btrfs: get rid of unique workqueue helper functions
authorOmar Sandoval <osandov@fb.com>
Mon, 16 Sep 2019 18:30:57 +0000 (11:30 -0700)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 9 Jan 2020 09:20:06 +0000 (10:20 +0100)
commit2bae3ee327c972896543c8f2f190568bbcf64b54
treeb210125fea67ba90ceaf84fdaa9480482984cebf
parent0c261ca281733557413a2d93c8fd9fd2eec6156e
btrfs: get rid of unique workqueue helper functions

[ Upstream commit a0cac0ec961f0d42828eeef196ac2246a2f07659 ]

Commit 9e0af2376434 ("Btrfs: fix task hang under heavy compressed
write") worked around the issue that a recycled work item could get a
false dependency on the original work item due to how the workqueue code
guarantees non-reentrancy. It did so by giving different work functions
to different types of work.

However, the fixes in the previous few patches are more complete, as
they prevent a work item from being recycled at all (except for a tiny
window that the kernel workqueue code handles for us). This obsoletes
the previous fix, so we don't need the unique helpers for correctness.
The only other reason to keep them would be so they show up in stack
traces, but they always seem to be optimized to a tail call, so they
don't show up anyways. So, let's just get rid of the extra indirection.

While we're here, rename normal_work_helper() to the more informative
btrfs_work_helper().

Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
12 files changed:
fs/btrfs/async-thread.c
fs/btrfs/async-thread.h
fs/btrfs/block-group.c
fs/btrfs/delayed-inode.c
fs/btrfs/disk-io.c
fs/btrfs/inode.c
fs/btrfs/ordered-data.c
fs/btrfs/qgroup.c
fs/btrfs/raid56.c
fs/btrfs/reada.c
fs/btrfs/scrub.c
fs/btrfs/volumes.c