Merge tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Don't destroy user event fields when removal fails

   User event fields are destroyed before the event is removed from
   visibility. But that can fail leaving the still visible event with no
   fields. Move the destroying of the fields to after the event is
   successfully removed from visibility.

 - Initialize function graph state is fork before calling
   copy_exec_state()

   For non-CLONE_VM forks, copy_exec_state() allocates a new
   task_exec_state. If that allocation fails, ftrace_graph_exit_task()
   will free the tasks ret_stack pointer. Since that pointer is still
   using the parent's ret_stack, it mistakenly frees the parent's
   pointer too.

   Call ftrace_graph_init() on the task first which will NULL out the
   new tasks's ret_stack and if the copy fails, it will not free
   anything.

 - Remove FGRAPH_MAX_INDEX

   The macro FGRAPH_MAX_INDEX was added but never used. Remove it.

 - Save ent_size in function graph printing of nested functions

   The function graph tracer needs to look at the next event to see if
   the next event is the return of the current function entry. If it is,
   it prints a single line:

	ktime_get();

   Otherwise it prints it like a nested function:

	tick_nohz_irq_exit() {
	    ktime_get();
	    kcpustat_irq_exit();
	}

   In order to look at the next event, it must save the current event so
   that it has the information to print from it. It saves the event in
   the iterator descriptor called "ent". What it doesn't save is the
   ent_size of the event which is now used to know if the function graph
   arguments are to be printed. The peek doesn't save the size so the
   size used happens to be that of the size of the last event that was
   seen.

   Save the entry event size in the iterator descriptor so that the
   correct size is used.

 - Fix several errors with freeing data in the histogram code

   The histogram code had a lot of leaked or or incorrect accounting
   when failures happen. Correct them.

 - Fix histogram regression of .percent and .graph modifiers

   Up until 6.3 histogram values could have "percent" or "graph"
   modifiers that changed how they were printed. But a change that added
   restricting histograms values from being strings, stack traces and
   other modifiers inadvertently prevented them from using the percent
   and graph modifiers, which were legal use cases for values.

   Put back the percent and graph modifiers.

 - Fix various typos in the comments

 - Set the trace_clock before initializing a histogram with clock
   argument

   The histogram API allows the user to specific which trace clock to
   use via a "clock=" string. The histogram is set up first before the
   clock is checked. If the passed in clock is not valid, it exits
   without fully fixing up the histogram leaving it on the list and a
   use-after-free can trigger.

   Update the clock argument first and if it fails then exit gracefully
   before the histogram trigger is placed on any lists.

 - Restore :mod: trailer after parsing in ftrace_set_clr_event

   The function ftrace_set_clr_event() modifies the parse string and
   needs to put it back to what was passed in. It searches for ":mod:"
   via a strsep() but fails to put back the first ':' in the string.

   Add back the ':' in the passed in string.

 - Take trace_array reference when opening a tracer options file

   The options files are dynamically created and some tracers add their
   own options. When a tracer adds their own list of options, the
   trace_array holding them has an array to hold the list of options for
   each tracer. This array increases in size via a krealloc(), and the
   new entry gets a newly allocated array to hold the options of the new
   tracer being added.

   The element in each entry of the tracer's option array holds a
   pointer back to the trace_array, a pointer to the tracer it is
   associated to, a pointer to the flags of the option.

   The issue is that these arrays are freed when the trace_array is
   freed when its instance it represents is removed from the instances
   directory. There's a race that an open of one of these options files
   can happen when the instance is being removed.

   Add a new helper function to be called by the open function of the
   options file to iterate all existing trace_arrays under a lock and
   find the one that has the given option element in one of it's tracer
   arrays. If found, then update the associated trace_array's reference
   counter to keep it from being freed. If not found, have the open call
   return -ENODEV.

 - Disable interrupts when acquiring the lock in rb_wake_up_waiters()

   The function rb_wake_up_waiters() assumes it will be called in
   interrupt context and does not disable irqs when taking
   cpu_buffer->reader_lock, which can be called in hard interrupt
   context. The issue is in PREEMPT_RT, this function is called in
   thread context leaving this lock open to a deadlock.

   Take the lock with interrupts disabled.

 - Use rcu_assign_pointer() for tmp_ops filter hash

   The tmp_ops used in update_ftrace_direct_mod() assigns its
   filter_hash field directly, but that field is annotated as __rcu and
   sparse complains. Assign it with rcu_assign_pointer()

 - Fix use-after-free in enable_trigger_private_data_free()

   The trace_event_call is accessed through the event_trigger_data's
   trace_event_file pointer to put the trace_event_call on freeing. The
   issue is that the trace_event_file data may have been freed already
   causing a use-after-free. Add a field to the event_trigger_data that
   points directly to the trace_event_call so that it can decrement its
   reference directly without needing to go through the
   trace_event_file.

 - Fix accounting of buffer data remote headers

   trace_buffer_desc_size() and trace_remote_alloc_buffer() undercount
   the number of pages is needed for the asked for size as it doesn't
   take into account the meta data on each page. Add a helper function
   to do the calculation properly and use that in these functions.

 - Catch nr_page_va overflow in ring_buffer_desc sizing

   The number of pages per remote ring buffer is capped by
   ring_buffer_desc::nr_page_va (32 bits). A buffer_size large enough to
   overflow that field would silently allocate a descriptor smaller than
   what was asked for.

 - Do not resize the subbuf order if any per_cpu buffer is disabled

   The mmapping of ring buffers disables resizing the subbuffers, but it
   is done per-cpu whereas the subbuf size change is done for all the
   per_cpu buffers under the buffer->mutex. It could change the size of
   some while the mapping is happening on others. Have the resize of the
   subbuf order check all the per_cpu buffers under the lock to see if
   any of them is disabled before starting and causing an inconsistency
   between buffers that are being mapped.

* tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (25 commits)
  ring-buffer: Check resize_disabled before publishing the new subbuf order
  tracing/remotes: Catch nr_page_va overflow in ring_buffer_desc sizing
  tracing/remotes: Account for ring buffer page header in size calculation
  tracing: Don't dereference trace_event_file in deferred trigger free
  ftrace: Use rcu_assign_pointer() for tmp_ops filter hash
  ring-buffer: Acquire the lock with irqsave in rb_wake_up_waiters()
  tracing: Take trace_array reference when opening a tracer options file
  tracing: Fix ring_buffer_read_page_size() kernel-doc
  tracing: Restore :mod: trailer after parsing in ftrace_set_clr_event()
  tracing: Fix memory corruption from a "STACKTRACE" histogram key
  tracing: Fix memory corruption from the histogram stacktrace modifier
  tracing: Undo the registration when enabling the histogram trigger fails
  tracing: Take the reference before publishing the named histogram trigger
  tracing: Set the trace clock before registering the histogram trigger
  tracing: Fix typo "preceeded" in comment
  tracing: Fix typo "availabe" in comment
  tracing: Let histogram values keep the percent and graph modifiers
  tracing: Keep the entry count when the histogram stats allocation fails
  tracing: Free histogram the field rejected for a bad modifier
  tracing: Free histogram the var ref when its initialization fails
  ...

1 files changed, 1262 insertions, 0 deletions

diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c
new file mode 100644
index 000000000..37cf9fa8d
--- /dev/null
+++ b/fs/f2fs/extent_cache.c
@@ -0,0 +1,1262 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * f2fs extent cache support
+ *
+ * Copyright (c) 2015 Motorola Mobility
+ * Copyright (c) 2015 Samsung Electronics
+ * Authors: Jaegeuk Kim <jaegeuk@kernel.org>
+ *          Chao Yu <chao2.yu@samsung.com>
+ *
+ * block_age-based extent cache added by:
+ * Copyright (c) 2022 xiaomi Co., Ltd.
+ *             http://www.xiaomi.com/
+ */
+
+#include <linux/fs.h>
+#include <linux/f2fs_fs.h>
+
+#include "f2fs.h"
+#include "node.h"
+#include "segment.h"
+#include <trace/events/f2fs.h>
+
+bool sanity_check_extent_cache(struct inode *inode, struct folio *ifolio)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct f2fs_extent *i_ext = &F2FS_INODE(ifolio)->i_ext;
+	struct extent_info ei;
+	int devi;
+
+	get_read_extent_info(&ei, i_ext);
+
+	if (!ei.len)
+		return true;
+
+	if (!f2fs_is_valid_blkaddr(sbi, ei.blk, DATA_GENERIC_ENHANCE) ||
+	    !f2fs_is_valid_blkaddr(sbi, ei.blk + ei.len - 1,
+					DATA_GENERIC_ENHANCE)) {
+		f2fs_warn(sbi, "%s: inode (ino=%llx) extent info [%u, %u, %u] is incorrect, run fsck to fix",
+			  __func__, inode->i_ino,
+			  ei.blk, ei.fofs, ei.len);
+		return false;
+	}
+
+	if (!IS_DEVICE_ALIASING(inode))
+		return true;
+
+	for (devi = 0; devi < sbi->s_ndevs; devi++) {
+		if (FDEV(devi).start_blk != ei.blk ||
+				FDEV(devi).end_blk != ei.blk + ei.len - 1)
+			continue;
+
+		if (devi == 0) {
+			f2fs_warn(sbi,
+			    "%s: inode (ino=%llx) is an alias of meta device",
+			    __func__, inode->i_ino);
+			return false;
+		}
+
+		if (bdev_is_zoned(FDEV(devi).bdev)) {
+			f2fs_warn(sbi,
+			    "%s: device alias inode (ino=%llx)'s extent info "
+			    "[%u, %u, %u] maps to zoned block device",
+			    __func__, inode->i_ino, ei.blk, ei.fofs, ei.len);
+			return false;
+		}
+
+		if ((GET_SEGOFF_FROM_SEG0(sbi, ei.blk) % BLKS_PER_SEC(sbi)) ||
+		    (ei.len % BLKS_PER_SEC(sbi))) {
+			f2fs_warn(sbi, "%s: device alias inode (ino=%llx)'s extent info [%u, %u, %u] is not aligned to section size %u",
+				  __func__, inode->i_ino, ei.blk, ei.fofs, ei.len,
+				  BLKS_PER_SEC(sbi));
+			return false;
+		}
+		return true;
+	}
+
+	f2fs_warn(sbi, "%s: device alias inode (ino=%llx)'s extent info "
+			"[%u, %u, %u] is inconsistent w/ any devices",
+			__func__, inode->i_ino, ei.blk, ei.fofs, ei.len);
+	return false;
+}
+
+static void __set_extent_info(struct extent_info *ei,
+				unsigned int fofs, unsigned int len,
+				block_t blk, bool keep_clen,
+				unsigned long age, unsigned long last_blocks,
+				enum extent_type type)
+{
+	ei->fofs = fofs;
+	ei->len = len;
+
+	if (type == EX_READ) {
+		ei->blk = blk;
+		if (keep_clen)
+			return;
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+		ei->c_len = 0;
+#endif
+	} else if (type == EX_BLOCK_AGE) {
+		ei->age = age;
+		ei->last_blocks = last_blocks;
+	}
+}
+
+static bool __init_may_extent_tree(struct inode *inode, enum extent_type type)
+{
+	if (type == EX_READ)
+		return test_opt(F2FS_I_SB(inode), READ_EXTENT_CACHE) &&
+			S_ISREG(inode->i_mode);
+	if (type == EX_BLOCK_AGE)
+		return test_opt(F2FS_I_SB(inode), AGE_EXTENT_CACHE) &&
+			(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode));
+	return false;
+}
+
+static bool __may_extent_tree(struct inode *inode, enum extent_type type)
+{
+	if (IS_DEVICE_ALIASING(inode) && type == EX_READ)
+		return true;
+
+	/*
+	 * for recovered files during mount do not create extents
+	 * if shrinker is not registered.
+	 */
+	if (list_empty(&F2FS_I_SB(inode)->s_list))
+		return false;
+
+	if (!__init_may_extent_tree(inode, type))
+		return false;
+
+	if (type == EX_READ) {
+		if (is_inode_flag_set(inode, FI_NO_EXTENT))
+			return false;
+		if (is_inode_flag_set(inode, FI_COMPRESSED_FILE) &&
+				 !f2fs_sb_has_readonly(F2FS_I_SB(inode)))
+			return false;
+	} else if (type == EX_BLOCK_AGE) {
+		if (is_inode_flag_set(inode, FI_COMPRESSED_FILE))
+			return false;
+		if (file_is_cold(inode))
+			return false;
+	}
+	return true;
+}
+
+static void __try_update_largest_extent(struct extent_tree *et,
+						struct extent_node *en)
+{
+	if (et->type != EX_READ)
+		return;
+	if (en->ei.len <= et->largest.len)
+		return;
+
+	et->largest = en->ei;
+	et->largest_updated = true;
+}
+
+static bool __is_extent_mergeable(struct extent_info *back,
+		struct extent_info *front, enum extent_type type)
+{
+	if (type == EX_READ) {
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+		if (back->c_len && back->len != back->c_len)
+			return false;
+		if (front->c_len && front->len != front->c_len)
+			return false;
+#endif
+		return (back->fofs + back->len == front->fofs &&
+				back->blk + back->len == front->blk);
+	} else if (type == EX_BLOCK_AGE) {
+		return (back->fofs + back->len == front->fofs &&
+			abs(back->age - front->age) <= SAME_AGE_REGION &&
+			abs(back->last_blocks - front->last_blocks) <=
+							SAME_AGE_REGION);
+	}
+	return false;
+}
+
+static bool __is_back_mergeable(struct extent_info *cur,
+		struct extent_info *back, enum extent_type type)
+{
+	return __is_extent_mergeable(back, cur, type);
+}
+
+static bool __is_front_mergeable(struct extent_info *cur,
+		struct extent_info *front, enum extent_type type)
+{
+	return __is_extent_mergeable(cur, front, type);
+}
+
+static struct extent_node *__lookup_extent_node(struct rb_root_cached *root,
+			struct extent_node *cached_en, unsigned int fofs)
+{
+	struct rb_node *node = root->rb_root.rb_node;
+	struct extent_node *en;
+
+	/* check a cached entry */
+	if (cached_en && cached_en->ei.fofs <= fofs &&
+			cached_en->ei.fofs + cached_en->ei.len > fofs)
+		return cached_en;
+
+	/* check rb_tree */
+	while (node) {
+		en = rb_entry(node, struct extent_node, rb_node);
+
+		if (fofs < en->ei.fofs)
+			node = node->rb_left;
+		else if (fofs >= en->ei.fofs + en->ei.len)
+			node = node->rb_right;
+		else
+			return en;
+	}
+	return NULL;
+}
+
+/*
+ * lookup rb entry in position of @fofs in rb-tree,
+ * if hit, return the entry, otherwise, return NULL
+ * @prev_ex: extent before fofs
+ * @next_ex: extent after fofs
+ * @insert_p: insert point for new extent at fofs
+ * in order to simplify the insertion after.
+ * tree must stay unchanged between lookup and insertion.
+ */
+static struct extent_node *__lookup_extent_node_ret(struct rb_root_cached *root,
+				struct extent_node *cached_en,
+				unsigned int fofs,
+				struct extent_node **prev_entry,
+				struct extent_node **next_entry,
+				struct rb_node ***insert_p,
+				struct rb_node **insert_parent,
+				bool *leftmost)
+{
+	struct rb_node **pnode = &root->rb_root.rb_node;
+	struct rb_node *parent = NULL, *tmp_node;
+	struct extent_node *en = cached_en;
+
+	*insert_p = NULL;
+	*insert_parent = NULL;
+	*prev_entry = NULL;
+	*next_entry = NULL;
+
+	if (RB_EMPTY_ROOT(&root->rb_root))
+		return NULL;
+
+	if (en && en->ei.fofs <= fofs && en->ei.fofs + en->ei.len > fofs)
+		goto lookup_neighbors;
+
+	*leftmost = true;
+
+	while (*pnode) {
+		parent = *pnode;
+		en = rb_entry(*pnode, struct extent_node, rb_node);
+
+		if (fofs < en->ei.fofs) {
+			pnode = &(*pnode)->rb_left;
+		} else if (fofs >= en->ei.fofs + en->ei.len) {
+			pnode = &(*pnode)->rb_right;
+			*leftmost = false;
+		} else {
+			goto lookup_neighbors;
+		}
+	}
+
+	*insert_p = pnode;
+	*insert_parent = parent;
+
+	en = rb_entry(parent, struct extent_node, rb_node);
+	tmp_node = parent;
+	if (parent && fofs > en->ei.fofs)
+		tmp_node = rb_next(parent);
+	*next_entry = rb_entry_safe(tmp_node, struct extent_node, rb_node);
+
+	tmp_node = parent;
+	if (parent && fofs < en->ei.fofs)
+		tmp_node = rb_prev(parent);
+	*prev_entry = rb_entry_safe(tmp_node, struct extent_node, rb_node);
+	return NULL;
+
+lookup_neighbors:
+	if (fofs == en->ei.fofs) {
+		/* lookup prev node for merging backward later */
+		tmp_node = rb_prev(&en->rb_node);
+		*prev_entry = rb_entry_safe(tmp_node,
+					struct extent_node, rb_node);
+	}
+	if (fofs == en->ei.fofs + en->ei.len - 1) {
+		/* lookup next node for merging frontward later */
+		tmp_node = rb_next(&en->rb_node);
+		*next_entry = rb_entry_safe(tmp_node,
+					struct extent_node, rb_node);
+	}
+	return en;
+}
+
+static struct kmem_cache *extent_tree_slab;
+static struct kmem_cache *extent_node_slab;
+
+static struct extent_node *__attach_extent_node(struct f2fs_sb_info *sbi,
+				struct extent_tree *et, struct extent_info *ei,
+				struct rb_node *parent, struct rb_node **p,
+				bool leftmost)
+{
+	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+	struct extent_node *en;
+
+	en = f2fs_kmem_cache_alloc(extent_node_slab, GFP_ATOMIC, false, sbi);
+	if (!en)
+		return NULL;
+
+	en->ei = *ei;
+	INIT_LIST_HEAD(&en->list);
+	en->et = et;
+
+	rb_link_node(&en->rb_node, parent, p);
+	rb_insert_color_cached(&en->rb_node, &et->root, leftmost);
+	atomic_inc(&et->node_cnt);
+	atomic_inc(&eti->total_ext_node);
+	return en;
+}
+
+static void __detach_extent_node(struct f2fs_sb_info *sbi,
+				struct extent_tree *et, struct extent_node *en)
+{
+	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+
+	rb_erase_cached(&en->rb_node, &et->root);
+	atomic_dec(&et->node_cnt);
+	atomic_dec(&eti->total_ext_node);
+
+	if (et->cached_en == en)
+		et->cached_en = NULL;
+	kmem_cache_free(extent_node_slab, en);
+}
+
+/*
+ * Flow to release an extent_node:
+ * 1. list_del_init
+ * 2. __detach_extent_node
+ * 3. kmem_cache_free.
+ */
+static void __release_extent_node(struct f2fs_sb_info *sbi,
+			struct extent_tree *et, struct extent_node *en)
+{
+	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+
+	spin_lock(&eti->extent_lock);
+	f2fs_bug_on(sbi, list_empty(&en->list));
+	list_del_init(&en->list);
+	spin_unlock(&eti->extent_lock);
+
+	__detach_extent_node(sbi, et, en);
+}
+
+static struct extent_tree *__grab_extent_tree(struct inode *inode,
+						enum extent_type type)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree_info *eti = &sbi->extent_tree[type];
+	struct extent_tree *et;
+	nid_t ino = inode->i_ino;
+
+	mutex_lock(&eti->extent_tree_lock);
+	et = radix_tree_lookup(&eti->extent_tree_root, ino);
+	if (!et) {
+		et = f2fs_kmem_cache_alloc(extent_tree_slab,
+					GFP_NOFS, true, NULL);
+		f2fs_radix_tree_insert(&eti->extent_tree_root, ino, et);
+		memset(et, 0, sizeof(struct extent_tree));
+		et->ino = ino;
+		et->type = type;
+		et->root = RB_ROOT_CACHED;
+		et->cached_en = NULL;
+		rwlock_init(&et->lock);
+		INIT_LIST_HEAD(&et->list);
+		atomic_set(&et->node_cnt, 0);
+		atomic_inc(&eti->total_ext_tree);
+	} else {
+		atomic_dec(&eti->total_zombie_tree);
+		list_del_init(&et->list);
+	}
+	mutex_unlock(&eti->extent_tree_lock);
+
+	/* never died until evict_inode */
+	F2FS_I(inode)->extent_tree[type] = et;
+
+	return et;
+}
+
+static unsigned int __free_extent_tree(struct f2fs_sb_info *sbi,
+				struct extent_tree *et, unsigned int nr_shrink)
+{
+	struct rb_node *node, *next;
+	struct extent_node *en;
+	unsigned int count;
+
+	node = rb_first_cached(&et->root);
+
+	for (count = 0; node && count < nr_shrink; count++) {
+		next = rb_next(node);
+		en = rb_entry(node, struct extent_node, rb_node);
+		__release_extent_node(sbi, et, en);
+		node = next;
+	}
+
+	return count;
+}
+
+static void __drop_largest_extent(struct extent_tree *et,
+					pgoff_t fofs, unsigned int len)
+{
+	if (fofs < (pgoff_t)et->largest.fofs + et->largest.len &&
+			fofs + len > et->largest.fofs) {
+		et->largest.len = 0;
+		et->largest_updated = true;
+	}
+}
+
+void f2fs_init_read_extent_tree(struct inode *inode, struct folio *ifolio)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree_info *eti = &sbi->extent_tree[EX_READ];
+	struct f2fs_extent *i_ext = &F2FS_INODE(ifolio)->i_ext;
+	struct extent_tree *et;
+	struct extent_node *en;
+	struct extent_info ei = {0};
+
+	if (!__may_extent_tree(inode, EX_READ)) {
+		/* drop largest read extent */
+		if (i_ext->len) {
+			f2fs_folio_wait_writeback(ifolio, NODE, true, true);
+			i_ext->len = 0;
+			folio_mark_dirty(ifolio);
+		}
+		set_inode_flag(inode, FI_NO_EXTENT);
+		return;
+	}
+
+	et = __grab_extent_tree(inode, EX_READ);
+
+	get_read_extent_info(&ei, i_ext);
+
+	write_lock(&et->lock);
+	if (atomic_read(&et->node_cnt) || !ei.len)
+		goto skip;
+
+	if (IS_DEVICE_ALIASING(inode)) {
+		et->largest = ei;
+		goto skip;
+	}
+
+	en = __attach_extent_node(sbi, et, &ei, NULL,
+				&et->root.rb_root.rb_node, true);
+	if (en) {
+		et->largest = en->ei;
+		et->cached_en = en;
+
+		spin_lock(&eti->extent_lock);
+		list_add_tail(&en->list, &eti->extent_list);
+		spin_unlock(&eti->extent_lock);
+	}
+skip:
+	/* Let's drop, if checkpoint got corrupted. */
+	if (f2fs_cp_error(sbi)) {
+		et->largest.len = 0;
+		et->largest_updated = true;
+	}
+	write_unlock(&et->lock);
+}
+
+void f2fs_init_age_extent_tree(struct inode *inode)
+{
+	if (!__init_may_extent_tree(inode, EX_BLOCK_AGE))
+		return;
+	__grab_extent_tree(inode, EX_BLOCK_AGE);
+}
+
+void f2fs_init_extent_tree(struct inode *inode)
+{
+	/* initialize read cache */
+	if (__init_may_extent_tree(inode, EX_READ))
+		__grab_extent_tree(inode, EX_READ);
+
+	/* initialize block age cache */
+	if (__init_may_extent_tree(inode, EX_BLOCK_AGE))
+		__grab_extent_tree(inode, EX_BLOCK_AGE);
+}
+
+static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs,
+			struct extent_info *ei, enum extent_type type)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree_info *eti = &sbi->extent_tree[type];
+	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
+	struct extent_node *en;
+	bool ret = false;
+
+	if (!et)
+		return false;
+
+	trace_f2fs_lookup_extent_tree_start(inode, pgofs, type);
+
+	read_lock(&et->lock);
+
+	if (type == EX_READ &&
+			et->largest.fofs <= pgofs &&
+			(pgoff_t)et->largest.fofs + et->largest.len > pgofs) {
+		*ei = et->largest;
+		ret = true;
+		stat_inc_largest_node_hit(sbi);
+		goto out;
+	}
+
+	if (IS_DEVICE_ALIASING(inode)) {
+		ret = false;
+		goto out;
+	}
+
+	en = __lookup_extent_node(&et->root, et->cached_en, pgofs);
+	if (!en)
+		goto out;
+
+	if (en == et->cached_en)
+		stat_inc_cached_node_hit(sbi, type);
+	else
+		stat_inc_rbtree_node_hit(sbi, type);
+
+	*ei = en->ei;
+	spin_lock(&eti->extent_lock);
+	if (!list_empty(&en->list)) {
+		list_move_tail(&en->list, &eti->extent_list);
+		et->cached_en = en;
+	}
+	spin_unlock(&eti->extent_lock);
+	ret = true;
+out:
+	stat_inc_total_hit(sbi, type);
+	read_unlock(&et->lock);
+
+	if (type == EX_READ)
+		trace_f2fs_lookup_read_extent_tree_end(inode, pgofs, ei);
+	else if (type == EX_BLOCK_AGE)
+		trace_f2fs_lookup_age_extent_tree_end(inode, pgofs, ei);
+	return ret;
+}
+
+static struct extent_node *__try_merge_extent_node(struct f2fs_sb_info *sbi,
+				struct extent_tree *et, struct extent_info *ei,
+				struct extent_node *prev_ex,
+				struct extent_node *next_ex)
+{
+	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+	struct extent_node *en = NULL;
+
+	if (prev_ex && __is_back_mergeable(ei, &prev_ex->ei, et->type)) {
+		prev_ex->ei.len += ei->len;
+		ei = &prev_ex->ei;
+		en = prev_ex;
+	}
+
+	if (next_ex && __is_front_mergeable(ei, &next_ex->ei, et->type)) {
+		next_ex->ei.fofs = ei->fofs;
+		next_ex->ei.len += ei->len;
+		if (et->type == EX_READ)
+			next_ex->ei.blk = ei->blk;
+		if (en)
+			__release_extent_node(sbi, et, prev_ex);
+
+		en = next_ex;
+	}
+
+	if (!en)
+		return NULL;
+
+	__try_update_largest_extent(et, en);
+
+	spin_lock(&eti->extent_lock);
+	if (!list_empty(&en->list)) {
+		list_move_tail(&en->list, &eti->extent_list);
+		et->cached_en = en;
+	}
+	spin_unlock(&eti->extent_lock);
+	return en;
+}
+
+static struct extent_node *__insert_extent_tree(struct f2fs_sb_info *sbi,
+				struct extent_tree *et, struct extent_info *ei,
+				struct rb_node **insert_p,
+				struct rb_node *insert_parent,
+				bool leftmost)
+{
+	struct extent_tree_info *eti = &sbi->extent_tree[et->type];
+	struct rb_node **p = &et->root.rb_root.rb_node;
+	struct rb_node *parent = NULL;
+	struct extent_node *en = NULL;
+
+	if (insert_p && insert_parent) {
+		parent = insert_parent;
+		p = insert_p;
+		goto do_insert;
+	}
+
+	leftmost = true;
+
+	/* look up extent_node in the rb tree */
+	while (*p) {
+		parent = *p;
+		en = rb_entry(parent, struct extent_node, rb_node);
+
+		if (ei->fofs < en->ei.fofs) {
+			p = &(*p)->rb_left;
+		} else if (ei->fofs >= en->ei.fofs + en->ei.len) {
+			p = &(*p)->rb_right;
+			leftmost = false;
+		} else {
+			f2fs_err_ratelimited(sbi, "%s: corrupted extent, type: %d, "
+				"extent node in rb tree [%u, %u, %u], age [%llu, %llu], "
+				"extent node to insert [%u, %u, %u], age [%llu, %llu]",
+				__func__, et->type, en->ei.fofs, en->ei.blk, en->ei.len, en->ei.age,
+				en->ei.last_blocks, ei->fofs, ei->blk, ei->len, ei->age, ei->last_blocks);
+			f2fs_bug_on(sbi, 1);
+			return NULL;
+		}
+	}
+
+do_insert:
+	en = __attach_extent_node(sbi, et, ei, parent, p, leftmost);
+	if (!en)
+		return NULL;
+
+	__try_update_largest_extent(et, en);
+
+	/* update in global extent list */
+	spin_lock(&eti->extent_lock);
+	list_add_tail(&en->list, &eti->extent_list);
+	et->cached_en = en;
+	spin_unlock(&eti->extent_lock);
+	return en;
+}
+
+static unsigned int __destroy_extent_node(struct inode *inode,
+					enum extent_type type)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
+	unsigned int nr_shrink = type == EX_READ ?
+				READ_EXTENT_CACHE_SHRINK_NUMBER :
+				AGE_EXTENT_CACHE_SHRINK_NUMBER;
+	unsigned int node_cnt = 0;
+
+	if (!et || !atomic_read(&et->node_cnt))
+		return 0;
+
+	while (atomic_read(&et->node_cnt)) {
+		write_lock(&et->lock);
+		node_cnt += __free_extent_tree(sbi, et, nr_shrink);
+		write_unlock(&et->lock);
+	}
+
+	return node_cnt;
+}
+
+static void __update_extent_tree_range(struct inode *inode,
+			struct extent_info *tei, enum extent_type type)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
+	struct extent_node *en = NULL, *en1 = NULL;
+	struct extent_node *prev_en = NULL, *next_en = NULL;
+	struct extent_info ei, dei, prev;
+	struct rb_node **insert_p = NULL, *insert_parent = NULL;
+	unsigned int fofs = tei->fofs, len = tei->len;
+	unsigned int end = fofs + len;
+	bool updated = false;
+	bool leftmost = false;
+
+	if (!et)
+		return;
+
+	if (unlikely(len == 0)) {
+		f2fs_err_ratelimited(sbi, "%s: extent len is zero, type: %d, "
+			"extent [%u, %u, %u], age [%llu, %llu]",
+			__func__, type, tei->fofs, tei->blk, tei->len,
+			tei->age, tei->last_blocks);
+		f2fs_bug_on(sbi, 1);
+		return;
+	}
+
+	if (type == EX_READ)
+		trace_f2fs_update_read_extent_tree_range(inode, fofs, len,
+						tei->blk, 0);
+	else if (type == EX_BLOCK_AGE)
+		trace_f2fs_update_age_extent_tree_range(inode, fofs, len,
+						tei->age, tei->last_blocks);
+
+	write_lock(&et->lock);
+
+	if (type == EX_READ) {
+		if (is_inode_flag_set(inode, FI_NO_EXTENT)) {
+			write_unlock(&et->lock);
+			return;
+		}
+
+		prev = et->largest;
+		dei.len = 0;
+
+		/*
+		 * drop largest extent before lookup, in case it's already
+		 * been shrunk from extent tree
+		 */
+		__drop_largest_extent(et, fofs, len);
+	}
+
+	/* 1. lookup first extent node in range [fofs, fofs + len - 1] */
+	en = __lookup_extent_node_ret(&et->root,
+					et->cached_en, fofs,
+					&prev_en, &next_en,
+					&insert_p, &insert_parent,
+					&leftmost);
+	if (!en)
+		en = next_en;
+
+	/* 2. invalidate all extent nodes in range [fofs, fofs + len - 1] */
+	while (en && en->ei.fofs < end) {
+		unsigned int org_end;
+		int parts = 0;	/* # of parts current extent split into */
+
+		next_en = en1 = NULL;
+
+		dei = en->ei;
+		org_end = dei.fofs + dei.len;
+		f2fs_bug_on(sbi, fofs >= org_end);
+
+		if (fofs > dei.fofs && (type != EX_READ ||
+				fofs - dei.fofs >= F2FS_MIN_EXTENT_LEN)) {
+			en->ei.len = fofs - en->ei.fofs;
+			prev_en = en;
+			parts = 1;
+		}
+
+		if (end < org_end && (type != EX_READ ||
+			(org_end - end >= F2FS_MIN_EXTENT_LEN &&
+			atomic_read(&et->node_cnt) <
+					sbi->max_read_extent_count))) {
+			if (parts) {
+				__set_extent_info(&ei,
+					end, org_end - end,
+					end - dei.fofs + dei.blk, false,
+					dei.age, dei.last_blocks,
+					type);
+				en1 = __insert_extent_tree(sbi, et, &ei,
+							NULL, NULL, true);
+				next_en = en1;
+			} else {
+				__set_extent_info(&en->ei,
+					end, en->ei.len - (end - dei.fofs),
+					en->ei.blk + (end - dei.fofs), true,
+					dei.age, dei.last_blocks,
+					type);
+				next_en = en;
+			}
+			parts++;
+		}
+
+		if (!next_en) {
+			struct rb_node *node = rb_next(&en->rb_node);
+
+			next_en = rb_entry_safe(node, struct extent_node,
+						rb_node);
+		}
+
+		if (parts)
+			__try_update_largest_extent(et, en);
+		else
+			__release_extent_node(sbi, et, en);
+
+		/*
+		 * if original extent is split into zero or two parts, extent
+		 * tree has been altered by deletion or insertion, therefore
+		 * invalidate pointers regard to tree.
+		 */
+		if (parts != 1) {
+			insert_p = NULL;
+			insert_parent = NULL;
+		}
+		en = next_en;
+	}
+
+	if (type == EX_BLOCK_AGE)
+		goto update_age_extent_cache;
+
+	/* 3. update extent in read extent cache */
+	BUG_ON(type != EX_READ);
+
+	if (tei->blk) {
+		__set_extent_info(&ei, fofs, len, tei->blk, false,
+				  0, 0, EX_READ);
+		if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
+			__insert_extent_tree(sbi, et, &ei,
+					insert_p, insert_parent, leftmost);
+
+		/* give up extent_cache, if split and small updates happen */
+		if (dei.len >= 1 &&
+				prev.len < F2FS_MIN_EXTENT_LEN &&
+				et->largest.len < F2FS_MIN_EXTENT_LEN) {
+			et->largest.len = 0;
+			et->largest_updated = true;
+			set_inode_flag(inode, FI_NO_EXTENT);
+		}
+	}
+
+	if (et->largest_updated) {
+		et->largest_updated = false;
+		updated = true;
+	}
+	goto out_read_extent_cache;
+update_age_extent_cache:
+	if (tei->last_blocks == F2FS_EXTENT_AGE_INVALID)
+		goto out_read_extent_cache;
+
+	__set_extent_info(&ei, fofs, len, 0, false,
+			tei->age, tei->last_blocks, EX_BLOCK_AGE);
+	if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
+		__insert_extent_tree(sbi, et, &ei,
+					insert_p, insert_parent, leftmost);
+out_read_extent_cache:
+	write_unlock(&et->lock);
+
+	if (is_inode_flag_set(inode, FI_NO_EXTENT))
+		__destroy_extent_node(inode, EX_READ);
+
+	if (updated)
+		f2fs_mark_inode_dirty_sync(inode, true);
+}
+
+#ifdef CONFIG_F2FS_FS_COMPRESSION
+void f2fs_update_read_extent_tree_range_compressed(struct inode *inode,
+				pgoff_t fofs, block_t blkaddr, unsigned int llen,
+				unsigned int c_len)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree *et = F2FS_I(inode)->extent_tree[EX_READ];
+	struct extent_node *en = NULL;
+	struct extent_node *prev_en = NULL, *next_en = NULL;
+	struct extent_info ei;
+	struct rb_node **insert_p = NULL, *insert_parent = NULL;
+	bool leftmost = false;
+
+	trace_f2fs_update_read_extent_tree_range(inode, fofs, llen,
+						blkaddr, c_len);
+
+	/* it is safe here to check FI_NO_EXTENT w/o et->lock in ro image */
+	if (is_inode_flag_set(inode, FI_NO_EXTENT))
+		return;
+
+	write_lock(&et->lock);
+
+	en = __lookup_extent_node_ret(&et->root,
+					et->cached_en, fofs,
+					&prev_en, &next_en,
+					&insert_p, &insert_parent,
+					&leftmost);
+	if (en)
+		goto unlock_out;
+
+	__set_extent_info(&ei, fofs, llen, blkaddr, true, 0, 0, EX_READ);
+	ei.c_len = c_len;
+
+	if (!__try_merge_extent_node(sbi, et, &ei, prev_en, next_en))
+		__insert_extent_tree(sbi, et, &ei,
+				insert_p, insert_parent, leftmost);
+unlock_out:
+	write_unlock(&et->lock);
+}
+#endif
+
+static unsigned long long __calculate_block_age(struct f2fs_sb_info *sbi,
+						unsigned long long new,
+						unsigned long long old)
+{
+	unsigned int rem_old, rem_new;
+	unsigned long long res;
+	unsigned int weight = sbi->last_age_weight;
+
+	res = div_u64_rem(new, 100, &rem_new) * (100 - weight)
+		+ div_u64_rem(old, 100, &rem_old) * weight;
+
+	if (rem_new)
+		res += rem_new * (100 - weight) / 100;
+	if (rem_old)
+		res += rem_old * weight / 100;
+
+	return res;
+}
+
+/* This returns a new age and allocated blocks in ei */
+static int __get_new_block_age(struct inode *inode, struct extent_info *ei,
+						block_t blkaddr)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	loff_t f_size = i_size_read(inode);
+	unsigned long long cur_blocks =
+				atomic64_read(&sbi->allocated_data_blocks);
+	struct extent_info tei = *ei;	/* only fofs and len are valid */
+
+	/*
+	 * When I/O is not aligned to a PAGE_SIZE, update will happen to the last
+	 * file block even in seq write. So don't record age for newly last file
+	 * block here.
+	 */
+	if ((f_size >> PAGE_SHIFT) == ei->fofs && f_size & (PAGE_SIZE - 1) &&
+			blkaddr == NEW_ADDR)
+		return -EINVAL;
+
+	if (__lookup_extent_tree(inode, ei->fofs, &tei, EX_BLOCK_AGE)) {
+		unsigned long long cur_age;
+
+		if (cur_blocks >= tei.last_blocks)
+			cur_age = cur_blocks - tei.last_blocks;
+		else
+			/* allocated_data_blocks overflow */
+			cur_age = (ULLONG_MAX - 1) - tei.last_blocks + cur_blocks;
+
+		if (tei.age)
+			ei->age = __calculate_block_age(sbi, cur_age, tei.age);
+		else
+			ei->age = cur_age;
+		ei->last_blocks = cur_blocks;
+		WARN_ON(ei->age > cur_blocks);
+		return 0;
+	}
+
+	f2fs_bug_on(sbi, blkaddr == NULL_ADDR);
+
+	/* the data block was allocated for the first time */
+	if (blkaddr == NEW_ADDR)
+		goto out;
+
+	if (__is_valid_data_blkaddr(blkaddr) &&
+	    !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE))
+		return -EINVAL;
+out:
+	/*
+	 * init block age with zero, this can happen when the block age extent
+	 * was reclaimed due to memory constraint or system reboot
+	 */
+	ei->age = 0;
+	ei->last_blocks = cur_blocks;
+	return 0;
+}
+
+static void __update_extent_cache(struct dnode_of_data *dn, enum extent_type type)
+{
+	struct extent_info ei = {};
+
+	if (!__may_extent_tree(dn->inode, type))
+		return;
+
+	ei.fofs = f2fs_start_bidx_of_node(ofs_of_node(dn->node_folio), dn->inode) +
+								dn->ofs_in_node;
+	ei.len = 1;
+
+	if (type == EX_READ) {
+		if (dn->data_blkaddr == NEW_ADDR)
+			ei.blk = NULL_ADDR;
+		else
+			ei.blk = dn->data_blkaddr;
+	} else if (type == EX_BLOCK_AGE) {
+		if (__get_new_block_age(dn->inode, &ei, dn->data_blkaddr))
+			return;
+	}
+	__update_extent_tree_range(dn->inode, &ei, type);
+}
+
+static unsigned int __shrink_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink,
+					enum extent_type type)
+{
+	struct extent_tree_info *eti = &sbi->extent_tree[type];
+	struct extent_tree *et, *next;
+	struct extent_node *en;
+	unsigned int node_cnt = 0, tree_cnt = 0;
+	int remained;
+
+	if (!atomic_read(&eti->total_zombie_tree))
+		goto free_node;
+
+	if (!mutex_trylock(&eti->extent_tree_lock))
+		goto out;
+
+	/* 1. remove unreferenced extent tree */
+	list_for_each_entry_safe(et, next, &eti->zombie_list, list) {
+		if (atomic_read(&et->node_cnt)) {
+			write_lock(&et->lock);
+			node_cnt += __free_extent_tree(sbi, et,
+					nr_shrink - node_cnt - tree_cnt);
+			write_unlock(&et->lock);
+		}
+
+		if (atomic_read(&et->node_cnt))
+			goto unlock_out;
+
+		list_del_init(&et->list);
+		radix_tree_delete(&eti->extent_tree_root, et->ino);
+		kmem_cache_free(extent_tree_slab, et);
+		atomic_dec(&eti->total_ext_tree);
+		atomic_dec(&eti->total_zombie_tree);
+		tree_cnt++;
+
+		if (node_cnt + tree_cnt >= nr_shrink)
+			goto unlock_out;
+		cond_resched();
+	}
+	mutex_unlock(&eti->extent_tree_lock);
+
+free_node:
+	/* 2. remove LRU extent entries */
+	if (!mutex_trylock(&eti->extent_tree_lock))
+		goto out;
+
+	remained = nr_shrink - (node_cnt + tree_cnt);
+
+	spin_lock(&eti->extent_lock);
+	for (; remained > 0; remained--) {
+		if (list_empty(&eti->extent_list))
+			break;
+		en = list_first_entry(&eti->extent_list,
+					struct extent_node, list);
+		et = en->et;
+		if (!write_trylock(&et->lock)) {
+			/* refresh this extent node's position in extent list */
+			list_move_tail(&en->list, &eti->extent_list);
+			continue;
+		}
+
+		list_del_init(&en->list);
+		spin_unlock(&eti->extent_lock);
+
+		__detach_extent_node(sbi, et, en);
+
+		write_unlock(&et->lock);
+		node_cnt++;
+		spin_lock(&eti->extent_lock);
+	}
+	spin_unlock(&eti->extent_lock);
+
+unlock_out:
+	mutex_unlock(&eti->extent_tree_lock);
+out:
+	trace_f2fs_shrink_extent_tree(sbi, node_cnt, tree_cnt, type);
+
+	return node_cnt + tree_cnt;
+}
+
+/* read extent cache operations */
+bool f2fs_lookup_read_extent_cache(struct inode *inode, pgoff_t pgofs,
+				struct extent_info *ei)
+{
+	if (!__may_extent_tree(inode, EX_READ))
+		return false;
+
+	return __lookup_extent_tree(inode, pgofs, ei, EX_READ);
+}
+
+bool f2fs_lookup_read_extent_cache_block(struct inode *inode, pgoff_t index,
+				block_t *blkaddr)
+{
+	struct extent_info ei = {};
+
+	if (!f2fs_lookup_read_extent_cache(inode, index, &ei))
+		return false;
+	*blkaddr = ei.blk + index - ei.fofs;
+	return true;
+}
+
+void f2fs_update_read_extent_cache(struct dnode_of_data *dn)
+{
+	return __update_extent_cache(dn, EX_READ);
+}
+
+void f2fs_update_read_extent_cache_range(struct dnode_of_data *dn,
+				pgoff_t fofs, block_t blkaddr, unsigned int len)
+{
+	struct extent_info ei = {
+		.fofs = fofs,
+		.len = len,
+		.blk = blkaddr,
+	};
+
+	if (!__may_extent_tree(dn->inode, EX_READ))
+		return;
+
+	__update_extent_tree_range(dn->inode, &ei, EX_READ);
+}
+
+unsigned int f2fs_shrink_read_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
+{
+	if (!test_opt(sbi, READ_EXTENT_CACHE))
+		return 0;
+
+	return __shrink_extent_tree(sbi, nr_shrink, EX_READ);
+}
+
+/* block age extent cache operations */
+bool f2fs_lookup_age_extent_cache(struct inode *inode, pgoff_t pgofs,
+				struct extent_info *ei)
+{
+	if (!__may_extent_tree(inode, EX_BLOCK_AGE))
+		return false;
+
+	return __lookup_extent_tree(inode, pgofs, ei, EX_BLOCK_AGE);
+}
+
+void f2fs_update_age_extent_cache(struct dnode_of_data *dn)
+{
+	return __update_extent_cache(dn, EX_BLOCK_AGE);
+}
+
+void f2fs_update_age_extent_cache_range(struct dnode_of_data *dn,
+				pgoff_t fofs, unsigned int len)
+{
+	struct extent_info ei = {
+		.fofs = fofs,
+		.len = len,
+		.last_blocks = F2FS_EXTENT_AGE_INVALID,
+	};
+
+	if (!__may_extent_tree(dn->inode, EX_BLOCK_AGE))
+		return;
+
+	__update_extent_tree_range(dn->inode, &ei, EX_BLOCK_AGE);
+}
+
+unsigned int f2fs_shrink_age_extent_tree(struct f2fs_sb_info *sbi, int nr_shrink)
+{
+	if (!test_opt(sbi, AGE_EXTENT_CACHE))
+		return 0;
+
+	return __shrink_extent_tree(sbi, nr_shrink, EX_BLOCK_AGE);
+}
+
+void f2fs_destroy_extent_node(struct inode *inode)
+{
+	__destroy_extent_node(inode, EX_READ);
+	__destroy_extent_node(inode, EX_BLOCK_AGE);
+}
+
+static void __drop_extent_tree(struct inode *inode, enum extent_type type)
+{
+	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
+	bool updated = false;
+
+	if (!__may_extent_tree(inode, type))
+		return;
+
+	write_lock(&et->lock);
+	if (type == EX_READ) {
+		set_inode_flag(inode, FI_NO_EXTENT);
+		if (et->largest.len) {
+			et->largest.len = 0;
+			updated = true;
+		}
+	}
+	write_unlock(&et->lock);
+
+	__destroy_extent_node(inode, type);
+
+	if (updated)
+		f2fs_mark_inode_dirty_sync(inode, true);
+}
+
+void f2fs_drop_extent_tree(struct inode *inode)
+{
+	__drop_extent_tree(inode, EX_READ);
+	__drop_extent_tree(inode, EX_BLOCK_AGE);
+}
+
+static void __destroy_extent_tree(struct inode *inode, enum extent_type type)
+{
+	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
+	struct extent_tree_info *eti = &sbi->extent_tree[type];
+	struct extent_tree *et = F2FS_I(inode)->extent_tree[type];
+	unsigned int node_cnt = 0;
+
+	if (!et)
+		return;
+
+	if (inode->i_nlink && !is_bad_inode(inode) &&
+					atomic_read(&et->node_cnt)) {
+		mutex_lock(&eti->extent_tree_lock);
+		list_add_tail(&et->list, &eti->zombie_list);
+		atomic_inc(&eti->total_zombie_tree);
+		mutex_unlock(&eti->extent_tree_lock);
+		return;
+	}
+
+	/* free all extent info belong to this extent tree */
+	node_cnt = __destroy_extent_node(inode, type);
+
+	/* delete extent tree entry in radix tree */
+	mutex_lock(&eti->extent_tree_lock);
+	f2fs_bug_on(sbi, atomic_read(&et->node_cnt));
+	radix_tree_delete(&eti->extent_tree_root, inode->i_ino);
+	kmem_cache_free(extent_tree_slab, et);
+	atomic_dec(&eti->total_ext_tree);
+	mutex_unlock(&eti->extent_tree_lock);
+
+	F2FS_I(inode)->extent_tree[type] = NULL;
+
+	trace_f2fs_destroy_extent_tree(inode, node_cnt, type);
+}
+
+void f2fs_destroy_extent_tree(struct inode *inode)
+{
+	__destroy_extent_tree(inode, EX_READ);
+	__destroy_extent_tree(inode, EX_BLOCK_AGE);
+}
+
+static void __init_extent_tree_info(struct extent_tree_info *eti)
+{
+	INIT_RADIX_TREE(&eti->extent_tree_root, GFP_NOIO);
+	mutex_init(&eti->extent_tree_lock);
+	INIT_LIST_HEAD(&eti->extent_list);
+	spin_lock_init(&eti->extent_lock);
+	atomic_set(&eti->total_ext_tree, 0);
+	INIT_LIST_HEAD(&eti->zombie_list);
+	atomic_set(&eti->total_zombie_tree, 0);
+	atomic_set(&eti->total_ext_node, 0);
+}
+
+void f2fs_init_extent_cache_info(struct f2fs_sb_info *sbi)
+{
+	__init_extent_tree_info(&sbi->extent_tree[EX_READ]);
+	__init_extent_tree_info(&sbi->extent_tree[EX_BLOCK_AGE]);
+
+	/* initialize for block age extents */
+	atomic64_set(&sbi->allocated_data_blocks, 0);
+	sbi->hot_data_age_threshold = DEF_HOT_DATA_AGE_THRESHOLD;
+	sbi->warm_data_age_threshold = DEF_WARM_DATA_AGE_THRESHOLD;
+	sbi->last_age_weight = LAST_AGE_WEIGHT;
+	sbi->max_read_extent_count = DEF_MAX_READ_EXTENT_COUNT;
+}
+
+int __init f2fs_create_extent_cache(void)
+{
+	extent_tree_slab = f2fs_kmem_cache_create("f2fs_extent_tree",
+			sizeof(struct extent_tree));
+	if (!extent_tree_slab)
+		return -ENOMEM;
+	extent_node_slab = f2fs_kmem_cache_create("f2fs_extent_node",
+			sizeof(struct extent_node));
+	if (!extent_node_slab) {
+		kmem_cache_destroy(extent_tree_slab);
+		return -ENOMEM;
+	}
+	return 0;
+}
+
+void f2fs_destroy_extent_cache(void)
+{
+	kmem_cache_destroy(extent_node_slab);
+	kmem_cache_destroy(extent_tree_slab);
+}