Merge branch 'jh/memihash-opt'

Signed-off-by: Johannes Schindelin <[email protected]>

Author: dscho

Documentation/technical/api-hashmap.txt

@@ -21,6 +21,9 @@ that the hashmap is initialized. It may also be useful for statistical purposes
 `cmpfn` stores the comparison function specified in `hashmap_init()`. In
 advanced scenarios, it may be useful to change this, e.g. to switch between
 case-sensitive and case-insensitive lookup.
++
+When `disallow_rehash` is set, automatic rehashes are prevented during inserts
+and deletes.
 
 `struct hashmap_entry`::
 
@@ -57,6 +60,7 @@ Functions
 `unsigned int strihash(const char *buf)`::
 `unsigned int memhash(const void *buf, size_t len)`::
 `unsigned int memihash(const void *buf, size_t len)`::
+`unsigned int memihash_cont(unsigned int hash_seed, const void *buf, size_t len)`::
 
 	Ready-to-use hash functions for strings, using the FNV-1 algorithm (see
 	http://www.isthe.com/chongo/tech/comp/fnv).
@@ -65,6 +69,9 @@ Functions
 `memihash` operate on arbitrary-length memory.
 +
 `strihash` and `memihash` are case insensitive versions.
++
+`memihash_cont` is a variant of `memihash` that allows a computation to be
+continued with another chunk of data.
 
 `unsigned int sha1hash(const unsigned char *sha1)`::
 
@@ -184,6 +191,21 @@ passed to `hashmap_cmp_fn` to decide whether the entry matches the key.
 +
 Returns the removed entry, or NULL if not found.
 
+`void hashmap_disallow_rehash(struct hashmap *map, unsigned value)`::
+
+	Disallow/allow automatic rehashing of the hashmap during inserts
+	and deletes.
++
+This is useful if the caller knows that the hashmap will be accessed
+by multiple threads.
++
+The caller is still responsible for any necessary locking; this simply
+prevents unexpected rehashing.  The caller is also responsible for properly
+sizing the initial hashmap to ensure good performance.
++
+A call to allow rehashing does not force a rehash; that might happen
+with the next insert or delete.
+
 `void hashmap_iter_init(struct hashmap *map, struct hashmap_iter *iter)`::
 `void *hashmap_iter_next(struct hashmap_iter *iter)`::
 `void *hashmap_iter_first(struct hashmap *map, struct hashmap_iter *iter)`::

Makefile

@@ -615,6 +615,7 @@ TEST_PROGRAMS_NEED_X += test-fake-ssh
 TEST_PROGRAMS_NEED_X += test-genrandom
 TEST_PROGRAMS_NEED_X += test-hashmap
 TEST_PROGRAMS_NEED_X += test-index-version
+TEST_PROGRAMS_NEED_X += test-lazy-init-name-hash
 TEST_PROGRAMS_NEED_X += test-line-buffer
 TEST_PROGRAMS_NEED_X += test-match-trees
 TEST_PROGRAMS_NEED_X += test-mergesort

cache.h

@@ -343,6 +343,7 @@ struct index_state {
 extern struct index_state the_index;
 
 /* Name hashing */
+extern int test_lazy_init_name_hash(struct index_state *istate, int try_threaded);
 extern void add_name_hash(struct index_state *istate, struct cache_entry *ce);
 extern void remove_name_hash(struct index_state *istate, struct cache_entry *ce);
 extern void free_name_hash(struct index_state *istate);

hashmap.c

@@ -50,6 +50,23 @@ unsigned int memihash(const void *buf, size_t len)
 	return hash;
 }
 
+/*
+ * Incoporate another chunk of data into a memihash
+ * computation.
+ */
+unsigned int memihash_cont(unsigned int hash_seed, const void *buf, size_t len)
+{
+	unsigned int hash = hash_seed;
+	unsigned char *ucbuf = (unsigned char *) buf;
+	while (len--) {
+		unsigned int c = *ucbuf++;
+		if (c >= 'a' && c <= 'z')
+			c -= 'a' - 'A';
+		hash = (hash * FNV32_PRIME) ^ c;
+	}
+	return hash;
+}
+
 #define HASHMAP_INITIAL_SIZE 64
 /* grow / shrink by 2^2 */
 #define HASHMAP_RESIZE_BITS 2
@@ -87,11 +104,19 @@ static inline unsigned int bucket(const struct hashmap *map,
 	return key->hash & (map->tablesize - 1);
 }
 
+int hashmap_bucket(const struct hashmap *map, unsigned int hash)
+{
+	return hash & (map->tablesize - 1);
+}
+
 static void rehash(struct hashmap *map, unsigned int newsize)
 {
 	unsigned int i, oldsize = map->tablesize;
 	struct hashmap_entry **oldtable = map->table;
 
+	if (map->disallow_rehash)
+		return;
+
 	alloc_table(map, newsize);
 	for (i = 0; i < oldsize; i++) {
 		struct hashmap_entry *e = oldtable[i];
@@ -124,7 +149,9 @@ void hashmap_init(struct hashmap *map, hashmap_cmp_fn equals_function,
 		size_t initial_size)
 {
 	unsigned int size = HASHMAP_INITIAL_SIZE;
-	map->size = 0;
+
+	memset(map, 0, sizeof(*map));
+
 	map->cmpfn = equals_function ? equals_function : always_equal;
 
 	/* calculate initial table size and allocate the table */

hashmap.h

@@ -12,6 +12,7 @@ extern unsigned int strhash(const char *buf);
 extern unsigned int strihash(const char *buf);
 extern unsigned int memhash(const void *buf, size_t len);
 extern unsigned int memihash(const void *buf, size_t len);
+extern unsigned int memihash_cont(unsigned int hash_seed, const void *buf, size_t len);
 
 static inline unsigned int sha1hash(const unsigned char *sha1)
 {
@@ -38,6 +39,7 @@ struct hashmap {
 	struct hashmap_entry **table;
 	hashmap_cmp_fn cmpfn;
 	unsigned int size, tablesize, grow_at, shrink_at;
+	unsigned disallow_rehash : 1;
 };
 
 struct hashmap_iter {
@@ -76,6 +78,29 @@ static inline void *hashmap_get_from_hash(const struct hashmap *map,
 	return hashmap_get(map, &key, keydata);
 }
 
+int hashmap_bucket(const struct hashmap *map, unsigned int hash);
+
+/*
+ * Disallow/allow rehashing of the hashmap.
+ * This is useful if the caller knows that the hashmap
+ * needs multi-threaded access.  The caller is still
+ * required to guard/lock searches and inserts in a
+ * manner appropriate to their usage.  This simply
+ * prevents the table from being unexpectedly re-mapped.
+ *
+ * If is up to the caller to ensure that the hashmap is
+ * initialized to a reasonable size to prevent poor
+ * performance.
+ *
+ * When value=1, prevent future rehashes on adds and deleted.
+ * When value=0, allow future rehahses.  This DOES NOT force
+ * a rehash now.
+ */
+static inline void hashmap_disallow_rehash(struct hashmap *map, unsigned value)
+{
+	map->disallow_rehash = value;
+}
+
 /* hashmap_iter functions */
 
 extern void hashmap_iter_init(struct hashmap *map, struct hashmap_iter *iter);