From d1b5c5671d010de1df78d3efddb84bf22bfafd1e Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Wed, 19 Aug 2015 14:17:40 +0200 Subject: [PATCH] btrfs: Prevent from early transaction abort Btrfs relies on GFP_NOFS allocation when committing the transaction but this allocation context is rather weak wrt. reclaim capabilities. The page allocator currently tries hard to not fail these allocations if they are small (<=PAGE_ALLOC_COSTLY_ORDER) so this is not a problem currently but there is an attempt to move away from the default no-fail behavior and allow these allocation to fail more eagerly. And this would lead to a pre-mature transaction abort as follows: [ 55.328093] Call Trace: [ 55.328890] [] dump_stack+0x4f/0x7b [ 55.330518] [] ? console_unlock+0x334/0x363 [ 55.332738] [] __alloc_pages_nodemask+0x81d/0x8d4 [ 55.334910] [] pagecache_get_page+0x10e/0x20c [ 55.336844] [] alloc_extent_buffer+0xd0/0x350 [btrfs] [ 55.338973] [] btrfs_find_create_tree_block+0x15/0x17 [btrfs] [ 55.341329] [] btrfs_alloc_tree_block+0x18c/0x405 [btrfs] [ 55.343566] [] split_leaf+0x1e4/0x6a6 [btrfs] [ 55.345577] [] btrfs_search_slot+0x671/0x831 [btrfs] [ 55.347679] [] ? get_parent_ip+0xe/0x3e [ 55.349434] [] btrfs_insert_empty_items+0x5d/0xa8 [btrfs] [ 55.351681] [] __btrfs_run_delayed_refs+0x7a6/0xf35 [btrfs] [ 55.353979] [] btrfs_run_delayed_refs+0x6e/0x226 [btrfs] [ 55.356212] [] ? start_transaction+0x192/0x534 [btrfs] [ 55.358378] [] ? start_transaction+0x192/0x534 [btrfs] [ 55.360626] [] btrfs_commit_transaction+0x4c/0xaba [btrfs] [ 55.362894] [] ? start_transaction+0x192/0x534 [btrfs] [ 55.365221] [] btrfs_sync_file+0x29c/0x310 [btrfs] [ 55.367273] [] vfs_fsync_range+0x8f/0x9e [ 55.369047] [] vfs_fsync+0x1c/0x1e [ 55.370654] [] do_fsync+0x34/0x4e [ 55.372246] [] SyS_fsync+0x10/0x14 [ 55.373851] [] system_call_fastpath+0x12/0x6f [ 55.381070] BTRFS: error (device hdb1) in btrfs_run_delayed_refs:2821: errno=-12 Out of memory [ 55.382431] BTRFS warning (device hdb1): Skipping commit of aborted transaction. [ 55.382433] BTRFS warning (device hdb1): cleanup_transaction:1692: Aborting unused transaction(IO failure). [ 55.384280] ------------[ cut here ]------------ [ 55.384312] WARNING: CPU: 0 PID: 3010 at fs/btrfs/delayed-ref.c:438 btrfs_select_ref_head+0xd9/0xfe [btrfs]() [...] [ 55.384337] Call Trace: [ 55.384353] [] dump_stack+0x4f/0x7b [ 55.384357] [] ? down_trylock+0x2d/0x37 [ 55.384359] [] warn_slowpath_common+0xa1/0xbb [ 55.384398] [] ? btrfs_select_ref_head+0xd9/0xfe [btrfs] [ 55.384400] [] warn_slowpath_null+0x1a/0x1c [ 55.384423] [] btrfs_select_ref_head+0xd9/0xfe [btrfs] [ 55.384446] [] ? __btrfs_run_delayed_refs+0xa2/0xf35 [btrfs] [ 55.384455] [] __btrfs_run_delayed_refs+0xab/0xf35 [btrfs] [ 55.384476] [] btrfs_run_delayed_refs+0x6e/0x226 [btrfs] [ 55.384499] [] ? start_transaction+0x192/0x534 [btrfs] [ 55.384521] [] ? start_transaction+0x192/0x534 [btrfs] [ 55.384543] [] btrfs_commit_transaction+0x4c/0xaba [btrfs] [ 55.384565] [] ? start_transaction+0x192/0x534 [btrfs] [ 55.384588] [] btrfs_sync_file+0x29c/0x310 [btrfs] [ 55.384591] [] vfs_fsync_range+0x8f/0x9e [ 55.384592] [] vfs_fsync+0x1c/0x1e [ 55.384593] [] do_fsync+0x34/0x4e [ 55.384594] [] SyS_fsync+0x10/0x14 [ 55.384595] [] system_call_fastpath+0x12/0x6f [...] [ 55.384608] ---[ end trace c29799da1d4dd621 ]--- [ 55.437323] BTRFS info (device hdb1): forced readonly [ 55.438815] BTRFS info (device hdb1): delayed_refs has NO entry Fix this by being explicit about the no-fail behavior of this allocation path and use __GFP_NOFAIL. Signed-off-by: Michal Hocko Signed-off-by: Chris Mason --- fs/btrfs/extent_io.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index b9755ce98218..3cfbd6261f9b 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4624,9 +4624,7 @@ __alloc_extent_buffer(struct btrfs_fs_info *fs_info, u64 start, { struct extent_buffer *eb = NULL; - eb = kmem_cache_zalloc(extent_buffer_cache, GFP_NOFS); - if (eb == NULL) - return NULL; + eb = kmem_cache_zalloc(extent_buffer_cache, GFP_NOFS|__GFP_NOFAIL); eb->start = start; eb->len = len; eb->fs_info = fs_info; @@ -4884,7 +4882,7 @@ struct extent_buffer *alloc_extent_buffer(struct btrfs_fs_info *fs_info, return NULL; for (i = 0; i < num_pages; i++, index++) { - p = find_or_create_page(mapping, index, GFP_NOFS); + p = find_or_create_page(mapping, index, GFP_NOFS|__GFP_NOFAIL); if (!p) goto free_eb;