On 30.01.25 14:46, Simona Vetter wrote:
On Wed, Jan 29, 2025 at 12:54:01PM +0100, David Hildenbrand wrote:
The single "real" user in the tree of make_device_exclusive_range() always
requests making only a single address exclusive. The current implementation
is hard to fix for properly supporting anonymous THP / large folios and
for avoiding messing with rmap walks in weird ways.
So let's always process a single address/page and return folio + page to
minimize page -> folio lookups. This is a preparation for further
changes.
Reject any non-anonymous or hugetlb folios early, directly after GUP.
Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
Yeah this makes sense. Even for pmd entries I think we want to make this
very explicit with an explicit hugetlb opt-in I think.
Acked-by: Simona Vetter <simona.vetter@xxxxxxxx>
Thanks, I'll fold in the following:
diff --git a/mm/rmap.c b/mm/rmap.c
index 676df4fba5b0..94256925682d 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -2525,6 +2525,10 @@ static bool folio_make_device_exclusive(struct folio *folio,
* programming is complete it should drop the page lock and reference after
* which point CPU access to the page will revoke the exclusive access.
*
+ * Note: This function always operates on individual PTEs mapping individual
+ * pages. PMD-sized THPs are first remapped to be mapped by PTEs before the
+ * conversion happens on a single PTE corresponding to @addr.
+ *
* Returns: pointer to mapped page on success, otherwise a negative error.
*/
struct page *make_device_exclusive(struct mm_struct *mm, unsigned long addr,
--
Cheers,
David / dhildenb
