[igt-dev] [PATCH igt] tests/kms_frontbuffer_tracking: increase FBC wait timeout to 5s

[Paulo Zanoni]

Em Qua, 2018-03-21 às 01:06 +0000, Chris Wilson escreveu:
> Quoting Paulo Zanoni (2018-03-21 00:29:45)
> > From: "Lofstedt, Marta" <marta.lofstedt at intel.com>
> > 
> > This is a partial revert of:
> > 64590c7b768dc8d8dd962f812d5ff5a39e7e8b54
> > ("kms_frontbuffer_tracking:
> > reduce the FBC wait timeout to 2s")
> > 
> > The subtests: igt at kms_frontbuffer_tracking@fbc-*draw* have
> > non-consistent results, pending between fail and pass. The fails
> > are
> > always due to "FBC disabled".  With this increase in timeout the
> > flip-flop behavior is no longer reproducible.
> 
> We haven't seen this behaviour for a few months, afaict. Currently we
> have FBC perma-disabled due to bad mode, underruns and whatnot, and
> flip flops from CRC mismatches. But no FBC timeouts waiting for 
> enable.

Then why is it being reported as highest-priority-needs-to-be-fixed-
immediately to me? The goal of the patch would be to try to reduce the
flip-flopping caused by the low timeout so we can better split this
type of problem from the others that may be happening.

> -Chris