Messages in this thread Patch in this message | | | Date | Mon, 15 Dec 2008 23:33:10 +0100 | From | Andi Kleen <> | Subject | Re: [PATCH 2/2] x86: re-enable MCE on secondary CPUS after suspend/resume |
| |
> void __cpuinit mcheck_init(struct cpuinfo_x86 *c) > { > static cpumask_t mce_cpus = CPU_MASK_NONE; > > mce_cpu_quirks(c); > > if (mce_dont_init || > cpu_test_and_set(smp_processor_id(), mce_cpus) || > !mce_available(c)) > => return; > > mce_init(NULL); > mce_cpu_features(c); > } > > But we need to call mce_init to clear all MCE state. > IMHO the best location to call mce_init for APs is the cpu notifier.
Ah got it. Thanks that makes sense.
But I think the better fix is to just drop the mce_cpus check and then handly it naturally in the standard bootup path. I'm not sure what it was good for anyways. I copied it into the 64bit code from 32bit, but I suppose even there it isn't really needed and on 32bit it is already gone even.
How about this patch. Does it fix the problem for you too?
-Andi
--
Don't prevent multiple initialization of MCEs.
Back from early prehistory mcheck_init() has a reentry check. Presumably that was needed in very old kernels to prevent it entering twice.
But as Andreas points out this prevents CPU hotplug (and therefore resume) to correctly reinitialize MCEs when a AP boots again after being offlined.
Just drop the check.
Based on a report from Andreas Herrmann
Signed-off-by: Andi Kleen <ak@linux.intel.com>
--- arch/x86/kernel/cpu/mcheck/mce_64.c | 3 --- 1 file changed, 3 deletions(-)
Index: linux/arch/x86/kernel/cpu/mcheck/mce_64.c =================================================================== --- linux.orig/arch/x86/kernel/cpu/mcheck/mce_64.c 2008-12-15 23:13:02.000000000 +0100 +++ linux/arch/x86/kernel/cpu/mcheck/mce_64.c 2008-12-15 23:13:44.000000000 +0100 @@ -510,12 +510,9 @@ */ void __cpuinit mcheck_init(struct cpuinfo_x86 *c) { - static cpumask_t mce_cpus = CPU_MASK_NONE; - mce_cpu_quirks(c); if (mce_dont_init || - cpu_test_and_set(smp_processor_id(), mce_cpus) || !mce_available(c)) return;
| |