Because of the possible alloc failure of the alloc_page(), it could
return NULL pointer.
And then it will cause the BUG_ON() in sg_assign_page().
Therefore, it should be better to check it before to avoid the bug.

Fixes: 2d97591ef43d ("crypto: af_alg - consolidation of duplicate code")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
---
 crypto/af_alg.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/crypto/af_alg.c b/crypto/af_alg.c
index 18cc82dc4a42..a1c0118e222d 100644
--- a/crypto/af_alg.c
+++ b/crypto/af_alg.c
@@ -931,11 +931,18 @@ int af_alg_sendmsg(struct socket *sock, struct msghdr *msg, size_t size,
 			sg_unmark_end(sg + sgl->cur - 1);
 
 		do {
+			struct page *pg;
 			unsigned int i = sgl->cur;
 
 			plen = min_t(size_t, len, PAGE_SIZE);
 
-			sg_assign_page(sg + i, alloc_page(GFP_KERNEL));
+			pg = alloc_page(GFP_KERNEL);
+			if (!pg) {
+				err = -ENOMEM;
+				goto unlock;
+			}
+
+			sg_assign_page(sg + i, pg);
 			if (!sg_page(sg + i)) {
 				err = -ENOMEM;
 				goto unlock;
-- 
2.25.1

On Wed, Dec 29, 2021 at 05:32:16PM +0800, Jiasheng Jiang wrote:
> And then it will cause the BUG_ON() in sg_assign_page().

That's not true.  The two BUG_ON()'s in sg_assign_page() are:

        BUG_ON((unsigned long) page & (SG_CHAIN | SG_END));
#ifdef CONFIG_DEBUG_SG
        BUG_ON(sg_is_chain(sg));
#endif

Neither will trigger due to page == NULL.

>  		do {
> +			struct page *pg;
>  			unsigned int i = sgl->cur;
>  
>  			plen = min_t(size_t, len, PAGE_SIZE);
>  
> -			sg_assign_page(sg + i, alloc_page(GFP_KERNEL));
> +			pg = alloc_page(GFP_KERNEL);
> +			if (!pg) {
> +				err = -ENOMEM;
> +				goto unlock;
> +			}
> +
> +			sg_assign_page(sg + i, pg);
>  			if (!sg_page(sg + i)) {
>  				err = -ENOMEM;
>  				goto unlock;

The NULL check is already done, just below the redundant one you're adding.

I do think it would be more logical to do the NULL check before
sg_assign_page().  So you could send a patch that moves it there.  But it would
be a cleanup, not a fix.

- Eric