* Pentium Pro and Pentium-II/Xeon MP machines.
* Original development of Linux SMP code supported by Caldera.
*
- * This code is released under the GNU General Public License version 2 or
- * later.
+ * This code is released under the GNU General Public License version 2
*
* Fixes
* Felix Koop : NR_CPUS used properly
* Tigran Aivazian : fixed "0.00 in /proc/uptime on SMP" bug.
* Maciej W. Rozycki : Bits for genuine 82489DX APICs
* Andi Kleen : Changed for SMP boot into long mode.
- * Rusty Russell : Hacked into shape for new "hotplug" boot process.
+ * Rusty Russell : Hacked into shape for new "hotplug" boot process.
+ * Andi Kleen : Converted to new state machine.
+ * Various cleanups.
+ * Probably mostly hotplug CPU ready now.
+ * Ashok Raj : CPU hotplug support
*/
+
#include <linux/config.h>
#include <linux/init.h>
#include <linux/mm.h>
#include <linux/kernel_stat.h>
#include <linux/smp_lock.h>
-#include <linux/irq.h>
#include <linux/bootmem.h>
#include <linux/thread_info.h>
+#include <linux/module.h>
#include <linux/delay.h>
#include <linux/mc146818rtc.h>
#include <asm/kdebug.h>
#include <asm/tlbflush.h>
#include <asm/proto.h>
+#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;
-char phys_proc_id[NR_CPUS]; /* Package ID of each logical CPU */
+/* Package ID of each logical CPU */
+u8 phys_proc_id[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID };
+/* core ID of each logical CPU */
+u8 cpu_core_id[NR_CPUS] __read_mostly = { [0 ... NR_CPUS-1] = BAD_APICID };
+
+/* Last level cache ID of each logical CPU */
+u8 cpu_llc_id[NR_CPUS] __cpuinitdata = {[0 ... NR_CPUS-1] = BAD_APICID};
/* Bitmask of currently online CPUs */
-cpumask_t cpu_online_map;
+cpumask_t cpu_online_map __read_mostly;
-/* which logical CPU number maps to which CPU (physical APIC ID) */
-volatile char x86_cpu_to_apicid[NR_CPUS];
+EXPORT_SYMBOL(cpu_online_map);
-static cpumask_t cpu_callin_map;
+/*
+ * Private maps to synchronize booting between AP and BP.
+ * Probably not needed anymore, but it makes for easier debugging. -AK
+ */
+cpumask_t cpu_callin_map;
cpumask_t cpu_callout_map;
-static cpumask_t smp_commenced_mask;
+
+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;
/* Set when the idlers are all forked */
int smp_threads_ready;
-char cpu_sibling_map[NR_CPUS] __cacheline_aligned;
+/* representing HT siblings of each logical CPU */
+cpumask_t cpu_sibling_map[NR_CPUS] __read_mostly;
+
+/* representing HT and core siblings of each logical CPU */
+cpumask_t cpu_core_map[NR_CPUS] __read_mostly;
+EXPORT_SYMBOL(cpu_core_map);
/*
* Trampoline 80x86 program as an array.
*/
-extern unsigned char trampoline_data [];
-extern unsigned char trampoline_end [];
+extern unsigned char trampoline_data[];
+extern unsigned char trampoline_end[];
+
+/* State of each CPU */
+DEFINE_PER_CPU(int, cpu_state) = { 0 };
+
+/*
+ * Store all idle threads, this can be reused instead of creating
+ * a new thread. Also avoids complicated thread destroy functionality
+ * for idle threads.
+ */
+struct task_struct *idle_thread_array[NR_CPUS] __cpuinitdata ;
+
+#define get_idle_for_cpu(x) (idle_thread_array[(x)])
+#define set_idle_for_cpu(x,p) (idle_thread_array[(x)] = (p))
/*
* Currently trivial. Write the real->protected mode
* has made sure it's suitably aligned.
*/
-static unsigned long __init setup_trampoline(void)
+static unsigned long __cpuinit setup_trampoline(void)
{
void *tramp = __va(SMP_TRAMPOLINE_BASE);
- extern volatile __u32 tramp_gdt_ptr;
- tramp_gdt_ptr = __pa_symbol(&cpu_gdt_table);
memcpy(tramp, trampoline_data, trampoline_end - trampoline_data);
return virt_to_phys(tramp);
}
* a given CPU
*/
-static void __init smp_store_cpu_info(int id)
+static void __cpuinit smp_store_cpu_info(int id)
{
struct cpuinfo_x86 *c = cpu_data + id;
*c = boot_cpu_data;
identify_cpu(c);
+ print_cpu_info(c);
}
/*
- * TSC synchronization.
+ * New Funky TSC sync algorithm borrowed from IA64.
+ * Main advantage is that it doesn't reset the TSCs fully and
+ * in general looks more robust and it works better than my earlier
+ * attempts. I believe it was written by David Mosberger. Some minor
+ * adjustments for x86-64 by me -AK
+ *
+ * Original comment reproduced below.
+ *
+ * Synchronize TSC of the current (slave) CPU with the TSC of the
+ * MASTER CPU (normally the time-keeper CPU). We use a closed loop to
+ * eliminate the possibility of unaccounted-for errors (such as
+ * getting a machine check in the middle of a calibration step). The
+ * basic idea is for the slave to ask the master what itc value it has
+ * and to read its own itc before and after the master responds. Each
+ * iteration gives us three timestamps:
+ *
+ * slave master
+ *
+ * t0 ---\
+ * ---\
+ * --->
+ * tm
+ * /---
+ * /---
+ * t1 <---
+ *
+ *
+ * The goal is to adjust the slave's TSC such that tm falls exactly
+ * half-way between t0 and t1. If we achieve this, the clocks are
+ * synchronized provided the interconnect between the slave and the
+ * master is symmetric. Even if the interconnect were asymmetric, we
+ * would still know that the synchronization error is smaller than the
+ * roundtrip latency (t0 - t1).
+ *
+ * When the interconnect is quiet and symmetric, this lets us
+ * synchronize the TSC to within one or two cycles. However, we can
+ * only *guarantee* that the synchronization is accurate to within a
+ * round-trip time, which is typically in the range of several hundred
+ * cycles (e.g., ~500 cycles). In practice, this means that the TSCs
+ * are usually almost perfectly synchronized, but we shouldn't assume
+ * that the accuracy is much better than half a micro second or so.
*
- * We first check whether all CPUs have their TSC's synchronized,
- * then we print a warning if not, and always resync.
+ * [there are other errors like the latency of RDTSC and of the
+ * WRMSR. These can also account to hundreds of cycles. So it's
+ * probably worse. It claims 153 cycles error on a dual Opteron,
+ * but I suspect the numbers are actually somewhat worse -AK]
*/
-static atomic_t tsc_start_flag = ATOMIC_INIT(0);
-static atomic_t tsc_count_start = ATOMIC_INIT(0);
-static atomic_t tsc_count_stop = ATOMIC_INIT(0);
-static unsigned long long tsc_values[NR_CPUS];
+#define MASTER 0
+#define SLAVE (SMP_CACHE_BYTES/8)
-#define NR_LOOPS 5
+/* Intentionally don't use cpu_relax() while TSC synchronization
+ because we don't want to go into funky power save modi or cause
+ hypervisors to schedule us away. Going to sleep would likely affect
+ latency and low latency is the primary objective here. -AK */
+#define no_cpu_relax() barrier()
-extern unsigned int fast_gettimeoffset_quotient;
+static __cpuinitdata DEFINE_SPINLOCK(tsc_sync_lock);
+static volatile __cpuinitdata unsigned long go[SLAVE + 1];
+static int notscsync __cpuinitdata;
-static void __init synchronize_tsc_bp (void)
+#undef DEBUG_TSC_SYNC
+
+#define NUM_ROUNDS 64 /* magic value */
+#define NUM_ITERS 5 /* likewise */
+
+/* Callback on boot CPU */
+static __cpuinit void sync_master(void *arg)
{
+ unsigned long flags, i;
+
+ go[MASTER] = 0;
+
+ local_irq_save(flags);
+ {
+ for (i = 0; i < NUM_ROUNDS*NUM_ITERS; ++i) {
+ while (!go[MASTER])
+ no_cpu_relax();
+ go[MASTER] = 0;
+ rdtscll(go[SLAVE]);
+ }
+ }
+ local_irq_restore(flags);
+}
+
+/*
+ * Return the number of cycles by which our tsc differs from the tsc
+ * on the master (time-keeper) CPU. A positive number indicates our
+ * tsc is ahead of the master, negative that it is behind.
+ */
+static inline long
+get_delta(long *rt, long *master)
+{
+ unsigned long best_t0 = 0, best_t1 = ~0UL, best_tm = 0;
+ unsigned long tcenter, t0, t1, tm;
int i;
- unsigned long long t0;
- unsigned long long sum, avg;
- long long delta;
- long one_usec;
- int buggy = 0;
- printk(KERN_INFO "checking TSC synchronization across %u CPUs: ",num_booting_cpus());
+ for (i = 0; i < NUM_ITERS; ++i) {
+ rdtscll(t0);
+ go[MASTER] = 1;
+ while (!(tm = go[SLAVE]))
+ no_cpu_relax();
+ go[SLAVE] = 0;
+ rdtscll(t1);
- one_usec = cpu_khz;
+ if (t1 - t0 < best_t1 - best_t0)
+ best_t0 = t0, best_t1 = t1, best_tm = tm;
+ }
- atomic_set(&tsc_start_flag, 1);
- wmb();
+ *rt = best_t1 - best_t0;
+ *master = best_tm - best_t0;
- /*
- * We loop a few times to get a primed instruction cache,
- * then the last pass is more or less synchronized and
- * the BP and APs set their cycle counters to zero all at
- * once. This reduces the chance of having random offsets
- * between the processors, and guarantees that the maximum
- * delay between the cycle counters is never bigger than
- * the latency of information-passing (cachelines) between
- * two CPUs.
- */
- for (i = 0; i < NR_LOOPS; i++) {
- /*
- * all APs synchronize but they loop on '== num_cpus'
- */
- while (atomic_read(&tsc_count_start) != num_booting_cpus()-1) mb();
- atomic_set(&tsc_count_stop, 0);
- wmb();
- /*
- * this lets the APs save their current TSC:
- */
- atomic_inc(&tsc_count_start);
+ /* average best_t0 and best_t1 without overflow: */
+ tcenter = (best_t0/2 + best_t1/2);
+ if (best_t0 % 2 + best_t1 % 2 == 2)
+ ++tcenter;
+ return tcenter - best_tm;
+}
- sync_core();
- rdtscll(tsc_values[smp_processor_id()]);
- /*
- * We clear the TSC in the last loop:
- */
- if (i == NR_LOOPS-1)
- write_tsc(0, 0);
+static __cpuinit void sync_tsc(unsigned int master)
+{
+ int i, done = 0;
+ long delta, adj, adjust_latency = 0;
+ unsigned long flags, rt, master_time_stamp, bound;
+#ifdef DEBUG_TSC_SYNC
+ static struct syncdebug {
+ long rt; /* roundtrip time */
+ long master; /* master's timestamp */
+ long diff; /* difference between midpoint and master's timestamp */
+ long lat; /* estimate of tsc adjustment latency */
+ } t[NUM_ROUNDS] __cpuinitdata;
+#endif
- /*
- * Wait for all APs to leave the synchronization point:
- */
- while (atomic_read(&tsc_count_stop) != num_booting_cpus()-1) mb();
- atomic_set(&tsc_count_start, 0);
- wmb();
- atomic_inc(&tsc_count_stop);
- }
+ printk(KERN_INFO "CPU %d: Syncing TSC to CPU %u.\n",
+ smp_processor_id(), master);
- sum = 0;
- for (i = 0; i < NR_CPUS; i++) {
- if (cpu_isset(i, cpu_callout_map)) {
- t0 = tsc_values[i];
- sum += t0;
- }
- }
- avg = sum / num_booting_cpus();
+ go[MASTER] = 1;
+
+ /* It is dangerous to broadcast IPI as cpus are coming up,
+ * as they may not be ready to accept them. So since
+ * we only need to send the ipi to the boot cpu direct
+ * the message, and avoid the race.
+ */
+ smp_call_function_single(master, sync_master, NULL, 1, 0);
- sum = 0;
- for (i = 0; i < NR_CPUS; i++) {
- if (!cpu_isset(i, cpu_callout_map))
- continue;
+ while (go[MASTER]) /* wait for master to be ready */
+ no_cpu_relax();
- delta = tsc_values[i] - avg;
- if (delta < 0)
- delta = -delta;
- /*
- * We report bigger than 2 microseconds clock differences.
- */
- if (delta > 2*one_usec) {
- long realdelta;
- if (!buggy) {
- buggy = 1;
- printk("\n");
+ spin_lock_irqsave(&tsc_sync_lock, flags);
+ {
+ for (i = 0; i < NUM_ROUNDS; ++i) {
+ delta = get_delta(&rt, &master_time_stamp);
+ if (delta == 0) {
+ done = 1; /* let's lock on to this... */
+ bound = rt;
}
- realdelta = delta / one_usec;
- if (tsc_values[i] < avg)
- realdelta = -realdelta;
- printk("BIOS BUG: CPU#%d improperly initialized, has %ld usecs TSC skew! FIXED.\n",
- i, realdelta);
- }
+ if (!done) {
+ unsigned long t;
+ if (i > 0) {
+ adjust_latency += -delta;
+ adj = -delta + adjust_latency/4;
+ } else
+ adj = -delta;
- sum += delta;
+ rdtscll(t);
+ wrmsrl(MSR_IA32_TSC, t + adj);
+ }
+#ifdef DEBUG_TSC_SYNC
+ t[i].rt = rt;
+ t[i].master = master_time_stamp;
+ t[i].diff = delta;
+ t[i].lat = adjust_latency/4;
+#endif
+ }
}
- if (!buggy)
- printk("passed.\n");
+ spin_unlock_irqrestore(&tsc_sync_lock, flags);
+
+#ifdef DEBUG_TSC_SYNC
+ for (i = 0; i < NUM_ROUNDS; ++i)
+ printk("rt=%5ld master=%5ld diff=%5ld adjlat=%5ld\n",
+ t[i].rt, t[i].master, t[i].diff, t[i].lat);
+#endif
+
+ printk(KERN_INFO
+ "CPU %d: synchronized TSC with CPU %u (last diff %ld cycles, "
+ "maxerr %lu cycles)\n",
+ smp_processor_id(), master, delta, rt);
}
-static void __init synchronize_tsc_ap (void)
+static void __cpuinit tsc_sync_wait(void)
{
- int i;
-
/*
- * Not every cpu is online at the time
- * this gets called, so we first wait for the BP to
- * finish SMP initialization:
+ * 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.
*/
- while (!atomic_read(&tsc_start_flag)) mb();
-
- for (i = 0; i < NR_LOOPS; i++) {
- atomic_inc(&tsc_count_start);
- while (atomic_read(&tsc_count_start) != num_booting_cpus()) mb();
-
- sync_core();
- rdtscll(tsc_values[smp_processor_id()]);
- if (i == NR_LOOPS-1)
- write_tsc(0, 0);
+ if (notscsync || !cpu_has_tsc || !unsynchronized_tsc())
+ return;
+ sync_tsc(0);
+}
- atomic_inc(&tsc_count_stop);
- while (atomic_read(&tsc_count_stop) != num_booting_cpus()) mb();
- }
+static __init int notscsync_setup(char *s)
+{
+ notscsync = 1;
+ return 1;
}
-#undef NR_LOOPS
+__setup("notscsync", notscsync_setup);
-static atomic_t init_deasserted;
+static atomic_t init_deasserted __cpuinitdata;
-void __init smp_callin(void)
+/*
+ * Report back to the Boot Processor.
+ * Running on AP.
+ */
+void __cpuinit smp_callin(void)
{
int cpuid, phys_id;
unsigned long timeout;
* our local APIC. We have to wait for the IPI or we'll
* lock up on an APIC access.
*/
- while (!atomic_read(&init_deasserted));
+ while (!atomic_read(&init_deasserted))
+ cpu_relax();
/*
* (This works even if the APIC is not enabled.)
*/
if (cpu_isset(cpuid, cpu_callout_map))
break;
- rep_nop();
+ cpu_relax();
}
if (!time_before(jiffies, timeout)) {
Dprintk("CALLIN, before setup_local_APIC().\n");
setup_local_APIC();
- local_irq_enable();
-
/*
* Get our bogomips.
+ *
+ * Need to enable IRQs because it can take longer and then
+ * the NMI watchdog might kill us.
*/
+ local_irq_enable();
calibrate_delay();
+ local_irq_disable();
Dprintk("Stack at about %p\n",&cpuid);
disable_APIC_timer();
*/
smp_store_cpu_info(cpuid);
- local_irq_disable();
-
/*
* Allow the master to continue.
*/
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;
/*
- * Synchronize the TSC with the BP
+ * For perf, we return last level cache shared map.
+ * TBD: when power saving sched policy is added, we will return
+ * cpu_core_map when power saving policy is enabled
*/
- if (cpu_has_tsc)
- synchronize_tsc_ap();
+ return c->llc_shared_map;
}
-int cpucount;
+/* 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_mask(i, cpu_sibling_setup_map) {
+ if (phys_proc_id[cpu] == phys_proc_id[i] &&
+ cpu_core_id[cpu] == cpu_core_id[i]) {
+ 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]);
+ }
+
+ 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 (phys_proc_id[cpu] == phys_proc_id[i]) {
+ 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;
+ }
+ }
+}
/*
- * Activate a secondary processor.
+ * Setup code on secondary processor (after comming out of the trampoline)
*/
-void __init start_secondary(void)
+void __cpuinit start_secondary(void)
{
/*
* Dont put anything before smp_callin(), SMP
* 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 */
barrier();
- Dprintk("cpu %d: waiting for commence\n", smp_processor_id());
- while (!cpu_isset(smp_processor_id(), smp_commenced_mask))
- rep_nop();
-
Dprintk("cpu %d: setting up apic clock\n", smp_processor_id());
setup_secondary_APIC_clock();
- Dprintk("cpu %d: enabling apic timer\n", smp_processor_id());
+ Dprintk("cpu %d: enabling apic timer\n", smp_processor_id());
if (nmi_watchdog == NMI_IO_APIC) {
disable_8259A_irq(0);
enable_8259A_irq(0);
}
+ enable_APIC_timer();
- enable_APIC_timer();
+ /*
+ * The sibling maps must be set before turing the online map on for
+ * this cpu
+ */
+ set_cpu_sibling_map(smp_processor_id());
+
+ /*
+ * Wait for TSC sync to not schedule things before.
+ * We still process interrupts, which could see an inconsistent
+ * time in that window unfortunately.
+ * Do this here because TSC sync has global unprotected state.
+ */
+ tsc_sync_wait();
/*
- * low-memory mappings have been cleared, flush them from
- * the local TLBs too.
+ * We need to hold call_lock, so there is no inconsistency
+ * between the time smp_call_function() determines number of
+ * IPI receipients, and the time when the determination is made
+ * for which cpus receive the IPI in genapic_flat.c. Holding this
+ * lock helps us to not include this cpu in a currently in progress
+ * smp_call_function().
*/
- local_flush_tlb();
+ lock_ipi_call_lock();
- Dprintk("cpu %d eSetting cpu_online_map\n", smp_processor_id());
+ /*
+ * Allow the master to continue.
+ */
cpu_set(smp_processor_id(), cpu_online_map);
- wmb();
-
+ per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;
+ unlock_ipi_call_lock();
+
cpu_idle();
}
-extern volatile unsigned long init_rsp;
+extern volatile unsigned long init_rsp;
extern void (*initial_code)(void);
-static struct task_struct * __init fork_by_hand(void)
-{
- struct pt_regs regs;
- /*
- * don't care about the eip and regs settings since
- * we'll never reschedule the forked task.
- */
- return copy_process(CLONE_VM|CLONE_IDLETASK, 0, ®s, 0, NULL, NULL);
-}
-
-#if APIC_DEBUG
-static inline void inquire_remote_apic(int apicid)
+#ifdef APIC_DEBUG
+static void inquire_remote_apic(int apicid)
{
unsigned i, regs[] = { APIC_ID >> 4, APIC_LVR >> 4, APIC_SPIV >> 4 };
char *names[] = { "ID", "VERSION", "SPIV" };
*/
apic_wait_icr_idle();
- apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(apicid));
- apic_write_around(APIC_ICR, APIC_DM_REMRD | regs[i]);
+ apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(apicid));
+ apic_write(APIC_ICR, APIC_DM_REMRD | regs[i]);
timeout = 0;
do {
}
#endif
-static int __init wakeup_secondary_via_INIT(int phys_apicid, unsigned int start_rip)
+/*
+ * Kick the secondary to wake up.
+ */
+static int __cpuinit wakeup_secondary_via_INIT(int phys_apicid, unsigned int start_rip)
{
unsigned long send_status = 0, accept_status = 0;
int maxlvt, timeout, num_starts, j;
/*
* Turn INIT on target chip
*/
- apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));
+ apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));
/*
* Send IPI
*/
- apic_write_around(APIC_ICR, APIC_INT_LEVELTRIG | APIC_INT_ASSERT
+ apic_write(APIC_ICR, APIC_INT_LEVELTRIG | APIC_INT_ASSERT
| APIC_DM_INIT);
Dprintk("Waiting for send to finish...\n");
Dprintk("Deasserting INIT.\n");
/* Target chip */
- apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));
+ apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));
/* Send IPI */
- apic_write_around(APIC_ICR, APIC_INT_LEVELTRIG | APIC_DM_INIT);
+ apic_write(APIC_ICR, APIC_INT_LEVELTRIG | APIC_DM_INIT);
Dprintk("Waiting for send to finish...\n");
timeout = 0;
send_status = apic_read(APIC_ICR) & APIC_ICR_BUSY;
} while (send_status && (timeout++ < 1000));
+ mb();
atomic_set(&init_deasserted, 1);
- /*
- * Should we send STARTUP IPIs ?
- *
- * Determine this based on the APIC version.
- * If we don't have an integrated APIC, don't send the STARTUP IPIs.
- */
- if (APIC_INTEGRATED(apic_version[phys_apicid]))
- num_starts = 2;
- else
- num_starts = 0;
+ num_starts = 2;
/*
* Run STARTUP IPI loop.
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");
*/
/* Target chip */
- apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));
+ apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));
/* Boot on the stack */
/* Kick the second */
- apic_write_around(APIC_ICR, APIC_DM_STARTUP
- | (start_rip >> 12));
+ apic_write(APIC_ICR, APIC_DM_STARTUP | (start_rip >> 12));
/*
* Give the other CPU some time to accept the IPI.
* 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);
return (send_status | accept_status);
}
-static void __init do_boot_cpu (int apicid)
-{
+struct create_idle {
struct task_struct *idle;
+ struct completion done;
+ int cpu;
+};
+
+void do_fork_idle(void *_c_idle)
+{
+ struct create_idle *c_idle = _c_idle;
+
+ c_idle->idle = fork_idle(c_idle->cpu);
+ complete(&c_idle->done);
+}
+
+/*
+ * Boot one CPU.
+ */
+static int __cpuinit do_boot_cpu(int cpu, int apicid)
+{
unsigned long boot_error;
- int timeout, cpu;
+ int timeout;
unsigned long start_rip;
+ struct create_idle c_idle = {
+ .cpu = cpu,
+ .done = COMPLETION_INITIALIZER(c_idle.done),
+ };
+ 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;
+ }
- cpu = ++cpucount;
- /*
- * We can't use kernel_thread since we must avoid to
- * reschedule the child.
- */
- idle = fork_by_hand();
- if (IS_ERR(idle))
- panic("failed fork for CPU %d", cpu);
- wake_up_forked_process(idle);
- x86_cpu_to_apicid[cpu] = apicid;
+ /* 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);
+ }
+
+
+ c_idle.idle = get_idle_for_cpu(cpu);
+
+ if (c_idle.idle) {
+ c_idle.idle->thread.rsp = (unsigned long) (((struct pt_regs *)
+ (THREAD_SIZE + task_stack_page(c_idle.idle))) - 1);
+ init_idle(c_idle.idle, cpu);
+ goto do_rest;
+ }
/*
- * We remove it from the pidhash and the runqueue
- * once we got the process:
+ * During cold boot process, keventd thread is not spun up yet.
+ * When we do cpu hot-add, we create idle threads on the fly, we should
+ * not acquire any attributes from the calling context. Hence the clean
+ * way to create kernel_threads() is to do that from keventd().
+ * We do the current_is_keventd() due to the fact that ACPI notifier
+ * was also queuing to keventd() and when the caller is already running
+ * in context of keventd(), we would end up with locking up the keventd
+ * thread.
*/
- init_idle(idle,cpu);
+ if (!keventd_up() || current_is_keventd())
+ work.func(work.data);
+ else {
+ schedule_work(&work);
+ wait_for_completion(&c_idle.done);
+ }
+
+ if (IS_ERR(c_idle.idle)) {
+ printk("failed fork for CPU %d\n", cpu);
+ return PTR_ERR(c_idle.idle);
+ }
+
+ set_idle_for_cpu(cpu, c_idle.idle);
- unhash_process(idle);
+do_rest:
- cpu_pda[cpu].pcurrent = idle;
+ cpu_pda(cpu)->pcurrent = c_idle.idle;
start_rip = setup_trampoline();
- init_rsp = idle->thread.rsp;
- init_tss[cpu].rsp0 = init_rsp;
+ init_rsp = c_idle.idle->thread.rsp;
+ per_cpu(init_tss,cpu).rsp0 = init_rsp;
initial_code = start_secondary;
- clear_ti_thread_flag(idle->thread_info, TIF_FORK);
+ clear_tsk_thread_flag(c_idle.idle, TIF_FORK);
- printk(KERN_INFO "Booting processor %d/%d rip %lx rsp %lx\n", cpu, apicid,
- start_rip, init_rsp);
+ printk(KERN_INFO "Booting processor %d/%d APIC 0x%x\n", cpu,
+ cpus_weight(cpu_present_map),
+ apicid);
/*
* This grunge runs the startup process for
/*
* 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
/*
* Starting actual IPI sequence...
*/
- boot_error = wakeup_secondary_via_INIT(apicid, start_rip);
+ boot_error = wakeup_secondary_via_INIT(apicid, start_rip);
if (!boot_error) {
/*
if (cpu_isset(cpu, cpu_callin_map)) {
/* number CPUs logically, starting from 1 (BSP is 0) */
- Dprintk("OK.\n");
- print_cpu_info(&cpu_data[cpu]);
Dprintk("CPU has booted.\n");
} else {
boot_error = 1;
- if (*((volatile unsigned char *)phys_to_virt(8192))
+ if (*((volatile unsigned char *)phys_to_virt(SMP_TRAMPOLINE_BASE))
== 0xA5)
/* trampoline started but...? */
printk("Stuck ??\n");
else
/* trampoline code not run */
printk("Not responding.\n");
-#if APIC_DEBUG
+#ifdef APIC_DEBUG
inquire_remote_apic(apicid);
#endif
}
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() */
- cpucount--;
+ 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;
+ x86_cpu_to_log_apicid[cpu] = BAD_APICID;
+ return -EIO;
}
- /* mark "stuck" area as not stuck */
- *((volatile unsigned *)phys_to_virt(8192)) = 0;
+ return 0;
}
cycles_t cacheflush_time;
unsigned long cache_decay_ticks;
-static void smp_tune_scheduling (void)
+/*
+ * Cleanup possible dangling ends...
+ */
+static __cpuinit void smp_cleanup_boot(void)
{
- int cachesize; /* kB */
- unsigned long bandwidth = 1000; /* MB/s */
/*
- * Rough estimation for SMP scheduling, this is the number of
- * cycles it takes for a fully memory-limited process to flush
- * the SMP-local cache.
- *
- * (For a P5 this pretty much means we will choose another idle
- * CPU almost always at wakeup time (this is due to the small
- * L1 cache), on PIIs it's around 50-100 usecs, depending on
- * the cache size)
+ * Paranoid: Set warm reset code and vector here back
+ * to default values.
*/
+ CMOS_WRITE(0, 0xf);
- if (!cpu_khz) {
- /*
- * this basically disables processor-affinity
- * scheduling on SMP without a TSC.
- */
- cacheflush_time = 0;
- return;
- } else {
- cachesize = boot_cpu_data.x86_cache_size;
- if (cachesize == -1) {
- cachesize = 16; /* Pentiums, 2x8kB cache */
- bandwidth = 100;
- }
+ /*
+ * Reset trampoline flag
+ */
+ *((volatile int *) phys_to_virt(0x467)) = 0;
+}
- cacheflush_time = (cpu_khz>>10) * (cachesize<<10) / bandwidth;
- }
+/*
+ * Fall back to non SMP mode after errors.
+ *
+ * RED-PEN audit/test this more. I bet there is more state messed up here.
+ */
+static __init void disable_smp(void)
+{
+ cpu_present_map = cpumask_of_cpu(0);
+ cpu_possible_map = cpumask_of_cpu(0);
+ if (smp_found_config)
+ phys_cpu_present_map = physid_mask_of_physid(boot_cpu_id);
+ else
+ phys_cpu_present_map = physid_mask_of_physid(0);
+ cpu_set(0, cpu_sibling_map[0]);
+ cpu_set(0, cpu_core_map[0]);
+}
- cache_decay_ticks = (long)cacheflush_time/cpu_khz * HZ / 1000;
+#ifdef CONFIG_HOTPLUG_CPU
- printk(KERN_INFO "per-CPU timeslice cutoff: %ld.%02ld usecs.\n",
- (long)cacheflush_time/(cpu_khz/1000),
- ((long)cacheflush_time*100/(cpu_khz/1000)) % 100);
- printk(KERN_INFO "task migration cache decay timeout: %ld msecs.\n",
- (cache_decay_ticks + 1) * 1000 / HZ);
-}
+int additional_cpus __initdata = -1;
/*
- * Cycle through the processors sending APIC IPIs to boot each.
+ * cpu_possible_map should be static, it cannot change as cpu's
+ * are onlined, or offlined. The reason is per-cpu data-structures
+ * are allocated by some modules at init time, and dont expect to
+ * do this dynamically on cpu arrival/departure.
+ * cpu_present_map on the other hand can change dynamically.
+ * In case when cpu_hotplug is not compiled, then we resort to current
+ * behaviour, which is cpu_possible == cpu_present.
+ * - Ashok Raj
+ *
+ * Three ways to find out the number of additional hotplug CPUs:
+ * - If the BIOS specified disabled CPUs in ACPI/mptables use that.
+ * - 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
*/
-
-static void __init smp_boot_cpus(unsigned int max_cpus)
+__init void prefill_possible_map(void)
{
- unsigned apicid, cpu, bit, kicked;
+ int i;
+ int possible;
+
+ if (additional_cpus == -1) {
+ if (disabled_cpus > 0)
+ additional_cpus = disabled_cpus;
+ else
+ additional_cpus = 0;
+ }
+ possible = num_processors + additional_cpus;
+ if (possible > NR_CPUS)
+ possible = NR_CPUS;
+
+ if (possible == 0) { /* Could be SMP kernel on UP hw with broken BIOS */
+ possible = 1;
+ printk (KERN_DEBUG "BIOS never enumerated boot CPU, fixing.\n");
+ }
- nmi_watchdog_default();
+ printk(KERN_INFO "SMP: Allowing %d CPUs, %d hotplug CPUs\n",
+ possible,
+ max_t(int, possible - num_processors, 0));
- /*
- * Setup boot CPU information
- */
- smp_store_cpu_info(0); /* Final full version of the data */
- printk(KERN_INFO "CPU%d: ", 0);
- print_cpu_info(&cpu_data[0]);
-
- current_thread_info()->cpu = 0;
- smp_tune_scheduling();
+ for (i = 0; i < possible; i++)
+ cpu_set(i, cpu_possible_map);
+}
+#endif
+/*
+ * Various sanity checks.
+ */
+static int __init smp_sanity_check(unsigned max_cpus)
+{
if (!physid_isset(hard_smp_processor_id(), phys_cpu_present_map)) {
printk("weird, boot CPU (#%d) not listed by the BIOS.\n",
hard_smp_processor_id());
*/
if (!smp_found_config) {
printk(KERN_NOTICE "SMP motherboard not detected.\n");
- io_apic_irqs = 0;
- cpu_online_map = cpumask_of_cpu(0);
- phys_cpu_present_map = physid_mask_of_physid(0);
+ disable_smp();
if (APIC_init_uniprocessor())
printk(KERN_NOTICE "Local APIC not detected."
" Using dummy APIC emulation.\n");
- goto smp_done;
+ return -1;
}
/*
/*
* 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");
- io_apic_irqs = 0;
- cpu_online_map = cpumask_of_cpu(0);
- phys_cpu_present_map = physid_mask_of_physid(0);
- disable_apic = 1;
- goto smp_done;
+ nr_ioapics = 0;
+ return -1;
}
- verify_local_APIC();
-
/*
* If SMP should be disabled, then really disable it!
*/
if (!max_cpus) {
- smp_found_config = 0;
printk(KERN_INFO "SMP mode deactivated, forcing use of dummy APIC emulation.\n");
- io_apic_irqs = 0;
- cpu_online_map = cpumask_of_cpu(0);
- phys_cpu_present_map = physid_mask_of_physid(0);
- disable_apic = 1;
- goto smp_done;
+ nr_ioapics = 0;
+ return -1;
}
+ return 0;
+}
+
+/*
+ * Prepare for SMP bootup. The MP table or ACPI has been read
+ * earlier. Just do some sanity checking here and enable APIC mode.
+ */
+void __init smp_prepare_cpus(unsigned int max_cpus)
+{
+ 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");
+ disable_smp();
+ return;
+ }
+
+
+ /*
+ * Switch from PIC to APIC mode.
+ */
connect_bsp_APIC();
setup_local_APIC();
- if (GET_APIC_ID(apic_read(APIC_ID)) != boot_cpu_id)
- BUG();
+ if (GET_APIC_ID(apic_read(APIC_ID)) != boot_cpu_id) {
+ panic("Boot APIC ID in local APIC unexpected (%d vs %d)",
+ GET_APIC_ID(apic_read(APIC_ID)), boot_cpu_id);
+ /* Or can we switch back to PIC here? */
+ }
- x86_cpu_to_apicid[0] = boot_cpu_id;
+ /*
+ * Now start the IO-APICs
+ */
+ if (!skip_ioapic_setup && nr_ioapics)
+ setup_IO_APIC();
+ else
+ nr_ioapics = 0;
/*
- * Now scan the CPU present map and fire up the other CPUs.
+ * Set up local APIC timer on boot CPU.
*/
- Dprintk("CPU present map: %lx\n", physids_coerce(phys_cpu_present_map));
- kicked = 1;
- for (bit = 0; kicked < NR_CPUS && bit < MAX_APICS; bit++) {
- apicid = cpu_present_to_apicid(bit);
- /*
- * Don't even attempt to start the boot CPU!
- */
- if (apicid == boot_cpu_id || (apicid == BAD_APICID))
- continue;
+ setup_boot_APIC_clock();
+}
- if (!cpu_isset(apicid, phys_cpu_present_map))
- continue;
- if ((max_cpus >= 0) && (max_cpus <= cpucount+1))
- continue;
+/*
+ * Early setup to make printk work.
+ */
+void __init smp_prepare_boot_cpu(void)
+{
+ int me = smp_processor_id();
+ cpu_set(me, cpu_online_map);
+ cpu_set(me, cpu_callout_map);
+ per_cpu(cpu_state, me) = CPU_ONLINE;
+}
- do_boot_cpu(apicid);
- ++kicked;
- }
+/*
+ * Entry point to boot a CPU.
+ */
+int __cpuinit __cpu_up(unsigned int cpu)
+{
+ int err;
+ int apicid = cpu_present_to_apicid(cpu);
- /*
- * Cleanup possible dangling ends...
- */
- {
- /*
- * Install writable page 0 entry to set BIOS data area.
- */
- local_flush_tlb();
+ WARN_ON(irqs_disabled());
- /*
- * Paranoid: Set warm reset code and vector here back
- * to default values.
- */
- CMOS_WRITE(0, 0xf);
+ Dprintk("++++++++++++++++++++=_---CPU UP %u\n", cpu);
- *((volatile int *) phys_to_virt(0x467)) = 0;
+ if (apicid == BAD_APICID || apicid == boot_cpu_id ||
+ !physid_isset(apicid, phys_cpu_present_map)) {
+ printk("__cpu_up: bad cpu %d\n", cpu);
+ return -EINVAL;
}
/*
- * Allow the user to impress friends.
+ * Already booted CPU?
*/
+ if (cpu_isset(cpu, cpu_callin_map)) {
+ Dprintk("do_boot_cpu %d Already started\n", cpu);
+ return -ENOSYS;
+ }
- Dprintk("Before bogomips.\n");
- if (!cpucount) {
- printk(KERN_INFO "Only one processor found.\n");
- } else {
- unsigned long bogosum = 0;
- for (cpu = 0; cpu < NR_CPUS; cpu++)
- if (cpu_isset(cpu, cpu_callout_map))
- bogosum += cpu_data[cpu].loops_per_jiffy;
- printk(KERN_INFO "Total of %d processors activated (%lu.%02lu BogoMIPS).\n",
- cpucount+1,
- bogosum/(500000/HZ),
- (bogosum/(5000/HZ))%100);
- Dprintk("Before bogocount - setting activated=1.\n");
- }
-
- /*
- * If Hyper-Threading is avaialble, construct cpu_sibling_map[], so
- * that we can tell the sibling CPU efficiently.
- */
- if (cpu_has_ht && smp_num_siblings > 1) {
- for (cpu = 0; cpu < NR_CPUS; cpu++)
- cpu_sibling_map[cpu] = NO_PROC_ID;
-
- for (cpu = 0; cpu < NR_CPUS; cpu++) {
- int i;
- if (!cpu_isset(cpu, cpu_callout_map))
- continue;
-
- for (i = 0; i < NR_CPUS; i++) {
- if (i == cpu || !cpu_isset(i, cpu_callout_map))
- continue;
- if (phys_proc_id[cpu] == phys_proc_id[i]) {
- cpu_sibling_map[cpu] = i;
- break;
- }
- }
- if (cpu_sibling_map[cpu] == (char)NO_PROC_ID) {
- smp_num_siblings = 1;
- printk(KERN_WARNING "WARNING: No sibling found for CPU %d.\n", cpu);
- }
- }
+ per_cpu(cpu_state, cpu) = CPU_UP_PREPARE;
+ /* Boot it! */
+ err = do_boot_cpu(cpu, apicid);
+ if (err < 0) {
+ Dprintk("do_boot_cpu failed %d\n", err);
+ return err;
}
- Dprintk("Boot done.\n");
+ /* Unleash the CPU! */
+ Dprintk("waiting for cpu %d\n", cpu);
+
+ while (!cpu_isset(cpu, cpu_online_map))
+ cpu_relax();
+ err = 0;
- /*
- * Here we can be sure that there is an IO-APIC in the system. Let's
- * go and set it up:
- */
- if (!skip_ioapic_setup && nr_ioapics)
- setup_IO_APIC();
- else
- nr_ioapics = 0;
+ return err;
+}
- setup_boot_APIC_clock();
+/*
+ * Finish the SMP boot.
+ */
+void __init smp_cpus_done(unsigned int max_cpus)
+{
+ smp_cleanup_boot();
- /*
- * Synchronize the TSC with the AP
- */
- if (cpu_has_tsc && cpucount)
- synchronize_tsc_bp();
+#ifdef CONFIG_X86_IO_APIC
+ setup_ioapic_dest();
+#endif
- smp_done:
- time_init_smp();
+ check_nmi_watchdog();
}
-/* These are wrappers to interface to the new boot process. Someone
- who understands all this stuff should rewrite it properly. --RR 15/Jul/02 */
-void __init smp_prepare_cpus(unsigned int max_cpus)
+#ifdef CONFIG_HOTPLUG_CPU
+
+static void remove_siblinginfo(int cpu)
{
- smp_boot_cpus(max_cpus);
+ 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]);
+ cpus_clear(cpu_sibling_map[cpu]);
+ cpus_clear(cpu_core_map[cpu]);
+ phys_proc_id[cpu] = BAD_APICID;
+ cpu_core_id[cpu] = BAD_APICID;
+ cpu_clear(cpu, cpu_sibling_setup_map);
}
-void __devinit smp_prepare_boot_cpu(void)
+void remove_cpu_from_maps(void)
{
- cpu_set(smp_processor_id(), cpu_online_map);
- cpu_set(smp_processor_id(), cpu_callout_map);
+ int cpu = smp_processor_id();
+
+ 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 __devinit __cpu_up(unsigned int cpu)
+int __cpu_disable(void)
{
- /* This only works at boot for x86. See "rewrite" above. */
- if (cpu_isset(cpu, smp_commenced_mask)) {
- local_irq_enable();
- return -ENOSYS;
- }
+ int cpu = smp_processor_id();
- /* In case one didn't come up */
- if (!cpu_isset(cpu, cpu_callin_map)) {
- local_irq_enable();
- return -EIO;
- }
+ /*
+ * Perhaps use cpufreq to drop frequency, but that could go
+ * into generic code.
+ *
+ * We won't take down the boot processor on i386 due to some
+ * interrupts only being able to be serviced by the BSP.
+ * Especially so if we're not using an IOAPIC -zwane
+ */
+ if (cpu == 0)
+ return -EBUSY;
+
+ clear_local_APIC();
+
+ /*
+ * HACK:
+ * Allow any queued timer interrupts to get serviced
+ * This is only a temporary solution until we cleanup
+ * fixup_irqs as we do for IA64.
+ */
local_irq_enable();
+ mdelay(1);
- /* Unleash the CPU! */
- Dprintk("waiting for cpu %d\n", cpu);
+ local_irq_disable();
+ remove_siblinginfo(cpu);
- cpu_set(cpu, smp_commenced_mask);
- while (!cpu_isset(cpu, cpu_online_map))
- mb();
+ /* It's now safe to remove this processor from the online map */
+ cpu_clear(cpu, cpu_online_map);
+ remove_cpu_from_maps();
+ fixup_irqs(cpu_online_map);
return 0;
}
-void __init smp_cpus_done(unsigned int max_cpus)
+void __cpu_die(unsigned int cpu)
{
- zap_low_mappings();
+ /* We don't do anything here: idle task is faking death itself. */
+ unsigned int i;
+
+ for (i = 0; i < 10; i++) {
+ /* 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);
+ return;
+ }
+ msleep(100);
+ }
+ printk(KERN_ERR "CPU %u didn't die...\n", cpu);
}
+__init int setup_additional_cpus(char *s)
+{
+ return get_option(&s, &additional_cpus);
+}
+__setup("additional_cpus=", setup_additional_cpus);
+
+#else /* ... !CONFIG_HOTPLUG_CPU */
+
+int __cpu_disable(void)
+{
+ return -ENOSYS;
+}
+
+void __cpu_die(unsigned int cpu)
+{
+ /* We said "no" in __cpu_disable */
+ BUG();
+}
+#endif /* CONFIG_HOTPLUG_CPU */