[PATCH net-2.6.26 3/3][TUN][NETNS]: Allow to register tun devices in namespace.

[PATCH net-2.6.26 3/3][TUN][NETNS]: Allow to register tun devices in namespace.

[Pavel Emelyanov]

This is basically means that a net is set for a new device, but
actually this involves two more steps:

1. mark the tun device as "local", i.e. do not allow for it to
   move across namespaces.

This is done so, since tun device is most often associated to some
file (and thus to some process) and moving the device alone is not
valid while keeping the file and the process outside.

2. get the tun device's net when tun becomes attached and put one
   when it becomes detached.

This is needed to handle the case when a task owning the tun dies,
but a files lives for some more time - in this case we must not
allow for net to be freed, since its exit hook will spoil that file's
private data by unregistering the tun from under tun_chr_close.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>

---
 drivers/net/tun.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index e3210bf..bcc4ced 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -438,6 +438,7 @@ static void tun_setup(struct net_device *dev)
 	dev->stop = tun_net_close;
 	dev->ethtool_ops = &tun_ethtool_ops;
 	dev->destructor = free_netdev;
+	dev->features |= NETIF_F_NETNS_LOCAL;
 }
 
 static struct tun_struct *tun_get_by_name(struct net *net, const char *name)
@@ -503,6 +504,7 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
 		if (!dev)
 			return -ENOMEM;
 
+		dev_net_set(dev, net);
 		tun = netdev_priv(dev);
 		tun->dev = dev;
 		tun->flags = flags;
@@ -542,6 +544,7 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
 
 	file->private_data = tun;
 	tun->attached = 1;
+	get_net(dev_net(tun->dev));
 
 	strcpy(ifr->ifr_name, tun->dev->name);
 	return 0;
@@ -757,6 +760,7 @@ static int tun_chr_close(struct inode *inode, struct file *file)
 	/* Detach from net device */
 	file->private_data = NULL;
 	tun->attached = 0;
+	put_net(dev_net(tun->dev));
 
 	/* Drop read queue */
 	skb_queue_purge(&tun->readq);
-- 
1.5.3.4

Re: [PATCH net-2.6.26 3/3][TUN][NETNS]: Allow to register tun devices in namespace.

[Max Krasnyanskiy]

Hi Pavel,

All three patches look perfectly fine to me.
Feel free to add Acked-by: Max Krasnyansky <maxk@qualcomm.com>

Pavel Emelyanov wrote:
> This is basically means that a net is set for a new device, but
> actually this involves two more steps:
> 
> 1. mark the tun device as "local", i.e. do not allow for it to
>    move across namespaces.
> 
> This is done so, since tun device is most often associated to some
> file (and thus to some process) and moving the device alone is not
> valid while keeping the file and the process outside.
> 
> 2. get the tun device's net when tun becomes attached and put one
>    when it becomes detached.
> 
> This is needed to handle the case when a task owning the tun dies,
> but a files lives for some more time - in this case we must not
> allow for net to be freed, since its exit hook will spoil that file's
> private data by unregistering the tun from under tun_chr_close.
I'm not sure what you mean "by file lives on" here. I believe you're talking 
about persistent tun devices. ie Those that exist in detached state and are 
not attached to any file descriptors.
"net" refcounting logic there looks fine to me.

Max

Re: [PATCH net-2.6.26 3/3][TUN][NETNS]: Allow to register tun devices in namespace.

[Pavel Emelyanov]

Max Krasnyanskiy wrote:
> Hi Pavel,
> 
> All three patches look perfectly fine to me.
> Feel free to add Acked-by: Max Krasnyansky <maxk@qualcomm.com>

:)

> Pavel Emelyanov wrote:
>> This is basically means that a net is set for a new device, but
>> actually this involves two more steps:
>>
>> 1. mark the tun device as "local", i.e. do not allow for it to
>>    move across namespaces.
>>
>> This is done so, since tun device is most often associated to some
>> file (and thus to some process) and moving the device alone is not
>> valid while keeping the file and the process outside.
>>
>> 2. get the tun device's net when tun becomes attached and put one
>>    when it becomes detached.
>>
>> This is needed to handle the case when a task owning the tun dies,
>> but a files lives for some more time - in this case we must not
>> allow for net to be freed, since its exit hook will spoil that file's
>> private data by unregistering the tun from under tun_chr_close.
> I'm not sure what you mean "by file lives on" here. I believe you're talking

I meant the following. When a task dies it releases it's net namespace
and thus this net cleanup job may start. But the file, that holds the
(not yet detached) tun device is not released immediately - so before this
file is being released the tun_exit_net callback may be called and the 
file's tun device will be unregistered in it. I catch this race by getting
the net by attached tun device, so that net is cleaned up only when there
are only detached tuns in it. In this case no tasks will pick up these tuns
and everything will be OK.
 
> about persistent tun devices. ie Those that exist in detached state and are 
> not attached to any file descriptors.
> "net" refcounting logic there looks fine to me.

Great!

> Max

Thanks,
Pavel