2019-10-29 07:22:47

by Jonas Karlman

Subject: [PATCH v2 05/10] media: hantro: Reduce H264 extra space for motion vectors

A decoded 8-bit 4:2:0 frame need memory for up to 448 bytes per
macroblock with additional 32 bytes on multi-core variants.

Memory layout is as follow:

+---------------------------+
| Y-plane 256 bytes x MBs |
+---------------------------+
| UV-plane 128 bytes x MBs |
+---------------------------+
| MV buffer 64 bytes x MBs |
+---------------------------+
| MC sync 32 bytes |
+---------------------------+

Reduce the extra space allocated now that motion vector buffer offset no
longer is based on the extra space.

Only allocate extra space for 64 bytes x MBs of motion vector buffer
and 32 bytes for multi-core sync.

Fixes: a9471e25629b ("media: hantro: Add core bits to support H264 decoding")
Signed-off-by: Jonas Karlman <[email protected]>
---
Changes in v2:
- updated commit message
---
drivers/staging/media/hantro/hantro_v4l2.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/staging/media/hantro/hantro_v4l2.c b/drivers/staging/media/hantro/hantro_v4l2.c
index 3dae52abb96c..b3a4368b37de 100644
--- a/drivers/staging/media/hantro/hantro_v4l2.c
+++ b/drivers/staging/media/hantro/hantro_v4l2.c
@@ -242,12 +242,12 @@ static int vidioc_try_fmt(struct file *file, void *priv, struct v4l2_format *f,
/*
* The H264 decoder needs extra space on the output buffers
* to store motion vectors. This is needed for reference
- * frames.
+ * frames. 32 extra bytes is used for multi-core sync.
*/
if (ctx->vpu_src_fmt->fourcc == V4L2_PIX_FMT_H264_SLICE)
pix_mp->plane_fmt[0].sizeimage +=
- 128 * DIV_ROUND_UP(pix_mp->width, 16) *
- DIV_ROUND_UP(pix_mp->height, 16);
+ 64 * MB_WIDTH(pix_mp->width) *
+ MB_WIDTH(pix_mp->height) + 32;
} else if (!pix_mp->plane_fmt[0].sizeimage) {
/*
* For coded formats the application can specify
--
2.17.1

2019-10-31 09:51:56

by Boris Brezillon

Subject: Re: [PATCH v2 05/10] media: hantro: Reduce H264 extra space for motion vectors

On Tue, 29 Oct 2019 01:24:50 +0000
Jonas Karlman <[email protected]> wrote:

> A decoded 8-bit 4:2:0 frame need memory for up to 448 bytes per
> macroblock with additional 32 bytes on multi-core variants.
>
> Memory layout is as follow:
>
> +---------------------------+
> | Y-plane 256 bytes x MBs |
> +---------------------------+
> | UV-plane 128 bytes x MBs |
> +---------------------------+
> | MV buffer 64 bytes x MBs |
> +---------------------------+
> | MC sync 32 bytes |
> +---------------------------+

As for patch 4, can we point this diagram as a comment in the code too?

>
> Reduce the extra space allocated now that motion vector buffer offset no
> longer is based on the extra space.
>
> Only allocate extra space for 64 bytes x MBs of motion vector buffer
> and 32 bytes for multi-core sync.
>
> Fixes: a9471e25629b ("media: hantro: Add core bits to support H264 decoding")
> Signed-off-by: Jonas Karlman <[email protected]>

Reviewed-by: Boris Brezillon <[email protected]>

> ---
> Changes in v2:
> - updated commit message
> ---
> drivers/staging/media/hantro/hantro_v4l2.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/staging/media/hantro/hantro_v4l2.c b/drivers/staging/media/hantro/hantro_v4l2.c
> index 3dae52abb96c..b3a4368b37de 100644
> --- a/drivers/staging/media/hantro/hantro_v4l2.c
> +++ b/drivers/staging/media/hantro/hantro_v4l2.c
> @@ -242,12 +242,12 @@ static int vidioc_try_fmt(struct file *file, void *priv, struct v4l2_format *f,
> /*
> * The H264 decoder needs extra space on the output buffers
> * to store motion vectors. This is needed for reference
> - * frames.
> + * frames. 32 extra bytes is used for multi-core sync.
> */
> if (ctx->vpu_src_fmt->fourcc == V4L2_PIX_FMT_H264_SLICE)
> pix_mp->plane_fmt[0].sizeimage +=
> - 128 * DIV_ROUND_UP(pix_mp->width, 16) *
> - DIV_ROUND_UP(pix_mp->height, 16);
> + 64 * MB_WIDTH(pix_mp->width) *
> + MB_WIDTH(pix_mp->height) + 32;
> } else if (!pix_mp->plane_fmt[0].sizeimage) {
> /*
> * For coded formats the application can specify