Skip to contentSkip to conversation

FAILED: patch "[PATCH] mm/hugetlb: add migration/hwpoisoned entry check in" failed to apply to 3.10-stable tree

DISCUSSION [email protected] 1 participant 1 message Apr 26, 2015 2 messages not in the archive Stable branch

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

[email protected]

author maintainer
#2 lore raw
Quoted 77 lines · Naoya Horiguchi
Quoted 10 lines · On Wed, Mar 11, 2015 at 02:54:04PM +0100, [email protected] wrote:

The patch below does not apply to the 3.10-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

Quoted 7 lines

------------------ original commit in Linus's tree ------------------

From a8bda28d87c38c6aa93de28ba5d30cc18e865a11 Mon Sep 17 00:00:00 2001
From: Naoya Horiguchi <[email protected]>
Date: Wed, 11 Feb 2015 15:25:28 -0800
Subject: [PATCH] mm/hugetlb: add migration/hwpoisoned entry check in
hugetlb_change_protection

The conflict is trivial enough to resolve, so please apply the below one.

Thanks,
Naoya Horiguchi
---

Quoted 1 line

From 5babc2f225937be13d7c040d6d850128bd4f7ebf Mon Sep 17 00:00:00 2001

From: Naoya Horiguchi <[email protected]>
Date: Mon, 16 Mar 2015 16:04:05 +0900
Subject: [PATCH] mm/hugetlb: add migration/hwpoisoned entry check in
hugetlb_change_protection

There is a race condition between hugepage migration and change_protection(), where hugetlb_change_protection() doesn't care about migration entries and wrongly overwrites them. That causes unexpected results like kernel crash. HWPoison entries also can cause the same problem.

This patch adds is_hugetlb_entry_(migration|hwpoisoned) check in this function to do proper actions.

[[email protected]: resolve conflict to apply to v3.10.71]
Fixes 290408d4a2 ("hugetlb: hugepage migration core")
Cc Hugh Dickins <[email protected]>
Cc James Hogan <[email protected]>
Cc David Rientjes <[email protected]>
Cc Mel Gorman <[email protected]>
Cc Johannes Weiner <[email protected]>
Cc Michal Hocko <[email protected]>
Cc Rik van Riel <[email protected]>
Cc Andrea Arcangeli <[email protected]>
Cc Luiz Capitulino <[email protected]>
Cc Nishanth Aravamudan <[email protected]>
Cc Lee Schermerhorn <[email protected]>
Cc Steve Capper <[email protected]>
Cc <[email protected]> [2.6.36+]
---
 mm/hugetlb.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)
mm/hugetlb.c +7 -1
index 7de4f67c81fe..e25e32f5e0e1 100644
@@ -3116,7 +3116,26 @@ unsigned long hugetlb_change_protection(struct vm_area_struct *vma,
 			pages++;
 			continue;
 		}
-		if (!huge_pte_none(huge_ptep_get(ptep))) {
+		pte = huge_ptep_get(ptep);
+		if (unlikely(is_hugetlb_entry_hwpoisoned(pte))) {
+			spin_unlock(ptl);
+			continue;
+		}
+		if (unlikely(is_hugetlb_entry_migration(pte))) {
+			swp_entry_t entry = pte_to_swp_entry(pte);

mm/hugetlb.c unsigned long hugetlb_change_protection(struct vm_area_struct *vma, · line 3135

Quoted 10 lines
mm/hugetlb.c +10 -0
+
+			if (is_write_migration_entry(entry)) {
+				pte_t newpte;
+
+				make_migration_entry_read(&entry);
+				newpte = swp_entry_to_pte(entry);
+				set_huge_pte_at(mm, address, ptep, newpte);
+				pages++;
+			}
+			spin_unlock(ptl);

Doesn't compile :(