node.c 57.9 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
}

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

252
	new = f2fs_kmem_cache_alloc(nat_entry_slab, GFP_NOFS);
253
	f2fs_radix_tree_insert(&nm_i->nat_root, nid, new);
J
Jaegeuk Kim 已提交
254 255
	memset(new, 0, sizeof(struct nat_entry));
	nat_set_nid(new, nid);
256
	nat_reset_flag(new);
J
Jaegeuk Kim 已提交
257 258 259 260 261
	list_add_tail(&new->list, &nm_i->nat_entries);
	nm_i->nat_cnt++;
	return new;
}

262
static void cache_nat_entry(struct f2fs_sb_info *sbi, nid_t nid,
J
Jaegeuk Kim 已提交
263 264
						struct f2fs_nat_entry *ne)
{
265
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
266
	struct nat_entry *e;
267

J
Jaegeuk Kim 已提交
268 269 270
	e = __lookup_nat_cache(nm_i, nid);
	if (!e) {
		e = grab_nat_entry(nm_i, nid);
271
		node_info_from_raw_nat(&e->ni, ne);
272
	} else {
E
Eric Biggers 已提交
273 274 275
		f2fs_bug_on(sbi, nat_get_ino(e) != le32_to_cpu(ne->ino) ||
				nat_get_blkaddr(e) !=
					le32_to_cpu(ne->block_addr) ||
276
				nat_get_version(e) != ne->version);
J
Jaegeuk Kim 已提交
277 278 279 280
	}
}

static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni,
281
			block_t new_blkaddr, bool fsync_done)
J
Jaegeuk Kim 已提交
282 283 284
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
285

286
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
287 288 289
	e = __lookup_nat_cache(nm_i, ni->nid);
	if (!e) {
		e = grab_nat_entry(nm_i, ni->nid);
290
		copy_node_info(&e->ni, ni);
291
		f2fs_bug_on(sbi, ni->blk_addr == NEW_ADDR);
J
Jaegeuk Kim 已提交
292 293 294 295 296 297
	} 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.
		 */
298
		copy_node_info(&e->ni, ni);
299
		f2fs_bug_on(sbi, ni->blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
300 301 302
	}

	/* sanity check */
303 304
	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
305
			new_blkaddr == NULL_ADDR);
306
	f2fs_bug_on(sbi, nat_get_blkaddr(e) == NEW_ADDR &&
J
Jaegeuk Kim 已提交
307
			new_blkaddr == NEW_ADDR);
308
	f2fs_bug_on(sbi, nat_get_blkaddr(e) != NEW_ADDR &&
J
Jaegeuk Kim 已提交
309 310 311
			nat_get_blkaddr(e) != NULL_ADDR &&
			new_blkaddr == NEW_ADDR);

A
arter97 已提交
312
	/* increment version no as node is removed */
J
Jaegeuk Kim 已提交
313 314 315
	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
316 317 318 319

		/* 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
320 321 322 323
	}

	/* change address */
	nat_set_blkaddr(e, new_blkaddr);
324 325
	if (new_blkaddr == NEW_ADDR || new_blkaddr == NULL_ADDR)
		set_nat_flag(e, IS_CHECKPOINTED, false);
J
Jaegeuk Kim 已提交
326
	__set_nat_cache_dirty(nm_i, e);
327 328

	/* update fsync_mark if its inode nat entry is still alive */
329 330
	if (ni->nid != ni->ino)
		e = __lookup_nat_cache(nm_i, ni->ino);
331 332 333 334 335
	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);
	}
336
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
337 338
}

339
int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
J
Jaegeuk Kim 已提交
340 341
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
342
	int nr = nr_shrink;
J
Jaegeuk Kim 已提交
343

344 345
	if (!down_write_trylock(&nm_i->nat_tree_lock))
		return 0;
J
Jaegeuk Kim 已提交
346 347 348 349 350 351 352 353

	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--;
	}
354
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
355
	return nr - nr_shrink;
J
Jaegeuk Kim 已提交
356 357
}

J
Jaegeuk Kim 已提交
358
/*
A
arter97 已提交
359
 * This function always returns success
J
Jaegeuk Kim 已提交
360 361 362 363 364
 */
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);
365
	struct f2fs_journal *journal = curseg->journal;
J
Jaegeuk Kim 已提交
366 367 368 369 370 371 372 373 374 375
	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 */
376
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
377 378 379 380 381
	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);
382
		up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
383
		return;
384
	}
J
Jaegeuk Kim 已提交
385

386 387
	memset(&ne, 0, sizeof(struct f2fs_nat_entry));

J
Jaegeuk Kim 已提交
388
	/* Check current segment summary */
389
	down_read(&curseg->journal_rwsem);
390
	i = lookup_journal_in_cursum(journal, NAT_JOURNAL, nid, 0);
J
Jaegeuk Kim 已提交
391
	if (i >= 0) {
392
		ne = nat_in_journal(journal, i);
J
Jaegeuk Kim 已提交
393 394
		node_info_from_raw_nat(ni, &ne);
	}
395
	up_read(&curseg->journal_rwsem);
J
Jaegeuk Kim 已提交
396 397 398 399 400 401 402 403 404 405
	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:
406
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
407
	/* cache nat entry */
408
	down_write(&nm_i->nat_tree_lock);
409
	cache_nat_entry(sbi, nid, &ne);
410
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
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
/*
 * 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);
}

436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466
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
467
/*
J
Jaegeuk Kim 已提交
468 469 470
 * The maximum depth is four.
 * Offset[0] will have raw inode offset.
 */
471
static int get_node_path(struct inode *inode, long block,
472
				int offset[4], unsigned int noffset[4])
J
Jaegeuk Kim 已提交
473
{
474
	const long direct_index = ADDRS_PER_INODE(inode);
J
Jaegeuk Kim 已提交
475 476 477 478 479 480 481 482 483 484
	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) {
485
		offset[n] = block;
J
Jaegeuk Kim 已提交
486 487 488 489 490 491
		goto got;
	}
	block -= direct_index;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR1_BLOCK;
		noffset[n] = 1;
492
		offset[n] = block;
J
Jaegeuk Kim 已提交
493 494 495 496 497 498 499
		level = 1;
		goto got;
	}
	block -= direct_blks;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR2_BLOCK;
		noffset[n] = 2;
500
		offset[n] = block;
J
Jaegeuk Kim 已提交
501 502 503 504 505 506 507 508 509
		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];
510
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
511 512 513 514 515 516 517 518 519
		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];
520
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
521 522 523 524 525 526 527 528 529 530 531 532 533 534
		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];
535
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
536 537 538 539 540 541 542 543 544 545 546
		level = 3;
		goto got;
	} else {
		BUG();
	}
got:
	return level;
}

/*
 * Caller should call f2fs_put_dnode(dn).
C
Chao Yu 已提交
547 548
 * 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.
549
 * In the case of RDONLY_NODE, we don't need to care about mutex.
J
Jaegeuk Kim 已提交
550
 */
551
int get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
J
Jaegeuk Kim 已提交
552
{
553
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
554
	struct page *npage[4];
555
	struct page *parent = NULL;
J
Jaegeuk Kim 已提交
556 557 558
	int offset[4];
	unsigned int noffset[4];
	nid_t nids[4];
559
	int level, i = 0;
J
Jaegeuk Kim 已提交
560 561
	int err = 0;

562
	level = get_node_path(dn->inode, index, offset, noffset);
J
Jaegeuk Kim 已提交
563 564

	nids[0] = dn->inode->i_ino;
565
	npage[0] = dn->inode_page;
J
Jaegeuk Kim 已提交
566

567 568 569 570 571
	if (!npage[0]) {
		npage[0] = get_node_page(sbi, nids[0]);
		if (IS_ERR(npage[0]))
			return PTR_ERR(npage[0]);
	}
572 573 574

	/* if inline_data is set, should not report any block indices */
	if (f2fs_has_inline_data(dn->inode) && index) {
575
		err = -ENOENT;
576 577 578 579
		f2fs_put_page(npage[0], 1);
		goto release_out;
	}

J
Jaegeuk Kim 已提交
580
	parent = npage[0];
581 582
	if (level != 0)
		nids[1] = get_nid(parent, offset[0], true);
J
Jaegeuk Kim 已提交
583 584 585 586 587 588 589
	dn->inode_page = npage[0];
	dn->inode_page_locked = true;

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

590
		if (!nids[i] && mode == ALLOC_NODE) {
J
Jaegeuk Kim 已提交
591 592 593 594 595 596 597
			/* alloc new node */
			if (!alloc_nid(sbi, &(nids[i]))) {
				err = -ENOSPC;
				goto release_pages;
			}

			dn->nid = nids[i];
598
			npage[i] = new_node_page(dn, noffset[i], NULL);
J
Jaegeuk Kim 已提交
599 600 601 602 603 604 605 606 607
			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;
608
		} else if (mode == LOOKUP_NODE_RA && i == level && level > 1) {
J
Jaegeuk Kim 已提交
609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648
			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;
649 650 651
	if (err == -ENOENT) {
		dn->cur_level = i;
		dn->max_level = level;
652
		dn->ofs_in_node = offset[level];
653
	}
J
Jaegeuk Kim 已提交
654 655 656 657 658
	return err;
}

static void truncate_node(struct dnode_of_data *dn)
{
659
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
660 661 662
	struct node_info ni;

	get_node_info(sbi, dn->nid, &ni);
663
	if (dn->inode->i_blocks == 0) {
664
		f2fs_bug_on(sbi, ni.blk_addr != NULL_ADDR);
665 666
		goto invalidate;
	}
667
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
J
Jaegeuk Kim 已提交
668 669

	/* Deallocate node address */
670
	invalidate_blocks(sbi, ni.blk_addr);
671
	dec_valid_node_count(sbi, dn->inode);
672
	set_node_addr(sbi, &ni, NULL_ADDR, false);
J
Jaegeuk Kim 已提交
673 674 675 676

	if (dn->nid == dn->inode->i_ino) {
		remove_orphan_inode(sbi, dn->nid);
		dec_valid_inode_count(sbi);
677
		f2fs_inode_synced(dn->inode);
J
Jaegeuk Kim 已提交
678
	}
679
invalidate:
J
Jaegeuk Kim 已提交
680
	clear_node_page_dirty(dn->node_page);
681
	set_sbi_flag(sbi, SBI_IS_DIRTY);
J
Jaegeuk Kim 已提交
682 683

	f2fs_put_page(dn->node_page, 1);
684 685 686 687

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

J
Jaegeuk Kim 已提交
688
	dn->node_page = NULL;
689
	trace_f2fs_truncate_node(dn->inode, dn->nid, ni.blk_addr);
J
Jaegeuk Kim 已提交
690 691 692 693 694 695 696 697 698 699
}

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

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

	/* get direct node */
700
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
J
Jaegeuk Kim 已提交
701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727
	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;

728 729
	trace_f2fs_truncate_nodes_enter(dn->inode, dn->nid, dn->data_blkaddr);

730
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
731 732
	if (IS_ERR(page)) {
		trace_f2fs_truncate_nodes_exit(dn->inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
733
		return PTR_ERR(page);
734
	}
J
Jaegeuk Kim 已提交
735

736 737
	ra_node_pages(page, ofs, NIDS_PER_BLOCK);

738
	rn = F2FS_NODE(page);
J
Jaegeuk Kim 已提交
739 740 741 742 743 744 745 746 747
	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;
748 749
			if (set_nid(page, i, 0, false))
				dn->node_changed = true;
J
Jaegeuk Kim 已提交
750 751 752 753 754 755 756 757 758 759 760 761
		}
	} 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)) {
762 763
				if (set_nid(page, i, 0, false))
					dn->node_changed = true;
J
Jaegeuk Kim 已提交
764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779
				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);
	}
780
	trace_f2fs_truncate_nodes_exit(dn->inode, freed);
J
Jaegeuk Kim 已提交
781 782 783 784
	return freed;

out_err:
	f2fs_put_page(page, 1);
785
	trace_f2fs_truncate_nodes_exit(dn->inode, ret);
J
Jaegeuk Kim 已提交
786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803
	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 */
804
	for (i = 0; i < idx + 1; i++) {
A
arter97 已提交
805
		/* reference count'll be increased */
806
		pages[i] = get_node_page(F2FS_I_SB(dn->inode), nid[i]);
J
Jaegeuk Kim 已提交
807 808
		if (IS_ERR(pages[i])) {
			err = PTR_ERR(pages[i]);
809
			idx = i - 1;
J
Jaegeuk Kim 已提交
810 811 812 813 814
			goto fail;
		}
		nid[i + 1] = get_nid(pages[i], offset[i + 1], false);
	}

815 816
	ra_node_pages(pages[idx], offset[idx + 1], NIDS_PER_BLOCK);

J
Jaegeuk Kim 已提交
817
	/* free direct nodes linked to a partial indirect node */
818
	for (i = offset[idx + 1]; i < NIDS_PER_BLOCK; i++) {
J
Jaegeuk Kim 已提交
819 820 821 822 823 824 825
		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;
826 827
		if (set_nid(pages[idx], i, 0, false))
			dn->node_changed = true;
J
Jaegeuk Kim 已提交
828 829
	}

830
	if (offset[idx + 1] == 0) {
J
Jaegeuk Kim 已提交
831 832 833 834 835 836 837
		dn->node_page = pages[idx];
		dn->nid = nid[idx];
		truncate_node(dn);
	} else {
		f2fs_put_page(pages[idx], 1);
	}
	offset[idx]++;
838 839
	offset[idx + 1] = 0;
	idx--;
J
Jaegeuk Kim 已提交
840
fail:
841
	for (i = idx; i >= 0; i--)
J
Jaegeuk Kim 已提交
842
		f2fs_put_page(pages[i], 1);
843 844 845

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

J
Jaegeuk Kim 已提交
846 847 848
	return err;
}

J
Jaegeuk Kim 已提交
849
/*
J
Jaegeuk Kim 已提交
850 851 852 853
 * All the block addresses of data and nodes should be nullified.
 */
int truncate_inode_blocks(struct inode *inode, pgoff_t from)
{
854
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
J
Jaegeuk Kim 已提交
855 856
	int err = 0, cont = 1;
	int level, offset[4], noffset[4];
J
Jaegeuk Kim 已提交
857
	unsigned int nofs = 0;
858
	struct f2fs_inode *ri;
J
Jaegeuk Kim 已提交
859 860 861
	struct dnode_of_data dn;
	struct page *page;

862 863
	trace_f2fs_truncate_inode_blocks_enter(inode, from);

864
	level = get_node_path(inode, from, offset, noffset);
865

J
Jaegeuk Kim 已提交
866
	page = get_node_page(sbi, inode->i_ino);
867 868
	if (IS_ERR(page)) {
		trace_f2fs_truncate_inode_blocks_exit(inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
869
		return PTR_ERR(page);
870
	}
J
Jaegeuk Kim 已提交
871 872 873 874

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

875
	ri = F2FS_INODE(page);
J
Jaegeuk Kim 已提交
876 877 878 879 880 881 882 883 884
	switch (level) {
	case 0:
	case 1:
		nofs = noffset[1];
		break;
	case 2:
		nofs = noffset[1];
		if (!offset[level - 1])
			goto skip_partial;
885
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
886 887 888 889 890 891 892 893
		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;
894
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
895 896 897 898 899 900 901 902 903
		if (err < 0 && err != -ENOENT)
			goto fail;
		break;
	default:
		BUG();
	}

skip_partial:
	while (cont) {
904
		dn.nid = le32_to_cpu(ri->i_nid[offset[0] - NODE_DIR1_BLOCK]);
J
Jaegeuk Kim 已提交
905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926
		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 &&
927
				ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) {
J
Jaegeuk Kim 已提交
928
			lock_page(page);
929
			BUG_ON(page->mapping != NODE_MAPPING(sbi));
930
			f2fs_wait_on_page_writeback(page, NODE, true);
931
			ri->i_nid[offset[0] - NODE_DIR1_BLOCK] = 0;
J
Jaegeuk Kim 已提交
932 933 934 935 936 937 938 939 940
			set_page_dirty(page);
			unlock_page(page);
		}
		offset[1] = 0;
		offset[0]++;
		nofs += err;
	}
fail:
	f2fs_put_page(page, 0);
941
	trace_f2fs_truncate_inode_blocks_exit(inode, err);
J
Jaegeuk Kim 已提交
942 943 944
	return err > 0 ? 0 : err;
}

945 946
int truncate_xattr_node(struct inode *inode, struct page *page)
{
947
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
948 949 950 951 952 953 954 955 956 957 958
	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);

959
	f2fs_i_xnid_write(inode, 0);
J
Jaegeuk Kim 已提交
960 961 962 963

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

964 965 966
	set_new_dnode(&dn, inode, page, npage, nid);

	if (page)
967
		dn.inode_page_locked = true;
968 969 970 971
	truncate_node(&dn);
	return 0;
}

972
/*
C
Chao Yu 已提交
973 974
 * Caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
975
 */
C
Chao Yu 已提交
976
int remove_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
977 978
{
	struct dnode_of_data dn;
C
Chao Yu 已提交
979
	int err;
J
Jaegeuk Kim 已提交
980

981
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
C
Chao Yu 已提交
982 983 984
	err = get_dnode_of_data(&dn, 0, LOOKUP_NODE);
	if (err)
		return err;
J
Jaegeuk Kim 已提交
985

C
Chao Yu 已提交
986 987
	err = truncate_xattr_node(inode, dn.inode_page);
	if (err) {
988
		f2fs_put_dnode(&dn);
C
Chao Yu 已提交
989
		return err;
J
Jaegeuk Kim 已提交
990
	}
991 992 993 994 995 996

	/* 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
997
	/* 0 is possible, after f2fs_new_inode() has failed */
998 999
	f2fs_bug_on(F2FS_I_SB(inode),
			inode->i_blocks != 0 && inode->i_blocks != 1);
1000 1001

	/* will put inode & node pages */
1002
	truncate_node(&dn);
C
Chao Yu 已提交
1003
	return 0;
J
Jaegeuk Kim 已提交
1004 1005
}

1006
struct page *new_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
1007 1008 1009 1010 1011
{
	struct dnode_of_data dn;

	/* allocate inode page for new inode */
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
1012 1013

	/* caller should f2fs_put_page(page, 1); */
1014
	return new_node_page(&dn, 0, NULL);
J
Jaegeuk Kim 已提交
1015 1016
}

1017 1018
struct page *new_node_page(struct dnode_of_data *dn,
				unsigned int ofs, struct page *ipage)
J
Jaegeuk Kim 已提交
1019
{
1020
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
1021 1022 1023 1024
	struct node_info old_ni, new_ni;
	struct page *page;
	int err;

1025
	if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC)))
J
Jaegeuk Kim 已提交
1026 1027
		return ERR_PTR(-EPERM);

1028
	page = f2fs_grab_cache_page(NODE_MAPPING(sbi), dn->nid, false);
J
Jaegeuk Kim 已提交
1029 1030 1031
	if (!page)
		return ERR_PTR(-ENOMEM);

1032
	if (unlikely(!inc_valid_node_count(sbi, dn->inode))) {
1033 1034 1035
		err = -ENOSPC;
		goto fail;
	}
J
Jaegeuk Kim 已提交
1036

1037
	get_node_info(sbi, dn->nid, &old_ni);
J
Jaegeuk Kim 已提交
1038 1039

	/* Reinitialize old_ni with new node page */
1040
	f2fs_bug_on(sbi, old_ni.blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
1041 1042
	new_ni = old_ni;
	new_ni.ino = dn->inode->i_ino;
1043
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
1044

1045
	f2fs_wait_on_page_writeback(page, NODE, true);
1046
	fill_node_footer(page, dn->nid, dn->inode->i_ino, ofs, true);
1047
	set_cold_node(dn->inode, page);
1048 1049
	if (!PageUptodate(page))
		SetPageUptodate(page);
1050 1051
	if (set_page_dirty(page))
		dn->node_changed = true;
J
Jaegeuk Kim 已提交
1052

1053
	if (f2fs_has_xattr_block(ofs))
1054
		f2fs_i_xnid_write(dn->inode, dn->nid);
1055

J
Jaegeuk Kim 已提交
1056 1057 1058 1059 1060
	if (ofs == 0)
		inc_valid_inode_count(sbi);
	return page;

fail:
1061
	clear_node_page_dirty(page);
J
Jaegeuk Kim 已提交
1062 1063 1064 1065
	f2fs_put_page(page, 1);
	return ERR_PTR(err);
}

1066 1067 1068
/*
 * Caller should do after getting the following values.
 * 0: f2fs_put_page(page, 0)
1069
 * LOCKED_PAGE or error: f2fs_put_page(page, 1)
1070
 */
M
Mike Christie 已提交
1071
static int read_node_page(struct page *page, int op_flags)
J
Jaegeuk Kim 已提交
1072
{
1073
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1074
	struct node_info ni;
1075
	struct f2fs_io_info fio = {
1076
		.sbi = sbi,
1077
		.type = NODE,
M
Mike Christie 已提交
1078 1079
		.op = REQ_OP_READ,
		.op_flags = op_flags,
1080
		.page = page,
1081
		.encrypted_page = NULL,
1082
	};
J
Jaegeuk Kim 已提交
1083

1084 1085 1086
	if (PageUptodate(page))
		return LOCKED_PAGE;

J
Jaegeuk Kim 已提交
1087 1088
	get_node_info(sbi, page->index, &ni);

1089
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1090
		ClearPageUptodate(page);
J
Jaegeuk Kim 已提交
1091
		return -ENOENT;
1092 1093
	}

1094
	fio.new_blkaddr = fio.old_blkaddr = ni.blk_addr;
1095
	return f2fs_submit_page_bio(&fio);
J
Jaegeuk Kim 已提交
1096 1097
}

J
Jaegeuk Kim 已提交
1098
/*
J
Jaegeuk Kim 已提交
1099 1100 1101 1102 1103
 * Readahead a node page
 */
void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid)
{
	struct page *apage;
1104
	int err;
J
Jaegeuk Kim 已提交
1105

1106 1107 1108 1109
	if (!nid)
		return;
	f2fs_bug_on(sbi, check_nid_range(sbi, nid));

1110 1111 1112 1113
	rcu_read_lock();
	apage = radix_tree_lookup(&NODE_MAPPING(sbi)->page_tree, nid);
	rcu_read_unlock();
	if (apage)
1114
		return;
J
Jaegeuk Kim 已提交
1115

1116
	apage = f2fs_grab_cache_page(NODE_MAPPING(sbi), nid, false);
J
Jaegeuk Kim 已提交
1117 1118 1119
	if (!apage)
		return;

1120
	err = read_node_page(apage, REQ_RAHEAD);
1121
	f2fs_put_page(apage, err ? 1 : 0);
J
Jaegeuk Kim 已提交
1122 1123
}

J
Jaegeuk Kim 已提交
1124
static struct page *__get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid,
1125
					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 = f2fs_grab_cache_page(NODE_MAPPING(sbi), nid, false);
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
	if (parent)
1147
		ra_node_pages(parent, start + 1, MAX_RA_NODE);
1148

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

1151
	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
1152 1153 1154
		f2fs_put_page(page, 1);
		goto repeat;
	}
1155 1156 1157

	if (unlikely(!PageUptodate(page)))
		goto out_err;
1158
page_hit:
1159 1160 1161 1162 1163 1164 1165
	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
1166 1167 1168
	return page;
}

1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
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);
}

1182 1183 1184 1185
static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino)
{
	struct inode *inode;
	struct page *page;
1186
	int ret;
1187 1188 1189 1190 1191 1192

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

1193
	page = pagecache_get_page(inode->i_mapping, 0, FGP_LOCK|FGP_NOWAIT, 0);
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
	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;

1206 1207
	ret = f2fs_write_inline_data(inode, page);
	inode_dec_dirty_pages(inode);
1208
	remove_dirty_inode(inode);
1209
	if (ret)
1210 1211
		set_page_dirty(page);
page_out:
1212
	f2fs_put_page(page, 1);
1213 1214 1215 1216
iput_out:
	iput(inode);
}

1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247
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);
}

1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 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
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;
}

1308
int fsync_node_pages(struct f2fs_sb_info *sbi, struct inode *inode,
1309
			struct writeback_control *wbc, bool atomic)
J
Jaegeuk Kim 已提交
1310 1311 1312
{
	pgoff_t index, end;
	struct pagevec pvec;
1313
	int ret = 0;
1314 1315
	struct page *last_page = NULL;
	bool marked = false;
1316
	nid_t ino = inode->i_ino;
1317
	int nwritten = 0;
1318

1319 1320 1321 1322 1323 1324
	if (atomic) {
		last_page = last_fsync_dnode(sbi, ino);
		if (IS_ERR_OR_NULL(last_page))
			return PTR_ERR_OR_ZERO(last_page);
	}
retry:
1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
	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))) {
1341
				f2fs_put_page(last_page, 0);
1342
				pagevec_release(&pvec);
1343 1344
				ret = -EIO;
				goto out;
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361
			}

			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;

1362
			if (!PageDirty(page) && page != last_page) {
1363 1364 1365 1366 1367 1368 1369
				/* someone wrote it for us */
				goto continue_unlock;
			}

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

1370 1371
			if (!atomic || page == last_page) {
				set_fsync_mark(page, 1);
1372 1373 1374 1375
				if (IS_INODE(page)) {
					if (is_inode_flag_set(inode,
								FI_DIRTY_INODE))
						update_inode(inode, page);
1376
					set_dentry_mark(page,
1377
						need_dentry_mark(sbi, ino));
1378
				}
1379 1380 1381 1382 1383 1384 1385
				/*  may be written by other thread */
				if (!PageDirty(page))
					set_page_dirty(page);
			}

			if (!clear_page_dirty_for_io(page))
				goto continue_unlock;
1386

1387 1388
			ret = NODE_MAPPING(sbi)->a_ops->writepage(page, wbc);
			if (ret) {
1389
				unlock_page(page);
1390 1391
				f2fs_put_page(last_page, 0);
				break;
1392 1393
			} else {
				nwritten++;
1394
			}
1395

1396 1397 1398
			if (page == last_page) {
				f2fs_put_page(page, 0);
				marked = true;
1399
				break;
1400
			}
1401 1402 1403 1404
		}
		pagevec_release(&pvec);
		cond_resched();

1405
		if (ret || marked)
1406 1407
			break;
	}
1408 1409 1410 1411 1412 1413 1414 1415 1416
	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);
		set_page_dirty(last_page);
		unlock_page(last_page);
		goto retry;
	}
1417
out:
1418 1419
	if (nwritten)
		f2fs_submit_merged_bio_cond(sbi, NULL, NULL, ino, NODE, WRITE);
1420
	return ret ? -EIO: 0;
1421 1422 1423 1424 1425 1426 1427
}

int sync_node_pages(struct f2fs_sb_info *sbi, struct writeback_control *wbc)
{
	pgoff_t index, end;
	struct pagevec pvec;
	int step = 0;
1428
	int nwritten = 0;
1429
	int ret = 0;
J
Jaegeuk Kim 已提交
1430 1431 1432 1433 1434

	pagevec_init(&pvec, 0);

next_step:
	index = 0;
1435
	end = ULONG_MAX;
J
Jaegeuk Kim 已提交
1436 1437 1438

	while (index <= end) {
		int i, nr_pages;
1439
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
J
Jaegeuk Kim 已提交
1440 1441 1442 1443 1444 1445 1446 1447
				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
1448 1449
			if (unlikely(f2fs_cp_error(sbi))) {
				pagevec_release(&pvec);
1450 1451
				ret = -EIO;
				goto out;
C
Chao Yu 已提交
1452 1453
			}

J
Jaegeuk Kim 已提交
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
			/*
			 * 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;
1468
lock_node:
1469
			if (!trylock_page(page))
J
Jaegeuk Kim 已提交
1470 1471
				continue;

1472
			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
J
Jaegeuk Kim 已提交
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482
continue_unlock:
				unlock_page(page);
				continue;
			}

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

1483
			/* flush inline_data */
1484
			if (is_inline_node(page)) {
1485 1486 1487
				clear_inline_node(page);
				unlock_page(page);
				flush_inline_data(sbi, ino_of_node(page));
1488
				goto lock_node;
1489 1490
			}

1491 1492 1493
			f2fs_wait_on_page_writeback(page, NODE, true);

			BUG_ON(PageWriteback(page));
J
Jaegeuk Kim 已提交
1494 1495 1496
			if (!clear_page_dirty_for_io(page))
				goto continue_unlock;

1497 1498
			set_fsync_mark(page, 0);
			set_dentry_mark(page, 0);
1499 1500 1501

			if (NODE_MAPPING(sbi)->a_ops->writepage(page, wbc))
				unlock_page(page);
1502 1503
			else
				nwritten++;
J
Jaegeuk Kim 已提交
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520

			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;
	}
1521 1522 1523 1524
out:
	if (nwritten)
		f2fs_submit_merged_bio(sbi, NODE, WRITE);
	return ret;
J
Jaegeuk Kim 已提交
1525 1526
}

1527 1528
int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
{
1529
	pgoff_t index = 0, end = ULONG_MAX;
1530
	struct pagevec pvec;
1531
	int ret2, ret = 0;
1532 1533

	pagevec_init(&pvec, 0);
1534 1535 1536 1537 1538 1539 1540 1541

	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;
1542 1543 1544 1545 1546

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

			/* until radix tree lookup accepts end_index */
1547
			if (unlikely(page->index > end))
1548 1549
				continue;

1550
			if (ino && ino_of_node(page) == ino) {
1551
				f2fs_wait_on_page_writeback(page, NODE, true);
1552 1553 1554
				if (TestClearPageError(page))
					ret = -EIO;
			}
1555 1556 1557 1558 1559
		}
		pagevec_release(&pvec);
		cond_resched();
	}

1560
	ret2 = filemap_check_errors(NODE_MAPPING(sbi));
1561 1562 1563 1564 1565
	if (!ret)
		ret = ret2;
	return ret;
}

J
Jaegeuk Kim 已提交
1566 1567 1568
static int f2fs_write_node_page(struct page *page,
				struct writeback_control *wbc)
{
1569
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1570 1571
	nid_t nid;
	struct node_info ni;
1572
	struct f2fs_io_info fio = {
1573
		.sbi = sbi,
1574
		.type = NODE,
M
Mike Christie 已提交
1575 1576
		.op = REQ_OP_WRITE,
		.op_flags = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : 0,
1577
		.page = page,
1578
		.encrypted_page = NULL,
1579
	};
J
Jaegeuk Kim 已提交
1580

1581 1582
	trace_f2fs_writepage(page, NODE);

1583
	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
1584
		goto redirty_out;
1585 1586
	if (unlikely(f2fs_cp_error(sbi)))
		goto redirty_out;
1587

J
Jaegeuk Kim 已提交
1588 1589
	/* get old block addr of this node page */
	nid = nid_of_node(page);
1590
	f2fs_bug_on(sbi, page->index != nid);
J
Jaegeuk Kim 已提交
1591

1592 1593 1594 1595 1596 1597 1598
	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
1599 1600 1601
	get_node_info(sbi, nid, &ni);

	/* This page is already truncated */
1602
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1603
		ClearPageUptodate(page);
1604
		dec_page_count(sbi, F2FS_DIRTY_NODES);
1605
		up_read(&sbi->node_write);
1606 1607 1608
		unlock_page(page);
		return 0;
	}
J
Jaegeuk Kim 已提交
1609 1610

	set_page_writeback(page);
1611
	fio.old_blkaddr = ni.blk_addr;
1612
	write_node_page(nid, &fio);
1613
	set_node_addr(sbi, &ni, fio.new_blkaddr, is_fsync_dnode(page));
J
Jaegeuk Kim 已提交
1614
	dec_page_count(sbi, F2FS_DIRTY_NODES);
1615
	up_read(&sbi->node_write);
1616 1617 1618 1619

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

J
Jaegeuk Kim 已提交
1620
	unlock_page(page);
1621

1622
	if (unlikely(f2fs_cp_error(sbi)))
1623 1624
		f2fs_submit_merged_bio(sbi, NODE, WRITE);

J
Jaegeuk Kim 已提交
1625
	return 0;
1626 1627

redirty_out:
1628
	redirty_page_for_writepage(wbc, page);
1629
	return AOP_WRITEPAGE_ACTIVATE;
J
Jaegeuk Kim 已提交
1630 1631 1632 1633 1634
}

static int f2fs_write_node_pages(struct address_space *mapping,
			    struct writeback_control *wbc)
{
1635
	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
1636
	struct blk_plug plug;
1637
	long diff;
J
Jaegeuk Kim 已提交
1638

1639 1640
	/* balancing f2fs's metadata in background */
	f2fs_balance_fs_bg(sbi);
J
Jaegeuk Kim 已提交
1641

1642
	/* collect a number of dirty node pages and write together */
1643
	if (get_pages(sbi, F2FS_DIRTY_NODES) < nr_pages_to_skip(sbi, NODE))
1644
		goto skip_write;
1645

Y
Yunlei He 已提交
1646 1647
	trace_f2fs_writepages(mapping->host, wbc, NODE);

1648
	diff = nr_pages_to_write(sbi, NODE, wbc);
1649
	wbc->sync_mode = WB_SYNC_NONE;
1650
	blk_start_plug(&plug);
1651
	sync_node_pages(sbi, wbc);
1652
	blk_finish_plug(&plug);
1653
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
J
Jaegeuk Kim 已提交
1654
	return 0;
1655 1656 1657

skip_write:
	wbc->pages_skipped += get_pages(sbi, F2FS_DIRTY_NODES);
Y
Yunlei He 已提交
1658
	trace_f2fs_writepages(mapping->host, wbc, NODE);
1659
	return 0;
J
Jaegeuk Kim 已提交
1660 1661 1662 1663
}

static int f2fs_set_node_page_dirty(struct page *page)
{
1664 1665
	trace_f2fs_set_page_dirty(page, NODE);

1666 1667
	if (!PageUptodate(page))
		SetPageUptodate(page);
J
Jaegeuk Kim 已提交
1668
	if (!PageDirty(page)) {
1669
		f2fs_set_page_dirty_nobuffers(page);
1670
		inc_page_count(F2FS_P_SB(page), F2FS_DIRTY_NODES);
J
Jaegeuk Kim 已提交
1671
		SetPagePrivate(page);
J
Jaegeuk Kim 已提交
1672
		f2fs_trace_pid(page);
J
Jaegeuk Kim 已提交
1673 1674 1675 1676 1677
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
1678
/*
J
Jaegeuk Kim 已提交
1679 1680 1681 1682 1683 1684
 * 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,
1685 1686
	.invalidatepage	= f2fs_invalidate_page,
	.releasepage	= f2fs_release_page,
1687 1688 1689
#ifdef CONFIG_MIGRATION
	.migratepage    = f2fs_migrate_page,
#endif
J
Jaegeuk Kim 已提交
1690 1691
};

1692 1693
static struct free_nid *__lookup_free_nid_list(struct f2fs_nm_info *nm_i,
						nid_t n)
J
Jaegeuk Kim 已提交
1694
{
1695
	return radix_tree_lookup(&nm_i->free_nid_root, n);
J
Jaegeuk Kim 已提交
1696 1697
}

1698 1699
static int __insert_nid_to_list(struct f2fs_sb_info *sbi,
			struct free_nid *i, enum nid_list list, bool new)
C
Chao Yu 已提交
1700 1701 1702
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);

1703 1704 1705 1706 1707 1708
	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
1709 1710 1711 1712
	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]);
1713
	return 0;
C
Chao Yu 已提交
1714 1715 1716
}

static void __remove_nid_from_list(struct f2fs_sb_info *sbi,
1717
			struct free_nid *i, enum nid_list list, bool reuse)
C
Chao Yu 已提交
1718 1719 1720 1721 1722 1723 1724
{
	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]--;
	list_del(&i->list);
1725 1726
	if (!reuse)
		radix_tree_delete(&nm_i->free_nid_root, i->nid);
C
Chao Yu 已提交
1727 1728
}

1729
static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build)
J
Jaegeuk Kim 已提交
1730
{
1731
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1732
	struct free_nid *i;
1733
	struct nat_entry *ne;
1734
	int err;
J
Jaegeuk Kim 已提交
1735

1736
	/* 0 nid should not be used */
1737
	if (unlikely(nid == 0))
1738
		return 0;
1739

1740 1741 1742
	if (build) {
		/* do not add allocated nids */
		ne = __lookup_nat_cache(nm_i, nid);
1743
		if (ne && (!get_nat_flag(ne, IS_CHECKPOINTED) ||
1744
				nat_get_blkaddr(ne) != NULL_ADDR))
1745
			return 0;
J
Jaegeuk Kim 已提交
1746
	}
1747 1748

	i = f2fs_kmem_cache_alloc(free_nid_slab, GFP_NOFS);
J
Jaegeuk Kim 已提交
1749 1750 1751
	i->nid = nid;
	i->state = NID_NEW;

1752 1753 1754 1755 1756
	if (radix_tree_preload(GFP_NOFS)) {
		kmem_cache_free(free_nid_slab, i);
		return 0;
	}

C
Chao Yu 已提交
1757
	spin_lock(&nm_i->nid_list_lock);
1758 1759 1760 1761
	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
1762 1763 1764 1765 1766 1767
		kmem_cache_free(free_nid_slab, i);
		return 0;
	}
	return 1;
}

C
Chao Yu 已提交
1768
static void remove_free_nid(struct f2fs_sb_info *sbi, nid_t nid)
J
Jaegeuk Kim 已提交
1769
{
C
Chao Yu 已提交
1770
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1771
	struct free_nid *i;
1772 1773
	bool need_free = false;

C
Chao Yu 已提交
1774
	spin_lock(&nm_i->nid_list_lock);
1775
	i = __lookup_free_nid_list(nm_i, nid);
J
Jaegeuk Kim 已提交
1776
	if (i && i->state == NID_NEW) {
1777
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, false);
1778
		need_free = true;
J
Jaegeuk Kim 已提交
1779
	}
C
Chao Yu 已提交
1780
	spin_unlock(&nm_i->nid_list_lock);
1781 1782 1783

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1784 1785
}

1786
static void scan_nat_page(struct f2fs_sb_info *sbi,
J
Jaegeuk Kim 已提交
1787 1788
			struct page *nat_page, nid_t start_nid)
{
1789
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1790 1791 1792 1793 1794 1795 1796
	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
1797

1798
		if (unlikely(start_nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1799
			break;
H
Haicheng Li 已提交
1800 1801

		blk_addr = le32_to_cpu(nat_blk->entries[i].block_addr);
1802
		f2fs_bug_on(sbi, blk_addr == NEW_ADDR);
1803 1804
		if (blk_addr == NULL_ADDR)
			add_free_nid(sbi, start_nid, true);
J
Jaegeuk Kim 已提交
1805 1806 1807
	}
}

1808
static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync)
J
Jaegeuk Kim 已提交
1809 1810 1811
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
1812
	struct f2fs_journal *journal = curseg->journal;
1813
	int i = 0;
1814
	nid_t nid = nm_i->next_scan_nid;
J
Jaegeuk Kim 已提交
1815

1816
	/* Enough entries */
C
Chao Yu 已提交
1817
	if (nm_i->nid_cnt[FREE_NID_LIST] >= NAT_ENTRY_PER_BLOCK)
1818
		return;
J
Jaegeuk Kim 已提交
1819

1820 1821 1822
	if (!sync && !available_free_memory(sbi, FREE_NIDS))
		return;

1823
	/* readahead nat pages to be scanned */
1824 1825
	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nid), FREE_NID_PAGES,
							META_NAT, true);
J
Jaegeuk Kim 已提交
1826

1827
	down_read(&nm_i->nat_tree_lock);
1828

J
Jaegeuk Kim 已提交
1829 1830 1831
	while (1) {
		struct page *page = get_current_nat_page(sbi, nid);

1832
		scan_nat_page(sbi, page, nid);
J
Jaegeuk Kim 已提交
1833 1834 1835
		f2fs_put_page(page, 1);

		nid += (NAT_ENTRY_PER_BLOCK - (nid % NAT_ENTRY_PER_BLOCK));
1836
		if (unlikely(nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1837
			nid = 0;
1838

1839
		if (++i >= FREE_NID_PAGES)
J
Jaegeuk Kim 已提交
1840 1841 1842
			break;
	}

1843 1844
	/* go to the next free nat pages to find free nids abundantly */
	nm_i->next_scan_nid = nid;
J
Jaegeuk Kim 已提交
1845 1846

	/* find free nids from current sum_pages */
1847
	down_read(&curseg->journal_rwsem);
1848 1849 1850 1851 1852
	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
1853
		if (addr == NULL_ADDR)
1854
			add_free_nid(sbi, nid, true);
J
Jaegeuk Kim 已提交
1855
		else
C
Chao Yu 已提交
1856
			remove_free_nid(sbi, nid);
J
Jaegeuk Kim 已提交
1857
	}
1858
	up_read(&curseg->journal_rwsem);
1859
	up_read(&nm_i->nat_tree_lock);
C
Chao Yu 已提交
1860 1861

	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nm_i->next_scan_nid),
C
Chao Yu 已提交
1862
					nm_i->ra_nid_pages, META_NAT, false);
J
Jaegeuk Kim 已提交
1863 1864
}

1865
void build_free_nids(struct f2fs_sb_info *sbi, bool sync)
1866 1867
{
	mutex_lock(&NM_I(sbi)->build_lock);
1868
	__build_free_nids(sbi, sync);
1869 1870 1871
	mutex_unlock(&NM_I(sbi)->build_lock);
}

J
Jaegeuk Kim 已提交
1872 1873 1874 1875 1876 1877 1878 1879 1880 1881
/*
 * 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
1882
#ifdef CONFIG_F2FS_FAULT_INJECTION
1883
	if (time_to_inject(sbi, FAULT_ALLOC_NID))
J
Jaegeuk Kim 已提交
1884 1885
		return false;
#endif
1886
	if (unlikely(sbi->total_valid_node_count + 1 > nm_i->available_nids))
1887
		return false;
J
Jaegeuk Kim 已提交
1888

C
Chao Yu 已提交
1889
	spin_lock(&nm_i->nid_list_lock);
J
Jaegeuk Kim 已提交
1890

1891
	/* We should not use stale free nids created by build_free_nids */
C
Chao Yu 已提交
1892 1893 1894 1895
	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);
1896
		*nid = i->nid;
C
Chao Yu 已提交
1897

1898
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, true);
1899
		i->state = NID_ALLOC;
1900
		__insert_nid_to_list(sbi, i, ALLOC_NID_LIST, false);
C
Chao Yu 已提交
1901
		spin_unlock(&nm_i->nid_list_lock);
1902 1903
		return true;
	}
C
Chao Yu 已提交
1904
	spin_unlock(&nm_i->nid_list_lock);
1905 1906

	/* Let's scan nat pages and its caches to get free nids */
1907
	build_free_nids(sbi, true);
1908
	goto retry;
J
Jaegeuk Kim 已提交
1909 1910
}

J
Jaegeuk Kim 已提交
1911
/*
J
Jaegeuk Kim 已提交
1912 1913 1914 1915 1916 1917 1918
 * 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
1919
	spin_lock(&nm_i->nid_list_lock);
1920
	i = __lookup_free_nid_list(nm_i, nid);
C
Chao Yu 已提交
1921
	f2fs_bug_on(sbi, !i);
1922
	__remove_nid_from_list(sbi, i, ALLOC_NID_LIST, false);
C
Chao Yu 已提交
1923
	spin_unlock(&nm_i->nid_list_lock);
1924 1925

	kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1926 1927
}

J
Jaegeuk Kim 已提交
1928
/*
J
Jaegeuk Kim 已提交
1929 1930 1931 1932
 * alloc_nid() should be called prior to this function.
 */
void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid)
{
1933 1934
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i;
1935
	bool need_free = false;
1936

J
Jaegeuk Kim 已提交
1937 1938 1939
	if (!nid)
		return;

C
Chao Yu 已提交
1940
	spin_lock(&nm_i->nid_list_lock);
1941
	i = __lookup_free_nid_list(nm_i, nid);
C
Chao Yu 已提交
1942 1943
	f2fs_bug_on(sbi, !i);

1944
	if (!available_free_memory(sbi, FREE_NIDS)) {
1945
		__remove_nid_from_list(sbi, i, ALLOC_NID_LIST, false);
1946
		need_free = true;
1947
	} else {
1948
		__remove_nid_from_list(sbi, i, ALLOC_NID_LIST, true);
1949
		i->state = NID_NEW;
1950
		__insert_nid_to_list(sbi, i, FREE_NID_LIST, false);
1951
	}
C
Chao Yu 已提交
1952
	spin_unlock(&nm_i->nid_list_lock);
1953 1954 1955

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1956 1957
}

C
Chao Yu 已提交
1958 1959 1960 1961 1962 1963
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
1964
	if (nm_i->nid_cnt[FREE_NID_LIST] <= MAX_FREE_NIDS)
1965 1966
		return 0;

C
Chao Yu 已提交
1967 1968 1969
	if (!mutex_trylock(&nm_i->build_lock))
		return 0;

C
Chao Yu 已提交
1970 1971 1972 1973 1974
	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
1975
			break;
C
Chao Yu 已提交
1976

1977
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, false);
C
Chao Yu 已提交
1978 1979 1980
		kmem_cache_free(free_nid_slab, i);
		nr_shrink--;
	}
C
Chao Yu 已提交
1981
	spin_unlock(&nm_i->nid_list_lock);
C
Chao Yu 已提交
1982 1983 1984 1985 1986
	mutex_unlock(&nm_i->build_lock);

	return nr - nr_shrink;
}

1987
void recover_inline_xattr(struct inode *inode, struct page *page)
1988 1989 1990 1991 1992 1993
{
	void *src_addr, *dst_addr;
	size_t inline_size;
	struct page *ipage;
	struct f2fs_inode *ri;

1994
	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
1995
	f2fs_bug_on(F2FS_I_SB(inode), IS_ERR(ipage));
1996

1997 1998
	ri = F2FS_INODE(page);
	if (!(ri->i_inline & F2FS_INLINE_XATTR)) {
1999
		clear_inode_flag(inode, FI_INLINE_XATTR);
2000 2001 2002
		goto update_inode;
	}

2003 2004 2005 2006
	dst_addr = inline_xattr_addr(ipage);
	src_addr = inline_xattr_addr(page);
	inline_size = inline_xattr_size(inode);

2007
	f2fs_wait_on_page_writeback(ipage, NODE, true);
2008
	memcpy(dst_addr, src_addr, inline_size);
2009
update_inode:
2010 2011 2012 2013
	update_inode(inode, ipage);
	f2fs_put_page(ipage, 1);
}

2014
void recover_xattr_data(struct inode *inode, struct page *page, block_t blkaddr)
2015
{
2016
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
2017 2018 2019 2020 2021 2022 2023 2024 2025 2026
	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);
2027
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
2028 2029
	invalidate_blocks(sbi, ni.blk_addr);
	dec_valid_node_count(sbi, inode);
2030
	set_node_addr(sbi, &ni, NULL_ADDR, false);
2031 2032 2033 2034

recover_xnid:
	/* 2: allocate new xattr nid */
	if (unlikely(!inc_valid_node_count(sbi, inode)))
2035
		f2fs_bug_on(sbi, 1);
2036

C
Chao Yu 已提交
2037
	remove_free_nid(sbi, new_xnid);
2038 2039
	get_node_info(sbi, new_xnid, &ni);
	ni.ino = inode->i_ino;
2040
	set_node_addr(sbi, &ni, NEW_ADDR, false);
2041
	f2fs_i_xnid_write(inode, new_xnid);
2042 2043 2044

	/* 3: update xattr blkaddr */
	refresh_sit_entry(sbi, NEW_ADDR, blkaddr);
2045
	set_node_addr(sbi, &ni, blkaddr, false);
2046 2047
}

J
Jaegeuk Kim 已提交
2048 2049
int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
{
2050
	struct f2fs_inode *src, *dst;
J
Jaegeuk Kim 已提交
2051 2052 2053 2054
	nid_t ino = ino_of_node(page);
	struct node_info old_ni, new_ni;
	struct page *ipage;

2055 2056 2057 2058
	get_node_info(sbi, ino, &old_ni);

	if (unlikely(old_ni.blk_addr != NULL_ADDR))
		return -EINVAL;
2059
retry:
2060
	ipage = f2fs_grab_cache_page(NODE_MAPPING(sbi), ino, false);
2061 2062 2063 2064
	if (!ipage) {
		congestion_wait(BLK_RW_ASYNC, HZ/50);
		goto retry;
	}
J
Jaegeuk Kim 已提交
2065

A
arter97 已提交
2066
	/* Should not use this inode from free nid list */
C
Chao Yu 已提交
2067
	remove_free_nid(sbi, ino);
J
Jaegeuk Kim 已提交
2068

2069 2070
	if (!PageUptodate(ipage))
		SetPageUptodate(ipage);
J
Jaegeuk Kim 已提交
2071 2072
	fill_node_footer(ipage, ino, ino, 0, true);

2073 2074
	src = F2FS_INODE(page);
	dst = F2FS_INODE(ipage);
J
Jaegeuk Kim 已提交
2075

2076 2077 2078 2079 2080
	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;
2081
	dst->i_inline = src->i_inline & F2FS_INLINE_XATTR;
J
Jaegeuk Kim 已提交
2082 2083 2084 2085

	new_ni = old_ni;
	new_ni.ino = ino;

2086
	if (unlikely(!inc_valid_node_count(sbi, NULL)))
2087
		WARN_ON(1);
2088
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
J
Jaegeuk Kim 已提交
2089
	inc_valid_inode_count(sbi);
2090
	set_page_dirty(ipage);
J
Jaegeuk Kim 已提交
2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
	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
2101
	int bio_blocks = MAX_BIO_BLOCKS(sbi);
2102
	int i, idx, last_offset, nrpages;
J
Jaegeuk Kim 已提交
2103 2104 2105 2106 2107 2108

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

2109
	for (i = 0; i < last_offset; i += nrpages, addr += nrpages) {
2110
		nrpages = min(last_offset - i, bio_blocks);
2111

A
arter97 已提交
2112
		/* readahead node pages */
2113
		ra_meta_pages(sbi, addr, nrpages, META_POR, true);
J
Jaegeuk Kim 已提交
2114

2115
		for (idx = addr; idx < addr + nrpages; idx++) {
2116
			struct page *page = get_tmp_page(sbi, idx);
2117

2118 2119 2120 2121 2122 2123
			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);
2124
		}
2125

2126
		invalidate_mapping_pages(META_MAPPING(sbi), addr,
2127
							addr + nrpages);
J
Jaegeuk Kim 已提交
2128
	}
2129
	return 0;
J
Jaegeuk Kim 已提交
2130 2131
}

2132
static void remove_nats_in_journal(struct f2fs_sb_info *sbi)
J
Jaegeuk Kim 已提交
2133 2134 2135
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
2136
	struct f2fs_journal *journal = curseg->journal;
J
Jaegeuk Kim 已提交
2137 2138
	int i;

2139
	down_write(&curseg->journal_rwsem);
2140
	for (i = 0; i < nats_in_cursum(journal); i++) {
J
Jaegeuk Kim 已提交
2141 2142
		struct nat_entry *ne;
		struct f2fs_nat_entry raw_ne;
2143
		nid_t nid = le32_to_cpu(nid_in_journal(journal, i));
J
Jaegeuk Kim 已提交
2144

2145
		raw_ne = nat_in_journal(journal, i);
2146

J
Jaegeuk Kim 已提交
2147 2148
		ne = __lookup_nat_cache(nm_i, nid);
		if (!ne) {
2149 2150
			ne = grab_nat_entry(nm_i, nid);
			node_info_from_raw_nat(&ne->ni, &raw_ne);
J
Jaegeuk Kim 已提交
2151 2152 2153
		}
		__set_nat_cache_dirty(nm_i, ne);
	}
2154
	update_nats_in_cursum(journal, -i);
2155
	up_write(&curseg->journal_rwsem);
J
Jaegeuk Kim 已提交
2156 2157
}

2158 2159
static void __adjust_nat_entry_set(struct nat_entry_set *nes,
						struct list_head *head, int max)
J
Jaegeuk Kim 已提交
2160
{
2161
	struct nat_entry_set *cur;
J
Jaegeuk Kim 已提交
2162

2163 2164
	if (nes->entry_cnt >= max)
		goto add_out;
J
Jaegeuk Kim 已提交
2165

2166 2167 2168 2169 2170
	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;
		}
2171
	}
2172 2173 2174
add_out:
	list_add_tail(&nes->set_list, head);
}
J
Jaegeuk Kim 已提交
2175

2176 2177 2178 2179
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);
2180
	struct f2fs_journal *journal = curseg->journal;
2181 2182 2183 2184 2185
	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
2186

2187 2188 2189 2190 2191
	/*
	 * 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.
	 */
2192
	if (!__has_cursum_space(journal, set->entry_cnt, NAT_JOURNAL))
2193 2194 2195
		to_journal = false;

	if (to_journal) {
2196
		down_write(&curseg->journal_rwsem);
2197 2198 2199 2200 2201
	} else {
		page = get_next_nat_page(sbi, start_nid);
		nat_blk = page_address(page);
		f2fs_bug_on(sbi, !nat_blk);
	}
2202

2203 2204 2205 2206 2207 2208 2209 2210
	/* 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;
2211 2212

		if (to_journal) {
2213
			offset = lookup_journal_in_cursum(journal,
2214 2215
							NAT_JOURNAL, nid, 1);
			f2fs_bug_on(sbi, offset < 0);
2216 2217
			raw_ne = &nat_in_journal(journal, offset);
			nid_in_journal(journal, offset) = cpu_to_le32(nid);
2218
		} else {
2219
			raw_ne = &nat_blk->entries[nid - start_nid];
J
Jaegeuk Kim 已提交
2220
		}
2221 2222 2223 2224 2225 2226
		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
2227

2228
	if (to_journal)
2229
		up_write(&curseg->journal_rwsem);
2230 2231
	else
		f2fs_put_page(page, 1);
2232

2233 2234 2235 2236
	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);
2237
}
2238

2239 2240 2241 2242 2243 2244 2245
/*
 * 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);
2246
	struct f2fs_journal *journal = curseg->journal;
2247
	struct nat_entry_set *setvec[SETVEC_SIZE];
2248 2249 2250 2251 2252
	struct nat_entry_set *set, *tmp;
	unsigned int found;
	nid_t set_idx = 0;
	LIST_HEAD(sets);

2253 2254
	if (!nm_i->dirty_nat_cnt)
		return;
2255

2256
	down_write(&nm_i->nat_tree_lock);
2257

2258 2259 2260 2261 2262
	/*
	 * 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.
	 */
2263
	if (!__has_cursum_space(journal, nm_i->dirty_nat_cnt, NAT_JOURNAL))
2264 2265 2266
		remove_nats_in_journal(sbi);

	while ((found = __gang_lookup_nat_set(nm_i,
2267
					set_idx, SETVEC_SIZE, setvec))) {
2268 2269 2270 2271
		unsigned idx;
		set_idx = setvec[found - 1]->set + 1;
		for (idx = 0; idx < found; idx++)
			__adjust_nat_entry_set(setvec[idx], &sets,
2272
						MAX_NAT_JENTRIES(journal));
J
Jaegeuk Kim 已提交
2273
	}
2274

2275 2276 2277 2278
	/* flush dirty nats in nat entry set */
	list_for_each_entry_safe(set, tmp, &sets, set_list)
		__flush_nat_entry_set(sbi, set);

2279
	up_write(&nm_i->nat_tree_lock);
2280

2281
	f2fs_bug_on(sbi, nm_i->dirty_nat_cnt);
J
Jaegeuk Kim 已提交
2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
}

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);
2296

2297 2298
	nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nat_blocks;

2299
	/* not used nids: 0, node, meta, (and root counted as valid node) */
2300
	nm_i->available_nids = nm_i->max_nid - F2FS_RESERVED_NODE_NUM;
C
Chao Yu 已提交
2301 2302
	nm_i->nid_cnt[FREE_NID_LIST] = 0;
	nm_i->nid_cnt[ALLOC_NID_LIST] = 0;
J
Jaegeuk Kim 已提交
2303
	nm_i->nat_cnt = 0;
2304
	nm_i->ram_thresh = DEF_RAM_THRESHOLD;
C
Chao Yu 已提交
2305
	nm_i->ra_nid_pages = DEF_RA_NID_PAGES;
C
Chao Yu 已提交
2306
	nm_i->dirty_nats_ratio = DEF_DIRTY_NAT_RATIO_THRESHOLD;
J
Jaegeuk Kim 已提交
2307

2308
	INIT_RADIX_TREE(&nm_i->free_nid_root, GFP_ATOMIC);
C
Chao Yu 已提交
2309 2310
	INIT_LIST_HEAD(&nm_i->nid_list[FREE_NID_LIST]);
	INIT_LIST_HEAD(&nm_i->nid_list[ALLOC_NID_LIST]);
2311 2312
	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
2313 2314 2315
	INIT_LIST_HEAD(&nm_i->nat_entries);

	mutex_init(&nm_i->build_lock);
C
Chao Yu 已提交
2316
	spin_lock_init(&nm_i->nid_list_lock);
2317
	init_rwsem(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2318 2319

	nm_i->next_scan_nid = le32_to_cpu(sbi->ckpt->next_free_nid);
A
Alexandru Gheorghiu 已提交
2320
	nm_i->bitmap_size = __bitmap_size(sbi, NAT_BITMAP);
J
Jaegeuk Kim 已提交
2321 2322 2323 2324
	version_bitmap = __bitmap_ptr(sbi, NAT_BITMAP);
	if (!version_bitmap)
		return -EFAULT;

A
Alexandru Gheorghiu 已提交
2325 2326 2327 2328
	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
2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343
	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;

2344
	build_free_nids(sbi, true);
J
Jaegeuk Kim 已提交
2345 2346 2347 2348 2349 2350 2351 2352
	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];
2353
	struct nat_entry_set *setvec[SETVEC_SIZE];
J
Jaegeuk Kim 已提交
2354 2355 2356 2357 2358 2359 2360
	nid_t nid = 0;
	unsigned int found;

	if (!nm_i)
		return;

	/* destroy free nid list */
C
Chao Yu 已提交
2361 2362 2363
	spin_lock(&nm_i->nid_list_lock);
	list_for_each_entry_safe(i, next_i, &nm_i->nid_list[FREE_NID_LIST],
									list) {
2364
		__remove_nid_from_list(sbi, i, FREE_NID_LIST, false);
C
Chao Yu 已提交
2365
		spin_unlock(&nm_i->nid_list_lock);
2366
		kmem_cache_free(free_nid_slab, i);
C
Chao Yu 已提交
2367
		spin_lock(&nm_i->nid_list_lock);
J
Jaegeuk Kim 已提交
2368
	}
C
Chao Yu 已提交
2369 2370 2371 2372
	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
2373 2374

	/* destroy nat cache */
2375
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2376 2377 2378
	while ((found = __gang_lookup_nat_cache(nm_i,
					nid, NATVEC_SIZE, natvec))) {
		unsigned idx;
2379

2380 2381 2382
		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
2383
	}
2384
	f2fs_bug_on(sbi, nm_i->nat_cnt);
2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399

	/* 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]);
		}
	}
2400
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2401 2402 2403 2404 2405 2406

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

2407
int __init create_node_manager_caches(void)
J
Jaegeuk Kim 已提交
2408 2409
{
	nat_entry_slab = f2fs_kmem_cache_create("nat_entry",
2410
			sizeof(struct nat_entry));
J
Jaegeuk Kim 已提交
2411
	if (!nat_entry_slab)
2412
		goto fail;
J
Jaegeuk Kim 已提交
2413 2414

	free_nid_slab = f2fs_kmem_cache_create("free_nid",
2415
			sizeof(struct free_nid));
2416
	if (!free_nid_slab)
2417
		goto destroy_nat_entry;
2418 2419 2420 2421

	nat_entry_set_slab = f2fs_kmem_cache_create("nat_entry_set",
			sizeof(struct nat_entry_set));
	if (!nat_entry_set_slab)
2422
		goto destroy_free_nid;
J
Jaegeuk Kim 已提交
2423
	return 0;
2424

2425
destroy_free_nid:
2426
	kmem_cache_destroy(free_nid_slab);
2427
destroy_nat_entry:
2428 2429 2430
	kmem_cache_destroy(nat_entry_slab);
fail:
	return -ENOMEM;
J
Jaegeuk Kim 已提交
2431 2432 2433 2434
}

void destroy_node_manager_caches(void)
{
2435
	kmem_cache_destroy(nat_entry_set_slab);
J
Jaegeuk Kim 已提交
2436 2437 2438
	kmem_cache_destroy(free_nid_slab);
	kmem_cache_destroy(nat_entry_slab);
}