Re: [PATCH 1/9] rt2800: prepare for unification of EEPROM support code

[Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>]

On Sunday 08 November 2009 15:16:00 Gertjan van Wingerde wrote:
> On 11/08/09 14:38, Bartlomiej Zolnierkiewicz wrote:
> > From: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
> > Subject: [PATCH] rt2800: prepare for unification of EEPROM support code
> > 
> > * Factor out common code from rt2800[pci,usb]_validate_eeprom()
> >   to rt2800_validate_eeprom().
> > 
> > * Fix interface specific comment in rt2800[pci,usb]_validate_eeprom().
> > 
> > * Enclose interface specific code in rt2800[pci,usb]_init_eeprom()
> >   with rt2x00_intf_is_[pci,usb]() checks.
> > 
> > Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
> > ---
> > on top of "rt2x00: Add dynamic detection of eFuse EEPROM in rt2800pci."
> > patch from Gertjan
> > 
> >  drivers/net/wireless/rt2x00/rt2800pci.c |   51 ++++++++++++++++++--------------
> >  drivers/net/wireless/rt2x00/rt2800usb.c |   40 +++++++++++++++----------
> >  2 files changed, 55 insertions(+), 36 deletions(-)
> > 
> 
> <snip>
> 
> > @@ -1234,8 +1241,10 @@ static int rt2800pci_init_eeprom(struct 
> >  	    !rt2x00_rf(&rt2x00dev->chip, RF2750) &&
> >  	    !rt2x00_rf(&rt2x00dev->chip, RF3020) &&
> >  	    !rt2x00_rf(&rt2x00dev->chip, RF2020) &&
> > -	    !rt2x00_rf(&rt2x00dev->chip, RF3021) &&
> > -	    !rt2x00_rf(&rt2x00dev->chip, RF3022)) {
> > +	    (rt2x00_intf_is_usb(rt2x00dev) ||
> > +	     (rt2x00_intf_is_pci(rt2x00dev) &&
> > +	      !rt2x00_rf(&rt2x00dev->chip, RF3021) &&
> > +	      !rt2x00_rf(&rt2x00dev->chip, RF3022)))) {
> >  		ERROR(rt2x00dev, "Invalid RF chipset detected.\n");
> >  		return -ENODEV;
> >  	}
> 
> Can we leave this as-is. There are rt2800usb devices that support these RF chipsets as well (I am now the proud owner of such a device), so there is no
> need to restrict the use of these RF chipsets to just PCI devices.

Please note that the patch doesn't change the current behavior of rt2800usb
and for bisectability reasons I would strongly prefer for it to stay this way.

> So, if we leave this as-is then the EEPROM unification ensures rt2800usb will "support" these RF chipsets as well.

Lets work in incremental fashion -- when there is a need to do such change
lets just do it against current tree and apply it as soon as possible (this
way we won't be spending time on micro-decisions).

IOW just make a post- (or pre-) patch and I'll deal with it.

Thanks.
-- 
Bartlomiej Zolnierkiewicz