You may be right - this doesn't necessarily sound OSol-specific (although it may have played a part in creating the environment for the hang).

On a plain snv system, I see this:

5605/1: ioctl(6, SIOCGTUNPARAM, 0x08073670) Err#122 EOPNOTSUPP

So, the question is why your system is sleeping in that ioctl. I would suggest either some dtrace or mdb to see where it got to. Something like this:

   # echo "::pgrep ifconfig | ::walk thread | ::findstack" | mdb -k

This will locate all ifconfig processes and should dump a kernel stack trace for each one. If there are multiple, find the one with "ioctl" somewhere in the stack, then search the bug databases for a similar stack trace. I can't spot anything based on the ioctl name, but you may find something against the stack trace.

You may be better taking this to network-discuss.
Do you use IPsec/punchin/vpn software? Could it be that an unclean disconnect left some stale information somewhere?

HTH
Brian


Brian Banister wrote:
On my most recent reboot, ifconfig worked for several hours.  I let ifconfig 
run over night to verify that it would not ever complete.  The tail end of 
truss shows the following:

so_socket(PF_INET6, SOCK_DGRAM, IPPROTO_IP, 0x00000000, SOV_XPG4_2) = 3
so_socket(PF_INET, SOCK_DGRAM, IPPROTO_IP, 0x00000000, SOV_XPG4_2) = 6
ioctl(6, SIOCGLIFFLAGS, 0x080728A0)             = 0
close(6)                                        = 0
close(3)                                        = 0
so_socket(PF_INET, SOCK_DGRAM, IPPROTO_IP, 0x00000000, SOV_XPG4_2) = 3
ioctl(3, SIOCGLIFADDR, 0x08047B30)              = 0
ioctl(3, SIOCGLIFFLAGS, 0x080728A0)             = 0
ioctl(3, SIOCGLIFFLAGS, 0x080728A0)             = 0
ioctl(3, SIOCGLIFMETRIC, 0x080728A0)            = 0
ioctl(3, SIOCGLIFMTU, 0x080728A0)               = 0
ioctl(3, SIOCGLIFINDEX, 0x080728A0)             = 0
ioctl(3, SIOCGLIFZONE, 0x080728A0)              = 0
zone_lookup(0x00000000)                         = 0
ioctl(3, SIOCGLIFINDEX, 0x080728A0)             = 0
ioctl(3, SIOCGLIFSRCOF, 0x08047A40)             = 0
ioctl(3, SIOCGLIFUSESRC, 0x080728A0)            = 0
rge1: flags=1001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,FIXEDMTU> mtu 1480 
index 2
write(1, " r g e 1 :   f l a g s =".., 86)      = 86
close(3)                                        = 0
so_socket(PF_INET, SOCK_DGRAM, IPPROTO_IP, 0x00000000, SOV_XPG4_2) = 3
so_socket(PF_INET, SOCK_DGRAM, IPPROTO_IP, 0x00000000, SOV_DEFAULT) = 6
ioctl(6, SIOCGTUNPARAM, 0x08073670) (sleeping...)
^C    Received signal #2, SIGINT, in ioctl() [default]

The dom0 is not using any tun/tap interfaces.  This is looking less likely to 
be Indiana related so assuming no responses, I will be moving the issue 
elsewhere.

--
Brian Ruthven                                        Sun Microsystems UK
Solaris Revenue Product Engineering             Tel: +44 (0)1252 422 312
Sparc House, Guillemont Park, Camberley, GU17 9QG

_______________________________________________
indiana-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/indiana-discuss

Reply via email to