[igt-dev] [PATCH i-g-t] lib: Handle chamelium failures gracefully

[Petri Latvala]

On Wed, Nov 11, 2020 at 09:56:55AM +0000, Chris Wilson wrote:
> Quoting Petri Latvala (2020-11-11 07:52:10)
> > On Mon, Nov 09, 2020 at 03:16:58PM +0000, Chris Wilson wrote:
> > > Don't cause CI to abort a run if an _unrelated_ fixture checks for a
> > > display and chamelium is dead.
> > 
> > If there's a chamelium configured and it's dead, we do want to abort
> > though.
> 
> If there's no display, we don't run display tests. The GPUs are still
> very interesting from my pov.

For CI results the display tests becoming skips is awfully painful. If
the chamelium is dead, we (CI) want to be notified and people tend to
spot new aborts easier than a "we have no display" skip that might
even be filtered by cibuglog.

It's a tradeoff and there is no solution that satisfies everyone, we
went with the one that leads us fixing the hw earlier.

>  
> > What's an unrelated fixture in this context?
> 
> The fixture is after we have completed the subtest. So we get
> 	SUCCESS; ABORT
> when we never intended to use the display at all.

Ah, the cleanup fixtures, I see.


-- 
Petri Latvala