2 * Copyright (C) 2001 Jens Axboe <axboe@suse.de>
4 * This program is free software; you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License version 2 as
6 * published by the Free Software Foundation.
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
14 * You should have received a copy of the GNU General Public Licens
15 * along with this program; if not, write to the Free Software
16 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-
21 #include <linux/highmem.h>
22 #include <linux/mempool.h>
23 #include <linux/ioprio.h>
24 #include <linux/bug.h>
30 /* struct bio, bio_vec and BIO_* flags are defined in blk_types.h */
31 #include <linux/blk_types.h>
36 #define BIO_BUG_ON BUG_ON
41 #ifdef CONFIG_THP_SWAP
42 #if HPAGE_PMD_NR > 256
43 #define BIO_MAX_PAGES HPAGE_PMD_NR
45 #define BIO_MAX_PAGES 256
48 #define BIO_MAX_PAGES 256
51 #define bio_prio(bio) (bio)->bi_ioprio
52 #define bio_set_prio(bio, prio) ((bio)->bi_ioprio = prio)
54 #define bio_iter_iovec(bio, iter) \
55 bvec_iter_bvec((bio)->bi_io_vec, (iter))
57 #define bio_iter_page(bio, iter) \
58 bvec_iter_page((bio)->bi_io_vec, (iter))
59 #define bio_iter_len(bio, iter) \
60 bvec_iter_len((bio)->bi_io_vec, (iter))
61 #define bio_iter_offset(bio, iter) \
62 bvec_iter_offset((bio)->bi_io_vec, (iter))
64 #define bio_page(bio) bio_iter_page((bio), (bio)->bi_iter)
65 #define bio_offset(bio) bio_iter_offset((bio), (bio)->bi_iter)
66 #define bio_iovec(bio) bio_iter_iovec((bio), (bio)->bi_iter)
68 #define bio_multiple_segments(bio) \
69 ((bio)->bi_iter.bi_size != bio_iovec(bio).bv_len)
71 #define bvec_iter_sectors(iter) ((iter).bi_size >> 9)
72 #define bvec_iter_end_sector(iter) ((iter).bi_sector + bvec_iter_sectors((iter)))
74 #define bio_sectors(bio) bvec_iter_sectors((bio)->bi_iter)
75 #define bio_end_sector(bio) bvec_iter_end_sector((bio)->bi_iter)
78 * Return the data direction, READ or WRITE.
80 #define bio_data_dir(bio) \
81 (op_is_write(bio_op(bio)) ? WRITE : READ)
84 * Check whether this bio carries any data or not. A NULL bio is allowed.
86 static inline bool bio_has_data(struct bio *bio)
89 bio->bi_iter.bi_size &&
90 bio_op(bio) != REQ_OP_DISCARD &&
91 bio_op(bio) != REQ_OP_SECURE_ERASE &&
92 bio_op(bio) != REQ_OP_WRITE_ZEROES)
98 static inline bool bio_no_advance_iter(struct bio *bio)
100 return bio_op(bio) == REQ_OP_DISCARD ||
101 bio_op(bio) == REQ_OP_SECURE_ERASE ||
102 bio_op(bio) == REQ_OP_WRITE_SAME ||
103 bio_op(bio) == REQ_OP_WRITE_ZEROES;
106 static inline bool bio_mergeable(struct bio *bio)
108 if (bio->bi_opf & REQ_NOMERGE_FLAGS)
114 static inline unsigned int bio_cur_bytes(struct bio *bio)
116 if (bio_has_data(bio))
117 return bio_iovec(bio).bv_len;
118 else /* dataless requests such as discard */
119 return bio->bi_iter.bi_size;
122 static inline void *bio_data(struct bio *bio)
124 if (bio_has_data(bio))
125 return page_address(bio_page(bio)) + bio_offset(bio);
130 static inline bool bio_full(struct bio *bio)
132 return bio->bi_vcnt >= bio->bi_max_vecs;
138 #define bvec_to_phys(bv) (page_to_phys((bv)->bv_page) + (unsigned long) (bv)->bv_offset)
141 * drivers should _never_ use the all version - the bio may have been split
142 * before it got to the driver and the driver won't own all of it
144 #define bio_for_each_segment_all(bvl, bio, i) \
145 for (i = 0, bvl = (bio)->bi_io_vec; i < (bio)->bi_vcnt; i++, bvl++)
147 static inline void bio_advance_iter(struct bio *bio, struct bvec_iter *iter,
150 iter->bi_sector += bytes >> 9;
152 if (bio_no_advance_iter(bio))
153 iter->bi_size -= bytes;
155 bvec_iter_advance(bio->bi_io_vec, iter, bytes);
156 /* TODO: It is reasonable to complete bio with error here. */
159 #define __bio_for_each_segment(bvl, bio, iter, start) \
160 for (iter = (start); \
162 ((bvl = bio_iter_iovec((bio), (iter))), 1); \
163 bio_advance_iter((bio), &(iter), (bvl).bv_len))
165 #define bio_for_each_segment(bvl, bio, iter) \
166 __bio_for_each_segment(bvl, bio, iter, (bio)->bi_iter)
168 #define bio_iter_last(bvec, iter) ((iter).bi_size == (bvec).bv_len)
170 static inline unsigned bio_segments(struct bio *bio)
174 struct bvec_iter iter;
177 * We special case discard/write same/write zeroes, because they
178 * interpret bi_size differently:
181 switch (bio_op(bio)) {
183 case REQ_OP_SECURE_ERASE:
184 case REQ_OP_WRITE_ZEROES:
186 case REQ_OP_WRITE_SAME:
192 bio_for_each_segment(bv, bio, iter)
199 * get a reference to a bio, so it won't disappear. the intended use is
203 * submit_bio(rw, bio);
204 * if (bio->bi_flags ...)
208 * without the bio_get(), it could potentially complete I/O before submit_bio
209 * returns. and then bio would be freed memory when if (bio->bi_flags ...)
212 static inline void bio_get(struct bio *bio)
214 bio->bi_flags |= (1 << BIO_REFFED);
215 smp_mb__before_atomic();
216 atomic_inc(&bio->__bi_cnt);
219 static inline void bio_cnt_set(struct bio *bio, unsigned int count)
222 bio->bi_flags |= (1 << BIO_REFFED);
223 smp_mb__before_atomic();
225 atomic_set(&bio->__bi_cnt, count);
228 static inline bool bio_flagged(struct bio *bio, unsigned int bit)
230 return (bio->bi_flags & (1U << bit)) != 0;
233 static inline void bio_set_flag(struct bio *bio, unsigned int bit)
235 bio->bi_flags |= (1U << bit);
238 static inline void bio_clear_flag(struct bio *bio, unsigned int bit)
240 bio->bi_flags &= ~(1U << bit);
243 static inline void bio_get_first_bvec(struct bio *bio, struct bio_vec *bv)
245 *bv = bio_iovec(bio);
248 static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
250 struct bvec_iter iter = bio->bi_iter;
253 if (unlikely(!bio_multiple_segments(bio))) {
254 *bv = bio_iovec(bio);
258 bio_advance_iter(bio, &iter, iter.bi_size);
260 if (!iter.bi_bvec_done)
261 idx = iter.bi_idx - 1;
262 else /* in the middle of bvec */
265 *bv = bio->bi_io_vec[idx];
268 * iter.bi_bvec_done records actual length of the last bvec
269 * if this bio ends in the middle of one io vector
271 if (iter.bi_bvec_done)
272 bv->bv_len = iter.bi_bvec_done;
275 static inline unsigned bio_pages_all(struct bio *bio)
277 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
281 static inline struct bio_vec *bio_first_bvec_all(struct bio *bio)
283 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
284 return bio->bi_io_vec;
287 static inline struct page *bio_first_page_all(struct bio *bio)
289 return bio_first_bvec_all(bio)->bv_page;
292 static inline struct bio_vec *bio_last_bvec_all(struct bio *bio)
294 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
295 return &bio->bi_io_vec[bio->bi_vcnt - 1];
299 BIP_BLOCK_INTEGRITY = 1 << 0, /* block layer owns integrity data */
300 BIP_MAPPED_INTEGRITY = 1 << 1, /* ref tag has been remapped */
301 BIP_CTRL_NOCHECK = 1 << 2, /* disable HBA integrity checking */
302 BIP_DISK_NOCHECK = 1 << 3, /* disable disk integrity checking */
303 BIP_IP_CHECKSUM = 1 << 4, /* IP checksum */
307 * bio integrity payload
309 struct bio_integrity_payload {
310 struct bio *bip_bio; /* parent bio */
312 struct bvec_iter bip_iter;
314 unsigned short bip_slab; /* slab the bip came from */
315 unsigned short bip_vcnt; /* # of integrity bio_vecs */
316 unsigned short bip_max_vcnt; /* integrity bio_vec slots */
317 unsigned short bip_flags; /* control flags */
319 struct bvec_iter bio_iter; /* for rewinding parent bio */
321 struct work_struct bip_work; /* I/O completion */
323 struct bio_vec *bip_vec;
324 struct bio_vec bip_inline_vecs[0];/* embedded bvec array */
327 #if defined(CONFIG_BLK_DEV_INTEGRITY)
329 static inline struct bio_integrity_payload *bio_integrity(struct bio *bio)
331 if (bio->bi_opf & REQ_INTEGRITY)
332 return bio->bi_integrity;
337 static inline bool bio_integrity_flagged(struct bio *bio, enum bip_flags flag)
339 struct bio_integrity_payload *bip = bio_integrity(bio);
342 return bip->bip_flags & flag;
347 static inline sector_t bip_get_seed(struct bio_integrity_payload *bip)
349 return bip->bip_iter.bi_sector;
352 static inline void bip_set_seed(struct bio_integrity_payload *bip,
355 bip->bip_iter.bi_sector = seed;
358 #endif /* CONFIG_BLK_DEV_INTEGRITY */
360 extern void bio_trim(struct bio *bio, int offset, int size);
361 extern struct bio *bio_split(struct bio *bio, int sectors,
362 gfp_t gfp, struct bio_set *bs);
365 * bio_next_split - get next @sectors from a bio, splitting if necessary
367 * @sectors: number of sectors to split from the front of @bio
369 * @bs: bio set to allocate from
371 * Returns a bio representing the next @sectors of @bio - if the bio is smaller
372 * than @sectors, returns the original bio unchanged.
374 static inline struct bio *bio_next_split(struct bio *bio, int sectors,
375 gfp_t gfp, struct bio_set *bs)
377 if (sectors >= bio_sectors(bio))
380 return bio_split(bio, sectors, gfp, bs);
384 BIOSET_NEED_BVECS = BIT(0),
385 BIOSET_NEED_RESCUER = BIT(1),
387 extern int bioset_init(struct bio_set *, unsigned int, unsigned int, int flags);
388 extern void bioset_exit(struct bio_set *);
389 extern int biovec_init_pool(mempool_t *pool, int pool_entries);
390 extern int bioset_init_from_src(struct bio_set *bs, struct bio_set *src);
392 extern struct bio *bio_alloc_bioset(gfp_t, unsigned int, struct bio_set *);
393 extern void bio_put(struct bio *);
395 extern void __bio_clone_fast(struct bio *, struct bio *);
396 extern struct bio *bio_clone_fast(struct bio *, gfp_t, struct bio_set *);
398 extern struct bio_set fs_bio_set;
400 static inline struct bio *bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs)
402 return bio_alloc_bioset(gfp_mask, nr_iovecs, &fs_bio_set);
405 static inline struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned int nr_iovecs)
407 return bio_alloc_bioset(gfp_mask, nr_iovecs, NULL);
410 extern blk_qc_t submit_bio(struct bio *);
412 extern void bio_endio(struct bio *);
414 static inline void bio_io_error(struct bio *bio)
416 bio->bi_status = BLK_STS_IOERR;
420 static inline void bio_wouldblock_error(struct bio *bio)
422 bio->bi_status = BLK_STS_AGAIN;
426 struct request_queue;
427 extern int bio_phys_segments(struct request_queue *, struct bio *);
429 extern int submit_bio_wait(struct bio *bio);
430 extern void bio_advance(struct bio *, unsigned);
432 extern void bio_init(struct bio *bio, struct bio_vec *table,
433 unsigned short max_vecs);
434 extern void bio_uninit(struct bio *);
435 extern void bio_reset(struct bio *);
436 void bio_chain(struct bio *, struct bio *);
438 extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
439 extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
440 unsigned int, unsigned int);
441 bool __bio_try_merge_page(struct bio *bio, struct page *page,
442 unsigned int len, unsigned int off);
443 void __bio_add_page(struct bio *bio, struct page *page,
444 unsigned int len, unsigned int off);
445 int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter);
447 extern struct bio *bio_map_user_iov(struct request_queue *,
448 struct iov_iter *, gfp_t);
449 extern void bio_unmap_user(struct bio *);
450 extern struct bio *bio_map_kern(struct request_queue *, void *, unsigned int,
452 extern struct bio *bio_copy_kern(struct request_queue *, void *, unsigned int,
454 extern void bio_set_pages_dirty(struct bio *bio);
455 extern void bio_check_pages_dirty(struct bio *bio);
457 void generic_start_io_acct(struct request_queue *q, int op,
458 unsigned long sectors, struct hd_struct *part);
459 void generic_end_io_acct(struct request_queue *q, int op,
460 struct hd_struct *part,
461 unsigned long start_time);
463 #ifndef ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
464 # error "You should define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE for your platform"
466 #if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
467 extern void bio_flush_dcache_pages(struct bio *bi);
469 static inline void bio_flush_dcache_pages(struct bio *bi)
474 extern void bio_copy_data_iter(struct bio *dst, struct bvec_iter *dst_iter,
475 struct bio *src, struct bvec_iter *src_iter);
476 extern void bio_copy_data(struct bio *dst, struct bio *src);
477 extern void bio_list_copy_data(struct bio *dst, struct bio *src);
478 extern void bio_free_pages(struct bio *bio);
480 extern struct bio *bio_copy_user_iov(struct request_queue *,
481 struct rq_map_data *,
484 extern int bio_uncopy_user(struct bio *);
485 void zero_fill_bio_iter(struct bio *bio, struct bvec_iter iter);
487 static inline void zero_fill_bio(struct bio *bio)
489 zero_fill_bio_iter(bio, bio->bi_iter);
492 extern struct bio_vec *bvec_alloc(gfp_t, int, unsigned long *, mempool_t *);
493 extern void bvec_free(mempool_t *, struct bio_vec *, unsigned int);
494 extern unsigned int bvec_nr_vecs(unsigned short idx);
495 extern const char *bio_devname(struct bio *bio, char *buffer);
497 #define bio_set_dev(bio, bdev) \
499 if ((bio)->bi_disk != (bdev)->bd_disk) \
500 bio_clear_flag(bio, BIO_THROTTLED);\
501 (bio)->bi_disk = (bdev)->bd_disk; \
502 (bio)->bi_partno = (bdev)->bd_partno; \
505 #define bio_copy_dev(dst, src) \
507 (dst)->bi_disk = (src)->bi_disk; \
508 (dst)->bi_partno = (src)->bi_partno; \
511 #define bio_dev(bio) \
512 disk_devt((bio)->bi_disk)
514 #if defined(CONFIG_MEMCG) && defined(CONFIG_BLK_CGROUP)
515 int bio_associate_blkg_from_page(struct bio *bio, struct page *page);
517 static inline int bio_associate_blkg_from_page(struct bio *bio,
518 struct page *page) { return 0; }
521 #ifdef CONFIG_BLK_CGROUP
522 int bio_associate_blkg(struct bio *bio, struct blkcg_gq *blkg);
523 int bio_associate_blkg_from_css(struct bio *bio,
524 struct cgroup_subsys_state *css);
525 int bio_associate_create_blkg(struct request_queue *q, struct bio *bio);
526 void bio_disassociate_task(struct bio *bio);
527 void bio_clone_blkg_association(struct bio *dst, struct bio *src);
528 #else /* CONFIG_BLK_CGROUP */
529 static inline int bio_associate_blkg_from_css(struct bio *bio,
530 struct cgroup_subsys_state *css)
532 static inline int bio_associate_create_blkg(struct request_queue *q,
533 struct bio *bio) { return 0; }
534 static inline void bio_disassociate_task(struct bio *bio) { }
535 static inline void bio_clone_blkg_association(struct bio *dst,
537 #endif /* CONFIG_BLK_CGROUP */
539 #ifdef CONFIG_HIGHMEM
541 * remember never ever reenable interrupts between a bvec_kmap_irq and
544 static inline char *bvec_kmap_irq(struct bio_vec *bvec, unsigned long *flags)
549 * might not be a highmem page, but the preempt/irq count
550 * balancing is a lot nicer this way
552 local_irq_save(*flags);
553 addr = (unsigned long) kmap_atomic(bvec->bv_page);
555 BUG_ON(addr & ~PAGE_MASK);
557 return (char *) addr + bvec->bv_offset;
560 static inline void bvec_kunmap_irq(char *buffer, unsigned long *flags)
562 unsigned long ptr = (unsigned long) buffer & PAGE_MASK;
564 kunmap_atomic((void *) ptr);
565 local_irq_restore(*flags);
569 static inline char *bvec_kmap_irq(struct bio_vec *bvec, unsigned long *flags)
571 return page_address(bvec->bv_page) + bvec->bv_offset;
574 static inline void bvec_kunmap_irq(char *buffer, unsigned long *flags)
581 * BIO list management for use by remapping drivers (e.g. DM or MD) and loop.
583 * A bio_list anchors a singly-linked list of bios chained through the bi_next
584 * member of the bio. The bio_list also caches the last list member to allow
585 * fast access to the tail.
592 static inline int bio_list_empty(const struct bio_list *bl)
594 return bl->head == NULL;
597 static inline void bio_list_init(struct bio_list *bl)
599 bl->head = bl->tail = NULL;
602 #define BIO_EMPTY_LIST { NULL, NULL }
604 #define bio_list_for_each(bio, bl) \
605 for (bio = (bl)->head; bio; bio = bio->bi_next)
607 static inline unsigned bio_list_size(const struct bio_list *bl)
612 bio_list_for_each(bio, bl)
618 static inline void bio_list_add(struct bio_list *bl, struct bio *bio)
623 bl->tail->bi_next = bio;
630 static inline void bio_list_add_head(struct bio_list *bl, struct bio *bio)
632 bio->bi_next = bl->head;
640 static inline void bio_list_merge(struct bio_list *bl, struct bio_list *bl2)
646 bl->tail->bi_next = bl2->head;
648 bl->head = bl2->head;
650 bl->tail = bl2->tail;
653 static inline void bio_list_merge_head(struct bio_list *bl,
654 struct bio_list *bl2)
660 bl2->tail->bi_next = bl->head;
662 bl->tail = bl2->tail;
664 bl->head = bl2->head;
667 static inline struct bio *bio_list_peek(struct bio_list *bl)
672 static inline struct bio *bio_list_pop(struct bio_list *bl)
674 struct bio *bio = bl->head;
677 bl->head = bl->head->bi_next;
687 static inline struct bio *bio_list_get(struct bio_list *bl)
689 struct bio *bio = bl->head;
691 bl->head = bl->tail = NULL;
697 * Increment chain count for the bio. Make sure the CHAIN flag update
698 * is visible before the raised count.
700 static inline void bio_inc_remaining(struct bio *bio)
702 bio_set_flag(bio, BIO_CHAIN);
703 smp_mb__before_atomic();
704 atomic_inc(&bio->__bi_remaining);
708 * bio_set is used to allow other portions of the IO system to
709 * allocate their own private memory pools for bio and iovec structures.
710 * These memory pools in turn all allocate from the bio_slab
711 * and the bvec_slabs[].
713 #define BIO_POOL_SIZE 2
716 struct kmem_cache *bio_slab;
717 unsigned int front_pad;
721 #if defined(CONFIG_BLK_DEV_INTEGRITY)
722 mempool_t bio_integrity_pool;
723 mempool_t bvec_integrity_pool;
727 * Deadlock avoidance for stacking block drivers: see comments in
728 * bio_alloc_bioset() for details
730 spinlock_t rescue_lock;
731 struct bio_list rescue_list;
732 struct work_struct rescue_work;
733 struct workqueue_struct *rescue_workqueue;
739 struct kmem_cache *slab;
742 static inline bool bioset_initialized(struct bio_set *bs)
744 return bs->bio_slab != NULL;
748 * a small number of entries is fine, not going to be performance critical.
749 * basically we just need to survive
751 #define BIO_SPLIT_ENTRIES 2
753 #if defined(CONFIG_BLK_DEV_INTEGRITY)
755 #define bip_for_each_vec(bvl, bip, iter) \
756 for_each_bvec(bvl, (bip)->bip_vec, iter, (bip)->bip_iter)
758 #define bio_for_each_integrity_vec(_bvl, _bio, _iter) \
760 bip_for_each_vec(_bvl, _bio->bi_integrity, _iter)
762 extern struct bio_integrity_payload *bio_integrity_alloc(struct bio *, gfp_t, unsigned int);
763 extern int bio_integrity_add_page(struct bio *, struct page *, unsigned int, unsigned int);
764 extern bool bio_integrity_prep(struct bio *);
765 extern void bio_integrity_advance(struct bio *, unsigned int);
766 extern void bio_integrity_trim(struct bio *);
767 extern int bio_integrity_clone(struct bio *, struct bio *, gfp_t);
768 extern int bioset_integrity_create(struct bio_set *, int);
769 extern void bioset_integrity_free(struct bio_set *);
770 extern void bio_integrity_init(void);
772 #else /* CONFIG_BLK_DEV_INTEGRITY */
774 static inline void *bio_integrity(struct bio *bio)
779 static inline int bioset_integrity_create(struct bio_set *bs, int pool_size)
784 static inline void bioset_integrity_free (struct bio_set *bs)
789 static inline bool bio_integrity_prep(struct bio *bio)
794 static inline int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
800 static inline void bio_integrity_advance(struct bio *bio,
801 unsigned int bytes_done)
806 static inline void bio_integrity_trim(struct bio *bio)
811 static inline void bio_integrity_init(void)
816 static inline bool bio_integrity_flagged(struct bio *bio, enum bip_flags flag)
821 static inline void *bio_integrity_alloc(struct bio * bio, gfp_t gfp,
824 return ERR_PTR(-EINVAL);
827 static inline int bio_integrity_add_page(struct bio *bio, struct page *page,
828 unsigned int len, unsigned int offset)
833 #endif /* CONFIG_BLK_DEV_INTEGRITY */
835 #endif /* CONFIG_BLOCK */
836 #endif /* __LINUX_BIO_H */