ioctl.c 31.7 KB
Newer Older
C
Christoph Hellwig 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
/*
 * 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/kernel.h>
#include <linux/bio.h>
#include <linux/buffer_head.h>
#include <linux/file.h>
#include <linux/fs.h>
24
#include <linux/fsnotify.h>
C
Christoph Hellwig 已提交
25 26 27 28 29 30
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/time.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/backing-dev.h>
31
#include <linux/mount.h>
C
Christoph Hellwig 已提交
32
#include <linux/mpage.h>
33
#include <linux/namei.h>
C
Christoph Hellwig 已提交
34 35 36 37 38
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/statfs.h>
#include <linux/compat.h>
#include <linux/bit_spinlock.h>
39
#include <linux/security.h>
C
Christoph Hellwig 已提交
40
#include <linux/xattr.h>
41
#include <linux/vmalloc.h>
C
Chris Mason 已提交
42
#include "compat.h"
C
Christoph Hellwig 已提交
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"
#include "volumes.h"
50
#include "locking.h"
C
Christoph Hellwig 已提交
51

52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113
/* Mask out flags that are inappropriate for the given type of inode. */
static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
{
	if (S_ISDIR(mode))
		return flags;
	else if (S_ISREG(mode))
		return flags & ~FS_DIRSYNC_FL;
	else
		return flags & (FS_NODUMP_FL | FS_NOATIME_FL);
}

/*
 * Export inode flags to the format expected by the FS_IOC_GETFLAGS ioctl.
 */
static unsigned int btrfs_flags_to_ioctl(unsigned int flags)
{
	unsigned int iflags = 0;

	if (flags & BTRFS_INODE_SYNC)
		iflags |= FS_SYNC_FL;
	if (flags & BTRFS_INODE_IMMUTABLE)
		iflags |= FS_IMMUTABLE_FL;
	if (flags & BTRFS_INODE_APPEND)
		iflags |= FS_APPEND_FL;
	if (flags & BTRFS_INODE_NODUMP)
		iflags |= FS_NODUMP_FL;
	if (flags & BTRFS_INODE_NOATIME)
		iflags |= FS_NOATIME_FL;
	if (flags & BTRFS_INODE_DIRSYNC)
		iflags |= FS_DIRSYNC_FL;

	return iflags;
}

/*
 * Update inode->i_flags based on the btrfs internal flags.
 */
void btrfs_update_iflags(struct inode *inode)
{
	struct btrfs_inode *ip = BTRFS_I(inode);

	inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);

	if (ip->flags & BTRFS_INODE_SYNC)
		inode->i_flags |= S_SYNC;
	if (ip->flags & BTRFS_INODE_IMMUTABLE)
		inode->i_flags |= S_IMMUTABLE;
	if (ip->flags & BTRFS_INODE_APPEND)
		inode->i_flags |= S_APPEND;
	if (ip->flags & BTRFS_INODE_NOATIME)
		inode->i_flags |= S_NOATIME;
	if (ip->flags & BTRFS_INODE_DIRSYNC)
		inode->i_flags |= S_DIRSYNC;
}

/*
 * Inherit flags from the parent inode.
 *
 * Unlike extN we don't have any flags we don't want to inherit currently.
 */
void btrfs_inherit_iflags(struct inode *inode, struct inode *dir)
{
114 115 116 117 118 119
	unsigned int flags;

	if (!dir)
		return;

	flags = BTRFS_I(dir)->flags;
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

	if (S_ISREG(inode->i_mode))
		flags &= ~BTRFS_INODE_DIRSYNC;
	else if (!S_ISDIR(inode->i_mode))
		flags &= (BTRFS_INODE_NODUMP | BTRFS_INODE_NOATIME);

	BTRFS_I(inode)->flags = flags;
	btrfs_update_iflags(inode);
}

static int btrfs_ioctl_getflags(struct file *file, void __user *arg)
{
	struct btrfs_inode *ip = BTRFS_I(file->f_path.dentry->d_inode);
	unsigned int flags = btrfs_flags_to_ioctl(ip->flags);

	if (copy_to_user(arg, &flags, sizeof(flags)))
		return -EFAULT;
	return 0;
}

static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
{
	struct inode *inode = file->f_path.dentry->d_inode;
	struct btrfs_inode *ip = BTRFS_I(inode);
	struct btrfs_root *root = ip->root;
	struct btrfs_trans_handle *trans;
	unsigned int flags, oldflags;
	int ret;

	if (copy_from_user(&flags, arg, sizeof(flags)))
		return -EFAULT;

	if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
		      FS_NOATIME_FL | FS_NODUMP_FL | \
		      FS_SYNC_FL | FS_DIRSYNC_FL))
		return -EOPNOTSUPP;
C
Christoph Hellwig 已提交
156

157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222
	if (!is_owner_or_cap(inode))
		return -EACCES;

	mutex_lock(&inode->i_mutex);

	flags = btrfs_mask_flags(inode->i_mode, flags);
	oldflags = btrfs_flags_to_ioctl(ip->flags);
	if ((flags ^ oldflags) & (FS_APPEND_FL | FS_IMMUTABLE_FL)) {
		if (!capable(CAP_LINUX_IMMUTABLE)) {
			ret = -EPERM;
			goto out_unlock;
		}
	}

	ret = mnt_want_write(file->f_path.mnt);
	if (ret)
		goto out_unlock;

	if (flags & FS_SYNC_FL)
		ip->flags |= BTRFS_INODE_SYNC;
	else
		ip->flags &= ~BTRFS_INODE_SYNC;
	if (flags & FS_IMMUTABLE_FL)
		ip->flags |= BTRFS_INODE_IMMUTABLE;
	else
		ip->flags &= ~BTRFS_INODE_IMMUTABLE;
	if (flags & FS_APPEND_FL)
		ip->flags |= BTRFS_INODE_APPEND;
	else
		ip->flags &= ~BTRFS_INODE_APPEND;
	if (flags & FS_NODUMP_FL)
		ip->flags |= BTRFS_INODE_NODUMP;
	else
		ip->flags &= ~BTRFS_INODE_NODUMP;
	if (flags & FS_NOATIME_FL)
		ip->flags |= BTRFS_INODE_NOATIME;
	else
		ip->flags &= ~BTRFS_INODE_NOATIME;
	if (flags & FS_DIRSYNC_FL)
		ip->flags |= BTRFS_INODE_DIRSYNC;
	else
		ip->flags &= ~BTRFS_INODE_DIRSYNC;


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

	ret = btrfs_update_inode(trans, root, inode);
	BUG_ON(ret);

	btrfs_update_iflags(inode);
	inode->i_ctime = CURRENT_TIME;
	btrfs_end_transaction(trans, root);

	mnt_drop_write(file->f_path.mnt);
 out_unlock:
	mutex_unlock(&inode->i_mutex);
	return 0;
}

static int btrfs_ioctl_getversion(struct file *file, int __user *arg)
{
	struct inode *inode = file->f_path.dentry->d_inode;

	return put_user(inode->i_generation, arg);
}
C
Christoph Hellwig 已提交
223

224 225 226
static noinline int create_subvol(struct btrfs_root *root,
				  struct dentry *dentry,
				  char *name, int namelen)
C
Christoph Hellwig 已提交
227 228 229 230 231 232 233 234 235 236 237 238
{
	struct btrfs_trans_handle *trans;
	struct btrfs_key key;
	struct btrfs_root_item root_item;
	struct btrfs_inode_item *inode_item;
	struct extent_buffer *leaf;
	struct btrfs_root *new_root = root;
	struct inode *dir;
	int ret;
	int err;
	u64 objectid;
	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
239
	u64 index = 0;
C
Christoph Hellwig 已提交
240 241
	unsigned long nr = 1;

J
Josef Bacik 已提交
242
	ret = btrfs_check_metadata_free_space(root);
C
Christoph Hellwig 已提交
243 244 245 246 247 248 249 250 251 252 253
	if (ret)
		goto fail_commit;

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

	ret = btrfs_find_free_objectid(trans, root->fs_info->tree_root,
				       0, &objectid);
	if (ret)
		goto fail;

254 255
	leaf = btrfs_alloc_free_block(trans, root, root->leafsize,
				      0, objectid, NULL, 0, 0, 0);
256 257 258 259
	if (IS_ERR(leaf)) {
		ret = PTR_ERR(leaf);
		goto fail;
	}
C
Christoph Hellwig 已提交
260

261
	memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header));
C
Christoph Hellwig 已提交
262 263
	btrfs_set_header_bytenr(leaf, leaf->start);
	btrfs_set_header_generation(leaf, trans->transid);
264
	btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
C
Christoph Hellwig 已提交
265 266 267 268 269
	btrfs_set_header_owner(leaf, objectid);

	write_extent_buffer(leaf, root->fs_info->fsid,
			    (unsigned long)btrfs_header_fsid(leaf),
			    BTRFS_FSID_SIZE);
270 271 272
	write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
			    (unsigned long)btrfs_header_chunk_tree_uuid(leaf),
			    BTRFS_UUID_SIZE);
C
Christoph Hellwig 已提交
273 274 275 276 277 278 279
	btrfs_mark_buffer_dirty(leaf);

	inode_item = &root_item.inode;
	memset(inode_item, 0, sizeof(*inode_item));
	inode_item->generation = cpu_to_le64(1);
	inode_item->size = cpu_to_le64(3);
	inode_item->nlink = cpu_to_le32(1);
280
	inode_item->nbytes = cpu_to_le64(root->leafsize);
C
Christoph Hellwig 已提交
281 282 283
	inode_item->mode = cpu_to_le32(S_IFDIR | 0755);

	btrfs_set_root_bytenr(&root_item, leaf->start);
284
	btrfs_set_root_generation(&root_item, trans->transid);
C
Christoph Hellwig 已提交
285 286 287
	btrfs_set_root_level(&root_item, 0);
	btrfs_set_root_refs(&root_item, 1);
	btrfs_set_root_used(&root_item, 0);
Y
Yan Zheng 已提交
288
	btrfs_set_root_last_snapshot(&root_item, 0);
C
Christoph Hellwig 已提交
289 290 291 292

	memset(&root_item.drop_progress, 0, sizeof(root_item.drop_progress));
	root_item.drop_level = 0;

293
	btrfs_tree_unlock(leaf);
C
Christoph Hellwig 已提交
294 295 296 297 298 299
	free_extent_buffer(leaf);
	leaf = NULL;

	btrfs_set_root_dirid(&root_item, new_dirid);

	key.objectid = objectid;
300
	key.offset = 0;
C
Christoph Hellwig 已提交
301 302 303 304 305 306 307 308 309 310
	btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
				&root_item);
	if (ret)
		goto fail;

	/*
	 * insert the directory item
	 */
	key.offset = (u64)-1;
311 312 313 314 315
	dir = dentry->d_parent->d_inode;
	ret = btrfs_set_inode_index(dir, &index);
	BUG_ON(ret);

	ret = btrfs_insert_dir_item(trans, root,
C
Christoph Hellwig 已提交
316
				    name, namelen, dir->i_ino, &key,
317
				    BTRFS_FT_DIR, index);
C
Christoph Hellwig 已提交
318 319
	if (ret)
		goto fail;
320

321 322 323 324
	btrfs_i_size_write(dir, dir->i_size + namelen * 2);
	ret = btrfs_update_inode(trans, root, dir);
	BUG_ON(ret);

325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
	/* add the backref first */
	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
				 objectid, BTRFS_ROOT_BACKREF_KEY,
				 root->root_key.objectid,
				 dir->i_ino, index, name, namelen);

	BUG_ON(ret);

	/* now add the forward ref */
	ret = btrfs_add_root_ref(trans, root->fs_info->tree_root,
				 root->root_key.objectid, BTRFS_ROOT_REF_KEY,
				 objectid,
				 dir->i_ino, index, name, namelen);

	BUG_ON(ret);

C
Christoph Hellwig 已提交
341 342 343 344
	ret = btrfs_commit_transaction(trans, root);
	if (ret)
		goto fail_commit;

345
	new_root = btrfs_read_fs_root_no_name(root->fs_info, &key);
C
Christoph Hellwig 已提交
346 347 348 349 350
	BUG_ON(!new_root);

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

351
	ret = btrfs_create_subvol_root(trans, new_root, dentry, new_dirid,
C
Christoph Hellwig 已提交
352 353 354 355 356 357 358 359 360 361 362 363 364 365
				       BTRFS_I(dir)->block_group);
	if (ret)
		goto fail;

fail:
	nr = trans->blocks_used;
	err = btrfs_commit_transaction(trans, new_root);
	if (err && !ret)
		ret = err;
fail_commit:
	btrfs_btree_balance_dirty(root, nr);
	return ret;
}

366 367
static int create_snapshot(struct btrfs_root *root, struct dentry *dentry,
			   char *name, int namelen)
C
Christoph Hellwig 已提交
368 369 370
{
	struct btrfs_pending_snapshot *pending_snapshot;
	struct btrfs_trans_handle *trans;
371
	int ret = 0;
C
Christoph Hellwig 已提交
372 373 374 375 376 377
	int err;
	unsigned long nr = 0;

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

J
Josef Bacik 已提交
378
	ret = btrfs_check_metadata_free_space(root);
C
Christoph Hellwig 已提交
379 380 381
	if (ret)
		goto fail_unlock;

382
	pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS);
C
Christoph Hellwig 已提交
383 384 385 386 387 388 389 390 391 392 393 394
	if (!pending_snapshot) {
		ret = -ENOMEM;
		goto fail_unlock;
	}
	pending_snapshot->name = kmalloc(namelen + 1, GFP_NOFS);
	if (!pending_snapshot->name) {
		ret = -ENOMEM;
		kfree(pending_snapshot);
		goto fail_unlock;
	}
	memcpy(pending_snapshot->name, name, namelen);
	pending_snapshot->name[namelen] = '\0';
395
	pending_snapshot->dentry = dentry;
C
Christoph Hellwig 已提交
396 397 398 399 400 401 402 403 404 405 406 407
	trans = btrfs_start_transaction(root, 1);
	BUG_ON(!trans);
	pending_snapshot->root = root;
	list_add(&pending_snapshot->list,
		 &trans->transaction->pending_snapshots);
	err = btrfs_commit_transaction(trans, root);

fail_unlock:
	btrfs_btree_balance_dirty(root, nr);
	return ret;
}

408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423
/* copy of may_create in fs/namei.c() */
static inline int btrfs_may_create(struct inode *dir, struct dentry *child)
{
	if (child->d_inode)
		return -EEXIST;
	if (IS_DEADDIR(dir))
		return -ENOENT;
	return inode_permission(dir, MAY_WRITE | MAY_EXEC);
}

/*
 * Create a new subvolume below @parent.  This is largely modeled after
 * sys_mkdirat and vfs_mkdir, but we only do a single component lookup
 * inside this filesystem so it's quite a bit simpler.
 */
static noinline int btrfs_mksubvol(struct path *parent, char *name,
424 425
				   int mode, int namelen,
				   struct btrfs_root *snap_src)
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441
{
	struct dentry *dentry;
	int error;

	mutex_lock_nested(&parent->dentry->d_inode->i_mutex, I_MUTEX_PARENT);

	dentry = lookup_one_len(name, parent->dentry, namelen);
	error = PTR_ERR(dentry);
	if (IS_ERR(dentry))
		goto out_unlock;

	error = -EEXIST;
	if (dentry->d_inode)
		goto out_dput;

	if (!IS_POSIXACL(parent->dentry->d_inode))
A
Al Viro 已提交
442
		mode &= ~current_umask();
443

444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461
	error = mnt_want_write(parent->mnt);
	if (error)
		goto out_dput;

	error = btrfs_may_create(parent->dentry->d_inode, dentry);
	if (error)
		goto out_drop_write;

	/*
	 * Actually perform the low-level subvolume creation after all
	 * this VFS fuzz.
	 *
	 * Eventually we want to pass in an inode under which we create this
	 * subvolume, but for now all are under the filesystem root.
	 *
	 * Also we should pass on the mode eventually to allow creating new
	 * subvolume with specific mode bits.
	 */
462
	if (snap_src) {
463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484
		struct dentry *dir = dentry->d_parent;
		struct dentry *test = dir->d_parent;
		struct btrfs_path *path = btrfs_alloc_path();
		int ret;
		u64 test_oid;
		u64 parent_oid = BTRFS_I(dir->d_inode)->root->root_key.objectid;

		test_oid = snap_src->root_key.objectid;

		ret = btrfs_find_root_ref(snap_src->fs_info->tree_root,
					  path, parent_oid, test_oid);
		if (ret == 0)
			goto create;
		btrfs_release_path(snap_src->fs_info->tree_root, path);

		/* we need to make sure we aren't creating a directory loop
		 * by taking a snapshot of something that has our current
		 * subvol in its directory tree.  So, this loops through
		 * the dentries and checks the forward refs for each subvolume
		 * to see if is references the subvolume where we are
		 * placing this new snapshot.
		 */
C
Chris Mason 已提交
485
		while (1) {
486 487 488 489 490 491 492
			if (!test ||
			    dir == snap_src->fs_info->sb->s_root ||
			    test == snap_src->fs_info->sb->s_root ||
			    test->d_inode->i_sb != snap_src->fs_info->sb) {
				break;
			}
			if (S_ISLNK(test->d_inode->i_mode)) {
C
Chris Mason 已提交
493 494
				printk(KERN_INFO "Btrfs symlink in snapshot "
				       "path, failed\n");
495 496 497 498 499 500 501 502 503
				error = -EMLINK;
				btrfs_free_path(path);
				goto out_drop_write;
			}
			test_oid =
				BTRFS_I(test->d_inode)->root->root_key.objectid;
			ret = btrfs_find_root_ref(snap_src->fs_info->tree_root,
				  path, test_oid, parent_oid);
			if (ret == 0) {
C
Chris Mason 已提交
504 505
				printk(KERN_INFO "Btrfs snapshot creation "
				       "failed, looping\n");
506 507 508 509 510 511 512 513 514
				error = -EMLINK;
				btrfs_free_path(path);
				goto out_drop_write;
			}
			btrfs_release_path(snap_src->fs_info->tree_root, path);
			test = test->d_parent;
		}
create:
		btrfs_free_path(path);
515 516 517 518 519
		error = create_snapshot(snap_src, dentry, name, namelen);
	} else {
		error = create_subvol(BTRFS_I(parent->dentry->d_inode)->root,
				      dentry, name, namelen);
	}
520 521 522 523 524 525 526 527 528 529 530 531 532 533
	if (error)
		goto out_drop_write;

	fsnotify_mkdir(parent->dentry->d_inode, dentry);
out_drop_write:
	mnt_drop_write(parent->mnt);
out_dput:
	dput(dentry);
out_unlock:
	mutex_unlock(&parent->dentry->d_inode->i_mutex);
	return error;
}


534
static int btrfs_defrag_file(struct file *file)
C
Christoph Hellwig 已提交
535 536 537 538
{
	struct inode *inode = fdentry(file)->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
539
	struct btrfs_ordered_extent *ordered;
C
Christoph Hellwig 已提交
540 541 542 543 544 545 546 547 548
	struct page *page;
	unsigned long last_index;
	unsigned long ra_pages = root->fs_info->bdi.ra_pages;
	unsigned long total_read = 0;
	u64 page_start;
	u64 page_end;
	unsigned long i;
	int ret;

J
Josef Bacik 已提交
549
	ret = btrfs_check_data_free_space(root, inode, inode->i_size);
C
Christoph Hellwig 已提交
550 551 552 553 554 555 556 557 558 559 560
	if (ret)
		return -ENOSPC;

	mutex_lock(&inode->i_mutex);
	last_index = inode->i_size >> PAGE_CACHE_SHIFT;
	for (i = 0; i <= last_index; i++) {
		if (total_read % ra_pages == 0) {
			btrfs_force_ra(inode->i_mapping, &file->f_ra, file, i,
				       min(last_index, i + ra_pages - 1));
		}
		total_read++;
561
again:
C
Christoph Hellwig 已提交
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579
		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;
			}
		}

		wait_on_page_writeback(page);

		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
		page_end = page_start + PAGE_CACHE_SIZE - 1;
		lock_extent(io_tree, page_start, page_end, GFP_NOFS);
580 581 582 583 584 585 586 587 588 589 590 591

		ordered = btrfs_lookup_ordered_extent(inode, page_start);
		if (ordered) {
			unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
			unlock_page(page);
			page_cache_release(page);
			btrfs_start_ordered_extent(inode, ordered, 1);
			btrfs_put_ordered_extent(ordered);
			goto again;
		}
		set_page_extent_mapped(page);

592 593 594 595 596 597
		/*
		 * this makes sure page_mkwrite is called on the
		 * page if it is dirtied again later
		 */
		clear_page_dirty_for_io(page);

598
		btrfs_set_extent_delalloc(inode, page_start, page_end);
C
Christoph Hellwig 已提交
599
		set_page_dirty(page);
C
Chris Mason 已提交
600
		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
C
Christoph Hellwig 已提交
601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
		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;
}

static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg)
{
	u64 new_size;
	u64 old_size;
	u64 devid = 1;
	struct btrfs_ioctl_vol_args *vol_args;
	struct btrfs_trans_handle *trans;
	struct btrfs_device *device = NULL;
	char *sizestr;
	char *devstr = NULL;
	int ret = 0;
	int namelen;
	int mod = 0;

Y
Yan Zheng 已提交
625 626 627
	if (root->fs_info->sb->s_flags & MS_RDONLY)
		return -EROFS;

628 629 630
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

L
Li Zefan 已提交
631 632 633
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);
634 635

	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
C
Christoph Hellwig 已提交
636 637
	namelen = strlen(vol_args->name);

638
	mutex_lock(&root->fs_info->volume_mutex);
C
Christoph Hellwig 已提交
639 640 641 642 643 644 645 646
	sizestr = vol_args->name;
	devstr = strchr(sizestr, ':');
	if (devstr) {
		char *end;
		sizestr = devstr + 1;
		*devstr = '\0';
		devstr = vol_args->name;
		devid = simple_strtoull(devstr, &end, 10);
647 648
		printk(KERN_INFO "resizing devid %llu\n",
		       (unsigned long long)devid);
C
Christoph Hellwig 已提交
649
	}
Y
Yan Zheng 已提交
650
	device = btrfs_find_device(root, devid, NULL, NULL);
C
Christoph Hellwig 已提交
651
	if (!device) {
652 653
		printk(KERN_INFO "resizer unable to find device %llu\n",
		       (unsigned long long)devid);
C
Christoph Hellwig 已提交
654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709
		ret = -EINVAL;
		goto out_unlock;
	}
	if (!strcmp(sizestr, "max"))
		new_size = device->bdev->bd_inode->i_size;
	else {
		if (sizestr[0] == '-') {
			mod = -1;
			sizestr++;
		} else if (sizestr[0] == '+') {
			mod = 1;
			sizestr++;
		}
		new_size = btrfs_parse_size(sizestr);
		if (new_size == 0) {
			ret = -EINVAL;
			goto out_unlock;
		}
	}

	old_size = device->total_bytes;

	if (mod < 0) {
		if (new_size > old_size) {
			ret = -EINVAL;
			goto out_unlock;
		}
		new_size = old_size - new_size;
	} else if (mod > 0) {
		new_size = old_size + new_size;
	}

	if (new_size < 256 * 1024 * 1024) {
		ret = -EINVAL;
		goto out_unlock;
	}
	if (new_size > device->bdev->bd_inode->i_size) {
		ret = -EFBIG;
		goto out_unlock;
	}

	do_div(new_size, root->sectorsize);
	new_size *= root->sectorsize;

	printk(KERN_INFO "new size for %s is %llu\n",
		device->name, (unsigned long long)new_size);

	if (new_size > old_size) {
		trans = btrfs_start_transaction(root, 1);
		ret = btrfs_grow_device(trans, device, new_size);
		btrfs_commit_transaction(trans, root);
	} else {
		ret = btrfs_shrink_device(device, new_size);
	}

out_unlock:
710
	mutex_unlock(&root->fs_info->volume_mutex);
C
Christoph Hellwig 已提交
711 712 713 714
	kfree(vol_args);
	return ret;
}

715
static noinline int btrfs_ioctl_snap_create(struct file *file,
716
					    void __user *arg, int subvol)
C
Christoph Hellwig 已提交
717
{
718
	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
C
Christoph Hellwig 已提交
719 720 721
	struct btrfs_ioctl_vol_args *vol_args;
	struct btrfs_dir_item *di;
	struct btrfs_path *path;
722
	struct file *src_file;
C
Christoph Hellwig 已提交
723 724
	u64 root_dirid;
	int namelen;
725
	int ret = 0;
C
Christoph Hellwig 已提交
726

Y
Yan Zheng 已提交
727 728 729
	if (root->fs_info->sb->s_flags & MS_RDONLY)
		return -EROFS;

L
Li Zefan 已提交
730 731 732
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);
C
Christoph Hellwig 已提交
733

734
	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
C
Christoph Hellwig 已提交
735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762
	namelen = strlen(vol_args->name);
	if (strchr(vol_args->name, '/')) {
		ret = -EINVAL;
		goto out;
	}

	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}

	root_dirid = root->fs_info->sb->s_root->d_inode->i_ino,
	di = btrfs_lookup_dir_item(NULL, root->fs_info->tree_root,
			    path, root_dirid,
			    vol_args->name, namelen, 0);
	btrfs_free_path(path);

	if (di && !IS_ERR(di)) {
		ret = -EEXIST;
		goto out;
	}

	if (IS_ERR(di)) {
		ret = PTR_ERR(di);
		goto out;
	}

763
	if (subvol) {
764 765
		ret = btrfs_mksubvol(&file->f_path, vol_args->name,
				     file->f_path.dentry->d_inode->i_mode,
766
				     namelen, NULL);
767
	} else {
768 769 770 771 772 773 774 775 776
		struct inode *src_inode;
		src_file = fget(vol_args->fd);
		if (!src_file) {
			ret = -EINVAL;
			goto out;
		}

		src_inode = src_file->f_path.dentry->d_inode;
		if (src_inode->i_sb != file->f_path.dentry->d_inode->i_sb) {
C
Chris Mason 已提交
777 778
			printk(KERN_INFO "btrfs: Snapshot src from "
			       "another FS\n");
779 780 781 782 783 784 785 786
			ret = -EINVAL;
			fput(src_file);
			goto out;
		}
		ret = btrfs_mksubvol(&file->f_path, vol_args->name,
			     file->f_path.dentry->d_inode->i_mode,
			     namelen, BTRFS_I(src_inode)->root);
		fput(src_file);
787 788
	}

C
Christoph Hellwig 已提交
789 790 791 792 793 794 795 796 797
out:
	kfree(vol_args);
	return ret;
}

static int btrfs_ioctl_defrag(struct file *file)
{
	struct inode *inode = fdentry(file)->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
Y
Yan Zheng 已提交
798 799 800 801 802
	int ret;

	ret = mnt_want_write(file->f_path.mnt);
	if (ret)
		return ret;
C
Christoph Hellwig 已提交
803 804 805

	switch (inode->i_mode & S_IFMT) {
	case S_IFDIR:
806 807 808 809
		if (!capable(CAP_SYS_ADMIN)) {
			ret = -EPERM;
			goto out;
		}
C
Christoph Hellwig 已提交
810 811 812 813
		btrfs_defrag_root(root, 0);
		btrfs_defrag_root(root->fs_info->extent_root, 0);
		break;
	case S_IFREG:
814 815 816 817
		if (!(file->f_mode & FMODE_WRITE)) {
			ret = -EINVAL;
			goto out;
		}
C
Christoph Hellwig 已提交
818 819 820
		btrfs_defrag_file(file);
		break;
	}
821
out:
822
	mnt_drop_write(file->f_path.mnt);
823
	return ret;
C
Christoph Hellwig 已提交
824 825
}

826
static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg)
C
Christoph Hellwig 已提交
827 828 829 830
{
	struct btrfs_ioctl_vol_args *vol_args;
	int ret;

831 832 833
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

L
Li Zefan 已提交
834 835 836
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);
C
Christoph Hellwig 已提交
837

838
	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
C
Christoph Hellwig 已提交
839 840 841 842 843 844
	ret = btrfs_init_new_device(root, vol_args->name);

	kfree(vol_args);
	return ret;
}

845
static long btrfs_ioctl_rm_dev(struct btrfs_root *root, void __user *arg)
C
Christoph Hellwig 已提交
846 847 848 849
{
	struct btrfs_ioctl_vol_args *vol_args;
	int ret;

850 851 852
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

Y
Yan Zheng 已提交
853 854 855
	if (root->fs_info->sb->s_flags & MS_RDONLY)
		return -EROFS;

L
Li Zefan 已提交
856 857 858
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);
C
Christoph Hellwig 已提交
859

860
	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
C
Christoph Hellwig 已提交
861 862 863 864 865 866
	ret = btrfs_rm_device(root, vol_args->name);

	kfree(vol_args);
	return ret;
}

867 868
static long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
		u64 off, u64 olen, u64 destoff)
C
Christoph Hellwig 已提交
869 870 871 872 873 874 875 876
{
	struct inode *inode = fdentry(file)->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct file *src_file;
	struct inode *src;
	struct btrfs_trans_handle *trans;
	struct btrfs_path *path;
	struct extent_buffer *leaf;
Y
Yan Zheng 已提交
877 878
	char *buf;
	struct btrfs_key key;
C
Christoph Hellwig 已提交
879 880
	u32 nritems;
	int slot;
Y
Yan Zheng 已提交
881
	int ret;
882 883 884
	u64 len = olen;
	u64 bs = root->fs_info->sb->s_blocksize;
	u64 hint_byte;
885

886 887 888 889 890 891 892 893 894 895
	/*
	 * TODO:
	 * - split compressed inline extents.  annoying: we need to
	 *   decompress into destination's address_space (the file offset
	 *   may change, so source mapping won't do), then recompress (or
	 *   otherwise reinsert) a subrange.
	 * - allow ranges within the same file to be cloned (provided
	 *   they don't overlap)?
	 */

896 897 898 899
	/* the destination must be opened for writing */
	if (!(file->f_mode & FMODE_WRITE))
		return -EINVAL;

Y
Yan Zheng 已提交
900 901 902 903
	ret = mnt_want_write(file->f_path.mnt);
	if (ret)
		return ret;

904
	src_file = fget(srcfd);
905 906 907 908
	if (!src_file) {
		ret = -EBADF;
		goto out_drop_write;
	}
C
Christoph Hellwig 已提交
909 910
	src = src_file->f_dentry->d_inode;

911 912 913 914
	ret = -EINVAL;
	if (src == inode)
		goto out_fput;

Y
Yan Zheng 已提交
915 916 917 918
	ret = -EISDIR;
	if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode))
		goto out_fput;

C
Christoph Hellwig 已提交
919
	ret = -EXDEV;
Y
Yan Zheng 已提交
920 921 922 923 924 925 926 927 928 929 930
	if (src->i_sb != inode->i_sb || BTRFS_I(src)->root != root)
		goto out_fput;

	ret = -ENOMEM;
	buf = vmalloc(btrfs_level_size(root, 0));
	if (!buf)
		goto out_fput;

	path = btrfs_alloc_path();
	if (!path) {
		vfree(buf);
C
Christoph Hellwig 已提交
931
		goto out_fput;
Y
Yan Zheng 已提交
932 933
	}
	path->reada = 2;
C
Christoph Hellwig 已提交
934 935 936 937 938 939 940 941 942

	if (inode < src) {
		mutex_lock(&inode->i_mutex);
		mutex_lock(&src->i_mutex);
	} else {
		mutex_lock(&src->i_mutex);
		mutex_lock(&inode->i_mutex);
	}

943 944 945
	/* determine range to clone */
	ret = -EINVAL;
	if (off >= src->i_size || off + len > src->i_size)
C
Christoph Hellwig 已提交
946
		goto out_unlock;
947 948 949 950 951 952 953 954 955 956 957 958
	if (len == 0)
		olen = len = src->i_size - off;
	/* if we extend to eof, continue to block boundary */
	if (off + len == src->i_size)
		len = ((src->i_size + bs-1) & ~(bs-1))
			- off;

	/* verify the end result is block aligned */
	if ((off & (bs-1)) ||
	    ((off + len) & (bs-1)))
		goto out_unlock;

C
Christoph Hellwig 已提交
959 960 961
	/* do any pending delalloc/csum calc on src, one way or
	   another, and lock file content */
	while (1) {
Z
Zheng Yan 已提交
962
		struct btrfs_ordered_extent *ordered;
963 964
		lock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
		ordered = btrfs_lookup_first_ordered_extent(inode, off+len);
Y
Yan Zheng 已提交
965
		if (BTRFS_I(src)->delalloc_bytes == 0 && !ordered)
C
Christoph Hellwig 已提交
966
			break;
967
		unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
Y
Yan Zheng 已提交
968 969
		if (ordered)
			btrfs_put_ordered_extent(ordered);
970
		btrfs_wait_ordered_range(src, off, off+len);
C
Christoph Hellwig 已提交
971 972
	}

Y
Yan Zheng 已提交
973 974 975
	trans = btrfs_start_transaction(root, 1);
	BUG_ON(!trans);

976
	/* punch hole in destination first */
977
	btrfs_drop_extents(trans, root, inode, off, off + len,
C
Chris Mason 已提交
978
			   off + len, 0, &hint_byte, 1);
979 980

	/* clone data */
C
Christoph Hellwig 已提交
981
	key.objectid = src->i_ino;
Y
Yan Zheng 已提交
982 983
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = 0;
C
Christoph Hellwig 已提交
984 985 986 987 988 989 990 991 992 993

	while (1) {
		/*
		 * note the key will change type as we walk through the
		 * tree.
		 */
		ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
		if (ret < 0)
			goto out;

Y
Yan Zheng 已提交
994 995
		nritems = btrfs_header_nritems(path->nodes[0]);
		if (path->slots[0] >= nritems) {
C
Christoph Hellwig 已提交
996 997 998 999 1000
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				goto out;
			if (ret > 0)
				break;
Y
Yan Zheng 已提交
1001
			nritems = btrfs_header_nritems(path->nodes[0]);
C
Christoph Hellwig 已提交
1002 1003 1004 1005
		}
		leaf = path->nodes[0];
		slot = path->slots[0];

Y
Yan Zheng 已提交
1006
		btrfs_item_key_to_cpu(leaf, &key, slot);
1007
		if (btrfs_key_type(&key) > BTRFS_EXTENT_DATA_KEY ||
C
Christoph Hellwig 已提交
1008 1009 1010
		    key.objectid != src->i_ino)
			break;

1011 1012 1013
		if (btrfs_key_type(&key) == BTRFS_EXTENT_DATA_KEY) {
			struct btrfs_file_extent_item *extent;
			int type;
Z
Zheng Yan 已提交
1014 1015
			u32 size;
			struct btrfs_key new_key;
1016 1017 1018
			u64 disko = 0, diskl = 0;
			u64 datao = 0, datal = 0;
			u8 comp;
Z
Zheng Yan 已提交
1019 1020 1021 1022 1023

			size = btrfs_item_size_nr(leaf, slot);
			read_extent_buffer(leaf, buf,
					   btrfs_item_ptr_offset(leaf, slot),
					   size);
1024 1025 1026 1027 1028

			extent = btrfs_item_ptr(leaf, slot,
						struct btrfs_file_extent_item);
			comp = btrfs_file_extent_compression(leaf, extent);
			type = btrfs_file_extent_type(leaf, extent);
1029 1030
			if (type == BTRFS_FILE_EXTENT_REG ||
			    type == BTRFS_FILE_EXTENT_PREALLOC) {
C
Chris Mason 已提交
1031 1032 1033 1034
				disko = btrfs_file_extent_disk_bytenr(leaf,
								      extent);
				diskl = btrfs_file_extent_disk_num_bytes(leaf,
								 extent);
1035
				datao = btrfs_file_extent_offset(leaf, extent);
C
Chris Mason 已提交
1036 1037
				datal = btrfs_file_extent_num_bytes(leaf,
								    extent);
1038 1039 1040 1041 1042
			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
				/* take upper bound, may be compressed */
				datal = btrfs_file_extent_ram_bytes(leaf,
								    extent);
			}
Z
Zheng Yan 已提交
1043 1044
			btrfs_release_path(root, path);

1045 1046 1047 1048
			if (key.offset + datal < off ||
			    key.offset >= off+len)
				goto next;

Z
Zheng Yan 已提交
1049 1050
			memcpy(&new_key, &key, sizeof(new_key));
			new_key.objectid = inode->i_ino;
1051
			new_key.offset = key.offset + destoff - off;
Z
Zheng Yan 已提交
1052

1053 1054
			if (type == BTRFS_FILE_EXTENT_REG ||
			    type == BTRFS_FILE_EXTENT_PREALLOC) {
1055 1056 1057 1058 1059 1060 1061 1062
				ret = btrfs_insert_empty_item(trans, root, path,
							      &new_key, size);
				if (ret)
					goto out;

				leaf = path->nodes[0];
				slot = path->slots[0];
				write_extent_buffer(leaf, buf,
Z
Zheng Yan 已提交
1063 1064
					    btrfs_item_ptr_offset(leaf, slot),
					    size);
Y
Yan Zheng 已提交
1065

1066
				extent = btrfs_item_ptr(leaf, slot,
C
Christoph Hellwig 已提交
1067
						struct btrfs_file_extent_item);
1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085

				if (off > key.offset) {
					datao += off - key.offset;
					datal -= off - key.offset;
				}
				if (key.offset + datao + datal + key.offset >
				    off + len)
					datal = off + len - key.offset - datao;
				/* disko == 0 means it's a hole */
				if (!disko)
					datao = 0;

				btrfs_set_file_extent_offset(leaf, extent,
							     datao);
				btrfs_set_file_extent_num_bytes(leaf, extent,
								datal);
				if (disko) {
					inode_add_bytes(inode, datal);
Y
Yan Zheng 已提交
1086
					ret = btrfs_inc_extent_ref(trans, root,
1087 1088 1089 1090
							disko, diskl, 0,
							root->root_key.objectid,
							inode->i_ino,
							new_key.offset - datao);
Z
Zheng Yan 已提交
1091
					BUG_ON(ret);
C
Christoph Hellwig 已提交
1092
				}
1093 1094 1095 1096 1097 1098 1099
			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
				u64 skip = 0;
				u64 trim = 0;
				if (off > key.offset) {
					skip = off - key.offset;
					new_key.offset += skip;
				}
C
Chris Mason 已提交
1100

1101 1102
				if (key.offset + datal > off+len)
					trim = key.offset + datal - (off+len);
C
Chris Mason 已提交
1103

1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
				if (comp && (skip || trim)) {
					ret = -EINVAL;
					goto out;
				}
				size -= skip + trim;
				datal -= skip + trim;
				ret = btrfs_insert_empty_item(trans, root, path,
							      &new_key, size);
				if (ret)
					goto out;

				if (skip) {
C
Chris Mason 已提交
1116 1117
					u32 start =
					  btrfs_file_extent_calc_inline_size(0);
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127
					memmove(buf+start, buf+start+skip,
						datal);
				}

				leaf = path->nodes[0];
				slot = path->slots[0];
				write_extent_buffer(leaf, buf,
					    btrfs_item_ptr_offset(leaf, slot),
					    size);
				inode_add_bytes(inode, datal);
C
Christoph Hellwig 已提交
1128
			}
1129 1130

			btrfs_mark_buffer_dirty(leaf);
Y
Yan Zheng 已提交
1131
		}
1132

C
Chris Mason 已提交
1133
next:
Z
Zheng Yan 已提交
1134
		btrfs_release_path(root, path);
C
Christoph Hellwig 已提交
1135 1136 1137 1138
		key.offset++;
	}
	ret = 0;
out:
Y
Yan Zheng 已提交
1139 1140 1141
	btrfs_release_path(root, path);
	if (ret == 0) {
		inode->i_mtime = inode->i_ctime = CURRENT_TIME;
1142 1143
		if (destoff + olen > inode->i_size)
			btrfs_i_size_write(inode, destoff + olen);
Y
Yan Zheng 已提交
1144 1145 1146
		BTRFS_I(inode)->flags = BTRFS_I(src)->flags;
		ret = btrfs_update_inode(trans, root, inode);
	}
C
Christoph Hellwig 已提交
1147
	btrfs_end_transaction(trans, root);
1148
	unlock_extent(&BTRFS_I(src)->io_tree, off, off+len, GFP_NOFS);
Y
Yan Zheng 已提交
1149 1150
	if (ret)
		vmtruncate(inode, 0);
C
Christoph Hellwig 已提交
1151 1152 1153
out_unlock:
	mutex_unlock(&src->i_mutex);
	mutex_unlock(&inode->i_mutex);
Y
Yan Zheng 已提交
1154 1155
	vfree(buf);
	btrfs_free_path(path);
C
Christoph Hellwig 已提交
1156 1157
out_fput:
	fput(src_file);
1158 1159
out_drop_write:
	mnt_drop_write(file->f_path.mnt);
C
Christoph Hellwig 已提交
1160 1161 1162
	return ret;
}

1163
static long btrfs_ioctl_clone_range(struct file *file, void __user *argp)
1164 1165 1166
{
	struct btrfs_ioctl_clone_range_args args;

1167
	if (copy_from_user(&args, argp, sizeof(args)))
1168 1169 1170 1171 1172
		return -EFAULT;
	return btrfs_ioctl_clone(file, args.src_fd, args.src_offset,
				 args.src_length, args.dest_offset);
}

C
Christoph Hellwig 已提交
1173 1174 1175 1176 1177 1178
/*
 * there are many ways the trans_start and trans_end ioctls can lead
 * to deadlocks.  They should only be used by applications that
 * basically own the machine, and have a very in depth understanding
 * of all the possible deadlocks and enospc problems.
 */
1179
static long btrfs_ioctl_trans_start(struct file *file)
C
Christoph Hellwig 已提交
1180 1181 1182 1183 1184 1185
{
	struct inode *inode = fdentry(file)->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	int ret = 0;

1186 1187 1188
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

C
Christoph Hellwig 已提交
1189 1190 1191 1192
	if (file->private_data) {
		ret = -EINPROGRESS;
		goto out;
	}
1193

Y
Yan Zheng 已提交
1194 1195 1196 1197
	ret = mnt_want_write(file->f_path.mnt);
	if (ret)
		goto out;

1198 1199 1200 1201 1202
	mutex_lock(&root->fs_info->trans_mutex);
	root->fs_info->open_ioctl_trans++;
	mutex_unlock(&root->fs_info->trans_mutex);

	trans = btrfs_start_ioctl_transaction(root, 0);
C
Christoph Hellwig 已提交
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
	if (trans)
		file->private_data = trans;
	else
		ret = -ENOMEM;
	/*printk(KERN_INFO "btrfs_ioctl_trans_start on %p\n", file);*/
out:
	return ret;
}

/*
 * there are many ways the trans_start and trans_end ioctls can lead
 * to deadlocks.  They should only be used by applications that
 * basically own the machine, and have a very in depth understanding
 * of all the possible deadlocks and enospc problems.
 */
long btrfs_ioctl_trans_end(struct file *file)
{
	struct inode *inode = fdentry(file)->d_inode;
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	int ret = 0;

	trans = file->private_data;
	if (!trans) {
		ret = -EINVAL;
		goto out;
	}
	btrfs_end_transaction(trans, root);
C
Christoph Hellwig 已提交
1231
	file->private_data = NULL;
1232 1233 1234 1235 1236

	mutex_lock(&root->fs_info->trans_mutex);
	root->fs_info->open_ioctl_trans--;
	mutex_unlock(&root->fs_info->trans_mutex);

1237 1238
	mnt_drop_write(file->f_path.mnt);

C
Christoph Hellwig 已提交
1239 1240 1241 1242 1243 1244 1245 1246
out:
	return ret;
}

long btrfs_ioctl(struct file *file, unsigned int
		cmd, unsigned long arg)
{
	struct btrfs_root *root = BTRFS_I(fdentry(file)->d_inode)->root;
1247
	void __user *argp = (void __user *)arg;
C
Christoph Hellwig 已提交
1248 1249

	switch (cmd) {
1250 1251 1252 1253 1254 1255
	case FS_IOC_GETFLAGS:
		return btrfs_ioctl_getflags(file, argp);
	case FS_IOC_SETFLAGS:
		return btrfs_ioctl_setflags(file, argp);
	case FS_IOC_GETVERSION:
		return btrfs_ioctl_getversion(file, argp);
C
Christoph Hellwig 已提交
1256
	case BTRFS_IOC_SNAP_CREATE:
1257
		return btrfs_ioctl_snap_create(file, argp, 0);
1258
	case BTRFS_IOC_SUBVOL_CREATE:
1259
		return btrfs_ioctl_snap_create(file, argp, 1);
C
Christoph Hellwig 已提交
1260 1261 1262
	case BTRFS_IOC_DEFRAG:
		return btrfs_ioctl_defrag(file);
	case BTRFS_IOC_RESIZE:
1263
		return btrfs_ioctl_resize(root, argp);
C
Christoph Hellwig 已提交
1264
	case BTRFS_IOC_ADD_DEV:
1265
		return btrfs_ioctl_add_dev(root, argp);
C
Christoph Hellwig 已提交
1266
	case BTRFS_IOC_RM_DEV:
1267
		return btrfs_ioctl_rm_dev(root, argp);
C
Christoph Hellwig 已提交
1268 1269 1270
	case BTRFS_IOC_BALANCE:
		return btrfs_balance(root->fs_info->dev_root);
	case BTRFS_IOC_CLONE:
1271 1272
		return btrfs_ioctl_clone(file, arg, 0, 0, 0);
	case BTRFS_IOC_CLONE_RANGE:
1273
		return btrfs_ioctl_clone_range(file, argp);
C
Christoph Hellwig 已提交
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284
	case BTRFS_IOC_TRANS_START:
		return btrfs_ioctl_trans_start(file);
	case BTRFS_IOC_TRANS_END:
		return btrfs_ioctl_trans_end(file);
	case BTRFS_IOC_SYNC:
		btrfs_sync_fs(file->f_dentry->d_sb, 1);
		return 0;
	}

	return -ENOTTY;
}