Re: CVS Update@oss.sgi.com: linux

On Mon, Aug 19, 2002 at 02:35:19PM +0200, Maciej W. Rozycki wrote:

quoted

Modified files:
	arch/mips      : config-shared.in defconfig defconfig-decstation 
	                 defconfig-ip22 defconfig-nino defconfig-osprey 
	                 defconfig-sb1250-swarm defconfig-sead 
	arch/mips64    : defconfig-ip22 defconfig-sb1250-swarm 
	                 defconfig-sead 

Log message:
	Make CONFIG_IDE selectable independant of the bus type.

 Hmm, what's the intent of the change?  IDE, or more properly ATA, was
originally an ISA-only device and is still only available as ISA-style
implementations, AFAIK.  I'd prefer it to be available only if any of
CONFIG_ISA, CONFIG_EISA, CONFIG_PCI (unsure about CONFIG_MCA) is set.

 That said, the place for such decisions seems to be inappropriate
currently.  It'd be much more elegant just to source all relevant drivers,
net, etc. Config.in scripts unconditionally and make global enable/disable
decisions at the top of the relevant script, like e.g. 
drivers/message/i2o/Config.in already does. 

Even before the changes Alan mentioned we had a bunch of machines like the
SWARM which had PCI but it's PIO IDE adapter wasn't living on the PCI bus.
So disabling PCI on that machine would disable IDE also.  Other machines
had IDE on PCMCIA with the PCMCIA bridge not hanging off an (E)ISA or PCI
bridge.  Basically I could have changed that if statement into an
increasingly obscure and braindamagedly complex if statement.

  Ralf