drivers/block/xen-blkfront.c

   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/mutex.h>
  46 #include <linux/scatterlist.h>
  47 #include <linux/bitmap.h>
  48 #include <linux/list.h>
  49
  50 #include <xen/xen.h>
  51 #include <xen/xenbus.h>
  52 #include <xen/grant_table.h>
  53 #include <xen/events.h>
  54 #include <xen/page.h>
  55 #include <xen/platform_pci.h>
  56
  57 #include <xen/interface/grant_table.h>
  58 #include <xen/interface/io/blkif.h>
  59 #include <xen/interface/io/protocols.h>
  60
  61 #include <asm/xen/hypervisor.h>
  62
  63 /*
  64  * The minimal size of segment supported by the block framework is PAGE_SIZE.
  65  * When Linux is using a different page size than Xen, it may not be possible
  66  * to put all the data in a single segment.
  67  * This can happen when the backend doesn't support indirect descriptor and
  68  * therefore the maximum amount of data that a request can carry is
  69  * BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE = 44KB
  70  *
  71  * Note that we only support one extra request. So the Linux page size
  72  * should be <= ( 2 * BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE) =
  73  * 88KB.
  74  */
  75 #define HAS_EXTRA_REQ (BLKIF_MAX_SEGMENTS_PER_REQUEST < XEN_PFN_PER_PAGE)
  76
  77 enum blkif_state {
  78         BLKIF_STATE_DISCONNECTED,
  79         BLKIF_STATE_CONNECTED,
  80         BLKIF_STATE_SUSPENDED,
  81 };
  82
  83 struct grant {
  84         grant_ref_t gref;
  85         struct page *page;
  86         struct list_head node;
  87 };
  88
  89 enum blk_req_status {
  90         REQ_WAITING,
  91         REQ_DONE,
  92         REQ_ERROR,
  93         REQ_EOPNOTSUPP,
  94 };
  95
  96 struct blk_shadow {
  97         struct blkif_request req;
  98         struct request *request;
  99         struct grant **grants_used;
 100         struct grant **indirect_grants;
 101         struct scatterlist *sg;
 102         unsigned int num_sg;
 103         enum blk_req_status status;
 104
 105         #define NO_ASSOCIATED_ID ~0UL
 106         /*
 107          * Id of the sibling if we ever need 2 requests when handling a
 108          * block I/O request
 109          */
 110         unsigned long associated_id;
 111 };
 112
 113 struct split_bio {
 114         struct bio *bio;
 115         atomic_t pending;
 116 };
 117
 118 static DEFINE_MUTEX(blkfront_mutex);
 119 static const struct block_device_operations xlvbd_block_fops;
 120
 121 /*
 122  * Maximum number of segments in indirect requests, the actual value used by
 123  * the frontend driver is the minimum of this value and the value provided
 124  * by the backend driver.
 125  */
 126
 127 static unsigned int xen_blkif_max_segments = 32;
 128 module_param_named(max_indirect_segments, xen_blkif_max_segments, uint,
 129                    S_IRUGO);
 130 MODULE_PARM_DESC(max_indirect_segments,
 131                  "Maximum amount of segments in indirect requests (default is 32)");
 132
 133 static unsigned int xen_blkif_max_queues = 4;
 134 module_param_named(max_queues, xen_blkif_max_queues, uint, S_IRUGO);
 135 MODULE_PARM_DESC(max_queues, "Maximum number of hardware queues/rings used per virtual disk");
 136
 137 /*
 138  * Maximum order of pages to be used for the shared ring between front and
 139  * backend, 4KB page granularity is used.
 140  */
 141 static unsigned int xen_blkif_max_ring_order;
 142 module_param_named(max_ring_page_order, xen_blkif_max_ring_order, int, S_IRUGO);
 143 MODULE_PARM_DESC(max_ring_page_order, "Maximum order of pages to be used for the shared ring");
 144
 145 #define BLK_RING_SIZE(info)     \
 146         __CONST_RING_SIZE(blkif, XEN_PAGE_SIZE * (info)->nr_ring_pages)
 147
 148 #define BLK_MAX_RING_SIZE       \
 149         __CONST_RING_SIZE(blkif, XEN_PAGE_SIZE * XENBUS_MAX_RING_GRANTS)
 150
 151 /*
 152  * ring-ref%u i=(-1UL) would take 11 characters + 'ring-ref' is 8, so 19
 153  * characters are enough. Define to 20 to keep consistent with backend.
 154  */
 155 #define RINGREF_NAME_LEN (20)
 156 /*
 157  * queue-%u would take 7 + 10(UINT_MAX) = 17 characters.
 158  */
 159 #define QUEUE_NAME_LEN (17)
 160
 161 /*
 162  *  Per-ring info.
 163  *  Every blkfront device can associate with one or more blkfront_ring_info,
 164  *  depending on how many hardware queues/rings to be used.
 165  */
 166 struct blkfront_ring_info {
 167         /* Lock to protect data in every ring buffer. */
 168         spinlock_t ring_lock;
 169         struct blkif_front_ring ring;
 170         unsigned int ring_ref[XENBUS_MAX_RING_GRANTS];
 171         unsigned int evtchn, irq;
 172         struct work_struct work;
 173         struct gnttab_free_callback callback;
 174         struct blk_shadow shadow[BLK_MAX_RING_SIZE];
 175         struct list_head indirect_pages;
 176         struct list_head grants;
 177         unsigned int persistent_gnts_c;
 178         unsigned long shadow_free;
 179         struct blkfront_info *dev_info;
 180 };
 181
 182 /*
 183  * We have one of these per vbd, whether ide, scsi or 'other'.  They
 184  * hang in private_data off the gendisk structure. We may end up
 185  * putting all kinds of interesting stuff here :-)
 186  */
 187 struct blkfront_info
 188 {
 189         struct mutex mutex;
 190         struct xenbus_device *xbdev;
 191         struct gendisk *gd;
 192         int vdevice;
 193         blkif_vdev_t handle;
 194         enum blkif_state connected;
 195         /* Number of pages per ring buffer. */
 196         unsigned int nr_ring_pages;
 197         struct request_queue *rq;
 198         unsigned int feature_flush;
 199         unsigned int feature_discard:1;
 200         unsigned int feature_secdiscard:1;
 201         unsigned int discard_granularity;
 202         unsigned int discard_alignment;
 203         unsigned int feature_persistent:1;
 204         /* Number of 4KB segments handled */
 205         unsigned int max_indirect_segments;
 206         int is_ready;
 207         struct blk_mq_tag_set tag_set;
 208         struct blkfront_ring_info *rinfo;
 209         unsigned int nr_rings;
 210 };
 211
 212 static unsigned int nr_minors;
 213 static unsigned long *minors;
 214 static DEFINE_SPINLOCK(minor_lock);
 215
 216 #define GRANT_INVALID_REF       0
 217
 218 #define PARTS_PER_DISK          16
 219 #define PARTS_PER_EXT_DISK      256
 220
 221 #define BLKIF_MAJOR(dev) ((dev)>>8)
 222 #define BLKIF_MINOR(dev) ((dev) & 0xff)
 223
 224 #define EXT_SHIFT 28
 225 #define EXTENDED (1<<EXT_SHIFT)
 226 #define VDEV_IS_EXTENDED(dev) ((dev)&(EXTENDED))
 227 #define BLKIF_MINOR_EXT(dev) ((dev)&(~EXTENDED))
 228 #define EMULATED_HD_DISK_MINOR_OFFSET (0)
 229 #define EMULATED_HD_DISK_NAME_OFFSET (EMULATED_HD_DISK_MINOR_OFFSET / 256)
 230 #define EMULATED_SD_DISK_MINOR_OFFSET (0)
 231 #define EMULATED_SD_DISK_NAME_OFFSET (EMULATED_SD_DISK_MINOR_OFFSET / 256)
 232
 233 #define DEV_NAME        "xvd"   /* name in /dev */
 234
 235 /*
 236  * Grants are always the same size as a Xen page (i.e 4KB).
 237  * A physical segment is always the same size as a Linux page.
 238  * Number of grants per physical segment
 239  */
 240 #define GRANTS_PER_PSEG (PAGE_SIZE / XEN_PAGE_SIZE)
 241
 242 #define GRANTS_PER_INDIRECT_FRAME \
 243         (XEN_PAGE_SIZE / sizeof(struct blkif_request_segment))
 244
 245 #define PSEGS_PER_INDIRECT_FRAME        \
 246         (GRANTS_INDIRECT_FRAME / GRANTS_PSEGS)
 247
 248 #define INDIRECT_GREFS(_grants)         \
 249         DIV_ROUND_UP(_grants, GRANTS_PER_INDIRECT_FRAME)
 250
 251 #define GREFS(_psegs)   ((_psegs) * GRANTS_PER_PSEG)
 252
 253 static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo);
 254 static void blkfront_gather_backend_features(struct blkfront_info *info);
 255
 256 static int get_id_from_freelist(struct blkfront_ring_info *rinfo)
 257 {
 258         unsigned long free = rinfo->shadow_free;
 259
 260         BUG_ON(free >= BLK_RING_SIZE(rinfo->dev_info));
 261         rinfo->shadow_free = rinfo->shadow[free].req.u.rw.id;
 262         rinfo->shadow[free].req.u.rw.id = 0x0fffffee; /* debug */
 263         return free;
 264 }
 265
 266 static int add_id_to_freelist(struct blkfront_ring_info *rinfo,
 267                               unsigned long id)
 268 {
 269         if (rinfo->shadow[id].req.u.rw.id != id)
 270                 return -EINVAL;
 271         if (rinfo->shadow[id].request == NULL)
 272                 return -EINVAL;
 273         rinfo->shadow[id].req.u.rw.id  = rinfo->shadow_free;
 274         rinfo->shadow[id].request = NULL;
 275         rinfo->shadow_free = id;
 276         return 0;
 277 }
 278
 279 static int fill_grant_buffer(struct blkfront_ring_info *rinfo, int num)
 280 {
 281         struct blkfront_info *info = rinfo->dev_info;
 282         struct page *granted_page;
 283         struct grant *gnt_list_entry, *n;
 284         int i = 0;
 285
 286         while (i < num) {
 287                 gnt_list_entry = kzalloc(sizeof(struct grant), GFP_NOIO);
 288                 if (!gnt_list_entry)
 289                         goto out_of_memory;
 290
 291                 if (info->feature_persistent) {
 292                         granted_page = alloc_page(GFP_NOIO);
 293                         if (!granted_page) {
 294                                 kfree(gnt_list_entry);
 295                                 goto out_of_memory;
 296                         }
 297                         gnt_list_entry->page = granted_page;
 298                 }
 299
 300                 gnt_list_entry->gref = GRANT_INVALID_REF;
 301                 list_add(&gnt_list_entry->node, &rinfo->grants);
 302                 i++;
 303         }
 304
 305         return 0;
 306
 307 out_of_memory:
 308         list_for_each_entry_safe(gnt_list_entry, n,
 309                                  &rinfo->grants, node) {
 310                 list_del(&gnt_list_entry->node);
 311                 if (info->feature_persistent)
 312                         __free_page(gnt_list_entry->page);
 313                 kfree(gnt_list_entry);
 314                 i--;
 315         }
 316         BUG_ON(i != 0);
 317         return -ENOMEM;
 318 }
 319
 320 static struct grant *get_free_grant(struct blkfront_ring_info *rinfo)
 321 {
 322         struct grant *gnt_list_entry;
 323
 324         BUG_ON(list_empty(&rinfo->grants));
 325         gnt_list_entry = list_first_entry(&rinfo->grants, struct grant,
 326                                           node);
 327         list_del(&gnt_list_entry->node);
 328
 329         if (gnt_list_entry->gref != GRANT_INVALID_REF)
 330                 rinfo->persistent_gnts_c--;
 331
 332         return gnt_list_entry;
 333 }
 334
 335 static inline void grant_foreign_access(const struct grant *gnt_list_entry,
 336                                         const struct blkfront_info *info)
 337 {
 338         gnttab_page_grant_foreign_access_ref_one(gnt_list_entry->gref,
 339                                                  info->xbdev->otherend_id,
 340                                                  gnt_list_entry->page,
 341                                                  0);
 342 }
 343
 344 static struct grant *get_grant(grant_ref_t *gref_head,
 345                                unsigned long gfn,
 346                                struct blkfront_ring_info *rinfo)
 347 {
 348         struct grant *gnt_list_entry = get_free_grant(rinfo);
 349         struct blkfront_info *info = rinfo->dev_info;
 350
 351         if (gnt_list_entry->gref != GRANT_INVALID_REF)
 352                 return gnt_list_entry;
 353
 354         /* Assign a gref to this page */
 355         gnt_list_entry->gref = gnttab_claim_grant_reference(gref_head);
 356         BUG_ON(gnt_list_entry->gref == -ENOSPC);
 357         if (info->feature_persistent)
 358                 grant_foreign_access(gnt_list_entry, info);
 359         else {
 360                 /* Grant access to the GFN passed by the caller */
 361                 gnttab_grant_foreign_access_ref(gnt_list_entry->gref,
 362                                                 info->xbdev->otherend_id,
 363                                                 gfn, 0);
 364         }
 365
 366         return gnt_list_entry;
 367 }
 368
 369 static struct grant *get_indirect_grant(grant_ref_t *gref_head,
 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                 struct page *indirect_page;
 383
 384                 /* Fetch a pre-allocated page to use for indirect grefs */
 385                 BUG_ON(list_empty(&rinfo->indirect_pages));
 386                 indirect_page = list_first_entry(&rinfo->indirect_pages,
 387                                                  struct page, lru);
 388                 list_del(&indirect_page->lru);
 389                 gnt_list_entry->page = indirect_page;
 390         }
 391         grant_foreign_access(gnt_list_entry, info);
 392
 393         return gnt_list_entry;
 394 }
 395
 396 static const char *op_name(int op)
 397 {
 398         static const char *const names[] = {
 399                 [BLKIF_OP_READ] = "read",
 400                 [BLKIF_OP_WRITE] = "write",
 401                 [BLKIF_OP_WRITE_BARRIER] = "barrier",
 402                 [BLKIF_OP_FLUSH_DISKCACHE] = "flush",
 403                 [BLKIF_OP_DISCARD] = "discard" };
 404
 405         if (op < 0 || op >= ARRAY_SIZE(names))
 406                 return "unknown";
 407
 408         if (!names[op])
 409                 return "reserved";
 410
 411         return names[op];
 412 }
 413 static int xlbd_reserve_minors(unsigned int minor, unsigned int nr)
 414 {
 415         unsigned int end = minor + nr;
 416         int rc;
 417
 418         if (end > nr_minors) {
 419                 unsigned long *bitmap, *old;
 420
 421                 bitmap = kcalloc(BITS_TO_LONGS(end), sizeof(*bitmap),
 422                                  GFP_KERNEL);
 423                 if (bitmap == NULL)
 424                         return -ENOMEM;
 425
 426                 spin_lock(&minor_lock);
 427                 if (end > nr_minors) {
 428                         old = minors;
 429                         memcpy(bitmap, minors,
 430                                BITS_TO_LONGS(nr_minors) * sizeof(*bitmap));
 431                         minors = bitmap;
 432                         nr_minors = BITS_TO_LONGS(end) * BITS_PER_LONG;
 433                 } else
 434                         old = bitmap;
 435                 spin_unlock(&minor_lock);
 436                 kfree(old);
 437         }
 438
 439         spin_lock(&minor_lock);
 440         if (find_next_bit(minors, end, minor) >= end) {
 441                 bitmap_set(minors, minor, nr);
 442                 rc = 0;
 443         } else
 444                 rc = -EBUSY;
 445         spin_unlock(&minor_lock);
 446
 447         return rc;
 448 }
 449
 450 static void xlbd_release_minors(unsigned int minor, unsigned int nr)
 451 {
 452         unsigned int end = minor + nr;
 453
 454         BUG_ON(end > nr_minors);
 455         spin_lock(&minor_lock);
 456         bitmap_clear(minors,  minor, nr);
 457         spin_unlock(&minor_lock);
 458 }
 459
 460 static void blkif_restart_queue_callback(void *arg)
 461 {
 462         struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)arg;
 463         schedule_work(&rinfo->work);
 464 }
 465
 466 static int blkif_getgeo(struct block_device *bd, struct hd_geometry *hg)
 467 {
 468         /* We don't have real geometry info, but let's at least return
 469            values consistent with the size of the device */
 470         sector_t nsect = get_capacity(bd->bd_disk);
 471         sector_t cylinders = nsect;
 472
 473         hg->heads = 0xff;
 474         hg->sectors = 0x3f;
 475         sector_div(cylinders, hg->heads * hg->sectors);
 476         hg->cylinders = cylinders;
 477         if ((sector_t)(hg->cylinders + 1) * hg->heads * hg->sectors < nsect)
 478                 hg->cylinders = 0xffff;
 479         return 0;
 480 }
 481
 482 static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
 483                        unsigned command, unsigned long argument)
 484 {
 485         struct blkfront_info *info = bdev->bd_disk->private_data;
 486         int i;
 487
 488         dev_dbg(&info->xbdev->dev, "command: 0x%x, argument: 0x%lx\n",
 489                 command, (long)argument);
 490
 491         switch (command) {
 492         case CDROMMULTISESSION:
 493                 dev_dbg(&info->xbdev->dev, "FIXME: support multisession CDs later\n");
 494                 for (i = 0; i < sizeof(struct cdrom_multisession); i++)
 495                         if (put_user(0, (char __user *)(argument + i)))
 496                                 return -EFAULT;
 497                 return 0;
 498
 499         case CDROM_GET_CAPABILITY: {
 500                 struct gendisk *gd = info->gd;
 501                 if (gd->flags & GENHD_FL_CD)
 502                         return 0;
 503                 return -EINVAL;
 504         }
 505
 506         default:
 507                 /*printk(KERN_ALERT "ioctl %08x not supported by Xen blkdev\n",
 508                   command);*/
 509                 return -EINVAL; /* same return as native Linux */
 510         }
 511
 512         return 0;
 513 }
 514
 515 static unsigned long blkif_ring_get_request(struct blkfront_ring_info *rinfo,
 516                                             struct request *req,
 517                                             struct blkif_request **ring_req)
 518 {
 519         unsigned long id;
 520
 521         *ring_req = RING_GET_REQUEST(&rinfo->ring, rinfo->ring.req_prod_pvt);
 522         rinfo->ring.req_prod_pvt++;
 523
 524         id = get_id_from_freelist(rinfo);
 525         rinfo->shadow[id].request = req;
 526         rinfo->shadow[id].status = REQ_WAITING;
 527         rinfo->shadow[id].associated_id = NO_ASSOCIATED_ID;
 528
 529         (*ring_req)->u.rw.id = id;
 530
 531         return id;
 532 }
 533
 534 static int blkif_queue_discard_req(struct request *req, struct blkfront_ring_info *rinfo)
 535 {
 536         struct blkfront_info *info = rinfo->dev_info;
 537         struct blkif_request *ring_req;
 538         unsigned long id;
 539
 540         /* Fill out a communications ring structure. */
 541         id = blkif_ring_get_request(rinfo, req, &ring_req);
 542
 543         ring_req->operation = BLKIF_OP_DISCARD;
 544         ring_req->u.discard.nr_sectors = blk_rq_sectors(req);
 545         ring_req->u.discard.id = id;
 546         ring_req->u.discard.sector_number = (blkif_sector_t)blk_rq_pos(req);
 547         if ((req->cmd_flags & REQ_SECURE) && info->feature_secdiscard)
 548                 ring_req->u.discard.flag = BLKIF_DISCARD_SECURE;
 549         else
 550                 ring_req->u.discard.flag = 0;
 551
 552         /* Keep a private copy so we can reissue requests when recovering. */
 553         rinfo->shadow[id].req = *ring_req;
 554
 555         return 0;
 556 }
 557
 558 struct setup_rw_req {
 559         unsigned int grant_idx;
 560         struct blkif_request_segment *segments;
 561         struct blkfront_ring_info *rinfo;
 562         struct blkif_request *ring_req;
 563         grant_ref_t gref_head;
 564         unsigned int id;
 565         /* Only used when persistent grant is used and it's a read request */
 566         bool need_copy;
 567         unsigned int bvec_off;
 568         char *bvec_data;
 569
 570         bool require_extra_req;
 571         struct blkif_request *extra_ring_req;
 572 };
 573
 574 static void blkif_setup_rw_req_grant(unsigned long gfn, unsigned int offset,
 575                                      unsigned int len, void *data)
 576 {
 577         struct setup_rw_req *setup = data;
 578         int n, ref;
 579         struct grant *gnt_list_entry;
 580         unsigned int fsect, lsect;
 581         /* Convenient aliases */
 582         unsigned int grant_idx = setup->grant_idx;
 583         struct blkif_request *ring_req = setup->ring_req;
 584         struct blkfront_ring_info *rinfo = setup->rinfo;
 585         /*
 586          * We always use the shadow of the first request to store the list
 587          * of grant associated to the block I/O request. This made the
 588          * completion more easy to handle even if the block I/O request is
 589          * split.
 590          */
 591         struct blk_shadow *shadow = &rinfo->shadow[setup->id];
 592
 593         if (unlikely(setup->require_extra_req &&
 594                      grant_idx >= BLKIF_MAX_SEGMENTS_PER_REQUEST)) {
 595                 /*
 596                  * We are using the second request, setup grant_idx
 597                  * to be the index of the segment array.
 598                  */
 599                 grant_idx -= BLKIF_MAX_SEGMENTS_PER_REQUEST;
 600                 ring_req = setup->extra_ring_req;
 601         }
 602
 603         if ((ring_req->operation == BLKIF_OP_INDIRECT) &&
 604             (grant_idx % GRANTS_PER_INDIRECT_FRAME == 0)) {
 605                 if (setup->segments)
 606                         kunmap_atomic(setup->segments);
 607
 608                 n = grant_idx / GRANTS_PER_INDIRECT_FRAME;
 609                 gnt_list_entry = get_indirect_grant(&setup->gref_head, rinfo);
 610                 shadow->indirect_grants[n] = gnt_list_entry;
 611                 setup->segments = kmap_atomic(gnt_list_entry->page);
 612                 ring_req->u.indirect.indirect_grefs[n] = gnt_list_entry->gref;
 613         }
 614
 615         gnt_list_entry = get_grant(&setup->gref_head, gfn, rinfo);
 616         ref = gnt_list_entry->gref;
 617         /*
 618          * All the grants are stored in the shadow of the first
 619          * request. Therefore we have to use the global index.
 620          */
 621         shadow->grants_used[setup->grant_idx] = gnt_list_entry;
 622
 623         if (setup->need_copy) {
 624                 void *shared_data;
 625
 626                 shared_data = kmap_atomic(gnt_list_entry->page);
 627                 /*
 628                  * this does not wipe data stored outside the
 629                  * range sg->offset..sg->offset+sg->length.
 630                  * Therefore, blkback *could* see data from
 631                  * previous requests. This is OK as long as
 632                  * persistent grants are shared with just one
 633                  * domain. It may need refactoring if this
 634                  * changes
 635                  */
 636                 memcpy(shared_data + offset,
 637                        setup->bvec_data + setup->bvec_off,
 638                        len);
 639
 640                 kunmap_atomic(shared_data);
 641                 setup->bvec_off += len;
 642         }
 643
 644         fsect = offset >> 9;
 645         lsect = fsect + (len >> 9) - 1;
 646         if (ring_req->operation != BLKIF_OP_INDIRECT) {
 647                 ring_req->u.rw.seg[grant_idx] =
 648                         (struct blkif_request_segment) {
 649                                 .gref       = ref,
 650                                 .first_sect = fsect,
 651                                 .last_sect  = lsect };
 652         } else {
 653                 setup->segments[grant_idx % GRANTS_PER_INDIRECT_FRAME] =
 654                         (struct blkif_request_segment) {
 655                                 .gref       = ref,
 656                                 .first_sect = fsect,
 657                                 .last_sect  = lsect };
 658         }
 659
 660         (setup->grant_idx)++;
 661 }
 662
 663 static void blkif_setup_extra_req(struct blkif_request *first,
 664                                   struct blkif_request *second)
 665 {
 666         uint16_t nr_segments = first->u.rw.nr_segments;
 667
 668         /*
 669          * The second request is only present when the first request uses
 670          * all its segments. It's always the continuity of the first one.
 671          */
 672         first->u.rw.nr_segments = BLKIF_MAX_SEGMENTS_PER_REQUEST;
 673
 674         second->u.rw.nr_segments = nr_segments - BLKIF_MAX_SEGMENTS_PER_REQUEST;
 675         second->u.rw.sector_number = first->u.rw.sector_number +
 676                 (BLKIF_MAX_SEGMENTS_PER_REQUEST * XEN_PAGE_SIZE) / 512;
 677
 678         second->u.rw.handle = first->u.rw.handle;
 679         second->operation = first->operation;
 680 }
 681
 682 static int blkif_queue_rw_req(struct request *req, struct blkfront_ring_info *rinfo)
 683 {
 684         struct blkfront_info *info = rinfo->dev_info;
 685         struct blkif_request *ring_req, *extra_ring_req = NULL;
 686         unsigned long id, extra_id = NO_ASSOCIATED_ID;
 687         bool require_extra_req = false;
 688         int i;
 689         struct setup_rw_req setup = {
 690                 .grant_idx = 0,
 691                 .segments = NULL,
 692                 .rinfo = rinfo,
 693                 .need_copy = rq_data_dir(req) && info->feature_persistent,
 694         };
 695
 696         /*
 697          * Used to store if we are able to queue the request by just using
 698          * existing persistent grants, or if we have to get new grants,
 699          * as there are not sufficiently many free.
 700          */
 701         struct scatterlist *sg;
 702         int num_sg, max_grefs, num_grant;
 703
 704         max_grefs = req->nr_phys_segments * GRANTS_PER_PSEG;
 705         if (max_grefs > BLKIF_MAX_SEGMENTS_PER_REQUEST)
 706                 /*
 707                  * If we are using indirect segments we need to account
 708                  * for the indirect grefs used in the request.
 709                  */
 710                 max_grefs += INDIRECT_GREFS(max_grefs);
 711
 712         /*
 713          * We have to reserve 'max_grefs' grants because persistent
 714          * grants are shared by all rings.
 715          */
 716         if (max_grefs > 0)
 717                 if (gnttab_alloc_grant_references(max_grefs, &setup.gref_head) < 0) {
 718                         gnttab_request_free_callback(
 719                                 &rinfo->callback,
 720                                 blkif_restart_queue_callback,
 721                                 rinfo,
 722                                 max_grefs);
 723                         return 1;
 724                 }
 725
 726         /* Fill out a communications ring structure. */
 727         id = blkif_ring_get_request(rinfo, req, &ring_req);
 728
 729         num_sg = blk_rq_map_sg(req->q, req, rinfo->shadow[id].sg);
 730         num_grant = 0;
 731         /* Calculate the number of grant used */
 732         for_each_sg(rinfo->shadow[id].sg, sg, num_sg, i)
 733                num_grant += gnttab_count_grant(sg->offset, sg->length);
 734
 735         require_extra_req = info->max_indirect_segments == 0 &&
 736                 num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST;
 737         BUG_ON(!HAS_EXTRA_REQ && require_extra_req);
 738
 739         rinfo->shadow[id].num_sg = num_sg;
 740         if (num_grant > BLKIF_MAX_SEGMENTS_PER_REQUEST &&
 741             likely(!require_extra_req)) {
 742                 /*
 743                  * The indirect operation can only be a BLKIF_OP_READ or
 744                  * BLKIF_OP_WRITE
 745                  */
 746                 BUG_ON(req->cmd_flags & (REQ_FLUSH | REQ_FUA));
 747                 ring_req->operation = BLKIF_OP_INDIRECT;
 748                 ring_req->u.indirect.indirect_op = rq_data_dir(req) ?
 749                         BLKIF_OP_WRITE : BLKIF_OP_READ;
 750                 ring_req->u.indirect.sector_number = (blkif_sector_t)blk_rq_pos(req);
 751                 ring_req->u.indirect.handle = info->handle;
 752                 ring_req->u.indirect.nr_segments = num_grant;
 753         } else {
 754                 ring_req->u.rw.sector_number = (blkif_sector_t)blk_rq_pos(req);
 755                 ring_req->u.rw.handle = info->handle;
 756                 ring_req->operation = rq_data_dir(req) ?
 757                         BLKIF_OP_WRITE : BLKIF_OP_READ;
 758                 if (req->cmd_flags & (REQ_FLUSH | REQ_FUA)) {
 759                         /*
 760                          * Ideally we can do an unordered flush-to-disk.
 761                          * In case the backend onlysupports barriers, use that.
 762                          * A barrier request a superset of FUA, so we can
 763                          * implement it the same way.  (It's also a FLUSH+FUA,
 764                          * since it is guaranteed ordered WRT previous writes.)
 765                          */
 766                         switch (info->feature_flush &
 767                                 ((REQ_FLUSH|REQ_FUA))) {
 768                         case REQ_FLUSH|REQ_FUA:
 769                                 ring_req->operation =
 770                                         BLKIF_OP_WRITE_BARRIER;
 771                                 break;
 772                         case REQ_FLUSH:
 773                                 ring_req->operation =
 774                                         BLKIF_OP_FLUSH_DISKCACHE;
 775                                 break;
 776                         default:
 777                                 ring_req->operation = 0;
 778                         }
 779                 }
 780                 ring_req->u.rw.nr_segments = num_grant;
 781                 if (unlikely(require_extra_req)) {
 782                         extra_id = blkif_ring_get_request(rinfo, req,
 783                                                           &extra_ring_req);
 784                         /*
 785                          * Only the first request contains the scatter-gather
 786                          * list.
 787                          */
 788                         rinfo->shadow[extra_id].num_sg = 0;
 789
 790                         blkif_setup_extra_req(ring_req, extra_ring_req);
 791
 792                         /* Link the 2 requests together */
 793                         rinfo->shadow[extra_id].associated_id = id;
 794                         rinfo->shadow[id].associated_id = extra_id;
 795                 }
 796         }
 797
 798         setup.ring_req = ring_req;
 799         setup.id = id;
 800
 801         setup.require_extra_req = require_extra_req;
 802         if (unlikely(require_extra_req))
 803                 setup.extra_ring_req = extra_ring_req;
 804
 805         for_each_sg(rinfo->shadow[id].sg, sg, num_sg, i) {
 806                 BUG_ON(sg->offset + sg->length > PAGE_SIZE);
 807
 808                 if (setup.need_copy) {
 809                         setup.bvec_off = sg->offset;
 810                         setup.bvec_data = kmap_atomic(sg_page(sg));
 811                 }
 812
 813                 gnttab_foreach_grant_in_range(sg_page(sg),
 814                                               sg->offset,
 815                                               sg->length,
 816                                               blkif_setup_rw_req_grant,
 817                                               &setup);
 818
 819                 if (setup.need_copy)
 820                         kunmap_atomic(setup.bvec_data);
 821         }
 822         if (setup.segments)
 823                 kunmap_atomic(setup.segments);
 824
 825         /* Keep a private copy so we can reissue requests when recovering. */
 826         rinfo->shadow[id].req = *ring_req;
 827         if (unlikely(require_extra_req))
 828                 rinfo->shadow[extra_id].req = *extra_ring_req;
 829
 830         if (max_grefs > 0)
 831                 gnttab_free_grant_references(setup.gref_head);
 832
 833         return 0;
 834 }
 835
 836 /*
 837  * Generate a Xen blkfront IO request from a blk layer request.  Reads
 838  * and writes are handled as expected.
 839  *
 840  * @req: a request struct
 841  */
 842 static int blkif_queue_request(struct request *req, struct blkfront_ring_info *rinfo)
 843 {
 844         if (unlikely(rinfo->dev_info->connected != BLKIF_STATE_CONNECTED))
 845                 return 1;
 846
 847         if (unlikely(req->cmd_flags & (REQ_DISCARD | REQ_SECURE)))
 848                 return blkif_queue_discard_req(req, rinfo);
 849         else
 850                 return blkif_queue_rw_req(req, rinfo);
 851 }
 852
 853 static inline void flush_requests(struct blkfront_ring_info *rinfo)
 854 {
 855         int notify;
 856
 857         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&rinfo->ring, notify);
 858
 859         if (notify)
 860                 notify_remote_via_irq(rinfo->irq);
 861 }
 862
 863 static inline bool blkif_request_flush_invalid(struct request *req,
 864                                                struct blkfront_info *info)
 865 {
 866         return ((req->cmd_type != REQ_TYPE_FS) ||
 867                 ((req->cmd_flags & REQ_FLUSH) &&
 868                  !(info->feature_flush & REQ_FLUSH)) ||
 869                 ((req->cmd_flags & REQ_FUA) &&
 870                  !(info->feature_flush & REQ_FUA)));
 871 }
 872
 873 static int blkif_queue_rq(struct blk_mq_hw_ctx *hctx,
 874                           const struct blk_mq_queue_data *qd)
 875 {
 876         unsigned long flags;
 877         struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)hctx->driver_data;
 878
 879         blk_mq_start_request(qd->rq);
 880         spin_lock_irqsave(&rinfo->ring_lock, flags);
 881         if (RING_FULL(&rinfo->ring))
 882                 goto out_busy;
 883
 884         if (blkif_request_flush_invalid(qd->rq, rinfo->dev_info))
 885                 goto out_err;
 886
 887         if (blkif_queue_request(qd->rq, rinfo))
 888                 goto out_busy;
 889
 890         flush_requests(rinfo);
 891         spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 892         return BLK_MQ_RQ_QUEUE_OK;
 893
 894 out_err:
 895         spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 896         return BLK_MQ_RQ_QUEUE_ERROR;
 897
 898 out_busy:
 899         spin_unlock_irqrestore(&rinfo->ring_lock, flags);
 900         blk_mq_stop_hw_queue(hctx);
 901         return BLK_MQ_RQ_QUEUE_BUSY;
 902 }
 903
 904 static int blk_mq_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
 905                             unsigned int index)
 906 {
 907         struct blkfront_info *info = (struct blkfront_info *)data;
 908
 909         BUG_ON(info->nr_rings <= index);
 910         hctx->driver_data = &info->rinfo[index];
 911         return 0;
 912 }
 913
 914 static struct blk_mq_ops blkfront_mq_ops = {
 915         .queue_rq = blkif_queue_rq,
 916         .map_queue = blk_mq_map_queue,
 917         .init_hctx = blk_mq_init_hctx,
 918 };
 919
 920 static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size,
 921                                 unsigned int physical_sector_size,
 922                                 unsigned int segments)
 923 {
 924         struct request_queue *rq;
 925         struct blkfront_info *info = gd->private_data;
 926
 927         memset(&info->tag_set, 0, sizeof(info->tag_set));
 928         info->tag_set.ops = &blkfront_mq_ops;
 929         info->tag_set.nr_hw_queues = info->nr_rings;
 930         if (HAS_EXTRA_REQ && info->max_indirect_segments == 0) {
 931                 /*
 932                  * When indirect descriptior is not supported, the I/O request
 933                  * will be split between multiple request in the ring.
 934                  * To avoid problems when sending the request, divide by
 935                  * 2 the depth of the queue.
 936                  */
 937                 info->tag_set.queue_depth =  BLK_RING_SIZE(info) / 2;
 938         } else
 939                 info->tag_set.queue_depth = BLK_RING_SIZE(info);
 940         info->tag_set.numa_node = NUMA_NO_NODE;
 941         info->tag_set.flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_SG_MERGE;
 942         info->tag_set.cmd_size = 0;
 943         info->tag_set.driver_data = info;
 944
 945         if (blk_mq_alloc_tag_set(&info->tag_set))
 946                 return -EINVAL;
 947         rq = blk_mq_init_queue(&info->tag_set);
 948         if (IS_ERR(rq)) {
 949                 blk_mq_free_tag_set(&info->tag_set);
 950                 return PTR_ERR(rq);
 951         }
 952
 953         queue_flag_set_unlocked(QUEUE_FLAG_VIRT, rq);
 954
 955         if (info->feature_discard) {
 956                 queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, rq);
 957                 blk_queue_max_discard_sectors(rq, get_capacity(gd));
 958                 rq->limits.discard_granularity = info->discard_granularity;
 959                 rq->limits.discard_alignment = info->discard_alignment;
 960                 if (info->feature_secdiscard)
 961                         queue_flag_set_unlocked(QUEUE_FLAG_SECDISCARD, rq);
 962         }
 963
 964         /* Hard sector size and max sectors impersonate the equiv. hardware. */
 965         blk_queue_logical_block_size(rq, sector_size);
 966         blk_queue_physical_block_size(rq, physical_sector_size);
 967         blk_queue_max_hw_sectors(rq, (segments * XEN_PAGE_SIZE) / 512);
 968
 969         /* Each segment in a request is up to an aligned page in size. */
 970         blk_queue_segment_boundary(rq, PAGE_SIZE - 1);
 971         blk_queue_max_segment_size(rq, PAGE_SIZE);
 972
 973         /* Ensure a merged request will fit in a single I/O ring slot. */
 974         blk_queue_max_segments(rq, segments / GRANTS_PER_PSEG);
 975
 976         /* Make sure buffer addresses are sector-aligned. */
 977         blk_queue_dma_alignment(rq, 511);
 978
 979         /* Make sure we don't use bounce buffers. */
 980         blk_queue_bounce_limit(rq, BLK_BOUNCE_ANY);
 981
 982         gd->queue = rq;
 983
 984         return 0;
 985 }
 986
 987 static const char *flush_info(unsigned int feature_flush)
 988 {
 989         switch (feature_flush & ((REQ_FLUSH | REQ_FUA))) {
 990         case REQ_FLUSH|REQ_FUA:
 991                 return "barrier: enabled;";
 992         case REQ_FLUSH:
 993                 return "flush diskcache: enabled;";
 994         default:
 995                 return "barrier or flush: disabled;";
 996         }
 997 }
 998
 999 static void xlvbd_flush(struct blkfront_info *info)
1000 {
1001         blk_queue_flush(info->rq, info->feature_flush);
1002         pr_info("blkfront: %s: %s %s %s %s %s\n",
1003                 info->gd->disk_name, flush_info(info->feature_flush),
1004                 "persistent grants:", info->feature_persistent ?
1005                 "enabled;" : "disabled;", "indirect descriptors:",
1006                 info->max_indirect_segments ? "enabled;" : "disabled;");
1007 }
1008
1009 static int xen_translate_vdev(int vdevice, int *minor, unsigned int *offset)
1010 {
1011         int major;
1012         major = BLKIF_MAJOR(vdevice);
1013         *minor = BLKIF_MINOR(vdevice);
1014         switch (major) {
1015                 case XEN_IDE0_MAJOR:
1016                         *offset = (*minor / 64) + EMULATED_HD_DISK_NAME_OFFSET;
1017                         *minor = ((*minor / 64) * PARTS_PER_DISK) +
1018                                 EMULATED_HD_DISK_MINOR_OFFSET;
1019                         break;
1020                 case XEN_IDE1_MAJOR:
1021                         *offset = (*minor / 64) + 2 + EMULATED_HD_DISK_NAME_OFFSET;
1022                         *minor = (((*minor / 64) + 2) * PARTS_PER_DISK) +
1023                                 EMULATED_HD_DISK_MINOR_OFFSET;
1024                         break;
1025                 case XEN_SCSI_DISK0_MAJOR:
1026                         *offset = (*minor / PARTS_PER_DISK) + EMULATED_SD_DISK_NAME_OFFSET;
1027                         *minor = *minor + EMULATED_SD_DISK_MINOR_OFFSET;
1028                         break;
1029                 case XEN_SCSI_DISK1_MAJOR:
1030                 case XEN_SCSI_DISK2_MAJOR:
1031                 case XEN_SCSI_DISK3_MAJOR:
1032                 case XEN_SCSI_DISK4_MAJOR:
1033                 case XEN_SCSI_DISK5_MAJOR:
1034                 case XEN_SCSI_DISK6_MAJOR:
1035                 case XEN_SCSI_DISK7_MAJOR:
1036                         *offset = (*minor / PARTS_PER_DISK) +
1037                                 ((major - XEN_SCSI_DISK1_MAJOR + 1) * 16) +
1038                                 EMULATED_SD_DISK_NAME_OFFSET;
1039                         *minor = *minor +
1040                                 ((major - XEN_SCSI_DISK1_MAJOR + 1) * 16 * PARTS_PER_DISK) +
1041                                 EMULATED_SD_DISK_MINOR_OFFSET;
1042                         break;
1043                 case XEN_SCSI_DISK8_MAJOR:
1044                 case XEN_SCSI_DISK9_MAJOR:
1045                 case XEN_SCSI_DISK10_MAJOR:
1046                 case XEN_SCSI_DISK11_MAJOR:
1047                 case XEN_SCSI_DISK12_MAJOR:
1048                 case XEN_SCSI_DISK13_MAJOR:
1049                 case XEN_SCSI_DISK14_MAJOR:
1050                 case XEN_SCSI_DISK15_MAJOR:
1051                         *offset = (*minor / PARTS_PER_DISK) +
1052                                 ((major - XEN_SCSI_DISK8_MAJOR + 8) * 16) +
1053                                 EMULATED_SD_DISK_NAME_OFFSET;
1054                         *minor = *minor +
1055                                 ((major - XEN_SCSI_DISK8_MAJOR + 8) * 16 * PARTS_PER_DISK) +
1056                                 EMULATED_SD_DISK_MINOR_OFFSET;
1057                         break;
1058                 case XENVBD_MAJOR:
1059                         *offset = *minor / PARTS_PER_DISK;
1060                         break;
1061                 default:
1062                         printk(KERN_WARNING "blkfront: your disk configuration is "
1063                                         "incorrect, please use an xvd device instead\n");
1064                         return -ENODEV;
1065         }
1066         return 0;
1067 }
1068
1069 static char *encode_disk_name(char *ptr, unsigned int n)
1070 {
1071         if (n >= 26)
1072                 ptr = encode_disk_name(ptr, n / 26 - 1);
1073         *ptr = 'a' + n % 26;
1074         return ptr + 1;
1075 }
1076
1077 static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
1078                                struct blkfront_info *info,
1079                                u16 vdisk_info, u16 sector_size,
1080                                unsigned int physical_sector_size)
1081 {
1082         struct gendisk *gd;
1083         int nr_minors = 1;
1084         int err;
1085         unsigned int offset;
1086         int minor;
1087         int nr_parts;
1088         char *ptr;
1089
1090         BUG_ON(info->gd != NULL);
1091         BUG_ON(info->rq != NULL);
1092
1093         if ((info->vdevice>>EXT_SHIFT) > 1) {
1094                 /* this is above the extended range; something is wrong */
1095                 printk(KERN_WARNING "blkfront: vdevice 0x%x is above the extended range; ignoring\n", info->vdevice);
1096                 return -ENODEV;
1097         }
1098
1099         if (!VDEV_IS_EXTENDED(info->vdevice)) {
1100                 err = xen_translate_vdev(info->vdevice, &minor, &offset);
1101                 if (err)
1102                         return err;
1103                 nr_parts = PARTS_PER_DISK;
1104         } else {
1105                 minor = BLKIF_MINOR_EXT(info->vdevice);
1106                 nr_parts = PARTS_PER_EXT_DISK;
1107                 offset = minor / nr_parts;
1108                 if (xen_hvm_domain() && offset < EMULATED_HD_DISK_NAME_OFFSET + 4)
1109                         printk(KERN_WARNING "blkfront: vdevice 0x%x might conflict with "
1110                                         "emulated IDE disks,\n\t choose an xvd device name"
1111                                         "from xvde on\n", info->vdevice);
1112         }
1113         if (minor >> MINORBITS) {
1114                 pr_warn("blkfront: %#x's minor (%#x) out of range; ignoring\n",
1115                         info->vdevice, minor);
1116                 return -ENODEV;
1117         }
1118
1119         if ((minor % nr_parts) == 0)
1120                 nr_minors = nr_parts;
1121
1122         err = xlbd_reserve_minors(minor, nr_minors);
1123         if (err)
1124                 goto out;
1125         err = -ENODEV;
1126
1127         gd = alloc_disk(nr_minors);
1128         if (gd == NULL)
1129                 goto release;
1130
1131         strcpy(gd->disk_name, DEV_NAME);
1132         ptr = encode_disk_name(gd->disk_name + sizeof(DEV_NAME) - 1, offset);
1133         BUG_ON(ptr >= gd->disk_name + DISK_NAME_LEN);
1134         if (nr_minors > 1)
1135                 *ptr = 0;
1136         else
1137                 snprintf(ptr, gd->disk_name + DISK_NAME_LEN - ptr,
1138                          "%d", minor & (nr_parts - 1));
1139
1140         gd->major = XENVBD_MAJOR;
1141         gd->first_minor = minor;
1142         gd->fops = &xlvbd_block_fops;
1143         gd->private_data = info;
1144         gd->driverfs_dev = &(info->xbdev->dev);
1145         set_capacity(gd, capacity);
1146
1147         if (xlvbd_init_blk_queue(gd, sector_size, physical_sector_size,
1148                                  info->max_indirect_segments ? :
1149                                  BLKIF_MAX_SEGMENTS_PER_REQUEST)) {
1150                 del_gendisk(gd);
1151                 goto release;
1152         }
1153
1154         info->rq = gd->queue;
1155         info->gd = gd;
1156
1157         xlvbd_flush(info);
1158
1159         if (vdisk_info & VDISK_READONLY)
1160                 set_disk_ro(gd, 1);
1161
1162         if (vdisk_info & VDISK_REMOVABLE)
1163                 gd->flags |= GENHD_FL_REMOVABLE;
1164
1165         if (vdisk_info & VDISK_CDROM)
1166                 gd->flags |= GENHD_FL_CD;
1167
1168         return 0;
1169
1170  release:
1171         xlbd_release_minors(minor, nr_minors);
1172  out:
1173         return err;
1174 }
1175
1176 static void xlvbd_release_gendisk(struct blkfront_info *info)
1177 {
1178         unsigned int minor, nr_minors, i;
1179
1180         if (info->rq == NULL)
1181                 return;
1182
1183         /* No more blkif_request(). */
1184         blk_mq_stop_hw_queues(info->rq);
1185
1186         for (i = 0; i < info->nr_rings; i++) {
1187                 struct blkfront_ring_info *rinfo = &info->rinfo[i];
1188
1189                 /* No more gnttab callback work. */
1190                 gnttab_cancel_free_callback(&rinfo->callback);
1191
1192                 /* Flush gnttab callback work. Must be done with no locks held. */
1193                 flush_work(&rinfo->work);
1194         }
1195
1196         del_gendisk(info->gd);
1197
1198         minor = info->gd->first_minor;
1199         nr_minors = info->gd->minors;
1200         xlbd_release_minors(minor, nr_minors);
1201
1202         blk_cleanup_queue(info->rq);
1203         blk_mq_free_tag_set(&info->tag_set);
1204         info->rq = NULL;
1205
1206         put_disk(info->gd);
1207         info->gd = NULL;
1208 }
1209
1210 /* Already hold rinfo->ring_lock. */
1211 static inline void kick_pending_request_queues_locked(struct blkfront_ring_info *rinfo)
1212 {
1213         if (!RING_FULL(&rinfo->ring))
1214                 blk_mq_start_stopped_hw_queues(rinfo->dev_info->rq, true);
1215 }
1216
1217 static void kick_pending_request_queues(struct blkfront_ring_info *rinfo)
1218 {
1219         unsigned long flags;
1220
1221         spin_lock_irqsave(&rinfo->ring_lock, flags);
1222         kick_pending_request_queues_locked(rinfo);
1223         spin_unlock_irqrestore(&rinfo->ring_lock, flags);
1224 }
1225
1226 static void blkif_restart_queue(struct work_struct *work)
1227 {
1228         struct blkfront_ring_info *rinfo = container_of(work, struct blkfront_ring_info, work);
1229
1230         if (rinfo->dev_info->connected == BLKIF_STATE_CONNECTED)
1231                 kick_pending_request_queues(rinfo);
1232 }
1233
1234 static void blkif_free_ring(struct blkfront_ring_info *rinfo)
1235 {
1236         struct grant *persistent_gnt, *n;
1237         struct blkfront_info *info = rinfo->dev_info;
1238         int i, j, segs;
1239
1240         /*
1241          * Remove indirect pages, this only happens when using indirect
1242          * descriptors but not persistent grants
1243          */
1244         if (!list_empty(&rinfo->indirect_pages)) {
1245                 struct page *indirect_page, *n;
1246
1247                 BUG_ON(info->feature_persistent);
1248                 list_for_each_entry_safe(indirect_page, n, &rinfo->indirect_pages, lru) {
1249                         list_del(&indirect_page->lru);
1250                         __free_page(indirect_page);
1251                 }
1252         }
1253
1254         /* Remove all persistent grants. */
1255         if (!list_empty(&rinfo->grants)) {
1256                 list_for_each_entry_safe(persistent_gnt, n,
1257                                          &rinfo->grants, node) {
1258                         list_del(&persistent_gnt->node);
1259                         if (persistent_gnt->gref != GRANT_INVALID_REF) {
1260                                 gnttab_end_foreign_access(persistent_gnt->gref,
1261                                                           0, 0UL);
1262                                 rinfo->persistent_gnts_c--;
1263                         }
1264                         if (info->feature_persistent)
1265                                 __free_page(persistent_gnt->page);
1266                         kfree(persistent_gnt);
1267                 }
1268         }
1269         BUG_ON(rinfo->persistent_gnts_c != 0);
1270
1271         for (i = 0; i < BLK_RING_SIZE(info); i++) {
1272                 /*
1273                  * Clear persistent grants present in requests already
1274                  * on the shared ring
1275                  */
1276                 if (!rinfo->shadow[i].request)
1277                         goto free_shadow;
1278
1279                 segs = rinfo->shadow[i].req.operation == BLKIF_OP_INDIRECT ?
1280                        rinfo->shadow[i].req.u.indirect.nr_segments :
1281                        rinfo->shadow[i].req.u.rw.nr_segments;
1282                 for (j = 0; j < segs; j++) {
1283                         persistent_gnt = rinfo->shadow[i].grants_used[j];
1284                         gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL);
1285                         if (info->feature_persistent)
1286                                 __free_page(persistent_gnt->page);
1287                         kfree(persistent_gnt);
1288                 }
1289
1290                 if (rinfo->shadow[i].req.operation != BLKIF_OP_INDIRECT)
1291                         /*
1292                          * If this is not an indirect operation don't try to
1293                          * free indirect segments
1294                          */
1295                         goto free_shadow;
1296
1297                 for (j = 0; j < INDIRECT_GREFS(segs); j++) {
1298                         persistent_gnt = rinfo->shadow[i].indirect_grants[j];
1299                         gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL);
1300                         __free_page(persistent_gnt->page);
1301                         kfree(persistent_gnt);
1302                 }
1303
1304 free_shadow:
1305                 kfree(rinfo->shadow[i].grants_used);
1306                 rinfo->shadow[i].grants_used = NULL;
1307                 kfree(rinfo->shadow[i].indirect_grants);
1308                 rinfo->shadow[i].indirect_grants = NULL;
1309                 kfree(rinfo->shadow[i].sg);
1310                 rinfo->shadow[i].sg = NULL;
1311         }
1312
1313         /* No more gnttab callback work. */
1314         gnttab_cancel_free_callback(&rinfo->callback);
1315
1316         /* Flush gnttab callback work. Must be done with no locks held. */
1317         flush_work(&rinfo->work);
1318
1319         /* Free resources associated with old device channel. */
1320         for (i = 0; i < info->nr_ring_pages; i++) {
1321                 if (rinfo->ring_ref[i] != GRANT_INVALID_REF) {
1322                         gnttab_end_foreign_access(rinfo->ring_ref[i], 0, 0);
1323                         rinfo->ring_ref[i] = GRANT_INVALID_REF;
1324                 }
1325         }
1326         free_pages((unsigned long)rinfo->ring.sring, get_order(info->nr_ring_pages * PAGE_SIZE));
1327         rinfo->ring.sring = NULL;
1328
1329         if (rinfo->irq)
1330                 unbind_from_irqhandler(rinfo->irq, rinfo);
1331         rinfo->evtchn = rinfo->irq = 0;
1332 }
1333
1334 static void blkif_free(struct blkfront_info *info, int suspend)
1335 {
1336         unsigned int i;
1337
1338         /* Prevent new requests being issued until we fix things up. */
1339         info->connected = suspend ?
1340                 BLKIF_STATE_SUSPENDED : BLKIF_STATE_DISCONNECTED;
1341         /* No more blkif_request(). */
1342         if (info->rq)
1343                 blk_mq_stop_hw_queues(info->rq);
1344
1345         for (i = 0; i < info->nr_rings; i++)
1346                 blkif_free_ring(&info->rinfo[i]);
1347
1348         kfree(info->rinfo);
1349         info->rinfo = NULL;
1350         info->nr_rings = 0;
1351 }
1352
1353 struct copy_from_grant {
1354         const struct blk_shadow *s;
1355         unsigned int grant_idx;
1356         unsigned int bvec_offset;
1357         char *bvec_data;
1358 };
1359
1360 static void blkif_copy_from_grant(unsigned long gfn, unsigned int offset,
1361                                   unsigned int len, void *data)
1362 {
1363         struct copy_from_grant *info = data;
1364         char *shared_data;
1365         /* Convenient aliases */
1366         const struct blk_shadow *s = info->s;
1367
1368         shared_data = kmap_atomic(s->grants_used[info->grant_idx]->page);
1369
1370         memcpy(info->bvec_data + info->bvec_offset,
1371                shared_data + offset, len);
1372
1373         info->bvec_offset += len;
1374         info->grant_idx++;
1375
1376         kunmap_atomic(shared_data);
1377 }
1378
1379 static enum blk_req_status blkif_rsp_to_req_status(int rsp)
1380 {
1381         switch (rsp)
1382         {
1383         case BLKIF_RSP_OKAY:
1384                 return REQ_DONE;
1385         case BLKIF_RSP_EOPNOTSUPP:
1386                 return REQ_EOPNOTSUPP;
1387         case BLKIF_RSP_ERROR:
1388                 /* Fallthrough. */
1389         default:
1390                 return REQ_ERROR;
1391         }
1392 }
1393
1394 /*
1395  * Get the final status of the block request based on two ring response
1396  */
1397 static int blkif_get_final_status(enum blk_req_status s1,
1398                                   enum blk_req_status s2)
1399 {
1400         BUG_ON(s1 == REQ_WAITING);
1401         BUG_ON(s2 == REQ_WAITING);
1402
1403         if (s1 == REQ_ERROR || s2 == REQ_ERROR)
1404                 return BLKIF_RSP_ERROR;
1405         else if (s1 == REQ_EOPNOTSUPP || s2 == REQ_EOPNOTSUPP)
1406                 return BLKIF_RSP_EOPNOTSUPP;
1407         return BLKIF_RSP_OKAY;
1408 }
1409
1410 static bool blkif_completion(unsigned long *id,
1411                              struct blkfront_ring_info *rinfo,
1412                              struct blkif_response *bret)
1413 {
1414         int i = 0;
1415         struct scatterlist *sg;
1416         int num_sg, num_grant;
1417         struct blkfront_info *info = rinfo->dev_info;
1418         struct blk_shadow *s = &rinfo->shadow[*id];
1419         struct copy_from_grant data = {
1420                 .grant_idx = 0,
1421         };
1422
1423         num_grant = s->req.operation == BLKIF_OP_INDIRECT ?
1424                 s->req.u.indirect.nr_segments : s->req.u.rw.nr_segments;
1425
1426         /* The I/O request may be split in two. */
1427         if (unlikely(s->associated_id != NO_ASSOCIATED_ID)) {
1428                 struct blk_shadow *s2 = &rinfo->shadow[s->associated_id];
1429
1430                 /* Keep the status of the current response in shadow. */
1431                 s->status = blkif_rsp_to_req_status(bret->status);
1432
1433                 /* Wait the second response if not yet here. */
1434                 if (s2->status == REQ_WAITING)
1435                         return 0;
1436
1437                 bret->status = blkif_get_final_status(s->status,
1438                                                       s2->status);
1439
1440                 /*
1441                  * All the grants is stored in the first shadow in order
1442                  * to make the completion code simpler.
1443                  */
1444                 num_grant += s2->req.u.rw.nr_segments;
1445
1446                 /*
1447                  * The two responses may not come in order. Only the
1448                  * first request will store the scatter-gather list.
1449                  */
1450                 if (s2->num_sg != 0) {
1451                         /* Update "id" with the ID of the first response. */
1452                         *id = s->associated_id;
1453                         s = s2;
1454                 }
1455
1456                 /*
1457                  * We don't need anymore the second request, so recycling
1458                  * it now.
1459                  */
1460                 if (add_id_to_freelist(rinfo, s->associated_id))
1461                         WARN(1, "%s: can't recycle the second part (id = %ld) of the request\n",
1462                              info->gd->disk_name, s->associated_id);
1463         }
1464
1465         data.s = s;
1466         num_sg = s->num_sg;
1467
1468         if (bret->operation == BLKIF_OP_READ && info->feature_persistent) {
1469                 for_each_sg(s->sg, sg, num_sg, i) {
1470                         BUG_ON(sg->offset + sg->length > PAGE_SIZE);
1471
1472                         data.bvec_offset = sg->offset;
1473                         data.bvec_data = kmap_atomic(sg_page(sg));
1474
1475                         gnttab_foreach_grant_in_range(sg_page(sg),
1476                                                       sg->offset,
1477                                                       sg->length,
1478                                                       blkif_copy_from_grant,
1479                                                       &data);
1480
1481                         kunmap_atomic(data.bvec_data);
1482                 }
1483         }
1484         /* Add the persistent grant into the list of free grants */
1485         for (i = 0; i < num_grant; i++) {
1486                 if (gnttab_query_foreign_access(s->grants_used[i]->gref)) {
1487                         /*
1488                          * If the grant is still mapped by the backend (the
1489                          * backend has chosen to make this grant persistent)
1490                          * we add it at the head of the list, so it will be
1491                          * reused first.
1492                          */
1493                         if (!info->feature_persistent)
1494                                 pr_alert_ratelimited("backed has not unmapped grant: %u\n",
1495                                                      s->grants_used[i]->gref);
1496                         list_add(&s->grants_used[i]->node, &rinfo->grants);
1497                         rinfo->persistent_gnts_c++;
1498                 } else {
1499                         /*
1500                          * If the grant is not mapped by the backend we end the
1501                          * foreign access and add it to the tail of the list,
1502                          * so it will not be picked again unless we run out of
1503                          * persistent grants.
1504                          */
1505                         gnttab_end_foreign_access(s->grants_used[i]->gref, 0, 0UL);
1506                         s->grants_used[i]->gref = GRANT_INVALID_REF;
1507                         list_add_tail(&s->grants_used[i]->node, &rinfo->grants);
1508                 }
1509         }
1510         if (s->req.operation == BLKIF_OP_INDIRECT) {
1511                 for (i = 0; i < INDIRECT_GREFS(num_grant); i++) {
1512                         if (gnttab_query_foreign_access(s->indirect_grants[i]->gref)) {
1513                                 if (!info->feature_persistent)
1514                                         pr_alert_ratelimited("backed has not unmapped grant: %u\n",
1515                                                              s->indirect_grants[i]->gref);
1516                                 list_add(&s->indirect_grants[i]->node, &rinfo->grants);
1517                                 rinfo->persistent_gnts_c++;
1518                         } else {
1519                                 struct page *indirect_page;
1520
1521                                 gnttab_end_foreign_access(s->indirect_grants[i]->gref, 0, 0UL);
1522                                 /*
1523                                  * Add the used indirect page back to the list of
1524                                  * available pages for indirect grefs.
1525                                  */
1526                                 if (!info->feature_persistent) {
1527                                         indirect_page = s->indirect_grants[i]->page;
1528                                         list_add(&indirect_page->lru, &rinfo->indirect_pages);
1529                                 }
1530                                 s->indirect_grants[i]->gref = GRANT_INVALID_REF;
1531                                 list_add_tail(&s->indirect_grants[i]->node, &rinfo->grants);
1532                         }
1533                 }
1534         }
1535
1536         return 1;
1537 }
1538
1539 static irqreturn_t blkif_interrupt(int irq, void *dev_id)
1540 {
1541         struct request *req;
1542         struct blkif_response *bret;
1543         RING_IDX i, rp;
1544         unsigned long flags;
1545         struct blkfront_ring_info *rinfo = (struct blkfront_ring_info *)dev_id;
1546         struct blkfront_info *info = rinfo->dev_info;
1547         int error;
1548
1549         if (unlikely(info->connected != BLKIF_STATE_CONNECTED))
1550                 return IRQ_HANDLED;
1551
1552         spin_lock_irqsave(&rinfo->ring_lock, flags);
1553  again:
1554         rp = rinfo->ring.sring->rsp_prod;
1555         rmb(); /* Ensure we see queued responses up to 'rp'. */
1556
1557         for (i = rinfo->ring.rsp_cons; i != rp; i++) {
1558                 unsigned long id;
1559
1560                 bret = RING_GET_RESPONSE(&rinfo->ring, i);
1561                 id   = bret->id;
1562                 /*
1563                  * The backend has messed up and given us an id that we would
1564                  * never have given to it (we stamp it up to BLK_RING_SIZE -
1565                  * look in get_id_from_freelist.
1566                  */
1567                 if (id >= BLK_RING_SIZE(info)) {
1568                         WARN(1, "%s: response to %s has incorrect id (%ld)\n",
1569                              info->gd->disk_name, op_name(bret->operation), id);
1570                         /* We can't safely get the 'struct request' as
1571                          * the id is busted. */
1572                         continue;
1573                 }
1574                 req  = rinfo->shadow[id].request;
1575
1576                 if (bret->operation != BLKIF_OP_DISCARD) {
1577                         /*
1578                          * We may need to wait for an extra response if the
1579                          * I/O request is split in 2
1580                          */
1581                         if (!blkif_completion(&id, rinfo, bret))
1582                                 continue;
1583                 }
1584
1585                 if (add_id_to_freelist(rinfo, id)) {
1586                         WARN(1, "%s: response to %s (id %ld) couldn't be recycled!\n",
1587                              info->gd->disk_name, op_name(bret->operation), id);
1588                         continue;
1589                 }
1590
1591                 error = (bret->status == BLKIF_RSP_OKAY) ? 0 : -EIO;
1592                 switch (bret->operation) {
1593                 case BLKIF_OP_DISCARD:
1594                         if (unlikely(bret->status == BLKIF_RSP_EOPNOTSUPP)) {
1595                                 struct request_queue *rq = info->rq;
1596                                 printk(KERN_WARNING "blkfront: %s: %s op failed\n",
1597                                            info->gd->disk_name, op_name(bret->operation));
1598                                 error = -EOPNOTSUPP;
1599                                 info->feature_discard = 0;
1600                                 info->feature_secdiscard = 0;
1601                                 queue_flag_clear(QUEUE_FLAG_DISCARD, rq);
1602                                 queue_flag_clear(QUEUE_FLAG_SECDISCARD, rq);
1603                         }
1604                         blk_mq_complete_request(req, error);
1605                         break;
1606                 case BLKIF_OP_FLUSH_DISKCACHE:
1607                 case BLKIF_OP_WRITE_BARRIER:
1608                         if (unlikely(bret->status == BLKIF_RSP_EOPNOTSUPP)) {
1609                                 printk(KERN_WARNING "blkfront: %s: %s op failed\n",
1610                                        info->gd->disk_name, op_name(bret->operation));
1611                                 error = -EOPNOTSUPP;
1612                         }
1613                         if (unlikely(bret->status == BLKIF_RSP_ERROR &&
1614                                      rinfo->shadow[id].req.u.rw.nr_segments == 0)) {
1615                                 printk(KERN_WARNING "blkfront: %s: empty %s op failed\n",
1616                                        info->gd->disk_name, op_name(bret->operation));
1617                                 error = -EOPNOTSUPP;
1618                         }
1619                         if (unlikely(error)) {
1620                                 if (error == -EOPNOTSUPP)
1621                                         error = 0;
1622                                 info->feature_flush = 0;
1623                                 xlvbd_flush(info);
1624                         }
1625                         /* fall through */
1626                 case BLKIF_OP_READ:
1627                 case BLKIF_OP_WRITE:
1628                         if (unlikely(bret->status != BLKIF_RSP_OKAY))
1629                                 dev_dbg(&info->xbdev->dev, "Bad return from blkdev data "
1630                                         "request: %x\n", bret->status);
1631
1632                         blk_mq_complete_request(req, error);
1633                         break;
1634                 default:
1635                         BUG();
1636                 }
1637         }
1638
1639         rinfo->ring.rsp_cons = i;
1640
1641         if (i != rinfo->ring.req_prod_pvt) {
1642                 int more_to_do;
1643                 RING_FINAL_CHECK_FOR_RESPONSES(&rinfo->ring, more_to_do);
1644                 if (more_to_do)
1645                         goto again;
1646         } else
1647                 rinfo->ring.sring->rsp_event = i + 1;
1648
1649         kick_pending_request_queues_locked(rinfo);
1650
1651         spin_unlock_irqrestore(&rinfo->ring_lock, flags);
1652
1653         return IRQ_HANDLED;
1654 }
1655
1656
1657 static int setup_blkring(struct xenbus_device *dev,
1658                          struct blkfront_ring_info *rinfo)
1659 {
1660         struct blkif_sring *sring;
1661         int err, i;
1662         struct blkfront_info *info = rinfo->dev_info;
1663         unsigned long ring_size = info->nr_ring_pages * XEN_PAGE_SIZE;
1664         grant_ref_t gref[XENBUS_MAX_RING_GRANTS];
1665
1666         for (i = 0; i < info->nr_ring_pages; i++)
1667                 rinfo->ring_ref[i] = GRANT_INVALID_REF;
1668
1669         sring = (struct blkif_sring *)__get_free_pages(GFP_NOIO | __GFP_HIGH,
1670                                                        get_order(ring_size));
1671         if (!sring) {
1672                 xenbus_dev_fatal(dev, -ENOMEM, "allocating shared ring");
1673                 return -ENOMEM;
1674         }
1675         SHARED_RING_INIT(sring);
1676         FRONT_RING_INIT(&rinfo->ring, sring, ring_size);
1677
1678         err = xenbus_grant_ring(dev, rinfo->ring.sring, info->nr_ring_pages, gref);
1679         if (err < 0) {
1680                 free_pages((unsigned long)sring, get_order(ring_size));
1681                 rinfo->ring.sring = NULL;
1682                 goto fail;
1683         }
1684         for (i = 0; i < info->nr_ring_pages; i++)
1685                 rinfo->ring_ref[i] = gref[i];
1686
1687         err = xenbus_alloc_evtchn(dev, &rinfo->evtchn);
1688         if (err)
1689                 goto fail;
1690
1691         err = bind_evtchn_to_irqhandler(rinfo->evtchn, blkif_interrupt, 0,
1692                                         "blkif", rinfo);
1693         if (err <= 0) {
1694                 xenbus_dev_fatal(dev, err,
1695                                  "bind_evtchn_to_irqhandler failed");
1696                 goto fail;
1697         }
1698         rinfo->irq = err;
1699
1700         return 0;
1701 fail:
1702         blkif_free(info, 0);
1703         return err;
1704 }
1705
1706 /*
1707  * Write out per-ring/queue nodes including ring-ref and event-channel, and each
1708  * ring buffer may have multi pages depending on ->nr_ring_pages.
1709  */
1710 static int write_per_ring_nodes(struct xenbus_transaction xbt,
1711                                 struct blkfront_ring_info *rinfo, const char *dir)
1712 {
1713         int err;
1714         unsigned int i;
1715         const char *message = NULL;
1716         struct blkfront_info *info = rinfo->dev_info;
1717
1718         if (info->nr_ring_pages == 1) {
1719                 err = xenbus_printf(xbt, dir, "ring-ref", "%u", rinfo->ring_ref[0]);
1720                 if (err) {
1721                         message = "writing ring-ref";
1722                         goto abort_transaction;
1723                 }
1724         } else {
1725                 for (i = 0; i < info->nr_ring_pages; i++) {
1726                         char ring_ref_name[RINGREF_NAME_LEN];
1727
1728                         snprintf(ring_ref_name, RINGREF_NAME_LEN, "ring-ref%u", i);
1729                         err = xenbus_printf(xbt, dir, ring_ref_name,
1730                                             "%u", rinfo->ring_ref[i]);
1731                         if (err) {
1732                                 message = "writing ring-ref";
1733                                 goto abort_transaction;
1734                         }
1735                 }
1736         }
1737
1738         err = xenbus_printf(xbt, dir, "event-channel", "%u", rinfo->evtchn);
1739         if (err) {
1740                 message = "writing event-channel";
1741                 goto abort_transaction;
1742         }
1743
1744         return 0;
1745
1746 abort_transaction:
1747         xenbus_transaction_end(xbt, 1);
1748         if (message)
1749                 xenbus_dev_fatal(info->xbdev, err, "%s", message);
1750
1751         return err;
1752 }
1753
1754 /* Common code used when first setting up, and when resuming. */
1755 static int talk_to_blkback(struct xenbus_device *dev,
1756                            struct blkfront_info *info)
1757 {
1758         const char *message = NULL;
1759         struct xenbus_transaction xbt;
1760         int err;
1761         unsigned int i, max_page_order = 0;
1762         unsigned int ring_page_order = 0;
1763
1764         err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
1765                            "max-ring-page-order", "%u", &max_page_order);
1766         if (err != 1)
1767                 info->nr_ring_pages = 1;
1768         else {
1769                 ring_page_order = min(xen_blkif_max_ring_order, max_page_order);
1770                 info->nr_ring_pages = 1 << ring_page_order;
1771         }
1772
1773         for (i = 0; i < info->nr_rings; i++) {
1774                 struct blkfront_ring_info *rinfo = &info->rinfo[i];
1775
1776                 /* Create shared ring, alloc event channel. */
1777                 err = setup_blkring(dev, rinfo);
1778                 if (err)
1779                         goto destroy_blkring;
1780         }
1781
1782 again:
1783         err = xenbus_transaction_start(&xbt);
1784         if (err) {
1785                 xenbus_dev_fatal(dev, err, "starting transaction");
1786                 goto destroy_blkring;
1787         }
1788
1789         if (info->nr_ring_pages > 1) {
1790                 err = xenbus_printf(xbt, dev->nodename, "ring-page-order", "%u",
1791                                     ring_page_order);
1792                 if (err) {
1793                         message = "writing ring-page-order";
1794                         goto abort_transaction;
1795                 }
1796         }
1797
1798         /* We already got the number of queues/rings in _probe */
1799         if (info->nr_rings == 1) {
1800                 err = write_per_ring_nodes(xbt, &info->rinfo[0], dev->nodename);
1801                 if (err)
1802                         goto destroy_blkring;
1803         } else {
1804                 char *path;
1805                 size_t pathsize;
1806
1807                 err = xenbus_printf(xbt, dev->nodename, "multi-queue-num-queues", "%u",
1808                                     info->nr_rings);
1809                 if (err) {
1810                         message = "writing multi-queue-num-queues";
1811                         goto abort_transaction;
1812                 }
1813
1814                 pathsize = strlen(dev->nodename) + QUEUE_NAME_LEN;
1815                 path = kmalloc(pathsize, GFP_KERNEL);
1816                 if (!path) {
1817                         err = -ENOMEM;
1818                         message = "ENOMEM while writing ring references";
1819                         goto abort_transaction;
1820                 }
1821
1822                 for (i = 0; i < info->nr_rings; i++) {
1823                         memset(path, 0, pathsize);
1824                         snprintf(path, pathsize, "%s/queue-%u", dev->nodename, i);
1825                         err = write_per_ring_nodes(xbt, &info->rinfo[i], path);
1826                         if (err) {
1827                                 kfree(path);
1828                                 goto destroy_blkring;
1829                         }
1830                 }
1831                 kfree(path);
1832         }
1833         err = xenbus_printf(xbt, dev->nodename, "protocol", "%s",
1834                             XEN_IO_PROTO_ABI_NATIVE);
1835         if (err) {
1836                 message = "writing protocol";
1837                 goto abort_transaction;
1838         }
1839         err = xenbus_printf(xbt, dev->nodename,
1840                             "feature-persistent", "%u", 1);
1841         if (err)
1842                 dev_warn(&dev->dev,
1843                          "writing persistent grants feature to xenbus");
1844
1845         err = xenbus_transaction_end(xbt, 0);
1846         if (err) {
1847                 if (err == -EAGAIN)
1848                         goto again;
1849                 xenbus_dev_fatal(dev, err, "completing transaction");
1850                 goto destroy_blkring;
1851         }
1852
1853         for (i = 0; i < info->nr_rings; i++) {
1854                 unsigned int j;
1855                 struct blkfront_ring_info *rinfo = &info->rinfo[i];
1856
1857                 for (j = 0; j < BLK_RING_SIZE(info); j++)
1858                         rinfo->shadow[j].req.u.rw.id = j + 1;
1859                 rinfo->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff;
1860         }
1861         xenbus_switch_state(dev, XenbusStateInitialised);
1862
1863         return 0;
1864
1865  abort_transaction:
1866         xenbus_transaction_end(xbt, 1);
1867         if (message)
1868                 xenbus_dev_fatal(dev, err, "%s", message);
1869  destroy_blkring:
1870         blkif_free(info, 0);
1871
1872         kfree(info);
1873         dev_set_drvdata(&dev->dev, NULL);
1874
1875         return err;
1876 }
1877
1878 static int negotiate_mq(struct blkfront_info *info)
1879 {
1880         unsigned int backend_max_queues = 0;
1881         int err;
1882         unsigned int i;
1883
1884         BUG_ON(info->nr_rings);
1885
1886         /* Check if backend supports multiple queues. */
1887         err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
1888                            "multi-queue-max-queues", "%u", &backend_max_queues);
1889         if (err < 0)
1890                 backend_max_queues = 1;
1891
1892         info->nr_rings = min(backend_max_queues, xen_blkif_max_queues);
1893         /* We need at least one ring. */
1894         if (!info->nr_rings)
1895                 info->nr_rings = 1;
1896
1897         info->rinfo = kzalloc(sizeof(struct blkfront_ring_info) * info->nr_rings, GFP_KERNEL);
1898         if (!info->rinfo) {
1899                 xenbus_dev_fatal(info->xbdev, -ENOMEM, "allocating ring_info structure");
1900                 return -ENOMEM;
1901         }
1902
1903         for (i = 0; i < info->nr_rings; i++) {
1904                 struct blkfront_ring_info *rinfo;
1905
1906                 rinfo = &info->rinfo[i];
1907                 INIT_LIST_HEAD(&rinfo->indirect_pages);
1908                 INIT_LIST_HEAD(&rinfo->grants);
1909                 rinfo->dev_info = info;
1910                 INIT_WORK(&rinfo->work, blkif_restart_queue);
1911                 spin_lock_init(&rinfo->ring_lock);
1912         }
1913         return 0;
1914 }
1915 /**
1916  * Entry point to this code when a new device is created.  Allocate the basic
1917  * structures and the ring buffer for communication with the backend, and
1918  * inform the backend of the appropriate details for those.  Switch to
1919  * Initialised state.
1920  */
1921 static int blkfront_probe(struct xenbus_device *dev,
1922                           const struct xenbus_device_id *id)
1923 {
1924         int err, vdevice;
1925         struct blkfront_info *info;
1926
1927         /* FIXME: Use dynamic device id if this is not set. */
1928         err = xenbus_scanf(XBT_NIL, dev->nodename,
1929                            "virtual-device", "%i", &vdevice);
1930         if (err != 1) {
1931                 /* go looking in the extended area instead */
1932                 err = xenbus_scanf(XBT_NIL, dev->nodename, "virtual-device-ext",
1933                                    "%i", &vdevice);
1934                 if (err != 1) {
1935                         xenbus_dev_fatal(dev, err, "reading virtual-device");
1936                         return err;
1937                 }
1938         }
1939
1940         if (xen_hvm_domain()) {
1941                 char *type;
1942                 int len;
1943                 /* no unplug has been done: do not hook devices != xen vbds */
1944                 if (xen_has_pv_and_legacy_disk_devices()) {
1945                         int major;
1946
1947                         if (!VDEV_IS_EXTENDED(vdevice))
1948                                 major = BLKIF_MAJOR(vdevice);
1949                         else
1950                                 major = XENVBD_MAJOR;
1951
1952                         if (major != XENVBD_MAJOR) {
1953                                 printk(KERN_INFO
1954                                                 "%s: HVM does not support vbd %d as xen block device\n",
1955                                                 __func__, vdevice);
1956                                 return -ENODEV;
1957                         }
1958                 }
1959                 /* do not create a PV cdrom device if we are an HVM guest */
1960                 type = xenbus_read(XBT_NIL, dev->nodename, "device-type", &len);
1961                 if (IS_ERR(type))
1962                         return -ENODEV;
1963                 if (strncmp(type, "cdrom", 5) == 0) {
1964                         kfree(type);
1965                         return -ENODEV;
1966                 }
1967                 kfree(type);
1968         }
1969         info = kzalloc(sizeof(*info), GFP_KERNEL);
1970         if (!info) {
1971                 xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure");
1972                 return -ENOMEM;
1973         }
1974
1975         info->xbdev = dev;
1976         err = negotiate_mq(info);
1977         if (err) {
1978                 kfree(info);
1979                 return err;
1980         }
1981
1982         mutex_init(&info->mutex);
1983         info->vdevice = vdevice;
1984         info->connected = BLKIF_STATE_DISCONNECTED;
1985
1986         /* Front end dir is a number, which is used as the id. */
1987         info->handle = simple_strtoul(strrchr(dev->nodename, '/')+1, NULL, 0);
1988         dev_set_drvdata(&dev->dev, info);
1989
1990         return 0;
1991 }
1992
1993 static void split_bio_end(struct bio *bio)
1994 {
1995         struct split_bio *split_bio = bio->bi_private;
1996
1997         if (atomic_dec_and_test(&split_bio->pending)) {
1998                 split_bio->bio->bi_phys_segments = 0;
1999                 split_bio->bio->bi_error = bio->bi_error;
2000                 bio_endio(split_bio->bio);
2001                 kfree(split_bio);
2002         }
2003         bio_put(bio);
2004 }
2005
2006 static int blkif_recover(struct blkfront_info *info)
2007 {
2008         unsigned int i, r_index;
2009         struct request *req, *n;
2010         struct blk_shadow *copy;
2011         int rc;
2012         struct bio *bio, *cloned_bio;
2013         struct bio_list bio_list, merge_bio;
2014         unsigned int segs, offset;
2015         int pending, size;
2016         struct split_bio *split_bio;
2017         struct list_head requests;
2018
2019         blkfront_gather_backend_features(info);
2020         segs = info->max_indirect_segments ? : BLKIF_MAX_SEGMENTS_PER_REQUEST;
2021         blk_queue_max_segments(info->rq, segs);
2022         bio_list_init(&bio_list);
2023         INIT_LIST_HEAD(&requests);
2024
2025         for (r_index = 0; r_index < info->nr_rings; r_index++) {
2026                 struct blkfront_ring_info *rinfo;
2027
2028                 rinfo = &info->rinfo[r_index];
2029                 /* Stage 1: Make a safe copy of the shadow state. */
2030                 copy = kmemdup(rinfo->shadow, sizeof(rinfo->shadow),
2031                                GFP_NOIO | __GFP_REPEAT | __GFP_HIGH);
2032                 if (!copy)
2033                         return -ENOMEM;
2034
2035                 /* Stage 2: Set up free list. */
2036                 memset(&rinfo->shadow, 0, sizeof(rinfo->shadow));
2037                 for (i = 0; i < BLK_RING_SIZE(info); i++)
2038                         rinfo->shadow[i].req.u.rw.id = i+1;
2039                 rinfo->shadow_free = rinfo->ring.req_prod_pvt;
2040                 rinfo->shadow[BLK_RING_SIZE(info)-1].req.u.rw.id = 0x0fffffff;
2041
2042                 rc = blkfront_setup_indirect(rinfo);
2043                 if (rc) {
2044                         kfree(copy);
2045                         return rc;
2046                 }
2047
2048                 for (i = 0; i < BLK_RING_SIZE(info); i++) {
2049                         /* Not in use? */
2050                         if (!copy[i].request)
2051                                 continue;
2052
2053                         /*
2054                          * Get the bios in the request so we can re-queue them.
2055                          */
2056                         if (copy[i].request->cmd_flags &
2057                             (REQ_FLUSH | REQ_FUA | REQ_DISCARD | REQ_SECURE)) {
2058                                 /*
2059                                  * Flush operations don't contain bios, so
2060                                  * we need to requeue the whole request
2061                                  */
2062                                 list_add(&copy[i].request->queuelist, &requests);
2063                                 continue;
2064                         }
2065                         merge_bio.head = copy[i].request->bio;
2066                         merge_bio.tail = copy[i].request->biotail;
2067                         bio_list_merge(&bio_list, &merge_bio);
2068                         copy[i].request->bio = NULL;
2069                         blk_end_request_all(copy[i].request, 0);
2070                 }
2071
2072                 kfree(copy);
2073         }
2074         xenbus_switch_state(info->xbdev, XenbusStateConnected);
2075
2076         /* Now safe for us to use the shared ring */
2077         info->connected = BLKIF_STATE_CONNECTED;
2078
2079         for (r_index = 0; r_index < info->nr_rings; r_index++) {
2080                 struct blkfront_ring_info *rinfo;
2081
2082                 rinfo = &info->rinfo[r_index];
2083                 /* Kick any other new requests queued since we resumed */
2084                 kick_pending_request_queues(rinfo);
2085         }
2086
2087         list_for_each_entry_safe(req, n, &requests, queuelist) {
2088                 /* Requeue pending requests (flush or discard) */
2089                 list_del_init(&req->queuelist);
2090                 BUG_ON(req->nr_phys_segments > segs);
2091                 blk_mq_requeue_request(req);
2092         }
2093         blk_mq_kick_requeue_list(info->rq);
2094
2095         while ((bio = bio_list_pop(&bio_list)) != NULL) {
2096                 /* Traverse the list of pending bios and re-queue them */
2097                 if (bio_segments(bio) > segs) {
2098                         /*
2099                          * This bio has more segments than what we can
2100                          * handle, we have to split it.
2101                          */
2102                         pending = (bio_segments(bio) + segs - 1) / segs;
2103                         split_bio = kzalloc(sizeof(*split_bio), GFP_NOIO);
2104                         BUG_ON(split_bio == NULL);
2105                         atomic_set(&split_bio->pending, pending);
2106                         split_bio->bio = bio;
2107                         for (i = 0; i < pending; i++) {
2108                                 offset = (i * segs * XEN_PAGE_SIZE) >> 9;
2109                                 size = min((unsigned int)(segs * XEN_PAGE_SIZE) >> 9,
2110                                            (unsigned int)bio_sectors(bio) - offset);
2111                                 cloned_bio = bio_clone(bio, GFP_NOIO);
2112                                 BUG_ON(cloned_bio == NULL);
2113                                 bio_trim(cloned_bio, offset, size);
2114                                 cloned_bio->bi_private = split_bio;
2115                                 cloned_bio->bi_end_io = split_bio_end;
2116                                 submit_bio(cloned_bio->bi_rw, cloned_bio);
2117                         }
2118                         /*
2119                          * Now we have to wait for all those smaller bios to
2120                          * end, so we can also end the "parent" bio.
2121                          */
2122                         continue;
2123                 }
2124                 /* We don't need to split this bio */
2125                 submit_bio(bio->bi_rw, bio);
2126         }
2127
2128         return 0;
2129 }
2130
2131 /**
2132  * We are reconnecting to the backend, due to a suspend/resume, or a backend
2133  * driver restart.  We tear down our blkif structure and recreate it, but
2134  * leave the device-layer structures intact so that this is transparent to the
2135  * rest of the kernel.
2136  */
2137 static int blkfront_resume(struct xenbus_device *dev)
2138 {
2139         struct blkfront_info *info = dev_get_drvdata(&dev->dev);
2140         int err = 0;
2141
2142         dev_dbg(&dev->dev, "blkfront_resume: %s\n", dev->nodename);
2143
2144         blkif_free(info, info->connected == BLKIF_STATE_CONNECTED);
2145
2146         err = negotiate_mq(info);
2147         if (err)
2148                 return err;
2149
2150         err = talk_to_blkback(dev, info);
2151
2152         /*
2153          * We have to wait for the backend to switch to
2154          * connected state, since we want to read which
2155          * features it supports.
2156          */
2157
2158         return err;
2159 }
2160
2161 static void blkfront_closing(struct blkfront_info *info)
2162 {
2163         struct xenbus_device *xbdev = info->xbdev;
2164         struct block_device *bdev = NULL;
2165
2166         mutex_lock(&info->mutex);
2167
2168         if (xbdev->state == XenbusStateClosing) {
2169                 mutex_unlock(&info->mutex);
2170                 return;
2171         }
2172
2173         if (info->gd)
2174                 bdev = bdget_disk(info->gd, 0);
2175
2176         mutex_unlock(&info->mutex);
2177
2178         if (!bdev) {
2179                 xenbus_frontend_closed(xbdev);
2180                 return;
2181         }
2182
2183         mutex_lock(&bdev->bd_mutex);
2184
2185         if (bdev->bd_openers) {
2186                 xenbus_dev_error(xbdev, -EBUSY,
2187                                  "Device in use; refusing to close");
2188                 xenbus_switch_state(xbdev, XenbusStateClosing);
2189         } else {
2190                 xlvbd_release_gendisk(info);
2191                 xenbus_frontend_closed(xbdev);
2192         }
2193
2194         mutex_unlock(&bdev->bd_mutex);
2195         bdput(bdev);
2196 }
2197
2198 static void blkfront_setup_discard(struct blkfront_info *info)
2199 {
2200         int err;
2201         unsigned int discard_granularity;
2202         unsigned int discard_alignment;
2203         unsigned int discard_secure;
2204
2205         info->feature_discard = 1;
2206         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2207                 "discard-granularity", "%u", &discard_granularity,
2208                 "discard-alignment", "%u", &discard_alignment,
2209                 NULL);
2210         if (!err) {
2211                 info->discard_granularity = discard_granularity;
2212                 info->discard_alignment = discard_alignment;
2213         }
2214         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2215                     "discard-secure", "%d", &discard_secure,
2216                     NULL);
2217         if (!err)
2218                 info->feature_secdiscard = !!discard_secure;
2219 }
2220
2221 static int blkfront_setup_indirect(struct blkfront_ring_info *rinfo)
2222 {
2223         unsigned int psegs, grants;
2224         int err, i;
2225         struct blkfront_info *info = rinfo->dev_info;
2226
2227         if (info->max_indirect_segments == 0) {
2228                 if (!HAS_EXTRA_REQ)
2229                         grants = BLKIF_MAX_SEGMENTS_PER_REQUEST;
2230                 else {
2231                         /*
2232                          * When an extra req is required, the maximum
2233                          * grants supported is related to the size of the
2234                          * Linux block segment.
2235                          */
2236                         grants = GRANTS_PER_PSEG;
2237                 }
2238         }
2239         else
2240                 grants = info->max_indirect_segments;
2241         psegs = grants / GRANTS_PER_PSEG;
2242
2243         err = fill_grant_buffer(rinfo,
2244                                 (grants + INDIRECT_GREFS(grants)) * BLK_RING_SIZE(info));
2245         if (err)
2246                 goto out_of_memory;
2247
2248         if (!info->feature_persistent && info->max_indirect_segments) {
2249                 /*
2250                  * We are using indirect descriptors but not persistent
2251                  * grants, we need to allocate a set of pages that can be
2252                  * used for mapping indirect grefs
2253                  */
2254                 int num = INDIRECT_GREFS(grants) * BLK_RING_SIZE(info);
2255
2256                 BUG_ON(!list_empty(&rinfo->indirect_pages));
2257                 for (i = 0; i < num; i++) {
2258                         struct page *indirect_page = alloc_page(GFP_NOIO);
2259                         if (!indirect_page)
2260                                 goto out_of_memory;
2261                         list_add(&indirect_page->lru, &rinfo->indirect_pages);
2262                 }
2263         }
2264
2265         for (i = 0; i < BLK_RING_SIZE(info); i++) {
2266                 rinfo->shadow[i].grants_used = kzalloc(
2267                         sizeof(rinfo->shadow[i].grants_used[0]) * grants,
2268                         GFP_NOIO);
2269                 rinfo->shadow[i].sg = kzalloc(sizeof(rinfo->shadow[i].sg[0]) * psegs, GFP_NOIO);
2270                 if (info->max_indirect_segments)
2271                         rinfo->shadow[i].indirect_grants = kzalloc(
2272                                 sizeof(rinfo->shadow[i].indirect_grants[0]) *
2273                                 INDIRECT_GREFS(grants),
2274                                 GFP_NOIO);
2275                 if ((rinfo->shadow[i].grants_used == NULL) ||
2276                         (rinfo->shadow[i].sg == NULL) ||
2277                      (info->max_indirect_segments &&
2278                      (rinfo->shadow[i].indirect_grants == NULL)))
2279                         goto out_of_memory;
2280                 sg_init_table(rinfo->shadow[i].sg, psegs);
2281         }
2282
2283
2284         return 0;
2285
2286 out_of_memory:
2287         for (i = 0; i < BLK_RING_SIZE(info); i++) {
2288                 kfree(rinfo->shadow[i].grants_used);
2289                 rinfo->shadow[i].grants_used = NULL;
2290                 kfree(rinfo->shadow[i].sg);
2291                 rinfo->shadow[i].sg = NULL;
2292                 kfree(rinfo->shadow[i].indirect_grants);
2293                 rinfo->shadow[i].indirect_grants = NULL;
2294         }
2295         if (!list_empty(&rinfo->indirect_pages)) {
2296                 struct page *indirect_page, *n;
2297                 list_for_each_entry_safe(indirect_page, n, &rinfo->indirect_pages, lru) {
2298                         list_del(&indirect_page->lru);
2299                         __free_page(indirect_page);
2300                 }
2301         }
2302         return -ENOMEM;
2303 }
2304
2305 /*
2306  * Gather all backend feature-*
2307  */
2308 static void blkfront_gather_backend_features(struct blkfront_info *info)
2309 {
2310         int err;
2311         int barrier, flush, discard, persistent;
2312         unsigned int indirect_segments;
2313
2314         info->feature_flush = 0;
2315
2316         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2317                         "feature-barrier", "%d", &barrier,
2318                         NULL);
2319
2320         /*
2321          * If there's no "feature-barrier" defined, then it means
2322          * we're dealing with a very old backend which writes
2323          * synchronously; nothing to do.
2324          *
2325          * If there are barriers, then we use flush.
2326          */
2327         if (!err && barrier)
2328                 info->feature_flush = REQ_FLUSH | REQ_FUA;
2329         /*
2330          * And if there is "feature-flush-cache" use that above
2331          * barriers.
2332          */
2333         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2334                         "feature-flush-cache", "%d", &flush,
2335                         NULL);
2336
2337         if (!err && flush)
2338                 info->feature_flush = REQ_FLUSH;
2339
2340         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2341                         "feature-discard", "%d", &discard,
2342                         NULL);
2343
2344         if (!err && discard)
2345                 blkfront_setup_discard(info);
2346
2347         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2348                         "feature-persistent", "%u", &persistent,
2349                         NULL);
2350         if (err)
2351                 info->feature_persistent = 0;
2352         else
2353                 info->feature_persistent = persistent;
2354
2355         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2356                             "feature-max-indirect-segments", "%u", &indirect_segments,
2357                             NULL);
2358         if (err)
2359                 info->max_indirect_segments = 0;
2360         else
2361                 info->max_indirect_segments = min(indirect_segments,
2362                                                   xen_blkif_max_segments);
2363 }
2364
2365 /*
2366  * Invoked when the backend is finally 'ready' (and has told produced
2367  * the details about the physical device - #sectors, size, etc).
2368  */
2369 static void blkfront_connect(struct blkfront_info *info)
2370 {
2371         unsigned long long sectors;
2372         unsigned long sector_size;
2373         unsigned int physical_sector_size;
2374         unsigned int binfo;
2375         int err, i;
2376
2377         switch (info->connected) {
2378         case BLKIF_STATE_CONNECTED:
2379                 /*
2380                  * Potentially, the back-end may be signalling
2381                  * a capacity change; update the capacity.
2382                  */
2383                 err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
2384                                    "sectors", "%Lu", &sectors);
2385                 if (XENBUS_EXIST_ERR(err))
2386                         return;
2387                 printk(KERN_INFO "Setting capacity to %Lu\n",
2388                        sectors);
2389                 set_capacity(info->gd, sectors);
2390                 revalidate_disk(info->gd);
2391
2392                 return;
2393         case BLKIF_STATE_SUSPENDED:
2394                 /*
2395                  * If we are recovering from suspension, we need to wait
2396                  * for the backend to announce it's features before
2397                  * reconnecting, at least we need to know if the backend
2398                  * supports indirect descriptors, and how many.
2399                  */
2400                 blkif_recover(info);
2401                 return;
2402
2403         default:
2404                 break;
2405         }
2406
2407         dev_dbg(&info->xbdev->dev, "%s:%s.\n",
2408                 __func__, info->xbdev->otherend);
2409
2410         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
2411                             "sectors", "%llu", &sectors,
2412                             "info", "%u", &binfo,
2413                             "sector-size", "%lu", &sector_size,
2414                             NULL);
2415         if (err) {
2416                 xenbus_dev_fatal(info->xbdev, err,
2417                                  "reading backend fields at %s",
2418                                  info->xbdev->otherend);
2419                 return;
2420         }
2421
2422         /*
2423          * physcial-sector-size is a newer field, so old backends may not
2424          * provide this. Assume physical sector size to be the same as
2425          * sector_size in that case.
2426          */
2427         err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
2428                            "physical-sector-size", "%u", &physical_sector_size);
2429         if (err != 1)
2430                 physical_sector_size = sector_size;
2431
2432         blkfront_gather_backend_features(info);
2433         for (i = 0; i < info->nr_rings; i++) {
2434                 err = blkfront_setup_indirect(&info->rinfo[i]);
2435                 if (err) {
2436                         xenbus_dev_fatal(info->xbdev, err, "setup_indirect at %s",
2437                                          info->xbdev->otherend);
2438                         blkif_free(info, 0);
2439                         break;
2440                 }
2441         }
2442
2443         err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size,
2444                                   physical_sector_size);
2445         if (err) {
2446                 xenbus_dev_fatal(info->xbdev, err, "xlvbd_add at %s",
2447                                  info->xbdev->otherend);
2448                 return;
2449         }
2450
2451         xenbus_switch_state(info->xbdev, XenbusStateConnected);
2452
2453         /* Kick pending requests. */
2454         info->connected = BLKIF_STATE_CONNECTED;
2455         for (i = 0; i < info->nr_rings; i++)
2456                 kick_pending_request_queues(&info->rinfo[i]);
2457
2458         add_disk(info->gd);
2459
2460         info->is_ready = 1;
2461 }
2462
2463 /**
2464  * Callback received when the backend's state changes.
2465  */
2466 static void blkback_changed(struct xenbus_device *dev,
2467                             enum xenbus_state backend_state)
2468 {
2469         struct blkfront_info *info = dev_get_drvdata(&dev->dev);
2470
2471         dev_dbg(&dev->dev, "blkfront:blkback_changed to state %d.\n", backend_state);
2472
2473         switch (backend_state) {
2474         case XenbusStateInitWait:
2475                 if (dev->state != XenbusStateInitialising)
2476                         break;
2477                 if (talk_to_blkback(dev, info))
2478                         break;
2479         case XenbusStateInitialising:
2480         case XenbusStateInitialised:
2481         case XenbusStateReconfiguring:
2482         case XenbusStateReconfigured:
2483         case XenbusStateUnknown:
2484                 break;
2485
2486         case XenbusStateConnected:
2487                 if (dev->state != XenbusStateInitialised) {
2488                         if (talk_to_blkback(dev, info))
2489                                 break;
2490                 }
2491                 blkfront_connect(info);
2492                 break;
2493
2494         case XenbusStateClosed:
2495                 if (dev->state == XenbusStateClosed)
2496                         break;
2497                 /* Missed the backend's Closing state -- fallthrough */
2498         case XenbusStateClosing:
2499                 if (info)
2500                         blkfront_closing(info);
2501                 break;
2502         }
2503 }
2504
2505 static int blkfront_remove(struct xenbus_device *xbdev)
2506 {
2507         struct blkfront_info *info = dev_get_drvdata(&xbdev->dev);
2508         struct block_device *bdev = NULL;
2509         struct gendisk *disk;
2510
2511         dev_dbg(&xbdev->dev, "%s removed", xbdev->nodename);
2512
2513         blkif_free(info, 0);
2514
2515         mutex_lock(&info->mutex);
2516
2517         disk = info->gd;
2518         if (disk)
2519                 bdev = bdget_disk(disk, 0);
2520
2521         info->xbdev = NULL;
2522         mutex_unlock(&info->mutex);
2523
2524         if (!bdev) {
2525                 kfree(info);
2526                 return 0;
2527         }
2528
2529         /*
2530          * The xbdev was removed before we reached the Closed
2531          * state. See if it's safe to remove the disk. If the bdev
2532          * isn't closed yet, we let release take care of it.
2533          */
2534
2535         mutex_lock(&bdev->bd_mutex);
2536         info = disk->private_data;
2537
2538         dev_warn(disk_to_dev(disk),
2539                  "%s was hot-unplugged, %d stale handles\n",
2540                  xbdev->nodename, bdev->bd_openers);
2541
2542         if (info && !bdev->bd_openers) {
2543                 xlvbd_release_gendisk(info);
2544                 disk->private_data = NULL;
2545                 kfree(info);
2546         }
2547
2548         mutex_unlock(&bdev->bd_mutex);
2549         bdput(bdev);
2550
2551         return 0;
2552 }
2553
2554 static int blkfront_is_ready(struct xenbus_device *dev)
2555 {
2556         struct blkfront_info *info = dev_get_drvdata(&dev->dev);
2557
2558         return info->is_ready && info->xbdev;
2559 }
2560
2561 static int blkif_open(struct block_device *bdev, fmode_t mode)
2562 {
2563         struct gendisk *disk = bdev->bd_disk;
2564         struct blkfront_info *info;
2565         int err = 0;
2566
2567         mutex_lock(&blkfront_mutex);
2568
2569         info = disk->private_data;
2570         if (!info) {
2571                 /* xbdev gone */
2572                 err = -ERESTARTSYS;
2573                 goto out;
2574         }
2575
2576         mutex_lock(&info->mutex);
2577
2578         if (!info->gd)
2579                 /* xbdev is closed */
2580                 err = -ERESTARTSYS;
2581
2582         mutex_unlock(&info->mutex);
2583
2584 out:
2585         mutex_unlock(&blkfront_mutex);
2586         return err;
2587 }
2588
2589 static void blkif_release(struct gendisk *disk, fmode_t mode)
2590 {
2591         struct blkfront_info *info = disk->private_data;
2592         struct block_device *bdev;
2593         struct xenbus_device *xbdev;
2594
2595         mutex_lock(&blkfront_mutex);
2596
2597         bdev = bdget_disk(disk, 0);
2598
2599         if (!bdev) {
2600                 WARN(1, "Block device %s yanked out from us!\n", disk->disk_name);
2601                 goto out_mutex;
2602         }
2603         if (bdev->bd_openers)
2604                 goto out;
2605
2606         /*
2607          * Check if we have been instructed to close. We will have
2608          * deferred this request, because the bdev was still open.
2609          */
2610
2611         mutex_lock(&info->mutex);
2612         xbdev = info->xbdev;
2613
2614         if (xbdev && xbdev->state == XenbusStateClosing) {
2615                 /* pending switch to state closed */
2616                 dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n");
2617                 xlvbd_release_gendisk(info);
2618                 xenbus_frontend_closed(info->xbdev);
2619         }
2620
2621         mutex_unlock(&info->mutex);
2622
2623         if (!xbdev) {
2624                 /* sudden device removal */
2625                 dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n");
2626                 xlvbd_release_gendisk(info);
2627                 disk->private_data = NULL;
2628                 kfree(info);
2629         }
2630
2631 out:
2632         bdput(bdev);
2633 out_mutex:
2634         mutex_unlock(&blkfront_mutex);
2635 }
2636
2637 static const struct block_device_operations xlvbd_block_fops =
2638 {
2639         .owner = THIS_MODULE,
2640         .open = blkif_open,
2641         .release = blkif_release,
2642         .getgeo = blkif_getgeo,
2643         .ioctl = blkif_ioctl,
2644 };
2645
2646
2647 static const struct xenbus_device_id blkfront_ids[] = {
2648         { "vbd" },
2649         { "" }
2650 };
2651
2652 static struct xenbus_driver blkfront_driver = {
2653         .ids  = blkfront_ids,
2654         .probe = blkfront_probe,
2655         .remove = blkfront_remove,
2656         .resume = blkfront_resume,
2657         .otherend_changed = blkback_changed,
2658         .is_ready = blkfront_is_ready,
2659 };
2660
2661 static int __init xlblk_init(void)
2662 {
2663         int ret;
2664         int nr_cpus = num_online_cpus();
2665
2666         if (!xen_domain())
2667                 return -ENODEV;
2668
2669         if (xen_blkif_max_ring_order > XENBUS_MAX_RING_GRANT_ORDER) {
2670                 pr_info("Invalid max_ring_order (%d), will use default max: %d.\n",
2671                         xen_blkif_max_ring_order, XENBUS_MAX_RING_GRANT_ORDER);
2672                 xen_blkif_max_ring_order = XENBUS_MAX_RING_GRANT_ORDER;
2673         }
2674
2675         if (xen_blkif_max_queues > nr_cpus) {
2676                 pr_info("Invalid max_queues (%d), will use default max: %d.\n",
2677                         xen_blkif_max_queues, nr_cpus);
2678                 xen_blkif_max_queues = nr_cpus;
2679         }
2680
2681         if (!xen_has_pv_disk_devices())
2682                 return -ENODEV;
2683
2684         if (register_blkdev(XENVBD_MAJOR, DEV_NAME)) {
2685                 printk(KERN_WARNING "xen_blk: can't get major %d with name %s\n",
2686                        XENVBD_MAJOR, DEV_NAME);
2687                 return -ENODEV;
2688         }
2689
2690         ret = xenbus_register_frontend(&blkfront_driver);
2691         if (ret) {
2692                 unregister_blkdev(XENVBD_MAJOR, DEV_NAME);
2693                 return ret;
2694         }
2695
2696         return 0;
2697 }
2698 module_init(xlblk_init);
2699
2700
2701 static void __exit xlblk_exit(void)
2702 {
2703         xenbus_unregister_driver(&blkfront_driver);
2704         unregister_blkdev(XENVBD_MAJOR, DEV_NAME);
2705         kfree(minors);
2706 }
2707 module_exit(xlblk_exit);
2708
2709 MODULE_DESCRIPTION("Xen virtual block device frontend");
2710 MODULE_LICENSE("GPL");
2711 MODULE_ALIAS_BLOCKDEV_MAJOR(XENVBD_MAJOR);
2712 MODULE_ALIAS("xen:vbd");
2713 MODULE_ALIAS("xenblk");