Re: [PATCH RFC 3/4] media: i2c: imx678: Add imx662 support to the driver
From: Jai Luthra
Date: Thu Sep 03 2026 - 00:21:08 EST
Hi Dave,
Quoting Dave Stevenson (2026-08-05 21:39:18)
> IMX662 is a 1080p Starvis 2 sensor that follows the same programming
> pattern as IMX678, so add it to the newly parameterised driver.
>
> FIXME: Update the common registers based on the Sony spreadsheet.
>
> Signed-off-by: Dave Stevenson <dave.stevenson@xxxxxxxxxxxxxxx>
> ---
> drivers/media/i2c/imx678.c | 64 ++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 64 insertions(+)
>
> diff --git a/drivers/media/i2c/imx678.c b/drivers/media/i2c/imx678.c
> index 40ba6661b366..9083f2744510 100644
> --- a/drivers/media/i2c/imx678.c
> +++ b/drivers/media/i2c/imx678.c
> @@ -673,6 +673,47 @@ const struct imx678_variant imx678_variant_def = {
> .hmax_min = imx678_min_hmax_4lane,
> };
>
> +static const struct cci_reg_sequence imx662_common_regs[] = {
> + { IMX678_REG_WDMODE, 0x00 },
> + { IMX678_REG_MDBIT, 0x01 },
> + { IMX678_REG_XXS_DRV, 0x00 },
> +};
> +
> +static const u16 imx662_min_hmax_4lane[] = {
> + [IMX678_LINK_FREQ_297MHZ] = 990,
> + [IMX678_LINK_FREQ_360MHZ] = 990,
> + [IMX678_LINK_FREQ_445MHZ] = 990,
> + [IMX678_LINK_FREQ_594MHZ] = 990,
> + [IMX678_LINK_FREQ_720MHZ] = 990,
> + [IMX678_LINK_FREQ_891MHZ] = 990,
> + [IMX678_LINK_FREQ_1039MHZ] = 990,
> + [IMX678_LINK_FREQ_1188MHZ] = 990,
> +};
> +
> +const struct imx678_variant imx662_variant_def = {
> + .name = "imx662",
> + .id_reg = IMX678_REG_MODULE_ID,
> + .id_value = 0x296,
> + .native_area = {
> + .top = 0,
> + .left = 0,
> + .width = 1937,
> + .height = 1101,
> + },
> + .active_area = {
> + .top = 0,
> + .left = 0,
> + .width = 1936,
> + .height = 1096,
> + },
> + .pixel_rate = 222750000,
> + .pix_per_clk = 3,
> + .common_regs = imx662_common_regs,
> + .num_common_regs = ARRAY_SIZE(imx662_common_regs),
> + .vmax_default = 1096 + 40,
>From the IMX662 SRM I see the following constraints for minimum:
VTTL >= PIX_VWIDTH + 70
VTTL >= 820
And the table giving round FPS numbers like 30/60/90 assumes VMAX = 1250.
Did you manually verify that setting blanking to 40 also works?
Thanks,
Jai
> + .hmax_min = imx662_min_hmax_4lane,
> +};
> +
> struct imx678_model_info {
> enum imx678_type type;
> const u32 *codes;
> @@ -703,6 +744,26 @@ static const struct imx678_model_info imx678_autodetect_info = {
> .auto_detect_mono = &imx678_aamr_info,
> };
>
> +static const struct imx678_model_info imx662_aaqr_info = {
> + .type = IMX678_COLOR,
> + .codes = codes_bayer,
> + .num_codes = ARRAY_SIZE(codes_bayer),
> + .variant = &imx662_variant_def,
> +};
> +
> +static const struct imx678_model_info imx662_aamr_info = {
> + .type = IMX678_MONOCHROME,
> + .codes = codes_monochrome,
> + .num_codes = ARRAY_SIZE(codes_monochrome),
> + .variant = &imx662_variant_def,
> +};
> +
> +static const struct imx678_model_info imx662_autodetect_info = {
> + .variant = &imx662_variant_def,
> + .auto_detect_colour = &imx662_aaqr_info,
> + .auto_detect_mono = &imx662_aamr_info,
> +};
> +
> static const char * const imx678_supply_name[] = {
> "avdd", /* Analog (3.3V) supply */
> "dvdd", /* Digital Core (1.1V) supply */
> @@ -1470,8 +1531,11 @@ static const struct dev_pm_ops imx678_pm_ops = {
> static const struct of_device_id imx678_of_match[] = {
> { .compatible = "sony,imx678-aamr", .data = &imx678_aamr_info },
> { .compatible = "sony,imx678-aaqr", .data = &imx678_aaqr_info },
> + { .compatible = "sony,imx662-aamr", .data = &imx662_aamr_info },
> + { .compatible = "sony,imx662-aaqr", .data = &imx662_aaqr_info },
> /* for non-conforming DTs that rely on runtime check */
> { .compatible = "sony,imx678", .data = &imx678_autodetect_info },
> + { .compatible = "sony,imx662", .data = &imx662_autodetect_info },
> { /* sentinel */ }
> };
>
>
> --
> 2.34.1
>