diff --git a/include/asm-ppc/pgtable.h b/include/asm-ppc/pgtable.h
index 19dfb7abaa2189e6d11a56da91ee8c8c9b3ccca3..2e88cd9feffe953ae2f447223f48237b80de869a 100644
--- a/include/asm-ppc/pgtable.h
+++ b/include/asm-ppc/pgtable.h
@@ -526,10 +526,10 @@ extern void add_hash_page(unsigned context, unsigned long va,
  * Atomic PTE updates.
  *
  * pte_update clears and sets bit atomically, and returns
- * the old pte value.
- * The ((unsigned long)(p+1) - 4) hack is to get to the least-significant
- * 32 bits of the PTE regardless of whether PTEs are 32 or 64 bits.
+ * the old pte value.  In the 64-bit PTE case we lock around the
+ * low PTE word since we expect ALL flag bits to be there
  */
+#ifndef CONFIG_PTE_64BIT
 static inline unsigned long pte_update(pte_t *p, unsigned long clr,
 				       unsigned long set)
 {
@@ -543,10 +543,31 @@ static inline unsigned long pte_update(pte_t *p, unsigned long clr,
 "	stwcx.	%1,0,%3\n\
 	bne-	1b"
 	: "=&r" (old), "=&r" (tmp), "=m" (*p)
-	: "r" ((unsigned long)(p+1) - 4), "r" (clr), "r" (set), "m" (*p)
+	: "r" (p), "r" (clr), "r" (set), "m" (*p)
 	: "cc" );
 	return old;
 }
+#else
+static inline unsigned long long pte_update(pte_t *p, unsigned long clr,
+				       unsigned long set)
+{
+	unsigned long long old;
+	unsigned long tmp;
+
+	__asm__ __volatile__("\
+1:	lwarx	%L0,0,%4\n\
+	lwzx	%0,0,%3\n\
+	andc	%1,%L0,%5\n\
+	or	%1,%1,%6\n"
+	PPC405_ERR77(0,%3)
+"	stwcx.	%1,0,%4\n\
+	bne-	1b"
+	: "=&r" (old), "=&r" (tmp), "=m" (*p)
+	: "r" (p), "r" ((unsigned long)(p) + 4), "r" (clr), "r" (set), "m" (*p)
+	: "cc" );
+	return old;
+}
+#endif
 
 /*
  * set_pte stores a linux PTE into the linux page table.