Building 2.6.0-test9 without ACPI

Building 2.6.0-test9 without ACPI

[Ian Wienand]

Hi,

My zx2000 keeps hitting a MCA seemingly the moment it gets to the
login: prompt, sometimes during boot with test9.

To try and debug I went to turn ACPI off in menuconfig, but it doesn't
seem to actually turn ACPI off properly.  The sub-menu doesn't
disappear as you would expect (and as happens when configuring an x86)
and trying to build ends up with

arch/ia64/kernel/iosapic.c: In function `iosapic_set_affinity':
arch/ia64/kernel/iosapic.c:290: warning: comparison is always false due to limited range of data type
arch/ia64/kernel/iosapic.c: In function `iosapic_register_platform_intr':
arch/ia64/kernel/iosapic.c:530: error: `ACPI_INTERRUPT_PMI' undeclared (first use in this function)
arch/ia64/kernel/iosapic.c:530: error: (Each undeclared identifier is reported only once
arch/ia64/kernel/iosapic.c:530: error: for each function it appears in.)
arch/ia64/kernel/iosapic.c:539: error: `ACPI_INTERRUPT_INIT' undeclared (first use in this function)
arch/ia64/kernel/iosapic.c:543: error: `ACPI_INTERRUPT_CPEI' undeclared (first use in this function)
make[1]: *** [arch/ia64/kernel/iosapic.o] Error 1

I think this is because all sorts of things are left in the config
that maybe shouldn't be there?

# CONFIG_ACPI is not set
CONFIG_ACPI_EFI=y
CONFIG_ACPI_INTERPRETER=y
CONFIG_ACPI_KERNEL_CONFIG=y

I had a look in drivers/acpi/Kconfig and things look mostly right to
me (ACPI_INTERPRETER depends on ACPI, everything else depends on
ACPI_INTERPRETER).  Is it just me?

-i
ianw@gelato.unsw.edu.au
http://www.gelato.unsw.edu.au

Re: Building 2.6.0-test9 without ACPI

[Grant Grundler]

On Wed, Nov 12, 2003 at 05:02:47PM +1100, Ian Wienand wrote:
> To try and debug I went to turn ACPI off in menuconfig, but it doesn't
> seem to actually turn ACPI off properly.

ACPI is required. It's how ia64 platform resources are discovered.
The fix would be to force ACPI on if any ia64 target is selected.

grant

Re: Building 2.6.0-test9 without ACPI

[Ian Wienand]

[-- Attachment #1: Type: text/plain, Size: 369 bytes --]

On Tue, Nov 11, 2003 at 10:06:38PM -0800, Grant Grundler wrote:
> ACPI is required. It's how ia64 platform resources are discovered.
> The fix would be to force ACPI on if any ia64 target is selected.

Doh! I did know that but got all excited when I saw you could turn it
off.  My fault as usual.  Maybe the attached could be slipped in
whenever it is appropriate.

-i

[-- Attachment #2: Kconfig-acpi.patch --]
[-- Type: text/plain, Size: 260 bytes --]

===== arch/ia64/Kconfig 1.48 vs edited =====
--- 1.48/arch/ia64/Kconfig	Wed Oct 15 18:35:33 2003
+++ edited/arch/ia64/Kconfig	Wed Nov 12 17:24:47 2003
@@ -10,6 +10,7 @@
 menu "Processor type and features"
 
 config IA64
+	select ACPI
 	bool
 	default y
 	help