diff mbox series

[v2,3/3] mm/rmap: sanity check that zeropages are not passed to RMAP

Message ID 20240522125713.775114-4-david@redhat.com (mailing list archive)
State New
Headers show
Series mm/memory: cleanly support zeropage in vm_insert_page*(), vm_map_pages*() and vmf_insert_mixed() | expand

Commit Message

David Hildenbrand May 22, 2024, 12:57 p.m. UTC 
Using insert_page() we might have previously ended up passing the zeropage
into rmap code. Make sure that won't happen again.

Note that we won't check the huge zeropage for now, which might still
end up in RMAP code.

Signed-off-by: David Hildenbrand <david@redhat.com>
---
 include/linux/rmap.h | 3 +++
 1 file changed, 3 insertions(+)
diff mbox series

Patch

diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index 7229b9baf20d..5cb0d419a1d7 100644
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -200,6 +200,9 @@  static inline void __folio_rmap_sanity_checks(struct folio *folio,
 	/* hugetlb folios are handled separately. */
 	VM_WARN_ON_FOLIO(folio_test_hugetlb(folio), folio);
 
+	/* When (un)mapping zeropages, we should never touch ref+mapcount. */
+	VM_WARN_ON_FOLIO(is_zero_folio(folio), folio);
+
 	/*
 	 * TODO: we get driver-allocated folios that have nothing to do with
 	 * the rmap using vm_insert_page(); therefore, we cannot assume that