[PATCH] mm/damon/vaddr: fix missing pte_unmap_unlock in damos_va_migrate_pmd_entry()

[PATCH] mm/damon/vaddr: fix missing pte_unmap_unlock in damos_va_migrate_pmd_entry()

[Akinobu Mita]

If the PTE page table lock is acquired by pte_offset_map_lock(), the lock
must be released via pte_unmap_unlock().

However, in damos_va_migrate_pmd_entry(), if damos_va_filter_out() returns
true, it immediately returns without releasing the lock.

This fixes the issue by not stopping page table traversal when
damos_va_filter_out() returns true and ensuring that the lock is released.

Fixes: 09efc56a3b1c ("mm/damon/vaddr: consistently use only pmd_entry for damos_migrate")
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
---
 mm/damon/vaddr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mm/damon/vaddr.c b/mm/damon/vaddr.c
index 2750c88e7225..23ed738a0bd6 100644
--- a/mm/damon/vaddr.c
+++ b/mm/damon/vaddr.c
@@ -743,7 +743,7 @@ static int damos_va_migrate_pmd_entry(pmd_t *pmd, unsigned long addr,
 		if (!folio)
 			continue;
 		if (damos_va_filter_out(s, folio, walk->vma, addr, pte, NULL))
-			return 0;
+			continue;
 		damos_va_migrate_dests_add(folio, walk->vma, addr, dests,
 				migration_lists);
 		nr = folio_nr_pages(folio);
-- 
2.43.0

Re: [PATCH] mm/damon/vaddr: fix missing pte_unmap_unlock in damos_va_migrate_pmd_entry()

[SeongJae Park]

On Wed, 10 Dec 2025 00:10:34 +0900 Akinobu Mita <akinobu.mita@gmail.com> wrote:

> If the PTE page table lock is acquired by pte_offset_map_lock(), the lock
> must be released via pte_unmap_unlock().
> 
> However, in damos_va_migrate_pmd_entry(), if damos_va_filter_out() returns
> true, it immediately returns without releasing the lock.
> 
> This fixes the issue by not stopping page table traversal when
> damos_va_filter_out() returns true and ensuring that the lock is released.

Thank you for finding and fixing this!

> 
> Fixes: 09efc56a3b1c ("mm/damon/vaddr: consistently use only pmd_entry for damos_migrate")
> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>

Reviewed-by: SeongJae Park <sj@kernel.org>


Thanks,
SJ

[...]