* [Patch net] ipv6: avoid unregistering inet6_dev for loopback
@ 2017-06-21 21:34 Cong Wang
2017-06-22 3:07 ` David Ahern
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Cong Wang @ 2017-06-21 21:34 UTC (permalink / raw)
To: netdev; +Cc: Cong Wang, Andrey Konovalov, Daniel Lezcano, David Ahern
The per netns loopback_dev->ip6_ptr is unregistered and set to
NULL when its mtu is set to smaller than IPV6_MIN_MTU, this
leads to that we could set rt->rt6i_idev NULL after a
rt6_uncached_list_flush_dev() and then crash after another
call.
In this case we should just bring its inet6_dev down, rather
than unregistering it, at least prior to commit 176c39af29bc
("netns: fix addrconf_ifdown kernel panic") we always
override the case for loopback.
Thanks a lot to Andrey for finding a reliable reproducer.
Fixes: 176c39af29bc ("netns: fix addrconf_ifdown kernel panic")
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Cc: Daniel Lezcano <dlezcano@fr.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
---
net/ipv6/addrconf.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 686c923..1d2dbac 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -3369,6 +3369,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
struct net_device *dev = netdev_notifier_info_to_dev(ptr);
struct netdev_notifier_changeupper_info *info;
struct inet6_dev *idev = __in6_dev_get(dev);
+ struct net *net = dev_net(dev);
int run_pending = 0;
int err;
@@ -3384,7 +3385,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
case NETDEV_CHANGEMTU:
/* if MTU under IPV6_MIN_MTU stop IPv6 on this interface. */
if (dev->mtu < IPV6_MIN_MTU) {
- addrconf_ifdown(dev, 1);
+ addrconf_ifdown(dev, dev != net->loopback_dev);
break;
}
@@ -3500,7 +3501,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
* IPV6_MIN_MTU stop IPv6 on this interface.
*/
if (dev->mtu < IPV6_MIN_MTU)
- addrconf_ifdown(dev, 1);
+ addrconf_ifdown(dev, dev != net->loopback_dev);
}
break;
--
2.5.5
^ permalink raw reply related [flat|nested] 5+ messages in thread* Re: [Patch net] ipv6: avoid unregistering inet6_dev for loopback
2017-06-21 21:34 [Patch net] ipv6: avoid unregistering inet6_dev for loopback Cong Wang
@ 2017-06-22 3:07 ` David Ahern
2017-06-22 10:10 ` Andrey Konovalov
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: David Ahern @ 2017-06-22 3:07 UTC (permalink / raw)
To: Cong Wang, netdev; +Cc: Andrey Konovalov, Daniel Lezcano
On 6/21/17 5:34 PM, Cong Wang wrote:
> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
> index 686c923..1d2dbac 100644
> --- a/net/ipv6/addrconf.c
> +++ b/net/ipv6/addrconf.c
> @@ -3369,6 +3369,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
> struct net_device *dev = netdev_notifier_info_to_dev(ptr);
> struct netdev_notifier_changeupper_info *info;
> struct inet6_dev *idev = __in6_dev_get(dev);
> + struct net *net = dev_net(dev);
Given the number of NETDEV events and the fact that only 2 need net,
perhaps the dev_net(dev) should be inline ...
> int run_pending = 0;
> int err;
>
> @@ -3384,7 +3385,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
> case NETDEV_CHANGEMTU:
> /* if MTU under IPV6_MIN_MTU stop IPv6 on this interface. */
> if (dev->mtu < IPV6_MIN_MTU) {
> - addrconf_ifdown(dev, 1);
> + addrconf_ifdown(dev, dev != net->loopback_dev);
so:
addrconf_ifdown(dev, dev != dev_net(dev)->loopback_dev);
> break;
> }
>
> @@ -3500,7 +3501,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
> * IPV6_MIN_MTU stop IPv6 on this interface.
> */
> if (dev->mtu < IPV6_MIN_MTU)
> - addrconf_ifdown(dev, 1);
> + addrconf_ifdown(dev, dev != net->loopback_dev);
> }
> break;
>
>
The overall idea of not destroying idev for loopback on a failure to set
MTU seems correct to me.
Acked-by: David Ahern <dsahern@gmail.com>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Patch net] ipv6: avoid unregistering inet6_dev for loopback
2017-06-21 21:34 [Patch net] ipv6: avoid unregistering inet6_dev for loopback Cong Wang
2017-06-22 3:07 ` David Ahern
@ 2017-06-22 10:10 ` Andrey Konovalov
2017-06-22 17:22 ` David Miller
2017-06-22 17:24 ` David Miller
3 siblings, 0 replies; 5+ messages in thread
From: Andrey Konovalov @ 2017-06-22 10:10 UTC (permalink / raw)
To: Cong Wang; +Cc: netdev, Daniel Lezcano, David Ahern
On Wed, Jun 21, 2017 at 11:34 PM, Cong Wang <xiyou.wangcong@gmail.com> wrote:
> The per netns loopback_dev->ip6_ptr is unregistered and set to
> NULL when its mtu is set to smaller than IPV6_MIN_MTU, this
> leads to that we could set rt->rt6i_idev NULL after a
> rt6_uncached_list_flush_dev() and then crash after another
> call.
>
> In this case we should just bring its inet6_dev down, rather
> than unregistering it, at least prior to commit 176c39af29bc
> ("netns: fix addrconf_ifdown kernel panic") we always
> override the case for loopback.
>
> Thanks a lot to Andrey for finding a reliable reproducer.
Thanks! This fixes the bug for me.
Tested-by: Andrey Konovalov <andreyknvl@google.com>
>
> Fixes: 176c39af29bc ("netns: fix addrconf_ifdown kernel panic")
> Reported-by: Andrey Konovalov <andreyknvl@google.com>
> Cc: Andrey Konovalov <andreyknvl@google.com>
> Cc: Daniel Lezcano <dlezcano@fr.ibm.com>
> Cc: David Ahern <dsahern@gmail.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
> ---
> net/ipv6/addrconf.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
> index 686c923..1d2dbac 100644
> --- a/net/ipv6/addrconf.c
> +++ b/net/ipv6/addrconf.c
> @@ -3369,6 +3369,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
> struct net_device *dev = netdev_notifier_info_to_dev(ptr);
> struct netdev_notifier_changeupper_info *info;
> struct inet6_dev *idev = __in6_dev_get(dev);
> + struct net *net = dev_net(dev);
> int run_pending = 0;
> int err;
>
> @@ -3384,7 +3385,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
> case NETDEV_CHANGEMTU:
> /* if MTU under IPV6_MIN_MTU stop IPv6 on this interface. */
> if (dev->mtu < IPV6_MIN_MTU) {
> - addrconf_ifdown(dev, 1);
> + addrconf_ifdown(dev, dev != net->loopback_dev);
> break;
> }
>
> @@ -3500,7 +3501,7 @@ static int addrconf_notify(struct notifier_block *this, unsigned long event,
> * IPV6_MIN_MTU stop IPv6 on this interface.
> */
> if (dev->mtu < IPV6_MIN_MTU)
> - addrconf_ifdown(dev, 1);
> + addrconf_ifdown(dev, dev != net->loopback_dev);
> }
> break;
>
> --
> 2.5.5
>
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Patch net] ipv6: avoid unregistering inet6_dev for loopback
2017-06-21 21:34 [Patch net] ipv6: avoid unregistering inet6_dev for loopback Cong Wang
2017-06-22 3:07 ` David Ahern
2017-06-22 10:10 ` Andrey Konovalov
@ 2017-06-22 17:22 ` David Miller
2017-06-22 17:24 ` David Miller
3 siblings, 0 replies; 5+ messages in thread
From: David Miller @ 2017-06-22 17:22 UTC (permalink / raw)
To: xiyou.wangcong; +Cc: netdev, andreyknvl, dlezcano, dsahern
From: Cong Wang <xiyou.wangcong@gmail.com>
Date: Wed, 21 Jun 2017 14:34:58 -0700
> The per netns loopback_dev->ip6_ptr is unregistered and set to
> NULL when its mtu is set to smaller than IPV6_MIN_MTU, this
> leads to that we could set rt->rt6i_idev NULL after a
> rt6_uncached_list_flush_dev() and then crash after another
> call.
>
> In this case we should just bring its inet6_dev down, rather
> than unregistering it, at least prior to commit 176c39af29bc
> ("netns: fix addrconf_ifdown kernel panic") we always
> override the case for loopback.
>
> Thanks a lot to Andrey for finding a reliable reproducer.
>
> Fixes: 176c39af29bc ("netns: fix addrconf_ifdown kernel panic")
> Reported-by: Andrey Konovalov <andreyknvl@google.com>
> Cc: Andrey Konovalov <andreyknvl@google.com>
> Cc: Daniel Lezcano <dlezcano@fr.ibm.com>
> Cc: David Ahern <dsahern@gmail.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Applied and queued up for -stable.
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [Patch net] ipv6: avoid unregistering inet6_dev for loopback
2017-06-21 21:34 [Patch net] ipv6: avoid unregistering inet6_dev for loopback Cong Wang
` (2 preceding siblings ...)
2017-06-22 17:22 ` David Miller
@ 2017-06-22 17:24 ` David Miller
3 siblings, 0 replies; 5+ messages in thread
From: David Miller @ 2017-06-22 17:24 UTC (permalink / raw)
To: xiyou.wangcong; +Cc: netdev, andreyknvl, dlezcano, dsahern
From: Cong Wang <xiyou.wangcong@gmail.com>
Date: Wed, 21 Jun 2017 14:34:58 -0700
> The per netns loopback_dev->ip6_ptr is unregistered and set to
> NULL when its mtu is set to smaller than IPV6_MIN_MTU, this
> leads to that we could set rt->rt6i_idev NULL after a
> rt6_uncached_list_flush_dev() and then crash after another
> call.
>
> In this case we should just bring its inet6_dev down, rather
> than unregistering it, at least prior to commit 176c39af29bc
> ("netns: fix addrconf_ifdown kernel panic") we always
> override the case for loopback.
>
> Thanks a lot to Andrey for finding a reliable reproducer.
>
> Fixes: 176c39af29bc ("netns: fix addrconf_ifdown kernel panic")
> Reported-by: Andrey Konovalov <andreyknvl@google.com>
> Cc: Andrey Konovalov <andreyknvl@google.com>
> Cc: Daniel Lezcano <dlezcano@fr.ibm.com>
> Cc: David Ahern <dsahern@gmail.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Applied, thanks.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2017-06-22 17:24 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-06-21 21:34 [Patch net] ipv6: avoid unregistering inet6_dev for loopback Cong Wang
2017-06-22 3:07 ` David Ahern
2017-06-22 10:10 ` Andrey Konovalov
2017-06-22 17:22 ` David Miller
2017-06-22 17:24 ` David Miller
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).