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