[PATCH net] macvtap: fix network header pointer for VLAN tagged pkts

Subsystems: networking drivers, the rest

STALE4113d

3 messages, 2 authors, 2015-07-21 · open the first message on its own page

[PATCH net] macvtap: fix network header pointer for VLAN tagged pkts

From: Ivan Vecera <ivecera@redhat.com>
Date: 2015-07-20 15:43:35

Network header is set with offset ETH_HLEN but it is not true for VLAN
(multiple-)tagged and results in checksum issues in lower devices.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
---
 drivers/net/macvtap.c | 6 ++++++
 1 file changed, 6 insertions(+)
diff --git a/drivers/net/macvtap.c b/drivers/net/macvtap.c
index 3b933bb..cdcbab4 100644
--- a/drivers/net/macvtap.c
+++ b/drivers/net/macvtap.c
@@ -796,6 +796,12 @@ static ssize_t macvtap_get_user(struct macvtap_queue *q, struct msghdr *m,
 	skb_reset_mac_header(skb);
 	skb->protocol = eth_hdr(skb)->h_proto;
 
+	if (skb_vlan_tagged(skb)) {
+		int depth;
+		skb->protocol = __vlan_get_protocol(skb, skb->protocol, &depth);
+		skb_set_network_header(skb, depth);
+	}
+
 	if (vnet_hdr_len) {
 		err = macvtap_skb_from_vnet_hdr(q, skb, &vnet_hdr);
 		if (err)
-- 
2.3.6

Re: [PATCH net] macvtap: fix network header pointer for VLAN tagged pkts

From: Vlad Yasevich <hidden>
Date: 2015-07-20 16:42:45

On 07/20/2015 11:44 AM, Ivan Vecera wrote:
quoted hunk
Network header is set with offset ETH_HLEN but it is not true for VLAN
(multiple-)tagged and results in checksum issues in lower devices.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
---
 drivers/net/macvtap.c | 6 ++++++
 1 file changed, 6 insertions(+)
diff --git a/drivers/net/macvtap.c b/drivers/net/macvtap.c
index 3b933bb..cdcbab4 100644
--- a/drivers/net/macvtap.c
+++ b/drivers/net/macvtap.c
@@ -796,6 +796,12 @@ static ssize_t macvtap_get_user(struct macvtap_queue *q, struct msghdr *m,
 	skb_reset_mac_header(skb);
 	skb->protocol = eth_hdr(skb)->h_proto;
 
+	if (skb_vlan_tagged(skb)) {
+		int depth;
+		skb->protocol = __vlan_get_protocol(skb, skb->protocol, &depth);
I don't think this is right.  This would reset the protocol to the encapsulated
protocol which isn't really the case since you are not really stripping vlan
encapsulations.

-vlad
quoted hunk
+		skb_set_network_header(skb, depth);
+	}
+
 	if (vnet_hdr_len) {
 		err = macvtap_skb_from_vnet_hdr(q, skb, &vnet_hdr);
 		if (err)

Re: [PATCH net] macvtap: fix network header pointer for VLAN tagged pkts

From: Ivan Vecera <ivecera@redhat.com>
Date: 2015-07-21 07:10:55

On 07/20/2015 06:42 PM, Vlad Yasevich wrote:
On 07/20/2015 11:44 AM, Ivan Vecera wrote:
quoted
Network header is set with offset ETH_HLEN but it is not true for VLAN
(multiple-)tagged and results in checksum issues in lower devices.

Signed-off-by: Ivan Vecera <ivecera@redhat.com>
---
  drivers/net/macvtap.c | 6 ++++++
  1 file changed, 6 insertions(+)
diff --git a/drivers/net/macvtap.c b/drivers/net/macvtap.c
index 3b933bb..cdcbab4 100644
--- a/drivers/net/macvtap.c
+++ b/drivers/net/macvtap.c
@@ -796,6 +796,12 @@ static ssize_t macvtap_get_user(struct macvtap_queue *q, struct msghdr *m,
  	skb_reset_mac_header(skb);
  	skb->protocol = eth_hdr(skb)->h_proto;

+	if (skb_vlan_tagged(skb)) {
+		int depth;
+		skb->protocol = __vlan_get_protocol(skb, skb->protocol, &depth);
I don't think this is right.  This would reset the protocol to the encapsulated
protocol which isn't really the case since you are not really stripping vlan
encapsulations.

-vlad
Yup, you are right, skb->protocol should be untouched. Will post v2.

Ivan
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help