MFC

4 files changed, 623 insertions, 1 deletions

diff --git a/sys/amd64/include/intr_machdep.h b/sys/amd64/include/intr_machdep.h
index 700e35f..8671605 100644
--- a/sys/amd64/include/intr_machdep.h
+++ b/sys/amd64/include/intr_machdep.h
@@ -94,7 +94,7 @@ struct pic {
 	int (*pic_config_intr)(struct intsrc *, enum intr_trigger,
 	    enum intr_polarity);
 	int (*pic_assign_cpu)(struct intsrc *, u_int apic_id);
-	STAILQ_ENTRY(pic) pics;
+	TAILQ_ENTRY(pic) pics;
 };
 
 /* Flags for pic_disable_source() */
diff --git a/sys/amd64/include/vmm.h b/sys/amd64/include/vmm.h
new file mode 100644
index 0000000..ec94083
--- /dev/null
+++ b/sys/amd64/include/vmm.h
@@ -0,0 +1,293 @@
+/*-
+ * Copyright (c) 2011 NetApp, Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef _VMM_H_
+#define	_VMM_H_
+
+#ifdef _KERNEL
+
+#define	VM_MAX_NAMELEN	32
+
+struct vm;
+struct vm_memory_segment;
+struct seg_desc;
+struct vm_exit;
+struct vm_run;
+struct vlapic;
+
+enum x2apic_state;
+
+typedef int	(*vmm_init_func_t)(void);
+typedef int	(*vmm_cleanup_func_t)(void);
+typedef void *	(*vmi_init_func_t)(struct vm *vm); /* instance specific apis */
+typedef int	(*vmi_run_func_t)(void *vmi, int vcpu, register_t rip);
+typedef void	(*vmi_cleanup_func_t)(void *vmi);
+typedef int	(*vmi_mmap_set_func_t)(void *vmi, vm_paddr_t gpa,
+				       vm_paddr_t hpa, size_t length,
+				       vm_memattr_t attr, int prot,
+				       boolean_t superpages_ok);
+typedef vm_paddr_t (*vmi_mmap_get_func_t)(void *vmi, vm_paddr_t gpa);
+typedef int	(*vmi_get_register_t)(void *vmi, int vcpu, int num,
+				      uint64_t *retval);
+typedef int	(*vmi_set_register_t)(void *vmi, int vcpu, int num,
+				      uint64_t val);
+typedef int	(*vmi_get_desc_t)(void *vmi, int vcpu, int num,
+				  struct seg_desc *desc);
+typedef int	(*vmi_set_desc_t)(void *vmi, int vcpu, int num,
+				  struct seg_desc *desc);
+typedef int	(*vmi_inject_event_t)(void *vmi, int vcpu,
+				      int type, int vector,
+				      uint32_t code, int code_valid);
+typedef int	(*vmi_get_cap_t)(void *vmi, int vcpu, int num, int *retval);
+typedef int	(*vmi_set_cap_t)(void *vmi, int vcpu, int num, int val);
+
+struct vmm_ops {
+	vmm_init_func_t		init;		/* module wide initialization */
+	vmm_cleanup_func_t	cleanup;
+
+	vmi_init_func_t		vminit;		/* vm-specific initialization */
+	vmi_run_func_t		vmrun;
+	vmi_cleanup_func_t	vmcleanup;
+	vmi_mmap_set_func_t	vmmmap_set;
+	vmi_mmap_get_func_t	vmmmap_get;
+	vmi_get_register_t	vmgetreg;
+	vmi_set_register_t	vmsetreg;
+	vmi_get_desc_t		vmgetdesc;
+	vmi_set_desc_t		vmsetdesc;
+	vmi_inject_event_t	vminject;
+	vmi_get_cap_t		vmgetcap;
+	vmi_set_cap_t		vmsetcap;
+};
+
+extern struct vmm_ops vmm_ops_intel;
+extern struct vmm_ops vmm_ops_amd;
+
+struct vm *vm_create(const char *name);
+void vm_destroy(struct vm *vm);
+const char *vm_name(struct vm *vm);
+int vm_malloc(struct vm *vm, vm_paddr_t gpa, size_t len);
+int vm_map_mmio(struct vm *vm, vm_paddr_t gpa, size_t len, vm_paddr_t hpa);
+int vm_unmap_mmio(struct vm *vm, vm_paddr_t gpa, size_t len);
+vm_paddr_t vm_gpa2hpa(struct vm *vm, vm_paddr_t gpa, size_t size);
+int vm_gpabase2memseg(struct vm *vm, vm_paddr_t gpabase,
+	      struct vm_memory_segment *seg);
+int vm_get_register(struct vm *vm, int vcpu, int reg, uint64_t *retval);
+int vm_set_register(struct vm *vm, int vcpu, int reg, uint64_t val);
+int vm_get_seg_desc(struct vm *vm, int vcpu, int reg,
+		    struct seg_desc *ret_desc);
+int vm_set_seg_desc(struct vm *vm, int vcpu, int reg,
+		    struct seg_desc *desc);
+int vm_get_pinning(struct vm *vm, int vcpu, int *cpuid);
+int vm_set_pinning(struct vm *vm, int vcpu, int cpuid);
+int vm_run(struct vm *vm, struct vm_run *vmrun);
+int vm_inject_event(struct vm *vm, int vcpu, int type,
+		    int vector, uint32_t error_code, int error_code_valid);
+int vm_inject_nmi(struct vm *vm, int vcpu);
+int vm_nmi_pending(struct vm *vm, int vcpuid);
+void vm_nmi_clear(struct vm *vm, int vcpuid);
+uint64_t *vm_guest_msrs(struct vm *vm, int cpu);
+struct vlapic *vm_lapic(struct vm *vm, int cpu);
+int vm_get_capability(struct vm *vm, int vcpu, int type, int *val);
+int vm_set_capability(struct vm *vm, int vcpu, int type, int val);
+int vm_get_x2apic_state(struct vm *vm, int vcpu, enum x2apic_state *state);
+int vm_set_x2apic_state(struct vm *vm, int vcpu, enum x2apic_state state);
+void vm_activate_cpu(struct vm *vm, int vcpu);
+cpuset_t vm_active_cpus(struct vm *vm);
+struct vm_exit *vm_exitinfo(struct vm *vm, int vcpuid);
+
+/*
+ * Return 1 if device indicated by bus/slot/func is supposed to be a
+ * pci passthrough device.
+ *
+ * Return 0 otherwise.
+ */
+int vmm_is_pptdev(int bus, int slot, int func);
+
+void *vm_iommu_domain(struct vm *vm);
+
+enum vcpu_state {
+	VCPU_IDLE,
+	VCPU_RUNNING,
+	VCPU_CANNOT_RUN,
+};
+
+int vcpu_set_state(struct vm *vm, int vcpu, enum vcpu_state state);
+enum vcpu_state vcpu_get_state(struct vm *vm, int vcpu);
+
+static int __inline
+vcpu_is_running(struct vm *vm, int vcpu)
+{
+	return (vcpu_get_state(vm, vcpu) == VCPU_RUNNING);
+}
+
+void *vcpu_stats(struct vm *vm, int vcpu);
+void vm_interrupt_hostcpu(struct vm *vm, int vcpu);
+
+#endif	/* KERNEL */
+
+#include <machine/vmm_instruction_emul.h>
+
+#define	VM_MAXCPU	8			/* maximum virtual cpus */
+
+/*
+ * Identifiers for events that can be injected into the VM
+ */
+enum vm_event_type {
+	VM_EVENT_NONE,
+	VM_HW_INTR,
+	VM_NMI,
+	VM_HW_EXCEPTION,
+	VM_SW_INTR,
+	VM_PRIV_SW_EXCEPTION,
+	VM_SW_EXCEPTION,
+	VM_EVENT_MAX
+};
+
+/*
+ * Identifiers for architecturally defined registers.
+ */
+enum vm_reg_name {
+	VM_REG_GUEST_RAX,
+	VM_REG_GUEST_RBX,
+	VM_REG_GUEST_RCX,
+	VM_REG_GUEST_RDX,
+	VM_REG_GUEST_RSI,
+	VM_REG_GUEST_RDI,
+	VM_REG_GUEST_RBP,
+	VM_REG_GUEST_R8,
+	VM_REG_GUEST_R9,
+	VM_REG_GUEST_R10,
+	VM_REG_GUEST_R11,
+	VM_REG_GUEST_R12,
+	VM_REG_GUEST_R13,
+	VM_REG_GUEST_R14,
+	VM_REG_GUEST_R15,
+	VM_REG_GUEST_CR0,
+	VM_REG_GUEST_CR3,
+	VM_REG_GUEST_CR4,
+	VM_REG_GUEST_DR7,
+	VM_REG_GUEST_RSP,
+	VM_REG_GUEST_RIP,
+	VM_REG_GUEST_RFLAGS,
+	VM_REG_GUEST_ES,
+	VM_REG_GUEST_CS,
+	VM_REG_GUEST_SS,
+	VM_REG_GUEST_DS,
+	VM_REG_GUEST_FS,
+	VM_REG_GUEST_GS,
+	VM_REG_GUEST_LDTR,
+	VM_REG_GUEST_TR,
+	VM_REG_GUEST_IDTR,
+	VM_REG_GUEST_GDTR,
+	VM_REG_GUEST_EFER,
+	VM_REG_LAST
+};
+
+/*
+ * Identifiers for optional vmm capabilities
+ */
+enum vm_cap_type {
+	VM_CAP_HALT_EXIT,
+	VM_CAP_MTRAP_EXIT,
+	VM_CAP_PAUSE_EXIT,
+	VM_CAP_UNRESTRICTED_GUEST,
+	VM_CAP_MAX
+};
+
+enum x2apic_state {
+	X2APIC_ENABLED,
+	X2APIC_AVAILABLE,
+	X2APIC_DISABLED,
+	X2APIC_STATE_LAST
+};
+
+/*
+ * The 'access' field has the format specified in Table 21-2 of the Intel
+ * Architecture Manual vol 3b.
+ *
+ * XXX The contents of the 'access' field are architecturally defined except
+ * bit 16 - Segment Unusable.
+ */
+struct seg_desc {
+	uint64_t	base;
+	uint32_t	limit;
+	uint32_t	access;
+};
+
+enum vm_exitcode {
+	VM_EXITCODE_INOUT,
+	VM_EXITCODE_VMX,
+	VM_EXITCODE_BOGUS,
+	VM_EXITCODE_RDMSR,
+	VM_EXITCODE_WRMSR,
+	VM_EXITCODE_HLT,
+	VM_EXITCODE_MTRAP,
+	VM_EXITCODE_PAUSE,
+	VM_EXITCODE_PAGING,
+	VM_EXITCODE_SPINUP_AP,
+	VM_EXITCODE_MAX
+};
+
+struct vm_exit {
+	enum vm_exitcode	exitcode;
+	int			inst_length;	/* 0 means unknown */
+	uint64_t		rip;
+	union {
+		struct {
+			uint16_t	bytes:3;	/* 1 or 2 or 4 */
+			uint16_t	in:1;		/* out is 0, in is 1 */
+			uint16_t	string:1;
+			uint16_t	rep:1;
+			uint16_t	port;
+			uint32_t	eax;		/* valid for out */
+		} inout;
+		struct {
+			uint64_t	gpa;
+			struct vie	vie;
+		} paging;
+		/*
+		 * VMX specific payload. Used when there is no "better"
+		 * exitcode to represent the VM-exit.
+		 */
+		struct {
+			int		error;		/* vmx inst error */
+			uint32_t	exit_reason;
+			uint64_t	exit_qualification;
+		} vmx;
+		struct {
+			uint32_t	code;		/* ecx value */
+			uint64_t	wval;
+		} msr;
+		struct {
+			int		vcpu;
+			uint64_t	rip;
+		} spinup_ap;
+	} u;
+};
+
+#endif	/* _VMM_H_ */
diff --git a/sys/amd64/include/vmm_dev.h b/sys/amd64/include/vmm_dev.h
new file mode 100644
index 0000000..2311673
--- /dev/null
+++ b/sys/amd64/include/vmm_dev.h
@@ -0,0 +1,215 @@
+/*-
+ * Copyright (c) 2011 NetApp, Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef	_VMM_DEV_H_
+#define	_VMM_DEV_H_
+
+#ifdef _KERNEL
+void	vmmdev_init(void);
+int	vmmdev_cleanup(void);
+#endif
+
+struct vm_memory_segment {
+	vm_paddr_t	gpa;	/* in */
+	size_t		len;	/* in */
+};
+
+struct vm_register {
+	int		cpuid;
+	int		regnum;		/* enum vm_reg_name */
+	uint64_t	regval;
+};
+
+struct vm_seg_desc {			/* data or code segment */
+	int		cpuid;
+	int		regnum;		/* enum vm_reg_name */
+	struct seg_desc desc;
+};
+
+struct vm_pin {
+	int		vm_cpuid;
+	int		host_cpuid;	/* -1 to unpin */
+};
+
+struct vm_run {
+	int		cpuid;
+	uint64_t	rip;		/* start running here */
+	struct vm_exit	vm_exit;
+};
+
+struct vm_event {
+	int		cpuid;
+	enum vm_event_type type;
+	int		vector;
+	uint32_t	error_code;
+	int		error_code_valid;
+};
+
+struct vm_lapic_irq {
+	int		cpuid;
+	int		vector;
+};
+
+struct vm_capability {
+	int		cpuid;
+	enum vm_cap_type captype;
+	int		capval;
+	int		allcpus;
+};
+
+struct vm_pptdev {
+	int		bus;
+	int		slot;
+	int		func;
+};
+
+struct vm_pptdev_mmio {
+	int		bus;
+	int		slot;
+	int		func;
+	vm_paddr_t	gpa;
+	vm_paddr_t	hpa;
+	size_t		len;
+};
+
+struct vm_pptdev_msi {
+	int		vcpu;
+	int		bus;
+	int		slot;
+	int		func;
+	int		numvec;		/* 0 means disabled */
+	int		vector;
+	int		destcpu;
+};
+
+struct vm_pptdev_msix {
+	int		vcpu;
+	int		bus;
+	int		slot;
+	int		func;
+	int		idx;
+	uint32_t	msg;
+	uint32_t	vector_control;
+	uint64_t	addr;
+};
+
+struct vm_nmi {
+	int		cpuid;
+};
+
+#define	MAX_VM_STATS	64
+struct vm_stats {
+	int		cpuid;				/* in */
+	int		num_entries;			/* out */
+	struct timeval	tv;
+	uint64_t	statbuf[MAX_VM_STATS];
+};
+
+struct vm_stat_desc {
+	int		index;				/* in */
+	char		desc[128];			/* out */
+};
+
+struct vm_x2apic {
+	int			cpuid;
+	enum x2apic_state	state;
+};
+
+enum {
+	IOCNUM_RUN,
+	IOCNUM_SET_PINNING,
+	IOCNUM_GET_PINNING,
+	IOCNUM_MAP_MEMORY,
+	IOCNUM_GET_MEMORY_SEG,
+	IOCNUM_SET_REGISTER,
+	IOCNUM_GET_REGISTER,
+	IOCNUM_SET_SEGMENT_DESCRIPTOR,
+	IOCNUM_GET_SEGMENT_DESCRIPTOR,
+	IOCNUM_INJECT_EVENT,
+	IOCNUM_LAPIC_IRQ,
+	IOCNUM_SET_CAPABILITY,
+	IOCNUM_GET_CAPABILITY,
+	IOCNUM_BIND_PPTDEV,
+	IOCNUM_UNBIND_PPTDEV,
+	IOCNUM_MAP_PPTDEV_MMIO,
+	IOCNUM_PPTDEV_MSI,
+	IOCNUM_PPTDEV_MSIX,
+	IOCNUM_INJECT_NMI,
+	IOCNUM_VM_STATS,
+	IOCNUM_VM_STAT_DESC,
+	IOCNUM_SET_X2APIC_STATE,
+	IOCNUM_GET_X2APIC_STATE,
+};
+
+#define	VM_RUN		\
+	_IOWR('v', IOCNUM_RUN, struct vm_run)
+#define	VM_SET_PINNING	\
+	_IOW('v', IOCNUM_SET_PINNING, struct vm_pin)
+#define	VM_GET_PINNING	\
+	_IOWR('v', IOCNUM_GET_PINNING, struct vm_pin)
+#define	VM_MAP_MEMORY	\
+	_IOWR('v', IOCNUM_MAP_MEMORY, struct vm_memory_segment)
+#define	VM_GET_MEMORY_SEG \
+	_IOWR('v', IOCNUM_GET_MEMORY_SEG, struct vm_memory_segment)
+#define	VM_SET_REGISTER \
+	_IOW('v', IOCNUM_SET_REGISTER, struct vm_register)
+#define	VM_GET_REGISTER \
+	_IOWR('v', IOCNUM_GET_REGISTER, struct vm_register)
+#define	VM_SET_SEGMENT_DESCRIPTOR \
+	_IOW('v', IOCNUM_SET_SEGMENT_DESCRIPTOR, struct vm_seg_desc)
+#define	VM_GET_SEGMENT_DESCRIPTOR \
+	_IOWR('v', IOCNUM_GET_SEGMENT_DESCRIPTOR, struct vm_seg_desc)
+#define	VM_INJECT_EVENT	\
+	_IOW('v', IOCNUM_INJECT_EVENT, struct vm_event)
+#define	VM_LAPIC_IRQ 		\
+	_IOW('v', IOCNUM_LAPIC_IRQ, struct vm_lapic_irq)
+#define	VM_SET_CAPABILITY \
+	_IOW('v', IOCNUM_SET_CAPABILITY, struct vm_capability)
+#define	VM_GET_CAPABILITY \
+	_IOWR('v', IOCNUM_GET_CAPABILITY, struct vm_capability)
+#define	VM_BIND_PPTDEV \
+	_IOW('v', IOCNUM_BIND_PPTDEV, struct vm_pptdev)
+#define	VM_UNBIND_PPTDEV \
+	_IOW('v', IOCNUM_UNBIND_PPTDEV, struct vm_pptdev)
+#define	VM_MAP_PPTDEV_MMIO \
+	_IOW('v', IOCNUM_MAP_PPTDEV_MMIO, struct vm_pptdev_mmio)
+#define	VM_PPTDEV_MSI \
+	_IOW('v', IOCNUM_PPTDEV_MSI, struct vm_pptdev_msi)
+#define	VM_PPTDEV_MSIX \
+	_IOW('v', IOCNUM_PPTDEV_MSIX, struct vm_pptdev_msix)
+#define VM_INJECT_NMI \
+	_IOW('v', IOCNUM_INJECT_NMI, struct vm_nmi)
+#define	VM_STATS \
+	_IOWR('v', IOCNUM_VM_STATS, struct vm_stats)
+#define	VM_STAT_DESC \
+	_IOWR('v', IOCNUM_VM_STAT_DESC, struct vm_stat_desc)
+#define	VM_SET_X2APIC_STATE \
+	_IOW('v', IOCNUM_SET_X2APIC_STATE, struct vm_x2apic)
+#define	VM_GET_X2APIC_STATE \
+	_IOWR('v', IOCNUM_GET_X2APIC_STATE, struct vm_x2apic)
+#endif
diff --git a/sys/amd64/include/vmm_instruction_emul.h b/sys/amd64/include/vmm_instruction_emul.h
new file mode 100644
index 0000000..4c7a346
--- /dev/null
+++ b/sys/amd64/include/vmm_instruction_emul.h
@@ -0,0 +1,114 @@
+/*-
+ * Copyright (c) 2012 NetApp, Inc.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef	_VMM_INSTRUCTION_EMUL_H_
+#define	_VMM_INSTRUCTION_EMUL_H_
+
+/*
+ * The data structures 'vie' and 'vie_op' are meant to be opaque to the
+ * consumers of instruction decoding. The only reason why their contents
+ * need to be exposed is because they are part of the 'vm_exit' structure.
+ */
+struct vie_op {
+	uint8_t		op_byte;	/* actual opcode byte */
+	uint8_t		op_type;	/* type of operation (e.g. MOV) */
+	uint16_t	op_flags;
+};
+
+#define	VIE_INST_SIZE	15
+struct vie {
+	uint8_t		inst[VIE_INST_SIZE];	/* instruction bytes */
+	uint8_t		num_valid;		/* size of the instruction */
+	uint8_t		num_processed;
+
+	uint8_t		rex_w:1,		/* REX prefix */
+			rex_r:1,
+			rex_x:1,
+			rex_b:1,
+			rex_present:1;
+
+	uint8_t		mod:2,			/* ModRM byte */
+			reg:4,
+			rm:4;
+
+	uint8_t		ss:2,			/* SIB byte */
+			index:4,
+			base:4;
+
+	uint8_t		disp_bytes;
+	uint8_t		imm_bytes;
+
+	uint8_t		scale;
+	int		base_register;		/* VM_REG_GUEST_xyz */
+	int		index_register;		/* VM_REG_GUEST_xyz */
+
+	int64_t		displacement;		/* optional addr displacement */
+	int64_t		immediate;		/* optional immediate operand */
+
+	uint8_t		decoded;	/* set to 1 if successfully decoded */
+
+	struct vie_op	op;			/* opcode description */
+};
+
+/*
+ * Callback functions to read and write memory regions.
+ */
+typedef int (*mem_region_read_t)(void *vm, int cpuid, uint64_t gpa,
+				 uint64_t *rval, int rsize, void *arg);
+
+typedef int (*mem_region_write_t)(void *vm, int cpuid, uint64_t gpa,
+				  uint64_t wval, int wsize, void *arg);
+
+/*
+ * Emulate the decoded 'vie' instruction.
+ *
+ * The callbacks 'mrr' and 'mrw' emulate reads and writes to the memory region
+ * containing 'gpa'. 'mrarg' is an opaque argument that is passed into the
+ * callback functions.
+ *
+ * 'void *vm' should be 'struct vm *' when called from kernel context and
+ * 'struct vmctx *' when called from user context.
+ * s
+ */
+int vmm_emulate_instruction(void *vm, int cpuid, uint64_t gpa, struct vie *vie,
+			    mem_region_read_t mrr, mem_region_write_t mrw,
+			    void *mrarg);
+
+#ifdef _KERNEL
+/*
+ * APIs to fetch and decode the instruction from nested page fault handler.
+ */
+int vmm_fetch_instruction(struct vm *vm, int cpuid,
+			  uint64_t rip, int inst_length, uint64_t cr3,
+			  struct vie *vie);
+
+int vmm_decode_instruction(struct vm *vm, int cpuid,
+			   uint64_t gla, struct vie *vie);
+#endif	/* _KERNEL */
+
+#endif	/* _VMM_INSTRUCTION_EMUL_H_ */