diff mbox series

[5/5] newrole: use DJB2a string hash function

Message ID 20230816123845.80171-5-cgzones@googlemail.com (mailing list archive)
State Accepted
Commit 3089f1f2fd92
Delegated to: Petr Lautrbach
Headers show
Series [1/5] libsepol: include length squared in hashtab_hash_eval() | expand

Commit Message

Christian Göttsche Aug. 16, 2023, 12:38 p.m. UTC
The hash table implementation uses `& (h->size - 1)` to truncate
generated hashes to the number of buckets.  This operation is equal to
`% h->size` if and only if the size is a power of two (which seems to be
always the case).  One property of the binary and with a power of two
(and probably a small one <=2048) is all higher bits are discarded.
Thus a hash function is needed with a good avalanche effect, which the
current one is not.

Signed-off-by: Christian Göttsche <cgzones@googlemail.com>
---
 policycoreutils/newrole/newrole.c | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)
diff mbox series

Patch

diff --git a/policycoreutils/newrole/newrole.c b/policycoreutils/newrole/newrole.c
index d9efa68a..5a1a1129 100644
--- a/policycoreutils/newrole/newrole.c
+++ b/policycoreutils/newrole/newrole.c
@@ -229,16 +229,13 @@  static int free_hashtab_entry(hashtab_key_t key, hashtab_datum_t d,
 
 static unsigned int reqsymhash(hashtab_t h, const_hashtab_key_t key)
 {
-	const char *p;
-	size_t size;
-	unsigned int val;
-
-	val = 0;
-	size = strlen(key);
-	for (p = key; ((size_t) (p - key)) < size; p++)
-		val =
-		    (val << 4 | (val >> (8 * sizeof(unsigned int) - 4))) ^ (*p);
-	return val & (h->size - 1);
+	unsigned int hash = 5381;
+	unsigned char c;
+
+	while ((c = *(unsigned const char *)key++))
+		hash = ((hash << 5) + hash) ^ c;
+
+	return hash & (h->size - 1);
 }
 
 static int reqsymcmp(hashtab_t h