summaryrefslogtreecommitdiff
path: root/arch/metag/kernel/smp.c
blob: 7c01131429817ab58ce63490cc97e5dba4f4b4c9 (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
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
/*
 *  Copyright (C) 2009,2010,2011 Imagination Technologies Ltd.
 *
 *  Copyright (C) 2002 ARM Limited, All Rights Reserved.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */
#include <linux/atomic.h>
#include <linux/completion.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/sched.h>
#include <linux/interrupt.h>
#include <linux/cache.h>
#include <linux/profile.h>
#include <linux/errno.h>
#include <linux/mm.h>
#include <linux/err.h>
#include <linux/cpu.h>
#include <linux/smp.h>
#include <linux/seq_file.h>
#include <linux/irq.h>
#include <linux/bootmem.h>

#include <asm/cacheflush.h>
#include <asm/cachepart.h>
#include <asm/core_reg.h>
#include <asm/cpu.h>
#include <asm/global_lock.h>
#include <asm/metag_mem.h>
#include <asm/mmu_context.h>
#include <asm/pgtable.h>
#include <asm/pgalloc.h>
#include <asm/processor.h>
#include <asm/setup.h>
#include <asm/tlbflush.h>
#include <asm/hwthread.h>
#include <asm/traps.h>

#define SYSC_DCPART(n)	(SYSC_DCPART0 + SYSC_xCPARTn_STRIDE * (n))
#define SYSC_ICPART(n)	(SYSC_ICPART0 + SYSC_xCPARTn_STRIDE * (n))

DECLARE_PER_CPU(PTBI, pTBI);

void *secondary_data_stack;

/*
 * structures for inter-processor calls
 * - A collection of single bit ipi messages.
 */
struct ipi_data {
	spinlock_t lock;
	unsigned long ipi_count;
	unsigned long bits;
};

static DEFINE_PER_CPU(struct ipi_data, ipi_data) = {
	.lock	= __SPIN_LOCK_UNLOCKED(ipi_data.lock),
};

static DEFINE_SPINLOCK(boot_lock);

static DECLARE_COMPLETION(cpu_running);

/*
 * "thread" is assumed to be a valid Meta hardware thread ID.
 */
int boot_secondary(unsigned int thread, struct task_struct *idle)
{
	u32 val;

	/*
	 * set synchronisation state between this boot processor
	 * and the secondary one
	 */
	spin_lock(&boot_lock);

	core_reg_write(TXUPC_ID, 0, thread, (unsigned int)secondary_startup);
	core_reg_write(TXUPC_ID, 1, thread, 0);

	/*
	 * Give the thread privilege (PSTAT) and clear potentially problematic
	 * bits in the process (namely ISTAT, CBMarker, CBMarkerI, LSM_STEP).
	 */
	core_reg_write(TXUCT_ID, TXSTATUS_REGNUM, thread, TXSTATUS_PSTAT_BIT);

	/* Clear the minim enable bit. */
	val = core_reg_read(TXUCT_ID, TXPRIVEXT_REGNUM, thread);
	core_reg_write(TXUCT_ID, TXPRIVEXT_REGNUM, thread, val & ~0x80);

	/*
	 * set the ThreadEnable bit (0x1) in the TXENABLE register
	 * for the specified thread - off it goes!
	 */
	val = core_reg_read(TXUCT_ID, TXENABLE_REGNUM, thread);
	core_reg_write(TXUCT_ID, TXENABLE_REGNUM, thread, val | 0x1);

	/*
	 * now the secondary core is starting up let it run its
	 * calibrations, then wait for it to finish
	 */
	spin_unlock(&boot_lock);

	return 0;
}

/**
 * describe_cachepart_change: describe a change to cache partitions.
 * @thread:	Hardware thread number.
 * @label:	Label of cache type, e.g. "dcache" or "icache".
 * @sz:		Total size of the cache.
 * @old:	Old cache partition configuration (*CPART* register).
 * @new:	New cache partition configuration (*CPART* register).
 *
 * If the cache partition has changed, prints a message to the log describing
 * those changes.
 */
static void describe_cachepart_change(unsigned int thread, const char *label,
				      unsigned int sz, unsigned int old,
				      unsigned int new)
{
	unsigned int lor1, land1, gor1, gand1;
	unsigned int lor2, land2, gor2, gand2;
	unsigned int diff = old ^ new;

	if (!diff)
		return;

	pr_info("Thread %d: %s partition changed:", thread, label);
	if (diff & (SYSC_xCPARTL_OR_BITS | SYSC_xCPARTL_AND_BITS)) {
		lor1   = (old & SYSC_xCPARTL_OR_BITS)  >> SYSC_xCPARTL_OR_S;
		lor2   = (new & SYSC_xCPARTL_OR_BITS)  >> SYSC_xCPARTL_OR_S;
		land1  = (old & SYSC_xCPARTL_AND_BITS) >> SYSC_xCPARTL_AND_S;
		land2  = (new & SYSC_xCPARTL_AND_BITS) >> SYSC_xCPARTL_AND_S;
		pr_cont(" L:%#x+%#x->%#x+%#x",
			(lor1 * sz) >> 4,
			((land1 + 1) * sz) >> 4,
			(lor2 * sz) >> 4,
			((land2 + 1) * sz) >> 4);
	}
	if (diff & (SYSC_xCPARTG_OR_BITS | SYSC_xCPARTG_AND_BITS)) {
		gor1   = (old & SYSC_xCPARTG_OR_BITS)  >> SYSC_xCPARTG_OR_S;
		gor2   = (new & SYSC_xCPARTG_OR_BITS)  >> SYSC_xCPARTG_OR_S;
		gand1  = (old & SYSC_xCPARTG_AND_BITS) >> SYSC_xCPARTG_AND_S;
		gand2  = (new & SYSC_xCPARTG_AND_BITS) >> SYSC_xCPARTG_AND_S;
		pr_cont(" G:%#x+%#x->%#x+%#x",
			(gor1 * sz) >> 4,
			((gand1 + 1) * sz) >> 4,
			(gor2 * sz) >> 4,
			((gand2 + 1) * sz) >> 4);
	}
	if (diff & SYSC_CWRMODE_BIT)
		pr_cont(" %sWR",
			(new & SYSC_CWRMODE_BIT) ? "+" : "-");
	if (diff & SYSC_DCPART_GCON_BIT)
		pr_cont(" %sGCOn",
			(new & SYSC_DCPART_GCON_BIT) ? "+" : "-");
	pr_cont("\n");
}

/**
 * setup_smp_cache: ensure cache coherency for new SMP thread.
 * @thread:	New hardware thread number.
 *
 * Ensures that coherency is enabled and that the threads share the same cache
 * partitions.
 */
static void setup_smp_cache(unsigned int thread)
{
	unsigned int this_thread, lflags;
	unsigned int dcsz, dcpart_this, dcpart_old, dcpart_new;
	unsigned int icsz, icpart_old, icpart_new;

	/*
	 * Copy over the current thread's cache partition configuration to the
	 * new thread so that they share cache partitions.
	 */
	__global_lock2(lflags);
	this_thread = hard_processor_id();
	/* Share dcache partition */
	dcpart_this = metag_in32(SYSC_DCPART(this_thread));
	dcpart_old = metag_in32(SYSC_DCPART(thread));
	dcpart_new = dcpart_this;
#if PAGE_OFFSET < LINGLOBAL_BASE
	/*
	 * For the local data cache to be coherent the threads must also have
	 * GCOn enabled.
	 */
	dcpart_new |= SYSC_DCPART_GCON_BIT;
	metag_out32(dcpart_new, SYSC_DCPART(this_thread));
#endif
	metag_out32(dcpart_new, SYSC_DCPART(thread));
	/* Share icache partition too */
	icpart_new = metag_in32(SYSC_ICPART(this_thread));
	icpart_old = metag_in32(SYSC_ICPART(thread));
	metag_out32(icpart_new, SYSC_ICPART(thread));
	__global_unlock2(lflags);

	/*
	 * Log if the cache partitions were altered so the user is aware of any
	 * potential unintentional cache wastage.
	 */
	dcsz = get_dcache_size();
	icsz = get_dcache_size();
	describe_cachepart_change(this_thread, "dcache", dcsz,
				  dcpart_this, dcpart_new);
	describe_cachepart_change(thread, "dcache", dcsz,
				  dcpart_old, dcpart_new);
	describe_cachepart_change(thread, "icache", icsz,
				  icpart_old, icpart_new);
}

int __cpu_up(unsigned int cpu, struct task_struct *idle)
{
	unsigned int thread = cpu_2_hwthread_id[cpu];
	int ret;

	load_pgd(swapper_pg_dir, thread);

	flush_tlb_all();

	setup_smp_cache(thread);

	/*
	 * Tell the secondary CPU where to find its idle thread's stack.
	 */
	secondary_data_stack = task_stack_page(idle);

	wmb();

	/*
	 * Now bring the CPU into our world.
	 */
	ret = boot_secondary(thread, idle);
	if (ret == 0) {
		/*
		 * CPU was successfully started, wait for it
		 * to come online or time out.
		 */
		wait_for_completion_timeout(&cpu_running,
					    msecs_to_jiffies(1000));

		if (!cpu_online(cpu))
			ret = -EIO;
	}

	secondary_data_stack = NULL;

	if (ret) {
		pr_crit("CPU%u: processor failed to boot\n", cpu);

		/*
		 * FIXME: We need to clean up the new idle thread. --rmk
		 */
	}

	return ret;
}

#ifdef CONFIG_HOTPLUG_CPU
static DECLARE_COMPLETION(cpu_killed);

/*
 * __cpu_disable runs on the processor to be shutdown.
 */
int __cpu_disable(void)
{
	unsigned int cpu = smp_processor_id();

	/*
	 * Take this CPU offline.  Once we clear this, we can't return,
	 * and we must not schedule until we're ready to give up the cpu.
	 */
	set_cpu_online(cpu, false);

	/*
	 * OK - migrate IRQs away from this CPU
	 */
	migrate_irqs();

	/*
	 * Flush user cache and TLB mappings, and then remove this CPU
	 * from the vm mask set of all processes.
	 */
	flush_cache_all();
	local_flush_tlb_all();

	clear_tasks_mm_cpumask(cpu);

	return 0;
}

/*
 * called on the thread which is asking for a CPU to be shutdown -
 * waits until shutdown has completed, or it is timed out.
 */
void __cpu_die(unsigned int cpu)
{
	if (!wait_for_completion_timeout(&cpu_killed, msecs_to_jiffies(1)))
		pr_err("CPU%u: unable to kill\n", cpu);
}

/*
 * Called from the idle thread for the CPU which has been shutdown.
 *
 * Note that we do not return from this function. If this cpu is
 * brought online again it will need to run secondary_startup().
 */
void cpu_die(void)
{
	local_irq_disable();
	idle_task_exit();

	complete(&cpu_killed);

	asm ("XOR	TXENABLE, D0Re0,D0Re0\n");
}
#endif /* CONFIG_HOTPLUG_CPU */

/*
 * Called by both boot and secondaries to move global data into
 * per-processor storage.
 */
void smp_store_cpu_info(unsigned int cpuid)
{
	struct cpuinfo_metag *cpu_info = &per_cpu(cpu_data, cpuid);

	cpu_info->loops_per_jiffy = loops_per_jiffy;
}

/*
 * This is the secondary CPU boot entry.  We're using this CPUs
 * idle thread stack and the global page tables.
 */
asmlinkage void secondary_start_kernel(void)
{
	struct mm_struct *mm = &init_mm;
	unsigned int cpu = smp_processor_id();

	/*
	 * All kernel threads share the same mm context; grab a
	 * reference and switch to it.
	 */
	atomic_inc(&mm->mm_users);
	atomic_inc(&mm->mm_count);
	current->active_mm = mm;
	cpumask_set_cpu(cpu, mm_cpumask(mm));
	enter_lazy_tlb(mm, current);
	local_flush_tlb_all();

	/*
	 * TODO: Some day it might be useful for each Linux CPU to
	 * have its own TBI structure. That would allow each Linux CPU
	 * to run different interrupt handlers for the same IRQ
	 * number.
	 *
	 * For now, simply copying the pointer to the boot CPU's TBI
	 * structure is sufficient because we always want to run the
	 * same interrupt handler whatever CPU takes the interrupt.
	 */
	per_cpu(pTBI, cpu) = __TBI(TBID_ISTAT_BIT);

	if (!per_cpu(pTBI, cpu))
		panic("No TBI found!");

	per_cpu_trap_init(cpu);

	preempt_disable();

	setup_priv();

	notify_cpu_starting(cpu);

	pr_info("CPU%u (thread %u): Booted secondary processor\n",
		cpu, cpu_2_hwthread_id[cpu]);

	calibrate_delay();
	smp_store_cpu_info(cpu);

	/*
	 * OK, now it's safe to let the boot CPU continue
	 */
	set_cpu_online(cpu, true);
	complete(&cpu_running);

	/*
	 * Enable local interrupts.
	 */
	tbi_startup_interrupt(TBID_SIGNUM_TRT);
	local_irq_enable();

	/*
	 * OK, it's off to the idle thread for us
	 */
	cpu_startup_entry(CPUHP_ONLINE);
}

void __init smp_cpus_done(unsigned int max_cpus)
{
	int cpu;
	unsigned long bogosum = 0;

	for_each_online_cpu(cpu)
		bogosum += per_cpu(cpu_data, cpu).loops_per_jiffy;

	pr_info("SMP: Total of %d processors activated (%lu.%02lu BogoMIPS).\n",
		num_online_cpus(),
		bogosum / (500000/HZ),
		(bogosum / (5000/HZ)) % 100);
}

void __init smp_prepare_cpus(unsigned int max_cpus)
{
	unsigned int cpu = smp_processor_id();

	init_new_context(current, &init_mm);
	current_thread_info()->cpu = cpu;

	smp_store_cpu_info(cpu);
	init_cpu_present(cpu_possible_mask);
}

void __init smp_prepare_boot_cpu(void)
{
	unsigned int cpu = smp_processor_id();

	per_cpu(pTBI, cpu) = __TBI(TBID_ISTAT_BIT);

	if (!per_cpu(pTBI, cpu))
		panic("No TBI found!");
}

static void smp_cross_call(cpumask_t callmap, enum ipi_msg_type msg);

static void send_ipi_message(const struct cpumask *mask, enum ipi_msg_type msg)
{
	unsigned long flags;
	unsigned int cpu;
	cpumask_t map;

	cpumask_clear(&map);
	local_irq_save(flags);

	for_each_cpu(cpu, mask) {
		struct ipi_data *ipi = &per_cpu(ipi_data, cpu);

		spin_lock(&ipi->lock);

		/*
		 * KICK interrupts are queued in hardware so we'll get
		 * multiple interrupts if we call smp_cross_call()
		 * multiple times for one msg. The problem is that we
		 * only have one bit for each message - we can't queue
		 * them in software.
		 *
		 * The first time through ipi_handler() we'll clear
		 * the msg bit, having done all the work. But when we
		 * return we'll get _another_ interrupt (and another,
		 * and another until we've handled all the queued
		 * KICKs). Running ipi_handler() when there's no work
		 * to do is bad because that's how kick handler
		 * chaining detects who the KICK was intended for.
		 * See arch/metag/kernel/kick.c for more details.
		 *
		 * So only add 'cpu' to 'map' if we haven't already
		 * queued a KICK interrupt for 'msg'.
		 */
		if (!(ipi->bits & (1 << msg))) {
			ipi->bits |= 1 << msg;
			cpumask_set_cpu(cpu, &map);
		}

		spin_unlock(&ipi->lock);
	}

	/*
	 * Call the platform specific cross-CPU call function.
	 */
	smp_cross_call(map, msg);

	local_irq_restore(flags);
}

void arch_send_call_function_ipi_mask(const struct cpumask *mask)
{
	send_ipi_message(mask, IPI_CALL_FUNC);
}

void arch_send_call_function_single_ipi(int cpu)
{
	send_ipi_message(cpumask_of(cpu), IPI_CALL_FUNC_SINGLE);
}

void show_ipi_list(struct seq_file *p)
{
	unsigned int cpu;

	seq_puts(p, "IPI:");

	for_each_present_cpu(cpu)
		seq_printf(p, " %10lu", per_cpu(ipi_data, cpu).ipi_count);

	seq_putc(p, '\n');
}

static DEFINE_SPINLOCK(stop_lock);

/*
 * Main handler for inter-processor interrupts
 *
 * For Meta, the ipimask now only identifies a single
 * category of IPI (Bit 1 IPIs have been replaced by a
 * different mechanism):
 *
 *  Bit 0 - Inter-processor function call
 */
static int do_IPI(struct pt_regs *regs)
{
	unsigned int cpu = smp_processor_id();
	struct ipi_data *ipi = &per_cpu(ipi_data, cpu);
	struct pt_regs *old_regs = set_irq_regs(regs);
	unsigned long msgs, nextmsg;
	int handled = 0;

	ipi->ipi_count++;

	spin_lock(&ipi->lock);
	msgs = ipi->bits;
	nextmsg = msgs & -msgs;
	ipi->bits &= ~nextmsg;
	spin_unlock(&ipi->lock);

	if (nextmsg) {
		handled = 1;

		nextmsg = ffz(~nextmsg);
		switch (nextmsg) {
		case IPI_RESCHEDULE:
			scheduler_ipi();
			break;

		case IPI_CALL_FUNC:
			generic_smp_call_function_interrupt();
			break;

		case IPI_CALL_FUNC_SINGLE:
			generic_smp_call_function_single_interrupt();
			break;

		default:
			pr_crit("CPU%u: Unknown IPI message 0x%lx\n",
				cpu, nextmsg);
			break;
		}
	}

	set_irq_regs(old_regs);

	return handled;
}

void smp_send_reschedule(int cpu)
{
	send_ipi_message(cpumask_of(cpu), IPI_RESCHEDULE);
}

static void stop_this_cpu(void *data)
{
	unsigned int cpu = smp_processor_id();

	if (system_state == SYSTEM_BOOTING ||
	    system_state == SYSTEM_RUNNING) {
		spin_lock(&stop_lock);
		pr_crit("CPU%u: stopping\n", cpu);
		dump_stack();
		spin_unlock(&stop_lock);
	}

	set_cpu_online(cpu, false);

	local_irq_disable();

	hard_processor_halt(HALT_OK);
}

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

/*
 * not supported here
 */
int setup_profiling_timer(unsigned int multiplier)
{
	return -EINVAL;
}

/*
 * We use KICKs for inter-processor interrupts.
 *
 * For every CPU in "callmap" the IPI data must already have been
 * stored in that CPU's "ipi_data" member prior to calling this
 * function.
 */
static void kick_raise_softirq(cpumask_t callmap, unsigned int irq)
{
	int cpu;

	for_each_cpu(cpu, &callmap) {
		unsigned int thread;

		thread = cpu_2_hwthread_id[cpu];

		BUG_ON(thread == BAD_HWTHREAD_ID);

		metag_out32(1, T0KICKI + (thread * TnXKICK_STRIDE));
	}
}

static TBIRES ipi_handler(TBIRES State, int SigNum, int Triggers,
		   int Inst, PTBI pTBI, int *handled)
{
	*handled = do_IPI((struct pt_regs *)State.Sig.pCtx);

	return State;
}

static struct kick_irq_handler ipi_irq = {
	.func = ipi_handler,
};

static void smp_cross_call(cpumask_t callmap, enum ipi_msg_type msg)
{
	kick_raise_softirq(callmap, 1);
}

static inline unsigned int get_core_count(void)
{
	int i;
	unsigned int ret = 0;

	for (i = 0; i < CONFIG_NR_CPUS; i++) {
		if (core_reg_read(TXUCT_ID, TXENABLE_REGNUM, i))
			ret++;
	}

	return ret;
}

/*
 * Initialise the CPU possible map early - this describes the CPUs
 * which may be present or become present in the system.
 */
void __init smp_init_cpus(void)
{
	unsigned int i, ncores = get_core_count();

	/* If no hwthread_map early param was set use default mapping */
	for (i = 0; i < NR_CPUS; i++)
		if (cpu_2_hwthread_id[i] == BAD_HWTHREAD_ID) {
			cpu_2_hwthread_id[i] = i;
			hwthread_id_2_cpu[i] = i;
		}

	for (i = 0; i < ncores; i++)
		set_cpu_possible(i, true);

	kick_register_func(&ipi_irq);
}