ipv6/addrconf: call ipv6_mc_up() for non-Ethernet interface
Subsystems (3)
PATCH 1/1 ipv6/addrconf: call ipv6_mc_up() for non-Ethernet interface
Rafał found an issue that for non-Ethernet interface, if we down and up frequently, the memory will be consumed slowly.
The reason is we add allnodes/allrouters addressed in multicast list in ipv6_add_dev(). When link down, we call ipv6_mc_down(), store all multicast addresses via mld_add_delrec(). But when link up, we don't call ipv6_mc_up() for non-Ethernet interface to remove the addresses. This makes idev->mc_tomb getting bigger and bigger. The call stack looks like:
addrconf_notify(NETDEV_REGISTER)
ipv6_add_dev ipv6_dev_mc_inc(ff01::1) ipv6_dev_mc_inc(ff02::1) ipv6_dev_mc_inc(ff02::2)
addrconf_notify(NETDEV_UP)
addrconf_dev_config /* Alas, we support only Ethernet autoconfiguration. */ return;
addrconf_notify(NETDEV_DOWN)
addrconf_ifdown ipv6_mc_down igmp6_group_dropped(ff02::2) mld_add_delrec(ff02::2) igmp6_group_dropped(ff02::1) igmp6_group_dropped(ff01::1)
After investigating, I can't found a rule to disable multicast on non-Ethernet interface. In RFC2460, the link could be Ethernet, PPP, ATM, tunnels, etc. In IPv4, it doesn't check the dev type when calls ip_mc_up() in inetdev_event(). Even for IPv6, we don't check the dev type and call ipv6_add_dev(), ipv6_dev_mc_inc() after register device.
So I think it's OK to fix this memory consumer by calling ipv6_mc_up() for non-Ethernet interface.
--- net/ipv6/addrconf.c | 3 +++ 1 file changed, 3 insertions(+)
net/ipv6/addrconf.c +3 -0
@@ -3347,6 +3347,9 @@ static void addrconf_dev_config(struct net_device *dev) (dev->type != ARPHRD_NONE) && (dev->type != ARPHRD_RAWIP)) {/* Alas, we support only Ethernet autoconfiguration. */ + idev =__in6_dev_get (dev); +if (!IS_ERR_OR_NULL (idev) && dev->flags &IFF_UP ) +ipv6_mc_up (idev);return ; }
Signature · 1 line
2.19.2
PATCH v2 1/1 ipv6/addrconf: call ipv6_mc_up() for non-Ethernet interface
Rafał found an issue that for non-Ethernet interface, if we down and up frequently, the memory will be consumed slowly.
The reason is we add allnodes/allrouters addressed in multicast list in ipv6_add_dev(). When link down, we call ipv6_mc_down(), store all multicast addresses via mld_add_delrec(). But when link up, we don't call ipv6_mc_up() for non-Ethernet interface to remove the addresses. This makes idev->mc_tomb getting bigger and bigger. The call stack looks like:
addrconf_notify(NETDEV_REGISTER)
ipv6_add_dev ipv6_dev_mc_inc(ff01::1) ipv6_dev_mc_inc(ff02::1) ipv6_dev_mc_inc(ff02::2)
addrconf_notify(NETDEV_UP)
addrconf_dev_config /* Alas, we support only Ethernet autoconfiguration. */ return;
addrconf_notify(NETDEV_DOWN)
addrconf_ifdown ipv6_mc_down igmp6_group_dropped(ff02::2) mld_add_delrec(ff02::2) igmp6_group_dropped(ff02::1) igmp6_group_dropped(ff01::1)
After investigating, I can't found a rule to disable multicast on non-Ethernet interface. In RFC2460, the link could be Ethernet, PPP, ATM, tunnels, etc. In IPv4, it doesn't check the dev type when calls ip_mc_up() in inetdev_event(). Even for IPv6, we don't check the dev type and call ipv6_add_dev(), ipv6_dev_mc_inc() after register device.
So I think it's OK to fix this memory consumer by calling ipv6_mc_up() for non-Ethernet interface.
v2: Also check IFF_MULTICAST flag to make sure the interface supports
multicast
--- net/ipv6/addrconf.c | 4 ++++ 1 file changed, 4 insertions(+)
net/ipv6/addrconf.c +4 -0
@@ -3347,6 +3347,10 @@ static void addrconf_dev_config(struct net_device *dev) (dev->type != ARPHRD_NONE) && (dev->type != ARPHRD_RAWIP)) {/* Alas, we support only Ethernet autoconfiguration. */ + idev =__in6_dev_get (dev); +if (!IS_ERR_OR_NULL (idev) && dev->flags &IFF_UP && + dev->flags &IFF_MULTICAST ) +ipv6_mc_up (idev);return ; }
Signature · 1 line
2.19.2
Quoted 37 lines · Hangbin Liu
Rafał found an issue that for non-Ethernet interface, if we down and up frequently, the memory will be consumed slowly.
The reason is we add allnodes/allrouters addressed in multicast list in
ipv6_add_dev(). When link down, we callipv6_mc_down(), store all multicast addresses viamld_add_delrec(). But when link up, we don't callipv6_mc_up()for non-Ethernet interface to remove the addresses. This makesidev->mc_tombgetting bigger and bigger. The call stack looks like:
addrconf_notify(NETDEV_REGISTER)ipv6_add_dev ipv6_dev_mc_inc(ff01::1) ipv6_dev_mc_inc(ff02::1) ipv6_dev_mc_inc(ff02::2)
addrconf_notify(NETDEV_UP)addrconf_dev_config /* Alas, we support only Ethernet autoconfiguration. */ return;
addrconf_notify(NETDEV_DOWN)addrconf_ifdown ipv6_mc_down igmp6_group_dropped(ff02::2) mld_add_delrec(ff02::2) igmp6_group_dropped(ff02::1) igmp6_group_dropped(ff01::1)After investigating, I can't found a rule to disable multicast on non-Ethernet interface. In RFC2460, the link could be Ethernet, PPP, ATM, tunnels, etc. In IPv4, it doesn't check the dev type when calls
ip_mc_up()ininetdev_event(). Even for IPv6, we don't check the dev type and callipv6_add_dev(),ipv6_dev_mc_inc()after register device.So I think it's OK to fix this memory consumer by calling
ipv6_mc_up()for non-Ethernet interface.
Quoted 7 lines
v2: Also check
IFF_MULTICASTflag to make sure the interface supports
multicastReported-by Rafał Miłecki <[email protected]>Fixes 74235a25c673 ("[IPV6] addrconf: Fix IPv6 on tuntap tunnels")Fixes 1666d49e1d41 ("mld: do not remove mld souce list info when set link down")Signed-off-by Hangbin Liu <[email protected]>
I'm really grateful for all your help. I verified this patch to fix reported issue. Thank you!
David Miller
maintainerFrom: Hangbin Liu <[email protected]>
Date: Tue, 10 Mar 2020 15:27:37 +0800
Quoted 37 lines
Rafał found an issue that for non-Ethernet interface, if we down and up frequently, the memory will be consumed slowly.
The reason is we add allnodes/allrouters addressed in multicast list in
ipv6_add_dev(). When link down, we callipv6_mc_down(), store all multicast addresses viamld_add_delrec(). But when link up, we don't callipv6_mc_up()for non-Ethernet interface to remove the addresses. This makesidev->mc_tombgetting bigger and bigger. The call stack looks like:
addrconf_notify(NETDEV_REGISTER)ipv6_add_dev ipv6_dev_mc_inc(ff01::1) ipv6_dev_mc_inc(ff02::1) ipv6_dev_mc_inc(ff02::2)
addrconf_notify(NETDEV_UP)addrconf_dev_config /* Alas, we support only Ethernet autoconfiguration. */ return;
addrconf_notify(NETDEV_DOWN)addrconf_ifdown ipv6_mc_down igmp6_group_dropped(ff02::2) mld_add_delrec(ff02::2) igmp6_group_dropped(ff02::1) igmp6_group_dropped(ff01::1)After investigating, I can't found a rule to disable multicast on non-Ethernet interface. In RFC2460, the link could be Ethernet, PPP, ATM, tunnels, etc. In IPv4, it doesn't check the dev type when calls
ip_mc_up()ininetdev_event(). Even for IPv6, we don't check the dev type and callipv6_add_dev(),ipv6_dev_mc_inc()after register device.So I think it's OK to fix this memory consumer by calling
ipv6_mc_up()for non-Ethernet interface.
Quoted 7 lines
v2: Also check
IFF_MULTICASTflag to make sure the interface supports
multicastReported-by Rafał Miłecki <[email protected]>Fixes 74235a25c673 ("[IPV6] addrconf: Fix IPv6 on tuntap tunnels")Fixes 1666d49e1d41 ("mld: do not remove mld souce list info when set link down")Signed-off-by Hangbin Liu <[email protected]>
Applied and queued up for -stable, thank you.