I thought I would write a short summary of the way in which three laptops which I am reponsible for and running F7 now have excellent wireless support with three different drivers.
For a long time I have been struggling with getting one laptop going on wireless using the rt2x00 driver recently supported in the kernel. On a second machine that was running FC6 and had an ipw3945 chipset support was solid with the ipw3945 driver and a real pain to get working in FC6 with iwl3945.
This situation has now improved vastly.
1) Laptop 1 is a Dell d610 with iw2200 wireless. Until recently this was set up in F7 using standard WEP wireless and avoiding the use of wpa_supplicant or NetworkManager. Recently I tried to set up wpa_supplicant by editing the files /etc/wpa_supplicant/wpa_supplicant.conf and /etc/sysconfig/wpa_supplicant appropriately. I found to me delight that running wpa_supplicant either manually or as a daemon worked well, and all that was necessary was to run dhcient to get an ip address. In order to get it to run at boot it was necessary to change the line in /etc/init.d/wpa_supplicant from # chkconfig: - 12 88 to # chkconfig: - 09 88
then I added two lines to /etc/rc.local /bin/sleep 2 /sbin/dhclient
Now wireless comes up nicely at boot and connects seemlessly to the AP.
2) Laptop 2 is an old Amilo D6800 which has no internal wireless. I was using the rt73 legacy driver manually compiled until recently. This has no wpa_supplicant support. However recently the development of the rt2x00 driver in the kernel has had a module rt73usb but has not worked. However the new test kernel for 2.6.23 was installed and it was found that provided /etc/modprobe.conf had a line alias wlan0 rt73usb
and in /etc/modules.d/blacklist the module rt2500usb is blacklisted then this modules works out of the box. Using the same technique as for the above I can now get this wireless to come up seemlessly in wpa_supplicant.
3) Laptop 3 is a Samsung Q35 machine running the ipw3945 chipset. In the latest released kernel for F7 the iwl3945 driver comes up seemlessly and using the same setup as in 1) wpa_supplicant also work very consistently in this machine.
So this is the first time that all three machines run well using wpa_supplicant and it is important to thank all the many people who have worked hard at getting the kernel modules going, and also for the good work on wpa_supplicant that has made this possible.
I am now looking forward to F8!
--
mike cohler
Mike Cohler <mike.cohler <at> gmail.com> writes:
So this is the first time that all three machines run well using wpa_supplicant and it is important to thank all the many people who have worked hard at getting the kernel modules going, and also for the good work on wpa_supplicant that has made this possible.
I am now looking forward to F8!
Apologies for the typos in the previous post.
It will be really nice when the new ath5k driver is made to work. I guess many have atheros based wireless cards and even though some people have had the MADWIFI driver working others, myself included, have not. So an in kernel driver for these chipsets would be even better - and at this stage a significant proportion of available wireless chipsets would be easy to get working. I believe that wireless access was one of the original goals for F7 - and it looks like at clean install F8 may have achieved this aim.
A quick followup to this. 1) ipw2200 and iwl3945 also run consistently and remain stable for WPA wireless encryption. This is very nice after spending a lot of time getting the systems set up. 2) On trying the in kernel rt73usb (from rt2x00) driver again this is not stable, for any real load on the link, and gives a lot of problems. Even with the new test kernel it is still very unstable.
Does anyone have any experience with rt73usb using rawhide kernels and/or rawhide wpa_supplicant within f7?
-
Mike C