smp.c

linux-2.6.15.6

/* smp.c: Sparc64 SMP support.
 *
 * Copyright (C) 1997 David S. Miller ([e-mail])
 */

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/threads.h>
#include <linux/smp.h>
#include <linux/smp_lock.h>
#include <linux/interrupt.h>
#include <linux/kernel_stat.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/fs.h>
#include <linux/seq_file.h>
#include <linux/cache.h>
#include <linux/jiffies.h>
#include <linux/profile.h>
#include <linux/bootmem.h>

#include <asm/head.h>
#include <asm/ptrace.h>
#include <asm/atomic.h>
#include <asm/tlbflush.h>
#include <asm/mmu_context.h>
#include <asm/cpudata.h>

#include <asm/irq.h>
#include <asm/page.h>
#include <asm/pgtable.h>
#include <asm/oplib.h>
#include <asm/uaccess.h>
#include <asm/timer.h>
#include <asm/starfire.h>
#include <asm/tlb.h>

extern void calibrate_delay(void);

/* Please don't make this stuff initdata!!!  --DaveM */
static unsigned char boot_cpu_id;

cpumask_t cpu_online_map __read_mostly = CPU_MASK_NONE;
cpumask_t phys_cpu_present_map __read_mostly = CPU_MASK_NONE;
static cpumask_t smp_commenced_mask;
static cpumask_t cpu_callout_map;

void smp_info(struct seq_file *m)
{
	int i;
	
	seq_printf(m, "State:\n");
	for (i = 0; i < NR_CPUS; i++) {
		if (cpu_online(i))
			seq_printf(m,
				   "CPU%d:\t\tonline\n", i);
	}
}

void smp_bogo(struct seq_file *m)
{
	int i;
	
	for (i = 0; i < NR_CPUS; i++)
		if (cpu_online(i))
			seq_printf(m,
				   "Cpu%dBogo\t: %lu.%02lu\n"
				   "Cpu%dClkTck\t: %016lx\n",
				   i, cpu_data(i).udelay_val / (500000/HZ),
				   (cpu_data(i).udelay_val / (5000/HZ)) % 100,
				   i, cpu_data(i).clock_tick);
}

void __init smp_store_cpu_info(int id)
{
	int cpu_node;

	/* multiplier and counter set by
	   smp_setup_percpu_timer()  */
	cpu_data(id).udelay_val			= loops_per_jiffy;

	cpu_find_by_mid(id, &cpu_node);
	cpu_data(id).clock_tick = prom_getintdefault(cpu_node,
						     "clock-frequency", 0);

	cpu_data(id).pgcache_size		= 0;
	cpu_data(id).pte_cache[0]		= NULL;
	cpu_data(id).pte_cache[1]		= NULL;
	cpu_data(id).pgd_cache			= NULL;
	cpu_data(id).idle_volume		= 1;

	cpu_data(id).dcache_size = prom_getintdefault(cpu_node, "dcache-size",
						      16 * 1024);
	cpu_data(id).dcache_line_size =
		prom_getintdefault(cpu_node, "dcache-line-size", 32);
	cpu_data(id).icache_size = prom_getintdefault(cpu_node, "icache-size",
						      16 * 1024);
	cpu_data(id).icache_line_size =
		prom_getintdefault(cpu_node, "icache-line-size", 32);
	cpu_data(id).ecache_size = prom_getintdefault(cpu_node, "ecache-size",
						      4 * 1024 * 1024);
	cpu_data(id).ecache_line_size =
		prom_getintdefault(cpu_node, "ecache-line-size", 64);
	printk("CPU[%d]: Caches "
	       "D[sz(%d):line_sz(%d)] "
	       "I[sz(%d):line_sz(%d)] "
	       "E[sz(%d):line_sz(%d)]\n",
	       id,
	       cpu_data(id).dcache_size, cpu_data(id).dcache_line_size,
	       cpu_data(id).icache_size, cpu_data(id).icache_line_size,
	       cpu_data(id).ecache_size, cpu_data(id).ecache_line_size);
}

static void smp_setup_percpu_timer(void);

static volatile unsigned long callin_flag = 0;

extern void inherit_locked_prom_mappings(int save_p);

static inline void cpu_setup_percpu_base(unsigned long cpu_id)
{
	__asm__ __volatile__("mov	%0, %%g5\n\t"
			     "stxa	%0, [%1] %2\n\t"
			     "membar	#Sync"
			     : /* no outputs */
			     : "r" (__per_cpu_offset(cpu_id)),
			       "r" (TSB_REG), "i" (ASI_IMMU));
}

void __init smp_callin(void)
{
	int cpuid = hard_smp_processor_id();

	inherit_locked_prom_mappings(0);

	__flush_tlb_all();

	cpu_setup_percpu_base(cpuid);

	smp_setup_percpu_timer();

	if (cheetah_pcache_forced_on)
		cheetah_enable_pcache();

	local_irq_enable();

	calibrate_delay();
	smp_store_cpu_info(cpuid);
	callin_flag = 1;
	__asm__ __volatile__("membar #Sync\n\t"
			     "flush  %%g6" : : : "memory");

	/* Clear this or we will die instantly when we
	 * schedule back to this idler...
	 */
	current_thread_info()->new_child = 0;

	/* Attach to the address space of init_task. */
	atomic_inc(&init_mm.mm_count);
	current->active_mm = &init_mm;

	while (!cpu_isset(cpuid, smp_commenced_mask))
		rmb();

	cpu_set(cpuid, cpu_online_map);

	/* idle thread is expected to have preempt disabled */
	preempt_disable();
}

void cpu_panic(void)
{
	printk("CPU[%d]: Returns from cpu_idle!\n", smp_processor_id());
	panic("SMP bolixed\n");
}

static unsigned long current_tick_offset __read_mostly;

/* This tick register synchronization scheme is taken entirely from
 * the ia64 port, see arch/ia64/kernel/smpboot.c for details and credit.
 *
 * The only change I've made is to rework it so that the master
 * initiates the synchonization instead of the slave. -DaveM
 */

#define MASTER	0
#define SLAVE	(SMP_CACHE_BYTES/sizeof(unsigned long))

#define NUM_ROUNDS	64	/* magic value */
#define NUM_ITERS	5	/* likewise */

static DEFINE_SPINLOCK(itc_sync_lock);
static unsigned long go[SLAVE + 1];

#define DEBUG_TICK_SYNC	0

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;
	unsigned long i;

	for (i = 0; i < NUM_ITERS; i++) {
		t0 = tick_ops->get_tick();
		go[MASTER] = 1;
		membar_storeload();
		while (!(tm = go[SLAVE]))
			rmb();
		go[SLAVE] = 0;
		wmb();
		t1 = tick_ops->get_tick();

		if (t1 - t0 < best_t1 - best_t0)
			best_t0 = t0, best_t1 = t1, best_tm = tm;
	}

	*rt = best_t1 - best_t0;
	*master = best_tm - best_t0;

	/* 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;
}

void smp_synchronize_tick_client(void)
{
	long i, delta, adj, adjust_latency = 0, done = 0;
	unsigned long flags, rt, master_time_stamp, bound;
#if DEBUG_TICK_SYNC
	struct {
		long rt;	/* roundtrip time */
		long master;	/* master's timestamp */
		long diff;	/* difference between midpoint and master's timestamp */
		long lat;	/* estimate of itc adjustment latency */
	} t[NUM_ROUNDS];
#endif

	go[MASTER] = 1;

	while (go[MASTER])
		rmb();

	local_irq_save(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;
			}

			if (!done) {
				if (i > 0) {
					adjust_latency += -delta;
					adj = -delta + adjust_latency/4;
				} else
					adj = -delta;

				tick_ops->add_tick(adj, current_tick_offset);
			}
#if DEBUG_TICK_SYNC
			t[i].rt = rt;
			t[i].master = master_time_stamp;
			t[i].diff = delta;
			t[i].lat = adjust_latency/4;
#endif
		}
	}
	local_irq_restore(flags);

#if DEBUG_TICK_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 TICK with master CPU (last diff %ld cycles,"
	       "maxerr %lu cycles)\n", smp_processor_id(), delta, rt);
}

static void smp_start_sync_tick_client(int cpu);

static void smp_synchronize_one_tick(int cpu)
{
	unsigned long flags, i;

	go[MASTER] = 0;

	smp_start_sync_tick_client(cpu);

	/* wait for client to be ready */
	while (!go[MASTER])
		rmb();

	/* now let the client proceed into his loop */
	go[MASTER] = 0;
	membar_storeload();

	spin_lock_irqsave(&itc_sync_lock, flags);
	{
		for (i = 0; i < NUM_ROUNDS*NUM_ITERS; i++) {
			while (!go[MASTER])
				rmb();
			go[MASTER] = 0;
			wmb();
			go[SLAVE] = tick_ops->get_tick();
			membar_storeload();
		}
	}
	spin_unlock_irqrestore(&itc_sync_lock, flags);
}

extern unsigned long sparc64_cpu_startup;

/* The OBP cpu startup callback truncates the 3rd arg cookie to
 * 32-bits (I think) so to be safe we have it read the pointer
 * contained here so we work on >4GB machines. -DaveM
 */
static struct thread_info *cpu_new_thread = NULL;

static int __devinit smp_boot_one_cpu(unsigned int cpu)
{
	unsigned long entry =
		(unsigned long)(&sparc64_cpu_startup);
	unsigned long cookie =
		(unsigned long)(&cpu_new_thread);
	struct task_struct *p;
	int timeout, ret, cpu_node;

	p = fork_idle(cpu);
	callin_flag = 0;
	cpu_new_thread = p->thread_info;
	cpu_set(cpu, cpu_callout_map);

	cpu_find_by_mid(cpu, &cpu_node);
	prom_startcpu(cpu_node, entry, cookie);

	for (timeout = 0; timeout < 5000000; timeout++) {
		if (callin_flag)
			break;
		udelay(100);
	}
	if (callin_flag) {
		ret = 0;
	} else {
		printk("Processor %d is stuck.\n", cpu);
		cpu_clear(cpu, cpu_callout_map);
		ret = -ENODEV;
	}
	cpu_new_thread = NULL;

	return ret;
}

static void spitfire_xcall_helper(u64 data0, u64 data1, u64 data2, u64 pstate, unsigned long cpu)
{
	u64 result, target;
	int stuck, tmp;

	if (this_is_starfire) {
		/* map to real upaid */
		cpu = (((cpu & 0x3c) << 1) |
			((cpu & 0x40) >> 4) |
			(cpu & 0x3));
	}

	target = (cpu << 14) | 0x70;
again:
	/* Ok, this is the real Spitfire Errata #54.
	 * One must read back from a UDB internal register
	 * after writes to the UDB interrupt dispatch, but
	 * before the membar Sync for that write.
	 * So we use the high UDB control register (ASI 0x7f,
	 * ADDR 0x20) for the dummy read. -DaveM
	 */
	tmp = 0x40;
	__asm__ __volatile__(
	"wrpr	%1, %2, %%pstate\n\t"
	"stxa	%4, [%0] %3\n\t"
	"stxa	%5, [%0+%8] %3\n\t"
	"add	%0, %8, %0\n\t"
	"stxa	%6, [%0+%8] %3\n\t"
	"membar	#Sync\n\t"
	"stxa	%%g0, [%7] %3\n\t"
	"membar	#Sync\n\t"
	"mov	0x20, %%g1\n\t"
	"ldxa	[%%g1] 0x7f, %%g0\n\t"
	"membar	#Sync"
	: "=r" (tmp)
	: "r" (pstate), "i" (PSTATE_IE), "i" (ASI_INTR_W),
	  "r" (data0), "r" (data1), "r" (data2), "r" (target),
	  "r" (0x10), "0" (tmp)
        : "g1");

	/* NOTE: PSTATE_IE is still clear. */
	stuck = 100000;
	do {
		__asm__ __volatile__("ldxa [%%g0] %1, %0"
			: "=r" (result)
			: "i" (ASI_INTR_DISPATCH_STAT));
		if (result == 0) {
			__asm__ __volatile__("wrpr %0, 0x0, %%pstate"
					     : : "r" (pstate));
			return;
		}
		stuck -= 1;
		if (stuck == 0)
			break;
	} while (result & 0x1);
	__asm__ __volatile__("wrpr %0, 0x0, %%pstate"
			     : : "r" (pstate));
	if (stuck == 0) {
		printk("CPU[%d]: mondo stuckage result[%016lx]\n",
		       smp_processor_id(), result);
	} else {
		udelay(2);
		goto again;
	}
}

static __inline__ void spitfire_xcall_deliver(u64 data0, u64 data1, u64 data2, cpumask_t mask)
{
	u64 pstate;
	int i;

	__asm__ __volatile__("rdpr %%pstate, %0" : "=r" (pstate));
	for_each_cpu_mask(i, mask)
		spitfire_xcall_helper(data0, data1, data2, pstate, i);
}

/* Cheetah now allows to send the whole 64-bytes of data in the interrupt
 * packet, but we have no use for that.  However we do take advantage of
 * the new pipelining feature (ie. dispatch to multiple cpus simultaneously).
 */
static void cheetah_xcall_deliver(u64 data0, u64 data1, u64 data2, cpumask_t mask)
{
	u64 pstate, ver;
	int nack_busy_id, is_jalapeno;

	if (cpus_empty(mask))
		return;

	/* Unfortunately, someone at Sun had the brilliant idea to make the
	 * busy/nack fields hard-coded by ITID number for this Ultra-III
	 * derivative processor.
	 */
	__asm__ ("rdpr %%ver, %0" : "=r" (ver));
	is_jalapeno = ((ver >> 32) == 0x003e0016);

	__asm__ __volatile__("rdpr %%pstate, %0" : "=r" (pstate));

retry:
	__asm__ __volatile__("wrpr %0, %1, %%pstate\n\t"
			     : : "r" (pstate), "i" (PSTATE_IE));

	/* Setup the dispatch data registers. */
	__asm__ __volatile__("stxa	%0, [%3] %6\n\t"
			     "stxa	%1, [%4] %6\n\t"
			     "stxa	%2, [%5] %6\n\t"
			     "membar	#Sync\n\t"
			     : /* no outputs */
			     : "r" (data0), "r" (data1), "r" (data2),
			       "r" (0x40), "r" (0x50), "r" (0x60),
			       "i" (ASI_INTR_W));

	nack_busy_id = 0;
	{
		int i;

		for_each_cpu_mask(i, mask) {
			u64 target = (i << 14) | 0x70;

			if (!is_jalapeno)
				target |= (nack_busy_id << 24);
			__asm__ __volatile__(
				"stxa	%%g0, [%0] %1\n\t"
				"membar	#Sync\n\t"
				: /* no outputs */
				: "r" (target), "i" (ASI_INTR_W));
			nack_busy_id++;
		}
	}

	/* Now, poll for completion. */
	{
		u64 dispatch_stat;
		long stuck;

		stuck = 100000 * nack_busy_id;
		do {
			__asm__ __volatile__("ldxa	[%%g0] %1, %0"
					     : "=r" (dispatch_stat)
					     : "i" (ASI_INTR_DISPATCH_STAT));
			if (dispatch_stat == 0UL) {
				__asm__ __volatile__("wrpr %0, 0x0, %%pstate"
						     : : "r" (pstate));
				return;
			}
			if (!--stuck)
				break;
		} while (dispatch_stat & 0x5555555555555555UL);

		__asm__ __volatile__("wrpr %0, 0x0, %%pstate"
				     : : "r" (pstate));

		if ((dispatch_stat & ~(0x5555555555555555UL)) == 0) {
			/* Busy bits will not clear, continue instead
			 * of freezing up on this cpu.
			 */
			printk("CPU[%d]: mondo stuckage result[%016lx]\n",
			       smp_processor_id(), dispatch_stat);
		} else {
			int i, this_busy_nack = 0;

			/* Delay some random time with interrupts enabled
			 * to prevent deadlock.
			 */
			udelay(2 * nack_busy_id);

			/* Clear out the mask bits for cpus which did not
			 * NACK us.
			 */
			for_each_cpu_mask(i, mask) {
				u64 check_mask;

				if (is_jalapeno)
					check_mask = (0x2UL << (2*i));
				else
					check_mask = (0x2UL <<
						      this_busy_nack);
				if ((dispatch_stat & check_mask) == 0)
					cpu_clear(i, mask);
				this_busy_nack += 2;
			}

			goto retry;
		}
	}
}

/* Send cross call to all processors mentioned in MASK
 * except self.
 */
static void smp_cross_call_masked(unsigned long *func, u32 ctx, u64 data1, u64 data2, cpumask_t mask)
{
	u64 data0 = (((u64)ctx)<<32 | (((u64)func) & 0xffffffff));
	int this_cpu = get_cpu();

	cpus_and(mask, mask, cpu_online_map);
	cpu_clear(this_cpu, mask);

	if (tlb_type == spitfire)
		spitfire_xcall_deliver(data0, data1, data2, mask);
	else
		cheetah_xcall_deliver(data0, data1, data2, mask);
	/* NOTE: Caller runs local copy on master. */

	put_cpu();
}

extern unsigned long xcall_sync_tick;

static void smp_start_sync_tick_client(int cpu)
{
	cpumask_t mask = cpumask_of_cpu(cpu);

	smp_cross_call_masked(&xcall_sync_tick,
			      0, 0, 0, mask);
}

/* Send cross call to all processors except self. */
#define smp_cross_call(func, ctx, data1, data2) \
	smp_cross_call_masked(func, ctx, data1, data2, cpu_online_map)

struct call_data_struct {
	void (*func) (void *info);
	void *info;
	atomic_t finished;
	int wait;
};

static DEFINE_SPINLOCK(call_lock);
static struct call_data_struct *call_data;

extern unsigned long xcall_call_function;

/*
 * You must not call this function with disabled interrupts or from a
 * hardware interrupt handler or from a bottom half handler.
 */
int smp_call_function(void (*func)(void *info), void *info,
		      int nonatomic, int wait)
{
	struct call_data_struct data;
	int cpus = num_online_cpus() - 1;
	long timeout;

	if (!cpus)
		return 0;

	/* Can deadlock when called with interrupts disabled */
	WARN_ON(irqs_disabled());