RE: ACPI failure (2.6.0-test<x> and 2.4.22-pre<x>)

RE: ACPI failure (2.6.0-test<x> and 2.4.22-pre<x>)

[Grover, Andrew]

> From: Richard A Nelson [mailto:cowboy@vnet.ibm.com] 
> On either 2.6, or 2.4, booting with ACPI enabled gets as far as
> parsing the ACPI EC table - at which point it oops with bad pointer
> and halts the system.  Sorry at this point I don't have the register
> contents; it took a while to narrow it this far - and have the
> screen in such a state that I can see any relevant information other
> than the the trying to kill init message :)

Could you try, say, 2.5.70? That would help me understand if it's a
regression or not.

If that also fails, then I think determining exactly where it is oopsing
(probably via printks) would be the next best way to go.

Regards -- Andy

RE: ACPI failure (2.6.0-test<x> and 2.4.22-pre<x>)

[Richard A Nelson]

On Mon, 28 Jul 2003, Grover, Andrew wrote:

> Could you try, say, 2.5.70? That would help me understand if it's a
> regression or not.

looks to be the same problem, I got a little more info - and compared
it to 2.6.0-test2-mm1; forgive the formatting, this was copied by hand
(its a pain being with only one box):

anything before this was cut of by the size of the screen

EIP: 0060: c01efc11   (2.6.0-test2-mm1 version)
EFLAGS: 00010202
EIP at: acpi_ut_remove_allocation+0xb5/0x126
EAX: 00000003  EBX: c136c59c  ECX: 00000000  EDX: 00000001
ESI: 00000000  EDI: 00000000  EBP: cff118c4  ESP: cff118c8
ds:  007b      es:  007b      ss:  0068
...
... acpi_ps_parse_aml+0xa2/0x250
... acpi_power_add+0x126/0x1c4
... acpi_bus_scan
...

> If that also fails, then I think determining exactly where it is oopsing
> (probably via printks) would be the next best way to go.

>From other messages, here I found and extracted the acpi_dsdt.aml if
that'll help with anything (can't make much sense of it myself, but
was surprised to see special casing for win98/nt/me in the init section)

-- 
Rick Nelson
I can saw a woman in two, but you won't want to look in the box when I do
'For My Next Trick I'll Need a Volunteer' -- Warren Zevon

Re: ACPI failure (2.6.0-test<x> and 2.4.22-pre<x>)

[Mikael Pettersson]

On Mon, 28 Jul 2003 14:09:40 -0400 (EDT), Richard A Nelson wrote:
>IBM T30 Laptop:
>IBM machine detected. Enabling interrupts during APM calls.
>IBM machine detected. Disabling SMBus accesses.  <- 2.6 only
>No local APIC present or hardware disabled       <- 2.4 only
>Local APIC disabled by BIOS -- reenabling.       <- 2.6 only
>Found and enabled local APIC!                    <- 2.6 only

This is the 2.5.74 patch for P4 kicking in. Without it
you wouldn't have access to the local APIC.

>On either 2.6, or 2.4, booting with ACPI enabled gets as far as
>parsing the ACPI EC table - at which point it oops with bad pointer
>and halts the system.  Sorry at this point I don't have the register
>contents; it took a while to narrow it this far - and have the
>screen in such a state that I can see any relevant information other
>than the the trying to kill init message :)

Since 2.4 also oopses it can't be the local APIC. I'm saying
this because ACPI + local APIC doesn't work with some BIOSen.

ACPI failure (2.6.0-test<x> and 2.4.22-pre<x>)

[Richard A Nelson]

IBM T30 Laptop:
IBM machine detected. Enabling interrupts during APM calls.
IBM machine detected. Disabling SMBus accesses.  <- 2.6 only
No local APIC present or hardware disabled       <- 2.4 only
Local APIC disabled by BIOS -- reenabling.       <- 2.6 only
Found and enabled local APIC!                    <- 2.6 only
ACPI:             have wakeup address 0xc0001000

model name      : Intel(R) Pentium(R) 4 Mobile CPU 1.80GHz
Bios Version:     1IET66WW (2.05 )               <- most recent available
BIOS32 Service Directory present.
ACPI 2.0 present.
        OEM ID: IBM
RSD table at 0x0FF63195.
PNP 1.0 present.
        Event Notification: Polling
        Event Notification Flag Address: 0x000004B4
        Real Mode Code Address: F000:9D36
        Real Mode Data Address: 0040:0000
        Protected Mode Code Address: 0x000F9D54
        Protected Mode Data Address: 0x00000400
PCI Interrupt Routing 1.0 present.
        Table Size: 256 bytes
        Router ID: 00:1f.0
        Exclusive IRQs: None
        Compatible Router: 8086:122e

On either 2.6, or 2.4, booting with ACPI enabled gets as far as
parsing the ACPI EC table - at which point it oops with bad pointer
and halts the system.  Sorry at this point I don't have the register
contents; it took a while to narrow it this far - and have the
screen in such a state that I can see any relevant information other
than the the trying to kill init message :)

-- 
Rick Nelson
I can saw a woman in two, but you won't want to look in the box when I do
'For My Next Trick I'll Need a Volunteer' -- Warren Zevon