* [PATCH net-next 0/5] ip6: Transmit tunneling fixes
@ 2016-05-10 0:12 Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 1/5] ip6_gre: Fix MTU setting Tom Herbert
` (5 more replies)
0 siblings, 6 replies; 8+ messages in thread
From: Tom Herbert @ 2016-05-10 0:12 UTC (permalink / raw)
To: davem, netdev, alexander.duyck; +Cc: kernel-team
Several fixes suggested by Alexander.
Tested: Running netperf TCP_STREAM with gretap and keyid configured.
Visually verified that MTU is correctly being set. Did not test HW
offload (Alexander plese try)
Tom Herbert (5):
ip6_gre: Fix MTU setting
gre6: Fix flag translations
ip6_gre: Set inner protocol correctly in __gre6_xmit
ip6: Don't set transport header in IPv6 tunneling
ip6_gre: Use correct flags for reading TUNNEL_SEQ
net/ipv6/ip6_gre.c | 55 ++++++++++++++++++++++++++-------------------------
net/ipv6/ip6_tunnel.c | 2 --
2 files changed, 28 insertions(+), 29 deletions(-)
--
2.8.0.rc2
^ permalink raw reply [flat|nested] 8+ messages in thread
* [PATCH net-next 1/5] ip6_gre: Fix MTU setting
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
@ 2016-05-10 0:12 ` Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 2/5] gre6: Fix flag translations Tom Herbert
` (4 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Tom Herbert @ 2016-05-10 0:12 UTC (permalink / raw)
To: davem, netdev, alexander.duyck; +Cc: kernel-team
In ip6gre_tnl_link_config set t->tun_len and t->hlen correctly for the
configuration. For hard_header_len and mtu calculation include
IPv6 header and encapsulation overhead.
In ip6gre_tunnel_init_common set t->tun_len and t->hlen correctly for
the configuration. Revert to setting hard_header_len instead of
needed_headroom.
Tested:
./ip link add name tun8 type ip6gretap remote \
2401:db00:20:911a:face:0:27:0 local \
2401:db00:20:911a:face:0:25:0 ttl 225
Gives MTU of 1434. That is equal to 1500 - 40 - 14 - 4 - 8.
./ip link add name tun8 type ip6gretap remote \
2401:db00:20:911a:face:0:27:0 local \
2401:db00:20:911a:face:0:25:0 ttl 225 okey 123
Gives MTU of 1430. That is equal to 1500 - 40 - 14 - 4 - 8 - 4.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
net/ipv6/ip6_gre.c | 29 +++++++++++++----------------
1 file changed, 13 insertions(+), 16 deletions(-)
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index 47b671a..6d0aa94 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -700,7 +700,7 @@ static void ip6gre_tnl_link_config(struct ip6_tnl *t, int set_mtu)
struct net_device *dev = t->dev;
struct __ip6_tnl_parm *p = &t->parms;
struct flowi6 *fl6 = &t->fl.u.ip6;
- int addend = sizeof(struct ipv6hdr) + 4;
+ int t_hlen;
if (dev->type != ARPHRD_ETHER) {
memcpy(dev->dev_addr, &p->laddr, sizeof(struct in6_addr));
@@ -727,16 +727,11 @@ static void ip6gre_tnl_link_config(struct ip6_tnl *t, int set_mtu)
else
dev->flags &= ~IFF_POINTOPOINT;
- /* Precalculate GRE options length */
- if (t->parms.o_flags&(GRE_CSUM|GRE_KEY|GRE_SEQ)) {
- if (t->parms.o_flags&GRE_CSUM)
- addend += 4;
- if (t->parms.o_flags&GRE_KEY)
- addend += 4;
- if (t->parms.o_flags&GRE_SEQ)
- addend += 4;
- }
- t->hlen = addend;
+ t->tun_hlen = gre_calc_hlen(t->parms.o_flags);
+
+ t->hlen = t->tun_hlen;
+
+ t_hlen = t->hlen + sizeof(struct ipv6hdr);
if (p->flags & IP6_TNL_F_CAP_XMIT) {
int strict = (ipv6_addr_type(&p->raddr) &
@@ -750,10 +745,11 @@ static void ip6gre_tnl_link_config(struct ip6_tnl *t, int set_mtu)
return;
if (rt->dst.dev) {
- dev->hard_header_len = rt->dst.dev->hard_header_len + addend;
+ dev->hard_header_len = rt->dst.dev->hard_header_len +
+ t_hlen;
if (set_mtu) {
- dev->mtu = rt->dst.dev->mtu - addend;
+ dev->mtu = rt->dst.dev->mtu - t_hlen;
if (!(t->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
dev->mtu -= 8;
if (dev->type == ARPHRD_ETHER)
@@ -1027,11 +1023,12 @@ static int ip6gre_tunnel_init_common(struct net_device *dev)
tunnel->tun_hlen = gre_calc_hlen(tunnel->parms.o_flags);
- t_hlen = tunnel->hlen + sizeof(struct ipv6hdr);
+ tunnel->hlen = tunnel->tun_hlen;
- dev->needed_headroom = LL_MAX_HEADER + t_hlen + 4;
- dev->mtu = ETH_DATA_LEN - t_hlen - 4;
+ t_hlen = tunnel->hlen + sizeof(struct ipv6hdr);
+ dev->hard_header_len = LL_MAX_HEADER + t_hlen;
+ dev->mtu = ETH_DATA_LEN - t_hlen;
if (!(tunnel->parms.flags & IP6_TNL_F_IGN_ENCAP_LIMIT))
dev->mtu -= 8;
--
2.8.0.rc2
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH net-next 2/5] gre6: Fix flag translations
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 1/5] ip6_gre: Fix MTU setting Tom Herbert
@ 2016-05-10 0:12 ` Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 3/5] ip6_gre: Set inner protocol correctly in __gre6_xmit Tom Herbert
` (3 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Tom Herbert @ 2016-05-10 0:12 UTC (permalink / raw)
To: davem, netdev, alexander.duyck; +Cc: kernel-team
GRE for IPv6 does not properly translate for GRE flags to tunnel
flags and vice versa. This patch fixes that.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
net/ipv6/ip6_gre.c | 20 ++++++++++++--------
1 file changed, 12 insertions(+), 8 deletions(-)
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index 6d0aa94..509fb92 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -795,8 +795,8 @@ static void ip6gre_tnl_parm_from_user(struct __ip6_tnl_parm *p,
p->link = u->link;
p->i_key = u->i_key;
p->o_key = u->o_key;
- p->i_flags = u->i_flags;
- p->o_flags = u->o_flags;
+ p->i_flags = gre_flags_to_tnl_flags(u->i_flags);
+ p->o_flags = gre_flags_to_tnl_flags(u->o_flags);
memcpy(p->name, u->name, sizeof(u->name));
}
@@ -813,8 +813,8 @@ static void ip6gre_tnl_parm_to_user(struct ip6_tnl_parm2 *u,
u->link = p->link;
u->i_key = p->i_key;
u->o_key = p->o_key;
- u->i_flags = p->i_flags;
- u->o_flags = p->o_flags;
+ u->i_flags = gre_tnl_flags_to_gre_flags(p->i_flags);
+ u->o_flags = gre_tnl_flags_to_gre_flags(p->o_flags);
memcpy(u->name, p->name, sizeof(u->name));
}
@@ -1214,10 +1214,12 @@ static void ip6gre_netlink_parms(struct nlattr *data[],
parms->link = nla_get_u32(data[IFLA_GRE_LINK]);
if (data[IFLA_GRE_IFLAGS])
- parms->i_flags = nla_get_be16(data[IFLA_GRE_IFLAGS]);
+ parms->i_flags = gre_flags_to_tnl_flags(
+ nla_get_be16(data[IFLA_GRE_IFLAGS]));
if (data[IFLA_GRE_OFLAGS])
- parms->o_flags = nla_get_be16(data[IFLA_GRE_OFLAGS]);
+ parms->o_flags = gre_flags_to_tnl_flags(
+ nla_get_be16(data[IFLA_GRE_OFLAGS]));
if (data[IFLA_GRE_IKEY])
parms->i_key = nla_get_be32(data[IFLA_GRE_IKEY]);
@@ -1409,8 +1411,10 @@ static int ip6gre_fill_info(struct sk_buff *skb, const struct net_device *dev)
struct __ip6_tnl_parm *p = &t->parms;
if (nla_put_u32(skb, IFLA_GRE_LINK, p->link) ||
- nla_put_be16(skb, IFLA_GRE_IFLAGS, p->i_flags) ||
- nla_put_be16(skb, IFLA_GRE_OFLAGS, p->o_flags) ||
+ nla_put_be16(skb, IFLA_GRE_IFLAGS,
+ gre_tnl_flags_to_gre_flags(p->i_flags)) ||
+ nla_put_be16(skb, IFLA_GRE_OFLAGS,
+ gre_tnl_flags_to_gre_flags(p->o_flags)) ||
nla_put_be32(skb, IFLA_GRE_IKEY, p->i_key) ||
nla_put_be32(skb, IFLA_GRE_OKEY, p->o_key) ||
nla_put_in6_addr(skb, IFLA_GRE_LOCAL, &p->laddr) ||
--
2.8.0.rc2
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH net-next 3/5] ip6_gre: Set inner protocol correctly in __gre6_xmit
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 1/5] ip6_gre: Fix MTU setting Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 2/5] gre6: Fix flag translations Tom Herbert
@ 2016-05-10 0:12 ` Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 4/5] ip6: Don't set transport header in IPv6 tunneling Tom Herbert
` (2 subsequent siblings)
5 siblings, 0 replies; 8+ messages in thread
From: Tom Herbert @ 2016-05-10 0:12 UTC (permalink / raw)
To: davem, netdev, alexander.duyck; +Cc: kernel-team
Need to use adjusted protocol value for setting inner protocol.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
net/ipv6/ip6_gre.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index 509fb92..ec209f4 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -519,7 +519,7 @@ static netdev_tx_t __gre6_xmit(struct sk_buff *skb,
gre_build_header(skb, tunnel->tun_hlen, tunnel->parms.o_flags,
protocol, tunnel->parms.o_key, htonl(tunnel->o_seqno));
- skb_set_inner_protocol(skb, proto);
+ skb_set_inner_protocol(skb, protocol);
return ip6_tnl_xmit(skb, dev, dsfield, fl6, encap_limit, pmtu,
NEXTHDR_GRE);
--
2.8.0.rc2
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH net-next 4/5] ip6: Don't set transport header in IPv6 tunneling
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
` (2 preceding siblings ...)
2016-05-10 0:12 ` [PATCH net-next 3/5] ip6_gre: Set inner protocol correctly in __gre6_xmit Tom Herbert
@ 2016-05-10 0:12 ` Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 5/5] ip6_gre: Use correct flags for reading TUNNEL_SEQ Tom Herbert
2016-05-10 3:30 ` [PATCH net-next 0/5] ip6: Transmit tunneling fixes Alexander Duyck
5 siblings, 0 replies; 8+ messages in thread
From: Tom Herbert @ 2016-05-10 0:12 UTC (permalink / raw)
To: davem, netdev, alexander.duyck; +Cc: kernel-team
We only need to reset network header here.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
net/ipv6/ip6_tunnel.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index ade55af..50af706 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -1114,8 +1114,6 @@ int ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev, __u8 dsfield,
dst_cache_set_ip6(&t->dst_cache, ndst, &fl6->saddr);
skb_dst_set(skb, dst);
- skb->transport_header = skb->network_header;
-
if (encap_limit >= 0) {
init_tel_txopt(&opt, encap_limit);
ipv6_push_nfrag_opts(skb, &opt.ops, &proto, NULL);
--
2.8.0.rc2
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH net-next 5/5] ip6_gre: Use correct flags for reading TUNNEL_SEQ
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
` (3 preceding siblings ...)
2016-05-10 0:12 ` [PATCH net-next 4/5] ip6: Don't set transport header in IPv6 tunneling Tom Herbert
@ 2016-05-10 0:12 ` Tom Herbert
2016-05-10 3:30 ` [PATCH net-next 0/5] ip6: Transmit tunneling fixes Alexander Duyck
5 siblings, 0 replies; 8+ messages in thread
From: Tom Herbert @ 2016-05-10 0:12 UTC (permalink / raw)
To: davem, netdev, alexander.duyck; +Cc: kernel-team
Fix two spots where o_flags in a tunnel are being compared to GRE_SEQ
instead of TUNNEL_SEQ.
Signed-off-by: Tom Herbert <tom@herbertland.com>
---
net/ipv6/ip6_gre.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index ec209f4..ee62ec4 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -343,7 +343,7 @@ static struct ip6_tnl *ip6gre_tunnel_locate(struct net *net,
goto failed_free;
/* Can use a lockless transmit, unless we generate output sequences */
- if (!(nt->parms.o_flags & GRE_SEQ))
+ if (!(nt->parms.o_flags & TUNNEL_SEQ))
dev->features |= NETIF_F_LLTX;
dev_hold(dev);
@@ -1314,7 +1314,7 @@ static int ip6gre_newlink(struct net *src_net, struct net_device *dev,
dev->features |= GRE6_FEATURES;
dev->hw_features |= GRE6_FEATURES;
- if (!(nt->parms.o_flags & GRE_SEQ)) {
+ if (!(nt->parms.o_flags & TUNNEL_SEQ)) {
/* TCP segmentation offload is not supported when we
* generate output sequences.
*/
--
2.8.0.rc2
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 0/5] ip6: Transmit tunneling fixes
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
` (4 preceding siblings ...)
2016-05-10 0:12 ` [PATCH net-next 5/5] ip6_gre: Use correct flags for reading TUNNEL_SEQ Tom Herbert
@ 2016-05-10 3:30 ` Alexander Duyck
2016-05-10 4:39 ` David Miller
5 siblings, 1 reply; 8+ messages in thread
From: Alexander Duyck @ 2016-05-10 3:30 UTC (permalink / raw)
To: Tom Herbert; +Cc: David Miller, Netdev, Kernel Team
On Mon, May 9, 2016 at 5:12 PM, Tom Herbert <tom@herbertland.com> wrote:
> Several fixes suggested by Alexander.
>
> Tested: Running netperf TCP_STREAM with gretap and keyid configured.
> Visually verified that MTU is correctly being set. Did not test HW
> offload (Alexander plese try)
This appears to have fixed the issues I saw. I can confirm the MTU is
the same value that it was before the regression and I am able to
send/receive at around 17Gb/s.
Tested-by: Alexander Duyck <aduyck@mirantis.com>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH net-next 0/5] ip6: Transmit tunneling fixes
2016-05-10 3:30 ` [PATCH net-next 0/5] ip6: Transmit tunneling fixes Alexander Duyck
@ 2016-05-10 4:39 ` David Miller
0 siblings, 0 replies; 8+ messages in thread
From: David Miller @ 2016-05-10 4:39 UTC (permalink / raw)
To: alexander.duyck; +Cc: tom, netdev, kernel-team
From: Alexander Duyck <alexander.duyck@gmail.com>
Date: Mon, 9 May 2016 20:30:46 -0700
> On Mon, May 9, 2016 at 5:12 PM, Tom Herbert <tom@herbertland.com> wrote:
>> Several fixes suggested by Alexander.
>>
>> Tested: Running netperf TCP_STREAM with gretap and keyid configured.
>> Visually verified that MTU is correctly being set. Did not test HW
>> offload (Alexander plese try)
>
> This appears to have fixed the issues I saw. I can confirm the MTU is
> the same value that it was before the regression and I am able to
> send/receive at around 17Gb/s.
>
> Tested-by: Alexander Duyck <aduyck@mirantis.com>
Series applied, thanks everyone.
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2016-05-10 4:39 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-05-10 0:12 [PATCH net-next 0/5] ip6: Transmit tunneling fixes Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 1/5] ip6_gre: Fix MTU setting Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 2/5] gre6: Fix flag translations Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 3/5] ip6_gre: Set inner protocol correctly in __gre6_xmit Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 4/5] ip6: Don't set transport header in IPv6 tunneling Tom Herbert
2016-05-10 0:12 ` [PATCH net-next 5/5] ip6_gre: Use correct flags for reading TUNNEL_SEQ Tom Herbert
2016-05-10 3:30 ` [PATCH net-next 0/5] ip6: Transmit tunneling fixes Alexander Duyck
2016-05-10 4:39 ` David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).