From f870c1ff65a6d1f3a083f277280802ee09a5b44d Mon Sep 17 00:00:00 2001 From: Alexey Kodanev Date: Thu, 14 Dec 2017 20:20:00 +0300 Subject: vxlan: restore dev->mtu setting based on lower device Stefano Brivio says: Commit a985343ba906 ("vxlan: refactor verification and application of configuration") introduced a change in the behaviour of initial MTU setting: earlier, the MTU for a link created on top of a given lower device, without an initial MTU specification, was set to the MTU of the lower device minus headroom as a result of this path in vxlan_dev_configure(): if (!conf->mtu) dev->mtu = lowerdev->mtu - (use_ipv6 ? VXLAN6_HEADROOM : VXLAN_HEADROOM); which is now gone. Now, the initial MTU, in absence of a configured value, is simply set by ether_setup() to ETH_DATA_LEN (1500 bytes). This breaks userspace expectations in case the MTU of the lower device is higher than 1500 bytes minus headroom. This patch restores the previous behaviour on newlink operation. Since max_mtu can be negative and we update dev->mtu directly, also check it for valid minimum. Reported-by: Junhan Yan Fixes: a985343ba906 ("vxlan: refactor verification and application of configuration") Signed-off-by: Alexey Kodanev Acked-by: Stefano Brivio Signed-off-by: Stefano Brivio Signed-off-by: David S. Miller --- drivers/net/vxlan.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'drivers/net/vxlan.c') diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c index 19b9cc51079e..1000b0e4ee01 100644 --- a/drivers/net/vxlan.c +++ b/drivers/net/vxlan.c @@ -3103,6 +3103,11 @@ static void vxlan_config_apply(struct net_device *dev, max_mtu = lowerdev->mtu - (use_ipv6 ? VXLAN6_HEADROOM : VXLAN_HEADROOM); + if (max_mtu < ETH_MIN_MTU) + max_mtu = ETH_MIN_MTU; + + if (!changelink && !conf->mtu) + dev->mtu = max_mtu; } if (dev->mtu > max_mtu) -- cgit v1.2.1 From a93bf0ff449064e6b7f44e58522e940f88c0d966 Mon Sep 17 00:00:00 2001 From: Xin Long Date: Mon, 18 Dec 2017 14:20:56 +0800 Subject: vxlan: update skb dst pmtu on tx path Unlike ip tunnels, now vxlan doesn't do any pmtu update for upper dst pmtu, even if it doesn't match the lower dst pmtu any more. The problem can be reproduced when reducing the vxlan lower dev's pmtu when running netperf. In jianlin's testing, the performance went to 1/7 of the previous. This patch is to update the upper dst pmtu to match the lower dst pmtu on tx path so that packets can be sent out even when lower dev's pmtu has been changed. It also works for metadata dst. Note that this patch doesn't process any pmtu icmp packet. But even in the future, the support for pmtu icmp packets process of udp tunnels will also needs this. The same thing will be done for geneve in another patch. Signed-off-by: Xin Long Signed-off-by: David S. Miller --- drivers/net/vxlan.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) (limited to 'drivers/net/vxlan.c') diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c index 1000b0e4ee01..31f4b7911ef8 100644 --- a/drivers/net/vxlan.c +++ b/drivers/net/vxlan.c @@ -2155,6 +2155,13 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev, } ndst = &rt->dst; + if (skb_dst(skb)) { + int mtu = dst_mtu(ndst) - VXLAN_HEADROOM; + + skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, + skb, mtu); + } + tos = ip_tunnel_ecn_encap(tos, old_iph, skb); ttl = ttl ? : ip4_dst_hoplimit(&rt->dst); err = vxlan_build_skb(skb, ndst, sizeof(struct iphdr), @@ -2190,6 +2197,13 @@ static void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev, goto out_unlock; } + if (skb_dst(skb)) { + int mtu = dst_mtu(ndst) - VXLAN6_HEADROOM; + + skb_dst(skb)->ops->update_pmtu(skb_dst(skb), NULL, + skb, mtu); + } + tos = ip_tunnel_ecn_encap(tos, old_iph, skb); ttl = ttl ? : ip6_dst_hoplimit(ndst); skb_scrub_packet(skb, xnet); -- cgit v1.2.1