kern/163701: ifconfig: group parameter makes impossible to restart interface

Dmitry Kazarov d.y.kazarov at mail.ru
Thu Dec 29 21:20:09 UTC 2011


>Number:         163701
>Category:       kern
>Synopsis:       ifconfig: group parameter makes impossible to restart interface
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Thu Dec 29 21:20:08 UTC 2011
>Closed-Date:
>Last-Modified:
>Originator:     Dmitry Kazarov
>Release:        8.2-STABLE
>Organization:
>Environment:
FreeBSD Trudiaga.localdomain 8.2-STABLE FreeBSD 8.2-STABLE #0: Mon Dec 12 01:42:49 MSK 2011     root at Trudiaga.localdomain:/var/obj/usr/src/sys/GENERIC  i386

>Description:
If group parameter is specified in /etc/rc.conf it makes /etc/rc.d/netif restart to fail:

# fgrep re0 /etc/rc.conf
ifconfig_re0='192.168.1.20/24 group ifgrp'

# ifconfig -v re0
re0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3898<VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,WOL_UCAST,WOL_MCAST,WOL_MAGIC>
        ether 00:1f:d0:5e:8f:57
        inet 192.168.1.20 netmask 0xffffff00 broadcast 192.168.1.255
        media: Ethernet autoselect (1000baseT <full-duplex>)
        status: active
        groups: ifgrp 

# /etc/rc.d/netif restart re0
Stopping Network: re0.
re0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3898<VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,WOL_UCAST,WOL_MCAST,WOL_MAGIC>
        ether 00:1f:d0:5e:8f:57
        media: Ethernet autoselect (none)
        status: no carrier
ifconfig:  SIOCAIFGROUP: File exists
Starting Network: re0.
re0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3898<VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,WOL_UCAST,WOL_MCAST,WOL_MAGIC>
        ether 00:1f:d0:5e:8f:57
        media: Ethernet autoselect (none)
        status: no carrier


The source of problem is that groups parameter persist when interface goes down:

# /etc/rc.d/netif stop re0
Stopping Network: re0.
re0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3898<VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,WOL_UCAST,WOL_MCAST,WOL_MAGIC>
        ether 00:1f:d0:5e:8f:57
        media: Ethernet autoselect (1000baseT <full-duplex>)
        status: active
# ifconfig -v re0
re0: flags=8802<BROADCAST,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=3898<VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,WOL_UCAST,WOL_MCAST,WOL_MAGIC>
        ether 00:1f:d0:5e:8f:57
        media: Ethernet autoselect (1000baseT <full-duplex>)
        status: active
        groups: ifgrp 


IMHO it's better to just ignore error when group param specifies groupname to which interface is already  assigned.
>How-To-Repeat:

>Fix:
--- src/sbin/ifconfig/ifgroup.c.orig    2011-12-30 01:07:54.000000000 +0400
+++ src/sbin/ifconfig/ifgroup.c 2011-12-30 01:09:09.000000000 +0400
@@ -58,7 +58,10 @@
        if (strlcpy(ifgr.ifgr_group, group_name, IFNAMSIZ) >= IFNAMSIZ)
                errx(1, "setifgroup: group name too long");
        if (ioctl(s, SIOCAIFGROUP, (caddr_t)&ifgr) == -1)
-               err(1," SIOCAIFGROUP");
+               if (errno == EEXIST)
+                       return;
+               else
+                       err(1," SIOCAIFGROUP");
 }
 
 /* ARGSUSED */


>Release-Note:
>Audit-Trail:
>Unformatted:


More information about the freebsd-bugs mailing list