[Bug 49571] [drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung

[bugzilla-daemon at bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=49571





--- Comment #23 from Daniel Vetter <daniel at ffwll.ch>  2012-12-08 17:20:58 ---
On Sat, Dec 8, 2012 at 5:49 PM,  <bugzilla-daemon at bugzilla.kernel.org> wrote:
> --- Comment #22 from Brad  Jackson <bjackson0971 at gmail.com>  2012-12-08 16:49:48 ---
> Created an attachment (id=88651)
 --> (https://bugzilla.kernel.org/attachment.cgi?id=88651)
>  --> (https://bugzilla.kernel.org/attachment.cgi?id=88651)
> Full bisect log
>
> One thing confuses me. I tried 3.7-rc1 and it had the crash, so my bisect was
> between 3.6.0 and 3.1-rc1. But part way through the bisect, I had an unbootable
> kernel that couldn't mount the partition, so I had to do a skip. After that, it
> started including rc3 and then rc2. So it's possible that my bisect is not
> right, or that there are two patches that cause the problem.

That's completely expected - the kernel version and also the one git
shows you just goes backwards in history until it finds the first tag
and uses that. Which means if the drm-intel-next tree for 3.7 is e.g.
based of 3.6-rc2, and your bisect is only in there (since all the
other branches merged into 3.7-rc1 seem to work) git lets you test
just commits on that branch.

It looks like git bisect went back in time before the 3.6 release, but
in reality the commits your testing with that 3.6-rc2 kernel have
never been in 3.6, so really need to be tested. You're not the first
to get totally confused by this though ;-)

-- 
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.