Re: [PATCH v5 3/3] media: i2c: imx471: Add Sony IMX471 image sensor driver

[Kate Hsuan]

Hi Kieran and Tarang,

Thank you for reviewing.

On Wed, Jun 24, 2026 at 10:09 AM Tarang Raval
<tarang.raval@xxxxxxxxxxxxxxxxx> wrote:
>
> Hi Kieran,
>
> > Quoting Kate Hsuan (2026-06-24 04:35:08)
> > > Add a new driver for Sony imx471 camera sensor. It is based on
> > > Jimmy Su <jimmy.su@xxxxxxxxx> implementation and the driver can be found
> > > in the following URL.
> > > https://github.com/intel/ipu6-drivers/commits/master/drivers/media/i2c/imx471.c
> > >
> > > This sensor can be found on Lenovo X1 Carbon G14, X9-14 and X9-15 laptops
> > > and it is a part of IPU7 solution. The driver was tested on Lenovo X1
> > > Carbon G14, X9-14 and X9-15 laptops.
> > >
> > > Signed-off-by: Kate Hsuan <hpa@xxxxxxxxxx>
> > > ---
> >
> > ... <snip>
> >
> > > +#define IMX471_REG_EXCK_FREQ                   CCI_REG16(0x0136)
> > > +#define IMX471_EXCK_FREQ(n)                    ((n) * 256)     /* n in MHz */
> > >
> >
> > <snip>
> >
> > > +
> > > +static const struct cci_reg_sequence imx471_global_regs[] = {
> > > +       { IMX471_REG_EXCK_FREQ, IMX471_EXCK_FREQ(19.2) },
> >
> >
> > Does this work? Is this a compile time constant multiplying floating
> > point with the compiler, or something happening in the kernel?
> >
> > I'll be happy if it does, but it catches my attention as something I
> > thought we couldn't do.
>
> I tested this with GCC. It evaluates the expression at compile time and
> folds it into the integer constant 4915. The generated assembly/object
> file contains .quad 4915 in the initialized data and no floating-point
> instructions are emitted, so the kernel only ever sees the integer value.

The compiler didn't complain anything and covert it to a integer but
putting a floating point here is not a good way.
I focued on finding a register value without noticing that it is a
floating point
I'll set a value IMX471_REG_EXCK_FREQ directly and  drop IMX471_EXCK_FREQ.


>
> Best Regards,
> Tarang
>


-- 
BR,
Kate