Re: [PATCH v2 1/2] i2c: add PEC error event

[Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>]

On Fri, Jul 24, 2020 at 1:46 AM Wolfram Sang <wsa@kernel.org> wrote:
>
> On Fri, Jul 17, 2020 at 02:31:54PM +0530, Rayagonda Kokatanur wrote:
> > Add new event I2C_SLAVE_PEC_ERR to list of slave events.
> > This event will be used by slave bus driver to indicate
> > PEC error to slave client or backend driver.
> >
> > Signed-off-by: Rayagonda Kokatanur <rayagonda.kokatanur@broadcom.com>
>
> Definately needs documentation in Documentation/i2c/slave-interface.rst.

Okay, I will update the doc.

>
> What is a backend supposed to do? Does 'value' have a meaning?

According to SMBUS spec, during slave receive transfer, it is
encouraged to issue NACK if the received PEC is not correct.
Also PEC errors discovered above the data link layer may also be
indicated with a NACK if the device is fast enough to discover and
indicate the error when the NACK is due.

If a device doesn't support issuing NACK and issuing NACK above the
data link layer is not possible because the device is not fast then in
that case we can send error code (I2C_SLAVE_PEC_ERR ) to the backend
driver. Backend drivers which support PEC should check for this error
code and take action such as discarding the data.

Best regards,
Rayagonda

>
> > ---
> >  include/linux/i2c.h | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/include/linux/i2c.h b/include/linux/i2c.h
> > index b8b8963f8bb9..e04acd04eb48 100644
> > --- a/include/linux/i2c.h
> > +++ b/include/linux/i2c.h
> > @@ -370,6 +370,7 @@ enum i2c_slave_event {
> >       I2C_SLAVE_READ_PROCESSED,
> >       I2C_SLAVE_WRITE_RECEIVED,
> >       I2C_SLAVE_STOP,
> > +     I2C_SLAVE_PEC_ERR,
> >  };
> >
> >  int i2c_slave_register(struct i2c_client *client, i2c_slave_cb_t slave_cb);
> > --
> > 2.17.1
> >