| Message ID | CAHA+R7Mhd2KYaU8THAd+O4HzQNRLxhpGccvEPOqWJRKAwZbpgQ@mail.gmail.com |
|---|---|
| State | RFC, archived |
| Delegated to: | David Miller |
| Headers | show |
On 05/22/2014 02:12 PM, Cong Wang wrote: > On Wed, May 21, 2014 at 9:42 PM, Jet Chen <jet.chen@intel.com> wrote: >> Hi Steffen, >> >> 0day kernel testing robot got the below dmesg and the first bad commit is >> >> git://git.kernel.org/pub/scm/linux/kernel/git/davem/net.git master >> commit 78ff4be45a4c51d8fb21ad92e4fabb467c6c3eeb >> Author: Steffen Klassert <steffen.klassert@secunet.com> >> AuthorDate: Mon May 19 11:36:56 2014 +0200 >> Commit: David S. Miller <davem@davemloft.net> >> CommitDate: Wed May 21 02:08:32 2014 -0400 >> >> ip_tunnel: Initialize the fallback device properly >> We need to initialize the fallback device to have a correct mtu >> set on this device. Otherwise the mtu is set to null and the device >> is unusable. >> Fixes: fd58156e456d ("IPIP: Use ip-tunneling code.") >> Cc: Pravin B Shelar <pshelar@nicira.com> >> Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com> >> Signed-off-by: David S. Miller <davem@davemloft.net> >> >> +--------------------------------------------------------------------------------+------------+------------+ >> | >> | d8d33c3b8a | 78ff4be45a | >> +--------------------------------------------------------------------------------+------------+------------+ >> | boot_successes >> | 60 | 0 | >> | boot_failures >> | 0 | 20 | >> | >> WARNING:CPU:PID:at_net/batman-adv/hard-interface.c:batadv_is_on_batman_iface() >> | 0 | 20 | >> | backtrace:register_netdevice_notifier >> | 0 | 20 | >> | backtrace:batadv_init >> | 0 | 20 | >> | backtrace:kernel_init_freeable >> | 0 | 20 | >> +--------------------------------------------------------------------------------+------------+------------+ >> > > batman needs to fix: > > diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c > index b851cc5..fbda6b5 100644 > --- a/net/batman-adv/hard-interface.c > +++ b/net/batman-adv/hard-interface.c > @@ -83,7 +83,7 @@ static bool batadv_is_on_batman_iface(const struct > net_device *net_dev) > return true; > > /* no more parents..stop recursion */ > - if (net_dev->iflink == net_dev->ifindex) > + if (net_dev->iflink == 0 || net_dev->iflink == net_dev->ifindex) > return false; > > /* recurse over the parent device */ > Your patch fixes that issue. Tested-by: Jet Chen <jet.chen@intel.com> Thanks, Jet -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, May 21, 2014 at 11:12:36PM -0700, Cong Wang wrote: > > batman needs to fix: > > diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c > index b851cc5..fbda6b5 100644 > --- a/net/batman-adv/hard-interface.c > +++ b/net/batman-adv/hard-interface.c > @@ -83,7 +83,7 @@ static bool batadv_is_on_batman_iface(const struct > net_device *net_dev) > return true; > > /* no more parents..stop recursion */ > - if (net_dev->iflink == net_dev->ifindex) > + if (net_dev->iflink == 0 || net_dev->iflink == net_dev->ifindex) > return false; > > /* recurse over the parent device */ Right, I think this is the correct fix. Please do a proper patch submission for this. I wondered why this issue does not show up with ipv6 tunnels and it appears to be a bug :) When a ipv6 tunnel interface is created, ip6_tnl_create2() calls ip6_tnl_dev_init() which calls ip6_tnl_link_config(). ip6_tnl_link_config() sets dev->iflink. Then register_netdev() overwrites this setting with dev->iflink = dev->ifindex. The output of the 'ip a' command looks like this: 9: ip6tnl0: <NOARP> mtu 1452 qdisc noop state DOWN link/tunnel6 :: brd :: 11: ip6tunnel1: <POINTOPOINT,NOARP> mtu 1452 qdisc noop state DOWN link/tunnel6 2001:470:1f00:ffff::191 peer 2001:470:1f00:ffff::189 When I change to avoid the overwriting of dev->iflink it looks like this: 9: ip6tnl0@NONE: <NOARP> mtu 1452 qdisc noop state DOWN link/tunnel6 :: brd :: 12: ip6tunnel1@eth0: <POINTOPOINT,NOARP> mtu 1452 qdisc noop state DOWN link/tunnel6 2001:470:1f00:ffff::191 peer 2001:470:1f00:ffff::189 So I think dev->iflink should be set in the ndo_init() function of the tunnel, then register_netdev() leaves dev->iflink alone. Thoughts, comments? -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/net/batman-adv/hard-interface.c b/net/batman-adv/hard-interface.c index b851cc5..fbda6b5 100644 --- a/net/batman-adv/hard-interface.c +++ b/net/batman-adv/hard-interface.c @@ -83,7 +83,7 @@ static bool batadv_is_on_batman_iface(const struct net_device *net_dev) return true; /* no more parents..stop recursion */ - if (net_dev->iflink == net_dev->ifindex) + if (net_dev->iflink == 0 || net_dev->iflink == net_dev->ifindex) return false;