Re: [ACPI-sppt] RAMB values

Hello,

Ducrot Bruno wrote:
> Something like that:
> 
> #include <stdio.h>
> #include <sys/io.h>

news on RAMB :-)

When testing with this tool, reading RAMB evaluates to 0x0f750064 (on
my system, maybe some others could test their values), while the value
of RAMB in the ACPI driver is 0x64646464. So I tried hardcoding the
value of RAMB in the DSDT to 0x0f750064, and now ACPI hangs on boot
like after running Windows.

As an immediate consequence, you can avoid the hang-on-reboot by
hardcoding RAMB to 0x64646464 in the DSDT (haven't tested this).

There seem to be two problems currently:

* Windows does some kind of initialization before starting the DSDT
that causes RAMB to give the correct value (0x0f750064), which Linux
doesn't do. Either the hardware/BIOS are broken in requiring this
initialization (and we need to work around it) or Linux is not doing
some initialization it should be doing.

* Given the correct value for RAMB, the DSDT hangs. Maybe there's
other wrongly initialized values or the DSDT is buggy.

For testing, it might be interesting to set RAMB to the new value
after successfully booting with the old value. However, the
OperationRegion that depends on RAMB is initialized statically, hence
this won't work, at least not easily.

Cheers
Robert


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf