bio.c 34.1 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 22 23 24 25 26 27
 *
 * 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>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/mempool.h>
#include <linux/workqueue.h>
28
#include <linux/blktrace_api.h>
29
#include <scsi/sg.h>		/* for struct sg_iovec */
L
Linus Torvalds 已提交
30

31
static struct kmem_cache *bio_slab __read_mostly;
L
Linus Torvalds 已提交
32

33
mempool_t *bio_split_pool __read_mostly;
L
Linus Torvalds 已提交
34 35 36 37 38 39 40 41

/*
 * 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) }
42
static struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
L
Linus Torvalds 已提交
43 44 45 46 47 48 49 50
	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.
 */
51
struct bio_set *fs_bio_set;
L
Linus Torvalds 已提交
52

53 54 55 56 57
unsigned int bvec_nr_vecs(unsigned short idx)
{
	return bvec_slabs[idx].nr_vecs;
}

58
struct bio_vec *bvec_alloc_bs(gfp_t gfp_mask, int nr, unsigned long *idx, struct bio_set *bs)
L
Linus Torvalds 已提交
59 60 61 62
{
	struct bio_vec *bvl;

	/*
J
Jens Axboe 已提交
63 64 65
	 * If 'bs' is given, lookup the pool and do the mempool alloc.
	 * If not, this is a bio_kmalloc() allocation and just do a
	 * kzalloc() for the exact number of vecs right away.
L
Linus Torvalds 已提交
66
	 */
J
Jens Axboe 已提交
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
	if (bs) {
		/*
		 * 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;
L
Linus Torvalds 已提交
90 91
		default:
			return NULL;
J
Jens Axboe 已提交
92
		}
L
Linus Torvalds 已提交
93

J
Jens Axboe 已提交
94 95 96 97 98 99 100 101 102
		/*
		 * idx now points to the pool we want to allocate from
		 */
		bvl = mempool_alloc(bs->bvec_pools[*idx], gfp_mask);
		if (bvl)
			memset(bvl, 0,
				bvec_nr_vecs(*idx) * sizeof(struct bio_vec));
	} else
		bvl = kzalloc(nr * sizeof(struct bio_vec), gfp_mask);
L
Linus Torvalds 已提交
103 104 105 106

	return bvl;
}

P
Peter Osterlund 已提交
107
void bio_free(struct bio *bio, struct bio_set *bio_set)
L
Linus Torvalds 已提交
108
{
109 110
	if (bio->bi_io_vec) {
		const int pool_idx = BIO_POOL_IDX(bio);
L
Linus Torvalds 已提交
111

112 113 114 115
		BIO_BUG_ON(pool_idx >= BIOVEC_NR_POOLS);

		mempool_free(bio->bi_io_vec, bio_set->bvec_pools[pool_idx]);
	}
L
Linus Torvalds 已提交
116

117 118 119
	if (bio_integrity(bio))
		bio_integrity_free(bio, bio_set);

P
Peter Osterlund 已提交
120 121 122 123 124 125 126 127 128
	mempool_free(bio, bio_set->bio_pool);
}

/*
 * default destructor for a bio allocated with bio_alloc_bioset()
 */
static void bio_fs_destructor(struct bio *bio)
{
	bio_free(bio, fs_bio_set);
L
Linus Torvalds 已提交
129 130
}

J
Jens Axboe 已提交
131 132 133 134 135 136
static void bio_kmalloc_destructor(struct bio *bio)
{
	kfree(bio->bi_io_vec);
	kfree(bio);
}

137
void bio_init(struct bio *bio)
L
Linus Torvalds 已提交
138
{
J
Jens Axboe 已提交
139
	memset(bio, 0, sizeof(*bio));
L
Linus Torvalds 已提交
140
	bio->bi_flags = 1 << BIO_UPTODATE;
141
	bio->bi_comp_cpu = -1;
L
Linus Torvalds 已提交
142 143 144 145 146 147 148
	atomic_set(&bio->bi_cnt, 1);
}

/**
 * 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
J
Jens Axboe 已提交
149
 * @bs:		the bio_set to allocate from. If %NULL, just use kmalloc
L
Linus Torvalds 已提交
150 151
 *
 * Description:
J
Jens Axboe 已提交
152
 *   bio_alloc_bioset will first try its own mempool to satisfy the allocation.
L
Linus Torvalds 已提交
153
 *   If %__GFP_WAIT is set then we will block on the internal pool waiting
J
Jens Axboe 已提交
154 155
 *   for a &struct bio to become free. If a %NULL @bs is passed in, we will
 *   fall back to just using @kmalloc to allocate the required memory.
L
Linus Torvalds 已提交
156 157
 *
 *   allocate bio and iovecs from the memory pools specified by the
J
Jens Axboe 已提交
158
 *   bio_set structure, or @kmalloc if none given.
L
Linus Torvalds 已提交
159
 **/
A
Al Viro 已提交
160
struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
L
Linus Torvalds 已提交
161
{
J
Jens Axboe 已提交
162 163 164 165 166 167
	struct bio *bio;

	if (bs)
		bio = mempool_alloc(bs->bio_pool, gfp_mask);
	else
		bio = kmalloc(sizeof(*bio), gfp_mask);
L
Linus Torvalds 已提交
168 169 170 171 172 173

	if (likely(bio)) {
		struct bio_vec *bvl = NULL;

		bio_init(bio);
		if (likely(nr_iovecs)) {
174
			unsigned long uninitialized_var(idx);
L
Linus Torvalds 已提交
175 176 177

			bvl = bvec_alloc_bs(gfp_mask, nr_iovecs, &idx, bs);
			if (unlikely(!bvl)) {
J
Jens Axboe 已提交
178 179 180 181
				if (bs)
					mempool_free(bio, bs->bio_pool);
				else
					kfree(bio);
L
Linus Torvalds 已提交
182 183 184 185
				bio = NULL;
				goto out;
			}
			bio->bi_flags |= idx << BIO_POOL_OFFSET;
D
Denis ChengRq 已提交
186
			bio->bi_max_vecs = bvec_nr_vecs(idx);
L
Linus Torvalds 已提交
187 188 189 190 191 192 193
		}
		bio->bi_io_vec = bvl;
	}
out:
	return bio;
}

A
Al Viro 已提交
194
struct bio *bio_alloc(gfp_t gfp_mask, int nr_iovecs)
L
Linus Torvalds 已提交
195
{
P
Peter Osterlund 已提交
196 197 198 199 200 201
	struct bio *bio = bio_alloc_bioset(gfp_mask, nr_iovecs, fs_bio_set);

	if (bio)
		bio->bi_destructor = bio_fs_destructor;

	return bio;
L
Linus Torvalds 已提交
202 203
}

J
Jens Axboe 已提交
204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
/*
 * Like bio_alloc(), but doesn't use a mempool backing. This means that
 * it CAN fail, but while bio_alloc() can only be used for allocations
 * that have a short (finite) life span, bio_kmalloc() should be used
 * for more permanent bio allocations (like allocating some bio's for
 * initalization or setup purposes).
 */
struct bio *bio_kmalloc(gfp_t gfp_mask, int nr_iovecs)
{
	struct bio *bio = bio_alloc_bioset(gfp_mask, nr_iovecs, NULL);

	if (bio)
		bio->bi_destructor = bio_kmalloc_destructor;

	return bio;
}

L
Linus Torvalds 已提交
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256
void zero_fill_bio(struct bio *bio)
{
	unsigned long flags;
	struct bio_vec *bv;
	int i;

	bio_for_each_segment(bv, bio, i) {
		char *data = bvec_kmap_irq(bv, &flags);
		memset(data, 0, bv->bv_len);
		flush_dcache_page(bv->bv_page);
		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
 *   bio_alloc or bio_get. The last put of a bio will free it.
 **/
void bio_put(struct bio *bio)
{
	BIO_BUG_ON(!atomic_read(&bio->bi_cnt));

	/*
	 * last put frees it
	 */
	if (atomic_dec_and_test(&bio->bi_cnt)) {
		bio->bi_next = NULL;
		bio->bi_destructor(bio);
	}
}

257
inline int bio_phys_segments(struct request_queue *q, struct bio *bio)
L
Linus Torvalds 已提交
258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
{
	if (unlikely(!bio_flagged(bio, BIO_SEG_VALID)))
		blk_recount_segments(q, bio);

	return bio->bi_phys_segments;
}

/**
 * 	__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.
 */
274
void __bio_clone(struct bio *bio, struct bio *bio_src)
L
Linus Torvalds 已提交
275
{
276 277
	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
278

279 280 281 282
	/*
	 * 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
283 284 285 286 287 288
	bio->bi_sector = bio_src->bi_sector;
	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;
	bio->bi_size = bio_src->bi_size;
A
Andrew Morton 已提交
289
	bio->bi_idx = bio_src->bi_idx;
L
Linus Torvalds 已提交
290 291 292 293 294 295 296 297 298
}

/**
 *	bio_clone	-	clone a bio
 *	@bio: bio to clone
 *	@gfp_mask: allocation priority
 *
 * 	Like __bio_clone, only also allocates the returned bio
 */
A
Al Viro 已提交
299
struct bio *bio_clone(struct bio *bio, gfp_t gfp_mask)
L
Linus Torvalds 已提交
300 301 302
{
	struct bio *b = bio_alloc_bioset(gfp_mask, bio->bi_max_vecs, fs_bio_set);

303 304 305 306 307 308 309 310 311 312 313 314 315
	if (!b)
		return NULL;

	b->bi_destructor = bio_fs_destructor;
	__bio_clone(b, bio);

	if (bio_integrity(bio)) {
		int ret;

		ret = bio_integrity_clone(b, bio, fs_bio_set);

		if (ret < 0)
			return NULL;
P
Peter Osterlund 已提交
316
	}
L
Linus Torvalds 已提交
317 318 319 320 321 322 323 324 325 326 327 328 329 330 331

	return b;
}

/**
 *	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)
{
332
	struct request_queue *q = bdev_get_queue(bdev);
L
Linus Torvalds 已提交
333 334 335 336 337 338 339 340 341 342 343
	int nr_pages;

	nr_pages = ((q->max_sectors << 9) + PAGE_SIZE - 1) >> PAGE_SHIFT;
	if (nr_pages > q->max_phys_segments)
		nr_pages = q->max_phys_segments;
	if (nr_pages > q->max_hw_segments)
		nr_pages = q->max_hw_segments;

	return nr_pages;
}

344
static int __bio_add_page(struct request_queue *q, struct bio *bio, struct page
345 346
			  *page, unsigned int len, unsigned int offset,
			  unsigned short max_sectors)
L
Linus Torvalds 已提交
347 348 349 350 351 352 353 354 355 356
{
	int retried_segments = 0;
	struct bio_vec *bvec;

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

357
	if (((bio->bi_size + len) >> 9) > max_sectors)
L
Linus Torvalds 已提交
358 359
		return 0;

360 361 362 363 364 365 366 367 368 369 370
	/*
	 * 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) {
			prev->bv_len += len;
371 372 373 374 375 376 377 378 379 380 381 382 383

			if (q->merge_bvec_fn) {
				struct bvec_merge_data bvm = {
					.bi_bdev = bio->bi_bdev,
					.bi_sector = bio->bi_sector,
					.bi_size = bio->bi_size,
					.bi_rw = bio->bi_rw,
				};

				if (q->merge_bvec_fn(q, &bvm, prev) < len) {
					prev->bv_len -= len;
					return 0;
				}
384 385 386 387 388 389 390
			}

			goto done;
		}
	}

	if (bio->bi_vcnt >= bio->bi_max_vecs)
L
Linus Torvalds 已提交
391 392 393 394 395 396 397 398
		return 0;

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

	while (bio->bi_phys_segments >= q->max_phys_segments
M
Mikulas Patocka 已提交
399
	       || bio->bi_phys_segments >= q->max_hw_segments) {
L
Linus Torvalds 已提交
400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422

		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) {
423 424 425 426 427 428 429
		struct bvec_merge_data bvm = {
			.bi_bdev = bio->bi_bdev,
			.bi_sector = bio->bi_sector,
			.bi_size = bio->bi_size,
			.bi_rw = bio->bi_rw,
		};

L
Linus Torvalds 已提交
430 431 432 433
		/*
		 * merge_bvec_fn() returns number of bytes it can accept
		 * at this offset
		 */
434
		if (q->merge_bvec_fn(q, &bvm, bvec) < len) {
L
Linus Torvalds 已提交
435 436 437 438 439 440 441 442
			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 */
443
	if (bio->bi_vcnt && (BIOVEC_PHYS_MERGEABLE(bvec-1, bvec)))
L
Linus Torvalds 已提交
444 445 446 447
		bio->bi_flags &= ~(1 << BIO_SEG_VALID);

	bio->bi_vcnt++;
	bio->bi_phys_segments++;
448
 done:
L
Linus Torvalds 已提交
449 450 451 452
	bio->bi_size += len;
	return len;
}

453 454
/**
 *	bio_add_pc_page	-	attempt to add page to bio
J
Jens Axboe 已提交
455
 *	@q: the target queue
456 457 458 459 460 461 462 463 464 465 466
 *	@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
 *	number of reasons, such as the bio being full or target block
 *	device limitations. The target block device must allow bio's
 *      smaller than 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.
 */
467
int bio_add_pc_page(struct request_queue *q, struct bio *bio, struct page *page,
468 469
		    unsigned int len, unsigned int offset)
{
470
	return __bio_add_page(q, bio, page, len, offset, q->max_hw_sectors);
471 472
}

L
Linus Torvalds 已提交
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488
/**
 *	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
 *	number of reasons, such as the bio being full or target block
 *	device limitations. The target block device must allow bio's
 *      smaller than PAGE_SIZE, so it is always possible to add a single
 *      page to an empty bio.
 */
int bio_add_page(struct bio *bio, struct page *page, unsigned int len,
		 unsigned int offset)
{
489 490
	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
	return __bio_add_page(q, bio, page, len, offset, q->max_sectors);
L
Linus Torvalds 已提交
491 492 493 494
}

struct bio_map_data {
	struct bio_vec *iovecs;
495
	struct sg_iovec *sgvecs;
496 497
	int nr_sgvecs;
	int is_our_pages;
L
Linus Torvalds 已提交
498 499
};

500
static void bio_set_map_data(struct bio_map_data *bmd, struct bio *bio,
501 502
			     struct sg_iovec *iov, int iov_count,
			     int is_our_pages)
L
Linus Torvalds 已提交
503 504
{
	memcpy(bmd->iovecs, bio->bi_io_vec, sizeof(struct bio_vec) * bio->bi_vcnt);
505 506
	memcpy(bmd->sgvecs, iov, sizeof(struct sg_iovec) * iov_count);
	bmd->nr_sgvecs = iov_count;
507
	bmd->is_our_pages = is_our_pages;
L
Linus Torvalds 已提交
508 509 510 511 512 513
	bio->bi_private = bmd;
}

static void bio_free_map_data(struct bio_map_data *bmd)
{
	kfree(bmd->iovecs);
514
	kfree(bmd->sgvecs);
L
Linus Torvalds 已提交
515 516 517
	kfree(bmd);
}

518 519
static struct bio_map_data *bio_alloc_map_data(int nr_segs, int iov_count,
					       gfp_t gfp_mask)
L
Linus Torvalds 已提交
520
{
521
	struct bio_map_data *bmd = kmalloc(sizeof(*bmd), gfp_mask);
L
Linus Torvalds 已提交
522 523 524 525

	if (!bmd)
		return NULL;

526
	bmd->iovecs = kmalloc(sizeof(struct bio_vec) * nr_segs, gfp_mask);
527 528 529 530 531
	if (!bmd->iovecs) {
		kfree(bmd);
		return NULL;
	}

532
	bmd->sgvecs = kmalloc(sizeof(struct sg_iovec) * iov_count, gfp_mask);
533
	if (bmd->sgvecs)
L
Linus Torvalds 已提交
534 535
		return bmd;

536
	kfree(bmd->iovecs);
L
Linus Torvalds 已提交
537 538 539 540
	kfree(bmd);
	return NULL;
}

541
static int __bio_copy_iov(struct bio *bio, struct bio_vec *iovecs,
542 543
			  struct sg_iovec *iov, int iov_count, int uncopy,
			  int do_free_page)
544 545 546 547 548 549 550 551 552
{
	int ret = 0, i;
	struct bio_vec *bvec;
	int iov_idx = 0;
	unsigned int iov_off = 0;
	int read = bio_data_dir(bio) == READ;

	__bio_for_each_segment(bvec, bio, i, 0) {
		char *bv_addr = page_address(bvec->bv_page);
553
		unsigned int bv_len = iovecs[i].bv_len;
554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585

		while (bv_len && iov_idx < iov_count) {
			unsigned int bytes;
			char *iov_addr;

			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) {
				if (!read && !uncopy)
					ret = copy_from_user(bv_addr, iov_addr,
							     bytes);
				if (read && uncopy)
					ret = copy_to_user(iov_addr, bv_addr,
							   bytes);

				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;
			}
		}

586
		if (do_free_page)
587 588 589 590 591 592
			__free_page(bvec->bv_page);
	}

	return ret;
}

L
Linus Torvalds 已提交
593 594 595 596 597 598 599 600 601 602
/**
 *	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;
603
	int ret = 0;
L
Linus Torvalds 已提交
604

605 606 607
	if (!bio_flagged(bio, BIO_NULL_MAPPED))
		ret = __bio_copy_iov(bio, bmd->iovecs, bmd->sgvecs,
				     bmd->nr_sgvecs, 1, bmd->is_our_pages);
L
Linus Torvalds 已提交
608 609 610 611 612 613
	bio_free_map_data(bmd);
	bio_put(bio);
	return ret;
}

/**
614
 *	bio_copy_user_iov	-	copy user data to bio
L
Linus Torvalds 已提交
615
 *	@q: destination block queue
616
 *	@map_data: pointer to the rq_map_data holding pages (if necessary)
617 618
 *	@iov:	the iovec.
 *	@iov_count: number of elements in the iovec
L
Linus Torvalds 已提交
619
 *	@write_to_vm: bool indicating writing to pages or not
620
 *	@gfp_mask: memory allocation flags
L
Linus Torvalds 已提交
621 622 623 624 625
 *
 *	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.
 */
626 627 628 629
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
630 631 632 633 634 635
{
	struct bio_map_data *bmd;
	struct bio_vec *bvec;
	struct page *page;
	struct bio *bio;
	int i, ret;
636 637
	int nr_pages = 0;
	unsigned int len = 0;
L
Linus Torvalds 已提交
638

639 640 641 642 643 644 645 646 647 648 649 650 651
	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;

		nr_pages += end - start;
		len += iov[i].iov_len;
	}

652
	bmd = bio_alloc_map_data(nr_pages, iov_count, gfp_mask);
L
Linus Torvalds 已提交
653 654 655 656
	if (!bmd)
		return ERR_PTR(-ENOMEM);

	ret = -ENOMEM;
657
	bio = bio_alloc(gfp_mask, nr_pages);
L
Linus Torvalds 已提交
658 659 660 661 662 663
	if (!bio)
		goto out_bmd;

	bio->bi_rw |= (!write_to_vm << BIO_RW);

	ret = 0;
664
	i = 0;
L
Linus Torvalds 已提交
665
	while (len) {
666 667 668 669 670 671
		unsigned int bytes;

		if (map_data)
			bytes = 1U << (PAGE_SHIFT + map_data->page_order);
		else
			bytes = PAGE_SIZE;
L
Linus Torvalds 已提交
672 673 674 675

		if (bytes > len)
			bytes = len;

676 677 678 679 680 681 682 683
		if (map_data) {
			if (i == map_data->nr_entries) {
				ret = -ENOMEM;
				break;
			}
			page = map_data->pages[i++];
		} else
			page = alloc_page(q->bounce_gfp | gfp_mask);
L
Linus Torvalds 已提交
684 685 686 687 688
		if (!page) {
			ret = -ENOMEM;
			break;
		}

689
		if (bio_add_pc_page(q, bio, page, bytes, 0) < bytes)
L
Linus Torvalds 已提交
690 691 692 693 694 695 696 697 698 699 700 701
			break;

		len -= bytes;
	}

	if (ret)
		goto cleanup;

	/*
	 * success
	 */
	if (!write_to_vm) {
702
		ret = __bio_copy_iov(bio, bio->bi_io_vec, iov, iov_count, 0, 0);
703 704
		if (ret)
			goto cleanup;
L
Linus Torvalds 已提交
705 706
	}

707
	bio_set_map_data(bmd, bio, iov, iov_count, map_data ? 0 : 1);
L
Linus Torvalds 已提交
708 709
	return bio;
cleanup:
710 711 712
	if (!map_data)
		bio_for_each_segment(bvec, bio, i)
			__free_page(bvec->bv_page);
L
Linus Torvalds 已提交
713 714 715 716 717 718 719

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

720 721 722
/**
 *	bio_copy_user	-	copy user data to bio
 *	@q: destination block queue
723
 *	@map_data: pointer to the rq_map_data holding pages (if necessary)
724 725 726
 *	@uaddr: start of user address
 *	@len: length in bytes
 *	@write_to_vm: bool indicating writing to pages or not
727
 *	@gfp_mask: memory allocation flags
728 729 730 731 732
 *
 *	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.
 */
733 734 735
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)
736 737 738 739 740 741
{
	struct sg_iovec iov;

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

742
	return bio_copy_user_iov(q, map_data, &iov, 1, write_to_vm, gfp_mask);
743 744
}

745
static struct bio *__bio_map_user_iov(struct request_queue *q,
746 747
				      struct block_device *bdev,
				      struct sg_iovec *iov, int iov_count,
748
				      int write_to_vm, gfp_t gfp_mask)
L
Linus Torvalds 已提交
749
{
750 751
	int i, j;
	int nr_pages = 0;
L
Linus Torvalds 已提交
752 753
	struct page **pages;
	struct bio *bio;
754 755
	int cur_page = 0;
	int ret, offset;
L
Linus Torvalds 已提交
756

757 758 759 760 761 762 763 764
	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;

		nr_pages += end - start;
		/*
765
		 * buffer must be aligned to at least hardsector size for now
766
		 */
767
		if (uaddr & queue_dma_alignment(q))
768 769 770 771
			return ERR_PTR(-EINVAL);
	}

	if (!nr_pages)
L
Linus Torvalds 已提交
772 773
		return ERR_PTR(-EINVAL);

774
	bio = bio_alloc(gfp_mask, nr_pages);
L
Linus Torvalds 已提交
775 776 777 778
	if (!bio)
		return ERR_PTR(-ENOMEM);

	ret = -ENOMEM;
779
	pages = kcalloc(nr_pages, sizeof(struct page *), gfp_mask);
L
Linus Torvalds 已提交
780 781 782
	if (!pages)
		goto out;

783 784 785 786 787 788 789 790
	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;
		
N
Nick Piggin 已提交
791 792
		ret = get_user_pages_fast(uaddr, local_nr_pages,
				write_to_vm, &pages[cur_page]);
793 794
		if (ret < local_nr_pages) {
			ret = -EFAULT;
795
			goto out_unmap;
796
		}
797 798 799 800 801 802 803 804 805 806 807 808 809 810

		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...
			 */
811 812
			if (bio_add_pc_page(q, bio, pages[j], bytes, offset) <
					    bytes)
813 814 815 816 817
				break;

			len -= bytes;
			offset = 0;
		}
L
Linus Torvalds 已提交
818

819
		cur_page = j;
L
Linus Torvalds 已提交
820
		/*
821
		 * release the pages we didn't map into the bio, if any
L
Linus Torvalds 已提交
822
		 */
823 824
		while (j < page_limit)
			page_cache_release(pages[j++]);
L
Linus Torvalds 已提交
825 826 827 828 829 830 831 832 833 834
	}

	kfree(pages);

	/*
	 * set data direction, and check if mapped pages need bouncing
	 */
	if (!write_to_vm)
		bio->bi_rw |= (1 << BIO_RW);

835
	bio->bi_bdev = bdev;
L
Linus Torvalds 已提交
836 837
	bio->bi_flags |= (1 << BIO_USER_MAPPED);
	return bio;
838 839 840 841 842 843 844 845

 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
846 847 848 849 850 851 852
	kfree(pages);
	bio_put(bio);
	return ERR_PTR(ret);
}

/**
 *	bio_map_user	-	map user address into bio
853
 *	@q: the struct request_queue for the bio
L
Linus Torvalds 已提交
854 855 856 857
 *	@bdev: destination block device
 *	@uaddr: start of user address
 *	@len: length in bytes
 *	@write_to_vm: bool indicating writing to pages or not
858
 *	@gfp_mask: memory allocation flags
L
Linus Torvalds 已提交
859 860 861 862
 *
 *	Map the user space address into a bio suitable for io to a block
 *	device. Returns an error pointer in case of error.
 */
863
struct bio *bio_map_user(struct request_queue *q, struct block_device *bdev,
864 865
			 unsigned long uaddr, unsigned int len, int write_to_vm,
			 gfp_t gfp_mask)
866 867 868
{
	struct sg_iovec iov;

869
	iov.iov_base = (void __user *)uaddr;
870 871
	iov.iov_len = len;

872
	return bio_map_user_iov(q, bdev, &iov, 1, write_to_vm, gfp_mask);
873 874 875 876
}

/**
 *	bio_map_user_iov - map user sg_iovec table into bio
877
 *	@q: the struct request_queue for the bio
878 879 880 881
 *	@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
882
 *	@gfp_mask: memory allocation flags
883 884 885 886
 *
 *	Map the user space address into a bio suitable for io to a block
 *	device. Returns an error pointer in case of error.
 */
887
struct bio *bio_map_user_iov(struct request_queue *q, struct block_device *bdev,
888
			     struct sg_iovec *iov, int iov_count,
889
			     int write_to_vm, gfp_t gfp_mask)
L
Linus Torvalds 已提交
890 891 892
{
	struct bio *bio;

893 894
	bio = __bio_map_user_iov(q, bdev, iov, iov_count, write_to_vm,
				 gfp_mask);
L
Linus Torvalds 已提交
895 896 897 898 899 900 901 902 903 904 905
	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);

906
	return bio;
L
Linus Torvalds 已提交
907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941
}

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

	/*
	 * make sure we dirty pages we wrote to
	 */
	__bio_for_each_segment(bvec, bio, i, 0) {
		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);
}

942
static void bio_map_kern_endio(struct bio *bio, int err)
943 944 945 946 947
{
	bio_put(bio);
}


948
static struct bio *__bio_map_kern(struct request_queue *q, void *data,
A
Al Viro 已提交
949
				  unsigned int len, gfp_t gfp_mask)
M
Mike Christie 已提交
950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971
{
	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;

	bio = bio_alloc(gfp_mask, nr_pages);
	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;

972 973
		if (bio_add_pc_page(q, bio, virt_to_page(data), bytes,
				    offset) < bytes)
M
Mike Christie 已提交
974 975 976 977 978 979 980
			break;

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

981
	bio->bi_end_io = bio_map_kern_endio;
M
Mike Christie 已提交
982 983 984 985 986
	return bio;
}

/**
 *	bio_map_kern	-	map kernel address into bio
987
 *	@q: the struct request_queue for the bio
M
Mike Christie 已提交
988 989 990 991 992 993 994
 *	@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.
 */
995
struct bio *bio_map_kern(struct request_queue *q, void *data, unsigned int len,
A
Al Viro 已提交
996
			 gfp_t gfp_mask)
M
Mike Christie 已提交
997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013
{
	struct bio *bio;

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

	if (bio->bi_size == len)
		return bio;

	/*
	 * Don't support partial mappings.
	 */
	bio_put(bio);
	return ERR_PTR(-EINVAL);
}

1014 1015 1016 1017
static void bio_copy_kern_endio(struct bio *bio, int err)
{
	struct bio_vec *bvec;
	const int read = bio_data_dir(bio) == READ;
1018
	struct bio_map_data *bmd = bio->bi_private;
1019
	int i;
1020
	char *p = bmd->sgvecs[0].iov_base;
1021 1022 1023

	__bio_for_each_segment(bvec, bio, i, 0) {
		char *addr = page_address(bvec->bv_page);
1024
		int len = bmd->iovecs[i].bv_len;
1025 1026

		if (read && !err)
1027
			memcpy(p, addr, len);
1028 1029

		__free_page(bvec->bv_page);
1030
		p += len;
1031 1032
	}

1033
	bio_free_map_data(bmd);
1034 1035 1036 1037 1038 1039 1040 1041 1042
	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
1043
 *	@reading: data direction is READ
1044 1045 1046 1047 1048 1049 1050 1051 1052
 *
 *	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;
1053
	int i;
1054

1055 1056 1057
	bio = bio_copy_user(q, NULL, (unsigned long)data, len, 1, gfp_mask);
	if (IS_ERR(bio))
		return bio;
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070

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

		bio_for_each_segment(bvec, bio, i) {
			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;
1071

1072 1073 1074
	return bio;
}

L
Linus Torvalds 已提交
1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
/*
 * 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.
 * But other code (eg, pdflush) could clean the pages if they are mapped
 * 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)
{
	struct bio_vec *bvec = bio->bi_io_vec;
	int i;

	for (i = 0; i < bio->bi_vcnt; i++) {
		struct page *page = bvec[i].bv_page;

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

1117
static void bio_release_pages(struct bio *bio)
L
Linus Torvalds 已提交
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140
{
	struct bio_vec *bvec = bio->bi_io_vec;
	int i;

	for (i = 0; i < bio->bi_vcnt; i++) {
		struct page *page = bvec[i].bv_page;

		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.
 */

1141
static void bio_dirty_fn(struct work_struct *work);
L
Linus Torvalds 已提交
1142

1143
static DECLARE_WORK(bio_dirty_work, bio_dirty_fn);
L
Linus Torvalds 已提交
1144 1145 1146 1147 1148 1149
static DEFINE_SPINLOCK(bio_dirty_lock);
static struct bio *bio_dirty_list;

/*
 * This runs in process context
 */
1150
static void bio_dirty_fn(struct work_struct *work)
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
{
	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)
{
	struct bio_vec *bvec = bio->bi_io_vec;
	int nr_clean_pages = 0;
	int i;

	for (i = 0; i < bio->bi_vcnt; i++) {
		struct page *page = bvec[i].bv_page;

		if (PageDirty(page) || PageCompound(page)) {
			page_cache_release(page);
			bvec[i].bv_page = NULL;
		} 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);
	}
}

/**
 * bio_endio - end I/O on a bio
 * @bio:	bio
 * @error:	error, if any
 *
 * Description:
1206
 *   bio_endio() will end I/O on the whole bio. bio_endio() is the
N
NeilBrown 已提交
1207 1208 1209 1210 1211 1212
 *   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
 *   something went wrong. Noone should call bi_end_io() directly on a
 *   bio unless they own it and thus know that it has an end_io
 *   function.
L
Linus Torvalds 已提交
1213
 **/
1214
void bio_endio(struct bio *bio, int error)
L
Linus Torvalds 已提交
1215 1216 1217
{
	if (error)
		clear_bit(BIO_UPTODATE, &bio->bi_flags);
N
NeilBrown 已提交
1218 1219
	else if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
		error = -EIO;
L
Linus Torvalds 已提交
1220

N
NeilBrown 已提交
1221
	if (bio->bi_end_io)
1222
		bio->bi_end_io(bio, error);
L
Linus Torvalds 已提交
1223 1224 1225 1226 1227 1228 1229
}

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

1230
		bio_endio(master, bp->error);
L
Linus Torvalds 已提交
1231 1232 1233 1234
		mempool_free(bp, bp->bio2.bi_private);
	}
}

1235
static void bio_pair_end_1(struct bio *bi, int err)
L
Linus Torvalds 已提交
1236 1237 1238 1239 1240 1241 1242 1243 1244
{
	struct bio_pair *bp = container_of(bi, struct bio_pair, bio1);

	if (err)
		bp->error = err;

	bio_pair_release(bp);
}

1245
static void bio_pair_end_2(struct bio *bi, int err)
L
Linus Torvalds 已提交
1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265
{
	struct bio_pair *bp = container_of(bi, struct bio_pair, bio2);

	if (err)
		bp->error = err;

	bio_pair_release(bp);
}

/*
 * split a bio - only worry about a bio with a single page
 * in it's iovec
 */
struct bio_pair *bio_split(struct bio *bi, mempool_t *pool, int first_sectors)
{
	struct bio_pair *bp = mempool_alloc(pool, GFP_NOIO);

	if (!bp)
		return bp;

1266 1267 1268
	blk_add_trace_pdu_int(bdev_get_queue(bi->bi_bdev), BLK_TA_SPLIT, bi,
				bi->bi_sector + first_sectors);

L
Linus Torvalds 已提交
1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
	BUG_ON(bi->bi_vcnt != 1);
	BUG_ON(bi->bi_idx != 0);
	atomic_set(&bp->cnt, 3);
	bp->error = 0;
	bp->bio1 = *bi;
	bp->bio2 = *bi;
	bp->bio2.bi_sector += first_sectors;
	bp->bio2.bi_size -= first_sectors << 9;
	bp->bio1.bi_size = first_sectors << 9;

	bp->bv1 = bi->bi_io_vec[0];
	bp->bv2 = bi->bi_io_vec[0];
	bp->bv2.bv_offset += first_sectors << 9;
	bp->bv2.bv_len -= first_sectors << 9;
	bp->bv1.bv_len = first_sectors << 9;

	bp->bio1.bi_io_vec = &bp->bv1;
	bp->bio2.bi_io_vec = &bp->bv2;

1288 1289 1290
	bp->bio1.bi_max_vecs = 1;
	bp->bio2.bi_max_vecs = 1;

L
Linus Torvalds 已提交
1291 1292 1293 1294 1295 1296
	bp->bio1.bi_end_io = bio_pair_end_1;
	bp->bio2.bi_end_io = bio_pair_end_2;

	bp->bio1.bi_private = bi;
	bp->bio2.bi_private = pool;

1297 1298 1299
	if (bio_integrity(bi))
		bio_integrity_split(bi, bp, first_sectors);

L
Linus Torvalds 已提交
1300 1301 1302
	return bp;
}

1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338
/**
 *      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)
{
	unsigned int sector_sz = queue_hardsect_size(bio->bi_bdev->bd_disk->queue);
	struct bio_vec *bv;
	sector_t sectors;
	int i;

	sectors = 0;

	if (index >= bio->bi_idx)
		index = bio->bi_vcnt - 1;

	__bio_for_each_segment(bv, bio, i, 0) {
		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
1339 1340 1341 1342 1343

/*
 * create memory pools for biovec's in a bio_set.
 * use the global biovec slabs created for general use.
 */
1344
static int biovec_create_pools(struct bio_set *bs, int pool_entries)
L
Linus Torvalds 已提交
1345 1346 1347 1348 1349 1350 1351
{
	int i;

	for (i = 0; i < BIOVEC_NR_POOLS; i++) {
		struct biovec_slab *bp = bvec_slabs + i;
		mempool_t **bvp = bs->bvec_pools + i;

1352
		*bvp = mempool_create_slab_pool(pool_entries, bp->slab);
L
Linus Torvalds 已提交
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376
		if (!*bvp)
			return -ENOMEM;
	}
	return 0;
}

static void biovec_free_pools(struct bio_set *bs)
{
	int i;

	for (i = 0; i < BIOVEC_NR_POOLS; i++) {
		mempool_t *bvp = bs->bvec_pools[i];

		if (bvp)
			mempool_destroy(bvp);
	}

}

void bioset_free(struct bio_set *bs)
{
	if (bs->bio_pool)
		mempool_destroy(bs->bio_pool);

1377
	bioset_integrity_free(bs);
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382
	biovec_free_pools(bs);

	kfree(bs);
}

1383
struct bio_set *bioset_create(int bio_pool_size, int bvec_pool_size)
L
Linus Torvalds 已提交
1384
{
1385
	struct bio_set *bs = kzalloc(sizeof(*bs), GFP_KERNEL);
L
Linus Torvalds 已提交
1386 1387 1388 1389

	if (!bs)
		return NULL;

1390
	bs->bio_pool = mempool_create_slab_pool(bio_pool_size, bio_slab);
L
Linus Torvalds 已提交
1391 1392 1393
	if (!bs->bio_pool)
		goto bad;

1394 1395 1396
	if (bioset_integrity_create(bs, bio_pool_size))
		goto bad;

1397
	if (!biovec_create_pools(bs, bvec_pool_size))
L
Linus Torvalds 已提交
1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414
		return bs;

bad:
	bioset_free(bs);
	return NULL;
}

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;

		size = bvs->nr_vecs * sizeof(struct bio_vec);
		bvs->slab = kmem_cache_create(bvs->name, size, 0,
1415
                                SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
1416 1417 1418 1419 1420
	}
}

static int __init init_bio(void)
{
1421
	bio_slab = KMEM_CACHE(bio, SLAB_HWCACHE_ALIGN|SLAB_PANIC);
L
Linus Torvalds 已提交
1422

1423
	bio_integrity_init_slab();
L
Linus Torvalds 已提交
1424 1425
	biovec_init_slabs();

1426
	fs_bio_set = bioset_create(BIO_POOL_SIZE, 2);
L
Linus Torvalds 已提交
1427 1428 1429
	if (!fs_bio_set)
		panic("bio: can't allocate bios\n");

1430 1431
	bio_split_pool = mempool_create_kmalloc_pool(BIO_SPLIT_ENTRIES,
						     sizeof(struct bio_pair));
L
Linus Torvalds 已提交
1432 1433 1434 1435 1436 1437 1438 1439 1440
	if (!bio_split_pool)
		panic("bio: can't create split pool\n");

	return 0;
}

subsys_initcall(init_bio);

EXPORT_SYMBOL(bio_alloc);
J
Jens Axboe 已提交
1441
EXPORT_SYMBOL(bio_kmalloc);
L
Linus Torvalds 已提交
1442
EXPORT_SYMBOL(bio_put);
P
Peter Osterlund 已提交
1443
EXPORT_SYMBOL(bio_free);
L
Linus Torvalds 已提交
1444 1445 1446 1447 1448 1449
EXPORT_SYMBOL(bio_endio);
EXPORT_SYMBOL(bio_init);
EXPORT_SYMBOL(__bio_clone);
EXPORT_SYMBOL(bio_clone);
EXPORT_SYMBOL(bio_phys_segments);
EXPORT_SYMBOL(bio_add_page);
1450
EXPORT_SYMBOL(bio_add_pc_page);
L
Linus Torvalds 已提交
1451
EXPORT_SYMBOL(bio_get_nr_vecs);
J
Jens Axboe 已提交
1452 1453
EXPORT_SYMBOL(bio_map_user);
EXPORT_SYMBOL(bio_unmap_user);
M
Mike Christie 已提交
1454
EXPORT_SYMBOL(bio_map_kern);
1455
EXPORT_SYMBOL(bio_copy_kern);
L
Linus Torvalds 已提交
1456 1457 1458 1459 1460 1461 1462 1463
EXPORT_SYMBOL(bio_pair_release);
EXPORT_SYMBOL(bio_split);
EXPORT_SYMBOL(bio_split_pool);
EXPORT_SYMBOL(bio_copy_user);
EXPORT_SYMBOL(bio_uncopy_user);
EXPORT_SYMBOL(bioset_create);
EXPORT_SYMBOL(bioset_free);
EXPORT_SYMBOL(bio_alloc_bioset);