core.c 34.7 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2 3 4 5 6 7
/*
 * Common code for the NVMe target.
 * Copyright (c) 2015-2016 HGST, a Western Digital Company.
 */
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/module.h>
8
#include <linux/random.h>
9
#include <linux/rculist.h>
10
#include <linux/pci-p2pdma.h>
11
#include <linux/scatterlist.h>
12

13 14
#include "nvmet.h"

15
struct workqueue_struct *buffered_io_wq;
16
static const struct nvmet_fabrics_ops *nvmet_transports[NVMF_TRTYPE_MAX];
17
static DEFINE_IDA(cntlid_ida);
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

/*
 * This read/write semaphore is used to synchronize access to configuration
 * information on a target system that will result in discovery log page
 * information change for at least one host.
 * The full list of resources to protected by this semaphore is:
 *
 *  - subsystems list
 *  - per-subsystem allowed hosts list
 *  - allow_any_host subsystem attribute
 *  - nvmet_genctr
 *  - the nvmet_transports array
 *
 * When updating any of those lists/structures write lock should be obtained,
 * while when reading (popolating discovery log page or checking host-subsystem
 * link) read lock is obtained to allow concurrent reads.
 */
DECLARE_RWSEM(nvmet_config_sem);

37 38 39 40
u32 nvmet_ana_group_enabled[NVMET_MAX_ANAGRPS + 1];
u64 nvmet_ana_chgcnt;
DECLARE_RWSEM(nvmet_ana_sem);

41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78
inline u16 errno_to_nvme_status(struct nvmet_req *req, int errno)
{
	u16 status;

	switch (errno) {
	case -ENOSPC:
		req->error_loc = offsetof(struct nvme_rw_command, length);
		status = NVME_SC_CAP_EXCEEDED | NVME_SC_DNR;
		break;
	case -EREMOTEIO:
		req->error_loc = offsetof(struct nvme_rw_command, slba);
		status = NVME_SC_LBA_RANGE | NVME_SC_DNR;
		break;
	case -EOPNOTSUPP:
		req->error_loc = offsetof(struct nvme_common_command, opcode);
		switch (req->cmd->common.opcode) {
		case nvme_cmd_dsm:
		case nvme_cmd_write_zeroes:
			status = NVME_SC_ONCS_NOT_SUPPORTED | NVME_SC_DNR;
			break;
		default:
			status = NVME_SC_INVALID_OPCODE | NVME_SC_DNR;
		}
		break;
	case -ENODATA:
		req->error_loc = offsetof(struct nvme_rw_command, nsid);
		status = NVME_SC_ACCESS_DENIED;
		break;
	case -EIO:
		/* FALLTHRU */
	default:
		req->error_loc = offsetof(struct nvme_common_command, opcode);
		status = NVME_SC_INTERNAL | NVME_SC_DNR;
	}

	return status;
}

79 80 81 82 83 84
static struct nvmet_subsys *nvmet_find_get_subsys(struct nvmet_port *port,
		const char *subsysnqn);

u16 nvmet_copy_to_sgl(struct nvmet_req *req, off_t off, const void *buf,
		size_t len)
{
85 86
	if (sg_pcopy_from_buffer(req->sg, req->sg_cnt, buf, len, off) != len) {
		req->error_loc = offsetof(struct nvme_common_command, dptr);
87
		return NVME_SC_SGL_INVALID_DATA | NVME_SC_DNR;
88
	}
89 90 91 92 93
	return 0;
}

u16 nvmet_copy_from_sgl(struct nvmet_req *req, off_t off, void *buf, size_t len)
{
94 95
	if (sg_pcopy_to_buffer(req->sg, req->sg_cnt, buf, len, off) != len) {
		req->error_loc = offsetof(struct nvme_common_command, dptr);
96
		return NVME_SC_SGL_INVALID_DATA | NVME_SC_DNR;
97
	}
98 99 100
	return 0;
}

101 102
u16 nvmet_zero_sgl(struct nvmet_req *req, off_t off, size_t len)
{
103 104
	if (sg_zero_buffer(req->sg, req->sg_cnt, len, off) != len) {
		req->error_loc = offsetof(struct nvme_common_command, dptr);
105
		return NVME_SC_SGL_INVALID_DATA | NVME_SC_DNR;
106
	}
107 108 109
	return 0;
}

110 111 112 113 114 115 116 117 118 119 120
static unsigned int nvmet_max_nsid(struct nvmet_subsys *subsys)
{
	struct nvmet_ns *ns;

	if (list_empty(&subsys->namespaces))
		return 0;

	ns = list_last_entry(&subsys->namespaces, struct nvmet_ns, dev_link);
	return ns->nsid;
}

121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
static u32 nvmet_async_event_result(struct nvmet_async_event *aen)
{
	return aen->event_type | (aen->event_info << 8) | (aen->log_page << 16);
}

static void nvmet_async_events_free(struct nvmet_ctrl *ctrl)
{
	struct nvmet_req *req;

	while (1) {
		mutex_lock(&ctrl->lock);
		if (!ctrl->nr_async_event_cmds) {
			mutex_unlock(&ctrl->lock);
			return;
		}

		req = ctrl->async_event_cmds[--ctrl->nr_async_event_cmds];
		mutex_unlock(&ctrl->lock);
		nvmet_req_complete(req, NVME_SC_INTERNAL | NVME_SC_DNR);
	}
}

static void nvmet_async_event_work(struct work_struct *work)
{
	struct nvmet_ctrl *ctrl =
		container_of(work, struct nvmet_ctrl, async_event_work);
	struct nvmet_async_event *aen;
	struct nvmet_req *req;

	while (1) {
		mutex_lock(&ctrl->lock);
		aen = list_first_entry_or_null(&ctrl->async_events,
				struct nvmet_async_event, entry);
		if (!aen || !ctrl->nr_async_event_cmds) {
			mutex_unlock(&ctrl->lock);
			return;
		}

		req = ctrl->async_event_cmds[--ctrl->nr_async_event_cmds];
		nvmet_set_result(req, nvmet_async_event_result(aen));

		list_del(&aen->entry);
		kfree(aen);

		mutex_unlock(&ctrl->lock);
		nvmet_req_complete(req, 0);
	}
}

170
void nvmet_add_async_event(struct nvmet_ctrl *ctrl, u8 event_type,
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
		u8 event_info, u8 log_page)
{
	struct nvmet_async_event *aen;

	aen = kmalloc(sizeof(*aen), GFP_KERNEL);
	if (!aen)
		return;

	aen->event_type = event_type;
	aen->event_info = event_info;
	aen->log_page = log_page;

	mutex_lock(&ctrl->lock);
	list_add_tail(&aen->entry, &ctrl->async_events);
	mutex_unlock(&ctrl->lock);

	schedule_work(&ctrl->async_event_work);
}

190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
static void nvmet_add_to_changed_ns_log(struct nvmet_ctrl *ctrl, __le32 nsid)
{
	u32 i;

	mutex_lock(&ctrl->lock);
	if (ctrl->nr_changed_ns > NVME_MAX_CHANGED_NAMESPACES)
		goto out_unlock;

	for (i = 0; i < ctrl->nr_changed_ns; i++) {
		if (ctrl->changed_ns_list[i] == nsid)
			goto out_unlock;
	}

	if (ctrl->nr_changed_ns == NVME_MAX_CHANGED_NAMESPACES) {
		ctrl->changed_ns_list[0] = cpu_to_le32(0xffffffff);
		ctrl->nr_changed_ns = U32_MAX;
		goto out_unlock;
	}

	ctrl->changed_ns_list[ctrl->nr_changed_ns++] = nsid;
out_unlock:
	mutex_unlock(&ctrl->lock);
}

214
void nvmet_ns_changed(struct nvmet_subsys *subsys, u32 nsid)
215 216 217
{
	struct nvmet_ctrl *ctrl;

218 219
	lockdep_assert_held(&subsys->lock);

220 221
	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) {
		nvmet_add_to_changed_ns_log(ctrl, cpu_to_le32(nsid));
222
		if (nvmet_aen_bit_disabled(ctrl, NVME_AEN_BIT_NS_ATTR))
223
			continue;
224 225 226 227 228 229
		nvmet_add_async_event(ctrl, NVME_AER_TYPE_NOTICE,
				NVME_AER_NOTICE_NS_CHANGED,
				NVME_LOG_CHANGED_NS);
	}
}

230 231 232 233 234 235 236 237 238
void nvmet_send_ana_event(struct nvmet_subsys *subsys,
		struct nvmet_port *port)
{
	struct nvmet_ctrl *ctrl;

	mutex_lock(&subsys->lock);
	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) {
		if (port && ctrl->port != port)
			continue;
239
		if (nvmet_aen_bit_disabled(ctrl, NVME_AEN_BIT_ANA_CHANGE))
240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
			continue;
		nvmet_add_async_event(ctrl, NVME_AER_TYPE_NOTICE,
				NVME_AER_NOTICE_ANA, NVME_LOG_ANA);
	}
	mutex_unlock(&subsys->lock);
}

void nvmet_port_send_ana_event(struct nvmet_port *port)
{
	struct nvmet_subsys_link *p;

	down_read(&nvmet_config_sem);
	list_for_each_entry(p, &port->subsystems, entry)
		nvmet_send_ana_event(p->subsys, port);
	up_read(&nvmet_config_sem);
}

257
int nvmet_register_transport(const struct nvmet_fabrics_ops *ops)
258 259 260 261 262 263 264 265 266 267 268 269 270 271
{
	int ret = 0;

	down_write(&nvmet_config_sem);
	if (nvmet_transports[ops->type])
		ret = -EINVAL;
	else
		nvmet_transports[ops->type] = ops;
	up_write(&nvmet_config_sem);

	return ret;
}
EXPORT_SYMBOL_GPL(nvmet_register_transport);

272
void nvmet_unregister_transport(const struct nvmet_fabrics_ops *ops)
273 274 275 276 277 278 279 280 281
{
	down_write(&nvmet_config_sem);
	nvmet_transports[ops->type] = NULL;
	up_write(&nvmet_config_sem);
}
EXPORT_SYMBOL_GPL(nvmet_unregister_transport);

int nvmet_enable_port(struct nvmet_port *port)
{
282
	const struct nvmet_fabrics_ops *ops;
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
	int ret;

	lockdep_assert_held(&nvmet_config_sem);

	ops = nvmet_transports[port->disc_addr.trtype];
	if (!ops) {
		up_write(&nvmet_config_sem);
		request_module("nvmet-transport-%d", port->disc_addr.trtype);
		down_write(&nvmet_config_sem);
		ops = nvmet_transports[port->disc_addr.trtype];
		if (!ops) {
			pr_err("transport type %d not supported\n",
				port->disc_addr.trtype);
			return -EINVAL;
		}
	}

	if (!try_module_get(ops->owner))
		return -EINVAL;

	ret = ops->add_port(port);
	if (ret) {
		module_put(ops->owner);
		return ret;
	}

309 310 311 312
	/* If the transport didn't set inline_data_size, then disable it. */
	if (port->inline_data_size < 0)
		port->inline_data_size = 0;

313
	port->enabled = true;
314
	port->tr_ops = ops;
315 316 317 318 319
	return 0;
}

void nvmet_disable_port(struct nvmet_port *port)
{
320
	const struct nvmet_fabrics_ops *ops;
321 322 323 324

	lockdep_assert_held(&nvmet_config_sem);

	port->enabled = false;
325
	port->tr_ops = NULL;
326 327 328 329 330 331 332 333 334 335

	ops = nvmet_transports[port->disc_addr.trtype];
	ops->remove_port(port);
	module_put(ops->owner);
}

static void nvmet_keep_alive_timer(struct work_struct *work)
{
	struct nvmet_ctrl *ctrl = container_of(to_delayed_work(work),
			struct nvmet_ctrl, ka_work);
336 337 338 339 340 341 342 343 344
	bool cmd_seen = ctrl->cmd_seen;

	ctrl->cmd_seen = false;
	if (cmd_seen) {
		pr_debug("ctrl %d reschedule traffic based keep-alive timer\n",
			ctrl->cntlid);
		schedule_delayed_work(&ctrl->ka_work, ctrl->kato * HZ);
		return;
	}
345 346 347 348

	pr_err("ctrl %d keep-alive timer (%d seconds) expired!\n",
		ctrl->cntlid, ctrl->kato);

349
	nvmet_ctrl_fatal_error(ctrl);
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
}

static void nvmet_start_keep_alive_timer(struct nvmet_ctrl *ctrl)
{
	pr_debug("ctrl %d start keep-alive timer for %d secs\n",
		ctrl->cntlid, ctrl->kato);

	INIT_DELAYED_WORK(&ctrl->ka_work, nvmet_keep_alive_timer);
	schedule_delayed_work(&ctrl->ka_work, ctrl->kato * HZ);
}

static void nvmet_stop_keep_alive_timer(struct nvmet_ctrl *ctrl)
{
	pr_debug("ctrl %d stop keep-alive\n", ctrl->cntlid);

	cancel_delayed_work_sync(&ctrl->ka_work);
}

static struct nvmet_ns *__nvmet_find_namespace(struct nvmet_ctrl *ctrl,
		__le32 nsid)
{
	struct nvmet_ns *ns;

	list_for_each_entry_rcu(ns, &ctrl->subsys->namespaces, dev_link) {
		if (ns->nsid == le32_to_cpu(nsid))
			return ns;
	}

	return NULL;
}

struct nvmet_ns *nvmet_find_namespace(struct nvmet_ctrl *ctrl, __le32 nsid)
{
	struct nvmet_ns *ns;

	rcu_read_lock();
	ns = __nvmet_find_namespace(ctrl, nsid);
	if (ns)
		percpu_ref_get(&ns->ref);
	rcu_read_unlock();

	return ns;
}

static void nvmet_destroy_namespace(struct percpu_ref *ref)
{
	struct nvmet_ns *ns = container_of(ref, struct nvmet_ns, ref);

	complete(&ns->disable_done);
}

void nvmet_put_namespace(struct nvmet_ns *ns)
{
	percpu_ref_put(&ns->ref);
}

406 407 408 409 410 411
static void nvmet_ns_dev_disable(struct nvmet_ns *ns)
{
	nvmet_bdev_ns_disable(ns);
	nvmet_file_ns_disable(ns);
}

412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
static int nvmet_p2pmem_ns_enable(struct nvmet_ns *ns)
{
	int ret;
	struct pci_dev *p2p_dev;

	if (!ns->use_p2pmem)
		return 0;

	if (!ns->bdev) {
		pr_err("peer-to-peer DMA is not supported by non-block device namespaces\n");
		return -EINVAL;
	}

	if (!blk_queue_pci_p2pdma(ns->bdev->bd_queue)) {
		pr_err("peer-to-peer DMA is not supported by the driver of %s\n",
		       ns->device_path);
		return -EINVAL;
	}

	if (ns->p2p_dev) {
		ret = pci_p2pdma_distance(ns->p2p_dev, nvmet_ns_dev(ns), true);
		if (ret < 0)
			return -EINVAL;
	} else {
		/*
		 * Right now we just check that there is p2pmem available so
		 * we can report an error to the user right away if there
		 * is not. We'll find the actual device to use once we
		 * setup the controller when the port's device is available.
		 */

		p2p_dev = pci_p2pmem_find(nvmet_ns_dev(ns));
		if (!p2p_dev) {
			pr_err("no peer-to-peer memory is available for %s\n",
			       ns->device_path);
			return -EINVAL;
		}

		pci_dev_put(p2p_dev);
	}

	return 0;
}

/*
 * Note: ctrl->subsys->lock should be held when calling this function
 */
static void nvmet_p2pmem_ns_add_p2p(struct nvmet_ctrl *ctrl,
				    struct nvmet_ns *ns)
{
	struct device *clients[2];
	struct pci_dev *p2p_dev;
	int ret;

466
	if (!ctrl->p2p_client || !ns->use_p2pmem)
467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494
		return;

	if (ns->p2p_dev) {
		ret = pci_p2pdma_distance(ns->p2p_dev, ctrl->p2p_client, true);
		if (ret < 0)
			return;

		p2p_dev = pci_dev_get(ns->p2p_dev);
	} else {
		clients[0] = ctrl->p2p_client;
		clients[1] = nvmet_ns_dev(ns);

		p2p_dev = pci_p2pmem_find_many(clients, ARRAY_SIZE(clients));
		if (!p2p_dev) {
			pr_err("no peer-to-peer memory is available that's supported by %s and %s\n",
			       dev_name(ctrl->p2p_client), ns->device_path);
			return;
		}
	}

	ret = radix_tree_insert(&ctrl->p2p_ns_map, ns->nsid, p2p_dev);
	if (ret < 0)
		pci_dev_put(p2p_dev);

	pr_info("using p2pmem on %s for nsid %d\n", pci_name(p2p_dev),
		ns->nsid);
}

495 496 497
int nvmet_ns_enable(struct nvmet_ns *ns)
{
	struct nvmet_subsys *subsys = ns->subsys;
498
	struct nvmet_ctrl *ctrl;
499
	int ret;
500 501

	mutex_lock(&subsys->lock);
502
	ret = 0;
503
	if (ns->enabled)
504 505
		goto out_unlock;

506 507 508 509
	ret = -EMFILE;
	if (subsys->nr_namespaces == NVMET_MAX_NAMESPACES)
		goto out_unlock;

510
	ret = nvmet_bdev_ns_enable(ns);
511
	if (ret == -ENOTBLK)
512 513
		ret = nvmet_file_ns_enable(ns);
	if (ret)
514 515
		goto out_unlock;

516 517
	ret = nvmet_p2pmem_ns_enable(ns);
	if (ret)
518
		goto out_dev_disable;
519 520 521 522

	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry)
		nvmet_p2pmem_ns_add_p2p(ctrl, ns);

523 524 525
	ret = percpu_ref_init(&ns->ref, nvmet_destroy_namespace,
				0, GFP_KERNEL);
	if (ret)
526
		goto out_dev_put;
527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547

	if (ns->nsid > subsys->max_nsid)
		subsys->max_nsid = ns->nsid;

	/*
	 * The namespaces list needs to be sorted to simplify the implementation
	 * of the Identify Namepace List subcommand.
	 */
	if (list_empty(&subsys->namespaces)) {
		list_add_tail_rcu(&ns->dev_link, &subsys->namespaces);
	} else {
		struct nvmet_ns *old;

		list_for_each_entry_rcu(old, &subsys->namespaces, dev_link) {
			BUG_ON(ns->nsid == old->nsid);
			if (ns->nsid < old->nsid)
				break;
		}

		list_add_tail_rcu(&ns->dev_link, &old->dev_link);
	}
548
	subsys->nr_namespaces++;
549

550
	nvmet_ns_changed(subsys, ns->nsid);
551
	ns->enabled = true;
552 553 554 555
	ret = 0;
out_unlock:
	mutex_unlock(&subsys->lock);
	return ret;
556
out_dev_put:
557 558
	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry)
		pci_dev_put(radix_tree_delete(&ctrl->p2p_ns_map, ns->nsid));
559
out_dev_disable:
560
	nvmet_ns_dev_disable(ns);
561 562 563 564 565 566
	goto out_unlock;
}

void nvmet_ns_disable(struct nvmet_ns *ns)
{
	struct nvmet_subsys *subsys = ns->subsys;
567
	struct nvmet_ctrl *ctrl;
568 569

	mutex_lock(&subsys->lock);
570 571 572 573 574
	if (!ns->enabled)
		goto out_unlock;

	ns->enabled = false;
	list_del_rcu(&ns->dev_link);
575 576
	if (ns->nsid == subsys->max_nsid)
		subsys->max_nsid = nvmet_max_nsid(subsys);
577 578 579 580

	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry)
		pci_dev_put(radix_tree_delete(&ctrl->p2p_ns_map, ns->nsid));

581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596
	mutex_unlock(&subsys->lock);

	/*
	 * Now that we removed the namespaces from the lookup list, we
	 * can kill the per_cpu ref and wait for any remaining references
	 * to be dropped, as well as a RCU grace period for anyone only
	 * using the namepace under rcu_read_lock().  Note that we can't
	 * use call_rcu here as we need to ensure the namespaces have
	 * been fully destroyed before unloading the module.
	 */
	percpu_ref_kill(&ns->ref);
	synchronize_rcu();
	wait_for_completion(&ns->disable_done);
	percpu_ref_exit(&ns->ref);

	mutex_lock(&subsys->lock);
597

598
	subsys->nr_namespaces--;
599
	nvmet_ns_changed(subsys, ns->nsid);
600
	nvmet_ns_dev_disable(ns);
601
out_unlock:
602 603 604 605 606 607 608
	mutex_unlock(&subsys->lock);
}

void nvmet_ns_free(struct nvmet_ns *ns)
{
	nvmet_ns_disable(ns);

609 610 611 612
	down_write(&nvmet_ana_sem);
	nvmet_ana_group_enabled[ns->anagrpid]--;
	up_write(&nvmet_ana_sem);

613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629
	kfree(ns->device_path);
	kfree(ns);
}

struct nvmet_ns *nvmet_ns_alloc(struct nvmet_subsys *subsys, u32 nsid)
{
	struct nvmet_ns *ns;

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

	INIT_LIST_HEAD(&ns->dev_link);
	init_completion(&ns->disable_done);

	ns->nsid = nsid;
	ns->subsys = subsys;
630 631 632 633 634 635

	down_write(&nvmet_ana_sem);
	ns->anagrpid = NVMET_DEFAULT_ANA_GRPID;
	nvmet_ana_group_enabled[ns->anagrpid]++;
	up_write(&nvmet_ana_sem);

636
	uuid_gen(&ns->uuid);
637
	ns->buffered_io = false;
638 639 640 641

	return ns;
}

642
static void nvmet_update_sq_head(struct nvmet_req *req)
643
{
J
James Smart 已提交
644
	if (req->sq->size) {
645 646
		u32 old_sqhd, new_sqhd;

J
James Smart 已提交
647 648 649 650 651 652
		do {
			old_sqhd = req->sq->sqhd;
			new_sqhd = (old_sqhd + 1) % req->sq->size;
		} while (cmpxchg(&req->sq->sqhd, old_sqhd, new_sqhd) !=
					old_sqhd);
	}
653
	req->cqe->sq_head = cpu_to_le16(req->sq->sqhd & 0x0000FFFF);
654 655
}

656 657 658 659 660 661
static void nvmet_set_error(struct nvmet_req *req, u16 status)
{
	struct nvmet_ctrl *ctrl = req->sq->ctrl;
	struct nvme_error_slot *new_error_slot;
	unsigned long flags;

662
	req->cqe->status = cpu_to_le16(status << 1);
663

664
	if (!ctrl || req->error_loc == NVMET_NO_ERROR_LOC)
665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681
		return;

	spin_lock_irqsave(&ctrl->error_lock, flags);
	ctrl->err_counter++;
	new_error_slot =
		&ctrl->slots[ctrl->err_counter % NVMET_ERROR_LOG_SLOTS];

	new_error_slot->error_count = cpu_to_le64(ctrl->err_counter);
	new_error_slot->sqid = cpu_to_le16(req->sq->qid);
	new_error_slot->cmdid = cpu_to_le16(req->cmd->common.command_id);
	new_error_slot->status_field = cpu_to_le16(status << 1);
	new_error_slot->param_error_location = cpu_to_le16(req->error_loc);
	new_error_slot->lba = cpu_to_le64(req->error_slba);
	new_error_slot->nsid = req->cmd->common.nsid;
	spin_unlock_irqrestore(&ctrl->error_lock, flags);

	/* set the more bit for this request */
682
	req->cqe->status |= cpu_to_le16(1 << 14);
683 684
}

685 686 687 688
static void __nvmet_req_complete(struct nvmet_req *req, u16 status)
{
	if (!req->sq->sqhd_disabled)
		nvmet_update_sq_head(req);
689 690
	req->cqe->sq_id = cpu_to_le16(req->sq->qid);
	req->cqe->command_id = req->cmd->common.command_id;
691

692
	if (unlikely(status))
693
		nvmet_set_error(req, status);
694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717
	if (req->ns)
		nvmet_put_namespace(req->ns);
	req->ops->queue_response(req);
}

void nvmet_req_complete(struct nvmet_req *req, u16 status)
{
	__nvmet_req_complete(req, status);
	percpu_ref_put(&req->sq->ref);
}
EXPORT_SYMBOL_GPL(nvmet_req_complete);

void nvmet_cq_setup(struct nvmet_ctrl *ctrl, struct nvmet_cq *cq,
		u16 qid, u16 size)
{
	cq->qid = qid;
	cq->size = size;

	ctrl->cqs[qid] = cq;
}

void nvmet_sq_setup(struct nvmet_ctrl *ctrl, struct nvmet_sq *sq,
		u16 qid, u16 size)
{
718
	sq->sqhd = 0;
719 720 721 722 723 724
	sq->qid = qid;
	sq->size = size;

	ctrl->sqs[qid] = sq;
}

725 726 727 728 729 730 731
static void nvmet_confirm_sq(struct percpu_ref *ref)
{
	struct nvmet_sq *sq = container_of(ref, struct nvmet_sq, ref);

	complete(&sq->confirm_done);
}

732 733 734 735 736 737 738 739
void nvmet_sq_destroy(struct nvmet_sq *sq)
{
	/*
	 * If this is the admin queue, complete all AERs so that our
	 * queue doesn't have outstanding requests on it.
	 */
	if (sq->ctrl && sq->ctrl->sqs && sq->ctrl->sqs[0] == sq)
		nvmet_async_events_free(sq->ctrl);
740 741
	percpu_ref_kill_and_confirm(&sq->ref, nvmet_confirm_sq);
	wait_for_completion(&sq->confirm_done);
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
	wait_for_completion(&sq->free_done);
	percpu_ref_exit(&sq->ref);

	if (sq->ctrl) {
		nvmet_ctrl_put(sq->ctrl);
		sq->ctrl = NULL; /* allows reusing the queue later */
	}
}
EXPORT_SYMBOL_GPL(nvmet_sq_destroy);

static void nvmet_sq_free(struct percpu_ref *ref)
{
	struct nvmet_sq *sq = container_of(ref, struct nvmet_sq, ref);

	complete(&sq->free_done);
}

int nvmet_sq_init(struct nvmet_sq *sq)
{
	int ret;

	ret = percpu_ref_init(&sq->ref, nvmet_sq_free, 0, GFP_KERNEL);
	if (ret) {
		pr_err("percpu_ref init failed!\n");
		return ret;
	}
	init_completion(&sq->free_done);
769
	init_completion(&sq->confirm_done);
770 771 772 773 774

	return 0;
}
EXPORT_SYMBOL_GPL(nvmet_sq_init);

775 776 777 778 779 780 781 782 783 784 785 786 787 788
static inline u16 nvmet_check_ana_state(struct nvmet_port *port,
		struct nvmet_ns *ns)
{
	enum nvme_ana_state state = port->ana_state[ns->anagrpid];

	if (unlikely(state == NVME_ANA_INACCESSIBLE))
		return NVME_SC_ANA_INACCESSIBLE;
	if (unlikely(state == NVME_ANA_PERSISTENT_LOSS))
		return NVME_SC_ANA_PERSISTENT_LOSS;
	if (unlikely(state == NVME_ANA_CHANGE))
		return NVME_SC_ANA_TRANSITION;
	return 0;
}

789 790 791 792 793 794 795 796 797 798 799 800 801 802 803
static inline u16 nvmet_io_cmd_check_access(struct nvmet_req *req)
{
	if (unlikely(req->ns->readonly)) {
		switch (req->cmd->common.opcode) {
		case nvme_cmd_read:
		case nvme_cmd_flush:
			break;
		default:
			return NVME_SC_NS_WRITE_PROTECTED;
		}
	}

	return 0;
}

804 805 806 807 808 809 810 811 812 813
static u16 nvmet_parse_io_cmd(struct nvmet_req *req)
{
	struct nvme_command *cmd = req->cmd;
	u16 ret;

	ret = nvmet_check_ctrl_status(req, cmd);
	if (unlikely(ret))
		return ret;

	req->ns = nvmet_find_namespace(req->sq->ctrl, cmd->rw.nsid);
814 815
	if (unlikely(!req->ns)) {
		req->error_loc = offsetof(struct nvme_common_command, nsid);
816
		return NVME_SC_INVALID_NS | NVME_SC_DNR;
817
	}
818
	ret = nvmet_check_ana_state(req->port, req->ns);
819 820
	if (unlikely(ret)) {
		req->error_loc = offsetof(struct nvme_common_command, nsid);
821
		return ret;
822
	}
823
	ret = nvmet_io_cmd_check_access(req);
824 825
	if (unlikely(ret)) {
		req->error_loc = offsetof(struct nvme_common_command, nsid);
826
		return ret;
827
	}
828 829 830 831 832 833 834

	if (req->ns->file)
		return nvmet_file_parse_io_cmd(req);
	else
		return nvmet_bdev_parse_io_cmd(req);
}

835
bool nvmet_req_init(struct nvmet_req *req, struct nvmet_cq *cq,
836
		struct nvmet_sq *sq, const struct nvmet_fabrics_ops *ops)
837 838 839 840 841 842 843 844 845
{
	u8 flags = req->cmd->common.flags;
	u16 status;

	req->cq = cq;
	req->sq = sq;
	req->ops = ops;
	req->sg = NULL;
	req->sg_cnt = 0;
846
	req->transfer_len = 0;
847 848
	req->cqe->status = 0;
	req->cqe->sq_head = 0;
849
	req->ns = NULL;
850
	req->error_loc = NVMET_NO_ERROR_LOC;
851
	req->error_slba = 0;
852 853 854

	/* no support for fused commands yet */
	if (unlikely(flags & (NVME_CMD_FUSE_FIRST | NVME_CMD_FUSE_SECOND))) {
855
		req->error_loc = offsetof(struct nvme_common_command, flags);
856 857 858 859
		status = NVME_SC_INVALID_FIELD | NVME_SC_DNR;
		goto fail;
	}

860 861 862 863 864 865
	/*
	 * For fabrics, PSDT field shall describe metadata pointer (MPTR) that
	 * contains an address of a single contiguous physical buffer that is
	 * byte aligned.
	 */
	if (unlikely((flags & NVME_CMD_SGL_ALL) != NVME_CMD_SGL_METABUF)) {
866
		req->error_loc = offsetof(struct nvme_common_command, flags);
867 868 869 870 871 872 873 874 875
		status = NVME_SC_INVALID_FIELD | NVME_SC_DNR;
		goto fail;
	}

	if (unlikely(!req->sq->ctrl))
		/* will return an error for any Non-connect command: */
		status = nvmet_parse_connect_cmd(req);
	else if (likely(req->sq->qid != 0))
		status = nvmet_parse_io_cmd(req);
876
	else if (nvme_is_fabrics(req->cmd))
877 878 879 880 881 882 883 884 885 886 887 888 889 890
		status = nvmet_parse_fabrics_cmd(req);
	else if (req->sq->ctrl->subsys->type == NVME_NQN_DISC)
		status = nvmet_parse_discovery_cmd(req);
	else
		status = nvmet_parse_admin_cmd(req);

	if (status)
		goto fail;

	if (unlikely(!percpu_ref_tryget_live(&sq->ref))) {
		status = NVME_SC_INVALID_FIELD | NVME_SC_DNR;
		goto fail;
	}

891 892 893
	if (sq->ctrl)
		sq->ctrl->cmd_seen = true;

894 895 896 897 898 899 900 901
	return true;

fail:
	__nvmet_req_complete(req, status);
	return false;
}
EXPORT_SYMBOL_GPL(nvmet_req_init);

902 903 904
void nvmet_req_uninit(struct nvmet_req *req)
{
	percpu_ref_put(&req->sq->ref);
905 906
	if (req->ns)
		nvmet_put_namespace(req->ns);
907 908 909
}
EXPORT_SYMBOL_GPL(nvmet_req_uninit);

910 911
void nvmet_req_execute(struct nvmet_req *req)
{
912 913
	if (unlikely(req->data_len != req->transfer_len)) {
		req->error_loc = offsetof(struct nvme_common_command, dptr);
914
		nvmet_req_complete(req, NVME_SC_SGL_INVALID_DATA | NVME_SC_DNR);
915
	} else
916 917 918 919
		req->execute(req);
}
EXPORT_SYMBOL_GPL(nvmet_req_execute);

920 921
int nvmet_req_alloc_sgl(struct nvmet_req *req)
{
922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944
	struct pci_dev *p2p_dev = NULL;

	if (IS_ENABLED(CONFIG_PCI_P2PDMA)) {
		if (req->sq->ctrl && req->ns)
			p2p_dev = radix_tree_lookup(&req->sq->ctrl->p2p_ns_map,
						    req->ns->nsid);

		req->p2p_dev = NULL;
		if (req->sq->qid && p2p_dev) {
			req->sg = pci_p2pmem_alloc_sgl(p2p_dev, &req->sg_cnt,
						       req->transfer_len);
			if (req->sg) {
				req->p2p_dev = p2p_dev;
				return 0;
			}
		}

		/*
		 * If no P2P memory was available we fallback to using
		 * regular memory
		 */
	}

945 946 947 948 949 950 951 952 953 954
	req->sg = sgl_alloc(req->transfer_len, GFP_KERNEL, &req->sg_cnt);
	if (!req->sg)
		return -ENOMEM;

	return 0;
}
EXPORT_SYMBOL_GPL(nvmet_req_alloc_sgl);

void nvmet_req_free_sgl(struct nvmet_req *req)
{
955 956 957 958 959
	if (req->p2p_dev)
		pci_p2pmem_free_sgl(req->p2p_dev, req->sg);
	else
		sgl_free(req->sg);

960 961 962 963 964
	req->sg = NULL;
	req->sg_cnt = 0;
}
EXPORT_SYMBOL_GPL(nvmet_req_free_sgl);

965 966
static inline bool nvmet_cc_en(u32 cc)
{
967
	return (cc >> NVME_CC_EN_SHIFT) & 0x1;
968 969 970 971
}

static inline u8 nvmet_cc_css(u32 cc)
{
972
	return (cc >> NVME_CC_CSS_SHIFT) & 0x7;
973 974 975 976
}

static inline u8 nvmet_cc_mps(u32 cc)
{
977
	return (cc >> NVME_CC_MPS_SHIFT) & 0xf;
978 979 980 981
}

static inline u8 nvmet_cc_ams(u32 cc)
{
982
	return (cc >> NVME_CC_AMS_SHIFT) & 0x7;
983 984 985 986
}

static inline u8 nvmet_cc_shn(u32 cc)
{
987
	return (cc >> NVME_CC_SHN_SHIFT) & 0x3;
988 989 990 991
}

static inline u8 nvmet_cc_iosqes(u32 cc)
{
992
	return (cc >> NVME_CC_IOSQES_SHIFT) & 0xf;
993 994 995 996
}

static inline u8 nvmet_cc_iocqes(u32 cc)
{
997
	return (cc >> NVME_CC_IOCQES_SHIFT) & 0xf;
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
}

static void nvmet_start_ctrl(struct nvmet_ctrl *ctrl)
{
	lockdep_assert_held(&ctrl->lock);

	if (nvmet_cc_iosqes(ctrl->cc) != NVME_NVM_IOSQES ||
	    nvmet_cc_iocqes(ctrl->cc) != NVME_NVM_IOCQES ||
	    nvmet_cc_mps(ctrl->cc) != 0 ||
	    nvmet_cc_ams(ctrl->cc) != 0 ||
	    nvmet_cc_css(ctrl->cc) != 0) {
		ctrl->csts = NVME_CSTS_CFS;
		return;
	}

	ctrl->csts = NVME_CSTS_RDY;
1014 1015 1016 1017 1018 1019 1020 1021

	/*
	 * Controllers that are not yet enabled should not really enforce the
	 * keep alive timeout, but we still want to track a timeout and cleanup
	 * in case a host died before it enabled the controller.  Hence, simply
	 * reset the keep alive timer when the controller is enabled.
	 */
	mod_delayed_work(system_wq, &ctrl->ka_work, ctrl->kato * HZ);
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
}

static void nvmet_clear_ctrl(struct nvmet_ctrl *ctrl)
{
	lockdep_assert_held(&ctrl->lock);

	/* XXX: tear down queues? */
	ctrl->csts &= ~NVME_CSTS_RDY;
	ctrl->cc = 0;
}

void nvmet_update_cc(struct nvmet_ctrl *ctrl, u32 new)
{
	u32 old;

	mutex_lock(&ctrl->lock);
	old = ctrl->cc;
	ctrl->cc = new;

	if (nvmet_cc_en(new) && !nvmet_cc_en(old))
		nvmet_start_ctrl(ctrl);
	if (!nvmet_cc_en(new) && nvmet_cc_en(old))
		nvmet_clear_ctrl(ctrl);
	if (nvmet_cc_shn(new) && !nvmet_cc_shn(old)) {
		nvmet_clear_ctrl(ctrl);
		ctrl->csts |= NVME_CSTS_SHST_CMPLT;
	}
	if (!nvmet_cc_shn(new) && nvmet_cc_shn(old))
		ctrl->csts &= ~NVME_CSTS_SHST_CMPLT;
	mutex_unlock(&ctrl->lock);
}

static void nvmet_init_cap(struct nvmet_ctrl *ctrl)
{
	/* command sets supported: NVMe command set: */
	ctrl->cap = (1ULL << 37);
	/* CC.EN timeout in 500msec units: */
	ctrl->cap |= (15ULL << 24);
	/* maximum queue entries supported: */
	ctrl->cap |= NVMET_QUEUE_SIZE - 1;
}

u16 nvmet_ctrl_find_get(const char *subsysnqn, const char *hostnqn, u16 cntlid,
		struct nvmet_req *req, struct nvmet_ctrl **ret)
{
	struct nvmet_subsys *subsys;
	struct nvmet_ctrl *ctrl;
	u16 status = 0;

	subsys = nvmet_find_get_subsys(req->port, subsysnqn);
	if (!subsys) {
		pr_warn("connect request for invalid subsystem %s!\n",
			subsysnqn);
1075
		req->cqe->result.u32 = IPO_IATTR_CONNECT_DATA(subsysnqn);
1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
		return NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
	}

	mutex_lock(&subsys->lock);
	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) {
		if (ctrl->cntlid == cntlid) {
			if (strncmp(hostnqn, ctrl->hostnqn, NVMF_NQN_SIZE)) {
				pr_warn("hostnqn mismatch.\n");
				continue;
			}
			if (!kref_get_unless_zero(&ctrl->ref))
				continue;

			*ret = ctrl;
			goto out;
		}
	}

	pr_warn("could not find controller %d for subsys %s / host %s\n",
		cntlid, subsysnqn, hostnqn);
1096
	req->cqe->result.u32 = IPO_IATTR_CONNECT_DATA(cntlid);
1097 1098 1099 1100 1101 1102 1103 1104
	status = NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;

out:
	mutex_unlock(&subsys->lock);
	nvmet_subsys_put(subsys);
	return status;
}

1105 1106 1107
u16 nvmet_check_ctrl_status(struct nvmet_req *req, struct nvme_command *cmd)
{
	if (unlikely(!(req->sq->ctrl->cc & NVME_CC_ENABLE))) {
1108
		pr_err("got cmd %d while CC.EN == 0 on qid = %d\n",
1109 1110 1111 1112 1113
		       cmd->common.opcode, req->sq->qid);
		return NVME_SC_CMD_SEQ_ERROR | NVME_SC_DNR;
	}

	if (unlikely(!(req->sq->ctrl->csts & NVME_CSTS_RDY))) {
1114
		pr_err("got cmd %d while CSTS.RDY == 0 on qid = %d\n",
1115 1116 1117 1118 1119 1120
		       cmd->common.opcode, req->sq->qid);
		return NVME_SC_CMD_SEQ_ERROR | NVME_SC_DNR;
	}
	return 0;
}

1121
bool nvmet_host_allowed(struct nvmet_subsys *subsys, const char *hostnqn)
1122 1123 1124
{
	struct nvmet_host_link *p;

1125 1126
	lockdep_assert_held(&nvmet_config_sem);

1127 1128 1129
	if (subsys->allow_any_host)
		return true;

1130 1131 1132
	if (subsys->type == NVME_NQN_DISC) /* allow all access to disc subsys */
		return true;

1133 1134 1135 1136 1137 1138 1139 1140
	list_for_each_entry(p, &subsys->hosts, entry) {
		if (!strcmp(nvmet_host_name(p->host), hostnqn))
			return true;
	}

	return false;
}

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
/*
 * Note: ctrl->subsys->lock should be held when calling this function
 */
static void nvmet_setup_p2p_ns_map(struct nvmet_ctrl *ctrl,
		struct nvmet_req *req)
{
	struct nvmet_ns *ns;

	if (!req->p2p_client)
		return;

	ctrl->p2p_client = get_device(req->p2p_client);

	list_for_each_entry_rcu(ns, &ctrl->subsys->namespaces, dev_link)
		nvmet_p2pmem_ns_add_p2p(ctrl, ns);
}

/*
 * Note: ctrl->subsys->lock should be held when calling this function
 */
static void nvmet_release_p2p_ns_map(struct nvmet_ctrl *ctrl)
{
	struct radix_tree_iter iter;
	void __rcu **slot;

	radix_tree_for_each_slot(slot, &ctrl->p2p_ns_map, &iter, 0)
		pci_dev_put(radix_tree_deref_slot(slot));

	put_device(ctrl->p2p_client);
}

1172 1173 1174 1175 1176 1177 1178 1179 1180
static void nvmet_fatal_error_handler(struct work_struct *work)
{
	struct nvmet_ctrl *ctrl =
			container_of(work, struct nvmet_ctrl, fatal_err_work);

	pr_err("ctrl %d fatal error occurred!\n", ctrl->cntlid);
	ctrl->ops->delete_ctrl(ctrl);
}

1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193
u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
		struct nvmet_req *req, u32 kato, struct nvmet_ctrl **ctrlp)
{
	struct nvmet_subsys *subsys;
	struct nvmet_ctrl *ctrl;
	int ret;
	u16 status;

	status = NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
	subsys = nvmet_find_get_subsys(req->port, subsysnqn);
	if (!subsys) {
		pr_warn("connect request for invalid subsystem %s!\n",
			subsysnqn);
1194
		req->cqe->result.u32 = IPO_IATTR_CONNECT_DATA(subsysnqn);
1195 1196 1197 1198 1199
		goto out;
	}

	status = NVME_SC_CONNECT_INVALID_PARAM | NVME_SC_DNR;
	down_read(&nvmet_config_sem);
1200
	if (!nvmet_host_allowed(subsys, hostnqn)) {
1201 1202
		pr_info("connect by host %s for subsystem %s not allowed\n",
			hostnqn, subsysnqn);
1203
		req->cqe->result.u32 = IPO_IATTR_CONNECT_DATA(hostnqn);
1204
		up_read(&nvmet_config_sem);
1205
		status = NVME_SC_CONNECT_INVALID_HOST | NVME_SC_DNR;
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
		goto out_put_subsystem;
	}
	up_read(&nvmet_config_sem);

	status = NVME_SC_INTERNAL;
	ctrl = kzalloc(sizeof(*ctrl), GFP_KERNEL);
	if (!ctrl)
		goto out_put_subsystem;
	mutex_init(&ctrl->lock);

	nvmet_init_cap(ctrl);

1218 1219
	ctrl->port = req->port;

1220 1221
	INIT_WORK(&ctrl->async_event_work, nvmet_async_event_work);
	INIT_LIST_HEAD(&ctrl->async_events);
1222
	INIT_RADIX_TREE(&ctrl->p2p_ns_map, GFP_KERNEL);
1223
	INIT_WORK(&ctrl->fatal_err_work, nvmet_fatal_error_handler);
1224 1225 1226 1227 1228 1229

	memcpy(ctrl->subsysnqn, subsysnqn, NVMF_NQN_SIZE);
	memcpy(ctrl->hostnqn, hostnqn, NVMF_NQN_SIZE);

	kref_init(&ctrl->ref);
	ctrl->subsys = subsys;
1230
	WRITE_ONCE(ctrl->aen_enabled, NVMET_AEN_CFG_OPTIONAL);
1231

1232 1233 1234 1235 1236
	ctrl->changed_ns_list = kmalloc_array(NVME_MAX_CHANGED_NAMESPACES,
			sizeof(__le32), GFP_KERNEL);
	if (!ctrl->changed_ns_list)
		goto out_free_ctrl;

1237 1238 1239 1240
	ctrl->cqs = kcalloc(subsys->max_qid + 1,
			sizeof(struct nvmet_cq *),
			GFP_KERNEL);
	if (!ctrl->cqs)
1241
		goto out_free_changed_ns_list;
1242 1243 1244 1245 1246 1247 1248

	ctrl->sqs = kcalloc(subsys->max_qid + 1,
			sizeof(struct nvmet_sq *),
			GFP_KERNEL);
	if (!ctrl->sqs)
		goto out_free_cqs;

1249
	ret = ida_simple_get(&cntlid_ida,
1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
			     NVME_CNTLID_MIN, NVME_CNTLID_MAX,
			     GFP_KERNEL);
	if (ret < 0) {
		status = NVME_SC_CONNECT_CTRL_BUSY | NVME_SC_DNR;
		goto out_free_sqs;
	}
	ctrl->cntlid = ret;

	ctrl->ops = req->ops;

1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
	/*
	 * Discovery controllers may use some arbitrary high value
	 * in order to cleanup stale discovery sessions
	 */
	if ((ctrl->subsys->type == NVME_NQN_DISC) && !kato)
		kato = NVMET_DISC_KATO_MS;

	/* keep-alive timeout in seconds */
	ctrl->kato = DIV_ROUND_UP(kato, 1000);

1270 1271 1272
	ctrl->err_counter = 0;
	spin_lock_init(&ctrl->error_lock);

1273 1274 1275 1276
	nvmet_start_keep_alive_timer(ctrl);

	mutex_lock(&subsys->lock);
	list_add_tail(&ctrl->subsys_entry, &subsys->ctrls);
1277
	nvmet_setup_p2p_ns_map(ctrl, req);
1278 1279 1280 1281 1282 1283 1284 1285 1286
	mutex_unlock(&subsys->lock);

	*ctrlp = ctrl;
	return 0;

out_free_sqs:
	kfree(ctrl->sqs);
out_free_cqs:
	kfree(ctrl->cqs);
1287 1288
out_free_changed_ns_list:
	kfree(ctrl->changed_ns_list);
1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302
out_free_ctrl:
	kfree(ctrl);
out_put_subsystem:
	nvmet_subsys_put(subsys);
out:
	return status;
}

static void nvmet_ctrl_free(struct kref *ref)
{
	struct nvmet_ctrl *ctrl = container_of(ref, struct nvmet_ctrl, ref);
	struct nvmet_subsys *subsys = ctrl->subsys;

	mutex_lock(&subsys->lock);
1303
	nvmet_release_p2p_ns_map(ctrl);
1304 1305 1306
	list_del(&ctrl->subsys_entry);
	mutex_unlock(&subsys->lock);

1307 1308
	nvmet_stop_keep_alive_timer(ctrl);

1309 1310 1311
	flush_work(&ctrl->async_event_work);
	cancel_work_sync(&ctrl->fatal_err_work);

1312
	ida_simple_remove(&cntlid_ida, ctrl->cntlid);
1313 1314 1315

	kfree(ctrl->sqs);
	kfree(ctrl->cqs);
1316
	kfree(ctrl->changed_ns_list);
1317
	kfree(ctrl);
1318 1319

	nvmet_subsys_put(subsys);
1320 1321 1322 1323 1324 1325 1326 1327 1328
}

void nvmet_ctrl_put(struct nvmet_ctrl *ctrl)
{
	kref_put(&ctrl->ref, nvmet_ctrl_free);
}

void nvmet_ctrl_fatal_error(struct nvmet_ctrl *ctrl)
{
1329 1330 1331 1332 1333 1334
	mutex_lock(&ctrl->lock);
	if (!(ctrl->csts & NVME_CSTS_CFS)) {
		ctrl->csts |= NVME_CSTS_CFS;
		schedule_work(&ctrl->fatal_err_work);
	}
	mutex_unlock(&ctrl->lock);
1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345
}
EXPORT_SYMBOL_GPL(nvmet_ctrl_fatal_error);

static struct nvmet_subsys *nvmet_find_get_subsys(struct nvmet_port *port,
		const char *subsysnqn)
{
	struct nvmet_subsys_link *p;

	if (!port)
		return NULL;

1346
	if (!strcmp(NVME_DISC_SUBSYS_NAME, subsysnqn)) {
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372
		if (!kref_get_unless_zero(&nvmet_disc_subsys->ref))
			return NULL;
		return nvmet_disc_subsys;
	}

	down_read(&nvmet_config_sem);
	list_for_each_entry(p, &port->subsystems, entry) {
		if (!strncmp(p->subsys->subsysnqn, subsysnqn,
				NVMF_NQN_SIZE)) {
			if (!kref_get_unless_zero(&p->subsys->ref))
				break;
			up_read(&nvmet_config_sem);
			return p->subsys;
		}
	}
	up_read(&nvmet_config_sem);
	return NULL;
}

struct nvmet_subsys *nvmet_subsys_alloc(const char *subsysnqn,
		enum nvme_subsys_type type)
{
	struct nvmet_subsys *subsys;

	subsys = kzalloc(sizeof(*subsys), GFP_KERNEL);
	if (!subsys)
1373
		return ERR_PTR(-ENOMEM);
1374

1375
	subsys->ver = NVME_VS(1, 3, 0); /* NVMe 1.3.0 */
1376 1377
	/* generate a random serial number as our controllers are ephemeral: */
	get_random_bytes(&subsys->serial, sizeof(subsys->serial));
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388

	switch (type) {
	case NVME_NQN_NVME:
		subsys->max_qid = NVMET_NR_QUEUES;
		break;
	case NVME_NQN_DISC:
		subsys->max_qid = 0;
		break;
	default:
		pr_err("%s: Unknown Subsystem type - %d\n", __func__, type);
		kfree(subsys);
1389
		return ERR_PTR(-EINVAL);
1390 1391 1392 1393
	}
	subsys->type = type;
	subsys->subsysnqn = kstrndup(subsysnqn, NVMF_NQN_SIZE,
			GFP_KERNEL);
1394
	if (!subsys->subsysnqn) {
1395
		kfree(subsys);
1396
		return ERR_PTR(-ENOMEM);
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419
	}

	kref_init(&subsys->ref);

	mutex_init(&subsys->lock);
	INIT_LIST_HEAD(&subsys->namespaces);
	INIT_LIST_HEAD(&subsys->ctrls);
	INIT_LIST_HEAD(&subsys->hosts);

	return subsys;
}

static void nvmet_subsys_free(struct kref *ref)
{
	struct nvmet_subsys *subsys =
		container_of(ref, struct nvmet_subsys, ref);

	WARN_ON_ONCE(!list_empty(&subsys->namespaces));

	kfree(subsys->subsysnqn);
	kfree(subsys);
}

1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
void nvmet_subsys_del_ctrls(struct nvmet_subsys *subsys)
{
	struct nvmet_ctrl *ctrl;

	mutex_lock(&subsys->lock);
	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry)
		ctrl->ops->delete_ctrl(ctrl);
	mutex_unlock(&subsys->lock);
}

1430 1431 1432 1433 1434 1435 1436 1437 1438
void nvmet_subsys_put(struct nvmet_subsys *subsys)
{
	kref_put(&subsys->ref, nvmet_subsys_free);
}

static int __init nvmet_init(void)
{
	int error;

1439 1440
	nvmet_ana_group_enabled[NVMET_DEFAULT_ANA_GRPID] = 1;

1441 1442 1443 1444 1445 1446
	buffered_io_wq = alloc_workqueue("nvmet-buffered-io-wq",
			WQ_MEM_RECLAIM, 0);
	if (!buffered_io_wq) {
		error = -ENOMEM;
		goto out;
	}
1447

1448 1449
	error = nvmet_init_discovery();
	if (error)
1450
		goto out_free_work_queue;
1451 1452 1453 1454 1455 1456 1457 1458

	error = nvmet_init_configfs();
	if (error)
		goto out_exit_discovery;
	return 0;

out_exit_discovery:
	nvmet_exit_discovery();
1459 1460
out_free_work_queue:
	destroy_workqueue(buffered_io_wq);
1461 1462 1463 1464 1465 1466 1467 1468
out:
	return error;
}

static void __exit nvmet_exit(void)
{
	nvmet_exit_configfs();
	nvmet_exit_discovery();
1469
	ida_destroy(&cntlid_ida);
1470
	destroy_workqueue(buffered_io_wq);
1471 1472 1473 1474 1475 1476 1477 1478 1479

	BUILD_BUG_ON(sizeof(struct nvmf_disc_rsp_page_entry) != 1024);
	BUILD_BUG_ON(sizeof(struct nvmf_disc_rsp_page_hdr) != 1024);
}

module_init(nvmet_init);
module_exit(nvmet_exit);

MODULE_LICENSE("GPL v2");