Re: [igt-dev] [PATCH i-g-t v2 3/4] lib/igt_kmod: add compatibility for KUnit

[Mauro Carvalho Chehab <mauro.chehab@linux.intel.com>]

On Thu, 3 Nov 2022 10:48:40 +0100
Mauro Carvalho Chehab <mauro.chehab@linux.intel.com> wrote:

> On Sun, 28 Aug 2022 21:09:19 -0300
> Isabella Basso <isabbasso@riseup.net> wrote:
> 
> > This adds functions for both executing the tests as well as parsing (K)TAP
> > kmsg output, as per the KTAP spec [1].
> > 
> > [1] https://www.kernel.org/doc/html/latest/dev-tools/ktap.html
> > 
> > Signed-off-by: Isabella Basso <isabbasso@riseup.net>  
> 
> Hi Isabella,
> 
> I'm doing some tests here with my i915 KUnit patch series.
> 
> There's a problem with the way it is currently parsing the KTAP logs on
> IGT. It sounds that it is parsing the data only at the end, and not as
> they arrive. That's bad, as if something goes wrong, the previously
> reported data is still useful.
> 
> Also, when there's a crash, the IGT process is getting segmentation
> fault. So, the end result is that nothing is actually reported.
> 
> 	$ sudo ./build/tests/i915_selftest
> 	IGT-Version: 1.26-gbab20eb65 (x86_64) (Linux: 6.1.0-rc2-drm-990037e9984e+ x86_64)
> 	Segmentation fault (core dumped)
> 
> However, the logs are there (see enclosed). In this specific case, there
> was a GPU hang while running one of the tests on a real hardware, which
> is something that it is actually expected at development time.

Btw, if you want to test it, the patches I'm using are at:
	https://patchwork.freedesktop.org/series/110483/

The latest patch there is a workaround that fixes the issue. So, if you
apply patches 1-7 only, skipping (or reverting) patch 8, you'll be able
to reproduce the Kernel crash (which seems to be caused by the way KUnit
is initialized).

Regards,
Mauro