dedup

commit 042e687f46e7c334def143a387c332e6884cd078
parent c86adacd4c9ec47d68823fbb768ec3fc7e5c23c0
Author: sin <sin@2f30.org>
Date:   Sun, 19 May 2019 18:52:14 +0300

Xor the given seed into the buzhash table

This is the first step to try and mitigate against fingerprinting
attacks.  The current code does not have any effect on the initial
state of the buzhash algorithm.

Diffstat:
M
chunker.c
 | 
10
++++++++--
M
chunker.h
 | 
2
+-
M
dup-pack.c
 | 
3
++-

3 files changed, 11 insertions(+), 4 deletions(-)
diff --git a/chunker.c b/chunker.c
@@ -17,6 +17,7 @@ struct chunker {
 	size_t maxsize;
 	size_t mask;
 	size_t winsize;
+	uint32_t seed;
 };
 
 /*
@@ -28,7 +29,7 @@ struct chunker {
  * exactly 50% chance that a XOR operation would flip all the bits in
  * the hash.
  */
-static const uint32_t buztbl[] = {
+static uint32_t buztbl[] = {
 	0xbc9fa594,0x30a8f827,0xced627a7,0xdb46a745,0xcfa4a9e8,0x77cccb59,0xddb66276,0x3adc532f,
 	0xfe8b67d3,0x8155b59e,0x0c893666,0x1d757009,0x17394ee4,0x85d94c07,0xcacd52da,0x076c6f79,
 	0xead0a798,0x6c7ccb4a,0x2639a1b8,0x3aa5ae32,0x3e6218d2,0xb290d980,0xa5149521,0x4b426119,
@@ -124,9 +125,10 @@ cgetsize(struct chunker *c)
 
 struct chunker *
 copen(int fd, size_t minsize, size_t maxsize,
-      size_t mask, size_t winsize)
+      size_t mask, size_t winsize, uint32_t seed)
 {
 	struct chunker *c;
+	size_t i;
 
 	c = calloc(1, sizeof(*c));
 	if (c == NULL) {
@@ -146,6 +148,10 @@ copen(int fd, size_t minsize, size_t maxsize,
 	c->maxsize = maxsize;
 	c->mask = mask;
 	c->winsize = winsize;
+	c->seed = seed;
+
+	for (i = 0; i < sizeof(buztbl) / sizeof(buztbl[0]); i++)
+		buztbl[i] ^= c->seed;
 	return c;
 }
 
diff --git a/chunker.h b/chunker.h
@@ -1,6 +1,6 @@
 struct chunker;
 
-extern struct chunker *copen(int, size_t, size_t, size_t, size_t);
+extern struct chunker *copen(int, size_t, size_t, size_t, size_t, uint32_t);
 extern void cclose(struct chunker *);
 extern ssize_t cfill(struct chunker *);
 extern void *cget(struct chunker *, size_t *);
diff --git a/dup-pack.c b/dup-pack.c
@@ -4,6 +4,7 @@
 #include <err.h>
 #include <fcntl.h>
 #include <limits.h>
+#include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <unistd.h>
@@ -62,7 +63,7 @@ pack(struct sctx *sctx, struct bctx *bctx)
 {
 	struct chunker *c;
 
-	if ((c = copen(0, BSIZEMIN, BSIZEMAX, HMASKBITS, WINSIZE)) == NULL)
+	if ((c = copen(0, BSIZEMIN, BSIZEMAX, HMASKBITS, WINSIZE, 0)) == NULL)
 		printerr("copen");
 
 	while (cfill(c) > 0) {