Hi,
On Fri, Apr 20, 2018 at 11:44:18AM +0200, Daniel Mack wrote:
> Add v4l2 controls to report the pixel and MIPI link rates of each mode.
> The camss camera subsystem needs them to set up the correct hardware
> clocks.
>
> Tested on msm8016 based hardware.
>
> Signed-off-by: Daniel Mack <daniel@xxxxxxxxxx>
> ---
> drivers/media/i2c/ov5640.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++
> 1 file changed, 77 insertions(+)
>
> diff --git a/drivers/media/i2c/ov5640.c b/drivers/media/i2c/ov5640.c
> index 96f1564abdf5..78669ed386cd 100644
> --- a/drivers/media/i2c/ov5640.c
> +++ b/drivers/media/i2c/ov5640.c
> @@ -91,6 +91,20 @@
> #define OV5640_REG_SDE_CTRL5 0x5585
> #define OV5640_REG_AVG_READOUT 0x56a1
>
> +#define OV5640_LINK_FREQ_111 0
> +#define OV5640_LINK_FREQ_166 1
> +#define OV5640_LINK_FREQ_222 2
> +#define OV5640_LINK_FREQ_333 3
> +#define OV5640_LINK_FREQ_666 4
> +
> +static const s64 link_freq_menu_items[] = {
> + 111066666,
> + 166600000,
> + 222133333,
> + 332200000,
> + 666400000,
> +};
> +
> enum ov5640_mode_id {
> OV5640_MODE_QCIF_176_144 = 0,
> OV5640_MODE_QVGA_320_240,
> @@ -167,12 +181,18 @@ struct ov5640_mode_info {
> enum ov5640_downsize_mode dn_mode;
> u32 width;
> u32 height;
> + u32 pixel_rate;
> + u32 link_freq_idx;
> const struct reg_value *reg_data;
> u32 reg_data_size;
> };
>
> struct ov5640_ctrls {
> struct v4l2_ctrl_handler handler;
> + struct {
> + struct v4l2_ctrl *link_freq;
> + struct v4l2_ctrl *pixel_rate;
> + };
> struct {
> struct v4l2_ctrl *auto_exp;
> struct v4l2_ctrl *exposure;
> @@ -732,6 +752,8 @@ static const struct ov5640_mode_info ov5640_mode_init_data = {
> .dn_mode = SUBSAMPLING,
> .width = 640,
> .height = 480,
> + .pixel_rate = 27766666,
> + .link_freq_idx = OV5640_LINK_FREQ_111,
I'm not sure where this is coming from, but on a parallel sensor I
have a quite different pixel rate.
I have a serie ongoing that tries to deal with this, hopefully in
order to get rid of all the clock setup done in the initialiasation
array.
See https://patchwork.linuxtv.org/patch/48710/ for the patch and
https://www.spinics.net/lists/linux-media/msg132201.html for a
discussion on what the clock tree might look like on a MIPI-CSI bus.
Feel free to step in the discussion.
Maxime
--
Maxime Ripard, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
Attachment:
signature.asc
Description: PGP signature
