Re: [PATCH]: b43: is there such a thing as radio 0x2062 rev 2?

Re: [PATCH]: b43: is there such a thing as radio 0x2062 rev 2?

[dimitri.gorokhovik]

----- "Gábor Stefanik" <netrolller.3d@gmail.com> a écrit :

> Does the card actually work if you make b43 believe that it has a
> BCM2063 radio?
The interface stays on (the WiFi LED lit), and I can scan and see 
the networks (Quality info seems OK as well with 2.6.35). However, 
I am still unable to connect to either an open or a protected 
network:

Aug 15 19:04:11 dimitri kernel: wlan0: authenticate with 00:17:33:1e:13:70 (try 1)
Aug 15 19:04:11 dimitri kernel: wlan0: authenticated
Aug 15 19:04:11 dimitri kernel: wlan0: associate with 00:17:33:1e:13:70 (try 1)
Aug 15 19:04:11 dimitri kernel: wlan0: RX AssocResp from 00:17:33:1e:13:70 (capab=0x411 status=0 aid=2)
Aug 15 19:04:11 dimitri kernel: wlan0: associated
Aug 15 19:04:16 dimitri kernel: No probe response from AP 00:17:33:1e:13:70 after 500ms, disconnecting.

> BTW, all BCM2062 radios seen so far are rev.2, and none have shown
> this error. Maybe your card actually has a BCM2063 radio, but for
> some reason b43 detects it as a BCM2062?
Sure, this is of course possible (I don't really believe my patch 
fixes the problem). The MAC address of the WiFi card is 
00:22:68:ba:e2:ef (maybe somebody knows know the manufacturing 
series).

For the record, I was able, only once, to run the unmodified driver
(2.6.32 x86+_64 debian kernel), but was never able to get it ever 
again after that. That time, the HDMI res negotiated on both my 
screens (though a dock) was rather unusual (something like 
11xx X 9xx) so I think my problem might be some sort of CEM 
problem.  

> 
> Also, does the Broadcom hybrid driver handle your card?
Last time I tried (more than a year ago), from memory, I believe,
I was able to scan, but I had the problems with Network Manager
support and didn't go for manual config.

Thanks,
Dimitri
> 
> Thanks,
> Gábor.
> 
> On Sun, Aug 15, 2010 at 8:27 PM,  <dimitri.gorokhovik@free.fr> wrote:
> > Hi,
> >
> > On a DELL D630 with BCM4312 (low-power PHY 4e4:4315 rev 01), I was
> able to fix the problem:
> >
> > Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: RC calib: Failed to
> switch to channel 7, error = -5
> >
> > by bumping the radio's version (0x2062 rev 2) to 0x2063 (kept rev
> 2).
> >
> > Before the fix, the hardware detected as:
> >
> > Aug 15 17:14:45 dimitri kernel: b43-phy0: Broadcom 4312 WLAN found
> (core revision 15)
> > Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: Found PHY: Analog 6,
> Type 5, Revision 1
> > Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: Found Radio: Manuf
> 0x17F, Version 0x2062, Revision 2
> >
> > The problem manifested itself (permanently) as:
> >
> > Aug 15 18:18:10 dimitri kernel: b43-phy0: Loading firmware version
> 478.104 (2008-07-01 00:50:23)
> > Aug 15 18:18:10 dimitri kernel: b43-phy0 debug: b2062: Using crystal
> tab entry 19200 kHz.
> > Aug 15 18:18:10 dimitri kernel: b43-phy0 debug: RC calib: Failed to
> switch to channel 7, error = -5
> >
> >
> > The patch follows, tested on 2.6.32.18 and 2.6.35.2, both built for
> x86_64.
> >
> > ---
> > b43: Bump the radio version 0x2063-rev-2 to 0x2063.
> >
> > Signed-off-by: <dimitri.gorokhovik@free.fr>
> > ---
> > diff --git a/drivers/net/wireless/b43/main.c
> b/drivers/net/wireless/b43/main.c
> > index 7965b70..d245a23 100644
> > --- a/drivers/net/wireless/b43/main.c
> > +++ b/drivers/net/wireless/b43/main.c
> > @@ -4084,6 +4084,7 @@ static int b43_phy_versioning(struct b43_wldev
> *dev)
> >        case B43_PHYTYPE_LP:
> >                if (radio_ver != 0x2062 && radio_ver != 0x2063)
> >                        unsupported = 1;
> > +               radio_ver += radio_rev > 1;
> >                break;
> >        default:
> >                B43_WARN_ON(1);
> > --
> > 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: [PATCH]: b43: is there such a thing as radio 0x2062 rev 2?

[Larry Finger]

On 08/15/2010 02:21 PM, dimitri.gorokhovik@free.fr wrote:
> 
> ----- "Gábor Stefanik" <netrolller.3d@gmail.com> a écrit :
> 
>> Does the card actually work if you make b43 believe that it has a
>> BCM2063 radio?
> The interface stays on (the WiFi LED lit), and I can scan and see 
> the networks (Quality info seems OK as well with 2.6.35). However, 
> I am still unable to connect to either an open or a protected 
> network:
> 
> Aug 15 19:04:11 dimitri kernel: wlan0: authenticate with 00:17:33:1e:13:70 (try 1)
> Aug 15 19:04:11 dimitri kernel: wlan0: authenticated
> Aug 15 19:04:11 dimitri kernel: wlan0: associate with 00:17:33:1e:13:70 (try 1)
> Aug 15 19:04:11 dimitri kernel: wlan0: RX AssocResp from 00:17:33:1e:13:70 (capab=0x411 status=0 aid=2)
> Aug 15 19:04:11 dimitri kernel: wlan0: associated
> Aug 15 19:04:16 dimitri kernel: No probe response from AP 00:17:33:1e:13:70 after 500ms, disconnecting.
> 
>> BTW, all BCM2062 radios seen so far are rev.2, and none have shown
>> this error. Maybe your card actually has a BCM2063 radio, but for
>> some reason b43 detects it as a BCM2062?
> Sure, this is of course possible (I don't really believe my patch 
> fixes the problem). The MAC address of the WiFi card is 
> 00:22:68:ba:e2:ef (maybe somebody knows know the manufacturing 
> series).
> 
> For the record, I was able, only once, to run the unmodified driver
> (2.6.32 x86+_64 debian kernel), but was never able to get it ever 
> again after that. That time, the HDMI res negotiated on both my 
> screens (though a dock) was rather unusual (something like 
> 11xx X 9xx) so I think my problem might be some sort of CEM 
> problem.  
> 
>>
>> Also, does the Broadcom hybrid driver handle your card?
> Last time I tried (more than a year ago), from memory, I believe,
> I was able to scan, but I had the problems with Network Manager
> support and didn't go for manual config.

>From a quick look at the latest Broadcom driver that we have for RE, for PHY rev
<= 2, there are times when a 2063 radio is treated as a 2062, but there are
others where it is a 2063.

When I get time, I'll try to sort it out and update the specs. In the meantime,
NACK on the patch. It is certainly incomplete, and does not appear to help.

Larry

Re: [PATCH]: b43: is there such a thing as radio 0x2062 rev 2?

[Gábor Stefanik]

Does the card actually work if you make b43 believe that it has a BCM2063 radio?

BTW, all BCM2062 radios seen so far are rev.2, and none have shown
this error. Maybe your card actually has a BCM2063 radio, but for some
reason b43 detects it as a BCM2062?

Also, does the Broadcom hybrid driver handle your card?

Thanks,
Gábor.

On Sun, Aug 15, 2010 at 8:27 PM,  <dimitri.gorokhovik@free.fr> wrote:
> Hi,
>
> On a DELL D630 with BCM4312 (low-power PHY 4e4:4315 rev 01), I was able to fix the problem:
>
> Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: RC calib: Failed to switch to channel 7, error = -5
>
> by bumping the radio's version (0x2062 rev 2) to 0x2063 (kept rev 2).
>
> Before the fix, the hardware detected as:
>
> Aug 15 17:14:45 dimitri kernel: b43-phy0: Broadcom 4312 WLAN found (core revision 15)
> Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: Found PHY: Analog 6, Type 5, Revision 1
> Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: Found Radio: Manuf 0x17F, Version 0x2062, Revision 2
>
> The problem manifested itself (permanently) as:
>
> Aug 15 18:18:10 dimitri kernel: b43-phy0: Loading firmware version 478.104 (2008-07-01 00:50:23)
> Aug 15 18:18:10 dimitri kernel: b43-phy0 debug: b2062: Using crystal tab entry 19200 kHz.
> Aug 15 18:18:10 dimitri kernel: b43-phy0 debug: RC calib: Failed to switch to channel 7, error = -5
>
>
> The patch follows, tested on 2.6.32.18 and 2.6.35.2, both built for x86_64.
>
> ---
> b43: Bump the radio version 0x2063-rev-2 to 0x2063.
>
> Signed-off-by: <dimitri.gorokhovik@free.fr>
> ---
> diff --git a/drivers/net/wireless/b43/main.c b/drivers/net/wireless/b43/main.c
> index 7965b70..d245a23 100644
> --- a/drivers/net/wireless/b43/main.c
> +++ b/drivers/net/wireless/b43/main.c
> @@ -4084,6 +4084,7 @@ static int b43_phy_versioning(struct b43_wldev *dev)
>        case B43_PHYTYPE_LP:
>                if (radio_ver != 0x2062 && radio_ver != 0x2063)
>                        unsupported = 1;
> +               radio_ver += radio_rev > 1;
>                break;
>        default:
>                B43_WARN_ON(1);
> --
> 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. :-)

[PATCH]: b43: is there such a thing as radio 0x2062 rev 2?

[dimitri.gorokhovik]

Hi,

On a DELL D630 with BCM4312 (low-power PHY 4e4:4315 rev 01), I was able to fix the problem:

Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: RC calib: Failed to switch to channel 7, error = -5

by bumping the radio's version (0x2062 rev 2) to 0x2063 (kept rev 2). 

Before the fix, the hardware detected as:

Aug 15 17:14:45 dimitri kernel: b43-phy0: Broadcom 4312 WLAN found (core revision 15)
Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: Found PHY: Analog 6, Type 5, Revision 1
Aug 15 17:14:45 dimitri kernel: b43-phy0 debug: Found Radio: Manuf 0x17F, Version 0x2062, Revision 2

The problem manifested itself (permanently) as:

Aug 15 18:18:10 dimitri kernel: b43-phy0: Loading firmware version 478.104 (2008-07-01 00:50:23)
Aug 15 18:18:10 dimitri kernel: b43-phy0 debug: b2062: Using crystal tab entry 19200 kHz.
Aug 15 18:18:10 dimitri kernel: b43-phy0 debug: RC calib: Failed to switch to channel 7, error = -5


The patch follows, tested on 2.6.32.18 and 2.6.35.2, both built for x86_64.

---
b43: Bump the radio version 0x2063-rev-2 to 0x2063. 

Signed-off-by: <dimitri.gorokhovik@free.fr>
---
diff --git a/drivers/net/wireless/b43/main.c b/drivers/net/wireless/b43/main.c
index 7965b70..d245a23 100644
--- a/drivers/net/wireless/b43/main.c
+++ b/drivers/net/wireless/b43/main.c
@@ -4084,6 +4084,7 @@ static int b43_phy_versioning(struct b43_wldev *dev)
        case B43_PHYTYPE_LP:
                if (radio_ver != 0x2062 && radio_ver != 0x2063)
                        unsupported = 1;
+		radio_ver += radio_rev > 1;
                break;
        default:
                B43_WARN_ON(1);