[Intel-gfx] [igt-dev] [PATCH i-g-t] [RFC] tests/kms_big_fb: Wait for vblank before collecting CRC

Tue Jun 8 11:50:45 UTC 2021

On 8.6.2021 12.19, Srinivas, Vidya wrote:
> Hello Juha-Pekka
> 
> Instead of wait for vblank, this also works
> igt_pipe_crc_start-> igt_pipe_crc_get_current for small fb after commit -> then igt_pipe_crc_get_current For big fb -> compare -> igt_pipe_crc_stop
> 
> Would this change be okay? Kindly suggest. igt_pipe_crc_collect_crc is not working. It gives CRC mismatch for few subtests like subtest y-tiled-32bpp-rotate-0

This change is ok. It kind of implies there maybe is some problem on 
your platform with starting of crc calculation but if this is only place 
where it will show I'm ok with that since crc will not affect normal 
users in any way.

I noticed your new patch, lets see how all ci machines behave on that 
before doing anything else.

/Juha-Pekka

> 
> Have submitted the change here https://patchwork.freedesktop.org/patch/437657/?series=90389&rev=6
> 
> Thank you so much.
> 
> Regards
> Vidya
> 
> -----Original Message-----
> From: Srinivas, Vidya
> Sent: Tuesday, June 8, 2021 1:19 PM
> To: juhapekka.heikkila at gmail.com; Modem, Bhanuprakash <Bhanuprakash.Modem at intel.com>; intel-gfx at lists.freedesktop.org; igt-dev at lists.freedesktop.org
> Cc: Lin, Charlton <Charlton.Lin at intel.com>
> Subject: RE: [igt-dev] [Intel-gfx] [PATCH i-g-t] [RFC] tests/kms_big_fb: Wait for vblank before collecting CRC
> 
> Hello Juha-Pekka and Bhanu
> 
> Thank you for the review comments. Apologies Juha-Pekka, I will incorporate your review comments and try out.
> 
> Regards
> Vidya
> 
> 
> -----Original Message-----
> From: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
> Sent: Tuesday, June 8, 2021 1:04 PM
> To: Modem, Bhanuprakash <bhanuprakash.modem at intel.com>; Srinivas, Vidya <vidya.srinivas at intel.com>; intel-gfx at lists.freedesktop.org; igt-dev at lists.freedesktop.org
> Cc: Lin, Charlton <charlton.lin at intel.com>
> Subject: Re: [igt-dev] [Intel-gfx] [PATCH i-g-t] [RFC] tests/kms_big_fb: Wait for vblank before collecting CRC
> 
> On 8.6.2021 10.01, Modem, Bhanuprakash wrote:
>>> From: Intel-gfx <intel-gfx-bounces at lists.freedesktop.org> On Behalf
>>> Of Vidya Srinivas
>>> Sent: Friday, May 28, 2021 9:57 AM
>>> To: intel-gfx at lists.freedesktop.org; igt-dev at lists.freedesktop.org
>>> Cc: markyacoub at chromium.org; Lin, Charlton <charlton.lin at intel.com>
>>> Subject: [Intel-gfx] [PATCH i-g-t] [RFC] tests/kms_big_fb: Wait for
>>> vblank before collecting CRC
>>>
>>> Without wait for vblank, CRC mismatch is seen between big and small
>>> CRC on few Gen11 systems.
>>>
>>> Signed-off-by: Vidya Srinivas <vidya.srinivas at intel.com>
>>> ---
>>>    tests/kms_big_fb.c | 6 ++++--
>>>    1 file changed, 4 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/tests/kms_big_fb.c b/tests/kms_big_fb.c index
>>> b35727a09bd0..f90363c3beb2 100644
>>> --- a/tests/kms_big_fb.c
>>> +++ b/tests/kms_big_fb.c
>>> @@ -254,6 +254,7 @@ static void unset_lut(data_t *data)
>>>    static bool test_plane(data_t *data)
>>>    {
>>>    	igt_plane_t *plane = data->plane;
>>> +	igt_display_t *display = &data->display;
>>
>> For code readability purpose, I think we need to update to use this
>> variable wherever we are using "&data->display" in this function.
>> s/"&data->display"/"display"/
>>
>> With above change, this patch LGTM
>> Reviewed-by: Bhanuprakash Modem <bhanuprakash.modem at intel.com>
>>
> 
> I still don't see benefit in this patch. For now all this look like is doing is slow down the test and if this actually helps there's a real bug somewhere which is not here. My earlier review comments were not addressed hence repeat here, see below.
> 
> 
>>>    	struct igt_fb *small_fb = &data->small_fb;
>>>    	struct igt_fb *big_fb = &data->big_fb;
>>>    	int w = data->big_fb_width - small_fb->width; @@ -337,16 +338,17
>>> @@ static bool test_plane(data_t *data)
>>>    		igt_display_commit2(&data->display, data->display.is_atomic ?
>>>    				    COMMIT_ATOMIC : COMMIT_UNIVERSAL);
>>>
>>> -
>>> +		igt_wait_for_vblank(data->drm_fd,
>>> +display->pipes[data->pipe].crtc_offset);
> 
> Above this line there's flip to different fb. Below this line crc calculation is restarted, get one crc and stop crc. There's several vblanks already spent here, if now adding one more somehow helps it sound like there's problems in crc calculation on some platform or kernel is saying too early framebuffer is ready to be shown. Am I missing something here?
> 
> /Juha-Pekka
> 
>>>    		igt_pipe_crc_collect_crc(data->pipe_crc, &small_crc);
>>>
>>>    		igt_plane_set_fb(plane, big_fb);
>>>    		igt_fb_set_position(big_fb, plane, x, y);
>>>    		igt_fb_set_size(big_fb, plane, small_fb->width,
>>> small_fb->height);
>>> +
> 
> spurious empty line need to be removed.
> 
>>>    		igt_plane_set_size(plane, data->width, data->height);
>>>    		igt_display_commit2(&data->display, data->display.is_atomic ?
>>>    				    COMMIT_ATOMIC : COMMIT_UNIVERSAL);
>>> -
>>> +		igt_wait_for_vblank(data->drm_fd,
>>> +display->pipes[data->pipe].crtc_offset);
>>>    		igt_pipe_crc_collect_crc(data->pipe_crc, &big_crc);
>>>
>>>    		igt_plane_set_fb(plane, NULL);
>>> --
>>> 2.7.4
>>>
>>> _______________________________________________
>>> Intel-gfx mailing list
>>> Intel-gfx at lists.freedesktop.org
>>> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
>> _______________________________________________
>> igt-dev mailing list
>> igt-dev at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/igt-dev
>>
> 