freebsd stable - Jun 2015 - suspend/resume regression

Kevin Oberman writes:
> On Sat, Jun 27, 2015 at 3:13 PM, Joseph Mingrone <jrm at ftfl.ca>
wrote:
>
>> Adrian Chadd <adrian.chadd at gmail.com> writes:
>> > does it always fail now?
>>
>> Yes.  The behaviour I described is, so far with 15-20 tries,
consistent.
>>
>
> I confirm that I see the similar behavior on my T520 ThinkPad, Sandy Bridge
> system.
> FreeBSD rogue 10.2-PRERELEASE FreeBSD 10.2-PRERELEASE #1 r284698: Mon Jun
> 22 09:25:11 PDT 2015     root at rogue:/usr/obj/usr/src/sys/GENERIC  amd64
>
> It has been working flawlessly since Adrian's ACPI update. As of now,
it
> suspends fine, but it won't resume. I get a fan spin-up, but the power
LED
> continues to pulse, indicating it is suspended and the logs show nothing
> after the suspend.
Just updated my T520 yesterday, and I get the same behavior with the
same machine here. Suspend is fine; resume fails. It seems to get a bit
farther along in the resume process if i915kms isn't loaded, though I've
got no way to really measure that. I've tried a GENERIC kernel, clean
source tree and empty src.conf, to no avail.

Several MFCs in r280434 from a couple weeks ago touched a whole bunch of
suspend/resume stuff. That might be a good place to start digging.

-- 
================================================================   		      ::
Brandon Wandersee ::
                  :: brandon.wandersee at gmail.com ::
================================================================='A common
mistake that people make when trying to design something
completely foolproof is to underestimate the ingenuity of complete
fools.'
                            - Douglas Adams
==================================================================

On Sat, Jun 27, 2015 at 8:03 PM, Brandon J. Wandersee <
brandon.wandersee at gmail.com> wrote:
>
> Kevin Oberman writes:
>
> > On Sat, Jun 27, 2015 at 3:13 PM, Joseph Mingrone <jrm at
ftfl.ca> wrote:
> >
> >> Adrian Chadd <adrian.chadd at gmail.com> writes:
> >> > does it always fail now?
> >>
> >> Yes.  The behaviour I described is, so far with 15-20 tries,
consistent.
> >>
> >
> > I confirm that I see the similar behavior on my T520 ThinkPad, Sandy
> Bridge
> > system.
> > FreeBSD rogue 10.2-PRERELEASE FreeBSD 10.2-PRERELEASE #1 r284698: Mon
Jun
> > 22 09:25:11 PDT 2015     root at rogue:/usr/obj/usr/src/sys/GENERIC 
amd64
> >
> > It has been working flawlessly since Adrian's ACPI update. As of
now, it
> > suspends fine, but it won't resume. I get a fan spin-up, but the
power
> LED
> > continues to pulse, indicating it is suspended and the logs show
nothing
> > after the suspend.
>
> Just updated my T520 yesterday, and I get the same behavior with the
> same machine here. Suspend is fine; resume fails. It seems to get a bit
> farther along in the resume process if i915kms isn't loaded, though
I've
> got no way to really measure that. I've tried a GENERIC kernel, clean
> source tree and empty src.conf, to no avail.
>
> Several MFCs in r280434 from a couple weeks ago touched a whole bunch of
> suspend/resume stuff. That might be a good place to start digging.
>
>
I have narrowed it to between 283990 and 284074. Have to crash, so maybe
someone else will track it down by tomorrow. I only see a handfull (<10)
commits in there that look like possible causes. 280434 is compiling right
now as the prime candidate.

Kevin Oberman, Network Engineer, Retired
E-mail: rkoberman at gmail.com
PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683