Richard Jonsson wrote:
> Larry Finger skrev:
>> Richard Jonsson wrote:
>>> Larry Finger skrev:
>>>> If Linus's latest git still has the problem, please do a bisection. I
>>> I've now completed the bisection, result below. I'm not really sure
>>> this is the bad commit, as v2.6.25 is fine, and this commit is from
>>> february. But then again, I don't fully understand how git does it's
>>> bisection.
>>> I've also tried a few kernels in the 2.6.25-rc series that don't show
>>> this problem.
>>
>> That part is not too difficult to understand. When the commit was made
>> in February, it was to the wireless-testing tree for inclusion in
>> 2.6.26. Only when Linus went to the 2.6.25-git kernels in the process
>> of getting to 2.6.26-rc1 did that commit make it to mainline.
>>
> Oh, I see.
>
>> Do you have the ssb and b43 debug info from one of the "bad" kernels?
>> Please post that info from /var/log/messages. It seems as though this
>> problem shows on your system because you have a multi-band device.
>>
>
> Attached. To my knowledge this card (4311) has B/G support=2.4GHz only,
> whereas 4312 has A/B/G support=2.4/5GHz.
In a previous post, you gave the following output:
[EMAIL PROTECTED]:~$ lspci -vv -d :4312
01:00.0 Network controller: Broadcom Corporation BCM4312 802.11a/b/g
(rev 01)
Subsystem: Hewlett-Packard Company Unknown device 1361
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR- FastB2B-
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast
>TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 0, Cache Line Size: 64 bytes
Interrupt: pin A routed to IRQ 19
Region 0: Memory at b3000000 (32-bit, non-prefetchable)
[size=16K]
Capabilities: <access denied>
Did this come from your computer? If so, the PCI information is that
it is as 4312, but the chip id, which is what b43 prints out,
identifies it as a 4311. Curious.
I have no idea as to why the interface is switching from 5 to 2.4 GHz,
and back endlessly. Perhaps Michael knows. It is also possible that
the bug is in the new mac80211 band API, which isn't used by b43 when
the "bad" commit is backed out.
The patch contained in
http://lists.berlios.de/pipermail/bcm43xx-dev/2008-April/007423.html
should fix the PHY transmission errors, at least it fixed the one that
was thrown by a BCM4311/2.
BTW, a patch to fix the LED configuration was submitted earlier today
on the linux-wireless mailing list. It should fix the problems you had
in getting the proper config.
Larry
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
