Here is my output from sudo /sbin/modinfo orinoco. I will try the experiment you suggested when I am at my wireless location later today.

SUCCESS!!!

THANK YOU.

I am at my wireless location logging this post right now. Here is what I typed in, and results:

While I am extremely happy everything worked, I want to ask one last favor. Help me with my education and understanding of which parts of what I typed in worked and what the rest of it means. When I saw the "FATAL" notifications come up, I decided to go ahead and enter the rest of the lines you recommended. I'd like to learn from this what FATAL meant, and what the likely outcome would have been if I had not gone ahead with the other commands after that, as well as what you might have recommended next if I had just posted that output. Could any of the lines have been omitted? Does the output suggest what was wrong from the very beginning when my wireless first disappeared?

When I first turned the computer on just a bit ago, I did get a message on the desktop saying "Disconnected" because I believe it was searching for connections. The last connections have all been wired connections at my home, and the system surely realized the wired connection was not attached this time. As soon as I typed in the last of the commands that "Disconnected" changed to the name of the connection (which just happens to be "Panera," my favorite stop for coffee and work).

I sincerely appreciate your help and perseverance with me.
David
Indianapolis, Indiana USA
20091221 1630 EST

I was looking at the output of lsusb and lsmod and couldn't think of why the "iwlist eth1 scan" failed. When I read the "force_monitor" option, it made me think that this possibly needed for scanning for APs. Also, a modprobe config option could easily be missed after a kernel security update, which I'm guessing is what caused your problem. Since iwlist scan still doesn't work, I'm still a bit stumped about that. I would need more particular details on the command you ran when you saw the FATAL message, and the exact wording of the entire message. Reading your /var/log/messages log may provide more details on the error.

I recommended that you shut down wireless, because if you try to unload a module that is still being used, you can lock up the kernel. Unloading and reloading the kernel was just for the test. It would be preferable to have the module loaded with that option in the future.

There is still some work to do. You need to make sure that this option is used when the moduler is modprobed in the future. Where to do this may depend on your distro and distro version.

• Check for orinoco & orinoco_cs lines in /etc/modprobe.conf
• Check in files in /etc/modprobe.d/. Perhaps a 50-orinoco_cs file. Try `grep orinoco /etc/modprobe.d/*'.
• Look in /etc/ for a file such as /etc/cardctl.conf. Hardware configurations like this have tended to move to udev rules, so you may not have one.

If you can't find anything, that is OK. Create a file "50-orinoco.conf" in /etc/modprobe.d/.

Add the lines:
# force_monitor allows scanning for Access Points
options orinoco_cs force_monitor=1

This should load the orinoco module with the option "force_monitor=1" when the orinoco_cs module is loaded. I based this on my systems /etc/modprobe.d/50-bluetooth.conf file. The btusb module depepends on the bluetooth module and the bluetooth module needs the "reset=1" option. So this has the same dependency relationship with an option as in your case.

My system (openSuSE 11.2) has an /etc/modprobe.d/99-local.config file. This may be a better file to add these lines. Just in case a security update replaces the 50-orinoco.conf file.


----

Some background information you requested.

Modinfo is a command used to report information about kernel modules. They don't need to be loaded to see these options. It reports things like dependencies, and what parameters the module can take.

Once I know the manufacturer's id of a device, I will search the kernel code for clues on which module is needed. When I bought a new Gateway netbook, it had an Atheros wireless device, so I knew from experience that the ath9k kernel driver was needed. However, looking in the source code, I saw that it didn't include the manufacture's ID of my device. I was using openSuSE 11.1 at the time. My solution was to add the "factory" suse repo that included a newer kernel. The newer kernel's ath9k module checked for my new wireless device.

Lspci is used to scan your pci bus and report what devices are connected to the bus. There are functions which interrogate these devices. The pciids package supplies a human readable database that has manufacture's IDs and Human readable text describing the device.

Each of the wireless tools has a manpage. I'd highly recommand scanning them. In answering your post, I rediscovered the "iwevent" program, which can help monitor the supplication process.


Good Luck!

1 members found this post helpful.

I'm just checking and I see that you saw my post. I've been running around all day. Glad to be part of spreading the love. I really don't know what I'm doing yet and I'm tired, so I understood about 4% of what you just said. You should see my face! But thanks, because I'm sure it will make sense to me some day. I will take another look at your post tomorrow. I believe I can get my understanding up to, maybe, 6%. I hope you're laughing because I'm trying to use humor to tell you how much I appreciate your help. But I'm serious about the 4%.

Keep that in mind when I reply again because I'm going to be asking what some things mean.

Seriously, I do want to make sure I have a good system that won't require extraordinary measures next time there is a major (or even a small) upgrade. So I'm interested and attentive about your saying that we're not done yet.

I was laughing at the "running around all day". I had some equipment break down at work on Friday night, so I can identify.

We have some new servers that run an OS based on Plan 9. I don't think I'm even at 4% yet.

Now I've gone and done it. For the moment, I can't get on the Internet at all, from my laptop. I don't have NetworkManager or WICD installed on it anymore. I was having trouble getting my wireless to detect at home, and I kept getting asked by WICD for my system logon password at bootup so. But the WICD tray icon never appeared and I couldn't get connected, so I thought I'd delete/uninstall WICD and reinstall it. However, the reinstall didn't work because it needed to download some files, which the system couldn't do because it would not connect to the Internet. Help! What do I do to at least be able to get online now? I'm typing this from another computer so copying and pasting command outputs won't be as easy. I'll have to transcribe. But I'll do that if it means getting back online.

I think I would start with "ps -ef wicd" to see if the process is running.
The command "lsof -c wicd" will print the files opened by a program starting with the letters "wicd". One of them might be a log.

Check the kernel logs, e.g. /var/log/messages and enter the command: "dmesg | less". Does the kernel module load successfully?

I'd recommend to always gather what information you can on a problem before taking action. Sometimes you may find a log file you didn't know about before has answers pinpointing the problem.

Also could you post the line you entered for modprobing the kernel device. I hope I didn't steer you wrong earlier.

"ps -ef wicd" gave me "ERROR: Unsupported option (BSD syntax)" ... with a whole list of possible commands. One of the ones I saw was simply "-e" for "all processes" so I typed "ps -e" and got a long list of information in column headings "PID," "TTY," "TIME," and "CMD." WICD was not among the listings.

When I typed "lsof -c wicd" there was no response; just return to the terminal prompt.

I don't think I know what to do with the suggestion to check the kernel logs... is that done right at the terminal prompt also? I tried typing "/var/log/messages" at the terminal prompt and got a response "bash: /var/log/messages: Permission denied" ... not sure what I'm doing there. Where/how does one check the kernel logs?

When I typed "dmesg | less" I got a whole bunch of stuff I didn't understand, which continued as I pressed the Enter key repeatedly.

Sorry I understand so little of this yet. Should I be looking at getting a disk to reload a network manager to my system, or something like that? Since I don't have access to the Internet, I can't download anything or load any program that depends on or initiates a download. What have I done? ? ? Dang!

Unless the actual binary that runs has a different name, it appears that the wicd program isn't running.

I don't have it available for my current distro. I downloaded an older rpm package and took a look inside. There are two programs /usr/sbin/wicd and /usr/bin/wicd-client. I presume that wicd runs as root and is started when you bootup, and wicd-client is a user program run in your desktop environment.

One thing to check is your kernel messages. Either run "dmesg" or "sudo less /var/log/messages". My guess is that the wicd program is failing to start. If that is the case, it may log a message telling you why.

When I ran "sudo less /var/log/messages" the only log entry that occurred before my last post, on Jan. 5, was:

Jan 5 23:33:05 customer-laptop rsyslogd: [origin software="rsyslogd" swVersion="4.2.0" x-pid="662" x-info="http://www.rsyslogd.com"] rsyslogd was HUPed, type 'lightweight'.

Everything else is dated after I used the system to remove the wicd program because I was having trouble getting my wireless to detect at home, and I kept getting asked by WICD for my system logon password at bootup (and the WICD tray icon never appeared).

I could make no sense of the output from "dmesg" and would have to be instructed about what to look for there.

I'm using a different computer but would like to know if any program to load a network manager is available on a hard device, like CD (since I can't download on the other computer).