Re: [PATCH] bdi: Fix another oops in wb_workfn()
From: Linus Torvalds <torvalds@linux-foundation.org>
Date: 2018-06-13 14:06:56
Also in:
linux-fsdevel, lkml
Possibly related (same subject, not in this thread)
- 2018-06-22 · Re: [PATCH] bdi: Fix another oops in wb_workfn() · Jens Axboe <axboe@kernel.dk>
- 2018-06-22 · Re: [PATCH] bdi: Fix another oops in wb_workfn() · Jan Kara <jack@suse.cz>
- 2018-06-19 · Re: [PATCH] bdi: Fix another oops in wb_workfn() · Jan Kara <jack@suse.cz>
- 2018-06-18 · Re: [PATCH] bdi: Fix another oops in wb_workfn() · Tejun Heo <tj@kernel.org>
- 2018-06-18 · Re: [PATCH] bdi: Fix another oops in wb_workfn() · Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
On Wed, Jun 13, 2018 at 3:44 AM Tetsuo Handa [off-list ref] wrote:
Can't we utilize RCU grace period (like shown below) ?
_Please_ don't do conditional locking. Particularly not the kind where
then a function drops the lock that was taken in another function -
and only does it based on some parameter.
Yes, we have cases where we do it, but it's seldom a great idea, and
it _really_ makes for subtle code (and subtle bugs).
Linus