search for: to_child_pipe

...-daemon mode, it forks the SSH process in pipe.c function. For the communication between the parent rsync and child SSH, it uses the socket pairs. As socket pairs are bidirectional, only one pair can be used for communication between the parent and child. e.g. socketpair(xx,xx,xx,xx) returns -> to_child_pipe[2]={3,4} then these two file descriptors 3 and 4 can be used one at each end for reading and writing. But in the rsync code, two pairs are created as to_child_pipe and from_child_pipe ------------------------------------------------------------ --------------------- if (fd_pair(to_child_pipe) &...

...0.6.fast/pipe.c 2009-09-23 14:40:42.000000000 +0300 @@ -29,6 +29,7 @@ extern mode_t orig_umask; extern char *logfile_name; extern struct chmod_mode_struct *chmod_modes; +int local_socket = 0; /** * Create a child connected to us via its stdin/stdout. @@ -111,11 +112,15 @@ pid_t pid; int to_child_pipe[2]; int from_child_pipe[2]; + int child_socket[2]; /* The parent process is always the sender for a local rsync. */ assert(am_sender); if (fd_pair(to_child_pipe) < 0 || +#ifdef HAVE_SOCKETPAIR + fd_pair(child_socket) < 0 || +#endif fd_pair(from_child_pipe) < 0) { rsys...

...bler - rsync: pipe: Operation not permitted (1) DEBUG: com.mddc.client.rsync.StreamGobbler - rsync error: error in IPC code (code 14) at /home/lapo/packaging/rsync-3.0.4-1/src/rsync-3.0.4/pipe.c(57) [sender=3.0.4] * The rsync code where the error happens is pipe.c(57) which is this. if (fd_pair(to_child_pipe) < 0 || fd_pair(from_child_pipe) < 0) { rsyserr(FERROR, errno, "pipe"); exit_cleanup(RERR_IPC); } So for some reason fd_pair(to_child_pipe) is < 0 or fd_pair(from_child_pipe) < 0. If anyone has any suggestions it would be great as I'm stuck now. Thank...

> Just curious: what about the rsync+ patch? Thanks for the reminder. I've just committed Jos's rsync+ patch onto the "branch_mbp_rsyncplus_merge" branch. If it works OK and nobody screams I will move it across onto the main tree tomorrow or Wednesday. I see the patch doesn't add documentation about the new options to the man page, so we should fix that in the future.

...* processes, so we must make sure that only one + * actually writes. It shouldn't matter which one, + * because they must produce the same file. + */ + write_batch = 0; + if (!am_sender) filesfrom_fd = -1; *************** *** 129,134 **** --- 150,156 ---- } if (to_child_pipe[0] != STDIN_FILENO) close(to_child_pipe[0]); if (from_child_pipe[1] != STDOUT_FILENO) close(from_child_pipe[1]); + if (read_batch) exit_cleanup(0); /* no reason to continue */ child_main(argc, argv); }

...ations like the one on * Solaris. + * + * If register_child is nonzero then the child is registered for autocleanup. **/ -pid_t piped_child(char **command, int *f_in, int *f_out) +pid_t piped_child(char **command, int *f_in, int *f_out, int blocking_io, int register_child) { pid_t pid; int to_child_pipe[2]; @@ -57,7 +58,7 @@ pid_t piped_child(char **command, int *f } - pid = do_fork(); + pid = register_child ? do_fork() : fork(); if (pid == -1) { rprintf(FERROR, "fork: %s\n", strerror(errno)); exit_cleanup(RERR_IPC); diff -rupNP rsync-2.6.2/syscall.c rsync-2.6.2-fnameconv.c...