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 29 struct module; 30 struct request_queue; 31 struct elevator_queue; 32 struct blk_trace; 33 struct request; 34 struct sg_io_hdr; 35 struct blkcg_gq; 36 struct blk_flush_queue; 37 struct kiocb; 38 struct pr_ops; 39 struct rq_qos; 40 struct blk_queue_stats; 41 struct blk_stat_callback; 42 struct blk_crypto_profile; 43 44 extern const struct device_type disk_type; 45 extern struct device_type part_type; 46 extern 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 70 struct 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 */ 91 enum { 92 GENHD_FL_REMOVABLE = 1 << 0, 93 GENHD_FL_HIDDEN = 1 << 1, 94 GENHD_FL_NO_PART = 1 << 2, 95 }; 96 97 enum { 98 DISK_EVENT_MEDIA_CHANGE = 1 << 0, /* media changed */ 99 DISK_EVENT_EJECT_REQUEST = 1 << 1, /* eject requested */ 100 }; 101 102 enum { 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 }; 110 111 struct disk_events; 112 struct badblocks; 113 114 struct 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 }; 121 122 struct 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 }; 173 174 static 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 193 194 static inline dev_t disk_devt(struct gendisk *disk) 195 { 196 return MKDEV(disk->major, disk->first_minor); 197 } 198 199 static 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 } 206 207 static 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 */ 219 enum 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 */ 229 enum blk_bounce { 230 BLK_BOUNCE_NONE, 231 BLK_BOUNCE_HIGH, 232 }; 233 234 struct 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_same_sectors; 252 unsigned int max_write_zeroes_sectors; 253 unsigned int max_zone_append_sectors; 254 unsigned int discard_granularity; 255 unsigned int discard_alignment; 256 unsigned int zone_write_granularity; 257 258 unsigned short max_segments; 259 unsigned short max_integrity_segments; 260 unsigned short max_discard_segments; 261 262 unsigned char misaligned; 263 unsigned char discard_misaligned; 264 unsigned char raid_partial_stripes_expensive; 265 enum blk_zoned_model zoned; 266 }; 267 268 typedef int (*report_zones_cb)(struct blk_zone *zone, unsigned int idx, 269 void *data); 270 271 void blk_queue_set_zoned(struct gendisk *disk, enum blk_zoned_model model); 272 273 #ifdef CONFIG_BLK_DEV_ZONED 274 275 #define BLK_ALL_ZONES ((unsigned int)-1) 276 int blkdev_report_zones(struct block_device *bdev, sector_t sector, 277 unsigned int nr_zones, report_zones_cb cb, void *data); 278 unsigned int blkdev_nr_zones(struct gendisk *disk); 279 extern int blkdev_zone_mgmt(struct block_device *bdev, enum req_opf op, 280 sector_t sectors, sector_t nr_sectors, 281 gfp_t gfp_mask); 282 int blk_revalidate_disk_zones(struct gendisk *disk, 283 void (*update_driver_data)(struct gendisk *disk)); 284 285 extern int blkdev_report_zones_ioctl(struct block_device *bdev, fmode_t mode, 286 unsigned int cmd, unsigned long arg); 287 extern int blkdev_zone_mgmt_ioctl(struct block_device *bdev, fmode_t mode, 288 unsigned int cmd, unsigned long arg); 289 290 #else /* CONFIG_BLK_DEV_ZONED */ 291 292 static inline unsigned int blkdev_nr_zones(struct gendisk *disk) 293 { 294 return 0; 295 } 296 297 static inline int blkdev_report_zones_ioctl(struct block_device *bdev, 298 fmode_t mode, unsigned int cmd, 299 unsigned long arg) 300 { 301 return -ENOTTY; 302 } 303 304 static inline int blkdev_zone_mgmt_ioctl(struct block_device *bdev, 305 fmode_t mode, unsigned int cmd, 306 unsigned long arg) 307 { 308 return -ENOTTY; 309 } 310 311 #endif /* CONFIG_BLK_DEV_ZONED */ 312 313 /* 314 * Independent access ranges: struct blk_independent_access_range describes 315 * a range of contiguous sectors that can be accessed using device command 316 * execution resources that are independent from the resources used for 317 * other access ranges. This is typically found with single-LUN multi-actuator 318 * HDDs where each access range is served by a different set of heads. 319 * The set of independent ranges supported by the device is defined using 320 * struct blk_independent_access_ranges. The independent ranges must not overlap 321 * and must include all sectors within the disk capacity (no sector holes 322 * allowed). 323 * For a device with multiple ranges, requests targeting sectors in different 324 * ranges can be executed in parallel. A request can straddle an access range 325 * boundary. 326 */ 327 struct blk_independent_access_range { 328 struct kobject kobj; 329 struct request_queue *queue; 330 sector_t sector; 331 sector_t nr_sectors; 332 }; 333 334 struct blk_independent_access_ranges { 335 struct kobject kobj; 336 bool sysfs_registered; 337 unsigned int nr_ia_ranges; 338 struct blk_independent_access_range ia_range[]; 339 }; 340 341 struct request_queue { 342 struct request *last_merge; 343 struct elevator_queue *elevator; 344 345 struct percpu_ref q_usage_counter; 346 347 struct blk_queue_stats *stats; 348 struct rq_qos *rq_qos; 349 350 const struct blk_mq_ops *mq_ops; 351 352 /* sw queues */ 353 struct blk_mq_ctx __percpu *queue_ctx; 354 355 unsigned int queue_depth; 356 357 /* hw dispatch queues */ 358 struct blk_mq_hw_ctx **queue_hw_ctx; 359 unsigned int nr_hw_queues; 360 361 /* 362 * The queue owner gets to use this for whatever they like. 363 * ll_rw_blk doesn't touch it. 364 */ 365 void *queuedata; 366 367 /* 368 * various queue flags, see QUEUE_* below 369 */ 370 unsigned long queue_flags; 371 /* 372 * Number of contexts that have called blk_set_pm_only(). If this 373 * counter is above zero then only RQF_PM requests are processed. 374 */ 375 atomic_t pm_only; 376 377 /* 378 * ida allocated id for this queue. Used to index queues from 379 * ioctx. 380 */ 381 int id; 382 383 spinlock_t queue_lock; 384 385 struct gendisk *disk; 386 387 /* 388 * queue kobject 389 */ 390 struct kobject kobj; 391 392 /* 393 * mq queue kobject 394 */ 395 struct kobject *mq_kobj; 396 397 #ifdef CONFIG_BLK_DEV_INTEGRITY 398 struct blk_integrity integrity; 399 #endif /* CONFIG_BLK_DEV_INTEGRITY */ 400 401 #ifdef CONFIG_PM 402 struct device *dev; 403 enum rpm_status rpm_status; 404 #endif 405 406 /* 407 * queue settings 408 */ 409 unsigned long nr_requests; /* Max # of requests */ 410 411 unsigned int dma_pad_mask; 412 unsigned int dma_alignment; 413 414 #ifdef CONFIG_BLK_INLINE_ENCRYPTION 415 struct blk_crypto_profile *crypto_profile; 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 #define BLK_MAX_WRITE_HINTS 5 521 u64 write_hints[BLK_MAX_WRITE_HINTS]; 522 523 /* 524 * Independent sector access ranges. This is always NULL for 525 * devices that do not have multiple independent access ranges. 526 */ 527 struct blk_independent_access_ranges *ia_ranges; 528 529 /** 530 * @srcu: Sleepable RCU. Use as lock when type of the request queue 531 * is blocking (BLK_MQ_F_BLOCKING). Must be the last member 532 */ 533 struct srcu_struct srcu[]; 534 }; 535 536 /* Keep blk_queue_flag_name[] in sync with the definitions below */ 537 #define QUEUE_FLAG_STOPPED 0 /* queue is stopped */ 538 #define QUEUE_FLAG_DYING 1 /* queue being torn down */ 539 #define QUEUE_FLAG_HAS_SRCU 2 /* SRCU is allocated */ 540 #define QUEUE_FLAG_NOMERGES 3 /* disable merge attempts */ 541 #define QUEUE_FLAG_SAME_COMP 4 /* complete on same CPU-group */ 542 #define QUEUE_FLAG_FAIL_IO 5 /* fake timeout */ 543 #define QUEUE_FLAG_NONROT 6 /* non-rotational device (SSD) */ 544 #define QUEUE_FLAG_VIRT QUEUE_FLAG_NONROT /* paravirt device */ 545 #define QUEUE_FLAG_IO_STAT 7 /* do disk/partitions IO accounting */ 546 #define QUEUE_FLAG_DISCARD 8 /* supports DISCARD */ 547 #define QUEUE_FLAG_NOXMERGES 9 /* No extended merges */ 548 #define QUEUE_FLAG_ADD_RANDOM 10 /* Contributes to random pool */ 549 #define QUEUE_FLAG_SECERASE 11 /* supports secure erase */ 550 #define QUEUE_FLAG_SAME_FORCE 12 /* force complete on same CPU */ 551 #define QUEUE_FLAG_DEAD 13 /* queue tear-down finished */ 552 #define QUEUE_FLAG_INIT_DONE 14 /* queue is initialized */ 553 #define QUEUE_FLAG_STABLE_WRITES 15 /* don't modify blks until WB is done */ 554 #define QUEUE_FLAG_POLL 16 /* IO polling enabled if set */ 555 #define QUEUE_FLAG_WC 17 /* Write back caching */ 556 #define QUEUE_FLAG_FUA 18 /* device supports FUA writes */ 557 #define QUEUE_FLAG_DAX 19 /* device supports DAX */ 558 #define QUEUE_FLAG_STATS 20 /* track IO start and completion times */ 559 #define QUEUE_FLAG_REGISTERED 22 /* queue has been registered to a disk */ 560 #define QUEUE_FLAG_QUIESCED 24 /* queue has been quiesced */ 561 #define QUEUE_FLAG_PCI_P2PDMA 25 /* device supports PCI p2p requests */ 562 #define QUEUE_FLAG_ZONE_RESETALL 26 /* supports Zone Reset All */ 563 #define QUEUE_FLAG_RQ_ALLOC_TIME 27 /* record rq->alloc_time_ns */ 564 #define QUEUE_FLAG_HCTX_ACTIVE 28 /* at least one blk-mq hctx is active */ 565 #define QUEUE_FLAG_NOWAIT 29 /* device supports NOWAIT */ 566 567 #define QUEUE_FLAG_MQ_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \ 568 (1 << QUEUE_FLAG_SAME_COMP) | \ 569 (1 << QUEUE_FLAG_NOWAIT)) 570 571 void blk_queue_flag_set(unsigned int flag, struct request_queue *q); 572 void blk_queue_flag_clear(unsigned int flag, struct request_queue *q); 573 bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q); 574 575 #define blk_queue_stopped(q) test_bit(QUEUE_FLAG_STOPPED, &(q)->queue_flags) 576 #define blk_queue_dying(q) test_bit(QUEUE_FLAG_DYING, &(q)->queue_flags) 577 #define blk_queue_has_srcu(q) test_bit(QUEUE_FLAG_HAS_SRCU, &(q)->queue_flags) 578 #define blk_queue_dead(q) test_bit(QUEUE_FLAG_DEAD, &(q)->queue_flags) 579 #define blk_queue_init_done(q) test_bit(QUEUE_FLAG_INIT_DONE, &(q)->queue_flags) 580 #define blk_queue_nomerges(q) test_bit(QUEUE_FLAG_NOMERGES, &(q)->queue_flags) 581 #define blk_queue_noxmerges(q) \ 582 test_bit(QUEUE_FLAG_NOXMERGES, &(q)->queue_flags) 583 #define blk_queue_nonrot(q) test_bit(QUEUE_FLAG_NONROT, &(q)->queue_flags) 584 #define blk_queue_stable_writes(q) \ 585 test_bit(QUEUE_FLAG_STABLE_WRITES, &(q)->queue_flags) 586 #define blk_queue_io_stat(q) test_bit(QUEUE_FLAG_IO_STAT, &(q)->queue_flags) 587 #define blk_queue_add_random(q) test_bit(QUEUE_FLAG_ADD_RANDOM, &(q)->queue_flags) 588 #define blk_queue_discard(q) test_bit(QUEUE_FLAG_DISCARD, &(q)->queue_flags) 589 #define blk_queue_zone_resetall(q) \ 590 test_bit(QUEUE_FLAG_ZONE_RESETALL, &(q)->queue_flags) 591 #define blk_queue_secure_erase(q) \ 592 (test_bit(QUEUE_FLAG_SECERASE, &(q)->queue_flags)) 593 #define blk_queue_dax(q) test_bit(QUEUE_FLAG_DAX, &(q)->queue_flags) 594 #define blk_queue_pci_p2pdma(q) \ 595 test_bit(QUEUE_FLAG_PCI_P2PDMA, &(q)->queue_flags) 596 #ifdef CONFIG_BLK_RQ_ALLOC_TIME 597 #define blk_queue_rq_alloc_time(q) \ 598 test_bit(QUEUE_FLAG_RQ_ALLOC_TIME, &(q)->queue_flags) 599 #else 600 #define blk_queue_rq_alloc_time(q) false 601 #endif 602 603 #define blk_noretry_request(rq) \ 604 ((rq)->cmd_flags & (REQ_FAILFAST_DEV|REQ_FAILFAST_TRANSPORT| \ 605 REQ_FAILFAST_DRIVER)) 606 #define blk_queue_quiesced(q) test_bit(QUEUE_FLAG_QUIESCED, &(q)->queue_flags) 607 #define blk_queue_pm_only(q) atomic_read(&(q)->pm_only) 608 #define blk_queue_fua(q) test_bit(QUEUE_FLAG_FUA, &(q)->queue_flags) 609 #define blk_queue_registered(q) test_bit(QUEUE_FLAG_REGISTERED, &(q)->queue_flags) 610 #define blk_queue_nowait(q) test_bit(QUEUE_FLAG_NOWAIT, &(q)->queue_flags) 611 612 extern void blk_set_pm_only(struct request_queue *q); 613 extern void blk_clear_pm_only(struct request_queue *q); 614 615 #define list_entry_rq(ptr) list_entry((ptr), struct request, queuelist) 616 617 #define dma_map_bvec(dev, bv, dir, attrs) \ 618 dma_map_page_attrs(dev, (bv)->bv_page, (bv)->bv_offset, (bv)->bv_len, \ 619 (dir), (attrs)) 620 621 static inline bool queue_is_mq(struct request_queue *q) 622 { 623 return q->mq_ops; 624 } 625 626 #ifdef CONFIG_PM 627 static inline enum rpm_status queue_rpm_status(struct request_queue *q) 628 { 629 return q->rpm_status; 630 } 631 #else 632 static inline enum rpm_status queue_rpm_status(struct request_queue *q) 633 { 634 return RPM_ACTIVE; 635 } 636 #endif 637 638 static inline enum blk_zoned_model 639 blk_queue_zoned_model(struct request_queue *q) 640 { 641 if (IS_ENABLED(CONFIG_BLK_DEV_ZONED)) 642 return q->limits.zoned; 643 return BLK_ZONED_NONE; 644 } 645 646 static inline bool blk_queue_is_zoned(struct request_queue *q) 647 { 648 switch (blk_queue_zoned_model(q)) { 649 case BLK_ZONED_HA: 650 case BLK_ZONED_HM: 651 return true; 652 default: 653 return false; 654 } 655 } 656 657 static inline sector_t blk_queue_zone_sectors(struct request_queue *q) 658 { 659 return blk_queue_is_zoned(q) ? q->limits.chunk_sectors : 0; 660 } 661 662 #ifdef CONFIG_BLK_DEV_ZONED 663 static inline unsigned int blk_queue_nr_zones(struct request_queue *q) 664 { 665 return blk_queue_is_zoned(q) ? q->nr_zones : 0; 666 } 667 668 static inline unsigned int blk_queue_zone_no(struct request_queue *q, 669 sector_t sector) 670 { 671 if (!blk_queue_is_zoned(q)) 672 return 0; 673 return sector >> ilog2(q->limits.chunk_sectors); 674 } 675 676 static inline bool blk_queue_zone_is_seq(struct request_queue *q, 677 sector_t sector) 678 { 679 if (!blk_queue_is_zoned(q)) 680 return false; 681 if (!q->conv_zones_bitmap) 682 return true; 683 return !test_bit(blk_queue_zone_no(q, sector), q->conv_zones_bitmap); 684 } 685 686 static inline void blk_queue_max_open_zones(struct request_queue *q, 687 unsigned int max_open_zones) 688 { 689 q->max_open_zones = max_open_zones; 690 } 691 692 static inline unsigned int queue_max_open_zones(const struct request_queue *q) 693 { 694 return q->max_open_zones; 695 } 696 697 static inline void blk_queue_max_active_zones(struct request_queue *q, 698 unsigned int max_active_zones) 699 { 700 q->max_active_zones = max_active_zones; 701 } 702 703 static inline unsigned int queue_max_active_zones(const struct request_queue *q) 704 { 705 return q->max_active_zones; 706 } 707 #else /* CONFIG_BLK_DEV_ZONED */ 708 static inline unsigned int blk_queue_nr_zones(struct request_queue *q) 709 { 710 return 0; 711 } 712 static inline bool blk_queue_zone_is_seq(struct request_queue *q, 713 sector_t sector) 714 { 715 return false; 716 } 717 static inline unsigned int blk_queue_zone_no(struct request_queue *q, 718 sector_t sector) 719 { 720 return 0; 721 } 722 static inline unsigned int queue_max_open_zones(const struct request_queue *q) 723 { 724 return 0; 725 } 726 static inline unsigned int queue_max_active_zones(const struct request_queue *q) 727 { 728 return 0; 729 } 730 #endif /* CONFIG_BLK_DEV_ZONED */ 731 732 static inline unsigned int blk_queue_depth(struct request_queue *q) 733 { 734 if (q->queue_depth) 735 return q->queue_depth; 736 737 return q->nr_requests; 738 } 739 740 /* 741 * default timeout for SG_IO if none specified 742 */ 743 #define BLK_DEFAULT_SG_TIMEOUT (60 * HZ) 744 #define BLK_MIN_SG_TIMEOUT (7 * HZ) 745 746 /* This should not be used directly - use rq_for_each_segment */ 747 #define for_each_bio(_bio) \ 748 for (; _bio; _bio = _bio->bi_next) 749 750 int __must_check device_add_disk(struct device *parent, struct gendisk *disk, 751 const struct attribute_group **groups); 752 static inline int __must_check add_disk(struct gendisk *disk) 753 { 754 return device_add_disk(NULL, disk, NULL); 755 } 756 void del_gendisk(struct gendisk *gp); 757 void invalidate_disk(struct gendisk *disk); 758 void set_disk_ro(struct gendisk *disk, bool read_only); 759 void disk_uevent(struct gendisk *disk, enum kobject_action action); 760 761 static inline int get_disk_ro(struct gendisk *disk) 762 { 763 return disk->part0->bd_read_only || 764 test_bit(GD_READ_ONLY, &disk->state); 765 } 766 767 static inline int bdev_read_only(struct block_device *bdev) 768 { 769 return bdev->bd_read_only || get_disk_ro(bdev->bd_disk); 770 } 771 772 bool set_capacity_and_notify(struct gendisk *disk, sector_t size); 773 bool disk_force_media_change(struct gendisk *disk, unsigned int events); 774 775 void add_disk_randomness(struct gendisk *disk) __latent_entropy; 776 void rand_initialize_disk(struct gendisk *disk); 777 778 static inline sector_t get_start_sect(struct block_device *bdev) 779 { 780 return bdev->bd_start_sect; 781 } 782 783 static inline sector_t bdev_nr_sectors(struct block_device *bdev) 784 { 785 return bdev->bd_nr_sectors; 786 } 787 788 static inline loff_t bdev_nr_bytes(struct block_device *bdev) 789 { 790 return (loff_t)bdev_nr_sectors(bdev) << SECTOR_SHIFT; 791 } 792 793 static inline sector_t get_capacity(struct gendisk *disk) 794 { 795 return bdev_nr_sectors(disk->part0); 796 } 797 798 static inline u64 sb_bdev_nr_blocks(struct super_block *sb) 799 { 800 return bdev_nr_sectors(sb->s_bdev) >> 801 (sb->s_blocksize_bits - SECTOR_SHIFT); 802 } 803 804 int bdev_disk_changed(struct gendisk *disk, bool invalidate); 805 806 struct gendisk *__alloc_disk_node(struct request_queue *q, int node_id, 807 struct lock_class_key *lkclass); 808 void put_disk(struct gendisk *disk); 809 struct gendisk *__blk_alloc_disk(int node, struct lock_class_key *lkclass); 810 811 /** 812 * blk_alloc_disk - allocate a gendisk structure 813 * @node_id: numa node to allocate on 814 * 815 * Allocate and pre-initialize a gendisk structure for use with BIO based 816 * drivers. 817 * 818 * Context: can sleep 819 */ 820 #define blk_alloc_disk(node_id) \ 821 ({ \ 822 static struct lock_class_key __key; \ 823 \ 824 __blk_alloc_disk(node_id, &__key); \ 825 }) 826 void blk_cleanup_disk(struct gendisk *disk); 827 828 int __register_blkdev(unsigned int major, const char *name, 829 void (*probe)(dev_t devt)); 830 #define register_blkdev(major, name) \ 831 __register_blkdev(major, name, NULL) 832 void unregister_blkdev(unsigned int major, const char *name); 833 834 bool bdev_check_media_change(struct block_device *bdev); 835 int __invalidate_device(struct block_device *bdev, bool kill_dirty); 836 void set_capacity(struct gendisk *disk, sector_t size); 837 838 #ifdef CONFIG_BLOCK_HOLDER_DEPRECATED 839 int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk); 840 void bd_unlink_disk_holder(struct block_device *bdev, struct gendisk *disk); 841 int bd_register_pending_holders(struct gendisk *disk); 842 #else 843 static inline int bd_link_disk_holder(struct block_device *bdev, 844 struct gendisk *disk) 845 { 846 return 0; 847 } 848 static inline void bd_unlink_disk_holder(struct block_device *bdev, 849 struct gendisk *disk) 850 { 851 } 852 static inline int bd_register_pending_holders(struct gendisk *disk) 853 { 854 return 0; 855 } 856 #endif /* CONFIG_BLOCK_HOLDER_DEPRECATED */ 857 858 dev_t part_devt(struct gendisk *disk, u8 partno); 859 void inc_diskseq(struct gendisk *disk); 860 dev_t blk_lookup_devt(const char *name, int partno); 861 void blk_request_module(dev_t devt); 862 863 extern int blk_register_queue(struct gendisk *disk); 864 extern void blk_unregister_queue(struct gendisk *disk); 865 void submit_bio_noacct(struct bio *bio); 866 867 extern int blk_lld_busy(struct request_queue *q); 868 extern void blk_queue_split(struct bio **); 869 extern int blk_queue_enter(struct request_queue *q, blk_mq_req_flags_t flags); 870 extern void blk_queue_exit(struct request_queue *q); 871 extern void blk_sync_queue(struct request_queue *q); 872 873 /* Helper to convert REQ_OP_XXX to its string format XXX */ 874 extern const char *blk_op_str(unsigned int op); 875 876 int blk_status_to_errno(blk_status_t status); 877 blk_status_t errno_to_blk_status(int errno); 878 879 /* only poll the hardware once, don't continue until a completion was found */ 880 #define BLK_POLL_ONESHOT (1 << 0) 881 /* do not sleep to wait for the expected completion time */ 882 #define BLK_POLL_NOSLEEP (1 << 1) 883 int bio_poll(struct bio *bio, struct io_comp_batch *iob, unsigned int flags); 884 int iocb_bio_iopoll(struct kiocb *kiocb, struct io_comp_batch *iob, 885 unsigned int flags); 886 887 static inline struct request_queue *bdev_get_queue(struct block_device *bdev) 888 { 889 return bdev->bd_queue; /* this is never NULL */ 890 } 891 892 #ifdef CONFIG_BLK_DEV_ZONED 893 894 /* Helper to convert BLK_ZONE_ZONE_XXX to its string format XXX */ 895 const char *blk_zone_cond_str(enum blk_zone_cond zone_cond); 896 897 static inline unsigned int bio_zone_no(struct bio *bio) 898 { 899 return blk_queue_zone_no(bdev_get_queue(bio->bi_bdev), 900 bio->bi_iter.bi_sector); 901 } 902 903 static inline unsigned int bio_zone_is_seq(struct bio *bio) 904 { 905 return blk_queue_zone_is_seq(bdev_get_queue(bio->bi_bdev), 906 bio->bi_iter.bi_sector); 907 } 908 #endif /* CONFIG_BLK_DEV_ZONED */ 909 910 static inline unsigned int blk_queue_get_max_sectors(struct request_queue *q, 911 int op) 912 { 913 if (unlikely(op == REQ_OP_DISCARD || op == REQ_OP_SECURE_ERASE)) 914 return min(q->limits.max_discard_sectors, 915 UINT_MAX >> SECTOR_SHIFT); 916 917 if (unlikely(op == REQ_OP_WRITE_SAME)) 918 return q->limits.max_write_same_sectors; 919 920 if (unlikely(op == REQ_OP_WRITE_ZEROES)) 921 return q->limits.max_write_zeroes_sectors; 922 923 return q->limits.max_sectors; 924 } 925 926 /* 927 * Return maximum size of a request at given offset. Only valid for 928 * file system requests. 929 */ 930 static inline unsigned int blk_max_size_offset(struct request_queue *q, 931 sector_t offset, 932 unsigned int chunk_sectors) 933 { 934 if (!chunk_sectors) { 935 if (q->limits.chunk_sectors) 936 chunk_sectors = q->limits.chunk_sectors; 937 else 938 return q->limits.max_sectors; 939 } 940 941 if (likely(is_power_of_2(chunk_sectors))) 942 chunk_sectors -= offset & (chunk_sectors - 1); 943 else 944 chunk_sectors -= sector_div(offset, chunk_sectors); 945 946 return min(q->limits.max_sectors, chunk_sectors); 947 } 948 949 /* 950 * Access functions for manipulating queue properties 951 */ 952 extern void blk_cleanup_queue(struct request_queue *); 953 void blk_queue_bounce_limit(struct request_queue *q, enum blk_bounce limit); 954 extern void blk_queue_max_hw_sectors(struct request_queue *, unsigned int); 955 extern void blk_queue_chunk_sectors(struct request_queue *, unsigned int); 956 extern void blk_queue_max_segments(struct request_queue *, unsigned short); 957 extern void blk_queue_max_discard_segments(struct request_queue *, 958 unsigned short); 959 extern void blk_queue_max_segment_size(struct request_queue *, unsigned int); 960 extern void blk_queue_max_discard_sectors(struct request_queue *q, 961 unsigned int max_discard_sectors); 962 extern void blk_queue_max_write_same_sectors(struct request_queue *q, 963 unsigned int max_write_same_sectors); 964 extern void blk_queue_max_write_zeroes_sectors(struct request_queue *q, 965 unsigned int max_write_same_sectors); 966 extern void blk_queue_logical_block_size(struct request_queue *, unsigned int); 967 extern void blk_queue_max_zone_append_sectors(struct request_queue *q, 968 unsigned int max_zone_append_sectors); 969 extern void blk_queue_physical_block_size(struct request_queue *, unsigned int); 970 void blk_queue_zone_write_granularity(struct request_queue *q, 971 unsigned int size); 972 extern void blk_queue_alignment_offset(struct request_queue *q, 973 unsigned int alignment); 974 void disk_update_readahead(struct gendisk *disk); 975 extern void blk_limits_io_min(struct queue_limits *limits, unsigned int min); 976 extern void blk_queue_io_min(struct request_queue *q, unsigned int min); 977 extern void blk_limits_io_opt(struct queue_limits *limits, unsigned int opt); 978 extern void blk_queue_io_opt(struct request_queue *q, unsigned int opt); 979 extern void blk_set_queue_depth(struct request_queue *q, unsigned int depth); 980 extern void blk_set_default_limits(struct queue_limits *lim); 981 extern void blk_set_stacking_limits(struct queue_limits *lim); 982 extern int blk_stack_limits(struct queue_limits *t, struct queue_limits *b, 983 sector_t offset); 984 extern void disk_stack_limits(struct gendisk *disk, struct block_device *bdev, 985 sector_t offset); 986 extern void blk_queue_update_dma_pad(struct request_queue *, unsigned int); 987 extern void blk_queue_segment_boundary(struct request_queue *, unsigned long); 988 extern void blk_queue_virt_boundary(struct request_queue *, unsigned long); 989 extern void blk_queue_dma_alignment(struct request_queue *, int); 990 extern void blk_queue_update_dma_alignment(struct request_queue *, int); 991 extern void blk_queue_rq_timeout(struct request_queue *, unsigned int); 992 extern void blk_queue_write_cache(struct request_queue *q, bool enabled, bool fua); 993 994 struct blk_independent_access_ranges * 995 disk_alloc_independent_access_ranges(struct gendisk *disk, int nr_ia_ranges); 996 void disk_set_independent_access_ranges(struct gendisk *disk, 997 struct blk_independent_access_ranges *iars); 998 999 /* 1000 * Elevator features for blk_queue_required_elevator_features: 1001 */ 1002 /* Supports zoned block devices sequential write constraint */ 1003 #define ELEVATOR_F_ZBD_SEQ_WRITE (1U << 0) 1004 /* Supports scheduling on multiple hardware queues */ 1005 #define ELEVATOR_F_MQ_AWARE (1U << 1) 1006 1007 extern void blk_queue_required_elevator_features(struct request_queue *q, 1008 unsigned int features); 1009 extern bool blk_queue_can_use_dma_map_merging(struct request_queue *q, 1010 struct device *dev); 1011 1012 bool __must_check blk_get_queue(struct request_queue *); 1013 extern void blk_put_queue(struct request_queue *); 1014 extern void blk_set_queue_dying(struct request_queue *); 1015 1016 #ifdef CONFIG_BLOCK 1017 /* 1018 * blk_plug permits building a queue of related requests by holding the I/O 1019 * fragments for a short period. This allows merging of sequential requests 1020 * into single larger request. As the requests are moved from a per-task list to 1021 * the device's request_queue in a batch, this results in improved scalability 1022 * as the lock contention for request_queue lock is reduced. 1023 * 1024 * It is ok not to disable preemption when adding the request to the plug list 1025 * or when attempting a merge. For details, please see schedule() where 1026 * blk_flush_plug() is called. 1027 */ 1028 struct blk_plug { 1029 struct request *mq_list; /* blk-mq requests */ 1030 1031 /* if ios_left is > 1, we can batch tag/rq allocations */ 1032 struct request *cached_rq; 1033 unsigned short nr_ios; 1034 1035 unsigned short rq_count; 1036 1037 bool multiple_queues; 1038 bool has_elevator; 1039 bool nowait; 1040 1041 struct list_head cb_list; /* md requires an unplug callback */ 1042 }; 1043 1044 struct blk_plug_cb; 1045 typedef void (*blk_plug_cb_fn)(struct blk_plug_cb *, bool); 1046 struct blk_plug_cb { 1047 struct list_head list; 1048 blk_plug_cb_fn callback; 1049 void *data; 1050 }; 1051 extern struct blk_plug_cb *blk_check_plugged(blk_plug_cb_fn unplug, 1052 void *data, int size); 1053 extern void blk_start_plug(struct blk_plug *); 1054 extern void blk_start_plug_nr_ios(struct blk_plug *, unsigned short); 1055 extern void blk_finish_plug(struct blk_plug *); 1056 1057 void __blk_flush_plug(struct blk_plug *plug, bool from_schedule); 1058 static inline void blk_flush_plug(struct blk_plug *plug, bool async) 1059 { 1060 if (plug) 1061 __blk_flush_plug(plug, async); 1062 } 1063 1064 int blkdev_issue_flush(struct block_device *bdev); 1065 long nr_blockdev_pages(void); 1066 #else /* CONFIG_BLOCK */ 1067 struct blk_plug { 1068 }; 1069 1070 static inline void blk_start_plug_nr_ios(struct blk_plug *plug, 1071 unsigned short nr_ios) 1072 { 1073 } 1074 1075 static inline void blk_start_plug(struct blk_plug *plug) 1076 { 1077 } 1078 1079 static inline void blk_finish_plug(struct blk_plug *plug) 1080 { 1081 } 1082 1083 static inline void blk_flush_plug(struct blk_plug *plug, bool async) 1084 { 1085 } 1086 1087 static inline int blkdev_issue_flush(struct block_device *bdev) 1088 { 1089 return 0; 1090 } 1091 1092 static inline long nr_blockdev_pages(void) 1093 { 1094 return 0; 1095 } 1096 #endif /* CONFIG_BLOCK */ 1097 1098 extern void blk_io_schedule(void); 1099 1100 extern int blkdev_issue_write_same(struct block_device *bdev, sector_t sector, 1101 sector_t nr_sects, gfp_t gfp_mask, struct page *page); 1102 1103 #define BLKDEV_DISCARD_SECURE (1 << 0) /* issue a secure erase */ 1104 1105 extern int blkdev_issue_discard(struct block_device *bdev, sector_t sector, 1106 sector_t nr_sects, gfp_t gfp_mask, unsigned long flags); 1107 extern int __blkdev_issue_discard(struct block_device *bdev, sector_t sector, 1108 sector_t nr_sects, gfp_t gfp_mask, int flags, 1109 struct bio **biop); 1110 1111 #define BLKDEV_ZERO_NOUNMAP (1 << 0) /* do not free blocks */ 1112 #define BLKDEV_ZERO_NOFALLBACK (1 << 1) /* don't write explicit zeroes */ 1113 1114 extern int __blkdev_issue_zeroout(struct block_device *bdev, sector_t sector, 1115 sector_t nr_sects, gfp_t gfp_mask, struct bio **biop, 1116 unsigned flags); 1117 extern int blkdev_issue_zeroout(struct block_device *bdev, sector_t sector, 1118 sector_t nr_sects, gfp_t gfp_mask, unsigned flags); 1119 1120 static inline int sb_issue_discard(struct super_block *sb, sector_t block, 1121 sector_t nr_blocks, gfp_t gfp_mask, unsigned long flags) 1122 { 1123 return blkdev_issue_discard(sb->s_bdev, 1124 block << (sb->s_blocksize_bits - 1125 SECTOR_SHIFT), 1126 nr_blocks << (sb->s_blocksize_bits - 1127 SECTOR_SHIFT), 1128 gfp_mask, flags); 1129 } 1130 static inline int sb_issue_zeroout(struct super_block *sb, sector_t block, 1131 sector_t nr_blocks, gfp_t gfp_mask) 1132 { 1133 return blkdev_issue_zeroout(sb->s_bdev, 1134 block << (sb->s_blocksize_bits - 1135 SECTOR_SHIFT), 1136 nr_blocks << (sb->s_blocksize_bits - 1137 SECTOR_SHIFT), 1138 gfp_mask, 0); 1139 } 1140 1141 static inline bool bdev_is_partition(struct block_device *bdev) 1142 { 1143 return bdev->bd_partno; 1144 } 1145 1146 enum blk_default_limits { 1147 BLK_MAX_SEGMENTS = 128, 1148 BLK_SAFE_MAX_SECTORS = 255, 1149 BLK_DEF_MAX_SECTORS = 2560, 1150 BLK_MAX_SEGMENT_SIZE = 65536, 1151 BLK_SEG_BOUNDARY_MASK = 0xFFFFFFFFUL, 1152 }; 1153 1154 static inline unsigned long queue_segment_boundary(const struct request_queue *q) 1155 { 1156 return q->limits.seg_boundary_mask; 1157 } 1158 1159 static inline unsigned long queue_virt_boundary(const struct request_queue *q) 1160 { 1161 return q->limits.virt_boundary_mask; 1162 } 1163 1164 static inline unsigned int queue_max_sectors(const struct request_queue *q) 1165 { 1166 return q->limits.max_sectors; 1167 } 1168 1169 static inline unsigned int queue_max_bytes(struct request_queue *q) 1170 { 1171 return min_t(unsigned int, queue_max_sectors(q), INT_MAX >> 9) << 9; 1172 } 1173 1174 static inline unsigned int queue_max_hw_sectors(const struct request_queue *q) 1175 { 1176 return q->limits.max_hw_sectors; 1177 } 1178 1179 static inline unsigned short queue_max_segments(const struct request_queue *q) 1180 { 1181 return q->limits.max_segments; 1182 } 1183 1184 static inline unsigned short queue_max_discard_segments(const struct request_queue *q) 1185 { 1186 return q->limits.max_discard_segments; 1187 } 1188 1189 static inline unsigned int queue_max_segment_size(const struct request_queue *q) 1190 { 1191 return q->limits.max_segment_size; 1192 } 1193 1194 static inline unsigned int queue_max_zone_append_sectors(const struct request_queue *q) 1195 { 1196 1197 const struct queue_limits *l = &q->limits; 1198 1199 return min(l->max_zone_append_sectors, l->max_sectors); 1200 } 1201 1202 static inline unsigned queue_logical_block_size(const struct request_queue *q) 1203 { 1204 int retval = 512; 1205 1206 if (q && q->limits.logical_block_size) 1207 retval = q->limits.logical_block_size; 1208 1209 return retval; 1210 } 1211 1212 static inline unsigned int bdev_logical_block_size(struct block_device *bdev) 1213 { 1214 return queue_logical_block_size(bdev_get_queue(bdev)); 1215 } 1216 1217 static inline unsigned int queue_physical_block_size(const struct request_queue *q) 1218 { 1219 return q->limits.physical_block_size; 1220 } 1221 1222 static inline unsigned int bdev_physical_block_size(struct block_device *bdev) 1223 { 1224 return queue_physical_block_size(bdev_get_queue(bdev)); 1225 } 1226 1227 static inline unsigned int queue_io_min(const struct request_queue *q) 1228 { 1229 return q->limits.io_min; 1230 } 1231 1232 static inline int bdev_io_min(struct block_device *bdev) 1233 { 1234 return queue_io_min(bdev_get_queue(bdev)); 1235 } 1236 1237 static inline unsigned int queue_io_opt(const struct request_queue *q) 1238 { 1239 return q->limits.io_opt; 1240 } 1241 1242 static inline int bdev_io_opt(struct block_device *bdev) 1243 { 1244 return queue_io_opt(bdev_get_queue(bdev)); 1245 } 1246 1247 static inline unsigned int 1248 queue_zone_write_granularity(const struct request_queue *q) 1249 { 1250 return q->limits.zone_write_granularity; 1251 } 1252 1253 static inline unsigned int 1254 bdev_zone_write_granularity(struct block_device *bdev) 1255 { 1256 return queue_zone_write_granularity(bdev_get_queue(bdev)); 1257 } 1258 1259 static inline int queue_alignment_offset(const struct request_queue *q) 1260 { 1261 if (q->limits.misaligned) 1262 return -1; 1263 1264 return q->limits.alignment_offset; 1265 } 1266 1267 static inline int queue_limit_alignment_offset(struct queue_limits *lim, sector_t sector) 1268 { 1269 unsigned int granularity = max(lim->physical_block_size, lim->io_min); 1270 unsigned int alignment = sector_div(sector, granularity >> SECTOR_SHIFT) 1271 << SECTOR_SHIFT; 1272 1273 return (granularity + lim->alignment_offset - alignment) % granularity; 1274 } 1275 1276 static inline int bdev_alignment_offset(struct block_device *bdev) 1277 { 1278 struct request_queue *q = bdev_get_queue(bdev); 1279 1280 if (q->limits.misaligned) 1281 return -1; 1282 if (bdev_is_partition(bdev)) 1283 return queue_limit_alignment_offset(&q->limits, 1284 bdev->bd_start_sect); 1285 return q->limits.alignment_offset; 1286 } 1287 1288 static inline int queue_discard_alignment(const struct request_queue *q) 1289 { 1290 if (q->limits.discard_misaligned) 1291 return -1; 1292 1293 return q->limits.discard_alignment; 1294 } 1295 1296 static inline int queue_limit_discard_alignment(struct queue_limits *lim, sector_t sector) 1297 { 1298 unsigned int alignment, granularity, offset; 1299 1300 if (!lim->max_discard_sectors) 1301 return 0; 1302 1303 /* Why are these in bytes, not sectors? */ 1304 alignment = lim->discard_alignment >> SECTOR_SHIFT; 1305 granularity = lim->discard_granularity >> SECTOR_SHIFT; 1306 if (!granularity) 1307 return 0; 1308 1309 /* Offset of the partition start in 'granularity' sectors */ 1310 offset = sector_div(sector, granularity); 1311 1312 /* And why do we do this modulus *again* in blkdev_issue_discard()? */ 1313 offset = (granularity + alignment - offset) % granularity; 1314 1315 /* Turn it back into bytes, gaah */ 1316 return offset << SECTOR_SHIFT; 1317 } 1318 1319 static inline int bdev_discard_alignment(struct block_device *bdev) 1320 { 1321 struct request_queue *q = bdev_get_queue(bdev); 1322 1323 if (bdev_is_partition(bdev)) 1324 return queue_limit_discard_alignment(&q->limits, 1325 bdev->bd_start_sect); 1326 return q->limits.discard_alignment; 1327 } 1328 1329 static inline unsigned int bdev_write_same(struct block_device *bdev) 1330 { 1331 struct request_queue *q = bdev_get_queue(bdev); 1332 1333 if (q) 1334 return q->limits.max_write_same_sectors; 1335 1336 return 0; 1337 } 1338 1339 static inline unsigned int bdev_write_zeroes_sectors(struct block_device *bdev) 1340 { 1341 struct request_queue *q = bdev_get_queue(bdev); 1342 1343 if (q) 1344 return q->limits.max_write_zeroes_sectors; 1345 1346 return 0; 1347 } 1348 1349 static inline enum blk_zoned_model bdev_zoned_model(struct block_device *bdev) 1350 { 1351 struct request_queue *q = bdev_get_queue(bdev); 1352 1353 if (q) 1354 return blk_queue_zoned_model(q); 1355 1356 return BLK_ZONED_NONE; 1357 } 1358 1359 static inline bool bdev_is_zoned(struct block_device *bdev) 1360 { 1361 struct request_queue *q = bdev_get_queue(bdev); 1362 1363 if (q) 1364 return blk_queue_is_zoned(q); 1365 1366 return false; 1367 } 1368 1369 static inline sector_t bdev_zone_sectors(struct block_device *bdev) 1370 { 1371 struct request_queue *q = bdev_get_queue(bdev); 1372 1373 if (q) 1374 return blk_queue_zone_sectors(q); 1375 return 0; 1376 } 1377 1378 static inline unsigned int bdev_max_open_zones(struct block_device *bdev) 1379 { 1380 struct request_queue *q = bdev_get_queue(bdev); 1381 1382 if (q) 1383 return queue_max_open_zones(q); 1384 return 0; 1385 } 1386 1387 static inline unsigned int bdev_max_active_zones(struct block_device *bdev) 1388 { 1389 struct request_queue *q = bdev_get_queue(bdev); 1390 1391 if (q) 1392 return queue_max_active_zones(q); 1393 return 0; 1394 } 1395 1396 static inline int queue_dma_alignment(const struct request_queue *q) 1397 { 1398 return q ? q->dma_alignment : 511; 1399 } 1400 1401 static inline int blk_rq_aligned(struct request_queue *q, unsigned long addr, 1402 unsigned int len) 1403 { 1404 unsigned int alignment = queue_dma_alignment(q) | q->dma_pad_mask; 1405 return !(addr & alignment) && !(len & alignment); 1406 } 1407 1408 /* assumes size > 256 */ 1409 static inline unsigned int blksize_bits(unsigned int size) 1410 { 1411 unsigned int bits = 8; 1412 do { 1413 bits++; 1414 size >>= 1; 1415 } while (size > 256); 1416 return bits; 1417 } 1418 1419 static inline unsigned int block_size(struct block_device *bdev) 1420 { 1421 return 1 << bdev->bd_inode->i_blkbits; 1422 } 1423 1424 int kblockd_schedule_work(struct work_struct *work); 1425 int kblockd_mod_delayed_work_on(int cpu, struct delayed_work *dwork, unsigned long delay); 1426 1427 #define MODULE_ALIAS_BLOCKDEV(major,minor) \ 1428 MODULE_ALIAS("block-major-" __stringify(major) "-" __stringify(minor)) 1429 #define MODULE_ALIAS_BLOCKDEV_MAJOR(major) \ 1430 MODULE_ALIAS("block-major-" __stringify(major) "-*") 1431 1432 #ifdef CONFIG_BLK_INLINE_ENCRYPTION 1433 1434 bool blk_crypto_register(struct blk_crypto_profile *profile, 1435 struct request_queue *q); 1436 1437 #else /* CONFIG_BLK_INLINE_ENCRYPTION */ 1438 1439 static inline bool blk_crypto_register(struct blk_crypto_profile *profile, 1440 struct request_queue *q) 1441 { 1442 return true; 1443 } 1444 1445 #endif /* CONFIG_BLK_INLINE_ENCRYPTION */ 1446 1447 enum blk_unique_id { 1448 /* these match the Designator Types specified in SPC */ 1449 BLK_UID_T10 = 1, 1450 BLK_UID_EUI64 = 2, 1451 BLK_UID_NAA = 3, 1452 }; 1453 1454 #define NFL4_UFLG_MASK 0x0000003F 1455 1456 struct block_device_operations { 1457 void (*submit_bio)(struct bio *bio); 1458 int (*open) (struct block_device *, fmode_t); 1459 void (*release) (struct gendisk *, fmode_t); 1460 int (*rw_page)(struct block_device *, sector_t, struct page *, unsigned int); 1461 int (*ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); 1462 int (*compat_ioctl) (struct block_device *, fmode_t, unsigned, unsigned long); 1463 unsigned int (*check_events) (struct gendisk *disk, 1464 unsigned int clearing); 1465 void (*unlock_native_capacity) (struct gendisk *); 1466 int (*getgeo)(struct block_device *, struct hd_geometry *); 1467 int (*set_read_only)(struct block_device *bdev, bool ro); 1468 void (*free_disk)(struct gendisk *disk); 1469 /* this callback is with swap_lock and sometimes page table lock held */ 1470 void (*swap_slot_free_notify) (struct block_device *, unsigned long); 1471 int (*report_zones)(struct gendisk *, sector_t sector, 1472 unsigned int nr_zones, report_zones_cb cb, void *data); 1473 char *(*devnode)(struct gendisk *disk, umode_t *mode); 1474 /* returns the length of the identifier or a negative errno: */ 1475 int (*get_unique_id)(struct gendisk *disk, u8 id[16], 1476 enum blk_unique_id id_type); 1477 struct module *owner; 1478 const struct pr_ops *pr_ops; 1479 1480 /* 1481 * Special callback for probing GPT entry at a given sector. 1482 * Needed by Android devices, used by GPT scanner and MMC blk 1483 * driver. 1484 */ 1485 int (*alternative_gpt_sector)(struct gendisk *disk, sector_t *sector); 1486 }; 1487 1488 #ifdef CONFIG_COMPAT 1489 extern int blkdev_compat_ptr_ioctl(struct block_device *, fmode_t, 1490 unsigned int, unsigned long); 1491 #else 1492 #define blkdev_compat_ptr_ioctl NULL 1493 #endif 1494 1495 extern int bdev_read_page(struct block_device *, sector_t, struct page *); 1496 extern int bdev_write_page(struct block_device *, sector_t, struct page *, 1497 struct writeback_control *); 1498 1499 static inline void blk_wake_io_task(struct task_struct *waiter) 1500 { 1501 /* 1502 * If we're polling, the task itself is doing the completions. For 1503 * that case, we don't need to signal a wakeup, it's enough to just 1504 * mark us as RUNNING. 1505 */ 1506 if (waiter == current) 1507 __set_current_state(TASK_RUNNING); 1508 else 1509 wake_up_process(waiter); 1510 } 1511 1512 unsigned long disk_start_io_acct(struct gendisk *disk, unsigned int sectors, 1513 unsigned int op); 1514 void disk_end_io_acct(struct gendisk *disk, unsigned int op, 1515 unsigned long start_time); 1516 1517 void bio_start_io_acct_time(struct bio *bio, unsigned long start_time); 1518 unsigned long bio_start_io_acct(struct bio *bio); 1519 void bio_end_io_acct_remapped(struct bio *bio, unsigned long start_time, 1520 struct block_device *orig_bdev); 1521 1522 /** 1523 * bio_end_io_acct - end I/O accounting for bio based drivers 1524 * @bio: bio to end account for 1525 * @start_time: start time returned by bio_start_io_acct() 1526 */ 1527 static inline void bio_end_io_acct(struct bio *bio, unsigned long start_time) 1528 { 1529 return bio_end_io_acct_remapped(bio, start_time, bio->bi_bdev); 1530 } 1531 1532 int bdev_read_only(struct block_device *bdev); 1533 int set_blocksize(struct block_device *bdev, int size); 1534 1535 const char *bdevname(struct block_device *bdev, char *buffer); 1536 int lookup_bdev(const char *pathname, dev_t *dev); 1537 1538 void blkdev_show(struct seq_file *seqf, off_t offset); 1539 1540 #define BDEVNAME_SIZE 32 /* Largest string for a blockdev identifier */ 1541 #define BDEVT_SIZE 10 /* Largest string for MAJ:MIN for blkdev */ 1542 #ifdef CONFIG_BLOCK 1543 #define BLKDEV_MAJOR_MAX 512 1544 #else 1545 #define BLKDEV_MAJOR_MAX 0 1546 #endif 1547 1548 struct block_device *blkdev_get_by_path(const char *path, fmode_t mode, 1549 void *holder); 1550 struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode, void *holder); 1551 int bd_prepare_to_claim(struct block_device *bdev, void *holder); 1552 void bd_abort_claiming(struct block_device *bdev, void *holder); 1553 void blkdev_put(struct block_device *bdev, fmode_t mode); 1554 1555 /* just for blk-cgroup, don't use elsewhere */ 1556 struct block_device *blkdev_get_no_open(dev_t dev); 1557 void blkdev_put_no_open(struct block_device *bdev); 1558 1559 struct block_device *bdev_alloc(struct gendisk *disk, u8 partno); 1560 void bdev_add(struct block_device *bdev, dev_t dev); 1561 struct block_device *I_BDEV(struct inode *inode); 1562 int truncate_bdev_range(struct block_device *bdev, fmode_t mode, loff_t lstart, 1563 loff_t lend); 1564 1565 #ifdef CONFIG_BLOCK 1566 void invalidate_bdev(struct block_device *bdev); 1567 int sync_blockdev(struct block_device *bdev); 1568 int sync_blockdev_nowait(struct block_device *bdev); 1569 void sync_bdevs(bool wait); 1570 void printk_all_partitions(void); 1571 #else 1572 static inline void invalidate_bdev(struct block_device *bdev) 1573 { 1574 } 1575 static inline int sync_blockdev(struct block_device *bdev) 1576 { 1577 return 0; 1578 } 1579 static inline int sync_blockdev_nowait(struct block_device *bdev) 1580 { 1581 return 0; 1582 } 1583 static inline void sync_bdevs(bool wait) 1584 { 1585 } 1586 static inline void printk_all_partitions(void) 1587 { 1588 } 1589 #endif /* CONFIG_BLOCK */ 1590 1591 int fsync_bdev(struct block_device *bdev); 1592 1593 int freeze_bdev(struct block_device *bdev); 1594 int thaw_bdev(struct block_device *bdev); 1595 1596 struct io_comp_batch { 1597 struct request *req_list; 1598 bool need_ts; 1599 void (*complete)(struct io_comp_batch *); 1600 }; 1601 1602 #define DEFINE_IO_COMP_BATCH(name) struct io_comp_batch name = { } 1603 1604 #endif /* _LINUX_BLKDEV_H */ 1605