Skip to contentSkip to conversation

Bug 10173 / Bluetooth on Toshiba & Compal laptops

DISCUSSION Robert M. Albrecht 3 participants 6 messages Dec 15, 2008 – Dec 23, 2008

lore
Share & tools
lore
#0 lore raw

Hi,

anyone working on toshiba_acpi or omnibook module ?

toshiba_acpi:

http://git.kernel.org/?p=linux/kernel/git/linville/wireless-next-2.6.git;a=commit;h=c41a40c58b623f97dd2e23aa3baab2cedb47d99f

The rfkill-switch does not work. The device is disabled when switched of, but it gets not reenabled. But the bluetooth dongle is enabled after booting.

AFAIK this only works for Toshiba manufactured devices (toshiba_acpi), not for Compal manufactured.

Omnibook:

Omnibook module functionality for Compal manufactured lptops should still be included in the kernel. The Compal supporting module omnibook additionaly works on many laptops from HP, ... The omnibook-module seems to be abandoned, the last commit is nearly two years ago.

Omnibook-module: http://sourceforge.net/projects/omnibook

cu romal

www.romal.de
blog.romal.de

#1 lore raw
Quoted 7 lines · Robert M. Albrecht

Hi,

anyone working on toshiba_acpi or omnibook module ?

toshiba_acpi:

http://git.kernel.org/?p=linux/kernel/git/linville/wireless-next-2.6.git;a=commit;h=c41a40c58b623f97dd2e23aa3baab2cedb47d99f

Quoted 7 lines

The rfkill-switch does not work. The device is disabled when switched of,

but it gets not reenabled. But the bluetooth dongle is enabled after booting.

AFAIK this only works for Toshiba manufactured devices (toshiba_acpi),

not

Quoted 1 line

for Compal manufactured.

The officially stated maintainer for toshiba_acpi seems to be out of contact,
but I'm responsible for the rfkill patch. I can only say that the patch is designed
to re-enable bluetooth when the hardware kill switch is de-asserted and that I tested
this to work before the change was merged. If it doesn't work now, it's either:

a) a regression introduced by one of the fixes for laptops with no bluetooth device. I
consider this unlikely.

b) different behaviour in your model of toshiba vs. mine

c) a difference in how you've got the kill switch configured in sysfs. Not sure if that's
even possible.

As you might conclude, I suspect B.

--phil

#2 lore raw

Hi Phil,

anything I can do ? Providing some logs or granting ssh access to the machine to have a look ?

cu romal

www.romal.de
blog.romal.de

Am 15.12.2008 23:35 Uhr, schrieb Philip Langdale:

Quoted 42 lines
Quoted 7 lines · Robert M. Albrecht

Hi,

anyone working on toshiba_acpi or omnibook module ?

toshiba_acpi:

http://git.kernel.org/?p=linux/kernel/git/linville/wireless-next-2.6.git;a=commit;h=c41a40c58b623f97dd2e23aa3baab2cedb47d99f

Quoted 6 lines

The rfkill-switch does not work. The device is disabled when switched of,

but it gets not reenabled. But the bluetooth dongle is enabled after booting.

AFAIK this only works for Toshiba manufactured devices (toshiba_acpi),

not

Quoted 1 line

for Compal manufactured.

The officially stated maintainer for toshiba_acpi seems to be out of contact,
but I'm responsible for the rfkill patch. I can only say that the patch is designed
to re-enable bluetooth when the hardware kill switch is de-asserted and that I tested
this to work before the change was merged. If it doesn't work now, it's either:

a) a regression introduced by one of the fixes for laptops with no bluetooth device. I
consider this unlikely.

b) different behaviour in your model of toshiba vs. mine

c) a difference in how you've got the kill switch configured in sysfs. Not sure if that's
even possible.

As you might conclude, I suspect B.

--phil

#3 lore raw
Quoted 4 lines · Robert M. Albrecht

Hi Phil,

anything I can do ? Providing some logs or granting ssh access to the machine to have a look ?

If necessary, that would be useful. However, the first question to ask is actually
related to my possibility C) - which I'm now re-evaluating as much more likely.

If you don't load the 'rfkill-input' module, then there is no policy applied to
killswitch handling. The core driver puts the device in the SOFT_BLOCKED state
when the de-assert the hardware switch. rfkill-input contains the policy that
decides to go one step further and re-enable the device. This module has no hardware dependencies so it is not automatically loaded - you must do it explicitly yourself.

If that's not your problem, then we'll have to dig deeper.

--phil

#4 lore raw

Hi,

the module is loaded:

[root@tycho ~]# lsmod | grep kill
rfkill 11288 2 toshiba_acpi
[root@tycho ~]#

The wlan-devices works ok, it get's disabled and enabled via rfkill. But the bluetooth-device does not get reeanbled after switching it off.

cu romal

Am 15.12.2008 23:47 Uhr, schrieb Philip Langdale:

Quoted 27 lines
Quoted 4 lines · Robert M. Albrecht

Hi Phil,

anything I can do ? Providing some logs or granting ssh access to the machine to have a look ?

If necessary, that would be useful. However, the first question to ask is actually
related to my possibility C) - which I'm now re-evaluating as much more likely.

If you don't load the 'rfkill-input' module, then there is no policy applied to
killswitch handling. The core driver puts the device in the SOFT_BLOCKED state
when the de-assert the hardware switch. rfkill-input contains the policy that
decides to go one step further and re-enable the device. This module has no hardware dependencies so it is not automatically loaded - you must do it explicitly yourself.

If that's not your problem, then we'll have to dig deeper.

--phil

#5 lore raw
Quoted 9 lines · Robert M. Albrecht

Hi,

the module is loaded:

[root@tycho ~]# lsmod | grep kill
rfkill 11288 2 toshiba_acpi
[root@tycho ~]#

The wlan-devices works ok, it get's disabled and enabled via rfkill. But the bluetooth-device does not get reeanbled after switching it off.

Umm. I'll quote you quoting me.

Quoted 5 lines

Am 15.12.2008 23:47 Uhr, schrieb Philip Langdale:

Quoted 3 lines

If you don't load the 'rfkill-input' module, then there is no policy applied to killswitch handling.

--phil