[PATCH hotfix] mm/compaction: handle free_pages_prepare() properly in compaction_free()

[PATCH hotfix] mm/compaction: handle free_pages_prepare() properly in compaction_free()

[Zi Yan]

free_pages_prepare() can fail but compaction_free() does not handle the
failure case. Failed pages should not be added back to cc->freepages for
future use, since they can be either PageHWPoison or free_page_is_bad()
and might cause data corruption.

Fixes: 733aea0b3a7bb ("mm/compaction: add support for >0 order folio memory compaction.")
Signed-off-by: Zi Yan <ziy@nvidia.com>
Cc: stable@vger.kernel.org
---
free_pages_prepare() can fail if a page is PageHWPoison or
free_page_is_bad(). compaction_free() needs to handle these cases to
prevent failed pages being reused in cc->freepages.
---
 mm/compaction.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/mm/compaction.c b/mm/compaction.c
index b776f35ad0200..f08765ade014c 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -1875,15 +1875,14 @@ static void compaction_free(struct folio *dst, unsigned long data)
 	int order = folio_order(dst);
 	struct page *page = &dst->page;
 
-	if (folio_put_testzero(dst)) {
-		free_pages_prepare(page, order);
+	if (folio_put_testzero(dst) && free_pages_prepare(page, order)) {
 		list_add(&dst->lru, &cc->freepages[order]);
 		cc->nr_freepages += 1 << order;
 	}
 	cc->nr_migratepages += 1 << order;
 	/*
-	 * someone else has referenced the page, we cannot take it back to our
-	 * free list.
+	 * someone else has referenced the page or free_pages_prepare() fails,
+	 * we cannot take it back to our free list.
 	 */
 }
 

---
base-commit: 13a1e1a618858407fa12c391f664ea750651f6b2
change-id: 20260621-handle_free_pages_prepare_in_compaction_free-7cca3ff08367

Best regards,
-- 
Zi Yan <ziy@nvidia.com>

Re: [PATCH hotfix] mm/compaction: handle free_pages_prepare() properly in compaction_free()

[Vlastimil Babka (SUSE)]

On 6/22/26 17:30, Zi Yan wrote:
> free_pages_prepare() can fail but compaction_free() does not handle the
> failure case. Failed pages should not be added back to cc->freepages for
> future use, since they can be either PageHWPoison or free_page_is_bad()
> and might cause data corruption.
> 
> Fixes: 733aea0b3a7bb ("mm/compaction: add support for >0 order folio memory compaction.")
> Signed-off-by: Zi Yan <ziy@nvidia.com>
> Cc: stable@vger.kernel.org

Reviewed-by: Vlastimil Babka (SUSE) <vbabka@kernel.org>

> ---
> free_pages_prepare() can fail if a page is PageHWPoison or
> free_page_is_bad(). compaction_free() needs to handle these cases to
> prevent failed pages being reused in cc->freepages.
> ---
>  mm/compaction.c | 7 +++----
>  1 file changed, 3 insertions(+), 4 deletions(-)
> 
> diff --git a/mm/compaction.c b/mm/compaction.c
> index b776f35ad0200..f08765ade014c 100644
> --- a/mm/compaction.c
> +++ b/mm/compaction.c
> @@ -1875,15 +1875,14 @@ static void compaction_free(struct folio *dst, unsigned long data)
>  	int order = folio_order(dst);
>  	struct page *page = &dst->page;
>  
> -	if (folio_put_testzero(dst)) {
> -		free_pages_prepare(page, order);
> +	if (folio_put_testzero(dst) && free_pages_prepare(page, order)) {
>  		list_add(&dst->lru, &cc->freepages[order]);
>  		cc->nr_freepages += 1 << order;
>  	}
>  	cc->nr_migratepages += 1 << order;
>  	/*
> -	 * someone else has referenced the page, we cannot take it back to our
> -	 * free list.
> +	 * someone else has referenced the page or free_pages_prepare() fails,
> +	 * we cannot take it back to our free list.
>  	 */
>  }
>  
> 
> ---
> base-commit: 13a1e1a618858407fa12c391f664ea750651f6b2
> change-id: 20260621-handle_free_pages_prepare_in_compaction_free-7cca3ff08367
> 
> Best regards,

Re: [PATCH hotfix] mm/compaction: handle free_pages_prepare() properly in compaction_free()

[Johannes Weiner]

On Mon, Jun 22, 2026 at 11:30:42AM -0400, Zi Yan wrote:
> free_pages_prepare() can fail but compaction_free() does not handle the
> failure case. Failed pages should not be added back to cc->freepages for
> future use, since they can be either PageHWPoison or free_page_is_bad()
> and might cause data corruption.
> 
> Fixes: 733aea0b3a7bb ("mm/compaction: add support for >0 order folio memory compaction.")
> Signed-off-by: Zi Yan <ziy@nvidia.com>
> Cc: stable@vger.kernel.org

Acked-by: Johannes Weiner <hannes@cmpxchg.org>