[SCM] The rsync repository. - branch master updated
Rsync CVS commit messages
rsync-cvs at lists.samba.org
Sat May 23 01:06:00 UTC 2020
The branch, master has been updated
via a28bc3eb Promoting xxhash support.
via 55bb4dab Some checksum improvements
from 5fa4209c AVX2 optimized version of get_checksum1() for x86-64
https://git.samba.org/?p=rsync.git;a=shortlog;h=master
- Log -----------------------------------------------------------------
commit a28bc3ebf672113cdd90432ef0270782d8ab6c2c
Author: Wayne Davison <wayne at opencoder.net>
Date: Fri May 22 17:40:08 2020 -0700
Promoting xxhash support.
commit 55bb4dab7a4f132537b5ed93e504ffc618b643b9
Author: Wayne Davison <wayne at opencoder.net>
Date: Fri May 22 15:06:20 2020 -0700
Some checksum improvements
- Improve csum negotation logic.
- Define the csum names in a single structure.
- Add --debug=CSUM.
-----------------------------------------------------------------------
Summary of changes:
checksum.c | 240 +++++++++++++++++++++++++++++++++++++++++++++++++++--------
compat.c | 21 +++---
configure.ac | 14 +++-
flist.c | 2 +-
io.c | 2 +-
options.c | 11 ++-
rsync.h | 3 +-
rsync.yo | 5 +-
8 files changed, 247 insertions(+), 51 deletions(-)
Changeset truncated at 500 lines:
diff --git a/checksum.c b/checksum.c
index 54e2c4aa..17a9507a 100644
--- a/checksum.c
+++ b/checksum.c
@@ -20,6 +20,9 @@
*/
#include "rsync.h"
+#ifdef SUPPORT_XXHASH
+#include "xxhash.h"
+#endif
extern int am_server;
extern int local_server;
@@ -36,9 +39,22 @@ extern char *checksum_choice;
#define CSUM_MD4_OLD 3
#define CSUM_MD4 4
#define CSUM_MD5 5
+#define CSUM_XXHASH 6
+
+#define CSUM_SAW_BUFLEN 10
-const char *default_checksum_list =
- "md5 md4";
+struct csum_struct {
+ int num;
+ const char *name;
+} valid_checksums[] = {
+#ifdef SUPPORT_XXHASH
+ { CSUM_XXHASH, "xxhash" },
+#endif
+ { CSUM_MD5, "md5" },
+ { CSUM_MD4, "md4" },
+ { CSUM_NONE, "none" },
+ { -1, NULL }
+};
#define MAX_CHECKSUM_LIST 1024
@@ -48,6 +64,8 @@ const char *negotiated_csum_name = NULL;
static int parse_csum_name(const char *name, int len, int allow_auto)
{
+ struct csum_struct *cs;
+
if (len < 0 && name)
len = strlen(name);
@@ -60,12 +78,11 @@ static int parse_csum_name(const char *name, int len, int allow_auto)
return CSUM_MD4_BUSTED;
return CSUM_MD4_ARCHAIC;
}
- if (len == 3 && strncasecmp(name, "md4", 3) == 0)
- return CSUM_MD4;
- if (len == 3 && strncasecmp(name, "md5", 3) == 0)
- return CSUM_MD5;
- if (len == 4 && strncasecmp(name, "none", 4) == 0)
- return CSUM_NONE;
+
+ for (cs = valid_checksums; cs->name; cs++) {
+ if (strncasecmp(name, cs->name, len) == 0 && cs->name[len] == '\0')
+ return cs->num;
+ }
if (allow_auto) {
rprintf(FERROR, "unknown checksum name: %s\n", name);
@@ -75,7 +92,22 @@ static int parse_csum_name(const char *name, int len, int allow_auto)
return -1;
}
-void parse_checksum_choice(void)
+static const char *checksum_name(int num)
+{
+ struct csum_struct *cs;
+
+ for (cs = valid_checksums; cs->name; cs++) {
+ if (num == cs->num)
+ return cs->name;
+ }
+
+ if (num < CSUM_MD4)
+ return "MD4";
+
+ return "UNKNOWN";
+}
+
+void parse_checksum_choice(int final_call)
{
if (!negotiated_csum_name) {
char *cp = checksum_choice ? strchr(checksum_choice, ',') : NULL;
@@ -85,46 +117,124 @@ void parse_checksum_choice(void)
} else
xfersum_type = checksum_type = parse_csum_name(checksum_choice, -1, 1);
}
+
if (xfersum_type == CSUM_NONE)
whole_file = 1;
+
+ if (final_call && DEBUG_GTE(CSUM, 1)) {
+ if (negotiated_csum_name)
+ rprintf(FINFO, "[%s] negotiated checksum: %s\n", who_am_i(), negotiated_csum_name);
+ else if (xfersum_type == checksum_type) {
+ rprintf(FINFO, "[%s] %s checksum: %s\n", who_am_i(),
+ checksum_choice ? "chosen" : "protocol-based",
+ checksum_name(xfersum_type));
+ } else {
+ rprintf(FINFO, "[%s] chosen transfer checksum: %s\n",
+ who_am_i(), checksum_name(xfersum_type));
+ rprintf(FINFO, "[%s] chosen pre-transfer checksum: %s\n",
+ who_am_i(), checksum_name(checksum_type));
+ }
+ }
+}
+
+static int parse_checksum_list(const char *from, char *sumbuf, int sumbuf_len, char *saw)
+{
+ char *to = sumbuf, *tok = NULL;
+ int cnt = 0;
+
+ memset(saw, 0, CSUM_SAW_BUFLEN);
+
+ while (1) {
+ if (*from == ' ' || !*from) {
+ if (tok) {
+ int sum_type = parse_csum_name(tok, to - tok, 0);
+ if (sum_type >= 0 && !saw[sum_type])
+ saw[sum_type] = ++cnt;
+ else
+ to = tok - (tok != sumbuf);
+ tok = NULL;
+ }
+ if (!*from++)
+ break;
+ continue;
+ }
+ if (!tok) {
+ if (to != sumbuf)
+ *to++ = ' ';
+ tok = to;
+ }
+ if (to - sumbuf >= sumbuf_len - 1) {
+ to = tok - (tok != sumbuf);
+ break;
+ }
+ *to++ = *from++;
+ }
+ *to = '\0';
+
+ return to - sumbuf;
}
-void negotiate_checksum(int f_in, int f_out, const char *csum_list)
+void negotiate_checksum(int f_in, int f_out, const char *csum_list, int saw_fail)
{
- char *tok, sumbuf[MAX_CHECKSUM_LIST];
+ char *tok, sumbuf[MAX_CHECKSUM_LIST], saw[CSUM_SAW_BUFLEN];
int sum_type, len;
- if (!am_server || local_server) {
- if (!csum_list || !*csum_list)
- csum_list = default_checksum_list;
- len = strlen(csum_list);
- if (len >= (int)sizeof sumbuf) {
- rprintf(FERROR, "The checksum list is too long.\n");
- exit_cleanup(RERR_UNSUPPORTED);
+ /* Simplify the user-provided string so that it contains valid
+ * checksum names without any duplicates. The client side also
+ * makes use of the saw values when scanning the server's list. */
+ if (csum_list && *csum_list && (!am_server || local_server)) {
+ len = parse_checksum_list(csum_list, sumbuf, sizeof sumbuf, saw);
+ if (saw_fail && !len)
+ len = strlcpy(sumbuf, "FAIL", sizeof sumbuf);
+ csum_list = sumbuf;
+ } else
+ csum_list = NULL;
+
+ if (!csum_list || !*csum_list) {
+ struct csum_struct *cs;
+ for (tok = sumbuf, cs = valid_checksums, len = 0; cs->name; cs++) {
+ if (cs->num == CSUM_NONE)
+ continue;
+ if (tok != sumbuf)
+ *tok++ = ' ';
+ tok += strlcpy(tok, cs->name, sizeof sumbuf - (tok - sumbuf));
+ saw[cs->num] = ++len;
}
- if (!local_server)
- write_vstring(f_out, csum_list, len);
+ *tok = '\0';
+ len = tok - sumbuf;
}
- if (local_server && !read_batch)
- memcpy(sumbuf, csum_list, len+1);
- else
+ /* Each side sends their list of valid checksum names to the other side and
+ * then both sides pick the first name in the client's list that is also in
+ * the server's list. */
+ if (!local_server)
+ write_vstring(f_out, sumbuf, len);
+
+ if (!local_server || read_batch)
len = read_vstring(f_in, sumbuf, sizeof sumbuf);
if (len > 0) {
+ int best = CSUM_SAW_BUFLEN; /* We want best == 1 from the client list */
+ if (am_server)
+ memset(saw, 1, CSUM_SAW_BUFLEN); /* The first client's choice is the best choice */
for (tok = strtok(sumbuf, " \t"); tok; tok = strtok(NULL, " \t")) {
- len = strlen(tok);
- sum_type = parse_csum_name(tok, len, 0);
- if (sum_type >= CSUM_NONE) {
- xfersum_type = checksum_type = sum_type;
- if (am_server && !local_server)
- write_vstring(f_out, tok, len);
- negotiated_csum_name = strdup(tok);
- return;
- }
+ sum_type = parse_csum_name(tok, -1, 0);
+ if (sum_type < 0 || !saw[sum_type] || best < saw[sum_type])
+ continue;
+ xfersum_type = checksum_type = sum_type;
+ negotiated_csum_name = tok;
+ best = saw[sum_type];
+ if (best == 1)
+ break;
+ }
+ if (negotiated_csum_name) {
+ negotiated_csum_name = strdup(negotiated_csum_name);
+ return;
}
}
+ if (!am_server)
+ msleep(20);
rprintf(FERROR, "Failed to negotiate a common checksum\n");
exit_cleanup(RERR_UNSUPPORTED);
}
@@ -144,6 +254,10 @@ int csum_len_for_type(int cst, BOOL flist_csum)
return MD4_DIGEST_LEN;
case CSUM_MD5:
return MD5_DIGEST_LEN;
+#ifdef SUPPORT_XXHASH
+ case CSUM_XXHASH:
+ return sizeof (XXH64_hash_t);
+#endif
default: /* paranoia to prevent missing case values */
exit_cleanup(RERR_UNSUPPORTED);
}
@@ -242,6 +356,11 @@ void get_checksum2(char *buf, int32 len, char *sum)
mdfour_result(&m, (uchar *)sum);
break;
}
+#ifdef SUPPORT_XXHASH
+ case CSUM_XXHASH:
+ SIVAL64(sum, 0, XXH64(buf, len, checksum_seed));
+ break;
+#endif
default: /* paranoia to prevent missing case values */
exit_cleanup(RERR_UNSUPPORTED);
}
@@ -298,6 +417,34 @@ void file_checksum(const char *fname, const STRUCT_STAT *st_p, char *sum)
mdfour_result(&m, (uchar *)sum);
break;
+#ifdef SUPPORT_XXHASH
+ case CSUM_XXHASH: {
+ XXH64_state_t* state = XXH64_createState();
+ if (state == NULL)
+ out_of_memory("file_checksum xx64");
+
+ if (XXH64_reset(state, 0) == XXH_ERROR) {
+ rprintf(FERROR, "error resetting XXH64 seed");
+ exit_cleanup(RERR_STREAMIO);
+ }
+
+ for (i = 0; i + CSUM_CHUNK <= len; i += CSUM_CHUNK) {
+ XXH_errorcode const updateResult =
+ XXH64_update(state, (uchar *)map_ptr(buf, i, CSUM_CHUNK), CSUM_CHUNK);
+ if (updateResult == XXH_ERROR) {
+ rprintf(FERROR, "error computing XX64 hash");
+ exit_cleanup(RERR_STREAMIO);
+ }
+ }
+ remainder = (int32)(len - i);
+ if (remainder > 0)
+ XXH64_update(state, (uchar *)map_ptr(buf, i, CSUM_CHUNK), remainder);
+ SIVAL64(sum, 0, XXH64_digest(state));
+
+ XXH64_freeState(state);
+ break;
+ }
+#endif
default:
rprintf(FERROR, "invalid checksum-choice for the --checksum option (%d)\n", checksum_type);
exit_cleanup(RERR_UNSUPPORTED);
@@ -310,6 +457,9 @@ void file_checksum(const char *fname, const STRUCT_STAT *st_p, char *sum)
static int32 sumresidue;
static md_context md;
static int cursum_type;
+#ifdef SUPPORT_XXHASH
+XXH64_state_t* xxh64_state = NULL;
+#endif
void sum_init(int csum_type, int seed)
{
@@ -335,6 +485,19 @@ void sum_init(int csum_type, int seed)
SIVAL(s, 0, seed);
sum_update(s, 4);
break;
+#ifdef SUPPORT_XXHASH
+ case CSUM_XXHASH:
+ if (xxh64_state == NULL) {
+ xxh64_state = XXH64_createState();
+ if (xxh64_state == NULL)
+ out_of_memory("sum_init xxh64");
+ }
+ if (XXH64_reset(xxh64_state, 0) == XXH_ERROR) {
+ rprintf(FERROR, "error resetting XXH64 state");
+ exit_cleanup(RERR_STREAMIO);
+ }
+ break;
+#endif
case CSUM_NONE:
break;
default: /* paranoia to prevent missing case values */
@@ -384,6 +547,14 @@ void sum_update(const char *p, int32 len)
if (sumresidue)
memcpy(md.buffer, p, sumresidue);
break;
+#ifdef SUPPORT_XXHASH
+ case CSUM_XXHASH:
+ if (XXH64_update(xxh64_state, p, len) == XXH_ERROR) {
+ rprintf(FERROR, "error computing XX64 hash");
+ exit_cleanup(RERR_STREAMIO);
+ }
+ break;
+#endif
case CSUM_NONE:
break;
default: /* paranoia to prevent missing case values */
@@ -412,6 +583,11 @@ int sum_end(char *sum)
mdfour_update(&md, (uchar *)md.buffer, sumresidue);
mdfour_result(&md, (uchar *)sum);
break;
+#ifdef SUPPORT_XXHASH
+ case CSUM_XXHASH:
+ SIVAL64(sum, 0, XXH64_digest(xxh64_state));
+ break;
+#endif
case CSUM_NONE:
*sum = '\0';
break;
diff --git a/compat.c b/compat.c
index b29b9637..8c77ea69 100644
--- a/compat.c
+++ b/compat.c
@@ -367,24 +367,25 @@ void setup_protocol(int f_out,int f_in)
}
#endif
- if (am_server) {
- if (!checksum_seed)
- checksum_seed = time(NULL) ^ (getpid() << 6);
- write_int(f_out, checksum_seed);
- } else {
- checksum_seed = read_int(f_in);
- }
-
if (!checksum_choice) {
const char *rcl = getenv("RSYNC_CHECKSUM_LIST");
+ int saw_fail = rcl && strstr(rcl, "FAIL");
if (csum_exchange)
- negotiate_checksum(f_in, f_out, rcl);
- else if (!am_server && rcl && *rcl && strstr(rcl, "FAIL")) {
+ negotiate_checksum(f_in, f_out, rcl, saw_fail);
+ else if (!am_server && saw_fail) {
rprintf(FERROR, "Remote rsync is too old for checksum negotation\n");
exit_cleanup(RERR_UNSUPPORTED);
}
}
+ if (am_server) {
+ if (!checksum_seed)
+ checksum_seed = time(NULL) ^ (getpid() << 6);
+ write_int(f_out, checksum_seed);
+ } else {
+ checksum_seed = read_int(f_in);
+ }
+
init_flist();
}
diff --git a/configure.ac b/configure.ac
index 554cf4ac..394f5b52 100644
--- a/configure.ac
+++ b/configure.ac
@@ -381,9 +381,21 @@ AC_CHECK_HEADERS(sys/fcntl.h sys/select.h fcntl.h sys/time.h sys/unistd.h \
netdb.h malloc.h float.h limits.h iconv.h libcharset.h langinfo.h \
sys/acl.h acl/libacl.h attr/xattr.h sys/xattr.h sys/extattr.h \
popt.h popt/popt.h linux/falloc.h netinet/in_systm.h netinet/ip.h \
- zlib.h)
+ zlib.h xxhash.h)
AC_HEADER_MAJOR_FIXED
+dnl Do you want to disable use of xxhash checksums
+AC_ARG_ENABLE([xxhash],
+ AS_HELP_STRING([--disable-xxhash],[disable xxhash checksums]))
+AH_TEMPLATE([SUPPORT_XXHASH],
+[Undefine if you do not want xxhash checksums. By default this is defined.])
+if test x"$enable_xxhash" != x"no"; then
+ if test x"$ac_cv_header_xxhash_h" = x"yes"; then
+ AC_SEARCH_LIBS(XXH64_createState, xxhash)
+ AC_DEFINE(SUPPORT_XXHASH)
+ fi
+fi
+
AC_CACHE_CHECK([if makedev takes 3 args],rsync_cv_MAKEDEV_TAKES_3_ARGS,[
AC_RUN_IFELSE([AC_LANG_SOURCE([[
#include <sys/types.h>
diff --git a/flist.c b/flist.c
index 1f2b278d..940071fd 100644
--- a/flist.c
+++ b/flist.c
@@ -142,7 +142,7 @@ void init_flist(void)
rprintf(FINFO, "FILE_STRUCT_LEN=%d, EXTRA_LEN=%d\n",
(int)FILE_STRUCT_LEN, (int)EXTRA_LEN);
}
- parse_checksum_choice(); /* Sets checksum_type && xfersum_type */
+ parse_checksum_choice(1); /* Sets checksum_type && xfersum_type */
flist_csum_len = csum_len_for_type(checksum_type, 1);
show_filelist_progress = INFO_GTE(FLIST, 1) && xfer_dirs && !am_server && !inc_recurse;
diff --git a/io.c b/io.c
index 446a5f34..189bc232 100644
--- a/io.c
+++ b/io.c
@@ -2369,8 +2369,8 @@ void start_write_batch(int fd)
write_int(batch_fd, protocol_version);
if (protocol_version >= 30)
write_varint(batch_fd, compat_flags);
- write_int(batch_fd, checksum_seed);
maybe_write_checksum(batch_fd);
+ write_int(batch_fd, checksum_seed);
if (am_sender)
write_batch_monitor_out = fd;
diff --git a/options.c b/options.c
index e2adcf83..959c4205 100644
--- a/options.c
+++ b/options.c
@@ -276,6 +276,7 @@ static struct output_struct debug_words[COUNT_DEBUG+1] = {
DEBUG_WORD(CHDIR, W_CLI|W_SRV, "Debug when the current directory changes"),
DEBUG_WORD(CONNECT, W_CLI, "Debug connection events (levels 1-2)"),
DEBUG_WORD(CMD, W_CLI, "Debug commands+options that are issued (levels 1-2)"),
+ DEBUG_WORD(CSUM, W_CLI|W_SRV, "Debug checksum negotiation"),
DEBUG_WORD(DEL, W_REC, "Debug delete actions (levels 1-3)"),
DEBUG_WORD(DELTASUM, W_SND|W_REC, "Debug delta-transfer checksumming (levels 1-4)"),
DEBUG_WORD(DUP, W_REC, "Debug weeding of duplicate names"),
@@ -579,6 +580,7 @@ static void print_rsync_version(enum logcode f)
char const *iconv = "no ";
char const *ipv6 = "no ";
char const *simd = "no ";
+ char const *xxhash = "no ";
STRUCT_STAT *dumstat;
#if SUBPROTOCOL_VERSION != 0
@@ -618,6 +620,9 @@ static void print_rsync_version(enum logcode f)
#ifdef HAVE_SIMD
simd = "";
#endif
+#ifdef SUPPORT_XXHASH
+ xxhash = "";
+#endif
rprintf(f, "%s version %s protocol version %d%s\n",
RSYNC_NAME, RSYNC_VERSION, PROTOCOL_VERSION, subprotocol);
@@ -631,8 +636,8 @@ static void print_rsync_version(enum logcode f)
(int)(sizeof (int64) * 8));
rprintf(f, " %ssocketpairs, %shardlinks, %ssymlinks, %sIPv6, batchfiles, %sinplace,\n",
got_socketpair, hardlinks, links, ipv6, have_inplace);
- rprintf(f, " %sappend, %sACLs, %sxattrs, %siconv, %ssymtimes, %sprealloc, %sSIMD\n",
- have_inplace, acls, xattrs, iconv, symtimes, prealloc, simd);
+ rprintf(f, " %sappend, %sACLs, %sxattrs, %siconv, %ssymtimes, %sprealloc, %sSIMD, %sxxhash\n",
+ have_inplace, acls, xattrs, iconv, symtimes, prealloc, simd, xxhash);
#ifdef MAINTAINER_MODE
rprintf(f, "Panic Action: \"%s\"\n", get_panic_action());
@@ -1932,7 +1937,7 @@ int parse_arguments(int *argc_p, const char ***argv_p)
/* Call this early to verify the args and figure out if we need to force
* --whole-file. Note that the parse function will get called again later,
* just in case an "auto" choice needs to know the protocol_version. */
- parse_checksum_choice();
+ parse_checksum_choice(0);
} else
checksum_choice = NULL;
diff --git a/rsync.h b/rsync.h
index 0eb4fbc0..d6c5de52 100644
--- a/rsync.h
--
The rsync repository.
More information about the rsync-cvs
mailing list