From: Anton Blanchard <anton@samba.org>

Paul wrote a patch to use some of the rmap infrastructure to flush TLB
entries on ppc64.  When testing it we found a problem in vmalloc where it
sets up the pte -> address mapping incorrectly.  We clear the top bits of
the address but then forget to pass in the full address to
pte_alloc_kernel.  The end result is the address in page->index is
truncated.

I fixed it in a similar way to how zeromap_pmd_range etc does it.  I'm
guessing no one uses the rmap hooks on vmalloc pages yet, so havent seen
this problem.



---

 mm/vmalloc.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff -puN mm/vmalloc.c~vmalloc-address-offset-fix mm/vmalloc.c
--- 25/mm/vmalloc.c~vmalloc-address-offset-fix	2004-01-20 20:34:28.000000000 -0800
+++ 25-akpm/mm/vmalloc.c	2004-01-20 20:34:28.000000000 -0800
@@ -114,15 +114,16 @@ static int map_area_pmd(pmd_t *pmd, unsi
 			       unsigned long size, pgprot_t prot,
 			       struct page ***pages)
 {
-	unsigned long end;
+	unsigned long base, end;
 
+	base = address & PGDIR_MASK;
 	address &= ~PGDIR_MASK;
 	end = address + size;
 	if (end > PGDIR_SIZE)
 		end = PGDIR_SIZE;
 
 	do {
-		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, address);
+		pte_t * pte = pte_alloc_kernel(&init_mm, pmd, base + address);
 		if (!pte)
 			return -ENOMEM;
 		if (map_area_pte(pte, address, end - address, prot, pages))

_