All of lore.kernel.org
 help / color / mirror / Atom feed
From: Oleg Nesterov <oleg@redhat.com>
To: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>,
	Ingo Molnar <mingo@elte.hu>,
	Alan Stern <stern@rowland.harvard.edu>,
	Arnaldo Carvalho de Melo <acme@redhat.com>,
	Paul Mackerras <paulus@samba.org>,
	Prasad <prasad@linux.vnet.ibm.com>,
	Roland McGrath <roland@redhat.com>,
	linux-kernel@vger.kernel.org
Subject: Re: Q: perf_install_in_context/perf_event_enable are racy?
Date: Thu, 27 Jan 2011 17:57:12 +0100	[thread overview]
Message-ID: <20110127165712.GC25060@redhat.com> (raw)
In-Reply-To: <1296134077.15234.161.camel@laptop>

On 01/27, Peter Zijlstra wrote:
>
> Right, so the fact of introducing extra scheduling makes me feel
> uncomfortable... the whole purpose is to observe without perturbing (as
> much as possible).

Yes, agreed.

Well, otoh the patch removes the code which sets ->task_ctx from
__perf_install_in_context() and __perf_event_enable(), and perhaps
we could simplify the things further, but anyway I agree.

> Should I think suffice to get the ctx in sync with the task state, we've
> got the following cases:
>  1) task is in the middle of scheduling in
>  2) task is in the middle of scheduling out
>  3) task is running
>
> Without __ARCH_WANT_INTERRUPT_ON_CTXSW everything is boring and works,
> 1: the IPI will be delayed until 3, 2: the IPI finds another task and
> the next schedule in will sort things.
>
> With, however, things are more interesting. 2 seems to be adequately
> covered by the patch I just send, the IPI will bail and the next
> sched-in of the relevant task will pick matters up. 1 otoh doesn't seem
> covered, the IPI will bail, leaving us stranded.

Hmm, yes... Strangely, I missed that when I was thinking about in_ctxsw.

Perhaps, we can change task_oncpu_function_call() so that it returns
-EAGAIN in case it hits in_ctxsw != 0? If the caller sees -EAGAIN, it
should always retry even if !ctx->is_active.

Oleg.


  parent reply	other threads:[~2011-01-27 17:05 UTC|newest]

Thread overview: 91+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-11-08 14:56 Q: perf_event && task->ptrace_bps[] Oleg Nesterov
2010-11-08 14:57 ` Q: sys_perf_event_open() && PF_EXITING Oleg Nesterov
2011-01-19 18:21   ` [PATCH 0/2] Was: " Oleg Nesterov
2011-01-19 18:22     ` [PATCH 1/2] perf: fix find_get_context() vs perf_event_exit_task() race Oleg Nesterov
2011-01-19 18:49       ` Peter Zijlstra
2011-01-19 19:18       ` [tip:perf/urgent] perf: Fix " tip-bot for Oleg Nesterov
2011-01-21 15:29         ` Ingo Molnar
2011-01-21 15:53           ` Oleg Nesterov
2011-01-21 17:45             ` [PATCH] perf: perf_event_exit_task_context: s/rcu_dereference/rcu_dereference_raw/ Oleg Nesterov
2011-01-21 17:53               ` Oleg Nesterov
2011-01-21 21:50                 ` Paul E. McKenney
2011-01-24 11:51                   ` Oleg Nesterov
2011-01-21 22:12               ` [tip:perf/urgent] " tip-bot for Oleg Nesterov
2011-01-19 18:22     ` [PATCH 2/2] perf: fix perf_event_init_task()/perf_event_free_task() interaction Oleg Nesterov
2011-01-19 18:51       ` Peter Zijlstra
2011-01-19 19:19       ` [tip:perf/urgent] perf: Fix " tip-bot for Oleg Nesterov
2011-01-20 19:30     ` Q: perf_install_in_context/perf_event_enable are racy? Oleg Nesterov
2011-01-21 12:11       ` Peter Zijlstra
2011-01-21 13:03         ` Ingo Molnar
2011-01-21 13:39           ` Peter Zijlstra
2011-01-21 14:26             ` Oleg Nesterov
2011-01-21 15:05               ` Peter Zijlstra
2011-01-21 20:40                 ` Frederic Weisbecker
2011-01-24 11:42                   ` Oleg Nesterov
2011-01-26 17:53                     ` Oleg Nesterov
2011-01-26 18:49                       ` Oleg Nesterov
2011-01-26 18:51                         ` [PATCH] fix the theoretical task_cpu/task_curr problem in kick_process/task_oncpu_function_call Oleg Nesterov
2011-01-26 19:05                         ` Q: perf_install_in_context/perf_event_enable are racy? Peter Zijlstra
2011-01-26 19:33                           ` Peter Zijlstra
2011-01-26 19:38                             ` Peter Zijlstra
2011-01-26 21:19                             ` Oleg Nesterov
2011-01-26 21:33                               ` Oleg Nesterov
2011-01-27 10:32                                 ` Peter Zijlstra
2011-01-27 12:29                                   ` Peter Zijlstra
2011-01-27 16:10                                     ` Oleg Nesterov
2011-01-27 16:27                                       ` Peter Zijlstra
2011-01-27 16:59                                         ` Oleg Nesterov
2011-01-27 15:52                                   ` Oleg Nesterov
2011-01-27 13:14                       ` Peter Zijlstra
2011-01-27 14:28                         ` Peter Zijlstra
2011-01-27 14:58                           ` Peter Zijlstra
2011-01-27 16:57                         ` Oleg Nesterov [this message]
2011-01-27 17:11                           ` Peter Zijlstra
2011-01-27 22:18                             ` Oleg Nesterov
2011-01-28 11:52                               ` Peter Zijlstra
2011-01-28 14:57                                 ` Peter Zijlstra
2011-01-28 16:28                                   ` Oleg Nesterov
2011-01-28 18:11                                     ` Peter Zijlstra
2011-01-31 17:26                                       ` Oleg Nesterov
2011-01-31 18:23                                         ` Peter Zijlstra
2011-01-31 19:11                                           ` Oleg Nesterov
2011-01-31 19:29                                             ` Peter Zijlstra
2011-02-01 14:03                                               ` [PATCH] perf: Cure task_oncpu_function_call() races Peter Zijlstra
2011-02-01 17:27                                                 ` Oleg Nesterov
2011-02-01 18:08                                                   ` Peter Zijlstra
2011-02-01 18:18                                                     ` Peter Zijlstra
2011-02-01 21:00                                                     ` Peter Zijlstra
2010-11-08 14:57 ` Q: perf_event && event->owner Oleg Nesterov
2010-11-08 20:11   ` Frederic Weisbecker
2010-11-08 20:41     ` Peter Zijlstra
2010-11-09 16:18       ` Oleg Nesterov
2010-11-09 15:57     ` Oleg Nesterov
2010-11-09 16:56       ` Peter Zijlstra
2010-11-09 16:58         ` Oleg Nesterov
2010-11-09 17:07           ` Peter Zijlstra
2010-11-09 17:42             ` Oleg Nesterov
2010-11-09 18:01               ` Peter Zijlstra
2010-11-09 18:57                 ` Oleg Nesterov
2010-11-09 19:16                   ` Peter Zijlstra
2010-11-10 15:17                   ` Peter Zijlstra
2010-11-10 15:44                     ` Oleg Nesterov
2010-11-12 15:48                       ` Peter Zijlstra
2010-11-12 18:49                         ` Oleg Nesterov
2010-11-18 14:09                         ` [tip:perf/urgent] perf: Fix owner-list vs exit tip-bot for Peter Zijlstra
2010-11-08 18:41 ` Q: perf_event && task->ptrace_bps[] Frederic Weisbecker
2010-11-08 19:18   ` Oleg Nesterov
2011-01-17 23:58     ` Frederic Weisbecker
2011-01-18  1:16       ` Roland McGrath
2011-01-17 20:34 ` Oleg Nesterov
2011-01-17 20:52   ` Peter Zijlstra
2011-01-17 21:01     ` Frederic Weisbecker
2011-01-18 16:09     ` [PATCH 0/2] perf: event->cpu checking fixes Oleg Nesterov
2011-01-18 16:10       ` [PATCH 1/2] perf: find_get_context: fix the per-cpu-counter check Oleg Nesterov
2011-01-18 19:06         ` [tip:perf/urgent] perf: Find_get_context: " tip-bot for Oleg Nesterov
2011-01-18 16:10       ` [PATCH 2/2] perf: validate cpu early in perf_event_alloc() Oleg Nesterov
2011-01-18 19:07         ` [tip:perf/urgent] perf: Validate " tip-bot for Oleg Nesterov
2011-01-18 18:42   ` Q: perf_event && task->ptrace_bps[] Frederic Weisbecker
2011-01-19 15:37     ` Oleg Nesterov
2011-01-19 20:05       ` Frederic Weisbecker
2011-01-20 17:28         ` Oleg Nesterov
2011-01-28 17:41           ` Frederic Weisbecker

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=20110127165712.GC25060@redhat.com \
    --to=oleg@redhat.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=acme@redhat.com \
    --cc=fweisbec@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=paulus@samba.org \
    --cc=prasad@linux.vnet.ibm.com \
    --cc=roland@redhat.com \
    --cc=stern@rowland.harvard.edu \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.