Commit 12cf9372 authored Feb 19, 2014 by Wang Shilong Committed by Josef Bacik Mar 10, 2014

Btrfs: device_replace: fix deadlock for nocow case



commit cb7ab021 cause a following deadlock found by
xfstests,btrfs/011:

Thread1 is commiting transaction which is blocked at
btrfs_scrub_pause().

Thread2 is calling btrfs_file_aio_write() which has held
inode's @i_mutex and commit transaction(blocked because
Thread1 is committing transaction).

Thread3 is copy_nocow_page worker which will also try to
hold inode @i_mutex, so thread3 will wait Thread1 finished.

Thread4 is waiting pending workers finished which will wait
Thread3 finished. So the problem is like this:

Thread1--->Thread4--->Thread3--->Thread2---->Thread1

Deadlock happens! we fix it by letting Thread1 go firstly,
which means we won't block transaction commit while we are
waiting pending workers finished.

Reported-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: Wang Shilong <wangsl.fnst@cn.fujitsu.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>

parent 6cf7f77e

fs/btrfs/scrub.c

+15 −2

Original line number	Diff line number	Diff line
		@@ -2686,10 +2686,23 @@ int scrub_enumerate_chunks(struct scrub_ctx *sctx,

		wait_event(sctx->list_wait,
		atomic_read(&sctx->bios_in_flight) == 0);
		atomic_set(&sctx->wr_ctx.flush_all_writes, 0);
		atomic_inc(&fs_info->scrubs_paused);
		wake_up(&fs_info->scrub_pause_wait);

		/*
		* must be called before we decrease @scrub_paused.
		* make sure we don't block transaction commit while
		* we are waiting pending workers finished.
		*/
		wait_event(sctx->list_wait,
		atomic_read(&sctx->workers_pending) == 0);
		scrub_blocked_if_needed(fs_info);
		atomic_set(&sctx->wr_ctx.flush_all_writes, 0);

		mutex_lock(&fs_info->scrub_lock);
		__scrub_blocked_if_needed(fs_info);
		atomic_dec(&fs_info->scrubs_paused);
		mutex_unlock(&fs_info->scrub_lock);
		wake_up(&fs_info->scrub_pause_wait);

		btrfs_put_block_group(cache);
		if (ret)

Admin message