Linux Perf Users
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Tengda Wu" <wutengda@huaweicloud.com>
Cc: linux-perf-users@vger.kernel.org
Subject: Re: [PATCH v5 2/2] perf/core: Disable context fast-swap for sigtrap events
Date: Thu, 19 Nov 2026 10:03:53 +0000	[thread overview]
Message-ID: <sashiko-outbox-222905@kernel.org> (raw)
In-Reply-To: <55b1a2845b0ad34a874a7653644e34af9eb068d3.1795053619.git.wutengda@huaweicloud.com>

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] perf/core: non-SMP-atomic local_t race on
  ctx->nr_no_switch_fast
--

commit 5e829db733e2fe49cdac7cb99bb9428db1c83359
Author: Tengda Wu <wutengda@huaweicloud.com>

perf/core: Disable context fast-swap for sigtrap events

This commit excludes sigtrap events from the context fast-swap path by
incrementing and decrementing ctx->nr_no_switch_fast when adding and
removing events with attr.sigtrap set.

> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index da86bc6e8be94..483ba8704d418 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -1999,7 +1999,7 @@ list_add_event(struct perf_event *event, struct perf_event_context *ctx)
>  		ctx->nr_user++;
>  	if (event->attr.inherit_stat)
>  		ctx->nr_stat++;
> -	if (has_inherit_and_sample_read(&event->attr))
> +	if (has_inherit_and_sample_read(&event->attr) || event->attr.sigtrap)
>  		local_inc(&ctx->nr_no_switch_fast);
>  
>  	if (event->state > PERF_EVENT_STATE_OFF)
> @@ -2225,7 +2225,7 @@ list_del_event(struct perf_event *event, struct perf_event_context *ctx)
>  		ctx->nr_user--;
>  	if (event->attr.inherit_stat)
>  		ctx->nr_stat--;
> -	if (has_inherit_and_sample_read(&event->attr))
> +	if (has_inherit_and_sample_read(&event->attr) || event->attr.sigtrap)
>  		local_dec(&ctx->nr_no_switch_fast);

[Severity: Medium]
Can concurrent local_inc() and local_dec() calls on ctx->nr_no_switch_fast
from different CPUs race and lose updates?

ctx->nr_no_switch_fast is a local_t, so local_inc() and local_dec() are
only atomic on a single CPU (for example, unlocked inc/dec on x86).  When
installing a disabled sigtrap event into a non-empty task context,
perf_install_in_context() runs add_event_to_ctx() directly on the
installer CPU under ctx->lock without an IPI, while an already-active
sigtrap event in that context can update ctx->nr_no_switch_fast on the
target task's CPU without holding ctx->lock:

CPU0                              CPU1 (running ctx->task)
perf_install_in_context()
  raw_spin_lock(&ctx->lock)
  add_event_to_ctx()
    list_add_event()
      local_inc(&ctx->nr_no_switch_fast)
                                  __perf_event_overflow()
                                    local_inc(&ctx->nr_no_switch_fast)
                                      <-- non-SMP-atomic RMW race
                                  ...
                                  perf_pending_task()
                                    local_dec(&ctx->nr_no_switch_fast)

A similar cross-CPU race is possible between list_del_event() in
perf_remove_from_context() (when !ctx->is_active) and perf_pending_task().
A lost increment allows ctx->nr_no_switch_fast to drop to 0 while a
sigtrap event or pending task_work is still present, re-enabling the
context fast-swap path and hitting WARN_ON_ONCE(event->ctx->task !=
current) in perf_sigtrap().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/cover.1795053619.git.wutengda@huaweicloud.com?part=2

      reply	other threads:[~2026-11-19 10:03 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-19  9:50 [PATCH v5 0/2] perf/core: Fix sigtrap task_work races Tengda Wu
2026-11-19  9:51 ` [PATCH v5 1/2] perf/core: Prevent duplicate sigtrap task work enqueue Tengda Wu
2026-11-19  9:59   ` sashiko-bot
2026-11-19  9:51 ` [PATCH v5 2/2] perf/core: Disable context fast-swap for sigtrap events Tengda Wu
2026-11-19 10:03   ` sashiko-bot [this message]

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=sashiko-outbox-222905@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=wutengda@huaweicloud.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