Re: [patch 0/7] [rfc] VM_MIXEDMAP, pte_special, xip work

[Martin Schwidefsky <schwidefsky@de.ibm.com>]

On Wed, 2008-03-12 at 16:33 +1100, Nick Piggin wrote:
> s390 is slightly different because it doesn't use a standard memory model
> but something more dynamic. It doesn't quite do the right thing here, so
> it uses pte_special. It could possibly tighten up pfn_valid, however I
> think there are various reasons why they don't want to (one is that they
> need to take a global lock in order to search their list of extents;
> which will suck for VM_MIXEDMAP performance).

Indeed. Under z/VM we have mappable memory segments named DCSS that can
have different types. Dependend on the type we either want to do
refcounting (exclusive write for memory hotplug) or not (read-only
shared for memory sharing). pfn_valid() would have to scan the list of
DCCSes to find out which type the segment has. That is way too
expensive.

-- 
blue skies,
  Martin.

"Reality continues to ruin my life." - Calvin.


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>