pre-release

Add basic dmsetup loop support.

VERSION

@@ -1 +1 @@
-2.02.08-cvs (2006-07-17)
+2.02.12-cvs (2006-10-12)

WHATS_NEW

@@ -1,5 +1,71 @@
-Version 2.02.08 - 
-================================
+Version 2.02.12 -
+===================================
+  Propogate clustered flag in vgsplit and require resizeable flag.
+
+Version 2.02.11 - 12th October 2006
+===================================
+  Add clvmd function to return the cluster name. not used by LVM yet.
+  Add cling allocation policy.
+  Change _check_contiguous() to use _for_each_pv().
+  Extend _for_each_pv() to allow termination without error.
+  Abstract _is_contiguous().
+  Remove duplicated pv arg from _check_contiguous().
+  Accept regionsize with lvconvert.
+  Add report columns with underscore before field names ending 'size'.
+  Correct regionsize default on lvcreate man page (MB).
+  Fix clvmd bug that could cause it to die when a node with a long name crashed.
+  Add device size to text metadata.
+  Fix format_text mda_setup pv->size and pv_setup pe_count calculations.
+  Fix _for_each_pv() for mirror with core log.
+  Add lvm_dump.sh script to create a tarball of debugging info from a system.
+  Capture error messages in clvmd and pass them back to the user.
+  Remove unused #defines from filter-md.c.
+  Make clvmd restart init script wait until clvmd has died before starting it.
+  Add -R to clvmd which tells running clvmds to reload their device cache.
+  Add LV column to reports listing kernel modules needed for activation.
+  Show available fields if report given invalid field. (e.g. lvs -o list)
+  Add timestamp functions with --disable-realtime configure option.
+  Add %VG, %LV and %FREE suffices to lvcreate/lvresize --extents arg.
+  Fix two potential NULL pointer derefs in error cases in vg_read().
+  Separate --enable-cluster from locking lib options in lvmconf.sh.
+  Add a missing comma in lvcreate man page.
+
+Version 2.02.10 - 19th September 2006
+=====================================
+  Fix lvconvert mirror change case detection logic.
+  Fix mirror log detachment so it correctly becomes a standalone LV.
+  Extend _check_contiguous() to detect single-area LVs.
+  Include mirror log (untested) in _for_each_pv() processing.
+  Use MIRROR_LOG_SIZE constant.
+  Remove struct seg_pvs from _for_each_pv() to generalise.
+  Avoid adding duplicates to list of parallel PVs to avoid.
+  Fix several incorrect comparisons in parallel area avoidance code.
+  Fix segment lengths when flattening existing parallel areas.
+  Log existing parallel areas prior to allocation.
+  Fix mirror log creation when activation disabled.
+  Don't attempt automatic recovery without proper locking.
+  When using local file locking, skip clustered VGs.
+  Add fallback_to_clustered_locking and fallback_to_local_locking parameters.
+  lvm.static uses built-in cluster locking instead of external locking.
+  Don't attempt to load shared libraries if built statically.
+  Change default locking_lib to liblvm2clusterlock.so.
+  Add skip_dev_dir() to process command line VGs.
+  Stop clvmd complaining about nodes that have left the cluster.
+  Move lvm_snprintf(), split_words() and split_dm_name() into libdevmapper.
+  Add lvconvert man page.
+  Add mirror options to man pages.
+  Prevent mirror renames.
+  Move CMDLIB code into separate file and record whether static build.
+
+Version 2.02.09 - 17th August 2006
+==================================
+  Fix PE_ALIGN for pagesize over 32KB.
+  Separate out LVM1_PE_ALIGN and pe_align().
+  Add lvm_getpagesize wrapper.
+  Add --maxphysicalvolumes to vgchange.
+
+Version 2.02.08 - 15th August 2006
+==================================
   Add checks for duplicate LV name, lvid and PV id before writing metadata.
   Report all sanity check failures, not just the first.
   Fix missing lockfs on first snapshot creation.

WHATS_NEW_DM

@@ -1,5 +1,20 @@
-Version 1.02.09 - 15 Aug 2006
+Version 1.02.12 - 13 Oct 2006
+=============================
+  Avoid deptree attempting to suspend a device that's already suspended.
+
+Version 1.02.11 -  12 Oct 2006
 ==============================
+  Add suspend noflush support.
+  Add basic dmsetup loop support.
+  Switch dmsetup to use dm_malloc and dm_free.
+
+Version 1.02.10 - 19 Sep 2006
+=============================
+  Add dm_snprintf(), dm_split_words() and dm_split_lvm_name() to libdevmapper.
+  Reorder mm bounds_check code to reduce window for a dmeventd race.
+
+Version 1.02.09 - 15 Aug 2006
+=============================
   Add --table argument to dmsetup for a one-line table.
   Abort if errors are found during cmdline option processing.
   Add lockfs indicator to debug output.

configure

@@ -310,7 +310,7 @@ ac_includes_default="\
 #endif"
 
 ac_default_prefix=/usr
-ac_subst_vars='SHELL PATH_SEPARATOR PACKAGE_NAME PACKAGE_TARNAME PACKAGE_VERSION PACKAGE_STRING PACKAGE_BUGREPORT exec_prefix prefix program_transform_name bindir sbindir libexecdir datadir sysconfdir sharedstatedir localstatedir libdir includedir oldincludedir infodir mandir build_alias host_alias target_alias DEFS ECHO_C ECHO_N ECHO_T LIBS build build_cpu build_vendor build_os host host_cpu host_vendor host_os target target_cpu target_vendor target_os AWK CC CFLAGS LDFLAGS CPPFLAGS ac_ct_CC EXEEXT OBJEXT INSTALL_PROGRAM INSTALL_SCRIPT INSTALL_DATA LN_S SET_MAKE RANLIB ac_ct_RANLIB CFLOW_CMD CSCOPE_CMD CPP EGREP ALLOCA LIBOBJS POW_LIB MSGFMT MODPROBE_CMD JOBS STATIC_LINK LVM1 POOL SNAPSHOTS MIRRORS OWNER GROUP COPTIMISE_FLAG CLDFLAGS CLDWHOLEARCHIVE CLDNOWHOLEARCHIVE LDDEPS LIB_SUFFIX LVM_VERSION LVM1_FALLBACK DEBUG DEVMAPPER HAVE_LIBDL HAVE_SELINUX CMDLIB LOCALEDIR CONFDIR STATICDIR INTL_PACKAGE INTL CLVMD CLUSTER FSADM DMEVENTD LTLIBOBJS'
+ac_subst_vars='SHELL PATH_SEPARATOR PACKAGE_NAME PACKAGE_TARNAME PACKAGE_VERSION PACKAGE_STRING PACKAGE_BUGREPORT exec_prefix prefix program_transform_name bindir sbindir libexecdir datadir sysconfdir sharedstatedir localstatedir libdir includedir oldincludedir infodir mandir build_alias host_alias target_alias DEFS ECHO_C ECHO_N ECHO_T LIBS build build_cpu build_vendor build_os host host_cpu host_vendor host_os target target_cpu target_vendor target_os AWK CC CFLAGS LDFLAGS CPPFLAGS ac_ct_CC EXEEXT OBJEXT INSTALL_PROGRAM INSTALL_SCRIPT INSTALL_DATA LN_S SET_MAKE RANLIB ac_ct_RANLIB CFLOW_CMD CSCOPE_CMD CPP EGREP ALLOCA LIBOBJS POW_LIB MSGFMT MODPROBE_CMD JOBS STATIC_LINK LVM1 POOL SNAPSHOTS MIRRORS OWNER GROUP COPTIMISE_FLAG CLDFLAGS CLDWHOLEARCHIVE CLDNOWHOLEARCHIVE LDDEPS LIB_SUFFIX LVM_VERSION LVM1_FALLBACK DEBUG DEVMAPPER HAVE_LIBDL HAVE_SELINUX HAVE_REALTIME CMDLIB LOCALEDIR CONFDIR STATICDIR INTL_PACKAGE INTL CLVMD CLUSTER FSADM DMEVENTD LTLIBOBJS'
 ac_subst_files=''
 
 # Initialize some variables set by options.
@@ -853,6 +853,7 @@ Optional Features:
                           statically.  Default is dynamic linking
   --enable-readline       Enable readline support
   --disable-selinux       Disable selinux support
+  --disable-realtime      Disable realtime clock support
   --enable-debug          Enable debugging
   --disable-devmapper     Disable device-mapper interaction
   --disable-o_direct      Disable O_DIRECT
@@ -1459,6 +1460,7 @@ case "$host_os" in
 		DEVMAPPER=yes
 		ODIRECT=yes
 		SELINUX=yes
+		REALTIME=yes
 		CLUSTER=internal
 		FSADM=no ;;
 	darwin*)
@@ -1473,6 +1475,7 @@ case "$host_os" in
 		DEVMAPPER=yes
 		ODIRECT=no
 		SELINUX=no
+		REALTIME=no
 		CLUSTER=none
 		FSADM=no ;;
 esac
@@ -7431,6 +7434,17 @@ fi;
 echo "$as_me:$LINENO: result: $SELINUX" >&5
 echo "${ECHO_T}$SELINUX" >&6
 
+################################################################################
+echo "$as_me:$LINENO: checking whether to enable realtime support" >&5
+echo $ECHO_N "checking whether to enable realtime support... $ECHO_C" >&6
+# Check whether --enable-realtime or --disable-realtime was given.
+if test "${enable_realtime+set}" = set; then
+  enableval="$enable_realtime"
+  REALTIME=$enableval
+fi;
+echo "$as_me:$LINENO: result: $REALTIME" >&5
+echo "${ECHO_T}$REALTIME" >&6
+
 ################################################################################
 echo "$as_me:$LINENO: checking whether to build cluster LVM daemon" >&5
 echo $ECHO_N "checking whether to build cluster LVM daemon... $ECHO_C" >&6
@@ -7534,14 +7548,6 @@ fi;
 echo "$as_me:$LINENO: result: $CMDLIB" >&5
 echo "${ECHO_T}$CMDLIB" >&6
 
-if test x$CMDLIB = xyes; then
-
-cat >>confdefs.h <<\_ACEOF
-#define CMDLIB 1
-_ACEOF
-
-fi
-
 ################################################################################
 echo "$as_me:$LINENO: checking whether to build fsadm" >&5
 echo $ECHO_N "checking whether to build fsadm... $ECHO_C" >&6
@@ -8428,6 +8434,96 @@ echo "$as_me: WARNING: Disabling selinux" >&2;}
 	fi
 fi
 
+################################################################################
+if test x$REALTIME = xyes; then
+	echo "$as_me:$LINENO: checking for clock_gettime function" >&5
+echo $ECHO_N "checking for clock_gettime function... $ECHO_C" >&6
+	echo "$as_me:$LINENO: checking for clock_gettime in -lrt" >&5
+echo $ECHO_N "checking for clock_gettime in -lrt... $ECHO_C" >&6
+if test "${ac_cv_lib_rt_clock_gettime+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  ac_check_lib_save_LIBS=$LIBS
+LIBS="-lrt  $LIBS"
+cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+
+/* Override any gcc2 internal prototype to avoid an error.  */
+#ifdef __cplusplus
+extern "C"
+#endif
+/* We use char because int might match the return type of a gcc2
+   builtin and then its argument prototype would still apply.  */
+char clock_gettime ();
+int
+main ()
+{
+clock_gettime ();
+  ;
+  return 0;
+}
+_ACEOF
+rm -f conftest.$ac_objext conftest$ac_exeext
+if { (eval echo "$as_me:$LINENO: \"$ac_link\"") >&5
+  (eval $ac_link) 2>conftest.er1
+  ac_status=$?
+  grep -v '^ *+' conftest.er1 >conftest.err
+  rm -f conftest.er1
+  cat conftest.err >&5
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } &&
+	 { ac_try='test -z "$ac_c_werror_flag"
+			 || test ! -s conftest.err'
+  { (eval echo "$as_me:$LINENO: \"$ac_try\"") >&5
+  (eval $ac_try) 2>&5
+  ac_status=$?
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); }; } &&
+	 { ac_try='test -s conftest$ac_exeext'
+  { (eval echo "$as_me:$LINENO: \"$ac_try\"") >&5
+  (eval $ac_try) 2>&5
+  ac_status=$?
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); }; }; then
+  ac_cv_lib_rt_clock_gettime=yes
+else
+  echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+ac_cv_lib_rt_clock_gettime=no
+fi
+rm -f conftest.err conftest.$ac_objext \
+      conftest$ac_exeext conftest.$ac_ext
+LIBS=$ac_check_lib_save_LIBS
+fi
+echo "$as_me:$LINENO: result: $ac_cv_lib_rt_clock_gettime" >&5
+echo "${ECHO_T}$ac_cv_lib_rt_clock_gettime" >&6
+if test $ac_cv_lib_rt_clock_gettime = yes; then
+  HAVE_REALTIME=yes
+else
+  HAVE_REALTIME=no
+fi
+
+	echo "$as_me:$LINENO: result: $HAVE_REALTIME" >&5
+echo "${ECHO_T}$HAVE_REALTIME" >&6
+
+	if test x$HAVE_REALTIME = xyes; then
+
+cat >>confdefs.h <<\_ACEOF
+#define HAVE_REALTIME 1
+_ACEOF
+
+		LIBS="-lrt $LIBS"
+	else
+		{ echo "$as_me:$LINENO: WARNING: Disabling realtime clock" >&5
+echo "$as_me: WARNING: Disabling realtime clock" >&2;}
+	fi
+fi
+
 ################################################################################
 
 for ac_header in getopt.h
@@ -11103,6 +11199,7 @@ fi
 
 
 
+
 
 
 ################################################################################
@@ -11797,6 +11894,7 @@ s,@DEBUG@,$DEBUG,;t t
 s,@DEVMAPPER@,$DEVMAPPER,;t t
 s,@HAVE_LIBDL@,$HAVE_LIBDL,;t t
 s,@HAVE_SELINUX@,$HAVE_SELINUX,;t t
+s,@HAVE_REALTIME@,$HAVE_REALTIME,;t t
 s,@CMDLIB@,$CMDLIB,;t t
 s,@LOCALEDIR@,$LOCALEDIR,;t t
 s,@CONFDIR@,$CONFDIR,;t t

configure.in

@@ -42,6 +42,7 @@ case "$host_os" in
 		DEVMAPPER=yes
 		ODIRECT=yes
 		SELINUX=yes
+		REALTIME=yes
 		CLUSTER=internal
 		FSADM=no ;;
 	darwin*)
@@ -56,6 +57,7 @@ case "$host_os" in
 		DEVMAPPER=yes
 		ODIRECT=no
 		SELINUX=no
+		REALTIME=no
 		CLUSTER=none
 		FSADM=no ;;
 esac
@@ -282,6 +284,13 @@ AC_ARG_ENABLE(selinux, [  --disable-selinux       Disable selinux support],
 SELINUX=$enableval)
 AC_MSG_RESULT($SELINUX)
 
+################################################################################
+dnl -- Disable realtime clock support
+AC_MSG_CHECKING(whether to enable realtime support)
+AC_ARG_ENABLE(realtime, [  --disable-realtime      Disable realtime clock support],
+REALTIME=$enableval)
+AC_MSG_RESULT($REALTIME)
+
 ################################################################################
 dnl -- Build cluster LVM daemon
 AC_MSG_CHECKING(whether to build cluster LVM daemon)
@@ -351,10 +360,6 @@ AC_ARG_ENABLE(cmdlib, [  --enable-cmdlib         Build shared command library],
 CMDLIB=$enableval, CMDLIB=no)
 AC_MSG_RESULT($CMDLIB)
 
-if test x$CMDLIB = xyes; then
-	AC_DEFINE([CMDLIB], 1, [Define to 1 to build the shared command library.])
-fi
-
 ################################################################################
 dnl -- Enable fsadm
 AC_MSG_CHECKING(whether to build fsadm)
@@ -453,6 +458,21 @@ if test x$SELINUX = xyes; then
 	fi
 fi
 
+################################################################################
+dnl -- Check for realtime clock support
+if test x$REALTIME = xyes; then
+	AC_MSG_CHECKING(for clock_gettime function)
+	AC_CHECK_LIB(rt, clock_gettime, HAVE_REALTIME=yes, HAVE_REALTIME=no)
+	AC_MSG_RESULT($HAVE_REALTIME)
+
+	if test x$HAVE_REALTIME = xyes; then
+		AC_DEFINE([HAVE_REALTIME], 1, [Define to 1 to include support for realtime clock.])
+		LIBS="-lrt $LIBS"
+	else
+		AC_MSG_WARN(Disabling realtime clock)
+	fi
+fi
+
 ################################################################################
 dnl -- Check for getopt
 AC_CHECK_HEADERS(getopt.h, AC_DEFINE([HAVE_GETOPTLONG], 1, [Define to 1 to if getopt_long is available.]))
@@ -582,6 +602,7 @@ AC_SUBST(DEBUG)
 AC_SUBST(DEVMAPPER)
 AC_SUBST(HAVE_LIBDL)
 AC_SUBST(HAVE_SELINUX)
+AC_SUBST(HAVE_REALTIME)
 AC_SUBST(CMDLIB)
 AC_SUBST(MSGFMT)
 AC_SUBST(LOCALEDIR)

daemons/clvmd/Makefile.in

@@ -19,6 +19,7 @@ SOURCES = \
 	clvmd-command.c  \
 	clvmd.c          \
 	lvm-functions.c  \
+	refresh_clvmd.c \
 	system-lv.c
 
 ifeq ("@CLVMD@", "gulm")

daemons/clvmd/clvm.h

@@ -63,4 +63,8 @@ static const char CLVMD_SOCKNAME[] = "\0clvmd";
 #define CLVMD_CMD_LOCK_LV           50
 #define CLVMD_CMD_LOCK_VG           51
 
+/* Misc functions */
+#define CLVMD_CMD_REFRESH	    40
+#define CLVMD_CMD_GET_CLUSTERNAME   41
+
 #endif

daemons/clvmd/clvmd-cman.c

@@ -137,7 +137,7 @@ static int _cluster_send_message(void *buf, int msglen, char *csid, const char *
 
 	if (cman_send_data(c_handle, buf, msglen, 0, CLUSTER_PORT_CLVMD, nodeid) <= 0)
 	{
-			log_error(errtext);
+		log_error(errtext);
 	}
 	return msglen;
 }
@@ -152,16 +152,18 @@ static void _get_our_csid(char *csid)
 
 /* Call a callback routine for each node is that known (down means not running a clvmd) */
 static int _cluster_do_node_callback(struct local_client *client,
-			     void (*callback) (struct local_client *, char *,
-					       int))
+				     void (*callback) (struct local_client *, char *,
+						       int))
 {
 	int i;
 	int somedown = 0;
 
 	for (i = 0; i < _get_num_nodes(); i++) {
-		callback(client, (char *)&nodes[i].cn_nodeid, node_updown[nodes[i].cn_nodeid]);
-		if (!node_updown[nodes[i].cn_nodeid])
-			somedown = -1;
+		if (nodes[i].cn_member) {
+			callback(client, (char *)&nodes[i].cn_nodeid, node_updown[nodes[i].cn_nodeid]);
+			if (!node_updown[nodes[i].cn_nodeid])
+				somedown = -1;
+		}
 	}
 	return somedown;
 }
@@ -170,7 +172,7 @@ static int _cluster_do_node_callback(struct local_client *client,
    this currently just means that a node has stopped listening on our port */
 static void event_callback(cman_handle_t handle, void *private, int reason, int arg)
 {
-	char namebuf[MAX_CLUSTER_NAME_LEN];
+	char namebuf[MAX_CLUSTER_MEMBER_NAME_LEN];
 
 	switch (reason) {
         case CMAN_REASON_PORTCLOSED:
@@ -205,7 +207,7 @@ static void event_callback(cman_handle_t handle, void *private, int reason, int
 
 static struct local_client *cman_client;
 static int _cluster_fd_callback(struct local_client *fd, char *buf, int len, char *csid,
-			struct local_client **new_client)
+				struct local_client **new_client)
 {
 
 	/* Save this for data_callback */
@@ -243,7 +245,7 @@ static void _add_up_node(char *csid)
 				 max_updown_nodes);
 		} else {
 			log_error
-			    ("Realloc failed. Node status for clvmd will be wrong. quitting\n");
+				("Realloc failed. Node status for clvmd will be wrong. quitting\n");
 			exit(999);
 		}
 	}
@@ -297,36 +299,37 @@ static void get_members()
 		return;
 	}
 
-	        /* Not enough room for new nodes list ? */
-	        if (num_nodes > count_nodes && nodes) {
-			free(nodes);
-			nodes = NULL;
-		}
+	/* Not enough room for new nodes list ? */
+	if (num_nodes > count_nodes && nodes) {
+		free(nodes);
+		nodes = NULL;
+	}
 
-		if (nodes == NULL) {
-		        count_nodes = num_nodes + 10; /* Overallocate a little */
+	if (nodes == NULL) {
+		count_nodes = num_nodes + 10; /* Overallocate a little */
 		nodes = malloc(count_nodes * sizeof(struct cman_node));
-			if (!nodes) {
-			        log_error("Unable to allocate nodes array\n");
-				exit(5);
-			}
+		if (!nodes) {
+			log_error("Unable to allocate nodes array\n");
+			exit(5);
 		}
+	}
 
 	status = cman_get_nodes(c_handle, count_nodes, &retnodes, nodes);
 	if (status < 0) {
-		        log_error("Unable to get node details");
-			exit(6);
-		}
-
-		if (node_updown == NULL) {
-			node_updown =
-			    (int *) malloc(sizeof(int) *
-					   max(num_nodes, max_updown_nodes));
-			memset(node_updown, 0,
-			       sizeof(int) * max(num_nodes, max_updown_nodes));
-		}
+		log_error("Unable to get node details");
+		exit(6);
 	}
 
+	if (node_updown == NULL) {
+		node_updown =
+			(int *) malloc(sizeof(int) *
+				       max(num_nodes, max_updown_nodes));
+		memset(node_updown, 0,
+		       sizeof(int) * max(num_nodes, max_updown_nodes));
+	}
+}
+
+
 /* Convert a node name to a CSID */
 static int _csid_from_name(char *csid, char *name)
 {
@@ -468,6 +471,18 @@ static int _sync_unlock(const char *resource /* UNUSED */, int lockid)
 
 }
 
+static int _get_cluster_name(char *buf, int buflen)
+{
+	cman_cluster_t cluster_info;
+	int status;
+
+	status = cman_get_cluster(c_handle, &cluster_info);
+	if (!status) {
+		strncpy(buf, cluster_info.ci_name, buflen);
+	}
+	return status;
+}
+
 static struct cluster_ops _cluster_cman_ops = {
 	.cluster_init_completed   = _cluster_init_completed,
 	.cluster_send_message     = _cluster_send_message,
@@ -481,6 +496,7 @@ static struct cluster_ops _cluster_cman_ops = {
 	.get_our_csid             = _get_our_csid,
 	.add_up_node              = _add_up_node,
 	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
 	.sync_lock                = _sync_lock,
 	.sync_unlock              = _sync_unlock,
 };

daemons/clvmd/clvmd-command.c

@@ -75,6 +75,8 @@
 #include "clvmd.h"
 #include "libdlm.h"
 
+extern struct cluster_ops *clops;
+
 /* This is where all the real work happens:
    NOTE: client will be NULL when this is executed on a remote node */
 int do_command(struct local_client *client, struct clvm_header *msg, int msglen,
@@ -117,11 +119,21 @@ int do_command(struct local_client *client, struct clvm_header *msg, int msglen,
 		if (status == EIO) {
 			*retlen =
 			    1 + snprintf(*buf, buflen,
-					 "Internal lvm error, check syslog");
+					 get_last_lvm_error());
 			return EIO;
 		}
 		break;
 
+	case CLVMD_CMD_REFRESH:
+		do_refresh_cache();
+		break;
+
+	case CLVMD_CMD_GET_CLUSTERNAME:
+		status = clops->get_cluster_name(*buf, buflen);
+		if (!status)
+			*retlen = strlen(*buf);
+		break;
+
 	default:
 		/* Won't get here because command is validated in pre_command */
 		break;
@@ -222,6 +234,10 @@ int do_pre_command(struct local_client *client)
 		status = pre_lock_lv(lock_cmd, lock_flags, lockname);
 		break;
 
+	case CLVMD_CMD_REFRESH:
+	case CLVMD_CMD_GET_CLUSTERNAME:
+		break;
+
 	default:
 		log_error("Unknown command %d received\n", header->cmd);
 		status = EINVAL;

daemons/clvmd/clvmd-comms.h

@@ -43,6 +43,8 @@ struct cluster_ops {
 	void (*reread_config) (void);
 	void (*cluster_closedown) (void);
 
+	int (*get_cluster_name)(char *buf, int buflen);
+
 	int (*sync_lock) (const char *resource, int mode, int flags, int *lockid);
 	int (*sync_unlock) (const char *resource, int lockid);
 

daemons/clvmd/clvmd-gulm.c

@@ -973,6 +973,12 @@ static int _cluster_send_message(void *buf, int msglen, char *csid, const char *
 	return gulm_cluster_send_message(buf, msglen, csid, errtext);
 }
 
+static int _get_cluster_name(char *buf, int buflen)
+{
+	strncpy(buf, cluster_name, buflen);
+	return 0;
+}
+
 static struct cluster_ops _cluster_gulm_ops = {
 	.cluster_init_completed   = NULL,
 	.cluster_send_message     = _cluster_send_message,
@@ -987,6 +993,7 @@ static struct cluster_ops _cluster_gulm_ops = {
 	.add_up_node              = gulm_add_up_node,
 	.reread_config            = _reread_config,
 	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
 	.sync_lock                = _sync_lock,
 	.sync_unlock              = _sync_unlock,
 };

daemons/clvmd/clvmd.c

@@ -42,6 +42,7 @@
 #include "clvm.h"
 #include "version.h"
 #include "clvmd.h"
+#include "refresh_clvmd.h"
 #include "libdlm.h"
 #include "system-lv.h"
 #include "list.h"
@@ -66,7 +67,7 @@ static struct local_client local_client_head;
 
 static unsigned short global_xid = 0;	/* Last transaction ID issued */
 
-static struct cluster_ops *clops = NULL;
+struct cluster_ops *clops = NULL;
 
 static char our_csid[MAX_CSID_LEN];
 static unsigned max_csid_len;
@@ -143,6 +144,7 @@ static void usage(char *prog, FILE *file)
 	fprintf(file, "   -V       Show version of clvmd\n");
 	fprintf(file, "   -h       Show this help information\n");
 	fprintf(file, "   -d       Don't fork, run in the foreground\n");
+	fprintf(file, "   -R       Tell all running clvmds in the cluster to reload their device cache\n");
 	fprintf(file, "   -t<secs> Command timeout (default 60 seconds)\n");
 	fprintf(file, "\n");
 }
@@ -173,7 +175,7 @@ int main(int argc, char *argv[])
 	/* Deal with command-line arguments */
 	opterr = 0;
 	optind = 0;
-	while ((opt = getopt(argc, argv, "?vVhdt:")) != EOF) {
+	while ((opt = getopt(argc, argv, "?vVhdt:R")) != EOF) {
 		switch (opt) {
 		case 'h':
 			usage(argv[0], stdout);
@@ -183,6 +185,9 @@ int main(int argc, char *argv[])
 			usage(argv[0], stderr);
 			exit(0);
 
+		case 'R':
+			return refresh_clvmd();
+
 		case 'd':
 			debug++;
 			break;
@@ -1684,6 +1689,7 @@ static void *lvm_thread_fn(void *arg)
 		}
 		pthread_mutex_unlock(&lvm_thread_mutex);
 	}
+	return NULL;
 }
 
 /* Pass down some work to the LVM thread */

daemons/clvmd/lvm-functions.c

@@ -50,12 +50,18 @@
 static struct cmd_context *cmd = NULL;
 static struct dm_hash_table *lv_hash = NULL;
 static pthread_mutex_t lv_hash_lock;
+static char last_error[1024];
 
 struct lv_info {
 	int lock_id;
 	int lock_mode;
 };
 
+char *get_last_lvm_error()
+{
+	return last_error;
+}
+
 /* Return the mode a lock is currently held at (or -1 if not held) */
 static int get_current_lock(char *resource)
 {
@@ -201,8 +207,17 @@ static int do_activate_lv(char *resource, unsigned char lock_flags, int mode)
 	/* Try to get the lock if it's a clustered volume group */
 	if (lock_flags & LCK_CLUSTER_VG) {
 		status = hold_lock(resource, mode, LKF_NOQUEUE);
-		if (status)
+		if (status) {
+			/* Return an LVM-sensible error for this.
+			 * Forcing EIO makes the upper level return this text
+			 * rather than the strerror text for EAGAIN.
+			 */
+			if (errno == EAGAIN) {
+				sprintf(last_error, "Volume is busy on another node");
+				errno = EIO;
+			}
 			return errno;
+		}
 	}
 
 	/* If it's suspended then resume it */
@@ -416,6 +431,13 @@ int do_check_lvm1(char *vgname)
 	return status == 1 ? 0 : EBUSY;
 }
 
+int do_refresh_cache()
+{
+	DEBUGLOG("Refreshing context\n");
+	log_notice("Refreshing context");
+	return refresh_toolcontext(cmd)==1?0:-1;
+}
+
 
 /* Only called at gulm startup. Drop any leftover VG or P_orphan locks
    that might be hanging around if we died for any reason
@@ -505,6 +527,20 @@ static void *get_initial_state()
 	return NULL;
 }
 
+static void lvm2_log_fn(int level, const char *file, int line,
+			const char *message)
+{
+	/*
+	 * Ignore non-error messages, but store the latest one for returning 
+	 * to the user.
+	 */
+	if (level != _LOG_ERR && level != _LOG_FATAL)
+		return;
+
+	strncpy(last_error, message, sizeof(last_error));
+	last_error[sizeof(last_error)-1] = '\0';
+}
+
 /* This checks some basic cluster-LVM configuration stuff */
 static void check_config()
 {
@@ -539,7 +575,7 @@ void init_lvhash()
 /* Called to initialise the LVM context of the daemon */
 int init_lvm(int using_gulm)
 {
-	if (!(cmd = create_toolcontext(NULL))) {
+	if (!(cmd = create_toolcontext(NULL, 0))) {
 		log_error("Failed to allocate command context");
 		return 0;
 	}
@@ -557,5 +593,8 @@ int init_lvm(int using_gulm)
 
 	get_initial_state();
 
+	/* Trap log messages so we can pass them back to the user */
+	init_log_fn(lvm2_log_fn);
+
 	return 1;
 }

daemons/clvmd/lvm-functions.h

@@ -25,11 +25,13 @@ extern int do_lock_lv(unsigned char lock_cmd, unsigned char lock_flags,
 extern int post_lock_lv(unsigned char lock_cmd, unsigned char lock_flags,
 			char *resource);
 extern int do_check_lvm1(char *vgname);
+extern int do_refresh_cache(void);
 extern int init_lvm(int using_gulm);
 extern void init_lvhash(void);
 
 extern int hold_unlock(char *resource);
 extern int hold_lock(char *resource, int mode, int flags);
 extern void unlock_all(void);
+extern char *get_last_lvm_error(void);
 
 #endif

daemons/clvmd/refresh_clvmd.c

@@ -0,0 +1,334 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
+ *
+ * This file is part of LVM2.
+ *
+ * This copyrighted material is made available to anyone wishing to use,
+ * modify, copy, or redistribute it subject to the terms and conditions
+ * of the GNU General Public License v.2.
+ *
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+/*
+ * Tell all clvmds in a cluster to refresh their toolcontext
+ *
+ */
+
+#include <stddef.h>
+#include <sys/socket.h>
+#include <sys/un.h>
+#include <errno.h>
+#include <unistd.h>
+#include <libdevmapper.h>
+#include <stdint.h>
+#include <stdio.h>
+
+#include "clvm.h"
+#include "refresh_clvmd.h"
+
+typedef struct lvm_response {
+	char node[255];
+	char *response;
+	int status;
+	int len;
+} lvm_response_t;
+
+/*
+ * This gets stuck at the start of memory we allocate so we
+ * can sanity-check it at deallocation time
+ */
+#define LVM_SIGNATURE 0x434C564D
+
+static int _clvmd_sock = -1;
+
+/* Open connection to the Cluster Manager daemon */
+static int _open_local_sock(void)
+{
+	int local_socket;
+	struct sockaddr_un sockaddr;
+
+	/* Open local socket */
+	if ((local_socket = socket(PF_UNIX, SOCK_STREAM, 0)) < 0) {
+		fprintf(stderr, "Local socket creation failed: %s", strerror(errno));
+		return -1;
+	}
+
+	memset(&sockaddr, 0, sizeof(sockaddr));
+	memcpy(sockaddr.sun_path, CLVMD_SOCKNAME, sizeof(CLVMD_SOCKNAME));
+
+	sockaddr.sun_family = AF_UNIX;
+
+	if (connect(local_socket,(struct sockaddr *) &sockaddr,
+		    sizeof(sockaddr))) {
+		int saved_errno = errno;
+
+		fprintf(stderr, "connect() failed on local socket: %s\n",
+			  strerror(errno));
+		if (close(local_socket))
+			return -1;
+
+		errno = saved_errno;
+		return -1;
+	}
+
+	return local_socket;
+}
+
+/* Send a request and return the status */
+static int _send_request(char *inbuf, int inlen, char **retbuf)
+{
+	char outbuf[PIPE_BUF];
+	struct clvm_header *outheader = (struct clvm_header *) outbuf;
+	int len;
+	int off;
+	int buflen;
+	int err;
+
+	/* Send it to CLVMD */
+ rewrite:
+	if ( (err = write(_clvmd_sock, inbuf, inlen)) != inlen) {
+	        if (err == -1 && errno == EINTR)
+		        goto rewrite;
+		fprintf(stderr, "Error writing data to clvmd: %s", strerror(errno));
+		return 0;
+	}
+
+	/* Get the response */
+ reread:
+	if ((len = read(_clvmd_sock, outbuf, sizeof(struct clvm_header))) < 0) {
+	        if (errno == EINTR)
+		        goto reread;
+		fprintf(stderr, "Error reading data from clvmd: %s", strerror(errno));
+		return 0;
+	}
+
+	if (len == 0) {
+		fprintf(stderr, "EOF reading CLVMD");
+		errno = ENOTCONN;
+		return 0;
+	}
+
+	/* Allocate buffer */
+	buflen = len + outheader->arglen;
+	*retbuf = dm_malloc(buflen);
+	if (!*retbuf) {
+		errno = ENOMEM;
+		return 0;
+	}
+
+	/* Copy the header */
+	memcpy(*retbuf, outbuf, len);
+	outheader = (struct clvm_header *) *retbuf;
+
+	/* Read the returned values */
+	off = 1;		/* we've already read the first byte */
+	while (off <= outheader->arglen && len > 0) {
+		len = read(_clvmd_sock, outheader->args + off,
+			   buflen - off - offsetof(struct clvm_header, args));
+		if (len > 0)
+			off += len;
+	}
+
+	/* Was it an error ? */
+	if (outheader->status != 0) {
+		errno = outheader->status;
+
+		/* Only return an error here if there are no node-specific
+		   errors present in the message that might have more detail */
+		if (!(outheader->flags & CLVMD_FLAG_NODEERRS)) {
+			fprintf(stderr, "cluster request failed: %s\n", strerror(errno));
+			return 0;
+		}
+
+	}
+
+	return 1;
+}
+
+/* Build the structure header and parse-out wildcard node names */
+static void _build_header(struct clvm_header *head, int cmd, const char *node,
+			  int len)
+{
+	head->cmd = cmd;
+	head->status = 0;
+	head->flags = 0;
+	head->clientid = 0;
+	head->arglen = len;
+
+	if (node) {
+		/*
+		 * Allow a couple of special node names:
+		 * "*" for all nodes,
+		 * "." for the local node only
+		 */
+		if (strcmp(node, "*") == 0) {
+			head->node[0] = '\0';
+		} else if (strcmp(node, ".") == 0) {
+			head->node[0] = '\0';
+			head->flags = CLVMD_FLAG_LOCAL;
+		} else
+			strcpy(head->node, node);
+	} else
+		head->node[0] = '\0';
+}
+
+/*
+ * Send a message to a(or all) node(s) in the cluster and wait for replies
+ */
+static int _cluster_request(char cmd, const char *node, void *data, int len,
+			   lvm_response_t ** response, int *num)
+{
+	char outbuf[sizeof(struct clvm_header) + len + strlen(node) + 1];
+	int *outptr;
+	char *inptr;
+	char *retbuf = NULL;
+	int status;
+	int i;
+	int num_responses = 0;
+	struct clvm_header *head = (struct clvm_header *) outbuf;
+	lvm_response_t *rarray;
+
+	*num = 0;
+
+	if (_clvmd_sock == -1)
+		_clvmd_sock = _open_local_sock();
+
+	if (_clvmd_sock == -1)
+		return 0;
+
+	_build_header(head, cmd, node, len);
+	memcpy(head->node + strlen(head->node) + 1, data, len);
+
+	status = _send_request(outbuf, sizeof(struct clvm_header) +
+			      strlen(head->node) + len, &retbuf);
+	if (!status)
+		goto out;
+
+	/* Count the number of responses we got */
+	head = (struct clvm_header *) retbuf;
+	inptr = head->args;
+	while (inptr[0]) {
+		num_responses++;
+		inptr += strlen(inptr) + 1;
+		inptr += sizeof(int);
+		inptr += strlen(inptr) + 1;
+	}
+
+	/*
+	 * Allocate response array.
+	 * With an extra pair of INTs on the front to sanity
+	 * check the pointer when we are given it back to free
+	 */
+	outptr = dm_malloc(sizeof(lvm_response_t) * num_responses +
+			    sizeof(int) * 2);
+	if (!outptr) {
+		errno = ENOMEM;
+		status = 0;
+		goto out;
+	}
+
+	*response = (lvm_response_t *) (outptr + 2);
+	outptr[0] = LVM_SIGNATURE;
+	outptr[1] = num_responses;
+	rarray = *response;
+
+	/* Unpack the response into an lvm_response_t array */
+	inptr = head->args;
+	i = 0;
+	while (inptr[0]) {
+		strcpy(rarray[i].node, inptr);
+		inptr += strlen(inptr) + 1;
+
+		memcpy(&rarray[i].status, inptr, sizeof(int));
+		inptr += sizeof(int);
+
+		rarray[i].response = dm_malloc(strlen(inptr) + 1);
+		if (rarray[i].response == NULL) {
+			/* Free up everything else and return error */
+			int j;
+			for (j = 0; j < i; j++)
+				dm_free(rarray[i].response);
+			free(outptr);
+			errno = ENOMEM;
+			status = -1;
+			goto out;
+		}
+
+		strcpy(rarray[i].response, inptr);
+		rarray[i].len = strlen(inptr);
+		inptr += strlen(inptr) + 1;
+		i++;
+	}
+	*num = num_responses;
+	*response = rarray;
+
+      out:
+	if (retbuf)
+		dm_free(retbuf);
+
+	return status;
+}
+
+/* Free reply array */
+static int _cluster_free_request(lvm_response_t * response)
+{
+	int *ptr = (int *) response - 2;
+	int i;
+	int num;
+
+	/* Check it's ours to free */
+	if (response == NULL || *ptr != LVM_SIGNATURE) {
+		errno = EINVAL;
+		return 0;
+	}
+
+	num = ptr[1];
+
+	for (i = 0; i < num; i++) {
+		dm_free(response[i].response);
+	}
+
+	dm_free(ptr);
+
+	return 1;
+}
+
+int refresh_clvmd()
+{
+	int num_responses;
+	char args[1]; // No args really.
+	lvm_response_t *response;
+	int saved_errno;
+	int status;
+	int i;
+
+	status = _cluster_request(CLVMD_CMD_REFRESH, "*", args, 0, &response, &num_responses);
+
+	/* If any nodes were down then display them and return an error */
+	for (i = 0; i < num_responses; i++) {
+		if (response[i].status == EHOSTDOWN) {
+			fprintf(stderr, "clvmd not running on node %s",
+				  response[i].node);
+			status = 0;
+			errno = response[i].status;
+		} else if (response[i].status) {
+			fprintf(stderr, "Error resetting node %s: %s",
+				  response[i].node,
+				  response[i].response[0] ?
+				  	response[i].response :
+				  	strerror(response[i].status));
+			status = 0;
+			errno = response[i].status;
+		}
+	}
+
+	saved_errno = errno;
+	_cluster_free_request(response);
+	errno = saved_errno;
+
+	return status;
+}

daemons/clvmd/refresh_clvmd.h

@@ -0,0 +1,2 @@
+int refresh_clvmd(void);
+

daemons/dmeventd/plugins/mirror/dmeventd_mirror.c

@@ -48,7 +48,7 @@ static int _get_mirror_event(char *params)
 	char *p;
 	int log_argc, num_devs, num_failures=0;
 
-	if (max_args <= split_words(params, max_args, args)) {
+	if (max_args <= dm_split_words(params, max_args, 0, args)) {
 		syslog(LOG_ERR, "Unable to split mirror parameters: Arg list too long");
 		return -E2BIG;
 	}
@@ -121,7 +121,7 @@ static int _remove_failed_devices(const char *device)
 	if (strlen(device) > 200)
 		return -ENAMETOOLONG;
 
-	if (!split_dm_name(mem_pool, device, &vg, &lv, &layer)) {
+	if (!dm_split_lvm_name(mem_pool, device, &vg, &lv, &layer)) {
 		syslog(LOG_ERR, "Unable to determine VG name from %s",
 		       device);
 		return -ENOMEM;

dmeventd/mirror/dmeventd_mirror.c

@@ -48,7 +48,7 @@ static int _get_mirror_event(char *params)
 	char *p;
 	int log_argc, num_devs, num_failures=0;
 
-	if (max_args <= split_words(params, max_args, args)) {
+	if (max_args <= dm_split_words(params, max_args, 0, args)) {
 		syslog(LOG_ERR, "Unable to split mirror parameters: Arg list too long");
 		return -E2BIG;
 	}
@@ -121,7 +121,7 @@ static int _remove_failed_devices(const char *device)
 	if (strlen(device) > 200)
 		return -ENAMETOOLONG;
 
-	if (!split_dm_name(mem_pool, device, &vg, &lv, &layer)) {
+	if (!dm_split_lvm_name(mem_pool, device, &vg, &lv, &layer)) {
 		syslog(LOG_ERR, "Unable to determine VG name from %s",
 		       device);
 		return -ENOMEM;

doc/example.conf

@@ -207,11 +207,26 @@ global {
     # Location of proc filesystem
     proc = "/proc"
 
-    # Type of locking to use. Defaults to file-based locking (1).
+    # Type of locking to use. Defaults to local file-based locking (1).
     # Turn locking off by setting to 0 (dangerous: risks metadata corruption
     # if LVM2 commands get run concurrently).
+    # Type 2 uses the external shared library locking_library.
+    # Type 3 uses built-in clustered locking.
     locking_type = 1
 
+    # If using external locking (type 2) and initialisation fails,
+    # with this set to 1 an attempt will be made to use the built-in
+    # clustered locking.
+    # If you are using a customised locking_library you should set this to 0.
+    fallback_to_clustered_locking = 1
+
+    # If an attempt to initialise type 2 or type 3 locking failed, perhaps
+    # because cluster components such as clvmd are not running, with this set
+    # to 1 an attempt will be made to use local file-based locking (type 1).
+    # If this succeeds, only commands against local volume groups will proceed.
+    # Volume Groups marked as clustered will be ignored.
+    fallback_to_local_locking = 1
+
     # Local non-LV directory that holds file-based locks while commands are
     # in progress.  A directory like /tmp that may get wiped on reboot is OK.
     locking_dir = "/var/lock/lvm"
@@ -223,6 +238,9 @@ global {
 
     # Search this directory first for shared libraries.
     #   library_dir = "/lib"
+
+    # The external locking library to load if locking_type is set to 2.
+    #   locking_library = "liblvm2clusterlock.so"
 }
 
 activation {

include/.symlinks

@@ -41,6 +41,7 @@
 ../lib/misc/lvm-exec.h
 ../lib/misc/lvm-file.h
 ../lib/misc/lvm-string.h
+../lib/misc/lvm-wrappers.h
 ../lib/misc/sharedlib.h
 ../lib/regex/matcher.h
 ../lib/report/report.h

lib/Makefile.in

@@ -78,6 +78,8 @@ SOURCES =\
 	misc/lvm-exec.c \
 	misc/lvm-file.c \
 	misc/lvm-string.c \
+	misc/lvm-wrappers.c \
+	misc/timestamp.c \
 	mm/memlock.c \
 	regex/matcher.c \
 	regex/parse_rx.c \

lib/activate/activate.c

@@ -39,7 +39,7 @@ int lvm1_present(struct cmd_context *cmd)
 {
 	char path[PATH_MAX];
 
-	if (lvm_snprintf(path, sizeof(path), "%s/lvm/global", cmd->proc_dir)
+	if (dm_snprintf(path, sizeof(path), "%s/lvm/global", cmd->proc_dir)
 	    < 0) {
 		log_error("LVM1 proc global snprintf failed");
 		return 0;
@@ -51,6 +51,66 @@ int lvm1_present(struct cmd_context *cmd)
 		return 0;
 }
 
+int list_segment_modules(struct dm_pool *mem, const struct lv_segment *seg,
+			 struct list *modules)
+{
+	unsigned int s;
+	struct lv_segment *seg2, *snap_seg;
+	struct list *snh;
+
+	if (seg->segtype->ops->modules_needed &&
+	    !seg->segtype->ops->modules_needed(mem, seg, modules)) {
+		log_error("module string allocation failed");
+		return 0;
+	}
+
+	if (lv_is_origin(seg->lv))
+		list_iterate(snh, &seg->lv->snapshot_segs)
+			if (!list_lv_modules(mem,
+					     list_struct_base(snh,
+							      struct lv_segment,
+							      origin_list)->cow,
+					     modules))
+				return_0;
+
+	if (lv_is_cow(seg->lv)) {
+		snap_seg = find_cow(seg->lv);
+		if (snap_seg->segtype->ops->modules_needed &&
+		    !snap_seg->segtype->ops->modules_needed(mem, snap_seg,
+							    modules)) {
+			log_error("snap_seg module string allocation failed");
+			return 0;
+		}
+	}
+
+	for (s = 0; s < seg->area_count; s++) {
+		switch (seg_type(seg, s)) {
+		case AREA_LV:
+			seg2 = find_seg_by_le(seg_lv(seg, s), seg_le(seg, s));
+			if (seg2 && !list_segment_modules(mem, seg2, modules))
+				return_0;
+			break;
+		case AREA_PV:
+		case AREA_UNASSIGNED:
+			;
+		}
+	}
+
+	return 1;
+}
+
+int list_lv_modules(struct dm_pool *mem, const struct logical_volume *lv,
+		    struct list *modules)
+{
+	struct lv_segment *seg;
+
+	list_iterate_items(seg, &lv->segments)
+		if (!list_segment_modules(mem, seg, modules))
+			return_0;
+
+	return 1;
+}
+
 #ifndef DEVMAPPER_SUPPORT
 void set_activation(int act)
 {
@@ -257,9 +317,9 @@ static int _passes_activation_filter(struct cmd_context *cmd,
 				continue;
 		}
 		/* vgname/lvname */
-		if (lvm_snprintf(path, sizeof(path), "%s/%s", lv->vg->name,
+		if (dm_snprintf(path, sizeof(path), "%s/%s", lv->vg->name,
 				 lv->name) < 0) {
-			log_error("lvm_snprintf error from %s/%s", lv->vg->name,
+			log_error("dm_snprintf error from %s/%s", lv->vg->name,
 				  lv->name);
 			continue;
 		}
@@ -342,7 +402,7 @@ int target_present(const char *target_name, int use_modprobe)
 		if (target_version(target_name, &maj, &min, &patchlevel))
 			return 1;
 
-		if (lvm_snprintf(module, sizeof(module), "dm-%s", target_name)
+		if (dm_snprintf(module, sizeof(module), "dm-%s", target_name)
 		    < 0) {
 			log_error("target_present module name too long: %s",
 				  target_name);

lib/activate/activate.h

@@ -39,6 +39,10 @@ int lvm1_present(struct cmd_context *cmd);
 int target_present(const char *target_name, int use_modprobe);
 int target_version(const char *target_name, uint32_t *maj,
                    uint32_t *min, uint32_t *patchlevel);
+int list_segment_modules(struct dm_pool *mem, const struct lv_segment *seg,
+			 struct list *modules);
+int list_lv_modules(struct dm_pool *mem, const struct logical_volume *lv,
+		    struct list *modules);
 
 void activation_release(void);
 void activation_exit(void);

lib/activate/dev_manager.c

@@ -912,7 +912,7 @@ static int _create_lv_symlinks(struct dev_manager *dm, struct dm_tree_node *root
 		name = dm_tree_node_get_name(child);
 
 		if (name && lvlayer->old_name && *lvlayer->old_name && strcmp(name, lvlayer->old_name)) {
-	        	if (!split_dm_name(dm->mem, lvlayer->old_name, &vgname, &lvname, &layer)) {
+	        	if (!dm_split_lvm_name(dm->mem, lvlayer->old_name, &vgname, &lvname, &layer)) {
                 		log_error("_create_lv_symlinks: Couldn't split up old device name %s", lvlayer->old_name);
                 		return 0;
         		}
@@ -938,7 +938,7 @@ static int _clean_tree(struct dev_manager *dm, struct dm_tree_node *root)
 		if (!(uuid = dm_tree_node_get_uuid(child)))
 			continue;
 
-        	if (!split_dm_name(dm->mem, name, &vgname, &lvname, &layer)) {
+        	if (!dm_split_lvm_name(dm->mem, name, &vgname, &lvname, &layer)) {
                 	log_error("_clean_tree: Couldn't split up device name %s.", name);
                 	return 0;
         	}

lib/activate/fs.c

@@ -30,7 +30,7 @@ static int _mk_dir(const char *dev_dir, const char *vg_name)
 {
 	char vg_path[PATH_MAX];
 
-	if (lvm_snprintf(vg_path, sizeof(vg_path), "%s%s",
+	if (dm_snprintf(vg_path, sizeof(vg_path), "%s%s",
 			 dev_dir, vg_name) == -1) {
 		log_error("Couldn't construct name of volume "
 			  "group directory.");
@@ -53,7 +53,7 @@ static int _rm_dir(const char *dev_dir, const char *vg_name)
 {
 	char vg_path[PATH_MAX];
 
-	if (lvm_snprintf(vg_path, sizeof(vg_path), "%s%s",
+	if (dm_snprintf(vg_path, sizeof(vg_path), "%s%s",
 			 dev_dir, vg_name) == -1) {
 		log_error("Couldn't construct name of volume "
 			  "group directory.");
@@ -87,7 +87,7 @@ static void _rm_blks(const char *dir)
 		if (!strcmp(name, ".") || !strcmp(name, ".."))
 			continue;
 
-		if (lvm_snprintf(path, sizeof(path), "%s/%s", dir, name) == -1) {
+		if (dm_snprintf(path, sizeof(path), "%s/%s", dir, name) == -1) {
 			log_error("Couldn't create path for %s", name);
 			continue;
 		}
@@ -109,28 +109,28 @@ static int _mk_link(const char *dev_dir, const char *vg_name,
 	char vg_path[PATH_MAX];
 	struct stat buf;
 
-	if (lvm_snprintf(vg_path, sizeof(vg_path), "%s%s",
+	if (dm_snprintf(vg_path, sizeof(vg_path), "%s%s",
 			 dev_dir, vg_name) == -1) {
 		log_error("Couldn't create path for volume group dir %s",
 			  vg_name);
 		return 0;
 	}
 
-	if (lvm_snprintf(lv_path, sizeof(lv_path), "%s/%s", vg_path,
+	if (dm_snprintf(lv_path, sizeof(lv_path), "%s/%s", vg_path,
 			 lv_name) == -1) {
 		log_error("Couldn't create source pathname for "
 			  "logical volume link %s", lv_name);
 		return 0;
 	}
 
-	if (lvm_snprintf(link_path, sizeof(link_path), "%s/%s",
+	if (dm_snprintf(link_path, sizeof(link_path), "%s/%s",
 			 dm_dir(), dev) == -1) {
 		log_error("Couldn't create destination pathname for "
 			  "logical volume link for %s", lv_name);
 		return 0;
 	}
 
-	if (lvm_snprintf(lvm1_group_path, sizeof(lvm1_group_path), "%s/group",
+	if (dm_snprintf(lvm1_group_path, sizeof(lvm1_group_path), "%s/group",
 			 vg_path) == -1) {
 		log_error("Couldn't create pathname for LVM1 group file for %s",
 			  vg_name);
@@ -190,7 +190,7 @@ static int _rm_link(const char *dev_dir, const char *vg_name,
 	struct stat buf;
 	char lv_path[PATH_MAX];
 
-	if (lvm_snprintf(lv_path, sizeof(lv_path), "%s%s/%s",
+	if (dm_snprintf(lv_path, sizeof(lv_path), "%s%s/%s",
 			 dev_dir, vg_name, lv_name) == -1) {
 		log_error("Couldn't determine link pathname.");
 		return 0;

lib/commands/toolcontext.c

@@ -67,7 +67,7 @@ static int _get_env_vars(struct cmd_context *cmd)
 
 	/* Set to "" to avoid using any system directory */
 	if ((e = getenv("LVM_SYSTEM_DIR"))) {
-		if (lvm_snprintf(cmd->sys_dir, sizeof(cmd->sys_dir),
+		if (dm_snprintf(cmd->sys_dir, sizeof(cmd->sys_dir),
 				 "%s", e) < 0) {
 			log_error("LVM_SYSTEM_DIR environment variable "
 				  "is too long.");
@@ -167,7 +167,7 @@ static int _process_config(struct cmd_context *cmd)
 		log_verbose("Set umask to %04o", cmd->default_settings.umask);
 
 	/* dev dir */
-	if (lvm_snprintf(cmd->dev_dir, sizeof(cmd->dev_dir), "%s/",
+	if (dm_snprintf(cmd->dev_dir, sizeof(cmd->dev_dir), "%s/",
 			 find_config_tree_str(cmd, "devices/dir",
 					 DEFAULT_DEV_DIR)) < 0) {
 		log_error("Device directory given in config file too long");
@@ -178,7 +178,7 @@ static int _process_config(struct cmd_context *cmd)
 #endif
 
 	/* proc dir */
-	if (lvm_snprintf(cmd->proc_dir, sizeof(cmd->proc_dir), "%s",
+	if (dm_snprintf(cmd->proc_dir, sizeof(cmd->proc_dir), "%s",
 			 find_config_tree_str(cmd, "global/proc",
 					 DEFAULT_PROC_DIR)) < 0) {
 		log_error("Device directory given in config file too long");
@@ -319,7 +319,7 @@ static int _load_config_file(struct cmd_context *cmd, const char *tag)
 	if (*tag)
 		filler = "_";
 
-	if (lvm_snprintf(config_file, sizeof(config_file), "%s/lvm%s%s.conf",
+	if (dm_snprintf(config_file, sizeof(config_file), "%s/lvm%s%s.conf",
 			 cmd->sys_dir, filler, tag) < 0) {
 		log_error("LVM_SYSTEM_DIR or tag was too long");
 		return 0;
@@ -587,7 +587,7 @@ static int _init_filters(struct cmd_context *cmd)
 	if (!(f3 = _init_filter_components(cmd)))
 		return 0;
 
-	if (lvm_snprintf(cache_file, sizeof(cache_file),
+	if (dm_snprintf(cache_file, sizeof(cache_file),
 			 "%s/.cache", cmd->sys_dir) < 0) {
 		log_error("Persistent cache filename too long ('%s/.cache').",
 			  cmd->sys_dir);
@@ -646,8 +646,9 @@ static int _init_formats(struct cmd_context *cmd)
 #endif
 
 #ifdef HAVE_LIBDL
-	/* Load any formats in shared libs */
-	if ((cn = find_config_tree_node(cmd, "global/format_libraries"))) {
+	/* Load any formats in shared libs if not static */
+	if (!cmd->is_static &&
+	    (cn = find_config_tree_node(cmd, "global/format_libraries"))) {
 
 		struct config_value *cv;
 		struct format_type *(*init_format_fn) (struct cmd_context *);
@@ -740,8 +741,9 @@ static int _init_segtypes(struct cmd_context *cmd)
 #endif
 
 #ifdef HAVE_LIBDL
-	/* Load any formats in shared libs */
-	if ((cn = find_config_tree_node(cmd, "global/segment_libraries"))) {
+	/* Load any formats in shared libs unless static */
+	if (!cmd->is_static &&
+	    (cn = find_config_tree_node(cmd, "global/segment_libraries"))) {
 
 		struct config_value *cv;
 		struct segment_type *(*init_segtype_fn) (struct cmd_context *);
@@ -839,7 +841,7 @@ static int _init_backup(struct cmd_context *cmd)
 	min = (uint32_t) find_config_tree_int(cmd, "backup/retain_min",
 					 DEFAULT_ARCHIVE_NUMBER);
 
-	if (lvm_snprintf
+	if (dm_snprintf
 	    (default_dir, sizeof(default_dir), "%s/%s", cmd->sys_dir,
 	     DEFAULT_ARCHIVE_SUBDIR) == -1) {
 		log_err("Couldn't create default archive path '%s/%s'.",
@@ -860,7 +862,7 @@ static int _init_backup(struct cmd_context *cmd)
 	    find_config_tree_bool(cmd, "backup/backup",
 			     DEFAULT_BACKUP_ENABLED);
 
-	if (lvm_snprintf
+	if (dm_snprintf
 	    (default_dir, sizeof(default_dir), "%s/%s", cmd->sys_dir,
 	     DEFAULT_BACKUP_SUBDIR) == -1) {
 		log_err("Couldn't create default backup path '%s/%s'.",
@@ -879,7 +881,7 @@ static int _init_backup(struct cmd_context *cmd)
 }
 
 /* Entry point */
-struct cmd_context *create_toolcontext(struct arg *the_args)
+struct cmd_context *create_toolcontext(struct arg *the_args, unsigned is_static)
 {
 	struct cmd_context *cmd;
 
@@ -902,6 +904,7 @@ struct cmd_context *create_toolcontext(struct arg *the_args)
 	}
 	memset(cmd, 0, sizeof(*cmd));
 	cmd->args = the_args;
+	cmd->is_static = is_static;
 	cmd->hosttags = 0;
 	list_init(&cmd->formats);
 	list_init(&cmd->segtypes);

lib/commands/toolcontext.h

@@ -64,6 +64,7 @@ struct cmd_context {
 	struct command *command;
 	struct arg *args;
 	char **argv;
+	unsigned is_static;	/* Static binary? */
 
 	struct dev_filter *filter;
 	int dump_filter;	/* Dump filter when exiting? */
@@ -87,7 +88,7 @@ struct cmd_context {
 	char proc_dir[PATH_MAX];
 };
 
-struct cmd_context *create_toolcontext(struct arg *the_args);
+struct cmd_context *create_toolcontext(struct arg *the_args, unsigned is_static);
 void destroy_toolcontext(struct cmd_context *cmd);
 int refresh_toolcontext(struct cmd_context *cmd);
 int config_files_changed(struct cmd_context *cmd);

lib/config/config.c

@@ -187,7 +187,7 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 		use_mmap = 0;
 
 	if (use_mmap) {
-		mmap_offset = offset % getpagesize();
+		mmap_offset = offset % lvm_getpagesize();
 		/* memory map the file */
 		p->fb = mmap((caddr_t) 0, size + mmap_offset, PROT_READ,
 			     MAP_PRIVATE, dev_fd(dev), offset - mmap_offset);

lib/config/defaults.h

@@ -32,7 +32,9 @@
 #define DEFAULT_MD_COMPONENT_DETECTION 1
 
 #define DEFAULT_LOCK_DIR "/var/lock/lvm"
-#define DEFAULT_LOCKING_LIB "lvm2_locking.so"
+#define DEFAULT_LOCKING_LIB "liblvm2clusterlock.so"
+#define DEFAULT_FALLBACK_TO_LOCAL_LOCKING 1
+#define DEFAULT_FALLBACK_TO_CLUSTERED_LOCKING 1
 
 #define DEFAULT_MIRROR_LOG_FAULT_POLICY "allocate"
 #define DEFAULT_MIRROR_DEV_FAULT_POLICY "remove"

lib/device/dev-io.c

@@ -176,7 +176,7 @@ static int _aligned_io(struct device_area *where, void *buffer,
 	}
 
 	if (!block_size)
-		block_size = getpagesize();
+		block_size = lvm_getpagesize();
 
 	_widen_region(block_size, where, &widened);
 

lib/display/display.c

@@ -30,6 +30,7 @@ static struct {
 } _policies[] = {
 	{
 	ALLOC_CONTIGUOUS, "contiguous"}, {
+	ALLOC_CLING, "cling"}, {
 	ALLOC_NORMAL, "normal"}, {
 	ALLOC_ANYWHERE, "anywhere"}, {
 	ALLOC_INHERIT, "inherit"}

lib/error/errseg.c

@@ -23,6 +23,7 @@
 #include "targets.h"
 #include "lvm-string.h"
 #include "activate.h"
+#include "str_list.h"
 
 static const char *_errseg_name(const struct lv_segment *seg)
 {
@@ -64,6 +65,18 @@ static int _errseg_target_present(void)
 }
 #endif
 
+static int _errseg_modules_needed(struct dm_pool *mem,
+				  const struct lv_segment *seg,
+				  struct list *modules)
+{
+	if (!str_list_add(mem, modules, "error")) {
+		log_error("error module string list allocation failed");
+		return 0;
+	}
+
+	return 1;
+}
+ 
 static void _errseg_destroy(const struct segment_type *segtype)
 {
 	dm_free((void *)segtype);
@@ -76,6 +89,7 @@ static struct segtype_handler _error_ops = {
 	.add_target_line = _errseg_add_target_line,
 	.target_present = _errseg_target_present,
 #endif
+	.modules_needed = _errseg_modules_needed,
 	.destroy = _errseg_destroy,
 };
 

lib/filters/filter-md.c

@@ -18,14 +18,6 @@
 
 #ifdef linux
 
-/* Lifted from <linux/raid/md_p.h> because of difficulty including it */
-
-#define MD_SB_MAGIC 0xa92b4efc
-#define MD_RESERVED_BYTES (64 * 1024)
-#define MD_RESERVED_SECTORS (MD_RESERVED_BYTES / 512)
-#define MD_NEW_SIZE_SECTORS(x) ((x & ~(MD_RESERVED_SECTORS - 1)) \
-				- MD_RESERVED_SECTORS)
-
 static int _ignore_md(struct dev_filter *f, struct device *dev)
 {
 	int ret;

lib/filters/filter-sysfs.c

@@ -32,7 +32,7 @@ static int _locate_sysfs_blocks(const char *proc, char *path, size_t len)
 		return 0;
 	}
 		
-	if (lvm_snprintf(proc_mounts, sizeof(proc_mounts),
+	if (dm_snprintf(proc_mounts, sizeof(proc_mounts),
 			 "%s/mounts", proc) < 0) {
 		log_error("Failed to create /proc/mounts string");
 		return 0;
@@ -44,9 +44,9 @@ static int _locate_sysfs_blocks(const char *proc, char *path, size_t len)
 	}
 
 	while (fgets(buffer, sizeof(buffer), fp)) {
-		if (split_words(buffer, 4, split) == 4 &&
+		if (dm_split_words(buffer, 4, 0, split) == 4 &&
 		    !strcmp(split[2], "sysfs")) {
-			if (lvm_snprintf(path, len, "%s/%s", split[1],
+			if (dm_snprintf(path, len, "%s/%s", split[1],
 					 "block") >= 0) {
 				r = 1;
 			}
@@ -183,7 +183,7 @@ static int _read_devs(struct dev_set *ds, const char *dir)
                 if (!strcmp(d->d_name, ".") || !strcmp(d->d_name, ".."))
 			continue;
 
-		if (lvm_snprintf(path, sizeof(path), "%s/%s", dir,
+		if (dm_snprintf(path, sizeof(path), "%s/%s", dir,
 				 d->d_name) < 0) {
 			log_error("sysfs path name too long: %s in %s",
 				  d->d_name, dir);

lib/filters/filter.c

@@ -145,7 +145,7 @@ static int _scan_proc_dev(const char *proc, const struct config_node *cn)
 	/* All types unrecognised initially */
 	memset(_max_partitions_by_major, 0, sizeof(int) * NUMBER_OF_MAJORS);
 
-	if (lvm_snprintf(proc_devices, sizeof(proc_devices),
+	if (dm_snprintf(proc_devices, sizeof(proc_devices),
 			 "%s/devices", proc) < 0) {
 		log_error("Failed to create /proc/devices string");
 		return 0;

lib/format1/disk-rep.h

@@ -172,6 +172,7 @@ struct disk_list {
  * Layout constants.
  */
 #define METADATA_ALIGN 4096UL
+#define LVM1_PE_ALIGN (65536UL >> SECTOR_SHIFT)      /* PE alignment */
 
 #define	METADATA_BASE 0UL
 #define	PV_SIZE 1024UL

lib/format1/format1.c

@@ -408,7 +408,7 @@ static int _format1_pv_write(const struct format_type *fmt, struct physical_volu
 
 	/* Ensure any residual PE structure is gone */
 	pv->pe_size = pv->pe_count = 0;
-	pv->pe_start = PE_ALIGN;
+	pv->pe_start = LVM1_PE_ALIGN;
 
 	if (!(mem = dm_pool_create("lvm1 pv_write", 1024))) {
 		stack;
@@ -431,7 +431,7 @@ static int _format1_pv_write(const struct format_type *fmt, struct physical_volu
 	   dev_write in order to make other disk tools happy */
 	dl->pvd.pv_on_disk.base = METADATA_BASE;
 	dl->pvd.pv_on_disk.size = PV_SIZE;
-	dl->pvd.pe_on_disk.base = PE_ALIGN << SECTOR_SHIFT;
+	dl->pvd.pe_on_disk.base = LVM1_PE_ALIGN << SECTOR_SHIFT;
 
 	list_add(&pvs, &dl->list);
 	if (!write_disks(fmt, &pvs)) {

lib/format1/import-export.c

@@ -103,7 +103,7 @@ int import_pv(struct dm_pool *mem, struct device *dev,
 static int _system_id(struct cmd_context *cmd, char *s, const char *prefix)
 {
 
-	if (lvm_snprintf(s, NAME_LEN, "%s%s%lu",
+	if (dm_snprintf(s, NAME_LEN, "%s%s%lu",
 			 prefix, cmd->hostname, time(NULL)) < 0) {
 		log_error("Generated system_id too long");
 		return 0;

lib/format1/layout.c

@@ -153,7 +153,7 @@ int calculate_extent_count(struct physical_volume *pv, uint32_t extent_size,
 		if (pe_start && end < pe_start)
 			end = pe_start;
 
-		pvd->pe_start = _round_up(end, PE_ALIGN);
+		pvd->pe_start = _round_up(end, LVM1_PE_ALIGN);
 
 	} while ((pvd->pe_start + (pvd->pe_total * extent_size))
 		 > pv->size);

lib/format_text/archive.c

@@ -277,7 +277,7 @@ int archive_vg(struct volume_group *vg,
 	}
 
 	for (i = 0; i < 10; i++) {
-		if (lvm_snprintf(archive_name, sizeof(archive_name),
+		if (dm_snprintf(archive_name, sizeof(archive_name),
 				 "%s/%s_%05u.vg", dir, vg->name, ix) < 0) {
 			log_error("Archive file name too long.");
 			return 0;

lib/format_text/archiver.c

@@ -190,7 +190,7 @@ static int __backup(struct volume_group *vg)
 		return 0;
 	}
 
-	if (lvm_snprintf(name, sizeof(name), "%s/%s",
+	if (dm_snprintf(name, sizeof(name), "%s/%s",
 			 vg->cmd->backup_params->dir, vg->name) < 0) {
 		log_error("Failed to generate volume group metadata backup "
 			  "filename.");
@@ -233,7 +233,7 @@ int backup_remove(struct cmd_context *cmd, const char *vg_name)
 {
 	char path[PATH_MAX];
 
-	if (lvm_snprintf(path, sizeof(path), "%s/%s",
+	if (dm_snprintf(path, sizeof(path), "%s/%s",
 			 cmd->backup_params->dir, vg_name) < 0) {
 		log_err("Failed to generate backup filename (for removal).");
 		return 0;
@@ -342,7 +342,7 @@ int backup_restore(struct cmd_context *cmd, const char *vg_name)
 {
 	char path[PATH_MAX];
 
-	if (lvm_snprintf(path, sizeof(path), "%s/%s",
+	if (dm_snprintf(path, sizeof(path), "%s/%s",
 			 cmd->backup_params->dir, vg_name) < 0) {
 		log_err("Failed to generate backup filename (for restore).");
 		return 0;
@@ -397,7 +397,7 @@ void check_current_backup(struct volume_group *vg)
 	if ((vg->status & PARTIAL_VG) || (vg->status & EXPORTED_VG))
 		return;
 
-	if (lvm_snprintf(path, sizeof(path), "%s/%s",
+	if (dm_snprintf(path, sizeof(path), "%s/%s",
 			 vg->cmd->backup_params->dir, vg->name) < 0) {
 		log_debug("Failed to generate backup filename.");
 		return;

lib/format_text/export.c

@@ -233,7 +233,7 @@ static int _sectors_to_units(uint64_t sectors, char *buffer, size_t s)
 	for (i = 0; (d > 1024.0) && _units[i]; i++)
 		d /= 1024.0;
 
-	return lvm_snprintf(buffer, s, "# %g %s", d, _units[i]) > 0;
+	return dm_snprintf(buffer, s, "# %g %s", d, _units[i]) > 0;
 }
 
 /*
@@ -409,6 +409,11 @@ static int _print_pvs(struct formatter *f, struct volume_group *vg)
 			outf(f, "tags = %s", buffer);
 		}
 
+		if (!out_size(f, pv->size, "dev_size = %" PRIu64, pv->size)) {
+			stack;
+			return 0;
+		}
+
 		outf(f, "pe_start = %" PRIu64, pv->pe_start);
 		if (!out_size(f, vg->extent_size * (uint64_t) pv->pe_count,
 			      "pe_count = %u", pv->pe_count)) {
@@ -623,7 +628,7 @@ static int _build_pv_names(struct formatter *f, struct volume_group *vg)
 		pv = pvl->pv;
 
 		/* FIXME But skip if there's already an LV called pv%d ! */
-		if (lvm_snprintf(buffer, sizeof(buffer), "pv%d", count++) < 0)
+		if (dm_snprintf(buffer, sizeof(buffer), "pv%d", count++) < 0)
 			return_0;
 
 		if (!(name = dm_pool_strdup(f->mem, buffer)))

lib/format_text/format-text.c

@@ -853,7 +853,7 @@ static int _scan_file(const struct format_type *fmt)
 			     tmp != dirent->d_name + strlen(dirent->d_name)
 			     - 4)) {
 				vgname = dirent->d_name;
-				if (lvm_snprintf(path, PATH_MAX, "%s/%s",
+				if (dm_snprintf(path, PATH_MAX, "%s/%s",
 						 dl->dir, vgname) < 0) {
 					log_error("Name too long %s/%s",
 						  dl->dir, vgname);
@@ -994,7 +994,7 @@ static int _text_scan(const struct format_type *fmt)
 }
 
 /* For orphan, creates new mdas according to policy.
-   Always have an mda between end-of-label and PE_ALIGN boundary */
+   Always have an mda between end-of-label and pe_align() boundary */
 static int _mda_setup(const struct format_type *fmt,
 		      uint64_t pe_start, uint64_t pe_end,
 		      int pvmetadatacopies,
@@ -1005,15 +1005,12 @@ static int _mda_setup(const struct format_type *fmt,
 	uint64_t start1, mda_size1;	/* First area - start of disk */
 	uint64_t start2, mda_size2;	/* Second area - end of disk */
 	uint64_t wipe_size = 8 << SECTOR_SHIFT;
-	size_t pagesize = getpagesize();
+	size_t pagesize = lvm_getpagesize();
 
-	if (!pvmetadatacopies) {
-		/* Space available for PEs */
-		pv->size -= PE_ALIGN;
+	if (!pvmetadatacopies)
 		return 1;
-	}
 
-	alignment = PE_ALIGN << SECTOR_SHIFT;
+	alignment = pe_align() << SECTOR_SHIFT;
 	disk_size = pv->size << SECTOR_SHIFT;
 	pe_start <<= SECTOR_SHIFT;
 	pe_end <<= SECTOR_SHIFT;
@@ -1027,9 +1024,6 @@ static int _mda_setup(const struct format_type *fmt,
 	/* Requested metadatasize */
 	mda_size1 = pvmetadatasize << SECTOR_SHIFT;
 
-	/* Space available for PEs (before any mdas created) */
-	pv->size -= LABEL_SCAN_SECTORS;
-
 	/* Place mda straight after label area at start of disk */
 	start1 = LABEL_SCAN_SIZE;
 
@@ -1037,11 +1031,8 @@ static int _mda_setup(const struct format_type *fmt,
 	if ((!pe_start && !pe_end) ||
 	    ((pe_start > start1) && (pe_start - start1 >= MDA_SIZE_MIN))) {
 		mda_adjustment = start1 % pagesize;
-		if (mda_adjustment) {
+		if (mda_adjustment)
 			start1 += (pagesize - mda_adjustment);
-			pv->size -= ((pagesize - mda_adjustment) >>
-				     SECTOR_SHIFT);
-		}
 	}
 
 	/* Ensure it's not going to be bigger than the disk! */
@@ -1055,7 +1046,7 @@ static int _mda_setup(const struct format_type *fmt,
 		pvmetadatacopies = 1;
 	}
 
-	/* Round up to PE_ALIGN boundary */
+	/* Round up to pe_align() boundary */
 	mda_adjustment = (mda_size1 + start1) % alignment;
 	if (mda_adjustment)
 		mda_size1 += (alignment - mda_adjustment);
@@ -1071,7 +1062,8 @@ static int _mda_setup(const struct format_type *fmt,
 	/* FIXME If creating new mdas, wipe them! */
 	if (mda_size1) {
 		if (!add_mda(fmt, fmt->cmd->mem, mdas, pv->dev, start1,
-			     mda_size1)) return 0;
+			     mda_size1))
+			return 0;
 
 		if (!dev_set((struct device *) pv->dev, start1,
 			     (size_t) (mda_size1 >
@@ -1080,7 +1072,6 @@ static int _mda_setup(const struct format_type *fmt,
 			return 0;
 		}
 
-		pv->size -= mda_size1 >> SECTOR_SHIFT;
 		if (pvmetadatacopies == 1)
 			return 1;
 	} else
@@ -1125,7 +1116,6 @@ static int _mda_setup(const struct format_type *fmt,
 			log_error("Failed to wipe new metadata area");
 			return 0;
 		}
-		pv->size -= mda_size2 >> SECTOR_SHIFT;
 	} else
 		return 0;
 
@@ -1189,18 +1179,18 @@ static int _text_pv_write(const struct format_type *fmt, struct physical_volume
 
 	/* Set pe_start to first aligned sector after any metadata 
 	 * areas that begin before pe_start */
-	pv->pe_start = PE_ALIGN;
+	pv->pe_start = pe_align();
 	list_iterate_items(mda, &info->mdas) {
 		mdac = (struct mda_context *) mda->metadata_locn;
 		if (pv->dev == mdac->area.dev &&
-		    (mdac->area.start < (pv->pe_start << SECTOR_SHIFT)) &&
+		    (mdac->area.start <= (pv->pe_start << SECTOR_SHIFT)) &&
 		    (mdac->area.start + mdac->area.size >
 		     (pv->pe_start << SECTOR_SHIFT))) {
 			pv->pe_start = (mdac->area.start + mdac->area.size)
 			    >> SECTOR_SHIFT;
-			adjustment = pv->pe_start % PE_ALIGN;
+			adjustment = pv->pe_start % pe_align();
 			if (adjustment)
-				pv->pe_start += (PE_ALIGN - adjustment);
+				pv->pe_start += (pe_align() - adjustment);
 		}
 	}
 	if (!add_da
@@ -1416,8 +1406,8 @@ static int _text_pv_setup(const struct format_type *fmt,
 	struct lvmcache_info *info;
 	int found;
 	uint64_t pe_end = 0;
-
-	/* FIXME if vg, adjust start/end of pe area to avoid mdas! */
+	unsigned mda_count = 0;
+	uint64_t mda_size2 = 0;
 
 	/* FIXME Cope with pvchange */
 	/* FIXME Merge code with _text_create_text_instance */
@@ -1428,11 +1418,16 @@ static int _text_pv_setup(const struct format_type *fmt,
 		if ((info = info_from_pvid(pv->dev->pvid))) {
 			pvmdas = &info->mdas;
 			list_iterate_items(mda, pvmdas) {
+				mda_count++;
 				mdac =
 				    (struct mda_context *) mda->metadata_locn;
 
 				/* FIXME Check it isn't already in use */
 
+				/* Reduce usable device size */
+				if (mda_count > 1)
+					mda_size2 = mdac->area.size >> SECTOR_SHIFT;
+
 				/* Ensure it isn't already on list */
 				found = 0;
 				list_iterate_items(mda2, mdas) {
@@ -1470,6 +1465,17 @@ static int _text_pv_setup(const struct format_type *fmt,
 			}
 		}
 
+		/* FIXME Cope with genuine pe_count 0 */
+
+		/* If missing, estimate pv->size from file-based metadata */
+		if (!pv->size && pv->pe_count)
+			pv->size = pv->pe_count * (uint64_t) vg->extent_size +
+				   pv->pe_start + mda_size2;
+
+		/* Recalculate number of extents that will fit */
+		if (!pv->pe_count)
+			pv->pe_count = (pv->size - pv->pe_start - mda_size2) / vg->extent_size;
+
 		/* Unlike LVM1, we don't store this outside a VG */
 		/* FIXME Default from config file? vgextend cmdline flag? */
 		pv->status |= ALLOCATABLE_PV;
@@ -1533,7 +1539,7 @@ static struct format_instance *_text_create_text_instance(const struct format_ty
 		dir_list = &((struct mda_lists *) fmt->private)->dirs;
 
 		list_iterate_items(dl, dir_list) {
-			if (lvm_snprintf(path, PATH_MAX, "%s/%s",
+			if (dm_snprintf(path, PATH_MAX, "%s/%s",
 					 dl->dir, vgname) < 0) {
 				log_error("Name too long %s/%s", dl->dir,
 					  vgname);

lib/format_text/import_vsn1.c

@@ -179,6 +179,9 @@ static int _read_pv(struct format_instance *fid, struct dm_pool *mem,
 		return 0;
 	}
 
+	/* Late addition */
+	_read_int64(pvn, "dev_size", &pv->size);
+
 	if (!_read_int64(pvn, "pe_start", &pv->pe_start)) {
 		log_error("Couldn't read extent size for volume group.");
 		return 0;
@@ -206,7 +209,7 @@ static int _read_pv(struct format_instance *fid, struct dm_pool *mem,
 	vg->free_count += pv->pe_count;
 
 	pv->pe_size = vg->extent_size;
-	pv->size = vg->extent_size * (uint64_t) pv->pe_count;
+
 	pv->pe_alloc_count = 0;
 	pv->fmt = fid->fmt;
 

lib/format_text/layout.h

@@ -83,6 +83,6 @@ struct mda_context {
 #define FMTT_VERSION 1
 #define MDA_HEADER_SIZE 512
 #define LVM2_LABEL "LVM2 001"
-#define MDA_SIZE_MIN (8 * (unsigned) getpagesize())
+#define MDA_SIZE_MIN (8 * (unsigned) lvm_getpagesize())
 
 #endif

lib/locking/cluster_locking.c

@@ -398,9 +398,9 @@ int lock_resource(struct cmd_context *cmd, const char *resource, int flags)
 	case LCK_VG:
 		/* If the VG name is empty then lock the unused PVs */
 		if (!*resource)
-			lvm_snprintf(lockname, sizeof(lockname), "P_orphans");
+			dm_snprintf(lockname, sizeof(lockname), "P_orphans");
 		else
-			lvm_snprintf(lockname, sizeof(lockname), "V_%s",
+			dm_snprintf(lockname, sizeof(lockname), "V_%s",
 				     resource);
 
 		cluster_cmd = CLVMD_CMD_LOCK_VG;

lib/locking/file_locking.c

@@ -212,10 +212,10 @@ static int _file_lock_resource(struct cmd_context *cmd, const char *resource,
 	switch (flags & LCK_SCOPE_MASK) {
 	case LCK_VG:
 		if (!*resource)
-			lvm_snprintf(lockfile, sizeof(lockfile),
+			dm_snprintf(lockfile, sizeof(lockfile),
 				     "%s/P_orphans", _lock_dir);
 		else
-			lvm_snprintf(lockfile, sizeof(lockfile),
+			dm_snprintf(lockfile, sizeof(lockfile),
 				     "%s/V_%s", _lock_dir, resource);
 
 		if (!_lock_file(lockfile, flags))

lib/locking/locking.c

@@ -20,6 +20,7 @@
 #include "activate.h"
 #include "toolcontext.h"
 #include "memlock.h"
+#include "defaults.h"
 
 #include <signal.h>
 #include <sys/stat.h>
@@ -134,24 +135,31 @@ int init_locking(int type, struct cmd_context *cmd)
 		return 1;
 
 	case 1:
+		log_very_verbose("File-based locking selected.");
 		if (!init_file_locking(&_locking, cmd))
 			break;
-		log_very_verbose("File-based locking enabled.");
 		return 1;
 
 #ifdef HAVE_LIBDL
 	case 2:
-		if (!init_external_locking(&_locking, cmd))
+		if (!cmd->is_static) {
+			log_very_verbose("External locking selected.");
+			if (!init_external_locking(&_locking, cmd))
+				break;
+			return 1;
+		}
+		if (!find_config_tree_int(cmd, "locking/fallback_to_clustered_locking",
+					  DEFAULT_FALLBACK_TO_CLUSTERED_LOCKING))
 			break;
-		log_very_verbose("External locking enabled.");
-		return 1;
+		log_very_verbose("Falling back to clustered locking.");
+		/* Fall through */
 #endif
 
 #ifdef CLUSTER_LOCKING_INTERNAL
 	case 3:
+		log_very_verbose("Cluster locking selected.");
 		if (!init_cluster_locking(&_locking, cmd))
 			break;
-		log_very_verbose("Cluster locking enabled.");
 		return 1;
 #endif
 
@@ -160,6 +168,16 @@ int init_locking(int type, struct cmd_context *cmd)
 		return 0;
 	}
 
+	if ((type == 2 || type == 3) &&
+            find_config_tree_int(cmd, "locking/fallback_to_local_locking",
+				 DEFAULT_FALLBACK_TO_LOCAL_LOCKING)) {
+		log_print("WARNING: Falling back to local file-based locking.");
+		log_print("Volume Groups with the clustered attribute will "
+			  "be inaccessible.");
+		if (init_file_locking(&_locking, cmd))
+			return 1;
+	}
+
 	if (!ignorelockingfailure())
 		return 0;
 
@@ -189,7 +207,7 @@ int check_lvm1_vg_inactive(struct cmd_context *cmd, const char *vgname)
 	if (!*vgname)
 		return 1;
 
-	if (lvm_snprintf(path, sizeof(path), "%s/lvm/VGs/%s", cmd->proc_dir,
+	if (dm_snprintf(path, sizeof(path), "%s/lvm/VGs/%s", cmd->proc_dir,
 			 vgname) < 0) {
 		log_error("LVM1 proc VG pathname too long for %s", vgname);
 		return 0;

lib/log/log.c

@@ -318,7 +318,7 @@ void print_log(int level, const char *file, int line, const char *format, ...)
       log_it:
 	if (!_log_suppress) {
 		if (_verbose_level > _LOG_DEBUG)
-			lvm_snprintf(locn, sizeof(locn), "#%s:%d ",
+			dm_snprintf(locn, sizeof(locn), "#%s:%d ",
 				     file, line);
 		else
 			locn[0] = '\0';
@@ -413,7 +413,7 @@ void print_log(int level, const char *file, int line, const char *format, ...)
 		_already_logging = 1;
 		memset(&buf, ' ', sizeof(buf));
 		bufused = 0;
-		if ((n = lvm_snprintf(buf, sizeof(buf) - bufused - 1,
+		if ((n = dm_snprintf(buf, sizeof(buf) - bufused - 1,
 				      "%s:%d %s%s", file, line, _cmd_name,
 				      _msg_prefix)) == -1)
 			goto done;

lib/metadata/lv_manip.c

@@ -158,7 +158,7 @@ void release_lv_segment_area(struct lv_segment *seg, uint32_t s,
 	}
 
 	if (seg_lv(seg, s)->status & MIRROR_IMAGE) {
-		lv_reduce(seg_lv(seg, s), area_reduction);	
+		lv_reduce(seg_lv(seg, s), area_reduction);
 		return;
 	}
 
@@ -398,6 +398,7 @@ struct alloced_area {
  * Details of an allocation attempt
  */
 struct alloc_handle {
+	struct cmd_context *cmd;
 	struct dm_pool *mem;
 
 	alloc_policy_t alloc;		/* Overall policy */
@@ -417,7 +418,8 @@ struct alloc_handle {
 /*
  * Preparation for a specific allocation attempt
  */
-static struct alloc_handle *_alloc_init(struct dm_pool *mem,
+static struct alloc_handle *_alloc_init(struct cmd_context *cmd,
+					struct dm_pool *mem,
 					const struct segment_type *segtype,
 					alloc_policy_t alloc,
 					uint32_t mirrors,
@@ -464,6 +466,8 @@ static struct alloc_handle *_alloc_init(struct dm_pool *mem,
 	if (segtype_is_virtual(segtype))
 		return ah;
 
+	ah->cmd = cmd;
+
 	if (!(ah->mem = dm_pool_create("allocation", 1024))) {
 		log_error("allocation pool creation failed");
 		return NULL;
@@ -490,6 +494,49 @@ void alloc_destroy(struct alloc_handle *ah)
 		dm_pool_destroy(ah->mem);
 }
 
+static int _log_parallel_areas(struct dm_pool *mem, struct list *parallel_areas)
+{
+	struct seg_pvs *spvs;
+	struct pv_list *pvl;
+	char *pvnames;
+
+	if (!parallel_areas)
+		return 1;
+
+	if (!dm_pool_begin_object(mem, 256)) {
+		log_error("dm_pool_begin_object failed");
+		return 0;
+	}
+
+	list_iterate_items(spvs, parallel_areas) {
+		list_iterate_items(pvl, &spvs->pvs) {
+			if (!dm_pool_grow_object(mem, dev_name(pvl->pv->dev), strlen(dev_name(pvl->pv->dev)))) {
+				log_error("dm_pool_grow_object failed");
+				dm_pool_abandon_object(mem);
+				return 0;
+			}
+			if (!dm_pool_grow_object(mem, " ", 1)) {
+				log_error("dm_pool_grow_object failed");
+				dm_pool_abandon_object(mem);
+				return 0;
+			}
+		}
+
+		if (!dm_pool_grow_object(mem, "\0", 1)) {
+			log_error("dm_pool_grow_object failed");
+			dm_pool_abandon_object(mem);
+			return 0;
+		}
+
+		pvnames = dm_pool_end_object(mem);
+		log_debug("Parallel PVs at LE %" PRIu32 " length %" PRIu32 ": %s",
+			  spvs->le, spvs->len, pvnames);
+		dm_pool_free(mem, pvnames);
+	}
+
+	return 1;
+}
+
 static int _setup_alloced_segment(struct logical_volume *lv, uint32_t status,
 				  uint32_t area_count,
 				  uint32_t stripe_size,
@@ -614,7 +661,7 @@ static int _alloc_parallel_area(struct alloc_handle *ah, uint32_t needed,
 	if (log_area) {
 		ah->log_area.pv = log_area->map->pv;
 		ah->log_area.pe = log_area->start;
-		ah->log_area.len = 1;	/* FIXME Calculate & check this */
+		ah->log_area.len = MIRROR_LOG_SIZE;	/* FIXME Calculate & check this */
 		consume_pv_area(log_area, ah->log_area.len);
 	}
 
@@ -623,6 +670,82 @@ static int _alloc_parallel_area(struct alloc_handle *ah, uint32_t needed,
 	return 1;
 }
 
+/*
+ * Call fn for each AREA_PV used by the LV segment at lv:le of length *max_seg_len.
+ * If any constituent area contains more than one segment, max_seg_len is
+ * reduced to cover only the first.
+ * fn should return 0 on error, 1 to continue scanning or >1 to terminate without error.
+ * In the last case, this function passes on the return code.
+ */
+static int _for_each_pv(struct cmd_context *cmd, struct logical_volume *lv,
+			uint32_t le, uint32_t len, uint32_t *max_seg_len,
+			uint32_t first_area, uint32_t max_areas,
+			int top_level_area_index,
+			int only_single_area_segments,
+			int (*fn)(struct cmd_context *cmd,
+				  struct pv_segment *peg, uint32_t s,
+				  void *data),
+			void *data)
+{
+	struct lv_segment *seg;
+	uint32_t s;
+	uint32_t remaining_seg_len, area_len, area_multiple;
+	int r = 1;
+
+	if (!(seg = find_seg_by_le(lv, le))) {
+		log_error("Failed to find segment for %s extent %" PRIu32,
+			  lv->name, le);
+		return 0;
+	}
+
+	/* Remaining logical length of segment */
+	remaining_seg_len = seg->len - (le - seg->le);
+
+	if (remaining_seg_len > len)
+		remaining_seg_len = len;
+
+	if (max_seg_len && *max_seg_len > remaining_seg_len)
+		*max_seg_len = remaining_seg_len;
+
+	area_multiple = segtype_is_striped(seg->segtype) ? seg->area_count : 1;
+	area_len = remaining_seg_len / area_multiple ? : 1;
+
+	for (s = first_area;
+	     s < seg->area_count && (!max_areas || s <= max_areas);
+	     s++) {
+		if (seg_type(seg, s) == AREA_LV) {
+			if (!(r = _for_each_pv(cmd, seg_lv(seg, s),
+					       seg_le(seg, s) +
+					       (le - seg->le) / area_multiple,
+					       area_len, max_seg_len,
+					       only_single_area_segments ? 0 : 0,
+					       only_single_area_segments ? 1 : 0,
+					       top_level_area_index != -1 ? top_level_area_index : s,
+					       only_single_area_segments, fn,
+					       data)))
+				stack;
+		} else if (seg_type(seg, s) == AREA_PV)
+			if (!(r = fn(cmd, seg_pvseg(seg, s), top_level_area_index != -1 ? top_level_area_index : s, data)))
+				stack;
+		if (r != 1)
+			return r;
+	}
+
+	/* FIXME only_single_area_segments used as workaround to skip log LV - needs new param? */
+	if (!only_single_area_segments && seg_is_mirrored(seg) && seg->log_lv) {
+		if (!(r = _for_each_pv(cmd, seg->log_lv, 0, MIRROR_LOG_SIZE,
+				       NULL, 0, 0, 0, only_single_area_segments,
+				       fn, data)))
+			stack;
+		if (r != 1)
+			return r;
+	}
+
+	/* FIXME Add snapshot cow LVs etc. */
+
+	return 1;
+}
+
 static int _comp_area(const void *l, const void *r)
 {
 	const struct pv_area *lhs = *((const struct pv_area **) l);
@@ -637,33 +760,114 @@ static int _comp_area(const void *l, const void *r)
 	return 0;
 }
 
+/*
+ * Search for pvseg that matches condition
+ */
+struct pv_match {
+	int (*condition)(struct pv_segment *pvseg, struct pv_area *pva);
+
+	struct pv_area **areas;
+	struct pv_area *pva;
+	uint32_t areas_size;
+	int s;	/* Area index of match */
+};
+
+/*
+ * Is PV area on the same PV?
+ */
+static int _is_same_pv(struct pv_segment *pvseg, struct pv_area *pva)
+{
+	if (pvseg->pv != pva->map->pv)
+		return 0;
+
+	return 1;
+}
+
+/*
+ * Is PV area contiguous to PV segment?
+ */
+static int _is_contiguous(struct pv_segment *pvseg, struct pv_area *pva)
+{
+	if (pvseg->pv != pva->map->pv)
+		return 0;
+
+	if (pvseg->pe + pvseg->len != pva->start)
+		return 0;
+
+	return 1;
+}
+
+static int _is_condition(struct cmd_context *cmd,
+			 struct pv_segment *pvseg, uint32_t s,
+			 void *data)
+{
+	struct pv_match *pvmatch = data;
+
+	if (!pvmatch->condition(pvseg, pvmatch->pva))
+		return 1;	/* Continue */
+
+	if (s >= pvmatch->areas_size)
+		return 1;
+
+	pvmatch->areas[s] = pvmatch->pva;
+
+	return 2;	/* Finished */
+}
+
+/*
+ * Is pva on same PV as any existing areas?
+ */
+static int _check_cling(struct cmd_context *cmd,
+			struct lv_segment *prev_lvseg, struct pv_area *pva,
+			struct pv_area **areas, uint32_t areas_size)
+{
+	struct pv_match pvmatch;
+	int r;
+
+	pvmatch.condition = _is_same_pv;
+	pvmatch.areas = areas;
+	pvmatch.areas_size = areas_size;
+	pvmatch.pva = pva;
+
+	/* FIXME Cope with stacks by flattening */
+	if (!(r = _for_each_pv(cmd, prev_lvseg->lv,
+			       prev_lvseg->le + prev_lvseg->len - 1, 1, NULL,
+			       0, 0, -1, 1,
+			       _is_condition, &pvmatch)))
+		stack;
+
+	if (r != 2)
+		return 0;
+
+	return 1;
+}
+
 /*
  * Is pva contiguous to any existing areas or on the same PV?
  */
-static int _check_contiguous(struct lv_segment *prev_lvseg,
-			     struct physical_volume *pv, struct pv_area *pva,
-			     struct pv_area **areas)
+static int _check_contiguous(struct cmd_context *cmd,
+			     struct lv_segment *prev_lvseg, struct pv_area *pva,
+			     struct pv_area **areas, uint32_t areas_size)
 {
-	struct pv_segment *prev_pvseg;
-	uint32_t s;
+	struct pv_match pvmatch;
+	int r;
 
-	for (s = 0; s < prev_lvseg->area_count; s++) {
-		if (seg_type(prev_lvseg, s) != AREA_PV)
-			continue;	/* FIXME Broken */
+	pvmatch.condition = _is_contiguous;
+	pvmatch.areas = areas;
+	pvmatch.areas_size = areas_size;
+	pvmatch.pva = pva;
 
-		if (!(prev_pvseg = seg_pvseg(prev_lvseg, s)))
-			continue; /* FIXME Broken */
+	/* FIXME Cope with stacks by flattening */
+	if (!(r = _for_each_pv(cmd, prev_lvseg->lv,
+			       prev_lvseg->le + prev_lvseg->len - 1, 1, NULL,
+			       0, 0, -1, 1,
+			       _is_condition, &pvmatch)))
+		stack;
 
-		if ((prev_pvseg->pv != pv))
-			continue;
+	if (r != 2)
+		return 0;
 
-		if (prev_pvseg->pe + prev_pvseg->len == pva->start) {
-			areas[s] = pva;
-			return 1;
-		}
-	}
-
-	return 0;
+	return 1;
 }
 
 /*
@@ -679,9 +883,9 @@ static int _find_parallel_space(struct alloc_handle *ah, alloc_policy_t alloc,
 	struct pv_area *pva;
 	struct pv_list *pvl;
 	unsigned already_found_one = 0;
-	unsigned contiguous = 0, contiguous_count = 0;
+	unsigned contiguous = 0, cling = 0, preferred_count = 0;
 	unsigned ix;
-	unsigned ix_offset = 0;	/* Offset for non-contiguous allocations */
+	unsigned ix_offset = 0;	/* Offset for non-preferred allocations */
 	uint32_t max_parallel;	/* Maximum extents to allocate */
 	uint32_t next_le;
 	struct seg_pvs *spvs;
@@ -691,9 +895,14 @@ static int _find_parallel_space(struct alloc_handle *ah, alloc_policy_t alloc,
 	/* FIXME Select log PV appropriately if there isn't one yet */
 
 	/* Are there any preceding segments we must follow on from? */
-	if ((alloc == ALLOC_CONTIGUOUS) && prev_lvseg) {
-		contiguous = 1;
+	if (prev_lvseg) {
 		ix_offset = prev_lvseg->area_count;
+		if ((alloc == ALLOC_CONTIGUOUS))
+			contiguous = 1;
+		else if ((alloc == ALLOC_CLING))
+			cling = 1;
+		else
+			ix_offset = 0;
 	}
 
 	/* FIXME This algorithm needs a lot of cleaning up! */
@@ -702,6 +911,7 @@ static int _find_parallel_space(struct alloc_handle *ah, alloc_policy_t alloc,
 	/* ix holds the number of areas found on other PVs */
 	do {
 		ix = 0;
+		preferred_count = 0;
 
 		parallel_pvs = NULL;
 		max_parallel = needed;
@@ -711,14 +921,15 @@ static int _find_parallel_space(struct alloc_handle *ah, alloc_policy_t alloc,
 		 * the maximum we can allocate in one go accordingly.
 		 */
 		if (ah->parallel_areas) {
+			next_le = (prev_lvseg ? prev_lvseg->le + prev_lvseg->len : 0) + *allocated / ah->area_multiple;
 			list_iterate_items(spvs, ah->parallel_areas) {
-				next_le = (prev_lvseg ? prev_lvseg->le + prev_lvseg->len : 0) + *allocated;
-				if (next_le >= spvs->le) {
-					if (next_le + max_parallel > spvs->le + spvs->len)
-						max_parallel = (spvs->le + spvs->len - next_le) * ah->area_multiple;
-					parallel_pvs = &spvs->pvs;
-					break;
-				}
+				if (next_le >= spvs->le + spvs->len)
+					continue;
+
+				if (max_parallel > (spvs->le + spvs->len) * ah->area_multiple)
+					max_parallel = (spvs->le + spvs->len) * ah->area_multiple;
+				parallel_pvs = &spvs->pvs;
+				break;
 			}
 		}
 
@@ -750,17 +961,30 @@ static int _find_parallel_space(struct alloc_handle *ah, alloc_policy_t alloc,
 			list_iterate_items(pva, &pvm->areas) {
 				if (contiguous) {
 					if (prev_lvseg &&
-					    _check_contiguous(prev_lvseg,
-							      pvm->pv,
-							      pva, areas)) {
-						contiguous_count++;
+					    _check_contiguous(ah->cmd,
+							      prev_lvseg,
+							      pva, areas,
+							      areas_size)) {
+						preferred_count++;
 						goto next_pv;
 					}
 					continue;
 				}
 
+				if (cling) {
+					if (prev_lvseg &&
+					    _check_cling(ah->cmd,
+							   prev_lvseg,
+							   pva, areas,
+							   areas_size)) {
+						preferred_count++;
+					}
+					goto next_pv;
+				}
+
 				/* Is it big enough on its own? */
-				if ((pva->count < max_parallel - *allocated) &&
+				if (pva->count * ah->area_multiple <
+				    max_parallel - *allocated &&
 				    ((!can_split && !ah->log_count) ||
 				     (already_found_one &&
 				      !(alloc == ALLOC_ANYWHERE))))
@@ -781,7 +1005,7 @@ static int _find_parallel_space(struct alloc_handle *ah, alloc_policy_t alloc,
 				break;
 		}
 
-		if (contiguous && (contiguous_count < ix_offset))
+		if ((contiguous || cling) && (preferred_count < ix_offset))
 			break;
 
 		/* Only allocate log_area the first time around */
@@ -853,6 +1077,9 @@ static int _allocate(struct alloc_handle *ah,
 		return 0;
 	}
 
+	if (!_log_parallel_areas(ah->mem, ah->parallel_areas))
+		stack;
+
 	areas_size = list_size(pvms);
 	if (areas_size < ah->area_count + ah->log_count) {
 		if (ah->alloc != ALLOC_ANYWHERE) {
@@ -887,6 +1114,18 @@ static int _allocate(struct alloc_handle *ah,
 	    (!can_split && (allocated != old_allocated)))
 		goto finished;
 
+	old_allocated = allocated;
+	if (!_find_parallel_space(ah, ALLOC_CLING, pvms, areas,
+				  areas_size, can_split,
+				  prev_lvseg, &allocated, new_extents)) {
+		stack;
+		goto out;
+	}
+
+	if ((allocated == new_extents) || (ah->alloc == ALLOC_CLING) ||
+	    (!can_split && (allocated != old_allocated)))
+		goto finished;
+
 	old_allocated = allocated;
 	if (!_find_parallel_space(ah, ALLOC_NORMAL, pvms, areas,
 				  areas_size, can_split,
@@ -982,7 +1221,7 @@ struct alloc_handle *allocate_extents(struct volume_group *vg,
 	if (alloc == ALLOC_INHERIT)
 		alloc = vg->alloc;
 
-	if (!(ah = _alloc_init(vg->cmd->mem, segtype, alloc, mirrors,
+	if (!(ah = _alloc_init(vg->cmd, vg->cmd->mem, segtype, alloc, mirrors,
 			       stripes, log_count, mirrored_pv,
 			       mirrored_pe, parallel_areas))) {
 		stack;
@@ -1213,7 +1452,7 @@ int lv_extend(struct logical_volume *lv,
 				log_error("Aborting. Failed to extend %s.",
 					  seg_lv(seg, m)->name);
 				return 0;
-                	}
+			}
 		}
 		seg->area_len += extents;
 		seg->len += extents;
@@ -1240,7 +1479,7 @@ char *generate_lv_name(struct volume_group *vg, const char *format,
 			high = i;
 	}
 
-	if (lvm_snprintf(buffer, len, format, high + 1) < 0)
+	if (dm_snprintf(buffer, len, format, high + 1) < 0)
 		return NULL;
 
 	return buffer;
@@ -1326,58 +1565,16 @@ struct logical_volume *lv_create_empty(struct format_instance *fi,
 	return lv;
 }
 
-/* Recursively process each PV used by part of an LV */
-static int _for_each_pv(struct cmd_context *cmd, struct logical_volume *lv,
-			uint32_t le, uint32_t len,
-			int (*fn)(struct cmd_context *cmd, struct pv_segment *peg, struct seg_pvs *spvs),
-			struct seg_pvs *spvs)
-{
-	struct lv_segment *seg;
-	uint32_t s;
-	uint32_t remaining_seg_len, area_len, area_multiple;
-
-	if (!(seg = find_seg_by_le(lv, le))) {
-		log_error("Failed to find segment for %s extent %" PRIu32,
-			  lv->name, le);
-		return 0;
-	}
-
-	/* Remaining logical length of segment */
-	remaining_seg_len = seg->len - (le - seg->le);
-
-	if (len > remaining_seg_len)
-		remaining_seg_len = len;
-
-	if (spvs->len > remaining_seg_len)
-		spvs->len = remaining_seg_len;
-
-	area_multiple = segtype_is_striped(seg->segtype) ? seg->area_count : 1;
-	area_len = remaining_seg_len / area_multiple;
-
-	for (s = 0; s < seg->area_count; s++) {
-		if (seg_type(seg, s) == AREA_LV) {
-			if (!_for_each_pv(cmd, seg_lv(seg, s),
-					  seg_le(seg, s) + (le - seg->le) / area_multiple,
-					  area_len, fn, spvs)) {
-				stack;
-				return 0;
-			}
-		} else if (seg_type(seg, s) == AREA_PV) {
-			if (!fn(cmd, seg_pvseg(seg, s), spvs)) {
-				stack;
-				return 0;
-			}
-		}
-	}
-
-	return 1;
-}
-
-static int _add_pvs(struct cmd_context *cmd, struct pv_segment *peg, struct seg_pvs *spvs)
+static int _add_pvs(struct cmd_context *cmd, struct pv_segment *peg,
+		    uint32_t s __attribute((unused)), void *data)
 {
+	struct seg_pvs *spvs = (struct seg_pvs *) data;
 	struct pv_list *pvl;
 
-	/* FIXME Don't add again if it's already on the list! */
+	/* Don't add again if it's already on list. */
+	list_iterate_items(pvl, &spvs->pvs)
+		if (pvl->pv == peg->pv)
+			return 1;
 
 	if (!(pvl = dm_pool_alloc(cmd->mem, sizeof(*pvl)))) {
 		log_error("pv_list allocation failed");
@@ -1386,11 +1583,8 @@ static int _add_pvs(struct cmd_context *cmd, struct pv_segment *peg, struct seg_
 
 	pvl->pv = peg->pv;
 
-	/* FIXME Use ordered list to facilitate comparison */
 	list_add(&spvs->pvs, &pvl->list);
 
-	/* FIXME Add mirror logs, snapshot cow LVs etc. */
-
 	return 1;
 }
 
@@ -1426,7 +1620,8 @@ struct list *build_parallel_areas_from_lv(struct cmd_context *cmd,
 
 		/* Find next segment end */
 		/* FIXME Unnecessary nesting! */
-		if (!_for_each_pv(cmd, lv, current_le, lv->le_count, _add_pvs, spvs)) {
+		if (!_for_each_pv(cmd, lv, current_le, spvs->len, &spvs->len,
+				  0, 0, -1, 0, _add_pvs, (void *) spvs)) {
 			stack;
 			return NULL;
 		}

lib/metadata/metadata.c

@@ -24,6 +24,13 @@
 #include "pv_alloc.h"
 #include "activate.h"
 
+#include <sys/param.h>
+
+unsigned long pe_align(void)
+{
+	return MAX(65536UL, lvm_getpagesize()) >> SECTOR_SHIFT;
+}
+
 static int _add_pv_to_vg(struct format_instance *fid, struct volume_group *vg,
 			 const char *pv_name)
 {
@@ -78,15 +85,12 @@ static int _add_pv_to_vg(struct format_instance *fid, struct volume_group *vg,
 
 	/* FIXME Do proper rounding-up alignment? */
 	/* Reserved space for label; this holds 0 for PVs created by LVM1 */
-	if (pv->pe_start < PE_ALIGN)
-		pv->pe_start = PE_ALIGN;
+	if (pv->pe_start < pe_align())
+		pv->pe_start = pe_align();
 
 	/*
-	 * The next two fields should be corrected
-	 * by fid->pv_setup.
+	 * pe_count must always be calculated by pv_setup
 	 */
-	pv->pe_count = (pv->size - pv->pe_start) / vg->extent_size;
-
 	pv->pe_alloc_count = 0;
 
 	if (!fid->fmt->ops->pv_setup(fid->fmt, UINT64_C(0), 0,
@@ -727,6 +731,8 @@ int vg_validate(struct volume_group *vg)
 	char uuid[64];
 	int r = 1;
 
+	/* FIXME Also check there's no data/metadata overlap */
+
 	list_iterate_items(pvl, &vg->pvs) {
 		list_iterate_items(pvl2, &vg->pvs) {
 			if (pvl == pvl2)
@@ -1037,7 +1043,7 @@ static struct volume_group *_vg_read(struct cmd_context *cmd,
 	if (correct_vg) {
 		if (list_size(&correct_vg->pvs) != list_size(pvids)) {
 			log_debug("Cached VG %s had incorrect PV list",
-				  vg->name);
+				  vgname);
 
 			if (memlock())
 				inconsistent = 1;
@@ -1046,7 +1052,7 @@ static struct volume_group *_vg_read(struct cmd_context *cmd,
 		} else list_iterate_items(pvl, &correct_vg->pvs) {
 			if (!str_list_match_item(pvids, pvl->pv->dev->pvid)) {
 				log_debug("Cached VG %s had incorrect PV list",
-					  vg->name);
+					  vgname);
 				correct_vg = NULL;
 				break;
 			}

lib/metadata/metadata.h

@@ -29,12 +29,12 @@
 #define MAX_STRIPES 128U
 #define SECTOR_SHIFT 9L
 #define SECTOR_SIZE ( 1L << SECTOR_SHIFT )
-#define STRIPE_SIZE_MIN ( (unsigned) getpagesize() >> SECTOR_SHIFT)	/* PAGESIZE in sectors */
+#define STRIPE_SIZE_MIN ( (unsigned) lvm_getpagesize() >> SECTOR_SHIFT)	/* PAGESIZE in sectors */
 #define STRIPE_SIZE_MAX ( 512L * 1024L >> SECTOR_SHIFT)	/* 512 KB in sectors */
 #define STRIPE_SIZE_LIMIT ((UINT_MAX >> 2) + 1)
 #define PV_MIN_SIZE ( 512L * 1024L >> SECTOR_SHIFT)	/* 512 KB in sectors */
-#define PE_ALIGN (65536UL >> SECTOR_SHIFT)	/* PE alignment */
 #define MAX_RESTRICTED_LVS 255	/* Used by FMT_RESTRICTED_LVIDS */
+#define MIRROR_LOG_SIZE 1	/* Extents */
 
 /* Various flags */
 /* Note that the bits no longer necessarily correspond to LVM1 disk format */
@@ -82,6 +82,7 @@ typedef enum {
 	ALLOC_INVALID = 0,
 	ALLOC_INHERIT,
 	ALLOC_CONTIGUOUS,
+	ALLOC_CLING,
 	ALLOC_NORMAL,
 	ALLOC_ANYWHERE
 } alloc_policy_t;
@@ -403,6 +404,7 @@ struct format_handler {
 /*
  * Utility functions
  */
+unsigned long pe_align(void);
 int vg_validate(struct volume_group *vg);
 int vg_write(struct volume_group *vg);
 int vg_commit(struct volume_group *vg);

lib/metadata/mirror.c

@@ -161,9 +161,11 @@ int remove_mirror_images(struct lv_segment *mirrored_seg, uint32_t num_mirrors,
 		remove_log = 1;
 	}
 
-	if (remove_log) {
+	if (remove_log && mirrored_seg->log_lv) {
 		log_lv = mirrored_seg->log_lv;
 		mirrored_seg->log_lv = NULL;
+		log_lv->status &= ~MIRROR_LOG;
+		log_lv->status |= VISIBLE_LV;
 	}
 
 	/*
@@ -428,7 +430,7 @@ static int _create_layers_for_mirror(struct alloc_handle *ah,
 		return 0;
 	}
 
-	if (lvm_snprintf(img_name, len, "%s_mimage_%%d", lv->name) < 0) {
+	if (dm_snprintf(img_name, len, "%s_mimage_%%d", lv->name) < 0) {
 		log_error("img_name allocation failed. "
 			  "Remove new LV and retry.");
 		return 0;

lib/metadata/segtype.h

@@ -77,6 +77,9 @@ struct segtype_handler {
 			       uint64_t *total_numerator,
 			       uint64_t *total_denominator, float *percent);
 	int (*target_present) (void);
+	int (*modules_needed) (struct dm_pool *mem,
+			       const struct lv_segment *seg,
+			       struct list *modules);
 	void (*destroy) (const struct segment_type * segtype);
 	int (*target_register_events) (struct lv_segment *seg, int events);
 	int (*target_unregister_events) (struct lv_segment *seg, int events);

lib/mirror/mirrored.c

@@ -26,6 +26,7 @@
 #include "targets.h"
 #include "activate.h"
 #include "sharedlib.h"
+#include "str_list.h"
 
 #ifdef DMEVENTD
 #  include <libdevmapper-event.h>
@@ -447,6 +448,28 @@ static int _target_unregister_events(struct lv_segment *seg,
 #endif /* DMEVENTD */
 #endif /* DEVMAPPER_SUPPORT */
 
+static int _mirrored_modules_needed(struct dm_pool *mem,
+				    const struct lv_segment *seg,
+				    struct list *modules)
+{
+	if (seg->log_lv &&
+	    !list_segment_modules(mem, first_seg(seg->log_lv), modules))
+		return_0;
+
+	if ((seg->lv->vg->status & CLUSTERED) &&
+	    !str_list_add(mem, modules, "clog")) {
+		log_error("cluster log string list allocation failed");
+		return 0;
+	}
+
+	if (!str_list_add(mem, modules, "mirror")) {
+		log_error("mirror string list allocation failed");
+		return 0;
+	}
+
+	return 1;
+}
+
 static void _mirrored_destroy(const struct segment_type *segtype)
 {
 	dm_free((void *) segtype);
@@ -467,6 +490,7 @@ static struct segtype_handler _mirrored_ops = {
 	.target_unregister_events = _target_unregister_events,
 #endif
 #endif
+	.modules_needed = _mirrored_modules_needed,
 	.destroy = _mirrored_destroy,
 };
 

lib/misc/configure.h.in

@@ -174,6 +174,9 @@
 /* Define to 1 if you have the <selinux/selinux.h> header file. */
 #undef HAVE_SELINUX_SELINUX_H
 
+/* define to 1 to include support for realtime clock */
+#undef HAVE_REALTIME
+
 /* Define to 1 if you have the `setlocale' function. */
 #undef HAVE_SETLOCALE
 

lib/misc/lib.h

@@ -28,6 +28,7 @@
 #include "log.h"
 #include "intl.h"
 #include "lvm-types.h"
+#include "lvm-wrappers.h"
 
 #include <libdevmapper.h>
 

lib/misc/lvm-file.c

@@ -50,7 +50,7 @@ int create_temp_name(const char *dir, char *buffer, size_t len, int *fd)
 
 	for (i = 0; i < 20; i++, num++) {
 
-		if (lvm_snprintf(buffer, len, "%s/.lvm_%s_%d_%d",
+		if (dm_snprintf(buffer, len, "%s/.lvm_%s_%d_%d",
 				 dir, hostname, pid, num) == -1) {
 			log_err("Not enough space to build temporary file "
 				"string.");

lib/misc/lvm-string.c

@@ -14,33 +14,10 @@
  */
 
 #include "lib.h"
-#include "lvm-types.h"
 #include "lvm-string.h"
 
 #include <ctype.h>
 
-/*
- * On error, up to glibc 2.0.6, snprintf returned -1 if buffer was too small;
- * From glibc 2.1 it returns number of chars (excl. trailing null) that would 
- * have been written had there been room.
- *
- * lvm_snprintf reverts to the old behaviour.
- */
-int lvm_snprintf(char *buf, size_t bufsize, const char *format, ...)
-{
-	int n;
-	va_list ap;
-
-	va_start(ap, format);
-	n = vsnprintf(buf, bufsize, format, ap);
-	va_end(ap);
-
-	if (n < 0 || (n > bufsize - 1))
-		return -1;
-
-	return n;
-}
-
 int emit_to_buffer(char **buffer, size_t *size, const char *fmt, ...)
 {
 	int n;
@@ -58,47 +35,6 @@ int emit_to_buffer(char **buffer, size_t *size, const char *fmt, ...)
 	return 1;
 }
 
-/*
- * consume characters while they match the predicate function.
- */
-static char *_consume(char *buffer, int (*fn) (int))
-{
-	while (*buffer && fn(*buffer))
-		buffer++;
-
-	return buffer;
-}
-
-static int _isword(int c)
-{
-	return !isspace(c);
-}
-
-/*
- * Split buffer into NULL-separated words in argv.
- * Returns number of words.
- */
-int split_words(char *buffer, unsigned max, char **argv)
-{
-	unsigned arg;
-
-	for (arg = 0; arg < max; arg++) {
-		buffer = _consume(buffer, isspace);
-		if (!*buffer)
-			break;
-
-		argv[arg] = buffer;
-		buffer = _consume(buffer, _isword);
-
-		if (*buffer) {
-			*buffer = '\0';
-			buffer++;
-		}
-	}
-
-	return arg;
-}
-
 /*
  * Device layer names are all of the form <vg>-<lv>-<layer>, any
  * other hyphens that appear in these names are quoted with yet
@@ -169,47 +105,6 @@ char *build_dm_name(struct dm_pool *mem, const char *vgname,
 	return r;
 }
 
-/*
- * Remove hyphen quoting from a component of a name.
- * NULL-terminates the component and returns start of next component.
- */
-static char *_unquote(char *component)
-{
-	char *c = component;
-	char *o = c;
-	char *r;
-
-	while (*c) {
-		if (*(c + 1)) {
-			if (*c == '-') {
-				if (*(c + 1) == '-')
-					c++;
-				else
-					break;
-			}
-		}
-		*o = *c;
-		o++;
-		c++;
-	}
-
-	r = (*c) ? c + 1 : c;
-	*o = '\0';
-
-	return r;
-}
-
-int split_dm_name(struct dm_pool *mem, const char *dmname,
-		  char **vgname, char **lvname, char **layer)
-{
-	if (!(*vgname = dm_pool_strdup(mem, dmname)))
-		return 0;
-
-	_unquote(*layer = _unquote(*lvname = _unquote(*vgname)));
-
-	return 1;
-}
-
 int validate_name(const char *n)
 {
 	register char c;

lib/misc/lvm-string.h

@@ -23,25 +23,11 @@
 
 struct pool;
 
-/*
- * On error, up to glibc 2.0.6, snprintf returned -1 if buffer was too small;
- * From glibc 2.1 it returns number of chars (excl. trailing null) that would 
- * have been written had there been room.
- *
- * lvm_snprintf reverts to the old behaviour.
- */
-int lvm_snprintf(char *buf, size_t bufsize, const char *format, ...);
-
 int emit_to_buffer(char **buffer, size_t *size, const char *fmt, ...);
 
-int split_words(char *buffer, unsigned max, char **argv);
-
 char *build_dm_name(struct dm_pool *mem, const char *vg,
                     const char *lv, const char *layer);
 
-int split_dm_name(struct dm_pool *mem, const char *dmname,
-                  char **vgname, char **lvname, char **layer);
-
 int validate_name(const char *n);
 
 #endif

lib/misc/lvm-wrappers.c

@@ -0,0 +1,22 @@
+/*
+ * Copyright (C) 2006 Red Hat, Inc. All rights reserved.
+ *
+ * This file is part of LVM2.
+ *
+ * This copyrighted material is made available to anyone wishing to use,
+ * modify, copy, or redistribute it subject to the terms and conditions
+ * of the GNU General Public License v.2.
+ *
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include "lib.h"
+
+#include <unistd.h>
+
+int lvm_getpagesize(void)
+{
+	return getpagesize();
+}

lib/misc/lvm-wrappers.h

@@ -0,0 +1,21 @@
+/*
+ * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
+ * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+ *
+ * This file is part of LVM2.
+ *
+ * This copyrighted material is made available to anyone wishing to use,
+ * modify, copy, or redistribute it subject to the terms and conditions
+ * of the GNU General Public License v.2.
+ *
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifndef _LVM_WRAPPERS_H
+#define _LVM_WRAPPERS_H
+
+int lvm_getpagesize(void);
+
+#endif

lib/misc/sharedlib.c

@@ -17,6 +17,7 @@
 #include "config.h"
 #include "lvm-string.h"
 #include "sharedlib.h"
+#include "toolcontext.h"
 
 #include <limits.h>
 #include <sys/stat.h>
@@ -32,7 +33,7 @@ void get_shared_library_path(struct cmd_context *cmd, const char *libname,
 	 * if present */
 	if (libname[0] == '/' ||
 	    !(lib_dir = find_config_tree_str(cmd, "global/library_dir", 0)) ||
-	    (lvm_snprintf(path, path_len, "%s/%s", lib_dir,
+	    (dm_snprintf(path, path_len, "%s/%s", lib_dir,
 			  libname) == -1) || stat(path, &info) == -1)
 		strncpy(path, libname, path_len);
 }
@@ -43,6 +44,12 @@ void *load_shared_library(struct cmd_context *cmd, const char *libname,
 	char path[PATH_MAX];
 	void *library;
 
+	if (cmd->is_static) {
+		log_error("Not loading shared %s library %s in static mode.",
+			  desc, libname);
+		return NULL;
+	}
+
 	get_shared_library_path(cmd, libname, path, sizeof(path));
 
 	log_very_verbose("Opening shared %s library %s", desc, path);

lib/misc/timestamp.c

@@ -0,0 +1,130 @@
+/*
+ * Copyright (C) 2006 Rackable Systems All rights reserved.
+ *
+ * This file is part of LVM2.
+ *
+ * This copyrighted material is made available to anyone wishing to use,
+ * modify, copy, or redistribute it subject to the terms and conditions
+ * of the GNU General Public License v.2.
+ *
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+/*
+ * Abstract out the time methods used so they can be adjusted later -
+ * the results of these routines should stay in-core.  This implementation
+ * requires librt.
+ */
+
+#include "lib.h"
+#include <stdlib.h>
+
+#include "timestamp.h"
+
+/*
+ * The realtime section uses clock_gettime with the CLOCK_MONOTONIC
+ * parameter to prevent issues with time warps
+ */
+#ifdef HAVE_REALTIME
+
+#include <time.h>
+#include <bits/time.h>
+
+struct timestamp {
+	struct timespec t;
+};
+
+struct timestamp *get_timestamp(void)
+{
+	struct timestamp *ts = NULL;
+
+	if (!(ts = dm_malloc(sizeof(*ts))))
+		return_NULL;
+
+	if (clock_gettime(CLOCK_MONOTONIC, &ts->t)) {
+		log_sys_error("clock_gettime", "get_timestamp");
+		return NULL;
+	}
+
+	return ts;
+}
+
+/* cmp_timestamp: Compare two timestamps
+ *
+ * Return: -1 if t1 is less than t2
+ *          0 if t1 is equal to t2
+ *          1 if t1 is greater than t2
+ */
+int cmp_timestamp(struct timestamp *t1, struct timestamp *t2)
+{
+	if(t1->t.tv_sec < t2->t.tv_sec)
+		return -1;
+	if(t1->t.tv_sec > t2->t.tv_sec)
+		return 1;
+
+	if(t1->t.tv_nsec < t2->t.tv_nsec)
+		return -1;
+	if(t1->t.tv_nsec > t2->t.tv_nsec)
+		return 1;
+
+	return 0;
+}
+
+#else /* ! HAVE_REALTIME */
+
+/*
+ * The !realtime section just uses gettimeofday and is therefore subject
+ * to ntp-type time warps - not sure if should allow that.
+ */
+
+#include <sys/time.h>
+
+struct timestamp {
+	struct timeval t;
+};
+
+struct timestamp *get_timestamp(void)
+{
+	struct timestamp *ts = NULL;
+
+	if (!(ts = dm_malloc(sizeof(*ts))))
+		return_NULL;
+
+	if (gettimeofday(&ts->t, NULL)) {
+		log_sys_error("gettimeofday", "get_timestamp");
+		return NULL;
+	}
+
+	return ts;
+}
+
+/* cmp_timestamp: Compare two timestamps
+ *
+ * Return: -1 if t1 is less than t2
+ *          0 if t1 is equal to t2
+ *          1 if t1 is greater than t2
+ */
+int cmp_timestamp(struct timestamp *t1, struct timestamp *t2)
+{
+	if(t1->t.tv_sec < t2->t.tv_sec)
+		return -1;
+	if(t1->t.tv_sec > t2->t.tv_sec)
+		return 1;
+
+	if(t1->t.tv_usec < t2->t.tv_usec)
+		return -1;
+	if(t1->t.tv_usec > t2->t.tv_usec)
+		return 1;
+
+	return 0;
+}
+
+#endif /* HAVE_REALTIME */
+
+void destroy_timestamp(struct timestamp *t)
+{
+	if (t)
+		dm_free(t);
+}

lib/misc/timestamp.h

@@ -0,0 +1,33 @@
+/*
+ * Copyright (C) 2006 Rackable Systems All rights reserved.  
+ *
+ * This file is part of LVM2.
+ *
+ * This copyrighted material is made available to anyone wishing to use,
+ * modify, copy, or redistribute it subject to the terms and conditions
+ * of the GNU General Public License v.2.
+ *
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#ifndef _LVM_TIMESTAMP_H
+#define _LVM_TIMESTAMP_H
+
+struct timestamp;
+
+struct timestamp *get_timestamp(void);
+
+/* cmp_timestamp: Compare two timestamps
+ * 
+ * Return: -1 if t1 is less than t2
+ *  	    0 if t1 is equal to t2
+ *          1 if t1 is greater than t2
+ */
+int cmp_timestamp(struct timestamp *t1, struct timestamp *t2);
+
+void destroy_timestamp(struct timestamp *t);
+
+#endif /* _LVM_TIMESTAMP_H */
+

lib/mm/memlock.c

@@ -58,7 +58,7 @@ static int _default_priority;
 
 static void _touch_memory(void *mem, size_t size)
 {
-	size_t pagesize = getpagesize();
+	size_t pagesize = lvm_getpagesize();
 	void *pos = mem;
 	void *end = mem + size - sizeof(long);
 

lib/report/columns.h

@@ -33,6 +33,7 @@ FIELD(LVS, lv, NUM, "Copy%", lvid, 6, copypercent, "copy_percent")
 FIELD(LVS, lv, STR, "Move", lvid, 4, movepv, "move_pv")
 FIELD(LVS, lv, STR, "LV Tags", tags, 7, tags, "lv_tags")
 FIELD(LVS, lv, STR, "Log", lvid, 3, loglv, "mirror_log")
+FIELD(LVS, lv, STR, "Modules", lvid, 7, modules, "modules")
 
 FIELD(PVS, pv, STR, "Fmt", id, 3, pvfmt, "pv_fmt")
 FIELD(PVS, pv, STR, "PV UUID", id, 38, uuid, "pv_uuid")
@@ -68,8 +69,11 @@ FIELD(VGS, vg, STR, "VG Tags", tags, 7, tags, "vg_tags")
 FIELD(SEGS, seg, STR, "Type", list, 4, segtype, "segtype")
 FIELD(SEGS, seg, NUM, "#Str", area_count, 4, uint32, "stripes")
 FIELD(SEGS, seg, NUM, "Stripe", stripe_size, 6, size32, "stripesize")
+FIELD(SEGS, seg, NUM, "Stripe", stripe_size, 6, size32, "stripe_size")
 FIELD(SEGS, seg, NUM, "Chunk", list, 5, chunksize, "chunksize")
+FIELD(SEGS, seg, NUM, "Chunk", list, 5, chunksize, "chunk_size")
 FIELD(SEGS, seg, NUM, "Region", region_size, 6, size32, "regionsize")
+FIELD(SEGS, seg, NUM, "Region", region_size, 6, size32, "region_size")
 FIELD(SEGS, seg, NUM, "Start", list, 5, segstart, "seg_start")
 FIELD(SEGS, seg, NUM, "SSize", list, 5, segsize, "seg_size")
 FIELD(SEGS, seg, STR, "Seg Tags", tags, 8, tags, "seg_tags")

lib/report/report.c

@@ -21,6 +21,7 @@
 #include "display.h"
 #include "activate.h"
 #include "segtype.h"
+#include "str_list.h"
 
 /* 
  * For macro use
@@ -103,6 +104,8 @@ static char _alloc_policy_char(alloc_policy_t alloc)
 	switch (alloc) {
 	case ALLOC_CONTIGUOUS:
 		return 'c';
+	case ALLOC_CLING:
+		return 'C';
 	case ALLOC_NORMAL:
 		return 'n';
 	case ALLOC_ANYWHERE:
@@ -172,9 +175,9 @@ static int _devices_disp(struct report_handle *rh, struct field *field,
 			return 0;
 		}
 
-		if (lvm_snprintf(extent_str, sizeof(extent_str), "(%" PRIu32
+		if (dm_snprintf(extent_str, sizeof(extent_str), "(%" PRIu32
 				 ")", extent) < 0) {
-			log_error("Extent number lvm_snprintf failed");
+			log_error("Extent number dm_snprintf failed");
 			return 0;
 		}
 
@@ -200,6 +203,7 @@ static int _devices_disp(struct report_handle *rh, struct field *field,
 
 	return 1;
 }
+
 static int _tags_disp(struct report_handle *rh, struct field *field,
 		      const void *data)
 {
@@ -230,6 +234,23 @@ static int _tags_disp(struct report_handle *rh, struct field *field,
 	return 1;
 }
 
+static int _modules_disp(struct report_handle *rh, struct field *field,
+			 const void *data)
+{
+	const struct logical_volume *lv = (const struct logical_volume *) data;
+	struct list *modules;
+
+	if (!(modules = str_list_create(rh->mem))) {
+		log_error("modules str_list allocation failed");
+		return 0;
+	}
+
+	if (!list_lv_modules(rh->mem, lv, modules))
+		return_0;
+
+	return _tags_disp(rh, field, modules);
+}
+
 static int _vgfmt_disp(struct report_handle *rh, struct field *field,
 		       const void *data)
 {
@@ -276,7 +297,7 @@ static int _int_disp(struct report_handle *rh, struct field *field,
 		return 0;
 	}
 
-	if (lvm_snprintf(repstr, 12, "%d", value) < 0) {
+	if (dm_snprintf(repstr, 12, "%d", value) < 0) {
 		log_error("int too big: %d", value);
 		return 0;
 	}
@@ -540,7 +561,7 @@ static int _lvname_disp(struct report_handle *rh, struct field *field,
 		return 0;
 	}
 
-	if (lvm_snprintf(repstr, len, "[%s]", lv->name) < 0) {
+	if (dm_snprintf(repstr, len, "[%s]", lv->name) < 0) {
 		log_error("lvname snprintf failed");
 		return 0;
 	}
@@ -784,7 +805,7 @@ static int _uint32_disp(struct report_handle *rh, struct field *field,
 		return 0;
 	}
 
-	if (lvm_snprintf(repstr, 11, "%u", value) < 0) {
+	if (dm_snprintf(repstr, 11, "%u", value) < 0) {
 		log_error("uint32 too big: %u", value);
 		return 0;
 	}
@@ -813,7 +834,7 @@ static int _int32_disp(struct report_handle *rh, struct field *field,
 		return 0;
 	}
 
-	if (lvm_snprintf(repstr, 12, "%d", value) < 0) {
+	if (dm_snprintf(repstr, 12, "%d", value) < 0) {
 		log_error("int32 too big: %d", value);
 		return 0;
 	}
@@ -870,7 +891,7 @@ static int _snpercent_disp(struct report_handle *rh, struct field *field,
 		return 0;
 	}
 
-	if (lvm_snprintf(repstr, 7, "%.2f", snap_percent) < 0) {
+	if (dm_snprintf(repstr, 7, "%.2f", snap_percent) < 0) {
 		log_error("snapshot percentage too large");
 		return 0;
 	}
@@ -910,7 +931,7 @@ static int _copypercent_disp(struct report_handle *rh, struct field *field,
 		return 0;
 	}
 
-	if (lvm_snprintf(repstr, 7, "%.2f", percent) < 0) {
+	if (dm_snprintf(repstr, 7, "%.2f", percent) < 0) {
 		log_error("copy percentage too large");
 		return 0;
 	}
@@ -948,6 +969,44 @@ static struct {
 
 const unsigned int _num_fields = sizeof(_fields) / sizeof(_fields[0]);
 
+static void _display_fields(void)
+{
+	uint32_t f;
+	const char *type, *last_type = "";
+
+	for (f = 0; f < _num_fields; f++) {
+		switch (_fields[f].type) {
+		case PVS:
+			type = "Physical Volume";
+			break;
+		case LVS:
+			type = "Logical Volume";
+			break;
+		case VGS:
+			type = "Volume Group";
+			break;
+		case SEGS:
+			type = "Logical Volume Segment";
+			break;
+		case PVSEGS:
+			type = "Physical Volume Segment";
+			break;
+		default:
+			type = " ";
+		}
+
+		if (type != last_type) {
+			if (*last_type)
+				log_print(" ");
+			log_print("%s Fields", type);
+		}
+
+		log_print("- %s", _fields[f].id);
+
+		last_type = type;
+	}
+}
+
 /*
  * Initialise report handle
  */
@@ -1080,6 +1139,8 @@ static int _parse_options(struct report_handle *rh, const char *format)
 		while (*we && *we != ',')
 			we++;
 		if (!_field_match(rh, ws, (size_t) (we - ws))) {
+			_display_fields();
+			log_print(" ");
 			log_error("Unrecognised field: %.*s", (int) (we - ws),
 				  ws);
 			return 0;
@@ -1324,7 +1385,7 @@ static int _report_headings(void *handle)
 
 		heading = _fields[fp->field_num].heading;
 		if (rh->flags & RH_ALIGNED) {
-			if (lvm_snprintf(buf, sizeof(buf), "%-*.*s",
+			if (dm_snprintf(buf, sizeof(buf), "%-*.*s",
 					 fp->width, fp->width, heading) < 0) {
 				log_error("snprintf heading failed");
 				dm_pool_end_object(rh->mem);
@@ -1467,7 +1528,7 @@ int report_output(void *handle)
 						      strlen(repstr)))
 					goto bad;
 			} else if (field->props->flags & FLD_ALIGN_LEFT) {
-				if (lvm_snprintf(buf, sizeof(buf), "%-*.*s",
+				if (dm_snprintf(buf, sizeof(buf), "%-*.*s",
 						 width, width, repstr) < 0) {
 					log_error("snprintf repstr failed");
 					dm_pool_end_object(rh->mem);
@@ -1476,7 +1537,7 @@ int report_output(void *handle)
 				if (!dm_pool_grow_object(rh->mem, buf, width))
 					goto bad;
 			} else if (field->props->flags & FLD_ALIGN_RIGHT) {
-				if (lvm_snprintf(buf, sizeof(buf), "%*.*s",
+				if (dm_snprintf(buf, sizeof(buf), "%*.*s",
 						 width, width, repstr) < 0) {
 					log_error("snprintf repstr failed");
 					dm_pool_end_object(rh->mem);

lib/snapshot/snapshot.c

@@ -20,6 +20,7 @@
 #include "text_export.h"
 #include "config.h"
 #include "activate.h"
+#include "str_list.h"
 
 static const char *_snap_name(const struct lv_segment *seg)
 {
@@ -126,6 +127,18 @@ static int _snap_target_present(void)
 }
 #endif
 
+static int _snap_modules_needed(struct dm_pool *mem,
+				const struct lv_segment *seg,
+				struct list *modules)
+{
+	if (!str_list_add(mem, modules, "snapshot")) {
+		log_error("snapshot string list allocation failed");
+		return 0;
+	}
+
+	return 1;
+}
+
 static void _snap_destroy(const struct segment_type *segtype)
 {
 	dm_free((void *)segtype);
@@ -139,6 +152,7 @@ static struct segtype_handler _snapshot_ops = {
 	.target_percent = _snap_target_percent,
 	.target_present = _snap_target_present,
 #endif
+	.modules_needed = _snap_modules_needed,
 	.destroy = _snap_destroy,
 };
 

lib/zero/zero.c

@@ -63,6 +63,18 @@ static int _zero_target_present(void)
 }
 #endif
 
+static int _zero_modules_needed(struct dm_pool *mem,
+				const struct lv_segment *seg,
+				struct list *modules)
+{
+	if (!str_list_add(mem, modules, "zero")) {
+		log_error("zero module string list allocation failed");
+		return 0;
+	}
+
+	return 1;
+}
+
 static void _zero_destroy(const struct segment_type *segtype)
 {
 	dm_free((void *) segtype);
@@ -75,6 +87,7 @@ static struct segtype_handler _zero_ops = {
 	.add_target_line = _zero_add_target_line,
 	.target_present = _zero_target_present,
 #endif
+	.modules_needed = _zero_modules_needed,
 	.destroy = _zero_destroy,
 };
 

libdm/.exported_symbols

@@ -28,6 +28,7 @@ dm_task_set_gid
 dm_task_set_mode
 dm_task_suppress_identical_reload
 dm_task_add_target
+dm_task_no_flush
 dm_task_no_open_count
 dm_task_skip_lockfs
 dm_task_update_nodes
@@ -109,3 +110,6 @@ dm_hash_get_first
 dm_hash_get_next
 dm_set_selinux_context
 dm_task_set_geometry
+dm_split_lvm_name
+dm_split_words
+dm_snprintf

libdm/Makefile.in

@@ -23,6 +23,7 @@ SOURCES =\
 	libdm-common.c \
 	libdm-file.c \
 	libdm-deptree.c \
+	libdm-string.c \
 	mm/dbg_malloc.c \
 	mm/pool.c \
 	$(interface)/libdm-iface.c

libdm/ioctl/libdm-iface.c

@@ -1026,6 +1026,13 @@ int dm_task_set_geometry(struct dm_task *dmt, const char *cylinders, const char
 	return 1;
 }
 
+int dm_task_no_flush(struct dm_task *dmt)
+{
+	dmt->no_flush = 1;
+
+	return 1;
+}
+
 int dm_task_no_open_count(struct dm_task *dmt)
 {
 	dmt->no_open_count = 1;
@@ -1270,6 +1277,8 @@ static struct dm_ioctl *_flatten(struct dm_task *dmt, unsigned repeat_count)
 
 	if (dmt->type == DM_DEVICE_SUSPEND)
 		dmi->flags |= DM_SUSPEND_FLAG;
+	if (dmt->no_flush)
+		dmi->flags |= DM_NOFLUSH_FLAG;
 	if (dmt->read_only)
 		dmi->flags |= DM_READONLY_FLAG;
 	if (dmt->skip_lockfs)
@@ -1543,7 +1552,7 @@ static struct dm_ioctl *_do_dm_ioctl(struct dm_task *dmt, unsigned command,
 		dmi->flags |= DM_SKIP_BDGET_FLAG;
 
 	log_debug("dm %s %s %s%s%s %s%.0d%s%.0d%s"
-		  "%s%c%s %.0llu %s [%u]",
+		  "%s%c%c%s %.0llu %s [%u]",
 		  _cmd_data_v4[dmt->type].name,
 		  dmi->name, dmi->uuid, dmt->newname ? " " : "",
 		  dmt->newname ? dmt->newname : "",
@@ -1554,6 +1563,7 @@ static struct dm_ioctl *_do_dm_ioctl(struct dm_task *dmt, unsigned command,
 		  dmt->major > 0 && dmt->minor == 0 ? "0" : "",
 		  dmt->major > 0 ? ") " : "",
 		  dmt->no_open_count ? 'N' : 'O',
+		  dmt->no_flush ? 'N' : 'F',
 		  dmt->skip_lockfs ? "S " : "",
 		  dmt->sector, dmt->message ? dmt->message : "",
 		  dmi->data_size);

libdm/ioctl/libdm-targets.h

@@ -52,6 +52,7 @@ struct dm_task {
 	char *message;
 	char *geometry;
 	uint64_t sector;
+	int no_flush;
 	int no_open_count;
 	int skip_lockfs;
 	int suppress_identical_reload;

libdm/libdevmapper.h

@@ -17,6 +17,7 @@
 #define LIB_DEVICE_MAPPER_H
 
 #include <inttypes.h>
+#include <stdarg.h>
 #include <sys/types.h>
 
 #ifdef linux
@@ -150,6 +151,7 @@ int dm_task_set_event_nr(struct dm_task *dmt, uint32_t event_nr);
 int dm_task_set_geometry(struct dm_task *dmt, const char *cylinders, const char *heads, const char *sectors, const char *start);
 int dm_task_set_message(struct dm_task *dmt, const char *message);
 int dm_task_set_sector(struct dm_task *dmt, uint64_t sector);
+int dm_task_no_flush(struct dm_task *dmt);
 int dm_task_no_open_count(struct dm_task *dmt);
 int dm_task_skip_lockfs(struct dm_task *dmt);
 int dm_task_suppress_identical_reload(struct dm_task *dmt);
@@ -582,4 +584,28 @@ struct dm_hash_node *dm_hash_get_next(struct dm_hash_table *t, struct dm_hash_no
  *********/
 int dm_set_selinux_context(const char *path, mode_t mode);
 
+/*********************
+ * string manipulation
+ *********************/
+
+/*
+ * Break up the name of a mapped device into its constituent
+ * Volume Group, Logical Volume and Layer (if present).
+ */
+int dm_split_lvm_name(struct dm_pool *mem, const char *dmname,
+		      char **vgname, char **lvname, char **layer);
+
+/*
+ * Destructively split buffer into NULL-separated words in argv.
+ * Returns number of words.
+ */
+int dm_split_words(char *buffer, unsigned max,
+		   unsigned ignore_comments, /* Not implemented */
+		   char **argv);
+
+/* 
+ * Returns -1 if buffer too small
+ */
+int dm_snprintf(char *buf, size_t bufsize, const char *format, ...);
+
 #endif				/* LIB_DEVICE_MAPPER_H */

libdm/libdm-deptree.c

@@ -1028,7 +1028,7 @@ int dm_tree_suspend_children(struct dm_tree_node *dnode,
 			continue;
 
 		if (!_info_by_dev(dinfo->major, dinfo->minor, 0, &info) ||
-		    !info.exists)
+		    !info.exists || info.suspended)
 			continue;
 
 		if (!_suspend_node(name, info.major, info.minor,

libdm/libdm-string.c

@@ -0,0 +1,123 @@
+/*
+ * Copyright (C) 2006 Red Hat, Inc. All rights reserved.
+ *
+ * This file is part of the device-mapper userspace tools.
+ *
+ * This copyrighted material is made available to anyone wishing to use,
+ * modify, copy, or redistribute it subject to the terms and conditions
+ * of the GNU General Public License v.2.
+ *
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include "lib.h"
+#include "libdevmapper.h"
+
+#include <ctype.h>
+
+/*
+ * consume characters while they match the predicate function.
+ */
+static char *_consume(char *buffer, int (*fn) (int))
+{
+	while (*buffer && fn(*buffer))
+		buffer++;
+
+	return buffer;
+}
+
+static int _isword(int c)
+{
+	return !isspace(c);
+}
+
+/*
+ * Split buffer into NULL-separated words in argv.
+ * Returns number of words.
+ */
+int dm_split_words(char *buffer, unsigned max, unsigned ignore_comments,
+		   char **argv)
+{
+	unsigned arg;
+
+	for (arg = 0; arg < max; arg++) {
+		buffer = _consume(buffer, isspace);
+		if (!*buffer)
+			break;
+
+		argv[arg] = buffer;
+		buffer = _consume(buffer, _isword);
+
+		if (*buffer) {
+			*buffer = '\0';
+			buffer++;
+		}
+	}
+
+	return arg;
+}
+
+/*
+ * Remove hyphen quoting from a component of a name.
+ * NULL-terminates the component and returns start of next component.
+ */
+static char *_unquote(char *component)
+{
+	char *c = component;
+	char *o = c;
+	char *r;
+
+	while (*c) {
+		if (*(c + 1)) {
+			if (*c == '-') {
+				if (*(c + 1) == '-')
+					c++;
+				else
+					break;
+			}
+		}
+		*o = *c;
+		o++;
+		c++;
+	}
+
+	r = (*c) ? c + 1 : c;
+	*o = '\0';
+
+	return r;
+}
+
+int dm_split_lvm_name(struct dm_pool *mem, const char *dmname,
+		      char **vgname, char **lvname, char **layer)
+{
+	if (!(*vgname = dm_pool_strdup(mem, dmname)))
+		return 0;
+
+	_unquote(*layer = _unquote(*lvname = _unquote(*vgname)));
+
+	return 1;
+}
+
+/*
+ * On error, up to glibc 2.0.6, snprintf returned -1 if buffer was too small;
+ * From glibc 2.1 it returns number of chars (excl. trailing null) that would 
+ * have been written had there been room.
+ *
+ * dm_snprintf reverts to the old behaviour.
+ */
+int dm_snprintf(char *buf, size_t bufsize, const char *format, ...)
+{
+	int n;
+	va_list ap;
+
+	va_start(ap, format);
+	n = vsnprintf(buf, bufsize, format, ap);
+	va_end(ap);
+
+	if (n < 0 || (n > bufsize - 1))
+		return -1;
+
+	return n;
+}

libdm/mm/dbg_malloc.c

@@ -77,15 +77,6 @@ void *dm_malloc_aux_debug(size_t s, const char *file, int line)
 	nb->length = s;
 	nb->id = ++_mem_stats.block_serialno;
 	nb->next = 0;
-	nb->prev = _tail;
-
-	/* link to tail of the list */
-	if (!_head)
-		_head = _tail = nb;
-	else {
-		_tail->next = nb;
-		_tail = nb;
-	}
 
 	/* stomp a pretty pattern across the new memory
 	   and fill in the boundary bytes */
@@ -99,6 +90,16 @@ void *dm_malloc_aux_debug(size_t s, const char *file, int line)
 			*ptr++ = (char) nb->id;
 	}
 
+	nb->prev = _tail;
+
+	/* link to tail of the list */
+	if (!_head)
+		_head = _tail = nb;
+	else {
+		_tail->next = nb;
+		_tail = nb;
+	}
+
 	_mem_stats.blocks_allocated++;
 	if (_mem_stats.blocks_allocated > _mem_stats.blocks_max)
 		_mem_stats.blocks_max = _mem_stats.blocks_allocated;
@@ -134,12 +135,6 @@ void dm_free_aux(void *p)
 
 	/* have we freed this before ? */
 	assert(mb->id != 0);
-	mb->id = 0;
-
-	/* stomp a different pattern across the memory */
-	ptr = ((char *) mb) + sizeof(struct memblock);
-	for (i = 0; i < mb->length; i++)
-		*ptr++ = i & 1 ? (char) 0xde : (char) 0xad;
 
 	/* unlink */
 	if (mb->prev)
@@ -152,6 +147,13 @@ void dm_free_aux(void *p)
 	else
 		_tail = mb->prev;
 
+	mb->id = 0;
+
+	/* stomp a different pattern across the memory */
+	ptr = ((char *) mb) + sizeof(struct memblock);
+	for (i = 0; i < mb->length; i++)
+		*ptr++ = i & 1 ? (char) 0xde : (char) 0xad;
+
 	assert(_mem_stats.blocks_allocated);
 	_mem_stats.blocks_allocated--;
 	_mem_stats.bytes -= mb->length;

man/Makefile.in

@@ -17,7 +17,8 @@ top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
 MAN5=lvm.conf.5
-MAN8=lvchange.8 lvcreate.8 lvdisplay.8 lvextend.8 lvm.8 lvmchange.8 \
+MAN8=lvchange.8 lvconvert.8 lvcreate.8 lvdisplay.8 lvextend.8 lvm.8 \
+	lvmchange.8 \
 	lvmdiskscan.8 lvreduce.8 lvremove.8 lvrename.8 lvresize.8 lvs.8 \
 	lvscan.8 pvchange.8 pvcreate.8 pvdisplay.8 pvmove.8 pvremove.8 \
 	pvresize.8 pvs.8 pvscan.8 vgcfgbackup.8 vgcfgrestore.8 vgchange.8 \

man/clvmd.8

@@ -4,6 +4,7 @@ clvmd \- cluster LVM daemon
 .SH SYNOPSIS
 .B clvmd
 [\-d] [\-h]
+[\-R]
 [\-t <timeout>]
 [\-V]
 .SH DESCRIPTION
@@ -22,6 +23,11 @@ be so small that commands with many disk updates to do will fail, so you
 may need to increase this on systems with very large disk farms. 
 The default is 30 seconds.
 .TP
+.I \-R
+Tells all the running clvmd in the cluster to reload their device cache and
+re-read the lvm configuration file. This command should be run whenever the
+devices on a cluster system are changed.
+.TP
 .I \-V
 Display the version of the cluster LVM daemon.
 .SH SEE ALSO

man/dmsetup.8.in

@@ -13,7 +13,7 @@ dmsetup \- low level logical volume management
 .I [-f|--force]
 .br
 .B dmsetup suspend
-.I [--nolockfs] device_name
+.I [--nolockfs] [--noflush] device_name
 .br
 .B dmsetup resume
 .I device_name
@@ -213,7 +213,7 @@ Outputs status information for each of the device's targets.
 With --target, only information relating to the specified target type
 is displayed.
 .IP \fBsuspend
-.I [--nolockfs]
+.I [--nolockfs] [--noflush]
 .I device_name
 .br
 Suspends a device.  Any I/O that has already been mapped by the device
@@ -221,6 +221,9 @@ but has not yet completed will be flushed.  Any further I/O to that
 device will be postponed for as long as the device is suspended.
 If there's a filesystem on the device which supports the operation, 
 an attempt will be made to sync it first unless --nolockfs is specified.
+Some targets such as recent (October 2006) versions of multipath may support
+the --noflush option.  This lets outstanding I/O that has not yet reached the
+device to remain unflushed.
 .IP \fBtable
 .I [--target target_type]
 .I [device_name]

man/lvchange.8

@@ -9,6 +9,7 @@ lvchange \- change attributes of a logical volume
 [\-C/\-\-contiguous y/n] [\-d/\-\-debug] [\-\-deltag Tag]
 [\-h/\-?/\-\-help]
 [\-\-ignorelockingfailure]
+[\-\-monitor {y|n}]
 [\-M/\-\-persistent y/n] [\-\-minor minor]
 [\-P/\-\-partial y/n]
 [\-p/\-\-permission r/w] [\-r/\-\-readahead ReadAheadSectors]
@@ -42,6 +43,14 @@ allocated physical extents are already contiguous.
 .I \-\-minor minor
 Set the minor number.
 .TP
+.I \-\-monitor y/n
+Controls whether or not a mirrored logical volume is monitored by
+dmeventd, if it is installed.
+If a device used by a monitored mirror reports an I/O error,
+the failure is handled according to 
+\fBmirror_image_fault_policy\fP and \fBmirror_log_fault_policy\fP
+set in \fBlvm.conf\fP.
+.TP
 .I \-M, \-\-persistent y/n
 Set to y to make the minor number specified persistent.
 .TP

man/lvconvert.8

@@ -0,0 +1,61 @@
+.TH LVCONVERT 8 "LVM TOOLS" "Red Hat, Inc" \" -*- nroff -*-
+.SH NAME
+lvconvert \- convert a logical volume between linear and mirror
+.SH SYNOPSIS
+.B lvconvert
+[\-m/\-\-mirrors Mirrors [\-\-corelog] [\-R/\-\-regionsize MirrorLogRegionSize]]
+[\-A/\-\-alloc AllocationPolicy]
+[\-h/\-?/\-\-help]
+[\-v/\-\-verbose]
+[\-\-version]
+LogicalVolume[Path] [PhysicalVolume[Path]...]
+.SH DESCRIPTION
+lvconvert will change a linear logical volume to a mirror
+logical volume or vis versa.  It is also used to add and
+remove disk logs from mirror devices.
+.SH OPTIONS
+See \fBlvm\fP for common options.
+.TP
+.I \-m, \-\-mirrors Mirrors
+Specifies the degree of the mirror you wish to create.
+For example, "-m 1" would convert the original logical
+volume to a mirror volume with 2-sides; that is, a
+linear volume plus one copy.
+.TP
+.I \-\-corelog
+This optional argument tells lvconvert to switch the
+mirror from using a disk-based (persistent) log to
+an in-memory log.  You may only specify this option
+when the \-\-mirror argument is the same degree of
+the mirror you are changing.
+.TP
+.I \-R, \-\-regionsize MirrorLogRegionSize
+A mirror is divided into regions of this size (in MB), and the mirror log
+uses this granularity to track which regions are in sync.
+.SH Examples
+"lvconvert -m1 vg00/lvol1"
+.br
+converts the linear logical volume "vg00/lvol1" to
+a mirror logical volume.  This command could also
+be used to convert a two-way mirror with an
+in-memory log to a two-way mirror with a disk log.
+
+"lvconvert -m1 --corelog vg00/lvol1"
+.br
+converts a two-way mirror with a disk log to a
+two-way mirror with an in-memory log.
+
+"lvconvert -m0 vg00/lvol1"
+.br
+converts a mirror logical volume to a linear logical
+volume.
+
+.SH SEE ALSO
+.BR lvm (8),
+.BR vgcreate (8),
+.BR lvremove (8),
+.BR lvrename (8),
+.BR lvextend (8),
+.BR lvreduce (8),
+.BR lvdisplay (8),
+.BR lvscan (8)

man/lvcreate.8

@@ -8,12 +8,13 @@ lvcreate \- create a logical volume in an existing volume group
 [\-A/\-\-autobackup y/n] [\-C/\-\-contiguous y/n] [\-d/\-\-debug]
 [\-h/\-?/\-\-help]
 [\-i/\-\-stripes Stripes [\-I/\-\-stripesize StripeSize]]
-{\-l/\-\-extents LogicalExtentsNumber |
+{\-l/\-\-extents LogicalExtentsNumber[%{VG|FREE}] |
  \-L/\-\-size LogicalVolumeSize[kKmMgGtT]}
 [\-M/\-\-persistent y/n] [\-\-minor minor]
+[\-m/\-\-mirrors Mirrors [\-\-nosync] [\-\-corelog]
+[\-R/\-\-regionsize MirrorLogRegionSize]]
 [\-n/\-\-name LogicalVolumeName]
 [\-p/\-\-permission r/rw] [\-r/\-\-readahead ReadAheadSectors]
-[-R|--regionsize MirrorLogRegionSize]
 [\-t/\-\-test]
 [\-v/\-\-verbose] [\-Z/\-\-zero y/n]
 VolumeGroupName [PhysicalVolumePath...]
@@ -21,7 +22,7 @@ VolumeGroupName [PhysicalVolumePath...]
 
 .br
 .B lvcreate
-{\-l/\-\-extents LogicalExtentsNumber |
+{\-l/\-\-extents LogicalExtentsNumber[%{VG|FREE}] |
  \-L/\-\-size LogicalVolumeSize[kKmMgGtT]}
 [\-c/\-\-chunksize ChunkSize]
 \-s/\-\-snapshot \-n/\-\-name SnapshotLogicalVolumeName OriginalLogicalVolumePath
@@ -62,9 +63,12 @@ StripeSize must be 2^n (n = 2 to 9) for metadata in LVM1 format.
 For metadata in LVM2 format, the stripe size may be a larger
 power of 2 but must not exceed the physical extent size.
 .TP
-.I \-l, \-\-extents LogicalExtentsNumber
+.I \-l, \-\-extents LogicalExtentsNumber[%{VG|FREE}]
 Gives the number of logical extents to allocate for the new
 logical volume.
+This can also be expressed as a percentage of the total space
+in the Volume Group with the suffix %VG or of the remaining free space
+with the suffix %FREE.
 .TP
 .I \-L, \-\-size LogicalVolumeSize[kKmMgGtT]
 Gives the size to allocate for the new logical volume.
@@ -79,6 +83,23 @@ Set the minor number.
 .I \-M, \-\-persistent y/n
 Set to y to make the minor number specified persistent.
 .TP
+.I \-m, \-\-mirrors Mirrors
+Creates a mirrored logical volume with "Mirrors" copies.  For example,
+specifying "-m 1" would result in a mirror with two-sides; that is, a
+linear volume plus one copy.
+
+Specifying the optional argument "--nosync" will cause the creation
+of the mirror to skip the initial resynchronization.  Any data written
+afterwards will be mirrored, but the original contents will not be
+copied.  This is useful for skipping a potentially long and resource
+intensive initial sync.
+
+Specifying the optional argument "--corelog" will create a mirror with
+an in-memory log verses a disk-based (persistent) log.  While this
+removes the need for an extra log device and *may* be slightly faster,
+it requires that the entire mirror be resynchronized upon each
+instantiation (e.g. a reboot).
+.TP
 .I \-n, \-\-name LogicalVolumeName
 The name for the new logical volume.
 .br
@@ -94,8 +115,8 @@ Default is read and write.
 Set read ahead sector count of this logical volume to a value between 2 and 120.
 Ignored by device-mapper.
 .TP
-.I \-R \-\-regionsize MirrorLogRegionSize
-A mirror is divided into regions of this size (in KB), and the mirror log 
+.I \-R, \-\-regionsize MirrorLogRegionSize
+A mirror is divided into regions of this size (in MB), and the mirror log 
 uses this granularity to track which regions are in sync.
 .TP
 .I \-s, \-\-snapshot
@@ -122,10 +143,15 @@ Default is yes.
 Warning: trying to mount an unzeroed logical volume can cause the system to
 hang.
 .SH Examples
-"lvcreate -i 3 -I 8 -L 100 vg00" tries to create a striped logical
+"lvcreate -i 3 -I 8 -L 100M vg00" tries to create a striped logical
 volume with 3 stripes, a stripesize of 8KB and a size of 100MB in the volume
 group named vg00. The logical volume name will be chosen by lvcreate.
 
+"lvcreate -m1 -L 500M vg00" tries to create a mirror logical volume
+with 2 sides with a useable size of 500 MiB.  This operation would
+require 3 devices - two for the mirror devices and one for the disk
+log.
+
 "lvcreate --size 100m --snapshot --name snap /dev/vg00/lvol1"
 .br
 creates a snapshot logical volume named /dev/vg00/snap which has access to the

man/lvextend.8

@@ -6,7 +6,7 @@ lvextend \- extend the size of a logical volume
 [\-\-alloc AllocationPolicy]
 [\-A/\-\-autobackup y/n] [\-d/\-\-debug] [\-h/\-?/\-\-help]
 [\-i/\-\-stripes Stripes [\-I/\-\-stripesize StripeSize]]
-{\-l/\-\-extents [+]LogicalExtentsNumber |
+{\-l/\-\-extents [+]LogicalExtentsNumber[%{VG|LV|FREE}] |
 \-L/\-\-size [+]LogicalVolumeSize[kKmMgGtT]}
 [\-t/\-\-test]
 [\-v/\-\-verbose] LogicalVolumePath [PhysicalVolumePath...]
@@ -15,13 +15,20 @@ lvextend allows you to extend the size of a logical volume.
 Extension of snapshot logical volumes (see
 .B lvcreate(8)
 for information to create snapshots) is supported as well.
+But to change the number of copies in a mirrored logical
+volume use 
+.BR lvconvert (8).
 .SH OPTIONS
 See \fBlvm\fP for common options.
 .TP
-.I \-l, \-\-extents [+]LogicalExtentsNumber
+.I \-l, \-\-extents [+]LogicalExtentsNumber[%{VG|LV|FREE}]
 Extend or set the logical volume size in units of logical extents.
 With the + sign the value is added to the actual size
 of the logical volume and without it, the value is taken as an absolute one.
+The number can also be expressed as a percentage of the total space
+in the Volume Group with the suffix %VG or relative to the existing
+size of the Logical Volume with the suffix %LV or as a percentage of the remaining
+free space in the Volume Group with the suffix %FREE.
 .TP
 .I \-L, \-\-size [+]LogicalVolumeSize[kKmMgGtT]
 Extend or set the logical volume size in units in units of megabytes.
@@ -48,6 +55,7 @@ there are enough free physical extents in it.
 .SH SEE ALSO
 .BR lvm (8), 
 .BR lvcreate (8), 
+.BR lvconvert (8), 
 .BR lvreduce (8), 
 .BR lvresize (8), 
 .BR lvchange (8)

man/lvm.8

@@ -38,7 +38,7 @@ The following commands are built into lvm without links normally
 being created in the filesystem for them.
 .TP
 \fBdumpconfig\fP \(em Display the configuration information after 
-loading \fBlvm.conf\fP (8) and any other configuration files.
+loading \fBlvm.conf\fP (5) and any other configuration files.
 .TP
 \fBformats\fP \(em Display recognised metadata formats.
 .TP
@@ -136,7 +136,7 @@ Characters allowed in tags are: A-Z a-z 0-9 _ + . -
 Delete the tag \fBtag\fP from a PV, VG or LV, if it's present.
 .TP
 \fB--alloc AllocationPolicy\fP
-The allocation policy to use: \fBcontiguous\fP, \fBnormal\fP, \fBanywhere\fP or \fBinherit\fP.
+The allocation policy to use: \fBcontiguous\fP, \fBcling\fP, \fBnormal\fP, \fBanywhere\fP or \fBinherit\fP.
 When a command needs to allocate physical extents from the volume group,
 the allocation policy controls how they are chosen.  
 Each volume group and logical volume has an allocation policy.
@@ -146,15 +146,18 @@ physical volume.  The default for a logical volume is \fBinherit\fP
 which applies the same policy as for the volume group.  These policies can
 be changed using \fBlvchange\fP (8) and \fBvgchange\fP (8) or over-ridden
 on the command line of any command that performs allocation.
-The \fBcontiguous\fP policy requires that new extents are adjacent to
-existing extents. If there are sufficient free extents to satisfy
+The \fBcontiguous\fP policy requires that new extents be placed adjacent
+to existing extents. 
+The \fBcling\fP policy places new extents on the same physical
+volume as existing extents in the same stripe of the Logical Volume.
+If there are sufficient free extents to satisfy
 an allocation request but \fBnormal\fP doesn't use them,
 \fBanywhere\fP will - even if that reduces performance by
 placing two stripes on the same physical volume.
 .IP
 N.B. The policies described above are not implemented fully yet.
-In particular, \fBcontiguous\fP does not place new extents adjacent to existing
-extents and \fBanywhere\fP is not implemented at all.
+In particular, contiguous free space cannot be broken up to
+satisfy allocation attempts.
 .SH ENVIRONMENT VARIABLES
 .TP
 \fBLVM_SYSTEM_DIR\fP 

man/lvm.conf.5

@@ -252,7 +252,7 @@ if \fBlocking_type\fP is set to 1.  The default is \fB/var/lock/lvm\fP.
 .IP
 \fBlocking_library\fP \(em The name of the external locking
 library to load if \fBlocking_type\fP is set to 2.
-The default is \fBlvm2_locking.so\fP.  If you need to write
+The default is \fBliblvm2clusterlock.so\fP.  If you need to write
 such a library, look at the lib/locking source code directory.
 .TP
 \fBtags\fP \(em Host tag settings

man/lvreduce.8

@@ -4,7 +4,8 @@ lvreduce \- reduce the size of a logical volume
 .SH SYNOPSIS
 .B lvreduce
 [\-A/\-\-autobackup y/n] [\-d/\-\-debug] [\-f/\-\-force]
-[\-h/\-?/\-\-help] {\-l/\-\-extents [\-]LogicalExtentsNumber |
+[\-h/\-?/\-\-help]
+{\-l/\-\-extents [\-]LogicalExtentsNumber[%{VG|LV|FREE}] |
 \-L/\-\-size [\-]LogicalVolumeSize[kKmMgGtT]}
 [\-t/\-\-test]
 [\-v/\-\-verbose] LogicalVolume[Path]
@@ -21,21 +22,29 @@ running lvreduce so that the extents that are to be removed are not in use.
 Shrinking snapshot logical volumes (see
 .B lvcreate(8)
 for information to create snapshots) is supported as well.
+But to change the number of copies in a mirrored logical
+volume use 
+.B lvconvert (8).
 .br
 Sizes will be rounded if necessary - for example, the volume size must
 be an exact number of extents and the size of a striped segment must
 be a multiple of the number of stripes.
+.br
 .SH OPTIONS
 See \fBlvm\fP for common options.
 .TP
 .I \-f, \-\-force
 Force size reduction without any question.
 .TP
-.I \-l, \-\-extents [\-]LogicalExtentsNumber
+.I \-l, \-\-extents [\-]LogicalExtentsNumber[%{VG|LV|FREE}]
 Reduce or set the logical volume size in units of logical extents.
 With the - sign the value will be subtracted from
 the logical volume's actual size and without it the will be taken as
 an absolute size.
+The number can also be expressed as a percentage of the total space
+in the Volume Group with the suffix %VG or relative to the existing
+size of the Logical Volume with the suffix %LV or as a percentage of the remaining
+free space in the Volume Group with the suffix %FREE.
 .TP
 .I \-L, \-\-size [\-]LogicalVolumeSize[kKmMgGtT]
 Reduce or set the logical volume size in units of megabyte by default.
@@ -49,6 +58,7 @@ an absolute size.
 in volume group vg00 by 3 logical extents.
 .SH SEE ALSO
 .BR lvchange (8),
+.BR lvconvert (8), 
 .BR lvcreate (8), 
 .BR lvextend (8), 
 .BR lvm (8), 

man/lvresize.8

@@ -6,7 +6,7 @@ lvresize \- resize a logical volume
 [\-\-alloc AllocationPolicy]
 [\-A/\-\-autobackup y/n] [\-d/\-\-debug] [\-h/\-?/\-\-help]
 [\-i/\-\-stripes Stripes [\-I/\-\-stripesize StripeSize]]
-{\-l/\-\-extents [+]LogicalExtentsNumber |
+{\-l/\-\-extents [+]LogicalExtentsNumber[%{VG|LV|FREE}] |
 \-L/\-\-size [+]LogicalVolumeSize[kKmMgGtT]}
 [\-t/\-\-test]
 [\-v/\-\-verbose] LogicalVolumePath [PhysicalVolumePath...]
@@ -19,13 +19,20 @@ shrunk first so that the extents that are to be removed are not in use.
 Resizing snapshot logical volumes (see
 .B lvcreate(8)
 for information about creating snapshots) is supported as well.
+But to change the number of copies in a mirrored logical
+volume use 
+.BR lvconvert (8).
 .SH OPTIONS
 See \fBlvm\fP for common options.
 .TP
-.I \-l, \-\-extents [+/-]LogicalExtentsNumber
+.I \-l, \-\-extents [+/-]LogicalExtentsNumber[%{VG|LV|FREE}]
 Change or set the logical volume size in units of logical extents.
 With the + or - sign the value is added to or subtracted from the actual size
 of the logical volume and without it, the value is taken as an absolute one.
+The number can also be expressed as a percentage of the total space
+in the Volume Group with the suffix %VG or relative to the existing
+size of the Logical Volume with the suffix %LV or as a percentage of the remaining 
+free space in the Volume Group with the suffix %FREE.
 .TP
 .I \-L, \-\-size [+/-]LogicalVolumeSize[kKmMgGtT]
 Change or set the logical volume size in units of megabytes.
@@ -49,6 +56,7 @@ must use a single value throughout.
 StripeSize must be 2^n (n = 2 to 9)
 .SH SEE ALSO
 .BR lvm (8), 
+.BR lvconvert (8),
 .BR lvcreate (8), 
 .BR lvreduce (8), 
 .BR lvchange (8)

man/vgchange.8

@@ -9,6 +9,7 @@ vgchange \- change attributes of a volume group
 .IR AllocationPolicy ]
 .RB [ \-A | \-\-autobackup " {" y | n }]
 .RB [ \-a | \-\-available " [e|l] {" y | n }]
+.RB [ \-\-monitor " {" y | n }]
 .RB [ \-d | \-\-debug]
 .RB [ \-\-deltag
 .IR Tag ]
@@ -16,6 +17,8 @@ vgchange \- change attributes of a volume group
 .RB [ \-\-ignorelockingfailure]
 .RB [ \-l | \-\-logicalvolume
 .IR MaxLogicalVolumes ]
+.RB [ -p | \-\-maxphysicalvolumes
+.IR MaxPhysicalVolumes ]
 .RB [ \-P | \-\-partial]
 .RB [ \-s | \-\-physicalextentsize
 .IR PhysicalExtentSize [ \fBkKmMgGtT\fR ]]
@@ -57,10 +60,33 @@ on the local node.
 Logical volumes with single-host snapshots are always activated
 exclusively because they can only be used on one node at once.
 .TP
+.BR \-\-monitor " " { y | n }
+Controls whether or not a mirrored logical volume is monitored by
+dmeventd, if it is installed.
+If a device used by a monitored mirror reports an I/O error,
+the failure is handled according to 
+.BR mirror_image_fault_policy
+and 
+.BR mirror_log_fault_policy
+set in 
+.BR lvm.conf (5).
+.TP
 .BR \-l ", " \-\-logicalvolume " " \fIMaxLogicalVolumes\fR
 Changes the maximum logical volume number of an existing inactive
 volume group.
 .TP
+.BR \-p ", " \-\-maxphysicalvolumes " " \fIMaxPhysicalVolumes\fR
+Changes the maximum number of physical volumes that can belong
+to this volume group.
+For volume groups with metadata in lvm1 format, the limit is 255.
+If the metadata uses lvm2 format, the value 0
+removes this restriction: there is then no limit.
+If you have a large number of physical volumes in
+a volume group with metadata in lvm2 format,
+for tool performance reasons, you should consider
+some use of \fB--metadatacopies 0\fP
+as described in \fBpvcreate(8)\fP.
+.TP
 .BR \-s ", " \-\-physicalextentsize " " \fIPhysicalExtentSize\fR[\fBkKmMgGtT\fR]
 Changes the physical extent size on physical volumes of this volume group.
 A size suffix (k for kilobytes up to t for terabytes) is optional, megabytes

man/vgscan.8

@@ -11,7 +11,7 @@ vgscan \- scan all disks for volume groups and rebuild caches
 .SH DESCRIPTION
 vgscan scans all SCSI, (E)IDE disks, multiple devices and a bunch
 of other disk devices in the system looking for LVM physical volumes
-and volume groups.  Define a filter in \fBlvm.conf\fP(8) to restrict 
+and volume groups.  Define a filter in \fBlvm.conf\fP(5) to restrict 
 the scan to avoid a CD ROM, for example.
 .LP
 In LVM2, vgscans take place automatically; but you might still need to

scripts/clvmd_init_rhel4

@@ -93,6 +93,24 @@ stop()
 	return $rtrn
 }
 
+wait_for_finish()
+{
+	count=0
+
+	while [ "$count" -le 10 -a -n "`pidof clvmd`" ]
+	do
+		sleep 1
+		count=$((count + 1))
+	done
+	
+	if [ `pidof clvmd` ]
+	then
+		return 1
+	else
+		return 0
+	fi
+}
+
 rtrn=1
 
 # See how we were called.
@@ -112,6 +130,7 @@ case "$1" in
   restart)
 	if stop
 	then
+		wait_for_finish
 		start
 	fi 
 	rtrn=$?

scripts/lvm_dump.sh

@@ -0,0 +1,118 @@
+#!/bin/bash
+#
+# lvm_dump: This script is used to collect pertinent information for
+#           the debugging of lvm issues.
+#
+
+function usage {
+	echo "$0 [options]"
+	echo "    -h print this message"
+	echo "    -a advanced collection - warning: if lvm is already hung,"
+	echo "       then this script may hang as well if -a is used"
+	echo "    -m gather LVM metadata from the PVs"
+	echo "    -d dump directory to place data in (default=/tmp/lvm_dump.\$\$)"
+	echo "    -c if running clvmd, gather cluster data as well"
+	echo ""
+	
+	exit 1
+}
+
+advanced=0
+clustered=0
+metadata=0
+while getopts :acd:hm opt; do
+	case $opt in 
+		a)	advanced=1 ;;
+		c)	clustered=1 ;;
+		d)	lvm_dir=$OPTARG ;;
+		h)	usage ;;
+		m)	metadata=1 ;;
+		:)	echo "$0: $OPTARG requires a value:"; usage ;;
+		\?)     echo "$0: unknown option $OPTARG"; usage ;;
+		*)	usage ;;
+	esac
+done
+
+dir=`mktemp -d -p /tmp lvm_dump.XXXXXX` || exit 2
+lvm_dir="$dir/lvm_dump"
+
+echo " "
+echo "Creating dump directory: $lvm_dir"
+echo " "
+
+mkdir -p $lvm_dir || exit 3
+
+if (( $advanced )); then
+	echo "Gathering LVM volume info..."
+
+	echo "  vgscan..."
+	vgscan -vvvv > $lvm_dir/vgscan 2>&1
+
+	echo "  pvscan..."
+	pvscan -v >> $lvm_dir/pvscan 2>/dev/null
+
+	echo "  lvs..."
+	lvs -a -o +devices >> $lvm_dir/lvs 2>/dev/null
+
+	echo "  pvs..."
+	pvs -a -v > $lvm_dir/pvs 2>/dev/null
+
+	echo "  vgs..."
+	vgs -v > $lvm_dir/vgs 2>/dev/null
+fi
+
+if (( $clustered )); then
+	echo "Gathering cluster info..."
+	echo "STATUS: " > $lvm_dir/cluster_info
+	echo "----------------------------------" >> $lvm_dir/cluster_info
+	cman_tool status >> $lvm_dir/cluster_info
+	echo " " >> $lvm_dir/lvm_info
+
+	echo "SERVICES: " >> $lvm_dir/cluster_info
+	echo "----------------------------------" >> $lvm_dir/cluster_info
+	cman_tool services >> $lvm_dir/cluster_info
+	echo " " >> $lvm_dir/lvm_info
+fi
+
+echo "Gathering LVM & device-mapper version info..."
+echo "LVM VERSION:" > $lvm_dir/versions
+lvs --version >> $lvm_dir/versions
+echo "DEVICE MAPPER VERSION:" >> $lvm_dir/versions
+dmsetup --version >> $lvm_dir/versions
+
+echo "Gathering dmsetup info..."
+dmsetup info -c > $lvm_dir/dmsetup_info
+dmsetup table > $lvm_dir/dmsetup_table
+dmsetup status > $lvm_dir/dmsetup_status
+
+echo "Gathering process info..."
+ps alx > $lvm_dir/ps_info
+
+echo "Gathering console messages..."
+tail -n 75 /var/log/messages > $lvm_dir/messages
+
+echo "Gathering /etc/lvm info..."
+cp -a /etc/lvm $lvm_dir/lvm
+
+echo "Gathering /dev listing..."
+ls -la /dev > $lvm_dir/dev_listing
+
+if (( $metadata )); then
+	echo "Gathering LVM metadata from Physical Volumes..."
+
+	mkdir -p $lvm_dir/metadata
+
+	for pv in `pvs --noheadings -o name`
+	do
+		echo "  $pv"
+		name=`basename $pv`
+		dd if=$pv of=$lvm_dir/metadata/$name bs=512 count=`pvs --noheadings --nosuffix --units s -o pe_start $pv | tr -d \ `
+	done 2>/dev/null
+fi
+
+lvm_dump=$lvm_dir.tgz
+echo "Creating tarball $lvm_dump..."
+tar czf $lvm_dump $lvm_dir 2>/dev/null
+
+exit 0
+

scripts/lvmconf.sh

@@ -1,6 +1,6 @@
 #!/bin/sh
 #
-# Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
 #
 # This file is part of the lvm2-cluster package.
 #
@@ -21,8 +21,9 @@ function usage
     echo "usage: $0 <command>"
     echo ""
     echo "Commands:"
-    echo "Enable clvm:  --enable-cluster --lockinglibdir <dir> [--lockinglib <lib>]"
+    echo "Enable clvm:  --enable-cluster [--lockinglibdir <dir>] [--lockinglib <lib>]"
     echo "Disable clvm: --disable-cluster"
+    echo "Set locking library: --lockinglibdir <dir> [--lockinglib <lib>]"
     echo ""
     echo "Global options:"
     echo "Config file location: --file <configfile>"
@@ -86,14 +87,13 @@ function validate_args
             exit 10
     fi
 
-    if [ -z "$LOCKING_TYPE" ]; then
+    if [ -z "$LOCKING_TYPE" ] && [ -z "$LOCKINGLIBDIR" ]; then
         usage
         exit 1
     fi
 
-    if [ "$LOCKING_TYPE" == "2" ]; then
+    if [ -n "$LOCKINGLIBDIR" ]; then
 
-        [ -z "$LOCKINGLIBDIR" ] && usage && exit 1    
         [ -z "$LOCKINGLIB" ] && LOCKINGLIB="liblvm2clusterlock.so"
             
         if [ "${LOCKINGLIBDIR:0:1}" != "/" ]
@@ -109,6 +109,10 @@ function validate_args
         fi
         
     fi
+
+    if [ "$LOCKING_TYPE" = "1" ] && [ -n "$LOCKINGLIBDIR" -o -n "$LOCKINGLIB" ]; then
+	echo "Superfluous locking lib parameter, ignoring"
+    fi
 }
 
 umask 0077
@@ -153,11 +157,19 @@ then
     fi
 fi
 
+if [ "$LOCKING_TYPE" = "2" ] && [ -z "$LOCKINGLIBDIR" ] && [ "$have_dir" = "1" ]; then
+	echo "no library_dir specified in $CONFIGFILE"
+	exit 16
+fi
+
 # So if we don't have "global {" we need to create one and 
 # populate it
 
 if [ "$have_global" = "1" ]
 then
+    if [ -z "$LOCKING_TYPE" ]; then
+	LOCKING_TYPE=1
+    fi
     if [ "$LOCKING_TYPE" = "2" ]; then
         cat $CONFIGFILE - <<EOF > $TMPFILE
 global {
@@ -180,14 +192,16 @@ else
     # locking entries as appropriate
     #
 
-    if [ "$have_type" = "0" ] 
-    then
-	SEDCMD=" s/^[[:blank:]]*locking_type[[:blank:]]*=.*/\ \ \ \ locking_type = $LOCKING_TYPE/g"
-    else
-	SEDCMD=" /global[[:blank:]]*{/a\ \ \ \ locking_type = $LOCKING_TYPE"
+    if [ -n "$LOCKING_TYPE" ]; then
+	if [ "$have_type" = "0" ] 
+	then
+	    SEDCMD=" s/^[[:blank:]]*locking_type[[:blank:]]*=.*/\ \ \ \ locking_type = $LOCKING_TYPE/g"
+	else
+	    SEDCMD=" /global[[:blank:]]*{/a\ \ \ \ locking_type = $LOCKING_TYPE"
+	fi
     fi
     
-    if [ "$LOCKING_TYPE" = "2" ]; then
+    if [ -n "$LOCKINGLIBDIR" ]; then
         if [ "$have_dir" = "0" ] 
             then
             SEDCMD="${SEDCMD}\ns'^[[:blank:]]*library_dir[[:blank:]]*=.*'\ \ \ \ library_dir = \"$LOCKINGLIBDIR\"'g"
@@ -201,7 +215,9 @@ else
         else
             SEDCMD="${SEDCMD}\n/global[[:blank:]]*{/a\ \ \ \ locking_library = \"$LOCKINGLIB\""
         fi
-    else
+    fi
+
+    if [ "$LOCKING_TYPE" = "1" ]; then
         # if we're not using cluster locking, remove the library dir and locking library name
         if [ "$have_dir" = "0" ] 
             then

tools/Makefile.in

@@ -80,7 +80,9 @@ endif
 
 LVMLIBS = -llvm
 
-CLEAN_TARGETS = liblvm2cmd.so liblvm2cmd.a lvm lvm.o lvm.static lvm.cflow lvm.xref lvm.tree lvm.rxref lvm.rtree
+CLEAN_TARGETS = liblvm2cmd.so liblvm2cmd.a liblvm2cmd-static.a lvm lvm.o \
+		lvm2cmd.o lvm2cmd-static.o lvm2cmdlib.o lvm.static \
+		lvm.cflow lvm.xref lvm.tree lvm.rxref lvm.rtree
 
 ifeq ("@CMDLIB@", "yes")
 	TARGETS += liblvm2cmd.so
@@ -104,9 +106,13 @@ lvm.static: $(OBJECTS) lvm-static.o $(top_srcdir)/lib/liblvm.a
 	$(CC) -o $@ $(OBJECTS) lvm-static.o -static $(LDFLAGS) $(LVMLIBS) \
 		$(LIBS) -rdynamic
 
-liblvm2cmd.a: $(top_srcdir)/lib/liblvm.a $(OBJECTS)
+liblvm2cmd.a: $(top_srcdir)/lib/liblvm.a $(OBJECTS) lvmcmdlib.o lvm2cmd.o
 	cat $(top_srcdir)/lib/liblvm.a > $@
-	$(AR) rs $@ $(OBJECTS)
+	$(AR) rs $@ $(OBJECTS) lvmcmdlib.o lvm2cmd.o
+
+liblvm2cmd-static.a: $(top_srcdir)/lib/liblvm.a $(OBJECTS) lvmcmdlib.o lvm2cmd-static.o
+	cat $(top_srcdir)/lib/liblvm.a > $@
+	$(AR) rs $@ $(OBJECTS) lvmcmdlib.o lvm2cmd-static.o
 
 liblvm2cmd.so: liblvm2cmd.a $(LDDEPS)
 
@@ -141,8 +147,8 @@ install_cmdlib_dynamic: liblvm2cmd.so
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 444 lvm2cmd.h \
 		$(includedir)/lvm2cmd.h
 
-install_cmdlib_static: liblvm2cmd.a
-	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) liblvm2cmd.a \
+install_cmdlib_static: liblvm2cmd-static.a
+	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) liblvm2cmd-static.a \
 		$(libdir)/liblvm2cmd.a.$(LIB_VERSION)
 	$(LN_S) -f liblvm2cmd.a.$(LIB_VERSION) $(libdir)/liblvm2cmd.a
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 444 lvm2cmd.h \

tools/args.h

@@ -84,7 +84,7 @@ arg(interval_ARG, 'i', "interval", int_arg)
 arg(iop_version_ARG, 'i', "iop_version", NULL)
 arg(logicalvolume_ARG, 'l', "logicalvolume", int_arg)
 arg(maxlogicalvolumes_ARG, 'l', "maxlogicalvolumes", int_arg)
-arg(extents_ARG, 'l', "extents", int_arg_with_sign)
+arg(extents_ARG, 'l', "extents", int_arg_with_sign_and_percent)
 arg(lvmpartition_ARG, 'l', "lvmpartition", NULL)
 arg(list_ARG, 'l', "list", NULL)
 arg(size_ARG, 'L', "size", size_mb_arg)

tools/commands.h

@@ -83,6 +83,7 @@ xx(lvconvert,
    "Change logical volume layout",
    "lvconvert "
    "[-m|--mirrors Mirrors [--corelog]]\n"
+   "\t[-R|--regionsize MirrorLogRegionSize]\n"
    "\t[--alloc AllocationPolicy]\n"
    "\t[-d|--debug]\n"
    "\t[-h|-?|--help]\n"
@@ -100,7 +101,7 @@ xx(lvconvert,
    "\t[--version]" "\n"
    "\tOriginalLogicalVolume[Path] SnapshotLogicalVolume[Path]\n",
 
-   alloc_ARG, chunksize_ARG, mirrors_ARG, corelog_ARG,
+   alloc_ARG, chunksize_ARG, mirrors_ARG, corelog_ARG, regionsize_ARG,
    snapshot_ARG, test_ARG, zero_ARG)
 
 xx(lvcreate,
@@ -137,7 +138,7 @@ xx(lvcreate,
    "\t[-d|--debug]\n"
    "\t[-h|-?|--help]\n"
    "\t[-i|--stripes Stripes [-I|--stripesize StripeSize]]\n"
-   "\t{-l|--extents LogicalExtentsNumber |\n"
+   "\t{-l|--extents LogicalExtentsNumber[%{VG|LV|FREE}] |\n"
    "\t -L|--size LogicalVolumeSize[kKmMgGtT]}\n"
    "\t[-M|--persistent {y|n}] [--major major] [--minor minor]\n"
    "\t[-n|--name LogicalVolumeName]\n"
@@ -202,7 +203,7 @@ xx(lvextend,
    "\t[-d|--debug]\n"
    "\t[-h|--help]\n"
    "\t[-i|--stripes Stripes [-I|--stripesize StripeSize]]\n"
-   "\t{-l|--extents [+]LogicalExtentsNumber |\n"
+   "\t{-l|--extents [+]LogicalExtentsNumber[%{VG|FREE}] |\n"
    "\t -L|--size [+]LogicalVolumeSize[kKmMgGtT]}\n"
    "\t[-m|--mirrors Mirrors]\n"
    "\t[-n|--nofsck]\n"
@@ -267,7 +268,7 @@ xx(lvreduce,
    "\t[-d|--debug]\n"
    "\t[-f|--force]\n"
    "\t[-h|--help]\n"
-   "\t{-l|--extents [-]LogicalExtentsNumber |\n"
+   "\t{-l|--extents [-]LogicalExtentsNumber[%{VG|LV|FREE}] |\n"
    "\t -L|--size [-]LogicalVolumeSize[kKmMgGtT]}\n"
    "\t[-n|--nofsck]\n"
    "\t[-r|--resizefs]\n"
@@ -315,7 +316,7 @@ xx(lvresize,
    "\t[-d|--debug]\n"
    "\t[-h|--help]\n"
    "\t[-i|--stripes Stripes [-I|--stripesize StripeSize]]\n"
-   "\t{-l|--extents [+|-]LogicalExtentsNumber |\n"
+   "\t{-l|--extents [+|-]LogicalExtentsNumber[%{VG|LV|FREE}] |\n"
    "\t -L|--size [+|-]LogicalVolumeSize[kKmMgGtT]}\n"
    "\t[-n|--nofsck]\n"
    "\t[-r|--resizefs]\n"
@@ -607,6 +608,7 @@ xx(vgchange,
    "\t -c|--clustered {y|n} |" "\n"
    "\t -x|--resizeable {y|n} |" "\n"
    "\t -l|--logicalvolume MaxLogicalVolumes |" "\n"
+   "\t -p|--maxphysicalvolumes MaxPhysicalVolumes |" "\n"
    "\t -s|--physicalextentsize PhysicalExtentSize[kKmMgGtT] |" "\n"
    "\t --addtag Tag |\n"
    "\t --deltag Tag}\n"
@@ -614,8 +616,8 @@ xx(vgchange,
 
    addtag_ARG, alloc_ARG, allocation_ARG, autobackup_ARG, available_ARG,
    clustered_ARG, deltag_ARG, ignorelockingfailure_ARG, logicalvolume_ARG,
-   monitor_ARG, partial_ARG, physicalextentsize_ARG, resizeable_ARG,
-   resizable_ARG, test_ARG, uuid_ARG)
+   maxphysicalvolumes_ARG, monitor_ARG, partial_ARG, physicalextentsize_ARG,
+   resizeable_ARG, resizable_ARG, test_ARG, uuid_ARG)
 
 xx(vgck,
    "Check the consistency of volume group(s)",

tools/dmsetup.c

@@ -1,6 +1,6 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
  * Copyright (C) 2005 NEC Corperation
  *
  * This file is part of the device-mapper userspace tools.
@@ -38,6 +38,16 @@
 #include <locale.h>
 #include <langinfo.h>
 
+#include <fcntl.h>
+#include <sys/stat.h>
+
+/* FIXME Unused so far */
+#undef HAVE_SYS_STATVFS_H
+
+#ifdef HAVE_SYS_STATVFS_H
+#  include <sys/statvfs.h>
+#endif
+
 #ifdef HAVE_SYS_IOCTL_H
 #  include <sys/ioctl.h>
 #endif
@@ -78,6 +88,11 @@ extern char *optarg;
 
 #define LINE_SIZE 4096
 #define ARGS_MAX 256
+#define LOOP_TABLE_SIZE (PATH_MAX + 255)
+
+/* FIXME Should be elsewhere */
+#define SECTOR_SHIFT 9L
+#define DEV_PATH "/dev/"
 
 #define err(msg, x...) fprintf(stderr, msg "\n", ##x)
 
@@ -93,6 +108,7 @@ enum {
 	MAJOR_ARG,
 	MINOR_ARG,
 	MODE_ARG,
+	NOFLUSH_ARG,
 	NOHEADINGS_ARG,
 	NOLOCKFS_ARG,
 	NOOPENCOUNT_ARG,
@@ -180,7 +196,7 @@ static int _parse_file(struct dm_task *dmt, const char *file)
 
 #ifndef HAVE_GETLINE
 	buffer_size = LINE_SIZE;
-	if (!(buffer = malloc(buffer_size))) {
+	if (!(buffer = dm_malloc(buffer_size))) {
 		err("Failed to malloc line buffer.");
 		return 0;
 	}
@@ -195,7 +211,7 @@ static int _parse_file(struct dm_task *dmt, const char *file)
 	r = 1;
 
       out:
-	free(buffer);
+	dm_free(buffer);
 	if (file)
 		fclose(fp);
 	return r;
@@ -512,7 +528,7 @@ static int _message(int argc, char **argv, void *data __attribute((unused)))
 	for (i = 0; i < argc; i++)
 		sz += strlen(argv[i]) + 1;
 
-	str = malloc(sz);
+	str = dm_malloc(sz);
 	memset(str, 0, sz);
 
 	for (i = 0; i < argc; i++) {
@@ -524,7 +540,7 @@ static int _message(int argc, char **argv, void *data __attribute((unused)))
 	if (!dm_task_set_message(dmt, str))
 		goto out;
 
-	free(str);
+	dm_free(str);
 
 	if (_switches[NOOPENCOUNT_ARG] && !dm_task_no_open_count(dmt))
 		goto out;
@@ -603,6 +619,9 @@ static int _simple(int task, const char *name, uint32_t event_nr, int display)
 	if (event_nr && !dm_task_set_event_nr(dmt, event_nr))
 		goto out;
 
+	if (_switches[NOFLUSH_ARG] && !dm_task_no_flush(dmt))
+		goto out;
+
 	if (_switches[NOOPENCOUNT_ARG] && !dm_task_no_open_count(dmt))
 		goto out;
 
@@ -744,28 +763,28 @@ static int _error_device(int argc __attribute((unused)), char **argv __attribute
 		return 0;
 
 	if (!_set_task_device(dmt, name, 0))
-		goto err;
+		goto error;
 
 	if (!dm_task_add_target(dmt, 0, size, "error", ""))
-		goto err;
+		goto error;
 
 	if (_switches[READ_ONLY] && !dm_task_set_ro(dmt))
-		goto err;
+		goto error;
 
 	if (_switches[NOOPENCOUNT_ARG] && !dm_task_no_open_count(dmt))
-		goto err;
+		goto error;
 
 	if (!dm_task_run(dmt))
-		goto err;
+		goto error;
 
 	if (!_simple(DM_DEVICE_RESUME, name, 0, 0)) {
 		_simple(DM_DEVICE_CLEAR, name, 0, 0);
-		goto err;
+		goto error;
 	}
 
 	r = 1;
 
-err:
+error:
 	dm_task_destroy(dmt);
 	return r;
 }
@@ -1487,12 +1506,12 @@ struct command {
 static struct command _commands[] = {
 	{"create", "<dev_name> [-j|--major <major> -m|--minor <minor>]\n"
 	  "\t                  [-U|--uid <uid>] [-G|--gid <gid>] [-M|--mode <octal_mode>]\n"
-	  "\t                  [-u|uuid <uuid>]"
+	  "\t                  [-u|uuid <uuid>]\n"
 	  "\t                  [--notable | --table <table> | <table_file>]",
 	 1, 2, _create},
 	{"remove", "[-f|--force] <device>", 0, 1, _remove},
 	{"remove_all", "[-f|--force]", 0, 0, _remove_all},
-	{"suspend", "<device>", 0, 1, _suspend},
+	{"suspend", "[--noflush] <device>", 0, 1, _suspend},
 	{"resume", "<device>", 0, 1, _resume},
 	{"load", "<device> [<table_file>]", 0, 2, _load},
 	{"clear", "<device>", 0, 1, _clear},
@@ -1529,6 +1548,13 @@ static void _usage(FILE *out)
 	return;
 }
 
+static void _losetup_usage(FILE *out)
+{
+	fprintf(out, "Usage:\n\n");
+	fprintf(out, "losetup [-d|-a] [-e encryption] "
+		     "[-o offset] [-f|loop_device] [file]\n\n");
+}
+
 static struct command *_find_command(const char *name)
 {
 	int i;
@@ -1606,11 +1632,225 @@ static int _process_tree_options(const char *options)
 	return 1;
 }
 
+/*
+ * Returns the full absolute path, or NULL if the path could
+ * not be resolved.
+ */
+static char *_get_abspath(char *path)
+{
+	char *_path;
+
+#ifdef HAVE_CANONICALIZE_FILE_NAME
+	_path = canonicalize_file_name(path);
+#else
+	/* FIXME Provide alternative */
+#endif
+	return _path;
+}
+
+static char *parse_loop_device_name(char *dev)
+{
+	char *buf;
+	char *device;
+
+	if (!(buf = dm_malloc(PATH_MAX)));
+		return NULL;
+
+	if (dev[0] == '/') {
+		if (!(device = _get_abspath(dev)))
+			goto error;
+
+		if (strncmp(device, DEV_PATH, strlen(DEV_PATH)))
+			goto error;
+
+		strncpy(buf, strrchr(device, '/') + 1, PATH_MAX);
+		dm_free(device);
+
+	} else {
+		/* check for device number */
+		if (!strncmp(dev, "loop", strlen("loop")))
+			strncpy(buf, dev, PATH_MAX);
+		else
+			goto error;
+	}
+
+	return buf;
+
+error:
+	return NULL;
+}
+
+/*
+ *  create a table for a mapped device using the loop target.
+ */
+static int _loop_table(char *table, size_t tlen, char *file, char *dev, off_t off)
+{
+	struct stat fbuf;
+	off_t size, sectors;
+	int fd = -1;
+#ifdef HAVE_SYS_STATVFS_H
+	struct statvfs fsbuf;
+	off_t blksize;
+#endif
+
+	if (!_switches[READ_ONLY])
+		fd = open(file, O_RDWR);
+
+	if (fd < 0) {
+		_switches[READ_ONLY]++;
+		fd = open(file, O_RDONLY);
+	}
+
+	if (fd < 0)
+		goto error;
+
+	if (fstat(fd, &fbuf))
+		goto error;
+
+	size = (fbuf.st_size - off);
+	sectors = size >> SECTOR_SHIFT;
+
+	if (_switches[VERBOSE_ARG])
+		fprintf(stderr, "losetup: set loop size to %llukB (%llu sectors)\n",
+			sectors >> 1, sectors);
+
+#ifdef HAVE_SYS_STATVFS_H
+	if (fstatvfs(fd, &fsbuf))
+		goto error;       
+
+	/* FIXME Fragment size currently unused */
+	blksize = fsbuf.f_frsize;
+#endif
+
+	close(fd);
+
+	if (dm_snprintf(table, tlen, "%llu %llu loop %s %llu\n", 0ULL, 
+			(long long unsigned)sectors, file, off) < 0)
+		return 0;
+
+	if (_switches[VERBOSE_ARG] > 1)
+		fprintf(stderr, "Table: %s\n", table);
+
+	return 1;
+
+error:
+	if (fd > -1)
+		close(fd);
+	return 0;
+}
+
+static int _process_losetup_switches(const char *base, int *argc, char ***argv)
+{
+	static int ind;
+	int c;
+	int encrypt_loop = 0, delete = 0, find = 0, show_all = 0;
+	char *device_name = NULL;
+	char *loop_file = NULL;
+	off_t offset = 0;
+
+#ifdef HAVE_GETOPTLONG
+	static struct option long_options[] = {
+		{0, 0, 0, 0}
+	};
+#endif
+
+	optarg = 0;
+	optind = OPTIND_INIT;
+	while ((ind = -1, c = GETOPTLONG_FN(*argc, *argv, "ade:fo:v",
+					    long_options, NULL)) != -1 ) {
+		if (c == ':' || c == '?')
+			return 0;
+		if (c == 'a')
+			show_all++;
+		if (c == 'd')
+			delete++;
+		if (c == 'e')
+			encrypt_loop++;
+		if (c == 'f')
+			find++;
+		if (c == 'o')
+			offset = atoi(optarg);
+		if (c == 'v')
+			_switches[VERBOSE_ARG]++;
+	}
+
+	*argv += optind ;
+	*argc -= optind ;
+
+	if (encrypt_loop){
+		fprintf(stderr, "%s: Sorry, cryptoloop is not yet implemented "
+				"in this version.\n", base);
+		return 0;
+	}
+
+	if (show_all) {
+		fprintf(stderr, "%s: Sorry, show all is not yet implemented "
+				"in this version.\n", base);
+		return 0;
+	}
+
+	if (find) {
+		fprintf(stderr, "%s: Sorry, find is not yet implemented "
+				"in this version.\n", base);
+		if (!*argc)
+			return 0;
+	}
+
+	if (!*argc) {
+		fprintf(stderr, "%s: Please specify loop_device.\n", base);
+		_losetup_usage(stderr);
+		return 0;
+	}
+
+	if (!(device_name = parse_loop_device_name((*argv)[0]))) {
+		fprintf(stderr, "%s: Could not parse loop_device %s\n",
+			base, (*argv)[0]);
+		_losetup_usage(stderr);
+		return 0;
+	}
+
+	if (delete) {
+		*argc = 2;
+
+		(*argv)[1] = device_name;
+		(*argv)[0] = (char *) "remove";
+
+		return 1;
+	}
+
+	if (*argc != 2) {
+		fprintf(stderr, "%s: Too few arguments\n", base);
+		_losetup_usage(stderr);
+		return 0;
+	}
+
+	/* FIXME move these to make them available to native dmsetup */
+	if (!(loop_file = _get_abspath((*argv)[(find) ? 0 : 1]))) {
+		fprintf(stderr, "%s: Could not parse loop file name %s\n",
+			base, (*argv)[1]);
+		_losetup_usage(stderr);
+		return 0;
+	}
+
+	/* FIXME Missing free */
+	_table = dm_malloc(LOOP_TABLE_SIZE);
+	if (!_loop_table(_table, LOOP_TABLE_SIZE, loop_file, device_name, offset)) {
+		fprintf(stderr, "Could not build device-mapper table for %s\n", (*argv)[0]);
+		return 0;
+	}
+	_switches[TABLE_ARG]++;
+
+	(*argv)[0] = (char *) "create";
+	(*argv)[1] = device_name ;
+
+	return 1;
+}
+
 static int _process_switches(int *argc, char ***argv)
 {
 	char *base, *namebase;
 	static int ind;
-	int c;
+	int c, r;
 
 #ifdef HAVE_GETOPTLONG
 	static struct option long_options[] = {
@@ -1622,6 +1862,7 @@ static int _process_switches(int *argc, char ***argv)
 		{"major", 1, &ind, MAJOR_ARG},
 		{"minor", 1, &ind, MINOR_ARG},
 		{"mode", 1, &ind, MODE_ARG},
+		{"noflush", 0, &ind, NOFLUSH_ARG},
 		{"noheadings", 0, &ind, NOHEADINGS_ARG},
 		{"nolockfs", 0, &ind, NOLOCKFS_ARG},
 		{"noopencount", 0, &ind, NOOPENCOUNT_ARG},
@@ -1678,6 +1919,12 @@ static int _process_switches(int *argc, char ***argv)
 		return 1;
 	}
 
+	if(!strcmp(base, "losetup") || !strcmp(base, "dmlosetup")){
+		r = _process_losetup_switches(base, argc, argv);
+		free(namebase);
+		return r;
+	}
+
 	free(namebase);
 
 	optarg = 0;
@@ -1733,6 +1980,8 @@ static int _process_switches(int *argc, char ***argv)
 			_switches[TARGET_ARG]++;
 			_target = optarg;
 		}
+		if ((ind == NOFLUSH_ARG))
+			_switches[NOFLUSH_ARG]++;
 		if ((ind == NOHEADINGS_ARG))
 			_switches[NOHEADINGS_ARG]++;
 		if ((ind == NOLOCKFS_ARG))

tools/lvconvert.c

@@ -102,7 +102,7 @@ static int _read_params(struct lvconvert_params *lp, struct cmd_context *cmd,
 			int argc, char **argv)
 {
 	int region_size;
-	int pagesize = getpagesize();
+	int pagesize = lvm_getpagesize();
 
 	memset(lp, 0, sizeof(*lp));
 
@@ -344,8 +344,7 @@ static int lvconvert_mirrors(struct cmd_context * cmd, struct logical_volume * l
 						  lp->mirrors - 1);
 					return 1;
 				}
-			}
-			if (lp->mirrors > existing_mirrors) {
+			} else if (lp->mirrors > existing_mirrors) {
 				/* FIXME Unless anywhere, remove PV of log_lv 
 				 * from allocatable_pvs & allocate 
 				 * (mirrors - existing_mirrors) new areas
@@ -565,6 +564,12 @@ int lvconvert(struct cmd_context * cmd, int argc, char **argv)
 		goto error;
 	}
 
+	if ((vg->status & CLUSTERED) && !locking_is_clustered() &&
+	    !lockingfailed()) {
+		log_error("Skipping clustered volume group %s", lp.vg_name);
+		goto error;
+	}
+
 	if (vg->status & EXPORTED_VG) {
 		log_error("Volume group \"%s\" is exported", lp.vg_name);
 		goto error;

tools/lvcreate.c

@@ -43,6 +43,7 @@ struct lvcreate_params {
 	/* size */
 	uint32_t extents;
 	uint64_t size;
+	percent_t percent;
 
 	uint32_t permission;
 	uint32_t read_ahead;
@@ -94,19 +95,7 @@ static int _lvcreate_name_params(struct lvcreate_params *lp,
 			}
 
 		} else {
-			vg_name = argv[0];
-			/* Strip dev_dir (optional) */
-			if (*vg_name == '/') {
-				while (*vg_name == '/')
-					vg_name++;
-				vg_name--;
-			}
-			if (!strncmp(vg_name, cmd->dev_dir,
-				     strlen(cmd->dev_dir))) {
-				vg_name += strlen(cmd->dev_dir);
-				while (*vg_name == '/')
-					vg_name++;
-			}
+			vg_name = skip_dev_dir(cmd, argv[0]);
 			if (strrchr(vg_name, '/')) {
 				log_error("Volume group name expected "
 					  "(no slash)");
@@ -169,6 +158,7 @@ static int _read_size_params(struct lvcreate_params *lp,
 			return 0;
 		}
 		lp->extents = arg_uint_value(cmd, extents_ARG, 0);
+		lp->percent = arg_percent_value(cmd, extents_ARG, PERCENT_NONE);
 	}
 
 	/* Size returned in kilobyte units; held in sectors */
@@ -178,6 +168,7 @@ static int _read_size_params(struct lvcreate_params *lp,
 			return 0;
 		}
 		lp->size = arg_uint64_value(cmd, size_ARG, UINT64_C(0)) * 2;
+		lp->percent = PERCENT_NONE;
 	}
 
 	return 1;
@@ -249,7 +240,7 @@ static int _read_mirror_params(struct lvcreate_params *lp,
 {
 	int argc = *pargc;
 	int region_size;
-	int pagesize = getpagesize();
+	int pagesize = lvm_getpagesize();
 
 	if (argc && (unsigned) argc < lp->mirrors) {
 		log_error("Too few physical volumes on "
@@ -498,6 +489,12 @@ static int _lvcreate(struct cmd_context *cmd, struct lvcreate_params *lp)
 		return 0;
 	}
 
+	if ((vg->status & CLUSTERED) && !locking_is_clustered() &&
+	    !lockingfailed()) {
+		log_error("Skipping clustered volume group %s", lp->vg_name);
+		return 0;
+	}
+
 	if (vg->status & EXPORTED_VG) {
 		log_error("Volume group \"%s\" is exported", lp->vg_name);
 		return 0;
@@ -562,6 +559,20 @@ static int _lvcreate(struct cmd_context *cmd, struct lvcreate_params *lp)
 		lp->extents = tmp_size / vg->extent_size;
 	}
 
+	switch(lp->percent) {
+		case PERCENT_VG:
+			lp->extents = lp->extents * vg->extent_count / 100;
+			break;
+		case PERCENT_FREE:
+			lp->extents = lp->extents * vg->free_count / 100;
+			break;
+		case PERCENT_LV:
+			log_error("Please express size as %%VG or %%FREE.");
+			return 0;
+		case PERCENT_NONE:
+			break;
+	}
+
 	if ((size_rest = lp->extents % lp->stripes)) {
 		log_print("Rounding size (%d extents) up to stripe boundary "
 			  "size (%d extents)", lp->extents,