Split hash_file into three steps, setup, add buf and final.

Setup inits the context and adds the seed. The buf function simply adds
a block from the file to the hash. The final function calls MD4_Final()
to close the context and generate the hash.

This will help to remove the mmap in the sender and should result in a
more atomic view of the file since hash_file() is now called together
with the other hash_functions.

OK deraadt@ tb@

diff --git a/usr.bin/rsync/blocks.c b/usr.bin/rsync/blocks.c
index 5679072..11f9e27 100644 (file)
--- a/usr.bin/rsync/blocks.c
+++ b/usr.bin/rsync/blocks.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: blocks.c,v 1.21 2021/11/03 14:42:12 deraadt Exp $ */
+/*     $OpenBSD: blocks.c,v 1.22 2024/02/27 11:28:30 claudio Exp $ */
 /*
  * Copyright (c) 2019 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -280,6 +280,8 @@ blk_match(struct sess *sess, const struct blkset *blks,
                            blk->len, blk->idx);
                        tok = -(blk->idx + 1);
 
+                       hash_file_buf(&st->ctx, st->map + last, sz + blk->len);
+
                        /*
                         * Write the data we have, then follow it with
                         * the tag of the block that matches.
@@ -293,6 +295,7 @@ blk_match(struct sess *sess, const struct blkset *blks,
                        st->total += blk->len;
                        st->offs += blk->len;
                        st->hint = blk->idx + 1;
+
                        return;
                }
 
@@ -308,6 +311,8 @@ blk_match(struct sess *sess, const struct blkset *blks,
                st->curlen = st->curpos + sz;
                st->curtok = 0;
                st->curst = sz ? BLKSTAT_DATA : BLKSTAT_TOK;
+
+               hash_file_buf(&st->ctx, st->map + st->curpos, sz);
        } else {
                st->curpos = 0;
                st->curlen = st->mapsz;
@@ -315,6 +320,8 @@ blk_match(struct sess *sess, const struct blkset *blks,
                st->curst = st->mapsz ? BLKSTAT_DATA : BLKSTAT_TOK;
                st->dirty = st->total = st->mapsz;
 
+               hash_file_buf(&st->ctx, st->map, st->mapsz);
+
                LOG4("%s: flushing whole file %zu B",
                    path, st->mapsz);
        }

diff --git a/usr.bin/rsync/extern.h b/usr.bin/rsync/extern.h
index a36d937..940db81 100644 (file)
--- a/usr.bin/rsync/extern.h
+++ b/usr.bin/rsync/extern.h
@@ -1,4 +1,4 @@
-/*     $OpenBSD: extern.h,v 1.47 2023/11/27 11:30:49 claudio Exp $ */
+/*     $OpenBSD: extern.h,v 1.48 2024/02/27 11:28:30 claudio Exp $ */
 /*
  * Copyright (c) 2019 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -17,6 +17,8 @@
 #ifndef EXTERN_H
 #define EXTERN_H
 
+#include <openssl/md4.h>
+
 /*
  * This is the rsync protocol version that we support.
  */
@@ -214,6 +216,7 @@ struct      blkstat {
        struct blktab   *blktab; /* hashtable of blocks */
        uint32_t         s1; /* partial sum for computing fast hash */
        uint32_t         s2; /* partial sum for computing fast hash */
+       MD4_CTX          ctx; /* context for hash_file */
 };
 
 /*
@@ -388,8 +391,10 @@ int                 blk_send_ack(struct sess *, int, struct blkset *);
 uint32_t        hash_fast(const void *, size_t);
 void            hash_slow(const void *, size_t, unsigned char *,
                    const struct sess *);
-void            hash_file(const void *, size_t, unsigned char *,
-                   const struct sess *);
+
+void            hash_file_start(MD4_CTX *, const struct sess *);
+void            hash_file_buf(MD4_CTX *, const void *, size_t);
+void            hash_file_final(MD4_CTX *, unsigned char *);
 
 void            copy_file(int, const char *, const struct flist *);
 

diff --git a/usr.bin/rsync/hash.c b/usr.bin/rsync/hash.c
index b87c56f..37529e9 100644 (file)
--- a/usr.bin/rsync/hash.c
+++ b/usr.bin/rsync/hash.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: hash.c,v 1.4 2021/06/30 13:10:04 claudio Exp $ */
+/*     $OpenBSD: hash.c,v 1.5 2024/02/27 11:28:30 claudio Exp $ */
 /*
  * Copyright (c) 2019 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -82,14 +82,21 @@ hash_slow(const void *buf, size_t len,
  * of the sequence, not the beginning.
  */
 void
-hash_file(const void *buf, size_t len,
-       unsigned char *md, const struct sess *sess)
+hash_file_start(MD4_CTX *ctx, const struct sess *sess)
 {
-       MD4_CTX          ctx;
        int32_t          seed = htole32(sess->seed);
 
-       MD4_Init(&ctx);
-       MD4_Update(&ctx, (unsigned char *)&seed, sizeof(int32_t));
-       MD4_Update(&ctx, buf, len);
-       MD4_Final(md, &ctx);
+       MD4_Init(ctx);
+       MD4_Update(ctx, (unsigned char *)&seed, sizeof(int32_t));
+}
+
+void
+hash_file_buf(MD4_CTX *ctx, const void *buf, size_t len)
+{
+       MD4_Update(ctx, buf, len);
+}
+void
+hash_file_final(MD4_CTX *ctx, unsigned char *md)
+{
+       MD4_Final(md, ctx);
 }

diff --git a/usr.bin/rsync/sender.c b/usr.bin/rsync/sender.c
index 8e49f4d..43ccfbd 100644 (file)
--- a/usr.bin/rsync/sender.c
+++ b/usr.bin/rsync/sender.c
@@ -1,4 +1,4 @@
-/*     $OpenBSD: sender.c,v 1.31 2024/02/19 16:39:18 claudio Exp $ */
+/*     $OpenBSD: sender.c,v 1.32 2024/02/27 11:28:30 claudio Exp $ */
 /*
  * Copyright (c) 2019 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -159,7 +159,7 @@ send_up_fsm(struct sess *sess, size_t *phase,
                 * finished with the file.
                 */
 
-               hash_file(up->stat.map, up->stat.mapsz, fmd, sess);
+               hash_file_final(&up->stat.ctx, fmd);
                if (!io_lowbuffer_alloc(sess, wb, wbsz, wbmax, dsz)) {
                        ERRX1("io_lowbuffer_alloc");
                        return 0;
@@ -619,6 +619,7 @@ rsync_sender(struct sess *sess, int fdin,
 
                        /* Hash our blocks. */
 
+                       hash_file_start(&up.stat.ctx, sess);
                        blkhash_set(up.stat.blktab, up.cur->blks);
 
                        /*