From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 5A453C433EF for ; Tue, 30 Nov 2021 15:42:44 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S243040AbhK3PqA (ORCPT ); Tue, 30 Nov 2021 10:46:00 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:45052 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S242750AbhK3Pph (ORCPT ); Tue, 30 Nov 2021 10:45:37 -0500 Received: from mail-lj1-x22c.google.com (mail-lj1-x22c.google.com [IPv6:2a00:1450:4864:20::22c]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 2D6C8C061574 for ; Tue, 30 Nov 2021 07:42:17 -0800 (PST) Received: by mail-lj1-x22c.google.com with SMTP id p8so28445154ljo.5 for ; Tue, 30 Nov 2021 07:42:17 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=WrERzcuCrxtLZTcbKN8avYqKjR8tbdUKn4Cb74q01M4=; b=MZua7H2uF4UG5Jve7ZuPf8BSEgCbsdVrVY1e7Pi+gqQYDCQ6RJmylLp8XhAQp81VuA Es5mJFMg+SbEEk1d54ZhvTtx2HLBaCuRxcdE8ZLQUF4U5f65eacXucyrucz8U2vUf1oa vP9EnmCyNSvhZYv25dw5NuluClmj9jSHtpP8lXh+5HB16xo0HDy/q5hRjalHXTmeoItN Yn4vzNUzA3+rb0J8zWHWh/aUC+5TAy43jSLJWkQGvTHLOHSjDiyLPVQHYRxXdSFC6j0t dbZCaDPwsU373AUEE0zLinKt0JqqBot8PWF3AIMiDjoGg6ADsW6xBJduIBWNQAT/Zu5M 2Otg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=WrERzcuCrxtLZTcbKN8avYqKjR8tbdUKn4Cb74q01M4=; b=lGY0Bsk7DBlBevT/i6wEy8GHHatewuiONVZASFQroBQ8dnwdxKnYczrS5q80XbJgB0 iEs4uQfzVLCQtQnk8r3r248QvrdwK0X9RUl3obSPUNliHDFoftnIrfxFcTr3K2emC4kc kX8Ug+ln3h6i/4UIR6oZp4lwct0jWSMinqS315NxRAH2BYScj8Zvcpii3BzHlDbQTZcN uNowjug6sV0C3/lBRPrNnFivgNfThqurQ7KK6CRT+WpHWEEOhAD0FSI/tShY+O99tmVn vyjwhWFJTGLjeTlySE4R3I3TYsouhil4w1eExG4h/AdWeDrSkehXPzxZCB0OIYOn/tb0 0AOA== X-Gm-Message-State: AOAM530Y0Gl7eViO0uYUE9W1tf267K7bjXsbA45aO8EyANCA+gr8i6dg ZyLm+fKuIINSsU2ATJRPGmgCKwdgK66r77i5ifMmfA== X-Google-Smtp-Source: ABdhPJxgCc/HfucoGypUL3oVKe7qh7LKQtNb68rDS7w9XORgw29qdTqWjIpeysOyzi/hH1FBsR7iPOaYRMh7W0q2RVY= X-Received: by 2002:a2e:94e:: with SMTP id 75mr15729578ljj.494.1638286935378; Tue, 30 Nov 2021 07:42:15 -0800 (PST) MIME-Version: 1.0 References: <8735nkcwov.mognet@arm.com> <87zgpsb6de.mognet@arm.com> <87sfvjavqk.mognet@arm.com> <87pmqmc16f.mognet@arm.com> <20211126171817.GA3798214@ubiquitous> <20211129164545.GA3981328@ubiquitous> In-Reply-To: <20211129164545.GA3981328@ubiquitous> From: Vincent Guittot Date: Tue, 30 Nov 2021 16:42:03 +0100 Message-ID: Subject: Re: [PATCH] sched/fair: Fix detection of per-CPU kthreads waking a task To: Vincent Donnefort Cc: Valentin Schneider , peterz@infradead.org, mingo@redhat.com, linux-kernel@vger.kernel.org, mgorman@techsingularity.net, dietmar.eggemann@arm.com Content-Type: text/plain; charset="UTF-8" Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 29 Nov 2021 at 17:54, Vincent Donnefort wrote: > > [...] > > > > > > > > > > > still i don't see the need of !is_idle_task(current) > > > > > > > > > > > > > Admittedly, belts and braces. The existing condition checks rq->nr_running <= 1 > > > > which can lead to coscheduling when the wakeup is issued by the idle task > > > > (or even if rq->nr_running == 0, you can have rq->ttwu_pending without > > > > having sent an IPI due to polling). Essentially this overrides the first > > > > check in sis() that uses idle_cpu(target) (prev == smp_processor_id() == > > > > target). > > > > > > > > I couldn't prove such wakeups can happen right now, but if/when they do > > > > (AIUI it would just take someone to add a wake_up_process() down some > > > > smp_call_function() callback) then we'll need the above. If you're still > > > > not convinced by now, I won't push it further. > > > > > > From a quick experiment, even with the asym_fits_capacity(), I can trigger > > > the following: > > > > > > [ 0.118855] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > > > [ 0.128214] select_idle_sibling: wakee=rcu_gp:3 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > > > [ 0.137327] select_idle_sibling: wakee=rcu_par_gp:4 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > > > [ 0.147221] select_idle_sibling: wakee=kworker/u16:0:7 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > > > [ 0.156994] select_idle_sibling: wakee=mm_percpu_wq:8 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > > > > Timestamp shows its booting phase and thread name above shows per cpu > > thread. Could it happen just while creating per cpu thread at boot and > > as a result not relevant ? > > I have more of those logs a bit later in the boot: > > [ 0.484791] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > [ 0.516495] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > [ 0.525758] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > [ 0.535078] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > [ 0.547486] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > [ 0.579192] select_idle_sibling: wakee=kthreadd:2 nr_cpus_allowed=8 current=swapper/0:1 in_task=1 > > The nr_cpus_allowed=8 suggest that none of the threads from the logs I > shared are per-CPU. Sorry if the format is confusing, I used: > > wakee=: current=:. > > > > > Can you see similar things later after booting ? > > I tried few scenarios other than the boot time but none of them produced > "current=swapper/X:1 in_task=1" > > > > > I have tried to trigger the situation but failed to get wrong > > sequence. All are coming from interrupt while idle. > > After adding in_task() condition, I haven't been able to trigger the > > warn() that I added to catch the wrong situations on SMP, Heterogenous > > or NUMA system. Could you share more details on your setup ? > > > > This is just my Hikey960 with the asym_fits_capacity() fix [1] to make sure I > don't simply hit the other issue with asym platforms. I ran my previous tests on dragonboard 845c which is dynamiQ and I have tried on my hikey960 since but without any success so far. This is what i use: --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -6397,9 +6397,12 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target) * essentially a sync wakeup. An obvious example of this * pattern is IO completions. */ - if (is_per_cpu_kthread(current) && + if (in_task() && + is_per_cpu_kthread(current) && prev == smp_processor_id() && this_rq()->nr_running <= 1) { + + WARN(is_idle_task(current), "idle per cpu kthread: cpu %d task: %s", prev, p->comm); return prev; } Without in_task() condition, i've got warnings from interrupt context but nothing else. Note that I don't even have the asym_fits_capacity() condition > > Then I just added my log in the per-CPU kthread wakee stacking exit path > > printk("%s: wakee=%s:%d nr_cpus_allowed=%d current=%s:%d in_task=%d\n", > __func__, p->comm, p->pid, p->nr_cpus_allowed, current->comm, current->pid, in_task()); > > > [1] https://lore.kernel.org/all/20211125101239.3248857-1-vincent.donnefort@arm.com/ > > > From the same logs I also see: > > wakee=xfsaild/mmcblk0:4855 nr_cpus_allowed=8 current=kworker/1:1:1070 in_task=0 > > Doesn't that look like a genuine wakeup that would escape the per-CPU kthread > stacking exit path because of the in_task test?