aboutsummaryrefslogtreecommitdiffstats
path: root/kernel/time/clocksource.c
Commit message (Expand)AuthorAgeFilesLines
* clocksource: Add clocksource_register_hz/khz interfaceJohn Stultz2010-05-101-0/+48
* timekeeping: Prevent oops when GENERIC_TIME=njohn stultz2010-03-021-0/+4
* Merge branch 'timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/...Linus Torvalds2010-03-011-1/+13
|\
| * clocksource: add suspend callbackMagnus Damm2010-02-051-0/+12
| * clocksource: add argument to resume callbackMagnus Damm2010-02-051-1/+1
* | clocksource: Prevent potential kgdb dead lockThomas Gleixner2010-01-261-3/+15
|/
* Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jik...Linus Torvalds2009-12-091-1/+1
|\
| * tree-wide: fix typos "selct" + "slect" -> "select"Uwe Kleine-König2009-11-091-1/+1
* | Merge branch 'timers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/...Linus Torvalds2009-12-081-0/+97
|\ \
| * | clocksource/events: Fix fallout of generic code changesThomas Gleixner2009-11-141-2/+2
| * | nohz: Prevent clocksource wrapping during idleJon Hunter2009-11-131-0/+44
| * | clocksource: Provide a generic mult/shift factor calculationThomas Gleixner2009-11-131-0/+53
| |/
* / clocksource/timecompare: Fix symbol exports to be GPL'd.David S. Miller2009-11-111-3/+3
|/
* clocksource: Resume clocksource without taking the clocksource mutexMartin Schwidefsky2009-09-241-4/+0
* clocksource: Delay clocksource down rating to late bootThomas Gleixner2009-09-141-10/+18
* clocksource: clocksource_select must be called with mutex lockedThomas Gleixner2009-09-141-0/+2
* clocksource: Resolve cpu hotplug dead lock with TSC unstable, fix crashMartin Schwidefsky2009-09-111-2/+3
* clocksource: Resolve cpu hotplug dead lock with TSC unstableThomas Gleixner2009-08-281-3/+30
* clocksource: Avoid clocksource watchdog circular locking dependencyMartin Schwidefsky2009-08-191-3/+16
* clocksource: Protect the watchdog rating changes with clocksource_mutexThomas Gleixner2009-08-191-5/+13
* clocksource: Call clocksource_change_rating() outside of watchdog_lockThomas Gleixner2009-08-151-2/+9
* timekeeping: Update clocksource with stop_machineMartin Schwidefsky2009-08-151-68/+44
* timekeeping: Introduce struct timekeeperMartin Schwidefsky2009-08-151-2/+4
* clocksource: Move watchdog downgrade to a work queue threadMartin Schwidefsky2009-08-151-17/+39
* clocksource: Refactor clocksource watchdogMartin Schwidefsky2009-08-151-28/+69
* clocksource: Simplify clocksource watchdog resume logicMartin Schwidefsky2009-08-151-8/+15
* clocksource: Delay clocksource watchdog highres enablementMartin Schwidefsky2009-08-151-25/+22
* clocksource: Cleanup clocksource selectionMartin Schwidefsky2009-08-151-83/+51
* clocksource: Prevent NULL pointer dereferenceThomas Gleixner2009-07-191-1/+1
* clocksource: prevent selection of low resolution clocksourse also for nohz=onThomas Gleixner2009-06-131-8/+10
* clocksource: sanity check sysfs clocksource changesjohn stultz2009-06-111-1/+17
* clocksource: setup mult_orig in clocksource_enable()Magnus Damm2009-05-021-3/+0
* clocksource: pass clocksource to read() callbackMagnus Damm2009-04-211-4/+4
* clocksource: allow usage independent of timekeeping.cPatrick Ohly2009-02-151-0/+76
* cpumask: replace for_each_cpu_mask_nr with for_each_cpu in kernel/time/Rusty Russell2009-01-011-1/+2
* cpumask: convert kernel time functionsRusty Russell2009-01-011-3/+3
* clocksource: keep track of original clocksource frequencyJohn Stultz2008-08-211-0/+3
* Merge branch 'cpus4096-for-linus' of git://git.kernel.org/pub/scm/linux/kerne...Linus Torvalds2008-07-231-2/+2
|\
| * clocksource/events: use performance variant for_each_cpu_mask_nrMike Travis2008-05-231-2/+2
* | sysdev: Pass the attribute to the low level sysdev show/store functionAndi Kleen2008-07-211-2/+6
|/
* clocksource: allow read access to available/current_clocksourceHeiko Carstens2008-05-031-2/+2
* clocksource: Fix permissions for available_clocksourceHeiko Carstens2008-05-031-1/+1
* Merge git://git.kernel.org/pub/scm/linux/kernel/git/tglx/linux-2.6-hrtLinus Torvalds2008-04-181-4/+14
|\
| * clocksource: make clocksource watchdog cycle through online CPUsAndi Kleen2008-04-171-4/+14
* | kgdb: clocksource watchdogJason Wessel2008-04-171-0/+12
|/
* clocksource: revert: use init_timer_deferrable for clocksource_watchdogThomas Gleixner2008-03-251-1/+1
* revert "clocksource: make clocksource watchdog cycle through online CPUs"Andrew Morton2008-03-191-10/+4
* clocksource: remove redundant codeLi Zefan2008-02-081-1/+0
* time: fix sysfs_show_{available,current}_clocksources() buffer overflow problemMiao Xie2008-02-061-9/+10
* clocksource: add unregister function to disable unusable clocksourcesThomas Gleixner2008-01-301-0/+15
id='n675' href='#n675'>675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
/*
 *  arch/s390/kernel/time.c
 *    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/sysdev.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/clocksource.h>
#include <linux/clockchips.h>
#include <linux/gfp.h>
#include <linux/kprobes.h>
#include <asm/uaccess.h>
#include <asm/delay.h>
#include <asm/s390_ext.h>
#include <asm/div64.h>
#include <asm/vdso.h>
#include <asm/irq.h>
#include <asm/irq_regs.h>
#include <asm/timer.h>
#include <asm/etr.h>
#include <asm/cio.h>

/* change this if you have some constant time drift */
#define USECS_PER_JIFFY     ((unsigned long) 1000000/HZ)
#define CLK_TICKS_PER_JIFFY ((unsigned long) USECS_PER_JIFFY << 12)

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);

/*
 * Scheduler clock - returns current time in nanosec units.
 */
unsigned long long notrace __kprobes sched_clock(void)
{
	return (get_clock_monotonic() * 125) >> 9;
}

/*
 * 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 timespec *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;
	set_clock_comparator(S390_lowcore.clock_comparator);
	cd = &__get_cpu_var(comparators);
	cd->event_handler(cd);
}

/*
 * Fixup the clock comparator.
 */
static void fixup_clock_comparator(unsigned long long delta)
{
	/* If nobody is waiting there's nothing to fix. */
	if (S390_lowcore.clock_comparator == -1ULL)
		return;
	S390_lowcore.clock_comparator += delta;
	set_clock_comparator(S390_lowcore.clock_comparator);
}

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

static void s390_set_mode(enum clock_event_mode mode,
			  struct clock_event_device *evt)
{
}

/*
 * 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;
	cd->set_mode		= s390_set_mode;

	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(unsigned int ext_int_code,
				       unsigned int param32,
				       unsigned long param64)
{
	kstat_cpu(smp_processor_id()).irqs[EXTINT_CLK]++;
	if (S390_lowcore.clock_comparator == -1ULL)
		set_clock_comparator(S390_lowcore.clock_comparator);
}

static void etr_timing_alert(struct etr_irq_parm *);
static void stp_timing_alert(struct stp_irq_parm *);

static void timing_alert_interrupt(unsigned int ext_int_code,
				   unsigned int param32, unsigned long param64)
{
	kstat_cpu(smp_processor_id()).irqs[EXTINT_TLA]++;
	if (param32 & 0x00c40000)
		etr_timing_alert((struct etr_irq_parm *) &param32);
	if (param32 & 0x00038000)
		stp_timing_alert((struct stp_irq_parm *) &param32);
}

static void etr_reset(void);
static void stp_reset(void);

void read_persistent_clock(struct timespec *ts)
{
	tod_to_timeval(get_clock() - TOD_UNIX_EPOCH, ts);
}

void read_boot_clock(struct timespec *ts)
{
	tod_to_timeval(sched_clock_base_cc - TOD_UNIX_EPOCH, ts);
}

static cycle_t read_tod_clock(struct clocksource *cs)
{
	return get_clock();
}

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 timespec *wall_time, struct timespec *wtm,
			struct clocksource *clock, u32 mult)
{
	if (clock != &clocksource_tod)
		return;

	/* Make userspace gettimeofday spin until we're done. */
	++vdso_data->tb_update_count;
	smp_wmb();
	vdso_data->xtime_tod_stamp = clock->cycle_last;
	vdso_data->xtime_clock_sec = wall_time->tv_sec;
	vdso_data->xtime_clock_nsec = wall_time->tv_nsec;
	vdso_data->wtom_clock_sec = wtm->tv_sec;
	vdso_data->wtom_clock_nsec = wtm->tv_nsec;
	vdso_data->ntp_mult = mult;
	smp_wmb();
	++vdso_data->tb_update_count;
}

extern struct timezone sys_tz;

void update_vsyscall_tz(void)
{
	/* Make userspace gettimeofday spin until we're done. */
	++vdso_data->tb_update_count;
	smp_wmb();
	vdso_data->tz_minuteswest = sys_tz.tz_minuteswest;
	vdso_data->tz_dsttime = sys_tz.tz_dsttime;
	smp_wmb();
	++vdso_data->tb_update_count;
}

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

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

	/* request the timing alert external interrupt */
	if (register_external_interrupt(0x1406, 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();
}

/*
 * The time is "clock". old is what we think the time is.
 * Adjust the value by a multiple of jiffies and add the delta to ntp.
 * "delay" is an approximation how long the synchronization took. If
 * the time correction is positive, then "delay" is subtracted from
 * the time difference and only the remaining part is passed to ntp.
 */
static unsigned long long adjust_time(unsigned long long old,
				      unsigned long long clock,
				      unsigned long long delay)
{
	unsigned long long delta, ticks;
	struct timex adjust;

	if (clock > old) {
		/* It is later than we thought. */
		delta = ticks = clock - old;
		delta = ticks = (delta < delay) ? 0 : delta - delay;
		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
		adjust.offset = ticks * (1000000 / HZ);
	} else {
		/* It is earlier than we thought. */
		delta = ticks = old - clock;
		delta -= do_div(ticks, CLK_TICKS_PER_JIFFY);
		delta = -delta;
		adjust.offset = -ticks * (1000000 / HZ);
	}
	sched_clock_base_cc += delta;
	if (adjust.offset != 0) {
		pr_notice("The ETR interface has adjusted the clock "
			  "by %li microseconds\n", adjust.offset);
		adjust.modes = ADJ_OFFSET_SINGLESHOT;
		do_adjtimex(&adjust);
	}
	return delta;
}

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_ETR	0
#define CLOCK_SYNC_HAS_STP	1
#define CLOCK_SYNC_ETR		2
#define CLOCK_SYNC_STP		3

/*
 * The synchronous get_clock function. It will write the current clock
 * value to the clock pointer and return 0 if the clock is in sync with
 * the external time source. If the clock mode is local it will return
 * -ENOSYS and -EAGAIN if the clock is not in sync with the external
 * reference.
 */
int get_sync_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_clock();
	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_ETR, &clock_sync_flags) &&
	    !test_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags))
		return -ENOSYS;
	if (!test_bit(CLOCK_SYNC_ETR, &clock_sync_flags) &&
	    !test_bit(CLOCK_SYNC_STP, &clock_sync_flags))
		return -EACCES;
	return -EAGAIN;
}
EXPORT_SYMBOL(get_sync_clock);

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

/*
 * Make get_sync_clock return 0 again.
 * Needs to be called from a context disabled for preemption.
 */
static void enable_sync_clock(void)
{
	atomic_t *sw_ptr = &__get_cpu_var(clock_sync_word);
	atomic_set_mask(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;
}

/* Single threaded workqueue used for etr and 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");
}

/*
 * External Time Reference (ETR) code.
 */
static int etr_port0_online;
static int etr_port1_online;
static int etr_steai_available;

static int __init early_parse_etr(char *p)
{
	if (strncmp(p, "off", 3) == 0)
		etr_port0_online = etr_port1_online = 0;
	else if (strncmp(p, "port0", 5) == 0)
		etr_port0_online = 1;
	else if (strncmp(p, "port1", 5) == 0)
		etr_port1_online = 1;
	else if (strncmp(p, "on", 2) == 0)
		etr_port0_online = etr_port1_online = 1;
	return 0;
}
early_param("etr", early_parse_etr);

enum etr_event {
	ETR_EVENT_PORT0_CHANGE,
	ETR_EVENT_PORT1_CHANGE,
	ETR_EVENT_PORT_ALERT,
	ETR_EVENT_SYNC_CHECK,
	ETR_EVENT_SWITCH_LOCAL,
	ETR_EVENT_UPDATE,
};

/*
 * Valid bit combinations of the eacr register are (x = don't care):
 * e0 e1 dp p0 p1 ea es sl
 *  0  0  x  0	0  0  0  0  initial, disabled state
 *  0  0  x  0	1  1  0  0  port 1 online
 *  0  0  x  1	0  1  0  0  port 0 online
 *  0  0  x  1	1  1  0  0  both ports online
 *  0  1  x  0	1  1  0  0  port 1 online and usable, ETR or PPS mode
 *  0  1  x  0	1  1  0  1  port 1 online, usable and ETR mode
 *  0  1  x  0	1  1  1  0  port 1 online, usable, PPS mode, in-sync
 *  0  1  x  0	1  1  1  1  port 1 online, usable, ETR mode, in-sync
 *  0  1  x  1	1  1  0  0  both ports online, port 1 usable
 *  0  1  x  1	1  1  1  0  both ports online, port 1 usable, PPS mode, in-sync
 *  0  1  x  1	1  1  1  1  both ports online, port 1 usable, ETR mode, in-sync
 *  1  0  x  1	0  1  0  0  port 0 online and usable, ETR or PPS mode
 *  1  0  x  1	0  1  0  1  port 0 online, usable and ETR mode
 *  1  0  x  1	0  1  1  0  port 0 online, usable, PPS mode, in-sync
 *  1  0  x  1	0  1  1  1  port 0 online, usable, ETR mode, in-sync
 *  1  0  x  1	1  1  0  0  both ports online, port 0 usable
 *  1  0  x  1	1  1  1  0  both ports online, port 0 usable, PPS mode, in-sync
 *  1  0  x  1	1  1  1  1  both ports online, port 0 usable, ETR mode, in-sync
 *  1  1  x  1	1  1  1  0  both ports online & usable, ETR, in-sync
 *  1  1  x  1	1  1  1  1  both ports online & usable, ETR, in-sync
 */
static struct etr_eacr etr_eacr;
static u64 etr_tolec;			/* time of last eacr update */
static struct etr_aib etr_port0;
static int etr_port0_uptodate;
static struct etr_aib etr_port1;
static int etr_port1_uptodate;
static unsigned long etr_events;
static struct timer_list etr_timer;

static void etr_timeout(unsigned long dummy);
static void etr_work_fn(struct work_struct *work);
static DEFINE_MUTEX(etr_work_mutex);
static DECLARE_WORK(etr_work, etr_work_fn);

/*
 * Reset ETR attachment.
 */
static void etr_reset(void)
{
	etr_eacr =  (struct etr_eacr) {
		.e0 = 0, .e1 = 0, ._pad0 = 4, .dp = 0,
		.p0 = 0, .p1 = 0, ._pad1 = 0, .ea = 0,
		.es = 0, .sl = 0 };
	if (etr_setr(&etr_eacr) == 0) {
		etr_tolec = get_clock();
		set_bit(CLOCK_SYNC_HAS_ETR, &clock_sync_flags);
		if (etr_port0_online && etr_port1_online)
			set_bit(CLOCK_SYNC_ETR, &clock_sync_flags);
	} else if (etr_port0_online || etr_port1_online) {
		pr_warning("The real or virtual hardware system does "
			   "not provide an ETR interface\n");
		etr_port0_online = etr_port1_online = 0;
	}
}

static int __init etr_init(void)
{
	struct etr_aib aib;

	if (!test_bit(CLOCK_SYNC_HAS_ETR, &clock_sync_flags))
		return 0;
	time_init_wq();
	/* Check if this machine has the steai instruction. */
	if (etr_steai(&aib, ETR_STEAI_STEPPING_PORT) == 0)
		etr_steai_available = 1;
	setup_timer(&etr_timer, etr_timeout, 0UL);
	if (etr_port0_online) {
		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
		queue_work(time_sync_wq, &etr_work);
	}
	if (etr_port1_online) {
		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
		queue_work(time_sync_wq, &etr_work);
	}
	return 0;
}

arch_initcall(etr_init);

/*
 * Two sorts of ETR machine checks. The architecture reads:
 * "When a machine-check niterruption occurs and if a switch-to-local or
 *  ETR-sync-check interrupt request is pending but disabled, this pending
 *  disabled interruption request is indicated and is cleared".
 * Which means that we can get etr_switch_to_local events from the machine
 * check handler although the interruption condition is disabled. Lovely..
 */

/*
 * Switch to local machine check. This is called when the last usable
 * ETR port goes inactive. After switch to local the clock is not in sync.
 */
void etr_switch_to_local(void)
{
	if (!etr_eacr.sl)
		return;
	disable_sync_clock(NULL);
	if (!test_and_set_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events)) {
		etr_eacr.es = etr_eacr.sl = 0;
		etr_setr(&etr_eacr);
		queue_work(time_sync_wq, &etr_work);
	}
}

/*
 * ETR sync check machine check. This is called when the ETR OTE and the
 * local clock OTE are farther apart than the ETR sync check tolerance.
 * After a ETR sync check the clock is not in sync. The machine check
 * is broadcasted to all cpus at the same time.
 */
void etr_sync_check(void)
{
	if (!etr_eacr.es)
		return;
	disable_sync_clock(NULL);
	if (!test_and_set_bit(ETR_EVENT_SYNC_CHECK, &etr_events)) {
		etr_eacr.es = 0;
		etr_setr(&etr_eacr);
		queue_work(time_sync_wq, &etr_work);
	}
}

/*
 * ETR timing alert. There are two causes:
 * 1) port state change, check the usability of the port
 * 2) port alert, one of the ETR-data-validity bits (v1-v2 bits of the
 *    sldr-status word) or ETR-data word 1 (edf1) or ETR-data word 3 (edf3)
 *    or ETR-data word 4 (edf4) has changed.
 */
static void etr_timing_alert(struct etr_irq_parm *intparm)
{
	if (intparm->pc0)
		/* ETR port 0 state change. */
		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
	if (intparm->pc1)
		/* ETR port 1 state change. */
		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
	if (intparm->eai)
		/*
		 * ETR port alert on either port 0, 1 or both.
		 * Both ports are not up-to-date now.
		 */
		set_bit(ETR_EVENT_PORT_ALERT, &etr_events);
	queue_work(time_sync_wq, &etr_work);
}

static void etr_timeout(unsigned long dummy)
{
	set_bit(ETR_EVENT_UPDATE, &etr_events);
	queue_work(time_sync_wq, &etr_work);
}

/*
 * Check if the etr mode is pss.
 */
static inline int etr_mode_is_pps(struct etr_eacr eacr)
{
	return eacr.es && !eacr.sl;
}

/*
 * Check if the etr mode is etr.
 */
static inline int etr_mode_is_etr(struct etr_eacr eacr)
{
	return eacr.es && eacr.sl;
}

/*
 * Check if the port can be used for TOD synchronization.
 * For PPS mode the port has to receive OTEs. For ETR mode
 * the port has to receive OTEs, the ETR stepping bit has to
 * be zero and the validity bits for data frame 1, 2, and 3
 * have to be 1.
 */
static int etr_port_valid(struct etr_aib *aib, int port)
{
	unsigned int psc;

	/* Check that this port is receiving OTEs. */
	if (aib->tsp == 0)
		return 0;

	psc = port ? aib->esw.psc1 : aib->esw.psc0;
	if (psc == etr_lpsc_pps_mode)
		return 1;
	if (psc == etr_lpsc_operational_step)
		return !aib->esw.y && aib->slsw.v1 &&
			aib->slsw.v2 && aib->slsw.v3;
	return 0;
}

/*
 * Check if two ports are on the same network.
 */
static int etr_compare_network(struct etr_aib *aib1, struct etr_aib *aib2)
{
	// FIXME: any other fields we have to compare?
	return aib1->edf1.net_id == aib2->edf1.net_id;
}

/*
 * Wrapper for etr_stei that converts physical port states
 * to logical port states to be consistent with the output
 * of stetr (see etr_psc vs. etr_lpsc).
 */
static void etr_steai_cv(struct etr_aib *aib, unsigned int func)
{
	BUG_ON(etr_steai(aib, func) != 0);
	/* Convert port state to logical port state. */
	if (aib->esw.psc0 == 1)
		aib->esw.psc0 = 2;
	else if (aib->esw.psc0 == 0 && aib->esw.p == 0)
		aib->esw.psc0 = 1;
	if (aib->esw.psc1 == 1)
		aib->esw.psc1 = 2;
	else if (aib->esw.psc1 == 0 && aib->esw.p == 1)
		aib->esw.psc1 = 1;
}

/*
 * Check if the aib a2 is still connected to the same attachment as
 * aib a1, the etv values differ by one and a2 is valid.
 */
static int etr_aib_follows(struct etr_aib *a1, struct etr_aib *a2, int p)
{
	int state_a1, state_a2;

	/* Paranoia check: e0/e1 should better be the same. */
	if (a1->esw.eacr.e0 != a2->esw.eacr.e0 ||
	    a1->esw.eacr.e1 != a2->esw.eacr.e1)
		return 0;

	/* Still connected to the same etr ? */
	state_a1 = p ? a1->esw.psc1 : a1->esw.psc0;
	state_a2 = p ? a2->esw.psc1 : a2->esw.psc0;
	if (state_a1 == etr_lpsc_operational_step) {
		if (state_a2 != etr_lpsc_operational_step ||
		    a1->edf1.net_id != a2->edf1.net_id ||
		    a1->edf1.etr_id != a2->edf1.etr_id ||
		    a1->edf1.etr_pn != a2->edf1.etr_pn)
			return 0;
	} else if (state_a2 != etr_lpsc_pps_mode)
		return 0;

	/* The ETV value of a2 needs to be ETV of a1 + 1. */
	if (a1->edf2.etv + 1 != a2->edf2.etv)
		return 0;

	if (!etr_port_valid(a2, p))
		return 0;

	return 1;
}

struct clock_sync_data {
	atomic_t cpus;
	int in_sync;
	unsigned long long fixup_cc;
	int etr_port;
	struct etr_aib *etr_aib;
};

static void clock_sync_cpu(struct clock_sync_data *sync)
{
	atomic_dec(&sync->cpus);
	enable_sync_clock();
	/*
	 * This looks like a busy wait loop but it isn't. etr_sync_cpus
	 * is called on all other cpus while the TOD clocks is stopped.
	 * __udelay will stop the cpu on an enabled wait psw until the
	 * TOD is running again.
	 */
	while (sync->in_sync == 0) {
		__udelay(1);
		/*
		 * A different cpu changes *in_sync. Therefore use
		 * barrier() to force memory access.
		 */
		barrier();
	}
	if (sync->in_sync != 1)
		/* Didn't work. Clear per-cpu in sync bit again. */
		disable_sync_clock(NULL);
	/*
	 * This round of TOD syncing is done. Set the clock comparator
	 * to the next tick and let the processor continue.
	 */
	fixup_clock_comparator(sync->fixup_cc);
}

/*
 * Sync the TOD clock using the port referred to by aibp. This port
 * has to be enabled and the other port has to be disabled. The
 * last eacr update has to be more than 1.6 seconds in the past.
 */
static int etr_sync_clock(void *data)
{
	static int first;
	unsigned long long clock, old_clock, delay, delta;
	struct clock_sync_data *etr_sync;
	struct etr_aib *sync_port, *aib;
	int port;
	int rc;

	etr_sync = data;

	if (xchg(&first, 1) == 1) {
		/* Slave */
		clock_sync_cpu(etr_sync);
		return 0;
	}

	/* Wait until all other cpus entered the sync function. */
	while (atomic_read(&etr_sync->cpus) != 0)
		cpu_relax();

	port = etr_sync->etr_port;
	aib = etr_sync->etr_aib;
	sync_port = (port == 0) ? &etr_port0 : &etr_port1;
	enable_sync_clock();

	/* Set clock to next OTE. */
	__ctl_set_bit(14, 21);
	__ctl_set_bit(0, 29);
	clock = ((unsigned long long) (aib->edf2.etv + 1)) << 32;
	old_clock = get_clock();
	if (set_clock(clock) == 0) {
		__udelay(1);	/* Wait for the clock to start. */
		__ctl_clear_bit(0, 29);
		__ctl_clear_bit(14, 21);
		etr_stetr(aib);
		/* Adjust Linux timing variables. */
		delay = (unsigned long long)
			(aib->edf2.etv - sync_port->edf2.etv) << 32;
		delta = adjust_time(old_clock, clock, delay);
		etr_sync->fixup_cc = delta;
		fixup_clock_comparator(delta);
		/* Verify that the clock is properly set. */
		if (!etr_aib_follows(sync_port, aib, port)) {
			/* Didn't work. */
			disable_sync_clock(NULL);
			etr_sync->in_sync = -EAGAIN;
			rc = -EAGAIN;
		} else {
			etr_sync->in_sync = 1;
			rc = 0;
		}
	} else {
		/* Could not set the clock ?!? */
		__ctl_clear_bit(0, 29);
		__ctl_clear_bit(14, 21);
		disable_sync_clock(NULL);
		etr_sync->in_sync = -EAGAIN;
		rc = -EAGAIN;
	}
	xchg(&first, 0);
	return rc;
}

static int etr_sync_clock_stop(struct etr_aib *aib, int port)
{
	struct clock_sync_data etr_sync;
	struct etr_aib *sync_port;
	int follows;
	int rc;

	/* Check if the current aib is adjacent to the sync port aib. */
	sync_port = (port == 0) ? &etr_port0 : &etr_port1;
	follows = etr_aib_follows(sync_port, aib, port);
	memcpy(sync_port, aib, sizeof(*aib));
	if (!follows)
		return -EAGAIN;
	memset(&etr_sync, 0, sizeof(etr_sync));
	etr_sync.etr_aib = aib;
	etr_sync.etr_port = port;
	get_online_cpus();
	atomic_set(&etr_sync.cpus, num_online_cpus() - 1);
	rc = stop_machine(etr_sync_clock, &etr_sync, cpu_online_mask);
	put_online_cpus();
	return rc;
}

/*
 * Handle the immediate effects of the different events.
 * The port change event is used for online/offline changes.
 */
static struct etr_eacr etr_handle_events(struct etr_eacr eacr)
{
	if (test_and_clear_bit(ETR_EVENT_SYNC_CHECK, &etr_events))
		eacr.es = 0;
	if (test_and_clear_bit(ETR_EVENT_SWITCH_LOCAL, &etr_events))
		eacr.es = eacr.sl = 0;
	if (test_and_clear_bit(ETR_EVENT_PORT_ALERT, &etr_events))
		etr_port0_uptodate = etr_port1_uptodate = 0;

	if (test_and_clear_bit(ETR_EVENT_PORT0_CHANGE, &etr_events)) {
		if (eacr.e0)
			/*
			 * Port change of an enabled port. We have to
			 * assume that this can have caused an stepping
			 * port switch.
			 */
			etr_tolec = get_clock();
		eacr.p0 = etr_port0_online;
		if (!eacr.p0)
			eacr.e0 = 0;
		etr_port0_uptodate = 0;
	}
	if (test_and_clear_bit(ETR_EVENT_PORT1_CHANGE, &etr_events)) {
		if (eacr.e1)
			/*
			 * Port change of an enabled port. We have to
			 * assume that this can have caused an stepping
			 * port switch.
			 */
			etr_tolec = get_clock();
		eacr.p1 = etr_port1_online;
		if (!eacr.p1)
			eacr.e1 = 0;
		etr_port1_uptodate = 0;
	}
	clear_bit(ETR_EVENT_UPDATE, &etr_events);
	return eacr;
}

/*
 * Set up a timer that expires after the etr_tolec + 1.6 seconds if
 * one of the ports needs an update.
 */
static void etr_set_tolec_timeout(unsigned long long now)
{
	unsigned long micros;

	if ((!etr_eacr.p0 || etr_port0_uptodate) &&
	    (!etr_eacr.p1 || etr_port1_uptodate))
		return;
	micros = (now > etr_tolec) ? ((now - etr_tolec) >> 12) : 0;
	micros = (micros > 1600000) ? 0 : 1600000 - micros;
	mod_timer(&etr_timer, jiffies + (micros * HZ) / 1000000 + 1);
}

/*
 * Set up a time that expires after 1/2 second.
 */
static void etr_set_sync_timeout(void)
{
	mod_timer(&etr_timer, jiffies + HZ/2);
}

/*
 * Update the aib information for one or both ports.
 */
static struct etr_eacr etr_handle_update(struct etr_aib *aib,
					 struct etr_eacr eacr)
{
	/* With both ports disabled the aib information is useless. */
	if (!eacr.e0 && !eacr.e1)
		return eacr;

	/* Update port0 or port1 with aib stored in etr_work_fn. */
	if (aib->esw.q == 0) {
		/* Information for port 0 stored. */
		if (eacr.p0 && !etr_port0_uptodate) {
			etr_port0 = *aib;
			if (etr_port0_online)
				etr_port0_uptodate = 1;
		}
	} else {
		/* Information for port 1 stored. */
		if (eacr.p1 && !etr_port1_uptodate) {
			etr_port1 = *aib;
			if (etr_port0_online)
				etr_port1_uptodate = 1;
		}
	}

	/*
	 * Do not try to get the alternate port aib if the clock
	 * is not in sync yet.
	 */
	if (!eacr.es || !check_sync_clock())
		return eacr;

	/*
	 * If steai is available we can get the information about
	 * the other port immediately. If only stetr is available the
	 * data-port bit toggle has to be used.
	 */
	if (etr_steai_available) {
		if (eacr.p0 && !etr_port0_uptodate) {
			etr_steai_cv(&etr_port0, ETR_STEAI_PORT_0);
			etr_port0_uptodate = 1;
		}
		if (eacr.p1 && !etr_port1_uptodate) {
			etr_steai_cv(&etr_port1, ETR_STEAI_PORT_1);
			etr_port1_uptodate = 1;
		}
	} else {
		/*
		 * One port was updated above, if the other
		 * port is not uptodate toggle dp bit.
		 */
		if ((eacr.p0 && !etr_port0_uptodate) ||
		    (eacr.p1 && !etr_port1_uptodate))
			eacr.dp ^= 1;
		else
			eacr.dp = 0;
	}
	return eacr;
}

/*
 * Write new etr control register if it differs from the current one.
 * Return 1 if etr_tolec has been updated as well.
 */
static void etr_update_eacr(struct etr_eacr eacr)
{
	int dp_changed;

	if (memcmp(&etr_eacr, &eacr, sizeof(eacr)) == 0)
		/* No change, return. */
		return;
	/*
	 * The disable of an active port of the change of the data port
	 * bit can/will cause a change in the data port.
	 */
	dp_changed = etr_eacr.e0 > eacr.e0 || etr_eacr.e1 > eacr.e1 ||
		(etr_eacr.dp ^ eacr.dp) != 0;
	etr_eacr = eacr;
	etr_setr(&etr_eacr);
	if (dp_changed)
		etr_tolec = get_clock();
}

/*
 * ETR work. In this function you'll find the main logic. In
 * particular this is the only function that calls etr_update_eacr(),
 * it "controls" the etr control register.
 */
static void etr_work_fn(struct work_struct *work)
{
	unsigned long long now;
	struct etr_eacr eacr;
	struct etr_aib aib;
	int sync_port;

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

	/* Create working copy of etr_eacr. */
	eacr = etr_eacr;

	/* Check for the different events and their immediate effects. */
	eacr = etr_handle_events(eacr);

	/* Check if ETR is supposed to be active. */
	eacr.ea = eacr.p0 || eacr.p1;
	if (!eacr.ea) {
		/* Both ports offline. Reset everything. */
		eacr.dp = eacr.es = eacr.sl = 0;
		on_each_cpu(disable_sync_clock, NULL, 1);
		del_timer_sync(&etr_timer);
		etr_update_eacr(eacr);
		goto out_unlock;
	}

	/* Store aib to get the current ETR status word. */
	BUG_ON(etr_stetr(&aib) != 0);
	etr_port0.esw = etr_port1.esw = aib.esw;	/* Copy status word. */
	now = get_clock();

	/*
	 * Update the port information if the last stepping port change
	 * or data port change is older than 1.6 seconds.
	 */
	if (now >= etr_tolec + (1600000 << 12))
		eacr = etr_handle_update(&aib, eacr);

	/*
	 * Select ports to enable. The preferred synchronization mode is PPS.
	 * If a port can be enabled depends on a number of things:
	 * 1) The port needs to be online and uptodate. A port is not
	 *    disabled just because it is not uptodate, but it is only
	 *    enabled if it is uptodate.
	 * 2) The port needs to have the same mode (pps / etr).
	 * 3) The port needs to be usable -> etr_port_valid() == 1
	 * 4) To enable the second port the clock needs to be in sync.
	 * 5) If both ports are useable and are ETR ports, the network id
	 *    has to be the same.
	 * The eacr.sl bit is used to indicate etr mode vs. pps mode.
	 */
	if (eacr.p0 && aib.esw.psc0 == etr_lpsc_pps_mode) {
		eacr.sl = 0;
		eacr.e0 = 1;
		if (!etr_mode_is_pps(etr_eacr))
			eacr.es = 0;
		if (!eacr.es || !eacr.p1 || aib.esw.psc1 != etr_lpsc_pps_mode)
			eacr.e1 = 0;
		// FIXME: uptodate checks ?
		else if (etr_port0_uptodate && etr_port1_uptodate)
			eacr.e1 = 1;
		sync_port = (etr_port0_uptodate &&
			     etr_port_valid(&etr_port0, 0)) ? 0 : -1;
	} else if (eacr.p1 && aib.esw.psc1 == etr_lpsc_pps_mode) {
		eacr.sl = 0;
		eacr.e0 = 0;
		eacr.e1 = 1;
		if (!etr_mode_is_pps(etr_eacr))
			eacr.es = 0;
		sync_port = (etr_port1_uptodate &&
			     etr_port_valid(&etr_port1, 1)) ? 1 : -1;
	} else if (eacr.p0 && aib.esw.psc0 == etr_lpsc_operational_step) {
		eacr.sl = 1;
		eacr.e0 = 1;
		if (!etr_mode_is_etr(etr_eacr))
			eacr.es = 0;
		if (!eacr.es || !eacr.p1 ||
		    aib.esw.psc1 != etr_lpsc_operational_alt)
			eacr.e1 = 0;
		else if (etr_port0_uptodate && etr_port1_uptodate &&
			 etr_compare_network(&etr_port0, &etr_port1))
			eacr.e1 = 1;
		sync_port = (etr_port0_uptodate &&
			     etr_port_valid(&etr_port0, 0)) ? 0 : -1;
	} else if (eacr.p1 && aib.esw.psc1 == etr_lpsc_operational_step) {
		eacr.sl = 1;
		eacr.e0 = 0;
		eacr.e1 = 1;
		if (!etr_mode_is_etr(etr_eacr))
			eacr.es = 0;
		sync_port = (etr_port1_uptodate &&
			     etr_port_valid(&etr_port1, 1)) ? 1 : -1;
	} else {
		/* Both ports not usable. */
		eacr.es = eacr.sl = 0;
		sync_port = -1;
	}

	/*
	 * If the clock is in sync just update the eacr and return.
	 * If there is no valid sync port wait for a port update.
	 */
	if ((eacr.es && check_sync_clock()) || sync_port < 0) {
		etr_update_eacr(eacr);
		etr_set_tolec_timeout(now);
		goto out_unlock;
	}

	/*
	 * Prepare control register for clock syncing
	 * (reset data port bit, set sync check control.
	 */
	eacr.dp = 0;
	eacr.es = 1;

	/*
	 * Update eacr and try to synchronize the clock. If the update
	 * of eacr caused a stepping port switch (or if we have to
	 * assume that a stepping port switch has occurred) or the
	 * clock syncing failed, reset the sync check control bit
	 * and set up a timer to try again after 0.5 seconds
	 */
	etr_update_eacr(eacr);
	if (now < etr_tolec + (1600000 << 12) ||
	    etr_sync_clock_stop(&aib, sync_port) != 0) {
		/* Sync failed. Try again in 1/2 second. */
		eacr.es = 0;
		etr_update_eacr(eacr);
		etr_set_sync_timeout();
	} else
		etr_set_tolec_timeout(now);
out_unlock:
	mutex_unlock(&etr_work_mutex);
}

/*
 * Sysfs interface functions
 */
static struct sysdev_class etr_sysclass = {
	.name	= "etr",
};

static struct sys_device etr_port0_dev = {
	.id	= 0,
	.cls	= &etr_sysclass,
};

static struct sys_device etr_port1_dev = {
	.id	= 1,
	.cls	= &etr_sysclass,
};

/*
 * ETR class attributes
 */
static ssize_t etr_stepping_port_show(struct sysdev_class *class,
					struct sysdev_class_attribute *attr,
					char *buf)
{
	return sprintf(buf, "%i\n", etr_port0.esw.p);
}

static SYSDEV_CLASS_ATTR(stepping_port, 0400, etr_stepping_port_show, NULL);

static ssize_t etr_stepping_mode_show(struct sysdev_class *class,
				      	struct sysdev_class_attribute *attr,
					char *buf)
{
	char *mode_str;

	if (etr_mode_is_pps(etr_eacr))
		mode_str = "pps";
	else if (etr_mode_is_etr(etr_eacr))
		mode_str = "etr";
	else
		mode_str = "local";
	return sprintf(buf, "%s\n", mode_str);
}

static SYSDEV_CLASS_ATTR(stepping_mode, 0400, etr_stepping_mode_show, NULL);

/*
 * ETR port attributes
 */
static inline struct etr_aib *etr_aib_from_dev(struct sys_device *dev)
{
	if (dev == &etr_port0_dev)
		return etr_port0_online ? &etr_port0 : NULL;
	else
		return etr_port1_online ? &etr_port1 : NULL;
}

static ssize_t etr_online_show(struct sys_device *dev,
				struct sysdev_attribute *attr,
				char *buf)
{
	unsigned int online;

	online = (dev == &etr_port0_dev) ? etr_port0_online : etr_port1_online;
	return sprintf(buf, "%i\n", online);
}

static ssize_t etr_online_store(struct sys_device *dev,
				struct sysdev_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_ETR, &clock_sync_flags))
		return -EOPNOTSUPP;
	mutex_lock(&clock_sync_mutex);
	if (dev == &etr_port0_dev) {
		if (etr_port0_online == value)
			goto out;	/* Nothing to do. */
		etr_port0_online = value;
		if (etr_port0_online && etr_port1_online)
			set_bit(CLOCK_SYNC_ETR, &clock_sync_flags);
		else
			clear_bit(CLOCK_SYNC_ETR, &clock_sync_flags);
		set_bit(ETR_EVENT_PORT0_CHANGE, &etr_events);
		queue_work(time_sync_wq, &etr_work);
	} else {
		if (etr_port1_online == value)
			goto out;	/* Nothing to do. */
		etr_port1_online = value;
		if (etr_port0_online && etr_port1_online)
			set_bit(CLOCK_SYNC_ETR, &clock_sync_flags);
		else
			clear_bit(CLOCK_SYNC_ETR, &clock_sync_flags);
		set_bit(ETR_EVENT_PORT1_CHANGE, &etr_events);
		queue_work(time_sync_wq, &etr_work);
	}
out:
	mutex_unlock(&clock_sync_mutex);
	return count;
}

static SYSDEV_ATTR(online, 0600, etr_online_show, etr_online_store);

static ssize_t etr_stepping_control_show(struct sys_device *dev,
					struct sysdev_attribute *attr,
					char *buf)
{
	return sprintf(buf, "%i\n", (dev == &etr_port0_dev) ?
		       etr_eacr.e0 : etr_eacr.e1);
}

static SYSDEV_ATTR(stepping_control, 0400, etr_stepping_control_show, NULL);

static ssize_t etr_mode_code_show(struct sys_device *dev,
				struct sysdev_attribute *attr, char *buf)
{
	if (!etr_port0_online && !etr_port1_online)
		/* Status word is not uptodate if both ports are offline. */
		return -ENODATA;
	return sprintf(buf, "%i\n", (dev == &etr_port0_dev) ?
		       etr_port0.esw.psc0 : etr_port0.esw.psc1);
}

static SYSDEV_ATTR(state_code, 0400, etr_mode_code_show, NULL);

static ssize_t etr_untuned_show(struct sys_device *dev,
				struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v1)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf1.u);
}

static SYSDEV_ATTR(untuned, 0400, etr_untuned_show, NULL);

static ssize_t etr_network_id_show(struct sys_device *dev,
				struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v1)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf1.net_id);
}

static SYSDEV_ATTR(network, 0400, etr_network_id_show, NULL);

static ssize_t etr_id_show(struct sys_device *dev,
			struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v1)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf1.etr_id);
}

static SYSDEV_ATTR(id, 0400, etr_id_show, NULL);

static ssize_t etr_port_number_show(struct sys_device *dev,
			struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v1)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf1.etr_pn);
}

static SYSDEV_ATTR(port, 0400, etr_port_number_show, NULL);

static ssize_t etr_coupled_show(struct sys_device *dev,
			struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v3)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf3.c);
}

static SYSDEV_ATTR(coupled, 0400, etr_coupled_show, NULL);

static ssize_t etr_local_time_show(struct sys_device *dev,
			struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v3)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf3.blto);
}

static SYSDEV_ATTR(local_time, 0400, etr_local_time_show, NULL);

static ssize_t etr_utc_offset_show(struct sys_device *dev,
			struct sysdev_attribute *attr, char *buf)
{
	struct etr_aib *aib = etr_aib_from_dev(dev);

	if (!aib || !aib->slsw.v3)
		return -ENODATA;
	return sprintf(buf, "%i\n", aib->edf3.buo);
}

static SYSDEV_ATTR(utc_offset, 0400, etr_utc_offset_show, NULL);

static struct sysdev_attribute *etr_port_attributes[] = {
	&attr_online,
	&attr_stepping_control,
	&attr_state_code,
	&attr_untuned,
	&attr_network,
	&attr_id,
	&attr_port,
	&attr_coupled,
	&attr_local_time,
	&attr_utc_offset,
	NULL
};

static int __init etr_register_port(struct sys_device *dev)
{
	struct sysdev_attribute **attr;
	int rc;

	rc = sysdev_register(dev);
	if (rc)
		goto out;
	for (attr = etr_port_attributes; *attr; attr++) {
		rc = sysdev_create_file(dev, *attr);
		if (rc)
			goto out_unreg;
	}
	return 0;
out_unreg:
	for (; attr >= etr_port_attributes; attr--)
		sysdev_remove_file(dev, *attr);
	sysdev_unregister(dev);
out:
	return rc;
}

static void __init etr_unregister_port(struct sys_device *dev)
{
	struct sysdev_attribute **attr;

	for (attr = etr_port_attributes; *attr; attr++)
		sysdev_remove_file(dev, *attr);
	sysdev_unregister(dev);
}

static int __init etr_init_sysfs(void)
{
	int rc;

	rc = sysdev_class_register(&etr_sysclass);
	if (rc)
		goto out;
	rc = sysdev_class_create_file(&etr_sysclass, &attr_stepping_port);
	if (rc)
		goto out_unreg_class;
	rc = sysdev_class_create_file(&etr_sysclass, &attr_stepping_mode);
	if (rc)
		goto out_remove_stepping_port;
	rc = etr_register_port(&etr_port0_dev);
	if (rc)
		goto out_remove_stepping_mode;
	rc = etr_register_port(&etr_port1_dev);
	if (rc)
		goto out_remove_port0;
	return 0;

out_remove_port0:
	etr_unregister_port(&etr_port0_dev);
out_remove_stepping_mode:
	sysdev_class_remove_file(&etr_sysclass, &attr_stepping_mode);
out_remove_stepping_port:
	sysdev_class_remove_file(&etr_sysclass, &attr_stepping_port);
out_unreg_class:
	sysdev_class_unregister(&etr_sysclass);
out:
	return rc;
}

device_initcall(etr_init_sysfs);

/*
 * Server Time Protocol (STP) code.
 */
static int 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)
{
	if (strncmp(p, "off", 3) == 0)
		stp_online = 0;
	else if (strncmp(p, "on", 2) == 0)
		stp_online = 1;
	return 0;
}
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);
	if (rc == 0)
		set_bit(CLOCK_SYNC_HAS_STP, &clock_sync_flags);
	else if (stp_online) {
		pr_warning("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.
 */
void stp_sync_check(void)
{
	disable_sync_clock(NULL);
	queue_work(time_sync_wq, &stp_work);
}

/*
 * 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.
 */
void stp_island_check(void)
{
	disable_sync_clock(NULL);
	queue_work(time_sync_wq, &stp_work);
}


static int stp_sync_clock(void *data)
{
	static int first;
	unsigned long long old_clock, delta;
	struct clock_sync_data *stp_sync;
	int rc;

	stp_sync = data;

	if (xchg(&first, 1) == 1) {
		/* Slave */
		clock_sync_cpu(stp_sync);
		return 0;
	}

	/* Wait until all other cpus entered the sync function. */
	while (atomic_read(&stp_sync->cpus) != 0)
		cpu_relax();

	enable_sync_clock();

	rc = 0;
	if (stp_info.todoff[0] || stp_info.todoff[1] ||
	    stp_info.todoff[2] || stp_info.todoff[3] ||
	    stp_info.tmd != 2) {
		old_clock = get_clock();
		rc = chsc_sstpc(stp_page, STP_OP_SYNC, 0);
		if (rc == 0) {
			delta = adjust_time(old_clock, get_clock(), 0);
			fixup_clock_comparator(delta);
			rc = chsc_sstpi(stp_page, &stp_info,
					sizeof(struct stp_sstpi));
			if (rc == 0 && stp_info.tmd != 2)
				rc = -EAGAIN;
		}
	}
	if (rc) {
		disable_sync_clock(NULL);
		stp_sync->in_sync = -EAGAIN;
	} else
		stp_sync->in_sync = 1;
	xchg(&first, 0);
	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);
		del_timer_sync(&stp_timer);
		goto out_unlock;
	}

	rc = chsc_sstpc(stp_page, STP_OP_CTRL, 0xb0e0);
	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 class sysfs interface functions
 */
static struct sysdev_class stp_sysclass = {
	.name	= "stp",
};

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

static ssize_t stp_online_show(struct sysdev_class *class,
				struct sysdev_class_attribute *attr,
				char *buf)
{
	return sprintf(buf, "%i\n", stp_online);
}

static ssize_t stp_online_store(struct sysdev_class *class,
				struct sysdev_class_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 SYSDEV_CLASS_ATTR because the attribute should be named
 * stp/online but attr_online already exists in this file ..
 */
static struct sysdev_class_attribute attr_stp_online = {
	.attr = { .name = "online", .mode = 0600 },
	.show	= stp_online_show,
	.store	= stp_online_store,
};

static struct sysdev_class_attribute *stp_attributes[] = {
	&attr_ctn_id,
	&attr_ctn_type,
	&attr_dst_offset,
	&attr_leap_seconds,
	&attr_stp_online,
	&attr_stratum,
	&attr_time_offset,
	&attr_time_zone_offset,
	&attr_timing_mode,
	&attr_timing_state,
	NULL
};

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

	rc = sysdev_class_register(&stp_sysclass);
	if (rc)
		goto out;
	for (attr = stp_attributes; *attr; attr++) {
		rc = sysdev_class_create_file(&stp_sysclass, *attr);
		if (rc)
			goto out_unreg;
	}
	return 0;
out_unreg:
	for (; attr >= stp_attributes; attr--)
		sysdev_class_remove_file(&stp_sysclass, *attr);
	sysdev_class_unregister(&stp_sysclass);
out:
	return rc;
}

device_initcall(stp_init_sysfs);