Re: [PATCH] tests/intel/xe_oa: Add test to ensure entire report has landed

["Dixit, Ashutosh" <ashutosh.dixit@intel.com>]

On Tue, 25 Aug 2026 21:42:10 -0700, Umesh Nerlige Ramappa wrote:
>

Hi Umesh,

> When running the load helper with a fast oa_exponent, sometimes the OA
> reports land out of order. Ensure that the kernel fix for such issues is
> working by configuring all OA counters and checking validity of the full
> report.
>
> Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
> ---
> v2:
> - add a comment on event definitions
> - rename report checker for clarity
> - remove unnecessary spaces
> - u32s in one line
> - s/576/report_size
> ---
>  tests/intel/xe_oa.c | 217 ++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 217 insertions(+)
>
> diff --git a/tests/intel/xe_oa.c b/tests/intel/xe_oa.c
> index 9ebb18690ec7..58e50bb182a6 100644
> --- a/tests/intel/xe_oa.c
> +++ b/tests/intel/xe_oa.c
> @@ -336,6 +336,88 @@ static struct oa_format lnl_oa_formats[XE_OA_FORMAT_MAX] = {
>		.bc_report = 0 },
>  };
>
> +/*
> + * test_oa_full defines only 2 events:
> + *
> + * 3e00 is "Number of GPU (gt) clock cycles at every clock period"
> + * 7e00 is "Number of GPU (gt) clock cycles at every other clock period"
> + *

Can you please also add here in the comment that we expect all counters in
a report to be non-zero with this metric set (or reg/val pairs).

Anyway, this test is failing in pre-merge CI, so let's see what we end up
with once we fix that issue. I'll hold the R-b till we have a working test.

Thanks.
--
Ashutosh