Bug#498141: Additional info - ACPI corruption?
I've now done a full Lenny install on the machine in question. So far 5 of
6 boots things have been fine, but one of the times I had this issue. I
did a diff of the kernel.log files between a good and a bad boot and found
this difference:
*** 422,428 ****
ACPI: ACPI0007:00 is registered as cooling_device1
ACPI: ACPI0007:01 is registered as cooling_device2
ACPI: LNXTHERM:01 is registered as thermal_zone0
! ACPI: Thermal Zone [THRM] (43 C)
ohci_hcd: 2006 August 04 USB 1.1 'Open' Host Controller (OHCI) Driver
ACPI: PCI Interrupt Link [APCF] enabled at IRQ 23
ACPI: PCI Interrupt 0000:00:02.0[A] -> Link [APCF] -> GSI 23 (level,
high) -> IRQ 23
--- 422,428 ----
ACPI: ACPI0007:00 is registered as cooling_device1
ACPI: ACPI0007:01 is registered as cooling_device2
ACPI: LNXTHERM:01 is registered as thermal_zone0
! ACPI: Thermal Zone [THRM] (38 C)
ohci_hcd: 2006 August 04 USB 1.1 'Open' Host Controller (OHCI) Driver
ACPI: PCI Interrupt Link [APCF] enabled at IRQ 23
ACPI: PCI Interrupt 0000:00:02.0[A] -> Link [APCF] -> GSI 23 (level,
high) -> IRQ 23
***************
In fact, looking across all of the logs from each boot since I installed
Lenny (yesterday), here is that line:
Jan 25 11:53:00 gerbil kernel: [ 1.059484] ACPI: Thermal Zone [THRM] (46
C)
Jan 25 12:21:32 gerbil kernel: [ 1.065176] ACPI: Thermal Zone [THRM] (45
C)
Jan 25 12:35:06 gerbil kernel: [ 1.059631] ACPI: Thermal Zone [THRM] (44
C)
Jan 25 13:08:00 gerbil kernel: [ 1.062476] ACPI: Thermal Zone [THRM] (45
C)
Jan 25 14:02:27 gerbil kernel: [ 1.060939] ACPI: Thermal Zone [THRM] (45 C)
Jan 26 05:36:39 gerbil kernel: [ 1.064245] ACPI: Thermal Zone [THRM] (43
C)
Jan 26 05:52:07 gerbil kernel: [ 1.077704] ACPI: Thermal Zone [THRM] (38
C)
(The 43 C boot was the session that had corruption, but there isn't
anything else too funny in the logs - just slight variations in CPU speed
and things like that.)
If I run acpi from the command line I get this:
~> acpi -t
Thermal 0: ok, 43.0 degrees C
~> acpi -t
Thermal 0: ok, 44.0 degrees C
I find the ACPI probe for the thermal zone being inconsistent to be
interesting as when this problem has surfaced in the past it has looked
correlated to acpi trying to turn a fan on. My guess is that there is an
ACPI bug lurking here. The machine is a Shuttle SN95G5 v2 with an NVidia
Nforce3 chipset in it. I'll go looking to see if there are BIOS updates
available although I'm guessing there won't be as I have upgraded it
previously and it's a 4 year old machine. (Also recall that the 2.4.22
kernel from backports was running fine on this hardware for several years,
so I'd suspect something in the kernel has changed since then.)
Anyways, I thought I would pass this additional data along.
Thanks,
Dale
--
Dale E. Martin - dale@the-martins.org
http://the-martins.org/~dmartin
Reply to: