Re: [patch 0/5] x86: mce: Bugfixes, cleanups and a new CMCI poll version

[Chen Gong <gong.chen@linux.intel.com>]

于 2012/6/7 5:53, Thomas Gleixner 写道:
> Sorry for the resend, I guess I need more sleep or vacation or both :(
>
> The following series fixes a few interesting bugs (found by review in
> context of the CMCI poll effort) and a cleanup to the timer/hotplug
> code followed by a consolidated version of the CMCI poll
> implementation. This series is based on
>
>    git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git
>
> which contains the bugfix for the dropped timer interval init.
>
> Thanks,
>
> 	tglx
>
>
>
>

I tested the latest patch series based on your tip tree. The basic logic
is correct as we expected :-).

But during the CPU online/offline test I found an issue. After *STORM*
mode is entered, it can't come back from *STORM* mode to normal
interrupt mode. At least there exists such an issue: when *STORM* is
entered, in the meanwhile, one CPU is offline during this period,
which means *cmci_storm_on_cpus* can't decrease to 0 because there
is one bit stuck on this offlined CPU. So we should detect such
situation and decrease on *cmci_storm_on_cpus* at proper time.
BTW, even I online the *CPU* in above situation, the normal CMCI
still doesn't come back, strange.

I still have another question: When we handle following case:
mce_cpu_callback(struct notifier_block *
		mce_device_remove(cpu);
		break;
	case CPU_DOWN_PREPARE:
-		del_timer_sync(t);
		smp_call_function_single(cpu, mce_disable_cpu, &action, 1);
+ 		del_timer_sync(t);
		break;

Where we add this timer back? I can't find it in "case CPU_ONLINE".