RE: [PATCH] i2c: rcar: add SMBus block read support

[Andrew Gabbasov <andrew_gabbasov@mentor.com>]

Hi Geert,

Thank you for your review!

> -----Original Message-----
> From: Geert Uytterhoeven <geert@linux-m68k.org>
> Sent: Tuesday, October 05, 2021 4:32 PM
> To: Gabbasov, Andrew <Andrew_Gabbasov@mentor.com>
> Cc: Linux-Renesas <linux-renesas-soc@vger.kernel.org>; Linux I2C <linux-i2c@vger.kernel.org>; Linux Kernel
> Mailing List <linux-kernel@vger.kernel.org>; Wolfram Sang <wsa+renesas@sang-engineering.com>; Surachari,
> Bhuvanesh <Bhuvanesh_Surachari@mentor.com>
> Subject: Re: [PATCH] i2c: rcar: add SMBus block read support
> 
> Hi Andrew,
> 
> On Wed, Sep 22, 2021 at 6:14 PM Andrew Gabbasov
> <andrew_gabbasov@mentor.com> wrote:
> > The smbus block read is not currently supported for rcar i2c devices.
> > This patchset adds the support to rcar i2c bus so that blocks of data
> > can be read using SMbus block reads.(using i2c_smbus_read_block_data()
> > function from the i2c-core-smbus.c).
> >
> > Inspired by commit 8e8782c71595 ("i2c: imx: add SMBus block read support")
> >
> > This patch (adapted) was tested with v4.14, but due to lack of real
> > hardware with SMBus block read operations support, using "simulation",
> > that is manual analysis of data, read from plain I2C devices with
> > SMBus block read request.
> >
> > Signed-off-by: Bhuvanesh Surachari <bhuvanesh_surachari@mentor.com>
> > Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
> 
> Thanks for your patch!
> 
> > --- a/drivers/i2c/busses/i2c-rcar.c
> > +++ b/drivers/i2c/busses/i2c-rcar.c
> > @@ -429,9 +431,16 @@ static bool rcar_i2c_dma(struct rcar_i2c_priv *priv)
> >                 /*
> >                  * The last two bytes needs to be fetched using PIO in
> >                  * order for the STOP phase to work.
> > +                *
> > +                * For SMBus block read the first byte was received using PIO.
> 
> So it might be easier to read, and more maintainable, to keep the
> old assignments:
> 
>     buf = priv->msg->buf;
>     len = priv->msg->len - 2;
> 
> and adjust them for SMBus afterwards:
> 
>     if (block_data) {
>             /* For SMBus block read the first byte was received using PIO */
>             buf++;
>             len--;
>     }
> 
> ?
> 
> >                  */
> > -               buf = priv->msg->buf;
> > -               len = priv->msg->len - 2;
> > +               if (block_data) {
> > +                       buf = priv->msg->buf + 1;
> > +                       len = priv->msg->len - 3;
> > +               } else {
> > +                       buf = priv->msg->buf;
> > +                       len = priv->msg->len - 2;
> > +               }
> >         } else {
> >                 /*
> >                  * First byte in message was sent using PIO.
> 
> And below we have another case handling buf and len :-(
> 
> So perhaps:
> 
>     buf = priv->msg->buf;
>     len = priv->msg->len;
> 
>     if (read) {
>             /*
>              * The last two bytes needs to be fetched using PIO in
>              * order for the STOP phase to work.
>              */
>             len -= 2;
>     }
>     if (!read || block_data) {
>             /* First byte in message was sent using PIO *
>             buf++;
>             len--;
>     }

Probably I was trying to minimize the changes ;-)

However, I agree with you that the whole code fragment can be simplified
and your variant indeed looks more clean and understandable.
Thank you for your suggestion, I'll submit version 2 of the patch
with this fragment changed.

Thanks!

Best regards,
Andrew