[igt-dev] [PATCH i-g-t] tests/kms_plane: add source clamping test

Juha-Pekka Heikkila juhapekka.heikkila at gmail.com
Wed Nov 21 09:45:31 UTC 2018


On 20.11.2018 13:30, Maarten Lankhorst wrote:
> Op 15-11-18 om 16:51 schreef Juha-Pekka Heikkila:
>> Add test which create fb bigger than plane and clamp fb from all sides.
>>
>> Signed-off-by: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
>> ---
>>   tests/kms_plane.c | 28 ++++++++++++++++++++++++++--
>>   1 file changed, 26 insertions(+), 2 deletions(-)
>>
>> diff --git a/tests/kms_plane.c b/tests/kms_plane.c
>> index 45e0a30..d79c22e 100644
>> --- a/tests/kms_plane.c
>> +++ b/tests/kms_plane.c
>> @@ -41,6 +41,7 @@ typedef struct {
>>   	int drm_fd;
>>   	igt_display_t display;
>>   	igt_pipe_crc_t *pipe_crc;
>> +	uint32_t crop;
>>   } data_t;
>>   
>>   static color_t red   = { 1.0f, 0.0f, 0.0f };
>> @@ -412,12 +413,22 @@ static void test_format_plane_color(data_t *data, enum pipe pipe,
>>   	const color_t *c = &colors[color];
>>   	struct igt_fb old_fb = *fb;
>>   
>> -	igt_create_color_fb(data->drm_fd, width, height,
>> -			    format, LOCAL_DRM_FORMAT_MOD_NONE,
>> +	igt_create_color_fb(data->drm_fd, width+data->crop*2,
>> +			    height+data->crop*2, format,
>> +			    LOCAL_DRM_FORMAT_MOD_NONE,
>>   			    c->red, c->green, c->blue, fb);
>>   
>>   	igt_plane_set_fb(plane, fb);
>>   
>> +	/*
>> +	 * if clamping test
>> +	 */
>> +	if (width-data->crop != 0) {
> if (data->crop) ?

Oops, I have been sleeping here.

>> +		igt_plane_set_size(plane, width, height);
>> +		igt_fb_set_position(fb, plane, data->crop, data->crop);
>> +		igt_fb_set_size(fb, plane, width, height);
>> +	}
>> +
> 
> If you don't fill the color fb outside the visible boundaries with a different color, I don't think you're testing anything..
> 

That's a good idea. Originally I just thought if source clamping would 
ever fail it would produce some garbage. This is how IFT version of this 
test went but I don't think it ever got any reasonable results anyway.

/Juha-Pekka


More information about the igt-dev mailing list