From: Mark Jackson Subject: Re: v3.9.0-rc8 : oops in tick_do_update_jiffies64+0xbc/0x110 Date: Mon, 29 Apr 2013 16:46:18 +0100 Message-ID: <517E95CA.4000708@newflow.co.uk> References: <517E3039.2070204@newflow.co.uk> <20130429154146.GB28721@atomide.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20130429154146.GB28721@atomide.com> Sender: linux-kernel-owner@vger.kernel.org To: Tony Lindgren Cc: "linux-omap@vger.kernel.org" , tglx@linutronix.de, lkml List-Id: linux-omap@vger.kernel.org On 29/04/13 16:41, Tony Lindgren wrote: > * Mark Jackson [130429 01:38]: >> I've been experiencing several crashes all pointing exactly the same place in the same tick routine (see below). >> >> The "exception stack" trace at the end changes depending on when the oops occurs. >> >> I've had the oops occur maybe 6 times in the last 50 reboots. >> >> Any ideas ? > > Sounds like it might be an issue with the physical memory or > the timings. It could also be related to the idle loop not > restoring something right for deeper idle states that corrupts > the memory. And that's why it would seem to appear while waking > to a timer. I am suspecting a borderline CPU. We only have "X" AM3359 parts which are all pre-production. > Maybe disable PM and run memtester to see if that runs reliably? I'll see if I can get it running on our platform. Thanks Mark J.