aboutsummaryrefslogtreecommitdiff
path: root/arch/i386/kernel/smp.c
blob: 68be7d0c7238c1feaf4eff4856fac04f111ae0dd (plain)
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
32
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
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
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
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
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
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
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
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
/*
 *	Intel SMP support routines.
 *
 *	(c) 1995 Alan Cox, Building #3 <alan@redhat.com>
 *	(c) 1998-99, 2000 Ingo Molnar <mingo@redhat.com>
 *
 *	This code is released under the GNU General Public License version 2 or
 *	later.
 */

#include <linux/init.h>

#include <linux/mm.h>
#include <linux/irq.h>
#include <linux/delay.h>
#include <linux/spinlock.h>
#include <linux/smp_lock.h>
#include <linux/kernel_stat.h>
#include <linux/mc146818rtc.h>
#include <linux/cache.h>
#include <linux/interrupt.h>
#include <linux/module.h>

#include <asm/mtrr.h>
#include <asm/tlbflush.h>
#include <mach_apic.h>

/*
 *	Some notes on x86 processor bugs affecting SMP operation:
 *
 *	Pentium, Pentium Pro, II, III (and all CPUs) have bugs.
 *	The Linux implications for SMP are handled as follows:
 *
 *	Pentium III / [Xeon]
 *		None of the E1AP-E3AP errata are visible to the user.
 *
 *	E1AP.	see PII A1AP
 *	E2AP.	see PII A2AP
 *	E3AP.	see PII A3AP
 *
 *	Pentium II / [Xeon]
 *		None of the A1AP-A3AP errata are visible to the user.
 *
 *	A1AP.	see PPro 1AP
 *	A2AP.	see PPro 2AP
 *	A3AP.	see PPro 7AP
 *
 *	Pentium Pro
 *		None of 1AP-9AP errata are visible to the normal user,
 *	except occasional delivery of 'spurious interrupt' as trap #15.
 *	This is very rare and a non-problem.
 *
 *	1AP.	Linux maps APIC as non-cacheable
 *	2AP.	worked around in hardware
 *	3AP.	fixed in C0 and above steppings microcode update.
 *		Linux does not use excessive STARTUP_IPIs.
 *	4AP.	worked around in hardware
 *	5AP.	symmetric IO mode (normal Linux operation) not affected.
 *		'noapic' mode has vector 0xf filled out properly.
 *	6AP.	'noapic' mode might be affected - fixed in later steppings
 *	7AP.	We do not assume writes to the LVT deassering IRQs
 *	8AP.	We do not enable low power mode (deep sleep) during MP bootup
 *	9AP.	We do not use mixed mode
 *
 *	Pentium
 *		There is a marginal case where REP MOVS on 100MHz SMP
 *	machines with B stepping processors can fail. XXX should provide
 *	an L1cache=Writethrough or L1cache=off option.
 *
 *		B stepping CPUs may hang. There are hardware work arounds
 *	for this. We warn about it in case your board doesn't have the work
 *	arounds. Basically thats so I can tell anyone with a B stepping
 *	CPU and SMP problems "tough".
 *
 *	Specific items [From Pentium Processor Specification Update]
 *
 *	1AP.	Linux doesn't use remote read
 *	2AP.	Linux doesn't trust APIC errors
 *	3AP.	We work around this
 *	4AP.	Linux never generated 3 interrupts of the same priority
 *		to cause a lost local interrupt.
 *	5AP.	Remote read is never used
 *	6AP.	not affected - worked around in hardware
 *	7AP.	not affected - worked around in hardware
 *	8AP.	worked around in hardware - we get explicit CS errors if not
 *	9AP.	only 'noapic' mode affected. Might generate spurious
 *		interrupts, we log only the first one and count the
 *		rest silently.
 *	10AP.	not affected - worked around in hardware
 *	11AP.	Linux reads the APIC between writes to avoid this, as per
 *		the documentation. Make sure you preserve this as it affects
 *		the C stepping chips too.
 *	12AP.	not affected - worked around in hardware
 *	13AP.	not affected - worked around in hardware
 *	14AP.	we always deassert INIT during bootup
 *	15AP.	not affected - worked around in hardware
 *	16AP.	not affected - worked around in hardware
 *	17AP.	not affected - worked around in hardware
 *	18AP.	not affected - worked around in hardware
 *	19AP.	not affected - worked around in BIOS
 *
 *	If this sounds worrying believe me these bugs are either ___RARE___,
 *	or are signal timing bugs worked around in hardware and there's
 *	about nothing of note with C stepping upwards.
 */

DEFINE_PER_CPU(struct tlb_state, cpu_tlbstate) ____cacheline_aligned = { &init_mm, 0, };

/*
 * the following functions deal with sending IPIs between CPUs.
 *
 * We use 'broadcast', CPU->CPU IPIs and self-IPIs too.
 */

static inline int __prepare_ICR (unsigned int shortcut, int vector)
{
	return APIC_DM_FIXED | shortcut | vector | APIC_DEST_LOGICAL;
}

static inline int __prepare_ICR2 (unsigned int mask)
{
	return SET_APIC_DEST_FIELD(mask);
}

void __send_IPI_shortcut(unsigned int shortcut, int vector)
{
	/*
	 * Subtle. In the case of the 'never do double writes' workaround
	 * we have to lock out interrupts to be safe.  As we don't care
	 * of the value read we use an atomic rmw access to avoid costly
	 * cli/sti.  Otherwise we use an even cheaper single atomic write
	 * to the APIC.
	 */
	unsigned int cfg;

	/*
	 * Wait for idle.
	 */
	apic_wait_icr_idle();

	/*
	 * No need to touch the target chip field
	 */
	cfg = __prepare_ICR(shortcut, vector);

	/*
	 * Send the IPI. The write to APIC_ICR fires this off.
	 */
	apic_write_around(APIC_ICR, cfg);
}

void fastcall send_IPI_self(int vector)
{
	__send_IPI_shortcut(APIC_DEST_SELF, vector);
}

/*
 * This is only used on smaller machines.
 */
void send_IPI_mask_bitmask(cpumask_t cpumask, int vector)
{
	unsigned long mask = cpus_addr(cpumask)[0];
	unsigned long cfg;
	unsigned long flags;

	local_irq_save(flags);
		
	/*
	 * Wait for idle.
	 */
	apic_wait_icr_idle();
		
	/*
	 * prepare target chip field
	 */
	cfg = __prepare_ICR2(mask);
	apic_write_around(APIC_ICR2, cfg);
		
	/*
	 * program the ICR 
	 */
	cfg = __prepare_ICR(0, vector);
			
	/*
	 * Send the IPI. The write to APIC_ICR fires this off.
	 */
	apic_write_around(APIC_ICR, cfg);

	local_irq_restore(flags);
}

void send_IPI_mask_sequence(cpumask_t mask, int vector)
{
	unsigned long cfg, flags;
	unsigned int query_cpu;

	/*
	 * Hack. The clustered APIC addressing mode doesn't allow us to send 
	 * to an arbitrary mask, so I do a unicasts to each CPU instead. This 
	 * should be modified to do 1 message per cluster ID - mbligh
	 */ 

	local_irq_save(flags);

	for (query_cpu = 0; query_cpu < NR_CPUS; ++query_cpu) {
		if (cpu_isset(query_cpu, mask)) {
		
			/*
			 * Wait for idle.
			 */
			apic_wait_icr_idle();
		
			/*
			 * prepare target chip field
			 */
			cfg = __prepare_ICR2(cpu_to_logical_apicid(query_cpu));
			apic_write_around(APIC_ICR2, cfg);
		
			/*
			 * program the ICR 
			 */
			cfg = __prepare_ICR(0, vector);
			
			/*
			 * Send the IPI. The write to APIC_ICR fires this off.
			 */
			apic_write_around(APIC_ICR, cfg);
		}
	}
	local_irq_restore(flags);
}

#include <mach_ipi.h> /* must come after the send_IPI functions above for inlining */

/*
 *	Smarter SMP flushing macros. 
 *		c/o Linus Torvalds.
 *
 *	These mean you can really definitely utterly forget about
 *	writing to user space from interrupts. (Its not allowed anyway).
 *
 *	Optimizations Manfred Spraul <manfred@colorfullife.com>
 */

static cpumask_t flush_cpumask;
static struct mm_struct * flush_mm;
static unsigned long flush_va;
static DEFINE_SPINLOCK(tlbstate_lock);
#define FLUSH_ALL	0xffffffff

/*
 * We cannot call mmdrop() because we are in interrupt context, 
 * instead update mm->cpu_vm_mask.
 *
 * We need to reload %cr3 since the page tables may be going
 * away from under us..
 */
static inline void leave_mm (unsigned long cpu)
{
	if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK)
		BUG();
	cpu_clear(cpu, per_cpu(cpu_tlbstate, cpu).active_mm->cpu_vm_mask);
	load_cr3(swapper_pg_dir);
}

/*
 *
 * The flush IPI assumes that a thread switch happens in this order:
 * [cpu0: the cpu that switches]
 * 1) switch_mm() either 1a) or 1b)
 * 1a) thread switch to a different mm
 * 1a1) cpu_clear(cpu, old_mm->cpu_vm_mask);
 * 	Stop ipi delivery for the old mm. This is not synchronized with
 * 	the other cpus, but smp_invalidate_interrupt ignore flush ipis
 * 	for the wrong mm, and in the worst case we perform a superflous
 * 	tlb flush.
 * 1a2) set cpu_tlbstate to TLBSTATE_OK
 * 	Now the smp_invalidate_interrupt won't call leave_mm if cpu0
 *	was in lazy tlb mode.
 * 1a3) update cpu_tlbstate[].active_mm
 * 	Now cpu0 accepts tlb flushes for the new mm.
 * 1a4) cpu_set(cpu, new_mm->cpu_vm_mask);
 * 	Now the other cpus will send tlb flush ipis.
 * 1a4) change cr3.
 * 1b) thread switch without mm change
 *	cpu_tlbstate[].active_mm is correct, cpu0 already handles
 *	flush ipis.
 * 1b1) set cpu_tlbstate to TLBSTATE_OK
 * 1b2) test_and_set the cpu bit in cpu_vm_mask.
 * 	Atomically set the bit [other cpus will start sending flush ipis],
 * 	and test the bit.
 * 1b3) if the bit was 0: leave_mm was called, flush the tlb.
 * 2) switch %%esp, ie current
 *
 * The interrupt must handle 2 special cases:
 * - cr3 is changed before %%esp, ie. it cannot use current->{active_,}mm.
 * - the cpu performs speculative tlb reads, i.e. even if the cpu only
 *   runs in kernel space, the cpu could load tlb entries for user space
 *   pages.
 *
 * The good news is that cpu_tlbstate is local to each cpu, no
 * write/read ordering problems.
 */

/*
 * TLB flush IPI:
 *
 * 1) Flush the tlb entries if the cpu uses the mm that's being flushed.
 * 2) Leave the mm if we are in the lazy tlb mode.
 */

fastcall void smp_invalidate_interrupt(struct pt_regs *regs)
{
	unsigned long cpu;

	cpu = get_cpu();

	if (!cpu_isset(cpu, flush_cpumask))
		goto out;
		/* 
		 * This was a BUG() but until someone can quote me the
		 * line from the intel manual that guarantees an IPI to
		 * multiple CPUs is retried _only_ on the erroring CPUs
		 * its staying as a return
		 *
		 * BUG();
		 */
		 
	if (flush_mm == per_cpu(cpu_tlbstate, cpu).active_mm) {
		if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) {
			if (flush_va == FLUSH_ALL)
				local_flush_tlb();
			else
				__flush_tlb_one(flush_va);
		} else
			leave_mm(cpu);
	}
	ack_APIC_irq();
	smp_mb__before_clear_bit();
	cpu_clear(cpu, flush_cpumask);
	smp_mb__after_clear_bit();
out:
	put_cpu_no_resched();
}

static void flush_tlb_others(cpumask_t cpumask, struct mm_struct *mm,
						unsigned long va)
{
	cpumask_t tmp;
	/*
	 * A couple of (to be removed) sanity checks:
	 *
	 * - we do not send IPIs to not-yet booted CPUs.
	 * - current CPU must not be in mask
	 * - mask must exist :)
	 */
	BUG_ON(cpus_empty(cpumask));

	cpus_and(tmp, cpumask, cpu_online_map);
	BUG_ON(!cpus_equal(cpumask, tmp));
	BUG_ON(cpu_isset(smp_processor_id(), cpumask));
	BUG_ON(!mm);

	/*
	 * i'm not happy about this global shared spinlock in the
	 * MM hot path, but we'll see how contended it is.
	 * Temporarily this turns IRQs off, so that lockups are
	 * detected by the NMI watchdog.
	 */
	spin_lock(&tlbstate_lock);
	
	flush_mm = mm;
	flush_va = va;
#if NR_CPUS <= BITS_PER_LONG
	atomic_set_mask(cpumask, &flush_cpumask);
#else
	{
		int k;
		unsigned long *flush_mask = (unsigned long *)&flush_cpumask;
		unsigned long *cpu_mask = (unsigned long *)&cpumask;
		for (k = 0; k < BITS_TO_LONGS(NR_CPUS); ++k)
			atomic_set_mask(cpu_mask[k], &flush_mask[k]);
	}
#endif
	/*
	 * We have to send the IPI only to
	 * CPUs affected.
	 */
	send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR);

	while (!cpus_empty(flush_cpumask))
		/* nothing. lockup detection does not belong here */
		mb();

	flush_mm = NULL;
	flush_va = 0;
	spin_unlock(&tlbstate_lock);
}
	
void flush_tlb_current_task(void)
{
	struct mm_struct *mm = current->mm;
	cpumask_t cpu_mask;

	preempt_disable();
	cpu_mask = mm->cpu_vm_mask;
	cpu_clear(smp_processor_id(), cpu_mask);

	local_flush_tlb();
	if (!cpus_empty(cpu_mask))
		flush_tlb_others(cpu_mask, mm, FLUSH_ALL);
	preempt_enable();
}

void flush_tlb_mm (struct mm_struct * mm)
{
	cpumask_t cpu_mask;

	preempt_disable();
	cpu_mask = mm->cpu_vm_mask;
	cpu_clear(smp_processor_id(), cpu_mask);

	if (current->active_mm == mm) {
		if (current->mm)
			local_flush_tlb();
		else
			leave_mm(smp_processor_id());
	}
	if (!cpus_empty(cpu_mask))
		flush_tlb_others(cpu_mask, mm, FLUSH_ALL);

	preempt_enable();
}

void flush_tlb_page(struct vm_area_struct * vma, unsigned long va)
{
	struct mm_struct *mm = vma->vm_mm;
	cpumask_t cpu_mask;

	preempt_disable();
	cpu_mask = mm->cpu_vm_mask;
	cpu_clear(smp_processor_id(), cpu_mask);

	if (current->active_mm == mm) {
		if(current->mm)
			__flush_tlb_one(va);
		 else
		 	leave_mm(smp_processor_id());
	}

	if (!cpus_empty(cpu_mask))
		flush_tlb_others(cpu_mask, mm, va);

	preempt_enable();
}
EXPORT_SYMBOL(flush_tlb_page);

static void do_flush_tlb_all(void* info)
{
	unsigned long cpu = smp_processor_id();

	__flush_tlb_all();
	if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY)
		leave_mm(cpu);
}

void flush_tlb_all(void)
{
	on_each_cpu(do_flush_tlb_all, NULL, 1, 1);
}

/*
 * this function sends a 'reschedule' IPI to another CPU.
 * it goes straight through and wastes no time serializing
 * anything. Worst case is that we lose a reschedule ...
 */
void smp_send_reschedule(int cpu)
{
	send_IPI_mask(cpumask_of_cpu(cpu), RESCHEDULE_VECTOR);
}

/*
 * Structure and data for smp_call_function(). This is designed to minimise
 * static memory requirements. It also looks cleaner.
 */
static DEFINE_SPINLOCK(call_lock);

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

static struct call_data_struct * call_data;

/*
 * this function sends a 'generic call function' IPI to all other CPUs
 * in the system.
 */

int smp_call_function (void (*func) (void *info), void *info, int nonatomic,
			int wait)
/*
 * [SUMMARY] 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.
 * <nonatomic> currently unused.
 * <wait> If true, wait (atomically) 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
 * hardware interrupt handler or from a bottom half handler.
 */
{
	struct call_data_struct data;
	int cpus = num_online_cpus()-1;

	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(&call_lock);
	call_data = &data;
	mb();
	
	/* Send a message to all other CPUs and wait for them to respond */
	send_IPI_allbutself(CALL_FUNCTION_VECTOR);

	/* Wait for response */
	while (atomic_read(&data.started) != cpus)
		cpu_relax();

	if (wait)
		while (atomic_read(&data.finished) != cpus)
			cpu_relax();
	spin_unlock(&call_lock);

	return 0;
}
EXPORT_SYMBOL(smp_call_function);

static void stop_this_cpu (void * dummy)
{
	/*
	 * Remove this CPU:
	 */
	cpu_clear(smp_processor_id(), cpu_online_map);
	local_irq_disable();
	disable_local_APIC();
	if (cpu_data[smp_processor_id()].hlt_works_ok)
		for(;;) __asm__("hlt");
	for (;;);
}

/*
 * this function calls the 'stop' function on all other CPUs in the system.
 */

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

	local_irq_disable();
	disable_local_APIC();
	local_irq_enable();
}

/*
 * Reschedule call back. Nothing to do,
 * all the work is done automatically when
 * we return from the interrupt.
 */
fastcall void smp_reschedule_interrupt(struct pt_regs *regs)
{
	ack_APIC_irq();
}

fastcall void smp_call_function_interrupt(struct pt_regs *regs)
{
	void (*func) (void *info) = call_data->func;
	void *info = call_data->info;
	int wait = call_data->wait;

	ack_APIC_irq();
	/*
	 * 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);
	}
}