Steve Scrimpshire

Now the question i should have asked before: Did you as root? modprobe ndiswrapper and add to /etc/modprobe.preload ndiswrapper and add to /etc/modprobe.conf alias wlan0 nidswrapper ?

Yeah, that.

ndiswrapper -m My router uses open WEP, so mine was iwconfig wlan0 open blahblah

Your card seems to be running, but it does not see your access point. I'm gonna show you the easy way and the hard way. Hard way: Find the hwaddr of your router (usually written on the bottom of the router), also known as MAC address. If you have another computer that can connect, you can also see it in the config of the web interface to the router. As root, in a console, run the command: iwconfig wlan0 ap <<insert router's hwaddr here>> Then, I am sure you have encryption set up on the router like a good wireless network user and ssid broadcast is turned off, so you need to run these two commands: iwconfig wlan0 essid <<insert essid here>> iwconfig wlan0 key <<encryption key>> You can do all these at one time if you wish (I think...not 100% sure): iwconfig wlan0 ap <<insert router's hwaddr here>> essid <<insert essid here>> key <<encryption key>> After doing this, check your /etc/sysconfig/network-scripts/ifcfg-wlan0 file and make sure it added the values there: DEVICE=wlan0 BOOTPROTO=dhcp NETMASK=255.255.255.0 #assuming your router is 192.168.1.something NETWORK=192.168.1.0 BROADCAST=192.168.1.255 ONBOOT=yes METRIC=10 #this depends on your card...I have the same chipset and yes/no seems to make no difference for me MII_NOT_SUPPORTED=no USERCTL=yes WIRELESS_MODE=Managed WIRELESS_ESSID=[b]<<your essid>>[/b] WIRELESS_RATE=54M WIRELESS_ENC_KEY=[b]<<your key>>[/b] #this is good to be set to 'no' IPV6INIT=no #so is this IPV6TO4INIT=no DHCP_CLIENT=dhclient DHCP_HOSTNAME=[b]<<hostname you want to use>>[/b] NEEDHOSTNAME=no PEERDNS=yes PEERYP=no PEERNTPD=no Do: service network restart and make sure you can surf. Easy way: Run Mandrake Control Center (mcc from a console) Select Hardware Select Look at and Configure Hardware Select your wireless card Click Run Config Tool Select Wireless Select wlan0 Configure everything there HTH

Ok, looks like I have it figured out. First, let me say that dhclient is run as a service with the directive that /etc/dhclient-wlan0.conf is the config file: send host-name "laptop.mdk"; # temporary ifup addition and I am too afraid to add directives to it...afraid to screw things up. I found this in /sbin/dhclient-script that I commented out and it seems to work: # if [ -n "$DOMAIN" ]; then # d="search $DOMAIN" # elif [ -n "$new_domain_name" ]; then # d="search $new_domain_name" # fi If this gives me trouble, I may do the chattr option or try m hand at adding some directives to /etc/dhclient-wlan0.conf. Thanks guys!

ian: I have already tried commenting that out. It still shows up.

Try adding your manfid line there: manfid 0x0105, 0x110a

Fixed the nvidia problem by installing the kernel *src.rpm and the kernel-source *.src.rpm and also had to install wireless-tools from Cooker and libiw from Cooker to support this, but the AirForceOne doesn't do 54M with this driver and drops lots of packets at 24M, 36M, 48M. This is still promising though as the driver matures.

Update: I've tried this with my Belkin Air Force One with the BCM4318 chipset on a non-Cooker system with kernel-2.6.16.1mdk and so far no luck, but I think that make installfw is just putting the firmware in a place the module cannot find it. On system shutdown, I got an error about not being able to load bcm<<something>>.fw. More on this later. SUCCESS (sort of): Before doing make installfw, do: export FIRMWARE_INSTALL_DIR=/usr/lib/hotplug/firmware But, installing the 2.6.16.1mdk kernel on my system broke the nvidia driver, since the version of gcc used to compile that kernel is differenet from the one on a non-Cooker system. Trying to update gcc gives a ton of dependency problems, so maybe installing the 2.6.16 kernel from source would be the way to go if you wish to do this.

How do I keep the "search blahblah" line from showing up in /etc/resolv.conf? The only reference I can find for the host there is in the router itself. I have NEEDHOSTNAME=no set in /etc/sysconfig/network-scripts/ifcfg-wlan0

Well, I sorta agree with both of you. While it may be relevant for those kernels that may be available to non-Cooker users, I don't know if my instructions will work fro non-Cooker with those kernels. I will have to check on my non-Cooker box as soon as I can. I will post back with results and see if we can't get this moved back. :D Apparently, this is not true: It just takes about 30 seconds to regain connectivity after doing 'service network restart', even though it says "Determining ip address for wlan0 OK"

Here's the details of my card: < root /home/omar > lspci -v 06:00.0 Network controller: Broadcom Corporation BCM4306 802.11b/g Wireless LAN Controller (rev 03) Subsystem: Motorola WN825G Flags: bus master, fast devsel, latency 64, IRQ 11 Memory at 36000000 (32-bit, non-prefetchable) [size=8K] I just discovered that, although I previously had to use ndiswrapper for this cardbus card, with the new kernel module, I can now run it natively. This is still considered an experimental module from what I've read, so your mileage may vary. First, I downloaded bcm43xx-fwcutter from here: http://developer.berlios.de/project/showfi...p?group_id=4547 Then I removed all traces of ndiswrapper from /lib/modules/2.6.16-1mdk/kernel//3rdparty and /lib/modules/2.6.16-1mdk/misc (if you don't, ndiswrapper still runs first, even if you have removed its alias from /etc/modprobe.conf). Then I installed bcm43xx-fwcutter, copied the *.sys file (in my case bcmwl5a.sys) to the bcm43xx-fwcutter-004 directory where I installed from and ran: bcm43xx-fwcutter bcmwl5a.sys Then I run make installfw as root, of course. I changed my alias in /etc/modprobe.conf to alias wlan0 bcm43xx (If you are doing this and you have never had ndiswrapper running, it may not be wlan0, but eth2 or some other ethX...I'm not sure.) I had to add these modules to /etc/modprobe.preload: ieee80211softmac bcm43xx (Due to some problems with Cooker, if that is what you are running previously I also had to add yenta_socket pcmcia pcmcia_core to /etc/modprobe.preload for some reason they had stopped loading automatically, even before switching to the new kernel module) bcm43xx defaults to 11 Mbps, so I had to also add this line to my /etc/sysconfig/network-scripts/ifcfg-wlan0: WIRELESS_RATE=54M For some reason after switching to this module if you do service network restart you lose internet connection and have to do rmmod bcm43xx && modprobe bcm43xx && ifup wlan0 to get it back. I hope this helps some people. Please comment on how these steps work for you. I am using Cooker last updated on Jun 2, 2006.

Oopsie. auto-updated and got X back. No major problems from GTK yet. Ugly black borders on GTK apps like sylpheed and firefox, but other than that, ok.

Possibly install libxorg-x11-devel. Probably what it is looking for.

Well, at least X was working for me, but now not at all. I am on a LiveCD right now and no time to locate the errors and post them, but I updated again this afternoon about 7 hours ago. Edit: Found the errors: (II) LoadModule: "vesa" (II) Loading /usr/lib/xorg/modules/drivers/vesa_drv.so (II) Module vesa: vendor="X.Org Foundation" compiled for 7.0.0, module version = 1.0.1 Module class: X.Org Video Driver ABI class: X.Org Video Driver, version 0.8 (EE) module ABI major version (0) doesn't match the server's version (1) (II) UnloadModule: "vesa" (II) Unloading /usr/lib/xorg/modules/drivers/vesa_drv.so (EE) Failed to load module "vesa" (module requirement mismatch, 0) This happens when I try any driver. Apparently this is what I had in ~/.xinitrc: [root@localhost omar]# cat .xinitrc /usr/X11R6/bin/fluxbox So that would explain that fluxbox problem..

I can use my machine just fine. :P I was basically wondering if anyone knew where to start as far as documenting these problems with the cooker folks.

All of these problems were 'repairable' to some degree, but ever since doing urpmi --auto --auto-select two days ago, I now have several major problems and I'm not sure where to report them, if they are interconnected, etc. To start with, I read on here that there was a possibility with the very latest kernel to be able to give up ndiswrapper and go native with my Motorola FD5700 (I think): Identification Vendor: ‎Broadcom Corp. Description: ‎BCM94306 802.11g NIC Media class: ‎NETWORK_OTHER Connection Bus: ‎PCI Bus PCI #: ‎6 PCI device #: ‎0 PCI function #: ‎0 Vendor ID: ‎5348 Device ID: ‎17184 Sub vendor ID: ‎4183 Sub device ID: ‎28709 Misc Module: ‎bcm43xx At first, it didn't work, so I did the urpmi --auto --auto-select and then things got really crazy. X no longer starts at boot (I had auto-login enabled) and I cannot get it to. Now, when I do startx, Fluxbox starts like it should, but it doesn't read ~/.fluxbox/startup like it used to and I have to source it manually. No wireless connection even after uninstalling ndiswrapper, reinstalling and reinstalling my driver from disk. I discovered that there is no longer a pcmcia service and pcmcia, pcmcia_core, and yenta_socket modules are not get loaded, so I had to put them in /etc/modprobe.preload. There is also no longer a cardctl command, but appears to be a pccardctl command that does the same thing....but that's trivial. Any thoughts on where to go from here with either problem? With all this I have completely gotten sidetracked on whether or not the new kernel's modules support my wireless card :D

I know this is reviving the dead, but I just happened to be browsing Workbench and stumbled on this. I go to that site now and all I see is: Are you gonna continue with this? I would've liked to have seen this.

Name suggestions: AlchemyLiveCD (though you may run into trouble with AMD, since they have a device called "Alchemy Au1200" that uses embedded linux) ValenceLiveCD TriplePointLiveCD

This is a great tip. I do have one question, though. Does anyone really need a whole gig of swap space? (I know yours was probably just an easy example.)

It is my understanding that the distro violates GPL by not providing source code. There is nothing in the GPL that states they have to give credit to the distro they derive it from...unless I have missed that part in the GPL.

Linus comes over to my house every day for lunch and chess. He's a sore loser, too.

Man, I am so glad I saw this before I did urpmi --auto-select next: < omar ~ > urpmq --sources xorg-x11-server ftp://ftp.proxad.net/pub/Distributions_Linux/MandrivaLinux/devel/cooker/i586//media/main/ xorg-x11-server-7.0.0-3mdk.noarch.rpm < omar ~ > Xorg -version X Window System Version 6.9.0 Release Date: 21 December 2005

I *think* I have solved it. Apparently, sylpheed's spamassassin plugin doesn't allow spamassassin to alter the message. I dropped this script: #!/bin/bash # # sc-spamc : a script for using SpamAssassin header modification / subject # tagging vi Sylpheed-Claws filtering actions. # # Author: Robert Story # Send questions or comments to the sylpheed-claws-users mailing list # # Based on a simpler script posted to the sylpheed-claws-users mailing list # by Richard A Downing. # # Should return the return code from spamc (1=spam, 0=nospam/error) # die() { : error "$@" : exit 0 } FILE=$* TEMP=sc-spamc.$$ # # call spamc, redirect output to temporary file, and save return value # /usr/bin/spamc -E -t 30 < "$FILE" > "$TEMP" RC=$? # # make sure temp file exists before we replace original message # if ! test -s "$TEMP"; then : die "spamc result file doesn't exist or is 0 bytes" fi # # backup original # mv "$FILE" "$FILE.bak" if ?$? -ne 0; then : die "'mv $@' failed!" fi # # rename filtered file # mv "$TEMP" "$FILE" if ?! 0 = $?; then : mv "$FILE.bak" "$FILE" : die "spam result rename failed, restoring backup" else : rm "$FILE.bak" fi # # return spamc exit code # exit $RC into ~/bin and created a filter rulle in spamassassin: test "!(sc-spamc %F)" move "#mh/Mailbox/Spam" And I am pretty sure it is working properly now. Found this info here: http://claws.sylpheed.org/faq/index.php/Fi...SpamAssassin.3F

I have spamassassin setup on my laptop and here is my /etc/mail/spamassassin/local.cf (with default comments removed): rewrite_header Subject *****SPAM***** required_score 1.0 use_bayes 1 ok_languages en ok_locales en bayes_auto_learn 1 required_hits 5 report_safe 0 auto_whitelist_path /var/spool/spamassassin/auto-whitelist auto_whitelist_file_mode 0666 spamd runs as a service, so I am not sure if it reads my ~/.spamassassin/user_prefs, but here it is (comments removed): required_score 1.0 rewrite_header subject *****SPAM***** report_safe 1 use_bayes 1 bayes_auto_learn 1 skip_rbl_checks 0 use_razor2 1 use_dcc 1 use_pyzor 1 ok_languages en ok_locales en Now my question: I am using the spamassassin plugin with sylpheed-claws and I use fetchyahoo to download my mail to /var/spool/mail/omar and sylpheed processes mail from there and runs it through spamassassin. The problem is that spam goes to the spam folder like it should, but the subject is not rewritten and the mail doesn't show as unread in the spam folder of sylpheed (I have to look to see if there is mail in there and sometimes I forget). Under the 'S' column there is a little black frowny emoticon, though (I'm assuming that is the way that sylpheed marks the spam). Here's a screenshot: Here's the headers of the message: X-RocketYMUMID: ADRVv9EAALqSRGvLdQoiPCoQFQc X-Apparently-To: omarserenity@yahoo.com via 209.191.85.52; Wed, 17 May 2006 18:18:45 -0700 X-YahooFilteredBulk: 83.199.7.146 X-Originating-IP: [83.199.7.146] Return-Path: <russelystovall@pjstar.com> Authentication-Results: mta268.mail.scd.yahoo.com from=pjstar.com; domainkeys=neutral (no sig) Received: from 83.199.7.146 (HELO hansbuba.com) (83.199.7.146) by mta268.mail.scd.yahoo.com with SMTP; Wed, 17 May 2006 18:18:45 -0700 Message-ID: <000001c67a18$5dd12830$87cda8c0@rnm73> Reply-To: "Russell Stovall" <russelystovall@pjstar.com> From: "Russell Stovall" <russelystovall@pjstar.com> To: omarserenity@yahoo.com Subject: Re: the meldtcations Date: Wed, 17 May 2006 18:14:09 -0700 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0001_01C679DD.B1725030" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2800.1106 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106 X-FetchYahoo: version 2.10.2 MsgId 7404_59813_1730_1348_5012_0_43498_8211_3056279695 Sylpheed sees it as spam, even though spamassassin didn't rewrite the subject. I'm assuming spamassassin is doing its job (or at least sylpheed's plugin is), but why is the subject not getting rewritten.? Is there a way to have the message marked as Unread when incorporated from the spool? Edit: Ok, it is just the Spam folder that the messages are getting received marked as read already. Apparently, sylpheed sees that they are spam from the spamassassin plugin and marks them as spam and so they are not marked as unread. Still don't know why spamassassin is not rewriting the subject line.