Hi All,
My actually System is:
HP Pavilion Laptop zv5265ea
Debian Lenny(Testing)
Linux Kernel 2.6.28 from kernel.org
Broadcom Wlan card BCM4303
My System doesnt work at all.
I also asked for support on this newsgroup, but without
any success. So i hope there will be sometime a Developer
reading this. :-)
Is a know there is a Problem to activate this card.
Regards Markus
Larry Finger schrieb:
> I looked at that mail again. Have you figured out how to get your radio switch on?
I have takled to Matthew Garret. And he told me to check whether i have
an empty directory </sys/bus/pci/slots/> or not.
Generally it should be not empty if the <acpiphp> modul works fine.
But thats all i know.
Further on here is what Matthgew Garret said:
"Some of these systems seem to use PCI express hotplug to handle the
rfkill. Can you try with the latest git tree, and make sure that you
have the acpiphp driver built in?"
AND
"Go to bus options/Support for PCI Hotplug/ACPI PCI Hotplug driver and
enable it. You'll probably want it built-in to the kernel."
> There are no developers that are working on b43legacy. Whenever a bug is found
> in b43, I will propagate it to b43legacy, but that is the extent of development.
>
> Unfortunately, none of us have a card with an RFKILL switch that uses b43legacy.
> Without hardware in hand, debugging is too difficult. You will need to put
> printk statements in your copy of the driver to find out where it fails.
I have some printk statements activated.
But i don't get more Debug messages than these in my
old threads, or do you know some special debug kernel
Options for wireless or general debug options ?
CONFIG_MAC80211_DEBUGFS
CONFIG_MAC80211_VERBOSE_DEBUG
CONFIG_MAC80211_VERBOSE_PS_DEBUG
Maybe some kernel_hacking Options ?
I can add a little more information. It turns out that I have one PCIe module
that will work with b43legacy as it uses ucode5 firmware. I modified my system
to load b43legacy, rather than b43, for this device. After I did that, the
RFKILL switch works correctly, just as it does in b43.
>From the above experiment, I conclude that the rfkill implementation in
b43legacy is correct. Any problems that you are having are the result of some
quirk in your hardware, therefore, none of us can debug it as we will not be
able to reproduce the problem.
Does your RFKILL switch generate a keycode? You can test that by switching to
console 1 (CTRL/ALT/F1), log in, run the command 'showkey', and operate the
radio switch. On my system, I get a keycode of 238. BTW, use CTRL/ALT/F7 to get
back to the X console.
Larry
Markus Feldmann wrote:
> Larry Finger schrieb:
>>
>> This card works in most cases.
>>
>> What does the command 'dmesg | grep b43' show?
> Hi Larry,
>
> nice to read you again. :-)
>
> The needed modul for this card is <b43-legacy>.
Yes, I know that b43legacy is the right module for this device; however, the
dmesg | grep b43
will work for both b43leagacy and b43, which is why I use that form.
> I talked to you some time ago.
> The subject was <Do not get BCM4303 to work>.
I looked at that mail again. Have you figured out how to get your radio switch on?
> As i understood, only some of the developers for the
> <acpiphp> can solve this Problem, because the card
> will not be activated if this doesn't work.
> For example <M a t t h e w_G a r r e t> i think could solve this.
>
> But he seems not to be online a long time.
>
> So maybe the card could work, but only if i can activate
> it.
>
> Do you know which Developer also works on this Modul ?
There are no developers that are working on b43legacy. Whenever a bug is found
in b43, I will propagate it to b43legacy, but that is the extent of development.
Unfortunately, none of us have a card with an RFKILL switch that uses b43legacy.
Without hardware in hand, debugging is too difficult. You will need to put
printk statements in your copy of the driver to find out where it fails.
Larry
Markus Feldmann wrote:
> Larry Finger schrieb:
>> I looked at that mail again. Have you figured out how to get your
>> radio switch on?
> I have takled to Matthew Garret. And he told me to check whether i have
> an empty directory </sys/bus/pci/slots/> or not.
> Generally it should be not empty if the <acpiphp> modul works fine.
>
> But thats all i know.
My system has an empty /sys/bus/pci/slots.
> Further on here is what Matthgew Garret said:
> "Some of these systems seem to use PCI express hotplug to handle the
> rfkill. Can you try with the latest git tree, and make sure that you
> have the acpiphp driver built in?"
> AND
> "Go to bus options/Support for PCI Hotplug/ACPI PCI Hotplug driver and
> enable it. You'll probably want it built-in to the kernel."
Did you do that?
> I have some printk statements activated.
> But i don't get more Debug messages than these in my
> old threads, or do you know some special debug kernel
> Options for wireless or general debug options ?
>
> CONFIG_MAC80211_DEBUGFS
> CONFIG_MAC80211_VERBOSE_DEBUG
> CONFIG_MAC80211_VERBOSE_PS_DEBUG
I meant new printk statements that are used to see if hitting your switch causes
any action with RFKILL. I don't know where to put them, nor do I want to take
much time. I have an old BCM4306 and a BCM4301 - both work with b43legacy.
Larry
Larry Finger schrieb:
> Markus Feldmann wrote:
>> But thats all i know.
>
> My system has an empty /sys/bus/pci/slots.
Ok, but thats what Matthew Garret told me.
> Did you do that?
Yes i built-in the modul.
>
>> I have some printk statements activated.
>> But i don't get more Debug messages than these in my
>> old threads, or do you know some special debug kernel
>> Options for wireless or general debug options ?
>>
> I meant new printk statements that are used to see if hitting your switch causes
> any action with RFKILL. I don't know where to put them, nor do I want to take
> much time. I have an old BCM4306 and a BCM4301 - both work with b43legacy.
That listens interesting. So if there is a possibility to get more debug
messages for the RFKILL Button, that would be nice. :-)
Regards Markus
Larry Finger schrieb:
>
> This card works in most cases.
>
> What does the command 'dmesg | grep b43' show?
Hi Larry,
nice to read you again. :-)
The needed modul for this card is <b43-legacy>.
I talked to you some time ago.
The subject was <Do not get BCM4303 to work>.
As i understood, only some of the developers for the
<acpiphp> can solve this Problem, because the card
will not be activated if this doesn't work.
For example <M a t t h e w_G a r r e t> i think could solve this.
But he seems not to be online a long time.
So maybe the card could work, but only if i can activate
it.
Do you know which Developer also works on this Modul ?
Regards Markus
Larry Finger schrieb:
> Markus Feldmann wrote:
>> With the kernel version 2.6.28.X it doesn't produce
>> any keycode. I will try out the current developement
>> version in the next days and wrote the results here.
>> Ont this Laptop are 2 Operating Systems. Win XP and Debian Lenny.
>> On WIn Xp the wlan card works fine. So i think the hardware
>> is ok.
>
> The keycode is a function of the hardware, not the kernel. If you don't get one
> now, you will _NEVER_ get one. Your only hope is the hp_wmi driver. I cannot
> help you there.
Thanks for your endevours,
Yesterday i tried out the kernel 2.6.29.7 without
success. <showkey> doesn't show any keycode by pressing
my wlan button.
i will try to contact M a t t h e w_G a r r e t. :-)
Do you know where to get more information about developing
the <hp-wmi> driver ?
Is there anotherone developing this(hp-wmi) driver ?
Regards Markus
Markus Feldmann wrote:
> Hi All,
>
> My actually System is:
> HP Pavilion Laptop zv5265ea
> Debian Lenny(Testing)
> Linux Kernel 2.6.28 from kernel.org
> Broadcom Wlan card BCM4303
This card works in most cases.
What does the command 'dmesg | grep b43' show?
Markus Feldmann wrote:
>
> With the kernel version 2.6.28.X it doesn't produce
> any keycode. I will try out the current developement
> version in the next days and wrote the results here.
> Ont this Laptop are 2 Operating Systems. Win XP and Debian Lenny.
> On WIn Xp the wlan card works fine. So i think the hardware
> is ok.
The keycode is a function of the hardware, not the kernel. If you don't get one
now, you will _NEVER_ get one. Your only hope is the hp_wmi driver. I cannot
help you there.
Larry
As Info,
i also talked about this wlan card in these threads:
1.) <Do not get BCM4303 to work>
2.) <show WMI magic Signals and activate hardware RF-kill button>
3.) <Does ACPIPHP work on my Machine>
Regards Markus
Larry Finger schrieb:
> I can add a little more information. It turns out that I have one PCIe module
> that will work with b43legacy as it uses ucode5 firmware. I modified my system
> to load b43legacy, rather than b43, for this device. After I did that, the
> RFKILL switch works correctly, just as it does in b43.
>
> From the above experiment, I conclude that the rfkill implementation in
> b43legacy is correct. Any problems that you are having are the result of some
> quirk in your hardware, therefore, none of us can debug it as we will not be
> able to reproduce the problem.
>
> Does your RFKILL switch generate a keycode? You can test that by switching to
> console 1 (CTRL/ALT/F1), log in, run the command 'showkey', and operate the
> radio switch. On my system, I get a keycode of 238. BTW, use CTRL/ALT/F7 to get
> back to the X console.
Thank you for your help.
With the kernel version 2.6.28.X it doesn't produce
any keycode. I will try out the current developement
version in the next days and wrote the results here.
Ont this Laptop are 2 Operating Systems. Win XP and Debian Lenny.
On WIn Xp the wlan card works fine. So i think the hardware
is ok.
So please watch here again next week. I think i could fit up
the newest kernel version until next weekend.
Regards Markus