summaryrefslogtreecommitdiffstats
path: root/fs/nfs/filelayout/filelayout.c
blob: a46bf6de9ce455a97ef18275efe4906ac10e12ed (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
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
/*
 *  Module for the pnfs nfs4 file layout driver.
 *  Defines all I/O and Policy interface operations, plus code
 *  to register itself with the pNFS client.
 *
 *  Copyright (c) 2002
 *  The Regents of the University of Michigan
 *  All Rights Reserved
 *
 *  Dean Hildebrand <dhildebz@umich.edu>
 *
 *  Permission is granted to use, copy, create derivative works, and
 *  redistribute this software and such derivative works for any purpose,
 *  so long as the name of the University of Michigan is not used in
 *  any advertising or publicity pertaining to the use or distribution
 *  of this software without specific, written prior authorization. If
 *  the above copyright notice or any other identification of the
 *  University of Michigan is included in any copy of any portion of
 *  this software, then the disclaimer below must also be included.
 *
 *  This software is provided as is, without representation or warranty
 *  of any kind either express or implied, including without limitation
 *  the implied warranties of merchantability, fitness for a particular
 *  purpose, or noninfringement.  The Regents of the University of
 *  Michigan shall not be liable for any damages, including special,
 *  indirect, incidental, or consequential damages, with respect to any
 *  claim arising out of or in connection with the use of the software,
 *  even if it has been or is hereafter advised of the possibility of
 *  such damages.
 */

#include <linux/nfs_fs.h>
#include <linux/nfs_page.h>
#include <linux/module.h>

#include <linux/sunrpc/metrics.h>

#include "../nfs4session.h"
#include "../internal.h"
#include "../delegation.h"
#include "filelayout.h"
#include "../nfs4trace.h"

#define NFSDBG_FACILITY         NFSDBG_PNFS_LD

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Dean Hildebrand <dhildebz@umich.edu>");
MODULE_DESCRIPTION("The NFSv4 file layout driver");

#define FILELAYOUT_POLL_RETRY_MAX     (15*HZ)

static loff_t
filelayout_get_dense_offset(struct nfs4_filelayout_segment *flseg,
			    loff_t offset)
{
	u32 stripe_width = flseg->stripe_unit * flseg->dsaddr->stripe_count;
	u64 stripe_no;
	u32 rem;

	offset -= flseg->pattern_offset;
	stripe_no = div_u64(offset, stripe_width);
	div_u64_rem(offset, flseg->stripe_unit, &rem);

	return stripe_no * flseg->stripe_unit + rem;
}

/* This function is used by the layout driver to calculate the
 * offset of the file on the dserver based on whether the
 * layout type is STRIPE_DENSE or STRIPE_SPARSE
 */
static loff_t
filelayout_get_dserver_offset(struct pnfs_layout_segment *lseg, loff_t offset)
{
	struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg);

	switch (flseg->stripe_type) {
	case STRIPE_SPARSE:
		return offset;

	case STRIPE_DENSE:
		return filelayout_get_dense_offset(flseg, offset);
	}

	BUG();
}

static void filelayout_reset_write(struct nfs_pgio_header *hdr)
{
	struct rpc_task *task = &hdr->task;

	if (!test_and_set_bit(NFS_IOHDR_REDO, &hdr->flags)) {
		dprintk("%s Reset task %5u for i/o through MDS "
			"(req %s/%llu, %u bytes @ offset %llu)\n", __func__,
			hdr->task.tk_pid,
			hdr->inode->i_sb->s_id,
			(unsigned long long)NFS_FILEID(hdr->inode),
			hdr->args.count,
			(unsigned long long)hdr->args.offset);

		task->tk_status = pnfs_write_done_resend_to_mds(hdr);
	}
}

static void filelayout_reset_read(struct nfs_pgio_header *hdr)
{
	struct rpc_task *task = &hdr->task;

	if (!test_and_set_bit(NFS_IOHDR_REDO, &hdr->flags)) {
		dprintk("%s Reset task %5u for i/o through MDS "
			"(req %s/%llu, %u bytes @ offset %llu)\n", __func__,
			hdr->task.tk_pid,
			hdr->inode->i_sb->s_id,
			(unsigned long long)NFS_FILEID(hdr->inode),
			hdr->args.count,
			(unsigned long long)hdr->args.offset);

		task->tk_status = pnfs_read_done_resend_to_mds(hdr);
	}
}

static int filelayout_async_handle_error(struct rpc_task *task,
					 struct nfs4_state *state,
					 struct nfs_client *clp,
					 struct pnfs_layout_segment *lseg)
{
	struct pnfs_layout_hdr *lo = lseg->pls_layout;
	struct inode *inode = lo->plh_inode;
	struct nfs_server *mds_server = NFS_SERVER(inode);
	struct nfs4_deviceid_node *devid = FILELAYOUT_DEVID_NODE(lseg);
	struct nfs_client *mds_client = mds_server->nfs_client;
	struct nfs4_slot_table *tbl = &clp->cl_session->fc_slot_table;

	if (task->tk_status >= 0)
		return 0;

	switch (task->tk_status) {
	/* MDS state errors */
	case -NFS4ERR_DELEG_REVOKED:
	case -NFS4ERR_ADMIN_REVOKED:
	case -NFS4ERR_BAD_STATEID:
	case -NFS4ERR_OPENMODE:
		if (state == NULL)
			break;
		if (nfs4_schedule_stateid_recovery(mds_server, state) < 0)
			goto out_bad_stateid;
		goto wait_on_recovery;
	case -NFS4ERR_EXPIRED:
		if (state != NULL) {
			if (nfs4_schedule_stateid_recovery(mds_server, state) < 0)
				goto out_bad_stateid;
		}
		nfs4_schedule_lease_recovery(mds_client);
		goto wait_on_recovery;
	/* DS session errors */
	case -NFS4ERR_BADSESSION:
	case -NFS4ERR_BADSLOT:
	case -NFS4ERR_BAD_HIGH_SLOT:
	case -NFS4ERR_DEADSESSION:
	case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
	case -NFS4ERR_SEQ_FALSE_RETRY:
	case -NFS4ERR_SEQ_MISORDERED:
		dprintk("%s ERROR %d, Reset session. Exchangeid "
			"flags 0x%x\n", __func__, task->tk_status,
			clp->cl_exchange_flags);
		nfs4_schedule_session_recovery(clp->cl_session, task->tk_status);
		break;
	case -NFS4ERR_DELAY:
	case -NFS4ERR_GRACE:
		rpc_delay(task, FILELAYOUT_POLL_RETRY_MAX);
		break;
	case -NFS4ERR_RETRY_UNCACHED_REP:
		break;
	/* Invalidate Layout errors */
	case -NFS4ERR_PNFS_NO_LAYOUT:
	case -ESTALE:           /* mapped NFS4ERR_STALE */
	case -EBADHANDLE:       /* mapped NFS4ERR_BADHANDLE */
	case -EISDIR:           /* mapped NFS4ERR_ISDIR */
	case -NFS4ERR_FHEXPIRED:
	case -NFS4ERR_WRONG_TYPE:
		dprintk("%s Invalid layout error %d\n", __func__,
			task->tk_status);
		/*
		 * Destroy layout so new i/o will get a new layout.
		 * Layout will not be destroyed until all current lseg
		 * references are put. Mark layout as invalid to resend failed
		 * i/o and all i/o waiting on the slot table to the MDS until
		 * layout is destroyed and a new valid layout is obtained.
		 */
		pnfs_destroy_layout(NFS_I(inode));
		rpc_wake_up(&tbl->slot_tbl_waitq);
		goto reset;
	/* RPC connection errors */
	case -ECONNREFUSED:
	case -EHOSTDOWN:
	case -EHOSTUNREACH:
	case -ENETUNREACH:
	case -EIO:
	case -ETIMEDOUT:
	case -EPIPE:
		dprintk("%s DS connection error %d\n", __func__,
			task->tk_status);
		nfs4_mark_deviceid_unavailable(devid);
		pnfs_error_mark_layout_for_return(inode, lseg);
		rpc_wake_up(&tbl->slot_tbl_waitq);
		/* fall through */
	default:
reset:
		dprintk("%s Retry through MDS. Error %d\n", __func__,
			task->tk_status);
		return -NFS4ERR_RESET_TO_MDS;
	}
out:
	task->tk_status = 0;
	return -EAGAIN;
out_bad_stateid:
	task->tk_status = -EIO;
	return 0;
wait_on_recovery:
	rpc_sleep_on(&mds_client->cl_rpcwaitq, task, NULL);
	if (test_bit(NFS4CLNT_MANAGER_RUNNING, &mds_client->cl_state) == 0)
		rpc_wake_up_queued_task(&mds_client->cl_rpcwaitq, task);
	goto out;
}

/* NFS_PROTO call done callback routines */

static int filelayout_read_done_cb(struct rpc_task *task,
				struct nfs_pgio_header *hdr)
{
	int err;

	trace_nfs4_pnfs_read(hdr, task->tk_status);
	err = filelayout_async_handle_error(task, hdr->args.context->state,
					    hdr->ds_clp, hdr->lseg);

	switch (err) {
	case -NFS4ERR_RESET_TO_MDS:
		filelayout_reset_read(hdr);
		return task->tk_status;
	case -EAGAIN:
		rpc_restart_call_prepare(task);
		return -EAGAIN;
	}

	return 0;
}

/*
 * We reference the rpc_cred of the first WRITE that triggers the need for
 * a LAYOUTCOMMIT, and use it to send the layoutcommit compound.
 * rfc5661 is not clear about which credential should be used.
 */
static void
filelayout_set_layoutcommit(struct nfs_pgio_header *hdr)
{

	if (FILELAYOUT_LSEG(hdr->lseg)->commit_through_mds ||
	    hdr->res.verf->committed != NFS_DATA_SYNC)
		return;

	pnfs_set_layoutcommit(hdr->inode, hdr->lseg,
			hdr->mds_offset + hdr->res.count);
	dprintk("%s inode %lu pls_end_pos %lu\n", __func__, hdr->inode->i_ino,
		(unsigned long) NFS_I(hdr->inode)->layout->plh_lwb);
}

bool
filelayout_test_devid_unavailable(struct nfs4_deviceid_node *node)
{
	return filelayout_test_devid_invalid(node) ||
		nfs4_test_deviceid_unavailable(node);
}

static bool
filelayout_reset_to_mds(struct pnfs_layout_segment *lseg)
{
	struct nfs4_deviceid_node *node = FILELAYOUT_DEVID_NODE(lseg);

	return filelayout_test_devid_unavailable(node);
}

/*
 * Call ops for the async read/write cases
 * In the case of dense layouts, the offset needs to be reset to its
 * original value.
 */
static void filelayout_read_prepare(struct rpc_task *task, void *data)
{
	struct nfs_pgio_header *hdr = data;

	if (unlikely(test_bit(NFS_CONTEXT_BAD, &hdr->args.context->flags))) {
		rpc_exit(task, -EIO);
		return;
	}
	if (filelayout_reset_to_mds(hdr->lseg)) {
		dprintk("%s task %u reset io to MDS\n", __func__, task->tk_pid);
		filelayout_reset_read(hdr);
		rpc_exit(task, 0);
		return;
	}
	hdr->pgio_done_cb = filelayout_read_done_cb;

	if (nfs41_setup_sequence(hdr->ds_clp->cl_session,
			&hdr->args.seq_args,
			&hdr->res.seq_res,
			task))
		return;
	if (nfs4_set_rw_stateid(&hdr->args.stateid, hdr->args.context,
			hdr->args.lock_context, FMODE_READ) == -EIO)
		rpc_exit(task, -EIO); /* lost lock, terminate I/O */
}

static void filelayout_read_call_done(struct rpc_task *task, void *data)
{
	struct nfs_pgio_header *hdr = data;

	dprintk("--> %s task->tk_status %d\n", __func__, task->tk_status);

	if (test_bit(NFS_IOHDR_REDO, &hdr->flags) &&
	    task->tk_status == 0) {
		nfs41_sequence_done(task, &hdr->res.seq_res);
		return;
	}

	/* Note this may cause RPC to be resent */
	hdr->mds_ops->rpc_call_done(task, data);
}

static void filelayout_read_count_stats(struct rpc_task *task, void *data)
{
	struct nfs_pgio_header *hdr = data;

	rpc_count_iostats(task, NFS_SERVER(hdr->inode)->client->cl_metrics);
}

static int filelayout_write_done_cb(struct rpc_task *task,
				struct nfs_pgio_header *hdr)
{
	int err;

	trace_nfs4_pnfs_write(hdr, task->tk_status);
	err = filelayout_async_handle_error(task, hdr->args.context->state,
					    hdr->ds_clp, hdr->lseg);

	switch (err) {
	case -NFS4ERR_RESET_TO_MDS:
		filelayout_reset_write(hdr);
		return task->tk_status;
	case -EAGAIN:
		rpc_restart_call_prepare(task);
		return -EAGAIN;
	}

	filelayout_set_layoutcommit(hdr);
	return 0;
}

static int filelayout_commit_done_cb(struct rpc_task *task,
				     struct nfs_commit_data *data)
{
	int err;

	trace_nfs4_pnfs_commit_ds(data, task->tk_status);
	err = filelayout_async_handle_error(task, NULL, data->ds_clp,
					    data->lseg);

	switch (err) {
	case -NFS4ERR_RESET_TO_MDS:
		pnfs_generic_prepare_to_resend_writes(data);
		return -EAGAIN;
	case -EAGAIN:
		rpc_restart_call_prepare(task);
		return -EAGAIN;
	}

	if (data->verf.committed == NFS_UNSTABLE)
		pnfs_set_layoutcommit(data->inode, data->lseg, data->lwb);

	return 0;
}

static void filelayout_write_prepare(struct rpc_task *task, void *data)
{
	struct nfs_pgio_header *hdr = data;

	if (unlikely(test_bit(NFS_CONTEXT_BAD, &hdr->args.context->flags))) {
		rpc_exit(task, -EIO);
		return;
	}
	if (filelayout_reset_to_mds(hdr->lseg)) {
		dprintk("%s task %u reset io to MDS\n", __func__, task->tk_pid);
		filelayout_reset_write(hdr);
		rpc_exit(task, 0);
		return;
	}
	if (nfs41_setup_sequence(hdr->ds_clp->cl_session,
			&hdr->args.seq_args,
			&hdr->res.seq_res,
			task))
		return;
	if (nfs4_set_rw_stateid(&hdr->args.stateid, hdr->args.context,
			hdr->args.lock_context, FMODE_WRITE) == -EIO)
		rpc_exit(task, -EIO); /* lost lock, terminate I/O */
}

static void filelayout_write_call_done(struct rpc_task *task, void *data)
{
	struct nfs_pgio_header *hdr = data;

	if (test_bit(NFS_IOHDR_REDO, &hdr->flags) &&
	    task->tk_status == 0) {
		nfs41_sequence_done(task, &hdr->res.seq_res);
		return;
	}

	/* Note this may cause RPC to be resent */
	hdr->mds_ops->rpc_call_done(task, data);
}

static void filelayout_write_count_stats(struct rpc_task *task, void *data)
{
	struct nfs_pgio_header *hdr = data;

	rpc_count_iostats(task, NFS_SERVER(hdr->inode)->client->cl_metrics);
}

static void filelayout_commit_prepare(struct rpc_task *task, void *data)
{
	struct nfs_commit_data *wdata = data;

	nfs41_setup_sequence(wdata->ds_clp->cl_session,
			&wdata->args.seq_args,
			&wdata->res.seq_res,
			task);
}

static void filelayout_commit_count_stats(struct rpc_task *task, void *data)
{
	struct nfs_commit_data *cdata = data;

	rpc_count_iostats(task, NFS_SERVER(cdata->inode)->client->cl_metrics);
}

static const struct rpc_call_ops filelayout_read_call_ops = {
	.rpc_call_prepare = filelayout_read_prepare,
	.rpc_call_done = filelayout_read_call_done,
	.rpc_count_stats = filelayout_read_count_stats,
	.rpc_release = pnfs_generic_rw_release,
};

static const struct rpc_call_ops filelayout_write_call_ops = {
	.rpc_call_prepare = filelayout_write_prepare,
	.rpc_call_done = filelayout_write_call_done,
	.rpc_count_stats = filelayout_write_count_stats,
	.rpc_release = pnfs_generic_rw_release,
};

static const struct rpc_call_ops filelayout_commit_call_ops = {
	.rpc_call_prepare = filelayout_commit_prepare,
	.rpc_call_done = pnfs_generic_write_commit_done,
	.rpc_count_stats = filelayout_commit_count_stats,
	.rpc_release = pnfs_generic_commit_release,
};

static enum pnfs_try_status
filelayout_read_pagelist(struct nfs_pgio_header *hdr)
{
	struct pnfs_layout_segment *lseg = hdr->lseg;
	struct nfs4_pnfs_ds *ds;
	struct rpc_clnt *ds_clnt;
	loff_t offset = hdr->args.offset;
	u32 j, idx;
	struct nfs_fh *fh;

	dprintk("--> %s ino %lu pgbase %u req %Zu@%llu\n",
		__func__, hdr->inode->i_ino,
		hdr->args.pgbase, (size_t)hdr->args.count, offset);

	/* Retrieve the correct rpc_client for the byte range */
	j = nfs4_fl_calc_j_index(lseg, offset);
	idx = nfs4_fl_calc_ds_index(lseg, j);
	ds = nfs4_fl_prepare_ds(lseg, idx);
	if (!ds)
		return PNFS_NOT_ATTEMPTED;

	ds_clnt = nfs4_find_or_create_ds_client(ds->ds_clp, hdr->inode);
	if (IS_ERR(ds_clnt))
		return PNFS_NOT_ATTEMPTED;

	dprintk("%s USE DS: %s cl_count %d\n", __func__,
		ds->ds_remotestr, atomic_read(&ds->ds_clp->cl_count));

	/* No multipath support. Use first DS */
	atomic_inc(&ds->ds_clp->cl_count);
	hdr->ds_clp = ds->ds_clp;
	hdr->ds_commit_idx = idx;
	fh = nfs4_fl_select_ds_fh(lseg, j);
	if (fh)
		hdr->args.fh = fh;

	hdr->args.offset = filelayout_get_dserver_offset(lseg, offset);
	hdr->mds_offset = offset;

	/* Perform an asynchronous read to ds */
	nfs_initiate_pgio(ds_clnt, hdr, hdr->cred,
			  NFS_PROTO(hdr->inode), &filelayout_read_call_ops,
			  0, RPC_TASK_SOFTCONN);
	return PNFS_ATTEMPTED;
}

/* Perform async writes. */
static enum pnfs_try_status
filelayout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
{
	struct pnfs_layout_segment *lseg = hdr->lseg;
	struct nfs4_pnfs_ds *ds;
	struct rpc_clnt *ds_clnt;
	loff_t offset = hdr->args.offset;
	u32 j, idx;
	struct nfs_fh *fh;

	/* Retrieve the correct rpc_client for the byte range */
	j = nfs4_fl_calc_j_index(lseg, offset);
	idx = nfs4_fl_calc_ds_index(lseg, j);
	ds = nfs4_fl_prepare_ds(lseg, idx);
	if (!ds)
		return PNFS_NOT_ATTEMPTED;

	ds_clnt = nfs4_find_or_create_ds_client(ds->ds_clp, hdr->inode);
	if (IS_ERR(ds_clnt))
		return PNFS_NOT_ATTEMPTED;

	dprintk("%s ino %lu sync %d req %Zu@%llu DS: %s cl_count %d\n",
		__func__, hdr->inode->i_ino, sync, (size_t) hdr->args.count,
		offset, ds->ds_remotestr, atomic_read(&ds->ds_clp->cl_count));

	hdr->pgio_done_cb = filelayout_write_done_cb;
	atomic_inc(&ds->ds_clp->cl_count);
	hdr->ds_clp = ds->ds_clp;
	hdr->ds_commit_idx = idx;
	fh = nfs4_fl_select_ds_fh(lseg, j);
	if (fh)
		hdr->args.fh = fh;
	hdr->args.offset = filelayout_get_dserver_offset(lseg, offset);

	/* Perform an asynchronous write */
	nfs_initiate_pgio(ds_clnt, hdr, hdr->cred,
			  NFS_PROTO(hdr->inode), &filelayout_write_call_ops,
			  sync, RPC_TASK_SOFTCONN);
	return PNFS_ATTEMPTED;
}

/*
 * filelayout_check_layout()
 *
 * Make sure layout segment parameters are sane WRT the device.
 * At this point no generic layer initialization of the lseg has occurred,
 * and nothing has been added to the layout_hdr cache.
 *
 */
static int
filelayout_check_layout(struct pnfs_layout_hdr *lo,
			struct nfs4_filelayout_segment *fl,
			struct nfs4_layoutget_res *lgr,
			struct nfs4_deviceid *id,
			gfp_t gfp_flags)
{
	struct nfs4_deviceid_node *d;
	struct nfs4_file_layout_dsaddr *dsaddr;
	int status = -EINVAL;

	dprintk("--> %s\n", __func__);

	/* FIXME: remove this check when layout segment support is added */
	if (lgr->range.offset != 0 ||
	    lgr->range.length != NFS4_MAX_UINT64) {
		dprintk("%s Only whole file layouts supported. Use MDS i/o\n",
			__func__);
		goto out;
	}

	if (fl->pattern_offset > lgr->range.offset) {
		dprintk("%s pattern_offset %lld too large\n",
				__func__, fl->pattern_offset);
		goto out;
	}

	if (!fl->stripe_unit) {
		dprintk("%s Invalid stripe unit (%u)\n",
			__func__, fl->stripe_unit);
		goto out;
	}

	/* find and reference the deviceid */
	d = nfs4_find_get_deviceid(NFS_SERVER(lo->plh_inode), id,
			lo->plh_lc_cred, gfp_flags);
	if (d == NULL)
		goto out;

	dsaddr = container_of(d, struct nfs4_file_layout_dsaddr, id_node);
	/* Found deviceid is unavailable */
	if (filelayout_test_devid_unavailable(&dsaddr->id_node))
		goto out_put;

	fl->dsaddr = dsaddr;

	if (fl->first_stripe_index >= dsaddr->stripe_count) {
		dprintk("%s Bad first_stripe_index %u\n",
				__func__, fl->first_stripe_index);
		goto out_put;
	}

	if ((fl->stripe_type == STRIPE_SPARSE &&
	    fl->num_fh > 1 && fl->num_fh != dsaddr->ds_num) ||
	    (fl->stripe_type == STRIPE_DENSE &&
	    fl->num_fh != dsaddr->stripe_count)) {
		dprintk("%s num_fh %u not valid for given packing\n",
			__func__, fl->num_fh);
		goto out_put;
	}

	status = 0;
out:
	dprintk("--> %s returns %d\n", __func__, status);
	return status;
out_put:
	nfs4_fl_put_deviceid(dsaddr);
	goto out;
}

static void filelayout_free_fh_array(struct nfs4_filelayout_segment *fl)
{
	int i;

	for (i = 0; i < fl->num_fh; i++) {
		if (!fl->fh_array[i])
			break;
		kfree(fl->fh_array[i]);
	}
	kfree(fl->fh_array);
	fl->fh_array = NULL;
}

static void
_filelayout_free_lseg(struct nfs4_filelayout_segment *fl)
{
	filelayout_free_fh_array(fl);
	kfree(fl);
}

static int
filelayout_decode_layout(struct pnfs_layout_hdr *flo,
			 struct nfs4_filelayout_segment *fl,
			 struct nfs4_layoutget_res *lgr,
			 struct nfs4_deviceid *id,
			 gfp_t gfp_flags)
{
	struct xdr_stream stream;
	struct xdr_buf buf;
	struct page *scratch;
	__be32 *p;
	uint32_t nfl_util;
	int i;

	dprintk("%s: set_layout_map Begin\n", __func__);

	scratch = alloc_page(gfp_flags);
	if (!scratch)
		return -ENOMEM;

	xdr_init_decode_pages(&stream, &buf, lgr->layoutp->pages, lgr->layoutp->len);
	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);

	/* 20 = ufl_util (4), first_stripe_index (4), pattern_offset (8),
	 * num_fh (4) */
	p = xdr_inline_decode(&stream, NFS4_DEVICEID4_SIZE + 20);
	if (unlikely(!p))
		goto out_err;

	memcpy(id, p, sizeof(*id));
	p += XDR_QUADLEN(NFS4_DEVICEID4_SIZE);
	nfs4_print_deviceid(id);

	nfl_util = be32_to_cpup(p++);
	if (nfl_util & NFL4_UFLG_COMMIT_THRU_MDS)
		fl->commit_through_mds = 1;
	if (nfl_util & NFL4_UFLG_DENSE)
		fl->stripe_type = STRIPE_DENSE;
	else
		fl->stripe_type = STRIPE_SPARSE;
	fl->stripe_unit = nfl_util & ~NFL4_UFLG_MASK;

	fl->first_stripe_index = be32_to_cpup(p++);
	p = xdr_decode_hyper(p, &fl->pattern_offset);
	fl->num_fh = be32_to_cpup(p++);

	dprintk("%s: nfl_util 0x%X num_fh %u fsi %u po %llu\n",
		__func__, nfl_util, fl->num_fh, fl->first_stripe_index,
		fl->pattern_offset);

	/* Note that a zero value for num_fh is legal for STRIPE_SPARSE.
	 * Futher checking is done in filelayout_check_layout */
	if (fl->num_fh >
	    max(NFS4_PNFS_MAX_STRIPE_CNT, NFS4_PNFS_MAX_MULTI_CNT))
		goto out_err;

	if (fl->num_fh > 0) {
		fl->fh_array = kcalloc(fl->num_fh, sizeof(fl->fh_array[0]),
				       gfp_flags);
		if (!fl->fh_array)
			goto out_err;
	}

	for (i = 0; i < fl->num_fh; i++) {
		/* Do we want to use a mempool here? */
		fl->fh_array[i] = kmalloc(sizeof(struct nfs_fh), gfp_flags);
		if (!fl->fh_array[i])
			goto out_err_free;

		p = xdr_inline_decode(&stream, 4);
		if (unlikely(!p))
			goto out_err_free;
		fl->fh_array[i]->size = be32_to_cpup(p++);
		if (sizeof(struct nfs_fh) < fl->fh_array[i]->size) {
			printk(KERN_ERR "NFS: Too big fh %d received %d\n",
			       i, fl->fh_array[i]->size);
			goto out_err_free;
		}

		p = xdr_inline_decode(&stream, fl->fh_array[i]->size);
		if (unlikely(!p))
			goto out_err_free;
		memcpy(fl->fh_array[i]->data, p, fl->fh_array[i]->size);
		dprintk("DEBUG: %s: fh len %d\n", __func__,
			fl->fh_array[i]->size);
	}

	__free_page(scratch);
	return 0;

out_err_free:
	filelayout_free_fh_array(fl);
out_err:
	__free_page(scratch);
	return -EIO;
}

static void
filelayout_free_lseg(struct pnfs_layout_segment *lseg)
{
	struct nfs4_filelayout_segment *fl = FILELAYOUT_LSEG(lseg);

	dprintk("--> %s\n", __func__);
	nfs4_fl_put_deviceid(fl->dsaddr);
	/* This assumes a single RW lseg */
	if (lseg->pls_range.iomode == IOMODE_RW) {
		struct nfs4_filelayout *flo;

		flo = FILELAYOUT_FROM_HDR(lseg->pls_layout);
		flo->commit_info.nbuckets = 0;
		kfree(flo->commit_info.buckets);
		flo->commit_info.buckets = NULL;
	}
	_filelayout_free_lseg(fl);
}

static int
filelayout_alloc_commit_info(struct pnfs_layout_segment *lseg,
			     struct nfs_commit_info *cinfo,
			     gfp_t gfp_flags)
{
	struct nfs4_filelayout_segment *fl = FILELAYOUT_LSEG(lseg);
	struct pnfs_commit_bucket *buckets;
	int size, i;

	if (fl->commit_through_mds)
		return 0;

	size = (fl->stripe_type == STRIPE_SPARSE) ?
		fl->dsaddr->ds_num : fl->dsaddr->stripe_count;

	if (cinfo->ds->nbuckets >= size) {
		/* This assumes there is only one IOMODE_RW lseg.  What
		 * we really want to do is have a layout_hdr level
		 * dictionary of <multipath_list4, fh> keys, each
		 * associated with a struct list_head, populated by calls
		 * to filelayout_write_pagelist().
		 * */
		return 0;
	}

	buckets = kcalloc(size, sizeof(struct pnfs_commit_bucket),
			  gfp_flags);
	if (!buckets)
		return -ENOMEM;
	for (i = 0; i < size; i++) {
		INIT_LIST_HEAD(&buckets[i].written);
		INIT_LIST_HEAD(&buckets[i].committing);
		/* mark direct verifier as unset */
		buckets[i].direct_verf.committed = NFS_INVALID_STABLE_HOW;
	}

	spin_lock(cinfo->lock);
	if (cinfo->ds->nbuckets >= size)
		goto out;
	for (i = 0; i < cinfo->ds->nbuckets; i++) {
		list_splice(&cinfo->ds->buckets[i].written,
			    &buckets[i].written);
		list_splice(&cinfo->ds->buckets[i].committing,
			    &buckets[i].committing);
		buckets[i].direct_verf.committed =
			cinfo->ds->buckets[i].direct_verf.committed;
		buckets[i].wlseg = cinfo->ds->buckets[i].wlseg;
		buckets[i].clseg = cinfo->ds->buckets[i].clseg;
	}
	swap(cinfo->ds->buckets, buckets);
	cinfo->ds->nbuckets = size;
out:
	spin_unlock(cinfo->lock);
	kfree(buckets);
	return 0;
}

static struct pnfs_layout_segment *
filelayout_alloc_lseg(struct pnfs_layout_hdr *layoutid,
		      struct nfs4_layoutget_res *lgr,
		      gfp_t gfp_flags)
{
	struct nfs4_filelayout_segment *fl;
	int rc;
	struct nfs4_deviceid id;

	dprintk("--> %s\n", __func__);
	fl = kzalloc(sizeof(*fl), gfp_flags);
	if (!fl)
		return NULL;

	rc = filelayout_decode_layout(layoutid, fl, lgr, &id, gfp_flags);
	if (rc != 0 || filelayout_check_layout(layoutid, fl, lgr, &id, gfp_flags)) {
		_filelayout_free_lseg(fl);
		return NULL;
	}
	return &fl->generic_hdr;
}

/*
 * filelayout_pg_test(). Called by nfs_can_coalesce_requests()
 *
 * Return 0 if @req cannot be coalesced into @pgio, otherwise return the number
 * of bytes (maximum @req->wb_bytes) that can be coalesced.
 */
static size_t
filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
		   struct nfs_page *req)
{
	unsigned int size;
	u64 p_stripe, r_stripe;
	u32 stripe_offset;
	u64 segment_offset = pgio->pg_lseg->pls_range.offset;
	u32 stripe_unit = FILELAYOUT_LSEG(pgio->pg_lseg)->stripe_unit;

	/* calls nfs_generic_pg_test */
	size = pnfs_generic_pg_test(pgio, prev, req);
	if (!size)
		return 0;

	/* see if req and prev are in the same stripe */
	if (prev) {
		p_stripe = (u64)req_offset(prev) - segment_offset;
		r_stripe = (u64)req_offset(req) - segment_offset;
		do_div(p_stripe, stripe_unit);
		do_div(r_stripe, stripe_unit);

		if (p_stripe != r_stripe)
			return 0;
	}

	/* calculate remaining bytes in the current stripe */
	div_u64_rem((u64)req_offset(req) - segment_offset,
			stripe_unit,
			&stripe_offset);
	WARN_ON_ONCE(stripe_offset > stripe_unit);
	if (stripe_offset >= stripe_unit)
		return 0;
	return min(stripe_unit - (unsigned int)stripe_offset, size);
}

static void
filelayout_pg_init_read(struct nfs_pageio_descriptor *pgio,
			struct nfs_page *req)
{
	if (!pgio->pg_lseg)
		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
					   req->wb_context,
					   0,
					   NFS4_MAX_UINT64,
					   IOMODE_READ,
					   GFP_KERNEL);
	/* If no lseg, fall back to read through mds */
	if (pgio->pg_lseg == NULL)
		nfs_pageio_reset_read_mds(pgio);
}

static void
filelayout_pg_init_write(struct nfs_pageio_descriptor *pgio,
			 struct nfs_page *req)
{
	struct nfs_commit_info cinfo;
	int status;

	if (!pgio->pg_lseg)
		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
					   req->wb_context,
					   0,
					   NFS4_MAX_UINT64,
					   IOMODE_RW,
					   GFP_NOFS);
	/* If no lseg, fall back to write through mds */
	if (pgio->pg_lseg == NULL)
		goto out_mds;
	nfs_init_cinfo(&cinfo, pgio->pg_inode, pgio->pg_dreq);
	status = filelayout_alloc_commit_info(pgio->pg_lseg, &cinfo, GFP_NOFS);
	if (status < 0) {
		pnfs_put_lseg(pgio->pg_lseg);
		pgio->pg_lseg = NULL;
		goto out_mds;
	}
	return;
out_mds:
	nfs_pageio_reset_write_mds(pgio);
}

static const struct nfs_pageio_ops filelayout_pg_read_ops = {
	.pg_init = filelayout_pg_init_read,
	.pg_test = filelayout_pg_test,
	.pg_doio = pnfs_generic_pg_readpages,
	.pg_cleanup = pnfs_generic_pg_cleanup,
};

static const struct nfs_pageio_ops filelayout_pg_write_ops = {
	.pg_init = filelayout_pg_init_write,
	.pg_test = filelayout_pg_test,
	.pg_doio = pnfs_generic_pg_writepages,
	.pg_cleanup = pnfs_generic_pg_cleanup,
};

static u32 select_bucket_index(struct nfs4_filelayout_segment *fl, u32 j)
{
	if (fl->stripe_type == STRIPE_SPARSE)
		return nfs4_fl_calc_ds_index(&fl->generic_hdr, j);
	else
		return j;
}

static void
filelayout_mark_request_commit(struct nfs_page *req,
			       struct pnfs_layout_segment *lseg,
			       struct nfs_commit_info *cinfo,
			       u32 ds_commit_idx)

{
	struct nfs4_filelayout_segment *fl = FILELAYOUT_LSEG(lseg);
	u32 i, j;

	if (fl->commit_through_mds) {
		nfs_request_add_commit_list(req, &cinfo->mds->list, cinfo);
	} else {
		/* Note that we are calling nfs4_fl_calc_j_index on each page
		 * that ends up being committed to a data server.  An attractive
		 * alternative is to add a field to nfs_write_data and nfs_page
		 * to store the value calculated in filelayout_write_pagelist
		 * and just use that here.
		 */
		j = nfs4_fl_calc_j_index(lseg, req_offset(req));
		i = select_bucket_index(fl, j);
		pnfs_layout_mark_request_commit(req, lseg, cinfo, i);
	}
}

static u32 calc_ds_index_from_commit(struct pnfs_layout_segment *lseg, u32 i)
{
	struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg);

	if (flseg->stripe_type == STRIPE_SPARSE)
		return i;
	else
		return nfs4_fl_calc_ds_index(lseg, i);
}

static struct nfs_fh *
select_ds_fh_from_commit(struct pnfs_layout_segment *lseg, u32 i)
{
	struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg);

	if (flseg->stripe_type == STRIPE_SPARSE) {
		if (flseg->num_fh == 1)
			i = 0;
		else if (flseg->num_fh == 0)
			/* Use the MDS OPEN fh set in nfs_read_rpcsetup */
			return NULL;
	}
	return flseg->fh_array[i];
}

static int filelayout_initiate_commit(struct nfs_commit_data *data, int how)
{
	struct pnfs_layout_segment *lseg = data->lseg;
	struct nfs4_pnfs_ds *ds;
	struct rpc_clnt *ds_clnt;
	u32 idx;
	struct nfs_fh *fh;

	idx = calc_ds_index_from_commit(lseg, data->ds_commit_index);
	ds = nfs4_fl_prepare_ds(lseg, idx);
	if (!ds)
		goto out_err;

	ds_clnt = nfs4_find_or_create_ds_client(ds->ds_clp, data->inode);
	if (IS_ERR(ds_clnt))
		goto out_err;

	dprintk("%s ino %lu, how %d cl_count %d\n", __func__,
		data->inode->i_ino, how, atomic_read(&ds->ds_clp->cl_count));
	data->commit_done_cb = filelayout_commit_done_cb;
	atomic_inc(&ds->ds_clp->cl_count);
	data->ds_clp = ds->ds_clp;
	fh = select_ds_fh_from_commit(lseg, data->ds_commit_index);
	if (fh)
		data->args.fh = fh;
	return nfs_initiate_commit(ds_clnt, data, NFS_PROTO(data->inode),
				   &filelayout_commit_call_ops, how,
				   RPC_TASK_SOFTCONN);
out_err:
	pnfs_generic_prepare_to_resend_writes(data);
	pnfs_generic_commit_release(data);
	return -EAGAIN;
}

/* filelayout_search_commit_reqs - Search lists in @cinfo for the head reqest
 *				   for @page
 * @cinfo - commit info for current inode
 * @page - page to search for matching head request
 *
 * Returns a the head request if one is found, otherwise returns NULL.
 */
static struct nfs_page *
filelayout_search_commit_reqs(struct nfs_commit_info *cinfo, struct page *page)
{
	struct nfs_page *freq, *t;
	struct pnfs_commit_bucket *b;
	int i;

	/* Linearly search the commit lists for each bucket until a matching
	 * request is found */
	for (i = 0, b = cinfo->ds->buckets; i < cinfo->ds->nbuckets; i++, b++) {
		list_for_each_entry_safe(freq, t, &b->written, wb_list) {
			if (freq->wb_page == page)
				return freq->wb_head;
		}
		list_for_each_entry_safe(freq, t, &b->committing, wb_list) {
			if (freq->wb_page == page)
				return freq->wb_head;
		}
	}

	return NULL;
}

static int
filelayout_commit_pagelist(struct inode *inode, struct list_head *mds_pages,
			   int how, struct nfs_commit_info *cinfo)
{
	return pnfs_generic_commit_pagelist(inode, mds_pages, how, cinfo,
					    filelayout_initiate_commit);
}

static struct nfs4_deviceid_node *
filelayout_alloc_deviceid_node(struct nfs_server *server,
		struct pnfs_device *pdev, gfp_t gfp_flags)
{
	struct nfs4_file_layout_dsaddr *dsaddr;

	dsaddr = nfs4_fl_alloc_deviceid_node(server, pdev, gfp_flags);
	if (!dsaddr)
		return NULL;
	return &dsaddr->id_node;
}

static void
filelayout_free_deviceid_node(struct nfs4_deviceid_node *d)
{
	nfs4_fl_free_deviceid(container_of(d, struct nfs4_file_layout_dsaddr, id_node));
}

static struct pnfs_layout_hdr *
filelayout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags)
{
	struct nfs4_filelayout *flo;

	flo = kzalloc(sizeof(*flo), gfp_flags);
	return flo != NULL ? &flo->generic_hdr : NULL;
}

static void
filelayout_free_layout_hdr(struct pnfs_layout_hdr *lo)
{
	kfree(FILELAYOUT_FROM_HDR(lo));
}

static struct pnfs_ds_commit_info *
filelayout_get_ds_info(struct inode *inode)
{
	struct pnfs_layout_hdr *layout = NFS_I(inode)->layout;

	if (layout == NULL)
		return NULL;
	else
		return &FILELAYOUT_FROM_HDR(layout)->commit_info;
}

static struct pnfs_layoutdriver_type filelayout_type = {
	.id			= LAYOUT_NFSV4_1_FILES,
	.name			= "LAYOUT_NFSV4_1_FILES",
	.owner			= THIS_MODULE,
	.alloc_layout_hdr	= filelayout_alloc_layout_hdr,
	.free_layout_hdr	= filelayout_free_layout_hdr,
	.alloc_lseg		= filelayout_alloc_lseg,
	.free_lseg		= filelayout_free_lseg,
	.pg_read_ops		= &filelayout_pg_read_ops,
	.pg_write_ops		= &filelayout_pg_write_ops,
	.get_ds_info		= &filelayout_get_ds_info,
	.mark_request_commit	= filelayout_mark_request_commit,
	.clear_request_commit	= pnfs_generic_clear_request_commit,
	.scan_commit_lists	= pnfs_generic_scan_commit_lists,
	.recover_commit_reqs	= pnfs_generic_recover_commit_reqs,
	.search_commit_reqs	= filelayout_search_commit_reqs,
	.commit_pagelist	= filelayout_commit_pagelist,
	.read_pagelist		= filelayout_read_pagelist,
	.write_pagelist		= filelayout_write_pagelist,
	.alloc_deviceid_node	= filelayout_alloc_deviceid_node,
	.free_deviceid_node	= filelayout_free_deviceid_node,
	.sync			= pnfs_nfs_generic_sync,
};

static int __init nfs4filelayout_init(void)
{
	printk(KERN_INFO "%s: NFSv4 File Layout Driver Registering...\n",
	       __func__);
	return pnfs_register_layoutdriver(&filelayout_type);
}

static void __exit nfs4filelayout_exit(void)
{
	printk(KERN_INFO "%s: NFSv4 File Layout Driver Unregistering...\n",
	       __func__);
	pnfs_unregister_layoutdriver(&filelayout_type);
}

MODULE_ALIAS("nfs-layouttype4-1");

module_init(nfs4filelayout_init);
module_exit(nfs4filelayout_exit);