs/dcookies.c +++ b/fs/dcookies.c @@ -178,6 +178,8 @@ SYSCALL_DEFINE(lookup_dcookie)(u64 cookie64, char __user * buf, size_t len) /* FIXME: (deleted) ? */ path = d_path(&dcs->path, kbuf, PAGE_SIZE); + mutex_unlock(&dcookie_mutex); + if (IS_ERR(path)) { err = PTR_ERR(path); goto out_free; @@ -194,6 +196,7 @@ SYSCALL_DEFINE(lookup_dcookie)(u64 cookie64, char __user * buf, size_t len) out_free: kfree(kbuf); + return err; out: mutex_unlock(&dcookie_mutex); return err; diff --git a/fs/ecryptfs/crypto.c b/fs/ecryptfs/crypto.c index b8d5c8091024..58609bde3b9f 100644 --- a/fs/ecryptfs/crypto.c +++ b/fs/ecryptfs/crypto.c @@ -1024,25 +1024,25 @@ out: } /** - * contains_ecryptfs_marker - check for the ecryptfs marker + * ecryptfs_validate_marker - check for the ecryptfs marker * @data: The data block in which to check * - * Returns one if marker found; zero if not found + * Returns zero if marker found; -EINVAL if not found */ -static int contains_ecryptfs_marker(char *data) +static int ecryptfs_validate_marker(char *data) { u32 m_1, m_2; m_1 = get_unaligned_be32(data); m_2 = get_unaligned_be32(data + 4); if ((m_1 ^ MAGIC_ECRYPTFS_MARKER) == m_2) - return 1; + return 0; ecryptfs_printk(KERN_DEBUG, "m_1 = [0x%.8x]; m_2 = [0x%.8x]; " "MAGIC_ECRYPTFS_MARKER = [0x%.8x]\n", m_1, m_2, MAGIC_ECRYPTFS_MARKER); ecryptfs_printk(KERN_DEBUG, "(m_1 ^ MAGIC_ECRYPTFS_MARKER) = " "[0x%.8x]\n", (m_1 ^ MAGIC_ECRYPTFS_MARKER)); - return 0; + return -EINVAL; } struct ecryptfs_flag_map_elem { @@ -1201,27 +1201,19 @@ int ecryptfs_cipher_code_to_string(char *str, u8 cipher_code) return rc; } -int ecryptfs_read_and_validate_header_region(char *data, - struct inode *ecryptfs_inode) +int ecryptfs_read_and_validate_header_region(struct inode *inode) { - struct ecryptfs_crypt_stat *crypt_stat = - &(ecryptfs_inode_to_private(ecryptfs_inode)->crypt_stat); + u8 file_size[ECRYPTFS_SIZE_AND_MARKER_BYTES]; + u8 *marker = file_size + ECRYPTFS_FILE_SIZE_BYTES; int rc; - if (crypt_stat->extent_size == 0) - crypt_stat->extent_size = ECRYPTFS_DEFAULT_EXTENT_SIZE; - rc = ecryptfs_read_lower(data, 0, crypt_stat->extent_size, - ecryptfs_inode); - if (rc < 0) { - printk(KERN_ERR "%s: Error reading header region; rc = [%d]\n", - __func__, rc); - goto out; - } - if (!contains_ecryptfs_marker(data + ECRYPTFS_FILE_SIZE_BYTES)) { - rc = -EINVAL; - } else - rc = 0; -out: + rc = ecryptfs_read_lower(file_size, 0, ECRYPTFS_SIZE_AND_MARKER_BYTES, + inode); + if (rc < ECRYPTFS_SIZE_AND_MARKER_BYTES) + return rc >= 0 ? -EINVAL : rc; + rc = ecryptfs_validate_marker(marker); + if (!rc) + ecryptfs_i_size_init(file_size, inode); return rc; } @@ -1242,8 +1234,7 @@ ecryptfs_write_header_metadata(char *virt, (*written) = 6; } -struct kmem_cache *ecryptfs_header_cache_1; -struct kmem_cache *ecryptfs_header_cache_2; +struct kmem_cache *ecryptfs_header_cache; /** * ecryptfs_write_headers_virt @@ -1496,11 +1487,9 @@ static int ecryptfs_read_headers_virt(char *page_virt, crypt_stat->mount_crypt_stat = &ecryptfs_superblock_to_private( ecryptfs_dentry->d_sb)->mount_crypt_stat; offset = ECRYPTFS_FILE_SIZE_BYTES; - rc = contains_ecryptfs_marker(page_virt + offset); - if (rc == 0) { - rc = -EINVAL; + rc = ecryptfs_validate_marker(page_virt + offset); + if (rc) goto out; - } if (!(crypt_stat->flags & ECRYPTFS_I_SIZE_INITIALIZED)) ecryptfs_i_size_init(page_virt, ecryptfs_dentry->d_inode); offset += MAGIC_ECRYPTFS_MARKER_SIZE_BYTES; @@ -1567,20 +1556,21 @@ out: return rc; } -int ecryptfs_read_and_validate_xattr_region(char *page_virt, - struct dentry *ecryptfs_dentry) +int ecryptfs_read_and_validate_xattr_region(struct dentry *dentry, + struct inode *inode) { + u8 file_size[ECRYPTFS_SIZE_AND_MARKER_BYTES]; + u8 *marker = file_size + ECRYPTFS_FILE_SIZE_BYTES; int rc; - rc = ecryptfs_read_xattr_region(page_virt, ecryptfs_dentry->d_inode); - if (rc) - goto out; - if (!contains_ecryptfs_marker(page_virt + ECRYPTFS_FILE_SIZE_BYTES)) { - printk(KERN_WARNING "Valid data found in [%s] xattr, but " - "the marker is invalid\n", ECRYPTFS_XATTR_NAME); - rc = -EINVAL; - } -out: + rc = ecryptfs_getxattr_lower(ecryptfs_dentry_to_lower(dentry), + ECRYPTFS_XATTR_NAME, file_size, + ECRYPTFS_SIZE_AND_MARKER_BYTES); + if (rc < ECRYPTFS_SIZE_AND_MARKER_BYTES) + return rc >= 0 ? -EINVAL : rc; + rc = ecryptfs_validate_marker(marker); + if (!rc) + ecryptfs_i_size_init(file_size, inode); return rc; } @@ -1610,7 +1600,7 @@ int ecryptfs_read_metadata(struct dentry *ecryptfs_dentry) ecryptfs_copy_mount_wide_flags_to_inode_flags(crypt_stat, mount_crypt_stat); /* Read the first page from the underlying file */ - page_virt = kmem_cache_alloc(ecryptfs_header_cache_1, GFP_USER); + page_virt = kmem_cache_alloc(ecryptfs_header_cache, GFP_USER); if (!page_virt) { rc = -ENOMEM; printk(KERN_ERR "%s: Unable to allocate page_virt\n", @@ -1655,7 +1645,7 @@ int ecryptfs_read_metadata(struct dentry *ecryptfs_dentry) out: if (page_virt) { memset(page_virt, 0, PAGE_CACHE_SIZE); - kmem_cache_free(ecryptfs_header_cache_1, page_virt); + kmem_cache_free(ecryptfs_header_cache, page_virt); } return rc; } diff --git a/fs/ecryptfs/ecryptfs_kernel.h b/fs/ecryptfs/ecryptfs_kernel.h index e70282775e2c..43c7c43b06f5 100644 --- a/fs/ecryptfs/ecryptfs_kernel.h +++ b/fs/ecryptfs/ecryptfs_kernel.h @@ -200,6 +200,8 @@ ecryptfs_get_key_payload_data(struct key *key) #define MAGIC_ECRYPTFS_MARKER 0x3c81b7f5 #define MAGIC_ECRYPTFS_MARKER_SIZE_BYTES 8 /* 4*2 */ #define ECRYPTFS_FILE_SIZE_BYTES (sizeof(u64)) +#define ECRYPTFS_SIZE_AND_MARKER_BYTES (ECRYPTFS_FILE_SIZE_BYTES \ + + MAGIC_ECRYPTFS_MARKER_SIZE_BYTES) #define ECRYPTFS_DEFAULT_CIPHER "aes" #define ECRYPTFS_DEFAULT_KEY_BYTES 16 #define ECRYPTFS_DEFAULT_HASH "md5" @@ -603,8 +605,7 @@ extern struct kmem_cache *ecryptfs_file_info_cache; extern struct kmem_cache *ecryptfs_dentry_info_cache; extern struct kmem_cache *ecryptfs_inode_info_cache; extern struct kmem_cache *ecryptfs_sb_info_cache; -extern struct kmem_cache *ecryptfs_header_cache_1; -extern struct kmem_cache *ecryptfs_header_cache_2; +extern struct kmem_cache *ecryptfs_header_cache; extern struct kmem_cache *ecryptfs_xattr_cache; extern struct kmem_cache *ecryptfs_key_record_cache; extern struct kmem_cache *ecryptfs_key_sig_cache; @@ -625,14 +626,9 @@ struct ecryptfs_open_req { struct list_head kthread_ctl_list; }; -#define ECRYPTFS_INTERPOSE_FLAG_D_ADD 0x00000001 -int ecryptfs_interpose(struct dentry *hidden_dentry, - struct dentry *this_dentry, struct super_block *sb, - u32 flags); +struct inode *ecryptfs_get_inode(struct inode *lower_inode, + struct super_block *sb); void ecryptfs_i_size_init(const char *page_virt, struct inode *inode); -int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry, - struct dentry *lower_dentry, - struct inode *ecryptfs_dir_inode); int ecryptfs_decode_and_decrypt_filename(char **decrypted_name, size_t *decrypted_name_size, struct dentry *ecryptfs_dentry, @@ -664,10 +660,9 @@ int ecryptfs_new_file_context(struct dentry *ecryptfs_dentry); void ecryptfs_write_crypt_stat_flags(char *page_virt, struct ecryptfs_crypt_stat *crypt_stat, size_t *written); -int ecryptfs_read_and_validate_header_region(char *data, - struct inode *ecryptfs_inode); -int ecryptfs_read_and_validate_xattr_region(char *page_virt, - struct dentry *ecryptfs_dentry); +int ecryptfs_read_and_validate_header_region(struct inode *inode); +int ecryptfs_read_and_validate_xattr_region(struct dentry *dentry, + struct inode *inode); u8 ecryptfs_code_for_cipher_string(char *cipher_name, size_t key_bytes); int ecryptfs_cipher_code_to_string(char *str, u8 cipher_code); void ecryptfs_set_default_sizes(struct ecryptfs_crypt_stat *crypt_stat); @@ -679,9 +674,6 @@ int ecryptfs_parse_packet_set(struct ecryptfs_crypt_stat *crypt_stat, unsigned char *src, struct dentry *ecryptfs_dentry); int ecryptfs_truncate(struct dentry *dentry, loff_t new_length); -int ecryptfs_inode_test(struct inode *inode, void *candidate_lower_inode); -int ecryptfs_inode_set(struct inode *inode, void *lower_inode); -void ecryptfs_init_inode(struct inode *inode, struct inode *lower_inode); ssize_t ecryptfs_getxattr_lower(struct dentry *lower_dentry, const char *name, void *value, size_t size); @@ -761,7 +753,7 @@ int ecryptfs_privileged_open(struct file **lower_file, struct dentry *lower_dentry, struct vfsmount *lower_mnt, const struct cred *cred); -int ecryptfs_get_lower_file(struct dentry *ecryptfs_dentry); +int ecryptfs_get_lower_file(struct dentry *dentry, struct inode *inode); void ecryptfs_put_lower_file(struct inode *inode); int ecryptfs_write_tag_70_packet(char *dest, size_t *remaining_bytes, diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c index 566e5472f78c..4ec9eb00a241 100644 --- a/fs/ecryptfs/file.c +++ b/fs/ecryptfs/file.c @@ -191,7 +191,7 @@ static int ecryptfs_open(struct inode *inode, struct file *file) | ECRYPTFS_ENCRYPTED); } mutex_unlock(&crypt_stat->cs_mutex); - rc = ecryptfs_get_lower_file(ecryptfs_dentry); + rc = ecryptfs_get_lower_file(ecryptfs_dentry, inode); if (rc) { printk(KERN_ERR "%s: Error attempting to initialize " "the lower file for the dentry with name " diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c index 227b409b8406..7349ade17de6 100644 --- a/fs/ecryptfs/inode.c +++ b/fs/ecryptfs/inode.c @@ -51,6 +51,97 @@ static void unlock_dir(struct dentry *dir) dput(dir); } +static int ecryptfs_inode_test(struct inode *inode, void *lower_inode) +{ + if (ecryptfs_inode_to_lower(inode) == (struct inode *)lower_inode) + return 1; + return 0; +} + +static int ecryptfs_inode_set(struct inode *inode, void *opaque) +{ + struct inode *lower_inode = opaque; + + ecryptfs_set_inode_lower(inode, lower_inode); + fsstack_copy_attr_all(inode, lower_inode); + /* i_size will be overwritten for encrypted regular files */ + fsstack_copy_inode_size(inode, lower_inode); + inode->i_ino = lower_inode->i_ino; + inode->i_version++; + inode->i_mapping->a_ops = &ecryptfs_aops; + + if (S_ISLNK(inode->i_mode)) + inode->i_op = &ecryptfs_symlink_iops; + else if (S_ISDIR(inode->i_mode)) + inode->i_op = &ecryptfs_dir_iops; + else + inode->i_op = &ecryptfs_main_iops; + + if (S_ISDIR(inode->i_mode)) + inode->i_fop = &ecryptfs_dir_fops; + else if (special_file(inode->i_mode)) + init_special_inode(inode, inode->i_mode, inode->i_rdev); + else + inode->i_fop = &ecryptfs_main_fops; + + return 0; +} + +static struct inode *__ecryptfs_get_inode(struct inode *lower_inode, + struct super_block *sb) +{ + struct inode *inode; + + if (lower_inode->i_sb != ecryptfs_superblock_to_lower(sb)) + return ERR_PTR(-EXDEV); + if (!igrab(lower_inode)) + return ERR_PTR(-ESTALE); + inode = iget5_locked(sb, (unsigned long)lower_inode, + ecryptfs_inode_test, ecryptfs_inode_set, + lower_inode); + if (!inode) { + iput(lower_inode); + return ERR_PTR(-EACCES); + } + if (!(inode->i_state & I_NEW)) + iput(lower_inode); + + return inode; +} + +struct inode *ecryptfs_get_inode(struct inode *lower_inode, + struct super_block *sb) +{ + struct inode *inode = __ecryptfs_get_inode(lower_inode, sb); + + if (!IS_ERR(inode) && (inode->i_state & I_NEW)) + unlock_new_inode(inode); + + return inode; +} + +/** + * ecryptfs_interpose + * @lower_dentry: Existing dentry in the lower filesystem + * @dentry: ecryptfs' dentry + * @sb: ecryptfs's super_block + * + * Interposes upper and lower dentries. + * + * Returns zero on success; non-zero otherwise + */ +static int ecryptfs_interpose(struct dentry *lower_dentry, + struct dentry *dentry, struct super_block *sb) +{ + struct inode *inode = ecryptfs_get_inode(lower_dentry->d_inode, sb); + + if (IS_ERR(inode)) + return PTR_ERR(inode); + d_instantiate(dentry, inode); + + return 0; +} + /** * ecryptfs_create_underlying_file * @lower_dir_inode: inode of the parent in the lower fs of the new file @@ -129,7 +220,7 @@ ecryptfs_do_create(struct inode *directory_inode, goto out_lock; } rc = ecryptfs_interpose(lower_dentry, ecryptfs_dentry, - directory_inode->i_sb, 0); + directory_inode->i_sb); if (rc) { ecryptfs_printk(KERN_ERR, "Failure in ecryptfs_interpose\n"); goto out_lock; @@ -168,7 +259,8 @@ static int ecryptfs_initialize_file(struct dentry *ecryptfs_dentry) "context; rc = [%d]\n", rc); goto out; } - rc = ecryptfs_get_lower_file(ecryptfs_dentry); + rc = ecryptfs_get_lower_file(ecryptfs_dentry, + ecryptfs_dentry->d_inode); if (rc) { printk(KERN_ERR "%s: Error attempting to initialize " "the lower file for the dentry with name " @@ -215,102 +307,90 @@ out: return rc; } +static int ecryptfs_i_size_read(struct dentry *dentry, struct inode *inode) +{ + struct ecryptfs_crypt_stat *crypt_stat; + int rc; + + rc = ecryptfs_get_lower_file(dentry, inode); + if (rc) { + printk(KERN_ERR "%s: Error attempting to initialize " + "the lower file for the dentry with name " + "[%s]; rc = [%d]\n", __func__, + dentry->d_name.name, rc); + return rc; + } + + crypt_stat = &ecryptfs_inode_to_private(inode)->crypt_stat; + /* TODO: lock for crypt_stat comparison */ + if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED)) + ecryptfs_set_default_sizes(crypt_stat); + + rc = ecryptfs_read_and_validate_header_region(inode); + ecryptfs_put_lower_file(inode); + if (rc) { + rc = ecryptfs_read_and_validate_xattr_region(dentry, inode); + if (!rc) + crypt_stat->flags |= ECRYPTFS_METADATA_IN_XATTR; + } + + /* Must return 0 to allow non-eCryptfs files to be looked up, too */ + return 0; +} + /** - * ecryptfs_lookup_and_interpose_lower - Perform a lookup + * ecryptfs_lookup_interpose - Dentry interposition for a lookup */ -int ecryptfs_lookup_and_interpose_lower(struct dentry *ecryptfs_dentry, - struct dentry *lower_dentry, - struct inode *ecryptfs_dir_inode) +static int ecryptfs_lookup_interpose(struct dentry *dentry, + struct dentry *lower_dentry, + struct inode *dir_inode) { - struct dentry *lower_dir_dentry; + struct inode *inode, *lower_inode = lower_dentry->d_inode; + struct ecryptfs_dentry_info *dentry_info; struct vfsmount *lower_mnt; - struct inode *lower_inode; - struct ecryptfs_crypt_stat *crypt_stat; - char *page_virt = NULL; - int put_lower = 0, rc = 0; - - lower_dir_dentry = lower_dentry->d_parent; - lower_mnt = mntget(ecryptfs_dentry_to_lower_mnt( - ecryptfs_dentry->d_parent)); - lower_inode = lower_dentry->d_inode; - fsstack_copy_attr_atime(ecryptfs_dir_inode, lower_dir_dentry->d_inode); + int rc = 0; + + lower_mnt = mntget(ecryptfs_dentry_to_lower_mnt(dentry->d_parent)); + fsstack_copy_attr_atime(dir_inode, lower_dentry->d_parent->d_inode); BUG_ON(!lower_dentry->d_count); - ecryptfs_set_dentry_private(ecryptfs_dentry, - kmem_cache_alloc(ecryptfs_dentry_info_cache, - GFP_KERNEL)); - if (!ecryptfs_dentry_to_private(ecryptfs_dentry)) { - rc = -ENOMEM; + + dentry_info = kmem_cache_alloc(ecryptfs_dentry_info_cache, GFP_KERNEL); + ecryptfs_set_dentry_private(dentry, dentry_info); + if (!dentry_info) { printk(KERN_ERR "%s: Out of memory whilst attempting " "to allocate ecryptfs_dentry_info struct\n", __func__); - goto out_put; + dput(lower_dentry); + mntput(lower_mnt); + d_drop(dentry); + return -ENOMEM; } - ecryptfs_set_dentry_lower(ecryptfs_dentry, lower_dentry); - ecryptfs_set_dentry_lower_mnt(ecryptfs_dentry, lower_mnt); + ecryptfs_set_dentry_lower(dentry, lower_dentry); + ecryptfs_set_dentry_lower_mnt(dentry, lower_mnt); + if (!lower_dentry->d_inode) { /* We want to add because we couldn't find in lower */ - d_add(ecryptfs_dentry, NULL); - goto out; - } - rc = ecryptfs_interpose(lower_dentry, ecryptfs_dentry, - ecryptfs_dir_inode->i_sb, - ECRYPTFS_INTERPOSE_FLAG_D_ADD); - if (rc) { - printk(KERN_ERR "%s: Error interposing; rc = [%d]\n", - __func__, rc); - goto out; - } - if (S_ISDIR(lower_inode->i_mode)) - goto out; - if (S_ISLNK(lower_inode->i_mode)) - goto out; - if (special_file(lower_inode->i_mode)) - goto out; - /* Released in this function */ - page_virt = kmem_cache_zalloc(ecryptfs_header_cache_2, GFP_USER); - if (!page_virt) { - printk(KERN_ERR "%s: Cannot kmem_cache_zalloc() a page\n", - __func__); - rc = -ENOMEM; - goto out; + d_add(dentry, NULL); + return 0; } - rc = ecryptfs_get_lower_file(ecryptfs_dentry); - if (rc) { - printk(KERN_ERR "%s: Error attempting to initialize " - "the lower file for the dentry with name " - "[%s]; rc = [%d]\n", __func__, - ecryptfs_dentry->d_name.name, rc); - goto out_free_kmem; + inode = __ecryptfs_get_inode(lower_inode, dir_inode->i_sb); + if (IS_ERR(inode)) { + printk(KERN_ERR "%s: Error interposing; rc = [%ld]\n", + __func__, PTR_ERR(inode)); + return PTR_ERR(inode); } - put_lower = 1; - crypt_stat = &ecryptfs_inode_to_private( - ecryptfs_dentry->d_inode)->crypt_stat; - /* TODO: lock for crypt_stat comparison */ - if (!(crypt_stat->flags & ECRYPTFS_POLICY_APPLIED)) - ecryptfs_set_default_sizes(crypt_stat); - rc = ecryptfs_read_and_validate_header_region(page_virt, - ecryptfs_dentry->d_inode); - if (rc) { - memset(page_virt, 0, PAGE_CACHE_SIZE); - rc = ecryptfs_read_and_validate_xattr_region(page_virt, - ecryptfs_dentry); + if (S_ISREG(inode->i_mode)) { + rc = ecryptfs_i_size_read(dentry, inode); if (rc) { - rc = 0; - goto out_free_kmem; + make_bad_inode(inode); + return rc; } - crypt_stat->flags |= ECRYPTFS_METADATA_IN_XATTR; } - ecryptfs_i_size_init(page_virt, ecryptfs_dentry->d_inode); -out_free_kmem: - kmem_cache_free(ecryptfs_header_cache_2, page_virt); - goto out; -out_put: - dput(lower_dentry); - mntput(lower_mnt); - d_drop(ecryptfs_dentry); -out: - if (put_lower) - ecryptfs_put_lower_file(ecryptfs_dentry->d_inode); + + if (inode->i_state & I_NEW) + unlock_new_inode(inode); + d_add(dentry, inode); + return rc; } @@ -353,12 +433,12 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode, goto out_d_drop; } if (lower_dentry->d_inode) - goto lookup_and_interpose; + goto interpose; mount_crypt_stat = &ecryptfs_superblock_to_private( ecryptfs_dentry->d_sb)->mount_crypt_stat; if (!(mount_crypt_stat && (mount_crypt_stat->flags & ECRYPTFS_GLOBAL_ENCRYPT_FILENAMES))) - goto lookup_and_interpose; + goto interpose; dput(lower_dentry); rc = ecryptfs_encrypt_and_encode_filename( &encrypted_and_encoded_name, &encrypted_and_encoded_name_size, @@ -381,9 +461,9 @@ static struct dentry *ecryptfs_lookup(struct inode *ecryptfs_dir_inode, encrypted_and_encoded_name); goto out_d_drop; } -lookup_and_interpose: - rc = ecryptfs_lookup_and_interpose_lower(ecryptfs_dentry, lower_dentry, - ecryptfs_dir_inode); +interpose: + rc = ecryptfs_lookup_interpose(ecryptfs_dentry, lower_dentry, + ecryptfs_dir_inode); goto out; out_d_drop: d_drop(ecryptfs_dentry); @@ -411,7 +491,7 @@ static int ecryptfs_link(struct dentry *old_dentry, struct inode *dir, lower_new_dentry); if (rc || !lower_new_dentry->d_inode) goto out_lock; - rc = ecryptfs_interpose(lower_new_dentry, new_dentry, dir->i_sb, 0); + rc = ecryptfs_interpose(lower_new_dentry, new_dentry, dir->i_sb); if (rc) goto out_lock; fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); @@ -478,7 +558,7 @@ static int ecryptfs_symlink(struct inode *dir, struct dentry *dentry, kfree(encoded_symname); if (rc || !lower_dentry->d_inode) goto out_lock; - rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); + rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb); if (rc) goto out_lock; fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); @@ -502,7 +582,7 @@ static int ecryptfs_mkdir(struct inode *dir, struct dentry *dentry, int mode) rc = vfs_mkdir(lower_dir_dentry->d_inode, lower_dentry, mode); if (rc || !lower_dentry->d_inode) goto out; - rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); + rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb); if (rc) goto out; fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); @@ -521,14 +601,14 @@ static int ecryptfs_rmdir(struct inode *dir, struct dentry *dentry) struct dentry *lower_dir_dentry; int rc; - dentry_unhash(dentry); - lower_dentry = ecryptfs_dentry_to_lower(dentry); dget(dentry); lower_dir_dentry = lock_parent(lower_dentry); dget(lower_dentry); rc = vfs_rmdir(lower_dir_dentry->d_inode, lower_dentry); dput(lower_dentry); + if (!rc && dentry->d_inode) + clear_nlink(dentry->d_inode); fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); dir->i_nlink = lower_dir_dentry->d_inode->i_nlink; unlock_dir(lower_dir_dentry); @@ -550,7 +630,7 @@ ecryptfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) rc = vfs_mknod(lower_dir_dentry->d_inode, lower_dentry, mode, dev); if (rc || !lower_dentry->d_inode) goto out; - rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb, 0); + rc = ecryptfs_interpose(lower_dentry, dentry, dir->i_sb); if (rc) goto out; fsstack_copy_attr_times(dir, lower_dir_dentry->d_inode); @@ -573,9 +653,6 @@ ecryptfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct dentry *lower_new_dir_dentry; struct dentry *trap = NULL; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - lower_old_dentry = ecryptfs_dentry_to_lower(old_dentry); lower_new_dentry = ecryptfs_dentry_to_lower(new_dentry); dget(lower_old_dentry); @@ -753,7 +830,7 @@ static int truncate_upper(struct dentry *dentry, struct iattr *ia, lower_ia->ia_valid &= ~ATTR_SIZE; return 0; } - rc = ecryptfs_get_lower_file(dentry); + rc = ecryptfs_get_lower_file(dentry, inode); if (rc) return rc; crypt_stat = &ecryptfs_inode_to_private(dentry->d_inode)->crypt_stat; @@ -909,7 +986,7 @@ static int ecryptfs_setattr(struct dentry *dentry, struct iattr *ia) mount_crypt_stat = &ecryptfs_superblock_to_private( dentry->d_sb)->mount_crypt_stat; - rc = ecryptfs_get_lower_file(dentry); + rc = ecryptfs_get_lower_file(dentry, inode); if (rc) { mutex_unlock(&crypt_stat->cs_mutex); goto out; @@ -1082,21 +1159,6 @@ out: return rc; } -int ecryptfs_inode_test(struct inode *inode, void *candidate_lower_inode) -{ - if ((ecryptfs_inode_to_lower(inode) - == (struct inode *)candidate_lower_inode)) - return 1; - else - return 0; -} - -int ecryptfs_inode_set(struct inode *inode, void *lower_inode) -{ - ecryptfs_init_inode(inode, (struct inode *)lower_inode); - return 0; -} - const struct inode_operations ecryptfs_symlink_iops = { .readlink = ecryptfs_readlink, .follow_link = ecryptfs_follow_link, diff --git a/fs/ecryptfs/keystore.c b/fs/ecryptfs/keystore.c index 03e609c45012..27a7fefb83eb 100644 --- a/fs/ecryptfs/keystore.c +++ b/fs/ecryptfs/keystore.c @@ -599,8 +599,8 @@ struct ecryptfs_write_tag_70_packet_silly_stack { struct mutex *tfm_mutex; char *block_aligned_filename; struct ecryptfs_auth_tok *auth_tok; - struct scatterlist src_sg; - struct scatterlist dst_sg; + struct scatterlist src_sg[2]; + struct scatterlist dst_sg[2]; struct blkcipher_desc desc; char iv[ECRYPTFS_MAX_IV_BYTES]; char hash[ECRYPTFS_TAG_70_DIGEST_SIZE]; @@ -816,23 +816,21 @@ ecryptfs_write_tag_70_packet(char *dest, size_t *remaining_bytes, memcpy(&s->block_aligned_filename[s->num_rand_bytes], filename, filename_size); rc = virt_to_scatterlist(s->block_aligned_filename, - s->block_aligned_filename_size, &s->src_sg, 1); - if (rc != 1) { + s->block_aligned_filename_size, s->src_sg, 2); + if (rc < 1) { printk(KERN_ERR "%s: Internal error whilst attempting to " - "convert filename memory to scatterlist; " - "expected rc = 1; got rc = [%d]. " + "convert filename memory to scatterlist; rc = [%d]. " "block_aligned_filename_size = [%zd]\n", __func__, rc, s->block_aligned_filename_size); goto out_release_free_unlock; } rc = virt_to_scatterlist(&dest[s->i], s->block_aligned_filename_size, - &s->dst_sg, 1); - if (rc != 1) { + s->dst_sg, 2); + if (rc < 1) { printk(KERN_ERR "%s: Internal error whilst attempting to " "convert encrypted filename memory to scatterlist; " - "expected rc = 1; got rc = [%d]. " - "block_aligned_filename_size = [%zd]\n", __func__, rc, - s->block_aligned_filename_size); + "rc = [%d]. block_aligned_filename_size = [%zd]\n", + __func__, rc, s->block_aligned_filename_size); goto out_release_free_unlock; } /* The characters in the first block effectively do the job @@ -855,7 +853,7 @@ ecryptfs_write_tag_70_packet(char *dest, size_t *remaining_bytes, mount_crypt_stat->global_default_fn_cipher_key_bytes); goto out_release_free_unlock; } - rc = crypto_blkcipher_encrypt_iv(&s->desc, &s->dst_sg, &s->src_sg, + rc = crypto_blkcipher_encrypt_iv(&s->desc, s->dst_sg, s->src_sg, s->block_aligned_filename_size); if (rc) { printk(KERN_ERR "%s: Error attempting to encrypt filename; " @@ -891,8 +889,8 @@ struct ecryptfs_parse_tag_70_packet_silly_stack { struct mutex *tfm_mutex; char *decrypted_filename; struct ecryptfs_auth_tok *auth_tok; - struct scatterlist src_sg; - struct scatterlist dst_sg; + struct scatterlist src_sg[2]; + struct scatterlist dst_sg[2]; struct blkcipher_desc desc; char fnek_sig_hex[ECRYPTFS_SIG_SIZE_HEX + 1]; char iv[ECRYPTFS_MAX_IV_BYTES]; @@ -1008,13 +1006,12 @@ ecryptfs_parse_tag_70_packet(char **filename, size_t *filename_size, } mutex_lock(s->tfm_mutex); rc = virt_to_scatterlist(&data[(*packet_size)], - s->block_aligned_filename_size, &s->src_sg, 1); - if (rc != 1) { + s->block_aligned_filename_size, s->src_sg, 2); + if (rc < 1) { printk(KERN_ERR "%s: Internal error whilst attempting to " "convert encrypted filename memory to scatterlist; " - "expected rc = 1; got rc = [%d]. " - "block_aligned_filename_size = [%zd]\n", __func__, rc, - s->block_aligned_filename_size); + "rc = [%d]. block_aligned_filename_size = [%zd]\n", + __func__, rc, s->block_aligned_filename_size); goto out_unlock; } (*packet_size) += s->block_aligned_filename_size; @@ -1028,13 +1025,12 @@ ecryptfs_parse_tag_70_packet(char **filename, size_t *filename_size, goto out_unlock; } rc = virt_to_scatterlist(s->decrypted_filename, - s->block_aligned_filename_size, &s->dst_sg, 1); - if (rc != 1) { + s->block_aligned_filename_size, s->dst_sg, 2); + if (rc < 1) { printk(KERN_ERR "%s: Internal error whilst attempting to " "convert decrypted filename memory to scatterlist; " - "expected rc = 1; got rc = [%d]. " - "block_aligned_filename_size = [%zd]\n", __func__, rc, - s->block_aligned_filename_size); + "rc = [%d]. block_aligned_filename_size = [%zd]\n", + __func__, rc, s->block_aligned_filename_size); goto out_free_unlock; } /* The characters in the first block effectively do the job of @@ -1065,7 +1061,7 @@ ecryptfs_parse_tag_70_packet(char **filename, size_t *filename_size, mount_crypt_stat->global_default_fn_cipher_key_bytes); goto out_free_unlock; } - rc = crypto_blkcipher_decrypt_iv(&s->desc, &s->dst_sg, &s->src_sg, + rc = crypto_blkcipher_decrypt_iv(&s->desc, s->dst_sg, s->src_sg, s->block_aligned_filename_size); if (rc) { printk(KERN_ERR "%s: Error attempting to decrypt filename; " diff --git a/fs/ecryptfs/main.c b/fs/ecryptfs/main.c index 89b93389af8e..9f1bb747d77d 100644 --- a/fs/ecryptfs/main.c +++ b/fs/ecryptfs/main.c @@ -135,12 +135,12 @@ static int ecryptfs_init_lower_file(struct dentry *dentry, return rc; } -int ecryptfs_get_lower_file(struct dentry *dentry) +int ecryptfs_get_lower_file(struct dentry *dentry, struct inode *inode) { - struct ecryptfs_inode_info *inode_info = - ecryptfs_inode_to_private(dentry->d_inode); + struct ecryptfs_inode_info *inode_info; int count, rc = 0; + inode_info = ecryptfs_inode_to_private(inode); mutex_lock(&inode_info->lower_file_mutex); count = atomic_inc_return(&inode_info->lower_file_count); if (WARN_ON_ONCE(count < 1)) @@ -168,75 +168,6 @@ void ecryptfs_put_lower_file(struct inode *inode) } } -static struct inode *ecryptfs_get_inode(struct inode *lower_inode, - struct super_block *sb) -{ - struct inode *inode; - int rc = 0; - - if (lower_inode->i_sb != ecryptfs_superblock_to_lower(sb)) { - rc = -EXDEV; - goto out; - } - if (!igrab(lower_inode)) { - rc = -ESTALE; - goto out; - } - inode = iget5_locked(sb, (unsigned long)lower_inode, - ecryptfs_inode_test, ecryptfs_inode_set, - lower_inode); - if (!inode) { - rc = -EACCES; - iput(lower_inode); - goto out; - } - if (inode->i_state & I_NEW) - unlock_new_inode(inode); - else - iput(lower_inode); - if (S_ISLNK(lower_inode->i_mode)) - inode->i_op = &ecryptfs_symlink_iops; - else if (S_ISDIR(lower_inode->i_mode)) - inode->i_op = &ecryptfs_dir_iops; - if (S_ISDIR(lower_inode->i_mode)) - inode->i_fop = &ecryptfs_dir_fops; - if (special_file(lower_inode->i_mode)) - init_special_inode(inode, lower_inode->i_mode, - lower_inode->i_rdev); - fsstack_copy_attr_all(inode, lower_inode); - /* This size will be overwritten for real files w/ headers and - * other metadata */ - fsstack_copy_inode_size(inode, lower_inode); - return inode; -out: - return ERR_PTR(rc); -} - -/** - * ecryptfs_interpose - * @lower_dentry: Existing dentry in the lower filesystem - * @dentry: ecryptfs' dentry - * @sb: ecryptfs's super_block - * @flags: flags to govern behavior of interpose procedure - * - * Interposes upper and lower dentries. - * - * Returns zero on success; non-zero otherwise - */ -int ecryptfs_interpose(struct dentry *lower_dentry, struct dentry *dentry, - struct super_block *sb, u32 flags) -{ - struct inode *lower_inode = lower_dentry->d_inode; - struct inode *inode = ecryptfs_get_inode(lower_inode, sb); - if (IS_ERR(inode)) - return PTR_ERR(inode); - if (flags & ECRYPTFS_INTERPOSE_FLAG_D_ADD) - d_add(dentry, inode); - else - d_instantiate(dentry, inode); - return 0; -} - enum { ecryptfs_opt_sig, ecryptfs_opt_ecryptfs_sig, ecryptfs_opt_cipher, ecryptfs_opt_ecryptfs_cipher, ecryptfs_opt_ecryptfs_key_bytes, @@ -704,13 +635,8 @@ static struct ecryptfs_cache_info { .size = sizeof(struct ecryptfs_sb_info), }, { - .cache = &ecryptfs_header_cache_1, - .name = "ecryptfs_headers_1", - .size = PAGE_CACHE_SIZE, - }, - { - .cache = &ecryptfs_header_cache_2, - .name = "ecryptfs_headers_2", + .cache = &ecryptfs_header_cache, + .name = "ecryptfs_headers", .size = PAGE_CACHE_SIZE, }, { diff --git a/fs/ecryptfs/super.c b/fs/ecryptfs/super.c index 245b517bf1b6..dbd52d40df4c 100644 --- a/fs/ecryptfs/super.c +++ b/fs/ecryptfs/super.c @@ -93,22 +93,6 @@ static void ecryptfs_destroy_inode(struct inode *inode) } /** - * ecryptfs_init_inode - * @inode: The ecryptfs inode - * - * Set up the ecryptfs inode. - */ -void ecryptfs_init_inode(struct inode *inode, struct inode *lower_inode) -{ - ecryptfs_set_inode_lower(inode, lower_inode); - inode->i_ino = lower_inode->i_ino; - inode->i_version++; - inode->i_op = &ecryptfs_main_iops; - inode->i_fop = &ecryptfs_main_fops; - inode->i_mapping->a_ops = &ecryptfs_aops; -} - -/** * ecryptfs_statfs * @sb: The ecryptfs super block * @buf: The struct kstatfs to fill in with stats diff --git a/fs/exec.c b/fs/exec.c index ea5f748906a8..6075a1e727ae 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1093,6 +1093,7 @@ int flush_old_exec(struct linux_binprm * bprm) bprm->mm = NULL; /* We're using it now */ + set_fs(USER_DS); current->flags &= ~(PF_RANDOMIZE | PF_KTHREAD); flush_thread(); current->personality &= ~bprm->per_clear; @@ -1357,10 +1358,6 @@ int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs) if (retval) return retval; - /* kernel module loader fixup */ - /* so we don't try to load run modprobe in kernel space. */ - set_fs(USER_DS); - retval = audit_bprm(bprm); if (retval) return retval; @@ -1999,7 +1996,7 @@ static void wait_for_dump_helpers(struct file *file) * is a special value that we use to trap recursive * core dumps */ -static int umh_pipe_setup(struct subprocess_info *info) +static int umh_pipe_setup(struct subprocess_info *info, struct cred *new) { struct file *rp, *wp; struct fdtable *fdt; diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index 68b2e43d7c35..3451d23c3bae 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -3392,7 +3392,7 @@ int ext3_mark_inode_dirty(handle_t *handle, struct inode *inode) * so would cause a commit on atime updates, which we don't bother doing. * We handle synchronous inodes at the highest possible level. */ -void ext3_dirty_inode(struct inode *inode) +void ext3_dirty_inode(struct inode *inode, int flags) { handle_t *current_handle = ext3_journal_current_handle(); handle_t *handle; diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index a74b89c09f90..1921392cd708 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1813,7 +1813,7 @@ extern int ext4_getattr(struct vfsmount *mnt, struct dentry *dentry, extern void ext4_evict_inode(struct inode *); extern void ext4_clear_inode(struct inode *); extern int ext4_sync_inode(handle_t *, struct inode *); -extern void ext4_dirty_inode(struct inode *); +extern void ext4_dirty_inode(struct inode *, int); extern int ext4_change_inode_journal_flag(struct inode *, int); extern int ext4_get_inode_loc(struct inode *, struct ext4_iloc *); extern int ext4_can_truncate(struct inode *inode); diff --git a/fs/ext4/ext4_extents.h b/fs/ext4/ext4_extents.h index 2e29abb30f76..095c36f3b612 100644 --- a/fs/ext4/ext4_extents.h +++ b/fs/ext4/ext4_extents.h @@ -125,7 +125,7 @@ struct ext4_ext_path { * positive retcode - signal for ext4_ext_walk_space(), see below * callback must return valid extent (passed or newly created) */ -typedef int (*ext_prepare_callback)(struct inode *, struct ext4_ext_path *, +typedef int (*ext_prepare_callback)(struct inode *, ext4_lblk_t, struct ext4_ext_cache *, struct ext4_extent *, void *); @@ -133,8 +133,11 @@ typedef int (*ext_prepare_callback)(struct inode *, struct ext4_ext_path *, #define EXT_BREAK 1 #define EXT_REPEAT 2 -/* Maximum logical block in a file; ext4_extent's ee_block is __le32 */ -#define EXT_MAX_BLOCK 0xffffffff +/* + * Maximum number of logical blocks in a file; ext4_extent's ee_block is + * __le32. + */ +#define EXT_MAX_BLOCKS 0xffffffff /* * EXT_INIT_MAX_LEN is the maximum number of blocks we can have in an diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 5199bac7fc62..f815cc81e7a2 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -1408,7 +1408,7 @@ got_index: /* * ext4_ext_next_allocated_block: - * returns allocated block in subsequent extent or EXT_MAX_BLOCK. + * returns allocated block in subsequent extent or EXT_MAX_BLOCKS. * NOTE: it considers block number from index entry as * allocated block. Thus, index entries have to be consistent * with leaves. @@ -1422,7 +1422,7 @@ ext4_ext_next_allocated_block(struct ext4_ext_path *path) depth = path->p_depth; if (depth == 0 && path->p_ext == NULL) - return EXT_MAX_BLOCK; + return EXT_MAX_BLOCKS; while (depth >= 0) { if (depth == path->p_depth) { @@ -1439,12 +1439,12 @@ ext4_ext_next_allocated_block(struct ext4_ext_path *path) depth--; } - return EXT_MAX_BLOCK; + return EXT_MAX_BLOCKS; } /* * ext4_ext_next_leaf_block: - * returns first allocated block from next leaf or EXT_MAX_BLOCK + * returns first allocated block from next leaf or EXT_MAX_BLOCKS */ static ext4_lblk_t ext4_ext_next_leaf_block(struct inode *inode, struct ext4_ext_path *path) @@ -1456,7 +1456,7 @@ static ext4_lblk_t ext4_ext_next_leaf_block(struct inode *inode, /* zero-tree has no leaf blocks at all */ if (depth == 0) - return EXT_MAX_BLOCK; + return EXT_MAX_BLOCKS; /* go to index block */ depth--; @@ -1469,7 +1469,7 @@ static ext4_lblk_t ext4_ext_next_leaf_block(struct inode *inode, depth--; } - return EXT_MAX_BLOCK; + return EXT_MAX_BLOCKS; } /* @@ -1677,13 +1677,13 @@ static unsigned int ext4_ext_check_overlap(struct inode *inode, */ if (b2 < b1) { b2 = ext4_ext_next_allocated_block(path); - if (b2 == EXT_MAX_BLOCK) + if (b2 == EXT_MAX_BLOCKS) goto out; } /* check for wrap through zero on extent logical start block*/ if (b1 + len1 < b1) { - len1 = EXT_MAX_BLOCK - b1; + len1 = EXT_MAX_BLOCKS - b1; newext->ee_len = cpu_to_le16(len1); ret = 1; } @@ -1767,7 +1767,7 @@ repeat: fex = EXT_LAST_EXTENT(eh); next = ext4_ext_next_leaf_block(inode, path); if (le32_to_cpu(newext->ee_block) > le32_to_cpu(fex->ee_block) - && next != EXT_MAX_BLOCK) { + && next != EXT_MAX_BLOCKS) { ext_debug("next leaf block - %d\n", next); BUG_ON(npath != NULL); npath = ext4_ext_find_extent(inode, next, NULL); @@ -1887,7 +1887,7 @@ static int ext4_ext_walk_space(struct inode *inode, ext4_lblk_t block, BUG_ON(func == NULL); BUG_ON(inode == NULL); - while (block < last && block != EXT_MAX_BLOCK) { + while (block < last && block != EXT_MAX_BLOCKS) { num = last - block; /* find extent for this block */ down_read(&EXT4_I(inode)->i_data_sem); @@ -1958,7 +1958,7 @@ static int ext4_ext_walk_space(struct inode *inode, ext4_lblk_t block, err = -EIO; break; } - err = func(inode, path, &cbex, ex, cbdata); + err = func(inode, next, &cbex, ex, cbdata); ext4_ext_drop_refs(path); if (err < 0) @@ -2020,7 +2020,7 @@ ext4_ext_put_gap_in_cache(struct inode *inode, struct ext4_ext_path *path, if (ex == NULL) { /* there is no extent yet, so gap is [0;-] */ lblock = 0; - len = EXT_MAX_BLOCK; + len = EXT_MAX_BLOCKS; ext_debug("cache gap(whole file):"); } else if (block < le32_to_cpu(ex->ee_block)) { lblock = block; @@ -2350,7 +2350,7 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode, * never happen because at least one of the end points * needs to be on the edge of the extent. */ - if (end == EXT_MAX_BLOCK) { + if (end == EXT_MAX_BLOCKS - 1) { ext_debug(" bad truncate %u:%u\n", start, end); block = 0; @@ -2398,7 +2398,7 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode, * If this is a truncate, this condition * should never happen */ - if (end == EXT_MAX_BLOCK) { + if (end == EXT_MAX_BLOCKS - 1) { ext_debug(" bad truncate %u:%u\n", start, end); err = -EIO; @@ -2478,7 +2478,7 @@ ext4_ext_rm_leaf(handle_t *handle, struct inode *inode, * we need to remove it from the leaf */ if (num == 0) { - if (end != EXT_MAX_BLOCK) { + if (end != EXT_MAX_BLOCKS - 1) { /* * For hole punching, we need to scoot all the * extents up when an extent is removed so that @@ -3699,7 +3699,7 @@ void ext4_ext_truncate(struct inode *inode) last_block = (inode->i_size + sb->s_blocksize - 1) >> EXT4_BLOCK_SIZE_BITS(sb); - err = ext4_ext_remove_space(inode, last_block, EXT_MAX_BLOCK); + err = ext4_ext_remove_space(inode, last_block, EXT_MAX_BLOCKS - 1); /* In a multi-transaction truncate, we only make the final * transaction synchronous. @@ -3914,14 +3914,13 @@ int ext4_convert_unwritten_extents(struct inode *inode, loff_t offset, /* * Callback function called for each extent to gather FIEMAP information. */ -static int ext4_ext_fiemap_cb(struct inode *inode, struct ext4_ext_path *path, +static int ext4_ext_fiemap_cb(struct inode *inode, ext4_lblk_t next, struct ext4_ext_cache *newex, struct ext4_extent *ex, void *data) { __u64 logical; __u64 physical; __u64 length; - loff_t size; __u32 flags = 0; int ret = 0; struct fiemap_extent_info *fieinfo = data; @@ -4103,8 +4102,7 @@ found_delayed_extent: if (ex && ext4_ext_is_uninitialized(ex)) flags |= FIEMAP_EXTENT_UNWRITTEN; - size = i_size_read(inode); - if (logical + length >= size) + if (next == EXT_MAX_BLOCKS) flags |= FIEMAP_EXTENT_LAST; ret = fiemap_fill_next_extent(fieinfo, logical, physical, @@ -4347,8 +4345,8 @@ int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, start_blk = start >> inode->i_sb->s_blocksize_bits; last_blk = (start + len - 1) >> inode->i_sb->s_blocksize_bits; - if (last_blk >= EXT_MAX_BLOCK) - last_blk = EXT_MAX_BLOCK-1; + if (last_blk >= EXT_MAX_BLOCKS) + last_blk = EXT_MAX_BLOCKS-1; len_blks = ((ext4_lblk_t) last_blk) - start_blk + 1; /* diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 50d0e9c64584..e3126c051006 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2634,7 +2634,7 @@ static int ext4_writepage(struct page *page, struct buffer_head *page_bufs = NULL; struct inode *inode = page->mapping->host; - trace_ext4_writepage(inode, page); + trace_ext4_writepage(page); size = i_size_read(inode); if (page->index == size >> PAGE_CACHE_SHIFT) len = size & ~PAGE_CACHE_MASK; @@ -5733,7 +5733,7 @@ int ext4_mark_inode_dirty(handle_t *handle, struct inode *inode) * so would cause a commit on atime updates, which we don't bother doing. * We handle synchronous inodes at the highest possible level. */ -void ext4_dirty_inode(struct inode *inode) +void ext4_dirty_inode(struct inode *inode, int flags) { handle_t *handle; diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 859f2ae8864e..6ed859d56850 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -3578,8 +3578,8 @@ ext4_mb_release_inode_pa(struct ext4_buddy *e4b, struct buffer_head *bitmap_bh, free += next - bit; trace_ext4_mballoc_discard(sb, NULL, group, bit, next - bit); - trace_ext4_mb_release_inode_pa(sb, pa->pa_inode, pa, - grp_blk_start + bit, next - bit); + trace_ext4_mb_release_inode_pa(pa, grp_blk_start + bit, + next - bit); mb_free_blocks(pa->pa_inode, e4b, bit, next - bit); bit = next + 1; } @@ -3608,7 +3608,7 @@ ext4_mb_release_group_pa(struct ext4_buddy *e4b, ext4_group_t group; ext4_grpblk_t bit; - trace_ext4_mb_release_group_pa(sb, pa); + trace_ext4_mb_release_group_pa(pa); BUG_ON(pa->pa_deleted == 0); ext4_get_group_no_and_offset(sb, pa->pa_pstart, &group, &bit); BUG_ON(group != e4b->bd_group && pa->pa_len != 0); @@ -4448,7 +4448,7 @@ ext4_mb_free_metadata(handle_t *handle, struct ext4_buddy *e4b, * @inode: inode * @block: start physical block to free * @count: number of blocks to count - * @metadata: Are these metadata blocks + * @flags: flags used by ext4_free_blocks */ void ext4_free_blocks(handle_t *handle, struct inode *inode, struct buffer_head *bh, ext4_fsblk_t block, diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c index 2b8304bf3c50..f57455a1b1b2 100644 --- a/fs/ext4/move_extent.c +++ b/fs/ext4/move_extent.c @@ -1002,12 +1002,12 @@ mext_check_arguments(struct inode *orig_inode, return -EINVAL; } - if ((orig_start > EXT_MAX_BLOCK) || - (donor_start > EXT_MAX_BLOCK) || - (*len > EXT_MAX_BLOCK) || - (orig_start + *len > EXT_MAX_BLOCK)) { + if ((orig_start >= EXT_MAX_BLOCKS) || + (donor_start >= EXT_MAX_BLOCKS) || + (*len > EXT_MAX_BLOCKS) || + (orig_start + *len >= EXT_MAX_BLOCKS)) { ext4_debug("ext4 move extent: Can't handle over [%u] blocks " - "[ino:orig %lu, donor %lu]\n", EXT_MAX_BLOCK, + "[ino:orig %lu, donor %lu]\n", EXT_MAX_BLOCKS, orig_inode->i_ino, donor_inode->i_ino); return -EINVAL; } diff --git a/fs/ext4/super.c b/fs/ext4/super.c index cc5c157aa11d..9ea71aa864b3 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -2243,6 +2243,12 @@ static void ext4_orphan_cleanup(struct super_block *sb, * in the vfs. ext4 inode has 48 bits of i_block in fsblock units, * so that won't be a limiting factor. * + * However there is other limiting factor. We do store extents in the form + * of starting block and length, hence the resulting length of the extent + * covering maximum file size must fit into on-disk format containers as + * well. Given that length is always by 1 unit bigger than max unit (because + * we count 0 as well) we have to lower the s_maxbytes by one fs block. + * * Note, this does *not* consider any metadata overhead for vfs i_blocks. */ static loff_t ext4_max_size(int blkbits, int has_huge_files) @@ -2264,10 +2270,13 @@ static loff_t ext4_max_size(int blkbits, int has_huge_files) upper_limit <<= blkbits; } - /* 32-bit extent-start container, ee_block */ - res = 1LL << 32; + /* + * 32-bit extent-start container, ee_block. We lower the maxbytes + * by one fs block, so ee_len can cover the extent of maximum file + * size + */ + res = (1LL << 32) - 1; res <<= blkbits; - res -= 1; /* Sanity check against vm- & vfs- imposed limits */ if (res > upper_limit) diff --git a/fs/fat/file.c b/fs/fat/file.c index 7257752b6d5d..7018e1d8902d 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -102,7 +102,7 @@ static int fat_ioctl_set_attributes(struct file *file, u32 __user *user_attr) if (attr & ATTR_SYS) inode->i_flags |= S_IMMUTABLE; else - inode->i_flags &= S_IMMUTABLE; + inode->i_flags &= ~S_IMMUTABLE; } fat_save_attrs(inode, attr); diff --git a/fs/fat/namei_msdos.c b/fs/fat/namei_msdos.c index be15437c272e..3b222dafd15b 100644 --- a/fs/fat/namei_msdos.c +++ b/fs/fat/namei_msdos.c @@ -326,8 +326,6 @@ static int msdos_rmdir(struct inode *dir, struct dentry *dentry) struct fat_slot_info sinfo; int err; - dentry_unhash(dentry); - lock_super(sb); /* * Check whether the directory is not in use, then check @@ -459,9 +457,6 @@ static int do_msdos_rename(struct inode *old_dir, unsigned char *old_name, old_inode = old_dentry->d_inode; new_inode = new_dentry->d_inode; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - err = fat_scan(old_dir, old_name, &old_sinfo); if (err) { err = -EIO; diff --git a/fs/fat/namei_vfat.c b/fs/fat/namei_vfat.c index c61a6789f36c..20b4ea53fdc4 100644 --- a/fs/fat/namei_vfat.c +++ b/fs/fat/namei_vfat.c @@ -824,8 +824,6 @@ static int vfat_rmdir(struct inode *dir, struct dentry *dentry) struct fat_slot_info sinfo; int err; - dentry_unhash(dentry); - lock_super(sb); err = fat_dir_empty(inode); @@ -933,9 +931,6 @@ static int vfat_rename(struct inode *old_dir, struct dentry *old_dentry, int err, is_dir, update_dotdot, corrupt = 0; struct super_block *sb = old_dir->i_sb; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - old_sinfo.bh = sinfo.bh = dotdot_bh = NULL; old_inode = old_dentry->d_inode; new_inode = new_dentry->d_inode; diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 34591ee804b5..0f015a0468de 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -1007,9 +1007,6 @@ static noinline void block_dump___mark_inode_dirty(struct inode *inode) * In short, make sure you hash any inodes _before_ you start marking * them dirty. * - * This function *must* be atomic for the I_DIRTY_PAGES case - - * set_page_dirty() is called under spinlock in several places. - * * Note that for blockdevs, inode->dirtied_when represents the dirtying time of * the block-special inode (/dev/hda1) itself. And the ->dirtied_when field of * the kernel-internal blockdev inode represents the dirtying time of the @@ -1028,7 +1025,7 @@ void __mark_inode_dirty(struct inode *inode, int flags) */ if (flags & (I_DIRTY_SYNC | I_DIRTY_DATASYNC)) { if (sb->s_op->dirty_inode) - sb->s_op->dirty_inode(inode); + sb->s_op->dirty_inode(inode, flags); } /* diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 0d0e3faddcfa..d50160714595 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -667,8 +667,6 @@ static int fuse_rmdir(struct inode *dir, struct dentry *entry) if (IS_ERR(req)) return PTR_ERR(req); - dentry_unhash(entry); - req->in.h.opcode = FUSE_RMDIR; req->in.h.nodeid = get_node_id(dir); req->in.numargs = 1; @@ -694,9 +692,6 @@ static int fuse_rename(struct inode *olddir, struct dentry *oldent, struct fuse_conn *fc = get_fuse_conn(olddir); struct fuse_req *req = fuse_get_req(fc); - if (newent->d_inode && S_ISDIR(newent->d_inode->i_mode)) - dentry_unhash(newent); - if (IS_ERR(req)) return PTR_ERR(req); diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index cc6ec4b2f0ff..38f84cd48b67 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -921,6 +921,8 @@ static int fuse_fill_super(struct super_block *sb, void *data, int silent) if (sb->s_flags & MS_MANDLOCK) goto err; + sb->s_flags &= ~MS_NOSEC; + if (!parse_fuse_opt((char *) data, &d, is_bdev)) goto err; diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 2792a790e50b..1c1336e7b3b2 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -663,14 +663,19 @@ static void glock_work_func(struct work_struct *work) drop_ref = 1; } spin_lock(&gl->gl_spin); - if (test_and_clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) && + if (test_bit(GLF_PENDING_DEMOTE, &gl->gl_flags) && gl->gl_state != LM_ST_UNLOCKED && gl->gl_demote_state != LM_ST_EXCLUSIVE) { unsigned long holdtime, now = jiffies; + holdtime = gl->gl_tchange + gl->gl_ops->go_min_hold_time; if (time_before(now, holdtime)) delay = holdtime - now; - set_bit(delay ? GLF_PENDING_DEMOTE : GLF_DEMOTE, &gl->gl_flags); + + if (!delay) { + clear_bit(GLF_PENDING_DEMOTE, &gl->gl_flags); + set_bit(GLF_DEMOTE, &gl->gl_flags); + } } run_queue(gl, 0); spin_unlock(&gl->gl_spin); diff --git a/fs/hfs/dir.c b/fs/hfs/dir.c index 1cb70cdba2c1..b4d70b13be92 100644 --- a/fs/hfs/dir.c +++ b/fs/hfs/dir.c @@ -253,9 +253,6 @@ static int hfs_remove(struct inode *dir, struct dentry *dentry) struct inode *inode = dentry->d_inode; int res; - if (S_ISDIR(inode->i_mode)) - dentry_unhash(dentry); - if (S_ISDIR(inode->i_mode) && inode->i_size != 2) return -ENOTEMPTY; res = hfs_cat_delete(inode->i_ino, dir, &dentry->d_name); @@ -286,9 +283,6 @@ static int hfs_rename(struct inode *old_dir, struct dentry *old_dentry, /* Unlink destination if it already exists */ if (new_dentry->d_inode) { - if (S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - res = hfs_remove(new_dir, new_dentry); if (res) return res; diff --git a/fs/hfsplus/dir.c b/fs/hfsplus/dir.c index b28835091dd0..4df5059c25da 100644 --- a/fs/hfsplus/dir.c +++ b/fs/hfsplus/dir.c @@ -370,8 +370,6 @@ static int hfsplus_rmdir(struct inode *dir, struct dentry *dentry) struct inode *inode = dentry->d_inode; int res; - dentry_unhash(dentry); - if (inode->i_size != 2) return -ENOTEMPTY; @@ -469,12 +467,10 @@ static int hfsplus_rename(struct inode *old_dir, struct dentry *old_dentry, /* Unlink destination if it already exists */ if (new_dentry->d_inode) { - if (S_ISDIR(new_dentry->d_inode->i_mode)) { - dentry_unhash(new_dentry); + if (S_ISDIR(new_dentry->d_inode->i_mode)) res = hfsplus_rmdir(new_dir, new_dentry); - } else { + else res = hfsplus_unlink(new_dir, new_dentry); - } if (res) return res; } diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c index e6816b9e6903..2638c834ed28 100644 --- a/fs/hostfs/hostfs_kern.c +++ b/fs/hostfs/hostfs_kern.c @@ -683,8 +683,6 @@ int hostfs_rmdir(struct inode *ino, struct dentry *dentry) char *file; int err; - dentry_unhash(dentry); - if ((file = dentry_name(dentry)) == NULL) return -ENOMEM; err = do_rmdir(file); @@ -738,9 +736,6 @@ int hostfs_rename(struct inode *from_ino, struct dentry *from, char *from_name, *to_name; int err; - if (to->d_inode && S_ISDIR(to->d_inode->i_mode)) - dentry_unhash(to); - if ((from_name = dentry_name(from)) == NULL) return -ENOMEM; if ((to_name = dentry_name(to)) == NULL) { diff --git a/fs/hpfs/namei.c b/fs/hpfs/namei.c index ff0ce21c0867..acf95dab2aac 100644 --- a/fs/hpfs/namei.c +++ b/fs/hpfs/namei.c @@ -439,8 +439,6 @@ static int hpfs_rmdir(struct inode *dir, struct dentry *dentry) int err; int r; - dentry_unhash(dentry); - hpfs_adjust_length(name, &len); hpfs_lock(dir->i_sb); err = -ENOENT; @@ -535,9 +533,6 @@ static int hpfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct fnode *fnode; int err; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - if ((err = hpfs_chk_name(new_name, &new_len))) return err; err = 0; hpfs_adjust_length(old_name, &old_len); diff --git a/fs/inode.c b/fs/inode.c index 990d284877a1..43566d17d1b8 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1,9 +1,7 @@ /* - * linux/fs/inode.c - * * (C) 1997 Linus Torvalds + * (C) 1999 Andrea Arcangeli <andrea@suse.de> (dynamic inode allocation) */ - #include <linux/fs.h> #include <linux/mm.h> #include <linux/dcache.h> @@ -27,10 +25,11 @@ #include <linux/prefetch.h> #include <linux/ima.h> #include <linux/cred.h> +#include <linux/buffer_head.h> /* for inode_has_buffers */ #include "internal.h" /* - * inode locking rules. + * Inode locking rules: * * inode->i_lock protects: * inode->i_state, inode->i_hash, __iget() @@ -60,54 +59,11 @@ * inode_hash_lock */ -/* - * This is needed for the following functions: - * - inode_has_buffers - * - invalidate_bdev - * - * FIXME: remove all knowledge of the buffer layer from this file - */ -#include <linux/buffer_head.h> - -/* - * New inode.c implementation. - * - * This implementation has the basic premise of trying - * to be extremely low-overhead and SMP-safe, yet be - * simple enough to be "obviously correct". - * - * Famous last words. - */ - -/* inode dynamic allocation 1999, Andrea Arcangeli <andrea@suse.de> */ - -/* #define INODE_PARANOIA 1 */ -/* #define INODE_DEBUG 1 */ - -/* - * Inode lookup is no longer as critical as it used to be: - * most of the lookups are going to be through the dcache. - */ -#define I_HASHBITS i_hash_shift -#define I_HASHMASK i_hash_mask - static unsigned int i_hash_mask __read_mostly; static unsigned int i_hash_shift __read_mostly; static struct hlist_head *inode_hashtable __read_mostly; static __cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_hash_lock); -/* - * Each inode can be on two separate lists. One is - * the hash list of the inode, used for lookups. The - * other linked list is the "type" list: - * "in_use" - valid inode, i_count > 0, i_nlink > 0 - * "dirty" - as "in_use" but also dirty - * "unused" - valid inode, i_count = 0 - * - * A "dirty" list is maintained for each super block, - * allowing for low-overhead inode sync() operations. - */ - static LIST_HEAD(inode_lru); static DEFINE_SPINLOCK(inode_lru_lock); @@ -424,8 +380,8 @@ static unsigned long hash(struct super_block *sb, unsigned long hashval) tmp = (hashval * (unsigned long)sb) ^ (GOLDEN_RATIO_PRIME + hashval) / L1_CACHE_BYTES; - tmp = tmp ^ ((tmp ^ GOLDEN_RATIO_PRIME) >> I_HASHBITS); - return tmp & I_HASHMASK; + tmp = tmp ^ ((tmp ^ GOLDEN_RATIO_PRIME) >> i_hash_shift); + return tmp & i_hash_mask; } /** @@ -467,7 +423,14 @@ EXPORT_SYMBOL(remove_inode_hash); void end_writeback(struct inode *inode) { might_sleep(); + /* + * We have to cycle tree_lock here because reclaim can be still in the + * process of removing the last page (in __delete_from_page_cache()) + * and we must not free mapping under it. + */ + spin_lock_irq(&inode->i_data.tree_lock); BUG_ON(inode->i_data.nrpages); + spin_unlock_irq(&inode->i_data.tree_lock); BUG_ON(!list_empty(&inode->i_data.private_list)); BUG_ON(!(inode->i_state & I_FREEING)); BUG_ON(inode->i_state & I_CLEAR); diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index 3db5ba4568fc..b3cc8586984e 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -974,7 +974,7 @@ out_no_inode: out_no_read: printk(KERN_WARNING "%s: bread failed, dev=%s, iso_blknum=%d, block=%d\n", __func__, s->s_id, iso_blknum, block); - goto out_freesbi; + goto out_freebh; out_bad_zone_size: printk(KERN_WARNING "ISOFS: Bad logical zone size %ld\n", sbi->s_log_zone_size); @@ -989,6 +989,7 @@ out_unknown_format: out_freebh: brelse(bh); + brelse(pri_bh); out_freesbi: kfree(opt.iocharset); kfree(sbi); diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c index 6a79fd0a1a32..2c62c5aae82f 100644 --- a/fs/jbd2/checkpoint.c +++ b/fs/jbd2/checkpoint.c @@ -97,10 +97,14 @@ static int __try_to_free_cp_buf(struct journal_head *jh) if (jh->b_jlist == BJ_None && !buffer_locked(bh) && !buffer_dirty(bh) && !buffer_write_io_error(bh)) { + /* + * Get our reference so that bh cannot be freed before + * we unlock it + */ + get_bh(bh); JBUFFER_TRACE(jh, "remove from checkpoint list"); ret = __jbd2_journal_remove_checkpoint(jh) + 1; jbd_unlock_bh_state(bh); - jbd2_journal_remove_journal_head(bh); BUFFER_TRACE(bh, "release"); __brelse(bh); } else { @@ -223,8 +227,8 @@ restart: spin_lock(&journal->j_list_lock); goto restart; } + get_bh(bh); if (buffer_locked(bh)) { - atomic_inc(&bh->b_count); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); wait_on_buffer(bh); @@ -243,7 +247,6 @@ restart: */ released = __jbd2_journal_remove_checkpoint(jh); jbd_unlock_bh_state(bh); - jbd2_journal_remove_journal_head(bh); __brelse(bh); } @@ -284,7 +287,7 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh, int ret = 0; if (buffer_locked(bh)) { - atomic_inc(&bh->b_count); + get_bh(bh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); wait_on_buffer(bh); @@ -316,12 +319,12 @@ static int __process_buffer(journal_t *journal, struct journal_head *jh, ret = 1; if (unlikely(buffer_write_io_error(bh))) ret = -EIO; + get_bh(bh); J_ASSERT_JH(jh, !buffer_jbddirty(bh)); BUFFER_TRACE(bh, "remove from checkpoint"); __jbd2_journal_remove_checkpoint(jh); spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); - jbd2_journal_remove_journal_head(bh); __brelse(bh); } else { /* @@ -554,7 +557,8 @@ int jbd2_cleanup_journal_tail(journal_t *journal) /* * journal_clean_one_cp_list * - * Find all the written-back checkpoint buffers in the given list and release them. + * Find all the written-back checkpoint buffers in the given list and + * release them. * * Called with the journal locked. * Called with j_list_lock held. @@ -663,8 +667,8 @@ out: * checkpoint lists. * * The function returns 1 if it frees the transaction, 0 otherwise. + * The function can free jh and bh. * - * This function is called with the journal locked. * This function is called with j_list_lock held. * This function is called with jbd_lock_bh_state(jh2bh(jh)) */ @@ -684,13 +688,14 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) } journal = transaction->t_journal; + JBUFFER_TRACE(jh, "removing from transaction"); __buffer_unlink(jh); jh->b_cp_transaction = NULL; + jbd2_journal_put_journal_head(jh); if (transaction->t_checkpoint_list != NULL || transaction->t_checkpoint_io_list != NULL) goto out; - JBUFFER_TRACE(jh, "transaction has no more buffers"); /* * There is one special case to worry about: if we have just pulled the @@ -701,10 +706,8 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) * The locking here around t_state is a bit sleazy. * See the comment at the end of jbd2_journal_commit_transaction(). */ - if (transaction->t_state != T_FINISHED) { - JBUFFER_TRACE(jh, "belongs to running/committing transaction"); + if (transaction->t_state != T_FINISHED) goto out; - } /* OK, that was the last buffer for the transaction: we can now safely remove this transaction from the log */ @@ -723,7 +726,6 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh) wake_up(&journal->j_wait_logspace); ret = 1; out: - JBUFFER_TRACE(jh, "exit"); return ret; } @@ -742,6 +744,8 @@ void __jbd2_journal_insert_checkpoint(struct journal_head *jh, J_ASSERT_JH(jh, buffer_dirty(jh2bh(jh)) || buffer_jbddirty(jh2bh(jh))); J_ASSERT_JH(jh, jh->b_cp_transaction == NULL); + /* Get reference for checkpointing transaction */ + jbd2_journal_grab_journal_head(jh2bh(jh)); jh->b_cp_transaction = transaction; if (!transaction->t_checkpoint_list) { diff --git a/fs/jbd2/commit.c b/fs/jbd2/commit.c index 7f21cf3aaf92..eef6979821a4 100644 --- a/fs/jbd2/commit.c +++ b/fs/jbd2/commit.c @@ -848,10 +848,16 @@ restart_loop: while (commit_transaction->t_forget) { transaction_t *cp_transaction; struct buffer_head *bh; + int try_to_free = 0; jh = commit_transaction->t_forget; spin_unlock(&journal->j_list_lock); bh = jh2bh(jh); + /* + * Get a reference so that bh cannot be freed before we are + * done with it. + */ + get_bh(bh); jbd_lock_bh_state(bh); J_ASSERT_JH(jh, jh->b_transaction == commit_transaction); @@ -914,28 +920,27 @@ restart_loop: __jbd2_journal_insert_checkpoint(jh, commit_transaction); if (is_journal_aborted(journal)) clear_buffer_jbddirty(bh); - JBUFFER_TRACE(jh, "refile for checkpoint writeback"); - __jbd2_journal_refile_buffer(jh); - jbd_unlock_bh_state(bh); } else { J_ASSERT_BH(bh, !buffer_dirty(bh)); - /* The buffer on BJ_Forget list and not jbddirty means + /* + * The buffer on BJ_Forget list and not jbddirty means * it has been freed by this transaction and hence it * could not have been reallocated until this * transaction has committed. *BUT* it could be * reallocated once we have written all the data to * disk and before we process the buffer on BJ_Forget - * list. */ - JBUFFER_TRACE(jh, "refile or unfile freed buffer"); - __jbd2_journal_refile_buffer(jh); - if (!jh->b_transaction) { - jbd_unlock_bh_state(bh); - /* needs a brelse */ - jbd2_journal_remove_journal_head(bh); - release_buffer_page(bh); - } else - jbd_unlock_bh_state(bh); + * list. + */ + if (!jh->b_next_transaction) + try_to_free = 1; } + JBUFFER_TRACE(jh, "refile or unfile buffer"); + __jbd2_journal_refile_buffer(jh); + jbd_unlock_bh_state(bh); + if (try_to_free) + release_buffer_page(bh); /* Drops bh reference */ + else + __brelse(bh); cond_resched_lock(&journal->j_list_lock); } spin_unlock(&journal->j_list_lock); diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c index 9a7826990304..0dfa5b598e68 100644 --- a/fs/jbd2/journal.c +++ b/fs/jbd2/journal.c @@ -2078,10 +2078,9 @@ static void journal_free_journal_head(struct journal_head *jh) * When a buffer has its BH_JBD bit set it is immune from being released by * core kernel code, mainly via ->b_count. * - * A journal_head may be detached from its buffer_head when the journal_head's - * b_transaction, b_cp_transaction and b_next_transaction pointers are NULL. - * Various places in JBD call jbd2_journal_remove_journal_head() to indicate that the - * journal_head can be dropped if needed. + * A journal_head is detached from its buffer_head when the journal_head's + * b_jcount reaches zero. Running transaction (b_transaction) and checkpoint + * transaction (b_cp_transaction) hold their references to b_jcount. * * Various places in the kernel want to attach a journal_head to a buffer_head * _before_ attaching the journal_head to a transaction. To protect the @@ -2094,17 +2093,16 @@ static void journal_free_journal_head(struct journal_head *jh) * (Attach a journal_head if needed. Increments b_jcount) * struct journal_head *jh = jbd2_journal_add_journal_head(bh); * ... + * (Get another reference for transaction) + * jbd2_journal_grab_journal_head(bh); * jh->b_transaction = xxx; + * (Put original reference) * jbd2_journal_put_journal_head(jh); - * - * Now, the journal_head's b_jcount is zero, but it is safe from being released - * because it has a non-zero b_transaction. */ /* * Give a buffer_head a journal_head. * - * Doesn't need the journal lock. * May sleep. */ struct journal_head *jbd2_journal_add_journal_head(struct buffer_head *bh) @@ -2168,61 +2166,29 @@ static void __journal_remove_journal_head(struct buffer_head *bh) struct journal_head *jh = bh2jh(bh); J_ASSERT_JH(jh, jh->b_jcount >= 0); - - get_bh(bh); - if (jh->b_jcount == 0) { - if (jh->b_transaction == NULL && - jh->b_next_transaction == NULL && - jh->b_cp_transaction == NULL) { - J_ASSERT_JH(jh, jh->b_jlist == BJ_None); - J_ASSERT_BH(bh, buffer_jbd(bh)); - J_ASSERT_BH(bh, jh2bh(jh) == bh); - BUFFER_TRACE(bh, "remove journal_head"); - if (jh->b_frozen_data) { - printk(KERN_WARNING "%s: freeing " - "b_frozen_data\n", - __func__); - jbd2_free(jh->b_frozen_data, bh->b_size); - } - if (jh->b_committed_data) { - printk(KERN_WARNING "%s: freeing " - "b_committed_data\n", - __func__); - jbd2_free(jh->b_committed_data, bh->b_size); - } - bh->b_private = NULL; - jh->b_bh = NULL; /* debug, really */ - clear_buffer_jbd(bh); - __brelse(bh); - journal_free_journal_head(jh); - } else { - BUFFER_TRACE(bh, "journal_head was locked"); - } + J_ASSERT_JH(jh, jh->b_transaction == NULL); + J_ASSERT_JH(jh, jh->b_next_transaction == NULL); + J_ASSERT_JH(jh, jh->b_cp_transaction == NULL); + J_ASSERT_JH(jh, jh->b_jlist == BJ_None); + J_ASSERT_BH(bh, buffer_jbd(bh)); + J_ASSERT_BH(bh, jh2bh(jh) == bh); + BUFFER_TRACE(bh, "remove journal_head"); + if (jh->b_frozen_data) { + printk(KERN_WARNING "%s: freeing b_frozen_data\n", __func__); + jbd2_free(jh->b_frozen_data, bh->b_size); } + if (jh->b_committed_data) { + printk(KERN_WARNING "%s: freeing b_committed_data\n", __func__); + jbd2_free(jh->b_committed_data, bh->b_size); + } + bh->b_private = NULL; + jh->b_bh = NULL; /* debug, really */ + clear_buffer_jbd(bh); + journal_free_journal_head(jh); } /* - * jbd2_journal_remove_journal_head(): if the buffer isn't attached to a transaction - * and has a zero b_jcount then remove and release its journal_head. If we did - * see that the buffer is not used by any transaction we also "logically" - * decrement ->b_count. - * - * We in fact take an additional increment on ->b_count as a convenience, - * because the caller usually wants to do additional things with the bh - * after calling here. - * The caller of jbd2_journal_remove_journal_head() *must* run __brelse(bh) at some - * time. Once the caller has run __brelse(), the buffer is eligible for - * reaping by try_to_free_buffers(). - */ -void jbd2_journal_remove_journal_head(struct buffer_head *bh) -{ - jbd_lock_bh_journal_head(bh); - __journal_remove_journal_head(bh); - jbd_unlock_bh_journal_head(bh); -} - -/* - * Drop a reference on the passed journal_head. If it fell to zero then try to + * Drop a reference on the passed journal_head. If it fell to zero then * release the journal_head from the buffer_head. */ void jbd2_journal_put_journal_head(struct journal_head *jh) @@ -2232,11 +2198,12 @@ void jbd2_journal_put_journal_head(struct journal_head *jh) jbd_lock_bh_journal_head(bh); J_ASSERT_JH(jh, jh->b_jcount > 0); --jh->b_jcount; - if (!jh->b_jcount && !jh->b_transaction) { + if (!jh->b_jcount) { __journal_remove_journal_head(bh); + jbd_unlock_bh_journal_head(bh); __brelse(bh); - } - jbd_unlock_bh_journal_head(bh); + } else + jbd_unlock_bh_journal_head(bh); } /* diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 3eec82d32fd4..2d7109414cdd 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -30,6 +30,7 @@ #include <linux/module.h> static void __jbd2_journal_temp_unlink_buffer(struct journal_head *jh); +static void __jbd2_journal_unfile_buffer(struct journal_head *jh); /* * jbd2_get_transaction: obtain a new transaction_t object. @@ -764,7 +765,6 @@ repeat: if (!jh->b_transaction) { JBUFFER_TRACE(jh, "no transaction"); J_ASSERT_JH(jh, !jh->b_next_transaction); - jh->b_transaction = transaction; JBUFFER_TRACE(jh, "file as BJ_Reserved"); spin_lock(&journal->j_list_lock); __jbd2_journal_file_buffer(jh, transaction, BJ_Reserved); @@ -814,7 +814,6 @@ out: * int jbd2_journal_get_write_access() - notify intent to modify a buffer for metadata (not data) update. * @handle: transaction to add buffer modifications to * @bh: bh to be used for metadata writes - * @credits: variable that will receive credits for the buffer * * Returns an error code or 0 on success. * @@ -896,8 +895,6 @@ int jbd2_journal_get_create_access(handle_t *handle, struct buffer_head *bh) * committed and so it's safe to clear the dirty bit. */ clear_buffer_dirty(jh2bh(jh)); - jh->b_transaction = transaction; - /* first access by this transaction */ jh->b_modified = 0; @@ -932,7 +929,6 @@ out: * non-rewindable consequences * @handle: transaction * @bh: buffer to undo - * @credits: store the number of taken credits here (if not NULL) * * Sometimes there is a need to distinguish between metadata which has * been committed to disk and that which has not. The ext3fs code uses @@ -1232,8 +1228,6 @@ int jbd2_journal_forget (handle_t *handle, struct buffer_head *bh) __jbd2_journal_file_buffer(jh, transaction, BJ_Forget); } else { __jbd2_journal_unfile_buffer(jh); - jbd2_journal_remove_journal_head(bh); - __brelse(bh); if (!buffer_jbd(bh)) { spin_unlock(&journal->j_list_lock); jbd_unlock_bh_state(bh); @@ -1556,19 +1550,32 @@ void __jbd2_journal_temp_unlink_buffer(struct journal_head *jh) mark_buffer_dirty(bh); /* Expose it to the VM */ } -void __jbd2_journal_unfile_buffer(struct journal_head *jh) +/* + * Remove buffer from all transactions. + * + * Called with bh_state lock and j_list_lock + * + * jh and bh may be already freed when this function returns. + */ +static void __jbd2_journal_unfile_buffer(struct journal_head *jh) { __jbd2_journal_temp_unlink_buffer(jh); jh->b_transaction = NULL; + jbd2_journal_put_journal_head(jh); } void jbd2_journal_unfile_buffer(journal_t *journal, struct journal_head *jh) { - jbd_lock_bh_state(jh2bh(jh)); + struct buffer_head *bh = jh2bh(jh); + + /* Get reference so that buffer cannot be freed before we unlock it */ + get_bh(bh); + jbd_lock_bh_state(bh); spin_lock(&journal->j_list_lock); __jbd2_journal_unfile_buffer(jh); spin_unlock(&journal->j_list_lock); - jbd_unlock_bh_state(jh2bh(jh)); + jbd_unlock_bh_state(bh); + __brelse(bh); } /* @@ -1595,8 +1602,6 @@ __journal_try_to_free_buffer(journal_t *journal, struct buffer_head *bh) if (jh->b_jlist == BJ_None) { JBUFFER_TRACE(jh, "remove from checkpoint list"); __jbd2_journal_remove_checkpoint(jh); - jbd2_journal_remove_journal_head(bh); - __brelse(bh); } } spin_unlock(&journal->j_list_lock); @@ -1659,7 +1664,6 @@ int jbd2_journal_try_to_free_buffers(journal_t *journal, /* * We take our own ref against the journal_head here to avoid * having to add tons of locking around each instance of - * jbd2_journal_remove_journal_head() and * jbd2_journal_put_journal_head(). */ jh = jbd2_journal_grab_journal_head(bh); @@ -1697,10 +1701,9 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction) int may_free = 1; struct buffer_head *bh = jh2bh(jh); - __jbd2_journal_unfile_buffer(jh); - if (jh->b_cp_transaction) { JBUFFER_TRACE(jh, "on running+cp transaction"); + __jbd2_journal_temp_unlink_buffer(jh); /* * We don't want to write the buffer anymore, clear the * bit so that we don't confuse checks in @@ -1711,8 +1714,7 @@ static int __dispose_buffer(struct journal_head *jh, transaction_t *transaction) may_free = 0; } else { JBUFFER_TRACE(jh, "on running transaction"); - jbd2_journal_remove_journal_head(bh); - __brelse(bh); + __jbd2_journal_unfile_buffer(jh); } return may_free; } @@ -1990,6 +1992,8 @@ void __jbd2_journal_file_buffer(struct journal_head *jh, if (jh->b_transaction) __jbd2_journal_temp_unlink_buffer(jh); + else + jbd2_journal_grab_journal_head(bh); jh->b_transaction = transaction; switch (jlist) { @@ -2041,9 +2045,10 @@ void jbd2_journal_file_buffer(struct journal_head *jh, * already started to be used by a subsequent transaction, refile the * buffer on that transaction's metadata list. * - * Called under journal->j_list_lock - * + * Called under j_list_lock * Called under jbd_lock_bh_state(jh2bh(jh)) + * + * jh and bh may be already free when this function returns */ void __jbd2_journal_refile_buffer(struct journal_head *jh) { @@ -2067,6 +2072,11 @@ void __jbd2_journal_refile_buffer(struct journal_head *jh) was_dirty = test_clear_buffer_jbddirty(bh); __jbd2_journal_temp_unlink_buffer(jh); + /* + * We set b_transaction here because b_next_transaction will inherit + * our jh reference and thus __jbd2_journal_file_buffer() must not + * take a new one. + */ jh->b_transaction = jh->b_next_transaction; jh->b_next_transaction = NULL; if (buffer_freed(bh)) @@ -2083,30 +2093,21 @@ void __jbd2_journal_refile_buffer(struct journal_head *jh) } /* - * For the unlocked version of this call, also make sure that any - * hanging journal_head is cleaned up if necessary. - * - * __jbd2_journal_refile_buffer is usually called as part of a single locked - * operation on a buffer_head, in which the caller is probably going to - * be hooking the journal_head onto other lists. In that case it is up - * to the caller to remove the journal_head if necessary. For the - * unlocked jbd2_journal_refile_buffer call, the caller isn't going to be - * doing anything else to the buffer so we need to do the cleanup - * ourselves to avoid a jh leak. - * - * *** The journal_head may be freed by this call! *** + * __jbd2_journal_refile_buffer() with necessary locking added. We take our + * bh reference so that we can safely unlock bh. + * + * The jh and bh may be freed by this call. */ void jbd2_journal_refile_buffer(journal_t *journal, struct journal_head *jh) { struct buffer_head *bh = jh2bh(jh); + /* Get reference so that buffer cannot be freed before we unlock it */ + get_bh(bh); jbd_lock_bh_state(bh); spin_lock(&journal->j_list_lock); - __jbd2_journal_refile_buffer(jh); jbd_unlock_bh_state(bh); - jbd2_journal_remove_journal_head(bh); - spin_unlock(&journal->j_list_lock); __brelse(bh); } diff --git a/fs/jffs2/dir.c b/fs/jffs2/dir.c index 05f73328b28b..4bca6a2e5c07 100644 --- a/fs/jffs2/dir.c +++ b/fs/jffs2/dir.c @@ -75,7 +75,6 @@ static struct dentry *jffs2_lookup(struct inode *dir_i, struct dentry *target, struct nameidata *nd) { struct jffs2_inode_info *dir_f; - struct jffs2_sb_info *c; struct jffs2_full_dirent *fd = NULL, *fd_list; uint32_t ino = 0; struct inode *inode = NULL; @@ -86,7 +85,6 @@ static struct dentry *jffs2_lookup(struct inode *dir_i, struct dentry *target, return ERR_PTR(-ENAMETOOLONG); dir_f = JFFS2_INODE_INFO(dir_i); - c = JFFS2_SB_INFO(dir_i->i_sb); mutex_lock(&dir_f->sem); @@ -119,7 +117,6 @@ static struct dentry *jffs2_lookup(struct inode *dir_i, struct dentry *target, static int jffs2_readdir(struct file *filp, void *dirent, filldir_t filldir) { struct jffs2_inode_info *f; - struct jffs2_sb_info *c; struct inode *inode = filp->f_path.dentry->d_inode; struct jffs2_full_dirent *fd; unsigned long offset, curofs; @@ -127,7 +124,6 @@ static int jffs2_readdir(struct file *filp, void *dirent, filldir_t filldir) D1(printk(KERN_DEBUG "jffs2_readdir() for dir_i #%lu\n", filp->f_path.dentry->d_inode->i_ino)); f = JFFS2_INODE_INFO(inode); - c = JFFS2_SB_INFO(inode->i_sb); offset = filp->f_pos; @@ -609,8 +605,6 @@ static int jffs2_rmdir (struct inode *dir_i, struct dentry *dentry) int ret; uint32_t now = get_seconds(); - dentry_unhash(dentry); - for (fd = f->dents ; fd; fd = fd->next) { if (fd->ino) return -ENOTEMPTY; @@ -786,9 +780,6 @@ static int jffs2_rename (struct inode *old_dir_i, struct dentry *old_dentry, uint8_t type; uint32_t now; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - /* The VFS will check for us and prevent trying to rename a * file over a directory and vice versa, but if it's a directory, * the VFS can't check whether the victim is empty. The filesystem diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c index e896e67767eb..46ad619b6124 100644 --- a/fs/jffs2/fs.c +++ b/fs/jffs2/fs.c @@ -357,7 +357,7 @@ error: return ERR_PTR(ret); } -void jffs2_dirty_inode(struct inode *inode) +void jffs2_dirty_inode(struct inode *inode, int flags) { struct iattr iattr; diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h index 00bae7cc2e48..65c6c43ca482 100644 --- a/fs/jffs2/os-linux.h +++ b/fs/jffs2/os-linux.h @@ -172,7 +172,7 @@ int jffs2_setattr (struct dentry *, struct iattr *); int jffs2_do_setattr (struct inode *, struct iattr *); struct inode *jffs2_iget(struct super_block *, unsigned long); void jffs2_evict_inode (struct inode *); -void jffs2_dirty_inode(struct inode *inode); +void jffs2_dirty_inode(struct inode *inode, int flags); struct inode *jffs2_new_inode (struct inode *dir_i, int mode, struct jffs2_raw_inode *ri); int jffs2_statfs (struct dentry *, struct kstatfs *); diff --git a/fs/jffs2/scan.c b/fs/jffs2/scan.c index b632dddcb482..8d8cd3419d02 100644 --- a/fs/jffs2/scan.c +++ b/fs/jffs2/scan.c @@ -94,7 +94,7 @@ int jffs2_scan_medium(struct jffs2_sb_info *c) uint32_t buf_size = 0; struct jffs2_summary *s = NULL; /* summary info collected by the scan process */ #ifndef __ECOS - size_t pointlen; + size_t pointlen, try_size; if (c->mtd->point) { ret = c->mtd->point(c->mtd, 0, c->mtd->size, &pointlen, @@ -113,18 +113,21 @@ int jffs2_scan_medium(struct jffs2_sb_info *c) /* For NAND it's quicker to read a whole eraseblock at a time, apparently */ if (jffs2_cleanmarker_oob(c)) - buf_size = c->sector_size; + try_size = c->sector_size; else - buf_size = PAGE_SIZE; + try_size = PAGE_SIZE; - /* Respect kmalloc limitations */ - if (buf_size > 128*1024) - buf_size = 128*1024; + D1(printk(KERN_DEBUG "Trying to allocate readbuf of %zu " + "bytes\n", try_size)); - D1(printk(KERN_DEBUG "Allocating readbuf of %d bytes\n", buf_size)); - flashbuf = kmalloc(buf_size, GFP_KERNEL); + flashbuf = mtd_kmalloc_up_to(c->mtd, &try_size); if (!flashbuf) return -ENOMEM; + + D1(printk(KERN_DEBUG "Allocated readbuf of %zu bytes\n", + try_size)); + + buf_size = (uint32_t)try_size; } if (jffs2_sum_active()) { diff --git a/fs/jfs/file.c b/fs/jfs/file.c index c5ce6c1d1ff4..2f3f531f3606 100644 --- a/fs/jfs/file.c +++ b/fs/jfs/file.c @@ -66,9 +66,9 @@ static int jfs_open(struct inode *inode, struct file *file) struct jfs_inode_info *ji = JFS_IP(inode); spin_lock_irq(&ji->ag_lock); if (ji->active_ag == -1) { - ji->active_ag = ji->agno; - atomic_inc( - &JFS_SBI(inode->i_sb)->bmap->db_active[ji->agno]); + struct jfs_sb_info *jfs_sb = JFS_SBI(inode->i_sb); + ji->active_ag = BLKTOAG(addressPXD(&ji->ixpxd), jfs_sb); + atomic_inc( &jfs_sb->bmap->db_active[ji->active_ag]); } spin_unlock_irq(&ji->ag_lock); } diff --git a/fs/jfs/inode.c b/fs/jfs/inode.c index eddbb373209e..109655904bbc 100644 --- a/fs/jfs/inode.c +++ b/fs/jfs/inode.c @@ -173,7 +173,7 @@ void jfs_evict_inode(struct inode *inode) dquot_drop(inode); } -void jfs_dirty_inode(struct inode *inode) +void jfs_dirty_inode(struct inode *inode, int flags) { static int noisy = 5; diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c index ed53a4740168..b78b2f978f04 100644 --- a/fs/jfs/jfs_imap.c +++ b/fs/jfs/jfs_imap.c @@ -397,7 +397,7 @@ int diRead(struct inode *ip) release_metapage(mp); /* set the ag for the inode */ - JFS_IP(ip)->agno = BLKTOAG(agstart, sbi); + JFS_IP(ip)->agstart = agstart; JFS_IP(ip)->active_ag = -1; return (rc); @@ -901,7 +901,7 @@ int diFree(struct inode *ip) /* get the allocation group for this ino. */ - agno = JFS_IP(ip)->agno; + agno = BLKTOAG(JFS_IP(ip)->agstart, JFS_SBI(ip->i_sb)); /* Lock the AG specific inode map information */ @@ -1315,12 +1315,11 @@ int diFree(struct inode *ip) static inline void diInitInode(struct inode *ip, int iagno, int ino, int extno, struct iag * iagp) { - struct jfs_sb_info *sbi = JFS_SBI(ip->i_sb); struct jfs_inode_info *jfs_ip = JFS_IP(ip); ip->i_ino = (iagno << L2INOSPERIAG) + ino; jfs_ip->ixpxd = iagp->inoext[extno]; - jfs_ip->agno = BLKTOAG(le64_to_cpu(iagp->agstart), sbi); + jfs_ip->agstart = le64_to_cpu(iagp->agstart); jfs_ip->active_ag = -1; } @@ -1379,7 +1378,7 @@ int diAlloc(struct inode *pip, bool dir, struct inode *ip) */ /* get the ag number of this iag */ - agno = JFS_IP(pip)->agno; + agno = BLKTOAG(JFS_IP(pip)->agstart, JFS_SBI(pip->i_sb)); if (atomic_read(&JFS_SBI(pip->i_sb)->bmap->db_active[agno])) { /* @@ -2921,10 +2920,9 @@ int diExtendFS(struct inode *ipimap, struct inode *ipbmap) continue; } - /* agstart that computes to the same ag is treated as same; */ agstart = le64_to_cpu(iagp->agstart); - /* iagp->agstart = agstart & ~(mp->db_agsize - 1); */ n = agstart >> mp->db_agl2size; + iagp->agstart = cpu_to_le64((s64)n << mp->db_agl2size); /* compute backed inodes */ numinos = (EXTSPERIAG - le32_to_cpu(iagp->nfreeexts)) diff --git a/fs/jfs/jfs_incore.h b/fs/jfs/jfs_incore.h index 1439f119ec83..584a4a1a6e81 100644 --- a/fs/jfs/jfs_incore.h +++ b/fs/jfs/jfs_incore.h @@ -50,8 +50,9 @@ struct jfs_inode_info { short btindex; /* btpage entry index*/ struct inode *ipimap; /* inode map */ unsigned long cflag; /* commit flags */ + u64 agstart; /* agstart of the containing IAG */ u16 bxflag; /* xflag of pseudo buffer? */ - unchar agno; /* ag number */ + unchar pad; signed char active_ag; /* ag currently allocating from */ lid_t blid; /* lid of pseudo buffer? */ lid_t atlhead; /* anonymous tlock list head */ diff --git a/fs/jfs/jfs_inode.h b/fs/jfs/jfs_inode.h index 155e91eff07d..ec2fb8b945fc 100644 --- a/fs/jfs/jfs_inode.h +++ b/fs/jfs/jfs_inode.h @@ -28,7 +28,7 @@ extern struct inode *jfs_iget(struct super_block *, unsigned long); extern int jfs_commit_inode(struct inode *, int); extern int jfs_write_inode(struct inode *, struct writeback_control *); extern void jfs_evict_inode(struct inode *); -extern void jfs_dirty_inode(struct inode *); +extern void jfs_dirty_inode(struct inode *, int); extern void jfs_truncate(struct inode *); extern void jfs_truncate_nolock(struct inode *, loff_t); extern void jfs_free_zero_link(struct inode *); diff --git a/fs/jfs/jfs_logmgr.c b/fs/jfs/jfs_logmgr.c index 278e3fb40b71..583636f745e5 100644 --- a/fs/jfs/jfs_logmgr.c +++ b/fs/jfs/jfs_logmgr.c @@ -1123,7 +1123,7 @@ int lmLogOpen(struct super_block *sb) bdev = blkdev_get_by_dev(sbi->logdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL, log); if (IS_ERR(bdev)) { - rc = -PTR_ERR(bdev); + rc = PTR_ERR(bdev); goto free; } diff --git a/fs/jfs/namei.c b/fs/jfs/namei.c index 865df16a6cf3..eaaf2b511e89 100644 --- a/fs/jfs/namei.c +++ b/fs/jfs/namei.c @@ -360,8 +360,6 @@ static int jfs_rmdir(struct inode *dip, struct dentry *dentry) jfs_info("jfs_rmdir: dip:0x%p name:%s", dip, dentry->d_name.name); - dentry_unhash(dentry); - /* Init inode for quota operations. */ dquot_initialize(dip); dquot_initialize(ip); @@ -1097,9 +1095,6 @@ static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry, jfs_info("jfs_rename: %s %s", old_dentry->d_name.name, new_dentry->d_name.name); - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - dquot_initialize(old_dir); dquot_initialize(new_dir); diff --git a/fs/jfs/resize.c b/fs/jfs/resize.c index 8ea5efb5a34e..8d0c1c7c0820 100644 --- a/fs/jfs/resize.c +++ b/fs/jfs/resize.c @@ -80,7 +80,7 @@ int jfs_extendfs(struct super_block *sb, s64 newLVSize, int newLogSize) int log_formatted = 0; struct inode *iplist[1]; struct jfs_superblock *j_sb, *j_sb2; - uint old_agsize; + s64 old_agsize; int agsizechanged = 0; struct buffer_head *bh, *bh2; diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index adb45ec9038c..e374050a911c 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -708,7 +708,13 @@ static void nlmclnt_unlock_callback(struct rpc_task *task, void *data) if (task->tk_status < 0) { dprintk("lockd: unlock failed (err = %d)\n", -task->tk_status); - goto retry_rebind; + switch (task->tk_status) { + case -EACCES: + case -EIO: + goto die; + default: + goto retry_rebind; + } } if (status == NLM_LCK_DENIED_GRACE_PERIOD) { rpc_delay(task, NLMCLNT_GRACE_WAIT); diff --git a/fs/logfs/dir.c b/fs/logfs/dir.c index f34c9cde9e94..1afae26cf236 100644 --- a/fs/logfs/dir.c +++ b/fs/logfs/dir.c @@ -273,8 +273,6 @@ static int logfs_rmdir(struct inode *dir, struct dentry *dentry) { struct inode *inode = dentry->d_inode; - dentry_unhash(dentry); - if (!logfs_empty_dir(inode)) return -ENOTEMPTY; @@ -557,13 +555,6 @@ static int logfs_symlink(struct inode *dir, struct dentry *dentry, return __logfs_create(dir, dentry, inode, target, destlen); } -static int logfs_permission(struct inode *inode, int mask, unsigned int flags) -{ - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - return generic_permission(inode, mask, flags, NULL); -} - static int logfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *dentry) { @@ -624,9 +615,6 @@ static int logfs_rename_cross(struct inode *old_dir, struct dentry *old_dentry, loff_t pos; int err; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - /* 1. locate source dd */ err = logfs_get_dd(old_dir, old_dentry, &dd, &pos); if (err) @@ -825,7 +813,6 @@ const struct inode_operations logfs_dir_iops = { .mknod = logfs_mknod, .rename = logfs_rename, .rmdir = logfs_rmdir, - .permission = logfs_permission, .symlink = logfs_symlink, .unlink = logfs_unlink, }; diff --git a/fs/minix/namei.c b/fs/minix/namei.c index f60aed8db9c4..6e6777f1b4b2 100644 --- a/fs/minix/namei.c +++ b/fs/minix/namei.c @@ -168,8 +168,6 @@ static int minix_rmdir(struct inode * dir, struct dentry *dentry) struct inode * inode = dentry->d_inode; int err = -ENOTEMPTY; - dentry_unhash(dentry); - if (minix_empty_dir(inode)) { err = minix_unlink(dir, dentry); if (!err) { @@ -192,9 +190,6 @@ static int minix_rename(struct inode * old_dir, struct dentry *old_dentry, struct minix_dir_entry * old_de; int err = -ENOENT; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - old_de = minix_find_entry(old_dentry, &old_page); if (!old_de) goto out; diff --git a/fs/namei.c b/fs/namei.c index 2358b326b221..0223c41fb114 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -238,7 +238,8 @@ int generic_permission(struct inode *inode, int mask, unsigned int flags, /* * Read/write DACs are always overridable. - * Executable DACs are overridable if at least one exec bit is set. + * Executable DACs are overridable for all directories and + * for non-directories that have least one exec bit set. */ if (!(mask & MAY_EXEC) || execute_ok(inode)) if (ns_capable(inode_userns(inode), CAP_DAC_OVERRIDE)) @@ -812,6 +813,11 @@ static int follow_automount(struct path *path, unsigned flags, if (!mnt) /* mount collision */ return 0; + if (!*need_mntput) { + /* lock_mount() may release path->mnt on error */ + mntget(path->mnt); + *need_mntput = true; + } err = finish_automount(mnt, path); switch (err) { @@ -819,12 +825,9 @@ static int follow_automount(struct path *path, unsigned flags, /* Someone else made a mount here whilst we were busy */ return 0; case 0: - dput(path->dentry); - if (*need_mntput) - mntput(path->mnt); + path_put(path); path->mnt = mnt; path->dentry = dget(mnt->mnt_root); - *need_mntput = true; return 0; default: return err; @@ -844,9 +847,10 @@ static int follow_automount(struct path *path, unsigned flags, */ static int follow_managed(struct path *path, unsigned flags) { + struct vfsmount *mnt = path->mnt; /* held by caller, must be left alone */ unsigned managed; bool need_mntput = false; - int ret; + int ret = 0; /* Given that we're not holding a lock here, we retain the value in a * local variable for each dentry as we look at it so that we don't see @@ -861,7 +865,7 @@ static int follow_managed(struct path *path, unsigned flags) BUG_ON(!path->dentry->d_op->d_manage); ret = path->dentry->d_op->d_manage(path->dentry, false); if (ret < 0) - return ret == -EISDIR ? 0 : ret; + break; } /* Transit to a mounted filesystem. */ @@ -887,14 +891,19 @@ static int follow_managed(struct path *path, unsigned flags) if (managed & DCACHE_NEED_AUTOMOUNT) { ret = follow_automount(path, flags, &need_mntput); if (ret < 0) - return ret == -EISDIR ? 0 : ret; + break; continue; } /* We didn't change the current path point */ break; } - return 0; + + if (need_mntput && path->mnt == mnt) + mntput(path->mnt); + if (ret == -EISDIR) + ret = 0; + return ret; } int follow_down_one(struct path *path) @@ -919,12 +928,11 @@ static inline bool managed_dentry_might_block(struct dentry *dentry) } /* - * Skip to top of mountpoint pile in rcuwalk mode. We abort the rcu-walk if we - * meet a managed dentry and we're not walking to "..". True is returned to - * continue, false to abort. + * Try to skip to top of mountpoint pile in rcuwalk mode. Fail if + * we meet a managed dentry that would need blocking. */ static bool __follow_mount_rcu(struct nameidata *nd, struct path *path, - struct inode **inode, bool reverse_transit) + struct inode **inode) { for (;;) { struct vfsmount *mounted; @@ -933,8 +941,7 @@ static bool __follow_mount_rcu(struct nameidata *nd, struct path *path, * that wants to block transit. */ *inode = path->dentry->d_inode; - if (!reverse_transit && - unlikely(managed_dentry_might_block(path->dentry))) + if (unlikely(managed_dentry_might_block(path->dentry))) return false; if (!d_mountpoint(path->dentry)) @@ -947,16 +954,24 @@ static bool __follow_mount_rcu(struct nameidata *nd, struct path *path, path->dentry = mounted->mnt_root; nd->seq = read_seqcount_begin(&path->dentry->d_seq); } - - if (unlikely(path->dentry->d_flags & DCACHE_NEED_AUTOMOUNT)) - return reverse_transit; return true; } -static int follow_dotdot_rcu(struct nameidata *nd) +static void follow_mount_rcu(struct nameidata *nd) { - struct inode *inode = nd->inode; + while (d_mountpoint(nd->path.dentry)) { + struct vfsmount *mounted; + mounted = __lookup_mnt(nd->path.mnt, nd->path.dentry, 1); + if (!mounted) + break; + nd->path.mnt = mounted; + nd->path.dentry = mounted->mnt_root; + nd->seq = read_seqcount_begin(&nd->path.dentry->d_seq); + } +} +static int follow_dotdot_rcu(struct nameidata *nd) +{ set_root_rcu(nd); while (1) { @@ -972,7 +987,6 @@ static int follow_dotdot_rcu(struct nameidata *nd) seq = read_seqcount_begin(&parent->d_seq); if (read_seqcount_retry(&old->d_seq, nd->seq)) goto failed; - inode = parent->d_inode; nd->path.dentry = parent; nd->seq = seq; break; @@ -980,10 +994,9 @@ static int follow_dotdot_rcu(struct nameidata *nd) if (!follow_up_rcu(&nd->path)) break; nd->seq = read_seqcount_begin(&nd->path.dentry->d_seq); - inode = nd->path.dentry->d_inode; } - __follow_mount_rcu(nd, &nd->path, &inode, true); - nd->inode = inode; + follow_mount_rcu(nd); + nd->inode = nd->path.dentry->d_inode; return 0; failed: @@ -999,9 +1012,6 @@ failed: * Follow down to the covering mount currently visible to userspace. At each * point, the filesystem owning that dentry may be queried as to whether the * caller is permitted to proceed or not. - * - * Care must be taken as namespace_sem may be held (indicated by mounting_here - * being true). */ int follow_down(struct path *path) { @@ -1157,8 +1167,11 @@ static int do_lookup(struct nameidata *nd, struct qstr *name, } path->mnt = mnt; path->dentry = dentry; - if (likely(__follow_mount_rcu(nd, path, inode, false))) - return 0; + if (unlikely(!__follow_mount_rcu(nd, path, inode))) + goto unlazy; + if (unlikely(path->dentry->d_flags & DCACHE_NEED_AUTOMOUNT)) + goto unlazy; + return 0; unlazy: if (unlazy_walk(nd, dentry)) return -ECHILD; @@ -2572,6 +2585,7 @@ int vfs_rmdir(struct inode *dir, struct dentry *dentry) if (error) goto out; + shrink_dcache_parent(dentry); error = dir->i_op->rmdir(dir, dentry); if (error) goto out; @@ -2616,6 +2630,10 @@ static long do_rmdir(int dfd, const char __user *pathname) error = PTR_ERR(dentry); if (IS_ERR(dentry)) goto exit2; + if (!dentry->d_inode) { + error = -ENOENT; + goto exit3; + } error = mnt_want_write(nd.path.mnt); if (error) goto exit3; @@ -2704,8 +2722,9 @@ static long do_unlinkat(int dfd, const char __user *pathname) if (nd.last.name[nd.last.len]) goto slashes; inode = dentry->d_inode; - if (inode) - ihold(inode); + if (!inode) + goto slashes; + ihold(inode); error = mnt_want_write(nd.path.mnt); if (error) goto exit2; @@ -2986,6 +3005,8 @@ static int vfs_rename_dir(struct inode *old_dir, struct dentry *old_dentry, if (d_mountpoint(old_dentry) || d_mountpoint(new_dentry)) goto out; + if (target) + shrink_dcache_parent(new_dentry); error = old_dir->i_op->rename(old_dir, old_dentry, new_dir, new_dentry); if (error) goto out; diff --git a/fs/ncpfs/dir.c b/fs/ncpfs/dir.c index e3e646b06404..9c51f621e901 100644 --- a/fs/ncpfs/dir.c +++ b/fs/ncpfs/dir.c @@ -1033,8 +1033,11 @@ static int ncp_rmdir(struct inode *dir, struct dentry *dentry) DPRINTK("ncp_rmdir: removing %s/%s\n", dentry->d_parent->d_name.name, dentry->d_name.name); + /* + * fail with EBUSY if there are still references to this + * directory. + */ dentry_unhash(dentry); - error = -EBUSY; if (!d_unhashed(dentry)) goto out; @@ -1141,8 +1144,16 @@ static int ncp_rename(struct inode *old_dir, struct dentry *old_dentry, old_dentry->d_parent->d_name.name, old_dentry->d_name.name, new_dentry->d_parent->d_name.name, new_dentry->d_name.name); - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) + if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) { + /* + * fail with EBUSY if there are still references to this + * directory. + */ dentry_unhash(new_dentry); + error = -EBUSY; + if (!d_unhashed(new_dentry)) + goto out; + } ncp_age_dentry(server, old_dentry); ncp_age_dentry(server, new_dentry); diff --git a/fs/nfs/Kconfig b/fs/nfs/Kconfig index ba306658a6db..81515545ba75 100644 --- a/fs/nfs/Kconfig +++ b/fs/nfs/Kconfig @@ -87,6 +87,16 @@ config NFS_V4_1 config PNFS_FILE_LAYOUT tristate +config PNFS_OBJLAYOUT + tristate "Provide support for the pNFS Objects Layout Driver for NFSv4.1 pNFS (EXPERIMENTAL)" + depends on NFS_FS && NFS_V4_1 && SCSI_OSD_ULD + help + Say M here if you want your pNFS client to support the Objects Layout Driver. + Requires the SCSI osd initiator library (SCSI_OSD_INITIATOR) and + upper level driver (SCSI_OSD_ULD). + + If unsure, say N. + config ROOT_NFS bool "Root file system on NFS" depends on NFS_FS=y && IP_PNP diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index 4776ff9e3814..6a34f7dd0e6f 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -15,9 +15,11 @@ nfs-$(CONFIG_NFS_V4) += nfs4proc.o nfs4xdr.o nfs4state.o nfs4renewd.o \ delegation.o idmap.o \ callback.o callback_xdr.o callback_proc.o \ nfs4namespace.o -nfs-$(CONFIG_NFS_V4_1) += pnfs.o +nfs-$(CONFIG_NFS_V4_1) += pnfs.o pnfs_dev.o nfs-$(CONFIG_SYSCTL) += sysctl.o nfs-$(CONFIG_NFS_FSCACHE) += fscache.o fscache-index.o obj-$(CONFIG_PNFS_FILE_LAYOUT) += nfs_layout_nfsv41_files.o nfs_layout_nfsv41_files-y := nfs4filelayout.o nfs4filelayoutdev.o + +obj-$(CONFIG_PNFS_OBJLAYOUT) += objlayout/ diff --git a/fs/nfs/callback.h b/fs/nfs/callback.h index 46d93ce7311b..b257383bb565 100644 --- a/fs/nfs/callback.h +++ b/fs/nfs/callback.h @@ -167,6 +167,23 @@ extern unsigned nfs4_callback_layoutrecall( extern void nfs4_check_drain_bc_complete(struct nfs4_session *ses); extern void nfs4_cb_take_slot(struct nfs_client *clp); + +struct cb_devicenotifyitem { + uint32_t cbd_notify_type; + uint32_t cbd_layout_type; + struct nfs4_deviceid cbd_dev_id; + uint32_t cbd_immediate; +}; + +struct cb_devicenotifyargs { + int ndevs; + struct cb_devicenotifyitem *devs; +}; + +extern __be32 nfs4_callback_devicenotify( + struct cb_devicenotifyargs *args, + void *dummy, struct cb_process_state *cps); + #endif /* CONFIG_NFS_V4_1 */ extern int check_gss_callback_principal(struct nfs_client *, struct svc_rqst *); extern __be32 nfs4_callback_getattr(struct cb_getattrargs *args, diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c index 2f41dccea18e..d4d1954e9bb9 100644 --- a/fs/nfs/callback_proc.c +++ b/fs/nfs/callback_proc.c @@ -139,7 +139,7 @@ static u32 initiate_file_draining(struct nfs_client *clp, spin_lock(&ino->i_lock); if (test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags) || mark_matching_lsegs_invalid(lo, &free_me_list, - args->cbl_range.iomode)) + &args->cbl_range)) rv = NFS4ERR_DELAY; else rv = NFS4ERR_NOMATCHING_LAYOUT; @@ -184,7 +184,7 @@ static u32 initiate_bulk_draining(struct nfs_client *clp, ino = lo->plh_inode; spin_lock(&ino->i_lock); set_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags); - if (mark_matching_lsegs_invalid(lo, &free_me_list, range.iomode)) + if (mark_matching_lsegs_invalid(lo, &free_me_list, &range)) rv = NFS4ERR_DELAY; list_del_init(&lo->plh_bulk_recall); spin_unlock(&ino->i_lock); @@ -241,6 +241,53 @@ static void pnfs_recall_all_layouts(struct nfs_client *clp) do_callback_layoutrecall(clp, &args); } +__be32 nfs4_callback_devicenotify(struct cb_devicenotifyargs *args, + void *dummy, struct cb_process_state *cps) +{ + int i; + __be32 res = 0; + struct nfs_client *clp = cps->clp; + struct nfs_server *server = NULL; + + dprintk("%s: -->\n", __func__); + + if (!clp) { + res = cpu_to_be32(NFS4ERR_OP_NOT_IN_SESSION); + goto out; + } + + for (i = 0; i < args->ndevs; i++) { + struct cb_devicenotifyitem *dev = &args->devs[i]; + + if (!server || + server->pnfs_curr_ld->id != dev->cbd_layout_type) { + rcu_read_lock(); + list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) + if (server->pnfs_curr_ld && + server->pnfs_curr_ld->id == dev->cbd_layout_type) { + rcu_read_unlock(); + goto found; + } + rcu_read_unlock(); + dprintk("%s: layout type %u not found\n", + __func__, dev->cbd_layout_type); + continue; + } + + found: + if (dev->cbd_notify_type == NOTIFY_DEVICEID4_CHANGE) + dprintk("%s: NOTIFY_DEVICEID4_CHANGE not supported, " + "deleting instead\n", __func__); + nfs4_delete_deviceid(server->pnfs_curr_ld, clp, &dev->cbd_dev_id); + } + +out: + kfree(args->devs); + dprintk("%s: exit with status = %u\n", + __func__, be32_to_cpu(res)); + return res; +} + int nfs41_validate_delegation_stateid(struct nfs_delegation *delegation, const nfs4_stateid *stateid) { if (delegation == NULL) diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c index 00ecf62ce7c1..c6c86a77e043 100644 --- a/fs/nfs/callback_xdr.c +++ b/fs/nfs/callback_xdr.c @@ -25,6 +25,7 @@ #if defined(CONFIG_NFS_V4_1) #define CB_OP_LAYOUTRECALL_RES_MAXSZ (CB_OP_HDR_RES_MAXSZ) +#define CB_OP_DEVICENOTIFY_RES_MAXSZ (CB_OP_HDR_RES_MAXSZ) #define CB_OP_SEQUENCE_RES_MAXSZ (CB_OP_HDR_RES_MAXSZ + \ 4 + 1 + 3) #define CB_OP_RECALLANY_RES_MAXSZ (CB_OP_HDR_RES_MAXSZ) @@ -284,6 +285,93 @@ out: return status; } +static +__be32 decode_devicenotify_args(struct svc_rqst *rqstp, + struct xdr_stream *xdr, + struct cb_devicenotifyargs *args) +{ + __be32 *p; + __be32 status = 0; + u32 tmp; + int n, i; + args->ndevs = 0; + + /* Num of device notifications */ + p = read_buf(xdr, sizeof(uint32_t)); + if (unlikely(p == NULL)) { + status = htonl(NFS4ERR_BADXDR); + goto out; + } + n = ntohl(*p++); + if (n <= 0) + goto out; + + args->devs = kmalloc(n * sizeof(*args->devs), GFP_KERNEL); + if (!args->devs) { + status = htonl(NFS4ERR_DELAY); + goto out; + } + + /* Decode each dev notification */ + for (i = 0; i < n; i++) { + struct cb_devicenotifyitem *dev = &args->devs[i]; + + p = read_buf(xdr, (4 * sizeof(uint32_t)) + NFS4_DEVICEID4_SIZE); + if (unlikely(p == NULL)) { + status = htonl(NFS4ERR_BADXDR); + goto err; + } + + tmp = ntohl(*p++); /* bitmap size */ + if (tmp != 1) { + status = htonl(NFS4ERR_INVAL); + goto err; + } + dev->cbd_notify_type = ntohl(*p++); + if (dev->cbd_notify_type != NOTIFY_DEVICEID4_CHANGE && + dev->cbd_notify_type != NOTIFY_DEVICEID4_DELETE) { + status = htonl(NFS4ERR_INVAL); + goto err; + } + + tmp = ntohl(*p++); /* opaque size */ + if (((dev->cbd_notify_type == NOTIFY_DEVICEID4_CHANGE) && + (tmp != NFS4_DEVICEID4_SIZE + 8)) || + ((dev->cbd_notify_type == NOTIFY_DEVICEID4_DELETE) && + (tmp != NFS4_DEVICEID4_SIZE + 4))) { + status = htonl(NFS4ERR_INVAL); + goto err; + } + dev->cbd_layout_type = ntohl(*p++); + memcpy(dev->cbd_dev_id.data, p, NFS4_DEVICEID4_SIZE); + p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE); + + if (dev->cbd_layout_type == NOTIFY_DEVICEID4_CHANGE) { + p = read_buf(xdr, sizeof(uint32_t)); + if (unlikely(p == NULL)) { + status = htonl(NFS4ERR_BADXDR); + goto err; + } + dev->cbd_immediate = ntohl(*p++); + } else { + dev->cbd_immediate = 0; + } + + args->ndevs++; + + dprintk("%s: type %d layout 0x%x immediate %d\n", + __func__, dev->cbd_notify_type, dev->cbd_layout_type, + dev->cbd_immediate); + } +out: + dprintk("%s: status %d ndevs %d\n", + __func__, ntohl(status), args->ndevs); + return status; +err: + kfree(args->devs); + goto out; +} + static __be32 decode_sessionid(struct xdr_stream *xdr, struct nfs4_sessionid *sid) { @@ -639,10 +727,10 @@ preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op) case OP_CB_RECALL_ANY: case OP_CB_RECALL_SLOT: case OP_CB_LAYOUTRECALL: + case OP_CB_NOTIFY_DEVICEID: *op = &callback_ops[op_nr]; break; - case OP_CB_NOTIFY_DEVICEID: case OP_CB_NOTIFY: case OP_CB_PUSH_DELEG: case OP_CB_RECALLABLE_OBJ_AVAIL: @@ -849,6 +937,12 @@ static struct callback_op callback_ops[] = { (callback_decode_arg_t)decode_layoutrecall_args, .res_maxsize = CB_OP_LAYOUTRECALL_RES_MAXSZ, }, + [OP_CB_NOTIFY_DEVICEID] = { + .process_op = (callback_process_op_t)nfs4_callback_devicenotify, + .decode_args = + (callback_decode_arg_t)decode_devicenotify_args, + .res_maxsize = CB_OP_DEVICENOTIFY_RES_MAXSZ, + }, [OP_CB_SEQUENCE] = { .process_op = (callback_process_op_t)nfs4_callback_sequence, .decode_args = (callback_decode_arg_t)decode_cb_sequence_args, diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 139be9647d80..b3dc2b88b65b 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -290,6 +290,8 @@ static void nfs_free_client(struct nfs_client *clp) if (clp->cl_machine_cred != NULL) put_rpccred(clp->cl_machine_cred); + nfs4_deviceid_purge_client(clp); + kfree(clp->cl_hostname); kfree(clp); diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index bbbc6bf5cb2e..dd25c2aec375 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -21,25 +21,13 @@ #include "delegation.h" #include "internal.h" -static void nfs_do_free_delegation(struct nfs_delegation *delegation) -{ - kfree(delegation); -} - -static void nfs_free_delegation_callback(struct rcu_head *head) -{ - struct nfs_delegation *delegation = container_of(head, struct nfs_delegation, rcu); - - nfs_do_free_delegation(delegation); -} - static void nfs_free_delegation(struct nfs_delegation *delegation) { if (delegation->cred) { put_rpccred(delegation->cred); delegation->cred = NULL; } - call_rcu(&delegation->rcu, nfs_free_delegation_callback); + kfree_rcu(delegation, rcu); } /** diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 424e47773a84..ededdbd0db38 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -512,12 +512,7 @@ int nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *en struct page **xdr_pages, struct page *page, unsigned int buflen) { struct xdr_stream stream; - struct xdr_buf buf = { - .pages = xdr_pages, - .page_len = buflen, - .buflen = buflen, - .len = buflen, - }; + struct xdr_buf buf; struct page *scratch; struct nfs_cache_array *array; unsigned int count = 0; @@ -527,7 +522,7 @@ int nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *en if (scratch == NULL) return -ENOMEM; - xdr_init_decode(&stream, &buf, NULL); + xdr_init_decode_pages(&stream, &buf, xdr_pages, buflen); xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE); do { diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 57bb31ad7a5e..6f4850deb272 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -256,7 +256,8 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr) nfs_attr_check_mountpoint(sb, fattr); - if ((fattr->valid & NFS_ATTR_FATTR_FILEID) == 0 && (fattr->valid & NFS_ATTR_FATTR_MOUNTPOINT) == 0) + if (((fattr->valid & NFS_ATTR_FATTR_FILEID) == 0) && + !nfs_attr_use_mounted_on_fileid(fattr)) goto out_no_inode; if ((fattr->valid & NFS_ATTR_FATTR_TYPE) == 0) goto out_no_inode; @@ -1294,12 +1295,17 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) if (new_isize != cur_isize) { /* Do we perhaps have any outstanding writes, or has * the file grown beyond our last write? */ - if (nfsi->npages == 0 || new_isize > cur_isize) { + if ((nfsi->npages == 0 && !test_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags)) || + new_isize > cur_isize) { i_size_write(inode, new_isize); invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA; } - dprintk("NFS: isize change on server for file %s/%ld\n", - inode->i_sb->s_id, inode->i_ino); + dprintk("NFS: isize change on server for file %s/%ld " + "(%Ld to %Ld)\n", + inode->i_sb->s_id, + inode->i_ino, + (long long)cur_isize, + (long long)new_isize); } } else invalid |= save_cache_validity & (NFS_INO_INVALID_ATTR @@ -1424,9 +1430,10 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr) */ void nfs4_evict_inode(struct inode *inode) { - pnfs_destroy_layout(NFS_I(inode)); truncate_inode_pages(&inode->i_data, 0); end_writeback(inode); + pnfs_return_layout(inode); + pnfs_destroy_layout(NFS_I(inode)); /* If we are holding a delegation, return it! */ nfs_inode_return_delegation_noreclaim(inode); /* First call standard NFS clear_inode() code */ diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 2df6ca7b5898..2a55347a2daa 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -45,6 +45,17 @@ static inline void nfs_attr_check_mountpoint(struct super_block *parent, struct fattr->valid |= NFS_ATTR_FATTR_MOUNTPOINT; } +static inline int nfs_attr_use_mounted_on_fileid(struct nfs_fattr *fattr) +{ + if (((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) == 0) || + (((fattr->valid & NFS_ATTR_FATTR_MOUNTPOINT) == 0) && + ((fattr->valid & NFS_ATTR_FATTR_V4_REFERRAL) == 0))) + return 0; + + fattr->fileid = fattr->mounted_on_fileid; + return 1; +} + struct nfs_clone_mount { const struct super_block *sb; const struct dentry *dentry; @@ -310,6 +321,7 @@ extern int nfs_migrate_page(struct address_space *, #endif /* nfs4proc.c */ +extern void __nfs4_read_done_cb(struct nfs_read_data *); extern void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data); extern int nfs4_init_client(struct nfs_client *clp, const struct rpc_timeout *timeparms, diff --git a/fs/nfs/nfs4filelayout.c b/fs/nfs/nfs4filelayout.c index be79dc9f386d..0bafcc91c27f 100644 --- a/fs/nfs/nfs4filelayout.c +++ b/fs/nfs/nfs4filelayout.c @@ -30,6 +30,7 @@ */ #include <linux/nfs_fs.h> +#include <linux/nfs_page.h> #include "internal.h" #include "nfs4filelayout.h" @@ -421,6 +422,7 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, struct nfs4_deviceid *id, gfp_t gfp_flags) { + struct nfs4_deviceid_node *d; struct nfs4_file_layout_dsaddr *dsaddr; int status = -EINVAL; struct nfs_server *nfss = NFS_SERVER(lo->plh_inode); @@ -428,7 +430,7 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, dprintk("--> %s\n", __func__); if (fl->pattern_offset > lgr->range.offset) { - dprintk("%s pattern_offset %lld to large\n", + dprintk("%s pattern_offset %lld too large\n", __func__, fl->pattern_offset); goto out; } @@ -440,12 +442,14 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo, } /* find and reference the deviceid */ - dsaddr = nfs4_fl_find_get_deviceid(id); - if (dsaddr == NULL) { + d = nfs4_find_get_deviceid(NFS_SERVER(lo->plh_inode)->pnfs_curr_ld, + NFS_SERVER(lo->plh_inode)->nfs_client, id); + if (d == NULL) { dsaddr = get_device_info(lo->plh_inode, id, gfp_flags); if (dsaddr == NULL) goto out; - } + } else + dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node); fl->dsaddr = dsaddr; if (fl->first_stripe_index < 0 || @@ -507,12 +511,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo, gfp_t gfp_flags) { struct xdr_stream stream; - struct xdr_buf buf = { - .pages = lgr->layoutp->pages, - .page_len = lgr->layoutp->len, - .buflen = lgr->layoutp->len, - .len = lgr->layoutp->len, - }; + struct xdr_buf buf; struct page *scratch; __be32 *p; uint32_t nfl_util; @@ -524,7 +523,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo, if (!scratch) return -ENOMEM; - xdr_init_decode(&stream, &buf, NULL); + xdr_init_decode_pages(&stream, &buf, lgr->layoutp->pages, lgr->layoutp->len); xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE); /* 20 = ufl_util (4), first_stripe_index (4), pattern_offset (8), @@ -535,7 +534,7 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo, memcpy(id, p, sizeof(*id)); p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE); - print_deviceid(id); + nfs4_print_deviceid(id); nfl_util = be32_to_cpup(p++); if (nfl_util & NFL4_UFLG_COMMIT_THRU_MDS) @@ -554,13 +553,18 @@ filelayout_decode_layout(struct pnfs_layout_hdr *flo, __func__, nfl_util, fl->num_fh, fl->first_stripe_index, fl->pattern_offset); - if (!fl->num_fh) + /* Note that a zero value for num_fh is legal for STRIPE_SPARSE. + * Futher checking is done in filelayout_check_layout */ + if (fl->num_fh < 0 || fl->num_fh > + max(NFS4_PNFS_MAX_STRIPE_CNT, NFS4_PNFS_MAX_MULTI_CNT)) goto out_err; - fl->fh_array = kzalloc(fl->num_fh * sizeof(struct nfs_fh *), - gfp_flags); - if (!fl->fh_array) - goto out_err; + if (fl->num_fh > 0) { + fl->fh_array = kzalloc(fl->num_fh * sizeof(struct nfs_fh *), + gfp_flags); + if (!fl->fh_array) + goto out_err; + } for (i = 0; i < fl->num_fh; i++) { /* Do we want to use a mempool here? */ @@ -653,16 +657,20 @@ filelayout_alloc_lseg(struct pnfs_layout_hdr *layoutid, /* * filelayout_pg_test(). Called by nfs_can_coalesce_requests() * - * return 1 : coalesce page - * return 0 : don't coalesce page + * return true : coalesce page + * return false : don't coalesce page */ -int +bool filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req) { u64 p_stripe, r_stripe; u32 stripe_unit; + if (!pnfs_generic_pg_test(pgio, prev, req) || + !nfs_generic_pg_test(pgio, prev, req)) + return false; + if (!pgio->pg_lseg) return 1; p_stripe = (u64)prev->wb_index << PAGE_CACHE_SHIFT; @@ -860,6 +868,12 @@ filelayout_commit_pagelist(struct inode *inode, struct list_head *mds_pages, return -ENOMEM; } +static void +filelayout_free_deveiceid_node(struct nfs4_deviceid_node *d) +{ + nfs4_fl_free_deviceid(container_of(d, struct nfs4_file_layout_dsaddr, id_node)); +} + static struct pnfs_layoutdriver_type filelayout_type = { .id = LAYOUT_NFSV4_1_FILES, .name = "LAYOUT_NFSV4_1_FILES", @@ -872,6 +886,7 @@ static struct pnfs_layoutdriver_type filelayout_type = { .commit_pagelist = filelayout_commit_pagelist, .read_pagelist = filelayout_read_pagelist, .write_pagelist = filelayout_write_pagelist, + .free_deviceid_node = filelayout_free_deveiceid_node, }; static int __init nfs4filelayout_init(void) diff --git a/fs/nfs/nfs4filelayout.h b/fs/nfs/nfs4filelayout.h index 2b461d77b43a..cebe01e3795e 100644 --- a/fs/nfs/nfs4filelayout.h +++ b/fs/nfs/nfs4filelayout.h @@ -59,9 +59,7 @@ struct nfs4_pnfs_ds { #define NFS4_DEVICE_ID_NEG_ENTRY 0x00000001 struct nfs4_file_layout_dsaddr { - struct hlist_node node; - struct nfs4_deviceid deviceid; - atomic_t ref; + struct nfs4_deviceid_node id_node; unsigned long flags; u32 stripe_count; u8 *stripe_indices; @@ -95,14 +93,12 @@ extern struct nfs_fh * nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j); extern void print_ds(struct nfs4_pnfs_ds *ds); -extern void print_deviceid(struct nfs4_deviceid *dev_id); u32 nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset); u32 nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j); struct nfs4_pnfs_ds *nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx); -extern struct nfs4_file_layout_dsaddr * -nfs4_fl_find_get_deviceid(struct nfs4_deviceid *dev_id); extern void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr); +extern void nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr); struct nfs4_file_layout_dsaddr * get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id, gfp_t gfp_flags); diff --git a/fs/nfs/nfs4filelayoutdev.c b/fs/nfs/nfs4filelayoutdev.c index db07c7af1395..3b7bf1377264 100644 --- a/fs/nfs/nfs4filelayoutdev.c +++ b/fs/nfs/nfs4filelayoutdev.c @@ -37,30 +37,6 @@ #define NFSDBG_FACILITY NFSDBG_PNFS_LD /* - * Device ID RCU cache. A device ID is unique per client ID and layout type. - */ -#define NFS4_FL_DEVICE_ID_HASH_BITS 5 -#define NFS4_FL_DEVICE_ID_HASH_SIZE (1 << NFS4_FL_DEVICE_ID_HASH_BITS) -#define NFS4_FL_DEVICE_ID_HASH_MASK (NFS4_FL_DEVICE_ID_HASH_SIZE - 1) - -static inline u32 -nfs4_fl_deviceid_hash(struct nfs4_deviceid *id) -{ - unsigned char *cptr = (unsigned char *)id->data; - unsigned int nbytes = NFS4_DEVICEID4_SIZE; - u32 x = 0; - - while (nbytes--) { - x *= 37; - x += *cptr++; - } - return x & NFS4_FL_DEVICE_ID_HASH_MASK; -} - -static struct hlist_head filelayout_deviceid_cache[NFS4_FL_DEVICE_ID_HASH_SIZE]; -static DEFINE_SPINLOCK(filelayout_deviceid_lock); - -/* * Data server cache * * Data servers can be mapped to different device ids. @@ -89,27 +65,6 @@ print_ds(struct nfs4_pnfs_ds *ds) ds->ds_clp ? ds->ds_clp->cl_exchange_flags : 0); } -void -print_ds_list(struct nfs4_file_layout_dsaddr *dsaddr) -{ - int i; - - ifdebug(FACILITY) { - printk("%s dsaddr->ds_num %d\n", __func__, - dsaddr->ds_num); - for (i = 0; i < dsaddr->ds_num; i++) - print_ds(dsaddr->ds_list[i]); - } -} - -void print_deviceid(struct nfs4_deviceid *id) -{ - u32 *p = (u32 *)id; - - dprintk("%s: device id= [%x%x%x%x]\n", __func__, - p[0], p[1], p[2], p[3]); -} - /* nfs4_ds_cache_lock is held */ static struct nfs4_pnfs_ds * _data_server_lookup_locked(u32 ip_addr, u32 port) @@ -201,13 +156,13 @@ destroy_ds(struct nfs4_pnfs_ds *ds) kfree(ds); } -static void +void nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr) { struct nfs4_pnfs_ds *ds; int i; - print_deviceid(&dsaddr->deviceid); + nfs4_print_deviceid(&dsaddr->id_node.deviceid); for (i = 0; i < dsaddr->ds_num; i++) { ds = dsaddr->ds_list[i]; @@ -353,12 +308,7 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags) u8 max_stripe_index; struct nfs4_file_layout_dsaddr *dsaddr = NULL; struct xdr_stream stream; - struct xdr_buf buf = { - .pages = pdev->pages, - .page_len = pdev->pglen, - .buflen = pdev->pglen, - .len = pdev->pglen, - }; + struct xdr_buf buf; struct page *scratch; /* set up xdr stream */ @@ -366,7 +316,7 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags) if (!scratch) goto out_err; - xdr_init_decode(&stream, &buf, NULL); + xdr_init_decode_pages(&stream, &buf, pdev->pages, pdev->pglen); xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE); /* Get the stripe count (number of stripe index) */ @@ -431,8 +381,10 @@ decode_device(struct inode *ino, struct pnfs_device *pdev, gfp_t gfp_flags) dsaddr->stripe_indices = stripe_indices; stripe_indices = NULL; dsaddr->ds_num = num; - - memcpy(&dsaddr->deviceid, &pdev->dev_id, sizeof(pdev->dev_id)); + nfs4_init_deviceid_node(&dsaddr->id_node, + NFS_SERVER(ino)->pnfs_curr_ld, + NFS_SERVER(ino)->nfs_client, + &pdev->dev_id); for (i = 0; i < dsaddr->ds_num; i++) { int j; @@ -505,8 +457,8 @@ out_err: static struct nfs4_file_layout_dsaddr * decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_flags) { - struct nfs4_file_layout_dsaddr *d, *new; - long hash; + struct nfs4_deviceid_node *d; + struct nfs4_file_layout_dsaddr *n, *new; new = decode_device(inode, dev, gfp_flags); if (!new) { @@ -515,20 +467,13 @@ decode_and_add_device(struct inode *inode, struct pnfs_device *dev, gfp_t gfp_fl return NULL; } - spin_lock(&filelayout_deviceid_lock); - d = nfs4_fl_find_get_deviceid(&new->deviceid); - if (d) { - spin_unlock(&filelayout_deviceid_lock); + d = nfs4_insert_deviceid_node(&new->id_node); + n = container_of(d, struct nfs4_file_layout_dsaddr, id_node); + if (n != new) { nfs4_fl_free_deviceid(new); - return d; + return n; } - INIT_HLIST_NODE(&new->node); - atomic_set(&new->ref, 1); - hash = nfs4_fl_deviceid_hash(&new->deviceid); - hlist_add_head_rcu(&new->node, &filelayout_deviceid_cache[hash]); - spin_unlock(&filelayout_deviceid_lock); - return new; } @@ -600,35 +545,7 @@ out_free: void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr) { - if (atomic_dec_and_lock(&dsaddr->ref, &filelayout_deviceid_lock)) { - hlist_del_rcu(&dsaddr->node); - spin_unlock(&filelayout_deviceid_lock); - - synchronize_rcu(); - nfs4_fl_free_deviceid(dsaddr); - } -} - -struct nfs4_file_layout_dsaddr * -nfs4_fl_find_get_deviceid(struct nfs4_deviceid *id) -{ - struct nfs4_file_layout_dsaddr *d; - struct hlist_node *n; - long hash = nfs4_fl_deviceid_hash(id); - - - rcu_read_lock(); - hlist_for_each_entry_rcu(d, n, &filelayout_deviceid_cache[hash], node) { - if (!memcmp(&d->deviceid, id, sizeof(*id))) { - if (!atomic_inc_not_zero(&d->ref)) - goto fail; - rcu_read_unlock(); - return d; - } - } -fail: - rcu_read_unlock(); - return NULL; + nfs4_put_deviceid_node(&dsaddr->id_node); } /* @@ -676,15 +593,15 @@ static void filelayout_mark_devid_negative(struct nfs4_file_layout_dsaddr *dsaddr, int err, u32 ds_addr) { - u32 *p = (u32 *)&dsaddr->deviceid; + u32 *p = (u32 *)&dsaddr->id_node.deviceid; printk(KERN_ERR "NFS: data server %x connection error %d." " Deviceid [%x%x%x%x] marked out of use.\n", ds_addr, err, p[0], p[1], p[2], p[3]); - spin_lock(&filelayout_deviceid_lock); + spin_lock(&nfs4_ds_cache_lock); dsaddr->flags |= NFS4_DEVICE_ID_NEG_ENTRY; - spin_unlock(&filelayout_deviceid_lock); + spin_unlock(&nfs4_ds_cache_lock); } struct nfs4_pnfs_ds * diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index cf1b339c3937..5879b23e0c99 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -267,9 +267,11 @@ static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struc break; nfs4_schedule_stateid_recovery(server, state); goto wait_on_recovery; + case -NFS4ERR_EXPIRED: + if (state != NULL) + nfs4_schedule_stateid_recovery(server, state); case -NFS4ERR_STALE_STATEID: case -NFS4ERR_STALE_CLIENTID: - case -NFS4ERR_EXPIRED: nfs4_schedule_lease_recovery(clp); goto wait_on_recovery; #if defined(CONFIG_NFS_V4_1) @@ -2263,12 +2265,14 @@ static int nfs4_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle, return nfs4_map_errors(status); } +static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr); /* * Get locations and (maybe) other attributes of a referral. * Note that we'll actually follow the referral later when * we detect fsid mismatch in inode revalidation */ -static int nfs4_get_referral(struct inode *dir, const struct qstr *name, struct nfs_fattr *fattr, struct nfs_fh *fhandle) +static int nfs4_get_referral(struct inode *dir, const struct qstr *name, + struct nfs_fattr *fattr, struct nfs_fh *fhandle) { int status = -ENOMEM; struct page *page = NULL; @@ -2286,15 +2290,16 @@ static int nfs4_get_referral(struct inode *dir, const struct qstr *name, struct goto out; /* Make sure server returned a different fsid for the referral */ if (nfs_fsid_equal(&NFS_SERVER(dir)->fsid, &locations->fattr.fsid)) { - dprintk("%s: server did not return a different fsid for a referral at %s\n", __func__, name->name); + dprintk("%s: server did not return a different fsid for" + " a referral at %s\n", __func__, name->name); status = -EIO; goto out; } + /* Fixup attributes for the nfs_lookup() call to nfs_fhget() */ + nfs_fixup_referral_attributes(&locations->fattr); + /* replace the lookup nfs_fattr with the locations nfs_fattr */ memcpy(fattr, &locations->fattr, sizeof(struct nfs_fattr)); - fattr->valid |= NFS_ATTR_FATTR_V4_REFERRAL; - if (!fattr->mode) - fattr->mode = S_IFDIR; memset(fhandle, 0, sizeof(struct nfs_fh)); out: if (page) @@ -2361,6 +2366,9 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr, struct nfs4_state *state = NULL; int status; + if (pnfs_ld_layoutret_on_setattr(inode)) + pnfs_return_layout(inode); + nfs_fattr_init(fattr); /* Search for an existing open(O_WRITE) file */ @@ -3175,6 +3183,11 @@ static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle, return err; } +void __nfs4_read_done_cb(struct nfs_read_data *data) +{ + nfs_invalidate_atime(data->inode); +} + static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data) { struct nfs_server *server = NFS_SERVER(data->inode); @@ -3184,7 +3197,7 @@ static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data) return -EAGAIN; } - nfs_invalidate_atime(data->inode); + __nfs4_read_done_cb(data); if (task->tk_status > 0) renew_lease(server, data->timestamp); return 0; @@ -3198,7 +3211,8 @@ static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data) if (!nfs4_sequence_done(task, &data->res.seq_res)) return -EAGAIN; - return data->read_done_cb(task, data); + return data->read_done_cb ? data->read_done_cb(task, data) : + nfs4_read_done_cb(task, data); } static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg) @@ -3243,7 +3257,8 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data) { if (!nfs4_sequence_done(task, &data->res.seq_res)) return -EAGAIN; - return data->write_done_cb(task, data); + return data->write_done_cb ? data->write_done_cb(task, data) : + nfs4_write_done_cb(task, data); } /* Reset the the nfs_write_data to send the write to the MDS. */ @@ -3670,9 +3685,11 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server, break; nfs4_schedule_stateid_recovery(server, state); goto wait_on_recovery; + case -NFS4ERR_EXPIRED: + if (state != NULL) + nfs4_schedule_stateid_recovery(server, state); case -NFS4ERR_STALE_STATEID: case -NFS4ERR_STALE_CLIENTID: - case -NFS4ERR_EXPIRED: nfs4_schedule_lease_recovery(clp); goto wait_on_recovery; #if defined(CONFIG_NFS_V4_1) @@ -4543,6 +4560,7 @@ int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl) case -ESTALE: goto out; case -NFS4ERR_EXPIRED: + nfs4_schedule_stateid_recovery(server, state); case -NFS4ERR_STALE_CLIENTID: case -NFS4ERR_STALE_STATEID: nfs4_schedule_lease_recovery(server->nfs_client); @@ -4652,11 +4670,15 @@ static size_t nfs4_xattr_list_nfs4_acl(struct dentry *dentry, char *list, return len; } +/* + * nfs_fhget will use either the mounted_on_fileid or the fileid + */ static void nfs_fixup_referral_attributes(struct nfs_fattr *fattr) { - if (!((fattr->valid & NFS_ATTR_FATTR_FILEID) && - (fattr->valid & NFS_ATTR_FATTR_FSID) && - (fattr->valid & NFS_ATTR_FATTR_V4_REFERRAL))) + if (!(((fattr->valid & NFS_ATTR_FATTR_MOUNTED_ON_FILEID) || + (fattr->valid & NFS_ATTR_FATTR_FILEID)) && + (fattr->valid & NFS_ATTR_FATTR_FSID) && + (fattr->valid & NFS_ATTR_FATTR_V4_REFERRAL))) return; fattr->valid |= NFS_ATTR_FATTR_TYPE | NFS_ATTR_FATTR_MODE | @@ -4671,7 +4693,6 @@ int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name, struct nfs_server *server = NFS_SERVER(dir); u32 bitmask[2] = { [0] = FATTR4_WORD0_FSID | FATTR4_WORD0_FS_LOCATIONS, - [1] = FATTR4_WORD1_MOUNTED_ON_FILEID, }; struct nfs4_fs_locations_arg args = { .dir_fh = NFS_FH(dir), @@ -4690,11 +4711,18 @@ int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name, int status; dprintk("%s: start\n", __func__); + + /* Ask for the fileid of the absent filesystem if mounted_on_fileid + * is not supported */ + if (NFS_SERVER(dir)->attr_bitmask[1] & FATTR4_WORD1_MOUNTED_ON_FILEID) + bitmask[1] |= FATTR4_WORD1_MOUNTED_ON_FILEID; + else + bitmask[0] |= FATTR4_WORD0_FILEID; + nfs_fattr_init(&fs_locations->fattr); fs_locations->server = server; fs_locations->nlocations = 0; status = nfs4_call_sync(server->client, server, &msg, &args.seq_args, &res.seq_res, 0); - nfs_fixup_referral_attributes(&fs_locations->fattr); dprintk("%s: returned status = %d\n", __func__, status); return status; } @@ -5083,7 +5111,6 @@ static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args) if (mxresp_sz == 0) mxresp_sz = NFS_MAX_FILE_IO_SIZE; /* Fore channel attributes */ - args->fc_attrs.headerpadsz = 0; args->fc_attrs.max_rqst_sz = mxrqst_sz; args->fc_attrs.max_resp_sz = mxresp_sz; args->fc_attrs.max_ops = NFS4_MAX_OPS; @@ -5096,7 +5123,6 @@ static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args) args->fc_attrs.max_ops, args->fc_attrs.max_reqs); /* Back channel attributes */ - args->bc_attrs.headerpadsz = 0; args->bc_attrs.max_rqst_sz = PAGE_SIZE; args->bc_attrs.max_resp_sz = PAGE_SIZE; args->bc_attrs.max_resp_sz_cached = 0; @@ -5116,8 +5142,6 @@ static int nfs4_verify_fore_channel_attrs(struct nfs41_create_session_args *args struct nfs4_channel_attrs *sent = &args->fc_attrs; struct nfs4_channel_attrs *rcvd = &session->fc_attrs; - if (rcvd->headerpadsz > sent->headerpadsz) - return -EINVAL; if (rcvd->max_resp_sz > sent->max_resp_sz) return -EINVAL; /* @@ -5666,6 +5690,88 @@ int nfs4_proc_layoutget(struct nfs4_layoutget *lgp) return status; } +static void +nfs4_layoutreturn_prepare(struct rpc_task *task, void *calldata) +{ + struct nfs4_layoutreturn *lrp = calldata; + + dprintk("--> %s\n", __func__); + if (nfs41_setup_sequence(lrp->clp->cl_session, &lrp->args.seq_args, + &lrp->res.seq_res, 0, task)) + return; + rpc_call_start(task); +} + +static void nfs4_layoutreturn_done(struct rpc_task *task, void *calldata) +{ + struct nfs4_layoutreturn *lrp = calldata; + struct nfs_server *server; + struct pnfs_layout_hdr *lo = NFS_I(lrp->args.inode)->layout; + + dprintk("--> %s\n", __func__); + + if (!nfs4_sequence_done(task, &lrp->res.seq_res)) + return; + + server = NFS_SERVER(lrp->args.inode); + if (nfs4_async_handle_error(task, server, NULL) == -EAGAIN) { + nfs_restart_rpc(task, lrp->clp); + return; + } + spin_lock(&lo->plh_inode->i_lock); + if (task->tk_status == 0) { + if (lrp->res.lrs_present) { + pnfs_set_layout_stateid(lo, &lrp->res.stateid, true); + } else + BUG_ON(!list_empty(&lo->plh_segs)); + } + lo->plh_block_lgets--; + spin_unlock(&lo->plh_inode->i_lock); + dprintk("<-- %s\n", __func__); +} + +static void nfs4_layoutreturn_release(void *calldata) +{ + struct nfs4_layoutreturn *lrp = calldata; + + dprintk("--> %s\n", __func__); + put_layout_hdr(NFS_I(lrp->args.inode)->layout); + kfree(calldata); + dprintk("<-- %s\n", __func__); +} + +static const struct rpc_call_ops nfs4_layoutreturn_call_ops = { + .rpc_call_prepare = nfs4_layoutreturn_prepare, + .rpc_call_done = nfs4_layoutreturn_done, + .rpc_release = nfs4_layoutreturn_release, +}; + +int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp) +{ + struct rpc_task *task; + struct rpc_message msg = { + .rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LAYOUTRETURN], + .rpc_argp = &lrp->args, + .rpc_resp = &lrp->res, + }; + struct rpc_task_setup task_setup_data = { + .rpc_client = lrp->clp->cl_rpcclient, + .rpc_message = &msg, + .callback_ops = &nfs4_layoutreturn_call_ops, + .callback_data = lrp, + }; + int status; + + dprintk("--> %s\n", __func__); + task = rpc_run_task(&task_setup_data); + if (IS_ERR(task)) + return PTR_ERR(task); + status = task->tk_status; + dprintk("<-- %s status=%d\n", __func__, status); + rpc_put_task(task); + return status; +} + static int _nfs4_proc_getdeviceinfo(struct nfs_server *server, struct pnfs_device *pdev) { diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 036f5adc9e1f..e97dd219f84f 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -1466,7 +1466,10 @@ static int nfs4_reclaim_lease(struct nfs_client *clp) #ifdef CONFIG_NFS_V4_1 void nfs4_schedule_session_recovery(struct nfs4_session *session) { - nfs4_schedule_lease_recovery(session->clp); + struct nfs_client *clp = session->clp; + + set_bit(NFS4CLNT_SESSION_RESET, &clp->cl_state); + nfs4_schedule_lease_recovery(clp); } EXPORT_SYMBOL_GPL(nfs4_schedule_session_recovery); @@ -1549,6 +1552,7 @@ static int nfs4_reset_session(struct nfs_client *clp) status = nfs4_recovery_handle_error(clp, status); goto out; } + clear_bit(NFS4CLNT_SESSION_RESET, &clp->cl_state); /* create_session negotiated new slot table */ clear_bit(NFS4CLNT_RECALL_SLOT, &clp->cl_state); diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index c3ccd2c46834..6870bc61ceec 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -255,7 +255,7 @@ static int nfs4_stat_to_errno(int); #define decode_fs_locations_maxsz \ (0) #define encode_secinfo_maxsz (op_encode_hdr_maxsz + nfs4_name_maxsz) -#define decode_secinfo_maxsz (op_decode_hdr_maxsz + 4 + (NFS_MAX_SECFLAVORS * (16 + GSS_OID_MAX_LEN))) +#define decode_secinfo_maxsz (op_decode_hdr_maxsz + 1 + ((NFS_MAX_SECFLAVORS * (16 + GSS_OID_MAX_LEN)) / 4)) #if defined(CONFIG_NFS_V4_1) #define NFS4_MAX_MACHINE_NAME_LEN (64) @@ -338,7 +338,11 @@ static int nfs4_stat_to_errno(int); 1 /* layoutupdate4 layout type */ + \ 1 /* NULL filelayout layoutupdate4 payload */) #define decode_layoutcommit_maxsz (op_decode_hdr_maxsz + 3) - +#define encode_layoutreturn_maxsz (8 + op_encode_hdr_maxsz + \ + encode_stateid_maxsz + \ + 1 /* FIXME: opaque lrf_body always empty at the moment */) +#define decode_layoutreturn_maxsz (op_decode_hdr_maxsz + \ + 1 + decode_stateid_maxsz) #else /* CONFIG_NFS_V4_1 */ #define encode_sequence_maxsz 0 #define decode_sequence_maxsz 0 @@ -760,7 +764,14 @@ static int nfs4_stat_to_errno(int); decode_putfh_maxsz + \ decode_layoutcommit_maxsz + \ decode_getattr_maxsz) - +#define NFS4_enc_layoutreturn_sz (compound_encode_hdr_maxsz + \ + encode_sequence_maxsz + \ + encode_putfh_maxsz + \ + encode_layoutreturn_maxsz) +#define NFS4_dec_layoutreturn_sz (compound_decode_hdr_maxsz + \ + decode_sequence_maxsz + \ + decode_putfh_maxsz + \ + decode_layoutreturn_maxsz) const u32 nfs41_maxwrite_overhead = ((RPC_MAX_HEADER_WITH_AUTH + compound_encode_hdr_maxsz + @@ -1714,7 +1725,7 @@ static void encode_create_session(struct xdr_stream *xdr, *p++ = cpu_to_be32(args->flags); /*flags */ /* Fore Channel */ - *p++ = cpu_to_be32(args->fc_attrs.headerpadsz); /* header padding size */ + *p++ = cpu_to_be32(0); /* header padding size */ *p++ = cpu_to_be32(args->fc_attrs.max_rqst_sz); /* max req size */ *p++ = cpu_to_be32(args->fc_attrs.max_resp_sz); /* max resp size */ *p++ = cpu_to_be32(max_resp_sz_cached); /* Max resp sz cached */ @@ -1723,7 +1734,7 @@ static void encode_create_session(struct xdr_stream *xdr, *p++ = cpu_to_be32(0); /* rdmachannel_attrs */ /* Back Channel */ - *p++ = cpu_to_be32(args->fc_attrs.headerpadsz); /* header padding size */ + *p++ = cpu_to_be32(0); /* header padding size */ *p++ = cpu_to_be32(args->bc_attrs.max_rqst_sz); /* max req size */ *p++ = cpu_to_be32(args->bc_attrs.max_resp_sz); /* max resp size */ *p++ = cpu_to_be32(args->bc_attrs.max_resp_sz_cached); /* Max resp sz cached */ @@ -1864,6 +1875,7 @@ encode_layoutget(struct xdr_stream *xdr, static int encode_layoutcommit(struct xdr_stream *xdr, + struct inode *inode, const struct nfs4_layoutcommit_args *args, struct compound_hdr *hdr) { @@ -1872,7 +1884,7 @@ encode_layoutcommit(struct xdr_stream *xdr, dprintk("%s: lbw: %llu type: %d\n", __func__, args->lastbytewritten, NFS_SERVER(args->inode)->pnfs_curr_ld->id); - p = reserve_space(xdr, 48 + NFS4_STATEID_SIZE); + p = reserve_space(xdr, 44 + NFS4_STATEID_SIZE); *p++ = cpu_to_be32(OP_LAYOUTCOMMIT); /* Only whole file layouts */ p = xdr_encode_hyper(p, 0); /* offset */ @@ -1883,12 +1895,49 @@ encode_layoutcommit(struct xdr_stream *xdr, p = xdr_encode_hyper(p, args->lastbytewritten); *p++ = cpu_to_be32(0); /* Never send time_modify_changed */ *p++ = cpu_to_be32(NFS_SERVER(args->inode)->pnfs_curr_ld->id);/* type */ - *p++ = cpu_to_be32(0); /* no file layout payload */ + + if (NFS_SERVER(inode)->pnfs_curr_ld->encode_layoutcommit) + NFS_SERVER(inode)->pnfs_curr_ld->encode_layoutcommit( + NFS_I(inode)->layout, xdr, args); + else { + p = reserve_space(xdr, 4); + *p = cpu_to_be32(0); /* no layout-type payload */ + } hdr->nops++; hdr->replen += decode_layoutcommit_maxsz; return 0; } + +static void +encode_layoutreturn(struct xdr_stream *xdr, + const struct nfs4_layoutreturn_args *args, + struct compound_hdr *hdr) +{ + __be32 *p; + + p = reserve_space(xdr, 20); + *p++ = cpu_to_be32(OP_LAYOUTRETURN); + *p++ = cpu_to_be32(0); /* reclaim. always 0 for now */ + *p++ = cpu_to_be32(args->layout_type); + *p++ = cpu_to_be32(IOMODE_ANY); + *p = cpu_to_be32(RETURN_FILE); + p = reserve_space(xdr, 16 + NFS4_STATEID_SIZE); + p = xdr_encode_hyper(p, 0); + p = xdr_encode_hyper(p, NFS4_MAX_UINT64); + spin_lock(&args->inode->i_lock); + xdr_encode_opaque_fixed(p, &args->stateid.data, NFS4_STATEID_SIZE); + spin_unlock(&args->inode->i_lock); + if (NFS_SERVER(args->inode)->pnfs_curr_ld->encode_layoutreturn) { + NFS_SERVER(args->inode)->pnfs_curr_ld->encode_layoutreturn( + NFS_I(args->inode)->layout, xdr, args); + } else { + p = reserve_space(xdr, 4); + *p = cpu_to_be32(0); + } + hdr->nops++; + hdr->replen += decode_layoutreturn_maxsz; +} #endif /* CONFIG_NFS_V4_1 */ /* @@ -2706,10 +2755,12 @@ static void nfs4_xdr_enc_layoutget(struct rpc_rqst *req, /* * Encode LAYOUTCOMMIT request */ -static int nfs4_xdr_enc_layoutcommit(struct rpc_rqst *req, - struct xdr_stream *xdr, - struct nfs4_layoutcommit_args *args) +static void nfs4_xdr_enc_layoutcommit(struct rpc_rqst *req, + struct xdr_stream *xdr, + struct nfs4_layoutcommit_args *args) { + struct nfs4_layoutcommit_data *data = + container_of(args, struct nfs4_layoutcommit_data, args); struct compound_hdr hdr = { .minorversion = nfs4_xdr_minorversion(&args->seq_args), }; @@ -2717,10 +2768,27 @@ static int nfs4_xdr_enc_layoutcommit(struct rpc_rqst *req, encode_compound_hdr(xdr, req, &hdr); encode_sequence(xdr, &args->seq_args, &hdr); encode_putfh(xdr, NFS_FH(args->inode), &hdr); - encode_layoutcommit(xdr, args, &hdr); + encode_layoutcommit(xdr, data->args.inode, args, &hdr); encode_getfattr(xdr, args->bitmask, &hdr); encode_nops(&hdr); - return 0; +} + +/* + * Encode LAYOUTRETURN request + */ +static void nfs4_xdr_enc_layoutreturn(struct rpc_rqst *req, + struct xdr_stream *xdr, + struct nfs4_layoutreturn_args *args) +{ + struct compound_hdr hdr = { + .minorversion = nfs4_xdr_minorversion(&args->seq_args), + }; + + encode_compound_hdr(xdr, req, &hdr); + encode_sequence(xdr, &args->seq_args, &hdr); + encode_putfh(xdr, NFS_FH(args->inode), &hdr); + encode_layoutreturn(xdr, args, &hdr); + encode_nops(&hdr); } #endif /* CONFIG_NFS_V4_1 */ @@ -3030,7 +3098,7 @@ out_overflow: return -EIO; } -static int decode_attr_error(struct xdr_stream *xdr, uint32_t *bitmap) +static int decode_attr_error(struct xdr_stream *xdr, uint32_t *bitmap, int32_t *res) { __be32 *p; @@ -3041,7 +3109,7 @@ static int decode_attr_error(struct xdr_stream *xdr, uint32_t *bitmap) if (unlikely(!p)) goto out_overflow; bitmap[0] &= ~FATTR4_WORD0_RDATTR_ERROR; - return -be32_to_cpup(p); + *res = -be32_to_cpup(p); } return 0; out_overflow: @@ -4002,6 +4070,7 @@ static int decode_getfattr_attrs(struct xdr_stream *xdr, uint32_t *bitmap, int status; umode_t fmode = 0; uint32_t type; + int32_t err; status = decode_attr_type(xdr, bitmap, &type); if (status < 0) @@ -4027,13 +4096,12 @@ static int decode_getfattr_attrs(struct xdr_stream *xdr, uint32_t *bitmap, goto xdr_error; fattr->valid |= status; - status = decode_attr_error(xdr, bitmap); - if (status == -NFS4ERR_WRONGSEC) { - nfs_fixup_secinfo_attributes(fattr, fh); - status = 0; - } + err = 0; + status = decode_attr_error(xdr, bitmap, &err); if (status < 0) goto xdr_error; + if (err == -NFS4ERR_WRONGSEC) + nfs_fixup_secinfo_attributes(fattr, fh); status = decode_attr_filehandle(xdr, bitmap, fh); if (status < 0) @@ -4929,12 +4997,14 @@ static int decode_chan_attrs(struct xdr_stream *xdr, struct nfs4_channel_attrs *attrs) { __be32 *p; - u32 nr_attrs; + u32 nr_attrs, val; p = xdr_inline_decode(xdr, 28); if (unlikely(!p)) goto out_overflow; - attrs->headerpadsz = be32_to_cpup(p++); + val = be32_to_cpup(p++); /* headerpadsz */ + if (val) + return -EINVAL; /* no support for header padding yet */ attrs->max_rqst_sz = be32_to_cpup(p++); attrs->max_resp_sz = be32_to_cpup(p++); attrs->max_resp_sz_cached = be32_to_cpup(p++); @@ -5203,6 +5273,27 @@ out_overflow: return -EIO; } +static int decode_layoutreturn(struct xdr_stream *xdr, + struct nfs4_layoutreturn_res *res) +{ + __be32 *p; + int status; + + status = decode_op_hdr(xdr, OP_LAYOUTRETURN); + if (status) + return status; + p = xdr_inline_decode(xdr, 4); + if (unlikely(!p)) + goto out_overflow; + res->lrs_present = be32_to_cpup(p); + if (res->lrs_present) + status = decode_stateid(xdr, &res->stateid); + return status; +out_overflow: + print_overflow_msg(__func__, xdr); + return -EIO; +} + static int decode_layoutcommit(struct xdr_stream *xdr, struct rpc_rqst *req, struct nfs4_layoutcommit_res *res) @@ -6320,6 +6411,30 @@ out: } /* + * Decode LAYOUTRETURN response + */ +static int nfs4_xdr_dec_layoutreturn(struct rpc_rqst *rqstp, + struct xdr_stream *xdr, + struct nfs4_layoutreturn_res *res) +{ + struct compound_hdr hdr; + int status; + + status = decode_compound_hdr(xdr, &hdr); + if (status) + goto out; + status = decode_sequence(xdr, &res->seq_res, rqstp); + if (status) + goto out; + status = decode_putfh(xdr); + if (status) + goto out; + status = decode_layoutreturn(xdr, res); +out: + return status; +} + +/* * Decode LAYOUTCOMMIT response */ static int nfs4_xdr_dec_layoutcommit(struct rpc_rqst *rqstp, @@ -6547,6 +6662,7 @@ struct rpc_procinfo nfs4_procedures[] = { PROC(GETDEVICEINFO, enc_getdeviceinfo, dec_getdeviceinfo), PROC(LAYOUTGET, enc_layoutget, dec_layoutget), PROC(LAYOUTCOMMIT, enc_layoutcommit, dec_layoutcommit), + PROC(LAYOUTRETURN, enc_layoutreturn, dec_layoutreturn), #endif /* CONFIG_NFS_V4_1 */ }; diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c index c541093a5bf2..c4744e1d513c 100644 --- a/fs/nfs/nfsroot.c +++ b/fs/nfs/nfsroot.c @@ -87,7 +87,7 @@ #define NFS_ROOT "/tftpboot/%s" /* Default NFSROOT mount options. */ -#define NFS_DEF_OPTIONS "udp" +#define NFS_DEF_OPTIONS "vers=2,udp,rsize=4096,wsize=4096" /* Parameters passed from the kernel command line */ static char nfs_root_parms[256] __initdata = ""; diff --git a/fs/nfs/objlayout/Kbuild b/fs/nfs/objlayout/Kbuild new file mode 100644 index 000000000000..ed30ea072bb8 --- /dev/null +++ b/fs/nfs/objlayout/Kbuild @@ -0,0 +1,5 @@ +# +# Makefile for the pNFS Objects Layout Driver kernel module +# +objlayoutdriver-y := objio_osd.o pnfs_osd_xdr_cli.o objlayout.o +obj-$(CONFIG_PNFS_OBJLAYOUT) += objlayoutdriver.o diff --git a/fs/nfs/objlayout/objio_osd.c b/fs/nfs/objlayout/objio_osd.c new file mode 100644 index 000000000000..8ff2ea3f10ef --- /dev/null +++ b/fs/nfs/objlayout/objio_osd.c @@ -0,0 +1,1059 @@ +/* + * pNFS Objects layout implementation over open-osd initiator library + * + * Copyright (C) 2009 Panasas Inc. [year of first publication] + * All rights reserved. + * + * Benny Halevy <bhalevy@panasas.com> + * Boaz Harrosh <bharrosh@panasas.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 + * See the file COPYING included with this distribution for more details. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the Panasas company nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED + * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE + * DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR + * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF + * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING + * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS + * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include <linux/module.h> +#include <scsi/osd_initiator.h> + +#include "objlayout.h" + +#define NFSDBG_FACILITY NFSDBG_PNFS_LD + +#define _LLU(x) ((unsigned long long)x) + +enum { BIO_MAX_PAGES_KMALLOC = + (PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec), +}; + +struct objio_dev_ent { + struct nfs4_deviceid_node id_node; + struct osd_dev *od; +}; + +static void +objio_free_deviceid_node(struct nfs4_deviceid_node *d) +{ + struct objio_dev_ent *de = container_of(d, struct objio_dev_ent, id_node); + + dprintk("%s: free od=%p\n", __func__, de->od); + osduld_put_device(de->od); + kfree(de); +} + +static struct objio_dev_ent *_dev_list_find(const struct nfs_server *nfss, + const struct nfs4_deviceid *d_id) +{ + struct nfs4_deviceid_node *d; + struct objio_dev_ent *de; + + d = nfs4_find_get_deviceid(nfss->pnfs_curr_ld, nfss->nfs_client, d_id); + if (!d) + return NULL; + + de = container_of(d, struct objio_dev_ent, id_node); + return de; +} + +static struct objio_dev_ent * +_dev_list_add(const struct nfs_server *nfss, + const struct nfs4_deviceid *d_id, struct osd_dev *od, + gfp_t gfp_flags) +{ + struct nfs4_deviceid_node *d; + struct objio_dev_ent *de = kzalloc(sizeof(*de), gfp_flags); + struct objio_dev_ent *n; + + if (!de) { + dprintk("%s: -ENOMEM od=%p\n", __func__, od); + return NULL; + } + + dprintk("%s: Adding od=%p\n", __func__, od); + nfs4_init_deviceid_node(&de->id_node, + nfss->pnfs_curr_ld, + nfss->nfs_client, + d_id); + de->od = od; + + d = nfs4_insert_deviceid_node(&de->id_node); + n = container_of(d, struct objio_dev_ent, id_node); + if (n != de) { + dprintk("%s: Race with other n->od=%p\n", __func__, n->od); + objio_free_deviceid_node(&de->id_node); + de = n; + } + + return de; +} + +struct caps_buffers { + u8 caps_key[OSD_CRYPTO_KEYID_SIZE]; + u8 creds[OSD_CAP_LEN]; +}; + +struct objio_segment { + struct pnfs_layout_segment lseg; + + struct pnfs_osd_object_cred *comps; + + unsigned mirrors_p1; + unsigned stripe_unit; + unsigned group_width; /* Data stripe_units without integrity comps */ + u64 group_depth; + unsigned group_count; + + unsigned max_io_size; + + unsigned comps_index; + unsigned num_comps; + /* variable length */ + struct objio_dev_ent *ods[]; +}; + +static inline struct objio_segment * +OBJIO_LSEG(struct pnfs_layout_segment *lseg) +{ + return container_of(lseg, struct objio_segment, lseg); +} + +struct objio_state; +typedef ssize_t (*objio_done_fn)(struct objio_state *ios); + +struct objio_state { + /* Generic layer */ + struct objlayout_io_state ol_state; + + struct objio_segment *layout; + + struct kref kref; + objio_done_fn done; + void *private; + + unsigned long length; + unsigned numdevs; /* Actually used devs in this IO */ + /* A per-device variable array of size numdevs */ + struct _objio_per_comp { + struct bio *bio; + struct osd_request *or; + unsigned long length; + u64 offset; + unsigned dev; + } per_dev[]; +}; + +/* Send and wait for a get_device_info of devices in the layout, + then look them up with the osd_initiator library */ +static struct objio_dev_ent *_device_lookup(struct pnfs_layout_hdr *pnfslay, + struct objio_segment *objio_seg, unsigned comp, + gfp_t gfp_flags) +{ + struct pnfs_osd_deviceaddr *deviceaddr; + struct nfs4_deviceid *d_id; + struct objio_dev_ent *ode; + struct osd_dev *od; + struct osd_dev_info odi; + int err; + + d_id = &objio_seg->comps[comp].oc_object_id.oid_device_id; + + ode = _dev_list_find(NFS_SERVER(pnfslay->plh_inode), d_id); + if (ode) + return ode; + + err = objlayout_get_deviceinfo(pnfslay, d_id, &deviceaddr, gfp_flags); + if (unlikely(err)) { + dprintk("%s: objlayout_get_deviceinfo dev(%llx:%llx) =>%d\n", + __func__, _DEVID_LO(d_id), _DEVID_HI(d_id), err); + return ERR_PTR(err); + } + + odi.systemid_len = deviceaddr->oda_systemid.len; + if (odi.systemid_len > sizeof(odi.systemid)) { + err = -EINVAL; + goto out; + } else if (odi.systemid_len) + memcpy(odi.systemid, deviceaddr->oda_systemid.data, + odi.systemid_len); + odi.osdname_len = deviceaddr->oda_osdname.len; + odi.osdname = (u8 *)deviceaddr->oda_osdname.data; + + if (!odi.osdname_len && !odi.systemid_len) { + dprintk("%s: !odi.osdname_len && !odi.systemid_len\n", + __func__); + err = -ENODEV; + goto out; + } + + od = osduld_info_lookup(&odi); + if (unlikely(IS_ERR(od))) { + err = PTR_ERR(od); + dprintk("%s: osduld_info_lookup => %d\n", __func__, err); + goto out; + } + + ode = _dev_list_add(NFS_SERVER(pnfslay->plh_inode), d_id, od, + gfp_flags); + +out: + dprintk("%s: return=%d\n", __func__, err); + objlayout_put_deviceinfo(deviceaddr); + return err ? ERR_PTR(err) : ode; +} + +static int objio_devices_lookup(struct pnfs_layout_hdr *pnfslay, + struct objio_segment *objio_seg, + gfp_t gfp_flags) +{ + unsigned i; + int err; + + /* lookup all devices */ + for (i = 0; i < objio_seg->num_comps; i++) { + struct objio_dev_ent *ode; + + ode = _device_lookup(pnfslay, objio_seg, i, gfp_flags); + if (unlikely(IS_ERR(ode))) { + err = PTR_ERR(ode); + goto out; + } + objio_seg->ods[i] = ode; + } + err = 0; + +out: + dprintk("%s: return=%d\n", __func__, err); + return err; +} + +static int _verify_data_map(struct pnfs_osd_layout *layout) +{ + struct pnfs_osd_data_map *data_map = &layout->olo_map; + u64 stripe_length; + u32 group_width; + +/* FIXME: Only raid0 for now. if not go through MDS */ + if (data_map->odm_raid_algorithm != PNFS_OSD_RAID_0) { + printk(KERN_ERR "Only RAID_0 for now\n"); + return -ENOTSUPP; + } + if (0 != (data_map->odm_num_comps % (data_map->odm_mirror_cnt + 1))) { + printk(KERN_ERR "Data Map wrong, num_comps=%u mirrors=%u\n", + data_map->odm_num_comps, data_map->odm_mirror_cnt); + return -EINVAL; + } + + if (data_map->odm_group_width) + group_width = data_map->odm_group_width; + else + group_width = data_map->odm_num_comps / + (data_map->odm_mirror_cnt + 1); + + stripe_length = (u64)data_map->odm_stripe_unit * group_width; + if (stripe_length >= (1ULL << 32)) { + printk(KERN_ERR "Total Stripe length(0x%llx)" + " >= 32bit is not supported\n", _LLU(stripe_length)); + return -ENOTSUPP; + } + + if (0 != (data_map->odm_stripe_unit & ~PAGE_MASK)) { + printk(KERN_ERR "Stripe Unit(0x%llx)" + " must be Multples of PAGE_SIZE(0x%lx)\n", + _LLU(data_map->odm_stripe_unit), PAGE_SIZE); + return -ENOTSUPP; + } + + return 0; +} + +static void copy_single_comp(struct pnfs_osd_object_cred *cur_comp, + struct pnfs_osd_object_cred *src_comp, + struct caps_buffers *caps_p) +{ + WARN_ON(src_comp->oc_cap_key.cred_len > sizeof(caps_p->caps_key)); + WARN_ON(src_comp->oc_cap.cred_len > sizeof(caps_p->creds)); + + *cur_comp = *src_comp; + + memcpy(caps_p->caps_key, src_comp->oc_cap_key.cred, + sizeof(caps_p->caps_key)); + cur_comp->oc_cap_key.cred = caps_p->caps_key; + + memcpy(caps_p->creds, src_comp->oc_cap.cred, + sizeof(caps_p->creds)); + cur_comp->oc_cap.cred = caps_p->creds; +} + +int objio_alloc_lseg(struct pnfs_layout_segment **outp, + struct pnfs_layout_hdr *pnfslay, + struct pnfs_layout_range *range, + struct xdr_stream *xdr, + gfp_t gfp_flags) +{ + struct objio_segment *objio_seg; + struct pnfs_osd_xdr_decode_layout_iter iter; + struct pnfs_osd_layout layout; + struct pnfs_osd_object_cred *cur_comp, src_comp; + struct caps_buffers *caps_p; + int err; + + err = pnfs_osd_xdr_decode_layout_map(&layout, &iter, xdr); + if (unlikely(err)) + return err; + + err = _verify_data_map(&layout); + if (unlikely(err)) + return err; + + objio_seg = kzalloc(sizeof(*objio_seg) + + sizeof(objio_seg->ods[0]) * layout.olo_num_comps + + sizeof(*objio_seg->comps) * layout.olo_num_comps + + sizeof(struct caps_buffers) * layout.olo_num_comps, + gfp_flags); + if (!objio_seg) + return -ENOMEM; + + objio_seg->comps = (void *)(objio_seg->ods + layout.olo_num_comps); + cur_comp = objio_seg->comps; + caps_p = (void *)(cur_comp + layout.olo_num_comps); + while (pnfs_osd_xdr_decode_layout_comp(&src_comp, &iter, xdr, &err)) + copy_single_comp(cur_comp++, &src_comp, caps_p++); + if (unlikely(err)) + goto err; + + objio_seg->num_comps = layout.olo_num_comps; + objio_seg->comps_index = layout.olo_comps_index; + err = objio_devices_lookup(pnfslay, objio_seg, gfp_flags); + if (err) + goto err; + + objio_seg->mirrors_p1 = layout.olo_map.odm_mirror_cnt + 1; + objio_seg->stripe_unit = layout.olo_map.odm_stripe_unit; + if (layout.olo_map.odm_group_width) { + objio_seg->group_width = layout.olo_map.odm_group_width; + objio_seg->group_depth = layout.olo_map.odm_group_depth; + objio_seg->group_count = layout.olo_map.odm_num_comps / + objio_seg->mirrors_p1 / + objio_seg->group_width; + } else { + objio_seg->group_width = layout.olo_map.odm_num_comps / + objio_seg->mirrors_p1; + objio_seg->group_depth = -1; + objio_seg->group_count = 1; + } + + /* Cache this calculation it will hit for every page */ + objio_seg->max_io_size = (BIO_MAX_PAGES_KMALLOC * PAGE_SIZE - + objio_seg->stripe_unit) * + objio_seg->group_width; + + *outp = &objio_seg->lseg; + return 0; + +err: + kfree(objio_seg); + dprintk("%s: Error: return %d\n", __func__, err); + *outp = NULL; + return err; +} + +void objio_free_lseg(struct pnfs_layout_segment *lseg) +{ + int i; + struct objio_segment *objio_seg = OBJIO_LSEG(lseg); + + for (i = 0; i < objio_seg->num_comps; i++) { + if (!objio_seg->ods[i]) + break; + nfs4_put_deviceid_node(&objio_seg->ods[i]->id_node); + } + kfree(objio_seg); +} + +int objio_alloc_io_state(struct pnfs_layout_segment *lseg, + struct objlayout_io_state **outp, + gfp_t gfp_flags) +{ + struct objio_segment *objio_seg = OBJIO_LSEG(lseg); + struct objio_state *ios; + const unsigned first_size = sizeof(*ios) + + objio_seg->num_comps * sizeof(ios->per_dev[0]); + const unsigned sec_size = objio_seg->num_comps * + sizeof(ios->ol_state.ioerrs[0]); + + ios = kzalloc(first_size + sec_size, gfp_flags); + if (unlikely(!ios)) + return -ENOMEM; + + ios->layout = objio_seg; + ios->ol_state.ioerrs = ((void *)ios) + first_size; + ios->ol_state.num_comps = objio_seg->num_comps; + + *outp = &ios->ol_state; + return 0; +} + +void objio_free_io_state(struct objlayout_io_state *ol_state) +{ + struct objio_state *ios = container_of(ol_state, struct objio_state, + ol_state); + + kfree(ios); +} + +enum pnfs_osd_errno osd_pri_2_pnfs_err(enum osd_err_priority oep) +{ + switch (oep) { + case OSD_ERR_PRI_NO_ERROR: + return (enum pnfs_osd_errno)0; + + case OSD_ERR_PRI_CLEAR_PAGES: + BUG_ON(1); + return 0; + + case OSD_ERR_PRI_RESOURCE: + return PNFS_OSD_ERR_RESOURCE; + case OSD_ERR_PRI_BAD_CRED: + return PNFS_OSD_ERR_BAD_CRED; + case OSD_ERR_PRI_NO_ACCESS: + return PNFS_OSD_ERR_NO_ACCESS; + case OSD_ERR_PRI_UNREACHABLE: + return PNFS_OSD_ERR_UNREACHABLE; + case OSD_ERR_PRI_NOT_FOUND: + return PNFS_OSD_ERR_NOT_FOUND; + case OSD_ERR_PRI_NO_SPACE: + return PNFS_OSD_ERR_NO_SPACE; + default: + WARN_ON(1); + /* fallthrough */ + case OSD_ERR_PRI_EIO: + return PNFS_OSD_ERR_EIO; + } +} + +static void _clear_bio(struct bio *bio) +{ + struct bio_vec *bv; + unsigned i; + + __bio_for_each_segment(bv, bio, i, 0) { + unsigned this_count = bv->bv_len; + + if (likely(PAGE_SIZE == this_count)) + clear_highpage(bv->bv_page); + else + zero_user(bv->bv_page, bv->bv_offset, this_count); + } +} + +static int _io_check(struct objio_state *ios, bool is_write) +{ + enum osd_err_priority oep = OSD_ERR_PRI_NO_ERROR; + int lin_ret = 0; + int i; + + for (i = 0; i < ios->numdevs; i++) { + struct osd_sense_info osi; + struct osd_request *or = ios->per_dev[i].or; + unsigned dev; + int ret; + + if (!or) + continue; + + ret = osd_req_decode_sense(or, &osi); + if (likely(!ret)) + continue; + + if (OSD_ERR_PRI_CLEAR_PAGES == osi.osd_err_pri) { + /* start read offset passed endof file */ + BUG_ON(is_write); + _clear_bio(ios->per_dev[i].bio); + dprintk("%s: start read offset passed end of file " + "offset=0x%llx, length=0x%lx\n", __func__, + _LLU(ios->per_dev[i].offset), + ios->per_dev[i].length); + + continue; /* we recovered */ + } + dev = ios->per_dev[i].dev; + objlayout_io_set_result(&ios->ol_state, dev, + &ios->layout->comps[dev].oc_object_id, + osd_pri_2_pnfs_err(osi.osd_err_pri), + ios->per_dev[i].offset, + ios->per_dev[i].length, + is_write); + + if (osi.osd_err_pri >= oep) { + oep = osi.osd_err_pri; + lin_ret = ret; + } + } + + return lin_ret; +} + +/* + * Common IO state helpers. + */ +static void _io_free(struct objio_state *ios) +{ + unsigned i; + + for (i = 0; i < ios->numdevs; i++) { + struct _objio_per_comp *per_dev = &ios->per_dev[i]; + + if (per_dev->or) { + osd_end_request(per_dev->or); + per_dev->or = NULL; + } + + if (per_dev->bio) { + bio_put(per_dev->bio); + per_dev->bio = NULL; + } + } +} + +struct osd_dev *_io_od(struct objio_state *ios, unsigned dev) +{ + unsigned min_dev = ios->layout->comps_index; + unsigned max_dev = min_dev + ios->layout->num_comps; + + BUG_ON(dev < min_dev || max_dev <= dev); + return ios->layout->ods[dev - min_dev]->od; +} + +struct _striping_info { + u64 obj_offset; + u64 group_length; + unsigned dev; + unsigned unit_off; +}; + +static void _calc_stripe_info(struct objio_state *ios, u64 file_offset, + struct _striping_info *si) +{ + u32 stripe_unit = ios->layout->stripe_unit; + u32 group_width = ios->layout->group_width; + u64 group_depth = ios->layout->group_depth; + u32 U = stripe_unit * group_width; + + u64 T = U * group_depth; + u64 S = T * ios->layout->group_count; + u64 M = div64_u64(file_offset, S); + + /* + G = (L - (M * S)) / T + H = (L - (M * S)) % T + */ + u64 LmodU = file_offset - M * S; + u32 G = div64_u64(LmodU, T); + u64 H = LmodU - G * T; + + u32 N = div_u64(H, U); + + div_u64_rem(file_offset, stripe_unit, &si->unit_off); + si->obj_offset = si->unit_off + (N * stripe_unit) + + (M * group_depth * stripe_unit); + + /* "H - (N * U)" is just "H % U" so it's bound to u32 */ + si->dev = (u32)(H - (N * U)) / stripe_unit + G * group_width; + si->dev *= ios->layout->mirrors_p1; + + si->group_length = T - H; +} + +static int _add_stripe_unit(struct objio_state *ios, unsigned *cur_pg, + unsigned pgbase, struct _objio_per_comp *per_dev, int cur_len, + gfp_t gfp_flags) +{ + unsigned pg = *cur_pg; + struct request_queue *q = + osd_request_queue(_io_od(ios, per_dev->dev)); + + per_dev->length += cur_len; + + if (per_dev->bio == NULL) { + unsigned stripes = ios->layout->num_comps / + ios->layout->mirrors_p1; + unsigned pages_in_stripe = stripes * + (ios->layout->stripe_unit / PAGE_SIZE); + unsigned bio_size = (ios->ol_state.nr_pages + pages_in_stripe) / + stripes; + + if (BIO_MAX_PAGES_KMALLOC < bio_size) + bio_size = BIO_MAX_PAGES_KMALLOC; + + per_dev->bio = bio_kmalloc(gfp_flags, bio_size); + if (unlikely(!per_dev->bio)) { + dprintk("Faild to allocate BIO size=%u\n", bio_size); + return -ENOMEM; + } + } + + while (cur_len > 0) { + unsigned pglen = min_t(unsigned, PAGE_SIZE - pgbase, cur_len); + unsigned added_len; + + BUG_ON(ios->ol_state.nr_pages <= pg); + cur_len -= pglen; + + added_len = bio_add_pc_page(q, per_dev->bio, + ios->ol_state.pages[pg], pglen, pgbase); + if (unlikely(pglen != added_len)) + return -ENOMEM; + pgbase = 0; + ++pg; + } + BUG_ON(cur_len); + + *cur_pg = pg; + return 0; +} + +static int _prepare_one_group(struct objio_state *ios, u64 length, + struct _striping_info *si, unsigned *last_pg, + gfp_t gfp_flags) +{ + unsigned stripe_unit = ios->layout->stripe_unit; + unsigned mirrors_p1 = ios->layout->mirrors_p1; + unsigned devs_in_group = ios->layout->group_width * mirrors_p1; + unsigned dev = si->dev; + unsigned first_dev = dev - (dev % devs_in_group); + unsigned max_comp = ios->numdevs ? ios->numdevs - mirrors_p1 : 0; + unsigned cur_pg = *last_pg; + int ret = 0; + + while (length) { + struct _objio_per_comp *per_dev = &ios->per_dev[dev]; + unsigned cur_len, page_off = 0; + + if (!per_dev->length) { + per_dev->dev = dev; + if (dev < si->dev) { + per_dev->offset = si->obj_offset + stripe_unit - + si->unit_off; + cur_len = stripe_unit; + } else if (dev == si->dev) { + per_dev->offset = si->obj_offset; + cur_len = stripe_unit - si->unit_off; + page_off = si->unit_off & ~PAGE_MASK; + BUG_ON(page_off && + (page_off != ios->ol_state.pgbase)); + } else { /* dev > si->dev */ + per_dev->offset = si->obj_offset - si->unit_off; + cur_len = stripe_unit; + } + + if (max_comp < dev) + max_comp = dev; + } else { + cur_len = stripe_unit; + } + if (cur_len >= length) + cur_len = length; + + ret = _add_stripe_unit(ios, &cur_pg, page_off , per_dev, + cur_len, gfp_flags); + if (unlikely(ret)) + goto out; + + dev += mirrors_p1; + dev = (dev % devs_in_group) + first_dev; + + length -= cur_len; + ios->length += cur_len; + } +out: + ios->numdevs = max_comp + mirrors_p1; + *last_pg = cur_pg; + return ret; +} + +static int _io_rw_pagelist(struct objio_state *ios, gfp_t gfp_flags) +{ + u64 length = ios->ol_state.count; + u64 offset = ios->ol_state.offset; + struct _striping_info si; + unsigned last_pg = 0; + int ret = 0; + + while (length) { + _calc_stripe_info(ios, offset, &si); + + if (length < si.group_length) + si.group_length = length; + + ret = _prepare_one_group(ios, si.group_length, &si, &last_pg, gfp_flags); + if (unlikely(ret)) + goto out; + + offset += si.group_length; + length -= si.group_length; + } + +out: + if (!ios->length) + return ret; + + return 0; +} + +static ssize_t _sync_done(struct objio_state *ios) +{ + struct completion *waiting = ios->private; + + complete(waiting); + return 0; +} + +static void _last_io(struct kref *kref) +{ + struct objio_state *ios = container_of(kref, struct objio_state, kref); + + ios->done(ios); +} + +static void _done_io(struct osd_request *or, void *p) +{ + struct objio_state *ios = p; + + kref_put(&ios->kref, _last_io); +} + +static ssize_t _io_exec(struct objio_state *ios) +{ + DECLARE_COMPLETION_ONSTACK(wait); + ssize_t status = 0; /* sync status */ + unsigned i; + objio_done_fn saved_done_fn = ios->done; + bool sync = ios->ol_state.sync; + + if (sync) { + ios->done = _sync_done; + ios->private = &wait; + } + + kref_init(&ios->kref); + + for (i = 0; i < ios->numdevs; i++) { + struct osd_request *or = ios->per_dev[i].or; + + if (!or) + continue; + + kref_get(&ios->kref); + osd_execute_request_async(or, _done_io, ios); + } + + kref_put(&ios->kref, _last_io); + + if (sync) { + wait_for_completion(&wait); + status = saved_done_fn(ios); + } + + return status; +} + +/* + * read + */ +static ssize_t _read_done(struct objio_state *ios) +{ + ssize_t status; + int ret = _io_check(ios, false); + + _io_free(ios); + + if (likely(!ret)) + status = ios->length; + else + status = ret; + + objlayout_read_done(&ios->ol_state, status, ios->ol_state.sync); + return status; +} + +static int _read_mirrors(struct objio_state *ios, unsigned cur_comp) +{ + struct osd_request *or = NULL; + struct _objio_per_comp *per_dev = &ios->per_dev[cur_comp]; + unsigned dev = per_dev->dev; + struct pnfs_osd_object_cred *cred = + &ios->layout->comps[dev]; + struct osd_obj_id obj = { + .partition = cred->oc_object_id.oid_partition_id, + .id = cred->oc_object_id.oid_object_id, + }; + int ret; + + or = osd_start_request(_io_od(ios, dev), GFP_KERNEL); + if (unlikely(!or)) { + ret = -ENOMEM; + goto err; + } + per_dev->or = or; + + osd_req_read(or, &obj, per_dev->offset, per_dev->bio, per_dev->length); + + ret = osd_finalize_request(or, 0, cred->oc_cap.cred, NULL); + if (ret) { + dprintk("%s: Faild to osd_finalize_request() => %d\n", + __func__, ret); + goto err; + } + + dprintk("%s:[%d] dev=%d obj=0x%llx start=0x%llx length=0x%lx\n", + __func__, cur_comp, dev, obj.id, _LLU(per_dev->offset), + per_dev->length); + +err: + return ret; +} + +static ssize_t _read_exec(struct objio_state *ios) +{ + unsigned i; + int ret; + + for (i = 0; i < ios->numdevs; i += ios->layout->mirrors_p1) { + if (!ios->per_dev[i].length) + continue; + ret = _read_mirrors(ios, i); + if (unlikely(ret)) + goto err; + } + + ios->done = _read_done; + return _io_exec(ios); /* In sync mode exec returns the io status */ + +err: + _io_free(ios); + return ret; +} + +ssize_t objio_read_pagelist(struct objlayout_io_state *ol_state) +{ + struct objio_state *ios = container_of(ol_state, struct objio_state, + ol_state); + int ret; + + ret = _io_rw_pagelist(ios, GFP_KERNEL); + if (unlikely(ret)) + return ret; + + return _read_exec(ios); +} + +/* + * write + */ +static ssize_t _write_done(struct objio_state *ios) +{ + ssize_t status; + int ret = _io_check(ios, true); + + _io_free(ios); + + if (likely(!ret)) { + /* FIXME: should be based on the OSD's persistence model + * See OSD2r05 Section 4.13 Data persistence model */ + ios->ol_state.committed = NFS_FILE_SYNC; + status = ios->length; + } else { + status = ret; + } + + objlayout_write_done(&ios->ol_state, status, ios->ol_state.sync); + return status; +} + +static int _write_mirrors(struct objio_state *ios, unsigned cur_comp) +{ + struct _objio_per_comp *master_dev = &ios->per_dev[cur_comp]; + unsigned dev = ios->per_dev[cur_comp].dev; + unsigned last_comp = cur_comp + ios->layout->mirrors_p1; + int ret; + + for (; cur_comp < last_comp; ++cur_comp, ++dev) { + struct osd_request *or = NULL; + struct pnfs_osd_object_cred *cred = + &ios->layout->comps[dev]; + struct osd_obj_id obj = { + .partition = cred->oc_object_id.oid_partition_id, + .id = cred->oc_object_id.oid_object_id, + }; + struct _objio_per_comp *per_dev = &ios->per_dev[cur_comp]; + struct bio *bio; + + or = osd_start_request(_io_od(ios, dev), GFP_NOFS); + if (unlikely(!or)) { + ret = -ENOMEM; + goto err; + } + per_dev->or = or; + + if (per_dev != master_dev) { + bio = bio_kmalloc(GFP_NOFS, + master_dev->bio->bi_max_vecs); + if (unlikely(!bio)) { + dprintk("Faild to allocate BIO size=%u\n", + master_dev->bio->bi_max_vecs); + ret = -ENOMEM; + goto err; + } + + __bio_clone(bio, master_dev->bio); + bio->bi_bdev = NULL; + bio->bi_next = NULL; + per_dev->bio = bio; + per_dev->dev = dev; + per_dev->length = master_dev->length; + per_dev->offset = master_dev->offset; + } else { + bio = master_dev->bio; + bio->bi_rw |= REQ_WRITE; + } + + osd_req_write(or, &obj, per_dev->offset, bio, per_dev->length); + + ret = osd_finalize_request(or, 0, cred->oc_cap.cred, NULL); + if (ret) { + dprintk("%s: Faild to osd_finalize_request() => %d\n", + __func__, ret); + goto err; + } + + dprintk("%s:[%d] dev=%d obj=0x%llx start=0x%llx length=0x%lx\n", + __func__, cur_comp, dev, obj.id, _LLU(per_dev->offset), + per_dev->length); + } + +err: + return ret; +} + +static ssize_t _write_exec(struct objio_state *ios) +{ + unsigned i; + int ret; + + for (i = 0; i < ios->numdevs; i += ios->layout->mirrors_p1) { + if (!ios->per_dev[i].length) + continue; + ret = _write_mirrors(ios, i); + if (unlikely(ret)) + goto err; + } + + ios->done = _write_done; + return _io_exec(ios); /* In sync mode exec returns the io->status */ + +err: + _io_free(ios); + return ret; +} + +ssize_t objio_write_pagelist(struct objlayout_io_state *ol_state, bool stable) +{ + struct objio_state *ios = container_of(ol_state, struct objio_state, + ol_state); + int ret; + + /* TODO: ios->stable = stable; */ + ret = _io_rw_pagelist(ios, GFP_NOFS); + if (unlikely(ret)) + return ret; + + return _write_exec(ios); +} + +static bool objio_pg_test(struct nfs_pageio_descriptor *pgio, + struct nfs_page *prev, struct nfs_page *req) +{ + if (!pnfs_generic_pg_test(pgio, prev, req)) + return false; + + if (pgio->pg_lseg == NULL) + return true; + + return pgio->pg_count + req->wb_bytes <= + OBJIO_LSEG(pgio->pg_lseg)->max_io_size; +} + +static struct pnfs_layoutdriver_type objlayout_type = { + .id = LAYOUT_OSD2_OBJECTS, + .name = "LAYOUT_OSD2_OBJECTS", + .flags = PNFS_LAYOUTRET_ON_SETATTR, + + .alloc_layout_hdr = objlayout_alloc_layout_hdr, + .free_layout_hdr = objlayout_free_layout_hdr, + + .alloc_lseg = objlayout_alloc_lseg, + .free_lseg = objlayout_free_lseg, + + .read_pagelist = objlayout_read_pagelist, + .write_pagelist = objlayout_write_pagelist, + .pg_test = objio_pg_test, + + .free_deviceid_node = objio_free_deviceid_node, + + .encode_layoutcommit = objlayout_encode_layoutcommit, + .encode_layoutreturn = objlayout_encode_layoutreturn, +}; + +MODULE_DESCRIPTION("pNFS Layout Driver for OSD2 objects"); +MODULE_AUTHOR("Benny Halevy <bhalevy@panasas.com>"); +MODULE_LICENSE("GPL"); + +static int __init +objlayout_init(void) +{ + int ret = pnfs_register_layoutdriver(&objlayout_type); + + if (ret) + printk(KERN_INFO + "%s: Registering OSD pNFS Layout Driver failed: error=%d\n", + __func__, ret); + else + printk(KERN_INFO "%s: Registered OSD pNFS Layout Driver\n", + __func__); + return ret; +} + +static void __exit +objlayout_exit(void) +{ + pnfs_unregister_layoutdriver(&objlayout_type); + printk(KERN_INFO "%s: Unregistered OSD pNFS Layout Driver\n", + __func__); +} + +module_init(objlayout_init); +module_exit(objlayout_exit); diff --git a/fs/nfs/objlayout/objlayout.c b/fs/nfs/objlayout/objlayout.c new file mode 100644 index 000000000000..1d06f8e2adea --- /dev/null +++ b/fs/nfs/objlayout/objlayout.c @@ -0,0 +1,712 @@ +/* + * pNFS Objects layout driver high level definitions + * + * Copyright (C) 2007 Panasas Inc. [year of first publication] + * All rights reserved. + * + * Benny Halevy <bhalevy@panasas.com> + * Boaz Harrosh <bharrosh@panasas.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 + * See the file COPYING included with this distribution for more details. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the Panasas company nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED + * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE + * DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR + * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF + * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING + * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS + * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include <scsi/osd_initiator.h> +#include "objlayout.h" + +#define NFSDBG_FACILITY NFSDBG_PNFS_LD +/* + * Create a objlayout layout structure for the given inode and return it. + */ +struct pnfs_layout_hdr * +objlayout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags) +{ + struct objlayout *objlay; + + objlay = kzalloc(sizeof(struct objlayout), gfp_flags); + if (objlay) { + spin_lock_init(&objlay->lock); + INIT_LIST_HEAD(&objlay->err_list); + } + dprintk("%s: Return %p\n", __func__, objlay); + return &objlay->pnfs_layout; +} + +/* + * Free an objlayout layout structure + */ +void +objlayout_free_layout_hdr(struct pnfs_layout_hdr *lo) +{ + struct objlayout *objlay = OBJLAYOUT(lo); + + dprintk("%s: objlay %p\n", __func__, objlay); + + WARN_ON(!list_empty(&objlay->err_list)); + kfree(objlay); +} + +/* + * Unmarshall layout and store it in pnfslay. + */ +struct pnfs_layout_segment * +objlayout_alloc_lseg(struct pnfs_layout_hdr *pnfslay, + struct nfs4_layoutget_res *lgr, + gfp_t gfp_flags) +{ + int status = -ENOMEM; + struct xdr_stream stream; + struct xdr_buf buf = { + .pages = lgr->layoutp->pages, + .page_len = lgr->layoutp->len, + .buflen = lgr->layoutp->len, + .len = lgr->layoutp->len, + }; + struct page *scratch; + struct pnfs_layout_segment *lseg; + + dprintk("%s: Begin pnfslay %p\n", __func__, pnfslay); + + scratch = alloc_page(gfp_flags); + if (!scratch) + goto err_nofree; + + xdr_init_decode(&stream, &buf, NULL); + xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE); + + status = objio_alloc_lseg(&lseg, pnfslay, &lgr->range, &stream, gfp_flags); + if (unlikely(status)) { + dprintk("%s: objio_alloc_lseg Return err %d\n", __func__, + status); + goto err; + } + + __free_page(scratch); + + dprintk("%s: Return %p\n", __func__, lseg); + return lseg; + +err: + __free_page(scratch); +err_nofree: + dprintk("%s: Err Return=>%d\n", __func__, status); + return ERR_PTR(status); +} + +/* + * Free a layout segement + */ +void +objlayout_free_lseg(struct pnfs_layout_segment *lseg) +{ + dprintk("%s: freeing layout segment %p\n", __func__, lseg); + + if (unlikely(!lseg)) + return; + + objio_free_lseg(lseg); +} + +/* + * I/O Operations + */ +static inline u64 +end_offset(u64 start, u64 len) +{ + u64 end; + + end = start + len; + return end >= start ? end : NFS4_MAX_UINT64; +} + +/* last octet in a range */ +static inline u64 +last_byte_offset(u64 start, u64 len) +{ + u64 end; + + BUG_ON(!len); + end = start + len; + return end > start ? end - 1 : NFS4_MAX_UINT64; +} + +static struct objlayout_io_state * +objlayout_alloc_io_state(struct pnfs_layout_hdr *pnfs_layout_type, + struct page **pages, + unsigned pgbase, + loff_t offset, + size_t count, + struct pnfs_layout_segment *lseg, + void *rpcdata, + gfp_t gfp_flags) +{ + struct objlayout_io_state *state; + u64 lseg_end_offset; + + dprintk("%s: allocating io_state\n", __func__); + if (objio_alloc_io_state(lseg, &state, gfp_flags)) + return NULL; + + BUG_ON(offset < lseg->pls_range.offset); + lseg_end_offset = end_offset(lseg->pls_range.offset, + lseg->pls_range.length); + BUG_ON(offset >= lseg_end_offset); + if (offset + count > lseg_end_offset) { + count = lseg->pls_range.length - + (offset - lseg->pls_range.offset); + dprintk("%s: truncated count %Zd\n", __func__, count); + } + + if (pgbase > PAGE_SIZE) { + pages += pgbase >> PAGE_SHIFT; + pgbase &= ~PAGE_MASK; + } + + INIT_LIST_HEAD(&state->err_list); + state->lseg = lseg; + state->rpcdata = rpcdata; + state->pages = pages; + state->pgbase = pgbase; + state->nr_pages = (pgbase + count + PAGE_SIZE - 1) >> PAGE_SHIFT; + state->offset = offset; + state->count = count; + state->sync = 0; + + return state; +} + +static void +objlayout_free_io_state(struct objlayout_io_state *state) +{ + dprintk("%s: freeing io_state\n", __func__); + if (unlikely(!state)) + return; + + objio_free_io_state(state); +} + +/* + * I/O done common code + */ +static void +objlayout_iodone(struct objlayout_io_state *state) +{ + dprintk("%s: state %p status\n", __func__, state); + + if (likely(state->status >= 0)) { + objlayout_free_io_state(state); + } else { + struct objlayout *objlay = OBJLAYOUT(state->lseg->pls_layout); + + spin_lock(&objlay->lock); + objlay->delta_space_valid = OBJ_DSU_INVALID; + list_add(&objlay->err_list, &state->err_list); + spin_unlock(&objlay->lock); + } +} + +/* + * objlayout_io_set_result - Set an osd_error code on a specific osd comp. + * + * The @index component IO failed (error returned from target). Register + * the error for later reporting at layout-return. + */ +void +objlayout_io_set_result(struct objlayout_io_state *state, unsigned index, + struct pnfs_osd_objid *pooid, int osd_error, + u64 offset, u64 length, bool is_write) +{ + struct pnfs_osd_ioerr *ioerr = &state->ioerrs[index]; + + BUG_ON(index >= state->num_comps); + if (osd_error) { + ioerr->oer_component = *pooid; + ioerr->oer_comp_offset = offset; + ioerr->oer_comp_length = length; + ioerr->oer_iswrite = is_write; + ioerr->oer_errno = osd_error; + + dprintk("%s: err[%d]: errno=%d is_write=%d dev(%llx:%llx) " + "par=0x%llx obj=0x%llx offset=0x%llx length=0x%llx\n", + __func__, index, ioerr->oer_errno, + ioerr->oer_iswrite, + _DEVID_LO(&ioerr->oer_component.oid_device_id), + _DEVID_HI(&ioerr->oer_component.oid_device_id), + ioerr->oer_component.oid_partition_id, + ioerr->oer_component.oid_object_id, + ioerr->oer_comp_offset, + ioerr->oer_comp_length); + } else { + /* User need not call if no error is reported */ + ioerr->oer_errno = 0; + } +} + +/* Function scheduled on rpc workqueue to call ->nfs_readlist_complete(). + * This is because the osd completion is called with ints-off from + * the block layer + */ +static void _rpc_read_complete(struct work_struct *work) +{ + struct rpc_task *task; + struct nfs_read_data *rdata; + + dprintk("%s enter\n", __func__); + task = container_of(work, struct rpc_task, u.tk_work); + rdata = container_of(task, struct nfs_read_data, task); + + pnfs_ld_read_done(rdata); +} + +void +objlayout_read_done(struct objlayout_io_state *state, ssize_t status, bool sync) +{ + int eof = state->eof; + struct nfs_read_data *rdata; + + state->status = status; + dprintk("%s: Begin status=%zd eof=%d\n", __func__, status, eof); + rdata = state->rpcdata; + rdata->task.tk_status = status; + if (status >= 0) { + rdata->res.count = status; + rdata->res.eof = eof; + } + objlayout_iodone(state); + /* must not use state after this point */ + + if (sync) + pnfs_ld_read_done(rdata); + else { + INIT_WORK(&rdata->task.u.tk_work, _rpc_read_complete); + schedule_work(&rdata->task.u.tk_work); + } +} + +/* + * Perform sync or async reads. + */ +enum pnfs_try_status +objlayout_read_pagelist(struct nfs_read_data *rdata) +{ + loff_t offset = rdata->args.offset; + size_t count = rdata->args.count; + struct objlayout_io_state *state; + ssize_t status = 0; + loff_t eof; + + dprintk("%s: Begin inode %p offset %llu count %d\n", + __func__, rdata->inode, offset, (int)count); + + eof = i_size_read(rdata->inode); + if (unlikely(offset + count > eof)) { + if (offset >= eof) { + status = 0; + rdata->res.count = 0; + rdata->res.eof = 1; + goto out; + } + count = eof - offset; + } + + state = objlayout_alloc_io_state(NFS_I(rdata->inode)->layout, + rdata->args.pages, rdata->args.pgbase, + offset, count, + rdata->lseg, rdata, + GFP_KERNEL); + if (unlikely(!state)) { + status = -ENOMEM; + goto out; + } + + state->eof = state->offset + state->count >= eof; + + status = objio_read_pagelist(state); + out: + dprintk("%s: Return status %Zd\n", __func__, status); + rdata->pnfs_error = status; + return PNFS_ATTEMPTED; +} + +/* Function scheduled on rpc workqueue to call ->nfs_writelist_complete(). + * This is because the osd completion is called with ints-off from + * the block layer + */ +static void _rpc_write_complete(struct work_struct *work) +{ + struct rpc_task *task; + struct nfs_write_data *wdata; + + dprintk("%s enter\n", __func__); + task = container_of(work, struct rpc_task, u.tk_work); + wdata = container_of(task, struct nfs_write_data, task); + + pnfs_ld_write_done(wdata); +} + +void +objlayout_write_done(struct objlayout_io_state *state, ssize_t status, + bool sync) +{ + struct nfs_write_data *wdata; + + dprintk("%s: Begin\n", __func__); + wdata = state->rpcdata; + state->status = status; + wdata->task.tk_status = status; + if (status >= 0) { + wdata->res.count = status; + wdata->verf.committed = state->committed; + dprintk("%s: Return status %d committed %d\n", + __func__, wdata->task.tk_status, + wdata->verf.committed); + } else + dprintk("%s: Return status %d\n", + __func__, wdata->task.tk_status); + objlayout_iodone(state); + /* must not use state after this point */ + + if (sync) + pnfs_ld_write_done(wdata); + else { + INIT_WORK(&wdata->task.u.tk_work, _rpc_write_complete); + schedule_work(&wdata->task.u.tk_work); + } +} + +/* + * Perform sync or async writes. + */ +enum pnfs_try_status +objlayout_write_pagelist(struct nfs_write_data *wdata, + int how) +{ + struct objlayout_io_state *state; + ssize_t status; + + dprintk("%s: Begin inode %p offset %llu count %u\n", + __func__, wdata->inode, wdata->args.offset, wdata->args.count); + + state = objlayout_alloc_io_state(NFS_I(wdata->inode)->layout, + wdata->args.pages, + wdata->args.pgbase, + wdata->args.offset, + wdata->args.count, + wdata->lseg, wdata, + GFP_NOFS); + if (unlikely(!state)) { + status = -ENOMEM; + goto out; + } + + state->sync = how & FLUSH_SYNC; + + status = objio_write_pagelist(state, how & FLUSH_STABLE); + out: + dprintk("%s: Return status %Zd\n", __func__, status); + wdata->pnfs_error = status; + return PNFS_ATTEMPTED; +} + +void +objlayout_encode_layoutcommit(struct pnfs_layout_hdr *pnfslay, + struct xdr_stream *xdr, + const struct nfs4_layoutcommit_args *args) +{ + struct objlayout *objlay = OBJLAYOUT(pnfslay); + struct pnfs_osd_layoutupdate lou; + __be32 *start; + + dprintk("%s: Begin\n", __func__); + + spin_lock(&objlay->lock); + lou.dsu_valid = (objlay->delta_space_valid == OBJ_DSU_VALID); + lou.dsu_delta = objlay->delta_space_used; + objlay->delta_space_used = 0; + objlay->delta_space_valid = OBJ_DSU_INIT; + lou.olu_ioerr_flag = !list_empty(&objlay->err_list); + spin_unlock(&objlay->lock); + + start = xdr_reserve_space(xdr, 4); + + BUG_ON(pnfs_osd_xdr_encode_layoutupdate(xdr, &lou)); + + *start = cpu_to_be32((xdr->p - start - 1) * 4); + + dprintk("%s: Return delta_space_used %lld err %d\n", __func__, + lou.dsu_delta, lou.olu_ioerr_flag); +} + +static int +err_prio(u32 oer_errno) +{ + switch (oer_errno) { + case 0: + return 0; + + case PNFS_OSD_ERR_RESOURCE: + return OSD_ERR_PRI_RESOURCE; + case PNFS_OSD_ERR_BAD_CRED: + return OSD_ERR_PRI_BAD_CRED; + case PNFS_OSD_ERR_NO_ACCESS: + return OSD_ERR_PRI_NO_ACCESS; + case PNFS_OSD_ERR_UNREACHABLE: + return OSD_ERR_PRI_UNREACHABLE; + case PNFS_OSD_ERR_NOT_FOUND: + return OSD_ERR_PRI_NOT_FOUND; + case PNFS_OSD_ERR_NO_SPACE: + return OSD_ERR_PRI_NO_SPACE; + default: + WARN_ON(1); + /* fallthrough */ + case PNFS_OSD_ERR_EIO: + return OSD_ERR_PRI_EIO; + } +} + +static void +merge_ioerr(struct pnfs_osd_ioerr *dest_err, + const struct pnfs_osd_ioerr *src_err) +{ + u64 dest_end, src_end; + + if (!dest_err->oer_errno) { + *dest_err = *src_err; + /* accumulated device must be blank */ + memset(&dest_err->oer_component.oid_device_id, 0, + sizeof(dest_err->oer_component.oid_device_id)); + + return; + } + + if (dest_err->oer_component.oid_partition_id != + src_err->oer_component.oid_partition_id) + dest_err->oer_component.oid_partition_id = 0; + + if (dest_err->oer_component.oid_object_id != + src_err->oer_component.oid_object_id) + dest_err->oer_component.oid_object_id = 0; + + if (dest_err->oer_comp_offset > src_err->oer_comp_offset) + dest_err->oer_comp_offset = src_err->oer_comp_offset; + + dest_end = end_offset(dest_err->oer_comp_offset, + dest_err->oer_comp_length); + src_end = end_offset(src_err->oer_comp_offset, + src_err->oer_comp_length); + if (dest_end < src_end) + dest_end = src_end; + + dest_err->oer_comp_length = dest_end - dest_err->oer_comp_offset; + + if ((src_err->oer_iswrite == dest_err->oer_iswrite) && + (err_prio(src_err->oer_errno) > err_prio(dest_err->oer_errno))) { + dest_err->oer_errno = src_err->oer_errno; + } else if (src_err->oer_iswrite) { + dest_err->oer_iswrite = true; + dest_err->oer_errno = src_err->oer_errno; + } +} + +static void +encode_accumulated_error(struct objlayout *objlay, __be32 *p) +{ + struct objlayout_io_state *state, *tmp; + struct pnfs_osd_ioerr accumulated_err = {.oer_errno = 0}; + + list_for_each_entry_safe(state, tmp, &objlay->err_list, err_list) { + unsigned i; + + for (i = 0; i < state->num_comps; i++) { + struct pnfs_osd_ioerr *ioerr = &state->ioerrs[i]; + + if (!ioerr->oer_errno) + continue; + + printk(KERN_ERR "%s: err[%d]: errno=%d is_write=%d " + "dev(%llx:%llx) par=0x%llx obj=0x%llx " + "offset=0x%llx length=0x%llx\n", + __func__, i, ioerr->oer_errno, + ioerr->oer_iswrite, + _DEVID_LO(&ioerr->oer_component.oid_device_id), + _DEVID_HI(&ioerr->oer_component.oid_device_id), + ioerr->oer_component.oid_partition_id, + ioerr->oer_component.oid_object_id, + ioerr->oer_comp_offset, + ioerr->oer_comp_length); + + merge_ioerr(&accumulated_err, ioerr); + } + list_del(&state->err_list); + objlayout_free_io_state(state); + } + + pnfs_osd_xdr_encode_ioerr(p, &accumulated_err); +} + +void +objlayout_encode_layoutreturn(struct pnfs_layout_hdr *pnfslay, + struct xdr_stream *xdr, + const struct nfs4_layoutreturn_args *args) +{ + struct objlayout *objlay = OBJLAYOUT(pnfslay); + struct objlayout_io_state *state, *tmp; + __be32 *start; + + dprintk("%s: Begin\n", __func__); + start = xdr_reserve_space(xdr, 4); + BUG_ON(!start); + + spin_lock(&objlay->lock); + + list_for_each_entry_safe(state, tmp, &objlay->err_list, err_list) { + __be32 *last_xdr = NULL, *p; + unsigned i; + int res = 0; + + for (i = 0; i < state->num_comps; i++) { + struct pnfs_osd_ioerr *ioerr = &state->ioerrs[i]; + + if (!ioerr->oer_errno) + continue; + + dprintk("%s: err[%d]: errno=%d is_write=%d " + "dev(%llx:%llx) par=0x%llx obj=0x%llx " + "offset=0x%llx length=0x%llx\n", + __func__, i, ioerr->oer_errno, + ioerr->oer_iswrite, + _DEVID_LO(&ioerr->oer_component.oid_device_id), + _DEVID_HI(&ioerr->oer_component.oid_device_id), + ioerr->oer_component.oid_partition_id, + ioerr->oer_component.oid_object_id, + ioerr->oer_comp_offset, + ioerr->oer_comp_length); + + p = pnfs_osd_xdr_ioerr_reserve_space(xdr); + if (unlikely(!p)) { + res = -E2BIG; + break; /* accumulated_error */ + } + + last_xdr = p; + pnfs_osd_xdr_encode_ioerr(p, &state->ioerrs[i]); + } + + /* TODO: use xdr_write_pages */ + if (unlikely(res)) { + /* no space for even one error descriptor */ + BUG_ON(!last_xdr); + + /* we've encountered a situation with lots and lots of + * errors and no space to encode them all. Use the last + * available slot to report the union of all the + * remaining errors. + */ + encode_accumulated_error(objlay, last_xdr); + goto loop_done; + } + list_del(&state->err_list); + objlayout_free_io_state(state); + } +loop_done: + spin_unlock(&objlay->lock); + + *start = cpu_to_be32((xdr->p - start - 1) * 4); + dprintk("%s: Return\n", __func__); +} + + +/* + * Get Device Info API for io engines + */ +struct objlayout_deviceinfo { + struct page *page; + struct pnfs_osd_deviceaddr da; /* This must be last */ +}; + +/* Initialize and call nfs_getdeviceinfo, then decode and return a + * "struct pnfs_osd_deviceaddr *" Eventually objlayout_put_deviceinfo() + * should be called. + */ +int objlayout_get_deviceinfo(struct pnfs_layout_hdr *pnfslay, + struct nfs4_deviceid *d_id, struct pnfs_osd_deviceaddr **deviceaddr, + gfp_t gfp_flags) +{ + struct objlayout_deviceinfo *odi; + struct pnfs_device pd; + struct super_block *sb; + struct page *page, **pages; + u32 *p; + int err; + + page = alloc_page(gfp_flags); + if (!page) + return -ENOMEM; + + pages = &page; + pd.pages = pages; + + memcpy(&pd.dev_id, d_id, sizeof(*d_id)); + pd.layout_type = LAYOUT_OSD2_OBJECTS; + pd.pages = &page; + pd.pgbase = 0; + pd.pglen = PAGE_SIZE; + pd.mincount = 0; + + sb = pnfslay->plh_inode->i_sb; + err = nfs4_proc_getdeviceinfo(NFS_SERVER(pnfslay->plh_inode), &pd); + dprintk("%s nfs_getdeviceinfo returned %d\n", __func__, err); + if (err) + goto err_out; + + p = page_address(page); + odi = kzalloc(sizeof(*odi), gfp_flags); + if (!odi) { + err = -ENOMEM; + goto err_out; + } + pnfs_osd_xdr_decode_deviceaddr(&odi->da, p); + odi->page = page; + *deviceaddr = &odi->da; + return 0; + +err_out: + __free_page(page); + return err; +} + +void objlayout_put_deviceinfo(struct pnfs_osd_deviceaddr *deviceaddr) +{ + struct objlayout_deviceinfo *odi = container_of(deviceaddr, + struct objlayout_deviceinfo, + da); + + __free_page(odi->page); + kfree(odi); +} diff --git a/fs/nfs/objlayout/objlayout.h b/fs/nfs/objlayout/objlayout.h new file mode 100644 index 000000000000..a8244c8e042d --- /dev/null +++ b/fs/nfs/objlayout/objlayout.h @@ -0,0 +1,187 @@ +/* + * Data types and function declerations for interfacing with the + * pNFS standard object layout driver. + * + * Copyright (C) 2007 Panasas Inc. [year of first publication] + * All rights reserved. + * + * Benny Halevy <bhalevy@panasas.com> + * Boaz Harrosh <bharrosh@panasas.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 + * See the file COPYING included with this distribution for more details. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the Panasas company nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED + * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE + * DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR + * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF + * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING + * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS + * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#ifndef _OBJLAYOUT_H +#define _OBJLAYOUT_H + +#include <linux/nfs_fs.h> +#include <linux/pnfs_osd_xdr.h> +#include "../pnfs.h" + +/* + * per-inode layout + */ +struct objlayout { + struct pnfs_layout_hdr pnfs_layout; + + /* for layout_commit */ + enum osd_delta_space_valid_enum { + OBJ_DSU_INIT = 0, + OBJ_DSU_VALID, + OBJ_DSU_INVALID, + } delta_space_valid; + s64 delta_space_used; /* consumed by write ops */ + + /* for layout_return */ + spinlock_t lock; + struct list_head err_list; +}; + +static inline struct objlayout * +OBJLAYOUT(struct pnfs_layout_hdr *lo) +{ + return container_of(lo, struct objlayout, pnfs_layout); +} + +/* + * per-I/O operation state + * embedded in objects provider io_state data structure + */ +struct objlayout_io_state { + struct pnfs_layout_segment *lseg; + + struct page **pages; + unsigned pgbase; + unsigned nr_pages; + unsigned long count; + loff_t offset; + bool sync; + + void *rpcdata; + int status; /* res */ + int eof; /* res */ + int committed; /* res */ + + /* Error reporting (layout_return) */ + struct list_head err_list; + unsigned num_comps; + /* Pointer to array of error descriptors of size num_comps. + * It should contain as many entries as devices in the osd_layout + * that participate in the I/O. It is up to the io_engine to allocate + * needed space and set num_comps. + */ + struct pnfs_osd_ioerr *ioerrs; +}; + +/* + * Raid engine I/O API + */ +extern int objio_alloc_lseg(struct pnfs_layout_segment **outp, + struct pnfs_layout_hdr *pnfslay, + struct pnfs_layout_range *range, + struct xdr_stream *xdr, + gfp_t gfp_flags); +extern void objio_free_lseg(struct pnfs_layout_segment *lseg); + +extern int objio_alloc_io_state( + struct pnfs_layout_segment *lseg, + struct objlayout_io_state **outp, + gfp_t gfp_flags); +extern void objio_free_io_state(struct objlayout_io_state *state); + +extern ssize_t objio_read_pagelist(struct objlayout_io_state *ol_state); +extern ssize_t objio_write_pagelist(struct objlayout_io_state *ol_state, + bool stable); + +/* + * callback API + */ +extern void objlayout_io_set_result(struct objlayout_io_state *state, + unsigned index, struct pnfs_osd_objid *pooid, + int osd_error, u64 offset, u64 length, bool is_write); + +static inline void +objlayout_add_delta_space_used(struct objlayout_io_state *state, s64 space_used) +{ + struct objlayout *objlay = OBJLAYOUT(state->lseg->pls_layout); + + /* If one of the I/Os errored out and the delta_space_used was + * invalid we render the complete report as invalid. Protocol mandate + * the DSU be accurate or not reported. + */ + spin_lock(&objlay->lock); + if (objlay->delta_space_valid != OBJ_DSU_INVALID) { + objlay->delta_space_valid = OBJ_DSU_VALID; + objlay->delta_space_used += space_used; + } + spin_unlock(&objlay->lock); +} + +extern void objlayout_read_done(struct objlayout_io_state *state, + ssize_t status, bool sync); +extern void objlayout_write_done(struct objlayout_io_state *state, + ssize_t status, bool sync); + +extern int objlayout_get_deviceinfo(struct pnfs_layout_hdr *pnfslay, + struct nfs4_deviceid *d_id, struct pnfs_osd_deviceaddr **deviceaddr, + gfp_t gfp_flags); +extern void objlayout_put_deviceinfo(struct pnfs_osd_deviceaddr *deviceaddr); + +/* + * exported generic objects function vectors + */ + +extern struct pnfs_layout_hdr *objlayout_alloc_layout_hdr(struct inode *, gfp_t gfp_flags); +extern void objlayout_free_layout_hdr(struct pnfs_layout_hdr *); + +extern struct pnfs_layout_segment *objlayout_alloc_lseg( + struct pnfs_layout_hdr *, + struct nfs4_layoutget_res *, + gfp_t gfp_flags); +extern void objlayout_free_lseg(struct pnfs_layout_segment *); + +extern enum pnfs_try_status objlayout_read_pagelist( + struct nfs_read_data *); + +extern enum pnfs_try_status objlayout_write_pagelist( + struct nfs_write_data *, + int how); + +extern void objlayout_encode_layoutcommit( + struct pnfs_layout_hdr *, + struct xdr_stream *, + const struct nfs4_layoutcommit_args *); + +extern void objlayout_encode_layoutreturn( + struct pnfs_layout_hdr *, + struct xdr_stream *, + const struct nfs4_layoutreturn_args *); + +#endif /* _OBJLAYOUT_H */ diff --git a/fs/nfs/objlayout/pnfs_osd_xdr_cli.c b/fs/nfs/objlayout/pnfs_osd_xdr_cli.c new file mode 100644 index 000000000000..16fc758e9123 --- /dev/null +++ b/fs/nfs/objlayout/pnfs_osd_xdr_cli.c @@ -0,0 +1,412 @@ +/* + * Object-Based pNFS Layout XDR layer + * + * Copyright (C) 2007 Panasas Inc. [year of first publication] + * All rights reserved. + * + * Benny Halevy <bhalevy@panasas.com> + * Boaz Harrosh <bharrosh@panasas.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 + * See the file COPYING included with this distribution for more details. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the Panasas company nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED + * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF + * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE + * DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR + * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF + * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING + * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS + * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include <linux/pnfs_osd_xdr.h> + +#define NFSDBG_FACILITY NFSDBG_PNFS_LD + +/* + * The following implementation is based on RFC5664 + */ + +/* + * struct pnfs_osd_objid { + * struct nfs4_deviceid oid_device_id; + * u64 oid_partition_id; + * u64 oid_object_id; + * }; // xdr size 32 bytes + */ +static __be32 * +_osd_xdr_decode_objid(__be32 *p, struct pnfs_osd_objid *objid) +{ + p = xdr_decode_opaque_fixed(p, objid->oid_device_id.data, + sizeof(objid->oid_device_id.data)); + + p = xdr_decode_hyper(p, &objid->oid_partition_id); + p = xdr_decode_hyper(p, &objid->oid_object_id); + return p; +} +/* + * struct pnfs_osd_opaque_cred { + * u32 cred_len; + * void *cred; + * }; // xdr size [variable] + * The return pointers are from the xdr buffer + */ +static int +_osd_xdr_decode_opaque_cred(struct pnfs_osd_opaque_cred *opaque_cred, + struct xdr_stream *xdr) +{ + __be32 *p = xdr_inline_decode(xdr, 1); + + if (!p) + return -EINVAL; + + opaque_cred->cred_len = be32_to_cpu(*p++); + + p = xdr_inline_decode(xdr, opaque_cred->cred_len); + if (!p) + return -EINVAL; + + opaque_cred->cred = p; + return 0; +} + +/* + * struct pnfs_osd_object_cred { + * struct pnfs_osd_objid oc_object_id; + * u32 oc_osd_version; + * u32 oc_cap_key_sec; + * struct pnfs_osd_opaque_cred oc_cap_key + * struct pnfs_osd_opaque_cred oc_cap; + * }; // xdr size 32 + 4 + 4 + [variable] + [variable] + */ +static int +_osd_xdr_decode_object_cred(struct pnfs_osd_object_cred *comp, + struct xdr_stream *xdr) +{ + __be32 *p = xdr_inline_decode(xdr, 32 + 4 + 4); + int ret; + + if (!p) + return -EIO; + + p = _osd_xdr_decode_objid(p, &comp->oc_object_id); + comp->oc_osd_version = be32_to_cpup(p++); + comp->oc_cap_key_sec = be32_to_cpup(p); + + ret = _osd_xdr_decode_opaque_cred(&comp->oc_cap_key, xdr); + if (unlikely(ret)) + return ret; + + ret = _osd_xdr_decode_opaque_cred(&comp->oc_cap, xdr); + return ret; +} + +/* + * struct pnfs_osd_data_map { + * u32 odm_num_comps; + * u64 odm_stripe_unit; + * u32 odm_group_width; + * u32 odm_group_depth; + * u32 odm_mirror_cnt; + * u32 odm_raid_algorithm; + * }; // xdr size 4 + 8 + 4 + 4 + 4 + 4 + */ +static inline int +_osd_data_map_xdr_sz(void) +{ + return 4 + 8 + 4 + 4 + 4 + 4; +} + +static __be32 * +_osd_xdr_decode_data_map(__be32 *p, struct pnfs_osd_data_map *data_map) +{ + data_map->odm_num_comps = be32_to_cpup(p++); + p = xdr_decode_hyper(p, &data_map->odm_stripe_unit); + data_map->odm_group_width = be32_to_cpup(p++); + data_map->odm_group_depth = be32_to_cpup(p++); + data_map->odm_mirror_cnt = be32_to_cpup(p++); + data_map->odm_raid_algorithm = be32_to_cpup(p++); + dprintk("%s: odm_num_comps=%u odm_stripe_unit=%llu odm_group_width=%u " + "odm_group_depth=%u odm_mirror_cnt=%u odm_raid_algorithm=%u\n", + __func__, + data_map->odm_num_comps, + (unsigned long long)data_map->odm_stripe_unit, + data_map->odm_group_width, + data_map->odm_group_depth, + data_map->odm_mirror_cnt, + data_map->odm_raid_algorithm); + return p; +} + +int pnfs_osd_xdr_decode_layout_map(struct pnfs_osd_layout *layout, + struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr) +{ + __be32 *p; + + memset(iter, 0, sizeof(*iter)); + + p = xdr_inline_decode(xdr, _osd_data_map_xdr_sz() + 4 + 4); + if (unlikely(!p)) + return -EINVAL; + + p = _osd_xdr_decode_data_map(p, &layout->olo_map); + layout->olo_comps_index = be32_to_cpup(p++); + layout->olo_num_comps = be32_to_cpup(p++); + iter->total_comps = layout->olo_num_comps; + return 0; +} + +bool pnfs_osd_xdr_decode_layout_comp(struct pnfs_osd_object_cred *comp, + struct pnfs_osd_xdr_decode_layout_iter *iter, struct xdr_stream *xdr, + int *err) +{ + BUG_ON(iter->decoded_comps > iter->total_comps); + if (iter->decoded_comps == iter->total_comps) + return false; + + *err = _osd_xdr_decode_object_cred(comp, xdr); + if (unlikely(*err)) { + dprintk("%s: _osd_xdr_decode_object_cred=>%d decoded_comps=%d " + "total_comps=%d\n", __func__, *err, + iter->decoded_comps, iter->total_comps); + return false; /* stop the loop */ + } + dprintk("%s: dev(%llx:%llx) par=0x%llx obj=0x%llx " + "key_len=%u cap_len=%u\n", + __func__, + _DEVID_LO(&comp->oc_object_id.oid_device_id), + _DEVID_HI(&comp->oc_object_id.oid_device_id), + comp->oc_object_id.oid_partition_id, + comp->oc_object_id.oid_object_id, + comp->oc_cap_key.cred_len, comp->oc_cap.cred_len); + + iter->decoded_comps++; + return true; +} + +/* + * Get Device Information Decoding + * + * Note: since Device Information is currently done synchronously, all + * variable strings fields are left inside the rpc buffer and are only + * pointed to by the pnfs_osd_deviceaddr members. So the read buffer + * should not be freed while the returned information is in use. + */ +/* + *struct nfs4_string { + * unsigned int len; + * char *data; + *}; // size [variable] + * NOTE: Returned string points to inside the XDR buffer + */ +static __be32 * +__read_u8_opaque(__be32 *p, struct nfs4_string *str) +{ + str->len = be32_to_cpup(p++); + str->data = (char *)p; + + p += XDR_QUADLEN(str->len); + return p; +} + +/* + * struct pnfs_osd_targetid { + * u32 oti_type; + * struct nfs4_string oti_scsi_device_id; + * };// size 4 + [variable] + */ +static __be32 * +__read_targetid(__be32 *p, struct pnfs_osd_targetid* targetid) +{ + u32 oti_type; + + oti_type = be32_to_cpup(p++); + targetid->oti_type = oti_type; + + switch (oti_type) { + case OBJ_TARGET_SCSI_NAME: + case OBJ_TARGET_SCSI_DEVICE_ID: + p = __read_u8_opaque(p, &targetid->oti_scsi_device_id); + } + + return p; +} + +/* + * struct pnfs_osd_net_addr { + * struct nfs4_string r_netid; + * struct nfs4_string r_addr; + * }; + */ +static __be32 * +__read_net_addr(__be32 *p, struct pnfs_osd_net_addr* netaddr) +{ + p = __read_u8_opaque(p, &netaddr->r_netid); + p = __read_u8_opaque(p, &netaddr->r_addr); + + return p; +} + +/* + * struct pnfs_osd_targetaddr { + * u32 ota_available; + * struct pnfs_osd_net_addr ota_netaddr; + * }; + */ +static __be32 * +__read_targetaddr(__be32 *p, struct pnfs_osd_targetaddr *targetaddr) +{ + u32 ota_available; + + ota_available = be32_to_cpup(p++); + targetaddr->ota_available = ota_available; + + if (ota_available) + p = __read_net_addr(p, &targetaddr->ota_netaddr); + + + return p; +} + +/* + * struct pnfs_osd_deviceaddr { + * struct pnfs_osd_targetid oda_targetid; + * struct pnfs_osd_targetaddr oda_targetaddr; + * u8 oda_lun[8]; + * struct nfs4_string oda_systemid; + * struct pnfs_osd_object_cred oda_root_obj_cred; + * struct nfs4_string oda_osdname; + * }; + */ + +/* We need this version for the pnfs_osd_xdr_decode_deviceaddr which does + * not have an xdr_stream + */ +static __be32 * +__read_opaque_cred(__be32 *p, + struct pnfs_osd_opaque_cred *opaque_cred) +{ + opaque_cred->cred_len = be32_to_cpu(*p++); + opaque_cred->cred = p; + return p + XDR_QUADLEN(opaque_cred->cred_len); +} + +static __be32 * +__read_object_cred(__be32 *p, struct pnfs_osd_object_cred *comp) +{ + p = _osd_xdr_decode_objid(p, &comp->oc_object_id); + comp->oc_osd_version = be32_to_cpup(p++); + comp->oc_cap_key_sec = be32_to_cpup(p++); + + p = __read_opaque_cred(p, &comp->oc_cap_key); + p = __read_opaque_cred(p, &comp->oc_cap); + return p; +} + +void pnfs_osd_xdr_decode_deviceaddr( + struct pnfs_osd_deviceaddr *deviceaddr, __be32 *p) +{ + p = __read_targetid(p, &deviceaddr->oda_targetid); + + p = __read_targetaddr(p, &deviceaddr->oda_targetaddr); + + p = xdr_decode_opaque_fixed(p, deviceaddr->oda_lun, + sizeof(deviceaddr->oda_lun)); + + p = __read_u8_opaque(p, &deviceaddr->oda_systemid); + + p = __read_object_cred(p, &deviceaddr->oda_root_obj_cred); + + p = __read_u8_opaque(p, &deviceaddr->oda_osdname); + + /* libosd likes this terminated in dbg. It's last, so no problems */ + deviceaddr->oda_osdname.data[deviceaddr->oda_osdname.len] = 0; +} + +/* + * struct pnfs_osd_layoutupdate { + * u32 dsu_valid; + * s64 dsu_delta; + * u32 olu_ioerr_flag; + * }; xdr size 4 + 8 + 4 + */ +int +pnfs_osd_xdr_encode_layoutupdate(struct xdr_stream *xdr, + struct pnfs_osd_layoutupdate *lou) +{ + __be32 *p = xdr_reserve_space(xdr, 4 + 8 + 4); + + if (!p) + return -E2BIG; + + *p++ = cpu_to_be32(lou->dsu_valid); + if (lou->dsu_valid) + p = xdr_encode_hyper(p, lou->dsu_delta); + *p++ = cpu_to_be32(lou->olu_ioerr_flag); + return 0; +} + +/* + * struct pnfs_osd_objid { + * struct nfs4_deviceid oid_device_id; + * u64 oid_partition_id; + * u64 oid_object_id; + * }; // xdr size 32 bytes + */ +static inline __be32 * +pnfs_osd_xdr_encode_objid(__be32 *p, struct pnfs_osd_objid *object_id) +{ + p = xdr_encode_opaque_fixed(p, &object_id->oid_device_id.data, + sizeof(object_id->oid_device_id.data)); + p = xdr_encode_hyper(p, object_id->oid_partition_id); + p = xdr_encode_hyper(p, object_id->oid_object_id); + + return p; +} + +/* + * struct pnfs_osd_ioerr { + * struct pnfs_osd_objid oer_component; + * u64 oer_comp_offset; + * u64 oer_comp_length; + * u32 oer_iswrite; + * u32 oer_errno; + * }; // xdr size 32 + 24 bytes + */ +void pnfs_osd_xdr_encode_ioerr(__be32 *p, struct pnfs_osd_ioerr *ioerr) +{ + p = pnfs_osd_xdr_encode_objid(p, &ioerr->oer_component); + p = xdr_encode_hyper(p, ioerr->oer_comp_offset); + p = xdr_encode_hyper(p, ioerr->oer_comp_length); + *p++ = cpu_to_be32(ioerr->oer_iswrite); + *p = cpu_to_be32(ioerr->oer_errno); +} + +__be32 *pnfs_osd_xdr_ioerr_reserve_space(struct xdr_stream *xdr) +{ + __be32 *p; + + p = xdr_reserve_space(xdr, 32 + 24); + if (unlikely(!p)) + dprintk("%s: out of xdr space\n", __func__); + + return p; +} diff --git a/fs/nfs/pagelist.c b/fs/nfs/pagelist.c index c80add6e2213..009855716286 100644 --- a/fs/nfs/pagelist.c +++ b/fs/nfs/pagelist.c @@ -204,6 +204,22 @@ nfs_wait_on_request(struct nfs_page *req) TASK_UNINTERRUPTIBLE); } +bool nfs_generic_pg_test(struct nfs_pageio_descriptor *desc, struct nfs_page *prev, struct nfs_page *req) +{ + /* + * FIXME: ideally we should be able to coalesce all requests + * that are not block boundary aligned, but currently this + * is problematic for the case of bsize < PAGE_CACHE_SIZE, + * since nfs_flush_multi and nfs_pagein_multi assume you + * can have only one struct nfs_page. + */ + if (desc->pg_bsize < PAGE_SIZE) + return 0; + + return desc->pg_count + req->wb_bytes <= desc->pg_bsize; +} +EXPORT_SYMBOL_GPL(nfs_generic_pg_test); + /** * nfs_pageio_init - initialise a page io descriptor * @desc: pointer to descriptor @@ -229,6 +245,8 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc, desc->pg_ioflags = io_flags; desc->pg_error = 0; desc->pg_lseg = NULL; + desc->pg_test = nfs_generic_pg_test; + pnfs_pageio_init(desc, inode); } /** @@ -242,29 +260,23 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc, * * Return 'true' if this is the case, else return 'false'. */ -static int nfs_can_coalesce_requests(struct nfs_page *prev, - struct nfs_page *req, - struct nfs_pageio_descriptor *pgio) +static bool nfs_can_coalesce_requests(struct nfs_page *prev, + struct nfs_page *req, + struct nfs_pageio_descriptor *pgio) { if (req->wb_context->cred != prev->wb_context->cred) - return 0; + return false; if (req->wb_lock_context->lockowner != prev->wb_lock_context->lockowner) - return 0; + return false; if (req->wb_context->state != prev->wb_context->state) - return 0; + return false; if (req->wb_index != (prev->wb_index + 1)) - return 0; + return false; if (req->wb_pgbase != 0) - return 0; + return false; if (prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE) - return 0; - /* - * Non-whole file layouts need to check that req is inside of - * pgio->pg_lseg. - */ - if (pgio->pg_test && !pgio->pg_test(pgio, prev, req)) - return 0; - return 1; + return false; + return pgio->pg_test(pgio, prev, req); } /** @@ -278,31 +290,18 @@ static int nfs_can_coalesce_requests(struct nfs_page *prev, static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc, struct nfs_page *req) { - size_t newlen = req->wb_bytes; - if (desc->pg_count != 0) { struct nfs_page *prev; - /* - * FIXME: ideally we should be able to coalesce all requests - * that are not block boundary aligned, but currently this - * is problematic for the case of bsize < PAGE_CACHE_SIZE, - * since nfs_flush_multi and nfs_pagein_multi assume you - * can have only one struct nfs_page. - */ - if (desc->pg_bsize < PAGE_SIZE) - return 0; - newlen += desc->pg_count; - if (newlen > desc->pg_bsize) - return 0; prev = nfs_list_entry(desc->pg_list.prev); if (!nfs_can_coalesce_requests(prev, req, desc)) return 0; - } else + } else { desc->pg_base = req->wb_pgbase; + } nfs_list_remove_request(req); nfs_list_add_request(req, &desc->pg_list); - desc->pg_count = newlen; + desc->pg_count += req->wb_bytes; return 1; } diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index f57f5281a520..29c0ca7fc347 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -177,13 +177,28 @@ get_layout_hdr(struct pnfs_layout_hdr *lo) atomic_inc(&lo->plh_refcount); } +static struct pnfs_layout_hdr * +pnfs_alloc_layout_hdr(struct inode *ino, gfp_t gfp_flags) +{ + struct pnfs_layoutdriver_type *ld = NFS_SERVER(ino)->pnfs_curr_ld; + return ld->alloc_layout_hdr ? ld->alloc_layout_hdr(ino, gfp_flags) : + kzalloc(sizeof(struct pnfs_layout_hdr), gfp_flags); +} + +static void +pnfs_free_layout_hdr(struct pnfs_layout_hdr *lo) +{ + struct pnfs_layoutdriver_type *ld = NFS_SERVER(lo->plh_inode)->pnfs_curr_ld; + return ld->alloc_layout_hdr ? ld->free_layout_hdr(lo) : kfree(lo); +} + static void destroy_layout_hdr(struct pnfs_layout_hdr *lo) { dprintk("%s: freeing layout cache %p\n", __func__, lo); BUG_ON(!list_empty(&lo->plh_layouts)); NFS_I(lo->plh_inode)->layout = NULL; - kfree(lo); + pnfs_free_layout_hdr(lo); } static void @@ -228,7 +243,7 @@ put_lseg_common(struct pnfs_layout_segment *lseg) { struct inode *inode = lseg->pls_layout->plh_inode; - BUG_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags)); + WARN_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags)); list_del_init(&lseg->pls_list); if (list_empty(&lseg->pls_layout->plh_segs)) { set_bit(NFS_LAYOUT_DESTROYED, &lseg->pls_layout->plh_flags); @@ -261,11 +276,72 @@ put_lseg(struct pnfs_layout_segment *lseg) } EXPORT_SYMBOL_GPL(put_lseg); +static inline u64 +end_offset(u64 start, u64 len) +{ + u64 end; + + end = start + len; + return end >= start ? end : NFS4_MAX_UINT64; +} + +/* last octet in a range */ +static inline u64 +last_byte_offset(u64 start, u64 len) +{ + u64 end; + + BUG_ON(!len); + end = start + len; + return end > start ? end - 1 : NFS4_MAX_UINT64; +} + +/* + * is l2 fully contained in l1? + * start1 end1 + * [----------------------------------) + * start2 end2 + * [----------------) + */ +static inline int +lo_seg_contained(struct pnfs_layout_range *l1, + struct pnfs_layout_range *l2) +{ + u64 start1 = l1->offset; + u64 end1 = end_offset(start1, l1->length); + u64 start2 = l2->offset; + u64 end2 = end_offset(start2, l2->length); + + return (start1 <= start2) && (end1 >= end2); +} + +/* + * is l1 and l2 intersecting? + * start1 end1 + * [----------------------------------) + * start2 end2 + * [----------------) + */ +static inline int +lo_seg_intersecting(struct pnfs_layout_range *l1, + struct pnfs_layout_range *l2) +{ + u64 start1 = l1->offset; + u64 end1 = end_offset(start1, l1->length); + u64 start2 = l2->offset; + u64 end2 = end_offset(start2, l2->length); + + return (end1 == NFS4_MAX_UINT64 || end1 > start2) && + (end2 == NFS4_MAX_UINT64 || end2 > start1); +} + static bool -should_free_lseg(u32 lseg_iomode, u32 recall_iomode) +should_free_lseg(struct pnfs_layout_range *lseg_range, + struct pnfs_layout_range *recall_range) { - return (recall_iomode == IOMODE_ANY || - lseg_iomode == recall_iomode); + return (recall_range->iomode == IOMODE_ANY || + lseg_range->iomode == recall_range->iomode) && + lo_seg_intersecting(lseg_range, recall_range); } /* Returns 1 if lseg is removed from list, 0 otherwise */ @@ -296,7 +372,7 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg, int mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, - u32 iomode) + struct pnfs_layout_range *recall_range) { struct pnfs_layout_segment *lseg, *next; int invalid = 0, removed = 0; @@ -309,7 +385,8 @@ mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, return 0; } list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list) - if (should_free_lseg(lseg->pls_range.iomode, iomode)) { + if (!recall_range || + should_free_lseg(&lseg->pls_range, recall_range)) { dprintk("%s: freeing lseg %p iomode %d " "offset %llu length %llu\n", __func__, lseg, lseg->pls_range.iomode, lseg->pls_range.offset, @@ -358,7 +435,7 @@ pnfs_destroy_layout(struct nfs_inode *nfsi) lo = nfsi->layout; if (lo) { lo->plh_block_lgets++; /* permanently block new LAYOUTGETs */ - mark_matching_lsegs_invalid(lo, &tmp_list, IOMODE_ANY); + mark_matching_lsegs_invalid(lo, &tmp_list, NULL); } spin_unlock(&nfsi->vfs_inode.i_lock); pnfs_free_lseg_list(&tmp_list); @@ -467,7 +544,7 @@ pnfs_choose_layoutget_stateid(nfs4_stateid *dst, struct pnfs_layout_hdr *lo, static struct pnfs_layout_segment * send_layoutget(struct pnfs_layout_hdr *lo, struct nfs_open_context *ctx, - u32 iomode, + struct pnfs_layout_range *range, gfp_t gfp_flags) { struct inode *ino = lo->plh_inode; @@ -499,11 +576,11 @@ send_layoutget(struct pnfs_layout_hdr *lo, goto out_err_free; } - lgp->args.minlength = NFS4_MAX_UINT64; + lgp->args.minlength = PAGE_CACHE_SIZE; + if (lgp->args.minlength > range->length) + lgp->args.minlength = range->length; lgp->args.maxcount = PNFS_LAYOUT_MAXSIZE; - lgp->args.range.iomode = iomode; - lgp->args.range.offset = 0; - lgp->args.range.length = NFS4_MAX_UINT64; + lgp->args.range = *range; lgp->args.type = server->pnfs_curr_ld->id; lgp->args.inode = ino; lgp->args.ctx = get_nfs_open_context(ctx); @@ -518,7 +595,7 @@ send_layoutget(struct pnfs_layout_hdr *lo, nfs4_proc_layoutget(lgp); if (!lseg) { /* remember that LAYOUTGET failed and suspend trying */ - set_bit(lo_fail_bit(iomode), &lo->plh_flags); + set_bit(lo_fail_bit(range->iomode), &lo->plh_flags); } /* free xdr pages */ @@ -542,6 +619,56 @@ out_err_free: return NULL; } +/* Initiates a LAYOUTRETURN(FILE) */ +int +_pnfs_return_layout(struct inode *ino) +{ + struct pnfs_layout_hdr *lo = NULL; + struct nfs_inode *nfsi = NFS_I(ino); + LIST_HEAD(tmp_list); + struct nfs4_layoutreturn *lrp; + nfs4_stateid stateid; + int status = 0; + + dprintk("--> %s\n", __func__); + + spin_lock(&ino->i_lock); + lo = nfsi->layout; + if (!lo) { + spin_unlock(&ino->i_lock); + dprintk("%s: no layout to return\n", __func__); + return status; + } + stateid = nfsi->layout->plh_stateid; + /* Reference matched in nfs4_layoutreturn_release */ + get_layout_hdr(lo); + mark_matching_lsegs_invalid(lo, &tmp_list, NULL); + lo->plh_block_lgets++; + spin_unlock(&ino->i_lock); + pnfs_free_lseg_list(&tmp_list); + + WARN_ON(test_bit(NFS_INO_LAYOUTCOMMIT, &nfsi->flags)); + + lrp = kzalloc(sizeof(*lrp), GFP_KERNEL); + if (unlikely(lrp == NULL)) { + status = -ENOMEM; + set_bit(NFS_LAYOUT_RW_FAILED, &lo->plh_flags); + set_bit(NFS_LAYOUT_RO_FAILED, &lo->plh_flags); + put_layout_hdr(lo); + goto out; + } + + lrp->args.stateid = stateid; + lrp->args.layout_type = NFS_SERVER(ino)->pnfs_curr_ld->id; + lrp->args.inode = ino; + lrp->clp = NFS_SERVER(ino)->nfs_client; + + status = nfs4_proc_layoutreturn(lrp); +out: + dprintk("<-- %s status: %d\n", __func__, status); + return status; +} + bool pnfs_roc(struct inode *ino) { struct pnfs_layout_hdr *lo; @@ -625,10 +752,23 @@ bool pnfs_roc_drain(struct inode *ino, u32 *barrier) * are seen first. */ static s64 -cmp_layout(u32 iomode1, u32 iomode2) +cmp_layout(struct pnfs_layout_range *l1, + struct pnfs_layout_range *l2) { + s64 d; + + /* high offset > low offset */ + d = l1->offset - l2->offset; + if (d) + return d; + + /* short length > long length */ + d = l2->length - l1->length; + if (d) + return d; + /* read > read/write */ - return (int)(iomode2 == IOMODE_READ) - (int)(iomode1 == IOMODE_READ); + return (int)(l1->iomode == IOMODE_READ) - (int)(l2->iomode == IOMODE_READ); } static void @@ -636,13 +776,12 @@ pnfs_insert_layout(struct pnfs_layout_hdr *lo, struct pnfs_layout_segment *lseg) { struct pnfs_layout_segment *lp; - int found = 0; dprintk("%s:Begin\n", __func__); assert_spin_locked(&lo->plh_inode->i_lock); list_for_each_entry(lp, &lo->plh_segs, pls_list) { - if (cmp_layout(lp->pls_range.iomode, lseg->pls_range.iomode) > 0) + if (cmp_layout(&lseg->pls_range, &lp->pls_range) > 0) continue; list_add_tail(&lseg->pls_list, &lp->pls_list); dprintk("%s: inserted lseg %p " @@ -652,16 +791,14 @@ pnfs_insert_layout(struct pnfs_layout_hdr *lo, lseg->pls_range.offset, lseg->pls_range.length, lp, lp->pls_range.iomode, lp->pls_range.offset, lp->pls_range.length); - found = 1; - break; - } - if (!found) { - list_add_tail(&lseg->pls_list, &lo->plh_segs); - dprintk("%s: inserted lseg %p " - "iomode %d offset %llu length %llu at tail\n", - __func__, lseg, lseg->pls_range.iomode, - lseg->pls_range.offset, lseg->pls_range.length); + goto out; } + list_add_tail(&lseg->pls_list, &lo->plh_segs); + dprintk("%s: inserted lseg %p " + "iomode %d offset %llu length %llu at tail\n", + __func__, lseg, lseg->pls_range.iomode, + lseg->pls_range.offset, lseg->pls_range.length); +out: get_layout_hdr(lo); dprintk("%s:Return\n", __func__); @@ -672,7 +809,7 @@ alloc_init_layout_hdr(struct inode *ino, gfp_t gfp_flags) { struct pnfs_layout_hdr *lo; - lo = kzalloc(sizeof(struct pnfs_layout_hdr), gfp_flags); + lo = pnfs_alloc_layout_hdr(ino, gfp_flags); if (!lo) return NULL; atomic_set(&lo->plh_refcount, 1); @@ -705,7 +842,7 @@ pnfs_find_alloc_layout(struct inode *ino, gfp_t gfp_flags) if (likely(nfsi->layout == NULL)) /* Won the race? */ nfsi->layout = new; else - kfree(new); + pnfs_free_layout_hdr(new); return nfsi->layout; } @@ -721,16 +858,28 @@ pnfs_find_alloc_layout(struct inode *ino, gfp_t gfp_flags) * READ RW true */ static int -is_matching_lseg(struct pnfs_layout_segment *lseg, u32 iomode) +is_matching_lseg(struct pnfs_layout_range *ls_range, + struct pnfs_layout_range *range) { - return (iomode != IOMODE_RW || lseg->pls_range.iomode == IOMODE_RW); + struct pnfs_layout_range range1; + + if ((range->iomode == IOMODE_RW && + ls_range->iomode != IOMODE_RW) || + !lo_seg_intersecting(ls_range, range)) + return 0; + + /* range1 covers only the first byte in the range */ + range1 = *range; + range1.length = 1; + return lo_seg_contained(ls_range, &range1); } /* * lookup range in layout */ static struct pnfs_layout_segment * -pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode) +pnfs_find_lseg(struct pnfs_layout_hdr *lo, + struct pnfs_layout_range *range) { struct pnfs_layout_segment *lseg, *ret = NULL; @@ -739,11 +888,11 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode) assert_spin_locked(&lo->plh_inode->i_lock); list_for_each_entry(lseg, &lo->plh_segs, pls_list) { if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags) && - is_matching_lseg(lseg, iomode)) { + is_matching_lseg(&lseg->pls_range, range)) { ret = get_lseg(lseg); break; } - if (cmp_layout(iomode, lseg->pls_range.iomode) > 0) + if (lseg->pls_range.offset > range->offset) break; } @@ -759,9 +908,17 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode) struct pnfs_layout_segment * pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx, + loff_t pos, + u64 count, enum pnfs_iomode iomode, gfp_t gfp_flags) { + struct pnfs_layout_range arg = { + .iomode = iomode, + .offset = pos, + .length = count, + }; + unsigned pg_offset; struct nfs_inode *nfsi = NFS_I(ino); struct nfs_client *clp = NFS_SERVER(ino)->nfs_client; struct pnfs_layout_hdr *lo; @@ -789,7 +946,7 @@ pnfs_update_layout(struct inode *ino, goto out_unlock; /* Check to see if the layout for the given range already exists */ - lseg = pnfs_find_lseg(lo, iomode); + lseg = pnfs_find_lseg(lo, &arg); if (lseg) goto out_unlock; @@ -811,7 +968,14 @@ pnfs_update_layout(struct inode *ino, spin_unlock(&clp->cl_lock); } - lseg = send_layoutget(lo, ctx, iomode, gfp_flags); + pg_offset = arg.offset & ~PAGE_CACHE_MASK; + if (pg_offset) { + arg.offset -= pg_offset; + arg.length += pg_offset; + } + arg.length = PAGE_CACHE_ALIGN(arg.length); + + lseg = send_layoutget(lo, ctx, &arg, gfp_flags); if (!lseg && first) { spin_lock(&clp->cl_lock); list_del_init(&lo->plh_layouts); @@ -838,17 +1002,6 @@ pnfs_layout_process(struct nfs4_layoutget *lgp) struct nfs_client *clp = NFS_SERVER(ino)->nfs_client; int status = 0; - /* Verify we got what we asked for. - * Note that because the xdr parsing only accepts a single - * element array, this can fail even if the server is behaving - * correctly. - */ - if (lgp->args.range.iomode > res->range.iomode || - res->range.offset != 0 || - res->range.length != NFS4_MAX_UINT64) { - status = -EINVAL; - goto out; - } /* Inject layout blob into I/O device driver */ lseg = NFS_SERVER(ino)->pnfs_curr_ld->alloc_lseg(lo, res, lgp->gfp_flags); if (!lseg || IS_ERR(lseg)) { @@ -895,51 +1048,77 @@ out_forget_reply: goto out; } -static int pnfs_read_pg_test(struct nfs_pageio_descriptor *pgio, - struct nfs_page *prev, - struct nfs_page *req) -{ - if (pgio->pg_count == prev->wb_bytes) { - /* This is first coelesce call for a series of nfs_pages */ - pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode, - prev->wb_context, - IOMODE_READ, - GFP_KERNEL); - } - return NFS_SERVER(pgio->pg_inode)->pnfs_curr_ld->pg_test(pgio, prev, req); -} - -void -pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *inode) +bool +pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, + struct nfs_page *req) { - struct pnfs_layoutdriver_type *ld; + enum pnfs_iomode access_type; + gfp_t gfp_flags; - ld = NFS_SERVER(inode)->pnfs_curr_ld; - pgio->pg_test = (ld && ld->pg_test) ? pnfs_read_pg_test : NULL; -} + /* We assume that pg_ioflags == 0 iff we're reading a page */ + if (pgio->pg_ioflags == 0) { + access_type = IOMODE_READ; + gfp_flags = GFP_KERNEL; + } else { + access_type = IOMODE_RW; + gfp_flags = GFP_NOFS; + } -static int pnfs_write_pg_test(struct nfs_pageio_descriptor *pgio, - struct nfs_page *prev, - struct nfs_page *req) -{ - if (pgio->pg_count == prev->wb_bytes) { + if (pgio->pg_lseg == NULL) { + if (pgio->pg_count != prev->wb_bytes) + return true; /* This is first coelesce call for a series of nfs_pages */ pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode, prev->wb_context, - IOMODE_RW, - GFP_NOFS); + req_offset(prev), + pgio->pg_count, + access_type, + gfp_flags); + if (pgio->pg_lseg == NULL) + return true; } - return NFS_SERVER(pgio->pg_inode)->pnfs_curr_ld->pg_test(pgio, prev, req); + + /* + * Test if a nfs_page is fully contained in the pnfs_layout_range. + * Note that this test makes several assumptions: + * - that the previous nfs_page in the struct nfs_pageio_descriptor + * is known to lie within the range. + * - that the nfs_page being tested is known to be contiguous with the + * previous nfs_page. + * - Layout ranges are page aligned, so we only have to test the + * start offset of the request. + * + * Please also note that 'end_offset' is actually the offset of the + * first byte that lies outside the pnfs_layout_range. FIXME? + * + */ + return req_offset(req) < end_offset(pgio->pg_lseg->pls_range.offset, + pgio->pg_lseg->pls_range.length); } +EXPORT_SYMBOL_GPL(pnfs_generic_pg_test); -void -pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode) +/* + * Called by non rpc-based layout drivers + */ +int +pnfs_ld_write_done(struct nfs_write_data *data) { - struct pnfs_layoutdriver_type *ld; + int status; - ld = NFS_SERVER(inode)->pnfs_curr_ld; - pgio->pg_test = (ld && ld->pg_test) ? pnfs_write_pg_test : NULL; + if (!data->pnfs_error) { + pnfs_set_layoutcommit(data); + data->mds_ops->rpc_call_done(&data->task, data); + data->mds_ops->rpc_release(data); + return 0; + } + + dprintk("%s: pnfs_error=%d, retry via MDS\n", __func__, + data->pnfs_error); + status = nfs_initiate_write(data, NFS_CLIENT(data->inode), + data->mds_ops, NFS_FILE_SYNC); + return status ? : -EAGAIN; } +EXPORT_SYMBOL_GPL(pnfs_ld_write_done); enum pnfs_try_status pnfs_try_to_write_data(struct nfs_write_data *wdata, @@ -966,6 +1145,29 @@ pnfs_try_to_write_data(struct nfs_write_data *wdata, } /* + * Called by non rpc-based layout drivers + */ +int +pnfs_ld_read_done(struct nfs_read_data *data) +{ + int status; + + if (!data->pnfs_error) { + __nfs4_read_done_cb(data); + data->mds_ops->rpc_call_done(&data->task, data); + data->mds_ops->rpc_release(data); + return 0; + } + + dprintk("%s: pnfs_error=%d, retry via MDS\n", __func__, + data->pnfs_error); + status = nfs_initiate_read(data, NFS_CLIENT(data->inode), + data->mds_ops); + return status ? : -EAGAIN; +} +EXPORT_SYMBOL_GPL(pnfs_ld_read_done); + +/* * Call the appropriate parallel I/O subsystem read function. */ enum pnfs_try_status @@ -1009,7 +1211,7 @@ void pnfs_set_layoutcommit(struct nfs_write_data *wdata) { struct nfs_inode *nfsi = NFS_I(wdata->inode); - loff_t end_pos = wdata->args.offset + wdata->res.count; + loff_t end_pos = wdata->mds_offset + wdata->res.count; bool mark_as_dirty = false; spin_lock(&nfsi->vfs_inode.i_lock); diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h index 0c015bad9e7a..96bf4e6f45be 100644 --- a/fs/nfs/pnfs.h +++ b/fs/nfs/pnfs.h @@ -30,6 +30,7 @@ #ifndef FS_NFS_PNFS_H #define FS_NFS_PNFS_H +#include <linux/nfs_fs.h> #include <linux/nfs_page.h> enum { @@ -64,17 +65,29 @@ enum { NFS_LAYOUT_DESTROYED, /* no new use of layout allowed */ }; +enum layoutdriver_policy_flags { + /* Should the pNFS client commit and return the layout upon a setattr */ + PNFS_LAYOUTRET_ON_SETATTR = 1 << 0, +}; + +struct nfs4_deviceid_node; + /* Per-layout driver specific registration structure */ struct pnfs_layoutdriver_type { struct list_head pnfs_tblid; const u32 id; const char *name; struct module *owner; + unsigned flags; + + struct pnfs_layout_hdr * (*alloc_layout_hdr) (struct inode *inode, gfp_t gfp_flags); + void (*free_layout_hdr) (struct pnfs_layout_hdr *); + struct pnfs_layout_segment * (*alloc_lseg) (struct pnfs_layout_hdr *layoutid, struct nfs4_layoutget_res *lgr, gfp_t gfp_flags); void (*free_lseg) (struct pnfs_layout_segment *lseg); /* test for nfs page cache coalescing */ - int (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *); + bool (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *); /* Returns true if layoutdriver wants to divert this request to * driver's commit routine. @@ -89,6 +102,16 @@ struct pnfs_layoutdriver_type { */ enum pnfs_try_status (*read_pagelist) (struct nfs_read_data *nfs_data); enum pnfs_try_status (*write_pagelist) (struct nfs_write_data *nfs_data, int how); + + void (*free_deviceid_node) (struct nfs4_deviceid_node *); + + void (*encode_layoutreturn) (struct pnfs_layout_hdr *layoutid, + struct xdr_stream *xdr, + const struct nfs4_layoutreturn_args *args); + + void (*encode_layoutcommit) (struct pnfs_layout_hdr *layoutid, + struct xdr_stream *xdr, + const struct nfs4_layoutcommit_args *args); }; struct pnfs_layout_hdr { @@ -120,21 +143,22 @@ extern void pnfs_unregister_layoutdriver(struct pnfs_layoutdriver_type *); extern int nfs4_proc_getdeviceinfo(struct nfs_server *server, struct pnfs_device *dev); extern int nfs4_proc_layoutget(struct nfs4_layoutget *lgp); +extern int nfs4_proc_layoutreturn(struct nfs4_layoutreturn *lrp); /* pnfs.c */ void get_layout_hdr(struct pnfs_layout_hdr *lo); void put_lseg(struct pnfs_layout_segment *lseg); struct pnfs_layout_segment * pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx, - enum pnfs_iomode access_type, gfp_t gfp_flags); + loff_t pos, u64 count, enum pnfs_iomode access_type, + gfp_t gfp_flags); void set_pnfs_layoutdriver(struct nfs_server *, u32 id); void unset_pnfs_layoutdriver(struct nfs_server *); enum pnfs_try_status pnfs_try_to_write_data(struct nfs_write_data *, const struct rpc_call_ops *, int); enum pnfs_try_status pnfs_try_to_read_data(struct nfs_read_data *, const struct rpc_call_ops *); -void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *); -void pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *); +bool pnfs_generic_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev, struct nfs_page *req); int pnfs_layout_process(struct nfs4_layoutget *lgp); void pnfs_free_lseg_list(struct list_head *tmp_list); void pnfs_destroy_layout(struct nfs_inode *); @@ -148,13 +172,38 @@ int pnfs_choose_layoutget_stateid(nfs4_stateid *dst, struct nfs4_state *open_state); int mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo, struct list_head *tmp_list, - u32 iomode); + struct pnfs_layout_range *recall_range); bool pnfs_roc(struct inode *ino); void pnfs_roc_release(struct inode *ino); void pnfs_roc_set_barrier(struct inode *ino, u32 barrier); bool pnfs_roc_drain(struct inode *ino, u32 *barrier); void pnfs_set_layoutcommit(struct nfs_write_data *wdata); int pnfs_layoutcommit_inode(struct inode *inode, bool sync); +int _pnfs_return_layout(struct inode *); +int pnfs_ld_write_done(struct nfs_write_data *); +int pnfs_ld_read_done(struct nfs_read_data *); + +/* pnfs_dev.c */ +struct nfs4_deviceid_node { + struct hlist_node node; + struct hlist_node tmpnode; + const struct pnfs_layoutdriver_type *ld; + const struct nfs_client *nfs_client; + struct nfs4_deviceid deviceid; + atomic_t ref; +}; + +void nfs4_print_deviceid(const struct nfs4_deviceid *dev_id); +struct nfs4_deviceid_node *nfs4_find_get_deviceid(const struct pnfs_layoutdriver_type *, const struct nfs_client *, const struct nfs4_deviceid *); +struct nfs4_deviceid_node *nfs4_unhash_put_deviceid(const struct pnfs_layoutdriver_type *, const struct nfs_client *, const struct nfs4_deviceid *); +void nfs4_delete_deviceid(const struct pnfs_layoutdriver_type *, const struct nfs_client *, const struct nfs4_deviceid *); +void nfs4_init_deviceid_node(struct nfs4_deviceid_node *, + const struct pnfs_layoutdriver_type *, + const struct nfs_client *, + const struct nfs4_deviceid *); +struct nfs4_deviceid_node *nfs4_insert_deviceid_node(struct nfs4_deviceid_node *); +bool nfs4_put_deviceid_node(struct nfs4_deviceid_node *); +void nfs4_deviceid_purge_client(const struct nfs_client *); static inline int lo_fail_bit(u32 iomode) { @@ -223,6 +272,36 @@ static inline void pnfs_clear_request_commit(struct nfs_page *req) put_lseg(req->wb_commit_lseg); } +/* Should the pNFS client commit and return the layout upon a setattr */ +static inline bool +pnfs_ld_layoutret_on_setattr(struct inode *inode) +{ + if (!pnfs_enabled_sb(NFS_SERVER(inode))) + return false; + return NFS_SERVER(inode)->pnfs_curr_ld->flags & + PNFS_LAYOUTRET_ON_SETATTR; +} + +static inline int pnfs_return_layout(struct inode *ino) +{ + struct nfs_inode *nfsi = NFS_I(ino); + struct nfs_server *nfss = NFS_SERVER(ino); + + if (pnfs_enabled_sb(nfss) && nfsi->layout) + return _pnfs_return_layout(ino); + + return 0; +} + +static inline void pnfs_pageio_init(struct nfs_pageio_descriptor *pgio, + struct inode *inode) +{ + struct pnfs_layoutdriver_type *ld = NFS_SERVER(inode)->pnfs_curr_ld; + + if (ld) + pgio->pg_test = ld->pg_test; +} + #else /* CONFIG_NFS_V4_1 */ static inline void pnfs_destroy_all_layouts(struct nfs_client *clp) @@ -245,7 +324,8 @@ static inline void put_lseg(struct pnfs_layout_segment *lseg) static inline struct pnfs_layout_segment * pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx, - enum pnfs_iomode access_type, gfp_t gfp_flags) + loff_t pos, u64 count, enum pnfs_iomode access_type, + gfp_t gfp_flags) { return NULL; } @@ -264,6 +344,17 @@ pnfs_try_to_write_data(struct nfs_write_data *data, return PNFS_NOT_ATTEMPTED; } +static inline int pnfs_return_layout(struct inode *ino) +{ + return 0; +} + +static inline bool +pnfs_ld_layoutret_on_setattr(struct inode *inode) +{ + return false; +} + static inline bool pnfs_roc(struct inode *ino) { @@ -294,16 +385,9 @@ static inline void unset_pnfs_layoutdriver(struct nfs_server *s) { } -static inline void -pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *ino) -{ - pgio->pg_test = NULL; -} - -static inline void -pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *ino) +static inline void pnfs_pageio_init(struct nfs_pageio_descriptor *pgio, + struct inode *inode) { - pgio->pg_test = NULL; } static inline void @@ -331,6 +415,10 @@ static inline int pnfs_layoutcommit_inode(struct inode *inode, bool sync) { return 0; } + +static inline void nfs4_deviceid_purge_client(struct nfs_client *ncl) +{ +} #endif /* CONFIG_NFS_V4_1 */ #endif /* FS_NFS_PNFS_H */ diff --git a/fs/nfs/pnfs_dev.c b/fs/nfs/pnfs_dev.c new file mode 100644 index 000000000000..f0f8e1e22f6c --- /dev/null +++ b/fs/nfs/pnfs_dev.c @@ -0,0 +1,277 @@ +/* + * Device operations for the pnfs client. + * + * Copyright (c) 2002 + * The Regents of the University of Michigan + * All Rights Reserved + * + * Dean Hildebrand <dhildebz@umich.edu> + * Garth Goodson <Garth.Goodson@netapp.com> + * + * Permission is granted to use, copy, create derivative works, and + * redistribute this software and such derivative works for any purpose, + * so long as the name of the University of Michigan is not used in + * any advertising or publicity pertaining to the use or distribution + * of this software without specific, written prior authorization. If + * the above copyright notice or any other identification of the + * University of Michigan is included in any copy of any portion of + * this software, then the disclaimer below must also be included. + * + * This software is provided as is, without representation or warranty + * of any kind either express or implied, including without limitation + * the implied warranties of merchantability, fitness for a particular + * purpose, or noninfringement. The Regents of the University of + * Michigan shall not be liable for any damages, including special, + * indirect, incidental, or consequential damages, with respect to any + * claim arising out of or in connection with the use of the software, + * even if it has been or is hereafter advised of the possibility of + * such damages. + */ + +#include "pnfs.h" + +#define NFSDBG_FACILITY NFSDBG_PNFS + +/* + * Device ID RCU cache. A device ID is unique per server and layout type. + */ +#define NFS4_DEVICE_ID_HASH_BITS 5 +#define NFS4_DEVICE_ID_HASH_SIZE (1 << NFS4_DEVICE_ID_HASH_BITS) +#define NFS4_DEVICE_ID_HASH_MASK (NFS4_DEVICE_ID_HASH_SIZE - 1) + +static struct hlist_head nfs4_deviceid_cache[NFS4_DEVICE_ID_HASH_SIZE]; +static DEFINE_SPINLOCK(nfs4_deviceid_lock); + +void +nfs4_print_deviceid(const struct nfs4_deviceid *id) +{ + u32 *p = (u32 *)id; + + dprintk("%s: device id= [%x%x%x%x]\n", __func__, + p[0], p[1], p[2], p[3]); +} +EXPORT_SYMBOL_GPL(nfs4_print_deviceid); + +static inline u32 +nfs4_deviceid_hash(const struct nfs4_deviceid *id) +{ + unsigned char *cptr = (unsigned char *)id->data; + unsigned int nbytes = NFS4_DEVICEID4_SIZE; + u32 x = 0; + + while (nbytes--) { + x *= 37; + x += *cptr++; + } + return x & NFS4_DEVICE_ID_HASH_MASK; +} + +static struct nfs4_deviceid_node * +_lookup_deviceid(const struct pnfs_layoutdriver_type *ld, + const struct nfs_client *clp, const struct nfs4_deviceid *id, + long hash) +{ + struct nfs4_deviceid_node *d; + struct hlist_node *n; + + hlist_for_each_entry_rcu(d, n, &nfs4_deviceid_cache[hash], node) + if (d->ld == ld && d->nfs_client == clp && + !memcmp(&d->deviceid, id, sizeof(*id))) { + if (atomic_read(&d->ref)) + return d; + else + continue; + } + return NULL; +} + +/* + * Lookup a deviceid in cache and get a reference count on it if found + * + * @clp nfs_client associated with deviceid + * @id deviceid to look up + */ +struct nfs4_deviceid_node * +_find_get_deviceid(const struct pnfs_layoutdriver_type *ld, + const struct nfs_client *clp, const struct nfs4_deviceid *id, + long hash) +{ + struct nfs4_deviceid_node *d; + + rcu_read_lock(); + d = _lookup_deviceid(ld, clp, id, hash); + if (d && !atomic_inc_not_zero(&d->ref)) + d = NULL; + rcu_read_unlock(); + return d; +} + +struct nfs4_deviceid_node * +nfs4_find_get_deviceid(const struct pnfs_layoutdriver_type *ld, + const struct nfs_client *clp, const struct nfs4_deviceid *id) +{ + return _find_get_deviceid(ld, clp, id, nfs4_deviceid_hash(id)); +} +EXPORT_SYMBOL_GPL(nfs4_find_get_deviceid); + +/* + * Unhash and put deviceid + * + * @clp nfs_client associated with deviceid + * @id the deviceid to unhash + * + * @ret the unhashed node, if found and dereferenced to zero, NULL otherwise. + */ +struct nfs4_deviceid_node * +nfs4_unhash_put_deviceid(const struct pnfs_layoutdriver_type *ld, + const struct nfs_client *clp, const struct nfs4_deviceid *id) +{ + struct nfs4_deviceid_node *d; + + spin_lock(&nfs4_deviceid_lock); + rcu_read_lock(); + d = _lookup_deviceid(ld, clp, id, nfs4_deviceid_hash(id)); + rcu_read_unlock(); + if (!d) { + spin_unlock(&nfs4_deviceid_lock); + return NULL; + } + hlist_del_init_rcu(&d->node); + spin_unlock(&nfs4_deviceid_lock); + synchronize_rcu(); + + /* balance the initial ref set in pnfs_insert_deviceid */ + if (atomic_dec_and_test(&d->ref)) + return d; + + return NULL; +} +EXPORT_SYMBOL_GPL(nfs4_unhash_put_deviceid); + +/* + * Delete a deviceid from cache + * + * @clp struct nfs_client qualifying the deviceid + * @id deviceid to delete + */ +void +nfs4_delete_deviceid(const struct pnfs_layoutdriver_type *ld, + const struct nfs_client *clp, const struct nfs4_deviceid *id) +{ + struct nfs4_deviceid_node *d; + + d = nfs4_unhash_put_deviceid(ld, clp, id); + if (!d) + return; + d->ld->free_deviceid_node(d); +} +EXPORT_SYMBOL_GPL(nfs4_delete_deviceid); + +void +nfs4_init_deviceid_node(struct nfs4_deviceid_node *d, + const struct pnfs_layoutdriver_type *ld, + const struct nfs_client *nfs_client, + const struct nfs4_deviceid *id) +{ + INIT_HLIST_NODE(&d->node); + INIT_HLIST_NODE(&d->tmpnode); + d->ld = ld; + d->nfs_client = nfs_client; + d->deviceid = *id; + atomic_set(&d->ref, 1); +} +EXPORT_SYMBOL_GPL(nfs4_init_deviceid_node); + +/* + * Uniquely initialize and insert a deviceid node into cache + * + * @new new deviceid node + * Note that the caller must set up the following members: + * new->ld + * new->nfs_client + * new->deviceid + * + * @ret the inserted node, if none found, otherwise, the found entry. + */ +struct nfs4_deviceid_node * +nfs4_insert_deviceid_node(struct nfs4_deviceid_node *new) +{ + struct nfs4_deviceid_node *d; + long hash; + + spin_lock(&nfs4_deviceid_lock); + hash = nfs4_deviceid_hash(&new->deviceid); + d = _find_get_deviceid(new->ld, new->nfs_client, &new->deviceid, hash); + if (d) { + spin_unlock(&nfs4_deviceid_lock); + return d; + } + + hlist_add_head_rcu(&new->node, &nfs4_deviceid_cache[hash]); + spin_unlock(&nfs4_deviceid_lock); + atomic_inc(&new->ref); + + return new; +} +EXPORT_SYMBOL_GPL(nfs4_insert_deviceid_node); + +/* + * Dereference a deviceid node and delete it when its reference count drops + * to zero. + * + * @d deviceid node to put + * + * @ret true iff the node was deleted + */ +bool +nfs4_put_deviceid_node(struct nfs4_deviceid_node *d) +{ + if (!atomic_dec_and_lock(&d->ref, &nfs4_deviceid_lock)) + return false; + hlist_del_init_rcu(&d->node); + spin_unlock(&nfs4_deviceid_lock); + synchronize_rcu(); + d->ld->free_deviceid_node(d); + return true; +} +EXPORT_SYMBOL_GPL(nfs4_put_deviceid_node); + +static void +_deviceid_purge_client(const struct nfs_client *clp, long hash) +{ + struct nfs4_deviceid_node *d; + struct hlist_node *n; + HLIST_HEAD(tmp); + + spin_lock(&nfs4_deviceid_lock); + rcu_read_lock(); + hlist_for_each_entry_rcu(d, n, &nfs4_deviceid_cache[hash], node) + if (d->nfs_client == clp && atomic_read(&d->ref)) { + hlist_del_init_rcu(&d->node); + hlist_add_head(&d->tmpnode, &tmp); + } + rcu_read_unlock(); + spin_unlock(&nfs4_deviceid_lock); + + if (hlist_empty(&tmp)) + return; + + synchronize_rcu(); + while (!hlist_empty(&tmp)) { + d = hlist_entry(tmp.first, struct nfs4_deviceid_node, tmpnode); + hlist_del(&d->tmpnode); + if (atomic_dec_and_test(&d->ref)) + d->ld->free_deviceid_node(d); + } +} + +void +nfs4_deviceid_purge_client(const struct nfs_client *clp) +{ + long h; + + if (!(clp->cl_exchange_flags & EXCHGID4_FLAG_USE_PNFS_MDS)) + return; + for (h = 0; h < NFS4_DEVICE_ID_HASH_SIZE; h++) + _deviceid_purge_client(clp, h); +} diff --git a/fs/nfs/read.c b/fs/nfs/read.c index 2bcf0dc306a1..20a7f952e244 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -288,7 +288,9 @@ static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc) atomic_set(&req->wb_complete, requests); BUG_ON(desc->pg_lseg != NULL); - lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_READ, GFP_KERNEL); + lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, + req_offset(req), desc->pg_count, + IOMODE_READ, GFP_KERNEL); ClearPageError(page); offset = 0; nbytes = desc->pg_count; @@ -351,7 +353,9 @@ static int nfs_pagein_one(struct nfs_pageio_descriptor *desc) } req = nfs_list_entry(data->pages.next); if ((!lseg) && list_is_singular(&data->pages)) - lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_READ, GFP_KERNEL); + lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, + req_offset(req), desc->pg_count, + IOMODE_READ, GFP_KERNEL); ret = nfs_read_rpcsetup(req, data, &nfs_read_full_ops, desc->pg_count, 0, lseg); @@ -660,7 +664,6 @@ int nfs_readpages(struct file *filp, struct address_space *mapping, if (ret == 0) goto read_complete; /* all pages were read */ - pnfs_pageio_init_read(&pgio, inode); if (rsize < PAGE_CACHE_SIZE) nfs_pageio_init(&pgio, inode, nfs_pagein_multi, rsize, 0); else diff --git a/fs/nfs/super.c b/fs/nfs/super.c index e288f06d3fa7..ce40e5c568ba 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -63,6 +63,7 @@ #include "iostat.h" #include "internal.h" #include "fscache.h" +#include "pnfs.h" #define NFSDBG_FACILITY NFSDBG_VFS @@ -732,6 +733,28 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt) return 0; } +#ifdef CONFIG_NFS_V4_1 +void show_sessions(struct seq_file *m, struct nfs_server *server) +{ + if (nfs4_has_session(server->nfs_client)) + seq_printf(m, ",sessions"); +} +#else +void show_sessions(struct seq_file *m, struct nfs_server *server) {} +#endif + +#ifdef CONFIG_NFS_V4_1 +void show_pnfs(struct seq_file *m, struct nfs_server *server) +{ + seq_printf(m, ",pnfs="); + if (server->pnfs_curr_ld) + seq_printf(m, "%s", server->pnfs_curr_ld->name); + else + seq_printf(m, "not configured"); +} +#else /* CONFIG_NFS_V4_1 */ +void show_pnfs(struct seq_file *m, struct nfs_server *server) {} +#endif /* CONFIG_NFS_V4_1 */ static int nfs_show_devname(struct seq_file *m, struct vfsmount *mnt) { @@ -792,6 +815,8 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) seq_printf(m, "bm0=0x%x", nfss->attr_bitmask[0]); seq_printf(m, ",bm1=0x%x", nfss->attr_bitmask[1]); seq_printf(m, ",acl=0x%x", nfss->acl_bitmask); + show_sessions(m, nfss); + show_pnfs(m, nfss); } #endif diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 49c715b4ac92..e268e3b23497 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -939,7 +939,9 @@ static int nfs_flush_multi(struct nfs_pageio_descriptor *desc) atomic_set(&req->wb_complete, requests); BUG_ON(desc->pg_lseg); - lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_RW, GFP_NOFS); + lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, + req_offset(req), desc->pg_count, + IOMODE_RW, GFP_NOFS); ClearPageError(page); offset = 0; nbytes = desc->pg_count; @@ -1013,7 +1015,9 @@ static int nfs_flush_one(struct nfs_pageio_descriptor *desc) } req = nfs_list_entry(data->pages.next); if ((!lseg) && list_is_singular(&data->pages)) - lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_RW, GFP_NOFS); + lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, + req_offset(req), desc->pg_count, + IOMODE_RW, GFP_NOFS); if ((desc->pg_ioflags & FLUSH_COND_STABLE) && (desc->pg_moreio || NFS_I(desc->pg_inode)->ncommit)) @@ -1032,8 +1036,6 @@ static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, { size_t wsize = NFS_SERVER(inode)->wsize; - pnfs_pageio_init_write(pgio, inode); - if (wsize < PAGE_CACHE_SIZE) nfs_pageio_init(pgio, inode, nfs_flush_multi, wsize, ioflags); else diff --git a/fs/nfsd/Kconfig b/fs/nfsd/Kconfig index 18b3e8975fe0..fbb2a5ef5817 100644 --- a/fs/nfsd/Kconfig +++ b/fs/nfsd/Kconfig @@ -82,6 +82,7 @@ config NFSD_V4 select NFSD_V3 select FS_POSIX_ACL select SUNRPC_GSS + select CRYPTO help This option enables support in your system's NFS server for version 4 of the NFS protocol (RFC 3530). diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index ad000aeb21a2..b9566e46219f 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -1354,12 +1354,6 @@ exp_pseudoroot(struct svc_rqst *rqstp, struct svc_fh *fhp) if (IS_ERR(exp)) return nfserrno(PTR_ERR(exp)); rv = fh_compose(fhp, exp, exp->ex_path.dentry, NULL); - if (rv) - goto out; - rv = check_nfsd_access(exp, rqstp); - if (rv) - fh_put(fhp); -out: exp_put(exp); return rv; } diff --git a/fs/nfsd/nfs3proc.c b/fs/nfsd/nfs3proc.c index 2247fc91d5e9..9095f3c21df9 100644 --- a/fs/nfsd/nfs3proc.c +++ b/fs/nfsd/nfs3proc.c @@ -245,7 +245,7 @@ nfsd3_proc_create(struct svc_rqst *rqstp, struct nfsd3_createargs *argp, } /* Now create the file and set attributes */ - nfserr = nfsd_create_v3(rqstp, dirfhp, argp->name, argp->len, + nfserr = do_nfsd_create(rqstp, dirfhp, argp->name, argp->len, attr, newfhp, argp->createmode, argp->verf, NULL, NULL); diff --git a/fs/nfsd/nfs3xdr.c b/fs/nfsd/nfs3xdr.c index ad48faca20fc..08c6e36ab2eb 100644 --- a/fs/nfsd/nfs3xdr.c +++ b/fs/nfsd/nfs3xdr.c @@ -842,7 +842,7 @@ out: return rv; } -__be32 *encode_entryplus_baggage(struct nfsd3_readdirres *cd, __be32 *p, const char *name, int namlen) +static __be32 *encode_entryplus_baggage(struct nfsd3_readdirres *cd, __be32 *p, const char *name, int namlen) { struct svc_fh fh; int err; diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index 5fcb1396a7e3..3a6dbd70b34b 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -196,9 +196,9 @@ do_open_lookup(struct svc_rqst *rqstp, struct svc_fh *current_fh, struct nfsd4_o /* * Note: create modes (UNCHECKED,GUARDED...) are the same - * in NFSv4 as in v3. + * in NFSv4 as in v3 except EXCLUSIVE4_1. */ - status = nfsd_create_v3(rqstp, current_fh, open->op_fname.data, + status = do_nfsd_create(rqstp, current_fh, open->op_fname.data, open->op_fname.len, &open->op_iattr, &resfh, open->op_createmode, (u32 *)open->op_verf.data, @@ -403,7 +403,7 @@ nfsd4_putfh(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, cstate->current_fh.fh_handle.fh_size = putfh->pf_fhlen; memcpy(&cstate->current_fh.fh_handle.fh_base, putfh->pf_fhval, putfh->pf_fhlen); - return fh_verify(rqstp, &cstate->current_fh, 0, NFSD_MAY_NOP); + return fh_verify(rqstp, &cstate->current_fh, 0, NFSD_MAY_BYPASS_GSS); } static __be32 @@ -762,6 +762,9 @@ nfsd4_secinfo(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, __be32 err; fh_init(&resfh, NFS4_FHSIZE); + err = fh_verify(rqstp, &cstate->current_fh, S_IFDIR, NFSD_MAY_EXEC); + if (err) + return err; err = nfsd_lookup_dentry(rqstp, &cstate->current_fh, secinfo->si_name, secinfo->si_namelen, &exp, &dentry); @@ -986,6 +989,9 @@ enum nfsd4_op_flags { ALLOWED_WITHOUT_FH = 1 << 0, /* No current filehandle required */ ALLOWED_ON_ABSENT_FS = 1 << 1, /* ops processed on absent fs */ ALLOWED_AS_FIRST_OP = 1 << 2, /* ops reqired first in compound */ + /* For rfc 5661 section 2.6.3.1.1: */ + OP_HANDLES_WRONGSEC = 1 << 3, + OP_IS_PUTFH_LIKE = 1 << 4, }; struct nfsd4_operation { @@ -1031,6 +1037,44 @@ static __be32 nfs41_check_op_ordering(struct nfsd4_compoundargs *args) return nfs_ok; } +static inline struct nfsd4_operation *OPDESC(struct nfsd4_op *op) +{ + return &nfsd4_ops[op->opnum]; +} + +static bool need_wrongsec_check(struct svc_rqst *rqstp) +{ + struct nfsd4_compoundres *resp = rqstp->rq_resp; + struct nfsd4_compoundargs *argp = rqstp->rq_argp; + struct nfsd4_op *this = &argp->ops[resp->opcnt - 1]; + struct nfsd4_op *next = &argp->ops[resp->opcnt]; + struct nfsd4_operation *thisd; + struct nfsd4_operation *nextd; + + thisd = OPDESC(this); + /* + * Most ops check wronsec on our own; only the putfh-like ops + * have special rules. + */ + if (!(thisd->op_flags & OP_IS_PUTFH_LIKE)) + return false; + /* + * rfc 5661 2.6.3.1.1.6: don't bother erroring out a + * put-filehandle operation if we're not going to use the + * result: + */ + if (argp->opcnt == resp->opcnt) + return false; + + nextd = OPDESC(next); + /* + * Rest of 2.6.3.1.1: certain operations will return WRONGSEC + * errors themselves as necessary; others should check for them + * now: + */ + return !(nextd->op_flags & OP_HANDLES_WRONGSEC); +} + /* * COMPOUND call. */ @@ -1108,7 +1152,7 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, goto encode_op; } - opdesc = &nfsd4_ops[op->opnum]; + opdesc = OPDESC(op); if (!cstate->current_fh.fh_dentry) { if (!(opdesc->op_flags & ALLOWED_WITHOUT_FH)) { @@ -1126,6 +1170,9 @@ nfsd4_proc_compound(struct svc_rqst *rqstp, else BUG_ON(op->status == nfs_ok); + if (!op->status && need_wrongsec_check(rqstp)) + op->status = check_nfsd_access(cstate->current_fh.fh_export, rqstp); + encode_op: /* Only from SEQUENCE */ if (resp->cstate.status == nfserr_replay_cache) { @@ -1217,10 +1264,12 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_LOOKUP] = { .op_func = (nfsd4op_func)nfsd4_lookup, + .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_LOOKUP", }, [OP_LOOKUPP] = { .op_func = (nfsd4op_func)nfsd4_lookupp, + .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_LOOKUPP", }, [OP_NVERIFY] = { @@ -1229,6 +1278,7 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_OPEN] = { .op_func = (nfsd4op_func)nfsd4_open, + .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_OPEN", }, [OP_OPEN_CONFIRM] = { @@ -1241,17 +1291,20 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_PUTFH] = { .op_func = (nfsd4op_func)nfsd4_putfh, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_IS_PUTFH_LIKE, .op_name = "OP_PUTFH", }, [OP_PUTPUBFH] = { .op_func = (nfsd4op_func)nfsd4_putrootfh, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_IS_PUTFH_LIKE, .op_name = "OP_PUTPUBFH", }, [OP_PUTROOTFH] = { .op_func = (nfsd4op_func)nfsd4_putrootfh, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_IS_PUTFH_LIKE, .op_name = "OP_PUTROOTFH", }, [OP_READ] = { @@ -1281,15 +1334,18 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_RESTOREFH] = { .op_func = (nfsd4op_func)nfsd4_restorefh, - .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS, + .op_flags = ALLOWED_WITHOUT_FH | ALLOWED_ON_ABSENT_FS + | OP_IS_PUTFH_LIKE, .op_name = "OP_RESTOREFH", }, [OP_SAVEFH] = { .op_func = (nfsd4op_func)nfsd4_savefh, + .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_SAVEFH", }, [OP_SECINFO] = { .op_func = (nfsd4op_func)nfsd4_secinfo, + .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_SECINFO", }, [OP_SETATTR] = { @@ -1353,6 +1409,7 @@ static struct nfsd4_operation nfsd4_ops[] = { }, [OP_SECINFO_NO_NAME] = { .op_func = (nfsd4op_func)nfsd4_secinfo_no_name, + .op_flags = OP_HANDLES_WRONGSEC, .op_name = "OP_SECINFO_NO_NAME", }, }; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 4cf04e11c66c..e98f3c2e9492 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1519,6 +1519,9 @@ nfsd4_create_session(struct svc_rqst *rqstp, bool confirm_me = false; int status = 0; + if (cr_ses->flags & ~SESSION4_FLAG_MASK_A) + return nfserr_inval; + nfs4_lock_state(); unconf = find_unconfirmed_client(&cr_ses->clientid); conf = find_confirmed_client(&cr_ses->clientid); @@ -1637,8 +1640,9 @@ __be32 nfsd4_bind_conn_to_session(struct svc_rqst *rqstp, return nfserr_badsession; status = nfsd4_map_bcts_dir(&bcts->dir); - nfsd4_new_conn(rqstp, cstate->session, bcts->dir); - return nfs_ok; + if (!status) + nfsd4_new_conn(rqstp, cstate->session, bcts->dir); + return status; } static bool nfsd4_compound_in_session(struct nfsd4_session *session, struct nfs4_sessionid *sid) @@ -1725,6 +1729,13 @@ static void nfsd4_sequence_check_conn(struct nfsd4_conn *new, struct nfsd4_sessi return; } +static bool nfsd4_session_too_many_ops(struct svc_rqst *rqstp, struct nfsd4_session *session) +{ + struct nfsd4_compoundargs *args = rqstp->rq_argp; + + return args->opcnt > session->se_fchannel.maxops; +} + __be32 nfsd4_sequence(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, @@ -1753,6 +1764,10 @@ nfsd4_sequence(struct svc_rqst *rqstp, if (!session) goto out; + status = nfserr_too_many_ops; + if (nfsd4_session_too_many_ops(rqstp, session)) + goto out; + status = nfserr_badslot; if (seq->slotid >= session->se_fchannel.maxreqs) goto out; @@ -1808,6 +1823,8 @@ out: __be32 nfsd4_reclaim_complete(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_reclaim_complete *rc) { + int status = 0; + if (rc->rca_one_fs) { if (!cstate->current_fh.fh_dentry) return nfserr_nofilehandle; @@ -1817,9 +1834,14 @@ nfsd4_reclaim_complete(struct svc_rqst *rqstp, struct nfsd4_compound_state *csta */ return nfs_ok; } + nfs4_lock_state(); - if (is_client_expired(cstate->session->se_client)) { - nfs4_unlock_state(); + status = nfserr_complete_already; + if (cstate->session->se_client->cl_firststate) + goto out; + + status = nfserr_stale_clientid; + if (is_client_expired(cstate->session->se_client)) /* * The following error isn't really legal. * But we only get here if the client just explicitly @@ -1827,11 +1849,13 @@ nfsd4_reclaim_complete(struct svc_rqst *rqstp, struct nfsd4_compound_state *csta * error it gets back on an operation for the dead * client. */ - return nfserr_stale_clientid; - } + goto out; + + status = nfs_ok; nfsd4_create_clid_dir(cstate->session->se_client); +out: nfs4_unlock_state(); - return nfs_ok; + return status; } __be32 @@ -2462,7 +2486,7 @@ find_delegation_file(struct nfs4_file *fp, stateid_t *stid) return NULL; } -int share_access_to_flags(u32 share_access) +static int share_access_to_flags(u32 share_access) { share_access &= ~NFS4_SHARE_WANT_MASK; @@ -2882,7 +2906,7 @@ out: return status; } -struct lock_manager nfsd4_manager = { +static struct lock_manager nfsd4_manager = { }; static void diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index c6766af00d98..990181103214 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -424,15 +424,12 @@ nfsd4_decode_access(struct nfsd4_compoundargs *argp, struct nfsd4_access *access static __be32 nfsd4_decode_bind_conn_to_session(struct nfsd4_compoundargs *argp, struct nfsd4_bind_conn_to_session *bcts) { DECODE_HEAD; - u32 dummy; READ_BUF(NFS4_MAX_SESSIONID_LEN + 8); COPYMEM(bcts->sessionid.data, NFS4_MAX_SESSIONID_LEN); READ32(bcts->dir); - /* XXX: Perhaps Tom Tucker could help us figure out how we - * should be using ctsa_use_conn_in_rdma_mode: */ - READ32(dummy); - + /* XXX: skipping ctsa_use_conn_in_rdma_mode. Perhaps Tom Tucker + * could help us figure out we should be using it. */ DECODE_TAIL; } @@ -588,8 +585,6 @@ nfsd4_decode_lockt(struct nfsd4_compoundargs *argp, struct nfsd4_lockt *lockt) READ_BUF(lockt->lt_owner.len); READMEM(lockt->lt_owner.data, lockt->lt_owner.len); - if (argp->minorversion && !zero_clientid(&lockt->lt_clientid)) - return nfserr_inval; DECODE_TAIL; } @@ -3120,7 +3115,7 @@ nfsd4_encode_destroy_session(struct nfsd4_compoundres *resp, int nfserr, return nfserr; } -__be32 +static __be32 nfsd4_encode_sequence(struct nfsd4_compoundres *resp, int nfserr, struct nfsd4_sequence *seq) { diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 1f5eae40f34e..2b1449dd2f49 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -13,6 +13,7 @@ #include <linux/lockd/lockd.h> #include <linux/sunrpc/clnt.h> #include <linux/sunrpc/gss_api.h> +#include <linux/sunrpc/gss_krb5_enctypes.h> #include "idmap.h" #include "nfsd.h" @@ -189,18 +190,10 @@ static struct file_operations export_features_operations = { .release = single_release, }; -#ifdef CONFIG_SUNRPC_GSS +#if defined(CONFIG_SUNRPC_GSS) || defined(CONFIG_SUNRPC_GSS_MODULE) static int supported_enctypes_show(struct seq_file *m, void *v) { - struct gss_api_mech *k5mech; - - k5mech = gss_mech_get_by_name("krb5"); - if (k5mech == NULL) - goto out; - if (k5mech->gm_upcall_enctypes != NULL) - seq_printf(m, k5mech->gm_upcall_enctypes); - gss_mech_put(k5mech); -out: + seq_printf(m, KRB5_SUPPORTED_ENCTYPES); return 0; } @@ -215,7 +208,7 @@ static struct file_operations supported_enctypes_ops = { .llseek = seq_lseek, .release = single_release, }; -#endif /* CONFIG_SUNRPC_GSS */ +#endif /* CONFIG_SUNRPC_GSS or CONFIG_SUNRPC_GSS_MODULE */ extern int nfsd_pool_stats_open(struct inode *inode, struct file *file); extern int nfsd_pool_stats_release(struct inode *inode, struct file *file); @@ -1427,9 +1420,9 @@ static int nfsd_fill_super(struct super_block * sb, void * data, int silent) [NFSD_Versions] = {"versions", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Ports] = {"portlist", &transaction_ops, S_IWUSR|S_IRUGO}, [NFSD_MaxBlkSize] = {"max_block_size", &transaction_ops, S_IWUSR|S_IRUGO}, -#ifdef CONFIG_SUNRPC_GSS +#if defined(CONFIG_SUNRPC_GSS) || defined(CONFIG_SUNRPC_GSS_MODULE) [NFSD_SupportedEnctypes] = {"supported_krb5_enctypes", &supported_enctypes_ops, S_IRUGO}, -#endif /* CONFIG_SUNRPC_GSS */ +#endif /* CONFIG_SUNRPC_GSS or CONFIG_SUNRPC_GSS_MODULE */ #ifdef CONFIG_NFSD_V4 [NFSD_Leasetime] = {"nfsv4leasetime", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Gracetime] = {"nfsv4gracetime", &transaction_ops, S_IWUSR|S_IRUSR}, diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index 55c8e63af0be..90c6aa6d5e0f 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -344,7 +344,7 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access) * which clients virtually always use auth_sys for, * even while using RPCSEC_GSS for NFS. */ - if (access & NFSD_MAY_LOCK) + if (access & NFSD_MAY_LOCK || access & NFSD_MAY_BYPASS_GSS) goto skip_pseudoflavor_check; /* * Clients may expect to be able to use auth_sys during mount, diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 129f3c9f62d5..fd0acca5370a 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -181,16 +181,10 @@ nfsd_lookup_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp, struct svc_export *exp; struct dentry *dparent; struct dentry *dentry; - __be32 err; int host_err; dprintk("nfsd: nfsd_lookup(fh %s, %.*s)\n", SVCFH_fmt(fhp), len,name); - /* Obtain dentry and export. */ - err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_EXEC); - if (err) - return err; - dparent = fhp->fh_dentry; exp = fhp->fh_export; exp_get(exp); @@ -254,6 +248,9 @@ nfsd_lookup(struct svc_rqst *rqstp, struct svc_fh *fhp, const char *name, struct dentry *dentry; __be32 err; + err = fh_verify(rqstp, fhp, S_IFDIR, NFSD_MAY_EXEC); + if (err) + return err; err = nfsd_lookup_dentry(rqstp, fhp, name, len, &exp, &dentry); if (err) return err; @@ -699,7 +696,15 @@ nfsd_access(struct svc_rqst *rqstp, struct svc_fh *fhp, u32 *access, u32 *suppor } #endif /* CONFIG_NFSD_V3 */ +static int nfsd_open_break_lease(struct inode *inode, int access) +{ + unsigned int mode; + if (access & NFSD_MAY_NOT_BREAK_LEASE) + return 0; + mode = (access & NFSD_MAY_WRITE) ? O_WRONLY : O_RDONLY; + return break_lease(inode, mode | O_NONBLOCK); +} /* * Open an existing file or directory. @@ -747,12 +752,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, if (!inode->i_fop) goto out; - /* - * Check to see if there are any leases on this file. - * This may block while leases are broken. - */ - if (!(access & NFSD_MAY_NOT_BREAK_LEASE)) - host_err = break_lease(inode, O_NONBLOCK | ((access & NFSD_MAY_WRITE) ? O_WRONLY : 0)); + host_err = nfsd_open_break_lease(inode, access); if (host_err) /* NOMEM or WOULDBLOCK */ goto out_nfserr; @@ -877,13 +877,11 @@ static __be32 nfsd_vfs_read(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, loff_t offset, struct kvec *vec, int vlen, unsigned long *count) { - struct inode *inode; mm_segment_t oldfs; __be32 err; int host_err; err = nfserr_perm; - inode = file->f_path.dentry->d_inode; if (file->f_op->splice_read && rqstp->rq_splice_ok) { struct splice_desc sd = { @@ -1340,11 +1338,18 @@ out_nfserr: } #ifdef CONFIG_NFSD_V3 + +static inline int nfsd_create_is_exclusive(int createmode) +{ + return createmode == NFS3_CREATE_EXCLUSIVE + || createmode == NFS4_CREATE_EXCLUSIVE4_1; +} + /* - * NFSv3 version of nfsd_create + * NFSv3 and NFSv4 version of nfsd_create */ __be32 -nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, +do_nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, char *fname, int flen, struct iattr *iap, struct svc_fh *resfhp, int createmode, u32 *verifier, int *truncp, int *created) @@ -1396,7 +1401,7 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, if (err) goto out; - if (createmode == NFS3_CREATE_EXCLUSIVE) { + if (nfsd_create_is_exclusive(createmode)) { /* solaris7 gets confused (bugid 4218508) if these have * the high bit set, so just clear the high bits. If this is * ever changed to use different attrs for storing the @@ -1437,6 +1442,11 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, && dchild->d_inode->i_atime.tv_sec == v_atime && dchild->d_inode->i_size == 0 ) break; + case NFS4_CREATE_EXCLUSIVE4_1: + if ( dchild->d_inode->i_mtime.tv_sec == v_mtime + && dchild->d_inode->i_atime.tv_sec == v_atime + && dchild->d_inode->i_size == 0 ) + goto set_attr; /* fallthru */ case NFS3_CREATE_GUARDED: err = nfserr_exist; @@ -1455,7 +1465,7 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, nfsd_check_ignore_resizing(iap); - if (createmode == NFS3_CREATE_EXCLUSIVE) { + if (nfsd_create_is_exclusive(createmode)) { /* Cram the verifier into atime/mtime */ iap->ia_valid = ATTR_MTIME|ATTR_ATIME | ATTR_MTIME_SET|ATTR_ATIME_SET; @@ -1653,8 +1663,10 @@ nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp, if (!dold->d_inode) goto out_drop_write; host_err = nfsd_break_lease(dold->d_inode); - if (host_err) + if (host_err) { + err = nfserrno(host_err); goto out_drop_write; + } host_err = vfs_link(dold, dirp, dnew); if (!host_err) { err = nfserrno(commit_metadata(ffhp)); @@ -2034,7 +2046,7 @@ nfsd_permission(struct svc_rqst *rqstp, struct svc_export *exp, struct inode *inode = dentry->d_inode; int err; - if (acc == NFSD_MAY_NOP) + if ((acc & NFSD_MAY_MASK) == NFSD_MAY_NOP) return 0; #if 0 dprintk("nfsd: permission 0x%x%s%s%s%s%s%s%s mode 0%o%s%s%s\n", diff --git a/fs/nfsd/vfs.h b/fs/nfsd/vfs.h index 9a370a5e36b7..e0bbac04d1dd 100644 --- a/fs/nfsd/vfs.h +++ b/fs/nfsd/vfs.h @@ -17,10 +17,14 @@ #define NFSD_MAY_SATTR 8 #define NFSD_MAY_TRUNC 16 #define NFSD_MAY_LOCK 32 +#define NFSD_MAY_MASK 63 + +/* extra hints to permission and open routines: */ #define NFSD_MAY_OWNER_OVERRIDE 64 #define NFSD_MAY_LOCAL_ACCESS 128 /* IRIX doing local access check on device special file*/ #define NFSD_MAY_BYPASS_GSS_ON_ROOT 256 #define NFSD_MAY_NOT_BREAK_LEASE 512 +#define NFSD_MAY_BYPASS_GSS 1024 #define NFSD_MAY_CREATE (NFSD_MAY_EXEC|NFSD_MAY_WRITE) #define NFSD_MAY_REMOVE (NFSD_MAY_EXEC|NFSD_MAY_WRITE|NFSD_MAY_TRUNC) @@ -54,7 +58,7 @@ __be32 nfsd_create(struct svc_rqst *, struct svc_fh *, int type, dev_t rdev, struct svc_fh *res); #ifdef CONFIG_NFSD_V3 __be32 nfsd_access(struct svc_rqst *, struct svc_fh *, u32 *, u32 *); -__be32 nfsd_create_v3(struct svc_rqst *, struct svc_fh *, +__be32 do_nfsd_create(struct svc_rqst *, struct svc_fh *, char *name, int len, struct iattr *attrs, struct svc_fh *res, int createmode, u32 *verifier, int *truncp, int *created); diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 7eafe468a29c..b2e3ff347620 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -1346,6 +1346,11 @@ static void nilfs_btree_shrink(struct nilfs_bmap *btree, path[level].bp_bh = NULL; } +static void nilfs_btree_nop(struct nilfs_bmap *btree, + struct nilfs_btree_path *path, + int level, __u64 *keyp, __u64 *ptrp) +{ +} static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree, struct nilfs_btree_path *path, @@ -1356,20 +1361,19 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree, struct buffer_head *bh; struct nilfs_btree_node *node, *parent, *sib; __u64 sibptr; - int pindex, level, ncmin, ncmax, ncblk, ret; + int pindex, dindex, level, ncmin, ncmax, ncblk, ret; ret = 0; stats->bs_nblocks = 0; ncmin = NILFS_BTREE_NODE_NCHILDREN_MIN(nilfs_btree_node_size(btree)); ncblk = nilfs_btree_nchildren_per_block(btree); - for (level = NILFS_BTREE_LEVEL_NODE_MIN; + for (level = NILFS_BTREE_LEVEL_NODE_MIN, dindex = path[level].bp_index; level < nilfs_btree_height(btree) - 1; level++) { node = nilfs_btree_get_nonroot_node(path, level); path[level].bp_oldreq.bpr_ptr = - nilfs_btree_node_get_ptr(node, path[level].bp_index, - ncblk); + nilfs_btree_node_get_ptr(node, dindex, ncblk); ret = nilfs_bmap_prepare_end_ptr(btree, &path[level].bp_oldreq, dat); if (ret < 0) @@ -1383,6 +1387,7 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree, parent = nilfs_btree_get_node(btree, path, level + 1, &ncmax); pindex = path[level + 1].bp_index; + dindex = pindex; if (pindex > 0) { /* left sibling */ @@ -1421,6 +1426,14 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree, path[level].bp_sib_bh = bh; path[level].bp_op = nilfs_btree_concat_right; stats->bs_nblocks++; + /* + * When merging right sibling node + * into the current node, pointer to + * the right sibling node must be + * terminated instead. The adjustment + * below is required for that. + */ + dindex = pindex + 1; /* continue; */ } } else { @@ -1431,29 +1444,31 @@ static int nilfs_btree_prepare_delete(struct nilfs_bmap *btree, NILFS_BTREE_ROOT_NCHILDREN_MAX) { path[level].bp_op = nilfs_btree_shrink; stats->bs_nblocks += 2; + level++; + path[level].bp_op = nilfs_btree_nop; + goto shrink_root_child; } else { path[level].bp_op = nilfs_btree_do_delete; stats->bs_nblocks++; + goto out; } - - goto out; - } } + /* child of the root node is deleted */ + path[level].bp_op = nilfs_btree_do_delete; + stats->bs_nblocks++; + +shrink_root_child: node = nilfs_btree_get_root(btree); path[level].bp_oldreq.bpr_ptr = - nilfs_btree_node_get_ptr(node, path[level].bp_index, + nilfs_btree_node_get_ptr(node, dindex, NILFS_BTREE_ROOT_NCHILDREN_MAX); ret = nilfs_bmap_prepare_end_ptr(btree, &path[level].bp_oldreq, dat); if (ret < 0) goto err_out_child_node; - /* child of the root node is deleted */ - path[level].bp_op = nilfs_btree_do_delete; - stats->bs_nblocks++; - /* success */ out: *levelp = level; diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 587f18432832..b9b45fc2903e 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -801,12 +801,7 @@ out_err: int nilfs_permission(struct inode *inode, int mask, unsigned int flags) { - struct nilfs_root *root; - - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - - root = NILFS_I(inode)->i_root; + struct nilfs_root *root = NILFS_I(inode)->i_root; if ((mask & MAY_WRITE) && root && root->cno != NILFS_CPTREE_CURRENT_CNO) return -EROFS; /* snapshot is not writable */ @@ -917,7 +912,7 @@ int nilfs_mark_inode_dirty(struct inode *inode) * construction. This function can be called both as a single operation * and as a part of indivisible file operations. */ -void nilfs_dirty_inode(struct inode *inode) +void nilfs_dirty_inode(struct inode *inode, int flags) { struct nilfs_transaction_info ti; struct nilfs_mdt_info *mdi = NILFS_MDT(inode); diff --git a/fs/nilfs2/namei.c b/fs/nilfs2/namei.c index 1102a5fbb744..546849b3e88f 100644 --- a/fs/nilfs2/namei.c +++ b/fs/nilfs2/namei.c @@ -334,8 +334,6 @@ static int nilfs_rmdir(struct inode *dir, struct dentry *dentry) struct nilfs_transaction_info ti; int err; - dentry_unhash(dentry); - err = nilfs_transaction_begin(dir->i_sb, &ti, 0); if (err) return err; @@ -371,9 +369,6 @@ static int nilfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct nilfs_transaction_info ti; int err; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - err = nilfs_transaction_begin(old_dir->i_sb, &ti, 1); if (unlikely(err)) return err; diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index a9c6a531f80c..f02b9ad43a21 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -269,7 +269,7 @@ int nilfs_load_inode_block(struct inode *inode, struct buffer_head **pbh); extern int nilfs_inode_dirty(struct inode *); int nilfs_set_file_dirty(struct inode *inode, unsigned nr_dirty); extern int nilfs_mark_inode_dirty(struct inode *); -extern void nilfs_dirty_inode(struct inode *); +extern void nilfs_dirty_inode(struct inode *, int flags); int nilfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, __u64 start, __u64 len); diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 141646e88fb5..bb24ab6c282f 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -2573,7 +2573,7 @@ static struct nilfs_sc_info *nilfs_segctor_new(struct super_block *sb, sci->sc_watermark = NILFS_SC_DEFAULT_WATERMARK; if (nilfs->ns_interval) - sci->sc_interval = nilfs->ns_interval; + sci->sc_interval = HZ * nilfs->ns_interval; if (nilfs->ns_watermark) sci->sc_watermark = nilfs->ns_watermark; return sci; diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c index 4c5488468c14..cd9427023d2e 100644 --- a/fs/ocfs2/move_extents.c +++ b/fs/ocfs2/move_extents.c @@ -368,7 +368,7 @@ static int ocfs2_find_victim_alloc_group(struct inode *inode, int *vict_bit, struct buffer_head **ret_bh) { - int ret, i, blocks_per_unit = 1; + int ret, i, bits_per_unit = 0; u64 blkno; char namebuf[40]; @@ -398,14 +398,14 @@ static int ocfs2_find_victim_alloc_group(struct inode *inode, rec = &(cl->cl_recs[0]); if (type == GLOBAL_BITMAP_SYSTEM_INODE) - blocks_per_unit <<= (osb->s_clustersize_bits - - inode->i_sb->s_blocksize_bits); + bits_per_unit = osb->s_clustersize_bits - + inode->i_sb->s_blocksize_bits; /* * 'vict_blkno' was out of the valid range. */ if ((vict_blkno < le64_to_cpu(rec->c_blkno)) || - (vict_blkno >= (le32_to_cpu(ac_dinode->id1.bitmap1.i_total) * - blocks_per_unit))) { + (vict_blkno >= (le32_to_cpu(ac_dinode->id1.bitmap1.i_total) << + bits_per_unit))) { ret = -EINVAL; goto out; } @@ -441,8 +441,8 @@ static int ocfs2_find_victim_alloc_group(struct inode *inode, le16_to_cpu(bg->bg_bits))) { *ret_bh = gd_bh; - *vict_bit = (vict_blkno - blkno) / - blocks_per_unit; + *vict_bit = (vict_blkno - blkno) >> + bits_per_unit; mlog(0, "find the victim group: #%llu, " "total_bits: %u, vict_bit: %u\n", blkno, le16_to_cpu(bg->bg_bits), @@ -472,12 +472,24 @@ static int ocfs2_validate_and_adjust_move_goal(struct inode *inode, int ret, goal_bit = 0; struct buffer_head *gd_bh = NULL; - struct ocfs2_group_desc *bg; + struct ocfs2_group_desc *bg = NULL; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); int c_to_b = 1 << (osb->s_clustersize_bits - inode->i_sb->s_blocksize_bits); /* + * make goal become cluster aligned. + */ + range->me_goal = ocfs2_block_to_cluster_start(inode->i_sb, + range->me_goal); + /* + * moving goal is not allowd to start with a group desc blok(#0 blk) + * let's compromise to the latter cluster. + */ + if (range->me_goal == le64_to_cpu(bg->bg_blkno)) + range->me_goal += c_to_b; + + /* * validate goal sits within global_bitmap, and return the victim * group desc */ @@ -491,19 +503,6 @@ static int ocfs2_validate_and_adjust_move_goal(struct inode *inode, bg = (struct ocfs2_group_desc *)gd_bh->b_data; /* - * make goal become cluster aligned. - */ - if (range->me_goal % c_to_b) - range->me_goal = range->me_goal / c_to_b * c_to_b; - - /* - * moving goal is not allowd to start with a group desc blok(#0 blk) - * let's compromise to the latter cluster. - */ - if (range->me_goal == le64_to_cpu(bg->bg_blkno)) - range->me_goal += c_to_b; - - /* * movement is not gonna cross two groups. */ if ((le16_to_cpu(bg->bg_bits) - goal_bit) * osb->s_clustersize < diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index cdbaf5e97308..56f61027236b 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1072,7 +1072,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) sb->s_magic = OCFS2_SUPER_MAGIC; - sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | + sb->s_flags = (sb->s_flags & ~(MS_POSIXACL | MS_NOSEC)) | ((osb->s_mount_opt & OCFS2_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); /* Hard readonly mode only if: bdev_read_only, MS_RDONLY, diff --git a/fs/omfs/dir.c b/fs/omfs/dir.c index c368360c35a1..3b8d3979e03b 100644 --- a/fs/omfs/dir.c +++ b/fs/omfs/dir.c @@ -241,11 +241,9 @@ static int omfs_remove(struct inode *dir, struct dentry *dentry) int ret; - if (S_ISDIR(inode->i_mode)) { - dentry_unhash(dentry); - if (!omfs_dir_is_empty(inode)) - return -ENOTEMPTY; - } + if (S_ISDIR(inode->i_mode) && + !omfs_dir_is_empty(inode)) + return -ENOTEMPTY; ret = omfs_delete_entry(dentry); if (ret) @@ -382,9 +380,6 @@ static int omfs_rename(struct inode *old_dir, struct dentry *old_dentry, int err; if (new_inode) { - if (S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - /* overwriting existing file/dir */ err = omfs_remove(new_dir, new_dentry); if (err) diff --git a/fs/omfs/file.c b/fs/omfs/file.c index d738a7e493dd..2c6d95257a4d 100644 --- a/fs/omfs/file.c +++ b/fs/omfs/file.c @@ -4,7 +4,6 @@ * Released under GPL v2. */ -#include <linux/version.h> #include <linux/module.h> #include <linux/fs.h> #include <linux/buffer_head.h> diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 8ed4d3433199..d545e97d99c3 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -255,11 +255,7 @@ ssize_t part_discard_alignment_show(struct device *dev, struct device_attribute *attr, char *buf) { struct hd_struct *p = dev_to_part(dev); - struct gendisk *disk = dev_to_disk(dev); - - return sprintf(buf, "%u\n", - queue_limit_discard_alignment(&disk->queue->limits, - p->start_sect)); + return sprintf(buf, "%u\n", p->discard_alignment); } ssize_t part_stat_show(struct device *dev, @@ -453,6 +449,8 @@ struct hd_struct *add_partition(struct gendisk *disk, int partno, p->start_sect = start; p->alignment_offset = queue_limit_alignment_offset(&disk->queue->limits, start); + p->discard_alignment = + queue_limit_discard_alignment(&disk->queue->limits, start); p->nr_sects = len; p->partno = partno; p->policy = get_disk_ro(disk); diff --git a/fs/proc/base.c b/fs/proc/base.c index 4ede550517a6..8a84210ca080 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -83,6 +83,9 @@ #include <linux/pid_namespace.h> #include <linux/fs_struct.h> #include <linux/slab.h> +#ifdef CONFIG_HARDWALL +#include <asm/hardwall.h> +#endif #include "internal.h" /* NOTE: @@ -2166,11 +2169,7 @@ static const struct file_operations proc_fd_operations = { */ static int proc_fd_permission(struct inode *inode, int mask, unsigned int flags) { - int rv; - - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - rv = generic_permission(inode, mask, flags, NULL); + int rv = generic_permission(inode, mask, flags, NULL); if (rv == 0) return 0; if (task_pid(current) == proc_pid(inode)) @@ -2842,6 +2841,9 @@ static const struct pid_entry tgid_base_stuff[] = { #ifdef CONFIG_TASK_IO_ACCOUNTING INF("io", S_IRUGO, proc_tgid_io_accounting), #endif +#ifdef CONFIG_HARDWALL + INF("hardwall", S_IRUGO, proc_pid_hardwall), +#endif }; static int proc_tgid_base_readdir(struct file * filp, @@ -3181,6 +3183,9 @@ static const struct pid_entry tid_base_stuff[] = { #ifdef CONFIG_TASK_IO_ACCOUNTING INF("io", S_IRUGO, proc_tid_io_accounting), #endif +#ifdef CONFIG_HARDWALL + INF("hardwall", S_IRUGO, proc_pid_hardwall), +#endif }; static int proc_tid_base_readdir(struct file * filp, diff --git a/fs/proc/namespaces.c b/fs/proc/namespaces.c index 781dec5bd682..be177f702acb 100644 --- a/fs/proc/namespaces.c +++ b/fs/proc/namespaces.c @@ -38,18 +38,21 @@ static struct dentry *proc_ns_instantiate(struct inode *dir, struct inode *inode; struct proc_inode *ei; struct dentry *error = ERR_PTR(-ENOENT); + void *ns; inode = proc_pid_make_inode(dir->i_sb, task); if (!inode) goto out; + ns = ns_ops->get(task); + if (!ns) + goto out_iput; + ei = PROC_I(inode); inode->i_mode = S_IFREG|S_IRUSR; inode->i_fop = &ns_file_operations; ei->ns_ops = ns_ops; - ei->ns = ns_ops->get(task); - if (!ei->ns) - goto out_iput; + ei->ns = ns; dentry->d_op = &pid_dentry_operations; d_add(dentry, inode); diff --git a/fs/proc/proc_sysctl.c b/fs/proc/proc_sysctl.c index f50133c11c24..d167de365a8d 100644 --- a/fs/proc/proc_sysctl.c +++ b/fs/proc/proc_sysctl.c @@ -304,9 +304,6 @@ static int proc_sys_permission(struct inode *inode, int mask,unsigned int flags) struct ctl_table *table; int error; - if (flags & IPERM_FLAG_RCU) - return -ECHILD; - /* Executable files are not allowed under /proc/sys/ */ if ((mask & MAY_EXEC) && S_ISREG(inode->i_mode)) return -EACCES; diff --git a/fs/proc/root.c b/fs/proc/root.c index a9000e9cfee5..d6c3b416529b 100644 --- a/fs/proc/root.c +++ b/fs/proc/root.c @@ -28,11 +28,12 @@ static int proc_test_super(struct super_block *sb, void *data) static int proc_set_super(struct super_block *sb, void *data) { - struct pid_namespace *ns; - - ns = (struct pid_namespace *)data; - sb->s_fs_info = get_pid_ns(ns); - return set_anon_super(sb, NULL); + int err = set_anon_super(sb, NULL); + if (!err) { + struct pid_namespace *ns = (struct pid_namespace *)data; + sb->s_fs_info = get_pid_ns(ns); + } + return err; } static struct dentry *proc_mount(struct file_system_type *fs_type, diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c index 76c8164d5651..118662690cdf 100644 --- a/fs/reiserfs/namei.c +++ b/fs/reiserfs/namei.c @@ -831,8 +831,6 @@ static int reiserfs_rmdir(struct inode *dir, struct dentry *dentry) INITIALIZE_PATH(path); struct reiserfs_dir_entry de; - dentry_unhash(dentry); - /* we will be doing 2 balancings and update 2 stat data, we change quotas * of the owner of the directory and of the owner of the parent directory. * The quota structure is possibly deleted only on last iput => outside @@ -1227,9 +1225,6 @@ static int reiserfs_rename(struct inode *old_dir, struct dentry *old_dentry, unsigned long savelink = 1; struct timespec ctime; - if (new_dentry->d_inode && S_ISDIR(new_dentry->d_inode->i_mode)) - dentry_unhash(new_dentry); - /* three balancings: (1) old name removal, (2) new name insertion and (3) maybe "save" link insertion stat data updates: (1) old directory, diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index b216ff6be1c9..aa91089162cb 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -568,7 +568,7 @@ static void destroy_inodecache(void) } /* we don't mark inodes dirty, we just log them */ -static void reiserfs_dirty_inode(struct inode *inode) +static void reiserfs_dirty_inode(struct inode *inode, int flags) { struct reiserfs_transaction_handle th; diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 50f1abccd1cd..d78089690965 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -98,7 +98,6 @@ static int xattr_rmdir(struct inode *dir, struct dentry *dentry) reiserfs_mutex_lock_nested_safe(&dentry->d_inode->i_mutex, I_MUTEX_CHILD, dir->i_sb); - dentry_unhash(dentry); error = dir->i_op->rmdir(dir, dentry); if (!error) dentry->d_inode->i_flags |= S_DEAD; @@ -955,8 +954,6 @@ static int xattr_mount_check(struct super_block *s) int reiserfs_permission(struct inode *inode, int mask, unsigned int flags) { - if (flags & IPERM_FLAG_RCU) - return -ECHILD; /* * We don't do permission checks on the internal objects. * Permissions are determined by the "owning" object. diff --git a/fs/romfs/mmap-nommu.c b/fs/romfs/mmap-nommu.c index f0511e816967..eed99428f104 100644 --- a/fs/romfs/mmap-nommu.c +++ b/fs/romfs/mmap-nommu.c @@ -27,14 +27,18 @@ static unsigned long romfs_get_unmapped_area(struct file *file, { struct inode *inode = file->f_mapping->host; struct mtd_info *mtd = inode->i_sb->s_mtd; - unsigned long isize, offset; + unsigned long isize, offset, maxpages, lpages; if (!mtd) goto cant_map_directly; + /* the mapping mustn't extend beyond the EOF */ + lpages = (len + PAGE_SIZE - 1) >> PAGE_SHIFT; isize = i_size_read(inode); offset = pgoff << PAGE_SHIFT; - if (offset > isize || len > isize || offset > isize - len) + + maxpages = (isize + PAGE_SIZE - 1) >> PAGE_SHIFT; + if ((pgoff >= maxpages) || (maxpages - pgoff < lpages)) return (unsigned long) -EINVAL; /* we need to call down to the MTD layer to do the actual mapping */ diff --git a/fs/squashfs/export.c b/fs/squashfs/export.c index 730c56248c9b..5e1101ff276f 100644 --- a/fs/squashfs/export.c +++ b/fs/squashfs/export.c @@ -147,7 +147,7 @@ __le64 *squashfs_read_inode_lookup_table(struct super_block *sb, * table[0] points to the first inode lookup table metadata block, * this should be less than lookup_table_start */ - if (!IS_ERR(table) && table[0] >= lookup_table_start) { + if (!IS_ERR(table) && le64_to_cpu(table[0]) >= lookup_table_start) { kfree(table); return ERR_PTR(-EINVAL); } diff --git a/fs/squashfs/fragment.c b/fs/squashfs/fragment.c index 1516a6490bfb..0ed6edbc5c71 100644 --- a/fs/squashfs/fragment.c +++ b/fs/squashfs/fragment.c @@ -90,7 +90,7 @@ __le64 *squashfs_read_fragment_index_table(struct super_block *sb, * table[0] points to the first fragment table metadata block, this * should be less than fragment_table_start */ - if (!IS_ERR(table) && table[0] >= fragment_table_start) { + if (!IS_ERR(table) && le64_to_cpu(table[0]) >= fragment_table_start) { kfree(table); return ERR_PTR(-EINVAL); } diff --git a/fs/squashfs/id.c b/fs/squashfs/id.c index a70858e0fb44..d38ea3dab951 100644 --- a/fs/squashfs/id.c +++ b/fs/squashfs/id.c @@ -93,7 +93,7 @@ __le64 *squashfs_read_id_index_table(struct super_block *sb, * table[0] points to the first id lookup table metadata block, this * should be less than id_table_start */ - if (!IS_ERR(table) && table[0] >= id_table_start) { + if (!IS_ERR(table) && le64_to_cpu(table[0]) >= id_table_start) { kfree(table); return ERR_PTR(-EINVAL); } diff --git a/fs/squashfs/super.c b/fs/squashfs/super.c index 6f26abee3597..7438850c62d0 100644 --- a/fs/squashfs/super.c +++ b/fs/squashfs/super.c @@ -245,7 +245,7 @@ allocate_id_index_table: msblk->id_table = NULL; goto failed_mount; } - next_table = msblk->id_table[0]; + next_table = le64_to_cpu(msblk->id_table[0]); /* Handle inode lookup table */ lookup_table_start = le64_to_cpu(sblk->lookup_table_start); @@ -261,7 +261,7 @@ allocate_id_index_table: msblk->inode_lookup_table = NULL; goto failed_mount; } - next_table = msblk->inode_lookup_table[0]; + next_table = le64_to_cpu(msblk->inode_lookup_table[0]); sb->s_export_op = &squashfs_export_ops; @@ -286,7 +286,7 @@ handle_fragments: msblk->fragment_index = NULL; goto failed_mount; } - next_table = msblk->fragment_index[0]; + next_table = le64_to_cpu(msblk->fragment_index[0]); check_directory_table: /* Sanity check directory_table */ diff --git a/fs/super.c b/fs/super.c index c75593953c52..ab3d672db0de 100644 --- a/fs/super.c +++ b/fs/super.c @@ -822,7 +822,7 @@ struct dentry *mount_bdev(struct file_system_type *fs_type, } else { char b[BDEVNAME_SIZE]; - s->s_flags = flags; + s->s_flags = flags | MS_NOSEC; s->s_mode = mode; strlcpy(s->s_id, bdevname(bdev, b), sizeof(s->s_id)); sb_set_blocksize(s, block_size(bdev)); diff --git a/fs/sysfs/mount.c b/fs/sysfs/mount.c index 266895783b47..e34f0d99ea4e 100644 --- a/fs/sysfs/mount.c +++ b/fs/sysfs/mount.c @@ -95,6 +95,14 @@ static int sysfs_set_super(struct super_block *sb, void *data) return error; } +static void free_sysfs_super_info(struct sysfs_super_info *info) +{ + int type; + for (type = KOBJ_NS_TYPE_NONE; type < KOBJ_NS_TYPES; type++) + kobj_ns_drop(type, info->ns[type]); + kfree(info); +} + static struct dentry *sysfs_mount(struct file_system_type *fs_type, int flags, const char *dev_name, void *data) { @@ -108,11 +116,11 @@ static struct dentry *sysfs_mount(struct file_system_type *fs_type, return ERR_PTR(-ENOMEM); for (type = KOBJ_NS_TYPE_NONE; type < KOBJ_NS_TYPES; type++) - info->ns[type] = kobj_ns_current(type); + info->ns[type] = kobj_ns_grab_current(type); sb = sget(fs_type, sysfs_test_super, sysfs_set_super, info); if (IS_ERR(sb) || sb->s_fs_info != info) - kfree(info); + free_sysfs_super_info(info); if (IS_ERR(sb)) return ERR_CAST(sb); if (!sb->s_root) { @@ -131,12 +139,11 @@ static struct dentry *sysfs_mount(struct file_system_type *fs_type, static void sysfs_kill_sb(struct super_block *sb) { struct sysfs_super_info *info = sysfs_info(sb); - /* Remove the superblock from fs_supers/s_instances * so we can't find it, before freeing sysfs_super_info. */ kill_anon_super(sb); - kfree(info); + free_sysfs_super_info(info); } static struct file_system_type sysfs_fs_type = { @@ -145,28 +152,6 @@ static struct file_system_type sysfs_fs_type = { .kill_sb = sysfs_kill_sb, }; -void sysfs_exit_ns(enum kobj_ns_type type, const void *ns) -{ - struct super_block *sb; - - mutex_lock(&sysfs_mutex); - spin_lock(&sb_lock); - list_for_each_entry(sb, &sysfs_fs_type.fs_supers, s_instances) { - struct sysfs_super_info *info = sysfs_info(sb); - /* - * If we see a superblock on the fs_supers/s_instances - * list the unmount has not completed and sb->s_fs_info - * points to a valid struct sysfs_super_info. - */ - /* Ignore superblocks with the wrong ns */ - if (info->ns[type] != ns) - continue; - info->ns[type] = NULL; - } - spin_unlock(&sb_lock); - mutex_unlock(&sysfs_mutex); -} - int __init sysfs_init(void) { int err = -ENOMEM; diff --git a/fs/sysfs/sysfs.h b/fs/sysfs/sysfs.h index 3d28af31d863..2ed2404f3113 100644 --- a/fs/sysfs/sysfs.h +++ b/fs/sysfs/sysfs.h @@ -136,7 +136,7 @@ struct sysfs_addrm_cxt { * instance). */ struct sysfs_super_info { - const void *ns[KOBJ_NS_TYPES]; + void *ns[KOBJ_NS_TYPES]; }; #define sysfs_info(SB) ((struct sysfs_super_info *)(SB->s_fs_info)) extern struct sysfs_dirent sysfs_root; diff --git a/fs/sysv/namei.c b/fs/sysv/namei.c index e2cc6756f3b1..e474fbcf8bde 100644 --- a/fs/sysv/namei.c +++ b/fs/sysv/namei.c @@ -196,8 +196,6 @@ static int sysv_rmdir(struct inode * dir, struct dentry * dentry) struct inode *inode = dentry->d_inode; int err = -ENOTEMPTY; - dentry_unhash(dentry); - if (sysv_empty_dir(inode)) { err = sysv_unlink(dir, dentry); if (!err) { @@ -224,9 +222,6 @@ static int sysv_rename(struct inode * old_dir, struct dentry * old_dentry, struct sysv_dir_entry * old_de; int err = -ENOENT; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - old_de = sysv_find_entry(old_dentry, &old_page); if (!old_de) goto out; diff --git a/fs/timerfd.c b/fs/timerfd.c index f67acbdda5e8..dffeb3795af1 100644 --- a/fs/timerfd.c +++ b/fs/timerfd.c @@ -61,7 +61,9 @@ static enum hrtimer_restart timerfd_tmrproc(struct hrtimer *htmr) /* * Called when the clock was set to cancel the timers in the cancel - * list. + * list. This will wake up processes waiting on these timers. The + * wake-up requires ctx->ticks to be non zero, therefore we increment + * it before calling wake_up_locked(). */ void timerfd_clock_was_set(void) { @@ -76,6 +78,7 @@ void timerfd_clock_was_set(void) spin_lock_irqsave(&ctx->wqh.lock, flags); if (ctx->moffs.tv64 != moffs.tv64) { ctx->moffs.tv64 = KTIME_MAX; + ctx->ticks++; wake_up_locked(&ctx->wqh); } spin_unlock_irqrestore(&ctx->wqh.lock, flags); diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c index c2b80943560d..ef5abd38f0bf 100644 --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -656,8 +656,6 @@ static int ubifs_rmdir(struct inode *dir, struct dentry *dentry) struct ubifs_inode *dir_ui = ubifs_inode(dir); struct ubifs_budget_req req = { .mod_dent = 1, .dirtied_ino = 2 }; - dentry_unhash(dentry); - /* * Budget request settings: deletion direntry, deletion inode and * changing the parent inode. If budgeting fails, go ahead anyway @@ -978,9 +976,6 @@ static int ubifs_rename(struct inode *old_dir, struct dentry *old_dentry, .dirtied_ino_d = ALIGN(old_inode_ui->data_len, 8) }; struct timespec time; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - /* * Budget request settings: deletion direntry, new direntry, removing * the old inode, and changing old and new parent directory inodes. diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c index 166951e0dcd3..3be645e012c9 100644 --- a/fs/ubifs/io.c +++ b/fs/ubifs/io.c @@ -581,6 +581,7 @@ int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len) ubifs_assert(wbuf->size % c->min_io_size == 0); ubifs_assert(mutex_is_locked(&wbuf->io_mutex)); ubifs_assert(!c->ro_media && !c->ro_mount); + ubifs_assert(!c->space_fixup); if (c->leb_size - wbuf->offs >= c->max_write_size) ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size)); @@ -759,6 +760,7 @@ int ubifs_write_node(struct ubifs_info *c, void *buf, int len, int lnum, ubifs_assert(lnum >= 0 && lnum < c->leb_cnt && offs >= 0); ubifs_assert(offs % c->min_io_size == 0 && offs < c->leb_size); ubifs_assert(!c->ro_media && !c->ro_mount); + ubifs_assert(!c->space_fixup); if (c->ro_error) return -EROFS; diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c index 34b1679e6e3a..cef0460f4c54 100644 --- a/fs/ubifs/journal.c +++ b/fs/ubifs/journal.c @@ -669,6 +669,7 @@ out_free: out_release: release_head(c, BASEHD); + kfree(dent); out_ro: ubifs_ro_mode(c, err); if (last_reference) diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c index bd644bf587a8..a5422fffbd69 100644 --- a/fs/ubifs/orphan.c +++ b/fs/ubifs/orphan.c @@ -674,7 +674,7 @@ static int kill_orphans(struct ubifs_info *c) if (IS_ERR(sleb)) { if (PTR_ERR(sleb) == -EUCLEAN) sleb = ubifs_recover_leb(c, lnum, 0, - c->sbuf, 0); + c->sbuf, -1); if (IS_ERR(sleb)) { err = PTR_ERR(sleb); break; diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c index 731d9e2e7b50..783d8e0beb76 100644 --- a/fs/ubifs/recovery.c +++ b/fs/ubifs/recovery.c @@ -564,19 +564,15 @@ static int fix_unclean_leb(struct ubifs_info *c, struct ubifs_scan_leb *sleb, } /** - * drop_last_node - drop the last node or group of nodes. + * drop_last_group - drop the last group of nodes. * @sleb: scanned LEB information * @offs: offset of dropped nodes is returned here - * @grouped: non-zero if whole group of nodes have to be dropped * * This is a helper function for 'ubifs_recover_leb()' which drops the last - * node of the scanned LEB or the last group of nodes if @grouped is not zero. - * This function returns %1 if a node was dropped and %0 otherwise. + * group of nodes of the scanned LEB. */ -static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped) +static void drop_last_group(struct ubifs_scan_leb *sleb, int *offs) { - int dropped = 0; - while (!list_empty(&sleb->nodes)) { struct ubifs_scan_node *snod; struct ubifs_ch *ch; @@ -585,17 +581,40 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped) list); ch = snod->node; if (ch->group_type != UBIFS_IN_NODE_GROUP) - return dropped; - dbg_rcvry("dropping node at %d:%d", sleb->lnum, snod->offs); + break; + + dbg_rcvry("dropping grouped node at %d:%d", + sleb->lnum, snod->offs); + *offs = snod->offs; + list_del(&snod->list); + kfree(snod); + sleb->nodes_cnt -= 1; + } +} + +/** + * drop_last_node - drop the last node. + * @sleb: scanned LEB information + * @offs: offset of dropped nodes is returned here + * @grouped: non-zero if whole group of nodes have to be dropped + * + * This is a helper function for 'ubifs_recover_leb()' which drops the last + * node of the scanned LEB. + */ +static void drop_last_node(struct ubifs_scan_leb *sleb, int *offs) +{ + struct ubifs_scan_node *snod; + + if (!list_empty(&sleb->nodes)) { + snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node, + list); + + dbg_rcvry("dropping last node at %d:%d", sleb->lnum, snod->offs); *offs = snod->offs; list_del(&snod->list); kfree(snod); sleb->nodes_cnt -= 1; - dropped = 1; - if (!grouped) - break; } - return dropped; } /** @@ -604,7 +623,8 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped) * @lnum: LEB number * @offs: offset * @sbuf: LEB-sized buffer to use - * @grouped: nodes may be grouped for recovery + * @jhead: journal head number this LEB belongs to (%-1 if the LEB does not + * belong to any journal head) * * This function does a scan of a LEB, but caters for errors that might have * been caused by the unclean unmount from which we are attempting to recover. @@ -612,13 +632,14 @@ static int drop_last_node(struct ubifs_scan_leb *sleb, int *offs, int grouped) * found, and a negative error code in case of failure. */ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum, - int offs, void *sbuf, int grouped) + int offs, void *sbuf, int jhead) { int ret = 0, err, len = c->leb_size - offs, start = offs, min_io_unit; + int grouped = jhead == -1 ? 0 : c->jheads[jhead].grouped; struct ubifs_scan_leb *sleb; void *buf = sbuf + offs; - dbg_rcvry("%d:%d", lnum, offs); + dbg_rcvry("%d:%d, jhead %d, grouped %d", lnum, offs, jhead, grouped); sleb = ubifs_start_scan(c, lnum, offs, sbuf); if (IS_ERR(sleb)) @@ -635,7 +656,7 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum, * Scan quietly until there is an error from which we cannot * recover */ - ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 0); + ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 1); if (ret == SCANNED_A_NODE) { /* A valid node, and not a padding node */ struct ubifs_ch *ch = buf; @@ -695,59 +716,62 @@ struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum, * If nodes are grouped, always drop the incomplete group at * the end. */ - drop_last_node(sleb, &offs, 1); + drop_last_group(sleb, &offs); - /* - * While we are in the middle of the same min. I/O unit keep dropping - * nodes. So basically, what we want is to make sure that the last min. - * I/O unit where we saw the corruption is dropped completely with all - * the uncorrupted node which may possibly sit there. - * - * In other words, let's name the min. I/O unit where the corruption - * starts B, and the previous min. I/O unit A. The below code tries to - * deal with a situation when half of B contains valid nodes or the end - * of a valid node, and the second half of B contains corrupted data or - * garbage. This means that UBIFS had been writing to B just before the - * power cut happened. I do not know how realistic is this scenario - * that half of the min. I/O unit had been written successfully and the - * other half not, but this is possible in our 'failure mode emulation' - * infrastructure at least. - * - * So what is the problem, why we need to drop those nodes? Whey can't - * we just clean-up the second half of B by putting a padding node - * there? We can, and this works fine with one exception which was - * reproduced with power cut emulation testing and happens extremely - * rarely. The description follows, but it is worth noting that that is - * only about the GC head, so we could do this trick only if the bud - * belongs to the GC head, but it does not seem to be worth an - * additional "if" statement. - * - * So, imagine the file-system is full, we run GC which is moving valid - * nodes from LEB X to LEB Y (obviously, LEB Y is the current GC head - * LEB). The @c->gc_lnum is -1, which means that GC will retain LEB X - * and will try to continue. Imagine that LEB X is currently the - * dirtiest LEB, and the amount of used space in LEB Y is exactly the - * same as amount of free space in LEB X. - * - * And a power cut happens when nodes are moved from LEB X to LEB Y. We - * are here trying to recover LEB Y which is the GC head LEB. We find - * the min. I/O unit B as described above. Then we clean-up LEB Y by - * padding min. I/O unit. And later 'ubifs_rcvry_gc_commit()' function - * fails, because it cannot find a dirty LEB which could be GC'd into - * LEB Y! Even LEB X does not match because the amount of valid nodes - * there does not fit the free space in LEB Y any more! And this is - * because of the padding node which we added to LEB Y. The - * user-visible effect of this which I once observed and analysed is - * that we cannot mount the file-system with -ENOSPC error. - * - * So obviously, to make sure that situation does not happen we should - * free min. I/O unit B in LEB Y completely and the last used min. I/O - * unit in LEB Y should be A. This is basically what the below code - * tries to do. - */ - while (min_io_unit == round_down(offs, c->min_io_size) && - min_io_unit != offs && - drop_last_node(sleb, &offs, grouped)); + if (jhead == GCHD) { + /* + * If this LEB belongs to the GC head then while we are in the + * middle of the same min. I/O unit keep dropping nodes. So + * basically, what we want is to make sure that the last min. + * I/O unit where we saw the corruption is dropped completely + * with all the uncorrupted nodes which may possibly sit there. + * + * In other words, let's name the min. I/O unit where the + * corruption starts B, and the previous min. I/O unit A. The + * below code tries to deal with a situation when half of B + * contains valid nodes or the end of a valid node, and the + * second half of B contains corrupted data or garbage. This + * means that UBIFS had been writing to B just before the power + * cut happened. I do not know how realistic is this scenario + * that half of the min. I/O unit had been written successfully + * and the other half not, but this is possible in our 'failure + * mode emulation' infrastructure at least. + * + * So what is the problem, why we need to drop those nodes? Why + * can't we just clean-up the second half of B by putting a + * padding node there? We can, and this works fine with one + * exception which was reproduced with power cut emulation + * testing and happens extremely rarely. + * + * Imagine the file-system is full, we run GC which starts + * moving valid nodes from LEB X to LEB Y (obviously, LEB Y is + * the current GC head LEB). The @c->gc_lnum is -1, which means + * that GC will retain LEB X and will try to continue. Imagine + * that LEB X is currently the dirtiest LEB, and the amount of + * used space in LEB Y is exactly the same as amount of free + * space in LEB X. + * + * And a power cut happens when nodes are moved from LEB X to + * LEB Y. We are here trying to recover LEB Y which is the GC + * head LEB. We find the min. I/O unit B as described above. + * Then we clean-up LEB Y by padding min. I/O unit. And later + * 'ubifs_rcvry_gc_commit()' function fails, because it cannot + * find a dirty LEB which could be GC'd into LEB Y! Even LEB X + * does not match because the amount of valid nodes there does + * not fit the free space in LEB Y any more! And this is + * because of the padding node which we added to LEB Y. The + * user-visible effect of this which I once observed and + * analysed is that we cannot mount the file-system with + * -ENOSPC error. + * + * So obviously, to make sure that situation does not happen we + * should free min. I/O unit B in LEB Y completely and the last + * used min. I/O unit in LEB Y should be A. This is basically + * what the below code tries to do. + */ + while (offs > min_io_unit) + drop_last_node(sleb, &offs); + } buf = sbuf + offs; len = c->leb_size - offs; @@ -881,7 +905,7 @@ struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum, } ubifs_scan_destroy(sleb); } - return ubifs_recover_leb(c, lnum, offs, sbuf, 0); + return ubifs_recover_leb(c, lnum, offs, sbuf, -1); } /** diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c index 6617280d1679..5e97161ce4d3 100644 --- a/fs/ubifs/replay.c +++ b/fs/ubifs/replay.c @@ -557,8 +557,7 @@ static int replay_bud(struct ubifs_info *c, struct bud_entry *b) * these LEBs could possibly be written to at the power cut * time. */ - sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf, - b->bud->jhead != GCHD); + sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf, b->bud->jhead); else sleb = ubifs_scan(c, lnum, offs, c->sbuf, 0); if (IS_ERR(sleb)) diff --git a/fs/ubifs/shrinker.c b/fs/ubifs/shrinker.c index 46961c003236..9e1d05666fed 100644 --- a/fs/ubifs/shrinker.c +++ b/fs/ubifs/shrinker.c @@ -277,13 +277,18 @@ static int kick_a_thread(void) return 0; } -int ubifs_shrinker(struct shrinker *shrink, int nr, gfp_t gfp_mask) +int ubifs_shrinker(struct shrinker *shrink, struct shrink_control *sc) { + int nr = sc->nr_to_scan; int freed, contention = 0; long clean_zn_cnt = atomic_long_read(&ubifs_clean_zn_cnt); if (nr == 0) - return clean_zn_cnt; + /* + * Due to the way UBIFS updates the clean znode counter it may + * temporarily be negative. + */ + return clean_zn_cnt >= 0 ? clean_zn_cnt : 1; if (!clean_zn_cnt) { /* diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 6db0bdaa9f74..529be0582029 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -382,7 +382,7 @@ done: end_writeback(inode); } -static void ubifs_dirty_inode(struct inode *inode) +static void ubifs_dirty_inode(struct inode *inode, int flags) { struct ubifs_inode *ui = ubifs_inode(inode); @@ -811,15 +811,18 @@ static int alloc_wbufs(struct ubifs_info *c) c->jheads[i].wbuf.sync_callback = &bud_wbuf_callback; c->jheads[i].wbuf.jhead = i; + c->jheads[i].grouped = 1; } c->jheads[BASEHD].wbuf.dtype = UBI_SHORTTERM; /* * Garbage Collector head likely contains long-term data and - * does not need to be synchronized by timer. + * does not need to be synchronized by timer. Also GC head nodes are + * not grouped. */ c->jheads[GCHD].wbuf.dtype = UBI_LONGTERM; c->jheads[GCHD].wbuf.no_timer = 1; + c->jheads[GCHD].grouped = 0; return 0; } @@ -1284,12 +1287,25 @@ static int mount_ubifs(struct ubifs_info *c) if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) { ubifs_msg("recovery needed"); c->need_recovery = 1; - if (!c->ro_mount) { - err = ubifs_recover_inl_heads(c, c->sbuf); - if (err) - goto out_master; - } - } else if (!c->ro_mount) { + } + + if (c->need_recovery && !c->ro_mount) { + err = ubifs_recover_inl_heads(c, c->sbuf); + if (err) + goto out_master; + } + + err = ubifs_lpt_init(c, 1, !c->ro_mount); + if (err) + goto out_master; + + if (!c->ro_mount && c->space_fixup) { + err = ubifs_fixup_free_space(c); + if (err) + goto out_master; + } + + if (!c->ro_mount) { /* * Set the "dirty" flag so that if we reboot uncleanly we * will notice this immediately on the next mount. @@ -1297,13 +1313,9 @@ static int mount_ubifs(struct ubifs_info *c) c->mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY); err = ubifs_write_master(c); if (err) - goto out_master; + goto out_lpt; } - err = ubifs_lpt_init(c, 1, !c->ro_mount); - if (err) - goto out_lpt; - err = dbg_check_idx_size(c, c->bi.old_idx_sz); if (err) goto out_lpt; @@ -1396,12 +1408,6 @@ static int mount_ubifs(struct ubifs_info *c) } else ubifs_assert(c->lst.taken_empty_lebs > 0); - if (!c->ro_mount && c->space_fixup) { - err = ubifs_fixup_free_space(c); - if (err) - goto out_infos; - } - err = dbg_check_filesystem(c); if (err) goto out_infos; @@ -1842,7 +1848,6 @@ static void ubifs_put_super(struct super_block *sb) bdi_destroy(&c->bdi); ubi_close_volume(c->ubi); mutex_unlock(&c->umount_mutex); - kfree(c); } static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data) @@ -1965,61 +1970,65 @@ static struct ubi_volume_desc *open_ubi(const char *name, int mode) return ERR_PTR(-EINVAL); } -static int ubifs_fill_super(struct super_block *sb, void *data, int silent) +static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi) { - struct ubi_volume_desc *ubi = sb->s_fs_info; struct ubifs_info *c; - struct inode *root; - int err; c = kzalloc(sizeof(struct ubifs_info), GFP_KERNEL); - if (!c) - return -ENOMEM; + if (c) { + spin_lock_init(&c->cnt_lock); + spin_lock_init(&c->cs_lock); + spin_lock_init(&c->buds_lock); + spin_lock_init(&c->space_lock); + spin_lock_init(&c->orphan_lock); + init_rwsem(&c->commit_sem); + mutex_init(&c->lp_mutex); + mutex_init(&c->tnc_mutex); + mutex_init(&c->log_mutex); + mutex_init(&c->mst_mutex); + mutex_init(&c->umount_mutex); + mutex_init(&c->bu_mutex); + mutex_init(&c->write_reserve_mutex); + init_waitqueue_head(&c->cmt_wq); + c->buds = RB_ROOT; + c->old_idx = RB_ROOT; + c->size_tree = RB_ROOT; + c->orph_tree = RB_ROOT; + INIT_LIST_HEAD(&c->infos_list); + INIT_LIST_HEAD(&c->idx_gc); + INIT_LIST_HEAD(&c->replay_list); + INIT_LIST_HEAD(&c->replay_buds); + INIT_LIST_HEAD(&c->uncat_list); + INIT_LIST_HEAD(&c->empty_list); + INIT_LIST_HEAD(&c->freeable_list); + INIT_LIST_HEAD(&c->frdi_idx_list); + INIT_LIST_HEAD(&c->unclean_leb_list); + INIT_LIST_HEAD(&c->old_buds); + INIT_LIST_HEAD(&c->orph_list); + INIT_LIST_HEAD(&c->orph_new); + c->no_chk_data_crc = 1; + + c->highest_inum = UBIFS_FIRST_INO; + c->lhead_lnum = c->ltail_lnum = UBIFS_LOG_LNUM; + + ubi_get_volume_info(ubi, &c->vi); + ubi_get_device_info(c->vi.ubi_num, &c->di); + } + return c; +} - spin_lock_init(&c->cnt_lock); - spin_lock_init(&c->cs_lock); - spin_lock_init(&c->buds_lock); - spin_lock_init(&c->space_lock); - spin_lock_init(&c->orphan_lock); - init_rwsem(&c->commit_sem); - mutex_init(&c->lp_mutex); - mutex_init(&c->tnc_mutex); - mutex_init(&c->log_mutex); - mutex_init(&c->mst_mutex); - mutex_init(&c->umount_mutex); - mutex_init(&c->bu_mutex); - mutex_init(&c->write_reserve_mutex); - init_waitqueue_head(&c->cmt_wq); - c->buds = RB_ROOT; - c->old_idx = RB_ROOT; - c->size_tree = RB_ROOT; - c->orph_tree = RB_ROOT; - INIT_LIST_HEAD(&c->infos_list); - INIT_LIST_HEAD(&c->idx_gc); - INIT_LIST_HEAD(&c->replay_list); - INIT_LIST_HEAD(&c->replay_buds); - INIT_LIST_HEAD(&c->uncat_list); - INIT_LIST_HEAD(&c->empty_list); - INIT_LIST_HEAD(&c->freeable_list); - INIT_LIST_HEAD(&c->frdi_idx_list); - INIT_LIST_HEAD(&c->unclean_leb_list); - INIT_LIST_HEAD(&c->old_buds); - INIT_LIST_HEAD(&c->orph_list); - INIT_LIST_HEAD(&c->orph_new); - c->no_chk_data_crc = 1; +static int ubifs_fill_super(struct super_block *sb, void *data, int silent) +{ + struct ubifs_info *c = sb->s_fs_info; + struct inode *root; + int err; c->vfs_sb = sb; - c->highest_inum = UBIFS_FIRST_INO; - c->lhead_lnum = c->ltail_lnum = UBIFS_LOG_LNUM; - - ubi_get_volume_info(ubi, &c->vi); - ubi_get_device_info(c->vi.ubi_num, &c->di); - /* Re-open the UBI device in read-write mode */ c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READWRITE); if (IS_ERR(c->ubi)) { err = PTR_ERR(c->ubi); - goto out_free; + goto out; } /* @@ -2085,24 +2094,29 @@ out_bdi: bdi_destroy(&c->bdi); out_close: ubi_close_volume(c->ubi); -out_free: - kfree(c); +out: return err; } static int sb_test(struct super_block *sb, void *data) { - dev_t *dev = data; + struct ubifs_info *c1 = data; struct ubifs_info *c = sb->s_fs_info; - return c->vi.cdev == *dev; + return c->vi.cdev == c1->vi.cdev; +} + +static int sb_set(struct super_block *sb, void *data) +{ + sb->s_fs_info = data; + return set_anon_super(sb, NULL); } static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags, const char *name, void *data) { struct ubi_volume_desc *ubi; - struct ubi_volume_info vi; + struct ubifs_info *c; struct super_block *sb; int err; @@ -2119,19 +2133,25 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags, name, (int)PTR_ERR(ubi)); return ERR_CAST(ubi); } - ubi_get_volume_info(ubi, &vi); - dbg_gen("opened ubi%d_%d", vi.ubi_num, vi.vol_id); + c = alloc_ubifs_info(ubi); + if (!c) { + err = -ENOMEM; + goto out_close; + } + + dbg_gen("opened ubi%d_%d", c->vi.ubi_num, c->vi.vol_id); - sb = sget(fs_type, &sb_test, &set_anon_super, &vi.cdev); + sb = sget(fs_type, sb_test, sb_set, c); if (IS_ERR(sb)) { err = PTR_ERR(sb); + kfree(c); goto out_close; } if (sb->s_root) { struct ubifs_info *c1 = sb->s_fs_info; - + kfree(c); /* A new mount point for already mounted UBIFS */ dbg_gen("this ubi volume is already mounted"); if (!!(flags & MS_RDONLY) != c1->ro_mount) { @@ -2140,11 +2160,6 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags, } } else { sb->s_flags = flags; - /* - * Pass 'ubi' to 'fill_super()' in sb->s_fs_info where it is - * replaced by 'c'. - */ - sb->s_fs_info = ubi; err = ubifs_fill_super(sb, data, flags & MS_SILENT ? 1 : 0); if (err) goto out_deact; @@ -2164,11 +2179,18 @@ out_close: return ERR_PTR(err); } +static void kill_ubifs_super(struct super_block *s) +{ + struct ubifs_info *c = s->s_fs_info; + kill_anon_super(s); + kfree(c); +} + static struct file_system_type ubifs_fs_type = { .name = "ubifs", .owner = THIS_MODULE, .mount = ubifs_mount, - .kill_sb = kill_anon_super, + .kill_sb = kill_ubifs_super, }; /* diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c index 8119b1fd8d94..91b4213dde84 100644 --- a/fs/ubifs/tnc.c +++ b/fs/ubifs/tnc.c @@ -2876,12 +2876,13 @@ static void tnc_destroy_cnext(struct ubifs_info *c) */ void ubifs_tnc_close(struct ubifs_info *c) { - long clean_freed; - tnc_destroy_cnext(c); if (c->zroot.znode) { - clean_freed = ubifs_destroy_tnc_subtree(c->zroot.znode); - atomic_long_sub(clean_freed, &ubifs_clean_zn_cnt); + long n; + + ubifs_destroy_tnc_subtree(c->zroot.znode); + n = atomic_long_read(&c->clean_zn_cnt); + atomic_long_sub(n, &ubifs_clean_zn_cnt); } kfree(c->gap_lebs); kfree(c->ilebs); diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 93d1412a06f0..f79983d6f860 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -722,12 +722,14 @@ struct ubifs_bud { * struct ubifs_jhead - journal head. * @wbuf: head's write-buffer * @buds_list: list of bud LEBs belonging to this journal head + * @grouped: non-zero if UBIFS groups nodes when writing to this journal head * * Note, the @buds list is protected by the @c->buds_lock. */ struct ubifs_jhead { struct ubifs_wbuf wbuf; struct list_head buds_list; + unsigned int grouped:1; }; /** @@ -1614,7 +1616,7 @@ int ubifs_tnc_start_commit(struct ubifs_info *c, struct ubifs_zbranch *zroot); int ubifs_tnc_end_commit(struct ubifs_info *c); /* shrinker.c */ -int ubifs_shrinker(struct shrinker *shrink, int nr_to_scan, gfp_t gfp_mask); +int ubifs_shrinker(struct shrinker *shrink, struct shrink_control *sc); /* commit.c */ int ubifs_bg_thread(void *info); @@ -1742,7 +1744,7 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum); int ubifs_recover_master_node(struct ubifs_info *c); int ubifs_write_rcvrd_mst_node(struct ubifs_info *c); struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum, - int offs, void *sbuf, int grouped); + int offs, void *sbuf, int jhead); struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum, int offs, void *sbuf); int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf); diff --git a/fs/udf/namei.c b/fs/udf/namei.c index 4d76594c2a8f..f1dce848ef96 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -783,8 +783,6 @@ static int udf_rmdir(struct inode *dir, struct dentry *dentry) struct fileIdentDesc *fi, cfi; struct kernel_lb_addr tloc; - dentry_unhash(dentry); - retval = -ENOENT; fi = udf_find_entry(dir, &dentry->d_name, &fibh, &cfi); if (!fi) @@ -1083,9 +1081,6 @@ static int udf_rename(struct inode *old_dir, struct dentry *old_dentry, struct kernel_lb_addr tloc; struct udf_inode_info *old_iinfo = UDF_I(old_inode); - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - ofi = udf_find_entry(old_dir, &old_dentry->d_name, &ofibh, &ocfi); if (ofi) { if (ofibh.sbh != ofibh.ebh) diff --git a/fs/ufs/namei.c b/fs/ufs/namei.c index 953ebdfc5bf7..29309e25417f 100644 --- a/fs/ufs/namei.c +++ b/fs/ufs/namei.c @@ -258,8 +258,6 @@ static int ufs_rmdir (struct inode * dir, struct dentry *dentry) struct inode * inode = dentry->d_inode; int err= -ENOTEMPTY; - dentry_unhash(dentry); - lock_ufs(dir->i_sb); if (ufs_empty_dir (inode)) { err = ufs_unlink(dir, dentry); @@ -284,9 +282,6 @@ static int ufs_rename(struct inode *old_dir, struct dentry *old_dentry, struct ufs_dir_entry *old_de; int err = -ENOENT; - if (new_inode && S_ISDIR(new_inode->i_mode)) - dentry_unhash(new_dentry); - old_de = ufs_find_entry(old_dir, &old_dentry->d_name, &old_page); if (!old_de) goto out; diff --git a/fs/xattr.c b/fs/xattr.c index f1ef94974dea..f060663ab70c 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -46,18 +46,22 @@ xattr_permission(struct inode *inode, const char *name, int mask) return 0; /* - * The trusted.* namespace can only be accessed by a privileged user. + * The trusted.* namespace can only be accessed by privileged users. */ - if (!strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN)) - return (capable(CAP_SYS_ADMIN) ? 0 : -EPERM); + if (!strncmp(name, XATTR_TRUSTED_PREFIX, XATTR_TRUSTED_PREFIX_LEN)) { + if (!capable(CAP_SYS_ADMIN)) + return (mask & MAY_WRITE) ? -EPERM : -ENODATA; + return 0; + } - /* In user.* namespace, only regular files and directories can have + /* + * In the user.* namespace, only regular files and directories can have * extended attributes. For sticky directories, only the owner and - * privileged user can write attributes. + * privileged users can write attributes. */ if (!strncmp(name, XATTR_USER_PREFIX, XATTR_USER_PREFIX_LEN)) { if (!S_ISREG(inode->i_mode) && !S_ISDIR(inode->i_mode)) - return -EPERM; + return (mask & MAY_WRITE) ? -EPERM : -ENODATA; if (S_ISDIR(inode->i_mode) && (inode->i_mode & S_ISVTX) && (mask & MAY_WRITE) && !inode_owner_or_capable(inode)) return -EPERM; @@ -87,7 +91,11 @@ int __vfs_setxattr_noperm(struct dentry *dentry, const char *name, { struct inode *inode = dentry->d_inode; int error = -EOPNOTSUPP; + int issec = !strncmp(name, XATTR_SECURITY_PREFIX, + XATTR_SECURITY_PREFIX_LEN); + if (issec) + inode->i_flags &= ~S_NOSEC; if (inode->i_op->setxattr) { error = inode->i_op->setxattr(dentry, name, value, size, flags); if (!error) { @@ -95,8 +103,7 @@ int __vfs_setxattr_noperm(struct dentry *dentry, const char *name, security_inode_post_setxattr(dentry, name, value, size, flags); } - } else if (!strncmp(name, XATTR_SECURITY_PREFIX, - XATTR_SECURITY_PREFIX_LEN)) { + } else if (issec) { const char *suffix = name + XATTR_SECURITY_PREFIX_LEN; error = security_inode_setsecurity(inode, suffix, value, size, flags); diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c index f4213ba1ff85..7f782af286bf 100644 --- a/fs/xfs/linux-2.6/xfs_file.c +++ b/fs/xfs/linux-2.6/xfs_file.c @@ -131,19 +131,34 @@ xfs_file_fsync( { struct inode *inode = file->f_mapping->host; struct xfs_inode *ip = XFS_I(inode); + struct xfs_mount *mp = ip->i_mount; struct xfs_trans *tp; int error = 0; int log_flushed = 0; trace_xfs_file_fsync(ip); - if (XFS_FORCED_SHUTDOWN(ip->i_mount)) + if (XFS_FORCED_SHUTDOWN(mp)) return -XFS_ERROR(EIO); xfs_iflags_clear(ip, XFS_ITRUNCATED); xfs_ioend_wait(ip); + if (mp->m_flags & XFS_MOUNT_BARRIER) { + /* + * If we have an RT and/or log subvolume we need to make sure + * to flush the write cache the device used for file data + * first. This is to ensure newly written file data make + * it to disk before logging the new inode size in case of + * an extending write. + */ + if (XFS_IS_REALTIME_INODE(ip)) + xfs_blkdev_issue_flush(mp->m_rtdev_targp); + else if (mp->m_logdev_targp != mp->m_ddev_targp) + xfs_blkdev_issue_flush(mp->m_ddev_targp); + } + /* * We always need to make sure that the required inode state is safe on * disk. The inode might be clean but we still might need to force the @@ -175,9 +190,9 @@ xfs_file_fsync( * updates. The sync transaction will also force the log. */ xfs_iunlock(ip, XFS_ILOCK_SHARED); - tp = xfs_trans_alloc(ip->i_mount, XFS_TRANS_FSYNC_TS); + tp = xfs_trans_alloc(mp, XFS_TRANS_FSYNC_TS); error = xfs_trans_reserve(tp, 0, - XFS_FSYNC_TS_LOG_RES(ip->i_mount), 0, 0, 0); + XFS_FSYNC_TS_LOG_RES(mp), 0, 0, 0); if (error) { xfs_trans_cancel(tp, 0); return -error; @@ -209,28 +224,25 @@ xfs_file_fsync( * force the log. */ if (xfs_ipincount(ip)) { - error = _xfs_log_force_lsn(ip->i_mount, + error = _xfs_log_force_lsn(mp, ip->i_itemp->ili_last_lsn, XFS_LOG_SYNC, &log_flushed); } xfs_iunlock(ip, XFS_ILOCK_SHARED); } - if (ip->i_mount->m_flags & XFS_MOUNT_BARRIER) { - /* - * If the log write didn't issue an ordered tag we need - * to flush the disk cache for the data device now. - */ - if (!log_flushed) - xfs_blkdev_issue_flush(ip->i_mount->m_ddev_targp); - - /* - * If this inode is on the RT dev we need to flush that - * cache as well. - */ - if (XFS_IS_REALTIME_INODE(ip)) - xfs_blkdev_issue_flush(ip->i_mount->m_rtdev_targp); - } + /* + * If we only have a single device, and the log force about was + * a no-op we might have to flush the data device cache here. + * This can only happen for fdatasync/O_DSYNC if we were overwriting + * an already allocated file and thus do not have any metadata to + * commit. + */ + if ((mp->m_flags & XFS_MOUNT_BARRIER) && + mp->m_logdev_targp == mp->m_ddev_targp && + !XFS_IS_REALTIME_INODE(ip) && + !log_flushed) + xfs_blkdev_issue_flush(mp->m_ddev_targp); return -error; } diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index dd21784525a8..d44d92cd12b1 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -182,7 +182,7 @@ xfs_vn_mknod( if (IS_POSIXACL(dir)) { default_acl = xfs_get_acl(dir, ACL_TYPE_DEFAULT); if (IS_ERR(default_acl)) - return -PTR_ERR(default_acl); + return PTR_ERR(default_acl); if (!default_acl) mode &= ~current_umask(); diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 98b9c91fcdf1..a1a881e68a9a 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -627,68 +627,6 @@ xfs_blkdev_put( blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL); } -/* - * Try to write out the superblock using barriers. - */ -STATIC int -xfs_barrier_test( - xfs_mount_t *mp) -{ - xfs_buf_t *sbp = xfs_getsb(mp, 0); - int error; - - XFS_BUF_UNDONE(sbp); - XFS_BUF_UNREAD(sbp); - XFS_BUF_UNDELAYWRITE(sbp); - XFS_BUF_WRITE(sbp); - XFS_BUF_UNASYNC(sbp); - XFS_BUF_ORDERED(sbp); - - xfsbdstrat(mp, sbp); - error = xfs_buf_iowait(sbp); - - /* - * Clear all the flags we set and possible error state in the - * buffer. We only did the write to try out whether barriers - * worked and shouldn't leave any traces in the superblock - * buffer. - */ - XFS_BUF_DONE(sbp); - XFS_BUF_ERROR(sbp, 0); - XFS_BUF_UNORDERED(sbp); - - xfs_buf_relse(sbp); - return error; -} - -STATIC void -xfs_mountfs_check_barriers(xfs_mount_t *mp) -{ - int error; - - if (mp->m_logdev_targp != mp->m_ddev_targp) { - xfs_notice(mp, - "Disabling barriers, not supported with external log device"); - mp->m_flags &= ~XFS_MOUNT_BARRIER; - return; - } - - if (xfs_readonly_buftarg(mp->m_ddev_targp)) { - xfs_notice(mp, - "Disabling barriers, underlying device is readonly"); - mp->m_flags &= ~XFS_MOUNT_BARRIER; - return; - } - - error = xfs_barrier_test(mp); - if (error) { - xfs_notice(mp, - "Disabling barriers, trial barrier write failed"); - mp->m_flags &= ~XFS_MOUNT_BARRIER; - return; - } -} - void xfs_blkdev_issue_flush( xfs_buftarg_t *buftarg) @@ -925,7 +863,8 @@ xfs_fs_inode_init_once( */ STATIC void xfs_fs_dirty_inode( - struct inode *inode) + struct inode *inode, + int flags) { barrier(); XFS_I(inode)->i_update_core = 1; @@ -1239,14 +1178,6 @@ xfs_fs_remount( switch (token) { case Opt_barrier: mp->m_flags |= XFS_MOUNT_BARRIER; - - /* - * Test if barriers are actually working if we can, - * else delay this check until the filesystem is - * marked writeable. - */ - if (!(mp->m_flags & XFS_MOUNT_RDONLY)) - xfs_mountfs_check_barriers(mp); break; case Opt_nobarrier: mp->m_flags &= ~XFS_MOUNT_BARRIER; @@ -1281,8 +1212,6 @@ xfs_fs_remount( /* ro -> rw */ if ((mp->m_flags & XFS_MOUNT_RDONLY) && !(*flags & MS_RDONLY)) { mp->m_flags &= ~XFS_MOUNT_RDONLY; - if (mp->m_flags & XFS_MOUNT_BARRIER) - xfs_mountfs_check_barriers(mp); /* * If this is the first remount to writeable state we @@ -1464,9 +1393,6 @@ xfs_fs_fill_super( if (error) goto out_free_sb; - if (mp->m_flags & XFS_MOUNT_BARRIER) - xfs_mountfs_check_barriers(mp); - error = xfs_filestream_mount(mp); if (error) goto out_free_sb; diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index c86375378810..01d2072fb6d4 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -490,6 +490,13 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags) args.whichfork = XFS_ATTR_FORK; /* + * we have no control over the attribute names that userspace passes us + * to remove, so we have to allow the name lookup prior to attribute + * removal to fail. + */ + args.op_flags = XFS_DA_OP_OKNOENT; + + /* * Attach the dquots to the inode. */ error = xfs_qm_dqattach(dp, 0); diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c index cb9b6d1469f7..3631783b2b53 100644 --- a/fs/xfs/xfs_iget.c +++ b/fs/xfs/xfs_iget.c @@ -253,16 +253,21 @@ xfs_iget_cache_hit( rcu_read_lock(); spin_lock(&ip->i_flags_lock); - ip->i_flags &= ~XFS_INEW; - ip->i_flags |= XFS_IRECLAIMABLE; - __xfs_inode_set_reclaim_tag(pag, ip); + ip->i_flags &= ~(XFS_INEW | XFS_IRECLAIM); + ASSERT(ip->i_flags & XFS_IRECLAIMABLE); trace_xfs_iget_reclaim_fail(ip); goto out_error; } spin_lock(&pag->pag_ici_lock); spin_lock(&ip->i_flags_lock); - ip->i_flags &= ~(XFS_IRECLAIMABLE | XFS_IRECLAIM); + + /* + * Clear the per-lifetime state in the inode as we are now + * effectively a new inode and need to return to the initial + * state before reuse occurs. + */ + ip->i_flags &= ~XFS_IRECLAIM_RESET_FLAGS; ip->i_flags |= XFS_INEW; __xfs_inode_clear_reclaim_tag(mp, pag, ip); inode->i_state = I_NEW; diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 3ae6d58e5473..964cfea77686 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -384,6 +384,16 @@ static inline void xfs_ifunlock(xfs_inode_t *ip) #define XFS_IDIRTY_RELEASE 0x0040 /* dirty release already seen */ /* + * Per-lifetime flags need to be reset when re-using a reclaimable inode during + * inode lookup. Thi prevents unintended behaviour on the new inode from + * ocurring. + */ +#define XFS_IRECLAIM_RESET_FLAGS \ + (XFS_IRECLAIMABLE | XFS_IRECLAIM | \ + XFS_IDIRTY_RELEASE | XFS_ITRUNCATED | \ + XFS_IFILESTREAM); + +/* * Flags for inode locking. * Bit ranges: 1<<1 - 1<<16-1 -- iolock/ilock modes (bitfield) * 1<<16 - 1<<32-1 -- lockdep annotation (integers) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 211930246f20..41d5b8f2bf92 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -1372,8 +1372,17 @@ xlog_sync(xlog_t *log, XFS_BUF_ASYNC(bp); bp->b_flags |= XBF_LOG_BUFFER; - if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) + if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) { + /* + * If we have an external log device, flush the data device + * before flushing the log to make sure all meta data + * written back from the AIL actually made it to disk + * before writing out the new log tail LSN in the log buffer. + */ + if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp) + xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp); XFS_BUF_ORDERED(bp); + } ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1); ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index b7a5fe7c52c8..619720705bc6 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -960,8 +960,11 @@ xfs_release( * be exposed to that problem. */ truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED); - if (truncated && VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0) - xfs_flush_pages(ip, 0, -1, XBF_ASYNC, FI_NONE); + if (truncated) { + xfs_iflags_clear(ip, XFS_IDIRTY_RELEASE); + if (VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0) + xfs_flush_pages(ip, 0, -1, XBF_ASYNC, FI_NONE); + } } if (ip->i_d.di_nlink == 0) |