aboutsummaryrefslogtreecommitdiffstats
path: root/arch/alpha/kernel/smp.c
blob: 0804b6abe20337eb02950d01e3320e4411a528c7 (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
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
/*
 *	linux/arch/alpha/kernel/smp.c
 *
 *      2001-07-09 Phil Ezolt (Phillip.Ezolt@compaq.com)
 *            Renamed modified smp_call_function to smp_call_function_on_cpu()
 *            Created an function that conforms to the old calling convention
 *            of smp_call_function().
 *
 *            This is helpful for DCPI.
 *
 */

#include <linux/errno.h>
#include <linux/kernel.h>
#include <linux/kernel_stat.h>
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/threads.h>
#include <linux/smp.h>
#include <linux/interrupt.h>
#include <linux/init.h>
#include <linux/delay.h>
#include <linux/spinlock.h>
#include <linux/irq.h>
#include <linux/cache.h>
#include <linux/profile.h>
#include <linux/bitops.h>

#include <asm/hwrpb.h>
#include <asm/ptrace.h>
#include <asm/atomic.h>

#include <asm/io.h>
#include <asm/irq.h>
#include <asm/pgtable.h>
#include <asm/pgalloc.h>
#include <asm/mmu_context.h>
#include <asm/tlbflush.h>

#include "proto.h"
#include "irq_impl.h"


#define DEBUG_SMP 0
#if DEBUG_SMP
#define DBGS(args)	printk args
#else
#define DBGS(args)
#endif

/* A collection of per-processor data.  */
struct cpuinfo_alpha cpu_data[NR_CPUS];
EXPORT_SYMBOL(cpu_data);

/* A collection of single bit ipi messages.  */
static struct {
	unsigned long bits ____cacheline_aligned;
} ipi_data[NR_CPUS] __cacheline_aligned;

enum ipi_message_type {
	IPI_RESCHEDULE,
	IPI_CALL_FUNC,
	IPI_CPU_STOP,
};

/* Set to a secondary's cpuid when it comes online.  */
static int smp_secondary_alive __devinitdata = 0;

/* Which cpus ids came online.  */
cpumask_t cpu_online_map;

EXPORT_SYMBOL(cpu_online_map);

int smp_num_probed;		/* Internal processor count */
int smp_num_cpus = 1;		/* Number that came online.  */
EXPORT_SYMBOL(smp_num_cpus);

extern void calibrate_delay(void);



/*
 * Called by both boot and secondaries to move global data into
 *  per-processor storage.
 */
static inline void __init
smp_store_cpu_info(int cpuid)
{
	cpu_data[cpuid].loops_per_jiffy = loops_per_jiffy;
	cpu_data[cpuid].last_asn = ASN_FIRST_VERSION;
	cpu_data[cpuid].need_new_asn = 0;
	cpu_data[cpuid].asn_lock = 0;
}

/*
 * Ideally sets up per-cpu profiling hooks.  Doesn't do much now...
 */
static inline void __init
smp_setup_percpu_timer(int cpuid)
{
	cpu_data[cpuid].prof_counter = 1;
	cpu_data[cpuid].prof_multiplier = 1;
}

static void __init
wait_boot_cpu_to_stop(int cpuid)
{
	unsigned long stop = jiffies + 10*HZ;

	while (time_before(jiffies, stop)) {
	        if (!smp_secondary_alive)
			return;
		barrier();
	}

	printk("wait_boot_cpu_to_stop: FAILED on CPU %d, hanging now\n", cpuid);
	for (;;)
		barrier();
}

/*
 * Where secondaries begin a life of C.
 */
void __init
smp_callin(void)
{
	int cpuid = hard_smp_processor_id();

	if (cpu_test_and_set(cpuid, cpu_online_map)) {
		printk("??, cpu 0x%x already present??\n", cpuid);
		BUG();
	}

	/* Turn on machine checks.  */
	wrmces(7);

	/* Set trap vectors.  */
	trap_init();

	/* Set interrupt vector.  */
	wrent(entInt, 0);

	/* Get our local ticker going. */
	smp_setup_percpu_timer(cpuid);

	/* Call platform-specific callin, if specified */
	if (alpha_mv.smp_callin) alpha_mv.smp_callin();

	/* All kernel threads share the same mm context.  */
	atomic_inc(&init_mm.mm_count);
	current->active_mm = &init_mm;

	/* Must have completely accurate bogos.  */
	local_irq_enable();

	/* Wait boot CPU to stop with irq enabled before running
	   calibrate_delay. */
	wait_boot_cpu_to_stop(cpuid);
	mb();
	calibrate_delay();

	smp_store_cpu_info(cpuid);
	/* Allow master to continue only after we written loops_per_jiffy.  */
	wmb();
	smp_secondary_alive = 1;

	DBGS(("smp_callin: commencing CPU %d current %p active_mm %p\n",
	      cpuid, current, current->active_mm));

	/* Do nothing.  */
	cpu_idle();
}

/* Wait until hwrpb->txrdy is clear for cpu.  Return -1 on timeout.  */
static int __devinit
wait_for_txrdy (unsigned long cpumask)
{
	unsigned long timeout;

	if (!(hwrpb->txrdy & cpumask))
		return 0;

	timeout = jiffies + 10*HZ;
	while (time_before(jiffies, timeout)) {
		if (!(hwrpb->txrdy & cpumask))
			return 0;
		udelay(10);
		barrier();
	}

	return -1;
}

/*
 * Send a message to a secondary's console.  "START" is one such
 * interesting message.  ;-)
 */
static void __init
send_secondary_console_msg(char *str, int cpuid)
{
	struct percpu_struct *cpu;
	register char *cp1, *cp2;
	unsigned long cpumask;
	size_t len;

	cpu = (struct percpu_struct *)
		((char*)hwrpb
		 + hwrpb->processor_offset
		 + cpuid * hwrpb->processor_size);

	cpumask = (1UL << cpuid);
	if (wait_for_txrdy(cpumask))
		goto timeout;

	cp2 = str;
	len = strlen(cp2);
	*(unsigned int *)&cpu->ipc_buffer[0] = len;
	cp1 = (char *) &cpu->ipc_buffer[1];
	memcpy(cp1, cp2, len);

	/* atomic test and set */
	wmb();
	set_bit(cpuid, &hwrpb->rxrdy);

	if (wait_for_txrdy(cpumask))
		goto timeout;
	return;

 timeout:
	printk("Processor %x not ready\n", cpuid);
}

/*
 * A secondary console wants to send a message.  Receive it.
 */
static void
recv_secondary_console_msg(void)
{
	int mycpu, i, cnt;
	unsigned long txrdy = hwrpb->txrdy;
	char *cp1, *cp2, buf[80];
	struct percpu_struct *cpu;

	DBGS(("recv_secondary_console_msg: TXRDY 0x%lx.\n", txrdy));

	mycpu = hard_smp_processor_id();

	for (i = 0; i < NR_CPUS; i++) {
		if (!(txrdy & (1UL << i)))
			continue;

		DBGS(("recv_secondary_console_msg: "
		      "TXRDY contains CPU %d.\n", i));

		cpu = (struct percpu_struct *)
		  ((char*)hwrpb
		   + hwrpb->processor_offset
		   + i * hwrpb->processor_size);

 		DBGS(("recv_secondary_console_msg: on %d from %d"
		      " HALT_REASON 0x%lx FLAGS 0x%lx\n",
		      mycpu, i, cpu->halt_reason, cpu->flags));

		cnt = cpu->ipc_buffer[0] >> 32;
		if (cnt <= 0 || cnt >= 80)
			strcpy(buf, "<<< BOGUS MSG >>>");
		else {
			cp1 = (char *) &cpu->ipc_buffer[11];
			cp2 = buf;
			strcpy(cp2, cp1);
			
			while ((cp2 = strchr(cp2, '\r')) != 0) {
				*cp2 = ' ';
				if (cp2[1] == '\n')
					cp2[1] = ' ';
			}
		}

		DBGS((KERN_INFO "recv_secondary_console_msg: on %d "
		      "message is '%s'\n", mycpu, buf));
	}

	hwrpb->txrdy = 0;
}

/*
 * Convince the console to have a secondary cpu begin execution.
 */
static int __init
secondary_cpu_start(int cpuid, struct task_struct *idle)
{
	struct percpu_struct *cpu;
	struct pcb_struct *hwpcb, *ipcb;
	unsigned long timeout;
	  
	cpu = (struct percpu_struct *)
		((char*)hwrpb
		 + hwrpb->processor_offset
		 + cpuid * hwrpb->processor_size);
	hwpcb = (struct pcb_struct *) cpu->hwpcb;
	ipcb = &task_thread_info(idle)->pcb;

	/* Initialize the CPU's HWPCB to something just good enough for
	   us to get started.  Immediately after starting, we'll swpctx
	   to the target idle task's pcb.  Reuse the stack in the mean
	   time.  Precalculate the target PCBB.  */
	hwpcb->ksp = (unsigned long)ipcb + sizeof(union thread_union) - 16;
	hwpcb->usp = 0;
	hwpcb->ptbr = ipcb->ptbr;
	hwpcb->pcc = 0;
	hwpcb->asn = 0;
	hwpcb->unique = virt_to_phys(ipcb);
	hwpcb->flags = ipcb->flags;
	hwpcb->res1 = hwpcb->res2 = 0;

#if 0
	DBGS(("KSP 0x%lx PTBR 0x%lx VPTBR 0x%lx UNIQUE 0x%lx\n",
	      hwpcb->ksp, hwpcb->ptbr, hwrpb->vptb, hwpcb->unique));
#endif
	DBGS(("Starting secondary cpu %d: state 0x%lx pal_flags 0x%lx\n",
	      cpuid, idle->state, ipcb->flags));

	/* Setup HWRPB fields that SRM uses to activate secondary CPU */
	hwrpb->CPU_restart = __smp_callin;
	hwrpb->CPU_restart_data = (unsigned long) __smp_callin;

	/* Recalculate and update the HWRPB checksum */
	hwrpb_update_checksum(hwrpb);

	/*
	 * Send a "start" command to the specified processor.
	 */

	/* SRM III 3.4.1.3 */
	cpu->flags |= 0x22;	/* turn on Context Valid and Restart Capable */
	cpu->flags &= ~1;	/* turn off Bootstrap In Progress */
	wmb();

	send_secondary_console_msg("START\r\n", cpuid);

	/* Wait 10 seconds for an ACK from the console.  */
	timeout = jiffies + 10*HZ;
	while (time_before(jiffies, timeout)) {
		if (cpu->flags & 1)
			goto started;
		udelay(10);
		barrier();
	}
	printk(KERN_ERR "SMP: Processor %d failed to start.\n", cpuid);
	return -1;

 started:
	DBGS(("secondary_cpu_start: SUCCESS for CPU %d!!!\n", cpuid));
	return 0;
}

/*
 * Bring one cpu online.
 */
static int __cpuinit
smp_boot_one_cpu(int cpuid)
{
	struct task_struct *idle;
	unsigned long timeout;

	/* Cook up an idler for this guy.  Note that the address we
	   give to kernel_thread is irrelevant -- it's going to start
	   where HWRPB.CPU_restart says to start.  But this gets all
	   the other task-y sort of data structures set up like we
	   wish.  We can't use kernel_thread since we must avoid
	   rescheduling the child.  */
	idle = fork_idle(cpuid);
	if (IS_ERR(idle))
		panic("failed fork for CPU %d", cpuid);

	DBGS(("smp_boot_one_cpu: CPU %d state 0x%lx flags 0x%lx\n",
	      cpuid, idle->state, idle->flags));

	/* Signal the secondary to wait a moment.  */
	smp_secondary_alive = -1;

	/* Whirrr, whirrr, whirrrrrrrrr... */
	if (secondary_cpu_start(cpuid, idle))
		return -1;

	/* Notify the secondary CPU it can run calibrate_delay.  */
	mb();
	smp_secondary_alive = 0;

	/* We've been acked by the console; wait one second for
	   the task to start up for real.  */
	timeout = jiffies + 1*HZ;
	while (time_before(jiffies, timeout)) {
		if (smp_secondary_alive == 1)
			goto alive;
		udelay(10);
		barrier();
	}

	/* We failed to boot the CPU.  */

	printk(KERN_ERR "SMP: Processor %d is stuck.\n", cpuid);
	return -1;

 alive:
	/* Another "Red Snapper". */
	return 0;
}

/*
 * Called from setup_arch.  Detect an SMP system and which processors
 * are present.
 */
void __init
setup_smp(void)
{
	struct percpu_struct *cpubase, *cpu;
	unsigned long i;

	if (boot_cpuid != 0) {
		printk(KERN_WARNING "SMP: Booting off cpu %d instead of 0?\n",
		       boot_cpuid);
	}

	if (hwrpb->nr_processors > 1) {
		int boot_cpu_palrev;

		DBGS(("setup_smp: nr_processors %ld\n",
		      hwrpb->nr_processors));

		cpubase = (struct percpu_struct *)
			((char*)hwrpb + hwrpb->processor_offset);
		boot_cpu_palrev = cpubase->pal_revision;

		for (i = 0; i < hwrpb->nr_processors; i++) {
			cpu = (struct percpu_struct *)
				((char *)cpubase + i*hwrpb->processor_size);
			if ((cpu->flags & 0x1cc) == 0x1cc) {
				smp_num_probed++;
				/* Assume here that "whami" == index */
				cpu_set(i, cpu_present_map);
				cpu->pal_revision = boot_cpu_palrev;
			}

			DBGS(("setup_smp: CPU %d: flags 0x%lx type 0x%lx\n",
			      i, cpu->flags, cpu->type));
			DBGS(("setup_smp: CPU %d: PAL rev 0x%lx\n",
			      i, cpu->pal_revision));
		}
	} else {
		smp_num_probed = 1;
	}

	printk(KERN_INFO "SMP: %d CPUs probed -- cpu_present_map = %lx\n",
	       smp_num_probed, cpu_present_map.bits[0]);
}

/*
 * Called by smp_init prepare the secondaries
 */
void __init
smp_prepare_cpus(unsigned int max_cpus)
{
	/* Take care of some initial bookkeeping.  */
	memset(ipi_data, 0, sizeof(ipi_data));

	current_thread_info()->cpu = boot_cpuid;

	smp_store_cpu_info(boot_cpuid);
	smp_setup_percpu_timer(boot_cpuid);

	/* Nothing to do on a UP box, or when told not to.  */
	if (smp_num_probed == 1 || max_cpus == 0) {
		cpu_present_map = cpumask_of_cpu(boot_cpuid);
		printk(KERN_INFO "SMP mode deactivated.\n");
		return;
	}

	printk(KERN_INFO "SMP starting up secondaries.\n");

	smp_num_cpus = smp_num_probed;
}

void __devinit
smp_prepare_boot_cpu(void)
{
}

int __cpuinit
__cpu_up(unsigned int cpu)
{
	smp_boot_one_cpu(cpu);

	return cpu_online(cpu) ? 0 : -ENOSYS;
}

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

	for(cpu = 0; cpu < NR_CPUS; cpu++) 
		if (cpu_online(cpu))
			bogosum += cpu_data[cpu].loops_per_jiffy;
	
	printk(KERN_INFO "SMP: Total of %d processors activated "
	       "(%lu.%02lu BogoMIPS).\n",
	       num_online_cpus(), 
	       (bogosum + 2500) / (500000/HZ),
	       ((bogosum + 2500) / (5000/HZ)) % 100);
}


void
smp_percpu_timer_interrupt(struct pt_regs *regs)
{
	struct pt_regs *old_regs;
	int cpu = smp_processor_id();
	unsigned long user = user_mode(regs);
	struct cpuinfo_alpha *data = &cpu_data[cpu];

	old_regs = set_irq_regs(regs);

	/* Record kernel PC.  */
	profile_tick(CPU_PROFILING);

	if (!--data->prof_counter) {
		/* We need to make like a normal interrupt -- otherwise
		   timer interrupts ignore the global interrupt lock,
		   which would be a Bad Thing.  */
		irq_enter();

		update_process_times(user);

		data->prof_counter = data->prof_multiplier;

		irq_exit();
	}
	set_irq_regs(old_regs);
}

int
setup_profiling_timer(unsigned int multiplier)
{
	return -EINVAL;
}


static void
send_ipi_message(cpumask_t to_whom, enum ipi_message_type operation)
{
	int i;

	mb();
	for_each_cpu_mask(i, to_whom)
		set_bit(operation, &ipi_data[i].bits);

	mb();
	for_each_cpu_mask(i, to_whom)
		wripir(i);
}

/* Structure and data for smp_call_function.  This is designed to 
   minimize static memory requirements.  Plus it looks cleaner.  */

struct smp_call_struct {
	void (*func) (void *info);
	void *info;
	long wait;
	atomic_t unstarted_count;
	atomic_t unfinished_count;
};

static struct smp_call_struct *smp_call_function_data;

/* Atomicly drop data into a shared pointer.  The pointer is free if
   it is initially locked.  If retry, spin until free.  */

static int
pointer_lock (void *lock, void *data, int retry)
{
	void *old, *tmp;

	mb();
 again:
	/* Compare and swap with zero.  */
	asm volatile (
	"1:	ldq_l	%0,%1\n"
	"	mov	%3,%2\n"
	"	bne	%0,2f\n"
	"	stq_c	%2,%1\n"
	"	beq	%2,1b\n"
	"2:"
	: "=&r"(old), "=m"(*(void **)lock), "=&r"(tmp)
	: "r"(data)
	: "memory");

	if (old == 0)
		return 0;
	if (! retry)
		return -EBUSY;

	while (*(void **)lock)
		barrier();
	goto again;
}

void
handle_ipi(struct pt_regs *regs)
{
	int this_cpu = smp_processor_id();
	unsigned long *pending_ipis = &ipi_data[this_cpu].bits;
	unsigned long ops;

#if 0
	DBGS(("handle_ipi: on CPU %d ops 0x%lx PC 0x%lx\n",
	      this_cpu, *pending_ipis, regs->pc));
#endif

	mb();	/* Order interrupt and bit testing. */
	while ((ops = xchg(pending_ipis, 0)) != 0) {
	  mb();	/* Order bit clearing and data access. */
	  do {
		unsigned long which;

		which = ops & -ops;
		ops &= ~which;
		which = __ffs(which);

		switch (which) {
		case IPI_RESCHEDULE:
			/* Reschedule callback.  Everything to be done
			   is done by the interrupt return path.  */
			break;

		case IPI_CALL_FUNC:
		    {
			struct smp_call_struct *data;
			void (*func)(void *info);
			void *info;
			int wait;

			data = smp_call_function_data;
			func = data->func;
			info = data->info;
			wait = data->wait;

			/* Notify the sending CPU that the data has been
			   received, and execution is about to begin.  */
			mb();
			atomic_dec (&data->unstarted_count);

			/* At this point the structure may be gone unless
			   wait is true.  */
			(*func)(info);

			/* Notify the sending CPU that the task is done.  */
			mb();
			if (wait) atomic_dec (&data->unfinished_count);
			break;
		    }

		case IPI_CPU_STOP:
			halt();

		default:
			printk(KERN_CRIT "Unknown IPI on CPU %d: %lu\n",
			       this_cpu, which);
			break;
		}
	  } while (ops);

	  mb();	/* Order data access and bit testing. */
	}

	cpu_data[this_cpu].ipi_count++;

	if (hwrpb->txrdy)
		recv_secondary_console_msg();
}

void
smp_send_reschedule(int cpu)
{
#ifdef DEBUG_IPI_MSG
	if (cpu == hard_smp_processor_id())
		printk(KERN_WARNING
		       "smp_send_reschedule: Sending IPI to self.\n");
#endif
	send_ipi_message(cpumask_of_cpu(cpu), IPI_RESCHEDULE);
}

void
smp_send_stop(void)
{
	cpumask_t to_whom = cpu_possible_map;
	cpu_clear(smp_processor_id(), to_whom);
#ifdef DEBUG_IPI_MSG
	if (hard_smp_processor_id() != boot_cpu_id)
		printk(KERN_WARNING "smp_send_stop: Not on boot cpu.\n");
#endif
	send_ipi_message(to_whom, IPI_CPU_STOP);
}

/*
 * 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
 * hardware interrupt handler or from a bottom half handler.
 */

int
smp_call_function_on_cpu (void (*func) (void *info), void *info, int retry,
			  int wait, cpumask_t to_whom)
{
	struct smp_call_struct data;
	unsigned long timeout;
	int num_cpus_to_call;
	
	/* Can deadlock when called with interrupts disabled */
	WARN_ON(irqs_disabled());

	data.func = func;
	data.info = info;
	data.wait = wait;

	cpu_clear(smp_processor_id(), to_whom);
	num_cpus_to_call = cpus_weight(to_whom);

	atomic_set(&data.unstarted_count, num_cpus_to_call);
	atomic_set(&data.unfinished_count, num_cpus_to_call);

	/* Acquire the smp_call_function_data mutex.  */
	if (pointer_lock(&smp_call_function_data, &data, retry))
		return -EBUSY;

	/* Send a message to the requested CPUs.  */
	send_ipi_message(to_whom, IPI_CALL_FUNC);

	/* Wait for a minimal response.  */
	timeout = jiffies + HZ;
	while (atomic_read (&data.unstarted_count) > 0
	       && time_before (jiffies, timeout))
		barrier();

	/* If there's no response yet, log a message but allow a longer
	 * timeout period -- if we get a response this time, log
	 * a message saying when we got it.. 
	 */
	if (atomic_read(&data.unstarted_count) > 0) {
		long start_time = jiffies;
		printk(KERN_ERR "%s: initial timeout -- trying long wait\n",
		       __FUNCTION__);
		timeout = jiffies + 30 * HZ;
		while (atomic_read(&data.unstarted_count) > 0
		       && time_before(jiffies, timeout))
			barrier();
		if (atomic_read(&data.unstarted_count) <= 0) {
			long delta = jiffies - start_time;
			printk(KERN_ERR 
			       "%s: response %ld.%ld seconds into long wait\n",
			       __FUNCTION__, delta / HZ,
			       (100 * (delta - ((delta / HZ) * HZ))) / HZ);
		}
	}

	/* We either got one or timed out -- clear the lock. */
	mb();
	smp_call_function_data = NULL;

	/* 
	 * If after both the initial and long timeout periods we still don't
	 * have a response, something is very wrong...
	 */
	BUG_ON(atomic_read (&data.unstarted_count) > 0);

	/* Wait for a complete response, if needed.  */
	if (wait) {
		while (atomic_read (&data.unfinished_count) > 0)
			barrier();
	}

	return 0;
}
EXPORT_SYMBOL(smp_call_function_on_cpu);

int
smp_call_function (void (*func) (void *info), void *info, int retry, int wait)
{
	return smp_call_function_on_cpu (func, info, retry, wait,
					 cpu_online_map);
}
EXPORT_SYMBOL(smp_call_function);

static void
ipi_imb(void *ignored)
{
	imb();
}

void
smp_imb(void)
{
	/* Must wait other processors to flush their icache before continue. */
	if (on_each_cpu(ipi_imb, NULL, 1, 1))
		printk(KERN_CRIT "smp_imb: timed out\n");
}
EXPORT_SYMBOL(smp_imb);

static void
ipi_flush_tlb_all(void *ignored)
{
	tbia();
}

void
flush_tlb_all(void)
{
	/* Although we don't have any data to pass, we do want to
	   synchronize with the other processors.  */
	if (on_each_cpu(ipi_flush_tlb_all, NULL, 1, 1)) {
		printk(KERN_CRIT "flush_tlb_all: timed out\n");
	}
}

#define asn_locked() (cpu_data[smp_processor_id()].asn_lock)

static void
ipi_flush_tlb_mm(void *x)
{
	struct mm_struct *mm = (struct mm_struct *) x;
	if (mm == current->active_mm && !asn_locked())
		flush_tlb_current(mm);
	else
		flush_tlb_other(mm);
}

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

	if (mm == current->active_mm) {
		flush_tlb_current(mm);
		if (atomic_read(&mm->mm_users) <= 1) {
			int cpu, this_cpu = smp_processor_id();
			for (cpu = 0; cpu < NR_CPUS; cpu++) {
				if (!cpu_online(cpu) || cpu == this_cpu)
					continue;
				if (mm->context[cpu])
					mm->context[cpu] = 0;
			}
			preempt_enable();
			return;
		}
	}

	if (smp_call_function(ipi_flush_tlb_mm, mm, 1, 1)) {
		printk(KERN_CRIT "flush_tlb_mm: timed out\n");
	}

	preempt_enable();
}
EXPORT_SYMBOL(flush_tlb_mm);

struct flush_tlb_page_struct {
	struct vm_area_struct *vma;
	struct mm_struct *mm;
	unsigned long addr;
};

static void
ipi_flush_tlb_page(void *x)
{
	struct flush_tlb_page_struct *data = (struct flush_tlb_page_struct *)x;
	struct mm_struct * mm = data->mm;

	if (mm == current->active_mm && !asn_locked())
		flush_tlb_current_page(mm, data->vma, data->addr);
	else
		flush_tlb_other(mm);
}

void
flush_tlb_page(struct vm_area_struct *vma, unsigned long addr)
{
	struct flush_tlb_page_struct data;
	struct mm_struct *mm = vma->vm_mm;

	preempt_disable();

	if (mm == current->active_mm) {
		flush_tlb_current_page(mm, vma, addr);
		if (atomic_read(&mm->mm_users) <= 1) {
			int cpu, this_cpu = smp_processor_id();
			for (cpu = 0; cpu < NR_CPUS; cpu++) {
				if (!cpu_online(cpu) || cpu == this_cpu)
					continue;
				if (mm->context[cpu])
					mm->context[cpu] = 0;
			}
			preempt_enable();
			return;
		}
	}

	data.vma = vma;
	data.mm = mm;
	data.addr = addr;

	if (smp_call_function(ipi_flush_tlb_page, &data, 1, 1)) {
		printk(KERN_CRIT "flush_tlb_page: timed out\n");
	}

	preempt_enable();
}
EXPORT_SYMBOL(flush_tlb_page);

void
flush_tlb_range(struct vm_area_struct *vma, unsigned long start, unsigned long end)
{
	/* On the Alpha we always flush the whole user tlb.  */
	flush_tlb_mm(vma->vm_mm);
}
EXPORT_SYMBOL(flush_tlb_range);

static void
ipi_flush_icache_page(void *x)
{
	struct mm_struct *mm = (struct mm_struct *) x;
	if (mm == current->active_mm && !asn_locked())
		__load_new_mm_context(mm);
	else
		flush_tlb_other(mm);
}

void
flush_icache_user_range(struct vm_area_struct *vma, struct page *page,
			unsigned long addr, int len)
{
	struct mm_struct *mm = vma->vm_mm;

	if ((vma->vm_flags & VM_EXEC) == 0)
		return;

	preempt_disable();

	if (mm == current->active_mm) {
		__load_new_mm_context(mm);
		if (atomic_read(&mm->mm_users) <= 1) {
			int cpu, this_cpu = smp_processor_id();
			for (cpu = 0; cpu < NR_CPUS; cpu++) {
				if (!cpu_online(cpu) || cpu == this_cpu)
					continue;
				if (mm->context[cpu])
					mm->context[cpu] = 0;
			}
			preempt_enable();
			return;
		}
	}

	if (smp_call_function(ipi_flush_icache_page, mm, 1, 1)) {
		printk(KERN_CRIT "flush_icache_page: timed out\n");
	}

	preempt_enable();
}