Most recent kernel where this bug did not occur: 2.6.20 Distribution: Gentoo Hardware Environment: broadcom 4306 rev02 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ Software Environment: Gentoo Base System release 1.12.9 gcc-4.1.2, glibc-2.5-r4 Problem Description: kernel 2.6.20 and earlier with gentoo patchset + ndiswrapper and kernel 2.6.20 with gentoo patchset + bcm43xx driver: wirelesscard worked perfectly kernel 2.6.21, 2.6.22 with gentoo-patchset + bcm43xx driver: poor connection, 10kbit downstream, several seconds ping time when loading many things at once through the connection notice: in the testings I noticed, that rebooting is the only way (ok I didn't try to wait several hours) to revise the problem. so when I start firefox, with many pages loading at once, it doesn't help to stop the loading, close the tabs, close firefox or even restart the init-service. only after a reboot everything works again. kernel 2.6.23_rc2 (plain vanilla) with bcm43xx driver: device doesn't work at all With patches from ftp://lwfinger.dynalias.org/patches which creates a bcm4301 driver, that should work with my card according to http://lists.berlios.de/pipermail/bcm43xx-dev/2007-July/005095.html : kernel 2.6.22 with gentoo-patchset + bcm4301 driver: kernelpanic on boot kernel 2.6.23_rc2 (plain vanilla) with bcm4301 driver: System hangs when trying to get up the interface (through init-script) For more info see original report at: https://bugs.gentoo.org/show_bug.cgi?id=188077 and this seems to be a similar bug report: https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/124159
This results from a typo in the specifications. When translated into code, it broke this particular model. It has now been reverted with the following commit: commit 829ef1f574a979f681959fd17f9bd698e878cda4 Author: John W. Linville <linville@tuxdriver.com> Date: Mon Aug 6 14:41:59 2007 -0400 Revert "[PATCH] bcm43xx: Fix deviation from specifications in set_baseband_attenuation" This reverts commit 77548f58070894cf5970a110981e511ffe793369. David Woodhouse wrote: >This broke my shinybook. I seem to get absolutely _no_ outgoing packets, >although I can receive OK. Larry Finger wrote: >Please revert this patch. Signed-off-by: John W. Linville <linville@tuxdriver.com> The complete failure of driver bcm43xx for this particular device should be fixed by the above reversion in 2.6.23-rc3. The temporary driver that was called bcm4301 has been abandoned. That effort has resulted in a driver named b43legacy that is in the wireless-dev git tree. On my system, that driver handles the BCM4306/2 device. Performance is not as good as other chips, but it is being worked on. I will certainly investigate all the changes since 2.6.20 to see which might be causing a regression in performance. If you have troubles with b43legacy, please report it on the Broadcom Linux mailing list at bcm43xx-dev@lists.berlios.de.
Fixed in 2.6.23-rc3...