vmm.h revision 263780
1/*-
2 * Copyright (c) 2011 NetApp, Inc.
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 *
14 * THIS SOFTWARE IS PROVIDED BY NETAPP, INC ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED.  IN NO EVENT SHALL NETAPP, INC OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24 * SUCH DAMAGE.
25 *
26 * $FreeBSD: head/sys/amd64/include/vmm.h 263780 2014-03-26 23:34:27Z neel $
27 */
28
29#ifndef _VMM_H_
30#define	_VMM_H_
31
32#ifdef _KERNEL
33
34#define	VM_MAX_NAMELEN	32
35
36struct vm;
37struct vm_exception;
38struct vm_memory_segment;
39struct seg_desc;
40struct vm_exit;
41struct vm_run;
42struct vhpet;
43struct vioapic;
44struct vlapic;
45struct vmspace;
46struct vm_object;
47struct pmap;
48
49enum x2apic_state;
50
51typedef int	(*vmm_init_func_t)(int ipinum);
52typedef int	(*vmm_cleanup_func_t)(void);
53typedef void	(*vmm_resume_func_t)(void);
54typedef void *	(*vmi_init_func_t)(struct vm *vm, struct pmap *pmap);
55typedef int	(*vmi_run_func_t)(void *vmi, int vcpu, register_t rip,
56				  struct pmap *pmap, void *rendezvous_cookie,
57				  void *suspend_cookie);
58typedef void	(*vmi_cleanup_func_t)(void *vmi);
59typedef int	(*vmi_get_register_t)(void *vmi, int vcpu, int num,
60				      uint64_t *retval);
61typedef int	(*vmi_set_register_t)(void *vmi, int vcpu, int num,
62				      uint64_t val);
63typedef int	(*vmi_get_desc_t)(void *vmi, int vcpu, int num,
64				  struct seg_desc *desc);
65typedef int	(*vmi_set_desc_t)(void *vmi, int vcpu, int num,
66				  struct seg_desc *desc);
67typedef int	(*vmi_get_cap_t)(void *vmi, int vcpu, int num, int *retval);
68typedef int	(*vmi_set_cap_t)(void *vmi, int vcpu, int num, int val);
69typedef struct vmspace * (*vmi_vmspace_alloc)(vm_offset_t min, vm_offset_t max);
70typedef void	(*vmi_vmspace_free)(struct vmspace *vmspace);
71typedef struct vlapic * (*vmi_vlapic_init)(void *vmi, int vcpu);
72typedef void	(*vmi_vlapic_cleanup)(void *vmi, struct vlapic *vlapic);
73
74struct vmm_ops {
75	vmm_init_func_t		init;		/* module wide initialization */
76	vmm_cleanup_func_t	cleanup;
77	vmm_resume_func_t	resume;
78
79	vmi_init_func_t		vminit;		/* vm-specific initialization */
80	vmi_run_func_t		vmrun;
81	vmi_cleanup_func_t	vmcleanup;
82	vmi_get_register_t	vmgetreg;
83	vmi_set_register_t	vmsetreg;
84	vmi_get_desc_t		vmgetdesc;
85	vmi_set_desc_t		vmsetdesc;
86	vmi_get_cap_t		vmgetcap;
87	vmi_set_cap_t		vmsetcap;
88	vmi_vmspace_alloc	vmspace_alloc;
89	vmi_vmspace_free	vmspace_free;
90	vmi_vlapic_init		vlapic_init;
91	vmi_vlapic_cleanup	vlapic_cleanup;
92};
93
94extern struct vmm_ops vmm_ops_intel;
95extern struct vmm_ops vmm_ops_amd;
96
97int vm_create(const char *name, struct vm **retvm);
98void vm_destroy(struct vm *vm);
99const char *vm_name(struct vm *vm);
100int vm_malloc(struct vm *vm, vm_paddr_t gpa, size_t len);
101int vm_map_mmio(struct vm *vm, vm_paddr_t gpa, size_t len, vm_paddr_t hpa);
102int vm_unmap_mmio(struct vm *vm, vm_paddr_t gpa, size_t len);
103void *vm_gpa_hold(struct vm *, vm_paddr_t gpa, size_t len, int prot,
104		  void **cookie);
105void vm_gpa_release(void *cookie);
106int vm_gpabase2memseg(struct vm *vm, vm_paddr_t gpabase,
107	      struct vm_memory_segment *seg);
108int vm_get_memobj(struct vm *vm, vm_paddr_t gpa, size_t len,
109		  vm_offset_t *offset, struct vm_object **object);
110boolean_t vm_mem_allocated(struct vm *vm, vm_paddr_t gpa);
111int vm_get_register(struct vm *vm, int vcpu, int reg, uint64_t *retval);
112int vm_set_register(struct vm *vm, int vcpu, int reg, uint64_t val);
113int vm_get_seg_desc(struct vm *vm, int vcpu, int reg,
114		    struct seg_desc *ret_desc);
115int vm_set_seg_desc(struct vm *vm, int vcpu, int reg,
116		    struct seg_desc *desc);
117int vm_run(struct vm *vm, struct vm_run *vmrun);
118int vm_suspend(struct vm *vm);
119int vm_inject_nmi(struct vm *vm, int vcpu);
120int vm_nmi_pending(struct vm *vm, int vcpuid);
121void vm_nmi_clear(struct vm *vm, int vcpuid);
122int vm_inject_extint(struct vm *vm, int vcpu);
123int vm_extint_pending(struct vm *vm, int vcpuid);
124void vm_extint_clear(struct vm *vm, int vcpuid);
125uint64_t *vm_guest_msrs(struct vm *vm, int cpu);
126struct vlapic *vm_lapic(struct vm *vm, int cpu);
127struct vioapic *vm_ioapic(struct vm *vm);
128struct vhpet *vm_hpet(struct vm *vm);
129int vm_get_capability(struct vm *vm, int vcpu, int type, int *val);
130int vm_set_capability(struct vm *vm, int vcpu, int type, int val);
131int vm_get_x2apic_state(struct vm *vm, int vcpu, enum x2apic_state *state);
132int vm_set_x2apic_state(struct vm *vm, int vcpu, enum x2apic_state state);
133int vm_apicid2vcpuid(struct vm *vm, int apicid);
134void vm_activate_cpu(struct vm *vm, int vcpu);
135cpuset_t vm_active_cpus(struct vm *vm);
136struct vm_exit *vm_exitinfo(struct vm *vm, int vcpuid);
137
138/*
139 * Rendezvous all vcpus specified in 'dest' and execute 'func(arg)'.
140 * The rendezvous 'func(arg)' is not allowed to do anything that will
141 * cause the thread to be put to sleep.
142 *
143 * If the rendezvous is being initiated from a vcpu context then the
144 * 'vcpuid' must refer to that vcpu, otherwise it should be set to -1.
145 *
146 * The caller cannot hold any locks when initiating the rendezvous.
147 *
148 * The implementation of this API may cause vcpus other than those specified
149 * by 'dest' to be stalled. The caller should not rely on any vcpus making
150 * forward progress when the rendezvous is in progress.
151 */
152typedef void (*vm_rendezvous_func_t)(struct vm *vm, int vcpuid, void *arg);
153void vm_smp_rendezvous(struct vm *vm, int vcpuid, cpuset_t dest,
154    vm_rendezvous_func_t func, void *arg);
155
156static __inline int
157vcpu_rendezvous_pending(void *rendezvous_cookie)
158{
159
160	return (*(uintptr_t *)rendezvous_cookie != 0);
161}
162
163static __inline int
164vcpu_suspended(void *suspend_cookie)
165{
166
167	return (*(int *)suspend_cookie);
168}
169
170/*
171 * Return 1 if device indicated by bus/slot/func is supposed to be a
172 * pci passthrough device.
173 *
174 * Return 0 otherwise.
175 */
176int vmm_is_pptdev(int bus, int slot, int func);
177
178void *vm_iommu_domain(struct vm *vm);
179
180enum vcpu_state {
181	VCPU_IDLE,
182	VCPU_FROZEN,
183	VCPU_RUNNING,
184	VCPU_SLEEPING,
185};
186
187int vcpu_set_state(struct vm *vm, int vcpu, enum vcpu_state state,
188    bool from_idle);
189enum vcpu_state vcpu_get_state(struct vm *vm, int vcpu, int *hostcpu);
190
191static int __inline
192vcpu_is_running(struct vm *vm, int vcpu, int *hostcpu)
193{
194	return (vcpu_get_state(vm, vcpu, hostcpu) == VCPU_RUNNING);
195}
196
197void *vcpu_stats(struct vm *vm, int vcpu);
198void vcpu_notify_event(struct vm *vm, int vcpuid, bool lapic_intr);
199struct vmspace *vm_get_vmspace(struct vm *vm);
200int vm_assign_pptdev(struct vm *vm, int bus, int slot, int func);
201int vm_unassign_pptdev(struct vm *vm, int bus, int slot, int func);
202struct vatpic *vm_atpic(struct vm *vm);
203struct vatpit *vm_atpit(struct vm *vm);
204
205/*
206 * Inject exception 'vme' into the guest vcpu. This function returns 0 on
207 * success and non-zero on failure.
208 *
209 * Wrapper functions like 'vm_inject_gp()' should be preferred to calling
210 * this function directly because they enforce the trap-like or fault-like
211 * behavior of an exception.
212 *
213 * This function should only be called in the context of the thread that is
214 * executing this vcpu.
215 */
216int vm_inject_exception(struct vm *vm, int vcpuid, struct vm_exception *vme);
217
218/*
219 * Returns 0 if there is no exception pending for this vcpu. Returns 1 if an
220 * exception is pending and also updates 'vme'. The pending exception is
221 * cleared when this function returns.
222 *
223 * This function should only be called in the context of the thread that is
224 * executing this vcpu.
225 */
226int vm_exception_pending(struct vm *vm, int vcpuid, struct vm_exception *vme);
227
228void vm_inject_gp(struct vm *vm, int vcpuid); /* general protection fault */
229void vm_inject_ud(struct vm *vm, int vcpuid); /* undefined instruction fault */
230
231#endif	/* KERNEL */
232
233#include <machine/vmm_instruction_emul.h>
234
235#define	VM_MAXCPU	16			/* maximum virtual cpus */
236
237/*
238 * Identifiers for architecturally defined registers.
239 */
240enum vm_reg_name {
241	VM_REG_GUEST_RAX,
242	VM_REG_GUEST_RBX,
243	VM_REG_GUEST_RCX,
244	VM_REG_GUEST_RDX,
245	VM_REG_GUEST_RSI,
246	VM_REG_GUEST_RDI,
247	VM_REG_GUEST_RBP,
248	VM_REG_GUEST_R8,
249	VM_REG_GUEST_R9,
250	VM_REG_GUEST_R10,
251	VM_REG_GUEST_R11,
252	VM_REG_GUEST_R12,
253	VM_REG_GUEST_R13,
254	VM_REG_GUEST_R14,
255	VM_REG_GUEST_R15,
256	VM_REG_GUEST_CR0,
257	VM_REG_GUEST_CR3,
258	VM_REG_GUEST_CR4,
259	VM_REG_GUEST_DR7,
260	VM_REG_GUEST_RSP,
261	VM_REG_GUEST_RIP,
262	VM_REG_GUEST_RFLAGS,
263	VM_REG_GUEST_ES,
264	VM_REG_GUEST_CS,
265	VM_REG_GUEST_SS,
266	VM_REG_GUEST_DS,
267	VM_REG_GUEST_FS,
268	VM_REG_GUEST_GS,
269	VM_REG_GUEST_LDTR,
270	VM_REG_GUEST_TR,
271	VM_REG_GUEST_IDTR,
272	VM_REG_GUEST_GDTR,
273	VM_REG_GUEST_EFER,
274	VM_REG_LAST
275};
276
277/*
278 * Identifiers for optional vmm capabilities
279 */
280enum vm_cap_type {
281	VM_CAP_HALT_EXIT,
282	VM_CAP_MTRAP_EXIT,
283	VM_CAP_PAUSE_EXIT,
284	VM_CAP_UNRESTRICTED_GUEST,
285	VM_CAP_ENABLE_INVPCID,
286	VM_CAP_MAX
287};
288
289enum x2apic_state {
290	X2APIC_DISABLED,
291	X2APIC_ENABLED,
292	X2APIC_STATE_LAST
293};
294
295/*
296 * The 'access' field has the format specified in Table 21-2 of the Intel
297 * Architecture Manual vol 3b.
298 *
299 * XXX The contents of the 'access' field are architecturally defined except
300 * bit 16 - Segment Unusable.
301 */
302struct seg_desc {
303	uint64_t	base;
304	uint32_t	limit;
305	uint32_t	access;
306};
307
308enum vm_exitcode {
309	VM_EXITCODE_INOUT,
310	VM_EXITCODE_VMX,
311	VM_EXITCODE_BOGUS,
312	VM_EXITCODE_RDMSR,
313	VM_EXITCODE_WRMSR,
314	VM_EXITCODE_HLT,
315	VM_EXITCODE_MTRAP,
316	VM_EXITCODE_PAUSE,
317	VM_EXITCODE_PAGING,
318	VM_EXITCODE_INST_EMUL,
319	VM_EXITCODE_SPINUP_AP,
320	VM_EXITCODE_SPINDOWN_CPU,
321	VM_EXITCODE_RENDEZVOUS,
322	VM_EXITCODE_IOAPIC_EOI,
323	VM_EXITCODE_SUSPENDED,
324	VM_EXITCODE_MAX
325};
326
327struct vm_exit {
328	enum vm_exitcode	exitcode;
329	int			inst_length;	/* 0 means unknown */
330	uint64_t		rip;
331	union {
332		struct {
333			uint16_t	bytes:3;	/* 1 or 2 or 4 */
334			uint16_t	in:1;		/* out is 0, in is 1 */
335			uint16_t	string:1;
336			uint16_t	rep:1;
337			uint16_t	port;
338			uint32_t	eax;		/* valid for out */
339		} inout;
340		struct {
341			uint64_t	gpa;
342			int		fault_type;
343		} paging;
344		struct {
345			uint64_t	gpa;
346			uint64_t	gla;
347			uint64_t	cr3;
348			enum vie_cpu_mode cpu_mode;
349			enum vie_paging_mode paging_mode;
350			struct vie	vie;
351		} inst_emul;
352		/*
353		 * VMX specific payload. Used when there is no "better"
354		 * exitcode to represent the VM-exit.
355		 */
356		struct {
357			int		status;		/* vmx inst status */
358			/*
359			 * 'exit_reason' and 'exit_qualification' are valid
360			 * only if 'status' is zero.
361			 */
362			uint32_t	exit_reason;
363			uint64_t	exit_qualification;
364			/*
365			 * 'inst_error' and 'inst_type' are valid
366			 * only if 'status' is non-zero.
367			 */
368			int		inst_type;
369			int		inst_error;
370		} vmx;
371		struct {
372			uint32_t	code;		/* ecx value */
373			uint64_t	wval;
374		} msr;
375		struct {
376			int		vcpu;
377			uint64_t	rip;
378		} spinup_ap;
379		struct {
380			uint64_t	rflags;
381		} hlt;
382		struct {
383			int		vector;
384		} ioapic_eoi;
385	} u;
386};
387
388#endif	/* _VMM_H_ */
389