From: Peter Zijlstra <peterz@infradead.org>
To: Thomas-Mich Richter <tmricht@linux.ibm.com>
Cc: Kees Cook <keescook@chromium.org>,
acme@redhat.com,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
Hendrik Brueckner <brueckner@linux.ibm.com>,
Martin Schwidefsky <schwidefsky@de.ibm.com>
Subject: Re: WARN_ON_ONCE() hit at kernel/events/core.c:330
Date: Thu, 4 Apr 2019 14:02:14 +0200 [thread overview]
Message-ID: <20190404120214.GD14281@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <20190404110909.GY4038@hirez.programming.kicks-ass.net>
On Thu, Apr 04, 2019 at 01:09:09PM +0200, Peter Zijlstra wrote:
> That is not entirely the scenario I talked about, but *groan*.
>
> So what I meant was:
>
> CPU-0 CPU-n
>
> __schedule()
> local_irq_disable()
>
> ...
> deactivate_task(prev);
>
> try_to_wake_up(@p)
> ...
> smp_cond_load_acquire(&p->on_cpu, !VAL);
>
> <PMI>
> ..
> perf_event_disable_inatomic()
> event->pending_disable = 1;
> irq_work_queue() /* self-IPI */
> </PMI>
>
> context_switch()
> prepare_task_switch()
> perf_event_task_sched_out()
> // the above chain that clears pending_disable
>
> finish_task_switch()
> finish_task()
> smp_store_release(prev->on_cpu, 0);
> /* finally.... */
> // take woken
> // context_switch to @p
> finish_lock_switch()
> raw_spin_unlock_irq()
> /* w00t, IRQs enabled, self-IPI time */
> <self-IPI>
> perf_pending_event()
> // event->pending_disable == 0
> </self-IPI>
>
>
> What you're suggesting, is that the time between:
>
> smp_store_release(prev->on_cpu, 0);
>
> and
>
> <self-IPI>
>
> on CPU-0 is sufficient for CPU-n to context switch to the task, enable
> the event there, trigger a PMI that calls perf_event_disable_inatomic()
> _again_ (this would mean irq_work_queue() failing, which we don't check)
> (and schedule out again, although that's not required).
>
> This being virt that might actually be possible if (v)CPU-0 takes a nap
> I suppose.
>
> Let me think about this a little more...
Arghh... s390 doesn't implement arch_irq_work_raise(), which makes it
far far worse.
I have a hack that might've cured it, were it not for that. Let me think
more still..
next prev parent reply other threads:[~2019-04-04 12:02 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-04-03 9:47 WARN_ON_ONCE() hit at kernel/events/core.c:330 Thomas-Mich Richter
2019-04-03 10:41 ` Peter Zijlstra
2019-04-03 11:26 ` Thomas-Mich Richter
2019-04-04 9:15 ` Thomas-Mich Richter
2019-04-04 11:09 ` Peter Zijlstra
2019-04-04 12:02 ` Peter Zijlstra [this message]
2019-04-04 12:13 ` Peter Zijlstra
2019-04-04 13:03 ` Peter Zijlstra
2019-04-04 13:21 ` Thomas-Mich Richter
2019-04-05 10:18 ` Thomas-Mich Richter
2019-04-05 11:46 ` Peter Zijlstra
2019-04-08 7:12 ` Thomas-Mich Richter
2019-04-08 8:22 ` Peter Zijlstra
2019-04-08 8:47 ` Thomas-Mich Richter
2019-04-08 9:50 ` Peter Zijlstra
2019-04-08 13:28 ` Thomas-Mich Richter
2019-04-09 6:07 ` Thomas-Mich Richter
2019-04-09 8:29 ` Peter Zijlstra
2019-04-09 8:53 ` Mark Rutland
2019-04-10 13:51 ` Thomas-Mich Richter
2019-04-10 14:33 ` Peter Zijlstra
2019-04-11 12:06 ` Alexander Shishkin
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=20190404120214.GD14281@hirez.programming.kicks-ass.net \
--to=peterz@infradead.org \
--cc=acme@redhat.com \
--cc=brueckner@linux.ibm.com \
--cc=heiko.carstens@de.ibm.com \
--cc=keescook@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=schwidefsky@de.ibm.com \
--cc=tmricht@linux.ibm.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).