Date: Thu, 19 Nov 2015 11:18:03 -0800
From: Andi Kleen <ak@linux.intel.com>
To: Peter Zijlstra <peterz@infradead.org>
Cc: kernel test robot <ying.huang@linux.intel.com>, lkp@01.org,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [lkp] [x86, perf] 7aba70e47c: BUG: unable to handle kernel
 paging request at 696d2f62
Message-ID: <20151119191803.GB8802@tassilo.jf.intel.com>
References: <87fv03u89v.fsf@yhuang-dev.intel.com>
 <20151118162742.GH3816@twins.programming.kicks-ass.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20151118162742.GH3816@twins.programming.kicks-ass.net>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1809
Lines: 42

On Wed, Nov 18, 2015 at 05:27:42PM +0100, Peter Zijlstra wrote:
> On Wed, Nov 18, 2015 at 02:33:00PM +0800, kernel test robot wrote:
> > FYI, we noticed the below changes on
> > 
> > https://git.kernel.org/pub/scm/linux/kernel/git/peterz/queue.git perf/core
> > commit 7aba70e47ca4e961acb5af96d5127e3fad651c7c ("x86, perf: Optimize stack walk user accesses")
> 
> Of course, that commit no longer exists. I re-create the tree every time
> I push it, this means that if you report something a few days later, its
> highly likely its against non-existant commits :/
> 
> > [   21.984049] BUG: unable to handle kernel paging request at 696d2f62
> > [   21.986759] IP: [<4110c023>] perf_prepare_sample+0xcc/0x51d
> > [   21.987859] *pdpt = 0000000001a93001 *pde = 0000000000000000 
> > [   21.988015] Oops: 0000 [#1] PREEMPT 
> > [   21.988015] Modules linked in:
> > [   21.988015] CPU: 0 PID: 496 Comm: trinity-main Not tainted 4.3.0-01147-g7aba70e #1
> 
> That doesn't actually look like something the fingered patch touches.
> And seeing how its trinity triggering it, I suspect bisection fail.

Ok. I assume it's not caused by my patch.  Let me know if that is wrong.

I also pushed the patch before to my tree (which is 0day tested) and there
was no such report (but of course trinity is somewhat random).

BTW if you're going to test trinity for perf it may be better to use
Vince Weaver's version here

https://github.com/deater/perf_event_tests

which has more coverage for perf than normal trinity.

-Andi

-- 
ak@linux.intel.com -- Speaking for myself only
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/