forked from Sricharanti/sricharan
-
Notifications
You must be signed in to change notification settings - Fork 0
/
xpc_uv.c
1813 lines (1497 loc) · 46.8 KB
/
xpc_uv.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
/*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*
* Copyright (c) 2008-2009 Silicon Graphics, Inc. All Rights Reserved.
*/
/*
* Cross Partition Communication (XPC) uv-based functions.
*
* Architecture specific implementation of common functions.
*
*/
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/delay.h>
#include <linux/device.h>
#include <linux/cpu.h>
#include <linux/module.h>
#include <linux/err.h>
#include <linux/slab.h>
#include <asm/uv/uv_hub.h>
#if defined CONFIG_X86_64
#include <asm/uv/bios.h>
#include <asm/uv/uv_irq.h>
#elif defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
#include <asm/sn/intr.h>
#include <asm/sn/sn_sal.h>
#endif
#include "../sgi-gru/gru.h"
#include "../sgi-gru/grukservices.h"
#include "xpc.h"
#if defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
struct uv_IO_APIC_route_entry {
__u64 vector : 8,
delivery_mode : 3,
dest_mode : 1,
delivery_status : 1,
polarity : 1,
__reserved_1 : 1,
trigger : 1,
mask : 1,
__reserved_2 : 15,
dest : 32;
};
#endif
static struct xpc_heartbeat_uv *xpc_heartbeat_uv;
#define XPC_ACTIVATE_MSG_SIZE_UV (1 * GRU_CACHE_LINE_BYTES)
#define XPC_ACTIVATE_MQ_SIZE_UV (4 * XP_MAX_NPARTITIONS_UV * \
XPC_ACTIVATE_MSG_SIZE_UV)
#define XPC_ACTIVATE_IRQ_NAME "xpc_activate"
#define XPC_NOTIFY_MSG_SIZE_UV (2 * GRU_CACHE_LINE_BYTES)
#define XPC_NOTIFY_MQ_SIZE_UV (4 * XP_MAX_NPARTITIONS_UV * \
XPC_NOTIFY_MSG_SIZE_UV)
#define XPC_NOTIFY_IRQ_NAME "xpc_notify"
static int xpc_mq_node = -1;
static struct xpc_gru_mq_uv *xpc_activate_mq_uv;
static struct xpc_gru_mq_uv *xpc_notify_mq_uv;
static int
xpc_setup_partitions_uv(void)
{
short partid;
struct xpc_partition_uv *part_uv;
for (partid = 0; partid < XP_MAX_NPARTITIONS_UV; partid++) {
part_uv = &xpc_partitions[partid].sn.uv;
mutex_init(&part_uv->cached_activate_gru_mq_desc_mutex);
spin_lock_init(&part_uv->flags_lock);
part_uv->remote_act_state = XPC_P_AS_INACTIVE;
}
return 0;
}
static void
xpc_teardown_partitions_uv(void)
{
short partid;
struct xpc_partition_uv *part_uv;
unsigned long irq_flags;
for (partid = 0; partid < XP_MAX_NPARTITIONS_UV; partid++) {
part_uv = &xpc_partitions[partid].sn.uv;
if (part_uv->cached_activate_gru_mq_desc != NULL) {
mutex_lock(&part_uv->cached_activate_gru_mq_desc_mutex);
spin_lock_irqsave(&part_uv->flags_lock, irq_flags);
part_uv->flags &= ~XPC_P_CACHED_ACTIVATE_GRU_MQ_DESC_UV;
spin_unlock_irqrestore(&part_uv->flags_lock, irq_flags);
kfree(part_uv->cached_activate_gru_mq_desc);
part_uv->cached_activate_gru_mq_desc = NULL;
mutex_unlock(&part_uv->
cached_activate_gru_mq_desc_mutex);
}
}
}
static int
xpc_get_gru_mq_irq_uv(struct xpc_gru_mq_uv *mq, int cpu, char *irq_name)
{
int mmr_pnode = uv_blade_to_pnode(mq->mmr_blade);
#if defined CONFIG_X86_64
mq->irq = uv_setup_irq(irq_name, cpu, mq->mmr_blade, mq->mmr_offset,
UV_AFFINITY_CPU);
if (mq->irq < 0)
return mq->irq;
mq->mmr_value = uv_read_global_mmr64(mmr_pnode, mq->mmr_offset);
#elif defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
if (strcmp(irq_name, XPC_ACTIVATE_IRQ_NAME) == 0)
mq->irq = SGI_XPC_ACTIVATE;
else if (strcmp(irq_name, XPC_NOTIFY_IRQ_NAME) == 0)
mq->irq = SGI_XPC_NOTIFY;
else
return -EINVAL;
mq->mmr_value = (unsigned long)cpu_physical_id(cpu) << 32 | mq->irq;
uv_write_global_mmr64(mmr_pnode, mq->mmr_offset, mq->mmr_value);
#else
#error not a supported configuration
#endif
return 0;
}
static void
xpc_release_gru_mq_irq_uv(struct xpc_gru_mq_uv *mq)
{
#if defined CONFIG_X86_64
uv_teardown_irq(mq->irq);
#elif defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
int mmr_pnode;
unsigned long mmr_value;
mmr_pnode = uv_blade_to_pnode(mq->mmr_blade);
mmr_value = 1UL << 16;
uv_write_global_mmr64(mmr_pnode, mq->mmr_offset, mmr_value);
#else
#error not a supported configuration
#endif
}
static int
xpc_gru_mq_watchlist_alloc_uv(struct xpc_gru_mq_uv *mq)
{
int ret;
#if defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
int mmr_pnode = uv_blade_to_pnode(mq->mmr_blade);
ret = sn_mq_watchlist_alloc(mmr_pnode, (void *)uv_gpa(mq->address),
mq->order, &mq->mmr_offset);
if (ret < 0) {
dev_err(xpc_part, "sn_mq_watchlist_alloc() failed, ret=%d\n",
ret);
return -EBUSY;
}
#elif defined CONFIG_X86_64
ret = uv_bios_mq_watchlist_alloc(uv_gpa(mq->address),
mq->order, &mq->mmr_offset);
if (ret < 0) {
dev_err(xpc_part, "uv_bios_mq_watchlist_alloc() failed, "
"ret=%d\n", ret);
return ret;
}
#else
#error not a supported configuration
#endif
mq->watchlist_num = ret;
return 0;
}
static void
xpc_gru_mq_watchlist_free_uv(struct xpc_gru_mq_uv *mq)
{
int ret;
int mmr_pnode = uv_blade_to_pnode(mq->mmr_blade);
#if defined CONFIG_X86_64
ret = uv_bios_mq_watchlist_free(mmr_pnode, mq->watchlist_num);
BUG_ON(ret != BIOS_STATUS_SUCCESS);
#elif defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
ret = sn_mq_watchlist_free(mmr_pnode, mq->watchlist_num);
BUG_ON(ret != SALRET_OK);
#else
#error not a supported configuration
#endif
}
static struct xpc_gru_mq_uv *
xpc_create_gru_mq_uv(unsigned int mq_size, int cpu, char *irq_name,
irq_handler_t irq_handler)
{
enum xp_retval xp_ret;
int ret;
int nid;
int nasid;
int pg_order;
struct page *page;
struct xpc_gru_mq_uv *mq;
struct uv_IO_APIC_route_entry *mmr_value;
mq = kmalloc(sizeof(struct xpc_gru_mq_uv), GFP_KERNEL);
if (mq == NULL) {
dev_err(xpc_part, "xpc_create_gru_mq_uv() failed to kmalloc() "
"a xpc_gru_mq_uv structure\n");
ret = -ENOMEM;
goto out_0;
}
mq->gru_mq_desc = kzalloc(sizeof(struct gru_message_queue_desc),
GFP_KERNEL);
if (mq->gru_mq_desc == NULL) {
dev_err(xpc_part, "xpc_create_gru_mq_uv() failed to kmalloc() "
"a gru_message_queue_desc structure\n");
ret = -ENOMEM;
goto out_1;
}
pg_order = get_order(mq_size);
mq->order = pg_order + PAGE_SHIFT;
mq_size = 1UL << mq->order;
mq->mmr_blade = uv_cpu_to_blade_id(cpu);
nid = cpu_to_node(cpu);
page = alloc_pages_exact_node(nid,
GFP_KERNEL | __GFP_ZERO | GFP_THISNODE,
pg_order);
if (page == NULL) {
dev_err(xpc_part, "xpc_create_gru_mq_uv() failed to alloc %d "
"bytes of memory on nid=%d for GRU mq\n", mq_size, nid);
ret = -ENOMEM;
goto out_2;
}
mq->address = page_address(page);
/* enable generation of irq when GRU mq operation occurs to this mq */
ret = xpc_gru_mq_watchlist_alloc_uv(mq);
if (ret != 0)
goto out_3;
ret = xpc_get_gru_mq_irq_uv(mq, cpu, irq_name);
if (ret != 0)
goto out_4;
ret = request_irq(mq->irq, irq_handler, 0, irq_name, NULL);
if (ret != 0) {
dev_err(xpc_part, "request_irq(irq=%d) returned error=%d\n",
mq->irq, -ret);
goto out_5;
}
nasid = UV_PNODE_TO_NASID(uv_cpu_to_pnode(cpu));
mmr_value = (struct uv_IO_APIC_route_entry *)&mq->mmr_value;
ret = gru_create_message_queue(mq->gru_mq_desc, mq->address, mq_size,
nasid, mmr_value->vector, mmr_value->dest);
if (ret != 0) {
dev_err(xpc_part, "gru_create_message_queue() returned "
"error=%d\n", ret);
ret = -EINVAL;
goto out_6;
}
/* allow other partitions to access this GRU mq */
xp_ret = xp_expand_memprotect(xp_pa(mq->address), mq_size);
if (xp_ret != xpSuccess) {
ret = -EACCES;
goto out_6;
}
return mq;
/* something went wrong */
out_6:
free_irq(mq->irq, NULL);
out_5:
xpc_release_gru_mq_irq_uv(mq);
out_4:
xpc_gru_mq_watchlist_free_uv(mq);
out_3:
free_pages((unsigned long)mq->address, pg_order);
out_2:
kfree(mq->gru_mq_desc);
out_1:
kfree(mq);
out_0:
return ERR_PTR(ret);
}
static void
xpc_destroy_gru_mq_uv(struct xpc_gru_mq_uv *mq)
{
unsigned int mq_size;
int pg_order;
int ret;
/* disallow other partitions to access GRU mq */
mq_size = 1UL << mq->order;
ret = xp_restrict_memprotect(xp_pa(mq->address), mq_size);
BUG_ON(ret != xpSuccess);
/* unregister irq handler and release mq irq/vector mapping */
free_irq(mq->irq, NULL);
xpc_release_gru_mq_irq_uv(mq);
/* disable generation of irq when GRU mq op occurs to this mq */
xpc_gru_mq_watchlist_free_uv(mq);
pg_order = mq->order - PAGE_SHIFT;
free_pages((unsigned long)mq->address, pg_order);
kfree(mq);
}
static enum xp_retval
xpc_send_gru_msg(struct gru_message_queue_desc *gru_mq_desc, void *msg,
size_t msg_size)
{
enum xp_retval xp_ret;
int ret;
while (1) {
ret = gru_send_message_gpa(gru_mq_desc, msg, msg_size);
if (ret == MQE_OK) {
xp_ret = xpSuccess;
break;
}
if (ret == MQE_QUEUE_FULL) {
dev_dbg(xpc_chan, "gru_send_message_gpa() returned "
"error=MQE_QUEUE_FULL\n");
/* !!! handle QLimit reached; delay & try again */
/* ??? Do we add a limit to the number of retries? */
(void)msleep_interruptible(10);
} else if (ret == MQE_CONGESTION) {
dev_dbg(xpc_chan, "gru_send_message_gpa() returned "
"error=MQE_CONGESTION\n");
/* !!! handle LB Overflow; simply try again */
/* ??? Do we add a limit to the number of retries? */
} else {
/* !!! Currently this is MQE_UNEXPECTED_CB_ERR */
dev_err(xpc_chan, "gru_send_message_gpa() returned "
"error=%d\n", ret);
xp_ret = xpGruSendMqError;
break;
}
}
return xp_ret;
}
static void
xpc_process_activate_IRQ_rcvd_uv(void)
{
unsigned long irq_flags;
short partid;
struct xpc_partition *part;
u8 act_state_req;
DBUG_ON(xpc_activate_IRQ_rcvd == 0);
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
for (partid = 0; partid < XP_MAX_NPARTITIONS_UV; partid++) {
part = &xpc_partitions[partid];
if (part->sn.uv.act_state_req == 0)
continue;
xpc_activate_IRQ_rcvd--;
BUG_ON(xpc_activate_IRQ_rcvd < 0);
act_state_req = part->sn.uv.act_state_req;
part->sn.uv.act_state_req = 0;
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (act_state_req == XPC_P_ASR_ACTIVATE_UV) {
if (part->act_state == XPC_P_AS_INACTIVE)
xpc_activate_partition(part);
else if (part->act_state == XPC_P_AS_DEACTIVATING)
XPC_DEACTIVATE_PARTITION(part, xpReactivating);
} else if (act_state_req == XPC_P_ASR_REACTIVATE_UV) {
if (part->act_state == XPC_P_AS_INACTIVE)
xpc_activate_partition(part);
else
XPC_DEACTIVATE_PARTITION(part, xpReactivating);
} else if (act_state_req == XPC_P_ASR_DEACTIVATE_UV) {
XPC_DEACTIVATE_PARTITION(part, part->sn.uv.reason);
} else {
BUG();
}
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (xpc_activate_IRQ_rcvd == 0)
break;
}
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
}
static void
xpc_handle_activate_mq_msg_uv(struct xpc_partition *part,
struct xpc_activate_mq_msghdr_uv *msg_hdr,
int part_setup,
int *wakeup_hb_checker)
{
unsigned long irq_flags;
struct xpc_partition_uv *part_uv = &part->sn.uv;
struct xpc_openclose_args *args;
part_uv->remote_act_state = msg_hdr->act_state;
switch (msg_hdr->type) {
case XPC_ACTIVATE_MQ_MSG_SYNC_ACT_STATE_UV:
/* syncing of remote_act_state was just done above */
break;
case XPC_ACTIVATE_MQ_MSG_ACTIVATE_REQ_UV: {
struct xpc_activate_mq_msg_activate_req_uv *msg;
/*
* ??? Do we deal here with ts_jiffies being different
* ??? if act_state != XPC_P_AS_INACTIVE instead of
* ??? below?
*/
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_activate_req_uv, hdr);
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (part_uv->act_state_req == 0)
xpc_activate_IRQ_rcvd++;
part_uv->act_state_req = XPC_P_ASR_ACTIVATE_UV;
part->remote_rp_pa = msg->rp_gpa; /* !!! _pa is _gpa */
part->remote_rp_ts_jiffies = msg_hdr->rp_ts_jiffies;
part_uv->heartbeat_gpa = msg->heartbeat_gpa;
if (msg->activate_gru_mq_desc_gpa !=
part_uv->activate_gru_mq_desc_gpa) {
spin_lock(&part_uv->flags_lock);
part_uv->flags &= ~XPC_P_CACHED_ACTIVATE_GRU_MQ_DESC_UV;
spin_unlock(&part_uv->flags_lock);
part_uv->activate_gru_mq_desc_gpa =
msg->activate_gru_mq_desc_gpa;
}
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
(*wakeup_hb_checker)++;
break;
}
case XPC_ACTIVATE_MQ_MSG_DEACTIVATE_REQ_UV: {
struct xpc_activate_mq_msg_deactivate_req_uv *msg;
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_deactivate_req_uv, hdr);
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (part_uv->act_state_req == 0)
xpc_activate_IRQ_rcvd++;
part_uv->act_state_req = XPC_P_ASR_DEACTIVATE_UV;
part_uv->reason = msg->reason;
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
(*wakeup_hb_checker)++;
return;
}
case XPC_ACTIVATE_MQ_MSG_CHCTL_CLOSEREQUEST_UV: {
struct xpc_activate_mq_msg_chctl_closerequest_uv *msg;
if (!part_setup)
break;
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_chctl_closerequest_uv,
hdr);
args = &part->remote_openclose_args[msg->ch_number];
args->reason = msg->reason;
spin_lock_irqsave(&part->chctl_lock, irq_flags);
part->chctl.flags[msg->ch_number] |= XPC_CHCTL_CLOSEREQUEST;
spin_unlock_irqrestore(&part->chctl_lock, irq_flags);
xpc_wakeup_channel_mgr(part);
break;
}
case XPC_ACTIVATE_MQ_MSG_CHCTL_CLOSEREPLY_UV: {
struct xpc_activate_mq_msg_chctl_closereply_uv *msg;
if (!part_setup)
break;
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_chctl_closereply_uv,
hdr);
spin_lock_irqsave(&part->chctl_lock, irq_flags);
part->chctl.flags[msg->ch_number] |= XPC_CHCTL_CLOSEREPLY;
spin_unlock_irqrestore(&part->chctl_lock, irq_flags);
xpc_wakeup_channel_mgr(part);
break;
}
case XPC_ACTIVATE_MQ_MSG_CHCTL_OPENREQUEST_UV: {
struct xpc_activate_mq_msg_chctl_openrequest_uv *msg;
if (!part_setup)
break;
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_chctl_openrequest_uv,
hdr);
args = &part->remote_openclose_args[msg->ch_number];
args->entry_size = msg->entry_size;
args->local_nentries = msg->local_nentries;
spin_lock_irqsave(&part->chctl_lock, irq_flags);
part->chctl.flags[msg->ch_number] |= XPC_CHCTL_OPENREQUEST;
spin_unlock_irqrestore(&part->chctl_lock, irq_flags);
xpc_wakeup_channel_mgr(part);
break;
}
case XPC_ACTIVATE_MQ_MSG_CHCTL_OPENREPLY_UV: {
struct xpc_activate_mq_msg_chctl_openreply_uv *msg;
if (!part_setup)
break;
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_chctl_openreply_uv, hdr);
args = &part->remote_openclose_args[msg->ch_number];
args->remote_nentries = msg->remote_nentries;
args->local_nentries = msg->local_nentries;
args->local_msgqueue_pa = msg->notify_gru_mq_desc_gpa;
spin_lock_irqsave(&part->chctl_lock, irq_flags);
part->chctl.flags[msg->ch_number] |= XPC_CHCTL_OPENREPLY;
spin_unlock_irqrestore(&part->chctl_lock, irq_flags);
xpc_wakeup_channel_mgr(part);
break;
}
case XPC_ACTIVATE_MQ_MSG_CHCTL_OPENCOMPLETE_UV: {
struct xpc_activate_mq_msg_chctl_opencomplete_uv *msg;
if (!part_setup)
break;
msg = container_of(msg_hdr, struct
xpc_activate_mq_msg_chctl_opencomplete_uv, hdr);
spin_lock_irqsave(&part->chctl_lock, irq_flags);
part->chctl.flags[msg->ch_number] |= XPC_CHCTL_OPENCOMPLETE;
spin_unlock_irqrestore(&part->chctl_lock, irq_flags);
xpc_wakeup_channel_mgr(part);
}
case XPC_ACTIVATE_MQ_MSG_MARK_ENGAGED_UV:
spin_lock_irqsave(&part_uv->flags_lock, irq_flags);
part_uv->flags |= XPC_P_ENGAGED_UV;
spin_unlock_irqrestore(&part_uv->flags_lock, irq_flags);
break;
case XPC_ACTIVATE_MQ_MSG_MARK_DISENGAGED_UV:
spin_lock_irqsave(&part_uv->flags_lock, irq_flags);
part_uv->flags &= ~XPC_P_ENGAGED_UV;
spin_unlock_irqrestore(&part_uv->flags_lock, irq_flags);
break;
default:
dev_err(xpc_part, "received unknown activate_mq msg type=%d "
"from partition=%d\n", msg_hdr->type, XPC_PARTID(part));
/* get hb checker to deactivate from the remote partition */
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (part_uv->act_state_req == 0)
xpc_activate_IRQ_rcvd++;
part_uv->act_state_req = XPC_P_ASR_DEACTIVATE_UV;
part_uv->reason = xpBadMsgType;
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
(*wakeup_hb_checker)++;
return;
}
if (msg_hdr->rp_ts_jiffies != part->remote_rp_ts_jiffies &&
part->remote_rp_ts_jiffies != 0) {
/*
* ??? Does what we do here need to be sensitive to
* ??? act_state or remote_act_state?
*/
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (part_uv->act_state_req == 0)
xpc_activate_IRQ_rcvd++;
part_uv->act_state_req = XPC_P_ASR_REACTIVATE_UV;
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
(*wakeup_hb_checker)++;
}
}
static irqreturn_t
xpc_handle_activate_IRQ_uv(int irq, void *dev_id)
{
struct xpc_activate_mq_msghdr_uv *msg_hdr;
short partid;
struct xpc_partition *part;
int wakeup_hb_checker = 0;
int part_referenced;
while (1) {
msg_hdr = gru_get_next_message(xpc_activate_mq_uv->gru_mq_desc);
if (msg_hdr == NULL)
break;
partid = msg_hdr->partid;
if (partid < 0 || partid >= XP_MAX_NPARTITIONS_UV) {
dev_err(xpc_part, "xpc_handle_activate_IRQ_uv() "
"received invalid partid=0x%x in message\n",
partid);
} else {
part = &xpc_partitions[partid];
part_referenced = xpc_part_ref(part);
xpc_handle_activate_mq_msg_uv(part, msg_hdr,
part_referenced,
&wakeup_hb_checker);
if (part_referenced)
xpc_part_deref(part);
}
gru_free_message(xpc_activate_mq_uv->gru_mq_desc, msg_hdr);
}
if (wakeup_hb_checker)
wake_up_interruptible(&xpc_activate_IRQ_wq);
return IRQ_HANDLED;
}
static enum xp_retval
xpc_cache_remote_gru_mq_desc_uv(struct gru_message_queue_desc *gru_mq_desc,
unsigned long gru_mq_desc_gpa)
{
enum xp_retval ret;
ret = xp_remote_memcpy(uv_gpa(gru_mq_desc), gru_mq_desc_gpa,
sizeof(struct gru_message_queue_desc));
if (ret == xpSuccess)
gru_mq_desc->mq = NULL;
return ret;
}
static enum xp_retval
xpc_send_activate_IRQ_uv(struct xpc_partition *part, void *msg, size_t msg_size,
int msg_type)
{
struct xpc_activate_mq_msghdr_uv *msg_hdr = msg;
struct xpc_partition_uv *part_uv = &part->sn.uv;
struct gru_message_queue_desc *gru_mq_desc;
unsigned long irq_flags;
enum xp_retval ret;
DBUG_ON(msg_size > XPC_ACTIVATE_MSG_SIZE_UV);
msg_hdr->type = msg_type;
msg_hdr->partid = xp_partition_id;
msg_hdr->act_state = part->act_state;
msg_hdr->rp_ts_jiffies = xpc_rsvd_page->ts_jiffies;
mutex_lock(&part_uv->cached_activate_gru_mq_desc_mutex);
again:
if (!(part_uv->flags & XPC_P_CACHED_ACTIVATE_GRU_MQ_DESC_UV)) {
gru_mq_desc = part_uv->cached_activate_gru_mq_desc;
if (gru_mq_desc == NULL) {
gru_mq_desc = kmalloc(sizeof(struct
gru_message_queue_desc),
GFP_KERNEL);
if (gru_mq_desc == NULL) {
ret = xpNoMemory;
goto done;
}
part_uv->cached_activate_gru_mq_desc = gru_mq_desc;
}
ret = xpc_cache_remote_gru_mq_desc_uv(gru_mq_desc,
part_uv->
activate_gru_mq_desc_gpa);
if (ret != xpSuccess)
goto done;
spin_lock_irqsave(&part_uv->flags_lock, irq_flags);
part_uv->flags |= XPC_P_CACHED_ACTIVATE_GRU_MQ_DESC_UV;
spin_unlock_irqrestore(&part_uv->flags_lock, irq_flags);
}
/* ??? Is holding a spin_lock (ch->lock) during this call a bad idea? */
ret = xpc_send_gru_msg(part_uv->cached_activate_gru_mq_desc, msg,
msg_size);
if (ret != xpSuccess) {
smp_rmb(); /* ensure a fresh copy of part_uv->flags */
if (!(part_uv->flags & XPC_P_CACHED_ACTIVATE_GRU_MQ_DESC_UV))
goto again;
}
done:
mutex_unlock(&part_uv->cached_activate_gru_mq_desc_mutex);
return ret;
}
static void
xpc_send_activate_IRQ_part_uv(struct xpc_partition *part, void *msg,
size_t msg_size, int msg_type)
{
enum xp_retval ret;
ret = xpc_send_activate_IRQ_uv(part, msg, msg_size, msg_type);
if (unlikely(ret != xpSuccess))
XPC_DEACTIVATE_PARTITION(part, ret);
}
static void
xpc_send_activate_IRQ_ch_uv(struct xpc_channel *ch, unsigned long *irq_flags,
void *msg, size_t msg_size, int msg_type)
{
struct xpc_partition *part = &xpc_partitions[ch->partid];
enum xp_retval ret;
ret = xpc_send_activate_IRQ_uv(part, msg, msg_size, msg_type);
if (unlikely(ret != xpSuccess)) {
if (irq_flags != NULL)
spin_unlock_irqrestore(&ch->lock, *irq_flags);
XPC_DEACTIVATE_PARTITION(part, ret);
if (irq_flags != NULL)
spin_lock_irqsave(&ch->lock, *irq_flags);
}
}
static void
xpc_send_local_activate_IRQ_uv(struct xpc_partition *part, int act_state_req)
{
unsigned long irq_flags;
struct xpc_partition_uv *part_uv = &part->sn.uv;
/*
* !!! Make our side think that the remote partition sent an activate
* !!! mq message our way by doing what the activate IRQ handler would
* !!! do had one really been sent.
*/
spin_lock_irqsave(&xpc_activate_IRQ_rcvd_lock, irq_flags);
if (part_uv->act_state_req == 0)
xpc_activate_IRQ_rcvd++;
part_uv->act_state_req = act_state_req;
spin_unlock_irqrestore(&xpc_activate_IRQ_rcvd_lock, irq_flags);
wake_up_interruptible(&xpc_activate_IRQ_wq);
}
static enum xp_retval
xpc_get_partition_rsvd_page_pa_uv(void *buf, u64 *cookie, unsigned long *rp_pa,
size_t *len)
{
s64 status;
enum xp_retval ret;
#if defined CONFIG_X86_64
status = uv_bios_reserved_page_pa((u64)buf, cookie, (u64 *)rp_pa,
(u64 *)len);
if (status == BIOS_STATUS_SUCCESS)
ret = xpSuccess;
else if (status == BIOS_STATUS_MORE_PASSES)
ret = xpNeedMoreInfo;
else
ret = xpBiosError;
#elif defined CONFIG_IA64_GENERIC || defined CONFIG_IA64_SGI_UV
status = sn_partition_reserved_page_pa((u64)buf, cookie, rp_pa, len);
if (status == SALRET_OK)
ret = xpSuccess;
else if (status == SALRET_MORE_PASSES)
ret = xpNeedMoreInfo;
else
ret = xpSalError;
#else
#error not a supported configuration
#endif
return ret;
}
static int
xpc_setup_rsvd_page_uv(struct xpc_rsvd_page *rp)
{
xpc_heartbeat_uv =
&xpc_partitions[sn_partition_id].sn.uv.cached_heartbeat;
rp->sn.uv.heartbeat_gpa = uv_gpa(xpc_heartbeat_uv);
rp->sn.uv.activate_gru_mq_desc_gpa =
uv_gpa(xpc_activate_mq_uv->gru_mq_desc);
return 0;
}
static void
xpc_allow_hb_uv(short partid)
{
}
static void
xpc_disallow_hb_uv(short partid)
{
}
static void
xpc_disallow_all_hbs_uv(void)
{
}
static void
xpc_increment_heartbeat_uv(void)
{
xpc_heartbeat_uv->value++;
}
static void
xpc_offline_heartbeat_uv(void)
{
xpc_increment_heartbeat_uv();
xpc_heartbeat_uv->offline = 1;
}
static void
xpc_online_heartbeat_uv(void)
{
xpc_increment_heartbeat_uv();
xpc_heartbeat_uv->offline = 0;
}
static void
xpc_heartbeat_init_uv(void)
{
xpc_heartbeat_uv->value = 1;
xpc_heartbeat_uv->offline = 0;
}
static void
xpc_heartbeat_exit_uv(void)
{
xpc_offline_heartbeat_uv();
}
static enum xp_retval
xpc_get_remote_heartbeat_uv(struct xpc_partition *part)
{
struct xpc_partition_uv *part_uv = &part->sn.uv;
enum xp_retval ret;
ret = xp_remote_memcpy(uv_gpa(&part_uv->cached_heartbeat),
part_uv->heartbeat_gpa,
sizeof(struct xpc_heartbeat_uv));
if (ret != xpSuccess)
return ret;
if (part_uv->cached_heartbeat.value == part->last_heartbeat &&
!part_uv->cached_heartbeat.offline) {
ret = xpNoHeartbeat;
} else {
part->last_heartbeat = part_uv->cached_heartbeat.value;
}
return ret;
}
static void
xpc_request_partition_activation_uv(struct xpc_rsvd_page *remote_rp,
unsigned long remote_rp_gpa, int nasid)
{
short partid = remote_rp->SAL_partid;
struct xpc_partition *part = &xpc_partitions[partid];
struct xpc_activate_mq_msg_activate_req_uv msg;
part->remote_rp_pa = remote_rp_gpa; /* !!! _pa here is really _gpa */
part->remote_rp_ts_jiffies = remote_rp->ts_jiffies;
part->sn.uv.heartbeat_gpa = remote_rp->sn.uv.heartbeat_gpa;
part->sn.uv.activate_gru_mq_desc_gpa =
remote_rp->sn.uv.activate_gru_mq_desc_gpa;
/*
* ??? Is it a good idea to make this conditional on what is
* ??? potentially stale state information?
*/
if (part->sn.uv.remote_act_state == XPC_P_AS_INACTIVE) {
msg.rp_gpa = uv_gpa(xpc_rsvd_page);
msg.heartbeat_gpa = xpc_rsvd_page->sn.uv.heartbeat_gpa;
msg.activate_gru_mq_desc_gpa =
xpc_rsvd_page->sn.uv.activate_gru_mq_desc_gpa;
xpc_send_activate_IRQ_part_uv(part, &msg, sizeof(msg),
XPC_ACTIVATE_MQ_MSG_ACTIVATE_REQ_UV);
}
if (part->act_state == XPC_P_AS_INACTIVE)
xpc_send_local_activate_IRQ_uv(part, XPC_P_ASR_ACTIVATE_UV);
}
static void
xpc_request_partition_reactivation_uv(struct xpc_partition *part)
{
xpc_send_local_activate_IRQ_uv(part, XPC_P_ASR_ACTIVATE_UV);
}
static void
xpc_request_partition_deactivation_uv(struct xpc_partition *part)
{
struct xpc_activate_mq_msg_deactivate_req_uv msg;
/*
* ??? Is it a good idea to make this conditional on what is
* ??? potentially stale state information?
*/
if (part->sn.uv.remote_act_state != XPC_P_AS_DEACTIVATING &&
part->sn.uv.remote_act_state != XPC_P_AS_INACTIVE) {
msg.reason = part->reason;
xpc_send_activate_IRQ_part_uv(part, &msg, sizeof(msg),
XPC_ACTIVATE_MQ_MSG_DEACTIVATE_REQ_UV);
}
}
static void
xpc_cancel_partition_deactivation_request_uv(struct xpc_partition *part)
{
/* nothing needs to be done */
return;
}
static void
xpc_init_fifo_uv(struct xpc_fifo_head_uv *head)
{
head->first = NULL;
head->last = NULL;
spin_lock_init(&head->lock);
head->n_entries = 0;
}
static void *
xpc_get_fifo_entry_uv(struct xpc_fifo_head_uv *head)
{
unsigned long irq_flags;
struct xpc_fifo_entry_uv *first;
spin_lock_irqsave(&head->lock, irq_flags);
first = head->first;
if (head->first != NULL) {
head->first = first->next;
if (head->first == NULL)
head->last = NULL;
head->n_entries--;
BUG_ON(head->n_entries < 0);
first->next = NULL;
}
spin_unlock_irqrestore(&head->lock, irq_flags);
return first;
}
static void
xpc_put_fifo_entry_uv(struct xpc_fifo_head_uv *head,
struct xpc_fifo_entry_uv *last)
{
unsigned long irq_flags;
last->next = NULL;
spin_lock_irqsave(&head->lock, irq_flags);
if (head->last != NULL)
head->last->next = last;
else
head->first = last;
head->last = last;
head->n_entries++;
spin_unlock_irqrestore(&head->lock, irq_flags);
}