Not your usual kernel topic but, I need an answer

[MickeyRat]

This is not about cool stuff like kernel programming but, I need an answer from someone that knows the kernel. Please run this through your clueless filter and try to translate what I mean if I get some terminology wrong.

This is Linux Mint 12 64 bit 3.0.0-12-generic kernel on an HP Probook 6560 with a Broadcom BCM43224 network card. There are a lot of problems with the driver for those Broadcom cards but, I'm not sure that's my problem.

This laptop does have a wireless button but, it looks like the wireless is bocked no matter what I do with it. Here's what it does:

> rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: yes
1: hp-wifi: Wireless LAN
	Soft blocked: no
	Hard blocked: yes

Press wireless button

> rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: yes
1: hp-wifi: Wireless LAN
	Soft blocked: no
	Hard blocked: no

Turn wireless button off in BIOS.

> rfkill list all
0: phy0: Wireless LAN
	Soft blocked: no
	Hard blocked: yes
1: hp-wifi: Wireless LAN
	Soft blocked: no
	Hard blocked: no

As expected pressing the button with it turned off in the BIOS has no effect. rfkill unblock all/wifi/wlan has no effect. I also tried blacklisting the hp_wmi. All that did was eliminate the hp-wifi status altogether. The phy0 remained hard blocked.

Everything that I can find on rfkill says it's reporting on some status registers in the kernel and that a hard block only occurs due to hardware. If that is the case, then there's no way a driver can influence it. Am I right about that? I need to know because there are literally thousands of threads on drivers for these cards but, if it's between the kernel and the hardware, I can quit looking at those.

[Rubberman]

What driver and firmware have you installed? In any case, DO visit Welcome - Linux Wireless - their docs and links (especially for Broadcom gear) have helped me tremendously in the past.

[MickeyRat]

What driver and firmware have you installed? In any case, DO visit (I have to have 15 posts to quote the link) - their docs and links (especially for Broadcom gear) have helped me tremendously in the past.

Thank you very much for the reply.

I've been to that site and still no joy. The shorter answer would be which driver haven't I installed. I've tried everything I can find up to and including the windows driver with ndiswrapper and downloading the driver source from Broadcom and compiling it myself. However, if the problem's with the kernel thinking that the hardware switch is off when it's not, nothing I do with the driver is going to change it. That's really the question I need answered. Can the driver cause a hardware block in rfkill with no software block?

Everything I can find about rfkill says no and I'm banging my head against a wall messing with the driver. However, the places that lead me to think that go over my head in a hurry. So, I could be reading it wrong.

My wife's laptop has a different Broadcom card (BCM4311) and loading the right driver got it working. This one has eluded me so far.

[Rubberman]

This may help:

On 09/09/2010, Broadcom released an open driver for their SSB-less chipsets. However, it only supported three chips (bcm4313, bcm43224, bcm43225). The driver is merged into kernel 2.6.37 under brcm80211. More information is available on this release at Broadcom releases an open-source driver for its wireless chipsets [LWN.net].