Re: [PATCH net-next V2 3/3] tun: rx batching
From: Stefan Hajnoczi <hidden>
Date: 2017-01-03 13:34:00
Also in:
kvm, lkml, netdev
Possibly related (same subject, not in this thread)
- 2016-12-30 · Re: [PATCH net-next V2 3/3] tun: rx batching · Jason Wang <jasowang@redhat.com>
- 2016-12-29 · Re: [PATCH net-next V2 3/3] tun: rx batching · David Miller <davem@davemloft.net>
- 2016-12-28 · [PATCH net-next V2 3/3] tun: rx batching · Jason Wang <jasowang@redhat.com>
On Wed, Dec 28, 2016 at 04:09:31PM +0800, Jason Wang wrote:
+static int tun_rx_batched(struct tun_file *tfile, struct sk_buff *skb,
+ int more)
+{
+ struct sk_buff_head *queue = &tfile->sk.sk_write_queue;
+ struct sk_buff_head process_queue;
+ int qlen;
+ bool rcv = false;
+
+ spin_lock(&queue->lock);Should this be spin_lock_bh()? Below and in tun_get_user() there are explicit local_bh_disable() calls so I guess BHs can interrupt us here and this would deadlock.