Hi,
On Wed, Jun 13, 2012 at 12:04:54AM -0400, Mike Frysinger wrote:
> On Tuesday 12 June 2012 04:08:05 Felipe Balbi wrote:
> > On Sun, Jun 10, 2012 at 11:16:25PM -0400, Mike Frysinger wrote:
> > > On Tue, Jun 5, 2012 at 7:30 AM, Felipe Balbi <balbi@xxxxxx> wrote:
> > > > On Wed, May 30, 2012 at 02:11:17PM +0800, Bob Liu wrote:
> > > >> --- a/drivers/usb/musb/blackfin.h
> > > >> +++ b/drivers/usb/musb/blackfin.h
> > > >> @@ -84,4 +84,10 @@ static void dump_fifo_data(u8 *buf, u16 len)
> > > >>
> > > >> static struct timer_list musb_conn_timer;
> > > >>
> > > >> +#ifndef CONFIG_BF60x
> > > >> +#define vrsel_set_value(val) gpio_set_value(musb->config->gpio_vrsel,
> > > >> val) +#else
> > > >> +#define vrsel_set_value(val) do { } while (0)
> > > >> +#endif
> > > >
> > > > this obfuscation is unnacceptable and it won't even work if you want to
> > > > have support for both blackfins on the same kernel.
> > >
> > > there are no plans to do that, and a lot more fundamental stuff would
> > > have to be sorted out before that could even be considered
> >
> > Still not taking it. It's just not right. Use a flag at least and don't
> > hide common frameworks under a macro.
>
> doesn't that add pointless runtime bloat for checking a flag that's always
> set/cleared for the targets ? for the structure in question, there is no gpio
> field for BF60x, so it has to be protected by an ifdef else we get a compile
> error.
just let the gpio to always be defined and drop the ifdef.
--
balbi
Attachment:
signature.asc
Description: Digital signature
