inode.c 62.7 KB
Newer Older
C
Chris Mason 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * 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.
 */

C
Chris Mason 已提交
19 20 21 22 23 24 25 26 27 28 29 30 31 32
#include <linux/buffer_head.h>
#include <linux/fs.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/smp_lock.h>
#include <linux/backing-dev.h>
#include <linux/mpage.h>
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/statfs.h>
#include <linux/compat.h>
C
Chris Mason 已提交
33
#include <linux/bit_spinlock.h>
34
#include <linux/version.h>
C
Chris Mason 已提交
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "ioctl.h"
#include "print-tree.h"

struct btrfs_iget_args {
	u64 ino;
	struct btrfs_root *root;
};

static struct inode_operations btrfs_dir_inode_operations;
static struct inode_operations btrfs_symlink_inode_operations;
static struct inode_operations btrfs_dir_ro_inode_operations;
J
Josef Bacik 已提交
50
static struct inode_operations btrfs_special_inode_operations;
C
Chris Mason 已提交
51 52 53 54
static struct inode_operations btrfs_file_inode_operations;
static struct address_space_operations btrfs_aops;
static struct address_space_operations btrfs_symlink_aops;
static struct file_operations btrfs_dir_file_operations;
55
static struct extent_map_ops btrfs_extent_map_ops;
C
Chris Mason 已提交
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73

static struct kmem_cache *btrfs_inode_cachep;
struct kmem_cache *btrfs_trans_handle_cachep;
struct kmem_cache *btrfs_transaction_cachep;
struct kmem_cache *btrfs_bit_radix_cachep;
struct kmem_cache *btrfs_path_cachep;

#define S_SHIFT 12
static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = {
	[S_IFREG >> S_SHIFT]	= BTRFS_FT_REG_FILE,
	[S_IFDIR >> S_SHIFT]	= BTRFS_FT_DIR,
	[S_IFCHR >> S_SHIFT]	= BTRFS_FT_CHRDEV,
	[S_IFBLK >> S_SHIFT]	= BTRFS_FT_BLKDEV,
	[S_IFIFO >> S_SHIFT]	= BTRFS_FT_FIFO,
	[S_IFSOCK >> S_SHIFT]	= BTRFS_FT_SOCK,
	[S_IFLNK >> S_SHIFT]	= BTRFS_FT_SYMLINK,
};

74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106
static int run_delalloc_range(struct inode *inode, u64 start, u64 end)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	struct btrfs_key ins;
	u64 alloc_hint = 0;
	u64 num_blocks;
	int ret;
	u64 blocksize = 1 << inode->i_blkbits;

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, inode);
	BUG_ON(!trans);
	num_blocks = (end - start + blocksize) & ~(blocksize - 1);
	ret = btrfs_drop_extents(trans, root, inode,
				 start, start + num_blocks, &alloc_hint);
	num_blocks = num_blocks >> inode->i_blkbits;
	ret = btrfs_alloc_extent(trans, root, inode->i_ino, num_blocks, 0,
				 alloc_hint, (u64)-1, &ins, 1);
	if (ret) {
		WARN_ON(1);
		goto out;
	}
	ret = btrfs_insert_file_extent(trans, root, inode->i_ino,
				       start, ins.objectid, ins.offset,
				       ins.offset);
out:
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;
}

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190
int btrfs_writepage_io_hook(struct page *page, u64 start, u64 end)
{
	struct inode *inode = page->mapping->host;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	char *kaddr;
	int ret;
	u64 page_start = page->index << PAGE_CACHE_SHIFT;
	size_t offset = start - page_start;

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, inode);
	kaddr = kmap(page);
	btrfs_csum_file_block(trans, root, inode->i_ino,
			      start, kaddr + offset, end - start + 1);
	kunmap(page);
	ret = btrfs_end_transaction(trans, root);
	BUG_ON(ret);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;
}

int btrfs_readpage_io_hook(struct page *page, u64 start, u64 end)
{
	int ret = 0;
	struct inode *inode = page->mapping->host;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	struct btrfs_csum_item *item;
	struct btrfs_path *path = NULL;
	u64 private;

	mutex_lock(&root->fs_info->fs_mutex);
	path = btrfs_alloc_path();
	item = btrfs_lookup_csum(NULL, root, path, inode->i_ino, start, 0);
	if (IS_ERR(item)) {
		ret = PTR_ERR(item);
		/* a csum that isn't present is a preallocated region. */
		if (ret == -ENOENT || ret == -EFBIG)
			ret = 0;
		private = 0;
		goto out;
	}
	memcpy((char *)&private, &item->csum, BTRFS_CRC32_SIZE);
	set_state_private(em_tree, start, private);
out:
	if (path)
		btrfs_free_path(path);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;
}

int btrfs_readpage_end_io_hook(struct page *page, u64 start, u64 end)
{
	char csum[BTRFS_CRC32_SIZE];
	size_t offset = start - (page->index << PAGE_CACHE_SHIFT);
	struct inode *inode = page->mapping->host;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	char *kaddr;
	u64 private;
	int ret;

	ret = get_state_private(em_tree, start, &private);
	kaddr = kmap_atomic(page, KM_IRQ0);
	if (ret) {
		goto zeroit;
	}
	ret = btrfs_csum_data(root, kaddr + offset, end - start + 1, csum);
	BUG_ON(ret);
	if (memcmp(csum, &private, BTRFS_CRC32_SIZE)) {
		goto zeroit;
	}
	kunmap_atomic(kaddr, KM_IRQ0);
	return 0;

zeroit:
	printk("btrfs csum failed ino %lu off %llu\n",
	       page->mapping->host->i_ino, (unsigned long long)start);
	memset(kaddr + offset, 1, end - start + 1); flush_dcache_page(page);
	kunmap_atomic(kaddr, KM_IRQ0);
	return 0;
}
191

C
Chris Mason 已提交
192 193 194 195 196 197 198
void btrfs_read_locked_inode(struct inode *inode)
{
	struct btrfs_path *path;
	struct btrfs_inode_item *inode_item;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_key location;
	u64 alloc_group_block;
J
Josef Bacik 已提交
199
	u32 rdev;
C
Chris Mason 已提交
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
	int ret;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	mutex_lock(&root->fs_info->fs_mutex);

	memcpy(&location, &BTRFS_I(inode)->location, sizeof(location));
	ret = btrfs_lookup_inode(NULL, root, path, &location, 0);
	if (ret) {
		btrfs_free_path(path);
		goto make_bad;
	}
	inode_item = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]),
				  path->slots[0],
				  struct btrfs_inode_item);

	inode->i_mode = btrfs_inode_mode(inode_item);
	inode->i_nlink = btrfs_inode_nlink(inode_item);
	inode->i_uid = btrfs_inode_uid(inode_item);
	inode->i_gid = btrfs_inode_gid(inode_item);
	inode->i_size = btrfs_inode_size(inode_item);
	inode->i_atime.tv_sec = btrfs_timespec_sec(&inode_item->atime);
	inode->i_atime.tv_nsec = btrfs_timespec_nsec(&inode_item->atime);
	inode->i_mtime.tv_sec = btrfs_timespec_sec(&inode_item->mtime);
	inode->i_mtime.tv_nsec = btrfs_timespec_nsec(&inode_item->mtime);
	inode->i_ctime.tv_sec = btrfs_timespec_sec(&inode_item->ctime);
	inode->i_ctime.tv_nsec = btrfs_timespec_nsec(&inode_item->ctime);
	inode->i_blocks = btrfs_inode_nblocks(inode_item);
	inode->i_generation = btrfs_inode_generation(inode_item);
J
Josef Bacik 已提交
229 230
	inode->i_rdev = 0;
	rdev = btrfs_inode_rdev(inode_item);
C
Chris Mason 已提交
231 232 233 234 235 236 237 238 239 240 241 242
	alloc_group_block = btrfs_inode_block_group(inode_item);
	BTRFS_I(inode)->block_group = btrfs_lookup_block_group(root->fs_info,
						       alloc_group_block);

	btrfs_free_path(path);
	inode_item = NULL;

	mutex_unlock(&root->fs_info->fs_mutex);

	switch (inode->i_mode & S_IFMT) {
	case S_IFREG:
		inode->i_mapping->a_ops = &btrfs_aops;
243
		BTRFS_I(inode)->extent_tree.ops = &btrfs_extent_map_ops;
C
Chris Mason 已提交
244 245 246 247 248 249 250 251 252 253 254 255 256 257
		inode->i_fop = &btrfs_file_operations;
		inode->i_op = &btrfs_file_inode_operations;
		break;
	case S_IFDIR:
		inode->i_fop = &btrfs_dir_file_operations;
		if (root == root->fs_info->tree_root)
			inode->i_op = &btrfs_dir_ro_inode_operations;
		else
			inode->i_op = &btrfs_dir_inode_operations;
		break;
	case S_IFLNK:
		inode->i_op = &btrfs_symlink_inode_operations;
		inode->i_mapping->a_ops = &btrfs_symlink_aops;
		break;
J
Josef Bacik 已提交
258 259 260
	default:
		init_special_inode(inode, inode->i_mode, rdev);
		break;
C
Chris Mason 已提交
261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
	}
	return;

make_bad:
	btrfs_release_path(root, path);
	btrfs_free_path(path);
	mutex_unlock(&root->fs_info->fs_mutex);
	make_bad_inode(inode);
}

static void fill_inode_item(struct btrfs_inode_item *item,
			    struct inode *inode)
{
	btrfs_set_inode_uid(item, inode->i_uid);
	btrfs_set_inode_gid(item, inode->i_gid);
	btrfs_set_inode_size(item, inode->i_size);
	btrfs_set_inode_mode(item, inode->i_mode);
	btrfs_set_inode_nlink(item, inode->i_nlink);
	btrfs_set_timespec_sec(&item->atime, inode->i_atime.tv_sec);
	btrfs_set_timespec_nsec(&item->atime, inode->i_atime.tv_nsec);
	btrfs_set_timespec_sec(&item->mtime, inode->i_mtime.tv_sec);
	btrfs_set_timespec_nsec(&item->mtime, inode->i_mtime.tv_nsec);
	btrfs_set_timespec_sec(&item->ctime, inode->i_ctime.tv_sec);
	btrfs_set_timespec_nsec(&item->ctime, inode->i_ctime.tv_nsec);
	btrfs_set_inode_nblocks(item, inode->i_blocks);
	btrfs_set_inode_generation(item, inode->i_generation);
J
Josef Bacik 已提交
287
	btrfs_set_inode_rdev(item, inode->i_rdev);
C
Chris Mason 已提交
288 289 290 291
	btrfs_set_inode_block_group(item,
				    BTRFS_I(inode)->block_group->key.objectid);
}

292
int btrfs_update_inode(struct btrfs_trans_handle *trans,
C
Chris Mason 已提交
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
			      struct btrfs_root *root,
			      struct inode *inode)
{
	struct btrfs_inode_item *inode_item;
	struct btrfs_path *path;
	int ret;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	ret = btrfs_lookup_inode(trans, root, path,
				 &BTRFS_I(inode)->location, 1);
	if (ret) {
		if (ret > 0)
			ret = -ENOENT;
		goto failed;
	}

	inode_item = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]),
				  path->slots[0],
				  struct btrfs_inode_item);

	fill_inode_item(inode_item, inode);
	btrfs_mark_buffer_dirty(path->nodes[0]);
316
	btrfs_set_inode_last_trans(trans, inode);
C
Chris Mason 已提交
317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
	ret = 0;
failed:
	btrfs_release_path(root, path);
	btrfs_free_path(path);
	return ret;
}


static int btrfs_unlink_trans(struct btrfs_trans_handle *trans,
			      struct btrfs_root *root,
			      struct inode *dir,
			      struct dentry *dentry)
{
	struct btrfs_path *path;
	const char *name = dentry->d_name.name;
	int name_len = dentry->d_name.len;
	int ret = 0;
	u64 objectid;
	struct btrfs_dir_item *di;

	path = btrfs_alloc_path();
338 339 340 341 342
	if (!path) {
		ret = -ENOMEM;
		goto err;
	}

C
Chris Mason 已提交
343 344 345 346 347 348 349 350 351 352 353 354
	di = btrfs_lookup_dir_item(trans, root, path, dir->i_ino,
				    name, name_len, -1);
	if (IS_ERR(di)) {
		ret = PTR_ERR(di);
		goto err;
	}
	if (!di) {
		ret = -ENOENT;
		goto err;
	}
	objectid = btrfs_disk_key_objectid(&di->location);
	ret = btrfs_delete_one_dir_name(trans, root, path, di);
355 356
	if (ret)
		goto err;
C
Chris Mason 已提交
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
	btrfs_release_path(root, path);

	di = btrfs_lookup_dir_index_item(trans, root, path, dir->i_ino,
					 objectid, name, name_len, -1);
	if (IS_ERR(di)) {
		ret = PTR_ERR(di);
		goto err;
	}
	if (!di) {
		ret = -ENOENT;
		goto err;
	}
	ret = btrfs_delete_one_dir_name(trans, root, path, di);

	dentry->d_inode->i_ctime = dir->i_ctime;
err:
	btrfs_free_path(path);
	if (!ret) {
		dir->i_size -= name_len * 2;
376
		dir->i_mtime = dir->i_ctime = CURRENT_TIME;
C
Chris Mason 已提交
377 378
		btrfs_update_inode(trans, root, dir);
		drop_nlink(dentry->d_inode);
379
		ret = btrfs_update_inode(trans, root, dentry->d_inode);
C
Chris Mason 已提交
380 381 382 383 384 385 386 387 388 389
		dir->i_sb->s_dirt = 1;
	}
	return ret;
}

static int btrfs_unlink(struct inode *dir, struct dentry *dentry)
{
	struct btrfs_root *root;
	struct btrfs_trans_handle *trans;
	int ret;
390
	unsigned long nr;
C
Chris Mason 已提交
391 392 393 394 395 396

	root = BTRFS_I(dir)->root;
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);
	ret = btrfs_unlink_trans(trans, root, dir, dentry);
397
	nr = trans->blocks_used;
C
Chris Mason 已提交
398 399
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
400
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416
	return ret;
}

static int btrfs_rmdir(struct inode *dir, struct dentry *dentry)
{
	struct inode *inode = dentry->d_inode;
	int err;
	int ret;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_trans_handle *trans;
	struct btrfs_key found_key;
	int found_type;
	struct btrfs_leaf *leaf;
	char *goodnames = "..";
417
	unsigned long nr;
C
Chris Mason 已提交
418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472

	path = btrfs_alloc_path();
	BUG_ON(!path);
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);
	key.objectid = inode->i_ino;
	key.offset = (u64)-1;
	key.flags = (u32)-1;
	while(1) {
		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
		if (ret < 0) {
			err = ret;
			goto out;
		}
		BUG_ON(ret == 0);
		if (path->slots[0] == 0) {
			err = -ENOENT;
			goto out;
		}
		path->slots[0]--;
		leaf = btrfs_buffer_leaf(path->nodes[0]);
		btrfs_disk_key_to_cpu(&found_key,
				      &leaf->items[path->slots[0]].key);
		found_type = btrfs_key_type(&found_key);
		if (found_key.objectid != inode->i_ino) {
			err = -ENOENT;
			goto out;
		}
		if ((found_type != BTRFS_DIR_ITEM_KEY &&
		     found_type != BTRFS_DIR_INDEX_KEY) ||
	            (!btrfs_match_dir_item_name(root, path, goodnames, 2) &&
	            !btrfs_match_dir_item_name(root, path, goodnames, 1))) {
			err = -ENOTEMPTY;
			goto out;
		}
		ret = btrfs_del_item(trans, root, path);
		BUG_ON(ret);

		if (found_type == BTRFS_DIR_ITEM_KEY && found_key.offset == 1)
			break;
		btrfs_release_path(root, path);
	}
	ret = 0;
	btrfs_release_path(root, path);

	/* now the directory is empty */
	err = btrfs_unlink_trans(trans, root, dir, dentry);
	if (!err) {
		inode->i_size = 0;
	}
out:
	btrfs_release_path(root, path);
	btrfs_free_path(path);
	mutex_unlock(&root->fs_info->fs_mutex);
473
	nr = trans->blocks_used;
C
Chris Mason 已提交
474
	ret = btrfs_end_transaction(trans, root);
475
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493
	if (ret && !err)
		err = ret;
	return err;
}

static int btrfs_free_inode(struct btrfs_trans_handle *trans,
			    struct btrfs_root *root,
			    struct inode *inode)
{
	struct btrfs_path *path;
	int ret;

	clear_inode(inode);

	path = btrfs_alloc_path();
	BUG_ON(!path);
	ret = btrfs_lookup_inode(trans, root, path,
				 &BTRFS_I(inode)->location, -1);
494 495 496 497
	if (ret > 0)
		ret = -ENOENT;
	if (!ret)
		ret = btrfs_del_item(trans, root, path);
C
Chris Mason 已提交
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
	btrfs_free_path(path);
	return ret;
}

/*
 * this can truncate away extent items, csum items and directory items.
 * It starts at a high offset and removes keys until it can't find
 * any higher than i_size.
 *
 * csum items that cross the new i_size are truncated to the new size
 * as well.
 */
static int btrfs_truncate_in_trans(struct btrfs_trans_handle *trans,
				   struct btrfs_root *root,
				   struct inode *inode)
{
	int ret;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct btrfs_disk_key *found_key;
	u32 found_type;
	struct btrfs_leaf *leaf;
	struct btrfs_file_extent_item *fi;
	u64 extent_start = 0;
	u64 extent_num_blocks = 0;
	u64 item_end = 0;
	int found_extent;
	int del_item;

527
	btrfs_drop_extent_cache(inode, inode->i_size, (u64)-1);
C
Chris Mason 已提交
528
	path = btrfs_alloc_path();
529
	path->reada = -1;
C
Chris Mason 已提交
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
	BUG_ON(!path);
	/* FIXME, add redo link to tree so we don't leak on crash */
	key.objectid = inode->i_ino;
	key.offset = (u64)-1;
	key.flags = (u32)-1;
	while(1) {
		btrfs_init_path(path);
		fi = NULL;
		ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
		if (ret < 0) {
			goto error;
		}
		if (ret > 0) {
			BUG_ON(path->slots[0] == 0);
			path->slots[0]--;
		}
		leaf = btrfs_buffer_leaf(path->nodes[0]);
		found_key = &leaf->items[path->slots[0]].key;
		found_type = btrfs_disk_key_type(found_key);

		if (btrfs_disk_key_objectid(found_key) != inode->i_ino)
			break;
		if (found_type != BTRFS_CSUM_ITEM_KEY &&
		    found_type != BTRFS_DIR_ITEM_KEY &&
		    found_type != BTRFS_DIR_INDEX_KEY &&
		    found_type != BTRFS_EXTENT_DATA_KEY)
			break;

		item_end = btrfs_disk_key_offset(found_key);
		if (found_type == BTRFS_EXTENT_DATA_KEY) {
			fi = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]),
					    path->slots[0],
					    struct btrfs_file_extent_item);
			if (btrfs_file_extent_type(fi) !=
			    BTRFS_FILE_EXTENT_INLINE) {
				item_end += btrfs_file_extent_num_blocks(fi) <<
						inode->i_blkbits;
			}
		}
		if (found_type == BTRFS_CSUM_ITEM_KEY) {
			ret = btrfs_csum_truncate(trans, root, path,
						  inode->i_size);
			BUG_ON(ret);
		}
		if (item_end < inode->i_size) {
575 576 577 578 579 580 581 582
			if (found_type == BTRFS_DIR_ITEM_KEY) {
				found_type = BTRFS_INODE_ITEM_KEY;
			} else if (found_type == BTRFS_EXTENT_ITEM_KEY) {
				found_type = BTRFS_CSUM_ITEM_KEY;
			} else if (found_type) {
				found_type--;
			} else {
				break;
C
Chris Mason 已提交
583
			}
584
			btrfs_set_key_type(&key, found_type);
585
			continue;
C
Chris Mason 已提交
586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608
		}
		if (btrfs_disk_key_offset(found_key) >= inode->i_size)
			del_item = 1;
		else
			del_item = 0;
		found_extent = 0;

		/* FIXME, shrink the extent if the ref count is only 1 */
		if (found_type == BTRFS_EXTENT_DATA_KEY &&
			   btrfs_file_extent_type(fi) !=
			   BTRFS_FILE_EXTENT_INLINE) {
			u64 num_dec;
			if (!del_item) {
				u64 orig_num_blocks =
					btrfs_file_extent_num_blocks(fi);
				extent_num_blocks = inode->i_size -
					btrfs_disk_key_offset(found_key) +
					root->blocksize - 1;
				extent_num_blocks >>= inode->i_blkbits;
				btrfs_set_file_extent_num_blocks(fi,
							 extent_num_blocks);
				inode->i_blocks -= (orig_num_blocks -
					extent_num_blocks) << 3;
C
Chris Mason 已提交
609
				btrfs_mark_buffer_dirty(path->nodes[0]);
C
Chris Mason 已提交
610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
			} else {
				extent_start =
					btrfs_file_extent_disk_blocknr(fi);
				extent_num_blocks =
					btrfs_file_extent_disk_num_blocks(fi);
				/* FIXME blocksize != 4096 */
				num_dec = btrfs_file_extent_num_blocks(fi) << 3;
				if (extent_start != 0) {
					found_extent = 1;
					inode->i_blocks -= num_dec;
				}
			}
		}
		if (del_item) {
			ret = btrfs_del_item(trans, root, path);
625 626
			if (ret)
				goto error;
C
Chris Mason 已提交
627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644
		} else {
			break;
		}
		btrfs_release_path(root, path);
		if (found_extent) {
			ret = btrfs_free_extent(trans, root, extent_start,
						extent_num_blocks, 0);
			BUG_ON(ret);
		}
	}
	ret = 0;
error:
	btrfs_release_path(root, path);
	btrfs_free_path(path);
	inode->i_sb->s_dirt = 1;
	return ret;
}

645
static int btrfs_cow_one_page(struct inode *inode, struct page *page,
646 647 648 649
			      size_t zero_start)
{
	char *kaddr;
	int ret = 0;
650
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
651
	u64 page_start = page->index << PAGE_CACHE_SHIFT;
652
	u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
653 654 655 656

	if (!PagePrivate(page)) {
		SetPagePrivate(page);
		set_page_private(page, 1);
657
		WARN_ON(!page->mapping->a_ops->invalidatepage);
658 659 660
		page_cache_get(page);
	}

661 662 663
	lock_extent(em_tree, page_start, page_end, GFP_NOFS);
	set_extent_delalloc(&BTRFS_I(inode)->extent_tree, page_start,
			    page_end, GFP_NOFS);
664
	if (zero_start != PAGE_CACHE_SIZE) {
665
		kaddr = kmap(page);
666 667
		memset(kaddr + zero_start, 0, PAGE_CACHE_SIZE - zero_start);
		flush_dcache_page(page);
668
		kunmap(page);
669
	}
670 671
	set_page_dirty(page);
	unlock_extent(em_tree, page_start, page_end, GFP_NOFS);
672 673 674 675

	return ret;
}

C
Chris Mason 已提交
676 677 678 679 680 681 682 683 684 685 686 687
/*
 * taken from block_truncate_page, but does cow as it zeros out
 * any bytes left in the last page in the file.
 */
static int btrfs_truncate_page(struct address_space *mapping, loff_t from)
{
	struct inode *inode = mapping->host;
	unsigned blocksize = 1 << inode->i_blkbits;
	pgoff_t index = from >> PAGE_CACHE_SHIFT;
	unsigned offset = from & (PAGE_CACHE_SIZE-1);
	struct page *page;
	int ret = 0;
688
	u64 page_start;
C
Chris Mason 已提交
689 690 691 692

	if ((offset & (blocksize - 1)) == 0)
		goto out;

693
	down_read(&BTRFS_I(inode)->root->snap_sem);
C
Chris Mason 已提交
694 695 696 697 698
	ret = -ENOMEM;
	page = grab_cache_page(mapping, index);
	if (!page)
		goto out;
	if (!PageUptodate(page)) {
C
Chris Mason 已提交
699
		ret = btrfs_readpage(NULL, page);
C
Chris Mason 已提交
700 701 702 703 704 705
		lock_page(page);
		if (!PageUptodate(page)) {
			ret = -EIO;
			goto out;
		}
	}
706 707
	page_start = page->index << PAGE_CACHE_SHIFT;

708
	ret = btrfs_cow_one_page(inode, page, offset);
C
Chris Mason 已提交
709 710 711

	unlock_page(page);
	page_cache_release(page);
712
	up_read(&BTRFS_I(inode)->root->snap_sem);
C
Chris Mason 已提交
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729
out:
	return ret;
}

static int btrfs_setattr(struct dentry *dentry, struct iattr *attr)
{
	struct inode *inode = dentry->d_inode;
	int err;

	err = inode_change_ok(inode, attr);
	if (err)
		return err;

	if (S_ISREG(inode->i_mode) &&
	    attr->ia_valid & ATTR_SIZE && attr->ia_size > inode->i_size) {
		struct btrfs_trans_handle *trans;
		struct btrfs_root *root = BTRFS_I(inode)->root;
C
Chris Mason 已提交
730 731
		struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;

C
Chris Mason 已提交
732 733
		u64 mask = root->blocksize - 1;
		u64 pos = (inode->i_size + mask) & ~mask;
C
Chris Mason 已提交
734
		u64 block_end = attr->ia_size | mask;
C
Chris Mason 已提交
735
		u64 hole_size;
C
Chris Mason 已提交
736
		u64 alloc_hint;
C
Chris Mason 已提交
737 738 739 740 741 742

		if (attr->ia_size <= pos)
			goto out;

		btrfs_truncate_page(inode->i_mapping, inode->i_size);

C
Chris Mason 已提交
743
		lock_extent(em_tree, pos, block_end, GFP_NOFS);
C
Chris Mason 已提交
744 745 746 747 748
		hole_size = (attr->ia_size - pos + mask) & ~mask;

		mutex_lock(&root->fs_info->fs_mutex);
		trans = btrfs_start_transaction(root, 1);
		btrfs_set_trans_block_group(trans, inode);
C
Chris Mason 已提交
749 750 751 752 753
		err = btrfs_drop_extents(trans, root, inode,
					 pos, pos + hole_size, &alloc_hint);

		hole_size >>= inode->i_blkbits;

C
Chris Mason 已提交
754 755 756 757
		err = btrfs_insert_file_extent(trans, root, inode->i_ino,
					       pos, 0, 0, hole_size);
		btrfs_end_transaction(trans, root);
		mutex_unlock(&root->fs_info->fs_mutex);
C
Chris Mason 已提交
758
		unlock_extent(em_tree, pos, block_end, GFP_NOFS);
759 760
		if (err)
			return err;
C
Chris Mason 已提交
761 762 763 764 765 766 767 768 769 770
	}
out:
	err = inode_setattr(inode, attr);

	return err;
}
void btrfs_delete_inode(struct inode *inode)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(inode)->root;
771
	unsigned long nr;
C
Chris Mason 已提交
772 773 774 775 776 777 778 779 780 781 782
	int ret;

	truncate_inode_pages(&inode->i_data, 0);
	if (is_bad_inode(inode)) {
		goto no_delete;
	}
	inode->i_size = 0;
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, inode);
	ret = btrfs_truncate_in_trans(trans, root, inode);
783 784 785 786 787
	if (ret)
		goto no_delete_lock;
	ret = btrfs_free_inode(trans, root, inode);
	if (ret)
		goto no_delete_lock;
788
	nr = trans->blocks_used;
C
Chris Mason 已提交
789 790
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
791
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
792
	return;
793 794

no_delete_lock:
795
	nr = trans->blocks_used;
796 797
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
798
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
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
no_delete:
	clear_inode(inode);
}

/*
 * this returns the key found in the dir entry in the location pointer.
 * If no dir entries were found, location->objectid is 0.
 */
static int btrfs_inode_by_name(struct inode *dir, struct dentry *dentry,
			       struct btrfs_key *location)
{
	const char *name = dentry->d_name.name;
	int namelen = dentry->d_name.len;
	struct btrfs_dir_item *di;
	struct btrfs_path *path;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	int ret;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	di = btrfs_lookup_dir_item(NULL, root, path, dir->i_ino, name,
				    namelen, 0);
	if (!di || IS_ERR(di)) {
		location->objectid = 0;
		ret = 0;
		goto out;
	}
	btrfs_disk_key_to_cpu(location, &di->location);
out:
	btrfs_release_path(root, path);
	btrfs_free_path(path);
	return ret;
}

/*
 * when we hit a tree root in a directory, the btrfs part of the inode
 * needs to be changed to reflect the root directory of the tree root.  This
 * is kind of like crossing a mount point.
 */
static int fixup_tree_root_location(struct btrfs_root *root,
			     struct btrfs_key *location,
840 841
			     struct btrfs_root **sub_root,
			     struct dentry *dentry)
C
Chris Mason 已提交
842 843 844 845 846 847 848 849 850 851 852 853 854
{
	struct btrfs_path *path;
	struct btrfs_root_item *ri;

	if (btrfs_key_type(location) != BTRFS_ROOT_ITEM_KEY)
		return 0;
	if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
		return 0;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	mutex_lock(&root->fs_info->fs_mutex);

855 856 857
	*sub_root = btrfs_read_fs_root(root->fs_info, location,
					dentry->d_name.name,
					dentry->d_name.len);
C
Chris Mason 已提交
858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876
	if (IS_ERR(*sub_root))
		return PTR_ERR(*sub_root);

	ri = &(*sub_root)->root_item;
	location->objectid = btrfs_root_dirid(ri);
	location->flags = 0;
	btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);
	location->offset = 0;

	btrfs_free_path(path);
	mutex_unlock(&root->fs_info->fs_mutex);
	return 0;
}

static int btrfs_init_locked_inode(struct inode *inode, void *p)
{
	struct btrfs_iget_args *args = p;
	inode->i_ino = args->ino;
	BTRFS_I(inode)->root = args->root;
877 878
	extent_map_tree_init(&BTRFS_I(inode)->extent_tree,
			     inode->i_mapping, GFP_NOFS);
C
Chris Mason 已提交
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921
	return 0;
}

static int btrfs_find_actor(struct inode *inode, void *opaque)
{
	struct btrfs_iget_args *args = opaque;
	return (args->ino == inode->i_ino &&
		args->root == BTRFS_I(inode)->root);
}

struct inode *btrfs_iget_locked(struct super_block *s, u64 objectid,
				struct btrfs_root *root)
{
	struct inode *inode;
	struct btrfs_iget_args args;
	args.ino = objectid;
	args.root = root;

	inode = iget5_locked(s, objectid, btrfs_find_actor,
			     btrfs_init_locked_inode,
			     (void *)&args);
	return inode;
}

static struct dentry *btrfs_lookup(struct inode *dir, struct dentry *dentry,
				   struct nameidata *nd)
{
	struct inode * inode;
	struct btrfs_inode *bi = BTRFS_I(dir);
	struct btrfs_root *root = bi->root;
	struct btrfs_root *sub_root = root;
	struct btrfs_key location;
	int ret;

	if (dentry->d_name.len > BTRFS_NAME_LEN)
		return ERR_PTR(-ENAMETOOLONG);
	mutex_lock(&root->fs_info->fs_mutex);
	ret = btrfs_inode_by_name(dir, dentry, &location);
	mutex_unlock(&root->fs_info->fs_mutex);
	if (ret < 0)
		return ERR_PTR(ret);
	inode = NULL;
	if (location.objectid) {
922 923
		ret = fixup_tree_root_location(root, &location, &sub_root,
						dentry);
C
Chris Mason 已提交
924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980
		if (ret < 0)
			return ERR_PTR(ret);
		if (ret > 0)
			return ERR_PTR(-ENOENT);
		inode = btrfs_iget_locked(dir->i_sb, location.objectid,
					  sub_root);
		if (!inode)
			return ERR_PTR(-EACCES);
		if (inode->i_state & I_NEW) {
			/* the inode and parent dir are two different roots */
			if (sub_root != root) {
				igrab(inode);
				sub_root->inode = inode;
			}
			BTRFS_I(inode)->root = sub_root;
			memcpy(&BTRFS_I(inode)->location, &location,
			       sizeof(location));
			btrfs_read_locked_inode(inode);
			unlock_new_inode(inode);
		}
	}
	return d_splice_alias(inode, dentry);
}

static unsigned char btrfs_filetype_table[] = {
	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
};

static int btrfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
{
	struct inode *inode = filp->f_path.dentry->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_item *item;
	struct btrfs_dir_item *di;
	struct btrfs_key key;
	struct btrfs_path *path;
	int ret;
	u32 nritems;
	struct btrfs_leaf *leaf;
	int slot;
	int advance;
	unsigned char d_type;
	int over = 0;
	u32 di_cur;
	u32 di_total;
	u32 di_len;
	int key_type = BTRFS_DIR_INDEX_KEY;

	/* FIXME, use a real flag for deciding about the key type */
	if (root->fs_info->tree_root == root)
		key_type = BTRFS_DIR_ITEM_KEY;
	mutex_lock(&root->fs_info->fs_mutex);
	key.objectid = inode->i_ino;
	key.flags = 0;
	btrfs_set_key_type(&key, key_type);
	key.offset = filp->f_pos;
	path = btrfs_alloc_path();
981
	path->reada = 2;
C
Chris Mason 已提交
982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 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 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0)
		goto err;
	advance = 0;
	while(1) {
		leaf = btrfs_buffer_leaf(path->nodes[0]);
		nritems = btrfs_header_nritems(&leaf->header);
		slot = path->slots[0];
		if (advance || slot >= nritems) {
			if (slot >= nritems -1) {
				ret = btrfs_next_leaf(root, path);
				if (ret)
					break;
				leaf = btrfs_buffer_leaf(path->nodes[0]);
				nritems = btrfs_header_nritems(&leaf->header);
				slot = path->slots[0];
			} else {
				slot++;
				path->slots[0]++;
			}
		}
		advance = 1;
		item = leaf->items + slot;
		if (btrfs_disk_key_objectid(&item->key) != key.objectid)
			break;
		if (btrfs_disk_key_type(&item->key) != key_type)
			break;
		if (btrfs_disk_key_offset(&item->key) < filp->f_pos)
			continue;
		filp->f_pos = btrfs_disk_key_offset(&item->key);
		advance = 1;
		di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
		di_cur = 0;
		di_total = btrfs_item_size(leaf->items + slot);
		while(di_cur < di_total) {
			d_type = btrfs_filetype_table[btrfs_dir_type(di)];
			over = filldir(dirent, (const char *)(di + 1),
				       btrfs_dir_name_len(di),
				       btrfs_disk_key_offset(&item->key),
				       btrfs_disk_key_objectid(&di->location),
				       d_type);
			if (over)
				goto nopos;
			di_len = btrfs_dir_name_len(di) + sizeof(*di);
			di_cur += di_len;
			di = (struct btrfs_dir_item *)((char *)di + di_len);
		}
	}
	filp->f_pos++;
nopos:
	ret = 0;
err:
	btrfs_release_path(root, path);
	btrfs_free_path(path);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;
}

int btrfs_write_inode(struct inode *inode, int wait)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	int ret = 0;

	if (wait) {
		mutex_lock(&root->fs_info->fs_mutex);
		trans = btrfs_start_transaction(root, 1);
		btrfs_set_trans_block_group(trans, inode);
		ret = btrfs_commit_transaction(trans, root);
		mutex_unlock(&root->fs_info->fs_mutex);
	}
	return ret;
}

/*
1057
 * This is somewhat expensive, updating the tree every time the
C
Chris Mason 已提交
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
 * inode changes.  But, it is most likely to find the inode in cache.
 * FIXME, needs more benchmarking...there are no reasons other than performance
 * to keep or drop this code.
 */
void btrfs_dirty_inode(struct inode *inode)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, inode);
	btrfs_update_inode(trans, root, inode);
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
}

static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
				     struct btrfs_root *root,
				     u64 objectid,
				     struct btrfs_block_group_cache *group,
				     int mode)
{
	struct inode *inode;
	struct btrfs_inode_item inode_item;
	struct btrfs_key *location;
	int ret;
	int owner;

	inode = new_inode(root->fs_info->sb);
	if (!inode)
		return ERR_PTR(-ENOMEM);

1091 1092
	extent_map_tree_init(&BTRFS_I(inode)->extent_tree,
			     inode->i_mapping, GFP_NOFS);
C
Chris Mason 已提交
1093
	BTRFS_I(inode)->root = root;
1094

C
Chris Mason 已提交
1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
	if (mode & S_IFDIR)
		owner = 0;
	else
		owner = 1;
	group = btrfs_find_block_group(root, group, 0, 0, owner);
	BTRFS_I(inode)->block_group = group;

	inode->i_uid = current->fsuid;
	inode->i_gid = current->fsgid;
	inode->i_mode = mode;
	inode->i_ino = objectid;
	inode->i_blocks = 0;
	inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
	fill_inode_item(&inode_item, inode);
	location = &BTRFS_I(inode)->location;
	location->objectid = objectid;
	location->flags = 0;
	location->offset = 0;
	btrfs_set_key_type(location, BTRFS_INODE_ITEM_KEY);

	ret = btrfs_insert_inode(trans, root, objectid, &inode_item);
1116 1117
	if (ret)
		return ERR_PTR(ret);
C
Chris Mason 已提交
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132
	insert_inode_hash(inode);
	return inode;
}

static inline u8 btrfs_inode_type(struct inode *inode)
{
	return btrfs_type_by_mode[(inode->i_mode & S_IFMT) >> S_SHIFT];
}

static int btrfs_add_link(struct btrfs_trans_handle *trans,
			    struct dentry *dentry, struct inode *inode)
{
	int ret;
	struct btrfs_key key;
	struct btrfs_root *root = BTRFS_I(dentry->d_parent->d_inode)->root;
1133
	struct inode *parent_inode;
C
Chris Mason 已提交
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
	key.objectid = inode->i_ino;
	key.flags = 0;
	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);
	key.offset = 0;

	ret = btrfs_insert_dir_item(trans, root,
				    dentry->d_name.name, dentry->d_name.len,
				    dentry->d_parent->d_inode->i_ino,
				    &key, btrfs_inode_type(inode));
	if (ret == 0) {
1144 1145 1146
		parent_inode = dentry->d_parent->d_inode;
		parent_inode->i_size += dentry->d_name.len * 2;
		parent_inode->i_mtime = parent_inode->i_ctime = CURRENT_TIME;
C
Chris Mason 已提交
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165
		ret = btrfs_update_inode(trans, root,
					 dentry->d_parent->d_inode);
	}
	return ret;
}

static int btrfs_add_nondir(struct btrfs_trans_handle *trans,
			    struct dentry *dentry, struct inode *inode)
{
	int err = btrfs_add_link(trans, dentry, inode);
	if (!err) {
		d_instantiate(dentry, inode);
		return 0;
	}
	if (err > 0)
		err = -EEXIST;
	return err;
}

J
Josef Bacik 已提交
1166 1167 1168 1169 1170 1171 1172 1173 1174
static int btrfs_mknod(struct inode *dir, struct dentry *dentry,
			int mode, dev_t rdev)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct inode *inode;
	int err;
	int drop_inode = 0;
	u64 objectid;
1175
	unsigned long nr;
J
Josef Bacik 已提交
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

	if (!new_valid_dev(rdev))
		return -EINVAL;

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);

	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
	if (err) {
		err = -ENOSPC;
		goto out_unlock;
	}

	inode = btrfs_new_inode(trans, root, objectid,
				BTRFS_I(dir)->block_group, mode);
	err = PTR_ERR(inode);
	if (IS_ERR(inode))
		goto out_unlock;

	btrfs_set_trans_block_group(trans, inode);
	err = btrfs_add_nondir(trans, dentry, inode);
	if (err)
		drop_inode = 1;
	else {
		inode->i_op = &btrfs_special_inode_operations;
		init_special_inode(inode, inode->i_mode, rdev);
1203
		btrfs_update_inode(trans, root, inode);
J
Josef Bacik 已提交
1204 1205 1206 1207 1208
	}
	dir->i_sb->s_dirt = 1;
	btrfs_update_inode_block_group(trans, inode);
	btrfs_update_inode_block_group(trans, dir);
out_unlock:
1209
	nr = trans->blocks_used;
J
Josef Bacik 已提交
1210 1211 1212 1213 1214 1215 1216
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);

	if (drop_inode) {
		inode_dec_link_count(inode);
		iput(inode);
	}
1217
	btrfs_btree_balance_dirty(root, nr);
J
Josef Bacik 已提交
1218 1219 1220
	return err;
}

C
Chris Mason 已提交
1221 1222 1223 1224 1225 1226 1227 1228
static int btrfs_create(struct inode *dir, struct dentry *dentry,
			int mode, struct nameidata *nd)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct inode *inode;
	int err;
	int drop_inode = 0;
1229
	unsigned long nr;
C
Chris Mason 已提交
1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255
	u64 objectid;

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);

	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
	if (err) {
		err = -ENOSPC;
		goto out_unlock;
	}

	inode = btrfs_new_inode(trans, root, objectid,
				BTRFS_I(dir)->block_group, mode);
	err = PTR_ERR(inode);
	if (IS_ERR(inode))
		goto out_unlock;

	btrfs_set_trans_block_group(trans, inode);
	err = btrfs_add_nondir(trans, dentry, inode);
	if (err)
		drop_inode = 1;
	else {
		inode->i_mapping->a_ops = &btrfs_aops;
		inode->i_fop = &btrfs_file_operations;
		inode->i_op = &btrfs_file_inode_operations;
1256 1257
		extent_map_tree_init(&BTRFS_I(inode)->extent_tree,
				     inode->i_mapping, GFP_NOFS);
1258
		BTRFS_I(inode)->extent_tree.ops = &btrfs_extent_map_ops;
C
Chris Mason 已提交
1259 1260 1261 1262 1263
	}
	dir->i_sb->s_dirt = 1;
	btrfs_update_inode_block_group(trans, inode);
	btrfs_update_inode_block_group(trans, dir);
out_unlock:
1264
	nr = trans->blocks_used;
C
Chris Mason 已提交
1265 1266 1267 1268 1269 1270 1271
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);

	if (drop_inode) {
		inode_dec_link_count(inode);
		iput(inode);
	}
1272
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
1273 1274 1275 1276 1277 1278 1279 1280 1281
	return err;
}

static int btrfs_link(struct dentry *old_dentry, struct inode *dir,
		      struct dentry *dentry)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct inode *inode = old_dentry->d_inode;
1282
	unsigned long nr;
C
Chris Mason 已提交
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
	int err;
	int drop_inode = 0;

	if (inode->i_nlink == 0)
		return -ENOENT;

	inc_nlink(inode);
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);
	atomic_inc(&inode->i_count);
	err = btrfs_add_nondir(trans, dentry, inode);
	if (err)
		drop_inode = 1;
	dir->i_sb->s_dirt = 1;
	btrfs_update_inode_block_group(trans, dir);
1299 1300 1301
	err = btrfs_update_inode(trans, root, inode);
	if (err)
		drop_inode = 1;
C
Chris Mason 已提交
1302

1303
	nr = trans->blocks_used;
C
Chris Mason 已提交
1304 1305 1306 1307 1308 1309 1310
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);

	if (drop_inode) {
		inode_dec_link_count(inode);
		iput(inode);
	}
1311
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
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
	return err;
}

static int btrfs_make_empty_dir(struct btrfs_trans_handle *trans,
				struct btrfs_root *root,
				u64 objectid, u64 dirid)
{
	int ret;
	char buf[2];
	struct btrfs_key key;

	buf[0] = '.';
	buf[1] = '.';

	key.objectid = objectid;
	key.offset = 0;
	key.flags = 0;
	btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY);

	ret = btrfs_insert_dir_item(trans, root, buf, 1, objectid,
				    &key, BTRFS_FT_DIR);
	if (ret)
		goto error;
	key.objectid = dirid;
	ret = btrfs_insert_dir_item(trans, root, buf, 2, objectid,
				    &key, BTRFS_FT_DIR);
	if (ret)
		goto error;
error:
	return ret;
}

static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
	struct inode *inode;
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	int err = 0;
	int drop_on_err = 0;
	u64 objectid;
1352
	unsigned long nr = 1;
C
Chris Mason 已提交
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);
	if (IS_ERR(trans)) {
		err = PTR_ERR(trans);
		goto out_unlock;
	}

	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
	if (err) {
		err = -ENOSPC;
		goto out_unlock;
	}

	inode = btrfs_new_inode(trans, root, objectid,
				BTRFS_I(dir)->block_group, S_IFDIR | mode);
	if (IS_ERR(inode)) {
		err = PTR_ERR(inode);
		goto out_fail;
	}
	drop_on_err = 1;
	inode->i_op = &btrfs_dir_inode_operations;
	inode->i_fop = &btrfs_dir_file_operations;
	btrfs_set_trans_block_group(trans, inode);

	err = btrfs_make_empty_dir(trans, root, inode->i_ino, dir->i_ino);
	if (err)
		goto out_fail;

	inode->i_size = 6;
	err = btrfs_update_inode(trans, root, inode);
	if (err)
		goto out_fail;
	err = btrfs_add_link(trans, dentry, inode);
	if (err)
		goto out_fail;
	d_instantiate(dentry, inode);
	drop_on_err = 0;
	dir->i_sb->s_dirt = 1;
	btrfs_update_inode_block_group(trans, inode);
	btrfs_update_inode_block_group(trans, dir);

out_fail:
1397
	nr = trans->blocks_used;
C
Chris Mason 已提交
1398 1399 1400 1401 1402
	btrfs_end_transaction(trans, root);
out_unlock:
	mutex_unlock(&root->fs_info->fs_mutex);
	if (drop_on_err)
		iput(inode);
1403
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
1404 1405 1406
	return err;
}

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 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479
struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
				    size_t page_offset, u64 start, u64 end,
				    int create)
{
	int ret;
	int err = 0;
	u64 blocknr;
	u64 extent_start = 0;
	u64 extent_end = 0;
	u64 objectid = inode->i_ino;
	u32 found_type;
	int failed_insert = 0;
	struct btrfs_path *path;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_file_extent_item *item;
	struct btrfs_leaf *leaf;
	struct btrfs_disk_key *found_key;
	struct extent_map *em = NULL;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	struct btrfs_trans_handle *trans = NULL;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	mutex_lock(&root->fs_info->fs_mutex);

again:
	em = lookup_extent_mapping(em_tree, start, end);
	if (em) {
		goto out;
	}
	if (!em) {
		em = alloc_extent_map(GFP_NOFS);
		if (!em) {
			err = -ENOMEM;
			goto out;
		}
		em->start = 0;
		em->end = 0;
	}
	em->bdev = inode->i_sb->s_bdev;
	ret = btrfs_lookup_file_extent(NULL, root, path,
				       objectid, start, 0);
	if (ret < 0) {
		err = ret;
		goto out;
	}

	if (ret != 0) {
		if (path->slots[0] == 0)
			goto not_found;
		path->slots[0]--;
	}

	item = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]), path->slots[0],
			      struct btrfs_file_extent_item);
	leaf = btrfs_buffer_leaf(path->nodes[0]);
	blocknr = btrfs_file_extent_disk_blocknr(item);
	blocknr += btrfs_file_extent_offset(item);

	/* are we inside the extent that was found? */
	found_key = &leaf->items[path->slots[0]].key;
	found_type = btrfs_disk_key_type(found_key);
	if (btrfs_disk_key_objectid(found_key) != objectid ||
	    found_type != BTRFS_EXTENT_DATA_KEY) {
		goto not_found;
	}

	found_type = btrfs_file_extent_type(item);
	extent_start = btrfs_disk_key_offset(&leaf->items[path->slots[0]].key);
	if (found_type == BTRFS_FILE_EXTENT_REG) {
		extent_end = extent_start +
		       (btrfs_file_extent_num_blocks(item) << inode->i_blkbits);
		err = 0;
1480
		if (start < extent_start || start >= extent_end) {
1481 1482
			em->start = start;
			if (start < extent_start) {
1483 1484
				if (end < extent_start)
					goto not_found;
1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512
				em->end = extent_end - 1;
			} else {
				em->end = end;
			}
			goto not_found_em;
		}
		if (btrfs_file_extent_disk_blocknr(item) == 0) {
			em->start = extent_start;
			em->end = extent_end - 1;
			em->block_start = 0;
			em->block_end = 0;
			goto insert;
		}
		em->block_start = blocknr << inode->i_blkbits;
		em->block_end = em->block_start +
			(btrfs_file_extent_num_blocks(item) <<
			 inode->i_blkbits) - 1;
		em->start = extent_start;
		em->end = extent_end - 1;
		goto insert;
	} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
		char *ptr;
		char *map;
		u32 size;

		size = btrfs_file_extent_inline_len(leaf->items +
						    path->slots[0]);
		extent_end = extent_start + size;
1513
		if (start < extent_start || start >= extent_end) {
1514 1515
			em->start = start;
			if (start < extent_start) {
1516 1517
				if (end < extent_start)
					goto not_found;
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551
				em->end = extent_end - 1;
			} else {
				em->end = end;
			}
			goto not_found_em;
		}
		em->block_start = EXTENT_MAP_INLINE;
		em->block_end = EXTENT_MAP_INLINE;
		em->start = extent_start;
		em->end = extent_end - 1;
		if (!page) {
			goto insert;
		}
		ptr = btrfs_file_extent_inline_start(item);
		map = kmap(page);
		memcpy(map + page_offset, ptr, size);
		flush_dcache_page(result->b_page);
		kunmap(page);
		set_extent_uptodate(em_tree, extent_start,
				    extent_end, GFP_NOFS);
		goto insert;
	} else {
		printk("unkknown found_type %d\n", found_type);
		WARN_ON(1);
	}
not_found:
	em->start = start;
	em->end = end;
not_found_em:
	em->block_start = 0;
	em->block_end = 0;
insert:
	btrfs_release_path(root, path);
	if (em->start > start || em->end < start) {
1552
		printk("bad extent! em: [%Lu %Lu] passed [%Lu %Lu]\n", em->start, em->end, start, end);
1553 1554 1555 1556 1557 1558
		err = -EIO;
		goto out;
	}
	ret = add_extent_mapping(em_tree, em);
	if (ret == -EEXIST) {
		free_extent_map(em);
C
Chris Mason 已提交
1559
		em = NULL;
1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584
		failed_insert++;
		if (failed_insert > 5) {
			printk("failing to insert %Lu %Lu\n", start, end);
			err = -EIO;
			goto out;
		}
		goto again;
	}
	err = 0;
out:
	btrfs_free_path(path);
	if (trans) {
		ret = btrfs_end_transaction(trans, root);
		if (!err)
			err = ret;
	}
	mutex_unlock(&root->fs_info->fs_mutex);
	if (err) {
		free_extent_map(em);
		WARN_ON(1);
		return ERR_PTR(err);
	}
	return em;
}

1585
static sector_t btrfs_bmap(struct address_space *mapping, sector_t iblock)
C
Chris Mason 已提交
1586
{
1587
	return extent_bmap(mapping, iblock, btrfs_get_extent);
C
Chris Mason 已提交
1588 1589 1590 1591 1592
}

static int btrfs_prepare_write(struct file *file, struct page *page,
			       unsigned from, unsigned to)
{
1593 1594 1595
	return extent_prepare_write(&BTRFS_I(page->mapping->host)->extent_tree,
				    page->mapping->host, page, from, to,
				    btrfs_get_extent);
C
Chris Mason 已提交
1596 1597
}

1598
int btrfs_readpage(struct file *file, struct page *page)
C
Chris Mason 已提交
1599
{
1600 1601 1602
	struct extent_map_tree *tree;
	tree = &BTRFS_I(page->mapping->host)->extent_tree;
	return extent_read_full_page(tree, page, btrfs_get_extent);
C
Chris Mason 已提交
1603
}
1604
static int btrfs_writepage(struct page *page, struct writeback_control *wbc)
C
Chris Mason 已提交
1605
{
1606
	struct extent_map_tree *tree;
1607 1608 1609 1610 1611 1612 1613


	if (current->flags & PF_MEMALLOC) {
		redirty_page_for_writepage(wbc, page);
		unlock_page(page);
		return 0;
	}
1614 1615
	tree = &BTRFS_I(page->mapping->host)->extent_tree;
	return extent_write_full_page(tree, page, btrfs_get_extent, wbc);
C
Chris Mason 已提交
1616 1617
}

1618
static int btrfs_releasepage(struct page *page, gfp_t unused_gfp_flags)
C
Chris Mason 已提交
1619
{
1620 1621
	struct extent_map_tree *tree;
	int ret;
1622

1623 1624 1625
	if (page->private != 1) {
		WARN_ON(1);
		return try_to_free_buffers(page);
C
Chris Mason 已提交
1626
	}
1627 1628 1629 1630 1631 1632
	tree = &BTRFS_I(page->mapping->host)->extent_tree;
	ret = try_release_extent_mapping(tree, page);
	if (ret == 1) {
		ClearPagePrivate(page);
		set_page_private(page, 0);
		page_cache_release(page);
C
Chris Mason 已提交
1633
	}
1634
	return ret;
C
Chris Mason 已提交
1635 1636
}

1637
static void btrfs_invalidatepage(struct page *page, unsigned long offset)
C
Chris Mason 已提交
1638
{
1639
	struct extent_map_tree *tree;
C
Chris Mason 已提交
1640

1641 1642 1643
	tree = &BTRFS_I(page->mapping->host)->extent_tree;
	extent_invalidatepage(tree, page, offset);
	btrfs_releasepage(page, GFP_NOFS);
C
Chris Mason 已提交
1644 1645
}

C
Chris Mason 已提交
1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
/*
 * btrfs_page_mkwrite() is not allowed to change the file size as it gets
 * called from a page fault handler when a page is first dirtied. Hence we must
 * be careful to check for EOF conditions here. We set the page up correctly
 * for a written page which means we get ENOSPC checking when writing into
 * holes and correct delalloc and unwritten extent mapping on filesystems that
 * support these features.
 *
 * We are not allowed to take the i_mutex here so we have to play games to
 * protect against truncate races as the page could now be beyond EOF.  Because
 * vmtruncate() writes the inode size before removing pages, once we have the
 * page lock we can determine safely if the page is beyond EOF. If it is not
 * beyond EOF, then the page is guaranteed safe against truncation until we
 * unlock the page.
 */
int btrfs_page_mkwrite(struct vm_area_struct *vma, struct page *page)
{
	struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
	unsigned long end;
	loff_t size;
	int ret = -EINVAL;
1667
	u64 page_start;
C
Chris Mason 已提交
1668

1669
	down_read(&BTRFS_I(inode)->root->snap_sem);
C
Chris Mason 已提交
1670 1671 1672
	lock_page(page);
	wait_on_page_writeback(page);
	size = i_size_read(inode);
1673 1674
	page_start = page->index << PAGE_CACHE_SHIFT;

C
Chris Mason 已提交
1675
	if ((page->mapping != inode->i_mapping) ||
1676
	    (page_start > size)) {
C
Chris Mason 已提交
1677 1678 1679 1680 1681
		/* page got truncated out from underneath us */
		goto out_unlock;
	}

	/* page is wholly or partially inside EOF */
1682
	if (page_start + PAGE_CACHE_SIZE > size)
C
Chris Mason 已提交
1683 1684 1685 1686
		end = size & ~PAGE_CACHE_MASK;
	else
		end = PAGE_CACHE_SIZE;

1687
	ret = btrfs_cow_one_page(inode, page, end);
C
Chris Mason 已提交
1688 1689

out_unlock:
1690
	up_read(&BTRFS_I(inode)->root->snap_sem);
C
Chris Mason 已提交
1691 1692 1693 1694
	unlock_page(page);
	return ret;
}

C
Chris Mason 已提交
1695 1696 1697 1698 1699
static void btrfs_truncate(struct inode *inode)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	int ret;
	struct btrfs_trans_handle *trans;
1700
	unsigned long nr;
C
Chris Mason 已提交
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715

	if (!S_ISREG(inode->i_mode))
		return;
	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
		return;

	btrfs_truncate_page(inode->i_mapping, inode->i_size);

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, inode);

	/* FIXME, add redo link to tree so we don't leak on crash */
	ret = btrfs_truncate_in_trans(trans, root, inode);
	btrfs_update_inode(trans, root, inode);
1716
	nr = trans->blocks_used;
C
Chris Mason 已提交
1717 1718 1719
	ret = btrfs_end_transaction(trans, root);
	BUG_ON(ret);
	mutex_unlock(&root->fs_info->fs_mutex);
1720
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
1721 1722 1723 1724 1725
}

int btrfs_commit_write(struct file *file, struct page *page,
		       unsigned from, unsigned to)
{
1726 1727
	return extent_commit_write(&BTRFS_I(page->mapping->host)->extent_tree,
				   page->mapping->host, page, from, to);
C
Chris Mason 已提交
1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
}

static int create_subvol(struct btrfs_root *root, char *name, int namelen)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_key key;
	struct btrfs_root_item root_item;
	struct btrfs_inode_item *inode_item;
	struct buffer_head *subvol;
	struct btrfs_leaf *leaf;
	struct btrfs_root *new_root;
	struct inode *inode;
	struct inode *dir;
	int ret;
1742
	int err;
C
Chris Mason 已提交
1743 1744
	u64 objectid;
	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
1745
	unsigned long nr = 1;
C
Chris Mason 已提交
1746 1747 1748 1749 1750

	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	BUG_ON(!trans);

1751
	subvol = btrfs_alloc_free_block(trans, root, 0, 0);
1752 1753
	if (IS_ERR(subvol))
		return PTR_ERR(subvol);
C
Chris Mason 已提交
1754 1755 1756 1757 1758 1759 1760 1761
	leaf = btrfs_buffer_leaf(subvol);
	btrfs_set_header_nritems(&leaf->header, 0);
	btrfs_set_header_level(&leaf->header, 0);
	btrfs_set_header_blocknr(&leaf->header, bh_blocknr(subvol));
	btrfs_set_header_generation(&leaf->header, trans->transid);
	btrfs_set_header_owner(&leaf->header, root->root_key.objectid);
	memcpy(leaf->header.fsid, root->fs_info->disk_super->fsid,
	       sizeof(leaf->header.fsid));
C
Chris Mason 已提交
1762
	btrfs_mark_buffer_dirty(subvol);
C
Chris Mason 已提交
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773

	inode_item = &root_item.inode;
	memset(inode_item, 0, sizeof(*inode_item));
	btrfs_set_inode_generation(inode_item, 1);
	btrfs_set_inode_size(inode_item, 3);
	btrfs_set_inode_nlink(inode_item, 1);
	btrfs_set_inode_nblocks(inode_item, 1);
	btrfs_set_inode_mode(inode_item, S_IFDIR | 0755);

	btrfs_set_root_blocknr(&root_item, bh_blocknr(subvol));
	btrfs_set_root_refs(&root_item, 1);
1774
	btrfs_set_root_blocks_used(&root_item, 0);
1775 1776
	memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress));
	root_item.drop_level = 0;
C
Chris Mason 已提交
1777 1778 1779 1780 1781
	brelse(subvol);
	subvol = NULL;

	ret = btrfs_find_free_objectid(trans, root->fs_info->tree_root,
				       0, &objectid);
1782 1783
	if (ret)
		goto fail;
C
Chris Mason 已提交
1784 1785 1786 1787 1788 1789 1790 1791 1792

	btrfs_set_root_dirid(&root_item, new_dirid);

	key.objectid = objectid;
	key.offset = 1;
	key.flags = 0;
	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
				&root_item);
1793 1794
	if (ret)
		goto fail;
C
Chris Mason 已提交
1795 1796 1797 1798 1799 1800 1801 1802 1803

	/*
	 * insert the directory item
	 */
	key.offset = (u64)-1;
	dir = root->fs_info->sb->s_root->d_inode;
	ret = btrfs_insert_dir_item(trans, root->fs_info->tree_root,
				    name, namelen, dir->i_ino, &key,
				    BTRFS_FT_DIR);
1804 1805
	if (ret)
		goto fail;
C
Chris Mason 已提交
1806 1807

	ret = btrfs_commit_transaction(trans, root);
1808 1809
	if (ret)
		goto fail_commit;
C
Chris Mason 已提交
1810

1811
	new_root = btrfs_read_fs_root(root->fs_info, &key, name, namelen);
C
Chris Mason 已提交
1812 1813 1814 1815 1816 1817 1818
	BUG_ON(!new_root);

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

	inode = btrfs_new_inode(trans, new_root, new_dirid,
				BTRFS_I(dir)->block_group, S_IFDIR | 0700);
1819 1820
	if (IS_ERR(inode))
		goto fail;
C
Chris Mason 已提交
1821 1822
	inode->i_op = &btrfs_dir_inode_operations;
	inode->i_fop = &btrfs_dir_file_operations;
C
Chris Mason 已提交
1823
	new_root->inode = inode;
C
Chris Mason 已提交
1824 1825

	ret = btrfs_make_empty_dir(trans, new_root, new_dirid, new_dirid);
1826 1827
	if (ret)
		goto fail;
C
Chris Mason 已提交
1828 1829 1830 1831

	inode->i_nlink = 1;
	inode->i_size = 6;
	ret = btrfs_update_inode(trans, new_root, inode);
1832 1833 1834
	if (ret)
		goto fail;
fail:
1835
	nr = trans->blocks_used;
1836 1837 1838 1839
	err = btrfs_commit_transaction(trans, root);
	if (err && !ret)
		ret = err;
fail_commit:
C
Chris Mason 已提交
1840
	mutex_unlock(&root->fs_info->fs_mutex);
1841
	btrfs_btree_balance_dirty(root, nr);
1842
	return ret;
C
Chris Mason 已提交
1843 1844 1845 1846 1847 1848 1849
}

static int create_snapshot(struct btrfs_root *root, char *name, int namelen)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_key key;
	struct btrfs_root_item new_root_item;
1850
	struct buffer_head *tmp;
C
Chris Mason 已提交
1851
	int ret;
1852
	int err;
C
Chris Mason 已提交
1853
	u64 objectid;
1854
	unsigned long nr;
C
Chris Mason 已提交
1855 1856 1857 1858

	if (!root->ref_cows)
		return -EINVAL;

1859 1860 1861 1862
	down_write(&root->snap_sem);
	freeze_bdev(root->fs_info->sb->s_bdev);
	thaw_bdev(root->fs_info->sb->s_bdev, root->fs_info->sb);

C
Chris Mason 已提交
1863 1864 1865 1866 1867
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	BUG_ON(!trans);

	ret = btrfs_update_inode(trans, root, root->inode);
1868 1869
	if (ret)
		goto fail;
C
Chris Mason 已提交
1870 1871 1872

	ret = btrfs_find_free_objectid(trans, root->fs_info->tree_root,
				       0, &objectid);
1873 1874
	if (ret)
		goto fail;
C
Chris Mason 已提交
1875 1876 1877 1878 1879 1880 1881 1882

	memcpy(&new_root_item, &root->root_item,
	       sizeof(new_root_item));

	key.objectid = objectid;
	key.offset = 1;
	key.flags = 0;
	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
1883
	btrfs_cow_block(trans, root, root->node, NULL, 0, &tmp);
C
Chris Mason 已提交
1884 1885 1886 1887
	btrfs_set_root_blocknr(&new_root_item, bh_blocknr(root->node));

	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
				&new_root_item);
1888 1889
	if (ret)
		goto fail;
C
Chris Mason 已提交
1890 1891 1892 1893 1894 1895 1896 1897 1898 1899

	/*
	 * insert the directory item
	 */
	key.offset = (u64)-1;
	ret = btrfs_insert_dir_item(trans, root->fs_info->tree_root,
				    name, namelen,
				    root->fs_info->sb->s_root->d_inode->i_ino,
				    &key, BTRFS_FT_DIR);

1900 1901
	if (ret)
		goto fail;
C
Chris Mason 已提交
1902 1903

	ret = btrfs_inc_root_ref(trans, root);
1904 1905 1906
	if (ret)
		goto fail;
fail:
1907
	nr = trans->blocks_used;
1908 1909 1910
	err = btrfs_commit_transaction(trans, root);
	if (err && !ret)
		ret = err;
C
Chris Mason 已提交
1911
	mutex_unlock(&root->fs_info->fs_mutex);
1912
	up_write(&root->snap_sem);
1913
	btrfs_btree_balance_dirty(root, nr);
1914
	return ret;
C
Chris Mason 已提交
1915 1916
}

1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
static unsigned long force_ra(struct address_space *mapping,
			      struct file_ra_state *ra, struct file *file,
			      pgoff_t offset, pgoff_t last_index)
{
	pgoff_t req_size;

#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23)
	req_size = last_index - offset + 1;
	offset = page_cache_readahead(mapping, ra, file, offset, req_size);
	return offset;
#else
	req_size = min(last_index - offset + 1, (pgoff_t)128);
	page_cache_sync_readahead(mapping, ra, file, offset, req_size);
	return offset + req_size;
#endif
}

int btrfs_defrag_file(struct file *file) {
	struct inode *inode = file->f_path.dentry->d_inode;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	struct page *page;
	unsigned long last_index;
	unsigned long ra_index = 0;
	u64 page_start;
	u64 page_end;
	unsigned long i;

	mutex_lock(&inode->i_mutex);
	last_index = inode->i_size >> PAGE_CACHE_SHIFT;
	for (i = 0; i <= last_index; i++) {
		if (i == ra_index) {
			ra_index = force_ra(inode->i_mapping, &file->f_ra,
					    file, ra_index, last_index);
		}
		page = grab_cache_page(inode->i_mapping, i);
		if (!page)
			goto out_unlock;
		if (!PageUptodate(page)) {
			btrfs_readpage(NULL, page);
			lock_page(page);
			if (!PageUptodate(page)) {
				unlock_page(page);
				page_cache_release(page);
				goto out_unlock;
			}
		}
		page_start = page->index << PAGE_CACHE_SHIFT;
		page_end = page_start + PAGE_CACHE_SIZE - 1;

		lock_extent(em_tree, page_start, page_end, GFP_NOFS);
		set_extent_delalloc(em_tree, page_start,
				    page_end, GFP_NOFS);
		unlock_extent(em_tree, page_start, page_end, GFP_NOFS);
		set_page_dirty(page);
		unlock_page(page);
		page_cache_release(page);
		balance_dirty_pages_ratelimited_nr(inode->i_mapping, 1);
	}

out_unlock:
	mutex_unlock(&inode->i_mutex);
	return 0;
}

C
Christoph Hellwig 已提交
1981
static int btrfs_ioctl_snap_create(struct btrfs_root *root, void __user *arg)
C
Chris Mason 已提交
1982 1983 1984 1985
{
	struct btrfs_ioctl_vol_args vol_args;
	struct btrfs_dir_item *di;
	struct btrfs_path *path;
C
Christoph Hellwig 已提交
1986
	int namelen;
C
Chris Mason 已提交
1987 1988
	u64 root_dirid;

C
Christoph Hellwig 已提交
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
	if (copy_from_user(&vol_args, arg, sizeof(vol_args)))
		return -EFAULT;
	
	namelen = strlen(vol_args.name);
	if (namelen > BTRFS_VOL_NAME_MAX)
		return -EINVAL;
	if (strchr(vol_args.name, '/'))
		return -EINVAL;

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

	root_dirid = root->fs_info->sb->s_root->d_inode->i_ino,
	mutex_lock(&root->fs_info->fs_mutex);
	di = btrfs_lookup_dir_item(NULL, root->fs_info->tree_root,
			    path, root_dirid,
			    vol_args.name, namelen, 0);
	mutex_unlock(&root->fs_info->fs_mutex);
	btrfs_free_path(path);
	if (di && !IS_ERR(di))
		return -EEXIST;
	if (IS_ERR(di))
		return PTR_ERR(di);

	if (root == root->fs_info->tree_root)
		return create_subvol(root, vol_args.name, namelen);
	return create_snapshot(root, vol_args.name, namelen);
}

static int btrfs_ioctl_defrag(struct file *file)
{
	struct inode *inode = file->f_path.dentry->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;

	switch (inode->i_mode & S_IFMT) {
	case S_IFDIR:
C
Chris Mason 已提交
2026
		mutex_lock(&root->fs_info->fs_mutex);
C
Christoph Hellwig 已提交
2027 2028
		btrfs_defrag_root(root, 0);
		btrfs_defrag_root(root->fs_info->extent_root, 0);
C
Chris Mason 已提交
2029 2030
		mutex_unlock(&root->fs_info->fs_mutex);
		break;
C
Christoph Hellwig 已提交
2031 2032 2033 2034 2035 2036 2037
	case S_IFREG:
		btrfs_defrag_file(file);
		break;
	}

	return 0;
}
2038

C
Christoph Hellwig 已提交
2039 2040 2041 2042 2043 2044 2045 2046
long btrfs_ioctl(struct file *file, unsigned int
		cmd, unsigned long arg)
{
	struct btrfs_root *root = BTRFS_I(file->f_path.dentry->d_inode)->root;

	switch (cmd) {
	case BTRFS_IOC_SNAP_CREATE:
		return btrfs_ioctl_snap_create(root, (void __user *)arg);
2047
	case BTRFS_IOC_DEFRAG:
C
Christoph Hellwig 已提交
2048
		return btrfs_ioctl_defrag(file);
C
Chris Mason 已提交
2049
	}
C
Christoph Hellwig 已提交
2050 2051

	return -ENOTTY;
C
Chris Mason 已提交
2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063
}

/*
 * Called inside transaction, so use GFP_NOFS
 */
struct inode *btrfs_alloc_inode(struct super_block *sb)
{
	struct btrfs_inode *ei;

	ei = kmem_cache_alloc(btrfs_inode_cachep, GFP_NOFS);
	if (!ei)
		return NULL;
2064
	ei->last_trans = 0;
C
Chris Mason 已提交
2065 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
	return &ei->vfs_inode;
}

void btrfs_destroy_inode(struct inode *inode)
{
	WARN_ON(!list_empty(&inode->i_dentry));
	WARN_ON(inode->i_data.nrpages);

	kmem_cache_free(btrfs_inode_cachep, BTRFS_I(inode));
}

static void init_once(void * foo, struct kmem_cache * cachep,
		      unsigned long flags)
{
	struct btrfs_inode *ei = (struct btrfs_inode *) foo;

	inode_init_once(&ei->vfs_inode);
}

void btrfs_destroy_cachep(void)
{
	if (btrfs_inode_cachep)
		kmem_cache_destroy(btrfs_inode_cachep);
	if (btrfs_trans_handle_cachep)
		kmem_cache_destroy(btrfs_trans_handle_cachep);
	if (btrfs_transaction_cachep)
		kmem_cache_destroy(btrfs_transaction_cachep);
	if (btrfs_bit_radix_cachep)
		kmem_cache_destroy(btrfs_bit_radix_cachep);
	if (btrfs_path_cachep)
		kmem_cache_destroy(btrfs_path_cachep);
}

2098
struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110
				       unsigned long extra_flags,
				       void (*ctor)(void *, struct kmem_cache *,
						    unsigned long))
{
	return kmem_cache_create(name, size, 0, (SLAB_RECLAIM_ACCOUNT |
				 SLAB_MEM_SPREAD | extra_flags), ctor
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23)
				 ,NULL
#endif
				);
}

C
Chris Mason 已提交
2111 2112
int btrfs_init_cachep(void)
{
2113
	btrfs_inode_cachep = btrfs_cache_create("btrfs_inode_cache",
2114 2115
					  sizeof(struct btrfs_inode),
					  0, init_once);
C
Chris Mason 已提交
2116 2117
	if (!btrfs_inode_cachep)
		goto fail;
2118 2119 2120 2121
	btrfs_trans_handle_cachep =
			btrfs_cache_create("btrfs_trans_handle_cache",
					   sizeof(struct btrfs_trans_handle),
					   0, NULL);
C
Chris Mason 已提交
2122 2123
	if (!btrfs_trans_handle_cachep)
		goto fail;
2124
	btrfs_transaction_cachep = btrfs_cache_create("btrfs_transaction_cache",
C
Chris Mason 已提交
2125
					     sizeof(struct btrfs_transaction),
2126
					     0, NULL);
C
Chris Mason 已提交
2127 2128
	if (!btrfs_transaction_cachep)
		goto fail;
2129
	btrfs_path_cachep = btrfs_cache_create("btrfs_path_cache",
2130 2131
					 sizeof(struct btrfs_transaction),
					 0, NULL);
C
Chris Mason 已提交
2132 2133
	if (!btrfs_path_cachep)
		goto fail;
2134
	btrfs_bit_radix_cachep = btrfs_cache_create("btrfs_radix", 256,
2135
					      SLAB_DESTROY_BY_RCU, NULL);
C
Chris Mason 已提交
2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263
	if (!btrfs_bit_radix_cachep)
		goto fail;
	return 0;
fail:
	btrfs_destroy_cachep();
	return -ENOMEM;
}

static int btrfs_getattr(struct vfsmount *mnt,
			 struct dentry *dentry, struct kstat *stat)
{
	struct inode *inode = dentry->d_inode;
	generic_fillattr(inode, stat);
	stat->blksize = 256 * 1024;
	return 0;
}

static int btrfs_rename(struct inode * old_dir, struct dentry *old_dentry,
			   struct inode * new_dir,struct dentry *new_dentry)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(old_dir)->root;
	struct inode *new_inode = new_dentry->d_inode;
	struct inode *old_inode = old_dentry->d_inode;
	struct timespec ctime = CURRENT_TIME;
	struct btrfs_path *path;
	struct btrfs_dir_item *di;
	int ret;

	if (S_ISDIR(old_inode->i_mode) && new_inode &&
	    new_inode->i_size > BTRFS_EMPTY_DIR_SIZE) {
		return -ENOTEMPTY;
	}
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, new_dir);
	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out_fail;
	}

	old_dentry->d_inode->i_nlink++;
	old_dir->i_ctime = old_dir->i_mtime = ctime;
	new_dir->i_ctime = new_dir->i_mtime = ctime;
	old_inode->i_ctime = ctime;
	if (S_ISDIR(old_inode->i_mode) && old_dir != new_dir) {
		struct btrfs_key *location = &BTRFS_I(new_dir)->location;
		u64 old_parent_oid;
		di = btrfs_lookup_dir_item(trans, root, path, old_inode->i_ino,
					   "..", 2, -1);
		if (IS_ERR(di)) {
			ret = PTR_ERR(di);
			goto out_fail;
		}
		if (!di) {
			ret = -ENOENT;
			goto out_fail;
		}
		old_parent_oid = btrfs_disk_key_objectid(&di->location);
		ret = btrfs_del_item(trans, root, path);
		if (ret) {
			goto out_fail;
		}
		btrfs_release_path(root, path);

		di = btrfs_lookup_dir_index_item(trans, root, path,
						 old_inode->i_ino,
						 old_parent_oid,
						 "..", 2, -1);
		if (IS_ERR(di)) {
			ret = PTR_ERR(di);
			goto out_fail;
		}
		if (!di) {
			ret = -ENOENT;
			goto out_fail;
		}
		ret = btrfs_del_item(trans, root, path);
		if (ret) {
			goto out_fail;
		}
		btrfs_release_path(root, path);

		ret = btrfs_insert_dir_item(trans, root, "..", 2,
					    old_inode->i_ino, location,
					    BTRFS_FT_DIR);
		if (ret)
			goto out_fail;
	}


	ret = btrfs_unlink_trans(trans, root, old_dir, old_dentry);
	if (ret)
		goto out_fail;

	if (new_inode) {
		new_inode->i_ctime = CURRENT_TIME;
		ret = btrfs_unlink_trans(trans, root, new_dir, new_dentry);
		if (ret)
			goto out_fail;
	}
	ret = btrfs_add_link(trans, new_dentry, old_inode);
	if (ret)
		goto out_fail;

out_fail:
	btrfs_free_path(path);
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
	return ret;
}

static int btrfs_symlink(struct inode *dir, struct dentry *dentry,
			 const char *symname)
{
	struct btrfs_trans_handle *trans;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct btrfs_path *path;
	struct btrfs_key key;
	struct inode *inode;
	int err;
	int drop_inode = 0;
	u64 objectid;
	int name_len;
	int datasize;
	char *ptr;
	struct btrfs_file_extent_item *ei;
2264
	unsigned long nr;
C
Chris Mason 已提交
2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292

	name_len = strlen(symname) + 1;
	if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root))
		return -ENAMETOOLONG;
	mutex_lock(&root->fs_info->fs_mutex);
	trans = btrfs_start_transaction(root, 1);
	btrfs_set_trans_block_group(trans, dir);

	err = btrfs_find_free_objectid(trans, root, dir->i_ino, &objectid);
	if (err) {
		err = -ENOSPC;
		goto out_unlock;
	}

	inode = btrfs_new_inode(trans, root, objectid,
				BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO);
	err = PTR_ERR(inode);
	if (IS_ERR(inode))
		goto out_unlock;

	btrfs_set_trans_block_group(trans, inode);
	err = btrfs_add_nondir(trans, dentry, inode);
	if (err)
		drop_inode = 1;
	else {
		inode->i_mapping->a_ops = &btrfs_aops;
		inode->i_fop = &btrfs_file_operations;
		inode->i_op = &btrfs_file_inode_operations;
2293 2294
		extent_map_tree_init(&BTRFS_I(inode)->extent_tree,
				     inode->i_mapping, GFP_NOFS);
2295
		BTRFS_I(inode)->extent_tree.ops = &btrfs_extent_map_ops;
C
Chris Mason 已提交
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311
	}
	dir->i_sb->s_dirt = 1;
	btrfs_update_inode_block_group(trans, inode);
	btrfs_update_inode_block_group(trans, dir);
	if (drop_inode)
		goto out_unlock;

	path = btrfs_alloc_path();
	BUG_ON(!path);
	key.objectid = inode->i_ino;
	key.offset = 0;
	key.flags = 0;
	btrfs_set_key_type(&key, BTRFS_EXTENT_DATA_KEY);
	datasize = btrfs_file_extent_calc_inline_size(name_len);
	err = btrfs_insert_empty_item(trans, root, path, &key,
				      datasize);
2312 2313 2314 2315
	if (err) {
		drop_inode = 1;
		goto out_unlock;
	}
C
Chris Mason 已提交
2316 2317 2318 2319 2320 2321 2322 2323
	ei = btrfs_item_ptr(btrfs_buffer_leaf(path->nodes[0]),
	       path->slots[0], struct btrfs_file_extent_item);
	btrfs_set_file_extent_generation(ei, trans->transid);
	btrfs_set_file_extent_type(ei,
				   BTRFS_FILE_EXTENT_INLINE);
	ptr = btrfs_file_extent_inline_start(ei);
	btrfs_memcpy(root, path->nodes[0]->b_data,
		     ptr, symname, name_len);
C
Chris Mason 已提交
2324
	btrfs_mark_buffer_dirty(path->nodes[0]);
C
Chris Mason 已提交
2325 2326 2327 2328
	btrfs_free_path(path);
	inode->i_op = &btrfs_symlink_inode_operations;
	inode->i_mapping->a_ops = &btrfs_symlink_aops;
	inode->i_size = name_len - 1;
2329 2330 2331
	err = btrfs_update_inode(trans, root, inode);
	if (err)
		drop_inode = 1;
C
Chris Mason 已提交
2332 2333

out_unlock:
2334
	nr = trans->blocks_used;
C
Chris Mason 已提交
2335 2336 2337 2338 2339 2340
	btrfs_end_transaction(trans, root);
	mutex_unlock(&root->fs_info->fs_mutex);
	if (drop_inode) {
		inode_dec_link_count(inode);
		iput(inode);
	}
2341
	btrfs_btree_balance_dirty(root, nr);
C
Chris Mason 已提交
2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354
	return err;
}

static struct inode_operations btrfs_dir_inode_operations = {
	.lookup		= btrfs_lookup,
	.create		= btrfs_create,
	.unlink		= btrfs_unlink,
	.link		= btrfs_link,
	.mkdir		= btrfs_mkdir,
	.rmdir		= btrfs_rmdir,
	.rename		= btrfs_rename,
	.symlink	= btrfs_symlink,
	.setattr	= btrfs_setattr,
J
Josef Bacik 已提交
2355
	.mknod		= btrfs_mknod,
C
Chris Mason 已提交
2356 2357 2358 2359 2360 2361 2362 2363 2364 2365
};

static struct inode_operations btrfs_dir_ro_inode_operations = {
	.lookup		= btrfs_lookup,
};

static struct file_operations btrfs_dir_file_operations = {
	.llseek		= generic_file_llseek,
	.read		= generic_read_dir,
	.readdir	= btrfs_readdir,
C
Christoph Hellwig 已提交
2366
	.unlocked_ioctl	= btrfs_ioctl,
C
Chris Mason 已提交
2367
#ifdef CONFIG_COMPAT
C
Christoph Hellwig 已提交
2368
	.compat_ioctl	= btrfs_ioctl,
C
Chris Mason 已提交
2369 2370 2371
#endif
};

2372 2373 2374 2375 2376 2377 2378
static struct extent_map_ops btrfs_extent_map_ops = {
	.fill_delalloc = run_delalloc_range,
	.writepage_io_hook = btrfs_writepage_io_hook,
	.readpage_io_hook = btrfs_readpage_io_hook,
	.readpage_end_io_hook = btrfs_readpage_end_io_hook,
};

C
Chris Mason 已提交
2379 2380 2381 2382 2383 2384 2385
static struct address_space_operations btrfs_aops = {
	.readpage	= btrfs_readpage,
	.writepage	= btrfs_writepage,
	.sync_page	= block_sync_page,
	.prepare_write	= btrfs_prepare_write,
	.commit_write	= btrfs_commit_write,
	.bmap		= btrfs_bmap,
2386 2387 2388
	.invalidatepage = btrfs_invalidatepage,
	.releasepage	= btrfs_releasepage,
	.set_page_dirty	= __set_page_dirty_nobuffers,
C
Chris Mason 已提交
2389 2390 2391 2392 2393
};

static struct address_space_operations btrfs_symlink_aops = {
	.readpage	= btrfs_readpage,
	.writepage	= btrfs_writepage,
C
Chris Mason 已提交
2394 2395
	.invalidatepage = btrfs_invalidatepage,
	.releasepage	= btrfs_releasepage,
C
Chris Mason 已提交
2396 2397 2398 2399 2400 2401 2402 2403
};

static struct inode_operations btrfs_file_inode_operations = {
	.truncate	= btrfs_truncate,
	.getattr	= btrfs_getattr,
	.setattr	= btrfs_setattr,
};

J
Josef Bacik 已提交
2404 2405 2406 2407 2408
static struct inode_operations btrfs_special_inode_operations = {
	.getattr	= btrfs_getattr,
	.setattr	= btrfs_setattr,
};

C
Chris Mason 已提交
2409 2410 2411 2412 2413
static struct inode_operations btrfs_symlink_inode_operations = {
	.readlink	= generic_readlink,
	.follow_link	= page_follow_link_light,
	.put_link	= page_put_link,
};