http://bugme.osdl.org/show_bug.cgi?id=464
Summary: 2.5.64: Dell Inspiron 8000 BIOS A04 EMERGENCY SHUTDOWN!
Kernel Version: 2.5.64
Status: NEW
Severity: blocking
Owner: [email protected]
Submitter: [email protected]
Distribution: Debian unstable
Hardware Environment: Dell Inspiron 8000 notebook, 512MB, ATI 32MB
Software Environment: 2.5.64, gcc 3.2.3
Problem Description:
I just got my new Dell Inspiron 8000 notebook (was defective, but repaired
quickly, by "cleaning" it :-).
I noticed two immediate emergency shutdowns on 2.5.64/ACPI, which I initially
suspected to be "notebook still slightly defective" issues.
But after several days of very reliable operation on 2.4.21/APM (it seems to do
passive management via CPU throttling), I'm quite
convinced that it's a problem of the 2.5.64 kernel ACPI management instead.
This notebook still has BIOS A04, despite a very advanced version of A21 (!)
being available (the previous owner obviously wasn't very eager to keep
it up to date).
In other words, I suspect that BIOS A04 has some "quirks" that cause 2.5.64
to horribly misinterprete ACPI thermal management.
I really don't think Dell would have released that machine if it had shown such
fatal power management behaviour on Windows, so I guess we do have an ACPI
management problem of some sort in Linux with slightly less kosher BIOS
implementations.
I intend to update my BIOS to the latest version VERY soon now (I don't feel
safe running such an old and broken BIOS version), so please try to handle this
bug report IMMEDIATELY.
Steps to reproduce:
I don't want to describe that here, since I really don't intend to put my
machine through yet another emergency shutdown...
You need to produce some high CPU load, though...
I was unsure of which Severity to assign to this bug (high or blocking),
but I believe we really shouldn't release 2.6 with any sorts of hardware damage
potential (at least as long as we can avoid it), so I made it blocking, for
a good reason, I think.
Hi!
> Steps to reproduce:
> I don't want to describe that here, since I really don't intend to put my
> machine through yet another emergency shutdown...
> You need to produce some high CPU load, though...
Heh, show us /proc/acpi/thermal/*.
Also don't fear emergency shutdowns -- they do NOT damage
hardware. [This machine survived >20 of them.]
Pavel
--
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]
Hi,
On Sun, Mar 16, 2003 at 09:36:29PM +0100, Pavel Machek wrote:
> Hi!
>
> > Steps to reproduce:
> > I don't want to describe that here, since I really don't intend to put my
> > machine through yet another emergency shutdown...
> > You need to produce some high CPU load, though...
>
> Heh, show us /proc/acpi/thermal/*.
Heh, gotcha!
It was in the attachments of bug #464 already... ;-)
(it's /proc/acpi/thermal_zone/*, BTW)
And don't expect any spectacular values just before shutdown there...
The thermal management is almost non-existent according to these values.
> Also don't fear emergency shutdowns -- they do NOT damage
> hardware. [This machine survived >20 of them.]
Yeah, and your 21st shutdown will kill it, for sure ;-)
Thanks for your reply!
Andreas Mohr
--
Help prevent Information Technology Fascism! - before it's too late...
http://www.againsttcpa.com