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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
|
/*-
* Copyright (c) 1996, by Steve Passe
* 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. The name of the developer may NOT be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``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 THE AUTHOR 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.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_apic.h"
#include "opt_cpu.h"
#include "opt_kstack_pages.h"
#include "opt_mp_watchdog.h"
#if !defined(lint)
#if !defined(SMP)
#error How did you get here?
#endif
#ifndef DEV_APIC
#error The apic device is required for SMP, add "device apic" to your config file.
#endif
#if defined(CPU_DISABLE_CMPXCHG) && !defined(COMPILING_LINT)
#error SMP not supported with CPU_DISABLE_CMPXCHG
#endif
#endif /* not lint */
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/bus.h>
#include <sys/cons.h> /* cngetc() */
#ifdef GPROF
#include <sys/gmon.h>
#endif
#include <sys/kernel.h>
#include <sys/ktr.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/memrange.h>
#include <sys/mutex.h>
#include <sys/pcpu.h>
#include <sys/proc.h>
#include <sys/smp.h>
#include <sys/sysctl.h>
#include <vm/vm.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
#include <vm/vm_kern.h>
#include <vm/vm_extern.h>
#include <machine/apicreg.h>
#include <machine/clock.h>
#include <machine/md_var.h>
#include <machine/mp_watchdog.h>
#include <machine/pcb.h>
#include <machine/smp.h>
#include <machine/smptests.h> /** COUNT_XINVLTLB_HITS */
#include <machine/specialreg.h>
#include <machine/privatespace.h>
#define WARMBOOT_TARGET 0
#define WARMBOOT_OFF (KERNBASE + 0x0467)
#define WARMBOOT_SEG (KERNBASE + 0x0469)
#define CMOS_REG (0x70)
#define CMOS_DATA (0x71)
#define BIOS_RESET (0x0f)
#define BIOS_WARM (0x0a)
/*
* this code MUST be enabled here and in mpboot.s.
* it follows the very early stages of AP boot by placing values in CMOS ram.
* it NORMALLY will never be needed and thus the primitive method for enabling.
*
#define CHECK_POINTS
*/
#if defined(CHECK_POINTS) && !defined(PC98)
#define CHECK_READ(A) (outb(CMOS_REG, (A)), inb(CMOS_DATA))
#define CHECK_WRITE(A,D) (outb(CMOS_REG, (A)), outb(CMOS_DATA, (D)))
#define CHECK_INIT(D); \
CHECK_WRITE(0x34, (D)); \
CHECK_WRITE(0x35, (D)); \
CHECK_WRITE(0x36, (D)); \
CHECK_WRITE(0x37, (D)); \
CHECK_WRITE(0x38, (D)); \
CHECK_WRITE(0x39, (D));
#define CHECK_PRINT(S); \
printf("%s: %d, %d, %d, %d, %d, %d\n", \
(S), \
CHECK_READ(0x34), \
CHECK_READ(0x35), \
CHECK_READ(0x36), \
CHECK_READ(0x37), \
CHECK_READ(0x38), \
CHECK_READ(0x39));
#else /* CHECK_POINTS */
#define CHECK_INIT(D)
#define CHECK_PRINT(S)
#define CHECK_WRITE(A, D)
#endif /* CHECK_POINTS */
/*
* Values to send to the POST hardware.
*/
#define MP_BOOTADDRESS_POST 0x10
#define MP_PROBE_POST 0x11
#define MPTABLE_PASS1_POST 0x12
#define MP_START_POST 0x13
#define MP_ENABLE_POST 0x14
#define MPTABLE_PASS2_POST 0x15
#define START_ALL_APS_POST 0x16
#define INSTALL_AP_TRAMP_POST 0x17
#define START_AP_POST 0x18
#define MP_ANNOUNCE_POST 0x19
/* lock region used by kernel profiling */
int mcount_lock;
/** XXX FIXME: where does this really belong, isa.h/isa.c perhaps? */
int current_postcode;
int mp_naps; /* # of Applications processors */
int boot_cpu_id = -1; /* designated BSP */
extern int nkpt;
/*
* CPU topology map datastructures for HTT.
*/
static struct cpu_group mp_groups[MAXCPU];
static struct cpu_top mp_top;
/* AP uses this during bootstrap. Do not staticize. */
char *bootSTK;
static int bootAP;
/* Hotwire a 0->4MB V==P mapping */
extern pt_entry_t *KPTphys;
/* SMP page table page */
extern pt_entry_t *SMPpt;
struct pcb stoppcbs[MAXCPU];
/* Variables needed for SMP tlb shootdown. */
vm_offset_t smp_tlb_addr1;
vm_offset_t smp_tlb_addr2;
volatile int smp_tlb_wait;
/*
* Local data and functions.
*/
static u_int logical_cpus;
/* used to hold the AP's until we are ready to release them */
static struct mtx ap_boot_mtx;
/* Set to 1 once we're ready to let the APs out of the pen. */
static volatile int aps_ready = 0;
/*
* Store data from cpu_add() until later in the boot when we actually setup
* the APs.
*/
struct cpu_info {
int cpu_present:1;
int cpu_bsp:1;
int cpu_disabled:1;
} static cpu_info[MAXCPU];
static int cpu_apic_ids[MAXCPU];
static u_int boot_address;
static void set_logical_apic_ids(void);
static int start_all_aps(void);
static void install_ap_tramp(void);
static int start_ap(int apic_id);
static void release_aps(void *dummy);
static int hlt_logical_cpus;
static struct sysctl_ctx_list logical_cpu_clist;
static void
mem_range_AP_init(void)
{
if (mem_range_softc.mr_op && mem_range_softc.mr_op->initAP)
mem_range_softc.mr_op->initAP(&mem_range_softc);
}
void
mp_topology(void)
{
struct cpu_group *group;
int logical_cpus;
int apic_id;
int groups;
int cpu;
/* Build the smp_topology map. */
/* Nothing to do if there is no HTT support. */
if ((cpu_feature & CPUID_HTT) == 0)
return;
logical_cpus = (cpu_procinfo & CPUID_HTT_CORES) >> 16;
if (logical_cpus <= 1)
return;
group = &mp_groups[0];
groups = 1;
for (cpu = 0, apic_id = 0; apic_id < MAXCPU; apic_id++) {
if (!cpu_info[apic_id].cpu_present)
continue;
/*
* If the current group has members and we're not a logical
* cpu, create a new group.
*/
if (group->cg_count != 0 && (apic_id % logical_cpus) == 0) {
group++;
groups++;
}
group->cg_count++;
group->cg_mask |= 1 << cpu;
cpu++;
}
mp_top.ct_count = groups;
mp_top.ct_group = mp_groups;
smp_topology = &mp_top;
}
/*
* Calculate usable address in base memory for AP trampoline code.
*/
u_int
mp_bootaddress(u_int basemem)
{
POSTCODE(MP_BOOTADDRESS_POST);
boot_address = trunc_page(basemem); /* round down to 4k boundary */
if ((basemem - boot_address) < bootMP_size)
boot_address -= PAGE_SIZE; /* not enough, lower by 4k */
return boot_address;
}
void
cpu_add(u_int apic_id, char boot_cpu)
{
if (apic_id >= MAXCPU) {
printf("SMP: CPU %d exceeds maximum CPU %d, ignoring\n",
apic_id, MAXCPU - 1);
return;
}
KASSERT(cpu_info[apic_id].cpu_present == 0, ("CPU %d added twice",
apic_id));
cpu_info[apic_id].cpu_present = 1;
if (boot_cpu) {
KASSERT(boot_cpu_id == -1,
("CPU %d claims to be BSP, but CPU %d already is", apic_id,
boot_cpu_id));
boot_cpu_id = apic_id;
cpu_info[apic_id].cpu_bsp = 1;
}
mp_ncpus++;
if (bootverbose)
printf("SMP: Added CPU %d (%s)\n", apic_id, boot_cpu ? "BSP" :
"AP");
}
void
cpu_mp_setmaxid(void)
{
mp_maxid = MAXCPU - 1;
}
int
cpu_mp_probe(void)
{
/*
* Always record BSP in CPU map so that the mbuf init code works
* correctly.
*/
all_cpus = 1;
if (mp_ncpus == 0) {
/*
* No CPUs were found, so this must be a UP system. Setup
* the variables to represent a system with a single CPU
* with an id of 0.
*/
mp_ncpus = 1;
return (0);
}
/* At least one CPU was found. */
if (mp_ncpus == 1) {
/*
* One CPU was found, so this must be a UP system with
* an I/O APIC.
*/
return (0);
}
/* At least two CPUs were found. */
return (1);
}
/*
* Initialize the IPI handlers and start up the AP's.
*/
void
cpu_mp_start(void)
{
int i;
POSTCODE(MP_START_POST);
/* Initialize the logical ID to APIC ID table. */
for (i = 0; i < MAXCPU; i++)
cpu_apic_ids[i] = -1;
/* Install an inter-CPU IPI for TLB invalidation */
setidt(IPI_INVLTLB, IDTVEC(invltlb),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
setidt(IPI_INVLPG, IDTVEC(invlpg),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
setidt(IPI_INVLRNG, IDTVEC(invlrng),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Install an inter-CPU IPI for forwarding hardclock() */
setidt(IPI_HARDCLOCK, IDTVEC(hardclock),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Install an inter-CPU IPI for forwarding statclock() */
setidt(IPI_STATCLOCK, IDTVEC(statclock),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Install an inter-CPU IPI for lazy pmap release */
setidt(IPI_LAZYPMAP, IDTVEC(lazypmap),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Install an inter-CPU IPI for all-CPU rendezvous */
setidt(IPI_RENDEZVOUS, IDTVEC(rendezvous),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Install an inter-CPU IPI for forcing an additional software trap */
setidt(IPI_AST, IDTVEC(cpuast),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Install an inter-CPU IPI for CPU stop/restart */
setidt(IPI_STOP, IDTVEC(cpustop),
SDT_SYS386IGT, SEL_KPL, GSEL(GCODE_SEL, SEL_KPL));
/* Set boot_cpu_id if needed. */
if (boot_cpu_id == -1) {
boot_cpu_id = PCPU_GET(apic_id);
cpu_info[boot_cpu_id].cpu_bsp = 1;
} else
KASSERT(boot_cpu_id == PCPU_GET(apic_id),
("BSP's APIC ID doesn't match boot_cpu_id"));
cpu_apic_ids[0] = boot_cpu_id;
/* Start each Application Processor */
start_all_aps();
/* Setup the initial logical CPUs info. */
logical_cpus = logical_cpus_mask = 0;
if (cpu_feature & CPUID_HTT)
logical_cpus = (cpu_procinfo & CPUID_HTT_CORES) >> 16;
set_logical_apic_ids();
}
/*
* Print various information about the SMP system hardware and setup.
*/
void
cpu_mp_announce(void)
{
int i, x;
POSTCODE(MP_ANNOUNCE_POST);
/* List CPUs */
printf(" cpu0 (BSP): APIC ID: %2d\n", boot_cpu_id);
for (i = 1, x = 0; x < MAXCPU; x++) {
if (!cpu_info[x].cpu_present || cpu_info[x].cpu_bsp)
continue;
if (cpu_info[x].cpu_disabled)
printf(" cpu (AP): APIC ID: %2d (disabled)\n", x);
else {
KASSERT(i < mp_ncpus,
("mp_ncpus and actual cpus are out of whack"));
printf(" cpu%d (AP): APIC ID: %2d\n", i++, x);
}
}
}
/*
* AP CPU's call this to initialize themselves.
*/
void
init_secondary(void)
{
int gsel_tss;
int x, myid;
u_int cr0;
/* bootAP is set in start_ap() to our ID. */
myid = bootAP;
gdt_segs[GPRIV_SEL].ssd_base = (int) &SMP_prvspace[myid];
gdt_segs[GPROC0_SEL].ssd_base =
(int) &SMP_prvspace[myid].pcpu.pc_common_tss;
SMP_prvspace[myid].pcpu.pc_prvspace =
&SMP_prvspace[myid].pcpu;
for (x = 0; x < NGDT; x++) {
ssdtosd(&gdt_segs[x], &gdt[myid * NGDT + x].sd);
}
r_gdt.rd_limit = NGDT * sizeof(gdt[0]) - 1;
r_gdt.rd_base = (int) &gdt[myid * NGDT];
lgdt(&r_gdt); /* does magic intra-segment return */
lidt(&r_idt);
lldt(_default_ldt);
PCPU_SET(currentldt, _default_ldt);
gsel_tss = GSEL(GPROC0_SEL, SEL_KPL);
gdt[myid * NGDT + GPROC0_SEL].sd.sd_type = SDT_SYS386TSS;
PCPU_SET(common_tss.tss_esp0, 0); /* not used until after switch */
PCPU_SET(common_tss.tss_ss0, GSEL(GDATA_SEL, SEL_KPL));
PCPU_SET(common_tss.tss_ioopt, (sizeof (struct i386tss)) << 16);
PCPU_SET(tss_gdt, &gdt[myid * NGDT + GPROC0_SEL].sd);
PCPU_SET(common_tssd, *PCPU_GET(tss_gdt));
ltr(gsel_tss);
/*
* Set to a known state:
* Set by mpboot.s: CR0_PG, CR0_PE
* Set by cpu_setregs: CR0_NE, CR0_MP, CR0_TS, CR0_WP, CR0_AM
*/
cr0 = rcr0();
cr0 &= ~(CR0_CD | CR0_NW | CR0_EM);
load_cr0(cr0);
CHECK_WRITE(0x38, 5);
/* Disable local APIC just to be sure. */
lapic_disable();
/* signal our startup to the BSP. */
mp_naps++;
CHECK_WRITE(0x39, 6);
/* Spin until the BSP releases the AP's. */
while (!aps_ready)
ia32_pause();
/* BSP may have changed PTD while we were waiting */
invltlb();
pmap_invalidate_range(kernel_pmap, 0, NKPT * NBPDR - 1);
#if defined(I586_CPU) && !defined(NO_F00F_HACK)
lidt(&r_idt);
#endif
/* set up CPU registers and state */
cpu_setregs();
/* set up FPU state on the AP */
npxinit(__INITIAL_NPXCW__);
/* set up SSE registers */
enable_sse();
/* A quick check from sanity claus */
if (PCPU_GET(apic_id) != lapic_id()) {
printf("SMP: cpuid = %d\n", PCPU_GET(cpuid));
printf("SMP: actual apic_id = %d\n", lapic_id());
printf("SMP: correct apic_id = %d\n", PCPU_GET(apic_id));
printf("PTD[MPPTDI] = %#jx\n", (uintmax_t)PTD[MPPTDI]);
panic("cpuid mismatch! boom!!");
}
mtx_lock_spin(&ap_boot_mtx);
/* Init local apic for irq's */
lapic_setup();
/* Set memory range attributes for this CPU to match the BSP */
mem_range_AP_init();
smp_cpus++;
CTR1(KTR_SMP, "SMP: AP CPU #%d Launched", PCPU_GET(cpuid));
printf("SMP: AP CPU #%d Launched!\n", PCPU_GET(cpuid));
/* Determine if we are a logical CPU. */
if (logical_cpus > 1 && PCPU_GET(apic_id) % logical_cpus != 0)
logical_cpus_mask |= PCPU_GET(cpumask);
/* Build our map of 'other' CPUs. */
PCPU_SET(other_cpus, all_cpus & ~PCPU_GET(cpumask));
if (bootverbose)
lapic_dump("AP");
if (smp_cpus == mp_ncpus) {
/* enable IPI's, tlb shootdown, freezes etc */
atomic_store_rel_int(&smp_started, 1);
smp_active = 1; /* historic */
}
mtx_unlock_spin(&ap_boot_mtx);
/* wait until all the AP's are up */
while (smp_started == 0)
ia32_pause();
/* ok, now grab sched_lock and enter the scheduler */
mtx_lock_spin(&sched_lock);
binuptime(PCPU_PTR(switchtime));
PCPU_SET(switchticks, ticks);
cpu_throw(NULL, choosethread()); /* doesn't return */
panic("scheduler returned us to %s", __func__);
/* NOTREACHED */
}
/*******************************************************************
* local functions and data
*/
/*
* Set the APIC logical IDs.
*
* We want to cluster logical CPU's within the same APIC ID cluster.
* Since logical CPU's are aligned simply filling in the clusters in
* APIC ID order works fine. Note that this does not try to balance
* the number of CPU's in each cluster. (XXX?)
*/
static void
set_logical_apic_ids(void)
{
u_int apic_id, cluster, cluster_id;
/* Force us to allocate cluster 0 at the start. */
cluster = -1;
cluster_id = APIC_MAX_INTRACLUSTER_ID;
for (apic_id = 0; apic_id < MAXCPU; apic_id++) {
if (!cpu_info[apic_id].cpu_present)
continue;
if (cluster_id == APIC_MAX_INTRACLUSTER_ID) {
cluster = ioapic_next_logical_cluster();
cluster_id = 0;
} else
cluster_id++;
if (bootverbose)
printf("APIC ID: physical %u, logical %u:%u\n",
apic_id, cluster, cluster_id);
lapic_set_logical_id(apic_id, cluster, cluster_id);
}
}
/*
* start each AP in our list
*/
static int
start_all_aps(void)
{
#ifndef PC98
u_char mpbiosreason;
#endif
u_long mpbioswarmvec;
struct pcpu *pc;
char *stack;
uintptr_t kptbase;
int i, pg, apic_id, cpu;
POSTCODE(START_ALL_APS_POST);
mtx_init(&ap_boot_mtx, "ap boot", NULL, MTX_SPIN);
/* install the AP 1st level boot code */
install_ap_tramp();
/* save the current value of the warm-start vector */
mpbioswarmvec = *((u_long *) WARMBOOT_OFF);
#ifndef PC98
outb(CMOS_REG, BIOS_RESET);
mpbiosreason = inb(CMOS_DATA);
#endif
/* set up temporary P==V mapping for AP boot */
/* XXX this is a hack, we should boot the AP on its own stack/PTD */
kptbase = (uintptr_t)(void *)KPTphys;
for (i = 0; i < NKPT; i++)
PTD[i] = (pd_entry_t)(PG_V | PG_RW |
((kptbase + i * PAGE_SIZE) & PG_FRAME));
invltlb();
/* start each AP */
for (cpu = 0, apic_id = 0; apic_id < MAXCPU; apic_id++) {
/* Ignore non-existent CPUs and the BSP. */
if (!cpu_info[apic_id].cpu_present ||
cpu_info[apic_id].cpu_bsp)
continue;
/* Don't use this CPU if it has been disabled by a tunable. */
if (resource_disabled("lapic", apic_id)) {
cpu_info[apic_id].cpu_disabled = 1;
mp_ncpus--;
continue;
}
cpu++;
/* save APIC ID for this logical ID */
cpu_apic_ids[cpu] = apic_id;
/* first page of AP's private space */
pg = cpu * i386_btop(sizeof(struct privatespace));
/* allocate a new private data page */
pc = (struct pcpu *)kmem_alloc(kernel_map, PAGE_SIZE);
/* wire it into the private page table page */
SMPpt[pg] = (pt_entry_t)(PG_V | PG_RW | vtophys(pc));
/* allocate and set up an idle stack data page */
stack = (char *)kmem_alloc(kernel_map, KSTACK_PAGES * PAGE_SIZE); /* XXXKSE */
for (i = 0; i < KSTACK_PAGES; i++)
SMPpt[pg + 1 + i] = (pt_entry_t)
(PG_V | PG_RW | vtophys(PAGE_SIZE * i + stack));
/* prime data page for it to use */
pcpu_init(pc, cpu, sizeof(struct pcpu));
pc->pc_apic_id = apic_id;
/* setup a vector to our boot code */
*((volatile u_short *) WARMBOOT_OFF) = WARMBOOT_TARGET;
*((volatile u_short *) WARMBOOT_SEG) = (boot_address >> 4);
#ifndef PC98
outb(CMOS_REG, BIOS_RESET);
outb(CMOS_DATA, BIOS_WARM); /* 'warm-start' */
#endif
bootSTK = &SMP_prvspace[cpu].idlekstack[KSTACK_PAGES *
PAGE_SIZE];
bootAP = cpu;
/* attempt to start the Application Processor */
CHECK_INIT(99); /* setup checkpoints */
if (!start_ap(apic_id)) {
printf("AP #%d (PHY# %d) failed!\n", cpu, apic_id);
CHECK_PRINT("trace"); /* show checkpoints */
/* better panic as the AP may be running loose */
printf("panic y/n? [y] ");
if (cngetc() != 'n')
panic("bye-bye");
}
CHECK_PRINT("trace"); /* show checkpoints */
all_cpus |= (1 << cpu); /* record AP in CPU map */
}
/* build our map of 'other' CPUs */
PCPU_SET(other_cpus, all_cpus & ~PCPU_GET(cpumask));
/* restore the warmstart vector */
*(u_long *) WARMBOOT_OFF = mpbioswarmvec;
#ifndef PC98
outb(CMOS_REG, BIOS_RESET);
outb(CMOS_DATA, mpbiosreason);
#endif
/*
* Set up the idle context for the BSP. Similar to above except
* that some was done by locore, some by pmap.c and some is implicit
* because the BSP is cpu#0 and the page is initially zero and also
* because we can refer to variables by name on the BSP..
*/
/* Allocate and setup BSP idle stack */
stack = (char *)kmem_alloc(kernel_map, KSTACK_PAGES * PAGE_SIZE);
for (i = 0; i < KSTACK_PAGES; i++)
SMPpt[1 + i] = (pt_entry_t)
(PG_V | PG_RW | vtophys(PAGE_SIZE * i + stack));
for (i = 0; i < NKPT; i++)
PTD[i] = 0;
pmap_invalidate_range(kernel_pmap, 0, NKPT * NBPDR - 1);
/* number of APs actually started */
return mp_naps;
}
/*
* load the 1st level AP boot code into base memory.
*/
/* targets for relocation */
extern void bigJump(void);
extern void bootCodeSeg(void);
extern void bootDataSeg(void);
extern void MPentry(void);
extern u_int MP_GDT;
extern u_int mp_gdtbase;
static void
install_ap_tramp(void)
{
int x;
int size = *(int *) ((u_long) & bootMP_size);
vm_offset_t va = boot_address + KERNBASE;
u_char *src = (u_char *) ((u_long) bootMP);
u_char *dst = (u_char *) va;
u_int boot_base = (u_int) bootMP;
u_int8_t *dst8;
u_int16_t *dst16;
u_int32_t *dst32;
POSTCODE(INSTALL_AP_TRAMP_POST);
KASSERT (size <= PAGE_SIZE,
("'size' do not fit into PAGE_SIZE, as expected."));
pmap_kenter(va, boot_address);
pmap_invalidate_page (kernel_pmap, va);
for (x = 0; x < size; ++x)
*dst++ = *src++;
/*
* modify addresses in code we just moved to basemem. unfortunately we
* need fairly detailed info about mpboot.s for this to work. changes
* to mpboot.s might require changes here.
*/
/* boot code is located in KERNEL space */
dst = (u_char *) va;
/* modify the lgdt arg */
dst32 = (u_int32_t *) (dst + ((u_int) & mp_gdtbase - boot_base));
*dst32 = boot_address + ((u_int) & MP_GDT - boot_base);
/* modify the ljmp target for MPentry() */
dst32 = (u_int32_t *) (dst + ((u_int) bigJump - boot_base) + 1);
*dst32 = ((u_int) MPentry - KERNBASE);
/* modify the target for boot code segment */
dst16 = (u_int16_t *) (dst + ((u_int) bootCodeSeg - boot_base));
dst8 = (u_int8_t *) (dst16 + 1);
*dst16 = (u_int) boot_address & 0xffff;
*dst8 = ((u_int) boot_address >> 16) & 0xff;
/* modify the target for boot data segment */
dst16 = (u_int16_t *) (dst + ((u_int) bootDataSeg - boot_base));
dst8 = (u_int8_t *) (dst16 + 1);
*dst16 = (u_int) boot_address & 0xffff;
*dst8 = ((u_int) boot_address >> 16) & 0xff;
}
/*
* This function starts the AP (application processor) identified
* by the APIC ID 'physicalCpu'. It does quite a "song and dance"
* to accomplish this. This is necessary because of the nuances
* of the different hardware we might encounter. It isn't pretty,
* but it seems to work.
*/
static int
start_ap(int apic_id)
{
int vector, ms;
int cpus;
POSTCODE(START_AP_POST);
/* calculate the vector */
vector = (boot_address >> 12) & 0xff;
/* used as a watchpoint to signal AP startup */
cpus = mp_naps;
/*
* first we do an INIT/RESET IPI this INIT IPI might be run, reseting
* and running the target CPU. OR this INIT IPI might be latched (P5
* bug), CPU waiting for STARTUP IPI. OR this INIT IPI might be
* ignored.
*/
/* do an INIT IPI: assert RESET */
lapic_ipi_raw(APIC_DEST_DESTFLD | APIC_TRIGMOD_EDGE |
APIC_LEVEL_ASSERT | APIC_DESTMODE_PHY | APIC_DELMODE_INIT, apic_id);
/* wait for pending status end */
lapic_ipi_wait(-1);
/* do an INIT IPI: deassert RESET */
lapic_ipi_raw(APIC_DEST_ALLESELF | APIC_TRIGMOD_LEVEL |
APIC_LEVEL_DEASSERT | APIC_DESTMODE_PHY | APIC_DELMODE_INIT, 0);
/* wait for pending status end */
DELAY(10000); /* wait ~10mS */
lapic_ipi_wait(-1);
/*
* next we do a STARTUP IPI: the previous INIT IPI might still be
* latched, (P5 bug) this 1st STARTUP would then terminate
* immediately, and the previously started INIT IPI would continue. OR
* the previous INIT IPI has already run. and this STARTUP IPI will
* run. OR the previous INIT IPI was ignored. and this STARTUP IPI
* will run.
*/
/* do a STARTUP IPI */
lapic_ipi_raw(APIC_DEST_DESTFLD | APIC_TRIGMOD_EDGE |
APIC_LEVEL_DEASSERT | APIC_DESTMODE_PHY | APIC_DELMODE_STARTUP |
vector, apic_id);
lapic_ipi_wait(-1);
DELAY(200); /* wait ~200uS */
/*
* finally we do a 2nd STARTUP IPI: this 2nd STARTUP IPI should run IF
* the previous STARTUP IPI was cancelled by a latched INIT IPI. OR
* this STARTUP IPI will be ignored, as only ONE STARTUP IPI is
* recognized after hardware RESET or INIT IPI.
*/
lapic_ipi_raw(APIC_DEST_DESTFLD | APIC_TRIGMOD_EDGE |
APIC_LEVEL_DEASSERT | APIC_DESTMODE_PHY | APIC_DELMODE_STARTUP |
vector, apic_id);
lapic_ipi_wait(-1);
DELAY(200); /* wait ~200uS */
/* Wait up to 5 seconds for it to start. */
for (ms = 0; ms < 5000; ms++) {
if (mp_naps > cpus)
return 1; /* return SUCCESS */
DELAY(1000);
}
return 0; /* return FAILURE */
}
#ifdef COUNT_XINVLTLB_HITS
u_int xhits_gbl[MAXCPU];
u_int xhits_pg[MAXCPU];
u_int xhits_rng[MAXCPU];
SYSCTL_NODE(_debug, OID_AUTO, xhits, CTLFLAG_RW, 0, "");
SYSCTL_OPAQUE(_debug_xhits, OID_AUTO, global, CTLFLAG_RW, &xhits_gbl,
sizeof(xhits_gbl), "IU", "");
SYSCTL_OPAQUE(_debug_xhits, OID_AUTO, page, CTLFLAG_RW, &xhits_pg,
sizeof(xhits_pg), "IU", "");
SYSCTL_OPAQUE(_debug_xhits, OID_AUTO, range, CTLFLAG_RW, &xhits_rng,
sizeof(xhits_rng), "IU", "");
u_int ipi_global;
u_int ipi_page;
u_int ipi_range;
u_int ipi_range_size;
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_global, CTLFLAG_RW, &ipi_global, 0, "");
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_page, CTLFLAG_RW, &ipi_page, 0, "");
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_range, CTLFLAG_RW, &ipi_range, 0, "");
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_range_size, CTLFLAG_RW, &ipi_range_size,
0, "");
u_int ipi_masked_global;
u_int ipi_masked_page;
u_int ipi_masked_range;
u_int ipi_masked_range_size;
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_masked_global, CTLFLAG_RW,
&ipi_masked_global, 0, "");
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_masked_page, CTLFLAG_RW,
&ipi_masked_page, 0, "");
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_masked_range, CTLFLAG_RW,
&ipi_masked_range, 0, "");
SYSCTL_INT(_debug_xhits, OID_AUTO, ipi_masked_range_size, CTLFLAG_RW,
&ipi_masked_range_size, 0, "");
#endif /* COUNT_XINVLTLB_HITS */
/*
* Flush the TLB on all other CPU's
*/
static void
smp_tlb_shootdown(u_int vector, vm_offset_t addr1, vm_offset_t addr2)
{
u_int ncpu;
ncpu = mp_ncpus - 1; /* does not shootdown self */
if (ncpu < 1)
return; /* no other cpus */
mtx_assert(&smp_ipi_mtx, MA_OWNED);
smp_tlb_addr1 = addr1;
smp_tlb_addr2 = addr2;
atomic_store_rel_int(&smp_tlb_wait, 0);
ipi_all_but_self(vector);
while (smp_tlb_wait < ncpu)
ia32_pause();
}
/*
* This is about as magic as it gets. fortune(1) has got similar code
* for reversing bits in a word. Who thinks up this stuff??
*
* Yes, it does appear to be consistently faster than:
* while (i = ffs(m)) {
* m >>= i;
* bits++;
* }
* and
* while (lsb = (m & -m)) { // This is magic too
* m &= ~lsb; // or: m ^= lsb
* bits++;
* }
* Both of these latter forms do some very strange things on gcc-3.1 with
* -mcpu=pentiumpro and/or -march=pentiumpro and/or -O or -O2.
* There is probably an SSE or MMX popcnt instruction.
*
* I wonder if this should be in libkern?
*
* XXX Stop the presses! Another one:
* static __inline u_int32_t
* popcnt1(u_int32_t v)
* {
* v -= ((v >> 1) & 0x55555555);
* v = (v & 0x33333333) + ((v >> 2) & 0x33333333);
* v = (v + (v >> 4)) & 0x0F0F0F0F;
* return (v * 0x01010101) >> 24;
* }
* The downside is that it has a multiply. With a pentium3 with
* -mcpu=pentiumpro and -march=pentiumpro then gcc-3.1 will use
* an imull, and in that case it is faster. In most other cases
* it appears slightly slower.
*
* Another variant (also from fortune):
* #define BITCOUNT(x) (((BX_(x)+(BX_(x)>>4)) & 0x0F0F0F0F) % 255)
* #define BX_(x) ((x) - (((x)>>1)&0x77777777) \
* - (((x)>>2)&0x33333333) \
* - (((x)>>3)&0x11111111))
*/
static __inline u_int32_t
popcnt(u_int32_t m)
{
m = (m & 0x55555555) + ((m & 0xaaaaaaaa) >> 1);
m = (m & 0x33333333) + ((m & 0xcccccccc) >> 2);
m = (m & 0x0f0f0f0f) + ((m & 0xf0f0f0f0) >> 4);
m = (m & 0x00ff00ff) + ((m & 0xff00ff00) >> 8);
m = (m & 0x0000ffff) + ((m & 0xffff0000) >> 16);
return m;
}
static void
smp_targeted_tlb_shootdown(u_int mask, u_int vector, vm_offset_t addr1, vm_offset_t addr2)
{
int ncpu, othercpus;
othercpus = mp_ncpus - 1;
if (mask == (u_int)-1) {
ncpu = othercpus;
if (ncpu < 1)
return;
} else {
mask &= ~PCPU_GET(cpumask);
if (mask == 0)
return;
ncpu = popcnt(mask);
if (ncpu > othercpus) {
/* XXX this should be a panic offence */
printf("SMP: tlb shootdown to %d other cpus (only have %d)\n",
ncpu, othercpus);
ncpu = othercpus;
}
/* XXX should be a panic, implied by mask == 0 above */
if (ncpu < 1)
return;
}
mtx_assert(&smp_ipi_mtx, MA_OWNED);
smp_tlb_addr1 = addr1;
smp_tlb_addr2 = addr2;
atomic_store_rel_int(&smp_tlb_wait, 0);
if (mask == (u_int)-1)
ipi_all_but_self(vector);
else
ipi_selected(mask, vector);
while (smp_tlb_wait < ncpu)
ia32_pause();
}
void
smp_invltlb(void)
{
if (smp_started) {
smp_tlb_shootdown(IPI_INVLTLB, 0, 0);
#ifdef COUNT_XINVLTLB_HITS
ipi_global++;
#endif
}
}
void
smp_invlpg(vm_offset_t addr)
{
if (smp_started) {
smp_tlb_shootdown(IPI_INVLPG, addr, 0);
#ifdef COUNT_XINVLTLB_HITS
ipi_page++;
#endif
}
}
void
smp_invlpg_range(vm_offset_t addr1, vm_offset_t addr2)
{
if (smp_started) {
smp_tlb_shootdown(IPI_INVLRNG, addr1, addr2);
#ifdef COUNT_XINVLTLB_HITS
ipi_range++;
ipi_range_size += (addr2 - addr1) / PAGE_SIZE;
#endif
}
}
void
smp_masked_invltlb(u_int mask)
{
if (smp_started) {
smp_targeted_tlb_shootdown(mask, IPI_INVLTLB, 0, 0);
#ifdef COUNT_XINVLTLB_HITS
ipi_masked_global++;
#endif
}
}
void
smp_masked_invlpg(u_int mask, vm_offset_t addr)
{
if (smp_started) {
smp_targeted_tlb_shootdown(mask, IPI_INVLPG, addr, 0);
#ifdef COUNT_XINVLTLB_HITS
ipi_masked_page++;
#endif
}
}
void
smp_masked_invlpg_range(u_int mask, vm_offset_t addr1, vm_offset_t addr2)
{
if (smp_started) {
smp_targeted_tlb_shootdown(mask, IPI_INVLRNG, addr1, addr2);
#ifdef COUNT_XINVLTLB_HITS
ipi_masked_range++;
ipi_masked_range_size += (addr2 - addr1) / PAGE_SIZE;
#endif
}
}
/*
* For statclock, we send an IPI to all CPU's to have them call this
* function.
*/
void
forwarded_statclock(struct clockframe frame)
{
struct thread *td;
CTR0(KTR_SMP, "forwarded_statclock");
td = curthread;
td->td_intr_nesting_level++;
if (profprocs != 0)
profclock(&frame);
if (pscnt == psdiv)
statclock(&frame);
td->td_intr_nesting_level--;
}
void
forward_statclock(void)
{
int map;
CTR0(KTR_SMP, "forward_statclock");
if (!smp_started || cold || panicstr)
return;
map = PCPU_GET(other_cpus) & ~(stopped_cpus|hlt_cpus_mask);
if (map != 0)
ipi_selected(map, IPI_STATCLOCK);
}
/*
* For each hardclock(), we send an IPI to all other CPU's to have them
* execute this function. It would be nice to reduce contention on
* sched_lock if we could simply peek at the CPU to determine the user/kernel
* state and call hardclock_process() on the CPU receiving the clock interrupt
* and then just use a simple IPI to handle any ast's if needed.
*/
void
forwarded_hardclock(struct clockframe frame)
{
struct thread *td;
CTR0(KTR_SMP, "forwarded_hardclock");
td = curthread;
td->td_intr_nesting_level++;
hardclock_process(&frame);
td->td_intr_nesting_level--;
}
void
forward_hardclock(void)
{
u_int map;
CTR0(KTR_SMP, "forward_hardclock");
if (!smp_started || cold || panicstr)
return;
map = PCPU_GET(other_cpus) & ~(stopped_cpus|hlt_cpus_mask);
if (map != 0)
ipi_selected(map, IPI_HARDCLOCK);
}
/*
* send an IPI to a set of cpus.
*/
void
ipi_selected(u_int32_t cpus, u_int ipi)
{
int cpu;
CTR3(KTR_SMP, "%s: cpus: %x ipi: %x", __func__, cpus, ipi);
while ((cpu = ffs(cpus)) != 0) {
cpu--;
KASSERT(cpu_apic_ids[cpu] != -1,
("IPI to non-existent CPU %d", cpu));
lapic_ipi_vectored(ipi, cpu_apic_ids[cpu]);
cpus &= ~(1 << cpu);
}
}
/*
* send an IPI INTerrupt containing 'vector' to all CPUs, including myself
*/
void
ipi_all(u_int ipi)
{
CTR2(KTR_SMP, "%s: ipi: %x", __func__, ipi);
lapic_ipi_vectored(ipi, APIC_IPI_DEST_ALL);
}
/*
* send an IPI to all CPUs EXCEPT myself
*/
void
ipi_all_but_self(u_int ipi)
{
CTR2(KTR_SMP, "%s: ipi: %x", __func__, ipi);
lapic_ipi_vectored(ipi, APIC_IPI_DEST_OTHERS);
}
/*
* send an IPI to myself
*/
void
ipi_self(u_int ipi)
{
CTR2(KTR_SMP, "%s: ipi: %x", __func__, ipi);
lapic_ipi_vectored(ipi, APIC_IPI_DEST_SELF);
}
/*
* This is called once the rest of the system is up and running and we're
* ready to let the AP's out of the pen.
*/
static void
release_aps(void *dummy __unused)
{
if (mp_ncpus == 1)
return;
mtx_lock_spin(&sched_lock);
atomic_store_rel_int(&aps_ready, 1);
while (smp_started == 0)
ia32_pause();
mtx_unlock_spin(&sched_lock);
}
SYSINIT(start_aps, SI_SUB_SMP, SI_ORDER_FIRST, release_aps, NULL);
static int
sysctl_hlt_cpus(SYSCTL_HANDLER_ARGS)
{
u_int mask;
int error;
mask = hlt_cpus_mask;
error = sysctl_handle_int(oidp, &mask, 0, req);
if (error || !req->newptr)
return (error);
if (logical_cpus_mask != 0 &&
(mask & logical_cpus_mask) == logical_cpus_mask)
hlt_logical_cpus = 1;
else
hlt_logical_cpus = 0;
if ((mask & all_cpus) == all_cpus)
mask &= ~(1<<0);
hlt_cpus_mask = mask;
return (error);
}
SYSCTL_PROC(_machdep, OID_AUTO, hlt_cpus, CTLTYPE_INT|CTLFLAG_RW,
0, 0, sysctl_hlt_cpus, "IU",
"Bitmap of CPUs to halt. 101 (binary) will halt CPUs 0 and 2.");
static int
sysctl_hlt_logical_cpus(SYSCTL_HANDLER_ARGS)
{
int disable, error;
disable = hlt_logical_cpus;
error = sysctl_handle_int(oidp, &disable, 0, req);
if (error || !req->newptr)
return (error);
if (disable)
hlt_cpus_mask |= logical_cpus_mask;
else
hlt_cpus_mask &= ~logical_cpus_mask;
if ((hlt_cpus_mask & all_cpus) == all_cpus)
hlt_cpus_mask &= ~(1<<0);
hlt_logical_cpus = disable;
return (error);
}
static void
cpu_hlt_setup(void *dummy __unused)
{
if (logical_cpus_mask != 0) {
TUNABLE_INT_FETCH("machdep.hlt_logical_cpus",
&hlt_logical_cpus);
sysctl_ctx_init(&logical_cpu_clist);
SYSCTL_ADD_PROC(&logical_cpu_clist,
SYSCTL_STATIC_CHILDREN(_machdep), OID_AUTO,
"hlt_logical_cpus", CTLTYPE_INT|CTLFLAG_RW, 0, 0,
sysctl_hlt_logical_cpus, "IU", "");
SYSCTL_ADD_UINT(&logical_cpu_clist,
SYSCTL_STATIC_CHILDREN(_machdep), OID_AUTO,
"logical_cpus_mask", CTLTYPE_INT|CTLFLAG_RD,
&logical_cpus_mask, 0, "");
if (hlt_logical_cpus)
hlt_cpus_mask |= logical_cpus_mask;
}
}
SYSINIT(cpu_hlt, SI_SUB_SMP, SI_ORDER_ANY, cpu_hlt_setup, NULL);
int
mp_grab_cpu_hlt(void)
{
u_int mask = PCPU_GET(cpumask);
#ifdef MP_WATCHDOG
u_int cpuid = PCPU_GET(cpuid);
#endif
int retval;
#ifdef MP_WATCHDOG
ap_watchdog(cpuid);
#endif
retval = mask & hlt_cpus_mask;
while (mask & hlt_cpus_mask)
__asm __volatile("sti; hlt" : : : "memory");
return (retval);
}
|