bio.c 48.4 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
 * Copyright (C) 2001 Jens Axboe <axboe@kernel.dk>
L
Linus Torvalds 已提交
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
 *
 * 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.
 *
 * You should have received a copy of the GNU General Public Licens
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-
 *
 */
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
22
#include <linux/uio.h>
23
#include <linux/iocontext.h>
L
Linus Torvalds 已提交
24 25 26
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/kernel.h>
27
#include <linux/export.h>
L
Linus Torvalds 已提交
28 29
#include <linux/mempool.h>
#include <linux/workqueue.h>
30
#include <linux/cgroup.h>
31
#include <scsi/sg.h>		/* for struct sg_iovec */
L
Linus Torvalds 已提交
32

33
#include <trace/events/block.h>
34

35 36 37 38 39 40
/*
 * Test patch to inline a certain number of bi_io_vec's inside the bio
 * itself, to shrink a bio data allocation from two mempool calls to one
 */
#define BIO_INLINE_VECS		4

D
Denis ChengRq 已提交
41
static mempool_t *bio_split_pool __read_mostly;
L
Linus Torvalds 已提交
42 43 44 45 46 47 48

/*
 * if you change this list, also change bvec_alloc or things will
 * break badly! cannot be bigger than what you can fit into an
 * unsigned short
 */
#define BV(x) { .nr_vecs = x, .name = "biovec-"__stringify(x) }
49
static struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
L
Linus Torvalds 已提交
50 51 52 53 54 55 56 57
	BV(1), BV(4), BV(16), BV(64), BV(128), BV(BIO_MAX_PAGES),
};
#undef BV

/*
 * fs_bio_set is the bio_set containing bio and iovec memory pools used by
 * IO code that does not need private memory pools.
 */
58
struct bio_set *fs_bio_set;
59
EXPORT_SYMBOL(fs_bio_set);
L
Linus Torvalds 已提交
60

61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77
/*
 * Our slab pool management
 */
struct bio_slab {
	struct kmem_cache *slab;
	unsigned int slab_ref;
	unsigned int slab_size;
	char name[8];
};
static DEFINE_MUTEX(bio_slab_lock);
static struct bio_slab *bio_slabs;
static unsigned int bio_slab_nr, bio_slab_max;

static struct kmem_cache *bio_find_or_create_slab(unsigned int extra_size)
{
	unsigned int sz = sizeof(struct bio) + extra_size;
	struct kmem_cache *slab = NULL;
78
	struct bio_slab *bslab, *new_bio_slabs;
79
	unsigned int new_bio_slab_max;
80 81 82 83 84 85
	unsigned int i, entry = -1;

	mutex_lock(&bio_slab_lock);

	i = 0;
	while (i < bio_slab_nr) {
86
		bslab = &bio_slabs[i];
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101

		if (!bslab->slab && entry == -1)
			entry = i;
		else if (bslab->slab_size == sz) {
			slab = bslab->slab;
			bslab->slab_ref++;
			break;
		}
		i++;
	}

	if (slab)
		goto out_unlock;

	if (bio_slab_nr == bio_slab_max && entry == -1) {
102
		new_bio_slab_max = bio_slab_max << 1;
103
		new_bio_slabs = krealloc(bio_slabs,
104
					 new_bio_slab_max * sizeof(struct bio_slab),
105 106
					 GFP_KERNEL);
		if (!new_bio_slabs)
107
			goto out_unlock;
108
		bio_slab_max = new_bio_slab_max;
109
		bio_slabs = new_bio_slabs;
110 111 112 113 114 115 116 117 118 119 120
	}
	if (entry == -1)
		entry = bio_slab_nr++;

	bslab = &bio_slabs[entry];

	snprintf(bslab->name, sizeof(bslab->name), "bio-%d", entry);
	slab = kmem_cache_create(bslab->name, sz, 0, SLAB_HWCACHE_ALIGN, NULL);
	if (!slab)
		goto out_unlock;

121
	printk(KERN_INFO "bio: create slab <%s> at %d\n", bslab->name, entry);
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
	bslab->slab = slab;
	bslab->slab_ref = 1;
	bslab->slab_size = sz;
out_unlock:
	mutex_unlock(&bio_slab_lock);
	return slab;
}

static void bio_put_slab(struct bio_set *bs)
{
	struct bio_slab *bslab = NULL;
	unsigned int i;

	mutex_lock(&bio_slab_lock);

	for (i = 0; i < bio_slab_nr; i++) {
		if (bs->bio_slab == bio_slabs[i].slab) {
			bslab = &bio_slabs[i];
			break;
		}
	}

	if (WARN(!bslab, KERN_ERR "bio: unable to find slab!\n"))
		goto out;

	WARN_ON(!bslab->slab_ref);

	if (--bslab->slab_ref)
		goto out;

	kmem_cache_destroy(bslab->slab);
	bslab->slab = NULL;

out:
	mutex_unlock(&bio_slab_lock);
}

159 160 161 162 163
unsigned int bvec_nr_vecs(unsigned short idx)
{
	return bvec_slabs[idx].nr_vecs;
}

164
void bvec_free(mempool_t *pool, struct bio_vec *bv, unsigned int idx)
165 166 167 168
{
	BIO_BUG_ON(idx >= BIOVEC_NR_POOLS);

	if (idx == BIOVEC_MAX_IDX)
169
		mempool_free(bv, pool);
170 171 172 173 174 175 176
	else {
		struct biovec_slab *bvs = bvec_slabs + idx;

		kmem_cache_free(bvs->slab, bv);
	}
}

177 178
struct bio_vec *bvec_alloc(gfp_t gfp_mask, int nr, unsigned long *idx,
			   mempool_t *pool)
L
Linus Torvalds 已提交
179 180 181
{
	struct bio_vec *bvl;

182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
	/*
	 * see comment near bvec_array define!
	 */
	switch (nr) {
	case 1:
		*idx = 0;
		break;
	case 2 ... 4:
		*idx = 1;
		break;
	case 5 ... 16:
		*idx = 2;
		break;
	case 17 ... 64:
		*idx = 3;
		break;
	case 65 ... 128:
		*idx = 4;
		break;
	case 129 ... BIO_MAX_PAGES:
		*idx = 5;
		break;
	default:
		return NULL;
	}

	/*
	 * idx now points to the pool we want to allocate from. only the
	 * 1-vec entry pool is mempool backed.
	 */
	if (*idx == BIOVEC_MAX_IDX) {
fallback:
214
		bvl = mempool_alloc(pool, gfp_mask);
215 216 217 218
	} else {
		struct biovec_slab *bvs = bvec_slabs + *idx;
		gfp_t __gfp_mask = gfp_mask & ~(__GFP_WAIT | __GFP_IO);

J
Jens Axboe 已提交
219
		/*
220 221 222
		 * Make this allocation restricted and don't dump info on
		 * allocation failures, since we'll fallback to the mempool
		 * in case of failure.
J
Jens Axboe 已提交
223
		 */
224
		__gfp_mask |= __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN;
L
Linus Torvalds 已提交
225

J
Jens Axboe 已提交
226
		/*
227 228
		 * Try a slab allocation. If this fails and __GFP_WAIT
		 * is set, retry with the 1-entry mempool
J
Jens Axboe 已提交
229
		 */
230 231 232 233 234 235 236
		bvl = kmem_cache_alloc(bvs->slab, __gfp_mask);
		if (unlikely(!bvl && (gfp_mask & __GFP_WAIT))) {
			*idx = BIOVEC_MAX_IDX;
			goto fallback;
		}
	}

L
Linus Torvalds 已提交
237 238 239
	return bvl;
}

K
Kent Overstreet 已提交
240
static void __bio_free(struct bio *bio)
L
Linus Torvalds 已提交
241
{
K
Kent Overstreet 已提交
242
	bio_disassociate_task(bio);
L
Linus Torvalds 已提交
243

244
	if (bio_integrity(bio))
245
		bio_integrity_free(bio);
K
Kent Overstreet 已提交
246
}
247

K
Kent Overstreet 已提交
248 249 250 251 252 253 254 255
static void bio_free(struct bio *bio)
{
	struct bio_set *bs = bio->bi_pool;
	void *p;

	__bio_free(bio);

	if (bs) {
256
		if (bio_flagged(bio, BIO_OWNS_VEC))
257
			bvec_free(bs->bvec_pool, bio->bi_io_vec, BIO_POOL_IDX(bio));
K
Kent Overstreet 已提交
258 259 260 261 262

		/*
		 * If we have front padding, adjust the bio pointer before freeing
		 */
		p = bio;
263 264
		p -= bs->front_pad;

K
Kent Overstreet 已提交
265 266 267 268 269
		mempool_free(p, bs->bio_pool);
	} else {
		/* Bio was allocated by bio_kmalloc() */
		kfree(bio);
	}
P
Peter Osterlund 已提交
270 271
}

272
void bio_init(struct bio *bio)
L
Linus Torvalds 已提交
273
{
J
Jens Axboe 已提交
274
	memset(bio, 0, sizeof(*bio));
L
Linus Torvalds 已提交
275 276 277
	bio->bi_flags = 1 << BIO_UPTODATE;
	atomic_set(&bio->bi_cnt, 1);
}
278
EXPORT_SYMBOL(bio_init);
L
Linus Torvalds 已提交
279

K
Kent Overstreet 已提交
280 281 282 283 284 285 286 287 288 289 290 291 292 293
/**
 * bio_reset - reinitialize a bio
 * @bio:	bio to reset
 *
 * Description:
 *   After calling bio_reset(), @bio will be in the same state as a freshly
 *   allocated bio returned bio bio_alloc_bioset() - the only fields that are
 *   preserved are the ones that are initialized by bio_alloc_bioset(). See
 *   comment in struct bio.
 */
void bio_reset(struct bio *bio)
{
	unsigned long flags = bio->bi_flags & (~0UL << BIO_RESET_BITS);

K
Kent Overstreet 已提交
294
	__bio_free(bio);
K
Kent Overstreet 已提交
295 296 297 298 299 300

	memset(bio, 0, BIO_RESET_BYTES);
	bio->bi_flags = flags|(1 << BIO_UPTODATE);
}
EXPORT_SYMBOL(bio_reset);

301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
static void bio_alloc_rescue(struct work_struct *work)
{
	struct bio_set *bs = container_of(work, struct bio_set, rescue_work);
	struct bio *bio;

	while (1) {
		spin_lock(&bs->rescue_lock);
		bio = bio_list_pop(&bs->rescue_list);
		spin_unlock(&bs->rescue_lock);

		if (!bio)
			break;

		generic_make_request(bio);
	}
}

static void punt_bios_to_rescuer(struct bio_set *bs)
{
	struct bio_list punt, nopunt;
	struct bio *bio;

	/*
	 * In order to guarantee forward progress we must punt only bios that
	 * were allocated from this bio_set; otherwise, if there was a bio on
	 * there for a stacking driver higher up in the stack, processing it
	 * could require allocating bios from this bio_set, and doing that from
	 * our own rescuer would be bad.
	 *
	 * Since bio lists are singly linked, pop them all instead of trying to
	 * remove from the middle of the list:
	 */

	bio_list_init(&punt);
	bio_list_init(&nopunt);

	while ((bio = bio_list_pop(current->bio_list)))
		bio_list_add(bio->bi_pool == bs ? &punt : &nopunt, bio);

	*current->bio_list = nopunt;

	spin_lock(&bs->rescue_lock);
	bio_list_merge(&bs->rescue_list, &punt);
	spin_unlock(&bs->rescue_lock);

	queue_work(bs->rescue_workqueue, &bs->rescue_work);
}

L
Linus Torvalds 已提交
349 350 351 352
/**
 * bio_alloc_bioset - allocate a bio for I/O
 * @gfp_mask:   the GFP_ mask given to the slab allocator
 * @nr_iovecs:	number of iovecs to pre-allocate
353
 * @bs:		the bio_set to allocate from.
L
Linus Torvalds 已提交
354 355
 *
 * Description:
356 357 358 359 360 361 362 363 364 365
 *   If @bs is NULL, uses kmalloc() to allocate the bio; else the allocation is
 *   backed by the @bs's mempool.
 *
 *   When @bs is not NULL, if %__GFP_WAIT is set then bio_alloc will always be
 *   able to allocate a bio. This is due to the mempool guarantees. To make this
 *   work, callers must never allocate more than 1 bio at a time from this pool.
 *   Callers that need to allocate more than 1 bio must always submit the
 *   previously allocated bio for IO before attempting to allocate a new one.
 *   Failure to do so can cause deadlocks under memory pressure.
 *
366 367 368 369 370 371 372 373 374 375 376 377 378 379 380
 *   Note that when running under generic_make_request() (i.e. any block
 *   driver), bios are not submitted until after you return - see the code in
 *   generic_make_request() that converts recursion into iteration, to prevent
 *   stack overflows.
 *
 *   This would normally mean allocating multiple bios under
 *   generic_make_request() would be susceptible to deadlocks, but we have
 *   deadlock avoidance code that resubmits any blocked bios from a rescuer
 *   thread.
 *
 *   However, we do not guarantee forward progress for allocations from other
 *   mempools. Doing multiple allocations from the same mempool under
 *   generic_make_request() should be avoided - instead, use bio_set's front_pad
 *   for per bio allocations.
 *
381 382 383
 *   RETURNS:
 *   Pointer to new bio on success, NULL on failure.
 */
A
Al Viro 已提交
384
struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
L
Linus Torvalds 已提交
385
{
386
	gfp_t saved_gfp = gfp_mask;
387 388
	unsigned front_pad;
	unsigned inline_vecs;
T
Tejun Heo 已提交
389
	unsigned long idx = BIO_POOL_NONE;
I
Ingo Molnar 已提交
390
	struct bio_vec *bvl = NULL;
T
Tejun Heo 已提交
391 392 393
	struct bio *bio;
	void *p;

394 395 396 397 398 399 400 401 402 403
	if (!bs) {
		if (nr_iovecs > UIO_MAXIOV)
			return NULL;

		p = kmalloc(sizeof(struct bio) +
			    nr_iovecs * sizeof(struct bio_vec),
			    gfp_mask);
		front_pad = 0;
		inline_vecs = nr_iovecs;
	} else {
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
		/*
		 * generic_make_request() converts recursion to iteration; this
		 * means if we're running beneath it, any bios we allocate and
		 * submit will not be submitted (and thus freed) until after we
		 * return.
		 *
		 * This exposes us to a potential deadlock if we allocate
		 * multiple bios from the same bio_set() while running
		 * underneath generic_make_request(). If we were to allocate
		 * multiple bios (say a stacking block driver that was splitting
		 * bios), we would deadlock if we exhausted the mempool's
		 * reserve.
		 *
		 * We solve this, and guarantee forward progress, with a rescuer
		 * workqueue per bio_set. If we go to allocate and there are
		 * bios on current->bio_list, we first try the allocation
		 * without __GFP_WAIT; if that fails, we punt those bios we
		 * would be blocking to the rescuer workqueue before we retry
		 * with the original gfp_flags.
		 */

		if (current->bio_list && !bio_list_empty(current->bio_list))
			gfp_mask &= ~__GFP_WAIT;

428
		p = mempool_alloc(bs->bio_pool, gfp_mask);
429 430 431 432 433 434
		if (!p && gfp_mask != saved_gfp) {
			punt_bios_to_rescuer(bs);
			gfp_mask = saved_gfp;
			p = mempool_alloc(bs->bio_pool, gfp_mask);
		}

435 436 437 438
		front_pad = bs->front_pad;
		inline_vecs = BIO_INLINE_VECS;
	}

T
Tejun Heo 已提交
439 440
	if (unlikely(!p))
		return NULL;
L
Linus Torvalds 已提交
441

442
	bio = p + front_pad;
I
Ingo Molnar 已提交
443 444
	bio_init(bio);

445
	if (nr_iovecs > inline_vecs) {
446
		bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx, bs->bvec_pool);
447 448 449
		if (!bvl && gfp_mask != saved_gfp) {
			punt_bios_to_rescuer(bs);
			gfp_mask = saved_gfp;
450
			bvl = bvec_alloc(gfp_mask, nr_iovecs, &idx, bs->bvec_pool);
451 452
		}

I
Ingo Molnar 已提交
453 454
		if (unlikely(!bvl))
			goto err_free;
455 456

		bio->bi_flags |= 1 << BIO_OWNS_VEC;
457 458
	} else if (nr_iovecs) {
		bvl = bio->bi_inline_vecs;
L
Linus Torvalds 已提交
459
	}
460 461

	bio->bi_pool = bs;
I
Ingo Molnar 已提交
462 463 464
	bio->bi_flags |= idx << BIO_POOL_OFFSET;
	bio->bi_max_vecs = nr_iovecs;
	bio->bi_io_vec = bvl;
L
Linus Torvalds 已提交
465
	return bio;
I
Ingo Molnar 已提交
466 467

err_free:
T
Tejun Heo 已提交
468
	mempool_free(p, bs->bio_pool);
I
Ingo Molnar 已提交
469
	return NULL;
L
Linus Torvalds 已提交
470
}
471
EXPORT_SYMBOL(bio_alloc_bioset);
L
Linus Torvalds 已提交
472 473 474 475

void zero_fill_bio(struct bio *bio)
{
	unsigned long flags;
476 477
	struct bio_vec bv;
	struct bvec_iter iter;
L
Linus Torvalds 已提交
478

479 480 481 482
	bio_for_each_segment(bv, bio, iter) {
		char *data = bvec_kmap_irq(&bv, &flags);
		memset(data, 0, bv.bv_len);
		flush_dcache_page(bv.bv_page);
L
Linus Torvalds 已提交
483 484 485 486 487 488 489 490 491 492 493
		bvec_kunmap_irq(data, &flags);
	}
}
EXPORT_SYMBOL(zero_fill_bio);

/**
 * bio_put - release a reference to a bio
 * @bio:   bio to release reference to
 *
 * Description:
 *   Put a reference to a &struct bio, either one you have gotten with
494
 *   bio_alloc, bio_get or bio_clone. The last put of a bio will free it.
L
Linus Torvalds 已提交
495 496 497 498 499 500 501 502
 **/
void bio_put(struct bio *bio)
{
	BIO_BUG_ON(!atomic_read(&bio->bi_cnt));

	/*
	 * last put frees it
	 */
K
Kent Overstreet 已提交
503 504
	if (atomic_dec_and_test(&bio->bi_cnt))
		bio_free(bio);
L
Linus Torvalds 已提交
505
}
506
EXPORT_SYMBOL(bio_put);
L
Linus Torvalds 已提交
507

508
inline int bio_phys_segments(struct request_queue *q, struct bio *bio)
L
Linus Torvalds 已提交
509 510 511 512 513 514
{
	if (unlikely(!bio_flagged(bio, BIO_SEG_VALID)))
		blk_recount_segments(q, bio);

	return bio->bi_phys_segments;
}
515
EXPORT_SYMBOL(bio_phys_segments);
L
Linus Torvalds 已提交
516 517 518 519 520 521 522 523 524 525

/**
 * 	__bio_clone	-	clone a bio
 * 	@bio: destination bio
 * 	@bio_src: bio to clone
 *
 *	Clone a &bio. Caller will own the returned bio, but not
 *	the actual data it points to. Reference count of returned
 * 	bio will be one.
 */
526
void __bio_clone(struct bio *bio, struct bio *bio_src)
L
Linus Torvalds 已提交
527
{
528 529
	memcpy(bio->bi_io_vec, bio_src->bi_io_vec,
		bio_src->bi_max_vecs * sizeof(struct bio_vec));
L
Linus Torvalds 已提交
530

531 532 533 534
	/*
	 * most users will be overriding ->bi_bdev with a new target,
	 * so we don't set nor calculate new physical/hw segment counts here
	 */
L
Linus Torvalds 已提交
535 536 537 538
	bio->bi_bdev = bio_src->bi_bdev;
	bio->bi_flags |= 1 << BIO_CLONED;
	bio->bi_rw = bio_src->bi_rw;
	bio->bi_vcnt = bio_src->bi_vcnt;
K
Kent Overstreet 已提交
539
	bio->bi_iter = bio_src->bi_iter;
L
Linus Torvalds 已提交
540
}
541
EXPORT_SYMBOL(__bio_clone);
L
Linus Torvalds 已提交
542 543

/**
544
 *	bio_clone_bioset -	clone a bio
L
Linus Torvalds 已提交
545 546
 *	@bio: bio to clone
 *	@gfp_mask: allocation priority
547
 *	@bs: bio_set to allocate from
L
Linus Torvalds 已提交
548 549 550
 *
 * 	Like __bio_clone, only also allocates the returned bio
 */
551 552
struct bio *bio_clone_bioset(struct bio *bio, gfp_t gfp_mask,
			     struct bio_set *bs)
L
Linus Torvalds 已提交
553
{
554
	struct bio *b;
L
Linus Torvalds 已提交
555

556
	b = bio_alloc_bioset(gfp_mask, bio->bi_max_vecs, bs);
557 558 559 560 561 562 563 564
	if (!b)
		return NULL;

	__bio_clone(b, bio);

	if (bio_integrity(bio)) {
		int ret;

565
		ret = bio_integrity_clone(b, bio, gfp_mask);
566

L
Li Zefan 已提交
567 568
		if (ret < 0) {
			bio_put(b);
569
			return NULL;
L
Li Zefan 已提交
570
		}
P
Peter Osterlund 已提交
571
	}
L
Linus Torvalds 已提交
572 573 574

	return b;
}
575
EXPORT_SYMBOL(bio_clone_bioset);
L
Linus Torvalds 已提交
576 577 578 579 580 581 582 583 584 585 586 587

/**
 *	bio_get_nr_vecs		- return approx number of vecs
 *	@bdev:  I/O target
 *
 *	Return the approximate number of pages we can send to this target.
 *	There's no guarantee that you will be able to fit this number of pages
 *	into a bio, it does not account for dynamic restrictions that vary
 *	on offset.
 */
int bio_get_nr_vecs(struct block_device *bdev)
{
588
	struct request_queue *q = bdev_get_queue(bdev);
589 590 591
	int nr_pages;

	nr_pages = min_t(unsigned,
592 593
		     queue_max_segments(q),
		     queue_max_sectors(q) / (PAGE_SIZE >> 9) + 1);
594 595 596

	return min_t(unsigned, nr_pages, BIO_MAX_PAGES);

L
Linus Torvalds 已提交
597
}
598
EXPORT_SYMBOL(bio_get_nr_vecs);
L
Linus Torvalds 已提交
599

600
static int __bio_add_page(struct request_queue *q, struct bio *bio, struct page
601
			  *page, unsigned int len, unsigned int offset,
602
			  unsigned int max_sectors)
L
Linus Torvalds 已提交
603 604 605 606 607 608 609 610 611 612
{
	int retried_segments = 0;
	struct bio_vec *bvec;

	/*
	 * cloned bio must not modify vec list
	 */
	if (unlikely(bio_flagged(bio, BIO_CLONED)))
		return 0;

613
	if (((bio->bi_iter.bi_size + len) >> 9) > max_sectors)
L
Linus Torvalds 已提交
614 615
		return 0;

616 617 618 619 620 621 622 623 624 625
	/*
	 * For filesystems with a blocksize smaller than the pagesize
	 * we will often be called with the same page as last time and
	 * a consecutive offset.  Optimize this special case.
	 */
	if (bio->bi_vcnt > 0) {
		struct bio_vec *prev = &bio->bi_io_vec[bio->bi_vcnt - 1];

		if (page == prev->bv_page &&
		    offset == prev->bv_offset + prev->bv_len) {
626
			unsigned int prev_bv_len = prev->bv_len;
627
			prev->bv_len += len;
628 629 630

			if (q->merge_bvec_fn) {
				struct bvec_merge_data bvm = {
631 632 633 634
					/* prev_bvec is already charged in
					   bi_size, discharge it in order to
					   simulate merging updated prev_bvec
					   as new bvec. */
635
					.bi_bdev = bio->bi_bdev,
636 637 638
					.bi_sector = bio->bi_iter.bi_sector,
					.bi_size = bio->bi_iter.bi_size -
						prev_bv_len,
639 640 641
					.bi_rw = bio->bi_rw,
				};

642
				if (q->merge_bvec_fn(q, &bvm, prev) < prev->bv_len) {
643 644 645
					prev->bv_len -= len;
					return 0;
				}
646 647 648 649 650 651 652
			}

			goto done;
		}
	}

	if (bio->bi_vcnt >= bio->bi_max_vecs)
L
Linus Torvalds 已提交
653 654 655 656 657 658 659
		return 0;

	/*
	 * we might lose a segment or two here, but rather that than
	 * make this too complex.
	 */

660
	while (bio->bi_phys_segments >= queue_max_segments(q)) {
L
Linus Torvalds 已提交
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683

		if (retried_segments)
			return 0;

		retried_segments = 1;
		blk_recount_segments(q, bio);
	}

	/*
	 * setup the new entry, we might clear it again later if we
	 * cannot add the page
	 */
	bvec = &bio->bi_io_vec[bio->bi_vcnt];
	bvec->bv_page = page;
	bvec->bv_len = len;
	bvec->bv_offset = offset;

	/*
	 * if queue has other restrictions (eg varying max sector size
	 * depending on offset), it can specify a merge_bvec_fn in the
	 * queue to get further control
	 */
	if (q->merge_bvec_fn) {
684 685
		struct bvec_merge_data bvm = {
			.bi_bdev = bio->bi_bdev,
686 687
			.bi_sector = bio->bi_iter.bi_sector,
			.bi_size = bio->bi_iter.bi_size,
688 689 690
			.bi_rw = bio->bi_rw,
		};

L
Linus Torvalds 已提交
691 692 693 694
		/*
		 * merge_bvec_fn() returns number of bytes it can accept
		 * at this offset
		 */
695
		if (q->merge_bvec_fn(q, &bvm, bvec) < bvec->bv_len) {
L
Linus Torvalds 已提交
696 697 698 699 700 701 702 703
			bvec->bv_page = NULL;
			bvec->bv_len = 0;
			bvec->bv_offset = 0;
			return 0;
		}
	}

	/* If we may be able to merge these biovecs, force a recount */
704
	if (bio->bi_vcnt && (BIOVEC_PHYS_MERGEABLE(bvec-1, bvec)))
L
Linus Torvalds 已提交
705 706 707 708
		bio->bi_flags &= ~(1 << BIO_SEG_VALID);

	bio->bi_vcnt++;
	bio->bi_phys_segments++;
709
 done:
710
	bio->bi_iter.bi_size += len;
L
Linus Torvalds 已提交
711 712 713
	return len;
}

714 715
/**
 *	bio_add_pc_page	-	attempt to add page to bio
J
Jens Axboe 已提交
716
 *	@q: the target queue
717 718 719 720 721 722
 *	@bio: destination bio
 *	@page: page to add
 *	@len: vec entry length
 *	@offset: vec entry offset
 *
 *	Attempt to add a page to the bio_vec maplist. This can fail for a
723 724 725 726 727
 *	number of reasons, such as the bio being full or target block device
 *	limitations. The target block device must allow bio's up to PAGE_SIZE,
 *	so it is always possible to add a single page to an empty bio.
 *
 *	This should only be used by REQ_PC bios.
728
 */
729
int bio_add_pc_page(struct request_queue *q, struct bio *bio, struct page *page,
730 731
		    unsigned int len, unsigned int offset)
{
732 733
	return __bio_add_page(q, bio, page, len, offset,
			      queue_max_hw_sectors(q));
734
}
735
EXPORT_SYMBOL(bio_add_pc_page);
736

L
Linus Torvalds 已提交
737 738 739 740 741 742 743 744
/**
 *	bio_add_page	-	attempt to add page to bio
 *	@bio: destination bio
 *	@page: page to add
 *	@len: vec entry length
 *	@offset: vec entry offset
 *
 *	Attempt to add a page to the bio_vec maplist. This can fail for a
745 746 747
 *	number of reasons, such as the bio being full or target block device
 *	limitations. The target block device must allow bio's up to PAGE_SIZE,
 *	so it is always possible to add a single page to an empty bio.
L
Linus Torvalds 已提交
748 749 750 751
 */
int bio_add_page(struct bio *bio, struct page *page, unsigned int len,
		 unsigned int offset)
{
752
	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
753
	return __bio_add_page(q, bio, page, len, offset, queue_max_sectors(q));
L
Linus Torvalds 已提交
754
}
755
EXPORT_SYMBOL(bio_add_page);
L
Linus Torvalds 已提交
756

757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
struct submit_bio_ret {
	struct completion event;
	int error;
};

static void submit_bio_wait_endio(struct bio *bio, int error)
{
	struct submit_bio_ret *ret = bio->bi_private;

	ret->error = error;
	complete(&ret->event);
}

/**
 * submit_bio_wait - submit a bio, and wait until it completes
 * @rw: whether to %READ or %WRITE, or maybe to %READA (read ahead)
 * @bio: The &struct bio which describes the I/O
 *
 * Simple wrapper around submit_bio(). Returns 0 on success, or the error from
 * bio_endio() on failure.
 */
int submit_bio_wait(int rw, struct bio *bio)
{
	struct submit_bio_ret ret;

	rw |= REQ_SYNC;
	init_completion(&ret.event);
	bio->bi_private = &ret;
	bio->bi_end_io = submit_bio_wait_endio;
	submit_bio(rw, bio);
	wait_for_completion(&ret.event);

	return ret.error;
}
EXPORT_SYMBOL(submit_bio_wait);

K
Kent Overstreet 已提交
793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808
/**
 * bio_advance - increment/complete a bio by some number of bytes
 * @bio:	bio to advance
 * @bytes:	number of bytes to complete
 *
 * This updates bi_sector, bi_size and bi_idx; if the number of bytes to
 * complete doesn't align with a bvec boundary, then bv_len and bv_offset will
 * be updated on the last bvec as well.
 *
 * @bio will then represent the remaining, uncompleted portion of the io.
 */
void bio_advance(struct bio *bio, unsigned bytes)
{
	if (bio_integrity(bio))
		bio_integrity_advance(bio, bytes);

K
Kent Overstreet 已提交
809
	bio_advance_iter(bio, &bio->bi_iter, bytes);
K
Kent Overstreet 已提交
810 811 812
}
EXPORT_SYMBOL(bio_advance);

K
Kent Overstreet 已提交
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840
/**
 * bio_alloc_pages - allocates a single page for each bvec in a bio
 * @bio: bio to allocate pages for
 * @gfp_mask: flags for allocation
 *
 * Allocates pages up to @bio->bi_vcnt.
 *
 * Returns 0 on success, -ENOMEM on failure. On failure, any allocated pages are
 * freed.
 */
int bio_alloc_pages(struct bio *bio, gfp_t gfp_mask)
{
	int i;
	struct bio_vec *bv;

	bio_for_each_segment_all(bv, bio, i) {
		bv->bv_page = alloc_page(gfp_mask);
		if (!bv->bv_page) {
			while (--bv >= bio->bi_io_vec)
				__free_page(bv->bv_page);
			return -ENOMEM;
		}
	}

	return 0;
}
EXPORT_SYMBOL(bio_alloc_pages);

K
Kent Overstreet 已提交
841 842 843 844 845 846 847 848 849 850 851 852 853 854
/**
 * bio_copy_data - copy contents of data buffers from one chain of bios to
 * another
 * @src: source bio list
 * @dst: destination bio list
 *
 * If @src and @dst are single bios, bi_next must be NULL - otherwise, treats
 * @src and @dst as linked lists of bios.
 *
 * Stops when it reaches the end of either @src or @dst - that is, copies
 * min(src->bi_size, dst->bi_size) bytes (or the equivalent for lists of bios).
 */
void bio_copy_data(struct bio *dst, struct bio *src)
{
855 856
	struct bvec_iter src_iter, dst_iter;
	struct bio_vec src_bv, dst_bv;
K
Kent Overstreet 已提交
857
	void *src_p, *dst_p;
858
	unsigned bytes;
K
Kent Overstreet 已提交
859

860 861
	src_iter = src->bi_iter;
	dst_iter = dst->bi_iter;
K
Kent Overstreet 已提交
862 863

	while (1) {
864 865 866 867
		if (!src_iter.bi_size) {
			src = src->bi_next;
			if (!src)
				break;
K
Kent Overstreet 已提交
868

869
			src_iter = src->bi_iter;
K
Kent Overstreet 已提交
870 871
		}

872 873 874 875
		if (!dst_iter.bi_size) {
			dst = dst->bi_next;
			if (!dst)
				break;
K
Kent Overstreet 已提交
876

877
			dst_iter = dst->bi_iter;
K
Kent Overstreet 已提交
878 879
		}

880 881 882 883
		src_bv = bio_iter_iovec(src, src_iter);
		dst_bv = bio_iter_iovec(dst, dst_iter);

		bytes = min(src_bv.bv_len, dst_bv.bv_len);
K
Kent Overstreet 已提交
884

885 886
		src_p = kmap_atomic(src_bv.bv_page);
		dst_p = kmap_atomic(dst_bv.bv_page);
K
Kent Overstreet 已提交
887

888 889
		memcpy(dst_p + dst_bv.bv_offset,
		       src_p + src_bv.bv_offset,
K
Kent Overstreet 已提交
890 891 892 893 894
		       bytes);

		kunmap_atomic(dst_p);
		kunmap_atomic(src_p);

895 896
		bio_advance_iter(src, &src_iter, bytes);
		bio_advance_iter(dst, &dst_iter, bytes);
K
Kent Overstreet 已提交
897 898 899 900
	}
}
EXPORT_SYMBOL(bio_copy_data);

L
Linus Torvalds 已提交
901 902
struct bio_map_data {
	struct bio_vec *iovecs;
903
	struct sg_iovec *sgvecs;
904 905
	int nr_sgvecs;
	int is_our_pages;
L
Linus Torvalds 已提交
906 907
};

908
static void bio_set_map_data(struct bio_map_data *bmd, struct bio *bio,
909 910
			     struct sg_iovec *iov, int iov_count,
			     int is_our_pages)
L
Linus Torvalds 已提交
911 912
{
	memcpy(bmd->iovecs, bio->bi_io_vec, sizeof(struct bio_vec) * bio->bi_vcnt);
913 914
	memcpy(bmd->sgvecs, iov, sizeof(struct sg_iovec) * iov_count);
	bmd->nr_sgvecs = iov_count;
915
	bmd->is_our_pages = is_our_pages;
L
Linus Torvalds 已提交
916 917 918 919 920 921
	bio->bi_private = bmd;
}

static void bio_free_map_data(struct bio_map_data *bmd)
{
	kfree(bmd->iovecs);
922
	kfree(bmd->sgvecs);
L
Linus Torvalds 已提交
923 924 925
	kfree(bmd);
}

926 927
static struct bio_map_data *bio_alloc_map_data(int nr_segs,
					       unsigned int iov_count,
928
					       gfp_t gfp_mask)
L
Linus Torvalds 已提交
929
{
930 931 932 933
	struct bio_map_data *bmd;

	if (iov_count > UIO_MAXIOV)
		return NULL;
L
Linus Torvalds 已提交
934

935
	bmd = kmalloc(sizeof(*bmd), gfp_mask);
L
Linus Torvalds 已提交
936 937 938
	if (!bmd)
		return NULL;

939
	bmd->iovecs = kmalloc(sizeof(struct bio_vec) * nr_segs, gfp_mask);
940 941 942 943 944
	if (!bmd->iovecs) {
		kfree(bmd);
		return NULL;
	}

945
	bmd->sgvecs = kmalloc(sizeof(struct sg_iovec) * iov_count, gfp_mask);
946
	if (bmd->sgvecs)
L
Linus Torvalds 已提交
947 948
		return bmd;

949
	kfree(bmd->iovecs);
L
Linus Torvalds 已提交
950 951 952 953
	kfree(bmd);
	return NULL;
}

954
static int __bio_copy_iov(struct bio *bio, struct bio_vec *iovecs,
955 956
			  struct sg_iovec *iov, int iov_count,
			  int to_user, int from_user, int do_free_page)
957 958 959 960 961 962
{
	int ret = 0, i;
	struct bio_vec *bvec;
	int iov_idx = 0;
	unsigned int iov_off = 0;

963
	bio_for_each_segment_all(bvec, bio, i) {
964
		char *bv_addr = page_address(bvec->bv_page);
965
		unsigned int bv_len = iovecs[i].bv_len;
966 967 968

		while (bv_len && iov_idx < iov_count) {
			unsigned int bytes;
969
			char __user *iov_addr;
970 971 972 973 974 975

			bytes = min_t(unsigned int,
				      iov[iov_idx].iov_len - iov_off, bv_len);
			iov_addr = iov[iov_idx].iov_base + iov_off;

			if (!ret) {
976
				if (to_user)
977 978 979
					ret = copy_to_user(iov_addr, bv_addr,
							   bytes);

980 981 982 983
				if (from_user)
					ret = copy_from_user(bv_addr, iov_addr,
							     bytes);

984 985 986 987 988 989 990 991 992 993 994 995 996 997 998
				if (ret)
					ret = -EFAULT;
			}

			bv_len -= bytes;
			bv_addr += bytes;
			iov_addr += bytes;
			iov_off += bytes;

			if (iov[iov_idx].iov_len == iov_off) {
				iov_idx++;
				iov_off = 0;
			}
		}

999
		if (do_free_page)
1000 1001 1002 1003 1004 1005
			__free_page(bvec->bv_page);
	}

	return ret;
}

L
Linus Torvalds 已提交
1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
/**
 *	bio_uncopy_user	-	finish previously mapped bio
 *	@bio: bio being terminated
 *
 *	Free pages allocated from bio_copy_user() and write back data
 *	to user space in case of a read.
 */
int bio_uncopy_user(struct bio *bio)
{
	struct bio_map_data *bmd = bio->bi_private;
1016 1017
	struct bio_vec *bvec;
	int ret = 0, i;
L
Linus Torvalds 已提交
1018

1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031
	if (!bio_flagged(bio, BIO_NULL_MAPPED)) {
		/*
		 * if we're in a workqueue, the request is orphaned, so
		 * don't copy into a random user address space, just free.
		 */
		if (current->mm)
			ret = __bio_copy_iov(bio, bmd->iovecs, bmd->sgvecs,
					     bmd->nr_sgvecs, bio_data_dir(bio) == READ,
					     0, bmd->is_our_pages);
		else if (bmd->is_our_pages)
			bio_for_each_segment_all(bvec, bio, i)
				__free_page(bvec->bv_page);
	}
L
Linus Torvalds 已提交
1032 1033 1034 1035
	bio_free_map_data(bmd);
	bio_put(bio);
	return ret;
}
1036
EXPORT_SYMBOL(bio_uncopy_user);
L
Linus Torvalds 已提交
1037 1038

/**
1039
 *	bio_copy_user_iov	-	copy user data to bio
L
Linus Torvalds 已提交
1040
 *	@q: destination block queue
1041
 *	@map_data: pointer to the rq_map_data holding pages (if necessary)
1042 1043
 *	@iov:	the iovec.
 *	@iov_count: number of elements in the iovec
L
Linus Torvalds 已提交
1044
 *	@write_to_vm: bool indicating writing to pages or not
1045
 *	@gfp_mask: memory allocation flags
L
Linus Torvalds 已提交
1046 1047 1048 1049 1050
 *
 *	Prepares and returns a bio for indirect user io, bouncing data
 *	to/from kernel pages as necessary. Must be paired with
 *	call bio_uncopy_user() on io completion.
 */
1051 1052 1053 1054
struct bio *bio_copy_user_iov(struct request_queue *q,
			      struct rq_map_data *map_data,
			      struct sg_iovec *iov, int iov_count,
			      int write_to_vm, gfp_t gfp_mask)
L
Linus Torvalds 已提交
1055 1056 1057 1058 1059 1060
{
	struct bio_map_data *bmd;
	struct bio_vec *bvec;
	struct page *page;
	struct bio *bio;
	int i, ret;
1061 1062
	int nr_pages = 0;
	unsigned int len = 0;
1063
	unsigned int offset = map_data ? map_data->offset & ~PAGE_MASK : 0;
L
Linus Torvalds 已提交
1064

1065 1066 1067 1068 1069 1070 1071 1072 1073
	for (i = 0; i < iov_count; i++) {
		unsigned long uaddr;
		unsigned long end;
		unsigned long start;

		uaddr = (unsigned long)iov[i].iov_base;
		end = (uaddr + iov[i].iov_len + PAGE_SIZE - 1) >> PAGE_SHIFT;
		start = uaddr >> PAGE_SHIFT;

1074 1075 1076 1077 1078 1079
		/*
		 * Overflow, abort
		 */
		if (end < start)
			return ERR_PTR(-EINVAL);

1080 1081 1082 1083
		nr_pages += end - start;
		len += iov[i].iov_len;
	}

1084 1085 1086
	if (offset)
		nr_pages++;

1087
	bmd = bio_alloc_map_data(nr_pages, iov_count, gfp_mask);
L
Linus Torvalds 已提交
1088 1089 1090 1091
	if (!bmd)
		return ERR_PTR(-ENOMEM);

	ret = -ENOMEM;
1092
	bio = bio_kmalloc(gfp_mask, nr_pages);
L
Linus Torvalds 已提交
1093 1094 1095
	if (!bio)
		goto out_bmd;

1096 1097
	if (!write_to_vm)
		bio->bi_rw |= REQ_WRITE;
L
Linus Torvalds 已提交
1098 1099

	ret = 0;
1100 1101

	if (map_data) {
1102
		nr_pages = 1 << map_data->page_order;
1103 1104
		i = map_data->offset / PAGE_SIZE;
	}
L
Linus Torvalds 已提交
1105
	while (len) {
1106
		unsigned int bytes = PAGE_SIZE;
L
Linus Torvalds 已提交
1107

1108 1109
		bytes -= offset;

L
Linus Torvalds 已提交
1110 1111 1112
		if (bytes > len)
			bytes = len;

1113
		if (map_data) {
1114
			if (i == map_data->nr_entries * nr_pages) {
1115 1116 1117
				ret = -ENOMEM;
				break;
			}
1118 1119 1120 1121 1122 1123

			page = map_data->pages[i / nr_pages];
			page += (i % nr_pages);

			i++;
		} else {
1124
			page = alloc_page(q->bounce_gfp | gfp_mask);
1125 1126 1127 1128
			if (!page) {
				ret = -ENOMEM;
				break;
			}
L
Linus Torvalds 已提交
1129 1130
		}

1131
		if (bio_add_pc_page(q, bio, page, bytes, offset) < bytes)
L
Linus Torvalds 已提交
1132 1133 1134
			break;

		len -= bytes;
1135
		offset = 0;
L
Linus Torvalds 已提交
1136 1137 1138 1139 1140 1141 1142 1143
	}

	if (ret)
		goto cleanup;

	/*
	 * success
	 */
1144 1145 1146
	if ((!write_to_vm && (!map_data || !map_data->null_mapped)) ||
	    (map_data && map_data->from_user)) {
		ret = __bio_copy_iov(bio, bio->bi_io_vec, iov, iov_count, 0, 1, 0);
1147 1148
		if (ret)
			goto cleanup;
L
Linus Torvalds 已提交
1149 1150
	}

1151
	bio_set_map_data(bmd, bio, iov, iov_count, map_data ? 0 : 1);
L
Linus Torvalds 已提交
1152 1153
	return bio;
cleanup:
1154
	if (!map_data)
1155
		bio_for_each_segment_all(bvec, bio, i)
1156
			__free_page(bvec->bv_page);
L
Linus Torvalds 已提交
1157 1158 1159 1160 1161 1162 1163

	bio_put(bio);
out_bmd:
	bio_free_map_data(bmd);
	return ERR_PTR(ret);
}

1164 1165 1166
/**
 *	bio_copy_user	-	copy user data to bio
 *	@q: destination block queue
1167
 *	@map_data: pointer to the rq_map_data holding pages (if necessary)
1168 1169 1170
 *	@uaddr: start of user address
 *	@len: length in bytes
 *	@write_to_vm: bool indicating writing to pages or not
1171
 *	@gfp_mask: memory allocation flags
1172 1173 1174 1175 1176
 *
 *	Prepares and returns a bio for indirect user io, bouncing data
 *	to/from kernel pages as necessary. Must be paired with
 *	call bio_uncopy_user() on io completion.
 */
1177 1178 1179
struct bio *bio_copy_user(struct request_queue *q, struct rq_map_data *map_data,
			  unsigned long uaddr, unsigned int len,
			  int write_to_vm, gfp_t gfp_mask)
1180 1181 1182 1183 1184 1185
{
	struct sg_iovec iov;

	iov.iov_base = (void __user *)uaddr;
	iov.iov_len = len;

1186
	return bio_copy_user_iov(q, map_data, &iov, 1, write_to_vm, gfp_mask);
1187
}
1188
EXPORT_SYMBOL(bio_copy_user);
1189

1190
static struct bio *__bio_map_user_iov(struct request_queue *q,
1191 1192
				      struct block_device *bdev,
				      struct sg_iovec *iov, int iov_count,
1193
				      int write_to_vm, gfp_t gfp_mask)
L
Linus Torvalds 已提交
1194
{
1195 1196
	int i, j;
	int nr_pages = 0;
L
Linus Torvalds 已提交
1197 1198
	struct page **pages;
	struct bio *bio;
1199 1200
	int cur_page = 0;
	int ret, offset;
L
Linus Torvalds 已提交
1201

1202 1203 1204 1205 1206 1207
	for (i = 0; i < iov_count; i++) {
		unsigned long uaddr = (unsigned long)iov[i].iov_base;
		unsigned long len = iov[i].iov_len;
		unsigned long end = (uaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
		unsigned long start = uaddr >> PAGE_SHIFT;

1208 1209 1210 1211 1212 1213
		/*
		 * Overflow, abort
		 */
		if (end < start)
			return ERR_PTR(-EINVAL);

1214 1215
		nr_pages += end - start;
		/*
1216
		 * buffer must be aligned to at least hardsector size for now
1217
		 */
1218
		if (uaddr & queue_dma_alignment(q))
1219 1220 1221 1222
			return ERR_PTR(-EINVAL);
	}

	if (!nr_pages)
L
Linus Torvalds 已提交
1223 1224
		return ERR_PTR(-EINVAL);

1225
	bio = bio_kmalloc(gfp_mask, nr_pages);
L
Linus Torvalds 已提交
1226 1227 1228 1229
	if (!bio)
		return ERR_PTR(-ENOMEM);

	ret = -ENOMEM;
1230
	pages = kcalloc(nr_pages, sizeof(struct page *), gfp_mask);
L
Linus Torvalds 已提交
1231 1232 1233
	if (!pages)
		goto out;

1234 1235 1236 1237 1238 1239 1240
	for (i = 0; i < iov_count; i++) {
		unsigned long uaddr = (unsigned long)iov[i].iov_base;
		unsigned long len = iov[i].iov_len;
		unsigned long end = (uaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
		unsigned long start = uaddr >> PAGE_SHIFT;
		const int local_nr_pages = end - start;
		const int page_limit = cur_page + local_nr_pages;
1241

N
Nick Piggin 已提交
1242 1243
		ret = get_user_pages_fast(uaddr, local_nr_pages,
				write_to_vm, &pages[cur_page]);
1244 1245
		if (ret < local_nr_pages) {
			ret = -EFAULT;
1246
			goto out_unmap;
1247
		}
1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261

		offset = uaddr & ~PAGE_MASK;
		for (j = cur_page; j < page_limit; j++) {
			unsigned int bytes = PAGE_SIZE - offset;

			if (len <= 0)
				break;
			
			if (bytes > len)
				bytes = len;

			/*
			 * sorry...
			 */
1262 1263
			if (bio_add_pc_page(q, bio, pages[j], bytes, offset) <
					    bytes)
1264 1265 1266 1267 1268
				break;

			len -= bytes;
			offset = 0;
		}
L
Linus Torvalds 已提交
1269

1270
		cur_page = j;
L
Linus Torvalds 已提交
1271
		/*
1272
		 * release the pages we didn't map into the bio, if any
L
Linus Torvalds 已提交
1273
		 */
1274 1275
		while (j < page_limit)
			page_cache_release(pages[j++]);
L
Linus Torvalds 已提交
1276 1277 1278 1279 1280 1281 1282 1283
	}

	kfree(pages);

	/*
	 * set data direction, and check if mapped pages need bouncing
	 */
	if (!write_to_vm)
1284
		bio->bi_rw |= REQ_WRITE;
L
Linus Torvalds 已提交
1285

1286
	bio->bi_bdev = bdev;
L
Linus Torvalds 已提交
1287 1288
	bio->bi_flags |= (1 << BIO_USER_MAPPED);
	return bio;
1289 1290 1291 1292 1293 1294 1295 1296

 out_unmap:
	for (i = 0; i < nr_pages; i++) {
		if(!pages[i])
			break;
		page_cache_release(pages[i]);
	}
 out:
L
Linus Torvalds 已提交
1297 1298 1299 1300 1301 1302 1303
	kfree(pages);
	bio_put(bio);
	return ERR_PTR(ret);
}

/**
 *	bio_map_user	-	map user address into bio
1304
 *	@q: the struct request_queue for the bio
L
Linus Torvalds 已提交
1305 1306 1307 1308
 *	@bdev: destination block device
 *	@uaddr: start of user address
 *	@len: length in bytes
 *	@write_to_vm: bool indicating writing to pages or not
1309
 *	@gfp_mask: memory allocation flags
L
Linus Torvalds 已提交
1310 1311 1312 1313
 *
 *	Map the user space address into a bio suitable for io to a block
 *	device. Returns an error pointer in case of error.
 */
1314
struct bio *bio_map_user(struct request_queue *q, struct block_device *bdev,
1315 1316
			 unsigned long uaddr, unsigned int len, int write_to_vm,
			 gfp_t gfp_mask)
1317 1318 1319
{
	struct sg_iovec iov;

1320
	iov.iov_base = (void __user *)uaddr;
1321 1322
	iov.iov_len = len;

1323
	return bio_map_user_iov(q, bdev, &iov, 1, write_to_vm, gfp_mask);
1324
}
1325
EXPORT_SYMBOL(bio_map_user);
1326 1327 1328

/**
 *	bio_map_user_iov - map user sg_iovec table into bio
1329
 *	@q: the struct request_queue for the bio
1330 1331 1332 1333
 *	@bdev: destination block device
 *	@iov:	the iovec.
 *	@iov_count: number of elements in the iovec
 *	@write_to_vm: bool indicating writing to pages or not
1334
 *	@gfp_mask: memory allocation flags
1335 1336 1337 1338
 *
 *	Map the user space address into a bio suitable for io to a block
 *	device. Returns an error pointer in case of error.
 */
1339
struct bio *bio_map_user_iov(struct request_queue *q, struct block_device *bdev,
1340
			     struct sg_iovec *iov, int iov_count,
1341
			     int write_to_vm, gfp_t gfp_mask)
L
Linus Torvalds 已提交
1342 1343 1344
{
	struct bio *bio;

1345 1346
	bio = __bio_map_user_iov(q, bdev, iov, iov_count, write_to_vm,
				 gfp_mask);
L
Linus Torvalds 已提交
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
	if (IS_ERR(bio))
		return bio;

	/*
	 * subtle -- if __bio_map_user() ended up bouncing a bio,
	 * it would normally disappear when its bi_end_io is run.
	 * however, we need it for the unmap, so grab an extra
	 * reference to it
	 */
	bio_get(bio);

1358
	return bio;
L
Linus Torvalds 已提交
1359 1360 1361 1362 1363 1364 1365 1366 1367 1368
}

static void __bio_unmap_user(struct bio *bio)
{
	struct bio_vec *bvec;
	int i;

	/*
	 * make sure we dirty pages we wrote to
	 */
1369
	bio_for_each_segment_all(bvec, bio, i) {
L
Linus Torvalds 已提交
1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392
		if (bio_data_dir(bio) == READ)
			set_page_dirty_lock(bvec->bv_page);

		page_cache_release(bvec->bv_page);
	}

	bio_put(bio);
}

/**
 *	bio_unmap_user	-	unmap a bio
 *	@bio:		the bio being unmapped
 *
 *	Unmap a bio previously mapped by bio_map_user(). Must be called with
 *	a process context.
 *
 *	bio_unmap_user() may sleep.
 */
void bio_unmap_user(struct bio *bio)
{
	__bio_unmap_user(bio);
	bio_put(bio);
}
1393
EXPORT_SYMBOL(bio_unmap_user);
L
Linus Torvalds 已提交
1394

1395
static void bio_map_kern_endio(struct bio *bio, int err)
1396 1397 1398 1399
{
	bio_put(bio);
}

1400
static struct bio *__bio_map_kern(struct request_queue *q, void *data,
A
Al Viro 已提交
1401
				  unsigned int len, gfp_t gfp_mask)
M
Mike Christie 已提交
1402 1403 1404 1405 1406 1407 1408 1409
{
	unsigned long kaddr = (unsigned long)data;
	unsigned long end = (kaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
	unsigned long start = kaddr >> PAGE_SHIFT;
	const int nr_pages = end - start;
	int offset, i;
	struct bio *bio;

1410
	bio = bio_kmalloc(gfp_mask, nr_pages);
M
Mike Christie 已提交
1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
	if (!bio)
		return ERR_PTR(-ENOMEM);

	offset = offset_in_page(kaddr);
	for (i = 0; i < nr_pages; i++) {
		unsigned int bytes = PAGE_SIZE - offset;

		if (len <= 0)
			break;

		if (bytes > len)
			bytes = len;

1424 1425
		if (bio_add_pc_page(q, bio, virt_to_page(data), bytes,
				    offset) < bytes)
M
Mike Christie 已提交
1426 1427 1428 1429 1430 1431 1432
			break;

		data += bytes;
		len -= bytes;
		offset = 0;
	}

1433
	bio->bi_end_io = bio_map_kern_endio;
M
Mike Christie 已提交
1434 1435 1436 1437 1438
	return bio;
}

/**
 *	bio_map_kern	-	map kernel address into bio
1439
 *	@q: the struct request_queue for the bio
M
Mike Christie 已提交
1440 1441 1442 1443 1444 1445 1446
 *	@data: pointer to buffer to map
 *	@len: length in bytes
 *	@gfp_mask: allocation flags for bio allocation
 *
 *	Map the kernel address into a bio suitable for io to a block
 *	device. Returns an error pointer in case of error.
 */
1447
struct bio *bio_map_kern(struct request_queue *q, void *data, unsigned int len,
A
Al Viro 已提交
1448
			 gfp_t gfp_mask)
M
Mike Christie 已提交
1449 1450 1451 1452 1453 1454 1455
{
	struct bio *bio;

	bio = __bio_map_kern(q, data, len, gfp_mask);
	if (IS_ERR(bio))
		return bio;

1456
	if (bio->bi_iter.bi_size == len)
M
Mike Christie 已提交
1457 1458 1459 1460 1461 1462 1463 1464
		return bio;

	/*
	 * Don't support partial mappings.
	 */
	bio_put(bio);
	return ERR_PTR(-EINVAL);
}
1465
EXPORT_SYMBOL(bio_map_kern);
M
Mike Christie 已提交
1466

1467 1468 1469 1470
static void bio_copy_kern_endio(struct bio *bio, int err)
{
	struct bio_vec *bvec;
	const int read = bio_data_dir(bio) == READ;
1471
	struct bio_map_data *bmd = bio->bi_private;
1472
	int i;
1473
	char *p = bmd->sgvecs[0].iov_base;
1474

1475
	bio_for_each_segment_all(bvec, bio, i) {
1476
		char *addr = page_address(bvec->bv_page);
1477
		int len = bmd->iovecs[i].bv_len;
1478

1479
		if (read)
1480
			memcpy(p, addr, len);
1481 1482

		__free_page(bvec->bv_page);
1483
		p += len;
1484 1485
	}

1486
	bio_free_map_data(bmd);
1487 1488 1489 1490 1491 1492 1493 1494 1495
	bio_put(bio);
}

/**
 *	bio_copy_kern	-	copy kernel address into bio
 *	@q: the struct request_queue for the bio
 *	@data: pointer to buffer to copy
 *	@len: length in bytes
 *	@gfp_mask: allocation flags for bio and page allocation
1496
 *	@reading: data direction is READ
1497 1498 1499 1500 1501 1502 1503 1504 1505
 *
 *	copy the kernel address into a bio suitable for io to a block
 *	device. Returns an error pointer in case of error.
 */
struct bio *bio_copy_kern(struct request_queue *q, void *data, unsigned int len,
			  gfp_t gfp_mask, int reading)
{
	struct bio *bio;
	struct bio_vec *bvec;
1506
	int i;
1507

1508 1509 1510
	bio = bio_copy_user(q, NULL, (unsigned long)data, len, 1, gfp_mask);
	if (IS_ERR(bio))
		return bio;
1511 1512 1513 1514

	if (!reading) {
		void *p = data;

1515
		bio_for_each_segment_all(bvec, bio, i) {
1516 1517 1518 1519 1520 1521 1522 1523
			char *addr = page_address(bvec->bv_page);

			memcpy(addr, p, bvec->bv_len);
			p += bvec->bv_len;
		}
	}

	bio->bi_end_io = bio_copy_kern_endio;
1524

1525 1526
	return bio;
}
1527
EXPORT_SYMBOL(bio_copy_kern);
1528

L
Linus Torvalds 已提交
1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
/*
 * bio_set_pages_dirty() and bio_check_pages_dirty() are support functions
 * for performing direct-IO in BIOs.
 *
 * The problem is that we cannot run set_page_dirty() from interrupt context
 * because the required locks are not interrupt-safe.  So what we can do is to
 * mark the pages dirty _before_ performing IO.  And in interrupt context,
 * check that the pages are still dirty.   If so, fine.  If not, redirty them
 * in process context.
 *
 * We special-case compound pages here: normally this means reads into hugetlb
 * pages.  The logic in here doesn't really work right for compound pages
 * because the VM does not uniformly chase down the head page in all cases.
 * But dirtiness of compound pages is pretty meaningless anyway: the VM doesn't
 * handle them at all.  So we skip compound pages here at an early stage.
 *
 * Note that this code is very hard to test under normal circumstances because
 * direct-io pins the pages with get_user_pages().  This makes
 * is_page_cache_freeable return false, and the VM will not clean the pages.
1548
 * But other code (eg, flusher threads) could clean the pages if they are mapped
L
Linus Torvalds 已提交
1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559
 * pagecache.
 *
 * Simply disabling the call to bio_set_pages_dirty() is a good way to test the
 * deferred bio dirtying paths.
 */

/*
 * bio_set_pages_dirty() will mark all the bio's pages as dirty.
 */
void bio_set_pages_dirty(struct bio *bio)
{
1560
	struct bio_vec *bvec;
L
Linus Torvalds 已提交
1561 1562
	int i;

1563 1564
	bio_for_each_segment_all(bvec, bio, i) {
		struct page *page = bvec->bv_page;
L
Linus Torvalds 已提交
1565 1566 1567 1568 1569 1570

		if (page && !PageCompound(page))
			set_page_dirty_lock(page);
	}
}

1571
static void bio_release_pages(struct bio *bio)
L
Linus Torvalds 已提交
1572
{
1573
	struct bio_vec *bvec;
L
Linus Torvalds 已提交
1574 1575
	int i;

1576 1577
	bio_for_each_segment_all(bvec, bio, i) {
		struct page *page = bvec->bv_page;
L
Linus Torvalds 已提交
1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594

		if (page)
			put_page(page);
	}
}

/*
 * bio_check_pages_dirty() will check that all the BIO's pages are still dirty.
 * If they are, then fine.  If, however, some pages are clean then they must
 * have been written out during the direct-IO read.  So we take another ref on
 * the BIO and the offending pages and re-dirty the pages in process context.
 *
 * It is expected that bio_check_pages_dirty() will wholly own the BIO from
 * here on.  It will run one page_cache_release() against each page and will
 * run one bio_put() against the BIO.
 */

1595
static void bio_dirty_fn(struct work_struct *work);
L
Linus Torvalds 已提交
1596

1597
static DECLARE_WORK(bio_dirty_work, bio_dirty_fn);
L
Linus Torvalds 已提交
1598 1599 1600 1601 1602 1603
static DEFINE_SPINLOCK(bio_dirty_lock);
static struct bio *bio_dirty_list;

/*
 * This runs in process context
 */
1604
static void bio_dirty_fn(struct work_struct *work)
L
Linus Torvalds 已提交
1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625
{
	unsigned long flags;
	struct bio *bio;

	spin_lock_irqsave(&bio_dirty_lock, flags);
	bio = bio_dirty_list;
	bio_dirty_list = NULL;
	spin_unlock_irqrestore(&bio_dirty_lock, flags);

	while (bio) {
		struct bio *next = bio->bi_private;

		bio_set_pages_dirty(bio);
		bio_release_pages(bio);
		bio_put(bio);
		bio = next;
	}
}

void bio_check_pages_dirty(struct bio *bio)
{
1626
	struct bio_vec *bvec;
L
Linus Torvalds 已提交
1627 1628 1629
	int nr_clean_pages = 0;
	int i;

1630 1631
	bio_for_each_segment_all(bvec, bio, i) {
		struct page *page = bvec->bv_page;
L
Linus Torvalds 已提交
1632 1633 1634

		if (PageDirty(page) || PageCompound(page)) {
			page_cache_release(page);
1635
			bvec->bv_page = NULL;
L
Linus Torvalds 已提交
1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
		} else {
			nr_clean_pages++;
		}
	}

	if (nr_clean_pages) {
		unsigned long flags;

		spin_lock_irqsave(&bio_dirty_lock, flags);
		bio->bi_private = bio_dirty_list;
		bio_dirty_list = bio;
		spin_unlock_irqrestore(&bio_dirty_lock, flags);
		schedule_work(&bio_dirty_work);
	} else {
		bio_put(bio);
	}
}

1654 1655 1656
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
void bio_flush_dcache_pages(struct bio *bi)
{
1657 1658
	struct bio_vec bvec;
	struct bvec_iter iter;
1659

1660 1661
	bio_for_each_segment(bvec, bi, iter)
		flush_dcache_page(bvec.bv_page);
1662 1663 1664 1665
}
EXPORT_SYMBOL(bio_flush_dcache_pages);
#endif

L
Linus Torvalds 已提交
1666 1667 1668 1669 1670 1671
/**
 * bio_endio - end I/O on a bio
 * @bio:	bio
 * @error:	error, if any
 *
 * Description:
1672
 *   bio_endio() will end I/O on the whole bio. bio_endio() is the
N
NeilBrown 已提交
1673 1674 1675
 *   preferred way to end I/O on a bio, it takes care of clearing
 *   BIO_UPTODATE on error. @error is 0 on success, and and one of the
 *   established -Exxxx (-EIO, for instance) error values in case
L
Lucas De Marchi 已提交
1676
 *   something went wrong. No one should call bi_end_io() directly on a
N
NeilBrown 已提交
1677 1678
 *   bio unless they own it and thus know that it has an end_io
 *   function.
L
Linus Torvalds 已提交
1679
 **/
1680
void bio_endio(struct bio *bio, int error)
L
Linus Torvalds 已提交
1681 1682 1683
{
	if (error)
		clear_bit(BIO_UPTODATE, &bio->bi_flags);
N
NeilBrown 已提交
1684 1685
	else if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
		error = -EIO;
L
Linus Torvalds 已提交
1686

N
NeilBrown 已提交
1687
	if (bio->bi_end_io)
1688
		bio->bi_end_io(bio, error);
L
Linus Torvalds 已提交
1689
}
1690
EXPORT_SYMBOL(bio_endio);
L
Linus Torvalds 已提交
1691 1692 1693 1694 1695 1696

void bio_pair_release(struct bio_pair *bp)
{
	if (atomic_dec_and_test(&bp->cnt)) {
		struct bio *master = bp->bio1.bi_private;

1697
		bio_endio(master, bp->error);
L
Linus Torvalds 已提交
1698 1699 1700
		mempool_free(bp, bp->bio2.bi_private);
	}
}
1701
EXPORT_SYMBOL(bio_pair_release);
L
Linus Torvalds 已提交
1702

1703
static void bio_pair_end_1(struct bio *bi, int err)
L
Linus Torvalds 已提交
1704 1705 1706 1707 1708 1709 1710 1711 1712
{
	struct bio_pair *bp = container_of(bi, struct bio_pair, bio1);

	if (err)
		bp->error = err;

	bio_pair_release(bp);
}

1713
static void bio_pair_end_2(struct bio *bi, int err)
L
Linus Torvalds 已提交
1714 1715 1716 1717 1718 1719 1720 1721 1722 1723
{
	struct bio_pair *bp = container_of(bi, struct bio_pair, bio2);

	if (err)
		bp->error = err;

	bio_pair_release(bp);
}

/*
1724
 * split a bio - only worry about a bio with a single page in its iovec
L
Linus Torvalds 已提交
1725
 */
D
Denis ChengRq 已提交
1726
struct bio_pair *bio_split(struct bio *bi, int first_sectors)
L
Linus Torvalds 已提交
1727
{
D
Denis ChengRq 已提交
1728
	struct bio_pair *bp = mempool_alloc(bio_split_pool, GFP_NOIO);
L
Linus Torvalds 已提交
1729 1730 1731 1732

	if (!bp)
		return bp;

1733
	trace_block_split(bdev_get_queue(bi->bi_bdev), bi,
1734
				bi->bi_iter.bi_sector + first_sectors);
1735

1736
	BUG_ON(bio_multiple_segments(bi));
L
Linus Torvalds 已提交
1737 1738 1739 1740
	atomic_set(&bp->cnt, 3);
	bp->error = 0;
	bp->bio1 = *bi;
	bp->bio2 = *bi;
1741 1742 1743
	bp->bio2.bi_iter.bi_sector += first_sectors;
	bp->bio2.bi_iter.bi_size -= first_sectors << 9;
	bp->bio1.bi_iter.bi_size = first_sectors << 9;
L
Linus Torvalds 已提交
1744

1745
	if (bi->bi_vcnt != 0) {
1746 1747
		bp->bv1 = bio_iovec(bi);
		bp->bv2 = bio_iovec(bi);
1748

1749 1750 1751 1752 1753
		if (bio_is_rw(bi)) {
			bp->bv2.bv_offset += first_sectors << 9;
			bp->bv2.bv_len -= first_sectors << 9;
			bp->bv1.bv_len = first_sectors << 9;
		}
L
Linus Torvalds 已提交
1754

1755 1756
		bp->bio1.bi_io_vec = &bp->bv1;
		bp->bio2.bi_io_vec = &bp->bv2;
L
Linus Torvalds 已提交
1757

1758 1759 1760
		bp->bio1.bi_max_vecs = 1;
		bp->bio2.bi_max_vecs = 1;
	}
1761

L
Linus Torvalds 已提交
1762 1763 1764 1765
	bp->bio1.bi_end_io = bio_pair_end_1;
	bp->bio2.bi_end_io = bio_pair_end_2;

	bp->bio1.bi_private = bi;
D
Denis ChengRq 已提交
1766
	bp->bio2.bi_private = bio_split_pool;
L
Linus Torvalds 已提交
1767

1768 1769 1770
	if (bio_integrity(bi))
		bio_integrity_split(bi, bp, first_sectors);

L
Linus Torvalds 已提交
1771 1772
	return bp;
}
1773
EXPORT_SYMBOL(bio_split);
L
Linus Torvalds 已提交
1774

1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
/**
 * bio_trim - trim a bio
 * @bio:	bio to trim
 * @offset:	number of sectors to trim from the front of @bio
 * @size:	size we want to trim @bio to, in sectors
 */
void bio_trim(struct bio *bio, int offset, int size)
{
	/* 'bio' is a cloned bio which we need to trim to match
	 * the given offset and size.
	 * This requires adjusting bi_sector, bi_size, and bi_io_vec
	 */
	int i;
	struct bio_vec *bvec;
	int sofar = 0;

	size <<= 9;
1792
	if (offset == 0 && size == bio->bi_iter.bi_size)
1793 1794 1795 1796 1797 1798
		return;

	clear_bit(BIO_SEG_VALID, &bio->bi_flags);

	bio_advance(bio, offset << 9);

1799
	bio->bi_iter.bi_size = size;
1800 1801

	/* avoid any complications with bi_idx being non-zero*/
1802 1803 1804 1805 1806 1807
	if (bio->bi_iter.bi_idx) {
		memmove(bio->bi_io_vec, bio->bi_io_vec+bio->bi_iter.bi_idx,
			(bio->bi_vcnt - bio->bi_iter.bi_idx) *
			sizeof(struct bio_vec));
		bio->bi_vcnt -= bio->bi_iter.bi_idx;
		bio->bi_iter.bi_idx = 0;
1808 1809
	}
	/* Make sure vcnt and last bv are not too big */
1810
	bio_for_each_segment_all(bvec, bio, i) {
1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821
		if (sofar + bvec->bv_len > size)
			bvec->bv_len = size - sofar;
		if (bvec->bv_len == 0) {
			bio->bi_vcnt = i;
			break;
		}
		sofar += bvec->bv_len;
	}
}
EXPORT_SYMBOL_GPL(bio_trim);

1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834
/**
 *      bio_sector_offset - Find hardware sector offset in bio
 *      @bio:           bio to inspect
 *      @index:         bio_vec index
 *      @offset:        offset in bv_page
 *
 *      Return the number of hardware sectors between beginning of bio
 *      and an end point indicated by a bio_vec index and an offset
 *      within that vector's page.
 */
sector_t bio_sector_offset(struct bio *bio, unsigned short index,
			   unsigned int offset)
{
1835
	unsigned int sector_sz;
1836 1837 1838 1839
	struct bio_vec *bv;
	sector_t sectors;
	int i;

1840
	sector_sz = queue_logical_block_size(bio->bi_bdev->bd_disk->queue);
1841 1842
	sectors = 0;

1843
	if (index >= bio->bi_iter.bi_idx)
1844 1845
		index = bio->bi_vcnt - 1;

1846
	bio_for_each_segment_all(bv, bio, i) {
1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858
		if (i == index) {
			if (offset > bv->bv_offset)
				sectors += (offset - bv->bv_offset) / sector_sz;
			break;
		}

		sectors += bv->bv_len / sector_sz;
	}

	return sectors;
}
EXPORT_SYMBOL(bio_sector_offset);
L
Linus Torvalds 已提交
1859 1860 1861 1862 1863

/*
 * create memory pools for biovec's in a bio_set.
 * use the global biovec slabs created for general use.
 */
1864
mempool_t *biovec_create_pool(struct bio_set *bs, int pool_entries)
L
Linus Torvalds 已提交
1865
{
1866
	struct biovec_slab *bp = bvec_slabs + BIOVEC_MAX_IDX;
L
Linus Torvalds 已提交
1867

1868
	return mempool_create_slab_pool(pool_entries, bp->slab);
L
Linus Torvalds 已提交
1869 1870 1871 1872
}

void bioset_free(struct bio_set *bs)
{
1873 1874 1875
	if (bs->rescue_workqueue)
		destroy_workqueue(bs->rescue_workqueue);

L
Linus Torvalds 已提交
1876 1877 1878
	if (bs->bio_pool)
		mempool_destroy(bs->bio_pool);

1879 1880 1881
	if (bs->bvec_pool)
		mempool_destroy(bs->bvec_pool);

1882
	bioset_integrity_free(bs);
1883
	bio_put_slab(bs);
L
Linus Torvalds 已提交
1884 1885 1886

	kfree(bs);
}
1887
EXPORT_SYMBOL(bioset_free);
L
Linus Torvalds 已提交
1888

1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902
/**
 * bioset_create  - Create a bio_set
 * @pool_size:	Number of bio and bio_vecs to cache in the mempool
 * @front_pad:	Number of bytes to allocate in front of the returned bio
 *
 * Description:
 *    Set up a bio_set to be used with @bio_alloc_bioset. Allows the caller
 *    to ask for a number of bytes to be allocated in front of the bio.
 *    Front pad allocation is useful for embedding the bio inside
 *    another structure, to avoid allocating extra data to go with the bio.
 *    Note that the bio must be embedded at the END of that structure always,
 *    or things will break badly.
 */
struct bio_set *bioset_create(unsigned int pool_size, unsigned int front_pad)
L
Linus Torvalds 已提交
1903
{
1904
	unsigned int back_pad = BIO_INLINE_VECS * sizeof(struct bio_vec);
1905
	struct bio_set *bs;
L
Linus Torvalds 已提交
1906

1907
	bs = kzalloc(sizeof(*bs), GFP_KERNEL);
L
Linus Torvalds 已提交
1908 1909 1910
	if (!bs)
		return NULL;

1911
	bs->front_pad = front_pad;
1912

1913 1914 1915 1916
	spin_lock_init(&bs->rescue_lock);
	bio_list_init(&bs->rescue_list);
	INIT_WORK(&bs->rescue_work, bio_alloc_rescue);

1917
	bs->bio_slab = bio_find_or_create_slab(front_pad + back_pad);
1918 1919 1920 1921 1922 1923
	if (!bs->bio_slab) {
		kfree(bs);
		return NULL;
	}

	bs->bio_pool = mempool_create_slab_pool(pool_size, bs->bio_slab);
L
Linus Torvalds 已提交
1924 1925 1926
	if (!bs->bio_pool)
		goto bad;

1927 1928
	bs->bvec_pool = biovec_create_pool(bs, pool_size);
	if (!bs->bvec_pool)
1929 1930 1931 1932 1933
		goto bad;

	bs->rescue_workqueue = alloc_workqueue("bioset", WQ_MEM_RECLAIM, 0);
	if (!bs->rescue_workqueue)
		goto bad;
L
Linus Torvalds 已提交
1934

1935
	return bs;
L
Linus Torvalds 已提交
1936 1937 1938 1939
bad:
	bioset_free(bs);
	return NULL;
}
1940
EXPORT_SYMBOL(bioset_create);
L
Linus Torvalds 已提交
1941

1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
#ifdef CONFIG_BLK_CGROUP
/**
 * bio_associate_current - associate a bio with %current
 * @bio: target bio
 *
 * Associate @bio with %current if it hasn't been associated yet.  Block
 * layer will treat @bio as if it were issued by %current no matter which
 * task actually issues it.
 *
 * This function takes an extra reference of @task's io_context and blkcg
 * which will be put when @bio is released.  The caller must own @bio,
 * ensure %current->io_context exists, and is responsible for synchronizing
 * calls to this function.
 */
int bio_associate_current(struct bio *bio)
{
	struct io_context *ioc;
	struct cgroup_subsys_state *css;

	if (bio->bi_ioc)
		return -EBUSY;

	ioc = current->io_context;
	if (!ioc)
		return -ENOENT;

	/* acquire active ref on @ioc and associate */
	get_io_context_active(ioc);
	bio->bi_ioc = ioc;

	/* associate blkcg if exists */
	rcu_read_lock();
1974
	css = task_css(current, blkio_subsys_id);
1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
	if (css && css_tryget(css))
		bio->bi_css = css;
	rcu_read_unlock();

	return 0;
}

/**
 * bio_disassociate_task - undo bio_associate_current()
 * @bio: target bio
 */
void bio_disassociate_task(struct bio *bio)
{
	if (bio->bi_ioc) {
		put_io_context(bio->bi_ioc);
		bio->bi_ioc = NULL;
	}
	if (bio->bi_css) {
		css_put(bio->bi_css);
		bio->bi_css = NULL;
	}
}

#endif /* CONFIG_BLK_CGROUP */

L
Linus Torvalds 已提交
2000 2001 2002 2003 2004 2005 2006 2007
static void __init biovec_init_slabs(void)
{
	int i;

	for (i = 0; i < BIOVEC_NR_POOLS; i++) {
		int size;
		struct biovec_slab *bvs = bvec_slabs + i;

2008 2009 2010 2011 2012
		if (bvs->nr_vecs <= BIO_INLINE_VECS) {
			bvs->slab = NULL;
			continue;
		}

L
Linus Torvalds 已提交
2013 2014
		size = bvs->nr_vecs * sizeof(struct bio_vec);
		bvs->slab = kmem_cache_create(bvs->name, size, 0,
2015
                                SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
2016 2017 2018 2019 2020
	}
}

static int __init init_bio(void)
{
2021 2022 2023 2024 2025
	bio_slab_max = 2;
	bio_slab_nr = 0;
	bio_slabs = kzalloc(bio_slab_max * sizeof(struct bio_slab), GFP_KERNEL);
	if (!bio_slabs)
		panic("bio: can't allocate bios\n");
L
Linus Torvalds 已提交
2026

2027
	bio_integrity_init();
L
Linus Torvalds 已提交
2028 2029
	biovec_init_slabs();

2030
	fs_bio_set = bioset_create(BIO_POOL_SIZE, 0);
L
Linus Torvalds 已提交
2031 2032 2033
	if (!fs_bio_set)
		panic("bio: can't allocate bios\n");

2034 2035 2036
	if (bioset_integrity_create(fs_bio_set, BIO_POOL_SIZE))
		panic("bio: can't create integrity pool\n");

2037 2038
	bio_split_pool = mempool_create_kmalloc_pool(BIO_SPLIT_ENTRIES,
						     sizeof(struct bio_pair));
L
Linus Torvalds 已提交
2039 2040 2041 2042 2043 2044
	if (!bio_split_pool)
		panic("bio: can't create split pool\n");

	return 0;
}
subsys_initcall(init_bio);