2006-08-07 05:25:08

by Randy Dunlap

[permalink] [raw]
Subject: Re: [patch] bootmem: use MAX_DMA_ADDRESS instead of LOW32LIMIT

On Sat, 29 Jul 2006 05:55:23 +0200 Ingo Molnar wrote:

>
> * Heiko Carstens <[email protected]> wrote:
>
> > Hm... how about this one then:
> >
> > From: Heiko Carstens <[email protected]>
> >
> > Introduce ARCH_LOW_ADDRESS_LIMIT which can be set per architecture to
> > override the 4GB default limit used by the bootmem allocater within
> > __alloc_bootmem_low() and __alloc_bootmem_low_node().
> > E.g. s390 needs a 2GB limit instead of 4GB.
> >
> > Cc: Ingo Molnar <[email protected]>
>
> Acked-by: Ingo Molnar <[email protected]>
>
> (although you might get some flak about using an ARCH* define. I'm not
> sure what the current upstream policy is - using an #ifndef default
> value is the most compact hence sanest thing to do, still it's sometimes
> being frowned upon in favor of sprinkling the default value into every
> architecture's processor.h. Putting the value into a Kconfig and
> combining it with #ifndef might be better.)

(sorry for the delay, too much travel/conferences)

I agree with your ordering. Linus wrote about the current
ARCH_HAS* (and HAVE_ARCH* I suppose):
"WE SHOULD GET RID OF ARCH_HAS_XYZZY. It's a disease."

I have patches for some of these that I will post soon (prob.
Monday), converting several ARCH_HAS* to CONFIG_ namespace.

---
~Randy