[PATCH i-g-t] tests/intel/xe_mmap: Skip pci-membarrier-parallel test on PVC

[Matthew Auld]

On 12/05/2025 08:36, Kolanupaka Naveena wrote:
> On xe, PVC RC6WAs are not there, hence the device is going to bad/sleep
> state. We are not going bring those WAs on xe as they are PVC specific.

Can you share some more info on the issue here? What is the relation to 
RC6WAs? What is the error? Also why do we only skip membarrier-parallel 
and not the other barrier subtests?

Is this not the issue with missing an rpm ref to keep the pci device 
awake since there is no display?  Real user would also be submitting 
work to keep the device awake, but we don't do that in this test IIRC.

> 
> So, decided to skip this test for PVC platform to avoid the error during
> the read.
> 
> Signed-off-by: Kolanupaka Naveena <kolanupaka.naveena at intel.com>
> ---
>   tests/intel/xe_mmap.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/tests/intel/xe_mmap.c b/tests/intel/xe_mmap.c
> index 5fd641075..a4915b37d 100644
> --- a/tests/intel/xe_mmap.c
> +++ b/tests/intel/xe_mmap.c
> @@ -429,6 +429,9 @@ igt_main
>   		unsigned int i;
>   		uint32_t *ptr;
>   
> +		if (IS_PONTEVECCHIO(intel_get_drm_devid(fd)))
> +			igt_skip("Test is not supported on this platform.\n");
> +
>   		igt_require(is_pci_membarrier_supported(fd));
>   		xe = drm_open_driver(DRIVER_XE);
>   		srand(time(0));