copy_up.c 20.4 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0-only
M
Miklos Szeredi 已提交
2 3 4 5 6
/*
 *
 * Copyright (C) 2011 Novell Inc.
 */

7
#include <linux/module.h>
M
Miklos Szeredi 已提交
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>
M
Miklos Szeredi 已提交
17
#include <linux/namei.h>
18 19
#include <linux/fdtable.h>
#include <linux/ratelimit.h>
20
#include <linux/exportfs.h>
M
Miklos Szeredi 已提交
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
{
27
	pr_warn("overlayfs: \"check_copy_up\" module option is obsolete\n");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

M
Miklos Szeredi 已提交
39 40
int ovl_copy_xattr(struct dentry *old, struct dentry *new)
{
41 42 43
	ssize_t list_size, size, value_size = 0;
	char *buf, *name, *value = NULL;
	int uninitialized_var(error);
44
	size_t slen;
M
Miklos Szeredi 已提交
45

46 47
	if (!(old->d_inode->i_opflags & IOP_XATTR) ||
	    !(new->d_inode->i_opflags & IOP_XATTR))
M
Miklos Szeredi 已提交
48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63
		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;
64
		goto out;
M
Miklos Szeredi 已提交
65 66
	}

67 68 69 70 71 72 73 74 75 76
	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;

M
Miklos Szeredi 已提交
77 78
		if (ovl_is_private_xattr(name))
			continue;
79 80 81 82 83
retry:
		size = vfs_getxattr(old, name, value, value_size);
		if (size == -ERANGE)
			size = vfs_getxattr(old, name, NULL, 0);

M
Miklos Szeredi 已提交
84
		if (size < 0) {
M
Miklos Szeredi 已提交
85
			error = size;
86
			break;
M
Miklos Szeredi 已提交
87
		}
88 89 90 91 92 93 94 95 96 97 98 99 100 101

		if (size > value_size) {
			void *new;

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

102 103 104 105 106 107 108
		error = security_inode_copy_up_xattr(name);
		if (error < 0 && error != -EOPNOTSUPP)
			break;
		if (error == 1) {
			error = 0;
			continue; /* Discard */
		}
M
Miklos Szeredi 已提交
109 110
		error = vfs_setxattr(new, name, value, size, 0);
		if (error)
111
			break;
M
Miklos Szeredi 已提交
112 113 114 115 116 117 118 119 120 121 122 123 124
	}
	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;
125
	loff_t cloned;
M
Miklos Szeredi 已提交
126 127 128 129 130
	int error = 0;

	if (len == 0)
		return 0;

131
	old_file = ovl_path_open(old, O_LARGEFILE | O_RDONLY);
M
Miklos Szeredi 已提交
132 133 134
	if (IS_ERR(old_file))
		return PTR_ERR(old_file);

135
	new_file = ovl_path_open(new, O_LARGEFILE | O_WRONLY);
M
Miklos Szeredi 已提交
136 137 138 139 140
	if (IS_ERR(new_file)) {
		error = PTR_ERR(new_file);
		goto out_fput;
	}

141
	/* Try to use clone_file_range to clone up within the same fs */
142
	cloned = do_clone_file_range(old_file, 0, new_file, 0, len, 0);
143
	if (cloned == len)
144 145 146
		goto out;
	/* Couldn't clone, so now we try to copy the data */

M
Miklos Szeredi 已提交
147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
	/* FIXME: copy up sparse files efficiently */
	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;
		}

		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;
	}
171
out:
M
Miklos Szeredi 已提交
172 173
	if (!error)
		error = vfs_fsync(new_file, 0);
M
Miklos Szeredi 已提交
174 175 176 177 178 179
	fput(new_file);
out_fput:
	fput(old_file);
	return error;
}

180 181 182 183 184 185 186 187 188 189
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);
}

M
Miklos Szeredi 已提交
190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
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;
}

227
struct ovl_fh *ovl_encode_real_fh(struct dentry *real, bool is_upper)
228 229
{
	struct ovl_fh *fh;
230
	int fh_type, dwords;
231
	int buflen = MAX_HANDLE_SZ;
232
	uuid_t *uuid = &real->d_sb->s_uuid;
233
	int err;
234

235 236 237 238 239 240
	/* 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)
241 242 243 244 245 246 247 248
		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;
249
	fh_type = exportfs_encode_fh(real, (void *)fh->fb.fid, &dwords, 0);
250 251
	buflen = (dwords << 2);

252
	err = -EIO;
253 254 255
	if (WARN_ON(fh_type < 0) ||
	    WARN_ON(buflen > MAX_HANDLE_SZ) ||
	    WARN_ON(fh_type == FILEID_INVALID))
256
		goto out_err;
257

258 259 260 261
	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;
262 263 264 265 266 267 268
	/*
	 * 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)
269
		fh->fb.flags |= OVL_FH_FLAG_PATH_UPPER;
270
	fh->fb.len = sizeof(fh->fb) + buflen;
271
	fh->fb.uuid = *uuid;
272 273

	return fh;
274 275 276 277

out_err:
	kfree(fh);
	return ERR_PTR(err);
278 279
}

280 281
int ovl_set_origin(struct dentry *dentry, struct dentry *lower,
		   struct dentry *upper)
282 283 284 285 286 287 288 289 290
{
	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.
	 */
291
	if (ovl_can_decode_fh(lower->d_sb)) {
292
		fh = ovl_encode_real_fh(lower, false);
293 294 295 296
		if (IS_ERR(fh))
			return PTR_ERR(fh);
	}

297 298 299
	/*
	 * Do not fail when upper doesn't support xattrs.
	 */
300 301
	err = ovl_check_setxattr(dentry, upper, OVL_XATTR_ORIGIN, fh->buf,
				 fh ? fh->fb.len : 0, 0);
302 303 304 305 306
	kfree(fh);

	return err;
}

307 308 309 310 311 312
/* 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;

313
	fh = ovl_encode_real_fh(upper, true);
314 315 316
	if (IS_ERR(fh))
		return PTR_ERR(fh);

317
	err = ovl_do_setxattr(index, OVL_XATTR_UPPER, fh->buf, fh->fb.len, 0);
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356

	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;

357
	temp = ovl_create_temp(indexdir, OVL_CATTR(S_IFDIR | 0));
358
	err = PTR_ERR(temp);
359
	if (IS_ERR(temp))
360
		goto free_name;
361 362 363

	err = ovl_set_upper_fh(upper, temp);
	if (err)
364
		goto out;
365 366 367 368 369 370 371 372 373

	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:
374 375
	if (err)
		ovl_cleanup(dir, temp);
376
	dput(temp);
377
free_name:
378 379 380 381
	kfree(name.name);
	return err;
}

382 383 384 385 386 387 388 389 390 391 392
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;
393
	bool indexed;
394
	bool metacopy;
395 396 397
};

static int ovl_link_up(struct ovl_copy_up_ctx *c)
398 399 400
{
	int err;
	struct dentry *upper;
401
	struct dentry *upperdir = ovl_dentry_upper(c->parent);
402 403
	struct inode *udir = d_inode(upperdir);

404 405 406 407 408 409
	/* 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);
410 411 412
	if (err)
		return err;

413
	inode_lock_nested(udir, I_MUTEX_PARENT);
414 415
	upper = lookup_one_len(c->dentry->d_name.name, upperdir,
			       c->dentry->d_name.len);
416 417
	err = PTR_ERR(upper);
	if (!IS_ERR(upper)) {
418
		err = ovl_do_link(ovl_dentry_upper(c->dentry), udir, upper);
419 420
		dput(upper);

421 422 423 424 425
		if (!err) {
			/* Restore timestamps on parent (best effort) */
			ovl_set_timestamps(upperdir, &c->pstat);
			ovl_dentry_set_upper_alias(c->dentry);
		}
426 427
	}
	inode_unlock(udir);
428 429 430 431
	if (err)
		return err;

	err = ovl_set_nlink_upper(c->dentry);
432 433 434 435

	return err;
}

436
static int ovl_copy_up_inode(struct ovl_copy_up_ctx *c, struct dentry *temp)
437 438 439
{
	int err;

440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
	/*
	 * 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;
	}

458
	err = ovl_copy_xattr(c->lowerpath.dentry, temp);
M
Miklos Szeredi 已提交
459
	if (err)
460
		return err;
M
Miklos Szeredi 已提交
461

462 463 464
	/*
	 * Store identifier of lower inode in upper inode xattr to
	 * allow lookup of the copy up origin inode.
465 466 467
	 *
	 * Don't set origin when we are breaking the association with a lower
	 * hard link.
468
	 */
469
	if (c->origin) {
470
		err = ovl_set_origin(c->dentry, c->lowerpath.dentry, temp);
471
		if (err)
472
			return err;
473
	}
474

475 476 477 478 479 480 481
	if (c->metacopy) {
		err = ovl_check_setxattr(c->dentry, temp, OVL_XATTR_METACOPY,
					 NULL, 0, -EOPNOTSUPP);
		if (err)
			return err;
	}

482
	inode_lock(temp->d_inode);
483 484 485 486
	if (c->metacopy)
		err = ovl_set_size(temp, &c->stat);
	if (!err)
		err = ovl_set_attr(temp, &c->stat);
487 488 489
	inode_unlock(temp->d_inode);

	return err;
490 491
}

492 493 494 495 496 497
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)
A
Amir Goldstein 已提交
498 499 500
{
	int err;

501 502
	cc->old = cc->new = NULL;
	err = security_inode_copy_up(dentry, &cc->new);
A
Amir Goldstein 已提交
503
	if (err < 0)
504
		return err;
A
Amir Goldstein 已提交
505

506 507
	if (cc->new)
		cc->old = override_creds(cc->new);
A
Amir Goldstein 已提交
508

509
	return 0;
A
Amir Goldstein 已提交
510 511
}

512
static void ovl_revert_cu_creds(struct ovl_cu_creds *cc)
A
Amir Goldstein 已提交
513
{
514 515 516 517
	if (cc->new) {
		revert_creds(cc->old);
		put_cred(cc->new);
	}
A
Amir Goldstein 已提交
518 519 520 521 522 523 524
}

/*
 * 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)
525
{
526
	struct inode *inode;
527 528 529
	struct inode *udir = d_inode(c->destdir), *wdir = d_inode(c->workdir);
	struct dentry *temp, *upper;
	struct ovl_cu_creds cc;
530
	int err;
531 532 533 534 535 536
	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
	};
537

A
Amir Goldstein 已提交
538 539 540 541
	err = ovl_lock_rename_workdir(c->workdir, c->destdir);
	if (err)
		return err;

542 543 544 545 546 547 548
	err = ovl_prep_cu_creds(c->dentry, &cc);
	if (err)
		goto unlock;

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

A
Amir Goldstein 已提交
549
	err = PTR_ERR(temp);
550
	if (IS_ERR(temp))
A
Amir Goldstein 已提交
551
		goto unlock;
552

553
	err = ovl_copy_up_inode(c, temp);
554
	if (err)
A
Amir Goldstein 已提交
555
		goto cleanup;
556

557 558 559
	if (S_ISDIR(c->stat.mode) && c->indexed) {
		err = ovl_create_index(c->dentry, c->lowerpath.dentry, temp);
		if (err)
A
Amir Goldstein 已提交
560
			goto cleanup;
561 562
	}

563 564 565 566 567 568 569
	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);
M
Miklos Szeredi 已提交
570
	if (err)
A
Amir Goldstein 已提交
571
		goto cleanup;
M
Miklos Szeredi 已提交
572

573 574
	if (!c->metacopy)
		ovl_set_upperdata(d_inode(c->dentry));
575
	inode = d_inode(c->dentry);
576
	ovl_inode_update(inode, temp);
577 578
	if (S_ISDIR(inode->i_mode))
		ovl_set_flag(OVL_WHITEOUTS, inode);
A
Amir Goldstein 已提交
579 580 581 582 583 584
unlock:
	unlock_rename(c->workdir, c->destdir);

	return err;

cleanup:
585 586 587
	ovl_cleanup(wdir, temp);
	dput(temp);
	goto unlock;
A
Amir Goldstein 已提交
588 589
}

590 591
/* Copyup using O_TMPFILE which does not require cross dir locking */
static int ovl_copy_up_tmpfile(struct ovl_copy_up_ctx *c)
A
Amir Goldstein 已提交
592
{
593 594 595
	struct inode *udir = d_inode(c->destdir);
	struct dentry *temp, *upper;
	struct ovl_cu_creds cc;
A
Amir Goldstein 已提交
596 597
	int err;

598 599 600
	err = ovl_prep_cu_creds(c->dentry, &cc);
	if (err)
		return err;
A
Amir Goldstein 已提交
601 602

	temp = ovl_do_tmpfile(c->workdir, c->stat.mode);
603
	ovl_revert_cu_creds(&cc);
A
Amir Goldstein 已提交
604

605 606
	if (IS_ERR(temp))
		return PTR_ERR(temp);
A
Amir Goldstein 已提交
607

608 609 610
	err = ovl_copy_up_inode(c, temp);
	if (err)
		goto out_dput;
A
Amir Goldstein 已提交
611 612 613 614 615

	inode_lock_nested(udir, I_MUTEX_PARENT);

	upper = lookup_one_len(c->destname.name, c->destdir, c->destname.len);
	err = PTR_ERR(upper);
616 617 618 619
	if (!IS_ERR(upper)) {
		err = ovl_do_link(temp, udir, upper);
		dput(upper);
	}
A
Amir Goldstein 已提交
620 621 622
	inode_unlock(udir);

	if (err)
623
		goto out_dput;
A
Amir Goldstein 已提交
624 625 626

	if (!c->metacopy)
		ovl_set_upperdata(d_inode(c->dentry));
627
	ovl_inode_update(d_inode(c->dentry), temp);
628

629 630 631
	return 0;

out_dput:
632
	dput(temp);
M
Miklos Szeredi 已提交
633 634 635 636 637 638
	return err;
}

/*
 * Copy up a single dentry
 *
M
Miklos Szeredi 已提交
639 640 641 642 643
 * 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.
M
Miklos Szeredi 已提交
644
 */
M
Miklos Szeredi 已提交
645
static int ovl_do_copy_up(struct ovl_copy_up_ctx *c)
M
Miklos Szeredi 已提交
646 647
{
	int err;
648
	struct ovl_fs *ofs = c->dentry->d_sb->s_fs_info;
649
	bool to_index = false;
650

651 652 653 654 655 656 657 658 659 660 661 662 663 664 665
	/*
	 * 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)
666 667
		c->origin = true;

668
	if (to_index) {
669 670 671 672
		c->destdir = ovl_indexdir(c->dentry->d_sb);
		err = ovl_get_index_name(c->lowerpath.dentry, &c->destname);
		if (err)
			return err;
673 674 675
	} else if (WARN_ON(!c->parent)) {
		/* Disconnected dentry must be copied up to index dir */
		return -EIO;
676 677 678 679 680 681 682 683 684
	} else {
		/*
		 * Mark parent "impure" because it may now contain non-pure
		 * upper
		 */
		err = ovl_set_impure(c->parent, c->destdir);
		if (err)
			return err;
	}
M
Miklos Szeredi 已提交
685

686
	/* Should we copyup with O_TMPFILE or with workdir? */
A
Amir Goldstein 已提交
687 688 689 690
	if (S_ISREG(c->stat.mode) && ofs->tmpfile)
		err = ovl_copy_up_tmpfile(c);
	else
		err = ovl_copy_up_workdir(c);
691 692 693 694
	if (err)
		goto out;

	if (c->indexed)
695 696 697
		ovl_set_flag(OVL_INDEX, d_inode(c->dentry));

	if (to_index) {
698 699 700
		/* Initialize nlink for copy up of disconnected dentry */
		err = ovl_set_nlink_upper(c->dentry);
	} else {
701 702 703 704 705 706 707 708
		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);
M
Miklos Szeredi 已提交
709 710
	}

711 712 713
out:
	if (to_index)
		kfree(c->destname.name);
M
Miklos Szeredi 已提交
714 715 716
	return err;
}

717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733
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;
}

734 735 736
/* 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)
{
737
	struct path upperpath, datapath;
738
	int err;
739 740
	char *capability = NULL;
	ssize_t uninitialized_var(cap_size);
741 742 743 744 745

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

746 747 748 749
	ovl_path_lowerdata(c->dentry, &datapath);
	if (WARN_ON(datapath.dentry == NULL))
		return -EIO;

750 751 752 753 754 755 756
	if (c->stat.size) {
		err = cap_size = ovl_getxattr(upperpath.dentry, XATTR_NAME_CAPS,
					      &capability, 0);
		if (err < 0 && err != -ENODATA)
			goto out;
	}

757
	err = ovl_copy_up_data(&datapath, &upperpath, c->stat.size);
758
	if (err)
759 760 761 762 763 764 765 766 767 768 769 770 771
		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;
	}

772 773 774

	err = vfs_removexattr(upperpath.dentry, OVL_XATTR_METACOPY);
	if (err)
775
		goto out_free;
776 777

	ovl_set_upperdata(d_inode(c->dentry));
778 779 780
out_free:
	kfree(capability);
out:
781 782 783
	return err;
}

M
Miklos Szeredi 已提交
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804
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;

805 806
	ctx.metacopy = ovl_need_meta_copy_up(dentry, ctx.stat.mode, flags);

807 808 809 810
	if (parent) {
		ovl_path_upper(parent, &parentpath);
		ctx.destdir = parentpath.dentry;
		ctx.destname = dentry->d_name;
M
Miklos Szeredi 已提交
811

812 813 814 815 816 817
		err = vfs_getattr(&parentpath, &ctx.pstat,
				  STATX_ATIME | STATX_MTIME,
				  AT_STATX_SYNC_AS_STAT);
		if (err)
			return err;
	}
M
Miklos Szeredi 已提交
818 819 820 821 822 823 824 825 826 827 828

	/* 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);
	}

829
	err = ovl_copy_up_start(dentry, flags);
M
Miklos Szeredi 已提交
830 831 832 833 834
	/* err < 0: interrupted, err > 0: raced with another copy-up */
	if (unlikely(err)) {
		if (err > 0)
			err = 0;
	} else {
835 836
		if (!ovl_dentry_upper(dentry))
			err = ovl_do_copy_up(&ctx);
837
		if (!err && parent && !ovl_dentry_has_upper_alias(dentry))
838
			err = ovl_link_up(&ctx);
839 840
		if (!err && ovl_dentry_needs_data_copy_up_locked(dentry, flags))
			err = ovl_copy_up_meta_inode_data(&ctx);
M
Miklos Szeredi 已提交
841 842
		ovl_copy_up_end(dentry);
	}
M
Miklos Szeredi 已提交
843
	do_delayed_call(&done);
M
Miklos Szeredi 已提交
844 845 846 847

	return err;
}

848
int ovl_copy_up_flags(struct dentry *dentry, int flags)
M
Miklos Szeredi 已提交
849
{
850 851
	int err = 0;
	const struct cred *old_cred = ovl_override_creds(dentry->d_sb);
852 853 854 855 856 857 858 859 860
	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;
M
Miklos Szeredi 已提交
861 862 863

	while (!err) {
		struct dentry *next;
864
		struct dentry *parent = NULL;
M
Miklos Szeredi 已提交
865

866
		if (ovl_already_copied_up(dentry, flags))
M
Miklos Szeredi 已提交
867 868 869 870
			break;

		next = dget(dentry);
		/* find the topmost dentry not yet copied up */
871
		for (; !disconnected;) {
M
Miklos Szeredi 已提交
872 873
			parent = dget_parent(next);

874
			if (ovl_dentry_upper(parent))
M
Miklos Szeredi 已提交
875 876 877 878 879 880
				break;

			dput(next);
			next = parent;
		}

M
Miklos Szeredi 已提交
881
		err = ovl_copy_up_one(parent, next, flags);
M
Miklos Szeredi 已提交
882 883 884 885

		dput(parent);
		dput(next);
	}
886
	revert_creds(old_cred);
M
Miklos Szeredi 已提交
887 888 889

	return err;
}
890

891 892 893
static bool ovl_open_need_copy_up(struct dentry *dentry, int flags)
{
	/* Copy up of disconnected dentry does not set upper alias */
894
	if (ovl_already_copied_up(dentry, flags))
895 896 897 898 899
		return false;

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

900
	if (!ovl_open_flags_need_copy_up(flags))
901 902 903 904 905
		return false;

	return true;
}

906
int ovl_maybe_copy_up(struct dentry *dentry, int flags)
907 908 909
{
	int err = 0;

910
	if (ovl_open_need_copy_up(dentry, flags)) {
911 912
		err = ovl_want_write(dentry);
		if (!err) {
913
			err = ovl_copy_up_flags(dentry, flags);
914 915 916 917 918 919 920
			ovl_drop_write(dentry);
		}
	}

	return err;
}

921 922 923 924 925
int ovl_copy_up_with_data(struct dentry *dentry)
{
	return ovl_copy_up_flags(dentry, O_WRONLY);
}

926 927 928 929
int ovl_copy_up(struct dentry *dentry)
{
	return ovl_copy_up_flags(dentry, 0);
}