node.c 66.6 KB
Newer Older
J
Jaegeuk Kim 已提交
1
/*
J
Jaegeuk Kim 已提交
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
 * fs/f2fs/node.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 <linux/mpage.h>
#include <linux/backing-dev.h>
#include <linux/blkdev.h>
#include <linux/pagevec.h>
#include <linux/swap.h>

#include "f2fs.h"
#include "node.h"
#include "segment.h"
J
Jaegeuk Kim 已提交
22
#include "trace.h"
23
#include <trace/events/f2fs.h>
J
Jaegeuk Kim 已提交
24

25 26
#define on_build_free_nids(nmi) mutex_is_locked(&nm_i->build_lock)

J
Jaegeuk Kim 已提交
27 28
static struct kmem_cache *nat_entry_slab;
static struct kmem_cache *free_nid_slab;
29
static struct kmem_cache *nat_entry_set_slab;
J
Jaegeuk Kim 已提交
30

31
bool available_free_memory(struct f2fs_sb_info *sbi, int type)
32
{
33
	struct f2fs_nm_info *nm_i = NM_I(sbi);
34
	struct sysinfo val;
35
	unsigned long avail_ram;
36
	unsigned long mem_size = 0;
37
	bool res = false;
38 39

	si_meminfo(&val);
40 41 42 43

	/* only uses low memory */
	avail_ram = val.totalram - val.totalhigh;

44 45 46
	/*
	 * give 25%, 25%, 50%, 50%, 50% memory for each components respectively
	 */
47
	if (type == FREE_NIDS) {
C
Chao Yu 已提交
48 49
		mem_size = (nm_i->nid_cnt[FREE_NID_LIST] *
				sizeof(struct free_nid)) >> PAGE_SHIFT;
50
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
51
	} else if (type == NAT_ENTRIES) {
52
		mem_size = (nm_i->nat_cnt * sizeof(struct nat_entry)) >>
53
							PAGE_SHIFT;
54
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
55 56
		if (excess_cached_nats(sbi))
			res = false;
57 58 59 60 61
	} else if (type == DIRTY_DENTS) {
		if (sbi->sb->s_bdi->wb.dirty_exceeded)
			return false;
		mem_size = get_pages(sbi, F2FS_DIRTY_DENTS);
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
62 63 64 65
	} else if (type == INO_ENTRIES) {
		int i;

		for (i = 0; i <= UPDATE_INO; i++)
66
			mem_size += (sbi->im[i].ino_num *
67
				sizeof(struct ino_entry)) >> PAGE_SHIFT;
68
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
69
	} else if (type == EXTENT_CACHE) {
70 71
		mem_size = (atomic_read(&sbi->total_ext_tree) *
				sizeof(struct extent_tree) +
72
				atomic_read(&sbi->total_ext_node) *
73
				sizeof(struct extent_node)) >> PAGE_SHIFT;
74
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
75
	} else {
76 77
		if (!sbi->sb->s_bdi->wb.dirty_exceeded)
			return true;
78 79
	}
	return res;
80 81
}

J
Jaegeuk Kim 已提交
82 83 84 85 86 87 88 89 90 91 92 93 94
static void clear_node_page_dirty(struct page *page)
{
	struct address_space *mapping = page->mapping;
	unsigned int long flags;

	if (PageDirty(page)) {
		spin_lock_irqsave(&mapping->tree_lock, flags);
		radix_tree_tag_clear(&mapping->page_tree,
				page_index(page),
				PAGECACHE_TAG_DIRTY);
		spin_unlock_irqrestore(&mapping->tree_lock, flags);

		clear_page_dirty_for_io(page);
95
		dec_page_count(F2FS_M_SB(mapping), F2FS_DIRTY_NODES);
J
Jaegeuk Kim 已提交
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
	}
	ClearPageUptodate(page);
}

static struct page *get_current_nat_page(struct f2fs_sb_info *sbi, nid_t nid)
{
	pgoff_t index = current_nat_addr(sbi, nid);
	return get_meta_page(sbi, index);
}

static struct page *get_next_nat_page(struct f2fs_sb_info *sbi, nid_t nid)
{
	struct page *src_page;
	struct page *dst_page;
	pgoff_t src_off;
	pgoff_t dst_off;
	void *src_addr;
	void *dst_addr;
	struct f2fs_nm_info *nm_i = NM_I(sbi);

	src_off = current_nat_addr(sbi, nid);
	dst_off = next_nat_addr(sbi, src_off);

	/* get current nat block page with lock */
	src_page = get_meta_page(sbi, src_off);
	dst_page = grab_meta_page(sbi, dst_off);
122
	f2fs_bug_on(sbi, PageDirty(src_page));
J
Jaegeuk Kim 已提交
123 124 125

	src_addr = page_address(src_page);
	dst_addr = page_address(dst_page);
126
	memcpy(dst_addr, src_addr, PAGE_SIZE);
J
Jaegeuk Kim 已提交
127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
	set_page_dirty(dst_page);
	f2fs_put_page(src_page, 1);

	set_to_next_nat(nm_i, nid);

	return dst_page;
}

static struct nat_entry *__lookup_nat_cache(struct f2fs_nm_info *nm_i, nid_t n)
{
	return radix_tree_lookup(&nm_i->nat_root, n);
}

static unsigned int __gang_lookup_nat_cache(struct f2fs_nm_info *nm_i,
		nid_t start, unsigned int nr, struct nat_entry **ep)
{
	return radix_tree_gang_lookup(&nm_i->nat_root, (void **)ep, start, nr);
}

static void __del_from_nat_cache(struct f2fs_nm_info *nm_i, struct nat_entry *e)
{
	list_del(&e->list);
	radix_tree_delete(&nm_i->nat_root, nat_get_nid(e));
	nm_i->nat_cnt--;
	kmem_cache_free(nat_entry_slab, e);
}

154 155 156 157 158 159 160 161
static void __set_nat_cache_dirty(struct f2fs_nm_info *nm_i,
						struct nat_entry *ne)
{
	nid_t set = NAT_BLOCK_OFFSET(ne->ni.nid);
	struct nat_entry_set *head;

	if (get_nat_flag(ne, IS_DIRTY))
		return;
162

163 164
	head = radix_tree_lookup(&nm_i->nat_set_root, set);
	if (!head) {
165
		head = f2fs_kmem_cache_alloc(nat_entry_set_slab, GFP_NOFS);
166 167 168 169 170

		INIT_LIST_HEAD(&head->entry_list);
		INIT_LIST_HEAD(&head->set_list);
		head->set = set;
		head->entry_cnt = 0;
171
		f2fs_radix_tree_insert(&nm_i->nat_set_root, set, head);
172 173 174 175 176 177 178 179 180 181
	}
	list_move_tail(&ne->list, &head->entry_list);
	nm_i->dirty_nat_cnt++;
	head->entry_cnt++;
	set_nat_flag(ne, IS_DIRTY, true);
}

static void __clear_nat_cache_dirty(struct f2fs_nm_info *nm_i,
						struct nat_entry *ne)
{
182
	nid_t set = NAT_BLOCK_OFFSET(ne->ni.nid);
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
	struct nat_entry_set *head;

	head = radix_tree_lookup(&nm_i->nat_set_root, set);
	if (head) {
		list_move_tail(&ne->list, &nm_i->nat_entries);
		set_nat_flag(ne, IS_DIRTY, false);
		head->entry_cnt--;
		nm_i->dirty_nat_cnt--;
	}
}

static unsigned int __gang_lookup_nat_set(struct f2fs_nm_info *nm_i,
		nid_t start, unsigned int nr, struct nat_entry_set **ep)
{
	return radix_tree_gang_lookup(&nm_i->nat_set_root, (void **)ep,
							start, nr);
}

J
Jaegeuk Kim 已提交
201
int need_dentry_mark(struct f2fs_sb_info *sbi, nid_t nid)
J
Jaegeuk Kim 已提交
202 203 204
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
J
Jaegeuk Kim 已提交
205
	bool need = false;
J
Jaegeuk Kim 已提交
206

207
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
208
	e = __lookup_nat_cache(nm_i, nid);
J
Jaegeuk Kim 已提交
209 210 211 212 213
	if (e) {
		if (!get_nat_flag(e, IS_CHECKPOINTED) &&
				!get_nat_flag(e, HAS_FSYNCED_INODE))
			need = true;
	}
214
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
215
	return need;
J
Jaegeuk Kim 已提交
216 217
}

J
Jaegeuk Kim 已提交
218
bool is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid)
219 220 221
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
J
Jaegeuk Kim 已提交
222
	bool is_cp = true;
223

224
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
225 226 227
	e = __lookup_nat_cache(nm_i, nid);
	if (e && !get_nat_flag(e, IS_CHECKPOINTED))
		is_cp = false;
228
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
229
	return is_cp;
230 231
}

232
bool need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino)
233 234 235
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
236
	bool need_update = true;
237

238
	down_read(&nm_i->nat_tree_lock);
239 240 241 242 243
	e = __lookup_nat_cache(nm_i, ino);
	if (e && get_nat_flag(e, HAS_LAST_FSYNC) &&
			(get_nat_flag(e, IS_CHECKPOINTED) ||
			 get_nat_flag(e, HAS_FSYNCED_INODE)))
		need_update = false;
244
	up_read(&nm_i->nat_tree_lock);
245
	return need_update;
246 247
}

248 249
static struct nat_entry *grab_nat_entry(struct f2fs_nm_info *nm_i, nid_t nid,
								bool no_fail)
J
Jaegeuk Kim 已提交
250 251 252
{
	struct nat_entry *new;

253 254 255 256 257 258 259 260 261 262 263 264 265
	if (no_fail) {
		new = f2fs_kmem_cache_alloc(nat_entry_slab, GFP_NOFS);
		f2fs_radix_tree_insert(&nm_i->nat_root, nid, new);
	} else {
		new = kmem_cache_alloc(nat_entry_slab, GFP_NOFS);
		if (!new)
			return NULL;
		if (radix_tree_insert(&nm_i->nat_root, nid, new)) {
			kmem_cache_free(nat_entry_slab, new);
			return NULL;
		}
	}

J
Jaegeuk Kim 已提交
266 267
	memset(new, 0, sizeof(struct nat_entry));
	nat_set_nid(new, nid);
268
	nat_reset_flag(new);
J
Jaegeuk Kim 已提交
269 270 271 272 273
	list_add_tail(&new->list, &nm_i->nat_entries);
	nm_i->nat_cnt++;
	return new;
}

274
static void cache_nat_entry(struct f2fs_sb_info *sbi, nid_t nid,
J
Jaegeuk Kim 已提交
275 276
						struct f2fs_nat_entry *ne)
{
277
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
278
	struct nat_entry *e;
279

J
Jaegeuk Kim 已提交
280 281
	e = __lookup_nat_cache(nm_i, nid);
	if (!e) {
282 283 284
		e = grab_nat_entry(nm_i, nid, false);
		if (e)
			node_info_from_raw_nat(&e->ni, ne);
285
	} else {
E
Eric Biggers 已提交
286 287 288
		f2fs_bug_on(sbi, nat_get_ino(e) != le32_to_cpu(ne->ino) ||
				nat_get_blkaddr(e) !=
					le32_to_cpu(ne->block_addr) ||
289
				nat_get_version(e) != ne->version);
J
Jaegeuk Kim 已提交
290 291 292 293
	}
}

static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni,
294
			block_t new_blkaddr, bool fsync_done)
J
Jaegeuk Kim 已提交
295 296 297
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
298

299
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
300 301
	e = __lookup_nat_cache(nm_i, ni->nid);
	if (!e) {
302
		e = grab_nat_entry(nm_i, ni->nid, true);
303
		copy_node_info(&e->ni, ni);
304
		f2fs_bug_on(sbi, ni->blk_addr == NEW_ADDR);
J
Jaegeuk Kim 已提交
305 306 307 308 309 310
	} else if (new_blkaddr == NEW_ADDR) {
		/*
		 * when nid is reallocated,
		 * previous nat entry can be remained in nat cache.
		 * So, reinitialize it with new information.
		 */
311
		copy_node_info(&e->ni, ni);
312
		f2fs_bug_on(sbi, ni->blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
313 314 315
	}

	/* sanity check */
316 317
	f2fs_bug_on(sbi, nat_get_blkaddr(e) != ni->blk_addr);
	f2fs_bug_on(sbi, nat_get_blkaddr(e) == NULL_ADDR &&
J
Jaegeuk Kim 已提交
318
			new_blkaddr == NULL_ADDR);
319
	f2fs_bug_on(sbi, nat_get_blkaddr(e) == NEW_ADDR &&
J
Jaegeuk Kim 已提交
320
			new_blkaddr == NEW_ADDR);
321
	f2fs_bug_on(sbi, nat_get_blkaddr(e) != NEW_ADDR &&
J
Jaegeuk Kim 已提交
322 323 324
			nat_get_blkaddr(e) != NULL_ADDR &&
			new_blkaddr == NEW_ADDR);

A
arter97 已提交
325
	/* increment version no as node is removed */
J
Jaegeuk Kim 已提交
326 327 328
	if (nat_get_blkaddr(e) != NEW_ADDR && new_blkaddr == NULL_ADDR) {
		unsigned char version = nat_get_version(e);
		nat_set_version(e, inc_node_version(version));
J
Jaegeuk Kim 已提交
329 330 331 332

		/* in order to reuse the nid */
		if (nm_i->next_scan_nid > ni->nid)
			nm_i->next_scan_nid = ni->nid;
J
Jaegeuk Kim 已提交
333 334 335 336
	}

	/* change address */
	nat_set_blkaddr(e, new_blkaddr);
337 338
	if (new_blkaddr == NEW_ADDR || new_blkaddr == NULL_ADDR)
		set_nat_flag(e, IS_CHECKPOINTED, false);
J
Jaegeuk Kim 已提交
339
	__set_nat_cache_dirty(nm_i, e);
340

341
	if (enabled_nat_bits(sbi, NULL) && new_blkaddr == NEW_ADDR)
J
Jaegeuk Kim 已提交
342
		__clear_bit_le(NAT_BLOCK_OFFSET(ni->nid), nm_i->empty_nat_bits);
343

344
	/* update fsync_mark if its inode nat entry is still alive */
345 346
	if (ni->nid != ni->ino)
		e = __lookup_nat_cache(nm_i, ni->ino);
347 348 349 350 351
	if (e) {
		if (fsync_done && ni->nid == ni->ino)
			set_nat_flag(e, HAS_FSYNCED_INODE, true);
		set_nat_flag(e, HAS_LAST_FSYNC, fsync_done);
	}
352
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
353 354
}

355
int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
J
Jaegeuk Kim 已提交
356 357
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
358
	int nr = nr_shrink;
J
Jaegeuk Kim 已提交
359

360 361
	if (!down_write_trylock(&nm_i->nat_tree_lock))
		return 0;
J
Jaegeuk Kim 已提交
362 363 364 365 366 367 368 369

	while (nr_shrink && !list_empty(&nm_i->nat_entries)) {
		struct nat_entry *ne;
		ne = list_first_entry(&nm_i->nat_entries,
					struct nat_entry, list);
		__del_from_nat_cache(nm_i, ne);
		nr_shrink--;
	}
370
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
371
	return nr - nr_shrink;
J
Jaegeuk Kim 已提交
372 373
}

J
Jaegeuk Kim 已提交
374
/*
A
arter97 已提交
375
 * This function always returns success
J
Jaegeuk Kim 已提交
376 377 378 379 380
 */
void get_node_info(struct f2fs_sb_info *sbi, nid_t nid, struct node_info *ni)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
381
	struct f2fs_journal *journal = curseg->journal;
J
Jaegeuk Kim 已提交
382 383 384 385 386 387 388 389 390 391
	nid_t start_nid = START_NID(nid);
	struct f2fs_nat_block *nat_blk;
	struct page *page = NULL;
	struct f2fs_nat_entry ne;
	struct nat_entry *e;
	int i;

	ni->nid = nid;

	/* Check nat cache */
392
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
393 394 395 396 397
	e = __lookup_nat_cache(nm_i, nid);
	if (e) {
		ni->ino = nat_get_ino(e);
		ni->blk_addr = nat_get_blkaddr(e);
		ni->version = nat_get_version(e);
398
		up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
399
		return;
400
	}
J
Jaegeuk Kim 已提交
401

402 403
	memset(&ne, 0, sizeof(struct f2fs_nat_entry));

J
Jaegeuk Kim 已提交
404
	/* Check current segment summary */
405
	down_read(&curseg->journal_rwsem);
406
	i = lookup_journal_in_cursum(journal, NAT_JOURNAL, nid, 0);
J
Jaegeuk Kim 已提交
407
	if (i >= 0) {
408
		ne = nat_in_journal(journal, i);
J
Jaegeuk Kim 已提交
409 410
		node_info_from_raw_nat(ni, &ne);
	}
411
	up_read(&curseg->journal_rwsem);
J
Jaegeuk Kim 已提交
412 413 414 415 416 417 418 419 420 421
	if (i >= 0)
		goto cache;

	/* Fill node_info from nat page */
	page = get_current_nat_page(sbi, start_nid);
	nat_blk = (struct f2fs_nat_block *)page_address(page);
	ne = nat_blk->entries[nid - start_nid];
	node_info_from_raw_nat(ni, &ne);
	f2fs_put_page(page, 1);
cache:
422
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
423
	/* cache nat entry */
424
	down_write(&nm_i->nat_tree_lock);
425
	cache_nat_entry(sbi, nid, &ne);
426
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
427 428
}

429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451
/*
 * readahead MAX_RA_NODE number of node pages.
 */
static void ra_node_pages(struct page *parent, int start, int n)
{
	struct f2fs_sb_info *sbi = F2FS_P_SB(parent);
	struct blk_plug plug;
	int i, end;
	nid_t nid;

	blk_start_plug(&plug);

	/* Then, try readahead for siblings of the desired node */
	end = start + n;
	end = min(end, NIDS_PER_BLOCK);
	for (i = start; i < end; i++) {
		nid = get_nid(parent, i, false);
		ra_node_page(sbi, nid);
	}

	blk_finish_plug(&plug);
}

452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482
pgoff_t get_next_page_offset(struct dnode_of_data *dn, pgoff_t pgofs)
{
	const long direct_index = ADDRS_PER_INODE(dn->inode);
	const long direct_blks = ADDRS_PER_BLOCK;
	const long indirect_blks = ADDRS_PER_BLOCK * NIDS_PER_BLOCK;
	unsigned int skipped_unit = ADDRS_PER_BLOCK;
	int cur_level = dn->cur_level;
	int max_level = dn->max_level;
	pgoff_t base = 0;

	if (!dn->max_level)
		return pgofs + 1;

	while (max_level-- > cur_level)
		skipped_unit *= NIDS_PER_BLOCK;

	switch (dn->max_level) {
	case 3:
		base += 2 * indirect_blks;
	case 2:
		base += 2 * direct_blks;
	case 1:
		base += direct_index;
		break;
	default:
		f2fs_bug_on(F2FS_I_SB(dn->inode), 1);
	}

	return ((pgofs - base) / skipped_unit + 1) * skipped_unit + base;
}

J
Jaegeuk Kim 已提交
483
/*
J
Jaegeuk Kim 已提交
484 485 486
 * The maximum depth is four.
 * Offset[0] will have raw inode offset.
 */
487
static int get_node_path(struct inode *inode, long block,
488
				int offset[4], unsigned int noffset[4])
J
Jaegeuk Kim 已提交
489
{
490
	const long direct_index = ADDRS_PER_INODE(inode);
J
Jaegeuk Kim 已提交
491 492 493 494 495 496 497 498 499 500
	const long direct_blks = ADDRS_PER_BLOCK;
	const long dptrs_per_blk = NIDS_PER_BLOCK;
	const long indirect_blks = ADDRS_PER_BLOCK * NIDS_PER_BLOCK;
	const long dindirect_blks = indirect_blks * NIDS_PER_BLOCK;
	int n = 0;
	int level = 0;

	noffset[0] = 0;

	if (block < direct_index) {
501
		offset[n] = block;
J
Jaegeuk Kim 已提交
502 503 504 505 506 507
		goto got;
	}
	block -= direct_index;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR1_BLOCK;
		noffset[n] = 1;
508
		offset[n] = block;
J
Jaegeuk Kim 已提交
509 510 511 512 513 514 515
		level = 1;
		goto got;
	}
	block -= direct_blks;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR2_BLOCK;
		noffset[n] = 2;
516
		offset[n] = block;
J
Jaegeuk Kim 已提交
517 518 519 520 521 522 523 524 525
		level = 1;
		goto got;
	}
	block -= direct_blks;
	if (block < indirect_blks) {
		offset[n++] = NODE_IND1_BLOCK;
		noffset[n] = 3;
		offset[n++] = block / direct_blks;
		noffset[n] = 4 + offset[n - 1];
526
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
527 528 529 530 531 532 533 534 535
		level = 2;
		goto got;
	}
	block -= indirect_blks;
	if (block < indirect_blks) {
		offset[n++] = NODE_IND2_BLOCK;
		noffset[n] = 4 + dptrs_per_blk;
		offset[n++] = block / direct_blks;
		noffset[n] = 5 + dptrs_per_blk + offset[n - 1];
536
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
537 538 539 540 541 542 543 544 545 546 547 548 549 550
		level = 2;
		goto got;
	}
	block -= indirect_blks;
	if (block < dindirect_blks) {
		offset[n++] = NODE_DIND_BLOCK;
		noffset[n] = 5 + (dptrs_per_blk * 2);
		offset[n++] = block / indirect_blks;
		noffset[n] = 6 + (dptrs_per_blk * 2) +
			      offset[n - 1] * (dptrs_per_blk + 1);
		offset[n++] = (block / direct_blks) % dptrs_per_blk;
		noffset[n] = 7 + (dptrs_per_blk * 2) +
			      offset[n - 2] * (dptrs_per_blk + 1) +
			      offset[n - 1];
551
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
552 553 554 555 556 557 558 559 560 561 562
		level = 3;
		goto got;
	} else {
		BUG();
	}
got:
	return level;
}

/*
 * Caller should call f2fs_put_dnode(dn).
C
Chao Yu 已提交
563 564
 * Also, it should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op() only if ro is not set RDONLY_NODE.
565
 * In the case of RDONLY_NODE, we don't need to care about mutex.
J
Jaegeuk Kim 已提交
566
 */
567
int get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
J
Jaegeuk Kim 已提交
568
{
569
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
570
	struct page *npage[4];
571
	struct page *parent = NULL;
J
Jaegeuk Kim 已提交
572 573 574
	int offset[4];
	unsigned int noffset[4];
	nid_t nids[4];
575
	int level, i = 0;
J
Jaegeuk Kim 已提交
576 577
	int err = 0;

578
	level = get_node_path(dn->inode, index, offset, noffset);
J
Jaegeuk Kim 已提交
579 580

	nids[0] = dn->inode->i_ino;
581
	npage[0] = dn->inode_page;
J
Jaegeuk Kim 已提交
582

583 584 585 586 587
	if (!npage[0]) {
		npage[0] = get_node_page(sbi, nids[0]);
		if (IS_ERR(npage[0]))
			return PTR_ERR(npage[0]);
	}
588 589 590

	/* if inline_data is set, should not report any block indices */
	if (f2fs_has_inline_data(dn->inode) && index) {
591
		err = -ENOENT;
592 593 594 595
		f2fs_put_page(npage[0], 1);
		goto release_out;
	}

J
Jaegeuk Kim 已提交
596
	parent = npage[0];
597 598
	if (level != 0)
		nids[1] = get_nid(parent, offset[0], true);
J
Jaegeuk Kim 已提交
599 600 601 602 603 604 605
	dn->inode_page = npage[0];
	dn->inode_page_locked = true;

	/* get indirect or direct nodes */
	for (i = 1; i <= level; i++) {
		bool done = false;

606
		if (!nids[i] && mode == ALLOC_NODE) {
J
Jaegeuk Kim 已提交
607 608 609 610 611 612 613
			/* alloc new node */
			if (!alloc_nid(sbi, &(nids[i]))) {
				err = -ENOSPC;
				goto release_pages;
			}

			dn->nid = nids[i];
614
			npage[i] = new_node_page(dn, noffset[i], NULL);
J
Jaegeuk Kim 已提交
615 616 617 618 619 620 621 622 623
			if (IS_ERR(npage[i])) {
				alloc_nid_failed(sbi, nids[i]);
				err = PTR_ERR(npage[i]);
				goto release_pages;
			}

			set_nid(parent, offset[i - 1], nids[i], i == 1);
			alloc_nid_done(sbi, nids[i]);
			done = true;
624
		} else if (mode == LOOKUP_NODE_RA && i == level && level > 1) {
J
Jaegeuk Kim 已提交
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
			npage[i] = get_node_page_ra(parent, offset[i - 1]);
			if (IS_ERR(npage[i])) {
				err = PTR_ERR(npage[i]);
				goto release_pages;
			}
			done = true;
		}
		if (i == 1) {
			dn->inode_page_locked = false;
			unlock_page(parent);
		} else {
			f2fs_put_page(parent, 1);
		}

		if (!done) {
			npage[i] = get_node_page(sbi, nids[i]);
			if (IS_ERR(npage[i])) {
				err = PTR_ERR(npage[i]);
				f2fs_put_page(npage[0], 0);
				goto release_out;
			}
		}
		if (i < level) {
			parent = npage[i];
			nids[i + 1] = get_nid(parent, offset[i], false);
		}
	}
	dn->nid = nids[level];
	dn->ofs_in_node = offset[level];
	dn->node_page = npage[level];
	dn->data_blkaddr = datablock_addr(dn->node_page, dn->ofs_in_node);
	return 0;

release_pages:
	f2fs_put_page(parent, 1);
	if (i > 1)
		f2fs_put_page(npage[0], 0);
release_out:
	dn->inode_page = NULL;
	dn->node_page = NULL;
665 666 667
	if (err == -ENOENT) {
		dn->cur_level = i;
		dn->max_level = level;
668
		dn->ofs_in_node = offset[level];
669
	}
J
Jaegeuk Kim 已提交
670 671 672 673 674
	return err;
}

static void truncate_node(struct dnode_of_data *dn)
{
675
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
676 677 678
	struct node_info ni;

	get_node_info(sbi, dn->nid, &ni);
679
	if (dn->inode->i_blocks == 0) {
680
		f2fs_bug_on(sbi, ni.blk_addr != NULL_ADDR);
681 682
		goto invalidate;
	}
683
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
J
Jaegeuk Kim 已提交
684 685

	/* Deallocate node address */
686
	invalidate_blocks(sbi, ni.blk_addr);
687
	dec_valid_node_count(sbi, dn->inode);
688
	set_node_addr(sbi, &ni, NULL_ADDR, false);
J
Jaegeuk Kim 已提交
689 690 691 692

	if (dn->nid == dn->inode->i_ino) {
		remove_orphan_inode(sbi, dn->nid);
		dec_valid_inode_count(sbi);
693
		f2fs_inode_synced(dn->inode);
J
Jaegeuk Kim 已提交
694
	}
695
invalidate:
J
Jaegeuk Kim 已提交
696
	clear_node_page_dirty(dn->node_page);
697
	set_sbi_flag(sbi, SBI_IS_DIRTY);
J
Jaegeuk Kim 已提交
698 699

	f2fs_put_page(dn->node_page, 1);
700 701 702 703

	invalidate_mapping_pages(NODE_MAPPING(sbi),
			dn->node_page->index, dn->node_page->index);

J
Jaegeuk Kim 已提交
704
	dn->node_page = NULL;
705
	trace_f2fs_truncate_node(dn->inode, dn->nid, ni.blk_addr);
J
Jaegeuk Kim 已提交
706 707 708 709 710 711 712 713 714 715
}

static int truncate_dnode(struct dnode_of_data *dn)
{
	struct page *page;

	if (dn->nid == 0)
		return 1;

	/* get direct node */
716
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
J
Jaegeuk Kim 已提交
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
	if (IS_ERR(page) && PTR_ERR(page) == -ENOENT)
		return 1;
	else if (IS_ERR(page))
		return PTR_ERR(page);

	/* Make dnode_of_data for parameter */
	dn->node_page = page;
	dn->ofs_in_node = 0;
	truncate_data_blocks(dn);
	truncate_node(dn);
	return 1;
}

static int truncate_nodes(struct dnode_of_data *dn, unsigned int nofs,
						int ofs, int depth)
{
	struct dnode_of_data rdn = *dn;
	struct page *page;
	struct f2fs_node *rn;
	nid_t child_nid;
	unsigned int child_nofs;
	int freed = 0;
	int i, ret;

	if (dn->nid == 0)
		return NIDS_PER_BLOCK + 1;

744 745
	trace_f2fs_truncate_nodes_enter(dn->inode, dn->nid, dn->data_blkaddr);

746
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
747 748
	if (IS_ERR(page)) {
		trace_f2fs_truncate_nodes_exit(dn->inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
749
		return PTR_ERR(page);
750
	}
J
Jaegeuk Kim 已提交
751

752 753
	ra_node_pages(page, ofs, NIDS_PER_BLOCK);

754
	rn = F2FS_NODE(page);
J
Jaegeuk Kim 已提交
755 756 757 758 759 760 761 762 763
	if (depth < 3) {
		for (i = ofs; i < NIDS_PER_BLOCK; i++, freed++) {
			child_nid = le32_to_cpu(rn->in.nid[i]);
			if (child_nid == 0)
				continue;
			rdn.nid = child_nid;
			ret = truncate_dnode(&rdn);
			if (ret < 0)
				goto out_err;
764 765
			if (set_nid(page, i, 0, false))
				dn->node_changed = true;
J
Jaegeuk Kim 已提交
766 767 768 769 770 771 772 773 774 775 776 777
		}
	} else {
		child_nofs = nofs + ofs * (NIDS_PER_BLOCK + 1) + 1;
		for (i = ofs; i < NIDS_PER_BLOCK; i++) {
			child_nid = le32_to_cpu(rn->in.nid[i]);
			if (child_nid == 0) {
				child_nofs += NIDS_PER_BLOCK + 1;
				continue;
			}
			rdn.nid = child_nid;
			ret = truncate_nodes(&rdn, child_nofs, 0, depth - 1);
			if (ret == (NIDS_PER_BLOCK + 1)) {
778 779
				if (set_nid(page, i, 0, false))
					dn->node_changed = true;
J
Jaegeuk Kim 已提交
780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795
				child_nofs += ret;
			} else if (ret < 0 && ret != -ENOENT) {
				goto out_err;
			}
		}
		freed = child_nofs;
	}

	if (!ofs) {
		/* remove current indirect node */
		dn->node_page = page;
		truncate_node(dn);
		freed++;
	} else {
		f2fs_put_page(page, 1);
	}
796
	trace_f2fs_truncate_nodes_exit(dn->inode, freed);
J
Jaegeuk Kim 已提交
797 798 799 800
	return freed;

out_err:
	f2fs_put_page(page, 1);
801
	trace_f2fs_truncate_nodes_exit(dn->inode, ret);
J
Jaegeuk Kim 已提交
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819
	return ret;
}

static int truncate_partial_nodes(struct dnode_of_data *dn,
			struct f2fs_inode *ri, int *offset, int depth)
{
	struct page *pages[2];
	nid_t nid[3];
	nid_t child_nid;
	int err = 0;
	int i;
	int idx = depth - 2;

	nid[0] = le32_to_cpu(ri->i_nid[offset[0] - NODE_DIR1_BLOCK]);
	if (!nid[0])
		return 0;

	/* get indirect nodes in the path */
820
	for (i = 0; i < idx + 1; i++) {
A
arter97 已提交
821
		/* reference count'll be increased */
822
		pages[i] = get_node_page(F2FS_I_SB(dn->inode), nid[i]);
J
Jaegeuk Kim 已提交
823 824
		if (IS_ERR(pages[i])) {
			err = PTR_ERR(pages[i]);
825
			idx = i - 1;
J
Jaegeuk Kim 已提交
826 827 828 829 830
			goto fail;
		}
		nid[i + 1] = get_nid(pages[i], offset[i + 1], false);
	}

831 832
	ra_node_pages(pages[idx], offset[idx + 1], NIDS_PER_BLOCK);

J
Jaegeuk Kim 已提交
833
	/* free direct nodes linked to a partial indirect node */
834
	for (i = offset[idx + 1]; i < NIDS_PER_BLOCK; i++) {
J
Jaegeuk Kim 已提交
835 836 837 838 839 840 841
		child_nid = get_nid(pages[idx], i, false);
		if (!child_nid)
			continue;
		dn->nid = child_nid;
		err = truncate_dnode(dn);
		if (err < 0)
			goto fail;
842 843
		if (set_nid(pages[idx], i, 0, false))
			dn->node_changed = true;
J
Jaegeuk Kim 已提交
844 845
	}

846
	if (offset[idx + 1] == 0) {
J
Jaegeuk Kim 已提交
847 848 849 850 851 852 853
		dn->node_page = pages[idx];
		dn->nid = nid[idx];
		truncate_node(dn);
	} else {
		f2fs_put_page(pages[idx], 1);
	}
	offset[idx]++;
854 855
	offset[idx + 1] = 0;
	idx--;
J
Jaegeuk Kim 已提交
856
fail:
857
	for (i = idx; i >= 0; i--)
J
Jaegeuk Kim 已提交
858
		f2fs_put_page(pages[i], 1);
859 860 861

	trace_f2fs_truncate_partial_nodes(dn->inode, nid, depth, err);

J
Jaegeuk Kim 已提交
862 863 864
	return err;
}

J
Jaegeuk Kim 已提交
865
/*
J
Jaegeuk Kim 已提交
866 867 868 869
 * All the block addresses of data and nodes should be nullified.
 */
int truncate_inode_blocks(struct inode *inode, pgoff_t from)
{
870
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
J
Jaegeuk Kim 已提交
871 872
	int err = 0, cont = 1;
	int level, offset[4], noffset[4];
J
Jaegeuk Kim 已提交
873
	unsigned int nofs = 0;
874
	struct f2fs_inode *ri;
J
Jaegeuk Kim 已提交
875 876 877
	struct dnode_of_data dn;
	struct page *page;

878 879
	trace_f2fs_truncate_inode_blocks_enter(inode, from);

880
	level = get_node_path(inode, from, offset, noffset);
881

J
Jaegeuk Kim 已提交
882
	page = get_node_page(sbi, inode->i_ino);
883 884
	if (IS_ERR(page)) {
		trace_f2fs_truncate_inode_blocks_exit(inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
885
		return PTR_ERR(page);
886
	}
J
Jaegeuk Kim 已提交
887 888 889 890

	set_new_dnode(&dn, inode, page, NULL, 0);
	unlock_page(page);

891
	ri = F2FS_INODE(page);
J
Jaegeuk Kim 已提交
892 893 894 895 896 897 898 899 900
	switch (level) {
	case 0:
	case 1:
		nofs = noffset[1];
		break;
	case 2:
		nofs = noffset[1];
		if (!offset[level - 1])
			goto skip_partial;
901
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
902 903 904 905 906 907 908 909
		if (err < 0 && err != -ENOENT)
			goto fail;
		nofs += 1 + NIDS_PER_BLOCK;
		break;
	case 3:
		nofs = 5 + 2 * NIDS_PER_BLOCK;
		if (!offset[level - 1])
			goto skip_partial;
910
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
911 912 913 914 915 916 917 918 919
		if (err < 0 && err != -ENOENT)
			goto fail;
		break;
	default:
		BUG();
	}

skip_partial:
	while (cont) {
920
		dn.nid = le32_to_cpu(ri->i_nid[offset[0] - NODE_DIR1_BLOCK]);
J
Jaegeuk Kim 已提交
921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942
		switch (offset[0]) {
		case NODE_DIR1_BLOCK:
		case NODE_DIR2_BLOCK:
			err = truncate_dnode(&dn);
			break;

		case NODE_IND1_BLOCK:
		case NODE_IND2_BLOCK:
			err = truncate_nodes(&dn, nofs, offset[1], 2);
			break;

		case NODE_DIND_BLOCK:
			err = truncate_nodes(&dn, nofs, offset[1], 3);
			cont = 0;
			break;

		default:
			BUG();
		}
		if (err < 0 && err != -ENOENT)
			goto fail;
		if (offset[1] == 0 &&
943
				ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) {
J
Jaegeuk Kim 已提交
944
			lock_page(page);
945
			BUG_ON(page->mapping != NODE_MAPPING(sbi));
946
			f2fs_wait_on_page_writeback(page, NODE, true);
947
			ri->i_nid[offset[0] - NODE_DIR1_BLOCK] = 0;
J
Jaegeuk Kim 已提交
948 949 950 951 952 953 954 955 956
			set_page_dirty(page);
			unlock_page(page);
		}
		offset[1] = 0;
		offset[0]++;
		nofs += err;
	}
fail:
	f2fs_put_page(page, 0);
957
	trace_f2fs_truncate_inode_blocks_exit(inode, err);
J
Jaegeuk Kim 已提交
958 959 960
	return err > 0 ? 0 : err;
}

961 962
int truncate_xattr_node(struct inode *inode, struct page *page)
{
963
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
964 965 966 967 968 969 970 971 972 973 974
	nid_t nid = F2FS_I(inode)->i_xattr_nid;
	struct dnode_of_data dn;
	struct page *npage;

	if (!nid)
		return 0;

	npage = get_node_page(sbi, nid);
	if (IS_ERR(npage))
		return PTR_ERR(npage);

975
	f2fs_i_xnid_write(inode, 0);
J
Jaegeuk Kim 已提交
976

977 978 979
	set_new_dnode(&dn, inode, page, npage, nid);

	if (page)
980
		dn.inode_page_locked = true;
981 982 983 984
	truncate_node(&dn);
	return 0;
}

985
/*
C
Chao Yu 已提交
986 987
 * Caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
988
 */
C
Chao Yu 已提交
989
int remove_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
990 991
{
	struct dnode_of_data dn;
C
Chao Yu 已提交
992
	int err;
J
Jaegeuk Kim 已提交
993

994
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
C
Chao Yu 已提交
995 996 997
	err = get_dnode_of_data(&dn, 0, LOOKUP_NODE);
	if (err)
		return err;
J
Jaegeuk Kim 已提交
998

C
Chao Yu 已提交
999 1000
	err = truncate_xattr_node(inode, dn.inode_page);
	if (err) {
1001
		f2fs_put_dnode(&dn);
C
Chao Yu 已提交
1002
		return err;
J
Jaegeuk Kim 已提交
1003
	}
1004 1005 1006 1007 1008 1009

	/* remove potential inline_data blocks */
	if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
				S_ISLNK(inode->i_mode))
		truncate_data_blocks_range(&dn, 1);

A
arter97 已提交
1010
	/* 0 is possible, after f2fs_new_inode() has failed */
1011 1012
	f2fs_bug_on(F2FS_I_SB(inode),
			inode->i_blocks != 0 && inode->i_blocks != 1);
1013 1014

	/* will put inode & node pages */
1015
	truncate_node(&dn);
C
Chao Yu 已提交
1016
	return 0;
J
Jaegeuk Kim 已提交
1017 1018
}

1019
struct page *new_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
1020 1021 1022 1023 1024
{
	struct dnode_of_data dn;

	/* allocate inode page for new inode */
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
1025 1026

	/* caller should f2fs_put_page(page, 1); */
1027
	return new_node_page(&dn, 0, NULL);
J
Jaegeuk Kim 已提交
1028 1029
}

1030 1031
struct page *new_node_page(struct dnode_of_data *dn,
				unsigned int ofs, struct page *ipage)
J
Jaegeuk Kim 已提交
1032
{
1033
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
1034
	struct node_info new_ni;
J
Jaegeuk Kim 已提交
1035 1036 1037
	struct page *page;
	int err;

1038
	if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC)))
J
Jaegeuk Kim 已提交
1039 1040
		return ERR_PTR(-EPERM);

1041
	page = f2fs_grab_cache_page(NODE_MAPPING(sbi), dn->nid, false);
J
Jaegeuk Kim 已提交
1042 1043 1044
	if (!page)
		return ERR_PTR(-ENOMEM);

1045
	if (unlikely(!inc_valid_node_count(sbi, dn->inode))) {
1046 1047 1048
		err = -ENOSPC;
		goto fail;
	}
1049 1050 1051 1052 1053
#ifdef CONFIG_F2FS_CHECK_FS
	get_node_info(sbi, dn->nid, &new_ni);
	f2fs_bug_on(sbi, new_ni.blk_addr != NULL_ADDR);
#endif
	new_ni.nid = dn->nid;
J
Jaegeuk Kim 已提交
1054
	new_ni.ino = dn->inode->i_ino;
1055 1056 1057
	new_ni.blk_addr = NULL_ADDR;
	new_ni.flag = 0;
	new_ni.version = 0;
1058
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
1059

1060
	f2fs_wait_on_page_writeback(page, NODE, true);
1061
	fill_node_footer(page, dn->nid, dn->inode->i_ino, ofs, true);
1062
	set_cold_node(dn->inode, page);
1063 1064
	if (!PageUptodate(page))
		SetPageUptodate(page);
1065 1066
	if (set_page_dirty(page))
		dn->node_changed = true;
J
Jaegeuk Kim 已提交
1067

1068
	if (f2fs_has_xattr_block(ofs))
1069
		f2fs_i_xnid_write(dn->inode, dn->nid);
1070

J
Jaegeuk Kim 已提交
1071 1072 1073 1074 1075
	if (ofs == 0)
		inc_valid_inode_count(sbi);
	return page;

fail:
1076
	clear_node_page_dirty(page);
J
Jaegeuk Kim 已提交
1077 1078 1079 1080
	f2fs_put_page(page, 1);
	return ERR_PTR(err);
}

1081 1082 1083
/*
 * Caller should do after getting the following values.
 * 0: f2fs_put_page(page, 0)
1084
 * LOCKED_PAGE or error: f2fs_put_page(page, 1)
1085
 */
M
Mike Christie 已提交
1086
static int read_node_page(struct page *page, int op_flags)
J
Jaegeuk Kim 已提交
1087
{
1088
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1089
	struct node_info ni;
1090
	struct f2fs_io_info fio = {
1091
		.sbi = sbi,
1092
		.type = NODE,
M
Mike Christie 已提交
1093 1094
		.op = REQ_OP_READ,
		.op_flags = op_flags,
1095
		.page = page,
1096
		.encrypted_page = NULL,
1097
	};
J
Jaegeuk Kim 已提交
1098

1099 1100 1101
	if (PageUptodate(page))
		return LOCKED_PAGE;

J
Jaegeuk Kim 已提交
1102 1103
	get_node_info(sbi, page->index, &ni);

1104
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1105
		ClearPageUptodate(page);
J
Jaegeuk Kim 已提交
1106
		return -ENOENT;
1107 1108
	}

1109
	fio.new_blkaddr = fio.old_blkaddr = ni.blk_addr;
1110
	return f2fs_submit_page_bio(&fio);
J
Jaegeuk Kim 已提交
1111 1112
}

J
Jaegeuk Kim 已提交
1113
/*
J
Jaegeuk Kim 已提交
1114 1115 1116 1117 1118
 * Readahead a node page
 */
void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid)
{
	struct page *apage;
1119
	int err;
J
Jaegeuk Kim 已提交
1120

1121 1122 1123 1124
	if (!nid)
		return;
	f2fs_bug_on(sbi, check_nid_range(sbi, nid));

1125 1126 1127 1128
	rcu_read_lock();
	apage = radix_tree_lookup(&NODE_MAPPING(sbi)->page_tree, nid);
	rcu_read_unlock();
	if (apage)
1129
		return;
J
Jaegeuk Kim 已提交
1130

1131
	apage = f2fs_grab_cache_page(NODE_MAPPING(sbi), nid, false);
J
Jaegeuk Kim 已提交
1132 1133 1134
	if (!apage)
		return;

1135
	err = read_node_page(apage, REQ_RAHEAD);
1136
	f2fs_put_page(apage, err ? 1 : 0);
J
Jaegeuk Kim 已提交
1137 1138
}

J
Jaegeuk Kim 已提交
1139
static struct page *__get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid,
1140
					struct page *parent, int start)
J
Jaegeuk Kim 已提交
1141 1142
{
	struct page *page;
1143
	int err;
J
Jaegeuk Kim 已提交
1144 1145 1146

	if (!nid)
		return ERR_PTR(-ENOENT);
1147
	f2fs_bug_on(sbi, check_nid_range(sbi, nid));
1148
repeat:
1149
	page = f2fs_grab_cache_page(NODE_MAPPING(sbi), nid, false);
J
Jaegeuk Kim 已提交
1150 1151 1152
	if (!page)
		return ERR_PTR(-ENOMEM);

1153
	err = read_node_page(page, 0);
1154 1155
	if (err < 0) {
		f2fs_put_page(page, 1);
J
Jaegeuk Kim 已提交
1156
		return ERR_PTR(err);
1157
	} else if (err == LOCKED_PAGE) {
1158
		goto page_hit;
1159
	}
J
Jaegeuk Kim 已提交
1160

1161
	if (parent)
1162
		ra_node_pages(parent, start + 1, MAX_RA_NODE);
1163

J
Jaegeuk Kim 已提交
1164
	lock_page(page);
1165

1166
	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
1167 1168 1169
		f2fs_put_page(page, 1);
		goto repeat;
	}
1170 1171 1172

	if (unlikely(!PageUptodate(page)))
		goto out_err;
1173
page_hit:
1174 1175 1176 1177 1178 1179 1180
	if(unlikely(nid != nid_of_node(page))) {
		f2fs_bug_on(sbi, 1);
		ClearPageUptodate(page);
out_err:
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
	}
J
Jaegeuk Kim 已提交
1181 1182 1183
	return page;
}

1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
struct page *get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid)
{
	return __get_node_page(sbi, nid, NULL, 0);
}

struct page *get_node_page_ra(struct page *parent, int start)
{
	struct f2fs_sb_info *sbi = F2FS_P_SB(parent);
	nid_t nid = get_nid(parent, start, false);

	return __get_node_page(sbi, nid, parent, start);
}

1197 1198 1199 1200
static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino)
{
	struct inode *inode;
	struct page *page;
1201
	int ret;
1202 1203 1204 1205 1206 1207

	/* should flush inline_data before evict_inode */
	inode = ilookup(sbi->sb, ino);
	if (!inode)
		return;

1208
	page = pagecache_get_page(inode->i_mapping, 0, FGP_LOCK|FGP_NOWAIT, 0);
1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220
	if (!page)
		goto iput_out;

	if (!PageUptodate(page))
		goto page_out;

	if (!PageDirty(page))
		goto page_out;

	if (!clear_page_dirty_for_io(page))
		goto page_out;

1221 1222
	ret = f2fs_write_inline_data(inode, page);
	inode_dec_dirty_pages(inode);
1223
	remove_dirty_inode(inode);
1224
	if (ret)
1225 1226
		set_page_dirty(page);
page_out:
1227
	f2fs_put_page(page, 1);
1228 1229 1230 1231
iput_out:
	iput(inode);
}

1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
void move_node_page(struct page *node_page, int gc_type)
{
	if (gc_type == FG_GC) {
		struct f2fs_sb_info *sbi = F2FS_P_SB(node_page);
		struct writeback_control wbc = {
			.sync_mode = WB_SYNC_ALL,
			.nr_to_write = 1,
			.for_reclaim = 0,
		};

		set_page_dirty(node_page);
		f2fs_wait_on_page_writeback(node_page, NODE, true);

		f2fs_bug_on(sbi, PageWriteback(node_page));
		if (!clear_page_dirty_for_io(node_page))
			goto out_page;

		if (NODE_MAPPING(sbi)->a_ops->writepage(node_page, &wbc))
			unlock_page(node_page);
		goto release_page;
	} else {
		/* set page dirty and write it */
		if (!PageWriteback(node_page))
			set_page_dirty(node_page);
	}
out_page:
	unlock_page(node_page);
release_page:
	f2fs_put_page(node_page, 0);
}

1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322
static struct page *last_fsync_dnode(struct f2fs_sb_info *sbi, nid_t ino)
{
	pgoff_t index, end;
	struct pagevec pvec;
	struct page *last_page = NULL;

	pagevec_init(&pvec, 0);
	index = 0;
	end = ULONG_MAX;

	while (index <= end) {
		int i, nr_pages;
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
				PAGECACHE_TAG_DIRTY,
				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
		if (nr_pages == 0)
			break;

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];

			if (unlikely(f2fs_cp_error(sbi))) {
				f2fs_put_page(last_page, 0);
				pagevec_release(&pvec);
				return ERR_PTR(-EIO);
			}

			if (!IS_DNODE(page) || !is_cold_node(page))
				continue;
			if (ino_of_node(page) != ino)
				continue;

			lock_page(page);

			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
continue_unlock:
				unlock_page(page);
				continue;
			}
			if (ino_of_node(page) != ino)
				goto continue_unlock;

			if (!PageDirty(page)) {
				/* someone wrote it for us */
				goto continue_unlock;
			}

			if (last_page)
				f2fs_put_page(last_page, 0);

			get_page(page);
			last_page = page;
			unlock_page(page);
		}
		pagevec_release(&pvec);
		cond_resched();
	}
	return last_page;
}

1323
static int __write_node_page(struct page *page, bool atomic, bool *submitted,
1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335
				struct writeback_control *wbc)
{
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
	nid_t nid;
	struct node_info ni;
	struct f2fs_io_info fio = {
		.sbi = sbi,
		.type = NODE,
		.op = REQ_OP_WRITE,
		.op_flags = wbc_to_write_flags(wbc),
		.page = page,
		.encrypted_page = NULL,
1336
		.submitted = false,
1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
	};

	trace_f2fs_writepage(page, NODE);

	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
		goto redirty_out;
	if (unlikely(f2fs_cp_error(sbi)))
		goto redirty_out;

	/* get old block addr of this node page */
	nid = nid_of_node(page);
	f2fs_bug_on(sbi, page->index != nid);

	if (wbc->for_reclaim) {
		if (!down_read_trylock(&sbi->node_write))
			goto redirty_out;
	} else {
		down_read(&sbi->node_write);
	}

	get_node_info(sbi, nid, &ni);

	/* This page is already truncated */
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
		ClearPageUptodate(page);
		dec_page_count(sbi, F2FS_DIRTY_NODES);
		up_read(&sbi->node_write);
		unlock_page(page);
		return 0;
	}

1368 1369 1370
	if (atomic && !test_opt(sbi, NOBARRIER))
		fio.op_flags |= REQ_PREFLUSH | REQ_FUA;

1371 1372 1373 1374 1375 1376 1377
	set_page_writeback(page);
	fio.old_blkaddr = ni.blk_addr;
	write_node_page(nid, &fio);
	set_node_addr(sbi, &ni, fio.new_blkaddr, is_fsync_dnode(page));
	dec_page_count(sbi, F2FS_DIRTY_NODES);
	up_read(&sbi->node_write);

1378
	if (wbc->for_reclaim) {
1379 1380
		f2fs_submit_merged_bio_cond(sbi, page->mapping->host, 0,
						page->index, NODE, WRITE);
1381 1382
		submitted = NULL;
	}
1383 1384 1385

	unlock_page(page);

1386
	if (unlikely(f2fs_cp_error(sbi))) {
1387
		f2fs_submit_merged_bio(sbi, NODE, WRITE);
1388 1389 1390 1391
		submitted = NULL;
	}
	if (submitted)
		*submitted = fio.submitted;
1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402

	return 0;

redirty_out:
	redirty_page_for_writepage(wbc, page);
	return AOP_WRITEPAGE_ACTIVATE;
}

static int f2fs_write_node_page(struct page *page,
				struct writeback_control *wbc)
{
1403
	return __write_node_page(page, false, NULL, wbc);
1404 1405
}

1406
int fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
1407
			struct writeback_control *wbc, bool atomic)
J
Jaegeuk Kim 已提交
1408 1409
{
	pgoff_t index, end;
1410
	pgoff_t last_idx = ULONG_MAX;
J
Jaegeuk Kim 已提交
1411
	struct pagevec pvec;
1412
	int ret = 0;
1413 1414
	struct page *last_page = NULL;
	bool marked = false;
1415
	nid_t ino = inode->i_ino;
1416

1417 1418 1419 1420 1421 1422
	if (atomic) {
		last_page = last_fsync_dnode(sbi, ino);
		if (IS_ERR_OR_NULL(last_page))
			return PTR_ERR_OR_ZERO(last_page);
	}
retry:
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436
	pagevec_init(&pvec, 0);
	index = 0;
	end = ULONG_MAX;

	while (index <= end) {
		int i, nr_pages;
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
				PAGECACHE_TAG_DIRTY,
				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
		if (nr_pages == 0)
			break;

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];
1437
			bool submitted = false;
1438 1439

			if (unlikely(f2fs_cp_error(sbi))) {
1440
				f2fs_put_page(last_page, 0);
1441
				pagevec_release(&pvec);
1442 1443
				ret = -EIO;
				goto out;
1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460
			}

			if (!IS_DNODE(page) || !is_cold_node(page))
				continue;
			if (ino_of_node(page) != ino)
				continue;

			lock_page(page);

			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
continue_unlock:
				unlock_page(page);
				continue;
			}
			if (ino_of_node(page) != ino)
				goto continue_unlock;

1461
			if (!PageDirty(page) && page != last_page) {
1462 1463 1464 1465 1466 1467 1468
				/* someone wrote it for us */
				goto continue_unlock;
			}

			f2fs_wait_on_page_writeback(page, NODE, true);
			BUG_ON(PageWriteback(page));

1469 1470
			if (!atomic || page == last_page) {
				set_fsync_mark(page, 1);
1471 1472 1473 1474
				if (IS_INODE(page)) {
					if (is_inode_flag_set(inode,
								FI_DIRTY_INODE))
						update_inode(inode, page);
1475
					set_dentry_mark(page,
1476
						need_dentry_mark(sbi, ino));
1477
				}
1478 1479 1480 1481 1482 1483 1484
				/*  may be written by other thread */
				if (!PageDirty(page))
					set_page_dirty(page);
			}

			if (!clear_page_dirty_for_io(page))
				goto continue_unlock;
1485

1486
			ret = __write_node_page(page, atomic &&
1487 1488
						page == last_page,
						&submitted, wbc);
1489
			if (ret) {
1490
				unlock_page(page);
1491 1492
				f2fs_put_page(last_page, 0);
				break;
1493
			} else if (submitted) {
1494
				last_idx = page->index;
1495
			}
1496

1497 1498 1499
			if (page == last_page) {
				f2fs_put_page(page, 0);
				marked = true;
1500
				break;
1501
			}
1502 1503 1504 1505
		}
		pagevec_release(&pvec);
		cond_resched();

1506
		if (ret || marked)
1507 1508
			break;
	}
1509 1510 1511 1512 1513
	if (!ret && atomic && !marked) {
		f2fs_msg(sbi->sb, KERN_DEBUG,
			"Retry to write fsync mark: ino=%u, idx=%lx",
					ino, last_page->index);
		lock_page(last_page);
1514
		f2fs_wait_on_page_writeback(last_page, NODE, true);
1515 1516 1517 1518
		set_page_dirty(last_page);
		unlock_page(last_page);
		goto retry;
	}
1519
out:
1520 1521 1522
	if (last_idx != ULONG_MAX)
		f2fs_submit_merged_bio_cond(sbi, NULL, ino, last_idx,
							NODE, WRITE);
1523
	return ret ? -EIO: 0;
1524 1525 1526 1527 1528 1529 1530
}

int sync_node_pages(struct f2fs_sb_info *sbi, struct writeback_control *wbc)
{
	pgoff_t index, end;
	struct pagevec pvec;
	int step = 0;
1531
	int nwritten = 0;
1532
	int ret = 0;
J
Jaegeuk Kim 已提交
1533 1534 1535 1536 1537

	pagevec_init(&pvec, 0);

next_step:
	index = 0;
1538
	end = ULONG_MAX;
J
Jaegeuk Kim 已提交
1539 1540 1541

	while (index <= end) {
		int i, nr_pages;
1542
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
J
Jaegeuk Kim 已提交
1543 1544 1545 1546 1547 1548 1549
				PAGECACHE_TAG_DIRTY,
				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
		if (nr_pages == 0)
			break;

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];
1550
			bool submitted = false;
J
Jaegeuk Kim 已提交
1551

C
Chao Yu 已提交
1552 1553
			if (unlikely(f2fs_cp_error(sbi))) {
				pagevec_release(&pvec);
1554 1555
				ret = -EIO;
				goto out;
C
Chao Yu 已提交
1556 1557
			}

J
Jaegeuk Kim 已提交
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571
			/*
			 * flushing sequence with step:
			 * 0. indirect nodes
			 * 1. dentry dnodes
			 * 2. file dnodes
			 */
			if (step == 0 && IS_DNODE(page))
				continue;
			if (step == 1 && (!IS_DNODE(page) ||
						is_cold_node(page)))
				continue;
			if (step == 2 && (!IS_DNODE(page) ||
						!is_cold_node(page)))
				continue;
1572
lock_node:
1573
			if (!trylock_page(page))
J
Jaegeuk Kim 已提交
1574 1575
				continue;

1576
			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
J
Jaegeuk Kim 已提交
1577 1578 1579 1580 1581 1582 1583 1584 1585 1586
continue_unlock:
				unlock_page(page);
				continue;
			}

			if (!PageDirty(page)) {
				/* someone wrote it for us */
				goto continue_unlock;
			}

1587
			/* flush inline_data */
1588
			if (is_inline_node(page)) {
1589 1590 1591
				clear_inline_node(page);
				unlock_page(page);
				flush_inline_data(sbi, ino_of_node(page));
1592
				goto lock_node;
1593 1594
			}

1595 1596 1597
			f2fs_wait_on_page_writeback(page, NODE, true);

			BUG_ON(PageWriteback(page));
J
Jaegeuk Kim 已提交
1598 1599 1600
			if (!clear_page_dirty_for_io(page))
				goto continue_unlock;

1601 1602
			set_fsync_mark(page, 0);
			set_dentry_mark(page, 0);
1603

1604 1605
			ret = __write_node_page(page, false, &submitted, wbc);
			if (ret)
1606
				unlock_page(page);
1607
			else if (submitted)
1608
				nwritten++;
J
Jaegeuk Kim 已提交
1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625

			if (--wbc->nr_to_write == 0)
				break;
		}
		pagevec_release(&pvec);
		cond_resched();

		if (wbc->nr_to_write == 0) {
			step = 2;
			break;
		}
	}

	if (step < 2) {
		step++;
		goto next_step;
	}
1626 1627 1628 1629
out:
	if (nwritten)
		f2fs_submit_merged_bio(sbi, NODE, WRITE);
	return ret;
J
Jaegeuk Kim 已提交
1630 1631
}

1632 1633
int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
{
1634
	pgoff_t index = 0, end = ULONG_MAX;
1635
	struct pagevec pvec;
1636
	int ret2, ret = 0;
1637 1638

	pagevec_init(&pvec, 0);
1639 1640 1641 1642 1643 1644 1645 1646

	while (index <= end) {
		int i, nr_pages;
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
				PAGECACHE_TAG_WRITEBACK,
				min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
		if (nr_pages == 0)
			break;
1647 1648 1649 1650 1651

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];

			/* until radix tree lookup accepts end_index */
1652
			if (unlikely(page->index > end))
1653 1654
				continue;

1655
			if (ino && ino_of_node(page) == ino) {
1656
				f2fs_wait_on_page_writeback(page, NODE, true);
1657 1658 1659
				if (TestClearPageError(page))
					ret = -EIO;
			}
1660 1661 1662 1663 1664
		}
		pagevec_release(&pvec);
		cond_resched();
	}

1665
	ret2 = filemap_check_errors(NODE_MAPPING(sbi));
1666 1667 1668 1669 1670
	if (!ret)
		ret = ret2;
	return ret;
}

J
Jaegeuk Kim 已提交
1671 1672 1673
static int f2fs_write_node_pages(struct address_space *mapping,
			    struct writeback_control *wbc)
{
1674
	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
1675
	struct blk_plug plug;
1676
	long diff;
J
Jaegeuk Kim 已提交
1677

1678 1679
	/* balancing f2fs's metadata in background */
	f2fs_balance_fs_bg(sbi);
J
Jaegeuk Kim 已提交
1680

1681
	/* collect a number of dirty node pages and write together */
1682
	if (get_pages(sbi, F2FS_DIRTY_NODES) < nr_pages_to_skip(sbi, NODE))
1683
		goto skip_write;
1684

Y
Yunlei He 已提交
1685 1686
	trace_f2fs_writepages(mapping->host, wbc, NODE);

1687
	diff = nr_pages_to_write(sbi, NODE, wbc);
1688
	wbc->sync_mode = WB_SYNC_NONE;
1689
	blk_start_plug(&plug);
1690
	sync_node_pages(sbi, wbc);
1691
	blk_finish_plug(&plug);
1692
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
J
Jaegeuk Kim 已提交
1693
	return 0;
1694 1695 1696

skip_write:
	wbc->pages_skipped += get_pages(sbi, F2FS_DIRTY_NODES);
Y
Yunlei He 已提交
1697
	trace_f2fs_writepages(mapping->host, wbc, NODE);
1698
	return 0;
J
Jaegeuk Kim 已提交
1699 1700 1701 1702
}

static int f2fs_set_node_page_dirty(struct page *page)
{
1703 1704
	trace_f2fs_set_page_dirty(page, NODE);

1705 1706
	if (!PageUptodate(page))
		SetPageUptodate(page);
J
Jaegeuk Kim 已提交
1707
	if (!PageDirty(page)) {
1708
		f2fs_set_page_dirty_nobuffers(page);
1709
		inc_page_count(F2FS_P_SB(page), F2FS_DIRTY_NODES);
J
Jaegeuk Kim 已提交
1710
		SetPagePrivate(page);
J
Jaegeuk Kim 已提交
1711
		f2fs_trace_pid(page);
J
Jaegeuk Kim 已提交
1712 1713 1714 1715 1716
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
1717
/*
J
Jaegeuk Kim 已提交
1718 1719 1720 1721 1722 1723
 * Structure of the f2fs node operations
 */
const struct address_space_operations f2fs_node_aops = {
	.writepage	= f2fs_write_node_page,
	.writepages	= f2fs_write_node_pages,
	.set_page_dirty	= f2fs_set_node_page_dirty,
1724 1725
	.invalidatepage	= f2fs_invalidate_page,
	.releasepage	= f2fs_release_page,
1726 1727 1728
#ifdef CONFIG_MIGRATION
	.migratepage    = f2fs_migrate_page,
#endif
J
Jaegeuk Kim 已提交
1729 1730
};

1731 1732
static struct free_nid *__lookup_free_nid_list(struct f2fs_nm_info *nm_i,
						nid_t n)
J
Jaegeuk Kim 已提交
1733
{
1734
	return radix_tree_lookup(&nm_i->free_nid_root, n);
J
Jaegeuk Kim 已提交
1735 1736
}

1737 1738
static int __insert_nid_to_list(struct f2fs_sb_info *sbi,
			struct free_nid *i, enum nid_list list, bool new)
J
Jaegeuk Kim 已提交
1739
{
C
Chao Yu 已提交
1740 1741
	struct f2fs_nm_info *nm_i = NM_I(sbi);

1742 1743 1744 1745 1746 1747
	if (new) {
		int err = radix_tree_insert(&nm_i->free_nid_root, i->nid, i);
		if (err)
			return err;
	}

C
Chao Yu 已提交
1748 1749 1750 1751
	f2fs_bug_on(sbi, list == FREE_NID_LIST ? i->state != NID_NEW :
						i->state != NID_ALLOC);
	nm_i->nid_cnt[list]++;
	list_add_tail(&i->list, &nm_i->nid_list[list]);
1752
	return 0;
C
Chao Yu 已提交
1753 1754 1755
}

static void __remove_nid_from_list(struct f2fs_sb_info *sbi,
1756
			struct free_nid *i, enum nid_list list, bool reuse)
C
Chao Yu 已提交
1757 1758 1759 1760 1761 1762
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);

	f2fs_bug_on(sbi, list == FREE_NID_LIST ? i->state != NID_NEW :
						i->state != NID_ALLOC);
	nm_i->nid_cnt[list]--;
J
Jaegeuk Kim 已提交
1763
	list_del(&i->list);
1764 1765
	if (!reuse)
		radix_tree_delete(&nm_i->free_nid_root, i->nid);
J
Jaegeuk Kim 已提交
1766 1767
}

C
Chao Yu 已提交
1768 1769
/* return if the nid is recognized as free */
static bool add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build)
J
Jaegeuk Kim 已提交
1770
{
1771
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1772
	struct free_nid *i;
1773
	struct nat_entry *ne;
1774
	int err;
1775 1776

	/* 0 nid should not be used */
1777
	if (unlikely(nid == 0))
C
Chao Yu 已提交
1778
		return false;
1779

1780 1781 1782
	if (build) {
		/* do not add allocated nids */
		ne = __lookup_nat_cache(nm_i, nid);
1783
		if (ne && (!get_nat_flag(ne, IS_CHECKPOINTED) ||
1784
				nat_get_blkaddr(ne) != NULL_ADDR))
C
Chao Yu 已提交
1785
			return false;
J
Jaegeuk Kim 已提交
1786
	}
1787 1788

	i = f2fs_kmem_cache_alloc(free_nid_slab, GFP_NOFS);
J
Jaegeuk Kim 已提交
1789 1790 1791
	i->nid = nid;
	i->state = NID_NEW;

1792 1793
	if (radix_tree_preload(GFP_NOFS)) {
		kmem_cache_free(free_nid_slab, i);
C
Chao Yu 已提交
1794
		return true;
1795 1796
	}

C
Chao Yu 已提交
1797
	spin_lock(&nm_i->nid_list_lock);
1798 1799 1800 1801
	err = __insert_nid_to_list(sbi, i, FREE_NID_LIST, true);
	spin_unlock(&nm_i->nid_list_lock);
	radix_tree_preload_end();
	if (err) {
J
Jaegeuk Kim 已提交
1802
		kmem_cache_free(free_nid_slab, i);
C
Chao Yu 已提交
1803
		return true;
J
Jaegeuk Kim 已提交
1804
	}
C
Chao Yu 已提交
1805
	return true;
J
Jaegeuk Kim 已提交
1806 1807
}

C
Chao Yu 已提交
1808
static void remove_free_nid(struct f2fs_sb_info *sbi, nid_t nid)
J
Jaegeuk Kim 已提交
1809
{
C
Chao Yu 已提交
1810
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1811
	struct free_nid *i;
1812 1813
	bool need_free = false;

C
Chao Yu 已提交
1814
	spin_lock(&nm_i->nid_list_lock);
1815
	i = __lookup_free_nid_list(nm_i, nid);
J
Jaegeuk Kim 已提交
1816
	if (i && i->state == NID_NEW) {
1817
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, false);
1818
		need_free = true;
J
Jaegeuk Kim 已提交
1819
	}
C
Chao Yu 已提交
1820
	spin_unlock(&nm_i->nid_list_lock);
1821 1822 1823

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1824 1825
}

J
Jaegeuk Kim 已提交
1826
static void update_free_nid_bitmap(struct f2fs_sb_info *sbi, nid_t nid,
1827
							bool set, bool build)
C
Chao Yu 已提交
1828 1829 1830 1831 1832 1833 1834 1835 1836
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	unsigned int nat_ofs = NAT_BLOCK_OFFSET(nid);
	unsigned int nid_ofs = nid - START_NID(nid);

	if (!test_bit_le(nat_ofs, nm_i->nat_block_bitmap))
		return;

	if (set)
J
Jaegeuk Kim 已提交
1837
		__set_bit_le(nid_ofs, nm_i->free_nid_bitmap[nat_ofs]);
C
Chao Yu 已提交
1838
	else
J
Jaegeuk Kim 已提交
1839
		__clear_bit_le(nid_ofs, nm_i->free_nid_bitmap[nat_ofs]);
1840 1841 1842 1843 1844 1845 1846

	spin_lock(&nm_i->free_nid_lock);
	if (set)
		nm_i->free_nid_count[nat_ofs]++;
	else if (!build)
		nm_i->free_nid_count[nat_ofs]--;
	spin_unlock(&nm_i->free_nid_lock);
C
Chao Yu 已提交
1847 1848
}

1849
static void scan_nat_page(struct f2fs_sb_info *sbi,
J
Jaegeuk Kim 已提交
1850 1851
			struct page *nat_page, nid_t start_nid)
{
1852
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1853 1854
	struct f2fs_nat_block *nat_blk = page_address(nat_page);
	block_t blk_addr;
C
Chao Yu 已提交
1855
	unsigned int nat_ofs = NAT_BLOCK_OFFSET(start_nid);
J
Jaegeuk Kim 已提交
1856 1857
	int i;

1858 1859 1860
	if (test_bit_le(nat_ofs, nm_i->nat_block_bitmap))
		return;

J
Jaegeuk Kim 已提交
1861
	__set_bit_le(nat_ofs, nm_i->nat_block_bitmap);
C
Chao Yu 已提交
1862

J
Jaegeuk Kim 已提交
1863 1864 1865
	i = start_nid % NAT_ENTRY_PER_BLOCK;

	for (; i < NAT_ENTRY_PER_BLOCK; i++, start_nid++) {
C
Chao Yu 已提交
1866
		bool freed = false;
H
Haicheng Li 已提交
1867

1868
		if (unlikely(start_nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1869
			break;
H
Haicheng Li 已提交
1870 1871

		blk_addr = le32_to_cpu(nat_blk->entries[i].block_addr);
1872
		f2fs_bug_on(sbi, blk_addr == NEW_ADDR);
1873
		if (blk_addr == NULL_ADDR)
C
Chao Yu 已提交
1874
			freed = add_free_nid(sbi, start_nid, true);
1875
		update_free_nid_bitmap(sbi, start_nid, freed, true);
C
Chao Yu 已提交
1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890
	}
}

static void scan_free_nid_bits(struct f2fs_sb_info *sbi)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
	struct f2fs_journal *journal = curseg->journal;
	unsigned int i, idx;

	down_read(&nm_i->nat_tree_lock);

	for (i = 0; i < nm_i->nat_blocks; i++) {
		if (!test_bit_le(i, nm_i->nat_block_bitmap))
			continue;
1891 1892
		if (!nm_i->free_nid_count[i])
			continue;
C
Chao Yu 已提交
1893 1894 1895 1896 1897 1898 1899 1900 1901
		for (idx = 0; idx < NAT_ENTRY_PER_BLOCK; idx++) {
			nid_t nid;

			if (!test_bit_le(idx, nm_i->free_nid_bitmap[i]))
				continue;

			nid = i * NAT_ENTRY_PER_BLOCK + idx;
			add_free_nid(sbi, nid, true);

1902
			if (nm_i->nid_cnt[FREE_NID_LIST] >= MAX_FREE_NIDS)
C
Chao Yu 已提交
1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917
				goto out;
		}
	}
out:
	down_read(&curseg->journal_rwsem);
	for (i = 0; i < nats_in_cursum(journal); i++) {
		block_t addr;
		nid_t nid;

		addr = le32_to_cpu(nat_in_journal(journal, i).block_addr);
		nid = le32_to_cpu(nid_in_journal(journal, i));
		if (addr == NULL_ADDR)
			add_free_nid(sbi, nid, true);
		else
			remove_free_nid(sbi, nid);
J
Jaegeuk Kim 已提交
1918
	}
C
Chao Yu 已提交
1919 1920
	up_read(&curseg->journal_rwsem);
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
1921 1922
}

1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947
static int scan_nat_bits(struct f2fs_sb_info *sbi)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct page *page;
	unsigned int i = 0;
	nid_t nid;

	if (!enabled_nat_bits(sbi, NULL))
		return -EAGAIN;

	down_read(&nm_i->nat_tree_lock);
check_empty:
	i = find_next_bit_le(nm_i->empty_nat_bits, nm_i->nat_blocks, i);
	if (i >= nm_i->nat_blocks) {
		i = 0;
		goto check_partial;
	}

	for (nid = i * NAT_ENTRY_PER_BLOCK; nid < (i + 1) * NAT_ENTRY_PER_BLOCK;
									nid++) {
		if (unlikely(nid >= nm_i->max_nid))
			break;
		add_free_nid(sbi, nid, true);
	}

1948
	if (nm_i->nid_cnt[FREE_NID_LIST] >= MAX_FREE_NIDS)
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
		goto out;
	i++;
	goto check_empty;

check_partial:
	i = find_next_zero_bit_le(nm_i->full_nat_bits, nm_i->nat_blocks, i);
	if (i >= nm_i->nat_blocks) {
		disable_nat_bits(sbi, true);
		up_read(&nm_i->nat_tree_lock);
		return -EINVAL;
	}

	nid = i * NAT_ENTRY_PER_BLOCK;
	page = get_current_nat_page(sbi, nid);
	scan_nat_page(sbi, page, nid);
	f2fs_put_page(page, 1);

1966
	if (nm_i->nid_cnt[FREE_NID_LIST] < MAX_FREE_NIDS) {
1967 1968 1969 1970 1971 1972 1973 1974 1975
		i++;
		goto check_partial;
	}
out:
	up_read(&nm_i->nat_tree_lock);
	return 0;
}

static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount)
J
Jaegeuk Kim 已提交
1976 1977 1978
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
1979
	struct f2fs_journal *journal = curseg->journal;
1980
	int i = 0;
1981
	nid_t nid = nm_i->next_scan_nid;
J
Jaegeuk Kim 已提交
1982

1983
	/* Enough entries */
C
Chao Yu 已提交
1984
	if (nm_i->nid_cnt[FREE_NID_LIST] >= NAT_ENTRY_PER_BLOCK)
1985
		return;
J
Jaegeuk Kim 已提交
1986

1987
	if (!sync && !available_free_memory(sbi, FREE_NIDS))
1988
		return;
J
Jaegeuk Kim 已提交
1989

C
Chao Yu 已提交
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000
	if (!mount) {
		/* try to find free nids in free_nid_bitmap */
		scan_free_nid_bits(sbi);

		if (nm_i->nid_cnt[FREE_NID_LIST])
			return;

		/* try to find free nids with nat_bits */
		if (!scan_nat_bits(sbi) && nm_i->nid_cnt[FREE_NID_LIST])
			return;
	}
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

	/* find next valid candidate */
	if (enabled_nat_bits(sbi, NULL)) {
		int idx = find_next_zero_bit_le(nm_i->full_nat_bits,
					nm_i->nat_blocks, 0);

		if (idx >= nm_i->nat_blocks)
			set_sbi_flag(sbi, SBI_NEED_FSCK);
		else
			nid = idx * NAT_ENTRY_PER_BLOCK;
	}

2013
	/* readahead nat pages to be scanned */
2014 2015
	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nid), FREE_NID_PAGES,
							META_NAT, true);
J
Jaegeuk Kim 已提交
2016

2017
	down_read(&nm_i->nat_tree_lock);
2018

J
Jaegeuk Kim 已提交
2019 2020 2021
	while (1) {
		struct page *page = get_current_nat_page(sbi, nid);

2022
		scan_nat_page(sbi, page, nid);
J
Jaegeuk Kim 已提交
2023 2024 2025
		f2fs_put_page(page, 1);

		nid += (NAT_ENTRY_PER_BLOCK - (nid % NAT_ENTRY_PER_BLOCK));
2026
		if (unlikely(nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
2027
			nid = 0;
2028

2029
		if (++i >= FREE_NID_PAGES)
J
Jaegeuk Kim 已提交
2030 2031 2032
			break;
	}

2033 2034
	/* go to the next free nat pages to find free nids abundantly */
	nm_i->next_scan_nid = nid;
J
Jaegeuk Kim 已提交
2035 2036

	/* find free nids from current sum_pages */
2037
	down_read(&curseg->journal_rwsem);
2038 2039 2040 2041 2042
	for (i = 0; i < nats_in_cursum(journal); i++) {
		block_t addr;

		addr = le32_to_cpu(nat_in_journal(journal, i).block_addr);
		nid = le32_to_cpu(nid_in_journal(journal, i));
J
Jaegeuk Kim 已提交
2043
		if (addr == NULL_ADDR)
2044
			add_free_nid(sbi, nid, true);
J
Jaegeuk Kim 已提交
2045
		else
C
Chao Yu 已提交
2046
			remove_free_nid(sbi, nid);
J
Jaegeuk Kim 已提交
2047
	}
2048
	up_read(&curseg->journal_rwsem);
2049
	up_read(&nm_i->nat_tree_lock);
C
Chao Yu 已提交
2050 2051

	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nm_i->next_scan_nid),
C
Chao Yu 已提交
2052
					nm_i->ra_nid_pages, META_NAT, false);
J
Jaegeuk Kim 已提交
2053 2054
}

2055
void build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount)
2056 2057
{
	mutex_lock(&NM_I(sbi)->build_lock);
2058
	__build_free_nids(sbi, sync, mount);
2059 2060 2061
	mutex_unlock(&NM_I(sbi)->build_lock);
}

J
Jaegeuk Kim 已提交
2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
/*
 * If this function returns success, caller can obtain a new nid
 * from second parameter of this function.
 * The returned nid could be used ino as well as nid when inode is created.
 */
bool alloc_nid(struct f2fs_sb_info *sbi, nid_t *nid)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i = NULL;
retry:
J
Jaegeuk Kim 已提交
2072
#ifdef CONFIG_F2FS_FAULT_INJECTION
2073 2074
	if (time_to_inject(sbi, FAULT_ALLOC_NID)) {
		f2fs_show_injection_info(FAULT_ALLOC_NID);
J
Jaegeuk Kim 已提交
2075
		return false;
2076
	}
J
Jaegeuk Kim 已提交
2077
#endif
C
Chao Yu 已提交
2078
	spin_lock(&nm_i->nid_list_lock);
J
Jaegeuk Kim 已提交
2079

2080 2081 2082 2083
	if (unlikely(nm_i->available_nids == 0)) {
		spin_unlock(&nm_i->nid_list_lock);
		return false;
	}
J
Jaegeuk Kim 已提交
2084

2085
	/* We should not use stale free nids created by build_free_nids */
C
Chao Yu 已提交
2086 2087 2088 2089
	if (nm_i->nid_cnt[FREE_NID_LIST] && !on_build_free_nids(nm_i)) {
		f2fs_bug_on(sbi, list_empty(&nm_i->nid_list[FREE_NID_LIST]));
		i = list_first_entry(&nm_i->nid_list[FREE_NID_LIST],
					struct free_nid, list);
2090
		*nid = i->nid;
C
Chao Yu 已提交
2091

2092
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, true);
2093
		i->state = NID_ALLOC;
2094
		__insert_nid_to_list(sbi, i, ALLOC_NID_LIST, false);
2095
		nm_i->available_nids--;
C
Chao Yu 已提交
2096

2097
		update_free_nid_bitmap(sbi, *nid, false, false);
C
Chao Yu 已提交
2098

C
Chao Yu 已提交
2099
		spin_unlock(&nm_i->nid_list_lock);
2100 2101
		return true;
	}
C
Chao Yu 已提交
2102
	spin_unlock(&nm_i->nid_list_lock);
2103 2104

	/* Let's scan nat pages and its caches to get free nids */
2105
	build_free_nids(sbi, true, false);
2106
	goto retry;
J
Jaegeuk Kim 已提交
2107 2108
}

J
Jaegeuk Kim 已提交
2109
/*
J
Jaegeuk Kim 已提交
2110 2111 2112 2113 2114 2115 2116
 * alloc_nid() should be called prior to this function.
 */
void alloc_nid_done(struct f2fs_sb_info *sbi, nid_t nid)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i;

C
Chao Yu 已提交
2117
	spin_lock(&nm_i->nid_list_lock);
2118
	i = __lookup_free_nid_list(nm_i, nid);
C
Chao Yu 已提交
2119
	f2fs_bug_on(sbi, !i);
2120
	__remove_nid_from_list(sbi, i, ALLOC_NID_LIST, false);
C
Chao Yu 已提交
2121
	spin_unlock(&nm_i->nid_list_lock);
2122 2123

	kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
2124 2125
}

J
Jaegeuk Kim 已提交
2126
/*
J
Jaegeuk Kim 已提交
2127 2128 2129 2130
 * alloc_nid() should be called prior to this function.
 */
void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid)
{
2131 2132
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i;
2133
	bool need_free = false;
2134

J
Jaegeuk Kim 已提交
2135 2136 2137
	if (!nid)
		return;

C
Chao Yu 已提交
2138
	spin_lock(&nm_i->nid_list_lock);
2139
	i = __lookup_free_nid_list(nm_i, nid);
C
Chao Yu 已提交
2140 2141
	f2fs_bug_on(sbi, !i);

2142
	if (!available_free_memory(sbi, FREE_NIDS)) {
2143
		__remove_nid_from_list(sbi, i, ALLOC_NID_LIST, false);
2144
		need_free = true;
2145
	} else {
2146
		__remove_nid_from_list(sbi, i, ALLOC_NID_LIST, true);
2147
		i->state = NID_NEW;
2148
		__insert_nid_to_list(sbi, i, FREE_NID_LIST, false);
2149
	}
2150 2151 2152

	nm_i->available_nids++;

2153
	update_free_nid_bitmap(sbi, nid, true, false);
C
Chao Yu 已提交
2154

C
Chao Yu 已提交
2155
	spin_unlock(&nm_i->nid_list_lock);
2156 2157 2158

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
2159 2160
}

C
Chao Yu 已提交
2161 2162 2163 2164 2165 2166
int try_to_free_nids(struct f2fs_sb_info *sbi, int nr_shrink)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i, *next;
	int nr = nr_shrink;

C
Chao Yu 已提交
2167
	if (nm_i->nid_cnt[FREE_NID_LIST] <= MAX_FREE_NIDS)
2168 2169
		return 0;

C
Chao Yu 已提交
2170 2171 2172
	if (!mutex_trylock(&nm_i->build_lock))
		return 0;

C
Chao Yu 已提交
2173 2174 2175 2176 2177
	spin_lock(&nm_i->nid_list_lock);
	list_for_each_entry_safe(i, next, &nm_i->nid_list[FREE_NID_LIST],
									list) {
		if (nr_shrink <= 0 ||
				nm_i->nid_cnt[FREE_NID_LIST] <= MAX_FREE_NIDS)
C
Chao Yu 已提交
2178
			break;
C
Chao Yu 已提交
2179

2180
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, false);
C
Chao Yu 已提交
2181 2182 2183
		kmem_cache_free(free_nid_slab, i);
		nr_shrink--;
	}
C
Chao Yu 已提交
2184
	spin_unlock(&nm_i->nid_list_lock);
C
Chao Yu 已提交
2185 2186 2187 2188 2189
	mutex_unlock(&nm_i->build_lock);

	return nr - nr_shrink;
}

2190
void recover_inline_xattr(struct inode *inode, struct page *page)
2191 2192 2193 2194 2195 2196
{
	void *src_addr, *dst_addr;
	size_t inline_size;
	struct page *ipage;
	struct f2fs_inode *ri;

2197
	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
2198
	f2fs_bug_on(F2FS_I_SB(inode), IS_ERR(ipage));
2199

2200 2201
	ri = F2FS_INODE(page);
	if (!(ri->i_inline & F2FS_INLINE_XATTR)) {
2202
		clear_inode_flag(inode, FI_INLINE_XATTR);
2203 2204 2205
		goto update_inode;
	}

2206 2207 2208 2209
	dst_addr = inline_xattr_addr(ipage);
	src_addr = inline_xattr_addr(page);
	inline_size = inline_xattr_size(inode);

2210
	f2fs_wait_on_page_writeback(ipage, NODE, true);
2211
	memcpy(dst_addr, src_addr, inline_size);
2212
update_inode:
2213 2214 2215 2216
	update_inode(inode, ipage);
	f2fs_put_page(ipage, 1);
}

2217
int recover_xattr_data(struct inode *inode, struct page *page, block_t blkaddr)
2218
{
2219
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2220 2221 2222
	nid_t prev_xnid = F2FS_I(inode)->i_xattr_nid;
	nid_t new_xnid = nid_of_node(page);
	struct node_info ni;
2223
	struct page *xpage;
2224 2225 2226 2227

	if (!prev_xnid)
		goto recover_xnid;

2228
	/* 1: invalidate the previous xattr nid */
2229
	get_node_info(sbi, prev_xnid, &ni);
2230
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
2231 2232
	invalidate_blocks(sbi, ni.blk_addr);
	dec_valid_node_count(sbi, inode);
2233
	set_node_addr(sbi, &ni, NULL_ADDR, false);
2234 2235

recover_xnid:
2236 2237 2238
	/* 2: update xattr nid in inode */
	remove_free_nid(sbi, new_xnid);
	f2fs_i_xnid_write(inode, new_xnid);
2239
	if (unlikely(!inc_valid_node_count(sbi, inode)))
2240
		f2fs_bug_on(sbi, 1);
2241 2242 2243 2244 2245 2246 2247 2248
	update_inode_page(inode);

	/* 3: update and set xattr node page dirty */
	xpage = grab_cache_page(NODE_MAPPING(sbi), new_xnid);
	if (!xpage)
		return -ENOMEM;

	memcpy(F2FS_NODE(xpage), F2FS_NODE(page), PAGE_SIZE);
2249 2250 2251

	get_node_info(sbi, new_xnid, &ni);
	ni.ino = inode->i_ino;
2252
	set_node_addr(sbi, &ni, NEW_ADDR, false);
2253 2254
	set_page_dirty(xpage);
	f2fs_put_page(xpage, 1);
2255

2256
	return 0;
2257 2258
}

J
Jaegeuk Kim 已提交
2259 2260
int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
{
2261
	struct f2fs_inode *src, *dst;
J
Jaegeuk Kim 已提交
2262 2263 2264 2265
	nid_t ino = ino_of_node(page);
	struct node_info old_ni, new_ni;
	struct page *ipage;

2266 2267 2268 2269
	get_node_info(sbi, ino, &old_ni);

	if (unlikely(old_ni.blk_addr != NULL_ADDR))
		return -EINVAL;
2270
retry:
2271
	ipage = f2fs_grab_cache_page(NODE_MAPPING(sbi), ino, false);
2272 2273 2274 2275
	if (!ipage) {
		congestion_wait(BLK_RW_ASYNC, HZ/50);
		goto retry;
	}
J
Jaegeuk Kim 已提交
2276

A
arter97 已提交
2277
	/* Should not use this inode from free nid list */
C
Chao Yu 已提交
2278
	remove_free_nid(sbi, ino);
J
Jaegeuk Kim 已提交
2279

2280 2281
	if (!PageUptodate(ipage))
		SetPageUptodate(ipage);
J
Jaegeuk Kim 已提交
2282 2283
	fill_node_footer(ipage, ino, ino, 0, true);

2284 2285
	src = F2FS_INODE(page);
	dst = F2FS_INODE(ipage);
J
Jaegeuk Kim 已提交
2286

2287 2288 2289 2290 2291
	memcpy(dst, src, (unsigned long)&src->i_ext - (unsigned long)src);
	dst->i_size = 0;
	dst->i_blocks = cpu_to_le64(1);
	dst->i_links = cpu_to_le32(1);
	dst->i_xattr_nid = 0;
2292
	dst->i_inline = src->i_inline & F2FS_INLINE_XATTR;
J
Jaegeuk Kim 已提交
2293 2294 2295 2296

	new_ni = old_ni;
	new_ni.ino = ino;

2297
	if (unlikely(!inc_valid_node_count(sbi, NULL)))
2298
		WARN_ON(1);
2299
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
J
Jaegeuk Kim 已提交
2300
	inc_valid_inode_count(sbi);
2301
	set_page_dirty(ipage);
J
Jaegeuk Kim 已提交
2302 2303 2304 2305 2306 2307 2308 2309 2310 2311
	f2fs_put_page(ipage, 1);
	return 0;
}

int restore_node_summary(struct f2fs_sb_info *sbi,
			unsigned int segno, struct f2fs_summary_block *sum)
{
	struct f2fs_node *rn;
	struct f2fs_summary *sum_entry;
	block_t addr;
2312
	int i, idx, last_offset, nrpages;
J
Jaegeuk Kim 已提交
2313 2314 2315 2316 2317 2318

	/* scan the node segment */
	last_offset = sbi->blocks_per_seg;
	addr = START_BLOCK(sbi, segno);
	sum_entry = &sum->entries[0];

2319
	for (i = 0; i < last_offset; i += nrpages, addr += nrpages) {
2320
		nrpages = min(last_offset - i, BIO_MAX_PAGES);
2321

A
arter97 已提交
2322
		/* readahead node pages */
2323
		ra_meta_pages(sbi, addr, nrpages, META_POR, true);
J
Jaegeuk Kim 已提交
2324

2325
		for (idx = addr; idx < addr + nrpages; idx++) {
2326
			struct page *page = get_tmp_page(sbi, idx);
2327

2328 2329 2330 2331 2332 2333
			rn = F2FS_NODE(page);
			sum_entry->nid = rn->footer.nid;
			sum_entry->version = 0;
			sum_entry->ofs_in_node = 0;
			sum_entry++;
			f2fs_put_page(page, 1);
2334
		}
2335

2336
		invalidate_mapping_pages(META_MAPPING(sbi), addr,
2337
							addr + nrpages);
J
Jaegeuk Kim 已提交
2338
	}
2339
	return 0;
J
Jaegeuk Kim 已提交
2340 2341
}

2342
static void remove_nats_in_journal(struct f2fs_sb_info *sbi)
J
Jaegeuk Kim 已提交
2343 2344 2345
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
2346
	struct f2fs_journal *journal = curseg->journal;
J
Jaegeuk Kim 已提交
2347 2348
	int i;

2349
	down_write(&curseg->journal_rwsem);
2350
	for (i = 0; i < nats_in_cursum(journal); i++) {
J
Jaegeuk Kim 已提交
2351 2352
		struct nat_entry *ne;
		struct f2fs_nat_entry raw_ne;
2353
		nid_t nid = le32_to_cpu(nid_in_journal(journal, i));
J
Jaegeuk Kim 已提交
2354

2355
		raw_ne = nat_in_journal(journal, i);
2356

J
Jaegeuk Kim 已提交
2357 2358
		ne = __lookup_nat_cache(nm_i, nid);
		if (!ne) {
2359
			ne = grab_nat_entry(nm_i, nid, true);
2360
			node_info_from_raw_nat(&ne->ni, &raw_ne);
J
Jaegeuk Kim 已提交
2361
		}
2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374

		/*
		 * if a free nat in journal has not been used after last
		 * checkpoint, we should remove it from available nids,
		 * since later we will add it again.
		 */
		if (!get_nat_flag(ne, IS_DIRTY) &&
				le32_to_cpu(raw_ne.block_addr) == NULL_ADDR) {
			spin_lock(&nm_i->nid_list_lock);
			nm_i->available_nids--;
			spin_unlock(&nm_i->nid_list_lock);
		}

J
Jaegeuk Kim 已提交
2375 2376
		__set_nat_cache_dirty(nm_i, ne);
	}
2377
	update_nats_in_cursum(journal, -i);
2378
	up_write(&curseg->journal_rwsem);
J
Jaegeuk Kim 已提交
2379 2380
}

2381 2382
static void __adjust_nat_entry_set(struct nat_entry_set *nes,
						struct list_head *head, int max)
J
Jaegeuk Kim 已提交
2383
{
2384
	struct nat_entry_set *cur;
J
Jaegeuk Kim 已提交
2385

2386 2387
	if (nes->entry_cnt >= max)
		goto add_out;
J
Jaegeuk Kim 已提交
2388

2389 2390 2391 2392 2393
	list_for_each_entry(cur, head, set_list) {
		if (cur->entry_cnt >= nes->entry_cnt) {
			list_add(&nes->set_list, cur->set_list.prev);
			return;
		}
2394
	}
2395 2396 2397
add_out:
	list_add_tail(&nes->set_list, head);
}
J
Jaegeuk Kim 已提交
2398

J
Jaegeuk Kim 已提交
2399
static void __update_nat_bits(struct f2fs_sb_info *sbi, nid_t start_nid,
2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417
						struct page *page)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	unsigned int nat_index = start_nid / NAT_ENTRY_PER_BLOCK;
	struct f2fs_nat_block *nat_blk = page_address(page);
	int valid = 0;
	int i;

	if (!enabled_nat_bits(sbi, NULL))
		return;

	for (i = 0; i < NAT_ENTRY_PER_BLOCK; i++) {
		if (start_nid == 0 && i == 0)
			valid++;
		if (nat_blk->entries[i].block_addr)
			valid++;
	}
	if (valid == 0) {
J
Jaegeuk Kim 已提交
2418 2419
		__set_bit_le(nat_index, nm_i->empty_nat_bits);
		__clear_bit_le(nat_index, nm_i->full_nat_bits);
2420 2421 2422
		return;
	}

J
Jaegeuk Kim 已提交
2423
	__clear_bit_le(nat_index, nm_i->empty_nat_bits);
2424
	if (valid == NAT_ENTRY_PER_BLOCK)
J
Jaegeuk Kim 已提交
2425
		__set_bit_le(nat_index, nm_i->full_nat_bits);
2426
	else
J
Jaegeuk Kim 已提交
2427
		__clear_bit_le(nat_index, nm_i->full_nat_bits);
2428 2429
}

2430
static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
2431
		struct nat_entry_set *set, struct cp_control *cpc)
2432 2433
{
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
2434
	struct f2fs_journal *journal = curseg->journal;
2435 2436 2437 2438 2439
	nid_t start_nid = set->set * NAT_ENTRY_PER_BLOCK;
	bool to_journal = true;
	struct f2fs_nat_block *nat_blk;
	struct nat_entry *ne, *cur;
	struct page *page = NULL;
J
Jaegeuk Kim 已提交
2440

2441 2442 2443 2444 2445
	/*
	 * there are two steps to flush nat entries:
	 * #1, flush nat entries to journal in current hot data summary block.
	 * #2, flush nat entries to nat page.
	 */
2446 2447
	if (enabled_nat_bits(sbi, cpc) ||
		!__has_cursum_space(journal, set->entry_cnt, NAT_JOURNAL))
2448 2449 2450
		to_journal = false;

	if (to_journal) {
2451
		down_write(&curseg->journal_rwsem);
2452 2453 2454 2455 2456
	} else {
		page = get_next_nat_page(sbi, start_nid);
		nat_blk = page_address(page);
		f2fs_bug_on(sbi, !nat_blk);
	}
2457

2458 2459 2460 2461 2462 2463 2464 2465
	/* flush dirty nats in nat entry set */
	list_for_each_entry_safe(ne, cur, &set->entry_list, list) {
		struct f2fs_nat_entry *raw_ne;
		nid_t nid = nat_get_nid(ne);
		int offset;

		if (nat_get_blkaddr(ne) == NEW_ADDR)
			continue;
2466 2467

		if (to_journal) {
2468
			offset = lookup_journal_in_cursum(journal,
2469 2470
							NAT_JOURNAL, nid, 1);
			f2fs_bug_on(sbi, offset < 0);
2471 2472
			raw_ne = &nat_in_journal(journal, offset);
			nid_in_journal(journal, offset) = cpu_to_le32(nid);
2473
		} else {
2474
			raw_ne = &nat_blk->entries[nid - start_nid];
J
Jaegeuk Kim 已提交
2475
		}
2476 2477 2478
		raw_nat_from_node_info(raw_ne, &ne->ni);
		nat_reset_flag(ne);
		__clear_nat_cache_dirty(NM_I(sbi), ne);
2479
		if (nat_get_blkaddr(ne) == NULL_ADDR) {
2480
			add_free_nid(sbi, nid, false);
2481 2482
			spin_lock(&NM_I(sbi)->nid_list_lock);
			NM_I(sbi)->available_nids++;
2483
			update_free_nid_bitmap(sbi, nid, true, false);
C
Chao Yu 已提交
2484 2485 2486
			spin_unlock(&NM_I(sbi)->nid_list_lock);
		} else {
			spin_lock(&NM_I(sbi)->nid_list_lock);
2487
			update_free_nid_bitmap(sbi, nid, false, false);
2488 2489
			spin_unlock(&NM_I(sbi)->nid_list_lock);
		}
2490
	}
J
Jaegeuk Kim 已提交
2491

2492
	if (to_journal) {
2493
		up_write(&curseg->journal_rwsem);
2494 2495
	} else {
		__update_nat_bits(sbi, start_nid, page);
2496
		f2fs_put_page(page, 1);
2497
	}
2498

2499 2500 2501 2502
	f2fs_bug_on(sbi, set->entry_cnt);

	radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
	kmem_cache_free(nat_entry_set_slab, set);
2503
}
2504

2505 2506 2507
/*
 * This function is called during the checkpointing process.
 */
2508
void flush_nat_entries(struct f2fs_sb_info *sbi, struct cp_control *cpc)
2509 2510 2511
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
2512
	struct f2fs_journal *journal = curseg->journal;
2513
	struct nat_entry_set *setvec[SETVEC_SIZE];
2514 2515 2516 2517 2518
	struct nat_entry_set *set, *tmp;
	unsigned int found;
	nid_t set_idx = 0;
	LIST_HEAD(sets);

2519 2520
	if (!nm_i->dirty_nat_cnt)
		return;
2521

2522
	down_write(&nm_i->nat_tree_lock);
2523

2524 2525 2526 2527 2528
	/*
	 * if there are no enough space in journal to store dirty nat
	 * entries, remove all entries from journal and merge them
	 * into nat entry set.
	 */
2529
	if (enabled_nat_bits(sbi, cpc) ||
2530
		!__has_cursum_space(journal, nm_i->dirty_nat_cnt, NAT_JOURNAL))
2531 2532 2533
		remove_nats_in_journal(sbi);

	while ((found = __gang_lookup_nat_set(nm_i,
2534
					set_idx, SETVEC_SIZE, setvec))) {
2535 2536 2537 2538
		unsigned idx;
		set_idx = setvec[found - 1]->set + 1;
		for (idx = 0; idx < found; idx++)
			__adjust_nat_entry_set(setvec[idx], &sets,
2539
						MAX_NAT_JENTRIES(journal));
J
Jaegeuk Kim 已提交
2540
	}
2541

2542 2543
	/* flush dirty nats in nat entry set */
	list_for_each_entry_safe(set, tmp, &sets, set_list)
2544
		__flush_nat_entry_set(sbi, set, cpc);
2545

2546
	up_write(&nm_i->nat_tree_lock);
2547

2548
	f2fs_bug_on(sbi, nm_i->dirty_nat_cnt);
J
Jaegeuk Kim 已提交
2549 2550
}

2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579
static int __get_nat_bitmaps(struct f2fs_sb_info *sbi)
{
	struct f2fs_checkpoint *ckpt = F2FS_CKPT(sbi);
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	unsigned int nat_bits_bytes = nm_i->nat_blocks / BITS_PER_BYTE;
	unsigned int i;
	__u64 cp_ver = cur_cp_version(ckpt);
	block_t nat_bits_addr;

	if (!enabled_nat_bits(sbi, NULL))
		return 0;

	nm_i->nat_bits_blocks = F2FS_BYTES_TO_BLK((nat_bits_bytes << 1) + 8 +
						F2FS_BLKSIZE - 1);
	nm_i->nat_bits = kzalloc(nm_i->nat_bits_blocks << F2FS_BLKSIZE_BITS,
						GFP_KERNEL);
	if (!nm_i->nat_bits)
		return -ENOMEM;

	nat_bits_addr = __start_cp_addr(sbi) + sbi->blocks_per_seg -
						nm_i->nat_bits_blocks;
	for (i = 0; i < nm_i->nat_bits_blocks; i++) {
		struct page *page = get_meta_page(sbi, nat_bits_addr++);

		memcpy(nm_i->nat_bits + (i << F2FS_BLKSIZE_BITS),
					page_address(page), F2FS_BLKSIZE);
		f2fs_put_page(page, 1);
	}

2580
	cp_ver |= (cur_cp_crc(ckpt) << 32);
2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592
	if (cpu_to_le64(cp_ver) != *(__le64 *)nm_i->nat_bits) {
		disable_nat_bits(sbi, true);
		return 0;
	}

	nm_i->full_nat_bits = nm_i->nat_bits + 8;
	nm_i->empty_nat_bits = nm_i->full_nat_bits + nat_bits_bytes;

	f2fs_msg(sbi->sb, KERN_NOTICE, "Found nat_bits in checkpoint");
	return 0;
}

J
Jaegeuk Kim 已提交
2593 2594 2595 2596 2597
static int init_node_manager(struct f2fs_sb_info *sbi)
{
	struct f2fs_super_block *sb_raw = F2FS_RAW_SUPER(sbi);
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	unsigned char *version_bitmap;
2598 2599
	unsigned int nat_segs;
	int err;
J
Jaegeuk Kim 已提交
2600 2601 2602 2603 2604

	nm_i->nat_blkaddr = le32_to_cpu(sb_raw->nat_blkaddr);

	/* segment_count_nat includes pair segment so divide to 2. */
	nat_segs = le32_to_cpu(sb_raw->segment_count_nat) >> 1;
2605 2606
	nm_i->nat_blocks = nat_segs << le32_to_cpu(sb_raw->log_blocks_per_seg);
	nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nm_i->nat_blocks;
2607

2608
	/* not used nids: 0, node, meta, (and root counted as valid node) */
2609 2610
	nm_i->available_nids = nm_i->max_nid - sbi->total_valid_node_count -
							F2FS_RESERVED_NODE_NUM;
C
Chao Yu 已提交
2611 2612
	nm_i->nid_cnt[FREE_NID_LIST] = 0;
	nm_i->nid_cnt[ALLOC_NID_LIST] = 0;
J
Jaegeuk Kim 已提交
2613
	nm_i->nat_cnt = 0;
2614
	nm_i->ram_thresh = DEF_RAM_THRESHOLD;
C
Chao Yu 已提交
2615
	nm_i->ra_nid_pages = DEF_RA_NID_PAGES;
C
Chao Yu 已提交
2616
	nm_i->dirty_nats_ratio = DEF_DIRTY_NAT_RATIO_THRESHOLD;
J
Jaegeuk Kim 已提交
2617

2618
	INIT_RADIX_TREE(&nm_i->free_nid_root, GFP_ATOMIC);
C
Chao Yu 已提交
2619 2620
	INIT_LIST_HEAD(&nm_i->nid_list[FREE_NID_LIST]);
	INIT_LIST_HEAD(&nm_i->nid_list[ALLOC_NID_LIST]);
2621 2622
	INIT_RADIX_TREE(&nm_i->nat_root, GFP_NOIO);
	INIT_RADIX_TREE(&nm_i->nat_set_root, GFP_NOIO);
J
Jaegeuk Kim 已提交
2623 2624 2625
	INIT_LIST_HEAD(&nm_i->nat_entries);

	mutex_init(&nm_i->build_lock);
C
Chao Yu 已提交
2626
	spin_lock_init(&nm_i->nid_list_lock);
2627
	init_rwsem(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2628 2629

	nm_i->next_scan_nid = le32_to_cpu(sbi->ckpt->next_free_nid);
A
Alexandru Gheorghiu 已提交
2630
	nm_i->bitmap_size = __bitmap_size(sbi, NAT_BITMAP);
J
Jaegeuk Kim 已提交
2631 2632 2633 2634
	version_bitmap = __bitmap_ptr(sbi, NAT_BITMAP);
	if (!version_bitmap)
		return -EFAULT;

A
Alexandru Gheorghiu 已提交
2635 2636 2637 2638
	nm_i->nat_bitmap = kmemdup(version_bitmap, nm_i->bitmap_size,
					GFP_KERNEL);
	if (!nm_i->nat_bitmap)
		return -ENOMEM;
2639

2640 2641 2642 2643
	err = __get_nat_bitmaps(sbi);
	if (err)
		return err;

2644 2645 2646 2647 2648 2649 2650
#ifdef CONFIG_F2FS_CHECK_FS
	nm_i->nat_bitmap_mir = kmemdup(version_bitmap, nm_i->bitmap_size,
					GFP_KERNEL);
	if (!nm_i->nat_bitmap_mir)
		return -ENOMEM;
#endif

J
Jaegeuk Kim 已提交
2651 2652 2653
	return 0;
}

J
Jaegeuk Kim 已提交
2654
static int init_free_nid_cache(struct f2fs_sb_info *sbi)
C
Chao Yu 已提交
2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);

	nm_i->free_nid_bitmap = f2fs_kvzalloc(nm_i->nat_blocks *
					NAT_ENTRY_BITMAP_SIZE, GFP_KERNEL);
	if (!nm_i->free_nid_bitmap)
		return -ENOMEM;

	nm_i->nat_block_bitmap = f2fs_kvzalloc(nm_i->nat_blocks / 8,
								GFP_KERNEL);
	if (!nm_i->nat_block_bitmap)
		return -ENOMEM;
2667 2668 2669 2670 2671 2672 2673 2674

	nm_i->free_nid_count = f2fs_kvzalloc(nm_i->nat_blocks *
					sizeof(unsigned short), GFP_KERNEL);
	if (!nm_i->free_nid_count)
		return -ENOMEM;

	spin_lock_init(&nm_i->free_nid_lock);

C
Chao Yu 已提交
2675 2676 2677
	return 0;
}

J
Jaegeuk Kim 已提交
2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689
int build_node_manager(struct f2fs_sb_info *sbi)
{
	int err;

	sbi->nm_info = kzalloc(sizeof(struct f2fs_nm_info), GFP_KERNEL);
	if (!sbi->nm_info)
		return -ENOMEM;

	err = init_node_manager(sbi);
	if (err)
		return err;

C
Chao Yu 已提交
2690 2691 2692 2693
	err = init_free_nid_cache(sbi);
	if (err)
		return err;

2694
	build_free_nids(sbi, true, true);
J
Jaegeuk Kim 已提交
2695 2696 2697 2698 2699 2700 2701 2702
	return 0;
}

void destroy_node_manager(struct f2fs_sb_info *sbi)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i, *next_i;
	struct nat_entry *natvec[NATVEC_SIZE];
2703
	struct nat_entry_set *setvec[SETVEC_SIZE];
J
Jaegeuk Kim 已提交
2704 2705 2706 2707 2708 2709 2710
	nid_t nid = 0;
	unsigned int found;

	if (!nm_i)
		return;

	/* destroy free nid list */
C
Chao Yu 已提交
2711 2712 2713
	spin_lock(&nm_i->nid_list_lock);
	list_for_each_entry_safe(i, next_i, &nm_i->nid_list[FREE_NID_LIST],
									list) {
2714
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, false);
C
Chao Yu 已提交
2715
		spin_unlock(&nm_i->nid_list_lock);
2716
		kmem_cache_free(free_nid_slab, i);
C
Chao Yu 已提交
2717
		spin_lock(&nm_i->nid_list_lock);
J
Jaegeuk Kim 已提交
2718
	}
C
Chao Yu 已提交
2719 2720 2721 2722
	f2fs_bug_on(sbi, nm_i->nid_cnt[FREE_NID_LIST]);
	f2fs_bug_on(sbi, nm_i->nid_cnt[ALLOC_NID_LIST]);
	f2fs_bug_on(sbi, !list_empty(&nm_i->nid_list[ALLOC_NID_LIST]));
	spin_unlock(&nm_i->nid_list_lock);
J
Jaegeuk Kim 已提交
2723 2724

	/* destroy nat cache */
2725
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2726 2727 2728
	while ((found = __gang_lookup_nat_cache(nm_i,
					nid, NATVEC_SIZE, natvec))) {
		unsigned idx;
2729

2730 2731 2732
		nid = nat_get_nid(natvec[found - 1]) + 1;
		for (idx = 0; idx < found; idx++)
			__del_from_nat_cache(nm_i, natvec[idx]);
J
Jaegeuk Kim 已提交
2733
	}
2734
	f2fs_bug_on(sbi, nm_i->nat_cnt);
2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749

	/* destroy nat set cache */
	nid = 0;
	while ((found = __gang_lookup_nat_set(nm_i,
					nid, SETVEC_SIZE, setvec))) {
		unsigned idx;

		nid = setvec[found - 1]->set + 1;
		for (idx = 0; idx < found; idx++) {
			/* entry_cnt is not zero, when cp_error was occurred */
			f2fs_bug_on(sbi, !list_empty(&setvec[idx]->entry_list));
			radix_tree_delete(&nm_i->nat_set_root, setvec[idx]->set);
			kmem_cache_free(nat_entry_set_slab, setvec[idx]);
		}
	}
2750
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2751

C
Chao Yu 已提交
2752 2753
	kvfree(nm_i->nat_block_bitmap);
	kvfree(nm_i->free_nid_bitmap);
2754
	kvfree(nm_i->free_nid_count);
C
Chao Yu 已提交
2755

J
Jaegeuk Kim 已提交
2756
	kfree(nm_i->nat_bitmap);
2757
	kfree(nm_i->nat_bits);
2758 2759 2760
#ifdef CONFIG_F2FS_CHECK_FS
	kfree(nm_i->nat_bitmap_mir);
#endif
J
Jaegeuk Kim 已提交
2761 2762 2763 2764
	sbi->nm_info = NULL;
	kfree(nm_i);
}

2765
int __init create_node_manager_caches(void)
J
Jaegeuk Kim 已提交
2766 2767
{
	nat_entry_slab = f2fs_kmem_cache_create("nat_entry",
2768
			sizeof(struct nat_entry));
J
Jaegeuk Kim 已提交
2769
	if (!nat_entry_slab)
2770
		goto fail;
J
Jaegeuk Kim 已提交
2771 2772

	free_nid_slab = f2fs_kmem_cache_create("free_nid",
2773
			sizeof(struct free_nid));
2774
	if (!free_nid_slab)
2775
		goto destroy_nat_entry;
2776 2777 2778 2779

	nat_entry_set_slab = f2fs_kmem_cache_create("nat_entry_set",
			sizeof(struct nat_entry_set));
	if (!nat_entry_set_slab)
2780
		goto destroy_free_nid;
J
Jaegeuk Kim 已提交
2781
	return 0;
2782

2783
destroy_free_nid:
2784
	kmem_cache_destroy(free_nid_slab);
2785
destroy_nat_entry:
2786 2787 2788
	kmem_cache_destroy(nat_entry_slab);
fail:
	return -ENOMEM;
J
Jaegeuk Kim 已提交
2789 2790 2791 2792
}

void destroy_node_manager_caches(void)
{
2793
	kmem_cache_destroy(nat_entry_set_slab);
J
Jaegeuk Kim 已提交
2794 2795 2796
	kmem_cache_destroy(free_nid_slab);
	kmem_cache_destroy(nat_entry_slab);
}