Re: [PATCH RFC 4/4] media: i2c: imx678: Add imx675 support to the driver

[Jai Luthra]

Quoting Dave Stevenson (2026-09-03 17:24:19)
[...]
> > > +const struct imx678_variant imx675_variant_def = {
> > > +       .name = "imx675",
> > > +       .id_reg = IMX678_REG_MODULE_ID,
> > > +       .id_value = 0x296,
> > > +       .native_area = {
> > > +               .top = 0,
> > > +               .left = 0,
> > > +               .width = 2609,
> > > +               .height = 1984,
> > > +       },
> > > +       .active_area = {
> > > +               .top = 0,
> > > +               .left = 0,
> > > +               .width = 2608,
> > > +               .height = 1984,
> > > +       },
> > > +       .pixel_rate = 594000000,
> > > +       .pix_per_clk = 8,
> > > +       .common_regs = imx675_common_regs,
> > > +       .num_common_regs = ARRAY_SIZE(imx675_common_regs),
> > > +       .vmax_default = 1984 + 40,
> >
> > Same doubt for the VBLANK = 40 here as IMX662, although I don't have a
> > datasheet for IMX675.
> 
> Datasheet lists
> 
> VTTL (1farame line length or VMAX) ≥ PIX_VWIDTH + 70
> Set VTTL to 940 or more.
> VTTL ≥ 940
> 
> But verified as working at +40. (I have a module from Soho Enterprises).
> We can go for the more conservative values if preferred. From memory,
> on IMX662 it'll lose about 10% (66fps at 40, ~60fps at 70).
> 

Ah if it's tested by you I'm fine with going with 40 for now.

One nitpick for your next version, maybe define VBLANK_MIN (40) and use that
along with active_area.height?

Thanks,
Jai