Re: 2.6.25-current-git hangs on boot unless CONFIG_CPU_IDLE=n-Apple

On Sun 6.Apr'08 at 22:31:17 -0700, Pallipadi, Venkatesh wrote:
On Fri 4.Apr'08 at 13:41:45 -0700, Pallipadi, Venkatesh wrote:
[...]
I've uploaded the picture where it hangs, it is in the very
beginning
of the boot process:

http://www.ift.unesp.br/users/crmafra/dsc04673.jpg


Can you try the latest git and see whether this still is
a problem?

I've just updated to latest git, but it may take some time for it
to happen again (the first 3 boots were okay). Last time
it happened
with 2.6.25-rc7-00149-gaf8be4e (I checked now).

Do you have reasons to believe this issue was fixed since then?

Yes. I think the patch here should have fixed the problem....
Patch : http://marc.info/?l=linux-kernel&m=120674502201007&w=4


Hmmmm.... Looking at latest git,
The above patch is not yet in git. Can you please apply the
above patch
and see whether do you still see the problem.....

I applied Venkatesh's patch from the link above on top of
v2.6.25-rc8-139-ge315c12, but my laptop hung at the 11th boot with the
same state show here:
http://www.ift.unesp.br/users/crmafra/dsc04673.jpg

I waited 3'30" but the boot process never resumed, so I pressed
the power button.

Any further patches I might try?

Carlos,

Hmm. So, we still have some bug hidden in there..
As this is similar to 10117, we should probably track this there.
10093 is a different cpuidle specific bug looks like, as it happens at a
different place while switching between governors.

Can you attach your kernel config, acpidump output and dmesg when kernel
boots ok to that bugzilla.

I uploaded the kernel config and dmesg to the bugzilla a few minutes ago.
I don't have a copy of the acpidump with me right now (laptop is at home,
I am at the physics institute now), but if there is urgency you can look
at the acpidump I've sent here http://lkml.org/lkml/2008/2/11/364

Also, if you can please try processor.max_cstate=2 boot option and see
whether you still can reproduce the problem. If you can, then try
processor.max_cstate=1. For this boot option to work, you have to have
ACPI_PROCESSOR built into your kernel (and should not be a module).

Ok, I will try them.

One more qn. Did you see this problem witj 2.6.24 as well or is this
something that you started seeing only recently?

It started more or less at the same time as 2.6.25-rc1 was released.
It was not there at 2.6.24.

Well, at first I thought it was related to drivers/acpi/thermal.c, because
the next kernel message (in a successfull boot) after the point in which it
fails is

ACPI: LNXTHERM:01 is registered as thermal_zone0

and there were some changes to that file a bit before -rc1 was out.
For example, commits 3f655ef8 and ce44e19701.

I don't know if this is recommended, but I added Zhang Rui to the Cc: list
to exclude or confirm this possibility.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/