On Fri, Jul 27, 2007 at 10:35:46AM -0400, Jeff Garzik wrote:

Also, general comment:

To be, or not to be:  CONFIG_PCI and CONFIG_EISA

Ah yes.  A can of worms I haven't got into yet.  I do intend to address
this, but it requires summoning up the will to build all kinds of crazy
(ie non-PCI) configs.

I don't have an opinion either way.  On one hand, the present setup is 
clean.  On the other hand, you are likely to run into non-ISA platforms 
that might not be happy with building non-PCI code.  It's not the 
eisa_register_driver() that is the problem, but other ISA-specific 
functions you wind up using in the process.

My hope is that once the driver is a bit cleaner, it'll be easier to
delinearate which bits are ISA, which EISA, which PCI, and which are
common to all three.

I think there's a lot of code which is currently common to all three
that doesn't need to be.

Plus, Andi likes to make a big fuss about building ISA code on x86-64 :)

Fortunately, this driver is currently limited to X86_32 by the Kconfig.
Again, I will address this ... but perhaps not in this patch series.

-- 
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."