From 448b643eb9e0bd64d0facfbd857e96135a95401b Mon Sep 17 00:00:00 2001
From: Vladimir Davydov <vdavydov.dev@gmail.com>
Date: Sat, 1 Jul 2017 18:15:46 +0300
Subject: [PATCH] vinyl: move vy_range to its own source file
Needed for #1906
---
src/box/CMakeLists.txt | 1 +
src/box/vinyl.c | 648 +----------------------------------------
src/box/vy_range.c | 549 ++++++++++++++++++++++++++++++++++
src/box/vy_range.h | 321 ++++++++++++++++++++
4 files changed, 872 insertions(+), 647 deletions(-)
create mode 100644 src/box/vy_range.c
create mode 100644 src/box/vy_range.h
diff --git a/src/box/CMakeLists.txt b/src/box/CMakeLists.txt
index 6f61c5b3bf..98ddc80000 100644
--- a/src/box/CMakeLists.txt
+++ b/src/box/CMakeLists.txt
@@ -54,6 +54,7 @@ add_library(box STATIC
vy_stmt.c
vy_mem.c
vy_run.c
+ vy_range.c
vy_write_iterator.c
vy_cache.c
vy_log.c
diff --git a/src/box/vinyl.c b/src/box/vinyl.c
index 0eb1e8d27b..3673091570 100644
--- a/src/box/vinyl.c
+++ b/src/box/vinyl.c
@@ -32,6 +32,7 @@
#include "vy_mem.h"
#include "vy_run.h"
+#include "vy_range.h"
#include "vy_cache.h"
#include "vy_log.h"
#include "vy_upsert.h"
@@ -263,68 +264,6 @@ vy_stat_tx_write_rate(struct vy_stat *s)
return rmean_mean(s->rmean, VY_STAT_TX_WRITE);
}
-struct vy_range {
- /** Unique ID of this range. */
- int64_t id;
- /**
- * Range lower bound. NULL if range is leftmost.
- * Both 'begin' and 'end' statements have SELECT type with the full
- * idexed key.
- */
- struct tuple *begin;
- /** Range upper bound. NULL if range is rightmost. */
- struct tuple *end;
- /** Key definition for comparing range boundaries. */
- const struct key_def *key_def;
- /** An estimate of the number of statements in this range. */
- struct vy_disk_stmt_counter count;
- /**
- * List of run slices in this range, linked by vy_slice->in_range.
- * The newer a slice, the closer it to the list head.
- */
- struct rlist slices;
- /** Number of entries in the ->slices list. */
- int slice_count;
- /**
- * The goal of compaction is to reduce read amplification.
- * All ranges for which the LSM tree has more runs per
- * level than run_count_per_level or run size larger than
- * one defined by run_size_ratio of this level are candidates
- * for compaction.
- * Unlike other LSM implementations, Vinyl can have many
- * sorted runs in a single level, and is able to compact
- * runs from any number of adjacent levels. Moreover,
- * higher levels are always taken in when compacting
- * a lower level - i.e. L1 is always included when
- * compacting L2, and both L1 and L2 are always included
- * when compacting L3.
- *
- * This variable contains the number of runs the next
- * compaction of this range will include.
- *
- * The lower the level is scheduled for compaction,
- * the bigger it tends to be because upper levels are
- * taken in.
- * @sa vy_range_update_compact_priority() to see
- * how we decide how many runs to compact next time.
- */
- int compact_priority;
- /** Number of times the range was compacted. */
- int n_compactions;
- /** Link in vy_index->tree. */
- rb_node(struct vy_range) tree_node;
- /** Link in vy_index->range_heap. */
- struct heap_node heap_node;
- /**
- * Incremented whenever an in-memory index or on disk
- * run is added to or deleted from this range. Used to
- * invalidate iterators.
- */
- uint32_t version;
-};
-
-typedef rb_tree(struct vy_range) vy_range_tree_t;
-
/**
* A single operation made by a transaction:
* a single read or write in a vy_index.
@@ -779,13 +718,6 @@ struct vy_merge_iterator {
bool range_ended;
};
-struct vy_range_iterator {
- vy_range_tree_t *tree;
- enum iterator_type iterator_type;
- const struct tuple *key;
- struct vy_range *curr_range;
-};
-
/**
* Complex read iterator over vinyl index and write_set of current tx
* Iterates over ranges, creates merge iterator for every range and outputs
@@ -1168,69 +1100,6 @@ vy_index_unacct_range(struct vy_index *index, struct vy_range *range)
histogram_discard(index->run_hist, range->slice_count);
}
-/** An snprint-style function to print a range's boundaries. */
-static int
-vy_range_snprint(char *buf, int size, const struct vy_range *range)
-{
- int total = 0;
- SNPRINT(total, snprintf, buf, size, "(");
- if (range->begin != NULL)
- SNPRINT(total, vy_key_snprint, buf, size,
- tuple_data(range->begin));
- else
- SNPRINT(total, snprintf, buf, size, "-inf");
- SNPRINT(total, snprintf, buf, size, "..");
- if (range->end != NULL)
- SNPRINT(total, vy_key_snprint, buf, size,
- tuple_data(range->end));
- else
- SNPRINT(total, snprintf, buf, size, "inf");
- SNPRINT(total, snprintf, buf, size, ")");
- return total;
-}
-
-/**
- * Helper function returning a human readable representation
- * of a range's boundaries.
- */
-static const char *
-vy_range_str(struct vy_range *range)
-{
- char *buf = tt_static_buf();
- vy_range_snprint(buf, TT_STATIC_BUF_LEN, range);
- return buf;
-}
-
-/** Add a run slice to the head of a range's list. */
-static void
-vy_range_add_slice(struct vy_range *range, struct vy_slice *slice)
-{
- rlist_add_entry(&range->slices, slice, in_range);
- range->slice_count++;
- vy_disk_stmt_counter_add(&range->count, &slice->count);
-}
-
-/** Add a run slice to a range's list before @next_slice. */
-static void
-vy_range_add_slice_before(struct vy_range *range, struct vy_slice *slice,
- struct vy_slice *next_slice)
-{
- rlist_add_tail(&next_slice->in_range, &slice->in_range);
- range->slice_count++;
- vy_disk_stmt_counter_add(&range->count, &slice->count);
-}
-
-/** Remove a run slice from a range's list. */
-static void
-vy_range_remove_slice(struct vy_range *range, struct vy_slice *slice)
-{
- assert(range->slice_count > 0);
- assert(!rlist_empty(&range->slices));
- rlist_del_entry(slice, in_range);
- range->slice_count--;
- vy_disk_stmt_counter_sub(&range->count, &slice->count);
-}
-
/**
* Allocate a new run for an index and write the information
* about it to the metadata log so that we could still find
@@ -1288,30 +1157,6 @@ vy_run_discard(struct vy_run *run)
}
}
-/** Return true if a task was scheduled for a given range. */
-static bool
-vy_range_is_scheduled(struct vy_range *range)
-{
- return range->heap_node.pos == UINT32_MAX;
-}
-
-#define HEAP_NAME vy_range_heap
-
-static bool
-vy_range_heap_less(struct heap_node *a, struct heap_node *b)
-{
- struct vy_range *left = container_of(a, struct vy_range, heap_node);
- struct vy_range *right = container_of(b, struct vy_range, heap_node);
- return left->compact_priority > right->compact_priority;
-}
-
-#define HEAP_LESS(h, l, r) vy_range_heap_less(l, r)
-
-#include "salad/heap.h"
-
-#undef HEAP_LESS
-#undef HEAP_NAME
-
/** Return max compact_priority among ranges of an index. */
static int
vy_index_compact_priority(struct vy_index *index)
@@ -1486,215 +1331,6 @@ vy_scheduler_remove_mem(struct vy_scheduler *scheduler, struct vy_mem *mem);
static bool
vy_scheduler_needs_dump(struct vy_scheduler *scheduler);
-static int
-vy_range_tree_cmp(struct vy_range *a, struct vy_range *b);
-
-static int
-vy_range_tree_key_cmp(const struct tuple *a, struct vy_range *b);
-
-rb_gen_ext_key(MAYBE_UNUSED static inline, vy_range_tree_, vy_range_tree_t,
- struct vy_range, tree_node, vy_range_tree_cmp,
- const struct tuple *, vy_range_tree_key_cmp);
-
-static int
-vy_range_tree_cmp(struct vy_range *range_a, struct vy_range *range_b)
-{
- if (range_a == range_b)
- return 0;
-
- /* Any key > -inf. */
- if (range_a->begin == NULL)
- return -1;
- if (range_b->begin == NULL)
- return 1;
-
- assert(range_a->key_def == range_b->key_def);
- return vy_key_compare(range_a->begin, range_b->begin,
- range_a->key_def);
-}
-
-static int
-vy_range_tree_key_cmp(const struct tuple *stmt, struct vy_range *range)
-{
- /* Any key > -inf. */
- if (range->begin == NULL)
- return 1;
- return vy_stmt_compare_with_key(stmt, range->begin, range->key_def);
-}
-
-static void
-vy_range_iterator_open(struct vy_range_iterator *itr, vy_range_tree_t *tree,
- enum iterator_type iterator_type,
- const struct tuple *key)
-{
- itr->tree = tree;
- itr->iterator_type = iterator_type;
- itr->key = key;
- itr->curr_range = NULL;
-}
-
-/*
- * Find the first range in which a given key should be looked up.
- */
-static struct vy_range *
-vy_range_tree_find_by_key(vy_range_tree_t *tree,
- enum iterator_type iterator_type,
- const struct tuple *key)
-{
- uint32_t key_field_count = tuple_field_count(key);
- if (key_field_count == 0) {
- switch (iterator_type) {
- case ITER_LT:
- case ITER_LE:
- return vy_range_tree_last(tree);
- case ITER_GT:
- case ITER_GE:
- case ITER_EQ:
- return vy_range_tree_first(tree);
- default:
- unreachable();
- return NULL;
- }
- }
- /* route */
- struct vy_range *range;
- if (iterator_type == ITER_GE || iterator_type == ITER_GT ||
- iterator_type == ITER_EQ) {
- /**
- * Case 1. part_count == 1, looking for [10]. ranges:
- * {1, 3, 5} {7, 8, 9} {10, 15 20} {22, 32, 42}
- * ^looking for this
- * Case 2. part_count == 1, looking for [10]. ranges:
- * {1, 2, 4} {5, 6, 7, 8} {50, 100, 200}
- * ^looking for this
- * Case 3. part_count == 2, looking for [10]. ranges:
- * {[1, 2], [2, 3]} {[9, 1], [10, 1], [10 2], [11 3]} {[12,..}
- * ^looking for this
- * Case 4. part_count == 2, looking for [10]. ranges:
- * {[1, 2], [10, 1]} {[10, 2] [10 3] [11 3]} {[12, 1]..}
- * ^looking for this
- * Case 5. part_count does not matter, looking for [10].
- * ranges:
- * {100, 200}, {300, 400}
- * ^looking for this
- */
- /**
- * vy_range_tree_psearch finds least range with begin == key
- * or previous if equal was not found
- */
- range = vy_range_tree_psearch(tree, key);
- /* switch to previous for case (4) */
- if (range != NULL && range->begin != NULL &&
- key_field_count < range->key_def->part_count &&
- vy_stmt_compare_with_key(key, range->begin,
- range->key_def) == 0)
- range = vy_range_tree_prev(tree, range);
- /* for case 5 or subcase of case 4 */
- if (range == NULL)
- range = vy_range_tree_first(tree);
- } else {
- assert(iterator_type == ITER_LT || iterator_type == ITER_LE);
- /**
- * Case 1. part_count == 1, looking for [10]. ranges:
- * {1, 3, 5} {7, 8, 9} {10, 15 20} {22, 32, 42}
- * ^looking for this
- * Case 2. part_count == 1, looking for [10]. ranges:
- * {1, 2, 4} {5, 6, 7, 8} {50, 100, 200}
- * ^looking for this
- * Case 3. part_count == 2, looking for [10]. ranges:
- * {[1, 2], [2, 3]} {[9, 1], [10, 1], [10 2], [11 3]} {[12,..}
- * ^looking for this
- * Case 4. part_count == 2, looking for [10]. ranges:
- * {[1, 2], [10, 1]} {[10, 2] [10 3] [11 3]} {[12, 1]..}
- * ^looking for this
- * Case 5. part_count does not matter, looking for [10].
- * ranges:
- * {1, 2}, {3, 4, ..}
- * ^looking for this
- */
- /**
- * vy_range_tree_nsearch finds most range with begin == key
- * or next if equal was not found
- */
- range = vy_range_tree_nsearch(tree, key);
- if (range != NULL) {
- /* fix curr_range for cases 2 and 3 */
- if (range->begin != NULL &&
- vy_stmt_compare_with_key(key, range->begin,
- range->key_def) != 0) {
- struct vy_range *prev;
- prev = vy_range_tree_prev(tree,
- range);
- if (prev != NULL)
- range = prev;
- }
- } else {
- /* Case 5 */
- range = vy_range_tree_last(tree);
- }
- }
- /* Range tree must span all possible keys. */
- assert(range != NULL);
- return range;
-}
-
-/**
- * Iterate to the next range. The next range is returned in @result.
- */
-static void
-vy_range_iterator_next(struct vy_range_iterator *itr, struct vy_range **result)
-{
- struct vy_range *curr = itr->curr_range;
- struct vy_range *next;
-
- if (curr == NULL) {
- /* First iteration */
- next = vy_range_tree_find_by_key(itr->tree, itr->iterator_type,
- itr->key);
- goto out;
- }
- switch (itr->iterator_type) {
- case ITER_LT:
- case ITER_LE:
- next = vy_range_tree_prev(itr->tree, curr);
- break;
- case ITER_GT:
- case ITER_GE:
- next = vy_range_tree_next(itr->tree, curr);
- break;
- case ITER_EQ:
- if (curr->end != NULL &&
- vy_stmt_compare_with_key(itr->key, curr->end,
- curr->key_def) >= 0) {
- /* A partial key can be found in more than one range. */
- next = vy_range_tree_next(itr->tree, curr);
- } else {
- next = NULL;
- }
- break;
- default:
- unreachable();
- }
-out:
- *result = itr->curr_range = next;
-}
-
-/**
- * Position iterator @itr to the range that contains @last_stmt and
- * return the current range in @result. If @last_stmt is NULL, restart
- * the iterator.
- */
-static void
-vy_range_iterator_restore(struct vy_range_iterator *itr,
- const struct tuple *last_stmt,
- struct vy_range **result)
-{
- struct vy_range *curr = vy_range_tree_find_by_key(itr->tree,
- itr->iterator_type,
- last_stmt != NULL ? last_stmt : itr->key);
- *result = itr->curr_range = curr;
-}
-
static void
vy_index_add_range(struct vy_index *index, struct vy_range *range)
{
@@ -1713,43 +1349,6 @@ vy_index_remove_range(struct vy_index *index, struct vy_range *range)
index->range_count--;
}
-/**
- * Allocate and initialize a range (either a new one or for
- * restore from disk).
- *
- * @param id Range id.
- * @param begin Range begin (inclusive) or NULL for -inf.
- * @param end Range end (exclusive) or NULL for +inf.
- * @param key_def Key definition for comparing range boundaries.
- *
- * @retval not NULL The new range.
- * @retval NULL Out of memory.
- */
-static struct vy_range *
-vy_range_new(int64_t id, struct tuple *begin, struct tuple *end,
- const struct key_def *key_def)
-{
- struct vy_range *range = (struct vy_range*) calloc(1, sizeof(*range));
- if (range == NULL) {
- diag_set(OutOfMemory, sizeof(struct vy_range), "malloc",
- "struct vy_range");
- return NULL;
- }
- range->id = id;
- if (begin != NULL) {
- tuple_ref(begin);
- range->begin = begin;
- }
- if (end != NULL) {
- tuple_ref(end);
- range->end = end;
- }
- range->key_def = key_def;
- rlist_create(&range->slices);
- range->heap_node.pos = UINT32_MAX;
- return range;
-}
-
/**
* Allocate a new active in-memory index for an index while moving
* the old one to the sealed list. Used by the dump task in order
@@ -1778,100 +1377,6 @@ vy_index_rotate_mem(struct vy_index *index)
return 0;
}
-static void
-vy_range_delete(struct vy_range *range)
-{
- if (range->begin != NULL)
- tuple_unref(range->begin);
- if (range->end != NULL)
- tuple_unref(range->end);
-
- /* Delete all run slices. */
- while (!rlist_empty(&range->slices)) {
- struct vy_slice *slice = rlist_shift_entry(&range->slices,
- struct vy_slice, in_range);
- vy_slice_delete(slice);
- }
-
- TRASH(range);
- free(range);
-}
-
-
-/**
- * Return true and set split_key accordingly if the range needs to be
- * split in two.
- *
- * - We should never split a range until it was merged at least once
- * (actually, it should be a function of run_count_per_level/number
- * of runs used for the merge: with low run_count_per_level it's more
- * than once, with high run_count_per_level it's once).
- * - We should use the last run size as the size of the range.
- * - We should split around the last run middle key.
- * - We should only split if the last run size is greater than
- * 4/3 * range_size.
- */
-static bool
-vy_range_needs_split(struct vy_range *range, const struct index_opts *opts,
- const char **p_split_key)
-{
- struct vy_slice *slice;
-
- /* The range hasn't been merged yet - too early to split it. */
- if (range->n_compactions < 1)
- return false;
-
- /* Find the oldest run. */
- assert(!rlist_empty(&range->slices));
- slice = rlist_last_entry(&range->slices, struct vy_slice, in_range);
-
- /* The range is too small to be split. */
- if (slice->count.bytes_compressed < opts->range_size * 4 / 3)
- return false;
-
- /* Find the median key in the oldest run (approximately). */
- struct vy_page_info *mid_page;
- mid_page = vy_run_page_info(slice->run, slice->first_page_no +
- (slice->last_page_no -
- slice->first_page_no) / 2);
-
- struct vy_page_info *first_page = vy_run_page_info(slice->run,
- slice->first_page_no);
-
- /* No point in splitting if a new range is going to be empty. */
- if (key_compare(first_page->min_key, mid_page->min_key,
- range->key_def) == 0)
- return false;
- /*
- * In extreme cases the median key can be < the beginning
- * of the slice, e.g.
- *
- * RUN:
- * ... |---- page N ----|-- page N + 1 --|-- page N + 2 --
- * | min_key = [10] | min_key = [50] | min_key = [100]
- *
- * SLICE:
- * begin = [30], end = [70]
- * first_page_no = N, last_page_no = N + 1
- *
- * which makes mid_page_no = N and mid_page->min_key = [10].
- *
- * In such cases there's no point in splitting the range.
- */
- if (slice->begin != NULL && key_compare(mid_page->min_key,
- tuple_data(slice->begin), range->key_def) <= 0)
- return false;
- /*
- * The median key can't be >= the end of the slice as we
- * take the min key of a page for the median key.
- */
- assert(slice->end == NULL || key_compare(mid_page->min_key,
- tuple_data(slice->end), range->key_def) < 0);
-
- *p_split_key = mid_page->min_key;
- return true;
-}
-
/**
* Split a range if it has grown too big, return true if the range
* was split. Splitting is done by making slices of the runs used
@@ -1991,157 +1496,6 @@ vy_index_split_range(struct vy_index *index, struct vy_range *range)
return false;
}
-/**
- * To reduce write amplification caused by compaction, we follow
- * the LSM tree design. Runs in each range are divided into groups
- * called levels:
- *
- * level 1: runs 1 .. L_1
- * level 2: runs L_1 + 1 .. L_2
- * ...
- * level N: runs L_{N-1} .. L_N
- *
- * where L_N is the total number of runs, N is the total number of
- * levels, older runs have greater numbers. Runs at each subsequent
- * are run_size_ratio times larger than on the previous one. When
- * the number of runs at a level exceeds run_count_per_level, we
- * compact all its runs along with all runs from the upper levels
- * and in-memory indexes. Including previous levels into
- * compaction is relatively cheap, because of the level size
- * ratio.
- *
- * Given a range, this function computes the maximal level that needs
- * to be compacted and sets @compact_priority to the number of runs in
- * this level and all preceding levels.
- */
-static void
-vy_range_update_compact_priority(struct vy_range *range,
- const struct index_opts *opts)
-{
- assert(opts->run_count_per_level > 0);
- assert(opts->run_size_ratio > 1);
-
- range->compact_priority = 0;
-
- /* Total number of checked runs. */
- uint32_t total_run_count = 0;
- /* The total size of runs checked so far. */
- uint64_t total_size = 0;
- /* Estimated size of a compacted run, if compaction is scheduled. */
- uint64_t est_new_run_size = 0;
- /* The number of runs at the current level. */
- uint32_t level_run_count = 0;
- /*
- * The target (perfect) size of a run at the current level.
- * For the first level, it's the size of the newest run.
- * For lower levels it's computed as first level run size
- * times run_size_ratio.
- */
- uint64_t target_run_size = 0;
-
- struct vy_slice *slice;
- rlist_foreach_entry(slice, &range->slices, in_range) {
- uint64_t size = slice->count.bytes_compressed;
- /*
- * The size of the first level is defined by
- * the size of the most recent run.
- */
- if (target_run_size == 0)
- target_run_size = size;
- total_size += size;
- level_run_count++;
- total_run_count++;
- while (size > target_run_size) {
- /*
- * The run size exceeds the threshold
- * set for the current level. Move this
- * run down to a lower level. Switch the
- * current level and reset the level run
- * count.
- */
- level_run_count = 1;
- /*
- * If we have already scheduled
- * a compaction of an upper level, and
- * estimated compacted run will end up at
- * this level, include the new run into
- * this level right away to avoid
- * a cascading compaction.
- */
- if (est_new_run_size > target_run_size)
- level_run_count++;
- /*
- * Calculate the target run size for this
- * level.
- */
- target_run_size *= opts->run_size_ratio;
- /*
- * Keep pushing the run down until
- * we find an appropriate level for it.
- */
- }
- if (level_run_count > opts->run_count_per_level) {
- /*
- * The number of runs at the current level
- * exceeds the configured maximum. Arrange
- * for compaction. We compact all runs at
- * this level and upper levels.
- */
- range->compact_priority = total_run_count;
- est_new_run_size = total_size;
- }
- }
-}
-
-/**
- * Check if a range should be coalesced with one or more its neighbors.
- * If it should, return true and set @p_first and @p_last to the first
- * and last ranges to coalesce, otherwise return false.
- *
- * We coalesce ranges together when they become too small, less than
- * half the target range size to avoid split-coalesce oscillations.
- */
-static bool
-vy_range_needs_coalesce(struct vy_range *range, vy_range_tree_t *tree,
- const struct index_opts *opts,
- struct vy_range **p_first, struct vy_range **p_last)
-{
- struct vy_range *it;
-
- /* Size of the coalesced range. */
- uint64_t total_size = range->count.bytes_compressed;
- /* Coalesce ranges until total_size > max_size. */
- uint64_t max_size = opts->range_size / 2;
-
- /*
- * We can't coalesce a range that was scheduled for dump
- * or compaction, because it is about to be processed by
- * a worker thread.
- */
- assert(!vy_range_is_scheduled(range));
-
- *p_first = *p_last = range;
- for (it = vy_range_tree_next(tree, range);
- it != NULL && !vy_range_is_scheduled(it);
- it = vy_range_tree_next(tree, it)) {
- uint64_t size = it->count.bytes_compressed;
- if (total_size + size > max_size)
- break;
- total_size += size;
- *p_last = it;
- }
- for (it = vy_range_tree_prev(tree, range);
- it != NULL && !vy_range_is_scheduled(it);
- it = vy_range_tree_prev(tree, it)) {
- uint64_t size = it->count.bytes_compressed;
- if (total_size + size > max_size)
- break;
- total_size += size;
- *p_first = it;
- }
- return *p_first != *p_last;
-}
-
/**
* Coalesce a range with one or more its neighbors if it is too small,
* return true if the range was coalesced. We coalesce ranges by
diff --git a/src/box/vy_range.c b/src/box/vy_range.c
new file mode 100644
index 0000000000..3a7401d962
--- /dev/null
+++ b/src/box/vy_range.c
@@ -0,0 +1,549 @@
+/*
+ * Copyright 2010-2017, Tarantool AUTHORS, please see AUTHORS file.
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the
+ * following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY AUTHORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
+ * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+ * AUTHORS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
+ * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+ * THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+#include "vy_range.h"
+
+#include <assert.h>
+#include <stdbool.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+#define RB_COMPACT 1
+#include <small/rb.h>
+#include <small/rlist.h>
+
+#include "diag.h"
+#include "iterator_type.h"
+#include "key_def.h"
+#include "trivia/util.h"
+#include "tuple.h"
+#include "tuple_compare.h"
+#include "vy_run.h"
+#include "vy_stat.h"
+#include "vy_stmt.h"
+
+int
+vy_range_tree_cmp(struct vy_range *range_a, struct vy_range *range_b)
+{
+ if (range_a == range_b)
+ return 0;
+
+ /* Any key > -inf. */
+ if (range_a->begin == NULL)
+ return -1;
+ if (range_b->begin == NULL)
+ return 1;
+
+ assert(range_a->key_def == range_b->key_def);
+ return vy_key_compare(range_a->begin, range_b->begin,
+ range_a->key_def);
+}
+
+int
+vy_range_tree_key_cmp(const struct tuple *stmt, struct vy_range *range)
+{
+ /* Any key > -inf. */
+ if (range->begin == NULL)
+ return 1;
+ return vy_stmt_compare_with_key(stmt, range->begin, range->key_def);
+}
+
+struct vy_range *
+vy_range_tree_find_by_key(vy_range_tree_t *tree,
+ enum iterator_type iterator_type,
+ const struct tuple *key)
+{
+ uint32_t key_field_count = tuple_field_count(key);
+ if (key_field_count == 0) {
+ switch (iterator_type) {
+ case ITER_LT:
+ case ITER_LE:
+ return vy_range_tree_last(tree);
+ case ITER_GT:
+ case ITER_GE:
+ case ITER_EQ:
+ return vy_range_tree_first(tree);
+ default:
+ unreachable();
+ return NULL;
+ }
+ }
+ struct vy_range *range;
+ if (iterator_type == ITER_GE || iterator_type == ITER_GT ||
+ iterator_type == ITER_EQ) {
+ /**
+ * Case 1. part_count == 1, looking for [10]. ranges:
+ * {1, 3, 5} {7, 8, 9} {10, 15 20} {22, 32, 42}
+ * ^looking for this
+ * Case 2. part_count == 1, looking for [10]. ranges:
+ * {1, 2, 4} {5, 6, 7, 8} {50, 100, 200}
+ * ^looking for this
+ * Case 3. part_count == 2, looking for [10]. ranges:
+ * {[1, 2], [2, 3]} {[9, 1], [10, 1], [10 2], [11 3]} {[12,..}
+ * ^looking for this
+ * Case 4. part_count == 2, looking for [10]. ranges:
+ * {[1, 2], [10, 1]} {[10, 2] [10 3] [11 3]} {[12, 1]..}
+ * ^looking for this
+ * Case 5. part_count does not matter, looking for [10].
+ * ranges:
+ * {100, 200}, {300, 400}
+ * ^looking for this
+ */
+ /**
+ * vy_range_tree_psearch finds least range with begin == key
+ * or previous if equal was not found
+ */
+ range = vy_range_tree_psearch(tree, key);
+ /* switch to previous for case (4) */
+ if (range != NULL && range->begin != NULL &&
+ key_field_count < range->key_def->part_count &&
+ vy_stmt_compare_with_key(key, range->begin,
+ range->key_def) == 0)
+ range = vy_range_tree_prev(tree, range);
+ /* for case 5 or subcase of case 4 */
+ if (range == NULL)
+ range = vy_range_tree_first(tree);
+ } else {
+ assert(iterator_type == ITER_LT || iterator_type == ITER_LE);
+ /**
+ * Case 1. part_count == 1, looking for [10]. ranges:
+ * {1, 3, 5} {7, 8, 9} {10, 15 20} {22, 32, 42}
+ * ^looking for this
+ * Case 2. part_count == 1, looking for [10]. ranges:
+ * {1, 2, 4} {5, 6, 7, 8} {50, 100, 200}
+ * ^looking for this
+ * Case 3. part_count == 2, looking for [10]. ranges:
+ * {[1, 2], [2, 3]} {[9, 1], [10, 1], [10 2], [11 3]} {[12,..}
+ * ^looking for this
+ * Case 4. part_count == 2, looking for [10]. ranges:
+ * {[1, 2], [10, 1]} {[10, 2] [10 3] [11 3]} {[12, 1]..}
+ * ^looking for this
+ * Case 5. part_count does not matter, looking for [10].
+ * ranges:
+ * {1, 2}, {3, 4, ..}
+ * ^looking for this
+ */
+ /**
+ * vy_range_tree_nsearch finds most range with begin == key
+ * or next if equal was not found
+ */
+ range = vy_range_tree_nsearch(tree, key);
+ if (range != NULL) {
+ /* fix curr_range for cases 2 and 3 */
+ if (range->begin != NULL &&
+ vy_stmt_compare_with_key(key, range->begin,
+ range->key_def) != 0) {
+ struct vy_range *prev;
+ prev = vy_range_tree_prev(tree, range);
+ if (prev != NULL)
+ range = prev;
+ }
+ } else {
+ /* Case 5 */
+ range = vy_range_tree_last(tree);
+ }
+ }
+ return range;
+}
+
+struct vy_range *
+vy_range_new(int64_t id, struct tuple *begin, struct tuple *end,
+ const struct key_def *key_def)
+{
+ struct vy_range *range = calloc(1, sizeof(*range));
+ if (range == NULL) {
+ diag_set(OutOfMemory, sizeof(*range),
+ "malloc", "struct vy_range");
+ return NULL;
+ }
+ range->id = id;
+ if (begin != NULL) {
+ tuple_ref(begin);
+ range->begin = begin;
+ }
+ if (end != NULL) {
+ tuple_ref(end);
+ range->end = end;
+ }
+ range->key_def = key_def;
+ rlist_create(&range->slices);
+ range->heap_node.pos = UINT32_MAX;
+ return range;
+}
+
+void
+vy_range_delete(struct vy_range *range)
+{
+ if (range->begin != NULL)
+ tuple_unref(range->begin);
+ if (range->end != NULL)
+ tuple_unref(range->end);
+
+ struct vy_slice *slice, *next_slice;
+ rlist_foreach_entry_safe(slice, &range->slices, in_range, next_slice)
+ vy_slice_delete(slice);
+
+ TRASH(range);
+ free(range);
+}
+
+int
+vy_range_snprint(char *buf, int size, const struct vy_range *range)
+{
+ int total = 0;
+ SNPRINT(total, snprintf, buf, size, "(");
+ if (range->begin != NULL)
+ SNPRINT(total, vy_key_snprint, buf, size,
+ tuple_data(range->begin));
+ else
+ SNPRINT(total, snprintf, buf, size, "-inf");
+ SNPRINT(total, snprintf, buf, size, "..");
+ if (range->end != NULL)
+ SNPRINT(total, vy_key_snprint, buf, size,
+ tuple_data(range->end));
+ else
+ SNPRINT(total, snprintf, buf, size, "inf");
+ SNPRINT(total, snprintf, buf, size, ")");
+ return total;
+}
+
+void
+vy_range_add_slice(struct vy_range *range, struct vy_slice *slice)
+{
+ rlist_add_entry(&range->slices, slice, in_range);
+ range->slice_count++;
+ vy_disk_stmt_counter_add(&range->count, &slice->count);
+}
+
+void
+vy_range_add_slice_before(struct vy_range *range, struct vy_slice *slice,
+ struct vy_slice *next_slice)
+{
+ rlist_add_tail(&next_slice->in_range, &slice->in_range);
+ range->slice_count++;
+ vy_disk_stmt_counter_add(&range->count, &slice->count);
+}
+
+void
+vy_range_remove_slice(struct vy_range *range, struct vy_slice *slice)
+{
+ assert(range->slice_count > 0);
+ assert(!rlist_empty(&range->slices));
+ rlist_del_entry(slice, in_range);
+ range->slice_count--;
+ vy_disk_stmt_counter_sub(&range->count, &slice->count);
+}
+
+/**
+ * To reduce write amplification caused by compaction, we follow
+ * the LSM tree design. Runs in each range are divided into groups
+ * called levels:
+ *
+ * level 1: runs 1 .. L_1
+ * level 2: runs L_1 + 1 .. L_2
+ * ...
+ * level N: runs L_{N-1} .. L_N
+ *
+ * where L_N is the total number of runs, N is the total number of
+ * levels, older runs have greater numbers. Runs at each subsequent
+ * are run_size_ratio times larger than on the previous one. When
+ * the number of runs at a level exceeds run_count_per_level, we
+ * compact all its runs along with all runs from the upper levels
+ * and in-memory indexes. Including previous levels into
+ * compaction is relatively cheap, because of the level size
+ * ratio.
+ *
+ * Given a range, this function computes the maximal level that needs
+ * to be compacted and sets @compact_priority to the number of runs in
+ * this level and all preceding levels.
+ */
+void
+vy_range_update_compact_priority(struct vy_range *range,
+ const struct index_opts *opts)
+{
+ assert(opts->run_count_per_level > 0);
+ assert(opts->run_size_ratio > 1);
+
+ range->compact_priority = 0;
+
+ /* Total number of checked runs. */
+ uint32_t total_run_count = 0;
+ /* The total size of runs checked so far. */
+ uint64_t total_size = 0;
+ /* Estimated size of a compacted run, if compaction is scheduled. */
+ uint64_t est_new_run_size = 0;
+ /* The number of runs at the current level. */
+ uint32_t level_run_count = 0;
+ /*
+ * The target (perfect) size of a run at the current level.
+ * For the first level, it's the size of the newest run.
+ * For lower levels it's computed as first level run size
+ * times run_size_ratio.
+ */
+ uint64_t target_run_size = 0;
+
+ struct vy_slice *slice;
+ rlist_foreach_entry(slice, &range->slices, in_range) {
+ uint64_t size = slice->count.bytes_compressed;
+ /*
+ * The size of the first level is defined by
+ * the size of the most recent run.
+ */
+ if (target_run_size == 0)
+ target_run_size = size;
+ total_size += size;
+ level_run_count++;
+ total_run_count++;
+ while (size > target_run_size) {
+ /*
+ * The run size exceeds the threshold
+ * set for the current level. Move this
+ * run down to a lower level. Switch the
+ * current level and reset the level run
+ * count.
+ */
+ level_run_count = 1;
+ /*
+ * If we have already scheduled
+ * a compaction of an upper level, and
+ * estimated compacted run will end up at
+ * this level, include the new run into
+ * this level right away to avoid
+ * a cascading compaction.
+ */
+ if (est_new_run_size > target_run_size)
+ level_run_count++;
+ /*
+ * Calculate the target run size for this
+ * level.
+ */
+ target_run_size *= opts->run_size_ratio;
+ /*
+ * Keep pushing the run down until
+ * we find an appropriate level for it.
+ */
+ }
+ if (level_run_count > opts->run_count_per_level) {
+ /*
+ * The number of runs at the current level
+ * exceeds the configured maximum. Arrange
+ * for compaction. We compact all runs at
+ * this level and upper levels.
+ */
+ range->compact_priority = total_run_count;
+ est_new_run_size = total_size;
+ }
+ }
+}
+
+/**
+ * Return true and set split_key accordingly if the range needs to be
+ * split in two.
+ *
+ * - We should never split a range until it was merged at least once
+ * (actually, it should be a function of run_count_per_level/number
+ * of runs used for the merge: with low run_count_per_level it's more
+ * than once, with high run_count_per_level it's once).
+ * - We should use the last run size as the size of the range.
+ * - We should split around the last run middle key.
+ * - We should only split if the last run size is greater than
+ * 4/3 * range_size.
+ */
+bool
+vy_range_needs_split(struct vy_range *range, const struct index_opts *opts,
+ const char **p_split_key)
+{
+ struct vy_slice *slice;
+
+ /* The range hasn't been merged yet - too early to split it. */
+ if (range->n_compactions < 1)
+ return false;
+
+ /* Find the oldest run. */
+ assert(!rlist_empty(&range->slices));
+ slice = rlist_last_entry(&range->slices, struct vy_slice, in_range);
+
+ /* The range is too small to be split. */
+ if (slice->count.bytes_compressed < opts->range_size * 4 / 3)
+ return false;
+
+ /* Find the median key in the oldest run (approximately). */
+ struct vy_page_info *mid_page;
+ mid_page = vy_run_page_info(slice->run, slice->first_page_no +
+ (slice->last_page_no -
+ slice->first_page_no) / 2);
+
+ struct vy_page_info *first_page = vy_run_page_info(slice->run,
+ slice->first_page_no);
+
+ /* No point in splitting if a new range is going to be empty. */
+ if (key_compare(first_page->min_key, mid_page->min_key,
+ range->key_def) == 0)
+ return false;
+ /*
+ * In extreme cases the median key can be < the beginning
+ * of the slice, e.g.
+ *
+ * RUN:
+ * ... |---- page N ----|-- page N + 1 --|-- page N + 2 --
+ * | min_key = [10] | min_key = [50] | min_key = [100]
+ *
+ * SLICE:
+ * begin = [30], end = [70]
+ * first_page_no = N, last_page_no = N + 1
+ *
+ * which makes mid_page_no = N and mid_page->min_key = [10].
+ *
+ * In such cases there's no point in splitting the range.
+ */
+ if (slice->begin != NULL && key_compare(mid_page->min_key,
+ tuple_data(slice->begin), range->key_def) <= 0)
+ return false;
+ /*
+ * The median key can't be >= the end of the slice as we
+ * take the min key of a page for the median key.
+ */
+ assert(slice->end == NULL || key_compare(mid_page->min_key,
+ tuple_data(slice->end), range->key_def) < 0);
+
+ *p_split_key = mid_page->min_key;
+ return true;
+}
+
+/**
+ * Check if a range should be coalesced with one or more its neighbors.
+ * If it should, return true and set @p_first and @p_last to the first
+ * and last ranges to coalesce, otherwise return false.
+ *
+ * We coalesce ranges together when they become too small, less than
+ * half the target range size to avoid split-coalesce oscillations.
+ */
+bool
+vy_range_needs_coalesce(struct vy_range *range, vy_range_tree_t *tree,
+ const struct index_opts *opts,
+ struct vy_range **p_first, struct vy_range **p_last)
+{
+ struct vy_range *it;
+
+ /* Size of the coalesced range. */
+ uint64_t total_size = range->count.bytes_compressed;
+ /* Coalesce ranges until total_size > max_size. */
+ uint64_t max_size = opts->range_size / 2;
+
+ /*
+ * We can't coalesce a range that was scheduled for dump
+ * or compaction, because it is about to be processed by
+ * a worker thread.
+ */
+ assert(!vy_range_is_scheduled(range));
+
+ *p_first = *p_last = range;
+ for (it = vy_range_tree_next(tree, range);
+ it != NULL && !vy_range_is_scheduled(it);
+ it = vy_range_tree_next(tree, it)) {
+ uint64_t size = it->count.bytes_compressed;
+ if (total_size + size > max_size)
+ break;
+ total_size += size;
+ *p_last = it;
+ }
+ for (it = vy_range_tree_prev(tree, range);
+ it != NULL && !vy_range_is_scheduled(it);
+ it = vy_range_tree_prev(tree, it)) {
+ uint64_t size = it->count.bytes_compressed;
+ if (total_size + size > max_size)
+ break;
+ total_size += size;
+ *p_first = it;
+ }
+ return *p_first != *p_last;
+}
+
+void
+vy_range_iterator_open(struct vy_range_iterator *itr, vy_range_tree_t *tree,
+ enum iterator_type iterator_type,
+ const struct tuple *key)
+{
+ itr->tree = tree;
+ itr->iterator_type = iterator_type;
+ itr->key = key;
+ itr->curr_range = NULL;
+}
+
+void
+vy_range_iterator_next(struct vy_range_iterator *itr, struct vy_range **result)
+{
+ struct vy_range *curr = itr->curr_range;
+ struct vy_range *next;
+
+ if (curr == NULL) {
+ /* First iteration */
+ next = vy_range_tree_find_by_key(itr->tree, itr->iterator_type,
+ itr->key);
+ goto out;
+ }
+ switch (itr->iterator_type) {
+ case ITER_LT:
+ case ITER_LE:
+ next = vy_range_tree_prev(itr->tree, curr);
+ break;
+ case ITER_GT:
+ case ITER_GE:
+ next = vy_range_tree_next(itr->tree, curr);
+ break;
+ case ITER_EQ:
+ if (curr->end != NULL &&
+ vy_stmt_compare_with_key(itr->key, curr->end,
+ curr->key_def) >= 0) {
+ /* A partial key can be found in more than one range. */
+ next = vy_range_tree_next(itr->tree, curr);
+ } else {
+ next = NULL;
+ }
+ break;
+ default:
+ unreachable();
+ }
+out:
+ *result = itr->curr_range = next;
+}
+
+void
+vy_range_iterator_restore(struct vy_range_iterator *itr,
+ const struct tuple *last_stmt,
+ struct vy_range **result)
+{
+ struct vy_range *curr = vy_range_tree_find_by_key(itr->tree,
+ itr->iterator_type,
+ last_stmt != NULL ? last_stmt : itr->key);
+ *result = itr->curr_range = curr;
+}
diff --git a/src/box/vy_range.h b/src/box/vy_range.h
new file mode 100644
index 0000000000..8bb6849a7f
--- /dev/null
+++ b/src/box/vy_range.h
@@ -0,0 +1,321 @@
+#ifndef INCLUDES_TARANTOOL_BOX_VY_RANGE_H
+#define INCLUDES_TARANTOOL_BOX_VY_RANGE_H
+/*
+ * Copyright 2010-2017, Tarantool AUTHORS, please see AUTHORS file.
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the
+ * following disclaimer.
+ *
+ * 2. Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY AUTHORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
+ * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
+ * AUTHORS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT,
+ * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR
+ * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+ * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF
+ * THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <stdbool.h>
+#include <stdint.h>
+
+#define RB_COMPACT 1
+#include <small/rb.h>
+#include <small/rlist.h>
+
+#include "iterator_type.h"
+#define HEAP_FORWARD_DECLARATION
+#include "salad/heap.h"
+#include "trivia/util.h"
+#include "vy_stat.h"
+
+#if defined(__cplusplus)
+extern "C" {
+#endif /* defined(__cplusplus) */
+
+struct index_opts;
+struct key_def;
+struct tuple;
+struct vy_slice;
+
+/**
+ * Range of keys in an index stored on disk.
+ */
+struct vy_range {
+ /** Unique ID of this range. */
+ int64_t id;
+ /**
+ * Range lower bound. NULL if range is leftmost.
+ * Both 'begin' and 'end' statements have SELECT type with
+ * the full idexed key.
+ */
+ struct tuple *begin;
+ /** Range upper bound. NULL if range is rightmost. */
+ struct tuple *end;
+ /** Key definition for comparing range boundaries. */
+ const struct key_def *key_def;
+ /** An estimate of the number of statements in this range. */
+ struct vy_disk_stmt_counter count;
+ /**
+ * List of run slices in this range, linked by vy_slice->in_range.
+ * The newer a slice, the closer it to the list head.
+ */
+ struct rlist slices;
+ /** Number of entries in the ->slices list. */
+ int slice_count;
+ /**
+ * The goal of compaction is to reduce read amplification.
+ * All ranges for which the LSM tree has more runs per
+ * level than run_count_per_level or run size larger than
+ * one defined by run_size_ratio of this level are candidates
+ * for compaction.
+ * Unlike other LSM implementations, Vinyl can have many
+ * sorted runs in a single level, and is able to compact
+ * runs from any number of adjacent levels. Moreover,
+ * higher levels are always taken in when compacting
+ * a lower level - i.e. L1 is always included when
+ * compacting L2, and both L1 and L2 are always included
+ * when compacting L3.
+ *
+ * This variable contains the number of runs the next
+ * compaction of this range will include.
+ *
+ * The lower the level is scheduled for compaction,
+ * the bigger it tends to be because upper levels are
+ * taken in.
+ * @sa vy_range_update_compact_priority() to see
+ * how we decide how many runs to compact next time.
+ */
+ int compact_priority;
+ /** Number of times the range was compacted. */
+ int n_compactions;
+ /** Link in vy_index->tree. */
+ rb_node(struct vy_range) tree_node;
+ /** Link in vy_index->range_heap. */
+ struct heap_node heap_node;
+ /**
+ * Incremented whenever an in-memory index or on disk
+ * run is added to or deleted from this range. Used to
+ * invalidate iterators.
+ */
+ uint32_t version;
+};
+
+/**
+ * Heap of all ranges of the same index, prioritized by
+ * vy_range->compact_priority.
+ */
+#define HEAP_NAME vy_range_heap
+static inline bool
+vy_range_heap_less(struct heap_node *a, struct heap_node *b)
+{
+ struct vy_range *r1 = container_of(a, struct vy_range, heap_node);
+ struct vy_range *r2 = container_of(b, struct vy_range, heap_node);
+ return r1->compact_priority > r2->compact_priority;
+}
+#define HEAP_LESS(h, l, r) vy_range_heap_less(l, r)
+#include "salad/heap.h"
+#undef HEAP_LESS
+#undef HEAP_NAME
+
+/** Return true if a task is scheduled for a given range. */
+static inline bool
+vy_range_is_scheduled(struct vy_range *range)
+{
+ return range->heap_node.pos == UINT32_MAX;
+}
+
+/**
+ * Search tree of all ranges of the same index, sorted by
+ * vy_range->begin. Ranges in a tree are supposed to span
+ * all possible keys without overlaps.
+ */
+int
+vy_range_tree_cmp(struct vy_range *range_a, struct vy_range *range_b);
+int
+vy_range_tree_key_cmp(const struct tuple *stmt, struct vy_range *range);
+
+typedef rb_tree(struct vy_range) vy_range_tree_t;
+rb_gen_ext_key(MAYBE_UNUSED static inline, vy_range_tree_, vy_range_tree_t,
+ struct vy_range, tree_node, vy_range_tree_cmp,
+ const struct tuple *, vy_range_tree_key_cmp);
+
+/**
+ * Find the first range in which a given key should be looked up.
+ *
+ * @param tree Range tree to search.
+ * @param iterator_type Iterator type.
+ * @param key Key to look up.
+ *
+ * @retval The first range to look up the key in.
+ */
+struct vy_range *
+vy_range_tree_find_by_key(vy_range_tree_t *tree,
+ enum iterator_type iterator_type,
+ const struct tuple *key);
+
+/**
+ * Allocate and initialize a range (either a new one or for
+ * restore from disk).
+ *
+ * @param id Range id.
+ * @param begin Range begin (inclusive) or NULL for -inf.
+ * @param end Range end (exclusive) or NULL for +inf.
+ * @param key_def Key definition for comparing range boundaries.
+ *
+ * @retval not NULL The new range.
+ * @retval NULL Out of memory.
+ */
+struct vy_range *
+vy_range_new(int64_t id, struct tuple *begin, struct tuple *end,
+ const struct key_def *key_def);
+
+/**
+ * Free a range and all its slices.
+ *
+ * @param range Range to free.
+ */
+void
+vy_range_delete(struct vy_range *range);
+
+/** An snprint-style function to print boundaries of a range. */
+int
+vy_range_snprint(char *buf, int size, const struct vy_range *range);
+
+static inline const char *
+vy_range_str(struct vy_range *range)
+{
+ char *buf = tt_static_buf();
+ vy_range_snprint(buf, TT_STATIC_BUF_LEN, range);
+ return buf;
+}
+
+/** Add a run slice to the head of a range's list. */
+void
+vy_range_add_slice(struct vy_range *range, struct vy_slice *slice);
+
+/** Add a run slice to a range's list before @next_slice. */
+void
+vy_range_add_slice_before(struct vy_range *range, struct vy_slice *slice,
+ struct vy_slice *next_slice);
+
+/** Remove a run slice from a range's list. */
+void
+vy_range_remove_slice(struct vy_range *range, struct vy_slice *slice);
+
+/**
+ * Update compaction priority of a range.
+ *
+ * @param range The range.
+ * @param opts Index options.
+ */
+void
+vy_range_update_compact_priority(struct vy_range *range,
+ const struct index_opts *opts);
+
+/**
+ * Check if a range needs to be split in two.
+ *
+ * @param range The range.
+ * @param opts Index options.
+ * @param[out] p_split_key Key to split the range by.
+ *
+ * @retval true If the range needs to be split.
+ */
+bool
+vy_range_needs_split(struct vy_range *range, const struct index_opts *opts,
+ const char **p_split_key);
+
+/**
+ * Check if a range needs to be coalesced with adjacent
+ * ranges in a range tree.
+ *
+ * @param range The range.
+ * @param tree The range tree.
+ * @param opts Index options.
+ * @param[out] p_first The first range in the tree to coalesce.
+ * @param[out] p_last The last range in the tree to coalesce.
+ *
+ * @retval true If the range needs to be coalesced.
+ */
+bool
+vy_range_needs_coalesce(struct vy_range *range, vy_range_tree_t *tree,
+ const struct index_opts *opts,
+ struct vy_range **p_first, struct vy_range **p_last);
+
+/**
+ * Iterator over ranges in a tree.
+ */
+struct vy_range_iterator {
+ /** Range tree to iterate. */
+ vy_range_tree_t *tree;
+ /** Iterator type. */
+ enum iterator_type iterator_type;
+ /** Search key. */
+ const struct tuple *key;
+ /**
+ * Current range or NULL if the iteration
+ * has stopped or has not been started.
+ */
+ struct vy_range *curr_range;
+};
+
+/**
+ * Initialize a range iterator.
+ *
+ * @param itr The iterator.
+ * @param tree Range tree to iterate.
+ * @param iterator_type Iterator type.
+ * @param key Search key.
+ */
+void
+vy_range_iterator_open(struct vy_range_iterator *itr, vy_range_tree_t *tree,
+ enum iterator_type iterator_type,
+ const struct tuple *key);
+
+/**
+ * Iterate to the next range.
+ *
+ * @param itr The iterator.
+ * @param[out] result Next range.
+ */
+void
+vy_range_iterator_next(struct vy_range_iterator *itr,
+ struct vy_range **result);
+
+/**
+ * Restore an iterator after a tree modification.
+ *
+ * @param itr The iterator.
+ * @param last_stmt The last iterated statement.
+ * @param[out] result Next range.
+ *
+ * This function positions the iterator to the range that
+ * contains @last_stmt. If @last_stmt is NULL, it restarts
+ * the iterator.
+ */
+void
+vy_range_iterator_restore(struct vy_range_iterator *itr,
+ const struct tuple *last_stmt,
+ struct vy_range **result);
+
+#if defined(__cplusplus)
+} /* extern "C" */
+#endif /* defined(__cplusplus) */
+
+#endif /* INCLUDES_TARANTOOL_BOX_VY_RANGE_H */
--
GitLab