Most recent kernel where this bug did *NOT* occur: 2.6.15 (Ubuntu's kernel) Distribution: Ubuntu Hardware Environment: ? Software Environment: ? Problem Description: After booting with the 2.6.19.1 kernel from kernel.org, the wireless that uses bcm43xx driver does not work Steps to reproduce: 1. compile 2.6.19.1 and install it as per http://www.ubuntuforums.org/showthread.php?t=157560 2. reboot 3. sudo ifconfig eth1 up sudo iwlist eth1 scan -> eth1: no scan results (even though there are many many ap's around) To fix: 4. reboot to Ubuntu's kernel 5. remove and purge 2.6.19.1 6. reboot to Windows and connect to some wireless ap there 7. reboot to Ubuntu's kernel (no need to boot to Windows ever again as long as you don't boot to 2.6.19.1)... For some reason, if I don't reboot to Windows and connect to an ap, bcm won't work in Ubuntu even with its own kernel... Errors encountered in dmesg, /var/log/syslog, and /var/log/messages with 2.6.19.1: ------------------------------- Dec 14 14:15:55 localhost kernel: [ 23.142000] cs: IO port probe Dec 14 13:39:19 localhost kernel: [ 23.414000] AC'97 2 access is not valid [0xffffffff], removing mixer. 0x3e0-0x4ff:<4>atiixp: codec read timeout (reg 0) Dec 14 14:15:55 localhost kernel: [ 23.908000] atiixp: codec read timeout (reg 0) Dec 14 14:15:55 localhost kernel: [ 23.913000] atiixp: codec read timeout (reg 3c) Dec 14 14:15:55 localhost kernel: [ 23.914000] atiixp: codec read timeout (reg 1c) (repeats itself for a while) Dec 14 14:15:55 localhost kernel: [ 23.915000] AC'97 2 does not respond - RESET (ac'97 2 must be the winmodem?) Dec 14 14:29:07 localhost kernel: [ 826.556000] ACPI: PCI Interrupt 0000:02:02.0[A] -> Link [LNK2] -> GSI 5 (level, low) -> IRQ 5 (02:02.0 is my broadcom, I'll attach the lspci files in a minute) Dec 14 14:28:50 localhost kernel: [ 809.447000] ieee80211_crypt: unregistered algorithm 'NULL' Dec 14 14:29:07 localhost kernel: [ 826.543000] ieee80211_crypt: registered algorithm 'NULL' Unfortunately, I won't be able to test code on this production machine, sorry. But pls let me know if I can provide more info.
Created attachment 9818 [details] /var/log/dmesg
Created attachment 9819 [details] /var/log/messages
Created attachment 9820 [details] /var/log/syslog
Created attachment 9821 [details] output of lspci
Created attachment 9822 [details] output of lspci -n
Created attachment 9823 [details] lspci -v
Created attachment 9824 [details] my config file for 2.6.19.1 I don't think this will be of use to you but... just in case...
correction: Distribution: Ubuntu 6.06.1 up-to-date
all that information and I forgot to give info on laptop model: HP Pavilion zv5120us laptop
Is this issue still present with kernel 2.6.20?
A BCM4306 should have worked with 2.6.19.1. For easing our efforts in debugging, please turn on the DEBUG option for both IEEE80211SOFTMAC CONFIG_IEEE80211_SOFTMAC=m # CONFIG_IEEE80211_SOFTMAC_DEBUG is not set CONFIG_WIRELESS_EXT=y and CONFIG_BCM43XX=m # CONFIG_BCM43XX_DEBUG is not set CONFIG_BCM43XX_DMA=y Without debug information for the bcm43xx, we don't know any of the details of your card, which are very important. I would also like to see the output of 'iwconfig eth1' when you are trying to scan. Having to boot to Windows to fix some situations sounds as if there is a firmware problem. With debugging turned on, we will get important information about that as well.
> I would also like to see the output of I truly appreciate your help. After enabling debugging that you mentioned, are there any specific files that you would like me to upload to the bug? thanks.
I need the new version of dmesg and the output of 'iwconfig eth1', or whatever your wireless interface is called.
sorry, can't try because can't boot with 2.6.20-1 I bump into a brick wall of "ERROR, PORTS ALREADY IN USE" for my hda hard disk.
How did you generate the .config for 2.6.20? You should have done a 'make oldconfig' while running your 2.6.19.1 system. You shold look for differences between the 2.6.19 and 2.6.20 versions of .config, particularly in the disk section.
Created attachment 10475 [details] config for 2.6.20-1 (brick wall) I looked at the difference but I cannot see anything (probably bc I have no idea where to look for). I made sure I had support for each hardware piece I had. I'm attaching it here in case you'd wanna see... I found that these people[1] had the same issue and though it might be related to an Ubuntu bug[2] but I've no idea... [1] http://ubuntuforums.org/showthread.php?t=321959 [2] https://launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/76872
From the second thread, it seems as if initramfs-tools. One thing that seems to work is to add add "ide0=noprobe ide1=noprobe" to your kernel command-line at the GRUB screen.
> add "ide0=noprobe ide1=noprobe" I tried that one too, but it didn't help either... frustrating indeed I posted there, maybe someone there can help. I'll see. I'll also try asking for a backport fix from Ubuntu.
Any updates on this problem? tu, have you tried newer kernels? Thanks.
Closing for no recent activity. Please reopen if problem confirmed with latest kernel.