[PATCH v2 2/2] mm/vmalloc: reject NULL pages in remap_vmalloc_range_partial

[Ye Liu <ye.liu@linux.dev>]

From: Ye Liu <liuye@kylinos.cn>

remap_vmalloc_range_partial() passes vmalloc_to_page()'s result to
vm_insert_page() unchecked, oopsing via page_folio() on a NULL page.
The NULL case cannot happen today because VM_USERMAP areas keep every
page mapped, but nothing enforces that invariant. Warn once and fail
the mapping with -EINVAL instead of oopsing if it is ever broken.

Suggested-by: Jose A. Perez de Azpillaga <azpijr@gmail.com>
Signed-off-by: Ye Liu <liuye@kylinos.cn>
---
 mm/vmalloc.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index a888eca016a8..dabef40ea88b 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -4906,6 +4906,9 @@ int remap_vmalloc_range_partial(struct vm_area_struct *vma, unsigned long uaddr,
 		struct page *page = vmalloc_to_page(kaddr);
 		int ret;
 
+		if (WARN_ON_ONCE(!page))
+			return -EINVAL;
+
 		ret = vm_insert_page(vma, uaddr, page);
 		if (ret)
 			return ret;

-- 
2.43.0