summaryrefslogtreecommitdiff
path: root/drivers/bus/arm-cci.c
blob: 602bf64fc6908d7f81f0d9dda1b45874144fad36 (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
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
/*
 * CCI cache coherent interconnect driver
 *
 * Copyright (C) 2013 ARM Ltd.
 * Author: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 * This program is distributed "as is" WITHOUT ANY WARRANTY of any
 * kind, whether express or implied; without even the implied warranty
 * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 */

#include <linux/arm-cci.h>
#include <linux/io.h>
#include <linux/interrupt.h>
#include <linux/module.h>
#include <linux/of_address.h>
#include <linux/of_irq.h>
#include <linux/of_platform.h>
#include <linux/perf_event.h>
#include <linux/platform_device.h>
#include <linux/slab.h>
#include <linux/spinlock.h>

#include <asm/cacheflush.h>
#include <asm/smp_plat.h>

static void __iomem *cci_ctrl_base;
static unsigned long cci_ctrl_phys;

#ifdef CONFIG_ARM_CCI400_PORT_CTRL
struct cci_nb_ports {
	unsigned int nb_ace;
	unsigned int nb_ace_lite;
};

static const struct cci_nb_ports cci400_ports = {
	.nb_ace = 2,
	.nb_ace_lite = 3
};

#define CCI400_PORTS_DATA	(&cci400_ports)
#else
#define CCI400_PORTS_DATA	(NULL)
#endif

static const struct of_device_id arm_cci_matches[] = {
#ifdef CONFIG_ARM_CCI400_COMMON
	{.compatible = "arm,cci-400", .data = CCI400_PORTS_DATA },
#endif
	{},
};

#ifdef CONFIG_ARM_CCI400_PMU

#define DRIVER_NAME		"CCI-400"
#define DRIVER_NAME_PMU		DRIVER_NAME " PMU"

#define CCI_PMCR		0x0100
#define CCI_PID2		0x0fe8

#define CCI_PMCR_CEN		0x00000001
#define CCI_PMCR_NCNT_MASK	0x0000f800
#define CCI_PMCR_NCNT_SHIFT	11

#define CCI_PID2_REV_MASK	0xf0
#define CCI_PID2_REV_SHIFT	4

#define CCI_PMU_EVT_SEL		0x000
#define CCI_PMU_CNTR		0x004
#define CCI_PMU_CNTR_CTRL	0x008
#define CCI_PMU_OVRFLW		0x00c

#define CCI_PMU_OVRFLW_FLAG	1

#define CCI_PMU_CNTR_SIZE(model)	((model)->cntr_size)
#define CCI_PMU_CNTR_BASE(model, idx)	((idx) * CCI_PMU_CNTR_SIZE(model))
#define CCI_PMU_CNTR_MASK		((1ULL << 32) -1)
#define CCI_PMU_CNTR_LAST(cci_pmu)	(cci_pmu->num_cntrs - 1)

#define CCI_PMU_EVENT_MASK		0xffUL
#define CCI_PMU_EVENT_SOURCE(event)	((event >> 5) & 0x7)
#define CCI_PMU_EVENT_CODE(event)	(event & 0x1f)

#define CCI_PMU_MAX_HW_CNTRS(model) \
	((model)->num_hw_cntrs + (model)->fixed_hw_cntrs)

/* Types of interfaces that can generate events */
enum {
	CCI_IF_SLAVE,
	CCI_IF_MASTER,
	CCI_IF_MAX,
};

struct event_range {
	u32 min;
	u32 max;
};

struct cci_pmu_hw_events {
	struct perf_event **events;
	unsigned long *used_mask;
	raw_spinlock_t pmu_lock;
};

/*
 * struct cci_pmu_model:
 * @fixed_hw_cntrs - Number of fixed event counters
 * @num_hw_cntrs - Maximum number of programmable event counters
 * @cntr_size - Size of an event counter mapping
 */
struct cci_pmu_model {
	char *name;
	u32 fixed_hw_cntrs;
	u32 num_hw_cntrs;
	u32 cntr_size;
	struct event_range event_ranges[CCI_IF_MAX];
};

static struct cci_pmu_model cci_pmu_models[];

struct cci_pmu {
	void __iomem *base;
	struct pmu pmu;
	int nr_irqs;
	int *irqs;
	unsigned long active_irqs;
	const struct cci_pmu_model *model;
	struct cci_pmu_hw_events hw_events;
	struct platform_device *plat_device;
	int num_cntrs;
	atomic_t active_events;
	struct mutex reserve_mutex;
	struct notifier_block cpu_nb;
	cpumask_t cpus;
};

#define to_cci_pmu(c)	(container_of(c, struct cci_pmu, pmu))

/* Port ids */
#define CCI_PORT_S0	0
#define CCI_PORT_S1	1
#define CCI_PORT_S2	2
#define CCI_PORT_S3	3
#define CCI_PORT_S4	4
#define CCI_PORT_M0	5
#define CCI_PORT_M1	6
#define CCI_PORT_M2	7

#define CCI_REV_R0		0
#define CCI_REV_R1		1
#define CCI_REV_R1_PX		5

/*
 * Instead of an event id to monitor CCI cycles, a dedicated counter is
 * provided. Use 0xff to represent CCI cycles and hope that no future revisions
 * make use of this event in hardware.
 */
enum cci400_perf_events {
	CCI_PMU_CYCLES = 0xff
};

#define CCI_PMU_CYCLE_CNTR_IDX		0
#define CCI_PMU_CNTR0_IDX		1

/*
 * CCI PMU event id is an 8-bit value made of two parts - bits 7:5 for one of 8
 * ports and bits 4:0 are event codes. There are different event codes
 * associated with each port type.
 *
 * Additionally, the range of events associated with the port types changed
 * between Rev0 and Rev1.
 *
 * The constants below define the range of valid codes for each port type for
 * the different revisions and are used to validate the event to be monitored.
 */

#define CCI_REV_R0_SLAVE_PORT_MIN_EV	0x00
#define CCI_REV_R0_SLAVE_PORT_MAX_EV	0x13
#define CCI_REV_R0_MASTER_PORT_MIN_EV	0x14
#define CCI_REV_R0_MASTER_PORT_MAX_EV	0x1a

#define CCI_REV_R1_SLAVE_PORT_MIN_EV	0x00
#define CCI_REV_R1_SLAVE_PORT_MAX_EV	0x14
#define CCI_REV_R1_MASTER_PORT_MIN_EV	0x00
#define CCI_REV_R1_MASTER_PORT_MAX_EV	0x11

static int pmu_validate_hw_event(struct cci_pmu *cci_pmu, unsigned long hw_event)
{
	u8 ev_source = CCI_PMU_EVENT_SOURCE(hw_event);
	u8 ev_code = CCI_PMU_EVENT_CODE(hw_event);
	int if_type;

	if (hw_event & ~CCI_PMU_EVENT_MASK)
		return -ENOENT;

	switch (ev_source) {
	case CCI_PORT_S0:
	case CCI_PORT_S1:
	case CCI_PORT_S2:
	case CCI_PORT_S3:
	case CCI_PORT_S4:
		/* Slave Interface */
		if_type = CCI_IF_SLAVE;
		break;
	case CCI_PORT_M0:
	case CCI_PORT_M1:
	case CCI_PORT_M2:
		/* Master Interface */
		if_type = CCI_IF_MASTER;
		break;
	default:
		return -ENOENT;
	}

	if (ev_code >= cci_pmu->model->event_ranges[if_type].min &&
		ev_code <= cci_pmu->model->event_ranges[if_type].max)
		return hw_event;

	return -ENOENT;
}

static int probe_cci_revision(void)
{
	int rev;
	rev = readl_relaxed(cci_ctrl_base + CCI_PID2) & CCI_PID2_REV_MASK;
	rev >>= CCI_PID2_REV_SHIFT;

	if (rev < CCI_REV_R1_PX)
		return CCI_REV_R0;
	else
		return CCI_REV_R1;
}

static const struct cci_pmu_model *probe_cci_model(struct platform_device *pdev)
{
	if (platform_has_secure_cci_access())
		return &cci_pmu_models[probe_cci_revision()];
	return NULL;
}

static int pmu_is_valid_counter(struct cci_pmu *cci_pmu, int idx)
{
	return 0 <= idx && idx <= CCI_PMU_CNTR_LAST(cci_pmu);
}

static u32 pmu_read_register(struct cci_pmu *cci_pmu, int idx, unsigned int offset)
{
	return readl_relaxed(cci_pmu->base +
			     CCI_PMU_CNTR_BASE(cci_pmu->model, idx) + offset);
}

static void pmu_write_register(struct cci_pmu *cci_pmu, u32 value,
			       int idx, unsigned int offset)
{
	return writel_relaxed(value, cci_pmu->base +
			      CCI_PMU_CNTR_BASE(cci_pmu->model, idx) + offset);
}

static void pmu_disable_counter(struct cci_pmu *cci_pmu, int idx)
{
	pmu_write_register(cci_pmu, 0, idx, CCI_PMU_CNTR_CTRL);
}

static void pmu_enable_counter(struct cci_pmu *cci_pmu, int idx)
{
	pmu_write_register(cci_pmu, 1, idx, CCI_PMU_CNTR_CTRL);
}

static void pmu_set_event(struct cci_pmu *cci_pmu, int idx, unsigned long event)
{
	pmu_write_register(cci_pmu, event, idx, CCI_PMU_EVT_SEL);
}

/*
 * Returns the number of programmable counters actually implemented
 * by the cci
 */
static u32 pmu_get_max_counters(void)
{
	return (readl_relaxed(cci_ctrl_base + CCI_PMCR) &
		CCI_PMCR_NCNT_MASK) >> CCI_PMCR_NCNT_SHIFT;
}

static int pmu_get_event_idx(struct cci_pmu_hw_events *hw, struct perf_event *event)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct hw_perf_event *hw_event = &event->hw;
	unsigned long cci_event = hw_event->config_base;
	int idx;

	if (cci_event == CCI_PMU_CYCLES) {
		if (test_and_set_bit(CCI_PMU_CYCLE_CNTR_IDX, hw->used_mask))
			return -EAGAIN;

		return CCI_PMU_CYCLE_CNTR_IDX;
	}

	for (idx = CCI_PMU_CNTR0_IDX; idx <= CCI_PMU_CNTR_LAST(cci_pmu); ++idx)
		if (!test_and_set_bit(idx, hw->used_mask))
			return idx;

	/* No counters available */
	return -EAGAIN;
}

static int pmu_map_event(struct perf_event *event)
{
	int mapping;
	unsigned long config = event->attr.config;

	if (event->attr.type < PERF_TYPE_MAX)
		return -ENOENT;

	if (config == CCI_PMU_CYCLES)
		mapping = config;
	else
		mapping = pmu_validate_hw_event(to_cci_pmu(event->pmu),
							config);

	return mapping;
}

static int pmu_request_irq(struct cci_pmu *cci_pmu, irq_handler_t handler)
{
	int i;
	struct platform_device *pmu_device = cci_pmu->plat_device;

	if (unlikely(!pmu_device))
		return -ENODEV;

	if (cci_pmu->nr_irqs < 1) {
		dev_err(&pmu_device->dev, "no irqs for CCI PMUs defined\n");
		return -ENODEV;
	}

	/*
	 * Register all available CCI PMU interrupts. In the interrupt handler
	 * we iterate over the counters checking for interrupt source (the
	 * overflowing counter) and clear it.
	 *
	 * This should allow handling of non-unique interrupt for the counters.
	 */
	for (i = 0; i < cci_pmu->nr_irqs; i++) {
		int err = request_irq(cci_pmu->irqs[i], handler, IRQF_SHARED,
				"arm-cci-pmu", cci_pmu);
		if (err) {
			dev_err(&pmu_device->dev, "unable to request IRQ%d for ARM CCI PMU counters\n",
				cci_pmu->irqs[i]);
			return err;
		}

		set_bit(i, &cci_pmu->active_irqs);
	}

	return 0;
}

static void pmu_free_irq(struct cci_pmu *cci_pmu)
{
	int i;

	for (i = 0; i < cci_pmu->nr_irqs; i++) {
		if (!test_and_clear_bit(i, &cci_pmu->active_irqs))
			continue;

		free_irq(cci_pmu->irqs[i], cci_pmu);
	}
}

static u32 pmu_read_counter(struct perf_event *event)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct hw_perf_event *hw_counter = &event->hw;
	int idx = hw_counter->idx;
	u32 value;

	if (unlikely(!pmu_is_valid_counter(cci_pmu, idx))) {
		dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx);
		return 0;
	}
	value = pmu_read_register(cci_pmu, idx, CCI_PMU_CNTR);

	return value;
}

static void pmu_write_counter(struct perf_event *event, u32 value)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct hw_perf_event *hw_counter = &event->hw;
	int idx = hw_counter->idx;

	if (unlikely(!pmu_is_valid_counter(cci_pmu, idx)))
		dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx);
	else
		pmu_write_register(cci_pmu, value, idx, CCI_PMU_CNTR);
}

static u64 pmu_event_update(struct perf_event *event)
{
	struct hw_perf_event *hwc = &event->hw;
	u64 delta, prev_raw_count, new_raw_count;

	do {
		prev_raw_count = local64_read(&hwc->prev_count);
		new_raw_count = pmu_read_counter(event);
	} while (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
		 new_raw_count) != prev_raw_count);

	delta = (new_raw_count - prev_raw_count) & CCI_PMU_CNTR_MASK;

	local64_add(delta, &event->count);

	return new_raw_count;
}

static void pmu_read(struct perf_event *event)
{
	pmu_event_update(event);
}

void pmu_event_set_period(struct perf_event *event)
{
	struct hw_perf_event *hwc = &event->hw;
	/*
	 * The CCI PMU counters have a period of 2^32. To account for the
	 * possiblity of extreme interrupt latency we program for a period of
	 * half that. Hopefully we can handle the interrupt before another 2^31
	 * events occur and the counter overtakes its previous value.
	 */
	u64 val = 1ULL << 31;
	local64_set(&hwc->prev_count, val);
	pmu_write_counter(event, val);
}

static irqreturn_t pmu_handle_irq(int irq_num, void *dev)
{
	unsigned long flags;
	struct cci_pmu *cci_pmu = dev;
	struct cci_pmu_hw_events *events = &cci_pmu->hw_events;
	int idx, handled = IRQ_NONE;

	raw_spin_lock_irqsave(&events->pmu_lock, flags);
	/*
	 * Iterate over counters and update the corresponding perf events.
	 * This should work regardless of whether we have per-counter overflow
	 * interrupt or a combined overflow interrupt.
	 */
	for (idx = CCI_PMU_CYCLE_CNTR_IDX; idx <= CCI_PMU_CNTR_LAST(cci_pmu); idx++) {
		struct perf_event *event = events->events[idx];
		struct hw_perf_event *hw_counter;

		if (!event)
			continue;

		hw_counter = &event->hw;

		/* Did this counter overflow? */
		if (!(pmu_read_register(cci_pmu, idx, CCI_PMU_OVRFLW) &
		      CCI_PMU_OVRFLW_FLAG))
			continue;

		pmu_write_register(cci_pmu, CCI_PMU_OVRFLW_FLAG, idx,
							CCI_PMU_OVRFLW);

		pmu_event_update(event);
		pmu_event_set_period(event);
		handled = IRQ_HANDLED;
	}
	raw_spin_unlock_irqrestore(&events->pmu_lock, flags);

	return IRQ_RETVAL(handled);
}

static int cci_pmu_get_hw(struct cci_pmu *cci_pmu)
{
	int ret = pmu_request_irq(cci_pmu, pmu_handle_irq);
	if (ret) {
		pmu_free_irq(cci_pmu);
		return ret;
	}
	return 0;
}

static void cci_pmu_put_hw(struct cci_pmu *cci_pmu)
{
	pmu_free_irq(cci_pmu);
}

static void hw_perf_event_destroy(struct perf_event *event)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	atomic_t *active_events = &cci_pmu->active_events;
	struct mutex *reserve_mutex = &cci_pmu->reserve_mutex;

	if (atomic_dec_and_mutex_lock(active_events, reserve_mutex)) {
		cci_pmu_put_hw(cci_pmu);
		mutex_unlock(reserve_mutex);
	}
}

static void cci_pmu_enable(struct pmu *pmu)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(pmu);
	struct cci_pmu_hw_events *hw_events = &cci_pmu->hw_events;
	int enabled = bitmap_weight(hw_events->used_mask, cci_pmu->num_cntrs);
	unsigned long flags;
	u32 val;

	if (!enabled)
		return;

	raw_spin_lock_irqsave(&hw_events->pmu_lock, flags);

	/* Enable all the PMU counters. */
	val = readl_relaxed(cci_ctrl_base + CCI_PMCR) | CCI_PMCR_CEN;
	writel(val, cci_ctrl_base + CCI_PMCR);
	raw_spin_unlock_irqrestore(&hw_events->pmu_lock, flags);

}

static void cci_pmu_disable(struct pmu *pmu)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(pmu);
	struct cci_pmu_hw_events *hw_events = &cci_pmu->hw_events;
	unsigned long flags;
	u32 val;

	raw_spin_lock_irqsave(&hw_events->pmu_lock, flags);

	/* Disable all the PMU counters. */
	val = readl_relaxed(cci_ctrl_base + CCI_PMCR) & ~CCI_PMCR_CEN;
	writel(val, cci_ctrl_base + CCI_PMCR);
	raw_spin_unlock_irqrestore(&hw_events->pmu_lock, flags);
}

static void cci_pmu_start(struct perf_event *event, int pmu_flags)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct cci_pmu_hw_events *hw_events = &cci_pmu->hw_events;
	struct hw_perf_event *hwc = &event->hw;
	int idx = hwc->idx;
	unsigned long flags;

	/*
	 * To handle interrupt latency, we always reprogram the period
	 * regardlesss of PERF_EF_RELOAD.
	 */
	if (pmu_flags & PERF_EF_RELOAD)
		WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));

	hwc->state = 0;

	if (unlikely(!pmu_is_valid_counter(cci_pmu, idx))) {
		dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx);
		return;
	}

	raw_spin_lock_irqsave(&hw_events->pmu_lock, flags);

	/* Configure the event to count, unless you are counting cycles */
	if (idx != CCI_PMU_CYCLE_CNTR_IDX)
		pmu_set_event(cci_pmu, idx, hwc->config_base);

	pmu_event_set_period(event);
	pmu_enable_counter(cci_pmu, idx);

	raw_spin_unlock_irqrestore(&hw_events->pmu_lock, flags);
}

static void cci_pmu_stop(struct perf_event *event, int pmu_flags)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct hw_perf_event *hwc = &event->hw;
	int idx = hwc->idx;

	if (hwc->state & PERF_HES_STOPPED)
		return;

	if (unlikely(!pmu_is_valid_counter(cci_pmu, idx))) {
		dev_err(&cci_pmu->plat_device->dev, "Invalid CCI PMU counter %d\n", idx);
		return;
	}

	/*
	 * We always reprogram the counter, so ignore PERF_EF_UPDATE. See
	 * cci_pmu_start()
	 */
	pmu_disable_counter(cci_pmu, idx);
	pmu_event_update(event);
	hwc->state |= PERF_HES_STOPPED | PERF_HES_UPTODATE;
}

static int cci_pmu_add(struct perf_event *event, int flags)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct cci_pmu_hw_events *hw_events = &cci_pmu->hw_events;
	struct hw_perf_event *hwc = &event->hw;
	int idx;
	int err = 0;

	perf_pmu_disable(event->pmu);

	/* If we don't have a space for the counter then finish early. */
	idx = pmu_get_event_idx(hw_events, event);
	if (idx < 0) {
		err = idx;
		goto out;
	}

	event->hw.idx = idx;
	hw_events->events[idx] = event;

	hwc->state = PERF_HES_STOPPED | PERF_HES_UPTODATE;
	if (flags & PERF_EF_START)
		cci_pmu_start(event, PERF_EF_RELOAD);

	/* Propagate our changes to the userspace mapping. */
	perf_event_update_userpage(event);

out:
	perf_pmu_enable(event->pmu);
	return err;
}

static void cci_pmu_del(struct perf_event *event, int flags)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	struct cci_pmu_hw_events *hw_events = &cci_pmu->hw_events;
	struct hw_perf_event *hwc = &event->hw;
	int idx = hwc->idx;

	cci_pmu_stop(event, PERF_EF_UPDATE);
	hw_events->events[idx] = NULL;
	clear_bit(idx, hw_events->used_mask);

	perf_event_update_userpage(event);
}

static int
validate_event(struct pmu *cci_pmu,
               struct cci_pmu_hw_events *hw_events,
               struct perf_event *event)
{
	if (is_software_event(event))
		return 1;

	/*
	 * Reject groups spanning multiple HW PMUs (e.g. CPU + CCI). The
	 * core perf code won't check that the pmu->ctx == leader->ctx
	 * until after pmu->event_init(event).
	 */
	if (event->pmu != cci_pmu)
		return 0;

	if (event->state < PERF_EVENT_STATE_OFF)
		return 1;

	if (event->state == PERF_EVENT_STATE_OFF && !event->attr.enable_on_exec)
		return 1;

	return pmu_get_event_idx(hw_events, event) >= 0;
}

static int
validate_group(struct perf_event *event)
{
	struct perf_event *sibling, *leader = event->group_leader;
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	unsigned long mask[BITS_TO_LONGS(cci_pmu->num_cntrs)];
	struct cci_pmu_hw_events fake_pmu = {
		/*
		 * Initialise the fake PMU. We only need to populate the
		 * used_mask for the purposes of validation.
		 */
		.used_mask = mask,
	};
	memset(mask, 0, BITS_TO_LONGS(cci_pmu->num_cntrs) * sizeof(unsigned long));

	if (!validate_event(event->pmu, &fake_pmu, leader))
		return -EINVAL;

	list_for_each_entry(sibling, &leader->sibling_list, group_entry) {
		if (!validate_event(event->pmu, &fake_pmu, sibling))
			return -EINVAL;
	}

	if (!validate_event(event->pmu, &fake_pmu, event))
		return -EINVAL;

	return 0;
}

static int
__hw_perf_event_init(struct perf_event *event)
{
	struct hw_perf_event *hwc = &event->hw;
	int mapping;

	mapping = pmu_map_event(event);

	if (mapping < 0) {
		pr_debug("event %x:%llx not supported\n", event->attr.type,
			 event->attr.config);
		return mapping;
	}

	/*
	 * We don't assign an index until we actually place the event onto
	 * hardware. Use -1 to signify that we haven't decided where to put it
	 * yet.
	 */
	hwc->idx		= -1;
	hwc->config_base	= 0;
	hwc->config		= 0;
	hwc->event_base		= 0;

	/*
	 * Store the event encoding into the config_base field.
	 */
	hwc->config_base	    |= (unsigned long)mapping;

	/*
	 * Limit the sample_period to half of the counter width. That way, the
	 * new counter value is far less likely to overtake the previous one
	 * unless you have some serious IRQ latency issues.
	 */
	hwc->sample_period  = CCI_PMU_CNTR_MASK >> 1;
	hwc->last_period    = hwc->sample_period;
	local64_set(&hwc->period_left, hwc->sample_period);

	if (event->group_leader != event) {
		if (validate_group(event) != 0)
			return -EINVAL;
	}

	return 0;
}

static int cci_pmu_event_init(struct perf_event *event)
{
	struct cci_pmu *cci_pmu = to_cci_pmu(event->pmu);
	atomic_t *active_events = &cci_pmu->active_events;
	int err = 0;
	int cpu;

	if (event->attr.type != event->pmu->type)
		return -ENOENT;

	/* Shared by all CPUs, no meaningful state to sample */
	if (is_sampling_event(event) || event->attach_state & PERF_ATTACH_TASK)
		return -EOPNOTSUPP;

	/* We have no filtering of any kind */
	if (event->attr.exclude_user	||
	    event->attr.exclude_kernel	||
	    event->attr.exclude_hv	||
	    event->attr.exclude_idle	||
	    event->attr.exclude_host	||
	    event->attr.exclude_guest)
		return -EINVAL;

	/*
	 * Following the example set by other "uncore" PMUs, we accept any CPU
	 * and rewrite its affinity dynamically rather than having perf core
	 * handle cpu == -1 and pid == -1 for this case.
	 *
	 * The perf core will pin online CPUs for the duration of this call and
	 * the event being installed into its context, so the PMU's CPU can't
	 * change under our feet.
	 */
	cpu = cpumask_first(&cci_pmu->cpus);
	if (event->cpu < 0 || cpu < 0)
		return -EINVAL;
	event->cpu = cpu;

	event->destroy = hw_perf_event_destroy;
	if (!atomic_inc_not_zero(active_events)) {
		mutex_lock(&cci_pmu->reserve_mutex);
		if (atomic_read(active_events) == 0)
			err = cci_pmu_get_hw(cci_pmu);
		if (!err)
			atomic_inc(active_events);
		mutex_unlock(&cci_pmu->reserve_mutex);
	}
	if (err)
		return err;

	err = __hw_perf_event_init(event);
	if (err)
		hw_perf_event_destroy(event);

	return err;
}

static ssize_t pmu_cpumask_attr_show(struct device *dev,
				     struct device_attribute *attr, char *buf)
{
	struct dev_ext_attribute *eattr = container_of(attr,
					struct dev_ext_attribute, attr);
	struct cci_pmu *cci_pmu = eattr->var;

	int n = scnprintf(buf, PAGE_SIZE - 1, "%*pbl",
			  cpumask_pr_args(&cci_pmu->cpus));
	buf[n++] = '\n';
	buf[n] = '\0';
	return n;
}

static struct dev_ext_attribute pmu_cpumask_attr = {
	__ATTR(cpumask, S_IRUGO, pmu_cpumask_attr_show, NULL),
	NULL,		/* Populated in cci_pmu_init */
};

static struct attribute *pmu_attrs[] = {
	&pmu_cpumask_attr.attr.attr,
	NULL,
};

static struct attribute_group pmu_attr_group = {
	.attrs = pmu_attrs,
};

static const struct attribute_group *pmu_attr_groups[] = {
	&pmu_attr_group,
	NULL
};

static int cci_pmu_init(struct cci_pmu *cci_pmu, struct platform_device *pdev)
{
	char *name = cci_pmu->model->name;
	u32 num_cntrs;

	pmu_cpumask_attr.var = cci_pmu;
	cci_pmu->pmu = (struct pmu) {
		.name		= cci_pmu->model->name,
		.task_ctx_nr	= perf_invalid_context,
		.pmu_enable	= cci_pmu_enable,
		.pmu_disable	= cci_pmu_disable,
		.event_init	= cci_pmu_event_init,
		.add		= cci_pmu_add,
		.del		= cci_pmu_del,
		.start		= cci_pmu_start,
		.stop		= cci_pmu_stop,
		.read		= pmu_read,
		.attr_groups	= pmu_attr_groups,
	};

	cci_pmu->plat_device = pdev;
	num_cntrs = pmu_get_max_counters();
	if (num_cntrs > cci_pmu->model->num_hw_cntrs) {
		dev_warn(&pdev->dev,
			"PMU implements more counters(%d) than supported by"
			" the model(%d), truncated.",
			num_cntrs, cci_pmu->model->num_hw_cntrs);
		num_cntrs = cci_pmu->model->num_hw_cntrs;
	}
	cci_pmu->num_cntrs = num_cntrs + cci_pmu->model->fixed_hw_cntrs;

	return perf_pmu_register(&cci_pmu->pmu, name, -1);
}

static int cci_pmu_cpu_notifier(struct notifier_block *self,
				unsigned long action, void *hcpu)
{
	struct cci_pmu *cci_pmu = container_of(self,
					struct cci_pmu, cpu_nb);
	unsigned int cpu = (long)hcpu;
	unsigned int target;

	switch (action & ~CPU_TASKS_FROZEN) {
	case CPU_DOWN_PREPARE:
		if (!cpumask_test_and_clear_cpu(cpu, &cci_pmu->cpus))
			break;
		target = cpumask_any_but(cpu_online_mask, cpu);
		if (target < 0) // UP, last CPU
			break;
		/*
		 * TODO: migrate context once core races on event->ctx have
		 * been fixed.
		 */
		cpumask_set_cpu(target, &cci_pmu->cpus);
	default:
		break;
	}

	return NOTIFY_OK;
}

static struct cci_pmu_model cci_pmu_models[] = {
	[CCI_REV_R0] = {
		.name = "CCI_400",
		.fixed_hw_cntrs = 1,	/* Cycle counter */
		.num_hw_cntrs = 4,
		.cntr_size = SZ_4K,
		.event_ranges = {
			[CCI_IF_SLAVE] = {
				CCI_REV_R0_SLAVE_PORT_MIN_EV,
				CCI_REV_R0_SLAVE_PORT_MAX_EV,
			},
			[CCI_IF_MASTER] = {
				CCI_REV_R0_MASTER_PORT_MIN_EV,
				CCI_REV_R0_MASTER_PORT_MAX_EV,
			},
		},
	},
	[CCI_REV_R1] = {
		.name = "CCI_400_r1",
		.fixed_hw_cntrs = 1,	/* Cycle counter */
		.num_hw_cntrs = 4,
		.cntr_size = SZ_4K,
		.event_ranges = {
			[CCI_IF_SLAVE] = {
				CCI_REV_R1_SLAVE_PORT_MIN_EV,
				CCI_REV_R1_SLAVE_PORT_MAX_EV,
			},
			[CCI_IF_MASTER] = {
				CCI_REV_R1_MASTER_PORT_MIN_EV,
				CCI_REV_R1_MASTER_PORT_MAX_EV,
			},
		},
	},
};

static const struct of_device_id arm_cci_pmu_matches[] = {
	{
		.compatible = "arm,cci-400-pmu",
		.data	= NULL,
	},
	{
		.compatible = "arm,cci-400-pmu,r0",
		.data	= &cci_pmu_models[CCI_REV_R0],
	},
	{
		.compatible = "arm,cci-400-pmu,r1",
		.data	= &cci_pmu_models[CCI_REV_R1],
	},
	{},
};

static inline const struct cci_pmu_model *get_cci_model(struct platform_device *pdev)
{
	const struct of_device_id *match = of_match_node(arm_cci_pmu_matches,
							pdev->dev.of_node);
	if (!match)
		return NULL;
	if (match->data)
		return match->data;

	dev_warn(&pdev->dev, "DEPRECATED compatible property,"
			 "requires secure access to CCI registers");
	return probe_cci_model(pdev);
}

static bool is_duplicate_irq(int irq, int *irqs, int nr_irqs)
{
	int i;

	for (i = 0; i < nr_irqs; i++)
		if (irq == irqs[i])
			return true;

	return false;
}

static struct cci_pmu *cci_pmu_alloc(struct platform_device *pdev)
{
	struct cci_pmu *cci_pmu;
	const struct cci_pmu_model *model;

	/*
	 * All allocations are devm_* hence we don't have to free
	 * them explicitly on an error, as it would end up in driver
	 * detach.
	 */
	model = get_cci_model(pdev);
	if (!model) {
		dev_warn(&pdev->dev, "CCI PMU version not supported\n");
		return ERR_PTR(-ENODEV);
	}

	cci_pmu = devm_kzalloc(&pdev->dev, sizeof(*cci_pmu), GFP_KERNEL);
	if (!cci_pmu)
		return ERR_PTR(-ENOMEM);

	cci_pmu->model = model;
	cci_pmu->irqs = devm_kcalloc(&pdev->dev, CCI_PMU_MAX_HW_CNTRS(model),
					sizeof(*cci_pmu->irqs), GFP_KERNEL);
	if (!cci_pmu->irqs)
		return ERR_PTR(-ENOMEM);
	cci_pmu->hw_events.events = devm_kcalloc(&pdev->dev,
					     CCI_PMU_MAX_HW_CNTRS(model),
					     sizeof(*cci_pmu->hw_events.events),
					     GFP_KERNEL);
	if (!cci_pmu->hw_events.events)
		return ERR_PTR(-ENOMEM);
	cci_pmu->hw_events.used_mask = devm_kcalloc(&pdev->dev,
						BITS_TO_LONGS(CCI_PMU_MAX_HW_CNTRS(model)),
						sizeof(*cci_pmu->hw_events.used_mask),
						GFP_KERNEL);
	if (!cci_pmu->hw_events.used_mask)
		return ERR_PTR(-ENOMEM);

	return cci_pmu;
}


static int cci_pmu_probe(struct platform_device *pdev)
{
	struct resource *res;
	struct cci_pmu *cci_pmu;
	int i, ret, irq;

	cci_pmu = cci_pmu_alloc(pdev);
	if (IS_ERR(cci_pmu))
		return PTR_ERR(cci_pmu);

	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
	cci_pmu->base = devm_ioremap_resource(&pdev->dev, res);
	if (IS_ERR(cci_pmu->base))
		return -ENOMEM;

	/*
	 * CCI PMU has one overflow interrupt per counter; but some may be tied
	 * together to a common interrupt.
	 */
	cci_pmu->nr_irqs = 0;
	for (i = 0; i < CCI_PMU_MAX_HW_CNTRS(cci_pmu->model); i++) {
		irq = platform_get_irq(pdev, i);
		if (irq < 0)
			break;

		if (is_duplicate_irq(irq, cci_pmu->irqs, cci_pmu->nr_irqs))
			continue;

		cci_pmu->irqs[cci_pmu->nr_irqs++] = irq;
	}

	/*
	 * Ensure that the device tree has as many interrupts as the number
	 * of counters.
	 */
	if (i < CCI_PMU_MAX_HW_CNTRS(cci_pmu->model)) {
		dev_warn(&pdev->dev, "In-correct number of interrupts: %d, should be %d\n",
			i, CCI_PMU_MAX_HW_CNTRS(cci_pmu->model));
		return -EINVAL;
	}

	raw_spin_lock_init(&cci_pmu->hw_events.pmu_lock);
	mutex_init(&cci_pmu->reserve_mutex);
	atomic_set(&cci_pmu->active_events, 0);
	cpumask_set_cpu(smp_processor_id(), &cci_pmu->cpus);

	cci_pmu->cpu_nb = (struct notifier_block) {
		.notifier_call	= cci_pmu_cpu_notifier,
		/*
		 * to migrate uncore events, our notifier should be executed
		 * before perf core's notifier.
		 */
		.priority	= CPU_PRI_PERF + 1,
	};

	ret = register_cpu_notifier(&cci_pmu->cpu_nb);
	if (ret)
		return ret;

	ret = cci_pmu_init(cci_pmu, pdev);
	if (ret) {
		unregister_cpu_notifier(&cci_pmu->cpu_nb);
		return ret;
	}

	pr_info("ARM %s PMU driver probed", cci_pmu->model->name);
	return 0;
}

static int cci_platform_probe(struct platform_device *pdev)
{
	if (!cci_probed())
		return -ENODEV;

	return of_platform_populate(pdev->dev.of_node, NULL, NULL, &pdev->dev);
}

static struct platform_driver cci_pmu_driver = {
	.driver = {
		   .name = DRIVER_NAME_PMU,
		   .of_match_table = arm_cci_pmu_matches,
		  },
	.probe = cci_pmu_probe,
};

static struct platform_driver cci_platform_driver = {
	.driver = {
		   .name = DRIVER_NAME,
		   .of_match_table = arm_cci_matches,
		  },
	.probe = cci_platform_probe,
};

static int __init cci_platform_init(void)
{
	int ret;

	ret = platform_driver_register(&cci_pmu_driver);
	if (ret)
		return ret;

	return platform_driver_register(&cci_platform_driver);
}

#else /* !CONFIG_ARM_CCI400_PMU */

static int __init cci_platform_init(void)
{
	return 0;
}

#endif /* CONFIG_ARM_CCI400_PMU */

#ifdef CONFIG_ARM_CCI400_PORT_CTRL

#define CCI_PORT_CTRL		0x0
#define CCI_CTRL_STATUS		0xc

#define CCI_ENABLE_SNOOP_REQ	0x1
#define CCI_ENABLE_DVM_REQ	0x2
#define CCI_ENABLE_REQ		(CCI_ENABLE_SNOOP_REQ | CCI_ENABLE_DVM_REQ)

enum cci_ace_port_type {
	ACE_INVALID_PORT = 0x0,
	ACE_PORT,
	ACE_LITE_PORT,
};

struct cci_ace_port {
	void __iomem *base;
	unsigned long phys;
	enum cci_ace_port_type type;
	struct device_node *dn;
};

static struct cci_ace_port *ports;
static unsigned int nb_cci_ports;

struct cpu_port {
	u64 mpidr;
	u32 port;
};

/*
 * Use the port MSB as valid flag, shift can be made dynamic
 * by computing number of bits required for port indexes.
 * Code disabling CCI cpu ports runs with D-cache invalidated
 * and SCTLR bit clear so data accesses must be kept to a minimum
 * to improve performance; for now shift is left static to
 * avoid one more data access while disabling the CCI port.
 */
#define PORT_VALID_SHIFT	31
#define PORT_VALID		(0x1 << PORT_VALID_SHIFT)

static inline void init_cpu_port(struct cpu_port *port, u32 index, u64 mpidr)
{
	port->port = PORT_VALID | index;
	port->mpidr = mpidr;
}

static inline bool cpu_port_is_valid(struct cpu_port *port)
{
	return !!(port->port & PORT_VALID);
}

static inline bool cpu_port_match(struct cpu_port *port, u64 mpidr)
{
	return port->mpidr == (mpidr & MPIDR_HWID_BITMASK);
}

static struct cpu_port cpu_port[NR_CPUS];

/**
 * __cci_ace_get_port - Function to retrieve the port index connected to
 *			a cpu or device.
 *
 * @dn: device node of the device to look-up
 * @type: port type
 *
 * Return value:
 *	- CCI port index if success
 *	- -ENODEV if failure
 */
static int __cci_ace_get_port(struct device_node *dn, int type)
{
	int i;
	bool ace_match;
	struct device_node *cci_portn;

	cci_portn = of_parse_phandle(dn, "cci-control-port", 0);
	for (i = 0; i < nb_cci_ports; i++) {
		ace_match = ports[i].type == type;
		if (ace_match && cci_portn == ports[i].dn)
			return i;
	}
	return -ENODEV;
}

int cci_ace_get_port(struct device_node *dn)
{
	return __cci_ace_get_port(dn, ACE_LITE_PORT);
}
EXPORT_SYMBOL_GPL(cci_ace_get_port);

static void cci_ace_init_ports(void)
{
	int port, cpu;
	struct device_node *cpun;

	/*
	 * Port index look-up speeds up the function disabling ports by CPU,
	 * since the logical to port index mapping is done once and does
	 * not change after system boot.
	 * The stashed index array is initialized for all possible CPUs
	 * at probe time.
	 */
	for_each_possible_cpu(cpu) {
		/* too early to use cpu->of_node */
		cpun = of_get_cpu_node(cpu, NULL);

		if (WARN(!cpun, "Missing cpu device node\n"))
			continue;

		port = __cci_ace_get_port(cpun, ACE_PORT);
		if (port < 0)
			continue;

		init_cpu_port(&cpu_port[cpu], port, cpu_logical_map(cpu));
	}

	for_each_possible_cpu(cpu) {
		WARN(!cpu_port_is_valid(&cpu_port[cpu]),
			"CPU %u does not have an associated CCI port\n",
			cpu);
	}
}
/*
 * Functions to enable/disable a CCI interconnect slave port
 *
 * They are called by low-level power management code to disable slave
 * interfaces snoops and DVM broadcast.
 * Since they may execute with cache data allocation disabled and
 * after the caches have been cleaned and invalidated the functions provide
 * no explicit locking since they may run with D-cache disabled, so normal
 * cacheable kernel locks based on ldrex/strex may not work.
 * Locking has to be provided by BSP implementations to ensure proper
 * operations.
 */

/**
 * cci_port_control() - function to control a CCI port
 *
 * @port: index of the port to setup
 * @enable: if true enables the port, if false disables it
 */
static void notrace cci_port_control(unsigned int port, bool enable)
{
	void __iomem *base = ports[port].base;

	writel_relaxed(enable ? CCI_ENABLE_REQ : 0, base + CCI_PORT_CTRL);
	/*
	 * This function is called from power down procedures
	 * and must not execute any instruction that might
	 * cause the processor to be put in a quiescent state
	 * (eg wfi). Hence, cpu_relax() can not be added to this
	 * read loop to optimize power, since it might hide possibly
	 * disruptive operations.
	 */
	while (readl_relaxed(cci_ctrl_base + CCI_CTRL_STATUS) & 0x1)
			;
}

/**
 * cci_disable_port_by_cpu() - function to disable a CCI port by CPU
 *			       reference
 *
 * @mpidr: mpidr of the CPU whose CCI port should be disabled
 *
 * Disabling a CCI port for a CPU implies disabling the CCI port
 * controlling that CPU cluster. Code disabling CPU CCI ports
 * must make sure that the CPU running the code is the last active CPU
 * in the cluster ie all other CPUs are quiescent in a low power state.
 *
 * Return:
 *	0 on success
 *	-ENODEV on port look-up failure
 */
int notrace cci_disable_port_by_cpu(u64 mpidr)
{
	int cpu;
	bool is_valid;
	for (cpu = 0; cpu < nr_cpu_ids; cpu++) {
		is_valid = cpu_port_is_valid(&cpu_port[cpu]);
		if (is_valid && cpu_port_match(&cpu_port[cpu], mpidr)) {
			cci_port_control(cpu_port[cpu].port, false);
			return 0;
		}
	}
	return -ENODEV;
}
EXPORT_SYMBOL_GPL(cci_disable_port_by_cpu);

/**
 * cci_enable_port_for_self() - enable a CCI port for calling CPU
 *
 * Enabling a CCI port for the calling CPU implies enabling the CCI
 * port controlling that CPU's cluster. Caller must make sure that the
 * CPU running the code is the first active CPU in the cluster and all
 * other CPUs are quiescent in a low power state  or waiting for this CPU
 * to complete the CCI initialization.
 *
 * Because this is called when the MMU is still off and with no stack,
 * the code must be position independent and ideally rely on callee
 * clobbered registers only.  To achieve this we must code this function
 * entirely in assembler.
 *
 * On success this returns with the proper CCI port enabled.  In case of
 * any failure this never returns as the inability to enable the CCI is
 * fatal and there is no possible recovery at this stage.
 */
asmlinkage void __naked cci_enable_port_for_self(void)
{
	asm volatile ("\n"
"	.arch armv7-a\n"
"	mrc	p15, 0, r0, c0, c0, 5	@ get MPIDR value \n"
"	and	r0, r0, #"__stringify(MPIDR_HWID_BITMASK)" \n"
"	adr	r1, 5f \n"
"	ldr	r2, [r1] \n"
"	add	r1, r1, r2		@ &cpu_port \n"
"	add	ip, r1, %[sizeof_cpu_port] \n"

	/* Loop over the cpu_port array looking for a matching MPIDR */
"1:	ldr	r2, [r1, %[offsetof_cpu_port_mpidr_lsb]] \n"
"	cmp	r2, r0 			@ compare MPIDR \n"
"	bne	2f \n"

	/* Found a match, now test port validity */
"	ldr	r3, [r1, %[offsetof_cpu_port_port]] \n"
"	tst	r3, #"__stringify(PORT_VALID)" \n"
"	bne	3f \n"

	/* no match, loop with the next cpu_port entry */
"2:	add	r1, r1, %[sizeof_struct_cpu_port] \n"
"	cmp	r1, ip			@ done? \n"
"	blo	1b \n"

	/* CCI port not found -- cheaply try to stall this CPU */
"cci_port_not_found: \n"
"	wfi \n"
"	wfe \n"
"	b	cci_port_not_found \n"

	/* Use matched port index to look up the corresponding ports entry */
"3:	bic	r3, r3, #"__stringify(PORT_VALID)" \n"
"	adr	r0, 6f \n"
"	ldmia	r0, {r1, r2} \n"
"	sub	r1, r1, r0 		@ virt - phys \n"
"	ldr	r0, [r0, r2] 		@ *(&ports) \n"
"	mov	r2, %[sizeof_struct_ace_port] \n"
"	mla	r0, r2, r3, r0		@ &ports[index] \n"
"	sub	r0, r0, r1		@ virt_to_phys() \n"

	/* Enable the CCI port */
"	ldr	r0, [r0, %[offsetof_port_phys]] \n"
"	mov	r3, %[cci_enable_req]\n"		   
"	str	r3, [r0, #"__stringify(CCI_PORT_CTRL)"] \n"

	/* poll the status reg for completion */
"	adr	r1, 7f \n"
"	ldr	r0, [r1] \n"
"	ldr	r0, [r0, r1]		@ cci_ctrl_base \n"
"4:	ldr	r1, [r0, #"__stringify(CCI_CTRL_STATUS)"] \n"
"	tst	r1, %[cci_control_status_bits] \n"			
"	bne	4b \n"

"	mov	r0, #0 \n"
"	bx	lr \n"

"	.align	2 \n"
"5:	.word	cpu_port - . \n"
"6:	.word	. \n"
"	.word	ports - 6b \n"
"7:	.word	cci_ctrl_phys - . \n"
	: :
	[sizeof_cpu_port] "i" (sizeof(cpu_port)),
	[cci_enable_req] "i" cpu_to_le32(CCI_ENABLE_REQ),
	[cci_control_status_bits] "i" cpu_to_le32(1),
#ifndef __ARMEB__
	[offsetof_cpu_port_mpidr_lsb] "i" (offsetof(struct cpu_port, mpidr)),
#else
	[offsetof_cpu_port_mpidr_lsb] "i" (offsetof(struct cpu_port, mpidr)+4),
#endif
	[offsetof_cpu_port_port] "i" (offsetof(struct cpu_port, port)),
	[sizeof_struct_cpu_port] "i" (sizeof(struct cpu_port)),
	[sizeof_struct_ace_port] "i" (sizeof(struct cci_ace_port)),
	[offsetof_port_phys] "i" (offsetof(struct cci_ace_port, phys)) );

	unreachable();
}

/**
 * __cci_control_port_by_device() - function to control a CCI port by device
 *				    reference
 *
 * @dn: device node pointer of the device whose CCI port should be
 *      controlled
 * @enable: if true enables the port, if false disables it
 *
 * Return:
 *	0 on success
 *	-ENODEV on port look-up failure
 */
int notrace __cci_control_port_by_device(struct device_node *dn, bool enable)
{
	int port;

	if (!dn)
		return -ENODEV;

	port = __cci_ace_get_port(dn, ACE_LITE_PORT);
	if (WARN_ONCE(port < 0, "node %s ACE lite port look-up failure\n",
				dn->full_name))
		return -ENODEV;
	cci_port_control(port, enable);
	return 0;
}
EXPORT_SYMBOL_GPL(__cci_control_port_by_device);

/**
 * __cci_control_port_by_index() - function to control a CCI port by port index
 *
 * @port: port index previously retrieved with cci_ace_get_port()
 * @enable: if true enables the port, if false disables it
 *
 * Return:
 *	0 on success
 *	-ENODEV on port index out of range
 *	-EPERM if operation carried out on an ACE PORT
 */
int notrace __cci_control_port_by_index(u32 port, bool enable)
{
	if (port >= nb_cci_ports || ports[port].type == ACE_INVALID_PORT)
		return -ENODEV;
	/*
	 * CCI control for ports connected to CPUS is extremely fragile
	 * and must be made to go through a specific and controlled
	 * interface (ie cci_disable_port_by_cpu(); control by general purpose
	 * indexing is therefore disabled for ACE ports.
	 */
	if (ports[port].type == ACE_PORT)
		return -EPERM;

	cci_port_control(port, enable);
	return 0;
}
EXPORT_SYMBOL_GPL(__cci_control_port_by_index);

static const struct of_device_id arm_cci_ctrl_if_matches[] = {
	{.compatible = "arm,cci-400-ctrl-if", },
	{},
};

static int cci_probe_ports(struct device_node *np)
{
	struct cci_nb_ports const *cci_config;
	int ret, i, nb_ace = 0, nb_ace_lite = 0;
	struct device_node *cp;
	struct resource res;
	const char *match_str;
	bool is_ace;


	cci_config = of_match_node(arm_cci_matches, np)->data;
	if (!cci_config)
		return -ENODEV;

	nb_cci_ports = cci_config->nb_ace + cci_config->nb_ace_lite;

	ports = kcalloc(nb_cci_ports, sizeof(*ports), GFP_KERNEL);
	if (!ports)
		return -ENOMEM;

	for_each_child_of_node(np, cp) {
		if (!of_match_node(arm_cci_ctrl_if_matches, cp))
			continue;

		i = nb_ace + nb_ace_lite;

		if (i >= nb_cci_ports)
			break;

		if (of_property_read_string(cp, "interface-type",
					&match_str)) {
			WARN(1, "node %s missing interface-type property\n",
				  cp->full_name);
			continue;
		}
		is_ace = strcmp(match_str, "ace") == 0;
		if (!is_ace && strcmp(match_str, "ace-lite")) {
			WARN(1, "node %s containing invalid interface-type property, skipping it\n",
					cp->full_name);
			continue;
		}

		ret = of_address_to_resource(cp, 0, &res);
		if (!ret) {
			ports[i].base = ioremap(res.start, resource_size(&res));
			ports[i].phys = res.start;
		}
		if (ret || !ports[i].base) {
			WARN(1, "unable to ioremap CCI port %d\n", i);
			continue;
		}

		if (is_ace) {
			if (WARN_ON(nb_ace >= cci_config->nb_ace))
				continue;
			ports[i].type = ACE_PORT;
			++nb_ace;
		} else {
			if (WARN_ON(nb_ace_lite >= cci_config->nb_ace_lite))
				continue;
			ports[i].type = ACE_LITE_PORT;
			++nb_ace_lite;
		}
		ports[i].dn = cp;
	}

	 /* initialize a stashed array of ACE ports to speed-up look-up */
	cci_ace_init_ports();

	/*
	 * Multi-cluster systems may need this data when non-coherent, during
	 * cluster power-up/power-down. Make sure it reaches main memory.
	 */
	sync_cache_w(&cci_ctrl_base);
	sync_cache_w(&cci_ctrl_phys);
	sync_cache_w(&ports);
	sync_cache_w(&cpu_port);
	__sync_cache_range_w(ports, sizeof(*ports) * nb_cci_ports);
	pr_info("ARM CCI driver probed\n");

	return 0;
}
#else /* !CONFIG_ARM_CCI400_PORT_CTRL */
static inline int cci_probe_ports(struct device_node *np)
{
	return 0;
}
#endif /* CONFIG_ARM_CCI400_PORT_CTRL */

static int cci_probe(void)
{
	int ret;
	struct device_node *np;
	struct resource res;

	np = of_find_matching_node(NULL, arm_cci_matches);
	if(!np || !of_device_is_available(np))
		return -ENODEV;

	ret = of_address_to_resource(np, 0, &res);
	if (!ret) {
		cci_ctrl_base = ioremap(res.start, resource_size(&res));
		cci_ctrl_phys =	res.start;
	}
	if (ret || !cci_ctrl_base) {
		WARN(1, "unable to ioremap CCI ctrl\n");
		return -ENXIO;
	}

	return cci_probe_ports(np);
}

static int cci_init_status = -EAGAIN;
static DEFINE_MUTEX(cci_probing);

static int cci_init(void)
{
	if (cci_init_status != -EAGAIN)
		return cci_init_status;

	mutex_lock(&cci_probing);
	if (cci_init_status == -EAGAIN)
		cci_init_status = cci_probe();
	mutex_unlock(&cci_probing);
	return cci_init_status;
}

/*
 * To sort out early init calls ordering a helper function is provided to
 * check if the CCI driver has beed initialized. Function check if the driver
 * has been initialized, if not it calls the init function that probes
 * the driver and updates the return value.
 */
bool cci_probed(void)
{
	return cci_init() == 0;
}
EXPORT_SYMBOL_GPL(cci_probed);

early_initcall(cci_init);
core_initcall(cci_platform_init);
MODULE_LICENSE("GPL");
MODULE_DESCRIPTION("ARM CCI support");