pblk-core.c 48.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
/*
 * Copyright (C) 2016 CNEX Labs
 * Initial release: Javier Gonzalez <javier@cnexlabs.com>
 *                  Matias Bjorling <matias@cnexlabs.com>
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License version
 * 2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * pblk-core.c - pblk's core functionality
 *
 */

#include "pblk.h"

21 22 23 24 25 26 27 28 29 30 31 32 33 34
static void pblk_line_mark_bb(struct work_struct *work)
{
	struct pblk_line_ws *line_ws = container_of(work, struct pblk_line_ws,
									ws);
	struct pblk *pblk = line_ws->pblk;
	struct nvm_tgt_dev *dev = pblk->dev;
	struct ppa_addr *ppa = line_ws->priv;
	int ret;

	ret = nvm_set_tgt_bb_tbl(dev, ppa, 1, NVM_BLK_T_GRWN_BAD);
	if (ret) {
		struct pblk_line *line;
		int pos;

35 36
		line = &pblk->lines[pblk_ppa_to_line(*ppa)];
		pos = pblk_ppa_to_pos(&dev->geo, *ppa);
37 38 39 40 41 42

		pr_err("pblk: failed to mark bb, line:%d, pos:%d\n",
				line->id, pos);
	}

	kfree(ppa);
43
	mempool_free(line_ws, &pblk->gen_ws_pool);
44 45
}

46
static void pblk_mark_bb(struct pblk *pblk, struct pblk_line *line,
47
			 struct ppa_addr ppa_addr)
48 49 50
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
51 52
	struct ppa_addr *ppa;
	int pos = pblk_ppa_to_pos(geo, ppa_addr);
53 54 55 56

	pr_debug("pblk: erase failed: line:%d, pos:%d\n", line->id, pos);
	atomic_long_inc(&pblk->erase_failed);

57
	atomic_dec(&line->blk_in_line);
58 59 60 61
	if (test_and_set_bit(pos, line->blk_bitmap))
		pr_err("pblk: attempted to erase bb: line:%d, pos:%d\n",
							line->id, pos);

62 63 64 65 66 67 68 69 70
	/* Not necessary to mark bad blocks on 2.0 spec. */
	if (geo->version == NVM_OCSSD_SPEC_20)
		return;

	ppa = kmalloc(sizeof(struct ppa_addr), GFP_ATOMIC);
	if (!ppa)
		return;

	*ppa = ppa_addr;
71 72
	pblk_gen_run_ws(pblk, NULL, ppa, pblk_line_mark_bb,
						GFP_ATOMIC, pblk->bb_wq);
73 74 75 76
}

static void __pblk_end_io_erase(struct pblk *pblk, struct nvm_rq *rqd)
{
77 78 79
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct nvm_chk_meta *chunk;
80
	struct pblk_line *line;
81
	int pos;
82

83
	line = &pblk->lines[pblk_ppa_to_line(rqd->ppa_addr)];
84 85 86
	pos = pblk_ppa_to_pos(geo, rqd->ppa_addr);
	chunk = &line->chks[pos];

87 88 89
	atomic_dec(&line->left_seblks);

	if (rqd->error) {
90 91 92 93
		chunk->state = NVM_CHK_ST_OFFLINE;
		pblk_mark_bb(pblk, line, rqd->ppa_addr);
	} else {
		chunk->state = NVM_CHK_ST_FREE;
94
	}
95 96

	atomic_dec(&pblk->inflight_io);
97 98 99 100 101 102 103 104
}

/* Erase completion assumes that only one block is erased at the time */
static void pblk_end_io_erase(struct nvm_rq *rqd)
{
	struct pblk *pblk = rqd->private;

	__pblk_end_io_erase(pblk, rqd);
105
	mempool_free(rqd, &pblk->e_rq_pool);
106 107
}

108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
/*
 * Get information for all chunks from the device.
 *
 * The caller is responsible for freeing the returned structure
 */
struct nvm_chk_meta *pblk_chunk_get_info(struct pblk *pblk)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct nvm_chk_meta *meta;
	struct ppa_addr ppa;
	unsigned long len;
	int ret;

	ppa.ppa = 0;

	len = geo->all_chunks * sizeof(*meta);
	meta = kzalloc(len, GFP_KERNEL);
	if (!meta)
		return ERR_PTR(-ENOMEM);

	ret = nvm_get_chunk_meta(dev, meta, ppa, geo->all_chunks);
	if (ret) {
		kfree(meta);
		return ERR_PTR(-EIO);
	}

	return meta;
}

struct nvm_chk_meta *pblk_chunk_get_off(struct pblk *pblk,
					      struct nvm_chk_meta *meta,
					      struct ppa_addr ppa)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	int ch_off = ppa.m.grp * geo->num_chk * geo->num_lun;
	int lun_off = ppa.m.pu * geo->num_chk;
	int chk_off = ppa.m.chk;

	return meta + ch_off + lun_off + chk_off;
}

151 152
void __pblk_map_invalidate(struct pblk *pblk, struct pblk_line *line,
			   u64 paddr)
153 154 155 156 157 158 159 160 161
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct list_head *move_list = NULL;

	/* Lines being reclaimed (GC'ed) cannot be invalidated. Before the L2P
	 * table is modified with reclaimed sectors, a check is done to endure
	 * that newer updates are not overwritten.
	 */
	spin_lock(&line->lock);
162
	WARN_ON(line->state == PBLK_LINESTATE_FREE);
163 164 165 166 167 168

	if (test_and_set_bit(paddr, line->invalid_bitmap)) {
		WARN_ONCE(1, "pblk: double invalidate\n");
		spin_unlock(&line->lock);
		return;
	}
169
	le32_add_cpu(line->vsc, -1);
170 171 172 173 174 175 176 177 178

	if (line->state == PBLK_LINESTATE_CLOSED)
		move_list = pblk_line_gc_list(pblk, line);
	spin_unlock(&line->lock);

	if (move_list) {
		spin_lock(&l_mg->gc_lock);
		spin_lock(&line->lock);
		/* Prevent moving a line that has just been chosen for GC */
179
		if (line->state == PBLK_LINESTATE_GC) {
180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202
			spin_unlock(&line->lock);
			spin_unlock(&l_mg->gc_lock);
			return;
		}
		spin_unlock(&line->lock);

		list_move_tail(&line->list, move_list);
		spin_unlock(&l_mg->gc_lock);
	}
}

void pblk_map_invalidate(struct pblk *pblk, struct ppa_addr ppa)
{
	struct pblk_line *line;
	u64 paddr;
	int line_id;

#ifdef CONFIG_NVM_DEBUG
	/* Callers must ensure that the ppa points to a device address */
	BUG_ON(pblk_addr_in_cache(ppa));
	BUG_ON(pblk_ppa_empty(ppa));
#endif

203
	line_id = pblk_ppa_to_line(ppa);
204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
	line = &pblk->lines[line_id];
	paddr = pblk_dev_ppa_to_line_addr(pblk, ppa);

	__pblk_map_invalidate(pblk, line, paddr);
}

static void pblk_invalidate_range(struct pblk *pblk, sector_t slba,
				  unsigned int nr_secs)
{
	sector_t lba;

	spin_lock(&pblk->trans_lock);
	for (lba = slba; lba < slba + nr_secs; lba++) {
		struct ppa_addr ppa;

		ppa = pblk_trans_map_get(pblk, lba);

		if (!pblk_addr_in_cache(ppa) && !pblk_ppa_empty(ppa))
			pblk_map_invalidate(pblk, ppa);

		pblk_ppa_set_empty(&ppa);
		pblk_trans_map_set(pblk, lba, ppa);
	}
	spin_unlock(&pblk->trans_lock);
}

230 231
/* Caller must guarantee that the request is a valid type */
struct nvm_rq *pblk_alloc_rqd(struct pblk *pblk, int type)
232 233 234 235 236
{
	mempool_t *pool;
	struct nvm_rq *rqd;
	int rq_size;

237 238 239
	switch (type) {
	case PBLK_WRITE:
	case PBLK_WRITE_INT:
240
		pool = &pblk->w_rq_pool;
241
		rq_size = pblk_w_rq_size;
242 243
		break;
	case PBLK_READ:
244
		pool = &pblk->r_rq_pool;
245
		rq_size = pblk_g_rq_size;
246 247
		break;
	default:
248
		pool = &pblk->e_rq_pool;
249
		rq_size = pblk_g_rq_size;
250 251 252 253 254 255 256 257
	}

	rqd = mempool_alloc(pool, GFP_KERNEL);
	memset(rqd, 0, rq_size);

	return rqd;
}

258 259
/* Typically used on completion path. Cannot guarantee request consistency */
void pblk_free_rqd(struct pblk *pblk, struct nvm_rq *rqd, int type)
260
{
261
	struct nvm_tgt_dev *dev = pblk->dev;
262 263
	mempool_t *pool;

264 265 266 267
	switch (type) {
	case PBLK_WRITE:
		kfree(((struct pblk_c_ctx *)nvm_rq_to_pdu(rqd))->lun_bitmap);
	case PBLK_WRITE_INT:
268
		pool = &pblk->w_rq_pool;
269 270
		break;
	case PBLK_READ:
271
		pool = &pblk->r_rq_pool;
272 273
		break;
	case PBLK_ERASE:
274
		pool = &pblk->e_rq_pool;
275 276 277 278 279
		break;
	default:
		pr_err("pblk: trying to free unknown rqd type\n");
		return;
	}
280

281
	nvm_dev_dma_free(dev->parent, rqd->meta_list, rqd->dma_meta_list);
282 283 284 285 286 287 288 289 290 291 292 293 294
	mempool_free(rqd, pool);
}

void pblk_bio_free_pages(struct pblk *pblk, struct bio *bio, int off,
			 int nr_pages)
{
	struct bio_vec bv;
	int i;

	WARN_ON(off + nr_pages != bio->bi_vcnt);

	for (i = off; i < nr_pages + off; i++) {
		bv = bio->bi_io_vec[i];
295
		mempool_free(bv.bv_page, &pblk->page_bio_pool);
296 297 298 299 300 301 302 303 304 305 306
	}
}

int pblk_bio_add_pages(struct pblk *pblk, struct bio *bio, gfp_t flags,
		       int nr_pages)
{
	struct request_queue *q = pblk->dev->q;
	struct page *page;
	int i, ret;

	for (i = 0; i < nr_pages; i++) {
307
		page = mempool_alloc(&pblk->page_bio_pool, flags);
308 309 310 311

		ret = bio_add_pc_page(q, bio, page, PBLK_EXPOSED_PAGE_SIZE, 0);
		if (ret != PBLK_EXPOSED_PAGE_SIZE) {
			pr_err("pblk: could not add page to bio\n");
312
			mempool_free(page, &pblk->page_bio_pool);
313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
			goto err;
		}
	}

	return 0;
err:
	pblk_bio_free_pages(pblk, bio, 0, i - 1);
	return -1;
}

static void pblk_write_kick(struct pblk *pblk)
{
	wake_up_process(pblk->writer_ts);
	mod_timer(&pblk->wtimer, jiffies + msecs_to_jiffies(1000));
}

329
void pblk_write_timer_fn(struct timer_list *t)
330
{
331
	struct pblk *pblk = from_timer(pblk, t, wtimer);
332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351

	/* kick the write thread every tick to flush outstanding data */
	pblk_write_kick(pblk);
}

void pblk_write_should_kick(struct pblk *pblk)
{
	unsigned int secs_avail = pblk_rb_read_count(&pblk->rwb);

	if (secs_avail >= pblk->min_write_pgs)
		pblk_write_kick(pblk);
}

void pblk_end_io_sync(struct nvm_rq *rqd)
{
	struct completion *waiting = rqd->private;

	complete(waiting);
}

352
static void pblk_wait_for_meta(struct pblk *pblk)
353
{
354 355 356
	do {
		if (!atomic_read(&pblk->inflight_io))
			break;
357

358 359 360
		schedule();
	} while (1);
}
361

362 363 364 365
static void pblk_flush_writer(struct pblk *pblk)
{
	pblk_rb_flush(&pblk->rwb);
	do {
366
		if (!pblk_rb_sync_count(&pblk->rwb))
367
			break;
368

369
		pblk_write_kick(pblk);
370 371
		schedule();
	} while (1);
372 373 374 375 376 377 378
}

struct list_head *pblk_line_gc_list(struct pblk *pblk, struct pblk_line *line)
{
	struct pblk_line_meta *lm = &pblk->lm;
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct list_head *move_list = NULL;
379
	int vsc = le32_to_cpu(*line->vsc);
380

381 382
	lockdep_assert_held(&line->lock);

383
	if (!vsc) {
384 385 386 387
		if (line->gc_group != PBLK_LINEGC_FULL) {
			line->gc_group = PBLK_LINEGC_FULL;
			move_list = &l_mg->gc_full_list;
		}
388
	} else if (vsc < lm->high_thrs) {
389 390 391 392
		if (line->gc_group != PBLK_LINEGC_HIGH) {
			line->gc_group = PBLK_LINEGC_HIGH;
			move_list = &l_mg->gc_high_list;
		}
393
	} else if (vsc < lm->mid_thrs) {
394 395 396 397
		if (line->gc_group != PBLK_LINEGC_MID) {
			line->gc_group = PBLK_LINEGC_MID;
			move_list = &l_mg->gc_mid_list;
		}
398
	} else if (vsc < line->sec_in_line) {
399 400 401 402
		if (line->gc_group != PBLK_LINEGC_LOW) {
			line->gc_group = PBLK_LINEGC_LOW;
			move_list = &l_mg->gc_low_list;
		}
403
	} else if (vsc == line->sec_in_line) {
404 405 406 407 408 409 410 411 412
		if (line->gc_group != PBLK_LINEGC_EMPTY) {
			line->gc_group = PBLK_LINEGC_EMPTY;
			move_list = &l_mg->gc_empty_list;
		}
	} else {
		line->state = PBLK_LINESTATE_CORRUPT;
		line->gc_group = PBLK_LINEGC_NONE;
		move_list =  &l_mg->corrupt_list;
		pr_err("pblk: corrupted vsc for line %d, vsc:%d (%d/%d/%d)\n",
413
						line->id, vsc,
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
						line->sec_in_line,
						lm->high_thrs, lm->mid_thrs);
	}

	return move_list;
}

void pblk_discard(struct pblk *pblk, struct bio *bio)
{
	sector_t slba = pblk_get_lba(bio);
	sector_t nr_secs = pblk_get_secs(bio);

	pblk_invalidate_range(pblk, slba, nr_secs);
}

void pblk_log_write_err(struct pblk *pblk, struct nvm_rq *rqd)
{
	atomic_long_inc(&pblk->write_failed);
#ifdef CONFIG_NVM_DEBUG
	pblk_print_failed_rqd(pblk, rqd, rqd->error);
#endif
}

void pblk_log_read_err(struct pblk *pblk, struct nvm_rq *rqd)
{
	/* Empty page read is not necessarily an error (e.g., L2P recovery) */
	if (rqd->error == NVM_RSP_ERR_EMPTYPAGE) {
		atomic_long_inc(&pblk->read_empty);
		return;
	}

	switch (rqd->error) {
	case NVM_RSP_WARN_HIGHECC:
		atomic_long_inc(&pblk->read_high_ecc);
		break;
	case NVM_RSP_ERR_FAILECC:
	case NVM_RSP_ERR_FAILCRC:
		atomic_long_inc(&pblk->read_failed);
		break;
	default:
		pr_err("pblk: unknown read error:%d\n", rqd->error);
	}
#ifdef CONFIG_NVM_DEBUG
	pblk_print_failed_rqd(pblk, rqd, rqd->error);
#endif
}

461 462 463 464 465
void pblk_set_sec_per_write(struct pblk *pblk, int sec_per_write)
{
	pblk->sec_per_write = sec_per_write;
}

466 467 468 469 470
int pblk_submit_io(struct pblk *pblk, struct nvm_rq *rqd)
{
	struct nvm_tgt_dev *dev = pblk->dev;

#ifdef CONFIG_NVM_DEBUG
471
	int ret;
472

473 474 475 476
	ret = pblk_check_io(pblk, rqd);
	if (ret)
		return ret;
#endif
477

478
	atomic_inc(&pblk->inflight_io);
479

480 481
	return nvm_submit_io(dev, rqd);
}
482

483 484 485 486 487 488 489 490 491 492
int pblk_submit_io_sync(struct pblk *pblk, struct nvm_rq *rqd)
{
	struct nvm_tgt_dev *dev = pblk->dev;

#ifdef CONFIG_NVM_DEBUG
	int ret;

	ret = pblk_check_io(pblk, rqd);
	if (ret)
		return ret;
493
#endif
494 495 496

	atomic_inc(&pblk->inflight_io);

497
	return nvm_submit_io_sync(dev, rqd);
498 499
}

500 501 502 503 504
static void pblk_bio_map_addr_endio(struct bio *bio)
{
	bio_put(bio);
}

505 506
struct bio *pblk_bio_map_addr(struct pblk *pblk, void *data,
			      unsigned int nr_secs, unsigned int len,
507
			      int alloc_type, gfp_t gfp_mask)
508 509 510 511 512 513 514
{
	struct nvm_tgt_dev *dev = pblk->dev;
	void *kaddr = data;
	struct page *page;
	struct bio *bio;
	int i, ret;

515
	if (alloc_type == PBLK_KMALLOC_META)
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540
		return bio_map_kern(dev->q, kaddr, len, gfp_mask);

	bio = bio_kmalloc(gfp_mask, nr_secs);
	if (!bio)
		return ERR_PTR(-ENOMEM);

	for (i = 0; i < nr_secs; i++) {
		page = vmalloc_to_page(kaddr);
		if (!page) {
			pr_err("pblk: could not map vmalloc bio\n");
			bio_put(bio);
			bio = ERR_PTR(-ENOMEM);
			goto out;
		}

		ret = bio_add_pc_page(dev->q, bio, page, PAGE_SIZE, 0);
		if (ret != PAGE_SIZE) {
			pr_err("pblk: could not add page to bio\n");
			bio_put(bio);
			bio = ERR_PTR(-ENOMEM);
			goto out;
		}

		kaddr += PAGE_SIZE;
	}
541 542

	bio->bi_end_io = pblk_bio_map_addr_endio;
543 544 545 546 547 548 549
out:
	return bio;
}

int pblk_calc_secs(struct pblk *pblk, unsigned long secs_avail,
		   unsigned long secs_to_flush)
{
550
	int max = pblk->sec_per_write;
551 552 553 554 555 556 557 558 559 560 561 562 563
	int min = pblk->min_write_pgs;
	int secs_to_sync = 0;

	if (secs_avail >= max)
		secs_to_sync = max;
	else if (secs_avail >= min)
		secs_to_sync = min * (secs_avail / min);
	else if (secs_to_flush)
		secs_to_sync = min;

	return secs_to_sync;
}

564 565 566 567 568
void pblk_dealloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs)
{
	u64 addr;
	int i;

569
	spin_lock(&line->lock);
570 571 572 573 574 575
	addr = find_next_zero_bit(line->map_bitmap,
					pblk->lm.sec_per_line, line->cur_sec);
	line->cur_sec = addr - nr_secs;

	for (i = 0; i < nr_secs; i++, line->cur_sec--)
		WARN_ON(!test_and_clear_bit(line->cur_sec, line->map_bitmap));
576
	spin_unlock(&line->lock);
577 578 579
}

u64 __pblk_alloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs)
580 581 582 583
{
	u64 addr;
	int i;

584 585
	lockdep_assert_held(&line->lock);

586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615
	/* logic error: ppa out-of-bounds. Prevent generating bad address */
	if (line->cur_sec + nr_secs > pblk->lm.sec_per_line) {
		WARN(1, "pblk: page allocation out of bounds\n");
		nr_secs = pblk->lm.sec_per_line - line->cur_sec;
	}

	line->cur_sec = addr = find_next_zero_bit(line->map_bitmap,
					pblk->lm.sec_per_line, line->cur_sec);
	for (i = 0; i < nr_secs; i++, line->cur_sec++)
		WARN_ON(test_and_set_bit(line->cur_sec, line->map_bitmap));

	return addr;
}

u64 pblk_alloc_page(struct pblk *pblk, struct pblk_line *line, int nr_secs)
{
	u64 addr;

	/* Lock needed in case a write fails and a recovery needs to remap
	 * failed write buffer entries
	 */
	spin_lock(&line->lock);
	addr = __pblk_alloc_page(pblk, line, nr_secs);
	line->left_msecs -= nr_secs;
	WARN(line->left_msecs < 0, "pblk: page allocation out of bounds\n");
	spin_unlock(&line->lock);

	return addr;
}

616 617 618 619 620 621 622 623 624 625 626 627
u64 pblk_lookup_page(struct pblk *pblk, struct pblk_line *line)
{
	u64 paddr;

	spin_lock(&line->lock);
	paddr = find_next_zero_bit(line->map_bitmap,
					pblk->lm.sec_per_line, line->cur_sec);
	spin_unlock(&line->lock);

	return paddr;
}

628 629 630 631 632
/*
 * Submit emeta to one LUN in the raid line at the time to avoid a deadlock when
 * taking the per LUN semaphore.
 */
static int pblk_line_submit_emeta_io(struct pblk *pblk, struct pblk_line *line,
633
				     void *emeta_buf, u64 paddr, int dir)
634 635 636
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
637
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
638
	struct pblk_line_meta *lm = &pblk->lm;
639
	void *ppa_list, *meta_list;
640 641
	struct bio *bio;
	struct nvm_rq rqd;
642
	dma_addr_t dma_ppa_list, dma_meta_list;
643
	int min = pblk->min_write_pgs;
644
	int left_ppas = lm->emeta_sec[0];
645 646 647 648 649 650
	int id = line->id;
	int rq_ppas, rq_len;
	int cmd_op, bio_op;
	int i, j;
	int ret;

651
	if (dir == PBLK_WRITE) {
652 653
		bio_op = REQ_OP_WRITE;
		cmd_op = NVM_OP_PWRITE;
654
	} else if (dir == PBLK_READ) {
655 656 657 658 659
		bio_op = REQ_OP_READ;
		cmd_op = NVM_OP_PREAD;
	} else
		return -EINVAL;

660 661 662
	meta_list = nvm_dev_dma_alloc(dev->parent, GFP_KERNEL,
							&dma_meta_list);
	if (!meta_list)
663 664
		return -ENOMEM;

665 666 667
	ppa_list = meta_list + pblk_dma_meta_size;
	dma_ppa_list = dma_meta_list + pblk_dma_meta_size;

668 669 670 671
next_rq:
	memset(&rqd, 0, sizeof(struct nvm_rq));

	rq_ppas = pblk_calc_secs(pblk, left_ppas, 0);
672
	rq_len = rq_ppas * geo->csecs;
673

674 675
	bio = pblk_bio_map_addr(pblk, emeta_buf, rq_ppas, rq_len,
					l_mg->emeta_alloc_type, GFP_KERNEL);
676 677 678 679 680 681 682 683 684
	if (IS_ERR(bio)) {
		ret = PTR_ERR(bio);
		goto free_rqd_dma;
	}

	bio->bi_iter.bi_sector = 0; /* internal bio */
	bio_set_op_attrs(bio, bio_op, 0);

	rqd.bio = bio;
685
	rqd.meta_list = meta_list;
686
	rqd.ppa_list = ppa_list;
687
	rqd.dma_meta_list = dma_meta_list;
688
	rqd.dma_ppa_list = dma_ppa_list;
689 690
	rqd.opcode = cmd_op;
	rqd.nr_ppas = rq_ppas;
691

692
	if (dir == PBLK_WRITE) {
693 694
		struct pblk_sec_meta *meta_list = rqd.meta_list;

695
		rqd.flags = pblk_set_progr_mode(pblk, PBLK_WRITE);
696 697 698 699
		for (i = 0; i < rqd.nr_ppas; ) {
			spin_lock(&line->lock);
			paddr = __pblk_alloc_page(pblk, line, min);
			spin_unlock(&line->lock);
700 701
			for (j = 0; j < min; j++, i++, paddr++) {
				meta_list[i].lba = cpu_to_le64(ADDR_EMPTY);
702 703
				rqd.ppa_list[i] =
					addr_to_gen_ppa(pblk, paddr, id);
704
			}
705 706 707 708
		}
	} else {
		for (i = 0; i < rqd.nr_ppas; ) {
			struct ppa_addr ppa = addr_to_gen_ppa(pblk, paddr, id);
709
			int pos = pblk_ppa_to_pos(geo, ppa);
710 711 712 713 714
			int read_type = PBLK_READ_RANDOM;

			if (pblk_io_aligned(pblk, rq_ppas))
				read_type = PBLK_READ_SEQUENTIAL;
			rqd.flags = pblk_set_read_mode(pblk, read_type);
715 716 717 718 719 720 721 722 723 724 725 726

			while (test_bit(pos, line->blk_bitmap)) {
				paddr += min;
				if (pblk_boundary_paddr_checks(pblk, paddr)) {
					pr_err("pblk: corrupt emeta line:%d\n",
								line->id);
					bio_put(bio);
					ret = -EINTR;
					goto free_rqd_dma;
				}

				ppa = addr_to_gen_ppa(pblk, paddr, id);
727
				pos = pblk_ppa_to_pos(geo, ppa);
728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
			}

			if (pblk_boundary_paddr_checks(pblk, paddr + min)) {
				pr_err("pblk: corrupt emeta line:%d\n",
								line->id);
				bio_put(bio);
				ret = -EINTR;
				goto free_rqd_dma;
			}

			for (j = 0; j < min; j++, i++, paddr++)
				rqd.ppa_list[i] =
					addr_to_gen_ppa(pblk, paddr, line->id);
		}
	}

744
	ret = pblk_submit_io_sync(pblk, &rqd);
745 746 747 748 749 750
	if (ret) {
		pr_err("pblk: emeta I/O submission failed: %d\n", ret);
		bio_put(bio);
		goto free_rqd_dma;
	}

751
	atomic_dec(&pblk->inflight_io);
752 753

	if (rqd.error) {
754
		if (dir == PBLK_WRITE)
755 756 757 758 759
			pblk_log_write_err(pblk, &rqd);
		else
			pblk_log_read_err(pblk, &rqd);
	}

760
	emeta_buf += rq_len;
761 762 763 764
	left_ppas -= rq_ppas;
	if (left_ppas)
		goto next_rq;
free_rqd_dma:
765
	nvm_dev_dma_free(dev->parent, rqd.meta_list, rqd.dma_meta_list);
766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
	return ret;
}

u64 pblk_line_smeta_start(struct pblk *pblk, struct pblk_line *line)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct pblk_line_meta *lm = &pblk->lm;
	int bit;

	/* This usually only happens on bad lines */
	bit = find_first_zero_bit(line->blk_bitmap, lm->blk_per_line);
	if (bit >= lm->blk_per_line)
		return -1;

781
	return bit * geo->ws_opt;
782 783 784 785 786 787 788 789 790 791 792 793 794 795
}

static int pblk_line_submit_smeta_io(struct pblk *pblk, struct pblk_line *line,
				     u64 paddr, int dir)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct pblk_line_meta *lm = &pblk->lm;
	struct bio *bio;
	struct nvm_rq rqd;
	__le64 *lba_list = NULL;
	int i, ret;
	int cmd_op, bio_op;
	int flags;

796
	if (dir == PBLK_WRITE) {
797 798
		bio_op = REQ_OP_WRITE;
		cmd_op = NVM_OP_PWRITE;
799
		flags = pblk_set_progr_mode(pblk, PBLK_WRITE);
800
		lba_list = emeta_to_lbas(pblk, line->emeta->buf);
801
	} else if (dir == PBLK_READ_RECOV || dir == PBLK_READ) {
802 803
		bio_op = REQ_OP_READ;
		cmd_op = NVM_OP_PREAD;
804
		flags = pblk_set_read_mode(pblk, PBLK_READ_SEQUENTIAL);
805 806 807 808 809
	} else
		return -EINVAL;

	memset(&rqd, 0, sizeof(struct nvm_rq));

810 811 812
	rqd.meta_list = nvm_dev_dma_alloc(dev->parent, GFP_KERNEL,
							&rqd.dma_meta_list);
	if (!rqd.meta_list)
813 814
		return -ENOMEM;

815 816 817
	rqd.ppa_list = rqd.meta_list + pblk_dma_meta_size;
	rqd.dma_ppa_list = rqd.dma_meta_list + pblk_dma_meta_size;

818 819 820 821 822 823 824 825 826 827 828 829 830 831 832
	bio = bio_map_kern(dev->q, line->smeta, lm->smeta_len, GFP_KERNEL);
	if (IS_ERR(bio)) {
		ret = PTR_ERR(bio);
		goto free_ppa_list;
	}

	bio->bi_iter.bi_sector = 0; /* internal bio */
	bio_set_op_attrs(bio, bio_op, 0);

	rqd.bio = bio;
	rqd.opcode = cmd_op;
	rqd.flags = flags;
	rqd.nr_ppas = lm->smeta_sec;

	for (i = 0; i < lm->smeta_sec; i++, paddr++) {
833 834
		struct pblk_sec_meta *meta_list = rqd.meta_list;

835
		rqd.ppa_list[i] = addr_to_gen_ppa(pblk, paddr, line->id);
836

837
		if (dir == PBLK_WRITE) {
838
			__le64 addr_empty = cpu_to_le64(ADDR_EMPTY);
839 840 841

			meta_list[i].lba = lba_list[paddr] = addr_empty;
		}
842 843 844 845 846 847 848
	}

	/*
	 * This I/O is sent by the write thread when a line is replace. Since
	 * the write thread is the only one sending write and erase commands,
	 * there is no need to take the LUN semaphore.
	 */
849
	ret = pblk_submit_io_sync(pblk, &rqd);
850 851 852 853 854 855
	if (ret) {
		pr_err("pblk: smeta I/O submission failed: %d\n", ret);
		bio_put(bio);
		goto free_ppa_list;
	}

856
	atomic_dec(&pblk->inflight_io);
857 858

	if (rqd.error) {
859
		if (dir == PBLK_WRITE)
860
			pblk_log_write_err(pblk, &rqd);
861
		else if (dir == PBLK_READ)
862 863 864 865
			pblk_log_read_err(pblk, &rqd);
	}

free_ppa_list:
866
	nvm_dev_dma_free(dev->parent, rqd.meta_list, rqd.dma_meta_list);
867 868 869 870 871 872 873 874

	return ret;
}

int pblk_line_read_smeta(struct pblk *pblk, struct pblk_line *line)
{
	u64 bpaddr = pblk_line_smeta_start(pblk, line);

875
	return pblk_line_submit_smeta_io(pblk, line, bpaddr, PBLK_READ_RECOV);
876 877
}

878 879
int pblk_line_read_emeta(struct pblk *pblk, struct pblk_line *line,
			 void *emeta_buf)
880
{
881
	return pblk_line_submit_emeta_io(pblk, line, emeta_buf,
882
						line->emeta_ssec, PBLK_READ);
883 884 885 886 887 888 889 890
}

static void pblk_setup_e_rq(struct pblk *pblk, struct nvm_rq *rqd,
			    struct ppa_addr ppa)
{
	rqd->opcode = NVM_OP_ERASE;
	rqd->ppa_addr = ppa;
	rqd->nr_ppas = 1;
891
	rqd->flags = pblk_set_progr_mode(pblk, PBLK_ERASE);
892 893 894 895 896 897
	rqd->bio = NULL;
}

static int pblk_blk_erase_sync(struct pblk *pblk, struct ppa_addr ppa)
{
	struct nvm_rq rqd;
898
	int ret = 0;
899 900 901 902 903 904 905 906

	memset(&rqd, 0, sizeof(struct nvm_rq));

	pblk_setup_e_rq(pblk, &rqd, ppa);

	/* The write thread schedules erases so that it minimizes disturbances
	 * with writes. Thus, there is no need to take the LUN semaphore.
	 */
907
	ret = pblk_submit_io_sync(pblk, &rqd);
908 909 910 911 912
	if (ret) {
		struct nvm_tgt_dev *dev = pblk->dev;
		struct nvm_geo *geo = &dev->geo;

		pr_err("pblk: could not sync erase line:%d,blk:%d\n",
913 914
					pblk_ppa_to_line(ppa),
					pblk_ppa_to_pos(geo, ppa));
915 916 917 918 919 920 921 922 923

		rqd.error = ret;
		goto out;
	}

out:
	rqd.private = pblk;
	__pblk_end_io_erase(pblk, &rqd);

924
	return ret;
925 926 927 928 929 930
}

int pblk_line_erase(struct pblk *pblk, struct pblk_line *line)
{
	struct pblk_line_meta *lm = &pblk->lm;
	struct ppa_addr ppa;
931
	int ret, bit = -1;
932

933 934 935 936 937 938 939 940 941 942
	/* Erase only good blocks, one at a time */
	do {
		spin_lock(&line->lock);
		bit = find_next_zero_bit(line->erase_bitmap, lm->blk_per_line,
								bit + 1);
		if (bit >= lm->blk_per_line) {
			spin_unlock(&line->lock);
			break;
		}

943
		ppa = pblk->luns[bit].bppa; /* set ch and lun */
944
		ppa.a.blk = line->id;
945

946
		atomic_dec(&line->left_eblks);
947
		WARN_ON(test_and_set_bit(bit, line->erase_bitmap));
948
		spin_unlock(&line->lock);
949

950 951
		ret = pblk_blk_erase_sync(pblk, ppa);
		if (ret) {
952
			pr_err("pblk: failed to erase line %d\n", line->id);
953
			return ret;
954
		}
955
	} while (1);
956 957 958 959

	return 0;
}

960 961 962 963 964 965
static void pblk_line_setup_metadata(struct pblk_line *line,
				     struct pblk_line_mgmt *l_mg,
				     struct pblk_line_meta *lm)
{
	int meta_line;

966 967
	lockdep_assert_held(&l_mg->free_lock);

968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989
retry_meta:
	meta_line = find_first_zero_bit(&l_mg->meta_bitmap, PBLK_DATA_LINES);
	if (meta_line == PBLK_DATA_LINES) {
		spin_unlock(&l_mg->free_lock);
		io_schedule();
		spin_lock(&l_mg->free_lock);
		goto retry_meta;
	}

	set_bit(meta_line, &l_mg->meta_bitmap);
	line->meta_line = meta_line;

	line->smeta = l_mg->sline_meta[meta_line];
	line->emeta = l_mg->eline_meta[meta_line];

	memset(line->smeta, 0, lm->smeta_len);
	memset(line->emeta->buf, 0, lm->emeta_len[0]);

	line->emeta->mem = 0;
	atomic_set(&line->emeta->sync, 0);
}

990 991 992
/* For now lines are always assumed full lines. Thus, smeta former and current
 * lun bitmaps are omitted.
 */
993
static int pblk_line_init_metadata(struct pblk *pblk, struct pblk_line *line,
994 995 996 997 998 999
				  struct pblk_line *cur)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct pblk_line_meta *lm = &pblk->lm;
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
1000 1001 1002
	struct pblk_emeta *emeta = line->emeta;
	struct line_emeta *emeta_buf = emeta->buf;
	struct line_smeta *smeta_buf = (struct line_smeta *)line->smeta;
1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024
	int nr_blk_line;

	/* After erasing the line, new bad blocks might appear and we risk
	 * having an invalid line
	 */
	nr_blk_line = lm->blk_per_line -
			bitmap_weight(line->blk_bitmap, lm->blk_per_line);
	if (nr_blk_line < lm->min_blk_line) {
		spin_lock(&l_mg->free_lock);
		spin_lock(&line->lock);
		line->state = PBLK_LINESTATE_BAD;
		spin_unlock(&line->lock);

		list_add_tail(&line->list, &l_mg->bad_list);
		spin_unlock(&l_mg->free_lock);

		pr_debug("pblk: line %d is bad\n", line->id);

		return 0;
	}

	/* Run-time metadata */
1025
	line->lun_bitmap = ((void *)(smeta_buf)) + sizeof(struct line_smeta);
1026 1027 1028 1029

	/* Mark LUNs allocated in this line (all for now) */
	bitmap_set(line->lun_bitmap, 0, lm->lun_bitmap_len);

1030 1031 1032 1033
	smeta_buf->header.identifier = cpu_to_le32(PBLK_MAGIC);
	memcpy(smeta_buf->header.uuid, pblk->instance_uuid, 16);
	smeta_buf->header.id = cpu_to_le32(line->id);
	smeta_buf->header.type = cpu_to_le16(line->type);
1034 1035
	smeta_buf->header.version_major = SMETA_VERSION_MAJOR;
	smeta_buf->header.version_minor = SMETA_VERSION_MINOR;
1036 1037

	/* Start metadata */
1038
	smeta_buf->seq_nr = cpu_to_le64(line->seq_nr);
1039
	smeta_buf->window_wr_lun = cpu_to_le32(geo->all_luns);
1040 1041 1042 1043

	/* Fill metadata among lines */
	if (cur) {
		memcpy(line->lun_bitmap, cur->lun_bitmap, lm->lun_bitmap_len);
1044 1045
		smeta_buf->prev_id = cpu_to_le32(cur->id);
		cur->emeta->buf->next_id = cpu_to_le32(line->id);
1046
	} else {
1047
		smeta_buf->prev_id = cpu_to_le32(PBLK_LINE_EMPTY);
1048 1049 1050
	}

	/* All smeta must be set at this point */
1051 1052 1053
	smeta_buf->header.crc = cpu_to_le32(
			pblk_calc_meta_header_crc(pblk, &smeta_buf->header));
	smeta_buf->crc = cpu_to_le32(pblk_calc_smeta_crc(pblk, smeta_buf));
1054 1055

	/* End metadata */
1056 1057
	memcpy(&emeta_buf->header, &smeta_buf->header,
						sizeof(struct line_header));
1058 1059 1060 1061 1062 1063

	emeta_buf->header.version_major = EMETA_VERSION_MAJOR;
	emeta_buf->header.version_minor = EMETA_VERSION_MINOR;
	emeta_buf->header.crc = cpu_to_le32(
			pblk_calc_meta_header_crc(pblk, &emeta_buf->header));

1064 1065 1066 1067 1068 1069
	emeta_buf->seq_nr = cpu_to_le64(line->seq_nr);
	emeta_buf->nr_lbas = cpu_to_le64(line->sec_in_line);
	emeta_buf->nr_valid_lbas = cpu_to_le64(0);
	emeta_buf->next_id = cpu_to_le32(PBLK_LINE_EMPTY);
	emeta_buf->crc = cpu_to_le32(0);
	emeta_buf->prev_id = smeta_buf->prev_id;
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085

	return 1;
}

/* For now lines are always assumed full lines. Thus, smeta former and current
 * lun bitmaps are omitted.
 */
static int pblk_line_init_bb(struct pblk *pblk, struct pblk_line *line,
			     int init)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct pblk_line_meta *lm = &pblk->lm;
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	u64 off;
	int bit = -1;
1086
	int emeta_secs;
1087 1088 1089 1090 1091 1092

	line->sec_in_line = lm->sec_per_line;

	/* Capture bad block information on line mapping bitmaps */
	while ((bit = find_next_bit(line->blk_bitmap, lm->blk_per_line,
					bit + 1)) < lm->blk_per_line) {
1093
		off = bit * geo->ws_opt;
1094 1095 1096 1097
		bitmap_shift_left(l_mg->bb_aux, l_mg->bb_template, off,
							lm->sec_per_line);
		bitmap_or(line->map_bitmap, line->map_bitmap, l_mg->bb_aux,
							lm->sec_per_line);
1098
		line->sec_in_line -= geo->clba;
1099 1100 1101 1102
	}

	/* Mark smeta metadata sectors as bad sectors */
	bit = find_first_zero_bit(line->blk_bitmap, lm->blk_per_line);
1103
	off = bit * geo->ws_opt;
1104 1105 1106 1107 1108
	bitmap_set(line->map_bitmap, off, lm->smeta_sec);
	line->sec_in_line -= lm->smeta_sec;
	line->smeta_ssec = off;
	line->cur_sec = off + lm->smeta_sec;

1109
	if (init && pblk_line_submit_smeta_io(pblk, line, off, PBLK_WRITE)) {
1110
		pr_debug("pblk: line smeta I/O failed. Retry\n");
1111
		return 1;
1112 1113 1114 1115 1116 1117 1118
	}

	bitmap_copy(line->invalid_bitmap, line->map_bitmap, lm->sec_per_line);

	/* Mark emeta metadata sectors as bad sectors. We need to consider bad
	 * blocks to make sure that there are enough sectors to store emeta
	 */
1119 1120 1121
	emeta_secs = lm->emeta_sec[0];
	off = lm->sec_per_line;
	while (emeta_secs) {
1122
		off -= geo->ws_opt;
1123
		if (!test_bit(off, line->invalid_bitmap)) {
1124 1125
			bitmap_set(line->invalid_bitmap, off, geo->ws_opt);
			emeta_secs -= geo->ws_opt;
1126 1127 1128 1129
		}
	}

	line->emeta_ssec = off;
1130
	line->sec_in_line -= lm->emeta_sec[0];
1131
	line->nr_valid_lbas = 0;
1132
	line->left_msecs = line->sec_in_line;
1133
	*line->vsc = cpu_to_le32(line->sec_in_line);
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149

	if (lm->sec_per_line - line->sec_in_line !=
		bitmap_weight(line->invalid_bitmap, lm->sec_per_line)) {
		spin_lock(&line->lock);
		line->state = PBLK_LINESTATE_BAD;
		spin_unlock(&line->lock);

		list_add_tail(&line->list, &l_mg->bad_list);
		pr_err("pblk: unexpected line %d is bad\n", line->id);

		return 0;
	}

	return 1;
}

1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173
static int pblk_prepare_new_line(struct pblk *pblk, struct pblk_line *line)
{
	struct pblk_line_meta *lm = &pblk->lm;
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	int blk_to_erase = atomic_read(&line->blk_in_line);
	int i;

	for (i = 0; i < lm->blk_per_line; i++) {
		struct pblk_lun *rlun = &pblk->luns[i];
		int pos = pblk_ppa_to_pos(geo, rlun->bppa);
		int state = line->chks[pos].state;

		/* Free chunks should not be erased */
		if (state & NVM_CHK_ST_FREE) {
			set_bit(pblk_ppa_to_pos(geo, rlun->bppa),
							line->erase_bitmap);
			blk_to_erase--;
		}
	}

	return blk_to_erase;
}

1174 1175 1176
static int pblk_line_prepare(struct pblk *pblk, struct pblk_line *line)
{
	struct pblk_line_meta *lm = &pblk->lm;
1177 1178
	int blk_in_line = atomic_read(&line->blk_in_line);
	int blk_to_erase, ret;
1179

1180
	line->map_bitmap = kzalloc(lm->sec_bitmap_len, GFP_ATOMIC);
1181 1182 1183
	if (!line->map_bitmap)
		return -ENOMEM;

1184 1185
	/* will be initialized using bb info from map_bitmap */
	line->invalid_bitmap = kmalloc(lm->sec_bitmap_len, GFP_ATOMIC);
1186
	if (!line->invalid_bitmap) {
1187 1188
		ret = -ENOMEM;
		goto fail_free_map_bitmap;
1189 1190
	}

1191 1192 1193
	/* Bad blocks do not need to be erased */
	bitmap_copy(line->erase_bitmap, line->blk_bitmap, lm->blk_per_line);

1194
	spin_lock(&line->lock);
1195 1196 1197 1198 1199 1200 1201 1202

	/* If we have not written to this line, we need to mark up free chunks
	 * as already erased
	 */
	if (line->state == PBLK_LINESTATE_NEW) {
		blk_to_erase = pblk_prepare_new_line(pblk, line);
		line->state = PBLK_LINESTATE_FREE;
	} else {
1203 1204 1205 1206 1207 1208
		blk_to_erase = blk_in_line;
	}

	if (blk_in_line < lm->min_blk_line) {
		ret = -EAGAIN;
		goto fail_free_invalid_bitmap;
1209 1210
	}

1211
	if (line->state != PBLK_LINESTATE_FREE) {
1212 1213
		WARN(1, "pblk: corrupted line %d, state %d\n",
							line->id, line->state);
1214 1215
		ret = -EINTR;
		goto fail_free_invalid_bitmap;
1216
	}
1217

1218
	line->state = PBLK_LINESTATE_OPEN;
1219

1220 1221
	atomic_set(&line->left_eblks, blk_to_erase);
	atomic_set(&line->left_seblks, blk_to_erase);
1222 1223

	line->meta_distance = lm->meta_distance;
1224 1225 1226 1227 1228
	spin_unlock(&line->lock);

	kref_init(&line->ref);

	return 0;
1229 1230 1231 1232 1233 1234 1235 1236 1237 1238

fail_free_invalid_bitmap:
	spin_unlock(&line->lock);
	kfree(line->invalid_bitmap);
	line->invalid_bitmap = NULL;
fail_free_map_bitmap:
	kfree(line->map_bitmap);
	line->map_bitmap = NULL;

	return ret;
1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
}

int pblk_line_recov_alloc(struct pblk *pblk, struct pblk_line *line)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	int ret;

	spin_lock(&l_mg->free_lock);
	l_mg->data_line = line;
	list_del(&line->list);

	ret = pblk_line_prepare(pblk, line);
	if (ret) {
		list_add(&line->list, &l_mg->free_list);
1253
		spin_unlock(&l_mg->free_lock);
1254 1255
		return ret;
	}
1256
	spin_unlock(&l_mg->free_lock);
1257

1258
	pblk_rl_free_lines_dec(&pblk->rl, line, true);
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269

	if (!pblk_line_init_bb(pblk, line, 0)) {
		list_add(&line->list, &l_mg->free_list);
		return -EINTR;
	}

	return 0;
}

void pblk_line_recov_close(struct pblk *pblk, struct pblk_line *line)
{
1270
	kfree(line->map_bitmap);
1271 1272 1273 1274 1275 1276 1277 1278 1279
	line->map_bitmap = NULL;
	line->smeta = NULL;
	line->emeta = NULL;
}

struct pblk_line *pblk_line_get(struct pblk *pblk)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct pblk_line_meta *lm = &pblk->lm;
1280 1281
	struct pblk_line *line;
	int ret, bit;
1282 1283 1284

	lockdep_assert_held(&l_mg->free_lock);

1285
retry:
1286 1287
	if (list_empty(&l_mg->free_list)) {
		pr_err("pblk: no free lines\n");
1288
		return NULL;
1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
	}

	line = list_first_entry(&l_mg->free_list, struct pblk_line, list);
	list_del(&line->list);
	l_mg->nr_free_lines--;

	bit = find_first_zero_bit(line->blk_bitmap, lm->blk_per_line);
	if (unlikely(bit >= lm->blk_per_line)) {
		spin_lock(&line->lock);
		line->state = PBLK_LINESTATE_BAD;
		spin_unlock(&line->lock);

		list_add_tail(&line->list, &l_mg->bad_list);

		pr_debug("pblk: line %d is bad\n", line->id);
1304
		goto retry;
1305 1306
	}

1307 1308
	ret = pblk_line_prepare(pblk, line);
	if (ret) {
1309 1310 1311 1312 1313
		switch (ret) {
		case -EAGAIN:
			list_add(&line->list, &l_mg->bad_list);
			goto retry;
		case -EINTR:
1314 1315
			list_add(&line->list, &l_mg->corrupt_list);
			goto retry;
1316
		default:
1317 1318 1319 1320 1321
			pr_err("pblk: failed to prepare line %d\n", line->id);
			list_add(&line->list, &l_mg->free_list);
			l_mg->nr_free_lines++;
			return NULL;
		}
1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
	}

	return line;
}

static struct pblk_line *pblk_line_retry(struct pblk *pblk,
					 struct pblk_line *line)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct pblk_line *retry_line;

1333
retry:
1334 1335 1336
	spin_lock(&l_mg->free_lock);
	retry_line = pblk_line_get(pblk);
	if (!retry_line) {
1337
		l_mg->data_line = NULL;
1338 1339 1340 1341 1342 1343 1344 1345
		spin_unlock(&l_mg->free_lock);
		return NULL;
	}

	retry_line->smeta = line->smeta;
	retry_line->emeta = line->emeta;
	retry_line->meta_line = line->meta_line;

1346
	pblk_line_free(pblk, line);
1347
	l_mg->data_line = retry_line;
1348 1349
	spin_unlock(&l_mg->free_lock);

1350
	pblk_rl_free_lines_dec(&pblk->rl, line, false);
1351

1352 1353 1354
	if (pblk_line_erase(pblk, retry_line))
		goto retry;

1355 1356 1357
	return retry_line;
}

1358 1359 1360 1361 1362 1363 1364
static void pblk_set_space_limit(struct pblk *pblk)
{
	struct pblk_rl *rl = &pblk->rl;

	atomic_set(&rl->rb_space, 0);
}

1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380
struct pblk_line *pblk_line_get_first_data(struct pblk *pblk)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct pblk_line *line;

	spin_lock(&l_mg->free_lock);
	line = pblk_line_get(pblk);
	if (!line) {
		spin_unlock(&l_mg->free_lock);
		return NULL;
	}

	line->seq_nr = l_mg->d_seq_nr++;
	line->type = PBLK_LINETYPE_DATA;
	l_mg->data_line = line;

1381
	pblk_line_setup_metadata(line, l_mg, &pblk->lm);
1382 1383 1384

	/* Allocate next line for preparation */
	l_mg->data_next = pblk_line_get(pblk);
1385 1386 1387 1388 1389 1390 1391 1392 1393
	if (!l_mg->data_next) {
		/* If we cannot get a new line, we need to stop the pipeline.
		 * Only allow as many writes in as we can store safely and then
		 * fail gracefully
		 */
		pblk_set_space_limit(pblk);

		l_mg->data_next = NULL;
	} else {
1394 1395 1396 1397 1398
		l_mg->data_next->seq_nr = l_mg->d_seq_nr++;
		l_mg->data_next->type = PBLK_LINETYPE_DATA;
	}
	spin_unlock(&l_mg->free_lock);

1399 1400 1401 1402 1403 1404
	if (pblk_line_erase(pblk, line)) {
		line = pblk_line_retry(pblk, line);
		if (!line)
			return NULL;
	}

1405
retry_setup:
1406
	if (!pblk_line_init_metadata(pblk, line, NULL)) {
1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
		line = pblk_line_retry(pblk, line);
		if (!line)
			return NULL;

		goto retry_setup;
	}

	if (!pblk_line_init_bb(pblk, line, 1)) {
		line = pblk_line_retry(pblk, line);
		if (!line)
			return NULL;

		goto retry_setup;
	}

1422 1423
	pblk_rl_free_lines_dec(&pblk->rl, line, true);

1424 1425 1426
	return line;
}

1427 1428 1429 1430 1431 1432 1433 1434
static void pblk_stop_writes(struct pblk *pblk, struct pblk_line *line)
{
	lockdep_assert_held(&pblk->l_mg.free_lock);

	pblk_set_space_limit(pblk);
	pblk->state = PBLK_STATE_STOPPING;
}

1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
static void pblk_line_close_meta_sync(struct pblk *pblk)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct pblk_line_meta *lm = &pblk->lm;
	struct pblk_line *line, *tline;
	LIST_HEAD(list);

	spin_lock(&l_mg->close_lock);
	if (list_empty(&l_mg->emeta_list)) {
		spin_unlock(&l_mg->close_lock);
		return;
	}

	list_cut_position(&list, &l_mg->emeta_list, l_mg->emeta_list.prev);
	spin_unlock(&l_mg->close_lock);

	list_for_each_entry_safe(line, tline, &list, list) {
		struct pblk_emeta *emeta = line->emeta;

		while (emeta->mem < lm->emeta_len[0]) {
			int ret;

			ret = pblk_submit_meta_io(pblk, line);
			if (ret) {
				pr_err("pblk: sync meta line %d failed (%d)\n",
							line->id, ret);
				return;
			}
		}
	}

	pblk_wait_for_meta(pblk);
	flush_workqueue(pblk->close_wq);
}

1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
void pblk_pipeline_stop(struct pblk *pblk)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	int ret;

	spin_lock(&l_mg->free_lock);
	if (pblk->state == PBLK_STATE_RECOVERING ||
					pblk->state == PBLK_STATE_STOPPED) {
		spin_unlock(&l_mg->free_lock);
		return;
	}
	pblk->state = PBLK_STATE_RECOVERING;
	spin_unlock(&l_mg->free_lock);

	pblk_flush_writer(pblk);
	pblk_wait_for_meta(pblk);

	ret = pblk_recov_pad(pblk);
	if (ret) {
		pr_err("pblk: could not close data on teardown(%d)\n", ret);
		return;
	}

1493
	flush_workqueue(pblk->bb_wq);
1494 1495 1496 1497 1498 1499 1500 1501 1502
	pblk_line_close_meta_sync(pblk);

	spin_lock(&l_mg->free_lock);
	pblk->state = PBLK_STATE_STOPPED;
	l_mg->data_line = NULL;
	l_mg->data_next = NULL;
	spin_unlock(&l_mg->free_lock);
}

1503
struct pblk_line *pblk_line_replace_data(struct pblk *pblk)
1504 1505
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
1506
	struct pblk_line *cur, *new = NULL;
1507 1508 1509 1510 1511
	unsigned int left_seblks;

	cur = l_mg->data_line;
	new = l_mg->data_next;
	if (!new)
1512
		goto out;
1513 1514
	l_mg->data_line = new;

1515 1516 1517 1518 1519
	spin_lock(&l_mg->free_lock);
	pblk_line_setup_metadata(new, l_mg, &pblk->lm);
	spin_unlock(&l_mg->free_lock);

retry_erase:
1520 1521 1522
	left_seblks = atomic_read(&new->left_seblks);
	if (left_seblks) {
		/* If line is not fully erased, erase it */
1523
		if (atomic_read(&new->left_eblks)) {
1524
			if (pblk_line_erase(pblk, new))
1525
				goto out;
1526 1527 1528
		} else {
			io_schedule();
		}
1529
		goto retry_erase;
1530 1531 1532
	}

retry_setup:
1533
	if (!pblk_line_init_metadata(pblk, new, cur)) {
1534
		new = pblk_line_retry(pblk, new);
1535
		if (!new)
1536
			goto out;
1537 1538 1539 1540 1541 1542 1543

		goto retry_setup;
	}

	if (!pblk_line_init_bb(pblk, new, 1)) {
		new = pblk_line_retry(pblk, new);
		if (!new)
1544
			goto out;
1545 1546 1547 1548

		goto retry_setup;
	}

1549 1550
	pblk_rl_free_lines_dec(&pblk->rl, new, true);

1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566
	/* Allocate next line for preparation */
	spin_lock(&l_mg->free_lock);
	l_mg->data_next = pblk_line_get(pblk);
	if (!l_mg->data_next) {
		/* If we cannot get a new line, we need to stop the pipeline.
		 * Only allow as many writes in as we can store safely and then
		 * fail gracefully
		 */
		pblk_stop_writes(pblk, new);
		l_mg->data_next = NULL;
	} else {
		l_mg->data_next->seq_nr = l_mg->d_seq_nr++;
		l_mg->data_next->type = PBLK_LINETYPE_DATA;
	}
	spin_unlock(&l_mg->free_lock);

1567 1568
out:
	return new;
1569 1570 1571 1572
}

void pblk_line_free(struct pblk *pblk, struct pblk_line *line)
{
1573 1574
	kfree(line->map_bitmap);
	kfree(line->invalid_bitmap);
1575

1576 1577
	*line->vsc = cpu_to_le32(EMPTY_ENTRY);

1578 1579
	line->map_bitmap = NULL;
	line->invalid_bitmap = NULL;
1580 1581
	line->smeta = NULL;
	line->emeta = NULL;
1582 1583
}

1584
static void __pblk_line_put(struct pblk *pblk, struct pblk_line *line)
1585 1586
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
1587
	struct pblk_gc *gc = &pblk->gc;
1588 1589 1590 1591 1592 1593 1594 1595

	spin_lock(&line->lock);
	WARN_ON(line->state != PBLK_LINESTATE_GC);
	line->state = PBLK_LINESTATE_FREE;
	line->gc_group = PBLK_LINEGC_NONE;
	pblk_line_free(pblk, line);
	spin_unlock(&line->lock);

1596 1597
	atomic_dec(&gc->pipeline_gc);

1598 1599 1600 1601 1602 1603 1604 1605
	spin_lock(&l_mg->free_lock);
	list_add_tail(&line->list, &l_mg->free_list);
	l_mg->nr_free_lines++;
	spin_unlock(&l_mg->free_lock);

	pblk_rl_free_lines_inc(&pblk->rl, line);
}

1606 1607 1608 1609 1610 1611 1612 1613
static void pblk_line_put_ws(struct work_struct *work)
{
	struct pblk_line_ws *line_put_ws = container_of(work,
						struct pblk_line_ws, ws);
	struct pblk *pblk = line_put_ws->pblk;
	struct pblk_line *line = line_put_ws->line;

	__pblk_line_put(pblk, line);
1614
	mempool_free(line_put_ws, &pblk->gen_ws_pool);
1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
}

void pblk_line_put(struct kref *ref)
{
	struct pblk_line *line = container_of(ref, struct pblk_line, ref);
	struct pblk *pblk = line->pblk;

	__pblk_line_put(pblk, line);
}

void pblk_line_put_wq(struct kref *ref)
{
	struct pblk_line *line = container_of(ref, struct pblk_line, ref);
	struct pblk *pblk = line->pblk;
	struct pblk_line_ws *line_put_ws;

1631
	line_put_ws = mempool_alloc(&pblk->gen_ws_pool, GFP_ATOMIC);
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
	if (!line_put_ws)
		return;

	line_put_ws->pblk = pblk;
	line_put_ws->line = line;
	line_put_ws->priv = NULL;

	INIT_WORK(&line_put_ws->ws, pblk_line_put_ws);
	queue_work(pblk->r_end_wq, &line_put_ws->ws);
}

1643 1644 1645 1646 1647
int pblk_blk_erase_async(struct pblk *pblk, struct ppa_addr ppa)
{
	struct nvm_rq *rqd;
	int err;

1648
	rqd = pblk_alloc_rqd(pblk, PBLK_ERASE);
1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663

	pblk_setup_e_rq(pblk, rqd, ppa);

	rqd->end_io = pblk_end_io_erase;
	rqd->private = pblk;

	/* The write thread schedules erases so that it minimizes disturbances
	 * with writes. Thus, there is no need to take the LUN semaphore.
	 */
	err = pblk_submit_io(pblk, rqd);
	if (err) {
		struct nvm_tgt_dev *dev = pblk->dev;
		struct nvm_geo *geo = &dev->geo;

		pr_err("pblk: could not async erase line:%d,blk:%d\n",
1664 1665
					pblk_ppa_to_line(ppa),
					pblk_ppa_to_pos(geo, ppa));
1666 1667 1668 1669 1670 1671 1672 1673 1674 1675
	}

	return err;
}

struct pblk_line *pblk_line_get_data(struct pblk *pblk)
{
	return pblk->l_mg.data_line;
}

1676 1677
/* For now, always erase next line */
struct pblk_line *pblk_line_get_erase(struct pblk *pblk)
1678 1679 1680 1681 1682 1683 1684 1685 1686
{
	return pblk->l_mg.data_next;
}

int pblk_line_is_full(struct pblk_line *line)
{
	return (line->left_msecs == 0);
}

1687 1688 1689 1690 1691 1692
static void pblk_line_should_sync_meta(struct pblk *pblk)
{
	if (pblk_rl_is_limit(&pblk->rl))
		pblk_line_close_meta_sync(pblk);
}

1693 1694
void pblk_line_close(struct pblk *pblk, struct pblk_line *line)
{
1695 1696 1697
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct pblk_line_meta *lm = &pblk->lm;
1698 1699
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct list_head *move_list;
1700
	int i;
1701

1702
#ifdef CONFIG_NVM_DEBUG
1703
	WARN(!bitmap_full(line->map_bitmap, lm->sec_per_line),
1704
				"pblk: corrupt closed line %d\n", line->id);
1705
#endif
1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718

	spin_lock(&l_mg->free_lock);
	WARN_ON(!test_and_clear_bit(line->meta_line, &l_mg->meta_bitmap));
	spin_unlock(&l_mg->free_lock);

	spin_lock(&l_mg->gc_lock);
	spin_lock(&line->lock);
	WARN_ON(line->state != PBLK_LINESTATE_OPEN);
	line->state = PBLK_LINESTATE_CLOSED;
	move_list = pblk_line_gc_list(pblk, line);

	list_add_tail(&line->list, move_list);

1719
	kfree(line->map_bitmap);
1720 1721 1722 1723
	line->map_bitmap = NULL;
	line->smeta = NULL;
	line->emeta = NULL;

1724 1725 1726 1727 1728 1729 1730 1731 1732
	for (i = 0; i < lm->blk_per_line; i++) {
		struct pblk_lun *rlun = &pblk->luns[i];
		int pos = pblk_ppa_to_pos(geo, rlun->bppa);
		int state = line->chks[pos].state;

		if (!(state & NVM_CHK_ST_OFFLINE))
			state = NVM_CHK_ST_CLOSED;
	}

1733 1734 1735 1736
	spin_unlock(&line->lock);
	spin_unlock(&l_mg->gc_lock);
}

1737 1738 1739 1740 1741 1742
void pblk_line_close_meta(struct pblk *pblk, struct pblk_line *line)
{
	struct pblk_line_mgmt *l_mg = &pblk->l_mg;
	struct pblk_line_meta *lm = &pblk->lm;
	struct pblk_emeta *emeta = line->emeta;
	struct line_emeta *emeta_buf = emeta->buf;
1743
	struct wa_counters *wa = emeta_to_wa(lm, emeta_buf);
1744

1745
	/* No need for exact vsc value; avoid a big line lock and take aprox. */
1746 1747 1748
	memcpy(emeta_to_vsc(pblk, emeta_buf), l_mg->vsc_list, lm->vsc_list_len);
	memcpy(emeta_to_bb(emeta_buf), line->blk_bitmap, lm->blk_bitmap_len);

1749 1750 1751 1752
	wa->user = cpu_to_le64(atomic64_read(&pblk->user_wa));
	wa->pad = cpu_to_le64(atomic64_read(&pblk->pad_wa));
	wa->gc = cpu_to_le64(atomic64_read(&pblk->gc_wa));

1753 1754 1755 1756 1757 1758 1759 1760
	emeta_buf->nr_valid_lbas = cpu_to_le64(line->nr_valid_lbas);
	emeta_buf->crc = cpu_to_le32(pblk_calc_emeta_crc(pblk, emeta_buf));

	spin_lock(&l_mg->close_lock);
	spin_lock(&line->lock);
	list_add_tail(&line->list, &l_mg->emeta_list);
	spin_unlock(&line->lock);
	spin_unlock(&l_mg->close_lock);
1761 1762

	pblk_line_should_sync_meta(pblk);
1763 1764
}

1765 1766 1767 1768 1769 1770 1771 1772
void pblk_line_close_ws(struct work_struct *work)
{
	struct pblk_line_ws *line_ws = container_of(work, struct pblk_line_ws,
									ws);
	struct pblk *pblk = line_ws->pblk;
	struct pblk_line *line = line_ws->line;

	pblk_line_close(pblk, line);
1773
	mempool_free(line_ws, &pblk->gen_ws_pool);
1774 1775
}

1776 1777
void pblk_gen_run_ws(struct pblk *pblk, struct pblk_line *line, void *priv,
		      void (*work)(struct work_struct *), gfp_t gfp_mask,
1778
		      struct workqueue_struct *wq)
1779 1780 1781
{
	struct pblk_line_ws *line_ws;

1782
	line_ws = mempool_alloc(&pblk->gen_ws_pool, gfp_mask);
1783 1784 1785 1786 1787 1788

	line_ws->pblk = pblk;
	line_ws->line = line;
	line_ws->priv = priv;

	INIT_WORK(&line_ws->ws, work);
1789
	queue_work(wq, &line_ws->ws);
1790 1791
}

1792 1793
static void __pblk_down_page(struct pblk *pblk, struct ppa_addr *ppa_list,
			     int nr_ppas, int pos)
1794
{
1795
	struct pblk_lun *rlun = &pblk->luns[pos];
1796 1797 1798 1799 1800 1801 1802 1803 1804 1805
	int ret;

	/*
	 * Only send one inflight I/O per LUN. Since we map at a page
	 * granurality, all ppas in the I/O will map to the same LUN
	 */
#ifdef CONFIG_NVM_DEBUG
	int i;

	for (i = 1; i < nr_ppas; i++)
1806 1807
		WARN_ON(ppa_list[0].a.lun != ppa_list[i].a.lun ||
				ppa_list[0].a.ch != ppa_list[i].a.ch);
1808 1809
#endif

1810
	ret = down_timeout(&rlun->wr_sem, msecs_to_jiffies(30000));
1811 1812
	if (ret == -ETIME || ret == -EINTR)
		pr_err("pblk: taking lun semaphore timed out: err %d\n", -ret);
1813 1814
}

1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850
void pblk_down_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);

	__pblk_down_page(pblk, ppa_list, nr_ppas, pos);
}

void pblk_down_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
		  unsigned long *lun_bitmap)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);

	/* If the LUN has been locked for this same request, do no attempt to
	 * lock it again
	 */
	if (test_and_set_bit(pos, lun_bitmap))
		return;

	__pblk_down_page(pblk, ppa_list, nr_ppas, pos);
}

void pblk_up_page(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct pblk_lun *rlun;
	int pos = pblk_ppa_to_pos(geo, ppa_list[0]);

#ifdef CONFIG_NVM_DEBUG
	int i;

	for (i = 1; i < nr_ppas; i++)
1851 1852
		WARN_ON(ppa_list[0].a.lun != ppa_list[i].a.lun ||
				ppa_list[0].a.ch != ppa_list[i].a.ch);
1853 1854 1855 1856 1857 1858
#endif

	rlun = &pblk->luns[pos];
	up(&rlun->wr_sem);
}

1859 1860 1861 1862 1863 1864
void pblk_up_rq(struct pblk *pblk, struct ppa_addr *ppa_list, int nr_ppas,
		unsigned long *lun_bitmap)
{
	struct nvm_tgt_dev *dev = pblk->dev;
	struct nvm_geo *geo = &dev->geo;
	struct pblk_lun *rlun;
1865
	int num_lun = geo->all_luns;
1866 1867
	int bit = -1;

1868
	while ((bit = find_next_bit(lun_bitmap, num_lun, bit + 1)) < num_lun) {
1869 1870 1871 1872 1873 1874 1875
		rlun = &pblk->luns[bit];
		up(&rlun->wr_sem);
	}
}

void pblk_update_map(struct pblk *pblk, sector_t lba, struct ppa_addr ppa)
{
1876
	struct ppa_addr ppa_l2p;
1877 1878 1879 1880 1881 1882 1883 1884

	/* logic error: lba out-of-bounds. Ignore update */
	if (!(lba < pblk->rl.nr_secs)) {
		WARN(1, "pblk: corrupted L2P map request\n");
		return;
	}

	spin_lock(&pblk->trans_lock);
1885
	ppa_l2p = pblk_trans_map_get(pblk, lba);
1886

1887 1888
	if (!pblk_addr_in_cache(ppa_l2p) && !pblk_ppa_empty(ppa_l2p))
		pblk_map_invalidate(pblk, ppa_l2p);
1889 1890 1891 1892 1893 1894 1895

	pblk_trans_map_set(pblk, lba, ppa);
	spin_unlock(&pblk->trans_lock);
}

void pblk_update_map_cache(struct pblk *pblk, sector_t lba, struct ppa_addr ppa)
{
1896

1897 1898 1899 1900 1901 1902 1903 1904 1905
#ifdef CONFIG_NVM_DEBUG
	/* Callers must ensure that the ppa points to a cache address */
	BUG_ON(!pblk_addr_in_cache(ppa));
	BUG_ON(pblk_rb_pos_oob(&pblk->rwb, pblk_addr_to_cacheline(ppa)));
#endif

	pblk_update_map(pblk, lba, ppa);
}

1906
int pblk_update_map_gc(struct pblk *pblk, sector_t lba, struct ppa_addr ppa_new,
1907
		       struct pblk_line *gc_line, u64 paddr_gc)
1908
{
1909
	struct ppa_addr ppa_l2p, ppa_gc;
1910 1911 1912 1913
	int ret = 1;

#ifdef CONFIG_NVM_DEBUG
	/* Callers must ensure that the ppa points to a cache address */
1914 1915
	BUG_ON(!pblk_addr_in_cache(ppa_new));
	BUG_ON(pblk_rb_pos_oob(&pblk->rwb, pblk_addr_to_cacheline(ppa_new)));
1916 1917 1918 1919 1920 1921 1922 1923 1924
#endif

	/* logic error: lba out-of-bounds. Ignore update */
	if (!(lba < pblk->rl.nr_secs)) {
		WARN(1, "pblk: corrupted L2P map request\n");
		return 0;
	}

	spin_lock(&pblk->trans_lock);
1925
	ppa_l2p = pblk_trans_map_get(pblk, lba);
1926
	ppa_gc = addr_to_gen_ppa(pblk, paddr_gc, gc_line->id);
1927

1928 1929 1930 1931 1932
	if (!pblk_ppa_comp(ppa_l2p, ppa_gc)) {
		spin_lock(&gc_line->lock);
		WARN(!test_bit(paddr_gc, gc_line->invalid_bitmap),
						"pblk: corrupted GC update");
		spin_unlock(&gc_line->lock);
1933

1934 1935 1936 1937
		ret = 0;
		goto out;
	}

1938
	pblk_trans_map_set(pblk, lba, ppa_new);
1939 1940 1941 1942 1943
out:
	spin_unlock(&pblk->trans_lock);
	return ret;
}

1944 1945
void pblk_update_map_dev(struct pblk *pblk, sector_t lba,
			 struct ppa_addr ppa_mapped, struct ppa_addr ppa_cache)
1946
{
1947
	struct ppa_addr ppa_l2p;
1948 1949 1950

#ifdef CONFIG_NVM_DEBUG
	/* Callers must ensure that the ppa points to a device address */
1951
	BUG_ON(pblk_addr_in_cache(ppa_mapped));
1952 1953 1954
#endif
	/* Invalidate and discard padded entries */
	if (lba == ADDR_EMPTY) {
1955
		atomic64_inc(&pblk->pad_wa);
1956 1957 1958
#ifdef CONFIG_NVM_DEBUG
		atomic_long_inc(&pblk->padded_wb);
#endif
1959 1960
		if (!pblk_ppa_empty(ppa_mapped))
			pblk_map_invalidate(pblk, ppa_mapped);
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970
		return;
	}

	/* logic error: lba out-of-bounds. Ignore update */
	if (!(lba < pblk->rl.nr_secs)) {
		WARN(1, "pblk: corrupted L2P map request\n");
		return;
	}

	spin_lock(&pblk->trans_lock);
1971
	ppa_l2p = pblk_trans_map_get(pblk, lba);
1972 1973 1974 1975

	/* Do not update L2P if the cacheline has been updated. In this case,
	 * the mapped ppa must be invalidated
	 */
1976 1977 1978
	if (!pblk_ppa_comp(ppa_l2p, ppa_cache)) {
		if (!pblk_ppa_empty(ppa_mapped))
			pblk_map_invalidate(pblk, ppa_mapped);
1979 1980 1981 1982
		goto out;
	}

#ifdef CONFIG_NVM_DEBUG
1983
	WARN_ON(!pblk_addr_in_cache(ppa_l2p) && !pblk_ppa_empty(ppa_l2p));
1984 1985
#endif

1986
	pblk_trans_map_set(pblk, lba, ppa_mapped);
1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
out:
	spin_unlock(&pblk->trans_lock);
}

void pblk_lookup_l2p_seq(struct pblk *pblk, struct ppa_addr *ppas,
			 sector_t blba, int nr_secs)
{
	int i;

	spin_lock(&pblk->trans_lock);
1997 1998 1999 2000 2001 2002 2003
	for (i = 0; i < nr_secs; i++) {
		struct ppa_addr ppa;

		ppa = ppas[i] = pblk_trans_map_get(pblk, blba + i);

		/* If the L2P entry maps to a line, the reference is valid */
		if (!pblk_ppa_empty(ppa) && !pblk_addr_in_cache(ppa)) {
2004
			int line_id = pblk_ppa_to_line(ppa);
2005 2006 2007 2008 2009
			struct pblk_line *line = &pblk->lines[line_id];

			kref_get(&line->ref);
		}
	}
2010 2011 2012 2013 2014 2015
	spin_unlock(&pblk->trans_lock);
}

void pblk_lookup_l2p_rand(struct pblk *pblk, struct ppa_addr *ppas,
			  u64 *lba_list, int nr_secs)
{
2016
	u64 lba;
2017 2018 2019 2020 2021
	int i;

	spin_lock(&pblk->trans_lock);
	for (i = 0; i < nr_secs; i++) {
		lba = lba_list[i];
2022
		if (lba != ADDR_EMPTY) {
2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
			/* logic error: lba out-of-bounds. Ignore update */
			if (!(lba < pblk->rl.nr_secs)) {
				WARN(1, "pblk: corrupted L2P map request\n");
				continue;
			}
			ppas[i] = pblk_trans_map_get(pblk, lba);
		}
	}
	spin_unlock(&pblk->trans_lock);
}