Re: [PATCH v7 2/3] drm/msm/dpu: allow using two SSPP blocks for a single plane

From: Abhinav Kumar
Date: Wed Dec 11 2024 - 17:42:01 EST




On 12/11/2024 2:24 PM, Dmitry Baryshkov wrote:
On Wed, Dec 11, 2024 at 01:51:51PM -0800, Abhinav Kumar wrote:


On 11/29/2024 5:55 PM, Dmitry Baryshkov wrote:
Virtual wide planes give high amount of flexibility, but it is not
always enough:

In parallel multirect case only the half of the usual width is supported
for tiled formats. Thus the whole width of two tiled multirect
rectangles can not be greater than max_linewidth, which is not enough
for some platforms/compositors.

Another example is as simple as wide YUV plane. YUV planes can not use
multirect, so currently they are limited to max_linewidth too.

Now that the planes are fully virtualized, add support for allocating
two SSPP blocks to drive a single DRM plane. This fixes both mentioned
cases and allows all planes to go up to 2*max_linewidth (at the cost of
making some of the planes unavailable to the user).


Overall looks so much cleaner after unification!

One small nit below,


You can still have,

Reviewed-by: Abhinav Kumar <quic_abhinavk@xxxxxxxxxxx>

Note: we have started testing this series with sc7180 CrOS, and will report
our findings/ give tested-by this week.


<snip>

+static bool dpu_plane_try_multirect_parallel(struct dpu_sw_pipe *pipe, struct dpu_sw_pipe_cfg *pipe_cfg,
+ struct dpu_sw_pipe *r_pipe, struct dpu_sw_pipe_cfg *r_pipe_cfg,
+ struct dpu_hw_sspp *sspp, const struct msm_format *fmt,
+ uint32_t max_linewidth)
+{
+ r_pipe->sspp = NULL;
+
+ pipe->multirect_index = DPU_SSPP_RECT_SOLO;
+ pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE;
+
+ r_pipe->multirect_index = DPU_SSPP_RECT_SOLO;
+ r_pipe->multirect_mode = DPU_SSPP_MULTIRECT_NONE;
+


There are two places where the multirect_index and multirect_mode are reset.
Would it be better to just have a small api dpu_plane_reset_multirect() and
do this there?

I'm not sure, what's the benefit. We can add an API to reset one pipe
(to also be able to use it in _dpu_plane_atomic_disable()), but then
it's just deduplication for the sake of deduplication.


Yeah I was thinking something like

dpu_plane_reset_multirect(pipe);
dpu_plane_reset_multirect(r_pipe);

But its only a minor benefit, hence as I wrote it as a nit. We can keep things as it is, if its unnecessary in your opinion.