From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Ahern Subject: [PATCH net-next 0/7] net/ipv6: Another followup to the fib6_info change Date: Fri, 20 Apr 2018 15:37:56 -0700 Message-ID: <20180420223803.15743-1-dsahern@gmail.com> Cc: davem@davemloft.net, idosch@idosch.org, roopa@cumulusnetworks.com, eric.dumazet@gmail.com, weiwan@google.com, kafai@fb.com, yoshfuji@linux-ipv6.org, David Ahern To: netdev@vger.kernel.org Return-path: Received: from mail-pf0-f194.google.com ([209.85.192.194]:35310 "EHLO mail-pf0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751169AbeDTWiL (ORCPT ); Fri, 20 Apr 2018 18:38:11 -0400 Received: by mail-pf0-f194.google.com with SMTP id j5so4881670pfh.2 for ; Fri, 20 Apr 2018 15:38:10 -0700 (PDT) Sender: netdev-owner@vger.kernel.org List-ID: Last one - for this week. Patches 1, 2 and 7 are more cleanup patches - removing dead code, moving code from a header to near its single caller, and updating function name. Patches 3-5 do some refactoring leading up to patch 6 which fixes a NULL dereference. I have only managed to trigger a panic once, so I can not definitively confirm it addresses the problem but it seems pretty clear that it is a race on removing a 'from' reference on an rt6_info and another path using that 'from' value to do cookie checking. David Ahern (7): net/ipv6: Clean up rt expires helpers net/ipv6: Rename rt6_get_cookie_safe net/ipv6: Move rcu_read_lock to callers of ip6_rt_cache_alloc net/ipv6: Move rcu locking to callers of fib6_get_cookie_safe net/ipv6: Move release of fib6_info from pcpu routes to helper net/ipv6: Make from in rt6_info rcu protected net/ipv6: Remove unncessary check on f6i in fib6_check include/net/ip6_fib.h | 41 +++++------------ net/ipv6/ip6_fib.c | 45 ++++++++++-------- net/ipv6/ip6_output.c | 9 +++- net/ipv6/route.c | 124 ++++++++++++++++++++++++++++++++++++-------------- 4 files changed, 136 insertions(+), 83 deletions(-) -- 2.11.0