Re: [PATCH] tests/intel/xe_pm_residency: Fix suspend duration calculation

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Mon, Mar 23, 2026 at 08:06:44PM +0530, Anirban, Sk wrote:
> Hi,
> 
> On 23-03-2026 03:31 pm, Ville Syrjälä wrote:
> > On Mon, Mar 23, 2026 at 12:32:29PM +0530, Sk Anirban wrote:
> >> Fix the unit mismatch caused by interpreting seconds as milliseconds
> >> during suspend duration calculations. Apply MSEC_PER_SEC multiplier
> >> to ensure correct unit handling.
> >>
> >> Signed-off-by: Sk Anirban <sk.anirban@intel.com>
> >> ---
> >>   tests/intel/xe_pm_residency.c | 2 +-
> >>   1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/tests/intel/xe_pm_residency.c b/tests/intel/xe_pm_residency.c
> >> index 61b63ed4f..46718a444 100644
> >> --- a/tests/intel/xe_pm_residency.c
> >> +++ b/tests/intel/xe_pm_residency.c
> >> @@ -223,7 +223,7 @@ static void test_idle_residency(int fd, int gt, enum test_type flag)
> >>   		 * elapsed time during suspend is approximately equal to autoresume delay
> >>   		 * when a full suspend cycle(SUSPEND_TEST_NONE) is used.
> >>   		 */
> >> -		elapsed_ms = igt_get_autoresume_delay(SUSPEND_STATE_FREEZE);
> >> +		elapsed_ms = igt_get_autoresume_delay(SUSPEND_STATE_FREEZE) * MSEC_PER_SEC;
> > Why doesn't the test actually measure this?
> 
> Since rtcwake is being used to enter s2idle, both paths end up using a 
> 15‑second delay as the command parameter
> 
> and this value is carried through all related calculations. Are you 
> suggesting switching to a monotonic clock–based approach during resume?

CLOCK_BOOTTIME

-- 
Ville Syrjälä
Intel