From 5d6ffd969d13d101d32cb9079dbadacc021b2ac8 Mon Sep 17 00:00:00 2001
From: Moinak Ghosh <moinakg@gmail.com>
Date: Wed, 24 Apr 2013 19:13:07 +0530
Subject: [PATCH] More tweaks to slightly improve segment dedupe efficiency.
 Use on average 8MB segments for all cases. Some minor cleanps.

---
 rabin/global/dedupe_config.c | 15 ++-------------
 rabin/global/dedupe_config.h |  4 +---
 rabin/global/index.c         |  2 +-
 3 files changed, 4 insertions(+), 17 deletions(-)

diff --git a/rabin/global/dedupe_config.c b/rabin/global/dedupe_config.c
index 707f3c5..69daac8 100644
--- a/rabin/global/dedupe_config.c
+++ b/rabin/global/dedupe_config.c
@@ -185,8 +185,8 @@ read_config(char *configfile, archive_config_t *cfg)
 	cfg->verify_chunks = 0;
 	cfg->algo = COMPRESS_LZ4;
 	cfg->chunk_cksum_type = DEFAULT_CHUNK_CKSUM;
-	cfg->similarity_cksum = DEFAULT_SIMILARITY_CKSUM;
-	cfg->pct_interval = DEFAULT_SIMILARITY_INTERVAL;
+	cfg->similarity_cksum = GLOBAL_SIM_CKSUM;
+	cfg->pct_interval = DEFAULT_PCT_INTERVAL;
 
 	fh = fopen(configfile, "r");
 	if (fh == NULL) {
@@ -272,13 +272,6 @@ read_config(char *configfile, archive_config_t *cfg)
 				fclose(fh);
 				return (1);
 			}
-		} else if (strncmp(line, "SIMILARITY_CKSUM", 16) == 0) {
-			cfg->chunk_cksum_type = get_cksum_type(pos);
-			if (cfg->chunk_cksum_type == CKSUM_INVALID) {
-				fprintf(stderr, "Invalid CHUNK_CKSUM setting.\n");
-				fclose(fh);
-				return (1);
-			}
 		}
 	}
 	fclose(fh);
@@ -373,10 +366,6 @@ set_config_s(archive_config_t *cfg, const char *algo, cksum_t ck, cksum_t ck_sim
 		cfg->dedupe_mode = MODE_SIMPLE;
 		cfg->segment_sz_bytes = user_chunk_sz;
 		cfg->similarity_cksum_sz = cfg->chunk_cksum_sz;
-
-	} else if (cfg->archive_sz < (ONE_TB * 100)) {
-		cfg->segment_sz_bytes = FOUR_MB;
-
 	} else {
 		cfg->segment_sz_bytes = EIGHT_MB;
 	}
diff --git a/rabin/global/dedupe_config.h b/rabin/global/dedupe_config.h
index afbed5f..3cd8e7a 100644
--- a/rabin/global/dedupe_config.h
+++ b/rabin/global/dedupe_config.h
@@ -33,11 +33,9 @@
 extern "C" {
 #endif
 
-#define	DEFAULT_SIMILARITY_INTERVAL	5
 #define	DEFAULT_CHUNK_CKSUM	CKSUM_SHA256
-#define	DEFAULT_SIMILARITY_CKSUM	CKSUM_BLAKE256
 #define	DEFAULT_COMPRESS		COMPRESS_LZ4
-#define	DEFAULT_PCT_INTERVAL	5
+#define	DEFAULT_PCT_INTERVAL	4
 #define	CONTAINER_ITEMS		2048
 #define	MIN_CK 1
 #define	MAX_CK 5
diff --git a/rabin/global/index.c b/rabin/global/index.c
index c8705eb..6a032ff 100644
--- a/rabin/global/index.c
+++ b/rabin/global/index.c
@@ -167,7 +167,7 @@ set_cfg:
 		*pct_interval = 0;
 	} else {
 		cfg->intervals = 100 / *pct_interval;
-		cfg->sub_intervals = (cfg->segment_sz / cfg->intervals) >> 1;
+		cfg->sub_intervals = cfg->intervals;
 		*hash_slots = file_sz / cfg->segment_sz_bytes + 1;
 		*hash_slots *= cfg->sub_intervals;
 	}