b43 / LCN PHY

b43 / LCN PHY

[Christopher Horler]

Hi All,

I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
not expecting it to work... at least not yet!

Is anyone looking at this?

Are there any obvious major blocking issues (so as to avoid unnecessary effort)?

Chris


Info:

(This is from my laptop)
For info # lspci -vvn -d 14e4:

03:00.0 Network controller [0280]: Broadcom Corporation BCM4313
802.11bgn Wireless Network Adapter [14e4:4727] (rev 01)
        Subsystem: Wistron NeWeb Corp. Device [185f:051a]
        Flags: bus master, fast devsel, latency 0, IRQ 16
        Memory at fea00000 (64-bit, non-prefetchable) [size=16K]
        Capabilities: [40] Power Management version 3
        Capabilities: [58] Vendor Specific Information: Len=78 <?>
        Capabilities: [48] MSI: Enable- Count=1/1 Maskable- 64bit+
        Capabilities: [d0] Express Endpoint, MSI 00
        Capabilities: [100] Advanced Error Reporting
        Capabilities: [13c] Virtual Channel
        Capabilities: [160] Device Serial Number 00-00-4a-ff-ff-4a-30-14
        Capabilities: [16c] Power Budgeting <?>
        Kernel driver in use: bcma-pci-bridge
        Kernel modules: bcma

After activating allhwsupport=1,  perhaps as expected the wireless doesn't work.

On boot, the code is logging the following message (full dmesg output
is attached)
[   23.058777] b43-phy0 ERROR: SW ctl table is unknown for this card

This seems to be generated on examination of board flags, is there any
way to read /interrogate the current board flags from sysfs or
debugfs?  (or do I need some printk's?)

The kernel log reports a Broadcom Bluetooth adapter showing up on the
USB bus, this is also on-board (perhaps the same module?)

tables_phy_lcn.c-       if (sprom->boardflags_lo & B43_BFL_FEM &&
tables_phy_lcn.c-           !(sprom->boardflags_hi & B43_BFH_FEM_BT))
tables_phy_lcn.c-               b43_lcntab_write_bulk(dev, B43_LCNTAB16(0xf, 0),
tables_phy_lcn.c-
ARRAY_SIZE(b43_lcntab_sw_ctl_4313_epa_rev0),
tables_phy_lcn.c-                       b43_lcntab_sw_ctl_4313_epa_rev0);
tables_phy_lcn.c-       else
tables_phy_lcn.c:               b43err(dev->wl, "SW ctl table is
unknown for this card\n");


dmesg (likely relevant output)
chorler at linux-foxtrot:~/projects/src> grep -iE
'bluetooth|bcm|ieee80211|b43' b43_lcn.log
[    0.000000] BIOS-e820: [mem 0x00000000afb43000-0x00000000afb45fff] ACPI NVS
[    0.000000] PM: Registered nosave memory: [mem 0xafb43000-0xafb45fff]
[    0.167662] PM: Registering ACPI NVS region [mem
0xafb43000-0xafb45fff] (12288 bytes)
[    0.294126] pci 0000:00:14.2: reg 0x10: [mem 0xfeb40000-0xfeb43fff 64bit]
[    3.437464] usb 3-3: Product: Broadcom BCM2070 Bluetooth Device
[   11.954429] bcma: bus0: Found chip with id 0x4313, rev 0x01 and package 0x08
[   11.954465] bcma: bus0: Core 0 found: ChipCommon (manuf 0x4BF, id
0x800, rev 0x24, class 0x0)
[   11.954489] bcma: bus0: Core 1 found: IEEE 802.11 (manuf 0x4BF, id
0x812, rev 0x18, class 0x0)
[   11.954533] bcma: bus0: Core 2 found: PCIe (manuf 0x4BF, id 0x820,
rev 0x11, class 0x0)
[   11.970475] bcma: bus0: Bus registered
[   12.535869] Bluetooth: Core ver 2.17
[   12.535938] Bluetooth: HCI device and connection manager initialized
[   12.535955] Bluetooth: HCI socket layer initialized
[   12.535961] Bluetooth: L2CAP socket layer initialized
[   12.535970] Bluetooth: SCO socket layer initialized
[   19.029813] b43-phy0: Broadcom 4313 WLAN found (core revision 24)
[   19.030736] b43-phy0: Found PHY: Analog 10, Type 8 (LCN), Revision 1
[   19.030752] b43-phy0 debug: Found Radio: Manuf 0x17F, Version
0x2064, Revision 1
[   19.601525] ieee80211 phy0: Selected rate control algorithm 'minstrel_ht'
[   20.457836] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[   20.457846] Bluetooth: BNEP filters: protocol multicast
[   20.457865] Bluetooth: BNEP socket layer initialized
[   23.055667] b43-phy0: Loading firmware version 666.2 (2011-02-23 01:15:07)
[   23.058777] b43-phy0 ERROR: SW ctl table is unknown for this card
[   23.085703] b43-phy0 debug: Chip initialized
[   23.087096] b43-phy0 debug: 64-bit DMA initialized
[   23.087196] b43-phy0 debug: QoS enabled
[   23.088439] b43-phy0 debug: Wireless interface started
[   23.088510] b43-phy0 debug: Adding Interface type 2
[   41.301198] Bluetooth: RFCOMM TTY layer initialized
[   41.301234] Bluetooth: RFCOMM socket layer initialized
[   41.301250] Bluetooth: RFCOMM ver 1.11
-------------- next part --------------
A non-text attachment was scrubbed...
Name: b43_lcn.log
Type: text/x-log
Size: 66088 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/b43-dev/attachments/20131125/4d593971/attachment-0001.bin>

b43 / LCN PHY

[Hauke Mehrtens]

On 11/25/2013 11:33 AM, Christopher Horler wrote:
> Hi All,
> 
> I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
> not expecting it to work... at least not yet!
> 
> Is anyone looking at this?
> 
> Are there any obvious major blocking issues (so as to avoid unnecessary effort)?
> 
> Chris

I do not think anybody will work on adding support for the LCN PHY
(BCM4313) to b43 any more, because brcmsmac already supports this device.

There is some code missing to make this card working in b43, it is not
working at all, you should use brcmsmac.

Hauke

b43 / LCN PHY

[Larry Finger]

On 11/25/2013 05:32 AM, Hauke Mehrtens wrote:
> On 11/25/2013 11:33 AM, Christopher Horler wrote:
>> Hi All,
>>
>> I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
>> not expecting it to work... at least not yet!
>>
>> Is anyone looking at this?
>>
>> Are there any obvious major blocking issues (so as to avoid unnecessary effort)?
>>
>> Chris
>
> I do not think anybody will work on adding support for the LCN PHY
> (BCM4313) to b43 any more, because brcmsmac already supports this device.
>
> There is some code missing to make this card working in b43, it is not
> working at all, you should use brcmsmac.

As the only person working on reverse engineering of Broadcom chips, I am 
certain that anything missing from the LCN specs is unlikely to be added. I 
agree that you should use brcmsmac and report all problems.

Larry

b43 / LCN PHY

[Christopher Horler]

On 25 November 2013 16:42, Larry Finger <Larry.Finger@lwfinger.net> wrote:
> On 11/25/2013 05:32 AM, Hauke Mehrtens wrote:
>>
>> On 11/25/2013 11:33 AM, Christopher Horler wrote:
>>>
>>> Hi All,
>>>
>>> I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
>>> not expecting it to work... at least not yet!
>>>
>>> Is anyone looking at this?
>>>
>>> Are there any obvious major blocking issues (so as to avoid unnecessary
>>> effort)?
>>>
>>> Chris
>>
>>
>> I do not think anybody will work on adding support for the LCN PHY
>> (BCM4313) to b43 any more, because brcmsmac already supports this device.
>>
>> There is some code missing to make this card working in b43, it is not
>> working at all, you should use brcmsmac.
>
>
> As the only person working on reverse engineering of Broadcom chips, I am
> certain that anything missing from the LCN specs is unlikely to be added. I
> agree that you should use brcmsmac and report all problems.
>
> Larry
>

Thanks for the info guys.

okay, will use brcmsmac - I've just noted the same info about LCN PHY
support is also here:
http://wireless.kernel.org/en/users/Drivers/b43#Comparison_of_recent_drivers

I also failed to notice that my system is automatically loading b43,
ssb, brcmsmac and bcma.  Is it normal (or necessary) for both drivers
to be loaded?

Things still seem to work correctly if I blacklist b43.

Chris

b43 / LCN PHY

[Larry Finger]

On 11/25/2013 11:06 AM, Christopher Horler wrote:
> On 25 November 2013 16:42, Larry Finger <Larry.Finger@lwfinger.net> wrote:
>> On 11/25/2013 05:32 AM, Hauke Mehrtens wrote:
>>>
>>> On 11/25/2013 11:33 AM, Christopher Horler wrote:
>>>>
>>>> Hi All,
>>>>
>>>> I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
>>>> not expecting it to work... at least not yet!
>>>>
>>>> Is anyone looking at this?
>>>>
>>>> Are there any obvious major blocking issues (so as to avoid unnecessary
>>>> effort)?
>>>>
>>>> Chris
>>>
>>>
>>> I do not think anybody will work on adding support for the LCN PHY
>>> (BCM4313) to b43 any more, because brcmsmac already supports this device.
>>>
>>> There is some code missing to make this card working in b43, it is not
>>> working at all, you should use brcmsmac.
>>
>>
>> As the only person working on reverse engineering of Broadcom chips, I am
>> certain that anything missing from the LCN specs is unlikely to be added. I
>> agree that you should use brcmsmac and report all problems.
>>
>> Larry
>>
>
> Thanks for the info guys.
>
> okay, will use brcmsmac - I've just noted the same info about LCN PHY
> support is also here:
> http://wireless.kernel.org/en/users/Drivers/b43#Comparison_of_recent_drivers
>
> I also failed to notice that my system is automatically loading b43,
> ssb, brcmsmac and bcma.  Is it normal (or necessary) for both drivers
> to be loaded?
>
> Things still seem to work correctly if I blacklist b43.

As long as both b43 and brcmsmac claim that PCI ID, both will be loaded in the 
absence of any blacklist info. They will likely interfere with each other, and 
blacklisting b43 is the correct step. You could also blacklist ssb, but that 
won't make much difference.

Larry

b43 / LCN PHY

[Rafał Miłecki]

2013/11/25 Christopher Horler <cshorler@googlemail.com>:
> On 25 November 2013 16:42, Larry Finger <Larry.Finger@lwfinger.net> wrote:
>> On 11/25/2013 05:32 AM, Hauke Mehrtens wrote:
>>>
>>> On 11/25/2013 11:33 AM, Christopher Horler wrote:
>>>>
>>>> Hi All,
>>>>
>>>> I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
>>>> not expecting it to work... at least not yet!
>>>>
>>>> Is anyone looking at this?
>>>>
>>>> Are there any obvious major blocking issues (so as to avoid unnecessary
>>>> effort)?
>>>>
>>>> Chris
>>>
>>>
>>> I do not think anybody will work on adding support for the LCN PHY
>>> (BCM4313) to b43 any more, because brcmsmac already supports this device.
>>>
>>> There is some code missing to make this card working in b43, it is not
>>> working at all, you should use brcmsmac.
>>
>>
>> As the only person working on reverse engineering of Broadcom chips, I am
>> certain that anything missing from the LCN specs is unlikely to be added. I
>> agree that you should use brcmsmac and report all problems.
>>
>> Larry
>>
>
> Thanks for the info guys.
>
> okay, will use brcmsmac - I've just noted the same info about LCN PHY
> support is also here:
> http://wireless.kernel.org/en/users/Drivers/b43#Comparison_of_recent_drivers
>
> I also failed to notice that my system is automatically loading b43,
> ssb, brcmsmac and bcma.  Is it normal (or necessary) for both drivers
> to be loaded?
>
> Things still seem to work correctly if I blacklist b43.

This is weird, even if b43 claims 802.11 device on your card, it
should simply exit after detecting unsupported hardware. That should
let brcmsmac load and work just fine.

If that doesn't work that way, I think it's something wrong there.

b43 / LCN PHY

[Larry Finger]

On 11/25/2013 01:54 PM, Rafa? Mi?ecki wrote:
> 2013/11/25 Christopher Horler <cshorler@googlemail.com>:
>> On 25 November 2013 16:42, Larry Finger <Larry.Finger@lwfinger.net> wrote:
>>> On 11/25/2013 05:32 AM, Hauke Mehrtens wrote:
>>>>
>>>> On 11/25/2013 11:33 AM, Christopher Horler wrote:
>>>>>
>>>>> Hi All,
>>>>>
>>>>> I enabled the (broken) LCN PHY support.  I understand it's broken, I'm
>>>>> not expecting it to work... at least not yet!
>>>>>
>>>>> Is anyone looking at this?
>>>>>
>>>>> Are there any obvious major blocking issues (so as to avoid unnecessary
>>>>> effort)?
>>>>>
>>>>> Chris
>>>>
>>>>
>>>> I do not think anybody will work on adding support for the LCN PHY
>>>> (BCM4313) to b43 any more, because brcmsmac already supports this device.
>>>>
>>>> There is some code missing to make this card working in b43, it is not
>>>> working at all, you should use brcmsmac.
>>>
>>>
>>> As the only person working on reverse engineering of Broadcom chips, I am
>>> certain that anything missing from the LCN specs is unlikely to be added. I
>>> agree that you should use brcmsmac and report all problems.
>>>
>>> Larry
>>>
>>
>> Thanks for the info guys.
>>
>> okay, will use brcmsmac - I've just noted the same info about LCN PHY
>> support is also here:
>> http://wireless.kernel.org/en/users/Drivers/b43#Comparison_of_recent_drivers
>>
>> I also failed to notice that my system is automatically loading b43,
>> ssb, brcmsmac and bcma.  Is it normal (or necessary) for both drivers
>> to be loaded?
>>
>> Things still seem to work correctly if I blacklist b43.
>
> This is weird, even if b43 claims 802.11 device on your card, it
> should simply exit after detecting unsupported hardware. That should
> let brcmsmac load and work just fine.
>
> If that doesn't work that way, I think it's something wrong there.

In the very first post in this thread, the OP said that he enabled the LCN PHY 
option. Does that change the behavior?

Larry

b43 / LCN PHY

[Rafał Miłecki]

2013/11/25 Larry Finger <Larry.Finger@lwfinger.net>:
> On 11/25/2013 01:54 PM, Rafa? Mi?ecki wrote:
>> This is weird, even if b43 claims 802.11 device on your card, it
>> should simply exit after detecting unsupported hardware. That should
>> let brcmsmac load and work just fine.
>>
>> If that doesn't work that way, I think it's something wrong there.
>
>
> In the very first post in this thread, the OP said that he enabled the LCN
> PHY option. Does that change the behavior?

Whoops, yeah. You're right. b43_phy_lcn_op_init returns 0 then and b43
tries to work with that card.

-- 
Rafa?