[PATCH] x86: adjust vmalloc_sync_all() for Xen (2nd try)

[PATCH] x86: adjust vmalloc_sync_all() for Xen (2nd try)

[Jan Beulich]

Since the fourth PDPT entry cannot be shared under Xen,
vmalloc_sync_all() must iterate over pmd-s rather than pgd-s here.
Luckily, the code isn't used for native PAE (SHARED_KERNEL_PMD is 1)
and the change is benign to non-PAE.

Also do a little more cleanup in that function.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>

---
 arch/x86/mm/fault.c |   14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

--- linux-2.6.27-rc5/arch/x86/mm/fault.c	2008-08-21 14:37:29.000000000 +0200
+++ 2.6.27-rc5-x86-vmalloc-sync-all/arch/x86/mm/fault.c	2008-08-22 15:24:33.000000000 +0200
@@ -915,15 +915,15 @@ LIST_HEAD(pgd_list);
 
 void vmalloc_sync_all(void)
 {
-#ifdef CONFIG_X86_32
-	unsigned long start = VMALLOC_START & PGDIR_MASK;
 	unsigned long address;
 
+#ifdef CONFIG_X86_32
 	if (SHARED_KERNEL_PMD)
 		return;
 
-	BUILD_BUG_ON(TASK_SIZE & ~PGDIR_MASK);
-	for (address = start; address >= TASK_SIZE; address += PGDIR_SIZE) {
+	for (address = VMALLOC_START & PMD_MASK;
+	     address >= TASK_SIZE && address < FIXADDR_TOP;
+	     address += PMD_SIZE) {
 		unsigned long flags;
 		struct page *page;
 
@@ -936,10 +936,8 @@ void vmalloc_sync_all(void)
 		spin_unlock_irqrestore(&pgd_lock, flags);
 	}
 #else /* CONFIG_X86_64 */
-	unsigned long start = VMALLOC_START & PGDIR_MASK;
-	unsigned long address;
-
-	for (address = start; address <= VMALLOC_END; address += PGDIR_SIZE) {
+	for (address = VMALLOC_START & PGDIR_MASK; address <= VMALLOC_END;
+	     address += PGDIR_SIZE) {
 		const pgd_t *pgd_ref = pgd_offset_k(address);
 		unsigned long flags;
 		struct page *page;

Re: [PATCH] x86: adjust vmalloc_sync_all() for Xen (2nd try)

[Ingo Molnar]

* Jan Beulich <jbeulich@novell.com> wrote:

> Since the fourth PDPT entry cannot be shared under Xen,
> vmalloc_sync_all() must iterate over pmd-s rather than pgd-s here.
> Luckily, the code isn't used for native PAE (SHARED_KERNEL_PMD is 1)
> and the change is benign to non-PAE.
> 
> Also do a little more cleanup in that function.
> 
> Signed-off-by: Jan Beulich <jbeulich@novell.com>
> Cc: Jeremy Fitzhardinge <jeremy@goop.org>

applied to tip/x86/xen, thanks Jan.

	Ingo