[Bug 101167] [BAT][SKL] igt at kms_flip@basic-flip-vs-dpms produces a dmesg-fail

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=101167

--- Comment #23 from Martin Peres <martin.peres at free.fr> ---
(In reply to Manasi from comment #22)
> (In reply to Marta Löfstedt from comment #18)
> > Created attachment 131731 [details]
> > dmesg for Manasi suggestion Comment 17
> 
> This dmesg log shows that we still do a fallback on eDP when link training
> fails the first time but then it seems to pass at reduced link rate to
> 162000.
> 
> Now why don't we see any dp_aux_ch timeouts at this econd attempt of modeset
> with reduced link rate?

The two are unrelated. It just happens that now, the panel has had enough time
to start.

> 
> So the test passes here with just that first warning right? Isnt that whats
> expected though?

No, the test fails because the mode cannot be set again, because it has been
pruned (making sure the lowest mode never gets pruned is another bug).

So, this bug is about IGT not checking that a mode is available before setting
it. This never was a problem because IGT simply sets the same mode that was
already-set when starting the test. In other word, the link-status patch broke
this assumption and IGT does not know what to do.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20170606/4dcdd046/attachment.html>