patch-2_6_7-vs1_9_1_12
[linux-2.6.git] / arch / ppc64 / kernel / smp.c
index 2447cce..7fa0e5a 100644 (file)
@@ -390,8 +390,7 @@ static inline int __devinit smp_startup_cpu(unsigned int lcpu)
        }
 
        /* Fixup atomic count: it exited inside IRQ handler. */
-       ((struct task_struct *)paca[lcpu].xCurrent)->thread_info->preempt_count
-               = 0;
+       paca[lcpu].xCurrent->thread_info->preempt_count = 0;
        /* Fixup SLB round-robin so next segment (kernel) goes in segment 0 */
        paca[lcpu].xStab_data.next_round_robin = 0;
 
@@ -618,7 +617,7 @@ void smp_message_recv(int msg, struct pt_regs *regs)
 #endif
 #ifdef CONFIG_DEBUGGER
        case PPC_MSG_DEBUGGER_BREAK:
-               debugger(regs);
+               debugger_ipi(regs);
                break;
 #endif
        default:
@@ -692,6 +691,9 @@ int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
        int ret = -1, cpus;
        unsigned long timeout;
 
+       /* Can deadlock when called with interrupts disabled */
+       WARN_ON(irqs_disabled());
+
        data.func = func;
        data.info = info;
        atomic_set(&data.started, 0);
@@ -814,7 +816,7 @@ static void __init smp_create_idle(unsigned int cpu)
        init_idle(p, cpu);
        unhash_process(p);
 
-       paca[cpu].xCurrent = (u64)p;
+       paca[cpu].xCurrent = p;
        current_set[cpu] = p->thread_info;
 }
 
@@ -834,11 +836,6 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
        paca[boot_cpuid].prof_counter = 1;
        paca[boot_cpuid].prof_multiplier = 1;
 
-       /*
-        * XXX very rough. 
-        */
-       cache_decay_ticks = HZ/100;
-
 #ifndef CONFIG_PPC_ISERIES
        paca[boot_cpuid].next_jiffy_update_tb = tb_last_stamp = get_tb();
 
@@ -871,7 +868,7 @@ void __devinit smp_prepare_boot_cpu(void)
        /* cpu_possible is set up in prom.c */
        cpu_set(boot_cpuid, cpu_online_map);
 
-       paca[boot_cpuid].xCurrent = (u64)current;
+       paca[boot_cpuid].xCurrent = current;
        current_set[boot_cpuid] = current->thread_info;
 }
 
@@ -914,8 +911,20 @@ int __devinit __cpu_up(unsigned int cpu)
         * use this value that I found through experimentation.
         * -- Cort
         */
-       for (c = 5000; c && !cpu_callin_map[cpu]; c--)
-               udelay(100);
+       if (system_state == SYSTEM_BOOTING)
+               for (c = 5000; c && !cpu_callin_map[cpu]; c--)
+                       udelay(100);
+#ifdef CONFIG_HOTPLUG_CPU
+       else
+               /*
+                * CPUs can take much longer to come up in the
+                * hotplug case.  Wait five seconds.
+                */
+               for (c = 25; c && !cpu_callin_map[cpu]; c--) {
+                       set_current_state(TASK_UNINTERRUPTIBLE);
+                       schedule_timeout(HZ/5);
+               }
+#endif
 
        if (!cpu_callin_map[cpu]) {
                printk("Processor %u is stuck.\n", cpu);
@@ -992,3 +1001,218 @@ void __init smp_cpus_done(unsigned int max_cpus)
 
        set_cpus_allowed(current, old_mask);
 }
+
+#ifdef CONFIG_SCHED_SMT
+#ifdef CONFIG_NUMA
+static struct sched_group sched_group_cpus[NR_CPUS];
+static struct sched_group sched_group_phys[NR_CPUS];
+static struct sched_group sched_group_nodes[MAX_NUMNODES];
+static DEFINE_PER_CPU(struct sched_domain, cpu_domains);
+static DEFINE_PER_CPU(struct sched_domain, phys_domains);
+static DEFINE_PER_CPU(struct sched_domain, node_domains);
+__init void arch_init_sched_domains(void)
+{
+       int i;
+       struct sched_group *first = NULL, *last = NULL;
+
+       /* Set up domains */
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               struct sched_domain *phys_domain = &per_cpu(phys_domains, i);
+               struct sched_domain *node_domain = &per_cpu(node_domains, i);
+               int node = cpu_to_node(i);
+               cpumask_t nodemask = node_to_cpumask(node);
+               cpumask_t my_cpumask = cpumask_of_cpu(i);
+               cpumask_t sibling_cpumask = cpumask_of_cpu(i ^ 0x1);
+
+               *cpu_domain = SD_SIBLING_INIT;
+               if (cur_cpu_spec->cpu_features & CPU_FTR_SMT)
+                       cpus_or(cpu_domain->span, my_cpumask, sibling_cpumask);
+               else
+                       cpu_domain->span = my_cpumask;
+               cpu_domain->parent = phys_domain;
+               cpu_domain->groups = &sched_group_cpus[i];
+
+               *phys_domain = SD_CPU_INIT;
+               phys_domain->span = nodemask;
+               phys_domain->parent = node_domain;
+               phys_domain->groups = &sched_group_phys[first_cpu(cpu_domain->span)];
+
+               *node_domain = SD_NODE_INIT;
+               node_domain->span = cpu_possible_map;
+               node_domain->groups = &sched_group_nodes[node];
+       }
+
+       /* Set up CPU (sibling) groups */
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               int j;
+               first = last = NULL;
+
+               if (i != first_cpu(cpu_domain->span))
+                       continue;
+
+               for_each_cpu_mask(j, cpu_domain->span) {
+                       struct sched_group *cpu = &sched_group_cpus[j];
+
+                       cpus_clear(cpu->cpumask);
+                       cpu_set(j, cpu->cpumask);
+                       cpu->cpu_power = SCHED_LOAD_SCALE;
+
+                       if (!first)
+                               first = cpu;
+                       if (last)
+                               last->next = cpu;
+                       last = cpu;
+               }
+               last->next = first;
+       }
+
+       for (i = 0; i < MAX_NUMNODES; i++) {
+               int j;
+               cpumask_t nodemask;
+               struct sched_group *node = &sched_group_nodes[i];
+               cpumask_t node_cpumask = node_to_cpumask(i);
+               cpus_and(nodemask, node_cpumask, cpu_possible_map);
+
+               if (cpus_empty(nodemask))
+                       continue;
+
+               first = last = NULL;
+               /* Set up physical groups */
+               for_each_cpu_mask(j, nodemask) {
+                       struct sched_domain *cpu_domain = &per_cpu(cpu_domains, j);
+                       struct sched_group *cpu = &sched_group_phys[j];
+
+                       if (j != first_cpu(cpu_domain->span))
+                               continue;
+
+                       cpu->cpumask = cpu_domain->span;
+                       /*
+                        * Make each extra sibling increase power by 10% of
+                        * the basic CPU. This is very arbitrary.
+                        */
+                       cpu->cpu_power = SCHED_LOAD_SCALE + SCHED_LOAD_SCALE*(cpus_weight(cpu->cpumask)-1) / 10;
+                       node->cpu_power += cpu->cpu_power;
+
+                       if (!first)
+                               first = cpu;
+                       if (last)
+                               last->next = cpu;
+                       last = cpu;
+               }
+               last->next = first;
+       }
+
+       /* Set up nodes */
+       first = last = NULL;
+       for (i = 0; i < MAX_NUMNODES; i++) {
+               struct sched_group *cpu = &sched_group_nodes[i];
+               cpumask_t nodemask;
+               cpumask_t node_cpumask = node_to_cpumask(i);
+               cpus_and(nodemask, node_cpumask, cpu_possible_map);
+
+               if (cpus_empty(nodemask))
+                       continue;
+
+               cpu->cpumask = nodemask;
+               /* ->cpu_power already setup */
+
+               if (!first)
+                       first = cpu;
+               if (last)
+                       last->next = cpu;
+               last = cpu;
+       }
+       last->next = first;
+
+       mb();
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               cpu_attach_domain(cpu_domain, i);
+       }
+}
+#else /* !CONFIG_NUMA */
+static struct sched_group sched_group_cpus[NR_CPUS];
+static struct sched_group sched_group_phys[NR_CPUS];
+static DEFINE_PER_CPU(struct sched_domain, cpu_domains);
+static DEFINE_PER_CPU(struct sched_domain, phys_domains);
+__init void arch_init_sched_domains(void)
+{
+       int i;
+       struct sched_group *first = NULL, *last = NULL;
+
+       /* Set up domains */
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               struct sched_domain *phys_domain = &per_cpu(phys_domains, i);
+               cpumask_t my_cpumask = cpumask_of_cpu(i);
+               cpumask_t sibling_cpumask = cpumask_of_cpu(i ^ 0x1);
+
+               *cpu_domain = SD_SIBLING_INIT;
+               if (cur_cpu_spec->cpu_features & CPU_FTR_SMT)
+                       cpus_or(cpu_domain->span, my_cpumask, sibling_cpumask);
+               else
+                       cpu_domain->span = my_cpumask;
+               cpu_domain->parent = phys_domain;
+               cpu_domain->groups = &sched_group_cpus[i];
+
+               *phys_domain = SD_CPU_INIT;
+               phys_domain->span = cpu_possible_map;
+               phys_domain->groups = &sched_group_phys[first_cpu(cpu_domain->span)];
+       }
+
+       /* Set up CPU (sibling) groups */
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               int j;
+               first = last = NULL;
+
+               if (i != first_cpu(cpu_domain->span))
+                       continue;
+
+               for_each_cpu_mask(j, cpu_domain->span) {
+                       struct sched_group *cpu = &sched_group_cpus[j];
+
+                       cpus_clear(cpu->cpumask);
+                       cpu_set(j, cpu->cpumask);
+                       cpu->cpu_power = SCHED_LOAD_SCALE;
+
+                       if (!first)
+                               first = cpu;
+                       if (last)
+                               last->next = cpu;
+                       last = cpu;
+               }
+               last->next = first;
+       }
+
+       first = last = NULL;
+       /* Set up physical groups */
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               struct sched_group *cpu = &sched_group_phys[i];
+
+               if (i != first_cpu(cpu_domain->span))
+                       continue;
+
+               cpu->cpumask = cpu_domain->span;
+               /* See SMT+NUMA setup for comment */
+               cpu->cpu_power = SCHED_LOAD_SCALE + SCHED_LOAD_SCALE*(cpus_weight(cpu->cpumask)-1) / 10;
+
+               if (!first)
+                       first = cpu;
+               if (last)
+                       last->next = cpu;
+               last = cpu;
+       }
+       last->next = first;
+
+       mb();
+       for_each_cpu(i) {
+               struct sched_domain *cpu_domain = &per_cpu(cpu_domains, i);
+               cpu_attach_domain(cpu_domain, i);
+       }
+}
+#endif /* CONFIG_NUMA */
+#endif /* CONFIG_SCHED_SMT */