[Mesa-ci-intel] Regressions on staging/19.1

[Juan A. Suarez Romero]

Hi.

I'm one week behind the scheduled release for 19.1.8, the last of the 19.1
releases, mainly because I'm getting several regressions that blocked.


Intel CI is showing regressions in two builds (supposedly I'm interpreting
correctly the results from the web):

* staging/19.1 (
https://mesa-ci.01.org/mesa_19_1_staging/builds/63/group/63a9f0ea7bb98050796b649e85481845
)


Main problems are in Iris and ICL. I'm ignoring those problems because those are
not stable in 19.1, and they shouldn't be tested in 19.1.

But there is one regression, though, affecting BDW in m64: 
https://mesa-ci.01.org/mesa_19_1_staging/builds/63/results/85860979 I've tried
to reproduce it locally, but everything works fine here. Could anyone verify
this is not a false regression, or what is the offending commit to remove it
from the queue?



* staging/19.1 daily (
https://mesa-ci.01.org/mesa_19_1_staging_daily/builds/87/group/63a9f0ea7bb98050796b649e85481845
)

This happens on a daily basis, but for some reason it wasn't triggered
yesterday.

In any case, the available results shows regressions in ICL and Iris, which I'm
ignoring as above, but also others in BSW (m32 and m64), gen9 (m32), gen9atom
(m32 and m64), and g965. I've tried some of them locally, but again, they work
fine locally.


Again, could anyone check what is the offending commit? In this case, it is even
harder to find because the daily happens... once per day, so there are no many
chances to find the problem.


As I said at the beginning, I'm quite delayed with this (last) release, and I
would like to remove the commits causing the regressions and make the release.

The other choice? Just ignore all of them and go with the release as it is. 19.2
is out, and afaik it doesn't have those regressions.



	J.A.