From patchwork Wed Oct 17 23:06:54 2012 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Kevin Hilman X-Patchwork-Id: 1607871 Return-Path: X-Original-To: patchwork-linux-omap@patchwork.kernel.org Delivered-To: patchwork-process-083081@patchwork1.kernel.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by patchwork1.kernel.org (Postfix) with ESMTP id 7246640135 for ; Wed, 17 Oct 2012 23:06:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752378Ab2JQXGn (ORCPT ); Wed, 17 Oct 2012 19:06:43 -0400 Received: from mail-pa0-f46.google.com ([209.85.220.46]:62279 "EHLO mail-pa0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752362Ab2JQXGl (ORCPT ); Wed, 17 Oct 2012 19:06:41 -0400 Received: by mail-pa0-f46.google.com with SMTP id hz1so7641666pad.19 for ; Wed, 17 Oct 2012 16:06:41 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=from:to:cc:subject:organization:references:date:in-reply-to :message-id:user-agent:mime-version:content-type:x-gm-message-state; bh=J7+Cly2OtdTT4NyuJINQbqm1ioHVB3AgqfjWQjkk6Ug=; b=jh1hvniMAokJ0jpw39j8qKzYLcePFsqKzcz9wJE7z4jQjjmevz9jXDiKCHsIkhQb98 pjdFOGaYSqVWzOHbY50OO439n3d+yqfgqY/WFln935Yo1skHNlT6Grl3OSi87UtvkV5t 9Cob0RQxK1dhWhLMQYNV5NKaRCPbsu2ptQw/SAiGHIpxagBM23NMf14JtYnemEzzKy0I t2wHiE3zfsZSFGKsP59kzuyXpGQvelUgCA92rPeYFN6uWQUfrTClxb7gMUaJFE7V/c1N TzKifQq+kfLZCXHz2l8zcrBfh7pITdFrEkv1ZygxPwfihC2g/5+Qe2YWNBLJwV84Wfxk fAoA== Received: by 10.66.77.201 with SMTP id u9mr54429811paw.6.1350515201021; Wed, 17 Oct 2012 16:06:41 -0700 (PDT) Received: from localhost (c-24-19-7-36.hsd1.wa.comcast.net. [24.19.7.36]) by mx.google.com with ESMTPS id k4sm13275227pax.7.2012.10.17.16.06.39 (version=TLSv1/SSLv3 cipher=OTHER); Wed, 17 Oct 2012 16:06:39 -0700 (PDT) From: Kevin Hilman To: Cc: Thomas Gleixner , Peter Zijlstra , Shubhrajyoti Datta , Paul Walmsley , , Shubhrajyoti D , Wolfram Sang , Ben Dooks , , Subject: Re: RT throttling and suspend/resume (was Re: [PATCH] i2c: omap: revert "i2c: omap: switch to threaded IRQ support") Organization: Deep Root Systems, LLC References: <20121016133356.GG21801@arwen.pp.htv.fi> <87ipaanljt.fsf_-_@deeprootsystems.com> <20121017140002.GI11394@arwen.pp.htv.fi> <20121017143534.GJ11394@arwen.pp.htv.fi> Date: Wed, 17 Oct 2012 16:06:54 -0700 In-Reply-To: <20121017143534.GJ11394@arwen.pp.htv.fi> (Felipe Balbi's message of "Wed, 17 Oct 2012 17:35:34 +0300") Message-ID: <87txtsitpt.fsf@deeprootsystems.com> User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/23.3 (gnu/linux) MIME-Version: 1.0 X-Gm-Message-State: ALoCoQloXqIUxbU/rwG+RUF/s+D0NXmfEdB+0Uh9RfjeA/IL23scO9ai5AJi25xSs+l6vAaZlC2K Sender: linux-omap-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-omap@vger.kernel.org Felipe Balbi writes: > On Wed, Oct 17, 2012 at 05:00:02PM +0300, Felipe Balbi wrote: >> >> On Tue, Oct 16, 2012 at 02:39:50PM -0700, Kevin Hilman wrote: >> > + peterz, tglx >> > >> > Felipe Balbi writes: >> > >> > [...] >> > >> > > The problem I see is that even though we properly return IRQ_WAKE_THREAD >> > > and wake_up_process() manages to wakeup the IRQ thread (it returns 1), >> > > the thread is never scheduled. To make things even worse, ouw irq thread >> > > runs once, but doesn't run on a consecutive call. Here's some (rather >> > > nasty) debug prints showing the problem: >> > >> > [...] >> > >> > >> [ 88.721923] try_to_wake_up 1411 >> > >> [ 88.725189] ===> irq_wake_thread 139: IRQ 72 wake_up_process 0 >> > >> [ 88.731292] [sched_delayed] sched: RT throttling activated >> > >> > This throttling message is the key one. >> > >> > With RT throttling activated, the IRQ thread will not be run (it >> > eventually will be allowed much later on, but by then, the I2C xfers >> > have timed out.) >> > >> > As a quick hack, the throttling can be disabled by seeting the >> > sched_rt_runtime to RUNTIME_INF: >> > >> > # sysctl -w kernel.sched_rt_runtime_us=-1 >> > >> > and a quick test shows that things go back to working as expected. But >> > we still need to figure out why the throttling is hapenning... >> > >> > So I started digging into why the RT runtime was so high, and noticed >> > that time spent in suspend was being counted as RT runtime! >> > >> > So spending time in suspend anywhere near sched_rt_runtime (0.95s) will >> > cause the RT throttling to always be triggered, and thus prevent IRQ >> > threads from running in the resume path. Ouch. >> > >> > I think I'm already in over my head in the RT runtime stuff, but >> > counting the time spent in suspend as RT runtime smells like a bug to >> > me. no? >> > >> > Peter? Thomas? >> >> it looks like removing console output completely (echo 0 > >> /proc/sysrq-trigger) I don't see the issue anymore. Let me just run for >> a few more iterations to make sure what I'm saying is correct. > > Yeah, really looks like removing console output makes the problem go > away. Ran a few iterations and it always worked fine. Full logs attached Removing console output during resume is going to significantly change the timing of what is happening during suspend/resume, so I suspect that combined with all your other debug prints is somehow masking the problem. How log are you letting the system stay in suspend? That being said, I can still easily reproduce the problem, even with console output disabled. With vanilla v3.7-rc1 + the debug patch below[1], with and without console output, I see RT throttling kicking in on resume, and the RT runtime on resume corresponds to the time spent in suspend. Here's an example of debug output of my patch below after ~3 sec in suspend: [ 43.198028] sched_rt_runtime_exceeded: rt_time 2671752930 > runtime 950000000 [ 43.198028] update_curr_rt: RT runtime exceeded: irq/72-omap_i2c [ 43.198059] update_curr_rt: RT runtime exceeded: irq/72-omap_i2c [ 43.203704] [sched_delayed] sched: RT throttling activated I see this rather consistently, and the rt_time value is always roughly the time I spent in suspend. So the primary question remains: is RT runtime supposed to include the time spent suspended? I suspect not. Kevin [1] --- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c index 418feb0..39de750 100644 --- a/kernel/sched/rt.c +++ b/kernel/sched/rt.c @@ -891,6 +891,8 @@ static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq) if (!once) { once = true; printk_sched("sched: RT throttling activated\n"); + pr_warn("%s: rt_time %llu > runtime %llu\n", + __func__, rt_rq->rt_time, runtime); } } else { /* @@ -948,8 +950,11 @@ static void update_curr_rt(struct rq *rq) if (sched_rt_runtime(rt_rq) != RUNTIME_INF) { raw_spin_lock(&rt_rq->rt_runtime_lock); rt_rq->rt_time += delta_exec; - if (sched_rt_runtime_exceeded(rt_rq)) + if (sched_rt_runtime_exceeded(rt_rq)) { + pr_warn("%s: RT runtime exceeded: %s\n", + __func__, curr->comm); resched_task(curr); + } raw_spin_unlock(&rt_rq->rt_runtime_lock); } }