From 4c17cdcb64e3f333f141b43b3193a19eece4c8a1 Mon Sep 17 00:00:00 2001
From: Wayne Davison <wayned@samba.org>
Date: Fri, 26 Oct 2007 22:11:19 +0000
Subject: [PATCH] Chunk a really large file to avoid sender-side hash-table
 overload.

---
 match.c | 30 ++++++++++++++++++++----------
 rsync.h |  2 +-
 2 files changed, 21 insertions(+), 11 deletions(-)

diff --git a/match.c b/match.c
index 8367d158..01e91173 100644
--- a/match.c
+++ b/match.c
@@ -46,9 +46,9 @@ static int32 *hash_table;
 #define SUM2HASH2(s1,s2) (((s1) + (s2)) & 0xFFFF)
 #define SUM2HASH(sum) SUM2HASH2((sum)&0xFFFF,(sum)>>16)
 
-static void build_hash_table(struct sum_struct *s)
+static int32 build_hash_table(struct sum_struct *s, int32 start)
 {
-	int32 i;
+	int32 i, end = s->count;
 
 	if (!hash_table) {
 		hash_table = new_array(int32, TABLESIZE);
@@ -58,11 +58,21 @@ static void build_hash_table(struct sum_struct *s)
 
 	memset(hash_table, 0xFF, TABLESIZE * sizeof hash_table[0]);
 
-	for (i = 0; i < s->count; i++) {
+	if (end - start > TABLESIZE*8/10)
+		end = start + TABLESIZE*8/10;
+
+	for (i = start; i < end; i++) {
 		uint32 t = SUM2HASH(s->sums[i].sum1);
 		s->sums[i].chain = hash_table[t];
 		hash_table[t] = i;
 	}
+
+	if (verbose > 2) {
+		rprintf(FINFO, "built hash table for entries %ld - %ld\n",
+			(long)start, (long)end - 1);
+	}
+
+	return end;
 }
 
 
@@ -120,8 +130,8 @@ static void matched(int f, struct sum_struct *s, struct map_struct *buf,
 static void hash_search(int f,struct sum_struct *s,
 			struct map_struct *buf, OFF_T len)
 {
-	OFF_T offset, end;
-	int32 k, want_i, backup;
+	OFF_T offset, end, reset = 0;
+	int32 k, want_i, backup, sum_pos = 0;
 	char sum2[SUM_LENGTH];
 	uint32 s1, s2, sum;
 	int more;
@@ -159,6 +169,11 @@ static void hash_search(int f,struct sum_struct *s,
 		int done_csum2 = 0;
 		int32 i;
 
+		if (offset >= reset) {
+			sum_pos = build_hash_table(s, sum_pos);
+			reset = sum_pos * s->blength;
+		}
+
 		if (verbose > 4) {
 			rprintf(FINFO, "offset=%.0f sum=%04x%04x\n",
 				(double)offset, s2 & 0xFFFF, s1 & 0xFFFF);
@@ -336,11 +351,6 @@ void match_sums(int f, struct sum_struct *s, struct map_struct *buf, OFF_T len)
 	}
 
 	if (len > 0 && s->count > 0) {
-		build_hash_table(s);
-
-		if (verbose > 2)
-			rprintf(FINFO,"built hash table\n");
-
 		hash_search(f, s, buf, len);
 
 		if (verbose > 2)
diff --git a/rsync.h b/rsync.h
index d5111b3c..78d51c1d 100644
--- a/rsync.h
+++ b/rsync.h
@@ -122,7 +122,7 @@
 #define CHUNK_SIZE (32*1024)
 #define MAX_MAP_SIZE (256*1024)
 #define IO_BUFFER_SIZE (4092)
-#define MAX_BLOCK_SIZE ((int32)1 << 29)
+#define MAX_BLOCK_SIZE ((int32)1 << 17)
 
 #define IOERR_GENERAL	(1<<0) /* For backward compatibility, this must == 1 */
 #define IOERR_VANISHED	(1<<1)
-- 
2.34.1