volumes.c 61.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */
#include <linux/sched.h>
#include <linux/bio.h>
20
#include <linux/buffer_head.h>
21
#include <linux/blkdev.h>
22
#include <linux/random.h>
23
#include <asm/div64.h>
24 25 26 27 28 29
#include "ctree.h"
#include "extent_map.h"
#include "disk-io.h"
#include "transaction.h"
#include "print-tree.h"
#include "volumes.h"
30
#include "async-thread.h"
31

32 33 34 35 36 37 38
struct map_lookup {
	u64 type;
	int io_align;
	int io_width;
	int stripe_len;
	int sector_size;
	int num_stripes;
C
Chris Mason 已提交
39
	int sub_stripes;
40
	struct btrfs_bio_stripe stripes[];
41 42 43
};

#define map_lookup_size(n) (sizeof(struct map_lookup) + \
44
			    (sizeof(struct btrfs_bio_stripe) * (n)))
45

46 47 48
static DEFINE_MUTEX(uuid_mutex);
static LIST_HEAD(fs_uuids);

49 50 51 52 53 54 55 56 57 58
void btrfs_lock_volumes(void)
{
	mutex_lock(&uuid_mutex);
}

void btrfs_unlock_volumes(void)
{
	mutex_unlock(&uuid_mutex);
}

59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74
int btrfs_cleanup_fs_uuids(void)
{
	struct btrfs_fs_devices *fs_devices;
	struct list_head *uuid_cur;
	struct list_head *devices_cur;
	struct btrfs_device *dev;

	list_for_each(uuid_cur, &fs_uuids) {
		fs_devices = list_entry(uuid_cur, struct btrfs_fs_devices,
					list);
		while(!list_empty(&fs_devices->devices)) {
			devices_cur = fs_devices->devices.next;
			dev = list_entry(devices_cur, struct btrfs_device,
					 dev_list);
			if (dev->bdev) {
				close_bdev_excl(dev->bdev);
75
				fs_devices->open_devices--;
76 77
			}
			list_del(&dev->dev_list);
78
			kfree(dev->name);
79 80 81 82 83 84
			kfree(dev);
		}
	}
	return 0;
}

85 86
static struct btrfs_device *__find_device(struct list_head *head, u64 devid,
					  u8 *uuid)
87 88 89 90 91 92
{
	struct btrfs_device *dev;
	struct list_head *cur;

	list_for_each(cur, head) {
		dev = list_entry(cur, struct btrfs_device, dev_list);
93
		if (dev->devid == devid &&
94
		    (!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
95
			return dev;
96
		}
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
	}
	return NULL;
}

static struct btrfs_fs_devices *find_fsid(u8 *fsid)
{
	struct list_head *cur;
	struct btrfs_fs_devices *fs_devices;

	list_for_each(cur, &fs_uuids) {
		fs_devices = list_entry(cur, struct btrfs_fs_devices, list);
		if (memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE) == 0)
			return fs_devices;
	}
	return NULL;
}

114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208
/*
 * we try to collect pending bios for a device so we don't get a large
 * number of procs sending bios down to the same device.  This greatly
 * improves the schedulers ability to collect and merge the bios.
 *
 * But, it also turns into a long list of bios to process and that is sure
 * to eventually make the worker thread block.  The solution here is to
 * make some progress and then put this work struct back at the end of
 * the list if the block device is congested.  This way, multiple devices
 * can make progress from a single worker thread.
 */
int run_scheduled_bios(struct btrfs_device *device)
{
	struct bio *pending;
	struct backing_dev_info *bdi;
	struct bio *tail;
	struct bio *cur;
	int again = 0;
	unsigned long num_run = 0;

	bdi = device->bdev->bd_inode->i_mapping->backing_dev_info;
loop:
	spin_lock(&device->io_lock);

	/* take all the bios off the list at once and process them
	 * later on (without the lock held).  But, remember the
	 * tail and other pointers so the bios can be properly reinserted
	 * into the list if we hit congestion
	 */
	pending = device->pending_bios;
	tail = device->pending_bio_tail;
	WARN_ON(pending && !tail);
	device->pending_bios = NULL;
	device->pending_bio_tail = NULL;

	/*
	 * if pending was null this time around, no bios need processing
	 * at all and we can stop.  Otherwise it'll loop back up again
	 * and do an additional check so no bios are missed.
	 *
	 * device->running_pending is used to synchronize with the
	 * schedule_bio code.
	 */
	if (pending) {
		again = 1;
		device->running_pending = 1;
	} else {
		again = 0;
		device->running_pending = 0;
	}
	spin_unlock(&device->io_lock);

	while(pending) {
		cur = pending;
		pending = pending->bi_next;
		cur->bi_next = NULL;
		atomic_dec(&device->dev_root->fs_info->nr_async_submits);
		submit_bio(cur->bi_rw, cur);
		num_run++;

		/*
		 * we made progress, there is more work to do and the bdi
		 * is now congested.  Back off and let other work structs
		 * run instead
		 */
		if (pending && num_run && bdi_write_congested(bdi)) {
			struct bio *old_head;

			spin_lock(&device->io_lock);
			old_head = device->pending_bios;
			device->pending_bios = pending;
			if (device->pending_bio_tail)
				tail->bi_next = old_head;
			else
				device->pending_bio_tail = tail;

			spin_unlock(&device->io_lock);
			btrfs_requeue_work(&device->work);
			goto done;
		}
	}
	if (again)
		goto loop;
done:
	return 0;
}

void pending_bios_fn(struct btrfs_work *work)
{
	struct btrfs_device *device;

	device = container_of(work, struct btrfs_device, work);
	run_scheduled_bios(device);
}

209 210 211 212 213 214 215 216 217 218
static int device_list_add(const char *path,
			   struct btrfs_super_block *disk_super,
			   u64 devid, struct btrfs_fs_devices **fs_devices_ret)
{
	struct btrfs_device *device;
	struct btrfs_fs_devices *fs_devices;
	u64 found_transid = btrfs_super_generation(disk_super);

	fs_devices = find_fsid(disk_super->fsid);
	if (!fs_devices) {
219
		fs_devices = kzalloc(sizeof(*fs_devices), GFP_NOFS);
220 221 222
		if (!fs_devices)
			return -ENOMEM;
		INIT_LIST_HEAD(&fs_devices->devices);
223
		INIT_LIST_HEAD(&fs_devices->alloc_list);
224 225 226 227 228 229
		list_add(&fs_devices->list, &fs_uuids);
		memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
		fs_devices->latest_devid = devid;
		fs_devices->latest_trans = found_transid;
		device = NULL;
	} else {
230 231
		device = __find_device(&fs_devices->devices, devid,
				       disk_super->dev_item.uuid);
232 233 234 235 236 237 238 239
	}
	if (!device) {
		device = kzalloc(sizeof(*device), GFP_NOFS);
		if (!device) {
			/* we can safely leave the fs_devices entry around */
			return -ENOMEM;
		}
		device->devid = devid;
240
		device->work.func = pending_bios_fn;
241 242
		memcpy(device->uuid, disk_super->dev_item.uuid,
		       BTRFS_UUID_SIZE);
243
		device->barriers = 1;
244
		spin_lock_init(&device->io_lock);
245 246 247 248 249 250
		device->name = kstrdup(path, GFP_NOFS);
		if (!device->name) {
			kfree(device);
			return -ENOMEM;
		}
		list_add(&device->dev_list, &fs_devices->devices);
251
		list_add(&device->dev_alloc_list, &fs_devices->alloc_list);
252 253 254 255 256 257 258 259 260 261 262
		fs_devices->num_devices++;
	}

	if (found_transid > fs_devices->latest_trans) {
		fs_devices->latest_devid = devid;
		fs_devices->latest_trans = found_transid;
	}
	*fs_devices_ret = fs_devices;
	return 0;
}

263 264 265 266 267 268 269 270 271 272 273
int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
{
	struct list_head *head = &fs_devices->devices;
	struct list_head *cur;
	struct btrfs_device *device;

	mutex_lock(&uuid_mutex);
again:
	list_for_each(cur, head) {
		device = list_entry(cur, struct btrfs_device, dev_list);
		if (!device->in_fs_metadata) {
274
			if (device->bdev) {
275
				close_bdev_excl(device->bdev);
276 277
				fs_devices->open_devices--;
			}
278 279 280 281 282 283 284 285 286 287 288
			list_del(&device->dev_list);
			list_del(&device->dev_alloc_list);
			fs_devices->num_devices--;
			kfree(device->name);
			kfree(device);
			goto again;
		}
	}
	mutex_unlock(&uuid_mutex);
	return 0;
}
289

290 291 292 293 294 295 296 297 298 299 300
int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
{
	struct list_head *head = &fs_devices->devices;
	struct list_head *cur;
	struct btrfs_device *device;

	mutex_lock(&uuid_mutex);
	list_for_each(cur, head) {
		device = list_entry(cur, struct btrfs_device, dev_list);
		if (device->bdev) {
			close_bdev_excl(device->bdev);
301
			fs_devices->open_devices--;
302 303
		}
		device->bdev = NULL;
304
		device->in_fs_metadata = 0;
305
	}
306
	fs_devices->mounted = 0;
307 308 309 310 311 312 313 314 315 316 317
	mutex_unlock(&uuid_mutex);
	return 0;
}

int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
		       int flags, void *holder)
{
	struct block_device *bdev;
	struct list_head *head = &fs_devices->devices;
	struct list_head *cur;
	struct btrfs_device *device;
318 319 320 321 322 323 324 325
	struct block_device *latest_bdev = NULL;
	struct buffer_head *bh;
	struct btrfs_super_block *disk_super;
	u64 latest_devid = 0;
	u64 latest_transid = 0;
	u64 transid;
	u64 devid;
	int ret = 0;
326 327

	mutex_lock(&uuid_mutex);
328 329 330
	if (fs_devices->mounted)
		goto out;

331 332
	list_for_each(cur, head) {
		device = list_entry(cur, struct btrfs_device, dev_list);
333 334 335
		if (device->bdev)
			continue;

336 337 338
		if (!device->name)
			continue;

339
		bdev = open_bdev_excl(device->name, flags, holder);
340

341 342
		if (IS_ERR(bdev)) {
			printk("open %s failed\n", device->name);
343
			goto error;
344
		}
345
		set_blocksize(bdev, 4096);
346 347 348 349 350 351 352 353 354 355 356 357 358 359 360

		bh = __bread(bdev, BTRFS_SUPER_INFO_OFFSET / 4096, 4096);
		if (!bh)
			goto error_close;

		disk_super = (struct btrfs_super_block *)bh->b_data;
		if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
		    sizeof(disk_super->magic)))
			goto error_brelse;

		devid = le64_to_cpu(disk_super->dev_item.devid);
		if (devid != device->devid)
			goto error_brelse;

		transid = btrfs_super_generation(disk_super);
361
		if (!latest_transid || transid > latest_transid) {
362 363 364 365 366
			latest_devid = devid;
			latest_transid = transid;
			latest_bdev = bdev;
		}

367
		device->bdev = bdev;
368
		device->in_fs_metadata = 0;
369 370
		fs_devices->open_devices++;
		continue;
371

372 373 374 375 376 377
error_brelse:
		brelse(bh);
error_close:
		close_bdev_excl(bdev);
error:
		continue;
378
	}
379 380 381 382 383 384 385 386 387
	if (fs_devices->open_devices == 0) {
		ret = -EIO;
		goto out;
	}
	fs_devices->mounted = 1;
	fs_devices->latest_bdev = latest_bdev;
	fs_devices->latest_devid = latest_devid;
	fs_devices->latest_trans = latest_transid;
out:
388 389 390 391 392 393 394 395 396 397 398 399
	mutex_unlock(&uuid_mutex);
	return ret;
}

int btrfs_scan_one_device(const char *path, int flags, void *holder,
			  struct btrfs_fs_devices **fs_devices_ret)
{
	struct btrfs_super_block *disk_super;
	struct block_device *bdev;
	struct buffer_head *bh;
	int ret;
	u64 devid;
400
	u64 transid;
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421

	mutex_lock(&uuid_mutex);

	bdev = open_bdev_excl(path, flags, holder);

	if (IS_ERR(bdev)) {
		ret = PTR_ERR(bdev);
		goto error;
	}

	ret = set_blocksize(bdev, 4096);
	if (ret)
		goto error_close;
	bh = __bread(bdev, BTRFS_SUPER_INFO_OFFSET / 4096, 4096);
	if (!bh) {
		ret = -EIO;
		goto error_close;
	}
	disk_super = (struct btrfs_super_block *)bh->b_data;
	if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
	    sizeof(disk_super->magic))) {
422
		ret = -EINVAL;
423 424 425
		goto error_brelse;
	}
	devid = le64_to_cpu(disk_super->dev_item.devid);
426
	transid = btrfs_super_generation(disk_super);
427 428 429 430 431 432 433 434 435
	if (disk_super->label[0])
		printk("device label %s ", disk_super->label);
	else {
		/* FIXME, make a readl uuid parser */
		printk("device fsid %llx-%llx ",
		       *(unsigned long long *)disk_super->fsid,
		       *(unsigned long long *)(disk_super->fsid + 8));
	}
	printk("devid %Lu transid %Lu %s\n", devid, transid, path);
436 437 438 439 440 441 442 443 444 445
	ret = device_list_add(path, disk_super, devid, fs_devices_ret);

error_brelse:
	brelse(bh);
error_close:
	close_bdev_excl(bdev);
error:
	mutex_unlock(&uuid_mutex);
	return ret;
}
446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473

/*
 * this uses a pretty simple search, the expectation is that it is
 * called very infrequently and that a given device has a small number
 * of extents
 */
static int find_free_dev_extent(struct btrfs_trans_handle *trans,
				struct btrfs_device *device,
				struct btrfs_path *path,
				u64 num_bytes, u64 *start)
{
	struct btrfs_key key;
	struct btrfs_root *root = device->dev_root;
	struct btrfs_dev_extent *dev_extent = NULL;
	u64 hole_size = 0;
	u64 last_byte = 0;
	u64 search_start = 0;
	u64 search_end = device->total_bytes;
	int ret;
	int slot = 0;
	int start_found;
	struct extent_buffer *l;

	start_found = 0;
	path->reada = 2;

	/* FIXME use last free of some kind */

474 475 476 477
	/* we don't want to overwrite the superblock on the drive,
	 * so we make sure to start at an offset of at least 1MB
	 */
	search_start = max((u64)1024 * 1024, search_start);
478 479 480 481

	if (root->fs_info->alloc_start + num_bytes <= device->total_bytes)
		search_start = max(root->fs_info->alloc_start, search_start);

482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556
	key.objectid = device->devid;
	key.offset = search_start;
	key.type = BTRFS_DEV_EXTENT_KEY;
	ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
	if (ret < 0)
		goto error;
	ret = btrfs_previous_item(root, path, 0, key.type);
	if (ret < 0)
		goto error;
	l = path->nodes[0];
	btrfs_item_key_to_cpu(l, &key, path->slots[0]);
	while (1) {
		l = path->nodes[0];
		slot = path->slots[0];
		if (slot >= btrfs_header_nritems(l)) {
			ret = btrfs_next_leaf(root, path);
			if (ret == 0)
				continue;
			if (ret < 0)
				goto error;
no_more_items:
			if (!start_found) {
				if (search_start >= search_end) {
					ret = -ENOSPC;
					goto error;
				}
				*start = search_start;
				start_found = 1;
				goto check_pending;
			}
			*start = last_byte > search_start ?
				last_byte : search_start;
			if (search_end <= *start) {
				ret = -ENOSPC;
				goto error;
			}
			goto check_pending;
		}
		btrfs_item_key_to_cpu(l, &key, slot);

		if (key.objectid < device->devid)
			goto next;

		if (key.objectid > device->devid)
			goto no_more_items;

		if (key.offset >= search_start && key.offset > last_byte &&
		    start_found) {
			if (last_byte < search_start)
				last_byte = search_start;
			hole_size = key.offset - last_byte;
			if (key.offset > last_byte &&
			    hole_size >= num_bytes) {
				*start = last_byte;
				goto check_pending;
			}
		}
		if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY) {
			goto next;
		}

		start_found = 1;
		dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
		last_byte = key.offset + btrfs_dev_extent_length(l, dev_extent);
next:
		path->slots[0]++;
		cond_resched();
	}
check_pending:
	/* we have to make sure we didn't find an extent that has already
	 * been allocated by the map tree or the original allocation
	 */
	btrfs_release_path(root, path);
	BUG_ON(*start < search_start);

557
	if (*start + num_bytes > search_end) {
558 559 560 561 562 563 564 565 566 567 568
		ret = -ENOSPC;
		goto error;
	}
	/* check for pending inserts here */
	return 0;

error:
	btrfs_release_path(root, path);
	return ret;
}

569 570 571 572 573 574 575 576
int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
			  struct btrfs_device *device,
			  u64 start)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_root *root = device->dev_root;
	struct btrfs_key key;
577 578 579
	struct btrfs_key found_key;
	struct extent_buffer *leaf = NULL;
	struct btrfs_dev_extent *extent = NULL;
580 581 582 583 584 585 586 587 588 589

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	key.objectid = device->devid;
	key.offset = start;
	key.type = BTRFS_DEV_EXTENT_KEY;

	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605
	if (ret > 0) {
		ret = btrfs_previous_item(root, path, key.objectid,
					  BTRFS_DEV_EXTENT_KEY);
		BUG_ON(ret);
		leaf = path->nodes[0];
		btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
		extent = btrfs_item_ptr(leaf, path->slots[0],
					struct btrfs_dev_extent);
		BUG_ON(found_key.offset > start || found_key.offset +
		       btrfs_dev_extent_length(leaf, extent) < start);
		ret = 0;
	} else if (ret == 0) {
		leaf = path->nodes[0];
		extent = btrfs_item_ptr(leaf, path->slots[0],
					struct btrfs_dev_extent);
	}
606 607
	BUG_ON(ret);

608 609
	if (device->bytes_used > 0)
		device->bytes_used -= btrfs_dev_extent_length(leaf, extent);
610 611 612 613 614 615 616
	ret = btrfs_del_item(trans, root, path);
	BUG_ON(ret);

	btrfs_free_path(path);
	return ret;
}

617 618
int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
			   struct btrfs_device *device,
619 620 621
			   u64 chunk_tree, u64 chunk_objectid,
			   u64 chunk_offset,
			   u64 num_bytes, u64 *start)
622 623 624 625 626 627 628 629
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_root *root = device->dev_root;
	struct btrfs_dev_extent *extent;
	struct extent_buffer *leaf;
	struct btrfs_key key;

630
	WARN_ON(!device->in_fs_metadata);
631 632 633 634 635
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	ret = find_free_dev_extent(trans, device, path, num_bytes, start);
636
	if (ret) {
637
		goto err;
638
	}
639 640 641 642 643 644 645 646 647 648 649

	key.objectid = device->devid;
	key.offset = *start;
	key.type = BTRFS_DEV_EXTENT_KEY;
	ret = btrfs_insert_empty_item(trans, root, path, &key,
				      sizeof(*extent));
	BUG_ON(ret);

	leaf = path->nodes[0];
	extent = btrfs_item_ptr(leaf, path->slots[0],
				struct btrfs_dev_extent);
650 651 652 653 654 655 656 657
	btrfs_set_dev_extent_chunk_tree(leaf, extent, chunk_tree);
	btrfs_set_dev_extent_chunk_objectid(leaf, extent, chunk_objectid);
	btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);

	write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
		    (unsigned long)btrfs_dev_extent_chunk_tree_uuid(extent),
		    BTRFS_UUID_SIZE);

658 659 660 661 662 663 664
	btrfs_set_dev_extent_length(leaf, extent, num_bytes);
	btrfs_mark_buffer_dirty(leaf);
err:
	btrfs_free_path(path);
	return ret;
}

665
static int find_next_chunk(struct btrfs_root *root, u64 objectid, u64 *offset)
666 667 668 669
{
	struct btrfs_path *path;
	int ret;
	struct btrfs_key key;
670
	struct btrfs_chunk *chunk;
671 672 673 674 675
	struct btrfs_key found_key;

	path = btrfs_alloc_path();
	BUG_ON(!path);

676
	key.objectid = objectid;
677 678 679 680 681 682 683 684 685 686 687
	key.offset = (u64)-1;
	key.type = BTRFS_CHUNK_ITEM_KEY;

	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto error;

	BUG_ON(ret == 0);

	ret = btrfs_previous_item(root, path, 0, BTRFS_CHUNK_ITEM_KEY);
	if (ret) {
688
		*offset = 0;
689 690 691
	} else {
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				      path->slots[0]);
692 693 694 695 696 697 698 699
		if (found_key.objectid != objectid)
			*offset = 0;
		else {
			chunk = btrfs_item_ptr(path->nodes[0], path->slots[0],
					       struct btrfs_chunk);
			*offset = found_key.offset +
				btrfs_chunk_length(path->nodes[0], chunk);
		}
700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752
	}
	ret = 0;
error:
	btrfs_free_path(path);
	return ret;
}

static int find_next_devid(struct btrfs_root *root, struct btrfs_path *path,
			   u64 *objectid)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_key found_key;

	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
	key.type = BTRFS_DEV_ITEM_KEY;
	key.offset = (u64)-1;

	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto error;

	BUG_ON(ret == 0);

	ret = btrfs_previous_item(root, path, BTRFS_DEV_ITEMS_OBJECTID,
				  BTRFS_DEV_ITEM_KEY);
	if (ret) {
		*objectid = 1;
	} else {
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				      path->slots[0]);
		*objectid = found_key.offset + 1;
	}
	ret = 0;
error:
	btrfs_release_path(root, path);
	return ret;
}

/*
 * the device information is stored in the chunk root
 * the btrfs_device struct should be fully filled in
 */
int btrfs_add_device(struct btrfs_trans_handle *trans,
		     struct btrfs_root *root,
		     struct btrfs_device *device)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_dev_item *dev_item;
	struct extent_buffer *leaf;
	struct btrfs_key key;
	unsigned long ptr;
753
	u64 free_devid = 0;
754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769

	root = root->fs_info->chunk_root;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	ret = find_next_devid(root, path, &free_devid);
	if (ret)
		goto out;

	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
	key.type = BTRFS_DEV_ITEM_KEY;
	key.offset = free_devid;

	ret = btrfs_insert_empty_item(trans, root, path, &key,
770
				      sizeof(*dev_item));
771 772 773 774 775 776
	if (ret)
		goto out;

	leaf = path->nodes[0];
	dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);

777
	device->devid = free_devid;
778 779 780 781 782 783 784
	btrfs_set_device_id(leaf, dev_item, device->devid);
	btrfs_set_device_type(leaf, dev_item, device->type);
	btrfs_set_device_io_align(leaf, dev_item, device->io_align);
	btrfs_set_device_io_width(leaf, dev_item, device->io_width);
	btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
	btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
	btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
785 786 787
	btrfs_set_device_group(leaf, dev_item, 0);
	btrfs_set_device_seek_speed(leaf, dev_item, 0);
	btrfs_set_device_bandwidth(leaf, dev_item, 0);
788 789

	ptr = (unsigned long)btrfs_device_uuid(dev_item);
790
	write_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
791 792 793 794 795 796 797
	btrfs_mark_buffer_dirty(leaf);
	ret = 0;

out:
	btrfs_free_path(path);
	return ret;
}
798

799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862
static int btrfs_rm_dev_item(struct btrfs_root *root,
			     struct btrfs_device *device)
{
	int ret;
	struct btrfs_path *path;
	struct block_device *bdev = device->bdev;
	struct btrfs_device *next_dev;
	struct btrfs_key key;
	u64 total_bytes;
	struct btrfs_fs_devices *fs_devices;
	struct btrfs_trans_handle *trans;

	root = root->fs_info->chunk_root;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	trans = btrfs_start_transaction(root, 1);
	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
	key.type = BTRFS_DEV_ITEM_KEY;
	key.offset = device->devid;

	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
	if (ret < 0)
		goto out;

	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	ret = btrfs_del_item(trans, root, path);
	if (ret)
		goto out;

	/*
	 * at this point, the device is zero sized.  We want to
	 * remove it from the devices list and zero out the old super
	 */
	list_del_init(&device->dev_list);
	list_del_init(&device->dev_alloc_list);
	fs_devices = root->fs_info->fs_devices;

	next_dev = list_entry(fs_devices->devices.next, struct btrfs_device,
			      dev_list);
	if (bdev == root->fs_info->sb->s_bdev)
		root->fs_info->sb->s_bdev = next_dev->bdev;
	if (bdev == fs_devices->latest_bdev)
		fs_devices->latest_bdev = next_dev->bdev;

	total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
	btrfs_set_super_num_devices(&root->fs_info->super_copy,
				    total_bytes - 1);
out:
	btrfs_free_path(path);
	btrfs_commit_transaction(trans, root);
	return ret;
}

int btrfs_rm_device(struct btrfs_root *root, char *device_path)
{
	struct btrfs_device *device;
	struct block_device *bdev;
863
	struct buffer_head *bh = NULL;
864 865 866 867 868 869 870 871 872 873 874 875 876
	struct btrfs_super_block *disk_super;
	u64 all_avail;
	u64 devid;
	int ret = 0;

	mutex_lock(&root->fs_info->fs_mutex);
	mutex_lock(&uuid_mutex);

	all_avail = root->fs_info->avail_data_alloc_bits |
		root->fs_info->avail_system_alloc_bits |
		root->fs_info->avail_metadata_alloc_bits;

	if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) &&
877
	    btrfs_super_num_devices(&root->fs_info->super_copy) <= 4) {
878 879 880 881 882 883
		printk("btrfs: unable to go below four devices on raid10\n");
		ret = -EINVAL;
		goto out;
	}

	if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) &&
884
	    btrfs_super_num_devices(&root->fs_info->super_copy) <= 2) {
885 886 887 888 889
		printk("btrfs: unable to go below two devices on raid1\n");
		ret = -EINVAL;
		goto out;
	}

890 891 892 893
	if (strcmp(device_path, "missing") == 0) {
		struct list_head *cur;
		struct list_head *devices;
		struct btrfs_device *tmp;
894

895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918
		device = NULL;
		devices = &root->fs_info->fs_devices->devices;
		list_for_each(cur, devices) {
			tmp = list_entry(cur, struct btrfs_device, dev_list);
			if (tmp->in_fs_metadata && !tmp->bdev) {
				device = tmp;
				break;
			}
		}
		bdev = NULL;
		bh = NULL;
		disk_super = NULL;
		if (!device) {
			printk("btrfs: no missing devices found to remove\n");
			goto out;
		}

	} else {
		bdev = open_bdev_excl(device_path, 0,
				      root->fs_info->bdev_holder);
		if (IS_ERR(bdev)) {
			ret = PTR_ERR(bdev);
			goto out;
		}
919

920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943
		bh = __bread(bdev, BTRFS_SUPER_INFO_OFFSET / 4096, 4096);
		if (!bh) {
			ret = -EIO;
			goto error_close;
		}
		disk_super = (struct btrfs_super_block *)bh->b_data;
		if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
		    sizeof(disk_super->magic))) {
			ret = -ENOENT;
			goto error_brelse;
		}
		if (memcmp(disk_super->fsid, root->fs_info->fsid,
			   BTRFS_FSID_SIZE)) {
			ret = -ENOENT;
			goto error_brelse;
		}
		devid = le64_to_cpu(disk_super->dev_item.devid);
		device = btrfs_find_device(root, devid, NULL);
		if (!device) {
			ret = -ENOENT;
			goto error_brelse;
		}

	}
944
	root->fs_info->fs_devices->num_devices--;
C
Chris Mason 已提交
945
	root->fs_info->fs_devices->open_devices--;
946 947 948 949 950 951 952 953 954 955

	ret = btrfs_shrink_device(device, 0);
	if (ret)
		goto error_brelse;


	ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device);
	if (ret)
		goto error_brelse;

956 957 958 959 960 961 962
	if (bh) {
		/* make sure this device isn't detected as part of
		 * the FS anymore
		 */
		memset(&disk_super->magic, 0, sizeof(disk_super->magic));
		set_buffer_dirty(bh);
		sync_dirty_buffer(bh);
963

964 965
		brelse(bh);
	}
966

967 968 969 970 971 972 973 974
	if (device->bdev) {
		/* one close for the device struct or super_block */
		close_bdev_excl(device->bdev);
	}
	if (bdev) {
		/* one close for us */
		close_bdev_excl(bdev);
	}
975 976 977 978 979 980 981 982
	kfree(device->name);
	kfree(device);
	ret = 0;
	goto out;

error_brelse:
	brelse(bh);
error_close:
983 984
	if (bdev)
		close_bdev_excl(bdev);
985 986 987 988 989 990
out:
	mutex_unlock(&uuid_mutex);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;
}

991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024
int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_device *device;
	struct block_device *bdev;
	struct list_head *cur;
	struct list_head *devices;
	u64 total_bytes;
	int ret = 0;


	bdev = open_bdev_excl(device_path, 0, root->fs_info->bdev_holder);
	if (!bdev) {
		return -EIO;
	}
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	devices = &root->fs_info->fs_devices->devices;
	list_for_each(cur, devices) {
		device = list_entry(cur, struct btrfs_device, dev_list);
		if (device->bdev == bdev) {
			ret = -EEXIST;
			goto out;
		}
	}

	device = kzalloc(sizeof(*device), GFP_NOFS);
	if (!device) {
		/* we can safely leave the fs_devices entry around */
		ret = -ENOMEM;
		goto out_close_bdev;
	}

	device->barriers = 1;
1025
	device->work.func = pending_bios_fn;
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
	generate_random_uuid(device->uuid);
	spin_lock_init(&device->io_lock);
	device->name = kstrdup(device_path, GFP_NOFS);
	if (!device->name) {
		kfree(device);
		goto out_close_bdev;
	}
	device->io_width = root->sectorsize;
	device->io_align = root->sectorsize;
	device->sector_size = root->sectorsize;
	device->total_bytes = i_size_read(bdev->bd_inode);
	device->dev_root = root->fs_info->dev_root;
	device->bdev = bdev;
1039
	device->in_fs_metadata = 1;
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056

	ret = btrfs_add_device(trans, root, device);
	if (ret)
		goto out_close_bdev;

	total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
	btrfs_set_super_total_bytes(&root->fs_info->super_copy,
				    total_bytes + device->total_bytes);

	total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
	btrfs_set_super_num_devices(&root->fs_info->super_copy,
				    total_bytes + 1);

	list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
	list_add(&device->dev_alloc_list,
		 &root->fs_info->fs_devices->alloc_list);
	root->fs_info->fs_devices->num_devices++;
1057
	root->fs_info->fs_devices->open_devices++;
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
out:
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;

out_close_bdev:
	close_bdev_excl(bdev);
	goto out;
}

1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
int btrfs_update_device(struct btrfs_trans_handle *trans,
			struct btrfs_device *device)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_root *root;
	struct btrfs_dev_item *dev_item;
	struct extent_buffer *leaf;
	struct btrfs_key key;

	root = device->dev_root->fs_info->chunk_root;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
	key.type = BTRFS_DEV_ITEM_KEY;
	key.offset = device->devid;

	ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
	if (ret < 0)
		goto out;

	if (ret > 0) {
		ret = -ENOENT;
		goto out;
	}

	leaf = path->nodes[0];
	dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);

	btrfs_set_device_id(leaf, dev_item, device->devid);
	btrfs_set_device_type(leaf, dev_item, device->type);
	btrfs_set_device_io_align(leaf, dev_item, device->io_align);
	btrfs_set_device_io_width(leaf, dev_item, device->io_width);
	btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
	btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
	btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
	btrfs_mark_buffer_dirty(leaf);

out:
	btrfs_free_path(path);
	return ret;
}

1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212
int btrfs_grow_device(struct btrfs_trans_handle *trans,
		      struct btrfs_device *device, u64 new_size)
{
	struct btrfs_super_block *super_copy =
		&device->dev_root->fs_info->super_copy;
	u64 old_total = btrfs_super_total_bytes(super_copy);
	u64 diff = new_size - device->total_bytes;

	btrfs_set_super_total_bytes(super_copy, old_total + diff);
	return btrfs_update_device(trans, device);
}

static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root,
			    u64 chunk_tree, u64 chunk_objectid,
			    u64 chunk_offset)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_key key;

	root = root->fs_info->chunk_root;
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	key.objectid = chunk_objectid;
	key.offset = chunk_offset;
	key.type = BTRFS_CHUNK_ITEM_KEY;

	ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
	BUG_ON(ret);

	ret = btrfs_del_item(trans, root, path);
	BUG_ON(ret);

	btrfs_free_path(path);
	return 0;
}

int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
			chunk_offset)
{
	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
	struct btrfs_disk_key *disk_key;
	struct btrfs_chunk *chunk;
	u8 *ptr;
	int ret = 0;
	u32 num_stripes;
	u32 array_size;
	u32 len = 0;
	u32 cur;
	struct btrfs_key key;

	array_size = btrfs_super_sys_array_size(super_copy);

	ptr = super_copy->sys_chunk_array;
	cur = 0;

	while (cur < array_size) {
		disk_key = (struct btrfs_disk_key *)ptr;
		btrfs_disk_key_to_cpu(&key, disk_key);

		len = sizeof(*disk_key);

		if (key.type == BTRFS_CHUNK_ITEM_KEY) {
			chunk = (struct btrfs_chunk *)(ptr + len);
			num_stripes = btrfs_stack_chunk_num_stripes(chunk);
			len += btrfs_chunk_item_size(num_stripes);
		} else {
			ret = -EIO;
			break;
		}
		if (key.objectid == chunk_objectid &&
		    key.offset == chunk_offset) {
			memmove(ptr, ptr + len, array_size - (cur + len));
			array_size -= len;
			btrfs_set_super_sys_array_size(super_copy, array_size);
		} else {
			ptr += len;
			cur += len;
		}
	}
	return ret;
}


int btrfs_relocate_chunk(struct btrfs_root *root,
			 u64 chunk_tree, u64 chunk_objectid,
			 u64 chunk_offset)
{
	struct extent_map_tree *em_tree;
	struct btrfs_root *extent_root;
	struct btrfs_trans_handle *trans;
	struct extent_map *em;
	struct map_lookup *map;
	int ret;
	int i;

1213 1214
	printk("btrfs relocating chunk %llu\n",
	       (unsigned long long)chunk_offset);
1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233
	root = root->fs_info->chunk_root;
	extent_root = root->fs_info->extent_root;
	em_tree = &root->fs_info->mapping_tree.map_tree;

	/* step one, relocate all the extents inside this chunk */
	ret = btrfs_shrink_extent_tree(extent_root, chunk_offset);
	BUG_ON(ret);

	trans = btrfs_start_transaction(root, 1);
	BUG_ON(!trans);

	/*
	 * step two, delete the device extents and the
	 * chunk tree entries
	 */
	spin_lock(&em_tree->lock);
	em = lookup_extent_mapping(em_tree, chunk_offset, 1);
	spin_unlock(&em_tree->lock);

1234 1235
	BUG_ON(em->start > chunk_offset ||
	       em->start + em->len < chunk_offset);
1236 1237 1238 1239 1240 1241
	map = (struct map_lookup *)em->bdev;

	for (i = 0; i < map->num_stripes; i++) {
		ret = btrfs_free_dev_extent(trans, map->stripes[i].dev,
					    map->stripes[i].physical);
		BUG_ON(ret);
1242

1243 1244 1245 1246
		if (map->stripes[i].dev) {
			ret = btrfs_update_device(trans, map->stripes[i].dev);
			BUG_ON(ret);
		}
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
	}
	ret = btrfs_free_chunk(trans, root, chunk_tree, chunk_objectid,
			       chunk_offset);

	BUG_ON(ret);

	if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
		ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset);
		BUG_ON(ret);
	}

	spin_lock(&em_tree->lock);
	remove_extent_mapping(em_tree, em);
	kfree(map);
	em->bdev = NULL;

	/* once for the tree */
	free_extent_map(em);
	spin_unlock(&em_tree->lock);

	/* once for us */
	free_extent_map(em);

	btrfs_end_transaction(trans, root);
	return 0;
}

1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
static u64 div_factor(u64 num, int factor)
{
	if (factor == 10)
		return num;
	num *= factor;
	do_div(num, 10);
	return num;
}


int btrfs_balance(struct btrfs_root *dev_root)
{
	int ret;
	struct list_head *cur;
	struct list_head *devices = &dev_root->fs_info->fs_devices->devices;
	struct btrfs_device *device;
	u64 old_size;
	u64 size_to_free;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_chunk *chunk;
	struct btrfs_root *chunk_root = dev_root->fs_info->chunk_root;
	struct btrfs_trans_handle *trans;
	struct btrfs_key found_key;


	dev_root = dev_root->fs_info->dev_root;

	mutex_lock(&dev_root->fs_info->fs_mutex);
	/* step one make some room on all the devices */
	list_for_each(cur, devices) {
		device = list_entry(cur, struct btrfs_device, dev_list);
		old_size = device->total_bytes;
		size_to_free = div_factor(old_size, 1);
		size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
		if (device->total_bytes - device->bytes_used > size_to_free)
			continue;

		ret = btrfs_shrink_device(device, old_size - size_to_free);
		BUG_ON(ret);

		trans = btrfs_start_transaction(dev_root, 1);
		BUG_ON(!trans);

		ret = btrfs_grow_device(trans, device, old_size);
		BUG_ON(ret);

		btrfs_end_transaction(trans, dev_root);
	}

	/* step two, relocate all the chunks */
	path = btrfs_alloc_path();
	BUG_ON(!path);

	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
	key.offset = (u64)-1;
	key.type = BTRFS_CHUNK_ITEM_KEY;

	while(1) {
		ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
		if (ret < 0)
			goto error;

		/*
		 * this shouldn't happen, it means the last relocate
		 * failed
		 */
		if (ret == 0)
			break;

		ret = btrfs_previous_item(chunk_root, path, 0,
					  BTRFS_CHUNK_ITEM_KEY);
		if (ret) {
			break;
		}
		btrfs_item_key_to_cpu(path->nodes[0], &found_key,
				      path->slots[0]);
		if (found_key.objectid != key.objectid)
			break;
		chunk = btrfs_item_ptr(path->nodes[0],
				       path->slots[0],
				       struct btrfs_chunk);
		key.offset = found_key.offset;
		/* chunk zero is special */
		if (key.offset == 0)
			break;

		ret = btrfs_relocate_chunk(chunk_root,
					   chunk_root->root_key.objectid,
					   found_key.objectid,
					   found_key.offset);
		BUG_ON(ret);
		btrfs_release_path(chunk_root, path);
	}
	ret = 0;
error:
	btrfs_free_path(path);
	mutex_unlock(&dev_root->fs_info->fs_mutex);
	return ret;
}

1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466
/*
 * shrinking a device means finding all of the device extents past
 * the new size, and then following the back refs to the chunks.
 * The chunk relocation code actually frees the device extent
 */
int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = device->dev_root;
	struct btrfs_dev_extent *dev_extent = NULL;
	struct btrfs_path *path;
	u64 length;
	u64 chunk_tree;
	u64 chunk_objectid;
	u64 chunk_offset;
	int ret;
	int slot;
	struct extent_buffer *l;
	struct btrfs_key key;
	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
	u64 old_total = btrfs_super_total_bytes(super_copy);
	u64 diff = device->total_bytes - new_size;


	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	trans = btrfs_start_transaction(root, 1);
	if (!trans) {
		ret = -ENOMEM;
		goto done;
	}

	path->reada = 2;

	device->total_bytes = new_size;
	ret = btrfs_update_device(trans, device);
	if (ret) {
		btrfs_end_transaction(trans, root);
		goto done;
	}
	WARN_ON(diff > old_total);
	btrfs_set_super_total_bytes(super_copy, old_total - diff);
	btrfs_end_transaction(trans, root);

	key.objectid = device->devid;
	key.offset = (u64)-1;
	key.type = BTRFS_DEV_EXTENT_KEY;

	while (1) {
		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
		if (ret < 0)
			goto done;

		ret = btrfs_previous_item(root, path, 0, key.type);
		if (ret < 0)
			goto done;
		if (ret) {
			ret = 0;
			goto done;
		}

		l = path->nodes[0];
		slot = path->slots[0];
		btrfs_item_key_to_cpu(l, &key, path->slots[0]);

		if (key.objectid != device->devid)
			goto done;

		dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
		length = btrfs_dev_extent_length(l, dev_extent);

		if (key.offset + length <= new_size)
			goto done;

		chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
		chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
		chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
		btrfs_release_path(root, path);

		ret = btrfs_relocate_chunk(root, chunk_tree, chunk_objectid,
					   chunk_offset);
		if (ret)
			goto done;
	}

done:
	btrfs_free_path(path);
	return ret;
}

1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490
int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
			   struct btrfs_root *root,
			   struct btrfs_key *key,
			   struct btrfs_chunk *chunk, int item_size)
{
	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
	struct btrfs_disk_key disk_key;
	u32 array_size;
	u8 *ptr;

	array_size = btrfs_super_sys_array_size(super_copy);
	if (array_size + item_size > BTRFS_SYSTEM_CHUNK_ARRAY_SIZE)
		return -EFBIG;

	ptr = super_copy->sys_chunk_array + array_size;
	btrfs_cpu_key_to_disk(&disk_key, key);
	memcpy(ptr, &disk_key, sizeof(disk_key));
	ptr += sizeof(disk_key);
	memcpy(ptr, chunk, item_size);
	item_size += sizeof(disk_key);
	btrfs_set_super_sys_array_size(super_copy, array_size + item_size);
	return 0;
}

1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502
static u64 chunk_bytes_by_type(u64 type, u64 calc_size, int num_stripes,
			       int sub_stripes)
{
	if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP))
		return calc_size;
	else if (type & BTRFS_BLOCK_GROUP_RAID10)
		return calc_size * (num_stripes / sub_stripes);
	else
		return calc_size * num_stripes;
}


1503 1504
int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
		      struct btrfs_root *extent_root, u64 *start,
1505
		      u64 *num_bytes, u64 type)
1506 1507
{
	u64 dev_offset;
1508
	struct btrfs_fs_info *info = extent_root->fs_info;
1509
	struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
1510
	struct btrfs_path *path;
1511 1512 1513
	struct btrfs_stripe *stripes;
	struct btrfs_device *device = NULL;
	struct btrfs_chunk *chunk;
1514
	struct list_head private_devs;
1515
	struct list_head *dev_list;
1516
	struct list_head *cur;
1517 1518 1519
	struct extent_map_tree *em_tree;
	struct map_lookup *map;
	struct extent_map *em;
1520
	int min_stripe_size = 1 * 1024 * 1024;
1521 1522
	u64 physical;
	u64 calc_size = 1024 * 1024 * 1024;
1523 1524
	u64 max_chunk_size = calc_size;
	u64 min_free;
1525 1526
	u64 avail;
	u64 max_avail = 0;
1527
	u64 percent_max;
1528
	int num_stripes = 1;
1529
	int min_stripes = 1;
C
Chris Mason 已提交
1530
	int sub_stripes = 0;
1531
	int looped = 0;
1532
	int ret;
1533
	int index;
1534
	int stripe_len = 64 * 1024;
1535 1536
	struct btrfs_key key;

1537 1538 1539 1540 1541
	if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
	    (type & BTRFS_BLOCK_GROUP_DUP)) {
		WARN_ON(1);
		type &= ~BTRFS_BLOCK_GROUP_DUP;
	}
1542
	dev_list = &extent_root->fs_info->fs_devices->alloc_list;
1543 1544
	if (list_empty(dev_list))
		return -ENOSPC;
1545

1546
	if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
C
Chris Mason 已提交
1547
		num_stripes = extent_root->fs_info->fs_devices->open_devices;
1548 1549 1550
		min_stripes = 2;
	}
	if (type & (BTRFS_BLOCK_GROUP_DUP)) {
1551
		num_stripes = 2;
1552 1553
		min_stripes = 2;
	}
1554 1555
	if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
		num_stripes = min_t(u64, 2,
C
Chris Mason 已提交
1556
			    extent_root->fs_info->fs_devices->open_devices);
1557 1558
		if (num_stripes < 2)
			return -ENOSPC;
1559
		min_stripes = 2;
1560
	}
C
Chris Mason 已提交
1561
	if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
C
Chris Mason 已提交
1562
		num_stripes = extent_root->fs_info->fs_devices->open_devices;
C
Chris Mason 已提交
1563 1564 1565 1566
		if (num_stripes < 4)
			return -ENOSPC;
		num_stripes &= ~(u32)1;
		sub_stripes = 2;
1567
		min_stripes = 4;
C
Chris Mason 已提交
1568
	}
1569 1570 1571

	if (type & BTRFS_BLOCK_GROUP_DATA) {
		max_chunk_size = 10 * calc_size;
1572
		min_stripe_size = 64 * 1024 * 1024;
1573 1574
	} else if (type & BTRFS_BLOCK_GROUP_METADATA) {
		max_chunk_size = 4 * calc_size;
1575 1576 1577 1578 1579
		min_stripe_size = 32 * 1024 * 1024;
	} else if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
		calc_size = 8 * 1024 * 1024;
		max_chunk_size = calc_size * 2;
		min_stripe_size = 1 * 1024 * 1024;
1580 1581
	}

1582 1583 1584 1585
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

1586 1587 1588 1589
	/* we don't want a chunk larger than 10% of the FS */
	percent_max = div_factor(btrfs_super_total_bytes(&info->super_copy), 1);
	max_chunk_size = min(percent_max, max_chunk_size);

1590
again:
1591 1592 1593 1594 1595 1596 1597
	if (calc_size * num_stripes > max_chunk_size) {
		calc_size = max_chunk_size;
		do_div(calc_size, num_stripes);
		do_div(calc_size, stripe_len);
		calc_size *= stripe_len;
	}
	/* we don't want tiny stripes */
1598
	calc_size = max_t(u64, min_stripe_size, calc_size);
1599 1600 1601 1602

	do_div(calc_size, stripe_len);
	calc_size *= stripe_len;

1603 1604 1605
	INIT_LIST_HEAD(&private_devs);
	cur = dev_list->next;
	index = 0;
1606 1607 1608

	if (type & BTRFS_BLOCK_GROUP_DUP)
		min_free = calc_size * 2;
1609 1610
	else
		min_free = calc_size;
1611

1612 1613 1614
	/* we add 1MB because we never use the first 1MB of the device */
	min_free += 1024 * 1024;

1615 1616
	/* build a private list of devices we will allocate from */
	while(index < num_stripes) {
1617
		device = list_entry(cur, struct btrfs_device, dev_alloc_list);
1618

1619 1620 1621 1622
		if (device->total_bytes > device->bytes_used)
			avail = device->total_bytes - device->bytes_used;
		else
			avail = 0;
1623
		cur = cur->next;
1624

1625
		if (device->in_fs_metadata && avail >= min_free) {
1626 1627 1628 1629 1630 1631 1632
			u64 ignored_start = 0;
			ret = find_free_dev_extent(trans, device, path,
						   min_free,
						   &ignored_start);
			if (ret == 0) {
				list_move_tail(&device->dev_alloc_list,
					       &private_devs);
1633
				index++;
1634 1635 1636
				if (type & BTRFS_BLOCK_GROUP_DUP)
					index++;
			}
1637
		} else if (device->in_fs_metadata && avail > max_avail)
1638
			max_avail = avail;
1639 1640 1641 1642 1643
		if (cur == dev_list)
			break;
	}
	if (index < num_stripes) {
		list_splice(&private_devs, dev_list);
1644 1645 1646 1647 1648 1649 1650 1651 1652
		if (index >= min_stripes) {
			num_stripes = index;
			if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
				num_stripes /= sub_stripes;
				num_stripes *= sub_stripes;
			}
			looped = 1;
			goto again;
		}
1653 1654 1655 1656 1657
		if (!looped && max_avail > 0) {
			looped = 1;
			calc_size = max_avail;
			goto again;
		}
1658
		btrfs_free_path(path);
1659 1660
		return -ENOSPC;
	}
1661 1662 1663 1664
	key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
	key.type = BTRFS_CHUNK_ITEM_KEY;
	ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
			      &key.offset);
1665 1666
	if (ret) {
		btrfs_free_path(path);
1667
		return ret;
1668
	}
1669 1670

	chunk = kmalloc(btrfs_chunk_item_size(num_stripes), GFP_NOFS);
1671 1672
	if (!chunk) {
		btrfs_free_path(path);
1673
		return -ENOMEM;
1674
	}
1675

1676 1677 1678
	map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
	if (!map) {
		kfree(chunk);
1679
		btrfs_free_path(path);
1680 1681
		return -ENOMEM;
	}
1682 1683
	btrfs_free_path(path);
	path = NULL;
1684

1685
	stripes = &chunk->stripe;
1686 1687
	*num_bytes = chunk_bytes_by_type(type, calc_size,
					 num_stripes, sub_stripes);
1688

1689
	index = 0;
1690
	while(index < num_stripes) {
1691
		struct btrfs_stripe *stripe;
1692 1693
		BUG_ON(list_empty(&private_devs));
		cur = private_devs.next;
1694
		device = list_entry(cur, struct btrfs_device, dev_alloc_list);
1695 1696 1697 1698

		/* loop over this device again if we're doing a dup group */
		if (!(type & BTRFS_BLOCK_GROUP_DUP) ||
		    (index == num_stripes - 1))
1699
			list_move_tail(&device->dev_alloc_list, dev_list);
1700 1701

		ret = btrfs_alloc_dev_extent(trans, device,
1702 1703 1704
			     info->chunk_root->root_key.objectid,
			     BTRFS_FIRST_CHUNK_TREE_OBJECTID, key.offset,
			     calc_size, &dev_offset);
1705 1706 1707 1708 1709
		BUG_ON(ret);
		device->bytes_used += calc_size;
		ret = btrfs_update_device(trans, device);
		BUG_ON(ret);

1710 1711
		map->stripes[index].dev = device;
		map->stripes[index].physical = dev_offset;
1712 1713 1714 1715
		stripe = stripes + index;
		btrfs_set_stack_stripe_devid(stripe, device->devid);
		btrfs_set_stack_stripe_offset(stripe, dev_offset);
		memcpy(stripe->dev_uuid, device->uuid, BTRFS_UUID_SIZE);
1716 1717 1718
		physical = dev_offset;
		index++;
	}
1719
	BUG_ON(!list_empty(&private_devs));
1720

1721 1722
	/* key was set above */
	btrfs_set_stack_chunk_length(chunk, *num_bytes);
1723
	btrfs_set_stack_chunk_owner(chunk, extent_root->root_key.objectid);
1724
	btrfs_set_stack_chunk_stripe_len(chunk, stripe_len);
1725 1726
	btrfs_set_stack_chunk_type(chunk, type);
	btrfs_set_stack_chunk_num_stripes(chunk, num_stripes);
1727 1728
	btrfs_set_stack_chunk_io_align(chunk, stripe_len);
	btrfs_set_stack_chunk_io_width(chunk, stripe_len);
1729
	btrfs_set_stack_chunk_sector_size(chunk, extent_root->sectorsize);
C
Chris Mason 已提交
1730
	btrfs_set_stack_chunk_sub_stripes(chunk, sub_stripes);
1731 1732 1733 1734 1735 1736
	map->sector_size = extent_root->sectorsize;
	map->stripe_len = stripe_len;
	map->io_align = stripe_len;
	map->io_width = stripe_len;
	map->type = type;
	map->num_stripes = num_stripes;
C
Chris Mason 已提交
1737
	map->sub_stripes = sub_stripes;
1738 1739 1740 1741

	ret = btrfs_insert_item(trans, chunk_root, &key, chunk,
				btrfs_chunk_item_size(num_stripes));
	BUG_ON(ret);
1742
	*start = key.offset;;
1743 1744 1745 1746 1747

	em = alloc_extent_map(GFP_NOFS);
	if (!em)
		return -ENOMEM;
	em->bdev = (struct block_device *)map;
1748 1749
	em->start = key.offset;
	em->len = *num_bytes;
1750 1751
	em->block_start = 0;

1752 1753 1754 1755 1756
	if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
		ret = btrfs_add_system_chunk(trans, chunk_root, &key,
				    chunk, btrfs_chunk_item_size(num_stripes));
		BUG_ON(ret);
	}
1757 1758 1759 1760 1761 1762
	kfree(chunk);

	em_tree = &extent_root->fs_info->mapping_tree.map_tree;
	spin_lock(&em_tree->lock);
	ret = add_extent_mapping(em_tree, em);
	spin_unlock(&em_tree->lock);
1763
	BUG_ON(ret);
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792
	free_extent_map(em);
	return ret;
}

void btrfs_mapping_init(struct btrfs_mapping_tree *tree)
{
	extent_map_tree_init(&tree->map_tree, GFP_NOFS);
}

void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree)
{
	struct extent_map *em;

	while(1) {
		spin_lock(&tree->map_tree.lock);
		em = lookup_extent_mapping(&tree->map_tree, 0, (u64)-1);
		if (em)
			remove_extent_mapping(&tree->map_tree, em);
		spin_unlock(&tree->map_tree.lock);
		if (!em)
			break;
		kfree(em->bdev);
		/* once for us */
		free_extent_map(em);
		/* once for the tree */
		free_extent_map(em);
	}
}

1793 1794 1795 1796 1797 1798 1799 1800 1801
int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len)
{
	struct extent_map *em;
	struct map_lookup *map;
	struct extent_map_tree *em_tree = &map_tree->map_tree;
	int ret;

	spin_lock(&em_tree->lock);
	em = lookup_extent_mapping(em_tree, logical, len);
1802
	spin_unlock(&em_tree->lock);
1803 1804 1805 1806 1807 1808
	BUG_ON(!em);

	BUG_ON(em->start > logical || em->start + em->len < logical);
	map = (struct map_lookup *)em->bdev;
	if (map->type & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1))
		ret = map->num_stripes;
C
Chris Mason 已提交
1809 1810
	else if (map->type & BTRFS_BLOCK_GROUP_RAID10)
		ret = map->sub_stripes;
1811 1812 1813 1814 1815 1816
	else
		ret = 1;
	free_extent_map(em);
	return ret;
}

1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832
static int find_live_mirror(struct map_lookup *map, int first, int num,
			    int optimal)
{
	int i;
	if (map->stripes[optimal].dev->bdev)
		return optimal;
	for (i = first; i < first + num; i++) {
		if (map->stripes[i].dev->bdev)
			return i;
	}
	/* we couldn't find one that doesn't fail.  Just return something
	 * and the io error handling code will clean up eventually
	 */
	return optimal;
}

1833 1834 1835 1836
static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
			     u64 logical, u64 *length,
			     struct btrfs_multi_bio **multi_ret,
			     int mirror_num, struct page *unplug_page)
1837 1838 1839 1840 1841
{
	struct extent_map *em;
	struct map_lookup *map;
	struct extent_map_tree *em_tree = &map_tree->map_tree;
	u64 offset;
1842 1843
	u64 stripe_offset;
	u64 stripe_nr;
1844
	int stripes_allocated = 8;
C
Chris Mason 已提交
1845
	int stripes_required = 1;
1846
	int stripe_index;
1847
	int i;
1848
	int num_stripes;
1849
	int max_errors = 0;
1850
	struct btrfs_multi_bio *multi = NULL;
1851

1852 1853 1854 1855 1856 1857 1858 1859 1860
	if (multi_ret && !(rw & (1 << BIO_RW))) {
		stripes_allocated = 1;
	}
again:
	if (multi_ret) {
		multi = kzalloc(btrfs_multi_bio_size(stripes_allocated),
				GFP_NOFS);
		if (!multi)
			return -ENOMEM;
1861 1862

		atomic_set(&multi->error, 0);
1863
	}
1864 1865 1866

	spin_lock(&em_tree->lock);
	em = lookup_extent_mapping(em_tree, logical, *length);
1867
	spin_unlock(&em_tree->lock);
1868 1869 1870 1871

	if (!em && unplug_page)
		return 0;

1872
	if (!em) {
1873
		printk("unable to find logical %Lu len %Lu\n", logical, *length);
1874
		BUG();
1875
	}
1876 1877 1878 1879

	BUG_ON(em->start > logical || em->start + em->len < logical);
	map = (struct map_lookup *)em->bdev;
	offset = logical - em->start;
1880

1881 1882 1883
	if (mirror_num > map->num_stripes)
		mirror_num = 0;

1884
	/* if our multi bio struct is too small, back off and try again */
C
Chris Mason 已提交
1885 1886 1887 1888
	if (rw & (1 << BIO_RW)) {
		if (map->type & (BTRFS_BLOCK_GROUP_RAID1 |
				 BTRFS_BLOCK_GROUP_DUP)) {
			stripes_required = map->num_stripes;
1889
			max_errors = 1;
C
Chris Mason 已提交
1890 1891
		} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
			stripes_required = map->sub_stripes;
1892
			max_errors = 1;
C
Chris Mason 已提交
1893 1894 1895 1896
		}
	}
	if (multi_ret && rw == WRITE &&
	    stripes_allocated < stripes_required) {
1897 1898 1899 1900 1901
		stripes_allocated = map->num_stripes;
		free_extent_map(em);
		kfree(multi);
		goto again;
	}
1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914
	stripe_nr = offset;
	/*
	 * stripe_nr counts the total number of stripes we have to stride
	 * to get to this block
	 */
	do_div(stripe_nr, map->stripe_len);

	stripe_offset = stripe_nr * map->stripe_len;
	BUG_ON(offset < stripe_offset);

	/* stripe_offset is the offset of this block in its stripe*/
	stripe_offset = offset - stripe_offset;

1915
	if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 |
C
Chris Mason 已提交
1916
			 BTRFS_BLOCK_GROUP_RAID10 |
1917 1918 1919 1920 1921 1922 1923
			 BTRFS_BLOCK_GROUP_DUP)) {
		/* we limit the length of each bio to what fits in a stripe */
		*length = min_t(u64, em->len - offset,
			      map->stripe_len - stripe_offset);
	} else {
		*length = em->len - offset;
	}
1924 1925

	if (!multi_ret && !unplug_page)
1926 1927
		goto out;

1928
	num_stripes = 1;
1929
	stripe_index = 0;
1930
	if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
1931 1932
		if (unplug_page || (rw & (1 << BIO_RW)))
			num_stripes = map->num_stripes;
1933
		else if (mirror_num)
1934
			stripe_index = mirror_num - 1;
1935 1936 1937 1938 1939
		else {
			stripe_index = find_live_mirror(map, 0,
					    map->num_stripes,
					    current->pid % map->num_stripes);
		}
1940

1941
	} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
1942
		if (rw & (1 << BIO_RW))
1943
			num_stripes = map->num_stripes;
1944 1945
		else if (mirror_num)
			stripe_index = mirror_num - 1;
1946

C
Chris Mason 已提交
1947 1948 1949 1950 1951 1952
	} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
		int factor = map->num_stripes / map->sub_stripes;

		stripe_index = do_div(stripe_nr, factor);
		stripe_index *= map->sub_stripes;

1953 1954
		if (unplug_page || (rw & (1 << BIO_RW)))
			num_stripes = map->sub_stripes;
C
Chris Mason 已提交
1955 1956
		else if (mirror_num)
			stripe_index += mirror_num - 1;
1957 1958 1959 1960 1961
		else {
			stripe_index = find_live_mirror(map, stripe_index,
					      map->sub_stripes, stripe_index +
					      current->pid % map->sub_stripes);
		}
1962 1963 1964 1965 1966 1967 1968 1969
	} else {
		/*
		 * after this do_div call, stripe_nr is the number of stripes
		 * on this device we have to walk to find the data, and
		 * stripe_index is the number of our device in the stripe array
		 */
		stripe_index = do_div(stripe_nr, map->num_stripes);
	}
1970
	BUG_ON(stripe_index >= map->num_stripes);
1971

1972 1973 1974 1975 1976 1977
	for (i = 0; i < num_stripes; i++) {
		if (unplug_page) {
			struct btrfs_device *device;
			struct backing_dev_info *bdi;

			device = map->stripes[stripe_index].dev;
1978 1979 1980 1981 1982
			if (device->bdev) {
				bdi = blk_get_backing_dev_info(device->bdev);
				if (bdi->unplug_io_fn) {
					bdi->unplug_io_fn(bdi, unplug_page);
				}
1983 1984 1985 1986 1987 1988 1989
			}
		} else {
			multi->stripes[i].physical =
				map->stripes[stripe_index].physical +
				stripe_offset + stripe_nr * map->stripe_len;
			multi->stripes[i].dev = map->stripes[stripe_index].dev;
		}
1990
		stripe_index++;
1991
	}
1992 1993 1994
	if (multi_ret) {
		*multi_ret = multi;
		multi->num_stripes = num_stripes;
1995
		multi->max_errors = max_errors;
1996
	}
1997
out:
1998 1999 2000 2001
	free_extent_map(em);
	return 0;
}

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
		      u64 logical, u64 *length,
		      struct btrfs_multi_bio **multi_ret, int mirror_num)
{
	return __btrfs_map_block(map_tree, rw, logical, length, multi_ret,
				 mirror_num, NULL);
}

int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree,
		      u64 logical, struct page *page)
{
	u64 length = PAGE_CACHE_SIZE;
	return __btrfs_map_block(map_tree, READ, logical, &length,
				 NULL, 0, page);
}


2019 2020 2021 2022 2023 2024 2025
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void end_bio_multi_stripe(struct bio *bio, int err)
#else
static int end_bio_multi_stripe(struct bio *bio,
				   unsigned int bytes_done, int err)
#endif
{
2026
	struct btrfs_multi_bio *multi = bio->bi_private;
2027 2028 2029 2030 2031 2032

#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	if (bio->bi_size)
		return 1;
#endif
	if (err)
2033
		atomic_inc(&multi->error);
2034

2035
	if (atomic_dec_and_test(&multi->stripes_pending)) {
2036 2037
		bio->bi_private = multi->private;
		bio->bi_end_io = multi->end_io;
2038 2039 2040
		/* only send an error to the higher layers if it is
		 * beyond the tolerance of the multi-bio
		 */
2041
		if (atomic_read(&multi->error) > multi->max_errors) {
2042
			err = -EIO;
2043 2044 2045 2046 2047 2048
		} else if (err) {
			/*
			 * this bio is actually up to date, we didn't
			 * go over the max number of errors
			 */
			set_bit(BIO_UPTODATE, &bio->bi_flags);
2049
			err = 0;
2050
		}
2051 2052
		kfree(multi);

2053 2054 2055
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
		bio_endio(bio, bio->bi_size, err);
#else
2056
		bio_endio(bio, err);
2057
#endif
2058 2059 2060 2061 2062 2063 2064 2065
	} else {
		bio_put(bio);
	}
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
	return 0;
#endif
}

2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118
struct async_sched {
	struct bio *bio;
	int rw;
	struct btrfs_fs_info *info;
	struct btrfs_work work;
};

/*
 * see run_scheduled_bios for a description of why bios are collected for
 * async submit.
 *
 * This will add one bio to the pending list for a device and make sure
 * the work struct is scheduled.
 */
int schedule_bio(struct btrfs_root *root, struct btrfs_device *device,
		 int rw, struct bio *bio)
{
	int should_queue = 1;

	/* don't bother with additional async steps for reads, right now */
	if (!(rw & (1 << BIO_RW))) {
		submit_bio(rw, bio);
		return 0;
	}

	/*
	 * nr_async_sumbits allows us to reliably return congestion to the
	 * higher layers.  Otherwise, the async bio makes it appear we have
	 * made progress against dirty pages when we've really just put it
	 * on a queue for later
	 */
	atomic_inc(&root->fs_info->nr_async_submits);
	bio->bi_next = NULL;
	bio->bi_rw |= rw;

	spin_lock(&device->io_lock);

	if (device->pending_bio_tail)
		device->pending_bio_tail->bi_next = bio;

	device->pending_bio_tail = bio;
	if (!device->pending_bios)
		device->pending_bios = bio;
	if (device->running_pending)
		should_queue = 0;

	spin_unlock(&device->io_lock);

	if (should_queue)
		btrfs_queue_worker(&root->fs_info->workers, &device->work);
	return 0;
}

2119
int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio,
2120
		  int mirror_num, int async_submit)
2121 2122 2123
{
	struct btrfs_mapping_tree *map_tree;
	struct btrfs_device *dev;
2124
	struct bio *first_bio = bio;
2125 2126 2127
	u64 logical = bio->bi_sector << 9;
	u64 length = 0;
	u64 map_length;
2128
	struct btrfs_multi_bio *multi = NULL;
2129
	int ret;
2130 2131
	int dev_nr = 0;
	int total_devs = 1;
2132

2133
	length = bio->bi_size;
2134 2135
	map_tree = &root->fs_info->mapping_tree;
	map_length = length;
2136

2137 2138
	ret = btrfs_map_block(map_tree, rw, logical, &map_length, &multi,
			      mirror_num);
2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150
	BUG_ON(ret);

	total_devs = multi->num_stripes;
	if (map_length < length) {
		printk("mapping failed logical %Lu bio len %Lu "
		       "len %Lu\n", logical, length, map_length);
		BUG();
	}
	multi->end_io = first_bio->bi_end_io;
	multi->private = first_bio->bi_private;
	atomic_set(&multi->stripes_pending, multi->num_stripes);

2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161
	while(dev_nr < total_devs) {
		if (total_devs > 1) {
			if (dev_nr < total_devs - 1) {
				bio = bio_clone(first_bio, GFP_NOFS);
				BUG_ON(!bio);
			} else {
				bio = first_bio;
			}
			bio->bi_private = multi;
			bio->bi_end_io = end_bio_multi_stripe;
		}
2162 2163
		bio->bi_sector = multi->stripes[dev_nr].physical >> 9;
		dev = multi->stripes[dev_nr].dev;
2164 2165
		if (dev && dev->bdev) {
			bio->bi_bdev = dev->bdev;
2166 2167 2168 2169
			if (async_submit)
				schedule_bio(root, dev, rw, bio);
			else
				submit_bio(rw, bio);
2170 2171 2172 2173 2174 2175 2176 2177 2178
		} else {
			bio->bi_bdev = root->fs_info->fs_devices->latest_bdev;
			bio->bi_sector = logical >> 9;
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
			bio_endio(bio, bio->bi_size, -EIO);
#else
			bio_endio(bio, -EIO);
#endif
		}
2179 2180
		dev_nr++;
	}
2181 2182
	if (total_devs == 1)
		kfree(multi);
2183 2184 2185
	return 0;
}

2186 2187
struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
				       u8 *uuid)
2188
{
2189
	struct list_head *head = &root->fs_info->fs_devices->devices;
2190

2191
	return __find_device(head, devid, uuid);
2192 2193
}

2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207
static struct btrfs_device *add_missing_dev(struct btrfs_root *root,
					    u64 devid, u8 *dev_uuid)
{
	struct btrfs_device *device;
	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;

	device = kzalloc(sizeof(*device), GFP_NOFS);
	list_add(&device->dev_list,
		 &fs_devices->devices);
	list_add(&device->dev_alloc_list,
		 &fs_devices->alloc_list);
	device->barriers = 1;
	device->dev_root = root->fs_info->dev_root;
	device->devid = devid;
2208
	device->work.func = pending_bios_fn;
2209 2210 2211 2212 2213 2214 2215
	fs_devices->num_devices++;
	spin_lock_init(&device->io_lock);
	memcpy(device->uuid, dev_uuid, BTRFS_UUID_SIZE);
	return device;
}


2216 2217 2218 2219 2220 2221 2222 2223 2224 2225
static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
			  struct extent_buffer *leaf,
			  struct btrfs_chunk *chunk)
{
	struct btrfs_mapping_tree *map_tree = &root->fs_info->mapping_tree;
	struct map_lookup *map;
	struct extent_map *em;
	u64 logical;
	u64 length;
	u64 devid;
2226
	u8 uuid[BTRFS_UUID_SIZE];
2227
	int num_stripes;
2228
	int ret;
2229
	int i;
2230

2231 2232
	logical = key->offset;
	length = btrfs_chunk_length(leaf, chunk);
2233

2234 2235
	spin_lock(&map_tree->map_tree.lock);
	em = lookup_extent_mapping(&map_tree->map_tree, logical, 1);
2236
	spin_unlock(&map_tree->map_tree.lock);
2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252

	/* already mapped? */
	if (em && em->start <= logical && em->start + em->len > logical) {
		free_extent_map(em);
		return 0;
	} else if (em) {
		free_extent_map(em);
	}

	map = kzalloc(sizeof(*map), GFP_NOFS);
	if (!map)
		return -ENOMEM;

	em = alloc_extent_map(GFP_NOFS);
	if (!em)
		return -ENOMEM;
2253 2254
	num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
	map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
2255 2256 2257 2258 2259 2260 2261 2262 2263 2264
	if (!map) {
		free_extent_map(em);
		return -ENOMEM;
	}

	em->bdev = (struct block_device *)map;
	em->start = logical;
	em->len = length;
	em->block_start = 0;

2265 2266 2267 2268 2269 2270
	map->num_stripes = num_stripes;
	map->io_width = btrfs_chunk_io_width(leaf, chunk);
	map->io_align = btrfs_chunk_io_align(leaf, chunk);
	map->sector_size = btrfs_chunk_sector_size(leaf, chunk);
	map->stripe_len = btrfs_chunk_stripe_len(leaf, chunk);
	map->type = btrfs_chunk_type(leaf, chunk);
C
Chris Mason 已提交
2271
	map->sub_stripes = btrfs_chunk_sub_stripes(leaf, chunk);
2272 2273 2274 2275
	for (i = 0; i < num_stripes; i++) {
		map->stripes[i].physical =
			btrfs_stripe_offset_nr(leaf, chunk, i);
		devid = btrfs_stripe_devid_nr(leaf, chunk, i);
2276 2277 2278 2279
		read_extent_buffer(leaf, uuid, (unsigned long)
				   btrfs_stripe_dev_uuid_nr(chunk, i),
				   BTRFS_UUID_SIZE);
		map->stripes[i].dev = btrfs_find_device(root, devid, uuid);
2280 2281

		if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) {
2282 2283 2284 2285
			kfree(map);
			free_extent_map(em);
			return -EIO;
		}
2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
		if (!map->stripes[i].dev) {
			map->stripes[i].dev =
				add_missing_dev(root, devid, uuid);
			if (!map->stripes[i].dev) {
				kfree(map);
				free_extent_map(em);
				return -EIO;
			}
		}
		map->stripes[i].dev->in_fs_metadata = 1;
2296 2297 2298 2299 2300
	}

	spin_lock(&map_tree->map_tree.lock);
	ret = add_extent_mapping(&map_tree->map_tree, em);
	spin_unlock(&map_tree->map_tree.lock);
2301
	BUG_ON(ret);
2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321
	free_extent_map(em);

	return 0;
}

static int fill_device_from_item(struct extent_buffer *leaf,
				 struct btrfs_dev_item *dev_item,
				 struct btrfs_device *device)
{
	unsigned long ptr;

	device->devid = btrfs_device_id(leaf, dev_item);
	device->total_bytes = btrfs_device_total_bytes(leaf, dev_item);
	device->bytes_used = btrfs_device_bytes_used(leaf, dev_item);
	device->type = btrfs_device_type(leaf, dev_item);
	device->io_align = btrfs_device_io_align(leaf, dev_item);
	device->io_width = btrfs_device_io_width(leaf, dev_item);
	device->sector_size = btrfs_device_sector_size(leaf, dev_item);

	ptr = (unsigned long)btrfs_device_uuid(dev_item);
2322
	read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
2323 2324 2325 2326

	return 0;
}

2327
static int read_one_dev(struct btrfs_root *root,
2328 2329 2330 2331 2332 2333
			struct extent_buffer *leaf,
			struct btrfs_dev_item *dev_item)
{
	struct btrfs_device *device;
	u64 devid;
	int ret;
2334 2335
	u8 dev_uuid[BTRFS_UUID_SIZE];

2336
	devid = btrfs_device_id(leaf, dev_item);
2337 2338 2339 2340
	read_extent_buffer(leaf, dev_uuid,
			   (unsigned long)btrfs_device_uuid(dev_item),
			   BTRFS_UUID_SIZE);
	device = btrfs_find_device(root, devid, dev_uuid);
2341
	if (!device) {
2342 2343
		printk("warning devid %Lu missing\n", devid);
		device = add_missing_dev(root, devid, dev_uuid);
2344 2345 2346
		if (!device)
			return -ENOMEM;
	}
2347 2348 2349

	fill_device_from_item(leaf, dev_item, device);
	device->dev_root = root->fs_info->dev_root;
2350
	device->in_fs_metadata = 1;
2351 2352 2353 2354 2355 2356 2357 2358 2359 2360
	ret = 0;
#if 0
	ret = btrfs_open_device(device);
	if (ret) {
		kfree(device);
	}
#endif
	return ret;
}

2361 2362 2363 2364 2365 2366 2367 2368 2369
int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf)
{
	struct btrfs_dev_item *dev_item;

	dev_item = (struct btrfs_dev_item *)offsetof(struct btrfs_super_block,
						     dev_item);
	return read_one_dev(root, buf, dev_item);
}

2370 2371 2372
int btrfs_read_sys_array(struct btrfs_root *root)
{
	struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
2373
	struct extent_buffer *sb;
2374 2375
	struct btrfs_disk_key *disk_key;
	struct btrfs_chunk *chunk;
2376 2377 2378
	u8 *ptr;
	unsigned long sb_ptr;
	int ret = 0;
2379 2380 2381 2382
	u32 num_stripes;
	u32 array_size;
	u32 len = 0;
	u32 cur;
2383
	struct btrfs_key key;
2384

2385 2386 2387 2388 2389 2390
	sb = btrfs_find_create_tree_block(root, BTRFS_SUPER_INFO_OFFSET,
					  BTRFS_SUPER_INFO_SIZE);
	if (!sb)
		return -ENOMEM;
	btrfs_set_buffer_uptodate(sb);
	write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE);
2391 2392 2393 2394 2395 2396 2397 2398 2399 2400
	array_size = btrfs_super_sys_array_size(super_copy);

	ptr = super_copy->sys_chunk_array;
	sb_ptr = offsetof(struct btrfs_super_block, sys_chunk_array);
	cur = 0;

	while (cur < array_size) {
		disk_key = (struct btrfs_disk_key *)ptr;
		btrfs_disk_key_to_cpu(&key, disk_key);

2401
		len = sizeof(*disk_key); ptr += len;
2402 2403 2404
		sb_ptr += len;
		cur += len;

2405
		if (key.type == BTRFS_CHUNK_ITEM_KEY) {
2406
			chunk = (struct btrfs_chunk *)sb_ptr;
2407
			ret = read_one_chunk(root, &key, sb, chunk);
2408 2409
			if (ret)
				break;
2410 2411 2412
			num_stripes = btrfs_chunk_num_stripes(sb, chunk);
			len = btrfs_chunk_item_size(num_stripes);
		} else {
2413 2414
			ret = -EIO;
			break;
2415 2416 2417 2418 2419
		}
		ptr += len;
		sb_ptr += len;
		cur += len;
	}
2420
	free_extent_buffer(sb);
2421
	return ret;
2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466
}

int btrfs_read_chunk_tree(struct btrfs_root *root)
{
	struct btrfs_path *path;
	struct extent_buffer *leaf;
	struct btrfs_key key;
	struct btrfs_key found_key;
	int ret;
	int slot;

	root = root->fs_info->chunk_root;

	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	/* first we search for all of the device items, and then we
	 * read in all of the chunk items.  This way we can create chunk
	 * mappings that reference all of the devices that are afound
	 */
	key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
	key.offset = 0;
	key.type = 0;
again:
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	while(1) {
		leaf = path->nodes[0];
		slot = path->slots[0];
		if (slot >= btrfs_header_nritems(leaf)) {
			ret = btrfs_next_leaf(root, path);
			if (ret == 0)
				continue;
			if (ret < 0)
				goto error;
			break;
		}
		btrfs_item_key_to_cpu(leaf, &found_key, slot);
		if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
			if (found_key.objectid != BTRFS_DEV_ITEMS_OBJECTID)
				break;
			if (found_key.type == BTRFS_DEV_ITEM_KEY) {
				struct btrfs_dev_item *dev_item;
				dev_item = btrfs_item_ptr(leaf, slot,
						  struct btrfs_dev_item);
2467
				ret = read_one_dev(root, leaf, dev_item);
2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488
				BUG_ON(ret);
			}
		} else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) {
			struct btrfs_chunk *chunk;
			chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
			ret = read_one_chunk(root, &found_key, leaf, chunk);
		}
		path->slots[0]++;
	}
	if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
		key.objectid = 0;
		btrfs_release_path(root, path);
		goto again;
	}

	btrfs_free_path(path);
	ret = 0;
error:
	return ret;
}