powerpc.c 26.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
/*
 * 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 in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.
 *
 * Copyright IBM Corp. 2007
 *
 * Authors: Hollis Blanchard <hollisb@us.ibm.com>
 *          Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
 */

#include <linux/errno.h>
#include <linux/err.h>
#include <linux/kvm_host.h>
#include <linux/vmalloc.h>
A
Alexander Graf 已提交
25
#include <linux/hrtimer.h>
26
#include <linux/fs.h>
27
#include <linux/slab.h>
S
Scott Wood 已提交
28
#include <linux/file.h>
29
#include <linux/module.h>
30 31 32
#include <asm/cputable.h>
#include <asm/uaccess.h>
#include <asm/kvm_ppc.h>
33
#include <asm/tlbflush.h>
34
#include <asm/cputhreads.h>
35
#include <asm/irqflags.h>
36
#include "timing.h"
37
#include "irq.h"
P
Paul Mackerras 已提交
38
#include "../mm/mmu_decl.h"
39

40 41 42
#define CREATE_TRACE_POINTS
#include "trace.h"

43 44 45 46 47
struct kvmppc_ops *kvmppc_hv_ops;
EXPORT_SYMBOL_GPL(kvmppc_hv_ops);
struct kvmppc_ops *kvmppc_pr_ops;
EXPORT_SYMBOL_GPL(kvmppc_pr_ops);

48

49 50
int kvm_arch_vcpu_runnable(struct kvm_vcpu *v)
{
51
	return !!(v->arch.pending_exceptions) ||
52
	       v->requests;
53 54
}

55 56 57 58 59
int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
{
	return 1;
}

60 61 62 63
/*
 * Common checks before entering the guest world.  Call with interrupts
 * disabled.
 *
64 65 66 67
 * returns:
 *
 * == 1 if we're ready to go into guest state
 * <= 0 if we need to go back to the host with return value
68 69 70
 */
int kvmppc_prepare_to_enter(struct kvm_vcpu *vcpu)
{
S
Scott Wood 已提交
71 72 73 74
	int r;

	WARN_ON(irqs_disabled());
	hard_irq_disable();
75 76 77 78 79

	while (true) {
		if (need_resched()) {
			local_irq_enable();
			cond_resched();
S
Scott Wood 已提交
80
			hard_irq_disable();
81 82 83 84
			continue;
		}

		if (signal_pending(current)) {
85 86 87
			kvmppc_account_exit(vcpu, SIGNAL_EXITS);
			vcpu->run->exit_reason = KVM_EXIT_INTR;
			r = -EINTR;
88 89 90
			break;
		}

91 92 93 94 95 96 97 98
		vcpu->mode = IN_GUEST_MODE;

		/*
		 * Reading vcpu->requests must happen after setting vcpu->mode,
		 * so we don't miss a request because the requester sees
		 * OUTSIDE_GUEST_MODE and assumes we'll be checking requests
		 * before next entering the guest (and thus doesn't IPI).
		 */
99
		smp_mb();
100

101 102 103 104
		if (vcpu->requests) {
			/* Make sure we process requests preemptable */
			local_irq_enable();
			trace_kvm_check_requests(vcpu);
105
			r = kvmppc_core_check_requests(vcpu);
S
Scott Wood 已提交
106
			hard_irq_disable();
107 108 109
			if (r > 0)
				continue;
			break;
110 111 112 113 114 115 116 117
		}

		if (kvmppc_core_prepare_to_enter(vcpu)) {
			/* interrupts got enabled in between, so we
			   are back at square 1 */
			continue;
		}

118
		kvm_guest_enter();
S
Scott Wood 已提交
119
		return 1;
120 121
	}

S
Scott Wood 已提交
122 123
	/* return to host */
	local_irq_enable();
124 125
	return r;
}
126
EXPORT_SYMBOL_GPL(kvmppc_prepare_to_enter);
127

128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148
#if defined(CONFIG_PPC_BOOK3S_64) && defined(CONFIG_KVM_BOOK3S_PR_POSSIBLE)
static void kvmppc_swab_shared(struct kvm_vcpu *vcpu)
{
	struct kvm_vcpu_arch_shared *shared = vcpu->arch.shared;
	int i;

	shared->sprg0 = swab64(shared->sprg0);
	shared->sprg1 = swab64(shared->sprg1);
	shared->sprg2 = swab64(shared->sprg2);
	shared->sprg3 = swab64(shared->sprg3);
	shared->srr0 = swab64(shared->srr0);
	shared->srr1 = swab64(shared->srr1);
	shared->dar = swab64(shared->dar);
	shared->msr = swab64(shared->msr);
	shared->dsisr = swab32(shared->dsisr);
	shared->int_pending = swab32(shared->int_pending);
	for (i = 0; i < ARRAY_SIZE(shared->sr); i++)
		shared->sr[i] = swab32(shared->sr[i]);
}
#endif

149 150 151 152 153 154 155 156 157 158
int kvmppc_kvm_pv(struct kvm_vcpu *vcpu)
{
	int nr = kvmppc_get_gpr(vcpu, 11);
	int r;
	unsigned long __maybe_unused param1 = kvmppc_get_gpr(vcpu, 3);
	unsigned long __maybe_unused param2 = kvmppc_get_gpr(vcpu, 4);
	unsigned long __maybe_unused param3 = kvmppc_get_gpr(vcpu, 5);
	unsigned long __maybe_unused param4 = kvmppc_get_gpr(vcpu, 6);
	unsigned long r2 = 0;

159
	if (!(kvmppc_get_msr(vcpu) & MSR_SF)) {
160 161 162 163 164 165 166 167
		/* 32 bit mode */
		param1 &= 0xffffffff;
		param2 &= 0xffffffff;
		param3 &= 0xffffffff;
		param4 &= 0xffffffff;
	}

	switch (nr) {
168
	case KVM_HCALL_TOKEN(KVM_HC_PPC_MAP_MAGIC_PAGE):
169
	{
170 171 172 173 174 175 176 177 178 179
#if defined(CONFIG_PPC_BOOK3S_64) && defined(CONFIG_KVM_BOOK3S_PR_POSSIBLE)
		/* Book3S can be little endian, find it out here */
		int shared_big_endian = true;
		if (vcpu->arch.intr_msr & MSR_LE)
			shared_big_endian = false;
		if (shared_big_endian != vcpu->arch.shared_big_endian)
			kvmppc_swab_shared(vcpu);
		vcpu->arch.shared_big_endian = shared_big_endian;
#endif

180 181 182 183 184 185 186 187 188 189 190 191
		if (!(param2 & MAGIC_PAGE_FLAG_NOT_MAPPED_NX)) {
			/*
			 * Older versions of the Linux magic page code had
			 * a bug where they would map their trampoline code
			 * NX. If that's the case, remove !PR NX capability.
			 */
			vcpu->arch.disable_kernel_nx = true;
			kvm_make_request(KVM_REQ_TLB_FLUSH, vcpu);
		}

		vcpu->arch.magic_page_pa = param1 & ~0xfffULL;
		vcpu->arch.magic_page_ea = param2 & ~0xfffULL;
192

193
		r2 = KVM_MAGIC_FEAT_SR | KVM_MAGIC_FEAT_MAS0_TO_SPRG7;
194

195
		r = EV_SUCCESS;
196 197
		break;
	}
198 199
	case KVM_HCALL_TOKEN(KVM_HC_FEATURES):
		r = EV_SUCCESS;
200
#if defined(CONFIG_PPC_BOOK3S) || defined(CONFIG_KVM_E500V2)
S
Scott Wood 已提交
201
		/* XXX Missing magic page on 44x */
202 203
		r2 |= (1 << KVM_FEATURE_MAGIC_PAGE);
#endif
204 205 206

		/* Second return value is in r4 */
		break;
207 208 209 210 211
	case EV_HCALL_TOKEN(EV_IDLE):
		r = EV_SUCCESS;
		kvm_vcpu_block(vcpu);
		clear_bit(KVM_REQ_UNHALT, &vcpu->requests);
		break;
212
	default:
213
		r = EV_UNIMPLEMENTED;
214 215 216
		break;
	}

217 218
	kvmppc_set_gpr(vcpu, 4, r2);

219 220
	return r;
}
221
EXPORT_SYMBOL_GPL(kvmppc_kvm_pv);
222

223 224 225 226 227 228 229 230 231 232 233 234 235
int kvmppc_sanity_check(struct kvm_vcpu *vcpu)
{
	int r = false;

	/* We have to know what CPU to virtualize */
	if (!vcpu->arch.pvr)
		goto out;

	/* PAPR only works with book3s_64 */
	if ((vcpu->arch.cpu_type != KVM_CPU_3S_64) && vcpu->arch.papr_enabled)
		goto out;

	/* HV KVM can only do PAPR mode for now */
236
	if (!vcpu->arch.papr_enabled && is_kvmppc_hv_enabled(vcpu->kvm))
237 238
		goto out;

239 240 241 242 243
#ifdef CONFIG_KVM_BOOKE_HV
	if (!cpu_has_feature(CPU_FTR_EMB_HV))
		goto out;
#endif

244 245 246 247 248 249
	r = true;

out:
	vcpu->arch.sane = r;
	return r ? 0 : -EINVAL;
}
250
EXPORT_SYMBOL_GPL(kvmppc_sanity_check);
251

252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274
int kvmppc_emulate_mmio(struct kvm_run *run, struct kvm_vcpu *vcpu)
{
	enum emulation_result er;
	int r;

	er = kvmppc_emulate_instruction(run, vcpu);
	switch (er) {
	case EMULATE_DONE:
		/* Future optimization: only reload non-volatiles if they were
		 * actually modified. */
		r = RESUME_GUEST_NV;
		break;
	case EMULATE_DO_MMIO:
		run->exit_reason = KVM_EXIT_MMIO;
		/* We must reload nonvolatiles because "update" load/store
		 * instructions modify register state. */
		/* Future optimization: only reload non-volatiles if they were
		 * actually modified. */
		r = RESUME_HOST_NV;
		break;
	case EMULATE_FAIL:
		/* XXX Deliver Program interrupt to guest. */
		printk(KERN_EMERG "%s: emulation failed (%08x)\n", __func__,
275
		       kvmppc_get_last_inst(vcpu));
276 277 278
		r = RESUME_HOST;
		break;
	default:
279 280
		WARN_ON(1);
		r = RESUME_GUEST;
281 282 283 284
	}

	return r;
}
285
EXPORT_SYMBOL_GPL(kvmppc_emulate_mmio);
286

287
int kvm_arch_hardware_enable(void *garbage)
288
{
289
	return 0;
290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
}

void kvm_arch_hardware_disable(void *garbage)
{
}

int kvm_arch_hardware_setup(void)
{
	return 0;
}

void kvm_arch_hardware_unsetup(void)
{
}

void kvm_arch_check_processor_compat(void *rtn)
{
307
	*(int *)rtn = kvmppc_core_check_processor_compat();
308 309
}

310
int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
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
	struct kvmppc_ops *kvm_ops = NULL;
	/*
	 * if we have both HV and PR enabled, default is HV
	 */
	if (type == 0) {
		if (kvmppc_hv_ops)
			kvm_ops = kvmppc_hv_ops;
		else
			kvm_ops = kvmppc_pr_ops;
		if (!kvm_ops)
			goto err_out;
	} else	if (type == KVM_VM_PPC_HV) {
		if (!kvmppc_hv_ops)
			goto err_out;
		kvm_ops = kvmppc_hv_ops;
	} else if (type == KVM_VM_PPC_PR) {
		if (!kvmppc_pr_ops)
			goto err_out;
		kvm_ops = kvmppc_pr_ops;
	} else
		goto err_out;

	if (kvm_ops->owner && !try_module_get(kvm_ops->owner))
		return -ENOENT;

	kvm->arch.kvm_ops = kvm_ops;
338
	return kvmppc_core_init_vm(kvm);
339 340
err_out:
	return -EINVAL;
341 342
}

343
void kvm_arch_destroy_vm(struct kvm *kvm)
344 345
{
	unsigned int i;
346
	struct kvm_vcpu *vcpu;
347

348 349 350 351 352 353 354 355
	kvm_for_each_vcpu(i, vcpu, kvm)
		kvm_arch_vcpu_free(vcpu);

	mutex_lock(&kvm->lock);
	for (i = 0; i < atomic_read(&kvm->online_vcpus); i++)
		kvm->vcpus[i] = NULL;

	atomic_set(&kvm->online_vcpus, 0);
356 357 358

	kvmppc_core_destroy_vm(kvm);

359
	mutex_unlock(&kvm->lock);
360 361 362

	/* drop the module reference */
	module_put(kvm->arch.kvm_ops->owner);
363 364
}

365 366 367 368
void kvm_arch_sync_events(struct kvm *kvm)
{
}

369 370 371
int kvm_dev_ioctl_check_extension(long ext)
{
	int r;
372 373 374 375
	/* FIXME!!
	 * Should some of this be vm ioctl ? is it possible now ?
	 */
	int hv_enabled = kvmppc_hv_ops ? 1 : 0;
376 377

	switch (ext) {
S
Scott Wood 已提交
378 379
#ifdef CONFIG_BOOKE
	case KVM_CAP_PPC_BOOKE_SREGS:
380
	case KVM_CAP_PPC_BOOKE_WATCHDOG:
381
	case KVM_CAP_PPC_EPR:
S
Scott Wood 已提交
382
#else
383
	case KVM_CAP_PPC_SEGSTATE:
384
	case KVM_CAP_PPC_HIOR:
385
	case KVM_CAP_PPC_PAPR:
S
Scott Wood 已提交
386
#endif
387
	case KVM_CAP_PPC_UNSET_IRQ:
388
	case KVM_CAP_PPC_IRQ_LEVEL:
389
	case KVM_CAP_ENABLE_CAP:
390
	case KVM_CAP_ONE_REG:
A
Alexander Graf 已提交
391
	case KVM_CAP_IOEVENTFD:
392
	case KVM_CAP_DEVICE_CTRL:
393 394 395
		r = 1;
		break;
	case KVM_CAP_PPC_PAIRED_SINGLES:
396
	case KVM_CAP_PPC_OSI:
397
	case KVM_CAP_PPC_GET_PVINFO:
398
#if defined(CONFIG_KVM_E500V2) || defined(CONFIG_KVM_E500MC)
S
Scott Wood 已提交
399
	case KVM_CAP_SW_TLB:
S
Scott Wood 已提交
400
#endif
401
		/* We support this only for PR */
402
		r = !hv_enabled;
403
		break;
404
#ifdef CONFIG_KVM_MMIO
405 406 407
	case KVM_CAP_COALESCED_MMIO:
		r = KVM_COALESCED_MMIO_PAGE_OFFSET;
		break;
408
#endif
409 410 411 412 413 414
#ifdef CONFIG_KVM_MPIC
	case KVM_CAP_IRQ_MPIC:
		r = 1;
		break;
#endif

415
#ifdef CONFIG_PPC_BOOK3S_64
416
	case KVM_CAP_SPAPR_TCE:
417
	case KVM_CAP_PPC_ALLOC_HTAB:
418
	case KVM_CAP_PPC_RTAS:
419 420 421
#ifdef CONFIG_KVM_XICS
	case KVM_CAP_IRQ_XICS:
#endif
422 423
		r = 1;
		break;
424
#endif /* CONFIG_PPC_BOOK3S_64 */
425
#ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
426
	case KVM_CAP_PPC_SMT:
427
		if (hv_enabled)
428 429 430
			r = threads_per_core;
		else
			r = 0;
431
		break;
432
	case KVM_CAP_PPC_RMA:
433
		r = hv_enabled;
434
		/* PPC970 requires an RMA */
435
		if (r && cpu_has_feature(CPU_FTR_ARCH_201))
436
			r = 2;
437
		break;
438
#endif
439
	case KVM_CAP_SYNC_MMU:
440
#ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
441
		if (hv_enabled)
442 443 444
			r = cpu_has_feature(CPU_FTR_ARCH_206) ? 1 : 0;
		else
			r = 0;
445 446 447 448
#elif defined(KVM_ARCH_WANT_MMU_NOTIFIER)
		r = 1;
#else
		r = 0;
449
#endif
450 451
		break;
#ifdef CONFIG_KVM_BOOK3S_HV_POSSIBLE
452
	case KVM_CAP_PPC_HTAB_FD:
453
		r = hv_enabled;
454
		break;
455
#endif
456 457 458 459 460 461 462
	case KVM_CAP_NR_VCPUS:
		/*
		 * Recommending a number of CPUs is somewhat arbitrary; we
		 * return the number of present CPUs for -HV (since a host
		 * will have secondary threads "offline"), and for other KVM
		 * implementations just count online CPUs.
		 */
463
		if (hv_enabled)
464 465 466
			r = num_present_cpus();
		else
			r = num_online_cpus();
467 468 469 470
		break;
	case KVM_CAP_MAX_VCPUS:
		r = KVM_MAX_VCPUS;
		break;
471 472 473 474 475
#ifdef CONFIG_PPC_BOOK3S_64
	case KVM_CAP_PPC_GET_SMMU_INFO:
		r = 1;
		break;
#endif
476 477 478 479 480 481 482 483 484 485 486 487 488 489
	default:
		r = 0;
		break;
	}
	return r;

}

long kvm_arch_dev_ioctl(struct file *filp,
                        unsigned int ioctl, unsigned long arg)
{
	return -EINVAL;
}

490
void kvm_arch_free_memslot(struct kvm *kvm, struct kvm_memory_slot *free,
491 492
			   struct kvm_memory_slot *dont)
{
493
	kvmppc_core_free_memslot(kvm, free, dont);
494 495
}

496 497
int kvm_arch_create_memslot(struct kvm *kvm, struct kvm_memory_slot *slot,
			    unsigned long npages)
498
{
499
	return kvmppc_core_create_memslot(kvm, slot, npages);
500 501
}

502 503 504 505
void kvm_arch_memslots_updated(struct kvm *kvm)
{
}

506
int kvm_arch_prepare_memory_region(struct kvm *kvm,
507
				   struct kvm_memory_slot *memslot,
508 509
				   struct kvm_userspace_memory_region *mem,
				   enum kvm_mr_change change)
510
{
511
	return kvmppc_core_prepare_memory_region(kvm, memslot, mem);
512 513
}

514
void kvm_arch_commit_memory_region(struct kvm *kvm,
515
				   struct kvm_userspace_memory_region *mem,
516 517
				   const struct kvm_memory_slot *old,
				   enum kvm_mr_change change)
518
{
519
	kvmppc_core_commit_memory_region(kvm, mem, old);
520 521
}

522 523 524
void kvm_arch_flush_shadow_all(struct kvm *kvm)
{
}
525

526 527
void kvm_arch_flush_shadow_memslot(struct kvm *kvm,
				   struct kvm_memory_slot *slot)
528
{
529
	kvmppc_core_flush_memslot(kvm, slot);
530 531
}

532 533
struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm, unsigned int id)
{
534 535
	struct kvm_vcpu *vcpu;
	vcpu = kvmppc_core_vcpu_create(kvm, id);
536 537
	if (!IS_ERR(vcpu)) {
		vcpu->arch.wqp = &vcpu->wq;
538
		kvmppc_create_vcpu_debugfs(vcpu, id);
539
	}
540
	return vcpu;
541 542
}

543 544 545 546 547
int kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
{
	return 0;
}

548 549
void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
{
550 551 552 553
	/* Make sure we're not using the vcpu anymore */
	hrtimer_cancel(&vcpu->arch.dec_timer);
	tasklet_kill(&vcpu->arch.tasklet);

554
	kvmppc_remove_vcpu_debugfs(vcpu);
S
Scott Wood 已提交
555 556 557 558 559

	switch (vcpu->arch.irq_type) {
	case KVMPPC_IRQ_MPIC:
		kvmppc_mpic_disconnect_vcpu(vcpu->arch.mpic, vcpu);
		break;
560 561 562
	case KVMPPC_IRQ_XICS:
		kvmppc_xics_free_icp(vcpu);
		break;
S
Scott Wood 已提交
563 564
	}

565
	kvmppc_core_vcpu_free(vcpu);
566 567 568 569 570 571 572 573 574
}

void kvm_arch_vcpu_destroy(struct kvm_vcpu *vcpu)
{
	kvm_arch_vcpu_free(vcpu);
}

int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu)
{
575
	return kvmppc_core_pending_dec(vcpu);
576 577
}

A
Alexander Graf 已提交
578 579 580 581 582 583 584 585 586 587 588 589 590 591
/*
 * low level hrtimer wake routine. Because this runs in hardirq context
 * we schedule a tasklet to do the real work.
 */
enum hrtimer_restart kvmppc_decrementer_wakeup(struct hrtimer *timer)
{
	struct kvm_vcpu *vcpu;

	vcpu = container_of(timer, struct kvm_vcpu, arch.dec_timer);
	tasklet_schedule(&vcpu->arch.tasklet);

	return HRTIMER_NORESTART;
}

592 593
int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
{
594 595
	int ret;

A
Alexander Graf 已提交
596 597 598
	hrtimer_init(&vcpu->arch.dec_timer, CLOCK_REALTIME, HRTIMER_MODE_ABS);
	tasklet_init(&vcpu->arch.tasklet, kvmppc_decrementer_func, (ulong)vcpu);
	vcpu->arch.dec_timer.function = kvmppc_decrementer_wakeup;
599
	vcpu->arch.dec_expires = ~(u64)0;
600

601 602 603
#ifdef CONFIG_KVM_EXIT_TIMING
	mutex_init(&vcpu->arch.exit_timing_lock);
#endif
604 605
	ret = kvmppc_subarch_vcpu_init(vcpu);
	return ret;
606 607 608 609
}

void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu)
{
610
	kvmppc_mmu_destroy(vcpu);
611
	kvmppc_subarch_vcpu_uninit(vcpu);
612 613 614 615
}

void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
{
616 617 618 619 620 621 622 623 624 625
#ifdef CONFIG_BOOKE
	/*
	 * vrsave (formerly usprg0) isn't used by Linux, but may
	 * be used by the guest.
	 *
	 * On non-booke this is associated with Altivec and
	 * is handled by code in book3s.c.
	 */
	mtspr(SPRN_VRSAVE, vcpu->arch.vrsave);
#endif
626
	kvmppc_core_vcpu_load(vcpu, cpu);
627 628 629 630
}

void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
631
	kvmppc_core_vcpu_put(vcpu);
632 633 634
#ifdef CONFIG_BOOKE
	vcpu->arch.vrsave = mfspr(SPRN_VRSAVE);
#endif
635 636 637 638 639
}

static void kvmppc_complete_dcr_load(struct kvm_vcpu *vcpu,
                                     struct kvm_run *run)
{
640
	kvmppc_set_gpr(vcpu, vcpu->arch.io_gpr, run->dcr.data);
641 642 643 644 645
}

static void kvmppc_complete_mmio_load(struct kvm_vcpu *vcpu,
                                      struct kvm_run *run)
{
646
	u64 uninitialized_var(gpr);
647

648
	if (run->mmio.len > sizeof(gpr)) {
649 650 651 652 653 654
		printk(KERN_ERR "bad MMIO length: %d\n", run->mmio.len);
		return;
	}

	if (vcpu->arch.mmio_is_bigendian) {
		switch (run->mmio.len) {
655
		case 8: gpr = *(u64 *)run->mmio.data; break;
656 657 658
		case 4: gpr = *(u32 *)run->mmio.data; break;
		case 2: gpr = *(u16 *)run->mmio.data; break;
		case 1: gpr = *(u8 *)run->mmio.data; break;
659 660 661 662
		}
	} else {
		/* Convert BE data from userland back to LE. */
		switch (run->mmio.len) {
663 664 665
		case 4: gpr = ld_le32((u32 *)run->mmio.data); break;
		case 2: gpr = ld_le16((u16 *)run->mmio.data); break;
		case 1: gpr = *(u8 *)run->mmio.data; break;
666 667
		}
	}
668

A
Alexander Graf 已提交
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684
	if (vcpu->arch.mmio_sign_extend) {
		switch (run->mmio.len) {
#ifdef CONFIG_PPC64
		case 4:
			gpr = (s64)(s32)gpr;
			break;
#endif
		case 2:
			gpr = (s64)(s16)gpr;
			break;
		case 1:
			gpr = (s64)(s8)gpr;
			break;
		}
	}

685
	kvmppc_set_gpr(vcpu, vcpu->arch.io_gpr, gpr);
686

687 688
	switch (vcpu->arch.io_gpr & KVM_MMIO_REG_EXT_MASK) {
	case KVM_MMIO_REG_GPR:
689 690
		kvmppc_set_gpr(vcpu, vcpu->arch.io_gpr, gpr);
		break;
691
	case KVM_MMIO_REG_FPR:
692
		VCPU_FPR(vcpu, vcpu->arch.io_gpr & KVM_MMIO_REG_MASK) = gpr;
693
		break;
694
#ifdef CONFIG_PPC_BOOK3S
695 696
	case KVM_MMIO_REG_QPR:
		vcpu->arch.qpr[vcpu->arch.io_gpr & KVM_MMIO_REG_MASK] = gpr;
697
		break;
698
	case KVM_MMIO_REG_FQPR:
699
		VCPU_FPR(vcpu, vcpu->arch.io_gpr & KVM_MMIO_REG_MASK) = gpr;
700
		vcpu->arch.qpr[vcpu->arch.io_gpr & KVM_MMIO_REG_MASK] = gpr;
701
		break;
702
#endif
703 704 705
	default:
		BUG();
	}
706 707 708
}

int kvmppc_handle_load(struct kvm_run *run, struct kvm_vcpu *vcpu,
709 710
		       unsigned int rt, unsigned int bytes,
		       int is_default_endian)
711
{
712
	int idx, ret;
713 714 715 716 717 718 719 720 721
	int is_bigendian;

	if (kvmppc_need_byteswap(vcpu)) {
		/* Default endianness is "little endian". */
		is_bigendian = !is_default_endian;
	} else {
		/* Default endianness is "big endian". */
		is_bigendian = is_default_endian;
	}
722

723 724 725 726 727 728 729 730 731 732 733 734 735
	if (bytes > sizeof(run->mmio.data)) {
		printk(KERN_ERR "%s: bad MMIO length: %d\n", __func__,
		       run->mmio.len);
	}

	run->mmio.phys_addr = vcpu->arch.paddr_accessed;
	run->mmio.len = bytes;
	run->mmio.is_write = 0;

	vcpu->arch.io_gpr = rt;
	vcpu->arch.mmio_is_bigendian = is_bigendian;
	vcpu->mmio_needed = 1;
	vcpu->mmio_is_write = 0;
A
Alexander Graf 已提交
736
	vcpu->arch.mmio_sign_extend = 0;
737

738 739 740 741 742 743 744 745
	idx = srcu_read_lock(&vcpu->kvm->srcu);

	ret = kvm_io_bus_read(vcpu->kvm, KVM_MMIO_BUS, run->mmio.phys_addr,
			      bytes, &run->mmio.data);

	srcu_read_unlock(&vcpu->kvm->srcu, idx);

	if (!ret) {
A
Alexander Graf 已提交
746 747 748 749 750
		kvmppc_complete_mmio_load(vcpu, run);
		vcpu->mmio_needed = 0;
		return EMULATE_DONE;
	}

751 752
	return EMULATE_DO_MMIO;
}
753
EXPORT_SYMBOL_GPL(kvmppc_handle_load);
754

A
Alexander Graf 已提交
755 756
/* Same as above, but sign extends */
int kvmppc_handle_loads(struct kvm_run *run, struct kvm_vcpu *vcpu,
757 758
			unsigned int rt, unsigned int bytes,
			int is_default_endian)
A
Alexander Graf 已提交
759 760 761 762
{
	int r;

	vcpu->arch.mmio_sign_extend = 1;
763
	r = kvmppc_handle_load(run, vcpu, rt, bytes, is_default_endian);
A
Alexander Graf 已提交
764 765 766 767

	return r;
}

768
int kvmppc_handle_store(struct kvm_run *run, struct kvm_vcpu *vcpu,
769
			u64 val, unsigned int bytes, int is_default_endian)
770 771
{
	void *data = run->mmio.data;
772
	int idx, ret;
773 774 775 776 777 778 779 780 781
	int is_bigendian;

	if (kvmppc_need_byteswap(vcpu)) {
		/* Default endianness is "little endian". */
		is_bigendian = !is_default_endian;
	} else {
		/* Default endianness is "big endian". */
		is_bigendian = is_default_endian;
	}
782 783 784 785 786 787 788 789 790 791 792 793 794 795 796

	if (bytes > sizeof(run->mmio.data)) {
		printk(KERN_ERR "%s: bad MMIO length: %d\n", __func__,
		       run->mmio.len);
	}

	run->mmio.phys_addr = vcpu->arch.paddr_accessed;
	run->mmio.len = bytes;
	run->mmio.is_write = 1;
	vcpu->mmio_needed = 1;
	vcpu->mmio_is_write = 1;

	/* Store the value at the lowest bytes in 'data'. */
	if (is_bigendian) {
		switch (bytes) {
797
		case 8: *(u64 *)data = val; break;
798 799 800 801 802 803 804 805 806 807 808 809 810
		case 4: *(u32 *)data = val; break;
		case 2: *(u16 *)data = val; break;
		case 1: *(u8  *)data = val; break;
		}
	} else {
		/* Store LE value into 'data'. */
		switch (bytes) {
		case 4: st_le32(data, val); break;
		case 2: st_le16(data, val); break;
		case 1: *(u8 *)data = val; break;
		}
	}

811 812 813 814 815 816 817 818
	idx = srcu_read_lock(&vcpu->kvm->srcu);

	ret = kvm_io_bus_write(vcpu->kvm, KVM_MMIO_BUS, run->mmio.phys_addr,
			       bytes, &run->mmio.data);

	srcu_read_unlock(&vcpu->kvm->srcu, idx);

	if (!ret) {
A
Alexander Graf 已提交
819 820 821 822
		vcpu->mmio_needed = 0;
		return EMULATE_DONE;
	}

823 824
	return EMULATE_DO_MMIO;
}
825
EXPORT_SYMBOL_GPL(kvmppc_handle_store);
826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842

int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
{
	int r;
	sigset_t sigsaved;

	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &vcpu->sigset, &sigsaved);

	if (vcpu->mmio_needed) {
		if (!vcpu->mmio_is_write)
			kvmppc_complete_mmio_load(vcpu, run);
		vcpu->mmio_needed = 0;
	} else if (vcpu->arch.dcr_needed) {
		if (!vcpu->arch.dcr_is_write)
			kvmppc_complete_dcr_load(vcpu, run);
		vcpu->arch.dcr_needed = 0;
843 844 845 846 847 848 849
	} else if (vcpu->arch.osi_needed) {
		u64 *gprs = run->osi.gprs;
		int i;

		for (i = 0; i < 32; i++)
			kvmppc_set_gpr(vcpu, i, gprs[i]);
		vcpu->arch.osi_needed = 0;
850 851 852 853 854 855 856
	} else if (vcpu->arch.hcall_needed) {
		int i;

		kvmppc_set_gpr(vcpu, 3, run->papr_hcall.ret);
		for (i = 0; i < 9; ++i)
			kvmppc_set_gpr(vcpu, 4 + i, run->papr_hcall.args[i]);
		vcpu->arch.hcall_needed = 0;
857 858 859 860 861
#ifdef CONFIG_BOOKE
	} else if (vcpu->arch.epr_needed) {
		kvmppc_set_epr(vcpu, run->epr.epr);
		vcpu->arch.epr_needed = 0;
#endif
862 863
	}

864
	r = kvmppc_vcpu_run(run, vcpu);
865 866 867 868 869 870 871 872 873

	if (vcpu->sigset_active)
		sigprocmask(SIG_SETMASK, &sigsaved, NULL);

	return r;
}

int kvm_vcpu_ioctl_interrupt(struct kvm_vcpu *vcpu, struct kvm_interrupt *irq)
{
874
	if (irq->irq == KVM_INTERRUPT_UNSET) {
875
		kvmppc_core_dequeue_external(vcpu);
876 877 878 879
		return 0;
	}

	kvmppc_core_queue_external(vcpu, irq);
880

881
	kvm_vcpu_kick(vcpu);
882

883 884 885
	return 0;
}

886 887 888 889 890 891 892 893 894
static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu,
				     struct kvm_enable_cap *cap)
{
	int r;

	if (cap->flags)
		return -EINVAL;

	switch (cap->cap) {
895 896 897 898
	case KVM_CAP_PPC_OSI:
		r = 0;
		vcpu->arch.osi_enabled = true;
		break;
899 900 901 902
	case KVM_CAP_PPC_PAPR:
		r = 0;
		vcpu->arch.papr_enabled = true;
		break;
903 904
	case KVM_CAP_PPC_EPR:
		r = 0;
905 906 907 908
		if (cap->args[0])
			vcpu->arch.epr_flags |= KVMPPC_EPR_USER;
		else
			vcpu->arch.epr_flags &= ~KVMPPC_EPR_USER;
909
		break;
910 911 912 913 914 915
#ifdef CONFIG_BOOKE
	case KVM_CAP_PPC_BOOKE_WATCHDOG:
		r = 0;
		vcpu->arch.watchdog_enabled = true;
		break;
#endif
916
#if defined(CONFIG_KVM_E500V2) || defined(CONFIG_KVM_E500MC)
S
Scott Wood 已提交
917 918 919 920 921 922 923 924 925 926
	case KVM_CAP_SW_TLB: {
		struct kvm_config_tlb cfg;
		void __user *user_ptr = (void __user *)(uintptr_t)cap->args[0];

		r = -EFAULT;
		if (copy_from_user(&cfg, user_ptr, sizeof(cfg)))
			break;

		r = kvm_vcpu_ioctl_config_tlb(vcpu, &cfg);
		break;
S
Scott Wood 已提交
927 928 929 930
	}
#endif
#ifdef CONFIG_KVM_MPIC
	case KVM_CAP_IRQ_MPIC: {
A
Al Viro 已提交
931
		struct fd f;
S
Scott Wood 已提交
932 933 934
		struct kvm_device *dev;

		r = -EBADF;
A
Al Viro 已提交
935 936
		f = fdget(cap->args[0]);
		if (!f.file)
S
Scott Wood 已提交
937 938 939
			break;

		r = -EPERM;
A
Al Viro 已提交
940
		dev = kvm_device_from_filp(f.file);
S
Scott Wood 已提交
941 942 943
		if (dev)
			r = kvmppc_mpic_connect_vcpu(dev, vcpu, cap->args[1]);

A
Al Viro 已提交
944
		fdput(f);
S
Scott Wood 已提交
945
		break;
S
Scott Wood 已提交
946 947
	}
#endif
948 949
#ifdef CONFIG_KVM_XICS
	case KVM_CAP_IRQ_XICS: {
A
Al Viro 已提交
950
		struct fd f;
951 952 953
		struct kvm_device *dev;

		r = -EBADF;
A
Al Viro 已提交
954 955
		f = fdget(cap->args[0]);
		if (!f.file)
956 957 958
			break;

		r = -EPERM;
A
Al Viro 已提交
959
		dev = kvm_device_from_filp(f.file);
960 961 962
		if (dev)
			r = kvmppc_xics_connect_vcpu(dev, vcpu, cap->args[1]);

A
Al Viro 已提交
963
		fdput(f);
964 965 966
		break;
	}
#endif /* CONFIG_KVM_XICS */
967 968 969 970 971
	default:
		r = -EINVAL;
		break;
	}

972 973 974
	if (!r)
		r = kvmppc_sanity_check(vcpu);

975 976 977
	return r;
}

978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996
int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu,
                                    struct kvm_mp_state *mp_state)
{
	return -EINVAL;
}

int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
                                    struct kvm_mp_state *mp_state)
{
	return -EINVAL;
}

long kvm_arch_vcpu_ioctl(struct file *filp,
                         unsigned int ioctl, unsigned long arg)
{
	struct kvm_vcpu *vcpu = filp->private_data;
	void __user *argp = (void __user *)arg;
	long r;

997 998
	switch (ioctl) {
	case KVM_INTERRUPT: {
999 1000 1001
		struct kvm_interrupt irq;
		r = -EFAULT;
		if (copy_from_user(&irq, argp, sizeof(irq)))
1002
			goto out;
1003
		r = kvm_vcpu_ioctl_interrupt(vcpu, &irq);
1004
		goto out;
1005
	}
1006

1007 1008 1009 1010 1011 1012 1013 1014 1015
	case KVM_ENABLE_CAP:
	{
		struct kvm_enable_cap cap;
		r = -EFAULT;
		if (copy_from_user(&cap, argp, sizeof(cap)))
			goto out;
		r = kvm_vcpu_ioctl_enable_cap(vcpu, &cap);
		break;
	}
S
Scott Wood 已提交
1016

1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030
	case KVM_SET_ONE_REG:
	case KVM_GET_ONE_REG:
	{
		struct kvm_one_reg reg;
		r = -EFAULT;
		if (copy_from_user(&reg, argp, sizeof(reg)))
			goto out;
		if (ioctl == KVM_SET_ONE_REG)
			r = kvm_vcpu_ioctl_set_one_reg(vcpu, &reg);
		else
			r = kvm_vcpu_ioctl_get_one_reg(vcpu, &reg);
		break;
	}

1031
#if defined(CONFIG_KVM_E500V2) || defined(CONFIG_KVM_E500MC)
S
Scott Wood 已提交
1032 1033 1034 1035 1036 1037 1038 1039 1040
	case KVM_DIRTY_TLB: {
		struct kvm_dirty_tlb dirty;
		r = -EFAULT;
		if (copy_from_user(&dirty, argp, sizeof(dirty)))
			goto out;
		r = kvm_vcpu_ioctl_dirty_tlb(vcpu, &dirty);
		break;
	}
#endif
1041 1042 1043 1044 1045 1046 1047 1048
	default:
		r = -EINVAL;
	}

out:
	return r;
}

1049 1050 1051 1052 1053
int kvm_arch_vcpu_fault(struct kvm_vcpu *vcpu, struct vm_fault *vmf)
{
	return VM_FAULT_SIGBUS;
}

1054 1055
static int kvm_vm_ioctl_get_pvinfo(struct kvm_ppc_pvinfo *pvinfo)
{
1056 1057 1058
	u32 inst_nop = 0x60000000;
#ifdef CONFIG_KVM_BOOKE_HV
	u32 inst_sc1 = 0x44000022;
1059 1060 1061 1062
	pvinfo->hcall[0] = cpu_to_be32(inst_sc1);
	pvinfo->hcall[1] = cpu_to_be32(inst_nop);
	pvinfo->hcall[2] = cpu_to_be32(inst_nop);
	pvinfo->hcall[3] = cpu_to_be32(inst_nop);
1063
#else
1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
	u32 inst_lis = 0x3c000000;
	u32 inst_ori = 0x60000000;
	u32 inst_sc = 0x44000002;
	u32 inst_imm_mask = 0xffff;

	/*
	 * The hypercall to get into KVM from within guest context is as
	 * follows:
	 *
	 *    lis r0, r0, KVM_SC_MAGIC_R0@h
	 *    ori r0, KVM_SC_MAGIC_R0@l
	 *    sc
	 *    nop
	 */
1078 1079 1080 1081
	pvinfo->hcall[0] = cpu_to_be32(inst_lis | ((KVM_SC_MAGIC_R0 >> 16) & inst_imm_mask));
	pvinfo->hcall[1] = cpu_to_be32(inst_ori | (KVM_SC_MAGIC_R0 & inst_imm_mask));
	pvinfo->hcall[2] = cpu_to_be32(inst_sc);
	pvinfo->hcall[3] = cpu_to_be32(inst_nop);
1082
#endif
1083

1084 1085
	pvinfo->flags = KVM_PPC_PVINFO_FLAGS_EV_IDLE;

1086 1087 1088
	return 0;
}

1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100
int kvm_vm_ioctl_irq_line(struct kvm *kvm, struct kvm_irq_level *irq_event,
			  bool line_status)
{
	if (!irqchip_in_kernel(kvm))
		return -ENXIO;

	irq_event->status = kvm_set_irq(kvm, KVM_USERSPACE_IRQ_SOURCE_ID,
					irq_event->irq, irq_event->level,
					line_status);
	return 0;
}

1101 1102 1103
long kvm_arch_vm_ioctl(struct file *filp,
                       unsigned int ioctl, unsigned long arg)
{
1104
	struct kvm *kvm __maybe_unused = filp->private_data;
1105
	void __user *argp = (void __user *)arg;
1106 1107 1108
	long r;

	switch (ioctl) {
1109 1110
	case KVM_PPC_GET_PVINFO: {
		struct kvm_ppc_pvinfo pvinfo;
1111
		memset(&pvinfo, 0, sizeof(pvinfo));
1112 1113 1114 1115 1116 1117 1118 1119
		r = kvm_vm_ioctl_get_pvinfo(&pvinfo);
		if (copy_to_user(argp, &pvinfo, sizeof(pvinfo))) {
			r = -EFAULT;
			goto out;
		}

		break;
	}
1120
#ifdef CONFIG_PPC_BOOK3S_64
1121 1122 1123 1124 1125 1126 1127 1128 1129
	case KVM_CREATE_SPAPR_TCE: {
		struct kvm_create_spapr_tce create_tce;

		r = -EFAULT;
		if (copy_from_user(&create_tce, argp, sizeof(create_tce)))
			goto out;
		r = kvm_vm_ioctl_create_spapr_tce(kvm, &create_tce);
		goto out;
	}
1130 1131
	case KVM_PPC_GET_SMMU_INFO: {
		struct kvm_ppc_smmu_info info;
1132
		struct kvm *kvm = filp->private_data;
1133 1134

		memset(&info, 0, sizeof(info));
1135
		r = kvm->arch.kvm_ops->get_smmu_info(kvm, &info);
1136 1137 1138 1139
		if (r >= 0 && copy_to_user(argp, &info, sizeof(info)))
			r = -EFAULT;
		break;
	}
1140 1141 1142 1143 1144 1145
	case KVM_PPC_RTAS_DEFINE_TOKEN: {
		struct kvm *kvm = filp->private_data;

		r = kvm_vm_ioctl_rtas_define_token(kvm, argp);
		break;
	}
1146 1147 1148 1149
	default: {
		struct kvm *kvm = filp->private_data;
		r = kvm->arch.kvm_ops->arch_vm_ioctl(filp, ioctl, arg);
	}
1150
#else /* CONFIG_PPC_BOOK3S_64 */
1151
	default:
1152
		r = -ENOTTY;
1153
#endif
1154
	}
1155
out:
1156 1157 1158
	return r;
}

1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175
static unsigned long lpid_inuse[BITS_TO_LONGS(KVMPPC_NR_LPIDS)];
static unsigned long nr_lpids;

long kvmppc_alloc_lpid(void)
{
	long lpid;

	do {
		lpid = find_first_zero_bit(lpid_inuse, KVMPPC_NR_LPIDS);
		if (lpid >= nr_lpids) {
			pr_err("%s: No LPIDs free\n", __func__);
			return -ENOMEM;
		}
	} while (test_and_set_bit(lpid, lpid_inuse));

	return lpid;
}
1176
EXPORT_SYMBOL_GPL(kvmppc_alloc_lpid);
1177 1178 1179 1180 1181

void kvmppc_claim_lpid(long lpid)
{
	set_bit(lpid, lpid_inuse);
}
1182
EXPORT_SYMBOL_GPL(kvmppc_claim_lpid);
1183 1184 1185 1186 1187

void kvmppc_free_lpid(long lpid)
{
	clear_bit(lpid, lpid_inuse);
}
1188
EXPORT_SYMBOL_GPL(kvmppc_free_lpid);
1189 1190 1191 1192 1193 1194

void kvmppc_init_lpid(unsigned long nr_lpids_param)
{
	nr_lpids = min_t(unsigned long, KVMPPC_NR_LPIDS, nr_lpids_param);
	memset(lpid_inuse, 0, sizeof(lpid_inuse));
}
1195
EXPORT_SYMBOL_GPL(kvmppc_init_lpid);
1196

1197 1198 1199 1200 1201 1202 1203
int kvm_arch_init(void *opaque)
{
	return 0;
}

void kvm_arch_exit(void)
{
1204

1205
}