[PATCH v5 00/15] Add MPEG-2 decoding to Rockchip VPU

[PATCH v5 00/15] Add MPEG-2 decoding to Rockchip VPU

[Boris Brezillon]

This series introduces the decoding infrastructure that will be
used to add support for other codecs such as VP8, VP9 and H.264.

As explained in the cover letter for the v1 patchset,
the driver is now exposing two video device nodes.
The VPU encoder is exposed on /dev/video0, and the VPU decoder
is exposed on /dev/video1. Both devices are tied to the same
memory-to-memory queue, and same media device /dev/media0.

Therefore there are two media graphs:

┌────────────────────────────────┐
│ rockchip,rk3399-vpu-enc-source │
│          /dev/video0           │
└────────────────────────────────┘
  ┃
  ┃
  ▼
┌────────────────────────────────┐
│  rockchip,rk3399-vpu-enc-proc  │
└────────────────────────────────┘
  ┃
  ┃
  ▼
┌────────────────────────────────┐
│  rockchip,rk3399-vpu-enc-sink  │
│          /dev/video0           │
└────────────────────────────────┘

┌────────────────────────────────┐
│ rockchip,rk3399-vpu-dec-source │
│          /dev/video1           │
└────────────────────────────────┘
  ┃
  ┃
  ▼
┌────────────────────────────────┐
│  rockchip,rk3399-vpu-dec-proc  │
└────────────────────────────────┘
  ┃
  ┃
  ▼
┌────────────────────────────────┐
│  rockchip,rk3399-vpu-dec-sink  │
│          /dev/video1           │
└────────────────────────────────┘

Of course, this work has been possible thanks to Jonas Karlman, who did
the initial MPEG-2 decoding work and also got mpv+ffmpeg working using
the Request API.

Regards,

Boris

v5:
 * Add helper to clamp height/width based on frmsize
 * Modify fill_fmt() helpers to keep user-defined sizeimage/bytesperline
   when they're big enough
 * Fixed a bunch of problems reported by Hans (see per-patch changelog
   for more details)

v4:
 * Fix issues in the probe/remove path preventing the driver from being
   unloaded/reloaded (Jonas)
 * Rework the media controller registration code to simplify the
   error/remove path (Boris)
 * Split the MPEG2 patches (one adding the generic code and one patch
   per platform)

v3:
  * Share more code between the encoder/decoder logic (VB2/V4L2
    interface implementation)
  * Add a patch to support MPEG2 decoding on RK3288 (Jonas)
  * Several fixes/improvements all over the place (see changelog
    in each patch for more details)

v2:
  * Fixed some minor issues brought up by v4l2-compliance.
  * Fixed bytesused wrongly assigned 0 on the MPEG-2 decoder.
  * Addressed comments from Hans and Tomasz on the pixel format
    helpers.

Boris Brezillon (8):
  media: v4l2-common: Fix v4l2_fill_pixfmt[_mp]() prototypes
  media: v4l2-common: Add an helper to apply frmsize constraints
  media: v4l2-common: Support custom imagesize/bytesperline in
    fill_pixfmt()
  rockchip/vpu: Use v4l2_apply_frmsize_constraints() where appropriate
  rockchip/vpu: Rename rockchip_vpu_common.h into rockchip_vpu_v4l2.h
  rockchip/vpu: Move encoder logic to a common place
  rockchip/vpu: Provide a helper to reset both src and dst formats
  rockchip/vpu: Prepare things to support decoders

Ezequiel Garcia (4):
  rockchip/vpu: Open-code media controller register
  rockchip/vpu: Support the Request API
  rockchip/vpu: Add decoder boilerplate
  rockchip/vpu: Add support for non-standard controls

Jonas Karlman (3):
  rockchip/vpu: Add infra to support MPEG-2 decoding
  rockchip/vpu: Add MPEG2 decoding support to RK3399
  rockchip/vpu: Add support for MPEG-2 decoding on RK3288

 drivers/media/v4l2-core/v4l2-common.c         |  86 ++-
 drivers/staging/media/rockchip/vpu/Kconfig    |   1 +
 drivers/staging/media/rockchip/vpu/Makefile   |   7 +-
 .../media/rockchip/vpu/rk3288_vpu_hw.c        |  59 +-
 .../rockchip/vpu/rk3288_vpu_hw_jpeg_enc.c     |   2 +-
 .../rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c    | 261 +++++++
 .../media/rockchip/vpu/rk3288_vpu_regs.h      |   1 +
 .../media/rockchip/vpu/rk3399_vpu_hw.c        |  59 +-
 .../rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c     |   8 +-
 .../rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c    | 267 +++++++
 .../staging/media/rockchip/vpu/rockchip_vpu.h | 110 ++-
 .../media/rockchip/vpu/rockchip_vpu_drv.c     | 467 ++++++++++--
 .../media/rockchip/vpu/rockchip_vpu_enc.c     | 571 ---------------
 .../media/rockchip/vpu/rockchip_vpu_hw.h      |  16 +
 .../media/rockchip/vpu/rockchip_vpu_mpeg2.c   |  61 ++
 .../media/rockchip/vpu/rockchip_vpu_v4l2.c    | 692 ++++++++++++++++++
 ...kchip_vpu_common.h => rockchip_vpu_v4l2.h} |  15 +-
 include/media/v4l2-common.h                   |  10 +-
 18 files changed, 2009 insertions(+), 684 deletions(-)
 create mode 100644 drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c
 create mode 100644 drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c
 delete mode 100644 drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
 create mode 100644 drivers/staging/media/rockchip/vpu/rockchip_vpu_mpeg2.c
 create mode 100644 drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
 rename drivers/staging/media/rockchip/vpu/{rockchip_vpu_common.h => rockchip_vpu_v4l2.h} (50%)

-- 
2.20.1

[PATCH v5 01/15] media: v4l2-common: Fix v4l2_fill_pixfmt[_mp]() prototypes

[Boris Brezillon]

Width/height and 4CC formats are expressed using u32 types everywhere,
let's fix the v4l2_fill_pixfmt[_mp]() prototypes to do the same.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes in v5:
* New patch
---
 drivers/media/v4l2-core/v4l2-common.c | 5 +++--
 include/media/v4l2-common.h           | 8 ++++----
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
index d869a2910435..6ae84aa1f7a9 100644
--- a/drivers/media/v4l2-core/v4l2-common.c
+++ b/drivers/media/v4l2-core/v4l2-common.c
@@ -520,7 +520,7 @@ static inline unsigned int v4l2_format_block_height(const struct v4l2_format_inf
 }
 
 int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
-			 int pixelformat, int width, int height)
+			u32 pixelformat, u32 width, u32 height)
 {
 	const struct v4l2_format_info *info;
 	struct v4l2_plane_pix_format *plane;
@@ -574,7 +574,8 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
 }
 EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt_mp);
 
-int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, int pixelformat, int width, int height)
+int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
+		     u32 width, u32 height)
 {
 	const struct v4l2_format_info *info;
 	int i;
diff --git a/include/media/v4l2-common.h b/include/media/v4l2-common.h
index 0a41bbecf3d3..3226bc8107cc 100644
--- a/include/media/v4l2-common.h
+++ b/include/media/v4l2-common.h
@@ -420,9 +420,9 @@ struct v4l2_format_info {
 
 const struct v4l2_format_info *v4l2_format_info(u32 format);
 
-int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, int pixelformat,
-		     int width, int height);
-int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt, int pixelformat,
-			int width, int height);
+int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
+		     u32 width, u32 height);
+int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt, u32 pixelformat,
+			u32 width, u32 height);
 
 #endif /* V4L2_COMMON_H_ */
-- 
2.20.1

[PATCH v5 02/15] media: v4l2-common: Add an helper to apply frmsize constraints

[Boris Brezillon]

The rockchip VPU driver is open-coding this logic which seems pretty
generic. Let's provide an helper to apply the min/max and alignment
constraints on width/height.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes in v5:
* New patch
---
 drivers/media/v4l2-core/v4l2-common.c | 27 +++++++++++++++++++++++++++
 include/media/v4l2-common.h           |  2 ++
 2 files changed, 29 insertions(+)

diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
index 6ae84aa1f7a9..3c6f5c115fc5 100644
--- a/drivers/media/v4l2-core/v4l2-common.c
+++ b/drivers/media/v4l2-core/v4l2-common.c
@@ -331,6 +331,16 @@ static unsigned int clamp_align(unsigned int x, unsigned int min,
 	return x;
 }
 
+static unsigned int clamp_roundup(unsigned int x, unsigned int min,
+				   unsigned int max, unsigned int alignment)
+{
+	x = clamp(x, min, max);
+	if (alignment)
+		x = round_up(x, alignment);
+
+	return x;
+}
+
 void v4l_bound_align_image(u32 *w, unsigned int wmin, unsigned int wmax,
 			   unsigned int walign,
 			   u32 *h, unsigned int hmin, unsigned int hmax,
@@ -519,6 +529,23 @@ static inline unsigned int v4l2_format_block_height(const struct v4l2_format_inf
 	return info->block_h[plane];
 }
 
+void v4l2_apply_frmsize_constraints(u32 *width, u32 *height,
+				    const struct v4l2_frmsize_stepwise *frmsize)
+{
+	if (!frmsize)
+		return;
+
+	/*
+	 * Clamp width/height to meet min/max constraints and round it up to
+	 * macroblock alignment.
+	 */
+	*width = clamp_roundup(*width, frmsize->min_width, frmsize->max_width,
+			       frmsize->step_width);
+        *height = clamp_roundup(*height, frmsize->min_height, frmsize->max_height,
+				frmsize->step_height);
+}
+EXPORT_SYMBOL_GPL(v4l2_apply_frmsize_constraints);
+
 int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
 			u32 pixelformat, u32 width, u32 height)
 {
diff --git a/include/media/v4l2-common.h b/include/media/v4l2-common.h
index 3226bc8107cc..e826b154bc35 100644
--- a/include/media/v4l2-common.h
+++ b/include/media/v4l2-common.h
@@ -420,6 +420,8 @@ struct v4l2_format_info {
 
 const struct v4l2_format_info *v4l2_format_info(u32 format);
 
+void v4l2_apply_frmsize_constraints(u32 *width, u32 *height,
+				    const struct v4l2_frmsize_stepwise *frmsize);
 int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
 		     u32 width, u32 height);
 int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt, u32 pixelformat,
-- 
2.20.1

[PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Boris Brezillon]

Users can define custom sizeimage and bytesperline as long as they're
big enough to store the amount of pixels required for a specific
width/height under a specific format. Avoid overriding those fields in
this case.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Hello Hans,

The sizeimage/bytesperline check on !MPLANE formats is still not 100%
sure, as custom bytesperline might induce bigger sizeimage than what
we calculate.

I tried implementing something smarter taking the per-component plane
bpp + hdiv param as we discussed the other day but decided to step
back after realizing the per-component plane macro block might also
differ at least in theory (not sure that's true in practice) and that
has an impact on bytesperline too.

Let me know how you want to handle that case.

Regards,

Boris

Changes in v5:
* New patch
---
 drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
 1 file changed, 39 insertions(+), 15 deletions(-)

diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
index 3c6f5c115fc5..37bfc984a8b5 100644
--- a/drivers/media/v4l2-core/v4l2-common.c
+++ b/drivers/media/v4l2-core/v4l2-common.c
@@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
 	pixfmt->num_planes = info->mem_planes;
 
 	if (info->mem_planes == 1) {
+		u32 bytesperline, sizeimage = 0;
+
 		plane = &pixfmt->plane_fmt[0];
-		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
-		plane->sizeimage = 0;
+		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
 
 		for (i = 0; i < info->comp_planes; i++) {
 			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
@@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
 			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
 			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
 
-			plane->sizeimage += info->bpp[i] *
-				DIV_ROUND_UP(aligned_width, hdiv) *
-				DIV_ROUND_UP(aligned_height, vdiv);
+			sizeimage += info->bpp[i] *
+				     DIV_ROUND_UP(aligned_width, hdiv) *
+				     DIV_ROUND_UP(aligned_height, vdiv);
 		}
+
+		/*
+		 * The user might have specified custom sizeimage/bytesperline,
+		 * only override them if they're not big enough.
+		 */
+		plane->sizeimage = max(sizeimage, plane->sizeimage);
+		plane->bytesperline = max(bytesperline, plane->bytesperline);
 	} else {
 		for (i = 0; i < info->comp_planes; i++) {
 			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
@@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
 			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
 
 			plane = &pixfmt->plane_fmt[i];
-			plane->bytesperline =
-				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
-			plane->sizeimage =
-				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
+
+			/*
+			 * The user might have specified custom
+			 * sizeimage/bytesperline, only override them if
+			 * they're not big enough.
+			 */
+			plane->bytesperline = max_t(u32,
+						    info->bpp[i] *
+						    DIV_ROUND_UP(aligned_width, hdiv),
+						    plane->bytesperline);
+			plane->sizeimage = max_t(u32,
+						 plane->bytesperline *
+						 DIV_ROUND_UP(aligned_height, vdiv),
+						 plane->sizeimage);
 		}
 	}
 	return 0;
@@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
 		     u32 width, u32 height)
 {
 	const struct v4l2_format_info *info;
+	u32 bytesperline, sizeimage = 0;
 	int i;
 
 	info = v4l2_format_info(pixelformat);
@@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
 	pixfmt->width = width;
 	pixfmt->height = height;
 	pixfmt->pixelformat = pixelformat;
-	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
-	pixfmt->sizeimage = 0;
+	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
 
 	for (i = 0; i < info->comp_planes; i++) {
 		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
@@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
 
 		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
 		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
-
-		pixfmt->sizeimage += info->bpp[i] *
-			DIV_ROUND_UP(aligned_width, hdiv) *
-			DIV_ROUND_UP(aligned_height, vdiv);
+		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
+			     DIV_ROUND_UP(aligned_height, vdiv);
 	}
+
+	/*
+	 * The user might have specified its own sizeimage/bytesperline values,
+	 * only override them if they're not big enough.
+	 */
+	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
+	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
+
 	return 0;
 }
 EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);
-- 
2.20.1

[PATCH v5 04/15] rockchip/vpu: Use v4l2_apply_frmsize_constraints() where appropriate

[Boris Brezillon]

Use the v4l2_apply_frmsize_constraints() helper instead of open-coding
it.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes in v5:
* New patch
---
 .../media/rockchip/vpu/rockchip_vpu_enc.c     | 44 ++++++-------------
 1 file changed, 14 insertions(+), 30 deletions(-)

diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
index 7c7c20ab2733..aa00df9a7ecb 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
@@ -197,15 +197,9 @@ vidioc_try_fmt_cap_mplane(struct file *file, void *priv, struct v4l2_format *f)
 
 	pix_mp->num_planes = 1;
 	pix_mp->field = V4L2_FIELD_NONE;
-	pix_mp->width = clamp(pix_mp->width,
-			      fmt->frmsize.min_width,
-			      fmt->frmsize.max_width);
-	pix_mp->height = clamp(pix_mp->height,
-			       fmt->frmsize.min_height,
-			       fmt->frmsize.max_height);
-	/* Round up to macroblocks. */
-	pix_mp->width = round_up(pix_mp->width, fmt->frmsize.step_width);
-	pix_mp->height = round_up(pix_mp->height, fmt->frmsize.step_height);
+
+	v4l2_apply_frmsize_constraints(&pix_mp->width, &pix_mp->height,
+				       &fmt->frmsize);
 
 	/*
 	 * For compressed formats the application can specify
@@ -226,7 +220,6 @@ vidioc_try_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
 	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
 	const struct rockchip_vpu_fmt *fmt;
-	unsigned int width, height;
 	int i;
 
 	vpu_debug(4, "%c%c%c%c\n",
@@ -242,18 +235,13 @@ vidioc_try_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 	}
 
 	pix_mp->field = V4L2_FIELD_NONE;
-	width = clamp(pix_mp->width,
-		      ctx->vpu_dst_fmt->frmsize.min_width,
-		      ctx->vpu_dst_fmt->frmsize.max_width);
-	height = clamp(pix_mp->height,
-		       ctx->vpu_dst_fmt->frmsize.min_height,
-		       ctx->vpu_dst_fmt->frmsize.max_height);
-	/* Round up to macroblocks. */
-	width = round_up(width, ctx->vpu_dst_fmt->frmsize.step_width);
-	height = round_up(height, ctx->vpu_dst_fmt->frmsize.step_height);
+
+	v4l2_apply_frmsize_constraints(&pix_mp->width, &pix_mp->height,
+				       &ctx->vpu_dst_fmt->frmsize);
 
 	/* Fill remaining fields */
-	v4l2_fill_pixfmt_mp(pix_mp, fmt->fourcc, width, height);
+	v4l2_fill_pixfmt_mp(pix_mp, fmt->fourcc, pix_mp->width,
+			    pix_mp->height);
 
 	for (i = 0; i < pix_mp->num_planes; i++) {
 		memset(pix_mp->plane_fmt[i].reserved, 0,
@@ -272,10 +260,8 @@ void rockchip_vpu_enc_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
 	memset(fmt, 0, sizeof(*fmt));
 
 	fmt->num_planes = 1;
-	fmt->width = clamp(fmt->width, ctx->vpu_dst_fmt->frmsize.min_width,
-			   ctx->vpu_dst_fmt->frmsize.max_width);
-	fmt->height = clamp(fmt->height, ctx->vpu_dst_fmt->frmsize.min_height,
-			    ctx->vpu_dst_fmt->frmsize.max_height);
+	v4l2_apply_frmsize_constraints(&fmt->width, &fmt->height,
+				       &ctx->vpu_dst_fmt->frmsize);
 	fmt->pixelformat = ctx->vpu_dst_fmt->fourcc;
 	fmt->field = V4L2_FIELD_NONE;
 	fmt->colorspace = V4L2_COLORSPACE_JPEG,
@@ -291,23 +277,21 @@ void rockchip_vpu_enc_reset_src_fmt(struct rockchip_vpu_dev *vpu,
 				    struct rockchip_vpu_ctx *ctx)
 {
 	struct v4l2_pix_format_mplane *fmt = &ctx->src_fmt;
-	unsigned int width, height;
 
 	ctx->vpu_src_fmt = rockchip_vpu_get_default_fmt(ctx, false);
 
 	memset(fmt, 0, sizeof(*fmt));
 
-	width = clamp(fmt->width, ctx->vpu_dst_fmt->frmsize.min_width,
-		      ctx->vpu_dst_fmt->frmsize.max_width);
-	height = clamp(fmt->height, ctx->vpu_dst_fmt->frmsize.min_height,
-		       ctx->vpu_dst_fmt->frmsize.max_height);
+	v4l2_apply_frmsize_constraints(&fmt->width, &fmt->height,
+				       &ctx->vpu_src_fmt->frmsize);
 	fmt->field = V4L2_FIELD_NONE;
 	fmt->colorspace = V4L2_COLORSPACE_JPEG,
 	fmt->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT;
 	fmt->quantization = V4L2_QUANTIZATION_DEFAULT;
 	fmt->xfer_func = V4L2_XFER_FUNC_DEFAULT;
 
-	v4l2_fill_pixfmt_mp(fmt, ctx->vpu_src_fmt->fourcc, width, height);
+	v4l2_fill_pixfmt_mp(fmt, ctx->vpu_src_fmt->fourcc, fmt->width,
+			    fmt->height);
 }
 
 static int
-- 
2.20.1

[PATCH v5 05/15] rockchip/vpu: Open-code media controller register

[Boris Brezillon]

From: Ezequiel Garcia <ezequiel@collabora.com>

In preparation to support decoders, using a single memory-to-memory
device, we need to roll our own media controller entities registration.

To do that, we define a rockchip_vpu_func object that embeds the
video_device object plus all the elements that are needed to attach this
vdev to the media device.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
--
Changes from v3:
* Rework the media controller registration logic (Boris)
* Fix media controller deregistration (Jonas)

Changes from v2:
* Use kvasprintf instead of kmalloc and snprintf.
* Fix missing kfree in error paths.
* Remove unneeded media_remove_intf_links on error paths.
---
 .../staging/media/rockchip/vpu/rockchip_vpu.h |  39 +++-
 .../media/rockchip/vpu/rockchip_vpu_drv.c     | 209 +++++++++++++++---
 2 files changed, 216 insertions(+), 32 deletions(-)

diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
index b15c02333a70..aba257c663a7 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
@@ -71,12 +71,47 @@ enum rockchip_vpu_codec_mode {
 	RK_VPU_MODE_JPEG_ENC,
 };
 
+/*
+ * struct rockchip_vpu_func - rockchip VPU functionality
+ *
+ * @id:			processing functionality ID (can be
+ *			%MEDIA_ENT_F_PROC_VIDEO_ENCODER or
+ *			%MEDIA_ENT_F_PROC_VIDEO_DECODER)
+ * @vdev:		&struct video_device that exposes the encoder or
+ *			decoder functionality
+ * @source_pad:		&struct media_pad with the source pad.
+ * @sink:		&struct media_entity pointer with the sink entity
+ * @sink_pad:		&struct media_pad with the sink pad.
+ * @proc:		&struct media_entity pointer with the M2M device itself.
+ * @proc_pads:		&struct media_pad with the @proc pads.
+ * @intf_devnode:	&struct media_intf devnode pointer with the interface
+ *			with controls the M2M device.
+ *
+ * Contains everything needed to attach the video device to the media device.
+ */
+struct rockchip_vpu_func {
+	unsigned int id;
+	struct video_device vdev;
+	struct media_pad source_pad;
+	struct media_entity sink;
+	struct media_pad sink_pad;
+	struct media_entity proc;
+	struct media_pad proc_pads[2];
+	struct media_intf_devnode *intf_devnode;
+};
+
+static inline struct rockchip_vpu_func *
+rockchip_vpu_vdev_to_func(struct video_device *vdev)
+{
+	return container_of(vdev, struct rockchip_vpu_func, vdev);
+}
+
 /**
  * struct rockchip_vpu_dev - driver data
  * @v4l2_dev:		V4L2 device to register video devices for.
  * @m2m_dev:		mem2mem device associated to this device.
  * @mdev:		media device associated to this device.
- * @vfd_enc:		Video device for encoder.
+ * @encoder:		encoder functionality.
  * @pdev:		Pointer to VPU platform device.
  * @dev:		Pointer to device for convenient logging using
  *			dev_ macros.
@@ -93,7 +128,7 @@ struct rockchip_vpu_dev {
 	struct v4l2_device v4l2_dev;
 	struct v4l2_m2m_dev *m2m_dev;
 	struct media_device mdev;
-	struct video_device *vfd_enc;
+	struct rockchip_vpu_func *encoder;
 	struct platform_device *pdev;
 	struct device *dev;
 	struct clk_bulk_data clocks[ROCKCHIP_VPU_MAX_CLOCKS];
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index 3c3ce3baeb6d..fa02354a0f8a 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -239,6 +239,7 @@ static int rockchip_vpu_open(struct file *filp)
 {
 	struct rockchip_vpu_dev *vpu = video_drvdata(filp);
 	struct video_device *vdev = video_devdata(filp);
+	struct rockchip_vpu_func *func = rockchip_vpu_vdev_to_func(vdev);
 	struct rockchip_vpu_ctx *ctx;
 	int ret;
 
@@ -256,7 +257,7 @@ static int rockchip_vpu_open(struct file *filp)
 		return -ENOMEM;
 
 	ctx->dev = vpu;
-	if (vdev == vpu->vfd_enc)
+	if (func->id == MEDIA_ENT_F_PROC_VIDEO_ENCODER)
 		ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(vpu->m2m_dev, ctx,
 						    &enc_queue_init);
 	else
@@ -324,52 +325,206 @@ static const struct of_device_id of_rockchip_vpu_match[] = {
 };
 MODULE_DEVICE_TABLE(of, of_rockchip_vpu_match);
 
-static int rockchip_vpu_video_device_register(struct rockchip_vpu_dev *vpu)
+static int rockchip_vpu_register_entity(struct media_device *mdev,
+					struct media_entity *entity,
+					const char *entity_name,
+					struct media_pad *pads, int num_pads,
+					int function,
+					struct video_device *vdev)
+{
+	char *name;
+	int ret;
+
+	entity->obj_type = MEDIA_ENTITY_TYPE_BASE;
+	if (function == MEDIA_ENT_F_IO_V4L) {
+		entity->info.dev.major = VIDEO_MAJOR;
+		entity->info.dev.minor = vdev->minor;
+	}
+
+	name = devm_kasprintf(mdev->dev, GFP_KERNEL, "%s-%s", vdev->name,
+			      entity_name);
+	if (!name)
+		return -ENOMEM;
+
+	entity->name = name;
+	entity->function = function;
+
+	ret = media_entity_pads_init(entity, num_pads, pads);
+	if (ret)
+		return ret;
+
+	ret = media_device_register_entity(mdev, entity);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int rockchip_attach_func(struct rockchip_vpu_dev *vpu,
+				struct rockchip_vpu_func *func)
+{
+	struct media_device *mdev = &vpu->mdev;
+	struct media_link *link;
+	int ret;
+
+	/* Create the three encoder entities with their pads */
+	func->source_pad.flags = MEDIA_PAD_FL_SOURCE;
+	ret = rockchip_vpu_register_entity(mdev, &func->vdev.entity,
+					   "source", &func->source_pad, 1,
+					   MEDIA_ENT_F_IO_V4L, &func->vdev);
+	if (ret)
+		return ret;
+
+	func->proc_pads[0].flags = MEDIA_PAD_FL_SINK;
+	func->proc_pads[1].flags = MEDIA_PAD_FL_SOURCE;
+	ret = rockchip_vpu_register_entity(mdev, &func->proc, "proc",
+					   func->proc_pads, 2, func->id,
+					   &func->vdev);
+	if (ret)
+		goto err_rel_entity0;
+
+	func->sink_pad.flags = MEDIA_PAD_FL_SINK;
+	ret = rockchip_vpu_register_entity(mdev, &func->sink, "sink",
+					   &func->sink_pad, 1,
+					   MEDIA_ENT_F_IO_V4L, &func->vdev);
+	if (ret)
+		goto err_rel_entity1;
+
+	/* Connect the three entities */
+	ret = media_create_pad_link(&func->vdev.entity, 0, &func->proc, 1,
+				    MEDIA_LNK_FL_IMMUTABLE |
+				    MEDIA_LNK_FL_ENABLED);
+	if (ret)
+		goto err_rel_entity2;
+
+	ret = media_create_pad_link(&func->proc, 0, &func->sink, 0,
+				    MEDIA_LNK_FL_IMMUTABLE |
+				    MEDIA_LNK_FL_ENABLED);
+	if (ret)
+		goto err_rm_links0;
+
+	/* Create video interface */
+	func->intf_devnode = media_devnode_create(mdev, MEDIA_INTF_T_V4L_VIDEO,
+						  0, VIDEO_MAJOR,
+						  func->vdev.minor);
+	if (!func->intf_devnode) {
+		ret = -ENOMEM;
+		goto err_rm_links1;
+	}
+
+	/* Connect the two DMA engines to the interface */
+	link = media_create_intf_link(&func->vdev.entity,
+				      &func->intf_devnode->intf,
+				      MEDIA_LNK_FL_IMMUTABLE |
+				      MEDIA_LNK_FL_ENABLED);
+	if (!link) {
+		ret = -ENOMEM;
+		goto err_rm_devnode;
+	}
+
+	link = media_create_intf_link(&func->sink, &func->intf_devnode->intf,
+				      MEDIA_LNK_FL_IMMUTABLE |
+				      MEDIA_LNK_FL_ENABLED);
+	if (!link) {
+		ret = -ENOMEM;
+		goto err_rm_devnode;
+	}
+	return 0;
+
+err_rm_devnode:
+	media_devnode_remove(func->intf_devnode);
+
+err_rm_links1:
+	media_entity_remove_links(&func->sink);
+
+err_rm_links0:
+	media_entity_remove_links(&func->proc);
+	media_entity_remove_links(&func->vdev.entity);
+
+err_rel_entity2:
+	media_device_unregister_entity(&func->sink);
+
+err_rel_entity1:
+	media_device_unregister_entity(&func->proc);
+
+err_rel_entity0:
+	media_device_unregister_entity(&func->vdev.entity);
+	return ret;
+}
+
+static void rockchip_detach_func(struct rockchip_vpu_func *func)
+{
+	media_devnode_remove(func->intf_devnode);
+	media_entity_remove_links(&func->sink);
+	media_entity_remove_links(&func->proc);
+	media_entity_remove_links(&func->vdev.entity);
+	media_device_unregister_entity(&func->sink);
+	media_device_unregister_entity(&func->proc);
+	media_device_unregister_entity(&func->vdev.entity);
+}
+
+static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
 {
 	const struct of_device_id *match;
+	struct rockchip_vpu_func *func;
 	struct video_device *vfd;
-	int function, ret;
+	int ret;
 
 	match = of_match_node(of_rockchip_vpu_match, vpu->dev->of_node);
-	vfd = video_device_alloc();
-	if (!vfd) {
+	func = devm_kzalloc(vpu->dev, sizeof(*func), GFP_KERNEL);
+	if (!func) {
 		v4l2_err(&vpu->v4l2_dev, "Failed to allocate video device\n");
 		return -ENOMEM;
 	}
 
+	func->id = MEDIA_ENT_F_PROC_VIDEO_ENCODER;
+
+	vfd = &func->vdev;
 	vfd->fops = &rockchip_vpu_fops;
-	vfd->release = video_device_release;
+	vfd->release = video_device_release_empty;
 	vfd->lock = &vpu->vpu_mutex;
 	vfd->v4l2_dev = &vpu->v4l2_dev;
 	vfd->vfl_dir = VFL_DIR_M2M;
 	vfd->device_caps = V4L2_CAP_STREAMING | V4L2_CAP_VIDEO_M2M_MPLANE;
 	vfd->ioctl_ops = &rockchip_vpu_enc_ioctl_ops;
 	snprintf(vfd->name, sizeof(vfd->name), "%s-enc", match->compatible);
-	vpu->vfd_enc = vfd;
+
+	vpu->encoder = func;
 	video_set_drvdata(vfd, vpu);
 
 	ret = video_register_device(vfd, VFL_TYPE_GRABBER, -1);
 	if (ret) {
 		v4l2_err(&vpu->v4l2_dev, "Failed to register video device\n");
-		goto err_free_dev;
+		return ret;
 	}
+
+	ret = rockchip_attach_func(vpu, func);
+	if (ret) {
+		v4l2_err(&vpu->v4l2_dev,
+			 "Failed to attach functionality to the media device\n");
+		goto err_unreg_dev;
+	}
+
 	v4l2_info(&vpu->v4l2_dev, "registered as /dev/video%d\n", vfd->num);
 
-	function = MEDIA_ENT_F_PROC_VIDEO_ENCODER;
-	ret = v4l2_m2m_register_media_controller(vpu->m2m_dev, vfd, function);
-	if (ret) {
-		v4l2_err(&vpu->v4l2_dev, "Failed to init mem2mem media controller\n");
-		goto err_unreg_video;
-	}
 	return 0;
 
-err_unreg_video:
+err_unreg_dev:
 	video_unregister_device(vfd);
-err_free_dev:
-	video_device_release(vfd);
 	return ret;
 }
 
+static void rockchip_vpu_remove_enc_func(struct rockchip_vpu_dev *vpu)
+{
+	struct rockchip_vpu_func *func = vpu->encoder;
+
+	if (!func)
+		return;
+
+	rockchip_detach_func(func);
+	video_unregister_device(&func->vdev);
+}
+
 static int rockchip_vpu_probe(struct platform_device *pdev)
 {
 	const struct of_device_id *match;
@@ -464,7 +619,7 @@ static int rockchip_vpu_probe(struct platform_device *pdev)
 	media_device_init(&vpu->mdev);
 	vpu->v4l2_dev.mdev = &vpu->mdev;
 
-	ret = rockchip_vpu_video_device_register(vpu);
+	ret = rockchip_vpu_add_enc_func(vpu);
 	if (ret) {
 		dev_err(&pdev->dev, "Failed to register encoder\n");
 		goto err_m2m_rel;
@@ -473,15 +628,13 @@ static int rockchip_vpu_probe(struct platform_device *pdev)
 	ret = media_device_register(&vpu->mdev);
 	if (ret) {
 		v4l2_err(&vpu->v4l2_dev, "Failed to register mem2mem media device\n");
-		goto err_video_dev_unreg;
+		goto err_rm_enc_func;
 	}
+
 	return 0;
-err_video_dev_unreg:
-	if (vpu->vfd_enc) {
-		v4l2_m2m_unregister_media_controller(vpu->m2m_dev);
-		video_unregister_device(vpu->vfd_enc);
-		video_device_release(vpu->vfd_enc);
-	}
+
+err_rm_enc_func:
+	rockchip_vpu_remove_enc_func(vpu);
 err_m2m_rel:
 	media_device_cleanup(&vpu->mdev);
 	v4l2_m2m_release(vpu->m2m_dev);
@@ -501,11 +654,7 @@ static int rockchip_vpu_remove(struct platform_device *pdev)
 	v4l2_info(&vpu->v4l2_dev, "Removing %s\n", pdev->name);
 
 	media_device_unregister(&vpu->mdev);
-	if (vpu->vfd_enc) {
-		v4l2_m2m_unregister_media_controller(vpu->m2m_dev);
-		video_unregister_device(vpu->vfd_enc);
-		video_device_release(vpu->vfd_enc);
-	}
+	rockchip_vpu_remove_enc_func(vpu);
 	media_device_cleanup(&vpu->mdev);
 	v4l2_m2m_release(vpu->m2m_dev);
 	v4l2_device_unregister(&vpu->v4l2_dev);
-- 
2.20.1

[PATCH v5 06/15] rockchip/vpu: Support the Request API

[Boris Brezillon]

From: Ezequiel Garcia <ezequiel@collabora.com>

Introduce support for the Request API. Although the JPEG encoder
does not mandate using the Request API, it's perfectly possible to
use it, if the application wants to.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
--
Changes from v3:
* None

Changes from v2:
* Drop unrelated mdev.bus_info changes.
* Remove the get_ref and get_ctrl helpers as they are not used yet.
---
 .../rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c     |  6 ++++++
 .../media/rockchip/vpu/rockchip_vpu_drv.c     |  7 +++++++
 .../media/rockchip/vpu/rockchip_vpu_enc.c     | 19 +++++++++++++++++++
 3 files changed, 32 insertions(+)

diff --git a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c
index 6f9f5158d193..74823d25cd8d 100644
--- a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c
@@ -113,11 +113,15 @@ void rk3399_vpu_jpeg_enc_run(struct rockchip_vpu_ctx *ctx)
 	struct rockchip_vpu_dev *vpu = ctx->dev;
 	struct vb2_v4l2_buffer *src_buf, *dst_buf;
 	struct rockchip_vpu_jpeg_ctx jpeg_ctx;
+	struct media_request *src_req;
 	u32 reg;
 
 	src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
 	dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx);
 
+	src_req = src_buf->vb2_buf.req_obj.req;
+	v4l2_ctrl_request_setup(src_req, &ctx->ctrl_handler);
+
 	memset(&jpeg_ctx, 0, sizeof(jpeg_ctx));
 	jpeg_ctx.buffer = vb2_plane_vaddr(&dst_buf->vb2_buf, 0);
 	jpeg_ctx.width = ctx->dst_fmt.width;
@@ -153,6 +157,8 @@ void rk3399_vpu_jpeg_enc_run(struct rockchip_vpu_ctx *ctx)
 		| VEPU_REG_ENCODE_FORMAT_JPEG
 		| VEPU_REG_ENCODE_ENABLE;
 
+	v4l2_ctrl_request_complete(src_req, &ctx->ctrl_handler);
+
 	/* Kick the watchdog and start encoding */
 	schedule_delayed_work(&vpu->watchdog_work, msecs_to_jiffies(2000));
 	vepu_write(vpu, reg, VEPU_REG_ENCODE_START);
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index fa02354a0f8a..f47fbd0f9545 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -162,6 +162,7 @@ enc_queue_init(void *priv, struct vb2_queue *src_vq, struct vb2_queue *dst_vq)
 	src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
 	src_vq->lock = &ctx->dev->vpu_mutex;
 	src_vq->dev = ctx->dev->v4l2_dev.dev;
+	src_vq->supports_requests = true;
 
 	ret = vb2_queue_init(src_vq);
 	if (ret)
@@ -525,6 +526,11 @@ static void rockchip_vpu_remove_enc_func(struct rockchip_vpu_dev *vpu)
 	video_unregister_device(&func->vdev);
 }
 
+static const struct media_device_ops rockchip_m2m_media_ops = {
+	.req_validate = vb2_request_validate,
+	.req_queue = v4l2_m2m_request_queue,
+};
+
 static int rockchip_vpu_probe(struct platform_device *pdev)
 {
 	const struct of_device_id *match;
@@ -617,6 +623,7 @@ static int rockchip_vpu_probe(struct platform_device *pdev)
 	strscpy(vpu->mdev.bus_info, "platform: " DRIVER_NAME,
 		sizeof(vpu->mdev.model));
 	media_device_init(&vpu->mdev);
+	vpu->mdev.ops = &rockchip_m2m_media_ops;
 	vpu->v4l2_dev.mdev = &vpu->mdev;
 
 	ret = rockchip_vpu_add_enc_func(vpu);
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
index aa00df9a7ecb..4512e94c3f32 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
@@ -540,14 +540,33 @@ static void rockchip_vpu_stop_streaming(struct vb2_queue *q)
 			vbuf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx);
 		if (!vbuf)
 			break;
+		v4l2_ctrl_request_complete(vbuf->vb2_buf.req_obj.req,
+					   &ctx->ctrl_handler);
 		v4l2_m2m_buf_done(vbuf, VB2_BUF_STATE_ERROR);
 	}
 }
 
+static void rockchip_vpu_buf_request_complete(struct vb2_buffer *vb)
+{
+	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+
+	v4l2_ctrl_request_complete(vb->req_obj.req, &ctx->ctrl_handler);
+}
+
+static int rockchip_vpu_buf_out_validate(struct vb2_buffer *vb)
+{
+	struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+
+	vbuf->field = V4L2_FIELD_NONE;
+	return 0;
+}
+
 const struct vb2_ops rockchip_vpu_enc_queue_ops = {
 	.queue_setup = rockchip_vpu_queue_setup,
 	.buf_prepare = rockchip_vpu_buf_prepare,
 	.buf_queue = rockchip_vpu_buf_queue,
+	.buf_out_validate = rockchip_vpu_buf_out_validate,
+	.buf_request_complete = rockchip_vpu_buf_request_complete,
 	.start_streaming = rockchip_vpu_start_streaming,
 	.stop_streaming = rockchip_vpu_stop_streaming,
 	.wait_prepare = vb2_ops_wait_prepare,
-- 
2.20.1

[PATCH v5 07/15] rockchip/vpu: Rename rockchip_vpu_common.h into rockchip_vpu_v4l2.h

[Boris Brezillon]

We're about to add prototypes for the vb2/v4l2 helpers shared by the
encoder/decoder logic in this file, so let's pick a name that reflects
that (rockchip_vpu_common.h was a bit to generic).

Suggested-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes from v3:
* None

Changes from v2:
* New patch
---
 drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_jpeg_enc.c | 2 +-
 drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c | 2 +-
 drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c       | 2 +-
 drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c       | 2 +-
 .../vpu/{rockchip_vpu_common.h => rockchip_vpu_v4l2.h}      | 6 +++---
 5 files changed, 7 insertions(+), 7 deletions(-)
 rename drivers/staging/media/rockchip/vpu/{rockchip_vpu_common.h => rockchip_vpu_v4l2.h} (88%)

diff --git a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_jpeg_enc.c b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_jpeg_enc.c
index 791353ae01e7..68176e91330a 100644
--- a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_jpeg_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_jpeg_enc.c
@@ -9,7 +9,7 @@
 #include <media/v4l2-mem2mem.h>
 #include "rockchip_vpu_jpeg.h"
 #include "rockchip_vpu.h"
-#include "rockchip_vpu_common.h"
+#include "rockchip_vpu_v4l2.h"
 #include "rockchip_vpu_hw.h"
 #include "rk3288_vpu_regs.h"
 
diff --git a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c
index 74823d25cd8d..460edc5ebe4d 100644
--- a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_jpeg_enc.c
@@ -27,7 +27,7 @@
 #include <media/v4l2-mem2mem.h>
 #include "rockchip_vpu_jpeg.h"
 #include "rockchip_vpu.h"
-#include "rockchip_vpu_common.h"
+#include "rockchip_vpu_v4l2.h"
 #include "rockchip_vpu_hw.h"
 #include "rk3399_vpu_regs.h"
 
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index f47fbd0f9545..59b72245fb07 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -24,7 +24,7 @@
 #include <media/videobuf2-core.h>
 #include <media/videobuf2-vmalloc.h>
 
-#include "rockchip_vpu_common.h"
+#include "rockchip_vpu_v4l2.h"
 #include "rockchip_vpu.h"
 #include "rockchip_vpu_hw.h"
 
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
index 4512e94c3f32..d2b4225516b5 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
@@ -28,7 +28,7 @@
 
 #include "rockchip_vpu.h"
 #include "rockchip_vpu_hw.h"
-#include "rockchip_vpu_common.h"
+#include "rockchip_vpu_v4l2.h"
 
 static const struct rockchip_vpu_fmt *
 rockchip_vpu_find_format(struct rockchip_vpu_ctx *ctx, u32 fourcc)
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_common.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
similarity index 88%
rename from drivers/staging/media/rockchip/vpu/rockchip_vpu_common.h
rename to drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
index ca77668d9579..50ad40dfb4f4 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_common.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
@@ -13,8 +13,8 @@
  * Copyright (C) 2011 Samsung Electronics Co., Ltd.
  */
 
-#ifndef ROCKCHIP_VPU_COMMON_H_
-#define ROCKCHIP_VPU_COMMON_H_
+#ifndef ROCKCHIP_VPU_V4L2_H_
+#define ROCKCHIP_VPU_V4L2_H_
 
 #include "rockchip_vpu.h"
 
@@ -26,4 +26,4 @@ void rockchip_vpu_enc_reset_src_fmt(struct rockchip_vpu_dev *vpu,
 void rockchip_vpu_enc_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
 				    struct rockchip_vpu_ctx *ctx);
 
-#endif /* ROCKCHIP_VPU_COMMON_H_ */
+#endif /* ROCKCHIP_VPU_V4L2_H_ */
-- 
2.20.1

[PATCH v5 08/15] rockchip/vpu: Move encoder logic to a common place

[Boris Brezillon]

The V4L2/VB2 implementation for the encoder and decoder logic are very
similar, so let's rename rockchip_vpu_enc.c file into
rockchip_vpu_v4l2.c and remove the _enc_ part in objects/functions
exposed in rockchip_vpu_v4l2.h. We also rename the enc_queue_init()
function (in rockchip_vpu_drv.c) queue_init() since it will be used
to initialize both type of queues.

The implementation itself will be patched to support the decoding case
when decoder support is added.

Suggested-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes from v3:
* None

Changes from v2:
* New patch
---
 drivers/staging/media/rockchip/vpu/Makefile        |  2 +-
 .../staging/media/rockchip/vpu/rockchip_vpu_drv.c  | 14 +++++++-------
 .../{rockchip_vpu_enc.c => rockchip_vpu_v4l2.c}    | 14 +++++++-------
 .../staging/media/rockchip/vpu/rockchip_vpu_v4l2.h | 12 ++++++------
 4 files changed, 21 insertions(+), 21 deletions(-)
 rename drivers/staging/media/rockchip/vpu/{rockchip_vpu_enc.c => rockchip_vpu_v4l2.c} (97%)

diff --git a/drivers/staging/media/rockchip/vpu/Makefile b/drivers/staging/media/rockchip/vpu/Makefile
index e9d733bb7632..ae512173e7e4 100644
--- a/drivers/staging/media/rockchip/vpu/Makefile
+++ b/drivers/staging/media/rockchip/vpu/Makefile
@@ -2,7 +2,7 @@ obj-$(CONFIG_VIDEO_ROCKCHIP_VPU) += rockchip-vpu.o
 
 rockchip-vpu-y += \
 		rockchip_vpu_drv.o \
-		rockchip_vpu_enc.o \
+		rockchip_vpu_v4l2.o \
 		rk3288_vpu_hw.o \
 		rk3288_vpu_hw_jpeg_enc.o \
 		rk3399_vpu_hw.o \
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index 59b72245fb07..ec18578d55d7 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -140,7 +140,7 @@ static struct v4l2_m2m_ops vpu_m2m_ops = {
 };
 
 static int
-enc_queue_init(void *priv, struct vb2_queue *src_vq, struct vb2_queue *dst_vq)
+queue_init(void *priv, struct vb2_queue *src_vq, struct vb2_queue *dst_vq)
 {
 	struct rockchip_vpu_ctx *ctx = priv;
 	int ret;
@@ -148,7 +148,7 @@ enc_queue_init(void *priv, struct vb2_queue *src_vq, struct vb2_queue *dst_vq)
 	src_vq->type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
 	src_vq->io_modes = VB2_MMAP | VB2_DMABUF;
 	src_vq->drv_priv = ctx;
-	src_vq->ops = &rockchip_vpu_enc_queue_ops;
+	src_vq->ops = &rockchip_vpu_queue_ops;
 	src_vq->mem_ops = &vb2_dma_contig_memops;
 
 	/*
@@ -179,7 +179,7 @@ enc_queue_init(void *priv, struct vb2_queue *src_vq, struct vb2_queue *dst_vq)
 	dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE;
 	dst_vq->io_modes = VB2_MMAP | VB2_DMABUF;
 	dst_vq->drv_priv = ctx;
-	dst_vq->ops = &rockchip_vpu_enc_queue_ops;
+	dst_vq->ops = &rockchip_vpu_queue_ops;
 	dst_vq->mem_ops = &vb2_vmalloc_memops;
 	dst_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer);
 	dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
@@ -260,7 +260,7 @@ static int rockchip_vpu_open(struct file *filp)
 	ctx->dev = vpu;
 	if (func->id == MEDIA_ENT_F_PROC_VIDEO_ENCODER)
 		ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(vpu->m2m_dev, ctx,
-						    &enc_queue_init);
+						    queue_init);
 	else
 		ctx->fh.m2m_ctx = ERR_PTR(-ENODEV);
 	if (IS_ERR(ctx->fh.m2m_ctx)) {
@@ -273,8 +273,8 @@ static int rockchip_vpu_open(struct file *filp)
 	filp->private_data = &ctx->fh;
 	v4l2_fh_add(&ctx->fh);
 
-	rockchip_vpu_enc_reset_dst_fmt(vpu, ctx);
-	rockchip_vpu_enc_reset_src_fmt(vpu, ctx);
+	rockchip_vpu_reset_dst_fmt(vpu, ctx);
+	rockchip_vpu_reset_src_fmt(vpu, ctx);
 
 	ret = rockchip_vpu_ctrls_setup(vpu, ctx);
 	if (ret) {
@@ -487,7 +487,7 @@ static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
 	vfd->v4l2_dev = &vpu->v4l2_dev;
 	vfd->vfl_dir = VFL_DIR_M2M;
 	vfd->device_caps = V4L2_CAP_STREAMING | V4L2_CAP_VIDEO_M2M_MPLANE;
-	vfd->ioctl_ops = &rockchip_vpu_enc_ioctl_ops;
+	vfd->ioctl_ops = &rockchip_vpu_ioctl_ops;
 	snprintf(vfd->name, sizeof(vfd->name), "%s-enc", match->compatible);
 
 	vpu->encoder = func;
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
similarity index 97%
rename from drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
rename to drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
index d2b4225516b5..3e8f6256e0ed 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_enc.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
@@ -250,8 +250,8 @@ vidioc_try_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 	return 0;
 }
 
-void rockchip_vpu_enc_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
-				    struct rockchip_vpu_ctx *ctx)
+void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
+				struct rockchip_vpu_ctx *ctx)
 {
 	struct v4l2_pix_format_mplane *fmt = &ctx->dst_fmt;
 
@@ -273,8 +273,8 @@ void rockchip_vpu_enc_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
 		fmt->width * fmt->height * ctx->vpu_dst_fmt->max_depth;
 }
 
-void rockchip_vpu_enc_reset_src_fmt(struct rockchip_vpu_dev *vpu,
-				    struct rockchip_vpu_ctx *ctx)
+void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
+				struct rockchip_vpu_ctx *ctx)
 {
 	struct v4l2_pix_format_mplane *fmt = &ctx->src_fmt;
 
@@ -373,11 +373,11 @@ vidioc_s_fmt_cap_mplane(struct file *file, void *priv, struct v4l2_format *f)
 	 * the raw format again after we return, so we don't need
 	 * anything smarter.
 	 */
-	rockchip_vpu_enc_reset_src_fmt(vpu, ctx);
+	rockchip_vpu_reset_src_fmt(vpu, ctx);
 	return 0;
 }
 
-const struct v4l2_ioctl_ops rockchip_vpu_enc_ioctl_ops = {
+const struct v4l2_ioctl_ops rockchip_vpu_ioctl_ops = {
 	.vidioc_querycap = vidioc_querycap,
 	.vidioc_enum_framesizes = vidioc_enum_framesizes,
 
@@ -561,7 +561,7 @@ static int rockchip_vpu_buf_out_validate(struct vb2_buffer *vb)
 	return 0;
 }
 
-const struct vb2_ops rockchip_vpu_enc_queue_ops = {
+const struct vb2_ops rockchip_vpu_queue_ops = {
 	.queue_setup = rockchip_vpu_queue_setup,
 	.buf_prepare = rockchip_vpu_buf_prepare,
 	.buf_queue = rockchip_vpu_buf_queue,
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
index 50ad40dfb4f4..816bd3988218 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
@@ -18,12 +18,12 @@
 
 #include "rockchip_vpu.h"
 
-extern const struct v4l2_ioctl_ops rockchip_vpu_enc_ioctl_ops;
-extern const struct vb2_ops rockchip_vpu_enc_queue_ops;
+extern const struct v4l2_ioctl_ops rockchip_vpu_ioctl_ops;
+extern const struct vb2_ops rockchip_vpu_queue_ops;
 
-void rockchip_vpu_enc_reset_src_fmt(struct rockchip_vpu_dev *vpu,
-				    struct rockchip_vpu_ctx *ctx);
-void rockchip_vpu_enc_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
-				    struct rockchip_vpu_ctx *ctx);
+void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
+				struct rockchip_vpu_ctx *ctx);
+void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
+				struct rockchip_vpu_ctx *ctx);
 
 #endif /* ROCKCHIP_VPU_V4L2_H_ */
-- 
2.20.1

[PATCH v5 09/15] rockchip/vpu: Provide a helper to reset both src and dst formats

[Boris Brezillon]

When initializing a context, the core wants to reset both src and dst
formats. Right now the order doesn't matter, but if we want to have a
valid default width/height on the non-coded/raw format side (src in
case of encoders, dst in case of decoders), we need to reset those
formats in the right order: first the coded-format side, then the other,
such that width and height on the raw format side can be taken from
the coded format.

Let's provide a helper that will reset both formats and make sure this
is done in the right order.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes from v2:
* None

Changes from v2:
* New patch
---
 .../staging/media/rockchip/vpu/rockchip_vpu_drv.c  |  3 +--
 .../staging/media/rockchip/vpu/rockchip_vpu_v4l2.c | 14 ++++++++++----
 .../staging/media/rockchip/vpu/rockchip_vpu_v4l2.h |  5 +----
 3 files changed, 12 insertions(+), 10 deletions(-)

diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index ec18578d55d7..d85b88067b03 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -273,8 +273,7 @@ static int rockchip_vpu_open(struct file *filp)
 	filp->private_data = &ctx->fh;
 	v4l2_fh_add(&ctx->fh);
 
-	rockchip_vpu_reset_dst_fmt(vpu, ctx);
-	rockchip_vpu_reset_src_fmt(vpu, ctx);
+	rockchip_vpu_reset_fmts(ctx);
 
 	ret = rockchip_vpu_ctrls_setup(vpu, ctx);
 	if (ret) {
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
index 3e8f6256e0ed..e30056dc6758 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
@@ -250,8 +250,8 @@ vidioc_try_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 	return 0;
 }
 
-void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
-				struct rockchip_vpu_ctx *ctx)
+static void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
+				       struct rockchip_vpu_ctx *ctx)
 {
 	struct v4l2_pix_format_mplane *fmt = &ctx->dst_fmt;
 
@@ -273,8 +273,8 @@ void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
 		fmt->width * fmt->height * ctx->vpu_dst_fmt->max_depth;
 }
 
-void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
-				struct rockchip_vpu_ctx *ctx)
+static void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
+				       struct rockchip_vpu_ctx *ctx)
 {
 	struct v4l2_pix_format_mplane *fmt = &ctx->src_fmt;
 
@@ -294,6 +294,12 @@ void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
 			    fmt->height);
 }
 
+void rockchip_vpu_reset_fmts(struct rockchip_vpu_ctx *ctx)
+{
+	rockchip_vpu_reset_dst_fmt(ctx->dev, ctx);
+	rockchip_vpu_reset_src_fmt(ctx->dev, ctx);
+}
+
 static int
 vidioc_s_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 {
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
index 816bd3988218..493e8751d22d 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.h
@@ -21,9 +21,6 @@
 extern const struct v4l2_ioctl_ops rockchip_vpu_ioctl_ops;
 extern const struct vb2_ops rockchip_vpu_queue_ops;
 
-void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
-				struct rockchip_vpu_ctx *ctx);
-void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
-				struct rockchip_vpu_ctx *ctx);
+void rockchip_vpu_reset_fmts(struct rockchip_vpu_ctx *ctx);
 
 #endif /* ROCKCHIP_VPU_V4L2_H_ */
-- 
2.20.1

[PATCH v5 10/15] rockchip/vpu: Prepare things to support decoders

[Boris Brezillon]

The code in rockchip_vpu_v4l2 was hardcoded for encoder support.
Modify it more generic to support the decoder case so that we can
re-use the same vb2/v4l2 ops for both devices.

Co-developed-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes from v4:
* Fix s_fmt() implems WRT to format properties propagation
* s/fmt->pix/fmt->pix_mp/ in try_fmt() (Reported by Hans)

Changes from v3:
* None

Changes from v2:
* New patch
---
 .../staging/media/rockchip/vpu/rockchip_vpu.h |   6 +
 .../media/rockchip/vpu/rockchip_vpu_v4l2.c    | 551 ++++++++++--------
 2 files changed, 325 insertions(+), 232 deletions(-)

diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
index aba257c663a7..0d24fd257a2b 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
@@ -262,4 +262,10 @@ static inline u32 vepu_read(struct rockchip_vpu_dev *vpu, u32 reg)
 	return val;
 }
 
+static inline bool
+rockchip_vpu_is_encoder_ctx(const struct rockchip_vpu_ctx *ctx)
+{
+	return true;
+}
+
 #endif /* ROCKCHIP_VPU_H_ */
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
index e30056dc6758..1ab558d6492d 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
@@ -31,14 +31,23 @@
 #include "rockchip_vpu_v4l2.h"
 
 static const struct rockchip_vpu_fmt *
-rockchip_vpu_find_format(struct rockchip_vpu_ctx *ctx, u32 fourcc)
+rockchip_vpu_get_formats(const struct rockchip_vpu_ctx *ctx,
+			 unsigned int *num_fmts)
 {
-	struct rockchip_vpu_dev *dev = ctx->dev;
 	const struct rockchip_vpu_fmt *formats;
-	unsigned int num_fmts, i;
 
-	formats = dev->variant->enc_fmts;
-	num_fmts = dev->variant->num_enc_fmts;
+	formats = ctx->dev->variant->enc_fmts;
+	*num_fmts = ctx->dev->variant->num_enc_fmts;
+
+	return formats;
+}
+
+static const struct rockchip_vpu_fmt *
+rockchip_vpu_find_format(const struct rockchip_vpu_fmt *formats,
+			 unsigned int num_fmts, u32 fourcc)
+{
+	unsigned int i;
+
 	for (i = 0; i < num_fmts; i++)
 		if (formats[i].fourcc == fourcc)
 			return &formats[i];
@@ -46,14 +55,11 @@ rockchip_vpu_find_format(struct rockchip_vpu_ctx *ctx, u32 fourcc)
 }
 
 static const struct rockchip_vpu_fmt *
-rockchip_vpu_get_default_fmt(struct rockchip_vpu_ctx *ctx, bool bitstream)
+rockchip_vpu_get_default_fmt(const struct rockchip_vpu_fmt *formats,
+			     unsigned int num_fmts, bool bitstream)
 {
-	struct rockchip_vpu_dev *dev = ctx->dev;
-	const struct rockchip_vpu_fmt *formats;
-	unsigned int num_fmts, i;
+	unsigned int i;
 
-	formats = dev->variant->enc_fmts;
-	num_fmts = dev->variant->num_enc_fmts;
 	for (i = 0; i < num_fmts; i++) {
 		if (bitstream == (formats[i].codec_mode != RK_VPU_MODE_NONE))
 			return &formats[i];
@@ -78,7 +84,8 @@ static int vidioc_enum_framesizes(struct file *file, void *priv,
 				  struct v4l2_frmsizeenum *fsize)
 {
 	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
-	const struct rockchip_vpu_fmt *fmt;
+	const struct rockchip_vpu_fmt *formats, *fmt;
+	unsigned int num_fmts;
 
 	if (fsize->index != 0) {
 		vpu_debug(0, "invalid frame size index (expected 0, got %d)\n",
@@ -86,7 +93,8 @@ static int vidioc_enum_framesizes(struct file *file, void *priv,
 		return -EINVAL;
 	}
 
-	fmt = rockchip_vpu_find_format(ctx, fsize->pixel_format);
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
+	fmt = rockchip_vpu_find_format(formats, num_fmts, fsize->pixel_format);
 	if (!fmt) {
 		vpu_debug(0, "unsupported bitstream format (%08x)\n",
 			  fsize->pixel_format);
@@ -103,19 +111,32 @@ static int vidioc_enum_framesizes(struct file *file, void *priv,
 	return 0;
 }
 
-static int vidioc_enum_fmt_vid_cap_mplane(struct file *file, void *priv,
-					  struct v4l2_fmtdesc *f)
-{
-	struct rockchip_vpu_dev *dev = video_drvdata(file);
-	const struct rockchip_vpu_fmt *fmt;
-	const struct rockchip_vpu_fmt *formats;
-	int num_fmts, i, j = 0;
+static int vidioc_enum_fmt(struct file *file, void *priv,
+			   struct v4l2_fmtdesc *f, bool capture)
 
-	formats = dev->variant->enc_fmts;
-	num_fmts = dev->variant->num_enc_fmts;
+{
+	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
+	const struct rockchip_vpu_fmt *fmt, *formats;
+	unsigned int num_fmts, i, j = 0;
+	bool skip_mode_none;
+
+	/*
+	 * When dealing with an encoder:
+	 *  - on the capture side we want to filter out all MODE_NONE formats.
+	 *  - on the output side we want to filter out all formats that are
+	 *    not MODE_NONE.
+	 * When dealing with a decoder:
+	 *  - on the capture side we want to filter out all formats that are
+	 *    not MODE_NONE.
+	 *  - on the output side we want to filter out all MODE_NONE formats.
+	 */
+	skip_mode_none = capture == rockchip_vpu_is_encoder_ctx(ctx);
+
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
 	for (i = 0; i < num_fmts; i++) {
-		/* Skip uncompressed formats */
-		if (formats[i].codec_mode == RK_VPU_MODE_NONE)
+		bool mode_none = formats[i].codec_mode == RK_VPU_MODE_NONE;
+
+		if (skip_mode_none == mode_none)
 			continue;
 		if (j == f->index) {
 			fmt = &formats[i];
@@ -127,27 +148,16 @@ static int vidioc_enum_fmt_vid_cap_mplane(struct file *file, void *priv,
 	return -EINVAL;
 }
 
+static int vidioc_enum_fmt_vid_cap_mplane(struct file *file, void *priv,
+					  struct v4l2_fmtdesc *f)
+{
+	return vidioc_enum_fmt(file, priv, f, true);
+}
+
 static int vidioc_enum_fmt_vid_out_mplane(struct file *file, void *priv,
 					  struct v4l2_fmtdesc *f)
 {
-	struct rockchip_vpu_dev *dev = video_drvdata(file);
-	const struct rockchip_vpu_fmt *formats;
-	const struct rockchip_vpu_fmt *fmt;
-	int num_fmts, i, j = 0;
-
-	formats = dev->variant->enc_fmts;
-	num_fmts = dev->variant->num_enc_fmts;
-	for (i = 0; i < num_fmts; i++) {
-		if (formats[i].codec_mode != RK_VPU_MODE_NONE)
-			continue;
-		if (j == f->index) {
-			fmt = &formats[i];
-			f->pixelformat = fmt->fourcc;
-			return 0;
-		}
-		++j;
-	}
-	return -EINVAL;
+	return vidioc_enum_fmt(file, priv, f, false);
 }
 
 static int vidioc_g_fmt_out_mplane(struct file *file, void *priv,
@@ -176,128 +186,149 @@ static int vidioc_g_fmt_cap_mplane(struct file *file, void *priv,
 	return 0;
 }
 
-static int
-vidioc_try_fmt_cap_mplane(struct file *file, void *priv, struct v4l2_format *f)
+static int vidioc_try_fmt(struct file *file, void *priv, struct v4l2_format *f,
+			  bool capture)
 {
 	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
 	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
-	const struct rockchip_vpu_fmt *fmt;
+	const struct rockchip_vpu_fmt *formats, *fmt, *vpu_fmt;
+	unsigned int num_fmts;
+	bool coded;
 
-	vpu_debug(4, "%c%c%c%c\n",
+	coded = capture == rockchip_vpu_is_encoder_ctx(ctx);
+
+	vpu_debug(4, "trying format %c%c%c%c\n",
 		  (pix_mp->pixelformat & 0x7f),
 		  (pix_mp->pixelformat >> 8) & 0x7f,
 		  (pix_mp->pixelformat >> 16) & 0x7f,
 		  (pix_mp->pixelformat >> 24) & 0x7f);
 
-	fmt = rockchip_vpu_find_format(ctx, pix_mp->pixelformat);
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
+	fmt = rockchip_vpu_find_format(formats, num_fmts, pix_mp->pixelformat);
 	if (!fmt) {
-		fmt = rockchip_vpu_get_default_fmt(ctx, true);
-		f->fmt.pix.pixelformat = fmt->fourcc;
+		fmt = rockchip_vpu_get_default_fmt(formats, num_fmts, coded);
+		f->fmt.pix_mp.pixelformat = fmt->fourcc;
+	}
+
+	if (coded) {
+		pix_mp->num_planes = 1;
+		vpu_fmt = fmt;
+	} else if (rockchip_vpu_is_encoder_ctx(ctx)) {
+		vpu_fmt = ctx->vpu_dst_fmt;
+	} else {
+		vpu_fmt = ctx->vpu_src_fmt;
+		/*
+		 * Width/height on the CAPTURE end of a decoder are ignored and
+		 * replaced by the OUTPUT ones.
+		 */
+		pix_mp->width = ctx->src_fmt.width;
+		pix_mp->height = ctx->src_fmt.height;
 	}
 
-	pix_mp->num_planes = 1;
 	pix_mp->field = V4L2_FIELD_NONE;
 
 	v4l2_apply_frmsize_constraints(&pix_mp->width, &pix_mp->height,
-				       &fmt->frmsize);
+				       &vpu_fmt->frmsize);
 
-	/*
-	 * For compressed formats the application can specify
-	 * sizeimage. If the application passes a zero sizeimage,
-	 * let's default to the maximum frame size.
-	 */
-	if (!pix_mp->plane_fmt[0].sizeimage)
+	if (!coded) {
+		/* Fill remaining fields */
+		v4l2_fill_pixfmt_mp(pix_mp, fmt->fourcc, pix_mp->width,
+				    pix_mp->height);
+	} else if (!pix_mp->plane_fmt[0].sizeimage) {
+		/*
+		 * For coded formats the application can specify
+		 * sizeimage. If the application passes a zero sizeimage,
+		 * let's default to the maximum frame size.
+		 */
 		pix_mp->plane_fmt[0].sizeimage = fmt->header_size +
 			pix_mp->width * pix_mp->height * fmt->max_depth;
-	memset(pix_mp->plane_fmt[0].reserved, 0,
-	       sizeof(pix_mp->plane_fmt[0].reserved));
+	}
+
 	return 0;
 }
 
-static int
-vidioc_try_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
+static int vidioc_try_fmt_cap_mplane(struct file *file, void *priv,
+				     struct v4l2_format *f)
 {
-	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
-	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
-	const struct rockchip_vpu_fmt *fmt;
-	int i;
-
-	vpu_debug(4, "%c%c%c%c\n",
-		  (pix_mp->pixelformat & 0x7f),
-		  (pix_mp->pixelformat >> 8) & 0x7f,
-		  (pix_mp->pixelformat >> 16) & 0x7f,
-		  (pix_mp->pixelformat >> 24) & 0x7f);
-
-	fmt = rockchip_vpu_find_format(ctx, pix_mp->pixelformat);
-	if (!fmt) {
-		fmt = rockchip_vpu_get_default_fmt(ctx, false);
-		f->fmt.pix.pixelformat = fmt->fourcc;
-	}
-
-	pix_mp->field = V4L2_FIELD_NONE;
-
-	v4l2_apply_frmsize_constraints(&pix_mp->width, &pix_mp->height,
-				       &ctx->vpu_dst_fmt->frmsize);
-
-	/* Fill remaining fields */
-	v4l2_fill_pixfmt_mp(pix_mp, fmt->fourcc, pix_mp->width,
-			    pix_mp->height);
-
-	for (i = 0; i < pix_mp->num_planes; i++) {
-		memset(pix_mp->plane_fmt[i].reserved, 0,
-		       sizeof(pix_mp->plane_fmt[i].reserved));
-	}
-	return 0;
+	return vidioc_try_fmt(file, priv, f, true);
 }
 
-static void rockchip_vpu_reset_dst_fmt(struct rockchip_vpu_dev *vpu,
-				       struct rockchip_vpu_ctx *ctx)
+static int vidioc_try_fmt_out_mplane(struct file *file, void *priv,
+				     struct v4l2_format *f)
 {
-	struct v4l2_pix_format_mplane *fmt = &ctx->dst_fmt;
-
-	ctx->vpu_dst_fmt = rockchip_vpu_get_default_fmt(ctx, true);
+	return vidioc_try_fmt(file, priv, f, false);
+}
 
+static void
+rockchip_vpu_reset_fmt(struct v4l2_pix_format_mplane *fmt,
+		       const struct rockchip_vpu_fmt *vpu_fmt)
+{
 	memset(fmt, 0, sizeof(*fmt));
 
+	fmt->pixelformat = vpu_fmt->fourcc;
+	fmt->field = V4L2_FIELD_NONE;
+	fmt->colorspace = V4L2_COLORSPACE_JPEG,
+	fmt->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT;
+	fmt->quantization = V4L2_QUANTIZATION_DEFAULT;
+	fmt->xfer_func = V4L2_XFER_FUNC_DEFAULT;
+}
+
+static void
+rockchip_vpu_reset_encoded_fmt(struct rockchip_vpu_ctx *ctx)
+{
+	const struct rockchip_vpu_fmt *vpu_fmt, *formats;
+	struct v4l2_pix_format_mplane *fmt;
+	unsigned int num_fmts;
+
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
+	vpu_fmt = rockchip_vpu_get_default_fmt(formats, num_fmts, true);
+
+	if (rockchip_vpu_is_encoder_ctx(ctx)) {
+		ctx->vpu_dst_fmt = vpu_fmt;
+		fmt = &ctx->dst_fmt;
+	} else {
+		ctx->vpu_src_fmt = vpu_fmt;
+		fmt = &ctx->src_fmt;
+	}
+
+	rockchip_vpu_reset_fmt(fmt, vpu_fmt);
 	fmt->num_planes = 1;
-	v4l2_apply_frmsize_constraints(&fmt->width, &fmt->height,
-				       &ctx->vpu_dst_fmt->frmsize);
-	fmt->pixelformat = ctx->vpu_dst_fmt->fourcc;
-	fmt->field = V4L2_FIELD_NONE;
-	fmt->colorspace = V4L2_COLORSPACE_JPEG,
-	fmt->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT;
-	fmt->quantization = V4L2_QUANTIZATION_DEFAULT;
-	fmt->xfer_func = V4L2_XFER_FUNC_DEFAULT;
-
-	fmt->plane_fmt[0].sizeimage = ctx->vpu_dst_fmt->header_size +
-		fmt->width * fmt->height * ctx->vpu_dst_fmt->max_depth;
+	fmt->width = vpu_fmt->frmsize.min_width;
+	fmt->height = vpu_fmt->frmsize.min_height;
+	fmt->plane_fmt[0].sizeimage = vpu_fmt->header_size +
+				fmt->width * fmt->height * vpu_fmt->max_depth;
 }
 
-static void rockchip_vpu_reset_src_fmt(struct rockchip_vpu_dev *vpu,
-				       struct rockchip_vpu_ctx *ctx)
+static void
+rockchip_vpu_reset_raw_fmt(struct rockchip_vpu_ctx *ctx)
 {
-	struct v4l2_pix_format_mplane *fmt = &ctx->src_fmt;
+	const struct rockchip_vpu_fmt *raw_vpu_fmt, *formats;
+	struct v4l2_pix_format_mplane *raw_fmt, *encoded_fmt;
+	unsigned int num_fmts;
 
-	ctx->vpu_src_fmt = rockchip_vpu_get_default_fmt(ctx, false);
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
+	raw_vpu_fmt = rockchip_vpu_get_default_fmt(formats, num_fmts, false);
 
-	memset(fmt, 0, sizeof(*fmt));
+	if (rockchip_vpu_is_encoder_ctx(ctx)) {
+		ctx->vpu_src_fmt = raw_vpu_fmt;
+		raw_fmt = &ctx->src_fmt;
+		encoded_fmt = &ctx->dst_fmt;
+	} else {
+		ctx->vpu_dst_fmt = raw_vpu_fmt;
+		raw_fmt = &ctx->dst_fmt;
+		encoded_fmt = &ctx->src_fmt;
+	}
 
-	v4l2_apply_frmsize_constraints(&fmt->width, &fmt->height,
-				       &ctx->vpu_src_fmt->frmsize);
-	fmt->field = V4L2_FIELD_NONE;
-	fmt->colorspace = V4L2_COLORSPACE_JPEG,
-	fmt->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT;
-	fmt->quantization = V4L2_QUANTIZATION_DEFAULT;
-	fmt->xfer_func = V4L2_XFER_FUNC_DEFAULT;
-
-	v4l2_fill_pixfmt_mp(fmt, ctx->vpu_src_fmt->fourcc, fmt->width,
-			    fmt->height);
+	rockchip_vpu_reset_fmt(raw_fmt, raw_vpu_fmt);
+	v4l2_fill_pixfmt_mp(raw_fmt, raw_vpu_fmt->fourcc,
+			    encoded_fmt->width,
+			    encoded_fmt->height);
 }
 
 void rockchip_vpu_reset_fmts(struct rockchip_vpu_ctx *ctx)
 {
-	rockchip_vpu_reset_dst_fmt(ctx->dev, ctx);
-	rockchip_vpu_reset_src_fmt(ctx->dev, ctx);
+	rockchip_vpu_reset_encoded_fmt(ctx);
+	rockchip_vpu_reset_raw_fmt(ctx);
 }
 
 static int
@@ -305,81 +336,126 @@ vidioc_s_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 {
 	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
 	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
+	const struct rockchip_vpu_fmt *formats;
+	unsigned int num_fmts;
 	struct vb2_queue *vq;
 	int ret;
 
-	/* Change not allowed if queue is streaming. */
+	/* Change not allowed if queue is busy. */
 	vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx, f->type);
-	if (vb2_is_streaming(vq))
+	if (vb2_is_busy(vq))
 		return -EBUSY;
 
+	if (!rockchip_vpu_is_encoder_ctx(ctx)) {
+		struct vb2_queue *peer_vq;
+
+		/*
+		 * Since format change on the OUTPUT queue will reset
+		 * the CAPTURE queue, we can't allow doing so
+		 * when the CAPTURE queue has buffers allocated.
+		 */
+		peer_vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx,
+					  V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+		if (vb2_is_busy(peer_vq))
+			return -EBUSY;
+	}
+
 	ret = vidioc_try_fmt_out_mplane(file, priv, f);
 	if (ret)
 		return ret;
 
-	ctx->vpu_src_fmt = rockchip_vpu_find_format(ctx, pix_mp->pixelformat);
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
+	ctx->vpu_src_fmt = rockchip_vpu_find_format(formats, num_fmts,
+						    pix_mp->pixelformat);
 	ctx->src_fmt = *pix_mp;
 
-	/* Propagate to the CAPTURE format */
-	ctx->dst_fmt.colorspace = pix_mp->colorspace;
-	ctx->dst_fmt.ycbcr_enc = pix_mp->ycbcr_enc;
-	ctx->dst_fmt.xfer_func = pix_mp->xfer_func;
-	ctx->dst_fmt.quantization = pix_mp->quantization;
-	ctx->dst_fmt.width = pix_mp->width;
-	ctx->dst_fmt.height = pix_mp->height;
-
-	vpu_debug(0, "OUTPUT codec mode: %d\n", ctx->vpu_src_fmt->codec_mode);
-	vpu_debug(0, "fmt - w: %d, h: %d\n",
-		  pix_mp->width, pix_mp->height);
-	return 0;
-}
-
-static int
-vidioc_s_fmt_cap_mplane(struct file *file, void *priv, struct v4l2_format *f)
-{
-	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
-	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
-	struct rockchip_vpu_dev *vpu = ctx->dev;
-	struct vb2_queue *vq, *peer_vq;
-	int ret;
-
-	/* Change not allowed if queue is streaming. */
-	vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx, f->type);
-	if (vb2_is_streaming(vq))
-		return -EBUSY;
-
-	/*
-	 * Since format change on the CAPTURE queue will reset
-	 * the OUTPUT queue, we can't allow doing so
-	 * when the OUTPUT queue has buffers allocated.
-	 */
-	peer_vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx,
-				  V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE);
-	if (vb2_is_busy(peer_vq) &&
-	    (pix_mp->pixelformat != ctx->dst_fmt.pixelformat ||
-	     pix_mp->height != ctx->dst_fmt.height ||
-	     pix_mp->width != ctx->dst_fmt.width))
-		return -EBUSY;
-
-	ret = vidioc_try_fmt_cap_mplane(file, priv, f);
-	if (ret)
-		return ret;
-
-	ctx->vpu_dst_fmt = rockchip_vpu_find_format(ctx, pix_mp->pixelformat);
-	ctx->dst_fmt = *pix_mp;
-
-	vpu_debug(0, "CAPTURE codec mode: %d\n", ctx->vpu_dst_fmt->codec_mode);
-	vpu_debug(0, "fmt - w: %d, h: %d\n",
-		  pix_mp->width, pix_mp->height);
-
 	/*
 	 * Current raw format might have become invalid with newly
 	 * selected codec, so reset it to default just to be safe and
 	 * keep internal driver state sane. User is mandated to set
 	 * the raw format again after we return, so we don't need
 	 * anything smarter.
+	 * Note that rockchip_vpu_reset_raw_fmt() also propagates size
+	 * changes to the raw format.
 	 */
-	rockchip_vpu_reset_src_fmt(vpu, ctx);
+	if (!rockchip_vpu_is_encoder_ctx(ctx))
+		rockchip_vpu_reset_raw_fmt(ctx);
+
+	/* Colorimetry information are always propagated. */
+	ctx->dst_fmt.colorspace = pix_mp->colorspace;
+	ctx->dst_fmt.ycbcr_enc = pix_mp->ycbcr_enc;
+	ctx->dst_fmt.xfer_func = pix_mp->xfer_func;
+	ctx->dst_fmt.quantization = pix_mp->quantization;
+
+	vpu_debug(0, "OUTPUT codec mode: %d\n", ctx->vpu_src_fmt->codec_mode);
+	vpu_debug(0, "fmt - w: %d, h: %d\n",
+		  pix_mp->width, pix_mp->height);
+	return 0;
+}
+
+static int vidioc_s_fmt_cap_mplane(struct file *file, void *priv,
+				   struct v4l2_format *f)
+{
+	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
+	struct rockchip_vpu_ctx *ctx = fh_to_ctx(priv);
+	const struct rockchip_vpu_fmt *formats;
+	struct vb2_queue *vq;
+	unsigned int num_fmts;
+	int ret;
+
+	/* Change not allowed if queue is busy. */
+	vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx, f->type);
+	if (vb2_is_busy(vq))
+		return -EBUSY;
+
+	if (rockchip_vpu_is_encoder_ctx(ctx)) {
+		struct vb2_queue *peer_vq;
+
+		/*
+		 * Since format change on the CAPTURE queue will reset
+		 * the OUTPUT queue, we can't allow doing so
+		 * when the OUTPUT queue has buffers allocated.
+		 */
+		peer_vq = v4l2_m2m_get_vq(ctx->fh.m2m_ctx,
+					  V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE);
+		if (vb2_is_busy(peer_vq) &&
+		    (pix_mp->pixelformat != ctx->dst_fmt.pixelformat ||
+		     pix_mp->height != ctx->dst_fmt.height ||
+		     pix_mp->width != ctx->dst_fmt.width))
+			return -EBUSY;
+	}
+
+	ret = vidioc_try_fmt_cap_mplane(file, priv, f);
+	if (ret)
+		return ret;
+
+	formats = rockchip_vpu_get_formats(ctx, &num_fmts);
+	ctx->vpu_dst_fmt = rockchip_vpu_find_format(formats, num_fmts,
+						    pix_mp->pixelformat);
+	ctx->dst_fmt = *pix_mp;
+
+	/*
+	 * Current raw format might have become invalid with newly
+	 * selected codec, so reset it to default just to be safe and
+	 * keep internal driver state sane. User is mandated to set
+	 * the raw format again after we return, so we don't need
+	 * anything smarter.
+	 * Note that rockchip_vpu_reset_raw_fmt() also propagates size
+	 * changes to the raw format.
+	 */
+	if (rockchip_vpu_is_encoder_ctx(ctx))
+		rockchip_vpu_reset_raw_fmt(ctx);
+
+	/* Colorimetry information are always propagated. */
+	ctx->src_fmt.colorspace = pix_mp->colorspace;
+	ctx->src_fmt.ycbcr_enc = pix_mp->ycbcr_enc;
+	ctx->src_fmt.xfer_func = pix_mp->xfer_func;
+	ctx->src_fmt.quantization = pix_mp->quantization;
+
+	vpu_debug(0, "CAPTURE codec mode: %d\n", ctx->vpu_dst_fmt->codec_mode);
+	vpu_debug(0, "fmt - w: %d, h: %d\n",
+		  pix_mp->width, pix_mp->height);
+
 	return 0;
 }
 
@@ -449,48 +525,37 @@ rockchip_vpu_queue_setup(struct vb2_queue *vq,
 	return 0;
 }
 
-static int rockchip_vpu_buf_prepare(struct vb2_buffer *vb)
+static int
+rockchip_vpu_buf_plane_check(struct vb2_buffer *vb,
+			     const struct rockchip_vpu_fmt *vpu_fmt,
+			     struct v4l2_pix_format_mplane *pixfmt)
 {
-	struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
-	struct vb2_queue *vq = vb->vb2_queue;
-	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(vq);
-	struct v4l2_pix_format_mplane *pixfmt;
 	unsigned int sz;
-	int ret = 0;
 	int i;
 
-	switch (vq->type) {
-	case V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE:
-		pixfmt = &ctx->dst_fmt;
-		break;
-	case V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE:
-		pixfmt = &ctx->src_fmt;
-
-		if (vbuf->field == V4L2_FIELD_ANY)
-			vbuf->field = V4L2_FIELD_NONE;
-		if (vbuf->field != V4L2_FIELD_NONE) {
-			vpu_debug(4, "field %d not supported\n",
-				  vbuf->field);
-			return -EINVAL;
-		}
-		break;
-	default:
-		vpu_err("invalid queue type: %d\n", vq->type);
-		return -EINVAL;
-	}
-
 	for (i = 0; i < pixfmt->num_planes; ++i) {
 		sz = pixfmt->plane_fmt[i].sizeimage;
 		vpu_debug(4, "plane %d size: %ld, sizeimage: %u\n",
 			  i, vb2_plane_size(vb, i), sz);
 		if (vb2_plane_size(vb, i) < sz) {
-			vpu_err("plane %d is too small\n", i);
-			ret = -EINVAL;
-			break;
+			vpu_err("plane %d is too small for output\n", i);
+			return -EINVAL;
 		}
 	}
+	return 0;
+}
 
-	return ret;
+static int rockchip_vpu_buf_prepare(struct vb2_buffer *vb)
+{
+	struct vb2_queue *vq = vb->vb2_queue;
+	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(vq);
+
+	if (V4L2_TYPE_IS_OUTPUT(vq->type))
+		return rockchip_vpu_buf_plane_check(vb, ctx->vpu_src_fmt,
+						    &ctx->src_fmt);
+
+	return rockchip_vpu_buf_plane_check(vb, ctx->vpu_dst_fmt,
+					    &ctx->dst_fmt);
 }
 
 static void rockchip_vpu_buf_queue(struct vb2_buffer *vb)
@@ -501,10 +566,17 @@ static void rockchip_vpu_buf_queue(struct vb2_buffer *vb)
 	v4l2_m2m_buf_queue(ctx->fh.m2m_ctx, vbuf);
 }
 
-static int rockchip_vpu_start_streaming(struct vb2_queue *q, unsigned int count)
+static bool rockchip_vpu_vq_is_coded(struct vb2_queue *q)
+{
+	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(q);
+
+	return rockchip_vpu_is_encoder_ctx(ctx) != V4L2_TYPE_IS_OUTPUT(q->type);
+}
+
+static int rockchip_vpu_start_streaming(struct vb2_queue *q,
+					unsigned int count)
 {
 	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(q);
-	enum rockchip_vpu_codec_mode codec_mode;
 	int ret = 0;
 
 	if (V4L2_TYPE_IS_OUTPUT(q->type))
@@ -512,38 +584,33 @@ static int rockchip_vpu_start_streaming(struct vb2_queue *q, unsigned int count)
 	else
 		ctx->sequence_cap = 0;
 
-	/* Set codec_ops for the chosen destination format */
-	codec_mode = ctx->vpu_dst_fmt->codec_mode;
+	if (rockchip_vpu_vq_is_coded(q)) {
+		enum rockchip_vpu_codec_mode codec_mode;
 
-	vpu_debug(4, "Codec mode = %d\n", codec_mode);
-	ctx->codec_ops = &ctx->dev->variant->codec_ops[codec_mode];
+		if (V4L2_TYPE_IS_OUTPUT(q->type))
+			codec_mode = ctx->vpu_src_fmt->codec_mode;
+		else
+			codec_mode = ctx->vpu_dst_fmt->codec_mode;
 
-	if (!V4L2_TYPE_IS_OUTPUT(q->type))
+		vpu_debug(4, "Codec mode = %d\n", codec_mode);
+		ctx->codec_ops = &ctx->dev->variant->codec_ops[codec_mode];
 		if (ctx->codec_ops && ctx->codec_ops->init)
 			ret = ctx->codec_ops->init(ctx);
+	}
+
 	return ret;
 }
 
-static void rockchip_vpu_stop_streaming(struct vb2_queue *q)
+static void
+rockchip_vpu_return_bufs(struct vb2_queue *q,
+			 struct vb2_v4l2_buffer *(*buf_remove)(struct v4l2_m2m_ctx *))
 {
 	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(q);
 
-	if (!V4L2_TYPE_IS_OUTPUT(q->type))
-		if (ctx->codec_ops && ctx->codec_ops->exit)
-			ctx->codec_ops->exit(ctx);
-
-	/*
-	 * The mem2mem framework calls v4l2_m2m_cancel_job before
-	 * .stop_streaming, so there isn't any job running and
-	 * it is safe to return all the buffers.
-	 */
 	for (;;) {
 		struct vb2_v4l2_buffer *vbuf;
 
-		if (V4L2_TYPE_IS_OUTPUT(q->type))
-			vbuf = v4l2_m2m_src_buf_remove(ctx->fh.m2m_ctx);
-		else
-			vbuf = v4l2_m2m_dst_buf_remove(ctx->fh.m2m_ctx);
+		vbuf = buf_remove(ctx->fh.m2m_ctx);
 		if (!vbuf)
 			break;
 		v4l2_ctrl_request_complete(vbuf->vb2_buf.req_obj.req,
@@ -552,6 +619,26 @@ static void rockchip_vpu_stop_streaming(struct vb2_queue *q)
 	}
 }
 
+static void rockchip_vpu_stop_streaming(struct vb2_queue *q)
+{
+	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(q);
+
+	if (rockchip_vpu_vq_is_coded(q)) {
+		if (ctx->codec_ops && ctx->codec_ops->exit)
+			ctx->codec_ops->exit(ctx);
+	}
+
+	/*
+	 * The mem2mem framework calls v4l2_m2m_cancel_job before
+	 * .stop_streaming, so there isn't any job running and
+	 * it is safe to return all the buffers.
+	 */
+	if (V4L2_TYPE_IS_OUTPUT(q->type))
+		rockchip_vpu_return_bufs(q, v4l2_m2m_src_buf_remove);
+	else
+		rockchip_vpu_return_bufs(q, v4l2_m2m_dst_buf_remove);
+}
+
 static void rockchip_vpu_buf_request_complete(struct vb2_buffer *vb)
 {
 	struct rockchip_vpu_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
-- 
2.20.1

[PATCH v5 11/15] rockchip/vpu: Add decoder boilerplate

[Boris Brezillon]

From: Ezequiel Garcia <ezequiel@collabora.com>

This commit adds the needed boilerplate code to support the VPU
in decoding operation. Two v4l2 interfaces are exposed, one for
encoding and one for decoding, but a single m2m device is shared
by them, so jobs are properly serialized.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
--
Changes from v4:
* Remove vpu_common.h inclusion
* Set ->requires_requests from s_fmt() instead of fixing it at queue
  creation time (suggested by Hans)

Changes from v3:
* Update things according to changes done in the patch introducing
  custom media controller support (Boris)
* Move the rk3399 vdpu irq definition in the commit introducing
  MPGEG2 support on rk3399
* Fix media controller deregistration code (Jonas)

Changes from v2:
* Use the common vb2/v4l2 implementation
* Use strscpy instead of strlcpy.
* Abstract vidioc v4l2 api implementations into generic code, creating
  helpers that can be used by the encoder and the decoder.
* Only prevent S_FMT on the coded format queue, if the peer queue has buffers allocated.
* Refactor the code, adding a buf_finish callback to rockchip_vpu_ctx.
  With this change, is_enc field is not needed.
* Separate OUTPUT and CAPTURE queue ops (vb2_ops), and create common helpers that
  can be used by both.
* Pass a no kernel mapping attribute on both ends of the decoder.
---
 drivers/staging/media/rockchip/vpu/Kconfig    |   1 +
 .../staging/media/rockchip/vpu/rockchip_vpu.h |  42 ++++-
 .../media/rockchip/vpu/rockchip_vpu_drv.c     | 169 ++++++++++++++----
 .../media/rockchip/vpu/rockchip_vpu_v4l2.c    |  29 ++-
 4 files changed, 203 insertions(+), 38 deletions(-)

diff --git a/drivers/staging/media/rockchip/vpu/Kconfig b/drivers/staging/media/rockchip/vpu/Kconfig
index 9a6fc1378242..f645c19f10d1 100644
--- a/drivers/staging/media/rockchip/vpu/Kconfig
+++ b/drivers/staging/media/rockchip/vpu/Kconfig
@@ -2,6 +2,7 @@ config VIDEO_ROCKCHIP_VPU
 	tristate "Rockchip VPU driver"
 	depends on ARCH_ROCKCHIP || COMPILE_TEST
 	depends on VIDEO_DEV && VIDEO_V4L2 && MEDIA_CONTROLLER
+	depends on MEDIA_CONTROLLER_REQUEST_API
 	select VIDEOBUF2_DMA_CONTIG
 	select VIDEOBUF2_VMALLOC
 	select V4L2_MEM2MEM_DEV
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
index 0d24fd257a2b..3d64f3e95c9b 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
@@ -40,23 +40,31 @@ struct rockchip_vpu_codec_ops;
  * struct rockchip_vpu_variant - information about VPU hardware variant
  *
  * @enc_offset:			Offset from VPU base to encoder registers.
+ * @dec_offset:			Offset from VPU base to decoder registers.
  * @enc_fmts:			Encoder formats.
  * @num_enc_fmts:		Number of encoder formats.
+ * @dec_fmts:			Decoder formats.
+ * @num_dec_fmts:		Number of decoder formats.
  * @codec:			Supported codecs
  * @codec_ops:			Codec ops.
  * @init:			Initialize hardware.
  * @vepu_irq:			encoder interrupt handler
+ * @vdpu_irq:			decoder interrupt handler
  * @clk_names:			array of clock names
  * @num_clocks:			number of clocks in the array
  */
 struct rockchip_vpu_variant {
 	unsigned int enc_offset;
+	unsigned int dec_offset;
 	const struct rockchip_vpu_fmt *enc_fmts;
 	unsigned int num_enc_fmts;
+	const struct rockchip_vpu_fmt *dec_fmts;
+	unsigned int num_dec_fmts;
 	unsigned int codec;
 	const struct rockchip_vpu_codec_ops *codec_ops;
 	int (*init)(struct rockchip_vpu_dev *vpu);
 	irqreturn_t (*vepu_irq)(int irq, void *priv);
+	irqreturn_t (*vdpu_irq)(int irq, void *priv);
 	const char *clk_names[ROCKCHIP_VPU_MAX_CLOCKS];
 	int num_clocks;
 };
@@ -112,12 +120,14 @@ rockchip_vpu_vdev_to_func(struct video_device *vdev)
  * @m2m_dev:		mem2mem device associated to this device.
  * @mdev:		media device associated to this device.
  * @encoder:		encoder functionality.
+ * @decoder:		decoder functionality.
  * @pdev:		Pointer to VPU platform device.
  * @dev:		Pointer to device for convenient logging using
  *			dev_ macros.
  * @clocks:		Array of clock handles.
  * @base:		Mapped address of VPU registers.
  * @enc_base:		Mapped address of VPU encoder register for convenience.
+ * @dec_base:		Mapped address of VPU decoder register for convenience.
  * @vpu_mutex:		Mutex to synchronize V4L2 calls.
  * @irqlock:		Spinlock to synchronize access to data structures
  *			shared with interrupt handlers.
@@ -129,11 +139,13 @@ struct rockchip_vpu_dev {
 	struct v4l2_m2m_dev *m2m_dev;
 	struct media_device mdev;
 	struct rockchip_vpu_func *encoder;
+	struct rockchip_vpu_func *decoder;
 	struct platform_device *pdev;
 	struct device *dev;
 	struct clk_bulk_data clocks[ROCKCHIP_VPU_MAX_CLOCKS];
 	void __iomem *base;
 	void __iomem *enc_base;
+	void __iomem *dec_base;
 
 	struct mutex vpu_mutex;	/* video_device lock */
 	spinlock_t irqlock;
@@ -158,6 +170,9 @@ struct rockchip_vpu_dev {
  * @ctrl_handler:	Control handler used to register controls.
  * @jpeg_quality:	User-specified JPEG compression quality.
  *
+ * @buf_finish:		Buffer finish. This depends on encoder or decoder
+ *			context, and it's called right before
+ *			calling v4l2_m2m_job_finish.
  * @codec_ops:		Set of operations related to codec mode.
  * @jpeg_enc:		JPEG-encoding context.
  */
@@ -176,6 +191,10 @@ struct rockchip_vpu_ctx {
 	struct v4l2_ctrl_handler ctrl_handler;
 	int jpeg_quality;
 
+	int (*buf_finish)(struct rockchip_vpu_ctx *ctx,
+			  struct vb2_buffer *buf,
+			  unsigned int bytesused);
+
 	const struct rockchip_vpu_codec_ops *codec_ops;
 
 	/* Specific for particular codec modes. */
@@ -262,10 +281,27 @@ static inline u32 vepu_read(struct rockchip_vpu_dev *vpu, u32 reg)
 	return val;
 }
 
-static inline bool
-rockchip_vpu_is_encoder_ctx(const struct rockchip_vpu_ctx *ctx)
+static inline void vdpu_write_relaxed(struct rockchip_vpu_dev *vpu,
+				      u32 val, u32 reg)
 {
-	return true;
+	vpu_debug(6, "0x%04x = 0x%08x\n", reg / 4, val);
+	writel_relaxed(val, vpu->dec_base + reg);
 }
 
+static inline void vdpu_write(struct rockchip_vpu_dev *vpu, u32 val, u32 reg)
+{
+	vpu_debug(6, "0x%04x = 0x%08x\n", reg / 4, val);
+	writel(val, vpu->dec_base + reg);
+}
+
+static inline u32 vdpu_read(struct rockchip_vpu_dev *vpu, u32 reg)
+{
+	u32 val = readl(vpu->dec_base + reg);
+
+	vpu_debug(6, "0x%04x = 0x%08x\n", reg / 4, val);
+	return val;
+}
+
+bool rockchip_vpu_is_encoder_ctx(const struct rockchip_vpu_ctx *ctx);
+
 #endif /* ROCKCHIP_VPU_H_ */
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index d85b88067b03..0a8d7fb8903a 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -35,13 +35,48 @@ module_param_named(debug, rockchip_vpu_debug, int, 0644);
 MODULE_PARM_DESC(debug,
 		 "Debug level - higher value produces more verbose messages");
 
+static int
+rockchip_vpu_enc_buf_finish(struct rockchip_vpu_ctx *ctx,
+			    struct vb2_buffer *buf,
+			    unsigned int bytesused)
+{
+	size_t avail_size;
+
+	avail_size = vb2_plane_size(buf, 0) - ctx->vpu_dst_fmt->header_size;
+	if (bytesused > avail_size)
+		return -EINVAL;
+	/*
+	 * The bounce buffer is only for the JPEG encoder.
+	 * TODO: Rework the JPEG encoder to eliminate the need
+	 * for a bounce buffer.
+	 */
+	if (ctx->jpeg_enc.bounce_buffer.cpu) {
+		memcpy(vb2_plane_vaddr(buf, 0) +
+		       ctx->vpu_dst_fmt->header_size,
+		       ctx->jpeg_enc.bounce_buffer.cpu, bytesused);
+	}
+	buf->planes[0].bytesused =
+		ctx->vpu_dst_fmt->header_size + bytesused;
+	return 0;
+}
+
+static int
+rockchip_vpu_dec_buf_finish(struct rockchip_vpu_ctx *ctx,
+			    struct vb2_buffer *buf,
+			    unsigned int bytesused)
+{
+	/* For decoders set bytesused as per the output picture. */
+	buf->planes[0].bytesused = ctx->dst_fmt.plane_fmt[0].sizeimage;
+	return 0;
+}
+
 static void rockchip_vpu_job_finish(struct rockchip_vpu_dev *vpu,
 				    struct rockchip_vpu_ctx *ctx,
 				    unsigned int bytesused,
 				    enum vb2_buffer_state result)
 {
 	struct vb2_v4l2_buffer *src, *dst;
-	size_t avail_size;
+	int ret;
 
 	pm_runtime_mark_last_busy(vpu->dev);
 	pm_runtime_put_autosuspend(vpu->dev);
@@ -60,24 +95,9 @@ static void rockchip_vpu_job_finish(struct rockchip_vpu_dev *vpu,
 
 	v4l2_m2m_buf_copy_metadata(src, dst, true);
 
-	avail_size = vb2_plane_size(&dst->vb2_buf, 0) -
-		     ctx->vpu_dst_fmt->header_size;
-	if (bytesused <= avail_size) {
-		/*
-		 * The bounce buffer is only for the JPEG encoder.
-		 * TODO: Rework the JPEG encoder to eliminate the need
-		 * for a bounce buffer.
-		 */
-		if (ctx->jpeg_enc.bounce_buffer.cpu) {
-			memcpy(vb2_plane_vaddr(&dst->vb2_buf, 0) +
-			       ctx->vpu_dst_fmt->header_size,
-			       ctx->jpeg_enc.bounce_buffer.cpu, bytesused);
-		}
-		dst->vb2_buf.planes[0].bytesused =
-			ctx->vpu_dst_fmt->header_size + bytesused;
-	} else {
+	ret = ctx->buf_finish(ctx, &dst->vb2_buf, bytesused);
+	if (ret)
 		result = VB2_BUF_STATE_ERROR;
-	}
 
 	v4l2_m2m_buf_done(src, result);
 	v4l2_m2m_buf_done(dst, result);
@@ -135,6 +155,11 @@ static void device_run(void *priv)
 	rockchip_vpu_job_finish(ctx->dev, ctx, 0, VB2_BUF_STATE_ERROR);
 }
 
+bool rockchip_vpu_is_encoder_ctx(const struct rockchip_vpu_ctx *ctx)
+{
+	return ctx->buf_finish == rockchip_vpu_enc_buf_finish;
+}
+
 static struct v4l2_m2m_ops vpu_m2m_ops = {
 	.device_run = device_run,
 };
@@ -169,18 +194,25 @@ queue_init(void *priv, struct vb2_queue *src_vq, struct vb2_queue *dst_vq)
 		return ret;
 
 	/*
-	 * The CAPTURE queue doesn't need dma memory,
-	 * as the CPU needs to create the JPEG frames,
-	 * from the hardware-produced JPEG payload.
+	 * When encoding, the CAPTURE queue doesn't need dma memory,
+	 * as the CPU needs to create the JPEG frames, from the
+	 * hardware-produced JPEG payload.
 	 *
-	 * For the DMA destination buffer, we use
-	 * a bounce buffer.
+	 * For the DMA destination buffer, we use a bounce buffer.
 	 */
+	if (rockchip_vpu_is_encoder_ctx(ctx)) {
+		dst_vq->mem_ops = &vb2_vmalloc_memops;
+	} else {
+		dst_vq->bidirectional = true;
+		dst_vq->mem_ops = &vb2_dma_contig_memops;
+		dst_vq->dma_attrs = DMA_ATTR_ALLOC_SINGLE_PAGES |
+				    DMA_ATTR_NO_KERNEL_MAPPING;
+	}
+
 	dst_vq->type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE;
 	dst_vq->io_modes = VB2_MMAP | VB2_DMABUF;
 	dst_vq->drv_priv = ctx;
 	dst_vq->ops = &rockchip_vpu_queue_ops;
-	dst_vq->mem_ops = &vb2_vmalloc_memops;
 	dst_vq->buf_struct_size = sizeof(struct v4l2_m2m_buffer);
 	dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
 	dst_vq->lock = &ctx->dev->vpu_mutex;
@@ -258,11 +290,17 @@ static int rockchip_vpu_open(struct file *filp)
 		return -ENOMEM;
 
 	ctx->dev = vpu;
-	if (func->id == MEDIA_ENT_F_PROC_VIDEO_ENCODER)
+	if (func->id == MEDIA_ENT_F_PROC_VIDEO_ENCODER) {
+		ctx->buf_finish = rockchip_vpu_enc_buf_finish;
 		ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(vpu->m2m_dev, ctx,
 						    queue_init);
-	else
+	} else if (func->id == MEDIA_ENT_F_PROC_VIDEO_DECODER) {
+		ctx->buf_finish = rockchip_vpu_dec_buf_finish;
+		ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(vpu->m2m_dev, ctx,
+						    queue_init);
+	} else {
 		ctx->fh.m2m_ctx = ERR_PTR(-ENODEV);
+	}
 	if (IS_ERR(ctx->fh.m2m_ctx)) {
 		ret = PTR_ERR(ctx->fh.m2m_ctx);
 		kfree(ctx);
@@ -463,7 +501,8 @@ static void rockchip_detach_func(struct rockchip_vpu_func *func)
 	media_device_unregister_entity(&func->vdev.entity);
 }
 
-static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
+static int rockchip_vpu_add_func(struct rockchip_vpu_dev *vpu,
+				 unsigned int funcid)
 {
 	const struct of_device_id *match;
 	struct rockchip_vpu_func *func;
@@ -477,7 +516,7 @@ static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
 		return -ENOMEM;
 	}
 
-	func->id = MEDIA_ENT_F_PROC_VIDEO_ENCODER;
+	func->id = funcid;
 
 	vfd = &func->vdev;
 	vfd->fops = &rockchip_vpu_fops;
@@ -487,9 +526,14 @@ static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
 	vfd->vfl_dir = VFL_DIR_M2M;
 	vfd->device_caps = V4L2_CAP_STREAMING | V4L2_CAP_VIDEO_M2M_MPLANE;
 	vfd->ioctl_ops = &rockchip_vpu_ioctl_ops;
-	snprintf(vfd->name, sizeof(vfd->name), "%s-enc", match->compatible);
+	snprintf(vfd->name, sizeof(vfd->name), "%s-%s", match->compatible,
+		 funcid == MEDIA_ENT_F_PROC_VIDEO_ENCODER ? "enc" : "dec");
+
+	if (funcid == MEDIA_ENT_F_PROC_VIDEO_ENCODER)
+		vpu->encoder = func;
+	else
+		vpu->decoder = func;
 
-	vpu->encoder = func;
 	video_set_drvdata(vfd, vpu);
 
 	ret = video_register_device(vfd, VFL_TYPE_GRABBER, -1);
@@ -514,9 +558,31 @@ static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
 	return ret;
 }
 
-static void rockchip_vpu_remove_enc_func(struct rockchip_vpu_dev *vpu)
+static int rockchip_vpu_add_enc_func(struct rockchip_vpu_dev *vpu)
 {
-	struct rockchip_vpu_func *func = vpu->encoder;
+	if (!vpu->variant->enc_fmts)
+		return 0;
+
+	return rockchip_vpu_add_func(vpu, MEDIA_ENT_F_PROC_VIDEO_ENCODER);
+}
+
+static int rockchip_vpu_add_dec_func(struct rockchip_vpu_dev *vpu)
+{
+	if (!vpu->variant->dec_fmts)
+		return 0;
+
+	return rockchip_vpu_add_func(vpu, MEDIA_ENT_F_PROC_VIDEO_DECODER);
+}
+
+static void rockchip_vpu_remove_func(struct rockchip_vpu_dev *vpu,
+				     unsigned int funcid)
+{
+	struct rockchip_vpu_func *func;
+
+	if (funcid == MEDIA_ENT_F_PROC_VIDEO_ENCODER)
+		func = vpu->encoder;
+	else
+		func = vpu->decoder;
 
 	if (!func)
 		return;
@@ -525,6 +591,16 @@ static void rockchip_vpu_remove_enc_func(struct rockchip_vpu_dev *vpu)
 	video_unregister_device(&func->vdev);
 }
 
+static void rockchip_vpu_remove_enc_func(struct rockchip_vpu_dev *vpu)
+{
+	rockchip_vpu_remove_func(vpu, MEDIA_ENT_F_PROC_VIDEO_ENCODER);
+}
+
+static void rockchip_vpu_remove_dec_func(struct rockchip_vpu_dev *vpu)
+{
+	rockchip_vpu_remove_func(vpu, MEDIA_ENT_F_PROC_VIDEO_DECODER);
+}
+
 static const struct media_device_ops rockchip_m2m_media_ops = {
 	.req_validate = vb2_request_validate,
 	.req_queue = v4l2_m2m_request_queue,
@@ -563,6 +639,7 @@ static int rockchip_vpu_probe(struct platform_device *pdev)
 	if (IS_ERR(vpu->base))
 		return PTR_ERR(vpu->base);
 	vpu->enc_base = vpu->base + vpu->variant->enc_offset;
+	vpu->dec_base = vpu->base + vpu->variant->dec_offset;
 
 	ret = dma_set_coherent_mask(vpu->dev, DMA_BIT_MASK(32));
 	if (ret) {
@@ -570,6 +647,23 @@ static int rockchip_vpu_probe(struct platform_device *pdev)
 		return ret;
 	}
 
+	if (vpu->variant->vdpu_irq) {
+		int irq;
+
+		irq = platform_get_irq_byname(vpu->pdev, "vdpu");
+		if (irq <= 0) {
+			dev_err(vpu->dev, "Could not get vdpu IRQ.\n");
+			return -ENXIO;
+		}
+
+		ret = devm_request_irq(vpu->dev, irq, vpu->variant->vdpu_irq,
+				       0, dev_name(vpu->dev), vpu);
+		if (ret) {
+			dev_err(vpu->dev, "Could not request vdpu IRQ.\n");
+			return ret;
+		}
+	}
+
 	if (vpu->variant->vepu_irq) {
 		int irq;
 
@@ -631,14 +725,22 @@ static int rockchip_vpu_probe(struct platform_device *pdev)
 		goto err_m2m_rel;
 	}
 
+	ret = rockchip_vpu_add_dec_func(vpu);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to register decoder\n");
+		goto err_rm_enc_func;
+	}
+
 	ret = media_device_register(&vpu->mdev);
 	if (ret) {
 		v4l2_err(&vpu->v4l2_dev, "Failed to register mem2mem media device\n");
-		goto err_rm_enc_func;
+		goto err_rm_dec_func;
 	}
 
 	return 0;
 
+err_rm_dec_func:
+	rockchip_vpu_remove_dec_func(vpu);
 err_rm_enc_func:
 	rockchip_vpu_remove_enc_func(vpu);
 err_m2m_rel:
@@ -660,6 +762,7 @@ static int rockchip_vpu_remove(struct platform_device *pdev)
 	v4l2_info(&vpu->v4l2_dev, "Removing %s\n", pdev->name);
 
 	media_device_unregister(&vpu->mdev);
+	rockchip_vpu_remove_dec_func(vpu);
 	rockchip_vpu_remove_enc_func(vpu);
 	media_device_cleanup(&vpu->mdev);
 	v4l2_m2m_release(vpu->m2m_dev);
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
index 1ab558d6492d..9768189148be 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_v4l2.c
@@ -36,8 +36,13 @@ rockchip_vpu_get_formats(const struct rockchip_vpu_ctx *ctx,
 {
 	const struct rockchip_vpu_fmt *formats;
 
-	formats = ctx->dev->variant->enc_fmts;
-	*num_fmts = ctx->dev->variant->num_enc_fmts;
+	if (rockchip_vpu_is_encoder_ctx(ctx)) {
+		formats = ctx->dev->variant->enc_fmts;
+		*num_fmts = ctx->dev->variant->num_enc_fmts;
+	} else {
+		formats = ctx->dev->variant->dec_fmts;
+		*num_fmts = ctx->dev->variant->num_dec_fmts;
+	}
 
 	return formats;
 }
@@ -331,6 +336,22 @@ void rockchip_vpu_reset_fmts(struct rockchip_vpu_ctx *ctx)
 	rockchip_vpu_reset_raw_fmt(ctx);
 }
 
+static void
+rockchip_vpu_update_requires_request(struct rockchip_vpu_ctx *ctx,
+				     u32 fourcc)
+{
+	switch (fourcc) {
+	case V4L2_PIX_FMT_JPEG:
+		ctx->fh.m2m_ctx->out_q_ctx.q.requires_requests = true;
+		break;
+	case V4L2_PIX_FMT_MPEG2_SLICE:
+		ctx->fh.m2m_ctx->out_q_ctx.q.requires_requests = false;
+		break;
+	default:
+		break;
+	}
+}
+
 static int
 vidioc_s_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 {
@@ -387,6 +408,8 @@ vidioc_s_fmt_out_mplane(struct file *file, void *priv, struct v4l2_format *f)
 	ctx->dst_fmt.xfer_func = pix_mp->xfer_func;
 	ctx->dst_fmt.quantization = pix_mp->quantization;
 
+	rockchip_vpu_update_requires_request(ctx, pix_mp->pixelformat);
+
 	vpu_debug(0, "OUTPUT codec mode: %d\n", ctx->vpu_src_fmt->codec_mode);
 	vpu_debug(0, "fmt - w: %d, h: %d\n",
 		  pix_mp->width, pix_mp->height);
@@ -456,6 +479,8 @@ static int vidioc_s_fmt_cap_mplane(struct file *file, void *priv,
 	vpu_debug(0, "fmt - w: %d, h: %d\n",
 		  pix_mp->width, pix_mp->height);
 
+	rockchip_vpu_update_requires_request(ctx, pix_mp->pixelformat);
+
 	return 0;
 }
 
-- 
2.20.1

[PATCH v5 12/15] rockchip/vpu: Add support for non-standard controls

[Boris Brezillon]

From: Ezequiel Garcia <ezequiel@collabora.com>

Rework the way controls are registered by the driver,
so it can support non-standard controls, such as those
used by stateless codecs.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
--
Changes from v4:
* Remove unused var in rockchip_vpu_ctrls_setup()

Changes from v3:
* None

Changes from v2:
* Got rid of unused ctrls related fields.
* Removed unndeeded is_std internal field.
---
 .../media/rockchip/vpu/rk3288_vpu_hw.c        |  2 +-
 .../media/rockchip/vpu/rk3399_vpu_hw.c        |  2 +-
 .../staging/media/rockchip/vpu/rockchip_vpu.h | 17 ++++++-
 .../media/rockchip/vpu/rockchip_vpu_drv.c     | 51 +++++++++++++++----
 4 files changed, 59 insertions(+), 13 deletions(-)

diff --git a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c
index a874a0d83c2d..da19f1cad957 100644
--- a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c
+++ b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c
@@ -112,7 +112,7 @@ const struct rockchip_vpu_variant rk3288_vpu_variant = {
 	.enc_fmts = rk3288_vpu_enc_fmts,
 	.num_enc_fmts = ARRAY_SIZE(rk3288_vpu_enc_fmts),
 	.codec_ops = rk3288_vpu_codec_ops,
-	.codec = RK_VPU_CODEC_JPEG,
+	.codec = RK_VPU_JPEG_ENCODER,
 	.vepu_irq = rk3288_vepu_irq,
 	.init = rk3288_vpu_hw_init,
 	.clk_names = {"aclk", "hclk"},
diff --git a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c
index f4effad00605..5c6c1e8b36d6 100644
--- a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c
+++ b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c
@@ -111,7 +111,7 @@ const struct rockchip_vpu_variant rk3399_vpu_variant = {
 	.enc_offset = 0x0,
 	.enc_fmts = rk3399_vpu_enc_fmts,
 	.num_enc_fmts = ARRAY_SIZE(rk3399_vpu_enc_fmts),
-	.codec = RK_VPU_CODEC_JPEG,
+	.codec = RK_VPU_JPEG_ENCODER,
 	.codec_ops = rk3399_vpu_codec_ops,
 	.vepu_irq = rk3399_vepu_irq,
 	.init = rk3399_vpu_hw_init,
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
index 3d64f3e95c9b..ec7557a98583 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
@@ -34,7 +34,10 @@
 struct rockchip_vpu_ctx;
 struct rockchip_vpu_codec_ops;
 
-#define RK_VPU_CODEC_JPEG BIT(0)
+#define RK_VPU_JPEG_ENCODER	BIT(0)
+#define RK_VPU_ENCODERS		0x0000ffff
+
+#define RK_VPU_DECODERS		0xffff0000
 
 /**
  * struct rockchip_vpu_variant - information about VPU hardware variant
@@ -79,6 +82,18 @@ enum rockchip_vpu_codec_mode {
 	RK_VPU_MODE_JPEG_ENC,
 };
 
+/*
+ * struct rockchip_vpu_ctrl - helper type to declare supported controls
+ * @id:		V4L2 control ID (V4L2_CID_xxx)
+ * @codec:	codec id this control belong to (RK_VPU_JPEG_ENCODER, etc.)
+ * @cfg:	control configuration
+ */
+struct rockchip_vpu_ctrl {
+	unsigned int id;
+	unsigned int codec;
+	struct v4l2_ctrl_config cfg;
+};
+
 /*
  * struct rockchip_vpu_func - rockchip VPU functionality
  *
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index 0a8d7fb8903a..75104174640b 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -245,22 +245,51 @@ static const struct v4l2_ctrl_ops rockchip_vpu_ctrl_ops = {
 	.s_ctrl = rockchip_vpu_s_ctrl,
 };
 
+static struct rockchip_vpu_ctrl controls[] = {
+	{
+		.id = V4L2_CID_JPEG_COMPRESSION_QUALITY,
+		.codec = RK_VPU_JPEG_ENCODER,
+		.cfg = {
+			.min = 5,
+			.max = 100,
+			.step = 1,
+			.def = 50,
+		},
+	},
+};
+
 static int rockchip_vpu_ctrls_setup(struct rockchip_vpu_dev *vpu,
-				    struct rockchip_vpu_ctx *ctx)
+				    struct rockchip_vpu_ctx *ctx,
+				    int allowed_codecs)
 {
-	v4l2_ctrl_handler_init(&ctx->ctrl_handler, 1);
-	if (vpu->variant->codec & RK_VPU_CODEC_JPEG) {
-		v4l2_ctrl_new_std(&ctx->ctrl_handler, &rockchip_vpu_ctrl_ops,
-				  V4L2_CID_JPEG_COMPRESSION_QUALITY,
-				  5, 100, 1, 50);
+	int i, num_ctrls = ARRAY_SIZE(controls);
+
+	v4l2_ctrl_handler_init(&ctx->ctrl_handler, num_ctrls);
+
+	for (i = 0; i < num_ctrls; i++) {
+		if (!(allowed_codecs & controls[i].codec))
+			continue;
+		if (!controls[i].cfg.elem_size) {
+			v4l2_ctrl_new_std(&ctx->ctrl_handler,
+					  &rockchip_vpu_ctrl_ops,
+					  controls[i].id, controls[i].cfg.min,
+					  controls[i].cfg.max,
+					  controls[i].cfg.step,
+					  controls[i].cfg.def);
+		} else {
+			controls[i].cfg.id = controls[i].id;
+			v4l2_ctrl_new_custom(&ctx->ctrl_handler,
+					     &controls[i].cfg, NULL);
+		}
+
 		if (ctx->ctrl_handler.error) {
-			vpu_err("Adding JPEG control failed %d\n",
+			vpu_err("Adding control (%d) failed %d\n",
+				controls[i].id,
 				ctx->ctrl_handler.error);
 			v4l2_ctrl_handler_free(&ctx->ctrl_handler);
 			return ctx->ctrl_handler.error;
 		}
 	}
-
 	return v4l2_ctrl_handler_setup(&ctx->ctrl_handler);
 }
 
@@ -274,7 +303,7 @@ static int rockchip_vpu_open(struct file *filp)
 	struct video_device *vdev = video_devdata(filp);
 	struct rockchip_vpu_func *func = rockchip_vpu_vdev_to_func(vdev);
 	struct rockchip_vpu_ctx *ctx;
-	int ret;
+	int allowed_codecs, ret;
 
 	/*
 	 * We do not need any extra locking here, because we operate only
@@ -291,10 +320,12 @@ static int rockchip_vpu_open(struct file *filp)
 
 	ctx->dev = vpu;
 	if (func->id == MEDIA_ENT_F_PROC_VIDEO_ENCODER) {
+		allowed_codecs = vpu->variant->codec & RK_VPU_ENCODERS;
 		ctx->buf_finish = rockchip_vpu_enc_buf_finish;
 		ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(vpu->m2m_dev, ctx,
 						    queue_init);
 	} else if (func->id == MEDIA_ENT_F_PROC_VIDEO_DECODER) {
+		allowed_codecs = vpu->variant->codec & RK_VPU_DECODERS;
 		ctx->buf_finish = rockchip_vpu_dec_buf_finish;
 		ctx->fh.m2m_ctx = v4l2_m2m_ctx_init(vpu->m2m_dev, ctx,
 						    queue_init);
@@ -313,7 +344,7 @@ static int rockchip_vpu_open(struct file *filp)
 
 	rockchip_vpu_reset_fmts(ctx);
 
-	ret = rockchip_vpu_ctrls_setup(vpu, ctx);
+	ret = rockchip_vpu_ctrls_setup(vpu, ctx, allowed_codecs);
 	if (ret) {
 		vpu_err("Failed to set up controls\n");
 		goto err_fh_free;
-- 
2.20.1

[PATCH v5 13/15] rockchip/vpu: Add infra to support MPEG-2 decoding

[Boris Brezillon]

From: Jonas Karlman <jonas@kwiboo.se>

Only adds structs and helpers to allow supporting MPEG-2 decoding on
rockchip SoCs. Support for RK3399 and RK3288 will be added in separate
commits

Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
--
Changes from v4:
* Define rockchip_vpu_get_ctrl() and rockchip_vpu_get_ref() prototypes
  in rockchip_vpu.h

Changes from v3:
* Coding style improvements (Jonas)
* Move RK3399 support in a separate commit

Changes from v2:
* Remove uneeded check from rockchip_vpu_get_ctrl.
* Use user negotiated resolution when programming the hardware.
---
 drivers/staging/media/rockchip/vpu/Makefile   |  3 +-
 .../staging/media/rockchip/vpu/rockchip_vpu.h | 12 ++++
 .../media/rockchip/vpu/rockchip_vpu_drv.c     | 30 +++++++++
 .../media/rockchip/vpu/rockchip_vpu_hw.h      | 14 +++++
 .../media/rockchip/vpu/rockchip_vpu_mpeg2.c   | 61 +++++++++++++++++++
 5 files changed, 119 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/media/rockchip/vpu/rockchip_vpu_mpeg2.c

diff --git a/drivers/staging/media/rockchip/vpu/Makefile b/drivers/staging/media/rockchip/vpu/Makefile
index ae512173e7e4..dc963936b9f9 100644
--- a/drivers/staging/media/rockchip/vpu/Makefile
+++ b/drivers/staging/media/rockchip/vpu/Makefile
@@ -7,4 +7,5 @@ rockchip-vpu-y += \
 		rk3288_vpu_hw_jpeg_enc.o \
 		rk3399_vpu_hw.o \
 		rk3399_vpu_hw_jpeg_enc.o \
-		rockchip_vpu_jpeg.o
+		rockchip_vpu_jpeg.o \
+		rockchip_vpu_mpeg2.o
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
index ec7557a98583..3093821440c0 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu.h
@@ -27,6 +27,10 @@
 
 #define ROCKCHIP_VPU_MAX_CLOCKS		4
 
+#define MPEG2_MB_DIM			16
+#define MPEG2_MB_WIDTH(w)		DIV_ROUND_UP(w, MPEG2_MB_DIM)
+#define MPEG2_MB_HEIGHT(h)		DIV_ROUND_UP(h, MPEG2_MB_DIM)
+
 #define JPEG_MB_DIM			16
 #define JPEG_MB_WIDTH(w)		DIV_ROUND_UP(w, JPEG_MB_DIM)
 #define JPEG_MB_HEIGHT(h)		DIV_ROUND_UP(h, JPEG_MB_DIM)
@@ -37,6 +41,7 @@ struct rockchip_vpu_codec_ops;
 #define RK_VPU_JPEG_ENCODER	BIT(0)
 #define RK_VPU_ENCODERS		0x0000ffff
 
+#define RK_VPU_MPEG2_DECODER	BIT(16)
 #define RK_VPU_DECODERS		0xffff0000
 
 /**
@@ -76,10 +81,12 @@ struct rockchip_vpu_variant {
  * enum rockchip_vpu_codec_mode - codec operating mode.
  * @RK_VPU_MODE_NONE:  No operating mode. Used for RAW video formats.
  * @RK_VPU_MODE_JPEG_ENC: JPEG encoder.
+ * @RK_VPU_MODE_MPEG2_DEC: MPEG-2 decoder.
  */
 enum rockchip_vpu_codec_mode {
 	RK_VPU_MODE_NONE = -1,
 	RK_VPU_MODE_JPEG_ENC,
+	RK_VPU_MODE_MPEG2_DEC,
 };
 
 /*
@@ -190,6 +197,7 @@ struct rockchip_vpu_dev {
  *			calling v4l2_m2m_job_finish.
  * @codec_ops:		Set of operations related to codec mode.
  * @jpeg_enc:		JPEG-encoding context.
+ * @mpeg2_dec:		MPEG-2-decoding context.
  */
 struct rockchip_vpu_ctx {
 	struct rockchip_vpu_dev *dev;
@@ -215,6 +223,7 @@ struct rockchip_vpu_ctx {
 	/* Specific for particular codec modes. */
 	union {
 		struct rockchip_vpu_jpeg_enc_hw_ctx jpeg_enc;
+		struct rockchip_vpu_mpeg2_dec_hw_ctx mpeg2_dec;
 	};
 };
 
@@ -319,4 +328,7 @@ static inline u32 vdpu_read(struct rockchip_vpu_dev *vpu, u32 reg)
 
 bool rockchip_vpu_is_encoder_ctx(const struct rockchip_vpu_ctx *ctx);
 
+void *rockchip_vpu_get_ctrl(struct rockchip_vpu_ctx *ctx, u32 id);
+dma_addr_t rockchip_vpu_get_ref(struct vb2_queue *q, u64 ts);
+
 #endif /* ROCKCHIP_VPU_H_ */
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
index 75104174640b..b94ff97451db 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_drv.c
@@ -35,6 +35,24 @@ module_param_named(debug, rockchip_vpu_debug, int, 0644);
 MODULE_PARM_DESC(debug,
 		 "Debug level - higher value produces more verbose messages");
 
+void *rockchip_vpu_get_ctrl(struct rockchip_vpu_ctx *ctx, u32 id)
+{
+	struct v4l2_ctrl *ctrl;
+
+	ctrl = v4l2_ctrl_find(&ctx->ctrl_handler, id);
+	return ctrl ? ctrl->p_cur.p : NULL;
+}
+
+dma_addr_t rockchip_vpu_get_ref(struct vb2_queue *q, u64 ts)
+{
+	int index;
+
+	index = vb2_find_timestamp(q, ts, 0);
+	if (index >= 0)
+		return vb2_dma_contig_plane_dma_addr(q->bufs[index], 0);
+	return 0;
+}
+
 static int
 rockchip_vpu_enc_buf_finish(struct rockchip_vpu_ctx *ctx,
 			    struct vb2_buffer *buf,
@@ -255,6 +273,18 @@ static struct rockchip_vpu_ctrl controls[] = {
 			.step = 1,
 			.def = 50,
 		},
+	}, {
+		.id = V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS,
+		.codec = RK_VPU_MPEG2_DECODER,
+		.cfg = {
+			.elem_size = sizeof(struct v4l2_ctrl_mpeg2_slice_params),
+		},
+	}, {
+		.id = V4L2_CID_MPEG_VIDEO_MPEG2_QUANTIZATION,
+		.codec = RK_VPU_MPEG2_DECODER,
+		.cfg = {
+			.elem_size = sizeof(struct v4l2_ctrl_mpeg2_quantization),
+		},
 	},
 };
 
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
index 46716d121538..1bdc5ceb956f 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
@@ -11,6 +11,7 @@
 
 #include <linux/interrupt.h>
 #include <linux/v4l2-controls.h>
+#include <media/mpeg2-ctrls.h>
 #include <media/videobuf2-core.h>
 
 struct rockchip_vpu_dev;
@@ -38,6 +39,14 @@ struct rockchip_vpu_jpeg_enc_hw_ctx {
 	struct rockchip_vpu_aux_buf bounce_buffer;
 };
 
+/**
+ * struct rockchip_vpu_mpeg2_dec_hw_ctx
+ * @qtable:		Quantization table
+ */
+struct rockchip_vpu_mpeg2_dec_hw_ctx {
+	struct rockchip_vpu_aux_buf qtable;
+};
+
 /**
  * struct rockchip_vpu_codec_ops - codec mode specific operations
  *
@@ -83,4 +92,9 @@ void rk3399_vpu_jpeg_enc_run(struct rockchip_vpu_ctx *ctx);
 int rockchip_vpu_jpeg_enc_init(struct rockchip_vpu_ctx *ctx);
 void rockchip_vpu_jpeg_enc_exit(struct rockchip_vpu_ctx *ctx);
 
+void rockchip_vpu_mpeg2_dec_copy_qtable(u8 *qtable,
+	const struct v4l2_ctrl_mpeg2_quantization *ctrl);
+int rockchip_vpu_mpeg2_dec_init(struct rockchip_vpu_ctx *ctx);
+void rockchip_vpu_mpeg2_dec_exit(struct rockchip_vpu_ctx *ctx);
+
 #endif /* ROCKCHIP_VPU_HW_H_ */
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_mpeg2.c b/drivers/staging/media/rockchip/vpu/rockchip_vpu_mpeg2.c
new file mode 100644
index 000000000000..5a5b9ea1f6b5
--- /dev/null
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_mpeg2.c
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Rockchip VPU codec driver
+ *
+ * Copyright (C) 2018 Rockchip Electronics Co., Ltd.
+ */
+
+#include "rockchip_vpu.h"
+
+static const u8 zigzag[64] = {
+	0,   1,  8, 16,  9,  2,  3, 10,
+	17, 24, 32, 25, 18, 11,  4,  5,
+	12, 19, 26, 33, 40, 48, 41, 34,
+	27, 20, 13,  6,  7, 14, 21, 28,
+	35, 42, 49, 56, 57, 50, 43, 36,
+	29, 22, 15, 23, 30, 37, 44, 51,
+	58, 59, 52, 45, 38, 31, 39, 46,
+	53, 60, 61, 54, 47, 55, 62, 63
+};
+
+void rockchip_vpu_mpeg2_dec_copy_qtable(u8 *qtable,
+	const struct v4l2_ctrl_mpeg2_quantization *ctrl)
+{
+	int i, n;
+
+	if (!qtable || !ctrl)
+		return;
+
+	for (i = 0; i < ARRAY_SIZE(zigzag); i++) {
+		n = zigzag[i];
+		qtable[n + 0] = ctrl->intra_quantiser_matrix[i];
+		qtable[n + 64] = ctrl->non_intra_quantiser_matrix[i];
+		qtable[n + 128] = ctrl->chroma_intra_quantiser_matrix[i];
+		qtable[n + 192] = ctrl->chroma_non_intra_quantiser_matrix[i];
+	}
+}
+
+int rockchip_vpu_mpeg2_dec_init(struct rockchip_vpu_ctx *ctx)
+{
+	struct rockchip_vpu_dev *vpu = ctx->dev;
+
+	ctx->mpeg2_dec.qtable.size = ARRAY_SIZE(zigzag) * 4;
+	ctx->mpeg2_dec.qtable.cpu =
+		dma_alloc_coherent(vpu->dev,
+				   ctx->mpeg2_dec.qtable.size,
+				   &ctx->mpeg2_dec.qtable.dma,
+				   GFP_KERNEL);
+	if (!ctx->mpeg2_dec.qtable.cpu)
+		return -ENOMEM;
+	return 0;
+}
+
+void rockchip_vpu_mpeg2_dec_exit(struct rockchip_vpu_ctx *ctx)
+{
+	struct rockchip_vpu_dev *vpu = ctx->dev;
+
+	dma_free_coherent(vpu->dev,
+			  ctx->mpeg2_dec.qtable.size,
+			  ctx->mpeg2_dec.qtable.cpu,
+			  ctx->mpeg2_dec.qtable.dma);
+}
-- 
2.20.1

[PATCH v5 14/15] rockchip/vpu: Add MPEG2 decoding support to RK3399

[Boris Brezillon]

From: Jonas Karlman <jonas@kwiboo.se>

Add the necessary bits to support MPEG2 decoding on RK3399.

Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes from v4:
* Stop including rockchip_vpu_common.h

Changes from v3:
* New patch (was part of another patch before)
---
 drivers/staging/media/rockchip/vpu/Makefile   |   1 +
 .../media/rockchip/vpu/rk3399_vpu_hw.c        |  59 +++-
 .../rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c    | 267 ++++++++++++++++++
 .../media/rockchip/vpu/rockchip_vpu_hw.h      |   1 +
 4 files changed, 327 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c

diff --git a/drivers/staging/media/rockchip/vpu/Makefile b/drivers/staging/media/rockchip/vpu/Makefile
index dc963936b9f9..fdf25c5d4b7e 100644
--- a/drivers/staging/media/rockchip/vpu/Makefile
+++ b/drivers/staging/media/rockchip/vpu/Makefile
@@ -7,5 +7,6 @@ rockchip-vpu-y += \
 		rk3288_vpu_hw_jpeg_enc.o \
 		rk3399_vpu_hw.o \
 		rk3399_vpu_hw_jpeg_enc.o \
+		rk3399_vpu_hw_mpeg2_dec.o \
 		rockchip_vpu_jpeg.o \
 		rockchip_vpu_mpeg2.o
diff --git a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c
index 5c6c1e8b36d6..2b3689968ef4 100644
--- a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c
+++ b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw.c
@@ -55,6 +55,26 @@ static const struct rockchip_vpu_fmt rk3399_vpu_enc_fmts[] = {
 	},
 };
 
+static const struct rockchip_vpu_fmt rk3399_vpu_dec_fmts[] = {
+	{
+		.fourcc = V4L2_PIX_FMT_NV12,
+		.codec_mode = RK_VPU_MODE_NONE,
+	},
+	{
+		.fourcc = V4L2_PIX_FMT_MPEG2_SLICE,
+		.codec_mode = RK_VPU_MODE_MPEG2_DEC,
+		.max_depth = 2,
+		.frmsize = {
+			.min_width = 48,
+			.max_width = 1920,
+			.step_width = MPEG2_MB_DIM,
+			.min_height = 48,
+			.max_height = 1088,
+			.step_height = MPEG2_MB_DIM,
+		},
+	},
+};
+
 static irqreturn_t rk3399_vepu_irq(int irq, void *dev_id)
 {
 	struct rockchip_vpu_dev *vpu = dev_id;
@@ -74,6 +94,24 @@ static irqreturn_t rk3399_vepu_irq(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
+static irqreturn_t rk3399_vdpu_irq(int irq, void *dev_id)
+{
+	struct rockchip_vpu_dev *vpu = dev_id;
+	enum vb2_buffer_state state;
+	u32 status;
+
+	status = vdpu_read(vpu, VDPU_REG_INTERRUPT);
+	state = (status & VDPU_REG_INTERRUPT_DEC_IRQ) ?
+		VB2_BUF_STATE_DONE : VB2_BUF_STATE_ERROR;
+
+	vdpu_write(vpu, 0, VDPU_REG_INTERRUPT);
+	vdpu_write(vpu, 0, VDPU_REG_AXI_CTRL);
+
+	rockchip_vpu_irq_done(vpu, 0, state);
+
+	return IRQ_HANDLED;
+}
+
 static int rk3399_vpu_hw_init(struct rockchip_vpu_dev *vpu)
 {
 	/* Bump ACLK to max. possible freq. to improve performance. */
@@ -90,6 +128,15 @@ static void rk3399_vpu_enc_reset(struct rockchip_vpu_ctx *ctx)
 	vepu_write(vpu, 0, VEPU_REG_AXI_CTRL);
 }
 
+static void rk3399_vpu_dec_reset(struct rockchip_vpu_ctx *ctx)
+{
+	struct rockchip_vpu_dev *vpu = ctx->dev;
+
+	vdpu_write(vpu, VDPU_REG_INTERRUPT_DEC_IRQ_DIS, VDPU_REG_INTERRUPT);
+	vdpu_write(vpu, 0, VDPU_REG_EN_FLAGS);
+	vdpu_write(vpu, 1, VDPU_REG_SOFT_RESET);
+}
+
 /*
  * Supported codec ops.
  */
@@ -101,6 +148,12 @@ static const struct rockchip_vpu_codec_ops rk3399_vpu_codec_ops[] = {
 		.init = rockchip_vpu_jpeg_enc_init,
 		.exit = rockchip_vpu_jpeg_enc_exit,
 	},
+	[RK_VPU_MODE_MPEG2_DEC] = {
+		.run = rk3399_vpu_mpeg2_dec_run,
+		.reset = rk3399_vpu_dec_reset,
+		.init = rockchip_vpu_mpeg2_dec_init,
+		.exit = rockchip_vpu_mpeg2_dec_exit,
+	},
 };
 
 /*
@@ -111,9 +164,13 @@ const struct rockchip_vpu_variant rk3399_vpu_variant = {
 	.enc_offset = 0x0,
 	.enc_fmts = rk3399_vpu_enc_fmts,
 	.num_enc_fmts = ARRAY_SIZE(rk3399_vpu_enc_fmts),
-	.codec = RK_VPU_JPEG_ENCODER,
+	.dec_offset = 0x400,
+	.dec_fmts = rk3399_vpu_dec_fmts,
+	.num_dec_fmts = ARRAY_SIZE(rk3399_vpu_dec_fmts),
+	.codec = RK_VPU_JPEG_ENCODER | RK_VPU_MPEG2_DECODER,
 	.codec_ops = rk3399_vpu_codec_ops,
 	.vepu_irq = rk3399_vepu_irq,
+	.vdpu_irq = rk3399_vdpu_irq,
 	.init = rk3399_vpu_hw_init,
 	.clk_names = {"aclk", "hclk"},
 	.num_clocks = 2
diff --git a/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c
new file mode 100644
index 000000000000..c4c092c2004a
--- /dev/null
+++ b/drivers/staging/media/rockchip/vpu/rk3399_vpu_hw_mpeg2_dec.c
@@ -0,0 +1,267 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Rockchip VPU codec driver
+ *
+ * Copyright (C) 2018 Rockchip Electronics Co., Ltd.
+ */
+
+#include <asm/unaligned.h>
+#include <linux/bitfield.h>
+#include <media/v4l2-mem2mem.h>
+#include "rockchip_vpu.h"
+#include "rockchip_vpu_hw.h"
+
+#define VDPU_SWREG(nr)			((nr) * 4)
+
+#define VDPU_REG_DEC_OUT_BASE		VDPU_SWREG(63)
+#define VDPU_REG_RLC_VLC_BASE		VDPU_SWREG(64)
+#define VDPU_REG_QTABLE_BASE		VDPU_SWREG(61)
+#define VDPU_REG_REFER0_BASE		VDPU_SWREG(131)
+#define VDPU_REG_REFER2_BASE		VDPU_SWREG(134)
+#define VDPU_REG_REFER3_BASE		VDPU_SWREG(135)
+#define VDPU_REG_REFER1_BASE		VDPU_SWREG(148)
+#define VDPU_REG_DEC_E(v)		((v) ? BIT(0) : 0)
+
+#define VDPU_REG_DEC_ADV_PRE_DIS(v)	((v) ? BIT(11) : 0)
+#define VDPU_REG_DEC_SCMD_DIS(v)	((v) ? BIT(10) : 0)
+#define VDPU_REG_FILTERING_DIS(v)	((v) ? BIT(8) : 0)
+#define VDPU_REG_DEC_LATENCY(v)		(((v) << 1) & GENMASK(6, 1))
+
+#define VDPU_REG_INIT_QP(v)		(((v) << 25) & GENMASK(30, 25))
+#define VDPU_REG_STREAM_LEN(v)		(((v) << 0) & GENMASK(23, 0))
+
+#define VDPU_REG_APF_THRESHOLD(v)	(((v) << 17) & GENMASK(30, 17))
+#define VDPU_REG_STARTMB_X(v)		(((v) << 8) & GENMASK(16, 8))
+#define VDPU_REG_STARTMB_Y(v)		(((v) << 0) & GENMASK(7, 0))
+
+#define VDPU_REG_DEC_MODE(v)		(((v) << 0) & GENMASK(3, 0))
+
+#define VDPU_REG_DEC_STRENDIAN_E(v)	((v) ? BIT(5) : 0)
+#define VDPU_REG_DEC_STRSWAP32_E(v)	((v) ? BIT(4) : 0)
+#define VDPU_REG_DEC_OUTSWAP32_E(v)	((v) ? BIT(3) : 0)
+#define VDPU_REG_DEC_INSWAP32_E(v)	((v) ? BIT(2) : 0)
+#define VDPU_REG_DEC_OUT_ENDIAN(v)	((v) ? BIT(1) : 0)
+#define VDPU_REG_DEC_IN_ENDIAN(v)	((v) ? BIT(0) : 0)
+
+#define VDPU_REG_DEC_DATA_DISC_E(v)	((v) ? BIT(22) : 0)
+#define VDPU_REG_DEC_MAX_BURST(v)	(((v) << 16) & GENMASK(20, 16))
+#define VDPU_REG_DEC_AXI_WR_ID(v)	(((v) << 8) & GENMASK(15, 8))
+#define VDPU_REG_DEC_AXI_RD_ID(v)	(((v) << 0) & GENMASK(7, 0))
+
+#define VDPU_REG_RLC_MODE_E(v)		((v) ? BIT(20) : 0)
+#define VDPU_REG_PIC_INTERLACE_E(v)	((v) ? BIT(17) : 0)
+#define VDPU_REG_PIC_FIELDMODE_E(v)	((v) ? BIT(16) : 0)
+#define VDPU_REG_PIC_B_E(v)		((v) ? BIT(15) : 0)
+#define VDPU_REG_PIC_INTER_E(v)		((v) ? BIT(14) : 0)
+#define VDPU_REG_PIC_TOPFIELD_E(v)	((v) ? BIT(13) : 0)
+#define VDPU_REG_FWD_INTERLACE_E(v)	((v) ? BIT(12) : 0)
+#define VDPU_REG_WRITE_MVS_E(v)		((v) ? BIT(10) : 0)
+#define VDPU_REG_DEC_TIMEOUT_E(v)	((v) ? BIT(5) : 0)
+#define VDPU_REG_DEC_CLK_GATE_E(v)	((v) ? BIT(4) : 0)
+
+#define VDPU_REG_PIC_MB_WIDTH(v)	(((v) << 23) & GENMASK(31, 23))
+#define VDPU_REG_PIC_MB_HEIGHT_P(v)	(((v) << 11) & GENMASK(18, 11))
+#define VDPU_REG_ALT_SCAN_E(v)		((v) ? BIT(6) : 0)
+#define VDPU_REG_TOPFIELDFIRST_E(v)	((v) ? BIT(5) : 0)
+
+#define VDPU_REG_STRM_START_BIT(v)	(((v) << 26) & GENMASK(31, 26))
+#define VDPU_REG_QSCALE_TYPE(v)		((v) ? BIT(24) : 0)
+#define VDPU_REG_CON_MV_E(v)		((v) ? BIT(4) : 0)
+#define VDPU_REG_INTRA_DC_PREC(v)	(((v) << 2) & GENMASK(3, 2))
+#define VDPU_REG_INTRA_VLC_TAB(v)	((v) ? BIT(1) : 0)
+#define VDPU_REG_FRAME_PRED_DCT(v)	((v) ? BIT(0) : 0)
+
+#define VDPU_REG_ALT_SCAN_FLAG_E(v)	((v) ? BIT(19) : 0)
+#define VDPU_REG_FCODE_FWD_HOR(v)	(((v) << 15) & GENMASK(18, 15))
+#define VDPU_REG_FCODE_FWD_VER(v)	(((v) << 11) & GENMASK(14, 11))
+#define VDPU_REG_FCODE_BWD_HOR(v)	(((v) << 7) & GENMASK(10, 7))
+#define VDPU_REG_FCODE_BWD_VER(v)	(((v) << 3) & GENMASK(6, 3))
+#define VDPU_REG_MV_ACCURACY_FWD(v)	((v) ? BIT(2) : 0)
+#define VDPU_REG_MV_ACCURACY_BWD(v)	((v) ? BIT(1) : 0)
+
+#define PICT_TOP_FIELD     1
+#define PICT_BOTTOM_FIELD  2
+#define PICT_FRAME         3
+
+static void
+rk3399_vpu_mpeg2_dec_set_quantization(struct rockchip_vpu_dev *vpu,
+				      struct rockchip_vpu_ctx *ctx)
+{
+	struct v4l2_ctrl_mpeg2_quantization *quantization;
+
+	quantization = rockchip_vpu_get_ctrl(ctx,
+				V4L2_CID_MPEG_VIDEO_MPEG2_QUANTIZATION);
+	rockchip_vpu_mpeg2_dec_copy_qtable(ctx->mpeg2_dec.qtable.cpu,
+					   quantization);
+	vdpu_write_relaxed(vpu, ctx->mpeg2_dec.qtable.dma,
+			   VDPU_REG_QTABLE_BASE);
+}
+
+static void
+rk3399_vpu_mpeg2_dec_set_buffers(struct rockchip_vpu_dev *vpu,
+				 struct rockchip_vpu_ctx *ctx,
+				 struct vb2_buffer *src_buf,
+				 struct vb2_buffer *dst_buf,
+				 const struct v4l2_mpeg2_sequence *sequence,
+				 const struct v4l2_mpeg2_picture *picture,
+				 const struct v4l2_ctrl_mpeg2_slice_params *slice_params)
+{
+	dma_addr_t forward_addr = 0, backward_addr = 0;
+	dma_addr_t current_addr, addr;
+	struct vb2_queue *vq;
+
+	vq = v4l2_m2m_get_dst_vq(ctx->fh.m2m_ctx);
+
+	switch (picture->picture_coding_type) {
+	case V4L2_MPEG2_PICTURE_CODING_TYPE_B:
+		backward_addr = rockchip_vpu_get_ref(vq,
+						slice_params->backward_ref_ts);
+		/* fall-through */
+	case V4L2_MPEG2_PICTURE_CODING_TYPE_P:
+		forward_addr = rockchip_vpu_get_ref(vq,
+						slice_params->forward_ref_ts);
+	}
+
+	/* Source bitstream buffer */
+	addr = vb2_dma_contig_plane_dma_addr(src_buf, 0);
+	vdpu_write_relaxed(vpu, addr, VDPU_REG_RLC_VLC_BASE);
+
+	/* Destination frame buffer */
+	addr = vb2_dma_contig_plane_dma_addr(dst_buf, 0);
+	current_addr = addr;
+
+	if (picture->picture_structure == PICT_BOTTOM_FIELD)
+		addr += ALIGN(ctx->dst_fmt.width, 16);
+	vdpu_write_relaxed(vpu, addr, VDPU_REG_DEC_OUT_BASE);
+
+	if (!forward_addr)
+		forward_addr = current_addr;
+	if (!backward_addr)
+		backward_addr = current_addr;
+
+	/* Set forward ref frame (top/bottom field) */
+	if (picture->picture_structure == PICT_FRAME ||
+	    picture->picture_coding_type == V4L2_MPEG2_PICTURE_CODING_TYPE_B ||
+	    (picture->picture_structure == PICT_TOP_FIELD &&
+	     picture->top_field_first) ||
+	    (picture->picture_structure == PICT_BOTTOM_FIELD &&
+	     !picture->top_field_first)) {
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER0_BASE);
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER1_BASE);
+	} else if (picture->picture_structure == PICT_TOP_FIELD) {
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER0_BASE);
+		vdpu_write_relaxed(vpu, current_addr, VDPU_REG_REFER1_BASE);
+	} else if (picture->picture_structure == PICT_BOTTOM_FIELD) {
+		vdpu_write_relaxed(vpu, current_addr, VDPU_REG_REFER0_BASE);
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER1_BASE);
+	}
+
+	/* Set backward ref frame (top/bottom field) */
+	vdpu_write_relaxed(vpu, backward_addr, VDPU_REG_REFER2_BASE);
+	vdpu_write_relaxed(vpu, backward_addr, VDPU_REG_REFER3_BASE);
+}
+
+void rk3399_vpu_mpeg2_dec_run(struct rockchip_vpu_ctx *ctx)
+{
+	struct rockchip_vpu_dev *vpu = ctx->dev;
+	struct vb2_v4l2_buffer *src_buf, *dst_buf;
+	const struct v4l2_ctrl_mpeg2_slice_params *slice_params;
+	const struct v4l2_mpeg2_sequence *sequence;
+	const struct v4l2_mpeg2_picture *picture;
+	u32 reg;
+
+	src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
+	dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx);
+
+	/* Apply request controls if any */
+	v4l2_ctrl_request_setup(src_buf->vb2_buf.req_obj.req,
+				&ctx->ctrl_handler);
+
+	slice_params = rockchip_vpu_get_ctrl(ctx,
+				V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS);
+	sequence = &slice_params->sequence;
+	picture = &slice_params->picture;
+
+	reg = VDPU_REG_DEC_ADV_PRE_DIS(0) |
+	      VDPU_REG_DEC_SCMD_DIS(0) |
+	      VDPU_REG_FILTERING_DIS(1) |
+	      VDPU_REG_DEC_LATENCY(0);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(50));
+
+	reg = VDPU_REG_INIT_QP(1) |
+	      VDPU_REG_STREAM_LEN(slice_params->bit_size >> 3);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(51));
+
+	reg = VDPU_REG_APF_THRESHOLD(8) |
+	      VDPU_REG_STARTMB_X(0) |
+	      VDPU_REG_STARTMB_Y(0);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(52));
+
+	reg = VDPU_REG_DEC_MODE(5);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(53));
+
+	reg = VDPU_REG_DEC_STRENDIAN_E(1) |
+	      VDPU_REG_DEC_STRSWAP32_E(1) |
+	      VDPU_REG_DEC_OUTSWAP32_E(1) |
+	      VDPU_REG_DEC_INSWAP32_E(1) |
+	      VDPU_REG_DEC_OUT_ENDIAN(1) |
+	      VDPU_REG_DEC_IN_ENDIAN(1);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(54));
+
+	reg = VDPU_REG_DEC_DATA_DISC_E(0) |
+	      VDPU_REG_DEC_MAX_BURST(16) |
+	      VDPU_REG_DEC_AXI_WR_ID(0) |
+	      VDPU_REG_DEC_AXI_RD_ID(0);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(56));
+
+	reg = VDPU_REG_RLC_MODE_E(0) |
+	      VDPU_REG_PIC_INTERLACE_E(!sequence->progressive_sequence) |
+	      VDPU_REG_PIC_FIELDMODE_E(picture->picture_structure != PICT_FRAME) |
+	      VDPU_REG_PIC_B_E(picture->picture_coding_type == V4L2_MPEG2_PICTURE_CODING_TYPE_B) |
+	      VDPU_REG_PIC_INTER_E(picture->picture_coding_type != V4L2_MPEG2_PICTURE_CODING_TYPE_I) |
+	      VDPU_REG_PIC_TOPFIELD_E(picture->picture_structure == PICT_TOP_FIELD) |
+	      VDPU_REG_FWD_INTERLACE_E(0) |
+	      VDPU_REG_WRITE_MVS_E(0) |
+	      VDPU_REG_DEC_TIMEOUT_E(1) |
+	      VDPU_REG_DEC_CLK_GATE_E(1);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(57));
+
+	reg = VDPU_REG_PIC_MB_WIDTH(MPEG2_MB_WIDTH(ctx->dst_fmt.width)) |
+	      VDPU_REG_PIC_MB_HEIGHT_P(MPEG2_MB_HEIGHT(ctx->dst_fmt.height)) |
+	      VDPU_REG_ALT_SCAN_E(picture->alternate_scan) |
+	      VDPU_REG_TOPFIELDFIRST_E(picture->top_field_first);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(120));
+
+	reg = VDPU_REG_STRM_START_BIT(slice_params->data_bit_offset) |
+	      VDPU_REG_QSCALE_TYPE(picture->q_scale_type) |
+	      VDPU_REG_CON_MV_E(picture->concealment_motion_vectors) |
+	      VDPU_REG_INTRA_DC_PREC(picture->intra_dc_precision) |
+	      VDPU_REG_INTRA_VLC_TAB(picture->intra_vlc_format) |
+	      VDPU_REG_FRAME_PRED_DCT(picture->frame_pred_frame_dct);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(122));
+
+	reg = VDPU_REG_ALT_SCAN_FLAG_E(picture->alternate_scan) |
+	      VDPU_REG_FCODE_FWD_HOR(picture->f_code[0][0]) |
+	      VDPU_REG_FCODE_FWD_VER(picture->f_code[0][1]) |
+	      VDPU_REG_FCODE_BWD_HOR(picture->f_code[1][0]) |
+	      VDPU_REG_FCODE_BWD_VER(picture->f_code[1][1]) |
+	      VDPU_REG_MV_ACCURACY_FWD(1) |
+	      VDPU_REG_MV_ACCURACY_BWD(1);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(136));
+
+	rk3399_vpu_mpeg2_dec_set_quantization(vpu, ctx);
+
+	rk3399_vpu_mpeg2_dec_set_buffers(vpu, ctx, &src_buf->vb2_buf,
+					 &dst_buf->vb2_buf,
+					 sequence, picture, slice_params);
+
+	/* Controls no longer in-use, we can complete them */
+	v4l2_ctrl_request_complete(src_buf->vb2_buf.req_obj.req,
+				   &ctx->ctrl_handler);
+
+	/* Kick the watchdog and start decoding */
+	schedule_delayed_work(&vpu->watchdog_work, msecs_to_jiffies(2000));
+
+	reg = vdpu_read(vpu, VDPU_SWREG(57)) | VDPU_REG_DEC_E(1);
+	vdpu_write(vpu, reg, VDPU_SWREG(57));
+}
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
index 1bdc5ceb956f..4268c3a8b924 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
@@ -92,6 +92,7 @@ void rk3399_vpu_jpeg_enc_run(struct rockchip_vpu_ctx *ctx);
 int rockchip_vpu_jpeg_enc_init(struct rockchip_vpu_ctx *ctx);
 void rockchip_vpu_jpeg_enc_exit(struct rockchip_vpu_ctx *ctx);
 
+void rk3399_vpu_mpeg2_dec_run(struct rockchip_vpu_ctx *ctx);
 void rockchip_vpu_mpeg2_dec_copy_qtable(u8 *qtable,
 	const struct v4l2_ctrl_mpeg2_quantization *ctrl);
 int rockchip_vpu_mpeg2_dec_init(struct rockchip_vpu_ctx *ctx);
-- 
2.20.1

[PATCH v5 15/15] rockchip/vpu: Add support for MPEG-2 decoding on RK3288

[Boris Brezillon]

From: Jonas Karlman <jonas@kwiboo.se>

Add necessary bits to support MPEG2 decoding on RK3288.

Signed-off-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
---
Changes from v4:
* Stop including rockchip_vpu_common.h

Changes from v3:
* Coding style improvements (Jonas)
* Use ALIGN() instead of div+shift (Jonas)

Changes from v2:
* New patch
---
 drivers/staging/media/rockchip/vpu/Makefile   |   1 +
 .../media/rockchip/vpu/rk3288_vpu_hw.c        |  59 +++-
 .../rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c    | 261 ++++++++++++++++++
 .../media/rockchip/vpu/rk3288_vpu_regs.h      |   1 +
 .../media/rockchip/vpu/rockchip_vpu_hw.h      |   1 +
 5 files changed, 322 insertions(+), 1 deletion(-)
 create mode 100644 drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c

diff --git a/drivers/staging/media/rockchip/vpu/Makefile b/drivers/staging/media/rockchip/vpu/Makefile
index fdf25c5d4b7e..d5b833a15653 100644
--- a/drivers/staging/media/rockchip/vpu/Makefile
+++ b/drivers/staging/media/rockchip/vpu/Makefile
@@ -5,6 +5,7 @@ rockchip-vpu-y += \
 		rockchip_vpu_v4l2.o \
 		rk3288_vpu_hw.o \
 		rk3288_vpu_hw_jpeg_enc.o \
+		rk3288_vpu_hw_mpeg2_dec.o \
 		rk3399_vpu_hw.o \
 		rk3399_vpu_hw_jpeg_enc.o \
 		rk3399_vpu_hw_mpeg2_dec.o \
diff --git a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c
index da19f1cad957..003143c77d37 100644
--- a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c
+++ b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw.c
@@ -55,6 +55,26 @@ static const struct rockchip_vpu_fmt rk3288_vpu_enc_fmts[] = {
 	},
 };
 
+static const struct rockchip_vpu_fmt rk3288_vpu_dec_fmts[] = {
+	{
+		.fourcc = V4L2_PIX_FMT_NV12,
+		.codec_mode = RK_VPU_MODE_NONE,
+	},
+	{
+		.fourcc = V4L2_PIX_FMT_MPEG2_SLICE,
+		.codec_mode = RK_VPU_MODE_MPEG2_DEC,
+		.max_depth = 2,
+		.frmsize = {
+			.min_width = 48,
+			.max_width = 1920,
+			.step_width = MPEG2_MB_DIM,
+			.min_height = 48,
+			.max_height = 1088,
+			.step_height = MPEG2_MB_DIM,
+		},
+	},
+};
+
 static irqreturn_t rk3288_vepu_irq(int irq, void *dev_id)
 {
 	struct rockchip_vpu_dev *vpu = dev_id;
@@ -74,6 +94,24 @@ static irqreturn_t rk3288_vepu_irq(int irq, void *dev_id)
 	return IRQ_HANDLED;
 }
 
+static irqreturn_t rk3288_vdpu_irq(int irq, void *dev_id)
+{
+	struct rockchip_vpu_dev *vpu = dev_id;
+	enum vb2_buffer_state state;
+	u32 status;
+
+	status = vdpu_read(vpu, VDPU_REG_INTERRUPT);
+	state = (status & VDPU_REG_INTERRUPT_DEC_RDY_INT) ?
+		VB2_BUF_STATE_DONE : VB2_BUF_STATE_ERROR;
+
+	vdpu_write(vpu, 0, VDPU_REG_INTERRUPT);
+	vdpu_write(vpu, VDPU_REG_CONFIG_DEC_CLK_GATE_E, VDPU_REG_CONFIG);
+
+	rockchip_vpu_irq_done(vpu, 0, state);
+
+	return IRQ_HANDLED;
+}
+
 static int rk3288_vpu_hw_init(struct rockchip_vpu_dev *vpu)
 {
 	/* Bump ACLK to max. possible freq. to improve performance. */
@@ -90,6 +128,15 @@ static void rk3288_vpu_enc_reset(struct rockchip_vpu_ctx *ctx)
 	vepu_write(vpu, 0, VEPU_REG_AXI_CTRL);
 }
 
+static void rk3288_vpu_dec_reset(struct rockchip_vpu_ctx *ctx)
+{
+	struct rockchip_vpu_dev *vpu = ctx->dev;
+
+	vdpu_write(vpu, VDPU_REG_INTERRUPT_DEC_IRQ_DIS, VDPU_REG_INTERRUPT);
+	vdpu_write(vpu, VDPU_REG_CONFIG_DEC_CLK_GATE_E, VDPU_REG_CONFIG);
+	vdpu_write(vpu, 1, VDPU_REG_SOFT_RESET);
+}
+
 /*
  * Supported codec ops.
  */
@@ -101,6 +148,12 @@ static const struct rockchip_vpu_codec_ops rk3288_vpu_codec_ops[] = {
 		.init = rockchip_vpu_jpeg_enc_init,
 		.exit = rockchip_vpu_jpeg_enc_exit,
 	},
+	[RK_VPU_MODE_MPEG2_DEC] = {
+		.run = rk3288_vpu_mpeg2_dec_run,
+		.reset = rk3288_vpu_dec_reset,
+		.init = rockchip_vpu_mpeg2_dec_init,
+		.exit = rockchip_vpu_mpeg2_dec_exit,
+	},
 };
 
 /*
@@ -111,9 +164,13 @@ const struct rockchip_vpu_variant rk3288_vpu_variant = {
 	.enc_offset = 0x0,
 	.enc_fmts = rk3288_vpu_enc_fmts,
 	.num_enc_fmts = ARRAY_SIZE(rk3288_vpu_enc_fmts),
+	.dec_offset = 0x400,
+	.dec_fmts = rk3288_vpu_dec_fmts,
+	.num_dec_fmts = ARRAY_SIZE(rk3288_vpu_dec_fmts),
+	.codec = RK_VPU_JPEG_ENCODER | RK_VPU_MPEG2_DECODER,
 	.codec_ops = rk3288_vpu_codec_ops,
-	.codec = RK_VPU_JPEG_ENCODER,
 	.vepu_irq = rk3288_vepu_irq,
+	.vdpu_irq = rk3288_vdpu_irq,
 	.init = rk3288_vpu_hw_init,
 	.clk_names = {"aclk", "hclk"},
 	.num_clocks = 2
diff --git a/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c
new file mode 100644
index 000000000000..e9eee47fcea1
--- /dev/null
+++ b/drivers/staging/media/rockchip/vpu/rk3288_vpu_hw_mpeg2_dec.c
@@ -0,0 +1,261 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Rockchip VPU codec driver
+ *
+ * Copyright (C) 2018 Rockchip Electronics Co., Ltd.
+ */
+
+#include <asm/unaligned.h>
+#include <linux/bitfield.h>
+#include <media/v4l2-mem2mem.h>
+#include "rockchip_vpu.h"
+#include "rockchip_vpu_hw.h"
+
+#define VDPU_SWREG(nr)			((nr) * 4)
+
+#define VDPU_REG_RLC_VLC_BASE		VDPU_SWREG(12)
+#define VDPU_REG_DEC_OUT_BASE		VDPU_SWREG(13)
+#define VDPU_REG_REFER0_BASE		VDPU_SWREG(14)
+#define VDPU_REG_REFER1_BASE		VDPU_SWREG(15)
+#define VDPU_REG_REFER2_BASE		VDPU_SWREG(16)
+#define VDPU_REG_REFER3_BASE		VDPU_SWREG(17)
+#define VDPU_REG_QTABLE_BASE		VDPU_SWREG(40)
+#define VDPU_REG_DEC_E(v)		((v) ? BIT(0) : 0)
+
+#define VDPU_REG_DEC_AXI_RD_ID(v)	(((v) << 24) & GENMASK(31, 24))
+#define VDPU_REG_DEC_TIMEOUT_E(v)	((v) ? BIT(23) : 0)
+#define VDPU_REG_DEC_STRSWAP32_E(v)	((v) ? BIT(22) : 0)
+#define VDPU_REG_DEC_STRENDIAN_E(v)	((v) ? BIT(21) : 0)
+#define VDPU_REG_DEC_INSWAP32_E(v)	((v) ? BIT(20) : 0)
+#define VDPU_REG_DEC_OUTSWAP32_E(v)	((v) ? BIT(19) : 0)
+#define VDPU_REG_DEC_DATA_DISC_E(v)	((v) ? BIT(18) : 0)
+#define VDPU_REG_DEC_LATENCY(v)		(((v) << 11) & GENMASK(16, 11))
+#define VDPU_REG_DEC_CLK_GATE_E(v)	((v) ? BIT(10) : 0)
+#define VDPU_REG_DEC_IN_ENDIAN(v)	((v) ? BIT(9) : 0)
+#define VDPU_REG_DEC_OUT_ENDIAN(v)	((v) ? BIT(8) : 0)
+#define VDPU_REG_DEC_ADV_PRE_DIS(v)	((v) ? BIT(6) : 0)
+#define VDPU_REG_DEC_SCMD_DIS(v)	((v) ? BIT(5) : 0)
+#define VDPU_REG_DEC_MAX_BURST(v)	(((v) << 0) & GENMASK(4, 0))
+
+#define VDPU_REG_DEC_MODE(v)		(((v) << 28) & GENMASK(31, 28))
+#define VDPU_REG_RLC_MODE_E(v)		((v) ? BIT(27) : 0)
+#define VDPU_REG_PIC_INTERLACE_E(v)	((v) ? BIT(23) : 0)
+#define VDPU_REG_PIC_FIELDMODE_E(v)	((v) ? BIT(22) : 0)
+#define VDPU_REG_PIC_B_E(v)		((v) ? BIT(21) : 0)
+#define VDPU_REG_PIC_INTER_E(v)		((v) ? BIT(20) : 0)
+#define VDPU_REG_PIC_TOPFIELD_E(v)	((v) ? BIT(19) : 0)
+#define VDPU_REG_FWD_INTERLACE_E(v)	((v) ? BIT(18) : 0)
+#define VDPU_REG_FILTERING_DIS(v)	((v) ? BIT(14) : 0)
+#define VDPU_REG_WRITE_MVS_E(v)		((v) ? BIT(12) : 0)
+#define VDPU_REG_DEC_AXI_WR_ID(v)	(((v) << 0) & GENMASK(7, 0))
+
+#define VDPU_REG_PIC_MB_WIDTH(v)	(((v) << 23) & GENMASK(31, 23))
+#define VDPU_REG_PIC_MB_HEIGHT_P(v)	(((v) << 11) & GENMASK(18, 11))
+#define VDPU_REG_ALT_SCAN_E(v)		((v) ? BIT(6) : 0)
+#define VDPU_REG_TOPFIELDFIRST_E(v)	((v) ? BIT(5) : 0)
+
+#define VDPU_REG_STRM_START_BIT(v)	(((v) << 26) & GENMASK(31, 26))
+#define VDPU_REG_QSCALE_TYPE(v)		((v) ? BIT(24) : 0)
+#define VDPU_REG_CON_MV_E(v)		((v) ? BIT(4) : 0)
+#define VDPU_REG_INTRA_DC_PREC(v)	(((v) << 2) & GENMASK(3, 2))
+#define VDPU_REG_INTRA_VLC_TAB(v)	((v) ? BIT(1) : 0)
+#define VDPU_REG_FRAME_PRED_DCT(v)	((v) ? BIT(0) : 0)
+
+#define VDPU_REG_INIT_QP(v)		(((v) << 25) & GENMASK(30, 25))
+#define VDPU_REG_STREAM_LEN(v)		(((v) << 0) & GENMASK(23, 0))
+
+#define VDPU_REG_ALT_SCAN_FLAG_E(v)	((v) ? BIT(19) : 0)
+#define VDPU_REG_FCODE_FWD_HOR(v)	(((v) << 15) & GENMASK(18, 15))
+#define VDPU_REG_FCODE_FWD_VER(v)	(((v) << 11) & GENMASK(14, 11))
+#define VDPU_REG_FCODE_BWD_HOR(v)	(((v) << 7) & GENMASK(10, 7))
+#define VDPU_REG_FCODE_BWD_VER(v)	(((v) << 3) & GENMASK(6, 3))
+#define VDPU_REG_MV_ACCURACY_FWD(v)	((v) ? BIT(2) : 0)
+#define VDPU_REG_MV_ACCURACY_BWD(v)	((v) ? BIT(1) : 0)
+
+#define VDPU_REG_STARTMB_X(v)		(((v) << 23) & GENMASK(31, 23))
+#define VDPU_REG_STARTMB_Y(v)		(((v) << 15) & GENMASK(22, 15))
+
+#define VDPU_REG_APF_THRESHOLD(v)	(((v) << 0) & GENMASK(13, 0))
+
+#define PICT_TOP_FIELD     1
+#define PICT_BOTTOM_FIELD  2
+#define PICT_FRAME         3
+
+static void
+rk3288_vpu_mpeg2_dec_set_quantization(struct rockchip_vpu_dev *vpu,
+				      struct rockchip_vpu_ctx *ctx)
+{
+	struct v4l2_ctrl_mpeg2_quantization *quantization;
+
+	quantization = rockchip_vpu_get_ctrl(ctx,
+				V4L2_CID_MPEG_VIDEO_MPEG2_QUANTIZATION);
+	rockchip_vpu_mpeg2_dec_copy_qtable(ctx->mpeg2_dec.qtable.cpu,
+					   quantization);
+	vdpu_write_relaxed(vpu, ctx->mpeg2_dec.qtable.dma,
+			   VDPU_REG_QTABLE_BASE);
+}
+
+static void
+rk3288_vpu_mpeg2_dec_set_buffers(struct rockchip_vpu_dev *vpu,
+				 struct rockchip_vpu_ctx *ctx,
+				 struct vb2_buffer *src_buf,
+				 struct vb2_buffer *dst_buf,
+				 const struct v4l2_mpeg2_sequence *sequence,
+				 const struct v4l2_mpeg2_picture *picture,
+				 const struct v4l2_ctrl_mpeg2_slice_params *slice_params)
+{
+	dma_addr_t forward_addr = 0, backward_addr = 0;
+	dma_addr_t current_addr, addr;
+	struct vb2_queue *vq;
+
+	vq = v4l2_m2m_get_dst_vq(ctx->fh.m2m_ctx);
+
+	switch (picture->picture_coding_type) {
+	case V4L2_MPEG2_PICTURE_CODING_TYPE_B:
+		backward_addr = rockchip_vpu_get_ref(vq,
+						slice_params->backward_ref_ts);
+		/* fall-through */
+	case V4L2_MPEG2_PICTURE_CODING_TYPE_P:
+		forward_addr = rockchip_vpu_get_ref(vq,
+						slice_params->forward_ref_ts);
+	}
+
+	/* Source bitstream buffer */
+	addr = vb2_dma_contig_plane_dma_addr(src_buf, 0);
+	vdpu_write_relaxed(vpu, addr, VDPU_REG_RLC_VLC_BASE);
+
+	/* Destination frame buffer */
+	addr = vb2_dma_contig_plane_dma_addr(dst_buf, 0);
+	current_addr = addr;
+
+	if (picture->picture_structure == PICT_BOTTOM_FIELD)
+		addr += ALIGN(ctx->dst_fmt.width, 16);
+	vdpu_write_relaxed(vpu, addr, VDPU_REG_DEC_OUT_BASE);
+
+	if (!forward_addr)
+		forward_addr = current_addr;
+	if (!backward_addr)
+		backward_addr = current_addr;
+
+	/* Set forward ref frame (top/bottom field) */
+	if (picture->picture_structure == PICT_FRAME ||
+	    picture->picture_coding_type == V4L2_MPEG2_PICTURE_CODING_TYPE_B ||
+	    (picture->picture_structure == PICT_TOP_FIELD &&
+	     picture->top_field_first) ||
+	    (picture->picture_structure == PICT_BOTTOM_FIELD &&
+	     !picture->top_field_first)) {
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER0_BASE);
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER1_BASE);
+	} else if (picture->picture_structure == PICT_TOP_FIELD) {
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER0_BASE);
+		vdpu_write_relaxed(vpu, current_addr, VDPU_REG_REFER1_BASE);
+	} else if (picture->picture_structure == PICT_BOTTOM_FIELD) {
+		vdpu_write_relaxed(vpu, current_addr, VDPU_REG_REFER0_BASE);
+		vdpu_write_relaxed(vpu, forward_addr, VDPU_REG_REFER1_BASE);
+	}
+
+	/* Set backward ref frame (top/bottom field) */
+	vdpu_write_relaxed(vpu, backward_addr, VDPU_REG_REFER2_BASE);
+	vdpu_write_relaxed(vpu, backward_addr, VDPU_REG_REFER3_BASE);
+}
+
+void rk3288_vpu_mpeg2_dec_run(struct rockchip_vpu_ctx *ctx)
+{
+	struct rockchip_vpu_dev *vpu = ctx->dev;
+	struct vb2_v4l2_buffer *src_buf, *dst_buf;
+	const struct v4l2_ctrl_mpeg2_slice_params *slice_params;
+	const struct v4l2_mpeg2_sequence *sequence;
+	const struct v4l2_mpeg2_picture *picture;
+	u32 reg;
+
+	src_buf = v4l2_m2m_next_src_buf(ctx->fh.m2m_ctx);
+	dst_buf = v4l2_m2m_next_dst_buf(ctx->fh.m2m_ctx);
+
+	/* Apply request controls if any */
+	v4l2_ctrl_request_setup(src_buf->vb2_buf.req_obj.req,
+				&ctx->ctrl_handler);
+
+	slice_params = rockchip_vpu_get_ctrl(ctx,
+				V4L2_CID_MPEG_VIDEO_MPEG2_SLICE_PARAMS);
+	sequence = &slice_params->sequence;
+	picture = &slice_params->picture;
+
+	reg = VDPU_REG_DEC_AXI_RD_ID(0) |
+	      VDPU_REG_DEC_TIMEOUT_E(1) |
+	      VDPU_REG_DEC_STRSWAP32_E(1) |
+	      VDPU_REG_DEC_STRENDIAN_E(1) |
+	      VDPU_REG_DEC_INSWAP32_E(1) |
+	      VDPU_REG_DEC_OUTSWAP32_E(1) |
+	      VDPU_REG_DEC_DATA_DISC_E(0) |
+	      VDPU_REG_DEC_LATENCY(0) |
+	      VDPU_REG_DEC_CLK_GATE_E(1) |
+	      VDPU_REG_DEC_IN_ENDIAN(1) |
+	      VDPU_REG_DEC_OUT_ENDIAN(1) |
+	      VDPU_REG_DEC_ADV_PRE_DIS(0) |
+	      VDPU_REG_DEC_SCMD_DIS(0) |
+	      VDPU_REG_DEC_MAX_BURST(16);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(2));
+
+	reg = VDPU_REG_DEC_MODE(5) |
+	      VDPU_REG_RLC_MODE_E(0) |
+	      VDPU_REG_PIC_INTERLACE_E(!sequence->progressive_sequence) |
+	      VDPU_REG_PIC_FIELDMODE_E(picture->picture_structure != PICT_FRAME) |
+	      VDPU_REG_PIC_B_E(picture->picture_coding_type == V4L2_MPEG2_PICTURE_CODING_TYPE_B) |
+	      VDPU_REG_PIC_INTER_E(picture->picture_coding_type != V4L2_MPEG2_PICTURE_CODING_TYPE_I) |
+	      VDPU_REG_PIC_TOPFIELD_E(picture->picture_structure == PICT_TOP_FIELD) |
+	      VDPU_REG_FWD_INTERLACE_E(0) |
+	      VDPU_REG_FILTERING_DIS(1) |
+	      VDPU_REG_WRITE_MVS_E(0) |
+	      VDPU_REG_DEC_AXI_WR_ID(0);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(3));
+
+	reg = VDPU_REG_PIC_MB_WIDTH(MPEG2_MB_WIDTH(ctx->dst_fmt.width)) |
+	      VDPU_REG_PIC_MB_HEIGHT_P(MPEG2_MB_HEIGHT(ctx->dst_fmt.height)) |
+	      VDPU_REG_ALT_SCAN_E(picture->alternate_scan) |
+	      VDPU_REG_TOPFIELDFIRST_E(picture->top_field_first);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(4));
+
+	reg = VDPU_REG_STRM_START_BIT(slice_params->data_bit_offset) |
+	      VDPU_REG_QSCALE_TYPE(picture->q_scale_type) |
+	      VDPU_REG_CON_MV_E(picture->concealment_motion_vectors) |
+	      VDPU_REG_INTRA_DC_PREC(picture->intra_dc_precision) |
+	      VDPU_REG_INTRA_VLC_TAB(picture->intra_vlc_format) |
+	      VDPU_REG_FRAME_PRED_DCT(picture->frame_pred_frame_dct);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(5));
+
+	reg = VDPU_REG_INIT_QP(1) |
+	      VDPU_REG_STREAM_LEN(slice_params->bit_size >> 3);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(6));
+
+	reg = VDPU_REG_ALT_SCAN_FLAG_E(picture->alternate_scan) |
+	      VDPU_REG_FCODE_FWD_HOR(picture->f_code[0][0]) |
+	      VDPU_REG_FCODE_FWD_VER(picture->f_code[0][1]) |
+	      VDPU_REG_FCODE_BWD_HOR(picture->f_code[1][0]) |
+	      VDPU_REG_FCODE_BWD_VER(picture->f_code[1][1]) |
+	      VDPU_REG_MV_ACCURACY_FWD(1) |
+	      VDPU_REG_MV_ACCURACY_BWD(1);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(18));
+
+	reg = VDPU_REG_STARTMB_X(0) |
+	      VDPU_REG_STARTMB_Y(0);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(48));
+
+	reg = VDPU_REG_APF_THRESHOLD(8);
+	vdpu_write_relaxed(vpu, reg, VDPU_SWREG(55));
+
+	rk3288_vpu_mpeg2_dec_set_quantization(vpu, ctx);
+
+	rk3288_vpu_mpeg2_dec_set_buffers(vpu, ctx, &src_buf->vb2_buf,
+					 &dst_buf->vb2_buf,
+					 sequence, picture, slice_params);
+
+	/* Controls no longer in-use, we can complete them */
+	v4l2_ctrl_request_complete(src_buf->vb2_buf.req_obj.req,
+				   &ctx->ctrl_handler);
+
+	/* Kick the watchdog and start decoding */
+	schedule_delayed_work(&vpu->watchdog_work, msecs_to_jiffies(2000));
+
+	reg = VDPU_REG_DEC_E(1);
+	vdpu_write(vpu, reg, VDPU_SWREG(1));
+}
diff --git a/drivers/staging/media/rockchip/vpu/rk3288_vpu_regs.h b/drivers/staging/media/rockchip/vpu/rk3288_vpu_regs.h
index 9d0b9bdf3297..c9631b713804 100644
--- a/drivers/staging/media/rockchip/vpu/rk3288_vpu_regs.h
+++ b/drivers/staging/media/rockchip/vpu/rk3288_vpu_regs.h
@@ -438,5 +438,6 @@
 #define     VDPU_REG_REF_BUF_CTRL2_REFBU2_THR(x)	(((x) & 0xfff) << 19)
 #define     VDPU_REG_REF_BUF_CTRL2_REFBU2_PICID(x)	(((x) & 0x1f) << 14)
 #define     VDPU_REG_REF_BUF_CTRL2_APF_THRESHOLD(x)	(((x) & 0x3fff) << 0)
+#define VDPU_REG_SOFT_RESET			0x194
 
 #endif /* RK3288_VPU_REGS_H_ */
diff --git a/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h b/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
index 4268c3a8b924..6cecb528f994 100644
--- a/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
+++ b/drivers/staging/media/rockchip/vpu/rockchip_vpu_hw.h
@@ -92,6 +92,7 @@ void rk3399_vpu_jpeg_enc_run(struct rockchip_vpu_ctx *ctx);
 int rockchip_vpu_jpeg_enc_init(struct rockchip_vpu_ctx *ctx);
 void rockchip_vpu_jpeg_enc_exit(struct rockchip_vpu_ctx *ctx);
 
+void rk3288_vpu_mpeg2_dec_run(struct rockchip_vpu_ctx *ctx);
 void rk3399_vpu_mpeg2_dec_run(struct rockchip_vpu_ctx *ctx);
 void rockchip_vpu_mpeg2_dec_copy_qtable(u8 *qtable,
 	const struct v4l2_ctrl_mpeg2_quantization *ctrl);
-- 
2.20.1

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Hans Verkuil]

On 5/3/19 1:47 PM, Boris Brezillon wrote:
> Users can define custom sizeimage and bytesperline as long as they're
> big enough to store the amount of pixels required for a specific
> width/height under a specific format. Avoid overriding those fields in
> this case.
> 
> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
> ---
> Hello Hans,
> 
> The sizeimage/bytesperline check on !MPLANE formats is still not 100%
> sure, as custom bytesperline might induce bigger sizeimage than what
> we calculate.
> 
> I tried implementing something smarter taking the per-component plane
> bpp + hdiv param as we discussed the other day but decided to step
> back after realizing the per-component plane macro block might also
> differ at least in theory (not sure that's true in practice) and that
> has an impact on bytesperline too.
> 
> Let me know how you want to handle that case.
> 
> Regards,
> 
> Boris
> 
> Changes in v5:
> * New patch
> ---
>  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
>  1 file changed, 39 insertions(+), 15 deletions(-)
> 
> diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
> index 3c6f5c115fc5..37bfc984a8b5 100644
> --- a/drivers/media/v4l2-core/v4l2-common.c
> +++ b/drivers/media/v4l2-core/v4l2-common.c
> @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>  	pixfmt->num_planes = info->mem_planes;
>  
>  	if (info->mem_planes == 1) {
> +		u32 bytesperline, sizeimage = 0;
> +
>  		plane = &pixfmt->plane_fmt[0];
> -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> -		plane->sizeimage = 0;
> +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>  
>  		for (i = 0; i < info->comp_planes; i++) {
>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>  
> -			plane->sizeimage += info->bpp[i] *
> -				DIV_ROUND_UP(aligned_width, hdiv) *
> -				DIV_ROUND_UP(aligned_height, vdiv);
> +			sizeimage += info->bpp[i] *
> +				     DIV_ROUND_UP(aligned_width, hdiv) *
> +				     DIV_ROUND_UP(aligned_height, vdiv);
>  		}
> +
> +		/*
> +		 * The user might have specified custom sizeimage/bytesperline,
> +		 * only override them if they're not big enough.
> +		 */
> +		plane->sizeimage = max(sizeimage, plane->sizeimage);
> +		plane->bytesperline = max(bytesperline, plane->bytesperline);

Let's just set bytesperline, ignoring the value the user supplied. There are very
few drivers that allow the user to set bytesperline anyway, so it's not a big deal
to drop support for that for now. We can add it back later.

Just add a comment that a user-defined bytesperline value isn't currently supported.

Regards,

	Hans

>  	} else {
>  		for (i = 0; i < info->comp_planes; i++) {
>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> @@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>  
>  			plane = &pixfmt->plane_fmt[i];
> -			plane->bytesperline =
> -				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
> -			plane->sizeimage =
> -				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
> +
> +			/*
> +			 * The user might have specified custom
> +			 * sizeimage/bytesperline, only override them if
> +			 * they're not big enough.
> +			 */
> +			plane->bytesperline = max_t(u32,
> +						    info->bpp[i] *
> +						    DIV_ROUND_UP(aligned_width, hdiv),
> +						    plane->bytesperline);
> +			plane->sizeimage = max_t(u32,
> +						 plane->bytesperline *
> +						 DIV_ROUND_UP(aligned_height, vdiv),
> +						 plane->sizeimage);
>  		}
>  	}
>  	return 0;
> @@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>  		     u32 width, u32 height)
>  {
>  	const struct v4l2_format_info *info;
> +	u32 bytesperline, sizeimage = 0;
>  	int i;
>  
>  	info = v4l2_format_info(pixelformat);
> @@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>  	pixfmt->width = width;
>  	pixfmt->height = height;
>  	pixfmt->pixelformat = pixelformat;
> -	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> -	pixfmt->sizeimage = 0;
> +	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>  
>  	for (i = 0; i < info->comp_planes; i++) {
>  		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> @@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>  
>  		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>  		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> -
> -		pixfmt->sizeimage += info->bpp[i] *
> -			DIV_ROUND_UP(aligned_width, hdiv) *
> -			DIV_ROUND_UP(aligned_height, vdiv);
> +		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
> +			     DIV_ROUND_UP(aligned_height, vdiv);
>  	}
> +
> +	/*
> +	 * The user might have specified its own sizeimage/bytesperline values,
> +	 * only override them if they're not big enough.
> +	 */
> +	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
> +	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
> +
>  	return 0;
>  }
>  EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);
> 

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Laurent Pinchart]

Hi Hans,

(CC'ing Kieran)

On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
> On 5/3/19 1:47 PM, Boris Brezillon wrote:
> > Users can define custom sizeimage and bytesperline as long as they're
> > big enough to store the amount of pixels required for a specific
> > width/height under a specific format. Avoid overriding those fields in
> > this case.
> > 
> > Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
> > ---
> > Hello Hans,
> > 
> > The sizeimage/bytesperline check on !MPLANE formats is still not 100%
> > sure, as custom bytesperline might induce bigger sizeimage than what
> > we calculate.
> > 
> > I tried implementing something smarter taking the per-component plane
> > bpp + hdiv param as we discussed the other day but decided to step
> > back after realizing the per-component plane macro block might also
> > differ at least in theory (not sure that's true in practice) and that
> > has an impact on bytesperline too.
> > 
> > Let me know how you want to handle that case.
> > 
> > Regards,
> > 
> > Boris
> > 
> > Changes in v5:
> > * New patch
> > ---
> >  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
> >  1 file changed, 39 insertions(+), 15 deletions(-)
> > 
> > diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
> > index 3c6f5c115fc5..37bfc984a8b5 100644
> > --- a/drivers/media/v4l2-core/v4l2-common.c
> > +++ b/drivers/media/v4l2-core/v4l2-common.c
> > @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> >  	pixfmt->num_planes = info->mem_planes;
> >  
> >  	if (info->mem_planes == 1) {
> > +		u32 bytesperline, sizeimage = 0;
> > +
> >  		plane = &pixfmt->plane_fmt[0];
> > -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > -		plane->sizeimage = 0;
> > +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> >  
> >  		for (i = 0; i < info->comp_planes; i++) {
> >  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> >  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> >  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> >  
> > -			plane->sizeimage += info->bpp[i] *
> > -				DIV_ROUND_UP(aligned_width, hdiv) *
> > -				DIV_ROUND_UP(aligned_height, vdiv);
> > +			sizeimage += info->bpp[i] *
> > +				     DIV_ROUND_UP(aligned_width, hdiv) *
> > +				     DIV_ROUND_UP(aligned_height, vdiv);
> >  		}
> > +
> > +		/*
> > +		 * The user might have specified custom sizeimage/bytesperline,
> > +		 * only override them if they're not big enough.
> > +		 */
> > +		plane->sizeimage = max(sizeimage, plane->sizeimage);
> > +		plane->bytesperline = max(bytesperline, plane->bytesperline);
> 
> Let's just set bytesperline, ignoring the value the user supplied. There are very
> few drivers that allow the user to set bytesperline anyway, so it's not a big deal
> to drop support for that for now. We can add it back later.
> 
> Just add a comment that a user-defined bytesperline value isn't currently supported.

Kieran recently ran into an issue related to this, when sharing buffers
between a CSI-2 receiver and an ISP. The ISP has alignment constraints
in both the horizontal and vertical directions on the line stride and
total image size. Out software framework currently allocates buffers
from the CSI-2 receiver which doesn't have those constraints, and not
being able to specify sizeimage is thus a problem.

We have worked around this by allocating buffers on the ISP side but
that departed from the normal way of operation, and it may not be a good
idea to carry that forward.

I wonder how we should solve these issues long term. A central memory
allocator is probably the way to go.

> >  	} else {
> >  		for (i = 0; i < info->comp_planes; i++) {
> >  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > @@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> >  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> >  
> >  			plane = &pixfmt->plane_fmt[i];
> > -			plane->bytesperline =
> > -				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
> > -			plane->sizeimage =
> > -				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
> > +
> > +			/*
> > +			 * The user might have specified custom
> > +			 * sizeimage/bytesperline, only override them if
> > +			 * they're not big enough.
> > +			 */
> > +			plane->bytesperline = max_t(u32,
> > +						    info->bpp[i] *
> > +						    DIV_ROUND_UP(aligned_width, hdiv),
> > +						    plane->bytesperline);
> > +			plane->sizeimage = max_t(u32,
> > +						 plane->bytesperline *
> > +						 DIV_ROUND_UP(aligned_height, vdiv),
> > +						 plane->sizeimage);
> >  		}
> >  	}
> >  	return 0;
> > @@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> >  		     u32 width, u32 height)
> >  {
> >  	const struct v4l2_format_info *info;
> > +	u32 bytesperline, sizeimage = 0;
> >  	int i;
> >  
> >  	info = v4l2_format_info(pixelformat);
> > @@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> >  	pixfmt->width = width;
> >  	pixfmt->height = height;
> >  	pixfmt->pixelformat = pixelformat;
> > -	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > -	pixfmt->sizeimage = 0;
> > +	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> >  
> >  	for (i = 0; i < info->comp_planes; i++) {
> >  		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > @@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> >  
> >  		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> >  		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> > -
> > -		pixfmt->sizeimage += info->bpp[i] *
> > -			DIV_ROUND_UP(aligned_width, hdiv) *
> > -			DIV_ROUND_UP(aligned_height, vdiv);
> > +		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
> > +			     DIV_ROUND_UP(aligned_height, vdiv);
> >  	}
> > +
> > +	/*
> > +	 * The user might have specified its own sizeimage/bytesperline values,
> > +	 * only override them if they're not big enough.
> > +	 */
> > +	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
> > +	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
> > +
> >  	return 0;
> >  }
> >  EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);

-- 
Regards,

Laurent Pinchart

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Hans Verkuil]

On 5/10/19 1:28 PM, Laurent Pinchart wrote:
> Hi Hans,
> 
> (CC'ing Kieran)
> 
> On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
>> On 5/3/19 1:47 PM, Boris Brezillon wrote:
>>> Users can define custom sizeimage and bytesperline as long as they're
>>> big enough to store the amount of pixels required for a specific
>>> width/height under a specific format. Avoid overriding those fields in
>>> this case.
>>>
>>> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
>>> ---
>>> Hello Hans,
>>>
>>> The sizeimage/bytesperline check on !MPLANE formats is still not 100%
>>> sure, as custom bytesperline might induce bigger sizeimage than what
>>> we calculate.
>>>
>>> I tried implementing something smarter taking the per-component plane
>>> bpp + hdiv param as we discussed the other day but decided to step
>>> back after realizing the per-component plane macro block might also
>>> differ at least in theory (not sure that's true in practice) and that
>>> has an impact on bytesperline too.
>>>
>>> Let me know how you want to handle that case.
>>>
>>> Regards,
>>>
>>> Boris
>>>
>>> Changes in v5:
>>> * New patch
>>> ---
>>>  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
>>>  1 file changed, 39 insertions(+), 15 deletions(-)
>>>
>>> diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
>>> index 3c6f5c115fc5..37bfc984a8b5 100644
>>> --- a/drivers/media/v4l2-core/v4l2-common.c
>>> +++ b/drivers/media/v4l2-core/v4l2-common.c
>>> @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>  	pixfmt->num_planes = info->mem_planes;
>>>  
>>>  	if (info->mem_planes == 1) {
>>> +		u32 bytesperline, sizeimage = 0;
>>> +
>>>  		plane = &pixfmt->plane_fmt[0];
>>> -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>> -		plane->sizeimage = 0;
>>> +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>  
>>>  		for (i = 0; i < info->comp_planes; i++) {
>>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>> @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>>  
>>> -			plane->sizeimage += info->bpp[i] *
>>> -				DIV_ROUND_UP(aligned_width, hdiv) *
>>> -				DIV_ROUND_UP(aligned_height, vdiv);
>>> +			sizeimage += info->bpp[i] *
>>> +				     DIV_ROUND_UP(aligned_width, hdiv) *
>>> +				     DIV_ROUND_UP(aligned_height, vdiv);
>>>  		}
>>> +
>>> +		/*
>>> +		 * The user might have specified custom sizeimage/bytesperline,
>>> +		 * only override them if they're not big enough.
>>> +		 */
>>> +		plane->sizeimage = max(sizeimage, plane->sizeimage);
>>> +		plane->bytesperline = max(bytesperline, plane->bytesperline);
>>
>> Let's just set bytesperline, ignoring the value the user supplied. There are very
>> few drivers that allow the user to set bytesperline anyway, so it's not a big deal
>> to drop support for that for now. We can add it back later.
>>
>> Just add a comment that a user-defined bytesperline value isn't currently supported.
> 
> Kieran recently ran into an issue related to this, when sharing buffers
> between a CSI-2 receiver and an ISP. The ISP has alignment constraints
> in both the horizontal and vertical directions on the line stride and
> total image size. Out software framework currently allocates buffers
> from the CSI-2 receiver which doesn't have those constraints, and not
> being able to specify sizeimage is thus a problem.

Not being able to specify sizeimage where? From userspace? Sorry, I don't
quite understand the specific issue here.

Regards,

	Hans

> 
> We have worked around this by allocating buffers on the ISP side but
> that departed from the normal way of operation, and it may not be a good
> idea to carry that forward.
> 
> I wonder how we should solve these issues long term. A central memory
> allocator is probably the way to go.
> 
>>>  	} else {
>>>  		for (i = 0; i < info->comp_planes; i++) {
>>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>> @@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>>  
>>>  			plane = &pixfmt->plane_fmt[i];
>>> -			plane->bytesperline =
>>> -				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
>>> -			plane->sizeimage =
>>> -				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
>>> +
>>> +			/*
>>> +			 * The user might have specified custom
>>> +			 * sizeimage/bytesperline, only override them if
>>> +			 * they're not big enough.
>>> +			 */
>>> +			plane->bytesperline = max_t(u32,
>>> +						    info->bpp[i] *
>>> +						    DIV_ROUND_UP(aligned_width, hdiv),
>>> +						    plane->bytesperline);
>>> +			plane->sizeimage = max_t(u32,
>>> +						 plane->bytesperline *
>>> +						 DIV_ROUND_UP(aligned_height, vdiv),
>>> +						 plane->sizeimage);
>>>  		}
>>>  	}
>>>  	return 0;
>>> @@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>>>  		     u32 width, u32 height)
>>>  {
>>>  	const struct v4l2_format_info *info;
>>> +	u32 bytesperline, sizeimage = 0;
>>>  	int i;
>>>  
>>>  	info = v4l2_format_info(pixelformat);
>>> @@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>>>  	pixfmt->width = width;
>>>  	pixfmt->height = height;
>>>  	pixfmt->pixelformat = pixelformat;
>>> -	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>> -	pixfmt->sizeimage = 0;
>>> +	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>  
>>>  	for (i = 0; i < info->comp_planes; i++) {
>>>  		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>> @@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>>>  
>>>  		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>>>  		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>> -
>>> -		pixfmt->sizeimage += info->bpp[i] *
>>> -			DIV_ROUND_UP(aligned_width, hdiv) *
>>> -			DIV_ROUND_UP(aligned_height, vdiv);
>>> +		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
>>> +			     DIV_ROUND_UP(aligned_height, vdiv);
>>>  	}
>>> +
>>> +	/*
>>> +	 * The user might have specified its own sizeimage/bytesperline values,
>>> +	 * only override them if they're not big enough.
>>> +	 */
>>> +	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
>>> +	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
>>> +
>>>  	return 0;
>>>  }
>>>  EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);
> 

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Laurent Pinchart]

Hi Hans,

On Fri, May 10, 2019 at 02:17:32PM +0200, Hans Verkuil wrote:
> On 5/10/19 1:28 PM, Laurent Pinchart wrote:
> > On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
> >> On 5/3/19 1:47 PM, Boris Brezillon wrote:
> >>> Users can define custom sizeimage and bytesperline as long as they're
> >>> big enough to store the amount of pixels required for a specific
> >>> width/height under a specific format. Avoid overriding those fields in
> >>> this case.
> >>>
> >>> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
> >>> ---
> >>> Hello Hans,
> >>>
> >>> The sizeimage/bytesperline check on !MPLANE formats is still not 100%
> >>> sure, as custom bytesperline might induce bigger sizeimage than what
> >>> we calculate.
> >>>
> >>> I tried implementing something smarter taking the per-component plane
> >>> bpp + hdiv param as we discussed the other day but decided to step
> >>> back after realizing the per-component plane macro block might also
> >>> differ at least in theory (not sure that's true in practice) and that
> >>> has an impact on bytesperline too.
> >>>
> >>> Let me know how you want to handle that case.
> >>>
> >>> Regards,
> >>>
> >>> Boris
> >>>
> >>> Changes in v5:
> >>> * New patch
> >>> ---
> >>>  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
> >>>  1 file changed, 39 insertions(+), 15 deletions(-)
> >>>
> >>> diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
> >>> index 3c6f5c115fc5..37bfc984a8b5 100644
> >>> --- a/drivers/media/v4l2-core/v4l2-common.c
> >>> +++ b/drivers/media/v4l2-core/v4l2-common.c
> >>> @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> >>>  	pixfmt->num_planes = info->mem_planes;
> >>>  
> >>>  	if (info->mem_planes == 1) {
> >>> +		u32 bytesperline, sizeimage = 0;
> >>> +
> >>>  		plane = &pixfmt->plane_fmt[0];
> >>> -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> >>> -		plane->sizeimage = 0;
> >>> +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> >>>  
> >>>  		for (i = 0; i < info->comp_planes; i++) {
> >>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> >>> @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> >>>  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> >>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> >>>  
> >>> -			plane->sizeimage += info->bpp[i] *
> >>> -				DIV_ROUND_UP(aligned_width, hdiv) *
> >>> -				DIV_ROUND_UP(aligned_height, vdiv);
> >>> +			sizeimage += info->bpp[i] *
> >>> +				     DIV_ROUND_UP(aligned_width, hdiv) *
> >>> +				     DIV_ROUND_UP(aligned_height, vdiv);
> >>>  		}
> >>> +
> >>> +		/*
> >>> +		 * The user might have specified custom sizeimage/bytesperline,
> >>> +		 * only override them if they're not big enough.
> >>> +		 */
> >>> +		plane->sizeimage = max(sizeimage, plane->sizeimage);
> >>> +		plane->bytesperline = max(bytesperline, plane->bytesperline);
> >>
> >> Let's just set bytesperline, ignoring the value the user supplied. There are very
> >> few drivers that allow the user to set bytesperline anyway, so it's not a big deal
> >> to drop support for that for now. We can add it back later.
> >>
> >> Just add a comment that a user-defined bytesperline value isn't currently supported.
> > 
> > Kieran recently ran into an issue related to this, when sharing buffers
> > between a CSI-2 receiver and an ISP. The ISP has alignment constraints
> > in both the horizontal and vertical directions on the line stride and
> > total image size. Out software framework currently allocates buffers
> > from the CSI-2 receiver which doesn't have those constraints, and not
> > being able to specify sizeimage is thus a problem.
> 
> Not being able to specify sizeimage where? From userspace? Sorry, I don't
> quite understand the specific issue here.

Yes, from userspace.

> > We have worked around this by allocating buffers on the ISP side but
> > that departed from the normal way of operation, and it may not be a good
> > idea to carry that forward.
> > 
> > I wonder how we should solve these issues long term. A central memory
> > allocator is probably the way to go.
> > 
> >>>  	} else {
> >>>  		for (i = 0; i < info->comp_planes; i++) {
> >>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> >>> @@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> >>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> >>>  
> >>>  			plane = &pixfmt->plane_fmt[i];
> >>> -			plane->bytesperline =
> >>> -				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
> >>> -			plane->sizeimage =
> >>> -				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
> >>> +
> >>> +			/*
> >>> +			 * The user might have specified custom
> >>> +			 * sizeimage/bytesperline, only override them if
> >>> +			 * they're not big enough.
> >>> +			 */
> >>> +			plane->bytesperline = max_t(u32,
> >>> +						    info->bpp[i] *
> >>> +						    DIV_ROUND_UP(aligned_width, hdiv),
> >>> +						    plane->bytesperline);
> >>> +			plane->sizeimage = max_t(u32,
> >>> +						 plane->bytesperline *
> >>> +						 DIV_ROUND_UP(aligned_height, vdiv),
> >>> +						 plane->sizeimage);
> >>>  		}
> >>>  	}
> >>>  	return 0;
> >>> @@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> >>>  		     u32 width, u32 height)
> >>>  {
> >>>  	const struct v4l2_format_info *info;
> >>> +	u32 bytesperline, sizeimage = 0;
> >>>  	int i;
> >>>  
> >>>  	info = v4l2_format_info(pixelformat);
> >>> @@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> >>>  	pixfmt->width = width;
> >>>  	pixfmt->height = height;
> >>>  	pixfmt->pixelformat = pixelformat;
> >>> -	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> >>> -	pixfmt->sizeimage = 0;
> >>> +	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> >>>  
> >>>  	for (i = 0; i < info->comp_planes; i++) {
> >>>  		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> >>> @@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> >>>  
> >>>  		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> >>>  		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> >>> -
> >>> -		pixfmt->sizeimage += info->bpp[i] *
> >>> -			DIV_ROUND_UP(aligned_width, hdiv) *
> >>> -			DIV_ROUND_UP(aligned_height, vdiv);
> >>> +		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
> >>> +			     DIV_ROUND_UP(aligned_height, vdiv);
> >>>  	}
> >>> +
> >>> +	/*
> >>> +	 * The user might have specified its own sizeimage/bytesperline values,
> >>> +	 * only override them if they're not big enough.
> >>> +	 */
> >>> +	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
> >>> +	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
> >>> +
> >>>  	return 0;
> >>>  }
> >>>  EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);

-- 
Regards,

Laurent Pinchart

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Hans Verkuil]

On 5/10/19 2:24 PM, Laurent Pinchart wrote:
> Hi Hans,
> 
> On Fri, May 10, 2019 at 02:17:32PM +0200, Hans Verkuil wrote:
>> On 5/10/19 1:28 PM, Laurent Pinchart wrote:
>>> On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
>>>> On 5/3/19 1:47 PM, Boris Brezillon wrote:
>>>>> Users can define custom sizeimage and bytesperline as long as they're
>>>>> big enough to store the amount of pixels required for a specific
>>>>> width/height under a specific format. Avoid overriding those fields in
>>>>> this case.
>>>>>
>>>>> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
>>>>> ---
>>>>> Hello Hans,
>>>>>
>>>>> The sizeimage/bytesperline check on !MPLANE formats is still not 100%
>>>>> sure, as custom bytesperline might induce bigger sizeimage than what
>>>>> we calculate.
>>>>>
>>>>> I tried implementing something smarter taking the per-component plane
>>>>> bpp + hdiv param as we discussed the other day but decided to step
>>>>> back after realizing the per-component plane macro block might also
>>>>> differ at least in theory (not sure that's true in practice) and that
>>>>> has an impact on bytesperline too.
>>>>>
>>>>> Let me know how you want to handle that case.
>>>>>
>>>>> Regards,
>>>>>
>>>>> Boris
>>>>>
>>>>> Changes in v5:
>>>>> * New patch
>>>>> ---
>>>>>  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
>>>>>  1 file changed, 39 insertions(+), 15 deletions(-)
>>>>>
>>>>> diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
>>>>> index 3c6f5c115fc5..37bfc984a8b5 100644
>>>>> --- a/drivers/media/v4l2-core/v4l2-common.c
>>>>> +++ b/drivers/media/v4l2-core/v4l2-common.c
>>>>> @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>>>  	pixfmt->num_planes = info->mem_planes;
>>>>>  
>>>>>  	if (info->mem_planes == 1) {
>>>>> +		u32 bytesperline, sizeimage = 0;
>>>>> +
>>>>>  		plane = &pixfmt->plane_fmt[0];
>>>>> -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>>> -		plane->sizeimage = 0;
>>>>> +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>>>  
>>>>>  		for (i = 0; i < info->comp_planes; i++) {
>>>>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>>>> @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>>>  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>>>>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>>>>  
>>>>> -			plane->sizeimage += info->bpp[i] *
>>>>> -				DIV_ROUND_UP(aligned_width, hdiv) *
>>>>> -				DIV_ROUND_UP(aligned_height, vdiv);
>>>>> +			sizeimage += info->bpp[i] *
>>>>> +				     DIV_ROUND_UP(aligned_width, hdiv) *
>>>>> +				     DIV_ROUND_UP(aligned_height, vdiv);
>>>>>  		}
>>>>> +
>>>>> +		/*
>>>>> +		 * The user might have specified custom sizeimage/bytesperline,
>>>>> +		 * only override them if they're not big enough.
>>>>> +		 */
>>>>> +		plane->sizeimage = max(sizeimage, plane->sizeimage);
>>>>> +		plane->bytesperline = max(bytesperline, plane->bytesperline);
>>>>
>>>> Let's just set bytesperline, ignoring the value the user supplied. There are very
>>>> few drivers that allow the user to set bytesperline anyway, so it's not a big deal
>>>> to drop support for that for now. We can add it back later.
>>>>
>>>> Just add a comment that a user-defined bytesperline value isn't currently supported.
>>>
>>> Kieran recently ran into an issue related to this, when sharing buffers
>>> between a CSI-2 receiver and an ISP. The ISP has alignment constraints
>>> in both the horizontal and vertical directions on the line stride and
>>> total image size. Out software framework currently allocates buffers
>>> from the CSI-2 receiver which doesn't have those constraints, and not
>>> being able to specify sizeimage is thus a problem.
>>
>> Not being able to specify sizeimage where? From userspace? Sorry, I don't
>> quite understand the specific issue here.
> 
> Yes, from userspace.

Ah, OK. But why not use CREATEBUFS? You can provide your own size when allocating
the buffers.

Also note this patch: https://patchwork.linuxtv.org/patch/55656/

Although this is specific for compressed formats.

Regards,

	Hans

> 
>>> We have worked around this by allocating buffers on the ISP side but
>>> that departed from the normal way of operation, and it may not be a good
>>> idea to carry that forward.
>>>
>>> I wonder how we should solve these issues long term. A central memory
>>> allocator is probably the way to go.
>>>
>>>>>  	} else {
>>>>>  		for (i = 0; i < info->comp_planes; i++) {
>>>>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>>>> @@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>>>>  
>>>>>  			plane = &pixfmt->plane_fmt[i];
>>>>> -			plane->bytesperline =
>>>>> -				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
>>>>> -			plane->sizeimage =
>>>>> -				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
>>>>> +
>>>>> +			/*
>>>>> +			 * The user might have specified custom
>>>>> +			 * sizeimage/bytesperline, only override them if
>>>>> +			 * they're not big enough.
>>>>> +			 */
>>>>> +			plane->bytesperline = max_t(u32,
>>>>> +						    info->bpp[i] *
>>>>> +						    DIV_ROUND_UP(aligned_width, hdiv),
>>>>> +						    plane->bytesperline);
>>>>> +			plane->sizeimage = max_t(u32,
>>>>> +						 plane->bytesperline *
>>>>> +						 DIV_ROUND_UP(aligned_height, vdiv),
>>>>> +						 plane->sizeimage);
>>>>>  		}
>>>>>  	}
>>>>>  	return 0;
>>>>> @@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>>>>>  		     u32 width, u32 height)
>>>>>  {
>>>>>  	const struct v4l2_format_info *info;
>>>>> +	u32 bytesperline, sizeimage = 0;
>>>>>  	int i;
>>>>>  
>>>>>  	info = v4l2_format_info(pixelformat);
>>>>> @@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>>>>>  	pixfmt->width = width;
>>>>>  	pixfmt->height = height;
>>>>>  	pixfmt->pixelformat = pixelformat;
>>>>> -	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>>> -	pixfmt->sizeimage = 0;
>>>>> +	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>>>  
>>>>>  	for (i = 0; i < info->comp_planes; i++) {
>>>>>  		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>>>> @@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
>>>>>  
>>>>>  		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>>>>>  		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>>>> -
>>>>> -		pixfmt->sizeimage += info->bpp[i] *
>>>>> -			DIV_ROUND_UP(aligned_width, hdiv) *
>>>>> -			DIV_ROUND_UP(aligned_height, vdiv);
>>>>> +		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
>>>>> +			     DIV_ROUND_UP(aligned_height, vdiv);
>>>>>  	}
>>>>> +
>>>>> +	/*
>>>>> +	 * The user might have specified its own sizeimage/bytesperline values,
>>>>> +	 * only override them if they're not big enough.
>>>>> +	 */
>>>>> +	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
>>>>> +	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
>>>>> +
>>>>>  	return 0;
>>>>>  }
>>>>>  EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);
> 

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Nicolas Dufresne]

Le vendredi 10 mai 2019 à 14:30 +0200, Hans Verkuil a écrit :
> On 5/10/19 2:24 PM, Laurent Pinchart wrote:
> > Hi Hans,
> > 
> > On Fri, May 10, 2019 at 02:17:32PM +0200, Hans Verkuil wrote:
> > > On 5/10/19 1:28 PM, Laurent Pinchart wrote:
> > > > On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
> > > > > On 5/3/19 1:47 PM, Boris Brezillon wrote:
> > > > > > Users can define custom sizeimage and bytesperline as long as they're
> > > > > > big enough to store the amount of pixels required for a specific
> > > > > > width/height under a specific format. Avoid overriding those fields in
> > > > > > this case.
> > > > > > 
> > > > > > Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
> > > > > > ---
> > > > > > Hello Hans,
> > > > > > 
> > > > > > The sizeimage/bytesperline check on !MPLANE formats is still not 100%
> > > > > > sure, as custom bytesperline might induce bigger sizeimage than what
> > > > > > we calculate.
> > > > > > 
> > > > > > I tried implementing something smarter taking the per-component plane
> > > > > > bpp + hdiv param as we discussed the other day but decided to step
> > > > > > back after realizing the per-component plane macro block might also
> > > > > > differ at least in theory (not sure that's true in practice) and that
> > > > > > has an impact on bytesperline too.
> > > > > > 
> > > > > > Let me know how you want to handle that case.
> > > > > > 
> > > > > > Regards,
> > > > > > 
> > > > > > Boris
> > > > > > 
> > > > > > Changes in v5:
> > > > > > * New patch
> > > > > > ---
> > > > > >  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
> > > > > >  1 file changed, 39 insertions(+), 15 deletions(-)
> > > > > > 
> > > > > > diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
> > > > > > index 3c6f5c115fc5..37bfc984a8b5 100644
> > > > > > --- a/drivers/media/v4l2-core/v4l2-common.c
> > > > > > +++ b/drivers/media/v4l2-core/v4l2-common.c
> > > > > > @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> > > > > >  	pixfmt->num_planes = info->mem_planes;
> > > > > >  
> > > > > >  	if (info->mem_planes == 1) {
> > > > > > +		u32 bytesperline, sizeimage = 0;
> > > > > > +
> > > > > >  		plane = &pixfmt->plane_fmt[0];
> > > > > > -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > > > > > -		plane->sizeimage = 0;
> > > > > > +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > > > > >  
> > > > > >  		for (i = 0; i < info->comp_planes; i++) {
> > > > > >  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > > > > > @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> > > > > >  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> > > > > >  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> > > > > >  
> > > > > > -			plane->sizeimage += info->bpp[i] *
> > > > > > -				DIV_ROUND_UP(aligned_width, hdiv) *
> > > > > > -				DIV_ROUND_UP(aligned_height, vdiv);
> > > > > > +			sizeimage += info->bpp[i] *
> > > > > > +				     DIV_ROUND_UP(aligned_width, hdiv) *
> > > > > > +				     DIV_ROUND_UP(aligned_height, vdiv);
> > > > > >  		}
> > > > > > +
> > > > > > +		/*
> > > > > > +		 * The user might have specified custom sizeimage/bytesperline,
> > > > > > +		 * only override them if they're not big enough.
> > > > > > +		 */
> > > > > > +		plane->sizeimage = max(sizeimage, plane->sizeimage);
> > > > > > +		plane->bytesperline = max(bytesperline, plane->bytesperline);
> > > > > 
> > > > > Let's just set bytesperline, ignoring the value the user supplied. There are very
> > > > > few drivers that allow the user to set bytesperline anyway, so it's not a big deal
> > > > > to drop support for that for now. We can add it back later.
> > > > > 
> > > > > Just add a comment that a user-defined bytesperline value isn't currently supported.
> > > > 
> > > > Kieran recently ran into an issue related to this, when sharing buffers
> > > > between a CSI-2 receiver and an ISP. The ISP has alignment constraints
> > > > in both the horizontal and vertical directions on the line stride and
> > > > total image size. Out software framework currently allocates buffers
> > > > from the CSI-2 receiver which doesn't have those constraints, and not
> > > > being able to specify sizeimage is thus a problem.
> > > 
> > > Not being able to specify sizeimage where? From userspace? Sorry, I don't
> > > quite understand the specific issue here.
> > 
> > Yes, from userspace.
> 
> Ah, OK. But why not use CREATEBUFS? You can provide your own size when allocating
> the buffers.
> 
> Also note this patch: https://patchwork.linuxtv.org/patch/55656/
> 
> Although this is specific for compressed formats.

While this work for compressed formats, it does not do anything for raw
image horizontal padding. The importation bit of V4L2 is pretty
difficult, so if we add helpers, we should load the way and simplify
things for userspace rather then enforcing the existing difficulty.

I think from now own we should design with the mindset that a DMABuf
that cannot be imported back into another driver due to software
limitations is a useless waste of FD.

Nicolas

> 
> Regards,
> 
> 	Hans
> 
> > > > We have worked around this by allocating buffers on the ISP side but
> > > > that departed from the normal way of operation, and it may not be a good
> > > > idea to carry that forward.
> > > > 
> > > > I wonder how we should solve these issues long term. A central memory
> > > > allocator is probably the way to go.
> > > > 
> > > > > >  	} else {
> > > > > >  		for (i = 0; i < info->comp_planes; i++) {
> > > > > >  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > > > > > @@ -591,10 +599,20 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> > > > > >  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> > > > > >  
> > > > > >  			plane = &pixfmt->plane_fmt[i];
> > > > > > -			plane->bytesperline =
> > > > > > -				info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv);
> > > > > > -			plane->sizeimage =
> > > > > > -				plane->bytesperline * DIV_ROUND_UP(aligned_height, vdiv);
> > > > > > +
> > > > > > +			/*
> > > > > > +			 * The user might have specified custom
> > > > > > +			 * sizeimage/bytesperline, only override them if
> > > > > > +			 * they're not big enough.
> > > > > > +			 */
> > > > > > +			plane->bytesperline = max_t(u32,
> > > > > > +						    info->bpp[i] *
> > > > > > +						    DIV_ROUND_UP(aligned_width, hdiv),
> > > > > > +						    plane->bytesperline);
> > > > > > +			plane->sizeimage = max_t(u32,
> > > > > > +						 plane->bytesperline *
> > > > > > +						 DIV_ROUND_UP(aligned_height, vdiv),
> > > > > > +						 plane->sizeimage);
> > > > > >  		}
> > > > > >  	}
> > > > > >  	return 0;
> > > > > > @@ -605,6 +623,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> > > > > >  		     u32 width, u32 height)
> > > > > >  {
> > > > > >  	const struct v4l2_format_info *info;
> > > > > > +	u32 bytesperline, sizeimage = 0;
> > > > > >  	int i;
> > > > > >  
> > > > > >  	info = v4l2_format_info(pixelformat);
> > > > > > @@ -618,8 +637,7 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> > > > > >  	pixfmt->width = width;
> > > > > >  	pixfmt->height = height;
> > > > > >  	pixfmt->pixelformat = pixelformat;
> > > > > > -	pixfmt->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > > > > > -	pixfmt->sizeimage = 0;
> > > > > > +	bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > > > > >  
> > > > > >  	for (i = 0; i < info->comp_planes; i++) {
> > > > > >  		unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > > > > > @@ -629,11 +647,17 @@ int v4l2_fill_pixfmt(struct v4l2_pix_format *pixfmt, u32 pixelformat,
> > > > > >  
> > > > > >  		aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> > > > > >  		aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> > > > > > -
> > > > > > -		pixfmt->sizeimage += info->bpp[i] *
> > > > > > -			DIV_ROUND_UP(aligned_width, hdiv) *
> > > > > > -			DIV_ROUND_UP(aligned_height, vdiv);
> > > > > > +		sizeimage += info->bpp[i] * DIV_ROUND_UP(aligned_width, hdiv) *
> > > > > > +			     DIV_ROUND_UP(aligned_height, vdiv);
> > > > > >  	}
> > > > > > +
> > > > > > +	/*
> > > > > > +	 * The user might have specified its own sizeimage/bytesperline values,
> > > > > > +	 * only override them if they're not big enough.
> > > > > > +	 */
> > > > > > +	pixfmt->sizeimage = max(sizeimage, pixfmt->sizeimage);
> > > > > > +	pixfmt->bytesperline = max(bytesperline, pixfmt->bytesperline);
> > > > > > +
> > > > > >  	return 0;
> > > > > >  }
> > > > > >  EXPORT_SYMBOL_GPL(v4l2_fill_pixfmt);

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Hans Verkuil]

On 5/10/19 3:51 PM, Nicolas Dufresne wrote:
> Le vendredi 10 mai 2019 à 14:30 +0200, Hans Verkuil a écrit :
>> On 5/10/19 2:24 PM, Laurent Pinchart wrote:
>>> Hi Hans,
>>>
>>> On Fri, May 10, 2019 at 02:17:32PM +0200, Hans Verkuil wrote:
>>>> On 5/10/19 1:28 PM, Laurent Pinchart wrote:
>>>>> On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
>>>>>> On 5/3/19 1:47 PM, Boris Brezillon wrote:
>>>>>>> Users can define custom sizeimage and bytesperline as long as they're
>>>>>>> big enough to store the amount of pixels required for a specific
>>>>>>> width/height under a specific format. Avoid overriding those fields in
>>>>>>> this case.
>>>>>>>
>>>>>>> Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
>>>>>>> ---
>>>>>>> Hello Hans,
>>>>>>>
>>>>>>> The sizeimage/bytesperline check on !MPLANE formats is still not 100%
>>>>>>> sure, as custom bytesperline might induce bigger sizeimage than what
>>>>>>> we calculate.
>>>>>>>
>>>>>>> I tried implementing something smarter taking the per-component plane
>>>>>>> bpp + hdiv param as we discussed the other day but decided to step
>>>>>>> back after realizing the per-component plane macro block might also
>>>>>>> differ at least in theory (not sure that's true in practice) and that
>>>>>>> has an impact on bytesperline too.
>>>>>>>
>>>>>>> Let me know how you want to handle that case.
>>>>>>>
>>>>>>> Regards,
>>>>>>>
>>>>>>> Boris
>>>>>>>
>>>>>>> Changes in v5:
>>>>>>> * New patch
>>>>>>> ---
>>>>>>>  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
>>>>>>>  1 file changed, 39 insertions(+), 15 deletions(-)
>>>>>>>
>>>>>>> diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
>>>>>>> index 3c6f5c115fc5..37bfc984a8b5 100644
>>>>>>> --- a/drivers/media/v4l2-core/v4l2-common.c
>>>>>>> +++ b/drivers/media/v4l2-core/v4l2-common.c
>>>>>>> @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>>>>>  	pixfmt->num_planes = info->mem_planes;
>>>>>>>  
>>>>>>>  	if (info->mem_planes == 1) {
>>>>>>> +		u32 bytesperline, sizeimage = 0;
>>>>>>> +
>>>>>>>  		plane = &pixfmt->plane_fmt[0];
>>>>>>> -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>>>>> -		plane->sizeimage = 0;
>>>>>>> +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
>>>>>>>  
>>>>>>>  		for (i = 0; i < info->comp_planes; i++) {
>>>>>>>  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
>>>>>>> @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
>>>>>>>  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
>>>>>>>  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
>>>>>>>  
>>>>>>> -			plane->sizeimage += info->bpp[i] *
>>>>>>> -				DIV_ROUND_UP(aligned_width, hdiv) *
>>>>>>> -				DIV_ROUND_UP(aligned_height, vdiv);
>>>>>>> +			sizeimage += info->bpp[i] *
>>>>>>> +				     DIV_ROUND_UP(aligned_width, hdiv) *
>>>>>>> +				     DIV_ROUND_UP(aligned_height, vdiv);
>>>>>>>  		}
>>>>>>> +
>>>>>>> +		/*
>>>>>>> +		 * The user might have specified custom sizeimage/bytesperline,
>>>>>>> +		 * only override them if they're not big enough.
>>>>>>> +		 */
>>>>>>> +		plane->sizeimage = max(sizeimage, plane->sizeimage);
>>>>>>> +		plane->bytesperline = max(bytesperline, plane->bytesperline);
>>>>>>
>>>>>> Let's just set bytesperline, ignoring the value the user supplied. There are very
>>>>>> few drivers that allow the user to set bytesperline anyway, so it's not a big deal
>>>>>> to drop support for that for now. We can add it back later.
>>>>>>
>>>>>> Just add a comment that a user-defined bytesperline value isn't currently supported.
>>>>>
>>>>> Kieran recently ran into an issue related to this, when sharing buffers
>>>>> between a CSI-2 receiver and an ISP. The ISP has alignment constraints
>>>>> in both the horizontal and vertical directions on the line stride and
>>>>> total image size. Out software framework currently allocates buffers
>>>>> from the CSI-2 receiver which doesn't have those constraints, and not
>>>>> being able to specify sizeimage is thus a problem.
>>>>
>>>> Not being able to specify sizeimage where? From userspace? Sorry, I don't
>>>> quite understand the specific issue here.
>>>
>>> Yes, from userspace.
>>
>> Ah, OK. But why not use CREATEBUFS? You can provide your own size when allocating
>> the buffers.
>>
>> Also note this patch: https://patchwork.linuxtv.org/patch/55656/
>>
>> Although this is specific for compressed formats.
> 
> While this work for compressed formats, it does not do anything for raw
> image horizontal padding. The importation bit of V4L2 is pretty
> difficult, so if we add helpers, we should load the way and simplify
> things for userspace rather then enforcing the existing difficulty.
> 
> I think from now own we should design with the mindset that a DMABuf
> that cannot be imported back into another driver due to software
> limitations is a useless waste of FD.

I agree, but I feel that this is part of the new fmt and streaming
ioctls project that Boris started. Doing further hacking of the existing API
is just complicating matters even more.

Creating new format ioctls that are much more flexible in describing image
formats (and closer to drm where possible) seems to be the right approach.

Hmmm... "waste of FD": Face Detection? File Descriptor? Probably not.

The urbandictionary doesn't help either: https://www.urbandictionary.com/define.php?term=FD

No idea what FD means :-), although I get the sentiment.

Regards,

	Hans

Re: [PATCH v5 03/15] media: v4l2-common: Support custom imagesize/bytesperline in fill_pixfmt()

[Nicolas Dufresne]

[-- Attachment #1: Type: text/plain, Size: 6719 bytes --]

Le vendredi 10 mai 2019 à 16:07 +0200, Hans Verkuil a écrit :
> On 5/10/19 3:51 PM, Nicolas Dufresne wrote:
> > Le vendredi 10 mai 2019 à 14:30 +0200, Hans Verkuil a écrit :
> > > On 5/10/19 2:24 PM, Laurent Pinchart wrote:
> > > > Hi Hans,
> > > > 
> > > > On Fri, May 10, 2019 at 02:17:32PM +0200, Hans Verkuil wrote:
> > > > > On 5/10/19 1:28 PM, Laurent Pinchart wrote:
> > > > > > On Fri, May 10, 2019 at 10:57:26AM +0200, Hans Verkuil wrote:
> > > > > > > On 5/3/19 1:47 PM, Boris Brezillon wrote:
> > > > > > > > Users can define custom sizeimage and bytesperline as long as they're
> > > > > > > > big enough to store the amount of pixels required for a specific
> > > > > > > > width/height under a specific format. Avoid overriding those fields in
> > > > > > > > this case.
> > > > > > > > 
> > > > > > > > Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
> > > > > > > > ---
> > > > > > > > Hello Hans,
> > > > > > > > 
> > > > > > > > The sizeimage/bytesperline check on !MPLANE formats is still not 100%
> > > > > > > > sure, as custom bytesperline might induce bigger sizeimage than what
> > > > > > > > we calculate.
> > > > > > > > 
> > > > > > > > I tried implementing something smarter taking the per-component plane
> > > > > > > > bpp + hdiv param as we discussed the other day but decided to step
> > > > > > > > back after realizing the per-component plane macro block might also
> > > > > > > > differ at least in theory (not sure that's true in practice) and that
> > > > > > > > has an impact on bytesperline too.
> > > > > > > > 
> > > > > > > > Let me know how you want to handle that case.
> > > > > > > > 
> > > > > > > > Regards,
> > > > > > > > 
> > > > > > > > Boris
> > > > > > > > 
> > > > > > > > Changes in v5:
> > > > > > > > * New patch
> > > > > > > > ---
> > > > > > > >  drivers/media/v4l2-core/v4l2-common.c | 54 +++++++++++++++++++--------
> > > > > > > >  1 file changed, 39 insertions(+), 15 deletions(-)
> > > > > > > > 
> > > > > > > > diff --git a/drivers/media/v4l2-core/v4l2-common.c b/drivers/media/v4l2-core/v4l2-common.c
> > > > > > > > index 3c6f5c115fc5..37bfc984a8b5 100644
> > > > > > > > --- a/drivers/media/v4l2-core/v4l2-common.c
> > > > > > > > +++ b/drivers/media/v4l2-core/v4l2-common.c
> > > > > > > > @@ -563,9 +563,10 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> > > > > > > >  	pixfmt->num_planes = info->mem_planes;
> > > > > > > >  
> > > > > > > >  	if (info->mem_planes == 1) {
> > > > > > > > +		u32 bytesperline, sizeimage = 0;
> > > > > > > > +
> > > > > > > >  		plane = &pixfmt->plane_fmt[0];
> > > > > > > > -		plane->bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > > > > > > > -		plane->sizeimage = 0;
> > > > > > > > +		bytesperline = ALIGN(width, v4l2_format_block_width(info, 0)) * info->bpp[0];
> > > > > > > >  
> > > > > > > >  		for (i = 0; i < info->comp_planes; i++) {
> > > > > > > >  			unsigned int hdiv = (i == 0) ? 1 : info->hdiv;
> > > > > > > > @@ -576,10 +577,17 @@ int v4l2_fill_pixfmt_mp(struct v4l2_pix_format_mplane *pixfmt,
> > > > > > > >  			aligned_width = ALIGN(width, v4l2_format_block_width(info, i));
> > > > > > > >  			aligned_height = ALIGN(height, v4l2_format_block_height(info, i));
> > > > > > > >  
> > > > > > > > -			plane->sizeimage += info->bpp[i] *
> > > > > > > > -				DIV_ROUND_UP(aligned_width, hdiv) *
> > > > > > > > -				DIV_ROUND_UP(aligned_height, vdiv);
> > > > > > > > +			sizeimage += info->bpp[i] *
> > > > > > > > +				     DIV_ROUND_UP(aligned_width, hdiv) *
> > > > > > > > +				     DIV_ROUND_UP(aligned_height, vdiv);
> > > > > > > >  		}
> > > > > > > > +
> > > > > > > > +		/*
> > > > > > > > +		 * The user might have specified custom sizeimage/bytesperline,
> > > > > > > > +		 * only override them if they're not big enough.
> > > > > > > > +		 */
> > > > > > > > +		plane->sizeimage = max(sizeimage, plane->sizeimage);
> > > > > > > > +		plane->bytesperline = max(bytesperline, plane->bytesperline);
> > > > > > > 
> > > > > > > Let's just set bytesperline, ignoring the value the user supplied. There are very
> > > > > > > few drivers that allow the user to set bytesperline anyway, so it's not a big deal
> > > > > > > to drop support for that for now. We can add it back later.
> > > > > > > 
> > > > > > > Just add a comment that a user-defined bytesperline value isn't currently supported.
> > > > > > 
> > > > > > Kieran recently ran into an issue related to this, when sharing buffers
> > > > > > between a CSI-2 receiver and an ISP. The ISP has alignment constraints
> > > > > > in both the horizontal and vertical directions on the line stride and
> > > > > > total image size. Out software framework currently allocates buffers
> > > > > > from the CSI-2 receiver which doesn't have those constraints, and not
> > > > > > being able to specify sizeimage is thus a problem.
> > > > > 
> > > > > Not being able to specify sizeimage where? From userspace? Sorry, I don't
> > > > > quite understand the specific issue here.
> > > > 
> > > > Yes, from userspace.
> > > 
> > > Ah, OK. But why not use CREATEBUFS? You can provide your own size when allocating
> > > the buffers.
> > > 
> > > Also note this patch: https://patchwork.linuxtv.org/patch/55656/
> > > 
> > > Although this is specific for compressed formats.
> > 
> > While this work for compressed formats, it does not do anything for raw
> > image horizontal padding. The importation bit of V4L2 is pretty
> > difficult, so if we add helpers, we should load the way and simplify
> > things for userspace rather then enforcing the existing difficulty.
> > 
> > I think from now own we should design with the mindset that a DMABuf
> > that cannot be imported back into another driver due to software
> > limitations is a useless waste of FD.
> 
> I agree, but I feel that this is part of the new fmt and streaming
> ioctls project that Boris started. Doing further hacking of the existing API
> is just complicating matters even more.
> 
> Creating new format ioctls that are much more flexible in describing image
> formats (and closer to drm where possible) seems to be the right approach.
> 
> Hmmm... "waste of FD": Face Detection? File Descriptor? Probably not.
> 
> The urbandictionary doesn't help either: https://www.urbandictionary.com/define.php?term=FD
> 
> No idea what FD means :-), although I get the sentiment.

I was not swearing. An FD on Unix systems like Linux is also well know
as a File Descriptor. DMABuf, memfd, an open file, etc.

regards,
Nicolas

> 
> Regards,
> 
> 	Hans
> 

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 195 bytes --]