tjbp wrote:Anyone else experiencing a regression with roaming support in the latest version? When in range of multiple APs on the same SSID the chip appears confused - it suddenly decides a random AP with not much signal has maximum signal strength and insists on connecting to it. The result is almost 100% packet loss. Confining myself to a part of the building with only one AP in range works fine.
I'm experiencing similar problems. High packet loss in the presence of multiple AP's. Not sure about the ssid bit though i think that might be it.
Has anyone found a fix to the problem ? Apart from the slightly inconvenient solution of having to confine yourself to a part of the building with only one AP
Unfortunately I don't have a solution besides configuring a BSSID to prevent it from connecting to a weaker AP. Interestingly the exact same behaviour was the reason I stopped using the b43 driver - so perhaps there's a wider problem higher in the stack...
]]>Anyone else experiencing a regression with roaming support in the latest version? When in range of multiple APs on the same SSID the chip appears confused - it suddenly decides a random AP with not much signal has maximum signal strength and insists on connecting to it. The result is almost 100% packet loss. Confining myself to a part of the building with only one AP in range works fine.
I'm experiencing similar problems. High packet loss in the presence of multiple AP's. Not sure about the ssid bit though i think that might be it.
Has anyone found a fix to the problem ? Apart from the slightly inconvenient solution of having to confine yourself to a part of the building with only one AP
@Legogris, did you rebuild the modules with dkms after the upgrade? I noticed that the other wl package (the 5.100.whatever maintained by Gaeton Bisson) didn't actually need to be rebuilt every kernel release. I was fine with the same module from 3.9 to 3.9.7, but when 3.9.8 hit, I had to rebuild.
That was it, thanks!
Kernel upgrade -> reboot -> dkms install broadcom-wl/6.30.223.30 ->reboot did the trick.
Jul 02 14:19:45 temperance systemd[1]: Job sys-subsystem-net-devices-eth0.device/start timed out.
Jul 02 14:19:45 temperance systemd[1]: Timed out waiting for device sys-subsystem-net-devices-eth0.device.
-- Subject: Unit sys-subsystem-net-devices-eth0.device has failed
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
-- Documentation: http://www.freedesktop.org/wiki/Software/systemd/catalog/be02cf6855d2428ba40df7e9d022f03d
--
-- Unit sys-subsystem-net-devices-eth0.device has failed.
--
-- The result is timeout.
Jul 02 14:19:45 temperance systemd[1]: Dependency failed for Automatic wireless network connection using netctl profiles.
-- Subject: Unit netctl-auto@eth0.service has failed
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
-- Documentation: http://www.freedesktop.org/wiki/Software/systemd/catalog/be02cf6855d2428ba40df7e9d022f03d
--
-- Unit netctl-auto@eth0.service has failed.
--
-- The result is dependency.
Downgrading to 3.9.7-1 works for now.
]]>Folks, I have created the new package broadcom-wl6-dkms: https://aur.archlinux.org/packages/broadcom-wl6-dkms/ . This package also includes the latest upstream version for 3.10 compatibility. I have not tested this against a 3.10 kernel, but the driver runs fine on the current kernel.
I think this is a much better package name for this version of braodcom-wl. I have a BCM43228, and it is working fine with 3.10.0-1-00147-gbd2931b-dirty. Though my machine works fine with the 5.100.* version, that wouldn't build against the 3.10 git tree, and lack of internet connectivity is kind of a show stopper. So I really appreciate your hard work in getting this ubuntu specific module working on Arch Linux.
It amazes me that since ubuntu apparently has the sway to get broadcom to release a new module version, that when they were give something ubuntu specific, they did not turn around and then complain and try to use that same influence to get a module that was distribution agnostic. Though then when I think about it, this does not suprise me at all.
]]>Please test it and I'll ask for a merge sometime next week.
To use the new version, I recommend first uninstalling the old wireless-bcm43142-dkms driver with pacman -R wireless-bcm43142-dkms. Then reboot. Then install the new broadcom-wl6-dkms driver. Reboot again.
]]>It corrects itself after a few minutes, but interestingly while it's confused a iwlist scan returns the inaccurate results instantly, instead of hanging for the usual second or so.
Curiously this exact behaviour was what drove me away from b43.
]]>As before, uninstall the current version first before installing this one. Upgrading doesn't seem to work so smooth.
]]>With the tip from @tjbp, I created a patch for the change in the cfg80211_put_bss() function for kernel 3.9.
I have NOT tested this build, but it does compile without errors on my 3.9 VM. Can someone please test this build 7 and post back with the results?
]]>This will give you more output:
NETCTL_DEBUG=yes netctl start wlp3s0-mywifi
Unfortunately, the output remains the same.
Then there's something else you could try - your wifi card is supported by the b43 driver, all you need to do is install b43-firmware from AUR.
I have been using b43 for a few months. It has a lousy signal strength compared to broadcom-wl (in my Ubuntu installation). I discovered twelveeighty's package today and i hoped i could make Archlinux have the same wireless range, but no game up to now.
EDIT:
NETCTL_DEBUG=yes netctl start wlp3s0-mywifi
Then there's something else you could try - your wifi card is supported by the b43 driver, all you need to do is install b43-firmware from AUR.
]]>