Re: [2.6 patch] PCMCIA mustn't select HAVE_IDE

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Tue, Apr 15, 2008 at 10:28:22PM +0100, Russell King wrote:
> On Wed, Apr 16, 2008 at 12:23:26AM +0300, Adrian Bunk wrote:
> > On Tue, Apr 15, 2008 at 10:15:34PM +0200, Bartlomiej Zolnierkiewicz wrote:
> > > On Monday 14 April 2008, Sam Ravnborg wrote:
> > > > On Mon, Apr 14, 2008 at 05:16:59PM +0300, Adrian Bunk wrote:
> > > > > It's plain wrong for PCMCIA to select HAVE_IDE that implies e.g. the 
> > > > > availability of an asm/ide.h
> > > > > 
> > > > > Signed-off-by: Adrian Bunk <bunk@kernel.org>
> > > > > 
> > > > > ---
> > > > > 9cdb66112488bc0c6e1d528444d3ba30d5b0487f diff --git a/drivers/pcmcia/Kconfig b/drivers/pcmcia/Kconfig
> > > > > index 8b22281..519b4ff 100644
> > > > > --- a/drivers/pcmcia/Kconfig
> > > > > +++ b/drivers/pcmcia/Kconfig
> > > > > @@ -38,7 +38,6 @@ config PCMCIA_DEBUG
> > > > >  config PCMCIA
> > > > >  	tristate "16-bit PCMCIA support"
> > > > >  	select CRC32
> > > > > -	select HAVE_IDE
> > > > >  	default y
> > > > 
> > > > I did this when introducing HAVE_IDE.
> > > > But I do not recall why and the rationale for removing it
> > > > seems fine to me.
> > > 
> > > IIRC it was needed for some arm platforms which don't select HAVE_IDE
> > > explicetely but I don't know if this is still the case, pinging Russell.
> > 
> > It's definitely bogus since it can cause compile breakage on 
> > architectures like avr32.
> > 
> > And whatever it should have fixed should be fixed properly.
> 
> I'd suggest backing out the entire change which introduced HAVE_IDE then -
> rather than doing it piecemeal and bringing up questions about it which
> are unanswerable (which is the case of Bart's question of me.)

Looking back at the original change:

-if PCMCIA || ARCH_CLPS7500 || ARCH_IOP32X || ARCH_IOP33X || ARCH_IXP4XX \
-       || ARCH_L7200 || ARCH_LH7A40X || ARCH_PXA || ARCH_RPC \
-       || ARCH_S3C2410 || ARCH_SA1100 || ARCH_SHARK || FOOTBRIDGE \
-       || ARCH_IXP23XX
 source "drivers/ide/Kconfig"
-endif

was what was done to ARM's Kconfig.  So, if we're going to be doing
anything with the 'select HAVE_IDE', it should be:

 config PCMCIA
        tristate "16-bit PCMCIA support"
        select CRC32
-	select HAVE_IDE
+	select HAVE_IDE if ARM

Or, as I've already said, we revert all of the HAVE_IDE change, and ARM
continues avoiding drivers/Kconfig.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of: