FAILED: patch "[PATCH] mm/hugetlb: separate path for hwpoison entry in" failed to apply to 5.18-stable tree

PATCH [email protected] 3 participants 4 messages Jul 30, 2022 – Aug 1, 2022

Subsystems (3)

Stable branch Memory management HugeTLB

lore

Share & tools

lore

[email protected]

author maintainer

2022-07-30 15:35 UTC #0 lore raw

The patch below does not apply to the 5.18-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

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

From c2cb0dcce9dd8b748b6ca8bb8d4a389f2e232307 Mon Sep 17 00:00:00 2001
From: Naoya Horiguchi <[email protected]>
Date: Mon, 4 Jul 2022 10:33:05 +0900
Subject: [PATCH] mm/hugetlb: separate path for hwpoison entry in
copy_hugetlb_page_range()

Originally copy_hugetlb_page_range() handles migration entries and hwpoisoned entries in similar manner. But recently the related code path has more code for migration entries, and when is_writable_migration_entry() was converted to !is_readable_migration_entry(), hwpoison entries on source processes got to be unexpectedly updated (which is legitimate for migration entries, but not for hwpoison entries). This results in unexpected serious issues like kernel panic when forking processes with hwpoison entries in pmd.

Separate the if branch into one for hwpoison entries and one for migration entries.

Link https://lkml.kernel.org/r/[email protected] open here

Fixes 6c287605fd56 ("mm: remember exclusively mapped anonymous pages with PG_anon_exclusive")

Signed-off-by Naoya Horiguchi <[email protected]>

Reviewed-by Miaohe Lin <[email protected]>

Reviewed-by Mike Kravetz <[email protected]>

Reviewed-by Muchun Song <[email protected]>

Cc <[email protected]> [5.18]

Cc David Hildenbrand <[email protected]>

Cc Liu Shixin <[email protected]>

Cc Oscar Salvador <[email protected]>

Cc Yang Shi <[email protected]>

Signed-off-by Andrew Morton <[email protected]>

mm/hugetlb.c +7 -2

index a57e1be41401..baf7f6b19ce6 100644
@@ -4788,8 +4788,13 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
 			 * sharing with another vma.
 			 */
 			;
-		} else if (unlikely(is_hugetlb_entry_migration(entry) ||
-				    is_hugetlb_entry_hwpoisoned(entry))) {
+		} else if (unlikely(is_hugetlb_entry_hwpoisoned(entry))) {
+			bool uffd_wp = huge_pte_uffd_wp(entry);
+
+			if (!userfaultfd_wp(dst_vma) && uffd_wp)
+				entry = huge_pte_clear_uffd_wp(entry);
+			set_huge_pte_at(dst, addr, dst_pte, entry);
+		} else if (unlikely(is_hugetlb_entry_migration(entry))) {
 			swp_entry_t swp_entry = pte_to_swp_entry(entry);
 			bool uffd_wp = huge_pte_uffd_wp(entry);

Naoya Horiguchi

Reviewed-by Miaohe Lin Reviewed-by Mike Kravetz Reviewed-by Muchun Song

2022-08-01 06:29 UTC #1 lore raw

Quoted 7 lines · On Sat, Jul 30, 2022 at 05:35:37PM +0200, [email protected] wrote:

The patch below does not apply to the 5.18-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,

Hello,

I updated the patch for 5.18-stable, could you apply this?

Thanks,
Naoya Horiguchi
---
From: Naoya Horiguchi <[email protected]>
Date: Mon, 1 Aug 2022 15:09:17 +0900
Subject: [PATCH] mm/hugetlb: separate path for hwpoison entry in
copy_hugetlb_page_range()

commit c2cb0dcce9dd8b748b6ca8bb8d4a389f2e232307 upstream.

Separate the if branch into one for hwpoison entries and one for migration entries.

[naoya.horiguchi: Resolved conflict to apply for 5.18-stable.]

Link https://lkml.kernel.org/r/[email protected] open here

Fixes 6c287605fd56 ("mm: remember exclusively mapped anonymous pages with PG_anon_exclusive")

Signed-off-by Naoya Horiguchi <[email protected]>

Reviewed-by Miaohe Lin <[email protected]>

Reviewed-by Mike Kravetz <[email protected]>

Reviewed-by Muchun Song <[email protected]>

Cc <[email protected]> [5.18]

Cc David Hildenbrand <[email protected]>

Cc Liu Shixin <[email protected]>

Cc Oscar Salvador <[email protected]>

Cc Yang Shi <[email protected]>

Signed-off-by Andrew Morton <[email protected]>

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

mm/hugetlb.c +3 -2

index 410bbb0aee32..0aa670addd8c 100644
@@ -4764,8 +4764,9 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
 			 * sharing with another vma.
 			 */
 			;
-		} else if (unlikely(is_hugetlb_entry_migration(entry) ||
-				    is_hugetlb_entry_hwpoisoned(entry))) {
+		} else if (unlikely(is_hugetlb_entry_hwpoisoned(entry))) {
+			set_huge_pte_at(dst, addr, dst_pte, entry);
+		} else if (unlikely(is_hugetlb_entry_migration(entry))) {
 			swp_entry_t swp_entry = pte_to_swp_entry(entry);
 
 			if (is_writable_migration_entry(swp_entry) && cow) {

Signature · 1 line

2.26.1.8815.g6a475b71f8

Quoted 56 lines

greg k-h
------------------ original commit in Linus's tree ------------------
From c2cb0dcce9dd8b748b6ca8bb8d4a389f2e232307 Mon Sep 17 00:00:00 2001
From: Naoya Horiguchi <[email protected]>
Date: Mon, 4 Jul 2022 10:33:05 +0900
Subject: [PATCH] mm/hugetlb: separate path for hwpoison entry in
copy_hugetlb_page_range()
Originally copy_hugetlb_page_range() handles migration entries and hwpoisoned entries in similar manner. But recently the related code path has more code for migration entries, and when is_writable_migration_entry() was converted to !is_readable_migration_entry(), hwpoison entries on source processes got to be unexpectedly updated (which is legitimate for migration entries, but not for hwpoison entries). This results in unexpected serious issues like kernel panic when forking processes with hwpoison entries in pmd.
Separate the if branch into one for hwpoison entries and one for migration entries.
Link https://lkml.kernel.org/r/[email protected] open here
Fixes 6c287605fd56 ("mm: remember exclusively mapped anonymous pages with PG_anon_exclusive")
Signed-off-by Naoya Horiguchi <[email protected]>
Reviewed-by Miaohe Lin <[email protected]>
Reviewed-by Mike Kravetz <[email protected]>
Reviewed-by Muchun Song <[email protected]>
Cc <[email protected]> [5.18]
Cc David Hildenbrand <[email protected]>
Cc Liu Shixin <[email protected]>
Cc Oscar Salvador <[email protected]>
Cc Yang Shi <[email protected]>
Signed-off-by Andrew Morton <[email protected]>
mm/hugetlb.c +7 -2
index a57e1be41401..baf7f6b19ce6 100644
@@ -4788,8 +4788,13 @@ int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src,
 			 * sharing with another vma.
 			 */
 			;
-		} else if (unlikely(is_hugetlb_entry_migration(entry) ||
-				    is_hugetlb_entry_hwpoisoned(entry))) {
+		} else if (unlikely(is_hugetlb_entry_hwpoisoned(entry))) {
+			bool uffd_wp = huge_pte_uffd_wp(entry);
+
+			if (!userfaultfd_wp(dst_vma) && uffd_wp)
+				entry = huge_pte_clear_uffd_wp(entry);
+			set_huge_pte_at(dst, addr, dst_pte, entry);
+		} else if (unlikely(is_hugetlb_entry_migration(entry))) {
 			swp_entry_t swp_entry = pte_to_swp_entry(entry);
 			bool uffd_wp = huge_pte_uffd_wp(entry);
 

Greg KH

author maintainer

2022-08-01 08:19 UTC #2 lore raw

Quoted 7 lines · Naoya Horiguchi

Quoted 6 lines · On Sat, Jul 30, 2022 at 05:35:37PM +0200, [email protected] wrote:
The patch below does not apply to the 5.18-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]>.

Quoted 5 lines

Quoted 1 line
thanks,
Hello,
I updated the patch for 5.18-stable, could you apply this?

I can, but are you sure you need/want this for 5.18-stable?

I ask because of this line in the changelog text:

Quoted 1 line

Fixes 6c287605fd56 ("mm: remember exclusively mapped anonymous pages with PG_anon_exclusive")

Which only showed up in 5.19-rc1 and is not backported anywhere.

So why did you have:

Quoted 1 line

Cc <[email protected]> [5.18]

In the changelog if the commit this fixes is not in 5.18?

confused,

greg k-h

HORIGUCHI NAOYA(堀口直也)

2022-08-01 08:32 UTC #3 lore raw

Quoted 14 lines · Greg KH

Quoted 7 lines · Naoya Horiguchi
Quoted 6 lines · On Sat, Jul 30, 2022 at 05:35:37PM +0200, [email protected] wrote:
The patch below does not apply to the 5.18-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]>.
Quoted 5 lines
Quoted 1 line
thanks,
Hello,
I updated the patch for 5.18-stable, could you apply this?

Quoted 7 lines

I can, but are you sure you need/want this for 5.18-stable?
I ask because of this line in the changelog text:
Quoted 1 line
Fixes 6c287605fd56 ("mm: remember exclusively mapped anonymous pages with PG_anon_exclusive")
Which only showed up in 5.19-rc1 and is not backported anywhere.

OK, So we don't have to backport this patch.

Quoted 6 lines

So why did you have:
Quoted 1 line
Cc <[email protected]> [5.18]
In the changelog if the commit this fixes is not in 5.18?

I wrongly judged that 6c287605fd56 was before 5.19, sorry about that. Please ignore my backport request.

Thanks,
Naoya Horiguchi

Quoted 4 lines

confused,
greg k-h