[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;
	}

	best = kvm_find_cpuid_entry(vcpu, 0xD, 0);
	if (!best)
		vcpu->arch.guest_supported_xcr0 = 0;
	else
		vcpu->arch.guest_supported_xcr0 =
			(best->eax | ((u64)best->edx << 32)) &
			host_xcr0 & KVM_SUPPORTED_XCR0;

	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 & KVM_SUPPORTED_XCR0 & 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;
	unsigned f_invpcid = kvm_x86_ops->invpcid_supported() ? F(INVPCID) : 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 */ |
		F(PCID) | 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_invpcid | 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 capability word 9 */
		if (index == 0) {
			entry->ebx &= kvm_supported_word9_x86_features;
			cpuid_mask(&entry->ebx, 9);
			// TSC_ADJUST is emulated
			entry->ebx |= F(TSC_ADJUST);
		} 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->eax &= host_xcr0 & KVM_SUPPORTED_XCR0;
		entry->edx &= (host_xcr0 & KVM_SUPPORTED_XCR0) >> 32;
		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: {
		static const char signature[12] = "KVMKVMKVM\0\0";
		const u32 *sigptr = (const 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_PV_EOI) |
			     (1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT) |
			     (1 << KVM_FEATURE_PV_UNHALT);

		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)(const struct kvm_cpuid_param *param);
};

static bool is_centaur_cpu(const 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 const 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++) {
		const 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_cpuid(struct kvm_vcpu *vcpu, u32 *eax, u32 *ebx, u32 *ecx, u32 *edx)
{
	u32 function = *eax, index = *ecx;
	struct kvm_cpuid_entry2 *best;

	best = kvm_find_cpuid_entry(vcpu, function, index);

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

	if (best) {
		*eax = best->eax;
		*ebx = best->ebx;
		*ecx = best->ecx;
		*edx = best->edx;
	} else
		*eax = *ebx = *ecx = *edx = 0;
}
EXPORT_SYMBOL_GPL(kvm_cpuid);

void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
{
	u32 function, eax, ebx, ecx, edx;

	function = eax = kvm_register_read(vcpu, VCPU_REGS_RAX);
	ecx = kvm_register_read(vcpu, VCPU_REGS_RCX);
	kvm_cpuid(vcpu, &eax, &ebx, &ecx, &edx);
	kvm_register_write(vcpu, VCPU_REGS_RAX, eax);
	kvm_register_write(vcpu, VCPU_REGS_RBX, ebx);
	kvm_register_write(vcpu, VCPU_REGS_RCX, ecx);
	kvm_register_write(vcpu, VCPU_REGS_RDX, edx);
	kvm_x86_ops->skip_emulated_instruction(vcpu);
	trace_kvm_cpuid(function, eax, ebx, ecx, edx);
}
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	48
d7876f1b PB	49	best = kvm_find_cpuid_entry(vcpu, 0xD, 0);
	50	if (!best)
	51	vcpu->arch.guest_supported_xcr0 = 0;
	52	else
	53	vcpu->arch.guest_supported_xcr0 =
	54	(best->eax \| ((u64)best->edx << 32)) &
	55	host_xcr0 & KVM_SUPPORTED_XCR0;
	56
f5132b01	57	kvm_pmu_cpuid_update(vcpu);
00b27a3e AK	58	}
	59
	60	static int is_efer_nx(void)
	61	{
	62	unsigned long long efer = 0;
	63
	64	rdmsrl_safe(MSR_EFER, &efer);
	65	return efer & EFER_NX;
	66	}
	67
	68	static void cpuid_fix_nx_cap(struct kvm_vcpu *vcpu)
	69	{
	70	int i;
	71	struct kvm_cpuid_entry2 e, entry;
	72
	73	entry = NULL;
	74	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
	75	e = &vcpu->arch.cpuid_entries[i];
	76	if (e->function == 0x80000001) {
	77	entry = e;
	78	break;
	79	}
	80	}
	81	if (entry && (entry->edx & (1 << 20)) && !is_efer_nx()) {
	82	entry->edx &= ~(1 << 20);
	83	printk(KERN_INFO "kvm: guest NX capability removed\n");
	84	}
	85	}
	86
	87	/* when an old userspace process fills a new kernel module */
	88	int kvm_vcpu_ioctl_set_cpuid(struct kvm_vcpu *vcpu,
	89	struct kvm_cpuid *cpuid,
	90	struct kvm_cpuid_entry __user *entries)
	91	{
	92	int r, i;
	93	struct kvm_cpuid_entry *cpuid_entries;
	94
	95	r = -E2BIG;
	96	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
	97	goto out;
	98	r = -ENOMEM;
	99	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry) * cpuid->nent);
	100	if (!cpuid_entries)
	101	goto out;
	102	r = -EFAULT;
	103	if (copy_from_user(cpuid_entries, entries,
	104	cpuid->nent * sizeof(struct kvm_cpuid_entry)))
	105	goto out_free;
	106	for (i = 0; i < cpuid->nent; i++) {
	107	vcpu->arch.cpuid_entries[i].function = cpuid_entries[i].function;
	108	vcpu->arch.cpuid_entries[i].eax = cpuid_entries[i].eax;
	109	vcpu->arch.cpuid_entries[i].ebx = cpuid_entries[i].ebx;
	110	vcpu->arch.cpuid_entries[i].ecx = cpuid_entries[i].ecx;
	111	vcpu->arch.cpuid_entries[i].edx = cpuid_entries[i].edx;
	112	vcpu->arch.cpuid_entries[i].index = 0;
	113	vcpu->arch.cpuid_entries[i].flags = 0;
	114	vcpu->arch.cpuid_entries[i].padding[0] = 0;
	115	vcpu->arch.cpuid_entries[i].padding[1] = 0;
	116	vcpu->arch.cpuid_entries[i].padding[2] = 0;
	117	}
	118	vcpu->arch.cpuid_nent = cpuid->nent;
	119	cpuid_fix_nx_cap(vcpu);
	120	r = 0;
	121	kvm_apic_set_version(vcpu);
122	kvm_x86_ops->cpuid_update(vcpu);
123	kvm_update_cpuid(vcpu);
124
125	out_free:
126	vfree(cpuid_entries);
127	out:
128	return r;
129	}
130
131	int kvm_vcpu_ioctl_set_cpuid2(struct kvm_vcpu *vcpu,
132	struct kvm_cpuid2 *cpuid,
133	struct kvm_cpuid_entry2 __user *entries)
134	{
135	int r;
136
137	r = -E2BIG;
138	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
139	goto out;
140	r = -EFAULT;
141	if (copy_from_user(&vcpu->arch.cpuid_entries, entries,
142	cpuid->nent * sizeof(struct kvm_cpuid_entry2)))
143	goto out;
144	vcpu->arch.cpuid_nent = cpuid->nent;
145	kvm_apic_set_version(vcpu);
146	kvm_x86_ops->cpuid_update(vcpu);
147	kvm_update_cpuid(vcpu);
148	return 0;
149
150	out:
151	return r;
152	}
153
154	int kvm_vcpu_ioctl_get_cpuid2(struct kvm_vcpu *vcpu,
155	struct kvm_cpuid2 *cpuid,
156	struct kvm_cpuid_entry2 __user *entries)
157	{
158	int r;
159
160	r = -E2BIG;
161	if (cpuid->nent < vcpu->arch.cpuid_nent)
162	goto out;
163	r = -EFAULT;
164	if (copy_to_user(entries, &vcpu->arch.cpuid_entries,
165	vcpu->arch.cpuid_nent * sizeof(struct kvm_cpuid_entry2)))
166	goto out;
167	return 0;
168
169	out:
170	cpuid->nent = vcpu->arch.cpuid_nent;
171	return r;
172	}
173
174	static void cpuid_mask(u32 *word, int wordnum)
175	{
176	*word &= boot_cpu_data.x86_capability[wordnum];
177	}
178
179	static void do_cpuid_1_ent(struct kvm_cpuid_entry2 *entry, u32 function,
180	u32 index)
181	{
182	entry->function = function;
183	entry->index = index;
184	cpuid_count(entry->function, entry->index,
185	&entry->eax, &entry->ebx, &entry->ecx, &entry->edx);
186	entry->flags = 0;
187	}
188
189	static bool supported_xcr0_bit(unsigned bit)
190	{
191	u64 mask = ((u64)1 << bit);
192
647e23bb	193	return mask & KVM_SUPPORTED_XCR0 & host_xcr0;
00b27a3e AK	194	}
	195
	196	#define F(x) bit(X86_FEATURE_##x)
	197
831bf664	198	static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
00b27a3e AK	199	u32 index, int *nent, int maxnent)
00b27a3e AK	200	{
831bf664	201	int r;
00b27a3e AK	202	unsigned f_nx = is_efer_nx() ? F(NX) : 0;
	203	#ifdef CONFIG_X86_64
	204	unsigned f_gbpages = (kvm_x86_ops->get_lpage_level() == PT_PDPE_LEVEL)
	205	? F(GBPAGES) : 0;
	206	unsigned f_lm = F(LM);
	207	#else
	208	unsigned f_gbpages = 0;
	209	unsigned f_lm = 0;
	210	#endif
	211	unsigned f_rdtscp = kvm_x86_ops->rdtscp_supported() ? F(RDTSCP) : 0;
ad756a16	212	unsigned f_invpcid = kvm_x86_ops->invpcid_supported() ? F(INVPCID) : 0;
00b27a3e AK	213
	214	/* cpuid 1.edx */
	215	const u32 kvm_supported_word0_x86_features =
	216	F(FPU) \| F(VME) \| F(DE) \| F(PSE) \|
	217	F(TSC) \| F(MSR) \| F(PAE) \| F(MCE) \|
	218	F(CX8) \| F(APIC) \| 0 /* Reserved */ \| F(SEP) \|
	219	F(MTRR) \| F(PGE) \| F(MCA) \| F(CMOV) \|
	220	F(PAT) \| F(PSE36) \| 0 /* PSN */ \| F(CLFLSH) \|
	221	0 /* Reserved, DS, ACPI */ \| F(MMX) \|
	222	F(FXSR) \| F(XMM) \| F(XMM2) \| F(SELFSNOOP) \|
	223	0 /* HTT, TM, Reserved, PBE */;
	224	/* cpuid 0x80000001.edx */
	225	const u32 kvm_supported_word1_x86_features =
	226	F(FPU) \| F(VME) \| F(DE) \| F(PSE) \|
	227	F(TSC) \| F(MSR) \| F(PAE) \| F(MCE) \|
	228	F(CX8) \| F(APIC) \| 0 /* Reserved */ \| F(SYSCALL) \|
	229	F(MTRR) \| F(PGE) \| F(MCA) \| F(CMOV) \|
	230	F(PAT) \| F(PSE36) \| 0 /* Reserved */ \|
	231	f_nx \| 0 /* Reserved */ \| F(MMXEXT) \| F(MMX) \|
	232	F(FXSR) \| F(FXSR_OPT) \| f_gbpages \| f_rdtscp \|
	233	0 /* Reserved */ \| f_lm \| F(3DNOWEXT) \| F(3DNOW);
	234	/* cpuid 1.ecx */
	235	const u32 kvm_supported_word4_x86_features =
	236	F(XMM3) \| F(PCLMULQDQ) \| 0 /* DTES64, MONITOR */ \|
	237	0 /* DS-CPL, VMX, SMX, EST */ \|
	238	0 /* TM2 / \| F(SSSE3) \| 0 / CNXT-ID / \| 0 / Reserved */ \|
fb215366	239	F(FMA) \| F(CX16) \| 0 /* xTPR Update, PDCM */ \|
ad756a16	240	F(PCID) \| 0 /* Reserved, DCA */ \| F(XMM4_1) \|
00b27a3e AK	241	F(XMM4_2) \| F(X2APIC) \| F(MOVBE) \| F(POPCNT) \|
	242	0 /* Reserved/ \| F(AES) \| F(XSAVE) \| 0 / OSXSAVE */ \| F(AVX) \|
	243	F(F16C) \| F(RDRAND);
	244	/* cpuid 0x80000001.ecx */
	245	const u32 kvm_supported_word6_x86_features =
	246	F(LAHF_LM) \| F(CMP_LEGACY) \| 0 /SVM/ \| 0 /* ExtApicSpace */ \|
	247	F(CR8_LEGACY) \| F(ABM) \| F(SSE4A) \| F(MISALIGNSSE) \|
2b036c6b	248	F(3DNOWPREFETCH) \| F(OSVW) \| 0 /* IBS */ \| F(XOP) \|
00b27a3e AK	249	0 /* SKINIT, WDT, LWP */ \| F(FMA4) \| F(TBM);
	250
	251	/* cpuid 0xC0000001.edx */
	252	const u32 kvm_supported_word5_x86_features =
	253	F(XSTORE) \| F(XSTORE_EN) \| F(XCRYPT) \| F(XCRYPT_EN) \|
	254	F(ACE2) \| F(ACE2_EN) \| F(PHE) \| F(PHE_EN) \|
	255	F(PMM) \| F(PMM_EN);
	256
	257	/* cpuid 7.0.ebx */
	258	const u32 kvm_supported_word9_x86_features =
83c52915	259	F(FSGSBASE) \| F(BMI1) \| F(HLE) \| F(AVX2) \| F(SMEP) \|
ad756a16	260	F(BMI2) \| F(ERMS) \| f_invpcid \| F(RTM);
00b27a3e AK	261
	262	/* all calls to cpuid_count() should be made on the same cpu */
	263	get_cpu();
831bf664 SL	264
	265	r = -E2BIG;
	266
	267	if (*nent >= maxnent)
	268	goto out;
	269
00b27a3e AK	270	do_cpuid_1_ent(entry, function, index);
	271	++*nent;
	272
	273	switch (function) {
	274	case 0:
	275	entry->eax = min(entry->eax, (u32)0xd);
	276	break;
	277	case 1:
	278	entry->edx &= kvm_supported_word0_x86_features;
	279	cpuid_mask(&entry->edx, 0);
	280	entry->ecx &= kvm_supported_word4_x86_features;
	281	cpuid_mask(&entry->ecx, 4);
	282	/* we support x2apic emulation even if host does not support
	283	* it since we emulate x2apic in software */
	284	entry->ecx \|= F(X2APIC);
	285	break;
	286	/* function 2 entries are STATEFUL. That is, repeated cpuid commands
	287	* may return different values. This forces us to get_cpu() before
	288	* issuing the first command, and also to emulate this annoying behavior
	289	* in kvm_emulate_cpuid() using KVM_CPUID_FLAG_STATE_READ_NEXT */
	290	case 2: {
	291	int t, times = entry->eax & 0xff;
	292
	293	entry->flags \|= KVM_CPUID_FLAG_STATEFUL_FUNC;
	294	entry->flags \|= KVM_CPUID_FLAG_STATE_READ_NEXT;
831bf664 SL	295	for (t = 1; t < times; ++t) {
	296	if (*nent >= maxnent)
	297	goto out;
	298
00b27a3e AK	299	do_cpuid_1_ent(&entry[t], function, 0);
	300	entry[t].flags \|= KVM_CPUID_FLAG_STATEFUL_FUNC;
	301	++*nent;
	302	}
	303	break;
	304	}
	305	/* function 4 has additional index. */
	306	case 4: {
	307	int i, cache_type;
	308
	309	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	310	/* read more entries until cache_type is zero */
831bf664 SL	311	for (i = 1; ; ++i) {
	312	if (*nent >= maxnent)
	313	goto out;
	314
00b27a3e AK	315	cache_type = entry[i - 1].eax & 0x1f;
	316	if (!cache_type)
	317	break;
	318	do_cpuid_1_ent(&entry[i], function, i);
	319	entry[i].flags \|=
	320	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	321	++*nent;
	322	}
	323	break;
	324	}
	325	case 7: {
	326	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
bbbda795	327	/* Mask ebx against host capability word 9 */
00b27a3e AK	328	if (index == 0) {
	329	entry->ebx &= kvm_supported_word9_x86_features;
	330	cpuid_mask(&entry->ebx, 9);
ba904635 WA	331	// TSC_ADJUST is emulated
ba904635 WA	332	entry->ebx \|= F(TSC_ADJUST);
00b27a3e AK	333	} else
	334	entry->ebx = 0;
	335	entry->eax = 0;
	336	entry->ecx = 0;
	337	entry->edx = 0;
	338	break;
	339	}
	340	case 9:
	341	break;
a6c06ed1 GN	342	case 0xa: { /* Architectural Performance Monitoring */
	343	struct x86_pmu_capability cap;
	344	union cpuid10_eax eax;
	345	union cpuid10_edx edx;
	346
	347	perf_get_x86_pmu_capability(&cap);
	348
	349	/*
	350	* Only support guest architectural pmu on a host
	351	* with architectural pmu.
	352	*/
	353	if (!cap.version)
	354	memset(&cap, 0, sizeof(cap));
	355
	356	eax.split.version_id = min(cap.version, 2);
	357	eax.split.num_counters = cap.num_counters_gp;
	358	eax.split.bit_width = cap.bit_width_gp;
	359	eax.split.mask_length = cap.events_mask_len;
	360
	361	edx.split.num_counters_fixed = cap.num_counters_fixed;
	362	edx.split.bit_width_fixed = cap.bit_width_fixed;
	363	edx.split.reserved = 0;
	364
	365	entry->eax = eax.full;
	366	entry->ebx = cap.events_mask;
	367	entry->ecx = 0;
	368	entry->edx = edx.full;
	369	break;
	370	}
00b27a3e AK	371	/* function 0xb has additional index. */
	372	case 0xb: {
	373	int i, level_type;
	374
	375	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	376	/* read more entries until level_type is zero */
831bf664 SL	377	for (i = 1; ; ++i) {
	378	if (*nent >= maxnent)
	379	goto out;
	380
00b27a3e AK	381	level_type = entry[i - 1].ecx & 0xff00;
	382	if (!level_type)
	383	break;
	384	do_cpuid_1_ent(&entry[i], function, i);
	385	entry[i].flags \|=
	386	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	387	++*nent;
	388	}
	389	break;
	390	}
	391	case 0xd: {
	392	int idx, i;
	393
647e23bb PB	394	entry->eax &= host_xcr0 & KVM_SUPPORTED_XCR0;
647e23bb PB	395	entry->edx &= (host_xcr0 & KVM_SUPPORTED_XCR0) >> 32;
00b27a3e	396	entry->flags \|= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
831bf664 SL	397	for (idx = 1, i = 1; idx < 64; ++idx) {
	398	if (*nent >= maxnent)
	399	goto out;
	400
00b27a3e AK	401	do_cpuid_1_ent(&entry[i], function, idx);
	402	if (entry[i].eax == 0 \|\| !supported_xcr0_bit(idx))
	403	continue;
	404	entry[i].flags \|=
	405	KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
	406	++*nent;
	407	++i;
	408	}
	409	break;
	410	}
	411	case KVM_CPUID_SIGNATURE: {
326d07cb MK	412	static const char signature[12] = "KVMKVMKVM\0\0";
326d07cb MK	413	const u32 sigptr = (const u32 )signature;
57c22e5f	414	entry->eax = KVM_CPUID_FEATURES;
00b27a3e AK	415	entry->ebx = sigptr[0];
	416	entry->ecx = sigptr[1];
	417	entry->edx = sigptr[2];
	418	break;
	419	}
	420	case KVM_CPUID_FEATURES:
	421	entry->eax = (1 << KVM_FEATURE_CLOCKSOURCE) \|
	422	(1 << KVM_FEATURE_NOP_IO_DELAY) \|
	423	(1 << KVM_FEATURE_CLOCKSOURCE2) \|
	424	(1 << KVM_FEATURE_ASYNC_PF) \|
ae7a2a3f	425	(1 << KVM_FEATURE_PV_EOI) \|
6aef266c SV	426	(1 << KVM_FEATURE_CLOCKSOURCE_STABLE_BIT) \|
6aef266c SV	427	(1 << KVM_FEATURE_PV_UNHALT);
00b27a3e AK	428
	429	if (sched_info_on())
	430	entry->eax \|= (1 << KVM_FEATURE_STEAL_TIME);
	431
	432	entry->ebx = 0;
	433	entry->ecx = 0;
	434	entry->edx = 0;
	435	break;
	436	case 0x80000000:
	437	entry->eax = min(entry->eax, 0x8000001a);
	438	break;
	439	case 0x80000001:
	440	entry->edx &= kvm_supported_word1_x86_features;
	441	cpuid_mask(&entry->edx, 1);
	442	entry->ecx &= kvm_supported_word6_x86_features;
	443	cpuid_mask(&entry->ecx, 6);
	444	break;
	445	case 0x80000008: {
	446	unsigned g_phys_as = (entry->eax >> 16) & 0xff;
	447	unsigned virt_as = max((entry->eax >> 8) & 0xff, 48U);
	448	unsigned phys_as = entry->eax & 0xff;
	449
	450	if (!g_phys_as)
	451	g_phys_as = phys_as;
	452	entry->eax = g_phys_as \| (virt_as << 8);
	453	entry->ebx = entry->edx = 0;
	454	break;
	455	}
	456	case 0x80000019:
	457	entry->ecx = entry->edx = 0;
	458	break;
	459	case 0x8000001a:
	460	break;
	461	case 0x8000001d:
	462	break;
	463	/Add support for Centaur's CPUID instruction/
	464	case 0xC0000000:
	465	/Just support up to 0xC0000004 now/
	466	entry->eax = min(entry->eax, 0xC0000004);
	467	break;
	468	case 0xC0000001:
	469	entry->edx &= kvm_supported_word5_x86_features;
	470	cpuid_mask(&entry->edx, 5);
	471	break;
	472	case 3: /* Processor serial number */
	473	case 5: /* MONITOR/MWAIT */
	474	case 6: /* Thermal management */
00b27a3e AK	475	case 0x80000007: /* Advanced power management */
	476	case 0xC0000002:
	477	case 0xC0000003:
	478	case 0xC0000004:
	479	default:
	480	entry->eax = entry->ebx = entry->ecx = entry->edx = 0;
	481	break;
	482	}
	483
	484	kvm_x86_ops->set_supported_cpuid(function, entry);
	485
831bf664 SL	486	r = 0;
	487
	488	out:
00b27a3e	489	put_cpu();
831bf664 SL	490
831bf664 SL	491	return r;
00b27a3e AK	492	}
	493
	494	#undef F
	495
831bf664 SL	496	struct kvm_cpuid_param {
	497	u32 func;
	498	u32 idx;
	499	bool has_leaf_count;
326d07cb	500	bool (qualifier)(const struct kvm_cpuid_param param);
831bf664 SL	501	};
831bf664 SL	502
326d07cb	503	static bool is_centaur_cpu(const struct kvm_cpuid_param *param)
831bf664 SL	504	{
	505	return boot_cpu_data.x86_vendor == X86_VENDOR_CENTAUR;
	506	}
	507
00b27a3e AK	508	int kvm_dev_ioctl_get_supported_cpuid(struct kvm_cpuid2 *cpuid,
	509	struct kvm_cpuid_entry2 __user *entries)
	510	{
	511	struct kvm_cpuid_entry2 *cpuid_entries;
831bf664	512	int limit, nent = 0, r = -E2BIG, i;
00b27a3e	513	u32 func;
326d07cb	514	static const struct kvm_cpuid_param param[] = {
831bf664 SL	515	{ .func = 0, .has_leaf_count = true },
	516	{ .func = 0x80000000, .has_leaf_count = true },
	517	{ .func = 0xC0000000, .qualifier = is_centaur_cpu, .has_leaf_count = true },
	518	{ .func = KVM_CPUID_SIGNATURE },
	519	{ .func = KVM_CPUID_FEATURES },
	520	};
00b27a3e AK	521
	522	if (cpuid->nent < 1)
	523	goto out;
	524	if (cpuid->nent > KVM_MAX_CPUID_ENTRIES)
	525	cpuid->nent = KVM_MAX_CPUID_ENTRIES;
	526	r = -ENOMEM;
	527	cpuid_entries = vmalloc(sizeof(struct kvm_cpuid_entry2) * cpuid->nent);
	528	if (!cpuid_entries)
	529	goto out;
	530
831bf664 SL	531	r = 0;
831bf664 SL	532	for (i = 0; i < ARRAY_SIZE(param); i++) {
326d07cb	533	const struct kvm_cpuid_param *ent = &param[i];
00b27a3e	534
831bf664 SL	535	if (ent->qualifier && !ent->qualifier(ent))
831bf664 SL	536	continue;
00b27a3e	537
831bf664	538	r = do_cpuid_ent(&cpuid_entries[nent], ent->func, ent->idx,
00b27a3e AK	539	&nent, cpuid->nent);
00b27a3e AK	540
831bf664	541	if (r)
00b27a3e AK	542	goto out_free;
00b27a3e AK	543
831bf664 SL	544	if (!ent->has_leaf_count)
	545	continue;
	546
00b27a3e	547	limit = cpuid_entries[nent - 1].eax;
831bf664 SL	548	for (func = ent->func + 1; func <= limit && nent < cpuid->nent && r == 0; ++func)
	549	r = do_cpuid_ent(&cpuid_entries[nent], func, ent->idx,
	550	&nent, cpuid->nent);
00b27a3e	551
831bf664	552	if (r)
00b27a3e AK	553	goto out_free;
	554	}
	555
00b27a3e AK	556	r = -EFAULT;
	557	if (copy_to_user(entries, cpuid_entries,
	558	nent * sizeof(struct kvm_cpuid_entry2)))
	559	goto out_free;
	560	cpuid->nent = nent;
	561	r = 0;
	562
	563	out_free:
	564	vfree(cpuid_entries);
	565	out:
	566	return r;
	567	}
	568
	569	static int move_to_next_stateful_cpuid_entry(struct kvm_vcpu *vcpu, int i)
	570	{
	571	struct kvm_cpuid_entry2 *e = &vcpu->arch.cpuid_entries[i];
	572	int j, nent = vcpu->arch.cpuid_nent;
	573
	574	e->flags &= ~KVM_CPUID_FLAG_STATE_READ_NEXT;
	575	/* when no next entry is found, the current entry[i] is reselected */
	576	for (j = i + 1; ; j = (j + 1) % nent) {
	577	struct kvm_cpuid_entry2 *ej = &vcpu->arch.cpuid_entries[j];
	578	if (ej->function == e->function) {
	579	ej->flags \|= KVM_CPUID_FLAG_STATE_READ_NEXT;
	580	return j;
	581	}
	582	}
	583	return 0; /* silence gcc, even though control never reaches here */
	584	}
	585
	586	/* find an entry with matching function, matching index (if needed), and that
	587	* should be read next (if it's stateful) */
	588	static int is_matching_cpuid_entry(struct kvm_cpuid_entry2 *e,
	589	u32 function, u32 index)
	590	{
	591	if (e->function != function)
	592	return 0;
	593	if ((e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) && e->index != index)
	594	return 0;
	595	if ((e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC) &&
	596	!(e->flags & KVM_CPUID_FLAG_STATE_READ_NEXT))
	597	return 0;
	598	return 1;
	599	}
	600
	601	struct kvm_cpuid_entry2 kvm_find_cpuid_entry(struct kvm_vcpu vcpu,
	602	u32 function, u32 index)
	603	{
	604	int i;
	605	struct kvm_cpuid_entry2 *best = NULL;
	606
	607	for (i = 0; i < vcpu->arch.cpuid_nent; ++i) {
	608	struct kvm_cpuid_entry2 *e;
	609
	610	e = &vcpu->arch.cpuid_entries[i];
	611	if (is_matching_cpuid_entry(e, function, index)) {
	612	if (e->flags & KVM_CPUID_FLAG_STATEFUL_FUNC)
	613	move_to_next_stateful_cpuid_entry(vcpu, i);
	614	best = e;
	615	break;
	616	}
	617	}
	618	return best;
	619	}
620	EXPORT_SYMBOL_GPL(kvm_find_cpuid_entry);
621
622	int cpuid_maxphyaddr(struct kvm_vcpu *vcpu)
623	{
624	struct kvm_cpuid_entry2 *best;
625
626	best = kvm_find_cpuid_entry(vcpu, 0x80000000, 0);
627	if (!best \|\| best->eax < 0x80000008)
628	goto not_found;
629	best = kvm_find_cpuid_entry(vcpu, 0x80000008, 0);
630	if (best)
631	return best->eax & 0xff;
632	not_found:
633	return 36;
634	}
635
636	/*
637	* If no match is found, check whether we exceed the vCPU's limit
638	* and return the content of the highest valid _standard_ leaf instead.
639	* This is to satisfy the CPUID specification.
640	*/
641	static struct kvm_cpuid_entry2* check_cpuid_limit(struct kvm_vcpu *vcpu,
642	u32 function, u32 index)
643	{
644	struct kvm_cpuid_entry2 *maxlevel;
645
646	maxlevel = kvm_find_cpuid_entry(vcpu, function & 0x80000000, 0);
647	if (!maxlevel \|\| maxlevel->eax >= function)
648	return NULL;
649	if (function & 0x80000000) {
650	maxlevel = kvm_find_cpuid_entry(vcpu, 0, 0);
651	if (!maxlevel)
652	return NULL;
653	}
654	return kvm_find_cpuid_entry(vcpu, maxlevel->eax, index);
655	}
656
62046e5a	657	void kvm_cpuid(struct kvm_vcpu vcpu, u32 eax, u32 ebx, u32 ecx, u32 *edx)
00b27a3e	658	{
62046e5a	659	u32 function = eax, index = ecx;
00b27a3e AK	660	struct kvm_cpuid_entry2 *best;
00b27a3e AK	661
00b27a3e AK	662	best = kvm_find_cpuid_entry(vcpu, function, index);
	663
	664	if (!best)
	665	best = check_cpuid_limit(vcpu, function, index);
	666
	667	if (best) {
62046e5a AK	668	*eax = best->eax;
	669	*ebx = best->ebx;
	670	*ecx = best->ecx;
	671	*edx = best->edx;
	672	} else
	673	eax = ebx = ecx = edx = 0;
	674	}
66f7b72e	675	EXPORT_SYMBOL_GPL(kvm_cpuid);
62046e5a AK	676
	677	void kvm_emulate_cpuid(struct kvm_vcpu *vcpu)
	678	{
	679	u32 function, eax, ebx, ecx, edx;
	680
	681	function = eax = kvm_register_read(vcpu, VCPU_REGS_RAX);
	682	ecx = kvm_register_read(vcpu, VCPU_REGS_RCX);
	683	kvm_cpuid(vcpu, &eax, &ebx, &ecx, &edx);
	684	kvm_register_write(vcpu, VCPU_REGS_RAX, eax);
	685	kvm_register_write(vcpu, VCPU_REGS_RBX, ebx);
	686	kvm_register_write(vcpu, VCPU_REGS_RCX, ecx);
	687	kvm_register_write(vcpu, VCPU_REGS_RDX, edx);
00b27a3e	688	kvm_x86_ops->skip_emulated_instruction(vcpu);
62046e5a	689	trace_kvm_cpuid(function, eax, ebx, ecx, edx);
00b27a3e AK	690	}
00b27a3e AK	691	EXPORT_SYMBOL_GPL(kvm_emulate_cpuid);