Problem with BCM943224

Problem with BCM943224

[Rafał Miłecki]

I've 14e4:4353 labeled as BCM943224HMS / DW1520

[  199.009268] brcm80211: module is from the staging directory, the
quality is unknown, you have been warned.
[  199.022493] brcm80211 0000:03:00.0: enabling device (0000 -> 0002)
[  199.022529] brcm80211 0000:03:00.0: PCI INT A -> GSI 28 (level,
low) -> IRQ 28
[  199.022543] brcm80211 0000:03:00.0: setting latency timer to 64
[  199.050090] cfg80211: World regulatory domain updated:
[  199.050096] cfg80211:     (start_freq - end_freq @ bandwidth),
(max_antenna_gain, max_eirp)
[  199.050101] cfg80211:     (2402000 KHz - 2472000 KHz @ 40000 KHz),
(300 mBi, 2000 mBm)
[  199.050105] cfg80211:     (2457000 KHz - 2482000 KHz @ 20000 KHz),
(300 mBi, 2000 mBm)
[  199.050109] cfg80211:     (2474000 KHz - 2494000 KHz @ 20000 KHz),
(300 mBi, 2000 mBm)
[  199.050113] cfg80211:     (5170000 KHz - 5250000 KHz @ 40000 KHz),
(300 mBi, 2000 mBm)
[  199.050117] cfg80211:     (5735000 KHz - 5835000 KHz @ 40000 KHz),
(300 mBi, 2000 mBm)
[  199.297803] Found chip of unknown type (0xffffffff)
[  199.297808] wl0: wlc_bmac_attach: si_attach failed
[  199.297813] wl0: wlc_bmac_attach: failed with err 11
[  199.297818] wl0: wlc_attach: failed with err 11
[  199.297856] brcm80211: wlc_attach() failed with code 11
[  199.297882] brcm80211: wl_pci_probe: wl_attach failed!

[  343.223842] wl: module license 'unspecified' taints kernel.
[  343.223849] Disabling lock debugging due to kernel taint
[  343.256069] wl 0000:03:00.0: setting latency timer to 64
[  343.258760] eth%d: 5.100.82.38 driver failed with code 11

It seems all reads are 1 (0xffffffff). Can this be some bug in driver?
Can this be because I use adapter mini PCIe to PCIe x1? I remember
such a situation when I connected adapter to PCIe x16 and used N-PHY
card SSB based.

-- 
Rafał

Re: Problem with BCM943224

[Larry Finger]

On 03/09/2011 04:59 AM, Rafał Miłecki wrote:
> I've 14e4:4353 labeled as BCM943224HMS / DW1520
>
> [  199.009268] brcm80211: module is from the staging directory, the
> quality is unknown, you have been warned.
> [  199.022493] brcm80211 0000:03:00.0: enabling device (0000 ->  0002)
> [  199.022529] brcm80211 0000:03:00.0: PCI INT A ->  GSI 28 (level,
> low) ->  IRQ 28
> [  199.022543] brcm80211 0000:03:00.0: setting latency timer to 64
> [  199.050090] cfg80211: World regulatory domain updated:
> [  199.050096] cfg80211:     (start_freq - end_freq @ bandwidth),
> (max_antenna_gain, max_eirp)
> [  199.050101] cfg80211:     (2402000 KHz - 2472000 KHz @ 40000 KHz),
> (300 mBi, 2000 mBm)
> [  199.050105] cfg80211:     (2457000 KHz - 2482000 KHz @ 20000 KHz),
> (300 mBi, 2000 mBm)
> [  199.050109] cfg80211:     (2474000 KHz - 2494000 KHz @ 20000 KHz),
> (300 mBi, 2000 mBm)
> [  199.050113] cfg80211:     (5170000 KHz - 5250000 KHz @ 40000 KHz),
> (300 mBi, 2000 mBm)
> [  199.050117] cfg80211:     (5735000 KHz - 5835000 KHz @ 40000 KHz),
> (300 mBi, 2000 mBm)
> [  199.297803] Found chip of unknown type (0xffffffff)
> [  199.297808] wl0: wlc_bmac_attach: si_attach failed
> [  199.297813] wl0: wlc_bmac_attach: failed with err 11
> [  199.297818] wl0: wlc_attach: failed with err 11
> [  199.297856] brcm80211: wlc_attach() failed with code 11
> [  199.297882] brcm80211: wl_pci_probe: wl_attach failed!
>
> [  343.223842] wl: module license 'unspecified' taints kernel.
> [  343.223849] Disabling lock debugging due to kernel taint
> [  343.256069] wl 0000:03:00.0: setting latency timer to 64
> [  343.258760] eth%d: 5.100.82.38 driver failed with code 11
>
> It seems all reads are 1 (0xffffffff). Can this be some bug in driver?
> Can this be because I use adapter mini PCIe to PCIe x1? I remember
> such a situation when I connected adapter to PCIe x16 and used N-PHY
> card SSB based.

A read of all 1's is a symptom of reading a non-existent register. I would 
certainly suspect the adapter.

I have a PCI Express extender for my HP laptop. Although it avoids the problem 
of the white-listed WLAN modules in the BIOS, not all cards work with it. 
Perhaps there is a critical timing issue.

I just tested again. My BCM4322 (14e4:432b) works with both b43 and wl in the 
extender. For the record, the upload speeds are 8 Mbps with b43 and 15 Mbps with 
wl in the 2.4 GHz band (router max 54 Mbps). In the 5 GHz band, the router is 
set for 270 Mbps max and wl gets 80 Mbps. Of course, the 5 GHz band in b43 is 
inactive.

Larry

Re: Problem with BCM943224

[Rafał Miłecki]

W dniu 9 marca 2011 16:44 użytkownik Larry Finger
<Larry.Finger@lwfinger.net> napisał:
> On 03/09/2011 04:59 AM, Rafał Miłecki wrote:
>>
>> I've 14e4:4353 labeled as BCM943224HMS / DW1520
>>
>> [  199.009268] brcm80211: module is from the staging directory, the
>> quality is unknown, you have been warned.
>> [  199.022493] brcm80211 0000:03:00.0: enabling device (0000 ->  0002)
>> [  199.022529] brcm80211 0000:03:00.0: PCI INT A ->  GSI 28 (level,
>> low) ->  IRQ 28
>> [  199.022543] brcm80211 0000:03:00.0: setting latency timer to 64
>> [  199.050090] cfg80211: World regulatory domain updated:
>> [  199.050096] cfg80211:     (start_freq - end_freq @ bandwidth),
>> (max_antenna_gain, max_eirp)
>> [  199.050101] cfg80211:     (2402000 KHz - 2472000 KHz @ 40000 KHz),
>> (300 mBi, 2000 mBm)
>> [  199.050105] cfg80211:     (2457000 KHz - 2482000 KHz @ 20000 KHz),
>> (300 mBi, 2000 mBm)
>> [  199.050109] cfg80211:     (2474000 KHz - 2494000 KHz @ 20000 KHz),
>> (300 mBi, 2000 mBm)
>> [  199.050113] cfg80211:     (5170000 KHz - 5250000 KHz @ 40000 KHz),
>> (300 mBi, 2000 mBm)
>> [  199.050117] cfg80211:     (5735000 KHz - 5835000 KHz @ 40000 KHz),
>> (300 mBi, 2000 mBm)
>> [  199.297803] Found chip of unknown type (0xffffffff)
>> [  199.297808] wl0: wlc_bmac_attach: si_attach failed
>> [  199.297813] wl0: wlc_bmac_attach: failed with err 11
>> [  199.297818] wl0: wlc_attach: failed with err 11
>> [  199.297856] brcm80211: wlc_attach() failed with code 11
>> [  199.297882] brcm80211: wl_pci_probe: wl_attach failed!
>>
>> [  343.223842] wl: module license 'unspecified' taints kernel.
>> [  343.223849] Disabling lock debugging due to kernel taint
>> [  343.256069] wl 0000:03:00.0: setting latency timer to 64
>> [  343.258760] eth%d: 5.100.82.38 driver failed with code 11
>>
>> It seems all reads are 1 (0xffffffff). Can this be some bug in driver?
>> Can this be because I use adapter mini PCIe to PCIe x1? I remember
>> such a situation when I connected adapter to PCIe x16 and used N-PHY
>> card SSB based.
>
> A read of all 1's is a symptom of reading a non-existent register. I would
> certainly suspect the adapter.
>
> I have a PCI Express extender for my HP laptop. Although it avoids the
> problem of the white-listed WLAN modules in the BIOS, not all cards work
> with it. Perhaps there is a critical timing issue.
>
> I just tested again. My BCM4322 (14e4:432b) works with both b43 and wl in
> the extender. For the record, the upload speeds are 8 Mbps with b43 and 15
> Mbps with wl in the 2.4 GHz band (router max 54 Mbps). In the 5 GHz band,
> the router is set for 270 Mbps max and wl gets 80 Mbps. Of course, the 5 GHz
> band in b43 is inactive.

BCM4322 is not a problem for me as well. I can use BCM4312, BCM4321,
BCM4322 with my adapter without problem.

It's BCM943224 that I hit this problem with.

AFAIR you have some AI based card. Could you test this with your
extender for me?

-- 
Rafał

Re: Problem with BCM943224

[Larry Finger]

On 03/09/2011 10:26 AM, Rafał Miłecki wrote:
>
> BCM4322 is not a problem for me as well. I can use BCM4312, BCM4321,
> BCM4322 with my adapter without problem.
>
> It's BCM943224 that I hit this problem with.
>
> AFAIR you have some AI based card. Could you test this with your
> extender for me?

Sorry, I have no AI-based card. I checked for a BCM4313 on Ebay a couple of 
weeks ago and couldn't find one. I know that Dell ships that card in some 
laptops, but I couldn't find the part on their web site. Does anyone know the 
Dell part number?

Larry

Re: Problem with BCM943224

[Rafał Miłecki]

W dniu 9 marca 2011 17:38 użytkownik Larry Finger
<Larry.Finger@lwfinger.net> napisał:
> On 03/09/2011 10:26 AM, Rafał Miłecki wrote:
>>
>> BCM4322 is not a problem for me as well. I can use BCM4312, BCM4321,
>> BCM4322 with my adapter without problem.
>>
>> It's BCM943224 that I hit this problem with.
>>
>> AFAIR you have some AI based card. Could you test this with your
>> extender for me?
>
> Sorry, I have no AI-based card. I checked for a BCM4313 on Ebay a couple of
> weeks ago and couldn't find one. I know that Dell ships that card in some
> laptops, but I couldn't find the part on their web site. Does anyone know
> the Dell part number?

For 4313 just type BCM94313 on ebay. Unfortunately price is quite
high, about $40 - $50.

For 43224, just type BCM943224, it's much cheaper, about $10 - $25.

In few days I can give you HP model I got shipped with 4313.

-- 
Rafał

Re: Problem with BCM943224

[Gábor Stefanik]

2011/3/9 Larry Finger <Larry.Finger@lwfinger.net>:
> On 03/09/2011 10:26 AM, Rafał Miłecki wrote:
>>
>> BCM4322 is not a problem for me as well. I can use BCM4312, BCM4321,
>> BCM4322 with my adapter without problem.
>>
>> It's BCM943224 that I hit this problem with.
>>
>> AFAIR you have some AI based card. Could you test this with your
>> extender for me?
>
> Sorry, I have no AI-based card. I checked for a BCM4313 on Ebay a couple of
> weeks ago and couldn't find one. I know that Dell ships that card in some
> laptops, but I couldn't find the part on their web site. Does anyone know
> the Dell part number?

Not Dell, but HP: 593836-001

http://shop.ebay.com/?_from=R40&_trksid=m570&_nkw=593836-001

Sells below $15.

>
> Larry
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>



-- 
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)

Re: Problem with BCM943224

[Larry Finger]

On 03/09/2011 10:53 AM, Gábor Stefanik wrote:
> 2011/3/9 Larry Finger<Larry.Finger@lwfinger.net>:
>> On 03/09/2011 10:26 AM, Rafał Miłecki wrote:
>>>
>>> BCM4322 is not a problem for me as well. I can use BCM4312, BCM4321,
>>> BCM4322 with my adapter without problem.
>>>
>>> It's BCM943224 that I hit this problem with.
>>>
>>> AFAIR you have some AI based card. Could you test this with your
>>> extender for me?
>>
>> Sorry, I have no AI-based card. I checked for a BCM4313 on Ebay a couple of
>> weeks ago and couldn't find one. I know that Dell ships that card in some
>> laptops, but I couldn't find the part on their web site. Does anyone know
>> the Dell part number?
>
> Not Dell, but HP: 593836-001
>
> http://shop.ebay.com/?_from=R40&_trksid=m570&_nkw=593836-001
>
> Sells below $15.

Gábor and Rafał,

Thanks for the links and search hints. I just bought a BCM94313 and a BCM943224. 
When they arrive, I'll let you know what I learn about using them on the extender.

Larry

Re: Problem with BCM943224

[Rafał Miłecki]

Do you think buying some more expensive adapter could help? The newest
offered by hwtools seems to be:
http://www.hwtools.net/Adapter/MP2H.html
(ver 2.2). It is not so expensive at all, $15 without shipping.

I've already written to them with question about BCM943224
compatibility (they don't have that chipset on their list).

Stanisław: I remember you telling that you have such a adapter. Do you
maybe have access to some BCM943224/BCM94313 card? Could you test that
combo for me, please?

The only other solution I can see is to buy motherboard with mini PCIe
slot. There are some *D410P* (BLKD410PT, BOXD410PT, BLKD410PTL,
D410PT) and D510MO having such a slow (and Atom CPU.... :/). Kernel
compilation would be disaster but I think I could live with that. Do
you think such a boards would be compatible with Broadcom's cards?

-- 
Rafał

Re: Problem with BCM943224

[Larry Finger]

On 03/09/2011 12:04 PM, Rafał Miłecki wrote:
>
> The only other solution I can see is to buy motherboard with mini PCIe
> slot. There are some *D410P* (BLKD410PT, BOXD410PT, BLKD410PTL,
> D410PT) and D510MO having such a slow (and Atom CPU.... :/). Kernel
> compilation would be disaster but I think I could live with that. Do
> you think such a boards would be compatible with Broadcom's cards?

I have a system with an Atom processor and one with an AMD 450 MHz CPU. I do not 
try to build kernels on either of them. I have their sources on a 2.4 GHz 
dual-core machine in an NFS-exported volume. I cross-build for i386 as the host 
is 64 bits and only do the "make modules_install install" on the target 
machines. It works quite well.

Larry

Re: Problem with BCM943224

[Stanislaw Gruszka]

On Wed, Mar 09, 2011 at 07:04:01PM +0100, Rafał Miłecki wrote:
> Do you think buying some more expensive adapter could help? The newest
> offered by hwtools seems to be:
> http://www.hwtools.net/Adapter/MP2H.html
> (ver 2.2). It is not so expensive at all, $15 without shipping.
> 
> I've already written to them with question about BCM943224
> compatibility (they don't have that chipset on their list).
> 
> Stanisław: I remember you telling that you have such a adapter. Do you
> maybe have access to some BCM943224/BCM94313 card? Could you test that
> combo for me, please?

I have new laptop at home with some bcm device, don't know exacly what
version, but most likely is somehting that you listed or similar. I'll
test and let you now next week. I planned to replace wireless adapter
on that laptop anyway.

> The only other solution I can see is to buy motherboard with mini PCIe
> slot. There are some *D410P* (BLKD410PT, BOXD410PT, BLKD410PTL,
> D410PT) and D510MO having such a slow (and Atom CPU.... :/). Kernel
> compilation would be disaster but I think I could live with that.

If you configure out all unnedded drivers/features, compilation time
should not be so bad.

> Do
> you think such a boards would be compatible with Broadcom's cards?

I think any miniPCIe host should be compatibile with any miniPCIe device.

Cheers
Stanislaw

Re: Problem with BCM943224

[Stanislaw Gruszka]

On Thu, Mar 10, 2011 at 10:27:53AM +0100, Stanislaw Gruszka wrote:
> > Stanisław: I remember you telling that you have such a adapter. Do you
> > maybe have access to some BCM943224/BCM94313 card? Could you test that
> > combo for me, please?
> 
> I have new laptop at home with some bcm device, don't know exacly what
> version, but most likely is somehting that you listed or similar. I'll
> test and let you now next week. I planned to replace wireless adapter
> on that laptop anyway.

BCM943224HMS works here with hwtools.net PCI-E Adapter Ver 2.03.

Stanislaw

Re: Problem with BCM943224

[Rafał Miłecki]

W dniu 9 marca 2011 17:38 użytkownik Larry Finger
<Larry.Finger@lwfinger.net> napisał:
> On 03/09/2011 10:26 AM, Rafał Miłecki wrote:
>>
>> BCM4322 is not a problem for me as well. I can use BCM4312, BCM4321,
>> BCM4322 with my adapter without problem.
>>
>> It's BCM943224 that I hit this problem with.
>>
>> AFAIR you have some AI based card. Could you test this with your
>> extender for me?
>
> Sorry, I have no AI-based card. I checked for a BCM4313 on Ebay a couple of
> weeks ago and couldn't find one. I know that Dell ships that card in some
> laptops, but I couldn't find the part on their web site. Does anyone know
> the Dell part number?

I guess you do not need that anymore, but I said to check, so... That
notebook that was shipped with BCM4313 is HP G62-A10sw.

I've taken BCM4313 from that notebook and put it in my cheap adapter.
It didn't work (again, all reads 1). It seems that my adapter does not
like any AI based card. It accepts SSB based cards only. At least now
I know it is not my BCM943223 that is faulty, but adapter. I'm sure
BCM4313 is fine, because it works OK in that HP notebook.

Bad news: that dumb, dumb, dumb HP has 14e4:4727 (BCM4313)
whitelisted, but 14e4:4353 (BCM43224) is blacklisted. Who the hell
designed that :|

-- 
Rafał

Re: Problem with BCM943224

[Larry Finger]

On 03/10/2011 11:40 AM, Rafał Miłecki wrote:
>
> I guess you do not need that anymore, but I said to check, so... That
> notebook that was shipped with BCM4313 is HP G62-A10sw.
>
> I've taken BCM4313 from that notebook and put it in my cheap adapter.
> It didn't work (again, all reads 1). It seems that my adapter does not
> like any AI based card. It accepts SSB based cards only. At least now
> I know it is not my BCM943223 that is faulty, but adapter. I'm sure
> BCM4313 is fine, because it works OK in that HP notebook.
>
> Bad news: that dumb, dumb, dumb HP has 14e4:4727 (BCM4313)
> whitelisted, but 14e4:4353 (BCM43224) is blacklisted. Who the hell
> designed that :|

The lawyers. HP's argument is that it cannot allow any combination that has not 
been tested for regulatory compliance. It carries about as much validity as 
Broadcom's lawyers not allowing redistribution of firmware for regulatory 
reasons. I've gotten quite good at hot plugging the PCIe card while at the GRUB 
prompt. Someday, I'll get mad enough to rewrite that part of the BIOS.

Other than these screwy BIOS problems, I like HP laptops, but I don't think I'll 
get another of theirs due to this whitelisting problem. I think Dell does not 
have this difficulty. Unfortunately, they seem to offer very few setups with AMD 
processors - my choice.

Larry

Re: Problem with BCM943224

[Arend van Spriel]

Rafał Miłecki <zajec5@...> writes:

> 
> It seems all reads are 1 (0xffffffff). Can this be some bug in driver?
> Can this be because I use adapter mini PCIe to PCIe x1? I remember
> such a situation when I connected adapter to PCIe x16 and used N-PHY
> card SSB based.
> 

I am using a Dell laptop (Latitude E6410) with 43224 installed and no
issues with it. We also have a test setup over here which is a desktop
model in which we use the 4313 in half-mini-pci to PCIe x1 adapter. Also
had 43224 in same adapter and both did not give issues with the si_attach,
which scans the chipset for the cores. Reading back 0xFFFFFFFF is reason
to abort this scan and end in a failure.

Gr. AvS

Re: Problem with BCM943224

[Rafał Miłecki]

2011/3/10 Arend van Spriel <arend@broadcom.com>:
> Rafał Miłecki <zajec5@...> writes:
>> It seems all reads are 1 (0xffffffff). Can this be some bug in driver?
>> Can this be because I use adapter mini PCIe to PCIe x1? I remember
>> such a situation when I connected adapter to PCIe x16 and used N-PHY
>> card SSB based.
>>
>
> I am using a Dell laptop (Latitude E6410) with 43224 installed and no
> issues with it. We also have a test setup over here which is a desktop
> model in which we use the 4313 in half-mini-pci to PCIe x1 adapter. Also
> had 43224 in same adapter and both did not give issues with the si_attach,
> which scans the chipset for the cores. Reading back 0xFFFFFFFF is reason
> to abort this scan and end in a failure.

Thank you for answer. Could you send me photo of your adapter? Do you
know where it was bought?

Could you (Broadcom) donate me with a one or two such a adapters?

-- 
Rafał

Re: Problem with BCM943224

[Arend van Spriel]

On Thu, 10 Mar 2011 11:56:44 +0100, Rafał Miłecki <zajec5@gmail.com> wrote:

> 2011/3/10 Arend van Spriel <arend@broadcom.com>:> Thank you for answer.  
> Could you send me photo of your adapter? Do you
> know where it was bought?

We are using extender from: www.adexelec.com.

PEX1-MINI+PEMINI-F2H: mini to PCIe 1x adapter + full to half mini adapter
PEX1-MINI-H: Half-mini to PCIe 1x adapter

> Could you (Broadcom) donate me with a one or two such a adapters?

I can bring it up over here, but no promises ;-)

Gr. AvS
-- 
"The most merciful thing in the world, I think, is the inability of the  
human
mind to correlate all its contents." - "The Call of Cthulhu"

Re: Problem with BCM943224

[Rafał Miłecki]

W dniu 10 marca 2011 12:15 użytkownik Arend van Spriel
<arend@broadcom.com> napisał:
> On Thu, 10 Mar 2011 11:56:44 +0100, Rafał Miłecki <zajec5@gmail.com> wrote:
>
>> 2011/3/10 Arend van Spriel <arend@broadcom.com>:> Thank you for answer.
>> Could you send me photo of your adapter? Do you
>> know where it was bought?
>
> We are using extender from: www.adexelec.com.
>
> PEX1-MINI+PEMINI-F2H: mini to PCIe 1x adapter + full to half mini adapter
> PEX1-MINI-H: Half-mini to PCIe 1x adapter

Huh, that adapter costs €80 without shipping without tax, I can not
afford it :( It'd be much easier to use adapter instead of new
netbook/motherboard, but not for such a price :|

Stanisław: let me know about your results, please.

-- 
Rafał

Re: Problem with BCM943224

[Rafał Miłecki]

W dniu 9 marca 2011 11:59 użytkownik Rafał Miłecki <zajec5@gmail.com> napisał:
> It seems all reads are 1 (0xffffffff). Can this be some bug in driver?
> Can this be because I use adapter mini PCIe to PCIe x1? I remember
> such a situation when I connected adapter to PCIe x16 and used N-PHY
> card SSB based.

Now, this is interesting. As my last step before ordering motherboard
with mini PCIe slot, I decided to test my adapter on another machine.

Motherboard in my main PC is P5V02-MX with some magic switch that let
me select between:
1) Ex-SATA
2) PCI-E*1

So I put my adapter with BCM943224 card into ASRock K8NF4G-SATA2
and... it worked with Ubuntu 11.04 LiveCD and brcm80211!

My conclusion:
It seems my P5V02-MX motherboard with that magic Ex-SATA/PCIe x1
switch is in conflict with AI boards. Maybe that switch generated some
timing delays that lead to all register reads being 1?

I'll try to exchange my PC with the one working with ASRock.

-- 
Rafał