Linux Media Controller development
 help / color / mirror / Atom feed
From: Vikash Garodia <vikash.garodia@oss.qualcomm.com>
To: Neil Armstrong <neil.armstrong@linaro.org>,
	Dikshita Agarwal <dikshita.agarwal@oss.qualcomm.com>,
	Abhinav Kumar <abhinav.kumar@linux.dev>,
	Bryan O'Donoghue <bod@kernel.org>,
	Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: linux-media@vger.kernel.org, linux-arm-msm@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 2/6] media: qcom: iris: add QC10C & P010 buffer size calculations
Date: Thu, 14 May 2026 00:13:34 +0530	[thread overview]
Message-ID: <a5e7a134-8b23-4cd1-a4c5-6f6c1954a983@oss.qualcomm.com> (raw)
In-Reply-To: <20260511-topic-sm8x50-iris-10bit-decoding-v3-2-7fc049b93042@linaro.org>


On 5/11/2026 2:50 PM, Neil Armstrong wrote:
> The P010 (YUV format with 16-bits per pixel with interleaved UV)
> and QC10C (P010 compressed mode similar to QC08C) requires specific
> buffer calculations to allocate the right buffer size for DPB frames

expand DPB

> and frames consumed by userspace.
> 
> Similar to 8bit, the 10bit DPB frames uses QC10C format.
> 
> Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
> ---
>   drivers/media/platform/qcom/iris/iris_buffer.c | 195 ++++++++++++++++++++++++-
>   1 file changed, 194 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/platform/qcom/iris/iris_buffer.c b/drivers/media/platform/qcom/iris/iris_buffer.c
> index 1d53c7414b75..713a63d0ca0c 100644
> --- a/drivers/media/platform/qcom/iris/iris_buffer.c
> +++ b/drivers/media/platform/qcom/iris/iris_buffer.c
> @@ -15,8 +15,11 @@
>   #define MAX_WIDTH 4096
>   #define MAX_HEIGHT 2304
>   #define Y_STRIDE_ALIGN 128
> +#define Y_STRIDE_ALIGN_P010 256
>   #define UV_STRIDE_ALIGN 128
> +#define UV_STRIDE_ALIGN_P010 256
>   #define Y_SCANLINE_ALIGN 32
> +#define Y_SCANLINE_ALIGN_QC10C 16
>   #define UV_SCANLINE_ALIGN 16
>   #define UV_SCANLINE_ALIGN_QC08C 32
>   #define META_STRIDE_ALIGNED 64
> @@ -80,6 +83,63 @@ static u32 iris_yuv_buffer_size_nv12(struct iris_inst *inst)
>   	return ALIGN(y_plane + uv_plane, PIXELS_4K);
>   }
>   
> +/*
> + * P010:
> + * YUV 4:2:0 image with a plane of 10 bit Y samples followed
> + * by an interleaved U/V plane containing 10 bit 2x2 subsampled
> + * colour difference samples.
> + *
> + * <-Y/UV_Stride (aligned to 256)->
> + * <----- Width*2 ------->
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  ^           ^
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  |           |
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  Height      |
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  |          y_scanlines (aligned to 32)
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  |           |
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  |           |
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  |           |
> + * Y Y Y Y Y Y Y Y Y Y Y Y . . . .  V           |
> + * . . . . . . . . . . . . . . . .              |
> + * . . . . . . . . . . . . . . . .              |
> + * . . . . . . . . . . . . . . . .              |
> + * . . . . . . . . . . . . . . . .              V
> + * U V U V U V U V U V U V . . . .  ^
> + * U V U V U V U V U V U V . . . .  |
> + * U V U V U V U V U V U V . . . .  |
> + * U V U V U V U V U V U V . . . .  uv_scanlines (aligned to 16)
> + * . . . . . . . . . . . . . . . .  |
> + * . . . . . . . . . . . . . . . .  V
> + * . . . . . . . . . . . . . . . .  --> Buffer size aligned to 4K
> + *
> + * y_stride : Width*2 aligned to 256
> + * uv_stride : Width*2 aligned to 256
> + * y_scanlines: Height aligned to 32
> + * uv_scanlines: Height/2 aligned to 16
> + * Total size = align((y_stride * y_scanlines
> + *          + uv_stride * uv_scanlines , 4096)
> + *
> + * Note: All the alignments are hardware requirements.
> + */
> +static u32 iris_yuv_buffer_size_p010(struct iris_inst *inst)
> +{
> +	u32 y_plane, uv_plane, y_stride, uv_stride, y_scanlines, uv_scanlines;
> +	struct v4l2_format *f;
> +
> +	if (inst->domain == DECODER)
> +		f = inst->fmt_dst;
> +	else
> +		f = inst->fmt_src;
> +
> +	y_stride = ALIGN(f->fmt.pix_mp.width * 2, Y_STRIDE_ALIGN_P010);
> +	uv_stride = ALIGN(f->fmt.pix_mp.width * 2, UV_STRIDE_ALIGN_P010);
> +	y_scanlines = ALIGN(f->fmt.pix_mp.height, Y_SCANLINE_ALIGN);
> +	uv_scanlines = ALIGN((f->fmt.pix_mp.height + 1) >> 1, UV_SCANLINE_ALIGN);
> +	y_plane = y_stride * y_scanlines;
> +	uv_plane = uv_stride * uv_scanlines;
> +
> +	return ALIGN(y_plane + uv_plane, PIXELS_4K);
> +}
> +
>   /*
>    * QC08C:
>    * Compressed Macro-tile format for NV12.
> @@ -204,6 +264,132 @@ static u32 iris_yuv_buffer_size_qc08c(struct iris_inst *inst)
>   	return ALIGN(y_meta_plane + y_plane + uv_meta_plane + uv_plane, PIXELS_4K);
>   }
>   
> +/*
> + * QC10C:
> + * UBWC-compressed format for P010.
> + * Contains 4 planes in the following order -
> + * (A) Y_Meta_Plane
> + * (B) Y_UBWC_Plane
> + * (C) UV_Meta_Plane
> + * (D) UV_UBWC_Plane
> + *
> + * Y_Meta_Plane consists of meta information to decode compressed
> + * tile data in Y_UBWC_Plane.
> + * Y_UBWC_Plane consists of Y data in compressed macro-tile format.
> + * UBWC decoder block will use the Y_Meta_Plane data together with
> + * Y_UBWC_Plane data to produce loss-less uncompressed 10 bit Y samples.
> + *
> + * UV_Meta_Plane consists of meta information to decode compressed
> + * tile data in UV_UBWC_Plane.
> + * UV_UBWC_Plane consists of UV data in compressed macro-tile format.
> + * UBWC decoder block will use UV_Meta_Plane data together with
> + * UV_UBWC_Plane data to produce loss-less uncompressed 10 bit 2x2
> + * subsampled color difference samples.
> + *
> + * Each tile in Y_UBWC_Plane/UV_UBWC_Plane is independently decodable
> + * and randomly accessible. There is no dependency between tiles.
> + *
> + * <----- Y Meta stride -----> (aligned to 64)
> + * <-------- Width ----------> (aligned to 48)
> + * M M M M M M M M M M M M . .      ^           ^
> + * M M M M M M M M M M M M . .      |           |
> + * M M M M M M M M M M M M . .      Height      |
> + * M M M M M M M M M M M M . .      |         Meta_Y_Scanlines (aligned to 16)
> + * M M M M M M M M M M M M . .      |           |
> + * M M M M M M M M M M M M . .      |           |
> + * M M M M M M M M M M M M . .      |           |
> + * M M M M M M M M M M M M . .      V           |
> + * . . . . . . . . . . . . . .                  |
> + * . . . . . . . . . . . . . .                  |
> + * . . . . . . . . . . . . . .      -------> Buffer size aligned to 4k
> + * . . . . . . . . . . . . . .                  V
> + * <--Compressed tile Y stride --> (aligned to 256)
> + * <------- Width * 4/3 ---------> (aligned to 48)
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  ^           ^
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  |           |
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  Height      |
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  |        Macro_tile_Y_Scanlines (aligned to 16)
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  |           |
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  |           |
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  |           |
> + * Y* Y* Y* Y* Y* Y* Y* Y* . . . .  V           |
> + * . . . . . . . . . . . . . . . .              |
> + * . . . . . . . . . . . . . . . .              |
> + * . . . . . . . . . . . . . . . .  -------> Buffer size aligned to 4k
> + * . . . . . . . . . . . . . . . .              V
> + * <---- UV Meta stride ----> (aligned to 64)
> + * <----- Width / 2 --------> (aligned to 24)
> + * M M M M M M M M M M M M . .    ^           ^
> + * M M M M M M M M M M M M . .    |           |
> + * M M M M M M M M M M M M . .    Height/2    |
> + * M M M M M M M M M M M M . .    V           M_UV_Scanlines (aligned to 16)
> + * . . . . . . . . . . . . . .                |
> + * . . . . . . . . . . . . . .                V
> + * . . . . . . . . . . . . . .      -------> Buffer size aligned to 4k
> + * <--Compressed tile UV stride--> (aligned to 256)
> + * <------- Width * 4/3 ---------> (aligned to 48)
> + * U* V* U* V* U* V* U* V* . . . .  ^
> + * U* V* U* V* U* V* U* V* . . . .  |
> + * U* V* U* V* U* V* U* V* . . . .  |
> + * U* V* U* V* U* V* U* V* . . . .  UV_Scanlines (aligned to 16)
> + * . . . . . . . . . . . . . . . .  |
> + * . . . . . . . . . . . . . . . .  V
> + * . . . . . . . . . . . . . . . .  -------> Buffer size aligned to 4k
> + *
> + * y_stride: width aligned to 256
> + * uv_stride: width aligned to 256

This does not match with above layout

> + * y_scanlines: height aligned to 16
> + * uv_scanlines: height aligned to 16
> + * y_plane: buffer size aligned to 4096
> + * uv_plane: buffer size aligned to 4096
> + * y_meta_stride: width aligned to 64
> + * y_meta_scanlines: height aligned to 16
> + * y_meta_plane: buffer size aligned to 4096
> + * uv_meta_stride: width aligned to 64
> + * uv_meta_scanlines: height aligned to 16
> + * uv_meta_plane: buffer size aligned to 4096
> + *
> + * Total size = align( y_plane + uv_plane +
> + *           y_meta_plane + uv_meta_plane, 4096)
> + *
> + * Note: All the alignments are hardware requirements.
> + */
> +static u32 iris_yuv_buffer_size_qc10c(struct iris_inst *inst)
> +{
> +	u32 y_plane, uv_plane, y_stride, uv_stride;
> +	u32 uv_meta_stride, uv_meta_plane;
> +	u32 y_meta_stride, y_meta_plane;
> +	struct v4l2_format *f = NULL;

drop NULL

> +
> +	if (inst->domain == DECODER)
> +		f = inst->fmt_dst;
> +	else
> +		f = inst->fmt_src;
> +
> +	y_meta_stride = ALIGN(DIV_ROUND_UP(f->fmt.pix_mp.width, 48),
> +			      META_STRIDE_ALIGNED);
> +	y_meta_plane = y_meta_stride * ALIGN(DIV_ROUND_UP(f->fmt.pix_mp.height, 4),
> +					     META_SCANLINE_ALIGNED);
> +	y_meta_plane = ALIGN(y_meta_plane, PIXELS_4K);
> +
> +	y_stride = ALIGN(f->fmt.pix_mp.width * 4 / 3, Y_STRIDE_ALIGN_P010);
> +	y_plane = ALIGN(y_stride * ALIGN(f->fmt.pix_mp.height, Y_SCANLINE_ALIGN_QC10C),
> +			PIXELS_4K);
> +
> +	uv_meta_stride = ALIGN(DIV_ROUND_UP((f->fmt.pix_mp.width + 1) / 2, 24),
> +			       META_STRIDE_ALIGNED);
> +	uv_meta_plane = uv_meta_stride *
> +			ALIGN(DIV_ROUND_UP((f->fmt.pix_mp.height + 1) / 2, 4),
> +			      META_SCANLINE_ALIGNED);

Please use >> 1 throughout for consistency. I see 
iris_yuv_buffer_size_p010() uses them, while 
iris_yuv_buffer_size_qc10c() dont.

> +	uv_meta_plane = ALIGN(uv_meta_plane, PIXELS_4K);
> +
> +	uv_stride = ALIGN(f->fmt.pix_mp.width * 4 / 3, UV_STRIDE_ALIGN_P010);
> +	uv_plane = ALIGN(uv_stride * ALIGN((f->fmt.pix_mp.height + 1) / 2, UV_SCANLINE_ALIGN),
> +			 PIXELS_4K);
> +
> +	return ALIGN(y_meta_plane + y_plane + uv_meta_plane + uv_plane, PIXELS_4K);
> +}
> +
>   static u32 iris_dec_bitstream_buffer_size(struct iris_inst *inst)
>   {
>   	struct platform_inst_caps *caps = inst->core->iris_platform_data->inst_caps;
> @@ -268,10 +454,17 @@ int iris_get_buffer_size(struct iris_inst *inst,
>   		case BUF_OUTPUT:
>   			if (inst->fmt_dst->fmt.pix_mp.pixelformat == V4L2_PIX_FMT_QC08C)
>   				return iris_yuv_buffer_size_qc08c(inst);
> +			else if (inst->fmt_dst->fmt.pix_mp.pixelformat == V4L2_PIX_FMT_QC10C)
> +				return iris_yuv_buffer_size_qc10c(inst);
> +			else if (inst->fmt_dst->fmt.pix_mp.pixelformat == V4L2_PIX_FMT_P010)
> +				return iris_yuv_buffer_size_p010(inst);
>   			else
>   				return iris_yuv_buffer_size_nv12(inst);
>   		case BUF_DPB:
> -			return iris_yuv_buffer_size_qc08c(inst);
> +			if (iris_fmt_is_10bit(inst->fmt_dst->fmt.pix_mp.pixelformat))
> +				return iris_yuv_buffer_size_qc10c(inst);
> +			else
> +				return iris_yuv_buffer_size_qc08c(inst);
>   		default:
>   			return 0;
>   		}
> 


  reply	other threads:[~2026-05-13 18:43 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-11  9:20 [PATCH v3 0/6] media: qcom: iris: add support for decoding 10bit formats Neil Armstrong
2026-05-11  9:20 ` [PATCH v3 1/6] media: qcom: iris: add helpers for 8bit and " Neil Armstrong
2026-05-13 10:25   ` Dmitry Baryshkov
2026-05-13 18:25   ` Vikash Garodia
2026-05-11  9:20 ` [PATCH v3 2/6] media: qcom: iris: add QC10C & P010 buffer size calculations Neil Armstrong
2026-05-13 18:43   ` Vikash Garodia [this message]
2026-05-11  9:20 ` [PATCH v3 3/6] media: qcom: iris: gen2: add support for 10bit decoding Neil Armstrong
2026-05-13 10:38   ` Dmitry Baryshkov
2026-05-13 18:50   ` Vikash Garodia
2026-05-11  9:20 ` [PATCH v3 4/6] media: qcom: iris: vdec: update size and stride calculations for 10bit formats Neil Armstrong
2026-05-13 10:39   ` Dmitry Baryshkov
2026-05-13 19:05   ` Vikash Garodia
2026-05-11  9:20 ` [PATCH v3 5/6] media: qcom: iris: vdec: update find_format to handle 8bit and " Neil Armstrong
2026-05-13 10:47   ` Dmitry Baryshkov
2026-05-13 19:27   ` Vikash Garodia
2026-05-11  9:20 ` [PATCH v3 6/6] media: qcom: iris: vdec: allow GEN2 decoding into 10bit format Neil Armstrong
2026-05-13 10:45   ` Dmitry Baryshkov
2026-05-13 11:59     ` Neil Armstrong
2026-05-13 12:04       ` Dmitry Baryshkov
2026-05-13 19:39   ` Vikash Garodia
2026-05-13  8:50 ` [PATCH v3 0/6] media: qcom: iris: add support for decoding 10bit formats Wangao Wang
2026-05-13 12:11   ` Neil Armstrong
2026-05-13 16:02     ` Neil Armstrong
2026-05-13 16:04       ` Dmitry Baryshkov
2026-05-13 17:33 ` Vikash Garodia

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=a5e7a134-8b23-4cd1-a4c5-6f6c1954a983@oss.qualcomm.com \
    --to=vikash.garodia@oss.qualcomm.com \
    --cc=abhinav.kumar@linux.dev \
    --cc=bod@kernel.org \
    --cc=dikshita.agarwal@oss.qualcomm.com \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=neil.armstrong@linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox