b43: regression on bcm4318

b43: regression on bcm4318

[Vitaly Bordug]

Hello John et al,
(previous email had a typo in ML addr, sorry for the extra noise)

Both b43 and b43legacy fail to load on bcm4318 for a while now, while it used to work (with known issues) more or less
fine. The issue started happening not long before ssb united the support of b44 as well. Therefor, I noticed the difference
soon, because I am not able to use ndiswrapper instead anymore (it does not like ssb, and my NIC, b44, does not like
ssb missing either).

dmesg looks like the below:

b43-phy1: Broadcom 4318 WLAN found
b43-phy1 debug: Found PHY: Analog 3, Type 2, Revision 7
b43-phy1 debug: Found Radio: Manuf 0x17F, Version 0x2050, Revision 8
phy1: Failed to select rate control algorithm
phy1: Failed to initialize rate control algorithm
b43: probe of ssb0:0 failed with error -2

Currently this is latest rawhide kernel, but has been happening for awhile
(since, at least, 2.6.22.5, where ssb was not yet merged and I still able to use ndiswrapper.

Any hint to debug/investigate it appreciated. It does not look like hw malfunction, since ndiswrapper,
when able to load, still works fine.

-- 
Sincerely, Vitaly

Re: b43: regression on bcm4318

[Michael Buesch]

On Saturday 13 October 2007 10:17:30 Vitaly Bordug wrote:
> phy1: Failed to select rate control algorithm
> phy1: Failed to initialize rate control algorithm

That's nothing b43 or ssb related.
Compile and load the rc80211_simple module.

-- 
Greetings Michael.

Re: b43: regression on bcm4318

[Vitaly Bordug]

Hello Michael,

On Sat, 13 Oct 2007 10:24:47 +0200
Michael Buesch wrote:

> On Saturday 13 October 2007 10:17:30 Vitaly Bordug wrote:
> > phy1: Failed to select rate control algorithm
> > phy1: Failed to initialize rate control algorithm
> 
> That's nothing b43 or ssb related.
> Compile and load the rc80211_simple module.
> 
oh, thanks, that was broken module dependencies. Well at least it is saved in archive now, no nobody else
would step into it.

Another question: is this driver known to work okay in ad-hoc or master mode? I gave it a try and there is no
driver-side whine, but it does not seem to work either. I see couple of packages turnaround, sometimes even
dhcp server is able to feed IP via wlan (host is acting as kinda AP), but nothing after that.

And, soon after actual b43 replaced bcm... in rawhide, the same worked, e.g. for at least 5 minutes my wifi-enabled
E61 was able stay up and connected (b43 was in ad-hoc mode that time).

I would appreciate any feedback.
TIA

-- 
Sincerely, Vitaly

Re: b43: regression on bcm4318

[Michael Buesch]

On Saturday 13 October 2007 16:14:31 Vitaly Bordug wrote:
> Hello Michael,
> 
> On Sat, 13 Oct 2007 10:24:47 +0200
> Michael Buesch wrote:
> 
> > On Saturday 13 October 2007 10:17:30 Vitaly Bordug wrote:
> > > phy1: Failed to select rate control algorithm
> > > phy1: Failed to initialize rate control algorithm
> > 
> > That's nothing b43 or ssb related.
> > Compile and load the rc80211_simple module.
> > 
> oh, thanks, that was broken module dependencies. Well at least it is saved in archive now, no nobody else
> would step into it.
> 
> Another question: is this driver known to work okay in ad-hoc or master mode? I gave it a try and there is no
> driver-side whine, but it does not seem to work either. I see couple of packages turnaround, sometimes even
> dhcp server is able to feed IP via wlan (host is acting as kinda AP), but nothing after that.

Master and adhoc is not stable, yet.
It might work, though. But I didn't really test that recently.

-- 
Greetings Michael.

Re: b43: regression on bcm4318

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 511 bytes --]

On Sat, 2007-10-13 at 18:14 +0400, Vitaly Bordug wrote:

> Another question: is this driver known to work okay in ad-hoc or master mode? I gave it a try and there is no
> driver-side whine, but it does not seem to work either. I see couple of packages turnaround, sometimes even
> dhcp server is able to feed IP via wlan (host is acting as kinda AP), but nothing after that.

b43 works in master mode with all my hostapd enablement patches
(johannes.sipsolutions.net/patches/{hostap,kernel}

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 828 bytes --]