Re: [PATCH] tests/intel/i915_pm_rc6_residency: make residency ratio tolerance symmetric

["Poosa, Karthik" <karthik.poosa@intel.com>]

On 29-09-2026 22:58, Sk Anirban wrote:
> Bump the upper ratio bound from 1.05 to 1.1 so residency_accuracy()
> allows the same ±10% margin on both sides of the expected duration.
>
> Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/17049
> Signed-off-by: Sk Anirban <sk.anirban@intel.com>
> ---
>   tests/intel/i915_pm_rc6_residency.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tests/intel/i915_pm_rc6_residency.c b/tests/intel/i915_pm_rc6_residency.c
> index 346796fb2..80c102e31 100644
> --- a/tests/intel/i915_pm_rc6_residency.c
> +++ b/tests/intel/i915_pm_rc6_residency.c
> @@ -112,7 +112,7 @@ static void residency_accuracy(unsigned int diff,
>   
>   	igt_info("Residency in %s or deeper state: %u ms (sleep duration %u ms) (%.1f%% of expected duration)\n",
>   		 name_of_rc6_residency, diff, duration, 100*ratio);
> -	igt_assert_f(ratio > 0.9 && ratio < 1.05,
> +	igt_assert_f(ratio > 0.9 && ratio < 1.1,

Looks fine,

but what is the reason upper bound was different previously ?

>   		     "Sysfs RC6 residency counter is inaccurate.\n");
>   }
>