From: Jeremiah Mahler <jmmahler@gmail.com>
To: David Ahern <dsa@cumulusnetworks.com>
Cc: netdev@vger.kernel.org, Andrey Wagin <avagin@gmail.com>,
Ying Huang <ying.huang@linux.intel.com>
Subject: Re: [PATCH net-next] net: ipv6: Fix refcnt on host routes
Date: Wed, 2 Mar 2016 17:23:40 -0800 [thread overview]
Message-ID: <20160303012340.GA1894@hudson.localdomain> (raw)
In-Reply-To: <1456947007-32444-1-git-send-email-dsa@cumulusnetworks.com>
Hi David,
On Wed, Mar 02, 2016 at 11:30:07AM -0800, David Ahern wrote:
> Andrew and Ying Huang's test robot both reported usage count problems that
> trace back to the 'keep address on ifdown' patch.
>
> From Andrew:
> We execute CRIU test on linux-next. On the current linux-next kernel
> they hangs on creating a network namespace.
>
> The kernel log contains many massages like this:
> [ 1036.122108] unregister_netdevice: waiting for lo to become free.
> Usage count = 2
> [ 1046.165156] unregister_netdevice: waiting for lo to become free.
> Usage count = 2
> [ 1056.210287] unregister_netdevice: waiting for lo to become free.
> Usage count = 2
>
> I tried to revert this patch and the bug disappeared.
>
> Here is a set of commands to reproduce this bug:
>
> [root@linux-next-test linux-next]# uname -a
> Linux linux-next-test 4.5.0-rc6-next-20160301+ #3 SMP Wed Mar 2
> 17:32:18 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
>
> [root@linux-next-test ~]# unshare -n
> [root@linux-next-test ~]# ip link set up dev lo
> [root@linux-next-test ~]# ip a
> 1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
> group default qlen 1
> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
> inet 127.0.0.1/8 scope host lo
> valid_lft forever preferred_lft forever
> inet6 ::1/128 scope host
> valid_lft forever preferred_lft forever
> [root@linux-next-test ~]# logout
> [root@linux-next-test ~]# unshare -n
>
> -----
>
> The problem is a change made to RTM_DELADDR case in __ipv6_ifa_notify that
> was added in an early version of the offending patch and is no longer
> needed.
>
> Fixes: f1705ec197e7 ("net: ipv6: Make address flushing on ifdown optional")
> Cc: Andrey Wagin <avagin@gmail.com>
> Cc: Ying Huang <ying.huang@linux.intel.com>
> Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
> ---
> net/ipv6/addrconf.c | 8 +++-----
> 1 file changed, 3 insertions(+), 5 deletions(-)
>
> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
> index b1169d511ab4..8c0dab2de5c9 100644
> --- a/net/ipv6/addrconf.c
> +++ b/net/ipv6/addrconf.c
> @@ -5319,12 +5319,10 @@ static void __ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
> if (rt)
> ip6_del_rt(rt);
> }
> - if (ifp->rt) {
> - dst_hold(&ifp->rt->dst);
> + dst_hold(&ifp->rt->dst);
> +
> + ip6_del_rt(ifp->rt);
>
> - ip6_del_rt(ifp->rt);
> - ifp->rt = NULL;
> - }
> rt_genid_bump_ipv6(net);
> break;
> }
> --
> 2.1.4
>
Tested-by: Jeremiah Mahler <jmmahler@gmail.com>
--
- Jeremiah Mahler
next prev parent reply other threads:[~2016-03-03 1:23 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-02 19:30 [PATCH net-next] net: ipv6: Fix refcnt on host routes David Ahern
2016-03-03 1:23 ` Jeremiah Mahler [this message]
2016-03-03 14:41 ` Daniel Borkmann
2016-03-03 22:26 ` David Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160303012340.GA1894@hudson.localdomain \
--to=jmmahler@gmail.com \
--cc=avagin@gmail.com \
--cc=dsa@cumulusnetworks.com \
--cc=netdev@vger.kernel.org \
--cc=ying.huang@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).