]> git.ipfire.org Git - thirdparty/kernel/stable.git/commit
sched: fix warning in sched_setaffinity
authorJosh Don <joshdon@google.com>
Mon, 11 Nov 2024 18:27:38 +0000 (10:27 -0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 14 Dec 2024 19:04:11 +0000 (20:04 +0100)
commit5c3fb75f538cfcb886f6dfeb497d99fc2f263ee6
tree3b98cde378b5fcf8d06933b780caa003888b4fda
parent4a2fcf88be76f883aeb2fa523a53d57d346fe9f1
sched: fix warning in sched_setaffinity

[ Upstream commit 70ee7947a29029736a1a06c73a48ff37674a851b ]

Commit 8f9ea86fdf99b added some logic to sched_setaffinity that included
a WARN when a per-task affinity assignment races with a cpuset update.

Specifically, we can have a race where a cpuset update results in the
task affinity no longer being a subset of the cpuset. That's fine; we
have a fallback to instead use the cpuset mask. However, we have a WARN
set up that will trigger if the cpuset mask has no overlap at all with
the requested task affinity. This shouldn't be a warning condition; its
trivial to create this condition.

Reproduced the warning by the following setup:

- $PID inside a cpuset cgroup
- another thread repeatedly switching the cpuset cpus from 1-2 to just 1
- another thread repeatedly setting the $PID affinity (via taskset) to 2

Fixes: 8f9ea86fdf99b ("sched: Always preserve the user requested cpumask")
Signed-off-by: Josh Don <joshdon@google.com>
Acked-and-tested-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Waiman Long <longman@redhat.com>
Tested-by: Madadi Vineeth Reddy <vineethr@linux.ibm.com>
Link: https://lkml.kernel.org/r/20241111182738.1832953-1-joshdon@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
kernel/sched/syscalls.c