[PATCH stable 3.11] sit: amend "allow to use rtnl ops on fb tunnel"

[PATCH stable 3.11] sit: amend "allow to use rtnl ops on fb tunnel"

[Willem de Bruijn]

Amend backport to 3.11.y of

 [ Upstream commit 205983c43700ac3a81e7625273a3fa83cd2759b5 ]

The discussion thread in the upstream commit mentions that in
backports to stable-* branches, the line

  - unregister_netdevice_queue(sitn->fb_tunnel_dev, &list);

must be omitted if that branch does not have commit 5e6700b3bf98
("sit: add support of x-netns"). This line has correctly been omitted
in the backport to 3.10, which indeed does not have that commit.

It was also removed in the backport to 3.11.y, which does have that
commit.

This causes the following steps to hit a BUG at net/core/dev.c:5039:

  `modprobe sit; rmmod sit`

The bug demonstrates that it causes a device to be unregistered twice.
The simple fix is to apply the one line in the upstream commit that
was dropped in the backport to 3.11 (3783100374653e2e7fbdf68c710f5).
This brings the logic in line with upstream linux, net and net-next
branches.

Signed-off-by: Willem de Bruijn <willemb@google.com>

---

The policy for networking patches is to queue stable patches up for
the davem/stable queue.

Since this patch only applies to one specific branch, I assumed
that creating it directly against 3.11.y is the right process. I
did *not* directly cc: the stable list. David, please let me know
if this should not go through your queue at all and I should
resubmit to stable.
---
 net/ipv6/sit.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c
index 86f639b..a51ad07 100644
--- a/net/ipv6/sit.c
+++ b/net/ipv6/sit.c
@@ -1708,7 +1708,6 @@ static void __net_exit sit_exit_net(struct net *net)
 
 	rtnl_lock();
 	sit_destroy_tunnels(sitn, &list);
-	unregister_netdevice_queue(sitn->fb_tunnel_dev, &list);
 	unregister_netdevice_many(&list);
 	rtnl_unlock();
 }
-- 
1.8.4

Re: [PATCH stable 3.11] sit: amend "allow to use rtnl ops on fb tunnel"

[Nicolas Dichtel]

Le 22/10/2013 16:59, Willem de Bruijn a écrit :
> Amend backport to 3.11.y of
>
>   [ Upstream commit 205983c43700ac3a81e7625273a3fa83cd2759b5 ]
>
> The discussion thread in the upstream commit mentions that in
> backports to stable-* branches, the line
>
>    - unregister_netdevice_queue(sitn->fb_tunnel_dev, &list);
>
> must be omitted if that branch does not have commit 5e6700b3bf98
> ("sit: add support of x-netns"). This line has correctly been omitted
> in the backport to 3.10, which indeed does not have that commit.
>
> It was also removed in the backport to 3.11.y, which does have that
> commit.
>
> This causes the following steps to hit a BUG at net/core/dev.c:5039:
>
>    `modprobe sit; rmmod sit`
>
> The bug demonstrates that it causes a device to be unregistered twice.
> The simple fix is to apply the one line in the upstream commit that
> was dropped in the backport to 3.11 (3783100374653e2e7fbdf68c710f5).
> This brings the logic in line with upstream linux, net and net-next
> branches.
>
> Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>

Re: [PATCH stable 3.11] sit: amend "allow to use rtnl ops on fb tunnel"

[Veaceslav Falico]

On Tue, Oct 22, 2013 at 10:59:18AM -0400, Willem de Bruijn wrote:
>Amend backport to 3.11.y of
>
> [ Upstream commit 205983c43700ac3a81e7625273a3fa83cd2759b5 ]
>
>The discussion thread in the upstream commit mentions that in
>backports to stable-* branches, the line
>
>  - unregister_netdevice_queue(sitn->fb_tunnel_dev, &list);
>
>must be omitted if that branch does not have commit 5e6700b3bf98
>("sit: add support of x-netns"). This line has correctly been omitted
>in the backport to 3.10, which indeed does not have that commit.
>
>It was also removed in the backport to 3.11.y, which does have that
>commit.
>
>This causes the following steps to hit a BUG at net/core/dev.c:5039:
>
>  `modprobe sit; rmmod sit`
>
>The bug demonstrates that it causes a device to be unregistered twice.
>The simple fix is to apply the one line in the upstream commit that
>was dropped in the backport to 3.11 (3783100374653e2e7fbdf68c710f5).
>This brings the logic in line with upstream linux, net and net-next
>branches.

Also seen that in stable, fixes it for me.

FWIW...

Reviewed-by: Veaceslav Falico <vfalico@redhat.com>

>
>Signed-off-by: Willem de Bruijn <willemb@google.com>
>
>---
>
>The policy for networking patches is to queue stable patches up for
>the davem/stable queue.
>
>Since this patch only applies to one specific branch, I assumed
>that creating it directly against 3.11.y is the right process. I
>did *not* directly cc: the stable list. David, please let me know
>if this should not go through your queue at all and I should
>resubmit to stable.
>---
> net/ipv6/sit.c | 1 -
> 1 file changed, 1 deletion(-)
>
>diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c
>index 86f639b..a51ad07 100644
>--- a/net/ipv6/sit.c
>+++ b/net/ipv6/sit.c
>@@ -1708,7 +1708,6 @@ static void __net_exit sit_exit_net(struct net *net)
>
> 	rtnl_lock();
> 	sit_destroy_tunnels(sitn, &list);
>-	unregister_netdevice_queue(sitn->fb_tunnel_dev, &list);
> 	unregister_netdevice_many(&list);
> 	rtnl_unlock();
> }
>-- 
>1.8.4
>
>--
>To unsubscribe from this list: send the line "unsubscribe netdev" in
>the body of a message to majordomo@vger.kernel.org
>More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH stable 3.11] sit: amend "allow to use rtnl ops on fb tunnel"

[David Miller]

From: Veaceslav Falico <vfalico@redhat.com>
Date: Tue, 22 Oct 2013 17:56:27 +0200

> On Tue, Oct 22, 2013 at 10:59:18AM -0400, Willem de Bruijn wrote:
>>Amend backport to 3.11.y of
>>
>> [ Upstream commit 205983c43700ac3a81e7625273a3fa83cd2759b5 ]
>>
>>The discussion thread in the upstream commit mentions that in
>>backports to stable-* branches, the line
>>
>>  - unregister_netdevice_queue(sitn->fb_tunnel_dev, &list);
>>
>>must be omitted if that branch does not have commit 5e6700b3bf98
>>("sit: add support of x-netns"). This line has correctly been omitted
>>in the backport to 3.10, which indeed does not have that commit.
>>
>>It was also removed in the backport to 3.11.y, which does have that
>>commit.
>>
>>This causes the following steps to hit a BUG at net/core/dev.c:5039:
>>
>>  `modprobe sit; rmmod sit`
>>
>>The bug demonstrates that it causes a device to be unregistered twice.
>>The simple fix is to apply the one line in the upstream commit that
>>was dropped in the backport to 3.11 (3783100374653e2e7fbdf68c710f5).
>>This brings the logic in line with upstream linux, net and net-next
>>branches.
> 
> Also seen that in stable, fixes it for me.
> 
> FWIW...
> 
> Reviewed-by: Veaceslav Falico <vfalico@redhat.com>
> 
>>
>>Signed-off-by: Willem de Bruijn <willemb@google.com>

I'll submit this to -stable, thanks everyone.