copy_up.c 21.9 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0-only
Miklos Szeredi's avatar
Miklos Szeredi committed
2 3 4 5 6
/*
 *
 * Copyright (C) 2011 Novell Inc.
 */

7
#include <linux/module.h>
Miklos Szeredi's avatar
Miklos Szeredi committed
8 9 10 11 12 13 14
#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/file.h>
#include <linux/splice.h>
#include <linux/xattr.h>
#include <linux/security.h>
#include <linux/uaccess.h>
15
#include <linux/sched/signal.h>
16
#include <linux/cred.h>
Miklos Szeredi's avatar
Miklos Szeredi committed
17
#include <linux/namei.h>
18 19
#include <linux/fdtable.h>
#include <linux/ratelimit.h>
20
#include <linux/exportfs.h>
Miklos Szeredi's avatar
Miklos Szeredi committed
21 22 23 24
#include "overlayfs.h"

#define OVL_COPY_UP_CHUNK_SIZE (1 << 20)

25
static int ovl_ccup_set(const char *buf, const struct kernel_param *param)
26
{
lijiazi's avatar
lijiazi committed
27
	pr_warn("\"check_copy_up\" module option is obsolete\n");
28 29 30
	return 0;
}

31
static int ovl_ccup_get(char *buf, const struct kernel_param *param)
32
{
33
	return sprintf(buf, "N\n");
34 35
}

36
module_param_call(check_copy_up, ovl_ccup_set, ovl_ccup_get, NULL, 0644);
37
MODULE_PARM_DESC(check_copy_up, "Obsolete; does nothing");
38

39 40 41 42 43 44 45
static bool ovl_must_copy_xattr(const char *name)
{
	return !strcmp(name, XATTR_POSIX_ACL_ACCESS) ||
	       !strcmp(name, XATTR_POSIX_ACL_DEFAULT) ||
	       !strncmp(name, XATTR_SECURITY_PREFIX, XATTR_SECURITY_PREFIX_LEN);
}

Miklos Szeredi's avatar
Miklos Szeredi committed
46 47
int ovl_copy_xattr(struct dentry *old, struct dentry *new)
{
48 49 50
	ssize_t list_size, size, value_size = 0;
	char *buf, *name, *value = NULL;
	int uninitialized_var(error);
51
	size_t slen;
Miklos Szeredi's avatar
Miklos Szeredi committed
52

53 54
	if (!(old->d_inode->i_opflags & IOP_XATTR) ||
	    !(new->d_inode->i_opflags & IOP_XATTR))
Miklos Szeredi's avatar
Miklos Szeredi committed
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70
		return 0;

	list_size = vfs_listxattr(old, NULL, 0);
	if (list_size <= 0) {
		if (list_size == -EOPNOTSUPP)
			return 0;
		return list_size;
	}

	buf = kzalloc(list_size, GFP_KERNEL);
	if (!buf)
		return -ENOMEM;

	list_size = vfs_listxattr(old, buf, list_size);
	if (list_size <= 0) {
		error = list_size;
71
		goto out;
Miklos Szeredi's avatar
Miklos Szeredi committed
72 73
	}

74 75 76 77 78 79 80 81 82 83
	for (name = buf; list_size; name += slen) {
		slen = strnlen(name, list_size) + 1;

		/* underlying fs providing us with an broken xattr list? */
		if (WARN_ON(slen > list_size)) {
			error = -EIO;
			break;
		}
		list_size -= slen;

Miklos Szeredi's avatar
Miklos Szeredi committed
84 85
		if (ovl_is_private_xattr(name))
			continue;
86 87 88 89 90
retry:
		size = vfs_getxattr(old, name, value, value_size);
		if (size == -ERANGE)
			size = vfs_getxattr(old, name, NULL, 0);

Miklos Szeredi's avatar
Miklos Szeredi committed
91
		if (size < 0) {
Miklos Szeredi's avatar
Miklos Szeredi committed
92
			error = size;
93
			break;
Miklos Szeredi's avatar
Miklos Szeredi committed
94
		}
95 96 97 98 99 100 101 102 103 104 105 106 107 108

		if (size > value_size) {
			void *new;

			new = krealloc(value, size, GFP_KERNEL);
			if (!new) {
				error = -ENOMEM;
				break;
			}
			value = new;
			value_size = size;
			goto retry;
		}

109 110 111 112 113 114 115
		error = security_inode_copy_up_xattr(name);
		if (error < 0 && error != -EOPNOTSUPP)
			break;
		if (error == 1) {
			error = 0;
			continue; /* Discard */
		}
Miklos Szeredi's avatar
Miklos Szeredi committed
116
		error = vfs_setxattr(new, name, value, size, 0);
117 118 119 120 121 122 123
		if (error) {
			if (error != -EOPNOTSUPP || ovl_must_copy_xattr(name))
				break;

			/* Ignore failure to copy unknown xattrs */
			error = 0;
		}
Miklos Szeredi's avatar
Miklos Szeredi committed
124 125 126 127 128 129 130 131 132 133 134 135 136
	}
	kfree(value);
out:
	kfree(buf);
	return error;
}

static int ovl_copy_up_data(struct path *old, struct path *new, loff_t len)
{
	struct file *old_file;
	struct file *new_file;
	loff_t old_pos = 0;
	loff_t new_pos = 0;
137
	loff_t cloned;
138 139 140
	loff_t data_pos = -1;
	loff_t hole_len;
	bool skip_hole = false;
Miklos Szeredi's avatar
Miklos Szeredi committed
141 142 143 144 145
	int error = 0;

	if (len == 0)
		return 0;

146
	old_file = ovl_path_open(old, O_LARGEFILE | O_RDONLY);
Miklos Szeredi's avatar
Miklos Szeredi committed
147 148 149
	if (IS_ERR(old_file))
		return PTR_ERR(old_file);

150
	new_file = ovl_path_open(new, O_LARGEFILE | O_WRONLY);
Miklos Szeredi's avatar
Miklos Szeredi committed
151 152 153 154 155
	if (IS_ERR(new_file)) {
		error = PTR_ERR(new_file);
		goto out_fput;
	}

156
	/* Try to use clone_file_range to clone up within the same fs */
157
	cloned = do_clone_file_range(old_file, 0, new_file, 0, len, 0);
158
	if (cloned == len)
159 160 161
		goto out;
	/* Couldn't clone, so now we try to copy the data */

162 163 164 165 166
	/* Check if lower fs supports seek operation */
	if (old_file->f_mode & FMODE_LSEEK &&
	    old_file->f_op->llseek)
		skip_hole = true;

Miklos Szeredi's avatar
Miklos Szeredi committed
167 168 169 170 171 172 173 174 175 176 177 178
	while (len) {
		size_t this_len = OVL_COPY_UP_CHUNK_SIZE;
		long bytes;

		if (len < this_len)
			this_len = len;

		if (signal_pending_state(TASK_KILLABLE, current)) {
			error = -EINTR;
			break;
		}

179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208
		/*
		 * Fill zero for hole will cost unnecessary disk space
		 * and meanwhile slow down the copy-up speed, so we do
		 * an optimization for hole during copy-up, it relies
		 * on SEEK_DATA implementation in lower fs so if lower
		 * fs does not support it, copy-up will behave as before.
		 *
		 * Detail logic of hole detection as below:
		 * When we detect next data position is larger than current
		 * position we will skip that hole, otherwise we copy
		 * data in the size of OVL_COPY_UP_CHUNK_SIZE. Actually,
		 * it may not recognize all kind of holes and sometimes
		 * only skips partial of hole area. However, it will be
		 * enough for most of the use cases.
		 */

		if (skip_hole && data_pos < old_pos) {
			data_pos = vfs_llseek(old_file, old_pos, SEEK_DATA);
			if (data_pos > old_pos) {
				hole_len = data_pos - old_pos;
				len -= hole_len;
				old_pos = new_pos = data_pos;
				continue;
			} else if (data_pos == -ENXIO) {
				break;
			} else if (data_pos < 0) {
				skip_hole = false;
			}
		}

Miklos Szeredi's avatar
Miklos Szeredi committed
209 210 211 212 213 214 215 216 217 218 219
		bytes = do_splice_direct(old_file, &old_pos,
					 new_file, &new_pos,
					 this_len, SPLICE_F_MOVE);
		if (bytes <= 0) {
			error = bytes;
			break;
		}
		WARN_ON(old_pos != new_pos);

		len -= bytes;
	}
220
out:
Miklos Szeredi's avatar
Miklos Szeredi committed
221 222
	if (!error)
		error = vfs_fsync(new_file, 0);
Miklos Szeredi's avatar
Miklos Szeredi committed
223 224 225 226 227 228
	fput(new_file);
out_fput:
	fput(old_file);
	return error;
}

229 230 231 232 233 234 235 236 237 238
static int ovl_set_size(struct dentry *upperdentry, struct kstat *stat)
{
	struct iattr attr = {
		.ia_valid = ATTR_SIZE,
		.ia_size = stat->size,
	};

	return notify_change(upperdentry, &attr, NULL);
}

Miklos Szeredi's avatar
Miklos Szeredi committed
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
static int ovl_set_timestamps(struct dentry *upperdentry, struct kstat *stat)
{
	struct iattr attr = {
		.ia_valid =
		     ATTR_ATIME | ATTR_MTIME | ATTR_ATIME_SET | ATTR_MTIME_SET,
		.ia_atime = stat->atime,
		.ia_mtime = stat->mtime,
	};

	return notify_change(upperdentry, &attr, NULL);
}

int ovl_set_attr(struct dentry *upperdentry, struct kstat *stat)
{
	int err = 0;

	if (!S_ISLNK(stat->mode)) {
		struct iattr attr = {
			.ia_valid = ATTR_MODE,
			.ia_mode = stat->mode,
		};
		err = notify_change(upperdentry, &attr, NULL);
	}
	if (!err) {
		struct iattr attr = {
			.ia_valid = ATTR_UID | ATTR_GID,
			.ia_uid = stat->uid,
			.ia_gid = stat->gid,
		};
		err = notify_change(upperdentry, &attr, NULL);
	}
	if (!err)
		ovl_set_timestamps(upperdentry, stat);

	return err;
}

276
struct ovl_fh *ovl_encode_real_fh(struct dentry *real, bool is_upper)
277 278
{
	struct ovl_fh *fh;
279
	int fh_type, dwords;
280
	int buflen = MAX_HANDLE_SZ;
281
	uuid_t *uuid = &real->d_sb->s_uuid;
282
	int err;
283

284 285 286 287 288 289
	/* Make sure the real fid stays 32bit aligned */
	BUILD_BUG_ON(OVL_FH_FID_OFFSET % 4);
	BUILD_BUG_ON(MAX_HANDLE_SZ + OVL_FH_FID_OFFSET > 255);

	fh = kzalloc(buflen + OVL_FH_FID_OFFSET, GFP_KERNEL);
	if (!fh)
290 291 292 293 294 295 296 297
		return ERR_PTR(-ENOMEM);

	/*
	 * We encode a non-connectable file handle for non-dir, because we
	 * only need to find the lower inode number and we don't want to pay
	 * the price or reconnecting the dentry.
	 */
	dwords = buflen >> 2;
298
	fh_type = exportfs_encode_fh(real, (void *)fh->fb.fid, &dwords, 0);
299 300
	buflen = (dwords << 2);

301
	err = -EIO;
302 303 304
	if (WARN_ON(fh_type < 0) ||
	    WARN_ON(buflen > MAX_HANDLE_SZ) ||
	    WARN_ON(fh_type == FILEID_INVALID))
305
		goto out_err;
306

307 308 309 310
	fh->fb.version = OVL_FH_VERSION;
	fh->fb.magic = OVL_FH_MAGIC;
	fh->fb.type = fh_type;
	fh->fb.flags = OVL_FH_FLAG_CPU_ENDIAN;
311 312 313 314 315 316 317
	/*
	 * When we will want to decode an overlay dentry from this handle
	 * and all layers are on the same fs, if we get a disconncted real
	 * dentry when we decode fid, the only way to tell if we should assign
	 * it to upperdentry or to lowerstack is by checking this flag.
	 */
	if (is_upper)
318
		fh->fb.flags |= OVL_FH_FLAG_PATH_UPPER;
319
	fh->fb.len = sizeof(fh->fb) + buflen;
320
	fh->fb.uuid = *uuid;
321 322

	return fh;
323 324 325 326

out_err:
	kfree(fh);
	return ERR_PTR(err);
327 328
}

329 330
int ovl_set_origin(struct dentry *dentry, struct dentry *lower,
		   struct dentry *upper)
331 332 333 334 335 336 337 338 339
{
	const struct ovl_fh *fh = NULL;
	int err;

	/*
	 * When lower layer doesn't support export operations store a 'null' fh,
	 * so we can use the overlay.origin xattr to distignuish between a copy
	 * up and a pure upper inode.
	 */
340
	if (ovl_can_decode_fh(lower->d_sb)) {
341
		fh = ovl_encode_real_fh(lower, false);
342 343 344 345
		if (IS_ERR(fh))
			return PTR_ERR(fh);
	}

346 347 348
	/*
	 * Do not fail when upper doesn't support xattrs.
	 */
349 350
	err = ovl_check_setxattr(dentry, upper, OVL_XATTR_ORIGIN, fh->buf,
				 fh ? fh->fb.len : 0, 0);
351 352 353 354 355
	kfree(fh);

	return err;
}

356 357 358 359 360 361
/* Store file handle of @upper dir in @index dir entry */
static int ovl_set_upper_fh(struct dentry *upper, struct dentry *index)
{
	const struct ovl_fh *fh;
	int err;

362
	fh = ovl_encode_real_fh(upper, true);
363 364 365
	if (IS_ERR(fh))
		return PTR_ERR(fh);

366
	err = ovl_do_setxattr(index, OVL_XATTR_UPPER, fh->buf, fh->fb.len, 0);
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405

	kfree(fh);
	return err;
}

/*
 * Create and install index entry.
 *
 * Caller must hold i_mutex on indexdir.
 */
static int ovl_create_index(struct dentry *dentry, struct dentry *origin,
			    struct dentry *upper)
{
	struct dentry *indexdir = ovl_indexdir(dentry->d_sb);
	struct inode *dir = d_inode(indexdir);
	struct dentry *index = NULL;
	struct dentry *temp = NULL;
	struct qstr name = { };
	int err;

	/*
	 * For now this is only used for creating index entry for directories,
	 * because non-dir are copied up directly to index and then hardlinked
	 * to upper dir.
	 *
	 * TODO: implement create index for non-dir, so we can call it when
	 * encoding file handle for non-dir in case index does not exist.
	 */
	if (WARN_ON(!d_is_dir(dentry)))
		return -EIO;

	/* Directory not expected to be indexed before copy up */
	if (WARN_ON(ovl_test_flag(OVL_INDEX, d_inode(dentry))))
		return -EIO;

	err = ovl_get_index_name(origin, &name);
	if (err)
		return err;

406
	temp = ovl_create_temp(indexdir, OVL_CATTR(S_IFDIR | 0));
407
	err = PTR_ERR(temp);
408
	if (IS_ERR(temp))
409
		goto free_name;
410 411 412

	err = ovl_set_upper_fh(upper, temp);
	if (err)
413
		goto out;
414 415 416 417 418 419 420 421 422

	index = lookup_one_len(name.name, indexdir, name.len);
	if (IS_ERR(index)) {
		err = PTR_ERR(index);
	} else {
		err = ovl_do_rename(dir, temp, dir, index, 0);
		dput(index);
	}
out:
423 424
	if (err)
		ovl_cleanup(dir, temp);
425
	dput(temp);
426
free_name:
427 428 429 430
	kfree(name.name);
	return err;
}

431 432 433 434 435 436 437 438 439 440 441
struct ovl_copy_up_ctx {
	struct dentry *parent;
	struct dentry *dentry;
	struct path lowerpath;
	struct kstat stat;
	struct kstat pstat;
	const char *link;
	struct dentry *destdir;
	struct qstr destname;
	struct dentry *workdir;
	bool origin;
442
	bool indexed;
443
	bool metacopy;
444 445 446
};

static int ovl_link_up(struct ovl_copy_up_ctx *c)
447 448 449
{
	int err;
	struct dentry *upper;
450
	struct dentry *upperdir = ovl_dentry_upper(c->parent);
451 452
	struct inode *udir = d_inode(upperdir);

453 454 455 456 457 458
	/* Mark parent "impure" because it may now contain non-pure upper */
	err = ovl_set_impure(c->parent, upperdir);
	if (err)
		return err;

	err = ovl_set_nlink_lower(c->dentry);
459 460 461
	if (err)
		return err;

462
	inode_lock_nested(udir, I_MUTEX_PARENT);
463 464
	upper = lookup_one_len(c->dentry->d_name.name, upperdir,
			       c->dentry->d_name.len);
465 466
	err = PTR_ERR(upper);
	if (!IS_ERR(upper)) {
467
		err = ovl_do_link(ovl_dentry_upper(c->dentry), udir, upper);
468 469
		dput(upper);

470 471 472 473 474
		if (!err) {
			/* Restore timestamps on parent (best effort) */
			ovl_set_timestamps(upperdir, &c->pstat);
			ovl_dentry_set_upper_alias(c->dentry);
		}
475 476
	}
	inode_unlock(udir);
477 478 479 480
	if (err)
		return err;

	err = ovl_set_nlink_upper(c->dentry);
481 482 483 484

	return err;
}

485
static int ovl_copy_up_inode(struct ovl_copy_up_ctx *c, struct dentry *temp)
486 487 488
{
	int err;

489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506
	/*
	 * Copy up data first and then xattrs. Writing data after
	 * xattrs will remove security.capability xattr automatically.
	 */
	if (S_ISREG(c->stat.mode) && !c->metacopy) {
		struct path upperpath, datapath;

		ovl_path_upper(c->dentry, &upperpath);
		if (WARN_ON(upperpath.dentry != NULL))
			return -EIO;
		upperpath.dentry = temp;

		ovl_path_lowerdata(c->dentry, &datapath);
		err = ovl_copy_up_data(&datapath, &upperpath, c->stat.size);
		if (err)
			return err;
	}

507
	err = ovl_copy_xattr(c->lowerpath.dentry, temp);
Miklos Szeredi's avatar
Miklos Szeredi committed
508
	if (err)
509
		return err;
Miklos Szeredi's avatar
Miklos Szeredi committed
510

511 512 513
	/*
	 * Store identifier of lower inode in upper inode xattr to
	 * allow lookup of the copy up origin inode.
514 515 516
	 *
	 * Don't set origin when we are breaking the association with a lower
	 * hard link.
517
	 */
518
	if (c->origin) {
519
		err = ovl_set_origin(c->dentry, c->lowerpath.dentry, temp);
520
		if (err)
521
			return err;
522
	}
523

524 525 526 527 528 529 530
	if (c->metacopy) {
		err = ovl_check_setxattr(c->dentry, temp, OVL_XATTR_METACOPY,
					 NULL, 0, -EOPNOTSUPP);
		if (err)
			return err;
	}

531
	inode_lock(temp->d_inode);
532
	if (S_ISREG(c->stat.mode))
533 534 535
		err = ovl_set_size(temp, &c->stat);
	if (!err)
		err = ovl_set_attr(temp, &c->stat);
536 537 538
	inode_unlock(temp->d_inode);

	return err;
539 540
}

541 542 543 544 545 546
struct ovl_cu_creds {
	const struct cred *old;
	struct cred *new;
};

static int ovl_prep_cu_creds(struct dentry *dentry, struct ovl_cu_creds *cc)
547 548 549
{
	int err;

550 551
	cc->old = cc->new = NULL;
	err = security_inode_copy_up(dentry, &cc->new);
552
	if (err < 0)
553
		return err;
554

555 556
	if (cc->new)
		cc->old = override_creds(cc->new);
557

558
	return 0;
559 560
}

561
static void ovl_revert_cu_creds(struct ovl_cu_creds *cc)
562
{
563 564 565 566
	if (cc->new) {
		revert_creds(cc->old);
		put_cred(cc->new);
	}
567 568 569 570 571 572 573
}

/*
 * Copyup using workdir to prepare temp file.  Used when copying up directories,
 * special files or when upper fs doesn't support O_TMPFILE.
 */
static int ovl_copy_up_workdir(struct ovl_copy_up_ctx *c)
574
{
575
	struct inode *inode;
576 577 578
	struct inode *udir = d_inode(c->destdir), *wdir = d_inode(c->workdir);
	struct dentry *temp, *upper;
	struct ovl_cu_creds cc;
579
	int err;
580 581 582 583 584 585
	struct ovl_cattr cattr = {
		/* Can't properly set mode on creation because of the umask */
		.mode = c->stat.mode & S_IFMT,
		.rdev = c->stat.rdev,
		.link = c->link
	};
586

587 588 589 590
	err = ovl_lock_rename_workdir(c->workdir, c->destdir);
	if (err)
		return err;

591 592 593 594 595 596 597
	err = ovl_prep_cu_creds(c->dentry, &cc);
	if (err)
		goto unlock;

	temp = ovl_create_temp(c->workdir, &cattr);
	ovl_revert_cu_creds(&cc);

598
	err = PTR_ERR(temp);
599
	if (IS_ERR(temp))
600
		goto unlock;
601

602
	err = ovl_copy_up_inode(c, temp);
603
	if (err)
604
		goto cleanup;
605

606 607 608
	if (S_ISDIR(c->stat.mode) && c->indexed) {
		err = ovl_create_index(c->dentry, c->lowerpath.dentry, temp);
		if (err)
609
			goto cleanup;
610 611
	}

612 613 614 615 616 617 618
	upper = lookup_one_len(c->destname.name, c->destdir, c->destname.len);
	err = PTR_ERR(upper);
	if (IS_ERR(upper))
		goto cleanup;

	err = ovl_do_rename(wdir, temp, udir, upper, 0);
	dput(upper);
Miklos Szeredi's avatar
Miklos Szeredi committed
619
	if (err)
620
		goto cleanup;
Miklos Szeredi's avatar
Miklos Szeredi committed
621

622 623
	if (!c->metacopy)
		ovl_set_upperdata(d_inode(c->dentry));
624
	inode = d_inode(c->dentry);
625
	ovl_inode_update(inode, temp);
626 627
	if (S_ISDIR(inode->i_mode))
		ovl_set_flag(OVL_WHITEOUTS, inode);
628 629 630 631 632 633
unlock:
	unlock_rename(c->workdir, c->destdir);

	return err;

cleanup:
634 635 636
	ovl_cleanup(wdir, temp);
	dput(temp);
	goto unlock;
637 638
}

639 640
/* Copyup using O_TMPFILE which does not require cross dir locking */
static int ovl_copy_up_tmpfile(struct ovl_copy_up_ctx *c)
641
{
642 643 644
	struct inode *udir = d_inode(c->destdir);
	struct dentry *temp, *upper;
	struct ovl_cu_creds cc;
645 646
	int err;

647 648 649
	err = ovl_prep_cu_creds(c->dentry, &cc);
	if (err)
		return err;
650 651

	temp = ovl_do_tmpfile(c->workdir, c->stat.mode);
652
	ovl_revert_cu_creds(&cc);
653

654 655
	if (IS_ERR(temp))
		return PTR_ERR(temp);
656

657 658 659
	err = ovl_copy_up_inode(c, temp);
	if (err)
		goto out_dput;
660 661 662 663 664

	inode_lock_nested(udir, I_MUTEX_PARENT);

	upper = lookup_one_len(c->destname.name, c->destdir, c->destname.len);
	err = PTR_ERR(upper);
665 666 667 668
	if (!IS_ERR(upper)) {
		err = ovl_do_link(temp, udir, upper);
		dput(upper);
	}
669 670 671
	inode_unlock(udir);

	if (err)
672
		goto out_dput;
673 674 675

	if (!c->metacopy)
		ovl_set_upperdata(d_inode(c->dentry));
676
	ovl_inode_update(d_inode(c->dentry), temp);
677

678 679 680
	return 0;

out_dput:
681
	dput(temp);
Miklos Szeredi's avatar
Miklos Szeredi committed
682 683 684 685 686 687
	return err;
}

/*
 * Copy up a single dentry
 *
Miklos Szeredi's avatar
Miklos Szeredi committed
688 689 690 691 692
 * All renames start with copy up of source if necessary.  The actual
 * rename will only proceed once the copy up was successful.  Copy up uses
 * upper parent i_mutex for exclusion.  Since rename can change d_parent it
 * is possible that the copy up will lock the old parent.  At that point
 * the file will have already been copied up anyway.
Miklos Szeredi's avatar
Miklos Szeredi committed
693
 */
Miklos Szeredi's avatar
Miklos Szeredi committed
694
static int ovl_do_copy_up(struct ovl_copy_up_ctx *c)
Miklos Szeredi's avatar
Miklos Szeredi committed
695 696
{
	int err;
697
	struct ovl_fs *ofs = c->dentry->d_sb->s_fs_info;
698
	bool to_index = false;
699

700 701 702 703 704 705 706 707 708 709 710 711 712 713 714
	/*
	 * Indexed non-dir is copied up directly to the index entry and then
	 * hardlinked to upper dir. Indexed dir is copied up to indexdir,
	 * then index entry is created and then copied up dir installed.
	 * Copying dir up to indexdir instead of workdir simplifies locking.
	 */
	if (ovl_need_index(c->dentry)) {
		c->indexed = true;
		if (S_ISDIR(c->stat.mode))
			c->workdir = ovl_indexdir(c->dentry->d_sb);
		else
			to_index = true;
	}

	if (S_ISDIR(c->stat.mode) || c->stat.nlink == 1 || to_index)
715 716
		c->origin = true;

717
	if (to_index) {
718 719 720 721
		c->destdir = ovl_indexdir(c->dentry->d_sb);
		err = ovl_get_index_name(c->lowerpath.dentry, &c->destname);
		if (err)
			return err;
722 723 724
	} else if (WARN_ON(!c->parent)) {
		/* Disconnected dentry must be copied up to index dir */
		return -EIO;
725 726 727 728 729 730 731 732 733
	} else {
		/*
		 * Mark parent "impure" because it may now contain non-pure
		 * upper
		 */
		err = ovl_set_impure(c->parent, c->destdir);
		if (err)
			return err;
	}
Miklos Szeredi's avatar
Miklos Szeredi committed
734

735
	/* Should we copyup with O_TMPFILE or with workdir? */
736 737 738 739
	if (S_ISREG(c->stat.mode) && ofs->tmpfile)
		err = ovl_copy_up_tmpfile(c);
	else
		err = ovl_copy_up_workdir(c);
740 741 742 743
	if (err)
		goto out;

	if (c->indexed)
744 745 746
		ovl_set_flag(OVL_INDEX, d_inode(c->dentry));

	if (to_index) {
747 748 749
		/* Initialize nlink for copy up of disconnected dentry */
		err = ovl_set_nlink_upper(c->dentry);
	} else {
750 751 752 753 754 755 756 757
		struct inode *udir = d_inode(c->destdir);

		/* Restore timestamps on parent (best effort) */
		inode_lock(udir);
		ovl_set_timestamps(c->destdir, &c->pstat);
		inode_unlock(udir);

		ovl_dentry_set_upper_alias(c->dentry);
Miklos Szeredi's avatar
Miklos Szeredi committed
758 759
	}

760 761 762
out:
	if (to_index)
		kfree(c->destname.name);
Miklos Szeredi's avatar
Miklos Szeredi committed
763 764 765
	return err;
}

766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
static bool ovl_need_meta_copy_up(struct dentry *dentry, umode_t mode,
				  int flags)
{
	struct ovl_fs *ofs = dentry->d_sb->s_fs_info;

	if (!ofs->config.metacopy)
		return false;

	if (!S_ISREG(mode))
		return false;

	if (flags && ((OPEN_FMODE(flags) & FMODE_WRITE) || (flags & O_TRUNC)))
		return false;

	return true;
}

783 784 785
/* Copy up data of an inode which was copied up metadata only in the past. */
static int ovl_copy_up_meta_inode_data(struct ovl_copy_up_ctx *c)
{
786
	struct path upperpath, datapath;
787
	int err;
788 789
	char *capability = NULL;
	ssize_t uninitialized_var(cap_size);
790 791 792 793 794

	ovl_path_upper(c->dentry, &upperpath);
	if (WARN_ON(upperpath.dentry == NULL))
		return -EIO;

795 796 797 798
	ovl_path_lowerdata(c->dentry, &datapath);
	if (WARN_ON(datapath.dentry == NULL))
		return -EIO;

799 800 801 802 803 804 805
	if (c->stat.size) {
		err = cap_size = ovl_getxattr(upperpath.dentry, XATTR_NAME_CAPS,
					      &capability, 0);
		if (err < 0 && err != -ENODATA)
			goto out;
	}

806
	err = ovl_copy_up_data(&datapath, &upperpath, c->stat.size);
807
	if (err)
808 809 810 811 812 813 814 815 816 817 818 819 820
		goto out_free;

	/*
	 * Writing to upper file will clear security.capability xattr. We
	 * don't want that to happen for normal copy-up operation.
	 */
	if (capability) {
		err = ovl_do_setxattr(upperpath.dentry, XATTR_NAME_CAPS,
				      capability, cap_size, 0);
		if (err)
			goto out_free;
	}

821 822 823

	err = vfs_removexattr(upperpath.dentry, OVL_XATTR_METACOPY);
	if (err)
824
		goto out_free;
825 826

	ovl_set_upperdata(d_inode(c->dentry));
827 828 829
out_free:
	kfree(capability);
out:
830 831 832
	return err;
}

Miklos Szeredi's avatar
Miklos Szeredi committed
833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853
static int ovl_copy_up_one(struct dentry *parent, struct dentry *dentry,
			   int flags)
{
	int err;
	DEFINE_DELAYED_CALL(done);
	struct path parentpath;
	struct ovl_copy_up_ctx ctx = {
		.parent = parent,
		.dentry = dentry,
		.workdir = ovl_workdir(dentry),
	};

	if (WARN_ON(!ctx.workdir))
		return -EROFS;

	ovl_path_lower(dentry, &ctx.lowerpath);
	err = vfs_getattr(&ctx.lowerpath, &ctx.stat,
			  STATX_BASIC_STATS, AT_STATX_SYNC_AS_STAT);
	if (err)
		return err;

854 855
	ctx.metacopy = ovl_need_meta_copy_up(dentry, ctx.stat.mode, flags);

856 857 858 859
	if (parent) {
		ovl_path_upper(parent, &parentpath);
		ctx.destdir = parentpath.dentry;
		ctx.destname = dentry->d_name;
Miklos Szeredi's avatar
Miklos Szeredi committed
860

861 862 863 864 865 866
		err = vfs_getattr(&parentpath, &ctx.pstat,
				  STATX_ATIME | STATX_MTIME,
				  AT_STATX_SYNC_AS_STAT);
		if (err)
			return err;
	}
Miklos Szeredi's avatar
Miklos Szeredi committed
867 868 869 870 871 872 873 874 875 876 877

	/* maybe truncate regular file. this has no effect on dirs */
	if (flags & O_TRUNC)
		ctx.stat.size = 0;

	if (S_ISLNK(ctx.stat.mode)) {
		ctx.link = vfs_get_link(ctx.lowerpath.dentry, &done);
		if (IS_ERR(ctx.link))
			return PTR_ERR(ctx.link);
	}

878
	err = ovl_copy_up_start(dentry, flags);
Miklos Szeredi's avatar
Miklos Szeredi committed
879 880 881 882 883
	/* err < 0: interrupted, err > 0: raced with another copy-up */
	if (unlikely(err)) {
		if (err > 0)
			err = 0;
	} else {
884 885
		if (!ovl_dentry_upper(dentry))
			err = ovl_do_copy_up(&ctx);
886
		if (!err && parent && !ovl_dentry_has_upper_alias(dentry))
887
			err = ovl_link_up(&ctx);
888 889
		if (!err && ovl_dentry_needs_data_copy_up_locked(dentry, flags))
			err = ovl_copy_up_meta_inode_data(&ctx);
Miklos Szeredi's avatar
Miklos Szeredi committed
890 891
		ovl_copy_up_end(dentry);
	}
Miklos Szeredi's avatar
Miklos Szeredi committed
892
	do_delayed_call(&done);
Miklos Szeredi's avatar
Miklos Szeredi committed
893 894 895 896

	return err;
}

897
int ovl_copy_up_flags(struct dentry *dentry, int flags)
Miklos Szeredi's avatar
Miklos Szeredi committed
898
{
899 900
	int err = 0;
	const struct cred *old_cred = ovl_override_creds(dentry->d_sb);
901 902 903 904 905 906 907 908 909
	bool disconnected = (dentry->d_flags & DCACHE_DISCONNECTED);

	/*
	 * With NFS export, copy up can get called for a disconnected non-dir.
	 * In this case, we will copy up lower inode to index dir without
	 * linking it to upper dir.
	 */
	if (WARN_ON(disconnected && d_is_dir(dentry)))
		return -EIO;
Miklos Szeredi's avatar
Miklos Szeredi committed
910 911 912

	while (!err) {
		struct dentry *next;
913
		struct dentry *parent = NULL;
Miklos Szeredi's avatar
Miklos Szeredi committed
914

915
		if (ovl_already_copied_up(dentry, flags))
Miklos Szeredi's avatar
Miklos Szeredi committed
916 917 918 919
			break;

		next = dget(dentry);
		/* find the topmost dentry not yet copied up */
920
		for (; !disconnected;) {
Miklos Szeredi's avatar
Miklos Szeredi committed
921 922
			parent = dget_parent(next);

923
			if (ovl_dentry_upper(parent))
Miklos Szeredi's avatar
Miklos Szeredi committed
924 925 926 927 928 929
				break;

			dput(next);
			next = parent;
		}

Miklos Szeredi's avatar
Miklos Szeredi committed
930
		err = ovl_copy_up_one(parent, next, flags);
Miklos Szeredi's avatar
Miklos Szeredi committed
931 932 933 934

		dput(parent);
		dput(next);
	}
935
	revert_creds(old_cred);
Miklos Szeredi's avatar
Miklos Szeredi committed
936 937 938

	return err;
}
939

940 941 942
static bool ovl_open_need_copy_up(struct dentry *dentry, int flags)
{
	/* Copy up of disconnected dentry does not set upper alias */
943
	if (ovl_already_copied_up(dentry, flags))
944 945 946 947 948
		return false;

	if (special_file(d_inode(dentry)->i_mode))
		return false;

949
	if (!ovl_open_flags_need_copy_up(flags))
950 951 952 953 954
		return false;

	return true;
}

955
int ovl_maybe_copy_up(struct dentry *dentry, int flags)
956 957 958
{
	int err = 0;

959
	if (ovl_open_need_copy_up(dentry, flags)) {
960 961
		err = ovl_want_write(dentry);
		if (!err) {
962
			err = ovl_copy_up_flags(dentry, flags);
963 964 965 966 967 968 969
			ovl_drop_write(dentry);
		}
	}

	return err;
}

970 971 972 973 974
int ovl_copy_up_with_data(struct dentry *dentry)
{
	return ovl_copy_up_flags(dentry, O_WRONLY);
}

975 976 977 978
int ovl_copy_up(struct dentry *dentry)
{
	return ovl_copy_up_flags(dentry, 0);
}