Date: Tue, 15 Mar 2016 16:09:07 +0300 From: Vasily Averin <vvs@...tuozzo.com> To: oss-security@...ts.openwall.com Cc: Solar Designer <solar@...nwall.com>, Cyrill Gorcunov <gorcunov@...tuozzo.com>, "David S. Miller" <davem@...emloft.net>, Konstantin Khorenko <khorenko@...tuozzo.com> Subject: CVE request: ipv4: Don't do expensive useless work during inetdev destroy Destroy of network interface with huge number of ipv4 addresses keeps rtnl_lock for a very long time (up to hour). It blocks many network related operations, including for example creation of new incoming ssh connections. The problem is especially important for containers, container owner have enough permission to enable this trigger and then can block network access on whole host node. The problem is fixed in net-next git by patch fbd40ea0180a2d328c5adc61414dc8bab9335ce2 (http://git.kernel.org/cgit/linux/kernel/git/davem/net-next.git/patch/?id=fbd40ea0180a2d328c5adc61414dc8bab9335ce2) From: David Miller <davem@...emloft.net> ipv4: Don't do expensive useless work during inetdev destroy. When an inetdev is destroyed, every address assigned to the interface is removed. And in this scenerio we do two pointless things which can be very expensive if the number of assigned interfaces is large: 1) Address promotion. We are deleting all addresses, so there is no point in doing this. 2) A full nf conntrack table purge for every address. We only need to do this once, as is already caught by the existing masq_dev_notifier so masq_inet_event() can skip this. Reported-by: Solar Designer <solar@...nwall.com> Signed-off-by: David S. Miller <davem@...emloft.net> Tested-by: Cyrill Gorcunov <gorcunov@...nvz.org>
Powered by blists - more mailing lists
Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.
Powered by Openwall GNU/*/Linux - Powered by OpenVZ