I'd like to do what I can helping out fixing this bug. (: At least for
the autodetection part.
(Would hotplug support replace starting wpa_supplicant as its own process
from /etc/init.d ? If anyone reading this has spare time enough to
explain how it works, please do. Because I think #304032 is a bad idea.)
The relevant line from my lspci is:
0000:02:02.0 Network controller: Intel Corp. PRO/Wireless LAN 2100 3B Mini PCI
Adapter (rev 04)
For what it's worth there does not seem to be a directory in my
/proc/net similar to those hostap/, p80211/ and ndiswrapper/ listed in
the script.
I do however have the following interesting files:
/sys/class/net/eth0/device/hardware
NIC entry [Address ] : Hex
IPW2100_CONTROL_REG [00220000] : 0707
0x210014 [00210014] : 72
0x210000 [00210000] : 82
/sys/class/net/eth0/device/device
0x1043
/sys/class/net/eth0/device/subsystem_device
0x2551
/sys/class/net/eth0/device/subsystem_vendor
0x8086
/sys/class/net/eth0/device/vendor
0x8086
Out of these I guess the vendor id is the best bet to use? Right?
A patch to the script posted in the mail to which I am replying is
attached.
--
/Martin
--- wpasupplicant 2005-05-01 13:39:07.626581862 +0200
+++ wpasupplicant.ipw 2005-05-01 14:27:30.455751948 +0200
@@ -51,6 +51,8 @@
DRIVER="wext"
elif [ -d "/proc/net/ndiswrapper/$1" ]; then
DRIVER="ndiswrapper"
+ elif [ "`cat /sys/class/net/eth0/device/vendor 2> /dev/null`" =
"0x8086" ]; then
+ DRIVER="ipw"
else
case "$1" in
ath*)
