Re: [PATCH v9 2/4] media: i2c: Add MAX9286 driver

[Manivannan Sadhasivam]

On Thu, May 14, 2020 at 01:59:35PM +0200, Geert Uytterhoeven wrote:
> Hi Kieran,
> 
> On Thu, May 14, 2020 at 1:47 PM Kieran Bingham
> <kieran.bingham+renesas@xxxxxxxxxxxxxxxx> wrote:
> > On 14/05/2020 11:13, Manivannan Sadhasivam wrote:
> > > On Thu, May 14, 2020 at 11:02:53AM +0100, Kieran Bingham wrote:
> > >> On 12/05/2020 19:17, Manivannan Sadhasivam wrote:
> > >>> On Tue, May 12, 2020 at 04:51:03PM +0100, Kieran Bingham wrote:
> > >>>> The MAX9286 is a 4-channel GMSL deserializer with coax or STP input and
> > >>>> CSI-2 output. The device supports multicamera streaming applications,
> > >>>> and features the ability to synchronise the attached cameras.
> > >>>>
> > >>>> CSI-2 output can be configured with 1 to 4 lanes, and a control channel
> > >>>> is supported over I2C, which implements an I2C mux to facilitate
> > >>>> communications with connected cameras across the reverse control
> > >>>> channel.
> > >>>>
> > >>>> Signed-off-by: Jacopo Mondi <jacopo+renesas@xxxxxxxxxx>
> > >>>> Signed-off-by: Kieran Bingham <kieran.bingham+renesas@xxxxxxxxxxxxxxxx>
> > >>>> Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@xxxxxxxxxxxxxxxx>
> > >>>> Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@xxxxxxxxxxxx>
> 
> > >>>> --- /dev/null
> > >>>> +++ b/drivers/media/i2c/max9286.c
> 
> > >>>> +static int max9286_register_gpio(struct max9286_priv *priv)
> > >>>> +{
> > >>>> +  struct device *dev = &priv->client->dev;
> > >>>> +  struct gpio_chip *gpio = &priv->gpio;
> > >>>> +  int ret;
> > >>>> +
> > >>>> +  static const char * const names[] = {
> > >>>> +          "GPIO0OUT",
> > >>>> +          "GPIO1OUT",
> > >>>> +  };
> > >>>> +
> > >>>> +  /* Configure the GPIO */
> > >>>> +  gpio->label = dev_name(dev);
> > >>>
> > >>> So if you have more than one MAX9286 in a system, all gpiochips will appear
> > >>> with the same name. I'd recommend to append the index to distinguish properly.
> > >>
> > >> Ah yes, that's a good point, and I think I've even seen that.
> > >>
> > >> I'll fix it now.
> >
> > Oh, in fact actually this doesn't.
> >
> > gpiodetect prints:
> >
> > gpiochip10 [4-004c] (2 lines)
> > gpiochip11 [4-006c] (2 lines)
> >
> > and mostly references them as gpiochip10 and gpiochip11.
> 
> Indeed, dev_name() should be different for each instance.
> 

Ah, my bad! Somehow I got confused that this delivers static name... Sorry for
the noise, Kieran.

> > However,
> >
> > > [    2.318533] gpio gpiochip11: Detected name collision for GPIO name 'GPIO0OUT'
> > > [    2.325739] gpio gpiochip11: Detected name collision for GPIO name 'GPIO1OUT'
> >
> > That seems to be more of a problem for the gpio library, so I think I'll
> > just drop the const names. I don't think they add much value.
> 

Well, I should've pointed this instead of above...

(lack of coffee)

> These are the line names.  If they're not unique, a warning is printed,
> but they are still registered.
> So probably you want to use kasprintf("%s.%s", dev_name(dev), names[i]) to
> generate names.
> 

Ack.

I think you should CC Linus W for next iteration to get review for gpiolib
implementation.

Thanks,
Mani

> See "[PATCH] gpiolib: Document that GPIO line names are not globally unique"
> (https://lore.kernel.org/linux-gpio/20200511101828.30046-1-geert+renesas@xxxxxxxxx/)
> to clear up the details.
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> -- 
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds