-
Notifications
You must be signed in to change notification settings - Fork 18
/
Copy pathvmm.c
1908 lines (1590 loc) · 43.8 KB
/
vmm.c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/*-
* Copyright (c) 2011 NetApp, Inc.
* Copyright (c) 2015 xhyve developers
* All rights reserved.
*
*
6DB6
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.
*
* $FreeBSD$
*/
#include <stdint.h>
#include <stdbool.h>
#include <string.h>
#include <errno.h>
#include <pthread.h>
#include <assert.h>
#include <libkern/OSAtomic.h>
#include <xhyve/support/misc.h>
#include <xhyve/support/atomic.h>
#include <xhyve/support/cpuset.h>
#include <xhyve/support/psl.h>
#include <xhyve/support/specialreg.h>
#include <xhyve/support/apicreg.h>
#include <xhyve/vmm/vmm.h>
#include <xhyve/vmm/vmm_lapic.h>
#include <xhyve/vmm/vmm_mem.h>
#include <xhyve/vmm/vmm_ioport.h>
#include <xhyve/vmm/vmm_instruction_emul.h>
#include <xhyve/vmm/vmm_callout.h>
#include <xhyve/vmm/vmm_host.h>
#include <xhyve/vmm/vmm_stat.h>
#include <xhyve/vmm/vmm_ktr.h>
#include <xhyve/vmm/io/vatpic.h>
#include <xhyve/vmm/io/vatpit.h>
#include <xhyve/vmm/io/vioapic.h>
#include <xhyve/vmm/io/vlapic.h>
#include <xhyve/vmm/io/vhpet.h>
#include <xhyve/vmm/io/vpmtmr.h>
#include <xhyve/vmm/io/vrtc.h>
struct vlapic;
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Wpadded"
/*
* Initialization:
* (a) allocated when vcpu is created
* (i) initialized when vcpu is created and when it is reinitialized
* (o) initialized the first time the vcpu is created
* (x) initialized before use
*/
struct vcpu {
OSSpinLock lock; /* (o) protects 'state' */
pthread_mutex_t state_sleep_mtx;
pthread_cond_t state_sleep_cnd;
pthread_mutex_t vcpu_sleep_mtx;
pthread_cond_t vcpu_sleep_cnd;
enum vcpu_state state; /* (o) vcpu state */
struct vlapic *vlapic; /* (i) APIC device model */
enum x2apic_state x2apic_state; /* (i) APIC mode */
uint64_t exitintinfo; /* (i) events pending at VM exit */
int nmi_pending; /* (i) NMI pending */
int extint_pending; /* (i) INTR pending */
int exception_pending; /* (i) exception pending */
int exc_vector; /* (x) exception collateral */
int exc_errcode_valid;
uint32_t exc_errcode;
uint64_t guest_xcr0; /* (i) guest %xcr0 register */
void *stats; /* (a,i) statistics */
struct vm_exit exitinfo; /* (x) exit reason and collateral */
uint64_t nextrip; /* (x) next instruction to execute */
};
#define vcpu_lock_init(v) (v)->lock = OS_SPINLOCK_INIT;
#define vcpu_lock(v) OSSpinLockLock(&(v)->lock)
#define vcpu_unlock(v) OSSpinLockUnlock(&(v)->lock)
struct mem_seg {
uint64_t gpa;
size_t len;
void *object;
};
#define VM_MAX_MEMORY_SEGMENTS 2
/*
* Initialization:
* (o) initialized the first time the VM is created
* (i) initialized when VM is created and when it is reinitialized
* (x) initialized before use
*/
struct vm {
void *cookie; /* (i) cpu-specific data */
struct vhpet *vhpet; /* (i) virtual HPET */
struct vioapic *vioapic; /* (i) virtual ioapic */
struct vatpic *vatpic; /* (i) virtual atpic */
struct vatpit *vatpit; /* (i) virtual atpit */
struct vpmtmr *vpmtmr; /* (i) virtual ACPI PM timer */
struct vrtc *vrtc; /* (o) virtual RTC */
volatile cpuset_t active_cpus; /* (i) active vcpus */
int suspend; /* (i) stop VM execution */
volatile cpuset_t suspended_cpus; /* (i) suspended vcpus */
volatile cpuset_t halted_cpus; /* (x) cpus in a hard halt */
cpuset_t rendezvous_req_cpus; /* (x) rendezvous requested */
cpuset_t rendezvous_done_cpus; /* (x) rendezvous finished */
void *rendezvous_arg; /* (x) rendezvous func/arg */
vm_rendezvous_func_t rendezvous_func;
pthread_mutex_t rendezvous_mtx; /* (o) rendezvous lock */
pthread_cond_t rendezvous_sleep_cnd;
int num_mem_segs; /* (o) guest memory segments */
struct mem_seg mem_segs[VM_MAX_MEMORY_SEGMENTS];
struct vcpu vcpu[VM_MAXCPU]; /* (i) guest vcpus */
};
#pragma clang diagnostic pop
static int vmm_initialized;
static struct vmm_ops *ops;
#define VMM_INIT() \
(*ops->init)()
#define VMM_CLEANUP() \
(*ops->cleanup)()
#define VM_INIT(vmi) \
(*ops->vm_init)(vmi)
#define VCPU_INIT(vmi, vcpu) \
(*ops->vcpu_init)(vmi, vcpu)
#define VMRUN(vmi, vcpu, rip, rptr, sptr) \
(*ops->vmrun)(vmi, vcpu, rip, rptr, sptr)
#define VM_CLEANUP(vmi) \
(*ops->vm_cleanup)(vmi)
#define VCPU_CLEANUP(vmi, vcpu) \
(*ops->vcpu_cleanup)(vmi, vcpu)
#define VMGETREG(vmi, vcpu, num, retval) \
(*ops->vmgetreg)(vmi, vcpu, num, retval)
#define VMSETREG(vmi, vcpu, num, val) \
(*ops->vmsetreg)(vmi, vcpu, num, val)
#define VMGETDESC(vmi, vcpu, num, desc) \
(*ops->vmgetdesc)(vmi, vcpu, num, desc)
#define VMSETDESC(vmi, vcpu, num, desc) \
(*ops->vmsetdesc)(vmi, vcpu, num, desc)
#define VMGETCAP(vmi, vcpu, num, retval) \
(*ops->vmgetcap)(vmi, vcpu, num, retval)
#define VMSETCAP(vmi, vcpu, num, val) \
(*ops->vmsetcap)(vmi, vcpu, num, val)
#define VLAPIC_INIT(vmi, vcpu) \
(*ops->vlapic_init)(vmi, vcpu)
#define VLAPIC_CLEANUP(vmi, vlapic) \
(*ops->vlapic_cleanup)(vmi, vlapic)
#define VCPU_INTERRUPT(vcpu) \
(*ops->vcpu_interrupt)(vcpu)
/* statistics */
//static VMM_STAT(VCPU_TOTAL_RUNTIME, "vcpu total runtime");
/*
* Halt the guest if all vcpus are executing a HLT instruction with
* interrupts disabled.
*/
static int halt_detection_enabled = 1;
static int trace_guest_exceptions = 0;
static void
vcpu_cleanup(struct vm *vm, int i, bool destroy)
{
struct vcpu *vcpu = &vm->vcpu[i];
VLAPIC_CLEANUP(vm->cookie, vcpu->vlapic);
if (destroy) {
vmm_stat_free(vcpu->stats);
}
}
static void
vcpu_init(struct vm *vm, int vcpu_id, bool create)
{
struct vcpu *vcpu;
KASSERT(vcpu_id >= 0 && vcpu_id < VM_MAXCPU,
("vcpu_init: invalid vcpu %d", vcpu_id));
vcpu = &vm->vcpu[vcpu_id];
if (create) {
vcpu_lock_init(vcpu);
pthread_mutex_init(&vcpu->state_sleep_mtx, NULL);
pthread_cond_init(&vcpu->state_sleep_cnd, NULL);
pthread_mutex_init(&vcpu->vcpu_sleep_mtx, NULL);
pthread_cond_init(&vcpu->vcpu_sleep_cnd, NULL);
vcpu->state = VCPU_IDLE;
vcpu->stats = vmm_stat_alloc();
}
vcpu->vlapic = VLAPIC_INIT(vm->cookie, vcpu_id);
vm_set_x2apic_state(vm, vcpu_id, X2APIC_DISABLED);
vcpu->exitintinfo = 0;
vcpu->nmi_pending = 0;
vcpu->extint_pending = 0;
vcpu->exception_pending = 0;
vcpu->guest_xcr0 = XFEATURE_ENABLED_X87;
vmm_stat_init(vcpu->stats);
}
int vcpu_create(struct vm *vm, int vcpu) {
if (vcpu < 0 || vcpu >= VM_MAXCPU)
xhyve_abort("vcpu_create: invalid cpuid %d\n", vcpu);
return VCPU_INIT(vm->cookie, vcpu);
}
void vcpu_destroy(struct vm *vm, int vcpu) {
if (vcpu < 0 || vcpu >= VM_MAXCPU)
xhyve_abort("vcpu_destroy: invalid cpuid %d\n", vcpu);
VCPU_CLEANUP(vm, vcpu);
}
int
vcpu_trace_exceptions(void)
{
return (trace_guest_exceptions);
}
struct vm_exit *
vm_exitinfo(struct vm *vm, int cpuid)
{
struct vcpu *vcpu;
if (cpuid < 0 || cpuid >= VM_MAXCPU)
xhyve_abort("vm_exitinfo: invalid cpuid %d\n", cpuid);
vcpu = &vm->vcpu[cpuid];
return (&vcpu->exitinfo);
}
int
vmm_init(void)
{
int error;
vmm_host_state_init();
error = vmm_mem_init();
if (error)
return (error);
ops = &vmm_ops_intel;
error = VMM_INIT();
if (error == 0)
vmm_initialized = 1;
return (error);
}
int
vmm_cleanup(void) {
int error;
error = VMM_CLEANUP();
if (error == 0)
vmm_initialized = 0;
return error;
}
static void
vm_init(struct vm *vm, bool create)
{
int vcpu;
if (create) {
callout_system_init();
}
vm->cookie = VM_INIT(vm);
vm->vioapic = vioapic_init(vm);
vm->vhpet = vhpet_init(vm);
vm->vatpic = vatpic_init(vm);
vm->vatpit = vatpit_init(vm);
vm->vpmtmr = vpmtmr_init(vm);
if (create) {
vm->vrtc = vrtc_init(vm);
}
CPU_ZERO(&vm->active_cpus);
vm->suspend = 0;
CPU_ZERO(&vm->suspended_cpus);
for (vcpu = 0; vcpu < VM_MAXCPU; vcpu++) {
vcpu_init(vm, vcpu, create);
}
}
int
vm_create(struct vm **retvm)
{
struct vm *vm;
if (!vmm_initialized)
return (ENXIO);
vm = malloc(sizeof(struct vm));
assert(vm);
bzero(vm, sizeof(struct vm));
vm->num_mem_segs = 0;
pthread_mutex_init(&vm->rendezvous_mtx, NULL);
pthread_cond_init(&vm->rendezvous_sleep_cnd, NULL);
vm_init(vm, true);
*retvm = vm;
return (0);
}
static void
vm_free_mem_seg(struct mem_seg *seg)
{
if (seg->object != NULL) {
vmm_mem_free(seg->gpa, seg->len, seg->object);
}
bzero(seg, sizeof(*seg));
}
static void
vm_cleanup(struct vm *vm, bool destroy)
{
int i, vcpu;
for (vcpu = 0; vcpu < VM_MAXCPU; vcpu++) {
vcpu_cleanup(vm, vcpu, destroy);
}
if (destroy) {
vrtc_cleanup(vm->vrtc);
} else {
vrtc_reset(vm->vrtc);
}
vpmtmr_cleanup(vm->vpmtmr);
vatpit_cleanup(vm->vatpit);
vhpet_cleanup(vm->vhpet);
vatpic_cleanup(vm->vatpic);
vioapic_cleanup(vm->vioapic);
VM_CLEANUP(vm->cookie);
if (destroy) {
for (i = 0; i < vm->num_mem_segs; i++) {
vm_free_mem_seg(&vm->mem_segs[i]);
}
vm->num_mem_segs = 0;
}
}
void
vm_destroy(struct vm *vm)
{
vm_cleanup(vm, true);
free(vm);
}
int
vm_reinit(struct vm *vm)
{
int error;
/*
* A virtual machine can be reset only if all vcpus are suspended.
*/
if (CPU_CMP(&vm->suspended_cpus, &vm->active_cpus) == 0) {
vm_cleanup(vm, false);
vm_init(vm, false);
error = 0;
} else {
error = EBUSY;
}
return (error);
}
const char *
vm_name(UNUSED struct vm *vm)
{
return "VM";
}
bool
vm_mem_allocated(struct vm *vm, uint64_t gpa)
{
int i;
uint64_t gpabase, gpalimit;
for (i = 0; i < vm->num_mem_segs; i++) {
gpabase = vm->mem_segs[i].gpa;
gpalimit = gpabase + vm->mem_segs[i].len;
if (gpa >= gpabase && gpa < gpalimit)
return (TRUE); /* 'gpa' is regular memory */
}
return (FALSE);
}
int
vm_malloc(struct vm *vm, uint64_t gpa, size_t len)
{
int available, allocated;
struct mem_seg *seg;
void *object;
uint64_t g;
if ((gpa & XHYVE_PAGE_MASK) || (len & XHYVE_PAGE_MASK) || len == 0)
return (EINVAL);
available = allocated = 0;
g = gpa;
while (g < gpa + len) {
if (vm_mem_allocated(vm, g))
allocated++;
else
available++;
g += XHYVE_PAGE_SIZE;
}
/*
* If there are some allocated and some available pages in the address
* range then it is an error.
*/
if (allocated && available)
return (EINVAL);
/*
* If the entire address range being requested has already been
* allocated then there isn't anything more to do.
*/
if (allocated && available == 0)
return (0);
if (vm->num_mem_segs >= VM_MAX_MEMORY_SEGMENTS)
return (E2BIG);
seg = &vm->mem_segs[vm->num_mem_segs];
if ((object = vmm_mem_alloc(gpa, len)) == NULL)
return (ENOMEM);
seg->gpa = gpa;
seg->len = len;
seg->object = object;
vm->num_mem_segs++;
return (0);
}
void *
vm_gpa2hva(struct vm *vm, uint64_t gpa, uint64_t len) {
void *base;
uint64_t offset;
if (vm_get_memobj(vm, gpa, len, &offset, &base)) {
return NULL;
}
return (void *) (((uintptr_t) base) + offset);
}
int
vm_gpabase2memseg(struct vm *vm, uint64_t gpabase,
struct vm_memory_segment *seg)
{
int i;
for (i = 0; i < vm->num_mem_segs; i++) {
if (gpabase == vm->mem_segs[i].gpa) {
seg->gpa = vm->mem_segs[i].gpa;
seg->len = vm->mem_segs[i].len;
return (0);
}
}
return (-1);
}
int
vm_get_memobj(struct vm *vm, uint64_t gpa, size_t len,
uint64_t *offset, void **object)
{
int i;
size_t seg_len;
uint64_t seg_gpa;
void *seg_obj;
for (i = 0; i < vm->num_mem_segs; i++) {
if ((seg_obj = vm->mem_segs[i].object) == NULL)
continue;
seg_gpa = vm->mem_segs[i].gpa;
seg_len = vm->mem_segs[i].len;
if ((gpa >= seg_gpa) && ((gpa + len) <= (seg_gpa + seg_len))) {
*offset = gpa - seg_gpa;
*object = seg_obj;
return (0);
}
}
return (EINVAL);
}
int
vm_get_register(struct vm *vm, int vcpu, int reg, uint64_t *retval)
{
if (vcpu < 0 || vcpu >= VM_MAXCPU)
return (EINVAL);
if (reg >= VM_REG_LAST)
return (EINVAL);
return (VMGETREG(vm->cookie, vcpu, reg, retval));
}
int
vm_set_register(struct vm *vm, int vcpuid, int reg, uint64_t val)
{
struct vcpu *vcpu;
int error;
if (vcpuid < 0 || vcpuid >= VM_MAXCPU)
return (EINVAL);
if (reg >= VM_REG_LAST)
return (EINVAL);
error = VMSETREG(vm->cookie, vcpuid, reg, val);
if (error || reg != VM_REG_GUEST_RIP)
return (error);
/* Set 'nextrip' to match the value of %rip */
VCPU_CTR1(vm, vcpuid, "Setting nextrip to %#llx", val);
vcpu = &vm->vcpu[vcpuid];
vcpu->nextrip = val;
return (0);
}
static bool
is_descriptor_table(int reg)
{
switch (reg) {
case VM_REG_GUEST_IDTR:
case VM_REG_GUEST_GDTR:
return (TRUE);
default:
return (FALSE);
}
}
static bool
is_segment_register(int reg)
{
switch (reg) {
case VM_REG_GUEST_ES:
case VM_REG_GUEST_CS:
case VM_REG_GUEST_SS:
case VM_REG_GUEST_DS:
case VM_REG_GUEST_FS:
case VM_REG_GUEST_GS:
case VM_REG_GUEST_TR:
case VM_REG_GUEST_LDTR:
return (TRUE);
default:
return (FALSE);
}
}
int
vm_get_seg_desc(struct vm *vm, int vcpu, int reg,
struct seg_desc *desc)
{
if (vcpu < 0 || vcpu >= VM_MAXCPU)
return (EINVAL);
if (!is_segment_register(reg) && !is_descriptor_table(reg))
return (EINVAL);
return (VMGETDESC(vm->cookie, vcpu, reg, desc));
}
int
vm_set_seg_desc(struct vm *vm, int vcpu, int reg,
struct seg_desc *desc)
{
if (vcpu < 0 || vcpu >= VM_MAXCPU)
return (EINVAL);
if (!is_segment_register(reg) && !is_descriptor_table(reg))
return (EINVAL);
return (VMSETDESC(vm->cookie, vcpu, reg, desc));
}
// static VMM_STAT(VCPU_IDLE_TICKS, "number of ticks vcpu was idle");
static int
vcpu_set_state_locked(struct vcpu *vcpu, enum vcpu_state newstate,
bool from_idle)
{
int error;
const struct timespec ts = {.tv_sec = 1, .tv_nsec = 0}; /* 1 second */
/*
* State transitions from the vmmdev_ioctl() must always begin from
* the VCPU_IDLE state. This guarantees that there is only a single
* ioctl() operating on a vcpu at any point.
*/
if (from_idle) {
while (vcpu->state != VCPU_IDLE) {
pthread_mutex_lock(&vcpu->state_sleep_mtx);
vcpu_unlock(vcpu);
pthread_cond_timedwait_relative_np(&vcpu->state_sleep_cnd,
&vcpu->state_sleep_mtx, &ts);
vcpu_lock(vcpu);
pthread_mutex_unlock(&vcpu->state_sleep_mtx);
//msleep_spin(&vcpu->state, &vcpu->mtx, "vmstat", hz);
}
} else {
KASSERT(vcpu->state != VCPU_IDLE, ("invalid transition from "
"vcpu idle state"));
}
/*
* The following state transitions are allowed:
* IDLE -> FROZEN -> IDLE
* FROZEN -> RUNNING -> FROZEN
* FROZEN -> SLEEPING -> FROZEN
*/
switch (vcpu->state) {
case VCPU_IDLE:
case VCPU_RUNNING:
case VCPU_SLEEPING:
error = (newstate != VCPU_FROZEN);
break;
case VCPU_FROZEN:
error = (newstate == VCPU_FROZEN);
break;
}
if (error)
return (EBUSY);
vcpu->state = newstate;
if (newstate == VCPU_IDLE)
pthread_cond_broadcast(&vcpu->state_sleep_cnd);
//wakeup(&vcpu->state);
return (0);
}
static void
vcpu_require_state(struct vm *vm, int vcpuid, enum vcpu_state newstate)
{
int error;
if ((error = vcpu_set_state(vm, vcpuid, newstate, false)) != 0)
xhyve_abort("Error %d setting state to %d\n", error, newstate);
}
static void
vcpu_require_state_locked(struct vcpu *vcpu, enum vcpu_state newstate)
{
int error;
if ((error = vcpu_set_state_locked(vcpu, newstate, false)) != 0)
xhyve_abort("Error %d setting state to %d", error, newstate);
}
static void
vm_set_rendezvous_func(struct vm *vm, vm_rendezvous_func_t func)
{
/*
* Update 'rendezvous_func' and execute a write memory barrier to
* ensure that it is visible across all host cpus. This is not needed
* for correctness but it does ensure that all the vcpus will notice
* that the rendezvous is requested immediately.
*/
vm->rendezvous_func = func;
wmb();
}
#define RENDEZVOUS_CTR0(vm, vcpuid, fmt) \
do { \
if (vcpuid >= 0) {\
VCPU_CTR0(vm, vcpuid, fmt); \
} else {\
VM_CTR0(vm, fmt); \
} \
} while (0)
static void
vm_handle_rendezvous(struct vm *vm, int vcpuid)
{
KASSERT(vcpuid == -1 || (vcpuid >= 0 && vcpuid < VM_MAXCPU),
("vm_handle_rendezvous: invalid vcpuid %d", vcpuid));
pthread_mutex_lock(&vm->rendezvous_mtx);
while (vm->rendezvous_func != NULL) {
/* 'rendezvous_req_cpus' must be a subset of 'active_cpus' */
CPU_AND(&vm->rendezvous_req_cpus, &vm->active_cpus);
if (vcpuid != -1 &&
CPU_ISSET(((unsigned) vcpuid), &vm->rendezvous_req_cpus) &&
!CPU_ISSET(((unsigned) vcpuid), &vm->rendezvous_done_cpus)) {
VCPU_CTR0(vm, vcpuid, "Calling rendezvous func");
(*vm->rendezvous_func)(vm, vcpuid, vm->rendezvous_arg);
CPU_SET(((unsigned) vcpuid), &vm->rendezvous_done_cpus);
}
if (CPU_CMP(&vm->rendezvous_req_cpus,
&vm->rendezvous_done_cpus) == 0) {
VCPU_CTR0(vm, vcpuid, "Rendezvous completed");
vm_set_rendezvous_func(vm, NULL);
pthread_cond_broadcast(&vm->rendezvous_sleep_cnd);
//wakeup(&vm->rendezvous_func);
break;
}
RENDEZVOUS_CTR0(vm, vcpuid, "Wait for rendezvous completion");
pthread_cond_wait(&vm->rendezvous_sleep_cnd, &vm->rendezvous_mtx);
//mtx_sleep(&vm->rendezvous_func, &vm->rendezvous_mtx, 0, "vmrndv", 0);
}
pthread_mutex_unlock(&vm->rendezvous_mtx);
}
/*
* Emulate a guest 'hlt' by sleeping until the vcpu is ready to run.
*/
static int
vm_handle_hlt(struct vm *vm, int vcpuid, bool intr_disabled)
{
struct vcpu *vcpu;
const char *wmesg;
int vcpu_halted, vm_halted;
const struct timespec ts = {.tv_sec = 1, .tv_nsec = 0}; /* 1 second */
KASSERT(!CPU_ISSET(((unsigned) vcpuid), &vm->halted_cpus),
("vcpu already halted"));
vcpu = &vm->vcpu[vcpuid];
vcpu_halted = 0;
vm_halted = 0;
vcpu_lock(vcpu);
while (1) {
/*
* Do a final check for pending NMI or interrupts before
* really putting this thread to sleep. Also check for
* software events that would cause this vcpu to wakeup.
*
* These interrupts/events could have happened after the
* vcpu returned from VMRUN() and before it acquired the
* vcpu lock above.
*/
if (vm->rendezvous_func != NULL || vm->suspend)
break;
if (vm_nmi_pending(vm, vcpuid))
break;
if (!intr_disabled) {
if (vm_extint_pending(vm, vcpuid) ||
vlapic_pending_intr(vcpu->vlapic, NULL)) {
break;
}
}
/*
* Some Linux guests implement "halt" by having all vcpus
* execute HLT with interrupts disabled. 'halted_cpus' keeps
* track of the vcpus that have entered this state. When all
* vcpus enter the halted state the virtual machine is halted.
*/
if (intr_disabled) {
wmesg = "vmhalt";
VCPU_CTR0(vm, vcpuid, "Halted");
if (!vcpu_halted && halt_detection_enabled) {
vcpu_halted = 1;
CPU_SET_ATOMIC(((unsigned) vcpuid), &vm->halted_cpus);
}
if (CPU_CMP(&vm->halted_cpus, &vm->active_cpus) == 0) {
vm_halted = 1;
break;
}
} else {
wmesg = "vmidle";
}
//t = ticks;
vcpu_require_state_locked(vcpu, VCPU_SLEEPING);
/*
* XXX msleep_spin() cannot be interrupted by signals so
* wake up periodically to check pending signals.
*/
pthread_mutex_lock(&vcpu->vcpu_sleep_mtx);
vcpu_unlock(vcpu);
pthread_cond_timedwait_relative_np(&vcpu->vcpu_sleep_cnd,
&vcpu->vcpu_sleep_mtx, &ts);
vcpu_lock(vcpu);
pthread_mutex_unlock(&vcpu->vcpu_sleep_mtx);
//msleep_spin(vcpu, &vcpu->mtx, wmesg, hz);
vcpu_require_state_locked(vcpu, VCPU_FROZEN);
//vmm_stat_incr(vm, vcpuid, VCPU_IDLE_TICKS, ticks - t);
}
if (vcpu_halted)
CPU_CLR_ATOMIC(((unsigned) vcpuid), &vm->halted_cpus);
vcpu_unlock(vcpu);
if (vm_halted)
vm_suspend(vm, VM_SUSPEND_HALT);
return (0);
}
static int
vm_handle_inst_emul(struct vm *vm, int vcpuid, bool *retu)
{
struct vie *vie;
struct vcpu *vcpu;
struct vm_exit *vme;
uint64_t gla, gpa, cs_base;
struct vm_guest_paging *paging;
mem_region_read_t mread;
mem_region_write_t mwrite;
enum vm_cpu_mode cpu_mode;
int cs_d, error, fault, length;
vcpu = &vm->vcpu[vcpuid];
vme = &vcpu->exitinfo;
gla = vme->u.inst_emul.gla;
gpa = vme->u.inst_emul.gpa;
cs_base = vme->u.inst_emul.cs_base;
cs_d = vme->u.inst_emul.cs_d;
vie = &vme->u.inst_emul.vie;
paging = &vme->u.inst_emul.paging;
cpu_mode = paging->cpu_mode;
VCPU_CTR1(vm, vcpuid, "inst_emul fault accessing gpa %#llx", gpa);
/* Fetch, decode and emulate the faulting instruction */
if (vie->num_valid == 0) {
/*
* If the instruction length is not known then assume a
* maximum size instruction.
*/
length = vme->inst_length ? vme->inst_length : VIE_INST_SIZE;
error = vmm_fetch_instruction(vm, vcpuid, paging, vme->rip +
cs_base, length, vie, &fault);
} else {
/*
* The instruction bytes have already been copied into 'vie'
*/
error = fault = 0;
}
if (error || fault)
return (error);
if (vmm_decode_instruction(vm, vcpuid, gla, cpu_mode, cs_d, vie) != 0) {
VCPU_CTR1(vm, vcpuid, "Error decoding instruction at %#llx",
vme->rip + cs_base);
*retu = true; /* dump instruction bytes in userspace */
return (0);
}
/*
* If the instruction length was not specified then update it now
* along with 'nextrip'.
*/
if (vme->inst_length == 0) {
vme->inst_length = vie->num_processed;
vcpu->nextrip += vie->num_processed;
}
/* return to userland unless this is an in-kernel emulated device */
if (gpa >= DEFAULT_APIC_BASE && gpa < DEFAULT_APIC_BASE + XHYVE_PAGE_SIZE) {
mread = lapic_mmio_read;
mwrite = lapic_mmio_write;
} else if (gpa >= VIOAPIC_BASE && gpa < VIOAPIC_BASE + VIOAPIC_SIZE) {
mread = vioapic_mmio_read;
mwrite = vioapic_mmio_write;
} else if (gpa >= VHPET_BASE && gpa < VHPET_BASE + VHPET_SIZE) {
mread = vhpet_mmio_read;
mwrite = vhpet_mmio_write;
} else {
*retu = true;
return (0);
}
error = vmm_emulate_instruction(vm, vcpuid, gpa, vie, paging,
mread, mwrite, retu);
return (error);
}
static int
vm_handle_suspend(struct vm *vm, int vcpuid, bool *retu)
{
int i, done;
struct vcpu *vcpu;
const struct timespec ts = {.tv_sec = 1, .tv_nsec = 0}; /* 1 second */
done = 0;
vcpu = &vm->vcpu[vcpuid];
CPU_SET_ATOMIC(((unsigned) vcpuid), &vm->suspended_cpus);
/*
* Wait until all 'active_cpus' have suspended themselves.
*
* Since a VM may be suspended at any time including when one or
* more vcpus are doing a rendezvous we need to call the rendezvous
* handler while we are waiting to prevent a deadlock.
*/
vcpu_lock(vcpu);
while (1) {
if (CPU_CMP(&vm->suspended_cpus, &vm->active_cpus) == 0) {
VCPU_CTR0(vm, vcpuid, "All vcpus suspended");
break;
}
if (vm->rendezvous_func == NULL) {
VCPU_CTR0(vm, vcpuid, "Sleeping during suspend");
vcpu_require_state_locked(vcpu, VCPU_SLEEPING);
pthread_mutex_lock(&vcpu->vcpu_sleep_mtx);
vcpu_unlock(vcpu);
pthread_cond_timedwait_relative_np(&vcpu->vcpu_sleep_cnd,
&vcpu->vcpu_sleep_mtx, &ts);
vcpu_lock(vcpu);
pthread_mutex_unlock(&vcpu->vcpu_sleep_mtx);
//msleep_spin(vcpu, &vcpu->mtx, "vmsusp", hz);
vcpu_require_state_locked(vcpu, VCPU_FROZEN);
} else {
VCPU_CTR0(vm, vcpuid, "Rendezvous during suspend");
vcpu_unlock(vcpu);
vm_handle_rendezvous(vm, vcpuid);
vcpu_lock(vcpu);
}
}
vcpu_unlock(vcpu);
/*
* Wakeup the other sleeping vcpus and return to userspace.
*/
for (i = 0; i < VM_MAXCPU; i++) {
if (CPU_ISSET(((unsigned) i), &vm->suspended_cpus)) {
vcpu_notify_event(vm, i, false);
}
}
*retu = true;
return (0);
}
int
vm_suspend(struct vm *vm, enum vm_suspend_how how)
{
int i;
if (how <= VM_SUSPEND_NONE || how >= VM_SUSPEND_LAST)
return (EINVAL);
if (atomic_cmpset_int(((volatile u_int *) &vm->suspend), 0, how) == 0) {