Skip to contentSkip to conversation

mm, hugetlbfs: Fix new warning in fault-time huge page allocation

PATCH New Dave Hansen 1 participant 1 message Oct 21, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm, hugetlbfs: Fix new warning in fault-time huge page allocation

From: Dave Hansen <[email protected]>

Kirill reported that he hit:

Quoted 4 lines

Review

Quoted 2 lines
(quoted diff) +2 -0
+	if (vma || addr) {
+		WARN_ON_ONCE(!addr || addr == -1);

Trinity triggered the WARN for me:

This was just a dumb mistake. I put the WARN_ON() in and planned to have addr=0 mean "use nid". But, I realized pretty quickly that addr=0 _is_ a valid place to fault. So I made it addr=-1 in __alloc_buddy_huge_page_no_mpol(), but I did not fix up the WARN_ON().

So hitting the warning in this case was harmless. But, fix up the warning condition.

Cc Kirill A. Shutemov <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Naoya Horiguchi <[email protected]>
Cc Mike Kravetz <[email protected]>
Cc Hillf Danton <[email protected]>
Cc David Rientjes <[email protected]>

---

 b/mm/hugetlb.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
mm/hugetlb.c renamed from mm/hugetlb.c~hugetlbfs-fix-warn +8 -2
diff -puN mm/hugetlb.c~hugetlbfs-fix-warn mm/hugetlb.c
@@ -1520,8 +1520,14 @@ static struct page *__alloc_buddy_huge_p
 	if (hstate_is_gigantic(h))
 		return NULL;
 
-	if (vma || addr) {
-		VM_WARN_ON_ONCE(!addr || addr == -1);
+	/*
+	 * Make sure that anyone specifying 'nid' is not also
+	 * specifying a VMA.  This makes sure the caller is
+	 * picking _one_ of the modes with which we can call this
+	 * function, not both.
+	 */
+	if (vma || (addr != -1)) {
+		VM_WARN_ON_ONCE(addr == -1);
 		VM_WARN_ON_ONCE(nid != NUMA_NO_NODE);
 	}
 	/*

_