mm/migrate_device: fix pgtable leak in migrate_vma_insert_huge_pmd_page
Subsystems (2)
PATCH 1/1 mm/migrate_device: fix pgtable leak in migrate_vma_insert_huge_pmd_page
When migrate_vma_insert_huge_pmd_page() jumps to unlock_abort due to a PMD check failure, the pgtable allocated earlier via pte_alloc_one() is never freed, causing a memory leak.
Add a pte_free() call in the unlock_abort error path to release the pgtable before returning.
--- mm/migrate_device.c | 1 + 1 file changed, 1 insertion(+)
mm/migrate_device.c +1 -0
@@ -893,6 +893,7 @@ static int migrate_vma_insert_huge_pmd_page(struct migrate_vma *migrate, unlock_abort:spin_unlock (ptl); +pte_free (vma->vm_mm, pgtable); abort:for (i =0 ; i <HPAGE_PMD_NR ; i++) src[i] &= ~MIGRATE_PFN_MIGRATE ;
Signature · 1 line
2.43.0
Quoted 4 lines · Sunny Patel
When
migrate_vma_insert_huge_pmd_page()jumps tounlock_abortdue to a PMD check failure, the pgtable allocated earlier viapte_alloc_one()is never freed, causing a memory leak.
Quoted 8 lines
Add a
pte_free()call in theunlock_aborterror path to release the pgtable before returning.Signed-off-by Sunny Patel <[email protected]>--- mm/migrate_device.c | 1 + 1 file changed, 1 insertion(+)
LGTM.
--
Best Regards,
Yan, Zi
Zi Yan
reviewerQuoted 26 lines · Sunny Patel
When
migrate_vma_insert_huge_pmd_page()jumps tounlock_abortdue to a PMD check failure, the pgtable allocated earlier viapte_alloc_one()is never freed, causing a memory leak.Add a
pte_free()call in theunlock_aborterror path to release the pgtable before returning.Signed-off-by Sunny Patel <[email protected]>--- mm/migrate_device.c | 1 + 1 file changed, 1 insertion(+)mm/migrate_device.c +1
-0@@ -893,6 +893,7 @@ static int migrate_vma_insert_huge_pmd_page(struct migrate_vma *migrate, unlock_abort:spin_unlock (ptl); +pte_free (vma->vm_mm, pgtable); abort:for (i =0 ; i <HPAGE_PMD_NR ; i++) src[i] &= ~MIGRATE_PFN_MIGRATE ;Signature · 1 line
2.43.0
Actually,
I think the "goto abort" above pmd_lock() need to do pte_free():
if (folio_is_device_private(folio)) {
...
} else {
if (folio_is_zone_device(folio) &&
!folio_is_device_coherent(folio)) {
goto abort; <-- this one
}
...
}The lock is not taken yet, so you might need to add a pte_free() before goto abort.
--
Best Regards,
Yan, Zi
Quoted 12 lines · Zi Yan
I think the "goto abort" above
pmd_lock()need to dopte_free():if (folio_is_device_private(folio)) { ... } else { if (folio_is_zone_device(folio) && !folio_is_device_coherent(folio)) { goto abort; <-- this one } ... }
Quoted 2 lines
The lock is not taken yet, so you might need to add a
pte_free()before goto abort.
You are right. The pgtable is allocated before this check and the lock has not been taken yet at this point, so goto unlock_abort would be incorrect here. A pte_free() call is needed before this goto abort.
Sent v2 with the changes.
https://lore.kernel.org/linux-mm/[email protected]/
Thanks for catching this.
Regards,
Sunny Patel
Huang, Ying
reviewerHi, Sunny,
Thanks for working on this!
Quoted 17 lines · Sunny Patel
When
migrate_vma_insert_huge_pmd_page()jumps tounlock_abortdue to a PMD check failure, the pgtable allocated earlier viapte_alloc_one()is never freed, causing a memory leak.Add a
pte_free()call in theunlock_aborterror path to release the pgtable before returning.Signed-off-by Sunny Patel <[email protected]>--- mm/migrate_device.c | 1 + 1 file changed, 1 insertion(+)mm/migrate_device.c +0
-0
mm/migrate_device.c static int migrate_vma_insert_huge_pmd_page(struct migrate_vma *migrate, · line 899
Quoted 7 lines
mm/migrate_device.c +1
-0@@ -893,6 +893,7 @@ static int migrate_vma_insert_huge_pmd_page(struct migrate_vma *migrate, unlock_abort:spin_unlock (ptl); +pte_free (vma->vm_mm, pgtable); abort:for (i =0 ; i <HPAGE_PMD_NR ; i++) src[i] &= ~MIGRATE_PFN_MIGRATE ;
Is it better to use guard based automatic memory freeing?
And check whether guard can help unlock case too?
---
Best Regards,
Huang, Ying