mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Guopeng Zhang <guopeng.zhang@linux.dev>
To: "Tejun Heo" <tj@kernel.org>,
	"Johannes Weiner" <hannes@cmpxchg.org>,
	"Michal Koutný" <mkoutny@suse.com>
Cc: cgroups@vger.kernel.org, linux-kernel@vger.kernel.org,
	Guopeng Zhang <zhangguopeng@kylinos.cn>
Subject: [PATCH] cgroup/pids: Restore pids.events notifications in local mode
Date: Thu, 24 Sep 2026 17:15:36 +0800	[thread overview]
Message-ID: <20260924091536.63572-1-guopeng.zhang@linux.dev> (raw)

From: Guopeng Zhang <zhangguopeng@kylinos.cn>

A fork rejected by the pids controller increments the counter reported by
pids.events. When local event accounting is selected, however, pids_event()
returns after notifying only events_local_file, leaving pids.events pollers
asleep.

On legacy hierarchies, pids.events.local does not exist. With
pids_localevents, pids.events reports the same local counter. In both
cases, pids.events changes without generating a notification.

This can be reproduced with a pids_localevents mount:

    mkdir /tmp/test
    mount -t cgroup2 -o pids_localevents none /tmp/test
    mkdir /tmp/test/t
    echo 1 > /tmp/test/t/pids.max
    cat /tmp/test/t/pids.events                 # max 0
    timeout 3 inotifywait -e modify /tmp/test/t/pids.events &
    sh -c 'echo $$ > /tmp/test/t/cgroup.procs; (true &)' 2>/dev/null
    wait
    cat /tmp/test/t/pids.events                 # max 1

Without this patch, inotifywait times out without reporting an event.
Notify pids.events before returning from the local event path.

Fixes: 3f26a885a068 ("cgroup/pids: Add pids.events.local")
Signed-off-by: Guopeng Zhang <zhangguopeng@kylinos.cn>
---
 kernel/cgroup/pids.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/kernel/cgroup/pids.c b/kernel/cgroup/pids.c
index ecbb839d2acb..8444e270554a 100644
--- a/kernel/cgroup/pids.c
+++ b/kernel/cgroup/pids.c
@@ -253,6 +253,11 @@ static void pids_event(struct pids_cgroup *pids_forking,
 	}
 	if (!cgroup_subsys_on_dfl(pids_cgrp_subsys) ||
 	    cgrp_dfl_root.flags & CGRP_ROOT_PIDS_LOCAL_EVENTS) {
+		/*
+		 * pids.events reports the local counter on legacy hierarchies and
+		 * when pids_localevents is enabled.
+		 */
+		cgroup_file_notify(&p->events_file);
 		cgroup_file_notify(&p->events_local_file);
 		return;
 	}
-- 
2.43.0


             reply	other threads:[~2026-09-24  9:16 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-24  9:15 Guopeng Zhang [this message]
2026-09-24 16:42 ` Tejun Heo

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=20260924091536.63572-1-guopeng.zhang@linux.dev \
    --to=guopeng.zhang@linux.dev \
    --cc=cgroups@vger.kernel.org \
    --cc=hannes@cmpxchg.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mkoutny@suse.com \
    --cc=tj@kernel.org \
    --cc=zhangguopeng@kylinos.cn \
    /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

all inboxes | Powered by JetHome®