diff options
Diffstat (limited to 'fs/ocfs2')
-rw-r--r-- | fs/ocfs2/acl.c | 1 | ||||
-rw-r--r-- | fs/ocfs2/alloc.c | 3 | ||||
-rw-r--r-- | fs/ocfs2/aops.c | 7 | ||||
-rw-r--r-- | fs/ocfs2/aops.h | 5 | ||||
-rw-r--r-- | fs/ocfs2/buffer_head_io.c | 2 | ||||
-rw-r--r-- | fs/ocfs2/cluster/sys.c | 2 | ||||
-rw-r--r-- | fs/ocfs2/cluster/tcp.c | 28 | ||||
-rw-r--r-- | fs/ocfs2/dcache.c | 61 | ||||
-rw-r--r-- | fs/ocfs2/dcache.h | 12 | ||||
-rw-r--r-- | fs/ocfs2/dir.c | 6 | ||||
-rw-r--r-- | fs/ocfs2/dlm/dlmdomain.c | 27 | ||||
-rw-r--r-- | fs/ocfs2/dlm/dlmrecovery.c | 29 | ||||
-rw-r--r-- | fs/ocfs2/dlmglue.c | 44 | ||||
-rw-r--r-- | fs/ocfs2/dlmglue.h | 3 | ||||
-rw-r--r-- | fs/ocfs2/file.c | 69 | ||||
-rw-r--r-- | fs/ocfs2/inode.c | 61 | ||||
-rw-r--r-- | fs/ocfs2/inode.h | 17 | ||||
-rw-r--r-- | fs/ocfs2/ioctl.c | 5 | ||||
-rw-r--r-- | fs/ocfs2/journal.c | 6 | ||||
-rw-r--r-- | fs/ocfs2/journal.h | 11 | ||||
-rw-r--r-- | fs/ocfs2/locks.c | 2 | ||||
-rw-r--r-- | fs/ocfs2/move_extents.c | 7 | ||||
-rw-r--r-- | fs/ocfs2/namei.c | 8 | ||||
-rw-r--r-- | fs/ocfs2/ocfs2.h | 33 | ||||
-rw-r--r-- | fs/ocfs2/quota.h | 2 | ||||
-rw-r--r-- | fs/ocfs2/quota_global.c | 35 | ||||
-rw-r--r-- | fs/ocfs2/stackglue.c | 22 | ||||
-rw-r--r-- | fs/ocfs2/suballoc.c | 29 | ||||
-rw-r--r-- | fs/ocfs2/suballoc.h | 4 | ||||
-rw-r--r-- | fs/ocfs2/super.c | 55 | ||||
-rw-r--r-- | fs/ocfs2/sysfile.c | 3 | ||||
-rw-r--r-- | fs/ocfs2/xattr.c | 35 |
32 files changed, 370 insertions, 264 deletions
diff --git a/fs/ocfs2/acl.c b/fs/ocfs2/acl.c index 555f4cddefe3..7e8282dcea2a 100644 --- a/fs/ocfs2/acl.c +++ b/fs/ocfs2/acl.c @@ -205,6 +205,7 @@ static int ocfs2_acl_set_mode(struct inode *inode, struct buffer_head *di_bh, di->i_mode = cpu_to_le16(inode->i_mode); di->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, di_bh); diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index e2edff38be52..b4deb5f750d9 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -5728,6 +5728,7 @@ int ocfs2_remove_btree_range(struct inode *inode, } ocfs2_et_update_clusters(et, -len); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, et->et_root_bh); @@ -6932,6 +6933,7 @@ int ocfs2_convert_inline_data_to_extents(struct inode *inode, di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features); spin_unlock(&oi->ip_lock); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_dinode_new_extent_list(inode, di); ocfs2_journal_dirty(handle, di_bh); @@ -7208,6 +7210,7 @@ int ocfs2_truncate_inline(struct inode *inode, struct buffer_head *di_bh, di->i_ctime = di->i_mtime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, di_bh); out_commit: diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index aeb44e879c51..d310d12a9adc 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -571,7 +571,6 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, { struct inode *inode = file_inode(iocb->ki_filp); int level; - wait_queue_head_t *wq = ocfs2_ioend_wq(inode); /* this io's submitter should not have unlocked this before we could */ BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); @@ -582,10 +581,7 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, if (ocfs2_iocb_is_unaligned_aio(iocb)) { ocfs2_iocb_clear_unaligned_aio(iocb); - if (atomic_dec_and_test(&OCFS2_I(inode)->ip_unaligned_aio) && - waitqueue_active(wq)) { - wake_up_all(wq); - } + mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio); } ocfs2_iocb_clear_rw_locked(iocb); @@ -2043,6 +2039,7 @@ out_write_size: inode->i_mtime = inode->i_ctime = CURRENT_TIME; di->i_mtime = di->i_ctime = cpu_to_le64(inode->i_mtime.tv_sec); di->i_mtime_nsec = di->i_ctime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, wc->w_di_bh); ocfs2_commit_trans(osb, handle); diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h index f671e49beb34..6cae155d54df 100644 --- a/fs/ocfs2/aops.h +++ b/fs/ocfs2/aops.h @@ -102,9 +102,4 @@ enum ocfs2_iocb_lock_bits { #define ocfs2_iocb_is_unaligned_aio(iocb) \ test_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) -#define OCFS2_IOEND_WQ_HASH_SZ 37 -#define ocfs2_ioend_wq(v) (&ocfs2__ioend_wq[((unsigned long)(v)) %\ - OCFS2_IOEND_WQ_HASH_SZ]) -extern wait_queue_head_t ocfs2__ioend_wq[OCFS2_IOEND_WQ_HASH_SZ]; - #endif /* OCFS2_FILE_H */ diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c index 5b704c63a103..1edcb141f639 100644 --- a/fs/ocfs2/buffer_head_io.c +++ b/fs/ocfs2/buffer_head_io.c @@ -90,7 +90,6 @@ int ocfs2_write_block(struct ocfs2_super *osb, struct buffer_head *bh, * information for this bh as it's not marked locally * uptodate. */ ret = -EIO; - put_bh(bh); mlog_errno(ret); } @@ -420,7 +419,6 @@ int ocfs2_write_super_or_backup(struct ocfs2_super *osb, if (!buffer_uptodate(bh)) { ret = -EIO; - put_bh(bh); mlog_errno(ret); } diff --git a/fs/ocfs2/cluster/sys.c b/fs/ocfs2/cluster/sys.c index a4b07730b2e1..b7f57271d49c 100644 --- a/fs/ocfs2/cluster/sys.c +++ b/fs/ocfs2/cluster/sys.c @@ -41,7 +41,7 @@ static ssize_t version_show(struct kobject *kobj, struct kobj_attribute *attr, return snprintf(buf, PAGE_SIZE, "%u\n", O2NM_API_VERSION); } static struct kobj_attribute attr_version = - __ATTR(interface_revision, S_IFREG | S_IRUGO, version_show, NULL); + __ATTR(interface_revision, S_IRUGO, version_show, NULL); static struct attribute *o2cb_attrs[] = { &attr_version.attr, diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index ea63d6461f55..dfda2ffdb16c 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -262,17 +262,17 @@ static void o2net_update_recv_stats(struct o2net_sock_container *sc) #endif /* CONFIG_OCFS2_FS_STATS */ -static inline int o2net_reconnect_delay(void) +static inline unsigned int o2net_reconnect_delay(void) { return o2nm_single_cluster->cl_reconnect_delay_ms; } -static inline int o2net_keepalive_delay(void) +static inline unsigned int o2net_keepalive_delay(void) { return o2nm_single_cluster->cl_keepalive_delay_ms; } -static inline int o2net_idle_timeout(void) +static inline unsigned int o2net_idle_timeout(void) { return o2nm_single_cluster->cl_idle_timeout_ms; } @@ -1937,18 +1937,30 @@ static void o2net_listen_data_ready(struct sock *sk, int bytes) goto out; } - /* ->sk_data_ready is also called for a newly established child socket - * before it has been accepted and the acceptor has set up their - * data_ready.. we only want to queue listen work for our listening - * socket */ + /* This callback may called twice when a new connection + * is being established as a child socket inherits everything + * from a parent LISTEN socket, including the data_ready cb of + * the parent. This leads to a hazard. In o2net_accept_one() + * we are still initializing the child socket but have not + * changed the inherited data_ready callback yet when + * data starts arriving. + * We avoid this hazard by checking the state. + * For the listening socket, the state will be TCP_LISTEN; for the new + * socket, will be TCP_ESTABLISHED. Also, in this case, + * sk->sk_user_data is not a valid function pointer. + */ + if (sk->sk_state == TCP_LISTEN) { mlog(ML_TCP, "bytes: %d\n", bytes); queue_work(o2net_wq, &o2net_listen_work); + } else { + ready = NULL; } out: read_unlock(&sk->sk_callback_lock); - ready(sk, bytes); + if (ready != NULL) + ready(sk, bytes); } static int o2net_open_listening_sock(__be32 addr, __be16 port) diff --git a/fs/ocfs2/dcache.c b/fs/ocfs2/dcache.c index 0d3a97d2d5f6..e2e05a106beb 100644 --- a/fs/ocfs2/dcache.c +++ b/fs/ocfs2/dcache.c @@ -37,7 +37,6 @@ #include "dlmglue.h" #include "file.h" #include "inode.h" -#include "super.h" #include "ocfs2_trace.h" void ocfs2_dentry_attach_gen(struct dentry *dentry) @@ -346,52 +345,6 @@ out_attach: return ret; } -DEFINE_SPINLOCK(dentry_list_lock); - -/* We limit the number of dentry locks to drop in one go. We have - * this limit so that we don't starve other users of ocfs2_wq. */ -#define DL_INODE_DROP_COUNT 64 - -/* Drop inode references from dentry locks */ -static void __ocfs2_drop_dl_inodes(struct ocfs2_super *osb, int drop_count) -{ - struct ocfs2_dentry_lock *dl; - - spin_lock(&dentry_list_lock); - while (osb->dentry_lock_list && (drop_count < 0 || drop_count--)) { - dl = osb->dentry_lock_list; - osb->dentry_lock_list = dl->dl_next; - spin_unlock(&dentry_list_lock); - iput(dl->dl_inode); - kfree(dl); - spin_lock(&dentry_list_lock); - } - spin_unlock(&dentry_list_lock); -} - -void ocfs2_drop_dl_inodes(struct work_struct *work) -{ - struct ocfs2_super *osb = container_of(work, struct ocfs2_super, - dentry_lock_work); - - __ocfs2_drop_dl_inodes(osb, DL_INODE_DROP_COUNT); - /* - * Don't queue dropping if umount is in progress. We flush the - * list in ocfs2_dismount_volume - */ - spin_lock(&dentry_list_lock); - if (osb->dentry_lock_list && - !ocfs2_test_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED)) - queue_work(ocfs2_wq, &osb->dentry_lock_work); - spin_unlock(&dentry_list_lock); -} - -/* Flush the whole work queue */ -void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb) -{ - __ocfs2_drop_dl_inodes(osb, -1); -} - /* * ocfs2_dentry_iput() and friends. * @@ -416,24 +369,16 @@ void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb) static void ocfs2_drop_dentry_lock(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl) { + iput(dl->dl_inode); ocfs2_simple_drop_lockres(osb, &dl->dl_lockres); ocfs2_lock_res_free(&dl->dl_lockres); - - /* We leave dropping of inode reference to ocfs2_wq as that can - * possibly lead to inode deletion which gets tricky */ - spin_lock(&dentry_list_lock); - if (!osb->dentry_lock_list && - !ocfs2_test_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED)) - queue_work(ocfs2_wq, &osb->dentry_lock_work); - dl->dl_next = osb->dentry_lock_list; - osb->dentry_lock_list = dl; - spin_unlock(&dentry_list_lock); + kfree(dl); } void ocfs2_dentry_lock_put(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl) { - int unlock; + int unlock = 0; BUG_ON(dl->dl_count == 0); diff --git a/fs/ocfs2/dcache.h b/fs/ocfs2/dcache.h index b79eff709958..55f58892b153 100644 --- a/fs/ocfs2/dcache.h +++ b/fs/ocfs2/dcache.h @@ -29,13 +29,8 @@ extern const struct dentry_operations ocfs2_dentry_ops; struct ocfs2_dentry_lock { - /* Use count of dentry lock */ unsigned int dl_count; - union { - /* Linked list of dentry locks to release */ - struct ocfs2_dentry_lock *dl_next; - u64 dl_parent_blkno; - }; + u64 dl_parent_blkno; /* * The ocfs2_dentry_lock keeps an inode reference until @@ -49,14 +44,9 @@ struct ocfs2_dentry_lock { int ocfs2_dentry_attach_lock(struct dentry *dentry, struct inode *inode, u64 parent_blkno); -extern spinlock_t dentry_list_lock; - void ocfs2_dentry_lock_put(struct ocfs2_super *osb, struct ocfs2_dentry_lock *dl); -void ocfs2_drop_dl_inodes(struct work_struct *work); -void ocfs2_drop_all_dl_inodes(struct ocfs2_super *osb); - struct dentry *ocfs2_find_local_alias(struct inode *inode, u64 parent_blkno, int skip_unhashed); diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c index 91a7e85ac8fd..0717662b4aef 100644 --- a/fs/ocfs2/dir.c +++ b/fs/ocfs2/dir.c @@ -2957,6 +2957,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, ocfs2_init_dir_trailer(dir, dirdata_bh, i); } + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, dirdata_bh); if (ocfs2_supports_indexed_dirs(osb) && !dx_inline) { @@ -3005,6 +3006,7 @@ static int ocfs2_expand_inline_dir(struct inode *dir, struct buffer_head *di_bh, di->i_size = cpu_to_le64(sb->s_blocksize); di->i_ctime = di->i_mtime = cpu_to_le64(dir->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(dir->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, dir, 1); /* * This should never fail as our extent list is empty and all @@ -3338,6 +3340,7 @@ do_extend: } else { de->rec_len = cpu_to_le16(sb->s_blocksize); } + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, new_bh); dir_i_size += dir->i_sb->s_blocksize; @@ -3896,6 +3899,7 @@ out_commit: dquot_free_space_nodirty(dir, ocfs2_clusters_to_bytes(dir->i_sb, 1)); + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_commit_trans(osb, handle); out: @@ -4134,6 +4138,7 @@ static int ocfs2_expand_inline_dx_root(struct inode *dir, mlog_errno(ret); did_quota = 0; + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, dx_root_bh); out_commit: @@ -4401,6 +4406,7 @@ static int ocfs2_dx_dir_remove_index(struct inode *dir, di->i_dyn_features = cpu_to_le16(OCFS2_I(dir)->ip_dyn_features); spin_unlock(&OCFS2_I(dir)->ip_lock); di->i_dx_root = cpu_to_le64(0ULL); + ocfs2_update_inode_fsync_trans(handle, dir, 1); ocfs2_journal_dirty(handle, di_bh); diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 33660a4a52fa..c973690dc0bc 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -1123,7 +1123,6 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, struct dlm_ctxt *dlm = NULL; char *local = NULL; int status = 0; - int locked = 0; qr = (struct dlm_query_region *) msg->buf; @@ -1132,10 +1131,8 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, /* buffer used in dlm_mast_regions() */ local = kmalloc(sizeof(qr->qr_regions), GFP_KERNEL); - if (!local) { - status = -ENOMEM; - goto bail; - } + if (!local) + return -ENOMEM; status = -EINVAL; @@ -1144,16 +1141,15 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, if (!dlm) { mlog(ML_ERROR, "Node %d queried hb regions on domain %s " "before join domain\n", qr->qr_node, qr->qr_domain); - goto bail; + goto out_domain_lock; } spin_lock(&dlm->spinlock); - locked = 1; if (dlm->joining_node != qr->qr_node) { mlog(ML_ERROR, "Node %d queried hb regions on domain %s " "but joining node is %d\n", qr->qr_node, qr->qr_domain, dlm->joining_node); - goto bail; + goto out_dlm_lock; } /* Support for global heartbeat was added in 1.1 */ @@ -1163,14 +1159,15 @@ static int dlm_query_region_handler(struct o2net_msg *msg, u32 len, "but active dlm protocol is %d.%d\n", qr->qr_node, qr->qr_domain, dlm->dlm_locking_proto.pv_major, dlm->dlm_locking_proto.pv_minor); - goto bail; + goto out_dlm_lock; } status = dlm_match_regions(dlm, qr, local, sizeof(qr->qr_regions)); -bail: - if (locked) - spin_unlock(&dlm->spinlock); +out_dlm_lock: + spin_unlock(&dlm->spinlock); + +out_domain_lock: spin_unlock(&dlm_domain_lock); kfree(local); @@ -1877,19 +1874,19 @@ static int dlm_join_domain(struct dlm_ctxt *dlm) goto bail; } - status = dlm_debug_init(dlm); + status = dlm_launch_thread(dlm); if (status < 0) { mlog_errno(status); goto bail; } - status = dlm_launch_thread(dlm); + status = dlm_launch_recovery_thread(dlm); if (status < 0) { mlog_errno(status); goto bail; } - status = dlm_launch_recovery_thread(dlm); + status = dlm_debug_init(dlm); if (status < 0) { mlog_errno(status); goto bail; diff --git a/fs/ocfs2/dlm/dlmrecovery.c b/fs/ocfs2/dlm/dlmrecovery.c index 7035af09cc03..fe29f7978f81 100644 --- a/fs/ocfs2/dlm/dlmrecovery.c +++ b/fs/ocfs2/dlm/dlmrecovery.c @@ -537,7 +537,10 @@ master_here: /* success! see if any other nodes need recovery */ mlog(0, "DONE mastering recovery of %s:%u here(this=%u)!\n", dlm->name, dlm->reco.dead_node, dlm->node_num); - dlm_reset_recovery(dlm); + spin_lock(&dlm->spinlock); + __dlm_reset_recovery(dlm); + dlm->reco.state &= ~DLM_RECO_STATE_FINALIZE; + spin_unlock(&dlm->spinlock); } dlm_end_recovery(dlm); @@ -695,6 +698,14 @@ static int dlm_remaster_locks(struct dlm_ctxt *dlm, u8 dead_node) if (all_nodes_done) { int ret; + /* Set this flag on recovery master to avoid + * a new recovery for another dead node start + * before the recovery is not done. That may + * cause recovery hung.*/ + spin_lock(&dlm->spinlock); + dlm->reco.state |= DLM_RECO_STATE_FINALIZE; + spin_unlock(&dlm->spinlock); + /* all nodes are now in DLM_RECO_NODE_DATA_DONE state * just send a finalize message to everyone and * clean up */ @@ -1750,13 +1761,13 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, struct dlm_migratable_lockres *mres) { struct dlm_migratable_lock *ml; - struct list_head *queue; + struct list_head *queue, *iter; struct list_head *tmpq = NULL; struct dlm_lock *newlock = NULL; struct dlm_lockstatus *lksb = NULL; int ret = 0; int i, j, bad; - struct dlm_lock *lock = NULL; + struct dlm_lock *lock; u8 from = O2NM_MAX_NODES; unsigned int added = 0; __be64 c; @@ -1791,14 +1802,16 @@ static int dlm_process_recovery_data(struct dlm_ctxt *dlm, /* MIGRATION ONLY! */ BUG_ON(!(mres->flags & DLM_MRES_MIGRATION)); + lock = NULL; spin_lock(&res->spinlock); for (j = DLM_GRANTED_LIST; j <= DLM_BLOCKED_LIST; j++) { tmpq = dlm_list_idx_to_ptr(res, j); - list_for_each_entry(lock, tmpq, list) { - if (lock->ml.cookie != ml->cookie) - lock = NULL; - else + list_for_each(iter, tmpq) { + lock = list_entry(iter, + struct dlm_lock, list); + if (lock->ml.cookie == ml->cookie) break; + lock = NULL; } if (lock) break; @@ -2882,8 +2895,8 @@ int dlm_finalize_reco_handler(struct o2net_msg *msg, u32 len, void *data, BUG(); } dlm->reco.state &= ~DLM_RECO_STATE_FINALIZE; + __dlm_reset_recovery(dlm); spin_unlock(&dlm->spinlock); - dlm_reset_recovery(dlm); dlm_kick_recovery_thread(dlm); break; default: diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 19986959d149..6bd690b5a061 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -3144,22 +3144,60 @@ out: return 0; } +static void ocfs2_process_blocked_lock(struct ocfs2_super *osb, + struct ocfs2_lock_res *lockres); + /* Mark the lockres as being dropped. It will no longer be * queued if blocking, but we still may have to wait on it * being dequeued from the downconvert thread before we can consider * it safe to drop. * * You can *not* attempt to call cluster_lock on this lockres anymore. */ -void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres) +void ocfs2_mark_lockres_freeing(struct ocfs2_super *osb, + struct ocfs2_lock_res *lockres) { int status; struct ocfs2_mask_waiter mw; - unsigned long flags; + unsigned long flags, flags2; ocfs2_init_mask_waiter(&mw); spin_lock_irqsave(&lockres->l_lock, flags); lockres->l_flags |= OCFS2_LOCK_FREEING; + if (lockres->l_flags & OCFS2_LOCK_QUEUED && current == osb->dc_task) { + /* + * We know the downconvert is queued but not in progress + * because we are the downconvert thread and processing + * different lock. So we can just remove the lock from the + * queue. This is not only an optimization but also a way + * to avoid the following deadlock: + * ocfs2_dentry_post_unlock() + * ocfs2_dentry_lock_put() + * ocfs2_drop_dentry_lock() + * iput() + * ocfs2_evict_inode() + * ocfs2_clear_inode() + * ocfs2_mark_lockres_freeing() + * ... blocks waiting for OCFS2_LOCK_QUEUED + * since we are the downconvert thread which + * should clear the flag. + */ + spin_unlock_irqrestore(&lockres->l_lock, flags); + spin_lock_irqsave(&osb->dc_task_lock, flags2); + list_del_init(&lockres->l_blocked_list); + osb->blocked_lock_count--; + spin_unlock_irqrestore(&osb->dc_task_lock, flags2); + /* + * Warn if we recurse into another post_unlock call. Strictly + * speaking it isn't a problem but we need to be careful if + * that happens (stack overflow, deadlocks, ...) so warn if + * ocfs2 grows a path for which this can happen. + */ + WARN_ON_ONCE(lockres->l_ops->post_unlock); + /* Since the lock is freeing we don't do much in the fn below */ + ocfs2_process_blocked_lock(osb, lockres); + return; + } while (lockres->l_flags & OCFS2_LOCK_QUEUED) { lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_QUEUED, 0); spin_unlock_irqrestore(&lockres->l_lock, flags); @@ -3180,7 +3218,7 @@ void ocfs2_simple_drop_lockres(struct ocfs2_super *osb, { int ret; - ocfs2_mark_lockres_freeing(lockres); + ocfs2_mark_lockres_freeing(osb, lockres); ret = ocfs2_drop_lock(osb, lockres); if (ret) mlog_errno(ret); diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index 1d596d8c4a4a..d293a22c32c5 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -157,7 +157,8 @@ int ocfs2_refcount_lock(struct ocfs2_refcount_tree *ref_tree, int ex); void ocfs2_refcount_unlock(struct ocfs2_refcount_tree *ref_tree, int ex); -void ocfs2_mark_lockres_freeing(struct ocfs2_lock_res *lockres); +void ocfs2_mark_lockres_freeing(struct ocfs2_super *osb, + struct ocfs2_lock_res *lockres); void ocfs2_simple_drop_lockres(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres); diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 9c27adf4ac72..8970dcf74de5 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -175,9 +175,13 @@ static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end, int datasync) { int err = 0; - journal_t *journal; struct inode *inode = file->f_mapping->host; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); + struct ocfs2_inode_info *oi = OCFS2_I(inode); + journal_t *journal = osb->journal->j_journal; + int ret; + tid_t commit_tid; + bool needs_barrier = false; trace_ocfs2_sync_file(inode, file, file->f_path.dentry, OCFS2_I(inode)->ip_blkno, @@ -192,29 +196,19 @@ static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end, if (err) return err; - /* - * Probably don't need the i_mutex at all in here, just putting it here - * to be consistent with how fsync used to be called, someone more - * familiar with the fs could possibly remove it. - */ - mutex_lock(&inode->i_mutex); - if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) { - /* - * We still have to flush drive's caches to get data to the - * platter - */ - if (osb->s_mount_opt & OCFS2_MOUNT_BARRIER) - blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); - goto bail; + commit_tid = datasync ? oi->i_datasync_tid : oi->i_sync_tid; + if (journal->j_flags & JBD2_BARRIER && + !jbd2_trans_will_send_data_barrier(journal, commit_tid)) + needs_barrier = true; + err = jbd2_complete_transaction(journal, commit_tid); + if (needs_barrier) { + ret = blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL); + if (!err) + err = ret; } - journal = osb->journal->j_journal; - err = jbd2_journal_force_commit(journal); - -bail: if (err) mlog_errno(err); - mutex_unlock(&inode->i_mutex); return (err < 0) ? -EIO : 0; } @@ -292,6 +286,7 @@ int ocfs2_update_inode_atime(struct inode *inode, inode->i_atime = CURRENT_TIME; di->i_atime = cpu_to_le64(inode->i_atime.tv_sec); di->i_atime_nsec = cpu_to_le32(inode->i_atime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, bh); out_commit: @@ -341,6 +336,7 @@ int ocfs2_simple_size_update(struct inode *inode, if (ret < 0) mlog_errno(ret); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_commit_trans(osb, handle); out: return ret; @@ -435,6 +431,7 @@ static int ocfs2_orphan_for_truncate(struct ocfs2_super *osb, di->i_size = cpu_to_le64(new_i_size); di->i_ctime = di->i_mtime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = di->i_mtime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, fe_bh); @@ -650,7 +647,7 @@ restarted_transaction: mlog_errno(status); goto leave; } - + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, bh); spin_lock(&OCFS2_I(inode)->ip_lock); @@ -743,6 +740,7 @@ static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode, OCFS2_JOURNAL_ACCESS_WRITE); if (ret) mlog_errno(ret); + ocfs2_update_inode_fsync_trans(handle, inode, 1); out: if (ret) { @@ -840,6 +838,7 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from, di->i_ctime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); di->i_mtime_nsec = di->i_ctime_nsec; ocfs2_journal_dirty(handle, di_bh); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle); } @@ -1344,6 +1343,7 @@ static int __ocfs2_write_remove_suid(struct inode *inode, di = (struct ocfs2_dinode *) bh->b_data; di->i_mode = cpu_to_le16(inode->i_mode); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, bh); @@ -1576,6 +1576,7 @@ static int ocfs2_zero_partial_clusters(struct inode *inode, if (ret) mlog_errno(ret); } + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_commit_trans(osb, handle); out: @@ -2061,13 +2062,6 @@ out: return ret; } -static void ocfs2_aiodio_wait(struct inode *inode) -{ - wait_queue_head_t *wq = ocfs2_ioend_wq(inode); - - wait_event(*wq, (atomic_read(&OCFS2_I(inode)->ip_unaligned_aio) == 0)); -} - static int ocfs2_is_io_unaligned(struct inode *inode, size_t count, loff_t pos) { int blockmask = inode->i_sb->s_blocksize - 1; @@ -2345,10 +2339,8 @@ relock: * Wait on previous unaligned aio to complete before * proceeding. */ - ocfs2_aiodio_wait(inode); - - /* Mark the iocb as needing a decrement in ocfs2_dio_end_io */ - atomic_inc(&OCFS2_I(inode)->ip_unaligned_aio); + mutex_lock(&OCFS2_I(inode)->ip_unaligned_aio); + /* Mark the iocb as needing an unlock in ocfs2_dio_end_io */ ocfs2_iocb_set_unaligned_aio(iocb); } @@ -2431,7 +2423,7 @@ out_dio: if (unaligned_dio) { ocfs2_iocb_clear_unaligned_aio(iocb); - atomic_dec(&OCFS2_I(inode)->ip_unaligned_aio); + mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio); } out: @@ -2648,7 +2640,16 @@ static loff_t ocfs2_file_llseek(struct file *file, loff_t offset, int whence) case SEEK_SET: break; case SEEK_END: - offset += inode->i_size; + /* SEEK_END requires the OCFS2 inode lock for the file + * because it references the file's size. + */ + ret = ocfs2_inode_lock(inode, NULL, 0); + if (ret < 0) { + mlog_errno(ret); + goto out; + } + offset += i_size_read(inode); + ocfs2_inode_unlock(inode, 0); break; case SEEK_CUR: if (offset == 0) { diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c index f29a90fde619..437de7f768c6 100644 --- a/fs/ocfs2/inode.c +++ b/fs/ocfs2/inode.c @@ -130,6 +130,7 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags, struct inode *inode = NULL; struct super_block *sb = osb->sb; struct ocfs2_find_inode_args args; + journal_t *journal = OCFS2_SB(sb)->journal->j_journal; trace_ocfs2_iget_begin((unsigned long long)blkno, flags, sysfile_type); @@ -169,6 +170,32 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags, goto bail; } + /* + * Set transaction id's of transactions that have to be committed + * to finish f[data]sync. We set them to currently running transaction + * as we cannot be sure that the inode or some of its metadata isn't + * part of the transaction - the inode could have been reclaimed and + * now it is reread from disk. + */ + if (journal) { + transaction_t *transaction; + tid_t tid; + struct ocfs2_inode_info *oi = OCFS2_I(inode); + + read_lock(&journal->j_state_lock); + if (journal->j_running_transaction) + transaction = journal->j_running_transaction; + else + transaction = journal->j_committing_transaction; + if (transaction) + tid = transaction->t_tid; + else + tid = journal->j_commit_sequence; + read_unlock(&journal->j_state_lock); + oi->i_sync_tid = tid; + oi->i_datasync_tid = tid; + } + bail: if (!IS_ERR(inode)) { trace_ocfs2_iget_end(inode, @@ -804,11 +831,13 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode) goto bail; } - /* If we're coming from downconvert_thread we can't go into our own - * voting [hello, deadlock city!], so unforuntately we just - * have to skip deleting this guy. That's OK though because - * the node who's doing the actual deleting should handle it - * anyway. */ + /* + * If we're coming from downconvert_thread we can't go into our own + * voting [hello, deadlock city!] so we cannot delete the inode. But + * since we dropped last inode ref when downconverting dentry lock, + * we cannot have the file open and thus the node doing unlink will + * take care of deleting the inode. + */ if (current == osb->dc_task) goto bail; @@ -822,12 +851,6 @@ static int ocfs2_inode_is_valid_to_delete(struct inode *inode) goto bail_unlock; } - /* If we have allowd wipe of this inode for another node, it - * will be marked here so we can safely skip it. Recovery will - * cleanup any inodes we might inadvertently skip here. */ - if (oi->ip_flags & OCFS2_INODE_SKIP_DELETE) - goto bail_unlock; - ret = 1; bail_unlock: spin_unlock(&oi->ip_lock); @@ -941,7 +964,7 @@ static void ocfs2_cleanup_delete_inode(struct inode *inode, (unsigned long long)OCFS2_I(inode)->ip_blkno, sync_data); if (sync_data) filemap_write_and_wait(inode->i_mapping); - truncate_inode_pages(&inode->i_data, 0); + truncate_inode_pages_final(&inode->i_data); } static void ocfs2_delete_inode(struct inode *inode) @@ -960,8 +983,6 @@ static void ocfs2_delete_inode(struct inode *inode) if (is_bad_inode(inode) || !OCFS2_I(inode)->ip_blkno) goto bail; - dquot_initialize(inode); - if (!ocfs2_inode_is_valid_to_delete(inode)) { /* It's probably not necessary to truncate_inode_pages * here but we do it for safety anyway (it will most @@ -970,6 +991,8 @@ static void ocfs2_delete_inode(struct inode *inode) goto bail; } + dquot_initialize(inode); + /* We want to block signals in delete_inode as the lock and * messaging paths may return us -ERESTARTSYS. Which would * cause us to exit early, resulting in inodes being orphaned @@ -1057,6 +1080,7 @@ static void ocfs2_clear_inode(struct inode *inode) { int status; struct ocfs2_inode_info *oi = OCFS2_I(inode); + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); clear_inode(inode); trace_ocfs2_clear_inode((unsigned long long)oi->ip_blkno, @@ -1073,9 +1097,9 @@ static void ocfs2_clear_inode(struct inode *inode) /* Do these before all the other work so that we don't bounce * the downconvert thread while waiting to destroy the locks. */ - ocfs2_mark_lockres_freeing(&oi->ip_rw_lockres); - ocfs2_mark_lockres_freeing(&oi->ip_inode_lockres); - ocfs2_mark_lockres_freeing(&oi->ip_open_lockres); + ocfs2_mark_lockres_freeing(osb, &oi->ip_rw_lockres); + ocfs2_mark_lockres_freeing(osb, &oi->ip_inode_lockres); + ocfs2_mark_lockres_freeing(osb, &oi->ip_open_lockres); ocfs2_resv_discard(&OCFS2_SB(inode->i_sb)->osb_la_resmap, &oi->ip_la_data_resv); @@ -1157,7 +1181,7 @@ void ocfs2_evict_inode(struct inode *inode) (OCFS2_I(inode)->ip_flags & OCFS2_INODE_MAYBE_ORPHANED)) { ocfs2_delete_inode(inode); } else { - truncate_inode_pages(&inode->i_data, 0); + truncate_inode_pages_final(&inode->i_data); } ocfs2_clear_inode(inode); } @@ -1260,6 +1284,7 @@ int ocfs2_mark_inode_dirty(handle_t *handle, fe->i_mtime_nsec = cpu_to_le32(inode->i_mtime.tv_nsec); ocfs2_journal_dirty(handle, bh); + ocfs2_update_inode_fsync_trans(handle, inode, 1); leave: return status; } diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h index 621fc73bf23d..a6c991c0fc98 100644 --- a/fs/ocfs2/inode.h +++ b/fs/ocfs2/inode.h @@ -44,7 +44,7 @@ struct ocfs2_inode_info struct rw_semaphore ip_xattr_sem; /* Number of outstanding AIO's which are not page aligned */ - atomic_t ip_unaligned_aio; + struct mutex ip_unaligned_aio; /* These fields are protected by ip_lock */ spinlock_t ip_lock; @@ -73,6 +73,13 @@ struct ocfs2_inode_info u32 ip_dir_lock_gen; struct ocfs2_alloc_reservation ip_la_data_resv; + + /* + * Transactions that contain inode's metadata needed to complete + * fsync and fdatasync, respectively. + */ + tid_t i_sync_tid; + tid_t i_datasync_tid; }; /* @@ -84,8 +91,6 @@ struct ocfs2_inode_info #define OCFS2_INODE_BITMAP 0x00000004 /* This inode has been wiped from disk */ #define OCFS2_INODE_DELETED 0x00000008 -/* Another node is deleting, so our delete is a nop */ -#define OCFS2_INODE_SKIP_DELETE 0x00000010 /* Has the inode been orphaned on another node? * * This hints to ocfs2_drop_inode that it should clear i_nlink before @@ -100,11 +105,11 @@ struct ocfs2_inode_info * rely on ocfs2_delete_inode to sort things out under the proper * cluster locks. */ -#define OCFS2_INODE_MAYBE_ORPHANED 0x00000020 +#define OCFS2_INODE_MAYBE_ORPHANED 0x00000010 /* Does someone have the file open O_DIRECT */ -#define OCFS2_INODE_OPEN_DIRECT 0x00000040 +#define OCFS2_INODE_OPEN_DIRECT 0x00000020 /* Tell the inode wipe code it's not in orphan dir */ -#define OCFS2_INODE_SKIP_ORPHAN_DIR 0x00000080 +#define OCFS2_INODE_SKIP_ORPHAN_DIR 0x00000040 static inline struct ocfs2_inode_info *OCFS2_I(struct inode *inode) { diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index 8ca3c29accbf..490229f43731 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -413,11 +413,12 @@ int ocfs2_info_handle_freeinode(struct inode *inode, } status = ocfs2_info_scan_inode_alloc(osb, inode_alloc, blkno, oifi, i); - if (status < 0) - goto bail; iput(inode_alloc); inode_alloc = NULL; + + if (status < 0) + goto bail; } o2info_set_request_filled(&oifi->ifi_req); diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 44fc3e530c3d..03ea9314fecd 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -2132,12 +2132,6 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb, iter = oi->ip_next_orphan; spin_lock(&oi->ip_lock); - /* The remote delete code may have set these on the - * assumption that the other node would wipe them - * successfully. If they are still in the node's - * orphan dir, we need to reset that state. */ - oi->ip_flags &= ~(OCFS2_INODE_DELETED|OCFS2_INODE_SKIP_DELETE); - /* Set the proper information to get us going into * ocfs2_delete_inode. */ oi->ip_flags |= OCFS2_INODE_MAYBE_ORPHANED; diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index 9ff4e8cf9d97..7f8cde94abfe 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -626,4 +626,15 @@ static inline int ocfs2_begin_ordered_truncate(struct inode *inode, new_size); } +static inline void ocfs2_update_inode_fsync_trans(handle_t *handle, + struct inode *inode, + int datasync) +{ + struct ocfs2_inode_info *oi = OCFS2_I(inode); + + oi->i_sync_tid = handle->h_transaction->t_tid; + if (datasync) + oi->i_datasync_tid = handle->h_transaction->t_tid; +} + #endif /* OCFS2_JOURNAL_H */ diff --git a/fs/ocfs2/locks.c b/fs/ocfs2/locks.c index e57c804069ea..6b6d092b0998 100644 --- a/fs/ocfs2/locks.c +++ b/fs/ocfs2/locks.c @@ -82,6 +82,8 @@ static int ocfs2_do_flock(struct file *file, struct inode *inode, } ret = flock_lock_file_wait(file, fl); + if (ret) + ocfs2_file_unlock(file); out: mutex_unlock(&fp->fp_mutex); diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c index 64c304d668f0..599eb4c4c8be 100644 --- a/fs/ocfs2/move_extents.c +++ b/fs/ocfs2/move_extents.c @@ -151,6 +151,7 @@ static int __ocfs2_move_extent(handle_t *handle, old_blkno, len); } + ocfs2_update_inode_fsync_trans(handle, inode, 0); out: ocfs2_free_path(path); return ret; @@ -690,8 +691,11 @@ static int ocfs2_move_extent(struct ocfs2_move_extents_context *context, ret = ocfs2_block_group_set_bits(handle, gb_inode, gd, gd_bh, goal_bit, len); - if (ret) + if (ret) { + ocfs2_rollback_alloc_dinode_counts(gb_inode, gb_bh, len, + le16_to_cpu(gd->bg_chain)); mlog_errno(ret); + } /* * Here we should write the new page out first if we are @@ -957,6 +961,7 @@ static int ocfs2_move_extents(struct ocfs2_move_extents_context *context) inode->i_ctime = CURRENT_TIME; di->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec); di->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, di_bh); diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index 3683643f3f0e..2060fc398445 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -450,7 +450,6 @@ leave: brelse(new_fe_bh); brelse(parent_fe_bh); - kfree(si.name); kfree(si.value); ocfs2_free_dir_lookup_result(&lookup); @@ -495,6 +494,7 @@ static int __ocfs2_mknod_locked(struct inode *dir, struct ocfs2_dinode *fe = NULL; struct ocfs2_extent_list *fel; u16 feat; + struct ocfs2_inode_info *oi = OCFS2_I(inode); *new_fe_bh = NULL; @@ -576,8 +576,8 @@ static int __ocfs2_mknod_locked(struct inode *dir, mlog_errno(status); } - status = 0; /* error in ocfs2_create_new_inode_locks is not - * critical */ + oi->i_sync_tid = handle->h_transaction->t_tid; + oi->i_datasync_tid = handle->h_transaction->t_tid; leave: if (status < 0) { @@ -1855,7 +1855,6 @@ bail: brelse(new_fe_bh); brelse(parent_fe_bh); - kfree(si.name); kfree(si.value); ocfs2_free_dir_lookup_result(&lookup); if (inode_ac) @@ -2481,6 +2480,7 @@ int ocfs2_mv_orphaned_inode_to_new(struct inode *dir, di->i_orphaned_slot = 0; set_nlink(inode, 1); ocfs2_set_links_count(di, inode->i_nlink); + ocfs2_update_inode_fsync_trans(handle, inode, 1); ocfs2_journal_dirty(handle, di_bh); status = ocfs2_add_entry(handle, dentry, inode, diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 553f53cc73ae..8d64a97a9d5e 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -30,6 +30,7 @@ #include <linux/sched.h> #include <linux/wait.h> #include <linux/list.h> +#include <linux/llist.h> #include <linux/rbtree.h> #include <linux/workqueue.h> #include <linux/kref.h> @@ -274,19 +275,16 @@ enum ocfs2_mount_options OCFS2_MOUNT_HB_GLOBAL = 1 << 14, /* Global heartbeat */ }; -#define OCFS2_OSB_SOFT_RO 0x0001 -#define OCFS2_OSB_HARD_RO 0x0002 -#define OCFS2_OSB_ERROR_FS 0x0004 -#define OCFS2_OSB_DROP_DENTRY_LOCK_IMMED 0x0008 - -#define OCFS2_DEFAULT_ATIME_QUANTUM 60 +#define OCFS2_OSB_SOFT_RO 0x0001 +#define OCFS2_OSB_HARD_RO 0x0002 +#define OCFS2_OSB_ERROR_FS 0x0004 +#define OCFS2_DEFAULT_ATIME_QUANTUM 60 struct ocfs2_journal; struct ocfs2_slot_info; struct ocfs2_recovery_map; struct ocfs2_replay_map; struct ocfs2_quota_recovery; -struct ocfs2_dentry_lock; struct ocfs2_super { struct task_struct *commit_task; @@ -414,10 +412,9 @@ struct ocfs2_super struct list_head blocked_lock_list; unsigned long blocked_lock_count; - /* List of dentry locks to release. Anyone can add locks to - * the list, ocfs2_wq processes the list */ - struct ocfs2_dentry_lock *dentry_lock_list; - struct work_struct dentry_lock_work; + /* List of dquot structures to drop last reference to */ + struct llist_head dquot_drop_list; + struct work_struct dquot_drop_work; wait_queue_head_t osb_mount_event; @@ -449,6 +446,8 @@ struct ocfs2_super /* rb tree root for refcount lock. */ struct rb_root osb_rf_lock_tree; struct ocfs2_refcount_tree *osb_ref_tree_lru; + + struct mutex system_file_mutex; }; #define OCFS2_SB(sb) ((struct ocfs2_super *)(sb)->s_fs_info) @@ -579,18 +578,6 @@ static inline void ocfs2_set_osb_flag(struct ocfs2_super *osb, spin_unlock(&osb->osb_lock); } - -static inline unsigned long ocfs2_test_osb_flag(struct ocfs2_super *osb, - unsigned long flag) -{ - unsigned long ret; - - spin_lock(&osb->osb_lock); - ret = osb->osb_flags & flag; - spin_unlock(&osb->osb_lock); - return ret; -} - static inline void ocfs2_set_ro_flag(struct ocfs2_super *osb, int hard) { diff --git a/fs/ocfs2/quota.h b/fs/ocfs2/quota.h index d5ab56cbe5c5..f266d67df3c6 100644 --- a/fs/ocfs2/quota.h +++ b/fs/ocfs2/quota.h @@ -28,6 +28,7 @@ struct ocfs2_dquot { unsigned int dq_use_count; /* Number of nodes having reference to this entry in global quota file */ s64 dq_origspace; /* Last globally synced space usage */ s64 dq_originodes; /* Last globally synced inode usage */ + struct llist_node list; /* Member of list of dquots to drop */ }; /* Description of one chunk to recover in memory */ @@ -110,6 +111,7 @@ int ocfs2_read_quota_phys_block(struct inode *inode, u64 p_block, int ocfs2_create_local_dquot(struct dquot *dquot); int ocfs2_local_release_dquot(handle_t *handle, struct dquot *dquot); int ocfs2_local_write_dquot(struct dquot *dquot); +void ocfs2_drop_dquot_refs(struct work_struct *work); extern const struct dquot_operations ocfs2_quota_operations; extern struct quota_format_type ocfs2_quota_format; diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index d7b5108789e2..b990a62cff50 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -10,6 +10,7 @@ #include <linux/jiffies.h> #include <linux/writeback.h> #include <linux/workqueue.h> +#include <linux/llist.h> #include <cluster/masklog.h> @@ -679,6 +680,27 @@ static int ocfs2_calc_qdel_credits(struct super_block *sb, int type) OCFS2_INODE_UPDATE_CREDITS; } +void ocfs2_drop_dquot_refs(struct work_struct *work) +{ + struct ocfs2_super *osb = container_of(work, struct ocfs2_super, + dquot_drop_work); + struct llist_node *list; + struct ocfs2_dquot *odquot, *next_odquot; + + list = llist_del_all(&osb->dquot_drop_list); + llist_for_each_entry_safe(odquot, next_odquot, list, list) { + /* Drop the reference we acquired in ocfs2_dquot_release() */ + dqput(&odquot->dq_dquot); + } +} + +/* + * Called when the last reference to dquot is dropped. If we are called from + * downconvert thread, we cannot do all the handling here because grabbing + * quota lock could deadlock (the node holding the quota lock could need some + * other cluster lock to proceed but with blocked downconvert thread we cannot + * release any lock). + */ static int ocfs2_release_dquot(struct dquot *dquot) { handle_t *handle; @@ -694,6 +716,19 @@ static int ocfs2_release_dquot(struct dquot *dquot) /* Check whether we are not racing with some other dqget() */ if (atomic_read(&dquot->dq_count) > 1) goto out; + /* Running from downconvert thread? Postpone quota processing to wq */ + if (current == osb->dc_task) { + /* + * Grab our own reference to dquot and queue it for delayed + * dropping. Quota code rechecks after calling + * ->release_dquot() and won't free dquot structure. + */ + dqgrab(dquot); + /* First entry on list -> queue work */ + if (llist_add(&OCFS2_DQUOT(dquot)->list, &osb->dquot_drop_list)) + queue_work(ocfs2_wq, &osb->dquot_drop_work); + goto out; + } status = ocfs2_lock_global_qf(oinfo, 1); if (status < 0) goto out; diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index ca5ce14cbddc..83f1a665ae97 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -496,7 +496,7 @@ static ssize_t ocfs2_max_locking_protocol_show(struct kobject *kobj, } static struct kobj_attribute ocfs2_attr_max_locking_protocol = - __ATTR(max_locking_protocol, S_IFREG | S_IRUGO, + __ATTR(max_locking_protocol, S_IRUGO, ocfs2_max_locking_protocol_show, NULL); static ssize_t ocfs2_loaded_cluster_plugins_show(struct kobject *kobj, @@ -528,7 +528,7 @@ static ssize_t ocfs2_loaded_cluster_plugins_show(struct kobject *kobj, } static struct kobj_attribute ocfs2_attr_loaded_cluster_plugins = - __ATTR(loaded_cluster_plugins, S_IFREG | S_IRUGO, + __ATTR(loaded_cluster_plugins, S_IRUGO, ocfs2_loaded_cluster_plugins_show, NULL); static ssize_t ocfs2_active_cluster_plugin_show(struct kobject *kobj, @@ -550,7 +550,7 @@ static ssize_t ocfs2_active_cluster_plugin_show(struct kobject *kobj, } static struct kobj_attribute ocfs2_attr_active_cluster_plugin = - __ATTR(active_cluster_plugin, S_IFREG | S_IRUGO, + __ATTR(active_cluster_plugin, S_IRUGO, ocfs2_active_cluster_plugin_show, NULL); static ssize_t ocfs2_cluster_stack_show(struct kobject *kobj, @@ -599,15 +599,29 @@ static ssize_t ocfs2_cluster_stack_store(struct kobject *kobj, static struct kobj_attribute ocfs2_attr_cluster_stack = - __ATTR(cluster_stack, S_IFREG | S_IRUGO | S_IWUSR, + __ATTR(cluster_stack, S_IRUGO | S_IWUSR, ocfs2_cluster_stack_show, ocfs2_cluster_stack_store); + + +static ssize_t ocfs2_dlm_recover_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + return snprintf(buf, PAGE_SIZE, "1\n"); +} + +static struct kobj_attribute ocfs2_attr_dlm_recover_support = + __ATTR(dlm_recover_callback_support, S_IRUGO, + ocfs2_dlm_recover_show, NULL); + static struct attribute *ocfs2_attrs[] = { &ocfs2_attr_max_locking_protocol.attr, &ocfs2_attr_loaded_cluster_plugins.attr, &ocfs2_attr_active_cluster_plugin.attr, &ocfs2_attr_cluster_stack.attr, + &ocfs2_attr_dlm_recover_support.attr, NULL, }; diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 47ae2663a6f5..0cb889a17ae1 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -771,6 +771,7 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb, spin_unlock(&OCFS2_I(alloc_inode)->ip_lock); i_size_write(alloc_inode, le64_to_cpu(fe->i_size)); alloc_inode->i_blocks = ocfs2_inode_sector_count(alloc_inode); + ocfs2_update_inode_fsync_trans(handle, alloc_inode, 0); status = 0; @@ -1607,6 +1608,21 @@ out: return ret; } +void ocfs2_rollback_alloc_dinode_counts(struct inode *inode, + struct buffer_head *di_bh, + u32 num_bits, + u16 chain) +{ + u32 tmp_used; + struct ocfs2_dinode *di = (struct ocfs2_dinode *) di_bh->b_data; + struct ocfs2_chain_list *cl; + + cl = (struct ocfs2_chain_list *)&di->id2.i_chain; + tmp_used = le32_to_cpu(di->id1.bitmap1.i_used); + di->id1.bitmap1.i_used = cpu_to_le32(tmp_used - num_bits); + le32_add_cpu(&cl->cl_recs[chain].c_free, num_bits); +} + static int ocfs2_bg_discontig_fix_by_rec(struct ocfs2_suballoc_result *res, struct ocfs2_extent_rec *rec, struct ocfs2_chain_list *cl) @@ -1707,8 +1723,12 @@ static int ocfs2_search_one_group(struct ocfs2_alloc_context *ac, ret = ocfs2_block_group_set_bits(handle, alloc_inode, gd, group_bh, res->sr_bit_offset, res->sr_bits); - if (ret < 0) + if (ret < 0) { + ocfs2_rollback_alloc_dinode_counts(alloc_inode, ac->ac_bh, + res->sr_bits, + le16_to_cpu(gd->bg_chain)); mlog_errno(ret); + } out_loc_only: *bits_left = le16_to_cpu(gd->bg_free_bits_count); @@ -1838,6 +1858,8 @@ static int ocfs2_search_chain(struct ocfs2_alloc_context *ac, res->sr_bit_offset, res->sr_bits); if (status < 0) { + ocfs2_rollback_alloc_dinode_counts(alloc_inode, + ac->ac_bh, res->sr_bits, chain); mlog_errno(status); goto bail; } @@ -2091,7 +2113,7 @@ int ocfs2_find_new_inode_loc(struct inode *dir, ac->ac_find_loc_priv = res; *fe_blkno = res->sr_blkno; - + ocfs2_update_inode_fsync_trans(handle, dir, 0); out: if (handle) ocfs2_commit_trans(OCFS2_SB(dir->i_sb), handle); @@ -2149,6 +2171,8 @@ int ocfs2_claim_new_inode_at_loc(handle_t *handle, res->sr_bit_offset, res->sr_bits); if (ret < 0) { + ocfs2_rollback_alloc_dinode_counts(ac->ac_inode, + ac->ac_bh, res->sr_bits, chain); mlog_errno(ret); goto out; } @@ -2870,6 +2894,7 @@ int ocfs2_test_inode_bit(struct ocfs2_super *osb, u64 blkno, int *res) status = ocfs2_inode_lock(inode_alloc_inode, &alloc_bh, 0); if (status < 0) { mutex_unlock(&inode_alloc_inode->i_mutex); + iput(inode_alloc_inode); mlog(ML_ERROR, "lock on alloc inode on slot %u failed %d\n", (u32)suballoc_slot, status); goto bail; diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h index 218d8036b3e7..2d2501767c0c 100644 --- a/fs/ocfs2/suballoc.h +++ b/fs/ocfs2/suballoc.h @@ -91,6 +91,10 @@ int ocfs2_alloc_dinode_update_counts(struct inode *inode, struct buffer_head *di_bh, u32 num_bits, u16 chain); +void ocfs2_rollback_alloc_dinode_counts(struct inode *inode, + struct buffer_head *di_bh, + u32 num_bits, + u16 chain); int ocfs2_block_group_set_bits(handle_t *handle, struct inode *alloc_inode, struct ocfs2_group_desc *bg, diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 49d84f80f36c..a7cdd56f4c79 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -561,6 +561,9 @@ static struct inode *ocfs2_alloc_inode(struct super_block *sb) if (!oi) return NULL; + oi->i_sync_tid = 0; + oi->i_datasync_tid = 0; + jbd2_journal_init_jbd_inode(&oi->ip_jinode, &oi->vfs_inode); return &oi->vfs_inode; } @@ -631,6 +634,8 @@ static int ocfs2_remount(struct super_block *sb, int *flags, char *data) struct ocfs2_super *osb = OCFS2_SB(sb); u32 tmp; + sync_filesystem(sb); + if (!ocfs2_parse_options(sb, data, &parsed_options, 1) || !ocfs2_check_set_options(sb, &parsed_options)) { ret = -EINVAL; @@ -1238,30 +1243,11 @@ static struct dentry *ocfs2_mount(struct file_system_type *fs_type, return mount_bdev(fs_type, flags, dev_name, data, ocfs2_fill_super); } -static void ocfs2_kill_sb(struct super_block *sb) -{ - struct ocfs2_super *osb = OCFS2_SB(sb); - - /* Failed mount? */ - if (!osb || atomic_read(&osb->vol_state) == VOLUME_DISABLED) - goto out; - - /* Prevent further queueing of inode drop events */ - spin_lock(&dentry_list_lock); - ocfs2_set_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED); - spin_unlock(&dentry_list_lock); - /* Wait for work to finish and/or remove it */ - cancel_work_sync(&osb->dentry_lock_work); -out: - kill_block_super(sb); -} - static struct file_system_type ocfs2_fs_type = { .owner = THIS_MODULE, .name = "ocfs2", .mount = ocfs2_mount, - .kill_sb = ocfs2_kill_sb, - + .kill_sb = kill_block_super, .fs_flags = FS_REQUIRES_DEV|FS_RENAME_DOES_D_MOVE, .next = NULL }; @@ -1612,14 +1598,9 @@ static int ocfs2_show_options(struct seq_file *s, struct dentry *root) return 0; } -wait_queue_head_t ocfs2__ioend_wq[OCFS2_IOEND_WQ_HASH_SZ]; - static int __init ocfs2_init(void) { - int status, i; - - for (i = 0; i < OCFS2_IOEND_WQ_HASH_SZ; i++) - init_waitqueue_head(&ocfs2__ioend_wq[i]); + int status; status = init_ocfs2_uptodate_cache(); if (status < 0) @@ -1761,7 +1742,7 @@ static void ocfs2_inode_init_once(void *data) ocfs2_extent_map_init(&oi->vfs_inode); INIT_LIST_HEAD(&oi->ip_io_markers); oi->ip_dir_start_lookup = 0; - atomic_set(&oi->ip_unaligned_aio, 0); + mutex_init(&oi->ip_unaligned_aio); init_rwsem(&oi->ip_alloc_sem); init_rwsem(&oi->ip_xattr_sem); mutex_init(&oi->ip_io_mutex); @@ -1932,17 +1913,16 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) debugfs_remove(osb->osb_ctxt); - /* - * Flush inode dropping work queue so that deletes are - * performed while the filesystem is still working - */ - ocfs2_drop_all_dl_inodes(osb); - /* Orphan scan should be stopped as early as possible */ ocfs2_orphan_scan_stop(osb); ocfs2_disable_quotas(osb); + /* All dquots should be freed by now */ + WARN_ON(!llist_empty(&osb->dquot_drop_list)); + /* Wait for worker to be done with the work structure in osb */ + cancel_work_sync(&osb->dquot_drop_work); + ocfs2_shutdown_local_alloc(osb); /* This will disable recovery and flush any recovery work. */ @@ -2077,7 +2057,6 @@ static int ocfs2_initialize_super(struct super_block *sb, struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data; struct inode *inode = NULL; struct ocfs2_journal *journal; - __le32 uuid_net_key; struct ocfs2_super *osb; u64 total_blocks; @@ -2123,6 +2102,8 @@ static int ocfs2_initialize_super(struct super_block *sb, spin_lock_init(&osb->osb_xattr_lock); ocfs2_init_steal_slots(osb); + mutex_init(&osb->system_file_mutex); + atomic_set(&osb->alloc_stats.moves, 0); atomic_set(&osb->alloc_stats.local_data, 0); atomic_set(&osb->alloc_stats.bitmap_data, 0); @@ -2276,8 +2257,8 @@ static int ocfs2_initialize_super(struct super_block *sb, INIT_WORK(&journal->j_recovery_work, ocfs2_complete_recovery); journal->j_state = OCFS2_JOURNAL_FREE; - INIT_WORK(&osb->dentry_lock_work, ocfs2_drop_dl_inodes); - osb->dentry_lock_list = NULL; + INIT_WORK(&osb->dquot_drop_work, ocfs2_drop_dquot_refs); + init_llist_head(&osb->dquot_drop_list); /* get some pseudo constants for clustersize bits */ osb->s_clustersize_bits = @@ -2311,8 +2292,6 @@ static int ocfs2_initialize_super(struct super_block *sb, goto bail; } - memcpy(&uuid_net_key, di->id2.i_super.s_uuid, sizeof(uuid_net_key)); - strncpy(osb->vol_label, di->id2.i_super.s_label, 63); osb->vol_label[63] = '\0'; osb->root_blkno = le64_to_cpu(di->id2.i_super.s_root_blkno); diff --git a/fs/ocfs2/sysfile.c b/fs/ocfs2/sysfile.c index f053688d22a3..af155c183123 100644 --- a/fs/ocfs2/sysfile.c +++ b/fs/ocfs2/sysfile.c @@ -113,9 +113,11 @@ struct inode *ocfs2_get_system_file_inode(struct ocfs2_super *osb, } else arr = get_local_system_inode(osb, type, slot); + mutex_lock(&osb->system_file_mutex); if (arr && ((inode = *arr) != NULL)) { /* get a ref in addition to the array ref */ inode = igrab(inode); + mutex_unlock(&osb->system_file_mutex); BUG_ON(!inode); return inode; @@ -129,6 +131,7 @@ struct inode *ocfs2_get_system_file_inode(struct ocfs2_super *osb, *arr = igrab(inode); BUG_ON(!*arr); } + mutex_unlock(&osb->system_file_mutex); return inode; } diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c index 185fa3b7f962..016f01df3825 100644 --- a/fs/ocfs2/xattr.c +++ b/fs/ocfs2/xattr.c @@ -369,7 +369,7 @@ static void ocfs2_xattr_bucket_free(struct ocfs2_xattr_bucket *bucket) * them fully. */ static int ocfs2_init_xattr_bucket(struct ocfs2_xattr_bucket *bucket, - u64 xb_blkno) + u64 xb_blkno, int new) { int i, rc = 0; @@ -383,9 +383,16 @@ static int ocfs2_init_xattr_bucket(struct ocfs2_xattr_bucket *bucket, } if (!ocfs2_buffer_uptodate(INODE_CACHE(bucket->bu_inode), - bucket->bu_bhs[i])) - ocfs2_set_new_buffer_uptodate(INODE_CACHE(bucket->bu_inode), - bucket->bu_bhs[i]); + bucket->bu_bhs[i])) { + if (new) + ocfs2_set_new_buffer_uptodate(INODE_CACHE(bucket->bu_inode), + bucket->bu_bhs[i]); + else { + set_buffer_uptodate(bucket->bu_bhs[i]); + ocfs2_set_buffer_uptodate(INODE_CACHE(bucket->bu_inode), + bucket->bu_bhs[i]); + } + } } if (rc) @@ -2602,6 +2609,7 @@ int ocfs2_xattr_remove(struct inode *inode, struct buffer_head *di_bh) oi->ip_dyn_features &= ~(OCFS2_INLINE_XATTR_FL | OCFS2_HAS_XATTR_FL); di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features); spin_unlock(&oi->ip_lock); + ocfs2_update_inode_fsync_trans(handle, inode, 0); ocfs2_journal_dirty(handle, di_bh); out_commit: @@ -3200,8 +3208,15 @@ meta_guess: clusters_add += 1; } } else { - meta_add += 1; credits += OCFS2_XATTR_BLOCK_CREATE_CREDITS; + if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) { + struct ocfs2_extent_list *el = &def_xv.xv.xr_list; + meta_add += ocfs2_extend_meta_needed(el); + credits += ocfs2_calc_extend_credits(inode->i_sb, + el); + } else { + meta_add += 1; + } } out: if (clusters_need) @@ -3614,6 +3629,7 @@ int ocfs2_xattr_set(struct inode *inode, } ret = __ocfs2_xattr_set_handle(inode, di, &xi, &xis, &xbs, &ctxt); + ocfs2_update_inode_fsync_trans(ctxt.handle, inode, 0); ocfs2_commit_trans(osb, ctxt.handle); @@ -4294,7 +4310,7 @@ static int ocfs2_xattr_create_index_block(struct inode *inode, trace_ocfs2_xattr_create_index_block((unsigned long long)blkno); - ret = ocfs2_init_xattr_bucket(xs->bucket, blkno); + ret = ocfs2_init_xattr_bucket(xs->bucket, blkno, 1); if (ret) { mlog_errno(ret); goto out; @@ -4638,7 +4654,7 @@ static int ocfs2_divide_xattr_bucket(struct inode *inode, * Even if !new_bucket_head, we're overwriting t_bucket. Thus, * there's no need to read it. */ - ret = ocfs2_init_xattr_bucket(t_bucket, new_blk); + ret = ocfs2_init_xattr_bucket(t_bucket, new_blk, new_bucket_head); if (ret) { mlog_errno(ret); goto out; @@ -4804,7 +4820,7 @@ static int ocfs2_cp_xattr_bucket(struct inode *inode, * Even if !t_is_new, we're overwriting t_bucket. Thus, * there's no need to read it. */ - ret = ocfs2_init_xattr_bucket(t_bucket, t_blkno); + ret = ocfs2_init_xattr_bucket(t_bucket, t_blkno, t_is_new); if (ret) goto out; @@ -5476,6 +5492,7 @@ static int ocfs2_rm_xattr_cluster(struct inode *inode, ret = ocfs2_truncate_log_append(osb, handle, blkno, len); if (ret) mlog_errno(ret); + ocfs2_update_inode_fsync_trans(handle, inode, 0); out_commit: ocfs2_commit_trans(osb, handle); @@ -6830,7 +6847,7 @@ static int ocfs2_reflink_xattr_bucket(handle_t *handle, break; } - ret = ocfs2_init_xattr_bucket(args->new_bucket, new_blkno); + ret = ocfs2_init_xattr_bucket(args->new_bucket, new_blkno, 1); if (ret) { mlog_errno(ret); break; |