sony-vaio-z-series team mailing list archive

[Matthias Welwarsky <matze@xxxxxxxxxxxx>]

On Sunday 15 February 2009 23:21:13 Eva Brucherseifer wrote:
> Hi Julius,
>
> in the meantime I also started Nvidia tests and indeed it doesn't work. In
> speed mode the card is enabled, the speed LED is correctly on, but
> according to the X log the output doesn't seem to be working.
> AFAIK there is a switch on the mainboard, routing either the output of the
> Intel card or the Nvidida card to VGA output. This switch seems to be the
> problem and we don't have an idea about the right now.

We think that there must be some kind of a switch matrix on the board that 
routes the output signal of the two graphics cards to any of the available 
output devices. Unfortunately I wasn't yet able to identify the necessary ACPI 
methods and arguments to take influence on the switch matrix. So the only thing 
I can do today is power down the nvidia card to not confuse X11 and save 
battery power.

One possible approach could be to start "Speed" mode using the XP trick and 
then dump the content of the various ACPI variables, then do the same for 
"Stamina" and afterwards compare the results. This could lead to information 
which of the ACPI variables are involved in Speed vs Stamina, which in turn 
helps with the reverse engineering of the DSDT. I _think_ it should be 
possible to evaluate the content of any ACPI object if only the path is known. 
The key is again the acpi_evaluate_object() call.

Unfortunately I'm travelling to Paris tomorrow evening, I won't be able to 
work on this project until Thursday. If you want to help and have some 
knowledge of ACPI and kernel programming (or know someone who has), you could 
try what I suggested above in the meantime. At this point, any help is more 
than welcome.

regards
	matze