Re: [PATCH 3/8 v3] i2c: i801: check INTR after every transaction

[Jean Delvare <khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org>]

On Thu, 28 Jun 2012 15:51:34 +0800, Daniel Kurtz wrote:
> On Thu, Jun 28, 2012 at 12:07 AM, Jean Delvare <khali-PUYAD+kWke1g9hUCZPvPmw@public.gmane.org> wrote:
> > On Wed, 27 Jun 2012 21:54:10 +0800, Daniel Kurtz wrote:
> >> +
> >> +     if (timeout > MAX_RETRIES)
> >> +             dev_dbg(&priv->pci_dev->dev, "INTR Timeout!\n");
> >> +
> >> +     outb_p(status, SMBHSTSTS(priv));
> >
> > Wouldn't it be more correct to only write the INTR bit? Writing back
> > the whole 8 bits frightens me a little especially because of bit
> > INUSE_STS. If we ever want to support this feature, I think we have to
> > first ban writing back the whole status to register HST_STS. And this
> > is the only place where we still do AFAICS.
> 
> It looks like the current code does it this way to clear any error
> bits that may be set in the timeout case (errors set, but no INTR).
> For example, if there was a bus / arbitration error while waiting for
> PEC.
> 
> Perhaps we can split the difference (I tested this and it has no
> obvious regression):
> 
> +     /* Clear INTR, or in case of timeout, any other status bits. */
> +     outb_p(status & STATUS_FLAGS, SMBHSTSTS(priv));
> 
> But in a separate patch...
> 

I agree, this would be a reasonable compromise.

-- 
Jean Delvare