Re: [PATCH 2/3] lx6464es: dump MAC in standard form

[Takashi Iwai <tiwai@suse.de>]

At Wed, 29 May 2013 14:31:07 +0300,
Andy Shevchenko wrote:
> 
> On Wed, 2013-05-29 at 13:03 +0200, Takashi Iwai wrote:
> 
> > > > > -	sprintf(card->shortname, "LX6464ES %02X.%02X.%02X.%02X.%02X.%02X",
> > > > > -		chip->mac_address[0], chip->mac_address[1], chip->mac_address[2],
> > > > > -		chip->mac_address[3], chip->mac_address[4], chip->mac_address[5]);
> > > > > +	sprintf(card->shortname, "LX6464ES %pM", chip->mac_address);
> > > > 
> > > > This will change from the upper to the lower case, thus it'll lead to
> > > > an incompatible name string, unfortunately.
> > > 
> > > Who is the user of this string?
> > 
> > It's exposed to the user-space via ioctl and it can be used as an id
> > string.  So this will break user-space compatibility.
> 
> MAC is theoretically set of arbitrary 6 bytes.
> Only what I can see here is the difference between '.' and ':' as a
> delimiter. But the MAC address form is defined in IEEE 802 standards.

It doesn't matter whether it's a valid MAC representation or not.
The problem is only that this patch will change the string
representation, thus it becomes incompatible with older kernels.
That's the only point.

> Anyway, I am okay if you reject this one.

Good :)


thanks,

Takashi

> -- 
> Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> Intel Finland Oy
>