Re: Debugging Thinkpad T430s occasional suspend failure.

From: Hugh Dickins
Date: Sat Feb 16 2013 - 21:22:08 EST

Next message: Peter Feuerer: "[PATCH] added two more trip points to acerhdf, this allows thermallayer to correctly handle the two point regulation of acerhdf. Trip point 0will be active from 0 degree to "fanoff" and is marked as passive, then trippoint 1 is valid from "fanoff" to "fanon" value and is marked as active,even if it's only really active in case temperature is going down from trippoint 2. Trip point 2 will be valid above "fanon" value and is also markedas active."
Previous message: Thadeu Lima de Souza Cascardo: "[PATCH] m68k/nommu: fix build when CPU is not coldfire"
In reply to: Hugh Dickins: "Re: Debugging Thinkpad T430s occasional suspend failure."
Next in thread: Daniel Vetter: "Re: Debugging Thinkpad T430s occasional suspend failure."
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Sat, 16 Feb 2013, Hugh Dickins wrote:
> On Sat, 16 Feb 2013, Linus Torvalds wrote:
> >
> > I think it's worth it to give them a heads-up already. So I've cc'd
> > the main suspects here..
>
> Okay, thanks.
>
> >
> > Daniel, Dave - any comments about a NULL fb in
> > intel_choose_pipe_bpp_dither() during either suspend or resume? Some
> > googling shows this:
> >
> > https://bugzilla.redhat.com/show_bug.cgi?id=895123
>
> Great, yes, I'm sure that's the same (though it says "suspend"
> and I say "resume").
>
> >
> > which sounds remarkably similar, and is also during a suspend attempt
> > (but apparently Satish got a full oops out).. Some timing race with a
> > worker entry?

Comparing Satish's backtrace and drivers/gpu/drm history, it's clear that
the oops comes from Daniel's 3.8-rc2 45e2b5f640b3 "drm/i915: force restore
on lid open", whose force_restore case now passes down crtc->base.fb. But
I wouldn't have a clue why that's usually non-NULL but occasionally NULL:
your timing race with a worker entry, perhaps.

And 45e2b5f640b3 contains a fine history of going back and forth, so I
wouldn't want to play further with it out of ignorance - though tempted
to replace the "if (force_restore) {" by an interim safe-seeming
compromise of "if (force_restore && crtc->base.fb) {".

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