[PATCH 0/2] mm/migrate: preserve soft dirty in remove_migration_pte()

[PATCH 0/2] mm/migrate: preserve soft dirty in remove_migration_pte()

[Ralph Campbell]

Two small patches for Andrew Morton's mm tree.
I happened to notice this from code inspection after seeing Alistair
Popple's patch ("mm/rmap: Fixup copying of soft dirty and uffd ptes").

Ralph Campbell (2):
  mm/migrate: remove unnecessary is_zone_device_page() check
  mm/migrate: preserve soft dirty in remove_migration_pte()

 mm/migrate.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

-- 
2.20.1

[PATCH 1/2] mm/migrate: remove unnecessary is_zone_device_page() check

[Ralph Campbell]

The check for is_zone_device_page() and is_device_private_page() is
unnecessary since the latter is sufficient to determine if the page
is a device private page. Simplify the code for easier reading.

Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
---
 mm/migrate.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/mm/migrate.c b/mm/migrate.c
index 659d3d8a3e1f..fe339a847328 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -246,13 +246,11 @@ static bool remove_migration_pte(struct page *page, struct vm_area_struct *vma,
 		else if (pte_swp_uffd_wp(*pvmw.pte))
 			pte = pte_mkuffd_wp(pte);
 
-		if (unlikely(is_zone_device_page(new))) {
-			if (is_device_private_page(new)) {
-				entry = make_device_private_entry(new, pte_write(pte));
-				pte = swp_entry_to_pte(entry);
-				if (pte_swp_uffd_wp(*pvmw.pte))
-					pte = pte_swp_mkuffd_wp(pte);
-			}
+		if (unlikely(is_device_private_page(new))) {
+			entry = make_device_private_entry(new, pte_write(pte));
+			pte = swp_entry_to_pte(entry);
+			if (pte_swp_uffd_wp(*pvmw.pte))
+				pte = pte_swp_mkuffd_wp(pte);
 		}
 
 #ifdef CONFIG_HUGETLB_PAGE
-- 
2.20.1

Re: [PATCH 1/2] mm/migrate: remove unnecessary is_zone_device_page() check

[Christoph Hellwig]

Looks good:

Reviewed-by: Christoph Hellwig <hch@lst.de>

[PATCH 2/2] mm/migrate: preserve soft dirty in remove_migration_pte()

[Ralph Campbell]

The code to remove a migration PTE and replace it with a device private
PTE was not copying the soft dirty bit from the migration entry.
This could lead to page contents not being marked dirty when faulting
the page back from device private memory.

Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
---
 mm/migrate.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/mm/migrate.c b/mm/migrate.c
index fe339a847328..4f89360d9e77 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -249,6 +249,8 @@ static bool remove_migration_pte(struct page *page, struct vm_area_struct *vma,
 		if (unlikely(is_device_private_page(new))) {
 			entry = make_device_private_entry(new, pte_write(pte));
 			pte = swp_entry_to_pte(entry);
+			if (pte_swp_soft_dirty(*pvmw.pte))
+				pte = pte_swp_mksoft_dirty(pte);
 			if (pte_swp_uffd_wp(*pvmw.pte))
 				pte = pte_swp_mkuffd_wp(pte);
 		}
-- 
2.20.1

Re: [PATCH 2/2] mm/migrate: preserve soft dirty in remove_migration_pte()

[Christoph Hellwig]

On Mon, Aug 31, 2020 at 02:22:22PM -0700, Ralph Campbell wrote:
> The code to remove a migration PTE and replace it with a device private
> PTE was not copying the soft dirty bit from the migration entry.
> This could lead to page contents not being marked dirty when faulting
> the page back from device private memory.
> 
> Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>

Looks good,

Reviewed-by: Christoph Hellwig <hch@lst.de>