Re: [PATCH v6 2/3] media: jpeg: Add reference quantization and huffman tables

From: Hans Verkuil
Date: Mon Mar 18 2024 - 09:40:15 EST


Hi Devarsh,

I was reviewing the pull request for this driver, but I have some concerns about
this patch.

First of all, it is generally bad practice to put data in a header, that really
belongs in a source. Why not just put this in drivers/media/v4l2-core/v4l2-jpeg.c?

On 28/02/2024 3:11 pm, Devarsh Thakkar wrote:
> Add reference quantization and huffman tables to a global header file so
> that they can be re-used by other JPEG drivers.
>
> These are example tables provided in ITU-T.81 as reference. The JPEG
> encoders are free to use either these or their own proprietary tables.
>
> These tables are being used already by verisilicon driver [1].

If the same tables are in use by another driver, then please convert that
driver to use the same tables defined in this patch. Bonus points if you
can check for other encoder JPEG drivers that might use the same tables :-)

Some more notes below:

>
> Also add necessary prefixes to be used for huffman tables in global header
> file.
>
> [1] :
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/media/platform/verisilicon/hantro_jpeg.c?h=v6.7#n30
>
> Signed-off-by: Devarsh Thakkar <devarsht@xxxxxx>
> ---
> V1->V4 : No change (Introduced in v4)
> V5 : Rename from jpeg_enc_reftables to jpeg-enc-reftables.h
> V6 : No change
>
> include/media/jpeg-enc-reftables.h | 112 +++++++++++++++++++++++++++++
> include/media/jpeg.h | 4 ++
> 2 files changed, 116 insertions(+)
> create mode 100644 include/media/jpeg-enc-reftables.h
>
> diff --git a/include/media/jpeg-enc-reftables.h b/include/media/jpeg-enc-reftables.h
> new file mode 100644
> index 000000000000..91959907113f
> --- /dev/null
> +++ b/include/media/jpeg-enc-reftables.h
> @@ -0,0 +1,112 @@
> +/* SPDX-License-Identifier: GPL-2.0-only */
> +/* Reference quantization, zig-zag scan and huffman tables as shared in ITU-T.81 */
> +
> +#ifndef _MEDIA_JPEG_ENC_REF_H_
> +#define _MEDIA_JPEG_ENC_REF_H_
> +
> +/* Luma and chroma qp table to acheive 50% compression quality

table -> tables
acheive -> achieve

> + * This is as per example in Annex K.1 of ITU-T.81
> + */
> +static const u8 luma_q_table[64] = {
> + 16, 11, 10, 16, 24, 40, 51, 61,
> + 12, 12, 14, 19, 26, 58, 60, 55,
> + 14, 13, 16, 24, 40, 57, 69, 56,
> + 14, 17, 22, 29, 51, 87, 80, 62,
> + 18, 22, 37, 56, 68, 109, 103, 77,
> + 24, 35, 55, 64, 81, 104, 113, 92,
> + 49, 64, 78, 87, 103, 121, 120, 101,
> + 72, 92, 95, 98, 112, 100, 103, 99
> +};
> +
> +static const u8 chroma_q_table[64] = {
> + 17, 18, 24, 47, 99, 99, 99, 99,
> + 18, 21, 26, 66, 99, 99, 99, 99,
> + 24, 26, 56, 99, 99, 99, 99, 99,
> + 47, 66, 99, 99, 99, 99, 99, 99,
> + 99, 99, 99, 99, 99, 99, 99, 99,
> + 99, 99, 99, 99, 99, 99, 99, 99,
> + 99, 99, 99, 99, 99, 99, 99, 99,
> + 99, 99, 99, 99, 99, 99, 99, 99
> +};
> +
> +/* Zigzag scan pattern */
> +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
> +};
> +
> +/*
> + * Contains the data that needs to be sent in the marker segment of an interchange format JPEG
> + * stream or an abbreviated format table specification data stream.
> + * Specifies the huffman table used for encoding the luminance DC coefficient differences.
> + * The table represents Table K.3 of ITU-T.81

Please keep the line length in this source within 80 chars max. Sometimes it is OK
to go with longer line lengths, but that is really not needed here.

> + */
> +static const u8 luma_dc_table[] = {
> + 0x00, 0x01, 0x05, 0x01, 0x01, 0x01, 0x01, 0x01,
> + 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
> + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0A, 0x0B
> +};
> +
> +/*
> + * Contains the data that needs to be sent in the marker segment of an interchange format JPEG
> + * stream or an abbreviated format table specification data stream.
> + * Specifies the huffman table used for encoding the luminance AC coefficients.
> + * The table represents Table K.5 of ITU-T.81
> + */
> +static const u8 luma_ac_table[] = {
> + 0x00, 0x02, 0x01, 0x03, 0x03, 0x02, 0x04, 0x03,
> + 0x05, 0x05, 0x04, 0x04, 0x00, 0x00, 0x01, 0x7D,
> + 0x01, 0x02, 0x03, 0x00, 0x04, 0x11, 0x05, 0x12, 0x21, 0x31, 0x41, 0x06, 0x13, 0x51, 0x61,
> + 0x07, 0x22, 0x71, 0x14, 0x32, 0x81, 0x91, 0xA1, 0x08, 0x23, 0x42, 0xB1, 0xC1, 0x15, 0x52,
> + 0xD1, 0xF0, 0x24, 0x33, 0x62, 0x72, 0x82, 0x09, 0x0A, 0x16, 0x17, 0x18, 0x19, 0x1A, 0x25,
> + 0x26, 0x27, 0x28, 0x29, 0x2A, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3A, 0x43, 0x44, 0x45,
> + 0x46, 0x47, 0x48, 0x49, 0x4A, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5A, 0x63, 0x64,
> + 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A, 0x83,
> + 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8A, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99,
> + 0x9A, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xAA, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6,
> + 0xB7, 0xB8, 0xB9, 0xBA, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA, 0xD2, 0xD3,
> + 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8,
> + 0xE9, 0xEA, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA
> +};
> +
> +/*
> + * Contains the data that needs to be sent in the marker segment of an interchange format JPEG
> + * stream or an abbreviated format table specification data stream.
> + * Specifies the huffman table used for encoding the chrominance DC coefficient differences.
> + * The table represents Table K.4 of ITU-T.81
> + */
> +static const u8 chroma_dc_table[] = {
> + 0x00, 0x03, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01,
> + 0x01, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00,
> + 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0A, 0x0B
> +};
> +
> +/*
> + * Contains the data that needs to be sent in the marker segment of an interchange format JPEG
> + * stream or an abbreviated format table specification data stream.
> + * Specifies the huffman table used for encoding the chrominance AC coefficients.
> + * The table represents Table K.6 of ITU-T.81
> + */
> +static const u8 chroma_ac_table[] = {
> + 0x00, 0x02, 0x01, 0x02, 0x04, 0x04, 0x03, 0x04,
> + 0x07, 0x05, 0x04, 0x04, 0x00, 0x01, 0x02, 0x77,
> + 0x00, 0x01, 0x02, 0x03, 0x11, 0x04, 0x05, 0x21, 0x31, 0x06, 0x12, 0x41, 0x51, 0x07, 0x61,
> + 0x71, 0x13, 0x22, 0x32, 0x81, 0x08, 0x14, 0x42, 0x91, 0xA1, 0xB1, 0xC1, 0x09, 0x23, 0x33,
> + 0x52, 0xF0, 0x15, 0x62, 0x72, 0xD1, 0x0A, 0x16, 0x24, 0x34, 0xE1, 0x25, 0xF1, 0x17, 0x18,
> + 0x19, 0x1A, 0x26, 0x27, 0x28, 0x29, 0x2A, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3A, 0x43, 0x44,
> + 0x45, 0x46, 0x47, 0x48, 0x49, 0x4A, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5A, 0x63,
> + 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, 0x7A,
> + 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8A, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97,
> + 0x98, 0x99, 0x9A, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, 0xA8, 0xA9, 0xAA, 0xB2, 0xB3, 0xB4,
> + 0xB5, 0xB6, 0xB7, 0xB8, 0xB9, 0xBA, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA,
> + 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, 0xD8, 0xD9, 0xDA, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7,
> + 0xE8, 0xE9, 0xEA, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA
> +};
> +
> +#endif /* _MEDIA_JPEG_ENC_REF_H_ */
> diff --git a/include/media/jpeg.h b/include/media/jpeg.h
> index a01e142e99a7..e07b20ef8665 100644
> --- a/include/media/jpeg.h
> +++ b/include/media/jpeg.h
> @@ -16,5 +16,9 @@
> #define JPEG_MARKER_DHP 0xde
> #define JPEG_MARKER_APP0 0xe0
> #define JPEG_MARKER_COM 0xfe
> +#define JPEG_LUM_HT 0x00
> +#define JPEG_CHR_HT 0x01
> +#define JPEG_DC_HT 0x00
> +#define JPEG_AC_HT 0x10

These are added without documentation. They are grouped as part of the markers, which
they are really not AFAICT. When the arrays are moved to v4l2-jpeg.c, then these defines
should be moved to media/v4l2-jpeg.h as well.

Actually, looking at the media/jpeg.h and media/v4l2-jpeg.h headers, I wonder if it
wouldn't make much more sense to first move the contents of jpeg.h to v4l2-jpeg.h and
drop the old jpeg.h header completely.

It makes no sense that we have two jpeg media headers. And that is also another
reason for not adding a third (jpeg-enc-reftables.h).

>
> #endif /* _MEDIA_JPEG_H_ */

Regards,

Hans