[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20090603052335.GA20823@sequoia.sous-sol.org>
Date: Tue, 2 Jun 2009 22:23:35 -0700
From: Chris Wright <chrisw@...s-sol.org>
To: David Miller <davem@...emloft.net>
Cc: netdev@...r.kernel.org, Herbert Xu <herbert@...dor.apana.org.au>,
Rusty Russell <rusty@...tcorp.com.au>
Subject: [PATCH] tun: virtio net is not compatible with LRO
The current failure mode for an LRO device bridged to a virtio device
(common for KVM) is BUG(). Rather than crashing the host, we can at
least limp along and give a warning. lro_flush() will set gso_size,
but no corresponding gso_type, unlike gro which will eventually call
->gro_complete and set up proper gso_type.
Signed-off-by: Chris Wright <chrisw@...s-sol.org>
---
drivers/net/tun.c | 11 +++++++++--
1 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index 735bf41..17c4516 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -717,8 +717,15 @@ static __inline__ ssize_t tun_put_user(struct tun_struct *tun,
gso.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
else if (sinfo->gso_type & SKB_GSO_TCPV6)
gso.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
- else
- BUG();
+ else {
+ WARN_ONCE(1, "%s: hdr_len %d gso_size %d "
+ "gso_type %x. virtio is not "
+ "compatible with LRO, verify NIC "
+ "isn't using LRO", __func__,
+ gso.hdr_len, gso.gso_size,
+ sinfo->gso_type);
+ return -EINVAL;
+ }
if (sinfo->gso_type & SKB_GSO_TCP_ECN)
gso.gso_type |= VIRTIO_NET_HDR_GSO_ECN;
} else
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists