arm.c 37.0 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Copyright (C) 2012 - Virtual Open Systems and Columbia University
 * Author: Christoffer Dall <c.dall@virtualopensystems.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 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.
 */

19
#include <linux/bug.h>
20
#include <linux/cpu_pm.h>
21 22 23
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/kvm_host.h>
24
#include <linux/list.h>
25 26 27 28 29
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/fs.h>
#include <linux/mman.h>
#include <linux/sched.h>
30
#include <linux/kvm.h>
31 32
#include <linux/kvm_irqfd.h>
#include <linux/irqbypass.h>
33
#include <linux/sched/stat.h>
34
#include <trace/events/kvm.h>
35
#include <kvm/arm_pmu.h>
36
#include <kvm/arm_psci.h>
37 38 39 40

#define CREATE_TRACE_POINTS
#include "trace.h"

41
#include <linux/uaccess.h>
42 43
#include <asm/ptrace.h>
#include <asm/mman.h>
44
#include <asm/tlbflush.h>
45
#include <asm/cacheflush.h>
46
#include <asm/cpufeature.h>
47 48 49 50
#include <asm/virt.h>
#include <asm/kvm_arm.h>
#include <asm/kvm_asm.h>
#include <asm/kvm_mmu.h>
51
#include <asm/kvm_emulate.h>
52
#include <asm/kvm_coproc.h>
53
#include <asm/sections.h>
54 55 56 57 58

#ifdef REQUIRES_VIRT
__asm__(".arch_extension	virt");
#endif

59
DEFINE_PER_CPU(kvm_cpu_context_t, kvm_host_cpu_state);
60 61
static DEFINE_PER_CPU(unsigned long, kvm_arm_hyp_stack_page);

62 63 64
/* Per-CPU variable containing the currently running vcpu. */
static DEFINE_PER_CPU(struct kvm_vcpu *, kvm_arm_running_vcpu);

65 66
/* The VMID used in the VTTBR */
static atomic64_t kvm_vmid_gen = ATOMIC64_INIT(1);
67 68
static u32 kvm_next_vmid;
static unsigned int kvm_vmid_bits __read_mostly;
69
static DEFINE_RWLOCK(kvm_vmid_lock);
70

71 72
static bool vgic_present;

73 74
static DEFINE_PER_CPU(unsigned char, kvm_arm_hardware_enabled);

75 76
static void kvm_arm_set_running_vcpu(struct kvm_vcpu *vcpu)
{
77
	__this_cpu_write(kvm_arm_running_vcpu, vcpu);
78 79
}

80 81
DEFINE_STATIC_KEY_FALSE(userspace_irqchip_in_use);

82 83 84 85 86 87
/**
 * kvm_arm_get_running_vcpu - get the vcpu running on the current CPU.
 * Must be called from non-preemptible context
 */
struct kvm_vcpu *kvm_arm_get_running_vcpu(void)
{
88
	return __this_cpu_read(kvm_arm_running_vcpu);
89 90 91 92 93
}

/**
 * kvm_arm_get_running_vcpus - get the per-CPU array of currently running vcpus.
 */
94
struct kvm_vcpu * __percpu *kvm_get_running_vcpus(void)
95 96 97 98
{
	return &kvm_arm_running_vcpu;
}

99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
int kvm_arch_vcpu_should_kick(struct kvm_vcpu *vcpu)
{
	return kvm_vcpu_exiting_guest_mode(vcpu) == IN_GUEST_MODE;
}

int kvm_arch_hardware_setup(void)
{
	return 0;
}

void kvm_arch_check_processor_compat(void *rtn)
{
	*(int *)rtn = 0;
}


115 116 117 118
/**
 * kvm_arch_init_vm - initializes a VM data structure
 * @kvm:	pointer to the KVM struct
 */
119 120
int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
{
121
	int ret, cpu;
122

123
	ret = kvm_arm_setup_stage2(kvm, type);
124 125
	if (ret)
		return ret;
126

127 128 129 130 131 132 133
	kvm->arch.last_vcpu_ran = alloc_percpu(typeof(*kvm->arch.last_vcpu_ran));
	if (!kvm->arch.last_vcpu_ran)
		return -ENOMEM;

	for_each_possible_cpu(cpu)
		*per_cpu_ptr(kvm->arch.last_vcpu_ran, cpu) = -1;

134 135 136 137
	ret = kvm_alloc_stage2_pgd(kvm);
	if (ret)
		goto out_fail_alloc;

138
	ret = create_hyp_mappings(kvm, kvm + 1, PAGE_HYP);
139 140 141
	if (ret)
		goto out_free_stage2_pgd;

142
	kvm_vgic_early_init(kvm);
143

144 145 146
	/* Mark the initial VMID generation invalid */
	kvm->arch.vmid_gen = 0;

147
	/* The maximum number of VCPUs is limited by the host's GIC model */
148 149
	kvm->arch.max_vcpus = vgic_present ?
				kvm_vgic_get_max_vcpus() : KVM_MAX_VCPUS;
150

151 152 153 154
	return ret;
out_free_stage2_pgd:
	kvm_free_stage2_pgd(kvm);
out_fail_alloc:
155 156
	free_percpu(kvm->arch.last_vcpu_ran);
	kvm->arch.last_vcpu_ran = NULL;
157
	return ret;
158 159
}

160 161 162 163 164 165 166 167 168 169
bool kvm_arch_has_vcpu_debugfs(void)
{
	return false;
}

int kvm_arch_create_vcpu_debugfs(struct kvm_vcpu *vcpu)
{
	return 0;
}

170
vm_fault_t kvm_arch_vcpu_fault(struct kvm_vcpu *vcpu, struct vm_fault *vmf)
171 172 173 174 175
{
	return VM_FAULT_SIGBUS;
}


176 177 178 179
/**
 * kvm_arch_destroy_vm - destroy the VM data structure
 * @kvm:	pointer to the KVM struct
 */
180 181 182 183
void kvm_arch_destroy_vm(struct kvm *kvm)
{
	int i;

184 185
	kvm_vgic_destroy(kvm);

186 187 188
	free_percpu(kvm->arch.last_vcpu_ran);
	kvm->arch.last_vcpu_ran = NULL;

189 190 191 192 193 194
	for (i = 0; i < KVM_MAX_VCPUS; ++i) {
		if (kvm->vcpus[i]) {
			kvm_arch_vcpu_free(kvm->vcpus[i]);
			kvm->vcpus[i] = NULL;
		}
	}
195
	atomic_set(&kvm->online_vcpus, 0);
196 197
}

198
int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
199 200 201
{
	int r;
	switch (ext) {
202
	case KVM_CAP_IRQCHIP:
203 204
		r = vgic_present;
		break;
205
	case KVM_CAP_IOEVENTFD:
206
	case KVM_CAP_DEVICE_CTRL:
207 208 209 210
	case KVM_CAP_USER_MEMORY:
	case KVM_CAP_SYNC_MMU:
	case KVM_CAP_DESTROY_MEMORY_REGION_WORKS:
	case KVM_CAP_ONE_REG:
211
	case KVM_CAP_ARM_PSCI:
212
	case KVM_CAP_ARM_PSCI_0_2:
213
	case KVM_CAP_READONLY_MEM:
214
	case KVM_CAP_MP_STATE:
215
	case KVM_CAP_IMMEDIATE_EXIT:
216 217
		r = 1;
		break;
218 219
	case KVM_CAP_ARM_SET_DEVICE_ADDR:
		r = 1;
220
		break;
221 222 223 224 225 226
	case KVM_CAP_NR_VCPUS:
		r = num_online_cpus();
		break;
	case KVM_CAP_MAX_VCPUS:
		r = KVM_MAX_VCPUS;
		break;
227 228 229
	case KVM_CAP_NR_MEMSLOTS:
		r = KVM_USER_MEM_SLOTS;
		break;
V
Vladimir Murzin 已提交
230 231 232 233 234 235
	case KVM_CAP_MSI_DEVID:
		if (!kvm)
			r = -EINVAL;
		else
			r = kvm->arch.vgic.msis_require_devid;
		break;
236 237 238 239 240 241 242
	case KVM_CAP_ARM_USER_IRQ:
		/*
		 * 1: EL1_VTIMER, EL1_PTIMER, and PMU.
		 * (bump this number if adding more devices)
		 */
		r = 1;
		break;
243
	default:
244
		r = kvm_arch_dev_ioctl_check_extension(kvm, ext);
245 246 247 248 249 250 251 252 253 254 255
		break;
	}
	return r;
}

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

256 257 258 259 260 261 262 263 264 265 266 267 268 269 270
struct kvm *kvm_arch_alloc_vm(void)
{
	if (!has_vhe())
		return kzalloc(sizeof(struct kvm), GFP_KERNEL);

	return vzalloc(sizeof(struct kvm));
}

void kvm_arch_free_vm(struct kvm *kvm)
{
	if (!has_vhe())
		kfree(kvm);
	else
		vfree(kvm);
}
271 272 273 274 275 276

struct kvm_vcpu *kvm_arch_vcpu_create(struct kvm *kvm, unsigned int id)
{
	int err;
	struct kvm_vcpu *vcpu;

277 278 279 280 281
	if (irqchip_in_kernel(kvm) && vgic_initialized(kvm)) {
		err = -EBUSY;
		goto out;
	}

282 283 284 285 286
	if (id >= kvm->arch.max_vcpus) {
		err = -EINVAL;
		goto out;
	}

287 288 289 290 291 292 293 294 295 296
	vcpu = kmem_cache_zalloc(kvm_vcpu_cache, GFP_KERNEL);
	if (!vcpu) {
		err = -ENOMEM;
		goto out;
	}

	err = kvm_vcpu_init(vcpu, kvm, id);
	if (err)
		goto free_vcpu;

297
	err = create_hyp_mappings(vcpu, vcpu + 1, PAGE_HYP);
298 299 300
	if (err)
		goto vcpu_uninit;

301
	return vcpu;
302 303
vcpu_uninit:
	kvm_vcpu_uninit(vcpu);
304 305 306 307 308 309
free_vcpu:
	kmem_cache_free(kvm_vcpu_cache, vcpu);
out:
	return ERR_PTR(err);
}

310
void kvm_arch_vcpu_postcreate(struct kvm_vcpu *vcpu)
311 312 313 314 315
{
}

void kvm_arch_vcpu_free(struct kvm_vcpu *vcpu)
{
316 317 318
	if (vcpu->arch.has_run_once && unlikely(!irqchip_in_kernel(vcpu->kvm)))
		static_branch_dec(&userspace_irqchip_in_use);

319
	kvm_mmu_free_memory_caches(vcpu);
320
	kvm_timer_vcpu_terminate(vcpu);
321
	kvm_pmu_vcpu_destroy(vcpu);
322
	kvm_vcpu_uninit(vcpu);
323
	kmem_cache_free(kvm_vcpu_cache, vcpu);
324 325 326 327 328 329 330 331 332
}

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

int kvm_cpu_has_pending_timer(struct kvm_vcpu *vcpu)
{
333
	return kvm_timer_is_pending(vcpu);
334 335
}

336 337 338
void kvm_arch_vcpu_blocking(struct kvm_vcpu *vcpu)
{
	kvm_timer_schedule(vcpu);
339
	kvm_vgic_v4_enable_doorbell(vcpu);
340 341 342 343 344
}

void kvm_arch_vcpu_unblocking(struct kvm_vcpu *vcpu)
{
	kvm_timer_unschedule(vcpu);
345
	kvm_vgic_v4_disable_doorbell(vcpu);
346 347
}

348 349
int kvm_arch_vcpu_init(struct kvm_vcpu *vcpu)
{
350 351
	/* Force users to call KVM_ARM_VCPU_INIT */
	vcpu->arch.target = -1;
352
	bitmap_zero(vcpu->arch.features, KVM_VCPU_MAX_FEATURES);
353

354 355 356
	/* Set up the timer */
	kvm_timer_vcpu_init(vcpu);

357 358
	kvm_arm_reset_debug_ptr(vcpu);

359
	return kvm_vgic_vcpu_init(vcpu);
360 361 362 363
}

void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
{
364 365 366 367 368 369 370 371 372 373 374 375 376
	int *last_ran;

	last_ran = this_cpu_ptr(vcpu->kvm->arch.last_vcpu_ran);

	/*
	 * We might get preempted before the vCPU actually runs, but
	 * over-invalidation doesn't affect correctness.
	 */
	if (*last_ran != vcpu->vcpu_id) {
		kvm_call_hyp(__kvm_tlb_flush_local_vmid, vcpu);
		*last_ran = vcpu->vcpu_id;
	}

377
	vcpu->cpu = cpu;
378
	vcpu->arch.host_cpu_context = this_cpu_ptr(&kvm_host_cpu_state);
379

380
	kvm_arm_set_running_vcpu(vcpu);
381
	kvm_vgic_load(vcpu);
382
	kvm_timer_vcpu_load(vcpu);
383
	kvm_vcpu_load_sysregs(vcpu);
384
	kvm_arch_vcpu_load_fp(vcpu);
385 386 387 388 389

	if (single_task_running())
		vcpu_clear_wfe_traps(vcpu);
	else
		vcpu_set_wfe_traps(vcpu);
390 391 392 393
}

void kvm_arch_vcpu_put(struct kvm_vcpu *vcpu)
{
394
	kvm_arch_vcpu_put_fp(vcpu);
395
	kvm_vcpu_put_sysregs(vcpu);
396
	kvm_timer_vcpu_put(vcpu);
397 398
	kvm_vgic_put(vcpu);

399 400
	vcpu->cpu = -1;

401
	kvm_arm_set_running_vcpu(NULL);
402 403
}

A
Andrew Jones 已提交
404 405 406
static void vcpu_power_off(struct kvm_vcpu *vcpu)
{
	vcpu->arch.power_off = true;
407
	kvm_make_request(KVM_REQ_SLEEP, vcpu);
A
Andrew Jones 已提交
408 409 410
	kvm_vcpu_kick(vcpu);
}

411 412 413
int kvm_arch_vcpu_ioctl_get_mpstate(struct kvm_vcpu *vcpu,
				    struct kvm_mp_state *mp_state)
{
414
	if (vcpu->arch.power_off)
415 416 417 418 419
		mp_state->mp_state = KVM_MP_STATE_STOPPED;
	else
		mp_state->mp_state = KVM_MP_STATE_RUNNABLE;

	return 0;
420 421 422 423 424
}

int kvm_arch_vcpu_ioctl_set_mpstate(struct kvm_vcpu *vcpu,
				    struct kvm_mp_state *mp_state)
{
425 426
	int ret = 0;

427 428
	switch (mp_state->mp_state) {
	case KVM_MP_STATE_RUNNABLE:
429
		vcpu->arch.power_off = false;
430 431
		break;
	case KVM_MP_STATE_STOPPED:
A
Andrew Jones 已提交
432
		vcpu_power_off(vcpu);
433 434
		break;
	default:
435
		ret = -EINVAL;
436 437
	}

438
	return ret;
439 440
}

441 442 443 444 445 446 447
/**
 * kvm_arch_vcpu_runnable - determine if the vcpu can be scheduled
 * @v:		The VCPU pointer
 *
 * If the guest CPU is not waiting for interrupts or an interrupt line is
 * asserted, the CPU is by definition runnable.
 */
448 449
int kvm_arch_vcpu_runnable(struct kvm_vcpu *v)
{
450 451
	bool irq_lines = *vcpu_hcr(v) & (HCR_VI | HCR_VF);
	return ((irq_lines || kvm_vgic_vcpu_pending_irq(v))
452
		&& !v->arch.power_off && !v->arch.pause);
453 454
}

455 456
bool kvm_arch_vcpu_in_kernel(struct kvm_vcpu *vcpu)
{
457
	return vcpu_mode_priv(vcpu);
458 459
}

460 461 462 463 464 465 466
/* Just ensure a guest exit from a particular CPU */
static void exit_vm_noop(void *info)
{
}

void force_vm_exit(const cpumask_t *mask)
{
467
	preempt_disable();
468
	smp_call_function_many(mask, exit_vm_noop, NULL, true);
469
	preempt_enable();
470 471 472 473
}

/**
 * need_new_vmid_gen - check that the VMID is still valid
A
Andrea Gelmini 已提交
474
 * @kvm: The VM's VMID to check
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
 *
 * return true if there is a new generation of VMIDs being used
 *
 * The hardware supports only 256 values with the value zero reserved for the
 * host, so we check if an assigned value belongs to a previous generation,
 * which which requires us to assign a new value. If we're the first to use a
 * VMID for the new generation, we must flush necessary caches and TLBs on all
 * CPUs.
 */
static bool need_new_vmid_gen(struct kvm *kvm)
{
	return unlikely(kvm->arch.vmid_gen != atomic64_read(&kvm_vmid_gen));
}

/**
 * update_vttbr - Update the VTTBR with a valid VMID before the guest runs
 * @kvm	The guest that we are about to run
 *
 * Called from kvm_arch_vcpu_ioctl_run before entering the guest to ensure the
 * VM has a valid VMID, otherwise assigns a new one and flushes corresponding
 * caches and TLBs.
 */
static void update_vttbr(struct kvm *kvm)
{
	phys_addr_t pgd_phys;
	u64 vmid;
501
	bool new_gen;
502

503 504 505 506 507
	read_lock(&kvm_vmid_lock);
	new_gen = need_new_vmid_gen(kvm);
	read_unlock(&kvm_vmid_lock);

	if (!new_gen)
508 509
		return;

510
	write_lock(&kvm_vmid_lock);
511 512 513 514 515 516 517

	/*
	 * We need to re-check the vmid_gen here to ensure that if another vcpu
	 * already allocated a valid vmid for this vm, then this vcpu should
	 * use the same vmid.
	 */
	if (!need_new_vmid_gen(kvm)) {
518
		write_unlock(&kvm_vmid_lock);
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
		return;
	}

	/* First user of a new VMID generation? */
	if (unlikely(kvm_next_vmid == 0)) {
		atomic64_inc(&kvm_vmid_gen);
		kvm_next_vmid = 1;

		/*
		 * On SMP we know no other CPUs can use this CPU's or each
		 * other's VMID after force_vm_exit returns since the
		 * kvm_vmid_lock blocks them from reentry to the guest.
		 */
		force_vm_exit(cpu_all_mask);
		/*
		 * Now broadcast TLB + ICACHE invalidation over the inner
		 * shareable domain to make sure all data structures are
		 * clean.
		 */
		kvm_call_hyp(__kvm_flush_vm_context);
	}

	kvm->arch.vmid_gen = atomic64_read(&kvm_vmid_gen);
	kvm->arch.vmid = kvm_next_vmid;
	kvm_next_vmid++;
544
	kvm_next_vmid &= (1 << kvm_vmid_bits) - 1;
545 546

	/* update vttbr to be used with the new vmid */
547
	pgd_phys = virt_to_phys(kvm->arch.pgd);
548
	BUG_ON(pgd_phys & ~kvm_vttbr_baddr_mask(kvm));
549
	vmid = ((u64)(kvm->arch.vmid) << VTTBR_VMID_SHIFT) & VTTBR_VMID_MASK(kvm_vmid_bits);
550
	kvm->arch.vttbr = kvm_phys_to_vttbr(pgd_phys) | vmid;
551

552
	write_unlock(&kvm_vmid_lock);
553 554 555 556
}

static int kvm_vcpu_first_run_init(struct kvm_vcpu *vcpu)
{
557
	struct kvm *kvm = vcpu->kvm;
558
	int ret = 0;
559

560 561 562 563
	if (likely(vcpu->arch.has_run_once))
		return 0;

	vcpu->arch.has_run_once = true;
564

565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580
	if (likely(irqchip_in_kernel(kvm))) {
		/*
		 * Map the VGIC hardware resources before running a vcpu the
		 * first time on this VM.
		 */
		if (unlikely(!vgic_ready(kvm))) {
			ret = kvm_vgic_map_resources(kvm);
			if (ret)
				return ret;
		}
	} else {
		/*
		 * Tell the rest of the code that there are userspace irqchip
		 * VMs in the wild.
		 */
		static_branch_inc(&userspace_irqchip_in_use);
581 582
	}

583
	ret = kvm_timer_enable(vcpu);
584 585 586 587
	if (ret)
		return ret;

	ret = kvm_arm_pmu_v3_enable(vcpu);
588

589
	return ret;
590 591
}

592 593 594 595 596
bool kvm_arch_intc_initialized(struct kvm *kvm)
{
	return vgic_initialized(kvm);
}

597
void kvm_arm_halt_guest(struct kvm *kvm)
598 599 600 601 602 603
{
	int i;
	struct kvm_vcpu *vcpu;

	kvm_for_each_vcpu(i, vcpu, kvm)
		vcpu->arch.pause = true;
604
	kvm_make_all_cpus_request(kvm, KVM_REQ_SLEEP);
605 606
}

607
void kvm_arm_resume_guest(struct kvm *kvm)
608 609 610 611
{
	int i;
	struct kvm_vcpu *vcpu;

612 613
	kvm_for_each_vcpu(i, vcpu, kvm) {
		vcpu->arch.pause = false;
614
		swake_up_one(kvm_arch_vcpu_wq(vcpu));
615
	}
616 617
}

618
static void vcpu_req_sleep(struct kvm_vcpu *vcpu)
619
{
620
	struct swait_queue_head *wq = kvm_arch_vcpu_wq(vcpu);
621

622
	swait_event_interruptible_exclusive(*wq, ((!vcpu->arch.power_off) &&
623
				       (!vcpu->arch.pause)));
624

A
Andrew Jones 已提交
625
	if (vcpu->arch.power_off || vcpu->arch.pause) {
626
		/* Awaken to handle a signal, request we sleep again later. */
627
		kvm_make_request(KVM_REQ_SLEEP, vcpu);
628
	}
629 630
}

631 632 633 634 635
static int kvm_vcpu_initialized(struct kvm_vcpu *vcpu)
{
	return vcpu->arch.target >= 0;
}

636 637 638
static void check_vcpu_requests(struct kvm_vcpu *vcpu)
{
	if (kvm_request_pending(vcpu)) {
639 640
		if (kvm_check_request(KVM_REQ_SLEEP, vcpu))
			vcpu_req_sleep(vcpu);
641 642 643 644 645 646

		/*
		 * Clear IRQ_PENDING requests that were made to guarantee
		 * that a VCPU sees new virtual interrupts.
		 */
		kvm_check_request(KVM_REQ_IRQ_PENDING, vcpu);
647 648 649
	}
}

650 651 652 653 654 655 656 657 658 659 660
/**
 * kvm_arch_vcpu_ioctl_run - the main VCPU run function to execute guest code
 * @vcpu:	The VCPU pointer
 * @run:	The kvm_run structure pointer used for userspace state exchange
 *
 * This function is called through the VCPU_RUN ioctl called from user space. It
 * will execute VM code in a loop until the time slice for the process is used
 * or some emulation is needed from user space in which case the function will
 * return with return value 0 and with the kvm_run structure filled in with the
 * required data for the requested emulation.
 */
661 662
int kvm_arch_vcpu_ioctl_run(struct kvm_vcpu *vcpu, struct kvm_run *run)
{
663 664
	int ret;

665
	if (unlikely(!kvm_vcpu_initialized(vcpu)))
666 667 668 669
		return -ENOEXEC;

	ret = kvm_vcpu_first_run_init(vcpu);
	if (ret)
670
		return ret;
671

C
Christoffer Dall 已提交
672 673 674
	if (run->exit_reason == KVM_EXIT_MMIO) {
		ret = kvm_handle_mmio_return(vcpu, vcpu->run);
		if (ret)
675 676 677
			return ret;
		if (kvm_arm_handle_step_debug(vcpu, vcpu->run))
			return 0;
C
Christoffer Dall 已提交
678 679
	}

680 681 682 683
	if (run->immediate_exit)
		return -EINTR;

	vcpu_load(vcpu);
684

685
	kvm_sigset_activate(vcpu);
686 687 688 689 690 691 692 693 694 695 696

	ret = 1;
	run->exit_reason = KVM_EXIT_UNKNOWN;
	while (ret > 0) {
		/*
		 * Check conditions before entering the guest
		 */
		cond_resched();

		update_vttbr(vcpu->kvm);

697 698
		check_vcpu_requests(vcpu);

699 700 701 702 703
		/*
		 * Preparing the interrupts to be injected also
		 * involves poking the GIC, which must be done in a
		 * non-preemptible context.
		 */
704
		preempt_disable();
705

706
		kvm_pmu_flush_hwstate(vcpu);
707

708 709
		local_irq_disable();

710 711
		kvm_vgic_flush_hwstate(vcpu);

712
		/*
713 714
		 * Exit if we have a signal pending so that we can deliver the
		 * signal to user space.
715
		 */
716
		if (signal_pending(current)) {
717 718 719 720
			ret = -EINTR;
			run->exit_reason = KVM_EXIT_INTR;
		}

721 722 723 724 725 726 727 728 729 730 731 732 733 734 735
		/*
		 * If we're using a userspace irqchip, then check if we need
		 * to tell a userspace irqchip about timer or PMU level
		 * changes and if so, exit to userspace (the actual level
		 * state gets updated in kvm_timer_update_run and
		 * kvm_pmu_update_run below).
		 */
		if (static_branch_unlikely(&userspace_irqchip_in_use)) {
			if (kvm_timer_should_notify_user(vcpu) ||
			    kvm_pmu_should_notify_user(vcpu)) {
				ret = -EINTR;
				run->exit_reason = KVM_EXIT_INTR;
			}
		}

736 737 738 739 740 741 742 743
		/*
		 * Ensure we set mode to IN_GUEST_MODE after we disable
		 * interrupts and before the final VCPU requests check.
		 * See the comment in kvm_vcpu_exiting_guest_mode() and
		 * Documentation/virtual/kvm/vcpu-requests.rst
		 */
		smp_store_mb(vcpu->mode, IN_GUEST_MODE);

744
		if (ret <= 0 || need_new_vmid_gen(vcpu->kvm) ||
A
Andrew Jones 已提交
745
		    kvm_request_pending(vcpu)) {
746
			vcpu->mode = OUTSIDE_GUEST_MODE;
747
			isb(); /* Ensure work in x_flush_hwstate is committed */
748
			kvm_pmu_sync_hwstate(vcpu);
749 750
			if (static_branch_unlikely(&userspace_irqchip_in_use))
				kvm_timer_sync_hwstate(vcpu);
751
			kvm_vgic_sync_hwstate(vcpu);
752
			local_irq_enable();
753
			preempt_enable();
754 755 756
			continue;
		}

757 758
		kvm_arm_setup_debug(vcpu);

759 760 761 762
		/**************************************************************
		 * Enter the guest
		 */
		trace_kvm_entry(*vcpu_pc(vcpu));
763
		guest_enter_irqoff();
764

765 766 767
		if (has_vhe()) {
			kvm_arm_vhe_guest_enter();
			ret = kvm_vcpu_run_vhe(vcpu);
768
			kvm_arm_vhe_guest_exit();
769 770 771 772
		} else {
			ret = kvm_call_hyp(__kvm_vcpu_run_nvhe, vcpu);
		}

773
		vcpu->mode = OUTSIDE_GUEST_MODE;
774
		vcpu->stat.exits++;
775 776 777 778
		/*
		 * Back from guest
		 *************************************************************/

779 780
		kvm_arm_clear_debug(vcpu);

781
		/*
782
		 * We must sync the PMU state before the vgic state so
783 784 785 786 787
		 * that the vgic can properly sample the updated state of the
		 * interrupt line.
		 */
		kvm_pmu_sync_hwstate(vcpu);

788 789 790 791 792
		/*
		 * Sync the vgic state before syncing the timer state because
		 * the timer code needs to know if the virtual timer
		 * interrupts are active.
		 */
793 794
		kvm_vgic_sync_hwstate(vcpu);

795 796 797 798 799
		/*
		 * Sync the timer hardware state before enabling interrupts as
		 * we don't want vtimer interrupts to race with syncing the
		 * timer virtual interrupt state.
		 */
800 801
		if (static_branch_unlikely(&userspace_irqchip_in_use))
			kvm_timer_sync_hwstate(vcpu);
802

803 804
		kvm_arch_vcpu_ctxsync_fp(vcpu);

805 806 807 808 809 810 811 812 813 814 815 816 817
		/*
		 * We may have taken a host interrupt in HYP mode (ie
		 * while executing the guest). This interrupt is still
		 * pending, as we haven't serviced it yet!
		 *
		 * We're now back in SVC mode, with interrupts
		 * disabled.  Enabling the interrupts now will have
		 * the effect of taking the interrupt again, in SVC
		 * mode this time.
		 */
		local_irq_enable();

		/*
818
		 * We do local_irq_enable() before calling guest_exit() so
819 820
		 * that if a timer interrupt hits while running the guest we
		 * account that tick as being spent in the guest.  We enable
821
		 * preemption after calling guest_exit() so that if we get
822 823 824
		 * preempted we make sure ticks after that is not counted as
		 * guest time.
		 */
825
		guest_exit();
826
		trace_kvm_exit(ret, kvm_vcpu_trap_get_class(vcpu), *vcpu_pc(vcpu));
827

828 829 830
		/* Exit types that need handling before we can be preempted */
		handle_exit_early(vcpu, run, ret);

831 832
		preempt_enable();

833 834 835
		ret = handle_exit(vcpu, run, ret);
	}

836
	/* Tell userspace about in-kernel device output levels */
837 838 839 840
	if (unlikely(!irqchip_in_kernel(vcpu->kvm))) {
		kvm_timer_update_run(vcpu);
		kvm_pmu_update_run(vcpu);
	}
841

842 843
	kvm_sigset_deactivate(vcpu);

844
	vcpu_put(vcpu);
845
	return ret;
846 847
}

848 849 850 851
static int vcpu_interrupt_line(struct kvm_vcpu *vcpu, int number, bool level)
{
	int bit_index;
	bool set;
852
	unsigned long *hcr;
853 854 855 856 857 858

	if (number == KVM_ARM_IRQ_CPU_IRQ)
		bit_index = __ffs(HCR_VI);
	else /* KVM_ARM_IRQ_CPU_FIQ */
		bit_index = __ffs(HCR_VF);

859
	hcr = vcpu_hcr(vcpu);
860
	if (level)
861
		set = test_and_set_bit(bit_index, hcr);
862
	else
863
		set = test_and_clear_bit(bit_index, hcr);
864 865 866 867 868 869 870 871 872 873 874 875

	/*
	 * If we didn't change anything, no need to wake up or kick other CPUs
	 */
	if (set == level)
		return 0;

	/*
	 * The vcpu irq_lines field was updated, wake up sleeping VCPUs and
	 * trigger a world-switch round on the running physical CPU to set the
	 * virtual IRQ/FIQ fields in the HCR appropriately.
	 */
876
	kvm_make_request(KVM_REQ_IRQ_PENDING, vcpu);
877 878 879 880 881
	kvm_vcpu_kick(vcpu);

	return 0;
}

882 883
int kvm_vm_ioctl_irq_line(struct kvm *kvm, struct kvm_irq_level *irq_level,
			  bool line_status)
884 885 886 887 888 889 890 891 892 893 894 895 896
{
	u32 irq = irq_level->irq;
	unsigned int irq_type, vcpu_idx, irq_num;
	int nrcpus = atomic_read(&kvm->online_vcpus);
	struct kvm_vcpu *vcpu = NULL;
	bool level = irq_level->level;

	irq_type = (irq >> KVM_ARM_IRQ_TYPE_SHIFT) & KVM_ARM_IRQ_TYPE_MASK;
	vcpu_idx = (irq >> KVM_ARM_IRQ_VCPU_SHIFT) & KVM_ARM_IRQ_VCPU_MASK;
	irq_num = (irq >> KVM_ARM_IRQ_NUM_SHIFT) & KVM_ARM_IRQ_NUM_MASK;

	trace_kvm_irq_line(irq_type, vcpu_idx, irq_num, irq_level->level);

897 898 899 900
	switch (irq_type) {
	case KVM_ARM_IRQ_TYPE_CPU:
		if (irqchip_in_kernel(kvm))
			return -ENXIO;
901

902 903
		if (vcpu_idx >= nrcpus)
			return -EINVAL;
904

905 906 907
		vcpu = kvm_get_vcpu(kvm, vcpu_idx);
		if (!vcpu)
			return -EINVAL;
908

909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925
		if (irq_num > KVM_ARM_IRQ_CPU_FIQ)
			return -EINVAL;

		return vcpu_interrupt_line(vcpu, irq_num, level);
	case KVM_ARM_IRQ_TYPE_PPI:
		if (!irqchip_in_kernel(kvm))
			return -ENXIO;

		if (vcpu_idx >= nrcpus)
			return -EINVAL;

		vcpu = kvm_get_vcpu(kvm, vcpu_idx);
		if (!vcpu)
			return -EINVAL;

		if (irq_num < VGIC_NR_SGIS || irq_num >= VGIC_NR_PRIVATE_IRQS)
			return -EINVAL;
926

927
		return kvm_vgic_inject_irq(kvm, vcpu->vcpu_id, irq_num, level, NULL);
928 929 930 931
	case KVM_ARM_IRQ_TYPE_SPI:
		if (!irqchip_in_kernel(kvm))
			return -ENXIO;

932
		if (irq_num < VGIC_NR_PRIVATE_IRQS)
933 934
			return -EINVAL;

935
		return kvm_vgic_inject_irq(kvm, 0, irq_num, level, NULL);
936 937 938
	}

	return -EINVAL;
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
static int kvm_vcpu_set_target(struct kvm_vcpu *vcpu,
			       const struct kvm_vcpu_init *init)
{
	unsigned int i;
	int phys_target = kvm_target_cpu();

	if (init->target != phys_target)
		return -EINVAL;

	/*
	 * Secondary and subsequent calls to KVM_ARM_VCPU_INIT must
	 * use the same target.
	 */
	if (vcpu->arch.target != -1 && vcpu->arch.target != init->target)
		return -EINVAL;

	/* -ENOENT for unknown features, -EINVAL for invalid combinations. */
	for (i = 0; i < sizeof(init->features) * 8; i++) {
		bool set = (init->features[i / 32] & (1 << (i % 32)));

		if (set && i >= KVM_VCPU_MAX_FEATURES)
			return -ENOENT;

		/*
		 * Secondary and subsequent calls to KVM_ARM_VCPU_INIT must
		 * use the same feature set.
		 */
		if (vcpu->arch.target != -1 && i < KVM_VCPU_MAX_FEATURES &&
		    test_bit(i, vcpu->arch.features) != set)
			return -EINVAL;

		if (set)
			set_bit(i, vcpu->arch.features);
	}

	vcpu->arch.target = phys_target;

	/* Now we know what it is, we can reset it. */
	return kvm_reset_vcpu(vcpu);
}


983 984 985 986 987 988 989 990 991
static int kvm_arch_vcpu_ioctl_vcpu_init(struct kvm_vcpu *vcpu,
					 struct kvm_vcpu_init *init)
{
	int ret;

	ret = kvm_vcpu_set_target(vcpu, init);
	if (ret)
		return ret;

992 993 994 995 996 997 998
	/*
	 * Ensure a rebooted VM will fault in RAM pages and detect if the
	 * guest MMU is turned off and flush the caches as needed.
	 */
	if (vcpu->arch.has_run_once)
		stage2_unmap_vm(vcpu->kvm);

999 1000
	vcpu_reset_hcr(vcpu);

1001
	/*
1002
	 * Handle the "start in power-off" case.
1003
	 */
1004
	if (test_bit(KVM_ARM_VCPU_POWER_OFF, vcpu->arch.features))
A
Andrew Jones 已提交
1005
		vcpu_power_off(vcpu);
1006
	else
1007
		vcpu->arch.power_off = false;
1008 1009 1010 1011

	return 0;
}

1012 1013 1014 1015 1016 1017 1018
static int kvm_arm_vcpu_set_attr(struct kvm_vcpu *vcpu,
				 struct kvm_device_attr *attr)
{
	int ret = -ENXIO;

	switch (attr->group) {
	default:
1019
		ret = kvm_arm_vcpu_arch_set_attr(vcpu, attr);
1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
		break;
	}

	return ret;
}

static int kvm_arm_vcpu_get_attr(struct kvm_vcpu *vcpu,
				 struct kvm_device_attr *attr)
{
	int ret = -ENXIO;

	switch (attr->group) {
	default:
1033
		ret = kvm_arm_vcpu_arch_get_attr(vcpu, attr);
1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046
		break;
	}

	return ret;
}

static int kvm_arm_vcpu_has_attr(struct kvm_vcpu *vcpu,
				 struct kvm_device_attr *attr)
{
	int ret = -ENXIO;

	switch (attr->group) {
	default:
1047
		ret = kvm_arm_vcpu_arch_has_attr(vcpu, attr);
1048 1049 1050 1051 1052 1053
		break;
	}

	return ret;
}

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
static int kvm_arm_vcpu_get_events(struct kvm_vcpu *vcpu,
				   struct kvm_vcpu_events *events)
{
	memset(events, 0, sizeof(*events));

	return __kvm_arm_vcpu_get_events(vcpu, events);
}

static int kvm_arm_vcpu_set_events(struct kvm_vcpu *vcpu,
				   struct kvm_vcpu_events *events)
{
	int i;

	/* check whether the reserved field is zero */
	for (i = 0; i < ARRAY_SIZE(events->reserved); i++)
		if (events->reserved[i])
			return -EINVAL;

	/* check whether the pad field is zero */
	for (i = 0; i < ARRAY_SIZE(events->exception.pad); i++)
		if (events->exception.pad[i])
			return -EINVAL;

	return __kvm_arm_vcpu_set_events(vcpu, events);
}

1080 1081 1082 1083 1084
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;
1085
	struct kvm_device_attr attr;
1086 1087
	long r;

1088 1089 1090 1091
	switch (ioctl) {
	case KVM_ARM_VCPU_INIT: {
		struct kvm_vcpu_init init;

1092
		r = -EFAULT;
1093
		if (copy_from_user(&init, argp, sizeof(init)))
1094
			break;
1095

1096 1097
		r = kvm_arch_vcpu_ioctl_vcpu_init(vcpu, &init);
		break;
1098 1099 1100 1101
	}
	case KVM_SET_ONE_REG:
	case KVM_GET_ONE_REG: {
		struct kvm_one_reg reg;
1102

1103
		r = -ENOEXEC;
1104
		if (unlikely(!kvm_vcpu_initialized(vcpu)))
1105
			break;
1106

1107
		r = -EFAULT;
1108
		if (copy_from_user(&reg, argp, sizeof(reg)))
1109 1110
			break;

1111
		if (ioctl == KVM_SET_ONE_REG)
1112
			r = kvm_arm_set_reg(vcpu, &reg);
1113
		else
1114 1115
			r = kvm_arm_get_reg(vcpu, &reg);
		break;
1116 1117 1118 1119 1120 1121
	}
	case KVM_GET_REG_LIST: {
		struct kvm_reg_list __user *user_list = argp;
		struct kvm_reg_list reg_list;
		unsigned n;

1122
		r = -ENOEXEC;
1123
		if (unlikely(!kvm_vcpu_initialized(vcpu)))
1124
			break;
1125

1126
		r = -EFAULT;
1127
		if (copy_from_user(&reg_list, user_list, sizeof(reg_list)))
1128
			break;
1129 1130 1131
		n = reg_list.n;
		reg_list.n = kvm_arm_num_regs(vcpu);
		if (copy_to_user(user_list, &reg_list, sizeof(reg_list)))
1132 1133
			break;
		r = -E2BIG;
1134
		if (n < reg_list.n)
1135 1136 1137
			break;
		r = kvm_arm_copy_reg_indices(vcpu, user_list->reg);
		break;
1138
	}
1139
	case KVM_SET_DEVICE_ATTR: {
1140
		r = -EFAULT;
1141
		if (copy_from_user(&attr, argp, sizeof(attr)))
1142 1143 1144
			break;
		r = kvm_arm_vcpu_set_attr(vcpu, &attr);
		break;
1145 1146
	}
	case KVM_GET_DEVICE_ATTR: {
1147
		r = -EFAULT;
1148
		if (copy_from_user(&attr, argp, sizeof(attr)))
1149 1150 1151
			break;
		r = kvm_arm_vcpu_get_attr(vcpu, &attr);
		break;
1152 1153
	}
	case KVM_HAS_DEVICE_ATTR: {
1154
		r = -EFAULT;
1155
		if (copy_from_user(&attr, argp, sizeof(attr)))
1156 1157 1158
			break;
		r = kvm_arm_vcpu_has_attr(vcpu, &attr);
		break;
1159
	}
1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
	case KVM_GET_VCPU_EVENTS: {
		struct kvm_vcpu_events events;

		if (kvm_arm_vcpu_get_events(vcpu, &events))
			return -EINVAL;

		if (copy_to_user(argp, &events, sizeof(events)))
			return -EFAULT;

		return 0;
	}
	case KVM_SET_VCPU_EVENTS: {
		struct kvm_vcpu_events events;

		if (copy_from_user(&events, argp, sizeof(events)))
			return -EFAULT;

		return kvm_arm_vcpu_set_events(vcpu, &events);
	}
1179
	default:
1180
		r = -EINVAL;
1181
	}
1182 1183

	return r;
1184 1185
}

1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
/**
 * kvm_vm_ioctl_get_dirty_log - get and clear the log of dirty pages in a slot
 * @kvm: kvm instance
 * @log: slot id and address to which we copy the log
 *
 * Steps 1-4 below provide general overview of dirty page logging. See
 * kvm_get_dirty_log_protect() function description for additional details.
 *
 * We call kvm_get_dirty_log_protect() to handle steps 1-3, upon return we
 * always flush the TLB (step 4) even if previous step failed  and the dirty
 * bitmap may be corrupt. Regardless of previous outcome the KVM logging API
 * does not preclude user space subsequent dirty log read. Flushing TLB ensures
 * writes will be marked dirty for next log read.
 *
 *   1. Take a snapshot of the bit and clear it if needed.
 *   2. Write protect the corresponding page.
 *   3. Copy the snapshot to the userspace.
 *   4. Flush TLB's if needed.
 */
1205 1206
int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm, struct kvm_dirty_log *log)
{
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218
	bool is_dirty = false;
	int r;

	mutex_lock(&kvm->slots_lock);

	r = kvm_get_dirty_log_protect(kvm, log, &is_dirty);

	if (is_dirty)
		kvm_flush_remote_tlbs(kvm);

	mutex_unlock(&kvm->slots_lock);
	return r;
1219 1220
}

1221 1222 1223
static int kvm_vm_ioctl_set_device_addr(struct kvm *kvm,
					struct kvm_arm_device_addr *dev_addr)
{
1224 1225 1226 1227 1228 1229 1230 1231 1232
	unsigned long dev_id, type;

	dev_id = (dev_addr->id & KVM_ARM_DEVICE_ID_MASK) >>
		KVM_ARM_DEVICE_ID_SHIFT;
	type = (dev_addr->id & KVM_ARM_DEVICE_TYPE_MASK) >>
		KVM_ARM_DEVICE_TYPE_SHIFT;

	switch (dev_id) {
	case KVM_ARM_DEVICE_VGIC_V2:
1233 1234
		if (!vgic_present)
			return -ENXIO;
1235
		return kvm_vgic_addr(kvm, type, &dev_addr->addr, true);
1236 1237 1238
	default:
		return -ENODEV;
	}
1239 1240
}

1241 1242 1243
long kvm_arch_vm_ioctl(struct file *filp,
		       unsigned int ioctl, unsigned long arg)
{
1244 1245 1246 1247
	struct kvm *kvm = filp->private_data;
	void __user *argp = (void __user *)arg;

	switch (ioctl) {
1248
	case KVM_CREATE_IRQCHIP: {
1249
		int ret;
1250 1251
		if (!vgic_present)
			return -ENXIO;
1252 1253 1254 1255
		mutex_lock(&kvm->lock);
		ret = kvm_vgic_create(kvm, KVM_DEV_TYPE_ARM_VGIC_V2);
		mutex_unlock(&kvm->lock);
		return ret;
1256
	}
1257 1258 1259 1260 1261 1262 1263
	case KVM_ARM_SET_DEVICE_ADDR: {
		struct kvm_arm_device_addr dev_addr;

		if (copy_from_user(&dev_addr, argp, sizeof(dev_addr)))
			return -EFAULT;
		return kvm_vm_ioctl_set_device_addr(kvm, &dev_addr);
	}
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
	case KVM_ARM_PREFERRED_TARGET: {
		int err;
		struct kvm_vcpu_init init;

		err = kvm_vcpu_preferred_target(&init);
		if (err)
			return err;

		if (copy_to_user(argp, &init, sizeof(init)))
			return -EFAULT;

		return 0;
	}
1277 1278 1279
	default:
		return -EINVAL;
	}
1280 1281
}

1282
static void cpu_init_hyp_mode(void *dummy)
1283
{
1284
	phys_addr_t pgd_ptr;
1285 1286 1287 1288 1289
	unsigned long hyp_stack_ptr;
	unsigned long stack_page;
	unsigned long vector_ptr;

	/* Switch from the HYP stub to our own HYP init vector */
1290
	__hyp_set_vectors(kvm_get_idmap_vector());
1291

1292
	pgd_ptr = kvm_mmu_get_httbr();
1293
	stack_page = __this_cpu_read(kvm_arm_hyp_stack_page);
1294
	hyp_stack_ptr = stack_page + PAGE_SIZE;
1295
	vector_ptr = (unsigned long)kvm_get_hyp_vector();
1296

M
Marc Zyngier 已提交
1297
	__cpu_init_hyp_mode(pgd_ptr, hyp_stack_ptr, vector_ptr);
1298
	__cpu_init_stage2();
1299 1300
}

1301 1302 1303 1304 1305 1306
static void cpu_hyp_reset(void)
{
	if (!is_kernel_in_hyp_mode())
		__hyp_reset_vectors();
}

1307 1308
static void cpu_hyp_reinit(void)
{
1309 1310
	cpu_hyp_reset();

1311
	if (is_kernel_in_hyp_mode())
1312
		kvm_timer_init_vhe();
1313
	else
1314
		cpu_init_hyp_mode(NULL);
1315

1316 1317
	kvm_arm_init_debug();

1318 1319
	if (vgic_present)
		kvm_vgic_init_cpu_hardware();
1320 1321
}

1322 1323 1324
static void _kvm_arch_hardware_enable(void *discard)
{
	if (!__this_cpu_read(kvm_arm_hardware_enabled)) {
1325
		cpu_hyp_reinit();
1326
		__this_cpu_write(kvm_arm_hardware_enabled, 1);
1327
	}
1328
}
1329

1330 1331 1332 1333
int kvm_arch_hardware_enable(void)
{
	_kvm_arch_hardware_enable(NULL);
	return 0;
1334 1335
}

1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
static void _kvm_arch_hardware_disable(void *discard)
{
	if (__this_cpu_read(kvm_arm_hardware_enabled)) {
		cpu_hyp_reset();
		__this_cpu_write(kvm_arm_hardware_enabled, 0);
	}
}

void kvm_arch_hardware_disable(void)
{
	_kvm_arch_hardware_disable(NULL);
}
1348

1349 1350 1351 1352 1353
#ifdef CONFIG_CPU_PM
static int hyp_init_cpu_pm_notifier(struct notifier_block *self,
				    unsigned long cmd,
				    void *v)
{
1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368
	/*
	 * kvm_arm_hardware_enabled is left with its old value over
	 * PM_ENTER->PM_EXIT. It is used to indicate PM_EXIT should
	 * re-enable hyp.
	 */
	switch (cmd) {
	case CPU_PM_ENTER:
		if (__this_cpu_read(kvm_arm_hardware_enabled))
			/*
			 * don't update kvm_arm_hardware_enabled here
			 * so that the hardware will be re-enabled
			 * when we resume. See below.
			 */
			cpu_hyp_reset();

1369
		return NOTIFY_OK;
1370
	case CPU_PM_ENTER_FAILED:
1371 1372 1373 1374
	case CPU_PM_EXIT:
		if (__this_cpu_read(kvm_arm_hardware_enabled))
			/* The hardware was enabled before suspend. */
			cpu_hyp_reinit();
1375

1376 1377 1378 1379 1380
		return NOTIFY_OK;

	default:
		return NOTIFY_DONE;
	}
1381 1382 1383 1384 1385 1386 1387 1388 1389 1390
}

static struct notifier_block hyp_init_cpu_pm_nb = {
	.notifier_call = hyp_init_cpu_pm_notifier,
};

static void __init hyp_cpu_pm_init(void)
{
	cpu_pm_register_notifier(&hyp_init_cpu_pm_nb);
}
1391 1392 1393 1394
static void __init hyp_cpu_pm_exit(void)
{
	cpu_pm_unregister_notifier(&hyp_init_cpu_pm_nb);
}
1395 1396 1397 1398
#else
static inline void hyp_cpu_pm_init(void)
{
}
1399 1400 1401
static inline void hyp_cpu_pm_exit(void)
{
}
1402 1403
#endif

1404 1405
static int init_common_resources(void)
{
1406 1407 1408 1409
	/* set size of VMID supported by CPU */
	kvm_vmid_bits = kvm_get_vmid_bits();
	kvm_info("%d-bit VMID\n", kvm_vmid_bits);

1410 1411
	kvm_set_ipa_limit();

1412 1413 1414 1415 1416
	return 0;
}

static int init_subsystems(void)
{
1417
	int err = 0;
1418

1419
	/*
1420
	 * Enable hardware so that subsystem initialisation can access EL2.
1421
	 */
1422
	on_each_cpu(_kvm_arch_hardware_enable, NULL, 1);
1423 1424 1425 1426 1427 1428

	/*
	 * Register CPU lower-power notifier
	 */
	hyp_cpu_pm_init();

1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439
	/*
	 * Init HYP view of VGIC
	 */
	err = kvm_vgic_hyp_init();
	switch (err) {
	case 0:
		vgic_present = true;
		break;
	case -ENODEV:
	case -ENXIO:
		vgic_present = false;
1440
		err = 0;
1441 1442
		break;
	default:
1443
		goto out;
1444 1445 1446 1447 1448
	}

	/*
	 * Init HYP architected timer support
	 */
1449
	err = kvm_timer_hyp_init(vgic_present);
1450
	if (err)
1451
		goto out;
1452 1453 1454 1455

	kvm_perf_init();
	kvm_coproc_table_init();

1456 1457 1458 1459
out:
	on_each_cpu(_kvm_arch_hardware_disable, NULL, 1);

	return err;
1460 1461 1462 1463 1464 1465 1466 1467 1468
}

static void teardown_hyp_mode(void)
{
	int cpu;

	free_hyp_pgds();
	for_each_possible_cpu(cpu)
		free_page(per_cpu(kvm_arm_hyp_stack_page, cpu));
1469
	hyp_cpu_pm_exit();
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
/**
 * Inits Hyp-mode on all online CPUs
 */
static int init_hyp_mode(void)
{
	int cpu;
	int err = 0;

	/*
	 * Allocate Hyp PGD and setup Hyp identity mapping
	 */
	err = kvm_mmu_init();
	if (err)
		goto out_err;

	/*
	 * Allocate stack pages for Hypervisor-mode
	 */
	for_each_possible_cpu(cpu) {
		unsigned long stack_page;

		stack_page = __get_free_page(GFP_KERNEL);
		if (!stack_page) {
			err = -ENOMEM;
1496
			goto out_err;
1497 1498 1499 1500 1501 1502 1503 1504
		}

		per_cpu(kvm_arm_hyp_stack_page, cpu) = stack_page;
	}

	/*
	 * Map the Hyp-code called directly from the host
	 */
1505
	err = create_hyp_mappings(kvm_ksym_ref(__hyp_text_start),
1506
				  kvm_ksym_ref(__hyp_text_end), PAGE_HYP_EXEC);
1507 1508
	if (err) {
		kvm_err("Cannot map world-switch code\n");
1509
		goto out_err;
1510 1511
	}

1512
	err = create_hyp_mappings(kvm_ksym_ref(__start_rodata),
1513
				  kvm_ksym_ref(__end_rodata), PAGE_HYP_RO);
1514 1515
	if (err) {
		kvm_err("Cannot map rodata section\n");
M
Marc Zyngier 已提交
1516 1517 1518 1519 1520 1521 1522
		goto out_err;
	}

	err = create_hyp_mappings(kvm_ksym_ref(__bss_start),
				  kvm_ksym_ref(__bss_stop), PAGE_HYP_RO);
	if (err) {
		kvm_err("Cannot map bss section\n");
1523
		goto out_err;
1524 1525
	}

1526 1527 1528 1529 1530 1531
	err = kvm_map_vectors();
	if (err) {
		kvm_err("Cannot map vectors\n");
		goto out_err;
	}

1532 1533 1534 1535 1536
	/*
	 * Map the Hyp stack pages
	 */
	for_each_possible_cpu(cpu) {
		char *stack_page = (char *)per_cpu(kvm_arm_hyp_stack_page, cpu);
1537 1538
		err = create_hyp_mappings(stack_page, stack_page + PAGE_SIZE,
					  PAGE_HYP);
1539 1540 1541

		if (err) {
			kvm_err("Cannot map hyp stack\n");
1542
			goto out_err;
1543 1544 1545 1546
		}
	}

	for_each_possible_cpu(cpu) {
1547
		kvm_cpu_context_t *cpu_ctxt;
1548

1549
		cpu_ctxt = per_cpu_ptr(&kvm_host_cpu_state, cpu);
1550
		err = create_hyp_mappings(cpu_ctxt, cpu_ctxt + 1, PAGE_HYP);
1551 1552

		if (err) {
1553
			kvm_err("Cannot map host CPU state: %d\n", err);
1554
			goto out_err;
1555 1556 1557
		}
	}

1558 1559 1560 1561
	err = hyp_map_aux_data();
	if (err)
		kvm_err("Cannot map host auxilary data: %d\n", err);

1562
	return 0;
1563

1564
out_err:
1565
	teardown_hyp_mode();
1566 1567 1568 1569
	kvm_err("error initializing Hyp mode: %d\n", err);
	return err;
}

1570 1571 1572 1573 1574
static void check_kvm_target_cpu(void *ret)
{
	*(int *)ret = kvm_target_cpu();
}

1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587
struct kvm_vcpu *kvm_mpidr_to_vcpu(struct kvm *kvm, unsigned long mpidr)
{
	struct kvm_vcpu *vcpu;
	int i;

	mpidr &= MPIDR_HWID_BITMASK;
	kvm_for_each_vcpu(i, vcpu, kvm) {
		if (mpidr == kvm_vcpu_get_mpidr_aff(vcpu))
			return vcpu;
	}
	return NULL;
}

1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598
bool kvm_arch_has_irq_bypass(void)
{
	return true;
}

int kvm_arch_irq_bypass_add_producer(struct irq_bypass_consumer *cons,
				      struct irq_bypass_producer *prod)
{
	struct kvm_kernel_irqfd *irqfd =
		container_of(cons, struct kvm_kernel_irqfd, consumer);

1599 1600
	return kvm_vgic_v4_set_forwarding(irqfd->kvm, prod->irq,
					  &irqfd->irq_entry);
1601 1602 1603 1604 1605 1606 1607
}
void kvm_arch_irq_bypass_del_producer(struct irq_bypass_consumer *cons,
				      struct irq_bypass_producer *prod)
{
	struct kvm_kernel_irqfd *irqfd =
		container_of(cons, struct kvm_kernel_irqfd, consumer);

1608 1609
	kvm_vgic_v4_unset_forwarding(irqfd->kvm, prod->irq,
				     &irqfd->irq_entry);
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627
}

void kvm_arch_irq_bypass_stop(struct irq_bypass_consumer *cons)
{
	struct kvm_kernel_irqfd *irqfd =
		container_of(cons, struct kvm_kernel_irqfd, consumer);

	kvm_arm_halt_guest(irqfd->kvm);
}

void kvm_arch_irq_bypass_start(struct irq_bypass_consumer *cons)
{
	struct kvm_kernel_irqfd *irqfd =
		container_of(cons, struct kvm_kernel_irqfd, consumer);

	kvm_arm_resume_guest(irqfd->kvm);
}

1628 1629 1630
/**
 * Initialize Hyp-mode and memory mappings on all CPUs.
 */
1631 1632
int kvm_arch_init(void *opaque)
{
1633
	int err;
1634
	int ret, cpu;
1635
	bool in_hyp_mode;
1636 1637

	if (!is_hyp_mode_available()) {
1638
		kvm_info("HYP mode not available\n");
1639 1640 1641
		return -ENODEV;
	}

1642 1643 1644 1645 1646
	if (!kvm_arch_check_sve_has_vhe()) {
		kvm_pr_unimpl("SVE system without VHE unsupported.  Broken cpu?");
		return -ENODEV;
	}

1647 1648 1649 1650 1651 1652
	for_each_online_cpu(cpu) {
		smp_call_function_single(cpu, check_kvm_target_cpu, &ret, 1);
		if (ret < 0) {
			kvm_err("Error, CPU %d not supported!\n", cpu);
			return -ENODEV;
		}
1653 1654
	}

1655
	err = init_common_resources();
1656
	if (err)
1657
		return err;
1658

1659 1660 1661
	in_hyp_mode = is_kernel_in_hyp_mode();

	if (!in_hyp_mode) {
1662
		err = init_hyp_mode();
1663 1664 1665
		if (err)
			goto out_err;
	}
1666

1667 1668 1669
	err = init_subsystems();
	if (err)
		goto out_hyp;
1670

1671 1672 1673 1674 1675
	if (in_hyp_mode)
		kvm_info("VHE mode initialized successfully\n");
	else
		kvm_info("Hyp mode initialized successfully\n");

1676
	return 0;
1677 1678

out_hyp:
1679 1680
	if (!in_hyp_mode)
		teardown_hyp_mode();
1681 1682
out_err:
	return err;
1683 1684 1685 1686 1687
}

/* NOP: Compiling as a module not supported */
void kvm_arch_exit(void)
{
1688
	kvm_perf_teardown();
1689 1690 1691 1692 1693 1694 1695 1696 1697
}

static int arm_init(void)
{
	int rc = kvm_init(NULL, sizeof(struct kvm_vcpu), 0, THIS_MODULE);
	return rc;
}

module_init(arm_init);