[PATCH] ppc64: Fix possible race with set_pte on a present PTE

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

(RESENT)

There is a subtle race which can cause set_pte to be called on ppc64 on
a PTE that is already present (that normally doesn't happen for us) and
which itself, in the proper race condition, can trigger a duplicate hash
entry to be added to the hash table (very bad).

This fixes it by making sure we trigger the actual flush of the batch
whenever set_pte is called on a present PTE, before putting the new PTE
in.

===== include/asm-ppc64/pgtable.h 1.32 vs edited =====
--- 1.32/include/asm-ppc64/pgtable.h	Fri Apr  9 03:30:57 2004
+++ edited/include/asm-ppc64/pgtable.h	Thu May 20 15:47:40 2004
@@ -396,8 +396,10 @@
  */
 static inline void set_pte(pte_t *ptep, pte_t pte)
 {
-	if (pte_present(*ptep))
+	if (pte_present(*ptep)) {
 		pte_clear(ptep);
+		flush_tlb_pending();
+	}
 	*ptep = __pte(pte_val(pte)) & ~_PAGE_HPTEFLAGS;
 }