ioctl.c 136.8 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
#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/compat.h>
#include <linux/bit_spinlock.h>
38
#include <linux/security.h>
C
Christoph Hellwig 已提交
39
#include <linux/xattr.h>
40
#include <linux/vmalloc.h>
41
#include <linux/slab.h>
42
#include <linux/blkdev.h>
43
#include <linux/uuid.h>
44
#include <linux/btrfs.h>
M
Mark Fasheh 已提交
45
#include <linux/uaccess.h>
C
Christoph Hellwig 已提交
46 47 48 49 50 51
#include "ctree.h"
#include "disk-io.h"
#include "transaction.h"
#include "btrfs_inode.h"
#include "print-tree.h"
#include "volumes.h"
52
#include "locking.h"
53
#include "inode-map.h"
54
#include "backref.h"
55
#include "rcu-string.h"
56
#include "send.h"
57
#include "dev-replace.h"
58
#include "props.h"
59
#include "sysfs.h"
J
Josef Bacik 已提交
60
#include "qgroup.h"
61
#include "tree-log.h"
62
#include "compression.h"
C
Christoph Hellwig 已提交
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
#ifdef CONFIG_64BIT
/* If we have a 32-bit userspace and 64-bit kernel, then the UAPI
 * structures are incorrect, as the timespec structure from userspace
 * is 4 bytes too small. We define these alternatives here to teach
 * the kernel about the 32-bit struct packing.
 */
struct btrfs_ioctl_timespec_32 {
	__u64 sec;
	__u32 nsec;
} __attribute__ ((__packed__));

struct btrfs_ioctl_received_subvol_args_32 {
	char	uuid[BTRFS_UUID_SIZE];	/* in */
	__u64	stransid;		/* in */
	__u64	rtransid;		/* out */
	struct btrfs_ioctl_timespec_32 stime; /* in */
	struct btrfs_ioctl_timespec_32 rtime; /* out */
	__u64	flags;			/* in */
	__u64	reserved[16];		/* in */
} __attribute__ ((__packed__));

#define BTRFS_IOC_SET_RECEIVED_SUBVOL_32 _IOWR(BTRFS_IOCTL_MAGIC, 37, \
				struct btrfs_ioctl_received_subvol_args_32)
#endif


M
Mark Fasheh 已提交
90
static int btrfs_clone(struct inode *src, struct inode *inode,
91 92
		       u64 off, u64 olen, u64 olen_aligned, u64 destoff,
		       int no_time_update);
M
Mark Fasheh 已提交
93

94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123
/* 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;
L
Li Zefan 已提交
124 125 126
	if (flags & BTRFS_INODE_NODATACOW)
		iflags |= FS_NOCOW_FL;

127
	if (flags & BTRFS_INODE_NOCOMPRESS)
L
Li Zefan 已提交
128
		iflags |= FS_NOCOMP_FL;
129 130
	else if (flags & BTRFS_INODE_COMPRESS)
		iflags |= FS_COMPR_FL;
131 132 133 134 135 136 137 138 139 140

	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);
141
	unsigned int new_fl = 0;
142 143

	if (ip->flags & BTRFS_INODE_SYNC)
144
		new_fl |= S_SYNC;
145
	if (ip->flags & BTRFS_INODE_IMMUTABLE)
146
		new_fl |= S_IMMUTABLE;
147
	if (ip->flags & BTRFS_INODE_APPEND)
148
		new_fl |= S_APPEND;
149
	if (ip->flags & BTRFS_INODE_NOATIME)
150
		new_fl |= S_NOATIME;
151
	if (ip->flags & BTRFS_INODE_DIRSYNC)
152 153 154 155 156
		new_fl |= S_DIRSYNC;

	set_mask_bits(&inode->i_flags,
		      S_SYNC | S_APPEND | S_IMMUTABLE | S_NOATIME | S_DIRSYNC,
		      new_fl);
157 158 159 160 161
}

/*
 * Inherit flags from the parent inode.
 *
162
 * Currently only the compression flags and the cow flags are inherited.
163 164 165
 */
void btrfs_inherit_iflags(struct inode *inode, struct inode *dir)
{
166 167 168 169 170 171
	unsigned int flags;

	if (!dir)
		return;

	flags = BTRFS_I(dir)->flags;
172

173 174 175 176 177 178 179 180
	if (flags & BTRFS_INODE_NOCOMPRESS) {
		BTRFS_I(inode)->flags &= ~BTRFS_INODE_COMPRESS;
		BTRFS_I(inode)->flags |= BTRFS_INODE_NOCOMPRESS;
	} else if (flags & BTRFS_INODE_COMPRESS) {
		BTRFS_I(inode)->flags &= ~BTRFS_INODE_NOCOMPRESS;
		BTRFS_I(inode)->flags |= BTRFS_INODE_COMPRESS;
	}

L
Liu Bo 已提交
181
	if (flags & BTRFS_INODE_NODATACOW) {
182
		BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW;
L
Liu Bo 已提交
183 184 185
		if (S_ISREG(inode->i_mode))
			BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM;
	}
186 187 188 189 190 191

	btrfs_update_iflags(inode);
}

static int btrfs_ioctl_getflags(struct file *file, void __user *arg)
{
A
Al Viro 已提交
192
	struct btrfs_inode *ip = BTRFS_I(file_inode(file));
193 194 195 196 197 198 199
	unsigned int flags = btrfs_flags_to_ioctl(ip->flags);

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

200 201 202 203 204
static int check_flags(unsigned int flags)
{
	if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
		      FS_NOATIME_FL | FS_NODUMP_FL | \
		      FS_SYNC_FL | FS_DIRSYNC_FL | \
L
Li Zefan 已提交
205 206
		      FS_NOCOMP_FL | FS_COMPR_FL |
		      FS_NOCOW_FL))
207 208 209 210 211 212 213 214
		return -EOPNOTSUPP;

	if ((flags & FS_NOCOMP_FL) && (flags & FS_COMPR_FL))
		return -EINVAL;

	return 0;
}

215 216
static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
{
A
Al Viro 已提交
217
	struct inode *inode = file_inode(file);
218 219 220 221 222
	struct btrfs_inode *ip = BTRFS_I(inode);
	struct btrfs_root *root = ip->root;
	struct btrfs_trans_handle *trans;
	unsigned int flags, oldflags;
	int ret;
223 224
	u64 ip_oldflags;
	unsigned int i_oldflags;
225
	umode_t mode;
226

227 228 229
	if (!inode_owner_or_capable(inode))
		return -EPERM;

L
Li Zefan 已提交
230 231 232
	if (btrfs_root_readonly(root))
		return -EROFS;

233 234 235
	if (copy_from_user(&flags, arg, sizeof(flags)))
		return -EFAULT;

236 237 238
	ret = check_flags(flags);
	if (ret)
		return ret;
C
Christoph Hellwig 已提交
239

240 241 242 243
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;

A
Al Viro 已提交
244
	inode_lock(inode);
245

246 247
	ip_oldflags = ip->flags;
	i_oldflags = inode->i_flags;
248
	mode = inode->i_mode;
249

250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
	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;
		}
	}

	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;
283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
	if (flags & FS_NOCOW_FL) {
		if (S_ISREG(mode)) {
			/*
			 * It's safe to turn csums off here, no extents exist.
			 * Otherwise we want the flag to reflect the real COW
			 * status of the file and will not set it.
			 */
			if (inode->i_size == 0)
				ip->flags |= BTRFS_INODE_NODATACOW
					   | BTRFS_INODE_NODATASUM;
		} else {
			ip->flags |= BTRFS_INODE_NODATACOW;
		}
	} else {
		/*
298
		 * Revert back under same assumptions as above
299 300 301 302 303 304 305 306 307
		 */
		if (S_ISREG(mode)) {
			if (inode->i_size == 0)
				ip->flags &= ~(BTRFS_INODE_NODATACOW
				             | BTRFS_INODE_NODATASUM);
		} else {
			ip->flags &= ~BTRFS_INODE_NODATACOW;
		}
	}
308

309 310 311 312 313 314 315 316
	/*
	 * The COMPRESS flag can only be changed by users, while the NOCOMPRESS
	 * flag may be changed automatically if compression code won't make
	 * things smaller.
	 */
	if (flags & FS_NOCOMP_FL) {
		ip->flags &= ~BTRFS_INODE_COMPRESS;
		ip->flags |= BTRFS_INODE_NOCOMPRESS;
317 318 319 320

		ret = btrfs_set_prop(inode, "btrfs.compression", NULL, 0, 0);
		if (ret && ret != -ENODATA)
			goto out_drop;
321
	} else if (flags & FS_COMPR_FL) {
322 323
		const char *comp;

324 325
		ip->flags |= BTRFS_INODE_COMPRESS;
		ip->flags &= ~BTRFS_INODE_NOCOMPRESS;
326 327 328 329 330 331 332 333 334 335

		if (root->fs_info->compress_type == BTRFS_COMPRESS_LZO)
			comp = "lzo";
		else
			comp = "zlib";
		ret = btrfs_set_prop(inode, "btrfs.compression",
				     comp, strlen(comp), 0);
		if (ret)
			goto out_drop;

L
Li Zefan 已提交
336
	} else {
337 338 339
		ret = btrfs_set_prop(inode, "btrfs.compression", NULL, 0, 0);
		if (ret && ret != -ENODATA)
			goto out_drop;
L
Li Zefan 已提交
340
		ip->flags &= ~(BTRFS_INODE_COMPRESS | BTRFS_INODE_NOCOMPRESS);
341
	}
342

343
	trans = btrfs_start_transaction(root, 1);
344 345 346 347
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out_drop;
	}
348

349
	btrfs_update_iflags(inode);
350
	inode_inc_iversion(inode);
351
	inode->i_ctime = current_time(inode);
352 353 354
	ret = btrfs_update_inode(trans, root, inode);

	btrfs_end_transaction(trans, root);
355 356 357 358 359
 out_drop:
	if (ret) {
		ip->flags = ip_oldflags;
		inode->i_flags = i_oldflags;
	}
360 361

 out_unlock:
A
Al Viro 已提交
362
	inode_unlock(inode);
363
	mnt_drop_write_file(file);
364
	return ret;
365 366 367 368
}

static int btrfs_ioctl_getversion(struct file *file, int __user *arg)
{
A
Al Viro 已提交
369
	struct inode *inode = file_inode(file);
370 371 372

	return put_user(inode->i_generation, arg);
}
C
Christoph Hellwig 已提交
373

374 375
static noinline int btrfs_ioctl_fitrim(struct file *file, void __user *arg)
{
A
Al Viro 已提交
376
	struct btrfs_fs_info *fs_info = btrfs_sb(file_inode(file)->i_sb);
377 378 379 380 381
	struct btrfs_device *device;
	struct request_queue *q;
	struct fstrim_range range;
	u64 minlen = ULLONG_MAX;
	u64 num_devices = 0;
382
	u64 total_bytes = btrfs_super_total_bytes(fs_info->super_copy);
383 384 385 386 387
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

388 389 390
	rcu_read_lock();
	list_for_each_entry_rcu(device, &fs_info->fs_devices->devices,
				dev_list) {
391 392 393 394 395 396 397 398 399
		if (!device->bdev)
			continue;
		q = bdev_get_queue(device->bdev);
		if (blk_queue_discard(q)) {
			num_devices++;
			minlen = min((u64)q->limits.discard_granularity,
				     minlen);
		}
	}
400
	rcu_read_unlock();
401

402 403 404 405
	if (!num_devices)
		return -EOPNOTSUPP;
	if (copy_from_user(&range, arg, sizeof(range)))
		return -EFAULT;
406 407
	if (range.start > total_bytes ||
	    range.len < fs_info->sb->s_blocksize)
408
		return -EINVAL;
409

410
	range.len = min(range.len, total_bytes - range.start);
411
	range.minlen = max(range.minlen, minlen);
412
	ret = btrfs_trim_fs(fs_info->tree_root, &range);
413 414 415 416 417 418 419 420 421
	if (ret < 0)
		return ret;

	if (copy_to_user(arg, &range, sizeof(range)))
		return -EFAULT;

	return 0;
}

422 423
int btrfs_is_empty_uuid(u8 *uuid)
{
C
Chris Mason 已提交
424 425 426 427 428 429 430
	int i;

	for (i = 0; i < BTRFS_UUID_SIZE; i++) {
		if (uuid[i])
			return 0;
	}
	return 1;
431 432
}

433
static noinline int create_subvol(struct inode *dir,
434
				  struct dentry *dentry,
S
Sage Weil 已提交
435
				  char *name, int namelen,
A
Arne Jansen 已提交
436
				  u64 *async_transid,
437
				  struct btrfs_qgroup_inherit *inherit)
C
Christoph Hellwig 已提交
438 439 440
{
	struct btrfs_trans_handle *trans;
	struct btrfs_key key;
441
	struct btrfs_root_item *root_item;
C
Christoph Hellwig 已提交
442 443
	struct btrfs_inode_item *inode_item;
	struct extent_buffer *leaf;
444
	struct btrfs_root *root = BTRFS_I(dir)->root;
445
	struct btrfs_root *new_root;
446
	struct btrfs_block_rsv block_rsv;
447
	struct timespec cur_time = current_time(dir);
448
	struct inode *inode;
C
Christoph Hellwig 已提交
449 450 451 452
	int ret;
	int err;
	u64 objectid;
	u64 new_dirid = BTRFS_FIRST_FREE_OBJECTID;
453
	u64 index = 0;
454
	u64 qgroup_reserved;
455
	uuid_le new_uuid;
C
Christoph Hellwig 已提交
456

457 458 459 460
	root_item = kzalloc(sizeof(*root_item), GFP_KERNEL);
	if (!root_item)
		return -ENOMEM;

461
	ret = btrfs_find_free_objectid(root->fs_info->tree_root, &objectid);
462
	if (ret)
463
		goto fail_free;
464

465 466
	/*
	 * Don't create subvolume whose level is not zero. Or qgroup will be
467
	 * screwed up since it assumes subvolume qgroup's level to be 0.
468
	 */
469 470 471 472
	if (btrfs_qgroup_level(objectid)) {
		ret = -ENOSPC;
		goto fail_free;
	}
473

474
	btrfs_init_block_rsv(&block_rsv, BTRFS_BLOCK_RSV_TEMP);
J
Josef Bacik 已提交
475
	/*
476 477
	 * The same as the snapshot creation, please see the comment
	 * of create_snapshot().
J
Josef Bacik 已提交
478
	 */
479
	ret = btrfs_subvolume_reserve_metadata(root, &block_rsv,
480
					       8, &qgroup_reserved, false);
481
	if (ret)
482
		goto fail_free;
483 484 485 486

	trans = btrfs_start_transaction(root, 0);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
487 488
		btrfs_subvolume_release_metadata(root, &block_rsv,
						 qgroup_reserved);
489
		goto fail_free;
490 491 492
	}
	trans->block_rsv = &block_rsv;
	trans->bytes_reserved = block_rsv.size;
C
Christoph Hellwig 已提交
493

494
	ret = btrfs_qgroup_inherit(trans, root->fs_info, 0, objectid, inherit);
A
Arne Jansen 已提交
495 496 497
	if (ret)
		goto fail;

498
	leaf = btrfs_alloc_tree_block(trans, root, 0, objectid, NULL, 0, 0, 0);
499 500 501 502
	if (IS_ERR(leaf)) {
		ret = PTR_ERR(leaf);
		goto fail;
	}
C
Christoph Hellwig 已提交
503

504
	memzero_extent_buffer(leaf, 0, sizeof(struct btrfs_header));
C
Christoph Hellwig 已提交
505 506
	btrfs_set_header_bytenr(leaf, leaf->start);
	btrfs_set_header_generation(leaf, trans->transid);
507
	btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
C
Christoph Hellwig 已提交
508 509
	btrfs_set_header_owner(leaf, objectid);

510 511 512
	write_extent_buffer_fsid(leaf, root->fs_info->fsid);
	write_extent_buffer_chunk_tree_uuid(leaf,
			root->fs_info->chunk_tree_uuid);
C
Christoph Hellwig 已提交
513 514
	btrfs_mark_buffer_dirty(leaf);

515
	inode_item = &root_item->inode;
516 517 518
	btrfs_set_stack_inode_generation(inode_item, 1);
	btrfs_set_stack_inode_size(inode_item, 3);
	btrfs_set_stack_inode_nlink(inode_item, 1);
519 520
	btrfs_set_stack_inode_nbytes(inode_item,
				     root->fs_info->nodesize);
521
	btrfs_set_stack_inode_mode(inode_item, S_IFDIR | 0755);
C
Christoph Hellwig 已提交
522

523 524
	btrfs_set_root_flags(root_item, 0);
	btrfs_set_root_limit(root_item, 0);
525
	btrfs_set_stack_inode_flags(inode_item, BTRFS_INODE_ROOT_ITEM_INIT);
526

527 528 529 530 531 532
	btrfs_set_root_bytenr(root_item, leaf->start);
	btrfs_set_root_generation(root_item, trans->transid);
	btrfs_set_root_level(root_item, 0);
	btrfs_set_root_refs(root_item, 1);
	btrfs_set_root_used(root_item, leaf->len);
	btrfs_set_root_last_snapshot(root_item, 0);
C
Christoph Hellwig 已提交
533

534 535
	btrfs_set_root_generation_v2(root_item,
			btrfs_root_generation(root_item));
536
	uuid_le_gen(&new_uuid);
537 538 539 540 541 542
	memcpy(root_item->uuid, new_uuid.b, BTRFS_UUID_SIZE);
	btrfs_set_stack_timespec_sec(&root_item->otime, cur_time.tv_sec);
	btrfs_set_stack_timespec_nsec(&root_item->otime, cur_time.tv_nsec);
	root_item->ctime = root_item->otime;
	btrfs_set_root_ctransid(root_item, trans->transid);
	btrfs_set_root_otransid(root_item, trans->transid);
C
Christoph Hellwig 已提交
543

544
	btrfs_tree_unlock(leaf);
C
Christoph Hellwig 已提交
545 546 547
	free_extent_buffer(leaf);
	leaf = NULL;

548
	btrfs_set_root_dirid(root_item, new_dirid);
C
Christoph Hellwig 已提交
549 550

	key.objectid = objectid;
551
	key.offset = 0;
552
	key.type = BTRFS_ROOT_ITEM_KEY;
C
Christoph Hellwig 已提交
553
	ret = btrfs_insert_root(trans, root->fs_info->tree_root, &key,
554
				root_item);
C
Christoph Hellwig 已提交
555 556 557
	if (ret)
		goto fail;

558 559
	key.offset = (u64)-1;
	new_root = btrfs_read_fs_root_no_name(root->fs_info, &key);
560 561
	if (IS_ERR(new_root)) {
		ret = PTR_ERR(new_root);
562
		btrfs_abort_transaction(trans, ret);
563 564
		goto fail;
	}
565 566 567

	btrfs_record_root_in_trans(trans, new_root);

568
	ret = btrfs_create_subvol_root(trans, new_root, root, new_dirid);
569 570
	if (ret) {
		/* We potentially lose an unused inode item here */
571
		btrfs_abort_transaction(trans, ret);
572 573 574
		goto fail;
	}

575 576 577 578
	mutex_lock(&new_root->objectid_mutex);
	new_root->highest_objectid = new_dirid;
	mutex_unlock(&new_root->objectid_mutex);

C
Christoph Hellwig 已提交
579 580 581
	/*
	 * insert the directory item
	 */
582
	ret = btrfs_set_inode_index(dir, &index);
583
	if (ret) {
584
		btrfs_abort_transaction(trans, ret);
585 586
		goto fail;
	}
587 588

	ret = btrfs_insert_dir_item(trans, root,
589
				    name, namelen, dir, &key,
590
				    BTRFS_FT_DIR, index);
591
	if (ret) {
592
		btrfs_abort_transaction(trans, ret);
C
Christoph Hellwig 已提交
593
		goto fail;
594
	}
595

596 597 598 599
	btrfs_i_size_write(dir, dir->i_size + namelen * 2);
	ret = btrfs_update_inode(trans, root, dir);
	BUG_ON(ret);

600
	ret = btrfs_add_root_ref(trans, root->fs_info,
601
				 objectid, root->root_key.objectid,
L
Li Zefan 已提交
602
				 btrfs_ino(dir), index, name, namelen);
603
	BUG_ON(ret);
C
Christoph Hellwig 已提交
604

605 606
	ret = btrfs_uuid_tree_add(trans, root->fs_info, root_item->uuid,
				  BTRFS_UUID_KEY_SUBVOL, objectid);
607
	if (ret)
608
		btrfs_abort_transaction(trans, ret);
609

C
Christoph Hellwig 已提交
610
fail:
611
	kfree(root_item);
612 613
	trans->block_rsv = NULL;
	trans->bytes_reserved = 0;
614 615
	btrfs_subvolume_release_metadata(root, &block_rsv, qgroup_reserved);

S
Sage Weil 已提交
616 617 618
	if (async_transid) {
		*async_transid = trans->transid;
		err = btrfs_commit_transaction_async(trans, root, 1);
619 620
		if (err)
			err = btrfs_commit_transaction(trans, root);
S
Sage Weil 已提交
621 622 623
	} else {
		err = btrfs_commit_transaction(trans, root);
	}
C
Christoph Hellwig 已提交
624 625
	if (err && !ret)
		ret = err;
626

627 628
	if (!ret) {
		inode = btrfs_lookup_dentry(dir, dentry);
629 630
		if (IS_ERR(inode))
			return PTR_ERR(inode);
631 632
		d_instantiate(dentry, inode);
	}
C
Christoph Hellwig 已提交
633
	return ret;
634 635 636 637

fail_free:
	kfree(root_item);
	return ret;
C
Christoph Hellwig 已提交
638 639
}

640
static void btrfs_wait_for_no_snapshoting_writes(struct btrfs_root *root)
641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656
{
	s64 writers;
	DEFINE_WAIT(wait);

	do {
		prepare_to_wait(&root->subv_writers->wait, &wait,
				TASK_UNINTERRUPTIBLE);

		writers = percpu_counter_sum(&root->subv_writers->counter);
		if (writers)
			schedule();

		finish_wait(&root->subv_writers->wait, &wait);
	} while (writers);
}

657 658 659 660
static int create_snapshot(struct btrfs_root *root, struct inode *dir,
			   struct dentry *dentry, char *name, int namelen,
			   u64 *async_transid, bool readonly,
			   struct btrfs_qgroup_inherit *inherit)
C
Christoph Hellwig 已提交
661
{
662
	struct inode *inode;
C
Christoph Hellwig 已提交
663 664
	struct btrfs_pending_snapshot *pending_snapshot;
	struct btrfs_trans_handle *trans;
665
	int ret;
C
Christoph Hellwig 已提交
666

667
	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
C
Christoph Hellwig 已提交
668 669
		return -EINVAL;

670 671 672 673
	pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_NOFS);
	if (!pending_snapshot)
		return -ENOMEM;

674 675
	pending_snapshot->root_item = kzalloc(sizeof(struct btrfs_root_item),
			GFP_NOFS);
676 677
	pending_snapshot->path = btrfs_alloc_path();
	if (!pending_snapshot->root_item || !pending_snapshot->path) {
678 679 680 681
		ret = -ENOMEM;
		goto free_pending;
	}

682
	atomic_inc(&root->will_be_snapshoted);
683
	smp_mb__after_atomic();
684
	btrfs_wait_for_no_snapshoting_writes(root);
685

686 687
	ret = btrfs_start_delalloc_inodes(root, 0);
	if (ret)
688
		goto dec_and_free;
689

690
	btrfs_wait_ordered_extents(root, -1, 0, (u64)-1);
691

692 693
	btrfs_init_block_rsv(&pending_snapshot->block_rsv,
			     BTRFS_BLOCK_RSV_TEMP);
694 695 696 697 698 699
	/*
	 * 1 - parent dir inode
	 * 2 - dir entries
	 * 1 - root item
	 * 2 - root ref/backref
	 * 1 - root of snapshot
700
	 * 1 - UUID item
701 702
	 */
	ret = btrfs_subvolume_reserve_metadata(BTRFS_I(dir)->root,
703
					&pending_snapshot->block_rsv, 8,
704 705
					&pending_snapshot->qgroup_reserved,
					false);
706
	if (ret)
707
		goto dec_and_free;
708

709
	pending_snapshot->dentry = dentry;
C
Christoph Hellwig 已提交
710
	pending_snapshot->root = root;
L
Li Zefan 已提交
711
	pending_snapshot->readonly = readonly;
712
	pending_snapshot->dir = dir;
713
	pending_snapshot->inherit = inherit;
714

715
	trans = btrfs_start_transaction(root, 0);
716 717 718 719 720
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto fail;
	}

721
	spin_lock(&root->fs_info->trans_lock);
C
Christoph Hellwig 已提交
722 723
	list_add(&pending_snapshot->list,
		 &trans->transaction->pending_snapshots);
724
	spin_unlock(&root->fs_info->trans_lock);
S
Sage Weil 已提交
725 726 727 728
	if (async_transid) {
		*async_transid = trans->transid;
		ret = btrfs_commit_transaction_async(trans,
				     root->fs_info->extent_root, 1);
729 730
		if (ret)
			ret = btrfs_commit_transaction(trans, root);
S
Sage Weil 已提交
731 732 733 734
	} else {
		ret = btrfs_commit_transaction(trans,
					       root->fs_info->extent_root);
	}
735
	if (ret)
736
		goto fail;
737 738 739 740 741

	ret = pending_snapshot->error;
	if (ret)
		goto fail;

742 743 744 745
	ret = btrfs_orphan_cleanup(pending_snapshot->snap);
	if (ret)
		goto fail;

746
	inode = btrfs_lookup_dentry(d_inode(dentry->d_parent), dentry);
747 748 749 750
	if (IS_ERR(inode)) {
		ret = PTR_ERR(inode);
		goto fail;
	}
751

752 753 754
	d_instantiate(dentry, inode);
	ret = 0;
fail:
755 756 757
	btrfs_subvolume_release_metadata(BTRFS_I(dir)->root,
					 &pending_snapshot->block_rsv,
					 pending_snapshot->qgroup_reserved);
758
dec_and_free:
759 760
	if (atomic_dec_and_test(&root->will_be_snapshoted))
		wake_up_atomic_t(&root->will_be_snapshoted);
761 762
free_pending:
	kfree(pending_snapshot->root_item);
763
	btrfs_free_path(pending_snapshot->path);
764 765
	kfree(pending_snapshot);

C
Christoph Hellwig 已提交
766 767 768
	return ret;
}

769 770 771 772 773 774 775 776 777 778 779
/*  copy of may_delete in fs/namei.c()
 *	Check whether we can remove a link victim from directory dir, check
 *  whether the type of victim is right.
 *  1. We can't do it if dir is read-only (done in permission())
 *  2. We should have write and exec permissions on dir
 *  3. We can't remove anything from append-only dir
 *  4. We can't do anything with immutable dir (done in permission())
 *  5. If the sticky bit on dir is set we should either
 *	a. be owner of dir, or
 *	b. be owner of victim, or
 *	c. have CAP_FOWNER capability
780
 *  6. If the victim is append-only or immutable we can't do anything with
781 782 783 784 785 786 787 788
 *     links pointing to it.
 *  7. If we were asked to remove a directory and victim isn't one - ENOTDIR.
 *  8. If we were asked to remove a non-directory and victim isn't one - EISDIR.
 *  9. We can't remove a root or mountpoint.
 * 10. We don't allow removal of NFS sillyrenamed files; it's handled by
 *     nfs_async_unlink().
 */

789
static int btrfs_may_delete(struct inode *dir, struct dentry *victim, int isdir)
790 791 792
{
	int error;

793
	if (d_really_is_negative(victim))
794 795
		return -ENOENT;

796
	BUG_ON(d_inode(victim->d_parent) != dir);
797
	audit_inode_child(dir, victim, AUDIT_TYPE_CHILD_DELETE);
798 799 800 801 802 803

	error = inode_permission(dir, MAY_WRITE | MAY_EXEC);
	if (error)
		return error;
	if (IS_APPEND(dir))
		return -EPERM;
804 805
	if (check_sticky(dir, d_inode(victim)) || IS_APPEND(d_inode(victim)) ||
	    IS_IMMUTABLE(d_inode(victim)) || IS_SWAPFILE(d_inode(victim)))
806 807
		return -EPERM;
	if (isdir) {
808
		if (!d_is_dir(victim))
809 810 811
			return -ENOTDIR;
		if (IS_ROOT(victim))
			return -EBUSY;
812
	} else if (d_is_dir(victim))
813 814 815 816 817 818 819 820
		return -EISDIR;
	if (IS_DEADDIR(dir))
		return -ENOENT;
	if (victim->d_flags & DCACHE_NFSFS_RENAMED)
		return -EBUSY;
	return 0;
}

821 822 823
/* copy of may_create in fs/namei.c() */
static inline int btrfs_may_create(struct inode *dir, struct dentry *child)
{
824
	if (d_really_is_positive(child))
825 826 827 828 829 830 831 832 833 834 835
		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.
 */
836 837
static noinline int btrfs_mksubvol(struct path *parent,
				   char *name, int namelen,
S
Sage Weil 已提交
838
				   struct btrfs_root *snap_src,
A
Arne Jansen 已提交
839
				   u64 *async_transid, bool readonly,
840
				   struct btrfs_qgroup_inherit *inherit)
841
{
842
	struct inode *dir  = d_inode(parent->dentry);
843 844 845
	struct dentry *dentry;
	int error;

846 847 848
	error = down_write_killable_nested(&dir->i_rwsem, I_MUTEX_PARENT);
	if (error == -EINTR)
		return error;
849 850 851 852 853 854

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

855
	error = btrfs_may_create(dir, dentry);
856
	if (error)
857
		goto out_dput;
858

C
Chris Mason 已提交
859 860 861 862 863 864 865 866 867 868
	/*
	 * even if this name doesn't exist, we may get hash collisions.
	 * check for them now when we can safely fail
	 */
	error = btrfs_check_dir_item_collision(BTRFS_I(dir)->root,
					       dir->i_ino, name,
					       namelen);
	if (error)
		goto out_dput;

869 870 871 872 873
	down_read(&BTRFS_I(dir)->root->fs_info->subvol_sem);

	if (btrfs_root_refs(&BTRFS_I(dir)->root->root_item) == 0)
		goto out_up_read;

874
	if (snap_src) {
875
		error = create_snapshot(snap_src, dir, dentry, name, namelen,
A
Arne Jansen 已提交
876
					async_transid, readonly, inherit);
877
	} else {
878 879
		error = create_subvol(dir, dentry, name, namelen,
				      async_transid, inherit);
880
	}
881 882 883 884
	if (!error)
		fsnotify_mkdir(dir, dentry);
out_up_read:
	up_read(&BTRFS_I(dir)->root->fs_info->subvol_sem);
885 886 887
out_dput:
	dput(dentry);
out_unlock:
A
Al Viro 已提交
888
	inode_unlock(dir);
889 890 891
	return error;
}

C
Chris Mason 已提交
892 893 894 895 896 897 898
/*
 * When we're defragging a range, we don't want to kick it off again
 * if it is really just waiting for delalloc to send it down.
 * If we find a nice big extent or delalloc range for the bytes in the
 * file you want to defrag, we return 0 to let you know to skip this
 * part of the file
 */
899
static int check_defrag_in_cache(struct inode *inode, u64 offset, u32 thresh)
C
Chris Mason 已提交
900 901 902 903 904 905 906
{
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct extent_map *em = NULL;
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	u64 end;

	read_lock(&em_tree->lock);
907
	em = lookup_extent_mapping(em_tree, offset, PAGE_SIZE);
C
Chris Mason 已提交
908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933
	read_unlock(&em_tree->lock);

	if (em) {
		end = extent_map_end(em);
		free_extent_map(em);
		if (end - offset > thresh)
			return 0;
	}
	/* if we already have a nice delalloc here, just stop */
	thresh /= 2;
	end = count_range_bits(io_tree, &offset, offset + thresh,
			       thresh, EXTENT_DELALLOC, 1);
	if (end >= thresh)
		return 0;
	return 1;
}

/*
 * helper function to walk through a file and find extents
 * newer than a specific transid, and smaller than thresh.
 *
 * This is used by the defragging code to find new and small
 * extents
 */
static int find_new_extents(struct btrfs_root *root,
			    struct inode *inode, u64 newer_than,
934
			    u64 *off, u32 thresh)
C
Chris Mason 已提交
935 936 937 938 939 940 941
{
	struct btrfs_path *path;
	struct btrfs_key min_key;
	struct extent_buffer *leaf;
	struct btrfs_file_extent_item *extent;
	int type;
	int ret;
942
	u64 ino = btrfs_ino(inode);
C
Chris Mason 已提交
943 944 945 946 947

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

948
	min_key.objectid = ino;
C
Chris Mason 已提交
949 950 951
	min_key.type = BTRFS_EXTENT_DATA_KEY;
	min_key.offset = *off;

952
	while (1) {
953
		ret = btrfs_search_forward(root, &min_key, path, newer_than);
C
Chris Mason 已提交
954 955
		if (ret != 0)
			goto none;
956
process_slot:
957
		if (min_key.objectid != ino)
C
Chris Mason 已提交
958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974
			goto none;
		if (min_key.type != BTRFS_EXTENT_DATA_KEY)
			goto none;

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

		type = btrfs_file_extent_type(leaf, extent);
		if (type == BTRFS_FILE_EXTENT_REG &&
		    btrfs_file_extent_num_bytes(leaf, extent) < thresh &&
		    check_defrag_in_cache(inode, min_key.offset, thresh)) {
			*off = min_key.offset;
			btrfs_free_path(path);
			return 0;
		}

975 976 977 978 979 980
		path->slots[0]++;
		if (path->slots[0] < btrfs_header_nritems(leaf)) {
			btrfs_item_key_to_cpu(leaf, &min_key, path->slots[0]);
			goto process_slot;
		}

C
Chris Mason 已提交
981 982 983 984 985 986 987 988 989 990 991
		if (min_key.offset == (u64)-1)
			goto none;

		min_key.offset++;
		btrfs_release_path(path);
	}
none:
	btrfs_free_path(path);
	return -ENOENT;
}

L
Li Zefan 已提交
992
static struct extent_map *defrag_lookup_extent(struct inode *inode, u64 start)
993 994
{
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
L
Li Zefan 已提交
995 996
	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
	struct extent_map *em;
997
	u64 len = PAGE_SIZE;
998

L
Li Zefan 已提交
999 1000 1001 1002
	/*
	 * hopefully we have this extent in the tree already, try without
	 * the full extent lock
	 */
1003
	read_lock(&em_tree->lock);
L
Li Zefan 已提交
1004
	em = lookup_extent_mapping(em_tree, start, len);
1005 1006
	read_unlock(&em_tree->lock);

L
Li Zefan 已提交
1007
	if (!em) {
1008 1009 1010
		struct extent_state *cached = NULL;
		u64 end = start + len - 1;

L
Li Zefan 已提交
1011
		/* get the big lock and read metadata off disk */
1012
		lock_extent_bits(io_tree, start, end, &cached);
L
Li Zefan 已提交
1013
		em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
1014
		unlock_extent_cached(io_tree, start, end, &cached, GFP_NOFS);
L
Li Zefan 已提交
1015 1016 1017 1018 1019 1020 1021

		if (IS_ERR(em))
			return NULL;
	}

	return em;
}
1022

L
Li Zefan 已提交
1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
static bool defrag_check_next_extent(struct inode *inode, struct extent_map *em)
{
	struct extent_map *next;
	bool ret = true;

	/* this is the last extent */
	if (em->start + em->len >= i_size_read(inode))
		return false;

	next = defrag_lookup_extent(inode, em->start + em->len);
1033 1034 1035
	if (!next || next->block_start >= EXTENT_MAP_LAST_BYTE)
		ret = false;
	else if ((em->block_start + em->block_len == next->block_start) &&
1036
		 (em->block_len > SZ_128K && next->block_len > SZ_128K))
L
Li Zefan 已提交
1037 1038 1039
		ret = false;

	free_extent_map(next);
1040 1041 1042
	return ret;
}

1043
static int should_defrag_range(struct inode *inode, u64 start, u32 thresh,
1044 1045
			       u64 *last_len, u64 *skip, u64 *defrag_end,
			       int compress)
1046
{
L
Li Zefan 已提交
1047
	struct extent_map *em;
1048
	int ret = 1;
L
Li Zefan 已提交
1049
	bool next_mergeable = true;
1050
	bool prev_mergeable = true;
1051 1052

	/*
1053
	 * make sure that once we start defragging an extent, we keep on
1054 1055 1056 1057 1058 1059 1060
	 * defragging it
	 */
	if (start < *defrag_end)
		return 1;

	*skip = 0;

L
Li Zefan 已提交
1061 1062 1063
	em = defrag_lookup_extent(inode, start);
	if (!em)
		return 0;
1064 1065

	/* this will cover holes, and inline extents */
1066
	if (em->block_start >= EXTENT_MAP_LAST_BYTE) {
1067
		ret = 0;
1068 1069 1070
		goto out;
	}

1071 1072 1073
	if (!*defrag_end)
		prev_mergeable = false;

L
Li Zefan 已提交
1074
	next_mergeable = defrag_check_next_extent(inode, em);
1075
	/*
L
Li Zefan 已提交
1076 1077
	 * we hit a real extent, if it is big or the next extent is not a
	 * real extent, don't bother defragging it
1078
	 */
1079
	if (!compress && (*last_len == 0 || *last_len >= thresh) &&
1080
	    (em->len >= thresh || (!next_mergeable && !prev_mergeable)))
1081
		ret = 0;
1082
out:
1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102
	/*
	 * last_len ends up being a counter of how many bytes we've defragged.
	 * every time we choose not to defrag an extent, we reset *last_len
	 * so that the next tiny extent will force a defrag.
	 *
	 * The end result of this is that tiny extents before a single big
	 * extent will force at least part of that big extent to be defragged.
	 */
	if (ret) {
		*defrag_end = extent_map_end(em);
	} else {
		*last_len = 0;
		*skip = extent_map_end(em);
		*defrag_end = 0;
	}

	free_extent_map(em);
	return ret;
}

C
Chris Mason 已提交
1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117
/*
 * it doesn't do much good to defrag one or two pages
 * at a time.  This pulls in a nice chunk of pages
 * to COW and defrag.
 *
 * It also makes sure the delalloc code has enough
 * dirty data to avoid making new small extents as part
 * of the defrag
 *
 * It's a good idea to start RA on this range
 * before calling this.
 */
static int cluster_pages_for_defrag(struct inode *inode,
				    struct page **pages,
				    unsigned long start_index,
1118
				    unsigned long num_pages)
C
Christoph Hellwig 已提交
1119
{
C
Chris Mason 已提交
1120 1121 1122 1123
	unsigned long file_end;
	u64 isize = i_size_read(inode);
	u64 page_start;
	u64 page_end;
1124
	u64 page_cnt;
C
Chris Mason 已提交
1125 1126 1127
	int ret;
	int i;
	int i_done;
1128
	struct btrfs_ordered_extent *ordered;
C
Chris Mason 已提交
1129
	struct extent_state *cached_state = NULL;
1130
	struct extent_io_tree *tree;
1131
	gfp_t mask = btrfs_alloc_write_mask(inode->i_mapping);
C
Chris Mason 已提交
1132

1133
	file_end = (isize - 1) >> PAGE_SHIFT;
1134 1135 1136 1137
	if (!isize || start_index > file_end)
		return 0;

	page_cnt = min_t(u64, (u64)num_pages, (u64)file_end - start_index + 1);
C
Chris Mason 已提交
1138 1139

	ret = btrfs_delalloc_reserve_space(inode,
1140 1141
			start_index << PAGE_SHIFT,
			page_cnt << PAGE_SHIFT);
C
Chris Mason 已提交
1142 1143 1144
	if (ret)
		return ret;
	i_done = 0;
1145
	tree = &BTRFS_I(inode)->io_tree;
C
Chris Mason 已提交
1146 1147

	/* step one, lock all the pages */
1148
	for (i = 0; i < page_cnt; i++) {
C
Chris Mason 已提交
1149
		struct page *page;
1150
again:
1151
		page = find_or_create_page(inode->i_mapping,
1152
					   start_index + i, mask);
C
Chris Mason 已提交
1153 1154 1155
		if (!page)
			break;

1156
		page_start = page_offset(page);
1157
		page_end = page_start + PAGE_SIZE - 1;
1158
		while (1) {
1159
			lock_extent_bits(tree, page_start, page_end,
1160
					 &cached_state);
1161 1162
			ordered = btrfs_lookup_ordered_extent(inode,
							      page_start);
1163 1164
			unlock_extent_cached(tree, page_start, page_end,
					     &cached_state, GFP_NOFS);
1165 1166 1167 1168 1169 1170 1171
			if (!ordered)
				break;

			unlock_page(page);
			btrfs_start_ordered_extent(inode, ordered, 1);
			btrfs_put_ordered_extent(ordered);
			lock_page(page);
1172 1173 1174 1175 1176 1177
			/*
			 * we unlocked the page above, so we need check if
			 * it was released or not.
			 */
			if (page->mapping != inode->i_mapping) {
				unlock_page(page);
1178
				put_page(page);
1179 1180
				goto again;
			}
1181 1182
		}

C
Chris Mason 已提交
1183 1184 1185 1186 1187
		if (!PageUptodate(page)) {
			btrfs_readpage(NULL, page);
			lock_page(page);
			if (!PageUptodate(page)) {
				unlock_page(page);
1188
				put_page(page);
C
Chris Mason 已提交
1189 1190 1191 1192
				ret = -EIO;
				break;
			}
		}
1193 1194 1195

		if (page->mapping != inode->i_mapping) {
			unlock_page(page);
1196
			put_page(page);
1197 1198 1199
			goto again;
		}

C
Chris Mason 已提交
1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
		pages[i] = page;
		i_done++;
	}
	if (!i_done || ret)
		goto out;

	if (!(inode->i_sb->s_flags & MS_ACTIVE))
		goto out;

	/*
	 * so now we have a nice long stream of locked
	 * and up to date pages, lets wait on them
	 */
	for (i = 0; i < i_done; i++)
		wait_on_page_writeback(pages[i]);

	page_start = page_offset(pages[0]);
1217
	page_end = page_offset(pages[i_done - 1]) + PAGE_SIZE;
C
Chris Mason 已提交
1218 1219

	lock_extent_bits(&BTRFS_I(inode)->io_tree,
1220
			 page_start, page_end - 1, &cached_state);
C
Chris Mason 已提交
1221 1222
	clear_extent_bit(&BTRFS_I(inode)->io_tree, page_start,
			  page_end - 1, EXTENT_DIRTY | EXTENT_DELALLOC |
1223 1224
			  EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG, 0, 0,
			  &cached_state, GFP_NOFS);
C
Chris Mason 已提交
1225

1226
	if (i_done != page_cnt) {
1227 1228 1229
		spin_lock(&BTRFS_I(inode)->lock);
		BTRFS_I(inode)->outstanding_extents++;
		spin_unlock(&BTRFS_I(inode)->lock);
C
Chris Mason 已提交
1230
		btrfs_delalloc_release_space(inode,
1231 1232
				start_index << PAGE_SHIFT,
				(page_cnt - i_done) << PAGE_SHIFT);
C
Chris Mason 已提交
1233 1234 1235
	}


1236
	set_extent_defrag(&BTRFS_I(inode)->io_tree, page_start, page_end - 1,
1237
			  &cached_state);
C
Chris Mason 已提交
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248

	unlock_extent_cached(&BTRFS_I(inode)->io_tree,
			     page_start, page_end - 1, &cached_state,
			     GFP_NOFS);

	for (i = 0; i < i_done; i++) {
		clear_page_dirty_for_io(pages[i]);
		ClearPageChecked(pages[i]);
		set_page_extent_mapped(pages[i]);
		set_page_dirty(pages[i]);
		unlock_page(pages[i]);
1249
		put_page(pages[i]);
C
Chris Mason 已提交
1250 1251 1252 1253 1254
	}
	return i_done;
out:
	for (i = 0; i < i_done; i++) {
		unlock_page(pages[i]);
1255
		put_page(pages[i]);
C
Chris Mason 已提交
1256
	}
1257
	btrfs_delalloc_release_space(inode,
1258 1259
			start_index << PAGE_SHIFT,
			page_cnt << PAGE_SHIFT);
C
Chris Mason 已提交
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
	return ret;

}

int btrfs_defrag_file(struct inode *inode, struct file *file,
		      struct btrfs_ioctl_defrag_range_args *range,
		      u64 newer_than, unsigned long max_to_defrag)
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct file_ra_state *ra = NULL;
C
Christoph Hellwig 已提交
1270
	unsigned long last_index;
1271
	u64 isize = i_size_read(inode);
1272 1273 1274
	u64 last_len = 0;
	u64 skip = 0;
	u64 defrag_end = 0;
C
Chris Mason 已提交
1275
	u64 newer_off = range->start;
C
Christoph Hellwig 已提交
1276
	unsigned long i;
1277
	unsigned long ra_index = 0;
C
Christoph Hellwig 已提交
1278
	int ret;
C
Chris Mason 已提交
1279
	int defrag_count = 0;
1280
	int compress_type = BTRFS_COMPRESS_ZLIB;
1281
	u32 extent_thresh = range->extent_thresh;
1282
	unsigned long max_cluster = SZ_256K >> PAGE_SHIFT;
1283
	unsigned long cluster = max_cluster;
1284
	u64 new_align = ~((u64)SZ_128K - 1);
C
Chris Mason 已提交
1285 1286
	struct page **pages = NULL;

1287 1288 1289 1290 1291
	if (isize == 0)
		return 0;

	if (range->start >= isize)
		return -EINVAL;
1292 1293 1294 1295 1296 1297 1298

	if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) {
		if (range->compress_type > BTRFS_COMPRESS_TYPES)
			return -EINVAL;
		if (range->compress_type)
			compress_type = range->compress_type;
	}
C
Christoph Hellwig 已提交
1299

1300
	if (extent_thresh == 0)
1301
		extent_thresh = SZ_256K;
1302

C
Chris Mason 已提交
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
	/*
	 * if we were not given a file, allocate a readahead
	 * context
	 */
	if (!file) {
		ra = kzalloc(sizeof(*ra), GFP_NOFS);
		if (!ra)
			return -ENOMEM;
		file_ra_state_init(ra, inode->i_mapping);
	} else {
		ra = &file->f_ra;
	}

1316
	pages = kmalloc_array(max_cluster, sizeof(struct page *),
C
Chris Mason 已提交
1317 1318 1319 1320 1321 1322 1323
			GFP_NOFS);
	if (!pages) {
		ret = -ENOMEM;
		goto out_ra;
	}

	/* find the last page to defrag */
C
Chris Mason 已提交
1324
	if (range->start + range->len > range->start) {
1325
		last_index = min_t(u64, isize - 1,
1326
			 range->start + range->len - 1) >> PAGE_SHIFT;
C
Chris Mason 已提交
1327
	} else {
1328
		last_index = (isize - 1) >> PAGE_SHIFT;
C
Chris Mason 已提交
1329 1330
	}

C
Chris Mason 已提交
1331 1332
	if (newer_than) {
		ret = find_new_extents(root, inode, newer_than,
1333
				       &newer_off, SZ_64K);
C
Chris Mason 已提交
1334 1335 1336 1337 1338 1339
		if (!ret) {
			range->start = newer_off;
			/*
			 * we always align our defrag to help keep
			 * the extents in the file evenly spaced
			 */
1340
			i = (newer_off & new_align) >> PAGE_SHIFT;
C
Chris Mason 已提交
1341 1342 1343
		} else
			goto out_ra;
	} else {
1344
		i = range->start >> PAGE_SHIFT;
C
Chris Mason 已提交
1345 1346
	}
	if (!max_to_defrag)
1347
		max_to_defrag = last_index - i + 1;
C
Chris Mason 已提交
1348

L
Li Zefan 已提交
1349 1350 1351 1352 1353 1354 1355
	/*
	 * make writeback starts from i, so the defrag range can be
	 * written sequentially.
	 */
	if (i < inode->i_mapping->writeback_index)
		inode->i_mapping->writeback_index = i;

1356
	while (i <= last_index && defrag_count < max_to_defrag &&
1357
	       (i < DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE))) {
C
Chris Mason 已提交
1358 1359 1360 1361 1362 1363 1364
		/*
		 * make sure we stop running if someone unmounts
		 * the FS
		 */
		if (!(inode->i_sb->s_flags & MS_ACTIVE))
			break;

1365
		if (btrfs_defrag_cancelled(root->fs_info)) {
1366
			btrfs_debug(root->fs_info, "defrag_file cancelled");
1367 1368 1369 1370
			ret = -EAGAIN;
			break;
		}

1371
		if (!should_defrag_range(inode, (u64)i << PAGE_SHIFT,
L
Li Zefan 已提交
1372
					 extent_thresh, &last_len, &skip,
1373 1374
					 &defrag_end, range->flags &
					 BTRFS_DEFRAG_RANGE_COMPRESS)) {
1375 1376 1377 1378 1379
			unsigned long next;
			/*
			 * the should_defrag function tells us how much to skip
			 * bump our counter by the suggested amount
			 */
1380
			next = DIV_ROUND_UP(skip, PAGE_SIZE);
1381 1382 1383
			i = max(i + 1, next);
			continue;
		}
1384 1385

		if (!newer_than) {
1386 1387
			cluster = (PAGE_ALIGN(defrag_end) >>
				   PAGE_SHIFT) - i;
1388 1389 1390 1391 1392 1393 1394 1395 1396
			cluster = min(cluster, max_cluster);
		} else {
			cluster = max_cluster;
		}

		if (i + cluster > ra_index) {
			ra_index = max(i, ra_index);
			btrfs_force_ra(inode->i_mapping, ra, file, ra_index,
				       cluster);
1397
			ra_index += cluster;
1398
		}
1399

A
Al Viro 已提交
1400
		inode_lock(inode);
1401 1402
		if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)
			BTRFS_I(inode)->force_compress = compress_type;
1403
		ret = cluster_pages_for_defrag(inode, pages, i, cluster);
1404
		if (ret < 0) {
A
Al Viro 已提交
1405
			inode_unlock(inode);
C
Chris Mason 已提交
1406
			goto out_ra;
1407
		}
C
Chris Mason 已提交
1408 1409

		defrag_count += ret;
1410
		balance_dirty_pages_ratelimited(inode->i_mapping);
A
Al Viro 已提交
1411
		inode_unlock(inode);
C
Chris Mason 已提交
1412 1413 1414 1415 1416

		if (newer_than) {
			if (newer_off == (u64)-1)
				break;

1417 1418 1419
			if (ret > 0)
				i += ret;

C
Chris Mason 已提交
1420
			newer_off = max(newer_off + 1,
1421
					(u64)i << PAGE_SHIFT);
C
Chris Mason 已提交
1422

1423 1424
			ret = find_new_extents(root, inode, newer_than,
					       &newer_off, SZ_64K);
C
Chris Mason 已提交
1425 1426
			if (!ret) {
				range->start = newer_off;
1427
				i = (newer_off & new_align) >> PAGE_SHIFT;
C
Chris Mason 已提交
1428 1429
			} else {
				break;
C
Christoph Hellwig 已提交
1430
			}
C
Chris Mason 已提交
1431
		} else {
1432
			if (ret > 0) {
L
Li Zefan 已提交
1433
				i += ret;
1434
				last_len += ret << PAGE_SHIFT;
1435
			} else {
L
Li Zefan 已提交
1436
				i++;
1437 1438
				last_len = 0;
			}
C
Christoph Hellwig 已提交
1439 1440 1441
		}
	}

1442
	if ((range->flags & BTRFS_DEFRAG_RANGE_START_IO)) {
C
Chris Mason 已提交
1443
		filemap_flush(inode->i_mapping);
1444 1445 1446 1447
		if (test_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
			     &BTRFS_I(inode)->runtime_flags))
			filemap_flush(inode->i_mapping);
	}
C
Chris Mason 已提交
1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463

	if ((range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)) {
		/* the filemap_flush will queue IO into the worker threads, but
		 * we have to make sure the IO is actually started and that
		 * ordered extents get created before we return
		 */
		atomic_inc(&root->fs_info->async_submit_draining);
		while (atomic_read(&root->fs_info->nr_async_submits) ||
		      atomic_read(&root->fs_info->async_delalloc_pages)) {
			wait_event(root->fs_info->async_submit_wait,
			   (atomic_read(&root->fs_info->nr_async_submits) == 0 &&
			    atomic_read(&root->fs_info->async_delalloc_pages) == 0));
		}
		atomic_dec(&root->fs_info->async_submit_draining);
	}

1464
	if (range->compress_type == BTRFS_COMPRESS_LZO) {
1465
		btrfs_set_fs_incompat(root->fs_info, COMPRESS_LZO);
1466 1467
	}

1468
	ret = defrag_count;
1469

C
Chris Mason 已提交
1470
out_ra:
1471
	if (range->flags & BTRFS_DEFRAG_RANGE_COMPRESS) {
A
Al Viro 已提交
1472
		inode_lock(inode);
1473
		BTRFS_I(inode)->force_compress = BTRFS_COMPRESS_NONE;
A
Al Viro 已提交
1474
		inode_unlock(inode);
1475
	}
C
Chris Mason 已提交
1476 1477 1478
	if (!file)
		kfree(ra);
	kfree(pages);
1479
	return ret;
C
Christoph Hellwig 已提交
1480 1481
}

1482
static noinline int btrfs_ioctl_resize(struct file *file,
1483
					void __user *arg)
C
Christoph Hellwig 已提交
1484 1485 1486 1487
{
	u64 new_size;
	u64 old_size;
	u64 devid = 1;
A
Al Viro 已提交
1488
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
C
Christoph Hellwig 已提交
1489 1490 1491 1492
	struct btrfs_ioctl_vol_args *vol_args;
	struct btrfs_trans_handle *trans;
	struct btrfs_device *device = NULL;
	char *sizestr;
1493
	char *retptr;
C
Christoph Hellwig 已提交
1494 1495 1496 1497
	char *devstr = NULL;
	int ret = 0;
	int mod = 0;

1498 1499 1500
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

1501 1502 1503 1504
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;

1505 1506
	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
			1)) {
1507
		mnt_drop_write_file(file);
1508
		return BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
1509 1510
	}

1511
	mutex_lock(&root->fs_info->volume_mutex);
L
Li Zefan 已提交
1512
	vol_args = memdup_user(arg, sizeof(*vol_args));
1513 1514 1515 1516
	if (IS_ERR(vol_args)) {
		ret = PTR_ERR(vol_args);
		goto out;
	}
1517 1518

	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
C
Christoph Hellwig 已提交
1519 1520 1521 1522 1523 1524 1525

	sizestr = vol_args->name;
	devstr = strchr(sizestr, ':');
	if (devstr) {
		sizestr = devstr + 1;
		*devstr = '\0';
		devstr = vol_args->name;
1526 1527 1528
		ret = kstrtoull(devstr, 10, &devid);
		if (ret)
			goto out_free;
1529 1530 1531 1532
		if (!devid) {
			ret = -EINVAL;
			goto out_free;
		}
1533
		btrfs_info(root->fs_info, "resizing devid %llu", devid);
C
Christoph Hellwig 已提交
1534
	}
M
Miao Xie 已提交
1535

1536
	device = btrfs_find_device(root->fs_info, devid, NULL, NULL);
C
Christoph Hellwig 已提交
1537
	if (!device) {
1538
		btrfs_info(root->fs_info, "resizer unable to find device %llu",
1539
		       devid);
1540
		ret = -ENODEV;
1541
		goto out_free;
C
Christoph Hellwig 已提交
1542
	}
M
Miao Xie 已提交
1543 1544

	if (!device->writeable) {
1545 1546
		btrfs_info(root->fs_info,
			   "resizer unable to apply on readonly device %llu",
1547
		       devid);
1548
		ret = -EPERM;
L
Liu Bo 已提交
1549 1550 1551
		goto out_free;
	}

C
Christoph Hellwig 已提交
1552 1553 1554 1555 1556 1557 1558 1559 1560 1561
	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++;
		}
1562 1563
		new_size = memparse(sizestr, &retptr);
		if (*retptr != '\0' || new_size == 0) {
C
Christoph Hellwig 已提交
1564
			ret = -EINVAL;
1565
			goto out_free;
C
Christoph Hellwig 已提交
1566 1567 1568
		}
	}

1569
	if (device->is_tgtdev_for_dev_replace) {
1570
		ret = -EPERM;
1571 1572 1573
		goto out_free;
	}

1574
	old_size = btrfs_device_get_total_bytes(device);
C
Christoph Hellwig 已提交
1575 1576 1577 1578

	if (mod < 0) {
		if (new_size > old_size) {
			ret = -EINVAL;
1579
			goto out_free;
C
Christoph Hellwig 已提交
1580 1581 1582
		}
		new_size = old_size - new_size;
	} else if (mod > 0) {
1583
		if (new_size > ULLONG_MAX - old_size) {
1584
			ret = -ERANGE;
1585 1586
			goto out_free;
		}
C
Christoph Hellwig 已提交
1587 1588 1589
		new_size = old_size + new_size;
	}

1590
	if (new_size < SZ_256M) {
C
Christoph Hellwig 已提交
1591
		ret = -EINVAL;
1592
		goto out_free;
C
Christoph Hellwig 已提交
1593 1594 1595
	}
	if (new_size > device->bdev->bd_inode->i_size) {
		ret = -EFBIG;
1596
		goto out_free;
C
Christoph Hellwig 已提交
1597 1598
	}

1599 1600
	new_size = div_u64(new_size, root->fs_info->sectorsize);
	new_size *= root->fs_info->sectorsize;
C
Christoph Hellwig 已提交
1601

1602
	btrfs_info_in_rcu(root->fs_info, "new size for %s is %llu",
1603
		      rcu_str_deref(device->name), new_size);
C
Christoph Hellwig 已提交
1604 1605

	if (new_size > old_size) {
1606
		trans = btrfs_start_transaction(root, 0);
1607 1608
		if (IS_ERR(trans)) {
			ret = PTR_ERR(trans);
1609
			goto out_free;
1610
		}
C
Christoph Hellwig 已提交
1611 1612
		ret = btrfs_grow_device(trans, device, new_size);
		btrfs_commit_transaction(trans, root);
1613
	} else if (new_size < old_size) {
C
Christoph Hellwig 已提交
1614
		ret = btrfs_shrink_device(device, new_size);
1615
	} /* equal, nothing need to do */
C
Christoph Hellwig 已提交
1616

1617
out_free:
C
Christoph Hellwig 已提交
1618
	kfree(vol_args);
1619 1620
out:
	mutex_unlock(&root->fs_info->volume_mutex);
1621
	atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0);
1622
	mnt_drop_write_file(file);
C
Christoph Hellwig 已提交
1623 1624 1625
	return ret;
}

S
Sage Weil 已提交
1626
static noinline int btrfs_ioctl_snap_create_transid(struct file *file,
A
Arne Jansen 已提交
1627 1628
				char *name, unsigned long fd, int subvol,
				u64 *transid, bool readonly,
1629
				struct btrfs_qgroup_inherit *inherit)
C
Christoph Hellwig 已提交
1630 1631
{
	int namelen;
1632
	int ret = 0;
C
Christoph Hellwig 已提交
1633

1634 1635 1636
	if (!S_ISDIR(file_inode(file)->i_mode))
		return -ENOTDIR;

1637 1638 1639 1640
	ret = mnt_want_write_file(file);
	if (ret)
		goto out;

S
Sage Weil 已提交
1641 1642
	namelen = strlen(name);
	if (strchr(name, '/')) {
C
Christoph Hellwig 已提交
1643
		ret = -EINVAL;
1644
		goto out_drop_write;
C
Christoph Hellwig 已提交
1645 1646
	}

1647 1648 1649
	if (name[0] == '.' &&
	   (namelen == 1 || (name[1] == '.' && namelen == 2))) {
		ret = -EEXIST;
1650
		goto out_drop_write;
1651 1652
	}

1653
	if (subvol) {
S
Sage Weil 已提交
1654
		ret = btrfs_mksubvol(&file->f_path, name, namelen,
A
Arne Jansen 已提交
1655
				     NULL, transid, readonly, inherit);
1656
	} else {
1657
		struct fd src = fdget(fd);
1658
		struct inode *src_inode;
1659
		if (!src.file) {
1660
			ret = -EINVAL;
1661
			goto out_drop_write;
1662 1663
		}

A
Al Viro 已提交
1664 1665
		src_inode = file_inode(src.file);
		if (src_inode->i_sb != file_inode(file)->i_sb) {
J
Josef Bacik 已提交
1666
			btrfs_info(BTRFS_I(file_inode(file))->root->fs_info,
1667
				   "Snapshot src from another FS");
1668
			ret = -EXDEV;
1669 1670 1671 1672 1673 1674
		} else if (!inode_owner_or_capable(src_inode)) {
			/*
			 * Subvolume creation is not restricted, but snapshots
			 * are limited to own subvolumes only
			 */
			ret = -EPERM;
1675 1676 1677 1678
		} else {
			ret = btrfs_mksubvol(&file->f_path, name, namelen,
					     BTRFS_I(src_inode)->root,
					     transid, readonly, inherit);
1679
		}
1680
		fdput(src);
1681
	}
1682 1683
out_drop_write:
	mnt_drop_write_file(file);
C
Christoph Hellwig 已提交
1684
out:
S
Sage Weil 已提交
1685 1686 1687 1688
	return ret;
}

static noinline int btrfs_ioctl_snap_create(struct file *file,
1689
					    void __user *arg, int subvol)
S
Sage Weil 已提交
1690
{
1691
	struct btrfs_ioctl_vol_args *vol_args;
S
Sage Weil 已提交
1692 1693
	int ret;

1694 1695 1696
	if (!S_ISDIR(file_inode(file)->i_mode))
		return -ENOTDIR;

1697 1698 1699 1700
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);
	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
S
Sage Weil 已提交
1701

1702
	ret = btrfs_ioctl_snap_create_transid(file, vol_args->name,
L
Li Zefan 已提交
1703
					      vol_args->fd, subvol,
A
Arne Jansen 已提交
1704
					      NULL, false, NULL);
1705

1706 1707 1708
	kfree(vol_args);
	return ret;
}
1709

1710 1711 1712 1713 1714 1715 1716
static noinline int btrfs_ioctl_snap_create_v2(struct file *file,
					       void __user *arg, int subvol)
{
	struct btrfs_ioctl_vol_args_v2 *vol_args;
	int ret;
	u64 transid = 0;
	u64 *ptr = NULL;
L
Li Zefan 已提交
1717
	bool readonly = false;
A
Arne Jansen 已提交
1718
	struct btrfs_qgroup_inherit *inherit = NULL;
1719

1720 1721 1722
	if (!S_ISDIR(file_inode(file)->i_mode))
		return -ENOTDIR;

1723 1724 1725 1726
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);
	vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0';
1727

L
Li Zefan 已提交
1728
	if (vol_args->flags &
A
Arne Jansen 已提交
1729 1730
	    ~(BTRFS_SUBVOL_CREATE_ASYNC | BTRFS_SUBVOL_RDONLY |
	      BTRFS_SUBVOL_QGROUP_INHERIT)) {
L
Li Zefan 已提交
1731
		ret = -EOPNOTSUPP;
D
Dan Carpenter 已提交
1732
		goto free_args;
S
Sage Weil 已提交
1733
	}
1734 1735 1736

	if (vol_args->flags & BTRFS_SUBVOL_CREATE_ASYNC)
		ptr = &transid;
L
Li Zefan 已提交
1737 1738
	if (vol_args->flags & BTRFS_SUBVOL_RDONLY)
		readonly = true;
A
Arne Jansen 已提交
1739
	if (vol_args->flags & BTRFS_SUBVOL_QGROUP_INHERIT) {
1740
		if (vol_args->size > PAGE_SIZE) {
A
Arne Jansen 已提交
1741
			ret = -EINVAL;
D
Dan Carpenter 已提交
1742
			goto free_args;
A
Arne Jansen 已提交
1743 1744 1745 1746
		}
		inherit = memdup_user(vol_args->qgroup_inherit, vol_args->size);
		if (IS_ERR(inherit)) {
			ret = PTR_ERR(inherit);
D
Dan Carpenter 已提交
1747
			goto free_args;
A
Arne Jansen 已提交
1748 1749
		}
	}
1750 1751

	ret = btrfs_ioctl_snap_create_transid(file, vol_args->name,
A
Arne Jansen 已提交
1752
					      vol_args->fd, subvol, ptr,
1753
					      readonly, inherit);
D
Dan Carpenter 已提交
1754 1755
	if (ret)
		goto free_inherit;
1756

D
Dan Carpenter 已提交
1757 1758 1759 1760
	if (ptr && copy_to_user(arg +
				offsetof(struct btrfs_ioctl_vol_args_v2,
					transid),
				ptr, sizeof(*ptr)))
1761
		ret = -EFAULT;
D
Dan Carpenter 已提交
1762 1763

free_inherit:
A
Arne Jansen 已提交
1764
	kfree(inherit);
D
Dan Carpenter 已提交
1765 1766
free_args:
	kfree(vol_args);
C
Christoph Hellwig 已提交
1767 1768 1769
	return ret;
}

1770 1771 1772
static noinline int btrfs_ioctl_subvol_getflags(struct file *file,
						void __user *arg)
{
A
Al Viro 已提交
1773
	struct inode *inode = file_inode(file);
1774 1775 1776 1777
	struct btrfs_root *root = BTRFS_I(inode)->root;
	int ret = 0;
	u64 flags = 0;

L
Li Zefan 已提交
1778
	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID)
1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
		return -EINVAL;

	down_read(&root->fs_info->subvol_sem);
	if (btrfs_root_readonly(root))
		flags |= BTRFS_SUBVOL_RDONLY;
	up_read(&root->fs_info->subvol_sem);

	if (copy_to_user(arg, &flags, sizeof(flags)))
		ret = -EFAULT;

	return ret;
}

static noinline int btrfs_ioctl_subvol_setflags(struct file *file,
					      void __user *arg)
{
A
Al Viro 已提交
1795
	struct inode *inode = file_inode(file);
1796 1797 1798 1799 1800 1801
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
	u64 root_flags;
	u64 flags;
	int ret = 0;

1802 1803 1804
	if (!inode_owner_or_capable(inode))
		return -EPERM;

1805 1806 1807
	ret = mnt_want_write_file(file);
	if (ret)
		goto out;
1808

1809 1810 1811 1812
	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) {
		ret = -EINVAL;
		goto out_drop_write;
	}
1813

1814 1815 1816 1817
	if (copy_from_user(&flags, arg, sizeof(flags))) {
		ret = -EFAULT;
		goto out_drop_write;
	}
1818

1819 1820 1821 1822
	if (flags & BTRFS_SUBVOL_CREATE_ASYNC) {
		ret = -EINVAL;
		goto out_drop_write;
	}
1823

1824 1825 1826 1827
	if (flags & ~BTRFS_SUBVOL_RDONLY) {
		ret = -EOPNOTSUPP;
		goto out_drop_write;
	}
1828 1829 1830 1831 1832

	down_write(&root->fs_info->subvol_sem);

	/* nothing to do */
	if (!!(flags & BTRFS_SUBVOL_RDONLY) == btrfs_root_readonly(root))
1833
		goto out_drop_sem;
1834 1835

	root_flags = btrfs_root_flags(&root->root_item);
1836
	if (flags & BTRFS_SUBVOL_RDONLY) {
1837 1838
		btrfs_set_root_flags(&root->root_item,
				     root_flags | BTRFS_ROOT_SUBVOL_RDONLY);
1839 1840 1841 1842 1843 1844 1845 1846
	} else {
		/*
		 * Block RO -> RW transition if this subvolume is involved in
		 * send
		 */
		spin_lock(&root->root_item_lock);
		if (root->send_in_progress == 0) {
			btrfs_set_root_flags(&root->root_item,
1847
				     root_flags & ~BTRFS_ROOT_SUBVOL_RDONLY);
1848 1849 1850 1851 1852 1853 1854 1855 1856 1857
			spin_unlock(&root->root_item_lock);
		} else {
			spin_unlock(&root->root_item_lock);
			btrfs_warn(root->fs_info,
			"Attempt to set subvolume %llu read-write during send",
					root->root_key.objectid);
			ret = -EPERM;
			goto out_drop_sem;
		}
	}
1858 1859 1860 1861 1862 1863 1864

	trans = btrfs_start_transaction(root, 1);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out_reset;
	}

1865
	ret = btrfs_update_root(trans, root->fs_info->tree_root,
1866 1867 1868 1869 1870 1871
				&root->root_key, &root->root_item);

	btrfs_commit_transaction(trans, root);
out_reset:
	if (ret)
		btrfs_set_root_flags(&root->root_item, root_flags);
1872
out_drop_sem:
1873
	up_write(&root->fs_info->subvol_sem);
1874 1875 1876
out_drop_write:
	mnt_drop_write_file(file);
out:
1877 1878 1879
	return ret;
}

1880 1881 1882 1883 1884 1885
/*
 * helper to check if the subvolume references other subvolumes
 */
static noinline int may_destroy_subvol(struct btrfs_root *root)
{
	struct btrfs_path *path;
1886
	struct btrfs_dir_item *di;
1887
	struct btrfs_key key;
1888
	u64 dir_id;
1889 1890 1891 1892 1893 1894
	int ret;

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

1895 1896 1897 1898 1899 1900 1901
	/* Make sure this root isn't set as the default subvol */
	dir_id = btrfs_super_root_dir(root->fs_info->super_copy);
	di = btrfs_lookup_dir_item(NULL, root->fs_info->tree_root, path,
				   dir_id, "default", 7, 0);
	if (di && !IS_ERR(di)) {
		btrfs_dir_item_key_to_cpu(path->nodes[0], di, &key);
		if (key.objectid == root->root_key.objectid) {
1902
			ret = -EPERM;
J
Jeff Mahoney 已提交
1903 1904 1905
			btrfs_err(root->fs_info,
				  "deleting default subvolume %llu is not allowed",
				  key.objectid);
1906 1907 1908 1909 1910
			goto out;
		}
		btrfs_release_path(path);
	}

1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933
	key.objectid = root->root_key.objectid;
	key.type = BTRFS_ROOT_REF_KEY;
	key.offset = (u64)-1;

	ret = btrfs_search_slot(NULL, root->fs_info->tree_root,
				&key, path, 0, 0);
	if (ret < 0)
		goto out;
	BUG_ON(ret == 0);

	ret = 0;
	if (path->slots[0] > 0) {
		path->slots[0]--;
		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
		if (key.objectid == root->root_key.objectid &&
		    key.type == BTRFS_ROOT_REF_KEY)
			ret = -ENOTEMPTY;
	}
out:
	btrfs_free_path(path);
	return ret;
}

1934 1935 1936
static noinline int key_in_sk(struct btrfs_key *key,
			      struct btrfs_ioctl_search_key *sk)
{
1937 1938 1939 1940 1941 1942 1943 1944 1945
	struct btrfs_key test;
	int ret;

	test.objectid = sk->min_objectid;
	test.type = sk->min_type;
	test.offset = sk->min_offset;

	ret = btrfs_comp_cpu_keys(key, &test);
	if (ret < 0)
1946
		return 0;
1947 1948 1949 1950 1951 1952 1953

	test.objectid = sk->max_objectid;
	test.type = sk->max_type;
	test.offset = sk->max_offset;

	ret = btrfs_comp_cpu_keys(key, &test);
	if (ret > 0)
1954 1955 1956 1957
		return 0;
	return 1;
}

1958
static noinline int copy_to_sk(struct btrfs_path *path,
1959 1960
			       struct btrfs_key *key,
			       struct btrfs_ioctl_search_key *sk,
1961
			       size_t *buf_size,
1962
			       char __user *ubuf,
1963 1964 1965 1966 1967 1968
			       unsigned long *sk_offset,
			       int *num_found)
{
	u64 found_transid;
	struct extent_buffer *leaf;
	struct btrfs_ioctl_search_header sh;
1969
	struct btrfs_key test;
1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
	unsigned long item_off;
	unsigned long item_len;
	int nritems;
	int i;
	int slot;
	int ret = 0;

	leaf = path->nodes[0];
	slot = path->slots[0];
	nritems = btrfs_header_nritems(leaf);

	if (btrfs_header_generation(leaf) > sk->max_transid) {
		i = nritems;
		goto advance_key;
	}
	found_transid = btrfs_header_generation(leaf);

	for (i = slot; i < nritems; i++) {
		item_off = btrfs_item_ptr_offset(leaf, i);
		item_len = btrfs_item_size_nr(leaf, i);

1991 1992 1993 1994
		btrfs_item_key_to_cpu(leaf, key, i);
		if (!key_in_sk(key, sk))
			continue;

1995
		if (sizeof(sh) + item_len > *buf_size) {
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
			if (*num_found) {
				ret = 1;
				goto out;
			}

			/*
			 * return one empty item back for v1, which does not
			 * handle -EOVERFLOW
			 */

2006
			*buf_size = sizeof(sh) + item_len;
2007
			item_len = 0;
2008 2009
			ret = -EOVERFLOW;
		}
2010

2011
		if (sizeof(sh) + item_len + *sk_offset > *buf_size) {
2012
			ret = 1;
2013
			goto out;
2014 2015 2016 2017 2018 2019 2020 2021 2022
		}

		sh.objectid = key->objectid;
		sh.offset = key->offset;
		sh.type = key->type;
		sh.len = item_len;
		sh.transid = found_transid;

		/* copy search result header */
2023 2024 2025 2026 2027
		if (copy_to_user(ubuf + *sk_offset, &sh, sizeof(sh))) {
			ret = -EFAULT;
			goto out;
		}

2028 2029 2030
		*sk_offset += sizeof(sh);

		if (item_len) {
2031
			char __user *up = ubuf + *sk_offset;
2032
			/* copy the item */
2033 2034 2035 2036 2037 2038
			if (read_extent_buffer_to_user(leaf, up,
						       item_off, item_len)) {
				ret = -EFAULT;
				goto out;
			}

2039 2040
			*sk_offset += item_len;
		}
2041
		(*num_found)++;
2042

2043 2044 2045
		if (ret) /* -EOVERFLOW from above */
			goto out;

2046 2047 2048 2049
		if (*num_found >= sk->nr_items) {
			ret = 1;
			goto out;
		}
2050 2051
	}
advance_key:
2052
	ret = 0;
2053 2054 2055 2056 2057 2058
	test.objectid = sk->max_objectid;
	test.type = sk->max_type;
	test.offset = sk->max_offset;
	if (btrfs_comp_cpu_keys(key, &test) >= 0)
		ret = 1;
	else if (key->offset < (u64)-1)
2059
		key->offset++;
2060
	else if (key->type < (u8)-1) {
2061
		key->offset = 0;
2062
		key->type++;
2063
	} else if (key->objectid < (u64)-1) {
2064 2065
		key->offset = 0;
		key->type = 0;
2066
		key->objectid++;
2067 2068
	} else
		ret = 1;
2069
out:
2070 2071 2072 2073 2074 2075 2076 2077 2078
	/*
	 *  0: all items from this leaf copied, continue with next
	 *  1: * more items can be copied, but unused buffer is too small
	 *     * all items were found
	 *     Either way, it will stops the loop which iterates to the next
	 *     leaf
	 *  -EOVERFLOW: item was to large for buffer
	 *  -EFAULT: could not copy extent buffer back to userspace
	 */
2079 2080 2081 2082
	return ret;
}

static noinline int search_ioctl(struct inode *inode,
2083
				 struct btrfs_ioctl_search_key *sk,
2084
				 size_t *buf_size,
2085
				 char __user *ubuf)
2086 2087 2088 2089 2090 2091 2092 2093 2094
{
	struct btrfs_root *root;
	struct btrfs_key key;
	struct btrfs_path *path;
	struct btrfs_fs_info *info = BTRFS_I(inode)->root->fs_info;
	int ret;
	int num_found = 0;
	unsigned long sk_offset = 0;

2095 2096
	if (*buf_size < sizeof(struct btrfs_ioctl_search_header)) {
		*buf_size = sizeof(struct btrfs_ioctl_search_header);
2097
		return -EOVERFLOW;
2098
	}
2099

2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121
	path = btrfs_alloc_path();
	if (!path)
		return -ENOMEM;

	if (sk->tree_id == 0) {
		/* search the root of the inode that was passed */
		root = BTRFS_I(inode)->root;
	} else {
		key.objectid = sk->tree_id;
		key.type = BTRFS_ROOT_ITEM_KEY;
		key.offset = (u64)-1;
		root = btrfs_read_fs_root_no_name(info, &key);
		if (IS_ERR(root)) {
			btrfs_free_path(path);
			return -ENOENT;
		}
	}

	key.objectid = sk->min_objectid;
	key.type = sk->min_type;
	key.offset = sk->min_offset;

2122
	while (1) {
2123
		ret = btrfs_search_forward(root, &key, path, sk->min_transid);
2124 2125 2126 2127 2128
		if (ret != 0) {
			if (ret > 0)
				ret = 0;
			goto err;
		}
2129
		ret = copy_to_sk(path, &key, sk, buf_size, ubuf,
2130
				 &sk_offset, &num_found);
2131
		btrfs_release_path(path);
2132
		if (ret)
2133 2134 2135
			break;

	}
2136 2137
	if (ret > 0)
		ret = 0;
2138 2139 2140 2141 2142 2143 2144 2145 2146
err:
	sk->nr_items = num_found;
	btrfs_free_path(path);
	return ret;
}

static noinline int btrfs_ioctl_tree_search(struct file *file,
					   void __user *argp)
{
2147 2148
	struct btrfs_ioctl_search_args __user *uargs;
	struct btrfs_ioctl_search_key sk;
2149 2150 2151
	struct inode *inode;
	int ret;
	size_t buf_size;
2152 2153 2154 2155

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

2156 2157 2158 2159
	uargs = (struct btrfs_ioctl_search_args __user *)argp;

	if (copy_from_user(&sk, &uargs->key, sizeof(sk)))
		return -EFAULT;
2160

2161
	buf_size = sizeof(uargs->buf);
2162

A
Al Viro 已提交
2163
	inode = file_inode(file);
2164
	ret = search_ioctl(inode, &sk, &buf_size, uargs->buf);
2165 2166 2167 2168 2169 2170 2171 2172

	/*
	 * In the origin implementation an overflow is handled by returning a
	 * search header with a len of zero, so reset ret.
	 */
	if (ret == -EOVERFLOW)
		ret = 0;

2173
	if (ret == 0 && copy_to_user(&uargs->key, &sk, sizeof(sk)))
2174 2175 2176 2177
		ret = -EFAULT;
	return ret;
}

G
Gerhard Heift 已提交
2178 2179 2180 2181 2182 2183 2184 2185
static noinline int btrfs_ioctl_tree_search_v2(struct file *file,
					       void __user *argp)
{
	struct btrfs_ioctl_search_args_v2 __user *uarg;
	struct btrfs_ioctl_search_args_v2 args;
	struct inode *inode;
	int ret;
	size_t buf_size;
2186
	const size_t buf_limit = SZ_16M;
G
Gerhard Heift 已提交
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

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	/* copy search header and buffer size */
	uarg = (struct btrfs_ioctl_search_args_v2 __user *)argp;
	if (copy_from_user(&args, uarg, sizeof(args)))
		return -EFAULT;

	buf_size = args.buf_size;

	if (buf_size < sizeof(struct btrfs_ioctl_search_header))
		return -EOVERFLOW;

	/* limit result size to 16MB */
	if (buf_size > buf_limit)
		buf_size = buf_limit;

	inode = file_inode(file);
	ret = search_ioctl(inode, &args.key, &buf_size,
			   (char *)(&uarg->buf[0]));
	if (ret == 0 && copy_to_user(&uarg->key, &args.key, sizeof(args.key)))
		ret = -EFAULT;
	else if (ret == -EOVERFLOW &&
		copy_to_user(&uarg->buf_size, &buf_size, sizeof(buf_size)))
		ret = -EFAULT;

2214 2215 2216
	return ret;
}

2217
/*
2218 2219 2220
 * Search INODE_REFs to identify path name of 'dirid' directory
 * in a 'tree_id' tree. and sets path name to 'name'.
 */
2221 2222 2223 2224 2225
static noinline int btrfs_search_path_in_tree(struct btrfs_fs_info *info,
				u64 tree_id, u64 dirid, char *name)
{
	struct btrfs_root *root;
	struct btrfs_key key;
2226
	char *ptr;
2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243
	int ret = -1;
	int slot;
	int len;
	int total_len = 0;
	struct btrfs_inode_ref *iref;
	struct extent_buffer *l;
	struct btrfs_path *path;

	if (dirid == BTRFS_FIRST_FREE_OBJECTID) {
		name[0]='\0';
		return 0;
	}

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

2244
	ptr = &name[BTRFS_INO_LOOKUP_PATH_MAX];
2245 2246 2247 2248 2249 2250

	key.objectid = tree_id;
	key.type = BTRFS_ROOT_ITEM_KEY;
	key.offset = (u64)-1;
	root = btrfs_read_fs_root_no_name(info, &key);
	if (IS_ERR(root)) {
2251
		btrfs_err(info, "could not find root %llu", tree_id);
2252 2253
		ret = -ENOENT;
		goto out;
2254 2255 2256 2257
	}

	key.objectid = dirid;
	key.type = BTRFS_INODE_REF_KEY;
2258
	key.offset = (u64)-1;
2259

2260
	while (1) {
2261 2262 2263
		ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
		if (ret < 0)
			goto out;
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273
		else if (ret > 0) {
			ret = btrfs_previous_item(root, path, dirid,
						  BTRFS_INODE_REF_KEY);
			if (ret < 0)
				goto out;
			else if (ret > 0) {
				ret = -ENOENT;
				goto out;
			}
		}
2274 2275 2276 2277 2278 2279 2280 2281 2282

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

		iref = btrfs_item_ptr(l, slot, struct btrfs_inode_ref);
		len = btrfs_inode_ref_name_len(l, iref);
		ptr -= len + 1;
		total_len += len + 1;
2283 2284
		if (ptr < name) {
			ret = -ENAMETOOLONG;
2285
			goto out;
2286
		}
2287 2288

		*(ptr + len) = '/';
2289
		read_extent_buffer(l, ptr, (unsigned long)(iref + 1), len);
2290 2291 2292 2293

		if (key.offset == BTRFS_FIRST_FREE_OBJECTID)
			break;

2294
		btrfs_release_path(path);
2295
		key.objectid = key.offset;
2296
		key.offset = (u64)-1;
2297 2298
		dirid = key.objectid;
	}
2299
	memmove(name, ptr, total_len);
2300
	name[total_len] = '\0';
2301 2302 2303
	ret = 0;
out:
	btrfs_free_path(path);
2304 2305 2306 2307 2308 2309 2310 2311
	return ret;
}

static noinline int btrfs_ioctl_ino_lookup(struct file *file,
					   void __user *argp)
{
	 struct btrfs_ioctl_ino_lookup_args *args;
	 struct inode *inode;
2312
	int ret = 0;
2313

J
Julia Lawall 已提交
2314 2315 2316
	args = memdup_user(argp, sizeof(*args));
	if (IS_ERR(args))
		return PTR_ERR(args);
2317

A
Al Viro 已提交
2318
	inode = file_inode(file);
2319

2320 2321 2322 2323
	/*
	 * Unprivileged query to obtain the containing subvolume root id. The
	 * path is reset so it's consistent with btrfs_search_path_in_tree.
	 */
2324 2325 2326
	if (args->treeid == 0)
		args->treeid = BTRFS_I(inode)->root->root_key.objectid;

2327 2328 2329 2330 2331 2332 2333 2334 2335 2336
	if (args->objectid == BTRFS_FIRST_FREE_OBJECTID) {
		args->name[0] = 0;
		goto out;
	}

	if (!capable(CAP_SYS_ADMIN)) {
		ret = -EPERM;
		goto out;
	}

2337 2338 2339 2340
	ret = btrfs_search_path_in_tree(BTRFS_I(inode)->root->fs_info,
					args->treeid, args->objectid,
					args->name);

2341
out:
2342 2343 2344 2345
	if (ret == 0 && copy_to_user(argp, args, sizeof(*args)))
		ret = -EFAULT;

	kfree(args);
2346 2347 2348
	return ret;
}

2349 2350 2351
static noinline int btrfs_ioctl_snap_destroy(struct file *file,
					     void __user *arg)
{
A
Al Viro 已提交
2352
	struct dentry *parent = file->f_path.dentry;
2353
	struct dentry *dentry;
2354
	struct inode *dir = d_inode(parent);
2355 2356 2357 2358 2359
	struct inode *inode;
	struct btrfs_root *root = BTRFS_I(dir)->root;
	struct btrfs_root *dest = NULL;
	struct btrfs_ioctl_vol_args *vol_args;
	struct btrfs_trans_handle *trans;
2360
	struct btrfs_block_rsv block_rsv;
2361
	u64 root_flags;
2362
	u64 qgroup_reserved;
2363 2364 2365 2366
	int namelen;
	int ret;
	int err = 0;

2367 2368 2369
	if (!S_ISDIR(dir->i_mode))
		return -ENOTDIR;

2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381
	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args))
		return PTR_ERR(vol_args);

	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
	namelen = strlen(vol_args->name);
	if (strchr(vol_args->name, '/') ||
	    strncmp(vol_args->name, "..", namelen) == 0) {
		err = -EINVAL;
		goto out;
	}

2382
	err = mnt_want_write_file(file);
2383 2384 2385
	if (err)
		goto out;

2386

2387 2388 2389
	err = down_write_killable_nested(&dir->i_rwsem, I_MUTEX_PARENT);
	if (err == -EINTR)
		goto out_drop_write;
2390 2391 2392 2393 2394 2395
	dentry = lookup_one_len(vol_args->name, parent, namelen);
	if (IS_ERR(dentry)) {
		err = PTR_ERR(dentry);
		goto out_unlock_dir;
	}

2396
	if (d_really_is_negative(dentry)) {
2397 2398 2399 2400
		err = -ENOENT;
		goto out_dput;
	}

2401
	inode = d_inode(dentry);
2402
	dest = BTRFS_I(inode)->root;
2403
	if (!capable(CAP_SYS_ADMIN)) {
2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417
		/*
		 * Regular user.  Only allow this with a special mount
		 * option, when the user has write+exec access to the
		 * subvol root, and when rmdir(2) would have been
		 * allowed.
		 *
		 * Note that this is _not_ check that the subvol is
		 * empty or doesn't contain data that we wouldn't
		 * otherwise be able to delete.
		 *
		 * Users who want to delete empty subvols should try
		 * rmdir(2).
		 */
		err = -EPERM;
2418
		if (!btrfs_test_opt(root->fs_info, USER_SUBVOL_RM_ALLOWED))
2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436
			goto out_dput;

		/*
		 * Do not allow deletion if the parent dir is the same
		 * as the dir to be deleted.  That means the ioctl
		 * must be called on the dentry referencing the root
		 * of the subvol, not a random directory contained
		 * within it.
		 */
		err = -EINVAL;
		if (root == dest)
			goto out_dput;

		err = inode_permission(inode, MAY_WRITE | MAY_EXEC);
		if (err)
			goto out_dput;
	}

2437 2438 2439 2440 2441
	/* check if subvolume may be deleted by a user */
	err = btrfs_may_delete(dir, dentry, 1);
	if (err)
		goto out_dput;

L
Li Zefan 已提交
2442
	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) {
2443 2444 2445 2446
		err = -EINVAL;
		goto out_dput;
	}

A
Al Viro 已提交
2447
	inode_lock(inode);
2448 2449 2450 2451 2452 2453 2454

	/*
	 * Don't allow to delete a subvolume with send in progress. This is
	 * inside the i_mutex so the error handling that has to drop the bit
	 * again is not run concurrently.
	 */
	spin_lock(&dest->root_item_lock);
2455 2456 2457
	root_flags = btrfs_root_flags(&dest->root_item);
	if (dest->send_in_progress == 0) {
		btrfs_set_root_flags(&dest->root_item,
2458 2459 2460 2461 2462 2463
				root_flags | BTRFS_ROOT_SUBVOL_DEAD);
		spin_unlock(&dest->root_item_lock);
	} else {
		spin_unlock(&dest->root_item_lock);
		btrfs_warn(root->fs_info,
			"Attempt to delete subvolume %llu during send",
2464
			dest->root_key.objectid);
2465
		err = -EPERM;
2466
		goto out_unlock_inode;
2467 2468
	}

2469 2470 2471 2472 2473 2474
	down_write(&root->fs_info->subvol_sem);

	err = may_destroy_subvol(dest);
	if (err)
		goto out_up_write;

2475 2476 2477 2478 2479 2480
	btrfs_init_block_rsv(&block_rsv, BTRFS_BLOCK_RSV_TEMP);
	/*
	 * One for dir inode, two for dir entries, two for root
	 * ref/backref.
	 */
	err = btrfs_subvolume_reserve_metadata(root, &block_rsv,
2481
					       5, &qgroup_reserved, true);
2482 2483 2484
	if (err)
		goto out_up_write;

2485 2486 2487
	trans = btrfs_start_transaction(root, 0);
	if (IS_ERR(trans)) {
		err = PTR_ERR(trans);
2488
		goto out_release;
2489
	}
2490 2491
	trans->block_rsv = &block_rsv;
	trans->bytes_reserved = block_rsv.size;
2492

2493 2494
	btrfs_record_snapshot_destroy(trans, dir);

2495 2496 2497 2498
	ret = btrfs_unlink_subvol(trans, root, dir,
				dest->root_key.objectid,
				dentry->d_name.name,
				dentry->d_name.len);
2499 2500
	if (ret) {
		err = ret;
2501
		btrfs_abort_transaction(trans, ret);
2502 2503
		goto out_end_trans;
	}
2504 2505 2506 2507 2508 2509 2510 2511

	btrfs_record_root_in_trans(trans, dest);

	memset(&dest->root_item.drop_progress, 0,
		sizeof(dest->root_item.drop_progress));
	dest->root_item.drop_level = 0;
	btrfs_set_root_refs(&dest->root_item, 0);

2512
	if (!test_and_set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &dest->state)) {
2513 2514 2515
		ret = btrfs_insert_orphan_item(trans,
					root->fs_info->tree_root,
					dest->root_key.objectid);
2516
		if (ret) {
2517
			btrfs_abort_transaction(trans, ret);
2518 2519 2520
			err = ret;
			goto out_end_trans;
		}
2521
	}
2522

2523 2524
	ret = btrfs_uuid_tree_rem(trans, root->fs_info, dest->root_item.uuid,
				  BTRFS_UUID_KEY_SUBVOL,
2525 2526
				  dest->root_key.objectid);
	if (ret && ret != -ENOENT) {
2527
		btrfs_abort_transaction(trans, ret);
2528 2529 2530 2531
		err = ret;
		goto out_end_trans;
	}
	if (!btrfs_is_empty_uuid(dest->root_item.received_uuid)) {
2532
		ret = btrfs_uuid_tree_rem(trans, root->fs_info,
2533 2534 2535 2536
					  dest->root_item.received_uuid,
					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
					  dest->root_key.objectid);
		if (ret && ret != -ENOENT) {
2537
			btrfs_abort_transaction(trans, ret);
2538 2539 2540 2541 2542
			err = ret;
			goto out_end_trans;
		}
	}

2543
out_end_trans:
2544 2545
	trans->block_rsv = NULL;
	trans->bytes_reserved = 0;
S
Sage Weil 已提交
2546
	ret = btrfs_end_transaction(trans, root);
2547 2548
	if (ret && !err)
		err = ret;
2549
	inode->i_flags |= S_DEAD;
2550 2551
out_release:
	btrfs_subvolume_release_metadata(root, &block_rsv, qgroup_reserved);
2552 2553
out_up_write:
	up_write(&root->fs_info->subvol_sem);
2554 2555
	if (err) {
		spin_lock(&dest->root_item_lock);
2556 2557
		root_flags = btrfs_root_flags(&dest->root_item);
		btrfs_set_root_flags(&dest->root_item,
2558 2559 2560
				root_flags & ~BTRFS_ROOT_SUBVOL_DEAD);
		spin_unlock(&dest->root_item_lock);
	}
2561
out_unlock_inode:
A
Al Viro 已提交
2562
	inode_unlock(inode);
2563
	if (!err) {
2564
		d_invalidate(dentry);
2565 2566
		btrfs_invalidate_inodes(dest);
		d_delete(dentry);
2567
		ASSERT(dest->send_in_progress == 0);
2568 2569

		/* the last ref */
2570 2571 2572
		if (dest->ino_cache_inode) {
			iput(dest->ino_cache_inode);
			dest->ino_cache_inode = NULL;
2573
		}
2574 2575 2576 2577
	}
out_dput:
	dput(dentry);
out_unlock_dir:
A
Al Viro 已提交
2578
	inode_unlock(dir);
2579
out_drop_write:
A
Al Viro 已提交
2580
	mnt_drop_write_file(file);
2581 2582 2583 2584 2585
out:
	kfree(vol_args);
	return err;
}

C
Chris Mason 已提交
2586
static int btrfs_ioctl_defrag(struct file *file, void __user *argp)
C
Christoph Hellwig 已提交
2587
{
A
Al Viro 已提交
2588
	struct inode *inode = file_inode(file);
C
Christoph Hellwig 已提交
2589
	struct btrfs_root *root = BTRFS_I(inode)->root;
C
Chris Mason 已提交
2590
	struct btrfs_ioctl_defrag_range_args *range;
Y
Yan Zheng 已提交
2591 2592
	int ret;

2593 2594 2595
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
L
Li Zefan 已提交
2596

2597 2598 2599
	if (btrfs_root_readonly(root)) {
		ret = -EROFS;
		goto out;
2600
	}
C
Christoph Hellwig 已提交
2601 2602 2603

	switch (inode->i_mode & S_IFMT) {
	case S_IFDIR:
2604 2605 2606 2607
		if (!capable(CAP_SYS_ADMIN)) {
			ret = -EPERM;
			goto out;
		}
2608
		ret = btrfs_defrag_root(root);
2609 2610
		if (ret)
			goto out;
2611
		ret = btrfs_defrag_root(root->fs_info->extent_root);
C
Christoph Hellwig 已提交
2612 2613
		break;
	case S_IFREG:
2614 2615 2616 2617
		if (!(file->f_mode & FMODE_WRITE)) {
			ret = -EINVAL;
			goto out;
		}
C
Chris Mason 已提交
2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629

		range = kzalloc(sizeof(*range), GFP_KERNEL);
		if (!range) {
			ret = -ENOMEM;
			goto out;
		}

		if (argp) {
			if (copy_from_user(range, argp,
					   sizeof(*range))) {
				ret = -EFAULT;
				kfree(range);
2630
				goto out;
C
Chris Mason 已提交
2631 2632 2633 2634 2635 2636 2637 2638 2639 2640
			}
			/* compression requires us to start the IO */
			if ((range->flags & BTRFS_DEFRAG_RANGE_COMPRESS)) {
				range->flags |= BTRFS_DEFRAG_RANGE_START_IO;
				range->extent_thresh = (u32)-1;
			}
		} else {
			/* the rest are all set to zero by kzalloc */
			range->len = (u64)-1;
		}
A
Al Viro 已提交
2641
		ret = btrfs_defrag_file(file_inode(file), file,
C
Chris Mason 已提交
2642 2643 2644
					range, 0, 0);
		if (ret > 0)
			ret = 0;
C
Chris Mason 已提交
2645
		kfree(range);
C
Christoph Hellwig 已提交
2646
		break;
2647 2648
	default:
		ret = -EINVAL;
C
Christoph Hellwig 已提交
2649
	}
2650
out:
2651
	mnt_drop_write_file(file);
2652
	return ret;
C
Christoph Hellwig 已提交
2653 2654
}

2655
static long btrfs_ioctl_add_dev(struct btrfs_root *root, void __user *arg)
C
Christoph Hellwig 已提交
2656 2657 2658 2659
{
	struct btrfs_ioctl_vol_args *vol_args;
	int ret;

2660 2661 2662
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

2663 2664
	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
			1)) {
2665
		return BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
2666 2667
	}

2668
	mutex_lock(&root->fs_info->volume_mutex);
L
Li Zefan 已提交
2669
	vol_args = memdup_user(arg, sizeof(*vol_args));
2670 2671 2672 2673
	if (IS_ERR(vol_args)) {
		ret = PTR_ERR(vol_args);
		goto out;
	}
C
Christoph Hellwig 已提交
2674

2675
	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
2676
	ret = btrfs_init_new_device(root->fs_info, vol_args->name);
C
Christoph Hellwig 已提交
2677

A
Anand Jain 已提交
2678 2679 2680
	if (!ret)
		btrfs_info(root->fs_info, "disk added %s",vol_args->name);

C
Christoph Hellwig 已提交
2681
	kfree(vol_args);
2682 2683
out:
	mutex_unlock(&root->fs_info->volume_mutex);
2684
	atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0);
C
Christoph Hellwig 已提交
2685 2686 2687
	return ret;
}

2688
static long btrfs_ioctl_rm_dev_v2(struct file *file, void __user *arg)
C
Christoph Hellwig 已提交
2689
{
A
Al Viro 已提交
2690
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
2691
	struct btrfs_ioctl_vol_args_v2 *vol_args;
C
Christoph Hellwig 已提交
2692 2693
	int ret;

2694 2695 2696
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

2697 2698 2699
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
Y
Yan Zheng 已提交
2700

L
Li Zefan 已提交
2701
	vol_args = memdup_user(arg, sizeof(*vol_args));
2702 2703
	if (IS_ERR(vol_args)) {
		ret = PTR_ERR(vol_args);
D
Dan Carpenter 已提交
2704
		goto err_drop;
2705
	}
C
Christoph Hellwig 已提交
2706

2707
	/* Check for compatibility reject unknown flags */
2708 2709
	if (vol_args->flags & ~BTRFS_VOL_ARG_V2_FLAGS_SUPPORTED)
		return -EOPNOTSUPP;
C
Christoph Hellwig 已提交
2710

2711 2712 2713 2714 2715 2716 2717
	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
			1)) {
		ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
		goto out;
	}

	mutex_lock(&root->fs_info->volume_mutex);
2718
	if (vol_args->flags & BTRFS_DEVICE_SPEC_BY_ID) {
2719 2720 2721 2722 2723
		ret = btrfs_rm_device(root, NULL, vol_args->devid);
	} else {
		vol_args->name[BTRFS_SUBVOL_NAME_MAX] = '\0';
		ret = btrfs_rm_device(root, vol_args->name, 0);
	}
2724
	mutex_unlock(&root->fs_info->volume_mutex);
2725
	atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0);
2726

2727
	if (!ret) {
2728
		if (vol_args->flags & BTRFS_DEVICE_SPEC_BY_ID)
2729 2730 2731 2732 2733 2734
			btrfs_info(root->fs_info, "device deleted: id %llu",
					vol_args->devid);
		else
			btrfs_info(root->fs_info, "device deleted: %s",
					vol_args->name);
	}
2735 2736
out:
	kfree(vol_args);
D
Dan Carpenter 已提交
2737
err_drop:
2738
	mnt_drop_write_file(file);
C
Christoph Hellwig 已提交
2739 2740 2741
	return ret;
}

2742
static long btrfs_ioctl_rm_dev(struct file *file, void __user *arg)
C
Christoph Hellwig 已提交
2743
{
A
Al Viro 已提交
2744
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
C
Christoph Hellwig 已提交
2745 2746 2747
	struct btrfs_ioctl_vol_args *vol_args;
	int ret;

2748 2749 2750
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

2751 2752 2753
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
Y
Yan Zheng 已提交
2754

2755 2756 2757
	if (atomic_xchg(&root->fs_info->mutually_exclusive_operation_running,
			1)) {
		ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
2758 2759 2760 2761 2762 2763
		goto out_drop_write;
	}

	vol_args = memdup_user(arg, sizeof(*vol_args));
	if (IS_ERR(vol_args)) {
		ret = PTR_ERR(vol_args);
2764 2765 2766
		goto out;
	}

2767
	vol_args->name[BTRFS_PATH_NAME_MAX] = '\0';
2768
	mutex_lock(&root->fs_info->volume_mutex);
2769
	ret = btrfs_rm_device(root, vol_args->name, 0);
2770
	mutex_unlock(&root->fs_info->volume_mutex);
2771

2772 2773
	if (!ret)
		btrfs_info(root->fs_info, "disk deleted %s",vol_args->name);
2774
	kfree(vol_args);
2775 2776 2777
out:
	atomic_set(&root->fs_info->mutually_exclusive_operation_running, 0);
out_drop_write:
2778
	mnt_drop_write_file(file);
2779

C
Christoph Hellwig 已提交
2780 2781 2782
	return ret;
}

J
Jan Schmidt 已提交
2783 2784
static long btrfs_ioctl_fs_info(struct btrfs_root *root, void __user *arg)
{
2785
	struct btrfs_ioctl_fs_info_args *fi_args;
J
Jan Schmidt 已提交
2786 2787
	struct btrfs_device *device;
	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
2788
	int ret = 0;
J
Jan Schmidt 已提交
2789

2790 2791 2792 2793
	fi_args = kzalloc(sizeof(*fi_args), GFP_KERNEL);
	if (!fi_args)
		return -ENOMEM;

2794
	mutex_lock(&fs_devices->device_list_mutex);
2795 2796
	fi_args->num_devices = fs_devices->num_devices;
	memcpy(&fi_args->fsid, root->fs_info->fsid, sizeof(fi_args->fsid));
J
Jan Schmidt 已提交
2797

2798
	list_for_each_entry(device, &fs_devices->devices, dev_list) {
2799 2800
		if (device->devid > fi_args->max_id)
			fi_args->max_id = device->devid;
J
Jan Schmidt 已提交
2801 2802 2803
	}
	mutex_unlock(&fs_devices->device_list_mutex);

2804 2805 2806 2807
	fi_args->nodesize = root->fs_info->super_copy->nodesize;
	fi_args->sectorsize = root->fs_info->super_copy->sectorsize;
	fi_args->clone_alignment = root->fs_info->super_copy->sectorsize;

2808 2809
	if (copy_to_user(arg, fi_args, sizeof(*fi_args)))
		ret = -EFAULT;
J
Jan Schmidt 已提交
2810

2811 2812
	kfree(fi_args);
	return ret;
J
Jan Schmidt 已提交
2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826
}

static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg)
{
	struct btrfs_ioctl_dev_info_args *di_args;
	struct btrfs_device *dev;
	struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
	int ret = 0;
	char *s_uuid = NULL;

	di_args = memdup_user(arg, sizeof(*di_args));
	if (IS_ERR(di_args))
		return PTR_ERR(di_args);

2827
	if (!btrfs_is_empty_uuid(di_args->uuid))
J
Jan Schmidt 已提交
2828 2829 2830
		s_uuid = di_args->uuid;

	mutex_lock(&fs_devices->device_list_mutex);
2831
	dev = btrfs_find_device(root->fs_info, di_args->devid, s_uuid, NULL);
J
Jan Schmidt 已提交
2832 2833 2834 2835 2836 2837 2838

	if (!dev) {
		ret = -ENODEV;
		goto out;
	}

	di_args->devid = dev->devid;
2839 2840
	di_args->bytes_used = btrfs_device_get_bytes_used(dev);
	di_args->total_bytes = btrfs_device_get_total_bytes(dev);
J
Jan Schmidt 已提交
2841
	memcpy(di_args->uuid, dev->uuid, sizeof(di_args->uuid));
2842
	if (dev->name) {
2843 2844 2845 2846 2847 2848
		struct rcu_string *name;

		rcu_read_lock();
		name = rcu_dereference(dev->name);
		strncpy(di_args->path, name->str, sizeof(di_args->path));
		rcu_read_unlock();
2849 2850
		di_args->path[sizeof(di_args->path) - 1] = 0;
	} else {
2851
		di_args->path[0] = '\0';
2852
	}
J
Jan Schmidt 已提交
2853 2854

out:
2855
	mutex_unlock(&fs_devices->device_list_mutex);
J
Jan Schmidt 已提交
2856 2857 2858 2859 2860 2861 2862
	if (ret == 0 && copy_to_user(arg, di_args, sizeof(*di_args)))
		ret = -EFAULT;

	kfree(di_args);
	return ret;
}

2863
static struct page *extent_same_get_page(struct inode *inode, pgoff_t index)
M
Mark Fasheh 已提交
2864 2865 2866 2867 2868
{
	struct page *page;

	page = grab_cache_page(inode->i_mapping, index);
	if (!page)
2869
		return ERR_PTR(-ENOMEM);
M
Mark Fasheh 已提交
2870 2871

	if (!PageUptodate(page)) {
2872 2873 2874 2875 2876
		int ret;

		ret = btrfs_readpage(NULL, page);
		if (ret)
			return ERR_PTR(ret);
M
Mark Fasheh 已提交
2877 2878 2879
		lock_page(page);
		if (!PageUptodate(page)) {
			unlock_page(page);
2880
			put_page(page);
2881 2882 2883 2884
			return ERR_PTR(-EIO);
		}
		if (page->mapping != inode->i_mapping) {
			unlock_page(page);
2885
			put_page(page);
2886
			return ERR_PTR(-EAGAIN);
M
Mark Fasheh 已提交
2887 2888 2889 2890 2891 2892
		}
	}

	return page;
}

2893 2894 2895 2896
static int gather_extent_pages(struct inode *inode, struct page **pages,
			       int num_pages, u64 off)
{
	int i;
2897
	pgoff_t index = off >> PAGE_SHIFT;
2898 2899

	for (i = 0; i < num_pages; i++) {
2900
again:
2901
		pages[i] = extent_same_get_page(inode, index + i);
2902 2903 2904 2905 2906 2907 2908 2909
		if (IS_ERR(pages[i])) {
			int err = PTR_ERR(pages[i]);

			if (err == -EAGAIN)
				goto again;
			pages[i] = NULL;
			return err;
		}
2910 2911 2912 2913
	}
	return 0;
}

2914 2915
static int lock_extent_range(struct inode *inode, u64 off, u64 len,
			     bool retry_range_locking)
2916
{
2917 2918 2919 2920 2921 2922 2923 2924
	/*
	 * Do any pending delalloc/csum calculations on inode, one way or
	 * another, and lock file content.
	 * The locking order is:
	 *
	 *   1) pages
	 *   2) range in the inode's io tree
	 */
2925 2926 2927 2928 2929
	while (1) {
		struct btrfs_ordered_extent *ordered;
		lock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
		ordered = btrfs_lookup_first_ordered_extent(inode,
							    off + len - 1);
2930 2931 2932
		if ((!ordered ||
		     ordered->file_offset + ordered->len <= off ||
		     ordered->file_offset >= off + len) &&
2933
		    !test_range_bit(&BTRFS_I(inode)->io_tree, off,
2934 2935 2936
				    off + len - 1, EXTENT_DELALLOC, 0, NULL)) {
			if (ordered)
				btrfs_put_ordered_extent(ordered);
2937
			break;
2938
		}
2939 2940 2941
		unlock_extent(&BTRFS_I(inode)->io_tree, off, off + len - 1);
		if (ordered)
			btrfs_put_ordered_extent(ordered);
2942 2943
		if (!retry_range_locking)
			return -EAGAIN;
2944 2945
		btrfs_wait_ordered_range(inode, off, len);
	}
2946
	return 0;
2947 2948
}

2949
static void btrfs_double_inode_unlock(struct inode *inode1, struct inode *inode2)
M
Mark Fasheh 已提交
2950
{
A
Al Viro 已提交
2951 2952
	inode_unlock(inode1);
	inode_unlock(inode2);
M
Mark Fasheh 已提交
2953 2954
}

2955 2956 2957 2958 2959
static void btrfs_double_inode_lock(struct inode *inode1, struct inode *inode2)
{
	if (inode1 < inode2)
		swap(inode1, inode2);

A
Al Viro 已提交
2960 2961
	inode_lock_nested(inode1, I_MUTEX_PARENT);
	inode_lock_nested(inode2, I_MUTEX_CHILD);
2962 2963 2964 2965 2966 2967 2968 2969 2970
}

static void btrfs_double_extent_unlock(struct inode *inode1, u64 loff1,
				      struct inode *inode2, u64 loff2, u64 len)
{
	unlock_extent(&BTRFS_I(inode1)->io_tree, loff1, loff1 + len - 1);
	unlock_extent(&BTRFS_I(inode2)->io_tree, loff2, loff2 + len - 1);
}

2971 2972 2973
static int btrfs_double_extent_lock(struct inode *inode1, u64 loff1,
				    struct inode *inode2, u64 loff2, u64 len,
				    bool retry_range_locking)
M
Mark Fasheh 已提交
2974
{
2975 2976
	int ret;

M
Mark Fasheh 已提交
2977 2978 2979 2980
	if (inode1 < inode2) {
		swap(inode1, inode2);
		swap(loff1, loff2);
	}
2981 2982 2983 2984 2985 2986 2987 2988
	ret = lock_extent_range(inode1, loff1, len, retry_range_locking);
	if (ret)
		return ret;
	ret = lock_extent_range(inode2, loff2, len, retry_range_locking);
	if (ret)
		unlock_extent(&BTRFS_I(inode1)->io_tree, loff1,
			      loff1 + len - 1);
	return ret;
2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003
}

struct cmp_pages {
	int		num_pages;
	struct page	**src_pages;
	struct page	**dst_pages;
};

static void btrfs_cmp_data_free(struct cmp_pages *cmp)
{
	int i;
	struct page *pg;

	for (i = 0; i < cmp->num_pages; i++) {
		pg = cmp->src_pages[i];
3004 3005
		if (pg) {
			unlock_page(pg);
3006
			put_page(pg);
3007
		}
3008
		pg = cmp->dst_pages[i];
3009 3010
		if (pg) {
			unlock_page(pg);
3011
			put_page(pg);
3012
		}
3013 3014 3015 3016 3017 3018 3019 3020 3021 3022
	}
	kfree(cmp->src_pages);
	kfree(cmp->dst_pages);
}

static int btrfs_cmp_data_prepare(struct inode *src, u64 loff,
				  struct inode *dst, u64 dst_loff,
				  u64 len, struct cmp_pages *cmp)
{
	int ret;
3023
	int num_pages = PAGE_ALIGN(len) >> PAGE_SHIFT;
3024 3025 3026 3027 3028 3029 3030 3031
	struct page **src_pgarr, **dst_pgarr;

	/*
	 * We must gather up all the pages before we initiate our
	 * extent locking. We use an array for the page pointers. Size
	 * of the array is bounded by len, which is in turn bounded by
	 * BTRFS_MAX_DEDUPE_LEN.
	 */
3032 3033
	src_pgarr = kcalloc(num_pages, sizeof(struct page *), GFP_KERNEL);
	dst_pgarr = kcalloc(num_pages, sizeof(struct page *), GFP_KERNEL);
3034 3035 3036 3037
	if (!src_pgarr || !dst_pgarr) {
		kfree(src_pgarr);
		kfree(dst_pgarr);
		return -ENOMEM;
M
Mark Fasheh 已提交
3038
	}
3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052
	cmp->num_pages = num_pages;
	cmp->src_pages = src_pgarr;
	cmp->dst_pages = dst_pgarr;

	ret = gather_extent_pages(src, cmp->src_pages, cmp->num_pages, loff);
	if (ret)
		goto out;

	ret = gather_extent_pages(dst, cmp->dst_pages, cmp->num_pages, dst_loff);

out:
	if (ret)
		btrfs_cmp_data_free(cmp);
	return 0;
M
Mark Fasheh 已提交
3053 3054 3055
}

static int btrfs_cmp_data(struct inode *src, u64 loff, struct inode *dst,
3056
			  u64 dst_loff, u64 len, struct cmp_pages *cmp)
M
Mark Fasheh 已提交
3057 3058
{
	int ret = 0;
3059
	int i;
M
Mark Fasheh 已提交
3060
	struct page *src_page, *dst_page;
3061
	unsigned int cmp_len = PAGE_SIZE;
M
Mark Fasheh 已提交
3062 3063
	void *addr, *dst_addr;

3064
	i = 0;
M
Mark Fasheh 已提交
3065
	while (len) {
3066
		if (len < PAGE_SIZE)
M
Mark Fasheh 已提交
3067 3068
			cmp_len = len;

3069 3070 3071 3072
		BUG_ON(i >= cmp->num_pages);

		src_page = cmp->src_pages[i];
		dst_page = cmp->dst_pages[i];
3073 3074
		ASSERT(PageLocked(src_page));
		ASSERT(PageLocked(dst_page));
3075

M
Mark Fasheh 已提交
3076 3077 3078 3079 3080 3081 3082
		addr = kmap_atomic(src_page);
		dst_addr = kmap_atomic(dst_page);

		flush_dcache_page(src_page);
		flush_dcache_page(dst_page);

		if (memcmp(addr, dst_addr, cmp_len))
3083
			ret = -EBADE;
M
Mark Fasheh 已提交
3084 3085 3086 3087 3088 3089 3090 3091

		kunmap_atomic(addr);
		kunmap_atomic(dst_addr);

		if (ret)
			break;

		len -= cmp_len;
3092
		i++;
M
Mark Fasheh 已提交
3093 3094 3095 3096 3097
	}

	return ret;
}

3098 3099
static int extent_same_check_offsets(struct inode *inode, u64 off, u64 *plen,
				     u64 olen)
M
Mark Fasheh 已提交
3100
{
3101
	u64 len = *plen;
M
Mark Fasheh 已提交
3102 3103
	u64 bs = BTRFS_I(inode)->root->fs_info->sb->s_blocksize;

3104
	if (off + olen > inode->i_size || off + olen < off)
M
Mark Fasheh 已提交
3105
		return -EINVAL;
3106 3107 3108 3109 3110

	/* if we extend to eof, continue to block boundary */
	if (off + len == inode->i_size)
		*plen = len = ALIGN(inode->i_size, bs) - off;

M
Mark Fasheh 已提交
3111 3112 3113 3114 3115 3116 3117
	/* Check that we are block aligned - btrfs_clone() requires this */
	if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs))
		return -EINVAL;

	return 0;
}

3118
static int btrfs_extent_same(struct inode *src, u64 loff, u64 olen,
M
Mark Fasheh 已提交
3119 3120 3121
			     struct inode *dst, u64 dst_loff)
{
	int ret;
3122
	u64 len = olen;
3123
	struct cmp_pages cmp;
M
Mark Fasheh 已提交
3124 3125 3126
	int same_inode = 0;
	u64 same_lock_start = 0;
	u64 same_lock_len = 0;
M
Mark Fasheh 已提交
3127 3128

	if (src == dst)
M
Mark Fasheh 已提交
3129
		same_inode = 1;
M
Mark Fasheh 已提交
3130

3131 3132 3133
	if (len == 0)
		return 0;

M
Mark Fasheh 已提交
3134
	if (same_inode) {
A
Al Viro 已提交
3135
		inode_lock(src);
M
Mark Fasheh 已提交
3136

M
Mark Fasheh 已提交
3137
		ret = extent_same_check_offsets(src, loff, &len, olen);
3138 3139 3140
		if (ret)
			goto out_unlock;
		ret = extent_same_check_offsets(src, dst_loff, &len, olen);
M
Mark Fasheh 已提交
3141 3142
		if (ret)
			goto out_unlock;
M
Mark Fasheh 已提交
3143

M
Mark Fasheh 已提交
3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181
		/*
		 * Single inode case wants the same checks, except we
		 * don't want our length pushed out past i_size as
		 * comparing that data range makes no sense.
		 *
		 * extent_same_check_offsets() will do this for an
		 * unaligned length at i_size, so catch it here and
		 * reject the request.
		 *
		 * This effectively means we require aligned extents
		 * for the single-inode case, whereas the other cases
		 * allow an unaligned length so long as it ends at
		 * i_size.
		 */
		if (len != olen) {
			ret = -EINVAL;
			goto out_unlock;
		}

		/* Check for overlapping ranges */
		if (dst_loff + len > loff && dst_loff < loff + len) {
			ret = -EINVAL;
			goto out_unlock;
		}

		same_lock_start = min_t(u64, loff, dst_loff);
		same_lock_len = max_t(u64, loff, dst_loff) + len - same_lock_start;
	} else {
		btrfs_double_inode_lock(src, dst);

		ret = extent_same_check_offsets(src, loff, &len, olen);
		if (ret)
			goto out_unlock;

		ret = extent_same_check_offsets(dst, dst_loff, &len, olen);
		if (ret)
			goto out_unlock;
	}
M
Mark Fasheh 已提交
3182 3183 3184 3185 3186 3187 3188 3189

	/* don't make the dst file partly checksummed */
	if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) !=
	    (BTRFS_I(dst)->flags & BTRFS_INODE_NODATASUM)) {
		ret = -EINVAL;
		goto out_unlock;
	}

3190
again:
3191 3192 3193 3194
	ret = btrfs_cmp_data_prepare(src, loff, dst, dst_loff, olen, &cmp);
	if (ret)
		goto out_unlock;

M
Mark Fasheh 已提交
3195
	if (same_inode)
3196 3197
		ret = lock_extent_range(src, same_lock_start, same_lock_len,
					false);
M
Mark Fasheh 已提交
3198
	else
3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229
		ret = btrfs_double_extent_lock(src, loff, dst, dst_loff, len,
					       false);
	/*
	 * If one of the inodes has dirty pages in the respective range or
	 * ordered extents, we need to flush dellaloc and wait for all ordered
	 * extents in the range. We must unlock the pages and the ranges in the
	 * io trees to avoid deadlocks when flushing delalloc (requires locking
	 * pages) and when waiting for ordered extents to complete (they require
	 * range locking).
	 */
	if (ret == -EAGAIN) {
		/*
		 * Ranges in the io trees already unlocked. Now unlock all
		 * pages before waiting for all IO to complete.
		 */
		btrfs_cmp_data_free(&cmp);
		if (same_inode) {
			btrfs_wait_ordered_range(src, same_lock_start,
						 same_lock_len);
		} else {
			btrfs_wait_ordered_range(src, loff, len);
			btrfs_wait_ordered_range(dst, dst_loff, len);
		}
		goto again;
	}
	ASSERT(ret == 0);
	if (WARN_ON(ret)) {
		/* ranges in the io trees already unlocked */
		btrfs_cmp_data_free(&cmp);
		return ret;
	}
3230

3231
	/* pass original length for comparison so we stay within i_size */
3232
	ret = btrfs_cmp_data(src, loff, dst, dst_loff, olen, &cmp);
M
Mark Fasheh 已提交
3233
	if (ret == 0)
3234
		ret = btrfs_clone(src, dst, loff, olen, len, dst_loff, 1);
M
Mark Fasheh 已提交
3235

M
Mark Fasheh 已提交
3236 3237 3238 3239 3240
	if (same_inode)
		unlock_extent(&BTRFS_I(src)->io_tree, same_lock_start,
			      same_lock_start + same_lock_len - 1);
	else
		btrfs_double_extent_unlock(src, loff, dst, dst_loff, len);
3241 3242

	btrfs_cmp_data_free(&cmp);
M
Mark Fasheh 已提交
3243
out_unlock:
M
Mark Fasheh 已提交
3244
	if (same_inode)
A
Al Viro 已提交
3245
		inode_unlock(src);
M
Mark Fasheh 已提交
3246 3247
	else
		btrfs_double_inode_unlock(src, dst);
M
Mark Fasheh 已提交
3248 3249 3250 3251

	return ret;
}

3252
#define BTRFS_MAX_DEDUPE_LEN	SZ_16M
M
Mark Fasheh 已提交
3253

3254 3255
ssize_t btrfs_dedupe_file_range(struct file *src_file, u64 loff, u64 olen,
				struct file *dst_file, u64 dst_loff)
M
Mark Fasheh 已提交
3256
{
3257 3258
	struct inode *src = file_inode(src_file);
	struct inode *dst = file_inode(dst_file);
M
Mark Fasheh 已提交
3259
	u64 bs = BTRFS_I(src)->root->fs_info->sb->s_blocksize;
3260
	ssize_t res;
M
Mark Fasheh 已提交
3261

3262 3263
	if (olen > BTRFS_MAX_DEDUPE_LEN)
		olen = BTRFS_MAX_DEDUPE_LEN;
M
Mark Fasheh 已提交
3264

3265
	if (WARN_ON_ONCE(bs < PAGE_SIZE)) {
M
Mark Fasheh 已提交
3266 3267 3268 3269 3270
		/*
		 * Btrfs does not support blocksize < page_size. As a
		 * result, btrfs_cmp_data() won't correctly handle
		 * this situation without an update.
		 */
3271
		return -EINVAL;
M
Mark Fasheh 已提交
3272 3273
	}

3274 3275 3276 3277
	res = btrfs_extent_same(src, loff, olen, dst, dst_loff);
	if (res)
		return res;
	return olen;
M
Mark Fasheh 已提交
3278 3279
}

3280 3281 3282 3283
static int clone_finish_inode_update(struct btrfs_trans_handle *trans,
				     struct inode *inode,
				     u64 endoff,
				     const u64 destoff,
3284 3285
				     const u64 olen,
				     int no_time_update)
3286 3287 3288 3289 3290
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
	int ret;

	inode_inc_iversion(inode);
3291
	if (!no_time_update)
3292
		inode->i_mtime = inode->i_ctime = current_time(inode);
3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303
	/*
	 * We round up to the block size at eof when determining which
	 * extents to clone above, but shouldn't round up the file size.
	 */
	if (endoff > destoff + olen)
		endoff = destoff + olen;
	if (endoff > inode->i_size)
		btrfs_i_size_write(inode, endoff);

	ret = btrfs_update_inode(trans, root, inode);
	if (ret) {
3304
		btrfs_abort_transaction(trans, ret);
3305 3306 3307 3308 3309 3310 3311 3312
		btrfs_end_transaction(trans, root);
		goto out;
	}
	ret = btrfs_end_transaction(trans, root);
out:
	return ret;
}

3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329
static void clone_update_extent_map(struct inode *inode,
				    const struct btrfs_trans_handle *trans,
				    const struct btrfs_path *path,
				    const u64 hole_offset,
				    const u64 hole_len)
{
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	struct extent_map *em;
	int ret;

	em = alloc_extent_map();
	if (!em) {
		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
			&BTRFS_I(inode)->runtime_flags);
		return;
	}

3330 3331 3332 3333 3334
	if (path) {
		struct btrfs_file_extent_item *fi;

		fi = btrfs_item_ptr(path->nodes[0], path->slots[0],
				    struct btrfs_file_extent_item);
3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364
		btrfs_extent_item_to_extent_map(inode, path, fi, false, em);
		em->generation = -1;
		if (btrfs_file_extent_type(path->nodes[0], fi) ==
		    BTRFS_FILE_EXTENT_INLINE)
			set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
				&BTRFS_I(inode)->runtime_flags);
	} else {
		em->start = hole_offset;
		em->len = hole_len;
		em->ram_bytes = em->len;
		em->orig_start = hole_offset;
		em->block_start = EXTENT_MAP_HOLE;
		em->block_len = 0;
		em->orig_block_len = 0;
		em->compress_type = BTRFS_COMPRESS_NONE;
		em->generation = trans->transid;
	}

	while (1) {
		write_lock(&em_tree->lock);
		ret = add_extent_mapping(em_tree, em, 1);
		write_unlock(&em_tree->lock);
		if (ret != -EEXIST) {
			free_extent_map(em);
			break;
		}
		btrfs_drop_extent_cache(inode, em->start,
					em->start + em->len - 1, 0);
	}

3365
	if (ret)
3366 3367 3368 3369
		set_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
			&BTRFS_I(inode)->runtime_flags);
}

3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407
/*
 * Make sure we do not end up inserting an inline extent into a file that has
 * already other (non-inline) extents. If a file has an inline extent it can
 * not have any other extents and the (single) inline extent must start at the
 * file offset 0. Failing to respect these rules will lead to file corruption,
 * resulting in EIO errors on read/write operations, hitting BUG_ON's in mm, etc
 *
 * We can have extents that have been already written to disk or we can have
 * dirty ranges still in delalloc, in which case the extent maps and items are
 * created only when we run delalloc, and the delalloc ranges might fall outside
 * the range we are currently locking in the inode's io tree. So we check the
 * inode's i_size because of that (i_size updates are done while holding the
 * i_mutex, which we are holding here).
 * We also check to see if the inode has a size not greater than "datal" but has
 * extents beyond it, due to an fallocate with FALLOC_FL_KEEP_SIZE (and we are
 * protected against such concurrent fallocate calls by the i_mutex).
 *
 * If the file has no extents but a size greater than datal, do not allow the
 * copy because we would need turn the inline extent into a non-inline one (even
 * with NO_HOLES enabled). If we find our destination inode only has one inline
 * extent, just overwrite it with the source inline extent if its size is less
 * than the source extent's size, or we could copy the source inline extent's
 * data into the destination inode's inline extent if the later is greater then
 * the former.
 */
static int clone_copy_inline_extent(struct inode *src,
				    struct inode *dst,
				    struct btrfs_trans_handle *trans,
				    struct btrfs_path *path,
				    struct btrfs_key *new_key,
				    const u64 drop_start,
				    const u64 datal,
				    const u64 skip,
				    const u64 size,
				    char *inline_data)
{
	struct btrfs_root *root = BTRFS_I(dst)->root;
	const u64 aligned_end = ALIGN(new_key->offset + datal,
3408
				      root->fs_info->sectorsize);
3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513
	int ret;
	struct btrfs_key key;

	if (new_key->offset > 0)
		return -EOPNOTSUPP;

	key.objectid = btrfs_ino(dst);
	key.type = BTRFS_EXTENT_DATA_KEY;
	key.offset = 0;
	ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
	if (ret < 0) {
		return ret;
	} else if (ret > 0) {
		if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
			ret = btrfs_next_leaf(root, path);
			if (ret < 0)
				return ret;
			else if (ret > 0)
				goto copy_inline_extent;
		}
		btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
		if (key.objectid == btrfs_ino(dst) &&
		    key.type == BTRFS_EXTENT_DATA_KEY) {
			ASSERT(key.offset > 0);
			return -EOPNOTSUPP;
		}
	} else if (i_size_read(dst) <= datal) {
		struct btrfs_file_extent_item *ei;
		u64 ext_len;

		/*
		 * If the file size is <= datal, make sure there are no other
		 * extents following (can happen do to an fallocate call with
		 * the flag FALLOC_FL_KEEP_SIZE).
		 */
		ei = btrfs_item_ptr(path->nodes[0], path->slots[0],
				    struct btrfs_file_extent_item);
		/*
		 * If it's an inline extent, it can not have other extents
		 * following it.
		 */
		if (btrfs_file_extent_type(path->nodes[0], ei) ==
		    BTRFS_FILE_EXTENT_INLINE)
			goto copy_inline_extent;

		ext_len = btrfs_file_extent_num_bytes(path->nodes[0], ei);
		if (ext_len > aligned_end)
			return -EOPNOTSUPP;

		ret = btrfs_next_item(root, path);
		if (ret < 0) {
			return ret;
		} else if (ret == 0) {
			btrfs_item_key_to_cpu(path->nodes[0], &key,
					      path->slots[0]);
			if (key.objectid == btrfs_ino(dst) &&
			    key.type == BTRFS_EXTENT_DATA_KEY)
				return -EOPNOTSUPP;
		}
	}

copy_inline_extent:
	/*
	 * We have no extent items, or we have an extent at offset 0 which may
	 * or may not be inlined. All these cases are dealt the same way.
	 */
	if (i_size_read(dst) > datal) {
		/*
		 * If the destination inode has an inline extent...
		 * This would require copying the data from the source inline
		 * extent into the beginning of the destination's inline extent.
		 * But this is really complex, both extents can be compressed
		 * or just one of them, which would require decompressing and
		 * re-compressing data (which could increase the new compressed
		 * size, not allowing the compressed data to fit anymore in an
		 * inline extent).
		 * So just don't support this case for now (it should be rare,
		 * we are not really saving space when cloning inline extents).
		 */
		return -EOPNOTSUPP;
	}

	btrfs_release_path(path);
	ret = btrfs_drop_extents(trans, root, dst, drop_start, aligned_end, 1);
	if (ret)
		return ret;
	ret = btrfs_insert_empty_item(trans, root, path, new_key, size);
	if (ret)
		return ret;

	if (skip) {
		const u32 start = btrfs_file_extent_calc_inline_size(0);

		memmove(inline_data + start, inline_data + start + skip, datal);
	}

	write_extent_buffer(path->nodes[0], inline_data,
			    btrfs_item_ptr_offset(path->nodes[0],
						  path->slots[0]),
			    size);
	inode_add_bytes(dst, datal);

	return 0;
}

3514 3515 3516 3517 3518 3519 3520
/**
 * btrfs_clone() - clone a range from inode file to another
 *
 * @src: Inode to clone from
 * @inode: Inode to clone to
 * @off: Offset within source to start clone from
 * @olen: Original length, passed by user, of range to clone
3521
 * @olen_aligned: Block-aligned value of olen
3522
 * @destoff: Offset within @inode to start clone
3523
 * @no_time_update: Whether to update mtime/ctime on the target inode
3524 3525
 */
static int btrfs_clone(struct inode *src, struct inode *inode,
3526
		       const u64 off, const u64 olen, const u64 olen_aligned,
3527
		       const u64 destoff, int no_time_update)
C
Christoph Hellwig 已提交
3528 3529
{
	struct btrfs_root *root = BTRFS_I(inode)->root;
3530
	struct btrfs_path *path = NULL;
C
Christoph Hellwig 已提交
3531
	struct extent_buffer *leaf;
3532 3533
	struct btrfs_trans_handle *trans;
	char *buf = NULL;
Y
Yan Zheng 已提交
3534
	struct btrfs_key key;
C
Christoph Hellwig 已提交
3535 3536
	u32 nritems;
	int slot;
Y
Yan Zheng 已提交
3537
	int ret;
3538 3539
	const u64 len = olen_aligned;
	u64 last_dest_end = destoff;
Y
Yan Zheng 已提交
3540 3541

	ret = -ENOMEM;
3542 3543
	buf = kmalloc(root->fs_info->nodesize,
		      GFP_KERNEL | __GFP_NOWARN);
3544
	if (!buf) {
3545
		buf = vmalloc(root->fs_info->nodesize);
3546 3547 3548
		if (!buf)
			return ret;
	}
Y
Yan Zheng 已提交
3549 3550 3551

	path = btrfs_alloc_path();
	if (!path) {
3552
		kvfree(buf);
3553
		return ret;
3554 3555
	}

3556
	path->reada = READA_FORWARD;
3557
	/* clone data */
L
Li Zefan 已提交
3558
	key.objectid = btrfs_ino(src);
Y
Yan Zheng 已提交
3559
	key.type = BTRFS_EXTENT_DATA_KEY;
3560
	key.offset = off;
C
Christoph Hellwig 已提交
3561 3562

	while (1) {
3563
		u64 next_key_min_offset = key.offset + 1;
3564

C
Christoph Hellwig 已提交
3565 3566 3567 3568
		/*
		 * note the key will change type as we walk through the
		 * tree.
		 */
3569
		path->leave_spinning = 1;
3570 3571
		ret = btrfs_search_slot(NULL, BTRFS_I(src)->root, &key, path,
				0, 0);
C
Christoph Hellwig 已提交
3572 3573
		if (ret < 0)
			goto out;
3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584
		/*
		 * First search, if no extent item that starts at offset off was
		 * found but the previous item is an extent item, it's possible
		 * it might overlap our target range, therefore process it.
		 */
		if (key.offset == off && ret > 0 && path->slots[0] > 0) {
			btrfs_item_key_to_cpu(path->nodes[0], &key,
					      path->slots[0] - 1);
			if (key.type == BTRFS_EXTENT_DATA_KEY)
				path->slots[0]--;
		}
C
Christoph Hellwig 已提交
3585

Y
Yan Zheng 已提交
3586
		nritems = btrfs_header_nritems(path->nodes[0]);
3587
process_slot:
Y
Yan Zheng 已提交
3588
		if (path->slots[0] >= nritems) {
3589
			ret = btrfs_next_leaf(BTRFS_I(src)->root, path);
C
Christoph Hellwig 已提交
3590 3591 3592 3593
			if (ret < 0)
				goto out;
			if (ret > 0)
				break;
Y
Yan Zheng 已提交
3594
			nritems = btrfs_header_nritems(path->nodes[0]);
C
Christoph Hellwig 已提交
3595 3596 3597 3598
		}
		leaf = path->nodes[0];
		slot = path->slots[0];

Y
Yan Zheng 已提交
3599
		btrfs_item_key_to_cpu(leaf, &key, slot);
3600
		if (key.type > BTRFS_EXTENT_DATA_KEY ||
L
Li Zefan 已提交
3601
		    key.objectid != btrfs_ino(src))
C
Christoph Hellwig 已提交
3602 3603
			break;

3604
		if (key.type == BTRFS_EXTENT_DATA_KEY) {
3605 3606
			struct btrfs_file_extent_item *extent;
			int type;
Z
Zheng Yan 已提交
3607 3608
			u32 size;
			struct btrfs_key new_key;
3609 3610 3611
			u64 disko = 0, diskl = 0;
			u64 datao = 0, datal = 0;
			u8 comp;
3612
			u64 drop_start;
Z
Zheng Yan 已提交
3613

3614 3615 3616 3617
			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);
3618 3619
			if (type == BTRFS_FILE_EXTENT_REG ||
			    type == BTRFS_FILE_EXTENT_PREALLOC) {
C
Chris Mason 已提交
3620 3621 3622 3623
				disko = btrfs_file_extent_disk_bytenr(leaf,
								      extent);
				diskl = btrfs_file_extent_disk_num_bytes(leaf,
								 extent);
3624
				datao = btrfs_file_extent_offset(leaf, extent);
C
Chris Mason 已提交
3625 3626
				datal = btrfs_file_extent_num_bytes(leaf,
								    extent);
3627 3628 3629 3630 3631
			} 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
3632

3633 3634 3635 3636 3637 3638
			/*
			 * The first search might have left us at an extent
			 * item that ends before our target range's start, can
			 * happen if we have holes and NO_HOLES feature enabled.
			 */
			if (key.offset + datal <= off) {
3639 3640
				path->slots[0]++;
				goto process_slot;
3641 3642
			} else if (key.offset >= off + len) {
				break;
3643
			}
3644
			next_key_min_offset = key.offset + datal;
3645 3646 3647 3648 3649 3650 3651
			size = btrfs_item_size_nr(leaf, slot);
			read_extent_buffer(leaf, buf,
					   btrfs_item_ptr_offset(leaf, slot),
					   size);

			btrfs_release_path(path);
			path->leave_spinning = 0;
3652

Z
Zheng Yan 已提交
3653
			memcpy(&new_key, &key, sizeof(new_key));
L
Li Zefan 已提交
3654
			new_key.objectid = btrfs_ino(inode);
3655 3656 3657 3658
			if (off <= key.offset)
				new_key.offset = key.offset + destoff - off;
			else
				new_key.offset = destoff;
Z
Zheng Yan 已提交
3659

3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671
			/*
			 * Deal with a hole that doesn't have an extent item
			 * that represents it (NO_HOLES feature enabled).
			 * This hole is either in the middle of the cloning
			 * range or at the beginning (fully overlaps it or
			 * partially overlaps it).
			 */
			if (new_key.offset != last_dest_end)
				drop_start = last_dest_end;
			else
				drop_start = new_key.offset;

3672 3673 3674 3675 3676 3677
			/*
			 * 1 - adjusting old extent (we may have to split it)
			 * 1 - add new extent
			 * 1 - inode update
			 */
			trans = btrfs_start_transaction(root, 3);
3678 3679 3680 3681 3682
			if (IS_ERR(trans)) {
				ret = PTR_ERR(trans);
				goto out;
			}

3683 3684
			if (type == BTRFS_FILE_EXTENT_REG ||
			    type == BTRFS_FILE_EXTENT_PREALLOC) {
3685 3686 3687 3688 3689
				/*
				 *    a  | --- range to clone ---|  b
				 * | ------------- extent ------------- |
				 */

3690
				/* subtract range b */
3691 3692 3693
				if (key.offset + datal > off + len)
					datal = off + len - key.offset;

3694
				/* subtract range a */
3695 3696 3697 3698 3699
				if (off > key.offset) {
					datao += off - key.offset;
					datal -= off - key.offset;
				}

J
Josef Bacik 已提交
3700
				ret = btrfs_drop_extents(trans, root, inode,
3701
							 drop_start,
3702
							 new_key.offset + datal,
3703
							 1);
3704
				if (ret) {
3705
					if (ret != -EOPNOTSUPP)
3706
						btrfs_abort_transaction(trans,
3707
									ret);
3708 3709 3710
					btrfs_end_transaction(trans, root);
					goto out;
				}
3711

3712 3713
				ret = btrfs_insert_empty_item(trans, root, path,
							      &new_key, size);
3714
				if (ret) {
3715
					btrfs_abort_transaction(trans, ret);
3716 3717 3718
					btrfs_end_transaction(trans, root);
					goto out;
				}
3719 3720 3721 3722

				leaf = path->nodes[0];
				slot = path->slots[0];
				write_extent_buffer(leaf, buf,
Z
Zheng Yan 已提交
3723 3724
					    btrfs_item_ptr_offset(leaf, slot),
					    size);
Y
Yan Zheng 已提交
3725

3726
				extent = btrfs_item_ptr(leaf, slot,
C
Christoph Hellwig 已提交
3727
						struct btrfs_file_extent_item);
3728 3729 3730 3731 3732 3733 3734 3735 3736

				/* 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);
J
Josef Bacik 已提交
3737

3738 3739
				if (disko) {
					inode_add_bytes(inode, datal);
Y
Yan Zheng 已提交
3740
					ret = btrfs_inc_extent_ref(trans, root,
3741 3742
							disko, diskl, 0,
							root->root_key.objectid,
L
Li Zefan 已提交
3743
							btrfs_ino(inode),
3744
							new_key.offset - datao);
3745 3746 3747 3748 3749 3750 3751 3752
					if (ret) {
						btrfs_abort_transaction(trans,
									ret);
						btrfs_end_transaction(trans,
								      root);
						goto out;

					}
C
Christoph Hellwig 已提交
3753
				}
3754 3755 3756
			} else if (type == BTRFS_FILE_EXTENT_INLINE) {
				u64 skip = 0;
				u64 trim = 0;
3757

3758 3759 3760 3761
				if (off > key.offset) {
					skip = off - key.offset;
					new_key.offset += skip;
				}
C
Chris Mason 已提交
3762

L
Liu Bo 已提交
3763 3764
				if (key.offset + datal > off + len)
					trim = key.offset + datal - (off + len);
C
Chris Mason 已提交
3765

3766 3767
				if (comp && (skip || trim)) {
					ret = -EINVAL;
3768
					btrfs_end_transaction(trans, root);
3769 3770 3771 3772
					goto out;
				}
				size -= skip + trim;
				datal -= skip + trim;
3773

3774 3775 3776 3777 3778 3779
				ret = clone_copy_inline_extent(src, inode,
							       trans, path,
							       &new_key,
							       drop_start,
							       datal,
							       skip, size, buf);
3780
				if (ret) {
3781
					if (ret != -EOPNOTSUPP)
3782
						btrfs_abort_transaction(trans,
3783
									ret);
3784 3785 3786
					btrfs_end_transaction(trans, root);
					goto out;
				}
3787 3788
				leaf = path->nodes[0];
				slot = path->slots[0];
C
Christoph Hellwig 已提交
3789
			}
3790

3791 3792 3793
			/* If we have an implicit hole (NO_HOLES feature). */
			if (drop_start < new_key.offset)
				clone_update_extent_map(inode, trans,
3794
						NULL, drop_start,
3795 3796
						new_key.offset - drop_start);

3797
			clone_update_extent_map(inode, trans, path, 0, 0);
3798

3799
			btrfs_mark_buffer_dirty(leaf);
3800
			btrfs_release_path(path);
3801

3802
			last_dest_end = ALIGN(new_key.offset + datal,
3803
					      root->fs_info->sectorsize);
3804 3805
			ret = clone_finish_inode_update(trans, inode,
							last_dest_end,
3806 3807
							destoff, olen,
							no_time_update);
3808
			if (ret)
3809
				goto out;
3810 3811
			if (new_key.offset + datal >= destoff + len)
				break;
3812
		}
3813
		btrfs_release_path(path);
3814
		key.offset = next_key_min_offset;
3815 3816 3817 3818 3819

		if (fatal_signal_pending(current)) {
			ret = -EINTR;
			goto out;
		}
C
Christoph Hellwig 已提交
3820 3821
	}
	ret = 0;
3822

3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 3841 3842
	if (last_dest_end < destoff + len) {
		/*
		 * We have an implicit hole (NO_HOLES feature is enabled) that
		 * fully or partially overlaps our cloning range at its end.
		 */
		btrfs_release_path(path);

		/*
		 * 1 - remove extent(s)
		 * 1 - inode update
		 */
		trans = btrfs_start_transaction(root, 2);
		if (IS_ERR(trans)) {
			ret = PTR_ERR(trans);
			goto out;
		}
		ret = btrfs_drop_extents(trans, root, inode,
					 last_dest_end, destoff + len, 1);
		if (ret) {
			if (ret != -EOPNOTSUPP)
3843
				btrfs_abort_transaction(trans, ret);
3844 3845 3846
			btrfs_end_transaction(trans, root);
			goto out;
		}
3847 3848
		clone_update_extent_map(inode, trans, NULL, last_dest_end,
					destoff + len - last_dest_end);
3849
		ret = clone_finish_inode_update(trans, inode, destoff + len,
3850
						destoff, olen, no_time_update);
3851 3852
	}

C
Christoph Hellwig 已提交
3853
out:
3854
	btrfs_free_path(path);
3855
	kvfree(buf);
3856 3857 3858
	return ret;
}

3859 3860
static noinline int btrfs_clone_files(struct file *file, struct file *file_src,
					u64 off, u64 olen, u64 destoff)
3861
{
A
Al Viro 已提交
3862
	struct inode *inode = file_inode(file);
3863
	struct inode *src = file_inode(file_src);
3864 3865 3866 3867
	struct btrfs_root *root = BTRFS_I(inode)->root;
	int ret;
	u64 len = olen;
	u64 bs = root->fs_info->sb->s_blocksize;
3868
	int same_inode = src == inode;
3869 3870 3871 3872 3873 3874 3875

	/*
	 * 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.
3876 3877 3878
	 *
	 * - split destination inode's inline extents.  The inline extents can
	 *   be either compressed or non-compressed.
3879 3880 3881 3882 3883
	 */

	if (btrfs_root_readonly(root))
		return -EROFS;

3884 3885 3886
	if (file_src->f_path.mnt != file->f_path.mnt ||
	    src->i_sb != inode->i_sb)
		return -EXDEV;
3887 3888 3889 3890

	/* don't make the dst file partly checksummed */
	if ((BTRFS_I(src)->flags & BTRFS_INODE_NODATASUM) !=
	    (BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM))
3891
		return -EINVAL;
3892 3893

	if (S_ISDIR(src->i_mode) || S_ISDIR(inode->i_mode))
3894
		return -EISDIR;
3895 3896

	if (!same_inode) {
3897
		btrfs_double_inode_lock(src, inode);
3898
	} else {
A
Al Viro 已提交
3899
		inode_lock(src);
3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911
	}

	/* determine range to clone */
	ret = -EINVAL;
	if (off + len > src->i_size || off + len < off)
		goto out_unlock;
	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 = ALIGN(src->i_size, bs) - off;

3912 3913 3914 3915 3916
	if (len == 0) {
		ret = 0;
		goto out_unlock;
	}

3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933
	/* verify the end result is block aligned */
	if (!IS_ALIGNED(off, bs) || !IS_ALIGNED(off + len, bs) ||
	    !IS_ALIGNED(destoff, bs))
		goto out_unlock;

	/* verify if ranges are overlapped within the same file */
	if (same_inode) {
		if (destoff + len > off && destoff < off + len)
			goto out_unlock;
	}

	if (destoff > inode->i_size) {
		ret = btrfs_cont_expand(inode, inode->i_size, destoff);
		if (ret)
			goto out_unlock;
	}

3934 3935 3936 3937 3938 3939 3940 3941 3942 3943
	/*
	 * Lock the target range too. Right after we replace the file extent
	 * items in the fs tree (which now point to the cloned data), we might
	 * have a worker replace them with extent items relative to a write
	 * operation that was issued before this clone operation (i.e. confront
	 * with inode.c:btrfs_finish_ordered_io).
	 */
	if (same_inode) {
		u64 lock_start = min_t(u64, off, destoff);
		u64 lock_len = max_t(u64, off, destoff) + len - lock_start;
3944

3945
		ret = lock_extent_range(src, lock_start, lock_len, true);
3946
	} else {
3947 3948 3949 3950 3951 3952 3953
		ret = btrfs_double_extent_lock(src, off, inode, destoff, len,
					       true);
	}
	ASSERT(ret == 0);
	if (WARN_ON(ret)) {
		/* ranges in the io trees already unlocked */
		goto out_unlock;
3954
	}
3955

3956
	ret = btrfs_clone(src, inode, off, olen, len, destoff, 0);
3957

3958 3959 3960 3961 3962 3963
	if (same_inode) {
		u64 lock_start = min_t(u64, off, destoff);
		u64 lock_end = max_t(u64, off, destoff) + len - 1;

		unlock_extent(&BTRFS_I(src)->io_tree, lock_start, lock_end);
	} else {
3964
		btrfs_double_extent_unlock(src, off, inode, destoff, len);
3965 3966 3967 3968 3969
	}
	/*
	 * Truncate page cache pages so that future reads will see the cloned
	 * data immediately and not the previous data.
	 */
3970
	truncate_inode_pages_range(&inode->i_data,
3971 3972
				round_down(destoff, PAGE_SIZE),
				round_up(destoff + len, PAGE_SIZE) - 1);
C
Christoph Hellwig 已提交
3973
out_unlock:
3974 3975 3976
	if (!same_inode)
		btrfs_double_inode_unlock(src, inode);
	else
A
Al Viro 已提交
3977
		inode_unlock(src);
3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992
	return ret;
}

ssize_t btrfs_copy_file_range(struct file *file_in, loff_t pos_in,
			      struct file *file_out, loff_t pos_out,
			      size_t len, unsigned int flags)
{
	ssize_t ret;

	ret = btrfs_clone_files(file_out, file_in, pos_in, len, pos_out);
	if (ret == 0)
		ret = len;
	return ret;
}

3993 3994
int btrfs_clone_file_range(struct file *src_file, loff_t off,
		struct file *dst_file, loff_t destoff, u64 len)
3995
{
3996
	return btrfs_clone_files(dst_file, src_file, off, len, destoff);
3997 3998
}

C
Christoph Hellwig 已提交
3999 4000 4001 4002 4003 4004
/*
 * 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.
 */
4005
static long btrfs_ioctl_trans_start(struct file *file)
C
Christoph Hellwig 已提交
4006
{
A
Al Viro 已提交
4007
	struct inode *inode = file_inode(file);
C
Christoph Hellwig 已提交
4008 4009
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;
4010
	int ret;
C
Christoph Hellwig 已提交
4011

4012
	ret = -EPERM;
4013
	if (!capable(CAP_SYS_ADMIN))
4014
		goto out;
4015

4016 4017
	ret = -EINPROGRESS;
	if (file->private_data)
C
Christoph Hellwig 已提交
4018
		goto out;
4019

L
Li Zefan 已提交
4020 4021 4022 4023
	ret = -EROFS;
	if (btrfs_root_readonly(root))
		goto out;

4024
	ret = mnt_want_write_file(file);
Y
Yan Zheng 已提交
4025 4026 4027
	if (ret)
		goto out;

J
Josef Bacik 已提交
4028
	atomic_inc(&root->fs_info->open_ioctl_trans);
4029

4030
	ret = -ENOMEM;
4031
	trans = btrfs_start_ioctl_transaction(root);
4032
	if (IS_ERR(trans))
4033 4034 4035 4036 4037 4038
		goto out_drop;

	file->private_data = trans;
	return 0;

out_drop:
J
Josef Bacik 已提交
4039
	atomic_dec(&root->fs_info->open_ioctl_trans);
A
Al Viro 已提交
4040
	mnt_drop_write_file(file);
C
Christoph Hellwig 已提交
4041 4042 4043 4044
out:
	return ret;
}

4045 4046
static long btrfs_ioctl_default_subvol(struct file *file, void __user *argp)
{
A
Al Viro 已提交
4047
	struct inode *inode = file_inode(file);
4048 4049 4050 4051 4052 4053 4054 4055 4056
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_root *new_root;
	struct btrfs_dir_item *di;
	struct btrfs_trans_handle *trans;
	struct btrfs_path *path;
	struct btrfs_key location;
	struct btrfs_disk_key disk_key;
	u64 objectid = 0;
	u64 dir_id;
4057
	int ret;
4058 4059 4060 4061

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4062 4063 4064 4065 4066 4067 4068 4069
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;

	if (copy_from_user(&objectid, argp, sizeof(objectid))) {
		ret = -EFAULT;
		goto out;
	}
4070 4071

	if (!objectid)
4072
		objectid = BTRFS_FS_TREE_OBJECTID;
4073 4074 4075 4076 4077 4078

	location.objectid = objectid;
	location.type = BTRFS_ROOT_ITEM_KEY;
	location.offset = (u64)-1;

	new_root = btrfs_read_fs_root_no_name(root->fs_info, &location);
4079 4080 4081 4082
	if (IS_ERR(new_root)) {
		ret = PTR_ERR(new_root);
		goto out;
	}
4083 4084

	path = btrfs_alloc_path();
4085 4086 4087 4088
	if (!path) {
		ret = -ENOMEM;
		goto out;
	}
4089 4090 4091
	path->leave_spinning = 1;

	trans = btrfs_start_transaction(root, 1);
4092
	if (IS_ERR(trans)) {
4093
		btrfs_free_path(path);
4094 4095
		ret = PTR_ERR(trans);
		goto out;
4096 4097
	}

4098
	dir_id = btrfs_super_root_dir(root->fs_info->super_copy);
4099 4100
	di = btrfs_lookup_dir_item(trans, root->fs_info->tree_root, path,
				   dir_id, "default", 7, 1);
4101
	if (IS_ERR_OR_NULL(di)) {
4102 4103
		btrfs_free_path(path);
		btrfs_end_transaction(trans, root);
J
Jeff Mahoney 已提交
4104 4105
		btrfs_err(new_root->fs_info,
			  "Umm, you don't have the default diritem, this isn't going to work");
4106 4107
		ret = -ENOENT;
		goto out;
4108 4109 4110 4111 4112 4113 4114
	}

	btrfs_cpu_key_to_disk(&disk_key, &new_root->root_key);
	btrfs_set_dir_item_key(path->nodes[0], di, &disk_key);
	btrfs_mark_buffer_dirty(path->nodes[0]);
	btrfs_free_path(path);

4115
	btrfs_set_fs_incompat(root->fs_info, DEFAULT_SUBVOL);
4116
	btrfs_end_transaction(trans, root);
4117 4118 4119
out:
	mnt_drop_write_file(file);
	return ret;
4120 4121
}

4122 4123
void btrfs_get_block_group_info(struct list_head *groups_list,
				struct btrfs_ioctl_space_info *space)
4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137
{
	struct btrfs_block_group_cache *block_group;

	space->total_bytes = 0;
	space->used_bytes = 0;
	space->flags = 0;
	list_for_each_entry(block_group, groups_list, list) {
		space->flags = block_group->flags;
		space->total_bytes += block_group->key.offset;
		space->used_bytes +=
			btrfs_block_group_used(&block_group->item);
	}
}

4138
static long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg)
J
Josef Bacik 已提交
4139 4140 4141 4142
{
	struct btrfs_ioctl_space_args space_args;
	struct btrfs_ioctl_space_info space;
	struct btrfs_ioctl_space_info *dest;
4143
	struct btrfs_ioctl_space_info *dest_orig;
4144
	struct btrfs_ioctl_space_info __user *user_dest;
J
Josef Bacik 已提交
4145
	struct btrfs_space_info *info;
4146 4147 4148 4149 4150
	u64 types[] = {BTRFS_BLOCK_GROUP_DATA,
		       BTRFS_BLOCK_GROUP_SYSTEM,
		       BTRFS_BLOCK_GROUP_METADATA,
		       BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA};
	int num_types = 4;
4151
	int alloc_size;
J
Josef Bacik 已提交
4152
	int ret = 0;
4153
	u64 slot_count = 0;
4154
	int i, c;
J
Josef Bacik 已提交
4155 4156 4157 4158 4159 4160

	if (copy_from_user(&space_args,
			   (struct btrfs_ioctl_space_args __user *)arg,
			   sizeof(space_args)))
		return -EFAULT;

4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184
	for (i = 0; i < num_types; i++) {
		struct btrfs_space_info *tmp;

		info = NULL;
		rcu_read_lock();
		list_for_each_entry_rcu(tmp, &root->fs_info->space_info,
					list) {
			if (tmp->flags == types[i]) {
				info = tmp;
				break;
			}
		}
		rcu_read_unlock();

		if (!info)
			continue;

		down_read(&info->groups_sem);
		for (c = 0; c < BTRFS_NR_RAID_TYPES; c++) {
			if (!list_empty(&info->block_groups[c]))
				slot_count++;
		}
		up_read(&info->groups_sem);
	}
4185

4186 4187 4188 4189 4190
	/*
	 * Global block reserve, exported as a space_info
	 */
	slot_count++;

4191 4192 4193 4194 4195
	/* space_slots == 0 means they are asking for a count */
	if (space_args.space_slots == 0) {
		space_args.total_spaces = slot_count;
		goto out;
	}
4196

4197
	slot_count = min_t(u64, space_args.space_slots, slot_count);
4198

4199
	alloc_size = sizeof(*dest) * slot_count;
4200

4201 4202 4203
	/* we generally have at most 6 or so space infos, one for each raid
	 * level.  So, a whole page should be more than enough for everyone
	 */
4204
	if (alloc_size > PAGE_SIZE)
4205 4206
		return -ENOMEM;

J
Josef Bacik 已提交
4207
	space_args.total_spaces = 0;
4208
	dest = kmalloc(alloc_size, GFP_KERNEL);
4209 4210 4211
	if (!dest)
		return -ENOMEM;
	dest_orig = dest;
J
Josef Bacik 已提交
4212

4213
	/* now we have a buffer to copy into */
4214 4215 4216
	for (i = 0; i < num_types; i++) {
		struct btrfs_space_info *tmp;

4217 4218 4219
		if (!slot_count)
			break;

4220 4221 4222 4223 4224 4225 4226 4227 4228 4229
		info = NULL;
		rcu_read_lock();
		list_for_each_entry_rcu(tmp, &root->fs_info->space_info,
					list) {
			if (tmp->flags == types[i]) {
				info = tmp;
				break;
			}
		}
		rcu_read_unlock();
4230

4231 4232 4233 4234 4235
		if (!info)
			continue;
		down_read(&info->groups_sem);
		for (c = 0; c < BTRFS_NR_RAID_TYPES; c++) {
			if (!list_empty(&info->block_groups[c])) {
4236 4237
				btrfs_get_block_group_info(
					&info->block_groups[c], &space);
4238 4239 4240
				memcpy(dest, &space, sizeof(space));
				dest++;
				space_args.total_spaces++;
4241
				slot_count--;
4242
			}
4243 4244
			if (!slot_count)
				break;
4245 4246
		}
		up_read(&info->groups_sem);
J
Josef Bacik 已提交
4247 4248
	}

4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263
	/*
	 * Add global block reserve
	 */
	if (slot_count) {
		struct btrfs_block_rsv *block_rsv = &root->fs_info->global_block_rsv;

		spin_lock(&block_rsv->lock);
		space.total_bytes = block_rsv->size;
		space.used_bytes = block_rsv->size - block_rsv->reserved;
		spin_unlock(&block_rsv->lock);
		space.flags = BTRFS_SPACE_INFO_GLOBAL_RSV;
		memcpy(dest, &space, sizeof(space));
		space_args.total_spaces++;
	}

D
Daniel J Blueman 已提交
4264
	user_dest = (struct btrfs_ioctl_space_info __user *)
4265 4266 4267 4268 4269 4270 4271 4272
		(arg + sizeof(struct btrfs_ioctl_space_args));

	if (copy_to_user(user_dest, dest_orig, alloc_size))
		ret = -EFAULT;

	kfree(dest_orig);
out:
	if (ret == 0 && copy_to_user(arg, &space_args, sizeof(space_args)))
J
Josef Bacik 已提交
4273 4274 4275 4276 4277
		ret = -EFAULT;

	return ret;
}

C
Christoph Hellwig 已提交
4278 4279 4280 4281 4282 4283 4284 4285
/*
 * 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)
{
A
Al Viro 已提交
4286
	struct inode *inode = file_inode(file);
C
Christoph Hellwig 已提交
4287 4288 4289 4290
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_trans_handle *trans;

	trans = file->private_data;
4291 4292
	if (!trans)
		return -EINVAL;
C
Christoph Hellwig 已提交
4293
	file->private_data = NULL;
4294

4295 4296
	btrfs_end_transaction(trans, root);

J
Josef Bacik 已提交
4297
	atomic_dec(&root->fs_info->open_ioctl_trans);
4298

A
Al Viro 已提交
4299
	mnt_drop_write_file(file);
4300
	return 0;
C
Christoph Hellwig 已提交
4301 4302
}

4303 4304
static noinline long btrfs_ioctl_start_sync(struct btrfs_root *root,
					    void __user *argp)
4305 4306 4307
{
	struct btrfs_trans_handle *trans;
	u64 transid;
T
Tsutomu Itoh 已提交
4308
	int ret;
4309

M
Miao Xie 已提交
4310
	trans = btrfs_attach_transaction_barrier(root);
4311 4312 4313 4314 4315 4316 4317 4318
	if (IS_ERR(trans)) {
		if (PTR_ERR(trans) != -ENOENT)
			return PTR_ERR(trans);

		/* No running transaction, don't bother */
		transid = root->fs_info->last_trans_committed;
		goto out;
	}
4319
	transid = trans->transid;
T
Tsutomu Itoh 已提交
4320
	ret = btrfs_commit_transaction_async(trans, root, 0);
4321 4322
	if (ret) {
		btrfs_end_transaction(trans, root);
T
Tsutomu Itoh 已提交
4323
		return ret;
4324
	}
4325
out:
4326 4327 4328 4329 4330 4331
	if (argp)
		if (copy_to_user(argp, &transid, sizeof(transid)))
			return -EFAULT;
	return 0;
}

4332 4333
static noinline long btrfs_ioctl_wait_sync(struct btrfs_root *root,
					   void __user *argp)
4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345
{
	u64 transid;

	if (argp) {
		if (copy_from_user(&transid, argp, sizeof(transid)))
			return -EFAULT;
	} else {
		transid = 0;  /* current trans */
	}
	return btrfs_wait_for_commit(root, transid);
}

M
Miao Xie 已提交
4346
static long btrfs_ioctl_scrub(struct file *file, void __user *arg)
J
Jan Schmidt 已提交
4347
{
A
Al Viro 已提交
4348
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
J
Jan Schmidt 已提交
4349
	struct btrfs_ioctl_scrub_args *sa;
M
Miao Xie 已提交
4350
	int ret;
J
Jan Schmidt 已提交
4351 4352 4353 4354 4355 4356 4357 4358

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	sa = memdup_user(arg, sizeof(*sa));
	if (IS_ERR(sa))
		return PTR_ERR(sa);

M
Miao Xie 已提交
4359 4360 4361 4362 4363 4364
	if (!(sa->flags & BTRFS_SCRUB_READONLY)) {
		ret = mnt_want_write_file(file);
		if (ret)
			goto out;
	}

4365
	ret = btrfs_scrub_dev(root->fs_info, sa->devid, sa->start, sa->end,
4366 4367
			      &sa->progress, sa->flags & BTRFS_SCRUB_READONLY,
			      0);
J
Jan Schmidt 已提交
4368 4369 4370 4371

	if (copy_to_user(arg, sa, sizeof(*sa)))
		ret = -EFAULT;

M
Miao Xie 已提交
4372 4373 4374
	if (!(sa->flags & BTRFS_SCRUB_READONLY))
		mnt_drop_write_file(file);
out:
J
Jan Schmidt 已提交
4375 4376 4377 4378 4379 4380 4381 4382 4383
	kfree(sa);
	return ret;
}

static long btrfs_ioctl_scrub_cancel(struct btrfs_root *root, void __user *arg)
{
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4384
	return btrfs_scrub_cancel(root->fs_info);
J
Jan Schmidt 已提交
4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408
}

static long btrfs_ioctl_scrub_progress(struct btrfs_root *root,
				       void __user *arg)
{
	struct btrfs_ioctl_scrub_args *sa;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	sa = memdup_user(arg, sizeof(*sa));
	if (IS_ERR(sa))
		return PTR_ERR(sa);

	ret = btrfs_scrub_progress(root, sa->devid, &sa->progress);

	if (copy_to_user(arg, sa, sizeof(*sa)))
		ret = -EFAULT;

	kfree(sa);
	return ret;
}

4409
static long btrfs_ioctl_get_dev_stats(struct btrfs_root *root,
4410
				      void __user *arg)
4411 4412 4413 4414 4415 4416 4417 4418
{
	struct btrfs_ioctl_get_dev_stats *sa;
	int ret;

	sa = memdup_user(arg, sizeof(*sa));
	if (IS_ERR(sa))
		return PTR_ERR(sa);

4419 4420 4421 4422 4423 4424
	if ((sa->flags & BTRFS_DEV_STATS_RESET) && !capable(CAP_SYS_ADMIN)) {
		kfree(sa);
		return -EPERM;
	}

	ret = btrfs_get_dev_stats(root, sa);
4425 4426 4427 4428 4429 4430 4431 4432

	if (copy_to_user(arg, sa, sizeof(*sa)))
		ret = -EFAULT;

	kfree(sa);
	return ret;
}

4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446
static long btrfs_ioctl_dev_replace(struct btrfs_root *root, void __user *arg)
{
	struct btrfs_ioctl_dev_replace_args *p;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	p = memdup_user(arg, sizeof(*p));
	if (IS_ERR(p))
		return PTR_ERR(p);

	switch (p->cmd) {
	case BTRFS_IOCTL_DEV_REPLACE_CMD_START:
4447 4448 4449 4450
		if (root->fs_info->sb->s_flags & MS_RDONLY) {
			ret = -EROFS;
			goto out;
		}
4451 4452 4453
		if (atomic_xchg(
			&root->fs_info->mutually_exclusive_operation_running,
			1)) {
4454
			ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
4455
		} else {
4456
			ret = btrfs_dev_replace_by_ioctl(root, p);
4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474 4475
			atomic_set(
			 &root->fs_info->mutually_exclusive_operation_running,
			 0);
		}
		break;
	case BTRFS_IOCTL_DEV_REPLACE_CMD_STATUS:
		btrfs_dev_replace_status(root->fs_info, p);
		ret = 0;
		break;
	case BTRFS_IOCTL_DEV_REPLACE_CMD_CANCEL:
		ret = btrfs_dev_replace_cancel(root->fs_info, p);
		break;
	default:
		ret = -EINVAL;
		break;
	}

	if (copy_to_user(arg, p, sizeof(*p)))
		ret = -EFAULT;
4476
out:
4477 4478 4479 4480
	kfree(p);
	return ret;
}

4481 4482 4483 4484
static long btrfs_ioctl_ino_to_path(struct btrfs_root *root, void __user *arg)
{
	int ret = 0;
	int i;
4485
	u64 rel_ptr;
4486
	int size;
4487
	struct btrfs_ioctl_ino_path_args *ipa = NULL;
4488 4489 4490
	struct inode_fs_paths *ipath = NULL;
	struct btrfs_path *path;

4491
	if (!capable(CAP_DAC_READ_SEARCH))
4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519
		return -EPERM;

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

	ipa = memdup_user(arg, sizeof(*ipa));
	if (IS_ERR(ipa)) {
		ret = PTR_ERR(ipa);
		ipa = NULL;
		goto out;
	}

	size = min_t(u32, ipa->size, 4096);
	ipath = init_ipath(size, root, path);
	if (IS_ERR(ipath)) {
		ret = PTR_ERR(ipath);
		ipath = NULL;
		goto out;
	}

	ret = paths_from_inode(ipa->inum, ipath);
	if (ret < 0)
		goto out;

	for (i = 0; i < ipath->fspath->elem_cnt; ++i) {
4520 4521
		rel_ptr = ipath->fspath->val[i] -
			  (u64)(unsigned long)ipath->fspath->val;
4522
		ipath->fspath->val[i] = rel_ptr;
4523 4524
	}

4525 4526
	ret = copy_to_user((void *)(unsigned long)ipa->fspath,
			   (void *)(unsigned long)ipath->fspath, size);
4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572
	if (ret) {
		ret = -EFAULT;
		goto out;
	}

out:
	btrfs_free_path(path);
	free_ipath(ipath);
	kfree(ipa);

	return ret;
}

static int build_ino_list(u64 inum, u64 offset, u64 root, void *ctx)
{
	struct btrfs_data_container *inodes = ctx;
	const size_t c = 3 * sizeof(u64);

	if (inodes->bytes_left >= c) {
		inodes->bytes_left -= c;
		inodes->val[inodes->elem_cnt] = inum;
		inodes->val[inodes->elem_cnt + 1] = offset;
		inodes->val[inodes->elem_cnt + 2] = root;
		inodes->elem_cnt += 3;
	} else {
		inodes->bytes_missing += c - inodes->bytes_left;
		inodes->bytes_left = 0;
		inodes->elem_missed += 3;
	}

	return 0;
}

static long btrfs_ioctl_logical_to_ino(struct btrfs_root *root,
					void __user *arg)
{
	int ret = 0;
	int size;
	struct btrfs_ioctl_logical_ino_args *loi;
	struct btrfs_data_container *inodes = NULL;
	struct btrfs_path *path = NULL;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	loi = memdup_user(arg, sizeof(*loi));
4573 4574
	if (IS_ERR(loi))
		return PTR_ERR(loi);
4575 4576 4577 4578 4579 4580 4581

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

4582
	size = min_t(u32, loi->size, SZ_64K);
4583 4584 4585 4586 4587 4588 4589
	inodes = init_data_container(size);
	if (IS_ERR(inodes)) {
		ret = PTR_ERR(inodes);
		inodes = NULL;
		goto out;
	}

L
Liu Bo 已提交
4590 4591 4592
	ret = iterate_inodes_from_logical(loi->logical, root->fs_info, path,
					  build_ino_list, inodes);
	if (ret == -EINVAL)
4593 4594 4595 4596
		ret = -ENOENT;
	if (ret < 0)
		goto out;

4597 4598
	ret = copy_to_user((void *)(unsigned long)loi->inodes,
			   (void *)(unsigned long)inodes, size);
4599 4600 4601 4602 4603
	if (ret)
		ret = -EFAULT;

out:
	btrfs_free_path(path);
4604
	vfree(inodes);
4605 4606 4607 4608 4609
	kfree(loi);

	return ret;
}

4610
void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock,
4611 4612 4613 4614 4615 4616
			       struct btrfs_ioctl_balance_args *bargs)
{
	struct btrfs_balance_control *bctl = fs_info->balance_ctl;

	bargs->flags = bctl->flags;

4617 4618 4619 4620
	if (atomic_read(&fs_info->balance_running))
		bargs->state |= BTRFS_BALANCE_STATE_RUNNING;
	if (atomic_read(&fs_info->balance_pause_req))
		bargs->state |= BTRFS_BALANCE_STATE_PAUSE_REQ;
4621 4622
	if (atomic_read(&fs_info->balance_cancel_req))
		bargs->state |= BTRFS_BALANCE_STATE_CANCEL_REQ;
4623

4624 4625 4626
	memcpy(&bargs->data, &bctl->data, sizeof(bargs->data));
	memcpy(&bargs->meta, &bctl->meta, sizeof(bargs->meta));
	memcpy(&bargs->sys, &bctl->sys, sizeof(bargs->sys));
4627 4628 4629 4630 4631 4632 4633 4634

	if (lock) {
		spin_lock(&fs_info->balance_lock);
		memcpy(&bargs->stat, &bctl->stat, sizeof(bargs->stat));
		spin_unlock(&fs_info->balance_lock);
	} else {
		memcpy(&bargs->stat, &bctl->stat, sizeof(bargs->stat));
	}
4635 4636
}

4637
static long btrfs_ioctl_balance(struct file *file, void __user *arg)
4638
{
A
Al Viro 已提交
4639
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
4640 4641 4642
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_ioctl_balance_args *bargs;
	struct btrfs_balance_control *bctl;
4643
	bool need_unlock; /* for mut. excl. ops lock */
4644 4645 4646 4647 4648
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4649
	ret = mnt_want_write_file(file);
4650 4651 4652
	if (ret)
		return ret;

4653 4654 4655 4656 4657 4658 4659 4660 4661
again:
	if (!atomic_xchg(&fs_info->mutually_exclusive_operation_running, 1)) {
		mutex_lock(&fs_info->volume_mutex);
		mutex_lock(&fs_info->balance_mutex);
		need_unlock = true;
		goto locked;
	}

	/*
4662
	 * mut. excl. ops lock is locked.  Three possibilities:
4663 4664 4665 4666
	 *   (1) some other op is running
	 *   (2) balance is running
	 *   (3) balance is paused -- special case (think resume)
	 */
4667
	mutex_lock(&fs_info->balance_mutex);
4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694
	if (fs_info->balance_ctl) {
		/* this is either (2) or (3) */
		if (!atomic_read(&fs_info->balance_running)) {
			mutex_unlock(&fs_info->balance_mutex);
			if (!mutex_trylock(&fs_info->volume_mutex))
				goto again;
			mutex_lock(&fs_info->balance_mutex);

			if (fs_info->balance_ctl &&
			    !atomic_read(&fs_info->balance_running)) {
				/* this is (3) */
				need_unlock = false;
				goto locked;
			}

			mutex_unlock(&fs_info->balance_mutex);
			mutex_unlock(&fs_info->volume_mutex);
			goto again;
		} else {
			/* this is (2) */
			mutex_unlock(&fs_info->balance_mutex);
			ret = -EINPROGRESS;
			goto out;
		}
	} else {
		/* this is (1) */
		mutex_unlock(&fs_info->balance_mutex);
4695
		ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
4696 4697 4698 4699 4700
		goto out;
	}

locked:
	BUG_ON(!atomic_read(&fs_info->mutually_exclusive_operation_running));
4701 4702 4703 4704 4705

	if (arg) {
		bargs = memdup_user(arg, sizeof(*bargs));
		if (IS_ERR(bargs)) {
			ret = PTR_ERR(bargs);
4706
			goto out_unlock;
4707
		}
4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718 4719 4720 4721

		if (bargs->flags & BTRFS_BALANCE_RESUME) {
			if (!fs_info->balance_ctl) {
				ret = -ENOTCONN;
				goto out_bargs;
			}

			bctl = fs_info->balance_ctl;
			spin_lock(&fs_info->balance_lock);
			bctl->flags |= BTRFS_BALANCE_RESUME;
			spin_unlock(&fs_info->balance_lock);

			goto do_balance;
		}
4722 4723 4724 4725
	} else {
		bargs = NULL;
	}

4726
	if (fs_info->balance_ctl) {
4727 4728 4729 4730
		ret = -EINPROGRESS;
		goto out_bargs;
	}

4731
	bctl = kzalloc(sizeof(*bctl), GFP_KERNEL);
4732 4733 4734 4735 4736 4737 4738 4739 4740 4741 4742 4743
	if (!bctl) {
		ret = -ENOMEM;
		goto out_bargs;
	}

	bctl->fs_info = fs_info;
	if (arg) {
		memcpy(&bctl->data, &bargs->data, sizeof(bctl->data));
		memcpy(&bctl->meta, &bargs->meta, sizeof(bctl->meta));
		memcpy(&bctl->sys, &bargs->sys, sizeof(bctl->sys));

		bctl->flags = bargs->flags;
4744 4745 4746
	} else {
		/* balance everything - no filters */
		bctl->flags |= BTRFS_BALANCE_TYPE_MASK;
4747 4748
	}

4749 4750
	if (bctl->flags & ~(BTRFS_BALANCE_ARGS_MASK | BTRFS_BALANCE_TYPE_MASK)) {
		ret = -EINVAL;
4751
		goto out_bctl;
4752 4753
	}

4754
do_balance:
4755
	/*
4756 4757 4758 4759 4760
	 * Ownership of bctl and mutually_exclusive_operation_running
	 * goes to to btrfs_balance.  bctl is freed in __cancel_balance,
	 * or, if restriper was paused all the way until unmount, in
	 * free_fs_info.  mutually_exclusive_operation_running is
	 * cleared in __cancel_balance.
4761
	 */
4762 4763 4764
	need_unlock = false;

	ret = btrfs_balance(bctl, bargs);
4765
	bctl = NULL;
4766

4767 4768 4769 4770 4771
	if (arg) {
		if (copy_to_user(arg, bargs, sizeof(*bargs)))
			ret = -EFAULT;
	}

4772 4773
out_bctl:
	kfree(bctl);
4774 4775
out_bargs:
	kfree(bargs);
4776
out_unlock:
4777 4778
	mutex_unlock(&fs_info->balance_mutex);
	mutex_unlock(&fs_info->volume_mutex);
4779 4780 4781
	if (need_unlock)
		atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
out:
4782
	mnt_drop_write_file(file);
4783 4784 4785
	return ret;
}

4786 4787 4788 4789 4790 4791 4792 4793
static long btrfs_ioctl_balance_ctl(struct btrfs_root *root, int cmd)
{
	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	switch (cmd) {
	case BTRFS_BALANCE_CTL_PAUSE:
		return btrfs_pause_balance(root->fs_info);
4794 4795
	case BTRFS_BALANCE_CTL_CANCEL:
		return btrfs_cancel_balance(root->fs_info);
4796 4797 4798 4799 4800
	}

	return -EINVAL;
}

4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813 4814 4815 4816
static long btrfs_ioctl_balance_progress(struct btrfs_root *root,
					 void __user *arg)
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	struct btrfs_ioctl_balance_args *bargs;
	int ret = 0;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	mutex_lock(&fs_info->balance_mutex);
	if (!fs_info->balance_ctl) {
		ret = -ENOTCONN;
		goto out;
	}

4817
	bargs = kzalloc(sizeof(*bargs), GFP_KERNEL);
4818 4819 4820 4821 4822 4823 4824 4825 4826 4827 4828 4829 4830 4831 4832 4833
	if (!bargs) {
		ret = -ENOMEM;
		goto out;
	}

	update_ioctl_balance_args(fs_info, 1, bargs);

	if (copy_to_user(arg, bargs, sizeof(*bargs)))
		ret = -EFAULT;

	kfree(bargs);
out:
	mutex_unlock(&fs_info->balance_mutex);
	return ret;
}

4834
static long btrfs_ioctl_quota_ctl(struct file *file, void __user *arg)
A
Arne Jansen 已提交
4835
{
A
Al Viro 已提交
4836
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
A
Arne Jansen 已提交
4837 4838 4839 4840 4841 4842 4843 4844
	struct btrfs_ioctl_quota_ctl_args *sa;
	struct btrfs_trans_handle *trans = NULL;
	int ret;
	int err;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4845 4846 4847
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
A
Arne Jansen 已提交
4848 4849

	sa = memdup_user(arg, sizeof(*sa));
4850 4851 4852 4853
	if (IS_ERR(sa)) {
		ret = PTR_ERR(sa);
		goto drop_write;
	}
A
Arne Jansen 已提交
4854

4855
	down_write(&root->fs_info->subvol_sem);
J
Jan Schmidt 已提交
4856 4857 4858 4859
	trans = btrfs_start_transaction(root->fs_info->tree_root, 2);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out;
A
Arne Jansen 已提交
4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873
	}

	switch (sa->cmd) {
	case BTRFS_QUOTA_CTL_ENABLE:
		ret = btrfs_quota_enable(trans, root->fs_info);
		break;
	case BTRFS_QUOTA_CTL_DISABLE:
		ret = btrfs_quota_disable(trans, root->fs_info);
		break;
	default:
		ret = -EINVAL;
		break;
	}

J
Jan Schmidt 已提交
4874 4875 4876
	err = btrfs_commit_transaction(trans, root->fs_info->tree_root);
	if (err && !ret)
		ret = err;
A
Arne Jansen 已提交
4877 4878
out:
	kfree(sa);
4879
	up_write(&root->fs_info->subvol_sem);
4880 4881
drop_write:
	mnt_drop_write_file(file);
A
Arne Jansen 已提交
4882 4883 4884
	return ret;
}

4885
static long btrfs_ioctl_qgroup_assign(struct file *file, void __user *arg)
A
Arne Jansen 已提交
4886
{
A
Al Viro 已提交
4887
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
A
Arne Jansen 已提交
4888 4889 4890 4891 4892 4893 4894 4895
	struct btrfs_ioctl_qgroup_assign_args *sa;
	struct btrfs_trans_handle *trans;
	int ret;
	int err;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4896 4897 4898
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
A
Arne Jansen 已提交
4899 4900

	sa = memdup_user(arg, sizeof(*sa));
4901 4902 4903 4904
	if (IS_ERR(sa)) {
		ret = PTR_ERR(sa);
		goto drop_write;
	}
A
Arne Jansen 已提交
4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917 4918 4919 4920

	trans = btrfs_join_transaction(root);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out;
	}

	/* FIXME: check if the IDs really exist */
	if (sa->assign) {
		ret = btrfs_add_qgroup_relation(trans, root->fs_info,
						sa->src, sa->dst);
	} else {
		ret = btrfs_del_qgroup_relation(trans, root->fs_info,
						sa->src, sa->dst);
	}

4921 4922 4923
	/* update qgroup status and info */
	err = btrfs_run_qgroups(trans, root->fs_info);
	if (err < 0)
4924 4925
		btrfs_handle_fs_error(root->fs_info, err,
			    "failed to update qgroup status and info");
A
Arne Jansen 已提交
4926 4927 4928 4929 4930 4931
	err = btrfs_end_transaction(trans, root);
	if (err && !ret)
		ret = err;

out:
	kfree(sa);
4932 4933
drop_write:
	mnt_drop_write_file(file);
A
Arne Jansen 已提交
4934 4935 4936
	return ret;
}

4937
static long btrfs_ioctl_qgroup_create(struct file *file, void __user *arg)
A
Arne Jansen 已提交
4938
{
A
Al Viro 已提交
4939
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
A
Arne Jansen 已提交
4940 4941 4942 4943 4944 4945 4946 4947
	struct btrfs_ioctl_qgroup_create_args *sa;
	struct btrfs_trans_handle *trans;
	int ret;
	int err;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4948 4949 4950
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
A
Arne Jansen 已提交
4951 4952

	sa = memdup_user(arg, sizeof(*sa));
4953 4954 4955 4956
	if (IS_ERR(sa)) {
		ret = PTR_ERR(sa);
		goto drop_write;
	}
A
Arne Jansen 已提交
4957

M
Miao Xie 已提交
4958 4959 4960 4961 4962
	if (!sa->qgroupid) {
		ret = -EINVAL;
		goto out;
	}

A
Arne Jansen 已提交
4963 4964 4965 4966 4967 4968 4969 4970
	trans = btrfs_join_transaction(root);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out;
	}

	/* FIXME: check if the IDs really exist */
	if (sa->create) {
4971
		ret = btrfs_create_qgroup(trans, root->fs_info, sa->qgroupid);
A
Arne Jansen 已提交
4972 4973 4974 4975 4976 4977 4978 4979 4980 4981
	} else {
		ret = btrfs_remove_qgroup(trans, root->fs_info, sa->qgroupid);
	}

	err = btrfs_end_transaction(trans, root);
	if (err && !ret)
		ret = err;

out:
	kfree(sa);
4982 4983
drop_write:
	mnt_drop_write_file(file);
A
Arne Jansen 已提交
4984 4985 4986
	return ret;
}

4987
static long btrfs_ioctl_qgroup_limit(struct file *file, void __user *arg)
A
Arne Jansen 已提交
4988
{
A
Al Viro 已提交
4989
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
A
Arne Jansen 已提交
4990 4991 4992 4993 4994 4995 4996 4997 4998
	struct btrfs_ioctl_qgroup_limit_args *sa;
	struct btrfs_trans_handle *trans;
	int ret;
	int err;
	u64 qgroupid;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

4999 5000 5001
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;
A
Arne Jansen 已提交
5002 5003

	sa = memdup_user(arg, sizeof(*sa));
5004 5005 5006 5007
	if (IS_ERR(sa)) {
		ret = PTR_ERR(sa);
		goto drop_write;
	}
A
Arne Jansen 已提交
5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029

	trans = btrfs_join_transaction(root);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out;
	}

	qgroupid = sa->qgroupid;
	if (!qgroupid) {
		/* take the current subvol as qgroup */
		qgroupid = root->root_key.objectid;
	}

	/* FIXME: check if the IDs really exist */
	ret = btrfs_limit_qgroup(trans, root->fs_info, qgroupid, &sa->lim);

	err = btrfs_end_transaction(trans, root);
	if (err && !ret)
		ret = err;

out:
	kfree(sa);
5030 5031
drop_write:
	mnt_drop_write_file(file);
A
Arne Jansen 已提交
5032 5033 5034
	return ret;
}

J
Jan Schmidt 已提交
5035 5036
static long btrfs_ioctl_quota_rescan(struct file *file, void __user *arg)
{
A
Al Viro 已提交
5037
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
J
Jan Schmidt 已提交
5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069
	struct btrfs_ioctl_quota_rescan_args *qsa;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	ret = mnt_want_write_file(file);
	if (ret)
		return ret;

	qsa = memdup_user(arg, sizeof(*qsa));
	if (IS_ERR(qsa)) {
		ret = PTR_ERR(qsa);
		goto drop_write;
	}

	if (qsa->flags) {
		ret = -EINVAL;
		goto out;
	}

	ret = btrfs_qgroup_rescan(root->fs_info);

out:
	kfree(qsa);
drop_write:
	mnt_drop_write_file(file);
	return ret;
}

static long btrfs_ioctl_quota_rescan_status(struct file *file, void __user *arg)
{
A
Al Viro 已提交
5070
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
J
Jan Schmidt 已提交
5071 5072 5073 5074 5075 5076
	struct btrfs_ioctl_quota_rescan_args *qsa;
	int ret = 0;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

5077
	qsa = kzalloc(sizeof(*qsa), GFP_KERNEL);
J
Jan Schmidt 已提交
5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 5090 5091 5092
	if (!qsa)
		return -ENOMEM;

	if (root->fs_info->qgroup_flags & BTRFS_QGROUP_STATUS_FLAG_RESCAN) {
		qsa->flags = 1;
		qsa->progress = root->fs_info->qgroup_rescan_progress.objectid;
	}

	if (copy_to_user(arg, qsa, sizeof(*qsa)))
		ret = -EFAULT;

	kfree(qsa);
	return ret;
}

5093 5094
static long btrfs_ioctl_quota_rescan_wait(struct file *file, void __user *arg)
{
A
Al Viro 已提交
5095
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
5096 5097 5098 5099

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

5100
	return btrfs_qgroup_wait_for_completion(root->fs_info, true);
5101 5102
}

5103 5104
static long _btrfs_ioctl_set_received_subvol(struct file *file,
					    struct btrfs_ioctl_received_subvol_args *sa)
5105
{
A
Al Viro 已提交
5106
	struct inode *inode = file_inode(file);
5107 5108 5109
	struct btrfs_root *root = BTRFS_I(inode)->root;
	struct btrfs_root_item *root_item = &root->root_item;
	struct btrfs_trans_handle *trans;
5110
	struct timespec ct = current_time(inode);
5111
	int ret = 0;
5112
	int received_uuid_changed;
5113

5114 5115 5116
	if (!inode_owner_or_capable(inode))
		return -EPERM;

5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132
	ret = mnt_want_write_file(file);
	if (ret < 0)
		return ret;

	down_write(&root->fs_info->subvol_sem);

	if (btrfs_ino(inode) != BTRFS_FIRST_FREE_OBJECTID) {
		ret = -EINVAL;
		goto out;
	}

	if (btrfs_root_readonly(root)) {
		ret = -EROFS;
		goto out;
	}

5133 5134 5135 5136 5137
	/*
	 * 1 - root item
	 * 2 - uuid items (received uuid + subvol uuid)
	 */
	trans = btrfs_start_transaction(root, 3);
5138 5139 5140 5141 5142 5143 5144 5145 5146 5147
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		trans = NULL;
		goto out;
	}

	sa->rtransid = trans->transid;
	sa->rtime.sec = ct.tv_sec;
	sa->rtime.nsec = ct.tv_nsec;

5148 5149 5150 5151
	received_uuid_changed = memcmp(root_item->received_uuid, sa->uuid,
				       BTRFS_UUID_SIZE);
	if (received_uuid_changed &&
	    !btrfs_is_empty_uuid(root_item->received_uuid))
5152
		btrfs_uuid_tree_rem(trans, root->fs_info,
5153 5154 5155
				    root_item->received_uuid,
				    BTRFS_UUID_KEY_RECEIVED_SUBVOL,
				    root->root_key.objectid);
5156 5157 5158
	memcpy(root_item->received_uuid, sa->uuid, BTRFS_UUID_SIZE);
	btrfs_set_root_stransid(root_item, sa->stransid);
	btrfs_set_root_rtransid(root_item, sa->rtransid);
5159 5160 5161 5162
	btrfs_set_stack_timespec_sec(&root_item->stime, sa->stime.sec);
	btrfs_set_stack_timespec_nsec(&root_item->stime, sa->stime.nsec);
	btrfs_set_stack_timespec_sec(&root_item->rtime, sa->rtime.sec);
	btrfs_set_stack_timespec_nsec(&root_item->rtime, sa->rtime.nsec);
5163 5164 5165 5166 5167 5168

	ret = btrfs_update_root(trans, root->fs_info->tree_root,
				&root->root_key, &root->root_item);
	if (ret < 0) {
		btrfs_end_transaction(trans, root);
		goto out;
5169 5170
	}
	if (received_uuid_changed && !btrfs_is_empty_uuid(sa->uuid)) {
5171
		ret = btrfs_uuid_tree_add(trans, root->fs_info, sa->uuid,
5172 5173 5174
					  BTRFS_UUID_KEY_RECEIVED_SUBVOL,
					  root->root_key.objectid);
		if (ret < 0 && ret != -EEXIST) {
5175
			btrfs_abort_transaction(trans, ret);
5176
			goto out;
5177 5178 5179 5180
		}
	}
	ret = btrfs_commit_transaction(trans, root);
	if (ret < 0) {
5181
		btrfs_abort_transaction(trans, ret);
5182
		goto out;
5183 5184
	}

5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199
out:
	up_write(&root->fs_info->subvol_sem);
	mnt_drop_write_file(file);
	return ret;
}

#ifdef CONFIG_64BIT
static long btrfs_ioctl_set_received_subvol_32(struct file *file,
						void __user *arg)
{
	struct btrfs_ioctl_received_subvol_args_32 *args32 = NULL;
	struct btrfs_ioctl_received_subvol_args *args64 = NULL;
	int ret = 0;

	args32 = memdup_user(arg, sizeof(*args32));
5200 5201
	if (IS_ERR(args32))
		return PTR_ERR(args32);
5202

5203
	args64 = kmalloc(sizeof(*args64), GFP_KERNEL);
5204 5205
	if (!args64) {
		ret = -ENOMEM;
5206 5207 5208 5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248
		goto out;
	}

	memcpy(args64->uuid, args32->uuid, BTRFS_UUID_SIZE);
	args64->stransid = args32->stransid;
	args64->rtransid = args32->rtransid;
	args64->stime.sec = args32->stime.sec;
	args64->stime.nsec = args32->stime.nsec;
	args64->rtime.sec = args32->rtime.sec;
	args64->rtime.nsec = args32->rtime.nsec;
	args64->flags = args32->flags;

	ret = _btrfs_ioctl_set_received_subvol(file, args64);
	if (ret)
		goto out;

	memcpy(args32->uuid, args64->uuid, BTRFS_UUID_SIZE);
	args32->stransid = args64->stransid;
	args32->rtransid = args64->rtransid;
	args32->stime.sec = args64->stime.sec;
	args32->stime.nsec = args64->stime.nsec;
	args32->rtime.sec = args64->rtime.sec;
	args32->rtime.nsec = args64->rtime.nsec;
	args32->flags = args64->flags;

	ret = copy_to_user(arg, args32, sizeof(*args32));
	if (ret)
		ret = -EFAULT;

out:
	kfree(args32);
	kfree(args64);
	return ret;
}
#endif

static long btrfs_ioctl_set_received_subvol(struct file *file,
					    void __user *arg)
{
	struct btrfs_ioctl_received_subvol_args *sa = NULL;
	int ret = 0;

	sa = memdup_user(arg, sizeof(*sa));
5249 5250
	if (IS_ERR(sa))
		return PTR_ERR(sa);
5251 5252 5253 5254 5255 5256

	ret = _btrfs_ioctl_set_received_subvol(file, sa);

	if (ret)
		goto out;

5257 5258 5259 5260 5261 5262 5263 5264 5265
	ret = copy_to_user(arg, sa, sizeof(*sa));
	if (ret)
		ret = -EFAULT;

out:
	kfree(sa);
	return ret;
}

5266 5267
static int btrfs_ioctl_get_fslabel(struct file *file, void __user *arg)
{
A
Al Viro 已提交
5268
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
5269
	size_t len;
5270
	int ret;
5271 5272 5273 5274 5275 5276 5277
	char label[BTRFS_LABEL_SIZE];

	spin_lock(&root->fs_info->super_lock);
	memcpy(label, root->fs_info->super_copy->label, BTRFS_LABEL_SIZE);
	spin_unlock(&root->fs_info->super_lock);

	len = strnlen(label, BTRFS_LABEL_SIZE);
5278 5279

	if (len == BTRFS_LABEL_SIZE) {
5280 5281
		btrfs_warn(root->fs_info,
			"label is too long, return the first %zu bytes", --len);
5282 5283 5284 5285 5286 5287 5288
	}

	ret = copy_to_user(arg, label, len);

	return ret ? -EFAULT : 0;
}

5289 5290
static int btrfs_ioctl_set_fslabel(struct file *file, void __user *arg)
{
A
Al Viro 已提交
5291
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303
	struct btrfs_super_block *super_block = root->fs_info->super_copy;
	struct btrfs_trans_handle *trans;
	char label[BTRFS_LABEL_SIZE];
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	if (copy_from_user(label, arg, sizeof(label)))
		return -EFAULT;

	if (strnlen(label, BTRFS_LABEL_SIZE) == BTRFS_LABEL_SIZE) {
J
Jeff Mahoney 已提交
5304 5305 5306
		btrfs_err(root->fs_info,
			  "unable to set label with more than %d bytes",
			  BTRFS_LABEL_SIZE - 1);
5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319
		return -EINVAL;
	}

	ret = mnt_want_write_file(file);
	if (ret)
		return ret;

	trans = btrfs_start_transaction(root, 0);
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out_unlock;
	}

5320
	spin_lock(&root->fs_info->super_lock);
5321
	strcpy(super_block->label, label);
5322
	spin_unlock(&root->fs_info->super_lock);
5323
	ret = btrfs_commit_transaction(trans, root);
5324 5325 5326 5327 5328 5329

out_unlock:
	mnt_drop_write_file(file);
	return ret;
}

5330 5331 5332 5333 5334
#define INIT_FEATURE_FLAGS(suffix) \
	{ .compat_flags = BTRFS_FEATURE_COMPAT_##suffix, \
	  .compat_ro_flags = BTRFS_FEATURE_COMPAT_RO_##suffix, \
	  .incompat_flags = BTRFS_FEATURE_INCOMPAT_##suffix }

5335
int btrfs_ioctl_get_supported_features(void __user *arg)
5336
{
D
David Sterba 已提交
5337
	static const struct btrfs_ioctl_feature_flags features[3] = {
5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 5363 5364
		INIT_FEATURE_FLAGS(SUPP),
		INIT_FEATURE_FLAGS(SAFE_SET),
		INIT_FEATURE_FLAGS(SAFE_CLEAR)
	};

	if (copy_to_user(arg, &features, sizeof(features)))
		return -EFAULT;

	return 0;
}

static int btrfs_ioctl_get_features(struct file *file, void __user *arg)
{
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
	struct btrfs_super_block *super_block = root->fs_info->super_copy;
	struct btrfs_ioctl_feature_flags features;

	features.compat_flags = btrfs_super_compat_flags(super_block);
	features.compat_ro_flags = btrfs_super_compat_ro_flags(super_block);
	features.incompat_flags = btrfs_super_incompat_flags(super_block);

	if (copy_to_user(arg, &features, sizeof(features)))
		return -EFAULT;

	return 0;
}

5365 5366
static int check_feature_bits(struct btrfs_root *root,
			      enum btrfs_feature_set set,
5367 5368 5369
			      u64 change_mask, u64 flags, u64 supported_flags,
			      u64 safe_set, u64 safe_clear)
{
5370 5371
	const char *type = btrfs_feature_set_names[set];
	char *names;
5372 5373 5374 5375 5376 5377
	u64 disallowed, unsupported;
	u64 set_mask = flags & change_mask;
	u64 clear_mask = ~flags & change_mask;

	unsupported = set_mask & ~supported_flags;
	if (unsupported) {
5378 5379 5380 5381 5382 5383 5384 5385
		names = btrfs_printable_features(set, unsupported);
		if (names) {
			btrfs_warn(root->fs_info,
			   "this kernel does not support the %s feature bit%s",
			   names, strchr(names, ',') ? "s" : "");
			kfree(names);
		} else
			btrfs_warn(root->fs_info,
5386 5387 5388 5389 5390 5391 5392
			   "this kernel does not support %s bits 0x%llx",
			   type, unsupported);
		return -EOPNOTSUPP;
	}

	disallowed = set_mask & ~safe_set;
	if (disallowed) {
5393 5394 5395 5396 5397 5398 5399 5400
		names = btrfs_printable_features(set, disallowed);
		if (names) {
			btrfs_warn(root->fs_info,
			   "can't set the %s feature bit%s while mounted",
			   names, strchr(names, ',') ? "s" : "");
			kfree(names);
		} else
			btrfs_warn(root->fs_info,
5401 5402 5403 5404 5405 5406 5407
			   "can't set %s bits 0x%llx while mounted",
			   type, disallowed);
		return -EPERM;
	}

	disallowed = clear_mask & ~safe_clear;
	if (disallowed) {
5408 5409 5410 5411 5412 5413 5414 5415
		names = btrfs_printable_features(set, disallowed);
		if (names) {
			btrfs_warn(root->fs_info,
			   "can't clear the %s feature bit%s while mounted",
			   names, strchr(names, ',') ? "s" : "");
			kfree(names);
		} else
			btrfs_warn(root->fs_info,
5416 5417 5418 5419 5420 5421 5422 5423 5424
			   "can't clear %s bits 0x%llx while mounted",
			   type, disallowed);
		return -EPERM;
	}

	return 0;
}

#define check_feature(root, change_mask, flags, mask_base)	\
5425
check_feature_bits(root, FEAT_##mask_base, change_mask, flags,	\
5426 5427 5428 5429 5430 5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445 5446 5447 5448 5449 5450 5451 5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 5463 5464
		   BTRFS_FEATURE_ ## mask_base ## _SUPP,	\
		   BTRFS_FEATURE_ ## mask_base ## _SAFE_SET,	\
		   BTRFS_FEATURE_ ## mask_base ## _SAFE_CLEAR)

static int btrfs_ioctl_set_features(struct file *file, void __user *arg)
{
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
	struct btrfs_super_block *super_block = root->fs_info->super_copy;
	struct btrfs_ioctl_feature_flags flags[2];
	struct btrfs_trans_handle *trans;
	u64 newflags;
	int ret;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

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

	/* Nothing to do */
	if (!flags[0].compat_flags && !flags[0].compat_ro_flags &&
	    !flags[0].incompat_flags)
		return 0;

	ret = check_feature(root, flags[0].compat_flags,
			    flags[1].compat_flags, COMPAT);
	if (ret)
		return ret;

	ret = check_feature(root, flags[0].compat_ro_flags,
			    flags[1].compat_ro_flags, COMPAT_RO);
	if (ret)
		return ret;

	ret = check_feature(root, flags[0].incompat_flags,
			    flags[1].incompat_flags, INCOMPAT);
	if (ret)
		return ret;

5465 5466 5467 5468
	ret = mnt_want_write_file(file);
	if (ret)
		return ret;

5469
	trans = btrfs_start_transaction(root, 0);
5470 5471 5472 5473
	if (IS_ERR(trans)) {
		ret = PTR_ERR(trans);
		goto out_drop_write;
	}
5474 5475 5476 5477 5478 5479 5480 5481 5482 5483 5484 5485 5486 5487 5488 5489 5490 5491

	spin_lock(&root->fs_info->super_lock);
	newflags = btrfs_super_compat_flags(super_block);
	newflags |= flags[0].compat_flags & flags[1].compat_flags;
	newflags &= ~(flags[0].compat_flags & ~flags[1].compat_flags);
	btrfs_set_super_compat_flags(super_block, newflags);

	newflags = btrfs_super_compat_ro_flags(super_block);
	newflags |= flags[0].compat_ro_flags & flags[1].compat_ro_flags;
	newflags &= ~(flags[0].compat_ro_flags & ~flags[1].compat_ro_flags);
	btrfs_set_super_compat_ro_flags(super_block, newflags);

	newflags = btrfs_super_incompat_flags(super_block);
	newflags |= flags[0].incompat_flags & flags[1].incompat_flags;
	newflags &= ~(flags[0].incompat_flags & ~flags[1].incompat_flags);
	btrfs_set_super_incompat_flags(super_block, newflags);
	spin_unlock(&root->fs_info->super_lock);

5492 5493 5494 5495 5496
	ret = btrfs_commit_transaction(trans, root);
out_drop_write:
	mnt_drop_write_file(file);

	return ret;
5497 5498
}

C
Christoph Hellwig 已提交
5499 5500 5501
long btrfs_ioctl(struct file *file, unsigned int
		cmd, unsigned long arg)
{
A
Al Viro 已提交
5502
	struct btrfs_root *root = BTRFS_I(file_inode(file))->root;
5503
	void __user *argp = (void __user *)arg;
C
Christoph Hellwig 已提交
5504 5505

	switch (cmd) {
5506 5507 5508 5509 5510 5511
	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);
5512 5513
	case FITRIM:
		return btrfs_ioctl_fitrim(file, argp);
C
Christoph Hellwig 已提交
5514
	case BTRFS_IOC_SNAP_CREATE:
5515
		return btrfs_ioctl_snap_create(file, argp, 0);
5516
	case BTRFS_IOC_SNAP_CREATE_V2:
5517
		return btrfs_ioctl_snap_create_v2(file, argp, 0);
5518
	case BTRFS_IOC_SUBVOL_CREATE:
5519
		return btrfs_ioctl_snap_create(file, argp, 1);
A
Arne Jansen 已提交
5520 5521
	case BTRFS_IOC_SUBVOL_CREATE_V2:
		return btrfs_ioctl_snap_create_v2(file, argp, 1);
5522 5523
	case BTRFS_IOC_SNAP_DESTROY:
		return btrfs_ioctl_snap_destroy(file, argp);
5524 5525 5526 5527
	case BTRFS_IOC_SUBVOL_GETFLAGS:
		return btrfs_ioctl_subvol_getflags(file, argp);
	case BTRFS_IOC_SUBVOL_SETFLAGS:
		return btrfs_ioctl_subvol_setflags(file, argp);
5528 5529
	case BTRFS_IOC_DEFAULT_SUBVOL:
		return btrfs_ioctl_default_subvol(file, argp);
C
Christoph Hellwig 已提交
5530
	case BTRFS_IOC_DEFRAG:
C
Chris Mason 已提交
5531 5532 5533
		return btrfs_ioctl_defrag(file, NULL);
	case BTRFS_IOC_DEFRAG_RANGE:
		return btrfs_ioctl_defrag(file, argp);
C
Christoph Hellwig 已提交
5534
	case BTRFS_IOC_RESIZE:
5535
		return btrfs_ioctl_resize(file, argp);
C
Christoph Hellwig 已提交
5536
	case BTRFS_IOC_ADD_DEV:
5537
		return btrfs_ioctl_add_dev(root, argp);
C
Christoph Hellwig 已提交
5538
	case BTRFS_IOC_RM_DEV:
5539
		return btrfs_ioctl_rm_dev(file, argp);
5540 5541
	case BTRFS_IOC_RM_DEV_V2:
		return btrfs_ioctl_rm_dev_v2(file, argp);
J
Jan Schmidt 已提交
5542 5543 5544 5545
	case BTRFS_IOC_FS_INFO:
		return btrfs_ioctl_fs_info(root, argp);
	case BTRFS_IOC_DEV_INFO:
		return btrfs_ioctl_dev_info(root, argp);
C
Christoph Hellwig 已提交
5546
	case BTRFS_IOC_BALANCE:
5547
		return btrfs_ioctl_balance(file, NULL);
C
Christoph Hellwig 已提交
5548 5549 5550 5551
	case BTRFS_IOC_TRANS_START:
		return btrfs_ioctl_trans_start(file);
	case BTRFS_IOC_TRANS_END:
		return btrfs_ioctl_trans_end(file);
5552 5553
	case BTRFS_IOC_TREE_SEARCH:
		return btrfs_ioctl_tree_search(file, argp);
G
Gerhard Heift 已提交
5554 5555
	case BTRFS_IOC_TREE_SEARCH_V2:
		return btrfs_ioctl_tree_search_v2(file, argp);
5556 5557
	case BTRFS_IOC_INO_LOOKUP:
		return btrfs_ioctl_ino_lookup(file, argp);
5558 5559 5560 5561
	case BTRFS_IOC_INO_PATHS:
		return btrfs_ioctl_ino_to_path(root, argp);
	case BTRFS_IOC_LOGICAL_INO:
		return btrfs_ioctl_logical_to_ino(root, argp);
J
Josef Bacik 已提交
5562 5563
	case BTRFS_IOC_SPACE_INFO:
		return btrfs_ioctl_space_info(root, argp);
5564 5565 5566
	case BTRFS_IOC_SYNC: {
		int ret;

5567
		ret = btrfs_start_delalloc_roots(root->fs_info, 0, -1);
5568 5569
		if (ret)
			return ret;
A
Al Viro 已提交
5570
		ret = btrfs_sync_fs(file_inode(file)->i_sb, 1);
5571 5572
		/*
		 * The transaction thread may want to do more work,
5573
		 * namely it pokes the cleaner kthread that will start
5574 5575 5576
		 * processing uncleaned subvols.
		 */
		wake_up_process(root->fs_info->transaction_kthread);
5577 5578
		return ret;
	}
5579
	case BTRFS_IOC_START_SYNC:
5580
		return btrfs_ioctl_start_sync(root, argp);
5581
	case BTRFS_IOC_WAIT_SYNC:
5582
		return btrfs_ioctl_wait_sync(root, argp);
J
Jan Schmidt 已提交
5583
	case BTRFS_IOC_SCRUB:
M
Miao Xie 已提交
5584
		return btrfs_ioctl_scrub(file, argp);
J
Jan Schmidt 已提交
5585 5586 5587 5588
	case BTRFS_IOC_SCRUB_CANCEL:
		return btrfs_ioctl_scrub_cancel(root, argp);
	case BTRFS_IOC_SCRUB_PROGRESS:
		return btrfs_ioctl_scrub_progress(root, argp);
5589
	case BTRFS_IOC_BALANCE_V2:
5590
		return btrfs_ioctl_balance(file, argp);
5591 5592
	case BTRFS_IOC_BALANCE_CTL:
		return btrfs_ioctl_balance_ctl(root, arg);
5593 5594
	case BTRFS_IOC_BALANCE_PROGRESS:
		return btrfs_ioctl_balance_progress(root, argp);
5595 5596
	case BTRFS_IOC_SET_RECEIVED_SUBVOL:
		return btrfs_ioctl_set_received_subvol(file, argp);
5597 5598 5599 5600
#ifdef CONFIG_64BIT
	case BTRFS_IOC_SET_RECEIVED_SUBVOL_32:
		return btrfs_ioctl_set_received_subvol_32(file, argp);
#endif
5601 5602
	case BTRFS_IOC_SEND:
		return btrfs_ioctl_send(file, argp);
5603
	case BTRFS_IOC_GET_DEV_STATS:
5604
		return btrfs_ioctl_get_dev_stats(root, argp);
A
Arne Jansen 已提交
5605
	case BTRFS_IOC_QUOTA_CTL:
5606
		return btrfs_ioctl_quota_ctl(file, argp);
A
Arne Jansen 已提交
5607
	case BTRFS_IOC_QGROUP_ASSIGN:
5608
		return btrfs_ioctl_qgroup_assign(file, argp);
A
Arne Jansen 已提交
5609
	case BTRFS_IOC_QGROUP_CREATE:
5610
		return btrfs_ioctl_qgroup_create(file, argp);
A
Arne Jansen 已提交
5611
	case BTRFS_IOC_QGROUP_LIMIT:
5612
		return btrfs_ioctl_qgroup_limit(file, argp);
J
Jan Schmidt 已提交
5613 5614 5615 5616
	case BTRFS_IOC_QUOTA_RESCAN:
		return btrfs_ioctl_quota_rescan(file, argp);
	case BTRFS_IOC_QUOTA_RESCAN_STATUS:
		return btrfs_ioctl_quota_rescan_status(file, argp);
5617 5618
	case BTRFS_IOC_QUOTA_RESCAN_WAIT:
		return btrfs_ioctl_quota_rescan_wait(file, argp);
5619 5620
	case BTRFS_IOC_DEV_REPLACE:
		return btrfs_ioctl_dev_replace(root, argp);
5621 5622
	case BTRFS_IOC_GET_FSLABEL:
		return btrfs_ioctl_get_fslabel(file, argp);
5623 5624
	case BTRFS_IOC_SET_FSLABEL:
		return btrfs_ioctl_set_fslabel(file, argp);
5625
	case BTRFS_IOC_GET_SUPPORTED_FEATURES:
5626
		return btrfs_ioctl_get_supported_features(argp);
5627 5628 5629 5630
	case BTRFS_IOC_GET_FEATURES:
		return btrfs_ioctl_get_features(file, argp);
	case BTRFS_IOC_SET_FEATURES:
		return btrfs_ioctl_set_features(file, argp);
C
Christoph Hellwig 已提交
5631 5632 5633 5634
	}

	return -ENOTTY;
}
5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655

#ifdef CONFIG_COMPAT
long btrfs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	switch (cmd) {
	case FS_IOC32_GETFLAGS:
		cmd = FS_IOC_GETFLAGS;
		break;
	case FS_IOC32_SETFLAGS:
		cmd = FS_IOC_SETFLAGS;
		break;
	case FS_IOC32_GETVERSION:
		cmd = FS_IOC_GETVERSION;
		break;
	default:
		return -ENOIOCTLCMD;
	}

	return btrfs_ioctl(file, cmd, (unsigned long) compat_ptr(arg));
}
#endif