[mirror_ubuntu-bionic-kernel.git] / arch / x86 / kvm / cpuid.c

/*
 * Kernel-based Virtual Machine driver for Linux
 * cpuid support routines
 *
 * derived from arch/x86/kvm/x86.c
 *
 * Copyright 2011 Red Hat, Inc. and/or its affiliates.
 * Copyright IBM Corporation, 2008
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 */

#include <linux/kvm_host.h>
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/uaccess.h>
#include <asm/user.h>
#include <asm/xsave.h>
#include "cpuid.h"
#include "lapic.h"
#include "mmu.h"
#include "trace.h"

void kvm_update_cpuid(struct kvm_vcpu *vcpu)
{
	struct kvm_cpuid_entry2 *best;
	struct kvm_lapic *apic = vcpu->arch.apic;

	best = kvm_find_cpuid_entry(vcpu, 1, 0);
	if (!best)
		return;

	/* Update OSXSAVE bit */
	if (cpu_has_xsave && best->function == 0x1) {
		best->ecx &= ~(bit(X86_FEATURE_OSXSAVE));
		if (kvm_read_cr4_bits(vcpu, X86_CR4_OSXSAVE))
			best->ecx |= bit(X86_FEATURE_OSXSAVE);
	}

	if (apic) {
		if (best->ecx & bit(X86_FEATURE_TSC_DEADLINE_TIMER))
			apic->lapic_timer.timer_mode_mask = 3 << 17;
		else
			apic->lapic_timer.timer_mode_mask = 1 << 17;
	}

	kvm_pmu_cpuid_update(vcpu);
}

static int is_efer_nx(void)
{
	unsigned long long efer = 0;

	rdmsrl_safe(MSR_EFER, &efer);
	return efer & EFER_NX;
}

static void cpuid_fix_nx_cap(struct kvm_vcpu *vcpu)
{
	int i;
	struct kvm_cpuid_entry2 *e, *entry;

	entry = NULL;
	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
		e = &vcpu->arch.cpuid_entries[i];
		if (e->function == 0x80000001) {
			entry = e;
			break;
		}
	}
	if (entry && (entry->edx & (1 << 20)) && !is_efer_nx()) {
		entry->edx &= ~(1 << 20);
		printk(KERN_INFO "kvm: guest NX capability removed\n");
	}
}

/* when an old userspace process fills a new kernel module */
int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
			     struct kvm_cpuid *cpuid,
			     struct kvm_cpuid_entry __user *entries)
{
	int r, i;
	struct kvm_cpuid_entry *cpuid_entries;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -ENOMEM;
	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry) * cpuid->nent);
	if (!cpuid_entries)
		goto out;
	r = -EFAULT;
	if (copy_from_user(cpuid_entries, entries,
			   cpuid->nent * sizeof(struct kvm_cpuid_entry)))
		goto out_free;
	for (i = 0; i < cpuid->nent; i++) {
		vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
		vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
		vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx;
		vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx;
		vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx;
		vcpu->arch.cpuid_entries[i].index = 0;
		vcpu->arch.cpuid_entries[i].flags = 0;
		vcpu->arch.cpuid_entries[i].padding[0] = 0;
		vcpu->arch.cpuid_entries[i].padding[1] = 0;
		vcpu->arch.cpuid_entries[i].padding[2] = 0;
	}
	vcpu->arch.cpuid_nent = cpuid->nent;
	cpuid_fix_nx_cap(vcpu);
	r = 0;
	kvm_apic_set_version(vcpu);
	kvm_x86_ops->cpuid_update(vcpu);
	kvm_update_cpuid(vcpu);

out_free:
	vfree(cpuid_entries);
out:
	return r;
}

int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
			      struct kvm_cpuid2 *cpuid,
			      struct kvm_cpuid_entry2 __user *entries)
{
	int r;

	r = -E2BIG;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		goto out;
	r = -EFAULT;
	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
			   cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
		goto out;
	vcpu->arch.cpuid_nent = cpuid->nent;
	kvm_apic_set_version(vcpu);
	kvm_x86_ops->cpuid_update(vcpu);
	kvm_update_cpuid(vcpu);
	return 0;

out:
	return r;
}

int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
			      struct kvm_cpuid2 *cpuid,
			      struct kvm_cpuid_entry2 __user *entries)
{
	int r;

	r = -E2BIG;
	if (cpuid->nent < vcpu->arch.cpuid_nent)
		goto out;
	r = -EFAULT;
	if (copy_to_user(entries, &vcpu->arch.cpuid_entries,
			 vcpu->arch.cpuid_nent * sizeof(struct kvm_cpuid_entry2)))
		goto out;
	return 0;

out:
	cpuid->nent = vcpu->arch.cpuid_nent;
	return r;
}

static void cpuid_mask(u32 *word, int wordnum)
{
	*word &= boot_cpu_data.x86_capability[wordnum];
}

static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
			   u32 index)
{
	entry->function = function;
	entry->index = index;
	cpuid_count(entry->function, entry->index,
		    &entry->eax, &entry->ebx, &entry->ecx, &entry->edx);
	entry->flags = 0;
}

static bool supported_xcr0_bit(unsigned bit)
{
	u64 mask = ((u64)1 << bit);

	return mask & (XSTATE_FP | XSTATE_SSE | XSTATE_YMM) & host_xcr0;
}

#define F(x) bit(X86_FEATURE_##x)

static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
			 u32 index, int *nent, int maxnent)
{
	int r;
	unsigned f_nx = is_efer_nx() ? F(NX) : 0;
#ifdef CONFIG_X86_64
	unsigned f_gbpages = (kvm_x86_ops->get_lpage_level() == PT_PDPE_LEVEL)
				? F(GBPAGES) : 0;
	unsigned f_lm = F(LM);
#else
	unsigned f_gbpages = 0;
	unsigned f_lm = 0;
#endif
	unsigned f_rdtscp = kvm_x86_ops->rdtscp_supported() ? F(RDTSCP) : 0;

	/* cpuid 1.edx */
	const u32 kvm_supported_word0_x86_features =
		F(FPU) | F(VME) | F(DE) | F(PSE) |
		F(TSC) | F(MSR) | F(PAE) | F(MCE) |
		F(CX8) | F(APIC) | 0 /* Reserved */ | F(SEP) |
		F(MTRR) | F(PGE) | F(MCA) | F(CMOV) |
		F(PAT) | F(PSE36) | 0 /* PSN */ | F(CLFLSH) |
		0 /* Reserved, DS, ACPI */ | F(MMX) |
		F(FXSR) | F(XMM) | F(XMM2) | F(SELFSNOOP) |
		0 /* HTT, TM, Reserved, PBE */;
	/* cpuid 0x80000001.edx */
	const u32 kvm_supported_word1_x86_features =
		F(FPU) | F(VME) | F(DE) | F(PSE) |
		F(TSC) | F(MSR) | F(PAE) | F(MCE) |
		F(CX8) | F(APIC) | 0 /* Reserved */ | F(SYSCALL) |
		F(MTRR) | F(PGE) | F(MCA) | F(CMOV) |
		F(PAT) | F(PSE36) | 0 /* Reserved */ |
		f_nx | 0 /* Reserved */ | F(MMXEXT) | F(MMX) |
		F(FXSR) | F(FXSR_OPT) | f_gbpages | f_rdtscp |
		0 /* Reserved */ | f_lm | F(3DNOWEXT) | F(3DNOW);
	/* cpuid 1.ecx */
	const u32 kvm_supported_word4_x86_features =
		F(XMM3) | F(PCLMULQDQ) | 0 /* DTES64, MONITOR */ |
		0 /* DS-CPL, VMX, SMX, EST */ |
		0 /* TM2 */ | F(SSSE3) | 0 /* CNXT-ID */ | 0 /* Reserved */ |
		F(FMA) | F(CX16) | 0 /* xTPR Update, PDCM */ |
		0 /* Reserved, DCA */ | F(XMM4_1) |
		F(XMM4_2) | F(X2APIC) | F(MOVBE) | F(POPCNT) |
		0 /* Reserved*/ | F(AES) | F(XSAVE) | 0 /* OSXSAVE */ | F(AVX) |
		F(F16C) | F(RDRAND);
	/* cpuid 0x80000001.ecx */
	const u32 kvm_supported_word6_x86_features =
		F(LAHF_LM) | F(CMP_LEGACY) | 0 /*SVM*/ | 0 /* ExtApicSpace */ |
		F(CR8_LEGACY) | F(ABM) | F(SSE4A) | F(MISALIGNSSE) |
		F(3DNOWPREFETCH) | F(OSVW) | 0 /* IBS */ | F(XOP) |
		0 /* SKINIT, WDT, LWP */ | F(FMA4) | F(TBM);

	/* cpuid 0xC0000001.edx */
	const u32 kvm_supported_word5_x86_features =
		F(XSTORE) | F(XSTORE_EN) | F(XCRYPT) | F(XCRYPT_EN) |
		F(ACE2) | F(ACE2_EN) | F(PHE) | F(PHE_EN) |
		F(PMM) | F(PMM_EN);

	/* cpuid 7.0.ebx */
	const u32 kvm_supported_word9_x86_features =
		F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) |
		F(BMI2) | F(ERMS) | F(RTM);

	/* all calls to cpuid_count() should be made on the same cpu */
	get_cpu();

	r = -E2BIG;

	if (*nent >= maxnent)
		goto out;

	do_cpuid_1_ent(entry, function, index);
	++*nent;

	switch (function) {
	case 0:
		entry->eax = min(entry->eax, (u32)0xd);
		break;
	case 1:
		entry->edx &= kvm_supported_word0_x86_features;
		cpuid_mask(&entry->edx, 0);
		entry->ecx &= kvm_supported_word4_x86_features;
		cpuid_mask(&entry->ecx, 4);
		/* we support x2apic emulation even if host does not support
		 * it since we emulate x2apic in software */
		entry->ecx |= F(X2APIC);
		break;
	/* function 2 entries are STATEFUL. That is, repeated cpuid commands
	 * may return different values. This forces us to get_cpu() before
	 * issuing the first command, and also to emulate this annoying behavior
	 * in kvm_emulate_cpuid() using KVM_CPUID_FLAG_STATE_READ_NEXT */
	case 2: {
		int t, times = entry->eax & 0xff;

		entry->flags |= KVM_CPUID_FLAG_STATEFUL_FUNC;
		entry->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT;
		for (t = 1; t < times; ++t) {
			if (*nent >= maxnent)
				goto out;

			do_cpuid_1_ent(&entry[t], function, 0);
			entry[t].flags |= KVM_CPUID_FLAG_STATEFUL_FUNC;
			++*nent;
		}
		break;
	}
	/* function 4 has additional index. */
	case 4: {
		int i, cache_type;

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until cache_type is zero */
		for (i = 1; ; ++i) {
			if (*nent >= maxnent)
				goto out;

			cache_type = entry[i - 1].eax & 0x1f;
			if (!cache_type)
				break;
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
	case 7: {
		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* Mask ebx against host capbability word 9 */
		if (index == 0) {
			entry->ebx &= kvm_supported_word9_x86_features;
			cpuid_mask(&entry->ebx, 9);
		} else
			entry->ebx = 0;
		entry->eax = 0;
		entry->ecx = 0;
		entry->edx = 0;
		break;
	}
	case 9:
		break;
	case 0xa: { /* Architectural Performance Monitoring */
		struct x86_pmu_capability cap;
		union cpuid10_eax eax;
		union cpuid10_edx edx;

		perf_get_x86_pmu_capability(&cap);

		/*
		 * Only support guest architectural pmu on a host
		 * with architectural pmu.
		 */
		if (!cap.version)
			memset(&cap, 0, sizeof(cap));

		eax.split.version_id = min(cap.version, 2);
		eax.split.num_counters = cap.num_counters_gp;
		eax.split.bit_width = cap.bit_width_gp;
		eax.split.mask_length = cap.events_mask_len;

		edx.split.num_counters_fixed = cap.num_counters_fixed;
		edx.split.bit_width_fixed = cap.bit_width_fixed;
		edx.split.reserved = 0;

		entry->eax = eax.full;
		entry->ebx = cap.events_mask;
		entry->ecx = 0;
		entry->edx = edx.full;
		break;
	}
	/* function 0xb has additional index. */
	case 0xb: {
		int i, level_type;

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		/* read more entries until level_type is zero */
		for (i = 1; ; ++i) {
			if (*nent >= maxnent)
				goto out;

			level_type = entry[i - 1].ecx & 0xff00;
			if (!level_type)
				break;
			do_cpuid_1_ent(&entry[i], function, i);
			entry[i].flags |=
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
		}
		break;
	}
	case 0xd: {
		int idx, i;

		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
		for (idx = 1, i = 1; idx < 64; ++idx) {
			if (*nent >= maxnent)
				goto out;

			do_cpuid_1_ent(&entry[i], function, idx);
			if (entry[i].eax == 0 || !supported_xcr0_bit(idx))
				continue;
			entry[i].flags |=
			       KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
			++*nent;
			++i;
		}
		break;
	}
	case KVM_CPUID_SIGNATURE: {
		char signature[12] = "KVMKVMKVM\0\0";
		u32 *sigptr = (u32 *)signature;
		entry->eax = KVM_CPUID_FEATURES;
		entry->ebx = sigptr[0];
		entry->ecx = sigptr[1];
		entry->edx = sigptr[2];
		break;
	}
	case KVM_CPUID_FEATURES:
		entry->eax = (1 << KVM_FEATURE_CLOCKSOURCE) |
			     (1 << KVM_FEATURE_NOP_IO_DELAY) |
			     (1 << KVM_FEATURE_CLOCKSOURCE2) |
			     (1 << KVM_FEATURE_ASYNC_PF) |
			     (1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT);

		if (sched_info_on())
			entry->eax |= (1 << KVM_FEATURE_STEAL_TIME);

		entry->ebx = 0;
		entry->ecx = 0;
		entry->edx = 0;
		break;
	case 0x80000000:
		entry->eax = min(entry->eax, 0x8000001a);
		break;
	case 0x80000001:
		entry->edx &= kvm_supported_word1_x86_features;
		cpuid_mask(&entry->edx, 1);
		entry->ecx &= kvm_supported_word6_x86_features;
		cpuid_mask(&entry->ecx, 6);
		break;
	case 0x80000008: {
		unsigned g_phys_as = (entry->eax >> 16) & 0xff;
		unsigned virt_as = max((entry->eax >> 8) & 0xff, 48U);
		unsigned phys_as = entry->eax & 0xff;

		if (!g_phys_as)
			g_phys_as = phys_as;
		entry->eax = g_phys_as | (virt_as << 8);
		entry->ebx = entry->edx = 0;
		break;
	}
	case 0x80000019:
		entry->ecx = entry->edx = 0;
		break;
	case 0x8000001a:
		break;
	case 0x8000001d:
		break;
	/*Add support for Centaur's CPUID instruction*/
	case 0xC0000000:
		/*Just support up to 0xC0000004 now*/
		entry->eax = min(entry->eax, 0xC0000004);
		break;
	case 0xC0000001:
		entry->edx &= kvm_supported_word5_x86_features;
		cpuid_mask(&entry->edx, 5);
		break;
	case 3: /* Processor serial number */
	case 5: /* MONITOR/MWAIT */
	case 6: /* Thermal management */
	case 0x80000007: /* Advanced power management */
	case 0xC0000002:
	case 0xC0000003:
	case 0xC0000004:
	default:
		entry->eax = entry->ebx = entry->ecx = entry->edx = 0;
		break;
	}

	kvm_x86_ops->set_supported_cpuid(function, entry);

	r = 0;

out:
	put_cpu();

	return r;
}

#undef F

struct kvm_cpuid_param {
	u32 func;
	u32 idx;
	bool has_leaf_count;
	bool (*qualifier)(struct kvm_cpuid_param *param);
};

static bool is_centaur_cpu(struct kvm_cpuid_param *param)
{
	return boot_cpu_data.x86_vendor == X86_VENDOR_CENTAUR;
}

int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
				      struct kvm_cpuid_entry2 __user *entries)
{
	struct kvm_cpuid_entry2 *cpuid_entries;
	int limit, nent = 0, r = -E2BIG, i;
	u32 func;
	static struct kvm_cpuid_param param[] = {
		{ .func = 0, .has_leaf_count = true },
		{ .func = 0x80000000, .has_leaf_count = true },
		{ .func = 0xC0000000, .qualifier = is_centaur_cpu, .has_leaf_count = true },
		{ .func = KVM_CPUID_SIGNATURE },
		{ .func = KVM_CPUID_FEATURES },
	};

	if (cpuid->nent < 1)
		goto out;
	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
		cpuid->nent = KVM_MAX_CPUID_ENTRIES;
	r = -ENOMEM;
	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
	if (!cpuid_entries)
		goto out;

	r = 0;
	for (i = 0; i < ARRAY_SIZE(param); i++) {
		struct kvm_cpuid_param *ent = &param[i];

		if (ent->qualifier && !ent->qualifier(ent))
			continue;

		r = do_cpuid_ent(&cpuid_entries[nent], ent->func, ent->idx,
				&nent, cpuid->nent);

		if (r)
			goto out_free;

		if (!ent->has_leaf_count)
			continue;

		limit = cpuid_entries[nent - 1].eax;
		for (func = ent->func + 1; func <= limit && nent < cpuid->nent && r == 0; ++func)
			r = do_cpuid_ent(&cpuid_entries[nent], func, ent->idx,
				     &nent, cpuid->nent);

		if (r)
			goto out_free;
	}

	r = -EFAULT;
	if (copy_to_user(entries, cpuid_entries,
			 nent * sizeof(struct kvm_cpuid_entry2)))
		goto out_free;
	cpuid->nent = nent;
	r = 0;

out_free:
	vfree(cpuid_entries);
out:
	return r;
}

static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
{
	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	int j, nent = vcpu->arch.cpuid_nent;

	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	/* when no next entry is found, the current entry[i] is reselected */
	for (j = i + 1; ; j = (j + 1) % nent) {
		struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
		if (ej->function == e->function) {
			ej->flags |= KVM_CPUID_FLAG_STATE_READ_NEXT;
			return j;
		}
	}
	return 0; /* silence gcc, even though control never reaches here */
}

/* find an entry with matching function, matching index (if needed), and that
 * should be read next (if it's stateful) */
static int is_matching_cpuid_entry(struct kvm_cpuid_entry2 *e,
	u32 function, u32 index)
{
	if (e->function != function)
		return 0;
	if ((e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) && e->index != index)
		return 0;
	if ((e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC) &&
	    !(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
		return 0;
	return 1;
}

struct kvm_cpuid_entry2 *kvm_find_cpuid_entry(struct kvm_vcpu *vcpu,
					      u32 function, u32 index)
{
	int i;
	struct kvm_cpuid_entry2 *best = NULL;

	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
		struct kvm_cpuid_entry2 *e;

		e = &vcpu->arch.cpuid_entries[i];
		if (is_matching_cpuid_entry(e, function, index)) {
			if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
				move_to_next_stateful_cpuid_entry(vcpu, i);
			best = e;
			break;
		}
	}
	return best;
}
EXPORT_SYMBOL_GPL(kvm_find_cpuid_entry);

int cpuid_maxphyaddr(struct kvm_vcpu *vcpu)
{
	struct kvm_cpuid_entry2 *best;

	best = kvm_find_cpuid_entry(vcpu, 0x80000000, 0);
	if (!best || best->eax < 0x80000008)
		goto not_found;
	best = kvm_find_cpuid_entry(vcpu, 0x80000008, 0);
	if (best)
		return best->eax & 0xff;
not_found:
	return 36;
}

/*
 * If no match is found, check whether we exceed the vCPU's limit
 * and return the content of the highest valid _standard_ leaf instead.
 * This is to satisfy the CPUID specification.
 */
static struct kvm_cpuid_entry2* check_cpuid_limit(struct kvm_vcpu *vcpu,
                                                  u32 function, u32 index)
{
	struct kvm_cpuid_entry2 *maxlevel;

	maxlevel = kvm_find_cpuid_entry(vcpu, function & 0x80000000, 0);
	if (!maxlevel || maxlevel->eax >= function)
		return NULL;
	if (function & 0x80000000) {
		maxlevel = kvm_find_cpuid_entry(vcpu, 0, 0);
		if (!maxlevel)
			return NULL;
	}
	return kvm_find_cpuid_entry(vcpu, maxlevel->eax, index);
}

void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
{
	u32 function, index;
	struct kvm_cpuid_entry2 *best;

	function = kvm_register_read(vcpu, VCPU_REGS_RAX);
	index = kvm_register_read(vcpu, VCPU_REGS_RCX);
	kvm_register_write(vcpu, VCPU_REGS_RAX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RBX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RCX, 0);
	kvm_register_write(vcpu, VCPU_REGS_RDX, 0);
	best = kvm_find_cpuid_entry(vcpu, function, index);

	if (!best)
		best = check_cpuid_limit(vcpu, function, index);

	if (best) {
		kvm_register_write(vcpu, VCPU_REGS_RAX, best->eax);
		kvm_register_write(vcpu, VCPU_REGS_RBX, best->ebx);
		kvm_register_write(vcpu, VCPU_REGS_RCX, best->ecx);
		kvm_register_write(vcpu, VCPU_REGS_RDX, best->edx);
	}
	kvm_x86_ops->skip_emulated_instruction(vcpu);
	trace_kvm_cpuid(function,
			kvm_register_read(vcpu, VCPU_REGS_RAX),
			kvm_register_read(vcpu, VCPU_REGS_RBX),
			kvm_register_read(vcpu, VCPU_REGS_RCX),
			kvm_register_read(vcpu, VCPU_REGS_RDX));
}
EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);
Commit	Line	Data
00b27a3e AK	1	/*
	2	* Kernel-based Virtual Machine driver for Linux
	3	* cpuid support routines
	4	*
	5	* derived from arch/x86/kvm/x86.c
	6	*
	7	* Copyright 2011 Red Hat, Inc. and/or its affiliates.
	8	* Copyright IBM Corporation, 2008
	9	*
	10	* This work is licensed under the terms of the GNU GPL, version 2. See
	11	* the COPYING file in the top-level directory.
	12	*
	13	*/
	14
	15	#include <linux/kvm_host.h>
	16	#include <linux/module.h>
bb5a798a JK	17	#include <linux/vmalloc.h>
bb5a798a JK	18	#include <linux/uaccess.h>
00b27a3e AK	19	#include <asm/user.h>
	20	#include <asm/xsave.h>
	21	#include "cpuid.h"
	22	#include "lapic.h"
	23	#include "mmu.h"
	24	#include "trace.h"
	25
	26	void kvm_update_cpuid(struct kvm_vcpu *vcpu)
	27	{
	28	struct kvm_cpuid_entry2 *best;
	29	struct kvm_lapic *apic = vcpu->arch.apic;
	30
	31	best = kvm_find_cpuid_entry(vcpu, 1, 0);
	32	if (!best)
	33	return;
	34
	35	/* Update OSXSAVE bit */
	36	if (cpu_has_xsave && best->function == 0x1) {
	37	best->ecx &= ~(bit(X86_FEATURE_OSXSAVE));
	38	if (kvm_read_cr4_bits(vcpu, X86_CR4_OSXSAVE))
	39	best->ecx \|= bit(X86_FEATURE_OSXSAVE);
	40	}
	41
	42	if (apic) {
	43	if (best->ecx & bit(X86_FEATURE_TSC_DEADLINE_TIMER))
	44	apic->lapic_timer.timer_mode_mask = 3 << 17;
	45	else
	46	apic->lapic_timer.timer_mode_mask = 1 << 17;
	47	}
f5132b01 GN	48
f5132b01 GN	49	kvm_pmu_cpuid_update(vcpu);
00b27a3e AK	50	}
	51
	52	static int is_efer_nx(void)
	53	{
	54	unsigned long long efer = 0;
	55
	56	rdmsrl_safe(MSR_EFER, &efer);
	57	return efer & EFER_NX;
	58	}
	59
	60	static void cpuid_fix_nx_cap(struct kvm_vcpu *vcpu)
	61	{
	62	int i;
	63	struct kvm_cpuid_entry2 e, entry;
	64
	65	entry = NULL;
	66	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
	67	e = &vcpu->arch.cpuid_entries[i];
	68	if (e->function == 0x80000001) {
	69	entry = e;
	70	break;
	71	}
	72	}
	73	if (entry && (entry->edx & (1 << 20)) && !is_efer_nx()) {
	74	entry->edx &= ~(1 << 20);
	75	printk(KERN_INFO "kvm: guest NX capability removed\n");
	76	}
	77	}
	78
	79	/* when an old userspace process fills a new kernel module */
	80	int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
	81	struct kvm_cpuid *cpuid,
	82	struct kvm_cpuid_entry __user *entries)
	83	{
	84	int r, i;
	85	struct kvm_cpuid_entry *cpuid_entries;
	86
	87	r = -E2BIG;
	88	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
	89	goto out;
	90	r = -ENOMEM;
	91	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry) * cpuid->nent);
	92	if (!cpuid_entries)
	93	goto out;
	94	r = -EFAULT;
	95	if (copy_from_user(cpuid_entries, entries,
	96	cpuid->nent * sizeof(struct kvm_cpuid_entry)))
	97	goto out_free;
	98	for (i = 0; i < cpuid->nent; i++) {
	99	vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
	100	vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
	101	vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx;
	102	vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx;
	103	vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx;
	104	vcpu->arch.cpuid_entries[i].index = 0;
	105	vcpu->arch.cpuid_entries[i].flags = 0;
	106	vcpu->arch.cpuid_entries[i].padding[0] = 0;
	107	vcpu->arch.cpuid_entries[i].padding[1] = 0;
	108	vcpu->arch.cpuid_entries[i].padding[2] = 0;
	109	}
	110	vcpu->arch.cpuid_nent = cpuid->nent;
	111	cpuid_fix_nx_cap(vcpu);
	112	r = 0;
	113	kvm_apic_set_version(vcpu);
114	kvm_x86_ops->cpuid_update(vcpu);
115	kvm_update_cpuid(vcpu);
116
117	out_free:
118	vfree(cpuid_entries);
119	out:
120	return r;
121	}
122
123	int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
124	struct kvm_cpuid2 *cpuid,
125	struct kvm_cpuid_entry2 __user *entries)
126	{
127	int r;
128
129	r = -E2BIG;
130	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
131	goto out;
132	r = -EFAULT;
133	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
134	cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
135	goto out;
136	vcpu->arch.cpuid_nent = cpuid->nent;
137	kvm_apic_set_version(vcpu);
138	kvm_x86_ops->cpuid_update(vcpu);
139	kvm_update_cpuid(vcpu);
140	return 0;
141
142	out:
143	return r;
144	}
145
146	int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
147	struct kvm_cpuid2 *cpuid,
148	struct kvm_cpuid_entry2 __user *entries)
149	{
150	int r;
151
152	r = -E2BIG;
153	if (cpuid->nent < vcpu->arch.cpuid_nent)
154	goto out;
155	r = -EFAULT;
156	if (copy_to_user(entries, &vcpu->arch.cpuid_entries,
157	vcpu->arch.cpuid_nent * sizeof(struct kvm_cpuid_entry2)))
158	goto out;
159	return 0;
160
161	out:
162	cpuid->nent = vcpu->arch.cpuid_nent;
163	return r;
164	}
165
166	static void cpuid_mask(u32 *word, int wordnum)
167	{
168	*word &= boot_cpu_data.x86_capability[wordnum];
169	}
170
171	static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
172	u32 index)
173	{
174	entry->function = function;
175	entry->index = index;
176	cpuid_count(entry->function, entry->index,
177	&entry->eax, &entry->ebx, &entry->ecx, &entry->edx);
178	entry->flags = 0;
179	}
180
181	static bool supported_xcr0_bit(unsigned bit)
182	{
183	u64 mask = ((u64)1 << bit);
184
185	return mask & (XSTATE_FP \| XSTATE_SSE \| XSTATE_YMM) & host_xcr0;
186	}
187
188	#define F(x) bit(X86_FEATURE_##x)
189
831bf664	190	static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
00b27a3e AK	191	u32 index, int *nent, int maxnent)
00b27a3e AK	192	{
831bf664	193	int r;
00b27a3e AK	194	unsigned f_nx = is_efer_nx() ? F(NX) : 0;
	195	#ifdef CONFIG_X86_64
	196	unsigned f_gbpages = (kvm_x86_ops->get_lpage_level() == PT_PDPE_LEVEL)
	197	? F(GBPAGES) : 0;
	198	unsigned f_lm = F(LM);
	199	#else
	200	unsigned f_gbpages = 0;
	201	unsigned f_lm = 0;
	202	#endif
	203	unsigned f_rdtscp = kvm_x86_ops->rdtscp_supported() ? F(RDTSCP) : 0;
	204
	205	/* cpuid 1.edx */
	206	const u32 kvm_supported_word0_x86_features =
	207	F(FPU) \| F(VME) \| F(DE) \| F(PSE) \|
	208	F(TSC) \| F(MSR) \| F(PAE) \| F(MCE) \|
	209	F(CX8) \| F(APIC) \| 0 /* Reserved */ \| F(SEP) \|
	210	F(MTRR) \| F(PGE) \| F(MCA) \| F(CMOV) \|
	211	F(PAT) \| F(PSE36) \| 0 /* PSN */ \| F(CLFLSH) \|
	212	0 /* Reserved, DS, ACPI */ \| F(MMX) \|
	213	F(FXSR) \| F(XMM) \| F(XMM2) \| F(SELFSNOOP) \|
	214	0 /* HTT, TM, Reserved, PBE */;
	215	/* cpuid 0x80000001.edx */
	216	const u32 kvm_supported_word1_x86_features =
	217	F(FPU) \| F(VME) \| F(DE) \| F(PSE) \|
	218	F(TSC) \| F(MSR) \| F(PAE) \| F(MCE) \|
	219	F(CX8) \| F(APIC) \| 0 /* Reserved */ \| F(SYSCALL) \|
	220	F(MTRR) \| F(PGE) \| F(MCA) \| F(CMOV) \|
	221	F(PAT) \| F(PSE36) \| 0 /* Reserved */ \|
	222	f_nx \| 0 /* Reserved */ \| F(MMXEXT) \| F(MMX) \|
	223	F(FXSR) \| F(FXSR_OPT) \| f_gbpages \| f_rdtscp \|
	224	0 /* Reserved */ \| f_lm \| F(3DNOWEXT) \| F(3DNOW);
	225	/* cpuid 1.ecx */
	226	const u32 kvm_supported_word4_x86_features =
	227	F(XMM3) \| F(PCLMULQDQ) \| 0 /* DTES64, MONITOR */ \|
	228	0 /* DS-CPL, VMX, SMX, EST */ \|
	229	0 /* TM2 / \| F(SSSE3) \| 0 / CNXT-ID / \| 0 / Reserved */ \|
fb215366	230	F(FMA) \| F(CX16) \| 0 /* xTPR Update, PDCM */ \|
00b27a3e AK	231	0 /* Reserved, DCA */ \| F(XMM4_1) \|
	232	F(XMM4_2) \| F(X2APIC) \| F(MOVBE) \| F(POPCNT) \|
	233	0 /* Reserved/ \| F(AES) \| F(XSAVE) \| 0 / OSXSAVE */ \| F(AVX) \|
	234	F(F16C) \| F(RDRAND);
	235	/* cpuid 0x80000001.ecx */
	236	const u32 kvm_supported_word6_x86_features =
	237	F(LAHF_LM) \| F(CMP_LEGACY) \| 0 /SVM/ \| 0 /* ExtApicSpace */ \|
	238	F(CR8_LEGACY) \| F(ABM) \| F(SSE4A) \| F(MISALIGNSSE) \|
2b036c6b	239	F(3DNOWPREFETCH) \| F(OSVW) \| 0 /* IBS */ \| F(XOP) \|
00b27a3e AK	240	0 /* SKINIT, WDT, LWP */ \| F(FMA4) \| F(TBM);
	241
	242	/* cpuid 0xC0000001.edx */
	243	const u32 kvm_supported_word5_x86_features =
	244	F(XSTORE) \| F(XSTORE_EN) \| F(XCRYPT) \| F(XCRYPT_EN) \|
	245	F(ACE2) \| F(ACE2_EN) \| F(PHE) \| F(PHE_EN) \|
	246	F(PMM) \| F(PMM_EN);
	247
	248	/* cpuid 7.0.ebx */
	249	const u32 kvm_supported_word9_x86_features =
83c52915 LJ	250	F(FSGSBASE) \| F(BMI1) \| F(HLE) \| F(AVX2) \| F(SMEP) \|
83c52915 LJ	251	F(BMI2) \| F(ERMS) \| F(RTM);
00b27a3e AK	252
	253	/* all calls to cpuid_count() should be made on the same cpu */
	254	get_cpu();
831bf664 SL	255
	256	r = -E2BIG;
	257
	258	if (*nent >= maxnent)
	259	goto out;
	260
00b27a3e AK	261	do_cpuid_1_ent(entry, function, index);
	262	++*nent;
	263
	264	switch (function) {
	265	case 0:
	266	entry->eax = min(entry->eax, (u32)0xd);
	267	break;
	268	case 1:
	269	entry->edx &= kvm_supported_word0_x86_features;
	270	cpuid_mask(&entry->edx, 0);
	271	entry->ecx &= kvm_supported_word4_x86_features;
	272	cpuid_mask(&entry->ecx, 4);
	273	/* we support x2apic emulation even if host does not support
	274	* it since we emulate x2apic in software */
	275	entry->ecx \|= F(X2APIC);
	276	break;
	277	/* function 2 entries are STATEFUL. That is, repeated cpuid commands
	278	* may return different values. This forces us to get_cpu() before
	279	* issuing the first command, and also to emulate this annoying behavior
	280	* in kvm_emulate_cpuid() using KVM_CPUID_FLAG_STATE_READ_NEXT */
	281	case 2: {
	282	int t, times = entry->eax & 0xff;
	283
	284	entry->flags \|= KVM_CPUID_FLAG_STATEFUL_FUNC;
	285	entry->flags \|= KVM_CPUID_FLAG_STATE_READ_NEXT;
831bf664 SL	286	for (t = 1; t < times; ++t) {
	287	if (*nent >= maxnent)
	288	goto out;
	289
00b27a3e AK	290	do_cpuid_1_ent(&entry[t], function, 0);
	291	entry[t].flags \|= KVM_CPUID_FLAG_STATEFUL_FUNC;
	292	++*nent;
	293	}
	294	break;
	295	}
	296	/* function 4 has additional index. */
	297	case 4: {
	298	int i, cache_type;
	299
	300	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	301	/* read more entries until cache_type is zero */
831bf664 SL	302	for (i = 1; ; ++i) {
	303	if (*nent >= maxnent)
	304	goto out;
	305
00b27a3e AK	306	cache_type = entry[i - 1].eax & 0x1f;
	307	if (!cache_type)
	308	break;
	309	do_cpuid_1_ent(&entry[i], function, i);
	310	entry[i].flags \|=
	311	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	312	++*nent;
	313	}
	314	break;
	315	}
	316	case 7: {
	317	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	318	/* Mask ebx against host capbability word 9 */
	319	if (index == 0) {
	320	entry->ebx &= kvm_supported_word9_x86_features;
	321	cpuid_mask(&entry->ebx, 9);
	322	} else
	323	entry->ebx = 0;
	324	entry->eax = 0;
	325	entry->ecx = 0;
	326	entry->edx = 0;
	327	break;
	328	}
	329	case 9:
	330	break;
a6c06ed1 GN	331	case 0xa: { /* Architectural Performance Monitoring */
	332	struct x86_pmu_capability cap;
	333	union cpuid10_eax eax;
	334	union cpuid10_edx edx;
	335
	336	perf_get_x86_pmu_capability(&cap);
	337
	338	/*
	339	* Only support guest architectural pmu on a host
	340	* with architectural pmu.
	341	*/
	342	if (!cap.version)
	343	memset(&cap, 0, sizeof(cap));
	344
	345	eax.split.version_id = min(cap.version, 2);
	346	eax.split.num_counters = cap.num_counters_gp;
	347	eax.split.bit_width = cap.bit_width_gp;
	348	eax.split.mask_length = cap.events_mask_len;
	349
	350	edx.split.num_counters_fixed = cap.num_counters_fixed;
	351	edx.split.bit_width_fixed = cap.bit_width_fixed;
	352	edx.split.reserved = 0;
	353
	354	entry->eax = eax.full;
	355	entry->ebx = cap.events_mask;
	356	entry->ecx = 0;
	357	entry->edx = edx.full;
	358	break;
	359	}
00b27a3e AK	360	/* function 0xb has additional index. */
	361	case 0xb: {
	362	int i, level_type;
	363
	364	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	365	/* read more entries until level_type is zero */
831bf664 SL	366	for (i = 1; ; ++i) {
	367	if (*nent >= maxnent)
	368	goto out;
	369
00b27a3e AK	370	level_type = entry[i - 1].ecx & 0xff00;
	371	if (!level_type)
	372	break;
	373	do_cpuid_1_ent(&entry[i], function, i);
	374	entry[i].flags \|=
	375	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	376	++*nent;
	377	}
	378	break;
	379	}
	380	case 0xd: {
	381	int idx, i;
	382
	383	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
831bf664 SL	384	for (idx = 1, i = 1; idx < 64; ++idx) {
	385	if (*nent >= maxnent)
	386	goto out;
	387
00b27a3e AK	388	do_cpuid_1_ent(&entry[i], function, idx);
	389	if (entry[i].eax == 0 \|\| !supported_xcr0_bit(idx))
	390	continue;
	391	entry[i].flags \|=
	392	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	393	++*nent;
	394	++i;
	395	}
	396	break;
	397	}
	398	case KVM_CPUID_SIGNATURE: {
	399	char signature[12] = "KVMKVMKVM\0\0";
	400	u32 sigptr = (u32 )signature;
57c22e5f	401	entry->eax = KVM_CPUID_FEATURES;
00b27a3e AK	402	entry->ebx = sigptr[0];
	403	entry->ecx = sigptr[1];
	404	entry->edx = sigptr[2];
	405	break;
	406	}
	407	case KVM_CPUID_FEATURES:
	408	entry->eax = (1 << KVM_FEATURE_CLOCKSOURCE) \|
	409	(1 << KVM_FEATURE_NOP_IO_DELAY) \|
	410	(1 << KVM_FEATURE_CLOCKSOURCE2) \|
	411	(1 << KVM_FEATURE_ASYNC_PF) \|
	412	(1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT);
	413
	414	if (sched_info_on())
	415	entry->eax \|= (1 << KVM_FEATURE_STEAL_TIME);
	416
	417	entry->ebx = 0;
	418	entry->ecx = 0;
	419	entry->edx = 0;
	420	break;
	421	case 0x80000000:
	422	entry->eax = min(entry->eax, 0x8000001a);
	423	break;
	424	case 0x80000001:
	425	entry->edx &= kvm_supported_word1_x86_features;
	426	cpuid_mask(&entry->edx, 1);
	427	entry->ecx &= kvm_supported_word6_x86_features;
	428	cpuid_mask(&entry->ecx, 6);
	429	break;
	430	case 0x80000008: {
	431	unsigned g_phys_as = (entry->eax >> 16) & 0xff;
	432	unsigned virt_as = max((entry->eax >> 8) & 0xff, 48U);
	433	unsigned phys_as = entry->eax & 0xff;
	434
	435	if (!g_phys_as)
	436	g_phys_as = phys_as;
	437	entry->eax = g_phys_as \| (virt_as << 8);
	438	entry->ebx = entry->edx = 0;
	439	break;
	440	}
	441	case 0x80000019:
	442	entry->ecx = entry->edx = 0;
	443	break;
	444	case 0x8000001a:
	445	break;
	446	case 0x8000001d:
	447	break;
	448	/Add support for Centaur's CPUID instruction/
	449	case 0xC0000000:
	450	/Just support up to 0xC0000004 now/
	451	entry->eax = min(entry->eax, 0xC0000004);
	452	break;
	453	case 0xC0000001:
	454	entry->edx &= kvm_supported_word5_x86_features;
	455	cpuid_mask(&entry->edx, 5);
	456	break;
	457	case 3: /* Processor serial number */
	458	case 5: /* MONITOR/MWAIT */
	459	case 6: /* Thermal management */
00b27a3e AK	460	case 0x80000007: /* Advanced power management */
	461	case 0xC0000002:
	462	case 0xC0000003:
	463	case 0xC0000004:
	464	default:
	465	entry->eax = entry->ebx = entry->ecx = entry->edx = 0;
	466	break;
	467	}
	468
	469	kvm_x86_ops->set_supported_cpuid(function, entry);
	470
831bf664 SL	471	r = 0;
	472
	473	out:
00b27a3e	474	put_cpu();
831bf664 SL	475
831bf664 SL	476	return r;
00b27a3e AK	477	}
	478
	479	#undef F
	480
831bf664 SL	481	struct kvm_cpuid_param {
	482	u32 func;
	483	u32 idx;
	484	bool has_leaf_count;
	485	bool (qualifier)(struct kvm_cpuid_param param);
	486	};
	487
	488	static bool is_centaur_cpu(struct kvm_cpuid_param *param)
	489	{
	490	return boot_cpu_data.x86_vendor == X86_VENDOR_CENTAUR;
	491	}
	492
00b27a3e AK	493	int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
	494	struct kvm_cpuid_entry2 __user *entries)
	495	{
	496	struct kvm_cpuid_entry2 *cpuid_entries;
831bf664	497	int limit, nent = 0, r = -E2BIG, i;
00b27a3e	498	u32 func;
831bf664 SL	499	static struct kvm_cpuid_param param[] = {
	500	{ .func = 0, .has_leaf_count = true },
	501	{ .func = 0x80000000, .has_leaf_count = true },
	502	{ .func = 0xC0000000, .qualifier = is_centaur_cpu, .has_leaf_count = true },
	503	{ .func = KVM_CPUID_SIGNATURE },
	504	{ .func = KVM_CPUID_FEATURES },
	505	};
00b27a3e AK	506
	507	if (cpuid->nent < 1)
	508	goto out;
	509	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
	510	cpuid->nent = KVM_MAX_CPUID_ENTRIES;
	511	r = -ENOMEM;
	512	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
	513	if (!cpuid_entries)
	514	goto out;
	515
831bf664 SL	516	r = 0;
	517	for (i = 0; i < ARRAY_SIZE(param); i++) {
	518	struct kvm_cpuid_param *ent = &param[i];
00b27a3e	519
831bf664 SL	520	if (ent->qualifier && !ent->qualifier(ent))
831bf664 SL	521	continue;
00b27a3e	522
831bf664	523	r = do_cpuid_ent(&cpuid_entries[nent], ent->func, ent->idx,
00b27a3e AK	524	&nent, cpuid->nent);
00b27a3e AK	525
831bf664	526	if (r)
00b27a3e AK	527	goto out_free;
00b27a3e AK	528
831bf664 SL	529	if (!ent->has_leaf_count)
	530	continue;
	531
00b27a3e	532	limit = cpuid_entries[nent - 1].eax;
831bf664 SL	533	for (func = ent->func + 1; func <= limit && nent < cpuid->nent && r == 0; ++func)
	534	r = do_cpuid_ent(&cpuid_entries[nent], func, ent->idx,
	535	&nent, cpuid->nent);
00b27a3e	536
831bf664	537	if (r)
00b27a3e AK	538	goto out_free;
	539	}
	540
00b27a3e AK	541	r = -EFAULT;
	542	if (copy_to_user(entries, cpuid_entries,
	543	nent * sizeof(struct kvm_cpuid_entry2)))
	544	goto out_free;
	545	cpuid->nent = nent;
	546	r = 0;
	547
	548	out_free:
	549	vfree(cpuid_entries);
	550	out:
	551	return r;
	552	}
	553
	554	static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
	555	{
	556	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	557	int j, nent = vcpu->arch.cpuid_nent;
	558
	559	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	560	/* when no next entry is found, the current entry[i] is reselected */
	561	for (j = i + 1; ; j = (j + 1) % nent) {
	562	struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
	563	if (ej->function == e->function) {
	564	ej->flags \|= KVM_CPUID_FLAG_STATE_READ_NEXT;
	565	return j;
	566	}
	567	}
	568	return 0; /* silence gcc, even though control never reaches here */
	569	}
	570
	571	/* find an entry with matching function, matching index (if needed), and that
	572	* should be read next (if it's stateful) */
	573	static int is_matching_cpuid_entry(struct kvm_cpuid_entry2 *e,
	574	u32 function, u32 index)
	575	{
	576	if (e->function != function)
	577	return 0;
	578	if ((e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) && e->index != index)
	579	return 0;
	580	if ((e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC) &&
	581	!(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
	582	return 0;
	583	return 1;
	584	}
	585
	586	struct kvm_cpuid_entry2 kvm_find_cpuid_entry(struct kvm_vcpu vcpu,
	587	u32 function, u32 index)
	588	{
	589	int i;
	590	struct kvm_cpuid_entry2 *best = NULL;
	591
	592	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
	593	struct kvm_cpuid_entry2 *e;
	594
	595	e = &vcpu->arch.cpuid_entries[i];
	596	if (is_matching_cpuid_entry(e, function, index)) {
	597	if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
	598	move_to_next_stateful_cpuid_entry(vcpu, i);
	599	best = e;
	600	break;
	601	}
	602	}
	603	return best;
	604	}
605	EXPORT_SYMBOL_GPL(kvm_find_cpuid_entry);
606
607	int cpuid_maxphyaddr(struct kvm_vcpu *vcpu)
608	{
609	struct kvm_cpuid_entry2 *best;
610
611	best = kvm_find_cpuid_entry(vcpu, 0x80000000, 0);
612	if (!best \|\| best->eax < 0x80000008)
613	goto not_found;
614	best = kvm_find_cpuid_entry(vcpu, 0x80000008, 0);
615	if (best)
616	return best->eax & 0xff;
617	not_found:
618	return 36;
619	}
620
621	/*
622	* If no match is found, check whether we exceed the vCPU's limit
623	* and return the content of the highest valid _standard_ leaf instead.
624	* This is to satisfy the CPUID specification.
625	*/
626	static struct kvm_cpuid_entry2* check_cpuid_limit(struct kvm_vcpu *vcpu,
627	u32 function, u32 index)
628	{
629	struct kvm_cpuid_entry2 *maxlevel;
630
631	maxlevel = kvm_find_cpuid_entry(vcpu, function & 0x80000000, 0);
632	if (!maxlevel \|\| maxlevel->eax >= function)
633	return NULL;
634	if (function & 0x80000000) {
635	maxlevel = kvm_find_cpuid_entry(vcpu, 0, 0);
636	if (!maxlevel)
637	return NULL;
638	}
639	return kvm_find_cpuid_entry(vcpu, maxlevel->eax, index);
640	}
641
642	void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
643	{
644	u32 function, index;
645	struct kvm_cpuid_entry2 *best;
646
647	function = kvm_register_read(vcpu, VCPU_REGS_RAX);
648	index = kvm_register_read(vcpu, VCPU_REGS_RCX);
649	kvm_register_write(vcpu, VCPU_REGS_RAX, 0);
650	kvm_register_write(vcpu, VCPU_REGS_RBX, 0);
651	kvm_register_write(vcpu, VCPU_REGS_RCX, 0);
652	kvm_register_write(vcpu, VCPU_REGS_RDX, 0);
653	best = kvm_find_cpuid_entry(vcpu, function, index);
654
655	if (!best)
656	best = check_cpuid_limit(vcpu, function, index);
657
658	if (best) {
659	kvm_register_write(vcpu, VCPU_REGS_RAX, best->eax);
660	kvm_register_write(vcpu, VCPU_REGS_RBX, best->ebx);
661	kvm_register_write(vcpu, VCPU_REGS_RCX, best->ecx);
662	kvm_register_write(vcpu, VCPU_REGS_RDX, best->edx);
663	}
664	kvm_x86_ops->skip_emulated_instruction(vcpu);
665	trace_kvm_cpuid(function,
666	kvm_register_read(vcpu, VCPU_REGS_RAX),
667	kvm_register_read(vcpu, VCPU_REGS_RBX),
668	kvm_register_read(vcpu, VCPU_REGS_RCX),
669	kvm_register_read(vcpu, VCPU_REGS_RDX));
670	}
671	EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);