HAL fails to initialize on ASUS A6 series notebooks

Bug #62990 reported by elele
10
Affects Status Importance Assigned to Milestone
linux-source-2.6.19 (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

When logging to Gnome desktop, a popup window appears every time with the text "Internal error. Failed to initialize HAL!". This is on a fresh install AMD64 Edge Beta on an ASUS A6Km notebook. Dapper never had this problem.

I found a similar bug report which suggested the problem is related to smbfs /cifs entry in fstab. I don't have any entries like that.

Strange thing is that hal daemon seems to be running just fine. I have processes for hald, hald-runner, hald-addon-acpi, hald-addon-keyboard and hald-addon-storage. I also tried adding --verbose=yes --use-syslog to DAEMON_OPTS in /etc/default/hal. That produced lots of debugging info in /var/log/daemon.log, but I couldn't find anything related to this problem.

I'm not sure if this is related, but Device Manager also does not work. If I run hal-device-manager from the command line, the following error is displayed:

<gtk.Menu object (GtkMenu) at 0x2b0a32f209b0>
Traceback (most recent call last):
  File "/usr/bin/hal-device-manager", line 20, in ?
    DeviceManager()
  File "/usr/share/hal/device-manager/DeviceManager.py", line 77, in __init__
    lambda *args: self.gdl_changed("DeviceAdded", *args))
  File "/var/lib/python-support/python2.4/dbus/_dbus.py", line 291, in connect_to_signal
    self._obj.connect_to_signal(signal_name, handler_function, dbus_interface, **keywords)
  File "/var/lib/python-support/python2.4/dbus/proxies.py", line 151, in connect_to_signal
    path=self._object_path,
  File "/var/lib/python-support/python2.4/dbus/_dbus.py", line 179, in add_signal_receiver
    named_service = bus_object.GetNameOwner(named_service, dbus_interface='org.freedesktop.DBus')
  File "/var/lib/python-support/python2.4/dbus/proxies.py", line 25, in __call__
    ret = self._proxy_method (*args, **keywords)
  File "/var/lib/python-support/python2.4/dbus/proxies.py", line 102, in __call__
    reply_message = self._connection.send_with_reply_and_block(message, timeout)
  File "dbus_bindings.pyx", line 455, in dbus_bindings.Connection.send_with_reply_and_block
dbus_bindings.DBusException: Could not get owner of name 'org.freedesktop.Hal': no such name

Revision history for this message
zekus (zekus-one) wrote :

Same error here on an Asus A6kt laptop with amd64 Edgy Beta.
Think that this error appeared after last dbus update.
Kernel updated to 2.6.17.

Revision history for this message
elele (elele-deactivatedaccount) wrote :

This bug still exists in Edgy RC.

Also I noticed that suspend / hibernate does not work because of this bug. These options are not displayed on the System > Quit... menu. Very problematic for notebook users.

On a custom compiled 2.6.18 kernel suspend and hibernate work fine and I also don't get the HAL error message. So maybe this is a kernel problem?

Revision history for this message
Matthew Sherborne (matt-sherborne) wrote :

Same happening for me on ASUS A6km (AMD64) too, after upgrade to edgy.

Also the gnome power manager always thinks I'm plugged in to mains power.

Also no suspend/hibernate, nor even power off when shut down.

I've attached the output of my dmesg...

Changed in hal:
status: Unconfirmed → Confirmed
Revision history for this message
zekus (zekus-one) wrote :

Anyone has found a solution to this?
Edgy is out and I still have this error...

Revision history for this message
elele (elele-deactivatedaccount) wrote :

I'm affraid not zekus. The only workaround I've found is to build your own custom 2.6.18 kernel. Whatever is causing this problem seems to be fixed there.

I did a some HAL debugging and found out the hald process is indeed defunct (marked Uninterruptible sleep / session leader). There's no way to kill it, not even with KILL signal. Here's the whole hald ps listing after system boot.

root 4211 0.0 0.1 10260 1632 ? Ss 03:28 0:00 /bin/sh /etc/dbus-1/event.d/20hal start
root 4221 0.0 0.1 14228 1092 ? S 03:28 0:00 /usr/sbin/hald
106 4222 8.4 0.5 21176 5816 ? Ds 03:28 0:02 /usr/sbin/hald
root 4223 0.1 0.1 11304 1140 ? S 03:28 0:00 hald-runner
106 4229 0.0 0.0 9288 876 ? S 03:28 0:00 /usr/lib/hal/hald-addon-acpi
106 4244 0.0 0.0 9288 872 ? S 03:28 0:00 /usr/lib/hal/hald-addon-keyboard
106 4260 0.0 0.0 9292 896 ? S 03:28 0:00 /usr/lib/hal/hald-addon-storage

I also managed to start HAL in verbose mode by first disabling it at /etc/dbus-1/event.d/20hal. This produced a big text log, which I'll add as an attachment. Seems like it's hanging at acpi LID... or at least that's the last log entry before the process hangs.

Oh well... I hope someone can make better sense out of this. I'm short on ideas as to what to try next. :(

Revision history for this message
zekus (zekus-one) wrote :

tried a fresh edgy installation... nothing new, same error!
This is really frustrating... can't use any external device via usb nor the internal dvd-writer.

Revision history for this message
elele (elele-deactivatedaccount) wrote :
Revision history for this message
elele (elele-deactivatedaccount) wrote :
Revision history for this message
elele (elele-deactivatedaccount) wrote :

Here's a partial solution that works at least on my A6Km. Fire up your text editor and point it at /etc/modprobe.d/blacklist. Add the following line and then reboot.

blacklist asus_acpi

For some reason the hal daemon in Edgy doesn't like this module (Asus notebook acpi extras) on the A6 series notebooks. This line will prevent the module from loading and make hald, suspend etc. work again.

The downside is you'll lose some acpi controlled functionality, like the LCD brightness buttons.

Revision history for this message
zekus (zekus-one) wrote :

I'm sorry but ... it doesnt worked! I disabled it as you have suggested but "failed to initialize HAL" is still there. I hope, however that you are next to the solution... maybe there is something else you have modified in your configuration that, with the previous workaround, made the world turn again!
I blacklisted also ibm_acpi and pcc_acpi after a first reboot with asus_acpi only disabled but... nothing.

Attached my dmesg.

Revision history for this message
elele (elele-deactivatedaccount) wrote :

That is strange. On my notebook blacklisting asus_acpi really does help... maybe it only works on the A6Km. :/

Could you try blacklisting some of the other acpi modules? You can find them all in the directory below. You also need to add seperate lines in /etc/modprobe.d/blacklist for every module... otherwise only the first module is counted for.

/lib/modules/2.6.17-10-generic/kernel/drivers/acpi

Since hald is hanging on acpi lid, the button acpi module is a good candidate. I tried this on my notebook and got rid of the hal error. This also disabled suspend / hibernate functions, which I guess is the expected behaviour.

Maybe you could even try booting with noacpi - just to see if the hal error message disappears.

Revision history for this message
zekus (zekus-one) wrote :

I've tried to start with pci=noacpi and pci=noapic but... nothing
the error is still there.
I'm really thinking to compile the kernel 2.6.18, but, before I'll try to start with hald manually like you have done to see where it stops.
Anyway, attached is my dmesg with acpi disabled.

Revision history for this message
Matthew Sherborne (matt-sherborne) wrote : Re: [Bug 62990] Re: HAL fails to initialize on ASUS A6 series notebooks

Hi Zekus,

in grub you can put on the end:

acpi=off

And get no ACPI, (which is practically the same as what we have, but with no
HAL message :)

Hope this helps.

Blessings
Matthew

On 11/9/06, zekus <email address hidden> wrote:
>
> I've tried to start with pci=noacpi and pci=noapic but... nothing
> the error is still there.
> I'm really thinking to compile the kernel 2.6.18, but, before I'll try to
> start with hald manually like you have done to see where it stops.
> Anyway, attached is my dmesg with acpi disabled.
>
> ** Attachment added: "pci=noacpi pci=noapic"
> http://librarian.launchpad.net/5000853/dmesg.txt
>
> --
> HAL fails to initialize on ASUS A6 series notebooks
> https://launchpad.net/bugs/62990
>

Revision history for this message
zekus (zekus-one) wrote :

Ok... acpi=off worked but...
it is not a solution! A laptop is not a laptop without ACPI so... let's find a real fix to this bug. How can I help?

Revision history for this message
Peter Salomonsen (petersalo) wrote :

I have a ASUS A6KM (AMD64) laptop and what I have noticed is that if I don't plug in the AC adapter, HAL seems to work, and I don't get this message. I can also run commands such as lshal.

However when I plug in the AC adapter, or if I boot with the AC adapter, I'll get this message "Failed to Initialize HAL". If running lshal I get this:

error: libhal_ctx_init: (null): (null)
Could not initialise connection to hald.
 Normally this mean the HAL daemon (hald) is not running or not ready.

if I do ps -ax, this is the state of my hal processes:

4359 ? Ds 0:02 /usr/sbin/hald
 4360 ? S 0:00 hald-runner
 4366 ? S 0:00 /usr/lib/hal/hald-addon-acpi
 4369 ? S 0:00 /usr/lib/hal/hald-addon-keyboard
 4372 ? S 0:00 /usr/lib/hal/hald-addon-keyboard
 4382 ? S 0:00 /usr/lib/hal/hald-addon-keyboard
 4396 ? S 0:00 /usr/lib/hal/hald-addon-storage

Revision history for this message
Martin Pitt (pitti) wrote :

If blacklisting the asus_acpi module worked, then it's a kernel bug. If it worked with 2.6.18, it should also work with 2.6.20, thus I'm closing this since Feisty has 2.6.20. Testing with Feisty appreciated!

Please cry out if it still does not work in Feisty, then I'll reopen.

Changed in hal:
status: Confirmed → Fix Released
Revision history for this message
Peter Salomonsen (petersalo) wrote :

Feisty works perfect with asus a6km!

cheers,

Peter

On Wednesday 10 January 2007 14:06, Martin Pitt wrote:
> If blacklisting the asus_acpi module worked, then it's a kernel bug. If
> it worked with 2.6.18, it should also work with 2.6.20, thus I'm closing
> this since Feisty has 2.6.20. Testing with Feisty appreciated!
>
> Please cry out if it still does not work in Feisty, then I'll reopen.
>
> ** Changed in: hal (Ubuntu)
> Sourcepackagename: hal => linux-source-2.6.19
> Status: Confirmed => Fix Released

Revision history for this message
zekus (zekus-one) wrote :

I was forced to manually upgrade kernel to 2.6.19 to make my laptop works.
I'll test Feisty...

Revision history for this message
batteryfast (batteryfast2001) wrote :

I have this type of laptop. Now, the laptop battery last only half a hour so I got a battery from http://www.batterygoshop.co.uk/asus/a6-series-battery.htm . I hope it can help you.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.