Re: [PATCH i-g-t] tests/device_reset: Cold reset doesn't depend on PCI reset

Hi Janusz,

On 2026-08-19 at 16:41:06 +0200, Janusz Krzysztofik wrote:
> Before commit 59e3bf83f6bae ("tests/device_reset: Add cold reset IGT
> test") all subtests required device PCI reset capability, with that
> dependency placed in an initial igt_fixture section.  New subtests
> introduced with that commit didn't depend on that capability but that
> requirement was left untouched.  That was OK as long as all devices
> being tested supported a PCI reset method, however, that's going to
> change with an upcoming kernel patch that addresses DG1 SKU 0x4905
> inability to perform a PCI reset without breaking the device.
> 
> Move the requirement from the initial igt_fixture section to the
> relevant subtest bodies, so cold reset subtest can be run even if PCI
> reset is not supported.  See commit 4a5fd4e7cb279 ("tests/device_reset:
> move later skips into subtests") for justification on why not to use
> an igt_fixture inside an igt_subtest_group for that purpose.
> 
> Signed-off-by: Janusz Krzysztofik <[email protected]>
> ---

Reviewed-by: Krzysztof Karas <[email protected]>

-- 
Best Regards,
Krzysztof