aboutsummaryrefslogtreecommitdiffstats
path: root/arch/x86/kernel/smpboot.c
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2014-06-07 17:50:38 -0400
committerLinus Torvalds <torvalds@linux-foundation.org>2014-06-07 17:50:38 -0400
commit813895f8dcb31bc6b0e9f5fc35e8c687a467f3dd (patch)
tree576996d2cae905f4af87a8604cfa5c59bdef54f5 /arch/x86/kernel/smpboot.c
parentd4c54919ed86302094c0ca7d48a8cbd4ee753e92 (diff)
parent745c51673e289acf4d9ffc2835524de73ef923fd (diff)
Merge branch 'x86/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Peter Anvin: "A significantly larger than I'd like set of patches for just below the wire. All of these, however, fix real problems. The one thing that is genuinely scary in here is the change of SMP initialization, but that *does* fix a confirmed hang when booting virtual machines. There is also a patch to actually do the right thing about not offlining a CPU when there are not enough interrupt vectors available in the system; the accounting was done incorrectly. The worst case for that patch is that we fail to offline CPUs when we should (the new code is strictly more conservative than the old), so is not particularly risky. Most of the rest is minor stuff; the EFI patches are all about exporting correct information to boot loaders and kexec" * 'x86/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86/boot: EFI_MIXED should not prohibit loading above 4G x86/smpboot: Initialize secondary CPU only if master CPU will wait for it x86/smpboot: Log error on secondary CPU wakeup failure at ERR level x86: Fix list/memory corruption on CPU hotplug x86: irq: Get correct available vectors for cpu disable x86/efi: Do not export efi runtime map in case old map x86/efi: earlyprintk=efi,keep fix
Diffstat (limited to 'arch/x86/kernel/smpboot.c')
-rw-r--r--arch/x86/kernel/smpboot.c104
1 files changed, 32 insertions, 72 deletions
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 34826934d4a7..bc52fac39dd3 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -111,7 +111,6 @@ atomic_t init_deasserted;
111static void smp_callin(void) 111static void smp_callin(void)
112{ 112{
113 int cpuid, phys_id; 113 int cpuid, phys_id;
114 unsigned long timeout;
115 114
116 /* 115 /*
117 * If waken up by an INIT in an 82489DX configuration 116 * If waken up by an INIT in an 82489DX configuration
@@ -130,37 +129,6 @@ static void smp_callin(void)
130 * (This works even if the APIC is not enabled.) 129 * (This works even if the APIC is not enabled.)
131 */ 130 */
132 phys_id = read_apic_id(); 131 phys_id = read_apic_id();
133 if (cpumask_test_cpu(cpuid, cpu_callin_mask)) {
134 panic("%s: phys CPU#%d, CPU#%d already present??\n", __func__,
135 phys_id, cpuid);
136 }
137 pr_debug("CPU#%d (phys ID: %d) waiting for CALLOUT\n", cpuid, phys_id);
138
139 /*
140 * STARTUP IPIs are fragile beasts as they might sometimes
141 * trigger some glue motherboard logic. Complete APIC bus
142 * silence for 1 second, this overestimates the time the
143 * boot CPU is spending to send the up to 2 STARTUP IPIs
144 * by a factor of two. This should be enough.
145 */
146
147 /*
148 * Waiting 2s total for startup (udelay is not yet working)
149 */
150 timeout = jiffies + 2*HZ;
151 while (time_before(jiffies, timeout)) {
152 /*
153 * Has the boot CPU finished it's STARTUP sequence?
154 */
155 if (cpumask_test_cpu(cpuid, cpu_callout_mask))
156 break;
157 cpu_relax();
158 }
159
160 if (!time_before(jiffies, timeout)) {
161 panic("%s: CPU%d started up but did not get a callout!\n",
162 __func__, cpuid);
163 }
164 132
165 /* 133 /*
166 * the boot CPU has finished the init stage and is spinning 134 * the boot CPU has finished the init stage and is spinning
@@ -750,8 +718,8 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
750 unsigned long start_ip = real_mode_header->trampoline_start; 718 unsigned long start_ip = real_mode_header->trampoline_start;
751 719
752 unsigned long boot_error = 0; 720 unsigned long boot_error = 0;
753 int timeout;
754 int cpu0_nmi_registered = 0; 721 int cpu0_nmi_registered = 0;
722 unsigned long timeout;
755 723
756 /* Just in case we booted with a single CPU. */ 724 /* Just in case we booted with a single CPU. */
757 alternatives_enable_smp(); 725 alternatives_enable_smp();
@@ -799,6 +767,15 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
799 } 767 }
800 768
801 /* 769 /*
770 * AP might wait on cpu_callout_mask in cpu_init() with
771 * cpu_initialized_mask set if previous attempt to online
772 * it timed-out. Clear cpu_initialized_mask so that after
773 * INIT/SIPI it could start with a clean state.
774 */
775 cpumask_clear_cpu(cpu, cpu_initialized_mask);
776 smp_mb();
777
778 /*
802 * Wake up a CPU in difference cases: 779 * Wake up a CPU in difference cases:
803 * - Use the method in the APIC driver if it's defined 780 * - Use the method in the APIC driver if it's defined
804 * Otherwise, 781 * Otherwise,
@@ -810,58 +787,41 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
810 boot_error = wakeup_cpu_via_init_nmi(cpu, start_ip, apicid, 787 boot_error = wakeup_cpu_via_init_nmi(cpu, start_ip, apicid,
811 &cpu0_nmi_registered); 788 &cpu0_nmi_registered);
812 789
790
813 if (!boot_error) { 791 if (!boot_error) {
814 /* 792 /*
815 * allow APs to start initializing. 793 * Wait 10s total for a response from AP
816 */ 794 */
817 pr_debug("Before Callout %d\n", cpu); 795 boot_error = -1;
818 cpumask_set_cpu(cpu, cpu_callout_mask); 796 timeout = jiffies + 10*HZ;
819 pr_debug("After Callout %d\n", cpu); 797 while (time_before(jiffies, timeout)) {
798 if (cpumask_test_cpu(cpu, cpu_initialized_mask)) {
799 /*
800 * Tell AP to proceed with initialization
801 */
802 cpumask_set_cpu(cpu, cpu_callout_mask);
803 boot_error = 0;
804 break;
805 }
806 udelay(100);
807 schedule();
808 }
809 }
820 810
811 if (!boot_error) {
821 /* 812 /*
822 * Wait 5s total for a response 813 * Wait till AP completes initial initialization
823 */ 814 */
824 for (timeout = 0; timeout < 50000; timeout++) { 815 while (!cpumask_test_cpu(cpu, cpu_callin_mask)) {
825 if (cpumask_test_cpu(cpu, cpu_callin_mask))
826 break; /* It has booted */
827 udelay(100);
828 /* 816 /*
829 * Allow other tasks to run while we wait for the 817 * Allow other tasks to run while we wait for the
830 * AP to come online. This also gives a chance 818 * AP to come online. This also gives a chance
831 * for the MTRR work(triggered by the AP coming online) 819 * for the MTRR work(triggered by the AP coming online)
832 * to be completed in the stop machine context. 820 * to be completed in the stop machine context.
833 */ 821 */
822 udelay(100);
834 schedule(); 823 schedule();
835 } 824 }
836
837 if (cpumask_test_cpu(cpu, cpu_callin_mask)) {
838 print_cpu_msr(&cpu_data(cpu));
839 pr_debug("CPU%d: has booted.\n", cpu);
840 } else {
841 boot_error = 1;
842 if (*trampoline_status == 0xA5A5A5A5)
843 /* trampoline started but...? */
844 pr_err("CPU%d: Stuck ??\n", cpu);
845 else
846 /* trampoline code not run */
847 pr_err("CPU%d: Not responding\n", cpu);
848 if (apic->inquire_remote_apic)
849 apic->inquire_remote_apic(apicid);
850 }
851 }
852
853 if (boot_error) {
854 /* Try to put things back the way they were before ... */
855 numa_remove_cpu(cpu); /* was set by numa_add_cpu */
856
857 /* was set by do_boot_cpu() */
858 cpumask_clear_cpu(cpu, cpu_callout_mask);
859
860 /* was set by cpu_init() */
861 cpumask_clear_cpu(cpu, cpu_initialized_mask);
862
863 set_cpu_present(cpu, false);
864 per_cpu(x86_cpu_to_apicid, cpu) = BAD_APICID;
865 } 825 }
866 826
867 /* mark "stuck" area as not stuck */ 827 /* mark "stuck" area as not stuck */
@@ -921,7 +881,7 @@ int native_cpu_up(unsigned int cpu, struct task_struct *tidle)
921 881
922 err = do_boot_cpu(apicid, cpu, tidle); 882 err = do_boot_cpu(apicid, cpu, tidle);
923 if (err) { 883 if (err) {
924 pr_debug("do_boot_cpu failed %d\n", err); 884 pr_err("do_boot_cpu failed(%d) to wakeup CPU#%u\n", err, cpu);
925 return -EIO; 885 return -EIO;
926 } 886 }
927 887