[igt-dev] [PATCH i-g-t] lib/igt_gt: Move assert with setting 'reset' parameter after IGT env check

[Daniel Vetter]

On Wed, Jan 23, 2019 at 3:12 PM Chris Wilson <chris at chris-wilson.co.uk> wrote:
>
> Quoting Katarzyna Dec (2019-01-23 14:04:04)
> > We are trying to set reset parameter earlier than checking IGT env.
> > Let's move it after checking IGT env to avoid setting reset parameter
> > when hang injection is disabled.
> >
> > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > Signed-off-by: Katarzyna Dec <katarzyna.dec at intel.com>
>
> I can see the sense in that. Can we kill the env var though?
> Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>
>
> If the system is not suitable for reset tests, can we not derive that
> from querying the system rather than rely on the user? We have better
> means of test selection now than just setting env var (and if not, why
> not) so afaik CI is not using it.

Yeah I thought the plan for these was to add some kernel patch (in
drm-intel-internal only ofc) which tells igt that reset isn't a thing.
We've definitely had this years ago, when there was no reset on gen2.
But maybe the wheels fell off on that one due to lack of use, dunno.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch