ib_isert.c 90.9 KB
Newer Older
1 2 3
/*******************************************************************************
 * This file contains iSCSI extentions for RDMA (iSER) Verbs
 *
4
 * (c) Copyright 2013 Datera, Inc.
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
 *
 * Nicholas A. Bellinger <nab@linux-iscsi.org>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * 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.
 ****************************************************************************/

#include <linux/string.h>
#include <linux/module.h>
#include <linux/scatterlist.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/in6.h>
#include <rdma/ib_verbs.h>
#include <rdma/rdma_cm.h>
#include <target/target_core_base.h>
#include <target/target_core_fabric.h>
#include <target/iscsi/iscsi_transport.h>
30
#include <linux/semaphore.h>
31 32 33 34 35 36 37

#include "isert_proto.h"
#include "ib_isert.h"

#define	ISERT_MAX_CONN		8
#define ISER_MAX_RX_CQ_LEN	(ISERT_QP_MAX_RECV_DTOS * ISERT_MAX_CONN)
#define ISER_MAX_TX_CQ_LEN	(ISERT_QP_MAX_REQ_DTOS  * ISERT_MAX_CONN)
38 39
#define ISER_MAX_CQ_LEN		(ISER_MAX_RX_CQ_LEN + ISER_MAX_TX_CQ_LEN + \
				 ISERT_MAX_CONN)
40 41 42 43

static DEFINE_MUTEX(device_list_mutex);
static LIST_HEAD(device_list);
static struct workqueue_struct *isert_comp_wq;
44
static struct workqueue_struct *isert_release_wq;
45

46 47 48 49 50
static void
isert_unmap_cmd(struct isert_cmd *isert_cmd, struct isert_conn *isert_conn);
static int
isert_map_rdma(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
	       struct isert_rdma_wr *wr);
51
static void
52
isert_unreg_rdma(struct isert_cmd *isert_cmd, struct isert_conn *isert_conn);
53
static int
54 55
isert_reg_rdma(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
	       struct isert_rdma_wr *wr);
56 57
static int
isert_put_response(struct iscsi_conn *conn, struct iscsi_cmd *cmd);
58 59 60 61
static int
isert_rdma_post_recvl(struct isert_conn *isert_conn);
static int
isert_rdma_accept(struct isert_conn *isert_conn);
62
struct rdma_cm_id *isert_setup_id(struct isert_np *isert_np);
63

64 65 66
static inline bool
isert_prot_cmd(struct isert_conn *conn, struct se_cmd *cmd)
{
67
	return (conn->pi_support &&
68 69 70 71
		cmd->prot_op != TARGET_PROT_NORMAL);
}


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
static void
isert_qp_event_callback(struct ib_event *e, void *context)
{
	struct isert_conn *isert_conn = (struct isert_conn *)context;

	pr_err("isert_qp_event_callback event: %d\n", e->event);
	switch (e->event) {
	case IB_EVENT_COMM_EST:
		rdma_notify(isert_conn->conn_cm_id, IB_EVENT_COMM_EST);
		break;
	case IB_EVENT_QP_LAST_WQE_REACHED:
		pr_warn("Reached TX IB_EVENT_QP_LAST_WQE_REACHED:\n");
		break;
	default:
		break;
	}
}

static int
isert_query_device(struct ib_device *ib_dev, struct ib_device_attr *devattr)
{
	int ret;

	ret = ib_query_device(ib_dev, devattr);
	if (ret) {
		pr_err("ib_query_device() failed: %d\n", ret);
		return ret;
	}
	pr_debug("devattr->max_sge: %d\n", devattr->max_sge);
	pr_debug("devattr->max_sge_rd: %d\n", devattr->max_sge_rd);

	return 0;
}

static int
107
isert_conn_setup_qp(struct isert_conn *isert_conn, struct rdma_cm_id *cma_id)
108 109 110
{
	struct isert_device *device = isert_conn->conn_device;
	struct ib_qp_init_attr attr;
111 112
	struct isert_comp *comp;
	int ret, i, min = 0;
113 114

	mutex_lock(&device_list_mutex);
115 116 117 118 119 120 121 122
	for (i = 0; i < device->comps_used; i++)
		if (device->comps[i].active_qps <
		    device->comps[min].active_qps)
			min = i;
	comp = &device->comps[min];
	comp->active_qps++;
	pr_info("conn %p, using comp %p min_index: %d\n",
		   isert_conn, comp, min);
123 124 125 126 127
	mutex_unlock(&device_list_mutex);

	memset(&attr, 0, sizeof(struct ib_qp_init_attr));
	attr.event_handler = isert_qp_event_callback;
	attr.qp_context = isert_conn;
128 129
	attr.send_cq = comp->cq;
	attr.recv_cq = comp->cq;
130
	attr.cap.max_send_wr = ISERT_QP_MAX_REQ_DTOS;
131
	attr.cap.max_recv_wr = ISERT_QP_MAX_RECV_DTOS + 1;
132 133
	/*
	 * FIXME: Use devattr.max_sge - 2 for max_send_sge as
134 135 136 137
	 * work-around for RDMA_READs with ConnectX-2.
	 *
	 * Also, still make sure to have at least two SGEs for
	 * outgoing control PDU responses.
138
	 */
139
	attr.cap.max_send_sge = max(2, device->dev_attr.max_sge - 2);
140 141 142 143 144
	isert_conn->max_sge = attr.cap.max_send_sge;

	attr.cap.max_recv_sge = 1;
	attr.sq_sig_type = IB_SIGNAL_REQ_WR;
	attr.qp_type = IB_QPT_RC;
145
	if (device->pi_capable)
146
		attr.create_flags |= IB_QP_CREATE_SIGNATURE_EN;
147 148 149 150 151 152 153 154 155

	pr_debug("isert_conn_setup_qp cma_id->device: %p\n",
		 cma_id->device);
	pr_debug("isert_conn_setup_qp conn_pd->device: %p\n",
		 isert_conn->conn_pd->device);

	ret = rdma_create_qp(cma_id, isert_conn->conn_pd, &attr);
	if (ret) {
		pr_err("rdma_create_qp failed for cma_id %d\n", ret);
156
		goto err;
157 158 159 160 161
	}
	isert_conn->conn_qp = cma_id->qp;
	pr_debug("rdma_create_qp() returned success >>>>>>>>>>>>>>>>>>>>>>>>>.\n");

	return 0;
162 163
err:
	mutex_lock(&device_list_mutex);
164
	comp->active_qps--;
165 166 167
	mutex_unlock(&device_list_mutex);

	return ret;
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
}

static void
isert_cq_event_callback(struct ib_event *e, void *context)
{
	pr_debug("isert_cq_event_callback event: %d\n", e->event);
}

static int
isert_alloc_rx_descriptors(struct isert_conn *isert_conn)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct iser_rx_desc *rx_desc;
	struct ib_sge *rx_sg;
	u64 dma_addr;
	int i, j;

	isert_conn->conn_rx_descs = kzalloc(ISERT_QP_MAX_RECV_DTOS *
				sizeof(struct iser_rx_desc), GFP_KERNEL);
	if (!isert_conn->conn_rx_descs)
		goto fail;

	rx_desc = isert_conn->conn_rx_descs;

	for (i = 0; i < ISERT_QP_MAX_RECV_DTOS; i++, rx_desc++)  {
		dma_addr = ib_dma_map_single(ib_dev, (void *)rx_desc,
					ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
		if (ib_dma_mapping_error(ib_dev, dma_addr))
			goto dma_map_fail;

		rx_desc->dma_addr = dma_addr;

		rx_sg = &rx_desc->rx_sg;
		rx_sg->addr = rx_desc->dma_addr;
		rx_sg->length = ISER_RX_PAYLOAD_SIZE;
		rx_sg->lkey = isert_conn->conn_mr->lkey;
	}

	isert_conn->conn_rx_desc_head = 0;
	return 0;

dma_map_fail:
	rx_desc = isert_conn->conn_rx_descs;
	for (j = 0; j < i; j++, rx_desc++) {
		ib_dma_unmap_single(ib_dev, rx_desc->dma_addr,
				    ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
	}
	kfree(isert_conn->conn_rx_descs);
	isert_conn->conn_rx_descs = NULL;
fail:
	return -ENOMEM;
}

static void
isert_free_rx_descriptors(struct isert_conn *isert_conn)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct iser_rx_desc *rx_desc;
	int i;

	if (!isert_conn->conn_rx_descs)
		return;

	rx_desc = isert_conn->conn_rx_descs;
	for (i = 0; i < ISERT_QP_MAX_RECV_DTOS; i++, rx_desc++)  {
		ib_dma_unmap_single(ib_dev, rx_desc->dma_addr,
				    ISER_RX_PAYLOAD_SIZE, DMA_FROM_DEVICE);
	}

	kfree(isert_conn->conn_rx_descs);
	isert_conn->conn_rx_descs = NULL;
}

241 242
static void isert_cq_work(struct work_struct *);
static void isert_cq_callback(struct ib_cq *, void *);
243 244 245 246 247

static int
isert_create_device_ib_res(struct isert_device *device)
{
	struct ib_device *ib_dev = device->ib_device;
248
	struct ib_device_attr *dev_attr;
249
	int ret = 0, i;
250
	int max_cqe;
251

252 253 254 255 256
	dev_attr = &device->dev_attr;
	ret = isert_query_device(ib_dev, dev_attr);
	if (ret)
		return ret;

257
	max_cqe = min(ISER_MAX_CQ_LEN, dev_attr->max_cqe);
258

259
	/* asign function handlers */
260 261
	if (dev_attr->device_cap_flags & IB_DEVICE_MEM_MGT_EXTENSIONS &&
	    dev_attr->device_cap_flags & IB_DEVICE_SIGNATURE_HANDOVER) {
262 263 264
		device->use_fastreg = 1;
		device->reg_rdma_mem = isert_reg_rdma;
		device->unreg_rdma_mem = isert_unreg_rdma;
265
	} else {
266
		device->use_fastreg = 0;
267 268 269
		device->reg_rdma_mem = isert_map_rdma;
		device->unreg_rdma_mem = isert_unmap_cmd;
	}
270

271 272 273 274
	/* Check signature cap */
	device->pi_capable = dev_attr->device_cap_flags &
			     IB_DEVICE_SIGNATURE_HANDOVER ? true : false;

275 276 277 278 279 280 281 282 283 284 285 286
	device->comps_used = min(ISERT_MAX_CQ, min_t(int, num_online_cpus(),
					device->ib_device->num_comp_vectors));
	pr_info("Using %d CQs, %s supports %d vectors support "
		   "Fast registration %d pi_capable %d\n",
		   device->comps_used, device->ib_device->name,
		   device->ib_device->num_comp_vectors, device->use_fastreg,
		   device->pi_capable);

	device->comps = kcalloc(device->comps_used, sizeof(struct isert_comp),
				GFP_KERNEL);
	if (!device->comps) {
		pr_err("Unable to allocate completion contexts\n");
287 288
		return -ENOMEM;
	}
289 290 291 292 293

	for (i = 0; i < device->comps_used; i++) {
		struct isert_comp *comp = &device->comps[i];

		comp->device = device;
294 295 296 297 298 299 300 301 302
		INIT_WORK(&comp->work, isert_cq_work);
		comp->cq = ib_create_cq(device->ib_device,
					isert_cq_callback,
					isert_cq_event_callback,
					(void *)comp,
					max_cqe, i);
		if (IS_ERR(comp->cq)) {
			ret = PTR_ERR(comp->cq);
			comp->cq = NULL;
303
			goto out_cq;
304
		}
305

306
		ret = ib_req_notify_cq(comp->cq, IB_CQ_NEXT_COMP);
307
		if (ret)
308 309 310 311 312 313
			goto out_cq;
	}

	return 0;

out_cq:
314 315
	for (i = 0; i < device->comps_used; i++) {
		struct isert_comp *comp = &device->comps[i];
316

317 318 319
		if (comp->cq) {
			cancel_work_sync(&comp->work);
			ib_destroy_cq(comp->cq);
320 321
		}
	}
322
	kfree(device->comps);
323 324 325 326 327 328 329 330 331

	return ret;
}

static void
isert_free_device_ib_res(struct isert_device *device)
{
	int i;

332
	pr_info("device %p\n", device);
333

334 335
	for (i = 0; i < device->comps_used; i++) {
		struct isert_comp *comp = &device->comps[i];
336

337 338 339
		cancel_work_sync(&comp->work);
		ib_destroy_cq(comp->cq);
		comp->cq = NULL;
340 341
	}
	kfree(device->comps);
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
}

static void
isert_device_try_release(struct isert_device *device)
{
	mutex_lock(&device_list_mutex);
	device->refcount--;
	if (!device->refcount) {
		isert_free_device_ib_res(device);
		list_del(&device->dev_node);
		kfree(device);
	}
	mutex_unlock(&device_list_mutex);
}

static struct isert_device *
isert_device_find_by_ib_dev(struct rdma_cm_id *cma_id)
{
	struct isert_device *device;
	int ret;

	mutex_lock(&device_list_mutex);
	list_for_each_entry(device, &device_list, dev_node) {
		if (device->ib_device->node_guid == cma_id->device->node_guid) {
			device->refcount++;
			mutex_unlock(&device_list_mutex);
			return device;
		}
	}

	device = kzalloc(sizeof(struct isert_device), GFP_KERNEL);
	if (!device) {
		mutex_unlock(&device_list_mutex);
		return ERR_PTR(-ENOMEM);
	}

	INIT_LIST_HEAD(&device->dev_node);

	device->ib_device = cma_id->device;
	ret = isert_create_device_ib_res(device);
	if (ret) {
		kfree(device);
		mutex_unlock(&device_list_mutex);
		return ERR_PTR(ret);
	}

	device->refcount++;
	list_add_tail(&device->dev_node, &device_list);
	mutex_unlock(&device_list_mutex);

	return device;
}

395
static void
396
isert_conn_free_fastreg_pool(struct isert_conn *isert_conn)
397 398 399 400
{
	struct fast_reg_descriptor *fr_desc, *tmp;
	int i = 0;

401
	if (list_empty(&isert_conn->conn_fr_pool))
402 403
		return;

404
	pr_debug("Freeing conn %p fastreg pool", isert_conn);
405 406

	list_for_each_entry_safe(fr_desc, tmp,
407
				 &isert_conn->conn_fr_pool, list) {
408 409 410
		list_del(&fr_desc->list);
		ib_free_fast_reg_page_list(fr_desc->data_frpl);
		ib_dereg_mr(fr_desc->data_mr);
411 412 413 414 415 416
		if (fr_desc->pi_ctx) {
			ib_free_fast_reg_page_list(fr_desc->pi_ctx->prot_frpl);
			ib_dereg_mr(fr_desc->pi_ctx->prot_mr);
			ib_destroy_mr(fr_desc->pi_ctx->sig_mr);
			kfree(fr_desc->pi_ctx);
		}
417 418 419 420
		kfree(fr_desc);
		++i;
	}

421
	if (i < isert_conn->conn_fr_pool_size)
422
		pr_warn("Pool still has %d regions registered\n",
423
			isert_conn->conn_fr_pool_size - i);
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
static int
isert_create_pi_ctx(struct fast_reg_descriptor *desc,
		    struct ib_device *device,
		    struct ib_pd *pd)
{
	struct ib_mr_init_attr mr_init_attr;
	struct pi_context *pi_ctx;
	int ret;

	pi_ctx = kzalloc(sizeof(*desc->pi_ctx), GFP_KERNEL);
	if (!pi_ctx) {
		pr_err("Failed to allocate pi context\n");
		return -ENOMEM;
	}

	pi_ctx->prot_frpl = ib_alloc_fast_reg_page_list(device,
					    ISCSI_ISER_SG_TABLESIZE);
	if (IS_ERR(pi_ctx->prot_frpl)) {
		pr_err("Failed to allocate prot frpl err=%ld\n",
			  PTR_ERR(pi_ctx->prot_frpl));
		ret = PTR_ERR(pi_ctx->prot_frpl);
		goto err_pi_ctx;
	}

	pi_ctx->prot_mr = ib_alloc_fast_reg_mr(pd, ISCSI_ISER_SG_TABLESIZE);
	if (IS_ERR(pi_ctx->prot_mr)) {
		pr_err("Failed to allocate prot frmr err=%ld\n",
			  PTR_ERR(pi_ctx->prot_mr));
		ret = PTR_ERR(pi_ctx->prot_mr);
		goto err_prot_frpl;
	}
	desc->ind |= ISERT_PROT_KEY_VALID;

	memset(&mr_init_attr, 0, sizeof(mr_init_attr));
	mr_init_attr.max_reg_descriptors = 2;
	mr_init_attr.flags |= IB_MR_SIGNATURE_EN;
	pi_ctx->sig_mr = ib_create_mr(pd, &mr_init_attr);
	if (IS_ERR(pi_ctx->sig_mr)) {
		pr_err("Failed to allocate signature enabled mr err=%ld\n",
			  PTR_ERR(pi_ctx->sig_mr));
		ret = PTR_ERR(pi_ctx->sig_mr);
		goto err_prot_mr;
	}

	desc->pi_ctx = pi_ctx;
	desc->ind |= ISERT_SIG_KEY_VALID;
	desc->ind &= ~ISERT_PROTECTED;

	return 0;

err_prot_mr:
	ib_dereg_mr(desc->pi_ctx->prot_mr);
err_prot_frpl:
	ib_free_fast_reg_page_list(desc->pi_ctx->prot_frpl);
err_pi_ctx:
	kfree(desc->pi_ctx);

	return ret;
}

486 487
static int
isert_create_fr_desc(struct ib_device *ib_device, struct ib_pd *pd,
488
		     struct fast_reg_descriptor *fr_desc)
489
{
490 491
	int ret;

492 493 494 495 496 497 498 499 500 501 502 503
	fr_desc->data_frpl = ib_alloc_fast_reg_page_list(ib_device,
							 ISCSI_ISER_SG_TABLESIZE);
	if (IS_ERR(fr_desc->data_frpl)) {
		pr_err("Failed to allocate data frpl err=%ld\n",
		       PTR_ERR(fr_desc->data_frpl));
		return PTR_ERR(fr_desc->data_frpl);
	}

	fr_desc->data_mr = ib_alloc_fast_reg_mr(pd, ISCSI_ISER_SG_TABLESIZE);
	if (IS_ERR(fr_desc->data_mr)) {
		pr_err("Failed to allocate data frmr err=%ld\n",
		       PTR_ERR(fr_desc->data_mr));
504 505
		ret = PTR_ERR(fr_desc->data_mr);
		goto err_data_frpl;
506
	}
507 508
	fr_desc->ind |= ISERT_DATA_KEY_VALID;

509
	pr_debug("Created fr_desc %p\n", fr_desc);
510 511

	return 0;
512

513 514 515 516
err_data_frpl:
	ib_free_fast_reg_page_list(fr_desc->data_frpl);

	return ret;
517 518 519
}

static int
520
isert_conn_create_fastreg_pool(struct isert_conn *isert_conn)
521 522 523
{
	struct fast_reg_descriptor *fr_desc;
	struct isert_device *device = isert_conn->conn_device;
524 525 526 527 528 529 530 531 532
	struct se_session *se_sess = isert_conn->conn->sess->se_sess;
	struct se_node_acl *se_nacl = se_sess->se_node_acl;
	int i, ret, tag_num;
	/*
	 * Setup the number of FRMRs based upon the number of tags
	 * available to session in iscsi_target_locate_portal().
	 */
	tag_num = max_t(u32, ISCSIT_MIN_TAGS, se_nacl->queue_depth);
	tag_num = (tag_num * 2) + ISCSIT_EXTRA_TAGS;
533

534
	isert_conn->conn_fr_pool_size = 0;
535
	for (i = 0; i < tag_num; i++) {
536 537 538 539 540 541 542
		fr_desc = kzalloc(sizeof(*fr_desc), GFP_KERNEL);
		if (!fr_desc) {
			pr_err("Failed to allocate fast_reg descriptor\n");
			ret = -ENOMEM;
			goto err;
		}

543
		ret = isert_create_fr_desc(device->ib_device,
544
					   isert_conn->conn_pd, fr_desc);
545 546 547
		if (ret) {
			pr_err("Failed to create fastreg descriptor err=%d\n",
			       ret);
548
			kfree(fr_desc);
549 550 551
			goto err;
		}

552 553
		list_add_tail(&fr_desc->list, &isert_conn->conn_fr_pool);
		isert_conn->conn_fr_pool_size++;
554 555
	}

556 557
	pr_debug("Creating conn %p fastreg pool size=%d",
		 isert_conn, isert_conn->conn_fr_pool_size);
558 559 560 561

	return 0;

err:
562
	isert_conn_free_fastreg_pool(isert_conn);
563 564 565
	return ret;
}

566 567 568
static int
isert_connect_request(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
{
569 570
	struct isert_np *isert_np = cma_id->context;
	struct iscsi_np *np = isert_np->np;
571 572 573 574
	struct isert_conn *isert_conn;
	struct isert_device *device;
	struct ib_device *ib_dev = cma_id->device;
	int ret = 0;
575 576 577 578 579 580 581 582

	spin_lock_bh(&np->np_thread_lock);
	if (!np->enabled) {
		spin_unlock_bh(&np->np_thread_lock);
		pr_debug("iscsi_np is not enabled, reject connect request\n");
		return rdma_reject(cma_id, NULL, 0);
	}
	spin_unlock_bh(&np->np_thread_lock);
583 584 585 586 587 588 589 590 591 592 593 594

	pr_debug("Entering isert_connect_request cma_id: %p, context: %p\n",
		 cma_id, cma_id->context);

	isert_conn = kzalloc(sizeof(struct isert_conn), GFP_KERNEL);
	if (!isert_conn) {
		pr_err("Unable to allocate isert_conn\n");
		return -ENOMEM;
	}
	isert_conn->state = ISER_CONN_INIT;
	INIT_LIST_HEAD(&isert_conn->conn_accept_node);
	init_completion(&isert_conn->conn_login_comp);
595
	init_completion(&isert_conn->login_req_comp);
596
	init_completion(&isert_conn->conn_wait);
597
	kref_init(&isert_conn->conn_kref);
598
	mutex_init(&isert_conn->conn_mutex);
599
	spin_lock_init(&isert_conn->conn_lock);
600
	INIT_LIST_HEAD(&isert_conn->conn_fr_pool);
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

	isert_conn->conn_cm_id = cma_id;

	isert_conn->login_buf = kzalloc(ISCSI_DEF_MAX_RECV_SEG_LEN +
					ISER_RX_LOGIN_SIZE, GFP_KERNEL);
	if (!isert_conn->login_buf) {
		pr_err("Unable to allocate isert_conn->login_buf\n");
		ret = -ENOMEM;
		goto out;
	}

	isert_conn->login_req_buf = isert_conn->login_buf;
	isert_conn->login_rsp_buf = isert_conn->login_buf +
				    ISCSI_DEF_MAX_RECV_SEG_LEN;
	pr_debug("Set login_buf: %p login_req_buf: %p login_rsp_buf: %p\n",
		 isert_conn->login_buf, isert_conn->login_req_buf,
		 isert_conn->login_rsp_buf);

	isert_conn->login_req_dma = ib_dma_map_single(ib_dev,
				(void *)isert_conn->login_req_buf,
				ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_FROM_DEVICE);

	ret = ib_dma_mapping_error(ib_dev, isert_conn->login_req_dma);
	if (ret) {
		pr_err("ib_dma_mapping_error failed for login_req_dma: %d\n",
		       ret);
		isert_conn->login_req_dma = 0;
		goto out_login_buf;
	}

	isert_conn->login_rsp_dma = ib_dma_map_single(ib_dev,
					(void *)isert_conn->login_rsp_buf,
					ISER_RX_LOGIN_SIZE, DMA_TO_DEVICE);

	ret = ib_dma_mapping_error(ib_dev, isert_conn->login_rsp_dma);
	if (ret) {
		pr_err("ib_dma_mapping_error failed for login_rsp_dma: %d\n",
		       ret);
		isert_conn->login_rsp_dma = 0;
		goto out_req_dma_map;
	}

	device = isert_device_find_by_ib_dev(cma_id);
	if (IS_ERR(device)) {
		ret = PTR_ERR(device);
		goto out_rsp_dma_map;
	}

649 650 651 652 653 654
	/* Set max inflight RDMA READ requests */
	isert_conn->initiator_depth = min_t(u8,
				event->param.conn.initiator_depth,
				device->dev_attr.max_qp_init_rd_atom);
	pr_debug("Using initiator_depth: %u\n", isert_conn->initiator_depth);

655
	isert_conn->conn_device = device;
656 657 658 659 660 661 662
	isert_conn->conn_pd = ib_alloc_pd(isert_conn->conn_device->ib_device);
	if (IS_ERR(isert_conn->conn_pd)) {
		ret = PTR_ERR(isert_conn->conn_pd);
		pr_err("ib_alloc_pd failed for conn %p: ret=%d\n",
		       isert_conn, ret);
		goto out_pd;
	}
663

664 665 666 667 668 669 670 671
	isert_conn->conn_mr = ib_get_dma_mr(isert_conn->conn_pd,
					   IB_ACCESS_LOCAL_WRITE);
	if (IS_ERR(isert_conn->conn_mr)) {
		ret = PTR_ERR(isert_conn->conn_mr);
		pr_err("ib_get_dma_mr failed for conn %p: ret=%d\n",
		       isert_conn, ret);
		goto out_mr;
	}
672

673
	ret = isert_conn_setup_qp(isert_conn, cma_id);
674 675 676
	if (ret)
		goto out_conn_dev;

677 678 679 680 681 682 683 684
	ret = isert_rdma_post_recvl(isert_conn);
	if (ret)
		goto out_conn_dev;

	ret = isert_rdma_accept(isert_conn);
	if (ret)
		goto out_conn_dev;

685
	mutex_lock(&isert_np->np_accept_mutex);
686
	list_add_tail(&isert_conn->conn_accept_node, &isert_np->np_accept_list);
687 688
	mutex_unlock(&isert_np->np_accept_mutex);

689 690
	pr_debug("isert_connect_request() up np_sem np: %p\n", np);
	up(&isert_np->np_sem);
691 692 693
	return 0;

out_conn_dev:
694 695 696 697
	ib_dereg_mr(isert_conn->conn_mr);
out_mr:
	ib_dealloc_pd(isert_conn->conn_pd);
out_pd:
698 699 700 701 702 703 704 705 706 707 708
	isert_device_try_release(device);
out_rsp_dma_map:
	ib_dma_unmap_single(ib_dev, isert_conn->login_rsp_dma,
			    ISER_RX_LOGIN_SIZE, DMA_TO_DEVICE);
out_req_dma_map:
	ib_dma_unmap_single(ib_dev, isert_conn->login_req_dma,
			    ISCSI_DEF_MAX_RECV_SEG_LEN, DMA_FROM_DEVICE);
out_login_buf:
	kfree(isert_conn->login_buf);
out:
	kfree(isert_conn);
709
	rdma_reject(cma_id, NULL, 0);
710 711 712 713 714 715 716 717 718 719 720
	return ret;
}

static void
isert_connect_release(struct isert_conn *isert_conn)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct isert_device *device = isert_conn->conn_device;

	pr_debug("Entering isert_connect_release(): >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>\n");

721 722
	if (device && device->use_fastreg)
		isert_conn_free_fastreg_pool(isert_conn);
723

724 725 726
	isert_free_rx_descriptors(isert_conn);
	rdma_destroy_id(isert_conn->conn_cm_id);

727
	if (isert_conn->conn_qp) {
728 729 730
		struct isert_comp *comp = isert_conn->conn_qp->recv_cq->cq_context;

		pr_debug("dec completion context %p active_qps\n", comp);
731
		mutex_lock(&device_list_mutex);
732
		comp->active_qps--;
733
		mutex_unlock(&device_list_mutex);
734

735
		ib_destroy_qp(isert_conn->conn_qp);
736 737
	}

738 739 740
	ib_dereg_mr(isert_conn->conn_mr);
	ib_dealloc_pd(isert_conn->conn_pd);

741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759
	if (isert_conn->login_buf) {
		ib_dma_unmap_single(ib_dev, isert_conn->login_rsp_dma,
				    ISER_RX_LOGIN_SIZE, DMA_TO_DEVICE);
		ib_dma_unmap_single(ib_dev, isert_conn->login_req_dma,
				    ISCSI_DEF_MAX_RECV_SEG_LEN,
				    DMA_FROM_DEVICE);
		kfree(isert_conn->login_buf);
	}
	kfree(isert_conn);

	if (device)
		isert_device_try_release(device);

	pr_debug("Leaving isert_connect_release >>>>>>>>>>>>\n");
}

static void
isert_connected_handler(struct rdma_cm_id *cma_id)
{
760
	struct isert_conn *isert_conn = cma_id->qp->qp_context;
761

762 763
	pr_info("conn %p\n", isert_conn);

764 765 766 767 768 769 770 771 772
	if (!kref_get_unless_zero(&isert_conn->conn_kref)) {
		pr_warn("conn %p connect_release is running\n", isert_conn);
		return;
	}

	mutex_lock(&isert_conn->conn_mutex);
	if (isert_conn->state != ISER_CONN_FULL_FEATURE)
		isert_conn->state = ISER_CONN_UP;
	mutex_unlock(&isert_conn->conn_mutex);
773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
}

static void
isert_release_conn_kref(struct kref *kref)
{
	struct isert_conn *isert_conn = container_of(kref,
				struct isert_conn, conn_kref);

	pr_debug("Calling isert_connect_release for final kref %s/%d\n",
		 current->comm, current->pid);

	isert_connect_release(isert_conn);
}

static void
isert_put_conn(struct isert_conn *isert_conn)
{
	kref_put(&isert_conn->conn_kref, isert_release_conn_kref);
}

793 794 795 796 797
/**
 * isert_conn_terminate() - Initiate connection termination
 * @isert_conn: isert connection struct
 *
 * Notes:
798
 * In case the connection state is FULL_FEATURE, move state
799
 * to TEMINATING and start teardown sequence (rdma_disconnect).
800
 * In case the connection state is UP, complete flush as well.
801 802 803 804 805 806 807 808 809
 *
 * This routine must be called with conn_mutex held. Thus it is
 * safe to call multiple times.
 */
static void
isert_conn_terminate(struct isert_conn *isert_conn)
{
	int err;

810 811 812 813 814
	switch (isert_conn->state) {
	case ISER_CONN_TERMINATING:
		break;
	case ISER_CONN_UP:
	case ISER_CONN_FULL_FEATURE: /* FALLTHRU */
815 816
		pr_info("Terminating conn %p state %d\n",
			   isert_conn, isert_conn->state);
817
		isert_conn->state = ISER_CONN_TERMINATING;
818 819 820 821
		err = rdma_disconnect(isert_conn->conn_cm_id);
		if (err)
			pr_warn("Failed rdma_disconnect isert_conn %p\n",
				   isert_conn);
822 823 824 825
		break;
	default:
		pr_warn("conn %p teminating in state %d\n",
			   isert_conn, isert_conn->state);
826 827 828
	}
}

829
static int
830 831
isert_np_cma_handler(struct isert_np *isert_np,
		     enum rdma_cm_event_type event)
832
{
833
	pr_debug("isert np %p, handling event %d\n", isert_np, event);
834

835 836
	switch (event) {
	case RDMA_CM_EVENT_DEVICE_REMOVAL:
837
		isert_np->np_cm_id = NULL;
838 839 840 841 842 843 844 845 846 847 848 849
		break;
	case RDMA_CM_EVENT_ADDR_CHANGE:
		isert_np->np_cm_id = isert_setup_id(isert_np);
		if (IS_ERR(isert_np->np_cm_id)) {
			pr_err("isert np %p setup id failed: %ld\n",
				 isert_np, PTR_ERR(isert_np->np_cm_id));
			isert_np->np_cm_id = NULL;
		}
		break;
	default:
		pr_err("isert np %p Unexpected event %d\n",
			  isert_np, event);
850 851
	}

852 853 854 855 856 857 858 859 860 861 862 863 864
	return -1;
}

static int
isert_disconnected_handler(struct rdma_cm_id *cma_id,
			   enum rdma_cm_event_type event)
{
	struct isert_np *isert_np = cma_id->context;
	struct isert_conn *isert_conn;

	if (isert_np->np_cm_id == cma_id)
		return isert_np_cma_handler(cma_id->context, event);

865
	isert_conn = cma_id->qp->qp_context;
866

867 868 869 870 871 872
	mutex_lock(&isert_conn->conn_mutex);
	isert_conn_terminate(isert_conn);
	mutex_unlock(&isert_conn->conn_mutex);

	pr_info("conn %p completing conn_wait\n", isert_conn);
	complete(&isert_conn->conn_wait);
873 874

	return 0;
875 876
}

877 878 879
static void
isert_connect_error(struct rdma_cm_id *cma_id)
{
880
	struct isert_conn *isert_conn = cma_id->qp->qp_context;
881 882 883 884

	isert_put_conn(isert_conn);
}

885 886 887 888 889 890 891 892 893 894 895
static int
isert_cma_handler(struct rdma_cm_id *cma_id, struct rdma_cm_event *event)
{
	int ret = 0;

	pr_debug("isert_cma_handler: event %d status %d conn %p id %p\n",
		 event->event, event->status, cma_id->context, cma_id);

	switch (event->event) {
	case RDMA_CM_EVENT_CONNECT_REQUEST:
		ret = isert_connect_request(cma_id, event);
896 897 898
		if (ret)
			pr_err("isert_cma_handler failed RDMA_CM_EVENT: 0x%08x %d\n",
				event->event, ret);
899 900 901 902
		break;
	case RDMA_CM_EVENT_ESTABLISHED:
		isert_connected_handler(cma_id);
		break;
903 904 905 906
	case RDMA_CM_EVENT_ADDR_CHANGE:    /* FALLTHRU */
	case RDMA_CM_EVENT_DISCONNECTED:   /* FALLTHRU */
	case RDMA_CM_EVENT_DEVICE_REMOVAL: /* FALLTHRU */
	case RDMA_CM_EVENT_TIMEWAIT_EXIT:  /* FALLTHRU */
907
		ret = isert_disconnected_handler(cma_id, event->event);
908
		break;
909 910
	case RDMA_CM_EVENT_REJECTED:       /* FALLTHRU */
	case RDMA_CM_EVENT_UNREACHABLE:    /* FALLTHRU */
911
	case RDMA_CM_EVENT_CONNECT_ERROR:
912 913
		isert_connect_error(cma_id);
		break;
914
	default:
915
		pr_err("Unhandled RDMA CMA event: %d\n", event->event);
916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931
		break;
	}

	return ret;
}

static int
isert_post_recv(struct isert_conn *isert_conn, u32 count)
{
	struct ib_recv_wr *rx_wr, *rx_wr_failed;
	int i, ret;
	unsigned int rx_head = isert_conn->conn_rx_desc_head;
	struct iser_rx_desc *rx_desc;

	for (rx_wr = isert_conn->conn_rx_wr, i = 0; i < count; i++, rx_wr++) {
		rx_desc		= &isert_conn->conn_rx_descs[rx_head];
932
		rx_wr->wr_id	= (uintptr_t)rx_desc;
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
		rx_wr->sg_list	= &rx_desc->rx_sg;
		rx_wr->num_sge	= 1;
		rx_wr->next	= rx_wr + 1;
		rx_head = (rx_head + 1) & (ISERT_QP_MAX_RECV_DTOS - 1);
	}

	rx_wr--;
	rx_wr->next = NULL; /* mark end of work requests list */

	isert_conn->post_recv_buf_count += count;
	ret = ib_post_recv(isert_conn->conn_qp, isert_conn->conn_rx_wr,
				&rx_wr_failed);
	if (ret) {
		pr_err("ib_post_recv() failed with ret: %d\n", ret);
		isert_conn->post_recv_buf_count -= count;
	} else {
		pr_debug("isert_post_recv(): Posted %d RX buffers\n", count);
		isert_conn->conn_rx_desc_head = rx_head;
	}
	return ret;
}

static int
isert_post_send(struct isert_conn *isert_conn, struct iser_tx_desc *tx_desc)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct ib_send_wr send_wr, *send_wr_failed;
	int ret;

	ib_dma_sync_single_for_device(ib_dev, tx_desc->dma_addr,
				      ISER_HEADERS_LEN, DMA_TO_DEVICE);

	send_wr.next	= NULL;
966
	send_wr.wr_id	= (uintptr_t)tx_desc;
967 968 969 970 971 972
	send_wr.sg_list	= tx_desc->tx_sg;
	send_wr.num_sge	= tx_desc->num_sge;
	send_wr.opcode	= IB_WR_SEND;
	send_wr.send_flags = IB_SEND_SIGNALED;

	ret = ib_post_send(isert_conn->conn_qp, &send_wr, &send_wr_failed);
973
	if (ret)
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
		pr_err("ib_post_send() failed, ret: %d\n", ret);

	return ret;
}

static void
isert_create_send_desc(struct isert_conn *isert_conn,
		       struct isert_cmd *isert_cmd,
		       struct iser_tx_desc *tx_desc)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;

	ib_dma_sync_single_for_cpu(ib_dev, tx_desc->dma_addr,
				   ISER_HEADERS_LEN, DMA_TO_DEVICE);

	memset(&tx_desc->iser_header, 0, sizeof(struct iser_hdr));
	tx_desc->iser_header.flags = ISER_VER;

	tx_desc->num_sge = 1;
	tx_desc->isert_cmd = isert_cmd;

	if (tx_desc->tx_sg[0].lkey != isert_conn->conn_mr->lkey) {
		tx_desc->tx_sg[0].lkey = isert_conn->conn_mr->lkey;
		pr_debug("tx_desc %p lkey mismatch, fixing\n", tx_desc);
	}
}

static int
isert_init_tx_hdrs(struct isert_conn *isert_conn,
		   struct iser_tx_desc *tx_desc)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	u64 dma_addr;

	dma_addr = ib_dma_map_single(ib_dev, (void *)tx_desc,
			ISER_HEADERS_LEN, DMA_TO_DEVICE);
	if (ib_dma_mapping_error(ib_dev, dma_addr)) {
		pr_err("ib_dma_mapping_error() failed\n");
		return -ENOMEM;
	}

	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 = isert_conn->conn_mr->lkey;

	pr_debug("isert_init_tx_hdrs: Setup tx_sg[0].addr: 0x%llx length: %u"
		 " lkey: 0x%08x\n", tx_desc->tx_sg[0].addr,
		 tx_desc->tx_sg[0].length, tx_desc->tx_sg[0].lkey);

	return 0;
}

static void
1028
isert_init_send_wr(struct isert_conn *isert_conn, struct isert_cmd *isert_cmd,
1029
		   struct ib_send_wr *send_wr)
1030
{
1031 1032
	struct iser_tx_desc *tx_desc = &isert_cmd->tx_desc;

1033
	isert_cmd->rdma_wr.iser_ib_op = ISER_IB_SEND;
1034
	send_wr->wr_id = (uintptr_t)&isert_cmd->tx_desc;
1035
	send_wr->opcode = IB_WR_SEND;
1036
	send_wr->sg_list = &tx_desc->tx_sg[0];
1037
	send_wr->num_sge = isert_cmd->tx_desc.num_sge;
1038
	send_wr->send_flags = IB_SEND_SIGNALED;
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056
}

static int
isert_rdma_post_recvl(struct isert_conn *isert_conn)
{
	struct ib_recv_wr rx_wr, *rx_wr_fail;
	struct ib_sge sge;
	int ret;

	memset(&sge, 0, sizeof(struct ib_sge));
	sge.addr = isert_conn->login_req_dma;
	sge.length = ISER_RX_LOGIN_SIZE;
	sge.lkey = isert_conn->conn_mr->lkey;

	pr_debug("Setup sge: addr: %llx length: %d 0x%08x\n",
		sge.addr, sge.length, sge.lkey);

	memset(&rx_wr, 0, sizeof(struct ib_recv_wr));
1057
	rx_wr.wr_id = (uintptr_t)isert_conn->login_req_buf;
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
	rx_wr.sg_list = &sge;
	rx_wr.num_sge = 1;

	isert_conn->post_recv_buf_count++;
	ret = ib_post_recv(isert_conn->conn_qp, &rx_wr, &rx_wr_fail);
	if (ret) {
		pr_err("ib_post_recv() failed: %d\n", ret);
		isert_conn->post_recv_buf_count--;
	}

	pr_debug("ib_post_recv(): returned success >>>>>>>>>>>>>>>>>>>>>>>>\n");
	return ret;
}

static int
isert_put_login_tx(struct iscsi_conn *conn, struct iscsi_login *login,
		   u32 length)
{
	struct isert_conn *isert_conn = conn->context;
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct iser_tx_desc *tx_desc = &isert_conn->conn_login_tx_desc;
	int ret;

	isert_create_send_desc(isert_conn, NULL, tx_desc);

	memcpy(&tx_desc->iscsi_header, &login->rsp[0],
	       sizeof(struct iscsi_hdr));

	isert_init_tx_hdrs(isert_conn, tx_desc);

	if (length > 0) {
		struct ib_sge *tx_dsg = &tx_desc->tx_sg[1];

		ib_dma_sync_single_for_cpu(ib_dev, isert_conn->login_rsp_dma,
					   length, DMA_TO_DEVICE);

		memcpy(isert_conn->login_rsp_buf, login->rsp_buf, length);

		ib_dma_sync_single_for_device(ib_dev, isert_conn->login_rsp_dma,
					      length, DMA_TO_DEVICE);

		tx_dsg->addr	= isert_conn->login_rsp_dma;
		tx_dsg->length	= length;
		tx_dsg->lkey	= isert_conn->conn_mr->lkey;
		tx_desc->num_sge = 2;
	}
	if (!login->login_failed) {
		if (login->login_complete) {
1106 1107
			if (!conn->sess->sess_ops->SessionType &&
			    isert_conn->conn_device->use_fastreg) {
1108
				ret = isert_conn_create_fastreg_pool(isert_conn);
1109 1110 1111 1112 1113 1114 1115
				if (ret) {
					pr_err("Conn: %p failed to create"
					       " fastreg pool\n", isert_conn);
					return ret;
				}
			}

1116 1117 1118 1119 1120 1121 1122 1123
			ret = isert_alloc_rx_descriptors(isert_conn);
			if (ret)
				return ret;

			ret = isert_post_recv(isert_conn, ISERT_MIN_POSTED_RX);
			if (ret)
				return ret;

1124
			/* Now we are in FULL_FEATURE phase */
1125
			mutex_lock(&isert_conn->conn_mutex);
1126
			isert_conn->state = ISER_CONN_FULL_FEATURE;
1127
			mutex_unlock(&isert_conn->conn_mutex);
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
			goto post_send;
		}

		ret = isert_rdma_post_recvl(isert_conn);
		if (ret)
			return ret;
	}
post_send:
	ret = isert_post_send(isert_conn, tx_desc);
	if (ret)
		return ret;

	return 0;
}

static void
1144
isert_rx_login_req(struct isert_conn *isert_conn)
1145
{
1146 1147
	struct iser_rx_desc *rx_desc = (void *)isert_conn->login_req_buf;
	int rx_buflen = isert_conn->login_req_len;
1148 1149 1150 1151
	struct iscsi_conn *conn = isert_conn->conn;
	struct iscsi_login *login = conn->conn_login;
	int size;

1152 1153 1154
	pr_info("conn %p\n", isert_conn);

	WARN_ON_ONCE(!login);
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183

	if (login->first_request) {
		struct iscsi_login_req *login_req =
			(struct iscsi_login_req *)&rx_desc->iscsi_header;
		/*
		 * Setup the initial iscsi_login values from the leading
		 * login request PDU.
		 */
		login->leading_connection = (!login_req->tsih) ? 1 : 0;
		login->current_stage =
			(login_req->flags & ISCSI_FLAG_LOGIN_CURRENT_STAGE_MASK)
			 >> 2;
		login->version_min	= login_req->min_version;
		login->version_max	= login_req->max_version;
		memcpy(login->isid, login_req->isid, 6);
		login->cmd_sn		= be32_to_cpu(login_req->cmdsn);
		login->init_task_tag	= login_req->itt;
		login->initial_exp_statsn = be32_to_cpu(login_req->exp_statsn);
		login->cid		= be16_to_cpu(login_req->cid);
		login->tsih		= be16_to_cpu(login_req->tsih);
	}

	memcpy(&login->req[0], (void *)&rx_desc->iscsi_header, ISCSI_HDR_LEN);

	size = min(rx_buflen, MAX_KEY_VALUE_PAIRS);
	pr_debug("Using login payload size: %d, rx_buflen: %d MAX_KEY_VALUE_PAIRS: %d\n",
		 size, rx_buflen, MAX_KEY_VALUE_PAIRS);
	memcpy(login->req_buf, &rx_desc->data[0], size);

1184 1185 1186 1187 1188
	if (login->first_request) {
		complete(&isert_conn->conn_login_comp);
		return;
	}
	schedule_delayed_work(&conn->login_work, 0);
1189 1190 1191
}

static struct iscsi_cmd
1192
*isert_allocate_cmd(struct iscsi_conn *conn)
1193 1194 1195
{
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct isert_cmd *isert_cmd;
1196
	struct iscsi_cmd *cmd;
1197

1198
	cmd = iscsit_allocate_cmd(conn, TASK_INTERRUPTIBLE);
1199 1200
	if (!cmd) {
		pr_err("Unable to allocate iscsi_cmd + isert_cmd\n");
1201 1202
		return NULL;
	}
1203
	isert_cmd = iscsit_priv_cmd(cmd);
1204
	isert_cmd->conn = isert_conn;
1205
	isert_cmd->iscsi_cmd = cmd;
1206

1207
	return cmd;
1208 1209 1210 1211
}

static int
isert_handle_scsi_cmd(struct isert_conn *isert_conn,
1212 1213
		      struct isert_cmd *isert_cmd, struct iscsi_cmd *cmd,
		      struct iser_rx_desc *rx_desc, unsigned char *buf)
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257
{
	struct iscsi_conn *conn = isert_conn->conn;
	struct iscsi_scsi_req *hdr = (struct iscsi_scsi_req *)buf;
	struct scatterlist *sg;
	int imm_data, imm_data_len, unsol_data, sg_nents, rc;
	bool dump_payload = false;

	rc = iscsit_setup_scsi_cmd(conn, cmd, buf);
	if (rc < 0)
		return rc;

	imm_data = cmd->immediate_data;
	imm_data_len = cmd->first_burst_len;
	unsol_data = cmd->unsolicited_data;

	rc = iscsit_process_scsi_cmd(conn, cmd, hdr);
	if (rc < 0) {
		return 0;
	} else if (rc > 0) {
		dump_payload = true;
		goto sequence_cmd;
	}

	if (!imm_data)
		return 0;

	sg = &cmd->se_cmd.t_data_sg[0];
	sg_nents = max(1UL, DIV_ROUND_UP(imm_data_len, PAGE_SIZE));

	pr_debug("Copying Immediate SG: %p sg_nents: %u from %p imm_data_len: %d\n",
		 sg, sg_nents, &rx_desc->data[0], imm_data_len);

	sg_copy_from_buffer(sg, sg_nents, &rx_desc->data[0], imm_data_len);

	cmd->write_data_done += imm_data_len;

	if (cmd->write_data_done == cmd->se_cmd.data_length) {
		spin_lock_bh(&cmd->istate_lock);
		cmd->cmd_flags |= ICF_GOT_LAST_DATAOUT;
		cmd->i_state = ISTATE_RECEIVED_LAST_DATAOUT;
		spin_unlock_bh(&cmd->istate_lock);
	}

sequence_cmd:
1258
	rc = iscsit_sequence_cmd(conn, cmd, buf, hdr->cmdsn);
1259 1260 1261

	if (!rc && dump_payload == false && unsol_data)
		iscsit_set_unsoliticed_dataout(cmd);
1262 1263
	else if (dump_payload && imm_data)
		target_put_sess_cmd(conn->sess->se_sess, &cmd->se_cmd);
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320

	return 0;
}

static int
isert_handle_iscsi_dataout(struct isert_conn *isert_conn,
			   struct iser_rx_desc *rx_desc, unsigned char *buf)
{
	struct scatterlist *sg_start;
	struct iscsi_conn *conn = isert_conn->conn;
	struct iscsi_cmd *cmd = NULL;
	struct iscsi_data *hdr = (struct iscsi_data *)buf;
	u32 unsol_data_len = ntoh24(hdr->dlength);
	int rc, sg_nents, sg_off, page_off;

	rc = iscsit_check_dataout_hdr(conn, buf, &cmd);
	if (rc < 0)
		return rc;
	else if (!cmd)
		return 0;
	/*
	 * FIXME: Unexpected unsolicited_data out
	 */
	if (!cmd->unsolicited_data) {
		pr_err("Received unexpected solicited data payload\n");
		dump_stack();
		return -1;
	}

	pr_debug("Unsolicited DataOut unsol_data_len: %u, write_data_done: %u, data_length: %u\n",
		 unsol_data_len, cmd->write_data_done, cmd->se_cmd.data_length);

	sg_off = cmd->write_data_done / PAGE_SIZE;
	sg_start = &cmd->se_cmd.t_data_sg[sg_off];
	sg_nents = max(1UL, DIV_ROUND_UP(unsol_data_len, PAGE_SIZE));
	page_off = cmd->write_data_done % PAGE_SIZE;
	/*
	 * FIXME: Non page-aligned unsolicited_data out
	 */
	if (page_off) {
		pr_err("Received unexpected non-page aligned data payload\n");
		dump_stack();
		return -1;
	}
	pr_debug("Copying DataOut: sg_start: %p, sg_off: %u sg_nents: %u from %p %u\n",
		 sg_start, sg_off, sg_nents, &rx_desc->data[0], unsol_data_len);

	sg_copy_from_buffer(sg_start, sg_nents, &rx_desc->data[0],
			    unsol_data_len);

	rc = iscsit_check_dataout_payload(cmd, hdr, false);
	if (rc < 0)
		return rc;

	return 0;
}

1321 1322
static int
isert_handle_nop_out(struct isert_conn *isert_conn, struct isert_cmd *isert_cmd,
1323 1324
		     struct iscsi_cmd *cmd, struct iser_rx_desc *rx_desc,
		     unsigned char *buf)
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339
{
	struct iscsi_conn *conn = isert_conn->conn;
	struct iscsi_nopout *hdr = (struct iscsi_nopout *)buf;
	int rc;

	rc = iscsit_setup_nop_out(conn, cmd, hdr);
	if (rc < 0)
		return rc;
	/*
	 * FIXME: Add support for NOPOUT payload using unsolicited RDMA payload
	 */

	return iscsit_process_nop_out(conn, cmd, hdr);
}

1340 1341
static int
isert_handle_text_cmd(struct isert_conn *isert_conn, struct isert_cmd *isert_cmd,
1342 1343
		      struct iscsi_cmd *cmd, struct iser_rx_desc *rx_desc,
		      struct iscsi_text *hdr)
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
{
	struct iscsi_conn *conn = isert_conn->conn;
	u32 payload_length = ntoh24(hdr->dlength);
	int rc;
	unsigned char *text_in;

	rc = iscsit_setup_text_cmd(conn, cmd, hdr);
	if (rc < 0)
		return rc;

	text_in = kzalloc(payload_length, GFP_KERNEL);
	if (!text_in) {
		pr_err("Unable to allocate text_in of payload_length: %u\n",
		       payload_length);
		return -ENOMEM;
	}
	cmd->text_in_ptr = text_in;

	memcpy(cmd->text_in_ptr, &rx_desc->data[0], payload_length);

	return iscsit_process_text_cmd(conn, cmd, hdr);
}

1367 1368 1369 1370 1371 1372 1373
static int
isert_rx_opcode(struct isert_conn *isert_conn, struct iser_rx_desc *rx_desc,
		uint32_t read_stag, uint64_t read_va,
		uint32_t write_stag, uint64_t write_va)
{
	struct iscsi_hdr *hdr = &rx_desc->iscsi_header;
	struct iscsi_conn *conn = isert_conn->conn;
1374
	struct iscsi_session *sess = conn->sess;
1375 1376 1377 1378 1379
	struct iscsi_cmd *cmd;
	struct isert_cmd *isert_cmd;
	int ret = -EINVAL;
	u8 opcode = (hdr->opcode & ISCSI_OPCODE_MASK);

1380 1381 1382 1383 1384 1385 1386
	if (sess->sess_ops->SessionType &&
	   (!(opcode & ISCSI_OP_TEXT) || !(opcode & ISCSI_OP_LOGOUT))) {
		pr_err("Got illegal opcode: 0x%02x in SessionType=Discovery,"
		       " ignoring\n", opcode);
		return 0;
	}

1387 1388
	switch (opcode) {
	case ISCSI_OP_SCSI_CMD:
1389
		cmd = isert_allocate_cmd(conn);
1390 1391 1392
		if (!cmd)
			break;

1393
		isert_cmd = iscsit_priv_cmd(cmd);
1394 1395 1396 1397 1398
		isert_cmd->read_stag = read_stag;
		isert_cmd->read_va = read_va;
		isert_cmd->write_stag = write_stag;
		isert_cmd->write_va = write_va;

1399
		ret = isert_handle_scsi_cmd(isert_conn, isert_cmd, cmd,
1400 1401 1402
					rx_desc, (unsigned char *)hdr);
		break;
	case ISCSI_OP_NOOP_OUT:
1403
		cmd = isert_allocate_cmd(conn);
1404 1405 1406
		if (!cmd)
			break;

1407 1408
		isert_cmd = iscsit_priv_cmd(cmd);
		ret = isert_handle_nop_out(isert_conn, isert_cmd, cmd,
1409
					   rx_desc, (unsigned char *)hdr);
1410 1411 1412 1413 1414 1415
		break;
	case ISCSI_OP_SCSI_DATA_OUT:
		ret = isert_handle_iscsi_dataout(isert_conn, rx_desc,
						(unsigned char *)hdr);
		break;
	case ISCSI_OP_SCSI_TMFUNC:
1416
		cmd = isert_allocate_cmd(conn);
1417 1418 1419 1420 1421 1422 1423
		if (!cmd)
			break;

		ret = iscsit_handle_task_mgt_cmd(conn, cmd,
						(unsigned char *)hdr);
		break;
	case ISCSI_OP_LOGOUT:
1424
		cmd = isert_allocate_cmd(conn);
1425 1426 1427 1428 1429
		if (!cmd)
			break;

		ret = iscsit_handle_logout_cmd(conn, cmd, (unsigned char *)hdr);
		break;
1430
	case ISCSI_OP_TEXT:
1431
		cmd = isert_allocate_cmd(conn);
1432 1433 1434
		if (!cmd)
			break;

1435 1436
		isert_cmd = iscsit_priv_cmd(cmd);
		ret = isert_handle_text_cmd(isert_conn, isert_cmd, cmd,
1437 1438
					    rx_desc, (struct iscsi_text *)hdr);
		break;
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486
	default:
		pr_err("Got unknown iSCSI OpCode: 0x%02x\n", opcode);
		dump_stack();
		break;
	}

	return ret;
}

static void
isert_rx_do_work(struct iser_rx_desc *rx_desc, struct isert_conn *isert_conn)
{
	struct iser_hdr *iser_hdr = &rx_desc->iser_header;
	uint64_t read_va = 0, write_va = 0;
	uint32_t read_stag = 0, write_stag = 0;
	int rc;

	switch (iser_hdr->flags & 0xF0) {
	case ISCSI_CTRL:
		if (iser_hdr->flags & ISER_RSV) {
			read_stag = be32_to_cpu(iser_hdr->read_stag);
			read_va = be64_to_cpu(iser_hdr->read_va);
			pr_debug("ISER_RSV: read_stag: 0x%08x read_va: 0x%16llx\n",
				 read_stag, (unsigned long long)read_va);
		}
		if (iser_hdr->flags & ISER_WSV) {
			write_stag = be32_to_cpu(iser_hdr->write_stag);
			write_va = be64_to_cpu(iser_hdr->write_va);
			pr_debug("ISER_WSV: write__stag: 0x%08x write_va: 0x%16llx\n",
				 write_stag, (unsigned long long)write_va);
		}

		pr_debug("ISER ISCSI_CTRL PDU\n");
		break;
	case ISER_HELLO:
		pr_err("iSER Hello message\n");
		break;
	default:
		pr_warn("Unknown iSER hdr flags: 0x%02x\n", iser_hdr->flags);
		break;
	}

	rc = isert_rx_opcode(isert_conn, rx_desc,
			     read_stag, read_va, write_stag, write_va);
}

static void
isert_rx_completion(struct iser_rx_desc *desc, struct isert_conn *isert_conn,
1487
		    u32 xfer_len)
1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct iscsi_hdr *hdr;
	u64 rx_dma;
	int rx_buflen, outstanding;

	if ((char *)desc == isert_conn->login_req_buf) {
		rx_dma = isert_conn->login_req_dma;
		rx_buflen = ISER_RX_LOGIN_SIZE;
		pr_debug("ISER login_buf: Using rx_dma: 0x%llx, rx_buflen: %d\n",
			 rx_dma, rx_buflen);
	} else {
		rx_dma = desc->dma_addr;
		rx_buflen = ISER_RX_PAYLOAD_SIZE;
		pr_debug("ISER req_buf: Using rx_dma: 0x%llx, rx_buflen: %d\n",
			 rx_dma, rx_buflen);
	}

	ib_dma_sync_single_for_cpu(ib_dev, rx_dma, rx_buflen, DMA_FROM_DEVICE);

	hdr = &desc->iscsi_header;
	pr_debug("iSCSI opcode: 0x%02x, ITT: 0x%08x, flags: 0x%02x dlen: %d\n",
		 hdr->opcode, hdr->itt, hdr->flags,
		 (int)(xfer_len - ISER_HEADERS_LEN));

1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
	if ((char *)desc == isert_conn->login_req_buf) {
		isert_conn->login_req_len = xfer_len - ISER_HEADERS_LEN;
		if (isert_conn->conn) {
			struct iscsi_login *login = isert_conn->conn->conn_login;

			if (login && !login->first_request)
				isert_rx_login_req(isert_conn);
		}
		mutex_lock(&isert_conn->conn_mutex);
		complete(&isert_conn->login_req_comp);
		mutex_unlock(&isert_conn->conn_mutex);
	} else {
1525
		isert_rx_do_work(desc, isert_conn);
1526
	}
1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549

	ib_dma_sync_single_for_device(ib_dev, rx_dma, rx_buflen,
				      DMA_FROM_DEVICE);

	isert_conn->post_recv_buf_count--;
	pr_debug("iSERT: Decremented post_recv_buf_count: %d\n",
		 isert_conn->post_recv_buf_count);

	if ((char *)desc == isert_conn->login_req_buf)
		return;

	outstanding = isert_conn->post_recv_buf_count;
	if (outstanding + ISERT_MIN_POSTED_RX <= ISERT_QP_MAX_RECV_DTOS) {
		int err, count = min(ISERT_QP_MAX_RECV_DTOS - outstanding,
				ISERT_MIN_POSTED_RX);
		err = isert_post_recv(isert_conn, count);
		if (err) {
			pr_err("isert_post_recv() count: %d failed, %d\n",
			       count, err);
		}
	}
}

1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
static int
isert_map_data_buf(struct isert_conn *isert_conn, struct isert_cmd *isert_cmd,
		   struct scatterlist *sg, u32 nents, u32 length, u32 offset,
		   enum iser_ib_op_code op, struct isert_data_buf *data)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;

	data->dma_dir = op == ISER_IB_RDMA_WRITE ?
			      DMA_TO_DEVICE : DMA_FROM_DEVICE;

	data->len = length - offset;
	data->offset = offset;
	data->sg_off = data->offset / PAGE_SIZE;

	data->sg = &sg[data->sg_off];
	data->nents = min_t(unsigned int, nents - data->sg_off,
					  ISCSI_ISER_SG_TABLESIZE);
	data->len = min_t(unsigned int, data->len, ISCSI_ISER_SG_TABLESIZE *
					PAGE_SIZE);

	data->dma_nents = ib_dma_map_sg(ib_dev, data->sg, data->nents,
					data->dma_dir);
	if (unlikely(!data->dma_nents)) {
		pr_err("Cmd: unable to dma map SGs %p\n", sg);
		return -EINVAL;
	}

	pr_debug("Mapped cmd: %p count: %u sg: %p sg_nents: %u rdma_len %d\n",
		 isert_cmd, data->dma_nents, data->sg, data->nents, data->len);

	return 0;
}

static void
isert_unmap_data_buf(struct isert_conn *isert_conn, struct isert_data_buf *data)
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;

	ib_dma_unmap_sg(ib_dev, data->sg, data->nents, data->dma_dir);
	memset(data, 0, sizeof(*data));
}



1594 1595 1596 1597 1598
static void
isert_unmap_cmd(struct isert_cmd *isert_cmd, struct isert_conn *isert_conn)
{
	struct isert_rdma_wr *wr = &isert_cmd->rdma_wr;

1599
	pr_debug("isert_unmap_cmd: %p\n", isert_cmd);
1600 1601

	if (wr->data.sg) {
1602
		pr_debug("isert_unmap_cmd: %p unmap_sg op\n", isert_cmd);
1603
		isert_unmap_data_buf(isert_conn, &wr->data);
1604 1605
	}

1606 1607 1608 1609 1610
	if (wr->send_wr) {
		pr_debug("isert_unmap_cmd: %p free send_wr\n", isert_cmd);
		kfree(wr->send_wr);
		wr->send_wr = NULL;
	}
1611

1612 1613 1614 1615 1616
	if (wr->ib_sge) {
		pr_debug("isert_unmap_cmd: %p free ib_sge\n", isert_cmd);
		kfree(wr->ib_sge);
		wr->ib_sge = NULL;
	}
1617 1618
}

1619
static void
1620
isert_unreg_rdma(struct isert_cmd *isert_cmd, struct isert_conn *isert_conn)
1621 1622 1623 1624
{
	struct isert_rdma_wr *wr = &isert_cmd->rdma_wr;
	LIST_HEAD(unmap_list);

1625
	pr_debug("unreg_fastreg_cmd: %p\n", isert_cmd);
1626 1627

	if (wr->fr_desc) {
1628
		pr_debug("unreg_fastreg_cmd: %p free fr_desc %p\n",
1629
			 isert_cmd, wr->fr_desc);
1630 1631 1632 1633
		if (wr->fr_desc->ind & ISERT_PROTECTED) {
			isert_unmap_data_buf(isert_conn, &wr->prot);
			wr->fr_desc->ind &= ~ISERT_PROTECTED;
		}
1634
		spin_lock_bh(&isert_conn->conn_lock);
1635
		list_add_tail(&wr->fr_desc->list, &isert_conn->conn_fr_pool);
1636 1637 1638 1639
		spin_unlock_bh(&isert_conn->conn_lock);
		wr->fr_desc = NULL;
	}

1640
	if (wr->data.sg) {
1641
		pr_debug("unreg_fastreg_cmd: %p unmap_sg op\n", isert_cmd);
1642
		isert_unmap_data_buf(isert_conn, &wr->data);
1643 1644 1645 1646 1647 1648
	}

	wr->ib_sge = NULL;
	wr->send_wr = NULL;
}

1649
static void
1650
isert_put_cmd(struct isert_cmd *isert_cmd, bool comp_err)
1651
{
1652
	struct iscsi_cmd *cmd = isert_cmd->iscsi_cmd;
1653
	struct isert_conn *isert_conn = isert_cmd->conn;
1654
	struct iscsi_conn *conn = isert_conn->conn;
1655
	struct isert_device *device = isert_conn->conn_device;
1656 1657 1658 1659 1660 1661 1662

	pr_debug("Entering isert_put_cmd: %p\n", isert_cmd);

	switch (cmd->iscsi_opcode) {
	case ISCSI_OP_SCSI_CMD:
		spin_lock_bh(&conn->cmd_lock);
		if (!list_empty(&cmd->i_conn_node))
1663
			list_del_init(&cmd->i_conn_node);
1664 1665
		spin_unlock_bh(&conn->cmd_lock);

1666
		if (cmd->data_direction == DMA_TO_DEVICE) {
1667
			iscsit_stop_dataout_timer(cmd);
1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680
			/*
			 * Check for special case during comp_err where
			 * WRITE_PENDING has been handed off from core,
			 * but requires an extra target_put_sess_cmd()
			 * before transport_generic_free_cmd() below.
			 */
			if (comp_err &&
			    cmd->se_cmd.t_state == TRANSPORT_WRITE_PENDING) {
				struct se_cmd *se_cmd = &cmd->se_cmd;

				target_put_sess_cmd(se_cmd->se_sess, se_cmd);
			}
		}
1681

1682
		device->unreg_rdma_mem(isert_cmd, isert_conn);
1683 1684
		transport_generic_free_cmd(&cmd->se_cmd, 0);
		break;
1685
	case ISCSI_OP_SCSI_TMFUNC:
1686 1687
		spin_lock_bh(&conn->cmd_lock);
		if (!list_empty(&cmd->i_conn_node))
1688
			list_del_init(&cmd->i_conn_node);
1689 1690
		spin_unlock_bh(&conn->cmd_lock);

1691 1692 1693 1694
		transport_generic_free_cmd(&cmd->se_cmd, 0);
		break;
	case ISCSI_OP_REJECT:
	case ISCSI_OP_NOOP_OUT:
1695
	case ISCSI_OP_TEXT:
1696 1697
		spin_lock_bh(&conn->cmd_lock);
		if (!list_empty(&cmd->i_conn_node))
1698
			list_del_init(&cmd->i_conn_node);
1699 1700 1701 1702 1703 1704 1705 1706
		spin_unlock_bh(&conn->cmd_lock);

		/*
		 * Handle special case for REJECT when iscsi_add_reject*() has
		 * overwritten the original iscsi_opcode assignment, and the
		 * associated cmd->se_cmd needs to be released.
		 */
		if (cmd->se_cmd.se_tfo != NULL) {
1707 1708 1709
			pr_debug("Calling transport_generic_free_cmd from"
				 " isert_put_cmd for 0x%02x\n",
				 cmd->iscsi_opcode);
1710 1711 1712 1713 1714 1715 1716
			transport_generic_free_cmd(&cmd->se_cmd, 0);
			break;
		}
		/*
		 * Fall-through
		 */
	default:
1717
		iscsit_release_cmd(cmd);
1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734
		break;
	}
}

static void
isert_unmap_tx_desc(struct iser_tx_desc *tx_desc, struct ib_device *ib_dev)
{
	if (tx_desc->dma_addr != 0) {
		pr_debug("Calling ib_dma_unmap_single for tx_desc->dma_addr\n");
		ib_dma_unmap_single(ib_dev, tx_desc->dma_addr,
				    ISER_HEADERS_LEN, DMA_TO_DEVICE);
		tx_desc->dma_addr = 0;
	}
}

static void
isert_completion_put(struct iser_tx_desc *tx_desc, struct isert_cmd *isert_cmd,
1735
		     struct ib_device *ib_dev, bool comp_err)
1736
{
1737 1738 1739 1740 1741
	if (isert_cmd->pdu_buf_dma != 0) {
		pr_debug("Calling ib_dma_unmap_single for isert_cmd->pdu_buf_dma\n");
		ib_dma_unmap_single(ib_dev, isert_cmd->pdu_buf_dma,
				    isert_cmd->pdu_buf_len, DMA_TO_DEVICE);
		isert_cmd->pdu_buf_dma = 0;
1742 1743 1744
	}

	isert_unmap_tx_desc(tx_desc, ib_dev);
1745
	isert_put_cmd(isert_cmd, comp_err);
1746 1747
}

1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
static int
isert_check_pi_status(struct se_cmd *se_cmd, struct ib_mr *sig_mr)
{
	struct ib_mr_status mr_status;
	int ret;

	ret = ib_check_mr_status(sig_mr, IB_MR_CHECK_SIG_STATUS, &mr_status);
	if (ret) {
		pr_err("ib_check_mr_status failed, ret %d\n", ret);
		goto fail_mr_status;
	}

	if (mr_status.fail_status & IB_MR_CHECK_SIG_STATUS) {
		u64 sec_offset_err;
		u32 block_size = se_cmd->se_dev->dev_attrib.block_size + 8;

		switch (mr_status.sig_err.err_type) {
		case IB_SIG_BAD_GUARD:
			se_cmd->pi_err = TCM_LOGICAL_BLOCK_GUARD_CHECK_FAILED;
			break;
		case IB_SIG_BAD_REFTAG:
			se_cmd->pi_err = TCM_LOGICAL_BLOCK_REF_TAG_CHECK_FAILED;
			break;
		case IB_SIG_BAD_APPTAG:
			se_cmd->pi_err = TCM_LOGICAL_BLOCK_APP_TAG_CHECK_FAILED;
			break;
		}
		sec_offset_err = mr_status.sig_err.sig_err_offset;
		do_div(sec_offset_err, block_size);
		se_cmd->bad_sector = sec_offset_err + se_cmd->t_task_lba;

		pr_err("isert: PI error found type %d at sector 0x%llx "
		       "expected 0x%x vs actual 0x%x\n",
		       mr_status.sig_err.err_type,
		       (unsigned long long)se_cmd->bad_sector,
		       mr_status.sig_err.expected,
		       mr_status.sig_err.actual);
		ret = 1;
	}

fail_mr_status:
	return ret;
}

1792 1793 1794 1795
static void
isert_completion_rdma_write(struct iser_tx_desc *tx_desc,
			    struct isert_cmd *isert_cmd)
{
1796
	struct isert_rdma_wr *wr = &isert_cmd->rdma_wr;
1797
	struct iscsi_cmd *cmd = isert_cmd->iscsi_cmd;
1798
	struct se_cmd *se_cmd = &cmd->se_cmd;
1799 1800
	struct isert_conn *isert_conn = isert_cmd->conn;
	struct isert_device *device = isert_conn->conn_device;
1801 1802 1803
	int ret = 0;

	if (wr->fr_desc && wr->fr_desc->ind & ISERT_PROTECTED) {
1804 1805 1806
		ret = isert_check_pi_status(se_cmd,
					    wr->fr_desc->pi_ctx->sig_mr);
		wr->fr_desc->ind &= ~ISERT_PROTECTED;
1807
	}
1808 1809

	device->unreg_rdma_mem(isert_cmd, isert_conn);
1810
	wr->send_wr_num = 0;
1811 1812 1813 1814 1815
	if (ret)
		transport_send_check_condition_and_sense(se_cmd,
							 se_cmd->pi_err, 0);
	else
		isert_put_response(isert_conn->conn, cmd);
1816 1817
}

1818 1819 1820 1821 1822
static void
isert_completion_rdma_read(struct iser_tx_desc *tx_desc,
			   struct isert_cmd *isert_cmd)
{
	struct isert_rdma_wr *wr = &isert_cmd->rdma_wr;
1823
	struct iscsi_cmd *cmd = isert_cmd->iscsi_cmd;
1824
	struct se_cmd *se_cmd = &cmd->se_cmd;
1825
	struct isert_conn *isert_conn = isert_cmd->conn;
1826
	struct isert_device *device = isert_conn->conn_device;
1827
	int ret = 0;
1828

1829
	if (wr->fr_desc && wr->fr_desc->ind & ISERT_PROTECTED) {
1830 1831 1832
		ret = isert_check_pi_status(se_cmd,
					    wr->fr_desc->pi_ctx->sig_mr);
		wr->fr_desc->ind &= ~ISERT_PROTECTED;
1833 1834
	}

1835
	iscsit_stop_dataout_timer(cmd);
1836
	device->unreg_rdma_mem(isert_cmd, isert_conn);
1837
	cmd->write_data_done = wr->data.len;
1838
	wr->send_wr_num = 0;
1839

1840
	pr_debug("Cmd: %p RDMA_READ comp calling execute_cmd\n", isert_cmd);
1841 1842 1843 1844 1845
	spin_lock_bh(&cmd->istate_lock);
	cmd->cmd_flags |= ICF_GOT_LAST_DATAOUT;
	cmd->i_state = ISTATE_RECEIVED_LAST_DATAOUT;
	spin_unlock_bh(&cmd->istate_lock);

1846 1847 1848 1849 1850
	if (ret)
		transport_send_check_condition_and_sense(se_cmd,
							 se_cmd->pi_err, 0);
	else
		target_execute_cmd(se_cmd);
1851 1852 1853 1854 1855 1856 1857 1858 1859
}

static void
isert_do_control_comp(struct work_struct *work)
{
	struct isert_cmd *isert_cmd = container_of(work,
			struct isert_cmd, comp_work);
	struct isert_conn *isert_conn = isert_cmd->conn;
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
1860
	struct iscsi_cmd *cmd = isert_cmd->iscsi_cmd;
1861 1862 1863 1864 1865 1866 1867 1868

	switch (cmd->i_state) {
	case ISTATE_SEND_TASKMGTRSP:
		pr_debug("Calling iscsit_tmr_post_handler >>>>>>>>>>>>>>>>>\n");

		iscsit_tmr_post_handler(cmd, cmd->conn);

		cmd->i_state = ISTATE_SENT_STATUS;
1869
		isert_completion_put(&isert_cmd->tx_desc, isert_cmd, ib_dev, false);
1870 1871 1872 1873 1874
		break;
	case ISTATE_SEND_REJECT:
		pr_debug("Got isert_do_control_comp ISTATE_SEND_REJECT: >>>\n");

		cmd->i_state = ISTATE_SENT_STATUS;
1875
		isert_completion_put(&isert_cmd->tx_desc, isert_cmd, ib_dev, false);
1876
		break;
1877 1878
	case ISTATE_SEND_LOGOUTRSP:
		pr_debug("Calling iscsit_logout_post_handler >>>>>>>>>>>>>>\n");
1879

1880 1881
		iscsit_logout_post_handler(cmd, cmd->conn);
		break;
1882 1883
	case ISTATE_SEND_TEXTRSP:
		cmd->i_state = ISTATE_SENT_STATUS;
1884
		isert_completion_put(&isert_cmd->tx_desc, isert_cmd, ib_dev, false);
1885
		break;
1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898
	default:
		pr_err("Unknown do_control_comp i_state %d\n", cmd->i_state);
		dump_stack();
		break;
	}
}

static void
isert_response_completion(struct iser_tx_desc *tx_desc,
			  struct isert_cmd *isert_cmd,
			  struct isert_conn *isert_conn,
			  struct ib_device *ib_dev)
{
1899
	struct iscsi_cmd *cmd = isert_cmd->iscsi_cmd;
1900 1901

	if (cmd->i_state == ISTATE_SEND_TASKMGTRSP ||
1902
	    cmd->i_state == ISTATE_SEND_LOGOUTRSP ||
1903 1904
	    cmd->i_state == ISTATE_SEND_REJECT ||
	    cmd->i_state == ISTATE_SEND_TEXTRSP) {
1905 1906 1907 1908 1909 1910
		isert_unmap_tx_desc(tx_desc, ib_dev);

		INIT_WORK(&isert_cmd->comp_work, isert_do_control_comp);
		queue_work(isert_comp_wq, &isert_cmd->comp_work);
		return;
	}
1911

1912
	cmd->i_state = ISTATE_SENT_STATUS;
1913
	isert_completion_put(tx_desc, isert_cmd, ib_dev, false);
1914 1915 1916
}

static void
1917 1918
isert_send_completion(struct iser_tx_desc *tx_desc,
		      struct isert_conn *isert_conn)
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct isert_cmd *isert_cmd = tx_desc->isert_cmd;
	struct isert_rdma_wr *wr;

	if (!isert_cmd) {
		isert_unmap_tx_desc(tx_desc, ib_dev);
		return;
	}
	wr = &isert_cmd->rdma_wr;

	switch (wr->iser_ib_op) {
	case ISER_IB_RECV:
		pr_err("isert_send_completion: Got ISER_IB_RECV\n");
		dump_stack();
		break;
	case ISER_IB_SEND:
		pr_debug("isert_send_completion: Got ISER_IB_SEND\n");
		isert_response_completion(tx_desc, isert_cmd,
					  isert_conn, ib_dev);
		break;
	case ISER_IB_RDMA_WRITE:
1941 1942
		pr_debug("isert_send_completion: Got ISER_IB_RDMA_WRITE\n");
		isert_completion_rdma_write(tx_desc, isert_cmd);
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955
		break;
	case ISER_IB_RDMA_READ:
		pr_debug("isert_send_completion: Got ISER_IB_RDMA_READ:\n");

		isert_completion_rdma_read(tx_desc, isert_cmd);
		break;
	default:
		pr_err("Unknown wr->iser_ib_op: 0x%02x\n", wr->iser_ib_op);
		dump_stack();
		break;
	}
}

1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978
/**
 * is_isert_tx_desc() - Indicate if the completion wr_id
 *     is a TX descriptor or not.
 * @isert_conn: iser connection
 * @wr_id: completion WR identifier
 *
 * Since we cannot rely on wc opcode in FLUSH errors
 * we must work around it by checking if the wr_id address
 * falls in the iser connection rx_descs buffer. If so
 * it is an RX descriptor, otherwize it is a TX.
 */
static inline bool
is_isert_tx_desc(struct isert_conn *isert_conn, void *wr_id)
{
	void *start = isert_conn->conn_rx_descs;
	int len = ISERT_QP_MAX_RECV_DTOS * sizeof(*isert_conn->conn_rx_descs);

	if (wr_id >= start && wr_id < start + len)
		return false;

	return true;
}

1979
static void
1980
isert_cq_comp_err(struct isert_conn *isert_conn, struct ib_wc *wc)
1981
{
1982 1983 1984 1985 1986
	if (wc->wr_id == ISER_BEACON_WRID) {
		pr_info("conn %p completing conn_wait_comp_err\n",
			   isert_conn);
		complete(&isert_conn->conn_wait_comp_err);
	} else
1987
	if (is_isert_tx_desc(isert_conn, (void *)wc->wr_id)) {
1988 1989
		struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
		struct isert_cmd *isert_cmd;
1990
		struct iser_tx_desc *desc;
1991

1992 1993
		desc = (struct iser_tx_desc *)(uintptr_t)wc->wr_id;
		isert_cmd = desc->isert_cmd;
1994 1995 1996 1997 1998 1999
		if (!isert_cmd)
			isert_unmap_tx_desc(desc, ib_dev);
		else
			isert_completion_put(desc, isert_cmd, ib_dev, true);
	} else {
		isert_conn->post_recv_buf_count--;
2000 2001
		if (!isert_conn->post_recv_buf_count)
			iscsit_cause_connection_reinstatement(isert_conn->conn, 0);
2002
	}
2003 2004 2005
}

static void
2006
isert_handle_wc(struct ib_wc *wc)
2007 2008 2009
{
	struct isert_conn *isert_conn;
	struct iser_tx_desc *tx_desc;
2010
	struct iser_rx_desc *rx_desc;
2011

2012 2013 2014 2015 2016
	isert_conn = wc->qp->qp_context;
	if (likely(wc->status == IB_WC_SUCCESS)) {
		if (wc->opcode == IB_WC_RECV) {
			rx_desc = (struct iser_rx_desc *)(uintptr_t)wc->wr_id;
			isert_rx_completion(rx_desc, isert_conn, wc->byte_len);
2017
		} else {
2018 2019
			tx_desc = (struct iser_tx_desc *)(uintptr_t)wc->wr_id;
			isert_send_completion(tx_desc, isert_conn);
2020
		}
2021 2022 2023 2024 2025 2026
	} else {
		if (wc->status != IB_WC_WR_FLUSH_ERR)
			pr_err("wr id %llx status %d vend_err %x\n",
				  wc->wr_id, wc->status, wc->vendor_err);
		else
			pr_debug("flush error: wr id %llx\n", wc->wr_id);
2027

2028 2029 2030
		if (wc->wr_id != ISER_FASTREG_LI_WRID)
			isert_cq_comp_err(isert_conn, wc);
	}
2031 2032 2033
}

static void
2034
isert_cq_work(struct work_struct *work)
2035
{
2036
	enum { isert_poll_budget = 65536 };
2037
	struct isert_comp *comp = container_of(work, struct isert_comp,
2038
					       work);
2039 2040
	struct ib_wc *const wcs = comp->wcs;
	int i, n, completed = 0;
2041

2042 2043 2044
	while ((n = ib_poll_cq(comp->cq, ARRAY_SIZE(comp->wcs), wcs)) > 0) {
		for (i = 0; i < n; i++)
			isert_handle_wc(&wcs[i]);
2045

2046 2047
		completed += n;
		if (completed >= isert_poll_budget)
2048 2049 2050
			break;
	}

2051
	ib_req_notify_cq(comp->cq, IB_CQ_NEXT_COMP);
2052 2053 2054
}

static void
2055
isert_cq_callback(struct ib_cq *cq, void *context)
2056
{
2057
	struct isert_comp *comp = context;
2058

2059
	queue_work(isert_comp_wq, &comp->work);
2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079
}

static int
isert_post_response(struct isert_conn *isert_conn, struct isert_cmd *isert_cmd)
{
	struct ib_send_wr *wr_failed;
	int ret;

	ret = ib_post_send(isert_conn->conn_qp, &isert_cmd->tx_desc.send_wr,
			   &wr_failed);
	if (ret) {
		pr_err("ib_post_send failed with %d\n", ret);
		return ret;
	}
	return ret;
}

static int
isert_put_response(struct iscsi_conn *conn, struct iscsi_cmd *cmd)
{
2080
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct ib_send_wr *send_wr = &isert_cmd->tx_desc.send_wr;
	struct iscsi_scsi_rsp *hdr = (struct iscsi_scsi_rsp *)
				&isert_cmd->tx_desc.iscsi_header;

	isert_create_send_desc(isert_conn, isert_cmd, &isert_cmd->tx_desc);
	iscsit_build_rsp_pdu(cmd, conn, true, hdr);
	isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
	/*
	 * Attach SENSE DATA payload to iSCSI Response PDU
	 */
	if (cmd->se_cmd.sense_buffer &&
	    ((cmd->se_cmd.se_cmd_flags & SCF_TRANSPORT_TASK_SENSE) ||
	    (cmd->se_cmd.se_cmd_flags & SCF_EMULATED_TASK_SENSE))) {
		struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
		struct ib_sge *tx_dsg = &isert_cmd->tx_desc.tx_sg[1];
2097
		u32 padding, pdu_len;
2098 2099 2100 2101 2102 2103 2104

		put_unaligned_be16(cmd->se_cmd.scsi_sense_length,
				   cmd->sense_buffer);
		cmd->se_cmd.scsi_sense_length += sizeof(__be16);

		padding = -(cmd->se_cmd.scsi_sense_length) & 3;
		hton24(hdr->dlength, (u32)cmd->se_cmd.scsi_sense_length);
2105
		pdu_len = cmd->se_cmd.scsi_sense_length + padding;
2106

2107 2108
		isert_cmd->pdu_buf_dma = ib_dma_map_single(ib_dev,
				(void *)cmd->sense_buffer, pdu_len,
2109 2110
				DMA_TO_DEVICE);

2111 2112 2113
		isert_cmd->pdu_buf_len = pdu_len;
		tx_dsg->addr	= isert_cmd->pdu_buf_dma;
		tx_dsg->length	= pdu_len;
2114 2115 2116 2117
		tx_dsg->lkey	= isert_conn->conn_mr->lkey;
		isert_cmd->tx_desc.num_sge = 2;
	}

2118
	isert_init_send_wr(isert_conn, isert_cmd, send_wr);
2119 2120 2121 2122 2123 2124

	pr_debug("Posting SCSI Response IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");

	return isert_post_response(isert_conn, isert_cmd);
}

2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142
static void
isert_aborted_task(struct iscsi_conn *conn, struct iscsi_cmd *cmd)
{
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct isert_device *device = isert_conn->conn_device;

	spin_lock_bh(&conn->cmd_lock);
	if (!list_empty(&cmd->i_conn_node))
		list_del_init(&cmd->i_conn_node);
	spin_unlock_bh(&conn->cmd_lock);

	if (cmd->data_direction == DMA_TO_DEVICE)
		iscsit_stop_dataout_timer(cmd);

	device->unreg_rdma_mem(isert_cmd, isert_conn);
}

2143 2144 2145 2146 2147 2148
static enum target_prot_op
isert_get_sup_prot_ops(struct iscsi_conn *conn)
{
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct isert_device *device = isert_conn->conn_device;

2149 2150 2151 2152 2153 2154 2155 2156 2157 2158
	if (conn->tpg->tpg_attrib.t10_pi) {
		if (device->pi_capable) {
			pr_info("conn %p PI offload enabled\n", isert_conn);
			isert_conn->pi_support = true;
			return TARGET_PROT_ALL;
		}
	}

	pr_info("conn %p PI offload disabled\n", isert_conn);
	isert_conn->pi_support = false;
2159 2160 2161 2162

	return TARGET_PROT_NORMAL;
}

2163 2164 2165 2166
static int
isert_put_nopin(struct iscsi_cmd *cmd, struct iscsi_conn *conn,
		bool nopout_response)
{
2167
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2168 2169 2170 2171 2172 2173 2174 2175
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct ib_send_wr *send_wr = &isert_cmd->tx_desc.send_wr;

	isert_create_send_desc(isert_conn, isert_cmd, &isert_cmd->tx_desc);
	iscsit_build_nopin_rsp(cmd, conn, (struct iscsi_nopin *)
			       &isert_cmd->tx_desc.iscsi_header,
			       nopout_response);
	isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
2176
	isert_init_send_wr(isert_conn, isert_cmd, send_wr);
2177

M
Masanari Iida 已提交
2178
	pr_debug("Posting NOPIN Response IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");
2179 2180 2181 2182 2183 2184 2185

	return isert_post_response(isert_conn, isert_cmd);
}

static int
isert_put_logout_rsp(struct iscsi_cmd *cmd, struct iscsi_conn *conn)
{
2186
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2187 2188 2189 2190 2191 2192 2193
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct ib_send_wr *send_wr = &isert_cmd->tx_desc.send_wr;

	isert_create_send_desc(isert_conn, isert_cmd, &isert_cmd->tx_desc);
	iscsit_build_logout_rsp(cmd, conn, (struct iscsi_logout_rsp *)
				&isert_cmd->tx_desc.iscsi_header);
	isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
2194
	isert_init_send_wr(isert_conn, isert_cmd, send_wr);
2195 2196 2197 2198 2199 2200 2201 2202 2203

	pr_debug("Posting Logout Response IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");

	return isert_post_response(isert_conn, isert_cmd);
}

static int
isert_put_tm_rsp(struct iscsi_cmd *cmd, struct iscsi_conn *conn)
{
2204
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2205 2206 2207 2208 2209 2210 2211
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct ib_send_wr *send_wr = &isert_cmd->tx_desc.send_wr;

	isert_create_send_desc(isert_conn, isert_cmd, &isert_cmd->tx_desc);
	iscsit_build_task_mgt_rsp(cmd, conn, (struct iscsi_tm_rsp *)
				  &isert_cmd->tx_desc.iscsi_header);
	isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
2212
	isert_init_send_wr(isert_conn, isert_cmd, send_wr);
2213 2214 2215 2216 2217 2218 2219 2220 2221

	pr_debug("Posting Task Management Response IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");

	return isert_post_response(isert_conn, isert_cmd);
}

static int
isert_put_reject(struct iscsi_cmd *cmd, struct iscsi_conn *conn)
{
2222
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2223 2224
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct ib_send_wr *send_wr = &isert_cmd->tx_desc.send_wr;
2225 2226 2227 2228
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	struct ib_sge *tx_dsg = &isert_cmd->tx_desc.tx_sg[1];
	struct iscsi_reject *hdr =
		(struct iscsi_reject *)&isert_cmd->tx_desc.iscsi_header;
2229 2230

	isert_create_send_desc(isert_conn, isert_cmd, &isert_cmd->tx_desc);
2231
	iscsit_build_reject(cmd, conn, hdr);
2232
	isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
2233 2234

	hton24(hdr->dlength, ISCSI_HDR_LEN);
2235
	isert_cmd->pdu_buf_dma = ib_dma_map_single(ib_dev,
2236 2237
			(void *)cmd->buf_ptr, ISCSI_HDR_LEN,
			DMA_TO_DEVICE);
2238 2239
	isert_cmd->pdu_buf_len = ISCSI_HDR_LEN;
	tx_dsg->addr	= isert_cmd->pdu_buf_dma;
2240 2241 2242 2243
	tx_dsg->length	= ISCSI_HDR_LEN;
	tx_dsg->lkey	= isert_conn->conn_mr->lkey;
	isert_cmd->tx_desc.num_sge = 2;

2244
	isert_init_send_wr(isert_conn, isert_cmd, send_wr);
2245 2246 2247 2248 2249 2250

	pr_debug("Posting Reject IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");

	return isert_post_response(isert_conn, isert_cmd);
}

2251 2252 2253
static int
isert_put_text_rsp(struct iscsi_cmd *cmd, struct iscsi_conn *conn)
{
2254
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2255 2256 2257 2258 2259 2260 2261 2262
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	struct ib_send_wr *send_wr = &isert_cmd->tx_desc.send_wr;
	struct iscsi_text_rsp *hdr =
		(struct iscsi_text_rsp *)&isert_cmd->tx_desc.iscsi_header;
	u32 txt_rsp_len;
	int rc;

	isert_create_send_desc(isert_conn, isert_cmd, &isert_cmd->tx_desc);
2263
	rc = iscsit_build_text_rsp(cmd, conn, hdr, ISCSI_INFINIBAND);
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283
	if (rc < 0)
		return rc;

	txt_rsp_len = rc;
	isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);

	if (txt_rsp_len) {
		struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
		struct ib_sge *tx_dsg = &isert_cmd->tx_desc.tx_sg[1];
		void *txt_rsp_buf = cmd->buf_ptr;

		isert_cmd->pdu_buf_dma = ib_dma_map_single(ib_dev,
				txt_rsp_buf, txt_rsp_len, DMA_TO_DEVICE);

		isert_cmd->pdu_buf_len = txt_rsp_len;
		tx_dsg->addr	= isert_cmd->pdu_buf_dma;
		tx_dsg->length	= txt_rsp_len;
		tx_dsg->lkey	= isert_conn->conn_mr->lkey;
		isert_cmd->tx_desc.num_sge = 2;
	}
2284
	isert_init_send_wr(isert_conn, isert_cmd, send_wr);
2285 2286 2287 2288 2289 2290

	pr_debug("Posting Text Response IB_WR_SEND >>>>>>>>>>>>>>>>>>>>>>\n");

	return isert_post_response(isert_conn, isert_cmd);
}

2291 2292 2293 2294 2295
static int
isert_build_rdma_wr(struct isert_conn *isert_conn, struct isert_cmd *isert_cmd,
		    struct ib_sge *ib_sge, struct ib_send_wr *send_wr,
		    u32 data_left, u32 offset)
{
2296
	struct iscsi_cmd *cmd = isert_cmd->iscsi_cmd;
2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308
	struct scatterlist *sg_start, *tmp_sg;
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
	u32 sg_off, page_off;
	int i = 0, sg_nents;

	sg_off = offset / PAGE_SIZE;
	sg_start = &cmd->se_cmd.t_data_sg[sg_off];
	sg_nents = min(cmd->se_cmd.t_data_nents - sg_off, isert_conn->max_sge);
	page_off = offset % PAGE_SIZE;

	send_wr->sg_list = ib_sge;
	send_wr->num_sge = sg_nents;
2309
	send_wr->wr_id = (uintptr_t)&isert_cmd->tx_desc;
2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322
	/*
	 * Perform mapping of TCM scatterlist memory ib_sge dma_addr.
	 */
	for_each_sg(sg_start, tmp_sg, sg_nents, i) {
		pr_debug("ISER RDMA from SGL dma_addr: 0x%16llx dma_len: %u, page_off: %u\n",
			 (unsigned long long)tmp_sg->dma_address,
			 tmp_sg->length, page_off);

		ib_sge->addr = ib_sg_dma_address(ib_dev, tmp_sg) + page_off;
		ib_sge->length = min_t(u32, data_left,
				ib_sg_dma_len(ib_dev, tmp_sg) - page_off);
		ib_sge->lkey = isert_conn->conn_mr->lkey;

2323 2324
		pr_debug("RDMA ib_sge: addr: 0x%16llx  length: %u lkey: %08x\n",
			 ib_sge->addr, ib_sge->length, ib_sge->lkey);
2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337
		page_off = 0;
		data_left -= ib_sge->length;
		ib_sge++;
		pr_debug("Incrementing ib_sge pointer to %p\n", ib_sge);
	}

	pr_debug("Set outgoing sg_list: %p num_sg: %u from TCM SGLs\n",
		 send_wr->sg_list, send_wr->num_sge);

	return sg_nents;
}

static int
2338 2339
isert_map_rdma(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
	       struct isert_rdma_wr *wr)
2340 2341
{
	struct se_cmd *se_cmd = &cmd->se_cmd;
2342
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2343
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
2344
	struct isert_data_buf *data = &wr->data;
2345
	struct ib_send_wr *send_wr;
2346
	struct ib_sge *ib_sge;
2347 2348
	u32 offset, data_len, data_left, rdma_write_max, va_offset = 0;
	int ret = 0, i, ib_sge_cnt;
2349

2350
	isert_cmd->tx_desc.isert_cmd = isert_cmd;
2351

2352 2353 2354 2355 2356 2357
	offset = wr->iser_ib_op == ISER_IB_RDMA_READ ? cmd->write_data_done : 0;
	ret = isert_map_data_buf(isert_conn, isert_cmd, se_cmd->t_data_sg,
				 se_cmd->t_data_nents, se_cmd->data_length,
				 offset, wr->iser_ib_op, &wr->data);
	if (ret)
		return ret;
2358

2359 2360
	data_left = data->len;
	offset = data->offset;
2361

2362
	ib_sge = kzalloc(sizeof(struct ib_sge) * data->nents, GFP_KERNEL);
2363
	if (!ib_sge) {
2364
		pr_warn("Unable to allocate ib_sge\n");
2365
		ret = -ENOMEM;
2366
		goto unmap_cmd;
2367
	}
2368
	wr->ib_sge = ib_sge;
2369

2370
	wr->send_wr_num = DIV_ROUND_UP(data->nents, isert_conn->max_sge);
2371 2372 2373
	wr->send_wr = kzalloc(sizeof(struct ib_send_wr) * wr->send_wr_num,
				GFP_KERNEL);
	if (!wr->send_wr) {
2374
		pr_debug("Unable to allocate wr->send_wr\n");
2375
		ret = -ENOMEM;
2376
		goto unmap_cmd;
2377 2378 2379 2380 2381 2382 2383 2384 2385 2386
	}

	wr->isert_cmd = isert_cmd;
	rdma_write_max = isert_conn->max_sge * PAGE_SIZE;

	for (i = 0; i < wr->send_wr_num; i++) {
		send_wr = &isert_cmd->rdma_wr.send_wr[i];
		data_len = min(data_left, rdma_write_max);

		send_wr->send_flags = 0;
2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403
		if (wr->iser_ib_op == ISER_IB_RDMA_WRITE) {
			send_wr->opcode = IB_WR_RDMA_WRITE;
			send_wr->wr.rdma.remote_addr = isert_cmd->read_va + offset;
			send_wr->wr.rdma.rkey = isert_cmd->read_stag;
			if (i + 1 == wr->send_wr_num)
				send_wr->next = &isert_cmd->tx_desc.send_wr;
			else
				send_wr->next = &wr->send_wr[i + 1];
		} else {
			send_wr->opcode = IB_WR_RDMA_READ;
			send_wr->wr.rdma.remote_addr = isert_cmd->write_va + va_offset;
			send_wr->wr.rdma.rkey = isert_cmd->write_stag;
			if (i + 1 == wr->send_wr_num)
				send_wr->send_flags = IB_SEND_SIGNALED;
			else
				send_wr->next = &wr->send_wr[i + 1];
		}
2404 2405 2406 2407 2408 2409

		ib_sge_cnt = isert_build_rdma_wr(isert_conn, isert_cmd, ib_sge,
					send_wr, data_len, offset);
		ib_sge += ib_sge_cnt;

		offset += data_len;
2410
		va_offset += data_len;
2411 2412
		data_left -= data_len;
	}
2413 2414

	return 0;
2415 2416 2417
unmap_cmd:
	isert_unmap_data_buf(isert_conn, data);

2418 2419 2420
	return ret;
}

2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460
static int
isert_map_fr_pagelist(struct ib_device *ib_dev,
		      struct scatterlist *sg_start, int sg_nents, u64 *fr_pl)
{
	u64 start_addr, end_addr, page, chunk_start = 0;
	struct scatterlist *tmp_sg;
	int i = 0, new_chunk, last_ent, n_pages;

	n_pages = 0;
	new_chunk = 1;
	last_ent = sg_nents - 1;
	for_each_sg(sg_start, tmp_sg, sg_nents, i) {
		start_addr = ib_sg_dma_address(ib_dev, tmp_sg);
		if (new_chunk)
			chunk_start = start_addr;
		end_addr = start_addr + ib_sg_dma_len(ib_dev, tmp_sg);

		pr_debug("SGL[%d] dma_addr: 0x%16llx len: %u\n",
			 i, (unsigned long long)tmp_sg->dma_address,
			 tmp_sg->length);

		if ((end_addr & ~PAGE_MASK) && i < last_ent) {
			new_chunk = 0;
			continue;
		}
		new_chunk = 1;

		page = chunk_start & PAGE_MASK;
		do {
			fr_pl[n_pages++] = page;
			pr_debug("Mapped page_list[%d] page_addr: 0x%16llx\n",
				 n_pages - 1, page);
			page += PAGE_SIZE;
		} while (page < end_addr);
	}

	return n_pages;
}

static int
2461 2462 2463
isert_fast_reg_mr(struct isert_conn *isert_conn,
		  struct fast_reg_descriptor *fr_desc,
		  struct isert_data_buf *mem,
2464
		  enum isert_indicator ind,
2465
		  struct ib_sge *sge)
2466 2467
{
	struct ib_device *ib_dev = isert_conn->conn_cm_id->device;
2468 2469
	struct ib_mr *mr;
	struct ib_fast_reg_page_list *frpl;
2470 2471
	struct ib_send_wr fr_wr, inv_wr;
	struct ib_send_wr *bad_wr, *wr = NULL;
2472 2473
	int ret, pagelist_len;
	u32 page_off;
2474 2475
	u8 key;

2476 2477 2478 2479
	if (mem->dma_nents == 1) {
		sge->lkey = isert_conn->conn_mr->lkey;
		sge->addr = ib_sg_dma_address(ib_dev, &mem->sg[0]);
		sge->length = ib_sg_dma_len(ib_dev, &mem->sg[0]);
2480 2481 2482
		pr_debug("%s:%d sge: addr: 0x%llx  length: %u lkey: %x\n",
			 __func__, __LINE__, sge->addr, sge->length,
			 sge->lkey);
2483 2484 2485
		return 0;
	}

2486 2487 2488 2489 2490 2491 2492 2493 2494 2495
	if (ind == ISERT_DATA_KEY_VALID) {
		/* Registering data buffer */
		mr = fr_desc->data_mr;
		frpl = fr_desc->data_frpl;
	} else {
		/* Registering protection buffer */
		mr = fr_desc->pi_ctx->prot_mr;
		frpl = fr_desc->pi_ctx->prot_frpl;
	}

2496
	page_off = mem->offset % PAGE_SIZE;
2497

2498
	pr_debug("Use fr_desc %p sg_nents %d offset %u\n",
2499
		 fr_desc, mem->nents, mem->offset);
2500

2501
	pagelist_len = isert_map_fr_pagelist(ib_dev, mem->sg, mem->nents,
2502
					     &frpl->page_list[0]);
2503

2504
	if (!(fr_desc->ind & ISERT_DATA_KEY_VALID)) {
2505
		memset(&inv_wr, 0, sizeof(inv_wr));
2506
		inv_wr.wr_id = ISER_FASTREG_LI_WRID;
2507
		inv_wr.opcode = IB_WR_LOCAL_INV;
2508
		inv_wr.ex.invalidate_rkey = mr->rkey;
2509 2510
		wr = &inv_wr;
		/* Bump the key */
2511 2512
		key = (u8)(mr->rkey & 0x000000FF);
		ib_update_fast_reg_key(mr, ++key);
2513 2514 2515 2516
	}

	/* Prepare FASTREG WR */
	memset(&fr_wr, 0, sizeof(fr_wr));
2517
	fr_wr.wr_id = ISER_FASTREG_LI_WRID;
2518
	fr_wr.opcode = IB_WR_FAST_REG_MR;
2519 2520
	fr_wr.wr.fast_reg.iova_start = frpl->page_list[0] + page_off;
	fr_wr.wr.fast_reg.page_list = frpl;
2521 2522
	fr_wr.wr.fast_reg.page_list_len = pagelist_len;
	fr_wr.wr.fast_reg.page_shift = PAGE_SHIFT;
2523
	fr_wr.wr.fast_reg.length = mem->len;
2524
	fr_wr.wr.fast_reg.rkey = mr->rkey;
2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536
	fr_wr.wr.fast_reg.access_flags = IB_ACCESS_LOCAL_WRITE;

	if (!wr)
		wr = &fr_wr;
	else
		wr->next = &fr_wr;

	ret = ib_post_send(isert_conn->conn_qp, wr, &bad_wr);
	if (ret) {
		pr_err("fast registration failed, ret:%d\n", ret);
		return ret;
	}
2537
	fr_desc->ind &= ~ind;
2538

2539 2540
	sge->lkey = mr->lkey;
	sge->addr = frpl->page_list[0] + page_off;
2541
	sge->length = mem->len;
2542

2543 2544 2545 2546 2547 2548 2549
	pr_debug("%s:%d sge: addr: 0x%llx  length: %u lkey: %x\n",
		 __func__, __LINE__, sge->addr, sge->length,
		 sge->lkey);

	return ret;
}

2550 2551 2552 2553
static inline void
isert_set_dif_domain(struct se_cmd *se_cmd, struct ib_sig_attrs *sig_attrs,
		     struct ib_sig_domain *domain)
{
2554
	domain->sig_type = IB_SIG_TYPE_T10_DIF;
2555 2556 2557
	domain->sig.dif.bg_type = IB_T10DIF_CRC;
	domain->sig.dif.pi_interval = se_cmd->se_dev->dev_attrib.block_size;
	domain->sig.dif.ref_tag = se_cmd->reftag_seed;
2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568
	/*
	 * At the moment we hard code those, but if in the future
	 * the target core would like to use it, we will take it
	 * from se_cmd.
	 */
	domain->sig.dif.apptag_check_mask = 0xffff;
	domain->sig.dif.app_escape = true;
	domain->sig.dif.ref_escape = true;
	if (se_cmd->prot_type == TARGET_DIF_TYPE1_PROT ||
	    se_cmd->prot_type == TARGET_DIF_TYPE2_PROT)
		domain->sig.dif.ref_remap = true;
2569 2570
};

2571 2572 2573 2574 2575 2576
static int
isert_set_sig_attrs(struct se_cmd *se_cmd, struct ib_sig_attrs *sig_attrs)
{
	switch (se_cmd->prot_op) {
	case TARGET_PROT_DIN_INSERT:
	case TARGET_PROT_DOUT_STRIP:
2577
		sig_attrs->mem.sig_type = IB_SIG_TYPE_NONE;
2578
		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->wire);
2579 2580 2581
		break;
	case TARGET_PROT_DOUT_INSERT:
	case TARGET_PROT_DIN_STRIP:
2582
		sig_attrs->wire.sig_type = IB_SIG_TYPE_NONE;
2583
		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->mem);
2584 2585 2586
		break;
	case TARGET_PROT_DIN_PASS:
	case TARGET_PROT_DOUT_PASS:
2587 2588
		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->wire);
		isert_set_dif_domain(se_cmd, sig_attrs, &sig_attrs->mem);
2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606
		break;
	default:
		pr_err("Unsupported PI operation %d\n", se_cmd->prot_op);
		return -EINVAL;
	}

	return 0;
}

static inline u8
isert_set_prot_checks(u8 prot_checks)
{
	return (prot_checks & TARGET_DIF_CHECK_GUARD  ? 0xc0 : 0) |
	       (prot_checks & TARGET_DIF_CHECK_REFTAG ? 0x30 : 0) |
	       (prot_checks & TARGET_DIF_CHECK_REFTAG ? 0x0f : 0);
}

static int
2607 2608 2609 2610
isert_reg_sig_mr(struct isert_conn *isert_conn,
		 struct se_cmd *se_cmd,
		 struct isert_rdma_wr *rdma_wr,
		 struct fast_reg_descriptor *fr_desc)
2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622
{
	struct ib_send_wr sig_wr, inv_wr;
	struct ib_send_wr *bad_wr, *wr = NULL;
	struct pi_context *pi_ctx = fr_desc->pi_ctx;
	struct ib_sig_attrs sig_attrs;
	int ret;
	u32 key;

	memset(&sig_attrs, 0, sizeof(sig_attrs));
	ret = isert_set_sig_attrs(se_cmd, &sig_attrs);
	if (ret)
		goto err;
2623

2624 2625 2626 2627 2628
	sig_attrs.check_mask = isert_set_prot_checks(se_cmd->prot_checks);

	if (!(fr_desc->ind & ISERT_SIG_KEY_VALID)) {
		memset(&inv_wr, 0, sizeof(inv_wr));
		inv_wr.opcode = IB_WR_LOCAL_INV;
2629
		inv_wr.wr_id = ISER_FASTREG_LI_WRID;
2630 2631 2632 2633 2634 2635 2636 2637 2638
		inv_wr.ex.invalidate_rkey = pi_ctx->sig_mr->rkey;
		wr = &inv_wr;
		/* Bump the key */
		key = (u8)(pi_ctx->sig_mr->rkey & 0x000000FF);
		ib_update_fast_reg_key(pi_ctx->sig_mr, ++key);
	}

	memset(&sig_wr, 0, sizeof(sig_wr));
	sig_wr.opcode = IB_WR_REG_SIG_MR;
2639
	sig_wr.wr_id = ISER_FASTREG_LI_WRID;
2640
	sig_wr.sg_list = &rdma_wr->ib_sg[DATA];
2641 2642 2643 2644 2645
	sig_wr.num_sge = 1;
	sig_wr.wr.sig_handover.access_flags = IB_ACCESS_LOCAL_WRITE;
	sig_wr.wr.sig_handover.sig_attrs = &sig_attrs;
	sig_wr.wr.sig_handover.sig_mr = pi_ctx->sig_mr;
	if (se_cmd->t_prot_sg)
2646
		sig_wr.wr.sig_handover.prot = &rdma_wr->ib_sg[PROT];
2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659

	if (!wr)
		wr = &sig_wr;
	else
		wr->next = &sig_wr;

	ret = ib_post_send(isert_conn->conn_qp, wr, &bad_wr);
	if (ret) {
		pr_err("fast registration failed, ret:%d\n", ret);
		goto err;
	}
	fr_desc->ind &= ~ISERT_SIG_KEY_VALID;

2660 2661 2662
	rdma_wr->ib_sg[SIG].lkey = pi_ctx->sig_mr->lkey;
	rdma_wr->ib_sg[SIG].addr = 0;
	rdma_wr->ib_sg[SIG].length = se_cmd->data_length;
2663 2664 2665 2666 2667 2668
	if (se_cmd->prot_op != TARGET_PROT_DIN_STRIP &&
	    se_cmd->prot_op != TARGET_PROT_DOUT_INSERT)
		/*
		 * We have protection guards on the wire
		 * so we need to set a larget transfer
		 */
2669
		rdma_wr->ib_sg[SIG].length += se_cmd->prot_length;
2670 2671

	pr_debug("sig_sge: addr: 0x%llx  length: %u lkey: %x\n",
2672 2673
		  rdma_wr->ib_sg[SIG].addr, rdma_wr->ib_sg[SIG].length,
		  rdma_wr->ib_sg[SIG].lkey);
2674
err:
2675 2676 2677
	return ret;
}

2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736
static int
isert_handle_prot_cmd(struct isert_conn *isert_conn,
		      struct isert_cmd *isert_cmd,
		      struct isert_rdma_wr *wr)
{
	struct isert_device *device = isert_conn->conn_device;
	struct se_cmd *se_cmd = &isert_cmd->iscsi_cmd->se_cmd;
	int ret;

	if (!wr->fr_desc->pi_ctx) {
		ret = isert_create_pi_ctx(wr->fr_desc,
					  device->ib_device,
					  isert_conn->conn_pd);
		if (ret) {
			pr_err("conn %p failed to allocate pi_ctx\n",
				  isert_conn);
			return ret;
		}
	}

	if (se_cmd->t_prot_sg) {
		ret = isert_map_data_buf(isert_conn, isert_cmd,
					 se_cmd->t_prot_sg,
					 se_cmd->t_prot_nents,
					 se_cmd->prot_length,
					 0, wr->iser_ib_op, &wr->prot);
		if (ret) {
			pr_err("conn %p failed to map protection buffer\n",
				  isert_conn);
			return ret;
		}

		memset(&wr->ib_sg[PROT], 0, sizeof(wr->ib_sg[PROT]));
		ret = isert_fast_reg_mr(isert_conn, wr->fr_desc, &wr->prot,
					ISERT_PROT_KEY_VALID, &wr->ib_sg[PROT]);
		if (ret) {
			pr_err("conn %p failed to fast reg mr\n",
				  isert_conn);
			goto unmap_prot_cmd;
		}
	}

	ret = isert_reg_sig_mr(isert_conn, se_cmd, wr, wr->fr_desc);
	if (ret) {
		pr_err("conn %p failed to fast reg mr\n",
			  isert_conn);
		goto unmap_prot_cmd;
	}
	wr->fr_desc->ind |= ISERT_PROTECTED;

	return 0;

unmap_prot_cmd:
	if (se_cmd->t_prot_sg)
		isert_unmap_data_buf(isert_conn, &wr->prot);

	return ret;
}

2737
static int
2738 2739
isert_reg_rdma(struct iscsi_conn *conn, struct iscsi_cmd *cmd,
	       struct isert_rdma_wr *wr)
2740 2741 2742
{
	struct se_cmd *se_cmd = &cmd->se_cmd;
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2743 2744
	struct isert_conn *isert_conn = conn->context;
	struct fast_reg_descriptor *fr_desc = NULL;
2745 2746
	struct ib_send_wr *send_wr;
	struct ib_sge *ib_sg;
2747 2748
	u32 offset;
	int ret = 0;
2749 2750
	unsigned long flags;

2751
	isert_cmd->tx_desc.isert_cmd = isert_cmd;
2752

2753 2754 2755 2756 2757 2758
	offset = wr->iser_ib_op == ISER_IB_RDMA_READ ? cmd->write_data_done : 0;
	ret = isert_map_data_buf(isert_conn, isert_cmd, se_cmd->t_data_sg,
				 se_cmd->t_data_nents, se_cmd->data_length,
				 offset, wr->iser_ib_op, &wr->data);
	if (ret)
		return ret;
2759

2760
	if (wr->data.dma_nents != 1 || isert_prot_cmd(isert_conn, se_cmd)) {
2761 2762 2763 2764 2765 2766
		spin_lock_irqsave(&isert_conn->conn_lock, flags);
		fr_desc = list_first_entry(&isert_conn->conn_fr_pool,
					   struct fast_reg_descriptor, list);
		list_del(&fr_desc->list);
		spin_unlock_irqrestore(&isert_conn->conn_lock, flags);
		wr->fr_desc = fr_desc;
2767 2768
	}

2769
	ret = isert_fast_reg_mr(isert_conn, fr_desc, &wr->data,
2770
				ISERT_DATA_KEY_VALID, &wr->ib_sg[DATA]);
2771 2772
	if (ret)
		goto unmap_cmd;
2773

2774
	if (isert_prot_cmd(isert_conn, se_cmd)) {
2775
		ret = isert_handle_prot_cmd(isert_conn, isert_cmd, wr);
2776
		if (ret)
2777
			goto unmap_cmd;
2778

2779 2780 2781 2782
		ib_sg = &wr->ib_sg[SIG];
	} else {
		ib_sg = &wr->ib_sg[DATA];
	}
2783

2784
	memcpy(&wr->s_ib_sge, ib_sg, sizeof(*ib_sg));
2785
	wr->ib_sge = &wr->s_ib_sge;
2786 2787 2788 2789 2790 2791
	wr->send_wr_num = 1;
	memset(&wr->s_send_wr, 0, sizeof(*send_wr));
	wr->send_wr = &wr->s_send_wr;
	wr->isert_cmd = isert_cmd;

	send_wr = &isert_cmd->rdma_wr.s_send_wr;
2792
	send_wr->sg_list = &wr->s_ib_sge;
2793
	send_wr->num_sge = 1;
2794
	send_wr->wr_id = (uintptr_t)&isert_cmd->tx_desc;
2795 2796 2797 2798
	if (wr->iser_ib_op == ISER_IB_RDMA_WRITE) {
		send_wr->opcode = IB_WR_RDMA_WRITE;
		send_wr->wr.rdma.remote_addr = isert_cmd->read_va;
		send_wr->wr.rdma.rkey = isert_cmd->read_stag;
2799
		send_wr->send_flags = !isert_prot_cmd(isert_conn, se_cmd) ?
2800
				      0 : IB_SEND_SIGNALED;
2801 2802 2803 2804 2805 2806 2807
	} else {
		send_wr->opcode = IB_WR_RDMA_READ;
		send_wr->wr.rdma.remote_addr = isert_cmd->write_va;
		send_wr->wr.rdma.rkey = isert_cmd->write_stag;
		send_wr->send_flags = IB_SEND_SIGNALED;
	}

2808
	return 0;
2809

2810 2811
unmap_cmd:
	if (fr_desc) {
2812
		spin_lock_irqsave(&isert_conn->conn_lock, flags);
2813
		list_add_tail(&fr_desc->list, &isert_conn->conn_fr_pool);
2814
		spin_unlock_irqrestore(&isert_conn->conn_lock, flags);
2815
	}
2816
	isert_unmap_data_buf(isert_conn, &wr->data);
2817 2818 2819 2820

	return ret;
}

2821 2822 2823 2824
static int
isert_put_datain(struct iscsi_conn *conn, struct iscsi_cmd *cmd)
{
	struct se_cmd *se_cmd = &cmd->se_cmd;
2825
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2826 2827
	struct isert_rdma_wr *wr = &isert_cmd->rdma_wr;
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
2828
	struct isert_device *device = isert_conn->conn_device;
2829 2830 2831 2832 2833 2834
	struct ib_send_wr *wr_failed;
	int rc;

	pr_debug("Cmd: %p RDMA_WRITE data_length: %u\n",
		 isert_cmd, se_cmd->data_length);
	wr->iser_ib_op = ISER_IB_RDMA_WRITE;
2835
	rc = device->reg_rdma_mem(conn, cmd, wr);
2836 2837 2838 2839 2840
	if (rc) {
		pr_err("Cmd: %p failed to prepare RDMA res\n", isert_cmd);
		return rc;
	}

2841
	if (!isert_prot_cmd(isert_conn, se_cmd)) {
2842 2843 2844 2845 2846 2847 2848 2849 2850
		/*
		 * Build isert_conn->tx_desc for iSCSI response PDU and attach
		 */
		isert_create_send_desc(isert_conn, isert_cmd,
				       &isert_cmd->tx_desc);
		iscsit_build_rsp_pdu(cmd, conn, true, (struct iscsi_scsi_rsp *)
				     &isert_cmd->tx_desc.iscsi_header);
		isert_init_tx_hdrs(isert_conn, &isert_cmd->tx_desc);
		isert_init_send_wr(isert_conn, isert_cmd,
2851
				   &isert_cmd->tx_desc.send_wr);
2852
		isert_cmd->rdma_wr.s_send_wr.next = &isert_cmd->tx_desc.send_wr;
2853
		wr->send_wr_num += 1;
2854
	}
2855 2856

	rc = ib_post_send(isert_conn->conn_qp, wr->send_wr, &wr_failed);
2857
	if (rc)
2858
		pr_warn("ib_post_send() failed for IB_WR_RDMA_WRITE\n");
2859

2860
	if (!isert_prot_cmd(isert_conn, se_cmd))
2861 2862 2863 2864 2865
		pr_debug("Cmd: %p posted RDMA_WRITE + Response for iSER Data "
			 "READ\n", isert_cmd);
	else
		pr_debug("Cmd: %p posted RDMA_WRITE for iSER Data READ\n",
			 isert_cmd);
2866

2867
	return 1;
2868 2869 2870 2871 2872 2873
}

static int
isert_get_dataout(struct iscsi_conn *conn, struct iscsi_cmd *cmd, bool recovery)
{
	struct se_cmd *se_cmd = &cmd->se_cmd;
2874
	struct isert_cmd *isert_cmd = iscsit_priv_cmd(cmd);
2875 2876
	struct isert_rdma_wr *wr = &isert_cmd->rdma_wr;
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
2877
	struct isert_device *device = isert_conn->conn_device;
2878 2879
	struct ib_send_wr *wr_failed;
	int rc;
2880

2881 2882
	pr_debug("Cmd: %p RDMA_READ data_length: %u write_data_done: %u\n",
		 isert_cmd, se_cmd->data_length, cmd->write_data_done);
2883
	wr->iser_ib_op = ISER_IB_RDMA_READ;
2884
	rc = device->reg_rdma_mem(conn, cmd, wr);
2885 2886 2887
	if (rc) {
		pr_err("Cmd: %p failed to prepare RDMA res\n", isert_cmd);
		return rc;
2888 2889 2890
	}

	rc = ib_post_send(isert_conn->conn_qp, wr->send_wr, &wr_failed);
2891
	if (rc)
2892
		pr_warn("ib_post_send() failed for IB_WR_RDMA_READ\n");
2893

2894 2895
	pr_debug("Cmd: %p posted RDMA_READ memory for ISER Data WRITE\n",
		 isert_cmd);
2896

2897
	return 0;
2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920
}

static int
isert_immediate_queue(struct iscsi_conn *conn, struct iscsi_cmd *cmd, int state)
{
	int ret;

	switch (state) {
	case ISTATE_SEND_NOPIN_WANT_RESPONSE:
		ret = isert_put_nopin(cmd, conn, false);
		break;
	default:
		pr_err("Unknown immediate state: 0x%02x\n", state);
		ret = -EINVAL;
		break;
	}

	return ret;
}

static int
isert_response_queue(struct iscsi_conn *conn, struct iscsi_cmd *cmd, int state)
{
S
Sagi Grimberg 已提交
2921
	struct isert_conn *isert_conn = conn->context;
2922 2923 2924 2925 2926
	int ret;

	switch (state) {
	case ISTATE_SEND_LOGOUTRSP:
		ret = isert_put_logout_rsp(cmd, conn);
S
Sagi Grimberg 已提交
2927 2928
		if (!ret)
			isert_conn->logout_posted = true;
2929 2930 2931 2932 2933 2934 2935 2936 2937 2938
		break;
	case ISTATE_SEND_NOPIN:
		ret = isert_put_nopin(cmd, conn, true);
		break;
	case ISTATE_SEND_TASKMGTRSP:
		ret = isert_put_tm_rsp(cmd, conn);
		break;
	case ISTATE_SEND_REJECT:
		ret = isert_put_reject(cmd, conn);
		break;
2939 2940 2941
	case ISTATE_SEND_TEXTRSP:
		ret = isert_put_text_rsp(cmd, conn);
		break;
2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957
	case ISTATE_SEND_STATUS:
		/*
		 * Special case for sending non GOOD SCSI status from TX thread
		 * context during pre se_cmd excecution failure.
		 */
		ret = isert_put_response(conn, cmd);
		break;
	default:
		pr_err("Unknown response state: 0x%02x\n", state);
		ret = -EINVAL;
		break;
	}

	return ret;
}

2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996
struct rdma_cm_id *
isert_setup_id(struct isert_np *isert_np)
{
	struct iscsi_np *np = isert_np->np;
	struct rdma_cm_id *id;
	struct sockaddr *sa;
	int ret;

	sa = (struct sockaddr *)&np->np_sockaddr;
	pr_debug("ksockaddr: %p, sa: %p\n", &np->np_sockaddr, sa);

	id = rdma_create_id(isert_cma_handler, isert_np,
			    RDMA_PS_TCP, IB_QPT_RC);
	if (IS_ERR(id)) {
		pr_err("rdma_create_id() failed: %ld\n", PTR_ERR(id));
		ret = PTR_ERR(id);
		goto out;
	}
	pr_debug("id %p context %p\n", id, id->context);

	ret = rdma_bind_addr(id, sa);
	if (ret) {
		pr_err("rdma_bind_addr() failed: %d\n", ret);
		goto out_id;
	}

	ret = rdma_listen(id, ISERT_RDMA_LISTEN_BACKLOG);
	if (ret) {
		pr_err("rdma_listen() failed: %d\n", ret);
		goto out_id;
	}

	return id;
out_id:
	rdma_destroy_id(id);
out:
	return ERR_PTR(ret);
}

2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009
static int
isert_setup_np(struct iscsi_np *np,
	       struct __kernel_sockaddr_storage *ksockaddr)
{
	struct isert_np *isert_np;
	struct rdma_cm_id *isert_lid;
	int ret;

	isert_np = kzalloc(sizeof(struct isert_np), GFP_KERNEL);
	if (!isert_np) {
		pr_err("Unable to allocate struct isert_np\n");
		return -ENOMEM;
	}
3010
	sema_init(&isert_np->np_sem, 0);
3011 3012 3013
	mutex_init(&isert_np->np_accept_mutex);
	INIT_LIST_HEAD(&isert_np->np_accept_list);
	init_completion(&isert_np->np_login_comp);
3014
	isert_np->np = np;
3015 3016 3017 3018 3019 3020 3021 3022

	/*
	 * Setup the np->np_sockaddr from the passed sockaddr setup
	 * in iscsi_target_configfs.c code..
	 */
	memcpy(&np->np_sockaddr, ksockaddr,
	       sizeof(struct __kernel_sockaddr_storage));

3023
	isert_lid = isert_setup_id(isert_np);
3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035
	if (IS_ERR(isert_lid)) {
		ret = PTR_ERR(isert_lid);
		goto out;
	}

	isert_np->np_cm_id = isert_lid;
	np->np_context = isert_np;

	return 0;

out:
	kfree(isert_np);
3036

3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070
	return ret;
}

static int
isert_rdma_accept(struct isert_conn *isert_conn)
{
	struct rdma_cm_id *cm_id = isert_conn->conn_cm_id;
	struct rdma_conn_param cp;
	int ret;

	memset(&cp, 0, sizeof(struct rdma_conn_param));
	cp.initiator_depth = isert_conn->initiator_depth;
	cp.retry_count = 7;
	cp.rnr_retry_count = 7;

	pr_debug("Before rdma_accept >>>>>>>>>>>>>>>>>>>>.\n");

	ret = rdma_accept(cm_id, &cp);
	if (ret) {
		pr_err("rdma_accept() failed with: %d\n", ret);
		return ret;
	}

	pr_debug("After rdma_accept >>>>>>>>>>>>>>>>>>>>>.\n");

	return 0;
}

static int
isert_get_login_rx(struct iscsi_conn *conn, struct iscsi_login *login)
{
	struct isert_conn *isert_conn = (struct isert_conn *)conn->context;
	int ret;

3071 3072 3073 3074 3075 3076 3077 3078 3079
	pr_info("before login_req comp conn: %p\n", isert_conn);
	ret = wait_for_completion_interruptible(&isert_conn->login_req_comp);
	if (ret) {
		pr_err("isert_conn %p interrupted before got login req\n",
			  isert_conn);
		return ret;
	}
	reinit_completion(&isert_conn->login_req_comp);

3080 3081 3082 3083 3084 3085 3086 3087
	/*
	 * For login requests after the first PDU, isert_rx_login_req() will
	 * kick schedule_delayed_work(&conn->login_work) as the packet is
	 * received, which turns this callback from iscsi_target_do_login_rx()
	 * into a NOP.
	 */
	if (!login->first_request)
		return 0;
3088

3089 3090 3091
	isert_rx_login_req(isert_conn);

	pr_info("before conn_login_comp conn: %p\n", conn);
3092 3093 3094 3095
	ret = wait_for_completion_interruptible(&isert_conn->conn_login_comp);
	if (ret)
		return ret;

3096 3097
	pr_info("processing login->req: %p\n", login->req);

3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142
	return 0;
}

static void
isert_set_conn_info(struct iscsi_np *np, struct iscsi_conn *conn,
		    struct isert_conn *isert_conn)
{
	struct rdma_cm_id *cm_id = isert_conn->conn_cm_id;
	struct rdma_route *cm_route = &cm_id->route;
	struct sockaddr_in *sock_in;
	struct sockaddr_in6 *sock_in6;

	conn->login_family = np->np_sockaddr.ss_family;

	if (np->np_sockaddr.ss_family == AF_INET6) {
		sock_in6 = (struct sockaddr_in6 *)&cm_route->addr.dst_addr;
		snprintf(conn->login_ip, sizeof(conn->login_ip), "%pI6c",
			 &sock_in6->sin6_addr.in6_u);
		conn->login_port = ntohs(sock_in6->sin6_port);

		sock_in6 = (struct sockaddr_in6 *)&cm_route->addr.src_addr;
		snprintf(conn->local_ip, sizeof(conn->local_ip), "%pI6c",
			 &sock_in6->sin6_addr.in6_u);
		conn->local_port = ntohs(sock_in6->sin6_port);
	} else {
		sock_in = (struct sockaddr_in *)&cm_route->addr.dst_addr;
		sprintf(conn->login_ip, "%pI4",
			&sock_in->sin_addr.s_addr);
		conn->login_port = ntohs(sock_in->sin_port);

		sock_in = (struct sockaddr_in *)&cm_route->addr.src_addr;
		sprintf(conn->local_ip, "%pI4",
			&sock_in->sin_addr.s_addr);
		conn->local_port = ntohs(sock_in->sin_port);
	}
}

static int
isert_accept_np(struct iscsi_np *np, struct iscsi_conn *conn)
{
	struct isert_np *isert_np = (struct isert_np *)np->np_context;
	struct isert_conn *isert_conn;
	int max_accept = 0, ret;

accept_wait:
3143
	ret = down_interruptible(&isert_np->np_sem);
S
Sagi Grimberg 已提交
3144
	if (ret || max_accept > 5)
3145 3146 3147
		return -ENODEV;

	spin_lock_bh(&np->np_thread_lock);
3148
	if (np->np_thread_state >= ISCSI_NP_THREAD_RESET) {
3149
		spin_unlock_bh(&np->np_thread_lock);
3150 3151 3152 3153 3154 3155
		pr_debug("np_thread_state %d for isert_accept_np\n",
			 np->np_thread_state);
		/**
		 * No point in stalling here when np_thread
		 * is in state RESET/SHUTDOWN/EXIT - bail
		 **/
3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176
		return -ENODEV;
	}
	spin_unlock_bh(&np->np_thread_lock);

	mutex_lock(&isert_np->np_accept_mutex);
	if (list_empty(&isert_np->np_accept_list)) {
		mutex_unlock(&isert_np->np_accept_mutex);
		max_accept++;
		goto accept_wait;
	}
	isert_conn = list_first_entry(&isert_np->np_accept_list,
			struct isert_conn, conn_accept_node);
	list_del_init(&isert_conn->conn_accept_node);
	mutex_unlock(&isert_np->np_accept_mutex);

	conn->context = isert_conn;
	isert_conn->conn = conn;
	max_accept = 0;

	isert_set_conn_info(np, conn, isert_conn);

3177 3178
	pr_debug("Processing isert_conn: %p\n", isert_conn);

3179 3180 3181 3182 3183 3184 3185
	return 0;
}

static void
isert_free_np(struct iscsi_np *np)
{
	struct isert_np *isert_np = (struct isert_np *)np->np_context;
3186
	struct isert_conn *isert_conn, *n;
3187

3188 3189
	if (isert_np->np_cm_id)
		rdma_destroy_id(isert_np->np_cm_id);
3190

3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210
	/*
	 * FIXME: At this point we don't have a good way to insure
	 * that at this point we don't have hanging connections that
	 * completed RDMA establishment but didn't start iscsi login
	 * process. So work-around this by cleaning up what ever piled
	 * up in np_accept_list.
	 */
	mutex_lock(&isert_np->np_accept_mutex);
	if (!list_empty(&isert_np->np_accept_list)) {
		pr_info("Still have isert connections, cleaning up...\n");
		list_for_each_entry_safe(isert_conn, n,
					 &isert_np->np_accept_list,
					 conn_accept_node) {
			pr_info("cleaning isert_conn %p state (%d)\n",
				   isert_conn, isert_conn->state);
			isert_connect_release(isert_conn);
		}
	}
	mutex_unlock(&isert_np->np_accept_mutex);

3211 3212 3213 3214
	np->np_context = NULL;
	kfree(isert_np);
}

3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232
static void isert_release_work(struct work_struct *work)
{
	struct isert_conn *isert_conn = container_of(work,
						     struct isert_conn,
						     release_work);

	pr_info("Starting release conn %p\n", isert_conn);

	wait_for_completion(&isert_conn->conn_wait);

	mutex_lock(&isert_conn->conn_mutex);
	isert_conn->state = ISER_CONN_DOWN;
	mutex_unlock(&isert_conn->conn_mutex);

	pr_info("Destroying conn %p\n", isert_conn);
	isert_put_conn(isert_conn);
}

S
Sagi Grimberg 已提交
3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244
static void
isert_wait4logout(struct isert_conn *isert_conn)
{
	struct iscsi_conn *conn = isert_conn->conn;

	if (isert_conn->logout_posted) {
		pr_info("conn %p wait for conn_logout_comp\n", isert_conn);
		wait_for_completion_timeout(&conn->conn_logout_comp,
					    SECONDS_FOR_LOGOUT_COMP * HZ);
	}
}

3245 3246 3247 3248 3249 3250 3251 3252 3253
static void
isert_wait4cmds(struct iscsi_conn *conn)
{
	if (conn->sess) {
		target_sess_cmd_list_set_waiting(conn->sess->se_sess);
		target_wait_for_sess_cmds(conn->sess->se_sess);
	}
}

3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269
static void
isert_wait4flush(struct isert_conn *isert_conn)
{
	struct ib_recv_wr *bad_wr;

	init_completion(&isert_conn->conn_wait_comp_err);
	isert_conn->beacon.wr_id = ISER_BEACON_WRID;
	/* post an indication that all flush errors were consumed */
	if (ib_post_recv(isert_conn->conn_qp, &isert_conn->beacon, &bad_wr)) {
		pr_err("conn %p failed to post beacon", isert_conn);
		return;
	}

	wait_for_completion(&isert_conn->conn_wait_comp_err);
}

3270
static void isert_wait_conn(struct iscsi_conn *conn)
3271 3272 3273
{
	struct isert_conn *isert_conn = conn->context;

3274
	pr_debug("isert_wait_conn: Starting \n");
3275

3276
	mutex_lock(&isert_conn->conn_mutex);
3277 3278 3279 3280
	/*
	 * Only wait for conn_wait_comp_err if the isert_conn made it
	 * into full feature phase..
	 */
3281 3282 3283
	if (isert_conn->state == ISER_CONN_INIT) {
		mutex_unlock(&isert_conn->conn_mutex);
		return;
3284
	}
3285
	isert_conn_terminate(isert_conn);
3286
	mutex_unlock(&isert_conn->conn_mutex);
3287

3288
	isert_wait4cmds(conn);
3289
	isert_wait4flush(isert_conn);
S
Sagi Grimberg 已提交
3290
	isert_wait4logout(isert_conn);
3291

3292 3293
	INIT_WORK(&isert_conn->release_work, isert_release_work);
	queue_work(isert_release_wq, &isert_conn->release_work);
3294 3295 3296 3297 3298
}

static void isert_free_conn(struct iscsi_conn *conn)
{
	struct isert_conn *isert_conn = conn->context;
3299 3300 3301 3302 3303 3304 3305

	isert_put_conn(isert_conn);
}

static struct iscsit_transport iser_target_transport = {
	.name			= "IB/iSER",
	.transport_type		= ISCSI_INFINIBAND,
3306
	.priv_size		= sizeof(struct isert_cmd),
3307 3308 3309 3310
	.owner			= THIS_MODULE,
	.iscsit_setup_np	= isert_setup_np,
	.iscsit_accept_np	= isert_accept_np,
	.iscsit_free_np		= isert_free_np,
3311
	.iscsit_wait_conn	= isert_wait_conn,
3312 3313 3314 3315 3316 3317 3318 3319
	.iscsit_free_conn	= isert_free_conn,
	.iscsit_get_login_rx	= isert_get_login_rx,
	.iscsit_put_login_tx	= isert_put_login_tx,
	.iscsit_immediate_queue	= isert_immediate_queue,
	.iscsit_response_queue	= isert_response_queue,
	.iscsit_get_dataout	= isert_get_dataout,
	.iscsit_queue_data_in	= isert_put_datain,
	.iscsit_queue_status	= isert_put_response,
3320
	.iscsit_aborted_task	= isert_aborted_task,
3321
	.iscsit_get_sup_prot_ops = isert_get_sup_prot_ops,
3322 3323 3324 3325 3326 3327 3328 3329 3330 3331
};

static int __init isert_init(void)
{
	int ret;

	isert_comp_wq = alloc_workqueue("isert_comp_wq", 0, 0);
	if (!isert_comp_wq) {
		pr_err("Unable to allocate isert_comp_wq\n");
		ret = -ENOMEM;
3332
		return -ENOMEM;
3333 3334
	}

3335 3336 3337 3338 3339 3340 3341 3342
	isert_release_wq = alloc_workqueue("isert_release_wq", WQ_UNBOUND,
					WQ_UNBOUND_MAX_ACTIVE);
	if (!isert_release_wq) {
		pr_err("Unable to allocate isert_release_wq\n");
		ret = -ENOMEM;
		goto destroy_comp_wq;
	}

3343
	iscsit_register_transport(&iser_target_transport);
3344 3345
	pr_info("iSER_TARGET[0] - Loaded iser_target_transport\n");

3346 3347
	return 0;

3348 3349
destroy_comp_wq:
	destroy_workqueue(isert_comp_wq);
3350

3351 3352 3353 3354 3355
	return ret;
}

static void __exit isert_exit(void)
{
3356
	flush_scheduled_work();
3357
	destroy_workqueue(isert_release_wq);
3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369
	destroy_workqueue(isert_comp_wq);
	iscsit_unregister_transport(&iser_target_transport);
	pr_debug("iSER_TARGET[0] - Released iser_target_transport\n");
}

MODULE_DESCRIPTION("iSER-Target for mainline target infrastructure");
MODULE_VERSION("0.1");
MODULE_AUTHOR("nab@Linux-iSCSI.org");
MODULE_LICENSE("GPL");

module_init(isert_init);
module_exit(isert_exit);