Got rid of RERR_LOG_FAILURE.
[rsync.git] / io.c
diff --git a/io.c b/io.c
index 2f4162cdac88ecef72a2a4ea4b617c2ab2dac9f8..fdaa71345ae07166087d8302de3a9f7a5b206b8d 100644 (file)
--- a/io.c
+++ b/io.c
-/* 
-   Copyright (C) Andrew Tridgell 1996
-   Copyright (C) Paul Mackerras 1996
-   
-   This program is free software; you can redistribute it and/or modify
-   it under the terms of the GNU General Public License as published by
-   the Free Software Foundation; either version 2 of the License, or
-   (at your option) any later version.
-   
-   This program is distributed in the hope that it will be useful,
-   but WITHOUT ANY WARRANTY; without even the implied warranty of
-   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-   GNU General Public License for more details.
-   
-   You should have received a copy of the GNU General Public License
-   along with this program; if not, write to the Free Software
-   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
-*/
+/* -*- c-file-style: "linux" -*-
+ *
+ * Copyright (C) 1996-2001 by Andrew Tridgell
+ * Copyright (C) Paul Mackerras 1996
+ * Copyright (C) 2001, 2002 by Martin Pool <mbp@samba.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+/**
+ * @file io.c
+ *
+ * Socket and pipe I/O utilities used in rsync.
+ *
+ * rsync provides its own multiplexing system, which is used to send
+ * stderr and stdout over a single socket.  We need this because
+ * stdout normally carries the binary data stream, and stderr all our
+ * error messages.
+ *
+ * For historical reasons this is off during the start of the
+ * connection, but it's switched on quite early using
+ * io_start_multiplex_out() and io_start_multiplex_in().
+ **/
 
-/*
-  socket and pipe IO utilities used in rsync 
-
-  tridge, June 1996
-  */
 #include "rsync.h"
 
-/* if no timeout is specified then use a 60 second select timeout */
+/** If no timeout is specified then use a 60 second select timeout */
 #define SELECT_TIMEOUT 60
 
 extern int bwlimit;
+extern size_t bwlimit_writemax;
+extern int verbose;
+extern int io_timeout;
+extern int allowed_lull;
+extern int am_server;
+extern int am_daemon;
+extern int am_sender;
+extern int am_generator;
+extern int eol_nulls;
+extern int csum_length;
+extern int checksum_seed;
+extern int protocol_version;
+extern int remove_sent_files;
+extern int preserve_hard_links;
+extern char *filesfrom_host;
+extern struct stats stats;
+extern struct file_list *the_file_list;
+
+const char phase_unknown[] = "unknown";
+int select_timeout = SELECT_TIMEOUT;
+int ignore_timeout = 0;
+int batch_fd = -1;
+int batch_gen_fd = -1;
+
+/**
+ * The connection might be dropped at some point; perhaps because the
+ * remote instance crashed.  Just giving the offset on the stream is
+ * not very helpful.  So instead we try to make io_phase_name point to
+ * something useful.
+ *
+ * For buffered/multiplexed I/O these names will be somewhat
+ * approximate; perhaps for ease of support we would rather make the
+ * buffer always flush when a single application-level I/O finishes.
+ *
+ * @todo Perhaps we want some simple stack functionality, but there's
+ * no need to overdo it.
+ **/
+const char *io_write_phase = phase_unknown;
+const char *io_read_phase = phase_unknown;
+
+/* Ignore an EOF error if non-zero. See whine_about_eof(). */
+int kluge_around_eof = 0;
+
+int msg_fd_in = -1;
+int msg_fd_out = -1;
+int sock_f_in = -1;
+int sock_f_out = -1;
 
 static int io_multiplexing_out;
 static int io_multiplexing_in;
-static int multiplex_in_fd;
-static int multiplex_out_fd;
 static time_t last_io;
-static int eof_error=1;
-extern int verbose;
-extern int io_timeout;
-extern struct stats stats;
+static int no_flush;
+
+static int write_batch_monitor_in = -1;
+static int write_batch_monitor_out = -1;
+
+static int io_filesfrom_f_in = -1;
+static int io_filesfrom_f_out = -1;
+static char io_filesfrom_buf[2048];
+static char *io_filesfrom_bp;
+static char io_filesfrom_lastchar;
+static int io_filesfrom_buflen;
+static size_t contiguous_write_len = 0;
+
+static void read_loop(int fd, char *buf, size_t len);
+
+struct flist_ndx_item {
+       struct flist_ndx_item *next;
+       int ndx;
+};
+
+struct flist_ndx_list {
+       struct flist_ndx_item *head, *tail;
+};
+
+static struct flist_ndx_list redo_list, hlink_list;
+
+struct msg_list {
+       struct msg_list *next;
+       char *buf;
+       int len;
+};
+
+static struct msg_list *msg_list_head;
+static struct msg_list *msg_list_tail;
 
-static int io_error_fd = -1;
+static void flist_ndx_push(struct flist_ndx_list *lp, int ndx)
+{
+       struct flist_ndx_item *item;
+
+       if (!(item = new(struct flist_ndx_item)))
+               out_of_memory("flist_ndx_push");
+       item->next = NULL;
+       item->ndx = ndx;
+       if (lp->tail)
+               lp->tail->next = item;
+       else
+               lp->head = item;
+       lp->tail = item;
+}
+
+static int flist_ndx_pop(struct flist_ndx_list *lp)
+{
+       struct flist_ndx_item *next;
+       int ndx;
+
+       if (!lp->head)
+               return -1;
 
-static void read_loop(int fd, char *buf, int len);
+       ndx = lp->head->ndx;
+       next = lp->head->next;
+       free(lp->head);
+       lp->head = next;
+       if (!next)
+               lp->tail = NULL;
+
+       return ndx;
+}
 
 static void check_timeout(void)
 {
-       extern int am_server, am_daemon;
        time_t t;
-       
-       if (!io_timeout) return;
+
+       if (!io_timeout || ignore_timeout)
+               return;
 
        if (!last_io) {
                last_io = time(NULL);
@@ -57,133 +173,525 @@ static void check_timeout(void)
 
        t = time(NULL);
 
-       if (last_io && io_timeout && (t-last_io) >= io_timeout) {
+       if (t - last_io >= io_timeout) {
                if (!am_server && !am_daemon) {
-                       rprintf(FERROR,"io timeout after %d second - exiting\n", 
+                       rprintf(FERROR, "io timeout after %d seconds -- exiting\n",
                                (int)(t-last_io));
                }
                exit_cleanup(RERR_TIMEOUT);
        }
 }
 
-/* setup the fd used to propogate errors */
-void io_set_error_fd(int fd)
+/* Note the fds used for the main socket (which might really be a pipe
+ * for a local transfer, but we can ignore that). */
+void io_set_sock_fds(int f_in, int f_out)
+{
+       sock_f_in = f_in;
+       sock_f_out = f_out;
+}
+
+/* Setup the fd used to receive MSG_* messages.  Only needed during the
+ * early stages of being a local sender (up through the sending of the
+ * file list) or when we're the generator (to fetch the messages from
+ * the receiver). */
+void set_msg_fd_in(int fd)
+{
+       msg_fd_in = fd;
+}
+
+/* Setup the fd used to send our MSG_* messages.  Only needed when
+ * we're the receiver (to send our messages to the generator). */
+void set_msg_fd_out(int fd)
+{
+       msg_fd_out = fd;
+       set_nonblocking(msg_fd_out);
+}
+
+/* Add a message to the pending MSG_* list. */
+static void msg_list_add(int code, char *buf, int len)
 {
-       io_error_fd = fd;
+       struct msg_list *ml;
+
+       if (!(ml = new(struct msg_list)))
+               out_of_memory("msg_list_add");
+       ml->next = NULL;
+       if (!(ml->buf = new_array(char, len+4)))
+               out_of_memory("msg_list_add");
+       SIVAL(ml->buf, 0, ((code+MPLEX_BASE)<<24) | len);
+       memcpy(ml->buf+4, buf, len);
+       ml->len = len+4;
+       if (msg_list_tail)
+               msg_list_tail->next = ml;
+       else
+               msg_list_head = ml;
+       msg_list_tail = ml;
 }
 
-/* read some data from the error fd and write it to the write log code */
-static void read_error_fd(void)
+void send_msg(enum msgcode code, char *buf, int len)
 {
-       char buf[200];
-       int n;
-       int fd = io_error_fd;
+       if (msg_fd_out < 0) {
+               io_multiplex_write(code, buf, len);
+               return;
+       }
+       msg_list_add(code, buf, len);
+       msg_list_push(NORMAL_FLUSH);
+}
+
+/* Read a message from the MSG_* fd and handle it.  This is called either
+ * during the early stages of being a local sender (up through the sending
+ * of the file list) or when we're the generator (to fetch the messages
+ * from the receiver). */
+static void read_msg_fd(void)
+{
+       char buf[2048];
+       size_t n;
+       int fd = msg_fd_in;
        int tag, len;
 
-       io_error_fd = -1;
+       /* Temporarily disable msg_fd_in.  This is needed to avoid looping back
+        * to this routine from writefd_unbuffered(). */
+       msg_fd_in = -1;
 
        read_loop(fd, buf, 4);
        tag = IVAL(buf, 0);
 
        len = tag & 0xFFFFFF;
-       tag = tag >> 24;
-       tag -= MPLEX_BASE;
+       tag = (tag >> 24) - MPLEX_BASE;
 
-       while (len) {
-               n = len;
-               if (n > (sizeof(buf)-1)) n = sizeof(buf)-1;
-               read_loop(fd, buf, n);
-               rwrite((enum logcode)tag, buf, n);
-               len -= n;
+       switch (tag) {
+       case MSG_DONE:
+               if (len != 0 || !am_generator) {
+                       rprintf(FERROR, "invalid message %d:%d\n", tag, len);
+                       exit_cleanup(RERR_STREAMIO);
+               }
+               flist_ndx_push(&redo_list, -1);
+               break;
+       case MSG_REDO:
+               if (len != 4 || !am_generator) {
+                       rprintf(FERROR, "invalid message %d:%d\n", tag, len);
+                       exit_cleanup(RERR_STREAMIO);
+               }
+               read_loop(fd, buf, 4);
+               flist_ndx_push(&redo_list, IVAL(buf,0));
+               break;
+       case MSG_DELETED:
+               if (len >= (int)sizeof buf || !am_generator) {
+                       rprintf(FERROR, "invalid message %d:%d\n", tag, len);
+                       exit_cleanup(RERR_STREAMIO);
+               }
+               read_loop(fd, buf, len);
+               io_multiplex_write(MSG_DELETED, buf, len);
+               break;
+       case MSG_SUCCESS:
+               if (len != 4 || !am_generator) {
+                       rprintf(FERROR, "invalid message %d:%d\n", tag, len);
+                       exit_cleanup(RERR_STREAMIO);
+               }
+               read_loop(fd, buf, len);
+               if (remove_sent_files)
+                       io_multiplex_write(MSG_SUCCESS, buf, len);
+               if (preserve_hard_links)
+                       flist_ndx_push(&hlink_list, IVAL(buf,0));
+               break;
+       case MSG_INFO:
+       case MSG_ERROR:
+       case MSG_LOG:
+               while (len) {
+                       n = len;
+                       if (n >= sizeof buf)
+                               n = sizeof buf - 1;
+                       read_loop(fd, buf, n);
+                       rwrite((enum logcode)tag, buf, n);
+                       len -= n;
+               }
+               break;
+       default:
+               rprintf(FERROR, "unknown message %d:%d\n", tag, len);
+               exit_cleanup(RERR_STREAMIO);
        }
 
-       io_error_fd = fd;
+       msg_fd_in = fd;
 }
 
+/* Try to push messages off the list onto the wire.  If we leave with more
+ * to do, return 0.  On error, return -1.  If everything flushed, return 1.
+ * This is only active in the receiver. */
+int msg_list_push(int flush_it_all)
+{
+       static int written = 0;
+       struct timeval tv;
+       fd_set fds;
 
-static int no_flush;
+       if (msg_fd_out < 0)
+               return -1;
+
+       while (msg_list_head) {
+               struct msg_list *ml = msg_list_head;
+               int n = write(msg_fd_out, ml->buf + written, ml->len - written);
+               if (n < 0) {
+                       if (errno == EINTR)
+                               continue;
+                       if (errno != EWOULDBLOCK && errno != EAGAIN)
+                               return -1;
+                       if (!flush_it_all)
+                               return 0;
+                       FD_ZERO(&fds);
+                       FD_SET(msg_fd_out, &fds);
+                       tv.tv_sec = select_timeout;
+                       tv.tv_usec = 0;
+                       if (!select(msg_fd_out+1, NULL, &fds, NULL, &tv))
+                               check_timeout();
+               } else if ((written += n) == ml->len) {
+                       free(ml->buf);
+                       msg_list_head = ml->next;
+                       if (!msg_list_head)
+                               msg_list_tail = NULL;
+                       free(ml);
+                       written = 0;
+               }
+       }
+       return 1;
+}
 
-/* read from a socket with IO timeout. return the number of
-   bytes read. If no bytes can be read then exit, never return
-   a number <= 0 */
-static int read_timeout(int fd, char *buf, int len)
+int get_redo_num(int itemizing, enum logcode code)
 {
-       int n, ret=0;
+       while (1) {
+               if (hlink_list.head)
+                       check_for_finished_hlinks(itemizing, code);
+               if (redo_list.head)
+                       break;
+               read_msg_fd();
+       }
 
-       io_flush();
+       return flist_ndx_pop(&redo_list);
+}
+
+int get_hlink_num(void)
+{
+       return flist_ndx_pop(&hlink_list);
+}
+
+/**
+ * When we're the receiver and we have a local --files-from list of names
+ * that needs to be sent over the socket to the sender, we have to do two
+ * things at the same time: send the sender a list of what files we're
+ * processing and read the incoming file+info list from the sender.  We do
+ * this by augmenting the read_timeout() function to copy this data.  It
+ * uses the io_filesfrom_buf to read a block of data from f_in (when it is
+ * ready, since it might be a pipe) and then blast it out f_out (when it
+ * is ready to receive more data).
+ */
+void io_set_filesfrom_fds(int f_in, int f_out)
+{
+       io_filesfrom_f_in = f_in;
+       io_filesfrom_f_out = f_out;
+       io_filesfrom_bp = io_filesfrom_buf;
+       io_filesfrom_lastchar = '\0';
+       io_filesfrom_buflen = 0;
+}
+
+/* It's almost always an error to get an EOF when we're trying to read from the
+ * network, because the protocol is (for the most part) self-terminating.
+ *
+ * There is one case for the receiver when it is at the end of the transfer
+ * (hanging around reading any keep-alive packets that might come its way): if
+ * the sender dies before the generator's kill-signal comes through, we can end
+ * up here needing to loop until the kill-signal arrives.  In this situation,
+ * kluge_around_eof will be < 0.
+ *
+ * There is another case for older protocol versions (< 24) where the module
+ * listing was not terminated, so we must ignore an EOF error in that case and
+ * exit.  In this situation, kluge_around_eof will be > 0. */
+static void whine_about_eof(int fd)
+{
+       if (kluge_around_eof && fd == sock_f_in) {
+               int i;
+               if (kluge_around_eof > 0)
+                       exit_cleanup(0);
+               /* If we're still here after 10 seconds, exit with an error. */
+               for (i = 10*1000/20; i--; )
+                       msleep(20);
+       }
+
+       rprintf(FERROR, RSYNC_NAME ": connection unexpectedly closed "
+               "(%.0f bytes received so far) [%s]\n",
+               (double)stats.total_read, who_am_i());
+
+       exit_cleanup(RERR_STREAMIO);
+}
+
+
+/**
+ * Read from a socket with I/O timeout. return the number of bytes
+ * read. If no bytes can be read then exit, never return a number <= 0.
+ *
+ * TODO: If the remote shell connection fails, then current versions
+ * actually report an "unexpected EOF" error here.  Since it's a
+ * fairly common mistake to try to use rsh when ssh is required, we
+ * should trap that: if we fail to read any data at all, we should
+ * give a better explanation.  We can tell whether the connection has
+ * started by looking e.g. at whether the remote version is known yet.
+ */
+static int read_timeout(int fd, char *buf, size_t len)
+{
+       int n, ret = 0;
+
+       io_flush(NORMAL_FLUSH);
 
        while (ret == 0) {
-               fd_set fds;
+               /* until we manage to read *something* */
+               fd_set r_fds, w_fds;
                struct timeval tv;
-               int fd_count = fd+1;
+               int maxfd = fd;
+               int count;
 
-               FD_ZERO(&fds);
-               FD_SET(fd, &fds);
-               if (io_error_fd != -1) {
-                       FD_SET(io_error_fd, &fds);
-                       if (io_error_fd > fd) fd_count = io_error_fd+1;
+               FD_ZERO(&r_fds);
+               FD_ZERO(&w_fds);
+               FD_SET(fd, &r_fds);
+               if (msg_list_head) {
+                       FD_SET(msg_fd_out, &w_fds);
+                       if (msg_fd_out > maxfd)
+                               maxfd = msg_fd_out;
+               }
+               if (io_filesfrom_f_out >= 0) {
+                       int new_fd;
+                       if (io_filesfrom_buflen == 0) {
+                               if (io_filesfrom_f_in >= 0) {
+                                       FD_SET(io_filesfrom_f_in, &r_fds);
+                                       new_fd = io_filesfrom_f_in;
+                               } else {
+                                       io_filesfrom_f_out = -1;
+                                       new_fd = -1;
+                               }
+                       } else {
+                               FD_SET(io_filesfrom_f_out, &w_fds);
+                               new_fd = io_filesfrom_f_out;
+                       }
+                       if (new_fd > maxfd)
+                               maxfd = new_fd;
                }
 
-               tv.tv_sec = io_timeout?io_timeout:SELECT_TIMEOUT;
+               tv.tv_sec = select_timeout;
                tv.tv_usec = 0;
 
                errno = 0;
 
-               if (select(fd_count, &fds, NULL, NULL, &tv) < 1) {
-                       if (errno == EBADF) {
+               count = select(maxfd + 1, &r_fds, &w_fds, NULL, &tv);
+
+               if (count <= 0) {
+                       if (errno == EBADF)
                                exit_cleanup(RERR_SOCKETIO);
-                       }
                        check_timeout();
                        continue;
                }
 
-               if (io_error_fd != -1 && FD_ISSET(io_error_fd, &fds)) {
-                       read_error_fd();
+               if (msg_list_head && FD_ISSET(msg_fd_out, &w_fds))
+                       msg_list_push(NORMAL_FLUSH);
+
+               if (io_filesfrom_f_out >= 0) {
+                       if (io_filesfrom_buflen) {
+                               if (FD_ISSET(io_filesfrom_f_out, &w_fds)) {
+                                       int l = write(io_filesfrom_f_out,
+                                                     io_filesfrom_bp,
+                                                     io_filesfrom_buflen);
+                                       if (l > 0) {
+                                               if (!(io_filesfrom_buflen -= l))
+                                                       io_filesfrom_bp = io_filesfrom_buf;
+                                               else
+                                                       io_filesfrom_bp += l;
+                                       } else {
+                                               /* XXX should we complain? */
+                                               io_filesfrom_f_out = -1;
+                                       }
+                               }
+                       } else if (io_filesfrom_f_in >= 0) {
+                               if (FD_ISSET(io_filesfrom_f_in, &r_fds)) {
+                                       int l = read(io_filesfrom_f_in,
+                                                    io_filesfrom_buf,
+                                                    sizeof io_filesfrom_buf);
+                                       if (l <= 0) {
+                                               /* Send end-of-file marker */
+                                               io_filesfrom_buf[0] = '\0';
+                                               io_filesfrom_buf[1] = '\0';
+                                               io_filesfrom_buflen = io_filesfrom_lastchar? 2 : 1;
+                                               io_filesfrom_f_in = -1;
+                                       } else {
+                                               if (!eol_nulls) {
+                                                       char *s = io_filesfrom_buf + l;
+                                                       /* Transform CR and/or LF into '\0' */
+                                                       while (s-- > io_filesfrom_buf) {
+                                                               if (*s == '\n' || *s == '\r')
+                                                                       *s = '\0';
+                                                       }
+                                               }
+                                               if (!io_filesfrom_lastchar) {
+                                                       /* Last buf ended with a '\0', so don't
+                                                        * let this buf start with one. */
+                                                       while (l && !*io_filesfrom_bp)
+                                                               io_filesfrom_bp++, l--;
+                                               }
+                                               if (!l)
+                                                       io_filesfrom_bp = io_filesfrom_buf;
+                                               else {
+                                                       char *f = io_filesfrom_bp;
+                                                       char *t = f;
+                                                       char *eob = f + l;
+                                                       /* Eliminate any multi-'\0' runs. */
+                                                       while (f != eob) {
+                                                               if (!(*t++ = *f++)) {
+                                                                       while (f != eob && !*f)
+                                                                               f++, l--;
+                                                               }
+                                                       }
+                                                       io_filesfrom_lastchar = f[-1];
+                                               }
+                                               io_filesfrom_buflen = l;
+                                       }
+                               }
+                       }
                }
 
-               if (!FD_ISSET(fd, &fds)) continue;
+               if (!FD_ISSET(fd, &r_fds))
+                       continue;
 
                n = read(fd, buf, len);
 
-               if (n > 0) {
-                       buf += n;
-                       len -= n;
-                       ret += n;
-                       if (io_timeout)
-                               last_io = time(NULL);
-                       continue;
-               }
+               if (n <= 0) {
+                       if (n == 0)
+                               whine_about_eof(fd); /* Doesn't return. */
+                       if (errno == EINTR || errno == EWOULDBLOCK
+                           || errno == EAGAIN)
+                               continue;
 
-               if (n == -1 && errno == EINTR) {
-                       continue;
+                       /* Don't write errors on a dead socket. */
+                       if (fd == sock_f_in)
+                               close_multiplexing_out();
+                       rsyserr(FERROR, errno, "read error");
+                       exit_cleanup(RERR_STREAMIO);
                }
 
-               if (n == -1 && 
-                   (errno == EWOULDBLOCK || errno == EAGAIN)) {
+               buf += n;
+               len -= n;
+               ret += n;
+
+               if (fd == sock_f_in && (io_timeout || am_generator))
+                       last_io = time(NULL);
+       }
+
+       return ret;
+}
+
+/**
+ * Read a line into the "fname" buffer (which must be at least MAXPATHLEN
+ * characters long).
+ */
+int read_filesfrom_line(int fd, char *fname)
+{
+       char ch, *s, *eob = fname + MAXPATHLEN - 1;
+       int cnt;
+       int reading_remotely = filesfrom_host != NULL;
+       int nulls = eol_nulls || reading_remotely;
+
+  start:
+       s = fname;
+       while (1) {
+               cnt = read(fd, &ch, 1);
+               if (cnt < 0 && (errno == EWOULDBLOCK
+                 || errno == EINTR || errno == EAGAIN)) {
+                       struct timeval tv;
+                       fd_set fds;
+                       FD_ZERO(&fds);
+                       FD_SET(fd, &fds);
+                       tv.tv_sec = select_timeout;
+                       tv.tv_usec = 0;
+                       if (!select(fd+1, &fds, NULL, NULL, &tv))
+                               check_timeout();
                        continue;
                }
+               if (cnt != 1)
+                       break;
+               if (nulls? !ch : (ch == '\r' || ch == '\n')) {
+                       /* Skip empty lines if reading locally. */
+                       if (!reading_remotely && s == fname)
+                               continue;
+                       break;
+               }
+               if (s < eob)
+                       *s++ = ch;
+       }
+       *s = '\0';
 
+       /* Dump comments. */
+       if (*fname == '#' || *fname == ';')
+               goto start;
 
-               if (n == 0) {
-                       if (eof_error) {
-                               rprintf(FERROR,"unexpected EOF in read_timeout\n");
-                       }
-                       exit_cleanup(RERR_STREAMIO);
-               }
+       return s - fname;
+}
 
-               /* this prevents us trying to write errors on a dead socket */
-               io_multiplexing_close();
 
-               rprintf(FERROR,"read error: %s\n", strerror(errno));
-               exit_cleanup(RERR_STREAMIO);
+static char *iobuf_out;
+static int iobuf_out_cnt;
+
+void io_start_buffering_out(void)
+{
+       if (iobuf_out)
+               return;
+       if (!(iobuf_out = new_array(char, IO_BUFFER_SIZE)))
+               out_of_memory("io_start_buffering_out");
+       iobuf_out_cnt = 0;
+}
+
+
+static char *iobuf_in;
+static size_t iobuf_in_siz;
+
+void io_start_buffering_in(void)
+{
+       if (iobuf_in)
+               return;
+       iobuf_in_siz = 2 * IO_BUFFER_SIZE;
+       if (!(iobuf_in = new_array(char, iobuf_in_siz)))
+               out_of_memory("io_start_buffering_in");
+}
+
+
+void io_end_buffering(void)
+{
+       io_flush(NORMAL_FLUSH);
+       if (!io_multiplexing_out) {
+               free(iobuf_out);
+               iobuf_out = NULL;
        }
+}
 
-       return ret;
+
+void maybe_flush_socket(void)
+{
+       if (iobuf_out && iobuf_out_cnt && time(NULL) - last_io >= 5)
+               io_flush(NORMAL_FLUSH);
 }
 
-/* continue trying to read len bytes - don't return until len
-   has been read */
-static void read_loop(int fd, char *buf, int len)
+
+void maybe_send_keepalive(void)
+{
+       if (time(NULL) - last_io >= allowed_lull) {
+               if (!iobuf_out || !iobuf_out_cnt) {
+                       if (protocol_version < 29)
+                               return; /* there's nothing we can do */
+                       write_int(sock_f_out, the_file_list->count);
+                       write_shortint(sock_f_out, ITEM_IS_NEW);
+               }
+               if (iobuf_out)
+                       io_flush(NORMAL_FLUSH);
+       }
+}
+
+
+/**
+ * Continue trying to read len bytes - don't return until len has been
+ * read.
+ **/
+static void read_loop(int fd, char *buf, size_t len)
 {
        while (len) {
                int n = read_timeout(fd, buf, len);
@@ -193,74 +701,141 @@ static void read_loop(int fd, char *buf, int len)
        }
 }
 
-/* read from the file descriptor handling multiplexing - 
-   return number of bytes read
-   never return <= 0 */
-static int read_unbuffered(int fd, char *buf, int len)
+
+/**
+ * Read from the file descriptor handling multiplexing - return number
+ * of bytes read.
+ *
+ * Never returns <= 0.
+ */
+static int readfd_unbuffered(int fd, char *buf, size_t len)
 {
-       static int remaining;
-       int tag, ret=0;
-       char line[1024];
+       static size_t remaining;
+       static size_t iobuf_in_ndx;
+       int tag, ret = 0;
+#if MAXPATHLEN < 4096
+       char line[4096+1024];
+#else
+       char line[MAXPATHLEN+1024];
+#endif
 
-       if (!io_multiplexing_in || fd != multiplex_in_fd) 
+       if (!iobuf_in || fd != sock_f_in)
                return read_timeout(fd, buf, len);
 
+       if (!io_multiplexing_in && remaining == 0) {
+               remaining = read_timeout(fd, iobuf_in, iobuf_in_siz);
+               iobuf_in_ndx = 0;
+       }
+
        while (ret == 0) {
                if (remaining) {
                        len = MIN(len, remaining);
-                       read_loop(fd, buf, len);
+                       memcpy(buf, iobuf_in + iobuf_in_ndx, len);
+                       iobuf_in_ndx += len;
                        remaining -= len;
                        ret = len;
-                       continue;
+                       break;
                }
 
                read_loop(fd, line, 4);
                tag = IVAL(line, 0);
 
                remaining = tag & 0xFFFFFF;
-               tag = tag >> 24;
-
-               if (tag == MPLEX_BASE) continue;
-
-               tag -= MPLEX_BASE;
-
-               if (tag != FERROR && tag != FINFO) {
-                       rprintf(FERROR,"unexpected tag %d\n", tag);
-                       exit_cleanup(RERR_STREAMIO);
-               }
-
-               if (remaining > sizeof(line)-1) {
-                       rprintf(FERROR,"multiplexing overflow %d\n\n", 
-                               remaining);
+               tag = (tag >> 24) - MPLEX_BASE;
+
+               switch (tag) {
+               case MSG_DATA:
+                       if (remaining > iobuf_in_siz) {
+                               if (!(iobuf_in = realloc_array(iobuf_in, char,
+                                                              remaining)))
+                                       out_of_memory("readfd_unbuffered");
+                               iobuf_in_siz = remaining;
+                       }
+                       read_loop(fd, iobuf_in, remaining);
+                       iobuf_in_ndx = 0;
+                       break;
+               case MSG_DELETED:
+                       if (remaining >= sizeof line) {
+                               rprintf(FERROR, "invalid multi-message %d:%ld\n",
+                                       tag, (long)remaining);
+                               exit_cleanup(RERR_STREAMIO);
+                       }
+                       read_loop(fd, line, remaining);
+                       line[remaining] = '\0';
+                       /* A directory name was sent with the trailing null */
+                       if (remaining > 0 && !line[remaining-1])
+                               log_delete(line, S_IFDIR);
+                       else
+                               log_delete(line, S_IFREG);
+                       remaining = 0;
+                       break;
+               case MSG_SUCCESS:
+                       if (remaining != 4) {
+                               rprintf(FERROR, "invalid multi-message %d:%ld [%s]\n",
+                                       tag, (long)remaining, who_am_i());
+                               exit_cleanup(RERR_STREAMIO);
+                       }
+                       read_loop(fd, line, remaining);
+                       successful_send(IVAL(line, 0));
+                       remaining = 0;
+                       break;
+               case MSG_INFO:
+               case MSG_ERROR:
+                       if (remaining >= sizeof line) {
+                               rprintf(FERROR,
+                                       "multiplexing overflow %d:%ld [%s]\n",
+                                       tag, (long)remaining, who_am_i());
+                               exit_cleanup(RERR_STREAMIO);
+                       }
+                       read_loop(fd, line, remaining);
+                       rwrite((enum logcode)tag, line, remaining);
+                       remaining = 0;
+                       break;
+               default:
+                       rprintf(FERROR, "unexpected tag %d [%s]\n",
+                               tag, who_am_i());
                        exit_cleanup(RERR_STREAMIO);
                }
-
-               read_loop(fd, line, remaining);
-               line[remaining] = 0;
-
-               rprintf((enum logcode)tag,"%s", line);
-               remaining = 0;
        }
 
+       if (remaining == 0)
+               io_flush(NORMAL_FLUSH);
+
        return ret;
 }
 
 
-/* do a buffered read from fd. don't return until all N bytes
-   have been read. If all N can't be read then exit with an error */
-static void readfd(int fd,char *buffer,int N)
+
+/**
+ * Do a buffered read from @p fd.  Don't return until all @p n bytes
+ * have been read.  If all @p n can't be read then exit with an
+ * error.
+ **/
+static void readfd(int fd, char *buffer, size_t N)
 {
        int  ret;
-       int total=0;  
-       
-       while (total < N) {
-               io_flush();
+       size_t total = 0;
 
-               ret = read_unbuffered(fd,buffer + total,N-total);
+       while (total < N) {
+               ret = readfd_unbuffered(fd, buffer + total, N-total);
                total += ret;
        }
 
-       stats.total_read += total;
+       if (fd == write_batch_monitor_in) {
+               if ((size_t)write(batch_fd, buffer, total) != total)
+                       exit_cleanup(RERR_FILEIO);
+       }
+
+       if (fd == sock_f_in)
+               stats.total_read += total;
+}
+
+
+int read_shortint(int f)
+{
+       uchar b[2];
+       readfd(f, (char *)b, 2);
+       return (b[1] << 8) + b[0];
 }
 
 
@@ -271,133 +846,259 @@ int32 read_int(int f)
 
        readfd(f,b,4);
        ret = IVAL(b,0);
-       if (ret == (int32)0xffffffff) return -1;
+       if (ret == (int32)0xffffffff)
+               return -1;
        return ret;
 }
 
 int64 read_longint(int f)
 {
-       extern int remote_version;
        int64 ret;
        char b[8];
        ret = read_int(f);
 
-       if ((int32)ret != (int32)0xffffffff) {
+       if ((int32)ret != (int32)0xffffffff)
                return ret;
-       }
 
-#ifdef NO_INT64
-       rprintf(FERROR,"Integer overflow - attempted 64 bit offset\n");
+#if SIZEOF_INT64 < 8
+       rprintf(FERROR, "Integer overflow: attempted 64-bit offset\n");
        exit_cleanup(RERR_UNSUPPORTED);
 #else
-       if (remote_version >= 16) {
-               readfd(f,b,8);
-               ret = IVAL(b,0) | (((int64)IVAL(b,4))<<32);
-       }
+       readfd(f,b,8);
+       ret = IVAL(b,0) | (((int64)IVAL(b,4))<<32);
 #endif
 
        return ret;
 }
 
-void read_buf(int f,char *buf,int len)
+void read_buf(int f,char *buf,size_t len)
 {
        readfd(f,buf,len);
 }
 
-void read_sbuf(int f,char *buf,int len)
+void read_sbuf(int f,char *buf,size_t len)
 {
-       read_buf(f,buf,len);
-       buf[len] = 0;
+       readfd(f, buf, len);
+       buf[len] = '\0';
 }
 
-unsigned char read_byte(int f)
+uchar read_byte(int f)
 {
-       unsigned char c;
-       read_buf(f,(char *)&c,1);
+       uchar c;
+       readfd(f, (char *)&c, 1);
        return c;
 }
 
-/* write len bytes to fd */
-static void writefd_unbuffered(int fd,char *buf,int len)
+int read_vstring(int f, char *buf, int bufsize)
+{
+       int len = read_byte(f);
+
+       if (len & 0x80)
+               len = (len & ~0x80) * 0x100 + read_byte(f);
+
+       if (len >= bufsize) {
+               rprintf(FERROR, "over-long vstring received (%d > %d)\n",
+                       len, bufsize - 1);
+               return -1;
+       }
+
+       if (len)
+               readfd(f, buf, len);
+       buf[len] = '\0';
+       return len;
+}
+
+/* Populate a sum_struct with values from the socket.  This is
+ * called by both the sender and the receiver. */
+void read_sum_head(int f, struct sum_struct *sum)
+{
+       sum->count = read_int(f);
+       sum->blength = read_int(f);
+       if (sum->blength < 0 || sum->blength > MAX_BLOCK_SIZE) {
+               rprintf(FERROR, "Invalid block length %ld [%s]\n",
+                       (long)sum->blength, who_am_i());
+               exit_cleanup(RERR_PROTOCOL);
+       }
+       sum->s2length = protocol_version < 27 ? csum_length : (int)read_int(f);
+       if (sum->s2length < 0 || sum->s2length > MD4_SUM_LENGTH) {
+               rprintf(FERROR, "Invalid checksum length %d [%s]\n",
+                       sum->s2length, who_am_i());
+               exit_cleanup(RERR_PROTOCOL);
+       }
+       sum->remainder = read_int(f);
+       if (sum->remainder < 0 || sum->remainder > sum->blength) {
+               rprintf(FERROR, "Invalid remainder length %ld [%s]\n",
+                       (long)sum->remainder, who_am_i());
+               exit_cleanup(RERR_PROTOCOL);
+       }
+}
+
+/* Send the values from a sum_struct over the socket.  Set sum to
+ * NULL if there are no checksums to send.  This is called by both
+ * the generator and the sender. */
+void write_sum_head(int f, struct sum_struct *sum)
+{
+       static struct sum_struct null_sum;
+
+       if (sum == NULL)
+               sum = &null_sum;
+
+       write_int(f, sum->count);
+       write_int(f, sum->blength);
+       if (protocol_version >= 27)
+               write_int(f, sum->s2length);
+       write_int(f, sum->remainder);
+}
+
+
+/**
+ * Sleep after writing to limit I/O bandwidth usage.
+ *
+ * @todo Rather than sleeping after each write, it might be better to
+ * use some kind of averaging.  The current algorithm seems to always
+ * use a bit less bandwidth than specified, because it doesn't make up
+ * for slow periods.  But arguably this is a feature.  In addition, we
+ * ought to take the time used to write the data into account.
+ *
+ * During some phases of big transfers (file FOO is uptodate) this is
+ * called with a small bytes_written every time.  As the kernel has to
+ * round small waits up to guarantee that we actually wait at least the
+ * requested number of microseconds, this can become grossly inaccurate.
+ * We therefore keep track of the bytes we've written over time and only
+ * sleep when the accumulated delay is at least 1 tenth of a second.
+ **/
+static void sleep_for_bwlimit(int bytes_written)
+{
+       static struct timeval prior_tv;
+       static long total_written = 0; 
+       struct timeval tv, start_tv;
+       long elapsed_usec, sleep_usec;
+
+#define ONE_SEC        1000000L /* # of microseconds in a second */
+
+       if (!bwlimit)
+               return;
+
+       total_written += bytes_written; 
+
+       gettimeofday(&start_tv, NULL);
+       if (prior_tv.tv_sec) {
+               elapsed_usec = (start_tv.tv_sec - prior_tv.tv_sec) * ONE_SEC
+                            + (start_tv.tv_usec - prior_tv.tv_usec);
+               total_written -= elapsed_usec * bwlimit / (ONE_SEC/1024);
+               if (total_written < 0)
+                       total_written = 0;
+       }
+
+       sleep_usec = total_written * (ONE_SEC/1024) / bwlimit;
+       if (sleep_usec < ONE_SEC / 10) {
+               prior_tv = start_tv;
+               return;
+       }
+
+       tv.tv_sec  = sleep_usec / ONE_SEC;
+       tv.tv_usec = sleep_usec % ONE_SEC;
+       select(0, NULL, NULL, NULL, &tv);
+
+       gettimeofday(&prior_tv, NULL);
+       elapsed_usec = (prior_tv.tv_sec - start_tv.tv_sec) * ONE_SEC
+                    + (prior_tv.tv_usec - start_tv.tv_usec);
+       total_written = (sleep_usec - elapsed_usec) * bwlimit / (ONE_SEC/1024);
+}
+
+
+/* Write len bytes to the file descriptor fd, looping as necessary to get
+ * the job done and also (in certain circumstnces) reading any data on
+ * msg_fd_in to avoid deadlock.
+ *
+ * This function underlies the multiplexing system.  The body of the
+ * application never calls this function directly. */
+static void writefd_unbuffered(int fd,char *buf,size_t len)
 {
-       int total = 0;
+       size_t n, total = 0;
        fd_set w_fds, r_fds;
-       int fd_count, count;
+       int maxfd, count, ret, using_r_fds;
        struct timeval tv;
 
        no_flush++;
 
        while (total < len) {
                FD_ZERO(&w_fds);
-               FD_ZERO(&r_fds);
                FD_SET(fd,&w_fds);
-               fd_count = fd;
-
-               if (io_error_fd != -1) {
-                       FD_SET(io_error_fd,&r_fds);
-                       if (io_error_fd > fd_count) 
-                               fd_count = io_error_fd;
+               maxfd = fd;
+
+               if (msg_fd_in >= 0 && len-total >= contiguous_write_len) {
+                       FD_ZERO(&r_fds);
+                       FD_SET(msg_fd_in,&r_fds);
+                       if (msg_fd_in > maxfd)
+                               maxfd = msg_fd_in;
+                       using_r_fds = 1;
+               } else
+                       using_r_fds = 0;
+               if (fd != sock_f_out && iobuf_out_cnt && no_flush == 1) {
+                       FD_SET(sock_f_out, &w_fds);
+                       if (sock_f_out > maxfd)
+                               maxfd = sock_f_out;
                }
 
-               tv.tv_sec = io_timeout?io_timeout:SELECT_TIMEOUT;
+               tv.tv_sec = select_timeout;
                tv.tv_usec = 0;
 
                errno = 0;
-
-               count = select(fd_count+1,
-                              io_error_fd != -1?&r_fds:NULL,
-                              &w_fds,NULL,
-                              &tv);
+               count = select(maxfd + 1, using_r_fds ? &r_fds : NULL,
+                              &w_fds, NULL, &tv);
 
                if (count <= 0) {
-                       if (errno == EBADF) {
+                       if (count < 0 && errno == EBADF)
                                exit_cleanup(RERR_SOCKETIO);
-                       }
                        check_timeout();
                        continue;
                }
 
-               if (io_error_fd != -1 && FD_ISSET(io_error_fd, &r_fds)) {
-                       read_error_fd();
-               }
+               if (using_r_fds && FD_ISSET(msg_fd_in, &r_fds))
+                       read_msg_fd();
 
-               if (FD_ISSET(fd, &w_fds)) {
-                       int ret, n = len-total;
-                       
-                       ret = write(fd,buf+total,n);
+               if (!FD_ISSET(fd, &w_fds))
+                       continue;
 
-                       if (ret == -1 && errno == EINTR) {
-                               continue;
+               n = len - total;
+               if (bwlimit && n > bwlimit_writemax)
+                       n = bwlimit_writemax;
+               ret = write(fd, buf + total, n);
+
+               if (ret <= 0) {
+                       if (ret < 0) {
+                               if (errno == EINTR)
+                                       continue;
+                               if (errno == EWOULDBLOCK || errno == EAGAIN) {
+                                       msleep(1);
+                                       continue;
+                               }
                        }
 
-                       if (ret == -1 && 
-                           (errno == EWOULDBLOCK || errno == EAGAIN)) {
-                               msleep(1);
-                               continue;
+                       /* Don't try to write errors back across the stream. */
+                       if (fd == sock_f_out)
+                               close_multiplexing_out();
+                       rsyserr(FERROR, errno,
+                               "writefd_unbuffered failed to write %ld bytes: phase \"%s\" [%s]",
+                               (long)len, io_write_phase, who_am_i());
+                       /* If the other side is sending us error messages, try
+                        * to grab any messages they sent before they died. */
+                       while (fd == sock_f_out && io_multiplexing_in) {
+                               io_timeout = select_timeout = 30;
+                               ignore_timeout = 0;
+                               readfd_unbuffered(sock_f_in, io_filesfrom_buf,
+                                                 sizeof io_filesfrom_buf);
                        }
+                       exit_cleanup(RERR_STREAMIO);
+               }
 
-                       if (ret <= 0) {
-                               rprintf(FERROR,"erroring writing %d bytes - exiting\n", len);
-                               exit_cleanup(RERR_STREAMIO);
-                       }
+               total += ret;
 
-                       /* Sleep after writing to limit I/O bandwidth */
-                       if (bwlimit)
-                       {
-                           tv.tv_sec = 0;
-                           tv.tv_usec = ret * 1000 / bwlimit;
-                           while (tv.tv_usec > 1000000)
-                           {
-                               tv.tv_sec++;
-                               tv.tv_usec -= 1000000;
-                           }
-                           select(0, NULL, NULL, NULL, &tv);
-                       }
-                       total += ret;
-
-                       if (io_timeout)
+               if (fd == sock_f_out) {
+                       if (io_timeout || am_generator)
                                last_io = time(NULL);
+                       sleep_for_bwlimit(ret);
                }
        }
 
@@ -405,99 +1106,100 @@ static void writefd_unbuffered(int fd,char *buf,int len)
 }
 
 
-static char *io_buffer;
-static int io_buffer_count;
-
-void io_start_buffering(int fd)
-{
-       if (io_buffer) return;
-       multiplex_out_fd = fd;
-       io_buffer = (char *)malloc(IO_BUFFER_SIZE);
-       if (!io_buffer) out_of_memory("writefd");
-       io_buffer_count = 0;
-}
-
-/* write an message to a multiplexed stream. If this fails then rsync
-   exits */
-static void mplex_write(int fd, enum logcode code, char *buf, int len)
+/**
+ * Write an message to a multiplexed stream. If this fails then rsync
+ * exits.
+ **/
+static void mplex_write(enum msgcode code, char *buf, size_t len)
 {
        char buffer[4096];
-       int n = len;
+       size_t n = len;
 
        SIVAL(buffer, 0, ((MPLEX_BASE + (int)code)<<24) + len);
 
-       if (n > (sizeof(buffer)-4)) {
-               n = sizeof(buffer)-4;
-       }
+       /* When the generator reads messages from the msg_fd_in pipe, it can
+        * cause output to occur down the socket.  Setting contiguous_write_len
+        * prevents the reading of msg_fd_in once we actually start to write
+        * this sequence of data (though we might read it before the start). */
+       if (am_generator && msg_fd_in >= 0)
+               contiguous_write_len = len + 4;
+
+       if (n > sizeof buffer - 4)
+               n = sizeof buffer - 4;
 
        memcpy(&buffer[4], buf, n);
-       writefd_unbuffered(fd, buffer, n+4);
+       writefd_unbuffered(sock_f_out, buffer, n+4);
 
        len -= n;
        buf += n;
 
-       if (len) {
-               writefd_unbuffered(fd, buf, len);
-       }
+       if (len)
+               writefd_unbuffered(sock_f_out, buf, len);
+
+       if (am_generator && msg_fd_in >= 0)
+               contiguous_write_len = 0;
 }
 
 
-void io_flush(void)
+void io_flush(int flush_it_all)
 {
-       int fd = multiplex_out_fd;
-       if (!io_buffer_count || no_flush) return;
+       msg_list_push(flush_it_all);
 
-       if (io_multiplexing_out) {
-               mplex_write(fd, FNONE, io_buffer, io_buffer_count);
-       } else {
-               writefd_unbuffered(fd, io_buffer, io_buffer_count);
-       }
-       io_buffer_count = 0;
-}
+       if (!iobuf_out_cnt || no_flush)
+               return;
 
-void io_end_buffering(int fd)
-{
-       io_flush();
-       if (!io_multiplexing_out) {
-               free(io_buffer);
-               io_buffer = NULL;
-       }
+       if (io_multiplexing_out)
+               mplex_write(MSG_DATA, iobuf_out, iobuf_out_cnt);
+       else
+               writefd_unbuffered(sock_f_out, iobuf_out, iobuf_out_cnt);
+       iobuf_out_cnt = 0;
 }
 
-/* some OSes have a bug where an exit causes the pending writes on
-   a socket to be flushed. Do an explicit shutdown to try to prevent this */
-void io_shutdown(void)
+
+static void writefd(int fd,char *buf,size_t len)
 {
-       if (multiplex_out_fd != -1) close(multiplex_out_fd);
-       if (io_error_fd != -1) close(io_error_fd);
-       multiplex_out_fd = -1;
-       io_error_fd = -1;
-}
+       if (fd == msg_fd_out) {
+               rprintf(FERROR, "Internal error: wrong write used in receiver.\n");
+               exit_cleanup(RERR_PROTOCOL);
+       }
 
+       if (fd == sock_f_out)
+               stats.total_written += len;
 
-static void writefd(int fd,char *buf,int len)
-{
-       stats.total_written += len;
+       if (fd == write_batch_monitor_out) {
+               if ((size_t)write(batch_fd, buf, len) != len)
+                       exit_cleanup(RERR_FILEIO);
+       }
 
-       if (!io_buffer || fd != multiplex_out_fd) {
+       if (!iobuf_out || fd != sock_f_out) {
                writefd_unbuffered(fd, buf, len);
                return;
        }
 
        while (len) {
-               int n = MIN(len, IO_BUFFER_SIZE-io_buffer_count);
+               int n = MIN((int)len, IO_BUFFER_SIZE - iobuf_out_cnt);
                if (n > 0) {
-                       memcpy(io_buffer+io_buffer_count, buf, n);
+                       memcpy(iobuf_out+iobuf_out_cnt, buf, n);
                        buf += n;
                        len -= n;
-                       io_buffer_count += n;
+                       iobuf_out_cnt += n;
                }
-               
-               if (io_buffer_count == IO_BUFFER_SIZE) io_flush();
+
+               if (iobuf_out_cnt == IO_BUFFER_SIZE)
+                       io_flush(NORMAL_FLUSH);
        }
 }
 
 
+void write_shortint(int f, int x)
+{
+       uchar b[2];
+       b[0] = x;
+       b[1] = x >> 8;
+       writefd(f, (char *)b, 2);
+}
+
+
 void write_int(int f,int32 x)
 {
        char b[4];
@@ -505,123 +1207,179 @@ void write_int(int f,int32 x)
        writefd(f,b,4);
 }
 
+
+void write_int_named(int f, int32 x, const char *phase)
+{
+       io_write_phase = phase;
+       write_int(f, x);
+       io_write_phase = phase_unknown;
+}
+
+
+/*
+ * Note: int64 may actually be a 32-bit type if ./configure couldn't find any
+ * 64-bit types on this platform.
+ */
 void write_longint(int f, int64 x)
 {
-       extern int remote_version;
        char b[8];
 
-       if (remote_version < 16 || x <= 0x7FFFFFFF) {
+       if (x <= 0x7FFFFFFF) {
                write_int(f, (int)x);
                return;
        }
 
+#if SIZEOF_INT64 < 8
+       rprintf(FERROR, "Integer overflow: attempted 64-bit offset\n");
+       exit_cleanup(RERR_UNSUPPORTED);
+#else
        write_int(f, (int32)0xFFFFFFFF);
        SIVAL(b,0,(x&0xFFFFFFFF));
        SIVAL(b,4,((x>>32)&0xFFFFFFFF));
 
        writefd(f,b,8);
+#endif
 }
 
-void write_buf(int f,char *buf,int len)
+void write_buf(int f,char *buf,size_t len)
 {
        writefd(f,buf,len);
 }
 
-/* write a string to the connection */
-static void write_sbuf(int f,char *buf)
+/** Write a string to the connection */
+void write_sbuf(int f, char *buf)
 {
-       write_buf(f, buf, strlen(buf));
+       writefd(f, buf, strlen(buf));
 }
 
-
-void write_byte(int f,unsigned char c)
+void write_byte(int f, uchar c)
 {
-       write_buf(f,(char *)&c,1);
+       writefd(f, (char *)&c, 1);
 }
 
-int read_line(int f, char *buf, int maxlen)
+void write_vstring(int f, char *str, int len)
 {
-       eof_error = 0;
+       uchar lenbuf[3], *lb = lenbuf;
+
+       if (len > 0x7F) {
+               if (len > 0x7FFF) {
+                       rprintf(FERROR,
+                               "attempting to send over-long vstring (%d > %d)\n",
+                               len, 0x7FFF);
+                       exit_cleanup(RERR_PROTOCOL);
+               }
+               *lb++ = len / 0x100 + 0x80;
+       }
+       *lb = len;
+
+       writefd(f, (char*)lenbuf, lb - lenbuf + 1);
+       if (len)
+               writefd(f, str, len);
+}
 
+
+/**
+ * Read a line of up to @p maxlen characters into @p buf (not counting
+ * the trailing null).  Strips the (required) trailing newline and all
+ * carriage returns.
+ *
+ * @return 1 for success; 0 for I/O error or truncation.
+ **/
+int read_line(int f, char *buf, size_t maxlen)
+{
        while (maxlen) {
                buf[0] = 0;
                read_buf(f, buf, 1);
-               if (buf[0] == 0) return 0;
-               if (buf[0] == '\n') {
-                       buf[0] = 0;
+               if (buf[0] == 0)
+                       return 0;
+               if (buf[0] == '\n')
                        break;
-               }
                if (buf[0] != '\r') {
                        buf++;
                        maxlen--;
                }
        }
-       if (maxlen == 0) {
-               *buf = 0;
-               return 0;
-       }
-
-       eof_error = 1;
-
-       return 1;
+       *buf = '\0';
+       return maxlen > 0;
 }
 
 
 void io_printf(int fd, const char *format, ...)
 {
-       va_list ap;  
+       va_list ap;
        char buf[1024];
        int len;
-       
+
        va_start(ap, format);
-       len = vslprintf(buf, sizeof(buf), format, ap);
+       len = vsnprintf(buf, sizeof buf, format, ap);
        va_end(ap);
 
-       if (len < 0) exit_cleanup(RERR_STREAMIO);
+       if (len < 0)
+               exit_cleanup(RERR_STREAMIO);
 
        write_sbuf(fd, buf);
 }
 
 
-/* setup for multiplexing an error stream with the data stream */
-void io_start_multiplex_out(int fd)
+/** Setup for multiplexing a MSG_* stream with the data stream. */
+void io_start_multiplex_out(void)
 {
-       multiplex_out_fd = fd;
-       io_flush();
-       io_start_buffering(fd);
+       io_flush(NORMAL_FLUSH);
+       io_start_buffering_out();
        io_multiplexing_out = 1;
 }
 
-/* setup for multiplexing an error stream with the data stream */
-void io_start_multiplex_in(int fd)
+/** Setup for multiplexing a MSG_* stream with the data stream. */
+void io_start_multiplex_in(void)
 {
-       multiplex_in_fd = fd;
-       io_flush();
+       io_flush(NORMAL_FLUSH);
+       io_start_buffering_in();
        io_multiplexing_in = 1;
 }
 
-/* write an message to the multiplexed error stream */
-int io_multiplex_write(enum logcode code, char *buf, int len)
+/** Write an message to the multiplexed data stream. */
+int io_multiplex_write(enum msgcode code, char *buf, size_t len)
 {
-       if (!io_multiplexing_out) return 0;
+       if (!io_multiplexing_out)
+               return 0;
 
-       io_flush();
+       io_flush(NORMAL_FLUSH);
        stats.total_written += (len+4);
-       mplex_write(multiplex_out_fd, code, buf, len);
+       mplex_write(code, buf, len);
        return 1;
 }
 
-/* write a message to the special error fd */
-int io_error_write(int f, enum logcode code, char *buf, int len)
+void close_multiplexing_in(void)
 {
-       if (f == -1) return 0;
-       mplex_write(f, code, buf, len);
-       return 1;
+       io_multiplexing_in = 0;
 }
 
-/* stop output multiplexing */
-void io_multiplexing_close(void)
+/** Stop output multiplexing. */
+void close_multiplexing_out(void)
 {
        io_multiplexing_out = 0;
 }
 
+void start_write_batch(int fd)
+{
+       write_stream_flags(batch_fd);
+
+       /* Some communication has already taken place, but we don't
+        * enable batch writing until here so that we can write a
+        * canonical record of the communication even though the
+        * actual communication so far depends on whether a daemon
+        * is involved. */
+       write_int(batch_fd, protocol_version);
+       write_int(batch_fd, checksum_seed);
+
+       if (am_sender)
+               write_batch_monitor_out = fd;
+       else
+               write_batch_monitor_in = fd;
+}
+
+void stop_write_batch(void)
+{
+       write_batch_monitor_out = -1;
+       write_batch_monitor_in = -1;
+}