[PATCH v3 4/5] mm/page_alloc: make sure tail_page->private is zero at page free time

[Zi Yan]

Any code using tail_page->private of a folio, a compound page or a
high-order page is supposed to reset it after use, otherwise ->private data
can leak to new page user and cause unexpected issues. Add a bad_page()
check at page free path for it.

Reviewed-by: Vlastimil Babka (SUSE) <vbabka@xxxxxxxxxx>
Signed-off-by: Zi Yan <ziy@xxxxxxxxxx>
---
 mm/page_alloc.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index a63733dac659e..cc45dfe89c76b 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1377,15 +1377,23 @@ static __always_inline bool __free_pages_prepare(struct page *page,
 #endif
 		}
 		for (i = 1; i < (1 << order); i++) {
+			struct page *tail_page = page + i;
+
 			if (compound)
-				bad += free_tail_page_prepare(page, page + i);
+				bad += free_tail_page_prepare(page, tail_page);
 			if (is_check_pages_enabled()) {
-				if (free_page_is_bad(page + i)) {
+				if (free_page_is_bad(tail_page)) {
+					bad++;
+					continue;
+				}
+
+				if (tail_page->private) {
+					bad_page(tail_page, "nonzero private");
 					bad++;
 					continue;
 				}
 			}
-			(page + i)->flags.f &= ~PAGE_FLAGS_CHECK_AT_PREP;
+			tail_page->flags.f &= ~PAGE_FLAGS_CHECK_AT_PREP;
 		}
 	}
 	if (folio_test_anon(folio)) {

-- 
2.53.0