Red Hat Bugzilla – Bug 238603
No wireless networks detected
Last modified: 2007-11-30 17:12:03 EST
Description of problem:
With the latest kernel and wireless-tools no wireless networks are detected and
running 'iwlist scan' causes a segmentation fault.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
No wireless networks detected.
My own network (and the other neighbouring networks) should be detected and I
should be able to choose one in NetworkManager.
Gonna need a bit more info. Let's start with the output of running 'dmesg'
and 'lspci -n'.
Created attachment 154001 [details]
Output of lspci -n
Created attachment 154002 [details]
Output of /var/log/dmesg
I suppose I ran out of steam on this one. Let me know what else would be useful.
It's a Vaio SZ3. Had pretty good results with the ipw3945 driver when running FC6.
Please try the latest kernels from here:
Do these work better for you?
Same result, but at least there's more information. I'm posting the dmesg
output as an attachment. Looks like trouble with NetworkManager?
Created attachment 154079 [details]
That's with davej kernel 3125
There still seems to be a problem with the latest kernel 3132. No wireless
networks shown in NM, and a segmentation fault when I run iwlist scan:
iwlist: segfault at 0000000003636cf0 rip 0000000000401c9a rsp
00007fff5a8cd7b0 error 4
Running 3132 w/ NM right now, seems to work OK here. I have had the driver
die and had to 'modprobe -r iwl3945 ; modprobe iwl3945'. Does that make
things any better for you?
Can you configure wireless manually? Does that work?
I'll try that later on. One weird thing with this kernel, having tried the
modprobe commands - ksoftirqd is now taking up 99% CPU time.
Tried it on a different access point this morning and it was the same. The
driver doesn't seem to be associating with any network. I won't have the chance
to try any non-WPA networks until Tuesday.
Must admit I'm not quite sure how to configure WPA manually.
Just tried with the latest davej kernel (3141) and the situation is still the
same: NM doesn't detect any networks and trying to connect to my access point
manually doesn't work as the driver never associates. Running 'iwlist scanning'
causes a segmentation fault.
Ditto with 3144: no networks detected with NM and manual setting doesn't work
Same effect with 2.6.21-1.3149.fc7:
Warning: Driver for device wlan0 has been compiled with version 22
of Wireless Extension, while this program supports up to version 20.
Some things may be broken...
wlan0 IEEE 802.11a ESSID:""
Mode:Managed Frequency:5.17 GHz Access Point: Not-Associated
Retry min limit:7 RTS thr:off Fragment thr=2346 B
Link Quality:0 Signal level:0 Noise level:0
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0
Have you tried using the latest wireless-tools package?
Does "iwlist wlan0 scan" still segfault?
Yes, still getting that segfault. Just about to try that wireless-tools package.
With that package (wireless-tools-28-3.fc7), "iwlist wlan0 scan" works and, once
I'd restarted the NM service, I could pick my own AP and authenticate to it.
Thanks for your persistence with this.
Meant to say 'hurrah' there. However, I was speaking a little too soon. While
wireless networks are detected now, the connection to my AP only lasts a minute
or so and then the laptop becomes very responsive, to the extent that I've had
to sync and reboot using Alt-SysRq twice now. I suppose such iwlwifi problems
are probably well covered in other bugs, though.
Please try the lastest firmware (released 5/16/2007):
Does that work any better for you?
That does seem better - the connection has been up for a few minutes now, though
I don't have the time to test more thoroughly tonight.
Created attachment 155132 [details]
latest dmesg output
This is with kernel 2.6.21-1.3167.fc7 and the newer firmware, as requested in
the previous comment.
Is 1.3194 working better for you? Can this be closed?
Yes. Networks are reliably detected now, though the operation of the driver is
still flaky (though I've referred to that in another bug report).