Skip to contentSkip to conversation

net: rtnetlink: Support alias interfaces with RTM_GETLINK

PATCH In review Phil Sutter 2 participants 6 messages Feb 7, 2019

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net: rtnetlink: Support alias interfaces with RTM_GETLINK

Align interface name handling regarding alias interfaces in rtnl_getlink() with dev_ioctl() treating SIOCGIFINDEX ioctl calls. The latter function strips any colon suffix before doing the interface lookup, do the same for RTM_GETLINK requests.

---
 net/core/rtnetlink.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
net/core/rtnetlink.c +3 -1
index a51cab95ba64c..aaee4df0ff00c 100644
@@ -3312,8 +3312,10 @@ static int rtnl_getlink(struct sk_buff *skb, struct nlmsghdr *nlh,
 			return PTR_ERR(tgt_net);
 	}
 
-	if (tb[IFLA_IFNAME])
+	if (tb[IFLA_IFNAME]) {
 		nla_strlcpy(ifname, tb[IFLA_IFNAME], IFNAMSIZ);
+		*strchrnul(ifname, ':') = '\0';
+	}
 
 	if (tb[IFLA_EXT_MASK])
 		ext_filter_mask = nla_get_u32(tb[IFLA_EXT_MASK]);
Signature · 1 line
2.20.1
#1 lore raw
Quoted 4 lines · Phil Sutter

Align interface name handling regarding alias interfaces in rtnl_getlink() with dev_ioctl() treating SIOCGIFINDEX ioctl calls. The latter function strips any colon suffix before doing the interface lookup, do the same for RTM_GETLINK requests.

After a second thought, I'll self-NACK this one: Given that netlink API is completely unrelated to ioctl one, there is no inherent need to do things the same way. Looking at RTM_NEWLINK handler, it seems possible to create interface names containing a colon via netlink. Of course those interfaces are not accessible via ioctl() then, but so are secondary interface addresses which don't have a properly chosen label.

Sorry for the noise, Phil

#2 lore raw
Quoted 10 lines · Phil Sutter
Quoted 4 lines · Phil Sutter

Align interface name handling regarding alias interfaces in rtnl_getlink() with dev_ioctl() treating SIOCGIFINDEX ioctl calls. The latter function strips any colon suffix before doing the interface lookup, do the same for RTM_GETLINK requests.

After a second thought, I'll self-NACK this one: Given that netlink API is completely unrelated to ioctl one, there is no inherent need to do things the same way. Looking at RTM_NEWLINK handler, it seems possible to create interface names containing a colon via netlink.

Not since commit a4176a939186 ("net: reject creation of netdev names with colons") which disallowed using such names in general.

But I still don't think it would be a good idea. It's bad enough that (as I just learned to my surprise) "ip link del dummy1:0" deletes dummy1 without any complaint because ip uses SIOCGIFINDEX ioctl for ifindex lookup.

Michal Kubecek

#3 lore raw

Hi,

Quoted 7 lines · Michal Kubecek
Quoted 6 lines · Phil Sutter
Quoted 4 lines · Phil Sutter

Align interface name handling regarding alias interfaces in rtnl_getlink() with dev_ioctl() treating SIOCGIFINDEX ioctl calls. The latter function strips any colon suffix before doing the interface lookup, do the same for RTM_GETLINK requests.

Quoted 7 lines
Quoted 4 lines

After a second thought, I'll self-NACK this one: Given that netlink API is completely unrelated to ioctl one, there is no inherent need to do things the same way. Looking at RTM_NEWLINK handler, it seems possible to create interface names containing a colon via netlink.

Not since commit a4176a939186 ("net: reject creation of netdev names with colons") which disallowed using such names in general.

In my typical afterthought I tried 'ip link add d0:1 type dummy' and was surprised to get EINVAL from kernel side. Just discovered that line in dev_valid_name(), too.

Quoted 4 lines

But I still don't think it would be a good idea. It's bad enough that (as I just learned to my surprise) "ip link del dummy1:0" deletes dummy1 without any complaint because ip uses SIOCGIFINDEX ioctl for ifindex lookup.

I'm struggling a bit with all this. The original problem is iproute2 commit 50b9950dd9011 ("link dump filter") which changed 'ip link show' to not use if_indextoname() when given just an interface name. So lookup happens by name (via RTM_GETLINK) and consequently 'ip link show eth0:1' doesn't give link stats of eth0 anymore.

Given that iproute2 is supposed to be backwards compatible, the only valid option I see is to make sure netlink API calls like the above behave identical to the ioctl ones they replace. Which means allowing for 'ip link show eth0:42' even if there's no address with that label assigned to eth0 as well as your example above.

Cheers, Phil

#4 lore raw
Quoted 7 lines · Phil Sutter
Quoted 4 lines · Michal Kubecek

But I still don't think it would be a good idea. It's bad enough that (as I just learned to my surprise) "ip link del dummy1:0" deletes dummy1 without any complaint because ip uses SIOCGIFINDEX ioctl for ifindex lookup.

Quoted 5 lines

I'm struggling a bit with all this. The original problem is iproute2 commit 50b9950dd9011 ("link dump filter") which changed 'ip link show' to not use if_indextoname() when given just an interface name. So lookup happens by name (via RTM_GETLINK) and consequently 'ip link show eth0:1' doesn't give link stats of eth0 anymore.

I would rather consider it a bug that it ever did. It's quite harmless with "show" but with "set" or "delete", the effect can be quite disastrous.

We want to preserve backward compatibility in general but iproute2 commit 50b9950dd901 is 4.5 years old and nobody seems to have complained about the change in behaviour until now.

Quoted 5 lines

Given that iproute2 is supposed to be backwards compatible, the only valid option I see is to make sure netlink API calls like the above behave identical to the ioctl ones they replace. Which means allowing for 'ip link show eth0:42' even if there's no address with that label assigned to eth0 as well as your example above.

One reason why I don't like this idea is that iproute2 is not the only user of rtnetlink interface. There is wicked and glibc for sure, most likely also NetworkManager (don't remember) and systemd-networkd (didn't check) and certainly many others I never heard of. Changing the logic in kernel rtnetlink implementation would affect all of them.

If we want to restore the old behaviour of ip (which I'm not convinced of), it would make more sense to me to strip the :* suffix in iproute2.

Michal Kubecek

#5 lore raw
Quoted 14 lines · Michal Kubecek
Quoted 7 lines · Phil Sutter
Quoted 4 lines · Michal Kubecek

But I still don't think it would be a good idea. It's bad enough that (as I just learned to my surprise) "ip link del dummy1:0" deletes dummy1 without any complaint because ip uses SIOCGIFINDEX ioctl for ifindex lookup.

Quoted 5 lines

I'm struggling a bit with all this. The original problem is iproute2 commit 50b9950dd9011 ("link dump filter") which changed 'ip link show' to not use if_indextoname() when given just an interface name. So lookup happens by name (via RTM_GETLINK) and consequently 'ip link show eth0:1' doesn't give link stats of eth0 anymore.

Quoted 3 lines

I would rather consider it a bug that it ever did. It's quite harmless with "show" but with "set" or "delete", the effect can be quite disastrous.

Yes, you're right. And unless maintainers care about this compatibility, I guess these monsters will slowly disappear as things progress towards netlink.

Quoted 3 lines

We want to preserve backward compatibility in general but iproute2 commit 50b9950dd901 is 4.5 years old and nobody seems to have complained about the change in behaviour until now.

Well, enterprise distributions deliberately try to prevent those changes from hitting their customers. And (perfect match), these users are the least tolerating ones. :)

Quoted 6 lines
Quoted 5 lines

Given that iproute2 is supposed to be backwards compatible, the only valid option I see is to make sure netlink API calls like the above behave identical to the ioctl ones they replace. Which means allowing for 'ip link show eth0:42' even if there's no address with that label assigned to eth0 as well as your example above.

Quoted 8 lines

One reason why I don't like this idea is that iproute2 is not the only user of rtnetlink interface. There is wicked and glibc for sure, most likely also NetworkManager (don't remember) and systemd-networkd (didn't check) and certainly many others I never heard of. Changing the logic in kernel rtnetlink implementation would affect all of them.

If we want to restore the old behaviour of ip (which I'm not convinced of), it would make more sense to me to strip the :* suffix in iproute2.

I just sent a patch, let's see what userspace has to say about it.

Thanks, Phil