1 /* 2 * blkfront.c 3 * 4 * XenLinux virtual block device driver. 5 * 6 * Copyright (c) 2003-2004, Keir Fraser & Steve Hand 7 * Modifications by Mark A. Williamson are (c) Intel Research Cambridge 8 * Copyright (c) 2004, Christian Limpach 9 * Copyright (c) 2004, Andrew Warfield 10 * Copyright (c) 2005, Christopher Clark 11 * Copyright (c) 2005, XenSource Ltd 12 * 13 * This program is free software; you can redistribute it and/or 14 * modify it under the terms of the GNU General Public License version 2 15 * as published by the Free Software Foundation; or, when distributed 16 * separately from the Linux kernel or incorporated into other 17 * software packages, subject to the following license: 18 * 19 * Permission is hereby granted, free of charge, to any person obtaining a copy 20 * of this source file (the "Software"), to deal in the Software without 21 * restriction, including without limitation the rights to use, copy, modify, 22 * merge, publish, distribute, sublicense, and/or sell copies of the Software, 23 * and to permit persons to whom the Software is furnished to do so, subject to 24 * the following conditions: 25 * 26 * The above copyright notice and this permission notice shall be included in 27 * all copies or substantial portions of the Software. 28 * 29 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 30 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 31 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 32 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 33 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING 34 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS 35 * IN THE SOFTWARE. 36 */ 37 38 #include <linux/interrupt.h> 39 #include <linux/blkdev.h> 40 #include <linux/blk-mq.h> 41 #include <linux/hdreg.h> 42 #include <linux/cdrom.h> 43 #include <linux/module.h> 44 #include <linux/slab.h> 45 #include <linux/major.h> 46 #include <linux/mutex.h> 47 #include <linux/scatterlist.h> 48 #include <linux/bitmap.h> 49 #include <linux/list.h> 50 #include <linux/workqueue.h> 51 #include <linux/sched/mm.h> 52 53 #include <xen/xen.h> 54 #include <xen/xenbus.h> 55 #include <xen/grant_table.h> 56 #include <xen/events.h> 57 #include <xen/page.h> 58 #include <xen/platform_pci.h> 59 60 #include <xen/interface/grant_table.h> 61 #include <xen/interface/io/blkif.h> 62 #include <xen/interface/io/protocols.h> 63 64 #include <asm/xen/hypervisor.h> 65 66 /* 67 * The minimal size of segment supported by the block framework is PAGE_SIZE. 68 * When Linux is using a different page size than Xen, it may not be possible 69 * to put all the data in a single segment. 70 * This can happen when the backend doesn't support indirect descriptor and 71 * therefore the maximum amount of data that a request can carry is 72 * BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE = 44KB 73 * 74 * Note that we only support one extra request. So the Linux page size 75 * should be <= ( 2 * BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE) = 76 * 88KB. 77 */ 78 #define HAS_EXTRA_REQ (BLKIF_MAX_SEGMENTS_PER_REQUEST < XEN_PFN_PER_PAGE) 79 80 enum blkif_state { 81 BLKIF_STATE_DISCONNECTED, 82 BLKIF_STATE_CONNECTED, 83 BLKIF_STATE_SUSPENDED, 84 BLKIF_STATE_ERROR, 85 }; 86 87 struct grant { 88 grant_ref_t gref; 89 struct page *page; 90 struct list_head node; 91 }; 92 93 enum blk_req_status { 94 REQ_PROCESSING, 95 REQ_WAITING, 96 REQ_DONE, 97 REQ_ERROR, 98 REQ_EOPNOTSUPP, 99 }; 100 101 struct blk_shadow { 102 struct blkif_request req; 103 struct request *request; 104 struct grant **grants_used; 105 struct grant **indirect_grants; 106 struct scatterlist *sg; 107 unsigned int num_sg; 108 enum blk_req_status status; 109 110 #define NO_ASSOCIATED_ID ~0UL 111 /* 112 * Id of the sibling if we ever need 2 requests when handling a 113 * block I/O request 114 */ 115 unsigned long associated_id; 116 }; 117 118 struct blkif_req { 119 blk_status_t error; 120 }; 121 122 static inline struct blkif_req *blkif_req(struct request *rq) 123 { 124 return blk_mq_rq_to_pdu(rq); 125 } 126 127 static DEFINE_MUTEX(blkfront_mutex); 128 static const struct block_device_operations xlvbd_block_fops; 129 static struct delayed_work blkfront_work; 130 static LIST_HEAD(info_list); 131 132 /* 133 * Maximum number of segments in indirect requests, the actual value used by 134 * the frontend driver is the minimum of this value and the value provided 135 * by the backend driver. 136 */ 137 138 static unsigned int xen_blkif_max_segments = 32; 139 module_param_named(max_indirect_segments, xen_blkif_max_segments, uint, 0444); 140 MODULE_PARM_DESC(max_indirect_segments, 141 "Maximum amount of segments in indirect requests (default is 32)"); 142 143 static unsigned int xen_blkif_max_queues = 4; 144 module_param_named(max_queues, xen_blkif_max_queues, uint, 0444); 145 MODULE_PARM_DESC(max_queues, "Maximum number of hardware queues/rings used per virtual disk"); 146 147 /* 148 * Maximum order of pages to be used for the shared ring between front and 149 * backend, 4KB page granularity is used. 150 */ 151 static unsigned int xen_blkif_max_ring_order; 152 module_param_named(max_ring_page_order, xen_blkif_max_ring_order, int, 0444); 153 MODULE_PARM_DESC(max_ring_page_order, "Maximum order of pages to be used for the shared ring"); 154 155 #define BLK_RING_SIZE(info) \ 156 __CONST_RING_SIZE(blkif, XEN_PAGE_SIZE * (info)->nr_ring_pages) 157 158 /* 159 * ring-ref%u i=(-1UL) would take 11 characters + 'ring-ref' is 8, so 19 160 * characters are enough. Define to 20 to keep consistent with backend. 161 */ 162 #define RINGREF_NAME_LEN (20) 163 /* 164 * queue-%u would take 7 + 10(UINT_MAX) = 17 characters. 165 */ 166 #define QUEUE_NAME_LEN (17) 167 168 /* 169 * Per-ring info. 170 * Every blkfront device can associate with one or more blkfront_ring_info, 171 * depending on how many hardware queues/rings to be used. 172 */ 173 struct blkfront_ring_info { 174 /* Lock to protect data in every ring buffer. */ 175 spinlock_t ring_lock; 176 struct blkif_front_ring ring; 177 unsigned int ring_ref[XENBUS_MAX_RING_GRANTS]; 178 unsigned int evtchn, irq; 179 struct work_struct work; 180 struct gnttab_free_callback callback; 181 struct list_head indirect_pages; 182 struct list_head grants; 183 unsigned int persistent_gnts_c; 184 unsigned long shadow_free; 185 struct blkfront_info *dev_info; 186 struct blk_shadow shadow[]; 187 }; 188 189 /* 190 * We have one of these per vbd, whether ide, scsi or 'other'. They 191 * hang in private_data off the gendisk structure. We may end up 192 * putting all kinds of interesting stuff here :-) 193 */ 194 struct blkfront_info 195 { 196 struct mutex mutex; 197 struct xenbus_device *xbdev; 198 struct gendisk *gd; 199 u16 sector_size; 200 unsigned int physical_sector_size; 201 unsigned long vdisk_info; 202 int vdevice; 203 blkif_vdev_t handle; 204 enum blkif_state connected; 205 /* Number of pages per ring buffer. */ 206 unsigned int nr_ring_pages; 207 struct request_queue *rq; 208 unsigned int feature_flush:1; 209 unsigned int feature_fua:1; 210 unsigned int feature_discard:1; 211 unsigned int feature_secdiscard:1; 212 unsigned int feature_persistent:1; 213 unsigned int discard_granularity; 214 unsigned int discard_alignment; 215 /* Number of 4KB segments handled */ 216 unsigned int max_indirect_segments; 217 int is_ready; 218 struct blk_mq_tag_set tag_set; 219 struct blkfront_ring_info *rinfo; 220 unsigned int nr_rings; 221 unsigned int rinfo_size; 222 /* Save uncomplete reqs and bios for migration. */ 223 struct list_head requests; 224 struct bio_list bio_list; 225 struct list_head info_list; 226 }; 227 228 static unsigned int nr_minors; 229 static unsigned long *minors; 230 static DEFINE_SPINLOCK(minor_lock); 231 232 #define GRANT_INVALID_REF 0 233 234 #define PARTS_PER_DISK 16 235 #define PARTS_PER_EXT_DISK 256 236 237 #define BLKIF_MAJOR(dev) ((dev)>>8) 238 #define BLKIF_MINOR(dev) ((dev) & 0xff) 239 240 #define EXT_SHIFT 28 241 #define EXTENDED (1<<EXT_SHIFT) 242 #define VDEV_IS_EXTENDED(dev) ((dev)&(EXTENDED)) 243 #define BLKIF_MINOR_EXT(dev) ((dev)&(~EXTENDED)) 244 #define EMULATED_HD_DISK_MINOR_OFFSET (0) 245 #define EMULATED_HD_DISK_NAME_OFFSET (EMULATED_HD_DISK_MINOR_OFFSET / 256) 246 #define EMULATED_SD_DISK_MINOR_OFFSET (0) 247 #define EMULATED_SD_DISK_NAME_OFFSET (EMULATED_SD_DISK_MINOR_OFFSET / 256) 248 249 #define DEV_NAME "xvd" /* name in /dev */ 250 251 /* 252 * Grants are always the same size as a Xen page (i.e 4KB). 253 * A physical segment is always the same size as a Linux page. 254 * Number of grants per physical segment 255 */ 256 #define GRANTS_PER_PSEG (PAGE_SIZE / XEN_PAGE_SIZE) 257 258 #define GRANTS_PER_INDIRECT_FRAME \ 259 (XEN_PAGE_SIZE / sizeof(struct blkif_request_segment)) 260 261 #define INDIRECT_GREFS(_grants) \ 262 DIV_ROUND_UP(_grants, GRANTS_PER_INDIRECT_FRAME) 263 264 static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo); 265 static void blkfront_gather_backend_features(struct blkfront_info *info); 266 static int negotiate_mq(struct blkfront_info *info); 267 268 #define for_each_rinfo(info, ptr, idx) \ 269 for ((ptr) = (info)->rinfo, (idx) = 0; \ 270 (idx) < (info)->nr_rings; \ 271 (idx)++, (ptr) = (void *)(ptr) + (info)->rinfo_size) 272 273 static inline struct blkfront_ring_info * 274 get_rinfo(const struct blkfront_info *info, unsigned int i) 275 { 276 BUG_ON(i >= info->nr_rings); 277 return (void *)info->rinfo + i * info->rinfo_size; 278 } 279 280 static int get_id_from_freelist(struct blkfront_ring_info *rinfo) 281 { 282 unsigned long free = rinfo->shadow_free; 283 284 BUG_ON(free >= BLK_RING_SIZE(rinfo->dev_info)); 285 rinfo->shadow_free = rinfo->shadow[free].req.u.rw.id; 286 rinfo->shadow[free].req.u.rw.id = 0x0fffffee; /* debug */ 287 return free; 288 } 289 290 static int add_id_to_freelist(struct blkfront_ring_info *rinfo, 291 unsigned long id) 292 { 293 if (rinfo->shadow[id].req.u.rw.id != id) 294 return -EINVAL; 295 if (rinfo->shadow[id].request == NULL) 296 return -EINVAL; 297 rinfo->shadow[id].req.u.rw.id = rinfo->shadow_free; 298 rinfo->shadow[id].request = NULL; 299 rinfo->shadow_free = id; 300 return 0; 301 } 302 303 static int fill_grant_buffer(struct blkfront_ring_info *rinfo, int num) 304 { 305 struct blkfront_info *info = rinfo->dev_info; 306 struct page *granted_page; 307 struct grant *gnt_list_entry, *n; 308 int i = 0; 309 310 while (i < num) { 311 gnt_list_entry = kzalloc(sizeof(struct grant), GFP_NOIO); 312 if (!gnt_list_entry) 313 goto out_of_memory; 314 315 if (info->feature_persistent) { 316 granted_page = alloc_page(GFP_NOIO); 317 if (!granted_page) { 318 kfree(gnt_list_entry); 319 goto out_of_memory; 320 } 321 gnt_list_entry->page = granted_page; 322 } 323 324 gnt_list_entry->gref = GRANT_INVALID_REF; 325 list_add(&gnt_list_entry->node, &rinfo->grants); 326 i++; 327 } 328 329 return 0; 330 331 out_of_memory: 332 list_for_each_entry_safe(gnt_list_entry, n, 333 &rinfo->grants, node) { 334 list_del(&gnt_list_entry->node); 335 if (info->feature_persistent) 336 __free_page(gnt_list_entry->page); 337 kfree(gnt_list_entry); 338 i--; 339 } 340 BUG_ON(i != 0); 341 return -ENOMEM; 342 } 343 344 static struct grant *get_free_grant(struct blkfront_ring_info *rinfo) 345 { 346 struct grant *gnt_list_entry; 347 348 BUG_ON(list_empty(&rinfo->grants)); 349 gnt_list_entry = list_first_entry(&rinfo->grants, struct grant, 350 node); 351 list_del(&gnt_list_entry->node); 352 353 if (gnt_list_entry->gref != GRANT_INVALID_REF) 354 rinfo->persistent_gnts_c--; 355 356 return gnt_list_entry; 357 } 358 359 static inline void grant_foreign_access(const struct grant *gnt_list_entry, 360 const struct blkfront_info *info) 361 { 362 gnttab_page_grant_foreign_access_ref_one(gnt_list_entry->gref, 363 info->xbdev->otherend_id, 364 gnt_list_entry->page, 365 0); 366 } 367 368 static struct grant *get_grant(grant_ref_t *gref_head, 369 unsigned long gfn, 370 struct blkfront_ring_info *rinfo) 371 { 372 struct grant *gnt_list_entry = get_free_grant(rinfo); 373 struct blkfront_info *info = rinfo->dev_info; 374 375 if (gnt_list_entry->gref != GRANT_INVALID_REF) 376 return gnt_list_entry; 377 378 /* Assign a gref to this page */ 379 gnt_list_entry->gref = gnttab_claim_grant_reference(gref_head); 380 BUG_ON(gnt_list_entry->gref == -ENOSPC); 381 if (info->feature_persistent) 382 grant_foreign_access(gnt_list_entry, info); 383 else { 384 /* Grant access to the GFN passed by the caller */ 385 gnttab_grant_foreign_access_ref(gnt_list_entry->gref, 386 info->xbdev->otherend_id, 387 gfn, 0); 388 } 389 390 return gnt_list_entry; 391 } 392 393 static struct grant *get_indirect_grant(grant_ref_t *gref_head, 394 struct blkfront_ring_info *rinfo) 395 { 396 struct grant *gnt_list_entry = get_free_grant(rinfo); 397 struct blkfront_info *info = rinfo->dev_info; 398 399 if (gnt_list_entry->gref != GRANT_INVALID_REF) 400 return gnt_list_entry; 401 402 /* Assign a gref to this page */ 403 gnt_list_entry->gref = gnttab_claim_grant_reference(gref_head); 404 BUG_ON(gnt_list_entry->gref == -ENOSPC); 405 if (!info->feature_persistent) { 406 struct page *indirect_page; 407 408 /* Fetch a pre-allocated page to use for indirect grefs */ 409 BUG_ON(list_empty(&rinfo->indirect_pages)); 410 indirect_page = list_first_entry(&rinfo->indirect_pages, 411 struct page, lru); 412 list_del(&indirect_page->lru); 413 gnt_list_entry->page = indirect_page; 414 } 415 grant_foreign_access(gnt_list_entry, info); 416 417 return gnt_list_entry; 418 } 419 420 static const char *op_name(int op) 421 { 422 static const char *const names[] = { 423 [BLKIF_OP_READ] = "read", 424 [BLKIF_OP_WRITE] = "write", 425 [BLKIF_OP_WRITE_BARRIER] = "barrier", 426 [BLKIF_OP_FLUSH_DISKCACHE] = "flush", 427 [BLKIF_OP_DISCARD] = "discard" }; 428 429 if (op < 0 || op >= ARRAY_SIZE(names)) 430 return "unknown"; 431 432 if (!names[op]) 433 return "reserved"; 434 435 return names[op]; 436 } 437 static int xlbd_reserve_minors(unsigned int minor, unsigned int nr) 438 { 439 unsigned int end = minor + nr; 440 int rc; 441 442 if (end > nr_minors) { 443 unsigned long *bitmap, *old; 444 445 bitmap = kcalloc(BITS_TO_LONGS(end), sizeof(*bitmap), 446 GFP_KERNEL); 447 if (bitmap == NULL) 448 return -ENOMEM; 449 450 spin_lock(&minor_lock); 451 if (end > nr_minors) { 452 old = minors; 453 memcpy(bitmap, minors, 454 BITS_TO_LONGS(nr_minors) * sizeof(*bitmap)); 455 minors = bitmap; 456 nr_minors = BITS_TO_LONGS(end) * BITS_PER_LONG; 457 } else 458 old = bitmap; 459 spin_unlock(&minor_lock); 460 kfree(old); 461 } 462 463 spin_lock(&minor_lock); 464 if (find_next_bit(minors, end, minor) >= end) { 465 bitmap_set(minors, minor, nr); 466 rc = 0; 467 } else 468 rc = -EBUSY; 469 spin_unlock(&minor_lock); 470 471 return rc; 472 } 473 474 static void xlbd_release_minors(unsigned int minor, unsigned int nr) 475 { 476 unsigned int end = minor + nr; 477 478 BUG_ON(end > nr_minors); 479 spin_lock(&minor_lock); 480 bitmap_clear(minors, minor, nr); 481 spin_unlock(&minor_lock); 482 } 483 484 static void blkif_restart_queue_callback(void *arg) 485 { 486 struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)arg; 487 schedule_work(&rinfo->work); 488 } 489 490 static int blkif_getgeo(struct block_device *bd, struct hd_geometry *hg) 491 { 492 /* We don't have real geometry info, but let's at least return 493 values consistent with the size of the device */ 494 sector_t nsect = get_capacity(bd->bd_disk); 495 sector_t cylinders = nsect; 496 497 hg->heads = 0xff; 498 hg->sectors = 0x3f; 499 sector_div(cylinders, hg->heads * hg->sectors); 500 hg->cylinders = cylinders; 501 if ((sector_t)(hg->cylinders + 1) * hg->heads * hg->sectors < nsect) 502 hg->cylinders = 0xffff; 503 return 0; 504 } 505 506 static int blkif_ioctl(struct block_device *bdev, fmode_t mode, 507 unsigned command, unsigned long argument) 508 { 509 struct blkfront_info *info = bdev->bd_disk->private_data; 510 int i; 511 512 switch (command) { 513 case CDROMMULTISESSION: 514 for (i = 0; i < sizeof(struct cdrom_multisession); i++) 515 if (put_user(0, (char __user *)(argument + i))) 516 return -EFAULT; 517 return 0; 518 case CDROM_GET_CAPABILITY: 519 if (!(info->vdisk_info & VDISK_CDROM)) 520 return -EINVAL; 521 return 0; 522 default: 523 return -EINVAL; 524 } 525 } 526 527 static unsigned long blkif_ring_get_request(struct blkfront_ring_info *rinfo, 528 struct request *req, 529 struct blkif_request **ring_req) 530 { 531 unsigned long id; 532 533 *ring_req = RING_GET_REQUEST(&rinfo->ring, rinfo->ring.req_prod_pvt); 534 rinfo->ring.req_prod_pvt++; 535 536 id = get_id_from_freelist(rinfo); 537 rinfo->shadow[id].request = req; 538 rinfo->shadow[id].status = REQ_PROCESSING; 539 rinfo->shadow[id].associated_id = NO_ASSOCIATED_ID; 540 541 rinfo->shadow[id].req.u.rw.id = id; 542 543 return id; 544 } 545 546 static int blkif_queue_discard_req(struct request *req, struct blkfront_ring_info *rinfo) 547 { 548 struct blkfront_info *info = rinfo->dev_info; 549 struct blkif_request *ring_req, *final_ring_req; 550 unsigned long id; 551 552 /* Fill out a communications ring structure. */ 553 id = blkif_ring_get_request(rinfo, req, &final_ring_req); 554 ring_req = &rinfo->shadow[id].req; 555 556 ring_req->operation = BLKIF_OP_DISCARD; 557 ring_req->u.discard.nr_sectors = blk_rq_sectors(req); 558 ring_req->u.discard.id = id; 559 ring_req->u.discard.sector_number = (blkif_sector_t)blk_rq_pos(req); 560 if (req_op(req) == REQ_OP_SECURE_ERASE && info->feature_secdiscard) 561 ring_req->u.discard.flag = BLKIF_DISCARD_SECURE; 562 else 563 ring_req->u.discard.flag = 0; 564 565 /* Copy the request to the ring page. */ 566 *final_ring_req = *ring_req; 567 rinfo->shadow[id].status = REQ_WAITING; 568 569 return 0; 570 } 571 572 struct setup_rw_req { 573 unsigned int grant_idx; 574 struct blkif_request_segment *segments; 575 struct blkfront_ring_info *rinfo; 576 struct blkif_request *ring_req; 577 grant_ref_t gref_head; 578 unsigned int id; 579 /* Only used when persistent grant is used and it's a read request */ 580 bool need_copy; 581 unsigned int bvec_off; 582 char *bvec_data; 583 584 bool require_extra_req; 585 struct blkif_request *extra_ring_req; 586 }; 587 588 static void blkif_setup_rw_req_grant(unsigned long gfn, unsigned int offset, 589 unsigned int len, void *data) 590 { 591 struct setup_rw_req *setup = data; 592 int n, ref; 593 struct grant *gnt_list_entry; 594 unsigned int fsect, lsect; 595 /* Convenient aliases */ 596 unsigned int grant_idx = setup->grant_idx; 597 struct blkif_request *ring_req = setup->ring_req; 598 struct blkfront_ring_info *rinfo = setup->rinfo; 599 /* 600 * We always use the shadow of the first request to store the list 601 * of grant associated to the block I/O request. This made the 602 * completion more easy to handle even if the block I/O request is 603 * split. 604 */ 605 struct blk_shadow *shadow = &rinfo->shadow[setup->id]; 606 607 if (unlikely(setup->require_extra_req && 608 grant_idx >= BLKIF_MAX_SEGMENTS_PER_REQUEST)) { 609 /* 610 * We are using the second request, setup grant_idx 611 * to be the index of the segment array. 612 */ 613 grant_idx -= BLKIF_MAX_SEGMENTS_PER_REQUEST; 614 ring_req = setup->extra_ring_req; 615 } 616 617 if ((ring_req->operation == BLKIF_OP_INDIRECT) && 618 (grant_idx % GRANTS_PER_INDIRECT_FRAME == 0)) { 619 if (setup->segments) 620 kunmap_atomic(setup->segments); 621 622 n = grant_idx / GRANTS_PER_INDIRECT_FRAME; 623 gnt_list_entry = get_indirect_grant(&setup->gref_head, rinfo); 624 shadow->indirect_grants[n] = gnt_list_entry; 625 setup->segments = kmap_atomic(gnt_list_entry->page); 626 ring_req->u.indirect.indirect_grefs[n] = gnt_list_entry->gref; 627 } 628 629 gnt_list_entry = get_grant(&setup->gref_head, gfn, rinfo); 630 ref = gnt_list_entry->gref; 631 /* 632 * All the grants are stored in the shadow of the first 633 * request. Therefore we have to use the global index. 634 */ 635 shadow->grants_used[setup->grant_idx] = gnt_list_entry; 636 637 if (setup->need_copy) { 638 void *shared_data; 639 640 shared_data = kmap_atomic(gnt_list_entry->page); 641 /* 642 * this does not wipe data stored outside the 643 * range sg->offset..sg->offset+sg->length. 644 * Therefore, blkback *could* see data from 645 * previous requests. This is OK as long as 646 * persistent grants are shared with just one 647 * domain. It may need refactoring if this 648 * changes 649 */ 650 memcpy(shared_data + offset, 651 setup->bvec_data + setup->bvec_off, 652 len); 653 654 kunmap_atomic(shared_data); 655 setup->bvec_off += len; 656 } 657 658 fsect = offset >> 9; 659 lsect = fsect + (len >> 9) - 1; 660 if (ring_req->operation != BLKIF_OP_INDIRECT) { 661 ring_req->u.rw.seg[grant_idx] = 662 (struct blkif_request_segment) { 663 .gref = ref, 664 .first_sect = fsect, 665 .last_sect = lsect }; 666 } else { 667 setup->segments[grant_idx % GRANTS_PER_INDIRECT_FRAME] = 668 (struct blkif_request_segment) { 669 .gref = ref, 670 .first_sect = fsect, 671 .last_sect = lsect }; 672 } 673 674 (setup->grant_idx)++; 675 } 676 677 static void blkif_setup_extra_req(struct blkif_request *first, 678 struct blkif_request *second) 679 { 680 uint16_t nr_segments = first->u.rw.nr_segments; 681 682 /* 683 * The second request is only present when the first request uses 684 * all its segments. It's always the continuity of the first one. 685 */ 686 first->u.rw.nr_segments = BLKIF_MAX_SEGMENTS_PER_REQUEST; 687 688 second->u.rw.nr_segments = nr_segments - BLKIF_MAX_SEGMENTS_PER_REQUEST; 689 second->u.rw.sector_number = first->u.rw.sector_number + 690 (BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE) / 512; 691 692 second->u.rw.handle = first->u.rw.handle; 693 second->operation = first->operation; 694 } 695 696 static int blkif_queue_rw_req(struct request *req, struct blkfront_ring_info *rinfo) 697 { 698 struct blkfront_info *info = rinfo->dev_info; 699 struct blkif_request *ring_req, *extra_ring_req = NULL; 700 struct blkif_request *final_ring_req, *final_extra_ring_req = NULL; 701 unsigned long id, extra_id = NO_ASSOCIATED_ID; 702 bool require_extra_req = false; 703 int i; 704 struct setup_rw_req setup = { 705 .grant_idx = 0, 706 .segments = NULL, 707 .rinfo = rinfo, 708 .need_copy = rq_data_dir(req) && info->feature_persistent, 709 }; 710 711 /* 712 * Used to store if we are able to queue the request by just using 713 * existing persistent grants, or if we have to get new grants, 714 * as there are not sufficiently many free. 715 */ 716 bool new_persistent_gnts = false; 717 struct scatterlist *sg; 718 int num_sg, max_grefs, num_grant; 719 720 max_grefs = req->nr_phys_segments * GRANTS_PER_PSEG; 721 if (max_grefs > BLKIF_MAX_SEGMENTS_PER_REQUEST) 722 /* 723 * If we are using indirect segments we need to account 724 * for the indirect grefs used in the request. 725 */ 726 max_grefs += INDIRECT_GREFS(max_grefs); 727 728 /* Check if we have enough persistent grants to allocate a requests */ 729 if (rinfo->persistent_gnts_c < max_grefs) { 730 new_persistent_gnts = true; 731 732 if (gnttab_alloc_grant_references( 733 max_grefs - rinfo->persistent_gnts_c, 734 &setup.gref_head) < 0) { 735 gnttab_request_free_callback( 736 &rinfo->callback, 737 blkif_restart_queue_callback, 738 rinfo, 739 max_grefs - rinfo->persistent_gnts_c); 740 return 1; 741 } 742 } 743 744 /* Fill out a communications ring structure. */ 745 id = blkif_ring_get_request(rinfo, req, &final_ring_req); 746 ring_req = &rinfo->shadow[id].req; 747 748 num_sg = blk_rq_map_sg(req->q, req, rinfo->shadow[id].sg); 749 num_grant = 0; 750 /* Calculate the number of grant used */ 751 for_each_sg(rinfo->shadow[id].sg, sg, num_sg, i) 752 num_grant += gnttab_count_grant(sg->offset, sg->length); 753 754 require_extra_req = info->max_indirect_segments == 0 && 755 num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST; 756 BUG_ON(!HAS_EXTRA_REQ && require_extra_req); 757 758 rinfo->shadow[id].num_sg = num_sg; 759 if (num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST && 760 likely(!require_extra_req)) { 761 /* 762 * The indirect operation can only be a BLKIF_OP_READ or 763 * BLKIF_OP_WRITE 764 */ 765 BUG_ON(req_op(req) == REQ_OP_FLUSH || req->cmd_flags & REQ_FUA); 766 ring_req->operation = BLKIF_OP_INDIRECT; 767 ring_req->u.indirect.indirect_op = rq_data_dir(req) ? 768 BLKIF_OP_WRITE : BLKIF_OP_READ; 769 ring_req->u.indirect.sector_number = (blkif_sector_t)blk_rq_pos(req); 770 ring_req->u.indirect.handle = info->handle; 771 ring_req->u.indirect.nr_segments = num_grant; 772 } else { 773 ring_req->u.rw.sector_number = (blkif_sector_t)blk_rq_pos(req); 774 ring_req->u.rw.handle = info->handle; 775 ring_req->operation = rq_data_dir(req) ? 776 BLKIF_OP_WRITE : BLKIF_OP_READ; 777 if (req_op(req) == REQ_OP_FLUSH || req->cmd_flags & REQ_FUA) { 778 /* 779 * Ideally we can do an unordered flush-to-disk. 780 * In case the backend onlysupports barriers, use that. 781 * A barrier request a superset of FUA, so we can 782 * implement it the same way. (It's also a FLUSH+FUA, 783 * since it is guaranteed ordered WRT previous writes.) 784 */ 785 if (info->feature_flush && info->feature_fua) 786 ring_req->operation = 787 BLKIF_OP_WRITE_BARRIER; 788 else if (info->feature_flush) 789 ring_req->operation = 790 BLKIF_OP_FLUSH_DISKCACHE; 791 else 792 ring_req->operation = 0; 793 } 794 ring_req->u.rw.nr_segments = num_grant; 795 if (unlikely(require_extra_req)) { 796 extra_id = blkif_ring_get_request(rinfo, req, 797 &final_extra_ring_req); 798 extra_ring_req = &rinfo->shadow[extra_id].req; 799 800 /* 801 * Only the first request contains the scatter-gather 802 * list. 803 */ 804 rinfo->shadow[extra_id].num_sg = 0; 805 806 blkif_setup_extra_req(ring_req, extra_ring_req); 807 808 /* Link the 2 requests together */ 809 rinfo->shadow[extra_id].associated_id = id; 810 rinfo->shadow[id].associated_id = extra_id; 811 } 812 } 813 814 setup.ring_req = ring_req; 815 setup.id = id; 816 817 setup.require_extra_req = require_extra_req; 818 if (unlikely(require_extra_req)) 819 setup.extra_ring_req = extra_ring_req; 820 821 for_each_sg(rinfo->shadow[id].sg, sg, num_sg, i) { 822 BUG_ON(sg->offset + sg->length > PAGE_SIZE); 823 824 if (setup.need_copy) { 825 setup.bvec_off = sg->offset; 826 setup.bvec_data = kmap_atomic(sg_page(sg)); 827 } 828 829 gnttab_foreach_grant_in_range(sg_page(sg), 830 sg->offset, 831 sg->length, 832 blkif_setup_rw_req_grant, 833 &setup); 834 835 if (setup.need_copy) 836 kunmap_atomic(setup.bvec_data); 837 } 838 if (setup.segments) 839 kunmap_atomic(setup.segments); 840 841 /* Copy request(s) to the ring page. */ 842 *final_ring_req = *ring_req; 843 rinfo->shadow[id].status = REQ_WAITING; 844 if (unlikely(require_extra_req)) { 845 *final_extra_ring_req = *extra_ring_req; 846 rinfo->shadow[extra_id].status = REQ_WAITING; 847 } 848 849 if (new_persistent_gnts) 850 gnttab_free_grant_references(setup.gref_head); 851 852 return 0; 853 } 854 855 /* 856 * Generate a Xen blkfront IO request from a blk layer request. Reads 857 * and writes are handled as expected. 858 * 859 * @req: a request struct 860 */ 861 static int blkif_queue_request(struct request *req, struct blkfront_ring_info *rinfo) 862 { 863 if (unlikely(rinfo->dev_info->connected != BLKIF_STATE_CONNECTED)) 864 return 1; 865 866 if (unlikely(req_op(req) == REQ_OP_DISCARD || 867 req_op(req) == REQ_OP_SECURE_ERASE)) 868 return blkif_queue_discard_req(req, rinfo); 869 else 870 return blkif_queue_rw_req(req, rinfo); 871 } 872 873 static inline void flush_requests(struct blkfront_ring_info *rinfo) 874 { 875 int notify; 876 877 RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&rinfo->ring, notify); 878 879 if (notify) 880 notify_remote_via_irq(rinfo->irq); 881 } 882 883 static inline bool blkif_request_flush_invalid(struct request *req, 884 struct blkfront_info *info) 885 { 886 return (blk_rq_is_passthrough(req) || 887 ((req_op(req) == REQ_OP_FLUSH) && 888 !info->feature_flush) || 889 ((req->cmd_flags & REQ_FUA) && 890 !info->feature_fua)); 891 } 892 893 static blk_status_t blkif_queue_rq(struct blk_mq_hw_ctx *hctx, 894 const struct blk_mq_queue_data *qd) 895 { 896 unsigned long flags; 897 int qid = hctx->queue_num; 898 struct blkfront_info *info = hctx->queue->queuedata; 899 struct blkfront_ring_info *rinfo = NULL; 900 901 rinfo = get_rinfo(info, qid); 902 blk_mq_start_request(qd->rq); 903 spin_lock_irqsave(&rinfo->ring_lock, flags); 904 if (RING_FULL(&rinfo->ring)) 905 goto out_busy; 906 907 if (blkif_request_flush_invalid(qd->rq, rinfo->dev_info)) 908 goto out_err; 909 910 if (blkif_queue_request(qd->rq, rinfo)) 911 goto out_busy; 912 913 flush_requests(rinfo); 914 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 915 return BLK_STS_OK; 916 917 out_err: 918 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 919 return BLK_STS_IOERR; 920 921 out_busy: 922 blk_mq_stop_hw_queue(hctx); 923 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 924 return BLK_STS_DEV_RESOURCE; 925 } 926 927 static void blkif_complete_rq(struct request *rq) 928 { 929 blk_mq_end_request(rq, blkif_req(rq)->error); 930 } 931 932 static const struct blk_mq_ops blkfront_mq_ops = { 933 .queue_rq = blkif_queue_rq, 934 .complete = blkif_complete_rq, 935 }; 936 937 static void blkif_set_queue_limits(struct blkfront_info *info) 938 { 939 struct request_queue *rq = info->rq; 940 struct gendisk *gd = info->gd; 941 unsigned int segments = info->max_indirect_segments ? : 942 BLKIF_MAX_SEGMENTS_PER_REQUEST; 943 944 blk_queue_flag_set(QUEUE_FLAG_VIRT, rq); 945 946 if (info->feature_discard) { 947 blk_queue_flag_set(QUEUE_FLAG_DISCARD, rq); 948 blk_queue_max_discard_sectors(rq, get_capacity(gd)); 949 rq->limits.discard_granularity = info->discard_granularity ?: 950 info->physical_sector_size; 951 rq->limits.discard_alignment = info->discard_alignment; 952 if (info->feature_secdiscard) 953 blk_queue_flag_set(QUEUE_FLAG_SECERASE, rq); 954 } 955 956 /* Hard sector size and max sectors impersonate the equiv. hardware. */ 957 blk_queue_logical_block_size(rq, info->sector_size); 958 blk_queue_physical_block_size(rq, info->physical_sector_size); 959 blk_queue_max_hw_sectors(rq, (segments * XEN_PAGE_SIZE) / 512); 960 961 /* Each segment in a request is up to an aligned page in size. */ 962 blk_queue_segment_boundary(rq, PAGE_SIZE - 1); 963 blk_queue_max_segment_size(rq, PAGE_SIZE); 964 965 /* Ensure a merged request will fit in a single I/O ring slot. */ 966 blk_queue_max_segments(rq, segments / GRANTS_PER_PSEG); 967 968 /* Make sure buffer addresses are sector-aligned. */ 969 blk_queue_dma_alignment(rq, 511); 970 } 971 972 static const char *flush_info(struct blkfront_info *info) 973 { 974 if (info->feature_flush && info->feature_fua) 975 return "barrier: enabled;"; 976 else if (info->feature_flush) 977 return "flush diskcache: enabled;"; 978 else 979 return "barrier or flush: disabled;"; 980 } 981 982 static void xlvbd_flush(struct blkfront_info *info) 983 { 984 blk_queue_write_cache(info->rq, info->feature_flush ? true : false, 985 info->feature_fua ? true : false); 986 pr_info("blkfront: %s: %s %s %s %s %s\n", 987 info->gd->disk_name, flush_info(info), 988 "persistent grants:", info->feature_persistent ? 989 "enabled;" : "disabled;", "indirect descriptors:", 990 info->max_indirect_segments ? "enabled;" : "disabled;"); 991 } 992 993 static int xen_translate_vdev(int vdevice, int *minor, unsigned int *offset) 994 { 995 int major; 996 major = BLKIF_MAJOR(vdevice); 997 *minor = BLKIF_MINOR(vdevice); 998 switch (major) { 999 case XEN_IDE0_MAJOR: 1000 *offset = (*minor / 64) + EMULATED_HD_DISK_NAME_OFFSET; 1001 *minor = ((*minor / 64) * PARTS_PER_DISK) + 1002 EMULATED_HD_DISK_MINOR_OFFSET; 1003 break; 1004 case XEN_IDE1_MAJOR: 1005 *offset = (*minor / 64) + 2 + EMULATED_HD_DISK_NAME_OFFSET; 1006 *minor = (((*minor / 64) + 2) * PARTS_PER_DISK) + 1007 EMULATED_HD_DISK_MINOR_OFFSET; 1008 break; 1009 case XEN_SCSI_DISK0_MAJOR: 1010 *offset = (*minor / PARTS_PER_DISK) + EMULATED_SD_DISK_NAME_OFFSET; 1011 *minor = *minor + EMULATED_SD_DISK_MINOR_OFFSET; 1012 break; 1013 case XEN_SCSI_DISK1_MAJOR: 1014 case XEN_SCSI_DISK2_MAJOR: 1015 case XEN_SCSI_DISK3_MAJOR: 1016 case XEN_SCSI_DISK4_MAJOR: 1017 case XEN_SCSI_DISK5_MAJOR: 1018 case XEN_SCSI_DISK6_MAJOR: 1019 case XEN_SCSI_DISK7_MAJOR: 1020 *offset = (*minor / PARTS_PER_DISK) + 1021 ((major - XEN_SCSI_DISK1_MAJOR + 1) * 16) + 1022 EMULATED_SD_DISK_NAME_OFFSET; 1023 *minor = *minor + 1024 ((major - XEN_SCSI_DISK1_MAJOR + 1) * 16 * PARTS_PER_DISK) + 1025 EMULATED_SD_DISK_MINOR_OFFSET; 1026 break; 1027 case XEN_SCSI_DISK8_MAJOR: 1028 case XEN_SCSI_DISK9_MAJOR: 1029 case XEN_SCSI_DISK10_MAJOR: 1030 case XEN_SCSI_DISK11_MAJOR: 1031 case XEN_SCSI_DISK12_MAJOR: 1032 case XEN_SCSI_DISK13_MAJOR: 1033 case XEN_SCSI_DISK14_MAJOR: 1034 case XEN_SCSI_DISK15_MAJOR: 1035 *offset = (*minor / PARTS_PER_DISK) + 1036 ((major - XEN_SCSI_DISK8_MAJOR + 8) * 16) + 1037 EMULATED_SD_DISK_NAME_OFFSET; 1038 *minor = *minor + 1039 ((major - XEN_SCSI_DISK8_MAJOR + 8) * 16 * PARTS_PER_DISK) + 1040 EMULATED_SD_DISK_MINOR_OFFSET; 1041 break; 1042 case XENVBD_MAJOR: 1043 *offset = *minor / PARTS_PER_DISK; 1044 break; 1045 default: 1046 printk(KERN_WARNING "blkfront: your disk configuration is " 1047 "incorrect, please use an xvd device instead\n"); 1048 return -ENODEV; 1049 } 1050 return 0; 1051 } 1052 1053 static char *encode_disk_name(char *ptr, unsigned int n) 1054 { 1055 if (n >= 26) 1056 ptr = encode_disk_name(ptr, n / 26 - 1); 1057 *ptr = 'a' + n % 26; 1058 return ptr + 1; 1059 } 1060 1061 static int xlvbd_alloc_gendisk(blkif_sector_t capacity, 1062 struct blkfront_info *info, u16 sector_size, 1063 unsigned int physical_sector_size) 1064 { 1065 struct gendisk *gd; 1066 int nr_minors = 1; 1067 int err; 1068 unsigned int offset; 1069 int minor; 1070 int nr_parts; 1071 char *ptr; 1072 1073 BUG_ON(info->gd != NULL); 1074 BUG_ON(info->rq != NULL); 1075 1076 if ((info->vdevice>>EXT_SHIFT) > 1) { 1077 /* this is above the extended range; something is wrong */ 1078 printk(KERN_WARNING "blkfront: vdevice 0x%x is above the extended range; ignoring\n", info->vdevice); 1079 return -ENODEV; 1080 } 1081 1082 if (!VDEV_IS_EXTENDED(info->vdevice)) { 1083 err = xen_translate_vdev(info->vdevice, &minor, &offset); 1084 if (err) 1085 return err; 1086 nr_parts = PARTS_PER_DISK; 1087 } else { 1088 minor = BLKIF_MINOR_EXT(info->vdevice); 1089 nr_parts = PARTS_PER_EXT_DISK; 1090 offset = minor / nr_parts; 1091 if (xen_hvm_domain() && offset < EMULATED_HD_DISK_NAME_OFFSET + 4) 1092 printk(KERN_WARNING "blkfront: vdevice 0x%x might conflict with " 1093 "emulated IDE disks,\n\t choose an xvd device name" 1094 "from xvde on\n", info->vdevice); 1095 } 1096 if (minor >> MINORBITS) { 1097 pr_warn("blkfront: %#x's minor (%#x) out of range; ignoring\n", 1098 info->vdevice, minor); 1099 return -ENODEV; 1100 } 1101 1102 if ((minor % nr_parts) == 0) 1103 nr_minors = nr_parts; 1104 1105 err = xlbd_reserve_minors(minor, nr_minors); 1106 if (err) 1107 return err; 1108 1109 memset(&info->tag_set, 0, sizeof(info->tag_set)); 1110 info->tag_set.ops = &blkfront_mq_ops; 1111 info->tag_set.nr_hw_queues = info->nr_rings; 1112 if (HAS_EXTRA_REQ && info->max_indirect_segments == 0) { 1113 /* 1114 * When indirect descriptior is not supported, the I/O request 1115 * will be split between multiple request in the ring. 1116 * To avoid problems when sending the request, divide by 1117 * 2 the depth of the queue. 1118 */ 1119 info->tag_set.queue_depth = BLK_RING_SIZE(info) / 2; 1120 } else 1121 info->tag_set.queue_depth = BLK_RING_SIZE(info); 1122 info->tag_set.numa_node = NUMA_NO_NODE; 1123 info->tag_set.flags = BLK_MQ_F_SHOULD_MERGE; 1124 info->tag_set.cmd_size = sizeof(struct blkif_req); 1125 info->tag_set.driver_data = info; 1126 1127 err = blk_mq_alloc_tag_set(&info->tag_set); 1128 if (err) 1129 goto out_release_minors; 1130 1131 gd = blk_mq_alloc_disk(&info->tag_set, info); 1132 if (IS_ERR(gd)) { 1133 err = PTR_ERR(gd); 1134 goto out_free_tag_set; 1135 } 1136 1137 strcpy(gd->disk_name, DEV_NAME); 1138 ptr = encode_disk_name(gd->disk_name + sizeof(DEV_NAME) - 1, offset); 1139 BUG_ON(ptr >= gd->disk_name + DISK_NAME_LEN); 1140 if (nr_minors > 1) 1141 *ptr = 0; 1142 else 1143 snprintf(ptr, gd->disk_name + DISK_NAME_LEN - ptr, 1144 "%d", minor & (nr_parts - 1)); 1145 1146 gd->major = XENVBD_MAJOR; 1147 gd->first_minor = minor; 1148 gd->minors = nr_minors; 1149 gd->fops = &xlvbd_block_fops; 1150 gd->private_data = info; 1151 set_capacity(gd, capacity); 1152 1153 info->rq = gd->queue; 1154 info->gd = gd; 1155 info->sector_size = sector_size; 1156 info->physical_sector_size = physical_sector_size; 1157 blkif_set_queue_limits(info); 1158 1159 xlvbd_flush(info); 1160 1161 if (info->vdisk_info & VDISK_READONLY) 1162 set_disk_ro(gd, 1); 1163 if (info->vdisk_info & VDISK_REMOVABLE) 1164 gd->flags |= GENHD_FL_REMOVABLE; 1165 1166 return 0; 1167 1168 out_free_tag_set: 1169 blk_mq_free_tag_set(&info->tag_set); 1170 out_release_minors: 1171 xlbd_release_minors(minor, nr_minors); 1172 return err; 1173 } 1174 1175 /* Already hold rinfo->ring_lock. */ 1176 static inline void kick_pending_request_queues_locked(struct blkfront_ring_info *rinfo) 1177 { 1178 if (!RING_FULL(&rinfo->ring)) 1179 blk_mq_start_stopped_hw_queues(rinfo->dev_info->rq, true); 1180 } 1181 1182 static void kick_pending_request_queues(struct blkfront_ring_info *rinfo) 1183 { 1184 unsigned long flags; 1185 1186 spin_lock_irqsave(&rinfo->ring_lock, flags); 1187 kick_pending_request_queues_locked(rinfo); 1188 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 1189 } 1190 1191 static void blkif_restart_queue(struct work_struct *work) 1192 { 1193 struct blkfront_ring_info *rinfo = container_of(work, struct blkfront_ring_info, work); 1194 1195 if (rinfo->dev_info->connected == BLKIF_STATE_CONNECTED) 1196 kick_pending_request_queues(rinfo); 1197 } 1198 1199 static void blkif_free_ring(struct blkfront_ring_info *rinfo) 1200 { 1201 struct grant *persistent_gnt, *n; 1202 struct blkfront_info *info = rinfo->dev_info; 1203 int i, j, segs; 1204 1205 /* 1206 * Remove indirect pages, this only happens when using indirect 1207 * descriptors but not persistent grants 1208 */ 1209 if (!list_empty(&rinfo->indirect_pages)) { 1210 struct page *indirect_page, *n; 1211 1212 BUG_ON(info->feature_persistent); 1213 list_for_each_entry_safe(indirect_page, n, &rinfo->indirect_pages, lru) { 1214 list_del(&indirect_page->lru); 1215 __free_page(indirect_page); 1216 } 1217 } 1218 1219 /* Remove all persistent grants. */ 1220 if (!list_empty(&rinfo->grants)) { 1221 list_for_each_entry_safe(persistent_gnt, n, 1222 &rinfo->grants, node) { 1223 list_del(&persistent_gnt->node); 1224 if (persistent_gnt->gref != GRANT_INVALID_REF) { 1225 gnttab_end_foreign_access(persistent_gnt->gref, 1226 0, 0UL); 1227 rinfo->persistent_gnts_c--; 1228 } 1229 if (info->feature_persistent) 1230 __free_page(persistent_gnt->page); 1231 kfree(persistent_gnt); 1232 } 1233 } 1234 BUG_ON(rinfo->persistent_gnts_c != 0); 1235 1236 for (i = 0; i < BLK_RING_SIZE(info); i++) { 1237 /* 1238 * Clear persistent grants present in requests already 1239 * on the shared ring 1240 */ 1241 if (!rinfo->shadow[i].request) 1242 goto free_shadow; 1243 1244 segs = rinfo->shadow[i].req.operation == BLKIF_OP_INDIRECT ? 1245 rinfo->shadow[i].req.u.indirect.nr_segments : 1246 rinfo->shadow[i].req.u.rw.nr_segments; 1247 for (j = 0; j < segs; j++) { 1248 persistent_gnt = rinfo->shadow[i].grants_used[j]; 1249 gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL); 1250 if (info->feature_persistent) 1251 __free_page(persistent_gnt->page); 1252 kfree(persistent_gnt); 1253 } 1254 1255 if (rinfo->shadow[i].req.operation != BLKIF_OP_INDIRECT) 1256 /* 1257 * If this is not an indirect operation don't try to 1258 * free indirect segments 1259 */ 1260 goto free_shadow; 1261 1262 for (j = 0; j < INDIRECT_GREFS(segs); j++) { 1263 persistent_gnt = rinfo->shadow[i].indirect_grants[j]; 1264 gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL); 1265 __free_page(persistent_gnt->page); 1266 kfree(persistent_gnt); 1267 } 1268 1269 free_shadow: 1270 kvfree(rinfo->shadow[i].grants_used); 1271 rinfo->shadow[i].grants_used = NULL; 1272 kvfree(rinfo->shadow[i].indirect_grants); 1273 rinfo->shadow[i].indirect_grants = NULL; 1274 kvfree(rinfo->shadow[i].sg); 1275 rinfo->shadow[i].sg = NULL; 1276 } 1277 1278 /* No more gnttab callback work. */ 1279 gnttab_cancel_free_callback(&rinfo->callback); 1280 1281 /* Flush gnttab callback work. Must be done with no locks held. */ 1282 flush_work(&rinfo->work); 1283 1284 /* Free resources associated with old device channel. */ 1285 for (i = 0; i < info->nr_ring_pages; i++) { 1286 if (rinfo->ring_ref[i] != GRANT_INVALID_REF) { 1287 gnttab_end_foreign_access(rinfo->ring_ref[i], 0, 0); 1288 rinfo->ring_ref[i] = GRANT_INVALID_REF; 1289 } 1290 } 1291 free_pages((unsigned long)rinfo->ring.sring, get_order(info->nr_ring_pages * XEN_PAGE_SIZE)); 1292 rinfo->ring.sring = NULL; 1293 1294 if (rinfo->irq) 1295 unbind_from_irqhandler(rinfo->irq, rinfo); 1296 rinfo->evtchn = rinfo->irq = 0; 1297 } 1298 1299 static void blkif_free(struct blkfront_info *info, int suspend) 1300 { 1301 unsigned int i; 1302 struct blkfront_ring_info *rinfo; 1303 1304 /* Prevent new requests being issued until we fix things up. */ 1305 info->connected = suspend ? 1306 BLKIF_STATE_SUSPENDED : BLKIF_STATE_DISCONNECTED; 1307 /* No more blkif_request(). */ 1308 if (info->rq) 1309 blk_mq_stop_hw_queues(info->rq); 1310 1311 for_each_rinfo(info, rinfo, i) 1312 blkif_free_ring(rinfo); 1313 1314 kvfree(info->rinfo); 1315 info->rinfo = NULL; 1316 info->nr_rings = 0; 1317 } 1318 1319 struct copy_from_grant { 1320 const struct blk_shadow *s; 1321 unsigned int grant_idx; 1322 unsigned int bvec_offset; 1323 char *bvec_data; 1324 }; 1325 1326 static void blkif_copy_from_grant(unsigned long gfn, unsigned int offset, 1327 unsigned int len, void *data) 1328 { 1329 struct copy_from_grant *info = data; 1330 char *shared_data; 1331 /* Convenient aliases */ 1332 const struct blk_shadow *s = info->s; 1333 1334 shared_data = kmap_atomic(s->grants_used[info->grant_idx]->page); 1335 1336 memcpy(info->bvec_data + info->bvec_offset, 1337 shared_data + offset, len); 1338 1339 info->bvec_offset += len; 1340 info->grant_idx++; 1341 1342 kunmap_atomic(shared_data); 1343 } 1344 1345 static enum blk_req_status blkif_rsp_to_req_status(int rsp) 1346 { 1347 switch (rsp) 1348 { 1349 case BLKIF_RSP_OKAY: 1350 return REQ_DONE; 1351 case BLKIF_RSP_EOPNOTSUPP: 1352 return REQ_EOPNOTSUPP; 1353 case BLKIF_RSP_ERROR: 1354 default: 1355 return REQ_ERROR; 1356 } 1357 } 1358 1359 /* 1360 * Get the final status of the block request based on two ring response 1361 */ 1362 static int blkif_get_final_status(enum blk_req_status s1, 1363 enum blk_req_status s2) 1364 { 1365 BUG_ON(s1 < REQ_DONE); 1366 BUG_ON(s2 < REQ_DONE); 1367 1368 if (s1 == REQ_ERROR || s2 == REQ_ERROR) 1369 return BLKIF_RSP_ERROR; 1370 else if (s1 == REQ_EOPNOTSUPP || s2 == REQ_EOPNOTSUPP) 1371 return BLKIF_RSP_EOPNOTSUPP; 1372 return BLKIF_RSP_OKAY; 1373 } 1374 1375 static bool blkif_completion(unsigned long *id, 1376 struct blkfront_ring_info *rinfo, 1377 struct blkif_response *bret) 1378 { 1379 int i = 0; 1380 struct scatterlist *sg; 1381 int num_sg, num_grant; 1382 struct blkfront_info *info = rinfo->dev_info; 1383 struct blk_shadow *s = &rinfo->shadow[*id]; 1384 struct copy_from_grant data = { 1385 .grant_idx = 0, 1386 }; 1387 1388 num_grant = s->req.operation == BLKIF_OP_INDIRECT ? 1389 s->req.u.indirect.nr_segments : s->req.u.rw.nr_segments; 1390 1391 /* The I/O request may be split in two. */ 1392 if (unlikely(s->associated_id != NO_ASSOCIATED_ID)) { 1393 struct blk_shadow *s2 = &rinfo->shadow[s->associated_id]; 1394 1395 /* Keep the status of the current response in shadow. */ 1396 s->status = blkif_rsp_to_req_status(bret->status); 1397 1398 /* Wait the second response if not yet here. */ 1399 if (s2->status < REQ_DONE) 1400 return false; 1401 1402 bret->status = blkif_get_final_status(s->status, 1403 s2->status); 1404 1405 /* 1406 * All the grants is stored in the first shadow in order 1407 * to make the completion code simpler. 1408 */ 1409 num_grant += s2->req.u.rw.nr_segments; 1410 1411 /* 1412 * The two responses may not come in order. Only the 1413 * first request will store the scatter-gather list. 1414 */ 1415 if (s2->num_sg != 0) { 1416 /* Update "id" with the ID of the first response. */ 1417 *id = s->associated_id; 1418 s = s2; 1419 } 1420 1421 /* 1422 * We don't need anymore the second request, so recycling 1423 * it now. 1424 */ 1425 if (add_id_to_freelist(rinfo, s->associated_id)) 1426 WARN(1, "%s: can't recycle the second part (id = %ld) of the request\n", 1427 info->gd->disk_name, s->associated_id); 1428 } 1429 1430 data.s = s; 1431 num_sg = s->num_sg; 1432 1433 if (bret->operation == BLKIF_OP_READ && info->feature_persistent) { 1434 for_each_sg(s->sg, sg, num_sg, i) { 1435 BUG_ON(sg->offset + sg->length > PAGE_SIZE); 1436 1437 data.bvec_offset = sg->offset; 1438 data.bvec_data = kmap_atomic(sg_page(sg)); 1439 1440 gnttab_foreach_grant_in_range(sg_page(sg), 1441 sg->offset, 1442 sg->length, 1443 blkif_copy_from_grant, 1444 &data); 1445 1446 kunmap_atomic(data.bvec_data); 1447 } 1448 } 1449 /* Add the persistent grant into the list of free grants */ 1450 for (i = 0; i < num_grant; i++) { 1451 if (gnttab_query_foreign_access(s->grants_used[i]->gref)) { 1452 /* 1453 * If the grant is still mapped by the backend (the 1454 * backend has chosen to make this grant persistent) 1455 * we add it at the head of the list, so it will be 1456 * reused first. 1457 */ 1458 if (!info->feature_persistent) 1459 pr_alert_ratelimited("backed has not unmapped grant: %u\n", 1460 s->grants_used[i]->gref); 1461 list_add(&s->grants_used[i]->node, &rinfo->grants); 1462 rinfo->persistent_gnts_c++; 1463 } else { 1464 /* 1465 * If the grant is not mapped by the backend we end the 1466 * foreign access and add it to the tail of the list, 1467 * so it will not be picked again unless we run out of 1468 * persistent grants. 1469 */ 1470 gnttab_end_foreign_access(s->grants_used[i]->gref, 0, 0UL); 1471 s->grants_used[i]->gref = GRANT_INVALID_REF; 1472 list_add_tail(&s->grants_used[i]->node, &rinfo->grants); 1473 } 1474 } 1475 if (s->req.operation == BLKIF_OP_INDIRECT) { 1476 for (i = 0; i < INDIRECT_GREFS(num_grant); i++) { 1477 if (gnttab_query_foreign_access(s->indirect_grants[i]->gref)) { 1478 if (!info->feature_persistent) 1479 pr_alert_ratelimited("backed has not unmapped grant: %u\n", 1480 s->indirect_grants[i]->gref); 1481 list_add(&s->indirect_grants[i]->node, &rinfo->grants); 1482 rinfo->persistent_gnts_c++; 1483 } else { 1484 struct page *indirect_page; 1485 1486 gnttab_end_foreign_access(s->indirect_grants[i]->gref, 0, 0UL); 1487 /* 1488 * Add the used indirect page back to the list of 1489 * available pages for indirect grefs. 1490 */ 1491 if (!info->feature_persistent) { 1492 indirect_page = s->indirect_grants[i]->page; 1493 list_add(&indirect_page->lru, &rinfo->indirect_pages); 1494 } 1495 s->indirect_grants[i]->gref = GRANT_INVALID_REF; 1496 list_add_tail(&s->indirect_grants[i]->node, &rinfo->grants); 1497 } 1498 } 1499 } 1500 1501 return true; 1502 } 1503 1504 static irqreturn_t blkif_interrupt(int irq, void *dev_id) 1505 { 1506 struct request *req; 1507 struct blkif_response bret; 1508 RING_IDX i, rp; 1509 unsigned long flags; 1510 struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)dev_id; 1511 struct blkfront_info *info = rinfo->dev_info; 1512 unsigned int eoiflag = XEN_EOI_FLAG_SPURIOUS; 1513 1514 if (unlikely(info->connected != BLKIF_STATE_CONNECTED)) { 1515 xen_irq_lateeoi(irq, XEN_EOI_FLAG_SPURIOUS); 1516 return IRQ_HANDLED; 1517 } 1518 1519 spin_lock_irqsave(&rinfo->ring_lock, flags); 1520 again: 1521 rp = READ_ONCE(rinfo->ring.sring->rsp_prod); 1522 virt_rmb(); /* Ensure we see queued responses up to 'rp'. */ 1523 if (RING_RESPONSE_PROD_OVERFLOW(&rinfo->ring, rp)) { 1524 pr_alert("%s: illegal number of responses %u\n", 1525 info->gd->disk_name, rp - rinfo->ring.rsp_cons); 1526 goto err; 1527 } 1528 1529 for (i = rinfo->ring.rsp_cons; i != rp; i++) { 1530 unsigned long id; 1531 unsigned int op; 1532 1533 eoiflag = 0; 1534 1535 RING_COPY_RESPONSE(&rinfo->ring, i, &bret); 1536 id = bret.id; 1537 1538 /* 1539 * The backend has messed up and given us an id that we would 1540 * never have given to it (we stamp it up to BLK_RING_SIZE - 1541 * look in get_id_from_freelist. 1542 */ 1543 if (id >= BLK_RING_SIZE(info)) { 1544 pr_alert("%s: response has incorrect id (%ld)\n", 1545 info->gd->disk_name, id); 1546 goto err; 1547 } 1548 if (rinfo->shadow[id].status != REQ_WAITING) { 1549 pr_alert("%s: response references no pending request\n", 1550 info->gd->disk_name); 1551 goto err; 1552 } 1553 1554 rinfo->shadow[id].status = REQ_PROCESSING; 1555 req = rinfo->shadow[id].request; 1556 1557 op = rinfo->shadow[id].req.operation; 1558 if (op == BLKIF_OP_INDIRECT) 1559 op = rinfo->shadow[id].req.u.indirect.indirect_op; 1560 if (bret.operation != op) { 1561 pr_alert("%s: response has wrong operation (%u instead of %u)\n", 1562 info->gd->disk_name, bret.operation, op); 1563 goto err; 1564 } 1565 1566 if (bret.operation != BLKIF_OP_DISCARD) { 1567 /* 1568 * We may need to wait for an extra response if the 1569 * I/O request is split in 2 1570 */ 1571 if (!blkif_completion(&id, rinfo, &bret)) 1572 continue; 1573 } 1574 1575 if (add_id_to_freelist(rinfo, id)) { 1576 WARN(1, "%s: response to %s (id %ld) couldn't be recycled!\n", 1577 info->gd->disk_name, op_name(bret.operation), id); 1578 continue; 1579 } 1580 1581 if (bret.status == BLKIF_RSP_OKAY) 1582 blkif_req(req)->error = BLK_STS_OK; 1583 else 1584 blkif_req(req)->error = BLK_STS_IOERR; 1585 1586 switch (bret.operation) { 1587 case BLKIF_OP_DISCARD: 1588 if (unlikely(bret.status == BLKIF_RSP_EOPNOTSUPP)) { 1589 struct request_queue *rq = info->rq; 1590 1591 pr_warn_ratelimited("blkfront: %s: %s op failed\n", 1592 info->gd->disk_name, op_name(bret.operation)); 1593 blkif_req(req)->error = BLK_STS_NOTSUPP; 1594 info->feature_discard = 0; 1595 info->feature_secdiscard = 0; 1596 blk_queue_flag_clear(QUEUE_FLAG_DISCARD, rq); 1597 blk_queue_flag_clear(QUEUE_FLAG_SECERASE, rq); 1598 } 1599 break; 1600 case BLKIF_OP_FLUSH_DISKCACHE: 1601 case BLKIF_OP_WRITE_BARRIER: 1602 if (unlikely(bret.status == BLKIF_RSP_EOPNOTSUPP)) { 1603 pr_warn_ratelimited("blkfront: %s: %s op failed\n", 1604 info->gd->disk_name, op_name(bret.operation)); 1605 blkif_req(req)->error = BLK_STS_NOTSUPP; 1606 } 1607 if (unlikely(bret.status == BLKIF_RSP_ERROR && 1608 rinfo->shadow[id].req.u.rw.nr_segments == 0)) { 1609 pr_warn_ratelimited("blkfront: %s: empty %s op failed\n", 1610 info->gd->disk_name, op_name(bret.operation)); 1611 blkif_req(req)->error = BLK_STS_NOTSUPP; 1612 } 1613 if (unlikely(blkif_req(req)->error)) { 1614 if (blkif_req(req)->error == BLK_STS_NOTSUPP) 1615 blkif_req(req)->error = BLK_STS_OK; 1616 info->feature_fua = 0; 1617 info->feature_flush = 0; 1618 xlvbd_flush(info); 1619 } 1620 fallthrough; 1621 case BLKIF_OP_READ: 1622 case BLKIF_OP_WRITE: 1623 if (unlikely(bret.status != BLKIF_RSP_OKAY)) 1624 dev_dbg_ratelimited(&info->xbdev->dev, 1625 "Bad return from blkdev data request: %#x\n", 1626 bret.status); 1627 1628 break; 1629 default: 1630 BUG(); 1631 } 1632 1633 if (likely(!blk_should_fake_timeout(req->q))) 1634 blk_mq_complete_request(req); 1635 } 1636 1637 rinfo->ring.rsp_cons = i; 1638 1639 if (i != rinfo->ring.req_prod_pvt) { 1640 int more_to_do; 1641 RING_FINAL_CHECK_FOR_RESPONSES(&rinfo->ring, more_to_do); 1642 if (more_to_do) 1643 goto again; 1644 } else 1645 rinfo->ring.sring->rsp_event = i + 1; 1646 1647 kick_pending_request_queues_locked(rinfo); 1648 1649 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 1650 1651 xen_irq_lateeoi(irq, eoiflag); 1652 1653 return IRQ_HANDLED; 1654 1655 err: 1656 info->connected = BLKIF_STATE_ERROR; 1657 1658 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 1659 1660 /* No EOI in order to avoid further interrupts. */ 1661 1662 pr_alert("%s disabled for further use\n", info->gd->disk_name); 1663 return IRQ_HANDLED; 1664 } 1665 1666 1667 static int setup_blkring(struct xenbus_device *dev, 1668 struct blkfront_ring_info *rinfo) 1669 { 1670 struct blkif_sring *sring; 1671 int err, i; 1672 struct blkfront_info *info = rinfo->dev_info; 1673 unsigned long ring_size = info->nr_ring_pages * XEN_PAGE_SIZE; 1674 grant_ref_t gref[XENBUS_MAX_RING_GRANTS]; 1675 1676 for (i = 0; i < info->nr_ring_pages; i++) 1677 rinfo->ring_ref[i] = GRANT_INVALID_REF; 1678 1679 sring = (struct blkif_sring *)__get_free_pages(GFP_NOIO | __GFP_HIGH, 1680 get_order(ring_size)); 1681 if (!sring) { 1682 xenbus_dev_fatal(dev, -ENOMEM, "allocating shared ring"); 1683 return -ENOMEM; 1684 } 1685 SHARED_RING_INIT(sring); 1686 FRONT_RING_INIT(&rinfo->ring, sring, ring_size); 1687 1688 err = xenbus_grant_ring(dev, rinfo->ring.sring, info->nr_ring_pages, gref); 1689 if (err < 0) { 1690 free_pages((unsigned long)sring, get_order(ring_size)); 1691 rinfo->ring.sring = NULL; 1692 goto fail; 1693 } 1694 for (i = 0; i < info->nr_ring_pages; i++) 1695 rinfo->ring_ref[i] = gref[i]; 1696 1697 err = xenbus_alloc_evtchn(dev, &rinfo->evtchn); 1698 if (err) 1699 goto fail; 1700 1701 err = bind_evtchn_to_irqhandler_lateeoi(rinfo->evtchn, blkif_interrupt, 1702 0, "blkif", rinfo); 1703 if (err <= 0) { 1704 xenbus_dev_fatal(dev, err, 1705 "bind_evtchn_to_irqhandler failed"); 1706 goto fail; 1707 } 1708 rinfo->irq = err; 1709 1710 return 0; 1711 fail: 1712 blkif_free(info, 0); 1713 return err; 1714 } 1715 1716 /* 1717 * Write out per-ring/queue nodes including ring-ref and event-channel, and each 1718 * ring buffer may have multi pages depending on ->nr_ring_pages. 1719 */ 1720 static int write_per_ring_nodes(struct xenbus_transaction xbt, 1721 struct blkfront_ring_info *rinfo, const char *dir) 1722 { 1723 int err; 1724 unsigned int i; 1725 const char *message = NULL; 1726 struct blkfront_info *info = rinfo->dev_info; 1727 1728 if (info->nr_ring_pages == 1) { 1729 err = xenbus_printf(xbt, dir, "ring-ref", "%u", rinfo->ring_ref[0]); 1730 if (err) { 1731 message = "writing ring-ref"; 1732 goto abort_transaction; 1733 } 1734 } else { 1735 for (i = 0; i < info->nr_ring_pages; i++) { 1736 char ring_ref_name[RINGREF_NAME_LEN]; 1737 1738 snprintf(ring_ref_name, RINGREF_NAME_LEN, "ring-ref%u", i); 1739 err = xenbus_printf(xbt, dir, ring_ref_name, 1740 "%u", rinfo->ring_ref[i]); 1741 if (err) { 1742 message = "writing ring-ref"; 1743 goto abort_transaction; 1744 } 1745 } 1746 } 1747 1748 err = xenbus_printf(xbt, dir, "event-channel", "%u", rinfo->evtchn); 1749 if (err) { 1750 message = "writing event-channel"; 1751 goto abort_transaction; 1752 } 1753 1754 return 0; 1755 1756 abort_transaction: 1757 xenbus_transaction_end(xbt, 1); 1758 if (message) 1759 xenbus_dev_fatal(info->xbdev, err, "%s", message); 1760 1761 return err; 1762 } 1763 1764 /* Common code used when first setting up, and when resuming. */ 1765 static int talk_to_blkback(struct xenbus_device *dev, 1766 struct blkfront_info *info) 1767 { 1768 const char *message = NULL; 1769 struct xenbus_transaction xbt; 1770 int err; 1771 unsigned int i, max_page_order; 1772 unsigned int ring_page_order; 1773 struct blkfront_ring_info *rinfo; 1774 1775 if (!info) 1776 return -ENODEV; 1777 1778 max_page_order = xenbus_read_unsigned(info->xbdev->otherend, 1779 "max-ring-page-order", 0); 1780 ring_page_order = min(xen_blkif_max_ring_order, max_page_order); 1781 info->nr_ring_pages = 1 << ring_page_order; 1782 1783 err = negotiate_mq(info); 1784 if (err) 1785 goto destroy_blkring; 1786 1787 for_each_rinfo(info, rinfo, i) { 1788 /* Create shared ring, alloc event channel. */ 1789 err = setup_blkring(dev, rinfo); 1790 if (err) 1791 goto destroy_blkring; 1792 } 1793 1794 again: 1795 err = xenbus_transaction_start(&xbt); 1796 if (err) { 1797 xenbus_dev_fatal(dev, err, "starting transaction"); 1798 goto destroy_blkring; 1799 } 1800 1801 if (info->nr_ring_pages > 1) { 1802 err = xenbus_printf(xbt, dev->nodename, "ring-page-order", "%u", 1803 ring_page_order); 1804 if (err) { 1805 message = "writing ring-page-order"; 1806 goto abort_transaction; 1807 } 1808 } 1809 1810 /* We already got the number of queues/rings in _probe */ 1811 if (info->nr_rings == 1) { 1812 err = write_per_ring_nodes(xbt, info->rinfo, dev->nodename); 1813 if (err) 1814 goto destroy_blkring; 1815 } else { 1816 char *path; 1817 size_t pathsize; 1818 1819 err = xenbus_printf(xbt, dev->nodename, "multi-queue-num-queues", "%u", 1820 info->nr_rings); 1821 if (err) { 1822 message = "writing multi-queue-num-queues"; 1823 goto abort_transaction; 1824 } 1825 1826 pathsize = strlen(dev->nodename) + QUEUE_NAME_LEN; 1827 path = kmalloc(pathsize, GFP_KERNEL); 1828 if (!path) { 1829 err = -ENOMEM; 1830 message = "ENOMEM while writing ring references"; 1831 goto abort_transaction; 1832 } 1833 1834 for_each_rinfo(info, rinfo, i) { 1835 memset(path, 0, pathsize); 1836 snprintf(path, pathsize, "%s/queue-%u", dev->nodename, i); 1837 err = write_per_ring_nodes(xbt, rinfo, path); 1838 if (err) { 1839 kfree(path); 1840 goto destroy_blkring; 1841 } 1842 } 1843 kfree(path); 1844 } 1845 err = xenbus_printf(xbt, dev->nodename, "protocol", "%s", 1846 XEN_IO_PROTO_ABI_NATIVE); 1847 if (err) { 1848 message = "writing protocol"; 1849 goto abort_transaction; 1850 } 1851 err = xenbus_printf(xbt, dev->nodename, "feature-persistent", "%u", 1852 info->feature_persistent); 1853 if (err) 1854 dev_warn(&dev->dev, 1855 "writing persistent grants feature to xenbus"); 1856 1857 err = xenbus_transaction_end(xbt, 0); 1858 if (err) { 1859 if (err == -EAGAIN) 1860 goto again; 1861 xenbus_dev_fatal(dev, err, "completing transaction"); 1862 goto destroy_blkring; 1863 } 1864 1865 for_each_rinfo(info, rinfo, i) { 1866 unsigned int j; 1867 1868 for (j = 0; j < BLK_RING_SIZE(info); j++) 1869 rinfo->shadow[j].req.u.rw.id = j + 1; 1870 rinfo->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff; 1871 } 1872 xenbus_switch_state(dev, XenbusStateInitialised); 1873 1874 return 0; 1875 1876 abort_transaction: 1877 xenbus_transaction_end(xbt, 1); 1878 if (message) 1879 xenbus_dev_fatal(dev, err, "%s", message); 1880 destroy_blkring: 1881 blkif_free(info, 0); 1882 return err; 1883 } 1884 1885 static int negotiate_mq(struct blkfront_info *info) 1886 { 1887 unsigned int backend_max_queues; 1888 unsigned int i; 1889 struct blkfront_ring_info *rinfo; 1890 1891 BUG_ON(info->nr_rings); 1892 1893 /* Check if backend supports multiple queues. */ 1894 backend_max_queues = xenbus_read_unsigned(info->xbdev->otherend, 1895 "multi-queue-max-queues", 1); 1896 info->nr_rings = min(backend_max_queues, xen_blkif_max_queues); 1897 /* We need at least one ring. */ 1898 if (!info->nr_rings) 1899 info->nr_rings = 1; 1900 1901 info->rinfo_size = struct_size(info->rinfo, shadow, 1902 BLK_RING_SIZE(info)); 1903 info->rinfo = kvcalloc(info->nr_rings, info->rinfo_size, GFP_KERNEL); 1904 if (!info->rinfo) { 1905 xenbus_dev_fatal(info->xbdev, -ENOMEM, "allocating ring_info structure"); 1906 info->nr_rings = 0; 1907 return -ENOMEM; 1908 } 1909 1910 for_each_rinfo(info, rinfo, i) { 1911 INIT_LIST_HEAD(&rinfo->indirect_pages); 1912 INIT_LIST_HEAD(&rinfo->grants); 1913 rinfo->dev_info = info; 1914 INIT_WORK(&rinfo->work, blkif_restart_queue); 1915 spin_lock_init(&rinfo->ring_lock); 1916 } 1917 return 0; 1918 } 1919 1920 /* Enable the persistent grants feature. */ 1921 static bool feature_persistent = true; 1922 module_param(feature_persistent, bool, 0644); 1923 MODULE_PARM_DESC(feature_persistent, 1924 "Enables the persistent grants feature"); 1925 1926 /* 1927 * Entry point to this code when a new device is created. Allocate the basic 1928 * structures and the ring buffer for communication with the backend, and 1929 * inform the backend of the appropriate details for those. Switch to 1930 * Initialised state. 1931 */ 1932 static int blkfront_probe(struct xenbus_device *dev, 1933 const struct xenbus_device_id *id) 1934 { 1935 int err, vdevice; 1936 struct blkfront_info *info; 1937 1938 /* FIXME: Use dynamic device id if this is not set. */ 1939 err = xenbus_scanf(XBT_NIL, dev->nodename, 1940 "virtual-device", "%i", &vdevice); 1941 if (err != 1) { 1942 /* go looking in the extended area instead */ 1943 err = xenbus_scanf(XBT_NIL, dev->nodename, "virtual-device-ext", 1944 "%i", &vdevice); 1945 if (err != 1) { 1946 xenbus_dev_fatal(dev, err, "reading virtual-device"); 1947 return err; 1948 } 1949 } 1950 1951 if (xen_hvm_domain()) { 1952 char *type; 1953 int len; 1954 /* no unplug has been done: do not hook devices != xen vbds */ 1955 if (xen_has_pv_and_legacy_disk_devices()) { 1956 int major; 1957 1958 if (!VDEV_IS_EXTENDED(vdevice)) 1959 major = BLKIF_MAJOR(vdevice); 1960 else 1961 major = XENVBD_MAJOR; 1962 1963 if (major != XENVBD_MAJOR) { 1964 printk(KERN_INFO 1965 "%s: HVM does not support vbd %d as xen block device\n", 1966 __func__, vdevice); 1967 return -ENODEV; 1968 } 1969 } 1970 /* do not create a PV cdrom device if we are an HVM guest */ 1971 type = xenbus_read(XBT_NIL, dev->nodename, "device-type", &len); 1972 if (IS_ERR(type)) 1973 return -ENODEV; 1974 if (strncmp(type, "cdrom", 5) == 0) { 1975 kfree(type); 1976 return -ENODEV; 1977 } 1978 kfree(type); 1979 } 1980 info = kzalloc(sizeof(*info), GFP_KERNEL); 1981 if (!info) { 1982 xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure"); 1983 return -ENOMEM; 1984 } 1985 1986 info->xbdev = dev; 1987 1988 mutex_init(&info->mutex); 1989 info->vdevice = vdevice; 1990 info->connected = BLKIF_STATE_DISCONNECTED; 1991 1992 info->feature_persistent = feature_persistent; 1993 1994 /* Front end dir is a number, which is used as the id. */ 1995 info->handle = simple_strtoul(strrchr(dev->nodename, '/')+1, NULL, 0); 1996 dev_set_drvdata(&dev->dev, info); 1997 1998 mutex_lock(&blkfront_mutex); 1999 list_add(&info->info_list, &info_list); 2000 mutex_unlock(&blkfront_mutex); 2001 2002 return 0; 2003 } 2004 2005 static int blkif_recover(struct blkfront_info *info) 2006 { 2007 unsigned int r_index; 2008 struct request *req, *n; 2009 int rc; 2010 struct bio *bio; 2011 unsigned int segs; 2012 struct blkfront_ring_info *rinfo; 2013 2014 blkfront_gather_backend_features(info); 2015 /* Reset limits changed by blk_mq_update_nr_hw_queues(). */ 2016 blkif_set_queue_limits(info); 2017 segs = info->max_indirect_segments ? : BLKIF_MAX_SEGMENTS_PER_REQUEST; 2018 blk_queue_max_segments(info->rq, segs / GRANTS_PER_PSEG); 2019 2020 for_each_rinfo(info, rinfo, r_index) { 2021 rc = blkfront_setup_indirect(rinfo); 2022 if (rc) 2023 return rc; 2024 } 2025 xenbus_switch_state(info->xbdev, XenbusStateConnected); 2026 2027 /* Now safe for us to use the shared ring */ 2028 info->connected = BLKIF_STATE_CONNECTED; 2029 2030 for_each_rinfo(info, rinfo, r_index) { 2031 /* Kick any other new requests queued since we resumed */ 2032 kick_pending_request_queues(rinfo); 2033 } 2034 2035 list_for_each_entry_safe(req, n, &info->requests, queuelist) { 2036 /* Requeue pending requests (flush or discard) */ 2037 list_del_init(&req->queuelist); 2038 BUG_ON(req->nr_phys_segments > segs); 2039 blk_mq_requeue_request(req, false); 2040 } 2041 blk_mq_start_stopped_hw_queues(info->rq, true); 2042 blk_mq_kick_requeue_list(info->rq); 2043 2044 while ((bio = bio_list_pop(&info->bio_list)) != NULL) { 2045 /* Traverse the list of pending bios and re-queue them */ 2046 submit_bio(bio); 2047 } 2048 2049 return 0; 2050 } 2051 2052 /* 2053 * We are reconnecting to the backend, due to a suspend/resume, or a backend 2054 * driver restart. We tear down our blkif structure and recreate it, but 2055 * leave the device-layer structures intact so that this is transparent to the 2056 * rest of the kernel. 2057 */ 2058 static int blkfront_resume(struct xenbus_device *dev) 2059 { 2060 struct blkfront_info *info = dev_get_drvdata(&dev->dev); 2061 int err = 0; 2062 unsigned int i, j; 2063 struct blkfront_ring_info *rinfo; 2064 2065 dev_dbg(&dev->dev, "blkfront_resume: %s\n", dev->nodename); 2066 2067 bio_list_init(&info->bio_list); 2068 INIT_LIST_HEAD(&info->requests); 2069 for_each_rinfo(info, rinfo, i) { 2070 struct bio_list merge_bio; 2071 struct blk_shadow *shadow = rinfo->shadow; 2072 2073 for (j = 0; j < BLK_RING_SIZE(info); j++) { 2074 /* Not in use? */ 2075 if (!shadow[j].request) 2076 continue; 2077 2078 /* 2079 * Get the bios in the request so we can re-queue them. 2080 */ 2081 if (req_op(shadow[j].request) == REQ_OP_FLUSH || 2082 req_op(shadow[j].request) == REQ_OP_DISCARD || 2083 req_op(shadow[j].request) == REQ_OP_SECURE_ERASE || 2084 shadow[j].request->cmd_flags & REQ_FUA) { 2085 /* 2086 * Flush operations don't contain bios, so 2087 * we need to requeue the whole request 2088 * 2089 * XXX: but this doesn't make any sense for a 2090 * write with the FUA flag set.. 2091 */ 2092 list_add(&shadow[j].request->queuelist, &info->requests); 2093 continue; 2094 } 2095 merge_bio.head = shadow[j].request->bio; 2096 merge_bio.tail = shadow[j].request->biotail; 2097 bio_list_merge(&info->bio_list, &merge_bio); 2098 shadow[j].request->bio = NULL; 2099 blk_mq_end_request(shadow[j].request, BLK_STS_OK); 2100 } 2101 } 2102 2103 blkif_free(info, info->connected == BLKIF_STATE_CONNECTED); 2104 2105 err = talk_to_blkback(dev, info); 2106 if (!err) 2107 blk_mq_update_nr_hw_queues(&info->tag_set, info->nr_rings); 2108 2109 /* 2110 * We have to wait for the backend to switch to 2111 * connected state, since we want to read which 2112 * features it supports. 2113 */ 2114 2115 return err; 2116 } 2117 2118 static void blkfront_closing(struct blkfront_info *info) 2119 { 2120 struct xenbus_device *xbdev = info->xbdev; 2121 struct blkfront_ring_info *rinfo; 2122 unsigned int i; 2123 2124 if (xbdev->state == XenbusStateClosing) 2125 return; 2126 2127 /* No more blkif_request(). */ 2128 blk_mq_stop_hw_queues(info->rq); 2129 blk_set_queue_dying(info->rq); 2130 set_capacity(info->gd, 0); 2131 2132 for_each_rinfo(info, rinfo, i) { 2133 /* No more gnttab callback work. */ 2134 gnttab_cancel_free_callback(&rinfo->callback); 2135 2136 /* Flush gnttab callback work. Must be done with no locks held. */ 2137 flush_work(&rinfo->work); 2138 } 2139 2140 xenbus_frontend_closed(xbdev); 2141 } 2142 2143 static void blkfront_setup_discard(struct blkfront_info *info) 2144 { 2145 info->feature_discard = 1; 2146 info->discard_granularity = xenbus_read_unsigned(info->xbdev->otherend, 2147 "discard-granularity", 2148 0); 2149 info->discard_alignment = xenbus_read_unsigned(info->xbdev->otherend, 2150 "discard-alignment", 0); 2151 info->feature_secdiscard = 2152 !!xenbus_read_unsigned(info->xbdev->otherend, "discard-secure", 2153 0); 2154 } 2155 2156 static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo) 2157 { 2158 unsigned int psegs, grants, memflags; 2159 int err, i; 2160 struct blkfront_info *info = rinfo->dev_info; 2161 2162 memflags = memalloc_noio_save(); 2163 2164 if (info->max_indirect_segments == 0) { 2165 if (!HAS_EXTRA_REQ) 2166 grants = BLKIF_MAX_SEGMENTS_PER_REQUEST; 2167 else { 2168 /* 2169 * When an extra req is required, the maximum 2170 * grants supported is related to the size of the 2171 * Linux block segment. 2172 */ 2173 grants = GRANTS_PER_PSEG; 2174 } 2175 } 2176 else 2177 grants = info->max_indirect_segments; 2178 psegs = DIV_ROUND_UP(grants, GRANTS_PER_PSEG); 2179 2180 err = fill_grant_buffer(rinfo, 2181 (grants + INDIRECT_GREFS(grants)) * BLK_RING_SIZE(info)); 2182 if (err) 2183 goto out_of_memory; 2184 2185 if (!info->feature_persistent && info->max_indirect_segments) { 2186 /* 2187 * We are using indirect descriptors but not persistent 2188 * grants, we need to allocate a set of pages that can be 2189 * used for mapping indirect grefs 2190 */ 2191 int num = INDIRECT_GREFS(grants) * BLK_RING_SIZE(info); 2192 2193 BUG_ON(!list_empty(&rinfo->indirect_pages)); 2194 for (i = 0; i < num; i++) { 2195 struct page *indirect_page = alloc_page(GFP_KERNEL); 2196 if (!indirect_page) 2197 goto out_of_memory; 2198 list_add(&indirect_page->lru, &rinfo->indirect_pages); 2199 } 2200 } 2201 2202 for (i = 0; i < BLK_RING_SIZE(info); i++) { 2203 rinfo->shadow[i].grants_used = 2204 kvcalloc(grants, 2205 sizeof(rinfo->shadow[i].grants_used[0]), 2206 GFP_KERNEL); 2207 rinfo->shadow[i].sg = kvcalloc(psegs, 2208 sizeof(rinfo->shadow[i].sg[0]), 2209 GFP_KERNEL); 2210 if (info->max_indirect_segments) 2211 rinfo->shadow[i].indirect_grants = 2212 kvcalloc(INDIRECT_GREFS(grants), 2213 sizeof(rinfo->shadow[i].indirect_grants[0]), 2214 GFP_KERNEL); 2215 if ((rinfo->shadow[i].grants_used == NULL) || 2216 (rinfo->shadow[i].sg == NULL) || 2217 (info->max_indirect_segments && 2218 (rinfo->shadow[i].indirect_grants == NULL))) 2219 goto out_of_memory; 2220 sg_init_table(rinfo->shadow[i].sg, psegs); 2221 } 2222 2223 memalloc_noio_restore(memflags); 2224 2225 return 0; 2226 2227 out_of_memory: 2228 for (i = 0; i < BLK_RING_SIZE(info); i++) { 2229 kvfree(rinfo->shadow[i].grants_used); 2230 rinfo->shadow[i].grants_used = NULL; 2231 kvfree(rinfo->shadow[i].sg); 2232 rinfo->shadow[i].sg = NULL; 2233 kvfree(rinfo->shadow[i].indirect_grants); 2234 rinfo->shadow[i].indirect_grants = NULL; 2235 } 2236 if (!list_empty(&rinfo->indirect_pages)) { 2237 struct page *indirect_page, *n; 2238 list_for_each_entry_safe(indirect_page, n, &rinfo->indirect_pages, lru) { 2239 list_del(&indirect_page->lru); 2240 __free_page(indirect_page); 2241 } 2242 } 2243 2244 memalloc_noio_restore(memflags); 2245 2246 return -ENOMEM; 2247 } 2248 2249 /* 2250 * Gather all backend feature-* 2251 */ 2252 static void blkfront_gather_backend_features(struct blkfront_info *info) 2253 { 2254 unsigned int indirect_segments; 2255 2256 info->feature_flush = 0; 2257 info->feature_fua = 0; 2258 2259 /* 2260 * If there's no "feature-barrier" defined, then it means 2261 * we're dealing with a very old backend which writes 2262 * synchronously; nothing to do. 2263 * 2264 * If there are barriers, then we use flush. 2265 */ 2266 if (xenbus_read_unsigned(info->xbdev->otherend, "feature-barrier", 0)) { 2267 info->feature_flush = 1; 2268 info->feature_fua = 1; 2269 } 2270 2271 /* 2272 * And if there is "feature-flush-cache" use that above 2273 * barriers. 2274 */ 2275 if (xenbus_read_unsigned(info->xbdev->otherend, "feature-flush-cache", 2276 0)) { 2277 info->feature_flush = 1; 2278 info->feature_fua = 0; 2279 } 2280 2281 if (xenbus_read_unsigned(info->xbdev->otherend, "feature-discard", 0)) 2282 blkfront_setup_discard(info); 2283 2284 if (info->feature_persistent) 2285 info->feature_persistent = 2286 !!xenbus_read_unsigned(info->xbdev->otherend, 2287 "feature-persistent", 0); 2288 2289 indirect_segments = xenbus_read_unsigned(info->xbdev->otherend, 2290 "feature-max-indirect-segments", 0); 2291 if (indirect_segments > xen_blkif_max_segments) 2292 indirect_segments = xen_blkif_max_segments; 2293 if (indirect_segments <= BLKIF_MAX_SEGMENTS_PER_REQUEST) 2294 indirect_segments = 0; 2295 info->max_indirect_segments = indirect_segments; 2296 2297 if (info->feature_persistent) { 2298 mutex_lock(&blkfront_mutex); 2299 schedule_delayed_work(&blkfront_work, HZ * 10); 2300 mutex_unlock(&blkfront_mutex); 2301 } 2302 } 2303 2304 /* 2305 * Invoked when the backend is finally 'ready' (and has told produced 2306 * the details about the physical device - #sectors, size, etc). 2307 */ 2308 static void blkfront_connect(struct blkfront_info *info) 2309 { 2310 unsigned long long sectors; 2311 unsigned long sector_size; 2312 unsigned int physical_sector_size; 2313 int err, i; 2314 struct blkfront_ring_info *rinfo; 2315 2316 switch (info->connected) { 2317 case BLKIF_STATE_CONNECTED: 2318 /* 2319 * Potentially, the back-end may be signalling 2320 * a capacity change; update the capacity. 2321 */ 2322 err = xenbus_scanf(XBT_NIL, info->xbdev->otherend, 2323 "sectors", "%Lu", §ors); 2324 if (XENBUS_EXIST_ERR(err)) 2325 return; 2326 printk(KERN_INFO "Setting capacity to %Lu\n", 2327 sectors); 2328 set_capacity_and_notify(info->gd, sectors); 2329 2330 return; 2331 case BLKIF_STATE_SUSPENDED: 2332 /* 2333 * If we are recovering from suspension, we need to wait 2334 * for the backend to announce it's features before 2335 * reconnecting, at least we need to know if the backend 2336 * supports indirect descriptors, and how many. 2337 */ 2338 blkif_recover(info); 2339 return; 2340 2341 default: 2342 break; 2343 } 2344 2345 dev_dbg(&info->xbdev->dev, "%s:%s.\n", 2346 __func__, info->xbdev->otherend); 2347 2348 err = xenbus_gather(XBT_NIL, info->xbdev->otherend, 2349 "sectors", "%llu", §ors, 2350 "info", "%u", &info->vdisk_info, 2351 "sector-size", "%lu", §or_size, 2352 NULL); 2353 if (err) { 2354 xenbus_dev_fatal(info->xbdev, err, 2355 "reading backend fields at %s", 2356 info->xbdev->otherend); 2357 return; 2358 } 2359 2360 /* 2361 * physical-sector-size is a newer field, so old backends may not 2362 * provide this. Assume physical sector size to be the same as 2363 * sector_size in that case. 2364 */ 2365 physical_sector_size = xenbus_read_unsigned(info->xbdev->otherend, 2366 "physical-sector-size", 2367 sector_size); 2368 blkfront_gather_backend_features(info); 2369 for_each_rinfo(info, rinfo, i) { 2370 err = blkfront_setup_indirect(rinfo); 2371 if (err) { 2372 xenbus_dev_fatal(info->xbdev, err, "setup_indirect at %s", 2373 info->xbdev->otherend); 2374 blkif_free(info, 0); 2375 break; 2376 } 2377 } 2378 2379 err = xlvbd_alloc_gendisk(sectors, info, sector_size, 2380 physical_sector_size); 2381 if (err) { 2382 xenbus_dev_fatal(info->xbdev, err, "xlvbd_add at %s", 2383 info->xbdev->otherend); 2384 goto fail; 2385 } 2386 2387 xenbus_switch_state(info->xbdev, XenbusStateConnected); 2388 2389 /* Kick pending requests. */ 2390 info->connected = BLKIF_STATE_CONNECTED; 2391 for_each_rinfo(info, rinfo, i) 2392 kick_pending_request_queues(rinfo); 2393 2394 err = device_add_disk(&info->xbdev->dev, info->gd, NULL); 2395 if (err) { 2396 blk_cleanup_disk(info->gd); 2397 blk_mq_free_tag_set(&info->tag_set); 2398 info->rq = NULL; 2399 goto fail; 2400 } 2401 2402 info->is_ready = 1; 2403 return; 2404 2405 fail: 2406 blkif_free(info, 0); 2407 return; 2408 } 2409 2410 /* 2411 * Callback received when the backend's state changes. 2412 */ 2413 static void blkback_changed(struct xenbus_device *dev, 2414 enum xenbus_state backend_state) 2415 { 2416 struct blkfront_info *info = dev_get_drvdata(&dev->dev); 2417 2418 dev_dbg(&dev->dev, "blkfront:blkback_changed to state %d.\n", backend_state); 2419 2420 switch (backend_state) { 2421 case XenbusStateInitWait: 2422 if (dev->state != XenbusStateInitialising) 2423 break; 2424 if (talk_to_blkback(dev, info)) 2425 break; 2426 break; 2427 case XenbusStateInitialising: 2428 case XenbusStateInitialised: 2429 case XenbusStateReconfiguring: 2430 case XenbusStateReconfigured: 2431 case XenbusStateUnknown: 2432 break; 2433 2434 case XenbusStateConnected: 2435 /* 2436 * talk_to_blkback sets state to XenbusStateInitialised 2437 * and blkfront_connect sets it to XenbusStateConnected 2438 * (if connection went OK). 2439 * 2440 * If the backend (or toolstack) decides to poke at backend 2441 * state (and re-trigger the watch by setting the state repeatedly 2442 * to XenbusStateConnected (4)) we need to deal with this. 2443 * This is allowed as this is used to communicate to the guest 2444 * that the size of disk has changed! 2445 */ 2446 if ((dev->state != XenbusStateInitialised) && 2447 (dev->state != XenbusStateConnected)) { 2448 if (talk_to_blkback(dev, info)) 2449 break; 2450 } 2451 2452 blkfront_connect(info); 2453 break; 2454 2455 case XenbusStateClosed: 2456 if (dev->state == XenbusStateClosed) 2457 break; 2458 fallthrough; 2459 case XenbusStateClosing: 2460 blkfront_closing(info); 2461 break; 2462 } 2463 } 2464 2465 static int blkfront_remove(struct xenbus_device *xbdev) 2466 { 2467 struct blkfront_info *info = dev_get_drvdata(&xbdev->dev); 2468 2469 dev_dbg(&xbdev->dev, "%s removed", xbdev->nodename); 2470 2471 del_gendisk(info->gd); 2472 2473 mutex_lock(&blkfront_mutex); 2474 list_del(&info->info_list); 2475 mutex_unlock(&blkfront_mutex); 2476 2477 blkif_free(info, 0); 2478 xlbd_release_minors(info->gd->first_minor, info->gd->minors); 2479 blk_cleanup_disk(info->gd); 2480 blk_mq_free_tag_set(&info->tag_set); 2481 2482 kfree(info); 2483 return 0; 2484 } 2485 2486 static int blkfront_is_ready(struct xenbus_device *dev) 2487 { 2488 struct blkfront_info *info = dev_get_drvdata(&dev->dev); 2489 2490 return info->is_ready && info->xbdev; 2491 } 2492 2493 static const struct block_device_operations xlvbd_block_fops = 2494 { 2495 .owner = THIS_MODULE, 2496 .getgeo = blkif_getgeo, 2497 .ioctl = blkif_ioctl, 2498 .compat_ioctl = blkdev_compat_ptr_ioctl, 2499 }; 2500 2501 2502 static const struct xenbus_device_id blkfront_ids[] = { 2503 { "vbd" }, 2504 { "" } 2505 }; 2506 2507 static struct xenbus_driver blkfront_driver = { 2508 .ids = blkfront_ids, 2509 .probe = blkfront_probe, 2510 .remove = blkfront_remove, 2511 .resume = blkfront_resume, 2512 .otherend_changed = blkback_changed, 2513 .is_ready = blkfront_is_ready, 2514 }; 2515 2516 static void purge_persistent_grants(struct blkfront_info *info) 2517 { 2518 unsigned int i; 2519 unsigned long flags; 2520 struct blkfront_ring_info *rinfo; 2521 2522 for_each_rinfo(info, rinfo, i) { 2523 struct grant *gnt_list_entry, *tmp; 2524 2525 spin_lock_irqsave(&rinfo->ring_lock, flags); 2526 2527 if (rinfo->persistent_gnts_c == 0) { 2528 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 2529 continue; 2530 } 2531 2532 list_for_each_entry_safe(gnt_list_entry, tmp, &rinfo->grants, 2533 node) { 2534 if (gnt_list_entry->gref == GRANT_INVALID_REF || 2535 gnttab_query_foreign_access(gnt_list_entry->gref)) 2536 continue; 2537 2538 list_del(&gnt_list_entry->node); 2539 gnttab_end_foreign_access(gnt_list_entry->gref, 0, 0UL); 2540 rinfo->persistent_gnts_c--; 2541 gnt_list_entry->gref = GRANT_INVALID_REF; 2542 list_add_tail(&gnt_list_entry->node, &rinfo->grants); 2543 } 2544 2545 spin_unlock_irqrestore(&rinfo->ring_lock, flags); 2546 } 2547 } 2548 2549 static void blkfront_delay_work(struct work_struct *work) 2550 { 2551 struct blkfront_info *info; 2552 bool need_schedule_work = false; 2553 2554 mutex_lock(&blkfront_mutex); 2555 2556 list_for_each_entry(info, &info_list, info_list) { 2557 if (info->feature_persistent) { 2558 need_schedule_work = true; 2559 mutex_lock(&info->mutex); 2560 purge_persistent_grants(info); 2561 mutex_unlock(&info->mutex); 2562 } 2563 } 2564 2565 if (need_schedule_work) 2566 schedule_delayed_work(&blkfront_work, HZ * 10); 2567 2568 mutex_unlock(&blkfront_mutex); 2569 } 2570 2571 static int __init xlblk_init(void) 2572 { 2573 int ret; 2574 int nr_cpus = num_online_cpus(); 2575 2576 if (!xen_domain()) 2577 return -ENODEV; 2578 2579 if (!xen_has_pv_disk_devices()) 2580 return -ENODEV; 2581 2582 if (register_blkdev(XENVBD_MAJOR, DEV_NAME)) { 2583 pr_warn("xen_blk: can't get major %d with name %s\n", 2584 XENVBD_MAJOR, DEV_NAME); 2585 return -ENODEV; 2586 } 2587 2588 if (xen_blkif_max_segments < BLKIF_MAX_SEGMENTS_PER_REQUEST) 2589 xen_blkif_max_segments = BLKIF_MAX_SEGMENTS_PER_REQUEST; 2590 2591 if (xen_blkif_max_ring_order > XENBUS_MAX_RING_GRANT_ORDER) { 2592 pr_info("Invalid max_ring_order (%d), will use default max: %d.\n", 2593 xen_blkif_max_ring_order, XENBUS_MAX_RING_GRANT_ORDER); 2594 xen_blkif_max_ring_order = XENBUS_MAX_RING_GRANT_ORDER; 2595 } 2596 2597 if (xen_blkif_max_queues > nr_cpus) { 2598 pr_info("Invalid max_queues (%d), will use default max: %d.\n", 2599 xen_blkif_max_queues, nr_cpus); 2600 xen_blkif_max_queues = nr_cpus; 2601 } 2602 2603 INIT_DELAYED_WORK(&blkfront_work, blkfront_delay_work); 2604 2605 ret = xenbus_register_frontend(&blkfront_driver); 2606 if (ret) { 2607 unregister_blkdev(XENVBD_MAJOR, DEV_NAME); 2608 return ret; 2609 } 2610 2611 return 0; 2612 } 2613 module_init(xlblk_init); 2614 2615 2616 static void __exit xlblk_exit(void) 2617 { 2618 cancel_delayed_work_sync(&blkfront_work); 2619 2620 xenbus_unregister_driver(&blkfront_driver); 2621 unregister_blkdev(XENVBD_MAJOR, DEV_NAME); 2622 kfree(minors); 2623 } 2624 module_exit(xlblk_exit); 2625 2626 MODULE_DESCRIPTION("Xen virtual block device frontend"); 2627 MODULE_LICENSE("GPL"); 2628 MODULE_ALIAS_BLOCKDEV_MAJOR(XENVBD_MAJOR); 2629 MODULE_ALIAS("xen:vbd"); 2630 MODULE_ALIAS("xenblk"); 2631