OpenWrt Forum Archive

Topic: Differences accross bcm43xx Wireless drivers

The content of this topic has been archived on 13 Apr 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

Hello,

Can anyone explain me about the main diferences, perfomance issues, and features acrros de various bcm43xx drivers..

In 8.09 we have a 2.4 ( Broadcom Proprietary driver ) and a 2.6 b43 driver..

In SVN 14658 there are various drivers :

2.4 Kernels :one  brcm proprietary

2.6 Kernels  4 drivers

Kmod-b43 ( stable and experimental versions )
OpenFirmware for wifi networks
Kmod-net-bcm43 :

I look for it in docs and foruns but not find any relelvant information

best regards.

Buildster:

In the trunk, there are three choices of Broadcom drivers... Stable, Experimental, and Open Firmware.

While the Open Firmware shows promise, it's pretty new and somewhat unstable. In my experience, the Stable and Experimental seem to work the same on the brcm63xx target with a BCM4318. The advantage with the Stable and Experimental firmware is support for QoS and CTS.

marc.

> While the Open Firmware shows promise, it's pretty new and somewhat unstable. In my experience, the Stable and Experimental

I am having problems with trunk b43 drivers:

- Poor performance
- can not by pass 1~11 channel maps ( I can use 1~13 here )

Regards,

MPG wrote:

> While the Open Firmware shows promise, it's pretty new and somewhat unstable.

The open firmware has been tested on many different Broadcom cards in AP mode and it works reliably on x86 in our testing, the most important feature missing is support for hardware crypto (which is really important for the slow CPUs found in the WRT family, as using WPA will limit throughput to around 300 KB/s).
It shouldn't be slow when running as an open AP, my 4306 was running pratically at the same speed (~ 2.0 MB/s) as the 2.4 driver.

Last time I tried kamikaze trunk (one of the first revisions with 2.6.28) the b43 driver was hanging randomly, even with the proprietary firmware. It's more reliable now? I'll give a try again next week, hopefully...

ath wrote:
MPG wrote:

> While the Open Firmware shows promise, it's pretty new and somewhat unstable.

The open firmware has been tested on many different Broadcom cards in AP mode and it works reliably on x86 in our testing, the most important feature missing is support for hardware crypto (which is really important for the slow CPUs found in the WRT family, as using WPA will limit throughput to around 300 KB/s).
It shouldn't be slow when running as an open AP, my 4306 was running pratically at the same speed (~ 2.0 MB/s) as the 2.4 driver.

Last time I tried kamikaze trunk (one of the first revisions with 2.6.28) the b43 driver was hanging randomly, even with the proprietary firmware. It's more reliable now? I'll give a try again next week, hopefully...

b43 is working fine for me, a wrt54GS as ap and wrt54GL as stations. I can see 15 Mbps transfer rates. my setup are without crypt

marca56 wrote:

In the trunk, there are three choices of Broadcom drivers... Stable, Experimental, and Open Firmware.

Is "trunk" different from "8.09" ? I was under the impression that the 8.09 branch would only change sporadically for really important backports but in the last two days there have been a flurry of commits.

Also, what is the functional difference between kmod-b43 and kmod-net-bcm43xx?

The discussion might have continued from here.