modify: don't send IFLA_INFO_KIND for Device types

[aojea]

Device links are links created by udev and are not a supported Type, sending the "device" string as IFLA_INFO_KIND throws an -EOPNOTSUPP error.

Tested witht the followed code against a phisical interface

package main

import (
	"fmt"
	"log"
	"os"

	"github.com/vishvananda/netlink"
	"github.com/vishvananda/netns"
)

func main() {

	if len(os.Args) != 3 {
		fmt.Printf("Usage: %s <interface> <namespace>\n", os.Args[0])
		os.Exit(1)
	}

	ifName := os.Args[1]
	ns := os.Args[2]

	link, err := netlink.LinkByName(ifName)
	if err != nil {
		log.Fatal("linkbyname failed", err)
	}

	err = netlink.LinkSetDown(link)
	if err != nil {
		log.Fatal("LinkSetDown failed", err)
	}

	attrs := netlink.NewLinkAttrs()
	attrs.Index = link.Attrs().Index
	attrs.MTU = link.Attrs().MTU
	attrs.Name = link.Attrs().Name

	containerNs, err := netns.GetFromPath(ns)
	if err != nil {
		log.Fatal("container from path", err)
	}
	attrs.Namespace = netlink.NsFd(containerNs)
	dev := &netlink.Device{
		LinkAttrs: attrs,
	}
	err = netlink.LinkModify(dev)
	if err != nil {
		log.Fatal("link modify", err)
	}
}

Without the patch:

link-setns eth3 /var/run/netns/cni-e79e4127-2e60-56fd-4ade-14b420c65de3
2024/12/17 12:53:38 operation not supported

With this patch:

$ link-setns eth4 /var/run/netns/cni-e79e4127-2e60-56fd-4ade-14b420c65de3
$ echo $?
0
$  ip netns exec cni-e79e4127-2e60-56fd-4ade-14b420c65de3  ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
2: eth0@if10: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1460 qdisc noqueue state UP group default
    link/ether f6:47:24:b8:93:44 brd ff:ff:ff:ff:ff:ff link-netnsid 0
    inet 10.52.3.110/24 brd 10.52.3.255 scope global eth0
       valid_lft forever preferred_lft forever
6: eth4: <BROADCAST,MULTICAST> mtu 8244 qdisc noop state DOWN group default qlen 1000
    link/ether 42:01:c0:a8:04:02 brd ff:ff:ff:ff:ff:ff

[aojea]

failed test is

    --- FAIL: TestRuleListFiltered/IPv4/returns_rules_filtered_by_fwmark_0xFFFFFFFF/0 (0.00s)

does not seem related, does it?

[aojea]

hmm, now it failed

    --- FAIL: TestRuleListFiltered/IPv6/returns_rules_with_specific_priority (0.00s)

two time in a row but this patch is not exercising this code

[aojea]

it failed again

    --- FAIL: TestRuleListFiltered/IPv4/returns_rules_filtered_by_fwmark_0x1234 (0.00s)

let me try locally now I'm worried I'm doing something wrong