media: hantro: Use H264_SCALING_MATRIX only when required

Baseline, Main and Extended profiles are specified to
not support a scaling matrix. Also, High profiles
can optionally specify a scaling matrix, using
SPS and PPS NAL units.

To meet this expectation, applications are required to
set the V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX control
and set the V4L2_H264_PPS_FLAG_SCALING_MATRIX_PRESENT
flag only when a scaling matrix is specified for a picture.

Implement this on hantro, which has hardware support for this
case.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Tested-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

drivers/staging/media/hantro/hantro_g1_h264_dec.c

@@ -59,9 +59,8 @@ static void set_params(struct hantro_ctx *ctx)
 	reg = G1_REG_DEC_CTRL2_CH_QP_OFFSET(pps->chroma_qp_index_offset) |
 	      G1_REG_DEC_CTRL2_CH_QP_OFFSET2(pps->second_chroma_qp_index_offset);
 
-	/* always use the matrix sent from userspace */
-	reg |= G1_REG_DEC_CTRL2_TYPE1_QUANT_E;
-
+	if (pps->flags & V4L2_H264_PPS_FLAG_SCALING_MATRIX_PRESENT)
+		reg |= G1_REG_DEC_CTRL2_TYPE1_QUANT_E;
 	if (!(sps->flags & V4L2_H264_SPS_FLAG_FRAME_MBS_ONLY))
 		reg |= G1_REG_DEC_CTRL2_FIELDPIC_FLAG_E;
 	vdpu_write_relaxed(vpu, reg, G1_REG_DEC_CTRL2);

drivers/staging/media/hantro/hantro_h264.c

@@ -197,6 +197,7 @@ assemble_scaling_list(struct hantro_ctx *ctx)
 {
 	const struct hantro_h264_dec_ctrls *ctrls = &ctx->h264_dec.ctrls;
 	const struct v4l2_ctrl_h264_scaling_matrix *scaling = ctrls->scaling;
+	const struct v4l2_ctrl_h264_pps *pps = ctrls->pps;
 	const size_t num_list_4x4 = ARRAY_SIZE(scaling->scaling_list_4x4);
 	const size_t list_len_4x4 = ARRAY_SIZE(scaling->scaling_list_4x4[0]);
 	const size_t list_len_8x8 = ARRAY_SIZE(scaling->scaling_list_8x8[0]);
@@ -205,6 +206,9 @@ assemble_scaling_list(struct hantro_ctx *ctx)
 	const u32 *src;
 	int i, j;
 
+	if (!(pps->flags & V4L2_H264_PPS_FLAG_SCALING_MATRIX_PRESENT))
+		return;
+
 	for (i = 0; i < num_list_4x4; i++) {
 		src = (u32 *)&scaling->scaling_list_4x4[i];
 		for (j = 0; j < list_len_4x4 / 4; j++)