ip_tunnel: fix over-mtu packet send fail without TUNNEL_DONT_FRAGMENT flags
authorwenxu <wenxu@ucloud.cn>
Fri, 30 Oct 2020 03:32:08 +0000 (11:32 +0800)
committerJakub Kicinski <kuba@kernel.org>
Sun, 1 Nov 2020 00:19:02 +0000 (17:19 -0700)
commit20149e9eb68c003eaa09e7c9a49023df40779552
treef3d533820c0a2d5c8905233edb5c961a55ca16ef
parent403dc16796f5516acf23d94a1cd9eba564d03210
ip_tunnel: fix over-mtu packet send fail without TUNNEL_DONT_FRAGMENT flags

The tunnel device such as vxlan, bareudp and geneve in the lwt mode set
the outer df only based TUNNEL_DONT_FRAGMENT.
And this was also the behavior for gre device before switching to use
ip_md_tunnel_xmit in commit 962924fa2b7a ("ip_gre: Refactor collect
metatdata mode tunnel xmit to ip_md_tunnel_xmit")

When the ip_gre in lwt mode xmit with ip_md_tunnel_xmi changed the rule and
make the discrepancy between handling of DF by different tunnels. So in the
ip_md_tunnel_xmit should follow the same rule like other tunnels.

Fixes: cfc7381b3002 ("ip_tunnel: add collect_md mode to IPIP tunnel")
Signed-off-by: wenxu <wenxu@ucloud.cn>
Link: https://lore.kernel.org/r/1604028728-31100-1-git-send-email-wenxu@ucloud.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
net/ipv4/ip_tunnel.c