An anonymous large folio can be split into non order-0 folios,
try_to_map_unused_to_zeropage() should not VM_BUG_ON compound pages but
just return false. This fixes the crash when splitting anonymous large
folios to non order-0 folios.

Fixes: b1f202060afe ("mm: remap unused subpages to shared zeropage when splitting isolated thp")
Signed-off-by: Zi Yan <ziy@nvidia.com>
---
 mm/migrate.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/mm/migrate.c b/mm/migrate.c
index e950fd62607f..7ffdbe078aa7 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -206,7 +206,8 @@ static bool try_to_map_unused_to_zeropage(struct page_vma_mapped_walk *pvmw,
 	pte_t newpte;
 	void *addr;
 
-	VM_BUG_ON_PAGE(PageCompound(page), page);
+	if (PageCompound(page))
+		return false;
 	VM_BUG_ON_PAGE(!PageAnon(page), page);
 	VM_BUG_ON_PAGE(!PageLocked(page), page);
 	VM_BUG_ON_PAGE(pte_present(*pvmw->pte), page);
-- 
2.45.2

On 23/10/2024 18:12, Zi Yan wrote:
> An anonymous large folio can be split into non order-0 folios,
> try_to_map_unused_to_zeropage() should not VM_BUG_ON compound pages but
> just return false. This fixes the crash when splitting anonymous large
> folios to non order-0 folios.
> 
> Fixes: b1f202060afe ("mm: remap unused subpages to shared zeropage when splitting isolated thp")
> Signed-off-by: Zi Yan <ziy@nvidia.com>
> ---
>  mm/migrate.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/migrate.c b/mm/migrate.c
> index e950fd62607f..7ffdbe078aa7 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -206,7 +206,8 @@ static bool try_to_map_unused_to_zeropage(struct page_vma_mapped_walk *pvmw,
>  	pte_t newpte;
>  	void *addr;
>  
> -	VM_BUG_ON_PAGE(PageCompound(page), page);
> +	if (PageCompound(page))
> +		return false;
>  	VM_BUG_ON_PAGE(!PageAnon(page), page);
>  	VM_BUG_ON_PAGE(!PageLocked(page), page);
>  	VM_BUG_ON_PAGE(pte_present(*pvmw->pte), page);

Thanks for fixing it!

Acked-by: Usama Arif <usamaarif642@gmail.com>

On 23.10.24 19:12, Zi Yan wrote:
> An anonymous large folio can be split into non order-0 folios,
> try_to_map_unused_to_zeropage() should not VM_BUG_ON compound pages but
> just return false. This fixes the crash when splitting anonymous large
> folios to non order-0 folios.
> 
> Fixes: b1f202060afe ("mm: remap unused subpages to shared zeropage when splitting isolated thp")
> Signed-off-by: Zi Yan <ziy@nvidia.com>
> ---
>   mm/migrate.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/migrate.c b/mm/migrate.c
> index e950fd62607f..7ffdbe078aa7 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -206,7 +206,8 @@ static bool try_to_map_unused_to_zeropage(struct page_vma_mapped_walk *pvmw,
>   	pte_t newpte;
>   	void *addr;
>   
> -	VM_BUG_ON_PAGE(PageCompound(page), page);
> +	if (PageCompound(page))
> +		return false;
>   	VM_BUG_ON_PAGE(!PageAnon(page), page);
>   	VM_BUG_ON_PAGE(!PageLocked(page), page);
>   	VM_BUG_ON_PAGE(pte_present(*pvmw->pte), page);

Would read nicer in 2024 when working on folios ;)

Acked-by: David Hildenbrand <david@redhat.com>

-- 
Cheers,

David / dhildenb