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/hdreg.h>
  41 #include <linux/cdrom.h>
  42 #include <linux/module.h>
  43 #include <linux/slab.h>
  44 #include <linux/mutex.h>
  45 #include <linux/scatterlist.h>
  46
  47 #include <xen/xen.h>
  48 #include <xen/xenbus.h>
  49 #include <xen/grant_table.h>
  50 #include <xen/events.h>
  51 #include <xen/page.h>
  52 #include <xen/platform_pci.h>
  53
  54 #include <xen/interface/grant_table.h>
  55 #include <xen/interface/io/blkif.h>
  56 #include <xen/interface/io/protocols.h>
  57
  58 #include <asm/xen/hypervisor.h>
  59
  60 enum blkif_state {
  61         BLKIF_STATE_DISCONNECTED,
  62         BLKIF_STATE_CONNECTED,
  63         BLKIF_STATE_SUSPENDED,
  64 };
  65
  66 struct blk_shadow {
  67         struct blkif_request req;
  68         struct request *request;
  69         unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST];
  70 };
  71
  72 static DEFINE_MUTEX(blkfront_mutex);
  73 static const struct block_device_operations xlvbd_block_fops;
  74
  75 #define BLK_RING_SIZE __CONST_RING_SIZE(blkif, PAGE_SIZE)
  76
  77 /*
  78  * We have one of these per vbd, whether ide, scsi or 'other'.  They
  79  * hang in private_data off the gendisk structure. We may end up
  80  * putting all kinds of interesting stuff here :-)
  81  */
  82 struct blkfront_info
  83 {
  84         struct mutex mutex;
  85         struct xenbus_device *xbdev;
  86         struct gendisk *gd;
  87         int vdevice;
  88         blkif_vdev_t handle;
  89         enum blkif_state connected;
  90         int ring_ref;
  91         struct blkif_front_ring ring;
  92         struct scatterlist sg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
  93         unsigned int evtchn, irq;
  94         struct request_queue *rq;
  95         struct work_struct work;
  96         struct gnttab_free_callback callback;
  97         struct blk_shadow shadow[BLK_RING_SIZE];
  98         unsigned long shadow_free;
  99         unsigned int feature_flush;
 100         unsigned int flush_op;
 101         unsigned int feature_discard;
 102         unsigned int discard_granularity;
 103         unsigned int discard_alignment;
 104         int is_ready;
 105 };
 106
 107 static DEFINE_SPINLOCK(blkif_io_lock);
 108
 109 static unsigned int nr_minors;
 110 static unsigned long *minors;
 111 static DEFINE_SPINLOCK(minor_lock);
 112
 113 #define MAXIMUM_OUTSTANDING_BLOCK_REQS \
 114         (BLKIF_MAX_SEGMENTS_PER_REQUEST * BLK_RING_SIZE)
 115 #define GRANT_INVALID_REF       0
 116
 117 #define PARTS_PER_DISK          16
 118 #define PARTS_PER_EXT_DISK      256
 119
 120 #define BLKIF_MAJOR(dev) ((dev)>>8)
 121 #define BLKIF_MINOR(dev) ((dev) & 0xff)
 122
 123 #define EXT_SHIFT 28
 124 #define EXTENDED (1<<EXT_SHIFT)
 125 #define VDEV_IS_EXTENDED(dev) ((dev)&(EXTENDED))
 126 #define BLKIF_MINOR_EXT(dev) ((dev)&(~EXTENDED))
 127 #define EMULATED_HD_DISK_MINOR_OFFSET (0)
 128 #define EMULATED_HD_DISK_NAME_OFFSET (EMULATED_HD_DISK_MINOR_OFFSET / 256)
 129 #define EMULATED_SD_DISK_MINOR_OFFSET (0)
 130 #define EMULATED_SD_DISK_NAME_OFFSET (EMULATED_SD_DISK_MINOR_OFFSET / 256)
 131
 132 #define DEV_NAME        "xvd"   /* name in /dev */
 133
 134 static int get_id_from_freelist(struct blkfront_info *info)
 135 {
 136         unsigned long free = info->shadow_free;
 137         BUG_ON(free >= BLK_RING_SIZE);
 138         info->shadow_free = info->shadow[free].req.id;
 139         info->shadow[free].req.id = 0x0fffffee; /* debug */
 140         return free;
 141 }
 142
 143 static void add_id_to_freelist(struct blkfront_info *info,
 144                                unsigned long id)
 145 {
 146         info->shadow[id].req.id  = info->shadow_free;
 147         info->shadow[id].request = NULL;
 148         info->shadow_free = id;
 149 }
 150
 151 static int xlbd_reserve_minors(unsigned int minor, unsigned int nr)
 152 {
 153         unsigned int end = minor + nr;
 154         int rc;
 155
 156         if (end > nr_minors) {
 157                 unsigned long *bitmap, *old;
 158
 159                 bitmap = kzalloc(BITS_TO_LONGS(end) * sizeof(*bitmap),
 160                                  GFP_KERNEL);
 161                 if (bitmap == NULL)
 162                         return -ENOMEM;
 163
 164                 spin_lock(&minor_lock);
 165                 if (end > nr_minors) {
 166                         old = minors;
 167                         memcpy(bitmap, minors,
 168                                BITS_TO_LONGS(nr_minors) * sizeof(*bitmap));
 169                         minors = bitmap;
 170                         nr_minors = BITS_TO_LONGS(end) * BITS_PER_LONG;
 171                 } else
 172                         old = bitmap;
 173                 spin_unlock(&minor_lock);
 174                 kfree(old);
 175         }
 176
 177         spin_lock(&minor_lock);
 178         if (find_next_bit(minors, end, minor) >= end) {
 179                 for (; minor < end; ++minor)
 180                         __set_bit(minor, minors);
 181                 rc = 0;
 182         } else
 183                 rc = -EBUSY;
 184         spin_unlock(&minor_lock);
 185
 186         return rc;
 187 }
 188
 189 static void xlbd_release_minors(unsigned int minor, unsigned int nr)
 190 {
 191         unsigned int end = minor + nr;
 192
 193         BUG_ON(end > nr_minors);
 194         spin_lock(&minor_lock);
 195         for (; minor < end; ++minor)
 196                 __clear_bit(minor, minors);
 197         spin_unlock(&minor_lock);
 198 }
 199
 200 static void blkif_restart_queue_callback(void *arg)
 201 {
 202         struct blkfront_info *info = (struct blkfront_info *)arg;
 203         schedule_work(&info->work);
 204 }
 205
 206 static int blkif_getgeo(struct block_device *bd, struct hd_geometry *hg)
 207 {
 208         /* We don't have real geometry info, but let's at least return
 209            values consistent with the size of the device */
 210         sector_t nsect = get_capacity(bd->bd_disk);
 211         sector_t cylinders = nsect;
 212
 213         hg->heads = 0xff;
 214         hg->sectors = 0x3f;
 215         sector_div(cylinders, hg->heads * hg->sectors);
 216         hg->cylinders = cylinders;
 217         if ((sector_t)(hg->cylinders + 1) * hg->heads * hg->sectors < nsect)
 218                 hg->cylinders = 0xffff;
 219         return 0;
 220 }
 221
 222 static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
 223                        unsigned command, unsigned long argument)
 224 {
 225         struct blkfront_info *info = bdev->bd_disk->private_data;
 226         int i;
 227
 228         dev_dbg(&info->xbdev->dev, "command: 0x%x, argument: 0x%lx\n",
 229                 command, (long)argument);
 230
 231         switch (command) {
 232         case CDROMMULTISESSION:
 233                 dev_dbg(&info->xbdev->dev, "FIXME: support multisession CDs later\n");
 234                 for (i = 0; i < sizeof(struct cdrom_multisession); i++)
 235                         if (put_user(0, (char __user *)(argument + i)))
 236                                 return -EFAULT;
 237                 return 0;
 238
 239         case CDROM_GET_CAPABILITY: {
 240                 struct gendisk *gd = info->gd;
 241                 if (gd->flags & GENHD_FL_CD)
 242                         return 0;
 243                 return -EINVAL;
 244         }
 245
 246         default:
 247                 /*printk(KERN_ALERT "ioctl %08x not supported by Xen blkdev\n",
 248                   command);*/
 249                 return -EINVAL; /* same return as native Linux */
 250         }
 251
 252         return 0;
 253 }
 254
 255 /*
 256  * Generate a Xen blkfront IO request from a blk layer request.  Reads
 257  * and writes are handled as expected.
 258  *
 259  * @req: a request struct
 260  */
 261 static int blkif_queue_request(struct request *req)
 262 {
 263         struct blkfront_info *info = req->rq_disk->private_data;
 264         unsigned long buffer_mfn;
 265         struct blkif_request *ring_req;
 266         unsigned long id;
 267         unsigned int fsect, lsect;
 268         int i, ref;
 269         grant_ref_t gref_head;
 270         struct scatterlist *sg;
 271
 272         if (unlikely(info->connected != BLKIF_STATE_CONNECTED))
 273                 return 1;
 274
 275         if (gnttab_alloc_grant_references(
 276                 BLKIF_MAX_SEGMENTS_PER_REQUEST, &gref_head) < 0) {
 277                 gnttab_request_free_callback(
 278                         &info->callback,
 279                         blkif_restart_queue_callback,
 280                         info,
 281                         BLKIF_MAX_SEGMENTS_PER_REQUEST);
 282                 return 1;
 283         }
 284
 285         /* Fill out a communications ring structure. */
 286         ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
 287         id = get_id_from_freelist(info);
 288         info->shadow[id].request = req;
 289
 290         ring_req->id = id;
 291         ring_req->u.rw.sector_number = (blkif_sector_t)blk_rq_pos(req);
 292         ring_req->handle = info->handle;
 293
 294         ring_req->operation = rq_data_dir(req) ?
 295                 BLKIF_OP_WRITE : BLKIF_OP_READ;
 296
 297         if (req->cmd_flags & (REQ_FLUSH | REQ_FUA)) {
 298                 /*
 299                  * Ideally we can do an unordered flush-to-disk. In case the
 300                  * backend onlysupports barriers, use that. A barrier request
 301                  * a superset of FUA, so we can implement it the same
 302                  * way.  (It's also a FLUSH+FUA, since it is
 303                  * guaranteed ordered WRT previous writes.)
 304                  */
 305                 ring_req->operation = info->flush_op;
 306         }
 307
 308         if (unlikely(req->cmd_flags & REQ_DISCARD)) {
 309                 /* id, sector_number and handle are set above. */
 310                 ring_req->operation = BLKIF_OP_DISCARD;
 311                 ring_req->nr_segments = 0;
 312                 ring_req->u.discard.nr_sectors = blk_rq_sectors(req);
 313         } else {
 314                 ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg);
 315                 BUG_ON(ring_req->nr_segments > BLKIF_MAX_SEGMENTS_PER_REQUEST);
 316
 317                 for_each_sg(info->sg, sg, ring_req->nr_segments, i) {
 318                         buffer_mfn = pfn_to_mfn(page_to_pfn(sg_page(sg)));
 319                         fsect = sg->offset >> 9;
 320                         lsect = fsect + (sg->length >> 9) - 1;
 321                         /* install a grant reference. */
 322                         ref = gnttab_claim_grant_reference(&gref_head);
 323                         BUG_ON(ref == -ENOSPC);
 324
 325                         gnttab_grant_foreign_access_ref(
 326                                         ref,
 327                                         info->xbdev->otherend_id,
 328                                         buffer_mfn,
 329                                         rq_data_dir(req));
 330
 331                         info->shadow[id].frame[i] = mfn_to_pfn(buffer_mfn);
 332                         ring_req->u.rw.seg[i] =
 333                                         (struct blkif_request_segment) {
 334                                                 .gref       = ref,
 335                                                 .first_sect = fsect,
 336                                                 .last_sect  = lsect };
 337                 }
 338         }
 339
 340         info->ring.req_prod_pvt++;
 341
 342         /* Keep a private copy so we can reissue requests when recovering. */
 343         info->shadow[id].req = *ring_req;
 344
 345         gnttab_free_grant_references(gref_head);
 346
 347         return 0;
 348 }
 349
 350
 351 static inline void flush_requests(struct blkfront_info *info)
 352 {
 353         int notify;
 354
 355         RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&info->ring, notify);
 356
 357         if (notify)
 358                 notify_remote_via_irq(info->irq);
 359 }
 360
 361 /*
 362  * do_blkif_request
 363  *  read a block; request is in a request queue
 364  */
 365 static void do_blkif_request(struct request_queue *rq)
 366 {
 367         struct blkfront_info *info = NULL;
 368         struct request *req;
 369         int queued;
 370
 371         pr_debug("Entered do_blkif_request\n");
 372
 373         queued = 0;
 374
 375         while ((req = blk_peek_request(rq)) != NULL) {
 376                 info = req->rq_disk->private_data;
 377
 378                 if (RING_FULL(&info->ring))
 379                         goto wait;
 380
 381                 blk_start_request(req);
 382
 383                 if ((req->cmd_type != REQ_TYPE_FS) ||
 384                     ((req->cmd_flags & (REQ_FLUSH | REQ_FUA)) &&
 385                     !info->flush_op)) {
 386                         __blk_end_request_all(req, -EIO);
 387                         continue;
 388                 }
 389
 390                 pr_debug("do_blk_req %p: cmd %p, sec %lx, "
 391                          "(%u/%u) buffer:%p [%s]\n",
 392                          req, req->cmd, (unsigned long)blk_rq_pos(req),
 393                          blk_rq_cur_sectors(req), blk_rq_sectors(req),
 394                          req->buffer, rq_data_dir(req) ? "write" : "read");
 395
 396                 if (blkif_queue_request(req)) {
 397                         blk_requeue_request(rq, req);
 398 wait:
 399                         /* Avoid pointless unplugs. */
 400                         blk_stop_queue(rq);
 401                         break;
 402                 }
 403
 404                 queued++;
 405         }
 406
 407         if (queued != 0)
 408                 flush_requests(info);
 409 }
 410
 411 static int xlvbd_init_blk_queue(struct gendisk *gd, u16 sector_size)
 412 {
 413         struct request_queue *rq;
 414         struct blkfront_info *info = gd->private_data;
 415
 416         rq = blk_init_queue(do_blkif_request, &blkif_io_lock);
 417         if (rq == NULL)
 418                 return -1;
 419
 420         queue_flag_set_unlocked(QUEUE_FLAG_VIRT, rq);
 421
 422         if (info->feature_discard) {
 423                 queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, rq);
 424                 blk_queue_max_discard_sectors(rq, get_capacity(gd));
 425                 rq->limits.discard_granularity = info->discard_granularity;
 426                 rq->limits.discard_alignment = info->discard_alignment;
 427         }
 428
 429         /* Hard sector size and max sectors impersonate the equiv. hardware. */
 430         blk_queue_logical_block_size(rq, sector_size);
 431         blk_queue_max_hw_sectors(rq, 512);
 432
 433         /* Each segment in a request is up to an aligned page in size. */
 434         blk_queue_segment_boundary(rq, PAGE_SIZE - 1);
 435         blk_queue_max_segment_size(rq, PAGE_SIZE);
 436
 437         /* Ensure a merged request will fit in a single I/O ring slot. */
 438         blk_queue_max_segments(rq, BLKIF_MAX_SEGMENTS_PER_REQUEST);
 439
 440         /* Make sure buffer addresses are sector-aligned. */
 441         blk_queue_dma_alignment(rq, 511);
 442
 443         /* Make sure we don't use bounce buffers. */
 444         blk_queue_bounce_limit(rq, BLK_BOUNCE_ANY);
 445
 446         gd->queue = rq;
 447
 448         return 0;
 449 }
 450
 451
 452 static void xlvbd_flush(struct blkfront_info *info)
 453 {
 454         blk_queue_flush(info->rq, info->feature_flush);
 455         printk(KERN_INFO "blkfront: %s: %s: %s\n",
 456                info->gd->disk_name,
 457                info->flush_op == BLKIF_OP_WRITE_BARRIER ?
 458                 "barrier" : (info->flush_op == BLKIF_OP_FLUSH_DISKCACHE ?
 459                 "flush diskcache" : "barrier or flush"),
 460                info->feature_flush ? "enabled" : "disabled");
 461 }
 462
 463 static int xen_translate_vdev(int vdevice, int *minor, unsigned int *offset)
 464 {
 465         int major;
 466         major = BLKIF_MAJOR(vdevice);
 467         *minor = BLKIF_MINOR(vdevice);
 468         switch (major) {
 469                 case XEN_IDE0_MAJOR:
 470                         *offset = (*minor / 64) + EMULATED_HD_DISK_NAME_OFFSET;
 471                         *minor = ((*minor / 64) * PARTS_PER_DISK) +
 472                                 EMULATED_HD_DISK_MINOR_OFFSET;
 473                         break;
 474                 case XEN_IDE1_MAJOR:
 475                         *offset = (*minor / 64) + 2 + EMULATED_HD_DISK_NAME_OFFSET;
 476                         *minor = (((*minor / 64) + 2) * PARTS_PER_DISK) +
 477                                 EMULATED_HD_DISK_MINOR_OFFSET;
 478                         break;
 479                 case XEN_SCSI_DISK0_MAJOR:
 480                         *offset = (*minor / PARTS_PER_DISK) + EMULATED_SD_DISK_NAME_OFFSET;
 481                         *minor = *minor + EMULATED_SD_DISK_MINOR_OFFSET;
 482                         break;
 483                 case XEN_SCSI_DISK1_MAJOR:
 484                 case XEN_SCSI_DISK2_MAJOR:
 485                 case XEN_SCSI_DISK3_MAJOR:
 486                 case XEN_SCSI_DISK4_MAJOR:
 487                 case XEN_SCSI_DISK5_MAJOR:
 488                 case XEN_SCSI_DISK6_MAJOR:
 489                 case XEN_SCSI_DISK7_MAJOR:
 490                         *offset = (*minor / PARTS_PER_DISK) +
 491                                 ((major - XEN_SCSI_DISK1_MAJOR + 1) * 16) +
 492                                 EMULATED_SD_DISK_NAME_OFFSET;
 493                         *minor = *minor +
 494                                 ((major - XEN_SCSI_DISK1_MAJOR + 1) * 16 * PARTS_PER_DISK) +
 495                                 EMULATED_SD_DISK_MINOR_OFFSET;
 496                         break;
 497                 case XEN_SCSI_DISK8_MAJOR:
 498                 case XEN_SCSI_DISK9_MAJOR:
 499                 case XEN_SCSI_DISK10_MAJOR:
 500                 case XEN_SCSI_DISK11_MAJOR:
 501                 case XEN_SCSI_DISK12_MAJOR:
 502                 case XEN_SCSI_DISK13_MAJOR:
 503                 case XEN_SCSI_DISK14_MAJOR:
 504                 case XEN_SCSI_DISK15_MAJOR:
 505                         *offset = (*minor / PARTS_PER_DISK) +
 506                                 ((major - XEN_SCSI_DISK8_MAJOR + 8) * 16) +
 507                                 EMULATED_SD_DISK_NAME_OFFSET;
 508                         *minor = *minor +
 509                                 ((major - XEN_SCSI_DISK8_MAJOR + 8) * 16 * PARTS_PER_DISK) +
 510                                 EMULATED_SD_DISK_MINOR_OFFSET;
 511                         break;
 512                 case XENVBD_MAJOR:
 513                         *offset = *minor / PARTS_PER_DISK;
 514                         break;
 515                 default:
 516                         printk(KERN_WARNING "blkfront: your disk configuration is "
 517                                         "incorrect, please use an xvd device instead\n");
 518                         return -ENODEV;
 519         }
 520         return 0;
 521 }
 522
 523 static int xlvbd_alloc_gendisk(blkif_sector_t capacity,
 524                                struct blkfront_info *info,
 525                                u16 vdisk_info, u16 sector_size)
 526 {
 527         struct gendisk *gd;
 528         int nr_minors = 1;
 529         int err;
 530         unsigned int offset;
 531         int minor;
 532         int nr_parts;
 533
 534         BUG_ON(info->gd != NULL);
 535         BUG_ON(info->rq != NULL);
 536
 537         if ((info->vdevice>>EXT_SHIFT) > 1) {
 538                 /* this is above the extended range; something is wrong */
 539                 printk(KERN_WARNING "blkfront: vdevice 0x%x is above the extended range; ignoring\n", info->vdevice);
 540                 return -ENODEV;
 541         }
 542
 543         if (!VDEV_IS_EXTENDED(info->vdevice)) {
 544                 err = xen_translate_vdev(info->vdevice, &minor, &offset);
 545                 if (err)
 546                         return err;
 547                 nr_parts = PARTS_PER_DISK;
 548         } else {
 549                 minor = BLKIF_MINOR_EXT(info->vdevice);
 550                 nr_parts = PARTS_PER_EXT_DISK;
 551                 offset = minor / nr_parts;
 552                 if (xen_hvm_domain() && offset < EMULATED_HD_DISK_NAME_OFFSET + 4)
 553                         printk(KERN_WARNING "blkfront: vdevice 0x%x might conflict with "
 554                                         "emulated IDE disks,\n\t choose an xvd device name"
 555                                         "from xvde on\n", info->vdevice);
 556         }
 557         err = -ENODEV;
 558
 559         if ((minor % nr_parts) == 0)
 560                 nr_minors = nr_parts;
 561
 562         err = xlbd_reserve_minors(minor, nr_minors);
 563         if (err)
 564                 goto out;
 565         err = -ENODEV;
 566
 567         gd = alloc_disk(nr_minors);
 568         if (gd == NULL)
 569                 goto release;
 570
 571         if (nr_minors > 1) {
 572                 if (offset < 26)
 573                         sprintf(gd->disk_name, "%s%c", DEV_NAME, 'a' + offset);
 574                 else
 575                         sprintf(gd->disk_name, "%s%c%c", DEV_NAME,
 576                                 'a' + ((offset / 26)-1), 'a' + (offset % 26));
 577         } else {
 578                 if (offset < 26)
 579                         sprintf(gd->disk_name, "%s%c%d", DEV_NAME,
 580                                 'a' + offset,
 581                                 minor & (nr_parts - 1));
 582                 else
 583                         sprintf(gd->disk_name, "%s%c%c%d", DEV_NAME,
 584                                 'a' + ((offset / 26) - 1),
 585                                 'a' + (offset % 26),
 586                                 minor & (nr_parts - 1));
 587         }
 588
 589         gd->major = XENVBD_MAJOR;
 590         gd->first_minor = minor;
 591         gd->fops = &xlvbd_block_fops;
 592         gd->private_data = info;
 593         gd->driverfs_dev = &(info->xbdev->dev);
 594         set_capacity(gd, capacity);
 595
 596         if (xlvbd_init_blk_queue(gd, sector_size)) {
 597                 del_gendisk(gd);
 598                 goto release;
 599         }
 600
 601         info->rq = gd->queue;
 602         info->gd = gd;
 603
 604         xlvbd_flush(info);
 605
 606         if (vdisk_info & VDISK_READONLY)
 607                 set_disk_ro(gd, 1);
 608
 609         if (vdisk_info & VDISK_REMOVABLE)
 610                 gd->flags |= GENHD_FL_REMOVABLE;
 611
 612         if (vdisk_info & VDISK_CDROM)
 613                 gd->flags |= GENHD_FL_CD;
 614
 615         return 0;
 616
 617  release:
 618         xlbd_release_minors(minor, nr_minors);
 619  out:
 620         return err;
 621 }
 622
 623 static void xlvbd_release_gendisk(struct blkfront_info *info)
 624 {
 625         unsigned int minor, nr_minors;
 626         unsigned long flags;
 627
 628         if (info->rq == NULL)
 629                 return;
 630
 631         spin_lock_irqsave(&blkif_io_lock, flags);
 632
 633         /* No more blkif_request(). */
 634         blk_stop_queue(info->rq);
 635
 636         /* No more gnttab callback work. */
 637         gnttab_cancel_free_callback(&info->callback);
 638         spin_unlock_irqrestore(&blkif_io_lock, flags);
 639
 640         /* Flush gnttab callback work. Must be done with no locks held. */
 641         flush_work_sync(&info->work);
 642
 643         del_gendisk(info->gd);
 644
 645         minor = info->gd->first_minor;
 646         nr_minors = info->gd->minors;
 647         xlbd_release_minors(minor, nr_minors);
 648
 649         blk_cleanup_queue(info->rq);
 650         info->rq = NULL;
 651
 652         put_disk(info->gd);
 653         info->gd = NULL;
 654 }
 655
 656 static void kick_pending_request_queues(struct blkfront_info *info)
 657 {
 658         if (!RING_FULL(&info->ring)) {
 659                 /* Re-enable calldowns. */
 660                 blk_start_queue(info->rq);
 661                 /* Kick things off immediately. */
 662                 do_blkif_request(info->rq);
 663         }
 664 }
 665
 666 static void blkif_restart_queue(struct work_struct *work)
 667 {
 668         struct blkfront_info *info = container_of(work, struct blkfront_info, work);
 669
 670         spin_lock_irq(&blkif_io_lock);
 671         if (info->connected == BLKIF_STATE_CONNECTED)
 672                 kick_pending_request_queues(info);
 673         spin_unlock_irq(&blkif_io_lock);
 674 }
 675
 676 static void blkif_free(struct blkfront_info *info, int suspend)
 677 {
 678         /* Prevent new requests being issued until we fix things up. */
 679         spin_lock_irq(&blkif_io_lock);
 680         info->connected = suspend ?
 681                 BLKIF_STATE_SUSPENDED : BLKIF_STATE_DISCONNECTED;
 682         /* No more blkif_request(). */
 683         if (info->rq)
 684                 blk_stop_queue(info->rq);
 685         /* No more gnttab callback work. */
 686         gnttab_cancel_free_callback(&info->callback);
 687         spin_unlock_irq(&blkif_io_lock);
 688
 689         /* Flush gnttab callback work. Must be done with no locks held. */
 690         flush_work_sync(&info->work);
 691
 692         /* Free resources associated with old device channel. */
 693         if (info->ring_ref != GRANT_INVALID_REF) {
 694                 gnttab_end_foreign_access(info->ring_ref, 0,
 695                                           (unsigned long)info->ring.sring);
 696                 info->ring_ref = GRANT_INVALID_REF;
 697                 info->ring.sring = NULL;
 698         }
 699         if (info->irq)
 700                 unbind_from_irqhandler(info->irq, info);
 701         info->evtchn = info->irq = 0;
 702
 703 }
 704
 705 static void blkif_completion(struct blk_shadow *s)
 706 {
 707         int i;
 708         for (i = 0; i < s->req.nr_segments; i++)
 709                 gnttab_end_foreign_access(s->req.u.rw.seg[i].gref, 0, 0UL);
 710 }
 711
 712 static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 713 {
 714         struct request *req;
 715         struct blkif_response *bret;
 716         RING_IDX i, rp;
 717         unsigned long flags;
 718         struct blkfront_info *info = (struct blkfront_info *)dev_id;
 719         int error;
 720
 721         spin_lock_irqsave(&blkif_io_lock, flags);
 722
 723         if (unlikely(info->connected != BLKIF_STATE_CONNECTED)) {
 724                 spin_unlock_irqrestore(&blkif_io_lock, flags);
 725                 return IRQ_HANDLED;
 726         }
 727
 728  again:
 729         rp = info->ring.sring->rsp_prod;
 730         rmb(); /* Ensure we see queued responses up to 'rp'. */
 731
 732         for (i = info->ring.rsp_cons; i != rp; i++) {
 733                 unsigned long id;
 734
 735                 bret = RING_GET_RESPONSE(&info->ring, i);
 736                 id   = bret->id;
 737                 req  = info->shadow[id].request;
 738
 739                 blkif_completion(&info->shadow[id]);
 740
 741                 add_id_to_freelist(info, id);
 742
 743                 error = (bret->status == BLKIF_RSP_OKAY) ? 0 : -EIO;
 744                 switch (bret->operation) {
 745                 case BLKIF_OP_DISCARD:
 746                         if (unlikely(bret->status == BLKIF_RSP_EOPNOTSUPP)) {
 747                                 struct request_queue *rq = info->rq;
 748                                 printk(KERN_WARNING "blkfront: %s: discard op failed\n",
 749                                            info->gd->disk_name);
 750                                 error = -EOPNOTSUPP;
 751                                 info->feature_discard = 0;
 752                                 queue_flag_clear(QUEUE_FLAG_DISCARD, rq);
 753                         }
 754                         __blk_end_request_all(req, error);
 755                         break;
 756                 case BLKIF_OP_FLUSH_DISKCACHE:
 757                 case BLKIF_OP_WRITE_BARRIER:
 758                         if (unlikely(bret->status == BLKIF_RSP_EOPNOTSUPP)) {
 759                                 printk(KERN_WARNING "blkfront: %s: write %s op failed\n",
 760                                        info->flush_op == BLKIF_OP_WRITE_BARRIER ?
 761                                        "barrier" :  "flush disk cache",
 762                                        info->gd->disk_name);
 763                                 error = -EOPNOTSUPP;
 764                         }
 765                         if (unlikely(bret->status == BLKIF_RSP_ERROR &&
 766                                      info->shadow[id].req.nr_segments == 0)) {
 767                                 printk(KERN_WARNING "blkfront: %s: empty write %s op failed\n",
 768                                        info->flush_op == BLKIF_OP_WRITE_BARRIER ?
 769                                        "barrier" :  "flush disk cache",
 770                                        info->gd->disk_name);
 771                                 error = -EOPNOTSUPP;
 772                         }
 773                         if (unlikely(error)) {
 774                                 if (error == -EOPNOTSUPP)
 775                                         error = 0;
 776                                 info->feature_flush = 0;
 777                                 info->flush_op = 0;
 778                                 xlvbd_flush(info);
 779                         }
 780                         /* fall through */
 781                 case BLKIF_OP_READ:
 782                 case BLKIF_OP_WRITE:
 783                         if (unlikely(bret->status != BLKIF_RSP_OKAY))
 784                                 dev_dbg(&info->xbdev->dev, "Bad return from blkdev data "
 785                                         "request: %x\n", bret->status);
 786
 787                         __blk_end_request_all(req, error);
 788                         break;
 789                 default:
 790                         BUG();
 791                 }
 792         }
 793
 794         info->ring.rsp_cons = i;
 795
 796         if (i != info->ring.req_prod_pvt) {
 797                 int more_to_do;
 798                 RING_FINAL_CHECK_FOR_RESPONSES(&info->ring, more_to_do);
 799                 if (more_to_do)
 800                         goto again;
 801         } else
 802                 info->ring.sring->rsp_event = i + 1;
 803
 804         kick_pending_request_queues(info);
 805
 806         spin_unlock_irqrestore(&blkif_io_lock, flags);
 807
 808         return IRQ_HANDLED;
 809 }
 810
 811
 812 static int setup_blkring(struct xenbus_device *dev,
 813                          struct blkfront_info *info)
 814 {
 815         struct blkif_sring *sring;
 816         int err;
 817
 818         info->ring_ref = GRANT_INVALID_REF;
 819
 820         sring = (struct blkif_sring *)__get_free_page(GFP_NOIO | __GFP_HIGH);
 821         if (!sring) {
 822                 xenbus_dev_fatal(dev, -ENOMEM, "allocating shared ring");
 823                 return -ENOMEM;
 824         }
 825         SHARED_RING_INIT(sring);
 826         FRONT_RING_INIT(&info->ring, sring, PAGE_SIZE);
 827
 828         sg_init_table(info->sg, BLKIF_MAX_SEGMENTS_PER_REQUEST);
 829
 830         err = xenbus_grant_ring(dev, virt_to_mfn(info->ring.sring));
 831         if (err < 0) {
 832                 free_page((unsigned long)sring);
 833                 info->ring.sring = NULL;
 834                 goto fail;
 835         }
 836         info->ring_ref = err;
 837
 838         err = xenbus_alloc_evtchn(dev, &info->evtchn);
 839         if (err)
 840                 goto fail;
 841
 842         err = bind_evtchn_to_irqhandler(info->evtchn,
 843                                         blkif_interrupt,
 844                                         IRQF_SAMPLE_RANDOM, "blkif", info);
 845         if (err <= 0) {
 846                 xenbus_dev_fatal(dev, err,
 847                                  "bind_evtchn_to_irqhandler failed");
 848                 goto fail;
 849         }
 850         info->irq = err;
 851
 852         return 0;
 853 fail:
 854         blkif_free(info, 0);
 855         return err;
 856 }
 857
 858
 859 /* Common code used when first setting up, and when resuming. */
 860 static int talk_to_blkback(struct xenbus_device *dev,
 861                            struct blkfront_info *info)
 862 {
 863         const char *message = NULL;
 864         struct xenbus_transaction xbt;
 865         int err;
 866
 867         /* Create shared ring, alloc event channel. */
 868         err = setup_blkring(dev, info);
 869         if (err)
 870                 goto out;
 871
 872 again:
 873         err = xenbus_transaction_start(&xbt);
 874         if (err) {
 875                 xenbus_dev_fatal(dev, err, "starting transaction");
 876                 goto destroy_blkring;
 877         }
 878
 879         err = xenbus_printf(xbt, dev->nodename,
 880                             "ring-ref", "%u", info->ring_ref);
 881         if (err) {
 882                 message = "writing ring-ref";
 883                 goto abort_transaction;
 884         }
 885         err = xenbus_printf(xbt, dev->nodename,
 886                             "event-channel", "%u", info->evtchn);
 887         if (err) {
 888                 message = "writing event-channel";
 889                 goto abort_transaction;
 890         }
 891         err = xenbus_printf(xbt, dev->nodename, "protocol", "%s",
 892                             XEN_IO_PROTO_ABI_NATIVE);
 893         if (err) {
 894                 message = "writing protocol";
 895                 goto abort_transaction;
 896         }
 897
 898         err = xenbus_transaction_end(xbt, 0);
 899         if (err) {
 900                 if (err == -EAGAIN)
 901                         goto again;
 902                 xenbus_dev_fatal(dev, err, "completing transaction");
 903                 goto destroy_blkring;
 904         }
 905
 906         xenbus_switch_state(dev, XenbusStateInitialised);
 907
 908         return 0;
 909
 910  abort_transaction:
 911         xenbus_transaction_end(xbt, 1);
 912         if (message)
 913                 xenbus_dev_fatal(dev, err, "%s", message);
 914  destroy_blkring:
 915         blkif_free(info, 0);
 916  out:
 917         return err;
 918 }
 919
 920 /**
 921  * Entry point to this code when a new device is created.  Allocate the basic
 922  * structures and the ring buffer for communication with the backend, and
 923  * inform the backend of the appropriate details for those.  Switch to
 924  * Initialised state.
 925  */
 926 static int blkfront_probe(struct xenbus_device *dev,
 927                           const struct xenbus_device_id *id)
 928 {
 929         int err, vdevice, i;
 930         struct blkfront_info *info;
 931
 932         /* FIXME: Use dynamic device id if this is not set. */
 933         err = xenbus_scanf(XBT_NIL, dev->nodename,
 934                            "virtual-device", "%i", &vdevice);
 935         if (err != 1) {
 936                 /* go looking in the extended area instead */
 937                 err = xenbus_scanf(XBT_NIL, dev->nodename, "virtual-device-ext",
 938                                    "%i", &vdevice);
 939                 if (err != 1) {
 940                         xenbus_dev_fatal(dev, err, "reading virtual-device");
 941                         return err;
 942                 }
 943         }
 944
 945         if (xen_hvm_domain()) {
 946                 char *type;
 947                 int len;
 948                 /* no unplug has been done: do not hook devices != xen vbds */
 949                 if (xen_platform_pci_unplug & XEN_UNPLUG_UNNECESSARY) {
 950                         int major;
 951
 952                         if (!VDEV_IS_EXTENDED(vdevice))
 953                                 major = BLKIF_MAJOR(vdevice);
 954                         else
 955                                 major = XENVBD_MAJOR;
 956
 957                         if (major != XENVBD_MAJOR) {
 958                                 printk(KERN_INFO
 959                                                 "%s: HVM does not support vbd %d as xen block device\n",
 960                                                 __FUNCTION__, vdevice);
 961                                 return -ENODEV;
 962                         }
 963                 }
 964                 /* do not create a PV cdrom device if we are an HVM guest */
 965                 type = xenbus_read(XBT_NIL, dev->nodename, "device-type", &len);
 966                 if (IS_ERR(type))
 967                         return -ENODEV;
 968                 if (strncmp(type, "cdrom", 5) == 0) {
 969                         kfree(type);
 970                         return -ENODEV;
 971                 }
 972                 kfree(type);
 973         }
 974         info = kzalloc(sizeof(*info), GFP_KERNEL);
 975         if (!info) {
 976                 xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure");
 977                 return -ENOMEM;
 978         }
 979
 980         mutex_init(&info->mutex);
 981         info->xbdev = dev;
 982         info->vdevice = vdevice;
 983         info->connected = BLKIF_STATE_DISCONNECTED;
 984         INIT_WORK(&info->work, blkif_restart_queue);
 985
 986         for (i = 0; i < BLK_RING_SIZE; i++)
 987                 info->shadow[i].req.id = i+1;
 988         info->shadow[BLK_RING_SIZE-1].req.id = 0x0fffffff;
 989
 990         /* Front end dir is a number, which is used as the id. */
 991         info->handle = simple_strtoul(strrchr(dev->nodename, '/')+1, NULL, 0);
 992         dev_set_drvdata(&dev->dev, info);
 993
 994         err = talk_to_blkback(dev, info);
 995         if (err) {
 996                 kfree(info);
 997                 dev_set_drvdata(&dev->dev, NULL);
 998                 return err;
 999         }
1000
1001         return 0;
1002 }
1003
1004
1005 static int blkif_recover(struct blkfront_info *info)
1006 {
1007         int i;
1008         struct blkif_request *req;
1009         struct blk_shadow *copy;
1010         int j;
1011
1012         /* Stage 1: Make a safe copy of the shadow state. */
1013         copy = kmalloc(sizeof(info->shadow),
1014                        GFP_NOIO | __GFP_REPEAT | __GFP_HIGH);
1015         if (!copy)
1016                 return -ENOMEM;
1017         memcpy(copy, info->shadow, sizeof(info->shadow));
1018
1019         /* Stage 2: Set up free list. */
1020         memset(&info->shadow, 0, sizeof(info->shadow));
1021         for (i = 0; i < BLK_RING_SIZE; i++)
1022                 info->shadow[i].req.id = i+1;
1023         info->shadow_free = info->ring.req_prod_pvt;
1024         info->shadow[BLK_RING_SIZE-1].req.id = 0x0fffffff;
1025
1026         /* Stage 3: Find pending requests and requeue them. */
1027         for (i = 0; i < BLK_RING_SIZE; i++) {
1028                 /* Not in use? */
1029                 if (!copy[i].request)
1030                         continue;
1031
1032                 /* Grab a request slot and copy shadow state into it. */
1033                 req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
1034                 *req = copy[i].req;
1035
1036                 /* We get a new request id, and must reset the shadow state. */
1037                 req->id = get_id_from_freelist(info);
1038                 memcpy(&info->shadow[req->id], &copy[i], sizeof(copy[i]));
1039
1040                 /* Rewrite any grant references invalidated by susp/resume. */
1041                 for (j = 0; j < req->nr_segments; j++)
1042                         gnttab_grant_foreign_access_ref(
1043                                 req->u.rw.seg[j].gref,
1044                                 info->xbdev->otherend_id,
1045                                 pfn_to_mfn(info->shadow[req->id].frame[j]),
1046                                 rq_data_dir(info->shadow[req->id].request));
1047                 info->shadow[req->id].req = *req;
1048
1049                 info->ring.req_prod_pvt++;
1050         }
1051
1052         kfree(copy);
1053
1054         xenbus_switch_state(info->xbdev, XenbusStateConnected);
1055
1056         spin_lock_irq(&blkif_io_lock);
1057
1058         /* Now safe for us to use the shared ring */
1059         info->connected = BLKIF_STATE_CONNECTED;
1060
1061         /* Send off requeued requests */
1062         flush_requests(info);
1063
1064         /* Kick any other new requests queued since we resumed */
1065         kick_pending_request_queues(info);
1066
1067         spin_unlock_irq(&blkif_io_lock);
1068
1069         return 0;
1070 }
1071
1072 /**
1073  * We are reconnecting to the backend, due to a suspend/resume, or a backend
1074  * driver restart.  We tear down our blkif structure and recreate it, but
1075  * leave the device-layer structures intact so that this is transparent to the
1076  * rest of the kernel.
1077  */
1078 static int blkfront_resume(struct xenbus_device *dev)
1079 {
1080         struct blkfront_info *info = dev_get_drvdata(&dev->dev);
1081         int err;
1082
1083         dev_dbg(&dev->dev, "blkfront_resume: %s\n", dev->nodename);
1084
1085         blkif_free(info, info->connected == BLKIF_STATE_CONNECTED);
1086
1087         err = talk_to_blkback(dev, info);
1088         if (info->connected == BLKIF_STATE_SUSPENDED && !err)
1089                 err = blkif_recover(info);
1090
1091         return err;
1092 }
1093
1094 static void
1095 blkfront_closing(struct blkfront_info *info)
1096 {
1097         struct xenbus_device *xbdev = info->xbdev;
1098         struct block_device *bdev = NULL;
1099
1100         mutex_lock(&info->mutex);
1101
1102         if (xbdev->state == XenbusStateClosing) {
1103                 mutex_unlock(&info->mutex);
1104                 return;
1105         }
1106
1107         if (info->gd)
1108                 bdev = bdget_disk(info->gd, 0);
1109
1110         mutex_unlock(&info->mutex);
1111
1112         if (!bdev) {
1113                 xenbus_frontend_closed(xbdev);
1114                 return;
1115         }
1116
1117         mutex_lock(&bdev->bd_mutex);
1118
1119         if (bdev->bd_openers) {
1120                 xenbus_dev_error(xbdev, -EBUSY,
1121                                  "Device in use; refusing to close");
1122                 xenbus_switch_state(xbdev, XenbusStateClosing);
1123         } else {
1124                 xlvbd_release_gendisk(info);
1125                 xenbus_frontend_closed(xbdev);
1126         }
1127
1128         mutex_unlock(&bdev->bd_mutex);
1129         bdput(bdev);
1130 }
1131
1132 static void blkfront_setup_discard(struct blkfront_info *info)
1133 {
1134         int err;
1135         char *type;
1136         unsigned int discard_granularity;
1137         unsigned int discard_alignment;
1138
1139         type = xenbus_read(XBT_NIL, info->xbdev->otherend, "type", NULL);
1140         if (IS_ERR(type))
1141                 return;
1142
1143         if (strncmp(type, "phy", 3) == 0) {
1144                 err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
1145                         "discard-granularity", "%u", &discard_granularity,
1146                         "discard-alignment", "%u", &discard_alignment,
1147                         NULL);
1148                 if (!err) {
1149                         info->feature_discard = 1;
1150                         info->discard_granularity = discard_granularity;
1151                         info->discard_alignment = discard_alignment;
1152                 }
1153         } else if (strncmp(type, "file", 4) == 0)
1154                 info->feature_discard = 1;
1155
1156         kfree(type);
1157 }
1158
1159 /*
1160  * Invoked when the backend is finally 'ready' (and has told produced
1161  * the details about the physical device - #sectors, size, etc).
1162  */
1163 static void blkfront_connect(struct blkfront_info *info)
1164 {
1165         unsigned long long sectors;
1166         unsigned long sector_size;
1167         unsigned int binfo;
1168         int err;
1169         int barrier, flush, discard;
1170
1171         switch (info->connected) {
1172         case BLKIF_STATE_CONNECTED:
1173                 /*
1174                  * Potentially, the back-end may be signalling
1175                  * a capacity change; update the capacity.
1176                  */
1177                 err = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
1178                                    "sectors", "%Lu", &sectors);
1179                 if (XENBUS_EXIST_ERR(err))
1180                         return;
1181                 printk(KERN_INFO "Setting capacity to %Lu\n",
1182                        sectors);
1183                 set_capacity(info->gd, sectors);
1184                 revalidate_disk(info->gd);
1185
1186                 /* fall through */
1187         case BLKIF_STATE_SUSPENDED:
1188                 return;
1189
1190         default:
1191                 break;
1192         }
1193
1194         dev_dbg(&info->xbdev->dev, "%s:%s.\n",
1195                 __func__, info->xbdev->otherend);
1196
1197         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
1198                             "sectors", "%llu", &sectors,
1199                             "info", "%u", &binfo,
1200                             "sector-size", "%lu", &sector_size,
1201                             NULL);
1202         if (err) {
1203                 xenbus_dev_fatal(info->xbdev, err,
1204                                  "reading backend fields at %s",
1205                                  info->xbdev->otherend);
1206                 return;
1207         }
1208
1209         info->feature_flush = 0;
1210         info->flush_op = 0;
1211
1212         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
1213                             "feature-barrier", "%d", &barrier,
1214                             NULL);
1215
1216         /*
1217          * If there's no "feature-barrier" defined, then it means
1218          * we're dealing with a very old backend which writes
1219          * synchronously; nothing to do.
1220          *
1221          * If there are barriers, then we use flush.
1222          */
1223         if (!err && barrier) {
1224                 info->feature_flush = REQ_FLUSH | REQ_FUA;
1225                 info->flush_op = BLKIF_OP_WRITE_BARRIER;
1226         }
1227         /*
1228          * And if there is "feature-flush-cache" use that above
1229          * barriers.
1230          */
1231         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
1232                             "feature-flush-cache", "%d", &flush,
1233                             NULL);
1234
1235         if (!err && flush) {
1236                 info->feature_flush = REQ_FLUSH;
1237                 info->flush_op = BLKIF_OP_FLUSH_DISKCACHE;
1238         }
1239
1240         err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
1241                             "feature-discard", "%d", &discard,
1242                             NULL);
1243
1244         if (!err && discard)
1245                 blkfront_setup_discard(info);
1246
1247         err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
1248         if (err) {
1249                 xenbus_dev_fatal(info->xbdev, err, "xlvbd_add at %s",
1250                                  info->xbdev->otherend);
1251                 return;
1252         }
1253
1254         xenbus_switch_state(info->xbdev, XenbusStateConnected);
1255
1256         /* Kick pending requests. */
1257         spin_lock_irq(&blkif_io_lock);
1258         info->connected = BLKIF_STATE_CONNECTED;
1259         kick_pending_request_queues(info);
1260         spin_unlock_irq(&blkif_io_lock);
1261
1262         add_disk(info->gd);
1263
1264         info->is_ready = 1;
1265 }
1266
1267 /**
1268  * Callback received when the backend's state changes.
1269  */
1270 static void blkback_changed(struct xenbus_device *dev,
1271                             enum xenbus_state backend_state)
1272 {
1273         struct blkfront_info *info = dev_get_drvdata(&dev->dev);
1274
1275         dev_dbg(&dev->dev, "blkfront:blkback_changed to state %d.\n", backend_state);
1276
1277         switch (backend_state) {
1278         case XenbusStateInitialising:
1279         case XenbusStateInitWait:
1280         case XenbusStateInitialised:
1281         case XenbusStateReconfiguring:
1282         case XenbusStateReconfigured:
1283         case XenbusStateUnknown:
1284         case XenbusStateClosed:
1285                 break;
1286
1287         case XenbusStateConnected:
1288                 blkfront_connect(info);
1289                 break;
1290
1291         case XenbusStateClosing:
1292                 blkfront_closing(info);
1293                 break;
1294         }
1295 }
1296
1297 static int blkfront_remove(struct xenbus_device *xbdev)
1298 {
1299         struct blkfront_info *info = dev_get_drvdata(&xbdev->dev);
1300         struct block_device *bdev = NULL;
1301         struct gendisk *disk;
1302
1303         dev_dbg(&xbdev->dev, "%s removed", xbdev->nodename);
1304
1305         blkif_free(info, 0);
1306
1307         mutex_lock(&info->mutex);
1308
1309         disk = info->gd;
1310         if (disk)
1311                 bdev = bdget_disk(disk, 0);
1312
1313         info->xbdev = NULL;
1314         mutex_unlock(&info->mutex);
1315
1316         if (!bdev) {
1317                 kfree(info);
1318                 return 0;
1319         }
1320
1321         /*
1322          * The xbdev was removed before we reached the Closed
1323          * state. See if it's safe to remove the disk. If the bdev
1324          * isn't closed yet, we let release take care of it.
1325          */
1326
1327         mutex_lock(&bdev->bd_mutex);
1328         info = disk->private_data;
1329
1330         dev_warn(disk_to_dev(disk),
1331                  "%s was hot-unplugged, %d stale handles\n",
1332                  xbdev->nodename, bdev->bd_openers);
1333
1334         if (info && !bdev->bd_openers) {
1335                 xlvbd_release_gendisk(info);
1336                 disk->private_data = NULL;
1337                 kfree(info);
1338         }
1339
1340         mutex_unlock(&bdev->bd_mutex);
1341         bdput(bdev);
1342
1343         return 0;
1344 }
1345
1346 static int blkfront_is_ready(struct xenbus_device *dev)
1347 {
1348         struct blkfront_info *info = dev_get_drvdata(&dev->dev);
1349
1350         return info->is_ready && info->xbdev;
1351 }
1352
1353 static int blkif_open(struct block_device *bdev, fmode_t mode)
1354 {
1355         struct gendisk *disk = bdev->bd_disk;
1356         struct blkfront_info *info;
1357         int err = 0;
1358
1359         mutex_lock(&blkfront_mutex);
1360
1361         info = disk->private_data;
1362         if (!info) {
1363                 /* xbdev gone */
1364                 err = -ERESTARTSYS;
1365                 goto out;
1366         }
1367
1368         mutex_lock(&info->mutex);
1369
1370         if (!info->gd)
1371                 /* xbdev is closed */
1372                 err = -ERESTARTSYS;
1373
1374         mutex_unlock(&info->mutex);
1375
1376 out:
1377         mutex_unlock(&blkfront_mutex);
1378         return err;
1379 }
1380
1381 static int blkif_release(struct gendisk *disk, fmode_t mode)
1382 {
1383         struct blkfront_info *info = disk->private_data;
1384         struct block_device *bdev;
1385         struct xenbus_device *xbdev;
1386
1387         mutex_lock(&blkfront_mutex);
1388
1389         bdev = bdget_disk(disk, 0);
1390         bdput(bdev);
1391
1392         if (bdev->bd_openers)
1393                 goto out;
1394
1395         /*
1396          * Check if we have been instructed to close. We will have
1397          * deferred this request, because the bdev was still open.
1398          */
1399
1400         mutex_lock(&info->mutex);
1401         xbdev = info->xbdev;
1402
1403         if (xbdev && xbdev->state == XenbusStateClosing) {
1404                 /* pending switch to state closed */
1405                 dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n");
1406                 xlvbd_release_gendisk(info);
1407                 xenbus_frontend_closed(info->xbdev);
1408         }
1409
1410         mutex_unlock(&info->mutex);
1411
1412         if (!xbdev) {
1413                 /* sudden device removal */
1414                 dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n");
1415                 xlvbd_release_gendisk(info);
1416                 disk->private_data = NULL;
1417                 kfree(info);
1418         }
1419
1420 out:
1421         mutex_unlock(&blkfront_mutex);
1422         return 0;
1423 }
1424
1425 static const struct block_device_operations xlvbd_block_fops =
1426 {
1427         .owner = THIS_MODULE,
1428         .open = blkif_open,
1429         .release = blkif_release,
1430         .getgeo = blkif_getgeo,
1431         .ioctl = blkif_ioctl,
1432 };
1433
1434
1435 static const struct xenbus_device_id blkfront_ids[] = {
1436         { "vbd" },
1437         { "" }
1438 };
1439
1440 static struct xenbus_driver blkfront = {
1441         .name = "vbd",
1442         .owner = THIS_MODULE,
1443         .ids = blkfront_ids,
1444         .probe = blkfront_probe,
1445         .remove = blkfront_remove,
1446         .resume = blkfront_resume,
1447         .otherend_changed = blkback_changed,
1448         .is_ready = blkfront_is_ready,
1449 };
1450
1451 static int __init xlblk_init(void)
1452 {
1453         int ret;
1454
1455         if (!xen_domain())
1456                 return -ENODEV;
1457
1458         if (register_blkdev(XENVBD_MAJOR, DEV_NAME)) {
1459                 printk(KERN_WARNING "xen_blk: can't get major %d with name %s\n",
1460                        XENVBD_MAJOR, DEV_NAME);
1461                 return -ENODEV;
1462         }
1463
1464         ret = xenbus_register_frontend(&blkfront);
1465         if (ret) {
1466                 unregister_blkdev(XENVBD_MAJOR, DEV_NAME);
1467                 return ret;
1468         }
1469
1470         return 0;
1471 }
1472 module_init(xlblk_init);
1473
1474
1475 static void __exit xlblk_exit(void)
1476 {
1477         return xenbus_unregister_driver(&blkfront);
1478 }
1479 module_exit(xlblk_exit);
1480
1481 MODULE_DESCRIPTION("Xen virtual block device frontend");
1482 MODULE_LICENSE("GPL");
1483 MODULE_ALIAS_BLOCKDEV_MAJOR(XENVBD_MAJOR);
1484 MODULE_ALIAS("xen:vbd");
1485 MODULE_ALIAS("xenblk");