HP Compaq nc6320

HP Compaq nc6320

[Peter Clifton]

I decided to revert the HP bios back to F.06, and now - to my surprise,
on boot, I get an extra SSDT! (The CpuPm one).

[17179569.184000] ACPI: RSDP (v000
HP                                    ) @ 0x000f7d70
[17179569.184000] ACPI: RSDT (v001 HP     30AA     0x28040620 HP
0x00000001) @ 0x3f7e5684
[17179569.184000] ACPI: FADT (v002 HP     30AA     0x00000002 HP
0x00000001) @ 0x3f7e5600
[17179569.184000] ACPI: MADT (v001 HP     30AA     0x00000001 HP
0x00000001) @ 0x3f7e56c0
[17179569.184000] ACPI: MCFG (v001 HP     30AA     0x00000001 HP
0x00000001) @ 0x3f7e5728
[17179569.184000] ACPI: TCPA (v002 HP     30AA     0x00000001 HP
0x00000001) @ 0x3f7e5764
[17179569.184000] ACPI: SSDT (v001 HP       HPQSAT 0x00000001 MSFT
0x0100000e) @ 0x3f7f4af8
[17179569.184000] ACPI: SSDT (v001 HP        CpuPm 0x00003000 INTL
0x20050624) @ 0x3f7f5327
[17179569.184000] ACPI: DSDT (v001 HP       nc6340 0x00010000 MSFT
0x0100000e) @ 0x00000000

The CpuPm SSDT table contains _PDC and _OSC methods, and generally looks
useful.

Still, linux doesn't load either speedstep-centrino, or acpi-speedstep.
I'll poke some more, but it looks promising.

I wounder why the BIOS F.08 doesn't have the CpuPm SSDT? It appeared to
flash correctly when I first installed it. Perhaps even booting windows
(to downgrade to F.06) had some effect on the bios to bring back the
SSDT? Perhaps it is only in the F.06.

Does anyone have any contacts with HP who might shed any light on this?

Regards

Peter Clifton

Re: HP Compaq nc6320

[Peter Clifton]

On Wed, 2006-10-11 at 23:08 +0100, Peter Clifton wrote:
> I decided to revert the HP bios back to F.06, and now - to my
surprise,
> on boot, I get an extra SSDT! (The CpuPm one).

[snip]

> Perhaps even booting windows
> (to downgrade to F.06) had some effect on the bios to bring back the
> SSDT? Perhaps it is only in the F.06.

Further to this.. the Battery / AC adapter status monitoring seems to be
working now aswell (if booting with ec_intr=0). I just booted with out
ec_intr=0, and it isn't working :(

I'm just about to check with the F.06 DSDT again, as my override DSDT is
based on F.08. (There isn't much difference between them though)

How odd!

Peter C.

Re: HP Compaq nc6320

[Peter Clifton]

On Wed, 2006-10-11 at 23:17 +0100, Peter Clifton wrote:
> On Wed, 2006-10-11 at 23:08 +0100, Peter Clifton wrote:
> > I decided to revert the HP bios back to F.06, and now - to my
> surprise,
> > on boot, I get an extra SSDT! (The CpuPm one).
> 
> [snip]
> 
> > Perhaps even booting windows
> > (to downgrade to F.06) had some effect on the bios to bring back the
> > SSDT? Perhaps it is only in the F.06.
> 
> Further to this.. the Battery / AC adapter status monitoring seems to
be
> working now aswell (if booting with ec_intr=0). I just booted with out
> ec_intr=0, and it isn't working :(
> 
> I'm just about to check with the F.06 DSDT again, as my override DSDT
is
> based on F.08. (There isn't much difference between them though)

Of course, this was _necessary_, as the namespaces between the two DSDT
is different, and the SSDT won't play nice if it can't find the correct
objects, methods, etc...

Wierdly, I've lost the Battery/AC adapter status monitoring again, with
or without ec_intr=0.

> How odd!

Gets odder!

Peter C.

Re: HP Compaq nc6320

[Peter Clifton]

> Wierdly, I've lost the Battery/AC adapter status monitoring again, with
> or without ec_intr=0.

> Gets odder!

Ok,

Nearing completion of this tale I think..

Turns out there is a known problem with psmouse, which leaves the
system / ACPI / something in a bad state if psmouse is loaded before
shutdown or reboot. Googling tells me that either removing batteries, AC
etc.. before powering up, booting Windows, or shutting down Linux with
psmouse removed, will "fix" the symptoms.

Magic...

Removing psmouse before boot from the init scripts, I reboot (no
ec_intr=0), and not only does the battery status work properly, the CPU
frequency scaling also works up to the CPU's maximum 2Ghz (which it
never did before, it only ever reached 1.6Ghz).

I'm now inclined to wonder if the F.08 BIOS might have worked ok, and
provided the appropriate SSDT for CPU speed control, had the psmouse
module been removed before reboot.

Unless anyone is keen to know the results, I'm not that motivated to
find out (by upgrading the bios again), as I'm not keen to either wear
out the bios flash chip, or have a random corruption during an update.

Thanks for all the help so far,

Regards

Peter Clifton.

Re: HP Compaq nc6320

[Rafael J. Wysocki]

On Thursday, 12 October 2006 02:08, Peter Clifton wrote:
> 
> > Wierdly, I've lost the Battery/AC adapter status monitoring again, with
> > or without ec_intr=0.
> 
> > Gets odder!
> 
> Ok,
> 
> Nearing completion of this tale I think..
> 
> Turns out there is a known problem with psmouse, which leaves the
> system / ACPI / something in a bad state if psmouse is loaded before
> shutdown or reboot. Googling tells me that either removing batteries, AC
> etc.. before powering up, booting Windows, or shutting down Linux with
> psmouse removed, will "fix" the symptoms.

Do you remember any URLs?

Rafael


-- 
You never change things by fighting the existing reality.
		R. Buckminster Fuller

Re: HP Compaq nc6320

[Thomas Renninger]

On Thu, 2006-10-12 at 16:53 +0200, Rafael J. Wysocki wrote:
> On Thursday, 12 October 2006 02:08, Peter Clifton wrote:
> > 
> > > Wierdly, I've lost the Battery/AC adapter status monitoring again, with
> > > or without ec_intr=0.
> > 
> > > Gets odder!
> > 
> > Ok,
> > 
> > Nearing completion of this tale I think..
> > 
> > Turns out there is a known problem with psmouse, which leaves the
> > system / ACPI / something in a bad state if psmouse is loaded before
> > shutdown or reboot. Googling tells me that either removing batteries, AC
> > etc.. before powering up, booting Windows, or shutting down Linux with
> > psmouse removed, will "fix" the symptoms.

I got several reports that compiling psmouse as module and unload it
explicitly on shutdown helps to fix things on newer HPs.

E.g. this report:
About the fan and battery issue, there are three workaround:
 - Turn off the machine and quit battery, Place the battery again.
 - Next time you turn on the machine, turn it off during post.
 - "modprobe -r psmouse" (after compiling psmouse as a module) before
shutdown the machine.
Next time you turn on the machine, battery status is working. Not sure
about the fans...


http://www.wolframschenck.de/nx6325.htm#sect3
Similar introduction in German:
http://home.no/slazz/nx6310/nx6310.html


   Thomas

Re: HP Compaq nc6320

[Peter Clifton]

[snip]

> I got several reports that compiling psmouse as module and unload it
> explicitly on shutdown helps to fix things on newer HPs.
> 
> E.g. this report:
> About the fan and battery issue, there are three workaround:
>  - Turn off the machine and quit battery, Place the battery again.
>  - Next time you turn on the machine, turn it off during post.
>  - "modprobe -r psmouse" (after compiling psmouse as a module) before
> shutdown the machine.
> Next time you turn on the machine, battery status is working. Not sure
> about the fans...

My fans seemed to work ok most of the time before, it was the battery
status which hung, and the speedstep would never go to full speed.

Everything is happy with Ubuntu's kernel 2.6.17-10 and when unloading
psmouse as part of the shutdown and reboot scripts.

I did dig though psmouse a little, but being a non-expert, couldn't see
anything odd.

It could be the synaptics driver part I guess, as most people have
laptops.

It could be that psmouse keeps a serial driver of some sort alive, or a
bus active. Perhaps it is still receiving and processing interrupts as
the bios tries to reboot? Does someone need to open a kernel bug for
this, and where should it go?

Regards

Peter Clifton.

Re: HP Compaq nc6320

[Rafael J. Wysocki]

On Thursday, 12 October 2006 18:39, Peter Clifton wrote:
> [snip]
> 
> > I got several reports that compiling psmouse as module and unload it
> > explicitly on shutdown helps to fix things on newer HPs.
> > 
> > E.g. this report:
> > About the fan and battery issue, there are three workaround:
> >  - Turn off the machine and quit battery, Place the battery again.
> >  - Next time you turn on the machine, turn it off during post.
> >  - "modprobe -r psmouse" (after compiling psmouse as a module) before
> > shutdown the machine.
> > Next time you turn on the machine, battery status is working. Not sure
> > about the fans...
> 
> My fans seemed to work ok most of the time before, it was the battery
> status which hung, and the speedstep would never go to full speed.
> 
> Everything is happy with Ubuntu's kernel 2.6.17-10 and when unloading
> psmouse as part of the shutdown and reboot scripts.
> 
> I did dig though psmouse a little, but being a non-expert, couldn't see
> anything odd.
> 
> It could be the synaptics driver part I guess, as most people have
> laptops.

Well, on the HPC nx6325 I'm using psmouse is apparently interfering with
ACPI during resume from disk in a destructive way.  Namely, with the
2.6.19-rc1-mm1 kernel the temperature in one of the thermal zones
is way off during the resume, so ACPI thinks the critical thermal
point has been passed and it tries to power off the system, which makes
it hang.  Unloading psmouse before the suspend apparently helps.

I think it's synaptics too, because it spits reset failure messages during the
resume, if not unloaded before.

> It could be that psmouse keeps a serial driver of some sort alive, or a
> bus active. Perhaps it is still receiving and processing interrupts as
> the bios tries to reboot? Does someone need to open a kernel bug for
> this, and where should it go?

Well I guess so, and it looks like ACPI is the right place to start from.

Greetings,
Rafael


-- 
You never change things by fighting the existing reality.
		R. Buckminster Fuller