projects / linux-block.git / blame_incremental
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (non-incremental) | history | HEAD
block: add a bdev_fua helper
[linux-block.git] / include / linux / blkdev.h
... / ...
CommitLineData
1/* SPDX-License-Identifier: GPL-2.0 */
2/*
3 * Portions Copyright (C) 1992 Drew Eckhardt
4 */
5#ifndef _LINUX_BLKDEV_H
6#define _LINUX_BLKDEV_H
7
8#include <linux/types.h>
9#include <linux/blk_types.h>
10#include <linux/device.h>
11#include <linux/list.h>
12#include <linux/llist.h>
13#include <linux/minmax.h>
14#include <linux/timer.h>
15#include <linux/workqueue.h>
16#include <linux/wait.h>
17#include <linux/bio.h>
18#include <linux/gfp.h>
19#include <linux/kdev_t.h>
20#include <linux/rcupdate.h>
21#include <linux/percpu-refcount.h>
22#include <linux/blkzoned.h>
23#include <linux/sched.h>
24#include <linux/sbitmap.h>
25#include <linux/srcu.h>
26#include <linux/uuid.h>
27#include <linux/xarray.h>
28
29struct module;
30struct request_queue;
31struct elevator_queue;
32struct blk_trace;
33struct request;
34struct sg_io_hdr;
35struct blkcg_gq;
36struct blk_flush_queue;
37struct kiocb;
38struct pr_ops;
39struct rq_qos;
40struct blk_queue_stats;
41struct blk_stat_callback;
42struct blk_crypto_profile;
43
44extern const struct device_type disk_type;
45extern struct device_type part_type;
46extern struct class block_class;
47
48/* Must be consistent with blk_mq_poll_stats_bkt() */
49#define BLK_MQ_POLL_STATS_BKTS 16
50
51/* Doing classic polling */
52#define BLK_MQ_POLL_CLASSIC -1
53
54/*
55 * Maximum number of blkcg policies allowed to be registered concurrently.
56 * Defined here to simplify include dependency.
57 */
58#define BLKCG_MAX_POLS 6
59
60#define DISK_MAX_PARTS 256
61#define DISK_NAME_LEN 32
62
63#define PARTITION_META_INFO_VOLNAMELTH 64
64/*
65 * Enough for the string representation of any kind of UUID plus NULL.
66 * EFI UUID is 36 characters. MSDOS UUID is 11 characters.
67 */
68#define PARTITION_META_INFO_UUIDLTH (UUID_STRING_LEN + 1)
69
70struct partition_meta_info {
71 char uuid[PARTITION_META_INFO_UUIDLTH];
72 u8 volname[PARTITION_META_INFO_VOLNAMELTH];
73};
74
75/**
76 * DOC: genhd capability flags
77 *
78 * ``GENHD_FL_REMOVABLE``: indicates that the block device gives access to
79 * removable media. When set, the device remains present even when media is not
80 * inserted. Shall not be set for devices which are removed entirely when the
81 * media is removed.
82 *
83 * ``GENHD_FL_HIDDEN``: the block device is hidden; it doesn't produce events,
84 * doesn't appear in sysfs, and can't be opened from userspace or using
85 * blkdev_get*. Used for the underlying components of multipath devices.
86 *
87 * ``GENHD_FL_NO_PART``: partition support is disabled. The kernel will not
88 * scan for partitions from add_disk, and users can't add partitions manually.
89 *
90 */
91enum {
92 GENHD_FL_REMOVABLE = 1 << 0,
93 GENHD_FL_HIDDEN = 1 << 1,
94 GENHD_FL_NO_PART = 1 << 2,
95};
96
97enum {
98 DISK_EVENT_MEDIA_CHANGE = 1 << 0, /* media changed */
99 DISK_EVENT_EJECT_REQUEST = 1 << 1, /* eject requested */
100};
101
102enum {
103 /* Poll even if events_poll_msecs is unset */
104 DISK_EVENT_FLAG_POLL = 1 << 0,
105 /* Forward events to udev */
106 DISK_EVENT_FLAG_UEVENT = 1 << 1,
107 /* Block event polling when open for exclusive write */
108 DISK_EVENT_FLAG_BLOCK_ON_EXCL_WRITE = 1 << 2,
109};
110
111struct disk_events;
112struct badblocks;
113
114struct blk_integrity {
115 const struct blk_integrity_profile *profile;
116 unsigned char flags;
117 unsigned char tuple_size;
118 unsigned char interval_exp;
119 unsigned char tag_size;
120};
121
122struct gendisk {
123 /*
124 * major/first_minor/minors should not be set by any new driver, the
125 * block core will take care of allocating them automatically.
126 */
127 int major;
128 int first_minor;
129 int minors;
130
131 char disk_name[DISK_NAME_LEN]; /* name of major driver */
132
133 unsigned short events; /* supported events */
134 unsigned short event_flags; /* flags related to event processing */
135
136 struct xarray part_tbl;
137 struct block_device *part0;
138
139 const struct block_device_operations *fops;
140 struct request_queue *queue;
141 void *private_data;
142
143 int flags;
144 unsigned long state;
145#define GD_NEED_PART_SCAN 0
146#define GD_READ_ONLY 1
147#define GD_DEAD 2
148#define GD_NATIVE_CAPACITY 3
149#define GD_ADDED 4
150
151 struct mutex open_mutex; /* open/close mutex */
152 unsigned open_partitions; /* number of open partitions */
153
154 struct backing_dev_info *bdi;
155 struct kobject *slave_dir;
156#ifdef CONFIG_BLOCK_HOLDER_DEPRECATED
157 struct list_head slave_bdevs;
158#endif
159 struct timer_rand_state *random;
160 atomic_t sync_io; /* RAID */
161 struct disk_events *ev;
162#ifdef CONFIG_BLK_DEV_INTEGRITY
163 struct kobject integrity_kobj;
164#endif /* CONFIG_BLK_DEV_INTEGRITY */
165#if IS_ENABLED(CONFIG_CDROM)
166 struct cdrom_device_info *cdi;
167#endif
168 int node_id;
169 struct badblocks *bb;
170 struct lockdep_map lockdep_map;
171 u64 diskseq;
172};
173
174static inline bool disk_live(struct gendisk *disk)
175{
176 return !inode_unhashed(disk->part0->bd_inode);
177}
178
179/*
180 * The gendisk is refcounted by the part0 block_device, and the bd_device
181 * therein is also used for device model presentation in sysfs.
182 */
183#define dev_to_disk(device) \
184 (dev_to_bdev(device)->bd_disk)
185#define disk_to_dev(disk) \
186 (&((disk)->part0->bd_device))
187
188#if IS_REACHABLE(CONFIG_CDROM)
189#define disk_to_cdi(disk) ((disk)->cdi)
190#else
191#define disk_to_cdi(disk) NULL
192#endif
193
194static inline dev_t disk_devt(struct gendisk *disk)
195{
196 return MKDEV(disk->major, disk->first_minor);
197}
198
199static inline int blk_validate_block_size(unsigned long bsize)
200{
201 if (bsize < 512 || bsize > PAGE_SIZE || !is_power_of_2(bsize))
202 return -EINVAL;
203
204 return 0;
205}
206
207static inline bool blk_op_is_passthrough(unsigned int op)
208{
209 op &= REQ_OP_MASK;
210 return op == REQ_OP_DRV_IN || op == REQ_OP_DRV_OUT;
211}
212
213/*
214 * Zoned block device models (zoned limit).
215 *
216 * Note: This needs to be ordered from the least to the most severe
217 * restrictions for the inheritance in blk_stack_limits() to work.
218 */
219enum blk_zoned_model {
220 BLK_ZONED_NONE = 0, /* Regular block device */
221 BLK_ZONED_HA, /* Host-aware zoned block device */
222 BLK_ZONED_HM, /* Host-managed zoned block device */
223};
224
225/*
226 * BLK_BOUNCE_NONE: never bounce (default)
227 * BLK_BOUNCE_HIGH: bounce all highmem pages
228 */
229enum blk_bounce {
230 BLK_BOUNCE_NONE,
231 BLK_BOUNCE_HIGH,
232};
233
234struct queue_limits {
235 enum blk_bounce bounce;
236 unsigned long seg_boundary_mask;
237 unsigned long virt_boundary_mask;
238
239 unsigned int max_hw_sectors;
240 unsigned int max_dev_sectors;
241 unsigned int chunk_sectors;
242 unsigned int max_sectors;
243 unsigned int max_segment_size;
244 unsigned int physical_block_size;
245 unsigned int logical_block_size;
246 unsigned int alignment_offset;
247 unsigned int io_min;
248 unsigned int io_opt;
249 unsigned int max_discard_sectors;
250 unsigned int max_hw_discard_sectors;
251 unsigned int max_write_zeroes_sectors;
252 unsigned int max_zone_append_sectors;
253 unsigned int discard_granularity;
254 unsigned int discard_alignment;
255 unsigned int zone_write_granularity;
256
257 unsigned short max_segments;
258 unsigned short max_integrity_segments;
259 unsigned short max_discard_segments;
260
261 unsigned char misaligned;
262 unsigned char discard_misaligned;
263 unsigned char raid_partial_stripes_expensive;
264 enum blk_zoned_model zoned;
265};
266
267typedef int (*report_zones_cb)(struct blk_zone *zone, unsigned int idx,
268 void *data);
269
270void blk_queue_set_zoned(struct gendisk *disk, enum blk_zoned_model model);
271
272#ifdef CONFIG_BLK_DEV_ZONED
273
274#define BLK_ALL_ZONES ((unsigned int)-1)
275int blkdev_report_zones(struct block_device *bdev, sector_t sector,
276 unsigned int nr_zones, report_zones_cb cb, void *data);
277unsigned int blkdev_nr_zones(struct gendisk *disk);
278extern int blkdev_zone_mgmt(struct block_device *bdev, enum req_opf op,
279 sector_t sectors, sector_t nr_sectors,
280 gfp_t gfp_mask);
281int blk_revalidate_disk_zones(struct gendisk *disk,
282 void (*update_driver_data)(struct gendisk *disk));
283
284extern int blkdev_report_zones_ioctl(struct block_device *bdev, fmode_t mode,
285 unsigned int cmd, unsigned long arg);
286extern int blkdev_zone_mgmt_ioctl(struct block_device *bdev, fmode_t mode,
287 unsigned int cmd, unsigned long arg);
288
289#else /* CONFIG_BLK_DEV_ZONED */
290
291static inline unsigned int blkdev_nr_zones(struct gendisk *disk)
292{
293 return 0;
294}
295
296static inline int blkdev_report_zones_ioctl(struct block_device *bdev,
297 fmode_t mode, unsigned int cmd,
298 unsigned long arg)
299{
300 return -ENOTTY;
301}
302
303static inline int blkdev_zone_mgmt_ioctl(struct block_device *bdev,
304 fmode_t mode, unsigned int cmd,
305 unsigned long arg)
306{
307 return -ENOTTY;
308}
309
310#endif /* CONFIG_BLK_DEV_ZONED */
311
312/*
313 * Independent access ranges: struct blk_independent_access_range describes
314 * a range of contiguous sectors that can be accessed using device command
315 * execution resources that are independent from the resources used for
316 * other access ranges. This is typically found with single-LUN multi-actuator
317 * HDDs where each access range is served by a different set of heads.
318 * The set of independent ranges supported by the device is defined using
319 * struct blk_independent_access_ranges. The independent ranges must not overlap
320 * and must include all sectors within the disk capacity (no sector holes
321 * allowed).
322 * For a device with multiple ranges, requests targeting sectors in different
323 * ranges can be executed in parallel. A request can straddle an access range
324 * boundary.
325 */
326struct blk_independent_access_range {
327 struct kobject kobj;
328 struct request_queue *queue;
329 sector_t sector;
330 sector_t nr_sectors;
331};
332
333struct blk_independent_access_ranges {
334 struct kobject kobj;
335 bool sysfs_registered;
336 unsigned int nr_ia_ranges;
337 struct blk_independent_access_range ia_range[];
338};
339
340struct request_queue {
341 struct request *last_merge;
342 struct elevator_queue *elevator;
343
344 struct percpu_ref q_usage_counter;
345
346 struct blk_queue_stats *stats;
347 struct rq_qos *rq_qos;
348
349 const struct blk_mq_ops *mq_ops;
350
351 /* sw queues */
352 struct blk_mq_ctx __percpu *queue_ctx;
353
354 unsigned int queue_depth;
355
356 /* hw dispatch queues */
357 struct xarray hctx_table;
358 unsigned int nr_hw_queues;
359
360 /*
361 * The queue owner gets to use this for whatever they like.
362 * ll_rw_blk doesn't touch it.
363 */
364 void *queuedata;
365
366 /*
367 * various queue flags, see QUEUE_* below
368 */
369 unsigned long queue_flags;
370 /*
371 * Number of contexts that have called blk_set_pm_only(). If this
372 * counter is above zero then only RQF_PM requests are processed.
373 */
374 atomic_t pm_only;
375
376 /*
377 * ida allocated id for this queue. Used to index queues from
378 * ioctx.
379 */
380 int id;
381
382 spinlock_t queue_lock;
383
384 struct gendisk *disk;
385
386 /*
387 * queue kobject
388 */
389 struct kobject kobj;
390
391 /*
392 * mq queue kobject
393 */
394 struct kobject *mq_kobj;
395
396#ifdef CONFIG_BLK_DEV_INTEGRITY
397 struct blk_integrity integrity;
398#endif /* CONFIG_BLK_DEV_INTEGRITY */
399
400#ifdef CONFIG_PM
401 struct device *dev;
402 enum rpm_status rpm_status;
403#endif
404
405 /*
406 * queue settings
407 */
408 unsigned long nr_requests; /* Max # of requests */
409
410 unsigned int dma_pad_mask;
411 unsigned int dma_alignment;
412
413#ifdef CONFIG_BLK_INLINE_ENCRYPTION
414 struct blk_crypto_profile *crypto_profile;
415 struct kobject *crypto_kobject;
416#endif
417
418 unsigned int rq_timeout;
419 int poll_nsec;
420
421 struct blk_stat_callback *poll_cb;
422 struct blk_rq_stat *poll_stat;
423
424 struct timer_list timeout;
425 struct work_struct timeout_work;
426
427 atomic_t nr_active_requests_shared_tags;
428
429 struct blk_mq_tags *sched_shared_tags;
430
431 struct list_head icq_list;
432#ifdef CONFIG_BLK_CGROUP
433 DECLARE_BITMAP (blkcg_pols, BLKCG_MAX_POLS);
434 struct blkcg_gq *root_blkg;
435 struct list_head blkg_list;
436#endif
437
438 struct queue_limits limits;
439
440 unsigned int required_elevator_features;
441
442#ifdef CONFIG_BLK_DEV_ZONED
443 /*
444 * Zoned block device information for request dispatch control.
445 * nr_zones is the total number of zones of the device. This is always
446 * 0 for regular block devices. conv_zones_bitmap is a bitmap of nr_zones
447 * bits which indicates if a zone is conventional (bit set) or
448 * sequential (bit clear). seq_zones_wlock is a bitmap of nr_zones
449 * bits which indicates if a zone is write locked, that is, if a write
450 * request targeting the zone was dispatched. All three fields are
451 * initialized by the low level device driver (e.g. scsi/sd.c).
452 * Stacking drivers (device mappers) may or may not initialize
453 * these fields.
454 *
455 * Reads of this information must be protected with blk_queue_enter() /
456 * blk_queue_exit(). Modifying this information is only allowed while
457 * no requests are being processed. See also blk_mq_freeze_queue() and
458 * blk_mq_unfreeze_queue().
459 */
460 unsigned int nr_zones;
461 unsigned long *conv_zones_bitmap;
462 unsigned long *seq_zones_wlock;
463 unsigned int max_open_zones;
464 unsigned int max_active_zones;
465#endif /* CONFIG_BLK_DEV_ZONED */
466
467 int node;
468 struct mutex debugfs_mutex;
469#ifdef CONFIG_BLK_DEV_IO_TRACE
470 struct blk_trace __rcu *blk_trace;
471#endif
472 /*
473 * for flush operations
474 */
475 struct blk_flush_queue *fq;
476
477 struct list_head requeue_list;
478 spinlock_t requeue_lock;
479 struct delayed_work requeue_work;
480
481 struct mutex sysfs_lock;
482 struct mutex sysfs_dir_lock;
483
484 /*
485 * for reusing dead hctx instance in case of updating
486 * nr_hw_queues
487 */
488 struct list_head unused_hctx_list;
489 spinlock_t unused_hctx_lock;
490
491 int mq_freeze_depth;
492
493#ifdef CONFIG_BLK_DEV_THROTTLING
494 /* Throttle data */
495 struct throtl_data *td;
496#endif
497 struct rcu_head rcu_head;
498 wait_queue_head_t mq_freeze_wq;
499 /*
500 * Protect concurrent access to q_usage_counter by
501 * percpu_ref_kill() and percpu_ref_reinit().
502 */
503 struct mutex mq_freeze_lock;
504
505 int quiesce_depth;
506
507 struct blk_mq_tag_set *tag_set;
508 struct list_head tag_set_list;
509 struct bio_set bio_split;
510
511 struct dentry *debugfs_dir;
512
513#ifdef CONFIG_BLK_DEBUG_FS
514 struct dentry *sched_debugfs_dir;
515 struct dentry *rqos_debugfs_dir;
516#endif
517
518 bool mq_sysfs_init_done;
519
520 /*
521 * Independent sector access ranges. This is always NULL for
522 * devices that do not have multiple independent access ranges.
523 */
524 struct blk_independent_access_ranges *ia_ranges;
525
526 /**
527 * @srcu: Sleepable RCU. Use as lock when type of the request queue
528 * is blocking (BLK_MQ_F_BLOCKING). Must be the last member
529 */
530 struct srcu_struct srcu[];
531};
532
533/* Keep blk_queue_flag_name[] in sync with the definitions below */
534#define QUEUE_FLAG_STOPPED 0 /* queue is stopped */
535#define QUEUE_FLAG_DYING 1 /* queue being torn down */
536#define QUEUE_FLAG_HAS_SRCU 2 /* SRCU is allocated */
537#define QUEUE_FLAG_NOMERGES 3 /* disable merge attempts */
538#define QUEUE_FLAG_SAME_COMP 4 /* complete on same CPU-group */
539#define QUEUE_FLAG_FAIL_IO 5 /* fake timeout */
540#define QUEUE_FLAG_NONROT 6 /* non-rotational device (SSD) */
541#define QUEUE_FLAG_VIRT QUEUE_FLAG_NONROT /* paravirt device */
542#define QUEUE_FLAG_IO_STAT 7 /* do disk/partitions IO accounting */
543#define QUEUE_FLAG_DISCARD 8 /* supports DISCARD */
544#define QUEUE_FLAG_NOXMERGES 9 /* No extended merges */
545#define QUEUE_FLAG_ADD_RANDOM 10 /* Contributes to random pool */
546#define QUEUE_FLAG_SECERASE 11 /* supports secure erase */
547#define QUEUE_FLAG_SAME_FORCE 12 /* force complete on same CPU */
548#define QUEUE_FLAG_DEAD 13 /* queue tear-down finished */
549#define QUEUE_FLAG_INIT_DONE 14 /* queue is initialized */
550#define QUEUE_FLAG_STABLE_WRITES 15 /* don't modify blks until WB is done */
551#define QUEUE_FLAG_POLL 16 /* IO polling enabled if set */
552#define QUEUE_FLAG_WC 17 /* Write back caching */
553#define QUEUE_FLAG_FUA 18 /* device supports FUA writes */
554#define QUEUE_FLAG_DAX 19 /* device supports DAX */
555#define QUEUE_FLAG_STATS 20 /* track IO start and completion times */
556#define QUEUE_FLAG_REGISTERED 22 /* queue has been registered to a disk */
557#define QUEUE_FLAG_QUIESCED 24 /* queue has been quiesced */
558#define QUEUE_FLAG_PCI_P2PDMA 25 /* device supports PCI p2p requests */
559#define QUEUE_FLAG_ZONE_RESETALL 26 /* supports Zone Reset All */
560#define QUEUE_FLAG_RQ_ALLOC_TIME 27 /* record rq->alloc_time_ns */
561#define QUEUE_FLAG_HCTX_ACTIVE 28 /* at least one blk-mq hctx is active */
562#define QUEUE_FLAG_NOWAIT 29 /* device supports NOWAIT */
563
564#define QUEUE_FLAG_MQ_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \
565 (1 << QUEUE_FLAG_SAME_COMP) | \
566 (1 << QUEUE_FLAG_NOWAIT))
567
568void blk_queue_flag_set(unsigned int flag, struct request_queue *q);
569void blk_queue_flag_clear(unsigned int flag, struct request_queue *q);
570bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q);
571
572#define blk_queue_stopped(q) test_bit(QUEUE_FLAG_STOPPED, &(q)->queue_flags)
573#define blk_queue_dying(q) test_bit(QUEUE_FLAG_DYING, &(q)->queue_flags)
574#define blk_queue_has_srcu(q) test_bit(QUEUE_FLAG_HAS_SRCU, &(q)->queue_flags)
575#define blk_queue_dead(q) test_bit(QUEUE_FLAG_DEAD, &(q)->queue_flags)
576#define blk_queue_init_done(q) test_bit(QUEUE_FLAG_INIT_DONE, &(q)->queue_flags)
577#define blk_queue_nomerges(q) test_bit(QUEUE_FLAG_NOMERGES, &(q)->queue_flags)
578#define blk_queue_noxmerges(q) \
579 test_bit(QUEUE_FLAG_NOXMERGES, &(q)->queue_flags)
580#define blk_queue_nonrot(q) test_bit(QUEUE_FLAG_NONROT, &(q)->queue_flags)
581#define blk_queue_stable_writes(q) \
582 test_bit(QUEUE_FLAG_STABLE_WRITES, &(q)->queue_flags)
583#define blk_queue_io_stat(q) test_bit(QUEUE_FLAG_IO_STAT, &(q)->queue_flags)
584#define blk_queue_add_random(q) test_bit(QUEUE_FLAG_ADD_RANDOM, &(q)->queue_flags)
585#define blk_queue_discard(q) test_bit(QUEUE_FLAG_DISCARD, &(q)->queue_flags)
586#define blk_queue_zone_resetall(q) \
587 test_bit(QUEUE_FLAG_ZONE_RESETALL, &(q)->queue_flags)
588#define blk_queue_secure_erase(q) \
589 (test_bit(QUEUE_FLAG_SECERASE, &(q)->queue_flags))
590#define blk_queue_dax(q) test_bit(QUEUE_FLAG_DAX, &(q)->queue_flags)
591#define blk_queue_pci_p2pdma(q) \
592 test_bit(QUEUE_FLAG_PCI_P2PDMA, &(q)->queue_flags)
593#ifdef CONFIG_BLK_RQ_ALLOC_TIME
594#define blk_queue_rq_alloc_time(q) \
595 test_bit(QUEUE_FLAG_RQ_ALLOC_TIME, &(q)->queue_flags)
596#else
597#define blk_queue_rq_alloc_time(q) false
598#endif
599
600#define blk_noretry_request(rq) \
601 ((rq)->cmd_flags & (REQ_FAILFAST_DEV|REQ_FAILFAST_TRANSPORT| \
602 REQ_FAILFAST_DRIVER))
603#define blk_queue_quiesced(q) test_bit(QUEUE_FLAG_QUIESCED, &(q)->queue_flags)
604#define blk_queue_pm_only(q) atomic_read(&(q)->pm_only)
605#define blk_queue_registered(q) test_bit(QUEUE_FLAG_REGISTERED, &(q)->queue_flags)
606#define blk_queue_nowait(q) test_bit(QUEUE_FLAG_NOWAIT, &(q)->queue_flags)
607
608extern void blk_set_pm_only(struct request_queue *q);
609extern void blk_clear_pm_only(struct request_queue *q);
610
611#define list_entry_rq(ptr) list_entry((ptr), struct request, queuelist)
612
613#define dma_map_bvec(dev, bv, dir, attrs) \
614 dma_map_page_attrs(dev, (bv)->bv_page, (bv)->bv_offset, (bv)->bv_len, \
615 (dir), (attrs))
616
617static inline bool queue_is_mq(struct request_queue *q)
618{
619 return q->mq_ops;
620}
621
622#ifdef CONFIG_PM
623static inline enum rpm_status queue_rpm_status(struct request_queue *q)
624{
625 return q->rpm_status;
626}
627#else
628static inline enum rpm_status queue_rpm_status(struct request_queue *q)
629{
630 return RPM_ACTIVE;
631}
632#endif
633
634static inline enum blk_zoned_model
635blk_queue_zoned_model(struct request_queue *q)
636{
637 if (IS_ENABLED(CONFIG_BLK_DEV_ZONED))
638 return q->limits.zoned;
639 return BLK_ZONED_NONE;
640}
641
642static inline bool blk_queue_is_zoned(struct request_queue *q)
643{
644 switch (blk_queue_zoned_model(q)) {
645 case BLK_ZONED_HA:
646 case BLK_ZONED_HM:
647 return true;
648 default:
649 return false;
650 }
651}
652
653static inline sector_t blk_queue_zone_sectors(struct request_queue *q)
654{
655 return blk_queue_is_zoned(q) ? q->limits.chunk_sectors : 0;
656}
657
658#ifdef CONFIG_BLK_DEV_ZONED
659static inline unsigned int blk_queue_nr_zones(struct request_queue *q)
660{
661 return blk_queue_is_zoned(q) ? q->nr_zones : 0;
662}
663
664static inline unsigned int blk_queue_zone_no(struct request_queue *q,
665 sector_t sector)
666{
667 if (!blk_queue_is_zoned(q))
668 return 0;
669 return sector >> ilog2(q->limits.chunk_sectors);
670}
671
672static inline bool blk_queue_zone_is_seq(struct request_queue *q,
673 sector_t sector)
674{
675 if (!blk_queue_is_zoned(q))
676 return false;
677 if (!q->conv_zones_bitmap)
678 return true;
679 return !test_bit(blk_queue_zone_no(q, sector), q->conv_zones_bitmap);
680}
681
682static inline void blk_queue_max_open_zones(struct request_queue *q,
683 unsigned int max_open_zones)
684{
685 q->max_open_zones = max_open_zones;
686}
687
688static inline unsigned int queue_max_open_zones(const struct request_queue *q)
689{
690 return q->max_open_zones;
691}
692
693static inline void blk_queue_max_active_zones(struct request_queue *q,
694 unsigned int max_active_zones)
695{
696 q->max_active_zones = max_active_zones;
697}
698
699static inline unsigned int queue_max_active_zones(const struct request_queue *q)
700{
701 return q->max_active_zones;
702}
703#else /* CONFIG_BLK_DEV_ZONED */
704static inline unsigned int blk_queue_nr_zones(struct request_queue *q)
705{
706 return 0;
707}
708static inline bool blk_queue_zone_is_seq(struct request_queue *q,
709 sector_t sector)
710{
711 return false;
712}
713static inline unsigned int blk_queue_zone_no(struct request_queue *q,
714 sector_t sector)
715{
716 return 0;
717}
718static inline unsigned int queue_max_open_zones(const struct request_queue *q)
719{
720 return 0;
721}
722static inline unsigned int queue_max_active_zones(const struct request_queue *q)
723{
724 return 0;
725}
726#endif /* CONFIG_BLK_DEV_ZONED */
727
728static inline unsigned int blk_queue_depth(struct request_queue *q)
729{
730 if (q->queue_depth)
731 return q->queue_depth;
732
733 return q->nr_requests;
734}
735
736/*
737 * default timeout for SG_IO if none specified
738 */
739#define BLK_DEFAULT_SG_TIMEOUT (60 * HZ)
740#define BLK_MIN_SG_TIMEOUT (7 * HZ)
741
742/* This should not be used directly - use rq_for_each_segment */
743#define for_each_bio(_bio) \
744 for (; _bio; _bio = _bio->bi_next)
745
746int __must_check device_add_disk(struct device *parent, struct gendisk *disk,
747 const struct attribute_group **groups);
748static inline int __must_check add_disk(struct gendisk *disk)
749{
750 return device_add_disk(NULL, disk, NULL);
751}
752void del_gendisk(struct gendisk *gp);
753void invalidate_disk(struct gendisk *disk);
754void set_disk_ro(struct gendisk *disk, bool read_only);
755void disk_uevent(struct gendisk *disk, enum kobject_action action);
756
757static inline int get_disk_ro(struct gendisk *disk)
758{
759 return disk->part0->bd_read_only ||
760 test_bit(GD_READ_ONLY, &disk->state);
761}
762
763static inline int bdev_read_only(struct block_device *bdev)
764{
765 return bdev->bd_read_only || get_disk_ro(bdev->bd_disk);
766}
767
768bool set_capacity_and_notify(struct gendisk *disk, sector_t size);
769bool disk_force_media_change(struct gendisk *disk, unsigned int events);
770
771void add_disk_randomness(struct gendisk *disk) __latent_entropy;
772void rand_initialize_disk(struct gendisk *disk);
773
774static inline sector_t get_start_sect(struct block_device *bdev)
775{
776 return bdev->bd_start_sect;
777}
778
779static inline sector_t bdev_nr_sectors(struct block_device *bdev)
780{
781 return bdev->bd_nr_sectors;
782}
783
784static inline loff_t bdev_nr_bytes(struct block_device *bdev)
785{
786 return (loff_t)bdev_nr_sectors(bdev) << SECTOR_SHIFT;
787}
788
789static inline sector_t get_capacity(struct gendisk *disk)
790{
791 return bdev_nr_sectors(disk->part0);
792}
793
794static inline u64 sb_bdev_nr_blocks(struct super_block *sb)
795{
796 return bdev_nr_sectors(sb->s_bdev) >>
797 (sb->s_blocksize_bits - SECTOR_SHIFT);
798}
799
800int bdev_disk_changed(struct gendisk *disk, bool invalidate);
801
802struct gendisk *__alloc_disk_node(struct request_queue *q, int node_id,
803 struct lock_class_key *lkclass);
804void put_disk(struct gendisk *disk);
805struct gendisk *__blk_alloc_disk(int node, struct lock_class_key *lkclass);
806
807/**
808 * blk_alloc_disk - allocate a gendisk structure
809 * @node_id: numa node to allocate on
810 *
811 * Allocate and pre-initialize a gendisk structure for use with BIO based
812 * drivers.
813 *
814 * Context: can sleep
815 */
816#define blk_alloc_disk(node_id) \
817({ \
818 static struct lock_class_key __key; \
819 \
820 __blk_alloc_disk(node_id, &__key); \
821})
822void blk_cleanup_disk(struct gendisk *disk);
823
824int __register_blkdev(unsigned int major, const char *name,
825 void (*probe)(dev_t devt));
826#define register_blkdev(major, name) \
827 __register_blkdev(major, name, NULL)
828void unregister_blkdev(unsigned int major, const char *name);
829
830bool bdev_check_media_change(struct block_device *bdev);
831int __invalidate_device(struct block_device *bdev, bool kill_dirty);
832void set_capacity(struct gendisk *disk, sector_t size);
833
834#ifdef CONFIG_BLOCK_HOLDER_DEPRECATED
835int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk);
836void bd_unlink_disk_holder(struct block_device *bdev, struct gendisk *disk);
837int bd_register_pending_holders(struct gendisk *disk);
838#else
839static inline int bd_link_disk_holder(struct block_device *bdev,
840 struct gendisk *disk)
841{
842 return 0;
843}
844static inline void bd_unlink_disk_holder(struct block_device *bdev,
845 struct gendisk *disk)
846{
847}
848static inline int bd_register_pending_holders(struct gendisk *disk)
849{
850 return 0;
851}
852#endif /* CONFIG_BLOCK_HOLDER_DEPRECATED */
853
854dev_t part_devt(struct gendisk *disk, u8 partno);
855void inc_diskseq(struct gendisk *disk);
856dev_t blk_lookup_devt(const char *name, int partno);
857void blk_request_module(dev_t devt);
858
859extern int blk_register_queue(struct gendisk *disk);
860extern void blk_unregister_queue(struct gendisk *disk);
861void submit_bio_noacct(struct bio *bio);
862
863extern int blk_lld_busy(struct request_queue *q);
864extern void blk_queue_split(struct bio **);
865extern int blk_queue_enter(struct request_queue *q, blk_mq_req_flags_t flags);
866extern void blk_queue_exit(struct request_queue *q);
867extern void blk_sync_queue(struct request_queue *q);
868
869/* Helper to convert REQ_OP_XXX to its string format XXX */
870extern const char *blk_op_str(unsigned int op);
871
872int blk_status_to_errno(blk_status_t status);
873blk_status_t errno_to_blk_status(int errno);
874
875/* only poll the hardware once, don't continue until a completion was found */
876#define BLK_POLL_ONESHOT (1 << 0)
877/* do not sleep to wait for the expected completion time */
878#define BLK_POLL_NOSLEEP (1 << 1)
879int bio_poll(struct bio *bio, struct io_comp_batch *iob, unsigned int flags);
880int iocb_bio_iopoll(struct kiocb *kiocb, struct io_comp_batch *iob,
881 unsigned int flags);
882
883static inline struct request_queue *bdev_get_queue(struct block_device *bdev)
884{
885 return bdev->bd_queue; /* this is never NULL */
886}
887
888#ifdef CONFIG_BLK_DEV_ZONED
889
890/* Helper to convert BLK_ZONE_ZONE_XXX to its string format XXX */
891const char *blk_zone_cond_str(enum blk_zone_cond zone_cond);
892
893static inline unsigned int bio_zone_no(struct bio *bio)
894{
895 return blk_queue_zone_no(bdev_get_queue(bio->bi_bdev),
896 bio->bi_iter.bi_sector);
897}
898
899static inline unsigned int bio_zone_is_seq(struct bio *bio)
900{
901 return blk_queue_zone_is_seq(bdev_get_queue(bio->bi_bdev),
902 bio->bi_iter.bi_sector);
903}
904#endif /* CONFIG_BLK_DEV_ZONED */
905
906static inline unsigned int blk_queue_get_max_sectors(struct request_queue *q,
907 int op)
908{
909 if (unlikely(op == REQ_OP_DISCARD || op == REQ_OP_SECURE_ERASE))
910 return min(q->limits.max_discard_sectors,
911 UINT_MAX >> SECTOR_SHIFT);
912
913 if (unlikely(op == REQ_OP_WRITE_ZEROES))
914 return q->limits.max_write_zeroes_sectors;
915
916 return q->limits.max_sectors;
917}
918
919/*
920 * Return maximum size of a request at given offset. Only valid for
921 * file system requests.
922 */
923static inline unsigned int blk_max_size_offset(struct request_queue *q,
924 sector_t offset,
925 unsigned int chunk_sectors)
926{
927 if (!chunk_sectors) {
928 if (q->limits.chunk_sectors)
929 chunk_sectors = q->limits.chunk_sectors;
930 else
931 return q->limits.max_sectors;
932 }
933
934 if (likely(is_power_of_2(chunk_sectors)))
935 chunk_sectors -= offset & (chunk_sectors - 1);
936 else
937 chunk_sectors -= sector_div(offset, chunk_sectors);
938
939 return min(q->limits.max_sectors, chunk_sectors);
940}
941
942/*
943 * Access functions for manipulating queue properties
944 */
945extern void blk_cleanup_queue(struct request_queue *);
946void blk_queue_bounce_limit(struct request_queue *q, enum blk_bounce limit);
947extern void blk_queue_max_hw_sectors(struct request_queue *, unsigned int);
948extern void blk_queue_chunk_sectors(struct request_queue *, unsigned int);
949extern void blk_queue_max_segments(struct request_queue *, unsigned short);
950extern void blk_queue_max_discard_segments(struct request_queue *,
951 unsigned short);
952extern void blk_queue_max_segment_size(struct request_queue *, unsigned int);
953extern void blk_queue_max_discard_sectors(struct request_queue *q,
954 unsigned int max_discard_sectors);
955extern void blk_queue_max_write_zeroes_sectors(struct request_queue *q,
956 unsigned int max_write_same_sectors);
957extern void blk_queue_logical_block_size(struct request_queue *, unsigned int);
958extern void blk_queue_max_zone_append_sectors(struct request_queue *q,
959 unsigned int max_zone_append_sectors);
960extern void blk_queue_physical_block_size(struct request_queue *, unsigned int);
961void blk_queue_zone_write_granularity(struct request_queue *q,
962 unsigned int size);
963extern void blk_queue_alignment_offset(struct request_queue *q,
964 unsigned int alignment);
965void disk_update_readahead(struct gendisk *disk);
966extern void blk_limits_io_min(struct queue_limits *limits, unsigned int min);
967extern void blk_queue_io_min(struct request_queue *q, unsigned int min);
968extern void blk_limits_io_opt(struct queue_limits *limits, unsigned int opt);
969extern void blk_queue_io_opt(struct request_queue *q, unsigned int opt);
970extern void blk_set_queue_depth(struct request_queue *q, unsigned int depth);
971extern void blk_set_default_limits(struct queue_limits *lim);
972extern void blk_set_stacking_limits(struct queue_limits *lim);
973extern int blk_stack_limits(struct queue_limits *t, struct queue_limits *b,
974 sector_t offset);
975extern void disk_stack_limits(struct gendisk *disk, struct block_device *bdev,
976 sector_t offset);
977extern void blk_queue_update_dma_pad(struct request_queue *, unsigned int);
978extern void blk_queue_segment_boundary(struct request_queue *, unsigned long);
979extern void blk_queue_virt_boundary(struct request_queue *, unsigned long);
980extern void blk_queue_dma_alignment(struct request_queue *, int);
981extern void blk_queue_update_dma_alignment(struct request_queue *, int);
982extern void blk_queue_rq_timeout(struct request_queue *, unsigned int);
983extern void blk_queue_write_cache(struct request_queue *q, bool enabled, bool fua);
984
985struct blk_independent_access_ranges *
986disk_alloc_independent_access_ranges(struct gendisk *disk, int nr_ia_ranges);
987void disk_set_independent_access_ranges(struct gendisk *disk,
988 struct blk_independent_access_ranges *iars);
989
990/*
991 * Elevator features for blk_queue_required_elevator_features:
992 */
993/* Supports zoned block devices sequential write constraint */
994#define ELEVATOR_F_ZBD_SEQ_WRITE (1U << 0)
995/* Supports scheduling on multiple hardware queues */
996#define ELEVATOR_F_MQ_AWARE (1U << 1)
997
998extern void blk_queue_required_elevator_features(struct request_queue *q,
999 unsigned int features);
1000extern bool blk_queue_can_use_dma_map_merging(struct request_queue *q,
1001 struct device *dev);
1002
1003bool __must_check blk_get_queue(struct request_queue *);
1004extern void blk_put_queue(struct request_queue *);
1005
1006void blk_mark_disk_dead(struct gendisk *disk);
1007
1008#ifdef CONFIG_BLOCK
1009/*
1010 * blk_plug permits building a queue of related requests by holding the I/O
1011 * fragments for a short period. This allows merging of sequential requests
1012 * into single larger request. As the requests are moved from a per-task list to
1013 * the device's request_queue in a batch, this results in improved scalability
1014 * as the lock contention for request_queue lock is reduced.
1015 *
1016 * It is ok not to disable preemption when adding the request to the plug list
1017 * or when attempting a merge. For details, please see schedule() where
1018 * blk_flush_plug() is called.
1019 */
1020struct blk_plug {
1021 struct request *mq_list; /* blk-mq requests */
1022
1023 /* if ios_left is > 1, we can batch tag/rq allocations */
1024 struct request *cached_rq;
1025 unsigned short nr_ios;
1026
1027 unsigned short rq_count;
1028
1029 bool multiple_queues;
1030 bool has_elevator;
1031 bool nowait;
1032
1033 struct list_head cb_list; /* md requires an unplug callback */
1034};
1035
1036struct blk_plug_cb;
1037typedef void (*blk_plug_cb_fn)(struct blk_plug_cb *, bool);
1038struct blk_plug_cb {
1039 struct list_head list;
1040 blk_plug_cb_fn callback;
1041 void *data;
1042};
1043extern struct blk_plug_cb *blk_check_plugged(blk_plug_cb_fn unplug,
1044 void *data, int size);
1045extern void blk_start_plug(struct blk_plug *);
1046extern void blk_start_plug_nr_ios(struct blk_plug *, unsigned short);
1047extern void blk_finish_plug(struct blk_plug *);
1048
1049void __blk_flush_plug(struct blk_plug *plug, bool from_schedule);
1050static inline void blk_flush_plug(struct blk_plug *plug, bool async)
1051{
1052 if (plug)
1053 __blk_flush_plug(plug, async);
1054}
1055
1056int blkdev_issue_flush(struct block_device *bdev);
1057long nr_blockdev_pages(void);
1058#else /* CONFIG_BLOCK */
1059struct blk_plug {
1060};
1061
1062static inline void blk_start_plug_nr_ios(struct blk_plug *plug,
1063 unsigned short nr_ios)
1064{
1065}
1066
1067static inline void blk_start_plug(struct blk_plug *plug)
1068{
1069}
1070
1071static inline void blk_finish_plug(struct blk_plug *plug)
1072{
1073}
1074
1075static inline void blk_flush_plug(struct blk_plug *plug, bool async)
1076{
1077}
1078
1079static inline int blkdev_issue_flush(struct block_device *bdev)
1080{
1081 return 0;
1082}
1083
1084static inline long nr_blockdev_pages(void)
1085{
1086 return 0;
1087}
1088#endif /* CONFIG_BLOCK */
1089
1090extern void blk_io_schedule(void);
1091
1092#define BLKDEV_DISCARD_SECURE (1 << 0) /* issue a secure erase */
1093
1094extern int blkdev_issue_discard(struct block_device *bdev, sector_t sector,
1095 sector_t nr_sects, gfp_t gfp_mask, unsigned long flags);
1096extern int __blkdev_issue_discard(struct block_device *bdev, sector_t sector,
1097 sector_t nr_sects, gfp_t gfp_mask, int flags,
1098 struct bio **biop);
1099
1100#define BLKDEV_ZERO_NOUNMAP (1 << 0) /* do not free blocks */
1101#define BLKDEV_ZERO_NOFALLBACK (1 << 1) /* don't write explicit zeroes */
1102
1103extern int __blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
1104 sector_t nr_sects, gfp_t gfp_mask, struct bio **biop,
1105 unsigned flags);
1106extern int blkdev_issue_zeroout(struct block_device *bdev, sector_t sector,
1107 sector_t nr_sects, gfp_t gfp_mask, unsigned flags);
1108
1109static inline int sb_issue_discard(struct super_block *sb, sector_t block,
1110 sector_t nr_blocks, gfp_t gfp_mask, unsigned long flags)
1111{
1112 return blkdev_issue_discard(sb->s_bdev,
1113 block << (sb->s_blocksize_bits -
1114 SECTOR_SHIFT),
1115 nr_blocks << (sb->s_blocksize_bits -
1116 SECTOR_SHIFT),
1117 gfp_mask, flags);
1118}
1119static inline int sb_issue_zeroout(struct super_block *sb, sector_t block,
1120 sector_t nr_blocks, gfp_t gfp_mask)
1121{
1122 return blkdev_issue_zeroout(sb->s_bdev,
1123 block << (sb->s_blocksize_bits -
1124 SECTOR_SHIFT),
1125 nr_blocks << (sb->s_blocksize_bits -
1126 SECTOR_SHIFT),
1127 gfp_mask, 0);
1128}
1129
1130static inline bool bdev_is_partition(struct block_device *bdev)
1131{
1132 return bdev->bd_partno;
1133}
1134
1135enum blk_default_limits {
1136 BLK_MAX_SEGMENTS = 128,
1137 BLK_SAFE_MAX_SECTORS = 255,
1138 BLK_DEF_MAX_SECTORS = 2560,
1139 BLK_MAX_SEGMENT_SIZE = 65536,
1140 BLK_SEG_BOUNDARY_MASK = 0xFFFFFFFFUL,
1141};
1142
1143static inline unsigned long queue_segment_boundary(const struct request_queue *q)
1144{
1145 return q->limits.seg_boundary_mask;
1146}
1147
1148static inline unsigned long queue_virt_boundary(const struct request_queue *q)
1149{
1150 return q->limits.virt_boundary_mask;
1151}
1152
1153static inline unsigned int queue_max_sectors(const struct request_queue *q)
1154{
1155 return q->limits.max_sectors;
1156}
1157
1158static inline unsigned int queue_max_bytes(struct request_queue *q)
1159{
1160 return min_t(unsigned int, queue_max_sectors(q), INT_MAX >> 9) << 9;
1161}
1162
1163static inline unsigned int queue_max_hw_sectors(const struct request_queue *q)
1164{
1165 return q->limits.max_hw_sectors;
1166}
1167
1168static inline unsigned short queue_max_segments(const struct request_queue *q)
1169{
1170 return q->limits.max_segments;
1171}
1172
1173static inline unsigned short queue_max_discard_segments(const struct request_queue *q)
1174{
1175 return q->limits.max_discard_segments;
1176}
1177
1178static inline unsigned int queue_max_segment_size(const struct request_queue *q)
1179{
1180 return q->limits.max_segment_size;
1181}
1182
1183static inline unsigned int queue_max_zone_append_sectors(const struct request_queue *q)
1184{
1185
1186 const struct queue_limits *l = &q->limits;
1187
1188 return min(l->max_zone_append_sectors, l->max_sectors);
1189}
1190
1191static inline unsigned queue_logical_block_size(const struct request_queue *q)
1192{
1193 int retval = 512;
1194
1195 if (q && q->limits.logical_block_size)
1196 retval = q->limits.logical_block_size;
1197
1198 return retval;
1199}
1200
1201static inline unsigned int bdev_logical_block_size(struct block_device *bdev)
1202{
1203 return queue_logical_block_size(bdev_get_queue(bdev));
1204}
1205
1206static inline unsigned int queue_physical_block_size(const struct request_queue *q)
1207{
1208 return q->limits.physical_block_size;
1209}
1210
1211static inline unsigned int bdev_physical_block_size(struct block_device *bdev)
1212{
1213 return queue_physical_block_size(bdev_get_queue(bdev));
1214}
1215
1216static inline unsigned int queue_io_min(const struct request_queue *q)
1217{
1218 return q->limits.io_min;
1219}
1220
1221static inline int bdev_io_min(struct block_device *bdev)
1222{
1223 return queue_io_min(bdev_get_queue(bdev));
1224}
1225
1226static inline unsigned int queue_io_opt(const struct request_queue *q)
1227{
1228 return q->limits.io_opt;
1229}
1230
1231static inline int bdev_io_opt(struct block_device *bdev)
1232{
1233 return queue_io_opt(bdev_get_queue(bdev));
1234}
1235
1236static inline unsigned int
1237queue_zone_write_granularity(const struct request_queue *q)
1238{
1239 return q->limits.zone_write_granularity;
1240}
1241
1242static inline unsigned int
1243bdev_zone_write_granularity(struct block_device *bdev)
1244{
1245 return queue_zone_write_granularity(bdev_get_queue(bdev));
1246}
1247
1248static inline int queue_alignment_offset(const struct request_queue *q)
1249{
1250 if (q->limits.misaligned)
1251 return -1;
1252
1253 return q->limits.alignment_offset;
1254}
1255
1256static inline int queue_limit_alignment_offset(struct queue_limits *lim, sector_t sector)
1257{
1258 unsigned int granularity = max(lim->physical_block_size, lim->io_min);
1259 unsigned int alignment = sector_div(sector, granularity >> SECTOR_SHIFT)
1260 << SECTOR_SHIFT;
1261
1262 return (granularity + lim->alignment_offset - alignment) % granularity;
1263}
1264
1265static inline int bdev_alignment_offset(struct block_device *bdev)
1266{
1267 struct request_queue *q = bdev_get_queue(bdev);
1268
1269 if (q->limits.misaligned)
1270 return -1;
1271 if (bdev_is_partition(bdev))
1272 return queue_limit_alignment_offset(&q->limits,
1273 bdev->bd_start_sect);
1274 return q->limits.alignment_offset;
1275}
1276
1277static inline int queue_discard_alignment(const struct request_queue *q)
1278{
1279 if (q->limits.discard_misaligned)
1280 return -1;
1281
1282 return q->limits.discard_alignment;
1283}
1284
1285static inline int queue_limit_discard_alignment(struct queue_limits *lim, sector_t sector)
1286{
1287 unsigned int alignment, granularity, offset;
1288
1289 if (!lim->max_discard_sectors)
1290 return 0;
1291
1292 /* Why are these in bytes, not sectors? */
1293 alignment = lim->discard_alignment >> SECTOR_SHIFT;
1294 granularity = lim->discard_granularity >> SECTOR_SHIFT;
1295 if (!granularity)
1296 return 0;
1297
1298 /* Offset of the partition start in 'granularity' sectors */
1299 offset = sector_div(sector, granularity);
1300
1301 /* And why do we do this modulus *again* in blkdev_issue_discard()? */
1302 offset = (granularity + alignment - offset) % granularity;
1303
1304 /* Turn it back into bytes, gaah */
1305 return offset << SECTOR_SHIFT;
1306}
1307
1308static inline int bdev_discard_alignment(struct block_device *bdev)
1309{
1310 struct request_queue *q = bdev_get_queue(bdev);
1311
1312 if (bdev_is_partition(bdev))
1313 return queue_limit_discard_alignment(&q->limits,
1314 bdev->bd_start_sect);
1315 return q->limits.discard_alignment;
1316}
1317
1318static inline unsigned int bdev_write_zeroes_sectors(struct block_device *bdev)
1319{
1320 struct request_queue *q = bdev_get_queue(bdev);
1321
1322 if (q)
1323 return q->limits.max_write_zeroes_sectors;
1324
1325 return 0;
1326}
1327
1328static inline bool bdev_nonrot(struct block_device *bdev)
1329{
1330 return blk_queue_nonrot(bdev_get_queue(bdev));
1331}
1332
1333static inline bool bdev_write_cache(struct block_device *bdev)
1334{
1335 return test_bit(QUEUE_FLAG_WC, &bdev_get_queue(bdev)->queue_flags);
1336}
1337
1338static inline bool bdev_fua(struct block_device *bdev)
1339{
1340 return test_bit(QUEUE_FLAG_FUA, &bdev_get_queue(bdev)->queue_flags);
1341}
1342
1343static inline enum blk_zoned_model bdev_zoned_model(struct block_device *bdev)
1344{
1345 struct request_queue *q = bdev_get_queue(bdev);
1346
1347 if (q)
1348 return blk_queue_zoned_model(q);
1349
1350 return BLK_ZONED_NONE;
1351}
1352
1353static inline bool bdev_is_zoned(struct block_device *bdev)
1354{
1355 struct request_queue *q = bdev_get_queue(bdev);
1356
1357 if (q)
1358 return blk_queue_is_zoned(q);
1359
1360 return false;
1361}
1362
1363static inline sector_t bdev_zone_sectors(struct block_device *bdev)
1364{
1365 struct request_queue *q = bdev_get_queue(bdev);
1366
1367 if (q)
1368 return blk_queue_zone_sectors(q);
1369 return 0;
1370}
1371
1372static inline unsigned int bdev_max_open_zones(struct block_device *bdev)
1373{
1374 struct request_queue *q = bdev_get_queue(bdev);
1375
1376 if (q)
1377 return queue_max_open_zones(q);
1378 return 0;
1379}
1380
1381static inline unsigned int bdev_max_active_zones(struct block_device *bdev)
1382{
1383 struct request_queue *q = bdev_get_queue(bdev);
1384
1385 if (q)
1386 return queue_max_active_zones(q);
1387 return 0;
1388}
1389
1390static inline int queue_dma_alignment(const struct request_queue *q)
1391{
1392 return q ? q->dma_alignment : 511;
1393}
1394
1395static inline int blk_rq_aligned(struct request_queue *q, unsigned long addr,
1396 unsigned int len)
1397{
1398 unsigned int alignment = queue_dma_alignment(q) | q->dma_pad_mask;
1399 return !(addr & alignment) && !(len & alignment);
1400}
1401
1402/* assumes size > 256 */
1403static inline unsigned int blksize_bits(unsigned int size)
1404{
1405 unsigned int bits = 8;
1406 do {
1407 bits++;
1408 size >>= 1;
1409 } while (size > 256);
1410 return bits;
1411}
1412
1413static inline unsigned int block_size(struct block_device *bdev)
1414{
1415 return 1 << bdev->bd_inode->i_blkbits;
1416}
1417
1418int kblockd_schedule_work(struct work_struct *work);
1419int kblockd_mod_delayed_work_on(int cpu, struct delayed_work *dwork, unsigned long delay);
1420
1421#define MODULE_ALIAS_BLOCKDEV(major,minor) \
1422 MODULE_ALIAS("block-major-" __stringify(major) "-" __stringify(minor))
1423#define MODULE_ALIAS_BLOCKDEV_MAJOR(major) \
1424 MODULE_ALIAS("block-major-" __stringify(major) "-*")
1425
1426#ifdef CONFIG_BLK_INLINE_ENCRYPTION
1427
1428bool blk_crypto_register(struct blk_crypto_profile *profile,
1429 struct request_queue *q);
1430
1431#else /* CONFIG_BLK_INLINE_ENCRYPTION */
1432
1433static inline bool blk_crypto_register(struct blk_crypto_profile *profile,
1434 struct request_queue *q)
1435{
1436 return true;
1437}
1438
1439#endif /* CONFIG_BLK_INLINE_ENCRYPTION */
1440
1441enum blk_unique_id {
1442 /* these match the Designator Types specified in SPC */
1443 BLK_UID_T10 = 1,
1444 BLK_UID_EUI64 = 2,
1445 BLK_UID_NAA = 3,
1446};
1447
1448#define NFL4_UFLG_MASK 0x0000003F
1449
1450struct block_device_operations {
1451 void (*submit_bio)(struct bio *bio);
1452 int (*poll_bio)(struct bio *bio, struct io_comp_batch *iob,
1453 unsigned int flags);
1454 int (*open) (struct block_device *, fmode_t);
1455 void (*release) (struct gendisk *, fmode_t);
1456 int (*rw_page)(struct block_device *, sector_t, struct page *, unsigned int);
1457 int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
1458 int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long);
1459 unsigned int (*check_events) (struct gendisk *disk,
1460 unsigned int clearing);
1461 void (*unlock_native_capacity) (struct gendisk *);
1462 int (*getgeo)(struct block_device *, struct hd_geometry *);
1463 int (*set_read_only)(struct block_device *bdev, bool ro);
1464 void (*free_disk)(struct gendisk *disk);
1465 /* this callback is with swap_lock and sometimes page table lock held */
1466 void (*swap_slot_free_notify) (struct block_device *, unsigned long);
1467 int (*report_zones)(struct gendisk *, sector_t sector,
1468 unsigned int nr_zones, report_zones_cb cb, void *data);
1469 char *(*devnode)(struct gendisk *disk, umode_t *mode);
1470 /* returns the length of the identifier or a negative errno: */
1471 int (*get_unique_id)(struct gendisk *disk, u8 id[16],
1472 enum blk_unique_id id_type);
1473 struct module *owner;
1474 const struct pr_ops *pr_ops;
1475
1476 /*
1477 * Special callback for probing GPT entry at a given sector.
1478 * Needed by Android devices, used by GPT scanner and MMC blk
1479 * driver.
1480 */
1481 int (*alternative_gpt_sector)(struct gendisk *disk, sector_t *sector);
1482};
1483
1484#ifdef CONFIG_COMPAT
1485extern int blkdev_compat_ptr_ioctl(struct block_device *, fmode_t,
1486 unsigned int, unsigned long);
1487#else
1488#define blkdev_compat_ptr_ioctl NULL
1489#endif
1490
1491extern int bdev_read_page(struct block_device *, sector_t, struct page *);
1492extern int bdev_write_page(struct block_device *, sector_t, struct page *,
1493 struct writeback_control *);
1494
1495static inline void blk_wake_io_task(struct task_struct *waiter)
1496{
1497 /*
1498 * If we're polling, the task itself is doing the completions. For
1499 * that case, we don't need to signal a wakeup, it's enough to just
1500 * mark us as RUNNING.
1501 */
1502 if (waiter == current)
1503 __set_current_state(TASK_RUNNING);
1504 else
1505 wake_up_process(waiter);
1506}
1507
1508unsigned long disk_start_io_acct(struct gendisk *disk, unsigned int sectors,
1509 unsigned int op);
1510void disk_end_io_acct(struct gendisk *disk, unsigned int op,
1511 unsigned long start_time);
1512
1513void bio_start_io_acct_time(struct bio *bio, unsigned long start_time);
1514unsigned long bio_start_io_acct(struct bio *bio);
1515void bio_end_io_acct_remapped(struct bio *bio, unsigned long start_time,
1516 struct block_device *orig_bdev);
1517
1518/**
1519 * bio_end_io_acct - end I/O accounting for bio based drivers
1520 * @bio: bio to end account for
1521 * @start_time: start time returned by bio_start_io_acct()
1522 */
1523static inline void bio_end_io_acct(struct bio *bio, unsigned long start_time)
1524{
1525 return bio_end_io_acct_remapped(bio, start_time, bio->bi_bdev);
1526}
1527
1528int bdev_read_only(struct block_device *bdev);
1529int set_blocksize(struct block_device *bdev, int size);
1530
1531const char *bdevname(struct block_device *bdev, char *buffer);
1532int lookup_bdev(const char *pathname, dev_t *dev);
1533
1534void blkdev_show(struct seq_file *seqf, off_t offset);
1535
1536#define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */
1537#define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */
1538#ifdef CONFIG_BLOCK
1539#define BLKDEV_MAJOR_MAX 512
1540#else
1541#define BLKDEV_MAJOR_MAX 0
1542#endif
1543
1544struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
1545 void *holder);
1546struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode, void *holder);
1547int bd_prepare_to_claim(struct block_device *bdev, void *holder);
1548void bd_abort_claiming(struct block_device *bdev, void *holder);
1549void blkdev_put(struct block_device *bdev, fmode_t mode);
1550
1551/* just for blk-cgroup, don't use elsewhere */
1552struct block_device *blkdev_get_no_open(dev_t dev);
1553void blkdev_put_no_open(struct block_device *bdev);
1554
1555struct block_device *bdev_alloc(struct gendisk *disk, u8 partno);
1556void bdev_add(struct block_device *bdev, dev_t dev);
1557struct block_device *I_BDEV(struct inode *inode);
1558int truncate_bdev_range(struct block_device *bdev, fmode_t mode, loff_t lstart,
1559 loff_t lend);
1560
1561#ifdef CONFIG_BLOCK
1562void invalidate_bdev(struct block_device *bdev);
1563int sync_blockdev(struct block_device *bdev);
1564int sync_blockdev_nowait(struct block_device *bdev);
1565void sync_bdevs(bool wait);
1566void printk_all_partitions(void);
1567#else
1568static inline void invalidate_bdev(struct block_device *bdev)
1569{
1570}
1571static inline int sync_blockdev(struct block_device *bdev)
1572{
1573 return 0;
1574}
1575static inline int sync_blockdev_nowait(struct block_device *bdev)
1576{
1577 return 0;
1578}
1579static inline void sync_bdevs(bool wait)
1580{
1581}
1582static inline void printk_all_partitions(void)
1583{
1584}
1585#endif /* CONFIG_BLOCK */
1586
1587int fsync_bdev(struct block_device *bdev);
1588
1589int freeze_bdev(struct block_device *bdev);
1590int thaw_bdev(struct block_device *bdev);
1591
1592struct io_comp_batch {
1593 struct request *req_list;
1594 bool need_ts;
1595 void (*complete)(struct io_comp_batch *);
1596};
1597
1598#define DEFINE_IO_COMP_BATCH(name) struct io_comp_batch name = { }
1599
1600#endif /* _LINUX_BLKDEV_H */
Linux 6.x block layer and io_uring tree(s)