aboutsummaryrefslogtreecommitdiffstats
path: root/arch/s390/kernel/time.c
blob: 52949df8852996985fd974c5f04d72c58d0e287d (plain) (blame)
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
/*
 *    Time of day based timer functions.
 *
 *  S390 version
 *    Copyright IBM Corp. 1999, 2008
 *    Author(s): Hartmut Penner (hp@de.ibm.com),
 *               Martin Schwidefsky (schwidefsky@de.ibm.com),
 *               Denis Joseph Barrow (djbarrow@de.ibm.com,barrow_dj@yahoo.com)
 *
 *  Derived from "arch/i386/kernel/time.c"
 *    Copyright (C) 1991, 1992, 1995  Linus Torvalds
 */

#define KMSG_COMPONENT "time"
#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt

#include <linux/kernel_stat.h>
#include <linux/errno.h>
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/param.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/cpu.h>
#include <linux/stop_machine.h>
#include <linux/time.h>
#include <linux/device.h>
#include <linux/delay.h>
#include <linux/init.h>
#include <linux/smp.h>
#include <linux/types.h>
#include <linux/profile.h>
#include <linux/timex.h>
#include <linux/notifier.h>
#include <linux/timekeeper_internal.h>
#include <linux/clockchips.h>
#include <linux/gfp.h>
#include <linux/kprobes.h>
#include <linux/uaccess.h>
#include <asm/facility.h>
#include <asm/delay.h>
#include <asm/div64.h>
#include <asm/vdso.h>
#include <asm/irq.h>
#include <asm/irq_regs.h>
#include <asm/vtimer.h>
#include <asm/stp.h>
#include <asm/cio.h>
#include "entry.h"

u64 sched_clock_base_cc = -1;	/* Force to data section. */
EXPORT_SYMBOL_GPL(sched_clock_base_cc);

static DEFINE_PER_CPU(struct clock_event_device, comparators);

ATOMIC_NOTIFIER_HEAD(s390_epoch_delta_notifier);
EXPORT_SYMBOL(s390_epoch_delta_notifier);

unsigned char ptff_function_mask[16];

static unsigned long long lpar_offset;
static unsigned long long initial_leap_seconds;
static unsigned long long tod_steering_end;
static long long tod_steering_delta;

/*
 * Get time offsets with PTFF
 */
void __init time_early_init(void)
{
	struct ptff_qto qto;
	struct ptff_qui qui;

	/* Initialize TOD steering parameters */
	tod_steering_end = sched_clock_base_cc;
	vdso_data->ts_end = tod_steering_end;

	if (!test_facility(28))
		return;

	ptff(&ptff_function_mask, sizeof(ptff_function_mask), PTFF_QAF);

	/* get LPAR offset */
	if (ptff_query(PTFF_QTO) && ptff(&qto, sizeof(qto), PTFF_QTO) == 0)
		lpar_offset = qto.tod_epoch_difference;

	/* get initial leap seconds */
	if (ptff_query(PTFF_QUI) && ptff(&qui, sizeof(qui), PTFF_QUI) == 0)
		initial_leap_seconds = (unsigned long long)
			((long) qui.old_leap * 4096000000L);
}

/*
 * Scheduler clock - returns current time in nanosec units.
 */
unsigned long long notrace sched_clock(void)
{
	return tod_to_ns(get_tod_clock_monotonic());
}
NOKPROBE_SYMBOL(sched_clock);

/*
 * Monotonic_clock - returns # of nanoseconds passed since time_init()
 */
unsigned long long monotonic_clock(void)
{
	return sched_clock();
}
EXPORT_SYMBOL(monotonic_clock);

void tod_to_timeval(__u64 todval, struct timespec64 *xt)
{
	unsigned long long sec;

	sec = todval >> 12;
	do_div(sec, 1000000);
	xt->tv_sec = sec;
	todval -= (sec * 1000000) << 12;
	xt->tv_nsec = ((todval * 1000) >> 12);
}
EXPORT_SYMBOL(tod_to_timeval);

void clock_comparator_work(void)
{
	struct clock_event_device *cd;

	S390_lowcore.clock_comparator = -1ULL;
	cd = this_cpu_ptr(&comparators);
	cd->event_handler(cd);
}

static int s390_next_event(unsigned long delta,
			   struct clock_event_device *evt)
{
	S390_lowcore.clock_comparator = get_tod_clock() + delta;
	set_clock_comparator(S390_lowcore.clock_comparator);
	return 0;
}

/*
 * Set up lowcore and control register of the current cpu to
 * enable TOD clock and clock comparator interrupts.
 */
void init_cpu_timer(void)
{
	struct clock_event_device *cd;
	int cpu;

	S390_lowcore.clock_comparator = -1ULL;
	set_clock_comparator(S390_lowcore.clock_comparator);

	cpu = smp_processor_id();
	cd = &per_cpu(comparators, cpu);
	cd->name		= "comparator";
	cd->features		= CLOCK_EVT_FEAT_ONESHOT;
	cd->mult		= 16777;
	cd->shift		= 12;
	cd->min_delta_ns	= 1;
	cd->max_delta_ns	= LONG_MAX;
	cd->rating		= 400;
	cd->cpumask		= cpumask_of(cpu);
	cd->set_next_event	= s390_next_event;

	clockevents_register_device(cd);

	/* Enable clock comparator timer interrupt. */
	__ctl_set_bit(0,11);

	/* Always allow the timing alert external interrupt. */
	__ctl_set_bit(0, 4);
}

static void clock_comparator_interrupt(struct ext_code ext_code,
				       unsigned int param32,
				       unsigned long param64)
{
	inc_irq_stat(IRQEXT_CLK);
	if (S390_lowcore.clock_comparator == -1ULL)
		set_clock_comparator(S390_lowcore.clock_comparator);
}

static void stp_timing_alert(struct stp_irq_parm *);

static void timing_alert_interrupt(struct ext_code ext_code,
				   unsigned int param32, unsigned long param64)
{
	inc_irq_stat(IRQEXT_TLA);
	if (param32 & 0x00038000)
		stp_timing_alert((struct stp_irq_parm *) &param32);
}

static void stp_reset(void);

void read_persistent_clock64(struct timespec64 *ts)
{
	__u64 clock;

	clock = get_tod_clock() - initial_leap_seconds;
	tod_to_timeval(clock - TOD_UNIX_EPOCH, ts);
}

void read_boot_clock64(struct timespec64 *ts)
{
	__u64 clock;

	clock = sched_clock_base_cc - initial_leap_seconds;
	tod_to_timeval(clock - TOD_UNIX_EPOCH, ts);
}

static u64 read_tod_clock(struct clocksource *cs)
{
	unsigned long long now, adj;

	preempt_disable(); /* protect from changes to steering parameters */
	now = get_tod_clock();
	adj = tod_steering_end - now;
	if (unlikely((s64) adj >= 0))
		/*
		 * manually steer by 1 cycle every 2^16 cycles. This
		 * corresponds to shifting the tod delta by 15. 1s is
		 * therefore steered in ~9h. The adjust will decrease
		 * over time, until it finally reaches 0.
		 */
		now += (tod_steering_delta < 0) ? (adj >> 15) : -(adj >> 15);
	preempt_enable();
	return now;
}

static struct clocksource clocksource_tod = {
	.name		= "tod",
	.rating		= 400,
	.read		= read_tod_clock,
	.mask		= -1ULL,
	.mult		= 1000,
	.shift		= 12,
	.flags		= CLOCK_SOURCE_IS_CONTINUOUS,
};

struct clocksource * __init clocksource_default_clock(void)
{
	return &clocksource_tod;
}

void update_vsyscall(struct timekeeper *tk)
{
	u64 nsecps;

	if (tk->tkr_mono.clock != &clocksource_tod)
		return;

	/* Make userspace gettimeofday spin until we're done. */
	++vdso_data->tb_update_count;
	smp_wmb();
	vdso_data->xtime_tod_stamp = tk->tkr_mono.cycle_last;
	vdso_data->xtime_clock_sec = tk->xtime_sec;
	vdso_data->xtime_clock_nsec = tk->tkr_mono.xtime_nsec;
	vdso_data->wtom_clock_sec =
		tk->xtime_sec + tk->wall_to_monotonic.tv_sec;
	vdso_data->wtom_clock_nsec = tk->tkr_mono.xtime_nsec +
		+ ((u64) tk->wall_to_monotonic.tv_nsec << tk->tkr_mono.shift);
	nsecps = (u64) NSEC_PER_SEC << tk->tkr_mono.shift;
	while (vdso_data->wtom_clock_nsec >= nsecps) {
		vdso_data->wtom_clock_nsec -= nsecps;
		vdso_data->wtom_clock_sec++;
	}

	vdso_data->xtime_coarse_sec = tk->xtime_sec;
	vdso_data->xtime_coarse_nsec =
		(long)(tk->tkr_mono.xtime_nsec >> tk->tkr_mono.shift);
	vdso_data->wtom_coarse_sec =
		vdso_data->xtime_coarse_sec + tk->wall_to_monotonic.tv_sec;
	vdso_data->wtom_coarse_nsec =
		vdso_data->xtime_coarse_nsec + tk->wall_to_monotonic.tv_nsec;
	while (vdso_data->wtom_coarse_nsec >= NSEC_PER_SEC) {
		vdso_data->wtom_coarse_nsec -= NSEC_PER_SEC;
		vdso_data->wtom_coarse_sec++;
	}

	vdso_data->tk_mult = tk->tkr_mono.mult;
	vdso_data->tk_shift = tk->tkr_mono.shift;
	smp_wmb();
	++vdso_data->tb_update_count;
}

extern struct timezone sys_tz;

void update_vsyscall_tz(void)
{
	vdso_data->tz_minuteswest = sys_tz.tz_minuteswest;
	vdso_data->tz_dsttime = sys_tz.tz_dsttime;
}

/*
 * Initialize the TOD clock and the CPU timer of
 * the boot cpu.
 */
void __init time_init(void)
{
	/* Reset time synchronization interfaces. */
	stp_reset();

	/* request the clock comparator external interrupt */
	if (register_external_irq(EXT_IRQ_CLK_COMP, clock_comparator_interrupt))
		panic("Couldn't request external interrupt 0x1004");

	/* request the timing alert external interrupt */
	if (register_external_irq(EXT_IRQ_TIMING_ALERT, timing_alert_interrupt))
		panic("Couldn't request external interrupt 0x1406");

	if (__clocksource_register(&clocksource_tod) != 0)
		panic("Could not register TOD clock source");

	/* Enable TOD clock interrupts on the boot cpu. */
	init_cpu_timer();

	/* Enable cpu timer interrupts on the boot cpu. */
	vtime_init();
}

static DEFINE_PER_CPU(atomic_t, clock_sync_word);
static DEFINE_MUTEX(clock_sync_mutex);
static unsigned long clock_sync_flags;

#define CLOCK_SYNC_HAS_STP	0
#define CLOCK_SYNC_STP		1

/*
 * The get_clock function for the physical clock. It will get the current
 * TOD clock, subtract the LPAR offset and write the result to *clock.
 * The function returns 0 if the clock is in sync with the external time
 * source. If the clock mode is local it will return -EOPNOTSUPP and
 * -EAGAIN if the clock is not in sync with the external reference.
 */
int get_phys_clock(unsigned long long *clock)
{
	atomic_t *sw_ptr;
	unsigned int sw0, sw1;

	sw_ptr = &get_cpu_var(clock_sync_word);
	sw0 = atomic_read(sw_ptr);
	*clock = get_tod_clock() - lpar_offset;
	sw1 = atomic_read(sw_ptr);
	put_cpu_var(clock_sync_word);
	if (sw0 == sw1 && (sw0 & 0x80000000U))
		/* Success: time is in sync. */
		return 0;
	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
		return -EOPNOTSUPP;
	if (!test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
		return -EACCES;
	return -EAGAIN;
}
EXPORT_SYMBOL(get_phys_clock);

/*
 * Make get_phys_clock() return -EAGAIN.
 */
static void disable_sync_clock(void *dummy)
{
	atomic_t *sw_ptr = this_cpu_ptr(&clock_sync_word);
	/*
	 * Clear the in-sync bit 2^31. All get_phys_clock calls will
	 * fail until the sync bit is turned back on. In addition
	 * increase the "sequence" counter to avoid the race of an
	 * stp event and the complete recovery against get_phys_clock.
	 */
	atomic_andnot(0x80000000, sw_ptr);
	atomic_inc(sw_ptr);
}

/*
 * Make get_phys_clock() return 0 again.
 * Needs to be called from a context disabled for preemption.
 */
static void enable_sync_clock(void)
{
	atomic_t *sw_ptr = this_cpu_ptr(&clock_sync_word);
	atomic_or(0x80000000, sw_ptr);
}

/*
 * Function to check if the clock is in sync.
 */
static inline int check_sync_clock(void)
{
	atomic_t *sw_ptr;
	int rc;

	sw_ptr = &get_cpu_var(clock_sync_word);
	rc = (atomic_read(sw_ptr) & 0x80000000U) != 0;
	put_cpu_var(clock_sync_word);
	return rc;
}

/*
 * Apply clock delta to the global data structures.
 * This is called once on the CPU that performed the clock sync.
 */
static void clock_sync_global(unsigned long long delta)
{
	unsigned long now, adj;
	struct ptff_qto qto;

	/* Fixup the monotonic sched clock. */
	sched_clock_base_cc += delta;
	/* Adjust TOD steering parameters. */
	vdso_data->tb_update_count++;
	now = get_tod_clock();
	adj = tod_steering_end - now;
	if (unlikely((s64) adj >= 0))
		/* Calculate how much of the old adjustment is left. */
		tod_steering_delta = (tod_steering_delta < 0) ?
			-(adj >> 15) : (adj >> 15);
	tod_steering_delta += delta;
	if ((abs(tod_steering_delta) >> 48) != 0)
		panic("TOD clock sync offset %lli is too large to drift\n",
		      tod_steering_delta);
	tod_steering_end = now + (abs(tod_steering_delta) << 15);
	vdso_data->ts_dir = (tod_steering_delta < 0) ? 0 : 1;
	vdso_data->ts_end = tod_steering_end;
	vdso_data->tb_update_count++;
	/* Update LPAR offset. */
	if (ptff_query(PTFF_QTO) && ptff(&qto, sizeof(qto), PTFF_QTO) == 0)
		lpar_offset = qto.tod_epoch_difference;
	/* Call the TOD clock change notifier. */
	atomic_notifier_call_chain(&s390_epoch_delta_notifier, 0, &delta);
}

/*
 * Apply clock delta to the per-CPU data structures of this CPU.
 * This is called for each online CPU after the call to clock_sync_global.
 */
static void clock_sync_local(unsigned long long delta)
{
	/* Add the delta to the clock comparator. */
	if (S390_lowcore.clock_comparator != -1ULL) {
		S390_lowcore.clock_comparator += delta;
		set_clock_comparator(S390_lowcore.clock_comparator);
	}
	/* Adjust the last_update_clock time-stamp. */
	S390_lowcore.last_update_clock += delta;
}

/* Single threaded workqueue used for stp sync events */
static struct workqueue_struct *time_sync_wq;

static void __init time_init_wq(void)
{
	if (time_sync_wq)
		return;
	time_sync_wq = create_singlethread_workqueue("timesync");
}

struct clock_sync_data {
	atomic_t cpus;
	int in_sync;
	unsigned long long clock_delta;
};

/*
 * Server Time Protocol (STP) code.
 */
static bool stp_online;
static struct stp_sstpi stp_info;
static void *stp_page;

static void stp_work_fn(struct work_struct *work);
static DEFINE_MUTEX(stp_work_mutex);
static DECLARE_WORK(stp_work, stp_work_fn);
static struct timer_list stp_timer;

static int __init early_parse_stp(char *p)
{
	return kstrtobool(p, &stp_online);
}
early_param("stp", early_parse_stp);

/*
 * Reset STP attachment.
 */
static void __init stp_reset(void)
{
	int rc;

	stp_page = (void *) get_zeroed_page(GFP_ATOMIC);
	rc = chsc_sstpc(stp_page, STP_OP_CTRL, 0x0000, NULL);
	if (rc == 0)
		set_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags);
	else if (stp_online) {
		pr_warn("The real or virtual hardware system does not provide an STP interface\n");
		free_page((unsigned long) stp_page);
		stp_page = NULL;
		stp_online = 0;
	}
}

static void stp_timeout(unsigned long dummy)
{
	queue_work(time_sync_wq, &stp_work);
}

static int __init stp_init(void)
{
	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
		return 0;
	setup_timer(&stp_timer, stp_timeout, 0UL);
	time_init_wq();
	if (!stp_online)
		return 0;
	queue_work(time_sync_wq, &stp_work);
	return 0;
}

arch_initcall(stp_init);

/*
 * STP timing alert. There are three causes:
 * 1) timing status change
 * 2) link availability change
 * 3) time control parameter change
 * In all three cases we are only interested in the clock source state.
 * If a STP clock source is now available use it.
 */
static void stp_timing_alert(struct stp_irq_parm *intparm)
{
	if (intparm->tsc || intparm->lac || intparm->tcpc)
		queue_work(time_sync_wq, &stp_work);
}

/*
 * STP sync check machine check. This is called when the timing state
 * changes from the synchronized state to the unsynchronized state.
 * After a STP sync check the clock is not in sync. The machine check
 * is broadcasted to all cpus at the same time.
 */
int stp_sync_check(void)
{
	disable_sync_clock(NULL);
	return 1;
}

/*
 * STP island condition machine check. This is called when an attached
 * server  attempts to communicate over an STP link and the servers
 * have matching CTN ids and have a valid stratum-1 configuration
 * but the configurations do not match.
 */
int stp_island_check(void)
{
	disable_sync_clock(NULL);
	return 1;
}

void stp_queue_work(void)
{
	queue_work(time_sync_wq, &stp_work);
}

static int stp_sync_clock(void *data)
{
	struct clock_sync_data *sync = data;
	unsigned long long clock_delta;
	static int first;
	int rc;

	enable_sync_clock();
	if (xchg(&first, 1) == 0) {
		/* Wait until all other cpus entered the sync function. */
		while (atomic_read(&sync->cpus) != 0)
			cpu_relax();
		rc = 0;
		if (stp_info.todoff[0] || stp_info.todoff[1] ||
		    stp_info.todoff[2] || stp_info.todoff[3] ||
		    stp_info.tmd != 2) {
			rc = chsc_sstpc(stp_page, STP_OP_SYNC, 0,
					&clock_delta);
			if (rc == 0) {
				sync->clock_delta = clock_delta;
				clock_sync_global(clock_delta);
				rc = chsc_sstpi(stp_page, &stp_info,
						sizeof(struct stp_sstpi));
				if (rc == 0 && stp_info.tmd != 2)
					rc = -EAGAIN;
			}
		}
		sync->in_sync = rc ? -EAGAIN : 1;
		xchg(&first, 0);
	} else {
		/* Slave */
		atomic_dec(&sync->cpus);
		/* Wait for in_sync to be set. */
		while (READ_ONCE(sync->in_sync) == 0)
			__udelay(1);
	}
	if (sync->in_sync != 1)
		/* Didn't work. Clear per-cpu in sync bit again. */
		disable_sync_clock(NULL);
	/* Apply clock delta to per-CPU fields of this CPU. */
	clock_sync_local(sync->clock_delta);

	return 0;
}

/*
 * STP work. Check for the STP state and take over the clock
 * synchronization if the STP clock source is usable.
 */
static void stp_work_fn(struct work_struct *work)
{
	struct clock_sync_data stp_sync;
	int rc;

	/* prevent multiple execution. */
	mutex_lock(&stp_work_mutex);

	if (!stp_online) {
		chsc_sstpc(stp_page, STP_OP_CTRL, 0x0000, NULL);
		del_timer_sync(&stp_timer);
		goto out_unlock;
	}

	rc = chsc_sstpc(stp_page, STP_OP_CTRL, 0xb0e0, NULL);
	if (rc)
		goto out_unlock;

	rc = chsc_sstpi(stp_page, &stp_info, sizeof(struct stp_sstpi));
	if (rc || stp_info.c == 0)
		goto out_unlock;

	/* Skip synchronization if the clock is already in sync. */
	if (check_sync_clock())
		goto out_unlock;

	memset(&stp_sync, 0, sizeof(stp_sync));
	get_online_cpus();
	atomic_set(&stp_sync.cpus, num_online_cpus() - 1);
	stop_machine(stp_sync_clock, &stp_sync, cpu_online_mask);
	put_online_cpus();

	if (!check_sync_clock())
		/*
		 * There is a usable clock but the synchonization failed.
		 * Retry after a second.
		 */
		mod_timer(&stp_timer, jiffies + HZ);

out_unlock:
	mutex_unlock(&stp_work_mutex);
}

/*
 * STP subsys sysfs interface functions
 */
static struct bus_type stp_subsys = {
	.name		= "stp",
	.dev_name	= "stp",
};

static ssize_t stp_ctn_id_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online)
		return -ENODATA;
	return sprintf(buf, "%016llx\n",
		       *(unsigned long long *) stp_info.ctnid);
}

static DEVICE_ATTR(ctn_id, 0400, stp_ctn_id_show, NULL);

static ssize_t stp_ctn_type_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online)
		return -ENODATA;
	return sprintf(buf, "%i\n", stp_info.ctn);
}

static DEVICE_ATTR(ctn_type, 0400, stp_ctn_type_show, NULL);

static ssize_t stp_dst_offset_show(struct device *dev,
				   struct device_attribute *attr,
				   char *buf)
{
	if (!stp_online || !(stp_info.vbits & 0x2000))
		return -ENODATA;
	return sprintf(buf, "%i\n", (int)(s16) stp_info.dsto);
}

static DEVICE_ATTR(dst_offset, 0400, stp_dst_offset_show, NULL);

static ssize_t stp_leap_seconds_show(struct device *dev,
					struct device_attribute *attr,
					char *buf)
{
	if (!stp_online || !(stp_info.vbits & 0x8000))
		return -ENODATA;
	return sprintf(buf, "%i\n", (int)(s16) stp_info.leaps);
}

static DEVICE_ATTR(leap_seconds, 0400, stp_leap_seconds_show, NULL);

static ssize_t stp_stratum_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online)
		return -ENODATA;
	return sprintf(buf, "%i\n", (int)(s16) stp_info.stratum);
}

static DEVICE_ATTR(stratum, 0400, stp_stratum_show, NULL);

static ssize_t stp_time_offset_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online || !(stp_info.vbits & 0x0800))
		return -ENODATA;
	return sprintf(buf, "%i\n", (int) stp_info.tto);
}

static DEVICE_ATTR(time_offset, 0400, stp_time_offset_show, NULL);

static ssize_t stp_time_zone_offset_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online || !(stp_info.vbits & 0x4000))
		return -ENODATA;
	return sprintf(buf, "%i\n", (int)(s16) stp_info.tzo);
}

static DEVICE_ATTR(time_zone_offset, 0400,
			 stp_time_zone_offset_show, NULL);

static ssize_t stp_timing_mode_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online)
		return -ENODATA;
	return sprintf(buf, "%i\n", stp_info.tmd);
}

static DEVICE_ATTR(timing_mode, 0400, stp_timing_mode_show, NULL);

static ssize_t stp_timing_state_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	if (!stp_online)
		return -ENODATA;
	return sprintf(buf, "%i\n", stp_info.tst);
}

static DEVICE_ATTR(timing_state, 0400, stp_timing_state_show, NULL);

static ssize_t stp_online_show(struct device *dev,
				struct device_attribute *attr,
				char *buf)
{
	return sprintf(buf, "%i\n", stp_online);
}

static ssize_t stp_online_store(struct device *dev,
				struct device_attribute *attr,
				const char *buf, size_t count)
{
	unsigned int value;

	value = simple_strtoul(buf, NULL, 0);
	if (value != 0 && value != 1)
		return -EINVAL;
	if (!test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
		return -EOPNOTSUPP;
	mutex_lock(&clock_sync_mutex);
	stp_online = value;
	if (stp_online)
		set_bit(CLOCK_SYNC_STP, &clock_sync_flags);
	else
		clear_bit(CLOCK_SYNC_STP, &clock_sync_flags);
	queue_work(time_sync_wq, &stp_work);
	mutex_unlock(&clock_sync_mutex);
	return count;
}

/*
 * Can't use DEVICE_ATTR because the attribute should be named
 * stp/online but dev_attr_online already exists in this file ..
 */
static struct device_attribute dev_attr_stp_online = {
	.attr = { .name = "online", .mode = 0600 },
	.show	= stp_online_show,
	.store	= stp_online_store,
};

static struct device_attribute *stp_attributes[] = {
	&dev_attr_ctn_id,
	&dev_attr_ctn_type,
	&dev_attr_dst_offset,
	&dev_attr_leap_seconds,
	&dev_attr_stp_online,
	&dev_attr_stratum,
	&dev_attr_time_offset,
	&dev_attr_time_zone_offset,
	&dev_attr_timing_mode,
	&dev_attr_timing_state,
	NULL
};

static int __init stp_init_sysfs(void)
{
	struct device_attribute **attr;
	int rc;

	rc = subsys_system_register(&stp_subsys, NULL);
	if (rc)
		goto out;
	for (attr = stp_attributes; *attr; attr++) {
		rc = device_create_file(stp_subsys.dev_root, *attr);
		if (rc)
			goto out_unreg;
	}
	return 0;
out_unreg:
	for (; attr >= stp_attributes; attr--)
		device_remove_file(stp_subsys.dev_root, *attr);
	bus_unregister(&stp_subsys);
out:
	return rc;
}

device_initcall(stp_init_sysfs);