Re: [igt-dev] [PATCH i-g-t 2/5] tests/intel/kms_pm_lpsp: Add XE support

["Modem, Bhanuprakash" <bhanuprakash.modem@intel.com>]

On Mon-11-09-2023 08:38 pm, Sharma, Swati2 wrote:
> On 11-Sep-23 8:14 PM, Modem, Bhanuprakash wrote:
>> Hi Swati,
>>
>> On Fri-08-09-2023 11:20 pm, Swati Sharma wrote:
>>> Add XE driver support for kms pm tests.
>>>
>>> Signed-off-by: Swati Sharma <swati2.sharma@intel.com>
>>> ---
>>>   tests/intel/kms_pm_lpsp.c | 6 +++---
>>>   1 file changed, 3 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/tests/intel/kms_pm_lpsp.c b/tests/intel/kms_pm_lpsp.c
>>> index 9b5547cda..7d697b1e2 100644
>>> --- a/tests/intel/kms_pm_lpsp.c
>>> +++ b/tests/intel/kms_pm_lpsp.c
>>> @@ -39,7 +39,7 @@
>>>    *
>>>    * SUBTEST: kms-lpsp
>>>    * Description: This test validates lpsp on all connected outputs 
>>> on low power PIPE_A
>>> - * Driver requirement: i915
>>> + * Driver requirement: i915, xe
>>>    * Functionality: pm_lpsp
>>>    * Mega feature: Display Power
>>>    * Run type: FULL
>>> @@ -194,8 +194,7 @@ igt_main
>>>       data_t data = {};
>>>       igt_fixture {
>>> -
>>> -        data.drm_fd = drm_open_driver_master(DRIVER_INTEL);
>>> +        data.drm_fd = drm_open_driver_master(DRIVER_INTEL | DRIVER_XE);
>>>           igt_require(data.drm_fd >= 0);
>>>           data.debugfs_fd = igt_debugfs_dir(data.drm_fd);
>>>           igt_require(data.debugfs_fd >= 0);
>>> @@ -208,6 +207,7 @@ igt_main
>>>       igt_describe("This test validates lpsp while all crtc are 
>>> disabled");
>>>       igt_subtest("screens-disabled") {
>>> +        igt_require_i915(data.drm_fd);
>>
>> Why it is i915 specific?
> 
> This test has no ROI if platform supports DC states.
> Test requirement not met in function __igt_unique____real_main192, file 
> ../../../usr/src/igt-gpu-tools/tests/intel/kms_pm_lpsp.c:211:
> Test requirement: !dmc_supported(data.debugfs_fd)
> DC states supported platform don't have ROI for this subtest
> Subtest screens-disabled: SKIP (0.000s)
> 
> xe+ we have all platforms supporting dc states, so we can skip this test 
> for xe.

Thanks for the clarification.

Reviewed-by: Bhanuprakash Modem <bhanuprakash.modem@intel.com>

> 
>>
>> - Bhanu
>>
>>>           igt_require_f(!dmc_supported(data.debugfs_fd),
>>>                     "DC states supported platform don't have ROI for 
>>> this subtest\n");
>>>           screens_disabled_subtest(&data);