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)
144 /* Default implementation of BIOVEC_PHYS_MERGEABLE */
145 #define __BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
146 ((bvec_to_phys((vec1)) + (vec1)->bv_len) == bvec_to_phys((vec2)))
149 * allow arch override, for eg virtualized architectures (put in asm/io.h)
151 #ifndef BIOVEC_PHYS_MERGEABLE
152 #define BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
153 __BIOVEC_PHYS_MERGEABLE(vec1, vec2)
156 #define __BIO_SEG_BOUNDARY(addr1, addr2, mask) \
157 (((addr1) | (mask)) == (((addr2) - 1) | (mask)))
158 #define BIOVEC_SEG_BOUNDARY(q, b1, b2) \
159 __BIO_SEG_BOUNDARY(bvec_to_phys((b1)), bvec_to_phys((b2)) + (b2)->bv_len, queue_segment_boundary((q)))
162 * drivers should _never_ use the all version - the bio may have been split
163 * before it got to the driver and the driver won't own all of it
165 #define bio_for_each_segment_all(bvl, bio, i) \
166 for (i = 0, bvl = (bio)->bi_io_vec; i < (bio)->bi_vcnt; i++, bvl++)
168 static inline void bio_advance_iter(struct bio *bio, struct bvec_iter *iter,
171 iter->bi_sector += bytes >> 9;
173 if (bio_no_advance_iter(bio)) {
174 iter->bi_size -= bytes;
175 iter->bi_done += bytes;
177 bvec_iter_advance(bio->bi_io_vec, iter, bytes);
178 /* TODO: It is reasonable to complete bio with error here. */
182 static inline bool bio_rewind_iter(struct bio *bio, struct bvec_iter *iter,
185 iter->bi_sector -= bytes >> 9;
187 if (bio_no_advance_iter(bio)) {
188 iter->bi_size += bytes;
189 iter->bi_done -= bytes;
193 return bvec_iter_rewind(bio->bi_io_vec, iter, bytes);
196 #define __bio_for_each_segment(bvl, bio, iter, start) \
197 for (iter = (start); \
199 ((bvl = bio_iter_iovec((bio), (iter))), 1); \
200 bio_advance_iter((bio), &(iter), (bvl).bv_len))
202 #define bio_for_each_segment(bvl, bio, iter) \
203 __bio_for_each_segment(bvl, bio, iter, (bio)->bi_iter)
205 #define bio_iter_last(bvec, iter) ((iter).bi_size == (bvec).bv_len)
207 static inline unsigned bio_segments(struct bio *bio)
211 struct bvec_iter iter;
214 * We special case discard/write same/write zeroes, because they
215 * interpret bi_size differently:
218 switch (bio_op(bio)) {
220 case REQ_OP_SECURE_ERASE:
221 case REQ_OP_WRITE_ZEROES:
223 case REQ_OP_WRITE_SAME:
229 bio_for_each_segment(bv, bio, iter)
236 * get a reference to a bio, so it won't disappear. the intended use is
240 * submit_bio(rw, bio);
241 * if (bio->bi_flags ...)
245 * without the bio_get(), it could potentially complete I/O before submit_bio
246 * returns. and then bio would be freed memory when if (bio->bi_flags ...)
249 static inline void bio_get(struct bio *bio)
251 bio->bi_flags |= (1 << BIO_REFFED);
252 smp_mb__before_atomic();
253 atomic_inc(&bio->__bi_cnt);
256 static inline void bio_cnt_set(struct bio *bio, unsigned int count)
259 bio->bi_flags |= (1 << BIO_REFFED);
260 smp_mb__before_atomic();
262 atomic_set(&bio->__bi_cnt, count);
265 static inline bool bio_flagged(struct bio *bio, unsigned int bit)
267 return (bio->bi_flags & (1U << bit)) != 0;
270 static inline void bio_set_flag(struct bio *bio, unsigned int bit)
272 bio->bi_flags |= (1U << bit);
275 static inline void bio_clear_flag(struct bio *bio, unsigned int bit)
277 bio->bi_flags &= ~(1U << bit);
280 static inline void bio_get_first_bvec(struct bio *bio, struct bio_vec *bv)
282 *bv = bio_iovec(bio);
285 static inline void bio_get_last_bvec(struct bio *bio, struct bio_vec *bv)
287 struct bvec_iter iter = bio->bi_iter;
290 if (unlikely(!bio_multiple_segments(bio))) {
291 *bv = bio_iovec(bio);
295 bio_advance_iter(bio, &iter, iter.bi_size);
297 if (!iter.bi_bvec_done)
298 idx = iter.bi_idx - 1;
299 else /* in the middle of bvec */
302 *bv = bio->bi_io_vec[idx];
305 * iter.bi_bvec_done records actual length of the last bvec
306 * if this bio ends in the middle of one io vector
308 if (iter.bi_bvec_done)
309 bv->bv_len = iter.bi_bvec_done;
312 static inline unsigned bio_pages_all(struct bio *bio)
314 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
318 static inline struct bio_vec *bio_first_bvec_all(struct bio *bio)
320 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
321 return bio->bi_io_vec;
324 static inline struct page *bio_first_page_all(struct bio *bio)
326 return bio_first_bvec_all(bio)->bv_page;
329 static inline struct bio_vec *bio_last_bvec_all(struct bio *bio)
331 WARN_ON_ONCE(bio_flagged(bio, BIO_CLONED));
332 return &bio->bi_io_vec[bio->bi_vcnt - 1];
336 BIP_BLOCK_INTEGRITY = 1 << 0, /* block layer owns integrity data */
337 BIP_MAPPED_INTEGRITY = 1 << 1, /* ref tag has been remapped */
338 BIP_CTRL_NOCHECK = 1 << 2, /* disable HBA integrity checking */
339 BIP_DISK_NOCHECK = 1 << 3, /* disable disk integrity checking */
340 BIP_IP_CHECKSUM = 1 << 4, /* IP checksum */
344 * bio integrity payload
346 struct bio_integrity_payload {
347 struct bio *bip_bio; /* parent bio */
349 struct bvec_iter bip_iter;
351 unsigned short bip_slab; /* slab the bip came from */
352 unsigned short bip_vcnt; /* # of integrity bio_vecs */
353 unsigned short bip_max_vcnt; /* integrity bio_vec slots */
354 unsigned short bip_flags; /* control flags */
356 struct work_struct bip_work; /* I/O completion */
358 struct bio_vec *bip_vec;
359 struct bio_vec bip_inline_vecs[0];/* embedded bvec array */
362 #if defined(CONFIG_BLK_DEV_INTEGRITY)
364 static inline struct bio_integrity_payload *bio_integrity(struct bio *bio)
366 if (bio->bi_opf & REQ_INTEGRITY)
367 return bio->bi_integrity;
372 static inline bool bio_integrity_flagged(struct bio *bio, enum bip_flags flag)
374 struct bio_integrity_payload *bip = bio_integrity(bio);
377 return bip->bip_flags & flag;
382 static inline sector_t bip_get_seed(struct bio_integrity_payload *bip)
384 return bip->bip_iter.bi_sector;
387 static inline void bip_set_seed(struct bio_integrity_payload *bip,
390 bip->bip_iter.bi_sector = seed;
393 #endif /* CONFIG_BLK_DEV_INTEGRITY */
395 extern void bio_trim(struct bio *bio, int offset, int size);
396 extern struct bio *bio_split(struct bio *bio, int sectors,
397 gfp_t gfp, struct bio_set *bs);
400 * bio_next_split - get next @sectors from a bio, splitting if necessary
402 * @sectors: number of sectors to split from the front of @bio
404 * @bs: bio set to allocate from
406 * Returns a bio representing the next @sectors of @bio - if the bio is smaller
407 * than @sectors, returns the original bio unchanged.
409 static inline struct bio *bio_next_split(struct bio *bio, int sectors,
410 gfp_t gfp, struct bio_set *bs)
412 if (sectors >= bio_sectors(bio))
415 return bio_split(bio, sectors, gfp, bs);
419 BIOSET_NEED_BVECS = BIT(0),
420 BIOSET_NEED_RESCUER = BIT(1),
422 extern int bioset_init(struct bio_set *, unsigned int, unsigned int, int flags);
423 extern void bioset_exit(struct bio_set *);
424 extern int biovec_init_pool(mempool_t *pool, int pool_entries);
425 extern int bioset_init_from_src(struct bio_set *bs, struct bio_set *src);
427 extern struct bio *bio_alloc_bioset(gfp_t, unsigned int, struct bio_set *);
428 extern void bio_put(struct bio *);
430 extern void __bio_clone_fast(struct bio *, struct bio *);
431 extern struct bio *bio_clone_fast(struct bio *, gfp_t, struct bio_set *);
432 extern struct bio *bio_clone_bioset(struct bio *, gfp_t, struct bio_set *bs);
434 extern struct bio_set fs_bio_set;
436 static inline struct bio *bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs)
438 return bio_alloc_bioset(gfp_mask, nr_iovecs, &fs_bio_set);
441 static inline struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned int nr_iovecs)
443 return bio_alloc_bioset(gfp_mask, nr_iovecs, NULL);
446 static inline struct bio *bio_clone_kmalloc(struct bio *bio, gfp_t gfp_mask)
448 return bio_clone_bioset(bio, gfp_mask, NULL);
452 extern blk_qc_t submit_bio(struct bio *);
454 extern void bio_endio(struct bio *);
456 static inline void bio_io_error(struct bio *bio)
458 bio->bi_status = BLK_STS_IOERR;
462 static inline void bio_wouldblock_error(struct bio *bio)
464 bio->bi_status = BLK_STS_AGAIN;
468 struct request_queue;
469 extern int bio_phys_segments(struct request_queue *, struct bio *);
471 extern int submit_bio_wait(struct bio *bio);
472 extern void bio_advance(struct bio *, unsigned);
474 extern void bio_init(struct bio *bio, struct bio_vec *table,
475 unsigned short max_vecs);
476 extern void bio_uninit(struct bio *);
477 extern void bio_reset(struct bio *);
478 void bio_chain(struct bio *, struct bio *);
480 extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
481 extern int bio_add_pc_page(struct request_queue *, struct bio *, struct page *,
482 unsigned int, unsigned int);
483 bool __bio_try_merge_page(struct bio *bio, struct page *page,
484 unsigned int len, unsigned int off);
485 void __bio_add_page(struct bio *bio, struct page *page,
486 unsigned int len, unsigned int off);
487 int bio_iov_iter_get_pages(struct bio *bio, struct iov_iter *iter);
489 extern struct bio *bio_map_user_iov(struct request_queue *,
490 struct iov_iter *, gfp_t);
491 extern void bio_unmap_user(struct bio *);
492 extern struct bio *bio_map_kern(struct request_queue *, void *, unsigned int,
494 extern struct bio *bio_copy_kern(struct request_queue *, void *, unsigned int,
496 extern void bio_set_pages_dirty(struct bio *bio);
497 extern void bio_check_pages_dirty(struct bio *bio);
499 void generic_start_io_acct(struct request_queue *q, int rw,
500 unsigned long sectors, struct hd_struct *part);
501 void generic_end_io_acct(struct request_queue *q, int rw,
502 struct hd_struct *part,
503 unsigned long start_time);
505 #ifndef ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
506 # error "You should define ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE for your platform"
508 #if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE
509 extern void bio_flush_dcache_pages(struct bio *bi);
511 static inline void bio_flush_dcache_pages(struct bio *bi)
516 extern void bio_copy_data_iter(struct bio *dst, struct bvec_iter *dst_iter,
517 struct bio *src, struct bvec_iter *src_iter);
518 extern void bio_copy_data(struct bio *dst, struct bio *src);
519 extern void bio_list_copy_data(struct bio *dst, struct bio *src);
520 extern void bio_free_pages(struct bio *bio);
522 extern struct bio *bio_copy_user_iov(struct request_queue *,
523 struct rq_map_data *,
526 extern int bio_uncopy_user(struct bio *);
527 void zero_fill_bio_iter(struct bio *bio, struct bvec_iter iter);
529 static inline void zero_fill_bio(struct bio *bio)
531 zero_fill_bio_iter(bio, bio->bi_iter);
534 extern struct bio_vec *bvec_alloc(gfp_t, int, unsigned long *, mempool_t *);
535 extern void bvec_free(mempool_t *, struct bio_vec *, unsigned int);
536 extern unsigned int bvec_nr_vecs(unsigned short idx);
537 extern const char *bio_devname(struct bio *bio, char *buffer);
539 #define bio_set_dev(bio, bdev) \
541 if ((bio)->bi_disk != (bdev)->bd_disk) \
542 bio_clear_flag(bio, BIO_THROTTLED);\
543 (bio)->bi_disk = (bdev)->bd_disk; \
544 (bio)->bi_partno = (bdev)->bd_partno; \
547 #define bio_copy_dev(dst, src) \
549 (dst)->bi_disk = (src)->bi_disk; \
550 (dst)->bi_partno = (src)->bi_partno; \
553 #define bio_dev(bio) \
554 disk_devt((bio)->bi_disk)
556 #ifdef CONFIG_BLK_CGROUP
557 int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css);
558 void bio_disassociate_task(struct bio *bio);
559 void bio_clone_blkcg_association(struct bio *dst, struct bio *src);
560 #else /* CONFIG_BLK_CGROUP */
561 static inline int bio_associate_blkcg(struct bio *bio,
562 struct cgroup_subsys_state *blkcg_css) { return 0; }
563 static inline void bio_disassociate_task(struct bio *bio) { }
564 static inline void bio_clone_blkcg_association(struct bio *dst,
566 #endif /* CONFIG_BLK_CGROUP */
568 #ifdef CONFIG_HIGHMEM
570 * remember never ever reenable interrupts between a bvec_kmap_irq and
573 static inline char *bvec_kmap_irq(struct bio_vec *bvec, unsigned long *flags)
578 * might not be a highmem page, but the preempt/irq count
579 * balancing is a lot nicer this way
581 local_irq_save(*flags);
582 addr = (unsigned long) kmap_atomic(bvec->bv_page);
584 BUG_ON(addr & ~PAGE_MASK);
586 return (char *) addr + bvec->bv_offset;
589 static inline void bvec_kunmap_irq(char *buffer, unsigned long *flags)
591 unsigned long ptr = (unsigned long) buffer & PAGE_MASK;
593 kunmap_atomic((void *) ptr);
594 local_irq_restore(*flags);
598 static inline char *bvec_kmap_irq(struct bio_vec *bvec, unsigned long *flags)
600 return page_address(bvec->bv_page) + bvec->bv_offset;
603 static inline void bvec_kunmap_irq(char *buffer, unsigned long *flags)
610 * BIO list management for use by remapping drivers (e.g. DM or MD) and loop.
612 * A bio_list anchors a singly-linked list of bios chained through the bi_next
613 * member of the bio. The bio_list also caches the last list member to allow
614 * fast access to the tail.
621 static inline int bio_list_empty(const struct bio_list *bl)
623 return bl->head == NULL;
626 static inline void bio_list_init(struct bio_list *bl)
628 bl->head = bl->tail = NULL;
631 #define BIO_EMPTY_LIST { NULL, NULL }
633 #define bio_list_for_each(bio, bl) \
634 for (bio = (bl)->head; bio; bio = bio->bi_next)
636 static inline unsigned bio_list_size(const struct bio_list *bl)
641 bio_list_for_each(bio, bl)
647 static inline void bio_list_add(struct bio_list *bl, struct bio *bio)
652 bl->tail->bi_next = bio;
659 static inline void bio_list_add_head(struct bio_list *bl, struct bio *bio)
661 bio->bi_next = bl->head;
669 static inline void bio_list_merge(struct bio_list *bl, struct bio_list *bl2)
675 bl->tail->bi_next = bl2->head;
677 bl->head = bl2->head;
679 bl->tail = bl2->tail;
682 static inline void bio_list_merge_head(struct bio_list *bl,
683 struct bio_list *bl2)
689 bl2->tail->bi_next = bl->head;
691 bl->tail = bl2->tail;
693 bl->head = bl2->head;
696 static inline struct bio *bio_list_peek(struct bio_list *bl)
701 static inline struct bio *bio_list_pop(struct bio_list *bl)
703 struct bio *bio = bl->head;
706 bl->head = bl->head->bi_next;
716 static inline struct bio *bio_list_get(struct bio_list *bl)
718 struct bio *bio = bl->head;
720 bl->head = bl->tail = NULL;
726 * Increment chain count for the bio. Make sure the CHAIN flag update
727 * is visible before the raised count.
729 static inline void bio_inc_remaining(struct bio *bio)
731 bio_set_flag(bio, BIO_CHAIN);
732 smp_mb__before_atomic();
733 atomic_inc(&bio->__bi_remaining);
737 * bio_set is used to allow other portions of the IO system to
738 * allocate their own private memory pools for bio and iovec structures.
739 * These memory pools in turn all allocate from the bio_slab
740 * and the bvec_slabs[].
742 #define BIO_POOL_SIZE 2
745 struct kmem_cache *bio_slab;
746 unsigned int front_pad;
750 #if defined(CONFIG_BLK_DEV_INTEGRITY)
751 mempool_t bio_integrity_pool;
752 mempool_t bvec_integrity_pool;
756 * Deadlock avoidance for stacking block drivers: see comments in
757 * bio_alloc_bioset() for details
759 spinlock_t rescue_lock;
760 struct bio_list rescue_list;
761 struct work_struct rescue_work;
762 struct workqueue_struct *rescue_workqueue;
768 struct kmem_cache *slab;
771 static inline bool bioset_initialized(struct bio_set *bs)
773 return bs->bio_slab != NULL;
777 * a small number of entries is fine, not going to be performance critical.
778 * basically we just need to survive
780 #define BIO_SPLIT_ENTRIES 2
782 #if defined(CONFIG_BLK_DEV_INTEGRITY)
784 #define bip_for_each_vec(bvl, bip, iter) \
785 for_each_bvec(bvl, (bip)->bip_vec, iter, (bip)->bip_iter)
787 #define bio_for_each_integrity_vec(_bvl, _bio, _iter) \
789 bip_for_each_vec(_bvl, _bio->bi_integrity, _iter)
791 extern struct bio_integrity_payload *bio_integrity_alloc(struct bio *, gfp_t, unsigned int);
792 extern int bio_integrity_add_page(struct bio *, struct page *, unsigned int, unsigned int);
793 extern bool bio_integrity_prep(struct bio *);
794 extern void bio_integrity_advance(struct bio *, unsigned int);
795 extern void bio_integrity_trim(struct bio *);
796 extern int bio_integrity_clone(struct bio *, struct bio *, gfp_t);
797 extern int bioset_integrity_create(struct bio_set *, int);
798 extern void bioset_integrity_free(struct bio_set *);
799 extern void bio_integrity_init(void);
801 #else /* CONFIG_BLK_DEV_INTEGRITY */
803 static inline void *bio_integrity(struct bio *bio)
808 static inline int bioset_integrity_create(struct bio_set *bs, int pool_size)
813 static inline void bioset_integrity_free (struct bio_set *bs)
818 static inline bool bio_integrity_prep(struct bio *bio)
823 static inline int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
829 static inline void bio_integrity_advance(struct bio *bio,
830 unsigned int bytes_done)
835 static inline void bio_integrity_trim(struct bio *bio)
840 static inline void bio_integrity_init(void)
845 static inline bool bio_integrity_flagged(struct bio *bio, enum bip_flags flag)
850 static inline void *bio_integrity_alloc(struct bio * bio, gfp_t gfp,
853 return ERR_PTR(-EINVAL);
856 static inline int bio_integrity_add_page(struct bio *bio, struct page *page,
857 unsigned int len, unsigned int offset)
862 #endif /* CONFIG_BLK_DEV_INTEGRITY */
864 #endif /* CONFIG_BLOCK */
865 #endif /* __LINUX_BIO_H */