Re: [RFC - V2] - Mapping ACPI tables as CACHED

[Jack Steiner <steiner@sgi.com>]

On Thu, Aug 26, 2010 at 11:08:02AM -0700, H. Peter Anvin wrote:
> On 08/26/2010 10:17 AM, Jack Steiner wrote:
> > Map ACPI tables as WB on x86_64. No substantive changes to IA64.
> > 
> > 
> > Signed-off-by: Jack Steiner <steiner@sgi.com>
> > 
> > ---
> > V2 - Change the patch to unconditionally map ACPI tables as WB on x86_64.
> >      I'm still some paranoid about this because of the potential imapct on
> >      some platforms with weird BIOSs. However, note that on EFI-enabled
> >      systems (like UV), the ACPI tables are already mapped as WB memory.
> >      This is done in the EFI function efi_enter_virtual_mode().
> >      The ACPI code in acpi_os_map_memory() is not currently aware of the
> >      EFI mapping & currently maps the memory as UC. This seems like a bug.
> > 
> 
> I really think that it's all a consequence of the fact that ioremap()
> became implicitly UC when we added PAT support (being the more
> conservative choice.)
> 
> Len, want to take this one or should I?
> 
> 	-hpa

What is the upstream status of this patch? See the LKML discussion at:

	http://marc.info/?l=linux-kernel&m=128206079905826&w=2
	http://marc.info/?l=linux-acpi&m=128284304032481&w=2

We would like to get this patch into the distros but need upstream acceptance first.

This patch substantially reduces the time to run ACPIDUMP on a large system:
 	527 seconds without the patch
 	  8 seconds with the patch

Is there something I should be doing? I can resend the patch if needed.


--- jack