Libguestfs - Jan 2011 - [PATCH 0/8 v2 DISCUSSION ONLY] Connecting to live virtual machines

NOTE: This is HIGHLY EXPERIMENTAL and can eat babies ...

Modify the guest libvirt XML to add the virtio-serial channel:

  <channel type='unix'>
    <source mode='bind' path='/tmp/socket'/>
    <target type='virtio'
name='org.libguestfs.channel.0'/>
  </channel>

In the guest, compile guestfsd and run it like this:

  sudo daemon/guestfsd -rv

(XXX - should current directory be / ...?)

In the host connect using:

  sudo guestfish -d GuestName --live
  ><fs> run
  ><fs> ll /
   etc.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine.  Supports Linux and Windows.
http://et.redhat.com/~rjones/virt-df/

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
New in Fedora 11: Fedora Windows cross-compiler. Compile Windows
programs, test, and build Windows installers. Over 70 libraries supprt'd
http://fedoraproject.org/wiki/MinGW http://www.annexia.org/fedora_mingw
-------------- next part --------------
>From 60e78e2ddcc38f752844a280ad3e29491402503e Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Thu, 27 Jan 2011 10:12:34 +0000
Subject: [PATCH 1/8] lib: Move appliance launching to separate function.

This is just code motion.
---
 src/launch.c |   38 ++++++++++++++++++++++++--------------
 1 files changed, 24 insertions(+), 14 deletions(-)

diff --git a/src/launch.c b/src/launch.c
index a7a86b9..775f7a6 100644
--- a/src/launch.c
+++ b/src/launch.c
@@ -70,6 +70,7 @@
 #include "guestfs-internal-actions.h"
 #include "guestfs_protocol.h"
 
+static int launch_appliance (guestfs_h *g);
 static int qemu_supports (guestfs_h *g, const char *option);
 
 /* Add a string to the current command line. */
@@ -350,32 +351,19 @@ static int is_openable (guestfs_h *g, const char *path,
int flags);
 int
 guestfs__launch (guestfs_h *g)
 {
-  int r;
-  int wfd[2], rfd[2];
-  char unixsock[256];
-  struct sockaddr_un addr;
-
   /* Configured? */
-  if (!g->cmdline) {
-    error (g, _("you must call guestfs_add_drive before
guestfs_launch"));
-    return -1;
-  }
-
   if (g->state != CONFIG) {
     error (g, _("the libguestfs handle has already been launched"));
     return -1;
   }
 
-  /* Start the clock ... */
-  gettimeofday (&g->launch_t, NULL);
-
   /* Make the temporary directory. */
   if (!g->tmpdir) {
     TMP_TEMPLATE_ON_STACK (dir_template);
     g->tmpdir = safe_strdup (g, dir_template);
     if (mkdtemp (g->tmpdir) == NULL) {
       perrorf (g, _("%s: cannot create temporary directory"),
dir_template);
-      goto cleanup0;
+      return -1;
     }
   }
 
@@ -386,6 +374,28 @@ guestfs__launch (guestfs_h *g)
   if (chmod (g->tmpdir, 0755) == -1)
     fprintf (stderr, "chmod: %s: %m (ignored)\n", g->tmpdir);
 
+  return launch_appliance (g);
+}
+
+static int
+launch_appliance (guestfs_h *g)
+{
+  int r;
+  int wfd[2], rfd[2];
+  char unixsock[256];
+  struct sockaddr_un addr;
+
+  /* At present you must add drives before starting the appliance.  In
+   * future when we enable hotplugging you won't need to do this.
+   */
+  if (!g->cmdline) {
+    error (g, _("you must call guestfs_add_drive before
guestfs_launch"));
+    return -1;
+  }
+
+  /* Start the clock ... */
+  gettimeofday (&g->launch_t, NULL);
+
   /* Locate and/or build the appliance. */
   char *kernel = NULL, *initrd = NULL, *appliance = NULL;
   if (guestfs___build_appliance (g, &kernel, &initrd, &appliance)
== -1)
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
libguestfs lets you edit virtual machines.  Supports shell scripting,
bindings from many languages.  http://libguestfs.org
-------------- next part --------------
>From d5411585bf02cb84136a523bb9a06db013fc3e24 Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Thu, 27 Jan 2011 11:20:43 +0000
Subject: [PATCH 2/8] New APIs: set-attach-method, get-attach-method.

These allow you to get and set the attach method.  The format
is one of:

* appliance
* unix:<path>

It's stored broken out into an enum and a string in the handle.
---
 generator/generator_actions.ml |   30 ++++++++++++++++++++++++++
 src/guestfs-internal.h         |    6 +++++
 src/guestfs.c                  |   45 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 81 insertions(+), 0 deletions(-)

diff --git a/generator/generator_actions.ml b/generator/generator_actions.ml
index 0ae8790..a2e7467 100644
--- a/generator/generator_actions.ml
+++ b/generator/generator_actions.ml
@@ -1368,6 +1368,36 @@ part of a set.
 
 Please read L<guestfs(3)/INSPECTION> for more details.");
 
+  ("set_attach_method", (RErr, [String "attachmethod"],
[]), -1, [FishAlias "attach-method"],
+   [],
+   "set the attach method",
+   "\
+Set the method that libguestfs uses to connect to the back end
+guestfsd daemon.  Possible methods are:
+
+=over 4
+
+=item C<appliance>
+
+Launch an appliance and connect to it.  This is the ordinary method
+and the default.
+
+=item C<unix:I<path>>
+
+Connect to the Unix domain socket I<path>.
+
+This method lets you connect to an existing daemon or (using
+virtio-serial) to a live guest.
+
+=back");
+
+  ("get_attach_method", (RString "attachmethod", [], []),
-1, [],
+   [InitNone, Always, TestOutput (
+      [["get_attach_method"]], "appliance")],
+   "get the attach method",
+   "\
+Return the current attach method.  See
C<guestfs_set_attach_method>.");
+
 ]
 
 (* daemon_functions are any functions which cause some action
diff --git a/src/guestfs-internal.h b/src/guestfs-internal.h
index 194c892..0eb395b 100644
--- a/src/guestfs-internal.h
+++ b/src/guestfs-internal.h
@@ -84,6 +84,9 @@
 /* GuestFS handle and connection. */
 enum state { CONFIG, LAUNCHING, READY, BUSY, NO_HANDLE };
 
+/* Attach method. */
+enum attach_method { ATTACH_METHOD_APPLIANCE = 0, ATTACH_METHOD_UNIX };
+
 struct guestfs_h
 {
   struct guestfs_h *next;	/* Linked list of open handles. */
@@ -116,6 +119,9 @@ struct guestfs_h
   char *qemu;			/* Qemu binary. */
   char *append;			/* Append to kernel command line. */
 
+  enum attach_method attach_method;
+  char *attach_method_arg;
+
   int memsize;			/* Size of RAM (megabytes). */
 
   int selinux;                  /* selinux enabled? */
diff --git a/src/guestfs.c b/src/guestfs.c
index 79fc5bf..8b7ab4d 100644
--- a/src/guestfs.c
+++ b/src/guestfs.c
@@ -645,6 +645,51 @@ guestfs__get_network (guestfs_h *g)
   return g->enable_network;
 }
 
+int
+guestfs__set_attach_method (guestfs_h *g, const char *method)
+{
+  if (STREQ (method, "appliance")) {
+    g->attach_method = ATTACH_METHOD_APPLIANCE;
+    free (g->attach_method_arg);
+    g->attach_method_arg = NULL;
+  }
+  else if (STRPREFIX (method, "unix:") && strlen (method)
> 5) {
+    g->attach_method = ATTACH_METHOD_UNIX;
+    free (g->attach_method_arg);
+    g->attach_method_arg = safe_strdup (g, method + 5);
+    /* Note that we don't check the path exists until launch is called. */
+  }
+  else {
+    error (g, "invalid attach method: %s", method);
+    return -1;
+  }
+
+  return 0;
+}
+
+char *
+guestfs__get_attach_method (guestfs_h *g)
+{
+  char *ret;
+
+  switch (g->attach_method) {
+  case ATTACH_METHOD_APPLIANCE:
+    ret = safe_strdup (g, "appliance");
+    break;
+
+  case ATTACH_METHOD_UNIX:
+    ret = safe_malloc (g, strlen (g->attach_method_arg) + 5 + 1);
+    strcpy (ret, "unix:");
+    strcat (ret, g->attach_method_arg);
+    break;
+
+  default: /* keep GCC happy - this is not reached */
+    abort ();
+  }
+
+  return ret;
+}
+
 void
 guestfs_set_log_message_callback (guestfs_h *g,
                                   guestfs_log_message_cb cb, void *opaque)
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
virt-top is 'top' for virtual machines.  Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://et.redhat.com/~rjones/virt-top
-------------- next part --------------
>From 1f20f266b7a1d63a53dc1782adedb9bcaef97d83 Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Thu, 27 Jan 2011 15:54:33 +0000
Subject: [PATCH 3/8] lib: Implement attach-method unix:<path>

Allow connections to a Unix domain socket which is connected
(via virtio-serial) to a guestfsd running free in an existing
guest.

In order to use this you have to add the following element
to the libvirt XML:

  <channel type='unix'>
    <source mode='bind' path='/tmp/socket'/>
    <target type='virtio'
name='org.libguestfs.channel.0'/>
  </channel>

(or perform the equivalent on the qemu command line).

Then in guestfish, you can do:

  guestfish -n \
    attach-method unix:/tmp/socket : \
    run : \
    ll /

(or any other commands as desired).

Note the '-n' option is necessary to stop autosync, which would
try to unmount all of the guest's disks.
---
 src/launch.c |   85 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 src/proto.c  |   19 +++++++-----
 2 files changed, 95 insertions(+), 9 deletions(-)

diff --git a/src/launch.c b/src/launch.c
index 775f7a6..e50985d 100644
--- a/src/launch.c
+++ b/src/launch.c
@@ -71,6 +71,7 @@
 #include "guestfs_protocol.h"
 
 static int launch_appliance (guestfs_h *g);
+static int connect_unix_socket (guestfs_h *g, const char *sock);
 static int qemu_supports (guestfs_h *g, const char *option);
 
 /* Add a string to the current command line. */
@@ -374,7 +375,17 @@ guestfs__launch (guestfs_h *g)
   if (chmod (g->tmpdir, 0755) == -1)
     fprintf (stderr, "chmod: %s: %m (ignored)\n", g->tmpdir);
 
-  return launch_appliance (g);
+  /* Launch the appliance or attach to an existing daemon. */
+  switch (g->attach_method) {
+  case ATTACH_METHOD_APPLIANCE:
+    return launch_appliance (g);
+
+  case ATTACH_METHOD_UNIX:
+    return connect_unix_socket (g, g->attach_method_arg);
+
+  default:
+    abort ();
+  }
 }
 
 static int
@@ -778,6 +789,78 @@ launch_appliance (guestfs_h *g)
   return -1;
 }
 
+/* Alternate attach method: instead of launching the appliance,
+ * connect to an existing unix socket.
+ */
+static int
+connect_unix_socket (guestfs_h *g, const char *sockpath)
+{
+  int r;
+  struct sockaddr_un addr;
+
+  /* Start the clock ... */
+  gettimeofday (&g->launch_t, NULL);
+
+  /* Set these to nothing so we don't try to kill random processes or
+   * read from random file descriptors.
+   */
+  g->pid = 0;
+  g->recoverypid = 0;
+  g->fd[0] = -1;
+  g->fd[1] = -1;
+
+  if (g->verbose)
+    guestfs___print_timestamped_message (g, "connecting to %s",
sockpath);
+
+  g->sock = socket (AF_UNIX, SOCK_STREAM, 0);
+  if (g->sock == -1) {
+    perrorf (g, "socket");
+    return -1;
+  }
+
+  addr.sun_family = AF_UNIX;
+  strncpy (addr.sun_path, sockpath, UNIX_PATH_MAX);
+  addr.sun_path[UNIX_PATH_MAX-1] = '\0';
+
+  g->state = LAUNCHING;
+
+  if (connect (g->sock, &addr, sizeof addr) == -1) {
+    perrorf (g, "bind");
+    goto cleanup;
+  }
+
+  if (fcntl (g->sock, F_SETFL, O_NONBLOCK) == -1) {
+    perrorf (g, "fcntl");
+    goto cleanup;
+  }
+
+  uint32_t size;
+  void *buf = NULL;
+  r = guestfs___recv_from_daemon (g, &size, &buf);
+  free (buf);
+
+  if (r == -1) return -1;
+
+  if (size != GUESTFS_LAUNCH_FLAG) {
+    error (g, _("guestfs_launch failed, unexpected initial message from
guestfsd"));
+    goto cleanup;
+  }
+
+  if (g->verbose)
+    guestfs___print_timestamped_message (g, "connected");
+
+  if (g->state != READY) {
+    error (g, _("contacted guestfsd, but state != READY"));
+    goto cleanup;
+  }
+
+  return 0;
+
+ cleanup:
+  close (g->sock);
+  return -1;
+}
+
 /* Return the location of the tmpdir (eg. "/tmp") and allow users
  * to override it at runtime using $TMPDIR.
  * http://www.pathname.com/fhs/pub/fhs-2.3.html#TMPTEMPORARYFILES
diff --git a/src/proto.c b/src/proto.c
index 0d63af6..549734b 100644
--- a/src/proto.c
+++ b/src/proto.c
@@ -183,8 +183,8 @@ child_cleanup (guestfs_h *g)
   if (g->recoverypid > 0) kill (g->recoverypid, 9);
   waitpid (g->pid, NULL, 0);
   if (g->recoverypid > 0) waitpid (g->recoverypid, NULL, 0);
-  close (g->fd[0]);
-  close (g->fd[1]);
+  if (g->fd[0] >= 0) close (g->fd[0]);
+  if (g->fd[1] >= 0) close (g->fd[1]);
   close (g->sock);
   g->fd[0] = -1;
   g->fd[1] = -1;
@@ -381,7 +381,8 @@ guestfs___send_to_daemon (guestfs_h *g, const void *v_buf,
size_t n)
   FD_ZERO (&rset);
   FD_ZERO (&wset);
 
-  FD_SET (g->fd[1], &rset);     /* Read qemu stdout for log messages
& EOF. */
+  if (g->fd[1] >= 0)            /* Read qemu stdout for log messages
& EOF. */
+    FD_SET (g->fd[1], &rset);
   FD_SET (g->sock, &rset);      /* Read socket for cancellation &
EOF. */
   FD_SET (g->sock, &wset);      /* Write to socket to send the data. */
 
@@ -398,7 +399,7 @@ guestfs___send_to_daemon (guestfs_h *g, const void *v_buf,
size_t n)
       return -1;
     }
 
-    if (FD_ISSET (g->fd[1], &rset2)) {
+    if (g->fd[1] >= 0 && FD_ISSET (g->fd[1], &rset2)) {
       if (read_log_message_or_eof (g, g->fd[1], 0) == -1)
         return -1;
     }
@@ -460,7 +461,8 @@ guestfs___recv_from_daemon (guestfs_h *g, uint32_t
*size_rtn, void **buf_rtn)
 
   FD_ZERO (&rset);
 
-  FD_SET (g->fd[1], &rset);     /* Read qemu stdout for log messages
& EOF. */
+  if (g->fd[1] >= 0)            /* Read qemu stdout for log messages
& EOF. */
+    FD_SET (g->fd[1], &rset);
   FD_SET (g->sock, &rset);      /* Read socket for data & EOF. */
 
   int max_fd = MAX (g->sock, g->fd[1]);
@@ -492,7 +494,7 @@ guestfs___recv_from_daemon (guestfs_h *g, uint32_t
*size_rtn, void **buf_rtn)
       return -1;
     }
 
-    if (FD_ISSET (g->fd[1], &rset2)) {
+    if (g->fd[1] >= 0 && FD_ISSET (g->fd[1], &rset2)) {
       if (read_log_message_or_eof (g, g->fd[1], 0) == -1) {
         free (*buf_rtn);
         *buf_rtn = NULL;
@@ -650,7 +652,8 @@ guestfs___accept_from_daemon (guestfs_h *g)
 
   FD_ZERO (&rset);
 
-  FD_SET (g->fd[1], &rset);     /* Read qemu stdout for log messages
& EOF. */
+  if (g->fd[1] >= 0)            /* Read qemu stdout for log messages
& EOF. */
+    FD_SET (g->fd[1], &rset);
   FD_SET (g->sock, &rset);      /* Read socket for accept. */
 
   int max_fd = MAX (g->sock, g->fd[1]);
@@ -674,7 +677,7 @@ guestfs___accept_from_daemon (guestfs_h *g)
       return -1;
     }
 
-    if (FD_ISSET (g->fd[1], &rset2)) {
+    if (g->fd[1] >= 0 && FD_ISSET (g->fd[1], &rset2)) {
       if (read_log_message_or_eof (g, g->fd[1], 1) == -1)
         return -1;
     }
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming blog: http://rwmj.wordpress.com
Fedora now supports 80 OCaml packages (the OPEN alternative to F#)
http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora
-------------- next part --------------
>From 5c5231800e53fc24f155a66effa768f30347f0eb Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Thu, 27 Jan 2011 16:54:48 +0000
Subject: [PATCH 4/8] daemon: Allow -r option to run daemon standalone.

This changes several aspects of the daemon.  Currently:

* sysroot will be "" (ie. operate directly on /)
* CHROOT_IN/CHROOT_OUT are disabled
* autosync doesn't try to unmount everything
---
 daemon/daemon.h   |   22 ++++++++++++++--------
 daemon/guestfsd.c |   16 ++++++++++++++--
 daemon/sync.c     |    5 +++--
 3 files changed, 31 insertions(+), 12 deletions(-)

diff --git a/daemon/daemon.h b/daemon/daemon.h
index da991b1..4a686c0 100644
--- a/daemon/daemon.h
+++ b/daemon/daemon.h
@@ -33,6 +33,8 @@
 /*-- in guestfsd.c --*/
 extern int verbose;
 
+extern int autosync_umount;
+
 extern const char *sysroot;
 extern int sysroot_len;
 
@@ -253,17 +255,21 @@ extern void notify_progress (uint64_t position, uint64_t
total);
  */
 #define CHROOT_IN				\
   do {						\
-    int __old_errno = errno;			\
-    if (chroot (sysroot) == -1)			\
-      perror ("CHROOT_IN: sysroot");		\
-    errno = __old_errno;			\
+    if (sysroot_len > 0) {                      \
+      int __old_errno = errno;			\
+      if (chroot (sysroot) == -1)               \
+        perror ("CHROOT_IN: sysroot");		\
+      errno = __old_errno;			\
+    }                                           \
   } while (0)
 #define CHROOT_OUT				\
   do {						\
-    int __old_errno = errno;			\
-    if (chroot (".") == -1)			\
-      perror ("CHROOT_OUT: .");			\
-    errno = __old_errno;			\
+    if (sysroot_len > 0) {                      \
+      int __old_errno = errno;			\
+      if (chroot (".") == -1)			\
+        perror ("CHROOT_OUT: .");               \
+      errno = __old_errno;			\
+    }                                           \
   } while (0)
 
 /* Marks functions which are not implemented.
diff --git a/daemon/guestfsd.c b/daemon/guestfsd.c
index 8d950fa..94e4994 100644
--- a/daemon/guestfsd.c
+++ b/daemon/guestfsd.c
@@ -118,6 +118,9 @@ winsock_init (void)
 const char *sysroot = "/sysroot"; /* No trailing slash. */
 int sysroot_len = 8;
 
+/* If set (the default), do 'umount-all' when performing autosync. */
+int autosync_umount = 1;
+
 /* Not used explicitly, but required by the gnulib 'error' module. */
 const char *program_name = "guestfsd";
 
@@ -125,13 +128,13 @@ static void
 usage (void)
 {
   fprintf (stderr,
-    "guestfsd [-f|--foreground] [-v|--verbose]\n");
+    "guestfsd [-f|--foreground] [-v|--verbose] [-r]\n");
 }
 
 int
 main (int argc, char *argv[])
 {
-  static const char *options = "fv?";
+  static const char *options = "frv?";
   static const struct option long_options[] = {
     { "foreground", 0, 0, 'f' },
     { "help", 0, 0, '?' },
@@ -171,6 +174,15 @@ main (int argc, char *argv[])
       dont_fork = 1;
       break;
 
+      /* The -r flag is used when running standalone.  It changes
+       * several aspects of the daemon.
+       */
+    case 'r':
+      sysroot = "";
+      sysroot_len = 0;
+      autosync_umount = 0;
+      break;
+
     case 'v':
       verbose = 1;
       break;
diff --git a/daemon/sync.c b/daemon/sync.c
index 25c7de8..8f5a461 100644
--- a/daemon/sync.c
+++ b/daemon/sync.c
@@ -51,9 +51,10 @@ do_sync (void)
 int
 do_internal_autosync (void)
 {
-  int r;
+  int r = 0;
 
-  r = do_umount_all ();
+  if (autosync_umount)
+    r = do_umount_all ();
 
   sync_disks ();
 
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
New in Fedora 11: Fedora Windows cross-compiler. Compile Windows
programs, test, and build Windows installers. Over 70 libraries supprt'd
http://fedoraproject.org/wiki/MinGW http://www.annexia.org/fedora_mingw
-------------- next part --------------
>From cfc2bf061165079ed02eba3f42a63e65d4f555b8 Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Fri, 28 Jan 2011 15:44:40 +0000
Subject: [PATCH 5/8] daemon: Remove -f (don't fork) option.

This option was not being used.
---
 appliance/init    |    2 +-
 daemon/guestfsd.c |   33 +++------------------------------
 2 files changed, 4 insertions(+), 31 deletions(-)

diff --git a/appliance/init b/appliance/init
index 4a9476d..1d31cd8 100755
--- a/appliance/init
+++ b/appliance/init
@@ -97,7 +97,7 @@ fi
 
 if ! grep -sq guestfs_rescue=1 /proc/cmdline; then
   # The host will kill qemu abruptly if guestfsd shuts down normally
-  guestfsd -f
+  guestfsd
 
   # Otherwise we try to clean up gracefully. For example, this ensures that a
   # core dump generated by the guest daemon will be written to disk.
diff --git a/daemon/guestfsd.c b/daemon/guestfsd.c
index 94e4994..2d10c82 100644
--- a/daemon/guestfsd.c
+++ b/daemon/guestfsd.c
@@ -1,5 +1,5 @@
 /* libguestfs - the guestfsd daemon
- * Copyright (C) 2009-2010 Red Hat Inc.
+ * Copyright (C) 2009-2011 Red Hat Inc.
  *
  * 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
@@ -18,8 +18,6 @@
 
 #include <config.h>
 
-#define _BSD_SOURCE		/* for daemon(3) */
-
 #ifdef HAVE_WINDOWS_H
 # include <windows.h>
 #endif
@@ -87,17 +85,6 @@ static int print_arginfo (const struct printf_info *info,
size_t n, int *argtype
 
 #ifdef WIN32
 static int
-daemon (int nochdir, int noclose)
-{
-  fprintf (stderr,
-           "On Windows the daemon does not support forking into the "
-           "background.\nYou *must* run the daemon with the -f
option.\n");
-  exit (EXIT_FAILURE);
-}
-#endif /* WIN32 */
-
-#ifdef WIN32
-static int
 winsock_init (void)
 {
   int r;
@@ -128,21 +115,19 @@ static void
 usage (void)
 {
   fprintf (stderr,
-    "guestfsd [-f|--foreground] [-v|--verbose] [-r]\n");
+    "guestfsd [-r] [-v|--verbose]\n");
 }
 
 int
 main (int argc, char *argv[])
 {
-  static const char *options = "frv?";
+  static const char *options = "rv?";
   static const struct option long_options[] = {
-    { "foreground", 0, 0, 'f' },
     { "help", 0, 0, '?' },
     { "verbose", 0, 0, 'v' },
     { 0, 0, 0, 0 }
   };
   int c;
-  int dont_fork = 0;
   char *cmdline;
 
   if (winsock_init () == -1)
@@ -170,10 +155,6 @@ main (int argc, char *argv[])
     if (c == -1) break;
 
     switch (c) {
-    case 'f':
-      dont_fork = 1;
-      break;
-
       /* The -r flag is used when running standalone.  It changes
        * several aspects of the daemon.
        */
@@ -290,14 +271,6 @@ main (int argc, char *argv[])
 
   xdr_destroy (&xdr);
 
-  /* Fork into the background. */
-  if (!dont_fork) {
-    if (daemon (0, 1) == -1) {
-      perror ("daemon");
-      exit (EXIT_FAILURE);
-    }
-  }
-
   /* Enter the main loop, reading and performing actions. */
   main_loop (sock);
 
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming blog: http://rwmj.wordpress.com
Fedora now supports 80 OCaml packages (the OPEN alternative to F#)
http://cocan.org/getting_started_with_ocaml_on_red_hat_and_fedora
-------------- next part --------------
>From 59d06155bf48f6656e5bc75dba99fc2b34515919 Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Fri, 28 Jan 2011 13:19:52 +0000
Subject: [PATCH 6/8] Add guestfs_add_domain 'live' flag.

This optional flag controls whether this API call will try to connect
to a running virtual machine 'guestfsd' process.

If the flag is given and the virtual machine is running, then the
libvirt XML is parsed looking for a suitable <channel> element, and
'guestfs_set_attach_method' is called with the corresponding
virtio-serial socket path.
---
 fish/fish.c                    |    2 +-
 generator/generator_actions.ml |   18 +++-
 src/virt.c                     |  192 ++++++++++++++++++++++++++++++++++-----
 3 files changed, 183 insertions(+), 29 deletions(-)

diff --git a/fish/fish.c b/fish/fish.c
index 10f645f..eb7c8fe 100644
--- a/fish/fish.c
+++ b/fish/fish.c
@@ -1,5 +1,5 @@
 /* guestfish - the filesystem interactive shell
- * Copyright (C) 2009-2010 Red Hat Inc.
+ * Copyright (C) 2009-2011 Red Hat Inc.
  *
  * 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
diff --git a/generator/generator_actions.ml b/generator/generator_actions.ml
index a2e7467..192df77 100644
--- a/generator/generator_actions.ml
+++ b/generator/generator_actions.ml
@@ -1074,7 +1074,7 @@ Please read L<guestfs(3)/INSPECTION> for more
details.");
 This returns the internal QEMU command line.  'debug' commands are
 not part of the formal API and can be removed or changed at any time.");
 
-  ("add_domain", (RInt "nrdisks", [String "dom"],
[String "libvirturi"; Bool "readonly"; String
"iface"]), -1, [FishAlias "domain"],
+  ("add_domain", (RInt "nrdisks", [String "dom"],
[String "libvirturi"; Bool "readonly"; String
"iface"; Bool "live"]), -1, [FishAlias "domain"],
    [],
    "add the disk(s) from a named libvirt domain",
    "\
@@ -1101,12 +1101,18 @@ we connect to the default libvirt URI (or one set
through an
 environment variable, see the libvirt documentation for full
 details).
 
+The optional C<live> flag controls whether this call will try
+to connect to a running virtual machine C<guestfsd> process if
+it sees a suitable E<lt>channelE<gt> element in the libvirt
+XML definition.  The default (if the flag is omitted) is never
+to try.
+
 The other optional parameters are passed directly through to
 C<guestfs_add_drive_opts>.");
 
 (*
 This interface is not quite baked yet. -- RWMJ 2010-11-11
-  ("add_libvirt_dom", (RInt "nrdisks", [Pointer
("virDomainPtr", "dom")], [Bool "readonly"; String
"iface"]), -1, [NotInFish],
+  ("add_libvirt_dom", (RInt "nrdisks", [Pointer
("virDomainPtr", "dom")], [Bool "readonly"; String
"iface"; Bool "live"]), -1, [NotInFish],
    [],
    "add the disk(s) from a libvirt domain",
    "\
@@ -1129,7 +1135,13 @@ from a remote libvirt connection (see
L<http://libvirt.org/remote.html>)
 will fail unless those disks are accessible via the same device path
 locally too.
 
-The optional parameters are passed directly through to
+The optional C<live> flag controls whether this call will try
+to connect to a running virtual machine C<guestfsd> process if
+it sees a suitable E<lt>channelE<gt> element in the libvirt
+XML definition.  The default (if the flag is omitted) is never
+to try.
+
+The other optional parameters are passed directly through to
 C<guestfs_add_drive_opts>.");
 *)
 
diff --git a/src/virt.c b/src/virt.c
index 9f84908..4b66cc8 100644
--- a/src/virt.c
+++ b/src/virt.c
@@ -59,6 +59,8 @@ struct guestfs___add_libvirt_dom_argv {
   int readonly;
 #define GUESTFS___ADD_LIBVIRT_DOM_IFACE_BITMASK (UINT64_C(1)<<1)
   const char *iface;
+#define GUESTFS___ADD_LIBVIRT_DOM_LIVE_BITMASK (UINT64_C(1)<<2)
+  int live;
 };
 
 static int guestfs___add_libvirt_dom (guestfs_h *g, virDomainPtr dom, const
struct guestfs___add_libvirt_dom_argv *optargs);
@@ -73,6 +75,7 @@ guestfs__add_domain (guestfs_h *g, const char *domain_name,
   int r = -1;
   const char *libvirturi;
   int readonly;
+  int live;
   const char *iface;
   struct guestfs___add_libvirt_dom_argv optargs2 = { .bitmask = 0 };
 
@@ -82,6 +85,13 @@ guestfs__add_domain (guestfs_h *g, const char *domain_name,
              ? optargs->readonly : 0;
   iface = optargs->bitmask & GUESTFS_ADD_DOMAIN_IFACE_BITMASK
           ? optargs->iface : NULL;
+  live = optargs->bitmask & GUESTFS_ADD_DOMAIN_LIVE_BITMASK
+         ? optargs->live : 0;
+
+  if (live && readonly) {
+    error (g, _("you cannot set both live and readonly flags"));
+    return -1;
+  }
 
   /* Connect to libvirt, find the domain. */
   conn = virConnectOpenReadOnly (libvirturi);
@@ -108,6 +118,10 @@ guestfs__add_domain (guestfs_h *g, const char *domain_name,
     optargs2.bitmask |= GUESTFS___ADD_LIBVIRT_DOM_IFACE_BITMASK;
     optargs2.iface = iface;
   }
+  if (live) {
+    optargs2.bitmask |= GUESTFS___ADD_LIBVIRT_DOM_LIVE_BITMASK;
+    optargs2.live = live;
+  }
 
   r = guestfs___add_libvirt_dom (g, dom, &optargs2);
 
@@ -282,49 +296,63 @@ guestfs___for_each_disk (guestfs_h *g,
   return r;
 }
 
-static int
-add_disk (guestfs_h *g, const char *filename, const char *format,
-          void *optargs_vp)
-{
-  struct guestfs_add_drive_opts_argv *optargs = optargs_vp;
-
-  if (format) {
-    optargs->bitmask |= GUESTFS_ADD_DRIVE_OPTS_FORMAT_BITMASK;
-    optargs->format = format;
-  } else
-    optargs->bitmask &= ~GUESTFS_ADD_DRIVE_OPTS_FORMAT_BITMASK;
+/* This was proposed as an external API, but it's not quite baked yet. */
 
-  return guestfs__add_drive_opts (g, filename, optargs);
-}
+static int add_disk (guestfs_h *g, const char *filename, const char *format,
void *optargs_vp);
+static int connect_live (guestfs_h *g, virDomainPtr dom);
 
-/* This was proposed as an external API, but it's not quite baked yet. */
 static int
 guestfs___add_libvirt_dom (guestfs_h *g, virDomainPtr dom,
                            const struct guestfs___add_libvirt_dom_argv
*optargs)
 {
-  int r = -1;
-  virErrorPtr err;
   int cmdline_pos;
+  int r;
+  int readonly;
+  const char *iface;
+  int live;
 
-  cmdline_pos = guestfs___checkpoint_cmdline (g);
-
-  int readonly +  readonly      optargs->bitmask &
GUESTFS___ADD_LIBVIRT_DOM_READONLY_BITMASK
     ? optargs->readonly : 0;
-  const char *iface +  iface      optargs->bitmask &
GUESTFS___ADD_LIBVIRT_DOM_IFACE_BITMASK
     ? optargs->iface : NULL;
+  live +    optargs->bitmask & GUESTFS___ADD_LIBVIRT_DOM_LIVE_BITMASK
+    ? optargs->live : 0;
+
+  if (live && readonly) {
+    error (g, _("you cannot set both live and readonly flags"));
+    return -1;
+  }
 
   if (!readonly) {
     virDomainInfo info;
+    virErrorPtr err;
+    int vm_running;
+
     if (virDomainGetInfo (dom, &info) == -1) {
       err = virGetLastError ();
       error (g, _("error getting domain info: %s"), err->message);
-      goto cleanup;
+      return -1;
     }
-    if (info.state != VIR_DOMAIN_SHUTOFF) {
-      error (g, _("error: domain is a live virtual machine.\nYou must use
readonly access because write access to a running virtual machine\ncan cause
disk corruption."));
-      goto cleanup;
+    vm_running = info.state != VIR_DOMAIN_SHUTOFF;
+
+    if (vm_running) {
+      /* If the caller specified the 'live' flag, then they want us to
+       * try to connect to guestfsd if the domain is running.  Note
+       * that live readonly connections are not possible.
+       */
+      if (live)
+        return connect_live (g, dom);
+
+      /* Dangerous to modify the disks of a running VM. */
+      error (g, _("error: domain is a live virtual machine.\n"
+                  "Writing to the disks of a running virtual machine can
cause disk corruption.\n"
+                  "Either use read-only access, or if the guest is running
the guestfsd daemon\n"
+                  "specify live access.  In most libguestfs tools these
options are --ro or\n"
+                  "--live respectively.  Consult the documentation for
further information."));
+      return -1;
     }
   }
 
@@ -339,10 +367,124 @@ guestfs___add_libvirt_dom (guestfs_h *g, virDomainPtr
dom,
     optargs2.iface = iface;
   }
 
+  /* Checkpoint the command line around the operation so that either
+   * all disks are added or none are added.
+   */
+  cmdline_pos = guestfs___checkpoint_cmdline (g);
   r = guestfs___for_each_disk (g, dom, add_disk, &optargs2);
+  if (r == -1)
+    guestfs___rollback_cmdline (g, cmdline_pos);
+
+  return r;
+}
+
+static int
+add_disk (guestfs_h *g, const char *filename, const char *format,
+          void *optargs_vp)
+{
+  struct guestfs_add_drive_opts_argv *optargs = optargs_vp;
+
+  if (format) {
+    optargs->bitmask |= GUESTFS_ADD_DRIVE_OPTS_FORMAT_BITMASK;
+    optargs->format = format;
+  } else
+    optargs->bitmask &= ~GUESTFS_ADD_DRIVE_OPTS_FORMAT_BITMASK;
+
+  return guestfs__add_drive_opts (g, filename, optargs);
+}
+
+static int
+connect_live (guestfs_h *g, virDomainPtr dom)
+{
+  int i, r = -1;
+  virErrorPtr err;
+  xmlDocPtr doc = NULL;
+  xmlXPathContextPtr xpathCtx = NULL;
+  xmlXPathObjectPtr xpathObj = NULL;
+  char *xml = NULL;
+  char *path = NULL;
+  char *attach_method = NULL;
+
+  /* Domain XML. */
+  xml = virDomainGetXMLDesc (dom, 0);
+
+  if (!xml) {
+    err = virGetLastError ();
+    error (g, _("error reading libvirt XML information: %s"),
+           err->message);
+    goto cleanup;
+  }
+
+  /* Parse XML to document. */
+  doc = xmlParseMemory (xml, strlen (xml));
+  if (doc == NULL) {
+    error (g, _("unable to parse XML information returned by
libvirt"));
+    goto cleanup;
+  }
+
+  xpathCtx = xmlXPathNewContext (doc);
+  if (xpathCtx == NULL) {
+    error (g, _("unable to create new XPath context"));
+    goto cleanup;
+  }
+
+  /* This gives us a set of all the <channel> nodes related to the
+   * guestfsd virtio-serial channel.
+   */
+  xpathObj = xmlXPathEvalExpression (BAD_CAST
+      "//devices/channel[@type=\"unix\" and "
+                        "./source/@mode=\"bind\" and "
+                        "./source/@path and "
+                        "./target/@type=\"virtio\" and "
+                       
"./target/@name=\"org.libguestfs.channel.0\"]",
+                                     xpathCtx);
+  if (xpathObj == NULL) {
+    error (g, _("unable to evaluate XPath expression"));
+    goto cleanup;
+  }
+
+  xmlNodeSetPtr nodes = xpathObj->nodesetval;
+  for (i = 0; i < nodes->nodeNr; ++i) {
+    xmlXPathObjectPtr xppath;
+
+    /* See note in function above. */
+    xpathCtx->node = nodes->nodeTab[i];
+
+    /* The path is in <source path=..> attribute. */
+    xppath = xmlXPathEvalExpression (BAD_CAST "./source/@path",
xpathCtx);
+    if (xppath == NULL ||
+        xppath->nodesetval == NULL ||
+        xppath->nodesetval->nodeNr == 0) {
+      xmlXPathFreeObject (xppath);
+      continue;                 /* no type attribute, skip it */
+    }
+    assert (xppath->nodesetval->nodeTab[0]);
+    assert (xppath->nodesetval->nodeTab[0]->type ==
XML_ATTRIBUTE_NODE);
+    xmlAttrPtr attr = (xmlAttrPtr) xppath->nodesetval->nodeTab[0];
+    path = (char *) xmlNodeListGetString (doc, attr->children, 1);
+    xmlXPathFreeObject (xppath);
+    break;
+  }
+
+  if (path == NULL) {
+    error (g, _("this guest has no libvirt <channel> definition for
guestfsd"));
+    goto cleanup;
+  }
+
+  /* Got a path. */
+  attach_method = safe_malloc (g, strlen (path) + 5 + 1);
+  strcpy (attach_method, "unix:");
+  strcat (attach_method, path);
+  r = guestfs_set_attach_method (g, attach_method);
 
  cleanup:
-  if (r == -1) guestfs___rollback_cmdline (g, cmdline_pos);
+  free (path);
+  free (attach_method);
+  free (xml);
+  if (xpathObj) xmlXPathFreeObject (xpathObj);
+  if (xpathCtx) xmlXPathFreeContext (xpathCtx);
+  if (doc) xmlFreeDoc (doc);
+
   return r;
 }
 
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
New in Fedora 11: Fedora Windows cross-compiler. Compile Windows
programs, test, and build Windows installers. Over 70 libraries supprt'd
http://fedoraproject.org/wiki/MinGW http://www.annexia.org/fedora_mingw
-------------- next part --------------
>From f07afe4767587dabf2a74da2c70509277267b0bb Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Thu, 27 Jan 2011 22:45:50 +0000
Subject: [PATCH 7/8] Add documentation for attach method.

---
 generator/generator_actions.ml |    9 ++++--
 src/guestfs.pod                |   57 ++++++++++++++++++++++++++++++++++++++++
 src/virt.c                     |    3 +-
 3 files changed, 65 insertions(+), 4 deletions(-)

diff --git a/generator/generator_actions.ml b/generator/generator_actions.ml
index 192df77..ca2dfe3 100644
--- a/generator/generator_actions.ml
+++ b/generator/generator_actions.ml
@@ -1105,7 +1105,8 @@ The optional C<live> flag controls whether this call
will try
 to connect to a running virtual machine C<guestfsd> process if
 it sees a suitable E<lt>channelE<gt> element in the libvirt
 XML definition.  The default (if the flag is omitted) is never
-to try.
+to try.  See L<guestfs(3)/ATTACHING TO RUNNING DAEMONS> for more
+information.
 
 The other optional parameters are passed directly through to
 C<guestfs_add_drive_opts>.");
@@ -1139,7 +1140,8 @@ The optional C<live> flag controls whether this call
will try
 to connect to a running virtual machine C<guestfsd> process if
 it sees a suitable E<lt>channelE<gt> element in the libvirt
 XML definition.  The default (if the flag is omitted) is never
-to try.
+to try.  See L<guestfs(3)/ATTACHING TO RUNNING DAEMONS> for more
+information.
 
 The other optional parameters are passed directly through to
 C<guestfs_add_drive_opts>.");
@@ -1399,7 +1401,8 @@ and the default.
 Connect to the Unix domain socket I<path>.
 
 This method lets you connect to an existing daemon or (using
-virtio-serial) to a live guest.
+virtio-serial) to a live guest.  For more information, see
+L<guestfs(3)/ATTACHING TO RUNNING DAEMONS>.
 
 =back");
 
diff --git a/src/guestfs.pod b/src/guestfs.pod
index b8b8ca2..c031bcc 100644
--- a/src/guestfs.pod
+++ b/src/guestfs.pod
@@ -972,6 +972,63 @@ For example:
 Note that libguestfs also calls qemu with the -help and -version
 options in order to determine features.
 
+=head2 ATTACHING TO RUNNING DAEMONS
+
+I<Note (1):> This is B<highly experimental> and has a tendency to
eat
+babies.  Use with caution.
+
+I<Note (2):> This section explains how to attach to a running daemon
+from a low level perspective.  For most users, simply using virt tools
+such as L<guestfish(1)> with the I<--live> option will "just
work".
+
+=head3 Using guestfs_set_attach_method
+
+By calling L</guestfs_set_attach_method> you can change how the
+library connects to the C<guestfsd> daemon in L</guestfs_launch>
+(read L</ARCHITECTURE> for some background).
+
+The normal attach method is C<appliance>, where a small appliance is
+created containing the daemon, and then the library connects to this.
+
+Setting attach method to C<unix:I<path>> (where I<path> is
the path of
+a Unix domain socket) causes L</guestfs_launch> to connect to an
+existing daemon over the Unix domain socket.
+
+The normal use for this is to connect to a running virtual machine
+that contains a C<guestfsd> daemon, and send commands so you can read
+and write files inside the live virtual machine.
+
+=head3 Using guestfs_add_domain with live flag
+
+L</guestfs_add_domain> provides some help for getting the
+correct attach method.  If you pass the C<live> option to this
+function, then (if the virtual machine is running) it will
+examine the libvirt XML looking for a virtio-serial channel
+to connect to:
+
+ <domain>
+   ...
+   <devices>
+     ...
+     <channel type='unix'>
+       <source mode='bind' path='/path/to/socket'/>
+       <target type='virtio'
name='org.libguestfs.channel.0'/>
+     </channel>
+     ...
+   </devices>
+ </domain>
+
+L</guestfs_add_domain> extracts C</path/to/socket> and sets the
attach
+method to C<unix:/path/to/socket>.
+
+Some of the libguestfs tools (including guestfish) support a I<--live>
+option which is passed through to L</guestfs_add_domain> thus allowing
+you to attach to and modify live virtual machines.
+
+The virtual machine needs to have been set up beforehand so that it
+has the virtio-serial channel and so that guestfsd is running inside
+it.
+
 =head2 ABI GUARANTEE
 
 We guarantee the libguestfs ABI (binary interface), for public,
diff --git a/src/virt.c b/src/virt.c
index 4b66cc8..a23ac0a 100644
--- a/src/virt.c
+++ b/src/virt.c
@@ -467,7 +467,8 @@ connect_live (guestfs_h *g, virDomainPtr dom)
   }
 
   if (path == NULL) {
-    error (g, _("this guest has no libvirt <channel> definition for
guestfsd"));
+    error (g, _("this guest has no libvirt <channel> definition for
guestfsd\n"
+                "See ATTACHING TO RUNNING DAEMONS in guestfs(3) for
further information."));
     goto cleanup;
   }
 
-- 
1.7.3.5

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
virt-top is 'top' for virtual machines.  Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://et.redhat.com/~rjones/virt-top
-------------- next part --------------
>From 4891a885aa7058faba901c059f08ddfbc5d43159 Mon Sep 17 00:00:00 2001
From: Richard W.M. Jones <rjones at redhat.com>
Date: Fri, 28 Jan 2011 19:25:02 +0000
Subject: [PATCH 8/8] fish: Add guestfish --live, guestmount --live options.

The other programs have the variable, but the flag is not enabled
either because it doesn't make sense or because the implications are
not well understood.
---
 cat/virt-cat.c             |    2 ++
 cat/virt-filesystems.c     |    2 ++
 cat/virt-ls.c              |    2 ++
 df/main.c                  |    2 ++
 fish/fish.c                |    5 +++++
 fish/guestfish.pod         |    5 +++++
 fish/inspect.c             |    6 ++++++
 fish/options.h             |    1 +
 fish/virt.c                |    4 ++++
 fuse/guestmount.c          |    5 +++++
 fuse/guestmount.pod        |    5 +++++
 inspector/virt-inspector.c |    2 ++
 rescue/virt-rescue.c       |    2 ++
 13 files changed, 43 insertions(+), 0 deletions(-)

diff --git a/cat/virt-cat.c b/cat/virt-cat.c
index 470c70c..834d490 100644
--- a/cat/virt-cat.c
+++ b/cat/virt-cat.c
@@ -34,6 +34,7 @@
 guestfs_h *g;
 
 int read_only = 1;
+int live = 0;
 int verbose = 0;
 int keys_from_stdin = 0;
 int echo_keys = 0;
@@ -213,6 +214,7 @@ main (int argc, char *argv[])
    */
   assert (read_only == 1);
   assert (inspector == 1);
+  assert (live == 0);
 
   /* User must specify at least one filename on the command line. */
   if (optind >= argc || argc - optind < 1)
diff --git a/cat/virt-filesystems.c b/cat/virt-filesystems.c
index cd991e8..7eca9eb 100644
--- a/cat/virt-filesystems.c
+++ b/cat/virt-filesystems.c
@@ -46,6 +46,7 @@
 guestfs_h *g;
 
 int read_only = 1;
+int live = 0;
 int verbose = 0;
 int keys_from_stdin = 0;
 int echo_keys = 0;
@@ -301,6 +302,7 @@ main (int argc, char *argv[])
    */
   assert (read_only == 1);
   assert (inspector == 0);
+  assert (live == 0);
 
   /* Must be no extra arguments on the command line. */
   if (optind != argc)
diff --git a/cat/virt-ls.c b/cat/virt-ls.c
index 64bae46..315fb1f 100644
--- a/cat/virt-ls.c
+++ b/cat/virt-ls.c
@@ -35,6 +35,7 @@
 guestfs_h *g;
 
 int read_only = 1;
+int live = 0;
 int verbose = 0;
 int keys_from_stdin = 0;
 int echo_keys = 0;
@@ -227,6 +228,7 @@ main (int argc, char *argv[])
    */
   assert (read_only == 1);
   assert (inspector == 1);
+  assert (live == 0);
 
   /* User must specify at least one directory name on the command line. */
   if (optind >= argc || argc - optind < 1)
diff --git a/df/main.c b/df/main.c
index 7ac5a6c..174692f 100644
--- a/df/main.c
+++ b/df/main.c
@@ -41,6 +41,7 @@
 guestfs_h *g;
 
 int read_only = 1;
+int live = 0;
 int verbose = 0;
 int keys_from_stdin = 0;
 int echo_keys = 0;
@@ -238,6 +239,7 @@ main (int argc, char *argv[])
    */
   assert (read_only == 1);
   assert (inspector == 0);
+  assert (live == 0);
 
   /* Must be no extra arguments on the command line. */
   if (optind != argc)
diff --git a/fish/fish.c b/fish/fish.c
index eb7c8fe..b62c098 100644
--- a/fish/fish.c
+++ b/fish/fish.c
@@ -74,6 +74,7 @@ static int override_progress_bars = -1;
 guestfs_h *g;
 
 int read_only = 0;
+int live = 0;
 int quit = 0;
 int verbose = 0;
 int remote_control_listen = 0;
@@ -124,6 +125,7 @@ usage (int status)
              "  -i|--inspector       Automatically mount
filesystems\n"
              "  --keys-from-stdin    Read passphrases from stdin\n"
              "  --listen             Listen for remote commands\n"
+             "  --live               Connect to a live virtual
machine\n"
              "  -m|--mount dev[:mnt] Mount dev on mnt (if omitted,
/)\n"
              "  -n|--no-sync         Don't autosync\n"
              "  -N|--new type        Create prepared disk (test1.img,
...)\n"
@@ -173,6 +175,7 @@ main (int argc, char *argv[])
     { "inspector", 0, 0, 'i' },
     { "keys-from-stdin", 0, 0, 0 },
     { "listen", 0, 0, 0 },
+    { "live", 0, 0, 0 },
     { "mount", 1, 0, 'm' },
     { "new", 1, 0, 'N' },
     { "no-dest-paths", 0, 0, 'D' },
@@ -277,6 +280,8 @@ main (int argc, char *argv[])
           format = optarg;
       } else if (STREQ (long_options[option_index].name, "csh")) {
         remote_control_csh = 1;
+      } else if (STREQ (long_options[option_index].name, "live")) {
+        live = 1;
       } else {
         fprintf (stderr, _("%s: unknown long option: %s (%d)\n"),
                  program_name, long_options[option_index].name, option_index);
diff --git a/fish/guestfish.pod b/fish/guestfish.pod
index 0f318f0..abf6d7a 100644
--- a/fish/guestfish.pod
+++ b/fish/guestfish.pod
@@ -294,6 +294,11 @@ to try to read passphrases from the user by opening
C</dev/tty>.
 Fork into the background and listen for remote commands.  See section
 L</REMOTE CONTROL GUESTFISH OVER A SOCKET> below.
 
+=item B<--live>
+
+Connect to a live virtual machine.
+(Experimental, see L<guestfs(3)/ATTACHING TO RUNNING DAEMONS>).
+
 =item B<-m dev[:mountpoint]>
 
 =item B<--mount dev[:mountpoint]>
diff --git a/fish/inspect.c b/fish/inspect.c
index 713501e..5e1948c 100644
--- a/fish/inspect.c
+++ b/fish/inspect.c
@@ -73,6 +73,12 @@ compare_keys (const void *p1, const void *p2)
 void
 inspect_mount (void)
 {
+  if (live) {
+    fprintf (stderr, _("%s: don't use --live and -i options
together\n"),
+             program_name);
+    exit (EXIT_FAILURE);
+  }
+
   inspect_do_decrypt ();
 
   char **roots = guestfs_inspect_os (g);
diff --git a/fish/options.h b/fish/options.h
index 728df04..b755d90 100644
--- a/fish/options.h
+++ b/fish/options.h
@@ -67,6 +67,7 @@
 /* Provided by guestfish or guestmount. */
 extern guestfs_h *g;
 extern int read_only;
+extern int live;
 extern int verbose;
 extern int inspector;
 extern int keys_from_stdin;
diff --git a/fish/virt.c b/fish/virt.c
index 13a6d12..b14cee2 100644
--- a/fish/virt.c
+++ b/fish/virt.c
@@ -43,6 +43,10 @@ add_libvirt_drives (const char *guest)
     optargs.bitmask |= GUESTFS_ADD_DOMAIN_READONLY_BITMASK;
     optargs.readonly = 1;
   }
+  if (live) {
+    optargs.bitmask |= GUESTFS_ADD_DOMAIN_LIVE_BITMASK;
+    optargs.live = 1;
+  }
 
   return guestfs_add_domain_argv (g, guest, &optargs);
 }
diff --git a/fuse/guestmount.c b/fuse/guestmount.c
index 3d3da39..6adf140 100644
--- a/fuse/guestmount.c
+++ b/fuse/guestmount.c
@@ -59,6 +59,7 @@
 
 guestfs_h *g = NULL;
 int read_only = 0;
+int live = 0;
 int verbose = 0;
 int inspector = 0;
 int keys_from_stdin = 0;
@@ -879,6 +880,7 @@ usage (int status)
              "  -i|--inspector       Automatically mount
filesystems\n"
              "  --help               Display help message and exit\n"
              "  --keys-from-stdin    Read passphrases from stdin\n"
+             "  --live               Connect to a live virtual
machine\n"
              "  -m|--mount dev[:mnt] Mount dev on mnt (if omitted,
/)\n"
              "  -n|--no-sync         Don't autosync\n"
              "  -o|--option opt      Pass extra option to FUSE\n"
@@ -917,6 +919,7 @@ main (int argc, char *argv[])
     { "help", 0, 0, HELP_OPTION },
     { "inspector", 0, 0, 'i' },
     { "keys-from-stdin", 0, 0, 0 },
+    { "live", 0, 0, 0 },
     { "mount", 1, 0, 'm' },
     { "no-sync", 0, 0, 'n' },
     { "option", 1, 0, 'o' },
@@ -1017,6 +1020,8 @@ main (int argc, char *argv[])
         keys_from_stdin = 1;
       } else if (STREQ (long_options[option_index].name,
"echo-keys")) {
         echo_keys = 1;
+      } else if (STREQ (long_options[option_index].name, "live")) {
+        live = 1;
       } else {
         fprintf (stderr, _("%s: unknown long option: %s (%d)\n"),
                  program_name, long_options[option_index].name, option_index);
diff --git a/fuse/guestmount.pod b/fuse/guestmount.pod
index 567a270..338af97 100644
--- a/fuse/guestmount.pod
+++ b/fuse/guestmount.pod
@@ -143,6 +143,11 @@ mounted on the real virtual machine.
 Read key or passphrase parameters from stdin.  The default is
 to try to read passphrases from the user by opening C</dev/tty>.
 
+=item B<--live>
+
+Connect to a live virtual machine.
+(Experimental, see L<guestfs(3)/ATTACHING TO RUNNING DAEMONS>).
+
 =item B<-m dev[:mnt]> | B<--mount dev[:mnt]>
 
 Mount the named partition or logical volume on the given mountpoint
diff --git a/inspector/virt-inspector.c b/inspector/virt-inspector.c
index 68f8b46..fda9787 100644
--- a/inspector/virt-inspector.c
+++ b/inspector/virt-inspector.c
@@ -38,6 +38,7 @@
 guestfs_h *g;
 
 int read_only = 1;
+int live = 0;
 int verbose = 0;
 int keys_from_stdin = 0;
 int echo_keys = 0;
@@ -226,6 +227,7 @@ main (int argc, char *argv[])
    */
   assert (read_only == 1);
   assert (inspector == 1);
+  assert (live == 0);
 
   /* Must be no extra arguments on the command line. */
   if (optind != argc)
diff --git a/rescue/virt-rescue.c b/rescue/virt-rescue.c
index 6224ad7..58be96b 100644
--- a/rescue/virt-rescue.c
+++ b/rescue/virt-rescue.c
@@ -36,6 +36,7 @@
 guestfs_h *g;
 
 int read_only = 0;
+int live = 0;
 int verbose = 0;
 int keys_from_stdin = 0;
 int echo_keys = 0;
@@ -239,6 +240,7 @@ main (int argc, char *argv[])
   assert (inspector == 0);
   assert (keys_from_stdin == 0);
   assert (echo_keys == 0);
+  assert (live == 0);
 
   /* Must be no extra arguments on the command line. */
   if (optind != argc)
-- 
1.7.3.5

On Fri, Jan 28, 2011 at 07:39:09PM +0000, Richard W.M. Jones
wrote:
> 
> -- 
> Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
> virt-top is 'top' for virtual machines.  Tiny program with many
> powerful monitoring features, net stats, disk stats, logging, etc.
> http://et.redhat.com/~rjones/virt-top
> >From 4891a885aa7058faba901c059f08ddfbc5d43159 Mon Sep 17 00:00:00 2001
> From: Richard W.M. Jones <rjones at redhat.com>
> Date: Fri, 28 Jan 2011 19:25:02 +0000
> Subject: [PATCH 8/8] fish: Add guestfish --live, guestmount --live options.
> 
> The other programs have the variable, but the flag is not enabled
> either because it doesn't make sense or because the implications are
> not well understood.
> ---
>  cat/virt-cat.c             |    2 ++
>  cat/virt-filesystems.c     |    2 ++
>  cat/virt-ls.c              |    2 ++
>  df/main.c                  |    2 ++
I would have thought it was ok to support --live in
these 4, since they're readonly apps in any case.
>  fish/fish.c                |    5 +++++
>  fish/guestfish.pod         |    5 +++++
>  fish/inspect.c             |    6 ++++++
>  fish/options.h             |    1 +
>  fish/virt.c                |    4 ++++
>  fuse/guestmount.c          |    5 +++++
>  fuse/guestmount.pod        |    5 +++++
>  inspector/virt-inspector.c |    2 ++
>  rescue/virt-rescue.c       |    2 ++
--live probably isn't useful for virt-rescue since if
you need to rescue a machine it likely isn't running
correctly & thus not have a guestfsd available.


Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|

On Fri, Jan 28, 2011 at 07:36:34PM +0000, Richard W.M. Jones
wrote:
> 
> NOTE: This is HIGHLY EXPERIMENTAL and can eat babies ...
> 
> Modify the guest libvirt XML to add the virtio-serial channel:
> 
>   <channel type='unix'>
>     <source mode='bind' path='/tmp/socket'/>
>     <target type='virtio'
name='org.libguestfs.channel.0'/>
>   </channel>
> 
> In the guest, compile guestfsd and run it like this:
> 
>   sudo daemon/guestfsd -rv
> 
> (XXX - should current directory be / ...?)
Yeah, I think it makes sense to have it chdir(/) so that
it doesn't accidentally cause a random FS to remain busy.
> In the host connect using:
> 
>   sudo guestfish -d GuestName --live
>   ><fs> run
>   ><fs> ll /
>    etc.
I think this series all looks fine. Obviously the main implication you
mentioned before is that we'll need to be strict about the RPC wire
back-compat support from this point forward.

Also need a init script (or systemd unit) to start the guestfsd if-only-if
it sees the virtio socket present inside the guest.

Regards,
Daniel
--
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|