[RFC 2/6] mm: keep dirty bit on anonymous page migration

[Minchan Kim]

Currently, If VM migrates anonymous page, we lose dirty bit of
page table entry. Instead, VM translates dirty bit of page table
as PG_dirty of page flags. It was okay because dirty bit of
page table for anonymous page was no matter to swap out.
Instead, VM took care of PG_dirty.

However, with introducing MADV_FREE, it's important to keep
page table's dirty bit because It could make MADV_FREE handling
logics more straighforward without taking care of PG_dirty.

This patch aims for preparing to remove PG_dirty check for MADV_FREE.

Signed-off-by: Minchan Kim <minchan@xxxxxxxxxx>
---
 mm/migrate.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/mm/migrate.c b/mm/migrate.c
index 236ee25e79d9..add30c3aaaa9 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -151,6 +151,10 @@ static int remove_migration_pte(struct page *new, struct vm_area_struct *vma,
 	if (is_write_migration_entry(entry))
 		pte = maybe_mkwrite(pte, vma);
 
+	/* MADV_FREE relies on pte_dirty. */
+	if (PageAnon(new))
+		pte = pte_mkdirty(pte);
+
 #ifdef CONFIG_HUGETLB_PAGE
 	if (PageHuge(new)) {
 		pte = pte_mkhuge(pte);
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/