bcm4322 success!

bcm4322 success!

[Rafał Miłecki]

W dniu 20 lutego 2011 13:40 u?ytkownik Christoph Thielecke
<christoph.thielecke@gmx.de> napisa?:
> Hello,
>
>> > I found your blog entry with the christmas gift:
>> > http://zajec.net/blog/view/2010-b43-and-nphy-rev-3-plus
>> > (Patch: http://marc.info/?l=linux-wireless&m=129296339527734&w=2)
>>
>> Your problems come from fact that you applied only final patch
>> enabling rev3+ support. That way you missed many important earlier
>> patches (for example the one enabling correct firmware loading).
>>
>> Please use sth like wireless-testing git tree to get all needed
>> patches. Using. compat-wireless may do the trick as well, however I
>> didn't try it.
> I can report success! I got it working at 2.6.38-rc5 by specifying options for
> b43 module.
> The specified options are:
> nohwcrypt=1 hwtkip=0 qos=0 pio=1 bad_frames_preempt=1
>
> I'm using it currently ~30 minutes and its working fine.

Thanks for your report! Do you need that all options at all? I think
hwtkip=0 is default one. Parameter qos should not matter. AFAIK pio is
needed for some BCM4321s only. I didn't check what bad_frames_preempt
does.


> wlan0 ? ? IEEE 802.11bg ?ESSID:"xxxxxxxxx"
> ? ? ? ? ?Mode:Managed ?Frequency:2.432 GHz ?Access Point: 00:14:BF:xx:xx:x
> ? ? ? ? ?Bit Rate=11 Mb/s ? Tx-Power=20 dBm
> ? ? ? ? ?Retry ?long limit:7 ? RTS thr:off ? Fragment thr:off
> ? ? ? ? ?Encryption key:off
> ? ? ? ? ?Power Management:off
> ? ? ? ? ?Link Quality=70/70 ?Signal level=0 dBm
> ? ? ? ? ?Rx invalid nwid:0 ?Rx invalid crypt:0 ?Rx invalid frag:0
> ? ? ? ? ?Tx excessive retries:4045 ?Invalid misc:1776 ? Missed beacon:0
>
> Its only 11Mbit (sure its a 54Mbit AP) but speed is constant ~ 500k/s.
>
> Btw: the driver from broadcom (wl) is NOT working but 4e3 does :-D

Hm, that's interesting. wl should work too.

As for speed, it's known issue. I still hope to get patches ready for
2.6.39 merge window.

-- 
Rafa?

bcm4322 success (WPA2)!

[Rafał Miłecki]

Please, include b43-dev in to/cc.

W dniu 20 lutego 2011 14:31 u?ytkownik Christoph Thielecke
<christoph.thielecke@gmx.de> napisa?:
>> Thanks for your report! Do you need that all options at all?
> I think so. I'll try to remove some and tell which is needed for working. Here
> the results:
>
>> I think
>> hwtkip=0 is default one.
> Yes, works without that.
>
>> Parameter qos should not matter.
> qos=0 is nessary, qos=1 or without doesnt work.

What does happen when it doesn't work? Do you get scan results? Can
you associate? Get IP? Transmit?


>> AFAIK pio is
>> needed for some BCM4321s only.
> pio=1 is needed, pio=0 or without doesnt work.
>
>> I didn't check what bad_frames_preempt
>> does.
> bad_frames_preempt=1 is needed, bad_frames_preempt=0 or without doesnt work.
>
> So the needed options are:
> nohwcrypt=1 qos=0 pio=1 bad_frames_preempt=1

What does happen when it doesn't work? Do you get scan results? Can
you associate? Get IP? Transmit?


One request: please blacklist b43, do cold boot (turn off machine for
few seconds), then boot and load b43 with:
pio=1

I ask you for that, because it's possible that using DMA locks up your
card. In such a case further tests can give you false results.
Do you have scan results when using pio=1 after cold boot? Can you associate?

-- 
Rafa?

bcm4322 success (WPA2)!

[Clemens Helfmeier]

Hello Developers,

I have just tried the current wireless-testing git tree with my bcm4322 device:

# lshw
0c:00.0 Network controller [0280]: Broadcom Corporation BCM4322 802.11a/b/g/n
    Wireless LAN Controller [14e4:432b] (rev 01)

# dmesg shows:
[   11.358762] b43-phy0: Broadcom 4322 WLAN found (core revision 16)
[   11.400231] b43-phy0 debug: Found PHY: Analog 8, Type 4, Revision 4 
[   11.400257] b43-phy0 debug: Found Radio: Manuf 0x17F, Version 0x2056, Revision 3

I had to apply this patch to b43-fwcutter in order to extract the firmware: 
http://lists.infradead.org/pipermail/b43-dev/2010-July/000225.html

Using WPA2, the connection is very stable, no connection break downs so far.

My dmesg log shows quite a lot of error messages (approx every 7 seconds in 
idle and more often when traffic is high):
[ 4015.044435] b43-phy0 ERROR: PHY transmission error
[ 4021.910720] b43-phy0 ERROR: PHY transmission error
[ 4028.866892] b43-phy0 ERROR: PHY transmission error

Thanks for your efforts to get my hardware working!

Best regards,
 Clemens

On Sun, Feb 20, 2011 at 02:41:42PM +0100, Rafa? Mi?ecki wrote:
> Please, include b43-dev in to/cc.
> 
> W dniu 20 lutego 2011 14:31 u?ytkownik Christoph Thielecke
> <christoph.thielecke@gmx.de> napisa?:
>>> Thanks for your report! Do you need that all options at all?
>> I think so. I'll try to remove some and tell which is needed for working. 
>> Here the results:
>> 
>>> I think
>>> hwtkip=0 is default one.
>> Yes, works without that.
>> 
>>> Parameter qos should not matter.
>> qos=0 is nessary, qos=1 or without doesnt work.
> 
> What does happen when it doesn't work? Do you get scan results? Can you 
> associate? Get IP? Transmit?
> 
> 
>>> AFAIK pio is
>>> needed for some BCM4321s only.
>> pio=1 is needed, pio=0 or without doesnt work.
>> 
>>> I didn't check what bad_frames_preempt
>>> does.
>> bad_frames_preempt=1 is needed, bad_frames_preempt=0 or without doesnt work.
>> 
>> So the needed options are:
>> nohwcrypt=1 qos=0 pio=1 bad_frames_preempt=1
> 
> What does happen when it doesn't work? Do you get scan results? Can you 
> associate? Get IP? Transmit?
> 
> 
> One request: please blacklist b43, do cold boot (turn off machine for few 
> seconds), then boot and load b43 with:
> pio=1
> 
> I ask you for that, because it's possible that using DMA locks up your card. 
> In such a case further tests can give you false results.
> Do you have scan results when using pio=1 after cold boot? Can you associate?
> 
> --
> Rafa?
> 
> _______________________________________________
> b43-dev mailing list
> b43-dev at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/b43-dev

 Clemens Helfmeier

bcm4322 success (WPA2)!

[Rafał Miłecki]

2011/2/20 Clemens Helfmeier <clemens@helfmeier.de>:
> I had to apply this patch to b43-fwcutter in order to extract the firmware:
> http://lists.infradead.org/pipermail/b43-dev/2010-July/000225.html

Yeah, there was some discussion about getting newer firmware. It's
unfortunately available only in huge firmware and Broadcom doesn't
give a sh* about providing open source community freely available
firmware. That's all about their openess and positive talks brcm80211
related.

For now you have to download huge firmware (and hacking extractor) or
try linking ucode16_mimo to ucode16.


> Using WPA2, the connection is very stable, no connection break downs so far.

We have some positive reports and not-so-much (coming for example from
Larry). Good to head there is one more person having this working.


> My dmesg log shows quite a lot of error messages (approx every 7 seconds in
> idle and more often when traffic is high):
> [ 4015.044435] b43-phy0 ERROR: PHY transmission error
> [ 4021.910720] b43-phy0 ERROR: PHY transmission error
> [ 4028.866892] b43-phy0 ERROR: PHY transmission error
>
> Thanks for your efforts to get my hardware working!

Just for sure: can you try forcing rate to 1M, 2M or 11M? Does it
noticeably lowers amount of reported PHY errors? The command you need
to use is like:
iwconfig wlan0 rate 1M
iwconfig wlan0 rate 2M
iwconfig wlan0 rate 11M

-- 
Rafa?

bcm4322 success (WPA2)!

[Tom Gundersen]

Hi Rafa?,

I also have a bcm4322 chip, which works with WPA2 using 2.6.38-rc5.

2011/2/20 Rafa? Mi?ecki <zajec5@gmail.com>:
> For now you have to download huge firmware (and hacking extractor) or
> try linking ucode16_mimo to ucode16.

This is the only thing I had to do to make it work.

> 2011/2/20 Clemens Helfmeier <clemens@helfmeier.de>:
>> My dmesg log shows quite a lot of error messages (approx every 7 seconds in
>> idle and more often when traffic is high):
>> [ 4015.044435] b43-phy0 ERROR: PHY transmission error
>> [ 4021.910720] b43-phy0 ERROR: PHY transmission error
>> [ 4028.866892] b43-phy0 ERROR: PHY transmission error

I'm seeing about the same here.

Furthermore, I get about 5% packet loss when pinging my wireless
router (with no other traffic).

> Just for sure: can you try forcing rate to 1M, 2M or 11M? Does it
> noticeably lowers amount of reported PHY errors? The command you need
> to use is like:
> iwconfig wlan0 rate 1M
> iwconfig wlan0 rate 2M
> iwconfig wlan0 rate 11M

iwconfig says that my rate is 11 Mb/s, but since I expected it to be
54, I figured this might be wrong, so I tried
% iwconfig wlan0 rate 11M

This appears to solve the problem completely! I have not seen any more
errors and my packet loss is 0%. (The download speed is still
suboptimal, but I read that you already have some patches for that, so
I'll just wait for when they are posted.)

Thanks for making this work! (It is infinitely better than the wl
driver...). If ever you need some testing of new patches related to my
card, let me know.

Cheers,

Tom


FYI:

% lspci -vvv -d 43
lspci -vvv -d :432b
0c:00.0 Network controller: Broadcom Corporation BCM4322 802.11a/b/g/n
Wireless LAN Controller (rev 01)
        Subsystem: Dell Wireless 1510 Wireless-N WLAN Mini-Card
        Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop-
ParErr- Stepping- SERR+ FastB2B- DisINTx-
        Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort-
<TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 0, Cache Line Size: 64 bytes
        Interrupt: pin A routed to IRQ 17
        Region 0: Memory@f69fc000 (64-bit, non-prefetchable) [size=16K]
        Capabilities: <access denied>
        Kernel driver in use: b43-pci-bridge
        Kernel modules: ssb

% dmesg | grep b43 | grep -v error
[    4.634710] b43-pci-bridge 0000:0c:00.0: PCI INT A -> GSI 17
(level, low) -> IRQ 17
[    4.634726] b43-pci-bridge 0000:0c:00.0: setting latency timer to 64
[    5.081082] b43-phy0: Broadcom 4322 WLAN found (core revision 16)
[    5.120182] b43-phy0 debug: Found PHY: Analog 8, Type 4, Revision 4
[    5.120211] b43-phy0 debug: Found Radio: Manuf 0x17F, Version
0x2056, Revision 3
[    5.153525] Registered led device: b43-phy0::tx
[    5.153598] Registered led device: b43-phy0::rx
[    5.153660] Registered led device: b43-phy0::radio
[    7.170069] b43-phy0: Loading firmware version 478.104 (2008-07-01 00:50:23)
[    7.300260] b43-phy0 debug: Chip initialized
[    7.300426] b43-phy0 debug: 64-bit DMA initialized
[    7.300514] b43-phy0 debug: QoS enabled
[    7.317592] b43-phy0 debug: Wireless interface started
[    7.317611] b43-phy0 debug: Adding Interface type 2
[   13.439529] b43-phy0 debug: Using hardware based encryption for
keyidx: 0, mac: 5c:33:8e:ea:24:f3
[   13.439619] b43-phy0 debug: Using hardware based encryption for
keyidx: 1, mac: ff:ff:ff:ff:ff:ff

bcm4322 success (WPA2)!

[Larry Finger]

On 02/20/2011 12:25 PM, Clemens Helfmeier wrote:
> Hello Developers,
>
> I have just tried the current wireless-testing git tree with my bcm4322 device:
>
> # lshw
> 0c:00.0 Network controller [0280]: Broadcom Corporation BCM4322 802.11a/b/g/n
>      Wireless LAN Controller [14e4:432b] (rev 01)
>
> # dmesg shows:
> [   11.358762] b43-phy0: Broadcom 4322 WLAN found (core revision 16)
> [   11.400231] b43-phy0 debug: Found PHY: Analog 8, Type 4, Revision 4
> [   11.400257] b43-phy0 debug: Found Radio: Manuf 0x17F, Version 0x2056, Revision 3
>
> I had to apply this patch to b43-fwcutter in order to extract the firmware:
> http://lists.infradead.org/pipermail/b43-dev/2010-July/000225.html
>
> Using WPA2, the connection is very stable, no connection break downs so far.
>
> My dmesg log shows quite a lot of error messages (approx every 7 seconds in
> idle and more often when traffic is high):
> [ 4015.044435] b43-phy0 ERROR: PHY transmission error
> [ 4021.910720] b43-phy0 ERROR: PHY transmission error
> [ 4028.866892] b43-phy0 ERROR: PHY transmission error
>
> Thanks for your efforts to get my hardware working!

The attached patch should fix the PHY transmission errors, which happens 
whenever the rate control code tries to set an OFDM rate. With the patch, those 
rates will not work, but no errors will be printed.

You can also lock the rate at 11M (iwconfig wlan0 rate 11M) to get rid of those 
errors.

Larry

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: b43_tx_set_control_word1
URL: <http://lists.infradead.org/pipermail/b43-dev/attachments/20110220/581de113/attachment.ksh>

bcm4322 success (WPA2)!

[Clemens Helfmeier]

Hello Larry,

On Sun, Feb 20, 2011 at 01:30:36PM -0600, Larry Finger wrote:
> On 02/20/2011 12:25 PM, Clemens Helfmeier wrote:
>> Hello Developers,
>> 
>> I have just tried the current wireless-testing git tree with my bcm4322 
>> device:
>> 
>> # lshw
>> 0c:00.0 Network controller [0280]: Broadcom Corporation BCM4322 
>> 802.11a/b/g/n
>>     Wireless LAN Controller [14e4:432b] (rev 01)
>> 
>> # dmesg shows:
>> [   11.358762] b43-phy0: Broadcom 4322 WLAN found (core revision 16) [   
>> 11.400231] b43-phy0 debug: Found PHY: Analog 8, Type 4, Revision 4 [   
>> 11.400257] b43-phy0 debug: Found Radio: Manuf 0x17F, Version 0x2056, 
>> Revision 3
>> 
>> I had to apply this patch to b43-fwcutter in order to extract the firmware: 
>> http://lists.infradead.org/pipermail/b43-dev/2010-July/000225.html
>> 
>> Using WPA2, the connection is very stable, no connection break downs so far.
>> 
>> My dmesg log shows quite a lot of error messages (approx every 7 seconds in 
>> idle and more often when traffic is high):
>> [ 4015.044435] b43-phy0 ERROR: PHY transmission error
>> [ 4021.910720] b43-phy0 ERROR: PHY transmission error
>> [ 4028.866892] b43-phy0 ERROR: PHY transmission error
>> 
>> Thanks for your efforts to get my hardware working!
> 
> The attached patch should fix the PHY transmission errors, which happens 
> whenever the rate control code tries to set an OFDM rate. With the patch, 
> those rates will not work, but no errors will be printed.
> 
> You can also lock the rate at 11M (iwconfig wlan0 rate 11M) to get rid of 
> those errors.

Locking the rate at 11M makes the driver not output any more error messages 
here ,too.
Thanks for your work!

Best regards,
 Clemens

bcm4322 success (WPA2)!

[Rafał Miłecki]

W dniu 20 lutego 2011 19:30 u?ytkownik Christoph Thielecke
<christoph.thielecke@gmx.de> napisa?:
> I played a little bit with the options and the only stable working case seems
> nohwcrypt=1 hwtkip=0 qos=0 pio=1 bad_frames_preempt=1
>
> I dont know why the other cases sometimes work.

Thanks for your tests, however I can not offer you anything right now.
I hope it will get more stable when we get TX patch, some general PHY
rev 3+ improvements and DMA fix. Will try to remember to ping you then
:)

-- 
Rafa?