Re: [RFC PATCH v6] media: mediatek: vcodec: support stateless AV1 decoder

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi again,

W dniu 17.11.2022 o 13:42, Andrzej Pietrasiewicz pisze:
Hi Xiaoyong Lu,

Sorry about chiming in only at v6. Please see inline below.

Andrzej

W dniu 17.11.2022 o 07:17, Xiaoyong Lu pisze:
Add mediatek av1 decoder linux driver which use the stateless API in
MT8195.

Signed-off-by: Xiaoyong Lu<xiaoyong.lu@xxxxxxxxxxxx>
---
Changes from v5:

- change av1 PROFILE and LEVEL cfg
- test by av1 fluster, result is 173/239

Changes from v4:

- convert vb2_find_timestamp to vb2_find_buffer
- test by av1 fluster, result is 173/239

Changes from v3:

- modify comment for struct vdec_av1_slice_slot
- add define SEG_LVL_ALT_Q
- change use_lr/use_chroma_lr parse from av1 spec
- use ARRAY_SIZE to replace size for loop_filter_level and loop_filter_mode_deltas
- change array size of loop_filter_mode_deltas from 4 to 2
- add define SECONDARY_FILTER_STRENGTH_NUM_BITS
- change some hex values from upper case to lower case
- change *dpb_sz equal to V4L2_AV1_TOTAL_REFS_PER_FRAME + 1
- test by av1 fluster, result is 173/239

Changes from v2:

- Match with av1 uapi v3 modify
- test by av1 fluster, result is 173/239

---
Reference series:
[1]: v3 of this series is presend by Daniel Almeida.
      message-id: 20220825225312.564619-1-daniel.almeida@xxxxxxxxxxxxx

  .../media/platform/mediatek/vcodec/Makefile   |    1 +
  .../vcodec/mtk_vcodec_dec_stateless.c         |   47 +-
  .../platform/mediatek/vcodec/mtk_vcodec_drv.h |    1 +
  .../vcodec/vdec/vdec_av1_req_lat_if.c         | 2234 +++++++++++++++++
  .../platform/mediatek/vcodec/vdec_drv_if.c    |    4 +
  .../platform/mediatek/vcodec/vdec_drv_if.h    |    1 +
  .../platform/mediatek/vcodec/vdec_msg_queue.c |   27 +
  .../platform/mediatek/vcodec/vdec_msg_queue.h |    4 +
  8 files changed, 2318 insertions(+), 1 deletion(-)
  create mode 100644 drivers/media/platform/mediatek/vcodec/vdec/vdec_av1_req_lat_if.c


<snip>

+
+static void *vdec_av1_get_ctrl_ptr(struct mtk_vcodec_ctx *ctx, int id)
+{
+    struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id);
+
+    if (!ctrl)
+        return ERR_PTR(-EINVAL);
+
+    return ctrl->p_cur.p;
+}

I see we keep repeating this kind of a v4l2_ctrl_find() wrapper in drivers.
The only reason this code cannot be factored out is the "context" struct pointer
pointing at structs of different types. Maybe we could

#define v4l2_get_ctrl_ptr(ctx, member, id) \
     __v4l2_get_ctrl_ptr((ctx), offsetof(typeof(*ctx), (member)), (id))

void *__v4l2_get_ctrl_ptr(void *ctx, size_t offset, u32 id)
{
     struct v4l2_ctrl_handler *hdl = (struct v4l2_ctrl_handler *)(ctx + offset);
     struct v4l2_ctrl *ctrl = v4l2_ctrl_find(hdl, id);

     if (!ctrl)
         return ERR_PTR(-EINVAL);

     return ctrl->p_cur.p;
}

and reuse v4l2_get_ctrl_ptr() in drivers?

A similar kind of void* arithmetic happens in container_of, only with '-'.


When I think of it it seems a bit over-engineered to me now, it would
be better to give up the macro and simply pass struct v4l2_ctrl_handler *hdl.

Another second thought is that including such a wrapper in this patch
would make it too noisy if all potential users were to be updated.
A separate series would make more sense.

Regards,

Andrzej




[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux