The malloc() function may return NULL when it fails,
which may cause null pointer deference in kmalloc(),
add Null check for return value of malloc().

Found by our static analysis tool.

Signed-off-by: Chenyuan Mi <cymi20@fudan.edu.cn>
---
 tools/io_uring/io_uring-bench.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/tools/io_uring/io_uring-bench.c b/tools/io_uring/io_uring-bench.c
index 7703f0118385..a7fedfdb9b84 100644
--- a/tools/io_uring/io_uring-bench.c
+++ b/tools/io_uring/io_uring-bench.c
@@ -560,6 +560,11 @@ int main(int argc, char *argv[])
 	pthread_create(&s->thread, NULL, submitter_fn, s);
 
 	fdepths = malloc(8 * s->nr_files);
+	if (!fdepths) {
+		printf("malloc failed");
+		return 1;
+	}
+
 	reap = calls = done = 0;
 	do {
 		unsigned long this_done = 0;
-- 
2.17.1

On 6/15/23 7:50 PM, Chenyuan Mi wrote:
> The malloc() function may return NULL when it fails,
> which may cause null pointer deference in kmalloc(),

It's a userspace app, there is no kmalloc(). Also, I don't think it's
worth to fix a missing ENOMEM handling for that old test program. But
anyway, let's wait for maintainers' comment on this.

-- 
Ammar Faizi

On 6/15/23 7:00?AM, Ammar Faizi wrote:
> On 6/15/23 7:50 PM, Chenyuan Mi wrote:
>> The malloc() function may return NULL when it fails,
>> which may cause null pointer deference in kmalloc(),
> 
> It's a userspace app, there is no kmalloc(). Also, I don't think it's
> worth to fix a missing ENOMEM handling for that old test program. But
> anyway, let's wait for maintainers' comment on this.

Definitely not worth it, and I find it odd how the author would target
just one of multiple allocations in that file. I'm guessing it's because
this checker only checks for malloc(), and no thought has otherwise gone
into a) if the patch makes any sense at all, and b) if it does make
sense, are there potentially other cases to consider?

-- 
Jens Axboe