fs/btrfs/volumes.c

   1 /*
   2  * Copyright (C) 2007 Oracle.  All rights reserved.
   3  *
   4  * This program is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU General Public
   6  * License v2 as published by the Free Software Foundation.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  11  * General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public
  14  * License along with this program; if not, write to the
  15  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  16  * Boston, MA 021110-1307, USA.
  17  */
  18 #include <linux/sched.h>
  19 #include <linux/bio.h>
  20 #include <linux/slab.h>
  21 #include <linux/buffer_head.h>
  22 #include <linux/blkdev.h>
  23 #include <linux/random.h>
  24 #include <linux/iocontext.h>
  25 #include <linux/capability.h>
  26 #include <asm/div64.h>
  27 #include "compat.h"
  28 #include "ctree.h"
  29 #include "extent_map.h"
  30 #include "disk-io.h"
  31 #include "transaction.h"
  32 #include "print-tree.h"
  33 #include "volumes.h"
  34 #include "async-thread.h"
  35
  36 static int init_first_rw_device(struct btrfs_trans_handle *trans,
  37                                 struct btrfs_root *root,
  38                                 struct btrfs_device *device);
  39 static int btrfs_relocate_sys_chunks(struct btrfs_root *root);
  40
  41 static DEFINE_MUTEX(uuid_mutex);
  42 static LIST_HEAD(fs_uuids);
  43
  44 static void lock_chunks(struct btrfs_root *root)
  45 {
  46         mutex_lock(&root->fs_info->chunk_mutex);
  47 }
  48
  49 static void unlock_chunks(struct btrfs_root *root)
  50 {
  51         mutex_unlock(&root->fs_info->chunk_mutex);
  52 }
  53
  54 static void free_fs_devices(struct btrfs_fs_devices *fs_devices)
  55 {
  56         struct btrfs_device *device;
  57         WARN_ON(fs_devices->opened);
  58         while (!list_empty(&fs_devices->devices)) {
  59                 device = list_entry(fs_devices->devices.next,
  60                                     struct btrfs_device, dev_list);
  61                 list_del(&device->dev_list);
  62                 kfree(device->name);
  63                 kfree(device);
  64         }
  65         kfree(fs_devices);
  66 }
  67
  68 int btrfs_cleanup_fs_uuids(void)
  69 {
  70         struct btrfs_fs_devices *fs_devices;
  71
  72         while (!list_empty(&fs_uuids)) {
  73                 fs_devices = list_entry(fs_uuids.next,
  74                                         struct btrfs_fs_devices, list);
  75                 list_del(&fs_devices->list);
  76                 free_fs_devices(fs_devices);
  77         }
  78         return 0;
  79 }
  80
  81 static noinline struct btrfs_device *__find_device(struct list_head *head,
  82                                                    u64 devid, u8 *uuid)
  83 {
  84         struct btrfs_device *dev;
  85
  86         list_for_each_entry(dev, head, dev_list) {
  87                 if (dev->devid == devid &&
  88                     (!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
  89                         return dev;
  90                 }
  91         }
  92         return NULL;
  93 }
  94
  95 static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid)
  96 {
  97         struct btrfs_fs_devices *fs_devices;
  98
  99         list_for_each_entry(fs_devices, &fs_uuids, list) {
 100                 if (memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE) == 0)
 101                         return fs_devices;
 102         }
 103         return NULL;
 104 }
 105
 106 static void requeue_list(struct btrfs_pending_bios *pending_bios,
 107                         struct bio *head, struct bio *tail)
 108 {
 109
 110         struct bio *old_head;
 111
 112         old_head = pending_bios->head;
 113         pending_bios->head = head;
 114         if (pending_bios->tail)
 115                 tail->bi_next = old_head;
 116         else
 117                 pending_bios->tail = tail;
 118 }
 119
 120 /*
 121  * we try to collect pending bios for a device so we don't get a large
 122  * number of procs sending bios down to the same device.  This greatly
 123  * improves the schedulers ability to collect and merge the bios.
 124  *
 125  * But, it also turns into a long list of bios to process and that is sure
 126  * to eventually make the worker thread block.  The solution here is to
 127  * make some progress and then put this work struct back at the end of
 128  * the list if the block device is congested.  This way, multiple devices
 129  * can make progress from a single worker thread.
 130  */
 131 static noinline int run_scheduled_bios(struct btrfs_device *device)
 132 {
 133         struct bio *pending;
 134         struct backing_dev_info *bdi;
 135         struct btrfs_fs_info *fs_info;
 136         struct btrfs_pending_bios *pending_bios;
 137         struct bio *tail;
 138         struct bio *cur;
 139         int again = 0;
 140         unsigned long num_run;
 141         unsigned long batch_run = 0;
 142         unsigned long limit;
 143         unsigned long last_waited = 0;
 144         int force_reg = 0;
 145         int sync_pending = 0;
 146         struct blk_plug plug;
 147
 148         /*
 149          * this function runs all the bios we've collected for
 150          * a particular device.  We don't want to wander off to
 151          * another device without first sending all of these down.
 152          * So, setup a plug here and finish it off before we return
 153          */
 154         blk_start_plug(&plug);
 155
 156         bdi = blk_get_backing_dev_info(device->bdev);
 157         fs_info = device->dev_root->fs_info;
 158         limit = btrfs_async_submit_limit(fs_info);
 159         limit = limit * 2 / 3;
 160
 161 loop:
 162         spin_lock(&device->io_lock);
 163
 164 loop_lock:
 165         num_run = 0;
 166
 167         /* take all the bios off the list at once and process them
 168          * later on (without the lock held).  But, remember the
 169          * tail and other pointers so the bios can be properly reinserted
 170          * into the list if we hit congestion
 171          */
 172         if (!force_reg && device->pending_sync_bios.head) {
 173                 pending_bios = &device->pending_sync_bios;
 174                 force_reg = 1;
 175         } else {
 176                 pending_bios = &device->pending_bios;
 177                 force_reg = 0;
 178         }
 179
 180         pending = pending_bios->head;
 181         tail = pending_bios->tail;
 182         WARN_ON(pending && !tail);
 183
 184         /*
 185          * if pending was null this time around, no bios need processing
 186          * at all and we can stop.  Otherwise it'll loop back up again
 187          * and do an additional check so no bios are missed.
 188          *
 189          * device->running_pending is used to synchronize with the
 190          * schedule_bio code.
 191          */
 192         if (device->pending_sync_bios.head == NULL &&
 193             device->pending_bios.head == NULL) {
 194                 again = 0;
 195                 device->running_pending = 0;
 196         } else {
 197                 again = 1;
 198                 device->running_pending = 1;
 199         }
 200
 201         pending_bios->head = NULL;
 202         pending_bios->tail = NULL;
 203
 204         spin_unlock(&device->io_lock);
 205
 206         while (pending) {
 207
 208                 rmb();
 209                 /* we want to work on both lists, but do more bios on the
 210                  * sync list than the regular list
 211                  */
 212                 if ((num_run > 32 &&
 213                     pending_bios != &device->pending_sync_bios &&
 214                     device->pending_sync_bios.head) ||
 215                    (num_run > 64 && pending_bios == &device->pending_sync_bios &&
 216                     device->pending_bios.head)) {
 217                         spin_lock(&device->io_lock);
 218                         requeue_list(pending_bios, pending, tail);
 219                         goto loop_lock;
 220                 }
 221
 222                 cur = pending;
 223                 pending = pending->bi_next;
 224                 cur->bi_next = NULL;
 225                 atomic_dec(&fs_info->nr_async_bios);
 226
 227                 if (atomic_read(&fs_info->nr_async_bios) < limit &&
 228                     waitqueue_active(&fs_info->async_submit_wait))
 229                         wake_up(&fs_info->async_submit_wait);
 230
 231                 BUG_ON(atomic_read(&cur->bi_cnt) == 0);
 232
 233                 /*
 234                  * if we're doing the sync list, record that our
 235                  * plug has some sync requests on it
 236                  *
 237                  * If we're doing the regular list and there are
 238                  * sync requests sitting around, unplug before
 239                  * we add more
 240                  */
 241                 if (pending_bios == &device->pending_sync_bios) {
 242                         sync_pending = 1;
 243                 } else if (sync_pending) {
 244                         blk_finish_plug(&plug);
 245                         blk_start_plug(&plug);
 246                         sync_pending = 0;
 247                 }
 248
 249                 submit_bio(cur->bi_rw, cur);
 250                 num_run++;
 251                 batch_run++;
 252                 if (need_resched())
 253                         cond_resched();
 254
 255                 /*
 256                  * we made progress, there is more work to do and the bdi
 257                  * is now congested.  Back off and let other work structs
 258                  * run instead
 259                  */
 260                 if (pending && bdi_write_congested(bdi) && batch_run > 8 &&
 261                     fs_info->fs_devices->open_devices > 1) {
 262                         struct io_context *ioc;
 263
 264                         ioc = current->io_context;
 265
 266                         /*
 267                          * the main goal here is that we don't want to
 268                          * block if we're going to be able to submit
 269                          * more requests without blocking.
 270                          *
 271                          * This code does two great things, it pokes into
 272                          * the elevator code from a filesystem _and_
 273                          * it makes assumptions about how batching works.
 274                          */
 275                         if (ioc && ioc->nr_batch_requests > 0 &&
 276                             time_before(jiffies, ioc->last_waited + HZ/50UL) &&
 277                             (last_waited == 0 ||
 278                              ioc->last_waited == last_waited)) {
 279                                 /*
 280                                  * we want to go through our batch of
 281                                  * requests and stop.  So, we copy out
 282                                  * the ioc->last_waited time and test
 283                                  * against it before looping
 284                                  */
 285                                 last_waited = ioc->last_waited;
 286                                 if (need_resched())
 287                                         cond_resched();
 288                                 continue;
 289                         }
 290                         spin_lock(&device->io_lock);
 291                         requeue_list(pending_bios, pending, tail);
 292                         device->running_pending = 1;
 293
 294                         spin_unlock(&device->io_lock);
 295                         btrfs_requeue_work(&device->work);
 296                         goto done;
 297                 }
 298         }
 299
 300         cond_resched();
 301         if (again)
 302                 goto loop;
 303
 304         spin_lock(&device->io_lock);
 305         if (device->pending_bios.head || device->pending_sync_bios.head)
 306                 goto loop_lock;
 307         spin_unlock(&device->io_lock);
 308
 309 done:
 310         blk_finish_plug(&plug);
 311         return 0;
 312 }
 313
 314 static void pending_bios_fn(struct btrfs_work *work)
 315 {
 316         struct btrfs_device *device;
 317
 318         device = container_of(work, struct btrfs_device, work);
 319         run_scheduled_bios(device);
 320 }
 321
 322 static noinline int device_list_add(const char *path,
 323                            struct btrfs_super_block *disk_super,
 324                            u64 devid, struct btrfs_fs_devices **fs_devices_ret)
 325 {
 326         struct btrfs_device *device;
 327         struct btrfs_fs_devices *fs_devices;
 328         u64 found_transid = btrfs_super_generation(disk_super);
 329         char *name;
 330
 331         fs_devices = find_fsid(disk_super->fsid);
 332         if (!fs_devices) {
 333                 fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
 334                 if (!fs_devices)
 335                         return -ENOMEM;
 336                 INIT_LIST_HEAD(&fs_devices->devices);
 337                 INIT_LIST_HEAD(&fs_devices->alloc_list);
 338                 list_add(&fs_devices->list, &fs_uuids);
 339                 memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
 340                 fs_devices->latest_devid = devid;
 341                 fs_devices->latest_trans = found_transid;
 342                 mutex_init(&fs_devices->device_list_mutex);
 343                 device = NULL;
 344         } else {
 345                 device = __find_device(&fs_devices->devices, devid,
 346                                        disk_super->dev_item.uuid);
 347         }
 348         if (!device) {
 349                 if (fs_devices->opened)
 350                         return -EBUSY;
 351
 352                 device = kzalloc(sizeof(*device), GFP_NOFS);
 353                 if (!device) {
 354                         /* we can safely leave the fs_devices entry around */
 355                         return -ENOMEM;
 356                 }
 357                 device->devid = devid;
 358                 device->work.func = pending_bios_fn;
 359                 memcpy(device->uuid, disk_super->dev_item.uuid,
 360                        BTRFS_UUID_SIZE);
 361                 spin_lock_init(&device->io_lock);
 362                 device->name = kstrdup(path, GFP_NOFS);
 363                 if (!device->name) {
 364                         kfree(device);
 365                         return -ENOMEM;
 366                 }
 367                 INIT_LIST_HEAD(&device->dev_alloc_list);
 368
 369                 mutex_lock(&fs_devices->device_list_mutex);
 370                 list_add_rcu(&device->dev_list, &fs_devices->devices);
 371                 mutex_unlock(&fs_devices->device_list_mutex);
 372
 373                 device->fs_devices = fs_devices;
 374                 fs_devices->num_devices++;
 375         } else if (!device->name || strcmp(device->name, path)) {
 376                 name = kstrdup(path, GFP_NOFS);
 377                 if (!name)
 378                         return -ENOMEM;
 379                 kfree(device->name);
 380                 device->name = name;
 381                 if (device->missing) {
 382                         fs_devices->missing_devices--;
 383                         device->missing = 0;
 384                 }
 385         }
 386
 387         if (found_transid > fs_devices->latest_trans) {
 388                 fs_devices->latest_devid = devid;
 389                 fs_devices->latest_trans = found_transid;
 390         }
 391         *fs_devices_ret = fs_devices;
 392         return 0;
 393 }
 394
 395 static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
 396 {
 397         struct btrfs_fs_devices *fs_devices;
 398         struct btrfs_device *device;
 399         struct btrfs_device *orig_dev;
 400
 401         fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
 402         if (!fs_devices)
 403                 return ERR_PTR(-ENOMEM);
 404
 405         INIT_LIST_HEAD(&fs_devices->devices);
 406         INIT_LIST_HEAD(&fs_devices->alloc_list);
 407         INIT_LIST_HEAD(&fs_devices->list);
 408         mutex_init(&fs_devices->device_list_mutex);
 409         fs_devices->latest_devid = orig->latest_devid;
 410         fs_devices->latest_trans = orig->latest_trans;
 411         memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid));
 412
 413         /* We have held the volume lock, it is safe to get the devices. */
 414         list_for_each_entry(orig_dev, &orig->devices, dev_list) {
 415                 device = kzalloc(sizeof(*device), GFP_NOFS);
 416                 if (!device)
 417                         goto error;
 418
 419                 device->name = kstrdup(orig_dev->name, GFP_NOFS);
 420                 if (!device->name) {
 421                         kfree(device);
 422                         goto error;
 423                 }
 424
 425                 device->devid = orig_dev->devid;
 426                 device->work.func = pending_bios_fn;
 427                 memcpy(device->uuid, orig_dev->uuid, sizeof(device->uuid));
 428                 spin_lock_init(&device->io_lock);
 429                 INIT_LIST_HEAD(&device->dev_list);
 430                 INIT_LIST_HEAD(&device->dev_alloc_list);
 431
 432                 list_add(&device->dev_list, &fs_devices->devices);
 433                 device->fs_devices = fs_devices;
 434                 fs_devices->num_devices++;
 435         }
 436         return fs_devices;
 437 error:
 438         free_fs_devices(fs_devices);
 439         return ERR_PTR(-ENOMEM);
 440 }
 441
 442 int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
 443 {
 444         struct btrfs_device *device, *next;
 445
 446         mutex_lock(&uuid_mutex);
 447 again:
 448         /* This is the initialized path, it is safe to release the devices. */
 449         list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
 450                 if (device->in_fs_metadata)
 451                         continue;
 452
 453                 if (device->bdev) {
 454                         blkdev_put(device->bdev, device->mode);
 455                         device->bdev = NULL;
 456                         fs_devices->open_devices--;
 457                 }
 458                 if (device->writeable) {
 459                         list_del_init(&device->dev_alloc_list);
 460                         device->writeable = 0;
 461                         fs_devices->rw_devices--;
 462                 }
 463                 list_del_init(&device->dev_list);
 464                 fs_devices->num_devices--;
 465                 kfree(device->name);
 466                 kfree(device);
 467         }
 468
 469         if (fs_devices->seed) {
 470                 fs_devices = fs_devices->seed;
 471                 goto again;
 472         }
 473
 474         mutex_unlock(&uuid_mutex);
 475         return 0;
 476 }
 477
 478 static void __free_device(struct work_struct *work)
 479 {
 480         struct btrfs_device *device;
 481
 482         device = container_of(work, struct btrfs_device, rcu_work);
 483
 484         if (device->bdev)
 485                 blkdev_put(device->bdev, device->mode);
 486
 487         kfree(device->name);
 488         kfree(device);
 489 }
 490
 491 static void free_device(struct rcu_head *head)
 492 {
 493         struct btrfs_device *device;
 494
 495         device = container_of(head, struct btrfs_device, rcu);
 496
 497         INIT_WORK(&device->rcu_work, __free_device);
 498         schedule_work(&device->rcu_work);
 499 }
 500
 501 static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
 502 {
 503         struct btrfs_device *device;
 504
 505         if (--fs_devices->opened > 0)
 506                 return 0;
 507
 508         mutex_lock(&fs_devices->device_list_mutex);
 509         list_for_each_entry(device, &fs_devices->devices, dev_list) {
 510                 struct btrfs_device *new_device;
 511
 512                 if (device->bdev)
 513                         fs_devices->open_devices--;
 514
 515                 if (device->writeable) {
 516                         list_del_init(&device->dev_alloc_list);
 517                         fs_devices->rw_devices--;
 518                 }
 519
 520                 if (device->can_discard)
 521                         fs_devices->num_can_discard--;
 522
 523                 new_device = kmalloc(sizeof(*new_device), GFP_NOFS);
 524                 BUG_ON(!new_device);
 525                 memcpy(new_device, device, sizeof(*new_device));
 526                 new_device->name = kstrdup(device->name, GFP_NOFS);
 527                 BUG_ON(device->name && !new_device->name);
 528                 new_device->bdev = NULL;
 529                 new_device->writeable = 0;
 530                 new_device->in_fs_metadata = 0;
 531                 new_device->can_discard = 0;
 532                 list_replace_rcu(&device->dev_list, &new_device->dev_list);
 533
 534                 call_rcu(&device->rcu, free_device);
 535         }
 536         mutex_unlock(&fs_devices->device_list_mutex);
 537
 538         WARN_ON(fs_devices->open_devices);
 539         WARN_ON(fs_devices->rw_devices);
 540         fs_devices->opened = 0;
 541         fs_devices->seeding = 0;
 542
 543         return 0;
 544 }
 545
 546 int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
 547 {
 548         struct btrfs_fs_devices *seed_devices = NULL;
 549         int ret;
 550
 551         mutex_lock(&uuid_mutex);
 552         ret = __btrfs_close_devices(fs_devices);
 553         if (!fs_devices->opened) {
 554                 seed_devices = fs_devices->seed;
 555                 fs_devices->seed = NULL;
 556         }
 557         mutex_unlock(&uuid_mutex);
 558
 559         while (seed_devices) {
 560                 fs_devices = seed_devices;
 561                 seed_devices = fs_devices->seed;
 562                 __btrfs_close_devices(fs_devices);
 563                 free_fs_devices(fs_devices);
 564         }
 565         return ret;
 566 }
 567
 568 static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 569                                 fmode_t flags, void *holder)
 570 {
 571         struct request_queue *q;
 572         struct block_device *bdev;
 573         struct list_head *head = &fs_devices->devices;
 574         struct btrfs_device *device;
 575         struct block_device *latest_bdev = NULL;
 576         struct buffer_head *bh;
 577         struct btrfs_super_block *disk_super;
 578         u64 latest_devid = 0;
 579         u64 latest_transid = 0;
 580         u64 devid;
 581         int seeding = 1;
 582         int ret = 0;
 583
 584         flags |= FMODE_EXCL;
 585
 586         list_for_each_entry(device, head, dev_list) {
 587                 if (device->bdev)
 588                         continue;
 589                 if (!device->name)
 590                         continue;
 591
 592                 bdev = blkdev_get_by_path(device->name, flags, holder);
 593                 if (IS_ERR(bdev)) {
 594                         printk(KERN_INFO "open %s failed\n", device->name);
 595                         goto error;
 596                 }
 597                 set_blocksize(bdev, 4096);
 598
 599                 bh = btrfs_read_dev_super(bdev);
 600                 if (!bh) {
 601                         ret = -EINVAL;
 602                         goto error_close;
 603                 }
 604
 605                 disk_super = (struct btrfs_super_block *)bh->b_data;
 606                 devid = btrfs_stack_device_id(&disk_super->dev_item);
 607                 if (devid != device->devid)
 608                         goto error_brelse;
 609
 610                 if (memcmp(device->uuid, disk_super->dev_item.uuid,
 611                            BTRFS_UUID_SIZE))
 612                         goto error_brelse;
 613
 614                 device->generation = btrfs_super_generation(disk_super);
 615                 if (!latest_transid || device->generation > latest_transid) {
 616                         latest_devid = devid;
 617                         latest_transid = device->generation;
 618                         latest_bdev = bdev;
 619                 }
 620
 621                 if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING) {
 622                         device->writeable = 0;
 623                 } else {
 624                         device->writeable = !bdev_read_only(bdev);
 625                         seeding = 0;
 626                 }
 627
 628                 q = bdev_get_queue(bdev);
 629                 if (blk_queue_discard(q)) {
 630                         device->can_discard = 1;
 631                         fs_devices->num_can_discard++;
 632                 }
 633
 634                 device->bdev = bdev;
 635                 device->in_fs_metadata = 0;
 636                 device->mode = flags;
 637
 638                 if (!blk_queue_nonrot(bdev_get_queue(bdev)))
 639                         fs_devices->rotating = 1;
 640
 641                 fs_devices->open_devices++;
 642                 if (device->writeable) {
 643                         fs_devices->rw_devices++;
 644                         list_add(&device->dev_alloc_list,
 645                                  &fs_devices->alloc_list);
 646                 }
 647                 brelse(bh);
 648                 continue;
 649
 650 error_brelse:
 651                 brelse(bh);
 652 error_close:
 653                 blkdev_put(bdev, flags);
 654 error:
 655                 continue;
 656         }
 657         if (fs_devices->open_devices == 0) {
 658                 ret = -EIO;
 659                 goto out;
 660         }
 661         fs_devices->seeding = seeding;
 662         fs_devices->opened = 1;
 663         fs_devices->latest_bdev = latest_bdev;
 664         fs_devices->latest_devid = latest_devid;
 665         fs_devices->latest_trans = latest_transid;
 666         fs_devices->total_rw_bytes = 0;
 667 out:
 668         return ret;
 669 }
 670
 671 int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
 672                        fmode_t flags, void *holder)
 673 {
 674         int ret;
 675
 676         mutex_lock(&uuid_mutex);
 677         if (fs_devices->opened) {
 678                 fs_devices->opened++;
 679                 ret = 0;
 680         } else {
 681                 ret = __btrfs_open_devices(fs_devices, flags, holder);
 682         }
 683         mutex_unlock(&uuid_mutex);
 684         return ret;
 685 }
 686
 687 int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
 688                           struct btrfs_fs_devices **fs_devices_ret)
 689 {
 690         struct btrfs_super_block *disk_super;
 691         struct block_device *bdev;
 692         struct buffer_head *bh;
 693         int ret;
 694         u64 devid;
 695         u64 transid;
 696
 697         mutex_lock(&uuid_mutex);
 698
 699         flags |= FMODE_EXCL;
 700         bdev = blkdev_get_by_path(path, flags, holder);
 701
 702         if (IS_ERR(bdev)) {
 703                 ret = PTR_ERR(bdev);
 704                 goto error;
 705         }
 706
 707         ret = set_blocksize(bdev, 4096);
 708         if (ret)
 709                 goto error_close;
 710         bh = btrfs_read_dev_super(bdev);
 711         if (!bh) {
 712                 ret = -EINVAL;
 713                 goto error_close;
 714         }
 715         disk_super = (struct btrfs_super_block *)bh->b_data;
 716         devid = btrfs_stack_device_id(&disk_super->dev_item);
 717         transid = btrfs_super_generation(disk_super);
 718         if (disk_super->label[0])
 719                 printk(KERN_INFO "device label %s ", disk_super->label);
 720         else
 721                 printk(KERN_INFO "device fsid %pU ", disk_super->fsid);
 722         printk(KERN_CONT "devid %llu transid %llu %s\n",
 723                (unsigned long long)devid, (unsigned long long)transid, path);
 724         ret = device_list_add(path, disk_super, devid, fs_devices_ret);
 725
 726         brelse(bh);
 727 error_close:
 728         blkdev_put(bdev, flags);
 729 error:
 730         mutex_unlock(&uuid_mutex);
 731         return ret;
 732 }
 733
 734 /* helper to account the used device space in the range */
 735 int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
 736                                    u64 end, u64 *length)
 737 {
 738         struct btrfs_key key;
 739         struct btrfs_root *root = device->dev_root;
 740         struct btrfs_dev_extent *dev_extent;
 741         struct btrfs_path *path;
 742         u64 extent_end;
 743         int ret;
 744         int slot;
 745         struct extent_buffer *l;
 746
 747         *length = 0;
 748
 749         if (start >= device->total_bytes)
 750                 return 0;
 751
 752         path = btrfs_alloc_path();
 753         if (!path)
 754                 return -ENOMEM;
 755         path->reada = 2;
 756
 757         key.objectid = device->devid;
 758         key.offset = start;
 759         key.type = BTRFS_DEV_EXTENT_KEY;
 760
 761         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 762         if (ret < 0)
 763                 goto out;
 764         if (ret > 0) {
 765                 ret = btrfs_previous_item(root, path, key.objectid, key.type);
 766                 if (ret < 0)
 767                         goto out;
 768         }
 769
 770         while (1) {
 771                 l = path->nodes[0];
 772                 slot = path->slots[0];
 773                 if (slot >= btrfs_header_nritems(l)) {
 774                         ret = btrfs_next_leaf(root, path);
 775                         if (ret == 0)
 776                                 continue;
 777                         if (ret < 0)
 778                                 goto out;
 779
 780                         break;
 781                 }
 782                 btrfs_item_key_to_cpu(l, &key, slot);
 783
 784                 if (key.objectid < device->devid)
 785                         goto next;
 786
 787                 if (key.objectid > device->devid)
 788                         break;
 789
 790                 if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
 791                         goto next;
 792
 793                 dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
 794                 extent_end = key.offset + btrfs_dev_extent_length(l,
 795                                                                   dev_extent);
 796                 if (key.offset <= start && extent_end > end) {
 797                         *length = end - start + 1;
 798                         break;
 799                 } else if (key.offset <= start && extent_end > start)
 800                         *length += extent_end - start;
 801                 else if (key.offset > start && extent_end <= end)
 802                         *length += extent_end - key.offset;
 803                 else if (key.offset > start && key.offset <= end) {
 804                         *length += end - key.offset + 1;
 805                         break;
 806                 } else if (key.offset > end)
 807                         break;
 808
 809 next:
 810                 path->slots[0]++;
 811         }
 812         ret = 0;
 813 out:
 814         btrfs_free_path(path);
 815         return ret;
 816 }
 817
 818 /*
 819  * find_free_dev_extent - find free space in the specified device
 820  * @trans:      transaction handler
 821  * @device:     the device which we search the free space in
 822  * @num_bytes:  the size of the free space that we need
 823  * @start:      store the start of the free space.
 824  * @len:        the size of the free space. that we find, or the size of the max
 825  *              free space if we don't find suitable free space
 826  *
 827  * this uses a pretty simple search, the expectation is that it is
 828  * called very infrequently and that a given device has a small number
 829  * of extents
 830  *
 831  * @start is used to store the start of the free space if we find. But if we
 832  * don't find suitable free space, it will be used to store the start position
 833  * of the max free space.
 834  *
 835  * @len is used to store the size of the free space that we find.
 836  * But if we don't find suitable free space, it is used to store the size of
 837  * the max free space.
 838  */
 839 int find_free_dev_extent(struct btrfs_trans_handle *trans,
 840                          struct btrfs_device *device, u64 num_bytes,
 841                          u64 *start, u64 *len)
 842 {
 843         struct btrfs_key key;
 844         struct btrfs_root *root = device->dev_root;
 845         struct btrfs_dev_extent *dev_extent;
 846         struct btrfs_path *path;
 847         u64 hole_size;
 848         u64 max_hole_start;
 849         u64 max_hole_size;
 850         u64 extent_end;
 851         u64 search_start;
 852         u64 search_end = device->total_bytes;
 853         int ret;
 854         int slot;
 855         struct extent_buffer *l;
 856
 857         /* FIXME use last free of some kind */
 858
 859         /* we don't want to overwrite the superblock on the drive,
 860          * so we make sure to start at an offset of at least 1MB
 861          */
 862         search_start = max(root->fs_info->alloc_start, 1024ull * 1024);
 863
 864         max_hole_start = search_start;
 865         max_hole_size = 0;
 866         hole_size = 0;
 867
 868         if (search_start >= search_end) {
 869                 ret = -ENOSPC;
 870                 goto error;
 871         }
 872
 873         path = btrfs_alloc_path();
 874         if (!path) {
 875                 ret = -ENOMEM;
 876                 goto error;
 877         }
 878         path->reada = 2;
 879
 880         key.objectid = device->devid;
 881         key.offset = search_start;
 882         key.type = BTRFS_DEV_EXTENT_KEY;
 883
 884         ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
 885         if (ret < 0)
 886                 goto out;
 887         if (ret > 0) {
 888                 ret = btrfs_previous_item(root, path, key.objectid, key.type);
 889                 if (ret < 0)
 890                         goto out;
 891         }
 892
 893         while (1) {
 894                 l = path->nodes[0];
 895                 slot = path->slots[0];
 896                 if (slot >= btrfs_header_nritems(l)) {
 897                         ret = btrfs_next_leaf(root, path);
 898                         if (ret == 0)
 899                                 continue;
 900                         if (ret < 0)
 901                                 goto out;
 902
 903                         break;
 904                 }
 905                 btrfs_item_key_to_cpu(l, &key, slot);
 906
 907                 if (key.objectid < device->devid)
 908                         goto next;
 909
 910                 if (key.objectid > device->devid)
 911                         break;
 912
 913                 if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
 914                         goto next;
 915
 916                 if (key.offset > search_start) {
 917                         hole_size = key.offset - search_start;
 918
 919                         if (hole_size > max_hole_size) {
 920                                 max_hole_start = search_start;
 921                                 max_hole_size = hole_size;
 922                         }
 923
 924                         /*
 925                          * If this free space is greater than which we need,
 926                          * it must be the max free space that we have found
 927                          * until now, so max_hole_start must point to the start
 928                          * of this free space and the length of this free space
 929                          * is stored in max_hole_size. Thus, we return
 930                          * max_hole_start and max_hole_size and go back to the
 931                          * caller.
 932                          */
 933                         if (hole_size >= num_bytes) {
 934                                 ret = 0;
 935                                 goto out;
 936                         }
 937                 }
 938
 939                 dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
 940                 extent_end = key.offset + btrfs_dev_extent_length(l,
 941                                                                   dev_extent);
 942                 if (extent_end > search_start)
 943                         search_start = extent_end;
 944 next:
 945                 path->slots[0]++;
 946                 cond_resched();
 947         }
 948
 949         /*
 950          * At this point, search_start should be the end of
 951          * allocated dev extents, and when shrinking the device,
 952          * search_end may be smaller than search_start.
 953          */
 954         if (search_end > search_start)
 955                 hole_size = search_end - search_start;
 956
 957         if (hole_size > max_hole_size) {
 958                 max_hole_start = search_start;
 959                 max_hole_size = hole_size;
 960         }
 961
 962         /* See above. */
 963         if (hole_size < num_bytes)
 964                 ret = -ENOSPC;
 965         else
 966                 ret = 0;
 967
 968 out:
 969         btrfs_free_path(path);
 970 error:
 971         *start = max_hole_start;
 972         if (len)
 973                 *len = max_hole_size;
 974         return ret;
 975 }
 976
 977 static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
 978                           struct btrfs_device *device,
 979                           u64 start)
 980 {
 981         int ret;
 982         struct btrfs_path *path;
 983         struct btrfs_root *root = device->dev_root;
 984         struct btrfs_key key;
 985         struct btrfs_key found_key;
 986         struct extent_buffer *leaf = NULL;
 987         struct btrfs_dev_extent *extent = NULL;
 988
 989         path = btrfs_alloc_path();
 990         if (!path)
 991                 return -ENOMEM;
 992
 993         key.objectid = device->devid;
 994         key.offset = start;
 995         key.type = BTRFS_DEV_EXTENT_KEY;
 996
 997         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 998         if (ret > 0) {
 999                 ret = btrfs_previous_item(root, path, key.objectid,
1000                                           BTRFS_DEV_EXTENT_KEY);
1001                 if (ret)
1002                         goto out;
1003                 leaf = path->nodes[0];
1004                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
1005                 extent = btrfs_item_ptr(leaf, path->slots[0],
1006                                         struct btrfs_dev_extent);
1007                 BUG_ON(found_key.offset > start || found_key.offset +
1008                        btrfs_dev_extent_length(leaf, extent) < start);
1009         } else if (ret == 0) {
1010                 leaf = path->nodes[0];
1011                 extent = btrfs_item_ptr(leaf, path->slots[0],
1012                                         struct btrfs_dev_extent);
1013         }
1014         BUG_ON(ret);
1015
1016         if (device->bytes_used > 0)
1017                 device->bytes_used -= btrfs_dev_extent_length(leaf, extent);
1018         ret = btrfs_del_item(trans, root, path);
1019
1020 out:
1021         btrfs_free_path(path);
1022         return ret;
1023 }
1024
1025 int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
1026                            struct btrfs_device *device,
1027                            u64 chunk_tree, u64 chunk_objectid,
1028                            u64 chunk_offset, u64 start, u64 num_bytes)
1029 {
1030         int ret;
1031         struct btrfs_path *path;
1032         struct btrfs_root *root = device->dev_root;
1033         struct btrfs_dev_extent *extent;
1034         struct extent_buffer *leaf;
1035         struct btrfs_key key;
1036
1037         WARN_ON(!device->in_fs_metadata);
1038         path = btrfs_alloc_path();
1039         if (!path)
1040                 return -ENOMEM;
1041
1042         key.objectid = device->devid;
1043         key.offset = start;
1044         key.type = BTRFS_DEV_EXTENT_KEY;
1045         ret = btrfs_insert_empty_item(trans, root, path, &key,
1046                                       sizeof(*extent));
1047         BUG_ON(ret);
1048
1049         leaf = path->nodes[0];
1050         extent = btrfs_item_ptr(leaf, path->slots[0],
1051                                 struct btrfs_dev_extent);
1052         btrfs_set_dev_extent_chunk_tree(leaf, extent, chunk_tree);
1053         btrfs_set_dev_extent_chunk_objectid(leaf, extent, chunk_objectid);
1054         btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);
1055
1056         write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
1057                     (unsigned long)btrfs_dev_extent_chunk_tree_uuid(extent),
1058                     BTRFS_UUID_SIZE);
1059
1060         btrfs_set_dev_extent_length(leaf, extent, num_bytes);
1061         btrfs_mark_buffer_dirty(leaf);
1062         btrfs_free_path(path);
1063         return ret;
1064 }
1065
1066 static noinline int find_next_chunk(struct btrfs_root *root,
1067                                     u64 objectid, u64 *offset)
1068 {
1069         struct btrfs_path *path;
1070         int ret;
1071         struct btrfs_key key;
1072         struct btrfs_chunk *chunk;
1073         struct btrfs_key found_key;
1074
1075         path = btrfs_alloc_path();
1076         if (!path)
1077                 return -ENOMEM;
1078
1079         key.objectid = objectid;
1080         key.offset = (u64)-1;
1081         key.type = BTRFS_CHUNK_ITEM_KEY;
1082
1083         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1084         if (ret < 0)
1085                 goto error;
1086
1087         BUG_ON(ret == 0);
1088
1089         ret = btrfs_previous_item(root, path, 0, BTRFS_CHUNK_ITEM_KEY);
1090         if (ret) {
1091                 *offset = 0;
1092         } else {
1093                 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
1094                                       path->slots[0]);
1095                 if (found_key.objectid != objectid)
1096                         *offset = 0;
1097                 else {
1098                         chunk = btrfs_item_ptr(path->nodes[0], path->slots[0],
1099                                                struct btrfs_chunk);
1100                         *offset = found_key.offset +
1101                                 btrfs_chunk_length(path->nodes[0], chunk);
1102                 }
1103         }
1104         ret = 0;
1105 error:
1106         btrfs_free_path(path);
1107         return ret;
1108 }
1109
1110 static noinline int find_next_devid(struct btrfs_root *root, u64 *objectid)
1111 {
1112         int ret;
1113         struct btrfs_key key;
1114         struct btrfs_key found_key;
1115         struct btrfs_path *path;
1116
1117         root = root->fs_info->chunk_root;
1118
1119         path = btrfs_alloc_path();
1120         if (!path)
1121                 return -ENOMEM;
1122
1123         key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
1124         key.type = BTRFS_DEV_ITEM_KEY;
1125         key.offset = (u64)-1;
1126
1127         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1128         if (ret < 0)
1129                 goto error;
1130
1131         BUG_ON(ret == 0);
1132
1133         ret = btrfs_previous_item(root, path, BTRFS_DEV_ITEMS_OBJECTID,
1134                                   BTRFS_DEV_ITEM_KEY);
1135         if (ret) {
1136                 *objectid = 1;
1137         } else {
1138                 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
1139                                       path->slots[0]);
1140                 *objectid = found_key.offset + 1;
1141         }
1142         ret = 0;
1143 error:
1144         btrfs_free_path(path);
1145         return ret;
1146 }
1147
1148 /*
1149  * the device information is stored in the chunk root
1150  * the btrfs_device struct should be fully filled in
1151  */
1152 int btrfs_add_device(struct btrfs_trans_handle *trans,
1153                      struct btrfs_root *root,
1154                      struct btrfs_device *device)
1155 {
1156         int ret;
1157         struct btrfs_path *path;
1158         struct btrfs_dev_item *dev_item;
1159         struct extent_buffer *leaf;
1160         struct btrfs_key key;
1161         unsigned long ptr;
1162
1163         root = root->fs_info->chunk_root;
1164
1165         path = btrfs_alloc_path();
1166         if (!path)
1167                 return -ENOMEM;
1168
1169         key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
1170         key.type = BTRFS_DEV_ITEM_KEY;
1171         key.offset = device->devid;
1172
1173         ret = btrfs_insert_empty_item(trans, root, path, &key,
1174                                       sizeof(*dev_item));
1175         if (ret)
1176                 goto out;
1177
1178         leaf = path->nodes[0];
1179         dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);
1180
1181         btrfs_set_device_id(leaf, dev_item, device->devid);
1182         btrfs_set_device_generation(leaf, dev_item, 0);
1183         btrfs_set_device_type(leaf, dev_item, device->type);
1184         btrfs_set_device_io_align(leaf, dev_item, device->io_align);
1185         btrfs_set_device_io_width(leaf, dev_item, device->io_width);
1186         btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
1187         btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
1188         btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
1189         btrfs_set_device_group(leaf, dev_item, 0);
1190         btrfs_set_device_seek_speed(leaf, dev_item, 0);
1191         btrfs_set_device_bandwidth(leaf, dev_item, 0);
1192         btrfs_set_device_start_offset(leaf, dev_item, 0);
1193
1194         ptr = (unsigned long)btrfs_device_uuid(dev_item);
1195         write_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
1196         ptr = (unsigned long)btrfs_device_fsid(dev_item);
1197         write_extent_buffer(leaf, root->fs_info->fsid, ptr, BTRFS_UUID_SIZE);
1198         btrfs_mark_buffer_dirty(leaf);
1199
1200         ret = 0;
1201 out:
1202         btrfs_free_path(path);
1203         return ret;
1204 }
1205
1206 static int btrfs_rm_dev_item(struct btrfs_root *root,
1207                              struct btrfs_device *device)
1208 {
1209         int ret;
1210         struct btrfs_path *path;
1211         struct btrfs_key key;
1212         struct btrfs_trans_handle *trans;
1213
1214         root = root->fs_info->chunk_root;
1215
1216         path = btrfs_alloc_path();
1217         if (!path)
1218                 return -ENOMEM;
1219
1220         trans = btrfs_start_transaction(root, 0);
1221         if (IS_ERR(trans)) {
1222                 btrfs_free_path(path);
1223                 return PTR_ERR(trans);
1224         }
1225         key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
1226         key.type = BTRFS_DEV_ITEM_KEY;
1227         key.offset = device->devid;
1228         lock_chunks(root);
1229
1230         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1231         if (ret < 0)
1232                 goto out;
1233
1234         if (ret > 0) {
1235                 ret = -ENOENT;
1236                 goto out;
1237         }
1238
1239         ret = btrfs_del_item(trans, root, path);
1240         if (ret)
1241                 goto out;
1242 out:
1243         btrfs_free_path(path);
1244         unlock_chunks(root);
1245         btrfs_commit_transaction(trans, root);
1246         return ret;
1247 }
1248
1249 int btrfs_rm_device(struct btrfs_root *root, char *device_path)
1250 {
1251         struct btrfs_device *device;
1252         struct btrfs_device *next_device;
1253         struct block_device *bdev;
1254         struct buffer_head *bh = NULL;
1255         struct btrfs_super_block *disk_super;
1256         struct btrfs_fs_devices *cur_devices;
1257         u64 all_avail;
1258         u64 devid;
1259         u64 num_devices;
1260         u8 *dev_uuid;
1261         int ret = 0;
1262         bool clear_super = false;
1263
1264         mutex_lock(&uuid_mutex);
1265         mutex_lock(&root->fs_info->volume_mutex);
1266
1267         all_avail = root->fs_info->avail_data_alloc_bits |
1268                 root->fs_info->avail_system_alloc_bits |
1269                 root->fs_info->avail_metadata_alloc_bits;
1270
1271         if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) &&
1272             root->fs_info->fs_devices->num_devices <= 4) {
1273                 printk(KERN_ERR "btrfs: unable to go below four devices "
1274                        "on raid10\n");
1275                 ret = -EINVAL;
1276                 goto out;
1277         }
1278
1279         if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) &&
1280             root->fs_info->fs_devices->num_devices <= 2) {
1281                 printk(KERN_ERR "btrfs: unable to go below two "
1282                        "devices on raid1\n");
1283                 ret = -EINVAL;
1284                 goto out;
1285         }
1286
1287         if (strcmp(device_path, "missing") == 0) {
1288                 struct list_head *devices;
1289                 struct btrfs_device *tmp;
1290
1291                 device = NULL;
1292                 devices = &root->fs_info->fs_devices->devices;
1293                 /*
1294                  * It is safe to read the devices since the volume_mutex
1295                  * is held.
1296                  */
1297                 list_for_each_entry(tmp, devices, dev_list) {
1298                         if (tmp->in_fs_metadata && !tmp->bdev) {
1299                                 device = tmp;
1300                                 break;
1301                         }
1302                 }
1303                 bdev = NULL;
1304                 bh = NULL;
1305                 disk_super = NULL;
1306                 if (!device) {
1307                         printk(KERN_ERR "btrfs: no missing devices found to "
1308                                "remove\n");
1309                         goto out;
1310                 }
1311         } else {
1312                 bdev = blkdev_get_by_path(device_path, FMODE_READ | FMODE_EXCL,
1313                                           root->fs_info->bdev_holder);
1314                 if (IS_ERR(bdev)) {
1315                         ret = PTR_ERR(bdev);
1316                         goto out;
1317                 }
1318
1319                 set_blocksize(bdev, 4096);
1320                 bh = btrfs_read_dev_super(bdev);
1321                 if (!bh) {
1322                         ret = -EINVAL;
1323                         goto error_close;
1324                 }
1325                 disk_super = (struct btrfs_super_block *)bh->b_data;
1326                 devid = btrfs_stack_device_id(&disk_super->dev_item);
1327                 dev_uuid = disk_super->dev_item.uuid;
1328                 device = btrfs_find_device(root, devid, dev_uuid,
1329                                            disk_super->fsid);
1330                 if (!device) {
1331                         ret = -ENOENT;
1332                         goto error_brelse;
1333                 }
1334         }
1335
1336         if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) {
1337                 printk(KERN_ERR "btrfs: unable to remove the only writeable "
1338                        "device\n");
1339                 ret = -EINVAL;
1340                 goto error_brelse;
1341         }
1342
1343         if (device->writeable) {
1344                 lock_chunks(root);
1345                 list_del_init(&device->dev_alloc_list);
1346                 unlock_chunks(root);
1347                 root->fs_info->fs_devices->rw_devices--;
1348                 clear_super = true;
1349         }
1350
1351         ret = btrfs_shrink_device(device, 0);
1352         if (ret)
1353                 goto error_undo;
1354
1355         ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device);
1356         if (ret)
1357                 goto error_undo;
1358
1359         device->in_fs_metadata = 0;
1360         btrfs_scrub_cancel_dev(root, device);
1361
1362         /*
1363          * the device list mutex makes sure that we don't change
1364          * the device list while someone else is writing out all
1365          * the device supers.
1366          */
1367
1368         cur_devices = device->fs_devices;
1369         mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
1370         list_del_rcu(&device->dev_list);
1371
1372         device->fs_devices->num_devices--;
1373
1374         if (device->missing)
1375                 root->fs_info->fs_devices->missing_devices--;
1376
1377         next_device = list_entry(root->fs_info->fs_devices->devices.next,
1378                                  struct btrfs_device, dev_list);
1379         if (device->bdev == root->fs_info->sb->s_bdev)
1380                 root->fs_info->sb->s_bdev = next_device->bdev;
1381         if (device->bdev == root->fs_info->fs_devices->latest_bdev)
1382                 root->fs_info->fs_devices->latest_bdev = next_device->bdev;
1383
1384         if (device->bdev)
1385                 device->fs_devices->open_devices--;
1386
1387         call_rcu(&device->rcu, free_device);
1388         mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
1389
1390         num_devices = btrfs_super_num_devices(&root->fs_info->super_copy) - 1;
1391         btrfs_set_super_num_devices(&root->fs_info->super_copy, num_devices);
1392
1393         if (cur_devices->open_devices == 0) {
1394                 struct btrfs_fs_devices *fs_devices;
1395                 fs_devices = root->fs_info->fs_devices;
1396                 while (fs_devices) {
1397                         if (fs_devices->seed == cur_devices)
1398                                 break;
1399                         fs_devices = fs_devices->seed;
1400                 }
1401                 fs_devices->seed = cur_devices->seed;
1402                 cur_devices->seed = NULL;
1403                 lock_chunks(root);
1404                 __btrfs_close_devices(cur_devices);
1405                 unlock_chunks(root);
1406                 free_fs_devices(cur_devices);
1407         }
1408
1409         /*
1410          * at this point, the device is zero sized.  We want to
1411          * remove it from the devices list and zero out the old super
1412          */
1413         if (clear_super) {
1414                 /* make sure this device isn't detected as part of
1415                  * the FS anymore
1416                  */
1417                 memset(&disk_super->magic, 0, sizeof(disk_super->magic));
1418                 set_buffer_dirty(bh);
1419                 sync_dirty_buffer(bh);
1420         }
1421
1422         ret = 0;
1423
1424 error_brelse:
1425         brelse(bh);
1426 error_close:
1427         if (bdev)
1428                 blkdev_put(bdev, FMODE_READ | FMODE_EXCL);
1429 out:
1430         mutex_unlock(&root->fs_info->volume_mutex);
1431         mutex_unlock(&uuid_mutex);
1432         return ret;
1433 error_undo:
1434         if (device->writeable) {
1435                 lock_chunks(root);
1436                 list_add(&device->dev_alloc_list,
1437                          &root->fs_info->fs_devices->alloc_list);
1438                 unlock_chunks(root);
1439                 root->fs_info->fs_devices->rw_devices++;
1440         }
1441         goto error_brelse;
1442 }
1443
1444 /*
1445  * does all the dirty work required for changing file system's UUID.
1446  */
1447 static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans,
1448                                 struct btrfs_root *root)
1449 {
1450         struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
1451         struct btrfs_fs_devices *old_devices;
1452         struct btrfs_fs_devices *seed_devices;
1453         struct btrfs_super_block *disk_super = &root->fs_info->super_copy;
1454         struct btrfs_device *device;
1455         u64 super_flags;
1456
1457         BUG_ON(!mutex_is_locked(&uuid_mutex));
1458         if (!fs_devices->seeding)
1459                 return -EINVAL;
1460
1461         seed_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
1462         if (!seed_devices)
1463                 return -ENOMEM;
1464
1465         old_devices = clone_fs_devices(fs_devices);
1466         if (IS_ERR(old_devices)) {
1467                 kfree(seed_devices);
1468                 return PTR_ERR(old_devices);
1469         }
1470
1471         list_add(&old_devices->list, &fs_uuids);
1472
1473         memcpy(seed_devices, fs_devices, sizeof(*seed_devices));
1474         seed_devices->opened = 1;
1475         INIT_LIST_HEAD(&seed_devices->devices);
1476         INIT_LIST_HEAD(&seed_devices->alloc_list);
1477         mutex_init(&seed_devices->device_list_mutex);
1478
1479         mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
1480         list_splice_init_rcu(&fs_devices->devices, &seed_devices->devices,
1481                               synchronize_rcu);
1482         mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
1483
1484         list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list);
1485         list_for_each_entry(device, &seed_devices->devices, dev_list) {
1486                 device->fs_devices = seed_devices;
1487         }
1488
1489         fs_devices->seeding = 0;
1490         fs_devices->num_devices = 0;
1491         fs_devices->open_devices = 0;
1492         fs_devices->seed = seed_devices;
1493
1494         generate_random_uuid(fs_devices->fsid);
1495         memcpy(root->fs_info->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
1496         memcpy(disk_super->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
1497         super_flags = btrfs_super_flags(disk_super) &
1498                       ~BTRFS_SUPER_FLAG_SEEDING;
1499         btrfs_set_super_flags(disk_super, super_flags);
1500
1501         return 0;
1502 }
1503
1504 /*
1505  * strore the expected generation for seed devices in device items.
1506  */
1507 static int btrfs_finish_sprout(struct btrfs_trans_handle *trans,
1508                                struct btrfs_root *root)
1509 {
1510         struct btrfs_path *path;
1511         struct extent_buffer *leaf;
1512         struct btrfs_dev_item *dev_item;
1513         struct btrfs_device *device;
1514         struct btrfs_key key;
1515         u8 fs_uuid[BTRFS_UUID_SIZE];
1516         u8 dev_uuid[BTRFS_UUID_SIZE];
1517         u64 devid;
1518         int ret;
1519
1520         path = btrfs_alloc_path();
1521         if (!path)
1522                 return -ENOMEM;
1523
1524         root = root->fs_info->chunk_root;
1525         key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
1526         key.offset = 0;
1527         key.type = BTRFS_DEV_ITEM_KEY;
1528
1529         while (1) {
1530                 ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
1531                 if (ret < 0)
1532                         goto error;
1533
1534                 leaf = path->nodes[0];
1535 next_slot:
1536                 if (path->slots[0] >= btrfs_header_nritems(leaf)) {
1537                         ret = btrfs_next_leaf(root, path);
1538                         if (ret > 0)
1539                                 break;
1540                         if (ret < 0)
1541                                 goto error;
1542                         leaf = path->nodes[0];
1543                         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1544                         btrfs_release_path(path);
1545                         continue;
1546                 }
1547
1548                 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1549                 if (key.objectid != BTRFS_DEV_ITEMS_OBJECTID ||
1550                     key.type != BTRFS_DEV_ITEM_KEY)
1551                         break;
1552
1553                 dev_item = btrfs_item_ptr(leaf, path->slots[0],
1554                                           struct btrfs_dev_item);
1555                 devid = btrfs_device_id(leaf, dev_item);
1556                 read_extent_buffer(leaf, dev_uuid,
1557                                    (unsigned long)btrfs_device_uuid(dev_item),
1558                                    BTRFS_UUID_SIZE);
1559                 read_extent_buffer(leaf, fs_uuid,
1560                                    (unsigned long)btrfs_device_fsid(dev_item),
1561                                    BTRFS_UUID_SIZE);
1562                 device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
1563                 BUG_ON(!device);
1564
1565                 if (device->fs_devices->seeding) {
1566                         btrfs_set_device_generation(leaf, dev_item,
1567                                                     device->generation);
1568                         btrfs_mark_buffer_dirty(leaf);
1569                 }
1570
1571                 path->slots[0]++;
1572                 goto next_slot;
1573         }
1574         ret = 0;
1575 error:
1576         btrfs_free_path(path);
1577         return ret;
1578 }
1579
1580 int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
1581 {
1582         struct request_queue *q;
1583         struct btrfs_trans_handle *trans;
1584         struct btrfs_device *device;
1585         struct block_device *bdev;
1586         struct list_head *devices;
1587         struct super_block *sb = root->fs_info->sb;
1588         u64 total_bytes;
1589         int seeding_dev = 0;
1590         int ret = 0;
1591
1592         if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding)
1593                 return -EINVAL;
1594
1595         bdev = blkdev_get_by_path(device_path, FMODE_EXCL,
1596                                   root->fs_info->bdev_holder);
1597         if (IS_ERR(bdev))
1598                 return PTR_ERR(bdev);
1599
1600         if (root->fs_info->fs_devices->seeding) {
1601                 seeding_dev = 1;
1602                 down_write(&sb->s_umount);
1603                 mutex_lock(&uuid_mutex);
1604         }
1605
1606         filemap_write_and_wait(bdev->bd_inode->i_mapping);
1607         mutex_lock(&root->fs_info->volume_mutex);
1608
1609         devices = &root->fs_info->fs_devices->devices;
1610         /*
1611          * we have the volume lock, so we don't need the extra
1612          * device list mutex while reading the list here.
1613          */
1614         list_for_each_entry(device, devices, dev_list) {
1615                 if (device->bdev == bdev) {
1616                         ret = -EEXIST;
1617                         goto error;
1618                 }
1619         }
1620
1621         device = kzalloc(sizeof(*device), GFP_NOFS);
1622         if (!device) {
1623                 /* we can safely leave the fs_devices entry around */
1624                 ret = -ENOMEM;
1625                 goto error;
1626         }
1627
1628         device->name = kstrdup(device_path, GFP_NOFS);
1629         if (!device->name) {
1630                 kfree(device);
1631                 ret = -ENOMEM;
1632                 goto error;
1633         }
1634
1635         ret = find_next_devid(root, &device->devid);
1636         if (ret) {
1637                 kfree(device->name);
1638                 kfree(device);
1639                 goto error;
1640         }
1641
1642         trans = btrfs_start_transaction(root, 0);
1643         if (IS_ERR(trans)) {
1644                 kfree(device->name);
1645                 kfree(device);
1646                 ret = PTR_ERR(trans);
1647                 goto error;
1648         }
1649
1650         lock_chunks(root);
1651
1652         q = bdev_get_queue(bdev);
1653         if (blk_queue_discard(q))
1654                 device->can_discard = 1;
1655         device->writeable = 1;
1656         device->work.func = pending_bios_fn;
1657         generate_random_uuid(device->uuid);
1658         spin_lock_init(&device->io_lock);
1659         device->generation = trans->transid;
1660         device->io_width = root->sectorsize;
1661         device->io_align = root->sectorsize;
1662         device->sector_size = root->sectorsize;
1663         device->total_bytes = i_size_read(bdev->bd_inode);
1664         device->disk_total_bytes = device->total_bytes;
1665         device->dev_root = root->fs_info->dev_root;
1666         device->bdev = bdev;
1667         device->in_fs_metadata = 1;
1668         device->mode = FMODE_EXCL;
1669         set_blocksize(device->bdev, 4096);
1670
1671         if (seeding_dev) {
1672                 sb->s_flags &= ~MS_RDONLY;
1673                 ret = btrfs_prepare_sprout(trans, root);
1674                 BUG_ON(ret);
1675         }
1676
1677         device->fs_devices = root->fs_info->fs_devices;
1678
1679         /*
1680          * we don't want write_supers to jump in here with our device
1681          * half setup
1682          */
1683         mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
1684         list_add_rcu(&device->dev_list, &root->fs_info->fs_devices->devices);
1685         list_add(&device->dev_alloc_list,
1686                  &root->fs_info->fs_devices->alloc_list);
1687         root->fs_info->fs_devices->num_devices++;
1688         root->fs_info->fs_devices->open_devices++;
1689         root->fs_info->fs_devices->rw_devices++;
1690         if (device->can_discard)
1691                 root->fs_info->fs_devices->num_can_discard++;
1692         root->fs_info->fs_devices->total_rw_bytes += device->total_bytes;
1693
1694         if (!blk_queue_nonrot(bdev_get_queue(bdev)))
1695                 root->fs_info->fs_devices->rotating = 1;
1696
1697         total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
1698         btrfs_set_super_total_bytes(&root->fs_info->super_copy,
1699                                     total_bytes + device->total_bytes);
1700
1701         total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
1702         btrfs_set_super_num_devices(&root->fs_info->super_copy,
1703                                     total_bytes + 1);
1704         mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
1705
1706         if (seeding_dev) {
1707                 ret = init_first_rw_device(trans, root, device);
1708                 BUG_ON(ret);
1709                 ret = btrfs_finish_sprout(trans, root);
1710                 BUG_ON(ret);
1711         } else {
1712                 ret = btrfs_add_device(trans, root, device);
1713         }
1714
1715         /*
1716          * we've got more storage, clear any full flags on the space
1717          * infos
1718          */
1719         btrfs_clear_space_info_full(root->fs_info);
1720
1721         unlock_chunks(root);
1722         btrfs_commit_transaction(trans, root);
1723
1724         if (seeding_dev) {
1725                 mutex_unlock(&uuid_mutex);
1726                 up_write(&sb->s_umount);
1727
1728                 ret = btrfs_relocate_sys_chunks(root);
1729                 BUG_ON(ret);
1730         }
1731 out:
1732         mutex_unlock(&root->fs_info->volume_mutex);
1733         return ret;
1734 error:
1735         blkdev_put(bdev, FMODE_EXCL);
1736         if (seeding_dev) {
1737                 mutex_unlock(&uuid_mutex);
1738                 up_write(&sb->s_umount);
1739         }
1740         goto out;
1741 }
1742
1743 static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
1744                                         struct btrfs_device *device)
1745 {
1746         int ret;
1747         struct btrfs_path *path;
1748         struct btrfs_root *root;
1749         struct btrfs_dev_item *dev_item;
1750         struct extent_buffer *leaf;
1751         struct btrfs_key key;
1752
1753         root = device->dev_root->fs_info->chunk_root;
1754
1755         path = btrfs_alloc_path();
1756         if (!path)
1757                 return -ENOMEM;
1758
1759         key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
1760         key.type = BTRFS_DEV_ITEM_KEY;
1761         key.offset = device->devid;
1762
1763         ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
1764         if (ret < 0)
1765                 goto out;
1766
1767         if (ret > 0) {
1768                 ret = -ENOENT;
1769                 goto out;
1770         }
1771
1772         leaf = path->nodes[0];
1773         dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);
1774
1775         btrfs_set_device_id(leaf, dev_item, device->devid);
1776         btrfs_set_device_type(leaf, dev_item, device->type);
1777         btrfs_set_device_io_align(leaf, dev_item, device->io_align);
1778         btrfs_set_device_io_width(leaf, dev_item, device->io_width);
1779         btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
1780         btrfs_set_device_total_bytes(leaf, dev_item, device->disk_total_bytes);
1781         btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
1782         btrfs_mark_buffer_dirty(leaf);
1783
1784 out:
1785         btrfs_free_path(path);
1786         return ret;
1787 }
1788
1789 static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
1790                       struct btrfs_device *device, u64 new_size)
1791 {
1792         struct btrfs_super_block *super_copy =
1793                 &device->dev_root->fs_info->super_copy;
1794         u64 old_total = btrfs_super_total_bytes(super_copy);
1795         u64 diff = new_size - device->total_bytes;
1796
1797         if (!device->writeable)
1798                 return -EACCES;
1799         if (new_size <= device->total_bytes)
1800                 return -EINVAL;
1801
1802         btrfs_set_super_total_bytes(super_copy, old_total + diff);
1803         device->fs_devices->total_rw_bytes += diff;
1804
1805         device->total_bytes = new_size;
1806         device->disk_total_bytes = new_size;
1807         btrfs_clear_space_info_full(device->dev_root->fs_info);
1808
1809         return btrfs_update_device(trans, device);
1810 }
1811
1812 int btrfs_grow_device(struct btrfs_trans_handle *trans,
1813                       struct btrfs_device *device, u64 new_size)
1814 {
1815         int ret;
1816         lock_chunks(device->dev_root);
1817         ret = __btrfs_grow_device(trans, device, new_size);
1818         unlock_chunks(device->dev_root);
1819         return ret;
1820 }
1821
1822 static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
1823                             struct btrfs_root *root,
1824                             u64 chunk_tree, u64 chunk_objectid,
1825                             u64 chunk_offset)
1826 {
1827         int ret;
1828         struct btrfs_path *path;
1829         struct btrfs_key key;
1830
1831         root = root->fs_info->chunk_root;
1832         path = btrfs_alloc_path();
1833         if (!path)
1834                 return -ENOMEM;
1835
1836         key.objectid = chunk_objectid;
1837         key.offset = chunk_offset;
1838         key.type = BTRFS_CHUNK_ITEM_KEY;
1839
1840         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1841         BUG_ON(ret);
1842
1843         ret = btrfs_del_item(trans, root, path);
1844
1845         btrfs_free_path(path);
1846         return ret;
1847 }
1848
1849 static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
1850                         chunk_offset)
1851 {
1852         struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
1853         struct btrfs_disk_key *disk_key;
1854         struct btrfs_chunk *chunk;
1855         u8 *ptr;
1856         int ret = 0;
1857         u32 num_stripes;
1858         u32 array_size;
1859         u32 len = 0;
1860         u32 cur;
1861         struct btrfs_key key;
1862
1863         array_size = btrfs_super_sys_array_size(super_copy);
1864
1865         ptr = super_copy->sys_chunk_array;
1866         cur = 0;
1867
1868         while (cur < array_size) {
1869                 disk_key = (struct btrfs_disk_key *)ptr;
1870                 btrfs_disk_key_to_cpu(&key, disk_key);
1871
1872                 len = sizeof(*disk_key);
1873
1874                 if (key.type == BTRFS_CHUNK_ITEM_KEY) {
1875                         chunk = (struct btrfs_chunk *)(ptr + len);
1876                         num_stripes = btrfs_stack_chunk_num_stripes(chunk);
1877                         len += btrfs_chunk_item_size(num_stripes);
1878                 } else {
1879                         ret = -EIO;
1880                         break;
1881                 }
1882                 if (key.objectid == chunk_objectid &&
1883                     key.offset == chunk_offset) {
1884                         memmove(ptr, ptr + len, array_size - (cur + len));
1885                         array_size -= len;
1886                         btrfs_set_super_sys_array_size(super_copy, array_size);
1887                 } else {
1888                         ptr += len;
1889                         cur += len;
1890                 }
1891         }
1892         return ret;
1893 }
1894
1895 static int btrfs_relocate_chunk(struct btrfs_root *root,
1896                          u64 chunk_tree, u64 chunk_objectid,
1897                          u64 chunk_offset)
1898 {
1899         struct extent_map_tree *em_tree;
1900         struct btrfs_root *extent_root;
1901         struct btrfs_trans_handle *trans;
1902         struct extent_map *em;
1903         struct map_lookup *map;
1904         int ret;
1905         int i;
1906
1907         root = root->fs_info->chunk_root;
1908         extent_root = root->fs_info->extent_root;
1909         em_tree = &root->fs_info->mapping_tree.map_tree;
1910
1911         ret = btrfs_can_relocate(extent_root, chunk_offset);
1912         if (ret)
1913                 return -ENOSPC;
1914
1915         /* step one, relocate all the extents inside this chunk */
1916         ret = btrfs_relocate_block_group(extent_root, chunk_offset);
1917         if (ret)
1918                 return ret;
1919
1920         trans = btrfs_start_transaction(root, 0);
1921         BUG_ON(IS_ERR(trans));
1922
1923         lock_chunks(root);
1924
1925         /*
1926          * step two, delete the device extents and the
1927          * chunk tree entries
1928          */
1929         read_lock(&em_tree->lock);
1930         em = lookup_extent_mapping(em_tree, chunk_offset, 1);
1931         read_unlock(&em_tree->lock);
1932
1933         BUG_ON(em->start > chunk_offset ||
1934                em->start + em->len < chunk_offset);
1935         map = (struct map_lookup *)em->bdev;
1936
1937         for (i = 0; i < map->num_stripes; i++) {
1938                 ret = btrfs_free_dev_extent(trans, map->stripes[i].dev,
1939                                             map->stripes[i].physical);
1940                 BUG_ON(ret);
1941
1942                 if (map->stripes[i].dev) {
1943                         ret = btrfs_update_device(trans, map->stripes[i].dev);
1944                         BUG_ON(ret);
1945                 }
1946         }
1947         ret = btrfs_free_chunk(trans, root, chunk_tree, chunk_objectid,
1948                                chunk_offset);
1949
1950         BUG_ON(ret);
1951
1952         trace_btrfs_chunk_free(root, map, chunk_offset, em->len);
1953
1954         if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
1955                 ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset);
1956                 BUG_ON(ret);
1957         }
1958
1959         ret = btrfs_remove_block_group(trans, extent_root, chunk_offset);
1960         BUG_ON(ret);
1961
1962         write_lock(&em_tree->lock);
1963         remove_extent_mapping(em_tree, em);
1964         write_unlock(&em_tree->lock);
1965
1966         kfree(map);
1967         em->bdev = NULL;
1968
1969         /* once for the tree */
1970         free_extent_map(em);
1971         /* once for us */
1972         free_extent_map(em);
1973
1974         unlock_chunks(root);
1975         btrfs_end_transaction(trans, root);
1976         return 0;
1977 }
1978
1979 static int btrfs_relocate_sys_chunks(struct btrfs_root *root)
1980 {
1981         struct btrfs_root *chunk_root = root->fs_info->chunk_root;
1982         struct btrfs_path *path;
1983         struct extent_buffer *leaf;
1984         struct btrfs_chunk *chunk;
1985         struct btrfs_key key;
1986         struct btrfs_key found_key;
1987         u64 chunk_tree = chunk_root->root_key.objectid;
1988         u64 chunk_type;
1989         bool retried = false;
1990         int failed = 0;
1991         int ret;
1992
1993         path = btrfs_alloc_path();
1994         if (!path)
1995                 return -ENOMEM;
1996
1997 again:
1998         key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
1999         key.offset = (u64)-1;
2000         key.type = BTRFS_CHUNK_ITEM_KEY;
2001
2002         while (1) {
2003                 ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
2004                 if (ret < 0)
2005                         goto error;
2006                 BUG_ON(ret == 0);
2007
2008                 ret = btrfs_previous_item(chunk_root, path, key.objectid,
2009                                           key.type);
2010                 if (ret < 0)
2011                         goto error;
2012                 if (ret > 0)
2013                         break;
2014
2015                 leaf = path->nodes[0];
2016                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
2017
2018                 chunk = btrfs_item_ptr(leaf, path->slots[0],
2019                                        struct btrfs_chunk);
2020                 chunk_type = btrfs_chunk_type(leaf, chunk);
2021                 btrfs_release_path(path);
2022
2023                 if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) {
2024                         ret = btrfs_relocate_chunk(chunk_root, chunk_tree,
2025                                                    found_key.objectid,
2026                                                    found_key.offset);
2027                         if (ret == -ENOSPC)
2028                                 failed++;
2029                         else if (ret)
2030                                 BUG();
2031                 }
2032
2033                 if (found_key.offset == 0)
2034                         break;
2035                 key.offset = found_key.offset - 1;
2036         }
2037         ret = 0;
2038         if (failed && !retried) {
2039                 failed = 0;
2040                 retried = true;
2041                 goto again;
2042         } else if (failed && retried) {
2043                 WARN_ON(1);
2044                 ret = -ENOSPC;
2045         }
2046 error:
2047         btrfs_free_path(path);
2048         return ret;
2049 }
2050
2051 static u64 div_factor(u64 num, int factor)
2052 {
2053         if (factor == 10)
2054                 return num;
2055         num *= factor;
2056         do_div(num, 10);
2057         return num;
2058 }
2059
2060 int btrfs_balance(struct btrfs_root *dev_root)
2061 {
2062         int ret;
2063         struct list_head *devices = &dev_root->fs_info->fs_devices->devices;
2064         struct btrfs_device *device;
2065         u64 old_size;
2066         u64 size_to_free;
2067         struct btrfs_path *path;
2068         struct btrfs_key key;
2069         struct btrfs_root *chunk_root = dev_root->fs_info->chunk_root;
2070         struct btrfs_trans_handle *trans;
2071         struct btrfs_key found_key;
2072
2073         if (dev_root->fs_info->sb->s_flags & MS_RDONLY)
2074                 return -EROFS;
2075
2076         if (!capable(CAP_SYS_ADMIN))
2077                 return -EPERM;
2078
2079         mutex_lock(&dev_root->fs_info->volume_mutex);
2080         dev_root = dev_root->fs_info->dev_root;
2081
2082         /* step one make some room on all the devices */
2083         list_for_each_entry(device, devices, dev_list) {
2084                 old_size = device->total_bytes;
2085                 size_to_free = div_factor(old_size, 1);
2086                 size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
2087                 if (!device->writeable ||
2088                     device->total_bytes - device->bytes_used > size_to_free)
2089                         continue;
2090
2091                 ret = btrfs_shrink_device(device, old_size - size_to_free);
2092                 if (ret == -ENOSPC)
2093                         break;
2094                 BUG_ON(ret);
2095
2096                 trans = btrfs_start_transaction(dev_root, 0);
2097                 BUG_ON(IS_ERR(trans));
2098
2099                 ret = btrfs_grow_device(trans, device, old_size);
2100                 BUG_ON(ret);
2101
2102                 btrfs_end_transaction(trans, dev_root);
2103         }
2104
2105         /* step two, relocate all the chunks */
2106         path = btrfs_alloc_path();
2107         if (!path) {
2108                 ret = -ENOMEM;
2109                 goto error;
2110         }
2111         key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
2112         key.offset = (u64)-1;
2113         key.type = BTRFS_CHUNK_ITEM_KEY;
2114
2115         while (1) {
2116                 ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
2117                 if (ret < 0)
2118                         goto error;
2119
2120                 /*
2121                  * this shouldn't happen, it means the last relocate
2122                  * failed
2123                  */
2124                 if (ret == 0)
2125                         break;
2126
2127                 ret = btrfs_previous_item(chunk_root, path, 0,
2128                                           BTRFS_CHUNK_ITEM_KEY);
2129                 if (ret)
2130                         break;
2131
2132                 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
2133                                       path->slots[0]);
2134                 if (found_key.objectid != key.objectid)
2135                         break;
2136
2137                 /* chunk zero is special */
2138                 if (found_key.offset == 0)
2139                         break;
2140
2141                 btrfs_release_path(path);
2142                 ret = btrfs_relocate_chunk(chunk_root,
2143                                            chunk_root->root_key.objectid,
2144                                            found_key.objectid,
2145                                            found_key.offset);
2146                 if (ret && ret != -ENOSPC)
2147                         goto error;
2148                 key.offset = found_key.offset - 1;
2149         }
2150         ret = 0;
2151 error:
2152         btrfs_free_path(path);
2153         mutex_unlock(&dev_root->fs_info->volume_mutex);
2154         return ret;
2155 }
2156
2157 /*
2158  * shrinking a device means finding all of the device extents past
2159  * the new size, and then following the back refs to the chunks.
2160  * The chunk relocation code actually frees the device extent
2161  */
2162 int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
2163 {
2164         struct btrfs_trans_handle *trans;
2165         struct btrfs_root *root = device->dev_root;
2166         struct btrfs_dev_extent *dev_extent = NULL;
2167         struct btrfs_path *path;
2168         u64 length;
2169         u64 chunk_tree;
2170         u64 chunk_objectid;
2171         u64 chunk_offset;
2172         int ret;
2173         int slot;
2174         int failed = 0;
2175         bool retried = false;
2176         struct extent_buffer *l;
2177         struct btrfs_key key;
2178         struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
2179         u64 old_total = btrfs_super_total_bytes(super_copy);
2180         u64 old_size = device->total_bytes;
2181         u64 diff = device->total_bytes - new_size;
2182
2183         if (new_size >= device->total_bytes)
2184                 return -EINVAL;
2185
2186         path = btrfs_alloc_path();
2187         if (!path)
2188                 return -ENOMEM;
2189
2190         path->reada = 2;
2191
2192         lock_chunks(root);
2193
2194         device->total_bytes = new_size;
2195         if (device->writeable)
2196                 device->fs_devices->total_rw_bytes -= diff;
2197         unlock_chunks(root);
2198
2199 again:
2200         key.objectid = device->devid;
2201         key.offset = (u64)-1;
2202         key.type = BTRFS_DEV_EXTENT_KEY;
2203
2204         while (1) {
2205                 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
2206                 if (ret < 0)
2207                         goto done;
2208
2209                 ret = btrfs_previous_item(root, path, 0, key.type);
2210                 if (ret < 0)
2211                         goto done;
2212                 if (ret) {
2213                         ret = 0;
2214                         btrfs_release_path(path);
2215                         break;
2216                 }
2217
2218                 l = path->nodes[0];
2219                 slot = path->slots[0];
2220                 btrfs_item_key_to_cpu(l, &key, path->slots[0]);
2221
2222                 if (key.objectid != device->devid) {
2223                         btrfs_release_path(path);
2224                         break;
2225                 }
2226
2227                 dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
2228                 length = btrfs_dev_extent_length(l, dev_extent);
2229
2230                 if (key.offset + length <= new_size) {
2231                         btrfs_release_path(path);
2232                         break;
2233                 }
2234
2235                 chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
2236                 chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
2237                 chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
2238                 btrfs_release_path(path);
2239
2240                 ret = btrfs_relocate_chunk(root, chunk_tree, chunk_objectid,
2241                                            chunk_offset);
2242                 if (ret && ret != -ENOSPC)
2243                         goto done;
2244                 if (ret == -ENOSPC)
2245                         failed++;
2246                 key.offset -= 1;
2247         }
2248
2249         if (failed && !retried) {
2250                 failed = 0;
2251                 retried = true;
2252                 goto again;
2253         } else if (failed && retried) {
2254                 ret = -ENOSPC;
2255                 lock_chunks(root);
2256
2257                 device->total_bytes = old_size;
2258                 if (device->writeable)
2259                         device->fs_devices->total_rw_bytes += diff;
2260                 unlock_chunks(root);
2261                 goto done;
2262         }
2263
2264         /* Shrinking succeeded, else we would be at "done". */
2265         trans = btrfs_start_transaction(root, 0);
2266         if (IS_ERR(trans)) {
2267                 ret = PTR_ERR(trans);
2268                 goto done;
2269         }
2270
2271         lock_chunks(root);
2272
2273         device->disk_total_bytes = new_size;
2274         /* Now btrfs_update_device() will change the on-disk size. */
2275         ret = btrfs_update_device(trans, device);
2276         if (ret) {
2277                 unlock_chunks(root);
2278                 btrfs_end_transaction(trans, root);
2279                 goto done;
2280         }
2281         WARN_ON(diff > old_total);
2282         btrfs_set_super_total_bytes(super_copy, old_total - diff);
2283         unlock_chunks(root);
2284         btrfs_end_transaction(trans, root);
2285 done:
2286         btrfs_free_path(path);
2287         return ret;
2288 }
2289
2290 static int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
2291                            struct btrfs_root *root,
2292                            struct btrfs_key *key,
2293                            struct btrfs_chunk *chunk, int item_size)
2294 {
2295         struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
2296         struct btrfs_disk_key disk_key;
2297         u32 array_size;
2298         u8 *ptr;
2299
2300         array_size = btrfs_super_sys_array_size(super_copy);
2301         if (array_size + item_size > BTRFS_SYSTEM_CHUNK_ARRAY_SIZE)
2302                 return -EFBIG;
2303
2304         ptr = super_copy->sys_chunk_array + array_size;
2305         btrfs_cpu_key_to_disk(&disk_key, key);
2306         memcpy(ptr, &disk_key, sizeof(disk_key));
2307         ptr += sizeof(disk_key);
2308         memcpy(ptr, chunk, item_size);
2309         item_size += sizeof(disk_key);
2310         btrfs_set_super_sys_array_size(super_copy, array_size + item_size);
2311         return 0;
2312 }
2313
2314 /*
2315  * sort the devices in descending order by max_avail, total_avail
2316  */
2317 static int btrfs_cmp_device_info(const void *a, const void *b)
2318 {
2319         const struct btrfs_device_info *di_a = a;
2320         const struct btrfs_device_info *di_b = b;
2321
2322         if (di_a->max_avail > di_b->max_avail)
2323                 return -1;
2324         if (di_a->max_avail < di_b->max_avail)
2325                 return 1;
2326         if (di_a->total_avail > di_b->total_avail)
2327                 return -1;
2328         if (di_a->total_avail < di_b->total_avail)
2329                 return 1;
2330         return 0;
2331 }
2332
2333 static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
2334                                struct btrfs_root *extent_root,
2335                                struct map_lookup **map_ret,
2336                                u64 *num_bytes_out, u64 *stripe_size_out,
2337                                u64 start, u64 type)
2338 {
2339         struct btrfs_fs_info *info = extent_root->fs_info;
2340         struct btrfs_fs_devices *fs_devices = info->fs_devices;
2341         struct list_head *cur;
2342         struct map_lookup *map = NULL;
2343         struct extent_map_tree *em_tree;
2344         struct extent_map *em;
2345         struct btrfs_device_info *devices_info = NULL;
2346         u64 total_avail;
2347         int num_stripes;        /* total number of stripes to allocate */
2348         int sub_stripes;        /* sub_stripes info for map */
2349         int dev_stripes;        /* stripes per dev */
2350         int devs_max;           /* max devs to use */
2351         int devs_min;           /* min devs needed */
2352         int devs_increment;     /* ndevs has to be a multiple of this */
2353         int ncopies;            /* how many copies to data has */
2354         int ret;
2355         u64 max_stripe_size;
2356         u64 max_chunk_size;
2357         u64 stripe_size;
2358         u64 num_bytes;
2359         int ndevs;
2360         int i;
2361         int j;
2362
2363         if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
2364             (type & BTRFS_BLOCK_GROUP_DUP)) {
2365                 WARN_ON(1);
2366                 type &= ~BTRFS_BLOCK_GROUP_DUP;
2367         }
2368
2369         if (list_empty(&fs_devices->alloc_list))
2370                 return -ENOSPC;
2371
2372         sub_stripes = 1;
2373         dev_stripes = 1;
2374         devs_increment = 1;
2375         ncopies = 1;
2376         devs_max = 0;   /* 0 == as many as possible */
2377         devs_min = 1;
2378
2379         /*
2380          * define the properties of each RAID type.
2381          * FIXME: move this to a global table and use it in all RAID
2382          * calculation code
2383          */
2384         if (type & (BTRFS_BLOCK_GROUP_DUP)) {
2385                 dev_stripes = 2;
2386                 ncopies = 2;
2387                 devs_max = 1;
2388         } else if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
2389                 devs_min = 2;
2390         } else if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
2391                 devs_increment = 2;
2392                 ncopies = 2;
2393                 devs_max = 2;
2394                 devs_min = 2;
2395         } else if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
2396                 sub_stripes = 2;
2397                 devs_increment = 2;
2398                 ncopies = 2;
2399                 devs_min = 4;
2400         } else {
2401                 devs_max = 1;
2402         }
2403
2404         if (type & BTRFS_BLOCK_GROUP_DATA) {
2405                 max_stripe_size = 1024 * 1024 * 1024;
2406                 max_chunk_size = 10 * max_stripe_size;
2407         } else if (type & BTRFS_BLOCK_GROUP_METADATA) {
2408                 max_stripe_size = 256 * 1024 * 1024;
2409                 max_chunk_size = max_stripe_size;
2410         } else if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
2411                 max_stripe_size = 8 * 1024 * 1024;
2412                 max_chunk_size = 2 * max_stripe_size;
2413         } else {
2414                 printk(KERN_ERR "btrfs: invalid chunk type 0x%llx requested\n",
2415                        type);
2416                 BUG_ON(1);
2417         }
2418
2419         /* we don't want a chunk larger than 10% of writeable space */
2420         max_chunk_size = min(div_factor(fs_devices->total_rw_bytes, 1),
2421                              max_chunk_size);
2422
2423         devices_info = kzalloc(sizeof(*devices_info) * fs_devices->rw_devices,
2424                                GFP_NOFS);
2425         if (!devices_info)
2426                 return -ENOMEM;
2427
2428         cur = fs_devices->alloc_list.next;
2429
2430         /*
2431          * in the first pass through the devices list, we gather information
2432          * about the available holes on each device.
2433          */
2434         ndevs = 0;
2435         while (cur != &fs_devices->alloc_list) {
2436                 struct btrfs_device *device;
2437                 u64 max_avail;
2438                 u64 dev_offset;
2439
2440                 device = list_entry(cur, struct btrfs_device, dev_alloc_list);
2441
2442                 cur = cur->next;
2443
2444                 if (!device->writeable) {
2445                         printk(KERN_ERR
2446                                "btrfs: read-only device in alloc_list\n");
2447                         WARN_ON(1);
2448                         continue;
2449                 }
2450
2451                 if (!device->in_fs_metadata)
2452                         continue;
2453
2454                 if (device->total_bytes > device->bytes_used)
2455                         total_avail = device->total_bytes - device->bytes_used;
2456                 else
2457                         total_avail = 0;
2458
2459                 /* If there is no space on this device, skip it. */
2460                 if (total_avail == 0)
2461                         continue;
2462
2463                 ret = find_free_dev_extent(trans, device,
2464                                            max_stripe_size * dev_stripes,
2465                                            &dev_offset, &max_avail);
2466                 if (ret && ret != -ENOSPC)
2467                         goto error;
2468
2469                 if (ret == 0)
2470                         max_avail = max_stripe_size * dev_stripes;
2471
2472                 if (max_avail < BTRFS_STRIPE_LEN * dev_stripes)
2473                         continue;
2474
2475                 devices_info[ndevs].dev_offset = dev_offset;
2476                 devices_info[ndevs].max_avail = max_avail;
2477                 devices_info[ndevs].total_avail = total_avail;
2478                 devices_info[ndevs].dev = device;
2479                 ++ndevs;
2480         }
2481
2482         /*
2483          * now sort the devices by hole size / available space
2484          */
2485         sort(devices_info, ndevs, sizeof(struct btrfs_device_info),
2486              btrfs_cmp_device_info, NULL);
2487
2488         /* round down to number of usable stripes */
2489         ndevs -= ndevs % devs_increment;
2490
2491         if (ndevs < devs_increment * sub_stripes || ndevs < devs_min) {
2492                 ret = -ENOSPC;
2493                 goto error;
2494         }
2495
2496         if (devs_max && ndevs > devs_max)
2497                 ndevs = devs_max;
2498         /*
2499          * the primary goal is to maximize the number of stripes, so use as many
2500          * devices as possible, even if the stripes are not maximum sized.
2501          */
2502         stripe_size = devices_info[ndevs-1].max_avail;
2503         num_stripes = ndevs * dev_stripes;
2504
2505         if (stripe_size * num_stripes > max_chunk_size * ncopies) {
2506                 stripe_size = max_chunk_size * ncopies;
2507                 do_div(stripe_size, num_stripes);
2508         }
2509
2510         do_div(stripe_size, dev_stripes);
2511         do_div(stripe_size, BTRFS_STRIPE_LEN);
2512         stripe_size *= BTRFS_STRIPE_LEN;
2513
2514         map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
2515         if (!map) {
2516                 ret = -ENOMEM;
2517                 goto error;
2518         }
2519         map->num_stripes = num_stripes;
2520
2521         for (i = 0; i < ndevs; ++i) {
2522                 for (j = 0; j < dev_stripes; ++j) {
2523                         int s = i * dev_stripes + j;
2524                         map->stripes[s].dev = devices_info[i].dev;
2525                         map->stripes[s].physical = devices_info[i].dev_offset +
2526                                                    j * stripe_size;
2527                 }
2528         }
2529         map->sector_size = extent_root->sectorsize;
2530         map->stripe_len = BTRFS_STRIPE_LEN;
2531         map->io_align = BTRFS_STRIPE_LEN;
2532         map->io_width = BTRFS_STRIPE_LEN;
2533         map->type = type;
2534         map->sub_stripes = sub_stripes;
2535
2536         *map_ret = map;
2537         num_bytes = stripe_size * (num_stripes / ncopies);
2538
2539         *stripe_size_out = stripe_size;
2540         *num_bytes_out = num_bytes;
2541
2542         trace_btrfs_chunk_alloc(info->chunk_root, map, start, num_bytes);
2543
2544         em = alloc_extent_map();
2545         if (!em) {
2546                 ret = -ENOMEM;
2547                 goto error;
2548         }
2549         em->bdev = (struct block_device *)map;
2550         em->start = start;
2551         em->len = num_bytes;
2552         em->block_start = 0;
2553         em->block_len = em->len;
2554
2555         em_tree = &extent_root->fs_info->mapping_tree.map_tree;
2556         write_lock(&em_tree->lock);
2557         ret = add_extent_mapping(em_tree, em);
2558         write_unlock(&em_tree->lock);
2559         BUG_ON(ret);
2560         free_extent_map(em);
2561
2562         ret = btrfs_make_block_group(trans, extent_root, 0, type,
2563                                      BTRFS_FIRST_CHUNK_TREE_OBJECTID,
2564                                      start, num_bytes);
2565         BUG_ON(ret);
2566
2567         for (i = 0; i < map->num_stripes; ++i) {
2568                 struct btrfs_device *device;
2569                 u64 dev_offset;
2570
2571                 device = map->stripes[i].dev;
2572                 dev_offset = map->stripes[i].physical;
2573
2574                 ret = btrfs_alloc_dev_extent(trans, device,
2575                                 info->chunk_root->root_key.objectid,
2576                                 BTRFS_FIRST_CHUNK_TREE_OBJECTID,
2577                                 start, dev_offset, stripe_size);
2578                 BUG_ON(ret);
2579         }
2580
2581         kfree(devices_info);
2582         return 0;
2583
2584 error:
2585         kfree(map);
2586         kfree(devices_info);
2587         return ret;
2588 }
2589
2590 static int __finish_chunk_alloc(struct btrfs_trans_handle *trans,
2591                                 struct btrfs_root *extent_root,
2592                                 struct map_lookup *map, u64 chunk_offset,
2593                                 u64 chunk_size, u64 stripe_size)
2594 {
2595         u64 dev_offset;
2596         struct btrfs_key key;
2597         struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
2598         struct btrfs_device *device;
2599         struct btrfs_chunk *chunk;
2600         struct btrfs_stripe *stripe;
2601         size_t item_size = btrfs_chunk_item_size(map->num_stripes);
2602         int index = 0;
2603         int ret;
2604
2605         chunk = kzalloc(item_size, GFP_NOFS);
2606         if (!chunk)
2607                 return -ENOMEM;
2608
2609         index = 0;
2610         while (index < map->num_stripes) {
2611                 device = map->stripes[index].dev;
2612                 device->bytes_used += stripe_size;
2613                 ret = btrfs_update_device(trans, device);
2614                 BUG_ON(ret);
2615                 index++;
2616         }
2617
2618         index = 0;
2619         stripe = &chunk->stripe;
2620         while (index < map->num_stripes) {
2621                 device = map->stripes[index].dev;
2622                 dev_offset = map->stripes[index].physical;
2623
2624                 btrfs_set_stack_stripe_devid(stripe, device->devid);
2625                 btrfs_set_stack_stripe_offset(stripe, dev_offset);
2626                 memcpy(stripe->dev_uuid, device->uuid, BTRFS_UUID_SIZE);
2627                 stripe++;
2628                 index++;
2629         }
2630
2631         btrfs_set_stack_chunk_length(chunk, chunk_size);
2632         btrfs_set_stack_chunk_owner(chunk, extent_root->root_key.objectid);
2633         btrfs_set_stack_chunk_stripe_len(chunk, map->stripe_len);
2634         btrfs_set_stack_chunk_type(chunk, map->type);
2635         btrfs_set_stack_chunk_num_stripes(chunk, map->num_stripes);
2636         btrfs_set_stack_chunk_io_align(chunk, map->stripe_len);
2637         btrfs_set_stack_chunk_io_width(chunk, map->stripe_len);
2638         btrfs_set_stack_chunk_sector_size(chunk, extent_root->sectorsize);
2639         btrfs_set_stack_chunk_sub_stripes(chunk, map->sub_stripes);
2640
2641         key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
2642         key.type = BTRFS_CHUNK_ITEM_KEY;
2643         key.offset = chunk_offset;
2644
2645         ret = btrfs_insert_item(trans, chunk_root, &key, chunk, item_size);
2646         BUG_ON(ret);
2647
2648         if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
2649                 ret = btrfs_add_system_chunk(trans, chunk_root, &key, chunk,
2650                                              item_size);
2651                 BUG_ON(ret);
2652         }
2653
2654         kfree(chunk);
2655         return 0;
2656 }
2657
2658 /*
2659  * Chunk allocation falls into two parts. The first part does works
2660  * that make the new allocated chunk useable, but not do any operation
2661  * that modifies the chunk tree. The second part does the works that
2662  * require modifying the chunk tree. This division is important for the
2663  * bootstrap process of adding storage to a seed btrfs.
2664  */
2665 int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
2666                       struct btrfs_root *extent_root, u64 type)
2667 {
2668         u64 chunk_offset;
2669         u64 chunk_size;
2670         u64 stripe_size;
2671         struct map_lookup *map;
2672         struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
2673         int ret;
2674
2675         ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
2676                               &chunk_offset);
2677         if (ret)
2678                 return ret;
2679
2680         ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
2681                                   &stripe_size, chunk_offset, type);
2682         if (ret)
2683                 return ret;
2684
2685         ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
2686                                    chunk_size, stripe_size);
2687         BUG_ON(ret);
2688         return 0;
2689 }
2690
2691 static noinline int init_first_rw_device(struct btrfs_trans_handle *trans,
2692                                          struct btrfs_root *root,
2693                                          struct btrfs_device *device)
2694 {
2695         u64 chunk_offset;
2696         u64 sys_chunk_offset;
2697         u64 chunk_size;
2698         u64 sys_chunk_size;
2699         u64 stripe_size;
2700         u64 sys_stripe_size;
2701         u64 alloc_profile;
2702         struct map_lookup *map;
2703         struct map_lookup *sys_map;
2704         struct btrfs_fs_info *fs_info = root->fs_info;
2705         struct btrfs_root *extent_root = fs_info->extent_root;
2706         int ret;
2707
2708         ret = find_next_chunk(fs_info->chunk_root,
2709                               BTRFS_FIRST_CHUNK_TREE_OBJECTID, &chunk_offset);
2710         if (ret)
2711                 return ret;
2712
2713         alloc_profile = BTRFS_BLOCK_GROUP_METADATA |
2714                         (fs_info->metadata_alloc_profile &
2715                          fs_info->avail_metadata_alloc_bits);
2716         alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile);
2717
2718         ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size,
2719                                   &stripe_size, chunk_offset, alloc_profile);
2720         BUG_ON(ret);
2721
2722         sys_chunk_offset = chunk_offset + chunk_size;
2723
2724         alloc_profile = BTRFS_BLOCK_GROUP_SYSTEM |
2725                         (fs_info->system_alloc_profile &
2726                          fs_info->avail_system_alloc_bits);
2727         alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile);
2728
2729         ret = __btrfs_alloc_chunk(trans, extent_root, &sys_map,
2730                                   &sys_chunk_size, &sys_stripe_size,
2731                                   sys_chunk_offset, alloc_profile);
2732         BUG_ON(ret);
2733
2734         ret = btrfs_add_device(trans, fs_info->chunk_root, device);
2735         BUG_ON(ret);
2736
2737         /*
2738          * Modifying chunk tree needs allocating new blocks from both
2739          * system block group and metadata block group. So we only can
2740          * do operations require modifying the chunk tree after both
2741          * block groups were created.
2742          */
2743         ret = __finish_chunk_alloc(trans, extent_root, map, chunk_offset,
2744                                    chunk_size, stripe_size);
2745         BUG_ON(ret);
2746
2747         ret = __finish_chunk_alloc(trans, extent_root, sys_map,
2748                                    sys_chunk_offset, sys_chunk_size,
2749                                    sys_stripe_size);
2750         BUG_ON(ret);
2751         return 0;
2752 }
2753
2754 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset)
2755 {
2756         struct extent_map *em;
2757         struct map_lookup *map;
2758         struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
2759         int readonly = 0;
2760         int i;
2761
2762         read_lock(&map_tree->map_tree.lock);
2763         em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1);
2764         read_unlock(&map_tree->map_tree.lock);
2765         if (!em)
2766                 return 1;
2767
2768         if (btrfs_test_opt(root, DEGRADED)) {
2769                 free_extent_map(em);
2770                 return 0;
2771         }
2772
2773         map = (struct map_lookup *)em->bdev;
2774         for (i = 0; i < map->num_stripes; i++) {
2775                 if (!map->stripes[i].dev->writeable) {
2776                         readonly = 1;
2777                         break;
2778                 }
2779         }
2780         free_extent_map(em);
2781         return readonly;
2782 }
2783
2784 void btrfs_mapping_init(struct btrfs_mapping_tree *tree)
2785 {
2786         extent_map_tree_init(&tree->map_tree);
2787 }
2788
2789 void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree)
2790 {
2791         struct extent_map *em;
2792
2793         while (1) {
2794                 write_lock(&tree->map_tree.lock);
2795                 em = lookup_extent_mapping(&tree->map_tree, 0, (u64)-1);
2796                 if (em)
2797                         remove_extent_mapping(&tree->map_tree, em);
2798                 write_unlock(&tree->map_tree.lock);
2799                 if (!em)
2800                         break;
2801                 kfree(em->bdev);
2802                 /* once for us */
2803                 free_extent_map(em);
2804                 /* once for the tree */
2805                 free_extent_map(em);
2806         }
2807 }
2808
2809 int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len)
2810 {
2811         struct extent_map *em;
2812         struct map_lookup *map;
2813         struct extent_map_tree *em_tree = &map_tree->map_tree;
2814         int ret;
2815
2816         read_lock(&em_tree->lock);
2817         em = lookup_extent_mapping(em_tree, logical, len);
2818         read_unlock(&em_tree->lock);
2819         BUG_ON(!em);
2820
2821         BUG_ON(em->start > logical || em->start + em->len < logical);
2822         map = (struct map_lookup *)em->bdev;
2823         if (map->type & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1))
2824                 ret = map->num_stripes;
2825         else if (map->type & BTRFS_BLOCK_GROUP_RAID10)
2826                 ret = map->sub_stripes;
2827         else
2828                 ret = 1;
2829         free_extent_map(em);
2830         return ret;
2831 }
2832
2833 static int find_live_mirror(struct map_lookup *map, int first, int num,
2834                             int optimal)
2835 {
2836         int i;
2837         if (map->stripes[optimal].dev->bdev)
2838                 return optimal;
2839         for (i = first; i < first + num; i++) {
2840                 if (map->stripes[i].dev->bdev)
2841                         return i;
2842         }
2843         /* we couldn't find one that doesn't fail.  Just return something
2844          * and the io error handling code will clean up eventually
2845          */
2846         return optimal;
2847 }
2848
2849 static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
2850                              u64 logical, u64 *length,
2851                              struct btrfs_multi_bio **multi_ret,
2852                              int mirror_num)
2853 {
2854         struct extent_map *em;
2855         struct map_lookup *map;
2856         struct extent_map_tree *em_tree = &map_tree->map_tree;
2857         u64 offset;
2858         u64 stripe_offset;
2859         u64 stripe_end_offset;
2860         u64 stripe_nr;
2861         u64 stripe_nr_orig;
2862         u64 stripe_nr_end;
2863         int stripes_allocated = 8;
2864         int stripes_required = 1;
2865         int stripe_index;
2866         int i;
2867         int num_stripes;
2868         int max_errors = 0;
2869         struct btrfs_multi_bio *multi = NULL;
2870
2871         if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD)))
2872                 stripes_allocated = 1;
2873 again:
2874         if (multi_ret) {
2875                 multi = kzalloc(btrfs_multi_bio_size(stripes_allocated),
2876                                 GFP_NOFS);
2877                 if (!multi)
2878                         return -ENOMEM;
2879
2880                 atomic_set(&multi->error, 0);
2881         }
2882
2883         read_lock(&em_tree->lock);
2884         em = lookup_extent_mapping(em_tree, logical, *length);
2885         read_unlock(&em_tree->lock);
2886
2887         if (!em) {
2888                 printk(KERN_CRIT "unable to find logical %llu len %llu\n",
2889                        (unsigned long long)logical,
2890                        (unsigned long long)*length);
2891                 BUG();
2892         }
2893
2894         BUG_ON(em->start > logical || em->start + em->len < logical);
2895         map = (struct map_lookup *)em->bdev;
2896         offset = logical - em->start;
2897
2898         if (mirror_num > map->num_stripes)
2899                 mirror_num = 0;
2900
2901         /* if our multi bio struct is too small, back off and try again */
2902         if (rw & REQ_WRITE) {
2903                 if (map->type & (BTRFS_BLOCK_GROUP_RAID1 |
2904                                  BTRFS_BLOCK_GROUP_DUP)) {
2905                         stripes_required = map->num_stripes;
2906                         max_errors = 1;
2907                 } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
2908                         stripes_required = map->sub_stripes;
2909                         max_errors = 1;
2910                 }
2911         }
2912         if (rw & REQ_DISCARD) {
2913                 if (map->type & (BTRFS_BLOCK_GROUP_RAID0 |
2914                                  BTRFS_BLOCK_GROUP_RAID1 |
2915                                  BTRFS_BLOCK_GROUP_DUP |
2916                                  BTRFS_BLOCK_GROUP_RAID10)) {
2917                         stripes_required = map->num_stripes;
2918                 }
2919         }
2920         if (multi_ret && (rw & (REQ_WRITE | REQ_DISCARD)) &&
2921             stripes_allocated < stripes_required) {
2922                 stripes_allocated = map->num_stripes;
2923                 free_extent_map(em);
2924                 kfree(multi);
2925                 goto again;
2926         }
2927         stripe_nr = offset;
2928         /*
2929          * stripe_nr counts the total number of stripes we have to stride
2930          * to get to this block
2931          */
2932         do_div(stripe_nr, map->stripe_len);
2933
2934         stripe_offset = stripe_nr * map->stripe_len;
2935         BUG_ON(offset < stripe_offset);
2936
2937         /* stripe_offset is the offset of this block in its stripe*/
2938         stripe_offset = offset - stripe_offset;
2939
2940         if (rw & REQ_DISCARD)
2941                 *length = min_t(u64, em->len - offset, *length);
2942         else if (map->type & (BTRFS_BLOCK_GROUP_RAID0 |
2943                               BTRFS_BLOCK_GROUP_RAID1 |
2944                               BTRFS_BLOCK_GROUP_RAID10 |
2945                               BTRFS_BLOCK_GROUP_DUP)) {
2946                 /* we limit the length of each bio to what fits in a stripe */
2947                 *length = min_t(u64, em->len - offset,
2948                                 map->stripe_len - stripe_offset);
2949         } else {
2950                 *length = em->len - offset;
2951         }
2952
2953         if (!multi_ret)
2954                 goto out;
2955
2956         num_stripes = 1;
2957         stripe_index = 0;
2958         stripe_nr_orig = stripe_nr;
2959         stripe_nr_end = (offset + *length + map->stripe_len - 1) &
2960                         (~(map->stripe_len - 1));
2961         do_div(stripe_nr_end, map->stripe_len);
2962         stripe_end_offset = stripe_nr_end * map->stripe_len -
2963                             (offset + *length);
2964         if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
2965                 if (rw & REQ_DISCARD)
2966                         num_stripes = min_t(u64, map->num_stripes,
2967                                             stripe_nr_end - stripe_nr_orig);
2968                 stripe_index = do_div(stripe_nr, map->num_stripes);
2969         } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
2970                 if (rw & (REQ_WRITE | REQ_DISCARD))
2971                         num_stripes = map->num_stripes;
2972                 else if (mirror_num)
2973                         stripe_index = mirror_num - 1;
2974                 else {
2975                         stripe_index = find_live_mirror(map, 0,
2976                                             map->num_stripes,
2977                                             current->pid % map->num_stripes);
2978                 }
2979
2980         } else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
2981                 if (rw & (REQ_WRITE | REQ_DISCARD))
2982                         num_stripes = map->num_stripes;
2983                 else if (mirror_num)
2984                         stripe_index = mirror_num - 1;
2985
2986         } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
2987                 int factor = map->num_stripes / map->sub_stripes;
2988
2989                 stripe_index = do_div(stripe_nr, factor);
2990                 stripe_index *= map->sub_stripes;
2991
2992                 if (rw & REQ_WRITE)
2993                         num_stripes = map->sub_stripes;
2994                 else if (rw & REQ_DISCARD)
2995                         num_stripes = min_t(u64, map->sub_stripes *
2996                                             (stripe_nr_end - stripe_nr_orig),
2997                                             map->num_stripes);
2998                 else if (mirror_num)
2999                         stripe_index += mirror_num - 1;
3000                 else {
3001                         stripe_index = find_live_mirror(map, stripe_index,
3002                                               map->sub_stripes, stripe_index +
3003                                               current->pid % map->sub_stripes);
3004                 }
3005         } else {
3006                 /*
3007                  * after this do_div call, stripe_nr is the number of stripes
3008                  * on this device we have to walk to find the data, and
3009                  * stripe_index is the number of our device in the stripe array
3010                  */
3011                 stripe_index = do_div(stripe_nr, map->num_stripes);
3012         }
3013         BUG_ON(stripe_index >= map->num_stripes);
3014
3015         if (rw & REQ_DISCARD) {
3016                 for (i = 0; i < num_stripes; i++) {
3017                         multi->stripes[i].physical =
3018                                 map->stripes[stripe_index].physical +
3019                                 stripe_offset + stripe_nr * map->stripe_len;
3020                         multi->stripes[i].dev = map->stripes[stripe_index].dev;
3021
3022                         if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
3023                                 u64 stripes;
3024                                 u32 last_stripe = 0;
3025                                 int j;
3026
3027                                 div_u64_rem(stripe_nr_end - 1,
3028                                             map->num_stripes,
3029                                             &last_stripe);
3030
3031                                 for (j = 0; j < map->num_stripes; j++) {
3032                                         u32 test;
3033
3034                                         div_u64_rem(stripe_nr_end - 1 - j,
3035                                                     map->num_stripes, &test);
3036                                         if (test == stripe_index)
3037                                                 break;
3038                                 }
3039                                 stripes = stripe_nr_end - 1 - j;
3040                                 do_div(stripes, map->num_stripes);
3041                                 multi->stripes[i].length = map->stripe_len *
3042                                         (stripes - stripe_nr + 1);
3043
3044                                 if (i == 0) {
3045                                         multi->stripes[i].length -=
3046                                                 stripe_offset;
3047                                         stripe_offset = 0;
3048                                 }
3049                                 if (stripe_index == last_stripe)
3050                                         multi->stripes[i].length -=
3051                                                 stripe_end_offset;
3052                         } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
3053                                 u64 stripes;
3054                                 int j;
3055                                 int factor = map->num_stripes /
3056                                              map->sub_stripes;
3057                                 u32 last_stripe = 0;
3058
3059                                 div_u64_rem(stripe_nr_end - 1,
3060                                             factor, &last_stripe);
3061                                 last_stripe *= map->sub_stripes;
3062
3063                                 for (j = 0; j < factor; j++) {
3064                                         u32 test;
3065
3066                                         div_u64_rem(stripe_nr_end - 1 - j,
3067                                                     factor, &test);
3068
3069                                         if (test ==
3070                                             stripe_index / map->sub_stripes)
3071                                                 break;
3072                                 }
3073                                 stripes = stripe_nr_end - 1 - j;
3074                                 do_div(stripes, factor);
3075                                 multi->stripes[i].length = map->stripe_len *
3076                                         (stripes - stripe_nr + 1);
3077
3078                                 if (i < map->sub_stripes) {
3079                                         multi->stripes[i].length -=
3080                                                 stripe_offset;
3081                                         if (i == map->sub_stripes - 1)
3082                                                 stripe_offset = 0;
3083                                 }
3084                                 if (stripe_index >= last_stripe &&
3085                                     stripe_index <= (last_stripe +
3086                                                      map->sub_stripes - 1)) {
3087                                         multi->stripes[i].length -=
3088                                                 stripe_end_offset;
3089                                 }
3090                         } else
3091                                 multi->stripes[i].length = *length;
3092
3093                         stripe_index++;
3094                         if (stripe_index == map->num_stripes) {
3095                                 /* This could only happen for RAID0/10 */
3096                                 stripe_index = 0;
3097                                 stripe_nr++;
3098                         }
3099                 }
3100         } else {
3101                 for (i = 0; i < num_stripes; i++) {
3102                         multi->stripes[i].physical =
3103                                 map->stripes[stripe_index].physical +
3104                                 stripe_offset +
3105                                 stripe_nr * map->stripe_len;
3106                         multi->stripes[i].dev =
3107                                 map->stripes[stripe_index].dev;
3108                         stripe_index++;
3109                 }
3110         }
3111         if (multi_ret) {
3112                 *multi_ret = multi;
3113                 multi->num_stripes = num_stripes;
3114                 multi->max_errors = max_errors;
3115         }
3116 out:
3117         free_extent_map(em);
3118         return 0;
3119 }
3120
3121 int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
3122                       u64 logical, u64 *length,
3123                       struct btrfs_multi_bio **multi_ret, int mirror_num)
3124 {
3125         return __btrfs_map_block(map_tree, rw, logical, length, multi_ret,
3126                                  mirror_num);
3127 }
3128
3129 int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
3130                      u64 chunk_start, u64 physical, u64 devid,
3131                      u64 **logical, int *naddrs, int *stripe_len)
3132 {
3133         struct extent_map_tree *em_tree = &map_tree->map_tree;
3134         struct extent_map *em;
3135         struct map_lookup *map;
3136         u64 *buf;
3137         u64 bytenr;
3138         u64 length;
3139         u64 stripe_nr;
3140         int i, j, nr = 0;
3141
3142         read_lock(&em_tree->lock);
3143         em = lookup_extent_mapping(em_tree, chunk_start, 1);
3144         read_unlock(&em_tree->lock);
3145
3146         BUG_ON(!em || em->start != chunk_start);
3147         map = (struct map_lookup *)em->bdev;
3148
3149         length = em->len;
3150         if (map->type & BTRFS_BLOCK_GROUP_RAID10)
3151                 do_div(length, map->num_stripes / map->sub_stripes);
3152         else if (map->type & BTRFS_BLOCK_GROUP_RAID0)
3153                 do_div(length, map->num_stripes);
3154
3155         buf = kzalloc(sizeof(u64) * map->num_stripes, GFP_NOFS);
3156         BUG_ON(!buf);
3157
3158         for (i = 0; i < map->num_stripes; i++) {
3159                 if (devid && map->stripes[i].dev->devid != devid)
3160                         continue;
3161                 if (map->stripes[i].physical > physical ||
3162                     map->stripes[i].physical + length <= physical)
3163                         continue;
3164
3165                 stripe_nr = physical - map->stripes[i].physical;
3166                 do_div(stripe_nr, map->stripe_len);
3167
3168                 if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
3169                         stripe_nr = stripe_nr * map->num_stripes + i;
3170                         do_div(stripe_nr, map->sub_stripes);
3171                 } else if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
3172                         stripe_nr = stripe_nr * map->num_stripes + i;
3173                 }
3174                 bytenr = chunk_start + stripe_nr * map->stripe_len;
3175                 WARN_ON(nr >= map->num_stripes);
3176                 for (j = 0; j < nr; j++) {
3177                         if (buf[j] == bytenr)
3178                                 break;
3179                 }
3180                 if (j == nr) {
3181                         WARN_ON(nr >= map->num_stripes);
3182                         buf[nr++] = bytenr;
3183                 }
3184         }
3185
3186         *logical = buf;
3187         *naddrs = nr;
3188         *stripe_len = map->stripe_len;
3189
3190         free_extent_map(em);
3191         return 0;
3192 }
3193
3194 static void end_bio_multi_stripe(struct bio *bio, int err)
3195 {
3196         struct btrfs_multi_bio *multi = bio->bi_private;
3197         int is_orig_bio = 0;
3198
3199         if (err)
3200                 atomic_inc(&multi->error);
3201
3202         if (bio == multi->orig_bio)
3203                 is_orig_bio = 1;
3204
3205         if (atomic_dec_and_test(&multi->stripes_pending)) {
3206                 if (!is_orig_bio) {
3207                         bio_put(bio);
3208                         bio = multi->orig_bio;
3209                 }
3210                 bio->bi_private = multi->private;
3211                 bio->bi_end_io = multi->end_io;
3212                 /* only send an error to the higher layers if it is
3213                  * beyond the tolerance of the multi-bio
3214                  */
3215                 if (atomic_read(&multi->error) > multi->max_errors) {
3216                         err = -EIO;
3217                 } else if (err) {
3218                         /*
3219                          * this bio is actually up to date, we didn't
3220                          * go over the max number of errors
3221                          */
3222                         set_bit(BIO_UPTODATE, &bio->bi_flags);
3223                         err = 0;
3224                 }
3225                 kfree(multi);
3226
3227                 bio_endio(bio, err);
3228         } else if (!is_orig_bio) {
3229                 bio_put(bio);
3230         }
3231 }
3232
3233 struct async_sched {
3234         struct bio *bio;
3235         int rw;
3236         struct btrfs_fs_info *info;
3237         struct btrfs_work work;
3238 };
3239
3240 /*
3241  * see run_scheduled_bios for a description of why bios are collected for
3242  * async submit.
3243  *
3244  * This will add one bio to the pending list for a device and make sure
3245  * the work struct is scheduled.
3246  */
3247 static noinline int schedule_bio(struct btrfs_root *root,
3248                                  struct btrfs_device *device,
3249                                  int rw, struct bio *bio)
3250 {
3251         int should_queue = 1;
3252         struct btrfs_pending_bios *pending_bios;
3253
3254         /* don't bother with additional async steps for reads, right now */
3255         if (!(rw & REQ_WRITE)) {
3256                 bio_get(bio);
3257                 submit_bio(rw, bio);
3258                 bio_put(bio);
3259                 return 0;
3260         }
3261
3262         /*
3263          * nr_async_bios allows us to reliably return congestion to the
3264          * higher layers.  Otherwise, the async bio makes it appear we have
3265          * made progress against dirty pages when we've really just put it
3266          * on a queue for later
3267          */
3268         atomic_inc(&root->fs_info->nr_async_bios);
3269         WARN_ON(bio->bi_next);
3270         bio->bi_next = NULL;
3271         bio->bi_rw |= rw;
3272
3273         spin_lock(&device->io_lock);
3274         if (bio->bi_rw & REQ_SYNC)
3275                 pending_bios = &device->pending_sync_bios;
3276         else
3277                 pending_bios = &device->pending_bios;
3278
3279         if (pending_bios->tail)
3280                 pending_bios->tail->bi_next = bio;
3281
3282         pending_bios->tail = bio;
3283         if (!pending_bios->head)
3284                 pending_bios->head = bio;
3285         if (device->running_pending)
3286                 should_queue = 0;
3287
3288         spin_unlock(&device->io_lock);
3289
3290         if (should_queue)
3291                 btrfs_queue_worker(&root->fs_info->submit_workers,
3292                                    &device->work);
3293         return 0;
3294 }
3295
3296 int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
3297                   int mirror_num, int async_submit)
3298 {
3299         struct btrfs_mapping_tree *map_tree;
3300         struct btrfs_device *dev;
3301         struct bio *first_bio = bio;
3302         u64 logical = (u64)bio->bi_sector << 9;
3303         u64 length = 0;
3304         u64 map_length;
3305         struct btrfs_multi_bio *multi = NULL;
3306         int ret;
3307         int dev_nr = 0;
3308         int total_devs = 1;
3309
3310         length = bio->bi_size;
3311         map_tree = &root->fs_info->mapping_tree;
3312         map_length = length;
3313
3314         ret = btrfs_map_block(map_tree, rw, logical, &map_length, &multi,
3315                               mirror_num);
3316         BUG_ON(ret);
3317
3318         total_devs = multi->num_stripes;
3319         if (map_length < length) {
3320                 printk(KERN_CRIT "mapping failed logical %llu bio len %llu "
3321                        "len %llu\n", (unsigned long long)logical,
3322                        (unsigned long long)length,
3323                        (unsigned long long)map_length);
3324                 BUG();
3325         }
3326         multi->end_io = first_bio->bi_end_io;
3327         multi->private = first_bio->bi_private;
3328         multi->orig_bio = first_bio;
3329         atomic_set(&multi->stripes_pending, multi->num_stripes);
3330
3331         while (dev_nr < total_devs) {
3332                 if (total_devs > 1) {
3333                         if (dev_nr < total_devs - 1) {
3334                                 bio = bio_clone(first_bio, GFP_NOFS);
3335                                 BUG_ON(!bio);
3336                         } else {
3337                                 bio = first_bio;
3338                         }
3339                         bio->bi_private = multi;
3340                         bio->bi_end_io = end_bio_multi_stripe;
3341                 }
3342                 bio->bi_sector = multi->stripes[dev_nr].physical >> 9;
3343                 dev = multi->stripes[dev_nr].dev;
3344                 if (dev && dev->bdev && (rw != WRITE || dev->writeable)) {
3345                         bio->bi_bdev = dev->bdev;
3346                         if (async_submit)
3347                                 schedule_bio(root, dev, rw, bio);
3348                         else
3349                                 submit_bio(rw, bio);
3350                 } else {
3351                         bio->bi_bdev = root->fs_info->fs_devices->latest_bdev;
3352                         bio->bi_sector = logical >> 9;
3353                         bio_endio(bio, -EIO);
3354                 }
3355                 dev_nr++;
3356         }
3357         if (total_devs == 1)
3358                 kfree(multi);
3359         return 0;
3360 }
3361
3362 struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
3363                                        u8 *uuid, u8 *fsid)
3364 {
3365         struct btrfs_device *device;
3366         struct btrfs_fs_devices *cur_devices;
3367
3368         cur_devices = root->fs_info->fs_devices;
3369         while (cur_devices) {
3370                 if (!fsid ||
3371                     !memcmp(cur_devices->fsid, fsid, BTRFS_UUID_SIZE)) {
3372                         device = __find_device(&cur_devices->devices,
3373                                                devid, uuid);
3374                         if (device)
3375                                 return device;
3376                 }
3377                 cur_devices = cur_devices->seed;
3378         }
3379         return NULL;
3380 }
3381
3382 static struct btrfs_device *add_missing_dev(struct btrfs_root *root,
3383                                             u64 devid, u8 *dev_uuid)
3384 {
3385         struct btrfs_device *device;
3386         struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
3387
3388         device = kzalloc(sizeof(*device), GFP_NOFS);
3389         if (!device)
3390                 return NULL;
3391         list_add(&device->dev_list,
3392                  &fs_devices->devices);
3393         device->dev_root = root->fs_info->dev_root;
3394         device->devid = devid;
3395         device->work.func = pending_bios_fn;
3396         device->fs_devices = fs_devices;
3397         device->missing = 1;
3398         fs_devices->num_devices++;
3399         fs_devices->missing_devices++;
3400         spin_lock_init(&device->io_lock);
3401         INIT_LIST_HEAD(&device->dev_alloc_list);
3402         memcpy(device->uuid, dev_uuid, BTRFS_UUID_SIZE);
3403         return device;
3404 }
3405
3406 static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
3407                           struct extent_buffer *leaf,
3408                           struct btrfs_chunk *chunk)
3409 {
3410         struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
3411         struct map_lookup *map;
3412         struct extent_map *em;
3413         u64 logical;
3414         u64 length;
3415         u64 devid;
3416         u8 uuid[BTRFS_UUID_SIZE];
3417         int num_stripes;
3418         int ret;
3419         int i;
3420
3421         logical = key->offset;
3422         length = btrfs_chunk_length(leaf, chunk);
3423
3424         read_lock(&map_tree->map_tree.lock);
3425         em = lookup_extent_mapping(&map_tree->map_tree, logical, 1);
3426         read_unlock(&map_tree->map_tree.lock);
3427
3428         /* already mapped? */
3429         if (em && em->start <= logical && em->start + em->len > logical) {
3430                 free_extent_map(em);
3431                 return 0;
3432         } else if (em) {
3433                 free_extent_map(em);
3434         }
3435
3436         em = alloc_extent_map();
3437         if (!em)
3438                 return -ENOMEM;
3439         num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
3440         map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
3441         if (!map) {
3442                 free_extent_map(em);
3443                 return -ENOMEM;
3444         }
3445
3446         em->bdev = (struct block_device *)map;
3447         em->start = logical;
3448         em->len = length;
3449         em->block_start = 0;
3450         em->block_len = em->len;
3451
3452         map->num_stripes = num_stripes;
3453         map->io_width = btrfs_chunk_io_width(leaf, chunk);
3454         map->io_align = btrfs_chunk_io_align(leaf, chunk);
3455         map->sector_size = btrfs_chunk_sector_size(leaf, chunk);
3456         map->stripe_len = btrfs_chunk_stripe_len(leaf, chunk);
3457         map->type = btrfs_chunk_type(leaf, chunk);
3458         map->sub_stripes = btrfs_chunk_sub_stripes(leaf, chunk);
3459         for (i = 0; i < num_stripes; i++) {
3460                 map->stripes[i].physical =
3461                         btrfs_stripe_offset_nr(leaf, chunk, i);
3462                 devid = btrfs_stripe_devid_nr(leaf, chunk, i);
3463                 read_extent_buffer(leaf, uuid, (unsigned long)
3464                                    btrfs_stripe_dev_uuid_nr(chunk, i),
3465                                    BTRFS_UUID_SIZE);
3466                 map->stripes[i].dev = btrfs_find_device(root, devid, uuid,
3467                                                         NULL);
3468                 if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) {
3469                         kfree(map);
3470                         free_extent_map(em);
3471                         return -EIO;
3472                 }
3473                 if (!map->stripes[i].dev) {
3474                         map->stripes[i].dev =
3475                                 add_missing_dev(root, devid, uuid);
3476                         if (!map->stripes[i].dev) {
3477                                 kfree(map);
3478                                 free_extent_map(em);
3479                                 return -EIO;
3480                         }
3481                 }
3482                 map->stripes[i].dev->in_fs_metadata = 1;
3483         }
3484
3485         write_lock(&map_tree->map_tree.lock);
3486         ret = add_extent_mapping(&map_tree->map_tree, em);
3487         write_unlock(&map_tree->map_tree.lock);
3488         BUG_ON(ret);
3489         free_extent_map(em);
3490
3491         return 0;
3492 }
3493
3494 static int fill_device_from_item(struct extent_buffer *leaf,
3495                                  struct btrfs_dev_item *dev_item,
3496                                  struct btrfs_device *device)
3497 {
3498         unsigned long ptr;
3499
3500         device->devid = btrfs_device_id(leaf, dev_item);
3501         device->disk_total_bytes = btrfs_device_total_bytes(leaf, dev_item);
3502         device->total_bytes = device->disk_total_bytes;
3503         device->bytes_used = btrfs_device_bytes_used(leaf, dev_item);
3504         device->type = btrfs_device_type(leaf, dev_item);
3505         device->io_align = btrfs_device_io_align(leaf, dev_item);
3506         device->io_width = btrfs_device_io_width(leaf, dev_item);
3507         device->sector_size = btrfs_device_sector_size(leaf, dev_item);
3508
3509         ptr = (unsigned long)btrfs_device_uuid(dev_item);
3510         read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
3511
3512         return 0;
3513 }
3514
3515 static int open_seed_devices(struct btrfs_root *root, u8 *fsid)
3516 {
3517         struct btrfs_fs_devices *fs_devices;
3518         int ret;
3519
3520         mutex_lock(&uuid_mutex);
3521
3522         fs_devices = root->fs_info->fs_devices->seed;
3523         while (fs_devices) {
3524                 if (!memcmp(fs_devices->fsid, fsid, BTRFS_UUID_SIZE)) {
3525                         ret = 0;
3526                         goto out;
3527                 }
3528                 fs_devices = fs_devices->seed;
3529         }
3530
3531         fs_devices = find_fsid(fsid);
3532         if (!fs_devices) {
3533                 ret = -ENOENT;
3534                 goto out;
3535         }
3536
3537         fs_devices = clone_fs_devices(fs_devices);
3538         if (IS_ERR(fs_devices)) {
3539                 ret = PTR_ERR(fs_devices);
3540                 goto out;
3541         }
3542
3543         ret = __btrfs_open_devices(fs_devices, FMODE_READ,
3544                                    root->fs_info->bdev_holder);
3545         if (ret)
3546                 goto out;
3547
3548         if (!fs_devices->seeding) {
3549                 __btrfs_close_devices(fs_devices);
3550                 free_fs_devices(fs_devices);
3551                 ret = -EINVAL;
3552                 goto out;
3553         }
3554
3555         fs_devices->seed = root->fs_info->fs_devices->seed;
3556         root->fs_info->fs_devices->seed = fs_devices;
3557 out:
3558         mutex_unlock(&uuid_mutex);
3559         return ret;
3560 }
3561
3562 static int read_one_dev(struct btrfs_root *root,
3563                         struct extent_buffer *leaf,
3564                         struct btrfs_dev_item *dev_item)
3565 {
3566         struct btrfs_device *device;
3567         u64 devid;
3568         int ret;
3569         u8 fs_uuid[BTRFS_UUID_SIZE];
3570         u8 dev_uuid[BTRFS_UUID_SIZE];
3571
3572         devid = btrfs_device_id(leaf, dev_item);
3573         read_extent_buffer(leaf, dev_uuid,
3574                            (unsigned long)btrfs_device_uuid(dev_item),
3575                            BTRFS_UUID_SIZE);
3576         read_extent_buffer(leaf, fs_uuid,
3577                            (unsigned long)btrfs_device_fsid(dev_item),
3578                            BTRFS_UUID_SIZE);
3579
3580         if (memcmp(fs_uuid, root->fs_info->fsid, BTRFS_UUID_SIZE)) {
3581                 ret = open_seed_devices(root, fs_uuid);
3582                 if (ret && !btrfs_test_opt(root, DEGRADED))
3583                         return ret;
3584         }
3585
3586         device = btrfs_find_device(root, devid, dev_uuid, fs_uuid);
3587         if (!device || !device->bdev) {
3588                 if (!btrfs_test_opt(root, DEGRADED))
3589                         return -EIO;
3590
3591                 if (!device) {
3592                         printk(KERN_WARNING "warning devid %llu missing\n",
3593                                (unsigned long long)devid);
3594                         device = add_missing_dev(root, devid, dev_uuid);
3595                         if (!device)
3596                                 return -ENOMEM;
3597                 } else if (!device->missing) {
3598                         /*
3599                          * this happens when a device that was properly setup
3600                          * in the device info lists suddenly goes bad.
3601                          * device->bdev is NULL, and so we have to set
3602                          * device->missing to one here
3603                          */
3604                         root->fs_info->fs_devices->missing_devices++;
3605                         device->missing = 1;
3606                 }
3607         }
3608
3609         if (device->fs_devices != root->fs_info->fs_devices) {
3610                 BUG_ON(device->writeable);
3611                 if (device->generation !=
3612                     btrfs_device_generation(leaf, dev_item))
3613                         return -EINVAL;
3614         }
3615
3616         fill_device_from_item(leaf, dev_item, device);
3617         device->dev_root = root->fs_info->dev_root;
3618         device->in_fs_metadata = 1;
3619         if (device->writeable)
3620                 device->fs_devices->total_rw_bytes += device->total_bytes;
3621         ret = 0;
3622         return ret;
3623 }
3624
3625 int btrfs_read_sys_array(struct btrfs_root *root)
3626 {
3627         struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
3628         struct extent_buffer *sb;
3629         struct btrfs_disk_key *disk_key;
3630         struct btrfs_chunk *chunk;
3631         u8 *ptr;
3632         unsigned long sb_ptr;
3633         int ret = 0;
3634         u32 num_stripes;
3635         u32 array_size;
3636         u32 len = 0;
3637         u32 cur;
3638         struct btrfs_key key;
3639
3640         sb = btrfs_find_create_tree_block(root, BTRFS_SUPER_INFO_OFFSET,
3641                                           BTRFS_SUPER_INFO_SIZE);
3642         if (!sb)
3643                 return -ENOMEM;
3644         btrfs_set_buffer_uptodate(sb);
3645         btrfs_set_buffer_lockdep_class(root->root_key.objectid, sb, 0);
3646
3647         write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE);
3648         array_size = btrfs_super_sys_array_size(super_copy);
3649
3650         ptr = super_copy->sys_chunk_array;
3651         sb_ptr = offsetof(struct btrfs_super_block, sys_chunk_array);
3652         cur = 0;
3653
3654         while (cur < array_size) {
3655                 disk_key = (struct btrfs_disk_key *)ptr;
3656                 btrfs_disk_key_to_cpu(&key, disk_key);
3657
3658                 len = sizeof(*disk_key); ptr += len;
3659                 sb_ptr += len;
3660                 cur += len;
3661
3662                 if (key.type == BTRFS_CHUNK_ITEM_KEY) {
3663                         chunk = (struct btrfs_chunk *)sb_ptr;
3664                         ret = read_one_chunk(root, &key, sb, chunk);
3665                         if (ret)
3666                                 break;
3667                         num_stripes = btrfs_chunk_num_stripes(sb, chunk);
3668                         len = btrfs_chunk_item_size(num_stripes);
3669                 } else {
3670                         ret = -EIO;
3671                         break;
3672                 }
3673                 ptr += len;
3674                 sb_ptr += len;
3675                 cur += len;
3676         }
3677         free_extent_buffer(sb);
3678         return ret;
3679 }
3680
3681 int btrfs_read_chunk_tree(struct btrfs_root *root)
3682 {
3683         struct btrfs_path *path;
3684         struct extent_buffer *leaf;
3685         struct btrfs_key key;
3686         struct btrfs_key found_key;
3687         int ret;
3688         int slot;
3689
3690         root = root->fs_info->chunk_root;
3691
3692         path = btrfs_alloc_path();
3693         if (!path)
3694                 return -ENOMEM;
3695
3696         /* first we search for all of the device items, and then we
3697          * read in all of the chunk items.  This way we can create chunk
3698          * mappings that reference all of the devices that are afound
3699          */
3700         key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
3701         key.offset = 0;
3702         key.type = 0;
3703 again:
3704         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
3705         if (ret < 0)
3706                 goto error;
3707         while (1) {
3708                 leaf = path->nodes[0];
3709                 slot = path->slots[0];
3710                 if (slot >= btrfs_header_nritems(leaf)) {
3711                         ret = btrfs_next_leaf(root, path);
3712                         if (ret == 0)
3713                                 continue;
3714                         if (ret < 0)
3715                                 goto error;
3716                         break;
3717                 }
3718                 btrfs_item_key_to_cpu(leaf, &found_key, slot);
3719                 if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
3720                         if (found_key.objectid != BTRFS_DEV_ITEMS_OBJECTID)
3721                                 break;
3722                         if (found_key.type == BTRFS_DEV_ITEM_KEY) {
3723                                 struct btrfs_dev_item *dev_item;
3724                                 dev_item = btrfs_item_ptr(leaf, slot,
3725                                                   struct btrfs_dev_item);
3726                                 ret = read_one_dev(root, leaf, dev_item);
3727                                 if (ret)
3728                                         goto error;
3729                         }
3730                 } else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) {
3731                         struct btrfs_chunk *chunk;
3732                         chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
3733                         ret = read_one_chunk(root, &found_key, leaf, chunk);
3734                         if (ret)
3735                                 goto error;
3736                 }
3737                 path->slots[0]++;
3738         }
3739         if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
3740                 key.objectid = 0;
3741                 btrfs_release_path(path);
3742                 goto again;
3743         }
3744         ret = 0;
3745 error:
3746         btrfs_free_path(path);
3747         return ret;
3748 }