node.c 53.1 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) {
48 49 50
		mem_size = (nm_i->fcnt * sizeof(struct free_nid)) >>
							PAGE_CACHE_SHIFT;
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
51
	} else if (type == NAT_ENTRIES) {
52 53 54
		mem_size = (nm_i->nat_cnt * sizeof(struct nat_entry)) >>
							PAGE_CACHE_SHIFT;
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 2);
55 56 57 58 59
	} 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);
60 61 62 63
	} else if (type == INO_ENTRIES) {
		int i;

		for (i = 0; i <= UPDATE_INO; i++)
64 65
			mem_size += (sbi->im[i].ino_num *
				sizeof(struct ino_entry)) >> PAGE_CACHE_SHIFT;
66
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
67
	} else if (type == EXTENT_CACHE) {
68 69
		mem_size = (atomic_read(&sbi->total_ext_tree) *
				sizeof(struct extent_tree) +
70 71 72
				atomic_read(&sbi->total_ext_node) *
				sizeof(struct extent_node)) >> PAGE_CACHE_SHIFT;
		res = mem_size < ((avail_ram * nm_i->ram_thresh / 100) >> 1);
73
	} else {
74 75
		if (!sbi->sb->s_bdi->wb.dirty_exceeded)
			return true;
76 77
	}
	return res;
78 79
}

J
Jaegeuk Kim 已提交
80 81 82 83 84 85 86 87 88 89 90 91 92
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);
93
		dec_page_count(F2FS_M_SB(mapping), F2FS_DIRTY_NODES);
J
Jaegeuk Kim 已提交
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
	}
	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);
120
	f2fs_bug_on(sbi, PageDirty(src_page));
J
Jaegeuk Kim 已提交
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151

	src_addr = page_address(src_page);
	dst_addr = page_address(dst_page);
	memcpy(dst_addr, src_addr, PAGE_CACHE_SIZE);
	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);
}

152 153 154 155 156 157 158 159
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;
160

161 162
	head = radix_tree_lookup(&nm_i->nat_set_root, set);
	if (!head) {
163
		head = f2fs_kmem_cache_alloc(nat_entry_set_slab, GFP_NOFS);
164 165 166 167 168

		INIT_LIST_HEAD(&head->entry_list);
		INIT_LIST_HEAD(&head->set_list);
		head->set = set;
		head->entry_cnt = 0;
169
		f2fs_radix_tree_insert(&nm_i->nat_set_root, set, head);
170 171 172 173 174 175 176 177 178 179
	}
	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)
{
180
	nid_t set = NAT_BLOCK_OFFSET(ne->ni.nid);
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198
	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
199
int need_dentry_mark(struct f2fs_sb_info *sbi, nid_t nid)
J
Jaegeuk Kim 已提交
200 201 202
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
J
Jaegeuk Kim 已提交
203
	bool need = false;
J
Jaegeuk Kim 已提交
204

205
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
206
	e = __lookup_nat_cache(nm_i, nid);
J
Jaegeuk Kim 已提交
207 208 209 210 211
	if (e) {
		if (!get_nat_flag(e, IS_CHECKPOINTED) &&
				!get_nat_flag(e, HAS_FSYNCED_INODE))
			need = true;
	}
212
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
213
	return need;
J
Jaegeuk Kim 已提交
214 215
}

J
Jaegeuk Kim 已提交
216
bool is_checkpointed_node(struct f2fs_sb_info *sbi, nid_t nid)
217 218 219
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
J
Jaegeuk Kim 已提交
220
	bool is_cp = true;
221

222
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
223 224 225
	e = __lookup_nat_cache(nm_i, nid);
	if (e && !get_nat_flag(e, IS_CHECKPOINTED))
		is_cp = false;
226
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
227
	return is_cp;
228 229
}

230
bool need_inode_block_update(struct f2fs_sb_info *sbi, nid_t ino)
231 232 233
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
234
	bool need_update = true;
235

236
	down_read(&nm_i->nat_tree_lock);
237 238 239 240 241
	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;
242
	up_read(&nm_i->nat_tree_lock);
243
	return need_update;
244 245
}

J
Jaegeuk Kim 已提交
246 247 248 249
static struct nat_entry *grab_nat_entry(struct f2fs_nm_info *nm_i, nid_t nid)
{
	struct nat_entry *new;

250
	new = f2fs_kmem_cache_alloc(nat_entry_slab, GFP_NOFS);
251
	f2fs_radix_tree_insert(&nm_i->nat_root, nid, new);
J
Jaegeuk Kim 已提交
252 253
	memset(new, 0, sizeof(struct nat_entry));
	nat_set_nid(new, nid);
254
	nat_reset_flag(new);
J
Jaegeuk Kim 已提交
255 256 257 258 259
	list_add_tail(&new->list, &nm_i->nat_entries);
	nm_i->nat_cnt++;
	return new;
}

260
static void cache_nat_entry(struct f2fs_sb_info *sbi, nid_t nid,
J
Jaegeuk Kim 已提交
261 262
						struct f2fs_nat_entry *ne)
{
263
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
264
	struct nat_entry *e;
265

J
Jaegeuk Kim 已提交
266 267 268
	e = __lookup_nat_cache(nm_i, nid);
	if (!e) {
		e = grab_nat_entry(nm_i, nid);
269
		node_info_from_raw_nat(&e->ni, ne);
270 271 272 273
	} else {
		f2fs_bug_on(sbi, nat_get_ino(e) != ne->ino ||
				nat_get_blkaddr(e) != ne->block_addr ||
				nat_get_version(e) != ne->version);
J
Jaegeuk Kim 已提交
274 275 276 277
	}
}

static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni,
278
			block_t new_blkaddr, bool fsync_done)
J
Jaegeuk Kim 已提交
279 280 281
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
282

283
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
284 285 286
	e = __lookup_nat_cache(nm_i, ni->nid);
	if (!e) {
		e = grab_nat_entry(nm_i, ni->nid);
287
		copy_node_info(&e->ni, ni);
288
		f2fs_bug_on(sbi, ni->blk_addr == NEW_ADDR);
J
Jaegeuk Kim 已提交
289 290 291 292 293 294
	} 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.
		 */
295
		copy_node_info(&e->ni, ni);
296
		f2fs_bug_on(sbi, ni->blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
297 298 299
	}

	/* sanity check */
300 301
	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
302
			new_blkaddr == NULL_ADDR);
303
	f2fs_bug_on(sbi, nat_get_blkaddr(e) == NEW_ADDR &&
J
Jaegeuk Kim 已提交
304
			new_blkaddr == NEW_ADDR);
305
	f2fs_bug_on(sbi, nat_get_blkaddr(e) != NEW_ADDR &&
J
Jaegeuk Kim 已提交
306 307 308
			nat_get_blkaddr(e) != NULL_ADDR &&
			new_blkaddr == NEW_ADDR);

A
arter97 已提交
309
	/* increment version no as node is removed */
J
Jaegeuk Kim 已提交
310 311 312
	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
313 314 315 316

		/* 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
317 318 319 320
	}

	/* change address */
	nat_set_blkaddr(e, new_blkaddr);
321 322
	if (new_blkaddr == NEW_ADDR || new_blkaddr == NULL_ADDR)
		set_nat_flag(e, IS_CHECKPOINTED, false);
J
Jaegeuk Kim 已提交
323
	__set_nat_cache_dirty(nm_i, e);
324 325

	/* update fsync_mark if its inode nat entry is still alive */
326 327
	if (ni->nid != ni->ino)
		e = __lookup_nat_cache(nm_i, ni->ino);
328 329 330 331 332
	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);
	}
333
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
334 335
}

336
int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
J
Jaegeuk Kim 已提交
337 338
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
339
	int nr = nr_shrink;
J
Jaegeuk Kim 已提交
340

J
Jaegeuk Kim 已提交
341
	if (!down_write_trylock(&nm_i->nat_tree_lock))
J
Jaegeuk Kim 已提交
342 343 344 345 346 347 348 349 350
		return 0;

	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--;
	}
351
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
352
	return nr - nr_shrink;
J
Jaegeuk Kim 已提交
353 354
}

J
Jaegeuk Kim 已提交
355
/*
A
arter97 已提交
356
 * This function always returns success
J
Jaegeuk Kim 已提交
357 358 359 360 361
 */
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);
362
	struct f2fs_journal *journal = curseg->journal;
J
Jaegeuk Kim 已提交
363 364 365 366 367 368 369 370 371 372
	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 */
373
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
374 375 376 377 378
	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);
379
		up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
380
		return;
381
	}
J
Jaegeuk Kim 已提交
382

383 384
	memset(&ne, 0, sizeof(struct f2fs_nat_entry));

J
Jaegeuk Kim 已提交
385
	/* Check current segment summary */
386
	down_read(&curseg->journal_rwsem);
387
	i = lookup_journal_in_cursum(journal, NAT_JOURNAL, nid, 0);
J
Jaegeuk Kim 已提交
388
	if (i >= 0) {
389
		ne = nat_in_journal(journal, i);
J
Jaegeuk Kim 已提交
390 391
		node_info_from_raw_nat(ni, &ne);
	}
392
	up_read(&curseg->journal_rwsem);
J
Jaegeuk Kim 已提交
393 394 395 396 397 398 399 400 401 402
	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:
403
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
404
	/* cache nat entry */
405 406
	down_write(&nm_i->nat_tree_lock);
	cache_nat_entry(sbi, nid, &ne);
407
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
408 409
}

410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
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
441
/*
J
Jaegeuk Kim 已提交
442 443 444
 * The maximum depth is four.
 * Offset[0] will have raw inode offset.
 */
445
static int get_node_path(struct inode *inode, long block,
446
				int offset[4], unsigned int noffset[4])
J
Jaegeuk Kim 已提交
447
{
448
	const long direct_index = ADDRS_PER_INODE(inode);
J
Jaegeuk Kim 已提交
449 450 451 452 453 454 455 456 457 458
	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) {
459
		offset[n] = block;
J
Jaegeuk Kim 已提交
460 461 462 463 464 465
		goto got;
	}
	block -= direct_index;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR1_BLOCK;
		noffset[n] = 1;
466
		offset[n] = block;
J
Jaegeuk Kim 已提交
467 468 469 470 471 472 473
		level = 1;
		goto got;
	}
	block -= direct_blks;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR2_BLOCK;
		noffset[n] = 2;
474
		offset[n] = block;
J
Jaegeuk Kim 已提交
475 476 477 478 479 480 481 482 483
		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];
484
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
485 486 487 488 489 490 491 492 493
		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];
494
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
495 496 497 498 499 500 501 502 503 504 505 506 507 508
		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];
509
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
510 511 512 513 514 515 516 517 518 519 520
		level = 3;
		goto got;
	} else {
		BUG();
	}
got:
	return level;
}

/*
 * Caller should call f2fs_put_dnode(dn).
C
Chao Yu 已提交
521 522
 * 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.
523
 * In the case of RDONLY_NODE, we don't need to care about mutex.
J
Jaegeuk Kim 已提交
524
 */
525
int get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
J
Jaegeuk Kim 已提交
526
{
527
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
528
	struct page *npage[4];
529
	struct page *parent = NULL;
J
Jaegeuk Kim 已提交
530 531 532
	int offset[4];
	unsigned int noffset[4];
	nid_t nids[4];
533
	int level, i = 0;
J
Jaegeuk Kim 已提交
534 535
	int err = 0;

536
	level = get_node_path(dn->inode, index, offset, noffset);
J
Jaegeuk Kim 已提交
537 538

	nids[0] = dn->inode->i_ino;
539
	npage[0] = dn->inode_page;
J
Jaegeuk Kim 已提交
540

541 542 543 544 545
	if (!npage[0]) {
		npage[0] = get_node_page(sbi, nids[0]);
		if (IS_ERR(npage[0]))
			return PTR_ERR(npage[0]);
	}
546 547 548

	/* if inline_data is set, should not report any block indices */
	if (f2fs_has_inline_data(dn->inode) && index) {
549
		err = -ENOENT;
550 551 552 553
		f2fs_put_page(npage[0], 1);
		goto release_out;
	}

J
Jaegeuk Kim 已提交
554
	parent = npage[0];
555 556
	if (level != 0)
		nids[1] = get_nid(parent, offset[0], true);
J
Jaegeuk Kim 已提交
557 558 559 560 561 562 563
	dn->inode_page = npage[0];
	dn->inode_page_locked = true;

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

564
		if (!nids[i] && mode == ALLOC_NODE) {
J
Jaegeuk Kim 已提交
565 566 567 568 569 570 571
			/* alloc new node */
			if (!alloc_nid(sbi, &(nids[i]))) {
				err = -ENOSPC;
				goto release_pages;
			}

			dn->nid = nids[i];
572
			npage[i] = new_node_page(dn, noffset[i], NULL);
J
Jaegeuk Kim 已提交
573 574 575 576 577 578 579 580 581
			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;
582
		} else if (mode == LOOKUP_NODE_RA && i == level && level > 1) {
J
Jaegeuk Kim 已提交
583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622
			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;
623 624 625 626
	if (err == -ENOENT) {
		dn->cur_level = i;
		dn->max_level = level;
	}
J
Jaegeuk Kim 已提交
627 628 629 630 631
	return err;
}

static void truncate_node(struct dnode_of_data *dn)
{
632
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
633 634 635
	struct node_info ni;

	get_node_info(sbi, dn->nid, &ni);
636
	if (dn->inode->i_blocks == 0) {
637
		f2fs_bug_on(sbi, ni.blk_addr != NULL_ADDR);
638 639
		goto invalidate;
	}
640
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
J
Jaegeuk Kim 已提交
641 642

	/* Deallocate node address */
643
	invalidate_blocks(sbi, ni.blk_addr);
644
	dec_valid_node_count(sbi, dn->inode);
645
	set_node_addr(sbi, &ni, NULL_ADDR, false);
J
Jaegeuk Kim 已提交
646 647 648 649 650 651 652

	if (dn->nid == dn->inode->i_ino) {
		remove_orphan_inode(sbi, dn->nid);
		dec_valid_inode_count(sbi);
	} else {
		sync_inode_page(dn);
	}
653
invalidate:
J
Jaegeuk Kim 已提交
654
	clear_node_page_dirty(dn->node_page);
655
	set_sbi_flag(sbi, SBI_IS_DIRTY);
J
Jaegeuk Kim 已提交
656 657

	f2fs_put_page(dn->node_page, 1);
658 659 660 661

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

J
Jaegeuk Kim 已提交
662
	dn->node_page = NULL;
663
	trace_f2fs_truncate_node(dn->inode, dn->nid, ni.blk_addr);
J
Jaegeuk Kim 已提交
664 665 666 667 668 669 670 671 672 673
}

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

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

	/* get direct node */
674
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
J
Jaegeuk Kim 已提交
675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701
	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;

702 703
	trace_f2fs_truncate_nodes_enter(dn->inode, dn->nid, dn->data_blkaddr);

704
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
705 706
	if (IS_ERR(page)) {
		trace_f2fs_truncate_nodes_exit(dn->inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
707
		return PTR_ERR(page);
708
	}
J
Jaegeuk Kim 已提交
709

710
	rn = F2FS_NODE(page);
J
Jaegeuk Kim 已提交
711 712 713 714 715 716 717 718 719
	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;
720 721
			if (set_nid(page, i, 0, false))
				dn->node_changed = true;
J
Jaegeuk Kim 已提交
722 723 724 725 726 727 728 729 730 731 732 733
		}
	} 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)) {
734 735
				if (set_nid(page, i, 0, false))
					dn->node_changed = true;
J
Jaegeuk Kim 已提交
736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751
				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);
	}
752
	trace_f2fs_truncate_nodes_exit(dn->inode, freed);
J
Jaegeuk Kim 已提交
753 754 755 756
	return freed;

out_err:
	f2fs_put_page(page, 1);
757
	trace_f2fs_truncate_nodes_exit(dn->inode, ret);
J
Jaegeuk Kim 已提交
758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775
	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 */
776
	for (i = 0; i < idx + 1; i++) {
A
arter97 已提交
777
		/* reference count'll be increased */
778
		pages[i] = get_node_page(F2FS_I_SB(dn->inode), nid[i]);
J
Jaegeuk Kim 已提交
779 780
		if (IS_ERR(pages[i])) {
			err = PTR_ERR(pages[i]);
781
			idx = i - 1;
J
Jaegeuk Kim 已提交
782 783 784 785 786 787
			goto fail;
		}
		nid[i + 1] = get_nid(pages[i], offset[i + 1], false);
	}

	/* free direct nodes linked to a partial indirect node */
788
	for (i = offset[idx + 1]; i < NIDS_PER_BLOCK; i++) {
J
Jaegeuk Kim 已提交
789 790 791 792 793 794 795
		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;
796 797
		if (set_nid(pages[idx], i, 0, false))
			dn->node_changed = true;
J
Jaegeuk Kim 已提交
798 799
	}

800
	if (offset[idx + 1] == 0) {
J
Jaegeuk Kim 已提交
801 802 803 804 805 806 807
		dn->node_page = pages[idx];
		dn->nid = nid[idx];
		truncate_node(dn);
	} else {
		f2fs_put_page(pages[idx], 1);
	}
	offset[idx]++;
808 809
	offset[idx + 1] = 0;
	idx--;
J
Jaegeuk Kim 已提交
810
fail:
811
	for (i = idx; i >= 0; i--)
J
Jaegeuk Kim 已提交
812
		f2fs_put_page(pages[i], 1);
813 814 815

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

J
Jaegeuk Kim 已提交
816 817 818
	return err;
}

J
Jaegeuk Kim 已提交
819
/*
J
Jaegeuk Kim 已提交
820 821 822 823
 * All the block addresses of data and nodes should be nullified.
 */
int truncate_inode_blocks(struct inode *inode, pgoff_t from)
{
824
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
J
Jaegeuk Kim 已提交
825 826
	int err = 0, cont = 1;
	int level, offset[4], noffset[4];
J
Jaegeuk Kim 已提交
827
	unsigned int nofs = 0;
828
	struct f2fs_inode *ri;
J
Jaegeuk Kim 已提交
829 830 831
	struct dnode_of_data dn;
	struct page *page;

832 833
	trace_f2fs_truncate_inode_blocks_enter(inode, from);

834
	level = get_node_path(inode, from, offset, noffset);
835
restart:
J
Jaegeuk Kim 已提交
836
	page = get_node_page(sbi, inode->i_ino);
837 838
	if (IS_ERR(page)) {
		trace_f2fs_truncate_inode_blocks_exit(inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
839
		return PTR_ERR(page);
840
	}
J
Jaegeuk Kim 已提交
841 842 843 844

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

845
	ri = F2FS_INODE(page);
J
Jaegeuk Kim 已提交
846 847 848 849 850 851 852 853 854
	switch (level) {
	case 0:
	case 1:
		nofs = noffset[1];
		break;
	case 2:
		nofs = noffset[1];
		if (!offset[level - 1])
			goto skip_partial;
855
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
856 857 858 859 860 861 862 863
		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;
864
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
865 866 867 868 869 870 871 872 873
		if (err < 0 && err != -ENOENT)
			goto fail;
		break;
	default:
		BUG();
	}

skip_partial:
	while (cont) {
874
		dn.nid = le32_to_cpu(ri->i_nid[offset[0] - NODE_DIR1_BLOCK]);
J
Jaegeuk Kim 已提交
875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896
		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 &&
897
				ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) {
J
Jaegeuk Kim 已提交
898
			lock_page(page);
899
			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
900 901 902
				f2fs_put_page(page, 1);
				goto restart;
			}
903
			f2fs_wait_on_page_writeback(page, NODE, true);
904
			ri->i_nid[offset[0] - NODE_DIR1_BLOCK] = 0;
J
Jaegeuk Kim 已提交
905 906 907 908 909 910 911 912 913
			set_page_dirty(page);
			unlock_page(page);
		}
		offset[1] = 0;
		offset[0]++;
		nofs += err;
	}
fail:
	f2fs_put_page(page, 0);
914
	trace_f2fs_truncate_inode_blocks_exit(inode, err);
J
Jaegeuk Kim 已提交
915 916 917
	return err > 0 ? 0 : err;
}

918 919
int truncate_xattr_node(struct inode *inode, struct page *page)
{
920
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
921 922 923 924 925 926 927 928 929 930 931 932
	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);

	F2FS_I(inode)->i_xattr_nid = 0;
J
Jaegeuk Kim 已提交
933 934 935 936

	/* need to do checkpoint during fsync */
	F2FS_I(inode)->xattr_ver = cur_cp_version(F2FS_CKPT(sbi));

937 938 939
	set_new_dnode(&dn, inode, page, npage, nid);

	if (page)
940
		dn.inode_page_locked = true;
941 942 943 944
	truncate_node(&dn);
	return 0;
}

945
/*
C
Chao Yu 已提交
946 947
 * Caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
948
 */
C
Chao Yu 已提交
949
int remove_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
950 951
{
	struct dnode_of_data dn;
C
Chao Yu 已提交
952
	int err;
J
Jaegeuk Kim 已提交
953

954
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
C
Chao Yu 已提交
955 956 957
	err = get_dnode_of_data(&dn, 0, LOOKUP_NODE);
	if (err)
		return err;
J
Jaegeuk Kim 已提交
958

C
Chao Yu 已提交
959 960
	err = truncate_xattr_node(inode, dn.inode_page);
	if (err) {
961
		f2fs_put_dnode(&dn);
C
Chao Yu 已提交
962
		return err;
J
Jaegeuk Kim 已提交
963
	}
964 965 966 967 968 969

	/* 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
970
	/* 0 is possible, after f2fs_new_inode() has failed */
971 972
	f2fs_bug_on(F2FS_I_SB(inode),
			inode->i_blocks != 0 && inode->i_blocks != 1);
973 974

	/* will put inode & node pages */
975
	truncate_node(&dn);
C
Chao Yu 已提交
976
	return 0;
J
Jaegeuk Kim 已提交
977 978
}

979
struct page *new_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
980 981 982 983 984
{
	struct dnode_of_data dn;

	/* allocate inode page for new inode */
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
985 986

	/* caller should f2fs_put_page(page, 1); */
987
	return new_node_page(&dn, 0, NULL);
J
Jaegeuk Kim 已提交
988 989
}

990 991
struct page *new_node_page(struct dnode_of_data *dn,
				unsigned int ofs, struct page *ipage)
J
Jaegeuk Kim 已提交
992
{
993
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
994 995 996 997
	struct node_info old_ni, new_ni;
	struct page *page;
	int err;

998
	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
J
Jaegeuk Kim 已提交
999 1000
		return ERR_PTR(-EPERM);

1001
	page = grab_cache_page(NODE_MAPPING(sbi), dn->nid);
J
Jaegeuk Kim 已提交
1002 1003 1004
	if (!page)
		return ERR_PTR(-ENOMEM);

1005
	if (unlikely(!inc_valid_node_count(sbi, dn->inode))) {
1006 1007 1008
		err = -ENOSPC;
		goto fail;
	}
J
Jaegeuk Kim 已提交
1009

1010
	get_node_info(sbi, dn->nid, &old_ni);
J
Jaegeuk Kim 已提交
1011 1012

	/* Reinitialize old_ni with new node page */
1013
	f2fs_bug_on(sbi, old_ni.blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
1014 1015
	new_ni = old_ni;
	new_ni.ino = dn->inode->i_ino;
1016
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
1017

1018
	f2fs_wait_on_page_writeback(page, NODE, true);
1019
	fill_node_footer(page, dn->nid, dn->inode->i_ino, ofs, true);
1020
	set_cold_node(dn->inode, page);
1021
	SetPageUptodate(page);
1022 1023
	if (set_page_dirty(page))
		dn->node_changed = true;
J
Jaegeuk Kim 已提交
1024

1025
	if (f2fs_has_xattr_block(ofs))
1026 1027
		F2FS_I(dn->inode)->i_xattr_nid = dn->nid;

J
Jaegeuk Kim 已提交
1028
	dn->node_page = page;
1029 1030 1031 1032
	if (ipage)
		update_inode(dn->inode, ipage);
	else
		sync_inode_page(dn);
J
Jaegeuk Kim 已提交
1033 1034 1035 1036 1037 1038
	if (ofs == 0)
		inc_valid_inode_count(sbi);

	return page;

fail:
1039
	clear_node_page_dirty(page);
J
Jaegeuk Kim 已提交
1040 1041 1042 1043
	f2fs_put_page(page, 1);
	return ERR_PTR(err);
}

1044 1045 1046
/*
 * Caller should do after getting the following values.
 * 0: f2fs_put_page(page, 0)
1047
 * LOCKED_PAGE or error: f2fs_put_page(page, 1)
1048
 */
1049
static int read_node_page(struct page *page, int rw)
J
Jaegeuk Kim 已提交
1050
{
1051
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1052
	struct node_info ni;
1053
	struct f2fs_io_info fio = {
1054
		.sbi = sbi,
1055 1056
		.type = NODE,
		.rw = rw,
1057
		.page = page,
1058
		.encrypted_page = NULL,
1059
	};
J
Jaegeuk Kim 已提交
1060 1061 1062

	get_node_info(sbi, page->index, &ni);

1063
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1064
		ClearPageUptodate(page);
J
Jaegeuk Kim 已提交
1065
		return -ENOENT;
1066 1067
	}

1068 1069
	if (PageUptodate(page))
		return LOCKED_PAGE;
1070

1071
	fio.new_blkaddr = fio.old_blkaddr = ni.blk_addr;
1072
	return f2fs_submit_page_bio(&fio);
J
Jaegeuk Kim 已提交
1073 1074
}

J
Jaegeuk Kim 已提交
1075
/*
J
Jaegeuk Kim 已提交
1076 1077 1078 1079 1080
 * Readahead a node page
 */
void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid)
{
	struct page *apage;
1081
	int err;
J
Jaegeuk Kim 已提交
1082

1083 1084 1085 1086
	if (!nid)
		return;
	f2fs_bug_on(sbi, check_nid_range(sbi, nid));

1087 1088 1089 1090
	rcu_read_lock();
	apage = radix_tree_lookup(&NODE_MAPPING(sbi)->page_tree, nid);
	rcu_read_unlock();
	if (apage)
1091
		return;
J
Jaegeuk Kim 已提交
1092

1093
	apage = grab_cache_page(NODE_MAPPING(sbi), nid);
J
Jaegeuk Kim 已提交
1094 1095 1096
	if (!apage)
		return;

1097
	err = read_node_page(apage, READA);
1098
	f2fs_put_page(apage, err ? 1 : 0);
J
Jaegeuk Kim 已提交
1099 1100
}

1101 1102 1103 1104
/*
 * readahead MAX_RA_NODE number of node pages.
 */
void ra_node_pages(struct page *parent, int start)
J
Jaegeuk Kim 已提交
1105
{
1106 1107 1108 1109
	struct f2fs_sb_info *sbi = F2FS_P_SB(parent);
	struct blk_plug plug;
	int i, end;
	nid_t nid;
1110

1111
	blk_start_plug(&plug);
J
Jaegeuk Kim 已提交
1112

1113 1114 1115 1116 1117 1118
	/* Then, try readahead for siblings of the desired node */
	end = start + MAX_RA_NODE;
	end = min(end, NIDS_PER_BLOCK);
	for (i = start; i < end; i++) {
		nid = get_nid(parent, i, false);
		ra_node_page(sbi, nid);
1119
	}
J
Jaegeuk Kim 已提交
1120

1121
	blk_finish_plug(&plug);
J
Jaegeuk Kim 已提交
1122 1123
}

1124 1125
struct page *__get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid,
					struct page *parent, int start)
J
Jaegeuk Kim 已提交
1126 1127
{
	struct page *page;
1128
	int err;
J
Jaegeuk Kim 已提交
1129 1130 1131

	if (!nid)
		return ERR_PTR(-ENOENT);
1132
	f2fs_bug_on(sbi, check_nid_range(sbi, nid));
1133
repeat:
1134
	page = grab_cache_page(NODE_MAPPING(sbi), nid);
J
Jaegeuk Kim 已提交
1135 1136 1137
	if (!page)
		return ERR_PTR(-ENOMEM);

1138
	err = read_node_page(page, READ_SYNC);
1139 1140
	if (err < 0) {
		f2fs_put_page(page, 1);
J
Jaegeuk Kim 已提交
1141
		return ERR_PTR(err);
1142
	} else if (err == LOCKED_PAGE) {
1143
		goto page_hit;
1144
	}
J
Jaegeuk Kim 已提交
1145

1146 1147
	if (parent)
		ra_node_pages(parent, start + 1);
1148

J
Jaegeuk Kim 已提交
1149
	lock_page(page);
1150

C
Chao Yu 已提交
1151 1152 1153 1154
	if (unlikely(!PageUptodate(page))) {
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
	}
1155
	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
1156 1157 1158
		f2fs_put_page(page, 1);
		goto repeat;
	}
1159
page_hit:
C
Chao Yu 已提交
1160
	f2fs_bug_on(sbi, nid != nid_of_node(page));
J
Jaegeuk Kim 已提交
1161 1162 1163
	return page;
}

1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176
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);
}

J
Jaegeuk Kim 已提交
1177 1178
void sync_inode_page(struct dnode_of_data *dn)
{
1179 1180
	int ret = 0;

J
Jaegeuk Kim 已提交
1181
	if (IS_INODE(dn->node_page) || dn->inode_page == dn->node_page) {
1182
		ret = update_inode(dn->inode, dn->node_page);
J
Jaegeuk Kim 已提交
1183 1184 1185
	} else if (dn->inode_page) {
		if (!dn->inode_page_locked)
			lock_page(dn->inode_page);
1186
		ret = update_inode(dn->inode, dn->inode_page);
J
Jaegeuk Kim 已提交
1187 1188 1189
		if (!dn->inode_page_locked)
			unlock_page(dn->inode_page);
	} else {
1190
		ret = update_inode_page(dn->inode);
J
Jaegeuk Kim 已提交
1191
	}
1192
	dn->node_changed = ret ? true: false;
J
Jaegeuk Kim 已提交
1193 1194
}

1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino)
{
	struct inode *inode;
	struct page *page;

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

1205
	page = pagecache_get_page(inode->i_mapping, 0, FGP_NOWAIT, 0);
1206 1207 1208
	if (!page)
		goto iput_out;

1209 1210 1211
	if (!trylock_page(page))
		goto release_out;

1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225
	if (!PageUptodate(page))
		goto page_out;

	if (!PageDirty(page))
		goto page_out;

	if (!clear_page_dirty_for_io(page))
		goto page_out;

	if (!f2fs_write_inline_data(inode, page))
		inode_dec_dirty_pages(inode);
	else
		set_page_dirty(page);
page_out:
1226 1227 1228
	unlock_page(page);
release_out:
	f2fs_put_page(page, 0);
1229 1230 1231 1232
iput_out:
	iput(inode);
}

J
Jaegeuk Kim 已提交
1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244
int sync_node_pages(struct f2fs_sb_info *sbi, nid_t ino,
					struct writeback_control *wbc)
{
	pgoff_t index, end;
	struct pagevec pvec;
	int step = ino ? 2 : 0;
	int nwritten = 0, wrote = 0;

	pagevec_init(&pvec, 0);

next_step:
	index = 0;
1245
	end = ULONG_MAX;
J
Jaegeuk Kim 已提交
1246 1247 1248

	while (index <= end) {
		int i, nr_pages;
1249
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
J
Jaegeuk Kim 已提交
1250 1251 1252 1253 1254 1255 1256 1257
				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];

C
Chao Yu 已提交
1258 1259 1260 1261 1262
			if (unlikely(f2fs_cp_error(sbi))) {
				pagevec_release(&pvec);
				return -EIO;
			}

J
Jaegeuk Kim 已提交
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281
			/*
			 * 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;

			/*
			 * If an fsync mode,
			 * we should not skip writing node pages.
			 */
1282
lock_node:
J
Jaegeuk Kim 已提交
1283 1284 1285 1286 1287
			if (ino && ino_of_node(page) == ino)
				lock_page(page);
			else if (!trylock_page(page))
				continue;

1288
			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
J
Jaegeuk Kim 已提交
1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
continue_unlock:
				unlock_page(page);
				continue;
			}
			if (ino && ino_of_node(page) != ino)
				goto continue_unlock;

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

1301 1302 1303 1304 1305
			/* flush inline_data */
			if (!ino && is_inline_node(page)) {
				clear_inline_node(page);
				unlock_page(page);
				flush_inline_data(sbi, ino_of_node(page));
1306
				goto lock_node;
1307 1308
			}

1309 1310 1311
			f2fs_wait_on_page_writeback(page, NODE, true);

			BUG_ON(PageWriteback(page));
J
Jaegeuk Kim 已提交
1312 1313 1314 1315 1316 1317
			if (!clear_page_dirty_for_io(page))
				goto continue_unlock;

			/* called by fsync() */
			if (ino && IS_DNODE(page)) {
				set_fsync_mark(page, 1);
J
Jaegeuk Kim 已提交
1318 1319 1320
				if (IS_INODE(page))
					set_dentry_mark(page,
						need_dentry_mark(sbi, ino));
J
Jaegeuk Kim 已提交
1321 1322 1323 1324 1325
				nwritten++;
			} else {
				set_fsync_mark(page, 0);
				set_dentry_mark(page, 0);
			}
1326 1327 1328 1329 1330

			if (NODE_MAPPING(sbi)->a_ops->writepage(page, wbc))
				unlock_page(page);
			else
				wrote++;
J
Jaegeuk Kim 已提交
1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348

			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;
	}

1349 1350 1351 1352 1353 1354 1355
	if (wrote) {
		if (ino)
			f2fs_submit_merged_bio_cond(sbi, NULL, NULL,
							ino, NODE, WRITE);
		else
			f2fs_submit_merged_bio(sbi, NODE, WRITE);
	}
J
Jaegeuk Kim 已提交
1356 1357 1358
	return nwritten;
}

1359 1360
int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
{
1361
	pgoff_t index = 0, end = ULONG_MAX;
1362 1363 1364 1365
	struct pagevec pvec;
	int ret2 = 0, ret = 0;

	pagevec_init(&pvec, 0);
1366 1367 1368 1369 1370 1371 1372 1373

	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;
1374 1375 1376 1377 1378

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

			/* until radix tree lookup accepts end_index */
1379
			if (unlikely(page->index > end))
1380 1381
				continue;

1382
			if (ino && ino_of_node(page) == ino) {
1383
				f2fs_wait_on_page_writeback(page, NODE, true);
1384 1385 1386
				if (TestClearPageError(page))
					ret = -EIO;
			}
1387 1388 1389 1390 1391
		}
		pagevec_release(&pvec);
		cond_resched();
	}

1392
	if (unlikely(test_and_clear_bit(AS_ENOSPC, &NODE_MAPPING(sbi)->flags)))
1393
		ret2 = -ENOSPC;
1394
	if (unlikely(test_and_clear_bit(AS_EIO, &NODE_MAPPING(sbi)->flags)))
1395 1396 1397 1398 1399 1400
		ret2 = -EIO;
	if (!ret)
		ret = ret2;
	return ret;
}

J
Jaegeuk Kim 已提交
1401 1402 1403
static int f2fs_write_node_page(struct page *page,
				struct writeback_control *wbc)
{
1404
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1405 1406
	nid_t nid;
	struct node_info ni;
1407
	struct f2fs_io_info fio = {
1408
		.sbi = sbi,
1409
		.type = NODE,
C
Chris Fries 已提交
1410
		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
1411
		.page = page,
1412
		.encrypted_page = NULL,
1413
	};
J
Jaegeuk Kim 已提交
1414

1415 1416
	trace_f2fs_writepage(page, NODE);

1417
	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
1418
		goto redirty_out;
1419 1420
	if (unlikely(f2fs_cp_error(sbi)))
		goto redirty_out;
1421

J
Jaegeuk Kim 已提交
1422 1423
	/* get old block addr of this node page */
	nid = nid_of_node(page);
1424
	f2fs_bug_on(sbi, page->index != nid);
J
Jaegeuk Kim 已提交
1425

1426 1427 1428 1429 1430 1431 1432
	if (wbc->for_reclaim) {
		if (!down_read_trylock(&sbi->node_write))
			goto redirty_out;
	} else {
		down_read(&sbi->node_write);
	}

J
Jaegeuk Kim 已提交
1433 1434 1435
	get_node_info(sbi, nid, &ni);

	/* This page is already truncated */
1436
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1437
		ClearPageUptodate(page);
1438
		dec_page_count(sbi, F2FS_DIRTY_NODES);
1439
		up_read(&sbi->node_write);
1440 1441 1442
		unlock_page(page);
		return 0;
	}
J
Jaegeuk Kim 已提交
1443 1444

	set_page_writeback(page);
1445
	fio.old_blkaddr = ni.blk_addr;
1446
	write_node_page(nid, &fio);
1447
	set_node_addr(sbi, &ni, fio.new_blkaddr, is_fsync_dnode(page));
J
Jaegeuk Kim 已提交
1448
	dec_page_count(sbi, F2FS_DIRTY_NODES);
1449
	up_read(&sbi->node_write);
1450 1451 1452 1453

	if (wbc->for_reclaim)
		f2fs_submit_merged_bio_cond(sbi, NULL, page, 0, NODE, WRITE);

J
Jaegeuk Kim 已提交
1454
	unlock_page(page);
1455

1456
	if (unlikely(f2fs_cp_error(sbi)))
1457 1458
		f2fs_submit_merged_bio(sbi, NODE, WRITE);

J
Jaegeuk Kim 已提交
1459
	return 0;
1460 1461

redirty_out:
1462
	redirty_page_for_writepage(wbc, page);
1463
	return AOP_WRITEPAGE_ACTIVATE;
J
Jaegeuk Kim 已提交
1464 1465 1466 1467 1468
}

static int f2fs_write_node_pages(struct address_space *mapping,
			    struct writeback_control *wbc)
{
1469
	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
1470
	long diff;
J
Jaegeuk Kim 已提交
1471

1472 1473
	/* balancing f2fs's metadata in background */
	f2fs_balance_fs_bg(sbi);
J
Jaegeuk Kim 已提交
1474

1475
	/* collect a number of dirty node pages and write together */
1476
	if (get_pages(sbi, F2FS_DIRTY_NODES) < nr_pages_to_skip(sbi, NODE))
1477
		goto skip_write;
1478

Y
Yunlei He 已提交
1479 1480
	trace_f2fs_writepages(mapping->host, wbc, NODE);

1481
	diff = nr_pages_to_write(sbi, NODE, wbc);
1482
	wbc->sync_mode = WB_SYNC_NONE;
J
Jaegeuk Kim 已提交
1483
	sync_node_pages(sbi, 0, wbc);
1484
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
J
Jaegeuk Kim 已提交
1485
	return 0;
1486 1487 1488

skip_write:
	wbc->pages_skipped += get_pages(sbi, F2FS_DIRTY_NODES);
Y
Yunlei He 已提交
1489
	trace_f2fs_writepages(mapping->host, wbc, NODE);
1490
	return 0;
J
Jaegeuk Kim 已提交
1491 1492 1493 1494
}

static int f2fs_set_node_page_dirty(struct page *page)
{
1495 1496
	trace_f2fs_set_page_dirty(page, NODE);

J
Jaegeuk Kim 已提交
1497 1498 1499
	SetPageUptodate(page);
	if (!PageDirty(page)) {
		__set_page_dirty_nobuffers(page);
1500
		inc_page_count(F2FS_P_SB(page), F2FS_DIRTY_NODES);
J
Jaegeuk Kim 已提交
1501
		SetPagePrivate(page);
J
Jaegeuk Kim 已提交
1502
		f2fs_trace_pid(page);
J
Jaegeuk Kim 已提交
1503 1504 1505 1506 1507
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
1508
/*
J
Jaegeuk Kim 已提交
1509 1510 1511 1512 1513 1514
 * 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,
1515 1516
	.invalidatepage	= f2fs_invalidate_page,
	.releasepage	= f2fs_release_page,
J
Jaegeuk Kim 已提交
1517 1518
};

1519 1520
static struct free_nid *__lookup_free_nid_list(struct f2fs_nm_info *nm_i,
						nid_t n)
J
Jaegeuk Kim 已提交
1521
{
1522
	return radix_tree_lookup(&nm_i->free_nid_root, n);
J
Jaegeuk Kim 已提交
1523 1524
}

1525 1526
static void __del_from_free_nid_list(struct f2fs_nm_info *nm_i,
						struct free_nid *i)
J
Jaegeuk Kim 已提交
1527 1528
{
	list_del(&i->list);
1529
	radix_tree_delete(&nm_i->free_nid_root, i->nid);
J
Jaegeuk Kim 已提交
1530 1531
}

1532
static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build)
J
Jaegeuk Kim 已提交
1533
{
1534
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1535
	struct free_nid *i;
1536 1537
	struct nat_entry *ne;
	bool allocated = false;
J
Jaegeuk Kim 已提交
1538

1539
	if (!available_free_memory(sbi, FREE_NIDS))
H
Haicheng Li 已提交
1540
		return -1;
1541 1542

	/* 0 nid should not be used */
1543
	if (unlikely(nid == 0))
1544
		return 0;
1545

1546 1547 1548
	if (build) {
		/* do not add allocated nids */
		ne = __lookup_nat_cache(nm_i, nid);
1549
		if (ne && (!get_nat_flag(ne, IS_CHECKPOINTED) ||
1550
				nat_get_blkaddr(ne) != NULL_ADDR))
1551 1552 1553
			allocated = true;
		if (allocated)
			return 0;
J
Jaegeuk Kim 已提交
1554
	}
1555 1556

	i = f2fs_kmem_cache_alloc(free_nid_slab, GFP_NOFS);
J
Jaegeuk Kim 已提交
1557 1558 1559
	i->nid = nid;
	i->state = NID_NEW;

1560 1561 1562 1563 1564
	if (radix_tree_preload(GFP_NOFS)) {
		kmem_cache_free(free_nid_slab, i);
		return 0;
	}

J
Jaegeuk Kim 已提交
1565
	spin_lock(&nm_i->free_nid_list_lock);
1566
	if (radix_tree_insert(&nm_i->free_nid_root, i->nid, i)) {
J
Jaegeuk Kim 已提交
1567
		spin_unlock(&nm_i->free_nid_list_lock);
1568
		radix_tree_preload_end();
J
Jaegeuk Kim 已提交
1569 1570 1571 1572 1573 1574
		kmem_cache_free(free_nid_slab, i);
		return 0;
	}
	list_add_tail(&i->list, &nm_i->free_nid_list);
	nm_i->fcnt++;
	spin_unlock(&nm_i->free_nid_list_lock);
1575
	radix_tree_preload_end();
J
Jaegeuk Kim 已提交
1576 1577 1578 1579 1580 1581
	return 1;
}

static void remove_free_nid(struct f2fs_nm_info *nm_i, nid_t nid)
{
	struct free_nid *i;
1582 1583
	bool need_free = false;

J
Jaegeuk Kim 已提交
1584
	spin_lock(&nm_i->free_nid_list_lock);
1585
	i = __lookup_free_nid_list(nm_i, nid);
J
Jaegeuk Kim 已提交
1586
	if (i && i->state == NID_NEW) {
1587
		__del_from_free_nid_list(nm_i, i);
J
Jaegeuk Kim 已提交
1588
		nm_i->fcnt--;
1589
		need_free = true;
J
Jaegeuk Kim 已提交
1590 1591
	}
	spin_unlock(&nm_i->free_nid_list_lock);
1592 1593 1594

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1595 1596
}

1597
static void scan_nat_page(struct f2fs_sb_info *sbi,
J
Jaegeuk Kim 已提交
1598 1599
			struct page *nat_page, nid_t start_nid)
{
1600
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1601 1602 1603 1604 1605 1606 1607
	struct f2fs_nat_block *nat_blk = page_address(nat_page);
	block_t blk_addr;
	int i;

	i = start_nid % NAT_ENTRY_PER_BLOCK;

	for (; i < NAT_ENTRY_PER_BLOCK; i++, start_nid++) {
H
Haicheng Li 已提交
1608

1609
		if (unlikely(start_nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1610
			break;
H
Haicheng Li 已提交
1611 1612

		blk_addr = le32_to_cpu(nat_blk->entries[i].block_addr);
1613
		f2fs_bug_on(sbi, blk_addr == NEW_ADDR);
H
Haicheng Li 已提交
1614
		if (blk_addr == NULL_ADDR) {
1615
			if (add_free_nid(sbi, start_nid, true) < 0)
H
Haicheng Li 已提交
1616 1617
				break;
		}
J
Jaegeuk Kim 已提交
1618 1619 1620 1621 1622 1623 1624
	}
}

static void build_free_nids(struct f2fs_sb_info *sbi)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
1625
	struct f2fs_journal *journal = curseg->journal;
1626
	int i = 0;
1627
	nid_t nid = nm_i->next_scan_nid;
J
Jaegeuk Kim 已提交
1628

1629 1630 1631
	/* Enough entries */
	if (nm_i->fcnt > NAT_ENTRY_PER_BLOCK)
		return;
J
Jaegeuk Kim 已提交
1632

1633
	/* readahead nat pages to be scanned */
1634 1635
	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nid), FREE_NID_PAGES,
							META_NAT, true);
J
Jaegeuk Kim 已提交
1636

1637 1638
	down_read(&nm_i->nat_tree_lock);

J
Jaegeuk Kim 已提交
1639 1640 1641
	while (1) {
		struct page *page = get_current_nat_page(sbi, nid);

1642
		scan_nat_page(sbi, page, nid);
J
Jaegeuk Kim 已提交
1643 1644 1645
		f2fs_put_page(page, 1);

		nid += (NAT_ENTRY_PER_BLOCK - (nid % NAT_ENTRY_PER_BLOCK));
1646
		if (unlikely(nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1647
			nid = 0;
1648

1649
		if (++i >= FREE_NID_PAGES)
J
Jaegeuk Kim 已提交
1650 1651 1652
			break;
	}

1653 1654
	/* go to the next free nat pages to find free nids abundantly */
	nm_i->next_scan_nid = nid;
J
Jaegeuk Kim 已提交
1655 1656

	/* find free nids from current sum_pages */
1657
	down_read(&curseg->journal_rwsem);
1658 1659 1660 1661 1662
	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
1663
		if (addr == NULL_ADDR)
1664
			add_free_nid(sbi, nid, true);
J
Jaegeuk Kim 已提交
1665 1666 1667
		else
			remove_free_nid(nm_i, nid);
	}
1668
	up_read(&curseg->journal_rwsem);
1669
	up_read(&nm_i->nat_tree_lock);
C
Chao Yu 已提交
1670 1671

	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nm_i->next_scan_nid),
C
Chao Yu 已提交
1672
					nm_i->ra_nid_pages, META_NAT, false);
J
Jaegeuk Kim 已提交
1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684
}

/*
 * 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:
1685
	if (unlikely(sbi->total_valid_node_count + 1 > nm_i->available_nids))
1686
		return false;
J
Jaegeuk Kim 已提交
1687 1688 1689

	spin_lock(&nm_i->free_nid_list_lock);

1690
	/* We should not use stale free nids created by build_free_nids */
1691
	if (nm_i->fcnt && !on_build_free_nids(nm_i)) {
1692
		f2fs_bug_on(sbi, list_empty(&nm_i->free_nid_list));
1693
		list_for_each_entry(i, &nm_i->free_nid_list, list)
1694 1695
			if (i->state == NID_NEW)
				break;
J
Jaegeuk Kim 已提交
1696

1697
		f2fs_bug_on(sbi, i->state != NID_NEW);
1698 1699 1700 1701 1702 1703
		*nid = i->nid;
		i->state = NID_ALLOC;
		nm_i->fcnt--;
		spin_unlock(&nm_i->free_nid_list_lock);
		return true;
	}
J
Jaegeuk Kim 已提交
1704
	spin_unlock(&nm_i->free_nid_list_lock);
1705 1706 1707 1708 1709 1710

	/* Let's scan nat pages and its caches to get free nids */
	mutex_lock(&nm_i->build_lock);
	build_free_nids(sbi);
	mutex_unlock(&nm_i->build_lock);
	goto retry;
J
Jaegeuk Kim 已提交
1711 1712
}

J
Jaegeuk Kim 已提交
1713
/*
J
Jaegeuk Kim 已提交
1714 1715 1716 1717 1718 1719 1720 1721
 * 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;

	spin_lock(&nm_i->free_nid_list_lock);
1722
	i = __lookup_free_nid_list(nm_i, nid);
1723
	f2fs_bug_on(sbi, !i || i->state != NID_ALLOC);
1724
	__del_from_free_nid_list(nm_i, i);
J
Jaegeuk Kim 已提交
1725
	spin_unlock(&nm_i->free_nid_list_lock);
1726 1727

	kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1728 1729
}

J
Jaegeuk Kim 已提交
1730
/*
J
Jaegeuk Kim 已提交
1731 1732 1733 1734
 * alloc_nid() should be called prior to this function.
 */
void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid)
{
1735 1736
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i;
1737
	bool need_free = false;
1738

J
Jaegeuk Kim 已提交
1739 1740 1741
	if (!nid)
		return;

1742
	spin_lock(&nm_i->free_nid_list_lock);
1743
	i = __lookup_free_nid_list(nm_i, nid);
1744
	f2fs_bug_on(sbi, !i || i->state != NID_ALLOC);
1745
	if (!available_free_memory(sbi, FREE_NIDS)) {
1746
		__del_from_free_nid_list(nm_i, i);
1747
		need_free = true;
1748 1749 1750 1751
	} else {
		i->state = NID_NEW;
		nm_i->fcnt++;
	}
1752
	spin_unlock(&nm_i->free_nid_list_lock);
1753 1754 1755

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1756 1757
}

C
Chao Yu 已提交
1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774
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;

	if (!mutex_trylock(&nm_i->build_lock))
		return 0;

	spin_lock(&nm_i->free_nid_list_lock);
	list_for_each_entry_safe(i, next, &nm_i->free_nid_list, list) {
		if (nr_shrink <= 0 || nm_i->fcnt <= NAT_ENTRY_PER_BLOCK)
			break;
		if (i->state == NID_ALLOC)
			continue;
		__del_from_free_nid_list(nm_i, i);
		kmem_cache_free(free_nid_slab, i);
1775
		nm_i->fcnt--;
C
Chao Yu 已提交
1776 1777 1778 1779 1780 1781 1782 1783
		nr_shrink--;
	}
	spin_unlock(&nm_i->free_nid_list_lock);
	mutex_unlock(&nm_i->build_lock);

	return nr - nr_shrink;
}

1784
void recover_inline_xattr(struct inode *inode, struct page *page)
1785 1786 1787 1788 1789 1790
{
	void *src_addr, *dst_addr;
	size_t inline_size;
	struct page *ipage;
	struct f2fs_inode *ri;

1791
	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
1792
	f2fs_bug_on(F2FS_I_SB(inode), IS_ERR(ipage));
1793

1794 1795 1796 1797 1798 1799
	ri = F2FS_INODE(page);
	if (!(ri->i_inline & F2FS_INLINE_XATTR)) {
		clear_inode_flag(F2FS_I(inode), FI_INLINE_XATTR);
		goto update_inode;
	}

1800 1801 1802 1803
	dst_addr = inline_xattr_addr(ipage);
	src_addr = inline_xattr_addr(page);
	inline_size = inline_xattr_size(inode);

1804
	f2fs_wait_on_page_writeback(ipage, NODE, true);
1805
	memcpy(dst_addr, src_addr, inline_size);
1806
update_inode:
1807 1808 1809 1810
	update_inode(inode, ipage);
	f2fs_put_page(ipage, 1);
}

1811
void recover_xattr_data(struct inode *inode, struct page *page, block_t blkaddr)
1812
{
1813
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1814 1815 1816 1817 1818 1819 1820 1821 1822 1823
	nid_t prev_xnid = F2FS_I(inode)->i_xattr_nid;
	nid_t new_xnid = nid_of_node(page);
	struct node_info ni;

	/* 1: invalidate the previous xattr nid */
	if (!prev_xnid)
		goto recover_xnid;

	/* Deallocate node address */
	get_node_info(sbi, prev_xnid, &ni);
1824
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
1825 1826
	invalidate_blocks(sbi, ni.blk_addr);
	dec_valid_node_count(sbi, inode);
1827
	set_node_addr(sbi, &ni, NULL_ADDR, false);
1828 1829 1830 1831

recover_xnid:
	/* 2: allocate new xattr nid */
	if (unlikely(!inc_valid_node_count(sbi, inode)))
1832
		f2fs_bug_on(sbi, 1);
1833 1834 1835 1836

	remove_free_nid(NM_I(sbi), new_xnid);
	get_node_info(sbi, new_xnid, &ni);
	ni.ino = inode->i_ino;
1837
	set_node_addr(sbi, &ni, NEW_ADDR, false);
1838 1839 1840 1841
	F2FS_I(inode)->i_xattr_nid = new_xnid;

	/* 3: update xattr blkaddr */
	refresh_sit_entry(sbi, NEW_ADDR, blkaddr);
1842
	set_node_addr(sbi, &ni, blkaddr, false);
1843 1844 1845 1846

	update_inode_page(inode);
}

J
Jaegeuk Kim 已提交
1847 1848
int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
{
1849
	struct f2fs_inode *src, *dst;
J
Jaegeuk Kim 已提交
1850 1851 1852 1853
	nid_t ino = ino_of_node(page);
	struct node_info old_ni, new_ni;
	struct page *ipage;

1854 1855 1856 1857 1858
	get_node_info(sbi, ino, &old_ni);

	if (unlikely(old_ni.blk_addr != NULL_ADDR))
		return -EINVAL;

1859
	ipage = grab_cache_page(NODE_MAPPING(sbi), ino);
J
Jaegeuk Kim 已提交
1860 1861 1862
	if (!ipage)
		return -ENOMEM;

A
arter97 已提交
1863
	/* Should not use this inode from free nid list */
J
Jaegeuk Kim 已提交
1864 1865 1866 1867 1868
	remove_free_nid(NM_I(sbi), ino);

	SetPageUptodate(ipage);
	fill_node_footer(ipage, ino, ino, 0, true);

1869 1870
	src = F2FS_INODE(page);
	dst = F2FS_INODE(ipage);
J
Jaegeuk Kim 已提交
1871

1872 1873 1874 1875 1876
	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;
1877
	dst->i_inline = src->i_inline & F2FS_INLINE_XATTR;
J
Jaegeuk Kim 已提交
1878 1879 1880 1881

	new_ni = old_ni;
	new_ni.ino = ino;

1882
	if (unlikely(!inc_valid_node_count(sbi, NULL)))
1883
		WARN_ON(1);
1884
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
J
Jaegeuk Kim 已提交
1885
	inc_valid_inode_count(sbi);
1886
	set_page_dirty(ipage);
J
Jaegeuk Kim 已提交
1887 1888 1889 1890 1891 1892 1893 1894 1895 1896
	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;
J
Jaegeuk Kim 已提交
1897
	int bio_blocks = MAX_BIO_BLOCKS(sbi);
1898
	int i, idx, last_offset, nrpages;
J
Jaegeuk Kim 已提交
1899 1900 1901 1902 1903 1904

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

1905
	for (i = 0; i < last_offset; i += nrpages, addr += nrpages) {
1906
		nrpages = min(last_offset - i, bio_blocks);
1907

A
arter97 已提交
1908
		/* readahead node pages */
1909
		ra_meta_pages(sbi, addr, nrpages, META_POR, true);
J
Jaegeuk Kim 已提交
1910

1911
		for (idx = addr; idx < addr + nrpages; idx++) {
1912
			struct page *page = get_tmp_page(sbi, idx);
1913

1914 1915 1916 1917 1918 1919
			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);
1920
		}
1921

1922
		invalidate_mapping_pages(META_MAPPING(sbi), addr,
1923
							addr + nrpages);
J
Jaegeuk Kim 已提交
1924
	}
1925
	return 0;
J
Jaegeuk Kim 已提交
1926 1927
}

1928
static void remove_nats_in_journal(struct f2fs_sb_info *sbi)
J
Jaegeuk Kim 已提交
1929 1930 1931
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
1932
	struct f2fs_journal *journal = curseg->journal;
J
Jaegeuk Kim 已提交
1933 1934
	int i;

1935
	down_write(&curseg->journal_rwsem);
1936
	for (i = 0; i < nats_in_cursum(journal); i++) {
J
Jaegeuk Kim 已提交
1937 1938
		struct nat_entry *ne;
		struct f2fs_nat_entry raw_ne;
1939
		nid_t nid = le32_to_cpu(nid_in_journal(journal, i));
J
Jaegeuk Kim 已提交
1940

1941
		raw_ne = nat_in_journal(journal, i);
1942

J
Jaegeuk Kim 已提交
1943 1944
		ne = __lookup_nat_cache(nm_i, nid);
		if (!ne) {
1945 1946
			ne = grab_nat_entry(nm_i, nid);
			node_info_from_raw_nat(&ne->ni, &raw_ne);
J
Jaegeuk Kim 已提交
1947 1948 1949
		}
		__set_nat_cache_dirty(nm_i, ne);
	}
1950
	update_nats_in_cursum(journal, -i);
1951
	up_write(&curseg->journal_rwsem);
J
Jaegeuk Kim 已提交
1952 1953
}

1954 1955
static void __adjust_nat_entry_set(struct nat_entry_set *nes,
						struct list_head *head, int max)
J
Jaegeuk Kim 已提交
1956
{
1957
	struct nat_entry_set *cur;
J
Jaegeuk Kim 已提交
1958

1959 1960
	if (nes->entry_cnt >= max)
		goto add_out;
J
Jaegeuk Kim 已提交
1961

1962 1963 1964 1965 1966
	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;
		}
1967
	}
1968 1969 1970
add_out:
	list_add_tail(&nes->set_list, head);
}
J
Jaegeuk Kim 已提交
1971

1972 1973 1974 1975
static void __flush_nat_entry_set(struct f2fs_sb_info *sbi,
					struct nat_entry_set *set)
{
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
1976
	struct f2fs_journal *journal = curseg->journal;
1977 1978 1979 1980 1981
	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
1982

1983 1984 1985 1986 1987
	/*
	 * 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.
	 */
1988
	if (!__has_cursum_space(journal, set->entry_cnt, NAT_JOURNAL))
1989 1990 1991
		to_journal = false;

	if (to_journal) {
1992
		down_write(&curseg->journal_rwsem);
1993 1994 1995 1996 1997
	} else {
		page = get_next_nat_page(sbi, start_nid);
		nat_blk = page_address(page);
		f2fs_bug_on(sbi, !nat_blk);
	}
1998

1999 2000 2001 2002 2003 2004 2005 2006
	/* 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;
2007 2008

		if (to_journal) {
2009
			offset = lookup_journal_in_cursum(journal,
2010 2011
							NAT_JOURNAL, nid, 1);
			f2fs_bug_on(sbi, offset < 0);
2012 2013
			raw_ne = &nat_in_journal(journal, offset);
			nid_in_journal(journal, offset) = cpu_to_le32(nid);
2014
		} else {
2015
			raw_ne = &nat_blk->entries[nid - start_nid];
J
Jaegeuk Kim 已提交
2016
		}
2017 2018 2019 2020 2021 2022
		raw_nat_from_node_info(raw_ne, &ne->ni);
		nat_reset_flag(ne);
		__clear_nat_cache_dirty(NM_I(sbi), ne);
		if (nat_get_blkaddr(ne) == NULL_ADDR)
			add_free_nid(sbi, nid, false);
	}
J
Jaegeuk Kim 已提交
2023

2024
	if (to_journal)
2025
		up_write(&curseg->journal_rwsem);
2026 2027
	else
		f2fs_put_page(page, 1);
2028

2029 2030 2031 2032
	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);
2033
}
2034

2035 2036 2037 2038 2039 2040 2041
/*
 * This function is called during the checkpointing process.
 */
void flush_nat_entries(struct f2fs_sb_info *sbi)
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
2042
	struct f2fs_journal *journal = curseg->journal;
2043
	struct nat_entry_set *setvec[SETVEC_SIZE];
2044 2045 2046 2047 2048
	struct nat_entry_set *set, *tmp;
	unsigned int found;
	nid_t set_idx = 0;
	LIST_HEAD(sets);

2049 2050
	if (!nm_i->dirty_nat_cnt)
		return;
2051 2052 2053

	down_write(&nm_i->nat_tree_lock);

2054 2055 2056 2057 2058
	/*
	 * 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.
	 */
2059
	if (!__has_cursum_space(journal, nm_i->dirty_nat_cnt, NAT_JOURNAL))
2060 2061 2062
		remove_nats_in_journal(sbi);

	while ((found = __gang_lookup_nat_set(nm_i,
2063
					set_idx, SETVEC_SIZE, setvec))) {
2064 2065 2066 2067
		unsigned idx;
		set_idx = setvec[found - 1]->set + 1;
		for (idx = 0; idx < found; idx++)
			__adjust_nat_entry_set(setvec[idx], &sets,
2068
						MAX_NAT_JENTRIES(journal));
J
Jaegeuk Kim 已提交
2069
	}
2070

2071 2072 2073 2074
	/* flush dirty nats in nat entry set */
	list_for_each_entry_safe(set, tmp, &sets, set_list)
		__flush_nat_entry_set(sbi, set);

2075 2076
	up_write(&nm_i->nat_tree_lock);

2077
	f2fs_bug_on(sbi, nm_i->dirty_nat_cnt);
J
Jaegeuk Kim 已提交
2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091
}

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;
	unsigned int nat_segs, nat_blocks;

	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;
	nat_blocks = nat_segs << le32_to_cpu(sb_raw->log_blocks_per_seg);
2092

2093 2094
	nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nat_blocks;

2095
	/* not used nids: 0, node, meta, (and root counted as valid node) */
2096
	nm_i->available_nids = nm_i->max_nid - F2FS_RESERVED_NODE_NUM;
J
Jaegeuk Kim 已提交
2097 2098
	nm_i->fcnt = 0;
	nm_i->nat_cnt = 0;
2099
	nm_i->ram_thresh = DEF_RAM_THRESHOLD;
C
Chao Yu 已提交
2100
	nm_i->ra_nid_pages = DEF_RA_NID_PAGES;
C
Chao Yu 已提交
2101
	nm_i->dirty_nats_ratio = DEF_DIRTY_NAT_RATIO_THRESHOLD;
J
Jaegeuk Kim 已提交
2102

2103
	INIT_RADIX_TREE(&nm_i->free_nid_root, GFP_ATOMIC);
J
Jaegeuk Kim 已提交
2104
	INIT_LIST_HEAD(&nm_i->free_nid_list);
2105 2106
	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
2107 2108 2109 2110
	INIT_LIST_HEAD(&nm_i->nat_entries);

	mutex_init(&nm_i->build_lock);
	spin_lock_init(&nm_i->free_nid_list_lock);
2111
	init_rwsem(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2112 2113

	nm_i->next_scan_nid = le32_to_cpu(sbi->ckpt->next_free_nid);
A
Alexandru Gheorghiu 已提交
2114
	nm_i->bitmap_size = __bitmap_size(sbi, NAT_BITMAP);
J
Jaegeuk Kim 已提交
2115 2116 2117 2118
	version_bitmap = __bitmap_ptr(sbi, NAT_BITMAP);
	if (!version_bitmap)
		return -EFAULT;

A
Alexandru Gheorghiu 已提交
2119 2120 2121 2122
	nm_i->nat_bitmap = kmemdup(version_bitmap, nm_i->bitmap_size,
					GFP_KERNEL);
	if (!nm_i->nat_bitmap)
		return -ENOMEM;
J
Jaegeuk Kim 已提交
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146
	return 0;
}

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;

	build_free_nids(sbi);
	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];
2147
	struct nat_entry_set *setvec[SETVEC_SIZE];
J
Jaegeuk Kim 已提交
2148 2149 2150 2151 2152 2153 2154 2155 2156
	nid_t nid = 0;
	unsigned int found;

	if (!nm_i)
		return;

	/* destroy free nid list */
	spin_lock(&nm_i->free_nid_list_lock);
	list_for_each_entry_safe(i, next_i, &nm_i->free_nid_list, list) {
2157
		f2fs_bug_on(sbi, i->state == NID_ALLOC);
2158
		__del_from_free_nid_list(nm_i, i);
J
Jaegeuk Kim 已提交
2159
		nm_i->fcnt--;
2160 2161 2162
		spin_unlock(&nm_i->free_nid_list_lock);
		kmem_cache_free(free_nid_slab, i);
		spin_lock(&nm_i->free_nid_list_lock);
J
Jaegeuk Kim 已提交
2163
	}
2164
	f2fs_bug_on(sbi, nm_i->fcnt);
J
Jaegeuk Kim 已提交
2165 2166 2167
	spin_unlock(&nm_i->free_nid_list_lock);

	/* destroy nat cache */
2168
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2169 2170 2171
	while ((found = __gang_lookup_nat_cache(nm_i,
					nid, NATVEC_SIZE, natvec))) {
		unsigned idx;
2172

2173 2174 2175
		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
2176
	}
2177
	f2fs_bug_on(sbi, nm_i->nat_cnt);
2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192

	/* 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]);
		}
	}
2193
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2194 2195 2196 2197 2198 2199

	kfree(nm_i->nat_bitmap);
	sbi->nm_info = NULL;
	kfree(nm_i);
}

2200
int __init create_node_manager_caches(void)
J
Jaegeuk Kim 已提交
2201 2202
{
	nat_entry_slab = f2fs_kmem_cache_create("nat_entry",
2203
			sizeof(struct nat_entry));
J
Jaegeuk Kim 已提交
2204
	if (!nat_entry_slab)
2205
		goto fail;
J
Jaegeuk Kim 已提交
2206 2207

	free_nid_slab = f2fs_kmem_cache_create("free_nid",
2208
			sizeof(struct free_nid));
2209
	if (!free_nid_slab)
2210
		goto destroy_nat_entry;
2211 2212 2213 2214

	nat_entry_set_slab = f2fs_kmem_cache_create("nat_entry_set",
			sizeof(struct nat_entry_set));
	if (!nat_entry_set_slab)
2215
		goto destroy_free_nid;
J
Jaegeuk Kim 已提交
2216
	return 0;
2217

2218
destroy_free_nid:
2219
	kmem_cache_destroy(free_nid_slab);
2220
destroy_nat_entry:
2221 2222 2223
	kmem_cache_destroy(nat_entry_slab);
fail:
	return -ENOMEM;
J
Jaegeuk Kim 已提交
2224 2225 2226 2227
}

void destroy_node_manager_caches(void)
{
2228
	kmem_cache_destroy(nat_entry_set_slab);
J
Jaegeuk Kim 已提交
2229 2230 2231
	kmem_cache_destroy(free_nid_slab);
	kmem_cache_destroy(nat_entry_slab);
}