Re: kernel (64bit) 4GB memory support II

[Andi Kleen <ak@muc.de>]

Rudolf Usselmann <rudi@asics.ws> writes:
>
> Just curious, did anybody ever look in to this at all ? I keep
> on downloading new kernels and trying 4GB of memory - still no
> luck.
>
> I did file a bug report but didn't get any notifications at all.
> I don't subscribe to the linux-kernel list so not sure if anything
> ever came up or not.
>
> Is there a way to get this fixed ?

Does the following patch (against a 2.6.12rc3 kernel) fix your problems?

-Andi

[...]

Please use this version instead, previous one was broken.

Don't look up struct page * of physical address in iounmap

it could be in a memory hole not mapped in mem_map

Signed-off-by: Andi Kleen <ak@suse.de>

Index: linux/arch/x86_64/mm/ioremap.c
===================================================================
--- linux.orig/arch/x86_64/mm/ioremap.c
+++ linux/arch/x86_64/mm/ioremap.c
@@ -272,7 +272,7 @@ void iounmap(volatile void __iomem *addr
 	if ((p->flags >> 20) &&
 		p->phys_addr + p->size - 1 < virt_to_phys(high_memory)) {
 		/* p->size includes the guard page, but cpa doesn't like that */
-		change_page_attr(virt_to_page(__va(p->phys_addr)),
+		change_page_attr_addr(__va(p->phys_addr),
 				 p->size >> PAGE_SHIFT,
 				 PAGE_KERNEL);
 		global_flush_tlb();