From bb82ab88dfdb12948af58989c75bfe904bc1b09d Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Fri, 10 Jun 2011 14:06:53 +0200 Subject: btrfs: add an extra wait mode to read_extent_buffer_pages read_extent_buffer_pages currently has two modes, either trigger a read without waiting for anything, or wait for the I/O to finish. The former also bails when it's unable to lock the page. This patch now adds an additional parameter to allow it to block on page lock, but don't wait for completion. Changes v5: - merge the 2 wait parameters into one and define WAIT_NONE, WAIT_COMPLETE and WAIT_PAGE_LOCK Change v6: - fix bug introduced in v5 Signed-off-by: Arne Jansen --- fs/btrfs/disk-io.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 07b3ac662e19..e6170e142cad 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -367,7 +367,8 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags); io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree; while (1) { - ret = read_extent_buffer_pages(io_tree, eb, start, 1, + ret = read_extent_buffer_pages(io_tree, eb, start, + WAIT_COMPLETE, btree_get_extent, mirror_num); if (!ret && !verify_parent_transid(io_tree, eb, parent_transid)) @@ -974,7 +975,7 @@ int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, if (!buf) return 0; read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, - buf, 0, 0, btree_get_extent, 0); + buf, 0, WAIT_NONE, btree_get_extent, 0); free_extent_buffer(buf); return ret; } -- cgit v1.2.3 From ab0fff03055d2d1b01a7581badeba18db9c4f55c Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Mon, 23 May 2011 14:25:41 +0200 Subject: btrfs: add READAHEAD extent buffer flag Add a READAHEAD extent buffer flag. Add a function to trigger a read with this flag set. Changes v2: - use extent buffer flags instead of extent state flags Changes v5: - adapt to changed read_extent_buffer_pages interface - don't return eb from reada_tree_block_flagged if it has CORRUPT flag set Signed-off-by: Arne Jansen --- fs/btrfs/disk-io.c | 32 ++++++++++++++++++++++++++++++++ fs/btrfs/disk-io.h | 2 ++ fs/btrfs/extent_io.h | 1 + 3 files changed, 35 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index e6170e142cad..1220d04072c8 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -980,6 +980,38 @@ int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, return ret; } +int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize, + int mirror_num, struct extent_buffer **eb) +{ + struct extent_buffer *buf = NULL; + struct inode *btree_inode = root->fs_info->btree_inode; + struct extent_io_tree *io_tree = &BTRFS_I(btree_inode)->io_tree; + int ret; + + buf = btrfs_find_create_tree_block(root, bytenr, blocksize); + if (!buf) + return 0; + + set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags); + + ret = read_extent_buffer_pages(io_tree, buf, 0, WAIT_PAGE_LOCK, + btree_get_extent, mirror_num); + if (ret) { + free_extent_buffer(buf); + return ret; + } + + if (test_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags)) { + free_extent_buffer(buf); + return -EIO; + } else if (extent_buffer_uptodate(io_tree, buf, NULL)) { + *eb = buf; + } else { + free_extent_buffer(buf); + } + return 0; +} + struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize) { diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index bec3ea4bd67f..b3bdb5c1390f 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -40,6 +40,8 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, u64 parent_transid); int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize, u64 parent_transid); +int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, u32 blocksize, + int mirror_num, struct extent_buffer **eb); struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize); int clean_tree_block(struct btrfs_trans_handle *trans, diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index 6d74c6b34691..fcaf49bcb880 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -32,6 +32,7 @@ #define EXTENT_BUFFER_BLOCKING 1 #define EXTENT_BUFFER_DIRTY 2 #define EXTENT_BUFFER_CORRUPT 3 +#define EXTENT_BUFFER_READAHEAD 4 /* this got triggered by readahead */ /* these are flags for extent_clear_unlock_delalloc */ #define EXTENT_CLEAR_UNLOCK_PAGE 0x1 -- cgit v1.2.3 From 90519d66abbccc251d14719ac76f191f70826e40 Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Mon, 23 May 2011 14:30:00 +0200 Subject: btrfs: state information for readahead Add state information for readahead to btrfs_fs_info and btrfs_device Changes v2: - don't wait in radix_trees - add own set of workers for readahead Reviewed-by: Josef Bacik Signed-off-by: Arne Jansen --- fs/btrfs/ctree.h | 5 +++++ fs/btrfs/disk-io.c | 10 ++++++++++ fs/btrfs/volumes.c | 8 ++++++++ fs/btrfs/volumes.h | 8 ++++++++ 4 files changed, 31 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 03912c5c6f49..f71fd24cc152 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1036,6 +1036,7 @@ struct btrfs_fs_info { struct btrfs_workers endio_freespace_worker; struct btrfs_workers submit_workers; struct btrfs_workers caching_workers; + struct btrfs_workers readahead_workers; /* * fixup workers take dirty pages that didn't properly go through @@ -1119,6 +1120,10 @@ struct btrfs_fs_info { u64 fs_state; struct btrfs_delayed_root *delayed_root; + + /* readahead tree */ + spinlock_t reada_lock; + struct radix_tree_root reada_tree; }; /* diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 1220d04072c8..fd28c40e9ec3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1711,6 +1711,10 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->defrag_inodes = RB_ROOT; fs_info->trans_no_join = 0; + /* readahead state */ + INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_WAIT); + spin_lock_init(&fs_info->reada_lock); + fs_info->thread_pool_size = min_t(unsigned long, num_online_cpus() + 2, 8); @@ -1903,6 +1907,9 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_init_workers(&fs_info->delayed_workers, "delayed-meta", fs_info->thread_pool_size, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->readahead_workers, "readahead", + fs_info->thread_pool_size, + &fs_info->generic_worker); /* * endios are largely parallel and should have a very @@ -1913,6 +1920,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, fs_info->endio_write_workers.idle_thresh = 2; fs_info->endio_meta_write_workers.idle_thresh = 2; + fs_info->readahead_workers.idle_thresh = 2; btrfs_start_workers(&fs_info->workers, 1); btrfs_start_workers(&fs_info->generic_worker, 1); @@ -1926,6 +1934,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, btrfs_start_workers(&fs_info->endio_freespace_worker, 1); btrfs_start_workers(&fs_info->delayed_workers, 1); btrfs_start_workers(&fs_info->caching_workers, 1); + btrfs_start_workers(&fs_info->readahead_workers, 1); fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super); fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages, @@ -2650,6 +2659,7 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->submit_workers); btrfs_stop_workers(&fs_info->delayed_workers); btrfs_stop_workers(&fs_info->caching_workers); + btrfs_stop_workers(&fs_info->readahead_workers); btrfs_close_devices(fs_info->fs_devices); btrfs_mapping_tree_free(&fs_info->mapping_tree); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index f2a4cc79da61..1dccce5bc93d 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -366,6 +366,14 @@ static noinline int device_list_add(const char *path, } INIT_LIST_HEAD(&device->dev_alloc_list); + /* init readahead state */ + spin_lock_init(&device->reada_lock); + device->reada_curr_zone = NULL; + atomic_set(&device->reada_in_flight, 0); + device->reada_next = 0; + INIT_RADIX_TREE(&device->reada_zones, GFP_NOFS & ~__GFP_WAIT); + INIT_RADIX_TREE(&device->reada_extents, GFP_NOFS & ~__GFP_WAIT); + mutex_lock(&fs_devices->device_list_mutex); list_add_rcu(&device->dev_list, &fs_devices->devices); mutex_unlock(&fs_devices->device_list_mutex); diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 6d866db4e177..2a751246188a 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -92,6 +92,14 @@ struct btrfs_device { struct btrfs_work work; struct rcu_head rcu; struct work_struct rcu_work; + + /* readahead state */ + spinlock_t reada_lock; + atomic_t reada_in_flight; + u64 reada_next; + struct reada_zone *reada_curr_zone; + struct radix_tree_root reada_zones; + struct radix_tree_root reada_extents; }; struct btrfs_fs_devices { -- cgit v1.2.3 From 4bb31e928d1a47f5bd046ecb176b8eff7c589fc0 Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Fri, 10 Jun 2011 13:55:54 +0200 Subject: btrfs: hooks for readahead This adds the hooks needed for readahead. In the readpage_end_io_hook, the extent state is checked for the EXTENT_READAHEAD flag. Only in this case the readahead hook is called, to keep the impact on non-ra as low as possible. Additionally, a hook for a failed IO is added, otherwise readahead would wait indefinitely for the extent to finish. Changes for v2: - eliminate race condition Signed-off-by: Arne Jansen --- fs/btrfs/disk-io.c | 37 +++++++++++++++++++++++++++++++++++++ fs/btrfs/extent_io.c | 2 +- 2 files changed, 38 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index fd28c40e9ec3..2151828aa142 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -609,11 +609,47 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end, end = min_t(u64, eb->len, PAGE_CACHE_SIZE); end = eb->start + end - 1; err: + if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) { + clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags); + btree_readahead_hook(root, eb, eb->start, ret); + } + free_extent_buffer(eb); out: return ret; } +static int btree_io_failed_hook(struct bio *failed_bio, + struct page *page, u64 start, u64 end, + struct extent_state *state) +{ + struct extent_io_tree *tree; + unsigned long len; + struct extent_buffer *eb; + struct btrfs_root *root = BTRFS_I(page->mapping->host)->root; + + tree = &BTRFS_I(page->mapping->host)->io_tree; + if (page->private == EXTENT_PAGE_PRIVATE) + goto out; + if (!page->private) + goto out; + + len = page->private >> 2; + WARN_ON(len == 0); + + eb = alloc_extent_buffer(tree, start, len, page); + if (eb == NULL) + goto out; + + if (test_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) { + clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags); + btree_readahead_hook(root, eb, eb->start, -EIO); + } + +out: + return -EIO; /* we fixed nothing */ +} + static void end_workqueue_bio(struct bio *bio, int err) { struct end_io_wq *end_io_wq = bio->bi_private; @@ -3166,6 +3202,7 @@ static int btrfs_cleanup_transaction(struct btrfs_root *root) static struct extent_io_ops btree_extent_io_ops = { .write_cache_pages_lock_hook = btree_lock_page_hook, .readpage_end_io_hook = btree_readpage_end_io_hook, + .readpage_io_failed_hook = btree_io_failed_hook, .submit_bio_hook = btree_submit_bio_hook, /* note we're sharing with inode.c for the merge bio hook */ .merge_bio_hook = btrfs_merge_bio_hook, diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 823028e73cf8..deba714236d7 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1731,7 +1731,7 @@ static void end_bio_extent_readpage(struct bio *bio, int err) if (!uptodate && tree->ops && tree->ops->readpage_io_failed_hook) { ret = tree->ops->readpage_io_failed_hook(bio, page, - start, end, NULL); + start, end, state); if (ret == 0) { uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags); -- cgit v1.2.3