Re: KASAN: use-after-free Read in io_uring_setup (2)

[Jens Axboe <axboe@kernel.dk>]

On 7/30/20 1:21 PM, syzbot wrote:
> Hello,
> 
> syzbot found the following issue on:
> 
> HEAD commit:    04b45717 Add linux-next specific files for 20200729
> git tree:       linux-next
> console output: https://syzkaller.appspot.com/x/log.txt?x=173774b8900000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=ec68f65b459f1ed
> dashboard link: https://syzkaller.appspot.com/bug?extid=9d46305e76057f30c74e
> compiler:       gcc (GCC) 10.1.0-syz 20200507
> 
> Unfortunately, I don't have any reproducer for this issue yet.
> 
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+9d46305e76057f30c74e@syzkaller.appspotmail.com
> 
> ==================================================================
> BUG: KASAN: use-after-free in io_account_mem fs/io_uring.c:7397 [inline]
> BUG: KASAN: use-after-free in io_uring_create fs/io_uring.c:8369 [inline]
> BUG: KASAN: use-after-free in io_uring_setup+0x2797/0x2910 fs/io_uring.c:8400
> Read of size 1 at addr ffff888087a41044 by task syz-executor.5/18145

Quick guess would be that the ring is closed in a race before we do the
accounting. The below should fix that, by ensuring that we account the
memory before we install the fd.

diff --git a/fs/io_uring.c b/fs/io_uring.c
index fabf0b692384..eb99994de5e2 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -8329,6 +8329,11 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p,
 		ret = -EFAULT;
 		goto err;
 	}
+
+	io_account_mem(ctx, ring_pages(p->sq_entries, p->cq_entries),
+		       ACCT_LOCKED);
+	ctx->limit_mem = limit_mem;
+
 	/*
 	 * Install ring fd as the very last thing, so we don't risk someone
 	 * having closed it before we finish setup
@@ -8338,9 +8343,6 @@ static int io_uring_create(unsigned entries, struct io_uring_params *p,
 		goto err;
 
 	trace_io_uring_create(ret, ctx, p->sq_entries, p->cq_entries, p->flags);
-	io_account_mem(ctx, ring_pages(p->sq_entries, p->cq_entries),
-		       ACCT_LOCKED);
-	ctx->limit_mem = limit_mem;
 	return ret;
 err:
 	io_ring_ctx_wait_and_kill(ctx);

-- 
Jens Axboe