Re: [PATCH] vme: Added NULL check for bridge

[Dan Carpenter]

On Sun, Jun 19, 2022 at 12:36:45PM +0530, Souptick Joarder wrote:
> From: "Souptick Joarder (HPE)" <jrdr.linux@xxxxxxxxx>
> 
> Kernel test robot throws below warning ->
> drivers/staging/vme_user/vme.c:662:20: warning: dereference
> of NULL 'bridge' [CWE-476] [-Wanalyzer-null-dereference]
> 
> Added a NULL check.
> 
> Reported-by: Kernel test robot <lkp@xxxxxxxxx>
> Signed-off-by: Souptick Joarder (HPE) <jrdr.linux@xxxxxxxxx>
> ---
>  drivers/staging/vme_user/vme.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/staging/vme_user/vme.c b/drivers/staging/vme_user/vme.c
> index b5555683a069..ede774f2fe5a 100644
> --- a/drivers/staging/vme_user/vme.c
> +++ b/drivers/staging/vme_user/vme.c
> @@ -659,7 +659,7 @@ ssize_t vme_master_read(struct vme_resource *resource, void *buf, size_t count,
>  	struct vme_master_resource *image;
>  	size_t length;
>  
> -	if (!bridge->master_read) {
> +	if (bridge && !bridge->master_read) {

This patch does not help anything.  If "bridge" is NULL here then the
kernel will just crash later in the function.  Smatch complains about it
now:

drivers/staging/vme_user/vme.c:688 vme_master_read() error: we previously assumed 'bridge' could be null (see line 666)

However, I don't believe that bridge can actually be NULL here.  We do
not write code just to silence that static checker false positives.  Fix
the checker instead.

regards,
dan carpenter