[tlinux-users 00216] Re: Toshiba Satellite A300, ACPI issues and a problematic ATI radeon 3650 HD card

[Ilyes Gouta]

Hi,

Well I gave a spin to the omnibook module and I got the bluetooth
enabled by forcing the ec_type param. to an arbitrary value. The said
module did advertize that the fan, lcd and a couple of other features
are detected but actually they didn't work since poking into the entry
file in /proc didn't change anything... Anyway, nice stuff but it
definitely needs a lifting IMHO. I think that Phoenix/Toshiba keep
somehow chaning the ACPI method names (this hold true at least for
controlling the brightness of the LCD panel in my case) so that one
needs a dedicated ACPI driver for the extra features. My Toshiba model
(Phoenix BIOS) DSDT table indcates that the TVAP is present plus an
ACPI WMI interface and I thought that these are only available on the
"pure" BIOSes made by Toshiba.

Now to my litte issue with my Satellite A300 PSAJ4E model, which comes
w/ a Core 2 Duo T8100 and an ATI Radeon 3650 HD card. Apperently the
gfx card has a fan attached to it and I read on the Toshiba forums
that they take the reference display driver as obtained from ATI and
customize it for their products... I bet that all that they're doing
is including the necessay piece of code to control that fan and to
throttle the chip depending on the load and the power management
policy. So not having that custom code running in Linux would mean
that the gfx card can potentially overheat which can lead to a failure
(think Nvidia and their broken chips) and render the whole laptop
useless, right? Do you guys think that this is a real issue? Any idea
on how to manage a GPU fan?

Regards,
Ilyes Gouta.

On Fri, Oct 10, 2008 at 4:34 PM, Jonathan Buzzard
<jonathan at buzzard.me.uk> wrote:
>
> On Fri, 2008-10-10 at 08:22 -0400, Thomas Charron wrote:
>> On Fri, Oct 10, 2008 at 4:36 AM, Ilyes Gouta <ilyes.gouta at gmail.com> wrote:
>> > Hi,
>> > You'll need to call some ACPI methods in order to enable/disable the
>> > LEDs. The new models from Toshiba (Statellite A/U 300 and up) are 100%
>> > designed for Windows. Reading the DSDT ACPI table was a shocker for me
>> > since it contained a lot of references to some Windows only
>> > capabilities such as a hidden ACPI device (TVAP, Toshiba Value Added
>> > Package), an ACPI - WMI interface which offer a way for windows to
>> > access BIOS functions through a WMI interface, an incomplete thermal
>> > zone specification in which there is no mention of a fan device,
>> > although the CPU (and/or GPU) fan keep spinning at full speed, all the
>> > time, and loudly as soon as I boot my laptop into Linux (Could you
>> > please check if you have the same symptoms as mine, regarding the last
>> > point?).
>> > Now, there is one thing I like you to do, could you please send me
>> > back your binary DSDT table? To do it, log as root on your machine,
>> > open a shell and do:
>> > $ cat  /proc/acpi/dsdt > acpi-dump.bin
>> > and then send me back acpi-dump.bin.
>
>
> The TVAP and the TVALD (Toshiba Value Added Logical Device) are
> effectively supported under Linux, for a *LONG* time now.
>
> I am not sure if the TVAP replaces the TVALD, or is only present on the
> Compal OEM laptops (as opposed to the puka 100% Toshiba ones). However
> it is just an interface to the SMM mode of the processor. The puka
> Toshiba laptops are supported under with /dev/toshiba for over 10 years
> now and the Compal ones with the Omnibook driver since around 2000 as I
> recall. Guess who OEM's laptops for HP :-)
>
> The TVALD device works in *exactly* the same manner as /dev/toshiba you
> open the device and do an ioctl... I have not looked into the TVAP
> closely, but it looked to be doing exactly the same thing.
>
> I guess the issue is that the toshiba_acpi driver does not present
> a /dev/toshiba device unless you specially patch it. That should really
> be fixed. There is some issues about who gets ownership of that device
> is it the toshiba module or the toshiba_acpi?
>
> It is my opinion that as APM and ACPI are mutually exclusive, it does
> not matter, as you cannot load both the toshiba and toshiba_acpi modules
> on any machine at the same time and if you can it is a bug that needs
> fixing.
>
> However I did not write the toshiba_acpi module, and for a *long* time
> did not run my laptop in ACPI mode. That said I replaced my ageing Tecra
> 8200 with a Tecra M5 late last year (the 8200 gave up the ghost after
> five and a bit years), and have been running with ACPI on the M5 ever
> since as it does not support APM. However the patched toshiba_acpi
> module does me fine.
>
>
> JAB.
>
> --
> Jonathan A. Buzzard                 Email: jonathan (at) buzzard.me.uk
> Fife, United Kingdom.
>
> _______________________________________________
> tlinux-users mailing list&lt;tlinux-users at ml.toshiba-dme.co.jp>
> http://linux.toshiba-dme.co.jp/mailman/listinfo/tlinux-users
>