[ltp] Bluetooth button not working after installing Debian 2.6.29 kernel

[Henrique de Moraes Holschuh]

On Thu, 26 Mar 2009, Bjørn Mork wrote:
> Henrique de Moraes Holschuh <hmh@hmh.eng.br> writes:
> > Yes.  Without CONFIG_ACPI_PROC_EVENT you will get hotkey events *ONLY*
> > through the thinkpad-acpi input device.
>
> But it's certainly something for the Debian kernel and acpid maintainers
> to be aware of.

Maybe.  See below.

> > Me, I will move my hotkey acpid scripts to HAL one of these days.  It is
> > where they belong in the new order (until it becomes the NewNew order and
> > HAL beats the dust, that is...).
> 
> Right.  I've been trying to avoid using HAL for this.  Guess there's no
> way around it anymore.

HAL ain't half bad, actually.  I just never bothered to learn how to write
.fdi files, and I probably won't like a lot of stuff I will find inside the
upstream .fdi files when I start looking into them.  But that's not HAL's
fault.

What I *really* dislike is udev.  Underengineered overcomplicated
easy-to-break beast...

> > In fact, CONFIG_ACPI_PROC_EVENT disabled makes Debian acpid USELESS, since
> > it has never been updated to deal with netlink events AFAIK.  My guess is
> > that it is a bug in the kernel config, and that it was not done on purpose.
> 
> I don't think so.  From the changelog:
> 
> linux-2.6 (2.6.29-1) unstable; urgency=low
>   ..
>   [ maximilian attems ]
>   ..
>   * topconfig unset legacy SCSI_PROC_FS, PCMCIA_IOCTL, ACPI_PROCFS_POWER,
>     ACPI_PROC_EVENT.

Hmm, let me check the latest acpid in Debian...  Hey, it has been updated,
my info was outdated.  Check acpid -n.  It now listens over netlink.

This reduces the problem to a thinkpad-acpi issue.  You will need to update
your config to use HAL or something else that can process input events :-(

Looks like I better do it myself ASAP as well, this is going to hit a lot of
users and I better be able to at least tell people how to change their
configs...

> I believe it was discussed on the debian-kernel list before the release,
> but I weren't aware of the acpid problems then.

Apparently, there are only minor bugs if acpid is not running with netlink
by default.  If it is, then there is just the thinkpad-acpi issue.

> Do you think this still should be reported as a kernel bug, or is the
> correct way forward to fix acpid ASAP?

Looks that there are no kernel bugs, and _maybe_ a minor acpid initscript
bug, if all that.  I will check.

-- 
  "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