[PATCH net V3 1/2] tuntap: fix dividing by zero in ebpf queue selection

[PATCH net V3 1/2] tuntap: fix dividing by zero in ebpf queue selection

[Jason Wang]

We need check if tun->numqueues is zero (e.g for the persist device)
before trying to use it for modular arithmetic.

Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Fixes: 96f84061620c6("tun: add eBPF based queue selection method")
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
 drivers/net/tun.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index e9ca1c0..dc62fc3 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -596,13 +596,18 @@ static u16 tun_automq_select_queue(struct tun_struct *tun, struct sk_buff *skb)
 static u16 tun_ebpf_select_queue(struct tun_struct *tun, struct sk_buff *skb)
 {
 	struct tun_prog *prog;
+	u32 numqueues;
 	u16 ret = 0;
 
+	numqueues = READ_ONCE(tun->numqueues);
+	if (!numqueues)
+		return 0;
+
 	prog = rcu_dereference(tun->steering_prog);
 	if (prog)
 		ret = bpf_prog_run_clear_cb(prog->prog, skb);
 
-	return ret % tun->numqueues;
+	return ret % numqueues;
 }
 
 static u16 tun_select_queue(struct net_device *dev, struct sk_buff *skb,
-- 
1.8.3.1

[PATCH net V3 2/2] tuntap: synchronize through tfiles array instead of tun->numqueues

[Jason Wang]

When a queue(tfile) is detached through __tun_detach(), we move the
last enabled tfile to the position where detached one sit but don't
NULL out last position. We expect to synchronize the datapath through
tun->numqueues. Unfortunately, this won't work since we're lacking
sufficient mechanism to order or synchronize the access to
tun->numqueues.

To fix this, NULL out the last position during detaching and check
RCU protected tfile against NULL instead of checking tun->numqueues in
datapath.

Cc: YueHaibing <yuehaibing@huawei.com>
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: weiyongjun (A) <weiyongjun1@huawei.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Fixes: c8d68e6be1c3b ("tuntap: multiqueue support")
Signed-off-by: Jason Wang <jasowang@redhat.com>
---
Changes from V2:
- resample during detach in tun_xdp_xmit()
Changes from V1:
- keep the check in tun_xdp_xmit()
---
 drivers/net/tun.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index dc62fc3..f4c933a 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -705,6 +705,8 @@ static void __tun_detach(struct tun_file *tfile, bool clean)
 				   tun->tfiles[tun->numqueues - 1]);
 		ntfile = rtnl_dereference(tun->tfiles[index]);
 		ntfile->queue_index = index;
+		rcu_assign_pointer(tun->tfiles[tun->numqueues - 1],
+				   NULL);
 
 		--tun->numqueues;
 		if (clean) {
@@ -1087,7 +1089,7 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
 	tfile = rcu_dereference(tun->tfiles[txq]);
 
 	/* Drop packet if interface is not attached */
-	if (txq >= tun->numqueues)
+	if (!tfile)
 		goto drop;
 
 	if (!rcu_dereference(tun->steering_prog))
@@ -1310,6 +1312,7 @@ static int tun_xdp_xmit(struct net_device *dev, int n,
 
 	rcu_read_lock();
 
+resample:
 	numqueues = READ_ONCE(tun->numqueues);
 	if (!numqueues) {
 		rcu_read_unlock();
@@ -1318,6 +1321,8 @@ static int tun_xdp_xmit(struct net_device *dev, int n,
 
 	tfile = rcu_dereference(tun->tfiles[smp_processor_id() %
 					    numqueues]);
+	if (unlikely(!tfile))
+		goto resample;
 
 	spin_lock(&tfile->tx_ring.producer_lock);
 	for (i = 0; i < n; i++) {
-- 
1.8.3.1

Re: [PATCH net V3 1/2] tuntap: fix dividing by zero in ebpf queue selection

[Eric Dumazet]

On 5/8/19 11:20 PM, Jason Wang wrote:
> We need check if tun->numqueues is zero (e.g for the persist device)
> before trying to use it for modular arithmetic.
> 
> Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
> Fixes: 96f84061620c6("tun: add eBPF based queue selection method")
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---

Reviewed-by: Eric Dumazet <edumazet@google.com>

Thanks.

RE: [PATCH net V3 2/2] tuntap: synchronize through tfiles array instead of tun->numqueues

[weiyongjun (A)]

> -----Original Message-----
> From: Jason Wang [mailto:jasowang@redhat.com]
> Sent: Thursday, May 09, 2019 11:20 AM
> To: netdev@vger.kernel.org; linux-kernel@vger.kernel.org
> Cc: yuehaibing <yuehaibing@huawei.com>; xiyou.wangcong@gmail.com;
> weiyongjun (A) <weiyongjun1@huawei.com>; eric.dumazet@gmail.com;
> Jason Wang <jasowang@redhat.com>
> Subject: [PATCH net V3 2/2] tuntap: synchronize through tfiles array instead
> of tun->numqueues
> 
> When a queue(tfile) is detached through __tun_detach(), we move the
> last enabled tfile to the position where detached one sit but don't
> NULL out last position. We expect to synchronize the datapath through
> tun->numqueues. Unfortunately, this won't work since we're lacking
> sufficient mechanism to order or synchronize the access to
> tun->numqueues.
> 
> To fix this, NULL out the last position during detaching and check
> RCU protected tfile against NULL instead of checking tun->numqueues in
> datapath.
> 
> Cc: YueHaibing <yuehaibing@huawei.com>
> Cc: Cong Wang <xiyou.wangcong@gmail.com>
> Cc: weiyongjun (A) <weiyongjun1@huawei.com>
> Cc: Eric Dumazet <eric.dumazet@gmail.com>
> Fixes: c8d68e6be1c3b ("tuntap: multiqueue support")
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
> Changes from V2:
> - resample during detach in tun_xdp_xmit()
> Changes from V1:
> - keep the check in tun_xdp_xmit()
> ---

Reviewed-by: Wei Yongjun <weiyongjun1@huawei.com>

Thanks

Re: [PATCH net V3 1/2] tuntap: fix dividing by zero in ebpf queue selection

[David Miller]

From: Jason Wang <jasowang@redhat.com>
Date: Wed,  8 May 2019 23:20:17 -0400

> We need check if tun->numqueues is zero (e.g for the persist device)
> before trying to use it for modular arithmetic.
> 
> Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
> Fixes: 96f84061620c6("tun: add eBPF based queue selection method")
> Signed-off-by: Jason Wang <jasowang@redhat.com>

Applied and queued up for -stable.

Re: [PATCH net V3 2/2] tuntap: synchronize through tfiles array instead of tun->numqueues

[David Miller]

From: Jason Wang <jasowang@redhat.com>
Date: Wed,  8 May 2019 23:20:18 -0400

> When a queue(tfile) is detached through __tun_detach(), we move the
> last enabled tfile to the position where detached one sit but don't
> NULL out last position. We expect to synchronize the datapath through
> tun->numqueues. Unfortunately, this won't work since we're lacking
> sufficient mechanism to order or synchronize the access to
> tun->numqueues.
> 
> To fix this, NULL out the last position during detaching and check
> RCU protected tfile against NULL instead of checking tun->numqueues in
> datapath.
> 
> Cc: YueHaibing <yuehaibing@huawei.com>
> Cc: Cong Wang <xiyou.wangcong@gmail.com>
> Cc: weiyongjun (A) <weiyongjun1@huawei.com>
> Cc: Eric Dumazet <eric.dumazet@gmail.com>
> Fixes: c8d68e6be1c3b ("tuntap: multiqueue support")
> Signed-off-by: Jason Wang <jasowang@redhat.com>
> ---
> Changes from V2:
> - resample during detach in tun_xdp_xmit()
> Changes from V1:
> - keep the check in tun_xdp_xmit()

Applied and queued up for -stable.