BPF List
 help / color / mirror / Atom feed
From: Subasri S <subasris1210@gmail.com>
To: Jakub Kicinski <kuba@kernel.org>,
	Andrew Lunn <andrew+netdev@lunn.ch>,
	 "David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	 Paolo Abeni <pabeni@redhat.com>,
	Alexei Starovoitov <ast@kernel.org>,
	 Daniel Borkmann <daniel@iogearbox.net>,
	 Jesper Dangaard Brouer <hawk@kernel.org>,
	 John Fastabend <john.fastabend@gmail.com>,
	 Stanislav Fomichev <sdf@fomichev.me>,
	Willem de Bruijn <willemb@google.com>,
	 Mina Almasry <almasrymina@google.com>
Cc: netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
	 bpf@vger.kernel.org,
	syzbot+c06674caba265dc61d46@syzkaller.appspotmail.com,
	 Subasri S <subasris1210@gmail.com>
Subject: [PATCH] netdevsim: update rxq->napi pointer during queue reset
Date: Wed, 12 Aug 2026 19:19:06 +0530	[thread overview]
Message-ID: <20260812-net-netdevsim-v1-1-521a85e4fe91@gmail.com> (raw)

In netdevsim, when queue reset is performed using debugfs, it triggers
these sequence of operations: nsim_queue_stop() -> nsim_queue_start()
-> nsim_queue_mem_free(). nsim_queue_mem_free() frees the old
nsim_rq struct which embeds the napi_struct. But the rxq->napi pointer
in the struct netdev_rx_queue still points to the old nsim_rq's embedded
napi_struct. So, any subsequent xsk_bind() which reads rxq->napi->napi_id
after a queue reset is a use-after-free.

Add netif_queue_set_napi() calls to nsim_queue_stop() and
nsim_queue_start() which clears the rxq->napi during stop
and sets it to the new napi instance during start.

KASAN report:

Call Trace:
 kasan_report+0xdf/0x1c0 mm/kasan/report.c:595
 xsk_bind+0x1582/0x16c0 net/xdp/xsk.c:1758
 __sys_bind_socket net/socket.c:1920 [inline]
 __sys_bind_socket net/socket.c:1912 [inline]
 __sys_bind+0x1a9/0x260 net/socket.c:1951

Allocated by task 5622:
 nsim_queue_alloc+0x3c/0x140 drivers/net/netdevsim/netdev.c:715
 nsim_queue_init drivers/net/netdevsim/netdev.c:1012 [inline]
 nsim_init_netdevsim drivers/net/netdevsim/netdev.c:1059 [inline]
 nsim_create+0xb13/0x1420 drivers/net/netdevsim/netdev.c:1152
 __nsim_dev_port_add+0x3ba/0x8f0 drivers/net/netdevsim/dev.c:1509
 nsim_dev_port_add_all drivers/net/netdevsim/dev.c:1570 [inline]
 nsim_drv_probe+0xdbd/0x13a0 drivers/net/netdevsim/dev.c:1731

Freed by task 5659:
 slab_free mm/slub.c:6377 [inline]
 kfree+0x22b/0x6c0 mm/slub.c:6692
 nsim_queue_mem_free+0xfe/0x190 drivers/net/netdevsim/netdev.c:796
 netdev_rx_queue_reconfig+0x405/0x630 net/core/netdev_rx_queue.c:144
 netdev_rx_queue_restart+0x8f/0xc0 net/core/netdev_rx_queue.c:183
 nsim_qreset_write+0x2e3/0x410 drivers/net/netdevsim/netdev.c:887

Reported-by: syzbot+c06674caba265dc61d46@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=c06674caba265dc61d46
Fixes: 5bc8e8dbef27 ("netdevsim: add queue management API support")
Tested-by: syzbot+c06674caba265dc61d46@syzkaller.appspotmail.com
Signed-off-by: Subasri S <subasris1210@gmail.com>
---
 drivers/net/netdevsim/netdev.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c
index 4e9d7e10b527..c18f9f540084 100644
--- a/drivers/net/netdevsim/netdev.c
+++ b/drivers/net/netdevsim/netdev.c
@@ -826,6 +826,8 @@ nsim_queue_start(struct net_device *dev, struct netdev_queue_config *qcfg,
 	}
 
 	ns->rq[idx] = qmem->rq;
+	netif_queue_set_napi(dev, idx, NETDEV_QUEUE_TYPE_RX,
+			     &ns->rq[idx]->napi);
 	napi_enable_locked(&ns->rq[idx]->napi);
 
 	return 0;
@@ -838,6 +840,7 @@ static int nsim_queue_stop(struct net_device *dev, void *per_queue_mem, int idx)
 
 	netdev_assert_locked(dev);
 
+	netif_queue_set_napi(dev, idx, NETDEV_QUEUE_TYPE_RX, NULL);
 	napi_disable_locked(&ns->rq[idx]->napi);
 
 	if (ns->rq_reset_mode == 1) {

---
base-commit: a59f57e2aa127c5354168d2ec4bac920df1be4f4
change-id: 20260812-net-netdevsim-aad6100834dd

Best regards,
-- 
Subasri S <subasris1210@gmail.com>


             reply	other threads:[~2026-08-12 13:49 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-12 13:49 Subasri S [this message]
2026-08-12 14:02 ` [PATCH] netdevsim: update rxq->napi pointer during queue reset sashiko-bot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260812-net-netdevsim-v1-1-521a85e4fe91@gmail.com \
    --to=subasris1210@gmail.com \
    --cc=almasrymina@google.com \
    --cc=andrew+netdev@lunn.ch \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=hawk@kernel.org \
    --cc=john.fastabend@gmail.com \
    --cc=kuba@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=sdf@fomichev.me \
    --cc=syzbot+c06674caba265dc61d46@syzkaller.appspotmail.com \
    --cc=willemb@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox