/*
* The removed flag needs to be updated atomically with the pointer.
+ * It indicates that no node must attach to the node scheduled for
+ * removal. The gc flag also needs to be updated atomically with the
+ * pointer. It indicates that node garbage collection must be performed.
+ * "removed" and "gc" flags are separate for the benefit of replacement
+ * operation.
* The dummy flag does not require to be updated atomically with the
* pointer, but it is added as a pointer low bit flag to save space.
*/
#define REMOVED_FLAG (1UL << 0)
-#define DUMMY_FLAG (1UL << 1)
-#define FLAGS_MASK ((1UL << 2) - 1)
+#define GC_FLAG (1UL << 1)
+#define DUMMY_FLAG (1UL << 2)
+#define FLAGS_MASK ((1UL << 3) - 1)
/* Value of the end pointer. Should not interact with flags. */
#define END_VALUE NULL
unsigned long start, unsigned long len);
};
+enum add_mode {
+ ADD_DEFAULT = 0,
+ ADD_UNIQUE = 1,
+ ADD_REPLACE = 2,
+};
+
static
struct cds_lfht_node *_cds_lfht_add(struct cds_lfht *ht,
unsigned long size,
struct cds_lfht_node *node,
- int unique, int dummy);
+ enum add_mode mode, int dummy);
+
+static
+int _cds_lfht_del(struct cds_lfht *ht, unsigned long size,
+ struct cds_lfht_node *node,
+ int dummy_removal, int do_gc);
/*
* Algorithm to reverse bits in a word by lookup table, extended to
return (struct cds_lfht_node *) (((unsigned long) node) | REMOVED_FLAG);
}
+static
+int is_gc(struct cds_lfht_node *node)
+{
+ return ((unsigned long) node) & GC_FLAG;
+}
+
+static
+struct cds_lfht_node *flag_gc(struct cds_lfht_node *node)
+{
+ return (struct cds_lfht_node *) (((unsigned long) node) | GC_FLAG);
+}
+
static
int is_dummy(struct cds_lfht_node *node)
{
struct cds_lfht_node *iter_prev, *iter, *next, *new_next;
assert(!is_dummy(dummy));
+ assert(!is_gc(dummy));
assert(!is_removed(dummy));
assert(!is_dummy(node));
+ assert(!is_gc(node));
assert(!is_removed(node));
for (;;) {
iter_prev = dummy;
if (likely(clear_flag(iter)->p.reverse_hash > node->p.reverse_hash))
return;
next = rcu_dereference(clear_flag(iter)->p.next);
- if (likely(is_removed(next)))
+ if (likely(is_gc(next)))
break;
iter_prev = clear_flag(iter);
iter = next;
}
- assert(!is_removed(iter));
+ assert(!is_gc(iter));
if (is_dummy(iter))
new_next = flag_dummy(clear_flag(next));
else
new_next = clear_flag(next);
+ if (is_removed(iter))
+ new_next = flag_removed(new_next);
(void) uatomic_cmpxchg(&iter_prev->p.next, iter, new_next);
}
return;
struct cds_lfht_node *_cds_lfht_add(struct cds_lfht *ht,
unsigned long size,
struct cds_lfht_node *node,
- int unique, int dummy)
+ enum add_mode mode, int dummy)
{
struct cds_lfht_node *iter_prev, *iter, *next, *new_node, *new_next,
- *dummy_node;
+ *dummy_node, *return_node, *replace_pinned = NULL;
struct _cds_lfht_node *lookup;
unsigned long hash, index, order;
assert(!is_dummy(node));
+ assert(!is_gc(node));
assert(!is_removed(node));
if (!size) {
assert(dummy);
}
hash = bit_reverse_ulong(node->p.reverse_hash);
for (;;) {
- uint32_t chain_len = 0;
+ uint32_t chain_len;
+ retry:
+ chain_len = 0;
/*
* iter_prev points to the non-removed node prior to the
* insert location.
if (likely(clear_flag(iter)->p.reverse_hash > node->p.reverse_hash))
goto insert;
next = rcu_dereference(clear_flag(iter)->p.next);
- if (unlikely(is_removed(next)))
+ if (unlikely(is_gc(next)))
goto gc_node;
- if (unique
+ if (unlikely(replace_pinned)) {
+ /*
+ * We're in the retry of a node
+ * replacement. Only get exact iter
+ * pointer match. We own it, so it
+ * _needs_ to be there at some point.
+ */
+ if (clear_flag(iter) == replace_pinned)
+ goto replace;
+ }
+ /*
+ * Next is removed but not gc'd. We need to
+ * busy-loop, because a concurrent replacement
+ * is keeping it temporarily pinned there but we
+ * cannot attach to it. The easiest solution is
+ * to retry.
+ */
+ if (unlikely(is_removed(next)))
+ goto retry;
+ if ((mode == ADD_UNIQUE || mode == ADD_REPLACE)
&& !is_dummy(next)
&& !ht->compare_fct(node->key, node->key_len,
clear_flag(iter)->key,
- clear_flag(iter)->key_len))
- return clear_flag(iter);
+ clear_flag(iter)->key_len)) {
+ if (mode == ADD_UNIQUE)
+ return clear_flag(iter);
+ else /* mode == ADD_REPLACE */
+ goto replace;
+ }
/* Only account for identical reverse hash once */
if (iter_prev->p.reverse_hash != clear_flag(iter)->p.reverse_hash
&& !is_dummy(next))
iter_prev = clear_flag(iter);
iter = next;
}
+
insert:
assert(node != clear_flag(iter));
assert(!is_removed(iter_prev));
assert(!is_removed(iter));
+ assert(!is_gc(iter_prev));
+ assert(!is_gc(iter));
assert(iter_prev != node);
+ assert(!replace_pinned);
if (!dummy)
node->p.next = clear_flag(iter);
else
else
new_node = node;
if (uatomic_cmpxchg(&iter_prev->p.next, iter,
- new_node) != iter)
+ new_node) != iter) {
continue; /* retry */
- else
+ } else {
+ if (mode == ADD_REPLACE)
+ return_node = NULL;
+ else /* ADD_DEFAULT and ADD_UNIQUE */
+ return_node = node;
goto gc_end;
+ }
+
+ replace:
+ assert(node != clear_flag(iter));
+ assert(!is_removed(iter_prev));
+ assert(!is_removed(iter));
+ assert(!is_gc(iter_prev));
+ assert(!is_gc(iter));
+ assert(iter_prev != node);
+ assert(!dummy);
+ node->p.next = clear_flag(next);
+ if (is_dummy(iter))
+ new_node = flag_dummy(node);
+ else
+ new_node = node;
+ /*
+ * Try to delete to-be-replaced node. Don't gc yet. Not
+ * performing gc here is important, because this lets
+ * concurrent lookups see the old node until we
+ * atomically swap the new node into its place.
+ *
+ * This algorithm is _not_ strictly lock-free between
+ * _cds_lfht_del and the uatomic_cmpxchg of the
+ * replacement operation, so a replacement should _not_
+ * crash here (which means: don't do replacements if you
+ * need strict lock-free guarantees).
+ */
+ if (!replace_pinned) {
+ if (_cds_lfht_del(ht, size, clear_flag(iter), 0, 0))
+ continue; /* concurrently removed. retry. */
+ }
+ /*
+ * After _cds_lfht_del succeeds, we have pinned the
+ * to-be-removed node in place by setting its removed
+ * flag, but not its gc flag. If we fail to cmpxchg our
+ * new node with this node, we need to retry everything
+ * from the initial lookup, and only stop when we reach
+ * the node we pinned into place.
+ */
+ return_node = uatomic_cmpxchg(&iter_prev->p.next,
+ iter, new_node);
+ if (return_node != iter) {
+ /*
+ * If cmpxchg fails, we need to do path
+ * compression, but end it by placing our own
+ * node into place.
+ */
+ replace_pinned = clear_flag(iter);
+ continue; /* retry */
+ } else {
+ /*
+ * cmpxchg succeeded. gc unnecessary, because we
+ * unlinked the return_node ourself with the
+ * cmpxchg.
+ */
+ return_node = clear_flag(return_node);
+ goto end;
+ }
+
gc_node:
assert(!is_removed(iter));
+ assert(!is_gc(iter));
if (is_dummy(iter))
new_next = flag_dummy(clear_flag(next));
else
lookup = &ht->t.tbl[order]->nodes[index & (!order ? 0 : ((1UL << (order - 1)) - 1))];
dummy_node = (struct cds_lfht_node *) lookup;
_cds_lfht_gc_bucket(dummy_node, node);
- return node;
+end:
+ return return_node;
}
static
int _cds_lfht_del(struct cds_lfht *ht, unsigned long size,
struct cds_lfht_node *node,
- int dummy_removal)
+ int dummy_removal, int do_gc)
{
struct cds_lfht_node *dummy, *next, *old;
struct _cds_lfht_node *lookup;
/* logically delete the node */
assert(!is_dummy(node));
+ assert(!is_gc(node));
assert(!is_removed(node));
old = rcu_dereference(node->p.next);
do {
+ struct cds_lfht_node *new_next;
+
next = old;
if (unlikely(is_removed(next)))
goto end;
assert(is_dummy(next));
else
assert(!is_dummy(next));
- old = uatomic_cmpxchg(&node->p.next, next,
- flag_removed(next));
+ new_next = flag_removed(next);
+ if (do_gc)
+ new_next = flag_gc(new_next);
+ old = uatomic_cmpxchg(&node->p.next, next, new_next);
} while (old != next);
/* We performed the (logical) deletion. */
flagged = 1;
+ if (!do_gc)
+ goto end;
+
/*
* Ensure that the node is not visible to readers anymore: lookup for
* the node, and remove it (along with any other logically removed node)
new_node->p.reverse_hash =
bit_reverse_ulong(!i ? 0 : (1UL << (i - 1)) + j);
(void) _cds_lfht_add(ht, !i ? 0 : (1UL << (i - 1)),
- new_node, 0, 1);
+ new_node, ADD_DEFAULT, 1);
if (CMM_LOAD_SHARED(ht->in_progress_destroy))
break;
}
fini_node->p.reverse_hash =
bit_reverse_ulong(!i ? 0 : (1UL << (i - 1)) + j);
(void) _cds_lfht_del(ht, !i ? 0 : (1UL << (i - 1)),
- fini_node, 1);
+ fini_node, 1, 1);
if (CMM_LOAD_SHARED(ht->in_progress_destroy))
break;
}
break;
}
next = rcu_dereference(node->p.next);
- if (likely(!is_removed(next))
+ /*
+ * We consider return nodes marked removed but not gc as
+ * hits for lookup vs replacement consistency.
+ */
+ if (likely(!is_gc(next))
&& !is_dummy(next)
&& likely(!ht->compare_fct(node->key, node->key_len, key, key_len))) {
break;
break;
}
next = rcu_dereference(node->p.next);
- if (likely(!is_removed(next))
+ /*
+ * We consider return nodes marked removed but not gc as
+ * hits for lookup vs replacement consistency.
+ */
+ if (likely(!is_gc(next))
&& !is_dummy(next)
&& likely(!ht->compare_fct(node->key, node->key_len, key, key_len))) {
break;
node->p.reverse_hash = bit_reverse_ulong((unsigned long) hash);
size = rcu_dereference(ht->t.size);
- (void) _cds_lfht_add(ht, size, node, 0, 0);
+ (void) _cds_lfht_add(ht, size, node, ADD_DEFAULT, 0);
ht_count_add(ht, size);
}
struct cds_lfht_node *cds_lfht_add_unique(struct cds_lfht *ht,
- struct cds_lfht_node *node)
+ struct cds_lfht_node *node)
{
unsigned long hash, size;
struct cds_lfht_node *ret;
node->p.reverse_hash = bit_reverse_ulong((unsigned long) hash);
size = rcu_dereference(ht->t.size);
- ret = _cds_lfht_add(ht, size, node, 1, 0);
+ ret = _cds_lfht_add(ht, size, node, ADD_UNIQUE, 0);
if (ret == node)
ht_count_add(ht, size);
return ret;
}
+struct cds_lfht_node *cds_lfht_replace(struct cds_lfht *ht,
+ struct cds_lfht_node *node)
+{
+ unsigned long hash, size;
+ struct cds_lfht_node *ret;
+
+ hash = ht->hash_fct(node->key, node->key_len, ht->hash_seed);
+ node->p.reverse_hash = bit_reverse_ulong((unsigned long) hash);
+
+ size = rcu_dereference(ht->t.size);
+ ret = _cds_lfht_add(ht, size, node, ADD_REPLACE, 0);
+ if (ret == NULL)
+ ht_count_add(ht, size);
+ return ret;
+}
+
int cds_lfht_del(struct cds_lfht *ht, struct cds_lfht_node *node)
{
unsigned long size;
int ret;
size = rcu_dereference(ht->t.size);
- ret = _cds_lfht_del(ht, size, node, 0);
+ ret = _cds_lfht_del(ht, size, node, 0, 1);
if (!ret)
ht_count_del(ht, size);
return ret;
if (!is_dummy(node))
return -EPERM;
assert(!is_removed(node));
+ assert(!is_gc(node));
} while (!is_end(node));
/*
* size accessed without rcu_dereference because hash table is
node = (struct cds_lfht_node *) lookup;
do {
next = rcu_dereference(node->p.next);
- if (is_removed(next)) {
+ if (is_removed(next) || is_gc(next)) {
assert(!is_dummy(next));
(*removed)++;
} else if (!is_dummy(next))
static unsigned long init_hash_size = DEFAULT_HASH_SIZE;
static unsigned long init_populate;
static int opt_auto_resize;
-static int add_only, add_unique;
+static int add_only, add_unique, add_replace;
static unsigned long init_pool_offset, lookup_pool_offset, write_pool_offset;
static unsigned long init_pool_size = DEFAULT_RAND_POOL,
cds_lfht_node_init(node,
(void *)(((unsigned long) rand_r(&rand_lookup) % write_pool_size) + write_pool_offset),
sizeof(void *));
- if (add_unique)
+ if (add_unique) {
ret_node = cds_lfht_add_unique(test_ht, node);
- else
- cds_lfht_add(test_ht, node);
+ } else {
+ if (add_replace)
+ ret_node = cds_lfht_replace(test_ht, node);
+ else
+ cds_lfht_add(test_ht, node);
+ }
rcu_read_unlock();
if (add_unique && ret_node != node) {
free(node);
nr_addexist++;
- } else
- nr_add++;
+ } else {
+ if (add_replace && ret_node) {
+ call_rcu(&ret_node->head, free_node_cb);
+ nr_addexist++;
+ } else {
+ nr_add++;
+ }
+ }
} else {
/* May delete */
rcu_read_lock();
if (!init_populate)
return 0;
- if (add_unique && init_populate * 10 > init_pool_size) {
+ if ((add_unique || add_replace) && init_populate * 10 > init_pool_size) {
printf("WARNING: required to populate %lu nodes (-k), but random "
-"pool is quite small (%lu values) and we are in add_unique (-u) mode. Try with a "
+"pool is quite small (%lu values) and we are in add_unique (-u) or add_replace (-s) mode. Try with a "
"larger random pool (-p option). This may take a while...\n", init_populate, init_pool_size);
}
cds_lfht_node_init(node,
(void *)(((unsigned long) rand_r(&rand_lookup) % init_pool_size) + init_pool_offset),
sizeof(void *));
- if (add_unique)
+ rcu_read_lock();
+ if (add_unique) {
ret_node = cds_lfht_add_unique(test_ht, node);
- else
- cds_lfht_add(test_ht, node);
+ } else {
+ if (add_replace)
+ ret_node = cds_lfht_replace(test_ht, node);
+ else
+ cds_lfht_add(test_ht, node);
+ }
+ rcu_read_unlock();
if (add_unique && ret_node != node) {
free(node);
nr_addexist++;
- } else
- nr_add++;
+ } else {
+ if (add_replace && ret_node) {
+ call_rcu(&ret_node->head, free_node_cb);
+ nr_addexist++;
+ } else {
+ nr_add++;
+ }
+ }
nr_writes++;
}
return 0;
void show_usage(int argc, char **argv)
{
- printf("Usage : %s nr_readers nr_writers duration (s)", argv[0]);
+ printf("Usage : %s nr_readers nr_writers duration (s)\n", argv[0]);
#ifdef DEBUG_YIELD
- printf(" [-r] [-w] (yield reader and/or writer)");
+ printf(" [-r] [-w] (yield reader and/or writer)\n");
#endif
- printf(" [-d delay] (writer period (us))");
- printf(" [-c duration] (reader C.S. duration (in loops))");
- printf(" [-v] (verbose output)");
- printf(" [-a cpu#] [-a cpu#]... (affinity)");
- printf(" [-h size] (initial hash table size)");
- printf(" [-u] Uniquify add.");
- printf(" [-i] Add only (no removal).");
- printf(" [-k nr_nodes] Number of nodes to insert initially.");
- printf(" [-A] Automatically resize hash table.");
- printf(" [-R offset] Lookup pool offset.");
- printf(" [-S offset] Write pool offset.");
- printf(" [-T offset] Init pool offset.");
- printf(" [-M size] Lookup pool size.");
- printf(" [-N size] Write pool size.");
- printf(" [-O size] Init pool size.");
- printf(" [-V] Validate lookups of init values (use with filled init pool, same lookup range, with different write range).");
- printf("\n");
+ printf(" [-d delay] (writer period (us))\n");
+ printf(" [-c duration] (reader C.S. duration (in loops))\n");
+ printf(" [-v] (verbose output)\n");
+ printf(" [-a cpu#] [-a cpu#]... (affinity)\n");
+ printf(" [-h size] (initial hash table size)\n");
+ printf(" [not -u nor -s] Add entries (supports redundant keys).\n");
+ printf(" [-u] Uniquify add (no redundant keys).\n");
+ printf(" [-s] Replace (swap) entries.\n");
+ printf(" [-i] Add only (no removal).\n");
+ printf(" [-k nr_nodes] Number of nodes to insert initially.\n");
+ printf(" [-A] Automatically resize hash table.\n");
+ printf(" [-R offset] Lookup pool offset.\n");
+ printf(" [-S offset] Write pool offset.\n");
+ printf(" [-T offset] Init pool offset.\n");
+ printf(" [-M size] Lookup pool size.\n");
+ printf(" [-N size] Write pool size.\n");
+ printf(" [-O size] Init pool size.\n");
+ printf(" [-V] Validate lookups of init values (use with filled init pool, same lookup range, with different write range).\n");
+ printf("\n\n");
}
int main(int argc, char **argv)
init_hash_size = atol(argv[++i]);
break;
case 'u':
+ if (add_replace) {
+ printf("Please specify at most one of -s or -u.\n");
+ exit(-1);
+ }
add_unique = 1;
break;
+ case 's':
+ if (add_unique) {
+ printf("Please specify at most one of -s or -u.\n");
+ exit(-1);
+ }
+ add_replace = 1;
+ break;
case 'i':
add_only = 1;
break;
printf_verbose("Reader duration : %lu loops.\n", rduration);
printf_verbose("Mode:%s%s.\n",
add_only ? " add only" : " add/remove",
- add_unique ? " uniquify" : "");
+ add_unique ? " uniquify" : ( add_replace ? " replace" : " insert"));
printf_verbose("Initial hash table size: %lu buckets.\n", init_hash_size);
printf_verbose("Init pool size offset %lu size %lu.\n",
init_pool_offset, init_pool_size);
tid_writer = malloc(sizeof(*tid_writer) * nr_writers);
count_reader = malloc(sizeof(*count_reader) * nr_readers);
count_writer = malloc(sizeof(*count_writer) * nr_writers);
+
+ err = create_all_cpu_call_rcu_data(0);
+ assert(!err);
+
+ /*
+ * Hash creation and population needs to be seen as a RCU reader
+ * thread from the point of view of resize.
+ */
+ rcu_register_thread();
test_ht = cds_lfht_new(test_hash, test_compare, 0x42UL,
init_hash_size,
opt_auto_resize ? CDS_LFHT_AUTO_RESIZE : 0, NULL);
- ret = populate_hash();
+ ret = populate_hash();
assert(!ret);
- err = create_all_cpu_call_rcu_data(0);
- assert(!err);
+ rcu_unregister_thread();
next_aff = 0;
/*
* struct cds_lfht_node and struct _cds_lfht_node should be aligned on
- * 4-bytes boundaries because the two lower bits are used as flags.
+ * 8-bytes boundaries because the two lower bits are used as flags.
*/
struct _cds_lfht_node {
- struct cds_lfht_node *next; /* ptr | DUMMY_FLAG | REMOVED_FLAG */
+ struct cds_lfht_node *next; /* ptr | DUMMY_FLAG | GC_FLAG | REMOVED_FLAG */
unsigned long reverse_hash;
-};
+} __attribute__((aligned(8)));
struct cds_lfht_node {
/* cache-hot for iteration */
* cds_lfht_add - add a node to the hash table.
*
* Call with rcu_read_lock held.
+ * This function supports adding redundant keys into the table.
*/
void cds_lfht_add(struct cds_lfht *ht, struct cds_lfht_node *node);
* cds_lfht_add_unique fails, the node passed as parameter should be
* freed by the caller.
* Call with rcu_read_lock held.
+ *
+ * The semantic of this function is that if only this function is used
+ * to add keys into the table, no duplicated keys should ever be
+ * observable in the table. The same guarantee apply for combination of
+ * add_unique and replace (see below).
*/
struct cds_lfht_node *cds_lfht_add_unique(struct cds_lfht *ht, struct cds_lfht_node *node);
+/*
+ * cds_lfht_replace - replace a node within hash table.
+ *
+ * Return the node replaced upon success. If no node matching the key
+ * was present, return NULL, which also means the operation succeeded.
+ * This replacement operation should never fail.
+ * Call with rcu_read_lock held.
+ * After successful replacement, a grace period must be waited for before
+ * freeing the memory reserved for the returned node.
+ *
+ * The semantic of replacement vs lookups is the following: if lookups
+ * are performed between a key insertion and its removal, we guarantee
+ * that the lookups will always find the key if it is replaced
+ * concurrently with the lookups. Providing this guarantee require us to
+ * pin the node to remove in place (disallowing any insertion after this
+ * node temporarily) before we can proceed to its exchange with the new
+ * node atomically. This renders the "replace" operation not strictly
+ * lock-free, because a thread crashing in the middle of the replace
+ * operation could stop progress for other updaters.
+ *
+ * Providing this semantic allows us to ensure that replacement-only
+ * schemes will never generate duplicated keys. It also allows us to
+ * guarantee that a combination of replacement and add_unique updates
+ * will never generate duplicated keys.
+ */
+struct cds_lfht_node *cds_lfht_replace(struct cds_lfht *ht, struct cds_lfht_node *node);
+
/*
* cds_lfht_del - remove node from hash table.
*
+ * Return 0 if the node is successfully removed.
* Node can be looked up with cds_lfht_lookup. RCU read-side lock must
* be held between lookup and removal.
* Call with rcu_read_lock held.
+ * After successful removal, a grace period must be waited for before
+ * freeing the memory reserved for node.
*/
int cds_lfht_del(struct cds_lfht *ht, struct cds_lfht_node *node);
projects / urcu.git / commitdiff
