netdev - Re: [syzbot ci] Re: netkit: Support for io_uring zero-copy and AF

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <3589f74a-ee6e-4f5d-a93e-8b0f8a3d1222@iogearbox.net>
Date: Thu, 16 Oct 2025 14:12:24 +0200
From: Daniel Borkmann <daniel@...earbox.net>
To: syzbot ci <syzbot+ci7c73a60f40f79ce2@...kaller.appspotmail.com>,
 bpf@...r.kernel.org, davem@...emloft.net, dw@...idwei.uk,
 john.fastabend@...il.com, jordan@...fe.io, kuba@...nel.org,
 maciej.fijalkowski@...el.com, magnus.karlsson@...el.com,
 martin.lau@...nel.org, netdev@...r.kernel.org, pabeni@...hat.com,
 razor@...ckwall.org, sdf@...ichev.me, toke@...hat.com,
 wangdongdong.6@...edance.com, willemb@...gle.com, yangzhenze@...edance.com
Cc: syzbot@...ts.linux.dev, syzkaller-bugs@...glegroups.com
Subject: Re: [syzbot ci] Re: netkit: Support for io_uring zero-copy and AF_XDP

On 10/15/25 9:13 PM, Daniel Borkmann wrote:
> On 10/15/25 8:46 PM, syzbot ci wrote:
>> syzbot ci has tested the following series
> [...]
>> WARNING: CPU: 1 PID: 5959 at ./include/net/netdev_lock.h:17 netdev_assert_locked include/net/netdev_lock.h:17 [inline]
>> WARNING: CPU: 1 PID: 5959 at ./include/net/netdev_lock.h:17 netif_get_rx_queue_peer_locked+0x2f1/0x3a0 net/core/netdev_rx_queue.c:71
> [...]
>>   <TASK>
>>   xsk_reg_pool_at_qid+0x20b/0x630 net/xdp/xsk.c:159
> 
> Hmm, looks like syzkaller tried using batman_adv interface. The lock in the
> xsk path uses netdev_lock_ops() which is why the assert hits. I'll change
> netif_get_rx_queue_peer_locked() to check for netdev_need_ops_lock() in v3.

(Just to clarify: I now locally replaced the netdev_assert_locked() into using
netdev_ops_assert_locked() to address it.)