Re: 2.6.10-rc2-mm2: spinlock problem in arch/i386/kernel/time.c on resume

[Jeremy Fitzhardinge <jeremy@goop.org>]

On Fri, 2004-11-19 at 02:53 -0800, Andrew Morton wrote:
> I can't imagine how that could happen.  I wonder if the spinlock debugging
> could be making a mistake.

When I disabled DEBUG_SPINLOCK, it works fine.  Does DEBUG_SPINLOCK have
any side effects?  I would have thought it wouldn't do anything on a UP,
non-PREEMPT system.

There is the comment in arch/i386/kernel/time.c saying that APM needs
get_cmos_time(), which is the site of the second message.  The code is
very simple, so its hard to see how this could be going wrong.  I'm
guessing, though, that this is a relatively untested path, since people
don't tend to use APM+SMP.

> > .config & lspci attached
> 
> I have but a lowly A21P.  With my .config it APM resumes happily.  With
> yours it refuses to even suspend.  And:
> 
> 
> 
> hdc: DMA disabled
> e100: Intel(R) PRO/100 Network Driver, 3.2.3-k2-NAPI
> e100: Copyright(c) 1999-2004 Intel Corporation

I've been using the eepro100 driver lately, which doesn't seem to have
this problem at all.

	J