aboutsummaryrefslogtreecommitdiffstats
path: root/arch
diff options
context:
space:
mode:
authorRafael J. Wysocki <rjw@sisk.pl>2008-08-22 16:23:09 -0400
committerIngo Molnar <mingo@elte.hu>2008-08-23 11:49:19 -0400
commit8735728ef8dc935c4fb351f913758fdbb62c308d (patch)
tree5c96e2fcb2b326d2ad1ad3aa8294e7cbeab3d007 /arch
parent91ede005d72df60d6b3f252be177a4743a6aa46a (diff)
x86 MCE: Fix CPU hotplug problem with multiple multicore AMD CPUs
During CPU hot-remove the sysfs directory created by threshold_create_bank(), defined in arch/x86/kernel/cpu/mcheck/mce_amd_64.c, has to be removed before its parent directory, created by mce_create_device(), defined in arch/x86/kernel/cpu/mcheck/mce_64.c . Moreover, when the CPU in question is hotplugged again, obviously the latter has to be created before the former. At present, the right ordering is not enforced, because all of these operations are carried out by CPU hotplug notifiers which are not appropriately ordered with respect to each other. This leads to serious problems on systems with two or more multicore AMD CPUs, among other things during suspend and hibernation. Fix the problem by placing threshold bank CPU hotplug callbacks in mce_cpu_callback(), so that they are invoked at the right places, if defined. Additionally, use kobject_del() to remove the sysfs directory associated with the kobject created by kobject_create_and_add() in threshold_create_bank(), to prevent the kernel from crashing during CPU hotplug operations on systems with two or more multicore AMD CPUs. This patch fixes bug #11337. Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl> Acked-by: Andi Kleen <andi@firstfloor.org> Tested-by: Mark Langsdorf <mark.langsdorf@amd.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Diffstat (limited to 'arch')
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_64.c5
-rw-r--r--arch/x86/kernel/cpu/mcheck/mce_amd_64.c18
2 files changed, 10 insertions, 13 deletions
diff --git a/arch/x86/kernel/cpu/mcheck/mce_64.c b/arch/x86/kernel/cpu/mcheck/mce_64.c
index 65a339678ece..726a5fcdf341 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_64.c
@@ -759,6 +759,7 @@ static struct sysdev_class mce_sysclass = {
759}; 759};
760 760
761DEFINE_PER_CPU(struct sys_device, device_mce); 761DEFINE_PER_CPU(struct sys_device, device_mce);
762void (*threshold_cpu_callback)(unsigned long action, unsigned int cpu) __cpuinitdata;
762 763
763/* Why are there no generic functions for this? */ 764/* Why are there no generic functions for this? */
764#define ACCESSOR(name, var, start) \ 765#define ACCESSOR(name, var, start) \
@@ -883,9 +884,13 @@ static int __cpuinit mce_cpu_callback(struct notifier_block *nfb,
883 case CPU_ONLINE: 884 case CPU_ONLINE:
884 case CPU_ONLINE_FROZEN: 885 case CPU_ONLINE_FROZEN:
885 mce_create_device(cpu); 886 mce_create_device(cpu);
887 if (threshold_cpu_callback)
888 threshold_cpu_callback(action, cpu);
886 break; 889 break;
887 case CPU_DEAD: 890 case CPU_DEAD:
888 case CPU_DEAD_FROZEN: 891 case CPU_DEAD_FROZEN:
892 if (threshold_cpu_callback)
893 threshold_cpu_callback(action, cpu);
889 mce_remove_device(cpu); 894 mce_remove_device(cpu);
890 break; 895 break;
891 } 896 }
diff --git a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
index 88736cadbaa6..5eb390a4b2e9 100644
--- a/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
+++ b/arch/x86/kernel/cpu/mcheck/mce_amd_64.c
@@ -628,6 +628,7 @@ static void threshold_remove_bank(unsigned int cpu, int bank)
628 deallocate_threshold_block(cpu, bank); 628 deallocate_threshold_block(cpu, bank);
629 629
630free_out: 630free_out:
631 kobject_del(b->kobj);
631 kobject_put(b->kobj); 632 kobject_put(b->kobj);
632 kfree(b); 633 kfree(b);
633 per_cpu(threshold_banks, cpu)[bank] = NULL; 634 per_cpu(threshold_banks, cpu)[bank] = NULL;
@@ -645,14 +646,11 @@ static void threshold_remove_device(unsigned int cpu)
645} 646}
646 647
647/* get notified when a cpu comes on/off */ 648/* get notified when a cpu comes on/off */
648static int __cpuinit threshold_cpu_callback(struct notifier_block *nfb, 649static void __cpuinit amd_64_threshold_cpu_callback(unsigned long action,
649 unsigned long action, void *hcpu) 650 unsigned int cpu)
650{ 651{
651 /* cpu was unsigned int to begin with */
652 unsigned int cpu = (unsigned long)hcpu;
653
654 if (cpu >= NR_CPUS) 652 if (cpu >= NR_CPUS)
655 goto out; 653 return;
656 654
657 switch (action) { 655 switch (action) {
658 case CPU_ONLINE: 656 case CPU_ONLINE:
@@ -666,14 +664,8 @@ static int __cpuinit threshold_cpu_callback(struct notifier_block *nfb,
666 default: 664 default:
667 break; 665 break;
668 } 666 }
669 out:
670 return NOTIFY_OK;
671} 667}
672 668
673static struct notifier_block threshold_cpu_notifier __cpuinitdata = {
674 .notifier_call = threshold_cpu_callback,
675};
676
677static __init int threshold_init_device(void) 669static __init int threshold_init_device(void)
678{ 670{
679 unsigned lcpu = 0; 671 unsigned lcpu = 0;
@@ -684,7 +676,7 @@ static __init int threshold_init_device(void)
684 if (err) 676 if (err)
685 return err; 677 return err;
686 } 678 }
687 register_hotcpu_notifier(&threshold_cpu_notifier); 679 threshold_cpu_callback = amd_64_threshold_cpu_callback;
688 return 0; 680 return 0;
689} 681}
690 682