Win7 on Intel Eagle Heights

[Thomas JOURDAN]

Hi guys

I'm trying to get Windows 7 booting on my Intel Eagle Heights evaluation
board. I tried to follow all the ACPI tips to get Window$ to boot but I
can't figure out the bug I'm facing. I'm using a checked build version of
Win 7 64-bits. When I start the installer, the first text screen is ok
(windows is copying files...) then it switches to graphics mode (green
progress bar with logo). Less than a second after switching from text to
graphics mode : BSOD.

With windbg I tried to dump as much as possible and the analyser tell me a
write operation occured at an invalid address. According to the stack frame,
the error occured in PowerPolicyProcessEvent when the kernel starts all
devices (enter D0 state).

Does the device states only applies to PCI devices, or all devices ? How can
I know which device causes this problem ? I can't even figure out if it's an
ACPI related problem, or something wrong in a device initialization
function.

Regards,
Thomas

[Stefan Reinauer]

On 5/13/11 8:57 AM, Thomas JOURDAN wrote:
> I'm trying to get Windows 7 booting on my Intel Eagle Heights 
> evaluation board. I tried to follow all the ACPI tips to get Window$ 
> to boot but I can't figure out the bug I'm facing. I'm using a checked 
> build version of Win 7 64-bits. When I start the installer, the first 
> text screen is ok (windows is copying files...) then it switches to 
> graphics mode (green progress bar with logo). Less than a second after 
> switching from text to graphics mode : BSOD.

Hi Thomas,

Looks like Windows 7 crashes in the "Windows Driver Foundation" with an 
illegal memory access. Possibly this is due to an incomplete ACPI 
implementation for the board.

Please have a look at the kontron/986lcd-m ACPI code for a modular 
sample implementation that can boot Windows 7 and 
http://www.coreboot.org/ACPI for more information on ACPI and ACPI 
debugging.

Stefan

[Thomas Jourdan]

Hi

Le vendredi 13 mai 2011 à 18:21 -0700, Stefan Reinauer a écrit :
> On 5/13/11 8:57 AM, Thomas JOURDAN wrote:
> > I'm trying to get Windows 7 booting on my Intel Eagle Heights 
> > evaluation board. I tried to follow all the ACPI tips to get Window$ 
> > to boot but I can't figure out the bug I'm facing. I'm using a checked 
> > build version of Win 7 64-bits. When I start the installer, the first 
> > text screen is ok (windows is copying files...) then it switches to 
> > graphics mode (green progress bar with logo). Less than a second after 
> > switching from text to graphics mode : BSOD.
>
> Hi Thomas,
>
> Looks like Windows 7 crashes in the "Windows Driver Foundation" with an 
> illegal memory access. Possibly this is due to an incomplete ACPI 
> implementation for the board.

Yes that what I suspect too but so far I can't figure out which ACPI
part could be missing. According to my probe, after the page
translation, the guilty address seems to point to valid memory area
(DDR).
> Please have a look at the kontron/986lcd-m ACPI code for a modular 
> sample implementation that can boot Windows 7 and 
> http://www.coreboot.org/ACPI for more information on ACPI and ACPI 
> debugging.

When I said I followed ACPI tips I mean I refered to kontron/986lcd-m,
the coreboot acpi page, and also some tricks disccussed on the mailing
list (64-bits alignment, fadt correct size according to version...).

I was hoping someone already met this issue and could give me a clue.

Regards, 
Thomas

[Marc Jones]

On Mon, May 16, 2011 at 1:54 AM, Thomas Jourdan
 wrote:
> Hi
>
> Le vendredi 13 mai 2011 à 18:21 -0700, Stefan Reinauer a écrit :
> > On 5/13/11 8:57 AM, Thomas JOURDAN wrote:
> > > I'm trying to get Windows 7 booting on my Intel Eagle Heights
> > > evaluation board. I tried to follow all the ACPI tips to get Window$
> > > to boot but I can't figure out the bug I'm facing. I'm using a checked
> > > build version of Win 7 64-bits. When I start the installer, the first
> > > text screen is ok (windows is copying files...) then it switches to
> > > graphics mode (green progress bar with logo). Less than a second after
> > > switching from text to graphics mode : BSOD.
> >
> > Hi Thomas,
> >
> > Looks like Windows 7 crashes in the "Windows Driver Foundation" with an
> > illegal memory access. Possibly this is due to an incomplete ACPI
> > implementation for the board.
>
> Yes that what I suspect too but so far I can't figure out which ACPI
> part could be missing. According to my probe, after the page
> translation, the guilty address seems to point to valid memory area
> (DDR).
> > Please have a look at the kontron/986lcd-m ACPI code for a modular
> > sample implementation that can boot Windows 7 and
> > http://www.coreboot.org/ACPI for more information on ACPI and ACPI
> > debugging.
>
> When I said I followed ACPI tips I mean I refered to kontron/986lcd-m,
> the coreboot acpi page, and also some tricks disccussed on the mailing
> list (64-bits alignment, fadt correct size according to version...).
>
> I was hoping someone already met this issue and could give me a clue.
>
> Regards,
> Thomas

Hi Thomas,

ScottD had a FADT fix for checked build on Persimmon that might be
related to your issue.

http://www.coreboot.org/pipermail/coreboot/2011-May/065115.html

Marc