Re: PROBLEM with 2.6.18: BUG: scheduling while atomic

From: Ingo Molnar
Date: Wed Sep 27 2006 - 16:15:21 EST



* Andrew Morton <akpm@xxxxxxxx> wrote:

> > Sep 24 13:29:24 gondolin kernel: [<ffffffff8026ef30>] default_idle+0x0/0x60
> > Sep 24 13:29:24 gondolin kernel: [<ffffffff8024f276>] cpu_idle+0x96/0xb0
> > Sep 24 13:29:24 gondolin kernel: [<ffffffff805e5641>] start_secondary+0x4f1/0x500
>
> This might indicate that some code somewhere forgot to do
> spin_unlock/preempt_enable/kunmap_atomic/whatever.
>
> Ingo, do you have a current version of the patch which allows us to
> locate the culprit?

first the nvidia thing needs to be removed from that kernel. (wouldnt be
the first time it messes up the preempt count)

i have the PREEMPT_TRACE feature in the latency-tracer patch-queue. I've
merged it to 2.6.18-mm1 and have uploaded a combo patch to:

http://redhat.com/~mingo/latency-tracing-patches/latency-tracing-2.6.18-mm1.patch

CONFIG_DEBUG_PREEMPT should be enabled, that will activate
CONFIG_PREEMPT_TRACE. No need to enable latency tracing itself.

Ingo
-
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/