/* * qht.c - QEMU Hash Table, designed to scale for read-mostly workloads. * * Copyright (C) 2016, Emilio G. Cota * * License: GNU GPL, version 2 or later. * See the COPYING file in the top-level directory. * * Assumptions: * - NULL cannot be inserted/removed as a pointer value. * - Trying to insert an already-existing hash-pointer pair is OK. However, * it is not OK to insert into the same hash table different hash-pointer * pairs that have the same pointer value, but not the hashes. * - Lookups are performed under an RCU read-critical section; removals * must wait for a grace period to elapse before freeing removed objects. * * Features: * - Reads (i.e. lookups and iterators) can be concurrent with other reads. * Lookups that are concurrent with writes to the same bucket will retry * via a seqlock; iterators acquire all bucket locks and therefore can be * concurrent with lookups and are serialized wrt writers. * - Writes (i.e. insertions/removals) can be concurrent with writes to * different buckets; writes to the same bucket are serialized through a lock. * - Optional auto-resizing: the hash table resizes up if the load surpasses * a certain threshold. Resizing is done concurrently with readers; writes * are serialized with the resize operation. * * The key structure is the bucket, which is cacheline-sized. Buckets * contain a few hash values and pointers; the u32 hash values are stored in * full so that resizing is fast. Having this structure instead of directly * chaining items has two advantages: * - Failed lookups fail fast, and touch a minimum number of cache lines. * - Resizing the hash table with concurrent lookups is easy. * * There are two types of buckets: * 1. "head" buckets are the ones allocated in the array of buckets in qht_map. * 2. all "non-head" buckets (i.e. all others) are members of a chain that * starts from a head bucket. * Note that the seqlock and spinlock of a head bucket applies to all buckets * chained to it; these two fields are unused in non-head buckets. * * On removals, we move the last valid item in the chain to the position of the * just-removed entry. This makes lookups slightly faster, since the moment an * invalid entry is found, the (failed) lookup is over. * * Resizing is done by taking all bucket spinlocks (so that no other writers can * race with us) and then copying all entries into a new hash map. Then, the * ht->map pointer is set, and the old map is freed once no RCU readers can see * it anymore. * * Writers check for concurrent resizes by comparing ht->map before and after * acquiring their bucket lock. If they don't match, a resize has occured * while the bucket spinlock was being acquired. * * Related Work: * - Idea of cacheline-sized buckets with full hashes taken from: * David, Guerraoui & Trigonakis, "Asynchronized Concurrency: * The Secret to Scaling Concurrent Search Data Structures", ASPLOS'15. * - Why not RCU-based hash tables? They would allow us to get rid of the * seqlock, but resizing would take forever since RCU read critical * sections in QEMU take quite a long time. * More info on relativistic hash tables: * + Triplett, McKenney & Walpole, "Resizable, Scalable, Concurrent Hash * Tables via Relativistic Programming", USENIX ATC'11. * + Corbet, "Relativistic hash tables, part 1: Algorithms", @ lwn.net, 2014. * https://lwn.net/Articles/612021/ */ #include "qemu/osdep.h" #include "qemu/qht.h" #include "qemu/atomic.h" #include "qemu/rcu.h" //#define QHT_DEBUG /* * We want to avoid false sharing of cache lines. Most systems have 64-byte * cache lines so we go with it for simplicity. * * Note that systems with smaller cache lines will be fine (the struct is * almost 64-bytes); systems with larger cache lines might suffer from * some false sharing. */ #define QHT_BUCKET_ALIGN 64 /* define these to keep sizeof(qht_bucket) within QHT_BUCKET_ALIGN */ #if HOST_LONG_BITS == 32 #define QHT_BUCKET_ENTRIES 6 #else /* 64-bit */ #define QHT_BUCKET_ENTRIES 4 #endif /* * Note: reading partially-updated pointers in @pointers could lead to * segfaults. We thus access them with atomic_read/set; this guarantees * that the compiler makes all those accesses atomic. We also need the * volatile-like behavior in atomic_read, since otherwise the compiler * might refetch the pointer. * atomic_read's are of course not necessary when the bucket lock is held. * * If both ht->lock and b->lock are grabbed, ht->lock should always * be grabbed first. */ struct qht_bucket { QemuSpin lock; QemuSeqLock sequence; uint32_t hashes[QHT_BUCKET_ENTRIES]; void *pointers[QHT_BUCKET_ENTRIES]; struct qht_bucket *next; } QEMU_ALIGNED(QHT_BUCKET_ALIGN); QEMU_BUILD_BUG_ON(sizeof(struct qht_bucket) > QHT_BUCKET_ALIGN); /** * struct qht_map - structure to track an array of buckets * @rcu: used by RCU. Keep it as the top field in the struct to help valgrind * find the whole struct. * @buckets: array of head buckets. It is constant once the map is created. * @n_buckets: number of head buckets. It is constant once the map is created. * @n_added_buckets: number of added (i.e. "non-head") buckets * @n_added_buckets_threshold: threshold to trigger an upward resize once the * number of added buckets surpasses it. * * Buckets are tracked in what we call a "map", i.e. this structure. */ struct qht_map { struct rcu_head rcu; struct qht_bucket *buckets; size_t n_buckets; size_t n_added_buckets; size_t n_added_buckets_threshold; }; /* trigger a resize when n_added_buckets > n_buckets / div */ #define QHT_NR_ADDED_BUCKETS_THRESHOLD_DIV 8 static void qht_do_resize_reset(struct qht *ht, struct qht_map *new, bool reset); static void qht_grow_maybe(struct qht *ht); #ifdef QHT_DEBUG #define qht_debug_assert(X) do { assert(X); } while (0) static void qht_bucket_debug__locked(struct qht_bucket *b) { bool seen_empty = false; bool corrupt = false; int i; do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { if (b->pointers[i] == NULL) { seen_empty = true; continue; } if (seen_empty) { fprintf(stderr, "%s: b: %p, pos: %i, hash: 0x%x, p: %p\n", __func__, b, i, b->hashes[i], b->pointers[i]); corrupt = true; } } b = b->next; } while (b); qht_debug_assert(!corrupt); } static void qht_map_debug__all_locked(struct qht_map *map) { int i; for (i = 0; i < map->n_buckets; i++) { qht_bucket_debug__locked(&map->buckets[i]); } } #else #define qht_debug_assert(X) do { (void)(X); } while (0) static inline void qht_bucket_debug__locked(struct qht_bucket *b) { } static inline void qht_map_debug__all_locked(struct qht_map *map) { } #endif /* QHT_DEBUG */ static inline size_t qht_elems_to_buckets(size_t n_elems) { return pow2ceil(n_elems / QHT_BUCKET_ENTRIES); } static inline void qht_head_init(struct qht_bucket *b) { memset(b, 0, sizeof(*b)); qemu_spin_init(&b->lock); seqlock_init(&b->sequence); } static inline struct qht_bucket *qht_map_to_bucket(struct qht_map *map, uint32_t hash) { return &map->buckets[hash & (map->n_buckets - 1)]; } /* acquire all bucket locks from a map */ static void qht_map_lock_buckets(struct qht_map *map) { size_t i; for (i = 0; i < map->n_buckets; i++) { struct qht_bucket *b = &map->buckets[i]; qemu_spin_lock(&b->lock); } } static void qht_map_unlock_buckets(struct qht_map *map) { size_t i; for (i = 0; i < map->n_buckets; i++) { struct qht_bucket *b = &map->buckets[i]; qemu_spin_unlock(&b->lock); } } /* * Call with at least a bucket lock held. * @map should be the value read before acquiring the lock (or locks). */ static inline bool qht_map_is_stale__locked(struct qht *ht, struct qht_map *map) { return map != ht->map; } /* * Grab all bucket locks, and set @pmap after making sure the map isn't stale. * * Pairs with qht_map_unlock_buckets(), hence the pass-by-reference. * * Note: callers cannot have ht->lock held. */ static inline void qht_map_lock_buckets__no_stale(struct qht *ht, struct qht_map **pmap) { struct qht_map *map; map = atomic_rcu_read(&ht->map); qht_map_lock_buckets(map); if (likely(!qht_map_is_stale__locked(ht, map))) { *pmap = map; return; } qht_map_unlock_buckets(map); /* we raced with a resize; acquire ht->lock to see the updated ht->map */ qemu_mutex_lock(&ht->lock); map = ht->map; qht_map_lock_buckets(map); qemu_mutex_unlock(&ht->lock); *pmap = map; return; } /* * Get a head bucket and lock it, making sure its parent map is not stale. * @pmap is filled with a pointer to the bucket's parent map. * * Unlock with qemu_spin_unlock(&b->lock). * * Note: callers cannot have ht->lock held. */ static inline struct qht_bucket *qht_bucket_lock__no_stale(struct qht *ht, uint32_t hash, struct qht_map **pmap) { struct qht_bucket *b; struct qht_map *map; map = atomic_rcu_read(&ht->map); b = qht_map_to_bucket(map, hash); qemu_spin_lock(&b->lock); if (likely(!qht_map_is_stale__locked(ht, map))) { *pmap = map; return b; } qemu_spin_unlock(&b->lock); /* we raced with a resize; acquire ht->lock to see the updated ht->map */ qemu_mutex_lock(&ht->lock); map = ht->map; b = qht_map_to_bucket(map, hash); qemu_spin_lock(&b->lock); qemu_mutex_unlock(&ht->lock); *pmap = map; return b; } static inline bool qht_map_needs_resize(struct qht_map *map) { return atomic_read(&map->n_added_buckets) > map->n_added_buckets_threshold; } static inline void qht_chain_destroy(struct qht_bucket *head) { struct qht_bucket *curr = head->next; struct qht_bucket *prev; while (curr) { prev = curr; curr = curr->next; qemu_vfree(prev); } } /* pass only an orphan map */ static void qht_map_destroy(struct qht_map *map) { size_t i; for (i = 0; i < map->n_buckets; i++) { qht_chain_destroy(&map->buckets[i]); } qemu_vfree(map->buckets); g_free(map); } static struct qht_map *qht_map_create(size_t n_buckets) { struct qht_map *map; size_t i; map = g_malloc(sizeof(*map)); map->n_buckets = n_buckets; map->n_added_buckets = 0; map->n_added_buckets_threshold = n_buckets / QHT_NR_ADDED_BUCKETS_THRESHOLD_DIV; /* let tiny hash tables to at least add one non-head bucket */ if (unlikely(map->n_added_buckets_threshold == 0)) { map->n_added_buckets_threshold = 1; } map->buckets = qemu_memalign(QHT_BUCKET_ALIGN, sizeof(*map->buckets) * n_buckets); for (i = 0; i < n_buckets; i++) { qht_head_init(&map->buckets[i]); } return map; } void qht_init(struct qht *ht, size_t n_elems, unsigned int mode) { struct qht_map *map; size_t n_buckets = qht_elems_to_buckets(n_elems); ht->mode = mode; qemu_mutex_init(&ht->lock); map = qht_map_create(n_buckets); atomic_rcu_set(&ht->map, map); } /* call only when there are no readers/writers left */ void qht_destroy(struct qht *ht) { qht_map_destroy(ht->map); memset(ht, 0, sizeof(*ht)); } static void qht_bucket_reset__locked(struct qht_bucket *head) { struct qht_bucket *b = head; int i; seqlock_write_begin(&head->sequence); do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { if (b->pointers[i] == NULL) { goto done; } atomic_set(&b->hashes[i], 0); atomic_set(&b->pointers[i], NULL); } b = b->next; } while (b); done: seqlock_write_end(&head->sequence); } /* call with all bucket locks held */ static void qht_map_reset__all_locked(struct qht_map *map) { size_t i; for (i = 0; i < map->n_buckets; i++) { qht_bucket_reset__locked(&map->buckets[i]); } qht_map_debug__all_locked(map); } void qht_reset(struct qht *ht) { struct qht_map *map; qht_map_lock_buckets__no_stale(ht, &map); qht_map_reset__all_locked(map); qht_map_unlock_buckets(map); } static inline void qht_do_resize(struct qht *ht, struct qht_map *new) { qht_do_resize_reset(ht, new, false); } static inline void qht_do_resize_and_reset(struct qht *ht, struct qht_map *new) { qht_do_resize_reset(ht, new, true); } bool qht_reset_size(struct qht *ht, size_t n_elems) { struct qht_map *new = NULL; struct qht_map *map; size_t n_buckets; n_buckets = qht_elems_to_buckets(n_elems); qemu_mutex_lock(&ht->lock); map = ht->map; if (n_buckets != map->n_buckets) { new = qht_map_create(n_buckets); } qht_do_resize_and_reset(ht, new); qemu_mutex_unlock(&ht->lock); return !!new; } static inline void *qht_do_lookup(struct qht_bucket *head, qht_lookup_func_t func, const void *userp, uint32_t hash) { struct qht_bucket *b = head; int i; do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { if (atomic_read(&b->hashes[i]) == hash) { /* The pointer is dereferenced before seqlock_read_retry, * so (unlike qht_insert__locked) we need to use * atomic_rcu_read here. */ void *p = atomic_rcu_read(&b->pointers[i]); if (likely(p) && likely(func(p, userp))) { return p; } } } b = atomic_rcu_read(&b->next); } while (b); return NULL; } static __attribute__((noinline)) void *qht_lookup__slowpath(struct qht_bucket *b, qht_lookup_func_t func, const void *userp, uint32_t hash) { unsigned int version; void *ret; do { version = seqlock_read_begin(&b->sequence); ret = qht_do_lookup(b, func, userp, hash); } while (seqlock_read_retry(&b->sequence, version)); return ret; } void *qht_lookup(struct qht *ht, qht_lookup_func_t func, const void *userp, uint32_t hash) { struct qht_bucket *b; struct qht_map *map; unsigned int version; void *ret; map = atomic_rcu_read(&ht->map); b = qht_map_to_bucket(map, hash); version = seqlock_read_begin(&b->sequence); ret = qht_do_lookup(b, func, userp, hash); if (likely(!seqlock_read_retry(&b->sequence, version))) { return ret; } /* * Removing the do/while from the fastpath gives a 4% perf. increase when * running a 100%-lookup microbenchmark. */ return qht_lookup__slowpath(b, func, userp, hash); } /* call with head->lock held */ static bool qht_insert__locked(struct qht *ht, struct qht_map *map, struct qht_bucket *head, void *p, uint32_t hash, bool *needs_resize) { struct qht_bucket *b = head; struct qht_bucket *prev = NULL; struct qht_bucket *new = NULL; int i; do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { if (b->pointers[i]) { if (unlikely(b->pointers[i] == p)) { return false; } } else { goto found; } } prev = b; b = b->next; } while (b); b = qemu_memalign(QHT_BUCKET_ALIGN, sizeof(*b)); memset(b, 0, sizeof(*b)); new = b; i = 0; atomic_inc(&map->n_added_buckets); if (unlikely(qht_map_needs_resize(map)) && needs_resize) { *needs_resize = true; } found: /* found an empty key: acquire the seqlock and write */ seqlock_write_begin(&head->sequence); if (new) { atomic_rcu_set(&prev->next, b); } /* smp_wmb() implicit in seqlock_write_begin. */ atomic_set(&b->hashes[i], hash); atomic_set(&b->pointers[i], p); seqlock_write_end(&head->sequence); return true; } static __attribute__((noinline)) void qht_grow_maybe(struct qht *ht) { struct qht_map *map; /* * If the lock is taken it probably means there's an ongoing resize, * so bail out. */ if (qemu_mutex_trylock(&ht->lock)) { return; } map = ht->map; /* another thread might have just performed the resize we were after */ if (qht_map_needs_resize(map)) { struct qht_map *new = qht_map_create(map->n_buckets * 2); qht_do_resize(ht, new); } qemu_mutex_unlock(&ht->lock); } bool qht_insert(struct qht *ht, void *p, uint32_t hash) { struct qht_bucket *b; struct qht_map *map; bool needs_resize = false; bool ret; /* NULL pointers are not supported */ qht_debug_assert(p); b = qht_bucket_lock__no_stale(ht, hash, &map); ret = qht_insert__locked(ht, map, b, p, hash, &needs_resize); qht_bucket_debug__locked(b); qemu_spin_unlock(&b->lock); if (unlikely(needs_resize) && ht->mode & QHT_MODE_AUTO_RESIZE) { qht_grow_maybe(ht); } return ret; } static inline bool qht_entry_is_last(struct qht_bucket *b, int pos) { if (pos == QHT_BUCKET_ENTRIES - 1) { if (b->next == NULL) { return true; } return b->next->pointers[0] == NULL; } return b->pointers[pos + 1] == NULL; } static void qht_entry_move(struct qht_bucket *to, int i, struct qht_bucket *from, int j) { qht_debug_assert(!(to == from && i == j)); qht_debug_assert(to->pointers[i]); qht_debug_assert(from->pointers[j]); atomic_set(&to->hashes[i], from->hashes[j]); atomic_set(&to->pointers[i], from->pointers[j]); atomic_set(&from->hashes[j], 0); atomic_set(&from->pointers[j], NULL); } /* * Find the last valid entry in @head, and swap it with @orig[pos], which has * just been invalidated. */ static inline void qht_bucket_remove_entry(struct qht_bucket *orig, int pos) { struct qht_bucket *b = orig; struct qht_bucket *prev = NULL; int i; if (qht_entry_is_last(orig, pos)) { orig->hashes[pos] = 0; atomic_set(&orig->pointers[pos], NULL); return; } do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { if (b->pointers[i]) { continue; } if (i > 0) { return qht_entry_move(orig, pos, b, i - 1); } qht_debug_assert(prev); return qht_entry_move(orig, pos, prev, QHT_BUCKET_ENTRIES - 1); } prev = b; b = b->next; } while (b); /* no free entries other than orig[pos], so swap it with the last one */ qht_entry_move(orig, pos, prev, QHT_BUCKET_ENTRIES - 1); } /* call with b->lock held */ static inline bool qht_remove__locked(struct qht_map *map, struct qht_bucket *head, const void *p, uint32_t hash) { struct qht_bucket *b = head; int i; do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { void *q = b->pointers[i]; if (unlikely(q == NULL)) { return false; } if (q == p) { qht_debug_assert(b->hashes[i] == hash); seqlock_write_begin(&head->sequence); qht_bucket_remove_entry(b, i); seqlock_write_end(&head->sequence); return true; } } b = b->next; } while (b); return false; } bool qht_remove(struct qht *ht, const void *p, uint32_t hash) { struct qht_bucket *b; struct qht_map *map; bool ret; /* NULL pointers are not supported */ qht_debug_assert(p); b = qht_bucket_lock__no_stale(ht, hash, &map); ret = qht_remove__locked(map, b, p, hash); qht_bucket_debug__locked(b); qemu_spin_unlock(&b->lock); return ret; } static inline void qht_bucket_iter(struct qht *ht, struct qht_bucket *b, qht_iter_func_t func, void *userp) { int i; do { for (i = 0; i < QHT_BUCKET_ENTRIES; i++) { if (b->pointers[i] == NULL) { return; } func(ht, b->pointers[i], b->hashes[i], userp); } b = b->next; } while (b); } /* call with all of the map's locks held */ static inline void qht_map_iter__all_locked(struct qht *ht, struct qht_map *map, qht_iter_func_t func, void *userp) { size_t i; for (i = 0; i < map->n_buckets; i++) { qht_bucket_iter(ht, &map->buckets[i], func, userp); } } void qht_iter(struct qht *ht, qht_iter_func_t func, void *userp) { struct qht_map *map; map = atomic_rcu_read(&ht->map); qht_map_lock_buckets(map); /* Note: ht here is merely for carrying ht->mode; ht->map won't be read */ qht_map_iter__all_locked(ht, map, func, userp); qht_map_unlock_buckets(map); } static void qht_map_copy(struct qht *ht, void *p, uint32_t hash, void *userp) { struct qht_map *new = userp; struct qht_bucket *b = qht_map_to_bucket(new, hash); /* no need to acquire b->lock because no thread has seen this map yet */ qht_insert__locked(ht, new, b, p, hash, NULL); } /* * Atomically perform a resize and/or reset. * Call with ht->lock held. */ static void qht_do_resize_reset(struct qht *ht, struct qht_map *new, bool reset) { struct qht_map *old; old = ht->map; qht_map_lock_buckets(old); if (reset) { qht_map_reset__all_locked(old); } if (new == NULL) { qht_map_unlock_buckets(old); return; } g_assert_cmpuint(new->n_buckets, !=, old->n_buckets); qht_map_iter__all_locked(ht, old, qht_map_copy, new); qht_map_debug__all_locked(new); atomic_rcu_set(&ht->map, new); qht_map_unlock_buckets(old); call_rcu(old, qht_map_destroy, rcu); } bool qht_resize(struct qht *ht, size_t n_elems) { size_t n_buckets = qht_elems_to_buckets(n_elems); size_t ret = false; qemu_mutex_lock(&ht->lock); if (n_buckets != ht->map->n_buckets) { struct qht_map *new; new = qht_map_create(n_buckets); qht_do_resize(ht, new); ret = true; } qemu_mutex_unlock(&ht->lock); return ret; } /* pass @stats to qht_statistics_destroy() when done */ void qht_statistics_init(struct qht *ht, struct qht_stats *stats) { struct qht_map *map; int i; map = atomic_rcu_read(&ht->map); stats->used_head_buckets = 0; stats->entries = 0; qdist_init(&stats->chain); qdist_init(&stats->occupancy); /* bail out if the qht has not yet been initialized */ if (unlikely(map == NULL)) { stats->head_buckets = 0; return; } stats->head_buckets = map->n_buckets; for (i = 0; i < map->n_buckets; i++) { struct qht_bucket *head = &map->buckets[i]; struct qht_bucket *b; unsigned int version; size_t buckets; size_t entries; int j; do { version = seqlock_read_begin(&head->sequence); buckets = 0; entries = 0; b = head; do { for (j = 0; j < QHT_BUCKET_ENTRIES; j++) { if (atomic_read(&b->pointers[j]) == NULL) { break; } entries++; } buckets++; b = atomic_rcu_read(&b->next); } while (b); } while (seqlock_read_retry(&head->sequence, version)); if (entries) { qdist_inc(&stats->chain, buckets); qdist_inc(&stats->occupancy, (double)entries / QHT_BUCKET_ENTRIES / buckets); stats->used_head_buckets++; stats->entries += entries; } else { qdist_inc(&stats->occupancy, 0); } } } void qht_statistics_destroy(struct qht_stats *stats) { qdist_destroy(&stats->occupancy); qdist_destroy(&stats->chain); }