Re: [PATCH v2 05/22] media: rockchip: rga: use stride for offset calculation

From: Nicolas Dufresne
Date: Wed Dec 24 2025 - 10:00:44 EST


Le mercredi 03 décembre 2025 à 16:52 +0100, Sven Püschel a écrit :
> Use the stride instead of the width for the offset calculation. This
> ensures that the bytesperline value doesn't need to match the width
> value of the image.
>
> Furthermore this patch removes the dependency on the uv_factor property
> and instead reuses the v4l2_format_info to determine the correct
> division factor.
>
> Signed-off-by: Sven Püschel <s.pueschel@xxxxxxxxxxxxxx>
> ---
>  drivers/media/platform/rockchip/rga/rga-buf.c | 14 +++++++++-----
>  1 file changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/media/platform/rockchip/rga/rga-buf.c b/drivers/media/platform/rockchip/rga/rga-buf.c
> index 730bdf98565a5..b5e6b1b527ca8 100644
> --- a/drivers/media/platform/rockchip/rga/rga-buf.c
> +++ b/drivers/media/platform/rockchip/rga/rga-buf.c
> @@ -14,7 +14,6 @@
>  #include <media/videobuf2-dma-sg.h>
>  #include <media/videobuf2-v4l2.h>
>  
> -#include "rga-hw.h"
>  #include "rga.h"
>  
>  static ssize_t fill_descriptors(struct rga_dma_desc *desc, size_t max_desc,
> @@ -92,14 +91,19 @@ static int rga_buf_init(struct vb2_buffer *vb)
>   return 0;
>  }
>  
> -static int get_plane_offset(struct rga_frame *f, int plane)
> +static int get_plane_offset(struct rga_frame *f,
> +     const struct v4l2_format_info *info,
> +     int plane)
>  {
> + u32 stride = f->pix.plane_fmt[0].bytesperline;
> +
>   if (plane == 0)
>   return 0;
>   if (plane == 1)
> - return f->width * f->height;
> + return stride * f->height;
>   if (plane == 2)
> - return f->width * f->height + (f->width * f->height / f->fmt->uv_factor);
> + return stride * f->height +
> +        (stride * f->height / info->hdiv / info->vdiv);

This was the only user of uv_factor, please cleanup rga.h/rga.c. I see there is
also redudant x_div/y_div, with this similar code re-implemented in
rga_get_addr_offset(), can you check for v3 if we can cleanup some more ?

One thought though is that this code is already hidden inside v4l2-common,
perhaps you could just expose it, this ensure that we don't accidently have
different information when configuring the HW and communicating with the user.

Nicolas

>  
>   return -EINVAL;
>  }
> @@ -145,7 +149,7 @@ static int rga_buf_prepare(struct vb2_buffer *vb)
>   /* Fill the remaining planes */
>   info = v4l2_format_info(f->fmt->fourcc);
>   for (i = info->mem_planes; i < info->comp_planes; i++)
> - offsets[i] = get_plane_offset(f, i);
> + offsets[i] = get_plane_offset(f, info, i);
>  
>   rbuf->offset.y_off = offsets[0];
>   rbuf->offset.u_off = offsets[1];

Attachment: signature.asc
Description: This is a digitally signed message part