xref: /openbmc/linux/fs/f2fs/extent_cache.c (revision c7e1962a)
1 // SPDX-License-Identifier: GPL-2.0
2 /*
3  * f2fs extent cache support
4  *
5  * Copyright (c) 2015 Motorola Mobility
6  * Copyright (c) 2015 Samsung Electronics
7  * Authors: Jaegeuk Kim <jaegeuk@kernel.org>
8  *          Chao Yu <chao2.yu@samsung.com>
9  *
10  * block_age-based extent cache added by:
11  * Copyright (c) 2022 xiaomi Co., Ltd.
12  *             http://www.xiaomi.com/
13  */
14 
15 #include <linux/fs.h>
16 #include <linux/f2fs_fs.h>
17 
18 #include "f2fs.h"
19 #include "node.h"
20 #include <trace/events/f2fs.h>
21 
22 bool sanity_check_extent_cache(struct inode *inode, struct page *ipage)
23 {
24 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
25 	struct f2fs_extent *i_ext = &F2FS_INODE(ipage)->i_ext;
26 	struct extent_info ei;
27 
28 	get_read_extent_info(&ei, i_ext);
29 
30 	if (!ei.len)
31 		return true;
32 
33 	if (!f2fs_is_valid_blkaddr(sbi, ei.blk, DATA_GENERIC_ENHANCE) ||
34 	    !f2fs_is_valid_blkaddr(sbi, ei.blk + ei.len - 1,
35 					DATA_GENERIC_ENHANCE)) {
36 		set_sbi_flag(sbi, SBI_NEED_FSCK);
37 		f2fs_warn(sbi, "%s: inode (ino=%lx) extent info [%u, %u, %u] is incorrect, run fsck to fix",
38 			  __func__, inode->i_ino,
39 			  ei.blk, ei.fofs, ei.len);
40 		return false;
41 	}
42 	return true;
43 }
44 
45 static void __set_extent_info(struct extent_info *ei,
46 				unsigned int fofs, unsigned int len,
47 				block_t blk, bool keep_clen,
48 				unsigned long age, unsigned long last_blocks,
49 				enum extent_type type)
50 {
51 	ei->fofs = fofs;
52 	ei->len = len;
53 
54 	if (type == EX_READ) {
55 		ei->blk = blk;
56 		if (keep_clen)
57 			return;
58 #ifdef CONFIG_F2FS_FS_COMPRESSION
59 		ei->c_len = 0;
60 #endif
61 	} else if (type == EX_BLOCK_AGE) {
62 		ei->age = age;
63 		ei->last_blocks = last_blocks;
64 	}
65 }
66 
67 static bool __init_may_extent_tree(struct inode *inode, enum extent_type type)
68 {
69 	if (type == EX_READ)
70 		return test_opt(F2FS_I_SB(inode), READ_EXTENT_CACHE) &&
71 			S_ISREG(inode->i_mode);
72 	if (type == EX_BLOCK_AGE)
73 		return test_opt(F2FS_I_SB(inode), AGE_EXTENT_CACHE) &&
74 			(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode));
75 	return false;
76 }
77 
78 static bool __may_extent_tree(struct inode *inode, enum extent_type type)
79 {
80 	/*
81 	 * for recovered files during mount do not create extents
82 	 * if shrinker is not registered.
83 	 */
84 	if (list_empty(&F2FS_I_SB(inode)->s_list))
85 		return false;
86 
87 	if (!__init_may_extent_tree(inode, type))
88 		return false;
89 
90 	if (type == EX_READ) {
91 		if (is_inode_flag_set(inode, FI_NO_EXTENT))
92 			return false;
93 		if (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
94 				 !f2fs_sb_has_readonly(F2FS_I_SB(inode)))
95 			return false;
96 	} else if (type == EX_BLOCK_AGE) {
97 		if (is_inode_flag_set(inode, FI_COMPRESSED_FILE))
98 			return false;
99 		if (file_is_cold(inode))
100 			return false;
101 	}
102 	return true;
103 }
104 
105 static void __try_update_largest_extent(struct extent_tree *et,
106 						struct extent_node *en)
107 {
108 	if (et->type != EX_READ)
109 		return;
110 	if (en->ei.len <= et->largest.len)
111 		return;
112 
113 	et->largest = en->ei;
114 	et->largest_updated = true;
115 }
116 
117 static bool __is_extent_mergeable(struct extent_info *back,
118 		struct extent_info *front, enum extent_type type)
119 {
120 	if (type == EX_READ) {
121 #ifdef CONFIG_F2FS_FS_COMPRESSION
122 		if (back->c_len && back->len != back->c_len)
123 			return false;
124 		if (front->c_len && front->len != front->c_len)
125 			return false;
126 #endif
127 		return (back->fofs + back->len == front->fofs &&
128 				back->blk + back->len == front->blk);
129 	} else if (type == EX_BLOCK_AGE) {
130 		return (back->fofs + back->len == front->fofs &&
131 			abs(back->age - front->age) <= SAME_AGE_REGION &&
132 			abs(back->last_blocks - front->last_blocks) <=
133 							SAME_AGE_REGION);
134 	}
135 	return false;
136 }
137 
138 static bool __is_back_mergeable(struct extent_info *cur,
139 		struct extent_info *back, enum extent_type type)
140 {
141 	return __is_extent_mergeable(back, cur, type);
142 }
143 
144 static bool __is_front_mergeable(struct extent_info *cur,
145 		struct extent_info *front, enum extent_type type)
146 {
147 	return __is_extent_mergeable(cur, front, type);
148 }
149 
150 static struct extent_node *__lookup_extent_node(struct rb_root_cached *root,
151 			struct extent_node *cached_en, unsigned int fofs)
152 {
153 	struct rb_node *node = root->rb_root.rb_node;
154 	struct extent_node *en;
155 
156 	/* check a cached entry */
157 	if (cached_en && cached_en->ei.fofs <= fofs &&
158 			cached_en->ei.fofs + cached_en->ei.len > fofs)
159 		return cached_en;
160 
161 	/* check rb_tree */
162 	while (node) {
163 		en = rb_entry(node, struct extent_node, rb_node);
164 
165 		if (fofs < en->ei.fofs)
166 			node = node->rb_left;
167 		else if (fofs >= en->ei.fofs + en->ei.len)
168 			node = node->rb_right;
169 		else
170 			return en;
171 	}
172 	return NULL;
173 }
174 
175 /*
176  * lookup rb entry in position of @fofs in rb-tree,
177  * if hit, return the entry, otherwise, return NULL
178  * @prev_ex: extent before fofs
179  * @next_ex: extent after fofs
180  * @insert_p: insert point for new extent at fofs
181  * in order to simplify the insertion after.
182  * tree must stay unchanged between lookup and insertion.
183  */
184 static struct extent_node *__lookup_extent_node_ret(struct rb_root_cached *root,
185 				struct extent_node *cached_en,
186 				unsigned int fofs,
187 				struct extent_node **prev_entry,
188 				struct extent_node **next_entry,
189 				struct rb_node ***insert_p,
190 				struct rb_node **insert_parent,
191 				bool *leftmost)
192 {
193 	struct rb_node **pnode = &root->rb_root.rb_node;
194 	struct rb_node *parent = NULL, *tmp_node;
195 	struct extent_node *en = cached_en;
196 
197 	*insert_p = NULL;
198 	*insert_parent = NULL;
199 	*prev_entry = NULL;
200 	*next_entry = NULL;
201 
202 	if (RB_EMPTY_ROOT(&root->rb_root))
203 		return NULL;
204 
205 	if (en && en->ei.fofs <= fofs && en->ei.fofs + en->ei.len > fofs)
206 		goto lookup_neighbors;
207 
208 	*leftmost = true;
209 
210 	while (*pnode) {
211 		parent = *pnode;
212 		en = rb_entry(*pnode, struct extent_node, rb_node);
213 
214 		if (fofs < en->ei.fofs) {
215 			pnode = &(*pnode)->rb_left;
216 		} else if (fofs >= en->ei.fofs + en->ei.len) {
217 			pnode = &(*pnode)->rb_right;
218 			*leftmost = false;
219 		} else {
220 			goto lookup_neighbors;
221 		}
222 	}
223 
224 	*insert_p = pnode;
225 	*insert_parent = parent;
226 
227 	en = rb_entry(parent, struct extent_node, rb_node);
228 	tmp_node = parent;
229 	if (parent && fofs > en->ei.fofs)
230 		tmp_node = rb_next(parent);
231 	*next_entry = rb_entry_safe(tmp_node, struct extent_node, rb_node);
232 
233 	tmp_node = parent;
234 	if (parent && fofs < en->ei.fofs)
235 		tmp_node = rb_prev(parent);
236 	*prev_entry = rb_entry_safe(tmp_node, struct extent_node, rb_node);
237 	return NULL;
238 
239 lookup_neighbors:
240 	if (fofs == en->ei.fofs) {
241 		/* lookup prev node for merging backward later */
242 		tmp_node = rb_prev(&en->rb_node);
243 		*prev_entry = rb_entry_safe(tmp_node,
244 					struct extent_node, rb_node);
245 	}
246 	if (fofs == en->ei.fofs + en->ei.len - 1) {
247 		/* lookup next node for merging frontward later */
248 		tmp_node = rb_next(&en->rb_node);
249 		*next_entry = rb_entry_safe(tmp_node,
250 					struct extent_node, rb_node);
251 	}
252 	return en;
253 }
254 
255 static struct kmem_cache *extent_tree_slab;
256 static struct kmem_cache *extent_node_slab;
257 
258 static struct extent_node *__attach_extent_node(struct f2fs_sb_info *sbi,
259 				struct extent_tree *et, struct extent_info *ei,
260 				struct rb_node *parent, struct rb_node **p,
261 				bool leftmost)
262 {
263 	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
264 	struct extent_node *en;
265 
266 	en = f2fs_kmem_cache_alloc(extent_node_slab, GFP_ATOMIC, false, sbi);
267 	if (!en)
268 		return NULL;
269 
270 	en->ei = *ei;
271 	INIT_LIST_HEAD(&en->list);
272 	en->et = et;
273 
274 	rb_link_node(&en->rb_node, parent, p);
275 	rb_insert_color_cached(&en->rb_node, &et->root, leftmost);
276 	atomic_inc(&et->node_cnt);
277 	atomic_inc(&eti->total_ext_node);
278 	return en;
279 }
280 
281 static void __detach_extent_node(struct f2fs_sb_info *sbi,
282 				struct extent_tree *et, struct extent_node *en)
283 {
284 	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
285 
286 	rb_erase_cached(&en->rb_node, &et->root);
287 	atomic_dec(&et->node_cnt);
288 	atomic_dec(&eti->total_ext_node);
289 
290 	if (et->cached_en == en)
291 		et->cached_en = NULL;
292 	kmem_cache_free(extent_node_slab, en);
293 }
294 
295 /*
296  * Flow to release an extent_node:
297  * 1. list_del_init
298  * 2. __detach_extent_node
299  * 3. kmem_cache_free.
300  */
301 static void __release_extent_node(struct f2fs_sb_info *sbi,
302 			struct extent_tree *et, struct extent_node *en)
303 {
304 	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
305 
306 	spin_lock(&eti->extent_lock);
307 	f2fs_bug_on(sbi, list_empty(&en->list));
308 	list_del_init(&en->list);
309 	spin_unlock(&eti->extent_lock);
310 
311 	__detach_extent_node(sbi, et, en);
312 }
313 
314 static struct extent_tree *__grab_extent_tree(struct inode *inode,
315 						enum extent_type type)
316 {
317 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
318 	struct extent_tree_info *eti = &sbi->extent_tree[type];
319 	struct extent_tree *et;
320 	nid_t ino = inode->i_ino;
321 
322 	mutex_lock(&eti->extent_tree_lock);
323 	et = radix_tree_lookup(&eti->extent_tree_root, ino);
324 	if (!et) {
325 		et = f2fs_kmem_cache_alloc(extent_tree_slab,
326 					GFP_NOFS, true, NULL);
327 		f2fs_radix_tree_insert(&eti->extent_tree_root, ino, et);
328 		memset(et, 0, sizeof(struct extent_tree));
329 		et->ino = ino;
330 		et->type = type;
331 		et->root = RB_ROOT_CACHED;
332 		et->cached_en = NULL;
333 		rwlock_init(&et->lock);
334 		INIT_LIST_HEAD(&et->list);
335 		atomic_set(&et->node_cnt, 0);
336 		atomic_inc(&eti->total_ext_tree);
337 	} else {
338 		atomic_dec(&eti->total_zombie_tree);
339 		list_del_init(&et->list);
340 	}
341 	mutex_unlock(&eti->extent_tree_lock);
342 
343 	/* never died until evict_inode */
344 	F2FS_I(inode)->extent_tree[type] = et;
345 
346 	return et;
347 }
348 
349 static unsigned int __free_extent_tree(struct f2fs_sb_info *sbi,
350 				struct extent_tree *et, unsigned int nr_shrink)
351 {
352 	struct rb_node *node, *next;
353 	struct extent_node *en;
354 	unsigned int count;
355 
356 	node = rb_first_cached(&et->root);
357 
358 	for (count = 0; node && count < nr_shrink; count++) {
359 		next = rb_next(node);
360 		en = rb_entry(node, struct extent_node, rb_node);
361 		__release_extent_node(sbi, et, en);
362 		node = next;
363 	}
364 
365 	return count;
366 }
367 
368 static void __drop_largest_extent(struct extent_tree *et,
369 					pgoff_t fofs, unsigned int len)
370 {
371 	if (fofs < (pgoff_t)et->largest.fofs + et->largest.len &&
372 			fofs + len > et->largest.fofs) {
373 		et->largest.len = 0;
374 		et->largest_updated = true;
375 	}
376 }
377 
378 void f2fs_init_read_extent_tree(struct inode *inode, struct page *ipage)
379 {
380 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
381 	struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
382 	struct f2fs_extent *i_ext = &F2FS_INODE(ipage)->i_ext;
383 	struct extent_tree *et;
384 	struct extent_node *en;
385 	struct extent_info ei;
386 
387 	if (!__may_extent_tree(inode, EX_READ)) {
388 		/* drop largest read extent */
389 		if (i_ext->len) {
390 			f2fs_wait_on_page_writeback(ipage, NODE, true, true);
391 			i_ext->len = 0;
392 			set_page_dirty(ipage);
393 		}
394 		set_inode_flag(inode, FI_NO_EXTENT);
395 		return;
396 	}
397 
398 	et = __grab_extent_tree(inode, EX_READ);
399 
400 	get_read_extent_info(&ei, i_ext);
401 
402 	write_lock(&et->lock);
403 	if (atomic_read(&et->node_cnt) || !ei.len)
404 		goto skip;
405 
406 	en = __attach_extent_node(sbi, et, &ei, NULL,
407 				&et->root.rb_root.rb_node, true);
408 	if (en) {
409 		et->largest = en->ei;
410 		et->cached_en = en;
411 
412 		spin_lock(&eti->extent_lock);
413 		list_add_tail(&en->list, &eti->extent_list);
414 		spin_unlock(&eti->extent_lock);
415 	}
416 skip:
417 	/* Let's drop, if checkpoint got corrupted. */
418 	if (f2fs_cp_error(sbi)) {
419 		et->largest.len = 0;
420 		et->largest_updated = true;
421 	}
422 	write_unlock(&et->lock);
423 }
424 
425 void f2fs_init_age_extent_tree(struct inode *inode)
426 {
427 	if (!__init_may_extent_tree(inode, EX_BLOCK_AGE))
428 		return;
429 	__grab_extent_tree(inode, EX_BLOCK_AGE);
430 }
431 
432 void f2fs_init_extent_tree(struct inode *inode)
433 {
434 	/* initialize read cache */
435 	if (__init_may_extent_tree(inode, EX_READ))
436 		__grab_extent_tree(inode, EX_READ);
437 
438 	/* initialize block age cache */
439 	if (__init_may_extent_tree(inode, EX_BLOCK_AGE))
440 		__grab_extent_tree(inode, EX_BLOCK_AGE);
441 }
442 
443 static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
444 			struct extent_info *ei, enum extent_type type)
445 {
446 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
447 	struct extent_tree_info *eti = &sbi->extent_tree[type];
448 	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
449 	struct extent_node *en;
450 	bool ret = false;
451 
452 	if (!et)
453 		return false;
454 
455 	trace_f2fs_lookup_extent_tree_start(inode, pgofs, type);
456 
457 	read_lock(&et->lock);
458 
459 	if (type == EX_READ &&
460 			et->largest.fofs <= pgofs &&
461 			(pgoff_t)et->largest.fofs + et->largest.len > pgofs) {
462 		*ei = et->largest;
463 		ret = true;
464 		stat_inc_largest_node_hit(sbi);
465 		goto out;
466 	}
467 
468 	en = __lookup_extent_node(&et->root, et->cached_en, pgofs);
469 	if (!en)
470 		goto out;
471 
472 	if (en == et->cached_en)
473 		stat_inc_cached_node_hit(sbi, type);
474 	else
475 		stat_inc_rbtree_node_hit(sbi, type);
476 
477 	*ei = en->ei;
478 	spin_lock(&eti->extent_lock);
479 	if (!list_empty(&en->list)) {
480 		list_move_tail(&en->list, &eti->extent_list);
481 		et->cached_en = en;
482 	}
483 	spin_unlock(&eti->extent_lock);
484 	ret = true;
485 out:
486 	stat_inc_total_hit(sbi, type);
487 	read_unlock(&et->lock);
488 
489 	if (type == EX_READ)
490 		trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
491 	else if (type == EX_BLOCK_AGE)
492 		trace_f2fs_lookup_age_extent_tree_end(inode, pgofs, ei);
493 	return ret;
494 }
495 
496 static struct extent_node *__try_merge_extent_node(struct f2fs_sb_info *sbi,
497 				struct extent_tree *et, struct extent_info *ei,
498 				struct extent_node *prev_ex,
499 				struct extent_node *next_ex)
500 {
501 	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
502 	struct extent_node *en = NULL;
503 
504 	if (prev_ex && __is_back_mergeable(ei, &prev_ex->ei, et->type)) {
505 		prev_ex->ei.len += ei->len;
506 		ei = &prev_ex->ei;
507 		en = prev_ex;
508 	}
509 
510 	if (next_ex && __is_front_mergeable(ei, &next_ex->ei, et->type)) {
511 		next_ex->ei.fofs = ei->fofs;
512 		next_ex->ei.len += ei->len;
513 		if (et->type == EX_READ)
514 			next_ex->ei.blk = ei->blk;
515 		if (en)
516 			__release_extent_node(sbi, et, prev_ex);
517 
518 		en = next_ex;
519 	}
520 
521 	if (!en)
522 		return NULL;
523 
524 	__try_update_largest_extent(et, en);
525 
526 	spin_lock(&eti->extent_lock);
527 	if (!list_empty(&en->list)) {
528 		list_move_tail(&en->list, &eti->extent_list);
529 		et->cached_en = en;
530 	}
531 	spin_unlock(&eti->extent_lock);
532 	return en;
533 }
534 
535 static struct extent_node *__insert_extent_tree(struct f2fs_sb_info *sbi,
536 				struct extent_tree *et, struct extent_info *ei,
537 				struct rb_node **insert_p,
538 				struct rb_node *insert_parent,
539 				bool leftmost)
540 {
541 	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
542 	struct rb_node **p = &et->root.rb_root.rb_node;
543 	struct rb_node *parent = NULL;
544 	struct extent_node *en = NULL;
545 
546 	if (insert_p && insert_parent) {
547 		parent = insert_parent;
548 		p = insert_p;
549 		goto do_insert;
550 	}
551 
552 	leftmost = true;
553 
554 	/* look up extent_node in the rb tree */
555 	while (*p) {
556 		parent = *p;
557 		en = rb_entry(parent, struct extent_node, rb_node);
558 
559 		if (ei->fofs < en->ei.fofs) {
560 			p = &(*p)->rb_left;
561 		} else if (ei->fofs >= en->ei.fofs + en->ei.len) {
562 			p = &(*p)->rb_right;
563 			leftmost = false;
564 		} else {
565 			f2fs_bug_on(sbi, 1);
566 		}
567 	}
568 
569 do_insert:
570 	en = __attach_extent_node(sbi, et, ei, parent, p, leftmost);
571 	if (!en)
572 		return NULL;
573 
574 	__try_update_largest_extent(et, en);
575 
576 	/* update in global extent list */
577 	spin_lock(&eti->extent_lock);
578 	list_add_tail(&en->list, &eti->extent_list);
579 	et->cached_en = en;
580 	spin_unlock(&eti->extent_lock);
581 	return en;
582 }
583 
584 static unsigned int __destroy_extent_node(struct inode *inode,
585 					enum extent_type type)
586 {
587 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
588 	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
589 	unsigned int nr_shrink = type == EX_READ ?
590 				READ_EXTENT_CACHE_SHRINK_NUMBER :
591 				AGE_EXTENT_CACHE_SHRINK_NUMBER;
592 	unsigned int node_cnt = 0;
593 
594 	if (!et || !atomic_read(&et->node_cnt))
595 		return 0;
596 
597 	while (atomic_read(&et->node_cnt)) {
598 		write_lock(&et->lock);
599 		node_cnt += __free_extent_tree(sbi, et, nr_shrink);
600 		write_unlock(&et->lock);
601 	}
602 
603 	f2fs_bug_on(sbi, atomic_read(&et->node_cnt));
604 
605 	return node_cnt;
606 }
607 
608 static void __update_extent_tree_range(struct inode *inode,
609 			struct extent_info *tei, enum extent_type type)
610 {
611 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
612 	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
613 	struct extent_node *en = NULL, *en1 = NULL;
614 	struct extent_node *prev_en = NULL, *next_en = NULL;
615 	struct extent_info ei, dei, prev;
616 	struct rb_node **insert_p = NULL, *insert_parent = NULL;
617 	unsigned int fofs = tei->fofs, len = tei->len;
618 	unsigned int end = fofs + len;
619 	bool updated = false;
620 	bool leftmost = false;
621 
622 	if (!et)
623 		return;
624 
625 	if (type == EX_READ)
626 		trace_f2fs_update_read_extent_tree_range(inode, fofs, len,
627 						tei->blk, 0);
628 	else if (type == EX_BLOCK_AGE)
629 		trace_f2fs_update_age_extent_tree_range(inode, fofs, len,
630 						tei->age, tei->last_blocks);
631 
632 	write_lock(&et->lock);
633 
634 	if (type == EX_READ) {
635 		if (is_inode_flag_set(inode, FI_NO_EXTENT)) {
636 			write_unlock(&et->lock);
637 			return;
638 		}
639 
640 		prev = et->largest;
641 		dei.len = 0;
642 
643 		/*
644 		 * drop largest extent before lookup, in case it's already
645 		 * been shrunk from extent tree
646 		 */
647 		__drop_largest_extent(et, fofs, len);
648 	}
649 
650 	/* 1. lookup first extent node in range [fofs, fofs + len - 1] */
651 	en = __lookup_extent_node_ret(&et->root,
652 					et->cached_en, fofs,
653 					&prev_en, &next_en,
654 					&insert_p, &insert_parent,
655 					&leftmost);
656 	if (!en)
657 		en = next_en;
658 
659 	/* 2. invalidate all extent nodes in range [fofs, fofs + len - 1] */
660 	while (en && en->ei.fofs < end) {
661 		unsigned int org_end;
662 		int parts = 0;	/* # of parts current extent split into */
663 
664 		next_en = en1 = NULL;
665 
666 		dei = en->ei;
667 		org_end = dei.fofs + dei.len;
668 		f2fs_bug_on(sbi, fofs >= org_end);
669 
670 		if (fofs > dei.fofs && (type != EX_READ ||
671 				fofs - dei.fofs >= F2FS_MIN_EXTENT_LEN)) {
672 			en->ei.len = fofs - en->ei.fofs;
673 			prev_en = en;
674 			parts = 1;
675 		}
676 
677 		if (end < org_end && (type != EX_READ ||
678 				org_end - end >= F2FS_MIN_EXTENT_LEN)) {
679 			if (parts) {
680 				__set_extent_info(&ei,
681 					end, org_end - end,
682 					end - dei.fofs + dei.blk, false,
683 					dei.age, dei.last_blocks,
684 					type);
685 				en1 = __insert_extent_tree(sbi, et, &ei,
686 							NULL, NULL, true);
687 				next_en = en1;
688 			} else {
689 				__set_extent_info(&en->ei,
690 					end, en->ei.len - (end - dei.fofs),
691 					en->ei.blk + (end - dei.fofs), true,
692 					dei.age, dei.last_blocks,
693 					type);
694 				next_en = en;
695 			}
696 			parts++;
697 		}
698 
699 		if (!next_en) {
700 			struct rb_node *node = rb_next(&en->rb_node);
701 
702 			next_en = rb_entry_safe(node, struct extent_node,
703 						rb_node);
704 		}
705 
706 		if (parts)
707 			__try_update_largest_extent(et, en);
708 		else
709 			__release_extent_node(sbi, et, en);
710 
711 		/*
712 		 * if original extent is split into zero or two parts, extent
713 		 * tree has been altered by deletion or insertion, therefore
714 		 * invalidate pointers regard to tree.
715 		 */
716 		if (parts != 1) {
717 			insert_p = NULL;
718 			insert_parent = NULL;
719 		}
720 		en = next_en;
721 	}
722 
723 	if (type == EX_BLOCK_AGE)
724 		goto update_age_extent_cache;
725 
726 	/* 3. update extent in read extent cache */
727 	BUG_ON(type != EX_READ);
728 
729 	if (tei->blk) {
730 		__set_extent_info(&ei, fofs, len, tei->blk, false,
731 				  0, 0, EX_READ);
732 		if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
733 			__insert_extent_tree(sbi, et, &ei,
734 					insert_p, insert_parent, leftmost);
735 
736 		/* give up extent_cache, if split and small updates happen */
737 		if (dei.len >= 1 &&
738 				prev.len < F2FS_MIN_EXTENT_LEN &&
739 				et->largest.len < F2FS_MIN_EXTENT_LEN) {
740 			et->largest.len = 0;
741 			et->largest_updated = true;
742 			set_inode_flag(inode, FI_NO_EXTENT);
743 		}
744 	}
745 
746 	if (et->largest_updated) {
747 		et->largest_updated = false;
748 		updated = true;
749 	}
750 	goto out_read_extent_cache;
751 update_age_extent_cache:
752 	if (!tei->last_blocks)
753 		goto out_read_extent_cache;
754 
755 	__set_extent_info(&ei, fofs, len, 0, false,
756 			tei->age, tei->last_blocks, EX_BLOCK_AGE);
757 	if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
758 		__insert_extent_tree(sbi, et, &ei,
759 					insert_p, insert_parent, leftmost);
760 out_read_extent_cache:
761 	write_unlock(&et->lock);
762 
763 	if (is_inode_flag_set(inode, FI_NO_EXTENT))
764 		__destroy_extent_node(inode, EX_READ);
765 
766 	if (updated)
767 		f2fs_mark_inode_dirty_sync(inode, true);
768 }
769 
770 #ifdef CONFIG_F2FS_FS_COMPRESSION
771 void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
772 				pgoff_t fofs, block_t blkaddr, unsigned int llen,
773 				unsigned int c_len)
774 {
775 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
776 	struct extent_tree *et = F2FS_I(inode)->extent_tree[EX_READ];
777 	struct extent_node *en = NULL;
778 	struct extent_node *prev_en = NULL, *next_en = NULL;
779 	struct extent_info ei;
780 	struct rb_node **insert_p = NULL, *insert_parent = NULL;
781 	bool leftmost = false;
782 
783 	trace_f2fs_update_read_extent_tree_range(inode, fofs, llen,
784 						blkaddr, c_len);
785 
786 	/* it is safe here to check FI_NO_EXTENT w/o et->lock in ro image */
787 	if (is_inode_flag_set(inode, FI_NO_EXTENT))
788 		return;
789 
790 	write_lock(&et->lock);
791 
792 	en = __lookup_extent_node_ret(&et->root,
793 					et->cached_en, fofs,
794 					&prev_en, &next_en,
795 					&insert_p, &insert_parent,
796 					&leftmost);
797 	if (en)
798 		goto unlock_out;
799 
800 	__set_extent_info(&ei, fofs, llen, blkaddr, true, 0, 0, EX_READ);
801 	ei.c_len = c_len;
802 
803 	if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
804 		__insert_extent_tree(sbi, et, &ei,
805 				insert_p, insert_parent, leftmost);
806 unlock_out:
807 	write_unlock(&et->lock);
808 }
809 #endif
810 
811 static unsigned long long __calculate_block_age(struct f2fs_sb_info *sbi,
812 						unsigned long long new,
813 						unsigned long long old)
814 {
815 	unsigned int rem_old, rem_new;
816 	unsigned long long res;
817 	unsigned int weight = sbi->last_age_weight;
818 
819 	res = div_u64_rem(new, 100, &rem_new) * (100 - weight)
820 		+ div_u64_rem(old, 100, &rem_old) * weight;
821 
822 	if (rem_new)
823 		res += rem_new * (100 - weight) / 100;
824 	if (rem_old)
825 		res += rem_old * weight / 100;
826 
827 	return res;
828 }
829 
830 /* This returns a new age and allocated blocks in ei */
831 static int __get_new_block_age(struct inode *inode, struct extent_info *ei,
832 						block_t blkaddr)
833 {
834 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
835 	loff_t f_size = i_size_read(inode);
836 	unsigned long long cur_blocks =
837 				atomic64_read(&sbi->allocated_data_blocks);
838 	struct extent_info tei = *ei;	/* only fofs and len are valid */
839 
840 	/*
841 	 * When I/O is not aligned to a PAGE_SIZE, update will happen to the last
842 	 * file block even in seq write. So don't record age for newly last file
843 	 * block here.
844 	 */
845 	if ((f_size >> PAGE_SHIFT) == ei->fofs && f_size & (PAGE_SIZE - 1) &&
846 			blkaddr == NEW_ADDR)
847 		return -EINVAL;
848 
849 	if (__lookup_extent_tree(inode, ei->fofs, &tei, EX_BLOCK_AGE)) {
850 		unsigned long long cur_age;
851 
852 		if (cur_blocks >= tei.last_blocks)
853 			cur_age = cur_blocks - tei.last_blocks;
854 		else
855 			/* allocated_data_blocks overflow */
856 			cur_age = ULLONG_MAX - tei.last_blocks + cur_blocks;
857 
858 		if (tei.age)
859 			ei->age = __calculate_block_age(sbi, cur_age, tei.age);
860 		else
861 			ei->age = cur_age;
862 		ei->last_blocks = cur_blocks;
863 		WARN_ON(ei->age > cur_blocks);
864 		return 0;
865 	}
866 
867 	f2fs_bug_on(sbi, blkaddr == NULL_ADDR);
868 
869 	/* the data block was allocated for the first time */
870 	if (blkaddr == NEW_ADDR)
871 		goto out;
872 
873 	if (__is_valid_data_blkaddr(blkaddr) &&
874 	    !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) {
875 		f2fs_bug_on(sbi, 1);
876 		return -EINVAL;
877 	}
878 out:
879 	/*
880 	 * init block age with zero, this can happen when the block age extent
881 	 * was reclaimed due to memory constraint or system reboot
882 	 */
883 	ei->age = 0;
884 	ei->last_blocks = cur_blocks;
885 	return 0;
886 }
887 
888 static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
889 {
890 	struct extent_info ei = {};
891 
892 	if (!__may_extent_tree(dn->inode, type))
893 		return;
894 
895 	ei.fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_page), dn->inode) +
896 								dn->ofs_in_node;
897 	ei.len = 1;
898 
899 	if (type == EX_READ) {
900 		if (dn->data_blkaddr == NEW_ADDR)
901 			ei.blk = NULL_ADDR;
902 		else
903 			ei.blk = dn->data_blkaddr;
904 	} else if (type == EX_BLOCK_AGE) {
905 		if (__get_new_block_age(dn->inode, &ei, dn->data_blkaddr))
906 			return;
907 	}
908 	__update_extent_tree_range(dn->inode, &ei, type);
909 }
910 
911 static unsigned int __shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink,
912 					enum extent_type type)
913 {
914 	struct extent_tree_info *eti = &sbi->extent_tree[type];
915 	struct extent_tree *et, *next;
916 	struct extent_node *en;
917 	unsigned int node_cnt = 0, tree_cnt = 0;
918 	int remained;
919 
920 	if (!atomic_read(&eti->total_zombie_tree))
921 		goto free_node;
922 
923 	if (!mutex_trylock(&eti->extent_tree_lock))
924 		goto out;
925 
926 	/* 1. remove unreferenced extent tree */
927 	list_for_each_entry_safe(et, next, &eti->zombie_list, list) {
928 		if (atomic_read(&et->node_cnt)) {
929 			write_lock(&et->lock);
930 			node_cnt += __free_extent_tree(sbi, et,
931 					nr_shrink - node_cnt - tree_cnt);
932 			write_unlock(&et->lock);
933 		}
934 
935 		if (atomic_read(&et->node_cnt))
936 			goto unlock_out;
937 
938 		list_del_init(&et->list);
939 		radix_tree_delete(&eti->extent_tree_root, et->ino);
940 		kmem_cache_free(extent_tree_slab, et);
941 		atomic_dec(&eti->total_ext_tree);
942 		atomic_dec(&eti->total_zombie_tree);
943 		tree_cnt++;
944 
945 		if (node_cnt + tree_cnt >= nr_shrink)
946 			goto unlock_out;
947 		cond_resched();
948 	}
949 	mutex_unlock(&eti->extent_tree_lock);
950 
951 free_node:
952 	/* 2. remove LRU extent entries */
953 	if (!mutex_trylock(&eti->extent_tree_lock))
954 		goto out;
955 
956 	remained = nr_shrink - (node_cnt + tree_cnt);
957 
958 	spin_lock(&eti->extent_lock);
959 	for (; remained > 0; remained--) {
960 		if (list_empty(&eti->extent_list))
961 			break;
962 		en = list_first_entry(&eti->extent_list,
963 					struct extent_node, list);
964 		et = en->et;
965 		if (!write_trylock(&et->lock)) {
966 			/* refresh this extent node's position in extent list */
967 			list_move_tail(&en->list, &eti->extent_list);
968 			continue;
969 		}
970 
971 		list_del_init(&en->list);
972 		spin_unlock(&eti->extent_lock);
973 
974 		__detach_extent_node(sbi, et, en);
975 
976 		write_unlock(&et->lock);
977 		node_cnt++;
978 		spin_lock(&eti->extent_lock);
979 	}
980 	spin_unlock(&eti->extent_lock);
981 
982 unlock_out:
983 	mutex_unlock(&eti->extent_tree_lock);
984 out:
985 	trace_f2fs_shrink_extent_tree(sbi, node_cnt, tree_cnt, type);
986 
987 	return node_cnt + tree_cnt;
988 }
989 
990 /* read extent cache operations */
991 bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
992 				struct extent_info *ei)
993 {
994 	if (!__may_extent_tree(inode, EX_READ))
995 		return false;
996 
997 	return __lookup_extent_tree(inode, pgofs, ei, EX_READ);
998 }
999 
1000 bool f2fs_lookup_read_extent_cache_block(struct inode *inode, pgoff_t index,
1001 				block_t *blkaddr)
1002 {
1003 	struct extent_info ei = {};
1004 
1005 	if (!f2fs_lookup_read_extent_cache(inode, index, &ei))
1006 		return false;
1007 	*blkaddr = ei.blk + index - ei.fofs;
1008 	return true;
1009 }
1010 
1011 void f2fs_update_read_extent_cache(struct dnode_of_data *dn)
1012 {
1013 	return __update_extent_cache(dn, EX_READ);
1014 }
1015 
1016 void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
1017 				pgoff_t fofs, block_t blkaddr, unsigned int len)
1018 {
1019 	struct extent_info ei = {
1020 		.fofs = fofs,
1021 		.len = len,
1022 		.blk = blkaddr,
1023 	};
1024 
1025 	if (!__may_extent_tree(dn->inode, EX_READ))
1026 		return;
1027 
1028 	__update_extent_tree_range(dn->inode, &ei, EX_READ);
1029 }
1030 
1031 unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
1032 {
1033 	if (!test_opt(sbi, READ_EXTENT_CACHE))
1034 		return 0;
1035 
1036 	return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
1037 }
1038 
1039 /* block age extent cache operations */
1040 bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
1041 				struct extent_info *ei)
1042 {
1043 	if (!__may_extent_tree(inode, EX_BLOCK_AGE))
1044 		return false;
1045 
1046 	return __lookup_extent_tree(inode, pgofs, ei, EX_BLOCK_AGE);
1047 }
1048 
1049 void f2fs_update_age_extent_cache(struct dnode_of_data *dn)
1050 {
1051 	return __update_extent_cache(dn, EX_BLOCK_AGE);
1052 }
1053 
1054 void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
1055 				pgoff_t fofs, unsigned int len)
1056 {
1057 	struct extent_info ei = {
1058 		.fofs = fofs,
1059 		.len = len,
1060 	};
1061 
1062 	if (!__may_extent_tree(dn->inode, EX_BLOCK_AGE))
1063 		return;
1064 
1065 	__update_extent_tree_range(dn->inode, &ei, EX_BLOCK_AGE);
1066 }
1067 
1068 unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
1069 {
1070 	if (!test_opt(sbi, AGE_EXTENT_CACHE))
1071 		return 0;
1072 
1073 	return __shrink_extent_tree(sbi, nr_shrink, EX_BLOCK_AGE);
1074 }
1075 
1076 void f2fs_destroy_extent_node(struct inode *inode)
1077 {
1078 	__destroy_extent_node(inode, EX_READ);
1079 	__destroy_extent_node(inode, EX_BLOCK_AGE);
1080 }
1081 
1082 static void __drop_extent_tree(struct inode *inode, enum extent_type type)
1083 {
1084 	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
1085 	bool updated = false;
1086 
1087 	if (!__may_extent_tree(inode, type))
1088 		return;
1089 
1090 	write_lock(&et->lock);
1091 	if (type == EX_READ) {
1092 		set_inode_flag(inode, FI_NO_EXTENT);
1093 		if (et->largest.len) {
1094 			et->largest.len = 0;
1095 			updated = true;
1096 		}
1097 	}
1098 	write_unlock(&et->lock);
1099 
1100 	__destroy_extent_node(inode, type);
1101 
1102 	if (updated)
1103 		f2fs_mark_inode_dirty_sync(inode, true);
1104 }
1105 
1106 void f2fs_drop_extent_tree(struct inode *inode)
1107 {
1108 	__drop_extent_tree(inode, EX_READ);
1109 	__drop_extent_tree(inode, EX_BLOCK_AGE);
1110 }
1111 
1112 static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
1113 {
1114 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
1115 	struct extent_tree_info *eti = &sbi->extent_tree[type];
1116 	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
1117 	unsigned int node_cnt = 0;
1118 
1119 	if (!et)
1120 		return;
1121 
1122 	if (inode->i_nlink && !is_bad_inode(inode) &&
1123 					atomic_read(&et->node_cnt)) {
1124 		mutex_lock(&eti->extent_tree_lock);
1125 		list_add_tail(&et->list, &eti->zombie_list);
1126 		atomic_inc(&eti->total_zombie_tree);
1127 		mutex_unlock(&eti->extent_tree_lock);
1128 		return;
1129 	}
1130 
1131 	/* free all extent info belong to this extent tree */
1132 	node_cnt = __destroy_extent_node(inode, type);
1133 
1134 	/* delete extent tree entry in radix tree */
1135 	mutex_lock(&eti->extent_tree_lock);
1136 	f2fs_bug_on(sbi, atomic_read(&et->node_cnt));
1137 	radix_tree_delete(&eti->extent_tree_root, inode->i_ino);
1138 	kmem_cache_free(extent_tree_slab, et);
1139 	atomic_dec(&eti->total_ext_tree);
1140 	mutex_unlock(&eti->extent_tree_lock);
1141 
1142 	F2FS_I(inode)->extent_tree[type] = NULL;
1143 
1144 	trace_f2fs_destroy_extent_tree(inode, node_cnt, type);
1145 }
1146 
1147 void f2fs_destroy_extent_tree(struct inode *inode)
1148 {
1149 	__destroy_extent_tree(inode, EX_READ);
1150 	__destroy_extent_tree(inode, EX_BLOCK_AGE);
1151 }
1152 
1153 static void __init_extent_tree_info(struct extent_tree_info *eti)
1154 {
1155 	INIT_RADIX_TREE(&eti->extent_tree_root, GFP_NOIO);
1156 	mutex_init(&eti->extent_tree_lock);
1157 	INIT_LIST_HEAD(&eti->extent_list);
1158 	spin_lock_init(&eti->extent_lock);
1159 	atomic_set(&eti->total_ext_tree, 0);
1160 	INIT_LIST_HEAD(&eti->zombie_list);
1161 	atomic_set(&eti->total_zombie_tree, 0);
1162 	atomic_set(&eti->total_ext_node, 0);
1163 }
1164 
1165 void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
1166 {
1167 	__init_extent_tree_info(&sbi->extent_tree[EX_READ]);
1168 	__init_extent_tree_info(&sbi->extent_tree[EX_BLOCK_AGE]);
1169 
1170 	/* initialize for block age extents */
1171 	atomic64_set(&sbi->allocated_data_blocks, 0);
1172 	sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD;
1173 	sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD;
1174 	sbi->last_age_weight = LAST_AGE_WEIGHT;
1175 }
1176 
1177 int __init f2fs_create_extent_cache(void)
1178 {
1179 	extent_tree_slab = f2fs_kmem_cache_create("f2fs_extent_tree",
1180 			sizeof(struct extent_tree));
1181 	if (!extent_tree_slab)
1182 		return -ENOMEM;
1183 	extent_node_slab = f2fs_kmem_cache_create("f2fs_extent_node",
1184 			sizeof(struct extent_node));
1185 	if (!extent_node_slab) {
1186 		kmem_cache_destroy(extent_tree_slab);
1187 		return -ENOMEM;
1188 	}
1189 	return 0;
1190 }
1191 
1192 void f2fs_destroy_extent_cache(void)
1193 {
1194 	kmem_cache_destroy(extent_node_slab);
1195 	kmem_cache_destroy(extent_tree_slab);
1196 }
1197