#include <asm/nmi.h>
#include <asm/irq.h>
#include <asm/hw_irq.h>
+#include <asm/numa.h>
/* Number of siblings per CPU package */
int smp_num_siblings = 1;
-/* Package ID of each logical CPU */
-u8 phys_proc_id[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID };
-u8 cpu_core_id[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID };
-EXPORT_SYMBOL(phys_proc_id);
-EXPORT_SYMBOL(cpu_core_id);
+EXPORT_SYMBOL(smp_num_siblings);
+
+/* Last level cache ID of each logical CPU */
+u8 cpu_llc_id[NR_CPUS] __cpuinitdata = {[0 ... NR_CPUS-1] = BAD_APICID};
+EXPORT_SYMBOL(cpu_llc_id);
/* Bitmask of currently online CPUs */
cpumask_t cpu_online_map __read_mostly;
*/
cpumask_t cpu_callin_map;
cpumask_t cpu_callout_map;
+EXPORT_SYMBOL(cpu_callout_map);
cpumask_t cpu_possible_map;
EXPORT_SYMBOL(cpu_possible_map);
/* Per CPU bogomips and other parameters */
struct cpuinfo_x86 cpu_data[NR_CPUS] __cacheline_aligned;
+EXPORT_SYMBOL(cpu_data);
/* Set when the idlers are all forked */
int smp_threads_ready;
+/* representing HT siblings of each logical CPU */
cpumask_t cpu_sibling_map[NR_CPUS] __read_mostly;
+EXPORT_SYMBOL(cpu_sibling_map);
+
+/* representing HT and core siblings of each logical CPU */
cpumask_t cpu_core_map[NR_CPUS] __read_mostly;
EXPORT_SYMBOL(cpu_core_map);
static void __cpuinit tsc_sync_wait(void)
{
- if (notscsync || !cpu_has_tsc)
+ /*
+ * When the CPU has synchronized TSCs assume the BIOS
+ * or the hardware already synced. Otherwise we could
+ * mess up a possible perfect synchronization with a
+ * not-quite-perfect algorithm.
+ */
+ if (notscsync || !cpu_has_tsc || !unsynchronized_tsc())
return;
sync_tsc(0);
}
static __init int notscsync_setup(char *s)
{
notscsync = 1;
- return 0;
+ return 1;
}
__setup("notscsync", notscsync_setup);
cpu_set(cpuid, cpu_callin_map);
}
+/* maps the cpu to the sched domain representing multi-core */
+cpumask_t cpu_coregroup_map(int cpu)
+{
+ struct cpuinfo_x86 *c = cpu_data + cpu;
+ /*
+ * For perf, we return last level cache shared map.
+ * And for power savings, we return cpu_core_map
+ */
+ if (sched_mc_power_savings || sched_smt_power_savings)
+ return cpu_core_map[cpu];
+ else
+ return c->llc_shared_map;
+}
+
+/* representing cpus for which sibling maps can be computed */
+static cpumask_t cpu_sibling_setup_map;
+
static inline void set_cpu_sibling_map(int cpu)
{
int i;
+ struct cpuinfo_x86 *c = cpu_data;
+
+ cpu_set(cpu, cpu_sibling_setup_map);
if (smp_num_siblings > 1) {
- for_each_cpu(i) {
- if (cpu_core_id[cpu] == cpu_core_id[i]) {
+ for_each_cpu_mask(i, cpu_sibling_setup_map) {
+ if (c[cpu].phys_proc_id == c[i].phys_proc_id &&
+ c[cpu].cpu_core_id == c[i].cpu_core_id) {
cpu_set(i, cpu_sibling_map[cpu]);
cpu_set(cpu, cpu_sibling_map[i]);
+ cpu_set(i, cpu_core_map[cpu]);
+ cpu_set(cpu, cpu_core_map[i]);
+ cpu_set(i, c[cpu].llc_shared_map);
+ cpu_set(cpu, c[i].llc_shared_map);
}
}
} else {
cpu_set(cpu, cpu_sibling_map[cpu]);
}
- if (current_cpu_data.x86_num_cores > 1) {
- for_each_cpu(i) {
- if (phys_proc_id[cpu] == phys_proc_id[i]) {
- cpu_set(i, cpu_core_map[cpu]);
- cpu_set(cpu, cpu_core_map[i]);
- }
- }
- } else {
+ cpu_set(cpu, c[cpu].llc_shared_map);
+
+ if (current_cpu_data.x86_max_cores == 1) {
cpu_core_map[cpu] = cpu_sibling_map[cpu];
+ c[cpu].booted_cores = 1;
+ return;
+ }
+
+ for_each_cpu_mask(i, cpu_sibling_setup_map) {
+ if (cpu_llc_id[cpu] != BAD_APICID &&
+ cpu_llc_id[cpu] == cpu_llc_id[i]) {
+ cpu_set(i, c[cpu].llc_shared_map);
+ cpu_set(cpu, c[i].llc_shared_map);
+ }
+ if (c[cpu].phys_proc_id == c[i].phys_proc_id) {
+ cpu_set(i, cpu_core_map[cpu]);
+ cpu_set(cpu, cpu_core_map[i]);
+ /*
+ * Does this new cpu bringup a new core?
+ */
+ if (cpus_weight(cpu_sibling_map[cpu]) == 1) {
+ /*
+ * for each core in package, increment
+ * the booted_cores for this new cpu
+ */
+ if (first_cpu(cpu_sibling_map[i]) == i)
+ c[cpu].booted_cores++;
+ /*
+ * increment the core count for all
+ * the other cpus in this package
+ */
+ if (i != cpu)
+ c[i].booted_cores++;
+ } else if (i != cpu && !c[cpu].booted_cores)
+ c[cpu].booted_cores = c[i].booted_cores;
+ }
}
}
* things done here to the most necessary things.
*/
cpu_init();
+ preempt_disable();
smp_callin();
/* otherwise gcc will move up the smp_processor_id before the cpu_init */
send_status = apic_read(APIC_ICR) & APIC_ICR_BUSY;
} while (send_status && (timeout++ < 1000));
+ mb();
atomic_set(&init_deasserted, 1);
num_starts = 2;
for (j = 1; j <= num_starts; j++) {
Dprintk("Sending STARTUP #%d.\n",j);
- apic_read_around(APIC_SPIV);
apic_write(APIC_ESR, 0);
apic_read(APIC_ESR);
Dprintk("After apic_write.\n");
* Due to the Pentium erratum 3AP.
*/
if (maxlvt > 3) {
- apic_read_around(APIC_SPIV);
apic_write(APIC_ESR, 0);
}
accept_status = (apic_read(APIC_ESR) & 0xEF);
};
DECLARE_WORK(work, do_fork_idle, &c_idle);
+ /* allocate memory for gdts of secondary cpus. Hotplug is considered */
+ if (!cpu_gdt_descr[cpu].address &&
+ !(cpu_gdt_descr[cpu].address = get_zeroed_page(GFP_KERNEL))) {
+ printk(KERN_ERR "Failed to allocate GDT for CPU %d\n", cpu);
+ return -1;
+ }
+
+ /* Allocate node local memory for AP pdas */
+ if (cpu_pda(cpu) == &boot_cpu_pda[cpu]) {
+ struct x8664_pda *newpda, *pda;
+ int node = cpu_to_node(cpu);
+ pda = cpu_pda(cpu);
+ newpda = kmalloc_node(sizeof (struct x8664_pda), GFP_ATOMIC,
+ node);
+ if (newpda) {
+ memcpy(newpda, pda, sizeof (struct x8664_pda));
+ cpu_pda(cpu) = newpda;
+ } else
+ printk(KERN_ERR
+ "Could not allocate node local PDA for CPU %d on node %d\n",
+ cpu, node);
+ }
+
+
+ alternatives_smp_switch(1);
+
c_idle.idle = get_idle_for_cpu(cpu);
if (c_idle.idle) {
c_idle.idle->thread.rsp = (unsigned long) (((struct pt_regs *)
- (THREAD_SIZE + (unsigned long) c_idle.idle->thread_info)) - 1);
+ (THREAD_SIZE + task_stack_page(c_idle.idle))) - 1);
init_idle(c_idle.idle, cpu);
goto do_rest;
}
do_rest:
- cpu_pda[cpu].pcurrent = c_idle.idle;
+ cpu_pda(cpu)->pcurrent = c_idle.idle;
start_rip = setup_trampoline();
init_rsp = c_idle.idle->thread.rsp;
per_cpu(init_tss,cpu).rsp0 = init_rsp;
initial_code = start_secondary;
- clear_ti_thread_flag(c_idle.idle->thread_info, TIF_FORK);
+ clear_tsk_thread_flag(c_idle.idle, TIF_FORK);
printk(KERN_INFO "Booting processor %d/%d APIC 0x%x\n", cpu,
cpus_weight(cpu_present_map),
/*
* Be paranoid about clearing APIC errors.
*/
- if (APIC_INTEGRATED(apic_version[apicid])) {
- apic_read_around(APIC_SPIV);
- apic_write(APIC_ESR, 0);
- apic_read(APIC_ESR);
- }
+ apic_write(APIC_ESR, 0);
+ apic_read(APIC_ESR);
/*
* Status is now clean
if (boot_error) {
cpu_clear(cpu, cpu_callout_map); /* was set here (do_boot_cpu()) */
clear_bit(cpu, &cpu_initialized); /* was set by cpu_init() */
+ clear_node_cpumask(cpu); /* was set by numa_add_cpu */
cpu_clear(cpu, cpu_present_map);
cpu_clear(cpu, cpu_possible_map);
x86_cpu_to_apicid[cpu] = BAD_APICID;
*
* Three ways to find out the number of additional hotplug CPUs:
* - If the BIOS specified disabled CPUs in ACPI/mptables use that.
- * - otherwise use half of the available CPUs or 2, whatever is more.
* - The user can overwrite it with additional_cpus=NUM
+ * - Otherwise don't reserve additional CPUs.
* We do this because additional CPUs waste a lot of memory.
* -AK
*/
int possible;
if (additional_cpus == -1) {
- if (disabled_cpus > 0) {
+ if (disabled_cpus > 0)
additional_cpus = disabled_cpus;
- } else {
- additional_cpus = num_processors / 2;
- if (additional_cpus == 0)
- additional_cpus = 2;
- }
+ else
+ additional_cpus = 0;
}
possible = num_processors + additional_cpus;
if (possible > NR_CPUS)
/*
* If we couldn't find a local APIC, then get out of here now!
*/
- if (APIC_INTEGRATED(apic_version[boot_cpu_id]) && !cpu_has_apic) {
+ if (!cpu_has_apic) {
printk(KERN_ERR "BIOS bug, local APIC #%d not detected!...\n",
boot_cpu_id);
printk(KERN_ERR "... forcing use of dummy APIC emulation. (tell your hw vendor)\n");
nmi_watchdog_default();
current_cpu_data = boot_cpu_data;
current_thread_info()->cpu = 0; /* needed? */
+ set_cpu_sibling_map(0);
if (smp_sanity_check(max_cpus) < 0) {
printk(KERN_INFO "SMP disabled\n");
int me = smp_processor_id();
cpu_set(me, cpu_online_map);
cpu_set(me, cpu_callout_map);
- cpu_set(0, cpu_sibling_map[0]);
- cpu_set(0, cpu_core_map[0]);
per_cpu(cpu_state, me) = CPU_ONLINE;
}
setup_ioapic_dest();
#endif
- time_init_gtod();
-
check_nmi_watchdog();
}
static void remove_siblinginfo(int cpu)
{
int sibling;
+ struct cpuinfo_x86 *c = cpu_data;
+ for_each_cpu_mask(sibling, cpu_core_map[cpu]) {
+ cpu_clear(cpu, cpu_core_map[sibling]);
+ /*
+ * last thread sibling in this cpu core going down
+ */
+ if (cpus_weight(cpu_sibling_map[cpu]) == 1)
+ c[sibling].booted_cores--;
+ }
+
for_each_cpu_mask(sibling, cpu_sibling_map[cpu])
cpu_clear(cpu, cpu_sibling_map[sibling]);
- for_each_cpu_mask(sibling, cpu_core_map[cpu])
- cpu_clear(cpu, cpu_core_map[sibling]);
cpus_clear(cpu_sibling_map[cpu]);
cpus_clear(cpu_core_map[cpu]);
- phys_proc_id[cpu] = BAD_APICID;
- cpu_core_id[cpu] = BAD_APICID;
+ c[cpu].phys_proc_id = 0;
+ c[cpu].cpu_core_id = 0;
+ cpu_clear(cpu, cpu_sibling_setup_map);
}
void remove_cpu_from_maps(void)
cpu_clear(cpu, cpu_callout_map);
cpu_clear(cpu, cpu_callin_map);
clear_bit(cpu, &cpu_initialized); /* was set by cpu_init() */
+ clear_node_cpumask(cpu);
}
int __cpu_disable(void)
if (cpu == 0)
return -EBUSY;
- disable_APIC_timer();
+ clear_local_APIC();
/*
* HACK:
/* They ack this in play_dead by setting CPU_DEAD */
if (per_cpu(cpu_state, cpu) == CPU_DEAD) {
printk ("CPU %d is now offline\n", cpu);
+ if (1 == num_online_cpus())
+ alternatives_smp_switch(0);
return;
}
msleep(100);
printk(KERN_ERR "CPU %u didn't die...\n", cpu);
}
-static __init int setup_additional_cpus(char *s)
+__init int setup_additional_cpus(char *s)
{
return get_option(&s, &additional_cpus);
}