// Copyright 2017 syzkaller project authors. All rights reserved.
// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.

#include <stdint.h>
#include <sys/utsname.h>

static unsigned host_kernel_version();
static void dump_cpu_state(int cpufd, char* vm_mem);

#ifdef GOARCH_amd64
static int cpu_feature_enabled(uint32_t function, uint32_t eax_bits, uint32_t ebx_bits, uint32_t ecx_bits, uint32_t edx_bits);

#define FEATURE_INTEL 0x00000001
#define FEATURE_INTEL_ECX_VMX (1 << 5)

#define FEATURE_AMD 0x800000001
#define FEATURE_AMD_ECX_SVM (1 << 2)
#endif

static int test_one(int text_type, const char* text, int text_size, int flags, unsigned reason, bool check_rax)
{
	printf("=== testing text %d, text size 0x%x, flags 0x%x\n", text_type, text_size, flags);
	int kvmfd = open("/dev/kvm", O_RDWR);
	if (kvmfd == -1) {
		if (errno == ENOENT) {
			printf("/dev/kvm is not present\n");
			return -1;
		}
		if (errno == EPERM || errno == EACCES) {
			printf("no permissions to open /dev/kvm\n");
			return -1;
		}
		printf("failed to open /dev/kvm (%d)\n", errno);
		return 1;
	}
	int vmfd = ioctl(kvmfd, KVM_CREATE_VM, 0);
	if (vmfd == -1) {
		printf("KVM_CREATE_VM failed (%d)\n", errno);
		return 1;
	}
	int cpufd = ioctl(vmfd, KVM_CREATE_VCPU, 0);
	if (cpufd == -1) {
		printf("KVM_CREATE_VCPU failed (%d)\n", errno);
		return 1;
	}
	int cpu_mem_size = ioctl(kvmfd, KVM_GET_VCPU_MMAP_SIZE, 0);
	if (cpu_mem_size <= 0) {
		printf("KVM_GET_VCPU_MMAP_SIZE failed (%d)\n", errno);
		return 1;
	}
	struct kvm_run* cpu_mem = (struct kvm_run*)mmap(0, cpu_mem_size,
							PROT_READ | PROT_WRITE, MAP_SHARED, cpufd, 0);
	if (cpu_mem == MAP_FAILED) {
		printf("cpu mmap failed (%d)\n", errno);
		return 1;
	}
	int vm_mem_size = 24 * SYZ_PAGE_SIZE; // Allocate what executor allocates for vma[24]
	void* vm_mem = mmap(0, vm_mem_size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
	if (vm_mem == MAP_FAILED) {
		printf("mmap failed (%d)\n", errno);
		return 1;
	}
	struct kvm_text kvm_text;
	kvm_text.typ = text_type;
	kvm_text.text = text;
	kvm_text.size = text_size;
	if (syz_kvm_setup_cpu(vmfd, cpufd, (uintptr_t)vm_mem, (uintptr_t)&kvm_text, 1, flags, 0, 0)) {
		printf("syz_kvm_setup_cpu failed (%d)\n", errno);
		return 1;
	}

	int ret = ioctl(cpufd, KVM_RUN, 0);
	// KVM_RUN returns positive values on PPC64
	if (ret < 0) {
		printf("KVM_RUN returned %d, errno=%d\n", ret, errno);
		return 1;
	}
	struct kvm_regs regs;
	if (ioctl(cpufd, KVM_GET_REGS, &regs)) {
		printf("KVM_GET_REGS failed (%d)\n", errno);
		dump_cpu_state(cpufd, (char*)vm_mem);
		return 1;
	}
	if (cpu_mem->exit_reason != reason) {
		printf("KVM_RUN exit reason %d, expect %d\n", cpu_mem->exit_reason, reason);
		if (cpu_mem->exit_reason == KVM_EXIT_FAIL_ENTRY)
			printf("hardware exit reason 0x%llx\n",
			       (unsigned long long)cpu_mem->fail_entry.hardware_entry_failure_reason);
		dump_cpu_state(cpufd, (char*)vm_mem);
		return 1;
	}
#ifdef GOARCH_amd64
	if (check_rax && regs.rax != 0xbadc0de) {
		printf("wrong result: rax=0x%llx\n", (long long)regs.rax);
		dump_cpu_state(cpufd, (char*)vm_mem);
		return 1;
	}
#elif GOARCH_ppc64le
	if (check_rax && regs.gpr[3] != 0xbadc0de) {
		printf("wrong result: gps[3]=0x%llx\n", (long long)regs.gpr[3]);
		dump_cpu_state(cpufd, (char*)vm_mem);
		return 1;
	}
#endif
	munmap(vm_mem, vm_mem_size);
	munmap(cpu_mem, cpu_mem_size);
	close(cpufd);
	close(vmfd);
	close(kvmfd);
	return 0;
}

static int test_kvm()
{
	int res = 0;

	unsigned ver = host_kernel_version();
	printf("host kernel version %u\n", ver);

	// TODO: test VM mode.
	// const char text16_vm[] = "\x48\xc7\xc3\xde\xc0\xad\x0b\x90\x90\x48\xc7\xc0\xef\xcd\xab\x00\xf4";
	// if (res = test_one(64, text16_vm, sizeof(text16_vm) - 1, KVM_SETUP_VM, KVM_EXIT_HLT, true))
	//	return res;

	// TODO: test code executed in interrupt handlers.
	// const char text32_div0[] = "\x31\xc0\xf7\xf0";
	// if (res = test_one(32, text32_div0, sizeof(text32_div0)-1, 0, KVM_EXIT_HLT, true))
	//	return res;

#ifdef GOARCH_amd64
	// Note: VIRT86 and CPL3 prefix seems to work with vmx only
	if (cpu_feature_enabled(FEATURE_INTEL, 0, 0, FEATURE_INTEL_ECX_VMX, 0) == 1) {
		const char text8[] = "\x66\xb8\xde\xc0\xad\x0b";
		if ((res = test_one(8, text8, sizeof(text8) - 1, 0, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(8, text8, sizeof(text8) - 1, KVM_SETUP_VIRT86, KVM_EXIT_SHUTDOWN, true)))
			return res;
		if ((res = test_one(8, text8, sizeof(text8) - 1, KVM_SETUP_VIRT86 | KVM_SETUP_PAGING, KVM_EXIT_SHUTDOWN, true)))
			return res;

		const char text16[] = "\x66\xb8\xde\xc0\xad\x0b";
		if ((res = test_one(16, text16, sizeof(text16) - 1, 0, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(16, text16, sizeof(text16) - 1, KVM_SETUP_CPL3, KVM_EXIT_SHUTDOWN, true)))
			return res;

		const char text32[] = "\xb8\xde\xc0\xad\x0b";
		if ((res = test_one(32, text32, sizeof(text32) - 1, 0, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(32, text32, sizeof(text32) - 1, KVM_SETUP_PAGING, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(32, text32, sizeof(text32) - 1, KVM_SETUP_CPL3, KVM_EXIT_SHUTDOWN, true)))
			return res;

		const char text64[] = "\x90\xb8\xde\xc0\xad\x0b";
		if ((res = test_one(64, text64, sizeof(text64) - 1, 0, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(64, text64, sizeof(text64) - 1, KVM_SETUP_PAGING, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(64, text64, sizeof(text64) - 1, KVM_SETUP_CPL3, KVM_EXIT_SHUTDOWN, true)))
			return res;

		const char text64_sysenter[] = "\xb8\xde\xc0\xad\x0b\x0f\x34";
		if ((res = test_one(64, text64_sysenter, sizeof(text64_sysenter) - 1, KVM_SETUP_CPL3, KVM_EXIT_SHUTDOWN, true)))
			return res;
	}

	// Note: SMM does not work on 3.13 kernels.
	if (ver >= 404) {
		const char text8_smm[] = "\x66\xb8\xde\xc0\xad\x0b";
		if ((res = test_one(8, text8_smm, sizeof(text8_smm) - 1, KVM_SETUP_SMM, KVM_EXIT_HLT, true)))
			return res;
		if ((res = test_one(8, text8_smm, sizeof(text8_smm) - 1, KVM_SETUP_SMM | KVM_SETUP_PROTECTED, KVM_EXIT_HLT, true)))
			return res;

		// const char text32_smm[] = "\xb8\xde\xc0\xad\x0b";
		if ((res = test_one(32, text8_smm, sizeof(text8_smm) - 1, KVM_SETUP_SMM, KVM_EXIT_HLT, true)))
			return res;

		// Also ensure that we are actually in SMM.
		// If we do MOV to RAX and then RSM, RAX will be restored to host value so RAX check will fail.
		// So instead we execute just RSM, if we are in SMM we will get KVM_EXIT_HLT,
		// otherwise KVM_EXIT_INTERNAL_ERROR.
		const char text_rsm[] = "\x0f\xaa";
		if ((res = test_one(8, text_rsm, sizeof(text_rsm) - 1, KVM_SETUP_SMM, KVM_EXIT_HLT, false)))
			return res;
		if ((res = test_one(32, text_rsm, sizeof(text_rsm) - 1, KVM_SETUP_SMM, KVM_EXIT_HLT, false)))
			return res;
	}
#elif GOARCH_ppc64le
	for (unsigned i = 0; i < (1 << 5); ++i) {
		res = test_one(8, kvm_ppc64_mr, sizeof(kvm_ppc64_mr) - 1, i, KVM_EXIT_DEBUG, true);
		if (res)
			return res;
		res = test_one(8, kvm_ppc64_ld, sizeof(kvm_ppc64_ld) - 1, i, KVM_EXIT_DEBUG, true);
		if (res)
			return res;
	}
#else
	// Keeping gcc happy
	const char text8[] = "\x66\xb8\xde\xc0\xad\x0b";
	if ((res = test_one(8, text8, sizeof(text8) - 1, 0, KVM_EXIT_HLT, true)))
		return res;
#endif

	return 0;
}

static unsigned host_kernel_version()
{
	struct utsname name;
	if (uname(&name)) {
		printf("uname failed (%d)\n", errno);
		doexit(1);
	}
	unsigned major = atoi(name.release);
	unsigned minor = 0;
	if (strchr(name.release, '.'))
		minor = atoi(strchr(name.release, '.') + 1);
	return major * 100 + minor;
}

#ifdef GOARCH_amd64
static void dump_seg(const char* name, struct kvm_segment* seg)
{
	printf("%s: base=0x%llx limit=0x%x sel=0x%x type=%d p=%d dpl=%d, db=%d s=%d l=%d g=%d\n",
	       name, seg->base, seg->limit, seg->selector, seg->type, seg->present, seg->dpl, seg->db, seg->s, seg->l, seg->g);
}
#endif

static void dump_cpu_state(int cpufd, char* vm_mem)
{
	struct kvm_sregs sregs;
	if (ioctl(cpufd, KVM_GET_SREGS, &sregs)) {
		printf("KVM_GET_SREGS failed (%d)\n", errno);
		return;
	}
	struct kvm_regs regs;
	if (ioctl(cpufd, KVM_GET_REGS, &regs)) {
		printf("KVM_GET_REGS failed (%d)\n", errno);
		return;
	}
#ifdef GOARCH_amd64
	printf("RIP=0x%llx RAX=0x%llx RDX=0x%llx RCX=0x%llx RBX=0x%llx CF=%d ZF=%d\n",
	       regs.rip, regs.rax, regs.rdx, regs.rcx, regs.rbx, !!(regs.rflags & (1 << 0)), !!(regs.rflags & (1 << 6)));
	printf("CR0=0x%llx CR2=0x%llx CR4=0x%llx EFER=0x%llx\n",
	       sregs.cr0, sregs.cr2, sregs.cr4, sregs.efer);
	dump_seg("CS", &sregs.cs);
	dump_seg("SS", &sregs.ss);
	dump_seg("DS", &sregs.ds);

	if (false) {
		printf("memory:\n");
		for (int i = 0; i < 0x80; i++)
			printf("0x%02x: 0x%02x\n", i, ((unsigned char*)vm_mem)[i]);
	}

	if (false) {
		printf("vmcs:\n");
		const int vmcs_size = 0x1000;
		for (int i = 0; i < vmcs_size / 8; i += 4) {
			printf("0x%04x: 0x%016llx 0x%016llx 0x%016llx 0x%016llx\n", i,
			       ((long long*)vm_mem)[i], ((long long*)vm_mem)[i + 1], ((long long*)vm_mem)[i + 2], ((long long*)vm_mem)[i + 3]);
		}
	}
#elif GOARCH_ppc64 || GOARCH_ppc64le
	printf("NIP %016lx\n", regs.pc);
	printf("MSR %016lx\n", regs.msr);
	printf("GPR00 %016lx %016lx %016lx %016lx\n", regs.gpr[0], regs.gpr[1], regs.gpr[2], regs.gpr[3]);
	printf("GPR04 %016lx %016lx %016lx %016lx\n", regs.gpr[4], regs.gpr[5], regs.gpr[6], regs.gpr[7]);
	printf("GPR08 %016lx %016lx %016lx %016lx\n", regs.gpr[8], regs.gpr[9], regs.gpr[10], regs.gpr[11]);
	printf("GPR12 %016lx %016lx %016lx %016lx\n", regs.gpr[12], regs.gpr[13], regs.gpr[14], regs.gpr[15]);
	printf("GPR16 %016lx %016lx %016lx %016lx\n", regs.gpr[16], regs.gpr[17], regs.gpr[18], regs.gpr[19]);
	printf("GPR20 %016lx %016lx %016lx %016lx\n", regs.gpr[20], regs.gpr[21], regs.gpr[22], regs.gpr[23]);
	printf("GPR24 %016lx %016lx %016lx %016lx\n", regs.gpr[24], regs.gpr[25], regs.gpr[26], regs.gpr[27]);
	printf("GPR28 %016lx %016lx %016lx %016lx\n", regs.gpr[28], regs.gpr[29], regs.gpr[30], regs.gpr[31]);
	printf(" SRR0 %016lx  SRR1 %016lx\n", regs.srr0, regs.srr1);
#endif
}

#ifdef GOARCH_amd64
// retcodes:
//  0 : feature disabled
//  1 : feature enabled
// -1 : error getting feature state
static int cpu_feature_enabled(uint32_t function, uint32_t eax_bits, uint32_t ebx_bits, uint32_t ecx_bits, uint32_t edx_bits)
{
	int kvmfd = open("/dev/kvm", O_RDWR);
	if (kvmfd == -1) {
		printf("failed to open /dev/kvm (%d)\n", errno);
		return -1;
	}
	char buf[sizeof(struct kvm_cpuid2) + 128 * sizeof(struct kvm_cpuid_entry2)];
	memset(buf, 0, sizeof(buf));
	struct kvm_cpuid2* cpuid = (struct kvm_cpuid2*)buf;
	cpuid->nent = 128;
	ioctl(kvmfd, KVM_GET_SUPPORTED_CPUID, cpuid);
	close(kvmfd);
	for (uint32_t i = 0; i < cpuid->nent; i++) {
		struct kvm_cpuid_entry2* entry = &cpuid->entries[i];
		if (entry->function == function && (!eax_bits || (entry->eax & eax_bits)) && (!ebx_bits || (entry->ebx & ebx_bits)) && (!ecx_bits || (entry->ecx & ecx_bits)) && (!edx_bits || (entry->edx & edx_bits)))
			return 1;
	}
	return 0;
}
#endif

#ifdef GOARCH_arm64
static int test_syzos()
{
	int mem_size = SYZ_PAGE_SIZE * 4;
	void* mem = mmap(0, mem_size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
	if (mem == MAP_FAILED) {
		printf("mmap failed (%d)\n", errno);
		return 1;
	}
	// Right now SyzOS testing just boils down to installing code into memory.
	install_syzos_code(mem, mem_size);
	munmap(mem, mem_size);
	return 0;
}
#endif