From: Chengming Zhou <zhouchengming@bytedance.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: mingo@redhat.com, acme@kernel.org, mark.rutland@arm.com,
alexander.shishkin@linux.intel.com, jolsa@kernel.org,
namhyung@kernel.org, eranian@google.com,
linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org,
duanxiongchun@bytedance.com, songmuchun@bytedance.com
Subject: Re: [External] Re: [PATCH v2 1/6] perf/core: Fix incosistency between cgroup sched_out and sched_in
Date: Tue, 22 Mar 2022 23:16:27 +0800 [thread overview]
Message-ID: <cdfb252e-9bfc-bee3-7ebe-b8ef401c85dd@bytedance.com> (raw)
In-Reply-To: <YjnjHPuNjo8YTwiQ@hirez.programming.kicks-ass.net>
Hi peter,
On 2022/3/22 10:54 下午, Peter Zijlstra wrote:
> On Tue, Mar 22, 2022 at 09:38:21PM +0800, Chengming Zhou wrote:
>> On 2022/3/22 8:59 下午, Peter Zijlstra wrote:
>>> On Tue, Mar 22, 2022 at 08:08:29PM +0800, Chengming Zhou wrote:
>>>> There is a race problem that can trigger WARN_ON_ONCE(cpuctx->cgrp)
>>>> in perf_cgroup_switch().
>>>>
>>>> CPU1 CPU2
>>>> (in context_switch) (attach running task)
>>>> perf_cgroup_sched_out(prev, next)
>>>> cgrp1 == cgrp2 is True
>>>> next->cgroups = cgrp3
>>>> perf_cgroup_attach()
>>>> perf_cgroup_sched_in(prev, next)
>>>> cgrp1 == cgrp3 is False
>>>>
>>>> The commit a8d757ef076f ("perf events: Fix slow and broken cgroup
>>>> context switch code") would save cpuctx switch out/in when the
>>>> perf_cgroup of "prev" and "next" are the same.
>>>>
>>>> But perf_cgroup of task can change in concurrent with context_switch.
>>>
>>> Can you clarify? IIRC then a task changes cgroup it goes throught the
>>> whole ->attach() dance, and that serializes against the context switch
>>> code.
>>>
>>
>> task->cgroups changed before perf_cgroup_attach(), and is not serialized
>> against the context switch, since task->cgroups can be changed without
>> rq lock held. (cgroup v1 or cgroup v2 with PSI disabled)
>>
>> So perf_cgroup_sched_out() in perf_cgroup_switch() may see the old or
>> new perf_cgroup when do context switch.
>
> __schedule()
> local_irq_disable(); <--- IRQ disable
> rq_lock();
>
> ...
>
> context_switch()
> prepare_task_switch()
> perf_event_task_sched_out()
> __perf_event_task_sched_out()
> perf_cgroup_sched_out();
here compare perf_cgroup_from_task(prev) and perf_cgroup_from_task(next)
>
> switch_to()
> finish_task_switch()
> perf_event_task_sched_in()
> __perf_event_task_sched_in()
> perf_cgroup_sched_in();
here compare perf_cgroup_from_task(prev) and perf_cgroup_from_task(next)
> finish_lock_switch()
> raw_spin_irq_unlock_irq(); <--- IRQ enable
>
>
> vs
>
rcu_assign_pointer(p->cgroups, to) <--- task perf_cgroup changed
task->cgroups has changed before sending IPI
> perf_event_cgrp_subsys.attach = perf_cgroup_attach()
> cgroup_taskset_for_each()
> task_function_call(task, __perf_cgroup_move) <--- sends IPI
>
>
> Please explain how this can interleave.
__perf_cgroup_move in IPI is of course serialized against context switch,
but the task->cgroups has changed before that, without rq lock held.
So perf_cgroup_from_task() may see the old or new perf_cgroup.
Thanks.
next prev parent reply other threads:[~2022-03-22 15:16 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-03-22 12:08 [PATCH v2 0/6] perf/core: Fixes and cleanup for cgroup events Chengming Zhou
2022-03-22 12:08 ` [PATCH v2 1/6] perf/core: Fix incosistency between cgroup sched_out and sched_in Chengming Zhou
2022-03-22 12:59 ` Peter Zijlstra
2022-03-22 13:38 ` [External] " Chengming Zhou
2022-03-22 14:54 ` Peter Zijlstra
2022-03-22 15:16 ` Chengming Zhou [this message]
2022-03-22 15:28 ` Chengming Zhou
2022-03-22 22:06 ` Namhyung Kim
2022-03-23 8:11 ` Peter Zijlstra
2022-03-22 12:08 ` [PATCH v2 2/6] perf/core: Introduce percpu perf_cgroup Chengming Zhou
2022-03-22 13:01 ` Peter Zijlstra
2022-03-22 16:33 ` [External] " Chengming Zhou
2022-03-23 8:13 ` Peter Zijlstra
2022-03-23 12:58 ` Chengming Zhou
2022-03-22 22:21 ` Namhyung Kim
2022-03-22 22:18 ` Namhyung Kim
2022-03-23 1:27 ` [Phishing Risk] [External] " Chengming Zhou
2022-03-23 12:51 ` Peter Zijlstra
2022-03-23 13:07 ` [External] " Chengming Zhou
2022-03-23 13:17 ` Peter Zijlstra
2022-03-23 13:37 ` Chengming Zhou
2022-03-23 14:05 ` Peter Zijlstra
2022-03-23 15:44 ` Chengming Zhou
2022-03-22 12:08 ` [PATCH v2 3/6] perf/core: Don't pass task around when ctx sched in Chengming Zhou
2022-03-22 13:01 ` Peter Zijlstra
2022-03-22 12:08 ` [PATCH v2 4/6] perf/core: Use stable cpuctx->cgrp when update perf cgroup time Chengming Zhou
2022-03-22 13:03 ` Peter Zijlstra
2022-03-22 12:08 ` [PATCH v2 5/6] perf/core: Always set cpuctx cgrp when enable cgroup event Chengming Zhou
2022-03-22 12:08 ` [PATCH v2 6/6] perf/core: Don't need event_filter_match when merge_sched_in() Chengming Zhou
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=cdfb252e-9bfc-bee3-7ebe-b8ef401c85dd@bytedance.com \
--to=zhouchengming@bytedance.com \
--cc=acme@kernel.org \
--cc=alexander.shishkin@linux.intel.com \
--cc=duanxiongchun@bytedance.com \
--cc=eranian@google.com \
--cc=jolsa@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mingo@redhat.com \
--cc=namhyung@kernel.org \
--cc=peterz@infradead.org \
--cc=songmuchun@bytedance.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).