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