Content-Type: multipart/mixed; boundary="===============3162489848785360295==" MIME-Version: 1.0 From: Peter Zijlstra To: lkp@lists.01.org Subject: Re: [sched] WARNING: CPU: 0 PID: 0 at kernel/time/hrtimer.c:802 hrtimer_forward() Date: Fri, 01 May 2015 12:10:07 +0200 Message-ID: <20150501101007.GB5029@twins.programming.kicks-ass.net> In-Reply-To: <20150501044713.GA4365@wfg-t540p.sh.intel.com> List-Id: --===============3162489848785360295== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable On Fri, May 01, 2015 at 12:47:13PM +0800, Fengguang Wu wrote: > commit 77a4d1a1b9a122ca1fa3507bd30aec1520d7a8a4 > sched: Cleanup bandwidth timers > [ 64.000438] WARNING: CPU: 0 PID: 0 at kernel/time/hrtimer.c:802 hrtime= r_forward+0x71/0x189() Argh, yes of course. The problem is that the timer can be re-started before the handler acquires the lock, and therefore the hrtimer_forwards in the handler will trigger that WARN. Lemme try and go convince this heap of snot that's occupying my brain cavity to produce coherent thought and come up with a nice fix for that. --===============3162489848785360295==--