[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(AVX2) | F(SMEP) | F(BMI2) | F(ERMS);

	/* 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 = 0;
		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 =
fb215366	250	F(FSGSBASE) \| F(BMI1) \| F(AVX2) \| F(SMEP) \| F(BMI2) \| F(ERMS);
00b27a3e AK	251
	252	/* all calls to cpuid_count() should be made on the same cpu */
	253	get_cpu();
831bf664 SL	254
	255	r = -E2BIG;
	256
	257	if (*nent >= maxnent)
	258	goto out;
	259
00b27a3e AK	260	do_cpuid_1_ent(entry, function, index);
	261	++*nent;
	262
	263	switch (function) {
	264	case 0:
	265	entry->eax = min(entry->eax, (u32)0xd);
	266	break;
	267	case 1:
	268	entry->edx &= kvm_supported_word0_x86_features;
	269	cpuid_mask(&entry->edx, 0);
	270	entry->ecx &= kvm_supported_word4_x86_features;
	271	cpuid_mask(&entry->ecx, 4);
	272	/* we support x2apic emulation even if host does not support
	273	* it since we emulate x2apic in software */
	274	entry->ecx \|= F(X2APIC);
	275	break;
	276	/* function 2 entries are STATEFUL. That is, repeated cpuid commands
	277	* may return different values. This forces us to get_cpu() before
	278	* issuing the first command, and also to emulate this annoying behavior
	279	* in kvm_emulate_cpuid() using KVM_CPUID_FLAG_STATE_READ_NEXT */
	280	case 2: {
	281	int t, times = entry->eax & 0xff;
	282
	283	entry->flags \|= KVM_CPUID_FLAG_STATEFUL_FUNC;
	284	entry->flags \|= KVM_CPUID_FLAG_STATE_READ_NEXT;
831bf664 SL	285	for (t = 1; t < times; ++t) {
	286	if (*nent >= maxnent)
	287	goto out;
	288
00b27a3e AK	289	do_cpuid_1_ent(&entry[t], function, 0);
	290	entry[t].flags \|= KVM_CPUID_FLAG_STATEFUL_FUNC;
	291	++*nent;
	292	}
	293	break;
	294	}
	295	/* function 4 has additional index. */
	296	case 4: {
	297	int i, cache_type;
	298
	299	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	300	/* read more entries until cache_type is zero */
831bf664 SL	301	for (i = 1; ; ++i) {
	302	if (*nent >= maxnent)
	303	goto out;
	304
00b27a3e AK	305	cache_type = entry[i - 1].eax & 0x1f;
	306	if (!cache_type)
	307	break;
	308	do_cpuid_1_ent(&entry[i], function, i);
	309	entry[i].flags \|=
	310	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	311	++*nent;
	312	}
	313	break;
	314	}
	315	case 7: {
	316	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	317	/* Mask ebx against host capbability word 9 */
	318	if (index == 0) {
	319	entry->ebx &= kvm_supported_word9_x86_features;
	320	cpuid_mask(&entry->ebx, 9);
	321	} else
	322	entry->ebx = 0;
	323	entry->eax = 0;
	324	entry->ecx = 0;
	325	entry->edx = 0;
	326	break;
	327	}
	328	case 9:
	329	break;
a6c06ed1 GN	330	case 0xa: { /* Architectural Performance Monitoring */
	331	struct x86_pmu_capability cap;
	332	union cpuid10_eax eax;
	333	union cpuid10_edx edx;
	334
	335	perf_get_x86_pmu_capability(&cap);
	336
	337	/*
	338	* Only support guest architectural pmu on a host
	339	* with architectural pmu.
	340	*/
	341	if (!cap.version)
	342	memset(&cap, 0, sizeof(cap));
	343
	344	eax.split.version_id = min(cap.version, 2);
	345	eax.split.num_counters = cap.num_counters_gp;
	346	eax.split.bit_width = cap.bit_width_gp;
	347	eax.split.mask_length = cap.events_mask_len;
	348
	349	edx.split.num_counters_fixed = cap.num_counters_fixed;
	350	edx.split.bit_width_fixed = cap.bit_width_fixed;
	351	edx.split.reserved = 0;
	352
	353	entry->eax = eax.full;
	354	entry->ebx = cap.events_mask;
	355	entry->ecx = 0;
	356	entry->edx = edx.full;
	357	break;
	358	}
00b27a3e AK	359	/* function 0xb has additional index. */
	360	case 0xb: {
	361	int i, level_type;
	362
	363	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	364	/* read more entries until level_type is zero */
831bf664 SL	365	for (i = 1; ; ++i) {
	366	if (*nent >= maxnent)
	367	goto out;
	368
00b27a3e AK	369	level_type = entry[i - 1].ecx & 0xff00;
	370	if (!level_type)
	371	break;
	372	do_cpuid_1_ent(&entry[i], function, i);
	373	entry[i].flags \|=
	374	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	375	++*nent;
	376	}
	377	break;
	378	}
	379	case 0xd: {
	380	int idx, i;
	381
	382	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
831bf664 SL	383	for (idx = 1, i = 1; idx < 64; ++idx) {
	384	if (*nent >= maxnent)
	385	goto out;
	386
00b27a3e AK	387	do_cpuid_1_ent(&entry[i], function, idx);
	388	if (entry[i].eax == 0 \|\| !supported_xcr0_bit(idx))
	389	continue;
	390	entry[i].flags \|=
	391	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	392	++*nent;
	393	++i;
	394	}
	395	break;
	396	}
	397	case KVM_CPUID_SIGNATURE: {
	398	char signature[12] = "KVMKVMKVM\0\0";
	399	u32 sigptr = (u32 )signature;
	400	entry->eax = 0;
	401	entry->ebx = sigptr[0];
	402	entry->ecx = sigptr[1];
	403	entry->edx = sigptr[2];
	404	break;
	405	}
	406	case KVM_CPUID_FEATURES:
	407	entry->eax = (1 << KVM_FEATURE_CLOCKSOURCE) \|
	408	(1 << KVM_FEATURE_NOP_IO_DELAY) \|
	409	(1 << KVM_FEATURE_CLOCKSOURCE2) \|
	410	(1 << KVM_FEATURE_ASYNC_PF) \|
	411	(1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT);
	412
	413	if (sched_info_on())
	414	entry->eax \|= (1 << KVM_FEATURE_STEAL_TIME);
	415
	416	entry->ebx = 0;
	417	entry->ecx = 0;
	418	entry->edx = 0;
	419	break;
	420	case 0x80000000:
	421	entry->eax = min(entry->eax, 0x8000001a);
	422	break;
	423	case 0x80000001:
	424	entry->edx &= kvm_supported_word1_x86_features;
	425	cpuid_mask(&entry->edx, 1);
	426	entry->ecx &= kvm_supported_word6_x86_features;
	427	cpuid_mask(&entry->ecx, 6);
	428	break;
	429	case 0x80000008: {
	430	unsigned g_phys_as = (entry->eax >> 16) & 0xff;
	431	unsigned virt_as = max((entry->eax >> 8) & 0xff, 48U);
	432	unsigned phys_as = entry->eax & 0xff;
	433
	434	if (!g_phys_as)
	435	g_phys_as = phys_as;
	436	entry->eax = g_phys_as \| (virt_as << 8);
	437	entry->ebx = entry->edx = 0;
	438	break;
	439	}
	440	case 0x80000019:
	441	entry->ecx = entry->edx = 0;
	442	break;
	443	case 0x8000001a:
	444	break;
	445	case 0x8000001d:
	446	break;
	447	/Add support for Centaur's CPUID instruction/
	448	case 0xC0000000:
	449	/Just support up to 0xC0000004 now/
	450	entry->eax = min(entry->eax, 0xC0000004);
451	break;
452	case 0xC0000001:
453	entry->edx &= kvm_supported_word5_x86_features;
454	cpuid_mask(&entry->edx, 5);
455	break;
456	case 3: /* Processor serial number */
457	case 5: /* MONITOR/MWAIT */
458	case 6: /* Thermal management */
00b27a3e AK	459	case 0x80000007: /* Advanced power management */
	460	case 0xC0000002:
	461	case 0xC0000003:
	462	case 0xC0000004:
	463	default:
	464	entry->eax = entry->ebx = entry->ecx = entry->edx = 0;
	465	break;
	466	}
	467
	468	kvm_x86_ops->set_supported_cpuid(function, entry);
	469
831bf664 SL	470	r = 0;
	471
	472	out:
00b27a3e	473	put_cpu();
831bf664 SL	474
831bf664 SL	475	return r;
00b27a3e AK	476	}
	477
	478	#undef F
	479
831bf664 SL	480	struct kvm_cpuid_param {
	481	u32 func;
	482	u32 idx;
	483	bool has_leaf_count;
	484	bool (qualifier)(struct kvm_cpuid_param param);
	485	};
	486
	487	static bool is_centaur_cpu(struct kvm_cpuid_param *param)
	488	{
	489	return boot_cpu_data.x86_vendor == X86_VENDOR_CENTAUR;
	490	}
	491
00b27a3e AK	492	int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
	493	struct kvm_cpuid_entry2 __user *entries)
	494	{
	495	struct kvm_cpuid_entry2 *cpuid_entries;
831bf664	496	int limit, nent = 0, r = -E2BIG, i;
00b27a3e	497	u32 func;
831bf664 SL	498	static struct kvm_cpuid_param param[] = {
	499	{ .func = 0, .has_leaf_count = true },
	500	{ .func = 0x80000000, .has_leaf_count = true },
	501	{ .func = 0xC0000000, .qualifier = is_centaur_cpu, .has_leaf_count = true },
	502	{ .func = KVM_CPUID_SIGNATURE },
	503	{ .func = KVM_CPUID_FEATURES },
	504	};
00b27a3e AK	505
	506	if (cpuid->nent < 1)
	507	goto out;
	508	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
	509	cpuid->nent = KVM_MAX_CPUID_ENTRIES;
	510	r = -ENOMEM;
	511	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
	512	if (!cpuid_entries)
	513	goto out;
	514
831bf664 SL	515	r = 0;
	516	for (i = 0; i < ARRAY_SIZE(param); i++) {
	517	struct kvm_cpuid_param *ent = &param[i];
00b27a3e	518
831bf664 SL	519	if (ent->qualifier && !ent->qualifier(ent))
831bf664 SL	520	continue;
00b27a3e	521
831bf664	522	r = do_cpuid_ent(&cpuid_entries[nent], ent->func, ent->idx,
00b27a3e AK	523	&nent, cpuid->nent);
00b27a3e AK	524
831bf664	525	if (r)
00b27a3e AK	526	goto out_free;
00b27a3e AK	527
831bf664 SL	528	if (!ent->has_leaf_count)
	529	continue;
	530
00b27a3e	531	limit = cpuid_entries[nent - 1].eax;
831bf664 SL	532	for (func = ent->func + 1; func <= limit && nent < cpuid->nent && r == 0; ++func)
	533	r = do_cpuid_ent(&cpuid_entries[nent], func, ent->idx,
	534	&nent, cpuid->nent);
00b27a3e	535
831bf664	536	if (r)
00b27a3e AK	537	goto out_free;
	538	}
	539
00b27a3e AK	540	r = -EFAULT;
	541	if (copy_to_user(entries, cpuid_entries,
	542	nent * sizeof(struct kvm_cpuid_entry2)))
	543	goto out_free;
	544	cpuid->nent = nent;
	545	r = 0;
	546
	547	out_free:
	548	vfree(cpuid_entries);
	549	out:
	550	return r;
	551	}
	552
	553	static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
	554	{
	555	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	556	int j, nent = vcpu->arch.cpuid_nent;
	557
	558	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	559	/* when no next entry is found, the current entry[i] is reselected */
	560	for (j = i + 1; ; j = (j + 1) % nent) {
	561	struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
	562	if (ej->function == e->function) {
	563	ej->flags \|= KVM_CPUID_FLAG_STATE_READ_NEXT;
	564	return j;
	565	}
	566	}
	567	return 0; /* silence gcc, even though control never reaches here */
	568	}
	569
	570	/* find an entry with matching function, matching index (if needed), and that
	571	* should be read next (if it's stateful) */
	572	static int is_matching_cpuid_entry(struct kvm_cpuid_entry2 *e,
	573	u32 function, u32 index)
	574	{
	575	if (e->function != function)
	576	return 0;
	577	if ((e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) && e->index != index)
	578	return 0;
	579	if ((e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC) &&
	580	!(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
	581	return 0;
	582	return 1;
	583	}
	584
	585	struct kvm_cpuid_entry2 kvm_find_cpuid_entry(struct kvm_vcpu vcpu,
	586	u32 function, u32 index)
	587	{
	588	int i;
	589	struct kvm_cpuid_entry2 *best = NULL;
	590
	591	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
	592	struct kvm_cpuid_entry2 *e;
	593
	594	e = &vcpu->arch.cpuid_entries[i];
	595	if (is_matching_cpuid_entry(e, function, index)) {
	596	if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
	597	move_to_next_stateful_cpuid_entry(vcpu, i);
	598	best = e;
	599	break;
	600	}
	601	}
	602	return best;
	603	}
604	EXPORT_SYMBOL_GPL(kvm_find_cpuid_entry);
605
606	int cpuid_maxphyaddr(struct kvm_vcpu *vcpu)
607	{
608	struct kvm_cpuid_entry2 *best;
609
610	best = kvm_find_cpuid_entry(vcpu, 0x80000000, 0);
611	if (!best \|\| best->eax < 0x80000008)
612	goto not_found;
613	best = kvm_find_cpuid_entry(vcpu, 0x80000008, 0);
614	if (best)
615	return best->eax & 0xff;
616	not_found:
617	return 36;
618	}
619
620	/*
621	* If no match is found, check whether we exceed the vCPU's limit
622	* and return the content of the highest valid _standard_ leaf instead.
623	* This is to satisfy the CPUID specification.
624	*/
625	static struct kvm_cpuid_entry2* check_cpuid_limit(struct kvm_vcpu *vcpu,
626	u32 function, u32 index)
627	{
628	struct kvm_cpuid_entry2 *maxlevel;
629
630	maxlevel = kvm_find_cpuid_entry(vcpu, function & 0x80000000, 0);
631	if (!maxlevel \|\| maxlevel->eax >= function)
632	return NULL;
633	if (function & 0x80000000) {
634	maxlevel = kvm_find_cpuid_entry(vcpu, 0, 0);
635	if (!maxlevel)
636	return NULL;
637	}
638	return kvm_find_cpuid_entry(vcpu, maxlevel->eax, index);
639	}
640
641	void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
642	{
643	u32 function, index;
644	struct kvm_cpuid_entry2 *best;
645
646	function = kvm_register_read(vcpu, VCPU_REGS_RAX);
647	index = kvm_register_read(vcpu, VCPU_REGS_RCX);
648	kvm_register_write(vcpu, VCPU_REGS_RAX, 0);
649	kvm_register_write(vcpu, VCPU_REGS_RBX, 0);
650	kvm_register_write(vcpu, VCPU_REGS_RCX, 0);
651	kvm_register_write(vcpu, VCPU_REGS_RDX, 0);
652	best = kvm_find_cpuid_entry(vcpu, function, index);
653
654	if (!best)
655	best = check_cpuid_limit(vcpu, function, index);
656
657	if (best) {
658	kvm_register_write(vcpu, VCPU_REGS_RAX, best->eax);
659	kvm_register_write(vcpu, VCPU_REGS_RBX, best->ebx);
660	kvm_register_write(vcpu, VCPU_REGS_RCX, best->ecx);
661	kvm_register_write(vcpu, VCPU_REGS_RDX, best->edx);
662	}
663	kvm_x86_ops->skip_emulated_instruction(vcpu);
664	trace_kvm_cpuid(function,
665	kvm_register_read(vcpu, VCPU_REGS_RAX),
666	kvm_register_read(vcpu, VCPU_REGS_RBX),
667	kvm_register_read(vcpu, VCPU_REGS_RCX),
668	kvm_register_read(vcpu, VCPU_REGS_RDX));
669	}
670	EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);