ib_srp.c 50.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41
/*
 * Copyright (c) 2005 Cisco Systems.  All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 * $Id: ib_srp.c 3932 2005-11-01 17:19:29Z roland $
 */

#include <linux/module.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/err.h>
#include <linux/string.h>
#include <linux/parser.h>
#include <linux/random.h>
42
#include <linux/jiffies.h>
43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64

#include <asm/atomic.h>

#include <scsi/scsi.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_dbg.h>
#include <scsi/srp.h>

#include <rdma/ib_cache.h>

#include "ib_srp.h"

#define DRV_NAME	"ib_srp"
#define PFX		DRV_NAME ": "
#define DRV_VERSION	"0.2"
#define DRV_RELDATE	"November 1, 2005"

MODULE_AUTHOR("Roland Dreier");
MODULE_DESCRIPTION("InfiniBand SCSI RDMA Protocol initiator "
		   "v" DRV_VERSION " (" DRV_RELDATE ")");
MODULE_LICENSE("Dual BSD/GPL");

65 66 67 68 69 70 71
static int srp_sg_tablesize = SRP_DEF_SG_TABLESIZE;
static int srp_max_iu_len;

module_param(srp_sg_tablesize, int, 0444);
MODULE_PARM_DESC(srp_sg_tablesize,
		 "Max number of gather/scatter entries per I/O (default is 12)");

72 73 74 75 76 77 78 79
static int topspin_workarounds = 1;

module_param(topspin_workarounds, int, 0444);
MODULE_PARM_DESC(topspin_workarounds,
		 "Enable workarounds for Topspin/Cisco SRP target bugs if != 0");

static const u8 topspin_oui[3] = { 0x00, 0x05, 0xad };

80 81 82 83 84 85 86 87
static int mellanox_workarounds = 1;

module_param(mellanox_workarounds, int, 0444);
MODULE_PARM_DESC(mellanox_workarounds,
		 "Enable workarounds for Mellanox SRP target bugs if != 0");

static const u8 mellanox_oui[3] = { 0x00, 0x02, 0xc9 };

88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
static void srp_add_one(struct ib_device *device);
static void srp_remove_one(struct ib_device *device);
static void srp_completion(struct ib_cq *cq, void *target_ptr);
static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event);

static struct ib_client srp_client = {
	.name   = "srp",
	.add    = srp_add_one,
	.remove = srp_remove_one
};

static inline struct srp_target_port *host_to_target(struct Scsi_Host *host)
{
	return (struct srp_target_port *) host->hostdata;
}

static const char *srp_target_info(struct Scsi_Host *host)
{
	return host_to_target(host)->target_name;
}

static struct srp_iu *srp_alloc_iu(struct srp_host *host, size_t size,
				   gfp_t gfp_mask,
				   enum dma_data_direction direction)
{
	struct srp_iu *iu;

	iu = kmalloc(sizeof *iu, gfp_mask);
	if (!iu)
		goto out;

	iu->buf = kzalloc(size, gfp_mask);
	if (!iu->buf)
		goto out_free_iu;

123 124
	iu->dma = dma_map_single(host->dev->dev->dma_device,
				 iu->buf, size, direction);
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145
	if (dma_mapping_error(iu->dma))
		goto out_free_buf;

	iu->size      = size;
	iu->direction = direction;

	return iu;

out_free_buf:
	kfree(iu->buf);
out_free_iu:
	kfree(iu);
out:
	return NULL;
}

static void srp_free_iu(struct srp_host *host, struct srp_iu *iu)
{
	if (!iu)
		return;

146 147
	dma_unmap_single(host->dev->dev->dma_device,
			 iu->dma, iu->size, iu->direction);
148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166
	kfree(iu->buf);
	kfree(iu);
}

static void srp_qp_event(struct ib_event *event, void *context)
{
	printk(KERN_ERR PFX "QP event %d\n", event->event);
}

static int srp_init_qp(struct srp_target_port *target,
		       struct ib_qp *qp)
{
	struct ib_qp_attr *attr;
	int ret;

	attr = kmalloc(sizeof *attr, GFP_KERNEL);
	if (!attr)
		return -ENOMEM;

167
	ret = ib_find_cached_pkey(target->srp_host->dev->dev,
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
				  target->srp_host->port,
				  be16_to_cpu(target->path.pkey),
				  &attr->pkey_index);
	if (ret)
		goto out;

	attr->qp_state        = IB_QPS_INIT;
	attr->qp_access_flags = (IB_ACCESS_REMOTE_READ |
				    IB_ACCESS_REMOTE_WRITE);
	attr->port_num        = target->srp_host->port;

	ret = ib_modify_qp(qp, attr,
			   IB_QP_STATE		|
			   IB_QP_PKEY_INDEX	|
			   IB_QP_ACCESS_FLAGS	|
			   IB_QP_PORT);

out:
	kfree(attr);
	return ret;
}

static int srp_create_target_ib(struct srp_target_port *target)
{
	struct ib_qp_init_attr *init_attr;
	int ret;

	init_attr = kzalloc(sizeof *init_attr, GFP_KERNEL);
	if (!init_attr)
		return -ENOMEM;

199
	target->cq = ib_create_cq(target->srp_host->dev->dev, srp_completion,
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
				  NULL, target, SRP_CQ_SIZE);
	if (IS_ERR(target->cq)) {
		ret = PTR_ERR(target->cq);
		goto out;
	}

	ib_req_notify_cq(target->cq, IB_CQ_NEXT_COMP);

	init_attr->event_handler       = srp_qp_event;
	init_attr->cap.max_send_wr     = SRP_SQ_SIZE;
	init_attr->cap.max_recv_wr     = SRP_RQ_SIZE;
	init_attr->cap.max_recv_sge    = 1;
	init_attr->cap.max_send_sge    = 1;
	init_attr->sq_sig_type         = IB_SIGNAL_ALL_WR;
	init_attr->qp_type             = IB_QPT_RC;
	init_attr->send_cq             = target->cq;
	init_attr->recv_cq             = target->cq;

218
	target->qp = ib_create_qp(target->srp_host->dev->pd, init_attr);
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
	if (IS_ERR(target->qp)) {
		ret = PTR_ERR(target->qp);
		ib_destroy_cq(target->cq);
		goto out;
	}

	ret = srp_init_qp(target, target->qp);
	if (ret) {
		ib_destroy_qp(target->qp);
		ib_destroy_cq(target->cq);
		goto out;
	}

out:
	kfree(init_attr);
	return ret;
}

static void srp_free_target_ib(struct srp_target_port *target)
{
	int i;

	ib_destroy_qp(target->qp);
	ib_destroy_cq(target->cq);

	for (i = 0; i < SRP_RQ_SIZE; ++i)
		srp_free_iu(target->srp_host, target->rx_ring[i]);
	for (i = 0; i < SRP_SQ_SIZE + 1; ++i)
		srp_free_iu(target->srp_host, target->tx_ring[i]);
}

static void srp_path_rec_completion(int status,
				    struct ib_sa_path_rec *pathrec,
				    void *target_ptr)
{
	struct srp_target_port *target = target_ptr;

	target->status = status;
	if (status)
		printk(KERN_ERR PFX "Got failed path rec status %d\n", status);
	else
		target->path = *pathrec;
	complete(&target->done);
}

static int srp_lookup_path(struct srp_target_port *target)
{
	target->path.numb_path = 1;

	init_completion(&target->done);

270
	target->path_query_id = ib_sa_path_rec_get(target->srp_host->dev->dev,
271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
						   target->srp_host->port,
						   &target->path,
						   IB_SA_PATH_REC_DGID		|
						   IB_SA_PATH_REC_SGID		|
						   IB_SA_PATH_REC_NUMB_PATH	|
						   IB_SA_PATH_REC_PKEY,
						   SRP_PATH_REC_TIMEOUT_MS,
						   GFP_KERNEL,
						   srp_path_rec_completion,
						   target, &target->path_query);
	if (target->path_query_id < 0)
		return target->path_query_id;

	wait_for_completion(&target->done);

	if (target->status < 0)
		printk(KERN_WARNING PFX "Path record query failed\n");

	return target->status;
}

static int srp_send_req(struct srp_target_port *target)
{
	struct {
		struct ib_cm_req_param param;
		struct srp_login_req   priv;
	} *req = NULL;
	int status;

	req = kzalloc(sizeof *req, GFP_KERNEL);
	if (!req)
		return -ENOMEM;

	req->param.primary_path 	      = &target->path;
	req->param.alternate_path 	      = NULL;
	req->param.service_id 		      = target->service_id;
	req->param.qp_num 		      = target->qp->qp_num;
	req->param.qp_type 		      = target->qp->qp_type;
	req->param.private_data 	      = &req->priv;
	req->param.private_data_len 	      = sizeof req->priv;
	req->param.flow_control 	      = 1;

	get_random_bytes(&req->param.starting_psn, 4);
	req->param.starting_psn 	     &= 0xffffff;

	/*
	 * Pick some arbitrary defaults here; we could make these
	 * module parameters if anyone cared about setting them.
	 */
	req->param.responder_resources	      = 4;
	req->param.remote_cm_response_timeout = 20;
	req->param.local_cm_response_timeout  = 20;
	req->param.retry_count 		      = 7;
	req->param.rnr_retry_count 	      = 7;
	req->param.max_cm_retries 	      = 15;

	req->priv.opcode     	= SRP_LOGIN_REQ;
	req->priv.tag        	= 0;
329
	req->priv.req_it_iu_len = cpu_to_be32(srp_max_iu_len);
330 331
	req->priv.req_buf_fmt 	= cpu_to_be16(SRP_BUF_FORMAT_DIRECT |
					      SRP_BUF_FORMAT_INDIRECT);
332
	/*
R
Roland Dreier 已提交
333
	 * In the published SRP specification (draft rev. 16a), the
334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
	 * port identifier format is 8 bytes of ID extension followed
	 * by 8 bytes of GUID.  Older drafts put the two halves in the
	 * opposite order, so that the GUID comes first.
	 *
	 * Targets conforming to these obsolete drafts can be
	 * recognized by the I/O Class they report.
	 */
	if (target->io_class == SRP_REV10_IB_IO_CLASS) {
		memcpy(req->priv.initiator_port_id,
		       target->srp_host->initiator_port_id + 8, 8);
		memcpy(req->priv.initiator_port_id + 8,
		       target->srp_host->initiator_port_id, 8);
		memcpy(req->priv.target_port_id,     &target->ioc_guid, 8);
		memcpy(req->priv.target_port_id + 8, &target->id_ext, 8);
	} else {
		memcpy(req->priv.initiator_port_id,
		       target->srp_host->initiator_port_id, 16);
		memcpy(req->priv.target_port_id,     &target->id_ext, 8);
		memcpy(req->priv.target_port_id + 8, &target->ioc_guid, 8);
	}

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
	/*
	 * Topspin/Cisco SRP targets will reject our login unless we
	 * zero out the first 8 bytes of our initiator port ID.  The
	 * second 8 bytes must be our local node GUID, but we always
	 * use that anyway.
	 */
	if (topspin_workarounds && !memcmp(&target->ioc_guid, topspin_oui, 3)) {
		printk(KERN_DEBUG PFX "Topspin/Cisco initiator port ID workaround "
		       "activated for target GUID %016llx\n",
		       (unsigned long long) be64_to_cpu(target->ioc_guid));
		memset(req->priv.initiator_port_id, 0, 8);
	}

	status = ib_send_cm_req(target->cm_id, &req->param);

	kfree(req);

	return status;
}

static void srp_disconnect_target(struct srp_target_port *target)
{
	/* XXX should send SRP_I_LOGOUT request */

	init_completion(&target->done);
380 381 382 383
	if (ib_send_cm_dreq(target->cm_id, NULL, 0)) {
		printk(KERN_DEBUG PFX "Sending CM DREQ failed\n");
		return;
	}
384 385 386 387 388 389 390 391 392 393 394 395 396 397 398
	wait_for_completion(&target->done);
}

static void srp_remove_work(void *target_ptr)
{
	struct srp_target_port *target = target_ptr;

	spin_lock_irq(target->scsi_host->host_lock);
	if (target->state != SRP_TARGET_DEAD) {
		spin_unlock_irq(target->scsi_host->host_lock);
		return;
	}
	target->state = SRP_TARGET_REMOVED;
	spin_unlock_irq(target->scsi_host->host_lock);

399
	spin_lock(&target->srp_host->target_lock);
400
	list_del(&target->list);
401
	spin_unlock(&target->srp_host->target_lock);
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448

	scsi_remove_host(target->scsi_host);
	ib_destroy_cm_id(target->cm_id);
	srp_free_target_ib(target);
	scsi_host_put(target->scsi_host);
}

static int srp_connect_target(struct srp_target_port *target)
{
	int ret;

	ret = srp_lookup_path(target);
	if (ret)
		return ret;

	while (1) {
		init_completion(&target->done);
		ret = srp_send_req(target);
		if (ret)
			return ret;
		wait_for_completion(&target->done);

		/*
		 * The CM event handling code will set status to
		 * SRP_PORT_REDIRECT if we get a port redirect REJ
		 * back, or SRP_DLID_REDIRECT if we get a lid/qp
		 * redirect REJ back.
		 */
		switch (target->status) {
		case 0:
			return 0;

		case SRP_PORT_REDIRECT:
			ret = srp_lookup_path(target);
			if (ret)
				return ret;
			break;

		case SRP_DLID_REDIRECT:
			break;

		default:
			return target->status;
		}
	}
}

449 450 451 452 453 454 455 456 457 458 459 460
static void srp_unmap_data(struct scsi_cmnd *scmnd,
			   struct srp_target_port *target,
			   struct srp_request *req)
{
	struct scatterlist *scat;
	int nents;

	if (!scmnd->request_buffer ||
	    (scmnd->sc_data_direction != DMA_TO_DEVICE &&
	     scmnd->sc_data_direction != DMA_FROM_DEVICE))
		return;

461 462 463 464 465
	if (req->fmr) {
		ib_fmr_pool_unmap(req->fmr);
		req->fmr = NULL;
	}

466 467 468 469 470 471 472 473 474 475 476 477
	/*
	 * This handling of non-SG commands can be killed when the
	 * SCSI midlayer no longer generates non-SG commands.
	 */
	if (likely(scmnd->use_sg)) {
		nents = scmnd->use_sg;
		scat  = scmnd->request_buffer;
	} else {
		nents = 1;
		scat  = &req->fake_sg;
	}

478
	dma_unmap_sg(target->srp_host->dev->dev->dma_device, scat, nents,
479 480 481
		     scmnd->sc_data_direction);
}

482 483 484 485 486 487 488 489 490 491 492 493 494
static void srp_remove_req(struct srp_target_port *target, struct srp_request *req)
{
	srp_unmap_data(req->scmnd, target, req);
	list_move_tail(&req->list, &target->free_reqs);
}

static void srp_reset_req(struct srp_target_port *target, struct srp_request *req)
{
	req->scmnd->result = DID_RESET << 16;
	req->scmnd->scsi_done(req->scmnd);
	srp_remove_req(target, req);
}

495 496 497 498
static int srp_reconnect_target(struct srp_target_port *target)
{
	struct ib_cm_id *new_cm_id;
	struct ib_qp_attr qp_attr;
499
	struct srp_request *req, *tmp;
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
	struct ib_wc wc;
	int ret;

	spin_lock_irq(target->scsi_host->host_lock);
	if (target->state != SRP_TARGET_LIVE) {
		spin_unlock_irq(target->scsi_host->host_lock);
		return -EAGAIN;
	}
	target->state = SRP_TARGET_CONNECTING;
	spin_unlock_irq(target->scsi_host->host_lock);

	srp_disconnect_target(target);
	/*
	 * Now get a new local CM ID so that we avoid confusing the
	 * target in case things are really fouled up.
	 */
516
	new_cm_id = ib_create_cm_id(target->srp_host->dev->dev,
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
				    srp_cm_handler, target);
	if (IS_ERR(new_cm_id)) {
		ret = PTR_ERR(new_cm_id);
		goto err;
	}
	ib_destroy_cm_id(target->cm_id);
	target->cm_id = new_cm_id;

	qp_attr.qp_state = IB_QPS_RESET;
	ret = ib_modify_qp(target->qp, &qp_attr, IB_QP_STATE);
	if (ret)
		goto err;

	ret = srp_init_qp(target, target->qp);
	if (ret)
		goto err;

	while (ib_poll_cq(target->cq, 1, &wc) > 0)
		; /* nothing */

537
	spin_lock_irq(target->scsi_host->host_lock);
538 539
	list_for_each_entry_safe(req, tmp, &target->req_queue, list)
		srp_reset_req(target, req);
540
	spin_unlock_irq(target->scsi_host->host_lock);
541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579

	target->rx_head	 = 0;
	target->tx_head	 = 0;
	target->tx_tail  = 0;

	ret = srp_connect_target(target);
	if (ret)
		goto err;

	spin_lock_irq(target->scsi_host->host_lock);
	if (target->state == SRP_TARGET_CONNECTING) {
		ret = 0;
		target->state = SRP_TARGET_LIVE;
	} else
		ret = -EAGAIN;
	spin_unlock_irq(target->scsi_host->host_lock);

	return ret;

err:
	printk(KERN_ERR PFX "reconnect failed (%d), removing target port.\n", ret);

	/*
	 * We couldn't reconnect, so kill our target port off.
	 * However, we have to defer the real removal because we might
	 * be in the context of the SCSI error handler now, which
	 * would deadlock if we call scsi_remove_host().
	 */
	spin_lock_irq(target->scsi_host->host_lock);
	if (target->state == SRP_TARGET_CONNECTING) {
		target->state = SRP_TARGET_DEAD;
		INIT_WORK(&target->work, srp_remove_work, target);
		schedule_work(&target->work);
	}
	spin_unlock_irq(target->scsi_host->host_lock);

	return ret;
}

580
static int srp_map_fmr(struct srp_target_port *target, struct scatterlist *scat,
581 582 583 584 585 586 587 588 589
		       int sg_cnt, struct srp_request *req,
		       struct srp_direct_buf *buf)
{
	u64 io_addr = 0;
	u64 *dma_pages;
	u32 len;
	int page_cnt;
	int i, j;
	int ret;
590
	struct srp_device *dev = target->srp_host->dev;
591 592 593 594

	if (!dev->fmr_pool)
		return -ENODEV;

595 596 597 598
	if ((sg_dma_address(&scat[0]) & ~dev->fmr_page_mask) &&
	    mellanox_workarounds && !memcmp(&target->ioc_guid, mellanox_oui, 3))
		return -EINVAL;

599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632
	len = page_cnt = 0;
	for (i = 0; i < sg_cnt; ++i) {
		if (sg_dma_address(&scat[i]) & ~dev->fmr_page_mask) {
			if (i > 0)
				return -EINVAL;
			else
				++page_cnt;
		}
		if ((sg_dma_address(&scat[i]) + sg_dma_len(&scat[i])) &
		    ~dev->fmr_page_mask) {
			if (i < sg_cnt - 1)
				return -EINVAL;
			else
				++page_cnt;
		}

		len += sg_dma_len(&scat[i]);
	}

	page_cnt += len >> dev->fmr_page_shift;
	if (page_cnt > SRP_FMR_SIZE)
		return -ENOMEM;

	dma_pages = kmalloc(sizeof (u64) * page_cnt, GFP_ATOMIC);
	if (!dma_pages)
		return -ENOMEM;

	page_cnt = 0;
	for (i = 0; i < sg_cnt; ++i)
		for (j = 0; j < sg_dma_len(&scat[i]); j += dev->fmr_page_size)
			dma_pages[page_cnt++] =
				(sg_dma_address(&scat[i]) & dev->fmr_page_mask) + j;

	req->fmr = ib_fmr_pool_map_phys(dev->fmr_pool,
633
					dma_pages, page_cnt, io_addr);
634 635
	if (IS_ERR(req->fmr)) {
		ret = PTR_ERR(req->fmr);
V
Vu Pham 已提交
636
		req->fmr = NULL;
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651
		goto out;
	}

	buf->va  = cpu_to_be64(sg_dma_address(&scat[0]) & ~dev->fmr_page_mask);
	buf->key = cpu_to_be32(req->fmr->fmr->rkey);
	buf->len = cpu_to_be32(len);

	ret = 0;

out:
	kfree(dma_pages);

	return ret;
}

652 653 654
static int srp_map_data(struct scsi_cmnd *scmnd, struct srp_target_port *target,
			struct srp_request *req)
{
655
	struct scatterlist *scat;
656
	struct srp_cmd *cmd = req->cmd->buf;
657
	int len, nents, count;
658
	u8 fmt = SRP_DATA_DESC_DIRECT;
659 660 661 662 663 664 665 666 667 668 669

	if (!scmnd->request_buffer || scmnd->sc_data_direction == DMA_NONE)
		return sizeof (struct srp_cmd);

	if (scmnd->sc_data_direction != DMA_FROM_DEVICE &&
	    scmnd->sc_data_direction != DMA_TO_DEVICE) {
		printk(KERN_WARNING PFX "Unhandled data direction %d\n",
		       scmnd->sc_data_direction);
		return -EINVAL;
	}

670 671 672 673 674 675 676 677 678 679 680 681
	/*
	 * This handling of non-SG commands can be killed when the
	 * SCSI midlayer no longer generates non-SG commands.
	 */
	if (likely(scmnd->use_sg)) {
		nents = scmnd->use_sg;
		scat  = scmnd->request_buffer;
	} else {
		nents = 1;
		scat  = &req->fake_sg;
		sg_init_one(scat, scmnd->request_buffer, scmnd->request_bufflen);
	}
682

683 684 685 686 687
	count = dma_map_sg(target->srp_host->dev->dev->dma_device,
			   scat, nents, scmnd->sc_data_direction);

	fmt = SRP_DATA_DESC_DIRECT;
	len = sizeof (struct srp_cmd) +	sizeof (struct srp_direct_buf);
688

689
	if (count == 1) {
690 691 692 693 694 695
		/*
		 * The midlayer only generated a single gather/scatter
		 * entry, or DMA mapping coalesced everything to a
		 * single entry.  So a direct descriptor along with
		 * the DMA MR suffices.
		 */
696
		struct srp_direct_buf *buf = (void *) cmd->add_data;
697

698
		buf->va  = cpu_to_be64(sg_dma_address(scat));
699
		buf->key = cpu_to_be32(target->srp_host->dev->mr->rkey);
700
		buf->len = cpu_to_be32(sg_dma_len(scat));
701
	} else if (srp_map_fmr(target, scat, count, req,
702 703 704 705 706 707
			       (void *) cmd->add_data)) {
		/*
		 * FMR mapping failed, and the scatterlist has more
		 * than one entry.  Generate an indirect memory
		 * descriptor.
		 */
708 709
		struct srp_indirect_buf *buf = (void *) cmd->add_data;
		u32 datalen = 0;
710
		int i;
711

712
		fmt = SRP_DATA_DESC_INDIRECT;
713 714 715 716 717 718 719 720 721 722 723 724 725
		len = sizeof (struct srp_cmd) +
			sizeof (struct srp_indirect_buf) +
			count * sizeof (struct srp_direct_buf);

		for (i = 0; i < count; ++i) {
			buf->desc_list[i].va  =
				cpu_to_be64(sg_dma_address(&scat[i]));
			buf->desc_list[i].key =
				cpu_to_be32(target->srp_host->dev->mr->rkey);
			buf->desc_list[i].len =
				cpu_to_be32(sg_dma_len(&scat[i]));
			datalen += sg_dma_len(&scat[i]);
		}
726

727 728 729 730 731
		if (scmnd->sc_data_direction == DMA_TO_DEVICE)
			cmd->data_out_desc_cnt = count;
		else
			cmd->data_in_desc_cnt = count;

732 733
		buf->table_desc.va  =
			cpu_to_be64(req->cmd->dma + sizeof *cmd + sizeof *buf);
734
		buf->table_desc.key =
735
			cpu_to_be32(target->srp_host->dev->mr->rkey);
736 737 738 739
		buf->table_desc.len =
			cpu_to_be32(count * sizeof (struct srp_direct_buf));

		buf->len = cpu_to_be32(datalen);
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
	}

	if (scmnd->sc_data_direction == DMA_TO_DEVICE)
		cmd->buf_fmt = fmt << 4;
	else
		cmd->buf_fmt = fmt;

	return len;
}

static void srp_process_rsp(struct srp_target_port *target, struct srp_rsp *rsp)
{
	struct srp_request *req;
	struct scsi_cmnd *scmnd;
	unsigned long flags;
	s32 delta;

	delta = (s32) be32_to_cpu(rsp->req_lim_delta);

	spin_lock_irqsave(target->scsi_host->host_lock, flags);

	target->req_lim += delta;

	req = &target->req_ring[rsp->tag & ~SRP_TAG_TSK_MGMT];

	if (unlikely(rsp->tag & SRP_TAG_TSK_MGMT)) {
		if (be32_to_cpu(rsp->resp_data_len) < 4)
			req->tsk_status = -1;
		else
			req->tsk_status = rsp->data[3];
		complete(&req->done);
	} else {
772
		scmnd = req->scmnd;
773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793
		if (!scmnd)
			printk(KERN_ERR "Null scmnd for RSP w/tag %016llx\n",
			       (unsigned long long) rsp->tag);
		scmnd->result = rsp->status;

		if (rsp->flags & SRP_RSP_FLAG_SNSVALID) {
			memcpy(scmnd->sense_buffer, rsp->data +
			       be32_to_cpu(rsp->resp_data_len),
			       min_t(int, be32_to_cpu(rsp->sense_data_len),
				     SCSI_SENSE_BUFFERSIZE));
		}

		if (rsp->flags & (SRP_RSP_FLAG_DOOVER | SRP_RSP_FLAG_DOUNDER))
			scmnd->resid = be32_to_cpu(rsp->data_out_res_cnt);
		else if (rsp->flags & (SRP_RSP_FLAG_DIOVER | SRP_RSP_FLAG_DIUNDER))
			scmnd->resid = be32_to_cpu(rsp->data_in_res_cnt);

		if (!req->tsk_mgmt) {
			scmnd->host_scribble = (void *) -1L;
			scmnd->scsi_done(scmnd);

794
			srp_remove_req(target, req);
795 796 797 798 799 800 801 802 803 804 805 806 807 808
		} else
			req->cmd_done = 1;
	}

	spin_unlock_irqrestore(target->scsi_host->host_lock, flags);
}

static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc)
{
	struct srp_iu *iu;
	u8 opcode;

	iu = target->rx_ring[wc->wr_id & ~SRP_OP_RECV];

809
	dma_sync_single_for_cpu(target->srp_host->dev->dev->dma_device, iu->dma,
810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845
				target->max_ti_iu_len, DMA_FROM_DEVICE);

	opcode = *(u8 *) iu->buf;

	if (0) {
		int i;

		printk(KERN_ERR PFX "recv completion, opcode 0x%02x\n", opcode);

		for (i = 0; i < wc->byte_len; ++i) {
			if (i % 8 == 0)
				printk(KERN_ERR "  [%02x] ", i);
			printk(" %02x", ((u8 *) iu->buf)[i]);
			if ((i + 1) % 8 == 0)
				printk("\n");
		}

		if (wc->byte_len % 8)
			printk("\n");
	}

	switch (opcode) {
	case SRP_RSP:
		srp_process_rsp(target, iu->buf);
		break;

	case SRP_T_LOGOUT:
		/* XXX Handle target logout */
		printk(KERN_WARNING PFX "Got target logout request\n");
		break;

	default:
		printk(KERN_WARNING PFX "Unhandled SRP opcode 0x%02x\n", opcode);
		break;
	}

846
	dma_sync_single_for_device(target->srp_host->dev->dev->dma_device, iu->dma,
847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
				   target->max_ti_iu_len, DMA_FROM_DEVICE);
}

static void srp_completion(struct ib_cq *cq, void *target_ptr)
{
	struct srp_target_port *target = target_ptr;
	struct ib_wc wc;

	ib_req_notify_cq(cq, IB_CQ_NEXT_COMP);
	while (ib_poll_cq(cq, 1, &wc) > 0) {
		if (wc.status) {
			printk(KERN_ERR PFX "failed %s status %d\n",
			       wc.wr_id & SRP_OP_RECV ? "receive" : "send",
			       wc.status);
			break;
		}

		if (wc.wr_id & SRP_OP_RECV)
			srp_handle_recv(target, &wc);
		else
			++target->tx_tail;
	}
}

static int __srp_post_recv(struct srp_target_port *target)
{
	struct srp_iu *iu;
	struct ib_sge list;
	struct ib_recv_wr wr, *bad_wr;
	unsigned int next;
	int ret;

	next 	 = target->rx_head & (SRP_RQ_SIZE - 1);
	wr.wr_id = next | SRP_OP_RECV;
	iu 	 = target->rx_ring[next];

	list.addr   = iu->dma;
	list.length = iu->size;
885
	list.lkey   = target->srp_host->dev->mr->lkey;
886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911

	wr.next     = NULL;
	wr.sg_list  = &list;
	wr.num_sge  = 1;

	ret = ib_post_recv(target->qp, &wr, &bad_wr);
	if (!ret)
		++target->rx_head;

	return ret;
}

static int srp_post_recv(struct srp_target_port *target)
{
	unsigned long flags;
	int ret;

	spin_lock_irqsave(target->scsi_host->host_lock, flags);
	ret = __srp_post_recv(target);
	spin_unlock_irqrestore(target->scsi_host->host_lock, flags);

	return ret;
}

/*
 * Must be called with target->scsi_host->host_lock held to protect
912 913
 * req_lim and tx_head.  Lock cannot be dropped between call here and
 * call to __srp_post_send().
914 915 916 917 918 919
 */
static struct srp_iu *__srp_get_tx_iu(struct srp_target_port *target)
{
	if (target->tx_head - target->tx_tail >= SRP_SQ_SIZE)
		return NULL;

920 921
	if (unlikely(target->req_lim < 1))
		++target->zero_req_lim;
922

923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938
	return target->tx_ring[target->tx_head & SRP_SQ_SIZE];
}

/*
 * Must be called with target->scsi_host->host_lock held to protect
 * req_lim and tx_head.
 */
static int __srp_post_send(struct srp_target_port *target,
			   struct srp_iu *iu, int len)
{
	struct ib_sge list;
	struct ib_send_wr wr, *bad_wr;
	int ret = 0;

	list.addr   = iu->dma;
	list.length = len;
939
	list.lkey   = target->srp_host->dev->mr->lkey;
940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980

	wr.next       = NULL;
	wr.wr_id      = target->tx_head & SRP_SQ_SIZE;
	wr.sg_list    = &list;
	wr.num_sge    = 1;
	wr.opcode     = IB_WR_SEND;
	wr.send_flags = IB_SEND_SIGNALED;

	ret = ib_post_send(target->qp, &wr, &bad_wr);

	if (!ret) {
		++target->tx_head;
		--target->req_lim;
	}

	return ret;
}

static int srp_queuecommand(struct scsi_cmnd *scmnd,
			    void (*done)(struct scsi_cmnd *))
{
	struct srp_target_port *target = host_to_target(scmnd->device->host);
	struct srp_request *req;
	struct srp_iu *iu;
	struct srp_cmd *cmd;
	int len;

	if (target->state == SRP_TARGET_CONNECTING)
		goto err;

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED) {
		scmnd->result = DID_BAD_TARGET << 16;
		done(scmnd);
		return 0;
	}

	iu = __srp_get_tx_iu(target);
	if (!iu)
		goto err;

981
	dma_sync_single_for_cpu(target->srp_host->dev->dev->dma_device, iu->dma,
982
				srp_max_iu_len, DMA_TO_DEVICE);
983

984
	req = list_entry(target->free_reqs.next, struct srp_request, list);
985 986 987

	scmnd->scsi_done     = done;
	scmnd->result        = 0;
988
	scmnd->host_scribble = (void *) (long) req->index;
989 990 991 992 993 994

	cmd = iu->buf;
	memset(cmd, 0, sizeof *cmd);

	cmd->opcode = SRP_CMD;
	cmd->lun    = cpu_to_be64((u64) scmnd->device->lun << 48);
995
	cmd->tag    = req->index;
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
	memcpy(cmd->cdb, scmnd->cmnd, scmnd->cmd_len);

	req->scmnd    = scmnd;
	req->cmd      = iu;
	req->cmd_done = 0;
	req->tsk_mgmt = NULL;

	len = srp_map_data(scmnd, target, req);
	if (len < 0) {
		printk(KERN_ERR PFX "Failed to map data\n");
		goto err;
	}

	if (__srp_post_recv(target)) {
		printk(KERN_ERR PFX "Recv failed\n");
		goto err_unmap;
	}

1014
	dma_sync_single_for_device(target->srp_host->dev->dev->dma_device, iu->dma,
1015
				   srp_max_iu_len, DMA_TO_DEVICE);
1016 1017 1018 1019 1020 1021

	if (__srp_post_send(target, iu, len)) {
		printk(KERN_ERR PFX "Send failed\n");
		goto err_unmap;
	}

1022
	list_move_tail(&req->list, &target->req_queue);
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046

	return 0;

err_unmap:
	srp_unmap_data(scmnd, target, req);

err:
	return SCSI_MLQUEUE_HOST_BUSY;
}

static int srp_alloc_iu_bufs(struct srp_target_port *target)
{
	int i;

	for (i = 0; i < SRP_RQ_SIZE; ++i) {
		target->rx_ring[i] = srp_alloc_iu(target->srp_host,
						  target->max_ti_iu_len,
						  GFP_KERNEL, DMA_FROM_DEVICE);
		if (!target->rx_ring[i])
			goto err;
	}

	for (i = 0; i < SRP_SQ_SIZE + 1; ++i) {
		target->tx_ring[i] = srp_alloc_iu(target->srp_host,
1047
						  srp_max_iu_len,
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
						  GFP_KERNEL, DMA_TO_DEVICE);
		if (!target->tx_ring[i])
			goto err;
	}

	return 0;

err:
	for (i = 0; i < SRP_RQ_SIZE; ++i) {
		srp_free_iu(target->srp_host, target->rx_ring[i]);
		target->rx_ring[i] = NULL;
	}

	for (i = 0; i < SRP_SQ_SIZE + 1; ++i) {
		srp_free_iu(target->srp_host, target->tx_ring[i]);
		target->tx_ring[i] = NULL;
	}

	return -ENOMEM;
}

static void srp_cm_rej_handler(struct ib_cm_id *cm_id,
			       struct ib_cm_event *event,
			       struct srp_target_port *target)
{
	struct ib_class_port_info *cpi;
	int opcode;

	switch (event->param.rej_rcvd.reason) {
	case IB_CM_REJ_PORT_CM_REDIRECT:
		cpi = event->param.rej_rcvd.ari;
		target->path.dlid = cpi->redirect_lid;
		target->path.pkey = cpi->redirect_pkey;
		cm_id->remote_cm_qpn = be32_to_cpu(cpi->redirect_qp) & 0x00ffffff;
		memcpy(target->path.dgid.raw, cpi->redirect_gid, 16);

		target->status = target->path.dlid ?
			SRP_DLID_REDIRECT : SRP_PORT_REDIRECT;
		break;

	case IB_CM_REJ_PORT_REDIRECT:
		if (topspin_workarounds &&
		    !memcmp(&target->ioc_guid, topspin_oui, 3)) {
			/*
			 * Topspin/Cisco SRP gateways incorrectly send
			 * reject reason code 25 when they mean 24
			 * (port redirect).
			 */
			memcpy(target->path.dgid.raw,
			       event->param.rej_rcvd.ari, 16);

			printk(KERN_DEBUG PFX "Topspin/Cisco redirect to target port GID %016llx%016llx\n",
			       (unsigned long long) be64_to_cpu(target->path.dgid.global.subnet_prefix),
			       (unsigned long long) be64_to_cpu(target->path.dgid.global.interface_id));

			target->status = SRP_PORT_REDIRECT;
		} else {
			printk(KERN_WARNING "  REJ reason: IB_CM_REJ_PORT_REDIRECT\n");
			target->status = -ECONNRESET;
		}
		break;

	case IB_CM_REJ_DUPLICATE_LOCAL_COMM_ID:
		printk(KERN_WARNING "  REJ reason: IB_CM_REJ_DUPLICATE_LOCAL_COMM_ID\n");
		target->status = -ECONNRESET;
		break;

	case IB_CM_REJ_CONSUMER_DEFINED:
		opcode = *(u8 *) event->private_data;
		if (opcode == SRP_LOGIN_REJ) {
			struct srp_login_rej *rej = event->private_data;
			u32 reason = be32_to_cpu(rej->reason);

			if (reason == SRP_LOGIN_REJ_REQ_IT_IU_LENGTH_TOO_LARGE)
				printk(KERN_WARNING PFX
				       "SRP_LOGIN_REJ: requested max_it_iu_len too large\n");
			else
				printk(KERN_WARNING PFX
				       "SRP LOGIN REJECTED, reason 0x%08x\n", reason);
		} else
			printk(KERN_WARNING "  REJ reason: IB_CM_REJ_CONSUMER_DEFINED,"
			       " opcode 0x%02x\n", opcode);
		target->status = -ECONNRESET;
		break;

	default:
		printk(KERN_WARNING "  REJ reason 0x%x\n",
		       event->param.rej_rcvd.reason);
		target->status = -ECONNRESET;
	}
}

static int srp_cm_handler(struct ib_cm_id *cm_id, struct ib_cm_event *event)
{
	struct srp_target_port *target = cm_id->context;
	struct ib_qp_attr *qp_attr = NULL;
	int attr_mask = 0;
	int comp = 0;
	int opcode = 0;

	switch (event->event) {
	case IB_CM_REQ_ERROR:
		printk(KERN_DEBUG PFX "Sending CM REQ failed\n");
		comp = 1;
		target->status = -ECONNRESET;
		break;

	case IB_CM_REP_RECEIVED:
		comp = 1;
		opcode = *(u8 *) event->private_data;

		if (opcode == SRP_LOGIN_RSP) {
			struct srp_login_rsp *rsp = event->private_data;

			target->max_ti_iu_len = be32_to_cpu(rsp->max_ti_iu_len);
			target->req_lim       = be32_to_cpu(rsp->req_lim_delta);

			target->scsi_host->can_queue = min(target->req_lim,
							   target->scsi_host->can_queue);
		} else {
			printk(KERN_WARNING PFX "Unhandled RSP opcode %#x\n", opcode);
			target->status = -ECONNRESET;
			break;
		}

		target->status = srp_alloc_iu_bufs(target);
		if (target->status)
			break;

		qp_attr = kmalloc(sizeof *qp_attr, GFP_KERNEL);
		if (!qp_attr) {
			target->status = -ENOMEM;
			break;
		}

		qp_attr->qp_state = IB_QPS_RTR;
		target->status = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask);
		if (target->status)
			break;

		target->status = ib_modify_qp(target->qp, qp_attr, attr_mask);
		if (target->status)
			break;

		target->status = srp_post_recv(target);
		if (target->status)
			break;

		qp_attr->qp_state = IB_QPS_RTS;
		target->status = ib_cm_init_qp_attr(cm_id, qp_attr, &attr_mask);
		if (target->status)
			break;

		target->status = ib_modify_qp(target->qp, qp_attr, attr_mask);
		if (target->status)
			break;

		target->status = ib_send_cm_rtu(cm_id, NULL, 0);
		if (target->status)
			break;

		break;

	case IB_CM_REJ_RECEIVED:
		printk(KERN_DEBUG PFX "REJ received\n");
		comp = 1;

		srp_cm_rej_handler(cm_id, event, target);
		break;

1218 1219 1220 1221
	case IB_CM_DREQ_RECEIVED:
		printk(KERN_WARNING PFX "DREQ received - connection closed\n");
		if (ib_send_cm_drep(cm_id, NULL, 0))
			printk(KERN_ERR PFX "Sending CM DREP failed\n");
1222 1223 1224 1225 1226 1227 1228 1229 1230
		break;

	case IB_CM_TIMEWAIT_EXIT:
		printk(KERN_ERR PFX "connection closed\n");

		comp = 1;
		target->status = 0;
		break;

1231 1232 1233 1234 1235
	case IB_CM_MRA_RECEIVED:
	case IB_CM_DREQ_ERROR:
	case IB_CM_DREP_RECEIVED:
		break;

1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
	default:
		printk(KERN_WARNING PFX "Unhandled CM event %d\n", event->event);
		break;
	}

	if (comp)
		complete(&target->done);

	kfree(qp_attr);

	return 0;
}

1249 1250
static int srp_send_tsk_mgmt(struct srp_target_port *target,
			     struct srp_request *req, u8 func)
1251 1252 1253 1254 1255 1256
{
	struct srp_iu *iu;
	struct srp_tsk_mgmt *tsk_mgmt;

	spin_lock_irq(target->scsi_host->host_lock);

1257 1258
	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED) {
1259
		req->scmnd->result = DID_BAD_TARGET << 16;
1260 1261 1262
		goto out;
	}

1263 1264 1265 1266 1267 1268 1269 1270 1271 1272
	init_completion(&req->done);

	iu = __srp_get_tx_iu(target);
	if (!iu)
		goto out;

	tsk_mgmt = iu->buf;
	memset(tsk_mgmt, 0, sizeof *tsk_mgmt);

	tsk_mgmt->opcode 	= SRP_TSK_MGMT;
1273 1274
	tsk_mgmt->lun 		= cpu_to_be64((u64) req->scmnd->device->lun << 48);
	tsk_mgmt->tag 		= req->index | SRP_TAG_TSK_MGMT;
1275
	tsk_mgmt->tsk_mgmt_func = func;
1276
	tsk_mgmt->task_tag 	= req->index;
1277 1278 1279 1280 1281 1282 1283

	if (__srp_post_send(target, iu, sizeof *tsk_mgmt))
		goto out;

	req->tsk_mgmt = iu;

	spin_unlock_irq(target->scsi_host->host_lock);
1284

1285 1286
	if (!wait_for_completion_timeout(&req->done,
					 msecs_to_jiffies(SRP_ABORT_TIMEOUT_MS)))
1287
		return -1;
1288

1289
	return 0;
1290 1291 1292

out:
	spin_unlock_irq(target->scsi_host->host_lock);
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305
	return -1;
}

static int srp_find_req(struct srp_target_port *target,
			struct scsi_cmnd *scmnd,
			struct srp_request **req)
{
	if (scmnd->host_scribble == (void *) -1L)
		return -1;

	*req = &target->req_ring[(long) scmnd->host_scribble];

	return 0;
1306 1307 1308 1309
}

static int srp_abort(struct scsi_cmnd *scmnd)
{
1310 1311 1312 1313
	struct srp_target_port *target = host_to_target(scmnd->device->host);
	struct srp_request *req;
	int ret = SUCCESS;

1314 1315
	printk(KERN_ERR "SRP abort called\n");

1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
	if (srp_find_req(target, scmnd, &req))
		return FAILED;
	if (srp_send_tsk_mgmt(target, req, SRP_TSK_ABORT_TASK))
		return FAILED;

	spin_lock_irq(target->scsi_host->host_lock);

	if (req->cmd_done) {
		srp_remove_req(target, req);
		scmnd->scsi_done(scmnd);
	} else if (!req->tsk_status) {
		srp_remove_req(target, req);
		scmnd->result = DID_ABORT << 16;
	} else
		ret = FAILED;

	spin_unlock_irq(target->scsi_host->host_lock);

	return ret;
1335 1336 1337 1338
}

static int srp_reset_device(struct scsi_cmnd *scmnd)
{
1339 1340 1341
	struct srp_target_port *target = host_to_target(scmnd->device->host);
	struct srp_request *req, *tmp;

1342 1343
	printk(KERN_ERR "SRP reset_device called\n");

1344 1345 1346 1347 1348 1349 1350 1351 1352 1353
	if (srp_find_req(target, scmnd, &req))
		return FAILED;
	if (srp_send_tsk_mgmt(target, req, SRP_TSK_LUN_RESET))
		return FAILED;
	if (req->tsk_status)
		return FAILED;

	spin_lock_irq(target->scsi_host->host_lock);

	list_for_each_entry_safe(req, tmp, &target->req_queue, list)
1354 1355
		if (req->scmnd->device == scmnd->device)
			srp_reset_req(target, req);
1356 1357 1358 1359

	spin_unlock_irq(target->scsi_host->host_lock);

	return SUCCESS;
1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
}

static int srp_reset_host(struct scsi_cmnd *scmnd)
{
	struct srp_target_port *target = host_to_target(scmnd->device->host);
	int ret = FAILED;

	printk(KERN_ERR PFX "SRP reset_host called\n");

	if (!srp_reconnect_target(target))
		ret = SUCCESS;

	return ret;
}

1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440
static ssize_t show_id_ext(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED)
		return -ENODEV;

	return sprintf(buf, "0x%016llx\n",
		       (unsigned long long) be64_to_cpu(target->id_ext));
}

static ssize_t show_ioc_guid(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED)
		return -ENODEV;

	return sprintf(buf, "0x%016llx\n",
		       (unsigned long long) be64_to_cpu(target->ioc_guid));
}

static ssize_t show_service_id(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED)
		return -ENODEV;

	return sprintf(buf, "0x%016llx\n",
		       (unsigned long long) be64_to_cpu(target->service_id));
}

static ssize_t show_pkey(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED)
		return -ENODEV;

	return sprintf(buf, "0x%04x\n", be16_to_cpu(target->path.pkey));
}

static ssize_t show_dgid(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED)
		return -ENODEV;

	return sprintf(buf, "%04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n",
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[0]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[1]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[2]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[3]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[4]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[5]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[6]),
		       be16_to_cpu(((__be16 *) target->path.dgid.raw)[7]));
}

1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
static ssize_t show_zero_req_lim(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	if (target->state == SRP_TARGET_DEAD ||
	    target->state == SRP_TARGET_REMOVED)
		return -ENODEV;

	return sprintf(buf, "%d\n", target->zero_req_lim);
}

1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473
static ssize_t show_local_ib_port(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	return sprintf(buf, "%d\n", target->srp_host->port);
}

static ssize_t show_local_ib_device(struct class_device *cdev, char *buf)
{
	struct srp_target_port *target = host_to_target(class_to_shost(cdev));

	return sprintf(buf, "%s\n", target->srp_host->dev->dev->name);
}

static CLASS_DEVICE_ATTR(id_ext,	  S_IRUGO, show_id_ext,		 NULL);
static CLASS_DEVICE_ATTR(ioc_guid,	  S_IRUGO, show_ioc_guid,	 NULL);
static CLASS_DEVICE_ATTR(service_id,	  S_IRUGO, show_service_id,	 NULL);
static CLASS_DEVICE_ATTR(pkey,		  S_IRUGO, show_pkey,		 NULL);
static CLASS_DEVICE_ATTR(dgid,		  S_IRUGO, show_dgid,		 NULL);
static CLASS_DEVICE_ATTR(zero_req_lim,	  S_IRUGO, show_zero_req_lim,	 NULL);
static CLASS_DEVICE_ATTR(local_ib_port,   S_IRUGO, show_local_ib_port,	 NULL);
static CLASS_DEVICE_ATTR(local_ib_device, S_IRUGO, show_local_ib_device, NULL);
1474 1475 1476 1477 1478 1479 1480

static struct class_device_attribute *srp_host_attrs[] = {
	&class_device_attr_id_ext,
	&class_device_attr_ioc_guid,
	&class_device_attr_service_id,
	&class_device_attr_pkey,
	&class_device_attr_dgid,
1481
	&class_device_attr_zero_req_lim,
1482 1483
	&class_device_attr_local_ib_port,
	&class_device_attr_local_ib_device,
1484 1485 1486
	NULL
};

1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497
static struct scsi_host_template srp_template = {
	.module				= THIS_MODULE,
	.name				= DRV_NAME,
	.info				= srp_target_info,
	.queuecommand			= srp_queuecommand,
	.eh_abort_handler		= srp_abort,
	.eh_device_reset_handler	= srp_reset_device,
	.eh_host_reset_handler		= srp_reset_host,
	.can_queue			= SRP_SQ_SIZE,
	.this_id			= -1,
	.cmd_per_lun			= SRP_SQ_SIZE,
1498 1499
	.use_clustering			= ENABLE_CLUSTERING,
	.shost_attrs			= srp_host_attrs
1500 1501 1502 1503 1504 1505 1506
};

static int srp_add_target(struct srp_host *host, struct srp_target_port *target)
{
	sprintf(target->target_name, "SRP.T10:%016llX",
		 (unsigned long long) be64_to_cpu(target->id_ext));

1507
	if (scsi_add_host(target->scsi_host, host->dev->dev->dma_device))
1508 1509
		return -ENODEV;

1510
	spin_lock(&host->target_lock);
1511
	list_add_tail(&target->list, &host->target_list);
1512
	spin_unlock(&host->target_lock);
1513 1514 1515 1516

	target->state = SRP_TARGET_LIVE;

	scsi_scan_target(&target->scsi_host->shost_gendev,
1517
			 0, target->scsi_id, SCAN_WILD_CARD, 0);
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550

	return 0;
}

static void srp_release_class_dev(struct class_device *class_dev)
{
	struct srp_host *host =
		container_of(class_dev, struct srp_host, class_dev);

	complete(&host->released);
}

static struct class srp_class = {
	.name    = "infiniband_srp",
	.release = srp_release_class_dev
};

/*
 * Target ports are added by writing
 *
 *     id_ext=<SRP ID ext>,ioc_guid=<SRP IOC GUID>,dgid=<dest GID>,
 *     pkey=<P_Key>,service_id=<service ID>
 *
 * to the add_target sysfs attribute.
 */
enum {
	SRP_OPT_ERR		= 0,
	SRP_OPT_ID_EXT		= 1 << 0,
	SRP_OPT_IOC_GUID	= 1 << 1,
	SRP_OPT_DGID		= 1 << 2,
	SRP_OPT_PKEY		= 1 << 3,
	SRP_OPT_SERVICE_ID	= 1 << 4,
	SRP_OPT_MAX_SECT	= 1 << 5,
1551
	SRP_OPT_MAX_CMD_PER_LUN	= 1 << 6,
1552
	SRP_OPT_IO_CLASS	= 1 << 7,
1553 1554 1555 1556 1557 1558 1559 1560
	SRP_OPT_ALL		= (SRP_OPT_ID_EXT	|
				   SRP_OPT_IOC_GUID	|
				   SRP_OPT_DGID		|
				   SRP_OPT_PKEY		|
				   SRP_OPT_SERVICE_ID),
};

static match_table_t srp_opt_tokens = {
1561 1562 1563 1564 1565 1566 1567
	{ SRP_OPT_ID_EXT,		"id_ext=%s" 		},
	{ SRP_OPT_IOC_GUID,		"ioc_guid=%s" 		},
	{ SRP_OPT_DGID,			"dgid=%s" 		},
	{ SRP_OPT_PKEY,			"pkey=%x" 		},
	{ SRP_OPT_SERVICE_ID,		"service_id=%s"		},
	{ SRP_OPT_MAX_SECT,		"max_sect=%d" 		},
	{ SRP_OPT_MAX_CMD_PER_LUN,	"max_cmd_per_lun=%d" 	},
1568
	{ SRP_OPT_IO_CLASS,		"io_class=%x"		},
1569
	{ SRP_OPT_ERR,			NULL 			}
1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611
};

static int srp_parse_options(const char *buf, struct srp_target_port *target)
{
	char *options, *sep_opt;
	char *p;
	char dgid[3];
	substring_t args[MAX_OPT_ARGS];
	int opt_mask = 0;
	int token;
	int ret = -EINVAL;
	int i;

	options = kstrdup(buf, GFP_KERNEL);
	if (!options)
		return -ENOMEM;

	sep_opt = options;
	while ((p = strsep(&sep_opt, ",")) != NULL) {
		if (!*p)
			continue;

		token = match_token(p, srp_opt_tokens, args);
		opt_mask |= token;

		switch (token) {
		case SRP_OPT_ID_EXT:
			p = match_strdup(args);
			target->id_ext = cpu_to_be64(simple_strtoull(p, NULL, 16));
			kfree(p);
			break;

		case SRP_OPT_IOC_GUID:
			p = match_strdup(args);
			target->ioc_guid = cpu_to_be64(simple_strtoull(p, NULL, 16));
			kfree(p);
			break;

		case SRP_OPT_DGID:
			p = match_strdup(args);
			if (strlen(p) != 32) {
				printk(KERN_WARNING PFX "bad dest GID parameter '%s'\n", p);
1612
				kfree(p);
1613 1614 1615 1616 1617 1618 1619
				goto out;
			}

			for (i = 0; i < 16; ++i) {
				strlcpy(dgid, p + i * 2, 3);
				target->path.dgid.raw[i] = simple_strtoul(dgid, NULL, 16);
			}
1620
			kfree(p);
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644
			break;

		case SRP_OPT_PKEY:
			if (match_hex(args, &token)) {
				printk(KERN_WARNING PFX "bad P_Key parameter '%s'\n", p);
				goto out;
			}
			target->path.pkey = cpu_to_be16(token);
			break;

		case SRP_OPT_SERVICE_ID:
			p = match_strdup(args);
			target->service_id = cpu_to_be64(simple_strtoull(p, NULL, 16));
			kfree(p);
			break;

		case SRP_OPT_MAX_SECT:
			if (match_int(args, &token)) {
				printk(KERN_WARNING PFX "bad max sect parameter '%s'\n", p);
				goto out;
			}
			target->scsi_host->max_sectors = token;
			break;

1645 1646 1647 1648 1649 1650 1651 1652
		case SRP_OPT_MAX_CMD_PER_LUN:
			if (match_int(args, &token)) {
				printk(KERN_WARNING PFX "bad max cmd_per_lun parameter '%s'\n", p);
				goto out;
			}
			target->scsi_host->cmd_per_lun = min(token, SRP_SQ_SIZE);
			break;

1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
		case SRP_OPT_IO_CLASS:
			if (match_hex(args, &token)) {
				printk(KERN_WARNING PFX "bad  IO class parameter '%s' \n", p);
				goto out;
			}
			if (token != SRP_REV10_IB_IO_CLASS &&
			    token != SRP_REV16A_IB_IO_CLASS) {
				printk(KERN_WARNING PFX "unknown IO class parameter value"
				       " %x specified (use %x or %x).\n",
				       token, SRP_REV10_IB_IO_CLASS, SRP_REV16A_IB_IO_CLASS);
				goto out;
			}
			target->io_class = token;
			break;

1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704
		default:
			printk(KERN_WARNING PFX "unknown parameter or missing value "
			       "'%s' in target creation request\n", p);
			goto out;
		}
	}

	if ((opt_mask & SRP_OPT_ALL) == SRP_OPT_ALL)
		ret = 0;
	else
		for (i = 0; i < ARRAY_SIZE(srp_opt_tokens); ++i)
			if ((srp_opt_tokens[i].token & SRP_OPT_ALL) &&
			    !(srp_opt_tokens[i].token & opt_mask))
				printk(KERN_WARNING PFX "target creation request is "
				       "missing parameter '%s'\n",
				       srp_opt_tokens[i].pattern);

out:
	kfree(options);
	return ret;
}

static ssize_t srp_create_target(struct class_device *class_dev,
				 const char *buf, size_t count)
{
	struct srp_host *host =
		container_of(class_dev, struct srp_host, class_dev);
	struct Scsi_Host *target_host;
	struct srp_target_port *target;
	int ret;
	int i;

	target_host = scsi_host_alloc(&srp_template,
				      sizeof (struct srp_target_port));
	if (!target_host)
		return -ENOMEM;

R
Roland Dreier 已提交
1705 1706
	target_host->max_lun = SRP_MAX_LUN;

1707 1708 1709
	target = host_to_target(target_host);
	memset(target, 0, sizeof *target);

1710
	target->io_class   = SRP_REV16A_IB_IO_CLASS;
1711 1712 1713
	target->scsi_host  = target_host;
	target->srp_host   = host;

1714
	INIT_LIST_HEAD(&target->free_reqs);
1715
	INIT_LIST_HEAD(&target->req_queue);
1716 1717 1718 1719
	for (i = 0; i < SRP_SQ_SIZE; ++i) {
		target->req_ring[i].index = i;
		list_add_tail(&target->req_ring[i].list, &target->free_reqs);
	}
1720 1721 1722 1723 1724

	ret = srp_parse_options(buf, target);
	if (ret)
		goto err;

1725
	ib_get_cached_gid(host->dev->dev, host->port, 0, &target->path.sgid);
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745

	printk(KERN_DEBUG PFX "new target: id_ext %016llx ioc_guid %016llx pkey %04x "
	       "service_id %016llx dgid %04x:%04x:%04x:%04x:%04x:%04x:%04x:%04x\n",
	       (unsigned long long) be64_to_cpu(target->id_ext),
	       (unsigned long long) be64_to_cpu(target->ioc_guid),
	       be16_to_cpu(target->path.pkey),
	       (unsigned long long) be64_to_cpu(target->service_id),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[0]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[2]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[4]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[6]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[8]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[10]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[12]),
	       (int) be16_to_cpu(*(__be16 *) &target->path.dgid.raw[14]));

	ret = srp_create_target_ib(target);
	if (ret)
		goto err;

1746
	target->cm_id = ib_create_cm_id(host->dev->dev, srp_cm_handler, target);
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
	if (IS_ERR(target->cm_id)) {
		ret = PTR_ERR(target->cm_id);
		goto err_free;
	}

	ret = srp_connect_target(target);
	if (ret) {
		printk(KERN_ERR PFX "Connection failed\n");
		goto err_cm_id;
	}

	ret = srp_add_target(host, target);
	if (ret)
		goto err_disconnect;

	return count;

err_disconnect:
	srp_disconnect_target(target);

err_cm_id:
	ib_destroy_cm_id(target->cm_id);

err_free:
	srp_free_target_ib(target);

err:
	scsi_host_put(target_host);

	return ret;
}

static CLASS_DEVICE_ATTR(add_target, S_IWUSR, NULL, srp_create_target);

static ssize_t show_ibdev(struct class_device *class_dev, char *buf)
{
	struct srp_host *host =
		container_of(class_dev, struct srp_host, class_dev);

1786
	return sprintf(buf, "%s\n", host->dev->dev->name);
1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800
}

static CLASS_DEVICE_ATTR(ibdev, S_IRUGO, show_ibdev, NULL);

static ssize_t show_port(struct class_device *class_dev, char *buf)
{
	struct srp_host *host =
		container_of(class_dev, struct srp_host, class_dev);

	return sprintf(buf, "%d\n", host->port);
}

static CLASS_DEVICE_ATTR(port, S_IRUGO, show_port, NULL);

1801
static struct srp_host *srp_add_port(struct srp_device *device, u8 port)
1802 1803 1804 1805 1806 1807 1808 1809
{
	struct srp_host *host;

	host = kzalloc(sizeof *host, GFP_KERNEL);
	if (!host)
		return NULL;

	INIT_LIST_HEAD(&host->target_list);
1810
	spin_lock_init(&host->target_lock);
1811 1812 1813 1814 1815
	init_completion(&host->released);
	host->dev  = device;
	host->port = port;

	host->initiator_port_id[7] = port;
1816
	memcpy(host->initiator_port_id + 8, &device->dev->node_guid, 8);
1817 1818

	host->class_dev.class = &srp_class;
1819
	host->class_dev.dev   = device->dev->dma_device;
1820
	snprintf(host->class_dev.class_id, BUS_ID_SIZE, "srp-%s-%d",
1821
		 device->dev->name, port);
1822 1823

	if (class_device_register(&host->class_dev))
1824
		goto free_host;
1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836
	if (class_device_create_file(&host->class_dev, &class_device_attr_add_target))
		goto err_class;
	if (class_device_create_file(&host->class_dev, &class_device_attr_ibdev))
		goto err_class;
	if (class_device_create_file(&host->class_dev, &class_device_attr_port))
		goto err_class;

	return host;

err_class:
	class_device_unregister(&host->class_dev);

1837
free_host:
1838 1839 1840 1841 1842 1843 1844
	kfree(host);

	return NULL;
}

static void srp_add_one(struct ib_device *device)
{
1845 1846 1847
	struct srp_device *srp_dev;
	struct ib_device_attr *dev_attr;
	struct ib_fmr_pool_param fmr_param;
1848 1849 1850
	struct srp_host *host;
	int s, e, p;

1851 1852
	dev_attr = kmalloc(sizeof *dev_attr, GFP_KERNEL);
	if (!dev_attr)
1853
		return;
1854

1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900
	if (ib_query_device(device, dev_attr)) {
		printk(KERN_WARNING PFX "Query device failed for %s\n",
		       device->name);
		goto free_attr;
	}

	srp_dev = kmalloc(sizeof *srp_dev, GFP_KERNEL);
	if (!srp_dev)
		goto free_attr;

	/*
	 * Use the smallest page size supported by the HCA, down to a
	 * minimum of 512 bytes (which is the smallest sector that a
	 * SCSI command will ever carry).
	 */
	srp_dev->fmr_page_shift = max(9, ffs(dev_attr->page_size_cap) - 1);
	srp_dev->fmr_page_size  = 1 << srp_dev->fmr_page_shift;
	srp_dev->fmr_page_mask  = ~((unsigned long) srp_dev->fmr_page_size - 1);

	INIT_LIST_HEAD(&srp_dev->dev_list);

	srp_dev->dev = device;
	srp_dev->pd  = ib_alloc_pd(device);
	if (IS_ERR(srp_dev->pd))
		goto free_dev;

	srp_dev->mr = ib_get_dma_mr(srp_dev->pd,
				    IB_ACCESS_LOCAL_WRITE |
				    IB_ACCESS_REMOTE_READ |
				    IB_ACCESS_REMOTE_WRITE);
	if (IS_ERR(srp_dev->mr))
		goto err_pd;

	memset(&fmr_param, 0, sizeof fmr_param);
	fmr_param.pool_size	    = SRP_FMR_POOL_SIZE;
	fmr_param.dirty_watermark   = SRP_FMR_DIRTY_SIZE;
	fmr_param.cache		    = 1;
	fmr_param.max_pages_per_fmr = SRP_FMR_SIZE;
	fmr_param.page_shift	    = srp_dev->fmr_page_shift;
	fmr_param.access	    = (IB_ACCESS_LOCAL_WRITE |
				       IB_ACCESS_REMOTE_WRITE |
				       IB_ACCESS_REMOTE_READ);

	srp_dev->fmr_pool = ib_create_fmr_pool(srp_dev->pd, &fmr_param);
	if (IS_ERR(srp_dev->fmr_pool))
		srp_dev->fmr_pool = NULL;
1901 1902 1903 1904 1905 1906 1907 1908 1909 1910

	if (device->node_type == IB_NODE_SWITCH) {
		s = 0;
		e = 0;
	} else {
		s = 1;
		e = device->phys_port_cnt;
	}

	for (p = s; p <= e; ++p) {
1911
		host = srp_add_port(srp_dev, p);
1912
		if (host)
1913
			list_add_tail(&host->list, &srp_dev->dev_list);
1914 1915
	}

1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927
	ib_set_client_data(device, &srp_client, srp_dev);

	goto free_attr;

err_pd:
	ib_dealloc_pd(srp_dev->pd);

free_dev:
	kfree(srp_dev);

free_attr:
	kfree(dev_attr);
1928 1929 1930 1931
}

static void srp_remove_one(struct ib_device *device)
{
1932
	struct srp_device *srp_dev;
1933 1934 1935 1936
	struct srp_host *host, *tmp_host;
	LIST_HEAD(target_list);
	struct srp_target_port *target, *tmp_target;

1937
	srp_dev = ib_get_client_data(device, &srp_client);
1938

1939
	list_for_each_entry_safe(host, tmp_host, &srp_dev->dev_list, list) {
1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950
		class_device_unregister(&host->class_dev);
		/*
		 * Wait for the sysfs entry to go away, so that no new
		 * target ports can be created.
		 */
		wait_for_completion(&host->released);

		/*
		 * Mark all target ports as removed, so we stop queueing
		 * commands and don't try to reconnect.
		 */
1951
		spin_lock(&host->target_lock);
1952
		list_for_each_entry(target, &host->target_list, list) {
1953 1954 1955
			spin_lock_irq(target->scsi_host->host_lock);
			target->state = SRP_TARGET_REMOVED;
			spin_unlock_irq(target->scsi_host->host_lock);
1956
		}
1957
		spin_unlock(&host->target_lock);
1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977

		/*
		 * Wait for any reconnection tasks that may have
		 * started before we marked our target ports as
		 * removed, and any target port removal tasks.
		 */
		flush_scheduled_work();

		list_for_each_entry_safe(target, tmp_target,
					 &host->target_list, list) {
			scsi_remove_host(target->scsi_host);
			srp_disconnect_target(target);
			ib_destroy_cm_id(target->cm_id);
			srp_free_target_ib(target);
			scsi_host_put(target->scsi_host);
		}

		kfree(host);
	}

1978 1979 1980 1981 1982 1983
	if (srp_dev->fmr_pool)
		ib_destroy_fmr_pool(srp_dev->fmr_pool);
	ib_dereg_mr(srp_dev->mr);
	ib_dealloc_pd(srp_dev->pd);

	kfree(srp_dev);
1984 1985 1986 1987 1988 1989
}

static int __init srp_init_module(void)
{
	int ret;

1990 1991 1992 1993 1994
	srp_template.sg_tablesize = srp_sg_tablesize;
	srp_max_iu_len = (sizeof (struct srp_cmd) +
			  sizeof (struct srp_indirect_buf) +
			  srp_sg_tablesize * 16);

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
	ret = class_register(&srp_class);
	if (ret) {
		printk(KERN_ERR PFX "couldn't register class infiniband_srp\n");
		return ret;
	}

	ret = ib_register_client(&srp_client);
	if (ret) {
		printk(KERN_ERR PFX "couldn't register IB client\n");
		class_unregister(&srp_class);
		return ret;
	}

	return 0;
}

static void __exit srp_cleanup_module(void)
{
	ib_unregister_client(&srp_client);
	class_unregister(&srp_class);
}

module_init(srp_init_module);
module_exit(srp_cleanup_module);