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