[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;
	}
}

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