Re: dm9000: Remove the 2 resources probe scheme.

[Ben Dooks <ben-linux@fluff.org>]

On Mon, Jun 16, 2008 at 10:46:48AM +0200, Laurent Pinchart wrote:
> On Sunday 15 June 2008 23:49, Ben Dooks wrote:
> > On Fri, Jun 13, 2008 at 10:29:45PM +0800, Bryan Wu wrote:
> > > On Fri, Jun 13, 2008 at 8:58 PM, Laurent Pinchart
> > > <laurentp@cse-semaphore.com> wrote:
> > > > The dm9000 driver accepts either 2 or 3 resources to describe the
> > > > platform devices. The 2 resources case abuses the ioresource mechanism
> > > > by passing ioremap()ed memory through the platform device resources.
> > > > This patch removes that case and converts boards that were using it to
> > > > the 3 resources scheme. 
> > > >
> > > > Signed-off-by: Laurent Pinchart <laurentp@cse-semaphore.com>
> > > > ---
> > > >  arch/blackfin/mach-bf527/boards/ezkit.c         |    7 ++-
> > > >  arch/blackfin/mach-bf533/boards/H8606.c         |    7 ++-
> > > >  arch/blackfin/mach-bf537/boards/generic_board.c |    7 ++-
> > > >  arch/blackfin/mach-bf537/boards/stamp.c         |    7 ++-
> > > 
> > > IMO, this patch can be split into two parts, one for Blackfin and the
> > > other for dm9000 driver.
> > 
> > I think I agree with that, although I would also like to make sure the
> > blackfin part of this is tested. I am also trying to get rid of the
> > defines for blackfin specific compile, as they shouldn't be in the
> > driver!
> > 
> > I've queued the dm9000 specific bits into my own little dm9000 tree
> > for the next kernel release.
> 
> Thanks. Should I resubmit a blackfin-specific patch (I obviously won't be able 
> to test it, as I have no blackfin hardware) or will you take care of it ?

I'll sort out the rest of the split and removing the other blackfin
cruft from the driver.

-- 
Ben (ben@fluff.org, http://www.fluff.org/)

  'a smiley only costs 4 bytes'