recovery.c 11.0 KB
Newer Older
J
Jaegeuk Kim 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 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
 * fs/f2fs/recovery.c
 *
 * Copyright (c) 2012 Samsung Electronics Co., Ltd.
 *             http://www.samsung.com/
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */
#include <linux/fs.h>
#include <linux/f2fs_fs.h>
#include "f2fs.h"
#include "node.h"
#include "segment.h"

static struct kmem_cache *fsync_entry_slab;

bool space_for_roll_forward(struct f2fs_sb_info *sbi)
{
	if (sbi->last_valid_block_count + sbi->alloc_valid_block_count
			> sbi->user_block_count)
		return false;
	return true;
}

static struct fsync_inode_entry *get_fsync_inode(struct list_head *head,
								nid_t ino)
{
	struct list_head *this;
	struct fsync_inode_entry *entry;

	list_for_each(this, head) {
		entry = list_entry(this, struct fsync_inode_entry, list);
		if (entry->inode->i_ino == ino)
			return entry;
	}
	return NULL;
}

static int recover_dentry(struct page *ipage, struct inode *inode)
{
43
	struct f2fs_inode *raw_inode = F2FS_INODE(ipage);
44
	nid_t pino = le32_to_cpu(raw_inode->i_pino);
J
Jaegeuk Kim 已提交
45
	struct f2fs_dir_entry *de;
46
	struct qstr name;
47
	struct page *page;
J
Jaegeuk Kim 已提交
48
	struct inode *dir, *einode;
49 50
	int err = 0;

51 52 53 54 55 56 57 58
	dir = check_dirty_dir_inode(F2FS_SB(inode->i_sb), pino);
	if (!dir) {
		dir = f2fs_iget(inode->i_sb, pino);
		if (IS_ERR(dir)) {
			err = PTR_ERR(dir);
			goto out;
		}
		set_inode_flag(F2FS_I(dir), FI_DELAY_IPUT);
59
		add_dirty_dir_inode(dir);
60 61
	}

62 63
	name.len = le32_to_cpu(raw_inode->i_namelen);
	name.name = raw_inode->i_name;
64 65 66 67 68 69

	if (unlikely(name.len > F2FS_NAME_LEN)) {
		WARN_ON(1);
		err = -ENAMETOOLONG;
		goto out;
	}
J
Jaegeuk Kim 已提交
70 71
retry:
	de = f2fs_find_entry(dir, &name, &page);
72 73
	if (de && inode->i_ino == le32_to_cpu(de->ino))
		goto out_unmap_put;
J
Jaegeuk Kim 已提交
74 75 76 77 78 79
	if (de) {
		einode = f2fs_iget(inode->i_sb, le32_to_cpu(de->ino));
		if (IS_ERR(einode)) {
			WARN_ON(1);
			if (PTR_ERR(einode) == -ENOENT)
				err = -EEXIST;
80 81 82 83 84 85
			goto out_unmap_put;
		}
		err = acquire_orphan_inode(F2FS_SB(inode->i_sb));
		if (err) {
			iput(einode);
			goto out_unmap_put;
J
Jaegeuk Kim 已提交
86 87 88 89
		}
		f2fs_delete_entry(de, page, einode);
		iput(einode);
		goto retry;
90
	}
J
Jaegeuk Kim 已提交
91
	err = __f2fs_add_link(dir, &name, inode);
92 93 94 95 96
	goto out;

out_unmap_put:
	kunmap(page);
	f2fs_put_page(page, 0);
97
out:
C
Chris Fries 已提交
98 99 100
	f2fs_msg(inode->i_sb, KERN_NOTICE,
			"%s: ino = %x, name = %s, dir = %lx, err = %d",
			__func__, ino_of_node(ipage), raw_inode->i_name,
D
Dan Carpenter 已提交
101
			IS_ERR(dir) ? 0 : dir->i_ino, err);
102 103 104 105 106
	return err;
}

static int recover_inode(struct inode *inode, struct page *node_page)
{
107
	struct f2fs_inode *raw_inode = F2FS_INODE(node_page);
108

109 110 111
	if (!IS_INODE(node_page))
		return 0;

112
	inode->i_mode = le16_to_cpu(raw_inode->i_mode);
113 114 115 116 117 118 119 120
	i_size_write(inode, le64_to_cpu(raw_inode->i_size));
	inode->i_atime.tv_sec = le64_to_cpu(raw_inode->i_mtime);
	inode->i_ctime.tv_sec = le64_to_cpu(raw_inode->i_ctime);
	inode->i_mtime.tv_sec = le64_to_cpu(raw_inode->i_mtime);
	inode->i_atime.tv_nsec = le32_to_cpu(raw_inode->i_mtime_nsec);
	inode->i_ctime.tv_nsec = le32_to_cpu(raw_inode->i_ctime_nsec);
	inode->i_mtime.tv_nsec = le32_to_cpu(raw_inode->i_mtime_nsec);

121 122 123 124 125 126
	if (is_dent_dnode(node_page))
		return recover_dentry(node_page, inode);

	f2fs_msg(inode->i_sb, KERN_NOTICE, "recover_inode: ino = %x, name = %s",
			ino_of_node(node_page), raw_inode->i_name);
	return 0;
127 128 129 130
}

static int find_fsync_dnodes(struct f2fs_sb_info *sbi, struct list_head *head)
{
131
	unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi));
132 133 134 135 136 137 138 139 140 141 142
	struct curseg_info *curseg;
	struct page *page;
	block_t blkaddr;
	int err = 0;

	/* get node pages in the current segment */
	curseg = CURSEG_I(sbi, CURSEG_WARM_NODE);
	blkaddr = START_BLOCK(sbi, curseg->segno) + curseg->next_blkoff;

	/* read node page */
	page = alloc_page(GFP_F2FS_ZERO);
143 144
	if (!page)
		return -ENOMEM;
145 146 147 148 149
	lock_page(page);

	while (1) {
		struct fsync_inode_entry *entry;

150
		err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
151
		if (err)
152
			return err;
153

154 155
		lock_page(page);

156
		if (cp_ver != cpver_of_node(page))
157
			break;
158 159 160 161 162 163 164 165 166 167 168

		if (!is_fsync_dnode(page))
			goto next;

		entry = get_fsync_inode(head, ino_of_node(page));
		if (entry) {
			if (IS_INODE(page) && is_dent_dnode(page))
				set_inode_flag(F2FS_I(entry->inode),
							FI_INC_LINK);
		} else {
			if (IS_INODE(page) && is_dent_dnode(page)) {
169 170
				err = recover_inode_page(sbi, page);
				if (err)
171
					break;
172 173 174 175 176 177
			}

			/* add this fsync inode to the list */
			entry = kmem_cache_alloc(fsync_entry_slab, GFP_NOFS);
			if (!entry) {
				err = -ENOMEM;
178
				break;
179 180 181 182 183
			}

			entry->inode = f2fs_iget(sbi->sb, ino_of_node(page));
			if (IS_ERR(entry->inode)) {
				err = PTR_ERR(entry->inode);
184
				kmem_cache_free(fsync_entry_slab, entry);
185
				break;
186
			}
187
			list_add_tail(&entry->list, head);
188
		}
J
Jaegeuk Kim 已提交
189 190
		entry->blkaddr = blkaddr;

191 192 193
		err = recover_inode(entry->inode, page);
		if (err && err != -ENOENT)
			break;
194 195 196 197
next:
		/* check next segment */
		blkaddr = next_blkaddr_of_node(page);
	}
198

199 200
	unlock_page(page);
	__free_pages(page, 0);
201

202 203 204
	return err;
}

205
static void destroy_fsync_dnodes(struct list_head *head)
206
{
207 208 209
	struct fsync_inode_entry *entry, *tmp;

	list_for_each_entry_safe(entry, tmp, head, list) {
210 211 212 213 214 215
		iput(entry->inode);
		list_del(&entry->list);
		kmem_cache_free(fsync_entry_slab, entry);
	}
}

216
static int check_index_in_prev_nodes(struct f2fs_sb_info *sbi,
217
			block_t blkaddr, struct dnode_of_data *dn)
218 219 220 221 222 223
{
	struct seg_entry *sentry;
	unsigned int segno = GET_SEGNO(sbi, blkaddr);
	unsigned short blkoff = GET_SEGOFF_FROM_SEG0(sbi, blkaddr) &
					(sbi->blocks_per_seg - 1);
	struct f2fs_summary sum;
224
	nid_t ino, nid;
225 226 227
	void *kaddr;
	struct inode *inode;
	struct page *node_page;
228
	unsigned int offset;
229 230 231 232 233
	block_t bidx;
	int i;

	sentry = get_seg_entry(sbi, segno);
	if (!f2fs_test_bit(blkoff, sentry->cur_valid_map))
234
		return 0;
235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252

	/* Get the previous summary */
	for (i = CURSEG_WARM_DATA; i <= CURSEG_COLD_DATA; i++) {
		struct curseg_info *curseg = CURSEG_I(sbi, i);
		if (curseg->segno == segno) {
			sum = curseg->sum_blk->entries[blkoff];
			break;
		}
	}
	if (i > CURSEG_COLD_DATA) {
		struct page *sum_page = get_sum_page(sbi, segno);
		struct f2fs_summary_block *sum_node;
		kaddr = page_address(sum_page);
		sum_node = (struct f2fs_summary_block *)kaddr;
		sum = sum_node->entries[blkoff];
		f2fs_put_page(sum_page, 1);
	}

253 254 255 256 257 258
	/* Use the locked dnode page and inode */
	nid = le32_to_cpu(sum.nid);
	if (dn->inode->i_ino == nid) {
		struct dnode_of_data tdn = *dn;
		tdn.nid = nid;
		tdn.node_page = dn->inode_page;
259
		tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
260
		truncate_data_blocks_range(&tdn, 1);
261
		return 0;
262 263
	} else if (dn->nid == nid) {
		struct dnode_of_data tdn = *dn;
264
		tdn.ofs_in_node = le16_to_cpu(sum.ofs_in_node);
265
		truncate_data_blocks_range(&tdn, 1);
266
		return 0;
267 268
	}

269
	/* Get the node page */
270
	node_page = get_node_page(sbi, nid);
271 272
	if (IS_ERR(node_page))
		return PTR_ERR(node_page);
273 274

	offset = ofs_of_node(node_page);
275 276 277 278
	ino = ino_of_node(node_page);
	f2fs_put_page(node_page, 1);

	/* Deallocate previous index in the node page */
279
	inode = f2fs_iget(sbi->sb, ino);
280
	if (IS_ERR(inode))
281
		return PTR_ERR(inode);
282

283 284 285
	bidx = start_bidx_of_node(offset, F2FS_I(inode)) +
					le16_to_cpu(sum.ofs_in_node);

286 287
	truncate_hole(inode, bidx, bidx + 1);
	iput(inode);
288
	return 0;
289 290
}

291
static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode,
292 293
					struct page *page, block_t blkaddr)
{
294
	struct f2fs_inode_info *fi = F2FS_I(inode);
295 296 297 298
	unsigned int start, end;
	struct dnode_of_data dn;
	struct f2fs_summary sum;
	struct node_info ni;
299
	int err = 0, recovered = 0;
300

301 302 303
	if (recover_inline_data(inode, page))
		goto out;

304 305 306
	if (recover_xattr_data(inode, page, blkaddr))
		goto out;

307
	start = start_bidx_of_node(ofs_of_node(page), fi);
308
	if (IS_INODE(page))
309
		end = start + ADDRS_PER_INODE(fi);
310 311 312
	else
		end = start + ADDRS_PER_BLOCK;

313
	f2fs_lock_op(sbi);
314

315
	set_new_dnode(&dn, inode, NULL, NULL, 0);
316

317
	err = get_dnode_of_data(&dn, start, ALLOC_NODE);
318
	if (err) {
319
		f2fs_unlock_op(sbi);
320
		goto out;
321
	}
322 323 324 325

	wait_on_page_writeback(dn.node_page);

	get_node_info(sbi, dn.nid, &ni);
326 327
	f2fs_bug_on(ni.ino != ino_of_node(page));
	f2fs_bug_on(ofs_of_node(dn.node_page) != ofs_of_node(page));
328 329 330 331 332 333 334 335 336

	for (; start < end; start++) {
		block_t src, dest;

		src = datablock_addr(dn.node_page, dn.ofs_in_node);
		dest = datablock_addr(page, dn.ofs_in_node);

		if (src != dest && dest != NEW_ADDR && dest != NULL_ADDR) {
			if (src == NULL_ADDR) {
337
				err = reserve_new_block(&dn);
338
				/* We should not get -ENOSPC */
339
				f2fs_bug_on(err);
340 341 342
			}

			/* Check the previous node page having this index */
343 344 345
			err = check_index_in_prev_nodes(sbi, dest, &dn);
			if (err)
				goto err;
346 347 348 349 350 351

			set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version);

			/* write dummy data page */
			recover_data_page(sbi, NULL, &sum, src, dest);
			update_extent_cache(dest, &dn);
352
			recovered++;
353 354 355 356 357 358 359 360 361 362 363 364 365 366 367
		}
		dn.ofs_in_node++;
	}

	/* write node page in place */
	set_summary(&sum, dn.nid, 0, 0);
	if (IS_INODE(dn.node_page))
		sync_inode_page(&dn);

	copy_node_footer(dn.node_page, page);
	fill_node_footer(dn.node_page, dn.nid, ni.ino,
					ofs_of_node(page), false);
	set_page_dirty(dn.node_page);

	recover_node_page(sbi, dn.node_page, &sum, &ni, blkaddr);
368
err:
369
	f2fs_put_dnode(&dn);
370
	f2fs_unlock_op(sbi);
371
out:
C
Chris Fries 已提交
372 373 374
	f2fs_msg(sbi->sb, KERN_NOTICE,
		"recover_data: ino = %lx, recovered = %d blocks, err = %d",
		inode->i_ino, recovered, err);
375
	return err;
376 377
}

378
static int recover_data(struct f2fs_sb_info *sbi,
379 380
				struct list_head *head, int type)
{
381
	unsigned long long cp_ver = cur_cp_version(F2FS_CKPT(sbi));
382 383
	struct curseg_info *curseg;
	struct page *page;
384
	int err = 0;
385 386 387 388 389 390 391
	block_t blkaddr;

	/* get node pages in the current segment */
	curseg = CURSEG_I(sbi, type);
	blkaddr = NEXT_FREE_BLKADDR(sbi, curseg);

	/* read node page */
392
	page = alloc_page(GFP_F2FS_ZERO);
393
	if (!page)
394 395
		return -ENOMEM;

396 397 398 399 400
	lock_page(page);

	while (1) {
		struct fsync_inode_entry *entry;

401
		err = f2fs_submit_page_bio(sbi, page, blkaddr, READ_SYNC);
402
		if (err)
403
			return err;
404

405 406
		lock_page(page);

407
		if (cp_ver != cpver_of_node(page))
408
			break;
409 410 411 412 413

		entry = get_fsync_inode(head, ino_of_node(page));
		if (!entry)
			goto next;

414 415
		err = do_recover_data(sbi, entry->inode, page, blkaddr);
		if (err)
416
			break;
417 418 419 420 421 422 423 424 425 426

		if (entry->blkaddr == blkaddr) {
			iput(entry->inode);
			list_del(&entry->list);
			kmem_cache_free(fsync_entry_slab, entry);
		}
next:
		/* check next segment */
		blkaddr = next_blkaddr_of_node(page);
	}
427

428 429 430
	unlock_page(page);
	__free_pages(page, 0);

431 432 433
	if (!err)
		allocate_new_segments(sbi);
	return err;
434 435
}

436
int recover_fsync_data(struct f2fs_sb_info *sbi)
437 438
{
	struct list_head inode_list;
439
	int err;
H
Haicheng Li 已提交
440
	bool need_writecp = false;
441 442 443

	fsync_entry_slab = f2fs_kmem_cache_create("f2fs_fsync_inode_entry",
			sizeof(struct fsync_inode_entry), NULL);
444
	if (!fsync_entry_slab)
445
		return -ENOMEM;
446 447 448 449

	INIT_LIST_HEAD(&inode_list);

	/* step #1: find fsynced inode numbers */
H
Haicheng Li 已提交
450
	sbi->por_doing = true;
451 452
	err = find_fsync_dnodes(sbi, &inode_list);
	if (err)
453 454 455 456 457
		goto out;

	if (list_empty(&inode_list))
		goto out;

H
Haicheng Li 已提交
458
	need_writecp = true;
459

460
	/* step #2: recover data */
461
	err = recover_data(sbi, &inode_list, CURSEG_WARM_NODE);
462
	f2fs_bug_on(!list_empty(&inode_list));
463
out:
464
	destroy_fsync_dnodes(&inode_list);
465
	kmem_cache_destroy(fsync_entry_slab);
H
Haicheng Li 已提交
466
	sbi->por_doing = false;
467
	if (!err && need_writecp)
468
		write_checkpoint(sbi, false);
469
	return err;
470
}