diff mbox

[net-next,v3] netpoll: fix a rtnl lock assertion failure

Message ID 1358160234-24996-1-git-send-email-amwang@redhat.com
State Changes Requested, archived
Delegated to: David Miller
Headers show

Commit Message

Amerigo Wang Jan. 14, 2013, 10:43 a.m. UTC 
From: Cong Wang <amwang@redhat.com>

This patch fixes the following warning:

[   72.013864] RTNL: assertion failed at net/core/dev.c (4955)
[   72.017758] Pid: 668, comm: netpoll-prep-v6 Not tainted 3.8.0-rc1+ #474
[   72.019582] Call Trace:
[   72.020295]  [<ffffffff8176653d>] netdev_master_upper_dev_get+0x35/0x58
[   72.022545]  [<ffffffff81784edd>] netpoll_setup+0x61/0x340
[   72.024846]  [<ffffffff815d837e>] store_enabled+0x82/0xc3
[   72.027466]  [<ffffffff815d7e51>] netconsole_target_attr_store+0x35/0x37
[   72.029348]  [<ffffffff811c3479>] configfs_write_file+0xe2/0x10c
[   72.030959]  [<ffffffff8115d239>] vfs_write+0xaf/0xf6
[   72.032359]  [<ffffffff81978a05>] ? sysret_check+0x22/0x5d
[   72.033824]  [<ffffffff8115d453>] sys_write+0x5c/0x84
[   72.035328]  [<ffffffff819789d9>] system_call_fastpath+0x16/0x1b

Just hold RCU read lock and call netdev_master_upper_dev_get_rcu(),
as suggested by Jiri.

Cc: Jiri Pirko <jiri@resnulli.us>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>

---
--
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

Comments

Jiri Pirko Jan. 14, 2013, 10:54 a.m. UTC | #1 
Mon, Jan 14, 2013 at 11:43:54AM CET, amwang@redhat.com wrote:
>From: Cong Wang <amwang@redhat.com>
>
>This patch fixes the following warning:
>
>[   72.013864] RTNL: assertion failed at net/core/dev.c (4955)
>[   72.017758] Pid: 668, comm: netpoll-prep-v6 Not tainted 3.8.0-rc1+ #474
>[   72.019582] Call Trace:
>[   72.020295]  [<ffffffff8176653d>] netdev_master_upper_dev_get+0x35/0x58
>[   72.022545]  [<ffffffff81784edd>] netpoll_setup+0x61/0x340
>[   72.024846]  [<ffffffff815d837e>] store_enabled+0x82/0xc3
>[   72.027466]  [<ffffffff815d7e51>] netconsole_target_attr_store+0x35/0x37
>[   72.029348]  [<ffffffff811c3479>] configfs_write_file+0xe2/0x10c
>[   72.030959]  [<ffffffff8115d239>] vfs_write+0xaf/0xf6
>[   72.032359]  [<ffffffff81978a05>] ? sysret_check+0x22/0x5d
>[   72.033824]  [<ffffffff8115d453>] sys_write+0x5c/0x84
>[   72.035328]  [<ffffffff819789d9>] system_call_fastpath+0x16/0x1b
>
>Just hold RCU read lock and call netdev_master_upper_dev_get_rcu(),
>as suggested by Jiri.
>
>Cc: Jiri Pirko <jiri@resnulli.us>
>Cc: David S. Miller <davem@davemloft.net>
>Signed-off-by: Cong Wang <amwang@redhat.com>

Reviewed-by: Jiri Pirko <jiri@resnulli.us>


>
>---
>diff --git a/net/core/netpoll.c b/net/core/netpoll.c
>index 9f05067..5d1f856 100644
>--- a/net/core/netpoll.c
>+++ b/net/core/netpoll.c
>@@ -1055,11 +1055,14 @@ int netpoll_setup(struct netpoll *np)
> 		return -ENODEV;
> 	}
> 
>-	if (netdev_master_upper_dev_get(ndev)) {
>+	rcu_read_lock();
>+	if (netdev_master_upper_dev_get_rcu(ndev)) {
>+		rcu_read_unlock();
> 		np_err(np, "%s is a slave device, aborting\n", np->dev_name);
> 		err = -EBUSY;
> 		goto put;
> 	}
>+	rcu_read_unlock();
> 
> 	if (!netif_running(ndev)) {
> 		unsigned long atmost, atleast;

Thanks Amerigo.

--
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
Eric Dumazet Jan. 14, 2013, 5:07 p.m. UTC | #2 
On Mon, 2013-01-14 at 18:43 +0800, Cong Wang wrote:
> From: Cong Wang <amwang@redhat.com>
> 
> This patch fixes the following warning:
> 
> [   72.013864] RTNL: assertion failed at net/core/dev.c (4955)
> [   72.017758] Pid: 668, comm: netpoll-prep-v6 Not tainted 3.8.0-rc1+ #474
> [   72.019582] Call Trace:
> [   72.020295]  [<ffffffff8176653d>] netdev_master_upper_dev_get+0x35/0x58
> [   72.022545]  [<ffffffff81784edd>] netpoll_setup+0x61/0x340
> [   72.024846]  [<ffffffff815d837e>] store_enabled+0x82/0xc3
> [   72.027466]  [<ffffffff815d7e51>] netconsole_target_attr_store+0x35/0x37
> [   72.029348]  [<ffffffff811c3479>] configfs_write_file+0xe2/0x10c
> [   72.030959]  [<ffffffff8115d239>] vfs_write+0xaf/0xf6
> [   72.032359]  [<ffffffff81978a05>] ? sysret_check+0x22/0x5d
> [   72.033824]  [<ffffffff8115d453>] sys_write+0x5c/0x84
> [   72.035328]  [<ffffffff819789d9>] system_call_fastpath+0x16/0x1b
> 
> Just hold RCU read lock and call netdev_master_upper_dev_get_rcu(),
> as suggested by Jiri.
> 
> Cc: Jiri Pirko <jiri@resnulli.us>
> Cc: David S. Miller <davem@davemloft.net>
> Signed-off-by: Cong Wang <amwang@redhat.com>
> 
> ---
> diff --git a/net/core/netpoll.c b/net/core/netpoll.c
> index 9f05067..5d1f856 100644
> --- a/net/core/netpoll.c
> +++ b/net/core/netpoll.c
> @@ -1055,11 +1055,14 @@ int netpoll_setup(struct netpoll *np)
>  		return -ENODEV;
>  	}
>  
> -	if (netdev_master_upper_dev_get(ndev)) {
> +	rcu_read_lock();
> +	if (netdev_master_upper_dev_get_rcu(ndev)) {
> +		rcu_read_unlock();
>  		np_err(np, "%s is a slave device, aborting\n", np->dev_name);
>  		err = -EBUSY;
>  		goto put;
>  	}
> +	rcu_read_unlock();
>  
>  	if (!netif_running(ndev)) {
>  		unsigned long atmost, atleast;
> --

Strange, why dont we call netdev_master_upper_dev_get() instead ?



--
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
Eric Dumazet Jan. 14, 2013, 5:09 p.m. UTC | #3 
On Mon, 2013-01-14 at 09:07 -0800, Eric Dumazet wrote:

> Strange, why dont we call netdev_master_upper_dev_get() instead ?
> 
> 

What I meant is : the result of this check might be of little use, if we
don't hold any lock preventing another thread to change things behind
us.



--
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
David Miller Jan. 14, 2013, 11:13 p.m. UTC | #4 
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Mon, 14 Jan 2013 09:09:56 -0800

> On Mon, 2013-01-14 at 09:07 -0800, Eric Dumazet wrote:
> 
>> Strange, why dont we call netdev_master_upper_dev_get() instead ?
>> 
>> 
> 
> What I meant is : the result of this check might be of little use, if we
> don't hold any lock preventing another thread to change things behind
> us.

Agreed, this test is going to be racey.

--
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
Amerigo Wang Jan. 15, 2013, 2:23 a.m. UTC | #5 
On Mon, 2013-01-14 at 09:09 -0800, Eric Dumazet wrote:
> On Mon, 2013-01-14 at 09:07 -0800, Eric Dumazet wrote:
> 
> > Strange, why dont we call netdev_master_upper_dev_get() instead ?
> > 
> > 
> 
> What I meant is : the result of this check might be of little use, if we
> don't hold any lock preventing another thread to change things behind
> us.
> 
> 

Even if so, enslaving this device can still happen after we release the
rtnl lock, unless we take this lock for the whole netpoll_setup().

Am I missing anything?

--
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
diff mbox

Patch

diff --git a/net/core/netpoll.c b/net/core/netpoll.c
index 9f05067..5d1f856 100644
--- a/net/core/netpoll.c
+++ b/net/core/netpoll.c
@@ -1055,11 +1055,14 @@  int netpoll_setup(struct netpoll *np)
 		return -ENODEV;
 	}
 
-	if (netdev_master_upper_dev_get(ndev)) {
+	rcu_read_lock();
+	if (netdev_master_upper_dev_get_rcu(ndev)) {
+		rcu_read_unlock();
 		np_err(np, "%s is a slave device, aborting\n", np->dev_name);
 		err = -EBUSY;
 		goto put;
 	}
+	rcu_read_unlock();
 
 	if (!netif_running(ndev)) {
 		unsigned long atmost, atleast;