Description of problem: The bcm43xx driver no longer associates with my access points in the rawhide kernels on my G4 ibook. I use kernel-2.6.16-1.2096_FC5 to get wireless going. Everything else in the rawhide tree up to kernel-2.6.16-1.2211_FC6 is b0rked for me. Version-Release number of selected component (if applicable): all rawhide kernels up to and including 2.6.16-1.2211_FC6 How reproducible: 100% Steps to Reproduce: 1. Boot to the kernel 2. Try to associate with a WAP 3. Actual results: Here is the relevant dmesg output: bcm43xx driver PCI: Enabling device 0001:10:12.0 (0004 -> 0006) bcm43xx: Chip ID 0x4318, rev 0x2 bcm43xx: Number of cores: 4 bcm43xx: Core 0: ID 0x800, rev 0xd, vendor 0x4243, enabled bcm43xx: Core 1: ID 0x812, rev 0x9, vendor 0x4243, disabled bcm43xx: Core 2: ID 0x804, rev 0xc, vendor 0x4243, enabled bcm43xx: Core 3: ID 0x80d, rev 0x7, vendor 0x4243, enabled bcm43xx: PHY connected bcm43xx: Detected PHY: Version: 3, Type 2, Revision 7 bcm43xx: Detected Radio: ID: 8205017f (Manuf: 17f Ver: 2050 Rev: 8) bcm43xx: Radio turned off bcm43xx: Radio turned off later.... bcm43xx: PHY connected bcm43xx: Radio turned on bcm43xx: ASSERTION FAILED (radio_attenuation < 10) at: drivers/net/wireless/bcm43xx/bcm43xx_phy.c:1485:bcm43xx_find_lopair() bcm43xx: ASSERTION FAILED (radio_attenuation < 10) at: drivers/net/wireless/bcm43xx/bcm43xx_phy.c:1485:bcm43xx_find_lopair() bcm43xx: ASSERTION FAILED (radio_attenuation < 10) at: drivers/net/wireless/bcm43xx/bcm43xx_phy.c:1485:bcm43xx_find_lopair() bcm43xx: Chip initialized bcm43xx: DMA initialized bcm43xx: 80211 cores initialized bcm43xx: Keys cleared SoftMAC: Associate: Scanning for networks first. SoftMAC: Start scanning with channel: 1 SoftMAC: Scanning 14 channels bcm43xx: ASSERTION FAILED (radio_attenuation < 10) at: drivers/net/wireless/bcm43xx/bcm43xx_phy.c:1485:bcm43xx_find_lopair() ADDRCONF(NETDEV_UP): eth1: link is not ready SoftMAC: Associate: Scanning for networks first. SoftMAC: Associate: failed to initiate scan. Is device up? SoftMAC: Scanning finished SoftMAC: Queueing Authentication Request to 02:12:f0:00:00:5e SoftMAC: cannot associate without being authenticated, requested authentication SoftMAC: Queueing Authentication Request to 02:12:f0:00:00:5e SoftMAC: cannot associate without being authenticated, requested authentication SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. bcm43xx: set security called bcm43xx: .level = 0 bcm43xx: .enabled = 0 bcm43xx: .encrypt = 0 ADDRCONF(NETDEV_UP): dev1804289383: link is not ready SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Start scanning with channel: 1 SoftMAC: Scanning 14 channels SoftMAC: Scanning finished SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Start scanning with channel: 1 SoftMAC: Scanning 14 channels SoftMAC: Scanning finished SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Sent Authentication Request to 02:12:f0:00:00:5e. SoftMAC: Authentication timed out with 02:12:f0:00:00:5e SoftMAC: cannot associate without being authenticated, requested authentication SoftMAC: cannot associate without being authenticated, requested authentication SoftMAC: Authentication timed out with 02:12:f0:00:00:5e SoftMAC: cannot associate without being authenticated, requested authentication SoftMAC: cannot associate without being authenticated, requested authentication SoftMAC: Start scanning with channel: 1 SoftMAC: Scanning 14 channels SoftMAC: Scanning finished then there are a whole swag of avc denied SELinux messages, mostly involving DBus (SELinux is permissive on my system), then a whole lot of this: SoftMAC: Start scanning with channel: 1 SoftMAC: Scanning 14 channels SoftMAC: Scanning finished SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. SoftMAC: Authentication response received from 00:12:f0:02:e3:45 but no queue item exists. Expected results: In the 2096 kernel I just boot up and it's associated and with an IP address, auto-magically. Additional info:
This is a regression which was introduced between the snapshot of bcm43xx which we put in FC5, and the version which was first merged into 2.6.16-git18. Joshua, you were going to try versions of the driver from SVN to see precisely which change was responsible -- how far did you get with that?
Created attachment 129973 [details] Full patch between the FC5 snapshot and the version which was merged.
Don't hold your breath for it, I'm trying to find someone who can show me how to do that, who has time to do it - when I have time to do it.... Am I the only one seeing this?
The patch from comment 2 is a little big to be helpful. Before digging too deeply, it might be nice for you to try the FC5.netdev.3 kernels from here: http://people.redhat.com/linville/kernels/fedora-netdev/5/rpms Could you give those a try to make sure the problem still exists there?
(In reply to comment #4) > The patch from comment 2 is a little big to be helpful. Yeah, just a little. Nevertheless, it's what we have -- that's the difference between the last known good, and the first known broken -- I talked Joshua through using git for that, and we established that bcm43xx was broken before it got merged to Linus. The next step is to build from the old bcm43xx SVN tree (or maybe the daily snapshots if they're still available) and track it down further. I cannot reproduce this here.
With 2.6.16-1.2122.2.3_FC5.netdev.3 wireless networking works. Resume from suspend is still problematic, but that's something else. The bcm43xx driver in this kernel functions for me.
I would expect the standard fedora kernels to work once they are based upon 2.6.17. I'm going to close this as UPSTREAM -- please reopen if the issues remains after fedora gets to 2.6.17.
umm, the standard Fedora kernels have been based on 2.6.17 for quite a few weeks now. I assume you meant .18 ?
Yes, 2.6.18 -- sorry for the confusion!