Skip to contentSkip to conversation

tuntap: purge socket error queue on detach

PATCH Applied Jason Wang 2 participants 4 messages Sep 5, 2013

Subsystems (3)
lore
Share & tools
lore

Jason Wang

author
#0 lore raw

PATCH v2 1/2 tuntap: purge socket error queue on detach

Commit eda297729171fe16bf34fe5b0419dfb69060f623
(tun: Support software transmit time stamping) will queue skbs into error queue when tx stamping is enabled. But it forgets to purge the error queue during detach. This patch fixes this.

---
 drivers/net/tun.c |   12 +++++++++---
 1 files changed, 9 insertions(+), 3 deletions(-)
drivers/net/tun.c +9 -3
index 60a1e93..2dddb1b 100644
@@ -409,6 +409,12 @@ static struct tun_struct *tun_enable_queue(struct tun_file *tfile)
 	return tun;
 }
 
+static void tun_queue_purge(struct tun_file *tfile)
+{
+	skb_queue_purge(&tfile->sk.sk_receive_queue);
+	skb_queue_purge(&tfile->sk.sk_error_queue);
+}
+
 static void __tun_detach(struct tun_file *tfile, bool clean)
 {
 	struct tun_file *ntfile;
@@ -435,7 +441,7 @@ static void __tun_detach(struct tun_file *tfile, bool clean)
 		synchronize_net();
 		tun_flow_delete_by_queue(tun, tun->numqueues + 1);
 		/* Drop read queue */
-		skb_queue_purge(&tfile->sk.sk_receive_queue);
+		tun_queue_purge(tfile);
 		tun_set_real_num_queues(tun);
 	} else if (tfile->detached && clean) {
 		tun = tun_enable_queue(tfile);
@@ -487,12 +493,12 @@ static void tun_detach_all(struct net_device *dev)
 	for (i = 0; i < n; i++) {
 		tfile = rtnl_dereference(tun->tfiles[i]);
 		/* Drop read queue */
-		skb_queue_purge(&tfile->sk.sk_receive_queue);
+		tun_queue_purge(tfile);
 		sock_put(&tfile->sk);
 	}
 	list_for_each_entry_safe(tfile, tmp, &tun->disabled, next) {
 		tun_enable_queue(tfile);
-		skb_queue_purge(&tfile->sk.sk_receive_queue);
+		tun_queue_purge(tfile);
 		sock_put(&tfile->sk);
 	}
 	BUG_ON(tun->numdisabled != 0);
Signature · 1 line
1.7.1
author
#1 lore raw

PATCH v2 2/2 tuntap: orphan frags before trying to set tx timestamp

sock_tx_timestamp() will clear all zerocopy flags of skb which may lead the frags never to be orphaned. This will break guest to guest traffic when zerocopy is enabled. Fix this by orphaning the frags before trying to set tx time stamp.

The issue were introduced by commit eda297729171fe16bf34fe5b0419dfb69060f623 (tun: Support software transmit time stamping).

---
Changes from v1:
- correct the comment location and style

---
 drivers/net/tun.c |    8 +++++---
 1 files changed, 5 insertions(+), 3 deletions(-)
drivers/net/tun.c +5 -3
index 2dddb1b..a639de8 100644
@@ -749,15 +749,17 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
 			  >= dev->tx_queue_len / tun->numqueues)
 		goto drop;
 
+	if (unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
+		goto drop;
+
 	if (skb->sk) {
 		sock_tx_timestamp(skb->sk, &skb_shinfo(skb)->tx_flags);
 		sw_tx_timestamp(skb);
 	}
 
 	/* Orphan the skb - required as we might hang on to it
-	 * for indefinite time. */
-	if (unlikely(skb_orphan_frags(skb, GFP_ATOMIC)))
-		goto drop;
+	 * for indefinite time.
+	 */
 	skb_orphan(skb);
 
 	nf_reset(skb);
Signature · 1 line
1.7.1
maintainer Applied
#2 lore raw

Re: [PATCH V2 net-next 2/2] tuntap: orphan frags before trying to set tx timestamp

From: Jason Wang <[email protected]>
Date: Thu, 5 Sep 2013 17:54:00 +0800

Quoted 11 lines

sock_tx_timestamp() will clear all zerocopy flags of skb which may lead the frags never to be orphaned. This will break guest to guest traffic when zerocopy is enabled. Fix this by orphaning the frags before trying to set tx time stamp.

The issue were introduced by commit eda297729171fe16bf34fe5b0419dfb69060f623 (tun: Support software transmit time stamping).

Applied.

I also queued both of these patches to -stable.

Thanks.

maintainer Applied
#3 lore raw
In reply to Jason Wang

From: Jason Wang <[email protected]>
Date: Thu, 5 Sep 2013 17:53:59 +0800

Quoted 8 lines

Commit eda297729171fe16bf34fe5b0419dfb69060f623
(tun: Support software transmit time stamping) will queue skbs into error queue when tx stamping is enabled. But it forgets to purge the error queue during detach. This patch fixes this.

Applied.