linux-kernel - Re: [PATCH v1 5/5] relayfs: uniformally use possible cpu iteration

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAL+tcoCN56fmQadhikyeg-PziM8OF7PitXufPR-rfGb1ko5Gcg@mail.gmail.com>
Date: Tue, 13 May 2025 13:52:22 +0800
From: Jason Xing <kerneljasonxing@...il.com>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: axboe@...nel.dk, rostedt@...dmis.org, mhiramat@...nel.org, 
	mathieu.desnoyers@...icios.com, linux-kernel@...r.kernel.org, 
	linux-block@...r.kernel.org, linux-trace-kernel@...r.kernel.org, 
	Jason Xing <kernelxing@...cent.com>, Yushan Zhou <katrinzhou@...cent.com>
Subject: Re: [PATCH v1 5/5] relayfs: uniformally use possible cpu iteration

On Tue, May 13, 2025 at 8:52 AM Andrew Morton <akpm@...ux-foundation.org> wrote:
>
> On Mon, 12 May 2025 10:49:35 +0800 Jason Xing <kerneljasonxing@...il.com> wrote:
>
> > From: Jason Xing <kernelxing@...cent.com>
> >
> > Use for_each_possible_cpu to create per-cpu relayfs file to avoid later
> > hotplug cpu which doesn't have its own file.
>
> I don't understand this.  Exactly what problem are we trying to solve?
>
> > Reviewed-by: Yushan Zhou <katrinzhou@...cent.com>
> > Signed-off-by: Jason Xing <kernelxing@...cent.com>
> > ---
> >  kernel/relay.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/kernel/relay.c b/kernel/relay.c
> > index 27f7e701724f..dcb099859e83 100644
> > --- a/kernel/relay.c
> > +++ b/kernel/relay.c
> > @@ -519,7 +519,7 @@ struct rchan *relay_open(const char *base_filename,
> >       kref_init(&chan->kref);
> >
> >       mutex_lock(&relay_channels_mutex);
> > -     for_each_online_cpu(i) {
> > +     for_each_possible_cpu(i) {
>
> num_possible_cpus() can sometimes greatly exceed num_online_cpus(), so
> this is an unfortunate change.  It would be better to implement the
> hotplug notifier?

For the record, the hotplug notifier has been retired by the commit
530e9b76ae8f("cpu/hotplug: Remove obsolete cpu hotplug
register/unregister functions"). And now in relay, a similar feature
called hotplug state machine has already been implemented by the
commit e6d4989a9ad1 ("relayfs: Convert to hotplug state machine"). So
the relay has hotplug support.

Sorry for missing this point. I would drop this patch.

Thanks,
Jason

>
> >               buf = relay_open_buf(chan, i);
> >               if (!buf)
> >                       goto free_bufs;
> > @@ -615,7 +615,7 @@ int relay_late_setup_files(struct rchan *chan,
> >        * no files associated. So it's safe to call relay_setup_buf_file()
> >        * on all currently online CPUs.
> >        */
> > -     for_each_online_cpu(i) {
> > +     for_each_possible_cpu(i) {
> >               buf = *per_cpu_ptr(chan->buf, i);
> >               if (unlikely(!buf)) {
> >                       WARN_ONCE(1, KERN_ERR "CPU has no buffer!\n");
> > --
> > 2.43.5