that's exactly what i have.
the script was from the leftover madwifi-old system.
It makes more sense now,
Daniel Wu wrote:
> (at this point, the debugging output will show ath_mode_init and ath_intr's
> been called)
> I was hoping that upon destroy, the interfaces will be stopped, and it won't
> "leak" to the new one.
> (After the last wlanconfig, the interface comes up automatically)
> Sometimes, when I "down" the interface before the destroy, it does the same
> thing after
> a new creation.
Yes, some distributions ship a package that automatically configures and
brings up devices upon hot-plugging them to the system. Since wlanconfig
create is much like a hotplug event, when the device appears the system
tries to enable the device. I ran into this problem here when i was
trying to hack up the fedora initscripts package to support madwifi-ng
type devices. Turns out that when wlanconfig create was called, the
script would start again. I haven't put a whole lot of effort into
figuring out why this is the way it is. One thing I do know is that if
(under a fedora/redhat-like system) a file
/etc/sysconfig/network-scripts/ifcfg-(subdevice-name) exists and
wlanconfig create is called, that script will be executed. Try moving or
deleting this file if it exists on your system and retry your tests, maybe.