* [PATCH net-next 1/1] net: sched: act_csum: Fix csum calc for tagged packets
@ 2019-02-26 9:57 Eli Britstein
2019-02-28 5:50 ` David Miller
0 siblings, 1 reply; 2+ messages in thread
From: Eli Britstein @ 2019-02-26 9:57 UTC (permalink / raw)
To: netdev
Cc: Jiri Pirko, Jamal Hadi Salim, Cong Wang, Roi Dayan, Maor Dickman,
Eli Britstein
The csum calculation is different for IPv4/6. For VLAN packets,
tc_skb_protocol returns the VLAN protocol rather than the packet's one
(e.g. IPv4/6), so csum is not calculated. Furthermore, VLAN may not be
stripped so csum is not calculated in this case too. Calculate the
csum for those cases.
Fixes: d8b9605d2697 ("net: sched: fix skb->protocol use in case of accelerated vlan path")
Signed-off-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
net/sched/act_csum.c | 31 +++++++++++++++++++++++++++++--
1 file changed, 29 insertions(+), 2 deletions(-)
diff --git a/net/sched/act_csum.c b/net/sched/act_csum.c
index 945fb34ae721..c79aca29505e 100644
--- a/net/sched/act_csum.c
+++ b/net/sched/act_csum.c
@@ -559,8 +559,11 @@ static int tcf_csum_act(struct sk_buff *skb, const struct tc_action *a,
struct tcf_result *res)
{
struct tcf_csum *p = to_tcf_csum(a);
+ bool orig_vlan_tag_present = false;
+ unsigned int vlan_hdr_count = 0;
struct tcf_csum_params *params;
u32 update_flags;
+ __be16 protocol;
int action;
params = rcu_dereference_bh(p->params);
@@ -573,7 +576,9 @@ static int tcf_csum_act(struct sk_buff *skb, const struct tc_action *a,
goto drop;
update_flags = params->update_flags;
- switch (tc_skb_protocol(skb)) {
+ protocol = tc_skb_protocol(skb);
+again:
+ switch (protocol) {
case cpu_to_be16(ETH_P_IP):
if (!tcf_csum_ipv4(skb, update_flags))
goto drop;
@@ -582,13 +587,35 @@ static int tcf_csum_act(struct sk_buff *skb, const struct tc_action *a,
if (!tcf_csum_ipv6(skb, update_flags))
goto drop;
break;
+ case cpu_to_be16(ETH_P_8021AD): /* fall through */
+ case cpu_to_be16(ETH_P_8021Q):
+ if (skb_vlan_tag_present(skb) && !orig_vlan_tag_present) {
+ protocol = skb->protocol;
+ orig_vlan_tag_present = true;
+ } else {
+ struct vlan_hdr *vlan = (struct vlan_hdr *)skb->data;
+
+ protocol = vlan->h_vlan_encapsulated_proto;
+ skb_pull(skb, VLAN_HLEN);
+ skb_reset_network_header(skb);
+ vlan_hdr_count++;
+ }
+ goto again;
+ }
+
+out:
+ /* Restore the skb for the pulled VLAN tags */
+ while (vlan_hdr_count--) {
+ skb_push(skb, VLAN_HLEN);
+ skb_reset_network_header(skb);
}
return action;
drop:
qstats_drop_inc(this_cpu_ptr(p->common.cpu_qstats));
- return TC_ACT_SHOT;
+ action = TC_ACT_SHOT;
+ goto out;
}
static int tcf_csum_dump(struct sk_buff *skb, struct tc_action *a, int bind,
--
2.17.2
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH net-next 1/1] net: sched: act_csum: Fix csum calc for tagged packets
2019-02-26 9:57 [PATCH net-next 1/1] net: sched: act_csum: Fix csum calc for tagged packets Eli Britstein
@ 2019-02-28 5:50 ` David Miller
0 siblings, 0 replies; 2+ messages in thread
From: David Miller @ 2019-02-28 5:50 UTC (permalink / raw)
To: elibr; +Cc: netdev, jiri, jhs, xiyou.wangcong, roid, maord
From: Eli Britstein <elibr@mellanox.com>
Date: Tue, 26 Feb 2019 09:57:34 +0000
> The csum calculation is different for IPv4/6. For VLAN packets,
> tc_skb_protocol returns the VLAN protocol rather than the packet's one
> (e.g. IPv4/6), so csum is not calculated. Furthermore, VLAN may not be
> stripped so csum is not calculated in this case too. Calculate the
> csum for those cases.
>
> Fixes: d8b9605d2697 ("net: sched: fix skb->protocol use in case of accelerated vlan path")
> Signed-off-by: Eli Britstein <elibr@mellanox.com>
> Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Applied.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2019-02-28 5:50 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-02-26 9:57 [PATCH net-next 1/1] net: sched: act_csum: Fix csum calc for tagged packets Eli Britstein
2019-02-28 5:50 ` David Miller
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.