Re: BUG: unable to handle kernel paging request in io_wq_cancel_all

[Jens Axboe]

On 10/30/19 1:44 AM, syzbot wrote:
> syzbot has bisected this bug to:
> 
> commit ef0524d3654628ead811f328af0a4a2953a8310f
> Author: Jens Axboe <axboe@xxxxxxxxx>
> Date:   Thu Oct 24 13:25:42 2019 +0000
> 
>       io_uring: replace workqueue usage with io-wq
> 
> bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=16acf5d0e00000
> start commit:   c57cf383 Add linux-next specific files for 20191029
> git tree:       linux-next
> final crash:    https://syzkaller.appspot.com/x/report.txt?x=15acf5d0e00000
> console output: https://syzkaller.appspot.com/x/log.txt?x=11acf5d0e00000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=cb86688f30db053d
> dashboard link: https://syzkaller.appspot.com/bug?extid=221cc24572a2fed23b6b
> syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=168671d4e00000
> C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=140f4898e00000
> 
> Reported-by: syzbot+221cc24572a2fed23b6b@xxxxxxxxxxxxxxxxxxxxxxxxx
> Fixes: ef0524d36546 ("io_uring: replace workqueue usage with io-wq")

Good catch, it's a case of NULL vs ERR_PTR() confusion. I'll fold in
the below fix.


diff --git a/fs/io_uring.c b/fs/io_uring.c
index af1937d66aee..76d653085987 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -3534,8 +3534,9 @@ static int io_sq_offload_start(struct io_ring_ctx *ctx,
 	/* Do QD, or 4 * CPUS, whatever is smallest */
 	concurrency = min(ctx->sq_entries, 4 * num_online_cpus());
 	ctx->io_wq = io_wq_create(concurrency, ctx->sqo_mm);
-	if (!ctx->io_wq) {
-		ret = -ENOMEM;
+	if (IS_ERR(ctx->io_wq)) {
+		ret = PTR_ERR(ctx->io_wq);
+		ctx->io_wq = NULL;
 		goto err;
 	}
 

-- 
Jens Axboe