Re: [PATCH 5/5] SRAT cleanup: make calculations and indenting level more sane

[keith <kmannth@us.ibm.com>]

On Thu, 2005-02-24 at 12:49, James Cleverdon wrote:
> No, I don't think we could rely on that.  Our BIOS did ascending 
> addresses, but I don't recall that being spelled out in the ACPI spec.
> 
> Of course, there's a new ACPI spec out.  Maybe it makes it a 
> requirement.  I'd take a look, but I can't afford the loss of sanity 
> caused by gazing on the dread visage of ACPI 3.0.   ;^)


The SRAT exists outside of the ACPI spec.  It is something made up by
folks in Kirkland.  I just reread the SRAT spec and I don't seen any
mention of requirements for linear order.  Still yet we have yet to find
a box/bios version that breaks this assumption. All I know of is the IBM
summit boxes but maybe there is something else.  

Maybe AMD x86_64 booting into 32 bit have SRATs as well?

Anyways maybe we could add some check to catch new hardware with less
friendly SRAT tables. 

after the  node_has_online_mem(nid) check 

if (node_start_pfn[nid] > node_memory_chunk[j].start_pfn) {
	printk (KERN_WARN "You need to rework the srat.c code\n");
	continue;
}

Keith 


--
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:"aart@kvack.org"> aart@kvack.org </a>