[coreboot] jetwaynf94-270-lf: ATOM-n270+i945 triplefault if activate the variable mtrr.

[bari]

Hi Urbez,

Do you have docs for the Atom cpu or did you hack your way into it for 
init and CAR?

-Bari

Urbez Santana Roma wrote:
> Hi, i'm working with these mainboard fanless:
> Jetway NF94-270-LF 
> 
> 
> The bios chip is one Winbond W25X80,
> i have released a simple circuit with one transistor
> BC547, and the rest connected directly to the
> paralel port in mode ECP, data buses are 3.3V
> Also i have release a simple program to erase, test
> and program it. If one are interesed in this, tell-me.
> 
> The flashprogram not works correctly with this chip, but
> i can try to solve.
> 
> The CPU is the Atom N270, normally i used VIA's but, i will
> test my aplications with it, with VIA i cannot use my SSE2
> optimizations, they are more slow that direct use of 32bit
> assembler code.
> 
> 
> 
> This have a Fintek SuperIO F71862FG that is similar to
> the F71805F.
> 
> The base is the ich7m and i945GSE.
> 
> I have tested and make that it works, except the VGA
> and the Audio, i have solved the problems with acpi,
> and linux kernel boots normally with the SATA.
> 
> I have only one problem, with the variable MTRR
> if i activate variable MTRR's for have cache, the
> kernel immediately makes a triple fault.
> 
> I not known if in the moment of activate MTRR this
> chip generates a spurious interrupt or similar, and
> when the kernel creates the IDT table, hangs.
> 
> Have one show this problem in the past?
> I have send to local APIC's the EOI, but the problem
> continues...
> 
> 
> 
> 
>