Board hangs without acpi=off

Board hangs without acpi=off

[Olivier Diotte]

Hi,

I have a development board here that hangs when trying to boot linux
(actually Ubuntu Precise Pangolin's kernel 3.5.7.6).
Supplying the kernel parameter 'acpi=off' allows the board to boot
correctly. The problem is that we need this board to boot with a
working ACPI stack.

I recompiled Ubuntu's kernel adding support for ACPI debugging
(CONFIG_PM_DEBUG=y, CONFIG_ACPI_DEBUG=y) and did some tests with
'acpi.debug_layer' and 'acpi.debug_level' and can provide these if
needed, though I am unsure if this is the right approach.

Is there something I can do to get this board's ACPI working? Any hint
and/or approach would be a good step forward.

Regards,
Olivier

Re: Board hangs without acpi=off

[Rafael J. Wysocki]

On Tuesday, March 26, 2013 11:09:55 AM Olivier Diotte wrote:
> Hi,
> 
> I have a development board here that hangs when trying to boot linux
> (actually Ubuntu Precise Pangolin's kernel 3.5.7.6).
> Supplying the kernel parameter 'acpi=off' allows the board to boot
> correctly. The problem is that we need this board to boot with a
> working ACPI stack.
> 
> I recompiled Ubuntu's kernel adding support for ACPI debugging
> (CONFIG_PM_DEBUG=y, CONFIG_ACPI_DEBUG=y) and did some tests with
> 'acpi.debug_layer' and 'acpi.debug_level' and can provide these if
> needed, though I am unsure if this is the right approach.
> 
> Is there something I can do to get this board's ACPI working? Any hint
> and/or approach would be a good step forward.

Can you get any debug output from boot attempts without acpi=off?

Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: Board hangs without acpi=off

[Olivier Diotte]

On Tue, Mar 26, 2013 at 6:47 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> On Tuesday, March 26, 2013 11:09:55 AM Olivier Diotte wrote:
> Can you get any debug output from boot attempts without acpi=off?
>
Yes, I did many tests with different values for both parameters. I
took pictures (as logging capacity is somewhat limited though maybe I
could try to test using a serial console) of the last screen of output
for some tests and a few notes (the latter are at work, I'll be able
to send them tomorrow if needed).

By now, the board should be done recompiling Linus' tree, which will
allow me to test with a more recent kernel.

I can post the pictures, but I am unsure what the etiquette is about
that on kernel mailing-lists.

I'll try to transcribe the pictures, the notes and try Linus' kernel
tomorrow when I have access to them all.

Please tell me what would be useful info and I'll happily oblige.

Regards,
Olivier