[09:00] <harmw> omgwtfbbqlol
[09:00] <harmw> libvirt_type = qemu...
[09:01] <harmw> no wonder my cloud instances are dead slow...
[09:01] <harmw> seriously, why didn't I figure that out like a year ago
[09:01] <harmw> (ok, the hardware is slow but still)
[09:01] <harmw> anyway, centos took a mere 40ish seconds, instead 5+ minutes to boot :p
[13:03] <smoser> harmw, my plan was to ditch the static/silly CONFIG_IFUPDOWN and make that configurable.
[13:03] <smoser> at "runtime".
[13:04] <smoser> ie, just have CONFIG_IFUPDOWN_UDHCPC_CMD_OPTIONS as ""
[13:04] <smoser> and then wrap udhcp or ifupdown or somethign with a shell that reads options from somewhere in the filesystem.
[19:01] <harmw> smoser: so let udhcpc run in way the user wants it to run, according to some config file
[19:01] <harmw> sounds like a plan
[19:04] <smoser> harmw, i think i might even consider patching busybox to call a different udchpc
[19:04] <smoser> ie, udhcpc.ifupdown
[19:04] <smoser> or something. 
[19:04] <smoser> then we can just provide that.
[19:04] <smoser> might have to atch busybox to do that
[19:05] <harmw> hmk
[19:05] <harmw> isn't a wrapperscript a little more convenient instead of patching busybox?
[19:05] <harmw> eg. working with a script+config file?
[19:06] <smoser> well, id consider patching busybox to invoke /sbin/udchpc.ifupdown
[19:06] <smoser> so that we'd know when *that*program was called it was from ifupdown
[19:07] <smoser> and not get in the user's way if they ran it : udchpc -i eth0
[19:07] <smoser> or something like that.
[19:07] <harmw> ah like that
[19:08] <harmw> I didnt venture into things that deep :p
[19:08] <harmw> though I'm starting to like the concept
[19:11] <harmw> I'm guessin' one hour of work
[19:11] <harmw> patching busybox... thats just a configuration file thingy I'd guess, right?
[19:12] <harmw> some of our own config in /etc/sysconfig/udhcpc.options
[23:09] <smoser> harmw, there is some infrastructure for quilt applying patches in cirros now
[23:10] <smoser> and i'd just do that to change the string that busybox invokes
[23:10] <smoser> not bother to do it properly (ie, not do a config option)
[23:10] <smoser> i think its networking/ifupdown.c