smp.c 11.1 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
/*
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version 2
 * of the License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
 *
 * Copyright (C) 2000, 2001 Kanoj Sarcar
 * Copyright (C) 2000, 2001 Ralf Baechle
 * Copyright (C) 2000, 2001 Silicon Graphics, Inc.
 * Copyright (C) 2000, 2001, 2003 Broadcom Corporation
 */
#include <linux/cache.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/spinlock.h>
#include <linux/threads.h>
#include <linux/module.h>
#include <linux/time.h>
#include <linux/timex.h>
#include <linux/sched.h>
#include <linux/cpumask.h>
R
Rojhalat Ibrahim 已提交
32
#include <linux/cpu.h>
L
Linus Torvalds 已提交
33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85

#include <asm/atomic.h>
#include <asm/cpu.h>
#include <asm/processor.h>
#include <asm/system.h>
#include <asm/mmu_context.h>
#include <asm/smp.h>

cpumask_t phys_cpu_present_map;		/* Bitmask of available CPUs */
volatile cpumask_t cpu_callin_map;	/* Bitmask of started secondaries */
cpumask_t cpu_online_map;		/* Bitmask of currently online CPUs */
int __cpu_number_map[NR_CPUS];		/* Map physical to logical */
int __cpu_logical_map[NR_CPUS];		/* Map logical to physical */

EXPORT_SYMBOL(phys_cpu_present_map);
EXPORT_SYMBOL(cpu_online_map);

static void smp_tune_scheduling (void)
{
	struct cache_desc *cd = &current_cpu_data.scache;
	unsigned long cachesize;       /* kB   */
	unsigned long cpu_khz;

	/*
	 * Crude estimate until we actually meassure ...
	 */
	cpu_khz = loops_per_jiffy * 2 * HZ / 1000;

	/*
	 * 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)
	 */
	if (!cpu_khz)
		return;

	cachesize = cd->linesz * cd->sets * cd->ways;
}

extern void __init calibrate_delay(void);
extern ATTRIB_NORET void cpu_idle(void);

/*
 * First C code run on the secondary CPUs after being started up by
 * the master.
 */
asmlinkage void start_secondary(void)
{
86
	unsigned int cpu;
L
Linus Torvalds 已提交
87 88 89 90 91 92 93 94 95 96 97 98

	cpu_probe();
	cpu_report();
	per_cpu_trap_init();
	prom_init_secondary();

	/*
	 * XXX parity protection should be folded in here when it's converted
	 * to an option instead of something based on .cputype
	 */

	calibrate_delay();
99 100
	preempt_disable();
	cpu = smp_processor_id();
L
Linus Torvalds 已提交
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
	cpu_data[cpu].udelay_val = loops_per_jiffy;

	prom_smp_finish();

	cpu_set(cpu, cpu_callin_map);

	cpu_idle();
}

DEFINE_SPINLOCK(smp_call_lock);

struct call_data_struct *call_data;

/*
 * Run a function on all other CPUs.
 *  <func>      The function to run. This must be fast and non-blocking.
 *  <info>      An arbitrary pointer to pass to the function.
 *  <retry>     If true, keep retrying until ready.
 *  <wait>      If true, wait until function has completed on other CPUs.
 *  [RETURNS]   0 on success, else a negative status code.
 *
 * Does not return until remote CPUs are nearly ready to execute <func>
 * or are or have executed.
 *
 * You must not call this function with disabled interrupts or from a
126 127 128 129 130 131 132 133 134 135 136 137 138
 * hardware interrupt handler or from a bottom half handler:
 *
 * CPU A                               CPU B
 * Disable interrupts
 *                                     smp_call_function()
 *                                     Take call_lock
 *                                     Send IPIs
 *                                     Wait for all cpus to acknowledge IPI
 *                                     CPU A has not responded, spin waiting
 *                                     for cpu A to respond, holding call_lock
 * smp_call_function()
 * Spin waiting for call_lock
 * Deadlock                            Deadlock
L
Linus Torvalds 已提交
139 140 141 142 143 144 145 146
 */
int smp_call_function (void (*func) (void *info), void *info, int retry,
								int wait)
{
	struct call_data_struct data;
	int i, cpus = num_online_cpus() - 1;
	int cpu = smp_processor_id();

147 148 149 150 151
	/*
	 * Can die spectacularly if this CPU isn't yet marked online
	 */
	BUG_ON(!cpu_online(cpu));

L
Linus Torvalds 已提交
152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
	if (!cpus)
		return 0;

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

	data.func = func;
	data.info = info;
	atomic_set(&data.started, 0);
	data.wait = wait;
	if (wait)
		atomic_set(&data.finished, 0);

	spin_lock(&smp_call_lock);
	call_data = &data;
	mb();

	/* Send a message to all other CPUs and wait for them to respond */
	for (i = 0; i < NR_CPUS; i++)
		if (cpu_online(i) && i != cpu)
			core_send_ipi(i, SMP_CALL_FUNCTION);

	/* Wait for response */
	/* FIXME: lock-up detection, backtrace on lock-up */
	while (atomic_read(&data.started) != cpus)
		barrier();

	if (wait)
		while (atomic_read(&data.finished) != cpus)
			barrier();
	spin_unlock(&smp_call_lock);

	return 0;
}

void smp_call_function_interrupt(void)
{
	void (*func) (void *info) = call_data->func;
	void *info = call_data->info;
	int wait = call_data->wait;

	/*
	 * Notify initiating CPU that I've grabbed the data and am
	 * about to execute the function.
	 */
	mb();
	atomic_inc(&call_data->started);

	/*
	 * At this point the info structure may be out of scope unless wait==1.
	 */
	irq_enter();
	(*func)(info);
	irq_exit();

	if (wait) {
		mb();
		atomic_inc(&call_data->finished);
	}
}

static void stop_this_cpu(void *dummy)
{
	/*
	 * Remove this CPU:
	 */
	cpu_clear(smp_processor_id(), cpu_online_map);
	local_irq_enable();	/* May need to service _machine_restart IPI */
	for (;;);		/* Wait if available. */
}

void smp_send_stop(void)
{
	smp_call_function(stop_this_cpu, NULL, 1, 0);
}

void __init smp_cpus_done(unsigned int max_cpus)
{
	prom_cpus_done();
}

/* called from main before smp_init() */
void __init smp_prepare_cpus(unsigned int max_cpus)
{
	init_new_context(current, &init_mm);
	current_thread_info()->cpu = 0;
	smp_tune_scheduling();
	prom_prepare_cpus(max_cpus);
}

/* preload SMP state for boot cpu */
void __devinit smp_prepare_boot_cpu(void)
{
	/*
	 * This assumes that bootup is always handled by the processor
	 * with the logic and physical number 0.
	 */
	__cpu_number_map[0] = 0;
	__cpu_logical_map[0] = 0;
	cpu_set(0, phys_cpu_present_map);
	cpu_set(0, cpu_online_map);
	cpu_set(0, cpu_callin_map);
}

/*
257 258 259
 * Called once for each "cpu_possible(cpu)".  Needs to spin up the cpu
 * and keep control until "cpu_online(cpu)" is set.  Note: cpu is
 * physical, not logical.
L
Linus Torvalds 已提交
260
 */
261
int __devinit __cpu_up(unsigned int cpu)
L
Linus Torvalds 已提交
262 263 264 265
{
	struct task_struct *idle;

	/*
266
	 * Processor goes to start_secondary(), sets online flag
L
Linus Torvalds 已提交
267 268 269 270 271
	 * The following code is purely to make sure
	 * Linux can schedule processes on this slave.
	 */
	idle = fork_idle(cpu);
	if (IS_ERR(idle))
272
		panic(KERN_ERR "Fork failed for CPU %d", cpu);
L
Linus Torvalds 已提交
273 274 275

	prom_boot_secondary(cpu, idle);

276 277 278
	/*
	 * Trust is futile.  We should really have timeouts ...
	 */
L
Linus Torvalds 已提交
279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
	while (!cpu_isset(cpu, cpu_callin_map))
		udelay(100);

	cpu_set(cpu, cpu_online_map);

	return 0;
}

/* Not really SMP stuff ... */
int setup_profiling_timer(unsigned int multiplier)
{
	return 0;
}

static void flush_tlb_all_ipi(void *info)
{
	local_flush_tlb_all();
}

void flush_tlb_all(void)
{
	on_each_cpu(flush_tlb_all_ipi, 0, 1, 1);
}

static void flush_tlb_mm_ipi(void *mm)
{
	local_flush_tlb_mm((struct mm_struct *)mm);
}

/*
 * The following tlb flush calls are invoked when old translations are
 * being torn down, or pte attributes are changing. For single threaded
 * address spaces, a new context is obtained on the current cpu, and tlb
 * context on other cpus are invalidated to force a new context allocation
 * at switch_mm time, should the mm ever be used on other cpus. For
 * multithreaded address spaces, intercpu interrupts have to be sent.
 * Another case where intercpu interrupts are required is when the target
 * mm might be active on another cpu (eg debuggers doing the flushes on
 * behalf of debugees, kswapd stealing pages from another process etc).
 * Kanoj 07/00.
 */

void flush_tlb_mm(struct mm_struct *mm)
{
	preempt_disable();

	if ((atomic_read(&mm->mm_users) != 1) || (current->mm != mm)) {
		smp_call_function(flush_tlb_mm_ipi, (void *)mm, 1, 1);
	} else {
		int i;
		for (i = 0; i < num_online_cpus(); i++)
			if (smp_processor_id() != i)
				cpu_context(i, mm) = 0;
	}
	local_flush_tlb_mm(mm);

	preempt_enable();
}

struct flush_tlb_data {
	struct vm_area_struct *vma;
	unsigned long addr1;
	unsigned long addr2;
};

static void flush_tlb_range_ipi(void *info)
{
	struct flush_tlb_data *fd = (struct flush_tlb_data *)info;

	local_flush_tlb_range(fd->vma, fd->addr1, fd->addr2);
}

void flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end)
{
	struct mm_struct *mm = vma->vm_mm;

	preempt_disable();
	if ((atomic_read(&mm->mm_users) != 1) || (current->mm != mm)) {
		struct flush_tlb_data fd;

		fd.vma = vma;
		fd.addr1 = start;
		fd.addr2 = end;
		smp_call_function(flush_tlb_range_ipi, (void *)&fd, 1, 1);
	} else {
		int i;
		for (i = 0; i < num_online_cpus(); i++)
			if (smp_processor_id() != i)
				cpu_context(i, mm) = 0;
	}
	local_flush_tlb_range(vma, start, end);
	preempt_enable();
}

static void flush_tlb_kernel_range_ipi(void *info)
{
	struct flush_tlb_data *fd = (struct flush_tlb_data *)info;

	local_flush_tlb_kernel_range(fd->addr1, fd->addr2);
}

void flush_tlb_kernel_range(unsigned long start, unsigned long end)
{
	struct flush_tlb_data fd;

	fd.addr1 = start;
	fd.addr2 = end;
	on_each_cpu(flush_tlb_kernel_range_ipi, (void *)&fd, 1, 1);
}

static void flush_tlb_page_ipi(void *info)
{
	struct flush_tlb_data *fd = (struct flush_tlb_data *)info;

	local_flush_tlb_page(fd->vma, fd->addr1);
}

void flush_tlb_page(struct vm_area_struct *vma, unsigned long page)
{
	preempt_disable();
	if ((atomic_read(&vma->vm_mm->mm_users) != 1) || (current->mm != vma->vm_mm)) {
		struct flush_tlb_data fd;

		fd.vma = vma;
		fd.addr1 = page;
		smp_call_function(flush_tlb_page_ipi, (void *)&fd, 1, 1);
	} else {
		int i;
		for (i = 0; i < num_online_cpus(); i++)
			if (smp_processor_id() != i)
				cpu_context(i, vma->vm_mm) = 0;
	}
	local_flush_tlb_page(vma, page);
	preempt_enable();
}

static void flush_tlb_one_ipi(void *info)
{
	unsigned long vaddr = (unsigned long) info;

	local_flush_tlb_one(vaddr);
}

void flush_tlb_one(unsigned long vaddr)
{
	smp_call_function(flush_tlb_one_ipi, (void *) vaddr, 1, 1);
	local_flush_tlb_one(vaddr);
}

R
Rojhalat Ibrahim 已提交
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446
static DEFINE_PER_CPU(struct cpu, cpu_devices);

static int __init topology_init(void)
{
	int cpu;
	int ret;

	for_each_cpu(cpu) {
		ret = register_cpu(&per_cpu(cpu_devices, cpu), cpu, NULL);
		if (ret)
			printk(KERN_WARNING "topology_init: register_cpu %d "
			       "failed (%d)\n", cpu, ret);
	}

	return 0;
}

subsys_initcall(topology_init);

L
Linus Torvalds 已提交
447 448 449 450
EXPORT_SYMBOL(flush_tlb_page);
EXPORT_SYMBOL(flush_tlb_one);
EXPORT_SYMBOL(cpu_data);
EXPORT_SYMBOL(synchronize_irq);