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);
133 #define bvec_to_phys(bv) (page_to_phys((bv)->bv_page) + (unsigned long) (bv)->bv_offset)
139 /* Default implementation of BIOVEC_PHYS_MERGEABLE */
140 #define __BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
141 ((bvec_to_phys((vec1)) + (vec1)->bv_len) == bvec_to_phys((vec2)))
144 * allow arch override, for eg virtualized architectures (put in asm/io.h)
146 #ifndef BIOVEC_PHYS_MERGEABLE
147 #define BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
148 __BIOVEC_PHYS_MERGEABLE(vec1, vec2)
151 #define __BIO_SEG_BOUNDARY(addr1, addr2, mask) \
152 (((addr1) | (mask)) == (((addr2) - 1) | (mask)))
153 #define BIOVEC_SEG_BOUNDARY(q, b1, b2) \
154 __BIO_SEG_BOUNDARY(bvec_to_phys((b1)), bvec_to_phys((b2)) + (b2)->bv_len, queue_segment_boundary((q)))
157 * drivers should _never_ use the all version - the bio may have been split
158 * before it got to the driver and the driver won't own all of it
160 #define bio_for_each_segment_all(bvl, bio, i) \
161 for (i = 0, bvl = (bio)->bi_io_vec; i < (bio)->bi_vcnt; i++, bvl++)
163 static inline void bio_advance_iter(struct bio *bio, struct bvec_iter *iter,
166 iter->bi_sector += bytes >> 9;
168 if (bio_no_advance_iter(bio)) {
169 iter->bi_size -= bytes;
170 iter->bi_done += bytes;
172 bvec_iter_advance(bio->bi_io_vec, iter, bytes);
173 /* TODO: It is reasonable to complete bio with error here. */
177 static inline bool bio_rewind_iter(struct bio *bio, struct bvec_iter *iter,
180 iter->bi_sector -= bytes >> 9;
182 if (bio_no_advance_iter(bio)) {
183 iter->bi_size += bytes;
184 iter->bi_done -= bytes;
188 return bvec_iter_rewind(bio->bi_io_vec, iter, bytes);
191 #define __bio_for_each_segment(bvl, bio, iter, start) \
192 for (iter = (start); \
194 ((bvl = bio_iter_iovec((bio), (iter))), 1); \
195 bio_advance_iter((bio), &(iter), (bvl).bv_len))
197 #define bio_for_each_segment(bvl, bio, iter) \
198 __bio_for_each_segment(bvl, bio, iter, (bio)->bi_iter)
200 #define bio_iter_last(bvec, iter) ((iter).bi_size == (bvec).bv_len)
202 static inline unsigned bio_segments(struct bio *bio)
206 struct bvec_iter iter;
209 * We special case discard/write same/write zeroes, because they
210 * interpret bi_size differently:
213 switch (bio_op(bio)) {
215 case REQ_OP_SECURE_ERASE:
216 case REQ_OP_WRITE_ZEROES:
218 case REQ_OP_WRITE_SAME:
224 bio_for_each_segment(bv, bio, iter)
231 * get a reference to a bio, so it won't disappear. the intended use is
235 * submit_bio(rw, bio);
236 * if (bio->bi_flags ...)
240 * without the bio_get(), it could potentially complete I/O before submit_bio
241 * returns. and then bio would be freed memory when if (bio->bi_flags ...)
244 static inline void bio_get(struct bio *bio)
246 bio->bi_flags |= (1 << BIO_REFFED);
247 smp_mb__before_atomic();
248 atomic_inc(&bio->__bi_cnt);
251 static inline void bio_cnt_set(struct bio *bio, unsigned int count)
254 bio->bi_flags |= (1 << BIO_REFFED);
255 smp_mb__before_atomic();
257 atomic_set(&bio->__bi_cnt, count);
260 static inline bool bio_flagged(struct bio *bio, unsigned int bit)
262 return (bio->bi_flags & (1U << bit)) != 0;
265 static inline void bio_set_flag(struct bio *bio, unsigned int bit)
267 bio->bi_flags |= (1U << bit);
270 static inline void bio_clear_flag(struct bio *bio, unsigned int bit)
272 bio->bi_flags &= ~(1U << bit);
275 static inline void bio_get_first_bvec(struct bio *bio, struct bio_vec *bv)
277 *bv = bio_iovec(bio);
280 static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
282 struct bvec_iter iter = bio->bi_iter;
285 if (unlikely(!bio_multiple_segments(bio))) {
286 *bv = bio_iovec(bio);
290 bio_advance_iter(bio, &iter, iter.bi_size);
292 if (!iter.bi_bvec_done)
293 idx = iter.bi_idx - 1;
294 else /* in the middle of bvec */
297 *bv = bio->bi_io_vec[idx];
300 * iter.bi_bvec_done records actual length of the last bvec
301 * if this bio ends in the middle of one io vector
303 if (iter.bi_bvec_done)
304 bv->bv_len = iter.bi_bvec_done;
307 static inline unsigned bio_pages_all(struct bio *bio)
309 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
313 static inline struct bio_vec *bio_first_bvec_all(struct bio *bio)
315 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
316 return bio->bi_io_vec;
319 static inline struct page *bio_first_page_all(struct bio *bio)
321 return bio_first_bvec_all(bio)->bv_page;
324 static inline struct bio_vec *bio_last_bvec_all(struct bio *bio)
326 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
327 return &bio->bi_io_vec[bio->bi_vcnt - 1];
331 BIP_BLOCK_INTEGRITY = 1 << 0, /* block layer owns integrity data */
332 BIP_MAPPED_INTEGRITY = 1 << 1, /* ref tag has been remapped */
333 BIP_CTRL_NOCHECK = 1 << 2, /* disable HBA integrity checking */
334 BIP_DISK_NOCHECK = 1 << 3, /* disable disk integrity checking */
335 BIP_IP_CHECKSUM = 1 << 4, /* IP checksum */
339 * bio integrity payload
341 struct bio_integrity_payload {
342 struct bio *bip_bio; /* parent bio */
344 struct bvec_iter bip_iter;
346 unsigned short bip_slab; /* slab the bip came from */
347 unsigned short bip_vcnt; /* # of integrity bio_vecs */
348 unsigned short bip_max_vcnt; /* integrity bio_vec slots */
349 unsigned short bip_flags; /* control flags */
351 struct work_struct bip_work; /* I/O completion */
353 struct bio_vec *bip_vec;
354 struct bio_vec bip_inline_vecs[0];/* embedded bvec array */
357 #if defined(CONFIG_BLK_DEV_INTEGRITY)
359 static inline struct bio_integrity_payload *bio_integrity(struct bio *bio)
361 if (bio->bi_opf & REQ_INTEGRITY)
362 return bio->bi_integrity;
367 static inline bool bio_integrity_flagged(struct bio *bio, enum bip_flags flag)
369 struct bio_integrity_payload *bip = bio_integrity(bio);
372 return bip->bip_flags & flag;
377 static inline sector_t bip_get_seed(struct bio_integrity_payload *bip)
379 return bip->bip_iter.bi_sector;
382 static inline void bip_set_seed(struct bio_integrity_payload *bip,
385 bip->bip_iter.bi_sector = seed;
388 #endif /* CONFIG_BLK_DEV_INTEGRITY */
390 extern void bio_trim(struct bio *bio, int offset, int size);
391 extern struct bio *bio_split(struct bio *bio, int sectors,
392 gfp_t gfp, struct bio_set *bs);
395 * bio_next_split - get next @sectors from a bio, splitting if necessary
397 * @sectors: number of sectors to split from the front of @bio
399 * @bs: bio set to allocate from
401 * Returns a bio representing the next @sectors of @bio - if the bio is smaller
402 * than @sectors, returns the original bio unchanged.
404 static inline struct bio *bio_next_split(struct bio *bio, int sectors,
405 gfp_t gfp, struct bio_set *bs)
407 if (sectors >= bio_sectors(bio))
410 return bio_split(bio, sectors, gfp, bs);
414 BIOSET_NEED_BVECS = BIT(0),
415 BIOSET_NEED_RESCUER = BIT(1),
417 extern int bioset_init(struct bio_set *, unsigned int, unsigned int, int flags);
418 extern void bioset_exit(struct bio_set *);
419 extern int biovec_init_pool(mempool_t *pool, int pool_entries);
421 extern struct bio *bio_alloc_bioset(gfp_t, unsigned int, struct bio_set *);
422 extern void bio_put(struct bio *);
424 extern void __bio_clone_fast(struct bio *, struct bio *);
425 extern struct bio *bio_clone_fast(struct bio *, gfp_t, struct bio_set *);
426 extern struct bio *bio_clone_bioset(struct bio *, gfp_t, struct bio_set *bs);
428 extern struct bio_set fs_bio_set;
430 static inline struct bio *bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs)
432 return bio_alloc_bioset(gfp_mask, nr_iovecs, &fs_bio_set);
435 static inline struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned int nr_iovecs)
437 return bio_alloc_bioset(gfp_mask, nr_iovecs, NULL);
440 static inline struct bio *bio_clone_kmalloc(struct bio *bio, gfp_t gfp_mask)
442 return bio_clone_bioset(bio, gfp_mask, NULL);
446 extern blk_qc_t submit_bio(struct bio *);
448 extern void bio_endio(struct bio *);
450 static inline void bio_io_error(struct bio *bio)
452 bio->bi_status = BLK_STS_IOERR;
456 static inline void bio_wouldblock_error(struct bio *bio)
458 bio->bi_status = BLK_STS_AGAIN;
462 struct request_queue;
463 extern int bio_phys_segments(struct request_queue *, struct bio *);
465 extern int submit_bio_wait(struct bio *bio);
466 extern void bio_advance(struct bio *, unsigned);
468 extern void bio_init(struct bio *bio, struct bio_vec *table,
469 unsigned short max_vecs);
470 extern void bio_uninit(struct bio *);
471 extern void bio_reset(struct bio *);
472 void bio_chain(struct bio *, struct bio *);
474 extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
475 extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
476 unsigned int, unsigned int);
477 int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter);
479 extern struct bio *bio_map_user_iov(struct request_queue *,
480 struct iov_iter *, gfp_t);
481 extern void bio_unmap_user(struct bio *);
482 extern struct bio *bio_map_kern(struct request_queue *, void *, unsigned int,
484 extern struct bio *bio_copy_kern(struct request_queue *, void *, unsigned int,
486 extern void bio_set_pages_dirty(struct bio *bio);
487 extern void bio_check_pages_dirty(struct bio *bio);
489 void generic_start_io_acct(struct request_queue *q, int rw,
490 unsigned long sectors, struct hd_struct *part);
491 void generic_end_io_acct(struct request_queue *q, int rw,
492 struct hd_struct *part,
493 unsigned long start_time);
495 #ifndef ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
496 # error "You should define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE for your platform"
498 #if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
499 extern void bio_flush_dcache_pages(struct bio *bi);
501 static inline void bio_flush_dcache_pages(struct bio *bi)
506 extern void bio_copy_data_iter(struct bio *dst, struct bvec_iter *dst_iter,
507 struct bio *src, struct bvec_iter *src_iter);
508 extern void bio_copy_data(struct bio *dst, struct bio *src);
509 extern void bio_list_copy_data(struct bio *dst, struct bio *src);
510 extern void bio_free_pages(struct bio *bio);
512 extern struct bio *bio_copy_user_iov(struct request_queue *,
513 struct rq_map_data *,
516 extern int bio_uncopy_user(struct bio *);
517 void zero_fill_bio_iter(struct bio *bio, struct bvec_iter iter);
519 static inline void zero_fill_bio(struct bio *bio)
521 zero_fill_bio_iter(bio, bio->bi_iter);
524 extern struct bio_vec *bvec_alloc(gfp_t, int, unsigned long *, mempool_t *);
525 extern void bvec_free(mempool_t *, struct bio_vec *, unsigned int);
526 extern unsigned int bvec_nr_vecs(unsigned short idx);
527 extern const char *bio_devname(struct bio *bio, char *buffer);
529 #define bio_set_dev(bio, bdev) \
531 if ((bio)->bi_disk != (bdev)->bd_disk) \
532 bio_clear_flag(bio, BIO_THROTTLED);\
533 (bio)->bi_disk = (bdev)->bd_disk; \
534 (bio)->bi_partno = (bdev)->bd_partno; \
537 #define bio_copy_dev(dst, src) \
539 (dst)->bi_disk = (src)->bi_disk; \
540 (dst)->bi_partno = (src)->bi_partno; \
543 #define bio_dev(bio) \
544 disk_devt((bio)->bi_disk)
546 #ifdef CONFIG_BLK_CGROUP
547 int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css);
548 void bio_disassociate_task(struct bio *bio);
549 void bio_clone_blkcg_association(struct bio *dst, struct bio *src);
550 #else /* CONFIG_BLK_CGROUP */
551 static inline int bio_associate_blkcg(struct bio *bio,
552 struct cgroup_subsys_state *blkcg_css) { return 0; }
553 static inline void bio_disassociate_task(struct bio *bio) { }
554 static inline void bio_clone_blkcg_association(struct bio *dst,
556 #endif /* CONFIG_BLK_CGROUP */
558 #ifdef CONFIG_HIGHMEM
560 * remember never ever reenable interrupts between a bvec_kmap_irq and
563 static inline char *bvec_kmap_irq(struct bio_vec *bvec, unsigned long *flags)
568 * might not be a highmem page, but the preempt/irq count
569 * balancing is a lot nicer this way
571 local_irq_save(*flags);
572 addr = (unsigned long) kmap_atomic(bvec->bv_page);
574 BUG_ON(addr & ~PAGE_MASK);
576 return (char *) addr + bvec->bv_offset;
579 static inline void bvec_kunmap_irq(char *buffer, unsigned long *flags)
581 unsigned long ptr = (unsigned long) buffer & PAGE_MASK;
583 kunmap_atomic((void *) ptr);
584 local_irq_restore(*flags);
588 static inline char *bvec_kmap_irq(struct bio_vec *bvec, unsigned long *flags)
590 return page_address(bvec->bv_page) + bvec->bv_offset;
593 static inline void bvec_kunmap_irq(char *buffer, unsigned long *flags)
600 * BIO list management for use by remapping drivers (e.g. DM or MD) and loop.
602 * A bio_list anchors a singly-linked list of bios chained through the bi_next
603 * member of the bio. The bio_list also caches the last list member to allow
604 * fast access to the tail.
611 static inline int bio_list_empty(const struct bio_list *bl)
613 return bl->head == NULL;
616 static inline void bio_list_init(struct bio_list *bl)
618 bl->head = bl->tail = NULL;
621 #define BIO_EMPTY_LIST { NULL, NULL }
623 #define bio_list_for_each(bio, bl) \
624 for (bio = (bl)->head; bio; bio = bio->bi_next)
626 static inline unsigned bio_list_size(const struct bio_list *bl)
631 bio_list_for_each(bio, bl)
637 static inline void bio_list_add(struct bio_list *bl, struct bio *bio)
642 bl->tail->bi_next = bio;
649 static inline void bio_list_add_head(struct bio_list *bl, struct bio *bio)
651 bio->bi_next = bl->head;
659 static inline void bio_list_merge(struct bio_list *bl, struct bio_list *bl2)
665 bl->tail->bi_next = bl2->head;
667 bl->head = bl2->head;
669 bl->tail = bl2->tail;
672 static inline void bio_list_merge_head(struct bio_list *bl,
673 struct bio_list *bl2)
679 bl2->tail->bi_next = bl->head;
681 bl->tail = bl2->tail;
683 bl->head = bl2->head;
686 static inline struct bio *bio_list_peek(struct bio_list *bl)
691 static inline struct bio *bio_list_pop(struct bio_list *bl)
693 struct bio *bio = bl->head;
696 bl->head = bl->head->bi_next;
706 static inline struct bio *bio_list_get(struct bio_list *bl)
708 struct bio *bio = bl->head;
710 bl->head = bl->tail = NULL;
716 * Increment chain count for the bio. Make sure the CHAIN flag update
717 * is visible before the raised count.
719 static inline void bio_inc_remaining(struct bio *bio)
721 bio_set_flag(bio, BIO_CHAIN);
722 smp_mb__before_atomic();
723 atomic_inc(&bio->__bi_remaining);
727 * bio_set is used to allow other portions of the IO system to
728 * allocate their own private memory pools for bio and iovec structures.
729 * These memory pools in turn all allocate from the bio_slab
730 * and the bvec_slabs[].
732 #define BIO_POOL_SIZE 2
735 struct kmem_cache *bio_slab;
736 unsigned int front_pad;
740 #if defined(CONFIG_BLK_DEV_INTEGRITY)
741 mempool_t bio_integrity_pool;
742 mempool_t bvec_integrity_pool;
746 * Deadlock avoidance for stacking block drivers: see comments in
747 * bio_alloc_bioset() for details
749 spinlock_t rescue_lock;
750 struct bio_list rescue_list;
751 struct work_struct rescue_work;
752 struct workqueue_struct *rescue_workqueue;
758 struct kmem_cache *slab;
761 static inline bool bioset_initialized(struct bio_set *bs)
763 return bs->bio_slab != NULL;
767 * a small number of entries is fine, not going to be performance critical.
768 * basically we just need to survive
770 #define BIO_SPLIT_ENTRIES 2
772 #if defined(CONFIG_BLK_DEV_INTEGRITY)
774 #define bip_for_each_vec(bvl, bip, iter) \
775 for_each_bvec(bvl, (bip)->bip_vec, iter, (bip)->bip_iter)
777 #define bio_for_each_integrity_vec(_bvl, _bio, _iter) \
779 bip_for_each_vec(_bvl, _bio->bi_integrity, _iter)
781 extern struct bio_integrity_payload *bio_integrity_alloc(struct bio *, gfp_t, unsigned int);
782 extern int bio_integrity_add_page(struct bio *, struct page *, unsigned int, unsigned int);
783 extern bool bio_integrity_prep(struct bio *);
784 extern void bio_integrity_advance(struct bio *, unsigned int);
785 extern void bio_integrity_trim(struct bio *);
786 extern int bio_integrity_clone(struct bio *, struct bio *, gfp_t);
787 extern int bioset_integrity_create(struct bio_set *, int);
788 extern void bioset_integrity_free(struct bio_set *);
789 extern void bio_integrity_init(void);
791 #else /* CONFIG_BLK_DEV_INTEGRITY */
793 static inline void *bio_integrity(struct bio *bio)
798 static inline int bioset_integrity_create(struct bio_set *bs, int pool_size)
803 static inline void bioset_integrity_free (struct bio_set *bs)
808 static inline bool bio_integrity_prep(struct bio *bio)
813 static inline int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
819 static inline void bio_integrity_advance(struct bio *bio,
820 unsigned int bytes_done)
825 static inline void bio_integrity_trim(struct bio *bio)
830 static inline void bio_integrity_init(void)
835 static inline bool bio_integrity_flagged(struct bio *bio, enum bip_flags flag)
840 static inline void *bio_integrity_alloc(struct bio * bio, gfp_t gfp,
843 return ERR_PTR(-EINVAL);
846 static inline int bio_integrity_add_page(struct bio *bio, struct page *page,
847 unsigned int len, unsigned int offset)
852 #endif /* CONFIG_BLK_DEV_INTEGRITY */
854 #endif /* CONFIG_BLOCK */
855 #endif /* __LINUX_BIO_H */