RE: [PATCH v8 4/4] media: i2c: imx334: update pixel and link frequency

[Shravan.Chippa]

> -----Original Message-----
> From: Jacopo Mondi <jacopo.mondi@xxxxxxxxxxxxxxxx>
> Sent: 06 January 2023 04:35 PM
> To: shravan Chippa - I35088 <Shravan.Chippa@xxxxxxxxxxxxx>
> Cc: paul.j.murphy@xxxxxxxxx; daniele.alessandrelli@xxxxxxxxx;
> mchehab@xxxxxxxxxx; linux-media@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; Sakari Ailus <sakari.ailus@xxxxxx>
> Subject: Re: [PATCH v8 4/4] media: i2c: imx334: update pixel and link frequency
> 
> [You don't often get email from jacopo.mondi@xxxxxxxxxxxxxxxx. Learn why
> this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the
> content is safe
> 
> Hi Shravan
> 
> On Fri, Jan 06, 2023 at 12:59:31PM +0530, shravan kumar wrote:
> > From: Shravan Chippa <shravan.chippa@xxxxxxxxxxxxx>
> >
> > Update pixel_rate and link frequency for 1920x1080@30 while changing
> > mode.
> >
> > Add dummy ctrl cases for pixel_rate and link frequency to avoid error
> > while changing the modes dynamically
> >
> > Suggested-by: Sakari Ailus <sakari.ailus@xxxxxx>
> > Signed-off-by: Shravan Chippa <shravan.chippa@xxxxxxxxxxxxx>
> > ---
> >  drivers/media/i2c/imx334.c | 11 ++++++++++-
> >  1 file changed, 10 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/media/i2c/imx334.c b/drivers/media/i2c/imx334.c
> > index 0315e1c9541d..8c3ba660abae 100644
> > --- a/drivers/media/i2c/imx334.c
> > +++ b/drivers/media/i2c/imx334.c
> > @@ -50,6 +50,7 @@
> >
> >  /* CSI2 HW configuration */
> >  #define IMX334_LINK_FREQ     891000000
> 
> I guess you want to rename this one to  IMX334_LINK_FREQ_891M
> 
> Give our previous discussion this seems correct for the following mode
> 
>         {
>                 .width = 3840,
>                 .height = 2160,
>                 .hblank = 560,
>                 .vblank = 2340,
>                 .vblank_min = 90,
>                 .vblank_max = 132840,
>                 .pclk = 594000000,
>                 .link_freq_idx = 0,
>                 .reg_list = {
>                         .num_of_regs = ARRAY_SIZE(mode_3840x2160_regs),
>                         .regs = mode_3840x2160_regs,
>                 },
>         }, {
> 
>         duration: (3840+560) * (2160+2340)  / 594000000 = 33sec = 30FPS
>         link_freq (3840+560) * (2160+2340)  * 30 * 12 / 8 = 891000000
> 
> Which works well if we use min_vblank = 90 for 60FPS
> 
>         duration: (3840+560) * (2160+90)  / 594000000 = 0.16 = 60 FPS
>         link_freq (3840+560) * (2160+90)  * 60 * 12 / 8 = 891000000
> 
> 
> > +#define IMX334_LINK_FREQ_445M        445500000
> 
> But this doesn't work well for me
> 
>         {
>                 .width = 1920,
>                 .height = 1080,
>                 .hblank = 280,
>                 .vblank = 1170,
>                 .vblank_min = 90,
>                 .vblank_max = 132840,
>                 .pclk = 74250000,
>                 .link_freq_idx = 1,
>                 .reg_list = {
>                         .num_of_regs = ARRAY_SIZE(mode_1920x1080_regs),
>                         .regs = mode_1920x1080_regs,
>                 },
>         },
> 
>         duration: (1920+280) * (1080+1170) / 74250000 = 66msec = 16FPS
>         link_freq = (1920+280) * (1080+1170) * 60 * 10 / 8 = 371250000
> 
> Do you agree with the above or have I missed something ?
> 
> I understand you get 30 FPS with the 1920*1080 mode so could you please
> check in the newly introduce mode register table what are the actual values for
> the blankings and compute the pixel_rate and link_freq accordingly ?

I will try to correct the hblank and vblank_min.

Thanks,
Shravan

> 
> >  #define IMX334_NUM_DATA_LANES        4
> >
> >  #define IMX334_REG_MIN               0x00
> > @@ -145,6 +146,7 @@ struct imx334 {
> >
> >  static const s64 link_freq[] = {
> >       IMX334_LINK_FREQ,
> > +     IMX334_LINK_FREQ_445M,
> >  };
> >
> >  /* Sensor mode registers */
> > @@ -468,7 +470,7 @@ static const struct imx334_mode supported_modes[]
> = {
> >               .vblank_min = 90,
> >               .vblank_max = 132840,
> >               .pclk = 74250000,
> > -             .link_freq_idx = 0,
> > +             .link_freq_idx = 1,
> >               .reg_list = {
> >                       .num_of_regs = ARRAY_SIZE(mode_1920x1080_regs),
> >                       .regs = mode_1920x1080_regs, @@ -598,6 +600,11
> > @@ static int imx334_update_controls(struct imx334 *imx334,
> >       if (ret)
> >               return ret;
> >
> > +     ret = __v4l2_ctrl_modify_range(imx334->pclk_ctrl, mode->pclk,
> > +                                    mode->pclk, 1, mode->pclk);
> > +     if (ret)
> > +             return ret;
> > +
> >       ret = __v4l2_ctrl_modify_range(imx334->hblank_ctrl, mode->hblank,
> >                                      mode->hblank, 1, mode->hblank);
> >       if (ret)
> > @@ -698,6 +705,8 @@ static int imx334_set_ctrl(struct v4l2_ctrl *ctrl)
> >               pm_runtime_put(imx334->dev);
> >
> >               break;
> > +     case V4L2_CID_PIXEL_RATE:
> > +     case V4L2_CID_LINK_FREQ:
> >       case V4L2_CID_HBLANK:
> 
> Same question as for patch 1/4: Do we need these safety checks for read-only
> controls ?
> 
> Thanks
>   j
> 
> >               ret = 0;
> >               break;
> > --
> > 2.34.1
> >