On 12/12/16 03:28 PM, Michel Dänzer wrote:
> On 12/12/16 05:39 AM, Frédéric Brière wrote:
>> On Sat, Dec 10, 2016 at 10:07:51PM -0500, Frédéric Brière wrote:
>>> Otherwise, I guess I'll have to muster up the courage to try to bisect
>>> this. (gulp!)
>>
>> It wasn't exactly a nice leisurely stroll through the park, but I did
>> manage to bisect and pinpoint f993091 as the guilty commit.
>
> Thanks for that!
Actually, this is a curious result of the bisection: I bisected an
immediate freeze of basically any client using DRI2 (with sync-to-vblank
enabled) to commit 8f1edf4bd3a1 ("dix: Use list for ready clients"),
which is immediately before the commit you bisected to. See
https://bugs.freedesktop.org/show_bug.cgi?id=98030 . Did you not run
into that while bisecting?
>> At this point, I guess there's not much left to do, apart from filing a
>> bug upstream? (Oh, great, Bugzilla. <shudder>)
>
> Yeah, that's the logical next step.
FWIW, the issue is likely specific to DRI2 vs DRI3, not EXA vs glamor.
--
Earthling Michel Dänzer | http://www.amd.com
Libre software enthusiast | Mesa and X developer
