+++ /dev/null
-This patch changes the way the --checksum option works by having the
-receiving side perform a checksum-read of every file in the file list
-as the list is received (if the sizes are equal), marking non-matching
-items with a flag. The idea is that the checksum pass on the sender and
-the receiver can then happen in parallel instead of having the reciever
-to its checksum pass during its normal find-the-different-files pass.
-
-I have benchmarked this a little, and it appears to slow things down
-for a local copy, so the old algorithm is used for local copies.
-
-To use this patch, run these commands for a successful build:
-
- patch -p1 <patches/early-checksum.diff
- ./configure (optional if already run)
- make
-
---- old/flist.c
-+++ new/flist.c
-@@ -34,6 +34,7 @@ extern int am_generator;
- extern int inc_recurse;
- extern int do_progress;
- extern int always_checksum;
-+extern int pre_checksum;
- extern int module_id;
- extern int ignore_errors;
- extern int numeric_ids;
-@@ -949,6 +950,15 @@ static struct file_struct *recv_file_ent
- memcpy(bp, F_SUM(first), checksum_len);
- } else
- read_buf(f, bp, checksum_len);
-+ if (pre_checksum && bp != tmp_sum) {
-+ STRUCT_STAT st;
-+ char *fname = f_name(file, NULL);
-+ if (stat(fname, &st) == 0 && st.st_size == file_length) {
-+ file_checksum(fname, tmp_sum, st.st_size);
-+ if (memcmp(bp, tmp_sum, checksum_len) != 0)
-+ file->flags |= FLAG_SUM_DIFFERS;
-+ }
-+ }
- }
-
- #ifdef SUPPORT_ACLS
---- old/generator.c
-+++ new/generator.c
-@@ -73,6 +73,7 @@ extern int protocol_version;
- extern int file_total;
- extern int fuzzy_basis;
- extern int always_checksum;
-+extern int pre_checksum;
- extern int checksum_len;
- extern char *partial_dir;
- extern char *basis_dir[];
-@@ -617,7 +618,8 @@ void itemize(const char *fnamecmp, struc
-
-
- /* Perform our quick-check heuristic for determining if a file is unchanged. */
--int unchanged_file(char *fn, struct file_struct *file, STRUCT_STAT *st)
-+int unchanged_file(char *fn, int fnamecmp_type, struct file_struct *file,
-+ STRUCT_STAT *st)
- {
- if (st->st_size != F_LENGTH(file))
- return 0;
-@@ -626,6 +628,8 @@ int unchanged_file(char *fn, struct file
- of the file time to determine whether to sync */
- if (always_checksum > 0 && S_ISREG(st->st_mode)) {
- char sum[MAX_DIGEST_LEN];
-+ if (pre_checksum && fnamecmp_type == FNAMECMP_FNAME)
-+ return !(file->flags & FLAG_SUM_DIFFERS);
- file_checksum(fn, sum, st->st_size);
- return memcmp(sum, F_SUM(file), checksum_len) == 0;
- }
-@@ -846,7 +850,7 @@ static int try_dests_reg(struct file_str
- match_level = 1;
- /* FALL THROUGH */
- case 1:
-- if (!unchanged_file(cmpbuf, file, &sxp->st))
-+ if (!unchanged_file(cmpbuf, 0, file, &sxp->st))
- continue;
- best_match = j;
- match_level = 2;
-@@ -1638,7 +1642,7 @@ static void recv_generator(char *fname,
- ;
- else if (fnamecmp_type == FNAMECMP_FUZZY)
- ;
-- else if (unchanged_file(fnamecmp, file, &sx.st)) {
-+ else if (unchanged_file(fnamecmp, fnamecmp_type, file, &sx.st)) {
- if (partialptr) {
- do_unlink(partialptr);
- handle_partial_dir(partialptr, PDIR_DELETE);
---- old/hlink.c
-+++ new/hlink.c
-@@ -353,7 +353,7 @@ int hard_link_check(struct file_struct *
- }
- break;
- }
-- if (!unchanged_file(cmpbuf, file, &alt_sx.st))
-+ if (!unchanged_file(cmpbuf, 0, file, &alt_sx.st))
- continue;
- statret = 1;
- if (unchanged_attrs(cmpbuf, file, &alt_sx))
---- old/main.c
-+++ new/main.c
-@@ -46,6 +46,7 @@ extern int module_id;
- extern int copy_links;
- extern int copy_dirlinks;
- extern int keep_dirlinks;
-+extern int always_checksum;
- extern int preserve_hard_links;
- extern int protocol_version;
- extern int file_total;
-@@ -78,6 +79,9 @@ extern struct filter_list_struct server_
- extern iconv_t ic_send;
- #endif
-
-+extern char curr_dir[MAXPATHLEN];
-+
-+int pre_checksum = 0;
- int local_server = 0;
- int daemon_over_rsh = 0;
- mode_t orig_umask = 0;
-@@ -850,6 +854,7 @@ static void do_server_recv(int f_in, int
- struct file_list *flist;
- char *local_name = NULL;
- char *dir = NULL;
-+ char olddir[sizeof curr_dir];
- int save_verbose = verbose;
-
- if (filesfrom_fd >= 0) {
-@@ -896,6 +901,10 @@ static void do_server_recv(int f_in, int
- filesfrom_fd = -1;
- }
-
-+ strlcpy(olddir, curr_dir, sizeof olddir);
-+ if (always_checksum && !local_server && argc > 0)
-+ pre_checksum = push_dir(argv[0], 0);
-+
- flist = recv_file_list(f_in);
- if (!flist) {
- rprintf(FERROR,"server_recv: recv_file_list error\n");
-@@ -905,6 +914,9 @@ static void do_server_recv(int f_in, int
- recv_additional_file_list(f_in);
- verbose = save_verbose;
-
-+ if (pre_checksum)
-+ pop_dir(olddir);
-+
- if (argc > 0)
- local_name = get_local_name(flist,argv[0]);
-
-@@ -981,6 +993,7 @@ int client_run(int f_in, int f_out, pid_
- {
- struct file_list *flist = NULL;
- int exit_code = 0, exit_code2 = 0;
-+ char olddir[sizeof curr_dir];
- char *local_name = NULL;
-
- cleanup_child_pid = pid;
-@@ -1058,12 +1071,19 @@ int client_run(int f_in, int f_out, pid_
- filesfrom_fd = -1;
- }
-
-+ strlcpy(olddir, curr_dir, sizeof olddir);
-+ if (always_checksum && !local_server)
-+ pre_checksum = push_dir(argv[0], 0);
-+
- if (write_batch && !am_server)
- start_write_batch(f_in);
- flist = recv_file_list(f_in);
- if (inc_recurse && file_total == 1)
- recv_additional_file_list(f_in);
-
-+ if (pre_checksum)
-+ pop_dir(olddir);
-+
- if (flist && flist->used > 0) {
- local_name = get_local_name(flist, argv[0]);
-
---- old/rsync.h
-+++ new/rsync.h
-@@ -73,6 +73,7 @@
- #define FLAG_HLINK_DONE (1<<8) /* receiver/generator */
- #define FLAG_LENGTH64 (1<<9) /* sender/receiver/generator */
- #define FLAG_SKIP_GROUP (1<<10) /* receiver/generator */
-+#define FLAG_SUM_DIFFERS (1<<11)/* receiver/generator */
-
- /* These flags are passed to functions but not stored. */
-
Matt McCutchen's Web Site / rsync / rsync-patches.git / blobdiff