Skip to contentSkip to conversation

mm/mempolicy: possible double-unlock in migrate_to_node() on 6.6

BUG [email protected] 3 participants 4 messages Mar 3

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

Hi Andrew,

I found that a fix is required for this submission in branch 6.6.
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a13b2b9b0b0b04612c7d81e3b3dfb485c5f7abc3

Regarding the difference between branches 6.6 and 6.12, in branch 6.6, mmap_read_lock() and mmap_read_unlock() are called outside the migrate_to_node() function. Therefore, mmap_read_unlock(mm) should not be invoked inside migrate_to_node(); otherwise, it would result in a double unlock of the mm.

mm/mempolicy.c +0 -0
index 54f1b78d1b2c0f..94c74c594d102a 100644
@@ -1071,6 +1071,10 @@ static long migrate_to_node(struct mm_struct *mm, int source, int dest,

VM_BUG_ON(!(flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))); vma = find_vma(mm, 0);
+ if (unlikely(!vma)) {
+ mmap_read_unlock(mm); // this line should be deleted
+ return 0;
+ }
/*
* This does not migrate the range, but isolates all pages that

Best Regards,
Li Ying , Light Storage Group. Senior Embedded Software Engineer
Tel: -; Fax: -
Mobile: +86-15651963981
[email protected]
www.sungrowpower.com
SUNGROW POWER SUPPLY Co., Ltd.
Add: No.1699 Xiyou Rd., New & High Technology Industrial Development Zone, Hefei, P. R. China., 230088

#1 lore raw
Quoted 4 lines · On Tue, Mar 03, 2026 at 10:52:51AM +0800, [email protected] wrote:

Hi Andrew,

I found that a fix is required for this submission in branch 6.6.
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a13b2b9b0b0b04612c7d81e3b3dfb485c5f7abc3

Agreed, although your email should more properly have been directed to David and possibly Greg. Are you going to submit a patch to fix this, or do you prefer that somebody else does it?

Quoted 26 lines

Regarding the difference between branches 6.6 and 6.12, in branch 6.6, mmap_read_lock() and mmap_read_unlock() are called outside the migrate_to_node() function. Therefore, mmap_read_unlock(mm) should not be invoked inside migrate_to_node(); otherwise, it would result in a double unlock of the mm.

mm/mempolicy.c +4 -0
index 54f1b78d1b2c0f..94c74c594d102a 100644
@@ -1071,6 +1071,10 @@ static long migrate_to_node(struct mm_struct *mm, int source, int dest,
 VM_BUG_ON(!(flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)));
 vma = find_vma(mm, 0);
+ if (unlikely(!vma)) {
+ mmap_read_unlock(mm); // this line should be deleted
+ return 0;
+ }
 /*
 * This does not migrate the range, but isolates all pages that
 
 

Best Regards,
Li Ying , Light Storage Group. Senior Embedded Software Engineer
Tel: -; Fax: -
Mobile: +86-15651963981
[email protected]
www.sungrowpower.com
SUNGROW POWER SUPPLY Co., Ltd.
Add: No.1699 Xiyou Rd., New & High Technology Industrial Development Zone, Hefei, P. R. China., 230088

#2 lore raw
Quoted 10 lines · Matthew Wilcox
Quoted 4 lines · On Tue, Mar 03, 2026 at 10:52:51AM +0800, [email protected] wrote:

Hi Andrew,

I found that a fix is required for this submission in branch 6.6.
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a13b2b9b0b0b04612c7d81e3b3dfb485c5f7abc3

Agreed, although your email should more properly have been directed to David and possibly Greg. Are you going to submit a patch to fix this, or do you prefer that somebody else does it?

Quoted 13 lines

mm/mempolicy.c static long migrate_to_node(struct mm_struct *mm, int source, int dest, · line 1076

Quoted 13 lines

Regarding the difference between branches 6.6 and 6.12, in branch 6.6, mmap_read_lock() and mmap_read_unlock() are called outside the migrate_to_node() function. Therefore, mmap_read_unlock(mm) should not be invoked inside migrate_to_node(); otherwise, it would result in a double unlock of the mm.

mm/mempolicy.c +4 -0
index 54f1b78d1b2c0f..94c74c594d102a 100644
@@ -1071,6 +1071,10 @@ static long migrate_to_node(struct mm_struct *mm, int source, int dest,
 VM_BUG_ON(!(flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)));
 vma = find_vma(mm, 0);
+ if (unlikely(!vma)) {
+ mmap_read_unlock(mm); // this line should be deleted
+ return 0;
+ }

Right, that unlock must go. Let me know if I should prepare a fix.

Signature · 3 lines
Cheers,

David
#3 lore raw

Thanks for your confirm, please help to process it.

Best Regards,
Li Ying , Light Storage Group. Senior Embedded Software Engineer
Tel: -; Fax: -
Mobile: +86-15651963981
[email protected]
www.sungrowpower.com
SUNGROW POWER SUPPLY Co., Ltd.
Add: No.1699 Xiyou Rd., New & High Technology Industrial Development Zone, Hefei, P. R. China., 230088

From: David Hildenbrand (Arm)
Date: 2026-03-03 17:04
To: Matthew Wilcox; [email protected]
CC: akpm; linux-mm; linux-kernel; victor7q; Greg Kroah-Hartman
Subject: Re: [BUG] mm/mempolicy: possible double-unlock in migrate_to_node() on 6.6

Quoted 10 lines · Matthew Wilcox
Quoted 4 lines · On Tue, Mar 03, 2026 at 10:52:51AM +0800, [email protected] wrote:

Hi Andrew,

I found that a fix is required for this submission in branch 6.6.
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=a13b2b9b0b0b04612c7d81e3b3dfb485c5f7abc3

Agreed, although your email should more properly have been directed to David and possibly Greg. Are you going to submit a patch to fix this, or do you prefer that somebody else does it?

Quoted 13 lines

mm/mempolicy.c static long migrate_to_node(struct mm_struct *mm, int source, int dest, · line 1076

Quoted 13 lines

Regarding the difference between branches 6.6 and 6.12, in branch 6.6, mmap_read_lock() and mmap_read_unlock() are called outside the migrate_to_node() function. Therefore, mmap_read_unlock(mm) should not be invoked inside migrate_to_node(); otherwise, it would result in a double unlock of the mm.

mm/mempolicy.c +4 -0
index 54f1b78d1b2c0f..94c74c594d102a 100644
@@ -1071,6 +1071,10 @@ static long migrate_to_node(struct mm_struct *mm, int source, int dest,
 VM_BUG_ON(!(flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)));
 vma = find_vma(mm, 0);
+ if (unlikely(!vma)) {
+ mmap_read_unlock(mm); // this line should be deleted
+ return 0;
+ }

Right, that unlock must go. Let me know if I should prepare a fix.

Signature · 3 lines
Cheers,
 
David