node.c 51.5 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
		if (sbi->sb->s_bdi->wb.dirty_exceeded)
75
			return false;
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
255 256 257 258 259 260 261 262 263
	list_add_tail(&new->list, &nm_i->nat_entries);
	nm_i->nat_cnt++;
	return new;
}

static void cache_nat_entry(struct f2fs_nm_info *nm_i, nid_t nid,
						struct f2fs_nat_entry *ne)
{
	struct nat_entry *e;
264

265
	down_write(&nm_i->nat_tree_lock);
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);
J
Jaegeuk Kim 已提交
270
	}
271
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
272 273 274
}

static void set_node_addr(struct f2fs_sb_info *sbi, struct node_info *ni,
275
			block_t new_blkaddr, bool fsync_done)
J
Jaegeuk Kim 已提交
276 277 278
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct nat_entry *e;
279

280
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
281 282 283
	e = __lookup_nat_cache(nm_i, ni->nid);
	if (!e) {
		e = grab_nat_entry(nm_i, ni->nid);
284
		copy_node_info(&e->ni, ni);
285
		f2fs_bug_on(sbi, ni->blk_addr == NEW_ADDR);
J
Jaegeuk Kim 已提交
286 287 288 289 290 291
	} 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.
		 */
292
		copy_node_info(&e->ni, ni);
293
		f2fs_bug_on(sbi, ni->blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
294 295 296
	}

	/* sanity check */
297 298
	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
299
			new_blkaddr == NULL_ADDR);
300
	f2fs_bug_on(sbi, nat_get_blkaddr(e) == NEW_ADDR &&
J
Jaegeuk Kim 已提交
301
			new_blkaddr == NEW_ADDR);
302
	f2fs_bug_on(sbi, nat_get_blkaddr(e) != NEW_ADDR &&
J
Jaegeuk Kim 已提交
303 304 305
			nat_get_blkaddr(e) != NULL_ADDR &&
			new_blkaddr == NEW_ADDR);

A
arter97 已提交
306
	/* increment version no as node is removed */
J
Jaegeuk Kim 已提交
307 308 309
	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
310 311 312 313

		/* 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
314 315 316 317
	}

	/* change address */
	nat_set_blkaddr(e, new_blkaddr);
318 319
	if (new_blkaddr == NEW_ADDR || new_blkaddr == NULL_ADDR)
		set_nat_flag(e, IS_CHECKPOINTED, false);
J
Jaegeuk Kim 已提交
320
	__set_nat_cache_dirty(nm_i, e);
321 322

	/* update fsync_mark if its inode nat entry is still alive */
323 324
	if (ni->nid != ni->ino)
		e = __lookup_nat_cache(nm_i, ni->ino);
325 326 327 328 329
	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);
	}
330
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
331 332
}

333
int try_to_free_nats(struct f2fs_sb_info *sbi, int nr_shrink)
J
Jaegeuk Kim 已提交
334 335
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
336
	int nr = nr_shrink;
J
Jaegeuk Kim 已提交
337

J
Jaegeuk Kim 已提交
338
	if (!down_write_trylock(&nm_i->nat_tree_lock))
J
Jaegeuk Kim 已提交
339 340 341 342 343 344 345 346 347
		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--;
	}
348
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
349
	return nr - nr_shrink;
J
Jaegeuk Kim 已提交
350 351
}

J
Jaegeuk Kim 已提交
352
/*
A
arter97 已提交
353
 * This function always returns success
J
Jaegeuk Kim 已提交
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
 */
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);
	struct f2fs_summary_block *sum = curseg->sum_blk;
	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 */
370
	down_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
371 372 373 374 375 376
	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);
	}
377
	up_read(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
378 379 380
	if (e)
		return;

381 382
	memset(&ne, 0, sizeof(struct f2fs_nat_entry));

J
Jaegeuk Kim 已提交
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404
	/* Check current segment summary */
	mutex_lock(&curseg->curseg_mutex);
	i = lookup_journal_in_cursum(sum, NAT_JOURNAL, nid, 0);
	if (i >= 0) {
		ne = nat_in_journal(sum, i);
		node_info_from_raw_nat(ni, &ne);
	}
	mutex_unlock(&curseg->curseg_mutex);
	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:
	/* cache nat entry */
	cache_nat_entry(NM_I(sbi), nid, &ne);
}

J
Jaegeuk Kim 已提交
405
/*
J
Jaegeuk Kim 已提交
406 407 408
 * The maximum depth is four.
 * Offset[0] will have raw inode offset.
 */
409 410
static int get_node_path(struct f2fs_inode_info *fi, long block,
				int offset[4], unsigned int noffset[4])
J
Jaegeuk Kim 已提交
411
{
412
	const long direct_index = ADDRS_PER_INODE(fi);
J
Jaegeuk Kim 已提交
413 414 415 416 417 418 419 420 421 422
	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) {
423
		offset[n] = block;
J
Jaegeuk Kim 已提交
424 425 426 427 428 429
		goto got;
	}
	block -= direct_index;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR1_BLOCK;
		noffset[n] = 1;
430
		offset[n] = block;
J
Jaegeuk Kim 已提交
431 432 433 434 435 436 437
		level = 1;
		goto got;
	}
	block -= direct_blks;
	if (block < direct_blks) {
		offset[n++] = NODE_DIR2_BLOCK;
		noffset[n] = 2;
438
		offset[n] = block;
J
Jaegeuk Kim 已提交
439 440 441 442 443 444 445 446 447
		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];
448
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
449 450 451 452 453 454 455 456 457
		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];
458
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
459 460 461 462 463 464 465 466 467 468 469 470 471 472
		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];
473
		offset[n] = block % direct_blks;
J
Jaegeuk Kim 已提交
474 475 476 477 478 479 480 481 482 483 484
		level = 3;
		goto got;
	} else {
		BUG();
	}
got:
	return level;
}

/*
 * Caller should call f2fs_put_dnode(dn).
C
Chao Yu 已提交
485 486
 * 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.
487
 * In the case of RDONLY_NODE, we don't need to care about mutex.
J
Jaegeuk Kim 已提交
488
 */
489
int get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
J
Jaegeuk Kim 已提交
490
{
491
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
492
	struct page *npage[4];
493
	struct page *parent = NULL;
J
Jaegeuk Kim 已提交
494 495 496 497 498 499
	int offset[4];
	unsigned int noffset[4];
	nid_t nids[4];
	int level, i;
	int err = 0;

500
	level = get_node_path(F2FS_I(dn->inode), index, offset, noffset);
J
Jaegeuk Kim 已提交
501 502

	nids[0] = dn->inode->i_ino;
503
	npage[0] = dn->inode_page;
J
Jaegeuk Kim 已提交
504

505 506 507 508 509
	if (!npage[0]) {
		npage[0] = get_node_page(sbi, nids[0]);
		if (IS_ERR(npage[0]))
			return PTR_ERR(npage[0]);
	}
510 511 512

	/* if inline_data is set, should not report any block indices */
	if (f2fs_has_inline_data(dn->inode) && index) {
513
		err = -ENOENT;
514 515 516 517
		f2fs_put_page(npage[0], 1);
		goto release_out;
	}

J
Jaegeuk Kim 已提交
518
	parent = npage[0];
519 520
	if (level != 0)
		nids[1] = get_nid(parent, offset[0], true);
J
Jaegeuk Kim 已提交
521 522 523 524 525 526 527
	dn->inode_page = npage[0];
	dn->inode_page_locked = true;

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

528
		if (!nids[i] && mode == ALLOC_NODE) {
J
Jaegeuk Kim 已提交
529 530 531 532 533 534 535
			/* alloc new node */
			if (!alloc_nid(sbi, &(nids[i]))) {
				err = -ENOSPC;
				goto release_pages;
			}

			dn->nid = nids[i];
536
			npage[i] = new_node_page(dn, noffset[i], NULL);
J
Jaegeuk Kim 已提交
537 538 539 540 541 542 543 544
			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]);
545
			dn->node_changed = true;
J
Jaegeuk Kim 已提交
546
			done = true;
547
		} else if (mode == LOOKUP_NODE_RA && i == level && level > 1) {
J
Jaegeuk Kim 已提交
548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592
			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;
	return err;
}

static void truncate_node(struct dnode_of_data *dn)
{
593
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
594 595 596
	struct node_info ni;

	get_node_info(sbi, dn->nid, &ni);
597
	if (dn->inode->i_blocks == 0) {
598
		f2fs_bug_on(sbi, ni.blk_addr != NULL_ADDR);
599 600
		goto invalidate;
	}
601
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
J
Jaegeuk Kim 已提交
602 603

	/* Deallocate node address */
604
	invalidate_blocks(sbi, ni.blk_addr);
605
	dec_valid_node_count(sbi, dn->inode);
606
	set_node_addr(sbi, &ni, NULL_ADDR, false);
J
Jaegeuk Kim 已提交
607 608 609 610 611 612 613

	if (dn->nid == dn->inode->i_ino) {
		remove_orphan_inode(sbi, dn->nid);
		dec_valid_inode_count(sbi);
	} else {
		sync_inode_page(dn);
	}
614
invalidate:
J
Jaegeuk Kim 已提交
615
	clear_node_page_dirty(dn->node_page);
616
	set_sbi_flag(sbi, SBI_IS_DIRTY);
J
Jaegeuk Kim 已提交
617 618

	f2fs_put_page(dn->node_page, 1);
619 620 621 622

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

J
Jaegeuk Kim 已提交
623
	dn->node_page = NULL;
624
	trace_f2fs_truncate_node(dn->inode, dn->nid, ni.blk_addr);
J
Jaegeuk Kim 已提交
625 626 627 628 629 630 631 632 633 634
}

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

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

	/* get direct node */
635
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
J
Jaegeuk Kim 已提交
636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662
	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;

663 664
	trace_f2fs_truncate_nodes_enter(dn->inode, dn->nid, dn->data_blkaddr);

665
	page = get_node_page(F2FS_I_SB(dn->inode), dn->nid);
666 667
	if (IS_ERR(page)) {
		trace_f2fs_truncate_nodes_exit(dn->inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
668
		return PTR_ERR(page);
669
	}
J
Jaegeuk Kim 已提交
670

671
	rn = F2FS_NODE(page);
J
Jaegeuk Kim 已提交
672 673 674 675 676 677 678 679 680 681
	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;
			set_nid(page, i, 0, false);
682
			dn->node_changed = true;
J
Jaegeuk Kim 已提交
683 684 685 686 687 688 689 690 691 692 693 694 695
		}
	} 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)) {
				set_nid(page, i, 0, false);
696
				dn->node_changed = true;
J
Jaegeuk Kim 已提交
697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
				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);
	}
713
	trace_f2fs_truncate_nodes_exit(dn->inode, freed);
J
Jaegeuk Kim 已提交
714 715 716 717
	return freed;

out_err:
	f2fs_put_page(page, 1);
718
	trace_f2fs_truncate_nodes_exit(dn->inode, ret);
J
Jaegeuk Kim 已提交
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736
	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 */
737
	for (i = 0; i < idx + 1; i++) {
A
arter97 已提交
738
		/* reference count'll be increased */
739
		pages[i] = get_node_page(F2FS_I_SB(dn->inode), nid[i]);
J
Jaegeuk Kim 已提交
740 741
		if (IS_ERR(pages[i])) {
			err = PTR_ERR(pages[i]);
742
			idx = i - 1;
J
Jaegeuk Kim 已提交
743 744 745 746 747 748
			goto fail;
		}
		nid[i + 1] = get_nid(pages[i], offset[i + 1], false);
	}

	/* free direct nodes linked to a partial indirect node */
749
	for (i = offset[idx + 1]; i < NIDS_PER_BLOCK; i++) {
J
Jaegeuk Kim 已提交
750 751 752 753 754 755 756 757
		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;
		set_nid(pages[idx], i, 0, false);
758
		dn->node_changed = true;
J
Jaegeuk Kim 已提交
759 760
	}

761
	if (offset[idx + 1] == 0) {
J
Jaegeuk Kim 已提交
762 763 764 765 766 767 768
		dn->node_page = pages[idx];
		dn->nid = nid[idx];
		truncate_node(dn);
	} else {
		f2fs_put_page(pages[idx], 1);
	}
	offset[idx]++;
769 770
	offset[idx + 1] = 0;
	idx--;
J
Jaegeuk Kim 已提交
771
fail:
772
	for (i = idx; i >= 0; i--)
J
Jaegeuk Kim 已提交
773
		f2fs_put_page(pages[i], 1);
774 775 776

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

J
Jaegeuk Kim 已提交
777 778 779
	return err;
}

J
Jaegeuk Kim 已提交
780
/*
J
Jaegeuk Kim 已提交
781 782 783 784
 * All the block addresses of data and nodes should be nullified.
 */
int truncate_inode_blocks(struct inode *inode, pgoff_t from)
{
785
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
J
Jaegeuk Kim 已提交
786 787
	int err = 0, cont = 1;
	int level, offset[4], noffset[4];
J
Jaegeuk Kim 已提交
788
	unsigned int nofs = 0;
789
	struct f2fs_inode *ri;
J
Jaegeuk Kim 已提交
790 791 792
	struct dnode_of_data dn;
	struct page *page;

793 794
	trace_f2fs_truncate_inode_blocks_enter(inode, from);

795
	level = get_node_path(F2FS_I(inode), from, offset, noffset);
796
restart:
J
Jaegeuk Kim 已提交
797
	page = get_node_page(sbi, inode->i_ino);
798 799
	if (IS_ERR(page)) {
		trace_f2fs_truncate_inode_blocks_exit(inode, PTR_ERR(page));
J
Jaegeuk Kim 已提交
800
		return PTR_ERR(page);
801
	}
J
Jaegeuk Kim 已提交
802 803 804 805

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

806
	ri = F2FS_INODE(page);
J
Jaegeuk Kim 已提交
807 808 809 810 811 812 813 814 815
	switch (level) {
	case 0:
	case 1:
		nofs = noffset[1];
		break;
	case 2:
		nofs = noffset[1];
		if (!offset[level - 1])
			goto skip_partial;
816
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
817 818 819 820 821 822 823 824
		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;
825
		err = truncate_partial_nodes(&dn, ri, offset, level);
J
Jaegeuk Kim 已提交
826 827 828 829 830 831 832 833 834
		if (err < 0 && err != -ENOENT)
			goto fail;
		break;
	default:
		BUG();
	}

skip_partial:
	while (cont) {
835
		dn.nid = le32_to_cpu(ri->i_nid[offset[0] - NODE_DIR1_BLOCK]);
J
Jaegeuk Kim 已提交
836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857
		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 &&
858
				ri->i_nid[offset[0] - NODE_DIR1_BLOCK]) {
J
Jaegeuk Kim 已提交
859
			lock_page(page);
860
			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
861 862 863
				f2fs_put_page(page, 1);
				goto restart;
			}
864
			f2fs_wait_on_page_writeback(page, NODE);
865
			ri->i_nid[offset[0] - NODE_DIR1_BLOCK] = 0;
J
Jaegeuk Kim 已提交
866 867 868 869 870 871 872 873 874
			set_page_dirty(page);
			unlock_page(page);
		}
		offset[1] = 0;
		offset[0]++;
		nofs += err;
	}
fail:
	f2fs_put_page(page, 0);
875
	trace_f2fs_truncate_inode_blocks_exit(inode, err);
J
Jaegeuk Kim 已提交
876 877 878
	return err > 0 ? 0 : err;
}

879 880
int truncate_xattr_node(struct inode *inode, struct page *page)
{
881
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
882 883 884 885 886 887 888 889 890 891 892 893
	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
894 895 896 897

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

898 899 900
	set_new_dnode(&dn, inode, page, npage, nid);

	if (page)
901
		dn.inode_page_locked = true;
902 903 904 905
	truncate_node(&dn);
	return 0;
}

906
/*
C
Chao Yu 已提交
907 908
 * Caller should grab and release a rwsem by calling f2fs_lock_op() and
 * f2fs_unlock_op().
909
 */
C
Chao Yu 已提交
910
int remove_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
911 912
{
	struct dnode_of_data dn;
C
Chao Yu 已提交
913
	int err;
J
Jaegeuk Kim 已提交
914

915
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
C
Chao Yu 已提交
916 917 918
	err = get_dnode_of_data(&dn, 0, LOOKUP_NODE);
	if (err)
		return err;
J
Jaegeuk Kim 已提交
919

C
Chao Yu 已提交
920 921
	err = truncate_xattr_node(inode, dn.inode_page);
	if (err) {
922
		f2fs_put_dnode(&dn);
C
Chao Yu 已提交
923
		return err;
J
Jaegeuk Kim 已提交
924
	}
925 926 927 928 929 930

	/* 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
931
	/* 0 is possible, after f2fs_new_inode() has failed */
932 933
	f2fs_bug_on(F2FS_I_SB(inode),
			inode->i_blocks != 0 && inode->i_blocks != 1);
934 935

	/* will put inode & node pages */
936
	truncate_node(&dn);
C
Chao Yu 已提交
937
	return 0;
J
Jaegeuk Kim 已提交
938 939
}

940
struct page *new_inode_page(struct inode *inode)
J
Jaegeuk Kim 已提交
941 942 943 944 945
{
	struct dnode_of_data dn;

	/* allocate inode page for new inode */
	set_new_dnode(&dn, inode, NULL, NULL, inode->i_ino);
946 947

	/* caller should f2fs_put_page(page, 1); */
948
	return new_node_page(&dn, 0, NULL);
J
Jaegeuk Kim 已提交
949 950
}

951 952
struct page *new_node_page(struct dnode_of_data *dn,
				unsigned int ofs, struct page *ipage)
J
Jaegeuk Kim 已提交
953
{
954
	struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
J
Jaegeuk Kim 已提交
955 956 957 958
	struct node_info old_ni, new_ni;
	struct page *page;
	int err;

959
	if (unlikely(is_inode_flag_set(F2FS_I(dn->inode), FI_NO_ALLOC)))
J
Jaegeuk Kim 已提交
960 961
		return ERR_PTR(-EPERM);

962
	page = grab_cache_page(NODE_MAPPING(sbi), dn->nid);
J
Jaegeuk Kim 已提交
963 964 965
	if (!page)
		return ERR_PTR(-ENOMEM);

966
	if (unlikely(!inc_valid_node_count(sbi, dn->inode))) {
967 968 969
		err = -ENOSPC;
		goto fail;
	}
J
Jaegeuk Kim 已提交
970

971
	get_node_info(sbi, dn->nid, &old_ni);
J
Jaegeuk Kim 已提交
972 973

	/* Reinitialize old_ni with new node page */
974
	f2fs_bug_on(sbi, old_ni.blk_addr != NULL_ADDR);
J
Jaegeuk Kim 已提交
975 976
	new_ni = old_ni;
	new_ni.ino = dn->inode->i_ino;
977
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
978

979
	f2fs_wait_on_page_writeback(page, NODE);
980
	fill_node_footer(page, dn->nid, dn->inode->i_ino, ofs, true);
981
	set_cold_node(dn->inode, page);
982 983
	SetPageUptodate(page);
	set_page_dirty(page);
J
Jaegeuk Kim 已提交
984

985
	if (f2fs_has_xattr_block(ofs))
986 987
		F2FS_I(dn->inode)->i_xattr_nid = dn->nid;

J
Jaegeuk Kim 已提交
988
	dn->node_page = page;
989 990 991 992
	if (ipage)
		update_inode(dn->inode, ipage);
	else
		sync_inode_page(dn);
J
Jaegeuk Kim 已提交
993 994 995 996 997 998
	if (ofs == 0)
		inc_valid_inode_count(sbi);

	return page;

fail:
999
	clear_node_page_dirty(page);
J
Jaegeuk Kim 已提交
1000 1001 1002 1003
	f2fs_put_page(page, 1);
	return ERR_PTR(err);
}

1004 1005 1006
/*
 * Caller should do after getting the following values.
 * 0: f2fs_put_page(page, 0)
1007
 * LOCKED_PAGE or error: f2fs_put_page(page, 1)
1008
 */
1009
static int read_node_page(struct page *page, int rw)
J
Jaegeuk Kim 已提交
1010
{
1011
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1012
	struct node_info ni;
1013
	struct f2fs_io_info fio = {
1014
		.sbi = sbi,
1015 1016
		.type = NODE,
		.rw = rw,
1017
		.page = page,
1018
		.encrypted_page = NULL,
1019
	};
J
Jaegeuk Kim 已提交
1020 1021 1022

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

1023
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1024
		ClearPageUptodate(page);
J
Jaegeuk Kim 已提交
1025
		return -ENOENT;
1026 1027
	}

1028 1029
	if (PageUptodate(page))
		return LOCKED_PAGE;
1030

1031
	fio.blk_addr = ni.blk_addr;
1032
	return f2fs_submit_page_bio(&fio);
J
Jaegeuk Kim 已提交
1033 1034
}

J
Jaegeuk Kim 已提交
1035
/*
J
Jaegeuk Kim 已提交
1036 1037 1038 1039 1040
 * Readahead a node page
 */
void ra_node_page(struct f2fs_sb_info *sbi, nid_t nid)
{
	struct page *apage;
1041
	int err;
J
Jaegeuk Kim 已提交
1042

1043
	apage = find_get_page(NODE_MAPPING(sbi), nid);
1044 1045 1046 1047
	if (apage && PageUptodate(apage)) {
		f2fs_put_page(apage, 0);
		return;
	}
J
Jaegeuk Kim 已提交
1048 1049
	f2fs_put_page(apage, 0);

1050
	apage = grab_cache_page(NODE_MAPPING(sbi), nid);
J
Jaegeuk Kim 已提交
1051 1052 1053
	if (!apage)
		return;

1054
	err = read_node_page(apage, READA);
1055
	f2fs_put_page(apage, err ? 1 : 0);
J
Jaegeuk Kim 已提交
1056 1057 1058 1059
}

struct page *get_node_page(struct f2fs_sb_info *sbi, pgoff_t nid)
{
1060 1061
	struct page *page;
	int err;
1062 1063 1064 1065

	if (!nid)
		return ERR_PTR(-ENOENT);
	f2fs_bug_on(sbi, check_nid_range(sbi, nid));
1066
repeat:
1067
	page = grab_cache_page(NODE_MAPPING(sbi), nid);
J
Jaegeuk Kim 已提交
1068 1069 1070 1071
	if (!page)
		return ERR_PTR(-ENOMEM);

	err = read_node_page(page, READ_SYNC);
1072 1073
	if (err < 0) {
		f2fs_put_page(page, 1);
J
Jaegeuk Kim 已提交
1074
		return ERR_PTR(err);
C
Chao Yu 已提交
1075 1076
	} else if (err == LOCKED_PAGE) {
		goto page_hit;
1077
	}
J
Jaegeuk Kim 已提交
1078

C
Chao Yu 已提交
1079 1080 1081
	lock_page(page);

	if (unlikely(!PageUptodate(page))) {
1082 1083 1084
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
	}
1085
	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
1086 1087 1088
		f2fs_put_page(page, 1);
		goto repeat;
	}
C
Chao Yu 已提交
1089 1090
page_hit:
	f2fs_bug_on(sbi, nid != nid_of_node(page));
J
Jaegeuk Kim 已提交
1091 1092 1093
	return page;
}

J
Jaegeuk Kim 已提交
1094
/*
J
Jaegeuk Kim 已提交
1095 1096 1097 1098 1099
 * Return a locked page for the desired node page.
 * And, readahead MAX_RA_NODE number of node pages.
 */
struct page *get_node_page_ra(struct page *parent, int start)
{
1100
	struct f2fs_sb_info *sbi = F2FS_P_SB(parent);
1101
	struct blk_plug plug;
J
Jaegeuk Kim 已提交
1102
	struct page *page;
1103 1104
	int err, i, end;
	nid_t nid;
J
Jaegeuk Kim 已提交
1105 1106 1107 1108 1109

	/* First, try getting the desired direct node. */
	nid = get_nid(parent, start, false);
	if (!nid)
		return ERR_PTR(-ENOENT);
1110
repeat:
1111
	page = grab_cache_page(NODE_MAPPING(sbi), nid);
J
Jaegeuk Kim 已提交
1112 1113 1114
	if (!page)
		return ERR_PTR(-ENOMEM);

1115
	err = read_node_page(page, READ_SYNC);
1116 1117
	if (err < 0) {
		f2fs_put_page(page, 1);
J
Jaegeuk Kim 已提交
1118
		return ERR_PTR(err);
1119
	} else if (err == LOCKED_PAGE) {
1120
		goto page_hit;
1121
	}
J
Jaegeuk Kim 已提交
1122

1123 1124
	blk_start_plug(&plug);

J
Jaegeuk Kim 已提交
1125 1126 1127 1128
	/* Then, try readahead for siblings of the desired node */
	end = start + MAX_RA_NODE;
	end = min(end, NIDS_PER_BLOCK);
	for (i = start + 1; i < end; i++) {
C
Chao Yu 已提交
1129 1130
		nid_t tnid = get_nid(parent, i, false);
		if (!tnid)
J
Jaegeuk Kim 已提交
1131
			continue;
C
Chao Yu 已提交
1132
		ra_node_page(sbi, tnid);
J
Jaegeuk Kim 已提交
1133 1134
	}

1135 1136
	blk_finish_plug(&plug);

J
Jaegeuk Kim 已提交
1137
	lock_page(page);
C
Chao Yu 已提交
1138 1139 1140 1141
	if (unlikely(!PageUptodate(page))) {
		f2fs_put_page(page, 1);
		return ERR_PTR(-EIO);
	}
1142
	if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
1143 1144 1145
		f2fs_put_page(page, 1);
		goto repeat;
	}
1146
page_hit:
C
Chao Yu 已提交
1147
	f2fs_bug_on(sbi, nid != nid_of_node(page));
J
Jaegeuk Kim 已提交
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161
	return page;
}

void sync_inode_page(struct dnode_of_data *dn)
{
	if (IS_INODE(dn->node_page) || dn->inode_page == dn->node_page) {
		update_inode(dn->inode, dn->node_page);
	} else if (dn->inode_page) {
		if (!dn->inode_page_locked)
			lock_page(dn->inode_page);
		update_inode(dn->inode, dn->inode_page);
		if (!dn->inode_page_locked)
			unlock_page(dn->inode_page);
	} else {
1162
		update_inode_page(dn->inode);
J
Jaegeuk Kim 已提交
1163
	}
1164
	dn->node_changed = true;
J
Jaegeuk Kim 已提交
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
}

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;
	end = LONG_MAX;

	while (index <= end) {
		int i, nr_pages;
1183
		nr_pages = pagevec_lookup_tag(&pvec, NODE_MAPPING(sbi), &index,
J
Jaegeuk Kim 已提交
1184 1185 1186 1187 1188 1189 1190 1191
				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
1192 1193 1194 1195 1196
			if (unlikely(f2fs_cp_error(sbi))) {
				pagevec_release(&pvec);
				return -EIO;
			}

J
Jaegeuk Kim 已提交
1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220
			/*
			 * 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.
			 */
			if (ino && ino_of_node(page) == ino)
				lock_page(page);
			else if (!trylock_page(page))
				continue;

1221
			if (unlikely(page->mapping != NODE_MAPPING(sbi))) {
J
Jaegeuk Kim 已提交
1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239
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;
			}

			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
1240 1241 1242
				if (IS_INODE(page))
					set_dentry_mark(page,
						need_dentry_mark(sbi, ino));
J
Jaegeuk Kim 已提交
1243 1244 1245 1246 1247
				nwritten++;
			} else {
				set_fsync_mark(page, 0);
				set_dentry_mark(page, 0);
			}
1248 1249 1250 1251 1252

			if (NODE_MAPPING(sbi)->a_ops->writepage(page, wbc))
				unlock_page(page);
			else
				wrote++;
J
Jaegeuk Kim 已提交
1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271

			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;
	}

	if (wrote)
J
Jaegeuk Kim 已提交
1272
		f2fs_submit_merged_bio(sbi, NODE, WRITE);
J
Jaegeuk Kim 已提交
1273 1274 1275
	return nwritten;
}

1276 1277 1278 1279 1280 1281 1282
int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
{
	pgoff_t index = 0, end = LONG_MAX;
	struct pagevec pvec;
	int ret2 = 0, ret = 0;

	pagevec_init(&pvec, 0);
1283 1284 1285 1286 1287 1288 1289 1290

	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;
1291 1292 1293 1294 1295

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

			/* until radix tree lookup accepts end_index */
1296
			if (unlikely(page->index > end))
1297 1298
				continue;

1299
			if (ino && ino_of_node(page) == ino) {
1300
				f2fs_wait_on_page_writeback(page, NODE);
1301 1302 1303
				if (TestClearPageError(page))
					ret = -EIO;
			}
1304 1305 1306 1307 1308
		}
		pagevec_release(&pvec);
		cond_resched();
	}

1309
	if (unlikely(test_and_clear_bit(AS_ENOSPC, &NODE_MAPPING(sbi)->flags)))
1310
		ret2 = -ENOSPC;
1311
	if (unlikely(test_and_clear_bit(AS_EIO, &NODE_MAPPING(sbi)->flags)))
1312 1313 1314 1315 1316 1317
		ret2 = -EIO;
	if (!ret)
		ret = ret2;
	return ret;
}

J
Jaegeuk Kim 已提交
1318 1319 1320
static int f2fs_write_node_page(struct page *page,
				struct writeback_control *wbc)
{
1321
	struct f2fs_sb_info *sbi = F2FS_P_SB(page);
J
Jaegeuk Kim 已提交
1322 1323
	nid_t nid;
	struct node_info ni;
1324
	struct f2fs_io_info fio = {
1325
		.sbi = sbi,
1326
		.type = NODE,
C
Chris Fries 已提交
1327
		.rw = (wbc->sync_mode == WB_SYNC_ALL) ? WRITE_SYNC : WRITE,
1328
		.page = page,
1329
		.encrypted_page = NULL,
1330
	};
J
Jaegeuk Kim 已提交
1331

1332 1333
	trace_f2fs_writepage(page, NODE);

1334
	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
1335
		goto redirty_out;
1336 1337
	if (unlikely(f2fs_cp_error(sbi)))
		goto redirty_out;
1338

1339
	f2fs_wait_on_page_writeback(page, NODE);
J
Jaegeuk Kim 已提交
1340 1341 1342

	/* get old block addr of this node page */
	nid = nid_of_node(page);
1343
	f2fs_bug_on(sbi, page->index != nid);
J
Jaegeuk Kim 已提交
1344

1345 1346 1347 1348 1349 1350 1351
	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
1352 1353 1354
	get_node_info(sbi, nid, &ni);

	/* This page is already truncated */
1355
	if (unlikely(ni.blk_addr == NULL_ADDR)) {
1356
		ClearPageUptodate(page);
1357
		dec_page_count(sbi, F2FS_DIRTY_NODES);
1358
		up_read(&sbi->node_write);
1359 1360 1361
		unlock_page(page);
		return 0;
	}
J
Jaegeuk Kim 已提交
1362 1363

	set_page_writeback(page);
1364
	fio.blk_addr = ni.blk_addr;
1365
	write_node_page(nid, &fio);
1366
	set_node_addr(sbi, &ni, fio.blk_addr, is_fsync_dnode(page));
J
Jaegeuk Kim 已提交
1367
	dec_page_count(sbi, F2FS_DIRTY_NODES);
1368
	up_read(&sbi->node_write);
J
Jaegeuk Kim 已提交
1369
	unlock_page(page);
1370 1371 1372 1373

	if (wbc->for_reclaim)
		f2fs_submit_merged_bio(sbi, NODE, WRITE);

J
Jaegeuk Kim 已提交
1374
	return 0;
1375 1376

redirty_out:
1377
	redirty_page_for_writepage(wbc, page);
1378
	return AOP_WRITEPAGE_ACTIVATE;
J
Jaegeuk Kim 已提交
1379 1380 1381 1382 1383
}

static int f2fs_write_node_pages(struct address_space *mapping,
			    struct writeback_control *wbc)
{
1384
	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
1385
	long diff;
J
Jaegeuk Kim 已提交
1386

1387 1388
	trace_f2fs_writepages(mapping->host, wbc, NODE);

1389 1390
	/* balancing f2fs's metadata in background */
	f2fs_balance_fs_bg(sbi);
J
Jaegeuk Kim 已提交
1391

1392
	/* collect a number of dirty node pages and write together */
1393
	if (get_pages(sbi, F2FS_DIRTY_NODES) < nr_pages_to_skip(sbi, NODE))
1394
		goto skip_write;
1395

1396
	diff = nr_pages_to_write(sbi, NODE, wbc);
1397
	wbc->sync_mode = WB_SYNC_NONE;
J
Jaegeuk Kim 已提交
1398
	sync_node_pages(sbi, 0, wbc);
1399
	wbc->nr_to_write = max((long)0, wbc->nr_to_write - diff);
J
Jaegeuk Kim 已提交
1400
	return 0;
1401 1402 1403 1404

skip_write:
	wbc->pages_skipped += get_pages(sbi, F2FS_DIRTY_NODES);
	return 0;
J
Jaegeuk Kim 已提交
1405 1406 1407 1408
}

static int f2fs_set_node_page_dirty(struct page *page)
{
1409 1410
	trace_f2fs_set_page_dirty(page, NODE);

J
Jaegeuk Kim 已提交
1411 1412 1413
	SetPageUptodate(page);
	if (!PageDirty(page)) {
		__set_page_dirty_nobuffers(page);
1414
		inc_page_count(F2FS_P_SB(page), F2FS_DIRTY_NODES);
J
Jaegeuk Kim 已提交
1415
		SetPagePrivate(page);
J
Jaegeuk Kim 已提交
1416
		f2fs_trace_pid(page);
J
Jaegeuk Kim 已提交
1417 1418 1419 1420 1421
		return 1;
	}
	return 0;
}

J
Jaegeuk Kim 已提交
1422
/*
J
Jaegeuk Kim 已提交
1423 1424 1425 1426 1427 1428
 * 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,
1429 1430
	.invalidatepage	= f2fs_invalidate_page,
	.releasepage	= f2fs_release_page,
J
Jaegeuk Kim 已提交
1431 1432
};

1433 1434
static struct free_nid *__lookup_free_nid_list(struct f2fs_nm_info *nm_i,
						nid_t n)
J
Jaegeuk Kim 已提交
1435
{
1436
	return radix_tree_lookup(&nm_i->free_nid_root, n);
J
Jaegeuk Kim 已提交
1437 1438
}

1439 1440
static void __del_from_free_nid_list(struct f2fs_nm_info *nm_i,
						struct free_nid *i)
J
Jaegeuk Kim 已提交
1441 1442
{
	list_del(&i->list);
1443
	radix_tree_delete(&nm_i->free_nid_root, i->nid);
J
Jaegeuk Kim 已提交
1444 1445
}

1446
static int add_free_nid(struct f2fs_sb_info *sbi, nid_t nid, bool build)
J
Jaegeuk Kim 已提交
1447
{
1448
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1449
	struct free_nid *i;
1450 1451
	struct nat_entry *ne;
	bool allocated = false;
J
Jaegeuk Kim 已提交
1452

1453
	if (!available_free_memory(sbi, FREE_NIDS))
H
Haicheng Li 已提交
1454
		return -1;
1455 1456

	/* 0 nid should not be used */
1457
	if (unlikely(nid == 0))
1458
		return 0;
1459

1460 1461
	if (build) {
		/* do not add allocated nids */
1462
		down_read(&nm_i->nat_tree_lock);
1463
		ne = __lookup_nat_cache(nm_i, nid);
1464
		if (ne &&
1465 1466
			(!get_nat_flag(ne, IS_CHECKPOINTED) ||
				nat_get_blkaddr(ne) != NULL_ADDR))
1467
			allocated = true;
1468
		up_read(&nm_i->nat_tree_lock);
1469 1470
		if (allocated)
			return 0;
J
Jaegeuk Kim 已提交
1471
	}
1472 1473

	i = f2fs_kmem_cache_alloc(free_nid_slab, GFP_NOFS);
J
Jaegeuk Kim 已提交
1474 1475 1476
	i->nid = nid;
	i->state = NID_NEW;

1477 1478 1479 1480 1481
	if (radix_tree_preload(GFP_NOFS)) {
		kmem_cache_free(free_nid_slab, i);
		return 0;
	}

J
Jaegeuk Kim 已提交
1482
	spin_lock(&nm_i->free_nid_list_lock);
1483
	if (radix_tree_insert(&nm_i->free_nid_root, i->nid, i)) {
J
Jaegeuk Kim 已提交
1484
		spin_unlock(&nm_i->free_nid_list_lock);
1485
		radix_tree_preload_end();
J
Jaegeuk Kim 已提交
1486 1487 1488 1489 1490 1491
		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);
1492
	radix_tree_preload_end();
J
Jaegeuk Kim 已提交
1493 1494 1495 1496 1497 1498
	return 1;
}

static void remove_free_nid(struct f2fs_nm_info *nm_i, nid_t nid)
{
	struct free_nid *i;
1499 1500
	bool need_free = false;

J
Jaegeuk Kim 已提交
1501
	spin_lock(&nm_i->free_nid_list_lock);
1502
	i = __lookup_free_nid_list(nm_i, nid);
J
Jaegeuk Kim 已提交
1503
	if (i && i->state == NID_NEW) {
1504
		__del_from_free_nid_list(nm_i, i);
J
Jaegeuk Kim 已提交
1505
		nm_i->fcnt--;
1506
		need_free = true;
J
Jaegeuk Kim 已提交
1507 1508
	}
	spin_unlock(&nm_i->free_nid_list_lock);
1509 1510 1511

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1512 1513
}

1514
static void scan_nat_page(struct f2fs_sb_info *sbi,
J
Jaegeuk Kim 已提交
1515 1516
			struct page *nat_page, nid_t start_nid)
{
1517
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1518 1519 1520 1521 1522 1523 1524
	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
1525

1526
		if (unlikely(start_nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1527
			break;
H
Haicheng Li 已提交
1528 1529

		blk_addr = le32_to_cpu(nat_blk->entries[i].block_addr);
1530
		f2fs_bug_on(sbi, blk_addr == NEW_ADDR);
H
Haicheng Li 已提交
1531
		if (blk_addr == NULL_ADDR) {
1532
			if (add_free_nid(sbi, start_nid, true) < 0)
H
Haicheng Li 已提交
1533 1534
				break;
		}
J
Jaegeuk Kim 已提交
1535 1536 1537 1538 1539 1540 1541 1542
	}
}

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);
	struct f2fs_summary_block *sum = curseg->sum_blk;
1543
	int i = 0;
1544
	nid_t nid = nm_i->next_scan_nid;
J
Jaegeuk Kim 已提交
1545

1546 1547 1548
	/* Enough entries */
	if (nm_i->fcnt > NAT_ENTRY_PER_BLOCK)
		return;
J
Jaegeuk Kim 已提交
1549

1550
	/* readahead nat pages to be scanned */
1551 1552
	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nid), FREE_NID_PAGES,
							META_NAT, true);
J
Jaegeuk Kim 已提交
1553 1554 1555 1556

	while (1) {
		struct page *page = get_current_nat_page(sbi, nid);

1557
		scan_nat_page(sbi, page, nid);
J
Jaegeuk Kim 已提交
1558 1559 1560
		f2fs_put_page(page, 1);

		nid += (NAT_ENTRY_PER_BLOCK - (nid % NAT_ENTRY_PER_BLOCK));
1561
		if (unlikely(nid >= nm_i->max_nid))
J
Jaegeuk Kim 已提交
1562
			nid = 0;
1563

1564
		if (++i >= FREE_NID_PAGES)
J
Jaegeuk Kim 已提交
1565 1566 1567
			break;
	}

1568 1569
	/* go to the next free nat pages to find free nids abundantly */
	nm_i->next_scan_nid = nid;
J
Jaegeuk Kim 已提交
1570 1571 1572 1573 1574 1575 1576

	/* find free nids from current sum_pages */
	mutex_lock(&curseg->curseg_mutex);
	for (i = 0; i < nats_in_cursum(sum); i++) {
		block_t addr = le32_to_cpu(nat_in_journal(sum, i).block_addr);
		nid = le32_to_cpu(nid_in_journal(sum, i));
		if (addr == NULL_ADDR)
1577
			add_free_nid(sbi, nid, true);
J
Jaegeuk Kim 已提交
1578 1579 1580 1581
		else
			remove_free_nid(nm_i, nid);
	}
	mutex_unlock(&curseg->curseg_mutex);
C
Chao Yu 已提交
1582 1583

	ra_meta_pages(sbi, NAT_BLOCK_OFFSET(nm_i->next_scan_nid),
C
Chao Yu 已提交
1584
					nm_i->ra_nid_pages, META_NAT, false);
J
Jaegeuk Kim 已提交
1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596
}

/*
 * 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:
1597
	if (unlikely(sbi->total_valid_node_count + 1 > nm_i->available_nids))
1598
		return false;
J
Jaegeuk Kim 已提交
1599 1600 1601

	spin_lock(&nm_i->free_nid_list_lock);

1602
	/* We should not use stale free nids created by build_free_nids */
1603
	if (nm_i->fcnt && !on_build_free_nids(nm_i)) {
1604 1605
		struct node_info ni;

1606
		f2fs_bug_on(sbi, list_empty(&nm_i->free_nid_list));
1607
		list_for_each_entry(i, &nm_i->free_nid_list, list)
1608 1609
			if (i->state == NID_NEW)
				break;
J
Jaegeuk Kim 已提交
1610

1611
		f2fs_bug_on(sbi, i->state != NID_NEW);
1612 1613 1614 1615
		*nid = i->nid;
		i->state = NID_ALLOC;
		nm_i->fcnt--;
		spin_unlock(&nm_i->free_nid_list_lock);
1616 1617 1618 1619 1620 1621 1622

		/* check nid is allocated already */
		get_node_info(sbi, *nid, &ni);
		if (ni.blk_addr != NULL_ADDR) {
			alloc_nid_done(sbi, *nid);
			goto retry;
		}
1623 1624
		return true;
	}
J
Jaegeuk Kim 已提交
1625
	spin_unlock(&nm_i->free_nid_list_lock);
1626 1627 1628 1629 1630 1631

	/* 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
1632 1633
}

J
Jaegeuk Kim 已提交
1634
/*
J
Jaegeuk Kim 已提交
1635 1636 1637 1638 1639 1640 1641 1642
 * 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);
1643
	i = __lookup_free_nid_list(nm_i, nid);
1644
	f2fs_bug_on(sbi, !i || i->state != NID_ALLOC);
1645
	__del_from_free_nid_list(nm_i, i);
J
Jaegeuk Kim 已提交
1646
	spin_unlock(&nm_i->free_nid_list_lock);
1647 1648

	kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1649 1650
}

J
Jaegeuk Kim 已提交
1651
/*
J
Jaegeuk Kim 已提交
1652 1653 1654 1655
 * alloc_nid() should be called prior to this function.
 */
void alloc_nid_failed(struct f2fs_sb_info *sbi, nid_t nid)
{
1656 1657
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct free_nid *i;
1658
	bool need_free = false;
1659

J
Jaegeuk Kim 已提交
1660 1661 1662
	if (!nid)
		return;

1663
	spin_lock(&nm_i->free_nid_list_lock);
1664
	i = __lookup_free_nid_list(nm_i, nid);
1665
	f2fs_bug_on(sbi, !i || i->state != NID_ALLOC);
1666
	if (!available_free_memory(sbi, FREE_NIDS)) {
1667
		__del_from_free_nid_list(nm_i, i);
1668
		need_free = true;
1669 1670 1671 1672
	} else {
		i->state = NID_NEW;
		nm_i->fcnt++;
	}
1673
	spin_unlock(&nm_i->free_nid_list_lock);
1674 1675 1676

	if (need_free)
		kmem_cache_free(free_nid_slab, i);
J
Jaegeuk Kim 已提交
1677 1678
}

C
Chao Yu 已提交
1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695
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);
1696
		nm_i->fcnt--;
C
Chao Yu 已提交
1697 1698 1699 1700 1701 1702 1703 1704
		nr_shrink--;
	}
	spin_unlock(&nm_i->free_nid_list_lock);
	mutex_unlock(&nm_i->build_lock);

	return nr - nr_shrink;
}

1705
void recover_inline_xattr(struct inode *inode, struct page *page)
1706 1707 1708 1709 1710 1711
{
	void *src_addr, *dst_addr;
	size_t inline_size;
	struct page *ipage;
	struct f2fs_inode *ri;

1712
	ipage = get_node_page(F2FS_I_SB(inode), inode->i_ino);
1713
	f2fs_bug_on(F2FS_I_SB(inode), IS_ERR(ipage));
1714

1715 1716 1717 1718 1719 1720
	ri = F2FS_INODE(page);
	if (!(ri->i_inline & F2FS_INLINE_XATTR)) {
		clear_inode_flag(F2FS_I(inode), FI_INLINE_XATTR);
		goto update_inode;
	}

1721 1722 1723 1724
	dst_addr = inline_xattr_addr(ipage);
	src_addr = inline_xattr_addr(page);
	inline_size = inline_xattr_size(inode);

1725
	f2fs_wait_on_page_writeback(ipage, NODE);
1726
	memcpy(dst_addr, src_addr, inline_size);
1727
update_inode:
1728 1729 1730 1731
	update_inode(inode, ipage);
	f2fs_put_page(ipage, 1);
}

1732
void recover_xattr_data(struct inode *inode, struct page *page, block_t blkaddr)
1733
{
1734
	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1735 1736 1737 1738 1739 1740 1741 1742 1743 1744
	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);
1745
	f2fs_bug_on(sbi, ni.blk_addr == NULL_ADDR);
1746 1747
	invalidate_blocks(sbi, ni.blk_addr);
	dec_valid_node_count(sbi, inode);
1748
	set_node_addr(sbi, &ni, NULL_ADDR, false);
1749 1750 1751 1752

recover_xnid:
	/* 2: allocate new xattr nid */
	if (unlikely(!inc_valid_node_count(sbi, inode)))
1753
		f2fs_bug_on(sbi, 1);
1754 1755 1756 1757

	remove_free_nid(NM_I(sbi), new_xnid);
	get_node_info(sbi, new_xnid, &ni);
	ni.ino = inode->i_ino;
1758
	set_node_addr(sbi, &ni, NEW_ADDR, false);
1759 1760 1761 1762
	F2FS_I(inode)->i_xattr_nid = new_xnid;

	/* 3: update xattr blkaddr */
	refresh_sit_entry(sbi, NEW_ADDR, blkaddr);
1763
	set_node_addr(sbi, &ni, blkaddr, false);
1764 1765 1766 1767

	update_inode_page(inode);
}

J
Jaegeuk Kim 已提交
1768 1769
int recover_inode_page(struct f2fs_sb_info *sbi, struct page *page)
{
1770
	struct f2fs_inode *src, *dst;
J
Jaegeuk Kim 已提交
1771 1772 1773 1774
	nid_t ino = ino_of_node(page);
	struct node_info old_ni, new_ni;
	struct page *ipage;

1775 1776 1777 1778 1779
	get_node_info(sbi, ino, &old_ni);

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

1780
	ipage = grab_cache_page(NODE_MAPPING(sbi), ino);
J
Jaegeuk Kim 已提交
1781 1782 1783
	if (!ipage)
		return -ENOMEM;

A
arter97 已提交
1784
	/* Should not use this inode from free nid list */
J
Jaegeuk Kim 已提交
1785 1786 1787 1788 1789
	remove_free_nid(NM_I(sbi), ino);

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

1790 1791
	src = F2FS_INODE(page);
	dst = F2FS_INODE(ipage);
J
Jaegeuk Kim 已提交
1792

1793 1794 1795 1796 1797
	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;
1798
	dst->i_inline = src->i_inline & F2FS_INLINE_XATTR;
J
Jaegeuk Kim 已提交
1799 1800 1801 1802

	new_ni = old_ni;
	new_ni.ino = ino;

1803
	if (unlikely(!inc_valid_node_count(sbi, NULL)))
1804
		WARN_ON(1);
1805
	set_node_addr(sbi, &new_ni, NEW_ADDR, false);
J
Jaegeuk Kim 已提交
1806
	inc_valid_inode_count(sbi);
1807
	set_page_dirty(ipage);
J
Jaegeuk Kim 已提交
1808 1809 1810 1811 1812 1813 1814 1815 1816 1817
	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
1818
	int bio_blocks = MAX_BIO_BLOCKS(sbi);
1819
	int i, idx, last_offset, nrpages;
J
Jaegeuk Kim 已提交
1820 1821 1822 1823 1824 1825

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

1826
	for (i = 0; i < last_offset; i += nrpages, addr += nrpages) {
1827
		nrpages = min(last_offset - i, bio_blocks);
1828

A
arter97 已提交
1829
		/* readahead node pages */
1830
		ra_meta_pages(sbi, addr, nrpages, META_POR, true);
J
Jaegeuk Kim 已提交
1831

1832
		for (idx = addr; idx < addr + nrpages; idx++) {
1833
			struct page *page = get_tmp_page(sbi, idx);
1834

1835 1836 1837 1838 1839 1840
			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);
1841
		}
1842

1843
		invalidate_mapping_pages(META_MAPPING(sbi), addr,
1844
							addr + nrpages);
J
Jaegeuk Kim 已提交
1845
	}
1846
	return 0;
J
Jaegeuk Kim 已提交
1847 1848
}

1849
static void remove_nats_in_journal(struct f2fs_sb_info *sbi)
J
Jaegeuk Kim 已提交
1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862
{
	struct f2fs_nm_info *nm_i = NM_I(sbi);
	struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_HOT_DATA);
	struct f2fs_summary_block *sum = curseg->sum_blk;
	int i;

	mutex_lock(&curseg->curseg_mutex);
	for (i = 0; i < nats_in_cursum(sum); i++) {
		struct nat_entry *ne;
		struct f2fs_nat_entry raw_ne;
		nid_t nid = le32_to_cpu(nid_in_journal(sum, i));

		raw_ne = nat_in_journal(sum, i);
1863

1864
		down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
1865 1866
		ne = __lookup_nat_cache(nm_i, nid);
		if (!ne) {
1867 1868
			ne = grab_nat_entry(nm_i, nid);
			node_info_from_raw_nat(&ne->ni, &raw_ne);
J
Jaegeuk Kim 已提交
1869 1870
		}
		__set_nat_cache_dirty(nm_i, ne);
1871
		up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
1872 1873 1874 1875 1876
	}
	update_nats_in_cursum(sum, -i);
	mutex_unlock(&curseg->curseg_mutex);
}

1877 1878
static void __adjust_nat_entry_set(struct nat_entry_set *nes,
						struct list_head *head, int max)
J
Jaegeuk Kim 已提交
1879
{
1880
	struct nat_entry_set *cur;
J
Jaegeuk Kim 已提交
1881

1882 1883
	if (nes->entry_cnt >= max)
		goto add_out;
J
Jaegeuk Kim 已提交
1884

1885 1886 1887 1888 1889
	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;
		}
1890
	}
1891 1892 1893
add_out:
	list_add_tail(&nes->set_list, head);
}
J
Jaegeuk Kim 已提交
1894

1895 1896 1897 1898 1899 1900 1901 1902 1903 1904
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);
	struct f2fs_summary_block *sum = curseg->sum_blk;
	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;
1905
	struct f2fs_nm_info *nm_i = NM_I(sbi);
J
Jaegeuk Kim 已提交
1906

1907 1908 1909 1910 1911
	/*
	 * 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.
	 */
1912 1913 1914 1915 1916 1917 1918 1919 1920 1921
	if (!__has_cursum_space(sum, set->entry_cnt, NAT_JOURNAL))
		to_journal = false;

	if (to_journal) {
		mutex_lock(&curseg->curseg_mutex);
	} else {
		page = get_next_nat_page(sbi, start_nid);
		nat_blk = page_address(page);
		f2fs_bug_on(sbi, !nat_blk);
	}
1922

1923 1924 1925 1926 1927 1928 1929 1930
	/* 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;
1931 1932

		if (to_journal) {
1933 1934 1935 1936 1937
			offset = lookup_journal_in_cursum(sum,
							NAT_JOURNAL, nid, 1);
			f2fs_bug_on(sbi, offset < 0);
			raw_ne = &nat_in_journal(sum, offset);
			nid_in_journal(sum, offset) = cpu_to_le32(nid);
1938
		} else {
1939
			raw_ne = &nat_blk->entries[nid - start_nid];
J
Jaegeuk Kim 已提交
1940
		}
1941
		raw_nat_from_node_info(raw_ne, &ne->ni);
J
Jaegeuk Kim 已提交
1942

1943
		down_write(&NM_I(sbi)->nat_tree_lock);
1944 1945
		nat_reset_flag(ne);
		__clear_nat_cache_dirty(NM_I(sbi), ne);
1946
		up_write(&NM_I(sbi)->nat_tree_lock);
1947

1948 1949 1950
		if (nat_get_blkaddr(ne) == NULL_ADDR)
			add_free_nid(sbi, nid, false);
	}
J
Jaegeuk Kim 已提交
1951

1952 1953 1954 1955
	if (to_journal)
		mutex_unlock(&curseg->curseg_mutex);
	else
		f2fs_put_page(page, 1);
1956

1957 1958
	f2fs_bug_on(sbi, set->entry_cnt);

1959
	down_write(&nm_i->nat_tree_lock);
1960
	radix_tree_delete(&NM_I(sbi)->nat_set_root, set->set);
1961
	up_write(&nm_i->nat_tree_lock);
1962
	kmem_cache_free(nat_entry_set_slab, set);
1963
}
1964

1965 1966 1967 1968 1969 1970 1971 1972
/*
 * 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);
	struct f2fs_summary_block *sum = curseg->sum_blk;
1973
	struct nat_entry_set *setvec[SETVEC_SIZE];
1974 1975 1976 1977 1978
	struct nat_entry_set *set, *tmp;
	unsigned int found;
	nid_t set_idx = 0;
	LIST_HEAD(sets);

1979 1980
	if (!nm_i->dirty_nat_cnt)
		return;
1981 1982 1983 1984 1985 1986 1987 1988
	/*
	 * 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.
	 */
	if (!__has_cursum_space(sum, nm_i->dirty_nat_cnt, NAT_JOURNAL))
		remove_nats_in_journal(sbi);

1989
	down_write(&nm_i->nat_tree_lock);
1990
	while ((found = __gang_lookup_nat_set(nm_i,
1991
					set_idx, SETVEC_SIZE, setvec))) {
1992 1993 1994 1995 1996
		unsigned idx;
		set_idx = setvec[found - 1]->set + 1;
		for (idx = 0; idx < found; idx++)
			__adjust_nat_entry_set(setvec[idx], &sets,
							MAX_NAT_JENTRIES(sum));
J
Jaegeuk Kim 已提交
1997
	}
1998
	up_write(&nm_i->nat_tree_lock);
1999

2000 2001 2002 2003
	/* flush dirty nats in nat entry set */
	list_for_each_entry_safe(set, tmp, &sets, set_list)
		__flush_nat_entry_set(sbi, set);

2004
	f2fs_bug_on(sbi, nm_i->dirty_nat_cnt);
J
Jaegeuk Kim 已提交
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
}

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);
2019

2020 2021
	nm_i->max_nid = NAT_ENTRY_PER_BLOCK * nat_blocks;

2022
	/* not used nids: 0, node, meta, (and root counted as valid node) */
2023
	nm_i->available_nids = nm_i->max_nid - F2FS_RESERVED_NODE_NUM;
J
Jaegeuk Kim 已提交
2024 2025
	nm_i->fcnt = 0;
	nm_i->nat_cnt = 0;
2026
	nm_i->ram_thresh = DEF_RAM_THRESHOLD;
C
Chao Yu 已提交
2027
	nm_i->ra_nid_pages = DEF_RA_NID_PAGES;
J
Jaegeuk Kim 已提交
2028

2029
	INIT_RADIX_TREE(&nm_i->free_nid_root, GFP_ATOMIC);
J
Jaegeuk Kim 已提交
2030
	INIT_LIST_HEAD(&nm_i->free_nid_list);
2031 2032
	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
2033 2034 2035 2036
	INIT_LIST_HEAD(&nm_i->nat_entries);

	mutex_init(&nm_i->build_lock);
	spin_lock_init(&nm_i->free_nid_list_lock);
2037
	init_rwsem(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2038 2039

	nm_i->next_scan_nid = le32_to_cpu(sbi->ckpt->next_free_nid);
A
Alexandru Gheorghiu 已提交
2040
	nm_i->bitmap_size = __bitmap_size(sbi, NAT_BITMAP);
J
Jaegeuk Kim 已提交
2041 2042 2043 2044
	version_bitmap = __bitmap_ptr(sbi, NAT_BITMAP);
	if (!version_bitmap)
		return -EFAULT;

A
Alexandru Gheorghiu 已提交
2045 2046 2047 2048
	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
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072
	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];
2073
	struct nat_entry_set *setvec[SETVEC_SIZE];
J
Jaegeuk Kim 已提交
2074 2075 2076 2077 2078 2079 2080 2081 2082
	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) {
2083
		f2fs_bug_on(sbi, i->state == NID_ALLOC);
2084
		__del_from_free_nid_list(nm_i, i);
J
Jaegeuk Kim 已提交
2085
		nm_i->fcnt--;
2086 2087 2088
		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
2089
	}
2090
	f2fs_bug_on(sbi, nm_i->fcnt);
J
Jaegeuk Kim 已提交
2091 2092 2093
	spin_unlock(&nm_i->free_nid_list_lock);

	/* destroy nat cache */
2094
	down_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2095 2096 2097
	while ((found = __gang_lookup_nat_cache(nm_i,
					nid, NATVEC_SIZE, natvec))) {
		unsigned idx;
2098

2099 2100 2101
		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
2102
	}
2103
	f2fs_bug_on(sbi, nm_i->nat_cnt);
2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118

	/* 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]);
		}
	}
2119
	up_write(&nm_i->nat_tree_lock);
J
Jaegeuk Kim 已提交
2120 2121 2122 2123 2124 2125

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

2126
int __init create_node_manager_caches(void)
J
Jaegeuk Kim 已提交
2127 2128
{
	nat_entry_slab = f2fs_kmem_cache_create("nat_entry",
2129
			sizeof(struct nat_entry));
J
Jaegeuk Kim 已提交
2130
	if (!nat_entry_slab)
2131
		goto fail;
J
Jaegeuk Kim 已提交
2132 2133

	free_nid_slab = f2fs_kmem_cache_create("free_nid",
2134
			sizeof(struct free_nid));
2135
	if (!free_nid_slab)
2136
		goto destroy_nat_entry;
2137 2138 2139 2140

	nat_entry_set_slab = f2fs_kmem_cache_create("nat_entry_set",
			sizeof(struct nat_entry_set));
	if (!nat_entry_set_slab)
2141
		goto destroy_free_nid;
J
Jaegeuk Kim 已提交
2142
	return 0;
2143

2144
destroy_free_nid:
2145
	kmem_cache_destroy(free_nid_slab);
2146
destroy_nat_entry:
2147 2148 2149
	kmem_cache_destroy(nat_entry_slab);
fail:
	return -ENOMEM;
J
Jaegeuk Kim 已提交
2150 2151 2152 2153
}

void destroy_node_manager_caches(void)
{
2154
	kmem_cache_destroy(nat_entry_set_slab);
J
Jaegeuk Kim 已提交
2155 2156 2157
	kmem_cache_destroy(free_nid_slab);
	kmem_cache_destroy(nat_entry_slab);
}