[PATCH] Cleanup/fix virt_to_maddr

[PATCH] Cleanup/fix virt_to_maddr

[Christoph Egger]

[-- Attachment #1: Type: text/plain, Size: 148 bytes --]



Hi!

Where virt_to_maddr() or __pa() is used, paddr_t is mostly expected rather
unsigned long. This may fix random issues in PAE mode.

Christoph

[-- Attachment #2: xen_paddr.diff --]
[-- Type: text/x-diff, Size: 663 bytes --]

diff -r bea505a69722 xen/include/asm-x86/page.h
--- a/xen/include/asm-x86/page.h	Tue Jan 23 15:58:05 2007 +0000
+++ b/xen/include/asm-x86/page.h	Wed Jan 24 09:32:26 2007 +0100
@@ -222,7 +222,7 @@ typedef struct { u64 pfn; } pagetable_t;
 
 /* Convert between Xen-heap virtual addresses and machine addresses. */
 #define PAGE_OFFSET         ((unsigned long)__PAGE_OFFSET)
-#define virt_to_maddr(va)   ((unsigned long)(va)-PAGE_OFFSET)
+#define virt_to_maddr(va)   ((paddr_t)(va)-PAGE_OFFSET)
 #define maddr_to_virt(ma)   ((void *)((unsigned long)(ma)+PAGE_OFFSET))
 /* Shorthand versions of the above functions. */
 #define __pa(x)             (virt_to_maddr(x))

[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: [PATCH] Cleanup/fix virt_to_maddr

[Keir Fraser]

On 24/1/07 09:06, "Christoph Egger" <Christoph.Egger@amd.com> wrote:

> Where virt_to_maddr() or __pa() is used, paddr_t is mostly expected rather
> unsigned long. This may fix random issues in PAE mode.

They're only used on Xen heap virtual addresses whose physical addresses are
always below 64MB. So there should be no issues in returning a ulong.

 -- Keir

Re: [PATCH] Cleanup/fix virt_to_maddr

[Christoph Egger]

On Wednesday 24 January 2007 11:48, Keir Fraser wrote:
> On 24/1/07 09:06, "Christoph Egger" <Christoph.Egger@amd.com> wrote:
> > Where virt_to_maddr() or __pa() is used, paddr_t is mostly expected
> > rather unsigned long. This may fix random issues in PAE mode.
>
> They're only used on Xen heap virtual addresses whose physical addresses
> are always below 64MB. So there should be no issues in returning a ulong.

Oh, I see. So it's absolutely not used for domains (which I assumed)?

Christoph

Re: [PATCH] Cleanup/fix virt_to_maddr

[Keir Fraser]

On 24/1/07 11:01, "Christoph Egger" <Christoph.Egger@amd.com> wrote:

>> They're only used on Xen heap virtual addresses whose physical addresses
>> are always below 64MB. So there should be no issues in returning a ulong.
> 
> Oh, I see. So it's absolutely not used for domains (which I assumed)?

No, all it does is a simple arithmetic transformation (subtracting
PAGE_OFFSET) which is only valid for non-guest pages on x86_32. It can be
used for any page on x86_64, but there we have paddr_t==ulong.

 -- Keir