summaryrefslogtreecommitdiffstats
path: root/drivers/infiniband/ulp/iser/iscsi_iser.c
blob: bb9aaff92ca3b03f4e7443b583921dd0952a3b8e (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
/*
 * iSCSI Initiator over iSER Data-Path
 *
 * Copyright (C) 2004 Dmitry Yusupov
 * Copyright (C) 2004 Alex Aizman
 * Copyright (C) 2005 Mike Christie
 * Copyright (c) 2005, 2006 Voltaire, Inc. All rights reserved.
 * Copyright (c) 2013-2014 Mellanox Technologies. All rights reserved.
 * maintained by openib-general@openib.org
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     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.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 * Credits:
 *	Christoph Hellwig
 *	FUJITA Tomonori
 *	Arne Redlich
 *	Zhenyu Wang
 * Modified by:
 *      Erez Zilber
 */

#include <linux/types.h>
#include <linux/list.h>
#include <linux/hardirq.h>
#include <linux/kfifo.h>
#include <linux/blkdev.h>
#include <linux/init.h>
#include <linux/ioctl.h>
#include <linux/cdev.h>
#include <linux/in.h>
#include <linux/net.h>
#include <linux/scatterlist.h>
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/module.h>

#include <net/sock.h>

#include <linux/uaccess.h>

#include <scsi/scsi_cmnd.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_eh.h>
#include <scsi/scsi_tcq.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi.h>
#include <scsi/scsi_transport_iscsi.h>

#include "iscsi_iser.h"

MODULE_DESCRIPTION("iSER (iSCSI Extensions for RDMA) Datamover");
MODULE_LICENSE("Dual BSD/GPL");
MODULE_AUTHOR("Alex Nezhinsky, Dan Bar Dov, Or Gerlitz");

static const struct scsi_host_template iscsi_iser_sht;
static struct iscsi_transport iscsi_iser_transport;
static struct scsi_transport_template *iscsi_iser_scsi_transport;
static struct workqueue_struct *release_wq;
static DEFINE_MUTEX(unbind_iser_conn_mutex);
struct iser_global ig;

int iser_debug_level = 0;
module_param_named(debug_level, iser_debug_level, int, S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(debug_level, "Enable debug tracing if > 0 (default:disabled)");

static int iscsi_iser_set(const char *val, const struct kernel_param *kp);
static const struct kernel_param_ops iscsi_iser_size_ops = {
	.set = iscsi_iser_set,
	.get = param_get_uint,
};

static unsigned int iscsi_max_lun = 512;
module_param_cb(max_lun, &iscsi_iser_size_ops, &iscsi_max_lun, S_IRUGO);
MODULE_PARM_DESC(max_lun, "Max LUNs to allow per session, should > 0 (default:512)");

unsigned int iser_max_sectors = ISER_DEF_MAX_SECTORS;
module_param_cb(max_sectors, &iscsi_iser_size_ops, &iser_max_sectors,
		S_IRUGO | S_IWUSR);
MODULE_PARM_DESC(max_sectors, "Max number of sectors in a single scsi command, should > 0 (default:1024)");

bool iser_always_reg = true;
module_param_named(always_register, iser_always_reg, bool, S_IRUGO);
MODULE_PARM_DESC(always_register,
		 "Always register memory, even for continuous memory regions (default:true)");

bool iser_pi_enable = false;
module_param_named(pi_enable, iser_pi_enable, bool, S_IRUGO);
MODULE_PARM_DESC(pi_enable, "Enable T10-PI offload support (default:disabled)");

static int iscsi_iser_set(const char *val, const struct kernel_param *kp)
{
	int ret;
	unsigned int n = 0;

	ret = kstrtouint(val, 10, &n);
	if (ret != 0 || n == 0)
		return -EINVAL;

	return param_set_uint(val, kp);
}

/*
 * iscsi_iser_recv() - Process a successful recv completion
 * @conn:         iscsi connection
 * @hdr:          iscsi header
 * @rx_data:      buffer containing receive data payload
 * @rx_data_len:  length of rx_data
 *
 * Notes: In case of data length errors or iscsi PDU completion failures
 *        this routine will signal iscsi layer of connection failure.
 */
void iscsi_iser_recv(struct iscsi_conn *conn, struct iscsi_hdr *hdr,
		     char *rx_data, int rx_data_len)
{
	int rc = 0;
	int datalen;

	/* verify PDU length */
	datalen = ntoh24(hdr->dlength);
	if (datalen > rx_data_len || (datalen + 4) < rx_data_len) {
		iser_err("wrong datalen %d (hdr), %d (IB)\n",
			datalen, rx_data_len);
		rc = ISCSI_ERR_DATALEN;
		goto error;
	}

	if (datalen != rx_data_len)
		iser_dbg("aligned datalen (%d) hdr, %d (IB)\n",
			datalen, rx_data_len);

	rc = iscsi_complete_pdu(conn, hdr, rx_data, rx_data_len);
	if (rc && rc != ISCSI_ERR_NO_SCSI_CMD)
		goto error;

	return;
error:
	iscsi_conn_failure(conn, rc);
}

/**
 * iscsi_iser_pdu_alloc() - allocate an iscsi-iser PDU
 * @task:     iscsi task
 * @opcode:   iscsi command opcode
 *
 * Netes: This routine can't fail, just assign iscsi task
 *        hdr and max hdr size.
 */
static int iscsi_iser_pdu_alloc(struct iscsi_task *task, uint8_t opcode)
{
	struct iscsi_iser_task *iser_task = task->dd_data;

	task->hdr = (struct iscsi_hdr *)&iser_task->desc.iscsi_header;
	task->hdr_max = sizeof(iser_task->desc.iscsi_header);

	return 0;
}

/**
 * iser_initialize_task_headers() - Initialize task headers
 * @task:       iscsi task
 * @tx_desc:    iser tx descriptor
 *
 * Notes:
 * This routine may race with iser teardown flow for scsi
 * error handling TMFs. So for TMF we should acquire the
 * state mutex to avoid dereferencing the IB device which
 * may have already been terminated.
 */
int iser_initialize_task_headers(struct iscsi_task *task,
				 struct iser_tx_desc *tx_desc)
{
	struct iser_conn *iser_conn = task->conn->dd_data;
	struct iser_device *device = iser_conn->ib_conn.device;
	struct iscsi_iser_task *iser_task = task->dd_data;
	u64 dma_addr;

	if (unlikely(iser_conn->state != ISER_CONN_UP))
		return -ENODEV;

	dma_addr = ib_dma_map_single(device->ib_device, (void *)tx_desc,
				ISER_HEADERS_LEN, DMA_TO_DEVICE);
	if (ib_dma_mapping_error(device->ib_device, dma_addr))
		return -ENOMEM;

	tx_desc->inv_wr.next = NULL;
	tx_desc->reg_wr.wr.next = NULL;
	tx_desc->mapped = true;
	tx_desc->dma_addr = dma_addr;
	tx_desc->tx_sg[0].addr   = tx_desc->dma_addr;
	tx_desc->tx_sg[0].length = ISER_HEADERS_LEN;
	tx_desc->tx_sg[0].lkey   = device->pd->local_dma_lkey;

	iser_task->iser_conn = iser_conn;

	return 0;
}

/**
 * iscsi_iser_task_init() - Initialize iscsi-iser task
 * @task: iscsi task
 *
 * Initialize the task for the scsi command or mgmt command.
 *
 * Return: Returns zero on success or -ENOMEM when failing
 *         to init task headers (dma mapping error).
 */
static int iscsi_iser_task_init(struct iscsi_task *task)
{
	struct iscsi_iser_task *iser_task = task->dd_data;
	int ret;

	ret = iser_initialize_task_headers(task, &iser_task->desc);
	if (ret) {
		iser_err("Failed to init task %p, err = %d\n",
			 iser_task, ret);
		return ret;
	}

	/* mgmt task */
	if (!task->sc)
		return 0;

	iser_task->command_sent = 0;
	iser_task_rdma_init(iser_task);
	iser_task->sc = task->sc;

	return 0;
}

/**
 * iscsi_iser_mtask_xmit() - xmit management (immediate) task
 * @conn: iscsi connection
 * @task: task management task
 *
 * Notes:
 *	The function can return -EAGAIN in which case caller must
 *	call it again later, or recover. '0' return code means successful
 *	xmit.
 *
 **/
static int iscsi_iser_mtask_xmit(struct iscsi_conn *conn,
				 struct iscsi_task *task)
{
	int error = 0;

	iser_dbg("mtask xmit [cid %d itt 0x%x]\n", conn->id, task->itt);

	error = iser_send_control(conn, task);

	/* since iser xmits control with zero copy, tasks can not be recycled
	 * right after sending them.
	 * The recycling scheme is based on whether a response is expected
	 * - if yes, the task is recycled at iscsi_complete_pdu
	 * - if no,  the task is recycled at iser_snd_completion
	 */
	return error;
}

static int iscsi_iser_task_xmit_unsol_data(struct iscsi_conn *conn,
					   struct iscsi_task *task)
{
	struct iscsi_r2t_info *r2t = &task->unsol_r2t;
	struct iscsi_data hdr;
	int error = 0;

	/* Send data-out PDUs while there's still unsolicited data to send */
	while (iscsi_task_has_unsol_data(task)) {
		iscsi_prep_data_out_pdu(task, r2t, &hdr);
		iser_dbg("Sending data-out: itt 0x%x, data count %d\n",
			   hdr.itt, r2t->data_count);

		/* the buffer description has been passed with the command */
		/* Send the command */
		error = iser_send_data_out(conn, task, &hdr);
		if (error) {
			r2t->datasn--;
			goto iscsi_iser_task_xmit_unsol_data_exit;
		}
		r2t->sent += r2t->data_count;
		iser_dbg("Need to send %d more as data-out PDUs\n",
			   r2t->data_length - r2t->sent);
	}

iscsi_iser_task_xmit_unsol_data_exit:
	return error;
}

/**
 * iscsi_iser_task_xmit() - xmit iscsi-iser task
 * @task: iscsi task
 *
 * Return: zero on success or escalates $error on failure.
 */
static int iscsi_iser_task_xmit(struct iscsi_task *task)
{
	struct iscsi_conn *conn = task->conn;
	struct iscsi_iser_task *iser_task = task->dd_data;
	int error = 0;

	if (!task->sc)
		return iscsi_iser_mtask_xmit(conn, task);

	if (task->sc->sc_data_direction == DMA_TO_DEVICE) {
		BUG_ON(scsi_bufflen(task->sc) == 0);

		iser_dbg("cmd [itt %x total %d imm %d unsol_data %d\n",
			   task->itt, scsi_bufflen(task->sc),
			   task->imm_count, task->unsol_r2t.data_length);
	}

	iser_dbg("ctask xmit [cid %d itt 0x%x]\n",
		   conn->id, task->itt);

	/* Send the cmd PDU */
	if (!iser_task->command_sent) {
		error = iser_send_command(conn, task);
		if (error)
			goto iscsi_iser_task_xmit_exit;
		iser_task->command_sent = 1;
	}

	/* Send unsolicited data-out PDU(s) if necessary */
	if (iscsi_task_has_unsol_data(task))
		error = iscsi_iser_task_xmit_unsol_data(conn, task);

 iscsi_iser_task_xmit_exit:
	return error;
}

/**
 * iscsi_iser_cleanup_task() - cleanup an iscsi-iser task
 * @task: iscsi task
 *
 * Notes: In case the RDMA device is already NULL (might have
 *        been removed in DEVICE_REMOVAL CM event it will bail-out
 *        without doing dma unmapping.
 */
static void iscsi_iser_cleanup_task(struct iscsi_task *task)
{
	struct iscsi_iser_task *iser_task = task->dd_data;
	struct iser_tx_desc *tx_desc = &iser_task->desc;
	struct iser_conn *iser_conn = task->conn->dd_data;
	struct iser_device *device = iser_conn->ib_conn.device;

	/* DEVICE_REMOVAL event might have already released the device */
	if (!device)
		return;

	if (likely(tx_desc->mapped)) {
		ib_dma_unmap_single(device->ib_device, tx_desc->dma_addr,
				    ISER_HEADERS_LEN, DMA_TO_DEVICE);
		tx_desc->mapped = false;
	}

	/* mgmt tasks do not need special cleanup */
	if (!task->sc)
		return;

	if (iser_task->status == ISER_TASK_STATUS_STARTED) {
		iser_task->status = ISER_TASK_STATUS_COMPLETED;
		iser_task_rdma_finalize(iser_task);
	}
}

/**
 * iscsi_iser_check_protection() - check protection information status of task.
 * @task:     iscsi task
 * @sector:   error sector if exsists (output)
 *
 * Return: zero if no data-integrity errors have occured
 *         0x1: data-integrity error occured in the guard-block
 *         0x2: data-integrity error occured in the reference tag
 *         0x3: data-integrity error occured in the application tag
 *
 *         In addition the error sector is marked.
 */
static u8 iscsi_iser_check_protection(struct iscsi_task *task, sector_t *sector)
{
	struct iscsi_iser_task *iser_task = task->dd_data;
	enum iser_data_dir dir = iser_task->dir[ISER_DIR_IN] ?
					ISER_DIR_IN : ISER_DIR_OUT;

	return iser_check_task_pi_status(iser_task, dir, sector);
}

/**
 * iscsi_iser_conn_create() - create a new iscsi-iser connection
 * @cls_session: iscsi class connection
 * @conn_idx:    connection index within the session (for MCS)
 *
 * Return: iscsi_cls_conn when iscsi_conn_setup succeeds or NULL
 *         otherwise.
 */
static struct iscsi_cls_conn *
iscsi_iser_conn_create(struct iscsi_cls_session *cls_session,
		       uint32_t conn_idx)
{
	struct iscsi_conn *conn;
	struct iscsi_cls_conn *cls_conn;

	cls_conn = iscsi_conn_setup(cls_session, 0, conn_idx);
	if (!cls_conn)
		return NULL;
	conn = cls_conn->dd_data;

	/*
	 * due to issues with the login code re iser sematics
	 * this not set in iscsi_conn_setup - FIXME
	 */
	conn->max_recv_dlength = ISER_RECV_DATA_SEG_LEN;

	return cls_conn;
}

/**
 * iscsi_iser_conn_bind() - bind iscsi and iser connection structures
 * @cls_session:     iscsi class session
 * @cls_conn:        iscsi class connection
 * @transport_eph:   transport end-point handle
 * @is_leading:      indicate if this is the session leading connection (MCS)
 *
 * Return: zero on success, $error if iscsi_conn_bind fails and
 *         -EINVAL in case end-point doesn't exists anymore or iser connection
 *         state is not UP (teardown already started).
 */
static int iscsi_iser_conn_bind(struct iscsi_cls_session *cls_session,
				struct iscsi_cls_conn *cls_conn,
				uint64_t transport_eph, int is_leading)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iser_conn *iser_conn;
	struct iscsi_endpoint *ep;
	int error;

	error = iscsi_conn_bind(cls_session, cls_conn, is_leading);
	if (error)
		return error;

	/* the transport ep handle comes from user space so it must be
	 * verified against the global ib connections list */
	ep = iscsi_lookup_endpoint(transport_eph);
	if (!ep) {
		iser_err("can't bind eph %llx\n",
			 (unsigned long long)transport_eph);
		return -EINVAL;
	}
	iser_conn = ep->dd_data;

	mutex_lock(&iser_conn->state_mutex);
	if (iser_conn->state != ISER_CONN_UP) {
		error = -EINVAL;
		iser_err("iser_conn %p state is %d, teardown started\n",
			 iser_conn, iser_conn->state);
		goto out;
	}

	error = iser_alloc_rx_descriptors(iser_conn, conn->session);
	if (error)
		goto out;

	/* binds the iSER connection retrieved from the previously
	 * connected ep_handle to the iSCSI layer connection. exchanges
	 * connection pointers */
	iser_info("binding iscsi conn %p to iser_conn %p\n", conn, iser_conn);

	conn->dd_data = iser_conn;
	iser_conn->iscsi_conn = conn;

out:
	iscsi_put_endpoint(ep);
	mutex_unlock(&iser_conn->state_mutex);
	return error;
}

/**
 * iscsi_iser_conn_start() - start iscsi-iser connection
 * @cls_conn: iscsi class connection
 *
 * Notes: Here iser intialize (or re-initialize) stop_completion as
 *        from this point iscsi must call conn_stop in session/connection
 *        teardown so iser transport must wait for it.
 */
static int iscsi_iser_conn_start(struct iscsi_cls_conn *cls_conn)
{
	struct iscsi_conn *iscsi_conn;
	struct iser_conn *iser_conn;

	iscsi_conn = cls_conn->dd_data;
	iser_conn = iscsi_conn->dd_data;
	reinit_completion(&iser_conn->stop_completion);

	return iscsi_conn_start(cls_conn);
}

/**
 * iscsi_iser_conn_stop() - stop iscsi-iser connection
 * @cls_conn:  iscsi class connection
 * @flag:      indicate if recover or terminate (passed as is)
 *
 * Notes: Calling iscsi_conn_stop might theoretically race with
 *        DEVICE_REMOVAL event and dereference a previously freed RDMA device
 *        handle, so we call it under iser the state lock to protect against
 *        this kind of race.
 */
static void iscsi_iser_conn_stop(struct iscsi_cls_conn *cls_conn, int flag)
{
	struct iscsi_conn *conn = cls_conn->dd_data;
	struct iser_conn *iser_conn = conn->dd_data;

	iser_info("stopping iscsi_conn: %p, iser_conn: %p\n", conn, iser_conn);

	/*
	 * Userspace may have goofed up and not bound the connection or
	 * might have only partially setup the connection.
	 */
	if (iser_conn) {
		mutex_lock(&iser_conn->state_mutex);
		mutex_lock(&unbind_iser_conn_mutex);
		iser_conn_terminate(iser_conn);
		iscsi_conn_stop(cls_conn, flag);

		/* unbind */
		iser_conn->iscsi_conn = NULL;
		conn->dd_data = NULL;
		mutex_unlock(&unbind_iser_conn_mutex);

		complete(&iser_conn->stop_completion);
		mutex_unlock(&iser_conn->state_mutex);
	} else {
		iscsi_conn_stop(cls_conn, flag);
	}
}

/**
 * iscsi_iser_session_destroy() - destroy iscsi-iser session
 * @cls_session: iscsi class session
 *
 * Removes and free iscsi host.
 */
static void iscsi_iser_session_destroy(struct iscsi_cls_session *cls_session)
{
	struct Scsi_Host *shost = iscsi_session_to_shost(cls_session);

	iscsi_session_teardown(cls_session);
	iscsi_host_remove(shost, false);
	iscsi_host_free(shost);
}

static inline unsigned int iser_dif_prot_caps(int prot_caps)
{
	int ret = 0;

	if (prot_caps & IB_PROT_T10DIF_TYPE_1)
		ret |= SHOST_DIF_TYPE1_PROTECTION |
		       SHOST_DIX_TYPE0_PROTECTION |
		       SHOST_DIX_TYPE1_PROTECTION;
	if (prot_caps & IB_PROT_T10DIF_TYPE_2)
		ret |= SHOST_DIF_TYPE2_PROTECTION |
		       SHOST_DIX_TYPE2_PROTECTION;
	if (prot_caps & IB_PROT_T10DIF_TYPE_3)
		ret |= SHOST_DIF_TYPE3_PROTECTION |
		       SHOST_DIX_TYPE3_PROTECTION;

	return ret;
}

/**
 * iscsi_iser_session_create() - create an iscsi-iser session
 * @ep:             iscsi end-point handle
 * @cmds_max:       maximum commands in this session
 * @qdepth:         session command queue depth
 * @initial_cmdsn:  initiator command sequnce number
 *
 * Allocates and adds a scsi host, expose DIF supprot if
 * exists, and sets up an iscsi session.
 */
static struct iscsi_cls_session *
iscsi_iser_session_create(struct iscsi_endpoint *ep,
			  uint16_t cmds_max, uint16_t qdepth,
			  uint32_t initial_cmdsn)
{
	struct iscsi_cls_session *cls_session;
	struct Scsi_Host *shost;
	struct iser_conn *iser_conn = NULL;
	struct ib_conn *ib_conn;
	struct ib_device *ib_dev;
	u32 max_fr_sectors;

	shost = iscsi_host_alloc(&iscsi_iser_sht, 0, 0);
	if (!shost)
		return NULL;
	shost->transportt = iscsi_iser_scsi_transport;
	shost->cmd_per_lun = qdepth;
	shost->max_lun = iscsi_max_lun;
	shost->max_id = 0;
	shost->max_channel = 0;
	shost->max_cmd_len = 16;

	/*
	 * older userspace tools (before 2.0-870) did not pass us
	 * the leading conn's ep so this will be NULL;
	 */
	if (ep) {
		iser_conn = ep->dd_data;
		shost->sg_tablesize = iser_conn->scsi_sg_tablesize;
		shost->can_queue = min_t(u16, cmds_max, iser_conn->max_cmds);

		mutex_lock(&iser_conn->state_mutex);
		if (iser_conn->state != ISER_CONN_UP) {
			iser_err("iser conn %p already started teardown\n",
				 iser_conn);
			mutex_unlock(&iser_conn->state_mutex);
			goto free_host;
		}

		ib_conn = &iser_conn->ib_conn;
		ib_dev = ib_conn->device->ib_device;
		if (ib_conn->pi_support) {
			u32 sig_caps = ib_dev->attrs.sig_prot_cap;

			shost->sg_prot_tablesize = shost->sg_tablesize;
			scsi_host_set_prot(shost, iser_dif_prot_caps(sig_caps));
			scsi_host_set_guard(shost, SHOST_DIX_GUARD_IP |
						   SHOST_DIX_GUARD_CRC);
		}

		if (!(ib_dev->attrs.kernel_cap_flags & IBK_SG_GAPS_REG))
			shost->virt_boundary_mask = SZ_4K - 1;

		if (iscsi_host_add(shost, ib_dev->dev.parent)) {
			mutex_unlock(&iser_conn->state_mutex);
			goto free_host;
		}
		mutex_unlock(&iser_conn->state_mutex);
	} else {
		shost->can_queue = min_t(u16, cmds_max, ISER_DEF_XMIT_CMDS_MAX);
		if (iscsi_host_add(shost, NULL))
			goto free_host;
	}

	max_fr_sectors = (shost->sg_tablesize * PAGE_SIZE) >> 9;
	shost->max_sectors = min(iser_max_sectors, max_fr_sectors);

	iser_dbg("iser_conn %p, sg_tablesize %u, max_sectors %u\n",
		 iser_conn, shost->sg_tablesize,
		 shost->max_sectors);

	if (shost->max_sectors < iser_max_sectors)
		iser_warn("max_sectors was reduced from %u to %u\n",
			  iser_max_sectors, shost->max_sectors);

	cls_session = iscsi_session_setup(&iscsi_iser_transport, shost,
					  shost->can_queue, 0,
					  sizeof(struct iscsi_iser_task),
					  initial_cmdsn, 0);
	if (!cls_session)
		goto remove_host;

	return cls_session;

remove_host:
	iscsi_host_remove(shost, false);
free_host:
	iscsi_host_free(shost);
	return NULL;
}

static int iscsi_iser_set_param(struct iscsi_cls_conn *cls_conn,
				enum iscsi_param param, char *buf, int buflen)
{
	int value;

	switch (param) {
	case ISCSI_PARAM_MAX_RECV_DLENGTH:
		/* TBD */
		break;
	case ISCSI_PARAM_HDRDGST_EN:
		sscanf(buf, "%d", &value);
		if (value) {
			iser_err("DataDigest wasn't negotiated to None\n");
			return -EPROTO;
		}
		break;
	case ISCSI_PARAM_DATADGST_EN:
		sscanf(buf, "%d", &value);
		if (value) {
			iser_err("DataDigest wasn't negotiated to None\n");
			return -EPROTO;
		}
		break;
	case ISCSI_PARAM_IFMARKER_EN:
		sscanf(buf, "%d", &value);
		if (value) {
			iser_err("IFMarker wasn't negotiated to No\n");
			return -EPROTO;
		}
		break;
	case ISCSI_PARAM_OFMARKER_EN:
		sscanf(buf, "%d", &value);
		if (value) {
			iser_err("OFMarker wasn't negotiated to No\n");
			return -EPROTO;
		}
		break;
	default:
		return iscsi_set_param(cls_conn, param, buf, buflen);
	}

	return 0;
}

/**
 * iscsi_iser_conn_get_stats() - get iscsi connection statistics
 * @cls_conn:    iscsi class connection
 * @stats:       iscsi stats to output
 *
 * Output connection statistics.
 */
static void iscsi_iser_conn_get_stats(struct iscsi_cls_conn *cls_conn,
				      struct iscsi_stats *stats)
{
	struct iscsi_conn *conn = cls_conn->dd_data;

	stats->txdata_octets = conn->txdata_octets;
	stats->rxdata_octets = conn->rxdata_octets;
	stats->scsicmd_pdus = conn->scsicmd_pdus_cnt;
	stats->dataout_pdus = conn->dataout_pdus_cnt;
	stats->scsirsp_pdus = conn->scsirsp_pdus_cnt;
	stats->datain_pdus = conn->datain_pdus_cnt; /* always 0 */
	stats->r2t_pdus = conn->r2t_pdus_cnt; /* always 0 */
	stats->tmfcmd_pdus = conn->tmfcmd_pdus_cnt;
	stats->tmfrsp_pdus = conn->tmfrsp_pdus_cnt;
	stats->custom_length = 0;
}

static int iscsi_iser_get_ep_param(struct iscsi_endpoint *ep,
				   enum iscsi_param param, char *buf)
{
	struct iser_conn *iser_conn = ep->dd_data;

	switch (param) {
	case ISCSI_PARAM_CONN_PORT:
	case ISCSI_PARAM_CONN_ADDRESS:
		if (!iser_conn || !iser_conn->ib_conn.cma_id)
			return -ENOTCONN;

		return iscsi_conn_get_addr_param((struct sockaddr_storage *)
				&iser_conn->ib_conn.cma_id->route.addr.dst_addr,
				param, buf);
	default:
		break;
	}
	return -ENOSYS;
}

/**
 * iscsi_iser_ep_connect() - Initiate iSER connection establishment
 * @shost:          scsi_host
 * @dst_addr:       destination address
 * @non_blocking:   indicate if routine can block
 *
 * Allocate an iscsi endpoint, an iser_conn structure and bind them.
 * After that start RDMA connection establishment via rdma_cm. We
 * don't allocate iser_conn embedded in iscsi_endpoint since in teardown
 * the endpoint will be destroyed at ep_disconnect while iser_conn will
 * cleanup its resources asynchronuously.
 *
 * Return: iscsi_endpoint created by iscsi layer or ERR_PTR(error)
 *         if fails.
 */
static struct iscsi_endpoint *iscsi_iser_ep_connect(struct Scsi_Host *shost,
						    struct sockaddr *dst_addr,
						    int non_blocking)
{
	int err;
	struct iser_conn *iser_conn;
	struct iscsi_endpoint *ep;

	ep = iscsi_create_endpoint(0);
	if (!ep)
		return ERR_PTR(-ENOMEM);

	iser_conn = kzalloc(sizeof(*iser_conn), GFP_KERNEL);
	if (!iser_conn) {
		err = -ENOMEM;
		goto failure;
	}

	ep->dd_data = iser_conn;
	iser_conn->ep = ep;
	iser_conn_init(iser_conn);

	err = iser_connect(iser_conn, NULL, dst_addr, non_blocking);
	if (err)
		goto failure;

	return ep;
failure:
	iscsi_destroy_endpoint(ep);
	return ERR_PTR(err);
}

/**
 * iscsi_iser_ep_poll() - poll for iser connection establishment to complete
 * @ep:            iscsi endpoint (created at ep_connect)
 * @timeout_ms:    polling timeout allowed in ms.
 *
 * This routine boils down to waiting for up_completion signaling
 * that cma_id got CONNECTED event.
 *
 * Return: 1 if succeeded in connection establishment, 0 if timeout expired
 *         (libiscsi will retry will kick in) or -1 if interrupted by signal
 *         or more likely iser connection state transitioned to TEMINATING or
 *         DOWN during the wait period.
 */
static int iscsi_iser_ep_poll(struct iscsi_endpoint *ep, int timeout_ms)
{
	struct iser_conn *iser_conn = ep->dd_data;
	int rc;

	rc = wait_for_completion_interruptible_timeout(&iser_conn->up_completion,
						       msecs_to_jiffies(timeout_ms));
	/* if conn establishment failed, return error code to iscsi */
	if (rc == 0) {
		mutex_lock(&iser_conn->state_mutex);
		if (iser_conn->state == ISER_CONN_TERMINATING ||
		    iser_conn->state == ISER_CONN_DOWN)
			rc = -1;
		mutex_unlock(&iser_conn->state_mutex);
	}

	iser_info("iser conn %p rc = %d\n", iser_conn, rc);

	if (rc > 0)
		return 1; /* success, this is the equivalent of EPOLLOUT */
	else if (!rc)
		return 0; /* timeout */
	else
		return rc; /* signal */
}

/**
 * iscsi_iser_ep_disconnect() - Initiate connection teardown process
 * @ep:    iscsi endpoint handle
 *
 * This routine is not blocked by iser and RDMA termination process
 * completion as we queue a deffered work for iser/RDMA destruction
 * and cleanup or actually call it immediately in case we didn't pass
 * iscsi conn bind/start stage, thus it is safe.
 */
static void iscsi_iser_ep_disconnect(struct iscsi_endpoint *ep)
{
	struct iser_conn *iser_conn = ep->dd_data;

	iser_info("ep %p iser conn %p\n", ep, iser_conn);

	mutex_lock(&iser_conn->state_mutex);
	iser_conn_terminate(iser_conn);

	/*
	 * if iser_conn and iscsi_conn are bound, we must wait for
	 * iscsi_conn_stop and flush errors completion before freeing
	 * the iser resources. Otherwise we are safe to free resources
	 * immediately.
	 */
	if (iser_conn->iscsi_conn) {
		INIT_WORK(&iser_conn->release_work, iser_release_work);
		queue_work(release_wq, &iser_conn->release_work);
		mutex_unlock(&iser_conn->state_mutex);
	} else {
		iser_conn->state = ISER_CONN_DOWN;
		mutex_unlock(&iser_conn->state_mutex);
		iser_conn_release(iser_conn);
	}

	iscsi_destroy_endpoint(ep);
}

static umode_t iser_attr_is_visible(int param_type, int param)
{
	switch (param_type) {
	case ISCSI_HOST_PARAM:
		switch (param) {
		case ISCSI_HOST_PARAM_NETDEV_NAME:
		case ISCSI_HOST_PARAM_HWADDRESS:
		case ISCSI_HOST_PARAM_INITIATOR_NAME:
			return S_IRUGO;
		default:
			return 0;
		}
	case ISCSI_PARAM:
		switch (param) {
		case ISCSI_PARAM_MAX_RECV_DLENGTH:
		case ISCSI_PARAM_MAX_XMIT_DLENGTH:
		case ISCSI_PARAM_HDRDGST_EN:
		case ISCSI_PARAM_DATADGST_EN:
		case ISCSI_PARAM_CONN_ADDRESS:
		case ISCSI_PARAM_CONN_PORT:
		case ISCSI_PARAM_EXP_STATSN:
		case ISCSI_PARAM_PERSISTENT_ADDRESS:
		case ISCSI_PARAM_PERSISTENT_PORT:
		case ISCSI_PARAM_PING_TMO:
		case ISCSI_PARAM_RECV_TMO:
		case ISCSI_PARAM_INITIAL_R2T_EN:
		case ISCSI_PARAM_MAX_R2T:
		case ISCSI_PARAM_IMM_DATA_EN:
		case ISCSI_PARAM_FIRST_BURST:
		case ISCSI_PARAM_MAX_BURST:
		case ISCSI_PARAM_PDU_INORDER_EN:
		case ISCSI_PARAM_DATASEQ_INORDER_EN:
		case ISCSI_PARAM_TARGET_NAME:
		case ISCSI_PARAM_TPGT:
		case ISCSI_PARAM_USERNAME:
		case ISCSI_PARAM_PASSWORD:
		case ISCSI_PARAM_USERNAME_IN:
		case ISCSI_PARAM_PASSWORD_IN:
		case ISCSI_PARAM_FAST_ABORT:
		case ISCSI_PARAM_ABORT_TMO:
		case ISCSI_PARAM_LU_RESET_TMO:
		case ISCSI_PARAM_TGT_RESET_TMO:
		case ISCSI_PARAM_IFACE_NAME:
		case ISCSI_PARAM_INITIATOR_NAME:
		case ISCSI_PARAM_DISCOVERY_SESS:
			return S_IRUGO;
		default:
			return 0;
		}
	}

	return 0;
}

static const struct scsi_host_template iscsi_iser_sht = {
	.module                 = THIS_MODULE,
	.name                   = "iSCSI Initiator over iSER",
	.queuecommand           = iscsi_queuecommand,
	.change_queue_depth	= scsi_change_queue_depth,
	.sg_tablesize           = ISCSI_ISER_DEF_SG_TABLESIZE,
	.cmd_per_lun            = ISER_DEF_CMD_PER_LUN,
	.eh_timed_out		= iscsi_eh_cmd_timed_out,
	.eh_abort_handler       = iscsi_eh_abort,
	.eh_device_reset_handler= iscsi_eh_device_reset,
	.eh_target_reset_handler = iscsi_eh_recover_target,
	.target_alloc		= iscsi_target_alloc,
	.proc_name              = "iscsi_iser",
	.this_id                = -1,
	.track_queue_depth	= 1,
	.cmd_size		= sizeof(struct iscsi_cmd),
};

static struct iscsi_transport iscsi_iser_transport = {
	.owner                  = THIS_MODULE,
	.name                   = "iser",
	.caps                   = CAP_RECOVERY_L0 | CAP_MULTI_R2T | CAP_TEXT_NEGO,
	/* session management */
	.create_session         = iscsi_iser_session_create,
	.destroy_session        = iscsi_iser_session_destroy,
	/* connection management */
	.create_conn            = iscsi_iser_conn_create,
	.bind_conn              = iscsi_iser_conn_bind,
	.unbind_conn		= iscsi_conn_unbind,
	.destroy_conn           = iscsi_conn_teardown,
	.attr_is_visible	= iser_attr_is_visible,
	.set_param              = iscsi_iser_set_param,
	.get_conn_param		= iscsi_conn_get_param,
	.get_ep_param		= iscsi_iser_get_ep_param,
	.get_session_param	= iscsi_session_get_param,
	.start_conn             = iscsi_iser_conn_start,
	.stop_conn              = iscsi_iser_conn_stop,
	/* iscsi host params */
	.get_host_param		= iscsi_host_get_param,
	.set_host_param		= iscsi_host_set_param,
	/* IO */
	.send_pdu		= iscsi_conn_send_pdu,
	.get_stats		= iscsi_iser_conn_get_stats,
	.init_task		= iscsi_iser_task_init,
	.xmit_task		= iscsi_iser_task_xmit,
	.cleanup_task		= iscsi_iser_cleanup_task,
	.alloc_pdu		= iscsi_iser_pdu_alloc,
	.check_protection	= iscsi_iser_check_protection,
	/* recovery */
	.session_recovery_timedout = iscsi_session_recovery_timedout,

	.ep_connect             = iscsi_iser_ep_connect,
	.ep_poll                = iscsi_iser_ep_poll,
	.ep_disconnect          = iscsi_iser_ep_disconnect
};

static int __init iser_init(void)
{
	int err;

	iser_dbg("Starting iSER datamover...\n");

	memset(&ig, 0, sizeof(struct iser_global));

	ig.desc_cache = kmem_cache_create("iser_descriptors",
					  sizeof(struct iser_tx_desc),
					  0, SLAB_HWCACHE_ALIGN,
					  NULL);
	if (ig.desc_cache == NULL)
		return -ENOMEM;

	/* device init is called only after the first addr resolution */
	mutex_init(&ig.device_list_mutex);
	INIT_LIST_HEAD(&ig.device_list);
	mutex_init(&ig.connlist_mutex);
	INIT_LIST_HEAD(&ig.connlist);

	release_wq = alloc_workqueue("release workqueue", 0, 0);
	if (!release_wq) {
		iser_err("failed to allocate release workqueue\n");
		err = -ENOMEM;
		goto err_alloc_wq;
	}

	iscsi_iser_scsi_transport = iscsi_register_transport(
							&iscsi_iser_transport);
	if (!iscsi_iser_scsi_transport) {
		iser_err("iscsi_register_transport failed\n");
		err = -EINVAL;
		goto err_reg;
	}

	return 0;

err_reg:
	destroy_workqueue(release_wq);
err_alloc_wq:
	kmem_cache_destroy(ig.desc_cache);

	return err;
}

static void __exit iser_exit(void)
{
	struct iser_conn *iser_conn, *n;
	int connlist_empty;

	iser_dbg("Removing iSER datamover...\n");
	destroy_workqueue(release_wq);

	mutex_lock(&ig.connlist_mutex);
	connlist_empty = list_empty(&ig.connlist);
	mutex_unlock(&ig.connlist_mutex);

	if (!connlist_empty) {
		iser_err("Error cleanup stage completed but we still have iser "
			 "connections, destroying them anyway\n");
		list_for_each_entry_safe(iser_conn, n, &ig.connlist,
					 conn_list) {
			iser_conn_release(iser_conn);
		}
	}

	iscsi_unregister_transport(&iscsi_iser_transport);
	kmem_cache_destroy(ig.desc_cache);
}

module_init(iser_init);
module_exit(iser_exit);