]> git.baikalelectronics.ru Git - kernel.git/commit
ipc: remove memcg accounting for sops objects in do_semtimedop()
authorVasily Averin <vvs@virtuozzo.com>
Sat, 11 Sep 2021 07:40:08 +0000 (10:40 +0300)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 14 Sep 2021 17:22:11 +0000 (10:22 -0700)
commit7a795cc63a02ac8e8a9eefc9d8076a4ff73df7e3
tree4b26d9e2065e934c718e62d24f08adf89ed730a0
parent547271d69249e4b5087364fcafdf35cc9d9f577e
ipc: remove memcg accounting for sops objects in do_semtimedop()

Linus proposes to revert an accounting for sops objects in
do_semtimedop() because it's really just a temporary buffer
for a single semtimedop() system call.

This object can consume up to 2 pages, syscall is sleeping
one, size and duration can be controlled by user, and this
allocation can be repeated by many thread at the same time.

However Shakeel Butt pointed that there are much more popular
objects with the same life time and similar memory
consumption, the accounting of which was decided to be
rejected for performance reasons.

Considering at least 2 pages for task_struct and 2 pages for
the kernel stack, a back of the envelope calculation gives a
footprint amplification of <1.5 so this temporal buffer can be
safely ignored.

The factor would IMO be interesting if it was >> 2 (from the
PoV of excessive (ab)use, fine-grained accounting seems to be
currently unfeasible due to performance impact).

Link: https://lore.kernel.org/lkml/90e254df-0dfe-f080-011e-b7c53ee7fd20@virtuozzo.com/
Fixes: ea9a3151b947 ("memcg: enable accounting of ipc resources")
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
Acked-by: Shakeel Butt <shakeelb@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
ipc/sem.c