diff options
| author | Zhaolei <zhaolei@cn.fujitsu.com> | 2015-08-05 16:43:30 +0800 |
|---|---|---|
| committer | Chris Mason <clm@fb.com> | 2015-08-09 07:07:13 -0700 |
| commit | 55e3a601c81cdca4497bf855fa4d331f8e830744 (patch) | |
| tree | 66d1f6312a3a5b3c2cdfaf42bab7f2dcb6ebde63 /fs/btrfs | |
| parent | b708ce969af3ceadca18aac5bffe48fe977473b1 (diff) | |
| download | linux-55e3a601c81cdca4497bf855fa4d331f8e830744.tar.bz2 | |
btrfs: Fix data checksum error cause by replace with io-load.
xfstests btrfs/070 sometimes failed.
In my test machine, its fail rate is about 30%.
In another vm(vmware), its fail rate is about 50%.
Reason:
btrfs/070 do replace and defrag with fsstress simultaneously,
after above operation, checksum error is found by scrub.
Actually, it have no relationship with defrag operation, only
replace with fsstress can trigger this bug.
New data writen to target device have possibility rewrited by
old data from source device by replace code in debug, to avoid
above problem, we can set target block group to readonly in
replace period, so new data requested by other operation will
not write to same place with replace code.
Before patch(4.1-rc3):
30% failed in 100 xfstests.
After patch:
0% failed in 300 xfstests.
It also happened in btrfs/071 as it's another scrub with IO load tests.
Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
Diffstat (limited to 'fs/btrfs')
| -rw-r--r-- | fs/btrfs/scrub.c | 34 | ||||
| -rw-r--r-- | fs/btrfs/volumes.c | 2 |
2 files changed, 29 insertions, 7 deletions
diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 281de3f2041d..7555ddc5289f 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -3407,7 +3407,7 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, u64 chunk_tree; u64 chunk_objectid; u64 chunk_offset; - int ret; + int ret = 0; int slot; struct extent_buffer *l; struct btrfs_key key; @@ -3435,8 +3435,14 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) { ret = btrfs_next_leaf(root, path); - if (ret) + if (ret < 0) + break; + if (ret > 0) { + ret = 0; break; + } + } else { + ret = 0; } } @@ -3478,6 +3484,22 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, if (!cache) goto skip; + /* + * we need call btrfs_inc_block_group_ro() with scrubs_paused, + * to avoid deadlock caused by: + * btrfs_inc_block_group_ro() + * -> btrfs_wait_for_commit() + * -> btrfs_commit_transaction() + * -> btrfs_scrub_pause() + */ + scrub_pause_on(fs_info); + ret = btrfs_inc_block_group_ro(root, cache); + scrub_pause_off(fs_info); + if (ret) { + btrfs_put_block_group(cache); + break; + } + dev_replace->cursor_right = found_key.offset + length; dev_replace->cursor_left = found_key.offset; dev_replace->item_needs_writeback = 1; @@ -3517,6 +3539,8 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx, scrub_pause_off(fs_info); + btrfs_dec_block_group_ro(root, cache); + btrfs_put_block_group(cache); if (ret) break; @@ -3539,11 +3563,7 @@ skip: 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; + return ret; } static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx, diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index fbe7c104531c..b7fb0b297f71 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -2785,7 +2785,9 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, return -ENOSPC; /* step one, relocate all the extents inside this chunk */ + btrfs_scrub_pause(root); ret = btrfs_relocate_block_group(extent_root, chunk_offset); + btrfs_scrub_continue(root); if (ret) return ret; |