Re: [PATCH v13 4/9] media: uapi: Add CAMSS ISP configuration definition

From: Loic Poulain

Date: Wed Nov 04 2026 - 16:34:17 EST

Hi Jacopo,

On Wed, Nov 4, 2026 at 6:05 PM Jacopo Mondi
<jacopo.mondi@xxxxxxxxxxxxxxxx> wrote:
>
> Hi Loic
>
> On Tue, Nov 03, 2026 at 09:05:48PM +0100, Loic Poulain wrote:
> > Add the uapi header qcom-camss-config.h defining the ISP parameter
> > structures used by the CAMSS Offline Processing Engine (OPE) driver.
> > This includes structures for white balance, chroma enhancement and
> > color correction configuration.
> >
> > The block type enum is shared across CAMSS ISPs so that userspace only
> > has to deal with a single set of block identifiers.
> >
> > Reviewed-by: Hans de Goede <johannes.goede@xxxxxxxxxxxxxxxx>
> > Tested-by: Hans de Goede <johannes.goede@xxxxxxxxxxxxxxxx>
> > Signed-off-by: Loic Poulain <loic.poulain@xxxxxxxxxxxxxxxx>
> > ---
> > include/uapi/linux/qcom-camss-config.h | 217 +++++++++++++++++++++++++++++++++
> > 1 file changed, 217 insertions(+)
> >
> > diff --git a/include/uapi/linux/qcom-camss-config.h b/include/uapi/linux/qcom-camss-config.h
> > new file mode 100644
> > index 0000000000000000000000000000000000000000..543c1682f8aed3b07d7085f54a24eb632ea190a8
> > --- /dev/null
> > +++ b/include/uapi/linux/qcom-camss-config.h
>
> We have (kind of) recently introduced include/uapi/linux/media
>
> This file should probably be:
> include/uapi/linux/media/qcom/camss-config.h

Thanks for pointing this out. I'll move it according to your suggestion.

Regards,
Loic

>
> > @@ -0,0 +1,217 @@
> > +/* SPDX-License-Identifier: GPL-2.0-only WITH Linux-syscall-note */
> > +/*
> > + * Qualcomm CAMSS ISP Driver - Userspace API
> > + *
> > + * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
> > + */
> > +
> > +#ifndef _UAPI_LINUX_QCOM_CAMSS_CONFIG_H
> > +#define _UAPI_LINUX_QCOM_CAMSS_CONFIG_H
> > +
> > +#include <linux/types.h>
> > +#include <linux/media/v4l2-isp.h>
> > +
> > +/**
> > + * enum camss_params_block_type - CAMSS ISP parameter block identifiers
> > + *
> > + * Each value identifies one ISP processing block. The value is placed in
> > + * the @type field of &struct v4l2_isp_block_header.
> > + *
> > + * @CAMSS_PARAMS_WB_GAIN: white balance gains and offsets (CLC_WB),
> > + * &struct camss_params_wb_gain
> > + * @CAMSS_PARAMS_CHROMA_ENHAN: RGB to YUV colour transfer matrix
> > + * (CLC_CHROMA_ENHAN), &struct camss_params_chroma_enhan
> > + * @CAMSS_PARAMS_COLOR_CORRECT: colour correction matrix (CLC_CC),
> > + * &struct camss_params_color_correct
> > + * @CAMSS_PARAMS_GAMMA: per-channel gamma correction curves (CLC_GLUT),
> > + * &struct camss_params_gamma
> > + */
> > +enum camss_params_block_type {
> > + CAMSS_PARAMS_WB_GAIN = 1,
> > + CAMSS_PARAMS_CHROMA_ENHAN = 2,
> > + CAMSS_PARAMS_COLOR_CORRECT = 3,
> > + CAMSS_PARAMS_GAMMA = 4,
> > +};
> > +
> > +/* Number of entries in each gamma channel LUT. */
> > +#define CAMSS_GAMMA_LUT_SIZE 256
> > +
> > +/**
> > + * struct camss_params_wb_gain - White Balance gains
> > + *
> > + * Implements the CLC_WB pipeline module. The pipeline applies three
> > + * sequential operations per channel::
> > + *
> > + * 1. Subtract sub-offset (black-level subtraction)
> > + * 2. Multiply by gain (colour balance)
> > + * 3. Add add-offset (output pedestal)
> > + *
> > + * Gains are 15uQ10 (15-bit unsigned, 10 fractional bits). Offsets
> > + * are 16-bit unsigned, normalised to full input scale (65535 = 1.0)
> > + *
> > + * @header: block header; @header.type = CAMSS_PARAMS_WB_GAIN
> > + * @g_gain: green channel gain (15uQ10, 1024 = 1.0)
> > + * @b_gain: blue channel gain (15uQ10, 1024 = 1.0)
> > + * @r_gain: red channel gain (15uQ10, 1024 = 1.0)
> > + * @g_sub: green sub-offset, subtracted before gain (16u)
> > + * @b_sub: blue sub-offset, subtracted before gain (16u)
> > + * @r_sub: red sub-offset, subtracted before gain (16u)
> > + * @g_add: green add-offset, added after gain (16u)
> > + * @b_add: blue add-offset, added after gain (16u)
> > + * @r_add: red add-offset, added after gain (16u)
> > + * @_pad: padding for 64-bit alignment, must be zero
> > + */
> > +struct camss_params_wb_gain {
> > + struct v4l2_isp_block_header header;
> > + __u16 g_gain;
> > + __u16 b_gain;
> > + __u16 r_gain;
> > + __u16 g_sub;
> > + __u16 b_sub;
> > + __u16 r_sub;
> > + __u16 g_add;
> > + __u16 b_add;
> > + __u16 r_add;
> > + __u16 _pad[3];
> > +} __attribute__((aligned(8)));
> > +
> > +/**
> > + * struct camss_params_chroma_enhan - RGB to YUV colour transfer matrix
> > + *
> > + * Implements the CLC_CHROMA_ENHAN pipeline module. All coefficients are
> > + * signed 12-bit fixed-point Q3.8 (range roughly -8.0 to +7.996).
> > + *
> > + * RGB2Y - Luma (Y) coefficients::
> > + *
> > + * Y = v0 * R + v1 * G + v2 * B
> > + *
> > + * RGB2Cb - Chroma (Cb) coefficients::
> > + *
> > + * Cb = a x ((B - G) + b(R - G)) + KCb
> > + * with:
> > + * a = ap, when (B-G) + b(R-G) > 0; a = am, when (B-G) + b(R-G) ≤ 0;
> > + * b = bp when (R-G) > 0; b = bm when (R-G) ≤ 0
> > + *
> > + * RGB2Cr - Chroma (Cr) coefficients::
> > + *
> > + * Cr = c x ((R - G) + d(B - G)) + KCr
> > + * with:
> > + * c = cp, when (R-G) + d(B-G) > 0; c = cm, when (R-G) + d(B-G) ≤ 0
> > + * d = dp when (B-G) > 0; d = dm when (B-G) ≤ 0
> > + *
> > + * @header: generic block header; @header.type = CAMSS_PARAMS_CHROMA_ENHAN
> > + * @luma_v0: R-to-Y coefficient (12sQ8)
> > + * @luma_v1: G-to-Y coefficient (12sQ8)
> > + * @luma_v2: B-to-Y coefficient (12sQ8)
> > + * @luma_k: Y output offset (9s, 0 = no offset)
> > + * @coeff_ap: Cb positive coefficient (12sQ8)
> > + * @coeff_am: Cb negative coefficient (12sQ8)
> > + * @coeff_bp: Cb positive coefficient (12sQ8)
> > + * @coeff_bm: Cb negative coefficient (12sQ8)
> > + * @coeff_cp: Cr positive coefficient (12sQ8)
> > + * @coeff_cm: Cr negative coefficient (12sQ8)
> > + * @coeff_dp: Cr positive coefficient (12sQ8)
> > + * @coeff_dm: Cr negative coefficient (12sQ8)
> > + * @kcb: Cb output offset (11s)
> > + * @kcr: Cr output offset (11s)
> > + * @_pad: padding for 64-bit alignment, must be zero
> > + */
> > +struct camss_params_chroma_enhan {
> > + struct v4l2_isp_block_header header;
> > + __s16 luma_v0;
> > + __s16 luma_v1;
> > + __s16 luma_v2;
> > + __s16 luma_k;
> > + __s16 coeff_ap;
> > + __s16 coeff_am;
> > + __s16 coeff_bp;
> > + __s16 coeff_bm;
> > + __s16 coeff_cp;
> > + __s16 coeff_cm;
> > + __s16 coeff_dp;
> > + __s16 coeff_dm;
> > + __s16 kcb;
> > + __s16 kcr;
> > + __u16 _pad[2];
> > +} __attribute__((aligned(8)));
> > +
> > +/**
> > + * struct camss_params_color_correct - colour correction matrix
> > + *
> > + * Implements the CLC_CC pipeline module. The matrix computes::
> > + *
> > + * Out_ch0 (G) = a0*G + b0*B + c0*R + k0
> > + * Out_ch1 (B) = a1*G + b1*B + c1*R + k1
> > + * Out_ch2 (R) = a2*G + b2*B + c2*R + k2
> > + *
> > + * The @qfactor member selects the fixed-point format of all coefficients::
> > + *
> > + * 0 = 12sQ7 (range ~-16.0 .. +15.992)
> > + * 1 = 12sQ8 (range ~-8.0 .. +7.996)
> > + * 2 = 12sQ9 (range ~-4.0 .. +3.998)
> > + * 3 = 12sQ10 (range ~-2.0 .. +1.999)
> > + *
> > + * @header: block header; @header.type = CAMSS_PARAMS_COLOR_CORRECT
> > + * @a: G-input coefficients per output channel (12s;
> > + * a[0]=Out_G, a[1]=Out_B, a[2]=Out_R)
> > + * @b: B-input coefficients (12s)
> > + * @c: R-input coefficients (12s)
> > + * @k: per-output-channel offsets (typically 9s effective)
> > + * @qfactor: Q-format selector (2u), see above
> > + * @_pad: padding for 64-bit alignment, must be zero
> > + */
> > +struct camss_params_color_correct {
> > + struct v4l2_isp_block_header header;
> > + __s16 a[3];
> > + __s16 b[3];
> > + __s16 c[3];
> > + __s16 k[3];
> > + __u16 qfactor;
> > + __u16 _pad[3];
> > +} __attribute__((aligned(8)));
> > +
> > +/**
> > + * struct camss_params_gamma - per-channel gamma correction curves
> > + *
> > + * Implements the CLC_GLUT pipeline module, applied in the RGB domain. It
> > + * holds one independent lookup table per colour channel. Each table is a
> > + * direct (not segmented) map of input level to output level.
> > + *
> > + * Each table has @CAMSS_GAMMA_LUT_SIZE (256) entries of 16-bit unsigned
> > + * output. The number of significant output bits follows the width of the
> > + * engine's internal bus; the upper bits are ignored by the hardware.
> > + *
> > + * The module maps a 12-bit input to an 8-bit output: the top 8 input bits
> > + * index the table and the low 4 bits are used to linearly interpolate
> > + * between adjacent entries. Entry i therefore represents the output for
> > + * input level i/255 of full scale.
> > + * Output range: 0x00 = black, 0xFF = white.
> > + *
> > + * How to fill a curve (same on all three channels for pure luminance
> > + * gamma; different curves per channel additionally shift colour balance)::
> > + *
> > + * Identity (pass-through, gamma 1.0):
> > + * lut[i] = i; // i = 0..255
> > + *
> > + * Encode with gamma g (e.g. sRGB-like, g = 2.2):
> > + * lut[i] = round(pow(i / 255.0, 1.0 / g) * 255.0);
> > + *
> > + * @header: block header; @header.type = CAMSS_PARAMS_GAMMA
> > + * @glut: green channel gamma curve
> > + * @blut: blue channel gamma curve
> > + * @rlut: red channel gamma curve
> > + */
> > +struct camss_params_gamma {
> > + struct v4l2_isp_block_header header;
> > + __u16 glut[CAMSS_GAMMA_LUT_SIZE];
> > + __u16 blut[CAMSS_GAMMA_LUT_SIZE];
> > + __u16 rlut[CAMSS_GAMMA_LUT_SIZE];
> > +} __attribute__((aligned(8)));
> > +
> > +#define CAMSS_PARAMS_OPE_MAX_PAYLOAD \
> > + (sizeof(struct camss_params_wb_gain) +\
> > + sizeof(struct camss_params_chroma_enhan) +\
> > + sizeof(struct camss_params_color_correct) +\
> > + sizeof(struct camss_params_gamma))
> > +
> > +#endif /* _UAPI_LINUX_QCOM_CAMSS_CONFIG_H */
> >
> > --
> > 2.34.1
> >
> >