media: s5p-mfc: Add YV12 and I420 multiplanar format support

YV12 and I420 format (3-plane) support is added. Stride information is
added to all formats and planes since it is necessary for YV12/I420
which are different from width.

Cc: linux-fsd@tesla.com
Signed-off-by: Smitha T Murthy <smithatmurthy@gmail.com>
Signed-off-by: Aakarsh Jain <aakarsh.jain@samsung.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>

1 files changed, 124 insertions, 33 deletions

diff --git a/drivers/media/platform/samsung/s5p-mfc/s5p_mfc_opr_v6.c b/drivers/media/platform/samsung/s5p-mfc/s5p_mfc_opr_v6.c
index a49aa6c1e30b..36aebca11313 100644
--- a/drivers/media/platform/samsung/s5p-mfc/s5p_mfc_opr_v6.c
+++ b/drivers/media/platform/samsung/s5p-mfc/s5p_mfc_opr_v6.c
@@ -488,16 +488,35 @@ static void s5p_mfc_dec_calc_dpb_size_v6(struct s5p_mfc_ctx *ctx)
 	struct s5p_mfc_dev *dev = ctx->dev;
 	ctx->buf_width = ALIGN(ctx->img_width, S5P_FIMV_NV12MT_HALIGN_V6);
 	ctx->buf_height = ALIGN(ctx->img_height, S5P_FIMV_NV12MT_VALIGN_V6);
+	ctx->chroma_size_1 = 0;
 	mfc_debug(2, "SEQ Done: Movie dimensions %dx%d,\n"
 			"buffer dimensions: %dx%d\n", ctx->img_width,
 			ctx->img_height, ctx->buf_width, ctx->buf_height);
 
-	ctx->luma_size = calc_plane(ctx->img_width, ctx->img_height);
-	ctx->chroma_size = calc_plane(ctx->img_width, (ctx->img_height >> 1));
+	switch (ctx->dst_fmt->fourcc) {
+	case V4L2_PIX_FMT_NV12M:
+	case V4L2_PIX_FMT_NV21M:
+		ctx->stride[0] = ALIGN(ctx->img_width, S5P_FIMV_NV12MT_HALIGN_V6);
+		ctx->stride[1] = ALIGN(ctx->img_width, S5P_FIMV_NV12MT_HALIGN_V6);
+		ctx->luma_size = calc_plane(ctx->stride[0], ctx->img_height);
+		ctx->chroma_size = calc_plane(ctx->stride[1], (ctx->img_height / 2));
+		break;
+	case V4L2_PIX_FMT_YUV420M:
+	case V4L2_PIX_FMT_YVU420M:
+		ctx->stride[0] = ALIGN(ctx->img_width, S5P_FIMV_NV12MT_HALIGN_V6);
+		ctx->stride[1] = ALIGN(ctx->img_width / 2, S5P_FIMV_NV12MT_HALIGN_V6);
+		ctx->stride[2] = ALIGN(ctx->img_width / 2, S5P_FIMV_NV12MT_HALIGN_V6);
+		ctx->luma_size = calc_plane(ctx->stride[0], ctx->img_height);
+		ctx->chroma_size = calc_plane(ctx->stride[1], (ctx->img_height / 2));
+		ctx->chroma_size_1 = calc_plane(ctx->stride[2], (ctx->img_height / 2));
+		break;
+	}
+
 	if (IS_MFCV8_PLUS(ctx->dev)) {
 		/* MFCv8 needs additional 64 bytes for luma,chroma dpb*/
 		ctx->luma_size += S5P_FIMV_D_ALIGN_PLANE_SIZE_V8;
 		ctx->chroma_size += S5P_FIMV_D_ALIGN_PLANE_SIZE_V8;
+		ctx->chroma_size_1 += S5P_FIMV_D_ALIGN_PLANE_SIZE_V8;
 	}
 
 	if (ctx->codec_mode == S5P_MFC_CODEC_H264_DEC ||
@@ -524,14 +543,40 @@ static void s5p_mfc_enc_calc_src_size_v6(struct s5p_mfc_ctx *ctx)
 	mb_width = MB_WIDTH(ctx->img_width);
 	mb_height = MB_HEIGHT(ctx->img_height);
 
-	ctx->buf_width = ALIGN(ctx->img_width, S5P_FIMV_NV12M_HALIGN_V6);
-	ctx->luma_size = ALIGN((mb_width * mb_height) * 256, 256);
-	ctx->chroma_size = ALIGN((mb_width * mb_height) * 128, 256);
-
-	/* MFCv7 needs pad bytes for Luma and Chroma */
-	if (IS_MFCV7_PLUS(ctx->dev)) {
+	if (IS_MFCV12(ctx->dev)) {
+		switch (ctx->src_fmt->fourcc) {
+		case V4L2_PIX_FMT_NV12M:
+		case V4L2_PIX_FMT_NV21M:
+			ctx->stride[0] = ALIGN(ctx->img_width, S5P_FIMV_NV12M_HALIGN_V6);
+			ctx->stride[1] = ALIGN(ctx->img_width, S5P_FIMV_NV12M_HALIGN_V6);
+			ctx->luma_size = ctx->stride[0] * ALIGN(ctx->img_height, 16);
+			ctx->chroma_size =  ctx->stride[0] * ALIGN(ctx->img_height / 2, 16);
+			break;
+		case V4L2_PIX_FMT_YUV420M:
+		case V4L2_PIX_FMT_YVU420M:
+			ctx->stride[0] = ALIGN(ctx->img_width, S5P_FIMV_NV12M_HALIGN_V6);
+			ctx->stride[1] = ALIGN(ctx->img_width / 2, S5P_FIMV_NV12M_HALIGN_V6);
+			ctx->stride[2] = ALIGN(ctx->img_width / 2, S5P_FIMV_NV12M_HALIGN_V6);
+			ctx->luma_size = ctx->stride[0] * ALIGN(ctx->img_height, 16);
+			ctx->chroma_size =  ctx->stride[1] * ALIGN(ctx->img_height / 2, 16);
+			ctx->chroma_size_1 =  ctx->stride[2] * ALIGN(ctx->img_height / 2, 16);
+			break;
+		}
 		ctx->luma_size += MFC_LUMA_PAD_BYTES_V7;
-		ctx->chroma_size += MFC_CHROMA_PAD_BYTES_V7;
+		ctx->chroma_size += MFC_CHROMA_PAD_BYTES_V12;
+		ctx->chroma_size_1 += MFC_CHROMA_PAD_BYTES_V12;
+	} else {
+		ctx->buf_width = ALIGN(ctx->img_width, S5P_FIMV_NV12M_HALIGN_V6);
+		ctx->stride[0] = ctx->buf_width;
+		ctx->stride[1] = ctx->buf_width;
+		ctx->luma_size = ALIGN((mb_width * mb_height) * 256, 256);
+		ctx->chroma_size = ALIGN((mb_width * mb_height) * 128, 256);
+		ctx->chroma_size_1 = 0;
+		/* MFCv7 needs pad bytes for Luma and Chroma */
+		if (IS_MFCV7_PLUS(ctx->dev)) {
+			ctx->luma_size += MFC_LUMA_PAD_BYTES_V7;
+			ctx->chroma_size += MFC_LUMA_PAD_BYTES_V7;
+		}
 	}
 }
 
@@ -578,15 +623,18 @@ static int s5p_mfc_set_dec_frame_buffer_v6(struct s5p_mfc_ctx *ctx)
 	writel(ctx->total_dpb_count, mfc_regs->d_num_dpb);
 	writel(ctx->luma_size, mfc_regs->d_first_plane_dpb_size);
 	writel(ctx->chroma_size, mfc_regs->d_second_plane_dpb_size);
-
+	if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_YUV420M || ctx->dst_fmt->fourcc ==
+			V4L2_PIX_FMT_YVU420M)
+		writel(ctx->chroma_size_1, mfc_regs->d_third_plane_dpb_size);
 	writel(buf_addr1, mfc_regs->d_scratch_buffer_addr);
 	writel(ctx->scratch_buf_size, mfc_regs->d_scratch_buffer_size);
 
 	if (IS_MFCV8_PLUS(dev)) {
-		writel(ctx->img_width,
-			mfc_regs->d_first_plane_dpb_stride_size);
-		writel(ctx->img_width,
-			mfc_regs->d_second_plane_dpb_stride_size);
+		writel(ctx->stride[0], mfc_regs->d_first_plane_dpb_stride_size);
+		writel(ctx->stride[1], mfc_regs->d_second_plane_dpb_stride_size);
+		if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_YUV420M || ctx->dst_fmt->fourcc ==
+				V4L2_PIX_FMT_YVU420M)
+			writel(ctx->stride[2], mfc_regs->d_third_plane_dpb_stride_size);
 	}
 
 	buf_addr1 += ctx->scratch_buf_size;
@@ -615,6 +663,13 @@ static int s5p_mfc_set_dec_frame_buffer_v6(struct s5p_mfc_ctx *ctx)
 					ctx->dst_bufs[i].cookie.raw.chroma);
 		writel(ctx->dst_bufs[i].cookie.raw.chroma,
 				mfc_regs->d_second_plane_dpb + i * 4);
+		if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_YUV420M || ctx->dst_fmt->fourcc ==
+				V4L2_PIX_FMT_YVU420M) {
+			mfc_debug(2, "\tChroma_1 %d: %zx\n", i, ctx
+					->dst_bufs[i].cookie.raw.chroma_1);
+			writel(ctx->dst_bufs[i].cookie.raw.chroma_1, mfc_regs->d_third_plane_dpb +
+					i * 4);
+		}
 	}
 	if (ctx->codec_mode == S5P_MFC_CODEC_H264_DEC ||
 			ctx->codec_mode == S5P_MFC_CODEC_H264_MVC_DEC ||
@@ -673,20 +728,24 @@ static int s5p_mfc_set_enc_stream_buffer_v6(struct s5p_mfc_ctx *ctx,
 }
 
 static void s5p_mfc_set_enc_frame_buffer_v6(struct s5p_mfc_ctx *ctx,
-		unsigned long y_addr, unsigned long c_addr)
+		unsigned long y_addr, unsigned long c_addr,
+		unsigned long c_1_addr)
 {
 	struct s5p_mfc_dev *dev = ctx->dev;
 	const struct s5p_mfc_regs *mfc_regs = dev->mfc_regs;
 
 	writel(y_addr, mfc_regs->e_source_first_plane_addr);
 	writel(c_addr, mfc_regs->e_source_second_plane_addr);
+	writel(c_1_addr, mfc_regs->e_source_third_plane_addr);
 
 	mfc_debug(2, "enc src y buf addr: 0x%08lx\n", y_addr);
 	mfc_debug(2, "enc src c buf addr: 0x%08lx\n", c_addr);
+	mfc_debug(2, "enc src cr buf addr: 0x%08lx\n", c_1_addr);
 }
 
 static void s5p_mfc_get_enc_frame_buffer_v6(struct s5p_mfc_ctx *ctx,
-		unsigned long *y_addr, unsigned long *c_addr)
+		unsigned long *y_addr, unsigned long *c_addr,
+		unsigned long *c_1_addr)
 {
 	struct s5p_mfc_dev *dev = ctx->dev;
 	const struct s5p_mfc_regs *mfc_regs = dev->mfc_regs;
@@ -694,12 +753,17 @@ static void s5p_mfc_get_enc_frame_buffer_v6(struct s5p_mfc_ctx *ctx,
 
 	*y_addr = readl(mfc_regs->e_encoded_source_first_plane_addr);
 	*c_addr = readl(mfc_regs->e_encoded_source_second_plane_addr);
+	if (ctx->src_fmt->fourcc == V4L2_PIX_FMT_YUV420M || ctx->src_fmt->fourcc ==
+			V4L2_PIX_FMT_YVU420M)
+		*c_1_addr = readl(mfc_regs->e_encoded_source_third_plane_addr);
+	else
+		*c_1_addr = 0;
 
 	enc_recon_y_addr = readl(mfc_regs->e_recon_luma_dpb_addr);
 	enc_recon_c_addr = readl(mfc_regs->e_recon_chroma_dpb_addr);
 
 	mfc_debug(2, "recon y addr: 0x%08lx y_addr: 0x%08lx\n", enc_recon_y_addr, *y_addr);
-	mfc_debug(2, "recon c addr: 0x%08lx\n", enc_recon_c_addr);
+	mfc_debug(2, "recon c addr: 0x%08lx c_addr: 0x%08lx\n", enc_recon_c_addr, *c_addr);
 }
 
 /* Set encoding ref & codec buffer */
@@ -876,6 +940,20 @@ static int s5p_mfc_set_enc_params(struct s5p_mfc_ctx *ctx)
 		writel(reg, mfc_regs->e_enc_options);
 		/* 0: NV12(CbCr), 1: NV21(CrCb) */
 		writel(0x0, mfc_regs->pixel_format);
+	} else if (ctx->src_fmt->fourcc == V4L2_PIX_FMT_YVU420M) {
+		/* 0: Linear, 1: 2D tiled*/
+		reg = readl(mfc_regs->e_enc_options);
+		reg &= ~(0x1 << 7);
+		writel(reg, mfc_regs->e_enc_options);
+		/* 2: YV12(CrCb), 3: I420(CrCb) */
+		writel(0x2, mfc_regs->pixel_format);
+	} else if (ctx->src_fmt->fourcc == V4L2_PIX_FMT_YUV420M) {
+		/* 0: Linear, 1: 2D tiled*/
+		reg = readl(mfc_regs->e_enc_options);
+		reg &= ~(0x1 << 7);
+		writel(reg, mfc_regs->e_enc_options);
+		/* 2: YV12(CrCb), 3: I420(CrCb) */
+		writel(0x3, mfc_regs->pixel_format);
 	}
 
 	/* memory structure recon. frame */
@@ -1686,8 +1764,12 @@ static int s5p_mfc_init_decode_v6(struct s5p_mfc_ctx *ctx)
 	else
 		writel(reg, mfc_regs->d_dec_options);
 
-	/* 0: NV12(CbCr), 1: NV21(CrCb) */
-	if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_NV21M)
+	/* 0: NV12(CbCr), 1: NV21(CrCb), 2: YV12(CrCb), 3: I420(CbCr) */
+	if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_YUV420M)
+		writel(0x3, mfc_regs->pixel_format);
+	else if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_YVU420M)
+		writel(0x2, mfc_regs->pixel_format);
+	else if (ctx->dst_fmt->fourcc == V4L2_PIX_FMT_NV21M)
 		writel(0x1, mfc_regs->pixel_format);
 	else
 		writel(0x0, mfc_regs->pixel_format);
@@ -1771,8 +1853,11 @@ static int s5p_mfc_init_encode_v6(struct s5p_mfc_ctx *ctx)
 
 	/* Set stride lengths for v7 & above */
 	if (IS_MFCV7_PLUS(dev)) {
-		writel(ctx->img_width, mfc_regs->e_source_first_plane_stride);
-		writel(ctx->img_width, mfc_regs->e_source_second_plane_stride);
+		writel(ctx->stride[0], mfc_regs->e_source_first_plane_stride);
+		writel(ctx->stride[1], mfc_regs->e_source_second_plane_stride);
+		if (ctx->src_fmt->fourcc == V4L2_PIX_FMT_YUV420M || ctx->src_fmt->fourcc ==
+				V4L2_PIX_FMT_YVU420M)
+			writel(ctx->stride[2], mfc_regs->e_source_third_plane_stride);
 	}
 
 	writel(ctx->inst_no, mfc_regs->instance_id);
@@ -1881,7 +1966,7 @@ static inline int s5p_mfc_run_enc_frame(struct s5p_mfc_ctx *ctx)
 	struct s5p_mfc_dev *dev = ctx->dev;
 	struct s5p_mfc_buf *dst_mb;
 	struct s5p_mfc_buf *src_mb;
-	unsigned long src_y_addr, src_c_addr, dst_addr;
+	unsigned long src_y_addr, src_c_addr, src_c_1_addr, dst_addr;
 	/*
 	unsigned int src_y_size, src_c_size;
 	*/
@@ -1899,22 +1984,28 @@ static inline int s5p_mfc_run_enc_frame(struct s5p_mfc_ctx *ctx)
 
 	if (list_empty(&ctx->src_queue)) {
 		/* send null frame */
-		s5p_mfc_set_enc_frame_buffer_v6(ctx, 0, 0);
+		s5p_mfc_set_enc_frame_buffer_v6(ctx, 0, 0, 0);
 		src_mb = NULL;
 	} else {
 		src_mb = list_entry(ctx->src_queue.next, struct s5p_mfc_buf, list);
 		src_mb->flags |= MFC_BUF_FLAG_USED;
 		if (src_mb->b->vb2_buf.planes[0].bytesused == 0) {
-			s5p_mfc_set_enc_frame_buffer_v6(ctx, 0, 0);
+			s5p_mfc_set_enc_frame_buffer_v6(ctx, 0, 0, 0);
 			ctx->state = MFCINST_FINISHING;
 		} else {
 			src_y_addr = vb2_dma_contig_plane_dma_addr(&src_mb->b->vb2_buf, 0);
 			src_c_addr = vb2_dma_contig_plane_dma_addr(&src_mb->b->vb2_buf, 1);
+			if (ctx->src_fmt->fourcc == V4L2_PIX_FMT_YUV420M || ctx->src_fmt->fourcc ==
+					V4L2_PIX_FMT_YVU420M)
+				src_c_1_addr = vb2_dma_contig_plane_dma_addr
+					(&src_mb->b->vb2_buf, 2);
+			else
+				src_c_1_addr = 0;
 
 			mfc_debug(2, "enc src y addr: 0x%08lx\n", src_y_addr);
 			mfc_debug(2, "enc src c addr: 0x%08lx\n", src_c_addr);
 
-			s5p_mfc_set_enc_frame_buffer_v6(ctx, src_y_addr, src_c_addr);
+			s5p_mfc_set_enc_frame_buffer_v6(ctx, src_y_addr, src_c_addr, src_c_1_addr);
 			if (src_mb->flags & MFC_BUF_FLAG_EOS)
 				ctx->state = MFCINST_FINISHING;
 		}
@@ -2436,10 +2527,9 @@ const struct s5p_mfc_regs *s5p_mfc_init_regs_v6_plus(struct s5p_mfc_dev *dev)
 	R(e_source_first_plane_stride, S5P_FIMV_E_SOURCE_FIRST_STRIDE_V7);
 	R(e_source_second_plane_stride, S5P_FIMV_E_SOURCE_SECOND_STRIDE_V7);
 	R(e_source_third_plane_stride, S5P_FIMV_E_SOURCE_THIRD_STRIDE_V7);
-	R(e_encoded_source_first_plane_addr,
-			S5P_FIMV_E_ENCODED_SOURCE_FIRST_ADDR_V7);
-	R(e_encoded_source_second_plane_addr,
-			S5P_FIMV_E_ENCODED_SOURCE_SECOND_ADDR_V7);
+	R(e_encoded_source_first_plane_addr, S5P_FIMV_E_ENCODED_SOURCE_FIRST_ADDR_V7);
+	R(e_encoded_source_second_plane_addr, S5P_FIMV_E_ENCODED_SOURCE_SECOND_ADDR_V7);
+	R(e_encoded_source_third_plane_addr, S5P_FIMV_E_ENCODED_SOURCE_THIRD_ADDR_V7);
 	R(e_vp8_options, S5P_FIMV_E_VP8_OPTIONS_V7);
 
 	if (!IS_MFCV8_PLUS(dev))
@@ -2454,16 +2544,17 @@ const struct s5p_mfc_regs *s5p_mfc_init_regs_v6_plus(struct s5p_mfc_dev *dev)
 	R(d_cpb_buffer_offset, S5P_FIMV_D_CPB_BUFFER_OFFSET_V8);
 	R(d_first_plane_dpb_size, S5P_FIMV_D_FIRST_PLANE_DPB_SIZE_V8);
 	R(d_second_plane_dpb_size, S5P_FIMV_D_SECOND_PLANE_DPB_SIZE_V8);
+	R(d_third_plane_dpb_size, S5P_FIMV_D_THIRD_PLANE_DPB_SIZE_V8);
 	R(d_scratch_buffer_addr, S5P_FIMV_D_SCRATCH_BUFFER_ADDR_V8);
 	R(d_scratch_buffer_size, S5P_FIMV_D_SCRATCH_BUFFER_SIZE_V8);
-	R(d_first_plane_dpb_stride_size,
-			S5P_FIMV_D_FIRST_PLANE_DPB_STRIDE_SIZE_V8);
-	R(d_second_plane_dpb_stride_size,
-			S5P_FIMV_D_SECOND_PLANE_DPB_STRIDE_SIZE_V8);
+	R(d_first_plane_dpb_stride_size, S5P_FIMV_D_FIRST_PLANE_DPB_STRIDE_SIZE_V8);
+	R(d_second_plane_dpb_stride_size, S5P_FIMV_D_SECOND_PLANE_DPB_STRIDE_SIZE_V8);
+	R(d_third_plane_dpb_stride_size, S5P_FIMV_D_THIRD_PLANE_DPB_STRIDE_SIZE_V8);
 	R(d_mv_buffer_size, S5P_FIMV_D_MV_BUFFER_SIZE_V8);
 	R(d_num_mv, S5P_FIMV_D_NUM_MV_V8);
 	R(d_first_plane_dpb, S5P_FIMV_D_FIRST_PLANE_DPB_V8);
 	R(d_second_plane_dpb, S5P_FIMV_D_SECOND_PLANE_DPB_V8);
+	R(d_third_plane_dpb, S5P_FIMV_D_THIRD_PLANE_DPB_V8);
 	R(d_mv_buffer, S5P_FIMV_D_MV_BUFFER_V8);
 	R(d_init_buffer_options, S5P_FIMV_D_INIT_BUFFER_OPTIONS_V8);
 	R(d_available_dpb_flag_lower, S5P_FIMV_D_AVAILABLE_DPB_FLAG_LOWER_V8);