summaryrefslogtreecommitdiff
path: root/drivers/infiniband/sw/rdmavt/mr.c
blob: 54f3f9c27552776229fc0d8f7896f6f766605390 (plain)
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
/*
 * Copyright(c) 2016 Intel Corporation.
 *
 * This file is provided under a dual BSD/GPLv2 license.  When using or
 * redistributing this file, you may do so under either license.
 *
 * GPL LICENSE SUMMARY
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of version 2 of the GNU General Public License as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * BSD LICENSE
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 *  - Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 *  - 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.
 *  - Neither the name of Intel Corporation nor the names of its
 *    contributors may be used to endorse or promote products derived
 *    from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS 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 COPYRIGHT
 * OWNER 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 <linux/slab.h>
#include <linux/vmalloc.h>
#include <rdma/ib_umem.h>
#include <rdma/rdma_vt.h>
#include "vt.h"
#include "mr.h"
#include "trace.h"

/**
 * rvt_driver_mr_init - Init MR resources per driver
 * @rdi: rvt dev struct
 *
 * Do any intilization needed when a driver registers with rdmavt.
 *
 * Return: 0 on success or errno on failure
 */
int rvt_driver_mr_init(struct rvt_dev_info *rdi)
{
	unsigned int lkey_table_size = rdi->dparms.lkey_table_size;
	unsigned lk_tab_size;
	int i;

	/*
	 * The top hfi1_lkey_table_size bits are used to index the
	 * table.  The lower 8 bits can be owned by the user (copied from
	 * the LKEY).  The remaining bits act as a generation number or tag.
	 */
	if (!lkey_table_size)
		return -EINVAL;

	spin_lock_init(&rdi->lkey_table.lock);

	/* ensure generation is at least 4 bits */
	if (lkey_table_size > RVT_MAX_LKEY_TABLE_BITS) {
		rvt_pr_warn(rdi, "lkey bits %u too large, reduced to %u\n",
			    lkey_table_size, RVT_MAX_LKEY_TABLE_BITS);
		rdi->dparms.lkey_table_size = RVT_MAX_LKEY_TABLE_BITS;
		lkey_table_size = rdi->dparms.lkey_table_size;
	}
	rdi->lkey_table.max = 1 << lkey_table_size;
	rdi->lkey_table.shift = 32 - lkey_table_size;
	lk_tab_size = rdi->lkey_table.max * sizeof(*rdi->lkey_table.table);
	rdi->lkey_table.table = (struct rvt_mregion __rcu **)
			       vmalloc_node(lk_tab_size, rdi->dparms.node);
	if (!rdi->lkey_table.table)
		return -ENOMEM;

	RCU_INIT_POINTER(rdi->dma_mr, NULL);
	for (i = 0; i < rdi->lkey_table.max; i++)
		RCU_INIT_POINTER(rdi->lkey_table.table[i], NULL);

	return 0;
}

/**
 *rvt_mr_exit: clean up MR
 *@rdi: rvt dev structure
 *
 * called when drivers have unregistered or perhaps failed to register with us
 */
void rvt_mr_exit(struct rvt_dev_info *rdi)
{
	if (rdi->dma_mr)
		rvt_pr_err(rdi, "DMA MR not null!\n");

	vfree(rdi->lkey_table.table);
}

static void rvt_deinit_mregion(struct rvt_mregion *mr)
{
	int i = mr->mapsz;

	mr->mapsz = 0;
	while (i)
		kfree(mr->map[--i]);
	percpu_ref_exit(&mr->refcount);
}

static void __rvt_mregion_complete(struct percpu_ref *ref)
{
	struct rvt_mregion *mr = container_of(ref, struct rvt_mregion,
					      refcount);

	complete(&mr->comp);
}

static int rvt_init_mregion(struct rvt_mregion *mr, struct ib_pd *pd,
			    int count, unsigned int percpu_flags)
{
	int m, i = 0;
	struct rvt_dev_info *dev = ib_to_rvt(pd->device);

	mr->mapsz = 0;
	m = (count + RVT_SEGSZ - 1) / RVT_SEGSZ;
	for (; i < m; i++) {
		mr->map[i] = kzalloc_node(sizeof(*mr->map[0]), GFP_KERNEL,
					  dev->dparms.node);
		if (!mr->map[i])
			goto bail;
		mr->mapsz++;
	}
	init_completion(&mr->comp);
	/* count returning the ptr to user */
	if (percpu_ref_init(&mr->refcount, &__rvt_mregion_complete,
			    percpu_flags, GFP_KERNEL))
		goto bail;

	atomic_set(&mr->lkey_invalid, 0);
	mr->pd = pd;
	mr->max_segs = count;
	return 0;
bail:
	rvt_deinit_mregion(mr);
	return -ENOMEM;
}

/**
 * rvt_alloc_lkey - allocate an lkey
 * @mr: memory region that this lkey protects
 * @dma_region: 0->normal key, 1->restricted DMA key
 *
 * Returns 0 if successful, otherwise returns -errno.
 *
 * Increments mr reference count as required.
 *
 * Sets the lkey field mr for non-dma regions.
 *
 */
static int rvt_alloc_lkey(struct rvt_mregion *mr, int dma_region)
{
	unsigned long flags;
	u32 r;
	u32 n;
	int ret = 0;
	struct rvt_dev_info *dev = ib_to_rvt(mr->pd->device);
	struct rvt_lkey_table *rkt = &dev->lkey_table;

	rvt_get_mr(mr);
	spin_lock_irqsave(&rkt->lock, flags);

	/* special case for dma_mr lkey == 0 */
	if (dma_region) {
		struct rvt_mregion *tmr;

		tmr = rcu_access_pointer(dev->dma_mr);
		if (!tmr) {
			mr->lkey_published = 1;
			/* Insure published written first */
			rcu_assign_pointer(dev->dma_mr, mr);
			rvt_get_mr(mr);
		}
		goto success;
	}

	/* Find the next available LKEY */
	r = rkt->next;
	n = r;
	for (;;) {
		if (!rcu_access_pointer(rkt->table[r]))
			break;
		r = (r + 1) & (rkt->max - 1);
		if (r == n)
			goto bail;
	}
	rkt->next = (r + 1) & (rkt->max - 1);
	/*
	 * Make sure lkey is never zero which is reserved to indicate an
	 * unrestricted LKEY.
	 */
	rkt->gen++;
	/*
	 * bits are capped to ensure enough bits for generation number
	 */
	mr->lkey = (r << (32 - dev->dparms.lkey_table_size)) |
		((((1 << (24 - dev->dparms.lkey_table_size)) - 1) & rkt->gen)
		 << 8);
	if (mr->lkey == 0) {
		mr->lkey |= 1 << 8;
		rkt->gen++;
	}
	mr->lkey_published = 1;
	/* Insure published written first */
	rcu_assign_pointer(rkt->table[r], mr);
success:
	spin_unlock_irqrestore(&rkt->lock, flags);
out:
	return ret;
bail:
	rvt_put_mr(mr);
	spin_unlock_irqrestore(&rkt->lock, flags);
	ret = -ENOMEM;
	goto out;
}

/**
 * rvt_free_lkey - free an lkey
 * @mr: mr to free from tables
 */
static void rvt_free_lkey(struct rvt_mregion *mr)
{
	unsigned long flags;
	u32 lkey = mr->lkey;
	u32 r;
	struct rvt_dev_info *dev = ib_to_rvt(mr->pd->device);
	struct rvt_lkey_table *rkt = &dev->lkey_table;
	int freed = 0;

	spin_lock_irqsave(&rkt->lock, flags);
	if (!lkey) {
		if (mr->lkey_published) {
			mr->lkey_published = 0;
			/* insure published is written before pointer */
			rcu_assign_pointer(dev->dma_mr, NULL);
			rvt_put_mr(mr);
		}
	} else {
		if (!mr->lkey_published)
			goto out;
		r = lkey >> (32 - dev->dparms.lkey_table_size);
		mr->lkey_published = 0;
		/* insure published is written before pointer */
		rcu_assign_pointer(rkt->table[r], NULL);
	}
	freed++;
out:
	spin_unlock_irqrestore(&rkt->lock, flags);
	if (freed)
		percpu_ref_kill(&mr->refcount);
}

static struct rvt_mr *__rvt_alloc_mr(int count, struct ib_pd *pd)
{
	struct rvt_mr *mr;
	int rval = -ENOMEM;
	int m;

	/* Allocate struct plus pointers to first level page tables. */
	m = (count + RVT_SEGSZ - 1) / RVT_SEGSZ;
	mr = kzalloc(struct_size(mr, mr.map, m), GFP_KERNEL);
	if (!mr)
		goto bail;

	rval = rvt_init_mregion(&mr->mr, pd, count, 0);
	if (rval)
		goto bail;
	/*
	 * ib_reg_phys_mr() will initialize mr->ibmr except for
	 * lkey and rkey.
	 */
	rval = rvt_alloc_lkey(&mr->mr, 0);
	if (rval)
		goto bail_mregion;
	mr->ibmr.lkey = mr->mr.lkey;
	mr->ibmr.rkey = mr->mr.lkey;
done:
	return mr;

bail_mregion:
	rvt_deinit_mregion(&mr->mr);
bail:
	kfree(mr);
	mr = ERR_PTR(rval);
	goto done;
}

static void __rvt_free_mr(struct rvt_mr *mr)
{
	rvt_free_lkey(&mr->mr);
	rvt_deinit_mregion(&mr->mr);
	kfree(mr);
}

/**
 * rvt_get_dma_mr - get a DMA memory region
 * @pd: protection domain for this memory region
 * @acc: access flags
 *
 * Return: the memory region on success, otherwise returns an errno.
 * Note that all DMA addresses should be created via the functions in
 * struct dma_virt_ops.
 */
struct ib_mr *rvt_get_dma_mr(struct ib_pd *pd, int acc)
{
	struct rvt_mr *mr;
	struct ib_mr *ret;
	int rval;

	if (ibpd_to_rvtpd(pd)->user)
		return ERR_PTR(-EPERM);

	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
	if (!mr) {
		ret = ERR_PTR(-ENOMEM);
		goto bail;
	}

	rval = rvt_init_mregion(&mr->mr, pd, 0, 0);
	if (rval) {
		ret = ERR_PTR(rval);
		goto bail;
	}

	rval = rvt_alloc_lkey(&mr->mr, 1);
	if (rval) {
		ret = ERR_PTR(rval);
		goto bail_mregion;
	}

	mr->mr.access_flags = acc;
	ret = &mr->ibmr;
done:
	return ret;

bail_mregion:
	rvt_deinit_mregion(&mr->mr);
bail:
	kfree(mr);
	goto done;
}

/**
 * rvt_reg_user_mr - register a userspace memory region
 * @pd: protection domain for this memory region
 * @start: starting userspace address
 * @length: length of region to register
 * @mr_access_flags: access flags for this memory region
 * @udata: unused by the driver
 *
 * Return: the memory region on success, otherwise returns an errno.
 */
struct ib_mr *rvt_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
			      u64 virt_addr, int mr_access_flags,
			      struct ib_udata *udata)
{
	struct rvt_mr *mr;
	struct ib_umem *umem;
	struct sg_page_iter sg_iter;
	int n, m;
	struct ib_mr *ret;

	if (length == 0)
		return ERR_PTR(-EINVAL);

	umem = ib_umem_get(udata, start, length, mr_access_flags, 0);
	if (IS_ERR(umem))
		return (void *)umem;

	n = ib_umem_num_pages(umem);

	mr = __rvt_alloc_mr(n, pd);
	if (IS_ERR(mr)) {
		ret = (struct ib_mr *)mr;
		goto bail_umem;
	}

	mr->mr.user_base = start;
	mr->mr.iova = virt_addr;
	mr->mr.length = length;
	mr->mr.offset = ib_umem_offset(umem);
	mr->mr.access_flags = mr_access_flags;
	mr->umem = umem;

	mr->mr.page_shift = PAGE_SHIFT;
	m = 0;
	n = 0;
	for_each_sg_page (umem->sg_head.sgl, &sg_iter, umem->nmap, 0) {
		void *vaddr;

		vaddr = page_address(sg_page_iter_page(&sg_iter));
		if (!vaddr) {
			ret = ERR_PTR(-EINVAL);
			goto bail_inval;
		}
		mr->mr.map[m]->segs[n].vaddr = vaddr;
		mr->mr.map[m]->segs[n].length = PAGE_SIZE;
		trace_rvt_mr_user_seg(&mr->mr, m, n, vaddr, PAGE_SIZE);
		if (++n == RVT_SEGSZ) {
			m++;
			n = 0;
		}
	}
	return &mr->ibmr;

bail_inval:
	__rvt_free_mr(mr);

bail_umem:
	ib_umem_release(umem);

	return ret;
}

/**
 * rvt_dereg_clean_qp_cb - callback from iterator
 * @qp - the qp
 * @v - the mregion (as u64)
 *
 * This routine fields the callback for all QPs and
 * for QPs in the same PD as the MR will call the
 * rvt_qp_mr_clean() to potentially cleanup references.
 */
static void rvt_dereg_clean_qp_cb(struct rvt_qp *qp, u64 v)
{
	struct rvt_mregion *mr = (struct rvt_mregion *)v;

	/* skip PDs that are not ours */
	if (mr->pd != qp->ibqp.pd)
		return;
	rvt_qp_mr_clean(qp, mr->lkey);
}

/**
 * rvt_dereg_clean_qps - find QPs for reference cleanup
 * @mr - the MR that is being deregistered
 *
 * This routine iterates RC QPs looking for references
 * to the lkey noted in mr.
 */
static void rvt_dereg_clean_qps(struct rvt_mregion *mr)
{
	struct rvt_dev_info *rdi = ib_to_rvt(mr->pd->device);

	rvt_qp_iter(rdi, (u64)mr, rvt_dereg_clean_qp_cb);
}

/**
 * rvt_check_refs - check references
 * @mr - the megion
 * @t - the caller identification
 *
 * This routine checks MRs holding a reference during
 * when being de-registered.
 *
 * If the count is non-zero, the code calls a clean routine then
 * waits for the timeout for the count to zero.
 */
static int rvt_check_refs(struct rvt_mregion *mr, const char *t)
{
	unsigned long timeout;
	struct rvt_dev_info *rdi = ib_to_rvt(mr->pd->device);

	if (mr->lkey) {
		/* avoid dma mr */
		rvt_dereg_clean_qps(mr);
		/* @mr was indexed on rcu protected @lkey_table */
		synchronize_rcu();
	}

	timeout = wait_for_completion_timeout(&mr->comp, 5 * HZ);
	if (!timeout) {
		rvt_pr_err(rdi,
			   "%s timeout mr %p pd %p lkey %x refcount %ld\n",
			   t, mr, mr->pd, mr->lkey,
			   atomic_long_read(&mr->refcount.count));
		rvt_get_mr(mr);
		return -EBUSY;
	}
	return 0;
}

/**
 * rvt_mr_has_lkey - is MR
 * @mr - the mregion
 * @lkey - the lkey
 */
bool rvt_mr_has_lkey(struct rvt_mregion *mr, u32 lkey)
{
	return mr && lkey == mr->lkey;
}

/**
 * rvt_ss_has_lkey - is mr in sge tests
 * @ss - the sge state
 * @lkey
 *
 * This code tests for an MR in the indicated
 * sge state.
 */
bool rvt_ss_has_lkey(struct rvt_sge_state *ss, u32 lkey)
{
	int i;
	bool rval = false;

	if (!ss->num_sge)
		return rval;
	/* first one */
	rval = rvt_mr_has_lkey(ss->sge.mr, lkey);
	/* any others */
	for (i = 0; !rval && i < ss->num_sge - 1; i++)
		rval = rvt_mr_has_lkey(ss->sg_list[i].mr, lkey);
	return rval;
}

/**
 * rvt_dereg_mr - unregister and free a memory region
 * @ibmr: the memory region to free
 *
 *
 * Note that this is called to free MRs created by rvt_get_dma_mr()
 * or rvt_reg_user_mr().
 *
 * Returns 0 on success.
 */
int rvt_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata)
{
	struct rvt_mr *mr = to_imr(ibmr);
	int ret;

	rvt_free_lkey(&mr->mr);

	rvt_put_mr(&mr->mr); /* will set completion if last */
	ret = rvt_check_refs(&mr->mr, __func__);
	if (ret)
		goto out;
	rvt_deinit_mregion(&mr->mr);
	if (mr->umem)
		ib_umem_release(mr->umem);
	kfree(mr);
out:
	return ret;
}

/**
 * rvt_alloc_mr - Allocate a memory region usable with the
 * @pd: protection domain for this memory region
 * @mr_type: mem region type
 * @max_num_sg: Max number of segments allowed
 *
 * Return: the memory region on success, otherwise return an errno.
 */
struct ib_mr *rvt_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type,
			   u32 max_num_sg, struct ib_udata *udata)
{
	struct rvt_mr *mr;

	if (mr_type != IB_MR_TYPE_MEM_REG)
		return ERR_PTR(-EINVAL);

	mr = __rvt_alloc_mr(max_num_sg, pd);
	if (IS_ERR(mr))
		return (struct ib_mr *)mr;

	return &mr->ibmr;
}

/**
 * rvt_set_page - page assignment function called by ib_sg_to_pages
 * @ibmr: memory region
 * @addr: dma address of mapped page
 *
 * Return: 0 on success
 */
static int rvt_set_page(struct ib_mr *ibmr, u64 addr)
{
	struct rvt_mr *mr = to_imr(ibmr);
	u32 ps = 1 << mr->mr.page_shift;
	u32 mapped_segs = mr->mr.length >> mr->mr.page_shift;
	int m, n;

	if (unlikely(mapped_segs == mr->mr.max_segs))
		return -ENOMEM;

	m = mapped_segs / RVT_SEGSZ;
	n = mapped_segs % RVT_SEGSZ;
	mr->mr.map[m]->segs[n].vaddr = (void *)addr;
	mr->mr.map[m]->segs[n].length = ps;
	trace_rvt_mr_page_seg(&mr->mr, m, n, (void *)addr, ps);
	mr->mr.length += ps;

	return 0;
}

/**
 * rvt_map_mr_sg - map sg list and set it the memory region
 * @ibmr: memory region
 * @sg: dma mapped scatterlist
 * @sg_nents: number of entries in sg
 * @sg_offset: offset in bytes into sg
 *
 * Overwrite rvt_mr length with mr length calculated by ib_sg_to_pages.
 *
 * Return: number of sg elements mapped to the memory region
 */
int rvt_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg,
		  int sg_nents, unsigned int *sg_offset)
{
	struct rvt_mr *mr = to_imr(ibmr);
	int ret;

	mr->mr.length = 0;
	mr->mr.page_shift = PAGE_SHIFT;
	ret = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, rvt_set_page);
	mr->mr.user_base = ibmr->iova;
	mr->mr.iova = ibmr->iova;
	mr->mr.offset = ibmr->iova - (u64)mr->mr.map[0]->segs[0].vaddr;
	mr->mr.length = (size_t)ibmr->length;
	return ret;
}

/**
 * rvt_fast_reg_mr - fast register physical MR
 * @qp: the queue pair where the work request comes from
 * @ibmr: the memory region to be registered
 * @key: updated key for this memory region
 * @access: access flags for this memory region
 *
 * Returns 0 on success.
 */
int rvt_fast_reg_mr(struct rvt_qp *qp, struct ib_mr *ibmr, u32 key,
		    int access)
{
	struct rvt_mr *mr = to_imr(ibmr);

	if (qp->ibqp.pd != mr->mr.pd)
		return -EACCES;

	/* not applicable to dma MR or user MR */
	if (!mr->mr.lkey || mr->umem)
		return -EINVAL;

	if ((key & 0xFFFFFF00) != (mr->mr.lkey & 0xFFFFFF00))
		return -EINVAL;

	ibmr->lkey = key;
	ibmr->rkey = key;
	mr->mr.lkey = key;
	mr->mr.access_flags = access;
	mr->mr.iova = ibmr->iova;
	atomic_set(&mr->mr.lkey_invalid, 0);

	return 0;
}
EXPORT_SYMBOL(rvt_fast_reg_mr);

/**
 * rvt_invalidate_rkey - invalidate an MR rkey
 * @qp: queue pair associated with the invalidate op
 * @rkey: rkey to invalidate
 *
 * Returns 0 on success.
 */
int rvt_invalidate_rkey(struct rvt_qp *qp, u32 rkey)
{
	struct rvt_dev_info *dev = ib_to_rvt(qp->ibqp.device);
	struct rvt_lkey_table *rkt = &dev->lkey_table;
	struct rvt_mregion *mr;

	if (rkey == 0)
		return -EINVAL;

	rcu_read_lock();
	mr = rcu_dereference(
		rkt->table[(rkey >> (32 - dev->dparms.lkey_table_size))]);
	if (unlikely(!mr || mr->lkey != rkey || qp->ibqp.pd != mr->pd))
		goto bail;

	atomic_set(&mr->lkey_invalid, 1);
	rcu_read_unlock();
	return 0;

bail:
	rcu_read_unlock();
	return -EINVAL;
}
EXPORT_SYMBOL(rvt_invalidate_rkey);

/**
 * rvt_alloc_fmr - allocate a fast memory region
 * @pd: the protection domain for this memory region
 * @mr_access_flags: access flags for this memory region
 * @fmr_attr: fast memory region attributes
 *
 * Return: the memory region on success, otherwise returns an errno.
 */
struct ib_fmr *rvt_alloc_fmr(struct ib_pd *pd, int mr_access_flags,
			     struct ib_fmr_attr *fmr_attr)
{
	struct rvt_fmr *fmr;
	int m;
	struct ib_fmr *ret;
	int rval = -ENOMEM;

	/* Allocate struct plus pointers to first level page tables. */
	m = (fmr_attr->max_pages + RVT_SEGSZ - 1) / RVT_SEGSZ;
	fmr = kzalloc(struct_size(fmr, mr.map, m), GFP_KERNEL);
	if (!fmr)
		goto bail;

	rval = rvt_init_mregion(&fmr->mr, pd, fmr_attr->max_pages,
				PERCPU_REF_INIT_ATOMIC);
	if (rval)
		goto bail;

	/*
	 * ib_alloc_fmr() will initialize fmr->ibfmr except for lkey &
	 * rkey.
	 */
	rval = rvt_alloc_lkey(&fmr->mr, 0);
	if (rval)
		goto bail_mregion;
	fmr->ibfmr.rkey = fmr->mr.lkey;
	fmr->ibfmr.lkey = fmr->mr.lkey;
	/*
	 * Resources are allocated but no valid mapping (RKEY can't be
	 * used).
	 */
	fmr->mr.access_flags = mr_access_flags;
	fmr->mr.max_segs = fmr_attr->max_pages;
	fmr->mr.page_shift = fmr_attr->page_shift;

	ret = &fmr->ibfmr;
done:
	return ret;

bail_mregion:
	rvt_deinit_mregion(&fmr->mr);
bail:
	kfree(fmr);
	ret = ERR_PTR(rval);
	goto done;
}

/**
 * rvt_map_phys_fmr - set up a fast memory region
 * @ibfmr: the fast memory region to set up
 * @page_list: the list of pages to associate with the fast memory region
 * @list_len: the number of pages to associate with the fast memory region
 * @iova: the virtual address of the start of the fast memory region
 *
 * This may be called from interrupt context.
 *
 * Return: 0 on success
 */

int rvt_map_phys_fmr(struct ib_fmr *ibfmr, u64 *page_list,
		     int list_len, u64 iova)
{
	struct rvt_fmr *fmr = to_ifmr(ibfmr);
	struct rvt_lkey_table *rkt;
	unsigned long flags;
	int m, n;
	unsigned long i;
	u32 ps;
	struct rvt_dev_info *rdi = ib_to_rvt(ibfmr->device);

	i = atomic_long_read(&fmr->mr.refcount.count);
	if (i > 2)
		return -EBUSY;

	if (list_len > fmr->mr.max_segs)
		return -EINVAL;

	rkt = &rdi->lkey_table;
	spin_lock_irqsave(&rkt->lock, flags);
	fmr->mr.user_base = iova;
	fmr->mr.iova = iova;
	ps = 1 << fmr->mr.page_shift;
	fmr->mr.length = list_len * ps;
	m = 0;
	n = 0;
	for (i = 0; i < list_len; i++) {
		fmr->mr.map[m]->segs[n].vaddr = (void *)page_list[i];
		fmr->mr.map[m]->segs[n].length = ps;
		trace_rvt_mr_fmr_seg(&fmr->mr, m, n, (void *)page_list[i], ps);
		if (++n == RVT_SEGSZ) {
			m++;
			n = 0;
		}
	}
	spin_unlock_irqrestore(&rkt->lock, flags);
	return 0;
}

/**
 * rvt_unmap_fmr - unmap fast memory regions
 * @fmr_list: the list of fast memory regions to unmap
 *
 * Return: 0 on success.
 */
int rvt_unmap_fmr(struct list_head *fmr_list)
{
	struct rvt_fmr *fmr;
	struct rvt_lkey_table *rkt;
	unsigned long flags;
	struct rvt_dev_info *rdi;

	list_for_each_entry(fmr, fmr_list, ibfmr.list) {
		rdi = ib_to_rvt(fmr->ibfmr.device);
		rkt = &rdi->lkey_table;
		spin_lock_irqsave(&rkt->lock, flags);
		fmr->mr.user_base = 0;
		fmr->mr.iova = 0;
		fmr->mr.length = 0;
		spin_unlock_irqrestore(&rkt->lock, flags);
	}
	return 0;
}

/**
 * rvt_dealloc_fmr - deallocate a fast memory region
 * @ibfmr: the fast memory region to deallocate
 *
 * Return: 0 on success.
 */
int rvt_dealloc_fmr(struct ib_fmr *ibfmr)
{
	struct rvt_fmr *fmr = to_ifmr(ibfmr);
	int ret = 0;

	rvt_free_lkey(&fmr->mr);
	rvt_put_mr(&fmr->mr); /* will set completion if last */
	ret = rvt_check_refs(&fmr->mr, __func__);
	if (ret)
		goto out;
	rvt_deinit_mregion(&fmr->mr);
	kfree(fmr);
out:
	return ret;
}

/**
 * rvt_sge_adjacent - is isge compressible
 * @last_sge: last outgoing SGE written
 * @sge: SGE to check
 *
 * If adjacent will update last_sge to add length.
 *
 * Return: true if isge is adjacent to last sge
 */
static inline bool rvt_sge_adjacent(struct rvt_sge *last_sge,
				    struct ib_sge *sge)
{
	if (last_sge && sge->lkey == last_sge->mr->lkey &&
	    ((uint64_t)(last_sge->vaddr + last_sge->length) == sge->addr)) {
		if (sge->lkey) {
			if (unlikely((sge->addr - last_sge->mr->user_base +
			      sge->length > last_sge->mr->length)))
				return false; /* overrun, caller will catch */
		} else {
			last_sge->length += sge->length;
		}
		last_sge->sge_length += sge->length;
		trace_rvt_sge_adjacent(last_sge, sge);
		return true;
	}
	return false;
}

/**
 * rvt_lkey_ok - check IB SGE for validity and initialize
 * @rkt: table containing lkey to check SGE against
 * @pd: protection domain
 * @isge: outgoing internal SGE
 * @last_sge: last outgoing SGE written
 * @sge: SGE to check
 * @acc: access flags
 *
 * Check the IB SGE for validity and initialize our internal version
 * of it.
 *
 * Increments the reference count when a new sge is stored.
 *
 * Return: 0 if compressed, 1 if added , otherwise returns -errno.
 */
int rvt_lkey_ok(struct rvt_lkey_table *rkt, struct rvt_pd *pd,
		struct rvt_sge *isge, struct rvt_sge *last_sge,
		struct ib_sge *sge, int acc)
{
	struct rvt_mregion *mr;
	unsigned n, m;
	size_t off;

	/*
	 * We use LKEY == zero for kernel virtual addresses
	 * (see rvt_get_dma_mr() and dma_virt_ops).
	 */
	if (sge->lkey == 0) {
		struct rvt_dev_info *dev = ib_to_rvt(pd->ibpd.device);

		if (pd->user)
			return -EINVAL;
		if (rvt_sge_adjacent(last_sge, sge))
			return 0;
		rcu_read_lock();
		mr = rcu_dereference(dev->dma_mr);
		if (!mr)
			goto bail;
		rvt_get_mr(mr);
		rcu_read_unlock();

		isge->mr = mr;
		isge->vaddr = (void *)sge->addr;
		isge->length = sge->length;
		isge->sge_length = sge->length;
		isge->m = 0;
		isge->n = 0;
		goto ok;
	}
	if (rvt_sge_adjacent(last_sge, sge))
		return 0;
	rcu_read_lock();
	mr = rcu_dereference(rkt->table[sge->lkey >> rkt->shift]);
	if (!mr)
		goto bail;
	rvt_get_mr(mr);
	if (!READ_ONCE(mr->lkey_published))
		goto bail_unref;

	if (unlikely(atomic_read(&mr->lkey_invalid) ||
		     mr->lkey != sge->lkey || mr->pd != &pd->ibpd))
		goto bail_unref;

	off = sge->addr - mr->user_base;
	if (unlikely(sge->addr < mr->user_base ||
		     off + sge->length > mr->length ||
		     (mr->access_flags & acc) != acc))
		goto bail_unref;
	rcu_read_unlock();

	off += mr->offset;
	if (mr->page_shift) {
		/*
		 * page sizes are uniform power of 2 so no loop is necessary
		 * entries_spanned_by_off is the number of times the loop below
		 * would have executed.
		*/
		size_t entries_spanned_by_off;

		entries_spanned_by_off = off >> mr->page_shift;
		off -= (entries_spanned_by_off << mr->page_shift);
		m = entries_spanned_by_off / RVT_SEGSZ;
		n = entries_spanned_by_off % RVT_SEGSZ;
	} else {
		m = 0;
		n = 0;
		while (off >= mr->map[m]->segs[n].length) {
			off -= mr->map[m]->segs[n].length;
			n++;
			if (n >= RVT_SEGSZ) {
				m++;
				n = 0;
			}
		}
	}
	isge->mr = mr;
	isge->vaddr = mr->map[m]->segs[n].vaddr + off;
	isge->length = mr->map[m]->segs[n].length - off;
	isge->sge_length = sge->length;
	isge->m = m;
	isge->n = n;
ok:
	trace_rvt_sge_new(isge, sge);
	return 1;
bail_unref:
	rvt_put_mr(mr);
bail:
	rcu_read_unlock();
	return -EINVAL;
}
EXPORT_SYMBOL(rvt_lkey_ok);

/**
 * rvt_rkey_ok - check the IB virtual address, length, and RKEY
 * @qp: qp for validation
 * @sge: SGE state
 * @len: length of data
 * @vaddr: virtual address to place data
 * @rkey: rkey to check
 * @acc: access flags
 *
 * Return: 1 if successful, otherwise 0.
 *
 * increments the reference count upon success
 */
int rvt_rkey_ok(struct rvt_qp *qp, struct rvt_sge *sge,
		u32 len, u64 vaddr, u32 rkey, int acc)
{
	struct rvt_dev_info *dev = ib_to_rvt(qp->ibqp.device);
	struct rvt_lkey_table *rkt = &dev->lkey_table;
	struct rvt_mregion *mr;
	unsigned n, m;
	size_t off;

	/*
	 * We use RKEY == zero for kernel virtual addresses
	 * (see rvt_get_dma_mr() and dma_virt_ops).
	 */
	rcu_read_lock();
	if (rkey == 0) {
		struct rvt_pd *pd = ibpd_to_rvtpd(qp->ibqp.pd);
		struct rvt_dev_info *rdi = ib_to_rvt(pd->ibpd.device);

		if (pd->user)
			goto bail;
		mr = rcu_dereference(rdi->dma_mr);
		if (!mr)
			goto bail;
		rvt_get_mr(mr);
		rcu_read_unlock();

		sge->mr = mr;
		sge->vaddr = (void *)vaddr;
		sge->length = len;
		sge->sge_length = len;
		sge->m = 0;
		sge->n = 0;
		goto ok;
	}

	mr = rcu_dereference(rkt->table[rkey >> rkt->shift]);
	if (!mr)
		goto bail;
	rvt_get_mr(mr);
	/* insure mr read is before test */
	if (!READ_ONCE(mr->lkey_published))
		goto bail_unref;
	if (unlikely(atomic_read(&mr->lkey_invalid) ||
		     mr->lkey != rkey || qp->ibqp.pd != mr->pd))
		goto bail_unref;

	off = vaddr - mr->iova;
	if (unlikely(vaddr < mr->iova || off + len > mr->length ||
		     (mr->access_flags & acc) == 0))
		goto bail_unref;
	rcu_read_unlock();

	off += mr->offset;
	if (mr->page_shift) {
		/*
		 * page sizes are uniform power of 2 so no loop is necessary
		 * entries_spanned_by_off is the number of times the loop below
		 * would have executed.
		*/
		size_t entries_spanned_by_off;

		entries_spanned_by_off = off >> mr->page_shift;
		off -= (entries_spanned_by_off << mr->page_shift);
		m = entries_spanned_by_off / RVT_SEGSZ;
		n = entries_spanned_by_off % RVT_SEGSZ;
	} else {
		m = 0;
		n = 0;
		while (off >= mr->map[m]->segs[n].length) {
			off -= mr->map[m]->segs[n].length;
			n++;
			if (n >= RVT_SEGSZ) {
				m++;
				n = 0;
			}
		}
	}
	sge->mr = mr;
	sge->vaddr = mr->map[m]->segs[n].vaddr + off;
	sge->length = mr->map[m]->segs[n].length - off;
	sge->sge_length = len;
	sge->m = m;
	sge->n = n;
ok:
	return 1;
bail_unref:
	rvt_put_mr(mr);
bail:
	rcu_read_unlock();
	return 0;
}
EXPORT_SYMBOL(rvt_rkey_ok);