[igt-dev] [v4 4/9] tests/kms_plane_scaling: Downscaling each plane
Jessica Zhang
quic_jesszhan at quicinc.com
Tue Feb 22 19:36:06 UTC 2022
Hi Swati,
Sorry for the late response -- had some access issues with my email
server so didn't get your reply until now.
I've looked over the rest of your patches and most of them are good to
be merged. The only patch I have a doubt about is "[v5,2/9]
tests/kms_plane_scaling: Add output name in dynamic subtests" as it
changes the output string of test_scaler_with_rotation_pipe.
This will conflict with our patch here [1] as we also change that output
string to specify the size of the buffer being passed in.
Apart from that, the rest of the patches are good to merge.
Thanks,
Jessica Zhang
[1] https://patchwork.freedesktop.org/patch/475216/?series=99364&rev=5
On 2/17/2022 10:37 AM, Sharma, Swati2 wrote:
> Okay. Rest of the patches can be merged?
>
> On 17-Feb-22 11:52 PM, Jessica Zhang wrote:
>>
>>
>> On 2/17/2022 4:15 AM, Sharma, Swati2 wrote:
>>> Hi Jessica/Petri,
>>>
>>> Sorry missed your mail.
>>>
>>> Do you think we should drop upscaling and downscaling single plane
>>> tests from
>>> my series
>>> https://patchwork.freedesktop.org/series/98333/ ?
>>> That should be covered by your tests?
>>>
>>> What do you think?
>>
>> Hi Swati,
>>
>> Yep, sounds good! We have upscale tests for single plane in Petri's
>> patch and we can add downscale tests too.
>>
>> Thanks,
>>
>> Jessica Zhang
>>
>>>
>>> On 16-Feb-22 6:15 AM, Jessica Zhang wrote:
>>>>
>>>>
>>>> On 2/15/2022 4:23 AM, Swati Sharma wrote:
>>>>> Subtest for testing downscaling for each plane
>>>>> individually (checked 1 plane per "class" like
>>>>> we do in other tests).
>>>>>
>>>>> v2: -set modifier as LINEAR (Ville)
>>>>> -shared code for upscaling and downscaling tests (Ville)
>>>>> -removed num_scaler() check and added try_commit() (Ville)
>>>>>
>>>>> Signed-off-by: Swati Sharma <swati2.sharma at intel.com>
>>>>
>>>> Hi Swati,
>>>>
>>>> FYI, Petri and I have been working on a similar patch: [1][2]
>>>>
>>>> For the latest version we've modified `check_scaling_pipe_plane_rot`
>>>> so that it takes in a custom buffer height/width and created
>>>> multiple subtests for different scaling factors.
>>>>
>>>> I see that you've added tests for both upscaling and downscaling
>>>> that do something similar. Maybe our changes to
>>>> `check_scaling_pipe_plane_rot` can be useful for your series to
>>>> reduce redundant code in both the upscale and downscale cases? Let
>>>> me know your thoughts on this.
>>>>
>>>> Thanks,
>>>>
>>>> Jessica Zhang
>>>>
>>>> [1] latest version:
>>>> https://patchwork.freedesktop.org/patch/470921/?series=99364&rev=2
>>>> [2] earlier version:
>>>> https://patchwork.freedesktop.org/patch/470856/?series=97584&rev=5
>>>>
>>>>> ---
>>>>> tests/kms_plane_scaling.c | 55
>>>>> ++++++++++++++++++++++++++++++++++++++-
>>>>> 1 file changed, 54 insertions(+), 1 deletion(-)
>>>>>
>>>>> diff --git a/tests/kms_plane_scaling.c b/tests/kms_plane_scaling.c
>>>>> index 39f2de3c..e960a6c5 100644
>>>>> --- a/tests/kms_plane_scaling.c
>>>>> +++ b/tests/kms_plane_scaling.c
>>>>> @@ -32,6 +32,7 @@ IGT_TEST_DESCRIPTION("Test display plane scaling");
>>>>> /* Test flags. */
>>>>> enum {
>>>>> TEST_UPSCALING = 1 << 0,
>>>>> + TEST_DOWNSCALING = 1 << 1,
>>>>> };
>>>>> typedef struct {
>>>>> @@ -266,6 +267,48 @@ __test_plane_upscaling(data_t *d, igt_plane_t
>>>>> *plane,
>>>>> igt_skip_on_f(ret == -EINVAL, "Scaling op not supported\n");
>>>>> }
>>>>> +static void
>>>>> +__test_plane_downscaling(data_t *d, igt_plane_t *plane,
>>>>> + enum pipe pipe, igt_output_t *output)
>>>>> +{
>>>>> + igt_display_t *display = &d->display;
>>>>> + int width, height;
>>>>> + drmModeModeInfo *mode;
>>>>> + bool test_complete = false;
>>>>> + int scaling = 50;
>>>>> + int ret;
>>>>> +
>>>>> + cleanup_crtc(d);
>>>>> +
>>>>> + igt_output_set_pipe(output, pipe);
>>>>> + mode = igt_output_get_mode(output);
>>>>> + width = mode->hdisplay;
>>>>> + height = mode->vdisplay;
>>>>> +
>>>>> + igt_create_color_pattern_fb(display->drm_fd,
>>>>> + width, height,
>>>>> + DRM_FORMAT_XRGB8888,
>>>>> + I915_TILING_NONE,
>>>>> + 1.0, 0.0, 0.0, &d->fb[0]);
>>>>> +
>>>>> + igt_plane_set_fb(plane, &d->fb[0]);
>>>>> +
>>>>> + while(!test_complete) {
>>>>> + igt_plane_set_size(plane, (width * scaling) / 100, (height
>>>>> * scaling) / 100);
>>>>> + ret = igt_display_try_commit_atomic(display,
>>>>> DRM_MODE_ATOMIC_ALLOW_MODESET, NULL);
>>>>> +
>>>>> + if (ret != -EINVAL)
>>>>> + test_complete = true;
>>>>> +
>>>>> + scaling += 5;
>>>>> + }
>>>>> +
>>>>> + igt_plane_set_fb(plane, NULL);
>>>>> + igt_remove_fb(display->drm_fd, &d->fb[0]);
>>>>> +
>>>>> + igt_skip_on_f(ret == -EINVAL, "Scaling op not supported\n");
>>>>> +}
>>>>> +
>>>>> static void
>>>>> test_plane_scaling(data_t *d, enum pipe pipe, igt_output_t
>>>>> *output, uint32_t flags)
>>>>> {
>>>>> @@ -285,9 +328,12 @@ test_plane_scaling(data_t *d, enum pipe pipe,
>>>>> igt_output_t *output, uint32_t fla
>>>>> uint32_t format = plane->drm_plane->formats[j];
>>>>> if (test_format(d, &tested_formats, format) &&
>>>>> igt_plane_has_format_mod(plane, format, modifier) &&
>>>>> - can_scale(d, format))
>>>>> + can_scale(d, format)) {
>>>>> if (flags & TEST_UPSCALING)
>>>>> __test_plane_upscaling(d, plane, pipe, output);
>>>>> + if (flags & TEST_DOWNSCALING)
>>>>> + __test_plane_downscaling(d, plane, pipe, output);
>>>>> + }
>>>>> }
>>>>> igt_vec_fini(&tested_formats);
>>>>> @@ -618,6 +664,13 @@ igt_main_args("", long_opts, help_str,
>>>>> opt_handler, &data)
>>>>> test_plane_scaling(&data, pipe, output,
>>>>> TEST_UPSCALING);
>>>>> }
>>>>> + igt_describe("Tests plane downscaling.");
>>>>> + igt_subtest_with_dynamic("plane-downscaling") {
>>>>> + for_each_pipe_with_single_output(&data.display, pipe,
>>>>> output)
>>>>> + igt_dynamic_f("pipe-%s-%s-downscaling",
>>>>> kmstest_pipe_name(pipe), igt_output_name(output))
>>>>> + test_plane_scaling(&data, pipe, output,
>>>>> TEST_DOWNSCALING);
>>>>> + }
>>>>> +
>>>>> igt_describe("Tests scaling with pixel formats.");
>>>>> igt_subtest_with_dynamic("scaler-with-pixel-format") {
>>>>> for_each_pipe_with_single_output(&data.display, pipe,
>>>>> output)
>>>>> --
>>>>> 2.25.1
>>>>>
>>>
>>> --
>>> ~Swati Sharma
>
> --
> ~Swati Sharma
More information about the igt-dev
mailing list