AW: [PATCH v3] i2c: i2c-gpio: Enhance driver for buses with shared SCL

["Markus Stockhausen" <markus.stockhausen@gmx.de>]

> Von: Linus Walleij <linusw@kernel.org> 
> Gesendet: Mittwoch, 27. Mai 2026 12:25
> An: Bartosz Golaszewski <brgl@kernel.org>
>  Betreff: Re: [PATCH v3] i2c: i2c-gpio: Enhance driver for buses with shared SCL
>
> On Thu, May 21, 2026 at 10:59 AM Bartosz Golaszewski <brgl@kernel.org> wrote:
>
> > >  /*
> > >   * Toggle SDA by changing the output value of the pin. This is only
> > >   * valid for pins configured as open drain (i.e. setting the value
> > > @@ -53,7 +72,7 @@ static void i2c_gpio_setscl_val(void *data, int state)
> > >  {
> > >         struct i2c_gpio_private_data *priv = data;
> > >
> > > -       gpiod_set_value_cansleep(priv->scl, state);
> > > +       gpiod_set_value_cansleep(priv->scl->gpio, state);
> >
> > That one bothers me a bit. We're driving a clock line but may end up
> > sleeping? That doesn't sound right. We typically do:
> >
> > setscl();
> > udelay();
> >
> > I know it's been like this before and maybe I'm not understanding the
> > whole picture so feel free to disregard the comment.
>
> I think there are people driving I2C on slow buses, i.e. a bit-banged
> GPIO I2C on a GPIO expander which is itself on I2C.
>
> I'm not saying such hardware design is a good idea, but I think
> I've seen it...
>
> The only way to make that work is to use gpiod_set_value_cansleep().

Gentle ping.

Thanks.

Markus