smpboot_64.c

linux 内核源代码

/*
 *	x86 SMP booting functions
 *
 *	(c) 1995 Alan Cox, Building #3 <[e-mail]>
 *	(c) 1998, 1999, 2000 Ingo Molnar <[e-mail]>
 *	Copyright 2001 Andi Kleen, SuSE Labs.
 *
 *	Much of the core SMP work is based on previous work by Thomas Radke, to
 *	whom a great many thanks are extended.
 *
 *	Thanks to Intel for making available several different Pentium,
 *	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
 *
 *	Fixes
 *		Felix Koop	:	NR_CPUS used properly
 *		Jose Renau	:	Handle single CPU case.
 *		Alan Cox	:	By repeated request 8) - Total BogoMIP report.
 *		Greg Wright	:	Fix for kernel stacks panic.
 *		Erich Boleyn	:	MP v1.4 and additional changes.
 *	Matthias Sattler	:	Changes for 2.1 kernel map.
 *	Michel Lespinasse	:	Changes for 2.1 kernel map.
 *	Michael Chastain	:	Change trampoline.S to gnu as.
 *		Alan Cox	:	Dumb bug: 'B' step PPro's are fine
 *		Ingo Molnar	:	Added APIC timers, based on code
 *					from Jose Renau
 *		Ingo Molnar	:	various cleanups and rewrites
 *		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.
 *      Andi Kleen              :       Converted to new state machine.
 *					Various cleanups.
 *					Probably mostly hotplug CPU ready now.
 *	Ashok Raj			: CPU hotplug support
 */


#include <linux/init.h>

#include <linux/mm.h>
#include <linux/kernel_stat.h>
#include <linux/bootmem.h>
#include <linux/thread_info.h>
#include <linux/module.h>
#include <linux/delay.h>
#include <linux/mc146818rtc.h>
#include <linux/smp.h>
#include <linux/kdebug.h>

#include <asm/mtrr.h>
#include <asm/pgalloc.h>
#include <asm/desc.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;
EXPORT_SYMBOL(smp_num_siblings);

/* Last level cache ID of each logical CPU */
DEFINE_PER_CPU(u8, cpu_llc_id) = BAD_APICID;

/* Bitmask of currently online CPUs */
cpumask_t cpu_online_map __read_mostly;

EXPORT_SYMBOL(cpu_online_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;
EXPORT_SYMBOL(cpu_callout_map);

cpumask_t cpu_possible_map;
EXPORT_SYMBOL(cpu_possible_map);

/* Per CPU bogomips and other parameters */
DEFINE_PER_CPU_SHARED_ALIGNED(struct cpuinfo_x86, cpu_info);
EXPORT_PER_CPU_SYMBOL(cpu_info);

/* Set when the idlers are all forked */
int smp_threads_ready;

/* representing HT siblings of each logical CPU */
DEFINE_PER_CPU(cpumask_t, cpu_sibling_map);
EXPORT_PER_CPU_SYMBOL(cpu_sibling_map);

/* representing HT and core siblings of each logical CPU */
DEFINE_PER_CPU(cpumask_t, cpu_core_map);
EXPORT_PER_CPU_SYMBOL(cpu_core_map);

/*
 * Trampoline 80x86 program as an array.
 */

extern const unsigned char trampoline_data[];
extern const 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
 * bootstrap into the page concerned. The caller
 * has made sure it's suitably aligned.
 */

static unsigned long __cpuinit setup_trampoline(void)
{
	void *tramp = __va(SMP_TRAMPOLINE_BASE); 
	memcpy(tramp, trampoline_data, trampoline_end - trampoline_data);
	return virt_to_phys(tramp);
}

/*
 * The bootstrap kernel entry code has set these up. Save them for
 * a given CPU
 */

static void __cpuinit smp_store_cpu_info(int id)
{
	struct cpuinfo_x86 *c = &cpu_data(id);

	*c = boot_cpu_data;
	c->cpu_index = id;
	identify_cpu(c);
	print_cpu_info(c);
}

static atomic_t init_deasserted __cpuinitdata;

/*
 * Report back to the Boot Processor.
 * Running on AP.
 */
void __cpuinit smp_callin(void)
{
	int cpuid, phys_id;
	unsigned long timeout;

	/*
	 * If waken up by an INIT in an 82489DX configuration
	 * we may get here before an INIT-deassert IPI reaches
	 * our local APIC.  We have to wait for the IPI or we'll
	 * lock up on an APIC access.
	 */
	while (!atomic_read(&init_deasserted))
		cpu_relax();

	/*
	 * (This works even if the APIC is not enabled.)
	 */
	phys_id = GET_APIC_ID(apic_read(APIC_ID));
	cpuid = smp_processor_id();
	if (cpu_isset(cpuid, cpu_callin_map)) {
		panic("smp_callin: phys CPU#%d, CPU#%d already present??\n",
					phys_id, cpuid);
	}
	Dprintk("CPU#%d (phys ID: %d) waiting for CALLOUT\n", cpuid, phys_id);

	/*
	 * STARTUP IPIs are fragile beasts as they might sometimes
	 * trigger some glue motherboard logic. Complete APIC bus
	 * silence for 1 second, this overestimates the time the
	 * boot CPU is spending to send the up to 2 STARTUP IPIs
	 * by a factor of two. This should be enough.
	 */

	/*
	 * Waiting 2s total for startup (udelay is not yet working)
	 */
	timeout = jiffies + 2*HZ;
	while (time_before(jiffies, timeout)) {
		/*
		 * Has the boot CPU finished it's STARTUP sequence?
		 */
		if (cpu_isset(cpuid, cpu_callout_map))
			break;
		cpu_relax();
	}

	if (!time_before(jiffies, timeout)) {
		panic("smp_callin: CPU%d started up but did not get a callout!\n",
			cpuid);
	}

	/*
	 * the boot CPU has finished the init stage and is spinning
	 * on callin_map until we finish. We are free to set up this
	 * CPU, first the APIC. (this is probably redundant on most
	 * boards)
	 */

	Dprintk("CALLIN, before setup_local_APIC().\n");
	setup_local_APIC();

	/*
	 * 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);

	/*
	 * Save our processor parameters
	 */
 	smp_store_cpu_info(cpuid);

	/*
	 * 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);
	/*
	 * 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 per_cpu(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);

	cpu_set(cpu, cpu_sibling_setup_map);

	if (smp_num_siblings > 1) {
		for_each_cpu_mask(i, cpu_sibling_setup_map) {
			if (c->phys_proc_id == cpu_data(i).phys_proc_id &&
			    c->cpu_core_id == cpu_data(i).cpu_core_id) {
				cpu_set(i, per_cpu(cpu_sibling_map, cpu));
				cpu_set(cpu, per_cpu(cpu_sibling_map, i));
				cpu_set(i, per_cpu(cpu_core_map, cpu));
				cpu_set(cpu, per_cpu(cpu_core_map, i));
				cpu_set(i, c->llc_shared_map);
				cpu_set(cpu, cpu_data(i).llc_shared_map);
			}
		}
	} else {
		cpu_set(cpu, per_cpu(cpu_sibling_map, cpu));
	}

	cpu_set(cpu, c->llc_shared_map);

	if (current_cpu_data.x86_max_cores == 1) {
		per_cpu(cpu_core_map, cpu) = per_cpu(cpu_sibling_map, cpu);
		c->booted_cores = 1;
		return;
	}

	for_each_cpu_mask(i, cpu_sibling_setup_map) {
		if (per_cpu(cpu_llc_id, cpu) != BAD_APICID &&
		    per_cpu(cpu_llc_id, cpu) == per_cpu(cpu_llc_id, i)) {
			cpu_set(i, c->llc_shared_map);
			cpu_set(cpu, cpu_data(i).llc_shared_map);
		}
		if (c->phys_proc_id == cpu_data(i).phys_proc_id) {
			cpu_set(i, per_cpu(cpu_core_map, cpu));
			cpu_set(cpu, per_cpu(cpu_core_map, i));
			/*
			 *  Does this new cpu bringup a new core?
			 */
			if (cpus_weight(per_cpu(cpu_sibling_map, cpu)) == 1) {
				/*
				 * for each core in package, increment
				 * the booted_cores for this new cpu
				 */
				if (first_cpu(per_cpu(cpu_sibling_map, i)) == i)
					c->booted_cores++;
				/*
				 * increment the core count for all
				 * the other cpus in this package
				 */
				if (i != cpu)
					cpu_data(i).booted_cores++;
			} else if (i != cpu && !c->booted_cores)
				c->booted_cores = cpu_data(i).booted_cores;
		}
	}
}

/*
 * Setup code on secondary processor (after comming out of the trampoline)
 */
void __cpuinit start_secondary(void)
{
	/*
	 * Dont put anything before smp_callin(), SMP
	 * booting is too fragile that we want to limit the
	 * 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();

	/*
  	 * Check TSC sync first:
 	 */
	check_tsc_sync_target();

	if (nmi_watchdog == NMI_IO_APIC) {
		disable_8259A_irq(0);
		enable_NMI_through_LVT0(NULL);
		enable_8259A_irq(0);
	}

	/*
	 * The sibling maps must be set before turing the online map on for
	 * this cpu
	 */
	set_cpu_sibling_map(smp_processor_id());

	/*
	 * We need to hold call_lock, so there is no inconsistency
	 * between the time smp_call_function() determines number of
	 * IPI recipients, 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().
	 */
	lock_ipi_call_lock();
	spin_lock(&vector_lock);

	/* Setup the per cpu irq handling data structures */
	__setup_vector_irq(smp_processor_id());
	/*
	 * Allow the master to continue.
	 */
	cpu_set(smp_processor_id(), cpu_online_map);
	per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;
	spin_unlock(&vector_lock);

	unlock_ipi_call_lock();

	setup_secondary_APIC_clock();

	cpu_idle();
}

extern volatile unsigned long init_rsp;
extern void (*initial_code)(void);

#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" };
	int timeout;
	unsigned int status;

	printk(KERN_INFO "Inquiring remote APIC #%d...\n", apicid);

	for (i = 0; i < ARRAY_SIZE(regs); i++) {
		printk("... APIC #%d %s: ", apicid, names[i]);

		/*
		 * Wait for idle.
		 */
		status = safe_apic_wait_icr_idle();
		if (status)
			printk("a previous APIC delivery may have failed\n");

		apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(apicid));
		apic_write(APIC_ICR, APIC_DM_REMRD | regs[i]);

		timeout = 0;
		do {
			udelay(100);
			status = apic_read(APIC_ICR) & APIC_ICR_RR_MASK;
		} while (status == APIC_ICR_RR_INPROG && timeout++ < 1000);

		switch (status) {
		case APIC_ICR_RR_VALID:
			status = apic_read(APIC_RRR);
			printk("%08x\n", status);
			break;
		default:
			printk("failed\n");
		}
	}
}
#endif

/*
 * Kick the secondary to wake up.
 */
static int __cpuinit wakeup_secondary_via_INIT(int phys_apicid, unsigned int start_rip)
{
	unsigned long send_status, accept_status = 0;
	int maxlvt, num_starts, j;

	Dprintk("Asserting INIT.\n");

	/*
	 * Turn INIT on target chip
	 */
	apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));

	/*
	 * Send IPI
	 */
	apic_write(APIC_ICR, APIC_INT_LEVELTRIG | APIC_INT_ASSERT
				| APIC_DM_INIT);

	Dprintk("Waiting for send to finish...\n");
	send_status = safe_apic_wait_icr_idle();

	mdelay(10);

	Dprintk("Deasserting INIT.\n");

	/* Target chip */
	apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));

	/* Send IPI */
	apic_write(APIC_ICR, APIC_INT_LEVELTRIG | APIC_DM_INIT);

	Dprintk("Waiting for send to finish...\n");
	send_status = safe_apic_wait_icr_idle();

	mb();
	atomic_set(&init_deasserted, 1);

	num_starts = 2;

	/*
	 * Run STARTUP IPI loop.
	 */
	Dprintk("#startup loops: %d.\n", num_starts);

	maxlvt = get_maxlvt();

	for (j = 1; j <= num_starts; j++) {
		Dprintk("Sending STARTUP #%d.\n",j);
		apic_write(APIC_ESR, 0);
		apic_read(APIC_ESR);
		Dprintk("After apic_write.\n");

		/*
		 * STARTUP IPI
		 */

		/* Target chip */
		apic_write(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid));

		/* Boot on the stack */
		/* Kick the second */
		apic_write(APIC_ICR, APIC_DM_STARTUP | (start_rip >> 12));

		/*
		 * Give the other CPU some time to accept the IPI.
		 */
		udelay(300);

		Dprintk("Startup point 1.\n");

		Dprintk("Waiting for send to finish...\n");
		send_status = safe_apic_wait_icr_idle();

		/*
		 * Give the other CPU some time to accept the IPI.
		 */
		udelay(200);
		/*
		 * Due to the Pentium erratum 3AP.
		 */
		if (maxlvt > 3) {
			apic_write(APIC_ESR, 0);
		}
		accept_status = (apic_read(APIC_ESR) & 0xEF);
		if (send_status || accept_status)
			break;
	}
	Dprintk("After Startup.\n");

	if (send_status)
		printk(KERN_ERR "APIC never delivered???\n");
	if (accept_status)
		printk(KERN_ERR "APIC delivery error (%lx).\n", accept_status);

	return (send_status | accept_status);
}

struct create_idle {
	struct work_struct work;
	struct task_struct *idle;
	struct completion done;
	int cpu;
};

static void __cpuinit do_fork_idle(struct work_struct *work)
{
	struct create_idle *c_idle =
		container_of(work, struct create_idle, work);

	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;
	unsigned long start_rip;
	struct create_idle c_idle = {
		.work = __WORK_INITIALIZER(c_idle.work, do_fork_idle),
		.cpu = cpu,
		.done = COMPLETION_INITIALIZER_ONSTACK(c_idle.done),
	};