v8-0003-Add-bytewise-interface.patch

From 54e6419a632b04d97cad847603035050ab48c84f Mon Sep 17 00:00:00 2001
From: John Naylor <john.naylor@postgresql.org>
Date: Sat, 9 Dec 2023 16:32:05 +0700
Subject: [PATCH v8 3/5] Add bytewise interface

This is useful for hashing values with unknown length,
like NUL-terminated strings. It should be faster than calling
strlen() first and passing the length, which most hash
functions require.

Note: This method can't give the same answer as
regular fasthash, so it will need to be evaluated. It's possible
we need to mix in the length at the finalization step (at which
time can know the length), in order to safeguard against
collisions.
---
 src/include/common/hashfn_unstable.h | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/src/include/common/hashfn_unstable.h b/src/include/common/hashfn_unstable.h
index fbae7a5522..80aec98dc9 100644
--- a/src/include/common/hashfn_unstable.h
+++ b/src/include/common/hashfn_unstable.h
@@ -49,6 +49,7 @@ typedef struct fasthash_state
 {
 	uint64 accum;
 #define FH_SIZEOF_ACCUM sizeof(uint64)
+	int8	accum_len;
 	uint64 hash;
 } fasthash_state;
 
@@ -69,6 +70,7 @@ fasthash_combine(fasthash_state* hs)
 
 	/* reset hash state for next input */
 	hs->accum = 0;
+	hs->accum_len = 0;
 }
 
 static inline void
@@ -82,6 +84,18 @@ fasthash_init(fasthash_state *hs, int len, uint64 seed)
 	hs->hash = seed ^ (len * 0x880355f21e6d1965ULL);
 }
 
+static inline void
+fasthash_accum_byte(fasthash_state *hs, const unsigned char ch)
+{
+	hs->accum <<= BITS_PER_BYTE;
+	hs->accum |= ch;
+	hs->accum_len++;
+
+	// wip: is there a better way to get sizeof struct member?
+	if (hs->accum_len == sizeof(((fasthash_state *) 0)->accum))
+		fasthash_combine(hs);
+}
+
 static inline void
 fasthash_accum(fasthash_state *hs, const unsigned char *k, int len)
 {
@@ -117,6 +131,11 @@ fasthash_accum(fasthash_state *hs, const unsigned char *k, int len)
 static inline uint64
 fasthash_final64(fasthash_state *hs)
 {
+	// check for remaining bytes to combine into hash
+	// should only be used by the bytewise interface
+	if (hs->accum_len > 0)
+		fasthash_combine(hs);
+
 	return fasthash_mix(hs->hash);
 }
 
-- 
2.43.0