Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Stefano Garzarella <sgarzare@redhat.com>
To:     netdev@vger.kernel.org
Cc:     Bobby Eshleman <bobby.eshleman@bytedance.com>,
        "David S. Miller" <davem@davemloft.net>,
        Paolo Abeni <pabeni@redhat.com>,
        Eric Dumazet <edumazet@google.com>,
        linux-kernel@vger.kernel.org, avkrasnov@sberdevices.ru,
        Jakub Kicinski <kuba@kernel.org>,
        virtualization@lists.linux-foundation.org,
        Stefano Garzarella <sgarzare@redhat.com>,
        syzbot+befff0a9536049e7902e@syzkaller.appspotmail.com
Subject: [PATCH net] vsock/loopback: use only sk_buff_head.lock to protect the packet queue
Date:   Fri, 24 Mar 2023 12:54:50 +0100
Message-Id: <20230324115450.11268-1-sgarzare@redhat.com>
MIME-Version: 1.0
Content-type: text/plain
Content-Transfer-Encoding: 8bit
Precedence: bulk

pkt_list_lock was used before commit 71dc9ec9ac7d ("virtio/vsock:
replace virtio_vsock_pkt with sk_buff") to protect the packet queue.
After that commit we switched to sk_buff and we are using
sk_buff_head.lock in almost every place to protect the packet queue
except in vsock_loopback_work() when we call skb_queue_splice_init().

As reported by syzbot, this caused unlocked concurrent access to the
packet queue between vsock_loopback_work() and
vsock_loopback_cancel_pkt() since it is not holding pkt_list_lock.

With the introduction of sk_buff_head, pkt_list_lock is redundant and
can cause confusion, so let's remove it and use sk_buff_head.lock
everywhere to protect the packet queue access.

Fixes: 71dc9ec9ac7d ("virtio/vsock: replace virtio_vsock_pkt with sk_buff")
Cc: bobby.eshleman@bytedance.com
Reported-and-tested-by: syzbot+befff0a9536049e7902e@syzkaller.appspotmail.com
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
---
 net/vmw_vsock/vsock_loopback.c | 10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/net/vmw_vsock/vsock_loopback.c b/net/vmw_vsock/vsock_loopback.c
index 671e03240fc5..89905c092645 100644
--- a/net/vmw_vsock/vsock_loopback.c
+++ b/net/vmw_vsock/vsock_loopback.c
@@ -15,7 +15,6 @@
 struct vsock_loopback {
 	struct workqueue_struct *workqueue;
 
-	spinlock_t pkt_list_lock; /* protects pkt_list */
 	struct sk_buff_head pkt_queue;
 	struct work_struct pkt_work;
 };
@@ -32,9 +31,7 @@ static int vsock_loopback_send_pkt(struct sk_buff *skb)
 	struct vsock_loopback *vsock = &the_vsock_loopback;
 	int len = skb->len;
 
-	spin_lock_bh(&vsock->pkt_list_lock);
 	skb_queue_tail(&vsock->pkt_queue, skb);
-	spin_unlock_bh(&vsock->pkt_list_lock);
 
 	queue_work(vsock->workqueue, &vsock->pkt_work);
 
@@ -113,9 +110,9 @@ static void vsock_loopback_work(struct work_struct *work)
 
 	skb_queue_head_init(&pkts);
 
-	spin_lock_bh(&vsock->pkt_list_lock);
+	spin_lock_bh(&vsock->pkt_queue.lock);
 	skb_queue_splice_init(&vsock->pkt_queue, &pkts);
-	spin_unlock_bh(&vsock->pkt_list_lock);
+	spin_unlock_bh(&vsock->pkt_queue.lock);
 
 	while ((skb = __skb_dequeue(&pkts))) {
 		virtio_transport_deliver_tap_pkt(skb);
@@ -132,7 +129,6 @@ static int __init vsock_loopback_init(void)
 	if (!vsock->workqueue)
 		return -ENOMEM;
 
-	spin_lock_init(&vsock->pkt_list_lock);
 	skb_queue_head_init(&vsock->pkt_queue);
 	INIT_WORK(&vsock->pkt_work, vsock_loopback_work);
 
@@ -156,9 +152,7 @@ static void __exit vsock_loopback_exit(void)
 
 	flush_work(&vsock->pkt_work);
 
-	spin_lock_bh(&vsock->pkt_list_lock);
 	virtio_vsock_skb_queue_purge(&vsock->pkt_queue);
-	spin_unlock_bh(&vsock->pkt_list_lock);
 
 	destroy_workqueue(vsock->workqueue);
 }
-- 
2.39.2