new link failing on duplicate names in different namespaces
Subsystems (2)
Hello,
rtnetlink's RTM_NEWLINK allows for specifying the network namespace in which the link is to be created via IFLA_NET_NS_PID or IFLA_NET_NS_FD. This not only saves a user a call to move the link into the target namespace, it can avoid a potential ifname conflict. For example, if creating eth0 in another namespace and the current one already has eth0.
Unfortunately, this is not the current behavior. If the user specifies IFLA_IFNAME, leaves ifinfomsg.ifi_index unspecified and sets NLM_F_EXCL flag, as in the case of creating a new link, the call will fail with EEXIST in cases where there's a name conflict.
rtnl_newlink() will:
if (ifname[0])
dev = __dev_get_by_name(net, ifname);
and later:
if (dev) {
int status = 0;
if (nlh->nlmsg_flags & NLM_F_EXCL)
return -EEXIST;...
}
I am not sure if this has been discussed before or if it's viewed as a real problem. Fixing this however, would make it easier to work with interfaces in different network namespaces. I don't have a good solution in mind though and so I can't propose a patch at this time.
Thanks,
Eugene
Quoted 14 lines · Eugene Yakubovich
Hello,
rtnetlink's
RTM_NEWLINKallows for specifying the network namespace in which the link is to be created viaIFLA_NET_NS_PIDorIFLA_NET_NS_FD. This not only saves a user a call to move the link into the target namespace, it can avoid a potential ifname conflict. For example, if creating eth0 in another namespace and the current one already has eth0.Unfortunately, this is not the current behavior. If the user specifies
IFLA_IFNAME, leaves ifinfomsg.ifi_index unspecified and setsNLM_F_EXCLflag, as in the case of creating a new link, the call will fail with EEXIST in cases where there's a name conflict.
Quoted 4 lines
rtnl_newlink()will:if (ifname[0])
dev = __dev_get_by_name(net, ifname);
Yes, it seems that this should be done in "dest_net" or "link_net".
Quoted 15 lines · Nicolas Dichtel
Quoted 14 lines · Eugene Yakubovich
Hello,
rtnetlink's
RTM_NEWLINKallows for specifying the network namespace in which the link is to be created viaIFLA_NET_NS_PIDorIFLA_NET_NS_FD. This not only saves a user a call to move the link into the target namespace, it can avoid a potential ifname conflict. For example, if creating eth0 in another namespace and the current one already has eth0.Unfortunately, this is not the current behavior. If the user specifies
IFLA_IFNAME, leaves ifinfomsg.ifi_index unspecified and setsNLM_F_EXCLflag, as in the case of creating a new link, the call will fail with EEXIST in cases where there's a name conflict.
Quoted 5 lines
Quoted 4 lines
rtnl_newlink()will:if (ifname[0])
dev = __dev_get_by_name(net, ifname);Yes, it seems that this should be done in "
dest_net" or "link_net".
Ugh. Looking at that code I think the link_net is at best extremely confusing most likely semantically broken the way it is being interpreted in rtnl_newlink. More in another email message.
The original semantics and what seems mostly reasonable because the network device in some sense lives in multiple network namespaces as it is being created is that the network device name be unique in both namespaces.
The code that creates a network devices fundamentally needs to know about both namespaces network devices like the macvlan driver need to find their lower devices in the original network namespace.
I think we could possibly remove the restriction for the names of newly created network devices being unique in both network namespaces but I don't think the code changes will be trivial.
Eric
Quoted 9 lines · Eric W. Biederman
The code that creates a network devices fundamentally needs to know about both namespaces network devices like the macvlan driver need to find their lower devices in the original network namespace.
I think we could possibly remove the restriction for the names of newly created network devices being unique in both network namespaces but I don't think the code changes will be trivial.
We should add a flag to network devices, so that some devices which don't have an underlying device, like veth, are allowed to be create right in the given netns. Or, of course, just modify rtnl_newlink() if we have some other clue (tb[IFLA_LINK] == NULL?).
I am working on a patch.
Quoted 12 lines · Cong Wang
Quoted 9 lines · Eric W. Biederman
The code that creates a network devices fundamentally needs to know about both namespaces network devices like the macvlan driver need to find their lower devices in the original network namespace.
I think we could possibly remove the restriction for the names of newly created network devices being unique in both network namespaces but I don't think the code changes will be trivial.
Quoted 4 lines
We should add a flag to network devices, so that some devices which don't have an underlying device, like veth, are allowed to be create right in the given netns. Or, of course, just modify
rtnl_newlink()if we have some other clue (tb[IFLA_LINK] == NULL?).
Never mind, seems we only need to move the dest_net and link_net up before checking duplicated.
Please give the attached patch a try.
net/core/rtnetlink.c +73 -51
@@ -1961,6 +1961,7 @@ static int rtnl_newlink(struct sk_buff *skb, struct nlmsghdr *nlh)struct nlattr *tb[IFLA_MAX +1 ];struct nlattr *linkinfo[IFLA_INFO_MAX +1 ];unsigned char name_assign_type =NET_NAME_USER ; +struct net *dest_net, *link_net =NULL ;int err;#ifdef CONFIG_MODULES
@@ -1976,11 +1977,26 @@ replay: ifname[0 ] ='\0' ; ifm =nlmsg_data (nlh); + + dest_net =rtnl_link_get_net (net, tb); +if (IS_ERR (dest_net)) +return PTR_ERR (dest_net); + +if (tb[IFLA_LINK_NETNSID ]) { +int id =nla_get_s32 (tb[IFLA_LINK_NETNSID ]); + + link_net =get_net_ns_by_id (dest_net, id); +if (!link_net) { + err = -EINVAL ; + goto out; + } + } +if (ifm->ifi_index >0 ) - dev =__dev_get_by_index (net, ifm->ifi_index); + dev =__dev_get_by_index (link_net ?: dest_net, ifm->ifi_index);else {if (ifname[0 ]) - dev =__dev_get_by_name (net, ifname); + dev =__dev_get_by_name (link_net ?: dest_net, ifname);else dev =NULL ; }
@@ -1993,13 +2009,13 @@ replay: err =validate_linkmsg (dev, tb);if (err <0 ) -return err; + goto err;if (tb[IFLA_LINKINFO ]) { err =nla_parse_nested (linkinfo,IFLA_INFO_MAX , tb[IFLA_LINKINFO ], ifla_info_policy);if (err <0 ) -return err; + goto err; }else memset (linkinfo,0 ,sizeof (linkinfo));
@@ -2016,7 +2032,6 @@ replay:struct nlattr *slave_attr[m_ops ? m_ops->slave_maxtype +1 :1 ];struct nlattr **data =NULL ;struct nlattr **slave_data =NULL ; -struct net *dest_net, *link_net =NULL ;if (ops) {if (ops->maxtype && linkinfo[IFLA_INFO_DATA ]) {
@@ -2024,13 +2039,13 @@ replay: linkinfo[IFLA_INFO_DATA ], ops->policy);if (err <0 ) -return err; + goto err; data = attr; }if (ops->validate) { err = ops->validate (tb, data);if (err <0 ) -return err; + goto err; } }
@@ -2042,59 +2057,71 @@ replay: linkinfo[IFLA_INFO_SLAVE_DATA ], m_ops->slave_policy);if (err <0 ) -return err; + goto err; slave_data = slave_attr; }if (m_ops->slave_validate) { err = m_ops->slave_validate (tb, slave_data);if (err <0 ) -return err; + goto err; } }if (dev) {int status =0 ; -if (nlh->nlmsg_flags &NLM_F_EXCL ) -return -EEXIST ; -if (nlh->nlmsg_flags &NLM_F_REPLACE ) -return -EOPNOTSUPP ; +if (nlh->nlmsg_flags &NLM_F_EXCL ) { + err = -EEXIST ; + goto err; + } +if (nlh->nlmsg_flags &NLM_F_REPLACE ) { + err = -EOPNOTSUPP ; + goto err; + }if (linkinfo[IFLA_INFO_DATA ]) {if (!ops || ops != dev->rtnl_link_ops || - !ops->changelink) -return -EOPNOTSUPP ; + !ops->changelink) { + err = -EOPNOTSUPP ; + goto err; + } err = ops->changelink (dev, tb, data);if (err <0 ) -return err; + goto err; status |=DO_SETLINK_NOTIFY ; }if (linkinfo[IFLA_INFO_SLAVE_DATA ]) { -if (!m_ops || !m_ops->slave_changelink) -return -EOPNOTSUPP ; +if (!m_ops || !m_ops->slave_changelink) { + err = -EOPNOTSUPP ; + goto err; + } err = m_ops->slave_changelink (master_dev, dev, tb, slave_data);if (err <0 ) -return err; + goto err; status |=DO_SETLINK_NOTIFY ; } -return do_setlink (skb, dev, ifm, tb, ifname, status); + err =do_setlink (skb, dev, ifm, tb, ifname, status); + goto err; }if (!(nlh->nlmsg_flags &NLM_F_CREATE )) { + err = -ENODEV ;if (ifm->ifi_index ==0 && tb[IFLA_GROUP ]) -return rtnl_group_changelink (skb, net, + err =rtnl_group_changelink (skb, dest_net,nla_get_u32 (tb[IFLA_GROUP ]), ifm, tb); -return -ENODEV ; + goto err; } -if (tb[IFLA_MAP ] || tb[IFLA_MASTER ] || tb[IFLA_PROTINFO ]) -return -EOPNOTSUPP ; +if (tb[IFLA_MAP ] || tb[IFLA_MASTER ] || tb[IFLA_PROTINFO ]) { + err = -EOPNOTSUPP ; + goto err; + }if (!ops) {#ifdef CONFIG_MODULES
@@ -2103,40 +2130,33 @@ replay:request_module ("rtnl-link-%s" , kind);rtnl_lock (); ops =rtnl_link_ops_get (kind); -if (ops) +if (ops) { +if (link_net) +put_net (link_net); +put_net (dest_net); goto replay; + } }#endif -return -EOPNOTSUPP ; + err = -EOPNOTSUPP ; + goto err; } -if (!ops->setup) -return -EOPNOTSUPP ; +if (!ops->setup) { + err = -EOPNOTSUPP ; + goto err; + }if (!ifname[0 ]) {snprintf (ifname,IFNAMSIZ ,"%s%%d" , ops->kind); name_assign_type =NET_NAME_ENUM ; } - dest_net =rtnl_link_get_net (net, tb); -if (IS_ERR (dest_net)) -return PTR_ERR (dest_net); - -if (tb[IFLA_LINK_NETNSID ]) { -int id =nla_get_s32 (tb[IFLA_LINK_NETNSID ]); - - link_net =get_net_ns_by_id (dest_net, id); -if (!link_net) { - err = -EINVAL ; - goto out; - } - } - dev =rtnl_create_link (link_net ? : dest_net, ifname, name_assign_type, ops, tb);if (IS_ERR (dev)) { err =PTR_ERR (dev); - goto out; + goto err; } dev->ifindex = ifm->ifi_index;
@@ -2151,13 +2171,13 @@ replay:/* If device is not registered at all, free it now */ if (dev->reg_state ==NETREG_UNINITIALIZED )free_netdev (dev); - goto out; + goto err; } }else { err =register_netdevice (dev);if (err <0 ) {free_netdev (dev); - goto out; + goto err; } } err =rtnl_configure_link (dev, ifm);
@@ -2170,7 +2190,7 @@ replay: }else {unregister_netdevice (dev); } - goto out; + goto err; }if (link_net) {
@@ -2178,12 +2198,14 @@ replay:if (err <0 )unregister_netdevice (dev); } -out: -if (link_net) -put_net (link_net); -put_net (dest_net); -return err; } + +err: +if (link_net) +put_net (link_net); +out: +put_net (dest_net); +return err; }static int rtnl_getlink (struct sk_buff *skb,struct nlmsghdr * nlh)
Quoted 1 line · Cong Wang
On Wed, Feb 25, 2015 at 9:29 PM, Cong Wang <[email protected]> wrote:
[snip]
Quoted 3 lines
Please give the attached patch a try.
It's hard to comment a patch which is sent in attachment
(see http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/SubmittingPatches#n333).
+ err = rtnl_group_changelink(skb, dest_net,
It should probably be something like 'link_net ? : dest_net' instead of dest_net.
I also don't understand why you need two label ('out' and 'err'). I think one is enough (link_net is initialized to NULL for this purpose). And why not keeping the same name as before, ie 'out'? It will minimize the
patch.
Quoted 14 lines · Nicolas Dichtel
Quoted 1 line · Cong Wang
On Wed, Feb 25, 2015 at 9:29 PM, Cong Wang <[email protected]> wrote:
[snip]
Quoted 3 lines
Please give the attached patch a try.
It's hard to comment a patch which is sent in attachment (see
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/SubmittingPatches#n333).net/core/rtnetlink.c +1
-0+ err = rtnl_group_changelink(skb, dest_net,It should probably be something like '
link_net? :dest_net' instead ofdest_net.
Quoted 4 lines
I also don't understand why you need two label ('out' and 'err'). I think one is enough (
link_netis initialized to NULL for this purpose).
And why not keeping the same name as before, ie 'out'? It will minimize the patch.
Sigh.
We can not give the guarantee that a new network device will only live in and have a unique name in a single network namespace until we get rid of dev_change_net_namespace in the code.
Cong your patch does not get rid of that and so is insufficient to solve this problem and buggy.
Eric
Le 26/02/2015 14:55, Eric W. Biederman a écrit :
[snip]
Quoted 4 lines
We can not give the guarantee that a new network device will only live in and have a unique name in a single network namespace until we get rid of
dev_change_net_namespacein the code.
In fact in case of IFLA_LINK_NETNSID, with Cong's patch, we will first check that the name is unique in link_net and after dev_change_net_namespace() will check that the name is also unique in the final net, else the command will fail (and the device is not created).
Quoted 4 lines · Nicolas Dichtel
Quoted 2 lines · Cong Wang
On Wed, Feb 25, 2015 at 9:29 PM, Cong Wang <[email protected]> wrote:
Quoted 9 lines
[snip]
Quoted 4 lines
Please give the attached patch a try.
It's hard to comment a patch which is sent in attachment (see
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/SubmittingPatches#n333).
I knew, it is just not yet ready to submit formally. :)
Quoted 3 lines
+ err = rtnl_group_changelink(skb,
dest_net,
It should probably be something like 'link_net? :dest_net' instead ofdest_net.
I think it should be "link_net ?: net", since it is the source net. And yes we need to pass the dest_net to this function as well as do_setlink() too.
Quoted 5 lines
I also don't understand why you need two label ('out' and 'err'). I think one is enough (
link_netis initialized to NULL for this purpose). And why not keeping the same name as before, ie 'out'? It will minimize the patch.
Ah, right, I missed it.
Thanks!
What are the intended semantics of IFLA_LINK_NETNSID?
In reply to Nicolas DichtelReading through the code of rtnl_newlink I am perplexed with what your intended semantics of IFLA_LINK_NETNSID are supposed to be.
My expectation was something with the same semantics IFLA_NET_NS_PID and IFLA_NET_NS_FD just a different data representation and something that could be used in more netlink messages, so you could report the network namespace where the other end of a tunnel or the the network namespace of an underlying device is.
Being very slow I would expect that ILFA_LINK_NETNSID would replace dest_net or possibly net in rtnl_newlink but it does not replace either of those completely. Which causes me to think that the implementation of IFLA_LINK_NETNSID in rtnl_newlink is broken.
I suspect the correct fix for rtnl_newlink is to just use IFLA_LINK_NETNSID in rtnl_link_get_net and have it be an alternative way of setting dest_net. But you may intend some different semantics that I don't understand.
Eric
Re: What are the intended semantics of IFLA_LINK_NETNSID?
Quoted 5 lines · Eric W. Biederman
I suspect the correct fix for
rtnl_newlinkis to just useIFLA_LINK_NETNSIDinrtnl_link_get_netand have it be an alternative way of settingdest_net. But you may intend some different semantics that I don't understand.
Looking at the example provided in iproute2, link-netnsid should refer the the netns, by the given ID, where the underlying link is. The difference is that this ID is only unique in a given netns, and refers to another peer netns.
Re: What are the intended semantics of IFLA_LINK_NETNSID?
In reply to Eric W. BiedermanQuoted 15 lines · Eric W. Biederman
Reading through the code of
rtnl_newlinkI am perplexed with what your intended semantics ofIFLA_LINK_NETNSIDare supposed to be.My expectation was something with the same semantics
IFLA_NET_NS_PIDandIFLA_NET_NS_FDjust a different data representation and something that could be used in more netlink messages, so you could report the network namespace where the other end of a tunnel or the the network namespace of an underlying device is.Being very slow I would expect that
ILFA_LINK_NETNSIDwould replacedest_netor possibly net inrtnl_newlinkbut it does not replace either of those completely. Which causes me to think that the implementation ofIFLA_LINK_NETNSIDinrtnl_newlinkis broken.
Quoted 4 lines
I suspect the correct fix for
rtnl_newlinkis to just useIFLA_LINK_NETNSIDinrtnl_link_get_netand have it be an alternative way of settingdest_net. But you may intend some different semantics that I don't understand.
ILFA_LINK_NETNSID is used to point to the i/o netns of the interface, ie the opposite netns of dest_net.
The interface is first created in link_net and moved at the end in dest_net.
IP tunnels interfaces (ipip, sit, ip6_tunnels, gre[v6]) does not use src_net, thus when you create an ipip interface by specifying an attribute IFLA_NET_NS_PID it will result to an interface which is not across two netns but only in the netns pointed by IFLA_NET_NS_PID. If you use IFLA_LINK_NETNSID, it allows you to create this kind of x-netns interfaces.
In fact, the goal of this attribute is to replace the two following command by only one:
ip link add foo ...
ip link set foo netns bar
=> ip link add foo link-netnsid barID
Regards,
Nicolas
Re: What are the intended semantics of IFLA_LINK_NETNSID?
Quoted 16 lines · Nicolas Dichtel
Quoted 15 lines · Eric W. Biederman
Reading through the code of
rtnl_newlinkI am perplexed with what your intended semantics ofIFLA_LINK_NETNSIDare supposed to be.My expectation was something with the same semantics
IFLA_NET_NS_PIDandIFLA_NET_NS_FDjust a different data representation and something that could be used in more netlink messages, so you could report the network namespace where the other end of a tunnel or the the network namespace of an underlying device is.Being very slow I would expect that
ILFA_LINK_NETNSIDwould replacedest_netor possibly net inrtnl_newlinkbut it does not replace either of those completely. Which causes me to think that the implementation ofIFLA_LINK_NETNSIDinrtnl_newlinkis broken.
Quoted 6 lines
Quoted 4 lines
I suspect the correct fix for
rtnl_newlinkis to just useIFLA_LINK_NETNSIDinrtnl_link_get_netand have it be an alternative way of settingdest_net. But you may intend some different semantics that I don't understand.
ILFA_LINK_NETNSIDis used to point to the i/o netns of the interface, ie the opposite netns ofdest_net.
Fair enough. That makes sense given how it is used in the reporting. With those semantics the code may arguably be correct.
Quoted 3 lines
The interface is first created in
link_netand moved at the end indest_net.IP tunnels interfaces (ipip, sit,
ip6_tunnels, gre[v6]) does not usesrc_net,
I am pretty certain that is simply something that was overlooked when cross network namespace support was added to those network device types. Right now I would be surprised if anything in userspace cares, so we can probably just change those device types to look at src_net from newlink.
Certainly for our sanity in maintaining rtnl_newlink finding a way to make that change would be preferable. Even if we ultimately have to add a flag that says only make src_net different from dev->net when IFLA_LINK_NETNSID is passed.
As making that change allows much more consistency in the code and allows us to get rid of an unnecessary dev_change_net_namespace.
Quoted 9 lines
thus when you create an ipip interface by specifying an attribute
IFLA_NET_NS_PIDit will result to an interface which is not across two netns but only in the netns pointed byIFLA_NET_NS_PID. If you useIFLA_LINK_NETNSID, it allows you to create this kind of x-netns interfaces.
In fact, the goal of this attribute is to replace the two following command by only one:
ip link add foo ...
ip link set foo netns bar
=> ip link add foo link-netnsid barID
Eric
Re: What are the intended semantics of IFLA_LINK_NETNSID?
Quoted 1 line · Eric W. Biederman
Nicolas Dichtel <[email protected]> writes:
[snip]
Quoted 8 lines
Quoted 3 lines
The interface is first created in
link_netand moved at the end indest_net.IP tunnels interfaces (ipip, sit,
ip6_tunnels, gre[v6]) does not usesrc_net,I am pretty certain that is simply something that was overlooked when cross network namespace support was added to those network device types. Right now I would be surprised if anything in userspace cares, so we can probably just change those device types to look at
src_netfrom newlink.
At least, depending of the interface type, the behavior is not the same. Thus, making all interfaces consistent seems good.
Quoted 5 lines
Certainly for our sanity in maintaining
rtnl_newlinkfinding a way to make that change would be preferable. Even if we ultimately have to add a flag that says only makesrc_netdifferent fromdev->netwhenIFLA_LINK_NETNSIDis passed.
A flag is probably not needed, it's just a matter of passing src_net through the functions that create the tunnel device.
Quoted 3 lines
As making that change allows much more consistency in the code and allows us to get rid of an unnecessary
dev_change_net_namespace.
Yes, I add this dev_change_net_namespace() to be independant of the interface implementation (and thus beeing consistent whatever the interface type is used).
Nicolas
PATCH 1/2 net: Verify permission to dest_net in newlink
When applicable verify that the caller has permision to create a network device in another network namespace. This check is already present when moving a network device between network namespaces in setlink so all that is needed is to duplicate that check in newlink.
This change almost backports cleanly, but there are context conflicts as the code that follows was added in v4.0-rc1
--- net/core/rtnetlink.c | 4 ++++ 1 file changed, 4 insertions(+)
net/core/rtnetlink.c +4 -0
@@ -2122,6 +2122,10 @@ replay:if (IS_ERR (dest_net))return PTR_ERR (dest_net); + err = -EPERM ; +if (!netlink_ns_capable (skb, dest_net->user_ns,CAP_NET_ADMIN )) + goto out; +if (tb[IFLA_LINK_NETNSID ]) {int id =nla_get_s32 (tb[IFLA_LINK_NETNSID ]);
Signature · 1 line
2.2.1
PATCH 2/2 net: Verify permission to link_net in newlink
When applicable verify that the caller has permisson to the underlying network namespace for a newly created network device.
Similary checks exist for the network namespace a network device will be created in.
--- net/core/rtnetlink.c | 3 +++ 1 file changed, 3 insertions(+)
net/core/rtnetlink.c +3 -0
@@ -2134,6 +2134,9 @@ replay: err = -EINVAL ; goto out; } + err = -EPERM ; +if (!netlink_ns_capable (skb, link_net->user_ns,CAP_NET_ADMIN )) + goto out; } dev = rtnl_create_link(link_net ? : dest_net, ifname,
Signature · 1 line
2.2.1
Re: [PATCH net 2/2] net: Verify permission to link_net in newlink
Quoted 8 lines · Eric W. Biederman
When applicable verify that the caller has permisson to the underlying network namespace for a newly created network device.
Similary checks exist for the network namespace a network device will be created in.
Fixes v4.0-rc1
Re: [PATCH net 2/2] net: Verify permission to link_net in newlink
In reply to Eric W. BiedermanFrom: [email protected] (Eric W. Biederman)
Date: Thu, 26 Feb 2015 16:20:07 -0600
Quoted 8 lines
When applicable verify that the caller has permisson to the underlying network namespace for a newly created network device.
Similary checks exist for the network namespace a network device will be created in.
Fixes v4.0-rc1Signed-off-by "Eric W. Biederman" <[email protected]>
Applied.
Re: [PATCH net 1/2] net: Verify permission to dest_net in newlink
In reply to Eric W. BiedermanQuoted 10 lines · Eric W. Biederman
When applicable verify that the caller has permision to create a network device in another network namespace. This check is already present when moving a network device between network namespaces in setlink so all that is needed is to duplicate that check in newlink.
This change almost backports cleanly, but there are context conflicts as the code that follows was added in v4.0-rc1
Fixes b51642f6d77b131dc85d1d71029c3cbb5b07c262 net: Enable a userns root rtnl calls that are safe for unprivilged users
12 digits is enough for the sha1 and commit title should be formatted like this
("commit title")
See http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/SubmittingPatches#n187
Re: [PATCH net 1/2] net: Verify permission to dest_net in newlink
In reply to Eric W. BiedermanFrom: [email protected] (Eric W. Biederman)
Date: Thu, 26 Feb 2015 16:19:00 -0600
Quoted 10 lines
When applicable verify that the caller has permision to create a network device in another network namespace. This check is already present when moving a network device between network namespaces in setlink so all that is needed is to duplicate that check in newlink.
This change almost backports cleanly, but there are context conflicts as the code that follows was added in v4.0-rc1
Fixes b51642f6d77b131dc85d1d71029c3cbb5b07c262 net: Enable a userns root rtnl calls that are safe for unprivilged usersSigned-off-by "Eric W. Biederman" <[email protected]>
Applied.