Re: [PATCH 1/7] sdhci: Rework some of the quirk behaviour

[Alan Cox <alan@linux.intel.com>]

> >  drivers/mmc/host/Kconfig           |   11 ++
> >  drivers/mmc/host/Makefile          |    1 
> >  drivers/mmc/host/sdhci-intel-mid.c |  163
> > ++++++++++++++++++++++++++++++++++++
> 
> Why are those added here and not in patch 3/7?

I thought it would be useful to add the hooks and show how they are
used in one - I can split that easily enough.


> > + * ADMA operation is disabled for Moorestown platform due to
> > + * hardware bugs.
> > + */
> > +static int mrst_hc0_probe(struct sdhci_pci_chip *chip)
> > +{
> > +	/*
> > +	 * slots number is fixed here for MRST as SDIO3 is never
> > used and has
> > +	 * hardware bugs.
> > +	 */
> > +	chip->num_slots = 1;
> > +	return 0;
> > +}
> 
> Why is this function here and not in sdhci-intel-mid.c?

So it follows the pattern of the other drivers, so it doesn't need an
extra exported symbol (at about ten times the size of the function).

We could move probe into host_ops I guess but that means tweaking all
the other drivers which I cannot test.

I guess the other way to do it would be not to have a separate file but
to stick it all in sdhci-pci ?



> >  		mmc->max_blk_size = (caps & SDHCI_MAX_BLOCK_MASK)
> > >> SDHCI_MAX_BLOCK_SHIFT;
> > -		if (mmc->max_blk_size >= 3) {
> > +		if (mmc->max_blk_size > 3) {
> 
> Why this change? Not mentioned in the changelog. And wrong according
> to the simplified standard v2. Is it V3 material?

That's got in by mistake. I thought I'd got all the spec related bits
stripped out.

I actually suspect for the more general case of cleaning up some of the
other driver quirks we need a couple more callbacks but its hard to
sure.