Re: [regression] Re: [ 049/149] rtc: Disable the alarm in the hardware (v2)

[Heinz Wiesinger]

On Monday 10 December 2012 17:56:02 John Stultz wrote:
> On 12/08/2012 03:10 AM, Heinz Wiesinger wrote:
> > On Tuesday 04 December 2012 10:34:39 John Stultz wrote:
> >> On 12/04/2012 02:34 AM, Heinz Wiesinger wrote:
> >>> On Monday 03 December 2012 15:14:12 you wrote:
> >>>> On 11/05/2012 02:55 PM, Heinz Wiesinger wrote:
> >>>>> On Monday 05 November 2012 11:13:31 Greg KH wrote:
> >>>>>> On Mon, Nov 05, 2012 at 01:11:18AM -0800, Jonathan Nieder wrote:
> >>>>>>> Hi,
> >>>>>>> 
> >>>>>>> In March, Greg KH wrote:
> >>>>>>>> 3.2-stable review patch.  If anyone has any objections, please let
> >>>>>>>> me
> >>>>>>>> know.
> >>>>>>> 
> >>>>>>> Sorry, I'm a little late.  This seems to be causing spurious wakeups
> >>>>>>> after shutdown on some systems just like v1 did. :/
> >>>>>>> 
> >>>>>>> Janne Boman reports[1], using an HP EliteBook 8530w:
> >>>>>>> | I'm unable to shutdown my system via normal means. The system
> >>>>>>> | completes the shutdown process, all hardware lights are off, but
> >>>>>>> | then
> >>>>>>> | somehow it restarts.
> >>>>>>> 
> >>>>>>> Based on a quick web search, Heinz Wiesinger, using the same model,
> >>>>>>> 
> >>>>>>> ran into the same problem and bisected it to this patch:
> >>>>>>>      https://bugzilla.kernel.org/show_bug.cgi?id=44261
> >>>>>>> 
> >>>>>>> Janne confirmed the same by testing kernels closely based on 3.2.13
> >>>>>>> and 3.2.14 --- 3.2.13 shut down fine, while on 3.2.14 "shutdown -h
> >>>>>>> now" behaved roughly speaking like "reboot".  A kernel close to
> >>>>>>> 3.6.4
> >>>>>>> was also affected.
> >>>>>> 
> >>>>>> So is 3.6.6 still and issue?  How about 3.7-rc4?  3.4.18?  Figuring
> >>>>>> out
> >>>>>> if this is still an issue in Linus's tree would be the best thing to
> >>>>>> do
> >>>>>> first.
> >>>>> 
> >>>>> I checked 3.7-rc4 and the issue is still there. Should I still test
> >>>>> the
> >>>>> other versions as well? I can do so but it might take a bit.
> >>>>> 
> >>>>> If you need anything else, I'll be happy to help wherever I can.
> >>>> 
> >>>> Hey Heinz,
> >>>> 
> >>>>      Sorry for being slow to respond to this, I was on leave when it
> >>>> 
> >>>> arrived, and it got buried under other items.
> >>>> 
> >>>> This is peculiar.  I'm guessing something quirky with your hardware is
> >>>> causing the alarm irq to trigger shortly after we supposedly disable
> >>>> it.
> >>>> Could you send me your dmesg and kernel config?
> >>>> 
> >>>> After that I'll likely have to send you some debugging instructions
> >>>> (commenting out various lines) to try to figure out what exactly is
> >>>> triggering it.
> >>>> 
> >>>> Sorry for the trouble!
> >>> 
> >>> Hey John,
> >>> 
> >>> Don't worry, I know how it is to be busy.
> >>> 
> >>> For the dmesg/kernel config is it fine if I send you the one from
> >>> 3.7-rc4
> >>> or should I update to a more recent version?
> >> 
> >> 3.7-rc4 is fine.
> >> 
> >>> I've also now seen this behaviour on another notebook with completely
> >>> different hardware (Asus VX1). I can send you dmesg output from that one
> >>> too if you want. Kernel config should be almost identical, but it's
> >>> running 3.2.29 still.
> >> 
> >> Yea, dmesg there would be good as well.
> > 
> > Hey John,
> > 
> > Sorry for the delay. I was quite busy this week myself.
> > Please find the kernel config and dmesg output attached.
> 
> Thanks for the details!
> 
> Quick test: Does the issue persist if you boot with hpet=disable?

Sorry, took a while longer. Issue persists if I boot with hpet=disable.

Grs,
Heinz

Attachment: signature.asc
Description: This is a digitally signed message part.