Add missing NULL check for kmalloc return value in
initramfs_test_fname_overrun(). Without this check, if kmalloc fails,
the subsequent memset() will dereference a NULL pointer, causing a
kernel crash.

Fixes: 83c0b27266ec ("initramfs_test: kunit tests for initramfs unpacking")
Signed-off-by: longlong yan <yanlonglong@kylinos.cn>
---
 init/initramfs_test.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/init/initramfs_test.c b/init/initramfs_test.c
index 1fc990a66563..c7aab7e5383e 100644
--- a/init/initramfs_test.c
+++ b/init/initramfs_test.c
@@ -189,6 +189,9 @@ static void __init initramfs_test_fname_overrun(struct kunit *test)
 	 * are already available (e.g. no compression).
 	 */
 	cpio_srcbuf = kmalloc(CPIO_HDRLEN + PATH_MAX + 3, GFP_KERNEL);
+	if (!cpio_srcbuf)
+		return;
+
 	memset(cpio_srcbuf, 'B', CPIO_HDRLEN + PATH_MAX + 3);
 	/* limit overrun to avoid crashes / filp_open() ENAMETOOLONG */
 	cpio_srcbuf[CPIO_HDRLEN + strlen(c[0].fname) + 20] = '\0';
-- 
2.43.0

Hi,

On Wed, 15 Jul 2026 10:08:19 +0800, longlong yan wrote:

> Add missing NULL check for kmalloc return value in
> initramfs_test_fname_overrun(). Without this check, if kmalloc fails,
> the subsequent memset() will dereference a NULL pointer, causing a
> kernel crash.
> 
> Fixes: 83c0b27266ec ("initramfs_test: kunit tests for initramfs unpacking")
> Signed-off-by: longlong yan <yanlonglong@kylinos.cn>
> ---
>  init/initramfs_test.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/init/initramfs_test.c b/init/initramfs_test.c
> index 1fc990a66563..c7aab7e5383e 100644
> --- a/init/initramfs_test.c
> +++ b/init/initramfs_test.c
> @@ -189,6 +189,9 @@ static void __init initramfs_test_fname_overrun(struct kunit *test)
>  	 * are already available (e.g. no compression).
>  	 */
>  	cpio_srcbuf = kmalloc(CPIO_HDRLEN + PATH_MAX + 3, GFP_KERNEL);
> +	if (!cpio_srcbuf)
> +		return;
> +
>  	memset(cpio_srcbuf, 'B', CPIO_HDRLEN + PATH_MAX + 3);

Panicking on kunit test alloc failure is reasonable behaviour IMO;
the Fixes tag isn't appropriate here. Also, your patch is only
handling one of the many allocations. KUNIT_ASSERT_NOT_NULL() would be
a better option if we wanted to explictly catch these failures.