summaryrefslogtreecommitdiff
path: root/arch/powerpc/platforms/pseries/mobility.c
blob: e83e0891272d3866fe7d3276101bda070f9da258 (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
// SPDX-License-Identifier: GPL-2.0-only
/*
 * Support for Partition Mobility/Migration
 *
 * Copyright (C) 2010 Nathan Fontenot
 * Copyright (C) 2010 IBM Corporation
 */


#define pr_fmt(fmt) "mobility: " fmt

#include <linux/cpu.h>
#include <linux/kernel.h>
#include <linux/kobject.h>
#include <linux/nmi.h>
#include <linux/sched.h>
#include <linux/smp.h>
#include <linux/stat.h>
#include <linux/stop_machine.h>
#include <linux/completion.h>
#include <linux/device.h>
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/stringify.h>

#include <asm/machdep.h>
#include <asm/rtas.h>
#include "pseries.h"
#include "../../kernel/cacheinfo.h"

static struct kobject *mobility_kobj;

struct update_props_workarea {
	__be32 phandle;
	__be32 state;
	__be64 reserved;
	__be32 nprops;
} __packed;

#define NODE_ACTION_MASK	0xff000000
#define NODE_COUNT_MASK		0x00ffffff

#define DELETE_DT_NODE	0x01000000
#define UPDATE_DT_NODE	0x02000000
#define ADD_DT_NODE	0x03000000

#define MIGRATION_SCOPE	(1)
#define PRRN_SCOPE -2

static int mobility_rtas_call(int token, char *buf, s32 scope)
{
	int rc;

	spin_lock(&rtas_data_buf_lock);

	memcpy(rtas_data_buf, buf, RTAS_DATA_BUF_SIZE);
	rc = rtas_call(token, 2, 1, NULL, rtas_data_buf, scope);
	memcpy(buf, rtas_data_buf, RTAS_DATA_BUF_SIZE);

	spin_unlock(&rtas_data_buf_lock);
	return rc;
}

static int delete_dt_node(struct device_node *dn)
{
	pr_debug("removing node %pOFfp\n", dn);
	dlpar_detach_node(dn);
	return 0;
}

static int update_dt_property(struct device_node *dn, struct property **prop,
			      const char *name, u32 vd, char *value)
{
	struct property *new_prop = *prop;
	int more = 0;

	/* A negative 'vd' value indicates that only part of the new property
	 * value is contained in the buffer and we need to call
	 * ibm,update-properties again to get the rest of the value.
	 *
	 * A negative value is also the two's compliment of the actual value.
	 */
	if (vd & 0x80000000) {
		vd = ~vd + 1;
		more = 1;
	}

	if (new_prop) {
		/* partial property fixup */
		char *new_data = kzalloc(new_prop->length + vd, GFP_KERNEL);
		if (!new_data)
			return -ENOMEM;

		memcpy(new_data, new_prop->value, new_prop->length);
		memcpy(new_data + new_prop->length, value, vd);

		kfree(new_prop->value);
		new_prop->value = new_data;
		new_prop->length += vd;
	} else {
		new_prop = kzalloc(sizeof(*new_prop), GFP_KERNEL);
		if (!new_prop)
			return -ENOMEM;

		new_prop->name = kstrdup(name, GFP_KERNEL);
		if (!new_prop->name) {
			kfree(new_prop);
			return -ENOMEM;
		}

		new_prop->length = vd;
		new_prop->value = kzalloc(new_prop->length, GFP_KERNEL);
		if (!new_prop->value) {
			kfree(new_prop->name);
			kfree(new_prop);
			return -ENOMEM;
		}

		memcpy(new_prop->value, value, vd);
		*prop = new_prop;
	}

	if (!more) {
		pr_debug("updating node %pOF property %s\n", dn, name);
		of_update_property(dn, new_prop);
		*prop = NULL;
	}

	return 0;
}

static int update_dt_node(struct device_node *dn, s32 scope)
{
	struct update_props_workarea *upwa;
	struct property *prop = NULL;
	int i, rc, rtas_rc;
	char *prop_data;
	char *rtas_buf;
	int update_properties_token;
	u32 nprops;
	u32 vd;

	update_properties_token = rtas_token("ibm,update-properties");
	if (update_properties_token == RTAS_UNKNOWN_SERVICE)
		return -EINVAL;

	rtas_buf = kzalloc(RTAS_DATA_BUF_SIZE, GFP_KERNEL);
	if (!rtas_buf)
		return -ENOMEM;

	upwa = (struct update_props_workarea *)&rtas_buf[0];
	upwa->phandle = cpu_to_be32(dn->phandle);

	do {
		rtas_rc = mobility_rtas_call(update_properties_token, rtas_buf,
					scope);
		if (rtas_rc < 0)
			break;

		prop_data = rtas_buf + sizeof(*upwa);
		nprops = be32_to_cpu(upwa->nprops);

		/* On the first call to ibm,update-properties for a node the
		 * the first property value descriptor contains an empty
		 * property name, the property value length encoded as u32,
		 * and the property value is the node path being updated.
		 */
		if (*prop_data == 0) {
			prop_data++;
			vd = be32_to_cpu(*(__be32 *)prop_data);
			prop_data += vd + sizeof(vd);
			nprops--;
		}

		for (i = 0; i < nprops; i++) {
			char *prop_name;

			prop_name = prop_data;
			prop_data += strlen(prop_name) + 1;
			vd = be32_to_cpu(*(__be32 *)prop_data);
			prop_data += sizeof(vd);

			switch (vd) {
			case 0x00000000:
				/* name only property, nothing to do */
				break;

			case 0x80000000:
				of_remove_property(dn, of_find_property(dn,
							prop_name, NULL));
				prop = NULL;
				break;

			default:
				rc = update_dt_property(dn, &prop, prop_name,
							vd, prop_data);
				if (rc) {
					pr_err("updating %s property failed: %d\n",
					       prop_name, rc);
				}

				prop_data += vd;
				break;
			}

			cond_resched();
		}

		cond_resched();
	} while (rtas_rc == 1);

	kfree(rtas_buf);
	return 0;
}

static int add_dt_node(struct device_node *parent_dn, __be32 drc_index)
{
	struct device_node *dn;
	int rc;

	dn = dlpar_configure_connector(drc_index, parent_dn);
	if (!dn)
		return -ENOENT;

	rc = dlpar_attach_node(dn, parent_dn);
	if (rc)
		dlpar_free_cc_nodes(dn);

	pr_debug("added node %pOFfp\n", dn);

	return rc;
}

int pseries_devicetree_update(s32 scope)
{
	char *rtas_buf;
	__be32 *data;
	int update_nodes_token;
	int rc;

	update_nodes_token = rtas_token("ibm,update-nodes");
	if (update_nodes_token == RTAS_UNKNOWN_SERVICE)
		return 0;

	rtas_buf = kzalloc(RTAS_DATA_BUF_SIZE, GFP_KERNEL);
	if (!rtas_buf)
		return -ENOMEM;

	do {
		rc = mobility_rtas_call(update_nodes_token, rtas_buf, scope);
		if (rc && rc != 1)
			break;

		data = (__be32 *)rtas_buf + 4;
		while (be32_to_cpu(*data) & NODE_ACTION_MASK) {
			int i;
			u32 action = be32_to_cpu(*data) & NODE_ACTION_MASK;
			u32 node_count = be32_to_cpu(*data) & NODE_COUNT_MASK;

			data++;

			for (i = 0; i < node_count; i++) {
				struct device_node *np;
				__be32 phandle = *data++;
				__be32 drc_index;

				np = of_find_node_by_phandle(be32_to_cpu(phandle));
				if (!np) {
					pr_warn("Failed lookup: phandle 0x%x for action 0x%x\n",
						be32_to_cpu(phandle), action);
					continue;
				}

				switch (action) {
				case DELETE_DT_NODE:
					delete_dt_node(np);
					break;
				case UPDATE_DT_NODE:
					update_dt_node(np, scope);
					break;
				case ADD_DT_NODE:
					drc_index = *data++;
					add_dt_node(np, drc_index);
					break;
				}

				of_node_put(np);
				cond_resched();
			}
		}

		cond_resched();
	} while (rc == 1);

	kfree(rtas_buf);
	return rc;
}

void post_mobility_fixup(void)
{
	int rc;

	rtas_activate_firmware();

	/*
	 * We don't want CPUs to go online/offline while the device
	 * tree is being updated.
	 */
	cpus_read_lock();

	/*
	 * It's common for the destination firmware to replace cache
	 * nodes.  Release all of the cacheinfo hierarchy's references
	 * before updating the device tree.
	 */
	cacheinfo_teardown();

	rc = pseries_devicetree_update(MIGRATION_SCOPE);
	if (rc)
		pr_err("device tree update failed: %d\n", rc);

	cacheinfo_rebuild();

	cpus_read_unlock();

	/* Possibly switch to a new L1 flush type */
	pseries_setup_security_mitigations();

	/* Reinitialise system information for hv-24x7 */
	read_24x7_sys_info();

	return;
}

static int poll_vasi_state(u64 handle, unsigned long *res)
{
	unsigned long retbuf[PLPAR_HCALL_BUFSIZE];
	long hvrc;
	int ret;

	hvrc = plpar_hcall(H_VASI_STATE, retbuf, handle);
	switch (hvrc) {
	case H_SUCCESS:
		ret = 0;
		*res = retbuf[0];
		break;
	case H_PARAMETER:
		ret = -EINVAL;
		break;
	case H_FUNCTION:
		ret = -EOPNOTSUPP;
		break;
	case H_HARDWARE:
	default:
		pr_err("unexpected H_VASI_STATE result %ld\n", hvrc);
		ret = -EIO;
		break;
	}
	return ret;
}

static int wait_for_vasi_session_suspending(u64 handle)
{
	unsigned long state;
	int ret;

	/*
	 * Wait for transition from H_VASI_ENABLED to
	 * H_VASI_SUSPENDING. Treat anything else as an error.
	 */
	while (true) {
		ret = poll_vasi_state(handle, &state);

		if (ret != 0 || state == H_VASI_SUSPENDING) {
			break;
		} else if (state == H_VASI_ENABLED) {
			ssleep(1);
		} else {
			pr_err("unexpected H_VASI_STATE result %lu\n", state);
			ret = -EIO;
			break;
		}
	}

	/*
	 * Proceed even if H_VASI_STATE is unavailable. If H_JOIN or
	 * ibm,suspend-me are also unimplemented, we'll recover then.
	 */
	if (ret == -EOPNOTSUPP)
		ret = 0;

	return ret;
}

static void prod_single(unsigned int target_cpu)
{
	long hvrc;
	int hwid;

	hwid = get_hard_smp_processor_id(target_cpu);
	hvrc = plpar_hcall_norets(H_PROD, hwid);
	if (hvrc == H_SUCCESS)
		return;
	pr_err_ratelimited("H_PROD of CPU %u (hwid %d) error: %ld\n",
			   target_cpu, hwid, hvrc);
}

static void prod_others(void)
{
	unsigned int cpu;

	for_each_online_cpu(cpu) {
		if (cpu != smp_processor_id())
			prod_single(cpu);
	}
}

static u16 clamp_slb_size(void)
{
	u16 prev = mmu_slb_size;

	slb_set_size(SLB_MIN_SIZE);

	return prev;
}

static int do_suspend(void)
{
	u16 saved_slb_size;
	int status;
	int ret;

	pr_info("calling ibm,suspend-me on CPU %i\n", smp_processor_id());

	/*
	 * The destination processor model may have fewer SLB entries
	 * than the source. We reduce mmu_slb_size to a safe minimum
	 * before suspending in order to minimize the possibility of
	 * programming non-existent entries on the destination. If
	 * suspend fails, we restore it before returning. On success
	 * the OF reconfig path will update it from the new device
	 * tree after resuming on the destination.
	 */
	saved_slb_size = clamp_slb_size();

	ret = rtas_ibm_suspend_me(&status);
	if (ret != 0) {
		pr_err("ibm,suspend-me error: %d\n", status);
		slb_set_size(saved_slb_size);
	}

	return ret;
}

/**
 * struct pseries_suspend_info - State shared between CPUs for join/suspend.
 * @counter: Threads are to increment this upon resuming from suspend
 *           or if an error is received from H_JOIN. The thread which performs
 *           the first increment (i.e. sets it to 1) is responsible for
 *           waking the other threads.
 * @done: False if join/suspend is in progress. True if the operation is
 *        complete (successful or not).
 */
struct pseries_suspend_info {
	atomic_t counter;
	bool done;
};

static int do_join(void *arg)
{
	struct pseries_suspend_info *info = arg;
	atomic_t *counter = &info->counter;
	long hvrc;
	int ret;

retry:
	/* Must ensure MSR.EE off for H_JOIN. */
	hard_irq_disable();
	hvrc = plpar_hcall_norets(H_JOIN);

	switch (hvrc) {
	case H_CONTINUE:
		/*
		 * All other CPUs are offline or in H_JOIN. This CPU
		 * attempts the suspend.
		 */
		ret = do_suspend();
		break;
	case H_SUCCESS:
		/*
		 * The suspend is complete and this cpu has received a
		 * prod, or we've received a stray prod from unrelated
		 * code (e.g. paravirt spinlocks) and we need to join
		 * again.
		 *
		 * This barrier orders the return from H_JOIN above vs
		 * the load of info->done. It pairs with the barrier
		 * in the wakeup/prod path below.
		 */
		smp_mb();
		if (READ_ONCE(info->done) == false) {
			pr_info_ratelimited("premature return from H_JOIN on CPU %i, retrying",
					    smp_processor_id());
			goto retry;
		}
		ret = 0;
		break;
	case H_BAD_MODE:
	case H_HARDWARE:
	default:
		ret = -EIO;
		pr_err_ratelimited("H_JOIN error %ld on CPU %i\n",
				   hvrc, smp_processor_id());
		break;
	}

	if (atomic_inc_return(counter) == 1) {
		pr_info("CPU %u waking all threads\n", smp_processor_id());
		WRITE_ONCE(info->done, true);
		/*
		 * This barrier orders the store to info->done vs subsequent
		 * H_PRODs to wake the other CPUs. It pairs with the barrier
		 * in the H_SUCCESS case above.
		 */
		smp_mb();
		prod_others();
	}
	/*
	 * Execution may have been suspended for several seconds, so
	 * reset the watchdog.
	 */
	touch_nmi_watchdog();
	return ret;
}

/*
 * Abort reason code byte 0. We use only the 'Migrating partition' value.
 */
enum vasi_aborting_entity {
	ORCHESTRATOR        = 1,
	VSP_SOURCE          = 2,
	PARTITION_FIRMWARE  = 3,
	PLATFORM_FIRMWARE   = 4,
	VSP_TARGET          = 5,
	MIGRATING_PARTITION = 6,
};

static void pseries_cancel_migration(u64 handle, int err)
{
	u32 reason_code;
	u32 detail;
	u8 entity;
	long hvrc;

	entity = MIGRATING_PARTITION;
	detail = abs(err) & 0xffffff;
	reason_code = (entity << 24) | detail;

	hvrc = plpar_hcall_norets(H_VASI_SIGNAL, handle,
				  H_VASI_SIGNAL_CANCEL, reason_code);
	if (hvrc)
		pr_err("H_VASI_SIGNAL error: %ld\n", hvrc);
}

static int pseries_suspend(u64 handle)
{
	const unsigned int max_attempts = 5;
	unsigned int retry_interval_ms = 1;
	unsigned int attempt = 1;
	int ret;

	while (true) {
		struct pseries_suspend_info info;
		unsigned long vasi_state;
		int vasi_err;

		info = (struct pseries_suspend_info) {
			.counter = ATOMIC_INIT(0),
			.done = false,
		};

		ret = stop_machine(do_join, &info, cpu_online_mask);
		if (ret == 0)
			break;
		/*
		 * Encountered an error. If the VASI stream is still
		 * in Suspending state, it's likely a transient
		 * condition related to some device in the partition
		 * and we can retry in the hope that the cause has
		 * cleared after some delay.
		 *
		 * A better design would allow drivers etc to prepare
		 * for the suspend and avoid conditions which prevent
		 * the suspend from succeeding. For now, we have this
		 * mitigation.
		 */
		pr_notice("Partition suspend attempt %u of %u error: %d\n",
			  attempt, max_attempts, ret);

		if (attempt == max_attempts)
			break;

		vasi_err = poll_vasi_state(handle, &vasi_state);
		if (vasi_err == 0) {
			if (vasi_state != H_VASI_SUSPENDING) {
				pr_notice("VASI state %lu after failed suspend\n",
					  vasi_state);
				break;
			}
		} else if (vasi_err != -EOPNOTSUPP) {
			pr_err("VASI state poll error: %d", vasi_err);
			break;
		}

		pr_notice("Will retry partition suspend after %u ms\n",
			  retry_interval_ms);

		msleep(retry_interval_ms);
		retry_interval_ms *= 10;
		attempt++;
	}

	return ret;
}

static int pseries_migrate_partition(u64 handle)
{
	int ret;

	ret = wait_for_vasi_session_suspending(handle);
	if (ret)
		return ret;

	ret = pseries_suspend(handle);
	if (ret == 0)
		post_mobility_fixup();
	else
		pseries_cancel_migration(handle, ret);

	return ret;
}

int rtas_syscall_dispatch_ibm_suspend_me(u64 handle)
{
	return pseries_migrate_partition(handle);
}

static ssize_t migration_store(struct class *class,
			       struct class_attribute *attr, const char *buf,
			       size_t count)
{
	u64 streamid;
	int rc;

	rc = kstrtou64(buf, 0, &streamid);
	if (rc)
		return rc;

	rc = pseries_migrate_partition(streamid);
	if (rc)
		return rc;

	return count;
}

/*
 * Used by drmgr to determine the kernel behavior of the migration interface.
 *
 * Version 1: Performs all PAPR requirements for migration including
 *	firmware activation and device tree update.
 */
#define MIGRATION_API_VERSION	1

static CLASS_ATTR_WO(migration);
static CLASS_ATTR_STRING(api_version, 0444, __stringify(MIGRATION_API_VERSION));

static int __init mobility_sysfs_init(void)
{
	int rc;

	mobility_kobj = kobject_create_and_add("mobility", kernel_kobj);
	if (!mobility_kobj)
		return -ENOMEM;

	rc = sysfs_create_file(mobility_kobj, &class_attr_migration.attr);
	if (rc)
		pr_err("unable to create migration sysfs file (%d)\n", rc);

	rc = sysfs_create_file(mobility_kobj, &class_attr_api_version.attr.attr);
	if (rc)
		pr_err("unable to create api_version sysfs file (%d)\n", rc);

	return 0;
}
machine_device_initcall(pseries, mobility_sysfs_init);