[ltp] Wireless hardware radio switch does not work on X60s

[Guillermo Juárez]

I have a x60s right here and I can assure that the kill swictch works
fine at least with the 2.6.20 kernel.

On the other hand, if I switch off wireless, when I switch it on it
tries to use the same ssid. That's a problems, since sometimes I
switch it off and go somewhere else. The only solution to this problem
is unloading the ipw3945 modules and loading it back.

On 5/31/07, Henrique de Moraes Holschuh <hmh@hmh.eng.br> wrote:
> On Wed, 30 May 2007, r8scq7b02@sneakemail.com wrote:
> > Yes. Wireless is working fine, switch is in off position. I even played around
> > with it turning it back and forth but nothing happened. And I must say that
> > it is very poor engineering. It is _very_ difficult to slide it back and
> > forth the way it is intended with finger over the switch, but _too easy_ to
> > switch it when pushing it from the two tiny edges. It almost makes me think
> > it is loose and broken.
>
> Hmm, we need a report from someone else with an Atheros card to know for
> sure, then.
>
> > > Lenovo is not IBM, and doesn't even come close.  I wish I could say they
> > > are working at it, but I don't see that, either.
> >
> > What do you mean by this? Is Lenovo doing a poor job at documentation? Do you
>
> Yes.
>
> > both, which is too bad. I still think that hardware (not compatibility) wise
> > they make the best machines on the market.
>
> We have already lost the good displays to Lenovo.
>
> Lenovo has already reduced the size of useful keys to cram in useless crap
> windows keys.  Not even using them as META or HYPER offsets that.  IBM cared
> for the keyboard a great deal, the only way it could be better would be if
> it used the patented A design, or if it was slightly U-shaped.
>
> Lenovo is starting to break the firmware backwards compatibility right and
> left, which has me worried (but it might be a good thing if they make it
> less braindead and then *stick* to the new interface for five years or so
> without breaking it).  Linux support for the X60 is schizophrenic because of
> that, right now.
>
> They have made the firmware/bios changelogs useless.
>
> They are playing the "old model, no fixes" game with the VT issue on a few
> thinkpads.
>
> Their promissed Linux support is something we were better off without: they
> made a half-hearted attempt at porting their thinkvantage suite to SuSE in
> such a bad way it was clearly rigged to fail (nobody is that incompetent),
> obviously so that some corporate infighting would end up favorable for
> someone in their ranks who doesn't want Linux to get a hold there (no, I am
> not even blaming Redmond for this one).  And now they "have a business case"
> for not doing it right.
>
> They have removed very relevant high-end functionality from the Windows
> thinkvantage suite (a lot of the crypto support being removed is the one I
> noticed immediately, but maybe other stuff was removed too).
>
> No, I really don't see a bright future for the thinkpad line under Lenovo's
> current strategies.
>
> > I believe that the madwifi drivers do have the rfkill attribute [2], but I am
> > not sure how to confirm this on my system. I tested with rfkill=0 and
> > rfkill=1 which produced the same results of the wireless always being on
> > regardless of the switch state.
>
> Well, Atheros drivers (madwifi/madwifi-ng) are *not* known for their
> quality, but since I avoid them like the plague, I don't know if their
> rfkill is supposed to be working or not.
>
> > However, according to the thinkwiki entry [1], a T60 also has a switch that
> > does not operate: "On at least one T60, model 2007-62U with Atheros AR5212,
> > running Ubuntu Edgy with the MadWifi driver, the wireless switch has no
> > effect. Wireless operates with the switch in either position. But there's an
> > experimental patch which adds support of the switch to ibm-acpi."
>
> That would mean the radio hardware has no hardware rf-kill support at all.
> Yuck.
>
> Nothing will help you if the atheros drivers don't implement rfkill. Nothing
> in the thinkpad firmware will kick a PCI/PCIe card off the bus, so if it is
> missing the hardware rf kill line, the firmware *cannot* control the radio
> state.
>
> thinkpad-acpi will, eventually, map that switch to KEY_WLAN and
> KEY_BLUETOOTH, so, *if* madwifi rfkill works, you will be able to get the
> switch to work in software.
>
> > Also, I cannot believe that I never noticed this, but the wireless light is
> > always off, even if the wireless card is turned on and in use. I take it that
> > this is a driver (madwifi) issue then?
>
> Yes.  The intel driver switches that led on/off on software, as well...
> (there is a gpio output in the intel card to mess with the led. I suppose
> atheros does the same).
>
> --
>   "One disk to rule them all, One disk to find them. One disk to bring
>   them all and in the darkness grind them. In the Land of Redmond
>   where the shadows lie." -- The Silicon Valley Tarot
>   Henrique Holschuh
> --
> The linux-thinkpad mailing list home page is at:
> http://mailman.linux-thinkpad.org/mailman/listinfo/linux-thinkpad
>