[PATCH] io_uring: NULL check before kfree() is not needed

[PATCH] io_uring: NULL check before kfree() is not needed

[Wu Bo]

NULL check before kfree() is not needed

Signed-off-by: Wu Bo <wubo40@huawei.com>
---
 fs/io_uring.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 2a3af95..59c19c8 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -1537,8 +1537,7 @@ static void io_dismantle_req(struct io_kiocb *req)
 {
 	io_clean_op(req);
 
-	if (req->io)
-		kfree(req->io);
+	kfree(req->io);
 	if (req->file)
 		io_put_file(req, req->file, (req->flags & REQ_F_FIXED_FILE));
 	io_req_clean_work(req);
@@ -3111,8 +3110,7 @@ static int io_read(struct io_kiocb *req, bool force_nonblock,
 		return -EAGAIN;
 	}
 out_free:
-	if (iovec)
-		kfree(iovec);
+	kfree(iovec);
 	return ret;
 }
 
@@ -3210,8 +3208,7 @@ static int io_write(struct io_kiocb *req, bool force_nonblock,
 		return -EAGAIN;
 	}
 out_free:
-	if (iovec)
-		kfree(iovec);
+	kfree(iovec);
 	return ret;
 }
 
-- 
1.8.3.1

Re: [PATCH] io_uring: NULL check before kfree() is not needed

[Jens Axboe]

On 8/14/20 8:25 PM, Wu Bo wrote:
> NULL check before kfree() is not needed

While it's not needed, it's also slower. Particularly on AMD it seems.
It's on my list to ensure that kfree() is an inline that checks for
non-NULL before calling the real freeing function.

Should get around to that soon, as this is probably the third time
someone has sent this patch :-)

-- 
Jens Axboe