[Intel-gfx] [igt-dev] [PATCH i-g-t] runner: Don't kill a test on taint if watching timeouts

[Janusz Krzysztofik]

On Fri, 2020-12-04 at 12:06 +0000, Chris Wilson wrote:
> Quoting Janusz Krzysztofik (2020-12-04 11:51:38)
> > We may still be interested in results of a test even if it has tainted
> > the kernel.  On the other hand, we need to kill the test on taint if no
> > other means of killing it on a jam is active.
> 
> Just so long as we are confident that the original error is preserved,
> and not lost under a deluge of what may come next. E.g. is there
> sufficient syncs after detecting the taint to record the kernel logs
> asap?
> 
> Although that the kernel taints is a strong indicator the test results
> are unreliable (or even irrelevant), imo.

Unless tainted with kernel warnings which result from bugs that can be
fixed, I believe.

Thanks,
Janusz

> -Chris