Re: [PATCH bpf-next] xsk: proper socket state check in xsk_poll

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 8/20/19 12:04 PM, Björn Töpel wrote:
From: Björn Töpel <bjorn.topel@xxxxxxxxx>

The poll() implementation for AF_XDP sockets did not perform the
proper state checks, prior accessing the socket umem. This patch fixes
that by performing a xsk_is_bound() check.

Suggested-by: Hillf Danton <hdanton@xxxxxxxx>
Reported-by: syzbot+c82697e3043781e08802@xxxxxxxxxxxxxxxxxxxxxxxxx
Fixes: 77cd0d7b3f25 ("xsk: add support for need_wakeup flag in AF_XDP rings")
Signed-off-by: Björn Töpel <bjorn.topel@xxxxxxxxx>
---
  net/xdp/xsk.c | 14 ++++++++++++--
  1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/net/xdp/xsk.c b/net/xdp/xsk.c
index ee4428a892fa..08bed5e92af4 100644
--- a/net/xdp/xsk.c
+++ b/net/xdp/xsk.c
@@ -356,13 +356,20 @@ static int xsk_generic_xmit(struct sock *sk, struct msghdr *m,
  	return err;
  }
+static bool xsk_is_bound(struct xdp_sock *xs)
+{
+	struct net_device *dev = READ_ONCE(xs->dev);
+
+	return dev && xs->state == XSK_BOUND;
+}
+
  static int xsk_sendmsg(struct socket *sock, struct msghdr *m, size_t total_len)
  {
  	bool need_wait = !(m->msg_flags & MSG_DONTWAIT);
  	struct sock *sk = sock->sk;
  	struct xdp_sock *xs = xdp_sk(sk);
- if (unlikely(!xs->dev))
+	if (unlikely(!xsk_is_bound(xs)))
  		return -ENXIO;
  	if (unlikely(!(xs->dev->flags & IFF_UP)))
  		return -ENETDOWN;
@@ -383,6 +390,9 @@ static unsigned int xsk_poll(struct file *file, struct socket *sock,
  	struct net_device *dev = xs->dev;
  	struct xdp_umem *umem = xs->umem;
+ if (unlikely(!xsk_is_bound(xs)))
+		return mask;
+
  	if (umem->need_wakeup)
  		dev->netdev_ops->ndo_xsk_wakeup(dev, xs->queue_id,
  						umem->need_wakeup);
@@ -417,7 +427,7 @@ static void xsk_unbind_dev(struct xdp_sock *xs)
  {
  	struct net_device *dev = xs->dev;
- if (!dev || xs->state != XSK_BOUND)
+	if (!xsk_is_bound(xs))
  		return;

I think I'm a bit confused by your READ_ONCE() usage. ;-/ I can see why you're
using it in xsk_is_bound() above, but then at the same time all the other callbacks
like xsk_poll() or xsk_unbind_dev() above have a struct net_device *dev = xs->dev
right before the test. Could you elaborate?

Thanks,
Daniel



[Index of Archives]     [Linux Networking Development]     [Fedora Linux Users]     [Linux SCTP]     [DCCP]     [Gimp]     [Yosemite Campsites]

  Powered by Linux