Hi Maxime,
Dne torek, 05. december 2017 ob 11:36:18 CET je Maxime Ripard napisal(a):
Hi,
On Fri, Dec 01, 2017 at 07:05:23AM +0100, Jernej Skrabec wrote:
Current DE2 driver is very basic and uses a lot of magic constants since there is no documentation and knowledge about it was limited at the time.
With studying BSP source code, deeper knowledge was gained which allows to improve mainline driver considerably.
At the beginning of this series, some code refactoring is done as well as adding some checks (patches 1-15).
Further patches add multi-plane support with HW scaling and all possible RGB formats (patches 16-21).
At last, support for YUV formats is added (patches 22-26).
At the end, I included patch which puts lowest plane before second lowest. This should help testing VI planes when mixer has configuration 1 VI plane and 1 or more UI planes (most SoCs except V3s).
This code was developed on H3, but it should work on every SoC if correct configuration structure is provided.
H3 code can be found here: https://github.com/jernejsk/linux-1/commits/de2_impr_for_next
Thanks a lot for that huge rework.
I've applied the patches 1 to 26, and will push them to drm-misc once the compilations are done.
In the future, if you happen to do such a huge rework again (which hopefully won't be needed :)), please use the -M option of format-patch. It will reduce a lot the verbosity of files renaming and will help the review.
Noted.
I think I missed initialization of min_scaler and max_scaler in sun8i_vi_layer_atomic_check() in sun8i_vi_layer.c when I was reworking patches.
Will you fix patch with those two lines min_scale = DRM_PLANE_HELPER_NO_SCALING; max_scale = DRM_PLANE_HELPER_NO_SCALING;
or should I send new patch which fixes that or should I send new version of original patch?
Best regards, Jernej