lvmcache: simplify metadata cache

The copy of VG metadata stored in lvmcache was not being used in general. It pretended to be a generic VG metadata cache, but was not being used except for clvmd activation. There it was used to avoid reading from disk while devices were suspended, i.e. in resume. This removes the code that attempted to make this look like a generic metadata cache, and replaces with with something narrowly targetted to what it's actually used for. This is a way of passing the VG from suspend to resume in clvmd. Since in the case of clvmd one caller can't simply pass the same VG to both suspend and resume, suspend needs to stash the VG somewhere that resume can grab it from. (resume doesn't want to read it from disk since devices are suspended.) The lvmcache vginfo struct is used as a convenient place to stash the VG to pass it from suspend to resume, even though it isn't related to the lvmcache or vginfo. These suspended_vg* vginfo fields should not be used or touched anywhere else, they are only to be used for passing the VG data from suspend to resume in clvmd. The VG data being passed between suspend and resume is never modified, and will only exist in the brief period between suspend and resume in clvmd. suspend has both old (current) and new (precommitted) copies of the VG metadata. It stashes both of these in the vginfo prior to suspending devices. When vg_commit is successful, it sets a flag in vginfo as before, signaling the transition from old to new metadata. resume grabs the VG stashed by suspend. If the vg_commit happened, it grabs the new VG, and if the vg_commit didn't happen it grabs the old VG. The VG is then used to resume LVs. This isolates clvmd-specific code and usage from the normal lvm vg_read code, making the code simpler and the behavior easier to verify. Sequence of operations: - lv_suspend() has both vg_old and vg_new and stashes a copy of each onto the vginfo: lvmcache_save_suspended_vg(vg_old); lvmcache_save_suspended_vg(vg_new); - vg_commit() happens, which causes all clvmd instances to call lvmcache_commit_metadata(vg). A flag is set in the vginfo indicating the transition from the old to new VG: vginfo->suspended_vg_committed = 1; - lv_resume() needs either vg_old or vg_new to use in resuming LVs. It doesn't want to read the VG from disk since devices are suspended, so it gets the VG stashed by lv_suspend: vg = lvmcache_get_suspended_vg(vgid); If the vg_commit did not happen, suspended_vg_committed will not be set, and in this case, lvmcache_get_suspended_vg() will return the old VG instead of the new VG, and it will resume LVs based on the old metadata.
vgcreate: improve the use of label_scan
2025-09-17 21:44:24 +03:00 · 2017-11-01 16:39:35 -05:00 · 2017-11-01 16:39:35 -05:00 · 2017-11-01 16:39:35 -05:00 · 2017-11-01 16:39:35 -05:00 · 2017-11-01 16:39:35 -05:00
975 changed files with 61199 additions and 70274 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -1,7 +1,6 @@
 *.5
 *.7
 *.8
-*.8_gen
 *.a
 *.d
 *.o
@@ -25,118 +24,9 @@ make.tmpl

 /autom4te.cache/
 /autoscan.log
-/build/
-/config.cache
 /config.log
 /config.status
 /configure.scan
 /cscope.out
-/html/
-/reports/
 /tags
 /tmp/
-
-coverity/coverity_model.xml
-
-tools/man-generator
-tools/man-generator.c
-
-test/.lib-dir-stamp
-test/.tests-stamp
-test/lib/dmsecuretest
-test/lib/lvchange
-test/lib/lvconvert
-test/lib/lvcreate
-test/lib/lvdisplay
-test/lib/lvextend
-test/lib/lvmconfig
-test/lib/lvmdiskscan
-test/lib/lvmsadc
-test/lib/lvmsar
-test/lib/lvreduce
-test/lib/lvremove
-test/lib/lvrename
-test/lib/lvresize
-test/lib/lvs
-test/lib/lvscan
-test/lib/pvchange
-test/lib/pvck
-test/lib/pvcreate
-test/lib/pvdisplay
-test/lib/pvmove
-test/lib/pvremove
-test/lib/pvresize
-test/lib/pvs
-test/lib/pvscan
-test/lib/securetest
-test/lib/vgcfgbackup
-test/lib/vgcfgrestore
-test/lib/vgchange
-test/lib/vgck
-test/lib/vgconvert
-test/lib/vgcreate
-test/lib/vgdisplay
-test/lib/vgexport
-test/lib/vgextend
-test/lib/vgimport
-test/lib/vgimportclone
-test/lib/vgmerge
-test/lib/vgmknodes
-test/lib/vgreduce
-test/lib/vgremove
-test/lib/vgrename
-test/lib/vgs
-test/lib/vgscan
-test/lib/vgsplit
-test/api/lvtest.t
-test/api/pe_start.t
-test/api/percent.t
-test/api/python_lvm_unit.py
-test/api/test
-test/api/thin_percent.t
-test/api/vglist.t
-test/api/vgtest.t
-test/lib/aux
-test/lib/check
-test/lib/clvmd
-test/lib/dm-version-expected
-test/lib/dmeventd
-test/lib/dmsetup
-test/lib/dmstats
-test/lib/fail
-test/lib/flavour-ndev-cluster
-test/lib/flavour-ndev-cluster-lvmpolld
-test/lib/flavour-ndev-lvmetad
-test/lib/flavour-ndev-lvmetad-lvmpolld
-test/lib/flavour-ndev-lvmpolld
-test/lib/flavour-ndev-vanilla
-test/lib/flavour-udev-cluster
-test/lib/flavour-udev-cluster-lvmpolld
-test/lib/flavour-udev-lvmetad
-test/lib/flavour-udev-lvmetad-lvmpolld
-test/lib/flavour-udev-lvmlockd-dlm
-test/lib/flavour-udev-lvmlockd-sanlock
-test/lib/flavour-udev-lvmlockd-test
-test/lib/flavour-udev-lvmpolld
-test/lib/flavour-udev-vanilla
-test/lib/fsadm
-test/lib/get
-test/lib/inittest
-test/lib/invalid
-test/lib/lvm
-test/lib/lvm-wrapper
-test/lib/lvmchange
-test/lib/lvmdbusd.profile
-test/lib/lvmetad
-test/lib/lvmpolld
-test/lib/not
-test/lib/paths
-test/lib/paths-common
-test/lib/runner
-test/lib/should
-test/lib/test
-test/lib/thin-performance.profile
-test/lib/utils
-test/lib/version-expected
-test/unit/dmraid_t.c
-test/unit/unit-test
--- a/COPYING.BSD
+++ b/COPYING.BSD
@@ -1,25 +0,0 @@
-BSD 2-Clause License
-
-Copyright (c) 2014, Red Hat, Inc.
-All rights reserved.
-
-Redistribution and use in source and binary forms, with or without
-modification, are permitted provided that the following conditions are met:
-
-1. Redistributions of source code must retain the above copyright notice, this
-   list of conditions and the following disclaimer.
-
-2. Redistributions in binary form must reproduce the above copyright notice,
-   this list of conditions and the following disclaimer in the documentation
-   and/or other materials provided with the distribution.
-
-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
-ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
-WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
-DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
-FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
-DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
-SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
-CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
-OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
--- a/Makefile.in
+++ b/Makefile.in
@@ -1,6 +1,6 @@
 #
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
-# Copyright (C) 2004-2018 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2015 Red Hat, Inc. All rights reserved.
 #
 # This file is part of LVM2.
 #
@@ -28,6 +28,14 @@ ifeq ("@INTL@", "yes")
  SUBDIRS += po
 endif

+ifeq ("@APPLIB@", "yes")
+  SUBDIRS += liblvm
+endif
+
+ifeq ("@PYTHON_BINDINGS@", "yes")
+  SUBDIRS += python
+endif
+
 ifeq ($(MAKECMDGOALS),clean)
  SUBDIRS += test
 endif
@@ -35,7 +43,8 @@ endif
 ifeq ($(MAKECMDGOALS),distclean)
  SUBDIRS = conf include man test scripts \
    libdaemon lib tools daemons libdm \
-    udev po
+    udev po liblvm python \
+    unit-tests/datastruct unit-tests/mm unit-tests/regex
 tools.distclean: test.distclean
 endif
 DISTCLEAN_DIRS += lcov_reports*
@@ -43,25 +52,20 @@ DISTCLEAN_TARGETS += config.cache config.log config.status make.tmpl

 include make.tmpl

-include $(top_srcdir)/base/Makefile
-include $(top_srcdir)/device_mapper/Makefile
-include $(top_srcdir)/test/unit/Makefile
-
 libdm: include
 libdaemon: include
-lib: libdaemon $(BASE_TARGET) $(DEVICE_MAPPER_TARGET)
+lib: libdm libdaemon
+liblvm: lib
 daemons: lib libdaemon tools
-scripts: lib
-tools: lib libdaemon
+tools: lib libdaemon device-mapper
 po: tools daemons
 man: tools
 all_man: tools
-scripts: libdm
-test: tools daemons
-unit-test  run-unit-test: test
+scripts: liblvm libdm

 lib.device-mapper: include.device-mapper
 libdm.device-mapper: include.device-mapper
+liblvm.device-mapper: include.device-mapper
 daemons.device-mapper: libdm.device-mapper
 tools.device-mapper: libdm.device-mapper
 scripts.device-mapper: include.device-mapper
@@ -75,6 +79,10 @@ po.pofile: tools.pofile daemons.pofile
 pofile: po.pofile
 endif

+ifeq ("@PYTHON_BINDINGS@", "yes")
+python: liblvm
+endif
+
 ifneq ("$(CFLOW_CMD)", "")
 tools.cflow: libdm.cflow lib.cflow
 daemons.cflow: tools.cflow
@@ -89,7 +97,7 @@ endif
 DISTCLEAN_TARGETS += cscope.out
 CLEAN_DIRS += autom4te.cache

-check check_system check_cluster check_local check_lvmpolld check_lvmlockd_test check_lvmlockd_dlm check_lvmlockd_sanlock: test
+check check_system check_cluster check_local check_lvmetad check_lvmpolld check_lvmlockd_test check_lvmlockd_dlm check_lvmlockd_sanlock unit: all
 	$(MAKE) -C test $(@)

 conf.generate man.generate: tools
@@ -138,7 +146,7 @@ install_system_dirs:
 	$(INSTALL_ROOT_DIR) $(DESTDIR)$(DEFAULT_RUN_DIR)
 	$(INSTALL_ROOT_DATA) /dev/null $(DESTDIR)$(DEFAULT_CACHE_DIR)/.cache

-install_initscripts:
+install_initscripts: 
 	$(MAKE) -C scripts install_initscripts

 install_systemd_generators:
@@ -151,11 +159,26 @@ install_systemd_units:
 install_all_man:
 	$(MAKE) -C man install_all_man

+ifeq ("@PYTHON_BINDINGS@", "yes")
+install_python_bindings:
+	$(MAKE) -C liblvm/python install_python_bindings
+endif
+
 install_tmpfiles_configuration:
 	$(MAKE) -C scripts install_tmpfiles_configuration

+LCOV_TRACES = libdm.info lib.info liblvm.info tools.info \
+	libdaemon/client.info libdaemon/server.info \
+	daemons/clvmd.info \
+	daemons/dmeventd.info \
+	daemons/lvmetad.info \
+	daemons/lvmlockd.info \
+	daemons/lvmpolld.info
+
+CLEAN_TARGETS += $(LCOV_TRACES)
+
 ifneq ("$(LCOV)", "")
-.PHONY: lcov-reset lcov lcov-dated
+.PHONY: lcov-reset lcov lcov-dated $(LCOV_TRACES)

 ifeq ($(MAKECMDGOALS),lcov-dated)
 LCOV_REPORTS_DIR := lcov_reports-$(shell date +%Y%m%d%k%M%S)
@@ -165,22 +188,55 @@ LCOV_REPORTS_DIR := lcov_reports
 endif

 lcov-reset:
-	$(LCOV) --zerocounters --directory $(top_builddir)
+	$(LCOV) --zerocounters $(addprefix -d , $(basename $(LCOV_TRACES)))
+
+# maybe use subdirs processing to create tracefiles...
+$(LCOV_TRACES):
+	$(LCOV) -b $(basename $@) -d $(basename $@) \
+		--ignore-errors source -c -o - | $(SED) \
+		-e "s/\(dmeventd_lvm.[ch]\)/plugins\/lvm2\/\1/" \
+		-e "s/dmeventd_\(mirror\|snapshot\|thin\|raid\)\.c/plugins\/\1\/dmeventd_\1\.c/" \
+		>$@

 ifneq ("$(GENHTML)", "")
-lcov:
-	$(RM) -rf $(LCOV_REPORTS_DIR)
+lcov: $(LCOV_TRACES)
+	$(RM) -r $(LCOV_REPORTS_DIR)
 	$(MKDIR_P) $(LCOV_REPORTS_DIR)
-	$(LCOV) --capture --directory $(top_builddir) --ignore-errors source \
-		--output-file $(LCOV_REPORTS_DIR)/out.info
-	-test ! -s $(LCOV_REPORTS_DIR)/out.info || \
-		$(GENHTML) -o $(LCOV_REPORTS_DIR) --ignore-errors source \
-		$(LCOV_REPORTS_DIR)/out.info
+	for i in $(LCOV_TRACES); do \
+		test -s $$i -a $$(wc -w <$$i) -ge 100 && lc="$$lc $$i"; \
+	done; \
+	test -z "$$lc" || $(GENHTML) -p @abs_top_builddir@ \
+		-o $(LCOV_REPORTS_DIR) $$lc
 endif

 endif

-ifneq ($(shell which ctags 2>/dev/null),)
+ifeq ("$(TESTING)", "yes")
+# testing and report generation
+RUBY=ruby1.9 -Ireport-generators/lib -Ireport-generators/test
+
+.PHONY: unit-test ruby-test test-programs
+
+# FIXME: put dependencies on libdm and liblvm
+# FIXME: Should be handled by Makefiles in subdirs, not here at top level.
+test-programs:
+	cd unit-tests/regex && $(MAKE)
+	cd unit-tests/datastruct && $(MAKE)
+	cd unit-tests/mm && $(MAKE)
+
+unit-test: test-programs
+	$(RUBY) report-generators/unit_test.rb $(shell find . -name TESTS)
+	$(RUBY) report-generators/title_page.rb
+
+memcheck: test-programs
+	$(RUBY) report-generators/memcheck.rb $(shell find . -name TESTS)
+	$(RUBY) report-generators/title_page.rb
+
+ruby-test:
+	$(RUBY) report-generators/test/ts.rb
+endif
+
+ifneq ($(shell which ctags),)
 .PHONY: tags
 tags:
 	test -z "$(shell find $(top_srcdir) -type f -name '*.[ch]' -newer tags 2>/dev/null | head -1)" || $(RM) tags
--- a/3
+++ b/3
@@ -1,8 +1,5 @@
 This tree contains the LVM2 and device-mapper tools and libraries.

-This is development branch, for stable 2.02 release see 2018-06-01-stable
-branch.
-
 For more information about LVM2 read the changelog in the WHATS_NEW file.
 Installation instructions are in INSTALL.

--- a/62
+++ b/62
@@ -1,62 +0,0 @@
-LVM2 Test Suite
-===============
-
-The codebase contains many tests in the test subdirectory.
-
-Before running tests
--------------------
-
-Keep in mind the testsuite MUST run under root user.
-
-It is recommended not to use LVM on the test machine, especially when running
-tests with udev (`make check_system`.)
-
-You MUST disable (or mask) any LVM daemons:
-
- lvmetad
- dmeventd
- lvmpolld
- lvmdbusd
- lvmlockd
- clvmd
- cmirrord
-
-For running cluster tests, we are using singlenode locking. Pass
-`--with-clvmd=singlenode` to configure.
-
-NOTE: This is useful only for testing, and should not be used in produciton
-code.
-
-To run D-Bus daemon tests, existing D-Bus session is required.
-
-Running tests
-------------
-
-As root run:
-
-    make check
-
-To run only tests matching a string:
-
-    make check T=test
-
-To skip tests matching a string:
-
-    make check S=test
-
-There are other targets and many environment variables can be used to tweak the
-testsuite - for full list and description run `make -C test help`.
-
-Installing testsuite
--------------------
-
-It is possible to install and run a testsuite against installed LVM. Run the
-following:
-
-    make -C test install
-
-Then lvm2-testsuite binary can be executed to test installed binaries.
-
-See `lvm2-testsuite --help` for options. The same environment variables can be
-used as with `make check`.
-
--- a/2
+++ b/2
@@ -1 +1 @@
-2.03.05(2)-git (2019-06-10)
+2.02.176(2)-git (2017-10-06)
--- a/2
+++ b/2
@@ -1 +1 @@
-1.02.163-git (2019-06-10)
+1.02.145-git (2017-10-06)
--- a/215
+++ b/215
@@ -1,218 +1,5 @@
-Version 2.03.05 - 
-================================
-  Fix command definition for pvchange -a.
-  Add vgck --updatemetadata command that will repair metadata problems.
-  Improve VG reading to work if one good copy of metadata is found.
-  Report/display/scan commands that read VGs will no longer write/repair.
-  Move metadata repairs from VG reading to VG writing.
-  Add config setting md_component_checks to control MD component checks.
-  Add end of device MD component checks when dev has no udev info.
-
-Version 2.03.04 - 10th June 2019
-================================
-  Remove unused_duplicate_devs from cmd causing segfault in dmeventd.
-
-Version 2.03.03 - 07th June 2019
-================================
-  Report no_discard_passdown for cache LVs with lvs -o+kernel_discards.
-  Add pvck --dump option to extract metadata.
-  Fix signal delivery checking race in libdaemon (lvmetad).
-  Add missing Before=shutdown.target to LVM2 services to fix shutdown ordering.
-  Skip autoactivation for a PV when PV size does not match device size.
-  Remove first-pvscan-initialization which should no longer be needed.
-  Add remote refresh through lvmlockd/dlm for shared LVs after lvextend.
-  Ignore foreign and shared PVs for pvscan online files.
-  Add config setting to control fields in debug file and verbose output.
-  Add command[pid] and timestamp to debug file and verbose output.
-  Fix missing growth of _pmsmare volume when extending _tmeta volume.
-  Automatically grow thin metadata, when thin data gets too big.
-  Add synchronization with udev before removing cached devices.
-  Add support for caching VDO LVs and VDOPOOL LVs.
-  Add support for vgsplit with cached devices.
-  Query mpath device only once per command for its state.
-  Use device INFO instead of STATUS when checking for mpath device uuid.
-  Change default io_memory_size from 4 to 8 MiB.
-  Add config setting io_memory_size to set bcache size.
-  Fix pvscan autoactivation for concurrent pvscans.
-  Change scan_lvs default to 0 so LVs are not scanned for PVs.
-  Thin-pool selects power-of-2 chunk size by default.
-  Cache selects power-of-2 chunk size by default.
-  Support reszing for VDOPoolLV and VDOLV.
-  Improve -lXXX%VG modifier which improves cache segment estimation.
-  Ensure migration_threshold for cache is at least 8 chunks.
-  Restore missing man info lvcreate --zero for thin-pools.
-  Drop misleadning comment for metadata minimum_io_size for VDO segment.
-  Add device hints to reduce scanning.
-  Introduce LVM_SUPPRESS_SYSLOG to suppress syslog usage by generator.
-  Fix generator quering lvmconfig unpresent config option.
-  Fix memleak on bcache error path code.
-  Fix missing unlock on lvm2 dmeventd plugin error path initialization.
-  Improve Makefile dependency tracking.
-  Move VDO support towards V2 target (6.2) support.
-
-Version 2.03.02 - 18th December 2018
-====================================
-  Fix missing proper initialization of pv_list struct when adding pv.
-  Fix (de)activation of RaidLVs with visible SubLVs.
-  Prohibit mirrored 'mirror' log via lvcreate and lvconvert.
-  Use sync io if async io_setup fails, or use_aio=0 is set in config.
-  Fix more issues reported by coverity scan.
-
-Version 2.03.01 - 31st October 2018
+Version 2.02.176 -
 ===================================
-
-Version 2.03.00 - 10th October 2018
-===================================
-  Add hot fix to avoiding locking collision when monitoring thin-pools.
-  Allow raid4 -> linear conversion request.
-  Fix lvconvert striped/raid0/raid0_meta -> raid6 regression.
-  Add 'lvm2-activation-generator:' prefix for kmsg messages logged by generator.
-  Add After=rbdmap.service to {lvm2-activation-net,blk-availability}.service.
-  Reduce max concurrent aios to avoid EMFILE with many devices.
-  Fix lvconvert conversion attempts to linear.
-  Fix lvconvert raid0/raid0_meta -> striped regression.
-  Fix lvconvert --splitmirror for mirror type (2.02.178).
-  Do not pair cache policy and cache metadata format.
-  lvconvert: reject conversions on raid1 LVs with split tracked SubLVs
-  lvconvert: reject conversions on raid1 split tracked SubLVs
-  Add basic creation support for VDO target.
-  Never send any discard ioctl with test mode.
-  Fix thin-pool alloc which needs same PV for data and metadata.
-  Extend list of non-memlocked areas with newly linked libs.
-  Enhance vgcfgrestore to check for active LVs in restored VG.
-  Configure supports --disable-silent-rules for verbose builds.
-  Fix unmonitoring of merging snapshots.
-  Cache can uses metadata format 2 with cleaner policy.
-  Fix check if resized PV can also fit metadata area.
-  Avoid showing internal error in lvs output or pvmoved LVs.
-  Remove clvmd
-  Remove lvmlib (api)
-  Remove lvmetad
-  lvconvert: provide possible layouts between linear and striped/raid
-  Use versionsort to fix archive file expiry beyond 100000 files.
-
-Version 2.02.178-rc1 - 24th May 2018
-====================================
-  Add libaio dependency for build.
-  Remove lvm1 and pool format handling and add filter to ignore them.
-  Move some filter checks to after disks are read.
-  Rework disk scanning and when it is used.
-  Add new io layer and shift code to using it.
-  Fix lvconvert's return code on degraded -m raid1 conversion.
-  --enable-testing switch for ./configure has been removed.
-  --with-snapshots switch for ./configure has been removed.
-  --with-mirrors switch for ./configure has been removed.
-  --with-raid switch for ./configure has been removed.
-  --with-thin switch for ./configure has been removed.
-  --with-cache switch for ./configure has been removed.
-  Include new unit-test framework and unit tests.
-  Extend validation of region_size for mirror segment.
-  Reload whole device stack when reinitilizing mirror log.
-  Mirrors without monitoring are WARNING and not blocking on error.
-  Detect too big region_size with clustered mirrors.
-  Fix evaluation of maximal region size for mirror log.
-  Enhance mirror log size estimation and use smaller size when possible.
-  Fix incorrect mirror log size calculation on 32bit arch.
-  Enhance preloading tree creating.
-  Fix regression on acceptance of any LV on lvconvert.
-  Restore usability of thin LV to be again external origin for another thin.
-  Keep systemd vars on change event in 69-dm-lvm-metad.rules for systemd reload.
-  Write systemd and non-systemd rule in 69-dm-lvm-metad.rules, GOTO active one.
-  Add test for activation/volume_list (Sub)LV remnants.
-  Disallow usage of cache format 2 with mq cache policy.
-  Again accept striped LV as COW LV with lvconvert -s (2.02.169).
-  Fix raid target version testing for supported features.
-  Allow activation of pools when thin/cache_check tool is missing.
-  Remove RaidLV on creation failure when rmeta devices can't be activated.
-  Add prioritized_section() to restore cookie boundaries (2.02.177).
-  Enhance error messages when read error happens.
-  Enhance mirror log initialization for old mirror target.
-  Skip private crypto and stratis devices.
-  Skip frozen raid devices from scanning.
-  Activate RAID SubLVs on read_only_volume_list readwrite.
-  Offer convenience type raid5_n converting to raid10.
-  Automatically avoid reading invalid snapshots during device scan.
-  Ensure COW device is writable even for read-only thick snapshots.
-  Support activation of component LVs in read-only mode.
-  Extend internal library to recognize and work with component LV.
-  Skip duplicate check for active LV when prompting for its removal.
-  Activate correct lock holding LV when it is cached.
-  Do not modify archived metadata when removing striped raid.
-  Fix memleak on error path when obtaining lv_raid_data_offset.
-  Fix compatibility size test of extended external origin.
-  Add external_origin visiting in for_each_sub_lv().
-  Ensure cluster commands drop their device cache before locking VG.
-  Do not report LV as remotely active when it's locally exclusive in cluster.
-  Add deprecate messages for usage of mirrors with mirrorlog.
-  Separate reporting of monitoring status and error status.
-  Improve validation of created strings in vgimportclone.
-  Add missing initialisation of mem pool in systemd generator.
-  Do not reopen output streams for multithreaded users of liblvm.
-  Configure ensures /usr/bin dir is checked for dmpd tools.
-  Restore pvmove support for wide-clustered active volumes (2.02.177).
-  Avoid non-exclusive activation of exclusive segment types.
-  Fix trimming sibling PVs when doing a pvmove of raid subLVs.
-  Preserve exclusive activation during thin snaphost merge.
-  Avoid exceeding array bounds in allocation tag processing.
-  Add --lockopt to common options and add option to skip selected locks.
-
-Version 2.02.177 - 18th December 2017
-=====================================
-  When writing text metadata content, use complete 4096 byte blocks.
-  Change text format metadata alignment from 512 to 4096 bytes.
-  When writing metadata, consistently skip mdas marked as failed.
-  Refactor and adjust text format metadata alignment calculation.
-  Fix python3 path in lvmdbusd to use value detected by configure.
-  Reduce checks for active LVs in vgchange before background polling.
-  Ensure _node_send_message always uses clean status of thin pool.
-  Fix lvmlockd to use pool lock when accessing _tmeta volume.
-  Report expected sanlock_convert errors only when retries fail.
-  Avoid blocking in sanlock_convert on SH to EX lock conversion.
-  Deactivate missing raid LV legs (_rimage_X-missing_Y_Z) on decativation.
-  Skip read-modify-write when entire block is replaced.
-  Categorise I/O with reason annotations in debug messages.
-  Allow extending of raid LVs created with --nosync after a failed repair.
-  Command will lock memory only when suspending volumes.
-  Merge segments when pvmove is finished.
-  Remove label_verify that has never been used.
-  Ensure very large numbers used as arguments are not casted to lower values. 
-  Enhance reading and validation of options stripes and stripes_size.
-  Fix printing of default stripe size when user is not using stripes.
-  Activation code for pvmove automatically discovers holding LVs for resume.
-  Make a pvmove LV locking holder.
-  Do not change critical section counter on resume path without real resume.
-  Enhance activation code to automatically suspend pvmove participants.
-  Prevent conversion of thin volumes to snapshot origin when lvmlockd is used.
-  Correct the steps to change lock type in lvmlockd man page.
-  Retry lock acquisition on recognized sanlock errors.
-  Fix lock manager error codes in lvmlockd.
-  Remove unnecessary single read from lvmdiskscan.
-  Check raid reshape flags in vg_validate().
-  Add support for pvmove of cache and snapshot origins.
-  Avoid using precommitted metadata for suspending pvmove tree.
-  Ehnance pvmove locking.
-  Deactivate activated LVs on error path when pvmove activation fails.
-  Add "io" to log/debug_classes for logging low-level I/O.
-  Eliminate redundant nested VG metadata in VG struct.
-  Avoid importing persistent filter in vgscan/pvscan/vgrename.
-  Fix memleak of string buffer when vgcfgbackup runs in secure mode.
-  Do not print error when clvmd cannot find running clvmd.
-  Prevent start of new merge of snapshot if origin is already being merged.
-  Fix offered type for raid6_n_6 to raid5 conversion (raid5_n).
-  Deactivate sub LVs when removing unused cache-pool.
-  Do not take backup with suspended devices.
-  Avoid RAID4 activation on incompatible kernels under all circumstances.
-  Reject conversion request to striped/raid0 on 2-legged raid4/5.
-
-Version 2.02.176 - 3rd November 2017
-====================================
-  Keep Install section only in lvm2-{lvmetad,lvmpolld}.socket systemd unit.
-  Fix segfault in lvm_pv_remove in liblvm. (2.02.173)
-  Do not allow storing VG metadata with LV without any segment.
-  Fix printed message when thin snapshot was already merged.
-  Remove created spare LV when creation of thin-pool failed.
-  Avoid reading ignored metadata when mda gets used again.
-  Fix detection of moved PVs in vgsplit. (2.02.175)
  Ignore --stripes/--stripesize on RAID takeover
  Improve used paths for generated systemd units and init shells.
  Disallow creation of snapshot of mirror/raid subLV (was never supported).
--- a/65
+++ b/65
@@ -1,66 +1,5 @@
-Version 1.02.163 - 
-=================================
-
-Version 1.02.161 - 10th June 2019
-=================================
-
-Version 1.02.159 - 07th June 2019
-=================================
-  Parsing of cache status understand no_discard_passdown.
-  Ensure migration_threshold for cache is at least 8 chunks.
-
-Version 1.02.155 - 18th December 2018
-=====================================
-  Include correct internal header inside libdm list.c.
-  Enhance ioctl flattening and add parameters only when needed.
-  Add DM_DEVICE_ARM_POLL for API completness matching kernel.
-  Do not add parameters for RESUME with DM_DEVICE_CREATE dm task.
-  Fix dmstats report printing no output.
-
-Version 1.02.153 - 31st October 2018
-====================================
-
-Version 1.02.151 - 10th October 2018
-====================================
-  Add hot fix to avoiding locking collision when monitoring thin-pools.
-
-Version 1.02.150 - 01 August 2018
-=================================
-  Add vdo plugin for monitoring VDO devices.
-
-Version 1.02.149 - 19th July 2018
-=================================
-
-Version 1.02.148 - 18th June 2018
-=================================
-
-Version 1.02.147 - 13th June 2018
-=================================
-
-Version 1.02.147-rc1 - 24th May 2018
-====================================
-  Reuse uname() result for mirror target.
-  Recognize also mounted btrfs through dm_device_has_mounted_fs().
-  Add missing log_error() into dm_stats_populate() returning 0.
-  Avoid calling dm_stats_populat() for DM devices without any stats regions.
-  Support DM_DEBUG_WITH_LINE_NUMBERS envvar for debug msg with source:line.
-  Configured command for thin pool threshold handling gets whole environment.
-  Fix tests for failing dm_snprintf() in stats code.
-  Parsing mirror status accepts 'userspace' keyword in status.
-  Introduce dm_malloc_aligned for page alignment of buffers.
-
-Version 1.02.146 - 18th December 2017
-=====================================
-  Activation tree of thin pool skips duplicated check of pool status.
-  Remove code supporting replicator target.
-  Do not ignore failure of _info_by_dev().
-  Propagate delayed resume for pvmove subvolumes.
-  Suppress integrity encryption keys in 'table' output unless --showkeys supplied.
-
-Version 1.02.145 - 3rd November 2017
-====================================
-  Keep Install section only in dm-event.socket systemd unit.
-  Issue a specific error with dmsetup status if device is unknown.
+Version 1.02.145 - 
+===================================
  Fix RT_LIBS reference in generated libdevmapper.pc for pkg-config

 Version 1.02.144 - 6th October 2017
--- a/acinclude.m4
+++ b/acinclude.m4
@@ -155,7 +155,7 @@ AC_DEFUN([AC_TRY_LDFLAGS],
 #   and this notice are preserved.  This file is offered as-is, without any
 #   warranty.

-serial 3
+#serial 3

 AC_DEFUN([AX_GCC_BUILTIN], [
    AS_VAR_PUSHDEF([ac_var], [ax_cv_have_$1])
--- a/aclocal.m4
+++ b/aclocal.m4
@@ -1,6 +1,6 @@
-# generated automatically by aclocal 1.15.1 -*- Autoconf -*-
+# generated automatically by aclocal 1.15 -*- Autoconf -*-

-# Copyright (C) 1996-2017 Free Software Foundation, Inc.
+# Copyright (C) 1996-2014 Free Software Foundation, Inc.

 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -13,7 +13,7 @@

 m4_ifndef([AC_CONFIG_MACRO_DIRS], [m4_defun([_AM_CONFIG_MACRO_DIRS], [])m4_defun([AC_CONFIG_MACRO_DIRS], [_AM_CONFIG_MACRO_DIRS($@)])])
 # ===========================================================================
-#     https://www.gnu.org/software/autoconf-archive/ax_python_module.html
+#     http://www.gnu.org/software/autoconf-archive/ax_python_module.html
 # ===========================================================================
 #
 # SYNOPSIS
@@ -37,7 +37,7 @@ m4_ifndef([AC_CONFIG_MACRO_DIRS], [m4_defun([_AM_CONFIG_MACRO_DIRS], [])m4_defun
 #   and this notice are preserved. This file is offered as-is, without any
 #   warranty.

-#serial 9
+#serial 8

 AU_ALIAS([AC_PYTHON_MODULE], [AX_PYTHON_MODULE])
 AC_DEFUN([AX_PYTHON_MODULE],[
@@ -69,63 +69,32 @@ AC_DEFUN([AX_PYTHON_MODULE],[
    fi
 ])

-# pkg.m4 - Macros to locate and utilise pkg-config.   -*- Autoconf -*-
-# serial 11 (pkg-config-0.29.1)
+# pkg.m4 - Macros to locate and utilise pkg-config.            -*- Autoconf -*-
+# serial 1 (pkg-config-0.24)
+# 
+# Copyright © 2004 Scott James Remnant <scott@netsplit.com>.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+# General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+#
+# As a special exception to the GNU General Public License, if you
+# distribute this file as part of a program that contains a
+# configuration script generated by Autoconf, you may include it under
+# the same distribution terms that you use for the rest of that program.

-dnl Copyright © 2004 Scott James Remnant <scott@netsplit.com>.
-dnl Copyright © 2012-2015 Dan Nicholson <dbn.lists@gmail.com>
-dnl
-dnl This program is free software; you can redistribute it and/or modify
-dnl it under the terms of the GNU General Public License as published by
-dnl the Free Software Foundation; either version 2 of the License, or
-dnl (at your option) any later version.
-dnl
-dnl This program is distributed in the hope that it will be useful, but
-dnl WITHOUT ANY WARRANTY; without even the implied warranty of
-dnl MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
-dnl General Public License for more details.
-dnl
-dnl You should have received a copy of the GNU General Public License
-dnl along with this program; if not, write to the Free Software
-dnl Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
-dnl 02111-1307, USA.
-dnl
-dnl As a special exception to the GNU General Public License, if you
-dnl distribute this file as part of a program that contains a
-dnl configuration script generated by Autoconf, you may include it under
-dnl the same distribution terms that you use for the rest of that
-dnl program.
-
-dnl PKG_PREREQ(MIN-VERSION)
-dnl -----------------------
-dnl Since: 0.29
-dnl
-dnl Verify that the version of the pkg-config macros are at least
-dnl MIN-VERSION. Unlike PKG_PROG_PKG_CONFIG, which checks the user's
-dnl installed version of pkg-config, this checks the developer's version
-dnl of pkg.m4 when generating configure.
-dnl
-dnl To ensure that this macro is defined, also add:
-dnl m4_ifndef([PKG_PREREQ],
-dnl     [m4_fatal([must install pkg-config 0.29 or later before running autoconf/autogen])])
-dnl
-dnl See the "Since" comment for each macro you use to see what version
-dnl of the macros you require.
-m4_defun([PKG_PREREQ],
-[m4_define([PKG_MACROS_VERSION], [0.29.1])
-m4_if(m4_version_compare(PKG_MACROS_VERSION, [$1]), -1,
-    [m4_fatal([pkg.m4 version $1 or higher is required but ]PKG_MACROS_VERSION[ found])])
-])dnl PKG_PREREQ
-
-dnl PKG_PROG_PKG_CONFIG([MIN-VERSION])
-dnl ----------------------------------
-dnl Since: 0.16
-dnl
-dnl Search for the pkg-config tool and set the PKG_CONFIG variable to
-dnl first found in the path. Checks that the version of pkg-config found
-dnl is at least MIN-VERSION. If MIN-VERSION is not specified, 0.9.0 is
-dnl used since that's the first version where most current features of
-dnl pkg-config existed.
+# PKG_PROG_PKG_CONFIG([MIN-VERSION])
+# ----------------------------------
 AC_DEFUN([PKG_PROG_PKG_CONFIG],
 [m4_pattern_forbid([^_?PKG_[A-Z_]+$])
 m4_pattern_allow([^PKG_CONFIG(_(PATH|LIBDIR|SYSROOT_DIR|ALLOW_SYSTEM_(CFLAGS|LIBS)))?$])
@@ -147,19 +116,18 @@ if test -n "$PKG_CONFIG"; then
 		PKG_CONFIG=""
 	fi
 fi[]dnl
-])dnl PKG_PROG_PKG_CONFIG
+])# PKG_PROG_PKG_CONFIG

-dnl PKG_CHECK_EXISTS(MODULES, [ACTION-IF-FOUND], [ACTION-IF-NOT-FOUND])
-dnl -------------------------------------------------------------------
-dnl Since: 0.18
-dnl
-dnl Check to see whether a particular set of modules exists. Similar to
-dnl PKG_CHECK_MODULES(), but does not set variables or print errors.
-dnl
-dnl Please remember that m4 expands AC_REQUIRE([PKG_PROG_PKG_CONFIG])
-dnl only at the first occurence in configure.ac, so if the first place
-dnl it's called might be skipped (such as if it is within an "if", you
-dnl have to call PKG_CHECK_EXISTS manually
+# PKG_CHECK_EXISTS(MODULES, [ACTION-IF-FOUND], [ACTION-IF-NOT-FOUND])
+#
+# Check to see whether a particular set of modules exists.  Similar
+# to PKG_CHECK_MODULES(), but does not set variables or print errors.
+#
+# Please remember that m4 expands AC_REQUIRE([PKG_PROG_PKG_CONFIG])
+# only at the first occurence in configure.ac, so if the first place
+# it's called might be skipped (such as if it is within an "if", you
+# have to call PKG_CHECK_EXISTS manually
+# --------------------------------------------------------------
 AC_DEFUN([PKG_CHECK_EXISTS],
 [AC_REQUIRE([PKG_PROG_PKG_CONFIG])dnl
 if test -n "$PKG_CONFIG" && \
@@ -169,10 +137,8 @@ m4_ifvaln([$3], [else
  $3])dnl
 fi])

-dnl _PKG_CONFIG([VARIABLE], [COMMAND], [MODULES])
-dnl ---------------------------------------------
-dnl Internal wrapper calling pkg-config via PKG_CONFIG and setting
-dnl pkg_failed based on the result.
+# _PKG_CONFIG([VARIABLE], [COMMAND], [MODULES])
+# ---------------------------------------------
 m4_define([_PKG_CONFIG],
 [if test -n "$$1"; then
    pkg_cv_[]$1="$$1"
@@ -184,11 +150,10 @@ m4_define([_PKG_CONFIG],
 else
    pkg_failed=untried
 fi[]dnl
-])dnl _PKG_CONFIG
+])# _PKG_CONFIG

-dnl _PKG_SHORT_ERRORS_SUPPORTED
-dnl ---------------------------
-dnl Internal check to see if pkg-config supports short errors.
+# _PKG_SHORT_ERRORS_SUPPORTED
+# -----------------------------
 AC_DEFUN([_PKG_SHORT_ERRORS_SUPPORTED],
 [AC_REQUIRE([PKG_PROG_PKG_CONFIG])
 if $PKG_CONFIG --atleast-pkgconfig-version 0.20; then
@@ -196,17 +161,19 @@ if $PKG_CONFIG --atleast-pkgconfig-version 0.20; then
 else
        _pkg_short_errors_supported=no
 fi[]dnl
-])dnl _PKG_SHORT_ERRORS_SUPPORTED
+])# _PKG_SHORT_ERRORS_SUPPORTED


-dnl PKG_CHECK_MODULES(VARIABLE-PREFIX, MODULES, [ACTION-IF-FOUND],
-dnl   [ACTION-IF-NOT-FOUND])
-dnl --------------------------------------------------------------
-dnl Since: 0.4.0
-dnl
-dnl Note that if there is a possibility the first call to
-dnl PKG_CHECK_MODULES might not happen, you should be sure to include an
-dnl explicit call to PKG_PROG_PKG_CONFIG in your configure.ac
+# PKG_CHECK_MODULES(VARIABLE-PREFIX, MODULES, [ACTION-IF-FOUND],
+# [ACTION-IF-NOT-FOUND])
+#
+#
+# Note that if there is a possibility the first call to
+# PKG_CHECK_MODULES might not happen, you should be sure to include an
+# explicit call to PKG_PROG_PKG_CONFIG in your configure.ac
+#
+#
+# --------------------------------------------------------------
 AC_DEFUN([PKG_CHECK_MODULES],
 [AC_REQUIRE([PKG_PROG_PKG_CONFIG])dnl
 AC_ARG_VAR([$1][_CFLAGS], [C compiler flags for $1, overriding pkg-config])dnl
@@ -260,40 +227,16 @@ else
        AC_MSG_RESULT([yes])
 	$3
 fi[]dnl
-])dnl PKG_CHECK_MODULES
+])# PKG_CHECK_MODULES


-dnl PKG_CHECK_MODULES_STATIC(VARIABLE-PREFIX, MODULES, [ACTION-IF-FOUND],
-dnl   [ACTION-IF-NOT-FOUND])
-dnl ---------------------------------------------------------------------
-dnl Since: 0.29
-dnl
-dnl Checks for existence of MODULES and gathers its build flags with
-dnl static libraries enabled. Sets VARIABLE-PREFIX_CFLAGS from --cflags
-dnl and VARIABLE-PREFIX_LIBS from --libs.
-dnl
-dnl Note that if there is a possibility the first call to
-dnl PKG_CHECK_MODULES_STATIC might not happen, you should be sure to
-dnl include an explicit call to PKG_PROG_PKG_CONFIG in your
-dnl configure.ac.
-AC_DEFUN([PKG_CHECK_MODULES_STATIC],
-[AC_REQUIRE([PKG_PROG_PKG_CONFIG])dnl
-_save_PKG_CONFIG=$PKG_CONFIG
-PKG_CONFIG="$PKG_CONFIG --static"
-PKG_CHECK_MODULES($@)
-PKG_CONFIG=$_save_PKG_CONFIG[]dnl
-])dnl PKG_CHECK_MODULES_STATIC
-
-
-dnl PKG_INSTALLDIR([DIRECTORY])
-dnl -------------------------
-dnl Since: 0.27
-dnl
-dnl Substitutes the variable pkgconfigdir as the location where a module
-dnl should install pkg-config .pc files. By default the directory is
-dnl $libdir/pkgconfig, but the default can be changed by passing
-dnl DIRECTORY. The user can override through the --with-pkgconfigdir
-dnl parameter.
+# PKG_INSTALLDIR(DIRECTORY)
+# -------------------------
+# Substitutes the variable pkgconfigdir as the location where a module
+# should install pkg-config .pc files. By default the directory is
+# $libdir/pkgconfig, but the default can be changed by passing
+# DIRECTORY. The user can override through the --with-pkgconfigdir
+# parameter.
 AC_DEFUN([PKG_INSTALLDIR],
 [m4_pushdef([pkg_default], [m4_default([$1], ['${libdir}/pkgconfig'])])
 m4_pushdef([pkg_description],
@@ -304,18 +247,16 @@ AC_ARG_WITH([pkgconfigdir],
 AC_SUBST([pkgconfigdir], [$with_pkgconfigdir])
 m4_popdef([pkg_default])
 m4_popdef([pkg_description])
-])dnl PKG_INSTALLDIR
+]) dnl PKG_INSTALLDIR


-dnl PKG_NOARCH_INSTALLDIR([DIRECTORY])
-dnl --------------------------------
-dnl Since: 0.27
-dnl
-dnl Substitutes the variable noarch_pkgconfigdir as the location where a
-dnl module should install arch-independent pkg-config .pc files. By
-dnl default the directory is $datadir/pkgconfig, but the default can be
-dnl changed by passing DIRECTORY. The user can override through the
-dnl --with-noarch-pkgconfigdir parameter.
+# PKG_NOARCH_INSTALLDIR(DIRECTORY)
+# -------------------------
+# Substitutes the variable noarch_pkgconfigdir as the location where a
+# module should install arch-independent pkg-config .pc files. By
+# default the directory is $datadir/pkgconfig, but the default can be
+# changed by passing DIRECTORY. The user can override through the
+# --with-noarch-pkgconfigdir parameter.
 AC_DEFUN([PKG_NOARCH_INSTALLDIR],
 [m4_pushdef([pkg_default], [m4_default([$1], ['${datadir}/pkgconfig'])])
 m4_pushdef([pkg_description],
@@ -326,15 +267,13 @@ AC_ARG_WITH([noarch-pkgconfigdir],
 AC_SUBST([noarch_pkgconfigdir], [$with_noarch_pkgconfigdir])
 m4_popdef([pkg_default])
 m4_popdef([pkg_description])
-])dnl PKG_NOARCH_INSTALLDIR
+]) dnl PKG_NOARCH_INSTALLDIR


-dnl PKG_CHECK_VAR(VARIABLE, MODULE, CONFIG-VARIABLE,
-dnl [ACTION-IF-FOUND], [ACTION-IF-NOT-FOUND])
-dnl -------------------------------------------
-dnl Since: 0.28
-dnl
-dnl Retrieves the value of the pkg-config variable for the given module.
+# PKG_CHECK_VAR(VARIABLE, MODULE, CONFIG-VARIABLE,
+# [ACTION-IF-FOUND], [ACTION-IF-NOT-FOUND])
+# -------------------------------------------
+# Retrieves the value of the pkg-config variable for the given module.
 AC_DEFUN([PKG_CHECK_VAR],
 [AC_REQUIRE([PKG_PROG_PKG_CONFIG])dnl
 AC_ARG_VAR([$1], [value of $3 for $2, overriding pkg-config])dnl
@@ -343,77 +282,9 @@ _PKG_CONFIG([$1], [variable="][$3]["], [$2])
 AS_VAR_COPY([$1], [pkg_cv_][$1])

 AS_VAR_IF([$1], [""], [$5], [$4])dnl
-])dnl PKG_CHECK_VAR
+])# PKG_CHECK_VAR

-dnl PKG_WITH_MODULES(VARIABLE-PREFIX, MODULES,
-dnl   [ACTION-IF-FOUND],[ACTION-IF-NOT-FOUND],
-dnl   [DESCRIPTION], [DEFAULT])
-dnl ------------------------------------------
-dnl
-dnl Prepare a "--with-" configure option using the lowercase
-dnl [VARIABLE-PREFIX] name, merging the behaviour of AC_ARG_WITH and
-dnl PKG_CHECK_MODULES in a single macro.
-AC_DEFUN([PKG_WITH_MODULES],
-[
-m4_pushdef([with_arg], m4_tolower([$1]))
-
-m4_pushdef([description],
-           [m4_default([$5], [build with ]with_arg[ support])])
-
-m4_pushdef([def_arg], [m4_default([$6], [auto])])
-m4_pushdef([def_action_if_found], [AS_TR_SH([with_]with_arg)=yes])
-m4_pushdef([def_action_if_not_found], [AS_TR_SH([with_]with_arg)=no])
-
-m4_case(def_arg,
-            [yes],[m4_pushdef([with_without], [--without-]with_arg)],
-            [m4_pushdef([with_without],[--with-]with_arg)])
-
-AC_ARG_WITH(with_arg,
-     AS_HELP_STRING(with_without, description[ @<:@default=]def_arg[@:>@]),,
-    [AS_TR_SH([with_]with_arg)=def_arg])
-
-AS_CASE([$AS_TR_SH([with_]with_arg)],
-            [yes],[PKG_CHECK_MODULES([$1],[$2],$3,$4)],
-            [auto],[PKG_CHECK_MODULES([$1],[$2],
-                                        [m4_n([def_action_if_found]) $3],
-                                        [m4_n([def_action_if_not_found]) $4])])
-
-m4_popdef([with_arg])
-m4_popdef([description])
-m4_popdef([def_arg])
-
-])dnl PKG_WITH_MODULES
-
-dnl PKG_HAVE_WITH_MODULES(VARIABLE-PREFIX, MODULES,
-dnl   [DESCRIPTION], [DEFAULT])
-dnl -----------------------------------------------
-dnl
-dnl Convenience macro to trigger AM_CONDITIONAL after PKG_WITH_MODULES
-dnl check._[VARIABLE-PREFIX] is exported as make variable.
-AC_DEFUN([PKG_HAVE_WITH_MODULES],
-[
-PKG_WITH_MODULES([$1],[$2],,,[$3],[$4])
-
-AM_CONDITIONAL([HAVE_][$1],
-               [test "$AS_TR_SH([with_]m4_tolower([$1]))" = "yes"])
-])dnl PKG_HAVE_WITH_MODULES
-
-dnl PKG_HAVE_DEFINE_WITH_MODULES(VARIABLE-PREFIX, MODULES,
-dnl   [DESCRIPTION], [DEFAULT])
-dnl ------------------------------------------------------
-dnl
-dnl Convenience macro to run AM_CONDITIONAL and AC_DEFINE after
-dnl PKG_WITH_MODULES check. HAVE_[VARIABLE-PREFIX] is exported as make
-dnl and preprocessor variable.
-AC_DEFUN([PKG_HAVE_DEFINE_WITH_MODULES],
-[
-PKG_HAVE_WITH_MODULES([$1],[$2],[$3],[$4])
-
-AS_IF([test "$AS_TR_SH([with_]m4_tolower([$1]))" = "yes"],
-        [AC_DEFINE([HAVE_][$1], 1, [Enable ]m4_tolower([$1])[ support])])
-])dnl PKG_HAVE_DEFINE_WITH_MODULES
-
-# Copyright (C) 1999-2017 Free Software Foundation, Inc.
+# Copyright (C) 1999-2014 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -446,9 +317,8 @@ AC_DEFUN([AM_PATH_PYTHON],
 [
  dnl Find a Python interpreter.  Python versions prior to 2.0 are not
  dnl supported. (2.0 was released on October 16, 2000).
-  dnl FIXME: Remove the need to hard-code Python versions here.
  m4_define_default([_AM_PYTHON_INTERPRETER_LIST],
-[python python2 python3 python3.5 python3.4 python3.3 python3.2 python3.1 python3.0 python2.7 dnl
+[python python2 python3 python3.3 python3.2 python3.1 python3.0 python2.7 dnl
 python2.6 python2.5 python2.4 python2.3 python2.2 python2.1 python2.0])

  AC_ARG_VAR([PYTHON], [the Python interpreter])
@@ -649,7 +519,7 @@ for i in list(range(0, 4)): minverhex = (minverhex << 8) + minver[[i]]
 sys.exit(sys.hexversion < minverhex)"
  AS_IF([AM_RUN_LOG([$1 -c "$prog"])], [$3], [$4])])

-# Copyright (C) 2001-2017 Free Software Foundation, Inc.
+# Copyright (C) 2001-2014 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -666,4 +536,5 @@ AC_DEFUN([AM_RUN_LOG],
   echo "$as_me:$LINENO: \$? = $ac_status" >&AS_MESSAGE_LOG_FD
   (exit $ac_status); }])

+
 m4_include([acinclude.m4])
--- a/base/Makefile
+++ b/base/Makefile
@@ -1,40 +0,0 @@
-# Copyright (C) 2018 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 Lesser General Public License v.2.1.
-#
-# You should have received a copy of the GNU Lesser General Public License
-# along with this program; if not, write to the Free Software Foundation,
-# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
-# Uncomment this to build the simple radix tree.  You'll need to make clean too.
-# Comment to build the advanced radix tree.
-#base/data-struct/radix-tree.o: CFLAGS += -DSIMPLE_RADIX_TREE
-
-# NOTE: this Makefile only works as 'include' for toplevel Makefile
-#       which defined all top_* variables
-
-BASE_SOURCE=\
-	base/data-struct/hash.c \
-	base/data-struct/list.c \
-	base/data-struct/radix-tree.c
-
-BASE_TARGET = base/libbase.a
-BASE_DEPENDS = $(BASE_SOURCE:%.c=%.d)
-BASE_OBJECTS = $(BASE_SOURCE:%.c=%.o)
-CLEAN_TARGETS += $(BASE_DEPENDS) $(BASE_OBJECTS) \
-	$(BASE_SOURCE:%.c=%.gcda) \
-	$(BASE_SOURCE:%.c=%.gcno) \
-	$(BASE_TARGET)
-
-$(BASE_TARGET): $(BASE_OBJECTS)
-	@echo "    [AR] $@"
-	$(Q) $(RM) $@
-	$(Q) $(AR) rsv $@ $(BASE_OBJECTS) > /dev/null
-
-ifeq ("$(DEPENDS)","yes")
-include $(BASE_DEPENDS)
-endif
--- a/base/data-struct/hash.c
+++ b/base/data-struct/hash.c
@@ -1,394 +0,0 @@
-/*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004-2011 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 Lesser General Public License v.2.1.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with this program; if not, write to the Free Software Foundation,
- * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include "device_mapper/misc/dmlib.h"
-#include "base/memory/zalloc.h"
-#include "hash.h"
-
-struct dm_hash_node {
-	struct dm_hash_node *next;
-	void *data;
-	unsigned data_len;
-	unsigned keylen;
-	char key[0];
-};
-
-struct dm_hash_table {
-	unsigned num_nodes;
-	unsigned num_slots;
-	struct dm_hash_node **slots;
-};
-
-/* Permutation of the Integers 0 through 255 */
-static unsigned char _nums[] = {
-	1, 14, 110, 25, 97, 174, 132, 119, 138, 170, 125, 118, 27, 233, 140, 51,
-	87, 197, 177, 107, 234, 169, 56, 68, 30, 7, 173, 73, 188, 40, 36, 65,
-	49, 213, 104, 190, 57, 211, 148, 223, 48, 115, 15, 2, 67, 186, 210, 28,
-	12, 181, 103, 70, 22, 58, 75, 78, 183, 167, 238, 157, 124, 147, 172,
-	144,
-	176, 161, 141, 86, 60, 66, 128, 83, 156, 241, 79, 46, 168, 198, 41, 254,
-	178, 85, 253, 237, 250, 154, 133, 88, 35, 206, 95, 116, 252, 192, 54,
-	221,
-	102, 218, 255, 240, 82, 106, 158, 201, 61, 3, 89, 9, 42, 155, 159, 93,
-	166, 80, 50, 34, 175, 195, 100, 99, 26, 150, 16, 145, 4, 33, 8, 189,
-	121, 64, 77, 72, 208, 245, 130, 122, 143, 55, 105, 134, 29, 164, 185,
-	194,
-	193, 239, 101, 242, 5, 171, 126, 11, 74, 59, 137, 228, 108, 191, 232,
-	139,
-	6, 24, 81, 20, 127, 17, 91, 92, 251, 151, 225, 207, 21, 98, 113, 112,
-	84, 226, 18, 214, 199, 187, 13, 32, 94, 220, 224, 212, 247, 204, 196,
-	43,
-	249, 236, 45, 244, 111, 182, 153, 136, 129, 90, 217, 202, 19, 165, 231,
-	71,
-	230, 142, 96, 227, 62, 179, 246, 114, 162, 53, 160, 215, 205, 180, 47,
-	109,
-	44, 38, 31, 149, 135, 0, 216, 52, 63, 23, 37, 69, 39, 117, 146, 184,
-	163, 200, 222, 235, 248, 243, 219, 10, 152, 131, 123, 229, 203, 76, 120,
-	209
-};
-
-static struct dm_hash_node *_create_node(const char *str, unsigned len)
-{
-	struct dm_hash_node *n = malloc(sizeof(*n) + len);
-
-	if (n) {
-		memcpy(n->key, str, len);
-		n->keylen = len;
-	}
-
-	return n;
-}
-
-static unsigned long _hash(const char *str, unsigned len)
-{
-	unsigned long h = 0, g;
-	unsigned i;
-
-	for (i = 0; i < len; i++) {
-		h <<= 4;
-		h += _nums[(unsigned char) *str++];
-		g = h & ((unsigned long) 0xf << 16u);
-		if (g) {
-			h ^= g >> 16u;
-			h ^= g >> 5u;
-		}
-	}
-
-	return h;
-}
-
-struct dm_hash_table *dm_hash_create(unsigned size_hint)
-{
-	size_t len;
-	unsigned new_size = 16u;
-	struct dm_hash_table *hc = zalloc(sizeof(*hc));
-
-	if (!hc)
-		return_0;
-
-	/* round size hint up to a power of two */
-	while (new_size < size_hint)
-		new_size = new_size << 1;
-
-	hc->num_slots = new_size;
-	len = sizeof(*(hc->slots)) * new_size;
-	if (!(hc->slots = zalloc(len)))
-		goto_bad;
-
-	return hc;
-
-      bad:
-	free(hc->slots);
-	free(hc);
-	return 0;
-}
-
-static void _free_nodes(struct dm_hash_table *t)
-{
-	struct dm_hash_node *c, *n;
-	unsigned i;
-
-	for (i = 0; i < t->num_slots; i++)
-		for (c = t->slots[i]; c; c = n) {
-			n = c->next;
-			free(c);
-		}
-}
-
-void dm_hash_destroy(struct dm_hash_table *t)
-{
-	_free_nodes(t);
-	free(t->slots);
-	free(t);
-}
-
-static struct dm_hash_node **_find(struct dm_hash_table *t, const void *key,
-				   uint32_t len)
-{
-	unsigned h = _hash(key, len) & (t->num_slots - 1);
-	struct dm_hash_node **c;
-
-	for (c = &t->slots[h]; *c; c = &((*c)->next)) {
-		if ((*c)->keylen != len)
-			continue;
-
-		if (!memcmp(key, (*c)->key, len))
-			break;
-	}
-
-	return c;
-}
-
-void *dm_hash_lookup_binary(struct dm_hash_table *t, const void *key,
-			    uint32_t len)
-{
-	struct dm_hash_node **c = _find(t, key, len);
-
-	return *c ? (*c)->data : 0;
-}
-
-int dm_hash_insert_binary(struct dm_hash_table *t, const void *key,
-			  uint32_t len, void *data)
-{
-	struct dm_hash_node **c = _find(t, key, len);
-
-	if (*c)
-		(*c)->data = data;
-	else {
-		struct dm_hash_node *n = _create_node(key, len);
-
-		if (!n)
-			return 0;
-
-		n->data = data;
-		n->next = 0;
-		*c = n;
-		t->num_nodes++;
-	}
-
-	return 1;
-}
-
-void dm_hash_remove_binary(struct dm_hash_table *t, const void *key,
-			uint32_t len)
-{
-	struct dm_hash_node **c = _find(t, key, len);
-
-	if (*c) {
-		struct dm_hash_node *old = *c;
-		*c = (*c)->next;
-		free(old);
-		t->num_nodes--;
-	}
-}
-
-void *dm_hash_lookup(struct dm_hash_table *t, const char *key)
-{
-	return dm_hash_lookup_binary(t, key, strlen(key) + 1);
-}
-
-int dm_hash_insert(struct dm_hash_table *t, const char *key, void *data)
-{
-	return dm_hash_insert_binary(t, key, strlen(key) + 1, data);
-}
-
-void dm_hash_remove(struct dm_hash_table *t, const char *key)
-{
-	dm_hash_remove_binary(t, key, strlen(key) + 1);
-}
-
-static struct dm_hash_node **_find_str_with_val(struct dm_hash_table *t,
-					        const void *key, const void *val,
-					        uint32_t len, uint32_t val_len)
-{
-	struct dm_hash_node **c;
-	unsigned h;
-       
-	h = _hash(key, len) & (t->num_slots - 1);
-
-	for (c = &t->slots[h]; *c; c = &((*c)->next)) {
-		if ((*c)->keylen != len)
-			continue;
-
-		if (!memcmp(key, (*c)->key, len) && (*c)->data) {
-			if (((*c)->data_len == val_len) &&
-			    !memcmp(val, (*c)->data, val_len))
-				return c;
-		}
-	}
-
-	return NULL;
-}
-
-int dm_hash_insert_allow_multiple(struct dm_hash_table *t, const char *key,
-				  const void *val, uint32_t val_len)
-{
-	struct dm_hash_node *n;
-	struct dm_hash_node *first;
-	int len = strlen(key) + 1;
-	unsigned h;
-
-	n = _create_node(key, len);
-	if (!n)
-		return 0;
-
-	n->data = (void *)val;
-	n->data_len = val_len;
-
-	h = _hash(key, len) & (t->num_slots - 1);
-
-	first = t->slots[h];
-
-	if (first)
-		n->next = first;
-	else
-		n->next = 0;
-	t->slots[h] = n;
-
-	t->num_nodes++;
-	return 1;
-}
-
-/*
- * Look through multiple entries with the same key for one that has a
- * matching val and return that.  If none have maching val, return NULL.
- */
-void *dm_hash_lookup_with_val(struct dm_hash_table *t, const char *key,
-			      const void *val, uint32_t val_len)
-{
-	struct dm_hash_node **c;
-
-	c = _find_str_with_val(t, key, val, strlen(key) + 1, val_len);
-
-	return (c && *c) ? (*c)->data : 0;
-}
-
-/*
- * Look through multiple entries with the same key for one that has a
- * matching val and remove that.
- */
-void dm_hash_remove_with_val(struct dm_hash_table *t, const char *key,
-			     const void *val, uint32_t val_len)
-{
-	struct dm_hash_node **c;
-
-	c = _find_str_with_val(t, key, val, strlen(key) + 1, val_len);
-
-	if (c && *c) {
-		struct dm_hash_node *old = *c;
-		*c = (*c)->next;
-		free(old);
-		t->num_nodes--;
-	}
-}
-
-/*
- * Look up the value for a key and count how many
- * entries have the same key.
- *
- * If no entries have key, return NULL and set count to 0.
- *
- * If one entry has the key, the function returns the val,
- * and sets count to 1.
- *
- * If N entries have the key, the function returns the val
- * from the first entry, and sets count to N.
- */
-void *dm_hash_lookup_with_count(struct dm_hash_table *t, const char *key, int *count)
-{
-	struct dm_hash_node **c;
-	struct dm_hash_node **c1 = NULL;
-	uint32_t len = strlen(key) + 1;
-	unsigned h;
-
-	*count = 0;
-
-	h = _hash(key, len) & (t->num_slots - 1);
-
-	for (c = &t->slots[h]; *c; c = &((*c)->next)) {
-		if ((*c)->keylen != len)
-			continue;
-
-		if (!memcmp(key, (*c)->key, len)) {
-			(*count)++;
-			if (!c1)
-				c1 = c;
-		}
-	}
-
-	if (!c1)
-		return NULL;
-	else
-		return *c1 ? (*c1)->data : 0;
-}
-
-unsigned dm_hash_get_num_entries(struct dm_hash_table *t)
-{
-	return t->num_nodes;
-}
-
-void dm_hash_iter(struct dm_hash_table *t, dm_hash_iterate_fn f)
-{
-	struct dm_hash_node *c, *n;
-	unsigned i;
-
-	for (i = 0; i < t->num_slots; i++)
-		for (c = t->slots[i]; c; c = n) {
-			n = c->next;
-			f(c->data);
-		}
-}
-
-void dm_hash_wipe(struct dm_hash_table *t)
-{
-	_free_nodes(t);
-	memset(t->slots, 0, sizeof(struct dm_hash_node *) * t->num_slots);
-	t->num_nodes = 0u;
-}
-
-char *dm_hash_get_key(struct dm_hash_table *t __attribute__((unused)),
-		      struct dm_hash_node *n)
-{
-	return n->key;
-}
-
-void *dm_hash_get_data(struct dm_hash_table *t __attribute__((unused)),
-		       struct dm_hash_node *n)
-{
-	return n->data;
-}
-
-static struct dm_hash_node *_next_slot(struct dm_hash_table *t, unsigned s)
-{
-	struct dm_hash_node *c = NULL;
-	unsigned i;
-
-	for (i = s; i < t->num_slots && !c; i++)
-		c = t->slots[i];
-
-	return c;
-}
-
-struct dm_hash_node *dm_hash_get_first(struct dm_hash_table *t)
-{
-	return _next_slot(t, 0);
-}
-
-struct dm_hash_node *dm_hash_get_next(struct dm_hash_table *t, struct dm_hash_node *n)
-{
-	unsigned h = _hash(n->key, n->keylen) & (t->num_slots - 1);
-
-	return n->next ? n->next : _next_slot(t, h + 1);
-}
--- a/base/data-struct/hash.h
+++ b/base/data-struct/hash.h
@@ -1,94 +0,0 @@
-#ifndef BASE_DATA_STRUCT_HASH_H
-#define BASE_DATA_STRUCT_HASH_H
-
-#include <stdint.h>
-
-//----------------------------------------------------------------
-
-struct dm_hash_table;
-struct dm_hash_node;
-
-typedef void (*dm_hash_iterate_fn) (void *data);
-
-struct dm_hash_table *dm_hash_create(unsigned size_hint)
-	__attribute__((__warn_unused_result__));
-void dm_hash_destroy(struct dm_hash_table *t);
-void dm_hash_wipe(struct dm_hash_table *t);
-
-void *dm_hash_lookup(struct dm_hash_table *t, const char *key);
-int dm_hash_insert(struct dm_hash_table *t, const char *key, void *data);
-void dm_hash_remove(struct dm_hash_table *t, const char *key);
-
-void *dm_hash_lookup_binary(struct dm_hash_table *t, const void *key, uint32_t len);
-int dm_hash_insert_binary(struct dm_hash_table *t, const void *key, uint32_t len,
-			  void *data);
-void dm_hash_remove_binary(struct dm_hash_table *t, const void *key, uint32_t len);
-
-unsigned dm_hash_get_num_entries(struct dm_hash_table *t);
-void dm_hash_iter(struct dm_hash_table *t, dm_hash_iterate_fn f);
-
-char *dm_hash_get_key(struct dm_hash_table *t, struct dm_hash_node *n);
-void *dm_hash_get_data(struct dm_hash_table *t, struct dm_hash_node *n);
-struct dm_hash_node *dm_hash_get_first(struct dm_hash_table *t);
-struct dm_hash_node *dm_hash_get_next(struct dm_hash_table *t, struct dm_hash_node *n);
-
-/*
- * dm_hash_insert() replaces the value of an existing
- * entry with a matching key if one exists.  Otherwise
- * it adds a new entry.
- *
- * dm_hash_insert_with_val() inserts a new entry if
- * another entry with the same key already exists.
- * val_len is the size of the data being inserted.
- *
- * If two entries with the same key exist,
- * (added using dm_hash_insert_allow_multiple), then:
- * . dm_hash_lookup() returns the first one it finds, and
- *   dm_hash_lookup_with_val() returns the one with a matching
- *   val_len/val.
- * . dm_hash_remove() removes the first one it finds, and
- *   dm_hash_remove_with_val() removes the one with a matching
- *   val_len/val.
- *
- * If a single entry with a given key exists, and it has
- * zero val_len, then:
- * . dm_hash_lookup() returns it
- * . dm_hash_lookup_with_val(val_len=0) returns it
- * . dm_hash_remove() removes it
- * . dm_hash_remove_with_val(val_len=0) removes it
- *
- * dm_hash_lookup_with_count() is a single call that will
- * both lookup a key's value and check if there is more
- * than one entry with the given key.
- *
- * (It is not meant to retrieve all the entries with the
- * given key.  In the common case where a single entry exists
- * for the key, it is useful to have a single call that will
- * both look up the value and indicate if multiple values
- * exist for the key.)
- *
- * dm_hash_lookup_with_count:
- * . If no entries exist, the function returns NULL, and
- *   the count is set to 0.
- * . If only one entry exists, the value of that entry is
- *   returned and count is set to 1.
- * . If N entries exists, the value of the first entry is
- *   returned and count is set to N.
- */
-
-void *dm_hash_lookup_with_val(struct dm_hash_table *t, const char *key,
-                              const void *val, uint32_t val_len);
-void dm_hash_remove_with_val(struct dm_hash_table *t, const char *key,
-                             const void *val, uint32_t val_len);
-int dm_hash_insert_allow_multiple(struct dm_hash_table *t, const char *key,
-                                  const void *val, uint32_t val_len);
-void *dm_hash_lookup_with_count(struct dm_hash_table *t, const char *key, int *count);
-
-
-#define dm_hash_iterate(v, h) \
-	for (v = dm_hash_get_first((h)); v; \
-	     v = dm_hash_get_next((h), v))
-
-//----------------------------------------------------------------
-
-#endif
--- a/base/data-struct/list.c
+++ b/base/data-struct/list.c
@@ -1,170 +0,0 @@
-/*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004-2010 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 Lesser General Public License v.2.1.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with this program; if not, write to the Free Software Foundation,
- * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include "list.h"
-
-#include <assert.h>
-#include <stdlib.h>
-
-/*
- * Initialise a list before use.
- * The list head's next and previous pointers point back to itself.
- */
-void dm_list_init(struct dm_list *head)
-{
-	head->n = head->p = head;
-}
-
-/*
- * Insert an element before 'head'.
- * If 'head' is the list head, this adds an element to the end of the list.
- */
-void dm_list_add(struct dm_list *head, struct dm_list *elem)
-{
-	assert(head->n);
-
-	elem->n = head;
-	elem->p = head->p;
-
-	head->p->n = elem;
-	head->p = elem;
-}
-
-/*
- * Insert an element after 'head'.
- * If 'head' is the list head, this adds an element to the front of the list.
- */
-void dm_list_add_h(struct dm_list *head, struct dm_list *elem)
-{
-	assert(head->n);
-
-	elem->n = head->n;
-	elem->p = head;
-
-	head->n->p = elem;
-	head->n = elem;
-}
-
-/*
- * Delete an element from its list.
- * Note that this doesn't change the element itself - it may still be safe
- * to follow its pointers.
- */
-void dm_list_del(struct dm_list *elem)
-{
-	elem->n->p = elem->p;
-	elem->p->n = elem->n;
-}
-
-/*
- * Remove an element from existing list and insert before 'head'.
- */
-void dm_list_move(struct dm_list *head, struct dm_list *elem)
-{
-        dm_list_del(elem);
-        dm_list_add(head, elem);
-}
-
-/*
- * Is the list empty?
- */
-int dm_list_empty(const struct dm_list *head)
-{
-	return head->n == head;
-}
-
-/*
- * Is this the first element of the list?
- */
-int dm_list_start(const struct dm_list *head, const struct dm_list *elem)
-{
-	return elem->p == head;
-}
-
-/*
- * Is this the last element of the list?
- */
-int dm_list_end(const struct dm_list *head, const struct dm_list *elem)
-{
-	return elem->n == head;
-}
-
-/*
- * Return first element of the list or NULL if empty
- */
-struct dm_list *dm_list_first(const struct dm_list *head)
-{
-	return (dm_list_empty(head) ? NULL : head->n);
-}
-
-/*
- * Return last element of the list or NULL if empty
- */
-struct dm_list *dm_list_last(const struct dm_list *head)
-{
-	return (dm_list_empty(head) ? NULL : head->p);
-}
-
-/*
- * Return the previous element of the list, or NULL if we've reached the start.
- */
-struct dm_list *dm_list_prev(const struct dm_list *head, const struct dm_list *elem)
-{
-	return (dm_list_start(head, elem) ? NULL : elem->p);
-}
-
-/*
- * Return the next element of the list, or NULL if we've reached the end.
- */
-struct dm_list *dm_list_next(const struct dm_list *head, const struct dm_list *elem)
-{
-	return (dm_list_end(head, elem) ? NULL : elem->n);
-}
-
-/*
- * Return the number of elements in a list by walking it.
- */
-unsigned int dm_list_size(const struct dm_list *head)
-{
-	unsigned int s = 0;
-	const struct dm_list *v;
-
-	dm_list_iterate(v, head)
-	    s++;
-
-	return s;
-}
-
-/*
- * Join two lists together.
- * This moves all the elements of the list 'head1' to the end of the list
- * 'head', leaving 'head1' empty.
- */
-void dm_list_splice(struct dm_list *head, struct dm_list *head1)
-{
-	assert(head->n);
-	assert(head1->n);
-
-	if (dm_list_empty(head1))
-	    return;
-
-	head1->p->n = head;
-	head1->n->p = head->p;
-
-	head->p->n = head1->n;
-	head->p = head1->p;
-
-	dm_list_init(head1);
-}
--- a/base/data-struct/list.h
+++ b/base/data-struct/list.h
@@ -1,209 +0,0 @@
-#ifndef BASE_DATA_STRUCT_LIST_H
-#define BASE_DATA_STRUCT_LIST_H
-
-//----------------------------------------------------------------
-
-/*
- * A list consists of a list head plus elements.
- * Each element has 'next' and 'previous' pointers.
- * The list head's pointers point to the first and the last element.
- */
-
-struct dm_list {
-	struct dm_list *n, *p;
-};
-
-/*
- * String list.
- */
-struct dm_str_list {
-	struct dm_list list;
-	const char *str;
-};
-
-/*
- * Initialise a list before use.
- * The list head's next and previous pointers point back to itself.
- */
-#define DM_LIST_HEAD_INIT(name)	 { &(name), &(name) }
-#define DM_LIST_INIT(name)	struct dm_list name = DM_LIST_HEAD_INIT(name)
-void dm_list_init(struct dm_list *head);
-
-/*
- * Insert an element before 'head'.
- * If 'head' is the list head, this adds an element to the end of the list.
- */
-void dm_list_add(struct dm_list *head, struct dm_list *elem);
-
-/*
- * Insert an element after 'head'.
- * If 'head' is the list head, this adds an element to the front of the list.
- */
-void dm_list_add_h(struct dm_list *head, struct dm_list *elem);
-
-/*
- * Delete an element from its list.
- * Note that this doesn't change the element itself - it may still be safe
- * to follow its pointers.
- */
-void dm_list_del(struct dm_list *elem);
-
-/*
- * Remove an element from existing list and insert before 'head'.
- */
-void dm_list_move(struct dm_list *head, struct dm_list *elem);
-
-/*
- * Join 'head1' to the end of 'head'.
- */
-void dm_list_splice(struct dm_list *head, struct dm_list *head1);
-
-/*
- * Is the list empty?
- */
-int dm_list_empty(const struct dm_list *head);
-
-/*
- * Is this the first element of the list?
- */
-int dm_list_start(const struct dm_list *head, const struct dm_list *elem);
-
-/*
- * Is this the last element of the list?
- */
-int dm_list_end(const struct dm_list *head, const struct dm_list *elem);
-
-/*
- * Return first element of the list or NULL if empty
- */
-struct dm_list *dm_list_first(const struct dm_list *head);
-
-/*
- * Return last element of the list or NULL if empty
- */
-struct dm_list *dm_list_last(const struct dm_list *head);
-
-/*
- * Return the previous element of the list, or NULL if we've reached the start.
- */
-struct dm_list *dm_list_prev(const struct dm_list *head, const struct dm_list *elem);
-
-/*
- * Return the next element of the list, or NULL if we've reached the end.
- */
-struct dm_list *dm_list_next(const struct dm_list *head, const struct dm_list *elem);
-
-/*
- * Given the address v of an instance of 'struct dm_list' called 'head'
- * contained in a structure of type t, return the containing structure.
- */
-#define dm_list_struct_base(v, t, head) \
-    ((t *)((const char *)(v) - (const char *)&((t *) 0)->head))
-
-/*
- * Given the address v of an instance of 'struct dm_list list' contained in
- * a structure of type t, return the containing structure.
- */
-#define dm_list_item(v, t) dm_list_struct_base((v), t, list)
-
-/*
- * Given the address v of one known element e in a known structure of type t,
- * return another element f.
- */
-#define dm_struct_field(v, t, e, f) \
-    (((t *)((uintptr_t)(v) - (uintptr_t)&((t *) 0)->e))->f)
-
-/*
- * Given the address v of a known element e in a known structure of type t,
- * return the list head 'list'
- */
-#define dm_list_head(v, t, e) dm_struct_field(v, t, e, list)
-
-/*
- * Set v to each element of a list in turn.
- */
-#define dm_list_iterate(v, head) \
-	for (v = (head)->n; v != head; v = v->n)
-
-/*
- * Set v to each element in a list in turn, starting from the element
- * in front of 'start'.
- * You can use this to 'unwind' a list_iterate and back out actions on
- * already-processed elements.
- * If 'start' is 'head' it walks the list backwards.
- */
-#define dm_list_uniterate(v, head, start) \
-	for (v = (start)->p; v != head; v = v->p)
-
-/*
- * A safe way to walk a list and delete and free some elements along
- * the way.
- * t must be defined as a temporary variable of the same type as v.
- */
-#define dm_list_iterate_safe(v, t, head) \
-	for (v = (head)->n, t = v->n; v != head; v = t, t = v->n)
-
-/*
- * Walk a list, setting 'v' in turn to the containing structure of each item.
- * The containing structure should be the same type as 'v'.
- * The 'struct dm_list' variable within the containing structure is 'field'.
- */
-#define dm_list_iterate_items_gen(v, head, field) \
-	for (v = dm_list_struct_base((head)->n, __typeof__(*v), field); \
-	     &v->field != (head); \
-	     v = dm_list_struct_base(v->field.n, __typeof__(*v), field))
-
-/*
- * Walk a list, setting 'v' in turn to the containing structure of each item.
- * The containing structure should be the same type as 'v'.
- * The list should be 'struct dm_list list' within the containing structure.
- */
-#define dm_list_iterate_items(v, head) dm_list_iterate_items_gen(v, (head), list)
-
-/*
- * Walk a list, setting 'v' in turn to the containing structure of each item.
- * The containing structure should be the same type as 'v'.
- * The 'struct dm_list' variable within the containing structure is 'field'.
- * t must be defined as a temporary variable of the same type as v.
- */
-#define dm_list_iterate_items_gen_safe(v, t, head, field) \
-	for (v = dm_list_struct_base((head)->n, __typeof__(*v), field), \
-	     t = dm_list_struct_base(v->field.n, __typeof__(*v), field); \
-	     &v->field != (head); \
-	     v = t, t = dm_list_struct_base(v->field.n, __typeof__(*v), field))
-/*
- * Walk a list, setting 'v' in turn to the containing structure of each item.
- * The containing structure should be the same type as 'v'.
- * The list should be 'struct dm_list list' within the containing structure.
- * t must be defined as a temporary variable of the same type as v.
- */
-#define dm_list_iterate_items_safe(v, t, head) \
-	dm_list_iterate_items_gen_safe(v, t, (head), list)
-
-/*
- * Walk a list backwards, setting 'v' in turn to the containing structure
- * of each item.
- * The containing structure should be the same type as 'v'.
- * The 'struct dm_list' variable within the containing structure is 'field'.
- */
-#define dm_list_iterate_back_items_gen(v, head, field) \
-	for (v = dm_list_struct_base((head)->p, __typeof__(*v), field); \
-	     &v->field != (head); \
-	     v = dm_list_struct_base(v->field.p, __typeof__(*v), field))
-
-/*
- * Walk a list backwards, setting 'v' in turn to the containing structure
- * of each item.
- * The containing structure should be the same type as 'v'.
- * The list should be 'struct dm_list list' within the containing structure.
- */
-#define dm_list_iterate_back_items(v, head) dm_list_iterate_back_items_gen(v, (head), list)
-
-/*
- * Return the number of elements in a list by walking it.
- */
-unsigned int dm_list_size(const struct dm_list *head);
-
-//----------------------------------------------------------------
-
-#endif
--- a/base/data-struct/radix-tree-adaptive.c
+++ b/base/data-struct/radix-tree-adaptive.c
--- a/base/data-struct/radix-tree-simple.c
+++ b/base/data-struct/radix-tree-simple.c
@@ -1,256 +0,0 @@
-// Copyright (C) 2018 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 Lesser General Public License v.2.1.
-//
-// You should have received a copy of the GNU Lesser General Public License
-// along with this program; if not, write to the Free Software Foundation,
-// Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
-#include "radix-tree.h"
-
-#include "base/memory/container_of.h"
-#include "base/memory/zalloc.h"
-
-#include <assert.h>
-#include <stdlib.h>
-#include <stdio.h>
-
-//----------------------------------------------------------------
-// This implementation is based around nested binary trees.  Very
-// simple (and hopefully correct).
-
-struct node {
-	struct node *left;
-	struct node *right;
-
-	uint8_t key;
-	struct node *center;
-
-	bool has_value;
-	union radix_value value;
-};
-
-struct radix_tree {
-	radix_value_dtr dtr;
-	void *dtr_context;
-
-	struct node *root;
-};
-
-struct radix_tree *
-radix_tree_create(radix_value_dtr dtr, void *dtr_context)
-{
-	struct radix_tree *rt = zalloc(sizeof(*rt));
-
-	if (rt) {
-		rt->dtr = dtr;
-		rt->dtr_context = dtr_context;
-	}
-
-	return rt;
-}
-
-// Returns the number of entries in the tree
-static unsigned _destroy_tree(struct node *n, radix_value_dtr dtr, void *context)
-{
-	unsigned r;
-
-	if (!n)
-		return 0;
-
-	r = _destroy_tree(n->left, dtr, context);
-	r += _destroy_tree(n->right, dtr, context);
-	r += _destroy_tree(n->center, dtr, context);
-
-	if (n->has_value) {
-		if (dtr)
-			dtr(context, n->value);
-		r++;
-	}
-
-	free(n);
-
-	return r;
-}
-
-void radix_tree_destroy(struct radix_tree *rt)
-{
-	_destroy_tree(rt->root, rt->dtr, rt->dtr_context);
-	free(rt);
-}
-
-static unsigned _count(struct node *n)
-{
-	unsigned r;
-
-	if (!n)
-		return 0;
-
-	r = _count(n->left);
-	r += _count(n->right);
-	r += _count(n->center);
-
-	if (n->has_value)
-		r++;
-
-	return r;
-}
-
-unsigned radix_tree_size(struct radix_tree *rt)
-{
-	return _count(rt->root);
-}
-
-static struct node **_lookup(struct node **pn, uint8_t *kb, uint8_t *ke)
-{
-	struct node *n = *pn;
-
-	if (!n || (kb == ke))
-		return pn;
-
-	if (*kb < n->key)
-		return _lookup(&n->left, kb, ke);
-
-	else if (*kb > n->key)
-		return _lookup(&n->right, kb, ke);
-
-	else
-		return _lookup(&n->center, kb + 1, ke);
-}
-
-static bool _insert(struct node **pn, uint8_t *kb, uint8_t *ke, union radix_value v)
-{
-	struct node *n = *pn;
-
-	if (!n) {
-		n = zalloc(sizeof(*n));
-		if (!n)
-			return false;
-
-		n->key = *kb;
-		*pn = n;
-	}
-
-	if (kb == ke) {
-		n->has_value = true;
-		n->value = v;
-		return true;
-	}
-
-	if (*kb < n->key)
-		return _insert(&n->left, kb, ke, v);
-
-	else if (*kb > n->key)
-		return _insert(&n->right, kb, ke, v);
-
-	else
-		return _insert(&n->center, kb + 1, ke, v);
-}
-
-bool radix_tree_insert(struct radix_tree *rt, uint8_t *kb, uint8_t *ke, union radix_value v)
-{
-	return _insert(&rt->root, kb, ke, v);
-}
-
-bool radix_tree_remove(struct radix_tree *rt, uint8_t *kb, uint8_t *ke)
-{
-	struct node **pn = _lookup(&rt->root, kb, ke);
-	struct node *n = *pn;
-
-	if (!n || !n->has_value)
-		return false;
-
-	else {
-		if (rt->dtr)
-			rt->dtr(rt->dtr_context, n->value);
-
-		if (n->left || n->center || n->right) {
-			n->has_value = false;
-			return true;
-
-		} else {
-			// FIXME: delete parent if this was the last entry
-			free(n);
-			*pn = NULL;
-		}
-
-		return true;
-	}
-}
-
-unsigned radix_tree_remove_prefix(struct radix_tree *rt, uint8_t *kb, uint8_t *ke)
-{
-	struct node **pn;
-	unsigned count;
-
-	pn = _lookup(&rt->root, kb, ke);
-
-	if (*pn) {
-		count = _destroy_tree(*pn, rt->dtr, rt->dtr_context);
-		*pn = NULL;
-	}
-
-	return count;
-}
-
-bool
-radix_tree_lookup(struct radix_tree *rt, uint8_t *kb, uint8_t *ke, union radix_value *result)
-{
-	struct node **pn = _lookup(&rt->root, kb, ke);
-	struct node *n = *pn;
-
-	if (n && n->has_value) {
-		*result = n->value;
-		return true;
-	} else
-		return false;
-}
-
-static void _iterate(struct node *n, struct radix_tree_iterator *it)
-{
-	if (!n)
-		return;
-
-	_iterate(n->left, it);
-
-	if (n->has_value)
-		// FIXME: fill out the key
-		it->visit(it, NULL, NULL, n->value);
-
-	_iterate(n->center, it);
-	_iterate(n->right, it);
-}
-
-void radix_tree_iterate(struct radix_tree *rt, uint8_t *kb, uint8_t *ke,
-                        struct radix_tree_iterator *it)
-{
-	if (kb == ke)
-		_iterate(rt->root, it);
-
-	else {
-		struct node **pn = _lookup(&rt->root, kb, ke);
-		struct node *n = *pn;
-
-		if (n) {
-			if (n->has_value)
-				it->visit(it, NULL, NULL, n->value);
-			_iterate(n->center, it);
-		}
-	}
-}
-
-bool radix_tree_is_well_formed(struct radix_tree *rt)
-{
-	return true;
-}
-
-void radix_tree_dump(struct radix_tree *rt, FILE *out)
-{
-}
-
-//----------------------------------------------------------------
-
--- a/base/data-struct/radix-tree.c
+++ b/base/data-struct/radix-tree.c
@@ -1,21 +0,0 @@
-// Copyright (C) 2018 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 Lesser General Public License v.2.1.
-//
-// You should have received a copy of the GNU Lesser General Public License
-// along with this program; if not, write to the Free Software Foundation,
-// Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
-//----------------------------------------------------------------
-
-#ifdef SIMPLE_RADIX_TREE
-#include "base/data-struct/radix-tree-simple.c"
-#else
-#include "base/data-struct/radix-tree-adaptive.c"
-#endif
-
-//----------------------------------------------------------------
--- a/base/data-struct/radix-tree.h
+++ b/base/data-struct/radix-tree.h
@@ -1,64 +0,0 @@
-// Copyright (C) 2018 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 Lesser General Public License v.2.1.
-//
-// You should have received a copy of the GNU Lesser General Public License
-// along with this program; if not, write to the Free Software Foundation,
-// Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- 
-#ifndef BASE_DATA_STRUCT_RADIX_TREE_H
-#define BASE_DATA_STRUCT_RADIX_TREE_H
-
-#include <stdbool.h>
-#include <stdint.h>
-#include <stdio.h>
-
-//----------------------------------------------------------------
-
-struct radix_tree;
-
-union radix_value {
-	void *ptr;
-	uint64_t n;
-};
-
-typedef void (*radix_value_dtr)(void *context, union radix_value v);
-
-// dtr will be called on any deleted entries.  dtr may be NULL.
-struct radix_tree *radix_tree_create(radix_value_dtr dtr, void *dtr_context);
-void radix_tree_destroy(struct radix_tree *rt);
-
-unsigned radix_tree_size(struct radix_tree *rt);
-bool radix_tree_insert(struct radix_tree *rt, uint8_t *kb, uint8_t *ke, union radix_value v);
-bool radix_tree_remove(struct radix_tree *rt, uint8_t *kb, uint8_t *ke);
-
-// Returns the number of values removed
-unsigned radix_tree_remove_prefix(struct radix_tree *rt, uint8_t *prefix_b, uint8_t *prefix_e);
-
-bool radix_tree_lookup(struct radix_tree *rt,
-		       uint8_t *kb, uint8_t *ke, union radix_value *result);
-
-// The radix tree stores entries in lexicographical order.  Which means
-// we can iterate entries, in order.  Or iterate entries with a particular
-// prefix.
-struct radix_tree_iterator {
-        // Returns false if the iteration should end.
-	bool (*visit)(struct radix_tree_iterator *it,
-                      uint8_t *kb, uint8_t *ke, union radix_value v);
-};
-
-void radix_tree_iterate(struct radix_tree *rt, uint8_t *kb, uint8_t *ke,
-                        struct radix_tree_iterator *it);
-
-// Checks that some constraints on the shape of the tree are
-// being held.  For debug only.
-bool radix_tree_is_well_formed(struct radix_tree *rt);
-void radix_tree_dump(struct radix_tree *rt, FILE *out);
-
-//----------------------------------------------------------------
-
-#endif
--- a/base/memory/container_of.h
+++ b/base/memory/container_of.h
@@ -1,23 +0,0 @@
-// Copyright (C) 2018 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 Lesser General Public License v.2.1.
-//
-// You should have received a copy of the GNU Lesser General Public License
-// along with this program; if not, write to the Free Software Foundation,
-// Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
-#ifndef BASE_MEMORY_CONTAINER_OF_H
-#define BASE_MEMORY_CONTAINER_OF_H
-
-//----------------------------------------------------------------
-
-#define container_of(v, t, head) \
-    ((t *)((const char *)(v) - (const char *)&((t *) 0)->head))
-
-//----------------------------------------------------------------
-
-#endif
--- a/base/memory/zalloc.h
+++ b/base/memory/zalloc.h
@@ -1,27 +0,0 @@
-// Copyright (C) 2018 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 Lesser General Public License v.2.1.
-//
-// You should have received a copy of the GNU Lesser General Public License
-// along with this program; if not, write to the Free Software Foundation,
-// Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
-
-#ifndef BASE_MEMORY_ZALLOC_H
-#define BASE_MEMORY_ZALLOC_H
-
-#include <stdlib.h>
-
-//----------------------------------------------------------------
-
-static inline void *zalloc(size_t len)
-{
-	return calloc(1, len);
-}
-
-//----------------------------------------------------------------
-
-#endif
--- a/conf/Makefile.in
+++ b/conf/Makefile.in
@@ -1,5 +1,5 @@
 #
-# Copyright (C) 2004-2018 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2015 Red Hat, Inc. All rights reserved.
 #
 # This file is part of LVM2.
 #
@@ -25,7 +25,6 @@ PROFILES=$(PROFILE_TEMPLATES) \
 	$(srcdir)/cache-smq.profile \
 	$(srcdir)/thin-generic.profile \
 	$(srcdir)/thin-performance.profile \
-	$(srcdir)/vdo-small.profile \
 	$(srcdir)/lvmdbusd.profile

 include $(top_builddir)/make.tmpl
@@ -33,8 +32,8 @@ include $(top_builddir)/make.tmpl
 .PHONY: install_conf install_localconf install_profiles

 generate:
-	$(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withgeneralpreamble --withcomments --ignorelocal --withspaces > example.conf.in
-	$(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withlocalpreamble --withcomments --withspaces local > lvmlocal.conf.in
+	LD_LIBRARY_PATH=$(top_builddir)/libdm:$(LD_LIBRARY_PATH) $(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withgeneralpreamble --withcomments --ignorelocal --withspaces > example.conf.in
+	LD_LIBRARY_PATH=$(top_builddir)/libdm:$(LD_LIBRARY_PATH) $(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withlocalpreamble --withcomments --withspaces local > lvmlocal.conf.in

 install_conf: $(CONFSRC)
 	@if [ ! -e $(confdir)/$(CONFDEST) ]; then \
@@ -49,9 +48,8 @@ install_localconf: $(CONFLOCAL)
 	fi

 install_profiles: $(PROFILES)
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_DIR) $(profiledir)
-	$(Q) $(INSTALL_DATA) $(PROFILES) $(profiledir)/
+	$(INSTALL_DIR) $(profiledir)
+	$(INSTALL_DATA) $(PROFILES) $(profiledir)/

 install_lvm2: install_conf install_localconf install_profiles

--- a/conf/example.conf.in
+++ b/conf/example.conf.in
@@ -88,22 +88,6 @@ devices {
 	# 
 	external_device_info_source = "none"

-	# Configuration option devices/hints.
-	# Use a local file to remember which devices have PVs on them.
-	# Some commands will use this as an optimization to reduce device
-	# scanning, and will only scan the listed PVs. Removing the hint file
-	# will cause lvm to generate a new one. Disable hints if PVs will
-	# be copied onto devices using non-lvm commands, like dd.
-	# 
-	# Accepted values:
-	#   all
-	#     Use all hints.
-	#   none
-	#     Use no hints.
-	# 
-	# This configuration option has an automatic default value.
-	# hints = "all"
-
 	# Configuration option devices/preferred_names.
 	# Select which path name to display for a block device.
 	# If multiple path names exist for a block device, and LVM needs to
@@ -139,6 +123,7 @@ devices {
 	# then the device is accepted. Be careful mixing 'a' and 'r' patterns,
 	# as the combination might produce unexpected results (test changes.)
 	# Run vgscan after changing the filter to regenerate the cache.
+	# See the use_lvmetad comment for a special case regarding filters.
 	# 
 	# Example
 	# Accept every block device:
@@ -158,13 +143,31 @@ devices {
 	# Configuration option devices/global_filter.
 	# Limit the block devices that are used by LVM system components.
 	# Because devices/filter may be overridden from the command line, it is
-	# not suitable for system-wide device filtering, e.g. udev.
+	# not suitable for system-wide device filtering, e.g. udev and lvmetad.
 	# Use global_filter to hide devices from these LVM system components.
 	# The syntax is the same as devices/filter. Devices rejected by
 	# global_filter are not opened by LVM.
 	# This configuration option has an automatic default value.
 	# global_filter = [ "a|.*/|" ]

+	# Configuration option devices/cache_dir.
+	# Directory in which to store the device cache file.
+	# The results of filtering are cached on disk to avoid rescanning dud
+	# devices (which can take a very long time). By default this cache is
+	# stored in a file named .cache. It is safe to delete this file; the
+	# tools regenerate it. If obtain_device_list_from_udev is enabled, the
+	# list of devices is obtained from udev and any existing .cache file
+	# is removed.
+	cache_dir = "@DEFAULT_SYS_DIR@/@DEFAULT_CACHE_SUBDIR@"
+
+	# Configuration option devices/cache_file_prefix.
+	# A prefix used before the .cache file name. See devices/cache_dir.
+	cache_file_prefix = ""
+
+	# Configuration option devices/write_cache_state.
+	# Enable/disable writing the cache file. See devices/cache_dir.
+	write_cache_state = 1
+
 	# Configuration option devices/types.
 	# List of additional acceptable block device types.
 	# These are of device type names from /proc/devices, followed by the
@@ -182,53 +185,14 @@ devices {
 	# present on the system. sysfs must be part of the kernel and mounted.)
 	sysfs_scan = 1

-	# Configuration option devices/scan_lvs.
-	# Scan LVM LVs for layered PVs, allowing LVs to be used as PVs.
-	# When 1, LVM will detect PVs layered on LVs, and caution must be
-	# taken to avoid a host accessing a layered VG that may not belong
-	# to it, e.g. from a guest image. This generally requires excluding
-	# the LVs with device filters. Also, when this setting is enabled,
-	# every LVM command will scan every active LV on the system (unless
-	# filtered), which can cause performance problems on systems with
-	# many active LVs. When this setting is 0, LVM will not detect or
-	# use PVs that exist on LVs, and will not allow a PV to be created on
-	# an LV. The LVs are ignored using a built in device filter that
-	# identifies and excludes LVs.
-	scan_lvs = 0
-
 	# Configuration option devices/multipath_component_detection.
 	# Ignore devices that are components of DM multipath devices.
 	multipath_component_detection = 1

 	# Configuration option devices/md_component_detection.
-	# Enable detection and exclusion of MD component devices.
-	# An MD component device is a block device that MD uses as part
-	# of a software RAID virtual device. When an LVM PV is created
-	# on an MD device, LVM must only use the top level MD device as
-	# the PV, and should ignore the underlying component devices.
-	# In cases where the MD superblock is located at the end of the
-	# component devices, it is more difficult for LVM to consistently
-	# identify an MD component, see the md_component_checks setting.
+	# Ignore devices that are components of software RAID (md) devices.
 	md_component_detection = 1

-	# Configuration option devices/md_component_checks.
-	# The checks LVM should use to detect MD component devices.
-	# MD component devices are block devices used by MD software RAID.
-	# 
-	# Accepted values:
-	#   auto
-	#     LVM will skip scanning the end of devices when it has other
-	#     indications that the device is not an MD component.
-	#   start
-	#     LVM will only scan the start of devices for MD superblocks.
-	#     This does not incur extra I/O by LVM.
-	#   full
-	#     LVM will scan the start and end of devices for MD superblocks.
-	#     This requires an extra read at the end of devices.
-	# 
-	# This configuration option has an automatic default value.
-	# md_component_checks = "auto"
-
 	# Configuration option devices/fw_raid_component_detection.
 	# Ignore devices that are components of firmware RAID devices.
 	# LVM must use an external_device_info_source other than none for this
@@ -236,24 +200,19 @@ devices {
 	fw_raid_component_detection = 0

 	# Configuration option devices/md_chunk_alignment.
-	# Align the start of a PV data area with md device's stripe-width.
+	# Align PV data blocks with md device's stripe-width.
 	# This applies if a PV is placed directly on an md device.
-	# default_data_alignment will be overriden if it is not aligned
-	# with the value detected for this setting.
-	# This setting is overriden by data_alignment_detection,
-	# data_alignment, and the --dataalignment option.
 	md_chunk_alignment = 1

 	# Configuration option devices/default_data_alignment.
-	# Align the start of a PV data area with this number of MiB.
-	# Set to 1 for 1MiB, 2 for 2MiB, etc. Set to 0 to disable.
-	# This setting is overriden by data_alignment and the --dataalignment
-	# option.
+	# Default alignment of the start of a PV data area in MB.
+	# If set to 0, a value of 64KiB will be used.
+	# Set to 1 for 1MiB, 2 for 2MiB, etc.
 	# This configuration option has an automatic default value.
 	# default_data_alignment = 1

 	# Configuration option devices/data_alignment_detection.
-	# Align the start of a PV data area with sysfs io properties.
+	# Detect PV data alignment based on sysfs device information.
 	# The start of a PV data area will be a multiple of minimum_io_size or
 	# optimal_io_size exposed in sysfs. minimum_io_size is the smallest
 	# request the device can perform without incurring a read-modify-write
@@ -261,29 +220,27 @@ devices {
 	# preferred unit of receiving I/O, e.g. MD stripe width.
 	# minimum_io_size is used if optimal_io_size is undefined (0).
 	# If md_chunk_alignment is enabled, that detects the optimal_io_size.
-	# default_data_alignment and md_chunk_alignment will be overriden
-	# if they are not aligned with the value detected for this setting.
-	# This setting is overriden by data_alignment and the --dataalignment
-	# option.
+	# This setting takes precedence over md_chunk_alignment.
 	data_alignment_detection = 1

 	# Configuration option devices/data_alignment.
-	# Align the start of a PV data area with this number of KiB.
-	# When non-zero, this setting overrides default_data_alignment.
-	# Set to 0 to disable, in which case default_data_alignment
-	# is used to align the first PE in units of MiB.
-	# This setting is overriden by the --dataalignment option.
+	# Alignment of the start of a PV data area in KiB.
+	# If a PV is placed directly on an md device and md_chunk_alignment or
+	# data_alignment_detection are enabled, then this setting is ignored.
+	# Otherwise, md_chunk_alignment and data_alignment_detection are
+	# disabled if this is set. Set to 0 to use the default alignment or the
+	# page size, if larger.
 	data_alignment = 0

 	# Configuration option devices/data_alignment_offset_detection.
-	# Shift the start of an aligned PV data area based on sysfs information.
-	# After a PV data area is aligned, it will be shifted by the
+	# Detect PV data alignment offset based on sysfs device information.
+	# The start of a PV aligned data area will be shifted by the
 	# alignment_offset exposed in sysfs. This offset is often 0, but may
 	# be non-zero. Certain 4KiB sector drives that compensate for windows
 	# partitioning will have an alignment_offset of 3584 bytes (sector 7
 	# is the lowest aligned logical block, the 4KiB sectors start at
 	# LBA -1, and consequently sector 63 is aligned on a 4KiB boundary).
-	# This setting is overriden by the --dataalignmentoffset option.
+	# pvcreate --dataalignmentoffset will skip this detection.
 	data_alignment_offset_detection = 1

 	# Configuration option devices/ignore_suspended_devices.
@@ -311,6 +268,14 @@ devices {
 	# different way, making them a better choice for VG stacking.
 	ignore_lvm_mirrors = 1

+	# Configuration option devices/disable_after_error_count.
+	# Number of I/O errors after which a device is skipped.
+	# During each LVM operation, errors received from each device are
+	# counted. If the counter of a device exceeds the limit set here,
+	# no further I/O is sent to that device for the remainder of the
+	# operation. Setting this to 0 disables the counters altogether.
+	disable_after_error_count = 0
+
 	# Configuration option devices/require_restorefile_with_uuid.
 	# Allow use of pvcreate --uuid without requiring --restorefile.
 	require_restorefile_with_uuid = 1
@@ -381,7 +346,7 @@ allocation {
 	maximise_cling = 1

 	# Configuration option allocation/use_blkid_wiping.
-	# Use blkid to detect and erase existing signatures on new PVs and LVs.
+	# Use blkid to detect existing signatures on new PVs and LVs.
 	# The blkid library can detect more signatures than the native LVM
 	# detection code, but may take longer. LVM needs to be compiled with
 	# blkid wiping support for this setting to apply. LVM native detection
@@ -533,154 +498,6 @@ allocation {
 	# Default physical extent size in KiB to use for new VGs.
 	# This configuration option has an automatic default value.
 	# physical_extent_size = 4096
-
-	# Configuration option allocation/vdo_use_compression.
-	# Enables or disables compression when creating a VDO volume.
-	# Compression may be disabled if necessary to maximize performance
-	# or to speed processing of data that is unlikely to compress.
-	# This configuration option has an automatic default value.
-	# vdo_use_compression = 1
-
-	# Configuration option allocation/vdo_use_deduplication.
-	# Enables or disables deduplication when creating a VDO volume.
-	# Deduplication may be disabled in instances where data is not expected
-	# to have good deduplication rates but compression is still desired.
-	# This configuration option has an automatic default value.
-	# vdo_use_deduplication = 1
-
-	# Configuration option allocation/vdo_use_metadata_hints.
-	# Enables or disables whether VDO volume should tag its latency-critical
-	# writes with the REQ_SYNC flag. Some device mapper targets such as dm-raid5
-	# process writes with this flag at a higher priority.
-	# Default is enabled.
-	# This configuration option has an automatic default value.
-	# vdo_use_metadata_hints = 1
-
-	# Configuration option allocation/vdo_minimum_io_size.
-	# The minimum IO size for VDO volume to accept, in bytes.
-	# Valid values are 512 or 4096. The recommended and default value is 4096.
-	# This configuration option has an automatic default value.
-	# vdo_minimum_io_size = 4096
-
-	# Configuration option allocation/vdo_block_map_cache_size_mb.
-	# Specifies the amount of memory in MiB allocated for caching block map
-	# pages for VDO volume. The value must be a multiple of 4096 and must be
-	# at least 128MiB and less than 16TiB. The cache must be at least 16MiB
-	# per logical thread. Note that there is a memory overhead of 15%.
-	# This configuration option has an automatic default value.
-	# vdo_block_map_cache_size_mb = 128
-
-	# Configuration option allocation/vdo_block_map_period.
-	# The speed with which the block map cache writes out modified block map pages.
-	# A smaller era length is likely to reduce the amount time spent rebuilding,
-	# at the cost of increased block map writes during normal operation.
-	# The maximum and recommended value is 16380; the minimum value is 1.
-	# This configuration option has an automatic default value.
-	# vdo_block_map_period = 16380
-
-	# Configuration option allocation/vdo_check_point_frequency.
-	# The default check point frequency for VDO volume.
-	# This configuration option has an automatic default value.
-	# vdo_check_point_frequency = 0
-
-	# Configuration option allocation/vdo_use_sparse_index.
-	# Enables sparse indexing for VDO volume.
-	# This configuration option has an automatic default value.
-	# vdo_use_sparse_index = 0
-
-	# Configuration option allocation/vdo_index_memory_size_mb.
-	# Specifies the amount of index memory in MiB for VDO volume.
-	# The value must be at least 256MiB and at most 1TiB.
-	# This configuration option has an automatic default value.
-	# vdo_index_memory_size_mb = 256
-
-	# Configuration option allocation/vdo_slab_size_mb.
-	# Specifies the size in MiB of the increment by which a VDO is grown.
-	# Using a smaller size constrains the total maximum physical size
-	# that can be accommodated. Must be a power of two between 128MiB and 32GiB.
-	# This configuration option has an automatic default value.
-	# vdo_slab_size_mb = 2048
-
-	# Configuration option allocation/vdo_ack_threads.
-	# Specifies the number of threads to use for acknowledging
-	# completion of requested VDO I/O operations.
-	# The value must be at in range [0..100].
-	# This configuration option has an automatic default value.
-	# vdo_ack_threads = 1
-
-	# Configuration option allocation/vdo_bio_threads.
-	# Specifies the number of threads to use for submitting I/O
-	# operations to the storage device of VDO volume.
-	# The value must be in range [1..100]
-	# Each additional thread after the first will use an additional 18MiB of RAM,
-	# plus 1.12 MiB of RAM per megabyte of configured read cache size.
-	# This configuration option has an automatic default value.
-	# vdo_bio_threads = 1
-
-	# Configuration option allocation/vdo_bio_rotation.
-	# Specifies the number of I/O operations to enqueue for each bio-submission
-	# thread before directing work to the next. The value must be in range [1..1024].
-	# This configuration option has an automatic default value.
-	# vdo_bio_rotation = 64
-
-	# Configuration option allocation/vdo_cpu_threads.
-	# Specifies the number of threads to use for CPU-intensive work such as
-	# hashing or compression for VDO volume. The value must be in range [1..100]
-	# This configuration option has an automatic default value.
-	# vdo_cpu_threads = 2
-
-	# Configuration option allocation/vdo_hash_zone_threads.
-	# Specifies the number of threads across which to subdivide parts of the VDO
-	# processing based on the hash value computed from the block data.
-	# The value must be at in range [0..100].
-	# vdo_hash_zone_threads, vdo_logical_threads and vdo_physical_threads must be
-	# either all zero or all non-zero.
-	# This configuration option has an automatic default value.
-	# vdo_hash_zone_threads = 1
-
-	# Configuration option allocation/vdo_logical_threads.
-	# Specifies the number of threads across which to subdivide parts of the VDO
-	# processing based on the hash value computed from the block data.
-	# A logical thread count of 9 or more will require explicitly specifying
-	# a sufficiently large block map cache size, as well.
-	# The value must be in range [0..100].
-	# vdo_hash_zone_threads, vdo_logical_threads and vdo_physical_threads must be
-	# either all zero or all non-zero.
-	# This configuration option has an automatic default value.
-	# vdo_logical_threads = 1
-
-	# Configuration option allocation/vdo_physical_threads.
-	# Specifies the number of threads across which to subdivide parts of the VDO
-	# processing based on physical block addresses.
-	# Each additional thread after the first will use an additional 10MiB of RAM.
-	# The value must be in range [0..16].
-	# vdo_hash_zone_threads, vdo_logical_threads and vdo_physical_threads must be
-	# either all zero or all non-zero.
-	# This configuration option has an automatic default value.
-	# vdo_physical_threads = 1
-
-	# Configuration option allocation/vdo_write_policy.
-	# Specifies the write policy:
-	# auto  - VDO will check the storage device and determine whether it supports flushes.
-	#         If it does, VDO will run in async mode, otherwise it will run in sync mode.
-	# sync  - Writes are acknowledged only after data is stably written.
-	#         This policy is not supported if the underlying storage is not also synchronous.
-	# async - Writes are acknowledged after data has been cached for writing to stable storage.
-	#         Data which has not been flushed is not guaranteed to persist in this mode.
-	# This configuration option has an automatic default value.
-	# vdo_write_policy = "auto"
-
-	# Configuration option allocation/vdo_max_discard.
-	# Specified te maximum size of discard bio accepted, in 4096 byte blocks.
-	# I/O requests to a VDO volume are normally split into 4096-byte blocks,
-	# and processed up to 2048 at a time. However, discard requests to a VDO volume
-	# can be automatically split to a larger size, up to <max discard> 4096-byte blocks
-	# in a single bio, and are limited to 1500 at a time.
-	# Increasing this value may provide better overall performance, at the cost of
-	# increased latency for the individual discard requests.
-	# The default and minimum is 1. The maximum is UINT_MAX / 4096.
-	# This configuration option has an automatic default value.
-	# vdo_max_discard = 1
 }

 # Configuration section log.
@@ -771,8 +588,7 @@ log {

 	# Configuration option log/indent.
 	# Indent messages according to their severity.
-	# This configuration option has an automatic default value.
-	# indent = 0
+	indent = 1

 	# Configuration option log/command_names.
 	# Display the command name on each line of output.
@@ -795,23 +611,9 @@ log {
 	# Select log messages by class.
 	# Some debugging messages are assigned to a class and only appear in
 	# debug output if the class is listed here. Classes currently
-	# available: memory, devices, io, activation, allocation,
+	# available: memory, devices, activation, allocation, lvmetad,
 	# metadata, cache, locking, lvmpolld. Use "all" to see everything.
-	debug_classes = [ "memory", "devices", "io", "activation", "allocation", "metadata", "cache", "locking", "lvmpolld", "dbus" ]
-
-	# Configuration option log/debug_file_fields.
-	# The fields included in debug output written to log file.
-	# Use "all" to include everything (the default).
-	# This configuration option is advanced.
-	# This configuration option has an automatic default value.
-	# debug_file_fields = [ "time", "command", "fileline", "message" ]
-
-	# Configuration option log/debug_output_fields.
-	# The fields included in debug output written to stderr.
-	# Use "all" to include everything (the default).
-	# This configuration option is advanced.
-	# This configuration option has an automatic default value.
-	# debug_output_fields = [ "time", "command", "fileline", "message" ]
+	debug_classes = [ "memory", "devices", "activation", "allocation", "lvmetad", "metadata", "cache", "locking", "lvmpolld", "dbus" ]
 }

 # Configuration section backup.
@@ -899,6 +701,32 @@ global {
 	# the error messages.
 	activation = 1

+	# Configuration option global/fallback_to_lvm1.
+	# Try running LVM1 tools if LVM cannot communicate with DM.
+	# This option only applies to 2.4 kernels and is provided to help
+	# switch between device-mapper kernels and LVM1 kernels. The LVM1
+	# tools need to be installed with .lvm1 suffices, e.g. vgscan.lvm1.
+	# They will stop working once the lvm2 on-disk metadata format is used.
+	# This configuration option has an automatic default value.
+	# fallback_to_lvm1 = @DEFAULT_FALLBACK_TO_LVM1@
+
+	# Configuration option global/format.
+	# The default metadata format that commands should use.
+	# The -M 1|2 option overrides this setting.
+	# 
+	# Accepted values:
+	#   lvm1
+	#   lvm2
+	# 
+	# This configuration option has an automatic default value.
+	# format = "lvm2"
+
+	# Configuration option global/format_libraries.
+	# Shared libraries that process different metadata formats.
+	# If support for LVM1 metadata was compiled as a shared library use
+	# format_libraries = "liblvm2format1.so"
+	# This configuration option does not have a default value defined.
+
 	# Configuration option global/segment_libraries.
 	# This configuration option does not have a default value defined.

@@ -911,10 +739,57 @@ global {
 	# Location of /etc system configuration directory.
 	etc = "@CONFDIR@"

+	# Configuration option global/locking_type.
+	# Type of locking to use.
+	# 
+	# Accepted values:
+	#   0
+	#     Turns off locking. Warning: this risks metadata corruption if
+	#     commands run concurrently.
+	#   1
+	#     LVM uses local file-based locking, the standard mode.
+	#   2
+	#     LVM uses the external shared library locking_library.
+	#   3
+	#     LVM uses built-in clustered locking with clvmd.
+	#     This is incompatible with lvmetad. If use_lvmetad is enabled,
+	#     LVM prints a warning and disables lvmetad use.
+	#   4
+	#     LVM uses read-only locking which forbids any operations that
+	#     might change metadata.
+	#   5
+	#     Offers dummy locking for tools that do not need any locks.
+	#     You should not need to set this directly; the tools will select
+	#     when to use it instead of the configured locking_type.
+	#     Do not use lvmetad or the kernel device-mapper driver with this
+	#     locking type. It is used by the --readonly option that offers
+	#     read-only access to Volume Group metadata that cannot be locked
+	#     safely because it belongs to an inaccessible domain and might be
+	#     in use, for example a virtual machine image or a disk that is
+	#     shared by a clustered machine.
+	# 
+	locking_type = 1
+
 	# Configuration option global/wait_for_locks.
 	# When disabled, fail if a lock request would block.
 	wait_for_locks = 1

+	# Configuration option global/fallback_to_clustered_locking.
+	# Attempt to use built-in cluster locking if locking_type 2 fails.
+	# If using external locking (type 2) and initialisation fails, with
+	# this enabled, an attempt will be made to use the built-in clustered
+	# locking. Disable this if using a customised locking_library.
+	fallback_to_clustered_locking = 1
+
+	# Configuration option global/fallback_to_local_locking.
+	# Use locking_type 1 (local) if locking_type 2 or 3 fail.
+	# If an attempt to initialise type 2 or type 3 locking failed, perhaps
+	# because cluster components such as clvmd are not running, with this
+	# enabled, an attempt will be made to use local file-based locking
+	# (type 1). If this succeeds, only commands against local VGs will
+	# proceed. VGs marked as clustered will be ignored.
+	fallback_to_local_locking = 1
+
 	# Configuration option global/locking_dir.
 	# Directory to use for LVM command file locks.
 	# Local non-LV directory that holds file-based locks while commands are
@@ -935,12 +810,24 @@ global {
 	# Search this directory first for shared libraries.
 	# This configuration option does not have a default value defined.

+	# Configuration option global/locking_library.
+	# The external locking library to use for locking_type 2.
+	# This configuration option has an automatic default value.
+	# locking_library = "liblvm2clusterlock.so"
+
 	# Configuration option global/abort_on_internal_errors.
 	# Abort a command that encounters an internal error.
 	# Treat any internal errors as fatal errors, aborting the process that
 	# encountered the internal error. Please only enable for debugging.
 	abort_on_internal_errors = 0

+	# Configuration option global/detect_internal_vg_cache_corruption.
+	# Internal verification of VG structures.
+	# Check if CRC matches when a parsed VG is used multiple times. This
+	# is useful to catch unexpected changes to cached VG structures.
+	# Please only enable for debugging.
+	detect_internal_vg_cache_corruption = 0
+
 	# Configuration option global/metadata_read_only.
 	# No operations that change on-disk metadata are permitted.
 	# Additionally, read-only commands that encounter metadata in need of
@@ -975,16 +862,6 @@ global {
 	# 
 	mirror_segtype_default = "@DEFAULT_MIRROR_SEGTYPE@"

-	# Configuration option global/support_mirrored_mirror_log.
-	# Enable mirrored 'mirror' log type for testing.
-	# 
-	# This type is deprecated to create or convert to but can
-	# be enabled to test that activation of existing mirrored
-	# logs and conversion to disk/core works.
-	# 
-	# Not supported for regular operation!
-	support_mirrored_mirror_log = 0
-
 	# Configuration option global/raid10_segtype_default.
 	# The segment type used by the -i -m combination.
 	# The --type raid10|mirror option overrides this setting.
@@ -1033,20 +910,41 @@ global {
 	# This configuration option has an automatic default value.
 	# lvdisplay_shows_full_device_path = 0

-	# Configuration option global/event_activation.
-	# Activate LVs based on system-generated device events.
-	# When a device appears on the system, a system-generated event runs
-	# the pvscan command to activate LVs if the new PV completes the VG.
-	# Use auto_activation_volume_list to select which LVs should be
-	# activated from these events (the default is all.)
-	# When event_activation is disabled, the system will generally run
-	# a direct activation command to activate LVs in complete VGs.
-	event_activation = 1
+	# Configuration option global/use_lvmetad.
+	# Use lvmetad to cache metadata and reduce disk scanning.
+	# When enabled (and running), lvmetad provides LVM commands with VG
+	# metadata and PV state. LVM commands then avoid reading this
+	# information from disks which can be slow. When disabled (or not
+	# running), LVM commands fall back to scanning disks to obtain VG
+	# metadata. lvmetad is kept updated via udev rules which must be set
+	# up for LVM to work correctly. (The udev rules should be installed
+	# by default.) Without a proper udev setup, changes in the system's
+	# block device configuration will be unknown to LVM, and ignored
+	# until a manual 'pvscan --cache' is run. If lvmetad was running
+	# while use_lvmetad was disabled, it must be stopped, use_lvmetad
+	# enabled, and then started. When using lvmetad, LV activation is
+	# switched to an automatic, event-based mode. In this mode, LVs are
+	# activated based on incoming udev events that inform lvmetad when
+	# PVs appear on the system. When a VG is complete (all PVs present),
+	# it is auto-activated. The auto_activation_volume_list setting
+	# controls which LVs are auto-activated (all by default.)
+	# When lvmetad is updated (automatically by udev events, or directly
+	# by pvscan --cache), devices/filter is ignored and all devices are
+	# scanned by default. lvmetad always keeps unfiltered information
+	# which is provided to LVM commands. Each LVM command then filters
+	# based on devices/filter. This does not apply to other, non-regexp,
+	# filtering settings: component filters such as multipath and MD
+	# are checked during pvscan --cache. To filter a device and prevent
+	# scanning from the LVM system entirely, including lvmetad, use
+	# devices/global_filter.
+	use_lvmetad = @DEFAULT_USE_LVMETAD@

-	# Configuration option global/use_aio.
-	# Use async I/O when reading and writing devices.
+	# Configuration option global/lvmetad_update_wait_time.
+	# Number of seconds a command will wait for lvmetad update to finish.
+	# After waiting for this period, a command will not use lvmetad, and
+	# will revert to disk scanning.
 	# This configuration option has an automatic default value.
-	# use_aio = 1
+	# lvmetad_update_wait_time = 10

 	# Configuration option global/use_lvmlockd.
 	# Use lvmlockd for locking among hosts using LVM on shared storage.
@@ -1172,17 +1070,6 @@ global {
 	# This configuration option has an automatic default value.
 	# cache_repair_options = [ "" ]

-	# Configuration option global/vdo_format_executable.
-	# The full path to the vdoformat command.
-	# LVM uses this command to initial data volume for VDO type logical volume
-	# This configuration option has an automatic default value.
-	# vdo_format_executable = "@VDO_FORMAT_CMD@"
-
-	# Configuration option global/vdo_format_options.
-	# List of options passed added to standard vdoformat command.
-	# This configuration option has an automatic default value.
-	# vdo_format_options = [ "" ]
-
 	# Configuration option global/fsadm_executable.
 	# The full path to the fsadm command.
 	# LVM uses this command to help with lvresize -r operations.
@@ -1240,16 +1127,6 @@ global {
 	# When enabled, an LVM command that changes PVs, changes VG metadata,
 	# or changes the activation state of an LV will send a notification.
 	notify_dbus = 1
-
-	# Configuration option global/io_memory_size.
-	# The amount of memory in KiB that LVM allocates to perform disk io.
-	# LVM performance may benefit from more io memory when there are many
-	# disks or VG metadata is large. Increasing this size may be necessary
-	# when a single copy of VG metadata is larger than the current setting.
-	# This value should usually not be decreased from the default; setting
-	# it too low can result in lvm failing to read VGs.
-	# This configuration option has an automatic default value.
-	# io_memory_size = 8192
 }

 # Configuration section activation.
@@ -1566,33 +1443,6 @@ activation {
 	# 
 	thin_pool_autoextend_percent = 20

-	# Configuration option activation/vdo_pool_autoextend_threshold.
-	# Auto-extend a VDO pool when its usage exceeds this percent.
-	# Setting this to 100 disables automatic extension.
-	# The minimum value is 50 (a smaller value is treated as 50.)
-	# Also see vdo_pool_autoextend_percent.
-	# Automatic extension requires dmeventd to be monitoring the LV.
-	# 
-	# Example
-	# Using 70% autoextend threshold and 20% autoextend size, when a 10G
-	# VDO pool exceeds 7G, it is extended to 12G, and when it exceeds
-	# 8.4G, it is extended to 14.4G:
-	# vdo_pool_autoextend_threshold = 70
-	# 
-	vdo_pool_autoextend_threshold = 100
-
-	# Configuration option activation/vdo_pool_autoextend_percent.
-	# Auto-extending a VDO pool adds this percent extra space.
-	# The amount of additional space added to a VDO pool is this
-	# percent of its current size.
-	# 
-	# Example
-	# Using 70% autoextend threshold and 20% autoextend size, when a 10G
-	# VDO pool exceeds 7G, it is extended to 12G, and when it exceeds
-	# 8.4G, it is extended to 14.4G:
-	# This configuration option has an automatic default value.
-	# vdo_pool_autoextend_percent = 20
-
 	# Configuration option activation/mlock_filter.
 	# Do not mlock these memory areas.
 	# While activating devices, I/O to devices being (re)configured is
@@ -1738,20 +1588,13 @@ activation {
 	# vgmetadatacopies = 0

 	# Configuration option metadata/pvmetadatasize.
-	# The default size of the metadata area in units of 512 byte sectors.
-	# The metadata area begins at an offset of the page size from the start
-	# of the device. The first PE is by default at 1 MiB from the start of
-	# the device. The space between these is the default metadata area size.
-	# The actual size of the metadata area may be larger than what is set
-	# here due to default_data_alignment making the first PE a MiB multiple.
-	# The metadata area begins with a 512 byte header and is followed by a
-	# circular buffer used for VG metadata text. The maximum size of the VG
-	# metadata is about half the size of the metadata buffer. VGs with large
-	# numbers of PVs or LVs, or VGs containing complex LV structures, may need
-	# additional space for VG metadata. The --metadatasize option overrides
-	# this setting.
-	# This configuration option does not have a default value defined.
+	# Approximate number of sectors to use for each metadata copy.
+	# VGs with large numbers of PVs or LVs, or VGs containing complex LV
+	# structures, may need additional space for VG metadata. The metadata
+	# areas are treated as circular buffers, so unused space becomes filled
+	# with an archive of the most recent previous versions of the metadata.
 	# This configuration option has an automatic default value.
+	# pvmetadatasize = 255

 	# Configuration option metadata/pvmetadataignore.
 	# Ignore metadata areas on a new PV.
@@ -1766,6 +1609,24 @@ activation {
 	# This configuration option is advanced.
 	# This configuration option has an automatic default value.
 	# stripesize = 64
+
+	# Configuration option metadata/dirs.
+	# Directories holding live copies of text format metadata.
+	# These directories must not be on logical volumes!
+	# It's possible to use LVM with a couple of directories here,
+	# preferably on different (non-LV) filesystems, and with no other
+	# on-disk metadata (pvmetadatacopies = 0). Or this can be in addition
+	# to on-disk metadata areas. The feature was originally added to
+	# simplify testing and is not supported under low memory situations -
+	# the machine could lock up. Never edit any files in these directories
+	# by hand unless you are absolutely sure you know what you are doing!
+	# Use the supplied toolset to make changes (e.g. vgcfgrestore).
+	# 
+	# Example
+	# dirs = [ "/etc/lvm/metadata", "/mnt/disk2/lvm/metadata2" ]
+	# 
+	# This configuration option is advanced.
+	# This configuration option does not have a default value defined.
 # }

 # Configuration section report.
@@ -2216,23 +2077,6 @@ dmeventd {
 	# This configuration option has an automatic default value.
 	# thin_command = "lvm lvextend --use-policies"

-	# Configuration option dmeventd/vdo_library.
-	# The library dmeventd uses when monitoring a VDO pool device.
-	# libdevmapper-event-lvm2vdo.so monitors the filling of a pool
-	# and emits a warning through syslog when the usage exceeds 80%. The
-	# warning is repeated when 85%, 90% and 95% of the pool is filled.
-	# This configuration option has an automatic default value.
-	# vdo_library = "libdevmapper-event-lvm2vdo.so"
-
-	# Configuration option dmeventd/vdo_command.
-	# The plugin runs command with each 5% increment when VDO pool volume
-	# gets above 50%.
-	# Command which starts with 'lvm ' prefix is internal lvm command.
-	# You can write your own handler to customise behaviour in more details.
-	# User handler is specified with the full path starting with '/'.
-	# This configuration option has an automatic default value.
-	# vdo_command = "lvm lvextend --use-policies"
-
 	# Configuration option dmeventd/executable.
 	# The full path to the dmeventd binary.
 	# This configuration option has an automatic default value.
--- a/conf/vdo-small.profile
+++ b/conf/vdo-small.profile
@@ -1,24 +0,0 @@
-# Demo configuration for 'VDO' using less memory.
-# ~lvmconfig --type full | grep vdo
-
-allocation {
-	vdo_use_compression=1
-	vdo_use_deduplication=1
-	vdo_use_metadata_hints=1
-	vdo_minimum_io_size=4096
-	vdo_block_map_cache_size_mb=128
-	vdo_block_map_period=16380
-	vdo_check_point_frequency=0
-	vdo_use_sparse_index=0
-	vdo_index_memory_size_mb=256
-	vdo_slab_size_mb=2048
-	vdo_ack_threads=1
-	vdo_bio_threads=1
-	vdo_bio_rotation=64
-	vdo_cpu_threads=2
-	vdo_hash_zone_threads=1
-	vdo_logical_threads=1
-	vdo_physical_threads=1
-	vdo_write_policy="auto"
-	vdo_max_discard=1
-}
--- a/2832
+++ b/2832
--- a/configure.in
+++ b/configure.in
--- a/daemons/Makefile.in
+++ b/daemons/Makefile.in
@@ -15,7 +15,11 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

-.PHONY: dmeventd cmirrord lvmpolld lvmlockd
+.PHONY: dmeventd clvmd cmirrord lvmetad lvmpolld lvmlockd
+
+ifneq ("@CLVMD@", "none")
+  SUBDIRS += clvmd
+endif

 ifeq ("@BUILD_CMIRRORD@", "yes")
  SUBDIRS += cmirrord
@@ -28,6 +32,10 @@ daemons.cflow: dmeventd.cflow
 endif
 endif

+ifeq ("@BUILD_LVMETAD@", "yes")
+  SUBDIRS += lvmetad
+endif
+
 ifeq ("@BUILD_LVMPOLLD@", "yes")
  SUBDIRS += lvmpolld
 endif
@@ -40,8 +48,12 @@ ifeq ("@BUILD_LVMDBUSD@", "yes")
  SUBDIRS += lvmdbusd
 endif

+ifeq ("@BUILD_DMFILEMAPD@", "yes")
+  SUBDIRS += dmfilemapd
+endif
+
 ifeq ($(MAKECMDGOALS),distclean)
-  SUBDIRS = cmirrord dmeventd lvmpolld lvmlockd lvmdbusd
+  SUBDIRS = clvmd cmirrord dmeventd lvmetad lvmpolld lvmlockd lvmdbusd dmfilemapd
 endif

 include $(top_builddir)/make.tmpl
--- a/daemons/clvmd/.gitignore
+++ b/daemons/clvmd/.gitignore
@@ -0,0 +1 @@
+clvmd
--- a/daemons/clvmd/Makefile.in
+++ b/daemons/clvmd/Makefile.in
@@ -0,0 +1,98 @@
+#
+# Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+#
+# This file is part of LVM2.
+#
+# This copyrighted material is made available to anyone wishing to use,
+# modify, copy, or redistribute it subject to the terms and conditions
+# of the GNU General Public License v.2.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+top_builddir = @top_builddir@
+
+CMAN_LIBS = @CMAN_LIBS@
+CMAN_CFLAGS = @CMAN_CFLAGS@
+CMAP_LIBS = @CMAP_LIBS@
+CMAP_CFLAGS = @CMAP_CFLAGS@
+CONFDB_LIBS = @CONFDB_LIBS@
+CONFDB_CFLAGS = @CONFDB_CFLAGS@
+CPG_LIBS = @CPG_LIBS@
+CPG_CFLAGS = @CPG_CFLAGS@
+DLM_LIBS = @DLM_LIBS@
+DLM_CFLAGS = @DLM_CFLAGS@
+QUORUM_LIBS = @QUORUM_LIBS@
+QUORUM_CFLAGS = @QUORUM_CFLAGS@
+SALCK_LIBS = @SALCK_LIBS@
+SALCK_CFLAGS = @SALCK_CFLAGS@
+
+SOURCES = \
+	clvmd-command.c\
+	clvmd.c\
+	lvm-functions.c\
+	refresh_clvmd.c
+
+ifneq (,$(findstring cman,, "@CLVMD@,"))
+	SOURCES += clvmd-cman.c
+	LMLIBS += $(CMAN_LIBS) $(CONFDB_LIBS) $(DLM_LIBS)
+	CFLAGS += $(CMAN_CFLAGS) $(CONFDB_CFLAGS) $(DLM_CFLAGS)
+	DEFS += -DUSE_CMAN
+endif
+
+ifneq (,$(findstring openais,, "@CLVMD@,"))
+	SOURCES += clvmd-openais.c
+	LMLIBS += $(CONFDB_LIBS) $(CPG_LIBS) $(SALCK_LIBS)
+	CFLAGS += $(CONFDB_CFLAGS) $(CPG_CFLAGS) $(SALCK_CFLAGS)
+	DEFS += -DUSE_OPENAIS
+endif
+
+ifneq (,$(findstring corosync,, "@CLVMD@,"))
+	SOURCES += clvmd-corosync.c
+	LMLIBS += $(CMAP_LIBS) $(CONFDB_LIBS) $(CPG_LIBS) $(DLM_LIBS) $(QUORUM_LIBS)
+	CFLAGS += $(CMAP_CFLAGS) $(CONFDB_CFLAGS) $(CPG_CFLAGS) $(DLM_CFLAGS) $(QUORUM_CFLAGS)
+	DEFS += -DUSE_COROSYNC
+endif
+
+ifneq (,$(findstring singlenode,, &quot;@CLVMD@,&quot;))
+	SOURCES += clvmd-singlenode.c
+	DEFS += -DUSE_SINGLENODE
+endif
+
+ifeq ($(MAKECMDGOALS),distclean)
+	SOURCES += clvmd-cman.c
+	SOURCES += clvmd-openais.c
+	SOURCES += clvmd-corosync.c
+	SOURCES += clvmd-singlenode.c
+endif
+
+TARGETS = \
+	clvmd
+
+include $(top_builddir)/make.tmpl
+
+LIBS += $(LVMINTERNAL_LIBS) -ldevmapper $(PTHREAD_LIBS)
+CFLAGS += -fno-strict-aliasing $(EXTRA_EXEC_CFLAGS)
+
+ifeq ("@AIO@", "yes")
+	LIBS += $(AIO_LIBS)
+endif
+
+INSTALL_TARGETS = \
+	install_clvmd
+
+clvmd: $(OBJECTS) $(top_builddir)/lib/liblvm-internal.a
+	$(CC) $(CFLAGS) $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) \
+	      -o clvmd $(OBJECTS) $(LMLIBS) $(LIBS)
+
+.PHONY: install_clvmd
+
+install_clvmd: $(TARGETS)
+	$(INSTALL_PROGRAM) -D clvmd $(usrsbindir)/clvmd
+
+install: $(INSTALL_TARGETS)
+
+install_cluster: $(INSTALL_TARGETS)
--- a/daemons/clvmd/clvm.h
+++ b/daemons/clvmd/clvm.h
@@ -0,0 +1,85 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2007 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/* Definitions for CLVMD server and clients */
+
+/*
+ * The protocol spoken over the cluster and across the local socket.
+ */
+
+#ifndef _CLVM_H
+#define _CLVM_H
+
+#include "configure.h"
+#include <inttypes.h>
+
+struct clvm_header {
+	uint8_t  cmd;	        /* See below */
+	uint8_t  flags;	        /* See below */
+	uint16_t xid;	        /* Transaction ID */
+	uint32_t clientid;	/* Only used in Daemon->Daemon comms */
+	int32_t  status;	/* For replies, whether request succeeded */
+	uint32_t arglen;	/* Length of argument below.
+				   If >1500 then it will be passed
+				   around the cluster in the system LV */
+	char node[1];		/* Actually a NUL-terminated string, node name.
+				   If this is empty then the command is
+				   forwarded to all cluster nodes unless
+				   FLAG_LOCAL or FLAG_REMOTE is also set. */
+	char args[1];		/* Arguments for the command follow the
+				   node name, This member is only
+				   valid if the node name is empty */
+} __attribute__ ((packed));
+
+/* Flags */
+#define CLVMD_FLAG_LOCAL	1	/* Only do this on the local node */
+#define CLVMD_FLAG_SYSTEMLV	2	/* Data in system LV under my node name */
+#define CLVMD_FLAG_NODEERRS	4	/* Reply has errors in node-specific portion */
+#define CLVMD_FLAG_REMOTE	8	/* Do this on all nodes except for the local node */
+
+/* Name of the local socket to communicate between lvm and clvmd */
+#define CLVMD_SOCKNAME		DEFAULT_RUN_DIR "/clvmd.sock"
+
+/* Internal commands & replies */
+#define CLVMD_CMD_REPLY    1
+#define CLVMD_CMD_VERSION  2	/* Send version around cluster when we start */
+#define CLVMD_CMD_GOAWAY   3	/* Die if received this - we are running
+				   an incompatible version */
+#define CLVMD_CMD_TEST     4	/* Just for mucking about */
+
+#define CLVMD_CMD_LOCK              30
+#define CLVMD_CMD_UNLOCK            31
+
+/* Lock/Unlock commands */
+#define CLVMD_CMD_LOCK_LV           50
+#define CLVMD_CMD_LOCK_VG           51
+#define CLVMD_CMD_LOCK_QUERY	    52
+
+/* Misc functions */
+#define CLVMD_CMD_REFRESH	    40
+#define CLVMD_CMD_GET_CLUSTERNAME   41
+#define CLVMD_CMD_SET_DEBUG	    42
+#define CLVMD_CMD_VG_BACKUP	    43
+#define CLVMD_CMD_RESTART	    44
+#define CLVMD_CMD_SYNC_NAMES	    45
+
+/* Used internally by some callers, but not part of the protocol.*/
+#ifndef NODE_ALL
+#  define NODE_ALL	"*"
+#  define NODE_LOCAL	"."
+#  define NODE_REMOTE	"^"
+#endif
+
+#endif
--- a/daemons/clvmd/clvmd-cman.c
+++ b/daemons/clvmd/clvmd-cman.c
@@ -0,0 +1,505 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+ * CMAN communication layer for clvmd.
+ */
+
+#include "clvmd-common.h"
+
+#include <pthread.h>
+
+#include "clvmd-comms.h"
+#include "clvm.h"
+#include "clvmd.h"
+#include "lvm-functions.h"
+
+#include <libdlm.h>
+
+#include <syslog.h>
+
+#define LOCKSPACE_NAME "clvmd"
+
+struct clvmd_node
+{
+	struct cman_node *node;
+	int clvmd_up;
+};
+
+static int num_nodes;
+static struct cman_node *nodes = NULL;
+static struct cman_node this_node;
+static int count_nodes; /* size of allocated nodes array */
+static struct dm_hash_table *node_updown_hash;
+static dlm_lshandle_t *lockspace;
+static cman_handle_t c_handle;
+
+static void count_clvmds_running(void);
+static void get_members(void);
+static int nodeid_from_csid(const char *csid);
+static int name_from_nodeid(int nodeid, char *name);
+static void event_callback(cman_handle_t handle, void *private, int reason, int arg);
+static void data_callback(cman_handle_t handle, void *private,
+			  char *buf, int len, uint8_t port, int nodeid);
+
+struct lock_wait {
+	pthread_cond_t cond;
+	pthread_mutex_t mutex;
+	struct dlm_lksb lksb;
+};
+
+static int _init_cluster(void)
+{
+	node_updown_hash = dm_hash_create(100);
+
+	/* Open the cluster communication socket */
+	c_handle = cman_init(NULL);
+	if (!c_handle) {
+		syslog(LOG_ERR, "Can't open cluster manager socket: %m");
+		return -1;
+	}
+	DEBUGLOG("Connected to CMAN\n");
+
+	if (cman_start_recv_data(c_handle, data_callback, CLUSTER_PORT_CLVMD)) {
+		syslog(LOG_ERR, "Can't bind cluster socket: %m");
+		return -1;
+	}
+
+	if (cman_start_notification(c_handle, event_callback)) {
+		syslog(LOG_ERR, "Can't start cluster event listening");
+		return -1;
+	}
+
+	/* Get the cluster members list */
+	get_members();
+	count_clvmds_running();
+
+	DEBUGLOG("CMAN initialisation complete\n");
+
+	/* Create a lockspace for LV & VG locks to live in */
+	lockspace = dlm_open_lockspace(LOCKSPACE_NAME);
+	if (!lockspace) {
+		lockspace = dlm_create_lockspace(LOCKSPACE_NAME, 0600);
+		if (!lockspace) {
+			syslog(LOG_ERR, "Unable to create DLM lockspace for CLVM: %m");
+			return -1;
+		}
+		DEBUGLOG("Created DLM lockspace for CLVMD.\n");
+	} else
+		DEBUGLOG("Opened existing DLM lockspace for CLVMD.\n");
+
+	dlm_ls_pthread_init(lockspace);
+	DEBUGLOG("DLM initialisation complete\n");
+	return 0;
+}
+
+static void _cluster_init_completed(void)
+{
+	clvmd_cluster_init_completed();
+}
+
+static int _get_main_cluster_fd(void)
+{
+	return cman_get_fd(c_handle);
+}
+
+static int _get_num_nodes(void)
+{
+	int i;
+	int nnodes = 0;
+
+	/* return number of ACTIVE nodes */
+	for (i=0; i<num_nodes; i++) {
+		if (nodes[i].cn_member && nodes[i].cn_nodeid)
+			nnodes++;
+	}
+	return nnodes;
+}
+
+/* send_message with the fd check removed */
+static int _cluster_send_message(const void *buf, int msglen, const char *csid,
+				 const char *errtext)
+{
+	int nodeid = 0;
+
+	if (csid)
+		memcpy(&nodeid, csid, CMAN_MAX_CSID_LEN);
+
+	if (cman_send_data(c_handle, buf, msglen, 0, CLUSTER_PORT_CLVMD, nodeid) <= 0)
+	{
+		log_error("%s", errtext);
+	}
+	return msglen;
+}
+
+static void _get_our_csid(char *csid)
+{
+	if (this_node.cn_nodeid == 0) {
+		cman_get_node(c_handle, 0, &this_node);
+	}
+	memcpy(csid, &this_node.cn_nodeid, CMAN_MAX_CSID_LEN);
+}
+
+/* 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 *,
+						       const char *,
+						       int))
+{
+	int i;
+	int somedown = 0;
+
+	for (i = 0; i < _get_num_nodes(); i++) {
+		if (nodes[i].cn_member && nodes[i].cn_nodeid) {
+			int up = (int)(long)dm_hash_lookup_binary(node_updown_hash, (char *)&nodes[i].cn_nodeid, sizeof(int));
+
+			callback(client, (char *)&nodes[i].cn_nodeid, up);
+			if (!up)
+				somedown = -1;
+		}
+	}
+	return somedown;
+}
+
+/* Process OOB messages from the cluster socket */
+static void event_callback(cman_handle_t handle, void *private, int reason, int arg)
+{
+	char namebuf[MAX_CLUSTER_MEMBER_NAME_LEN];
+
+	switch (reason) {
+        case CMAN_REASON_PORTCLOSED:
+		name_from_nodeid(arg, namebuf);
+		log_notice("clvmd on node %s has died\n", namebuf);
+		DEBUGLOG("Got port closed message, removing node %s\n", namebuf);
+
+		dm_hash_insert_binary(node_updown_hash, (char *)&arg, sizeof(int), (void *)0);
+		break;
+
+	case CMAN_REASON_STATECHANGE:
+		DEBUGLOG("Got state change message, re-reading members list\n");
+		get_members();
+		break;
+
+#if defined(LIBCMAN_VERSION) && LIBCMAN_VERSION >= 2
+	case CMAN_REASON_PORTOPENED:
+		/* Ignore this, wait for startup message from clvmd itself */
+		break;
+
+	case CMAN_REASON_TRY_SHUTDOWN:
+		DEBUGLOG("Got try shutdown, sending OK\n");
+		cman_replyto_shutdown(c_handle, 1);
+		break;
+#endif
+	default:
+		/* ERROR */
+		DEBUGLOG("Got unknown event callback message: %d\n", reason);
+		break;
+	}
+}
+
+static struct local_client *cman_client;
+static int _cluster_fd_callback(struct local_client *fd, char *buf, int len,
+				const char *csid,
+				struct local_client **new_client)
+{
+
+	/* Save this for data_callback */
+	cman_client = fd;
+
+	/* We never return a new client */
+	*new_client = NULL;
+
+	return cman_dispatch(c_handle, 0);
+}
+
+
+static void data_callback(cman_handle_t handle, void *private,
+			  char *buf, int len, uint8_t port, int nodeid)
+{
+	/* Ignore looped back messages */
+	if (nodeid == this_node.cn_nodeid)
+		return;
+	process_message(cman_client, buf, len, (char *)&nodeid);
+}
+
+static void _add_up_node(const char *csid)
+{
+	/* It's up ! */
+	int nodeid = nodeid_from_csid(csid);
+
+	dm_hash_insert_binary(node_updown_hash, (char *)&nodeid, sizeof(int), (void *)1);
+	DEBUGLOG("Added new node %d to updown list\n", nodeid);
+}
+
+static void _cluster_closedown(void)
+{
+	dlm_release_lockspace(LOCKSPACE_NAME, lockspace, 1);
+	cman_finish(c_handle);
+}
+
+static int is_listening(int nodeid)
+{
+	int status;
+
+	do {
+		status = cman_is_listening(c_handle, nodeid, CLUSTER_PORT_CLVMD);
+		if (status < 0 && errno == EBUSY) {	/* Don't busywait */
+			sleep(1);
+			errno = EBUSY;	/* In case sleep trashes it */
+		}
+	}
+	while (status < 0 && errno == EBUSY);
+
+	return status;
+}
+
+/* Populate the list of CLVMDs running.
+   called only at startup time */
+static void count_clvmds_running(void)
+{
+	int i;
+
+	for (i = 0; i < num_nodes; i++) {
+		int nodeid = nodes[i].cn_nodeid;
+
+		if (is_listening(nodeid) == 1)
+			dm_hash_insert_binary(node_updown_hash, (void *)&nodeid, sizeof(int), (void*)1);
+		else
+			dm_hash_insert_binary(node_updown_hash, (void *)&nodeid, sizeof(int), (void*)0);
+	}
+}
+
+/* Get a list of active cluster members */
+static void get_members(void)
+{
+	int retnodes;
+	int status;
+	int i;
+	int high_nodeid = 0;
+
+	num_nodes = cman_get_node_count(c_handle);
+	if (num_nodes == -1) {
+		log_error("Unable to get node count");
+		return;
+	}
+
+	/* 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 */
+		nodes = malloc(count_nodes * sizeof(struct cman_node));
+		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);
+	}
+
+	/* Get the highest nodeid */
+	for (i=0; i<retnodes; i++) {
+		if (nodes[i].cn_nodeid > high_nodeid)
+			high_nodeid = nodes[i].cn_nodeid;
+	}
+}
+
+
+/* Convert a node name to a CSID */
+static int _csid_from_name(char *csid, const char *name)
+{
+	int i;
+
+	for (i = 0; i < num_nodes; i++) {
+		if (strcmp(name, nodes[i].cn_name) == 0) {
+			memcpy(csid, &nodes[i].cn_nodeid, CMAN_MAX_CSID_LEN);
+			return 0;
+		}
+	}
+	return -1;
+}
+
+/* Convert a CSID to a node name */
+static int _name_from_csid(const char *csid, char *name)
+{
+	int i;
+
+	for (i = 0; i < num_nodes; i++) {
+		if (memcmp(csid, &nodes[i].cn_nodeid, CMAN_MAX_CSID_LEN) == 0) {
+			strcpy(name, nodes[i].cn_name);
+			return 0;
+		}
+	}
+	/* Who?? */
+	strcpy(name, "Unknown");
+	return -1;
+}
+
+/* Convert a node ID to a node name */
+static int name_from_nodeid(int nodeid, char *name)
+{
+	int i;
+
+	for (i = 0; i < num_nodes; i++) {
+		if (nodeid == nodes[i].cn_nodeid) {
+			strcpy(name, nodes[i].cn_name);
+			return 0;
+		}
+	}
+	/* Who?? */
+	strcpy(name, "Unknown");
+	return -1;
+}
+
+/* Convert a CSID to a node ID */
+static int nodeid_from_csid(const char *csid)
+{
+        int nodeid;
+
+	memcpy(&nodeid, csid, CMAN_MAX_CSID_LEN);
+
+	return nodeid;
+}
+
+static int _is_quorate(void)
+{
+	return cman_is_quorate(c_handle);
+}
+
+static void sync_ast_routine(void *arg)
+{
+	struct lock_wait *lwait = arg;
+
+	pthread_mutex_lock(&lwait->mutex);
+	pthread_cond_signal(&lwait->cond);
+	pthread_mutex_unlock(&lwait->mutex);
+}
+
+static int _sync_lock(const char *resource, int mode, int flags, int *lockid)
+{
+	int status;
+	struct lock_wait lwait;
+
+	if (!lockid) {
+		errno = EINVAL;
+		return -1;
+	}
+
+	DEBUGLOG("sync_lock: '%s' mode:%d flags=%d\n", resource,mode,flags);
+	/* Conversions need the lockid in the LKSB */
+	if (flags & LKF_CONVERT)
+		lwait.lksb.sb_lkid = *lockid;
+
+	pthread_cond_init(&lwait.cond, NULL);
+	pthread_mutex_init(&lwait.mutex, NULL);
+	pthread_mutex_lock(&lwait.mutex);
+
+	status = dlm_ls_lock(lockspace,
+			     mode,
+			     &lwait.lksb,
+			     flags,
+			     resource,
+			     strlen(resource),
+			     0, sync_ast_routine, &lwait, NULL, NULL);
+	if (status)
+		return status;
+
+	/* Wait for it to complete */
+	pthread_cond_wait(&lwait.cond, &lwait.mutex);
+	pthread_mutex_unlock(&lwait.mutex);
+
+	*lockid = lwait.lksb.sb_lkid;
+
+	errno = lwait.lksb.sb_status;
+	DEBUGLOG("sync_lock: returning lkid %x\n", *lockid);
+	if (lwait.lksb.sb_status)
+		return -1;
+	else
+		return 0;
+}
+
+static int _sync_unlock(const char *resource /* UNUSED */, int lockid)
+{
+	int status;
+	struct lock_wait lwait;
+
+	DEBUGLOG("sync_unlock: '%s' lkid:%x\n", resource, lockid);
+
+	pthread_cond_init(&lwait.cond, NULL);
+	pthread_mutex_init(&lwait.mutex, NULL);
+	pthread_mutex_lock(&lwait.mutex);
+
+	status = dlm_ls_unlock(lockspace, lockid, 0, &lwait.lksb, &lwait);
+
+	if (status)
+		return status;
+
+	/* Wait for it to complete */
+	pthread_cond_wait(&lwait.cond, &lwait.mutex);
+	pthread_mutex_unlock(&lwait.mutex);
+
+	errno = lwait.lksb.sb_status;
+	if (lwait.lksb.sb_status != EUNLOCK)
+		return -1;
+	else
+		return 0;
+
+}
+
+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 = {
+	.name                     = "cman",
+	.cluster_init_completed   = _cluster_init_completed,
+	.cluster_send_message     = _cluster_send_message,
+	.name_from_csid           = _name_from_csid,
+	.csid_from_name           = _csid_from_name,
+	.get_num_nodes            = _get_num_nodes,
+	.cluster_fd_callback      = _cluster_fd_callback,
+	.get_main_cluster_fd      = _get_main_cluster_fd,
+	.cluster_do_node_callback = _cluster_do_node_callback,
+	.is_quorate               = _is_quorate,
+	.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,
+};
+
+struct cluster_ops *init_cman_cluster(void)
+{
+	if (!_init_cluster())
+		return &_cluster_cman_ops;
+	else
+		return NULL;
+}
--- a/daemons/clvmd/clvmd-command.c
+++ b/daemons/clvmd/clvmd-command.c
@@ -0,0 +1,416 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2011 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+
+  CLVMD Cluster LVM daemon command processor.
+
+  To add commands to the daemon simply add a processor in do_command and return
+  and messages back in buf and the length in *retlen. The initial value of
+  buflen is the maximum size of the buffer. if buf is not large enough then it
+  may be reallocated by the functions in here to a suitable size bearing in
+  mind that anything larger than the passed-in size will have to be returned
+  using the system LV and so performance will suffer.
+
+  The status return will be negated and passed back to the originating node.
+
+  pre- and post- command routines are called only on the local node. The
+  purpose is primarily to get and release locks, though the pre- routine should
+  also do any other local setups required by the command (if any) and can
+  return a failure code that prevents the command from being distributed around
+  the cluster
+
+  The pre- and post- routines are run in their own thread so can block as long
+  they like, do_command is run in the main clvmd thread so should not block for
+  too long. If the pre-command returns an error code (!=0) then the command
+  will not be propogated around the cluster but the post-command WILL be called
+
+  Also note that the pre and post routine are *always* called on the local
+  node, even if the command to be executed was only requested to run on a
+  remote node. It may peek inside the client structure to check the status of
+  the command.
+
+  The clients of the daemon must, naturally, understand the return messages and
+  codes.
+
+  Routines in here may only READ the values in the client structure passed in
+  apart from client->private which they are free to do what they like with.
+
+*/
+
+#include "clvmd-common.h"
+#include "clvmd-comms.h"
+#include "clvm.h"
+#include "clvmd.h"
+#include "lvm-globals.h"
+#include "lvm-functions.h"
+
+#include "locking.h"
+
+#include <sys/utsname.h>
+
+extern struct cluster_ops *clops;
+static int restart_clvmd(void);
+
+/* 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,
+	       char **buf, int buflen, int *retlen)
+{
+	char *args = msg->node + strlen(msg->node) + 1;
+	int arglen = msglen - sizeof(struct clvm_header) - strlen(msg->node);
+	int status = 0;
+	char *lockname;
+	const char *locktype;
+	struct utsname nodeinfo;
+	unsigned char lock_cmd;
+	unsigned char lock_flags;
+
+	/* Do the command */
+	switch (msg->cmd) {
+		/* Just a test message */
+	case CLVMD_CMD_TEST:
+		if (arglen > buflen) {
+			char *new_buf;
+			buflen = arglen + 200;
+			new_buf = realloc(*buf, buflen);
+			if (new_buf == NULL) {
+				status = errno;
+				free (*buf);
+			}
+			*buf = new_buf;
+		}
+		if (*buf) {
+			if (uname(&nodeinfo))
+				memset(&nodeinfo, 0, sizeof(nodeinfo));
+
+			*retlen = 1 + dm_snprintf(*buf, buflen,
+						  "TEST from %s: %s v%s",
+						  nodeinfo.nodename, args,
+						  nodeinfo.release);
+		}
+		break;
+
+	case CLVMD_CMD_LOCK_VG:
+		lock_cmd = args[0];
+		lock_flags = args[1];
+		lockname = &args[2];
+		/* Check to see if the VG is in use by LVM1 */
+		status = do_check_lvm1(lockname);
+		do_lock_vg(lock_cmd, lock_flags, lockname);
+		break;
+
+	case CLVMD_CMD_LOCK_LV:
+		/* This is the biggie */
+		lock_cmd = args[0];
+		lock_flags = args[1];
+		lockname = &args[2];
+		status = do_lock_lv(lock_cmd, lock_flags, lockname);
+		/* Replace EIO with something less scary */
+		if (status == EIO) {
+			*retlen = 1 + dm_snprintf(*buf, buflen, "%s",
+						  get_last_lvm_error());
+			return EIO;
+		}
+		break;
+
+	case CLVMD_CMD_LOCK_QUERY:
+		lockname = &args[2];
+		if (buflen < 3)
+			return EIO;
+		if ((locktype = do_lock_query(lockname)))
+			*retlen = 1 + dm_snprintf(*buf, buflen, "%s", locktype);
+		break;
+
+	case CLVMD_CMD_REFRESH:
+		do_refresh_cache();
+		break;
+
+	case CLVMD_CMD_SYNC_NAMES:
+		lvm_do_fs_unlock();
+		break;
+
+	case CLVMD_CMD_SET_DEBUG:
+		clvmd_set_debug((debug_t) args[0]);
+		break;
+
+	case CLVMD_CMD_RESTART:
+		status = restart_clvmd();
+		break;
+
+	case CLVMD_CMD_GET_CLUSTERNAME:
+		status = clops->get_cluster_name(*buf, buflen);
+		if (!status)
+			*retlen = strlen(*buf)+1;
+		break;
+
+	case CLVMD_CMD_VG_BACKUP:
+		/*
+		 * Do not run backup on local node, caller should do that.
+		 */
+		if (!client)
+			lvm_do_backup(&args[2]);
+		break;
+
+	default:
+		/* Won't get here because command is validated in pre_command */
+		break;
+	}
+
+	/* Check the status of the command and return the error text */
+	if (status) {
+		if (*buf)
+			*retlen = dm_snprintf(*buf, buflen, "%s", strerror(status)) + 1;
+		else
+			*retlen = 0;
+	}
+
+	return status;
+}
+
+static int lock_vg(struct local_client *client)
+{
+	struct dm_hash_table *lock_hash;
+	struct clvm_header *header =
+		(struct clvm_header *) client->bits.localsock.cmd;
+	unsigned char lock_cmd;
+	int lock_mode;
+	char *args = header->node + strlen(header->node) + 1;
+	int lkid;
+	int status;
+	char *lockname;
+
+	/*
+	 * Keep a track of VG locks in our own hash table. In current
+	 * practice there should only ever be more than two VGs locked
+	 * if a user tries to merge lots of them at once
+	 */
+	if (!client->bits.localsock.private) {
+		if (!(lock_hash = dm_hash_create(3)))
+			return ENOMEM;
+		client->bits.localsock.private = (void *) lock_hash;
+	} else
+		lock_hash = (struct dm_hash_table *) client->bits.localsock.private;
+
+	lock_cmd = args[0] & (LCK_NONBLOCK | LCK_HOLD | LCK_SCOPE_MASK | LCK_TYPE_MASK);
+	lock_mode = ((int) lock_cmd & LCK_TYPE_MASK);
+	/* lock_flags = args[1]; */
+	lockname = &args[2];
+	DEBUGLOG("(%p) doing PRE command LOCK_VG '%s' at %x\n", client, lockname, lock_cmd);
+
+	if (lock_mode == LCK_UNLOCK) {
+		if (!(lkid = (int) (long) dm_hash_lookup(lock_hash, lockname)))
+			return EINVAL;
+
+		if ((status = sync_unlock(lockname, lkid)))
+			status = errno;
+		else
+			dm_hash_remove(lock_hash, lockname);
+	} else {
+		/* Read locks need to be PR; other modes get passed through */
+		if (lock_mode == LCK_READ)
+			lock_mode = LCK_PREAD;
+
+		if ((status = sync_lock(lockname, lock_mode, (lock_cmd & LCK_NONBLOCK) ? LCKF_NOQUEUE : 0, &lkid)))
+			status = errno;
+		else if (!dm_hash_insert(lock_hash, lockname, (void *) (long) lkid))
+			return ENOMEM;
+	}
+
+	return status;
+}
+
+
+/* Pre-command is a good place to get locks that are needed only for the duration
+   of the commands around the cluster (don't forget to free them in post-command),
+   and to sanity check the command arguments */
+int do_pre_command(struct local_client *client)
+{
+	struct clvm_header *header =
+	    (struct clvm_header *) client->bits.localsock.cmd;
+	unsigned char lock_cmd;
+	unsigned char lock_flags;
+	char *args = header->node + strlen(header->node) + 1;
+	int lockid = 0;
+	int status = 0;
+	char *lockname;
+
+	switch (header->cmd) {
+	case CLVMD_CMD_TEST:
+		status = sync_lock("CLVMD_TEST", LCK_EXCL, 0, &lockid);
+		client->bits.localsock.private = (void *)(long)lockid;
+		break;
+
+	case CLVMD_CMD_LOCK_VG:
+		lockname = &args[2];
+		/* We take out a real lock unless LCK_CACHE was set */
+		if (!strncmp(lockname, "V_", 2) ||
+		    !strncmp(lockname, "P_#", 3))
+			status = lock_vg(client);
+		break;
+
+	case CLVMD_CMD_LOCK_LV:
+		lock_cmd = args[0];
+		lock_flags = args[1];
+		lockname = &args[2];
+		status = pre_lock_lv(lock_cmd, lock_flags, lockname);
+		break;
+
+	case CLVMD_CMD_REFRESH:
+	case CLVMD_CMD_GET_CLUSTERNAME:
+	case CLVMD_CMD_SET_DEBUG:
+	case CLVMD_CMD_VG_BACKUP:
+	case CLVMD_CMD_SYNC_NAMES:
+	case CLVMD_CMD_LOCK_QUERY:
+	case CLVMD_CMD_RESTART:
+		break;
+
+	default:
+		log_error("Unknown command %d received\n", header->cmd);
+		status = EINVAL;
+	}
+	return status;
+}
+
+/* Note that the post-command routine is called even if the pre-command or the real command
+   failed */
+int do_post_command(struct local_client *client)
+{
+	struct clvm_header *header =
+	    (struct clvm_header *) client->bits.localsock.cmd;
+	int status = 0;
+	unsigned char lock_cmd;
+	unsigned char lock_flags;
+	char *args = header->node + strlen(header->node) + 1;
+	char *lockname;
+
+	switch (header->cmd) {
+	case CLVMD_CMD_TEST:
+		status = sync_unlock("CLVMD_TEST", (int) (long) client->bits.localsock.private);
+		client->bits.localsock.private = NULL;
+		break;
+
+	case CLVMD_CMD_LOCK_LV:
+		lock_cmd = args[0];
+		lock_flags = args[1];
+		lockname = &args[2];
+		status = post_lock_lv(lock_cmd, lock_flags, lockname);
+		break;
+
+	default:
+		/* Nothing to do here */
+		break;
+	}
+	return status;
+}
+
+
+/* Called when the client is about to be deleted */
+void cmd_client_cleanup(struct local_client *client)
+{
+	struct dm_hash_node *v;
+	struct dm_hash_table *lock_hash;
+	int lkid;
+	char *lockname;
+
+	DEBUGLOG("(%p) Client thread cleanup\n", client);
+	if (!client->bits.localsock.private)
+		return;
+
+	lock_hash = (struct dm_hash_table *)client->bits.localsock.private;
+
+	dm_hash_iterate(v, lock_hash) {
+		lkid = (int)(long)dm_hash_get_data(lock_hash, v);
+		lockname = dm_hash_get_key(lock_hash, v);
+		DEBUGLOG("(%p) Cleanup: Unlocking lock %s %x\n", client, lockname, lkid);
+		(void) sync_unlock(lockname, lkid);
+	}
+
+	dm_hash_destroy(lock_hash);
+	client->bits.localsock.private = NULL;
+}
+
+static int restart_clvmd(void)
+{
+	const char **argv;
+	char *lv_name;
+	int argc = 0, max_locks = 0;
+	struct dm_hash_node *hn = NULL;
+	char debug_arg[16];
+	const char *clvmd = getenv("LVM_CLVMD_BINARY") ? : CLVMD_PATH;
+
+	DEBUGLOG("clvmd restart requested\n");
+
+	/* Count exclusively-open LVs */
+	do {
+		hn = get_next_excl_lock(hn, &lv_name);
+		if (lv_name) {
+			max_locks++;
+			if (!*lv_name)
+				break; /* FIXME: Is this error ? */
+		}
+	} while (hn);
+
+	/* clvmd + locks (-E uuid) + debug (-d X) + NULL */
+	if (!(argv = malloc((max_locks * 2 + 6) * sizeof(*argv))))
+		goto_out;
+
+	/*
+	 * Build the command-line
+	 */
+	argv[argc++] = "clvmd";
+
+	/* Propagate debug options */
+	if (clvmd_get_debug()) {
+		if (dm_snprintf(debug_arg, sizeof(debug_arg), "-d%u", clvmd_get_debug()) < 0)
+			goto_out;
+		argv[argc++] = debug_arg;
+	}
+
+	/* Propagate foreground options */
+	if (clvmd_get_foreground())
+		argv[argc++] = "-f";
+
+	argv[argc++] = "-I";
+	argv[argc++] = clops->name;
+
+	/* Now add the exclusively-open LVs */
+	hn = NULL;
+	do {
+		hn = get_next_excl_lock(hn, &lv_name);
+		if (lv_name) {
+			if (!*lv_name)
+				break; /* FIXME: Is this error ? */
+			argv[argc++] = "-E";
+			argv[argc++] = lv_name;
+			DEBUGLOG("excl lock: %s\n", lv_name);
+		}
+	} while (hn);
+	argv[argc] = NULL;
+
+	/* Exec new clvmd */
+	DEBUGLOG("--- Restarting %s ---\n", clvmd);
+	for (argc = 1; argv[argc]; argc++) DEBUGLOG("--- %d: %s\n", argc, argv[argc]);
+
+	/* NOTE: This will fail when downgrading! */
+	execvp(clvmd, (char **)argv);
+out:
+	/* We failed */
+	DEBUGLOG("Restart of clvmd failed.\n");
+
+	free(argv);
+
+	return EIO;
+}
--- a/daemons/clvmd/clvmd-common.h
+++ b/daemons/clvmd/clvmd-common.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2010 Red Hat, Inc. All rights reserved.
 *
 * This file is part of LVM2.
 *
@@ -12,11 +12,16 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#ifndef _LIBDM_KDEV_H
-#define _LIBDM_KDEV_H
+/*
+ * This file must be included first by every clvmd source file.
+ */
+#ifndef _LVM_CLVMD_COMMON_H
+#define _LVM_CLVMD_COMMON_H

-#define MAJOR(dev)      ((dev & 0xfff00) >> 8)
-#define MINOR(dev)      ((dev & 0xff) | ((dev >> 12) & 0xfff00))
-#define MKDEV(ma,mi)    (((dev_t)mi & 0xff) | ((dev_t)ma << 8) | (((dev_t)mi & ~0xff) << 12))
+#define _REENTRANT
+
+#include "tool.h"
+
+#include "lvm-logging.h"

 #endif
--- a/daemons/clvmd/clvmd-comms.h
+++ b/daemons/clvmd/clvmd-comms.h
@@ -0,0 +1,119 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2011 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+ * Abstraction layer for clvmd cluster communications
+ */
+
+#ifndef _CLVMD_COMMS_H
+#define _CLVMD_COMMS_H
+
+struct local_client;
+
+struct cluster_ops {
+	const char *name;
+	void (*cluster_init_completed) (void);
+
+	int (*cluster_send_message) (const void *buf, int msglen,
+				     const char *csid,
+				     const char *errtext);
+	int (*name_from_csid) (const char *csid, char *name);
+	int (*csid_from_name) (char *csid, const char *name);
+	int (*get_num_nodes) (void);
+	int (*cluster_fd_callback) (struct local_client *fd, char *buf, int len,
+				    const char *csid,
+				    struct local_client **new_client);
+	int (*get_main_cluster_fd) (void);	/* gets accept FD or cman cluster socket */
+	int (*cluster_do_node_callback) (struct local_client *client,
+					 void (*callback) (struct local_client *,
+							   const char *csid,
+							   int node_up));
+	int (*is_quorate) (void);
+
+	void (*get_our_csid) (char *csid);
+	void (*add_up_node) (const char *csid);
+	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);
+
+};
+
+#ifdef USE_CMAN
+#  include <netinet/in.h>
+#  include "libcman.h"
+#  define CMAN_MAX_CSID_LEN 4
+#  ifndef MAX_CSID_LEN
+#    define MAX_CSID_LEN CMAN_MAX_CSID_LEN
+#  endif
+#  undef MAX_CLUSTER_MEMBER_NAME_LEN
+#  define MAX_CLUSTER_MEMBER_NAME_LEN   CMAN_MAX_NODENAME_LEN
+#  define CMAN_MAX_CLUSTER_MESSAGE 1500
+#  define CLUSTER_PORT_CLVMD 11
+struct cluster_ops *init_cman_cluster(void);
+#endif
+
+#ifdef USE_OPENAIS
+#  include <openais/saAis.h>
+#  include <corosync/totem/totem.h>
+#  define OPENAIS_CSID_LEN (sizeof(int))
+#  define OPENAIS_MAX_CLUSTER_MESSAGE         MESSAGE_SIZE_MAX
+#  define OPENAIS_MAX_CLUSTER_MEMBER_NAME_LEN SA_MAX_NAME_LENGTH
+#  ifndef MAX_CLUSTER_MEMBER_NAME_LEN
+#    define MAX_CLUSTER_MEMBER_NAME_LEN       SA_MAX_NAME_LENGTH
+#  endif
+#  ifndef CMAN_MAX_CLUSTER_MESSAGE
+#    define CMAN_MAX_CLUSTER_MESSAGE          MESSAGE_SIZE_MAX
+#  endif
+#  ifndef MAX_CSID_LEN
+#    define MAX_CSID_LEN sizeof(int)
+#  endif
+struct cluster_ops *init_openais_cluster(void);
+#endif
+
+#ifdef USE_COROSYNC
+#  include <corosync/corotypes.h>
+#  define COROSYNC_CSID_LEN (sizeof(int))
+#  define COROSYNC_MAX_CLUSTER_MESSAGE         65535
+#  define COROSYNC_MAX_CLUSTER_MEMBER_NAME_LEN CS_MAX_NAME_LENGTH
+#  ifndef MAX_CLUSTER_MEMBER_NAME_LEN
+#    define MAX_CLUSTER_MEMBER_NAME_LEN       CS_MAX_NAME_LENGTH
+#  endif
+#  ifndef CMAN_MAX_CLUSTER_MESSAGE
+#    define CMAN_MAX_CLUSTER_MESSAGE          65535
+#  endif
+#  ifndef MAX_CSID_LEN
+#    define MAX_CSID_LEN sizeof(int)
+#  endif
+struct cluster_ops *init_corosync_cluster(void);
+#endif
+
+#ifdef USE_SINGLENODE
+#  define SINGLENODE_CSID_LEN (sizeof(int))
+#  ifndef MAX_CLUSTER_MEMBER_NAME_LEN
+#    define MAX_CLUSTER_MEMBER_NAME_LEN       64
+#  endif
+#  define SINGLENODE_MAX_CLUSTER_MESSAGE          65535
+#  ifndef MAX_CSID_LEN
+#    define MAX_CSID_LEN sizeof(int)
+#  endif
+struct cluster_ops *init_singlenode_cluster(void);
+#endif
+
+#endif
--- a/daemons/clvmd/clvmd-corosync.c
+++ b/daemons/clvmd/clvmd-corosync.c
@@ -0,0 +1,662 @@
+/*
+ * Copyright (C) 2009-2012 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 Lesser General Public License v.2.1.
+ *
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+ * This provides the interface between clvmd and corosync/DLM as the cluster
+ * and lock manager.
+ */
+
+#include "clvmd-common.h"
+
+#include <pthread.h>
+
+#include "clvm.h"
+#include "clvmd-comms.h"
+#include "clvmd.h"
+#include "lvm-functions.h"
+
+#include "locking.h"
+
+#include <corosync/cpg.h>
+#include <corosync/quorum.h>
+
+#ifdef HAVE_COROSYNC_CONFDB_H
+#  include <corosync/confdb.h>
+#elif defined HAVE_COROSYNC_CMAP_H
+#  include <corosync/cmap.h>
+#else
+#  error "Either HAVE_COROSYNC_CONFDB_H or HAVE_COROSYNC_CMAP_H must be defined."
+#endif
+
+#include <libdlm.h>
+
+#include <syslog.h>
+
+/* Timeout value for several corosync calls */
+#define LOCKSPACE_NAME "clvmd"
+
+static void corosync_cpg_deliver_callback (cpg_handle_t handle,
+				  const struct cpg_name *groupName,
+				  uint32_t nodeid,
+				  uint32_t pid,
+				  void *msg,
+				  size_t msg_len);
+static void corosync_cpg_confchg_callback(cpg_handle_t handle,
+				 const struct cpg_name *groupName,
+				 const struct cpg_address *member_list, size_t member_list_entries,
+				 const struct cpg_address *left_list, size_t left_list_entries,
+				 const struct cpg_address *joined_list, size_t joined_list_entries);
+static void _cluster_closedown(void);
+
+/* Hash list of nodes in the cluster */
+static struct dm_hash_table *node_hash;
+
+/* Number of active nodes */
+static int num_nodes;
+static unsigned int our_nodeid;
+
+static struct local_client *cluster_client;
+
+/* Corosync handles */
+static cpg_handle_t cpg_handle;
+static quorum_handle_t quorum_handle;
+
+/* DLM Handle */
+static dlm_lshandle_t *lockspace;
+
+static struct cpg_name cpg_group_name;
+
+/* Corosync callback structs */
+cpg_callbacks_t corosync_cpg_callbacks = {
+	.cpg_deliver_fn =            corosync_cpg_deliver_callback,
+	.cpg_confchg_fn =            corosync_cpg_confchg_callback,
+};
+
+quorum_callbacks_t quorum_callbacks = {
+	.quorum_notify_fn = NULL,
+};
+
+struct node_info
+{
+	enum {NODE_DOWN, NODE_CLVMD} state;
+	int nodeid;
+};
+
+
+/* Set errno to something approximating the right value and return 0 or -1 */
+static int cs_to_errno(cs_error_t err)
+{
+	switch(err)
+	{
+	case CS_OK:
+		return 0;
+        case CS_ERR_LIBRARY:
+		errno = EINVAL;
+		break;
+        case CS_ERR_VERSION:
+		errno = EINVAL;
+		break;
+        case CS_ERR_INIT:
+		errno = EINVAL;
+		break;
+        case CS_ERR_TIMEOUT:
+		errno = ETIME;
+		break;
+        case CS_ERR_TRY_AGAIN:
+		errno = EAGAIN;
+		break;
+        case CS_ERR_INVALID_PARAM:
+		errno = EINVAL;
+		break;
+        case CS_ERR_NO_MEMORY:
+		errno = ENOMEM;
+		break;
+        case CS_ERR_BAD_HANDLE:
+		errno = EINVAL;
+		break;
+        case CS_ERR_BUSY:
+		errno = EBUSY;
+		break;
+        case CS_ERR_ACCESS:
+		errno = EPERM;
+		break;
+        case CS_ERR_NOT_EXIST:
+		errno = ENOENT;
+		break;
+        case CS_ERR_NAME_TOO_LONG:
+		errno = ENAMETOOLONG;
+		break;
+        case CS_ERR_EXIST:
+		errno = EEXIST;
+		break;
+        case CS_ERR_NO_SPACE:
+		errno = ENOSPC;
+		break;
+        case CS_ERR_INTERRUPT:
+		errno = EINTR;
+		break;
+	case CS_ERR_NAME_NOT_FOUND:
+		errno = ENOENT;
+		break;
+        case CS_ERR_NO_RESOURCES:
+		errno = ENOMEM;
+		break;
+        case CS_ERR_NOT_SUPPORTED:
+		errno = EOPNOTSUPP;
+		break;
+        case CS_ERR_BAD_OPERATION:
+		errno = EINVAL;
+		break;
+        case CS_ERR_FAILED_OPERATION:
+		errno = EIO;
+		break;
+        case CS_ERR_MESSAGE_ERROR:
+		errno = EIO;
+		break;
+        case CS_ERR_QUEUE_FULL:
+		errno = EXFULL;
+		break;
+        case CS_ERR_QUEUE_NOT_AVAILABLE:
+		errno = EINVAL;
+		break;
+        case CS_ERR_BAD_FLAGS:
+		errno = EINVAL;
+		break;
+        case CS_ERR_TOO_BIG:
+		errno = E2BIG;
+		break;
+        case CS_ERR_NO_SECTIONS:
+		errno = ENOMEM;
+		break;
+	default:
+		errno = EINVAL;
+		break;
+	}
+	return -1;
+}
+
+static char *print_corosync_csid(const char *csid)
+{
+	static char buf[128];
+	int id;
+
+	memcpy(&id, csid, sizeof(int));
+	sprintf(buf, "%d", id);
+	return buf;
+}
+
+static void corosync_cpg_deliver_callback (cpg_handle_t handle,
+				  const struct cpg_name *groupName,
+				  uint32_t nodeid,
+				  uint32_t pid,
+				  void *msg,
+				  size_t msg_len)
+{
+	int target_nodeid;
+
+	memcpy(&target_nodeid, msg, COROSYNC_CSID_LEN);
+
+	DEBUGLOG("%u got message from nodeid %d for %d. len %zd\n",
+		 our_nodeid, nodeid, target_nodeid, msg_len-4);
+
+	if (nodeid != our_nodeid)
+		if (target_nodeid == our_nodeid || target_nodeid == 0)
+			process_message(cluster_client, (char *)msg+COROSYNC_CSID_LEN,
+					msg_len-COROSYNC_CSID_LEN, (char*)&nodeid);
+}
+
+static void corosync_cpg_confchg_callback(cpg_handle_t handle,
+				 const struct cpg_name *groupName,
+				 const struct cpg_address *member_list, size_t member_list_entries,
+				 const struct cpg_address *left_list, size_t left_list_entries,
+				 const struct cpg_address *joined_list, size_t joined_list_entries)
+{
+	int i;
+	struct node_info *ninfo;
+
+	DEBUGLOG("confchg callback. %zd joined, %zd left, %zd members\n",
+		 joined_list_entries, left_list_entries, member_list_entries);
+
+	for (i=0; i<joined_list_entries; i++) {
+		ninfo = dm_hash_lookup_binary(node_hash,
+					      (char *)&joined_list[i].nodeid,
+					      COROSYNC_CSID_LEN);
+		if (!ninfo) {
+			ninfo = malloc(sizeof(struct node_info));
+			if (!ninfo) {
+				break;
+			}
+			else {
+				ninfo->nodeid = joined_list[i].nodeid;
+				dm_hash_insert_binary(node_hash,
+						      (char *)&ninfo->nodeid,
+						      COROSYNC_CSID_LEN, ninfo);
+			}
+		}
+		ninfo->state = NODE_CLVMD;
+	}
+
+	for (i=0; i<left_list_entries; i++) {
+		ninfo = dm_hash_lookup_binary(node_hash,
+					      (char *)&left_list[i].nodeid,
+					      COROSYNC_CSID_LEN);
+		if (ninfo)
+			ninfo->state = NODE_DOWN;
+	}
+
+	num_nodes = member_list_entries;
+}
+
+static int _init_cluster(void)
+{
+	cs_error_t err;
+
+#ifdef QUORUM_SET	/* corosync/quorum.h */
+	uint32_t quorum_type;
+#endif
+
+	node_hash = dm_hash_create(100);
+
+	err = cpg_initialize(&cpg_handle,
+			     &corosync_cpg_callbacks);
+	if (err != CS_OK) {
+		syslog(LOG_ERR, "Cannot initialise Corosync CPG service: %d",
+		       err);
+		DEBUGLOG("Cannot initialise Corosync CPG service: %d", err);
+		return cs_to_errno(err);
+	}
+
+#ifdef QUORUM_SET
+	err = quorum_initialize(&quorum_handle,
+				&quorum_callbacks,
+				&quorum_type);
+
+	if (quorum_type != QUORUM_SET) {
+		syslog(LOG_ERR, "Corosync quorum service is not configured");
+		DEBUGLOG("Corosync quorum service is not configured");
+		return EINVAL;
+	}
+#else
+	err = quorum_initialize(&quorum_handle,
+				&quorum_callbacks);
+#endif
+
+	if (err != CS_OK) {
+		syslog(LOG_ERR, "Cannot initialise Corosync quorum service: %d",
+		       err);
+		DEBUGLOG("Cannot initialise Corosync quorum service: %d", err);
+		return cs_to_errno(err);
+	}
+
+	/* Create a lockspace for LV & VG locks to live in */
+	lockspace = dlm_open_lockspace(LOCKSPACE_NAME);
+	if (!lockspace) {
+		lockspace = dlm_create_lockspace(LOCKSPACE_NAME, 0600);
+		if (!lockspace) {
+			syslog(LOG_ERR, "Unable to create DLM lockspace for CLVM: %m");
+			return -1;
+		}
+		DEBUGLOG("Created DLM lockspace for CLVMD.\n");
+	} else
+		DEBUGLOG("Opened existing DLM lockspace for CLVMD.\n");
+
+	dlm_ls_pthread_init(lockspace);
+	DEBUGLOG("DLM initialisation complete\n");
+
+	/* Connect to the clvmd group */
+	strcpy((char *)cpg_group_name.value, "clvmd");
+	cpg_group_name.length = strlen((char *)cpg_group_name.value);
+	err = cpg_join(cpg_handle, &cpg_group_name);
+	if (err != CS_OK) {
+		cpg_finalize(cpg_handle);
+		quorum_finalize(quorum_handle);
+		dlm_release_lockspace(LOCKSPACE_NAME, lockspace, 1);
+		syslog(LOG_ERR, "Cannot join clvmd process group");
+		DEBUGLOG("Cannot join clvmd process group: %d\n", err);
+		return cs_to_errno(err);
+	}
+
+	err = cpg_local_get(cpg_handle,
+			    &our_nodeid);
+	if (err != CS_OK) {
+		cpg_finalize(cpg_handle);
+		quorum_finalize(quorum_handle);
+		dlm_release_lockspace(LOCKSPACE_NAME, lockspace, 1);
+		syslog(LOG_ERR, "Cannot get local node id\n");
+		return cs_to_errno(err);
+	}
+	DEBUGLOG("Our local node id is %d\n", our_nodeid);
+
+	DEBUGLOG("Connected to Corosync\n");
+
+	return 0;
+}
+
+static void _cluster_closedown(void)
+{
+	dlm_release_lockspace(LOCKSPACE_NAME, lockspace, 1);
+	cpg_finalize(cpg_handle);
+	quorum_finalize(quorum_handle);
+}
+
+static void _get_our_csid(char *csid)
+{
+	memcpy(csid, &our_nodeid, sizeof(int));
+}
+
+/* Corosync doesn't really have nmode names so we
+   just use the node ID in hex instead */
+static int _csid_from_name(char *csid, const char *name)
+{
+	int nodeid;
+	struct node_info *ninfo;
+
+	if (sscanf(name, "%x", &nodeid) == 1) {
+		ninfo = dm_hash_lookup_binary(node_hash, csid, COROSYNC_CSID_LEN);
+		if (ninfo)
+			return nodeid;
+	}
+	return -1;
+}
+
+static int _name_from_csid(const char *csid, char *name)
+{
+	struct node_info *ninfo;
+
+	ninfo = dm_hash_lookup_binary(node_hash, csid, COROSYNC_CSID_LEN);
+	if (!ninfo)
+	{
+		sprintf(name, "UNKNOWN %s", print_corosync_csid(csid));
+		return -1;
+	}
+
+	sprintf(name, "%x", ninfo->nodeid);
+	return 0;
+}
+
+static int _get_num_nodes(void)
+{
+	DEBUGLOG("num_nodes = %d\n", num_nodes);
+	return num_nodes;
+}
+
+/* Node is now known to be running a clvmd */
+static void _add_up_node(const char *csid)
+{
+	struct node_info *ninfo;
+
+	ninfo = dm_hash_lookup_binary(node_hash, csid, COROSYNC_CSID_LEN);
+	if (!ninfo) {
+		DEBUGLOG("corosync_add_up_node no node_hash entry for csid %s\n",
+			 print_corosync_csid(csid));
+		return;
+	}
+
+	DEBUGLOG("corosync_add_up_node %d\n", ninfo->nodeid);
+
+	ninfo->state = NODE_CLVMD;
+
+	return;
+}
+
+/* Call a callback for each node, so the caller knows whether it's up or down */
+static int _cluster_do_node_callback(struct local_client *master_client,
+				     void (*callback)(struct local_client *,
+						      const char *csid, int node_up))
+{
+	struct dm_hash_node *hn;
+	struct node_info *ninfo;
+
+	dm_hash_iterate(hn, node_hash)
+	{
+		char csid[COROSYNC_CSID_LEN];
+
+		ninfo = dm_hash_get_data(node_hash, hn);
+		memcpy(csid, dm_hash_get_key(node_hash, hn), COROSYNC_CSID_LEN);
+
+		DEBUGLOG("down_callback. node %d, state = %d\n", ninfo->nodeid,
+			 ninfo->state);
+
+		if (ninfo->state == NODE_CLVMD)
+			callback(master_client, csid, 1);
+	}
+	return 0;
+}
+
+/* Real locking */
+static int _lock_resource(const char *resource, int mode, int flags, int *lockid)
+{
+	struct dlm_lksb lksb;
+	int err;
+
+	DEBUGLOG("lock_resource '%s', flags=%d, mode=%d\n", resource, flags, mode);
+
+	if (flags & LKF_CONVERT)
+		lksb.sb_lkid = *lockid;
+
+	err = dlm_ls_lock_wait(lockspace,
+			       mode,
+			       &lksb,
+			       flags,
+			       resource,
+			       strlen(resource),
+			       0,
+			       NULL, NULL, NULL);
+
+	if (err != 0)
+	{
+		DEBUGLOG("dlm_ls_lock returned %d\n", errno);
+		return err;
+	}
+	if (lksb.sb_status != 0)
+	{
+		DEBUGLOG("dlm_ls_lock returns lksb.sb_status %d\n", lksb.sb_status);
+		errno = lksb.sb_status;
+		return -1;
+	}
+
+	DEBUGLOG("lock_resource returning %d, lock_id=%x\n", err, lksb.sb_lkid);
+
+	*lockid = lksb.sb_lkid;
+
+	return 0;
+}
+
+
+static int _unlock_resource(const char *resource, int lockid)
+{
+	struct dlm_lksb lksb;
+	int err;
+
+	DEBUGLOG("unlock_resource: %s lockid: %x\n", resource, lockid);
+	lksb.sb_lkid = lockid;
+
+	err = dlm_ls_unlock_wait(lockspace,
+				 lockid,
+				 0,
+				 &lksb);
+	if (err != 0)
+	{
+		DEBUGLOG("Unlock returned %d\n", err);
+		return err;
+	}
+	if (lksb.sb_status != EUNLOCK)
+	{
+		DEBUGLOG("dlm_ls_unlock_wait returns lksb.sb_status: %d\n", lksb.sb_status);
+		errno = lksb.sb_status;
+		return -1;
+	}   
+
+
+	return 0;
+}
+
+static int _is_quorate(void)
+{
+	int quorate;
+	if (quorum_getquorate(quorum_handle, &quorate) == CS_OK)
+		return quorate;
+	else
+		return 0;
+}
+
+static int _get_main_cluster_fd(void)
+{
+	int select_fd;
+
+	cpg_fd_get(cpg_handle, &select_fd);
+	return select_fd;
+}
+
+static int _cluster_fd_callback(struct local_client *fd, char *buf, int len,
+				const char *csid,
+				struct local_client **new_client)
+{
+	cluster_client = fd;
+	*new_client = NULL;
+	cpg_dispatch(cpg_handle, CS_DISPATCH_ONE);
+	return 1;
+}
+
+static int _cluster_send_message(const void *buf, int msglen, const char *csid,
+				 const char *errtext)
+{
+	static pthread_mutex_t _mutex = PTHREAD_MUTEX_INITIALIZER;
+	struct iovec iov[2];
+	cs_error_t err;
+	int target_node;
+
+	if (csid)
+		memcpy(&target_node, csid, COROSYNC_CSID_LEN);
+	else
+		target_node = 0;
+
+	iov[0].iov_base = &target_node;
+	iov[0].iov_len = sizeof(int);
+	iov[1].iov_base = (char *)buf;
+	iov[1].iov_len = msglen;
+
+	pthread_mutex_lock(&_mutex);
+	err = cpg_mcast_joined(cpg_handle, CPG_TYPE_AGREED, iov, 2);
+	pthread_mutex_unlock(&_mutex);
+
+	return cs_to_errno(err);
+}
+
+#ifdef HAVE_COROSYNC_CONFDB_H
+/*
+ * We are not necessarily connected to a Red Hat Cluster system,
+ * but if we are, this returns the cluster name from cluster.conf.
+ * I've used confdb rather than ccs to reduce the inter-package
+ * dependancies as well as to allow people to set a cluster name
+ * for themselves even if they are not running on RH cluster.
+ */
+static int _get_cluster_name(char *buf, int buflen)
+{
+	confdb_handle_t handle;
+	int result;
+	size_t namelen = buflen;
+	hdb_handle_t cluster_handle;
+	confdb_callbacks_t callbacks = {
+		.confdb_key_change_notify_fn = NULL,
+		.confdb_object_create_change_notify_fn = NULL,
+		.confdb_object_delete_change_notify_fn = NULL
+	};
+
+	/* This is a default in case everything else fails */
+	strncpy(buf, "Corosync", buflen);
+
+	/* Look for a cluster name in confdb */
+	result = confdb_initialize (&handle, &callbacks);
+        if (result != CS_OK)
+		return 0;
+
+        result = confdb_object_find_start(handle, OBJECT_PARENT_HANDLE);
+	if (result != CS_OK)
+		goto out;
+
+        result = confdb_object_find(handle, OBJECT_PARENT_HANDLE, (void *)"cluster", strlen("cluster"), &cluster_handle);
+        if (result != CS_OK)
+		goto out;
+
+        result = confdb_key_get(handle, cluster_handle, (void *)"name", strlen("name"), buf, &namelen);
+        if (result != CS_OK)
+		goto out;
+
+	buf[namelen] = '\0';
+
+out:
+	confdb_finalize(handle);
+	return 0;
+}
+
+#elif defined HAVE_COROSYNC_CMAP_H
+
+static int _get_cluster_name(char *buf, int buflen)
+{
+	cmap_handle_t cmap_handle = 0;
+	int result;
+	char *name = NULL;
+
+	/* This is a default in case everything else fails */
+	strncpy(buf, "Corosync", buflen);
+
+	/* Look for a cluster name in cmap */
+	result = cmap_initialize(&cmap_handle);
+	if (result != CS_OK)
+		return 0;
+
+	result = cmap_get_string(cmap_handle, "totem.cluster_name", &name);
+	if (result != CS_OK)
+		goto out;
+
+	memset(buf, 0, buflen);
+	strncpy(buf, name, buflen - 1);
+
+out:
+	if (name)
+		free(name);
+	cmap_finalize(cmap_handle);
+	return 0;
+}
+
+#endif
+
+static struct cluster_ops _cluster_corosync_ops = {
+	.name                     = "corosync",
+	.cluster_init_completed   = NULL,
+	.cluster_send_message     = _cluster_send_message,
+	.name_from_csid           = _name_from_csid,
+	.csid_from_name           = _csid_from_name,
+	.get_num_nodes            = _get_num_nodes,
+	.cluster_fd_callback      = _cluster_fd_callback,
+	.get_main_cluster_fd      = _get_main_cluster_fd,
+	.cluster_do_node_callback = _cluster_do_node_callback,
+	.is_quorate               = _is_quorate,
+	.get_our_csid             = _get_our_csid,
+	.add_up_node              = _add_up_node,
+	.reread_config            = NULL,
+	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
+	.sync_lock                = _lock_resource,
+	.sync_unlock              = _unlock_resource,
+};
+
+struct cluster_ops *init_corosync_cluster(void)
+{
+	if (!_init_cluster())
+		return &_cluster_corosync_ops;
+	else
+		return NULL;
+}
--- a/daemons/clvmd/clvmd-openais.c
+++ b/daemons/clvmd/clvmd-openais.c
@@ -0,0 +1,687 @@
+/*
+ * Copyright (C) 2007-2009 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 Lesser General Public License v.2.1.
+ *
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/*
+ * This provides the interface between clvmd and OpenAIS as the cluster
+ * and lock manager.
+ */
+
+#include "clvmd-common.h"
+
+#include <pthread.h>
+#include <fcntl.h>
+#include <syslog.h>
+
+#include <openais/saAis.h>
+#include <openais/saLck.h>
+
+#include <corosync/corotypes.h>
+#include <corosync/cpg.h>
+
+#include "locking.h"
+#include "clvm.h"
+#include "clvmd-comms.h"
+#include "lvm-functions.h"
+#include "clvmd.h"
+
+/* Timeout value for several openais calls */
+#define TIMEOUT 10
+
+static void openais_cpg_deliver_callback (cpg_handle_t handle,
+				  const struct cpg_name *groupName,
+				  uint32_t nodeid,
+				  uint32_t pid,
+				  void *msg,
+				  size_t msg_len);
+static void openais_cpg_confchg_callback(cpg_handle_t handle,
+				 const struct cpg_name *groupName,
+				 const struct cpg_address *member_list, size_t member_list_entries,
+				 const struct cpg_address *left_list, size_t left_list_entries,
+				 const struct cpg_address *joined_list, size_t joined_list_entries);
+
+static void _cluster_closedown(void);
+
+/* Hash list of nodes in the cluster */
+static struct dm_hash_table *node_hash;
+
+/* For associating lock IDs & resource handles */
+static struct dm_hash_table *lock_hash;
+
+/* Number of active nodes */
+static int num_nodes;
+static unsigned int our_nodeid;
+
+static struct local_client *cluster_client;
+
+/* OpenAIS handles */
+static cpg_handle_t cpg_handle;
+static SaLckHandleT lck_handle;
+
+static struct cpg_name cpg_group_name;
+
+/* Openais callback structs */
+cpg_callbacks_t openais_cpg_callbacks = {
+	.cpg_deliver_fn =            openais_cpg_deliver_callback,
+	.cpg_confchg_fn =            openais_cpg_confchg_callback,
+};
+
+struct node_info
+{
+	enum {NODE_UNKNOWN, NODE_DOWN, NODE_UP, NODE_CLVMD} state;
+	int nodeid;
+};
+
+struct lock_info
+{
+	SaLckResourceHandleT res_handle;
+	SaLckLockIdT         lock_id;
+	SaNameT              lock_name;
+};
+
+/* Set errno to something approximating the right value and return 0 or -1 */
+static int ais_to_errno(SaAisErrorT err)
+{
+	switch(err)
+	{
+	case SA_AIS_OK:
+		return 0;
+        case SA_AIS_ERR_LIBRARY:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_VERSION:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_INIT:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_TIMEOUT:
+		errno = ETIME;
+		break;
+        case SA_AIS_ERR_TRY_AGAIN:
+		errno = EAGAIN;
+		break;
+        case SA_AIS_ERR_INVALID_PARAM:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_NO_MEMORY:
+		errno = ENOMEM;
+		break;
+        case SA_AIS_ERR_BAD_HANDLE:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_BUSY:
+		errno = EBUSY;
+		break;
+        case SA_AIS_ERR_ACCESS:
+		errno = EPERM;
+		break;
+        case SA_AIS_ERR_NOT_EXIST:
+		errno = ENOENT;
+		break;
+        case SA_AIS_ERR_NAME_TOO_LONG:
+		errno = ENAMETOOLONG;
+		break;
+        case SA_AIS_ERR_EXIST:
+		errno = EEXIST;
+		break;
+        case SA_AIS_ERR_NO_SPACE:
+		errno = ENOSPC;
+		break;
+        case SA_AIS_ERR_INTERRUPT:
+		errno = EINTR;
+		break;
+	case SA_AIS_ERR_NAME_NOT_FOUND:
+		errno = ENOENT;
+		break;
+        case SA_AIS_ERR_NO_RESOURCES:
+		errno = ENOMEM;
+		break;
+        case SA_AIS_ERR_NOT_SUPPORTED:
+		errno = EOPNOTSUPP;
+		break;
+        case SA_AIS_ERR_BAD_OPERATION:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_FAILED_OPERATION:
+		errno = EIO;
+		break;
+        case SA_AIS_ERR_MESSAGE_ERROR:
+		errno = EIO;
+		break;
+        case SA_AIS_ERR_QUEUE_FULL:
+		errno = EXFULL;
+		break;
+        case SA_AIS_ERR_QUEUE_NOT_AVAILABLE:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_BAD_FLAGS:
+		errno = EINVAL;
+		break;
+        case SA_AIS_ERR_TOO_BIG:
+		errno = E2BIG;
+		break;
+        case SA_AIS_ERR_NO_SECTIONS:
+		errno = ENOMEM;
+		break;
+	default:
+		errno = EINVAL;
+		break;
+	}
+	return -1;
+}
+
+static char *print_openais_csid(const char *csid)
+{
+	static char buf[128];
+	int id;
+
+	memcpy(&id, csid, sizeof(int));
+	sprintf(buf, "%d", id);
+	return buf;
+}
+
+static int add_internal_client(int fd, fd_callback_t callback)
+{
+	struct local_client *client;
+
+	DEBUGLOG("Add_internal_client, fd = %d\n", fd);
+
+	if (!(client = dm_zalloc(sizeof(*client)))) {
+		DEBUGLOG("malloc failed\n");
+		return -1;
+	}
+
+	client->fd = fd;
+	client->type = CLUSTER_INTERNAL;
+	client->callback = callback;
+	add_client(client);
+
+	/* Set Close-on-exec */
+	fcntl(fd, F_SETFD, 1);
+
+	return 0;
+}
+
+static void openais_cpg_deliver_callback (cpg_handle_t handle,
+				  const struct cpg_name *groupName,
+				  uint32_t nodeid,
+				  uint32_t pid,
+				  void *msg,
+				  size_t msg_len)
+{
+	int target_nodeid;
+
+	memcpy(&target_nodeid, msg, OPENAIS_CSID_LEN);
+
+	DEBUGLOG("%u got message from nodeid %d for %d. len %" PRIsize_t "\n",
+		 our_nodeid, nodeid, target_nodeid, msg_len-4);
+
+	if (nodeid != our_nodeid)
+		if (target_nodeid == our_nodeid || target_nodeid == 0)
+			process_message(cluster_client, (char *)msg+OPENAIS_CSID_LEN,
+					msg_len-OPENAIS_CSID_LEN, (char*)&nodeid);
+}
+
+static void openais_cpg_confchg_callback(cpg_handle_t handle,
+				 const struct cpg_name *groupName,
+				 const struct cpg_address *member_list, size_t member_list_entries,
+				 const struct cpg_address *left_list, size_t left_list_entries,
+				 const struct cpg_address *joined_list, size_t joined_list_entries)
+{
+	int i;
+	struct node_info *ninfo;
+
+	DEBUGLOG("confchg callback. %" PRIsize_t " joined, "
+		 FMTsize_t " left, %" PRIsize_t " members\n",
+		 joined_list_entries, left_list_entries, member_list_entries);
+
+	for (i=0; i<joined_list_entries; i++) {
+		ninfo = dm_hash_lookup_binary(node_hash,
+					      (char *)&joined_list[i].nodeid,
+					      OPENAIS_CSID_LEN);
+		if (!ninfo) {
+			ninfo = malloc(sizeof(struct node_info));
+			if (!ninfo) {
+				break;
+			}
+			else {
+				ninfo->nodeid = joined_list[i].nodeid;
+				dm_hash_insert_binary(node_hash,
+						      (char *)&ninfo->nodeid,
+						      OPENAIS_CSID_LEN, ninfo);
+			}
+		}
+		ninfo->state = NODE_CLVMD;
+	}
+
+	for (i=0; i<left_list_entries; i++) {
+		ninfo = dm_hash_lookup_binary(node_hash,
+					      (char *)&left_list[i].nodeid,
+					      OPENAIS_CSID_LEN);
+		if (ninfo)
+			ninfo->state = NODE_DOWN;
+	}
+
+	for (i=0; i<member_list_entries; i++) {
+		if (member_list[i].nodeid == 0) continue;
+		ninfo = dm_hash_lookup_binary(node_hash,
+				(char *)&member_list[i].nodeid,
+				OPENAIS_CSID_LEN);
+		if (!ninfo) {
+			ninfo = malloc(sizeof(struct node_info));
+			if (!ninfo) {
+				break;
+			}
+			else {
+				ninfo->nodeid = member_list[i].nodeid;
+				dm_hash_insert_binary(node_hash,
+						(char *)&ninfo->nodeid,
+						OPENAIS_CSID_LEN, ninfo);
+			}
+		}
+		ninfo->state = NODE_CLVMD;
+	}
+
+	num_nodes = member_list_entries;
+}
+
+static int lck_dispatch(struct local_client *client, char *buf, int len,
+			const char *csid, struct local_client **new_client)
+{
+	*new_client = NULL;
+	saLckDispatch(lck_handle, SA_DISPATCH_ONE);
+	return 1;
+}
+
+static int _init_cluster(void)
+{
+	SaAisErrorT err;
+	SaVersionT  ver = { 'B', 1, 1 };
+	int select_fd;
+
+	node_hash = dm_hash_create(100);
+	lock_hash = dm_hash_create(10);
+
+	err = cpg_initialize(&cpg_handle,
+			     &openais_cpg_callbacks);
+	if (err != SA_AIS_OK) {
+		syslog(LOG_ERR, "Cannot initialise OpenAIS CPG service: %d",
+		       err);
+		DEBUGLOG("Cannot initialise OpenAIS CPG service: %d", err);
+		return ais_to_errno(err);
+	}
+
+	err = saLckInitialize(&lck_handle,
+					NULL,
+			      &ver);
+	if (err != SA_AIS_OK) {
+		cpg_initialize(&cpg_handle, &openais_cpg_callbacks);
+		syslog(LOG_ERR, "Cannot initialise OpenAIS lock service: %d",
+		       err);
+		DEBUGLOG("Cannot initialise OpenAIS lock service: %d\n\n", err);
+		return ais_to_errno(err);
+	}
+
+	/* Connect to the clvmd group */
+	strcpy((char *)cpg_group_name.value, "clvmd");
+	cpg_group_name.length = strlen((char *)cpg_group_name.value);
+	err = cpg_join(cpg_handle, &cpg_group_name);
+	if (err != SA_AIS_OK) {
+		cpg_finalize(cpg_handle);
+		saLckFinalize(lck_handle);
+		syslog(LOG_ERR, "Cannot join clvmd process group");
+		DEBUGLOG("Cannot join clvmd process group: %d\n", err);
+		return ais_to_errno(err);
+	}
+
+	err = cpg_local_get(cpg_handle,
+			    &our_nodeid);
+	if (err != SA_AIS_OK) {
+		cpg_finalize(cpg_handle);
+		saLckFinalize(lck_handle);
+		syslog(LOG_ERR, "Cannot get local node id\n");
+		return ais_to_errno(err);
+	}
+	DEBUGLOG("Our local node id is %d\n", our_nodeid);
+
+	saLckSelectionObjectGet(lck_handle, (SaSelectionObjectT *)&select_fd);
+	add_internal_client(select_fd, lck_dispatch);
+
+	DEBUGLOG("Connected to OpenAIS\n");
+
+	return 0;
+}
+
+static void _cluster_closedown(void)
+{
+	saLckFinalize(lck_handle);
+	cpg_finalize(cpg_handle);
+}
+
+static void _get_our_csid(char *csid)
+{
+	memcpy(csid, &our_nodeid, sizeof(int));
+}
+
+/* OpenAIS doesn't really have nmode names so we
+   just use the node ID in hex instead */
+static int _csid_from_name(char *csid, const char *name)
+{
+	int nodeid;
+	struct node_info *ninfo;
+
+	if (sscanf(name, "%x", &nodeid) == 1) {
+		ninfo = dm_hash_lookup_binary(node_hash, csid, OPENAIS_CSID_LEN);
+		if (ninfo)
+			return nodeid;
+	}
+	return -1;
+}
+
+static int _name_from_csid(const char *csid, char *name)
+{
+	struct node_info *ninfo;
+
+	ninfo = dm_hash_lookup_binary(node_hash, csid, OPENAIS_CSID_LEN);
+	if (!ninfo)
+	{
+		sprintf(name, "UNKNOWN %s", print_openais_csid(csid));
+		return -1;
+	}
+
+	sprintf(name, "%x", ninfo->nodeid);
+	return 0;
+}
+
+static int _get_num_nodes()
+{
+	DEBUGLOG("num_nodes = %d\n", num_nodes);
+	return num_nodes;
+}
+
+/* Node is now known to be running a clvmd */
+static void _add_up_node(const char *csid)
+{
+	struct node_info *ninfo;
+
+	ninfo = dm_hash_lookup_binary(node_hash, csid, OPENAIS_CSID_LEN);
+	if (!ninfo) {
+		DEBUGLOG("openais_add_up_node no node_hash entry for csid %s\n",
+			 print_openais_csid(csid));
+		return;
+	}
+
+	DEBUGLOG("openais_add_up_node %d\n", ninfo->nodeid);
+
+	ninfo->state = NODE_CLVMD;
+}
+
+/* Call a callback for each node, so the caller knows whether it's up or down */
+static int _cluster_do_node_callback(struct local_client *master_client,
+				     void (*callback)(struct local_client *,
+						      const char *csid, int node_up))
+{
+	struct dm_hash_node *hn;
+	struct node_info *ninfo;
+	int somedown = 0;
+
+	dm_hash_iterate(hn, node_hash)
+	{
+		char csid[OPENAIS_CSID_LEN];
+
+		ninfo = dm_hash_get_data(node_hash, hn);
+		memcpy(csid, dm_hash_get_key(node_hash, hn), OPENAIS_CSID_LEN);
+
+		DEBUGLOG("down_callback. node %d, state = %d\n", ninfo->nodeid,
+			 ninfo->state);
+
+		if (ninfo->state != NODE_DOWN)
+			callback(master_client, csid, ninfo->state == NODE_CLVMD);
+		if (ninfo->state != NODE_CLVMD)
+			somedown = -1;
+	}
+	return somedown;
+}
+
+/* Real locking */
+static int _lock_resource(char *resource, int mode, int flags, int *lockid)
+{
+	struct lock_info *linfo;
+	SaLckResourceHandleT res_handle;
+	SaAisErrorT err;
+	SaLckLockIdT lock_id;
+	SaLckLockStatusT lockStatus;
+
+	/* This needs to be converted from DLM/LVM2 value for OpenAIS LCK */
+	if (flags & LCK_NONBLOCK) flags = SA_LCK_LOCK_NO_QUEUE;
+
+	linfo = malloc(sizeof(struct lock_info));
+	if (!linfo)
+		return -1;
+
+	DEBUGLOG("lock_resource '%s', flags=%d, mode=%d\n", resource, flags, mode);
+
+	linfo->lock_name.length = strlen(resource)+1;
+	strcpy((char *)linfo->lock_name.value, resource);
+
+	err = saLckResourceOpen(lck_handle, &linfo->lock_name,
+				SA_LCK_RESOURCE_CREATE, TIMEOUT, &res_handle);
+	if (err != SA_AIS_OK)
+	{
+		DEBUGLOG("ResourceOpen returned %d\n", err);
+		free(linfo);
+		return ais_to_errno(err);
+	}
+
+	err = saLckResourceLock(
+			res_handle,
+			&lock_id,
+			mode,
+			flags,
+			0,
+			SA_TIME_END,
+			&lockStatus);
+	if (err != SA_AIS_OK && lockStatus != SA_LCK_LOCK_GRANTED)
+	{
+		free(linfo);
+		saLckResourceClose(res_handle);
+		return ais_to_errno(err);
+	}
+
+	/* Wait for it to complete */
+
+	DEBUGLOG("lock_resource returning %d, lock_id=%" PRIx64 "\n",
+		 err, lock_id);
+
+	linfo->lock_id = lock_id;
+	linfo->res_handle = res_handle;
+
+	dm_hash_insert(lock_hash, resource, linfo);
+
+	return ais_to_errno(err);
+}
+
+
+static int _unlock_resource(char *resource, int lockid)
+{
+	SaAisErrorT err;
+	struct lock_info *linfo;
+
+	DEBUGLOG("unlock_resource %s\n", resource);
+	linfo = dm_hash_lookup(lock_hash, resource);
+	if (!linfo)
+		return 0;
+
+	DEBUGLOG("unlock_resource: lockid: %" PRIx64 "\n", linfo->lock_id);
+	err = saLckResourceUnlock(linfo->lock_id, SA_TIME_END);
+	if (err != SA_AIS_OK)
+	{
+		DEBUGLOG("Unlock returned %d\n", err);
+		return ais_to_errno(err);
+	}
+
+	/* Release the resource */
+	dm_hash_remove(lock_hash, resource);
+	saLckResourceClose(linfo->res_handle);
+	free(linfo);
+
+	return ais_to_errno(err);
+}
+
+static int _sync_lock(const char *resource, int mode, int flags, int *lockid)
+{
+	int status;
+	char lock1[strlen(resource)+3];
+	char lock2[strlen(resource)+3];
+
+	snprintf(lock1, sizeof(lock1), "%s-1", resource);
+	snprintf(lock2, sizeof(lock2), "%s-2", resource);
+
+	switch (mode)
+	{
+	case LCK_EXCL:
+		status = _lock_resource(lock1, SA_LCK_EX_LOCK_MODE, flags, lockid);
+		if (status)
+			goto out;
+
+		/* If we can't get this lock too then bail out */
+		status = _lock_resource(lock2, SA_LCK_EX_LOCK_MODE, LCK_NONBLOCK,
+					lockid);
+		if (status == SA_LCK_LOCK_NOT_QUEUED)
+		{
+			_unlock_resource(lock1, *lockid);
+			status = -1;
+			errno = EAGAIN;
+		}
+		break;
+
+	case LCK_PREAD:
+	case LCK_READ:
+		status = _lock_resource(lock1, SA_LCK_PR_LOCK_MODE, flags, lockid);
+		if (status)
+			goto out;
+		_unlock_resource(lock2, *lockid);
+		break;
+
+	case LCK_WRITE:
+		status = _lock_resource(lock2, SA_LCK_EX_LOCK_MODE, flags, lockid);
+		if (status)
+			goto out;
+		_unlock_resource(lock1, *lockid);
+		break;
+
+	default:
+		status = -1;
+		errno = EINVAL;
+		break;
+	}
+out:
+	*lockid = mode;
+	return status;
+}
+
+static int _sync_unlock(const char *resource, int lockid)
+{
+	int status = 0;
+	char lock1[strlen(resource)+3];
+	char lock2[strlen(resource)+3];
+
+	snprintf(lock1, sizeof(lock1), "%s-1", resource);
+	snprintf(lock2, sizeof(lock2), "%s-2", resource);
+
+	_unlock_resource(lock1, lockid);
+	_unlock_resource(lock2, lockid);
+
+	return status;
+}
+
+/* We are always quorate ! */
+static int _is_quorate()
+{
+	return 1;
+}
+
+static int _get_main_cluster_fd(void)
+{
+	int select_fd;
+
+	cpg_fd_get(cpg_handle, &select_fd);
+	return select_fd;
+}
+
+static int _cluster_fd_callback(struct local_client *fd, char *buf, int len,
+				const char *csid,
+				struct local_client **new_client)
+{
+	cluster_client = fd;
+	*new_client = NULL;
+	cpg_dispatch(cpg_handle, SA_DISPATCH_ONE);
+	return 1;
+}
+
+static int _cluster_send_message(const void *buf, int msglen, const char *csid,
+				 const char *errtext)
+{
+	struct iovec iov[2];
+	SaAisErrorT err;
+	int target_node;
+
+	if (csid)
+		memcpy(&target_node, csid, OPENAIS_CSID_LEN);
+	else
+		target_node = 0;
+
+	iov[0].iov_base = &target_node;
+	iov[0].iov_len = sizeof(int);
+	iov[1].iov_base = (char *)buf;
+	iov[1].iov_len = msglen;
+
+	err = cpg_mcast_joined(cpg_handle, CPG_TYPE_AGREED, iov, 2);
+	return ais_to_errno(err);
+}
+
+/* We don't have a cluster name to report here */
+static int _get_cluster_name(char *buf, int buflen)
+{
+	strncpy(buf, "OpenAIS", buflen);
+	return 0;
+}
+
+static struct cluster_ops _cluster_openais_ops = {
+	.name                     = "openais",
+	.cluster_init_completed   = NULL,
+	.cluster_send_message     = _cluster_send_message,
+	.name_from_csid           = _name_from_csid,
+	.csid_from_name           = _csid_from_name,
+	.get_num_nodes            = _get_num_nodes,
+	.cluster_fd_callback      = _cluster_fd_callback,
+	.get_main_cluster_fd      = _get_main_cluster_fd,
+	.cluster_do_node_callback = _cluster_do_node_callback,
+	.is_quorate               = _is_quorate,
+	.get_our_csid             = _get_our_csid,
+	.add_up_node              = _add_up_node,
+	.reread_config            = NULL,
+	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
+	.sync_lock                = _sync_lock,
+	.sync_unlock              = _sync_unlock,
+};
+
+struct cluster_ops *init_openais_cluster(void)
+{
+	if (!_init_cluster())
+		return &_cluster_openais_ops;
+
+	return NULL;
+}
--- a/daemons/clvmd/clvmd-singlenode.c
+++ b/daemons/clvmd/clvmd-singlenode.c
@@ -0,0 +1,382 @@
+/*
+ * Copyright (C) 2009-2013 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 Lesser General Public License v.2.1.
+ *
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "clvmd-common.h"
+
+#include <pthread.h>
+
+#include "locking.h"
+#include "clvm.h"
+#include "clvmd-comms.h"
+#include "clvmd.h"
+
+#include <sys/un.h>
+#include <sys/socket.h>
+#include <fcntl.h>
+
+static const char SINGLENODE_CLVMD_SOCKNAME[] = DEFAULT_RUN_DIR "/clvmd_singlenode.sock";
+static int listen_fd = -1;
+
+static struct dm_hash_table *_locks;
+static int _lockid;
+
+static pthread_mutex_t _lock_mutex = PTHREAD_MUTEX_INITIALIZER;
+/* Using one common condition for all locks for simplicity */
+static pthread_cond_t _lock_cond = PTHREAD_COND_INITIALIZER;
+
+struct lock {
+	struct dm_list list;
+	int lockid;
+	int mode;
+};
+
+static void close_comms(void)
+{
+	if (listen_fd != -1 && close(listen_fd))
+		stack;
+	(void)unlink(SINGLENODE_CLVMD_SOCKNAME);
+	listen_fd = -1;
+}
+
+static int init_comms(void)
+{
+	mode_t old_mask;
+	struct sockaddr_un addr = { .sun_family = AF_UNIX };
+
+	if (!dm_strncpy(addr.sun_path, SINGLENODE_CLVMD_SOCKNAME,
+			sizeof(addr.sun_path))) {
+		DEBUGLOG("%s: singlenode socket name too long.",
+			 SINGLENODE_CLVMD_SOCKNAME);
+		return -1;
+	}
+
+	close_comms();
+
+	(void) dm_prepare_selinux_context(SINGLENODE_CLVMD_SOCKNAME, S_IFSOCK);
+	old_mask = umask(0077);
+
+	listen_fd = socket(PF_UNIX, SOCK_STREAM, 0);
+	if (listen_fd < 0) {
+		DEBUGLOG("Can't create local socket: %s\n", strerror(errno));
+		goto error;
+	}
+	/* Set Close-on-exec */
+	if (fcntl(listen_fd, F_SETFD, 1)) {
+		DEBUGLOG("Setting CLOEXEC on client fd failed: %s\n", strerror(errno));
+		goto error;
+	}
+
+	if (bind(listen_fd, (struct sockaddr *)&addr, sizeof(addr)) < 0) {
+		DEBUGLOG("Can't bind local socket: %s\n", strerror(errno));
+		goto error;
+	}
+	if (listen(listen_fd, 10) < 0) {
+		DEBUGLOG("Can't listen local socket: %s\n", strerror(errno));
+		goto error;
+	}
+
+	umask(old_mask);
+	(void) dm_prepare_selinux_context(NULL, 0);
+	return 0;
+error:
+	umask(old_mask);
+	(void) dm_prepare_selinux_context(NULL, 0);
+	close_comms();
+	return -1;
+}
+
+static int _init_cluster(void)
+{
+	int r;
+
+	if (!(_locks = dm_hash_create(128))) {
+		DEBUGLOG("Failed to allocate single-node hash table.\n");
+		return 1;
+	}
+
+	r = init_comms();
+	if (r) {
+		dm_hash_destroy(_locks);
+		_locks = NULL;
+		return r;
+	}
+
+	DEBUGLOG("Single-node cluster initialised.\n");
+	return 0;
+}
+
+static void _cluster_closedown(void)
+{
+	close_comms();
+
+	/* If there is any awaited resource, kill it softly */
+	pthread_mutex_lock(&_lock_mutex);
+	dm_hash_destroy(_locks);
+	_locks = NULL;
+	_lockid = 0;
+	pthread_cond_broadcast(&_lock_cond); /* wakeup waiters */
+	pthread_mutex_unlock(&_lock_mutex);
+}
+
+static void _get_our_csid(char *csid)
+{
+	int nodeid = 1;
+	memcpy(csid, &nodeid, sizeof(int));
+}
+
+static int _csid_from_name(char *csid, const char *name)
+{
+	return 1;
+}
+
+static int _name_from_csid(const char *csid, char *name)
+{
+	strcpy(name, "SINGLENODE");
+	return 0;
+}
+
+static int _get_num_nodes(void)
+{
+	return 1;
+}
+
+/* Node is now known to be running a clvmd */
+static void _add_up_node(const char *csid)
+{
+}
+
+/* Call a callback for each node, so the caller knows whether it's up or down */
+static int _cluster_do_node_callback(struct local_client *master_client,
+				     void (*callback)(struct local_client *,
+				     const char *csid, int node_up))
+{
+	return 0;
+}
+
+int _lock_file(const char *file, uint32_t flags);
+
+static const char *_get_mode(int mode)
+{
+	switch (mode) {
+	case LCK_NULL: return "NULL";
+	case LCK_READ: return "READ";
+	case LCK_PREAD: return "PREAD";
+	case LCK_WRITE: return "WRITE";
+	case LCK_EXCL: return "EXCLUSIVE";
+	case LCK_UNLOCK: return "UNLOCK";
+	default: return "????";
+	}
+}
+
+/* Real locking */
+static int _lock_resource(const char *resource, int mode, int flags, int *lockid)
+{
+	/* DLM table of allowed transition states */
+	static const int _dlm_table[6][6] = {
+	/* Mode	   NL	CR	CW	PR	PW	EX */
+	/* NL */ { 1,	 1,	 1,	 1,	 1,	 1},
+	/* CR */ { 1,	 1,	 1,	 1,	 1,	 0},
+	/* CW */ { 1,	 1,	 1,	 0,	 0,	 0},
+	/* PR */ { 1,	 1,	 0,	 1,	 0,	 0},
+	/* PW */ { 1,	 1,	 0,	 0,	 0,	 0},
+	/* EX */ { 1,	 0,	 0,	 0,	 0,	 0}
+	};
+
+	struct lock *lck = NULL, *lckt;
+	struct dm_list *head;
+
+	DEBUGLOG("Locking resource %s, flags=0x%02x (%s%s%s), mode=%s (%d)\n",
+		 resource, flags,
+		 (flags & LCKF_NOQUEUE) ? "NOQUEUE" : "",
+		 ((flags & (LCKF_NOQUEUE | LCKF_CONVERT)) ==
+		  (LCKF_NOQUEUE | LCKF_CONVERT)) ? "|" : "",
+		 (flags & LCKF_CONVERT) ? "CONVERT" : "",
+		 _get_mode(mode), mode);
+
+	mode &= LCK_TYPE_MASK;
+	pthread_mutex_lock(&_lock_mutex);
+
+retry:
+	if (!(head = dm_hash_lookup(_locks, resource))) {
+		if (flags & LCKF_CONVERT) {
+			/* In real DLM, lock is identified only by lockid, resource is not used */
+			DEBUGLOG("Unlocked resource %s cannot be converted\n", resource);
+			goto_bad;
+		}
+		/* Add new locked resource */
+		if (!(head = dm_malloc(sizeof(struct dm_list))) ||
+		    !dm_hash_insert(_locks, resource, head)) {
+			dm_free(head);
+			goto_bad;
+		}
+
+		dm_list_init(head);
+	} else	/* Update/convert locked resource */
+		dm_list_iterate_items(lck, head) {
+			/* Check is all locks are compatible with requested lock */
+			if (flags & LCKF_CONVERT) {
+				if (lck->lockid != *lockid)
+					continue;
+
+				DEBUGLOG("Converting resource %s lockid=%d mode:%s -> %s...\n",
+					 resource, lck->lockid, _get_mode(lck->mode), _get_mode(mode));
+				dm_list_iterate_items(lckt, head) {
+					if ((lckt->lockid != *lockid) &&
+					    !_dlm_table[mode][lckt->mode]) {
+						if (!(flags & LCKF_NOQUEUE) &&
+						    /* TODO: Real dlm uses here conversion queues */
+						    !pthread_cond_wait(&_lock_cond, &_lock_mutex) &&
+						    _locks) /* End of the game? */
+							goto retry;
+						goto bad;
+					}
+				}
+				lck->mode = mode; /* Lock is now converted */
+				goto out;
+			} else if (!_dlm_table[mode][lck->mode]) {
+				DEBUGLOG("Resource %s already locked lockid=%d, mode:%s\n",
+					 resource, lck->lockid, _get_mode(lck->mode));
+				if (!(flags & LCKF_NOQUEUE) &&
+				    !pthread_cond_wait(&_lock_cond, &_lock_mutex) &&
+				    _locks) { /* End of the game? */
+					DEBUGLOG("Resource %s retrying lock in mode:%s...\n",
+						 resource, _get_mode(mode));
+					goto retry;
+				}
+				goto bad;
+			}
+		}
+
+	if (!(flags & LCKF_CONVERT)) {
+		if (!(lck = dm_malloc(sizeof(struct lock))))
+			goto_bad;
+
+		*lockid = lck->lockid = ++_lockid;
+		lck->mode = mode;
+		dm_list_add(head, &lck->list);
+	}
+out:
+	pthread_cond_broadcast(&_lock_cond); /* to wakeup waiters */
+	pthread_mutex_unlock(&_lock_mutex);
+	DEBUGLOG("Locked resource %s, lockid=%d, mode=%s\n",
+		 resource, lck->lockid, _get_mode(lck->mode));
+
+	return 0;
+bad:
+	pthread_cond_broadcast(&_lock_cond); /* to wakeup waiters */
+	pthread_mutex_unlock(&_lock_mutex);
+	DEBUGLOG("Failed to lock resource %s\n", resource);
+
+	return 1; /* fail */
+}
+
+static int _unlock_resource(const char *resource, int lockid)
+{
+	struct lock *lck;
+	struct dm_list *head;
+	int r = 1;
+
+	if (lockid < 0) {
+		DEBUGLOG("Not tracking unlock of lockid -1: %s, lockid=%d\n",
+			 resource, lockid);
+		return 1;
+	}
+
+	DEBUGLOG("Unlocking resource %s, lockid=%d\n", resource, lockid);
+	pthread_mutex_lock(&_lock_mutex);
+	pthread_cond_broadcast(&_lock_cond); /* wakeup waiters */
+
+	if (!(head = dm_hash_lookup(_locks, resource))) {
+		pthread_mutex_unlock(&_lock_mutex);
+		DEBUGLOG("Resource %s is not locked.\n", resource);
+		return 1;
+	}
+
+	dm_list_iterate_items(lck, head)
+		if (lck->lockid == lockid) {
+			dm_list_del(&lck->list);
+			dm_free(lck);
+			r = 0;
+			goto out;
+		}
+
+	DEBUGLOG("Resource %s has wrong lockid %d.\n", resource, lockid);
+out:
+	if (dm_list_empty(head)) {
+		//DEBUGLOG("Resource %s is no longer hashed (lockid=%d).\n", resource, lockid);
+		dm_hash_remove(_locks, resource);
+		dm_free(head);
+	}
+
+	pthread_mutex_unlock(&_lock_mutex);
+
+	return r;
+}
+
+static int _is_quorate(void)
+{
+	return 1;
+}
+
+static int _get_main_cluster_fd(void)
+{
+	return listen_fd;
+}
+
+static int _cluster_fd_callback(struct local_client *fd, char *buf, int len,
+				const char *csid,
+				struct local_client **new_client)
+{
+	return 1;
+}
+
+static int _cluster_send_message(const void *buf, int msglen,
+				 const char *csid,
+				 const char *errtext)
+{
+	return 0;
+}
+
+static int _get_cluster_name(char *buf, int buflen)
+{
+	return dm_strncpy(buf, "localcluster", buflen) ? 0 : 1;
+}
+
+static struct cluster_ops _cluster_singlenode_ops = {
+	.name                     = "singlenode",
+	.cluster_init_completed   = NULL,
+	.cluster_send_message     = _cluster_send_message,
+	.name_from_csid           = _name_from_csid,
+	.csid_from_name           = _csid_from_name,
+	.get_num_nodes            = _get_num_nodes,
+	.cluster_fd_callback      = _cluster_fd_callback,
+	.get_main_cluster_fd      = _get_main_cluster_fd,
+	.cluster_do_node_callback = _cluster_do_node_callback,
+	.is_quorate               = _is_quorate,
+	.get_our_csid             = _get_our_csid,
+	.add_up_node              = _add_up_node,
+	.reread_config            = NULL,
+	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
+	.sync_lock                = _lock_resource,
+	.sync_unlock              = _unlock_resource,
+};
+
+struct cluster_ops *init_singlenode_cluster(void)
+{
+	if (!_init_cluster())
+		return &_cluster_singlenode_ops;
+
+	return NULL;
+}
--- a/daemons/clvmd/clvmd.c
+++ b/daemons/clvmd/clvmd.c
--- a/daemons/clvmd/clvmd.h
+++ b/daemons/clvmd/clvmd.h
@@ -0,0 +1,126 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef _CLVMD_H
+#define _CLVMD_H
+
+#define CLVMD_MAJOR_VERSION 0
+#define CLVMD_MINOR_VERSION 2
+#define CLVMD_PATCH_VERSION 1
+
+/* Default time (in seconds) we will wait for all remote commands to execute
+   before declaring them dead */
+#define DEFAULT_CMD_TIMEOUT 60
+
+/* One of these for each reply we get from command execution on a node */
+struct node_reply {
+	char node[MAX_CLUSTER_MEMBER_NAME_LEN];
+	char *replymsg;
+	int status;
+	struct node_reply *next;
+};
+
+typedef enum {DEBUG_OFF, DEBUG_STDERR, DEBUG_SYSLOG} debug_t;
+
+/*
+ * These exist for the use of local sockets only when we are
+ * collecting responses from all cluster nodes
+ */
+struct localsock_bits {
+	struct node_reply *replies;
+	int num_replies;
+	int expected_replies;
+	time_t sent_time;	/* So we can check for timeouts */
+	int in_progress;	/* Only execute one cmd at a time per client */
+	int sent_out;		/* Flag to indicate that a command was sent
+				   to remote nodes */
+	void *private;		/* Private area for command processor use */
+	void *cmd;		/* Whole command as passed down local socket */
+	int cmd_len;		/* Length of above */
+	int pipe;		/* Pipe to send PRE completion status down */
+	int finished;		/* Flag to tell subthread to exit */
+	int all_success;	/* Set to 0 if any node (or the pre_command)
+				   failed */
+	int cleanup_needed;     /* helper for cleanup_zombie */
+	struct local_client *pipe_client;
+	pthread_t threadid;
+	enum { PRE_COMMAND, POST_COMMAND } state;
+	pthread_mutex_t mutex;	/* Main thread and worker synchronisation */
+	pthread_cond_t cond;
+};
+
+/* Entries for PIPE clients */
+struct pipe_bits {
+	struct local_client *client;	/* Actual (localsock) client */
+	pthread_t threadid;		/* Our own copy of the thread id */
+};
+
+/* Entries for Network socket clients */
+struct netsock_bits {
+	void *private;
+	int flags;
+};
+
+typedef int (*fd_callback_t) (struct local_client * fd, char *buf, int len,
+			      const char *csid,
+			      struct local_client ** new_client);
+
+/* One of these for each fd we are listening on */
+struct local_client {
+	int fd;
+	enum { CLUSTER_MAIN_SOCK, CLUSTER_DATA_SOCK, LOCAL_RENDEZVOUS,
+		    LOCAL_SOCK, THREAD_PIPE, CLUSTER_INTERNAL } type;
+	struct local_client *next;
+	unsigned short xid;
+	fd_callback_t callback;
+	uint8_t removeme;
+
+	union {
+		struct localsock_bits localsock;
+		struct pipe_bits pipe;
+		struct netsock_bits net;
+	} bits;
+};
+
+#define DEBUGLOG(fmt, args...) debuglog(fmt, ## args)
+
+#ifndef max
+#define max(a,b) ((a)>(b)?(a):(b))
+#endif
+
+/* The real command processor is in clvmd-command.c */
+extern int do_command(struct local_client *client, struct clvm_header *msg,
+		      int msglen, char **buf, int buflen, int *retlen);
+
+/* Pre and post command routines are called only on the local node */
+extern int do_pre_command(struct local_client *client);
+extern int do_post_command(struct local_client *client);
+extern void cmd_client_cleanup(struct local_client *client);
+extern int add_client(struct local_client *new_client);
+
+extern void clvmd_cluster_init_completed(void);
+extern void process_message(struct local_client *client, char *buf,
+			    int len, const char *csid);
+extern void debuglog(const char *fmt, ... )
+  __attribute__ ((format(printf, 1, 2)));
+
+void clvmd_set_debug(debug_t new_de);
+debug_t clvmd_get_debug(void);
+int clvmd_get_foreground(void);
+
+int sync_lock(const char *resource, int mode, int flags, int *lockid);
+int sync_unlock(const char *resource, int lockid);
+
+#endif
--- a/daemons/clvmd/lvm-functions.c
+++ b/daemons/clvmd/lvm-functions.c
@@ -0,0 +1,941 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2012 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "clvmd-common.h"
+
+#include <pthread.h>
+
+#include "clvm.h"
+#include "clvmd-comms.h"
+#include "clvmd.h"
+#include "lvm-functions.h"
+
+/* LVM2 headers */
+#include "toolcontext.h"
+#include "lvmcache.h"
+#include "lvm-globals.h"
+#include "activate.h"
+#include "archiver.h"
+#include "memlock.h"
+
+#include <syslog.h>
+
+static struct cmd_context *cmd = NULL;
+static struct dm_hash_table *lv_hash = NULL;
+static pthread_mutex_t lv_hash_lock;
+static pthread_mutex_t lvm_lock;
+static char last_error[1024];
+
+struct lv_info {
+	int lock_id;
+	int lock_mode;
+};
+
+static const char *decode_full_locking_cmd(uint32_t cmdl)
+{
+	static char buf[128];
+	const char *type;
+	const char *scope;
+	const char *command;
+
+	switch (cmdl & LCK_TYPE_MASK) {
+	case LCK_NULL:
+		type = "NULL";
+		break;
+	case LCK_READ:
+		type = "READ";
+		break;
+	case LCK_PREAD:
+		type = "PREAD";
+		break;
+	case LCK_WRITE:
+		type = "WRITE";
+		break;
+	case LCK_EXCL:
+		type = "EXCL";
+		break;
+	case LCK_UNLOCK:
+		type = "UNLOCK";
+		break;
+	default:
+		type = "unknown";
+		break;
+	}
+
+	switch (cmdl & LCK_SCOPE_MASK) {
+	case LCK_VG:
+		scope = "VG";
+		command = "LCK_VG";
+		break;
+	case LCK_LV:
+		scope = "LV";
+		switch (cmdl & LCK_MASK) {
+		case LCK_LV_EXCLUSIVE & LCK_MASK:
+			command = "LCK_LV_EXCLUSIVE";
+			break;
+		case LCK_LV_SUSPEND & LCK_MASK:
+			command = "LCK_LV_SUSPEND";
+			break;
+		case LCK_LV_RESUME & LCK_MASK:
+			command = "LCK_LV_RESUME";
+			break;
+		case LCK_LV_ACTIVATE & LCK_MASK:
+			command = "LCK_LV_ACTIVATE";
+			break;
+		case LCK_LV_DEACTIVATE & LCK_MASK:
+			command = "LCK_LV_DEACTIVATE";
+			break;
+		default:
+			command = "unknown";
+			break;
+		}
+		break;
+	default:
+		scope = "unknown";
+		command = "unknown";
+		break;
+	}
+
+	sprintf(buf, "0x%x %s (%s|%s%s%s%s%s)", cmdl, command, type, scope,
+		cmdl & LCK_NONBLOCK   ? "|NONBLOCK" : "",
+		cmdl & LCK_HOLD       ? "|HOLD" : "",
+		cmdl & LCK_CLUSTER_VG ? "|CLUSTER_VG" : "",
+		cmdl & LCK_CACHE      ? "|CACHE" : "");
+
+	return buf;
+}
+
+/*
+ * Only processes 8 bits: excludes LCK_CACHE.
+ */
+static const char *decode_locking_cmd(unsigned char cmdl)
+{
+	return decode_full_locking_cmd((uint32_t) cmdl);
+}
+
+static const char *decode_flags(unsigned char flags)
+{
+	static char buf[128];
+	int len;
+
+	len = sprintf(buf, "0x%x ( %s%s%s%s%s%s%s%s)", flags,
+		flags & LCK_PARTIAL_MODE	  ? "PARTIAL_MODE|" : "",
+		flags & LCK_MIRROR_NOSYNC_MODE	  ? "MIRROR_NOSYNC|" : "",
+		flags & LCK_DMEVENTD_MONITOR_MODE ? "DMEVENTD_MONITOR|" : "",
+		flags & LCK_ORIGIN_ONLY_MODE ? "ORIGIN_ONLY|" : "",
+		flags & LCK_TEST_MODE ? "TEST|" : "",
+		flags & LCK_CONVERT_MODE ? "CONVERT|" : "",
+		flags & LCK_DMEVENTD_MONITOR_IGNORE ? "DMEVENTD_MONITOR_IGNORE|" : "",
+		flags & LCK_REVERT_MODE ? "REVERT|" : "");
+
+	if (len > 1)
+		buf[len - 2] = ' ';
+	else
+		buf[0] = '\0';
+
+	return buf;
+}
+
+char *get_last_lvm_error(void)
+{
+	return last_error;
+}
+
+/*
+ * Hash lock info helpers
+ */
+static struct lv_info *lookup_info(const char *resource)
+{
+	struct lv_info *lvi;
+
+	pthread_mutex_lock(&lv_hash_lock);
+	lvi = dm_hash_lookup(lv_hash, resource);
+	pthread_mutex_unlock(&lv_hash_lock);
+
+	return lvi;
+}
+
+static int insert_info(const char *resource, struct lv_info *lvi)
+{
+	int ret;
+
+	pthread_mutex_lock(&lv_hash_lock);
+	ret = dm_hash_insert(lv_hash, resource, lvi);
+	pthread_mutex_unlock(&lv_hash_lock);
+
+	return ret;
+}
+
+static void remove_info(const char *resource)
+{
+	int num_open;
+
+	pthread_mutex_lock(&lv_hash_lock);
+	dm_hash_remove(lv_hash, resource);
+
+	/* When last lock is remove, validate there are not left opened devices */
+	if (!dm_hash_get_first(lv_hash)) {
+		if (critical_section())
+			log_error(INTERNAL_ERROR "No volumes are locked however clvmd is in activation mode critical section.");
+		if ((num_open = dev_cache_check_for_open_devices()))
+			log_error(INTERNAL_ERROR "No volumes are locked however %d devices are still open.", num_open);
+	}
+
+	pthread_mutex_unlock(&lv_hash_lock);
+}
+
+/*
+ * Return the mode a lock is currently held at (or -1 if not held)
+ */
+static int get_current_lock(char *resource)
+{
+	struct lv_info *lvi;
+
+	if ((lvi = lookup_info(resource)))
+		return lvi->lock_mode;
+
+	return -1;
+}
+
+
+void init_lvhash(void)
+{
+	/* Create hash table for keeping LV locks & status */
+	lv_hash = dm_hash_create(1024);
+	pthread_mutex_init(&lv_hash_lock, NULL);
+	pthread_mutex_init(&lvm_lock, NULL);
+}
+
+/* Called at shutdown to tidy the lockspace */
+void destroy_lvhash(void)
+{
+	struct dm_hash_node *v;
+	struct lv_info *lvi;
+	char *resource;
+	int status;
+
+	pthread_mutex_lock(&lv_hash_lock);
+
+	dm_hash_iterate(v, lv_hash) {
+		lvi = dm_hash_get_data(lv_hash, v);
+		resource = dm_hash_get_key(lv_hash, v);
+
+		if ((status = sync_unlock(resource, lvi->lock_id)))
+			DEBUGLOG("unlock_all. unlock failed(%d): %s\n",
+				 status,  strerror(errno));
+		dm_free(lvi);
+	}
+
+	dm_hash_destroy(lv_hash);
+	lv_hash = NULL;
+
+	pthread_mutex_unlock(&lv_hash_lock);
+}
+
+/* Gets a real lock and keeps the info in the hash table */
+static int hold_lock(char *resource, int mode, int flags)
+{
+	int status;
+	int saved_errno;
+	struct lv_info *lvi;
+
+	/* Mask off invalid options */
+	flags &= LCKF_NOQUEUE | LCKF_CONVERT;
+
+	lvi = lookup_info(resource);
+
+	if (lvi) {
+		if (lvi->lock_mode == mode) {
+			DEBUGLOG("hold_lock, lock mode %d already held\n",
+				 mode);
+			return 0;
+		}
+		if ((lvi->lock_mode == LCK_EXCL) && (mode == LCK_WRITE)) {
+			DEBUGLOG("hold_lock, lock already held LCK_EXCL, "
+				 "ignoring LCK_WRITE request\n");
+			return 0;
+		}
+	}
+
+	/* Only allow explicit conversions */
+	if (lvi && !(flags & LCKF_CONVERT)) {
+		errno = EBUSY;
+		return -1;
+	}
+	if (lvi) {
+		/* Already exists - convert it */
+		status = sync_lock(resource, mode, flags, &lvi->lock_id);
+		saved_errno = errno;
+		if (!status)
+			lvi->lock_mode = mode;
+		else
+			DEBUGLOG("hold_lock. convert to %d failed: %s\n", mode,
+				 strerror(errno));
+		errno = saved_errno;
+	} else {
+		if (!(lvi = dm_malloc(sizeof(struct lv_info)))) {
+			errno = ENOMEM;
+			return -1;
+		}
+
+		lvi->lock_mode = mode;
+		lvi->lock_id = 0;
+		status = sync_lock(resource, mode, flags & ~LCKF_CONVERT, &lvi->lock_id);
+		saved_errno = errno;
+		if (status) {
+			dm_free(lvi);
+			DEBUGLOG("hold_lock. lock at %d failed: %s\n", mode,
+				 strerror(errno));
+		} else
+			if (!insert_info(resource, lvi)) {
+				errno = ENOMEM;
+				return -1;
+			}
+
+		errno = saved_errno;
+	}
+	return status;
+}
+
+/* Unlock and remove it from the hash table */
+static int hold_unlock(char *resource)
+{
+	struct lv_info *lvi;
+	int status;
+	int saved_errno;
+
+	if (!(lvi = lookup_info(resource))) {
+		DEBUGLOG("hold_unlock, lock not already held\n");
+		return 0;
+	}
+
+	status = sync_unlock(resource, lvi->lock_id);
+	saved_errno = errno;
+	if (!status) {
+		remove_info(resource);
+		dm_free(lvi);
+	} else {
+		DEBUGLOG("hold_unlock. unlock failed(%d): %s\n", status,
+			 strerror(errno));
+	}
+
+	errno = saved_errno;
+	return status;
+}
+
+/* Watch the return codes here.
+   liblvm API functions return 1(true) for success, 0(false) for failure and don't set errno.
+   libdlm API functions return 0 for success, -1 for failure and do set errno.
+   These functions here return 0 for success or >0 for failure (where the retcode is errno)
+*/
+
+/* Activate LV exclusive or non-exclusive */
+static int do_activate_lv(char *resource, unsigned char command, unsigned char lock_flags, int mode)
+{
+	int oldmode;
+	int status;
+	int activate_lv;
+	int exclusive = 0;
+	struct lvinfo lvi;
+
+	/* Is it already open ? */
+	oldmode = get_current_lock(resource);
+	if (oldmode == mode && (command & LCK_CLUSTER_VG)) {
+		DEBUGLOG("do_activate_lv, lock already held at %d\n", oldmode);
+		return 0;	/* Nothing to do */
+	}
+
+	/* Does the config file want us to activate this LV ? */
+	if (!lv_activation_filter(cmd, resource, &activate_lv, NULL))
+		return EIO;
+
+	if (!activate_lv)
+		return 0;	/* Success, we did nothing! */
+
+	/* Do we need to activate exclusively? */
+	if ((activate_lv == 2) || (mode == LCK_EXCL)) {
+		exclusive = 1;
+		mode = LCK_EXCL;
+	}
+
+	/*
+	 * Try to get the lock if it's a clustered volume group.
+	 * Use lock conversion only if requested, to prevent implicit conversion
+	 * of exclusive lock to shared one during activation.
+	 */
+	if (!test_mode() && command & LCK_CLUSTER_VG) {
+		status = hold_lock(resource, mode, LCKF_NOQUEUE | ((lock_flags & LCK_CONVERT_MODE) ? LCKF_CONVERT:0));
+		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 */
+	if (!lv_info_by_lvid(cmd, resource, 0, &lvi, 0, 0))
+		goto error;
+
+	if (lvi.suspended) {
+		critical_section_inc(cmd, "resuming");
+		if (!lv_resume(cmd, resource, 0, NULL)) {
+			critical_section_dec(cmd, "resumed");
+			goto error;
+		}
+	}
+
+	/* Now activate it */
+	if (!lv_activate(cmd, resource, exclusive, 0, 0, NULL))
+		goto error;
+
+	return 0;
+
+error:
+	if (!test_mode() && (oldmode == -1 || oldmode != mode))
+		(void)hold_unlock(resource);
+	return EIO;
+}
+
+/* Resume the LV if it was active */
+static int do_resume_lv(char *resource, unsigned char command, unsigned char lock_flags)
+{
+	int oldmode, origin_only, exclusive, revert;
+
+	/* Is it open ? */
+	oldmode = get_current_lock(resource);
+	if (oldmode == -1 && (command & LCK_CLUSTER_VG)) {
+		DEBUGLOG("do_resume_lv, lock not already held\n");
+		return 0;	/* We don't need to do anything */
+	}
+	origin_only = (lock_flags & LCK_ORIGIN_ONLY_MODE) ? 1 : 0;
+	exclusive = (oldmode == LCK_EXCL) ? 1 : 0;
+	revert = (lock_flags & LCK_REVERT_MODE) ? 1 : 0;
+
+	if (!lv_resume_if_active(cmd, resource, origin_only, exclusive, revert, NULL))
+		return EIO;
+
+	return 0;
+}
+
+/* Suspend the device if active */
+static int do_suspend_lv(char *resource, unsigned char command, unsigned char lock_flags)
+{
+	int oldmode;
+	unsigned origin_only = (lock_flags & LCK_ORIGIN_ONLY_MODE) ? 1 : 0;
+	unsigned exclusive;
+
+	/* Is it open ? */
+	oldmode = get_current_lock(resource);
+	if (oldmode == -1 && (command & LCK_CLUSTER_VG)) {
+		DEBUGLOG("do_suspend_lv, lock not already held\n");
+		return 0; /* Not active, so it's OK */
+	}
+
+	exclusive = (oldmode == LCK_EXCL) ? 1 : 0;
+
+	/* Always call lv_suspend to read commited and precommited data */
+	if (!lv_suspend_if_active(cmd, resource, origin_only, exclusive, NULL, NULL))
+		return EIO;
+
+	return 0;
+}
+
+static int do_deactivate_lv(char *resource, unsigned char command, unsigned char lock_flags)
+{
+	int oldmode;
+	int status;
+
+	/* Is it open ? */
+	oldmode = get_current_lock(resource);
+	if (oldmode == -1 && (command & LCK_CLUSTER_VG)) {
+		DEBUGLOG("do_deactivate_lock, lock not already held\n");
+		return 0;	/* We don't need to do anything */
+	}
+
+	if (!lv_deactivate(cmd, resource, NULL))
+		return EIO;
+
+	if (!test_mode() && command & LCK_CLUSTER_VG) {
+		status = hold_unlock(resource);
+		if (status)
+			return errno;
+	}
+
+	return 0;
+}
+
+const char *do_lock_query(char *resource)
+{
+	int mode;
+	const char *type;
+
+	mode = get_current_lock(resource);
+	switch (mode) {
+	case LCK_NULL: type = "NL"; break;
+	case LCK_READ: type = "CR"; break;
+	case LCK_PREAD:type = "PR"; break;
+	case LCK_WRITE:type = "PW"; break;
+	case LCK_EXCL: type = "EX"; break;
+	default: type = NULL;
+	}
+
+	DEBUGLOG("do_lock_query: resource '%s', mode %i (%s)\n", resource, mode, type ?: "--");
+
+	return type;
+}
+
+/* This is the LOCK_LV part that happens on all nodes in the cluster -
+   it is responsible for the interaction with device-mapper and LVM */
+int do_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
+{
+	int status = 0;
+
+	DEBUGLOG("do_lock_lv: resource '%s', cmd = %s, flags = %s, critical_section = %d\n",
+		 resource, decode_locking_cmd(command), decode_flags(lock_flags), critical_section());
+
+	if (!cmd->initialized.config || config_files_changed(cmd)) {
+		/* Reinitialise various settings inc. logging, filters */
+		if (do_refresh_cache()) {
+			log_error("Updated config file invalid. Aborting.");
+			return EINVAL;
+		}
+	}
+
+	pthread_mutex_lock(&lvm_lock);
+	init_test((lock_flags & LCK_TEST_MODE) ? 1 : 0);
+
+	if (lock_flags & LCK_MIRROR_NOSYNC_MODE)
+		init_mirror_in_sync(1);
+
+	if (lock_flags & LCK_DMEVENTD_MONITOR_IGNORE)
+		init_dmeventd_monitor(DMEVENTD_MONITOR_IGNORE);
+	else {
+		if (lock_flags & LCK_DMEVENTD_MONITOR_MODE)
+			init_dmeventd_monitor(1);
+		else
+			init_dmeventd_monitor(0);
+	}
+
+	cmd->partial_activation = (lock_flags & LCK_PARTIAL_MODE) ? 1 : 0;
+
+	/* clvmd should never try to read suspended device */
+	init_ignore_suspended_devices(1);
+
+	switch (command & LCK_MASK) {
+	case LCK_LV_EXCLUSIVE:
+		status = do_activate_lv(resource, command, lock_flags, LCK_EXCL);
+		break;
+
+	case LCK_LV_SUSPEND:
+		status = do_suspend_lv(resource, command, lock_flags);
+		break;
+
+	case LCK_UNLOCK:
+	case LCK_LV_RESUME:	/* if active */
+		status = do_resume_lv(resource, command, lock_flags);
+		break;
+
+	case LCK_LV_ACTIVATE:
+		status = do_activate_lv(resource, command, lock_flags, LCK_READ);
+		break;
+
+	case LCK_LV_DEACTIVATE:
+		status = do_deactivate_lv(resource, command, lock_flags);
+		break;
+
+	default:
+		DEBUGLOG("Invalid LV command 0x%x\n", command);
+		status = EINVAL;
+		break;
+	}
+
+	if (lock_flags & LCK_MIRROR_NOSYNC_MODE)
+		init_mirror_in_sync(0);
+
+	cmd->partial_activation = 0;
+
+	/* clean the pool for another command */
+	dm_pool_empty(cmd->mem);
+	init_test(0);
+	pthread_mutex_unlock(&lvm_lock);
+
+	DEBUGLOG("Command return is %d, critical_section is %d\n", status, critical_section());
+	return status;
+}
+
+/* Functions to do on the local node only BEFORE the cluster-wide stuff above happens */
+int pre_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
+{
+	/* Nearly all the stuff happens cluster-wide. Apart from SUSPEND. Here we get the
+	   lock out on this node (because we are the node modifying the metadata)
+	   before suspending cluster-wide.
+	   LCKF_CONVERT is used always, local node is going to modify metadata
+	 */
+	if ((command & (LCK_SCOPE_MASK | LCK_TYPE_MASK)) == LCK_LV_SUSPEND &&
+	    (command & LCK_CLUSTER_VG)) {
+		DEBUGLOG("pre_lock_lv: resource '%s', cmd = %s, flags = %s\n",
+			 resource, decode_locking_cmd(command), decode_flags(lock_flags));
+
+		if (!(lock_flags & LCK_TEST_MODE) &&
+		    hold_lock(resource, LCK_WRITE, LCKF_NOQUEUE | LCKF_CONVERT))
+			return errno;
+	}
+	return 0;
+}
+
+/* Functions to do on the local node only AFTER the cluster-wide stuff above happens */
+int post_lock_lv(unsigned char command, unsigned char lock_flags,
+		 char *resource)
+{
+	int status;
+	unsigned origin_only = (lock_flags & LCK_ORIGIN_ONLY_MODE) ? 1 : 0;
+
+	/* Opposite of above, done on resume after a metadata update */
+	if ((command & (LCK_SCOPE_MASK | LCK_TYPE_MASK)) == LCK_LV_RESUME &&
+	    (command & LCK_CLUSTER_VG)) {
+		int oldmode;
+
+		DEBUGLOG("post_lock_lv: resource '%s', cmd = %s, flags = %s\n",
+			 resource, decode_locking_cmd(command), decode_flags(lock_flags));
+
+		/* If the lock state is PW then restore it to what it was */
+		oldmode = get_current_lock(resource);
+		if (oldmode == LCK_WRITE) {
+			struct lvinfo lvi;
+
+			pthread_mutex_lock(&lvm_lock);
+			status = lv_info_by_lvid(cmd, resource, origin_only, &lvi, 0, 0);
+			pthread_mutex_unlock(&lvm_lock);
+			if (!status)
+				return EIO;
+
+			if (!(lock_flags & LCK_TEST_MODE)) {
+				if (lvi.exists) {
+					if (hold_lock(resource, LCK_READ, LCKF_CONVERT))
+						return errno;
+				} else if (hold_unlock(resource))
+					return errno;
+			}
+		}
+	}
+	return 0;
+}
+
+/* Check if a VG is in use by LVM1 so we don't stomp on it */
+int do_check_lvm1(const char *vgname)
+{
+	int status;
+
+	status = check_lvm1_vg_inactive(cmd, vgname);
+
+	return status == 1 ? 0 : EBUSY;
+}
+
+int do_refresh_cache(void)
+{
+	DEBUGLOG("Refreshing context\n");
+	log_notice("Refreshing context");
+
+	pthread_mutex_lock(&lvm_lock);
+
+	if (!refresh_toolcontext(cmd)) {
+		pthread_mutex_unlock(&lvm_lock);
+		return -1;
+	}
+
+	cmd->use_aio = 0;
+	init_full_scan_done(0);
+	init_ignore_suspended_devices(1);
+	lvmcache_force_next_label_scan();
+	lvmcache_label_scan(cmd);
+	dm_pool_empty(cmd->mem);
+
+	pthread_mutex_unlock(&lvm_lock);
+
+	return 0;
+}
+
+/*
+ * Handle VG lock - drop metadata or update lvmcache state
+ */
+void do_lock_vg(unsigned char command, unsigned char lock_flags, char *resource)
+{
+	uint32_t lock_cmd = command;
+	char *vgname = resource + 2;
+
+	lock_cmd &= (LCK_SCOPE_MASK | LCK_TYPE_MASK | LCK_HOLD);
+
+	/*
+	 * Check if LCK_CACHE should be set. All P_ locks except # are cache related.
+	 */
+	if (strncmp(resource, "P_#", 3) && !strncmp(resource, "P_", 2))
+		lock_cmd |= LCK_CACHE;
+
+	DEBUGLOG("do_lock_vg: resource '%s', cmd = %s, flags = %s, critical_section = %d\n",
+		 resource, decode_full_locking_cmd(lock_cmd), decode_flags(lock_flags), critical_section());
+
+	/* P_#global causes a full cache refresh */
+	if (!strcmp(resource, "P_" VG_GLOBAL)) {
+		do_refresh_cache();
+		return;
+	}
+
+	pthread_mutex_lock(&lvm_lock);
+	init_test((lock_flags & LCK_TEST_MODE) ? 1 : 0);
+
+	switch (lock_cmd) {
+		case LCK_VG_COMMIT:
+			DEBUGLOG("vg_commit notification for VG %s\n", vgname);
+			lvmcache_commit_metadata(vgname);
+			break;
+		case LCK_VG_REVERT:
+			DEBUGLOG("vg_revert notification for VG %s\n", vgname);
+			lvmcache_drop_metadata(vgname, 1);
+			break;
+		case LCK_VG_DROP_CACHE:
+		default:
+			DEBUGLOG("Invalidating cached metadata for VG %s\n", vgname);
+			lvmcache_drop_metadata(vgname, 0);
+	}
+
+	init_test(0);
+	pthread_mutex_unlock(&lvm_lock);
+}
+
+/*
+ * Ideally, clvmd should be started before any LVs are active
+ * but this may not be the case...
+ * I suppose this also comes in handy if clvmd crashes, not that it would!
+ */
+static int get_initial_state(struct dm_hash_table *excl_uuid)
+{
+	int lock_mode;
+	char lv[65], vg[65], flags[26], vg_flags[26]; /* with space for '\0' */
+	char uuid[65];
+	char line[255];
+	char *lvs_cmd;
+	const char *lvm_binary = getenv("LVM_BINARY") ? : LVM_PATH;
+	FILE *lvs;
+
+	if (dm_asprintf(&lvs_cmd, "%s lvs  --config 'log{command_names=0 prefix=\"\"}' "
+			"--nolocking --noheadings -o vg_uuid,lv_uuid,lv_attr,vg_attr",
+			lvm_binary) < 0)
+		return_0;
+
+	/* FIXME: Maybe link and use liblvm2cmd directly instead of fork */
+	if (!(lvs = popen(lvs_cmd, "r"))) {
+		dm_free(lvs_cmd);
+		return 0;
+	}
+
+	while (fgets(line, sizeof(line), lvs)) {
+	        if (sscanf(line, "%64s %64s %25s %25s\n", vg, lv, flags, vg_flags) == 4) {
+
+			/* States: s:suspended a:active S:dropped snapshot I:invalid snapshot */
+		        if (strlen(vg) == 38 &&                         /* is is a valid UUID ? */
+			    (flags[4] == 'a' || flags[4] == 's') &&	/* is it active or suspended? */
+			    vg_flags[5] == 'c') {			/* is it clustered ? */
+				/* Convert hyphen-separated UUIDs into one */
+				memcpy(&uuid[0], &vg[0], 6);
+				memcpy(&uuid[6], &vg[7], 4);
+				memcpy(&uuid[10], &vg[12], 4);
+				memcpy(&uuid[14], &vg[17], 4);
+				memcpy(&uuid[18], &vg[22], 4);
+				memcpy(&uuid[22], &vg[27], 4);
+				memcpy(&uuid[26], &vg[32], 6);
+				memcpy(&uuid[32], &lv[0], 6);
+				memcpy(&uuid[38], &lv[7], 4);
+				memcpy(&uuid[42], &lv[12], 4);
+				memcpy(&uuid[46], &lv[17], 4);
+				memcpy(&uuid[50], &lv[22], 4);
+				memcpy(&uuid[54], &lv[27], 4);
+				memcpy(&uuid[58], &lv[32], 6);
+				uuid[64] = '\0';
+
+				/* Look for this lock in the list of EX locks
+				   we were passed on the command-line */
+				lock_mode = (dm_hash_lookup(excl_uuid, uuid)) ?
+					LCK_EXCL : LCK_READ;
+
+				DEBUGLOG("getting initial lock for %s\n", uuid);
+				if (hold_lock(uuid, lock_mode, LCKF_NOQUEUE))
+					DEBUGLOG("Failed to hold lock %s\n", uuid);
+			}
+		}
+	}
+	if (pclose(lvs))
+		DEBUGLOG("lvs pclose failed: %s\n", strerror(errno));
+
+	dm_free(lvs_cmd);
+
+	return 1;
+}
+
+static void lvm2_log_fn(int level, const char *file, int line, int dm_errno,
+			const char *message)
+{
+
+	/* Send messages to the normal LVM2 logging system too,
+	   so we get debug output when it's asked for.
+ 	   We need to NULL the function ptr otherwise it will just call
+	   back into here! */
+	init_log_fn(NULL);
+	print_log(level, file, line, dm_errno, "%s", message);
+	init_log_fn(lvm2_log_fn);
+
+	/*
+	 * 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(void)
+{
+	int locking_type;
+
+	locking_type = find_config_tree_int(cmd, global_locking_type_CFG, NULL);
+
+	if (locking_type == 3) /* compiled-in cluster support */
+		return;
+
+	if (locking_type == 2) { /* External library, check name */
+		const char *libname;
+
+		libname = find_config_tree_str(cmd, global_locking_library_CFG, NULL);
+		if (libname && strstr(libname, "liblvm2clusterlock.so"))
+			return;
+
+		log_error("Incorrect LVM locking library specified in lvm.conf, cluster operations may not work.");
+		return;
+	}
+	log_error("locking_type not set correctly in lvm.conf, cluster operations will not work.");
+}
+
+/* Backups up the LVM metadata if it's changed */
+void lvm_do_backup(const char *vgname)
+{
+	struct volume_group * vg;
+	int consistent = 0;
+
+	DEBUGLOG("Triggering backup of VG metadata for %s.\n", vgname);
+
+	pthread_mutex_lock(&lvm_lock);
+
+	vg = vg_read_internal(cmd, vgname, NULL /*vgid*/, WARN_PV_READ, &consistent);
+
+	if (vg && consistent)
+		check_current_backup(vg);
+	else
+		log_error("Error backing up metadata, can't find VG for group %s", vgname);
+
+	release_vg(vg);
+	dm_pool_empty(cmd->mem);
+
+	pthread_mutex_unlock(&lvm_lock);
+}
+
+struct dm_hash_node *get_next_excl_lock(struct dm_hash_node *v, char **name)
+{
+	struct lv_info *lvi;
+
+	*name = NULL;
+	if (!v)
+		v = dm_hash_get_first(lv_hash);
+
+	do {
+		if (v) {
+			lvi = dm_hash_get_data(lv_hash, v);
+			DEBUGLOG("Looking for EX locks. found %x mode %d\n", lvi->lock_id, lvi->lock_mode);
+
+			if (lvi->lock_mode == LCK_EXCL) {
+				*name = dm_hash_get_key(lv_hash, v);
+			}
+			v = dm_hash_get_next(lv_hash, v);
+		}
+	} while (v && !*name);
+
+	if (*name)
+		DEBUGLOG("returning EXclusive UUID %s\n", *name);
+	return v;
+}
+
+void lvm_do_fs_unlock(void)
+{
+	pthread_mutex_lock(&lvm_lock);
+	DEBUGLOG("Syncing device names\n");
+	fs_unlock();
+	pthread_mutex_unlock(&lvm_lock);
+}
+
+/* Called to initialise the LVM context of the daemon */
+int init_clvm(struct dm_hash_table *excl_uuid)
+{
+	/* Use LOG_DAEMON for syslog messages instead of LOG_USER */
+	init_syslog(LOG_DAEMON);
+	openlog("clvmd", LOG_PID, LOG_DAEMON);
+
+	/* Initialise already held locks */
+	if (!get_initial_state(excl_uuid))
+		log_error("Cannot load initial lock states.");
+
+	if (!udev_init_library_context())
+		stack;
+
+	if (!(cmd = create_toolcontext(1, NULL, 0, 1, 1, 1))) {
+		log_error("Failed to allocate command context");
+		udev_fin_library_context();
+		return 0;
+	}
+
+	if (stored_errno()) {
+		destroy_toolcontext(cmd);
+		return 0;
+	}
+
+	cmd->cmd_line = "clvmd";
+
+	/* Check lvm.conf is setup for cluster-LVM */
+	check_config();
+	init_ignore_suspended_devices(1);
+	cmd->use_aio = 0;
+
+	/* Trap log messages so we can pass them back to the user */
+	init_log_fn(lvm2_log_fn);
+	memlock_inc_daemon(cmd);
+
+	return 1;
+}
+
+void destroy_lvm(void)
+{
+	if (cmd) {
+		memlock_dec_daemon(cmd);
+		destroy_toolcontext(cmd);
+		udev_fin_library_context();
+		cmd = NULL;
+	}
+}
--- a/daemons/clvmd/lvm-functions.h
+++ b/daemons/clvmd/lvm-functions.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2010 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/* Functions in lvm-functions.c */
+
+#ifndef _LVM_FUNCTIONS_H
+#define _LVM_FUNCTIONS_H
+
+extern int pre_lock_lv(unsigned char lock_cmd, unsigned char lock_flags,
+		       char *resource);
+extern int do_lock_lv(unsigned char lock_cmd, unsigned char lock_flags,
+		      char *resource);
+extern const char *do_lock_query(char *resource);
+extern int post_lock_lv(unsigned char lock_cmd, unsigned char lock_flags,
+			char *resource);
+extern int do_check_lvm1(const char *vgname);
+extern int do_refresh_cache(void);
+extern int init_clvm(struct dm_hash_table *excl_uuid);
+extern void destroy_lvm(void);
+extern void init_lvhash(void);
+extern void destroy_lvhash(void);
+extern void lvm_do_backup(const char *vgname);
+extern char *get_last_lvm_error(void);
+extern void do_lock_vg(unsigned char command, unsigned char lock_flags,
+		      char *resource);
+extern struct dm_hash_node *get_next_excl_lock(struct dm_hash_node *v, char **name);
+void lvm_do_fs_unlock(void);
+
+#endif
--- a/daemons/clvmd/refresh_clvmd.c
+++ b/daemons/clvmd/refresh_clvmd.c
@@ -0,0 +1,382 @@
+/*
+ * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2010 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+/* FIXME Remove duplicated functions from this file. */
+
+/*
+ * Send a command to a running clvmd from the command-line
+ */
+
+#include "clvmd-common.h"
+
+#include "clvm.h"
+#include "refresh_clvmd.h"
+
+#include <stddef.h>
+#include <sys/socket.h>
+#include <sys/un.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 clvm daemon */
+static int _open_local_sock(void)
+{
+	int local_socket;
+	struct sockaddr_un sockaddr = { .sun_family = AF_UNIX };
+
+	if (!dm_strncpy(sockaddr.sun_path, CLVMD_SOCKNAME, sizeof(sockaddr.sun_path))) {
+		fprintf(stderr, "%s: clvmd socket name too long.", CLVMD_SOCKNAME);
+		return -1;
+	}
+
+	/* Open local socket */
+	if ((local_socket = socket(PF_UNIX, SOCK_STREAM, 0)) < 0) {
+		fprintf(stderr, "Local socket creation failed: %s", strerror(errno));
+		return -1;
+	}
+
+	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(const char *inbuf, int inlen, char **retbuf, int no_response)
+{
+	char outbuf[PIPE_BUF];
+	struct clvm_header *outheader = (struct clvm_header *) outbuf;
+	int len;
+	unsigned 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;
+	}
+	if (no_response)
+		return 1;
+
+	/* 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,
+			  unsigned int len)
+{
+	head->cmd = cmd;
+	head->status = 0;
+	head->flags = 0;
+	head->xid = 0;
+	head->clientid = 0;
+	if (len)
+		/* 1 byte is used from struct clvm_header.args[1], so -> len - 1 */
+		head->arglen = len - 1;
+	else {
+		head->arglen = 0;
+		*head->args = '\0';
+	}
+
+	/*
+	 * Translate special node names.
+	 */
+	if (!node || !strcmp(node, NODE_ALL))
+		head->node[0] = '\0';
+	else if (!strcmp(node, NODE_LOCAL)) {
+		head->node[0] = '\0';
+		head->flags = CLVMD_FLAG_LOCAL;
+	} else
+		strcpy(head->node, node);
+}
+
+/*
+ * 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, int no_response)
+{
+	char outbuf[sizeof(struct clvm_header) + len + strlen(node) + 1];
+	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);
+	if (len)
+		memcpy(head->node + strlen(head->node) + 1, data, len);
+
+	status = _send_request(outbuf, sizeof(struct clvm_header) +
+			       strlen(head->node) + len, &retbuf, no_response);
+	if (!status || no_response)
+		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
+	 */
+	*response = NULL;
+	if (!(rarray = dm_malloc(sizeof(lvm_response_t) * num_responses +
+				 sizeof(int) * 2))) {
+		errno = ENOMEM;
+		status = 0;
+		goto out;
+	}
+
+	/* 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);
+			dm_free(rarray);
+			errno = ENOMEM;
+			status = 0;
+			goto out;
+		}
+
+		strcpy(rarray[i].response, inptr);
+		rarray[i].len = strlen(inptr);
+		inptr += strlen(inptr) + 1;
+		i++;
+	}
+	*num = num_responses;
+	*response = rarray;
+
+      out:
+	dm_free(retbuf);
+
+	return status;
+}
+
+/* Free reply array */
+static int _cluster_free_request(lvm_response_t * response, int num)
+{
+	int i;
+
+	for (i = 0; i < num; i++) {
+		dm_free(response[i].response);
+	}
+
+	dm_free(response);
+
+	return 1;
+}
+
+int refresh_clvmd(int all_nodes)
+{
+	int num_responses;
+	char args[1]; // No args really.
+	lvm_response_t *response = NULL;
+	int saved_errno;
+	int status;
+	int i;
+
+	status = _cluster_request(CLVMD_CMD_REFRESH, all_nodes ? NODE_ALL : NODE_LOCAL, args, 0, &response, &num_responses, 0);
+
+	/* 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, num_responses);
+	errno = saved_errno;
+
+	return status;
+}
+
+int restart_clvmd(int all_nodes)
+{
+	int dummy, status;
+
+	status = _cluster_request(CLVMD_CMD_RESTART, all_nodes ? NODE_ALL : NODE_LOCAL, NULL, 0, NULL, &dummy, 1);
+
+	/*
+	 * FIXME: we cannot receive response, clvmd re-exec before it.
+	 *        but also should not close socket too early (the whole rq is dropped then).
+	 * FIXME: This should be handled this way:
+	 *  - client waits for RESTART ack (and socket close)
+	 *  - server restarts
+	 *  - client checks that server is ready again (VERSION command?)
+	 */
+	usleep(500000);
+
+	return status;
+}
+
+int debug_clvmd(int level, int clusterwide)
+{
+	int num_responses;
+	char args[1];
+	const char *nodes;
+	lvm_response_t *response = NULL;
+	int saved_errno;
+	int status;
+	int i;
+
+	args[0] = level;
+	if (clusterwide)
+		nodes = NODE_ALL;
+	else
+		nodes = NODE_LOCAL;
+
+	status = _cluster_request(CLVMD_CMD_SET_DEBUG, nodes, args, 1, &response, &num_responses, 0);
+
+	/* 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 setting debug on 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, num_responses);
+	errno = saved_errno;
+
+	return status;
+}
--- a/daemons/clvmd/refresh_clvmd.h
+++ b/daemons/clvmd/refresh_clvmd.h
@@ -0,0 +1,19 @@
+/*
+ * Copyright (C) 2007 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+
+int refresh_clvmd(int all_nodes);
+int restart_clvmd(int all_nodes);
+int debug_clvmd(int level, int clusterwide);
+
--- a/daemons/cmirrord/Makefile.in
+++ b/daemons/cmirrord/Makefile.in
@@ -17,6 +17,8 @@ top_builddir = @top_builddir@

 CPG_LIBS = @CPG_LIBS@
 CPG_CFLAGS = @CPG_CFLAGS@
+SACKPT_LIBS = @SACKPT_LIBS@
+SACKPT_CFLAGS = @SACKPT_CFLAGS@

 SOURCES = clogd.c cluster.c compat.c functions.c link_mon.c local.c logging.c

@@ -24,15 +26,14 @@ TARGETS = cmirrord

 include $(top_builddir)/make.tmpl

-LMLIBS += $(CPG_LIBS)
-CFLAGS += $(CPG_CFLAGS) $(EXTRA_EXEC_CFLAGS)
+LIBS += -ldevmapper
+LMLIBS += $(CPG_LIBS) $(SACKPT_LIBS)
+CFLAGS += $(CPG_CFLAGS) $(SACKPT_CFLAGS) $(EXTRA_EXEC_CFLAGS)
 LDFLAGS += $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)

-cmirrord: $(OBJECTS)
-	@echo "    [CC] $@"
-	$(Q) $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) \
-		$(LMLIBS) -L$(top_builddir)/libdm -ldevmapper $(LIBS)
+cmirrord: $(OBJECTS) $(top_builddir)/lib/liblvm-internal.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) \
+		$(LVMLIBS) $(LMLIBS) $(LIBS)

 install: $(TARGETS)
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_PROGRAM) -D cmirrord $(usrsbindir)/cmirrord
+	$(INSTALL_PROGRAM) -D cmirrord $(usrsbindir)/cmirrord
--- a/daemons/cmirrord/cluster.c
+++ b/daemons/cmirrord/cluster.c
@@ -16,11 +16,7 @@
 #include "functions.h"
 #include "link_mon.h"
 #include "local.h"
-#include "lib/mm/xlate.h"
-#include "base/memory/zalloc.h"
-
-/* FIXME: remove this and the code */
-#define CMIRROR_HAS_CHECKPOINT 0
+#include "xlate.h"

 #include <corosync/cpg.h>
 #include <errno.h>
@@ -170,9 +166,6 @@ int cluster_send(struct clog_request *rq)
 {
 	int r;
 	int found = 0;
-#if CMIRROR_HAS_CHECKPOINT
-	int count = 0;
-#endif
 	struct iovec iov;
 	struct clog_cpg *entry;

@@ -210,6 +203,8 @@ int cluster_send(struct clog_request *rq)

 #if CMIRROR_HAS_CHECKPOINT
 	do {
+		int count = 0;
+
 		r = cpg_mcast_joined(entry->handle, CPG_TYPE_AGREED, &iov, 1);
 		if (r != SA_AIS_ERR_TRY_AGAIN)
 			break;
@@ -403,12 +398,13 @@ static struct checkpoint_data *prepare_checkpoint(struct clog_cpg *entry,
 		return NULL;
 	}

-	new = zalloc(sizeof(*new));
+	new = malloc(sizeof(*new));
 	if (!new) {
 		LOG_ERROR("Unable to create checkpoint data for %u",
 			  cp_requester);
 		return NULL;
 	}
+	memset(new, 0, sizeof(*new));
 	new->requester = cp_requester;
 	strncpy(new->uuid, entry->name.value, entry->name.length);

@@ -643,12 +639,13 @@ static int export_checkpoint(struct checkpoint_data *cp)
 	rq_size += RECOVERING_REGION_SECTION_SIZE;
 	rq_size += cp->bitmap_size * 2; /* clean|sync_bits */

-	rq = zalloc(rq_size);
+	rq = malloc(rq_size);
 	if (!rq) {
 		LOG_ERROR("export_checkpoint: "
 			  "Unable to allocate transfer structs");
 		return -ENOMEM;
 	}
+	memset(rq, 0, rq_size);

 	dm_list_init(&rq->u.list);
 	rq->u_rq.request_type = DM_ULOG_CHECKPOINT_READY;
@@ -1620,11 +1617,12 @@ int create_cluster_cpg(char *uuid, uint64_t luid)
 			return -EEXIST;
 		}

-	new = zalloc(sizeof(*new));
+	new = malloc(sizeof(*new));
 	if (!new) {
 		LOG_ERROR("Unable to allocate memory for clog_cpg");
 		return -ENOMEM;
 	}
+	memset(new, 0, sizeof(*new));
 	dm_list_init(&new->list);
 	new->lowest_id = 0xDEAD;
 	dm_list_init(&new->startup_list);
@@ -1632,7 +1630,7 @@ int create_cluster_cpg(char *uuid, uint64_t luid)

 	size = ((strlen(uuid) + 1) > CPG_MAX_NAME_LENGTH) ?
 		CPG_MAX_NAME_LENGTH : (strlen(uuid) + 1);
-	(void) dm_strncpy(new->name.value, uuid, size);
+	strncpy(new->name.value, uuid, size);
 	new->name.length = (uint32_t)size;
 	new->luid = luid;

--- a/daemons/cmirrord/cluster.h
+++ b/daemons/cmirrord/cluster.h
@@ -12,8 +12,8 @@
 #ifndef _LVM_CLOG_CLUSTER_H
 #define _LVM_CLOG_CLUSTER_H

-#include "libdm/misc/dm-log-userspace.h"
-#include "libdm/libdevmapper.h"
+#include "dm-log-userspace.h"
+#include "libdevmapper.h"

 #define DM_ULOG_RESPONSE 0x1000U /* in last byte of 32-bit value */
 #define DM_ULOG_CHECKPOINT_READY 21
--- a/daemons/cmirrord/compat.c
+++ b/daemons/cmirrord/compat.c
@@ -8,7 +8,7 @@
 #include "logging.h"
 #include "cluster.h"
 #include "compat.h"
-#include "lib/mm/xlate.h"
+#include "xlate.h"

 #include <errno.h>

--- a/daemons/cmirrord/functions.c
+++ b/daemons/cmirrord/functions.c
@@ -11,7 +11,6 @@
 */
 #include "logging.h"
 #include "functions.h"
-#include "base/memory/zalloc.h"

 #include <sys/sysmacros.h>
 #include <dirent.h>
@@ -436,7 +435,7 @@ static int _clog_ctr(char *uuid, uint64_t luid,
 			block_on_error = 1;
 	}

-	lc = zalloc(sizeof(*lc));
+	lc = dm_zalloc(sizeof(*lc));
 	if (!lc) {
 		LOG_ERROR("Unable to allocate cluster log context");
 		r = -ENOMEM;
@@ -452,19 +451,15 @@ static int _clog_ctr(char *uuid, uint64_t luid,
 	lc->skip_bit_warning = region_count;
 	lc->disk_fd = -1;
 	lc->log_dev_failed = 0;
-	if (!dm_strncpy(lc->uuid, uuid, DM_UUID_LEN)) {
-		LOG_ERROR("Cannot use too long UUID %s.", uuid);
-		r = -EINVAL;
-		goto fail;
-	}
+	strncpy(lc->uuid, uuid, DM_UUID_LEN);
 	lc->luid = luid;

 	if (get_log(lc->uuid, lc->luid) ||
 	    get_pending_log(lc->uuid, lc->luid)) {
 		LOG_ERROR("[%s/%" PRIu64 "u] Log already exists, unable to create.",
 			  SHORT_UUID(lc->uuid), lc->luid);
-		r = -EINVAL;
-		goto fail;
+		dm_free(lc);
+		return -EINVAL;
 	}

 	dm_list_init(&lc->mark_list);
@@ -533,9 +528,9 @@ fail:
 			LOG_ERROR("Close device error, %s: %s",
 				  disk_path, strerror(errno));
 		free(lc->disk_buffer);
-		free(lc->sync_bits);
-		free(lc->clean_bits);
-		free(lc);
+		dm_free(lc->sync_bits);
+		dm_free(lc->clean_bits);
+		dm_free(lc);
 	}
 	return r;
 }
@@ -660,9 +655,9 @@ static int clog_dtr(struct dm_ulog_request *rq)
 			  strerror(errno));
 	if (lc->disk_buffer)
 		free(lc->disk_buffer);
-	free(lc->clean_bits);
-	free(lc->sync_bits);
-	free(lc);
+	dm_free(lc->clean_bits);
+	dm_free(lc->sync_bits);
+	dm_free(lc);

 	return 0;
 }
--- a/daemons/cmirrord/functions.h
+++ b/daemons/cmirrord/functions.h
@@ -12,7 +12,7 @@
 #ifndef _LVM_CLOG_FUNCTIONS_H
 #define _LVM_CLOG_FUNCTIONS_H

-#include "device_mapper/misc/dm-log-userspace.h"
+#include "dm-log-userspace.h"
 #include "cluster.h"

 #define LOG_RESUMED   1
--- a/daemons/cmirrord/logging.h
+++ b/daemons/cmirrord/logging.h
@@ -13,6 +13,10 @@
 #ifndef _LVM_CLOG_LOGGING_H
 #define _LVM_CLOG_LOGGING_H

+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include "configure.h"
 #include <stdio.h>
 #include <stdint.h>
 #include <syslog.h>
--- a/daemons/dmeventd/Makefile.in
+++ b/daemons/dmeventd/Makefile.in
@@ -57,16 +57,14 @@ all: device-mapper
 device-mapper: $(TARGETS)

 CFLAGS_dmeventd.o += $(EXTRA_EXEC_CFLAGS)
-LIBS += $(PTHREAD_LIBS) -L$(top_builddir)/libdm -ldevmapper
+LIBS += -ldevmapper $(PTHREAD_LIBS)

 dmeventd: $(LIB_SHARED) dmeventd.o
-	@echo "    [CC] $@"
-	$(Q) $(CC) $(CFLAGS) -L. $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) dmeventd.o \
-		-o $@ $(DL_LIBS) $(DMEVENT_LIBS) $(LIBS) -lm
+	$(CC) $(CFLAGS) -L. $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) dmeventd.o \
+		-o $@ $(DL_LIBS) $(DMEVENT_LIBS) $(LIBS)

-dmeventd.static: $(LIB_STATIC) dmeventd.o
-	@echo "    [CC] $@"
-	$(Q) $(CC) $(CFLAGS) $(LDFLAGS) -static -L. -L$(interfacebuilddir) dmeventd.o \
+dmeventd.static: $(LIB_STATIC) dmeventd.o $(interfacebuilddir)/libdevmapper.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -static -L. -L$(interfacebuilddir) dmeventd.o \
 		-o $@ $(DL_LIBS) $(DMEVENT_LIBS) $(LIBS) $(STATIC_LIBS)

 ifeq ("@PKGCONFIG@", "yes")
@@ -75,6 +73,7 @@ endif

 ifneq ("$(CFLOW_CMD)", "")
 CFLOW_SOURCES = $(addprefix $(srcdir)/, $(SOURCES))
+-include $(top_builddir)/libdm/libdevmapper.cflow
 -include $(top_builddir)/lib/liblvm-internal.cflow
 -include $(top_builddir)/lib/liblvm2cmd.cflow
 -include $(top_builddir)/daemons/dmeventd/$(LIB_NAME).cflow
@@ -82,28 +81,23 @@ CFLOW_SOURCES = $(addprefix $(srcdir)/, $(SOURCES))
 endif

 install_include: $(srcdir)/libdevmapper-event.h
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_DATA) -D $< $(includedir)/$(<F)
+	$(INSTALL_DATA) -D $< $(includedir)/$(<F)

 install_pkgconfig: libdevmapper-event.pc
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_DATA) -D $< $(pkgconfigdir)/devmapper-event.pc
+	$(INSTALL_DATA) -D $< $(pkgconfigdir)/devmapper-event.pc

 install_lib_dynamic: install_lib_shared

 install_lib_static: $(LIB_STATIC)
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_DATA) -D $< $(usrlibdir)/$(<F)
+	$(INSTALL_DATA) -D $< $(usrlibdir)/$(<F)

 install_lib: $(INSTALL_LIB_TARGETS)

 install_dmeventd_dynamic: dmeventd
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
+	$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)

 install_dmeventd_static: dmeventd.static
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_PROGRAM) -D $< $(staticdir)/$(<F)
+	$(INSTALL_PROGRAM) -D $< $(staticdir)/$(<F)

 install_dmeventd: $(INSTALL_DMEVENTD_TARGETS)

--- a/daemons/dmeventd/dmeventd.c
+++ b/daemons/dmeventd/dmeventd.c
@@ -16,12 +16,12 @@
 * dmeventd - dm event daemon to monitor active mapped devices
 */

+#include "dm-logging.h"

 #include "libdevmapper-event.h"
 #include "dmeventd.h"

-#include "libdm/misc/dm-logging.h"
-#include "base/memory/zalloc.h"
+#include "tool.h"

 #include <dlfcn.h>
 #include <pthread.h>
@@ -33,8 +33,6 @@
 #include <signal.h>
 #include <arpa/inet.h>		/* for htonl, ntohl */
 #include <fcntl.h>		/* for musl libc */
-#include <unistd.h>
-#include <syslog.h>

 #ifdef __linux__
 /*
@@ -62,6 +60,8 @@

 #endif

+#include <syslog.h>
+
 #define DM_SIGNALED_EXIT  1
 #define DM_SCHEDULED_EXIT 2
 static volatile sig_atomic_t _exit_now = 0;	/* set to '1' when signal is given to exit */
@@ -264,19 +264,19 @@ static pthread_cond_t _timeout_cond = PTHREAD_COND_INITIALIZER;
 /* DSO data allocate/free. */
 static void _free_dso_data(struct dso_data *data)
 {
-	free(data->dso_name);
-	free(data);
+	dm_free(data->dso_name);
+	dm_free(data);
 }

 static struct dso_data *_alloc_dso_data(struct message_data *data)
 {
-	struct dso_data *ret = (typeof(ret)) zalloc(sizeof(*ret));
+	struct dso_data *ret = (typeof(ret)) dm_zalloc(sizeof(*ret));

 	if (!ret)
 		return_NULL;

-	if (!(ret->dso_name = strdup(data->dso_name))) {
-		free(ret);
+	if (!(ret->dso_name = dm_strdup(data->dso_name))) {
+		dm_free(ret);
 		return_NULL;
 	}

@@ -397,9 +397,9 @@ static void _free_thread_status(struct thread_status *thread)
 	_lib_put(thread->dso_data);
 	if (thread->wait_task)
 		dm_task_destroy(thread->wait_task);
-	free(thread->device.uuid);
-	free(thread->device.name);
-	free(thread);
+	dm_free(thread->device.uuid);
+	dm_free(thread->device.name);
+	dm_free(thread);
 }

 /* Note: events_field must not be 0, ensured by caller */
@@ -408,7 +408,7 @@ static struct thread_status *_alloc_thread_status(const struct message_data *dat
 {
 	struct thread_status *thread;

-	if (!(thread = zalloc(sizeof(*thread)))) {
+	if (!(thread = dm_zalloc(sizeof(*thread)))) {
 		log_error("Cannot create new thread, out of memory.");
 		return NULL;
 	}
@@ -422,11 +422,11 @@ static struct thread_status *_alloc_thread_status(const struct message_data *dat
 	if (!dm_task_set_uuid(thread->wait_task, data->device_uuid))
 		goto_out;

-	if (!(thread->device.uuid = strdup(data->device_uuid)))
+	if (!(thread->device.uuid = dm_strdup(data->device_uuid)))
 		goto_out;

 	/* Until real name resolved, use UUID */
-	if (!(thread->device.name = strdup(data->device_uuid)))
+	if (!(thread->device.name = dm_strdup(data->device_uuid)))
 		goto_out;

 	/* runs ioctl and may register lvm2 pluging */
@@ -515,7 +515,7 @@ static int _fetch_string(char **ptr, char **src, const int delimiter)
 	if ((p = strchr(*src, delimiter))) {
 		if (*src < p) {
 			*p = 0; /* Temporary exit with \0 */
-			if (!(*ptr = strdup(*src))) {
+			if (!(*ptr = dm_strdup(*src))) {
 				log_error("Failed to fetch item %s.", *src);
 				ret = 0; /* Allocation fail */
 			}
@@ -525,7 +525,7 @@ static int _fetch_string(char **ptr, char **src, const int delimiter)
 		(*src)++; /* Skip delmiter, next field */
 	} else if ((len = strlen(*src))) {
 		/* No delimiter, item ends with '\0' */
-		if (!(*ptr = strdup(*src))) {
+		if (!(*ptr = dm_strdup(*src))) {
 			log_error("Failed to fetch last item %s.", *src);
 			ret = 0; /* Fail */
 		}
@@ -538,11 +538,11 @@ out:
 /* Free message memory. */
 static void _free_message(struct message_data *message_data)
 {
-	free(message_data->id);
-	free(message_data->dso_name);
-	free(message_data->device_uuid);
-	free(message_data->events_str);
-	free(message_data->timeout_str);
+	dm_free(message_data->id);
+	dm_free(message_data->dso_name);
+	dm_free(message_data->device_uuid);
+	dm_free(message_data->events_str);
+	dm_free(message_data->timeout_str);
 }

 /* Parse a register message from the client. */
@@ -574,7 +574,7 @@ static int _parse_message(struct message_data *message_data)
 		ret = 1;
 	}

-	free(msg->data);
+	dm_free(msg->data);
 	msg->data = NULL;

 	return ret;
@@ -608,8 +608,8 @@ static int _fill_device_data(struct thread_status *ts)
 	if (!dm_task_run(dmt))
 		goto fail;

-	free(ts->device.name);
-	if (!(ts->device.name = strdup(dm_task_get_name(dmt))))
+	dm_free(ts->device.name);
+	if (!(ts->device.name = dm_strdup(dm_task_get_name(dmt))))
 		goto fail;

 	if (!dm_task_get_info(dmt, &dmi))
@@ -696,8 +696,8 @@ static int _get_status(struct message_data *message_data)

 	len = strlen(message_data->id);
 	msg->size = size + len + 1;
-	free(msg->data);
-	if (!(msg->data = malloc(msg->size)))
+	dm_free(msg->data);
+	if (!(msg->data = dm_malloc(msg->size)))
 		goto out;

 	memcpy(msg->data, message_data->id, len);
@@ -712,7 +712,7 @@ static int _get_status(struct message_data *message_data)
 	ret = 0;
 out:
 	for (j = 0; j < i; ++j)
-		free(buffers[j]);
+		dm_free(buffers[j]);

 	return ret;
 }
@@ -721,7 +721,7 @@ static int _get_parameters(struct message_data *message_data) {
 	struct dm_event_daemon_message *msg = message_data->msg;
 	int size;

-	free(msg->data);
+	dm_free(msg->data);
 	if ((size = dm_asprintf(&msg->data, "%s pid=%d daemon=%s exec_method=%s",
 				message_data->id, getpid(),
 				_foreground ? "no" : "yes",
@@ -754,7 +754,6 @@ static void *_timeout_thread(void *unused __attribute__((unused)))
 	struct thread_status *thread;
 	struct timespec timeout;
 	time_t curr_time;
-	int ret;

 	DEBUGLOG("Timeout thread starting.");
 	pthread_cleanup_push(_exit_timeout, NULL);
@@ -776,10 +775,7 @@ static void *_timeout_thread(void *unused __attribute__((unused)))
 				} else {
 					DEBUGLOG("Sending SIGALRM to Thr %x for timeout.",
 						 (int) thread->thread);
-					ret = pthread_kill(thread->thread, SIGALRM);
-					if (ret && (ret != ESRCH))
-						log_error("Unable to wakeup Thr %x for timeout: %s.",
-							  (int) thread->thread, strerror(ret));
+					pthread_kill(thread->thread, SIGALRM);
 				}
 				_unlock_mutex();
 			}
@@ -869,7 +865,6 @@ static int _event_wait(struct thread_status *thread)
 	 * This is so that you can break out of waiting on an event,
 	 * either for a timeout event, or to cancel the thread.
 	 */
-	sigemptyset(&old);
 	sigemptyset(&set);
 	sigaddset(&set, SIGALRM);
 	if (pthread_sigmask(SIG_UNBLOCK, &set, &old) != 0) {
@@ -1225,7 +1220,7 @@ static int _registered_device(struct message_data *message_data,
 	int r;
 	struct dm_event_daemon_message *msg = message_data->msg;

-	free(msg->data);
+	dm_free(msg->data);

 	if ((r = dm_asprintf(&(msg->data), "%s %s %s %u",
 			     message_data->id,
@@ -1365,7 +1360,7 @@ static int _get_timeout(struct message_data *message_data)
 	if (!thread)
 		return -ENODEV;

-	free(msg->data);
+	dm_free(msg->data);
 	msg->size = dm_asprintf(&(msg->data), "%s %" PRIu32,
 				message_data->id, thread->timeout);

@@ -1502,7 +1497,7 @@ static int _client_read(struct dm_event_fifos *fifos,
 			bytes = 0;
 			if (!size)
 				break; /* No data -> error */
-			buf = msg->data = malloc(msg->size);
+			buf = msg->data = dm_malloc(msg->size);
 			if (!buf)
 				break; /* No mem -> error */
 			header = 0;
@@ -1510,7 +1505,7 @@ static int _client_read(struct dm_event_fifos *fifos,
 	}

 	if (bytes != size) {
-		free(msg->data);
+		dm_free(msg->data);
 		msg->data = NULL;
 		return 0;
 	}
@@ -1530,7 +1525,7 @@ static int _client_write(struct dm_event_fifos *fifos,
 	fd_set fds;

 	size_t size = 2 * sizeof(uint32_t) + ((msg->data) ? msg->size : 0);
-	uint32_t *header = malloc(size);
+	uint32_t *header = dm_malloc(size);
 	char *buf = (char *)header;

 	if (!header) {
@@ -1560,7 +1555,7 @@ static int _client_write(struct dm_event_fifos *fifos,
 	}

 	if (header != temp)
-		free(header);
+		dm_free(header);

 	return (bytes == size);
 }
@@ -1622,7 +1617,7 @@ static int _do_process_request(struct dm_event_daemon_message *msg)
 			msg->size = dm_asprintf(&(msg->data), "%s %s %d", answer,
 						(msg->cmd == DM_EVENT_CMD_DIE) ? "DYING" : "HELLO",
 						DM_EVENT_PROTOCOL_VERSION);
-			free(answer);
+			dm_free(answer);
 		}
 	} else if (msg->cmd != DM_EVENT_CMD_ACTIVE && !_parse_message(&message_data)) {
 		stack;
@@ -1664,7 +1659,7 @@ static void _process_request(struct dm_event_fifos *fifos)

 	DEBUGLOG("<<< CMD:%s (0x%x) completed (result %d).", decode_cmd(cmd), cmd, msg.cmd);

-	free(msg.data);
+	dm_free(msg.data);

 	if (cmd == DM_EVENT_CMD_DIE) {
 		if (unlink(DMEVENTD_PIDFILE))
@@ -1975,7 +1970,7 @@ static int _reinstate_registrations(struct dm_event_fifos *fifos)
 	int i, ret;

 	ret = daemon_talk(fifos, &msg, DM_EVENT_CMD_HELLO, NULL, NULL, 0, 0);
-	free(msg.data);
+	dm_free(msg.data);
 	msg.data = NULL;

 	if (ret) {
@@ -2061,13 +2056,13 @@ static void _restart_dmeventd(void)
 			++count;
 		}

-	if (!(_initial_registrations = malloc(sizeof(char*) * (count + 1)))) {
+	if (!(_initial_registrations = dm_malloc(sizeof(char*) * (count + 1)))) {
 		fprintf(stderr, "Memory allocation registration failed.\n");
 		goto bad;
 	}

 	for (i = 0; i < count; ++i) {
-		if (!(_initial_registrations[i] = strdup(message))) {
+		if (!(_initial_registrations[i] = dm_strdup(message))) {
 			fprintf(stderr, "Memory allocation for message failed.\n");
 			goto bad;
 		}
--- a/daemons/dmeventd/libdevmapper-event.c
+++ b/daemons/dmeventd/libdevmapper-event.c
@@ -12,12 +12,10 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

+#include "dm-logging.h"
+#include "dmlib.h"
 #include "libdevmapper-event.h"
 #include "dmeventd.h"
-#include "libdm/misc/dm-logging.h"
-#include "base/memory/zalloc.h"
-
-#include "lib/misc/intl.h"

 #include <fcntl.h>
 #include <sys/file.h>
@@ -27,7 +25,6 @@
 #include <arpa/inet.h>		/* for htonl, ntohl */
 #include <pthread.h>
 #include <syslog.h>
-#include <unistd.h>

 static int _debug_level = 0;
 static int _use_syslog = 0;
@@ -50,8 +47,8 @@ struct dm_event_handler {

 static void _dm_event_handler_clear_dev_info(struct dm_event_handler *dmevh)
 {
-	free(dmevh->dev_name);
-	free(dmevh->uuid);
+	dm_free(dmevh->dev_name);
+	dm_free(dmevh->uuid);
 	dmevh->dev_name = dmevh->uuid = NULL;
 	dmevh->major = dmevh->minor = 0;
 }
@@ -60,7 +57,7 @@ struct dm_event_handler *dm_event_handler_create(void)
 {
 	struct dm_event_handler *dmevh;

-	if (!(dmevh = zalloc(sizeof(*dmevh)))) {
+	if (!(dmevh = dm_zalloc(sizeof(*dmevh)))) {
 		log_error("Failed to allocate event handler.");
 		return NULL;
 	}
@@ -71,9 +68,9 @@ struct dm_event_handler *dm_event_handler_create(void)
 void dm_event_handler_destroy(struct dm_event_handler *dmevh)
 {
 	_dm_event_handler_clear_dev_info(dmevh);
-	free(dmevh->dso);
-	free(dmevh->dmeventd_path);
-	free(dmevh);
+	dm_free(dmevh->dso);
+	dm_free(dmevh->dmeventd_path);
+	dm_free(dmevh);
 }

 int dm_event_handler_set_dmeventd_path(struct dm_event_handler *dmevh, const char *dmeventd_path)
@@ -81,9 +78,9 @@ int dm_event_handler_set_dmeventd_path(struct dm_event_handler *dmevh, const cha
 	if (!dmeventd_path) /* noop */
 		return 0;

-	free(dmevh->dmeventd_path);
+	dm_free(dmevh->dmeventd_path);

-	if (!(dmevh->dmeventd_path = strdup(dmeventd_path)))
+	if (!(dmevh->dmeventd_path = dm_strdup(dmeventd_path)))
 		return -ENOMEM;

 	return 0;
@@ -94,9 +91,9 @@ int dm_event_handler_set_dso(struct dm_event_handler *dmevh, const char *path)
 	if (!path) /* noop */
 		return 0;

-	free(dmevh->dso);
+	dm_free(dmevh->dso);

-	if (!(dmevh->dso = strdup(path)))
+	if (!(dmevh->dso = dm_strdup(path)))
 		return -ENOMEM;

 	return 0;
@@ -109,7 +106,7 @@ int dm_event_handler_set_dev_name(struct dm_event_handler *dmevh, const char *de

 	_dm_event_handler_clear_dev_info(dmevh);

-	if (!(dmevh->dev_name = strdup(dev_name)))
+	if (!(dmevh->dev_name = dm_strdup(dev_name)))
 		return -ENOMEM;

 	return 0;
@@ -122,7 +119,7 @@ int dm_event_handler_set_uuid(struct dm_event_handler *dmevh, const char *uuid)

 	_dm_event_handler_clear_dev_info(dmevh);

-	if (!(dmevh->uuid = strdup(uuid)))
+	if (!(dmevh->uuid = dm_strdup(uuid)))
 		return -ENOMEM;

 	return 0;
@@ -262,7 +259,7 @@ static int _daemon_read(struct dm_event_fifos *fifos,
 		if (header && (bytes == 2 * sizeof(uint32_t))) {
 			msg->cmd = ntohl(header[0]);
 			msg->size = ntohl(header[1]);
-			buf = msg->data = malloc(msg->size);
+			buf = msg->data = dm_malloc(msg->size);
 			size = msg->size;
 			bytes = 0;
 			header = 0;
@@ -270,7 +267,7 @@ static int _daemon_read(struct dm_event_fifos *fifos,
 	}

 	if (bytes != size) {
-		free(msg->data);
+		dm_free(msg->data);
 		msg->data = NULL;
 	}
 	return bytes == size;
@@ -373,13 +370,13 @@ int daemon_talk(struct dm_event_fifos *fifos,
 	 */
 	if (!_daemon_write(fifos, msg)) {
 		stack;
-		free(msg->data);
+		dm_free(msg->data);
 		msg->data = NULL;
 		return -EIO;
 	}

 	do {
-		free(msg->data);
+		dm_free(msg->data);
 		msg->data = NULL;

 		if (!_daemon_read(fifos, msg)) {
@@ -622,7 +619,7 @@ static int _do_event(int cmd, char *dmeventd_path, struct dm_event_daemon_messag

 	ret = daemon_talk(&fifos, msg, DM_EVENT_CMD_HELLO, NULL, NULL, 0, 0);

-	free(msg->data);
+	dm_free(msg->data);
 	msg->data = 0;

 	if (!ret)
@@ -648,7 +645,6 @@ int dm_event_register_handler(const struct dm_event_handler *dmevh)
 	uuid = dm_task_get_uuid(dmt);

 	if (!strstr(dmevh->dso, "libdevmapper-event-lvm2thin.so") &&
-	    !strstr(dmevh->dso, "libdevmapper-event-lvm2vdo.so") &&
 	    !strstr(dmevh->dso, "libdevmapper-event-lvm2snapshot.so") &&
 	    !strstr(dmevh->dso, "libdevmapper-event-lvm2mirror.so") &&
 	    !strstr(dmevh->dso, "libdevmapper-event-lvm2raid.so"))
@@ -663,7 +659,7 @@ int dm_event_register_handler(const struct dm_event_handler *dmevh)
 		ret = 0;
 	}

-	free(msg.data);
+	dm_free(msg.data);

 	dm_task_destroy(dmt);

@@ -690,7 +686,7 @@ int dm_event_unregister_handler(const struct dm_event_handler *dmevh)
 		ret = 0;
 	}

-	free(msg.data);
+	dm_free(msg.data);

 	dm_task_destroy(dmt);

@@ -706,7 +702,7 @@ static char *_fetch_string(char **src, const int delimiter)
 	if ((p = strchr(*src, delimiter)))
 		*p = 0;

-	if ((ret = strdup(*src)))
+	if ((ret = dm_strdup(*src)))
 		*src += strlen(ret) + 1;

 	if (p)
@@ -726,11 +722,11 @@ static int _parse_message(struct dm_event_daemon_message *msg, char **dso_name,
 	    (*dso_name = _fetch_string(&p, ' ')) &&
 	    (*uuid = _fetch_string(&p, ' '))) {
 		*evmask = atoi(p);
-		free(id);
+		dm_free(id);
 		return 0;
 	}

-	free(id);
+	dm_free(id);
 	return -ENOMEM;
 }

@@ -758,10 +754,11 @@ int dm_event_get_registered_device(struct dm_event_handler *dmevh, int next)
 	uuid = dm_task_get_uuid(dmt);

 	/* FIXME Distinguish errors connecting to daemon */
-	if ((ret = _do_event(next ? DM_EVENT_CMD_GET_NEXT_REGISTERED_DEVICE :
-			    DM_EVENT_CMD_GET_REGISTERED_DEVICE, dmevh->dmeventd_path,
-			    &msg, dmevh->dso, uuid, dmevh->mask, 0))) {
+	if (_do_event(next ? DM_EVENT_CMD_GET_NEXT_REGISTERED_DEVICE :
+		      DM_EVENT_CMD_GET_REGISTERED_DEVICE, dmevh->dmeventd_path,
+		      &msg, dmevh->dso, uuid, dmevh->mask, 0)) {
 		log_debug("%s: device not registered.", dm_task_get_name(dmt));
+		ret = -ENOENT;
 		goto fail;
 	}

@@ -772,7 +769,7 @@ int dm_event_get_registered_device(struct dm_event_handler *dmevh, int next)
 	dm_task_destroy(dmt);
 	dmt = NULL;

-	free(msg.data);
+	dm_free(msg.data);
 	msg.data = NULL;

 	_dm_event_handler_clear_dev_info(dmevh);
@@ -781,7 +778,7 @@ int dm_event_get_registered_device(struct dm_event_handler *dmevh, int next)
 		goto fail;
 	}

-	if (!(dmevh->uuid = strdup(reply_uuid))) {
+	if (!(dmevh->uuid = dm_strdup(reply_uuid))) {
 		ret = -ENOMEM;
 		goto fail;
 	}
@@ -794,13 +791,13 @@ int dm_event_get_registered_device(struct dm_event_handler *dmevh, int next)
 	dm_event_handler_set_dso(dmevh, reply_dso);
 	dm_event_handler_set_event_mask(dmevh, reply_mask);

-	free(reply_dso);
+	dm_free(reply_dso);
 	reply_dso = NULL;

-	free(reply_uuid);
+	dm_free(reply_uuid);
 	reply_uuid = NULL;

-	if (!(dmevh->dev_name = strdup(dm_task_get_name(dmt)))) {
+	if (!(dmevh->dev_name = dm_strdup(dm_task_get_name(dmt)))) {
 		ret = -ENOMEM;
 		goto fail;
 	}
@@ -818,9 +815,9 @@ int dm_event_get_registered_device(struct dm_event_handler *dmevh, int next)
 	return ret;

 fail:
-	free(msg.data);
-	free(reply_dso);
-	free(reply_uuid);
+	dm_free(msg.data);
+	dm_free(reply_dso);
+	dm_free(reply_uuid);
 	_dm_event_handler_clear_dev_info(dmevh);
 	if (dmt)
 		dm_task_destroy(dmt);
@@ -985,12 +982,12 @@ int dm_event_get_timeout(const char *device_path, uint32_t *timeout)
 		if (!p) {
 			log_error("Malformed reply from dmeventd '%s'.",
 				  msg.data);
-			free(msg.data);
+			dm_free(msg.data);
 			return -EIO;
 		}
 		*timeout = atoi(p);
 	}
-	free(msg.data);
+	dm_free(msg.data);

 	return ret;
 }
--- a/daemons/dmeventd/libdevmapper-event.h
+++ b/daemons/dmeventd/libdevmapper-event.h
@@ -21,7 +21,6 @@
 #ifndef LIB_DMEVENT_H
 #define LIB_DMEVENT_H

-#include <stdarg.h>
 #include <stdint.h>

 /*
--- a/daemons/dmeventd/libdevmapper-event.pc.in
+++ b/daemons/dmeventd/libdevmapper-event.pc.in
@@ -8,3 +8,4 @@ Description: device-mapper event library
 Version: @DM_LIB_PATCHLEVEL@
 Cflags: -I${includedir}
 Libs: -L${libdir} -ldevmapper-event
+Requires.private: devmapper
--- a/daemons/dmeventd/plugins/Makefile.in
+++ b/daemons/dmeventd/plugins/Makefile.in
@@ -1,6 +1,6 @@
 #
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
-# Copyright (C) 2004-2018 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2005, 2011 Red Hat, Inc. All rights reserved.
 #
 # This file is part of LVM2.
 #
@@ -16,7 +16,27 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

-SUBDIRS += lvm2 snapshot raid thin mirror vdo
+SUBDIRS += lvm2
+
+ifneq ("@MIRRORS@", "none")
+  SUBDIRS += mirror
+endif
+
+ifneq ("@SNAPSHOTS@", "none")
+  SUBDIRS += snapshot
+endif
+
+ifneq ("@RAID@", "none")
+  SUBDIRS += raid
+endif
+
+ifneq ("@THIN@", "none")
+  SUBDIRS += thin
+endif
+
+ifeq ($(MAKECMDGOALS),distclean)
+  SUBDIRS = lvm2 mirror snapshot raid thin
+endif

 include $(top_builddir)/make.tmpl

@@ -24,4 +44,3 @@ snapshot: lvm2
 mirror: lvm2
 raid: lvm2
 thin: lvm2
-vdo: lvm2
--- a/daemons/dmeventd/plugins/lvm2/Makefile.in
+++ b/daemons/dmeventd/plugins/lvm2/Makefile.in
@@ -16,7 +16,6 @@ top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

 CLDFLAGS += -L$(top_builddir)/tools
-LIBS += $(DMEVENT_LIBS) $(PTHREAD_LIBS) @LVM2CMD_LIB@

 SOURCES = dmeventd_lvm.c

@@ -25,6 +24,8 @@ LIB_VERSION = $(LIB_VERSION_LVM)

 include $(top_builddir)/make.tmpl

+LIBS += @LVM2CMD_LIB@ -ldevmapper $(PTHREAD_LIBS)
+
 install_lvm2: install_lib_shared

 install: install_lvm2
--- a/daemons/dmeventd/plugins/lvm2/dmeventd_lvm.c
+++ b/daemons/dmeventd/plugins/lvm2/dmeventd_lvm.c
@@ -12,10 +12,10 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#include "lib/misc/lib.h"
+#include "lib.h"
 #include "dmeventd_lvm.h"
-#include "daemons/dmeventd/libdevmapper-event.h"
-#include "tools/lvm2cmd.h"
+#include "libdevmapper-event.h"
+#include "lvm2cmd.h"

 #include <pthread.h>

@@ -31,13 +31,6 @@ static pthread_mutex_t _register_mutex = PTHREAD_MUTEX_INITIALIZER;
 static int _register_count = 0;
 static struct dm_pool *_mem_pool = NULL;
 static void *_lvm_handle = NULL;
-static DM_LIST_INIT(_env_registry);
-
-struct env_data {
-	struct dm_list list;
-	const char *cmd;
-	const char *data;
-};

 DM_EVENT_LOG_FN("#lvm")

@@ -107,7 +100,6 @@ void dmeventd_lvm2_exit(void)
 		lvm2_run(_lvm_handle, "_memlock_dec");
 		dm_pool_destroy(_mem_pool);
 		_mem_pool = NULL;
-		dm_list_init(&_env_registry);
 		lvm2_exit(_lvm_handle);
 		_lvm_handle = NULL;
 		log_debug("lvm plugin exited.");
@@ -132,8 +124,6 @@ int dmeventd_lvm2_command(struct dm_pool *mem, char *buffer, size_t size,
 	static char _internal_prefix[] =  "_dmeventd_";
 	char *vg = NULL, *lv = NULL, *layer;
 	int r;
-	struct env_data *env_data;
-	const char *env = NULL;

 	if (!dm_split_lvm_name(mem, device, &vg, &lv, &layer)) {
 		log_error("Unable to determine VG name from %s.",
@@ -147,36 +137,18 @@ int dmeventd_lvm2_command(struct dm_pool *mem, char *buffer, size_t size,
 		*layer = '\0';

 	if (!strncmp(cmd, _internal_prefix, sizeof(_internal_prefix) - 1)) {
-		/* check if ENVVAR wasn't already resolved */
-		dm_list_iterate_items(env_data, &_env_registry)
-			if (!strcmp(cmd, env_data->cmd)) {
-				env = env_data->data;
-				break;
-			}
+		dmeventd_lvm2_lock();
+		/* output of internal command passed via env var */
+		if (!dmeventd_lvm2_run(cmd))
+			cmd = NULL;
+		else if ((cmd = getenv(cmd)))
+			cmd = dm_pool_strdup(mem, cmd); /* copy with lock */
+		dmeventd_lvm2_unlock();

-		if (!env) {
-			/* run lvm2 command to find out setting value */
-			dmeventd_lvm2_lock();
-			if (!dmeventd_lvm2_run(cmd) ||
-			    !(env = getenv(cmd))) {
-				dmeventd_lvm2_unlock();
-				log_error("Unable to find configured command.");
-				return 0;
-			}
-			/* output of internal command passed via env var */
-			env = dm_pool_strdup(_mem_pool, env); /* copy with lock */
-			dmeventd_lvm2_unlock();
-			if (!env ||
-			    !(env_data = dm_pool_zalloc(_mem_pool, sizeof(*env_data))) ||
-			    !(env_data->cmd = dm_pool_strdup(_mem_pool, cmd))) {
-				log_error("Unable to allocate env memory.");
-				return 0;
-			}
-			env_data->data = env;
-			/* add to ENVVAR registry */
-			dm_list_add(&_env_registry, &env_data->list);
+		if (!cmd) {
+			log_error("Unable to find configured command.");
+			return 0;
 		}
-		cmd = env;
 	}

 	r = dm_snprintf(buffer, size, "%s %s/%s", cmd, vg, lv);
--- a/daemons/dmeventd/plugins/mirror/Makefile.in
+++ b/daemons/dmeventd/plugins/mirror/Makefile.in
@@ -16,8 +16,8 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

+INCLUDES += -I$(top_srcdir)/daemons/dmeventd/plugins/lvm2
 CLDFLAGS += -L$(top_builddir)/daemons/dmeventd/plugins/lvm2
-LIBS += -ldevmapper-event-lvm2

 SOURCES = dmeventd_mirror.c

@@ -30,6 +30,8 @@ CFLOW_LIST_TARGET = $(LIB_NAME).cflow

 include $(top_builddir)/make.tmpl

+LIBS += -ldevmapper-event-lvm2 -ldevmapper
+
 install_lvm2: install_dm_plugin

 install: install_lvm2
--- a/daemons/dmeventd/plugins/mirror/dmeventd_mirror.c
+++ b/daemons/dmeventd/plugins/mirror/dmeventd_mirror.c
@@ -12,10 +12,10 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#include "lib/misc/lib.h"
-#include "daemons/dmeventd/plugins/lvm2/dmeventd_lvm.h"
-#include "daemons/dmeventd/libdevmapper-event.h"
-#include "lib/activate/activate.h"
+#include "lib.h"
+#include "libdevmapper-event.h"
+#include "dmeventd_lvm.h"
+#include "activate.h"	/* For TARGET_NAME* */

 /* FIXME Reformat to 80 char lines. */

--- a/daemons/dmeventd/plugins/raid/Makefile.in
+++ b/daemons/dmeventd/plugins/raid/Makefile.in
@@ -15,8 +15,8 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

+INCLUDES += -I$(top_srcdir)/daemons/dmeventd/plugins/lvm2
 CLDFLAGS += -L$(top_builddir)/daemons/dmeventd/plugins/lvm2
-LIBS += -ldevmapper-event-lvm2

 SOURCES = dmeventd_raid.c

@@ -29,6 +29,8 @@ CFLOW_LIST_TARGET = $(LIB_NAME).cflow

 include $(top_builddir)/make.tmpl

+LIBS += -ldevmapper-event-lvm2 -ldevmapper
+
 install_lvm2: install_dm_plugin

 install: install_lvm2
--- a/daemons/dmeventd/plugins/raid/dmeventd_raid.c
+++ b/daemons/dmeventd/plugins/raid/dmeventd_raid.c
@@ -12,10 +12,10 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#include "lib/misc/lib.h"
-#include "daemons/dmeventd/plugins/lvm2/dmeventd_lvm.h"
-#include "daemons/dmeventd/libdevmapper-event.h"
-#include "lib/config/defaults.h"
+#include "lib.h"
+#include "defaults.h"
+#include "dmeventd_lvm.h"
+#include "libdevmapper-event.h"

 /* Hold enough elements for the mximum number of RAID images */
 #define	RAID_DEVS_ELEMS	((DEFAULT_RAID_MAX_IMAGES + 63) / 64)
--- a/daemons/dmeventd/plugins/snapshot/Makefile.in
+++ b/daemons/dmeventd/plugins/snapshot/Makefile.in
@@ -16,8 +16,8 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

+INCLUDES += -I$(top_srcdir)/daemons/dmeventd/plugins/lvm2
 CLDFLAGS += -L$(top_builddir)/daemons/dmeventd/plugins/lvm2
-LIBS += -ldevmapper-event-lvm2

 SOURCES = dmeventd_snapshot.c

@@ -26,6 +26,8 @@ LIB_VERSION = $(LIB_VERSION_LVM)

 include $(top_builddir)/make.tmpl

+LIBS += -ldevmapper-event-lvm2 -ldevmapper
+
 install_lvm2: install_dm_plugin

 install: install_lvm2
--- a/daemons/dmeventd/plugins/snapshot/dmeventd_snapshot.c
+++ b/daemons/dmeventd/plugins/snapshot/dmeventd_snapshot.c
@@ -12,9 +12,9 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#include "lib/misc/lib.h"
-#include "daemons/dmeventd/plugins/lvm2/dmeventd_lvm.h"
-#include "daemons/dmeventd/libdevmapper-event.h"
+#include "lib.h"
+#include "dmeventd_lvm.h"
+#include "libdevmapper-event.h"

 #include <sys/sysmacros.h>
 #include <sys/wait.h>
@@ -175,7 +175,6 @@ void process_event(struct dm_task *dmt,
 	const char *device = dm_task_get_name(dmt);
 	int percent;
 	struct dm_info info;
-	int ret;

 	/* No longer monitoring, waiting for remove */
 	if (!state->percent_check)
@@ -206,8 +205,7 @@ void process_event(struct dm_task *dmt,
 		/* Maybe configurable ? */
 		_remove(dm_task_get_uuid(dmt));
 #endif
-		if ((ret = pthread_kill(pthread_self(), SIGALRM)) && (ret != ESRCH))
-			log_sys_error("pthread_kill", "self");
+		pthread_kill(pthread_self(), SIGALRM);
 		goto out;
 	}

@@ -215,8 +213,7 @@ void process_event(struct dm_task *dmt,
 		/* TODO eventually recognize earlier when room is enough */
 		log_info("Dropping monitoring of fully provisioned snapshot %s.",
 			 device);
-		if ((ret = pthread_kill(pthread_self(), SIGALRM)) && (ret != ESRCH))
-			log_sys_error("pthread_kill", "self");
+		pthread_kill(pthread_self(), SIGALRM);
 		goto out;
 	}

--- a/daemons/dmeventd/plugins/thin/Makefile.in
+++ b/daemons/dmeventd/plugins/thin/Makefile.in
@@ -15,8 +15,8 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

+INCLUDES += -I$(top_srcdir)/daemons/dmeventd/plugins/lvm2
 CLDFLAGS += -L$(top_builddir)/daemons/dmeventd/plugins/lvm2
-LIBS += -ldevmapper-event-lvm2

 SOURCES = dmeventd_thin.c

@@ -29,6 +29,8 @@ CFLOW_LIST_TARGET = $(LIB_NAME).cflow

 include $(top_builddir)/make.tmpl

+LIBS += -ldevmapper-event-lvm2 -ldevmapper
+
 install_lvm2: install_dm_plugin

 install: install_lvm2
--- a/daemons/dmeventd/plugins/thin/dmeventd_thin.c
+++ b/daemons/dmeventd/plugins/thin/dmeventd_thin.c
@@ -12,16 +12,16 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#include "lib/misc/lib.h"
-#include "daemons/dmeventd/plugins/lvm2/dmeventd_lvm.h"
-#include "daemons/dmeventd/libdevmapper-event.h"
+#include "lib.h"	/* using here lvm log */
+#include "dmeventd_lvm.h"
+#include "libdevmapper-event.h"

 #include <sys/wait.h>
 #include <stdarg.h>

 /* TODO - move this mountinfo code into library to be reusable */
 #ifdef __linux__
-#  include "libdm/misc/kdev_t.h"
+#  include "kdev_t.h"
 #else
 #  define MAJOR(x) major((x))
 #  define MINOR(x) minor((x))
@@ -62,25 +62,27 @@ struct dso_state {

 DM_EVENT_LOG_FN("thin")

+#define UUID_PREFIX "LVM-"
+
 static int _run_command(struct dso_state *state)
 {
-	char val[16];
+	char val[3][36];
+	char *env[] = { val[0], val[1], val[2], NULL };
 	int i;

 	/* Mark for possible lvm2 command we are running from dmeventd
 	 * lvm2 will not try to talk back to dmeventd while processing it */
-	(void) setenv("LVM_RUN_BY_DMEVENTD", "1", 1);
+	(void) dm_snprintf(val[0], sizeof(val[0]), "LVM_RUN_BY_DMEVENTD=1");

 	if (state->data_percent) {
 		/* Prepare some known data to env vars for easy use */
-		if (dm_snprintf(val, sizeof(val), "%d",
-				state->data_percent / DM_PERCENT_1) != -1)
-			(void) setenv("DMEVENTD_THIN_POOL_DATA", val, 1);
-		if (dm_snprintf(val, sizeof(val), "%d",
-				state->metadata_percent / DM_PERCENT_1) != -1)
-			(void) setenv("DMEVENTD_THIN_POOL_METADATA", val, 1);
+		(void) dm_snprintf(val[1], sizeof(val[1]), "DMEVENTD_THIN_POOL_DATA=%d",
+				   state->data_percent / DM_PERCENT_1);
+		(void) dm_snprintf(val[2], sizeof(val[2]), "DMEVENTD_THIN_POOL_METADATA=%d",
+				   state->metadata_percent / DM_PERCENT_1);
 	} else {
 		/* For an error event it's for a user to check status and decide */
+		env[1] = NULL;
 		log_debug("Error event processing.");
 	}

@@ -95,7 +97,7 @@ static int _run_command(struct dso_state *state)
 		/* child */
 		(void) close(0);
 		for (i = 3; i < 255; ++i) (void) close(i);
-		execvp(state->argv[0], state->argv);
+		execve(state->argv[0], state->argv, env);
 		_exit(errno);
 	} else if (state->pid == -1) {
 		log_error("Can't fork command %s.", state->cmd_str);
@@ -286,7 +288,7 @@ void process_event(struct dm_task *dmt,
 		if (state->fails++ <= state->max_fails) {
 			log_debug("Postponing frequently failing policy (%u <= %u).",
 				  state->fails - 1, state->max_fails);
-			goto out;
+			return;
 		}
 		if (state->max_fails < MAX_FAILS)
 			state->max_fails <<= 1;
--- a/daemons/dmeventd/plugins/vdo/.exported_symbols
+++ b/daemons/dmeventd/plugins/vdo/.exported_symbols
@@ -1,3 +0,0 @@
-process_event
-register_device
-unregister_device
--- a/daemons/dmeventd/plugins/vdo/dmeventd_vdo.c
+++ b/daemons/dmeventd/plugins/vdo/dmeventd_vdo.c
@@ -1,407 +0,0 @@
-/*
- * Copyright (C) 2018 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 Lesser General Public License v.2.1.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with this program; if not, write to the Free Software Foundation,
- * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include "lib/misc/lib.h"
-#include "daemons/dmeventd/plugins/lvm2/dmeventd_lvm.h"
-#include "daemons/dmeventd/libdevmapper-event.h"
-
-/* Use parser from new device_mapper library */
-#include "device_mapper/vdo/status.c"
-
-#include <sys/wait.h>
-#include <stdarg.h>
-
-/* First warning when VDO pool is 80% full. */
-#define WARNING_THRESH	(DM_PERCENT_1 * 80)
-/* Run a check every 5%. */
-#define CHECK_STEP	(DM_PERCENT_1 *  5)
-/* Do not bother checking VDO pool is less than 50% full. */
-#define CHECK_MINIMUM	(DM_PERCENT_1 * 50)
-
-#define MAX_FAILS	(256)  /* ~42 mins between cmd call retry with 10s delay */
-
-#define VDO_DEBUG 0
-
-struct dso_state {
-	struct dm_pool *mem;
-	int percent_check;
-	int percent;
-	uint64_t known_data_size;
-	unsigned fails;
-	unsigned max_fails;
-	int restore_sigset;
-	sigset_t old_sigset;
-	pid_t pid;
-	char *argv[3];
-	const char *cmd_str;
-	const char *name;
-};
-
-DM_EVENT_LOG_FN("vdo")
-
-static int _run_command(struct dso_state *state)
-{
-	char val[16];
-	int i;
-
-	/* Mark for possible lvm2 command we are running from dmeventd
-	 * lvm2 will not try to talk back to dmeventd while processing it */
-	(void) setenv("LVM_RUN_BY_DMEVENTD", "1", 1);
-
-	if (state->percent) {
-		/* Prepare some known data to env vars for easy use */
-		if (dm_snprintf(val, sizeof(val), "%d",
-				state->percent / DM_PERCENT_1) != -1)
-			(void) setenv("DMEVENTD_VDO_POOL", val, 1);
-	} else {
-		/* For an error event it's for a user to check status and decide */
-		log_debug("Error event processing.");
-	}
-
-	log_verbose("Executing command: %s", state->cmd_str);
-
-	/* TODO:
-	 *   Support parallel run of 'task' and it's waitpid maintainence
-	 *   ATM we can't handle signaling of  SIGALRM
-	 *   as signalling is not allowed while 'process_event()' is running
-	 */
-	if (!(state->pid = fork())) {
-		/* child */
-		(void) close(0);
-		for (i = 3; i < 255; ++i) (void) close(i);
-		execvp(state->argv[0], state->argv);
-		_exit(errno);
-	} else if (state->pid == -1) {
-		log_error("Can't fork command %s.", state->cmd_str);
-		state->fails = 1;
-		return 0;
-	}
-
-	return 1;
-}
-
-static int _use_policy(struct dm_task *dmt, struct dso_state *state)
-{
-#if VDO_DEBUG
-	log_debug("dmeventd executes: %s.", state->cmd_str);
-#endif
-	if (state->argv[0])
-		return _run_command(state);
-
-	if (!dmeventd_lvm2_run_with_lock(state->cmd_str)) {
-		log_error("Failed command for %s.", dm_task_get_name(dmt));
-		state->fails = 1;
-		return 0;
-	}
-
-	state->fails = 0;
-
-	return 1;
-}
-
-/* Check if executed command has finished
- * Only 1 command may run */
-static int _wait_for_pid(struct dso_state *state)
-{
-	int status = 0;
-
-	if (state->pid == -1)
-		return 1;
-
-	if (!waitpid(state->pid, &status, WNOHANG))
-		return 0;
-
-	/* Wait for finish */
-	if (WIFEXITED(status)) {
-		log_verbose("Child %d exited with status %d.",
-			    state->pid, WEXITSTATUS(status));
-		state->fails = WEXITSTATUS(status) ? 1 : 0;
-	} else {
-		if (WIFSIGNALED(status))
-			log_verbose("Child %d was terminated with status %d.",
-				    state->pid, WTERMSIG(status));
-		state->fails = 1;
-	}
-
-	state->pid = -1;
-
-	return 1;
-}
-
-void process_event(struct dm_task *dmt,
-		   enum dm_event_mask event __attribute__((unused)),
-		   void **user)
-{
-	const char *device = dm_task_get_name(dmt);
-	struct dso_state *state = *user;
-	void *next = NULL;
-	uint64_t start, length;
-	char *target_type = NULL;
-	char *params;
-	int needs_policy = 0;
-	struct dm_task *new_dmt = NULL;
-	struct dm_vdo_status_parse_result vdop = { .status = NULL };
-
-#if VDO_DEBUG
-	log_debug("Watch for VDO %s:%.2f%%.", state->name,
-		  dm_percent_to_round_float(state->percent_check, 2));
-#endif
-	if (!_wait_for_pid(state)) {
-		log_warn("WARNING: Skipping event, child %d is still running (%s).",
-			 state->pid, state->cmd_str);
-		return;
-	}
-
-	if (event & DM_EVENT_DEVICE_ERROR) {
-#if VDO_DEBUG
-		log_debug("VDO event error.");
-#endif
-		/* Error -> no need to check and do instant resize */
-		state->percent = 0;
-		if (_use_policy(dmt, state))
-			goto out;
-
-		stack;
-
-		if (!(new_dmt = dm_task_create(DM_DEVICE_STATUS)))
-			goto_out;
-
-		if (!dm_task_set_uuid(new_dmt, dm_task_get_uuid(dmt)))
-			goto_out;
-
-		/* Non-blocking status read */
-		if (!dm_task_no_flush(new_dmt))
-			log_warn("WARNING: Can't set no_flush for dm status.");
-
-		if (!dm_task_run(new_dmt))
-			goto_out;
-
-		dmt = new_dmt;
-	}
-
-	dm_get_next_target(dmt, next, &start, &length, &target_type, &params);
-
-	if (!target_type || (strcmp(target_type, "vdo") != 0)) {
-		log_error("Invalid target type.");
-		goto out;
-	}
-
-	if (!dm_vdo_status_parse(state->mem, params, &vdop)) {
-		log_error("Failed to parse status.");
-		goto out;
-	}
-
-	state->percent = dm_make_percent(vdop.status->used_blocks,
-					 vdop.status->total_blocks);
-
-#if VDO_DEBUG
-	log_debug("VDO %s status  %.2f%% " FMTu64 "/" FMTu64 ".",
-		  state->name, dm_percent_to_round_float(state->percent, 2),
-		  vdop.status->used_blocks, vdop.status->total_blocks);
-#endif
-
-	/* VDO pool size had changed. Clear the threshold. */
-	if (state->known_data_size != vdop.status->total_blocks) {
-		state->percent_check = CHECK_MINIMUM;
-		state->known_data_size = vdop.status->total_blocks;
-		state->fails = 0;
-	}
-
-	/*
-	 * Trigger action when threshold boundary is exceeded.
-	 * Report 80% threshold warning when it's used above 80%.
-	 * Only 100% is exception as it cannot be surpased so policy
-	 * action is called for:  >50%, >55% ... >95%, 100%
-	 */
-	if ((state->percent > WARNING_THRESH) &&
-	    (state->percent > state->percent_check))
-		log_warn("WARNING: VDO %s %s is now %.2f%% full.",
-			 state->name, device,
-			 dm_percent_to_round_float(state->percent, 2));
-	if (state->percent > CHECK_MINIMUM) {
-		/* Run action when usage raised more than CHECK_STEP since the last time */
-		if (state->percent > state->percent_check)
-			needs_policy = 1;
-		state->percent_check = (state->percent / CHECK_STEP + 1) * CHECK_STEP;
-		if (state->percent_check == DM_PERCENT_100)
-			state->percent_check--; /* Can't get bigger then 100% */
-	} else
-		state->percent_check = CHECK_MINIMUM;
-
-	/* Reduce number of _use_policy() calls by power-of-2 factor till frequency of MAX_FAILS is reached.
-	 * Avoids too high number of error retries, yet shows some status messages in log regularly.
-	 * i.e. PV could have been pvmoved and VG/LV was locked for a while...
-	 */
-	if (state->fails) {
-		if (state->fails++ <= state->max_fails) {
-			log_debug("Postponing frequently failing policy (%u <= %u).",
-				  state->fails - 1, state->max_fails);
-			goto out;
-		}
-		if (state->max_fails < MAX_FAILS)
-			state->max_fails <<= 1;
-		state->fails = needs_policy = 1; /* Retry failing command */
-	} else
-		state->max_fails = 1; /* Reset on success */
-
-	if (needs_policy)
-		_use_policy(dmt, state);
-out:
-	if (vdop.status)
-		dm_pool_free(state->mem, vdop.status);
-
-	if (new_dmt)
-		dm_task_destroy(new_dmt);
-}
-
-/* Handle SIGCHLD for a thread */
-static void _sig_child(int signum __attribute__((unused)))
-{
-	/* empty SIG_IGN */;
-}
-
-/* Setup handler for SIGCHLD when executing external command
- * to get quick 'waitpid()' reaction
- * It will interrupt syscall just like SIGALRM and
- * invoke process_event().
- */
-static void _init_thread_signals(struct dso_state *state)
-{
-	struct sigaction act = { .sa_handler = _sig_child };
-	sigset_t my_sigset;
-
-	sigemptyset(&my_sigset);
-
-	if (sigaction(SIGCHLD, &act, NULL))
-		log_warn("WARNING: Failed to set SIGCHLD action.");
-	else if (sigaddset(&my_sigset, SIGCHLD))
-		log_warn("WARNING: Failed to add SIGCHLD to set.");
-	else if (pthread_sigmask(SIG_UNBLOCK, &my_sigset, &state->old_sigset))
-		log_warn("WARNING: Failed to unblock SIGCHLD.");
-	else
-		state->restore_sigset = 1;
-}
-
-static void _restore_thread_signals(struct dso_state *state)
-{
-	if (state->restore_sigset &&
-	    pthread_sigmask(SIG_SETMASK, &state->old_sigset, NULL))
-		log_warn("WARNING: Failed to block SIGCHLD.");
-}
-
-int register_device(const char *device,
-		    const char *uuid,
-		    int major __attribute__((unused)),
-		    int minor __attribute__((unused)),
-		    void **user)
-{
-	struct dso_state *state;
-	const char *cmd;
-	char *str;
-	char cmd_str[PATH_MAX + 128 + 2]; /* cmd ' ' vg/lv \0 */
-        const char *name = "pool";
-
-	if (!dmeventd_lvm2_init_with_pool("vdo_pool_state", state))
-		goto_bad;
-
-	state->cmd_str = "";
-
-	/* Search for command for LVM- prefixed devices only */
-	cmd = (strncmp(uuid, "LVM-", 4) == 0) ? "_dmeventd_vdo_command" : "";
-
-	if (!dmeventd_lvm2_command(state->mem, cmd_str, sizeof(cmd_str), cmd, device))
-		goto_bad;
-
-	if (strncmp(cmd_str, "lvm ", 4) == 0) {
-		if (!(state->cmd_str = dm_pool_strdup(state->mem, cmd_str + 4))) {
-			log_error("Failed to copy lvm VDO command.");
-				goto bad;
-		}
-	} else if (cmd_str[0] == '/') {
-		if (!(state->cmd_str = dm_pool_strdup(state->mem, cmd_str))) {
-			log_error("Failed to copy VDO command.");
-			goto bad;
-		}
-
-		/* Find last space before 'vg/lv' */
-		if (!(str = strrchr(state->cmd_str, ' ')))
-			goto inval;
-
-		if (!(state->argv[0] = dm_pool_strndup(state->mem, state->cmd_str,
-						       str - state->cmd_str))) {
-			log_error("Failed to copy command.");
-			goto bad;
-		}
-
-		state->argv[1] = str + 1;  /* 1 argument - vg/lv */
-		_init_thread_signals(state);
-	} else if (cmd[0] == 0) {
-		state->name = "volume"; /* What to use with 'others?' */
-	} else/* Unuspported command format */
-		goto inval;
-
-	state->pid = -1;
-	state->name = name;
-	*user = state;
-
-	log_info("Monitoring VDO %s %s.", name, device);
-
-	return 1;
-inval:
-	log_error("Invalid command for monitoring: %s.", cmd_str);
-bad:
-	log_error("Failed to monitor VDO %s %s.", name, device);
-
-	if (state)
-		dmeventd_lvm2_exit_with_pool(state);
-
-	return 0;
-}
-
-int unregister_device(const char *device,
-		      const char *uuid __attribute__((unused)),
-		      int major __attribute__((unused)),
-		      int minor __attribute__((unused)),
-		      void **user)
-{
-	struct dso_state *state = *user;
-	const char *name = state->name;
-	int i;
-
-	for (i = 0; !_wait_for_pid(state) && (i < 6); ++i) {
-		if (i == 0)
-			/* Give it 2 seconds, then try to terminate & kill it */
-			log_verbose("Child %d still not finished (%s) waiting.",
-				    state->pid, state->cmd_str);
-		else if (i == 3) {
-			log_warn("WARNING: Terminating child %d.", state->pid);
-			kill(state->pid, SIGINT);
-			kill(state->pid, SIGTERM);
-		} else if (i == 5) {
-			log_warn("WARNING: Killing child %d.", state->pid);
-			kill(state->pid, SIGKILL);
-		}
-		sleep(1);
-	}
-
-	if (state->pid != -1)
-		log_warn("WARNING: Cannot kill child %d!", state->pid);
-
-	_restore_thread_signals(state);
-
-	dmeventd_lvm2_exit_with_pool(state);
-	log_info("No longer monitoring VDO %s %s.", name, device);
-
-	return 1;
-}
--- a/daemons/dmfilemapd/.gitignore
+++ b/daemons/dmfilemapd/.gitignore
@@ -1,2 +1 @@
-dmsetup
 dmfilemapd
--- a/daemons/dmfilemapd/Makefile.in
+++ b/daemons/dmfilemapd/Makefile.in
@@ -0,0 +1,66 @@
+#
+# Copyright (C) 2016 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 Lesser General Public License v.2.1.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+top_builddir = @top_builddir@
+
+SOURCES = dmfilemapd.c
+
+TARGETS = dmfilemapd
+
+.PHONY: install_dmfilemapd install_dmfilemapd_static
+
+INSTALL_DMFILEMAPD_TARGETS = install_dmfilemapd_dynamic
+
+CLEAN_TARGETS = dmfilemapd.static
+
+CFLOW_LIST = $(SOURCES)
+CFLOW_LIST_TARGET = $(LIB_NAME).cflow
+CFLOW_TARGET = dmfilemapd
+
+include $(top_builddir)/make.tmpl
+
+all: device-mapper
+device-mapper: $(TARGETS)
+
+CFLAGS_dmfilemapd.o += $(EXTRA_EXEC_CFLAGS)
+LIBS += -ldevmapper
+
+dmfilemapd: $(LIB_SHARED) dmfilemapd.o
+	$(CC) $(CFLAGS) $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) \
+		-o $@ dmfilemapd.o $(DL_LIBS) $(LIBS)
+
+dmfilemapd.static: $(LIB_STATIC) dmfilemapd.o $(interfacebuilddir)/libdevmapper.a
+	$(CC) $(CFLAGS) $(LDFLAGS) $(ELDFLAGS) -static -L$(interfacebuilddir) \
+		-o $@ dmfilemapd.o $(DL_LIBS) $(LIBS) $(STATIC_LIBS)
+
+ifneq ("$(CFLOW_CMD)", "")
+CFLOW_SOURCES = $(addprefix $(srcdir)/, $(SOURCES))
+-include $(top_builddir)/libdm/libdevmapper.cflow
+-include $(top_builddir)/lib/liblvm-internal.cflow
+-include $(top_builddir)/lib/liblvm2cmd.cflow
+-include $(top_builddir)/daemons/dmfilemapd/$(LIB_NAME).cflow
+endif
+
+install_dmfilemapd_dynamic: dmfilemapd
+	$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
+
+install_dmfilemapd_static: dmfilemapd.static
+	$(INSTALL_PROGRAM) -D $< $(staticdir)/$(<F)
+
+install_dmfilemapd: $(INSTALL_DMFILEMAPD_TARGETS)
+
+install: install_dmfilemapd
+
+install_device-mapper: install_dmfilemapd
--- a/daemons/dmfilemapd/dmfilemapd.c
+++ b/daemons/dmfilemapd/dmfilemapd.c
@@ -14,8 +14,11 @@
 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

-#include "util.h"
-#include "libdm/misc/dm-logging.h"
+#include "tool.h"
+
+#include "dm-logging.h"
+
+#include "defaults.h"

 #include <sys/types.h>
 #include <sys/stat.h>
@@ -26,15 +29,13 @@
 #include <ctype.h>

 #ifdef __linux__
-#  include "libdm/misc/kdev_t.h"
+#  include "kdev_t.h"
 #else
 #  define MAJOR(x) major((x))
 #  define MINOR(x) minor((x))
 #  define MKDEV(x,y) makedev((x),(y))
 #endif

-#define DEFAULT_PROC_DIR "/proc"
-
 /* limit to two updates/sec */
 #define FILEMAPD_WAIT_USECS 500000

@@ -310,7 +311,7 @@ static int _parse_args(int argc, char **argv, struct filemap_monitor *fm)
 		return 0;
 	}

-	fm->path = strdup(argv[0]);
+	fm->path = dm_strdup(argv[0]);
 	if (!fm->path) {
 		_early_log("Could not allocate memory for path argument.");
 		return 0;
@@ -537,8 +538,8 @@ static void _filemap_monitor_destroy(struct filemap_monitor *fm)
 		_filemap_monitor_end_notify(fm);
 		_filemap_monitor_close_fd(fm);
 	}
-	free((void *) fm->program_id);
-	free(fm->path);
+	dm_free((void *) fm->program_id);
+	dm_free(fm->path);
 }

 static int _filemap_monitor_check_same_file(int fd1, int fd2)
@@ -628,10 +629,10 @@ check_unlinked:

 static int _daemonise(struct filemap_monitor *fm)
 {
-	pid_t pid = 0;
+	pid_t pid = 0, sid;
 	int fd;

-	if (!setsid()) {
+	if (!(sid = setsid())) {
 		_early_log("setsid failed.");
 		return 0;
 	}
@@ -698,7 +699,7 @@ static int _update_regions(struct dm_stats *dms, struct filemap_monitor *fm)
 			 fm->group_id, regions[0]);
 		fm->group_id = regions[0];
 	}
-	free(regions);
+	dm_free(regions);
 	fm->nr_regions = nr_regions;
 	return 1;
 }
@@ -740,7 +741,7 @@ static int _dmfilemapd(struct filemap_monitor *fm)
 	 */
 	program_id = dm_stats_get_region_program_id(dms, fm->group_id);
 	if (program_id)
-		fm->program_id = strdup(program_id);
+		fm->program_id = dm_strdup(program_id);
 	else
 		fm->program_id = NULL;
 	dm_stats_set_program_id(dms, 1, program_id);
@@ -801,7 +802,7 @@ bad:
 	return 1;
 }

-static const char * const _mode_names[] = {
+static const char * _mode_names[] = {
 	"inode",
 	"path"
 };
@@ -816,7 +817,7 @@ int main(int argc, char **argv)
 	memset(&fm, 0, sizeof(fm));

 	if (!_parse_args(argc, argv, &fm)) {
-		free(fm.path);
+		dm_free(fm.path);
 		return 1;
 	}

@@ -826,10 +827,8 @@ int main(int argc, char **argv)
 		 "mode=%s, path=%s", fm.fd, fm.group_id,
 		 _mode_names[fm.mode], fm.path);

-	if (!_foreground && !_daemonise(&fm)) {
-		free(fm.path);
+	if (!_foreground && !_daemonise(&fm))
 		return 1;
-	}

 	return _dmfilemapd(&fm);
 }
--- a/daemons/lvmdbusd/.gitignore
+++ b/daemons/lvmdbusd/.gitignore
@@ -1,4 +1 @@
 path.py
-lvmdbusd
-lvmdb.py
-lvm_shell_proxy.py
--- a/daemons/lvmdbusd/Makefile.in
+++ b/daemons/lvmdbusd/Makefile.in
@@ -23,10 +23,13 @@ LVMDBUS_SRCDIR_FILES = \
 	cfg.py \
 	cmdhandler.py \
 	fetch.py \
+	__init__.py \
 	job.py \
 	loader.py \
-	lv.py \
+	lvmdb.py \
 	main.py \
+	lvm_shell_proxy.py \
+	lv.py \
 	manager.py \
 	objectmanager.py \
 	pv.py \
@@ -34,40 +37,30 @@ LVMDBUS_SRCDIR_FILES = \
 	state.py \
 	udevwatch.py \
 	utils.py \
-	vg.py \
-	__init__.py
+	vg.py

 LVMDBUS_BUILDDIR_FILES = \
-	lvmdb.py \
-	lvm_shell_proxy.py \
 	path.py

-LVMDBUSD = lvmdbusd
-
-CLEAN_DIRS += __pycache__
+LVMDBUSD = $(srcdir)/lvmdbusd

 include $(top_builddir)/make.tmpl

 .PHONY: install_lvmdbusd

-all:
-	$(Q) test -x $(LVMDBUSD) || chmod 755 $(LVMDBUSD)
-
 install_lvmdbusd:
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_DIR) $(sbindir)
-	$(Q) $(INSTALL_SCRIPT) $(LVMDBUSD) $(sbindir)
-	$(Q) $(INSTALL_DIR) $(DESTDIR)$(lvmdbusdir)
-	$(Q) (cd $(srcdir); $(INSTALL_DATA) $(LVMDBUS_SRCDIR_FILES) $(DESTDIR)$(lvmdbusdir))
-	$(Q) $(INSTALL_DATA) $(LVMDBUS_BUILDDIR_FILES) $(DESTDIR)$(lvmdbusdir)
-	$(Q) PYTHON=$(PYTHON3) $(PYCOMPILE) --destdir "$(DESTDIR)" --basedir "$(lvmdbusdir)" $(LVMDBUS_SRCDIR_FILES) $(LVMDBUS_BUILDDIR_FILES)
-	$(Q) $(CHMOD) 755 $(DESTDIR)$(lvmdbusdir)/__pycache__
-	$(Q) $(CHMOD) 444 $(DESTDIR)$(lvmdbusdir)/__pycache__/*.py[co]
+	$(INSTALL_DIR) $(sbindir)
+	$(INSTALL_SCRIPT) $(LVMDBUSD) $(sbindir)
+	$(INSTALL_DIR) $(DESTDIR)$(lvmdbusdir)
+	(cd $(srcdir); $(INSTALL_DATA) $(LVMDBUS_SRCDIR_FILES) $(DESTDIR)$(lvmdbusdir))
+	$(INSTALL_DATA) $(LVMDBUS_BUILDDIR_FILES) $(DESTDIR)$(lvmdbusdir)
+	PYTHON=$(PYTHON3) $(PYCOMPILE) --destdir "$(DESTDIR)" --basedir "$(lvmdbusdir)" $(LVMDBUS_SRCDIR_FILES) $(LVMDBUS_BUILDDIR_FILES)
+	$(CHMOD) 755 $(DESTDIR)$(lvmdbusdir)/__pycache__
+	$(CHMOD) 444 $(DESTDIR)$(lvmdbusdir)/__pycache__/*.py[co]

 install_lvm2: install_lvmdbusd

 install: install_lvm2

 DISTCLEAN_TARGETS+= \
-	$(LVMDBUS_BUILDDIR_FILES) \
-	$(LVMDBUSD)
+	$(LVMDBUS_BUILDDIR_FILES)
--- a/daemons/lvmdbusd/automatedproperties.py
+++ b/daemons/lvmdbusd/automatedproperties.py
@@ -155,7 +155,7 @@ class AutomatedProperties(dbus.service.Object):
 		# through all dbus objects as some don't have a search method, like
 		# 'Manager' object.
 		if not self._ap_search_method:
-			return 0
+			return

 		search = self.lvm_id
 		if search_key:
--- a/daemons/lvmdbusd/cfg.py
+++ b/daemons/lvmdbusd/cfg.py
@@ -87,13 +87,3 @@ blackbox = None

 # RequestEntry ctor
 create_request_entry = None
-
-
-def exit_daemon():
-    """
-    Exit the daemon cleanly
-    :return:
-    """
-    if run and loop:
-        run.value = 0
-        loop.quit()
--- a/daemons/lvmdbusd/cmdhandler.py
+++ b/daemons/lvmdbusd/cmdhandler.py
@@ -67,7 +67,7 @@ class LvmFlightRecorder(object):
 		with cmd_lock:
 			if len(self.queue):
 				log_error("LVM dbus flight recorder START")
-				for c in reversed(self.queue):
+				for c in self.queue:
 					log_error(str(c))
 				log_error("LVM dbus flight recorder END")

@@ -263,10 +263,10 @@ def lv_tag(lv_name, add, rm, tag_options):
 	return _tag('lvchange', lv_name, add, rm, tag_options)


-def vg_rename(vg_uuid, new_name, rename_options):
+def vg_rename(vg, new_name, rename_options):
 	cmd = ['vgrename']
 	cmd.extend(options_to_cli_args(rename_options))
-	cmd.extend([vg_uuid, new_name])
+	cmd.extend([vg, new_name])
 	return call(cmd)


@@ -497,8 +497,7 @@ def lvm_full_report_json():
 	])

 	rc, out, err = call(cmd)
-	# When we have an exported vg the exit code of lvs or fullreport will be 5
-	if rc == 0 or rc == 5:
+	if rc == 0:
 		# With the current implementation, if we are using the shell then we
 		# are using JSON and JSON is returned back to us as it was parsed to
 		# figure out if we completed OK or not
--- a/daemons/lvmdbusd/fetch.py
+++ b/daemons/lvmdbusd/fetch.py
@@ -14,7 +14,6 @@ from . import cfg
 from .utils import MThreadRunner, log_debug, log_error
 import threading
 import queue
-import time
 import traceback


@@ -83,8 +82,6 @@ class StateUpdate(object):

 	@staticmethod
 	def update_thread(obj):
-		exception_count = 0
-
 		queued_requests = []
 		while cfg.run.value != 0:
 			# noinspection PyBroadException
@@ -139,26 +136,12 @@ class StateUpdate(object):
 				# wake up if we get an exception
 				queued_requests = []

-				# We retrieved OK, clear exception count
-				exception_count = 0
-
 			except queue.Empty:
 				pass
-			except Exception as e:
+			except Exception:
 				st = traceback.format_exc()
 				log_error("update_thread exception: \n%s" % st)
 				cfg.blackbox.dump()
-				exception_count += 1
-				if exception_count >= 5:
-					for i in queued_requests:
-						i.set_result(e)
-
-					log_error("Too many errors in update_thread, exiting daemon")
-					cfg.exit_daemon()
-
-				else:
-					# Slow things down when encountering errors
-					time.sleep(1)

 	def __init__(self):
 		self.lock = threading.RLock()
--- a/daemons/lvmdbusd/lv.py
+++ b/daemons/lvmdbusd/lv.py
@@ -10,7 +10,7 @@
 from .automatedproperties import AutomatedProperties

 from . import utils
-from .utils import vg_obj_path_generate, log_error
+from .utils import vg_obj_path_generate
 import dbus
 from . import cmdhandler
 from . import cfg
@@ -24,8 +24,6 @@ from . import background
 from .utils import round_size, mt_remove_dbus_objects
 from .job import JobState

-import traceback
-

 # Try and build a key for a LV, so that we sort the LVs with least dependencies
 # first.  This may be error prone because of the flexibility LVM
@@ -234,6 +232,7 @@ class LvState(State):
@utils.dbus_property(LV_COMMON_INTERFACE, 'Attr', 's')
@utils.dbus_property(LV_COMMON_INTERFACE, 'DataPercent', 'u')
@utils.dbus_property(LV_COMMON_INTERFACE, 'SnapPercent', 'u')
+@utils.dbus_property(LV_COMMON_INTERFACE, 'DataPercent', 'u')
@utils.dbus_property(LV_COMMON_INTERFACE, 'MetaDataPercent', 'u')
@utils.dbus_property(LV_COMMON_INTERFACE, 'CopyPercent', 'u')
@utils.dbus_property(LV_COMMON_INTERFACE, 'SyncPercent', 'u')
@@ -293,22 +292,6 @@ class LvCommon(AutomatedProperties):
 				(lv_uuid, lv_name))
 		return dbo

-	def attr_struct(self, index, type_map, default='undisclosed'):
-		try:
-			if self.state.Attr[index] not in type_map:
-				log_error("LV %s %s with lv_attr %s, lv_attr[%d] = "
-					"'%s' is not known" %
-					(self.Uuid, self.Name, self.Attr, index,
-					self.state.Attr[index]))
-
-			return dbus.Struct((self.state.Attr[index],
-				type_map.get(self.state.Attr[index], default)),
-								signature="(ss)")
-		except BaseException:
-			st = traceback.format_exc()
-			log_error("attr_struct: \n%s" % st)
-			return dbus.Struct(('?', 'Unavailable'), signature="(ss)")
-
 	@property
 	def VolumeType(self):
 		type_map = {'C': 'Cache', 'm': 'mirrored',
@@ -322,14 +305,16 @@ class LvCommon(AutomatedProperties):
 					'V': 'thin Volume', 't': 'thin pool', 'T': 'Thin pool data',
 					'e': 'raid or pool metadata or pool metadata spare',
 					'-': 'Unspecified'}
-		return self.attr_struct(0, type_map)
+		return dbus.Struct((self.state.Attr[0], type_map[self.state.Attr[0]]),
+						signature="as")

 	@property
 	def Permissions(self):
 		type_map = {'w': 'writable', 'r': 'read-only',
 					'R': 'Read-only activation of non-read-only volume',
 					'-': 'Unspecified'}
-		return self.attr_struct(1, type_map)
+		return dbus.Struct((self.state.Attr[1], type_map[self.state.Attr[1]]),
+						signature="(ss)")

 	@property
 	def AllocationPolicy(self):
@@ -338,7 +323,8 @@ class LvCommon(AutomatedProperties):
 					'i': 'inherited', 'I': 'inherited locked',
 					'l': 'cling', 'L': 'cling locked',
 					'n': 'normal', 'N': 'normal locked', '-': 'Unspecified'}
-		return self.attr_struct(2, type_map)
+		return dbus.Struct((self.state.Attr[2], type_map[self.state.Attr[2]]),
+						signature="(ss)")

 	@property
 	def FixedMinor(self):
@@ -346,20 +332,15 @@ class LvCommon(AutomatedProperties):

 	@property
 	def State(self):
-		type_map = {'a': 'active',
-					's': 'suspended',
-					'I': 'Invalid snapshot',
+		type_map = {'a': 'active', 's': 'suspended', 'I': 'Invalid snapshot',
 					'S': 'invalid Suspended snapshot',
 					'm': 'snapshot merge failed',
 					'M': 'suspended snapshot (M)erge failed',
 					'd': 'mapped device present without  tables',
 					'i': 'mapped device present with inactive table',
-					'h': 'historical',
-					'c': 'check needed suspended thin-pool',
-					'C': 'check needed',
-					'X': 'unknown',
-					'-': 'Unspecified'}
-		return self.attr_struct(4, type_map)
+					'X': 'unknown', '-': 'Unspecified'}
+		return dbus.Struct((self.state.Attr[4], type_map[self.state.Attr[4]]),
+						signature="(ss)")

 	@property
 	def TargetType(self):
@@ -375,18 +356,11 @@ class LvCommon(AutomatedProperties):

 	@property
 	def Health(self):
-		type_map = {'p': 'partial',
-					'r': 'refresh needed',
-					'm': 'mismatches',
-					'w': 'writemostly',
-					'X': 'unknown',
-					'-': 'unspecified',
-					's': 'reshaping',
-					'F': 'failed',
-					'D': 'Data space',
-					'R': 'Remove',
-					'M': 'Metadata'}
-		return self.attr_struct(8, type_map)
+		type_map = {'p': 'partial', 'r': 'refresh',
+					'm': 'mismatches', 'w': 'writemostly',
+					'X': 'X unknown', '-': 'Unspecified'}
+		return dbus.Struct((self.state.Attr[8], type_map[self.state.Attr[8]]),
+					signature="(ss)")

 	@property
 	def SkipActivation(self):
@@ -524,7 +498,7 @@ class Lv(LvCommon):
 		# it is a thin lv
 		if not dbo.IsThinVolume:
 			if optional_size == 0:
-				space = dbo.SizeBytes // 80
+				space = dbo.SizeBytes / 80
 				remainder = space % 512
 				optional_size = space + 512 - remainder

--- a/daemons/lvmdbusd/lvm_shell_proxy.py.in
+++ b/daemons/lvmdbusd/lvm_shell_proxy.py.in
@@ -1,4 +1,4 @@
-#!@PYTHON3@
+#!/usr/bin/env python3

 # Copyright (C) 2015-2016 Red Hat, Inc. All rights reserved.
 #
@@ -220,10 +220,7 @@ class LVMShellProxy(object):

 		# Parse the report to see what happened
 		if 'log' in report_json:
-			ret_code = int(report_json['log'][-1:][0]['log_ret_code'])
-			# If we have an exported vg we get a log_ret_code == 5 when
-			# we do a 'fullreport'
-			if (ret_code == 1) or (ret_code == 5 and argv[0] == 'fullreport'):
+			if report_json['log'][-1:][0]['log_ret_code'] == '1':
 				rc = 0
 			else:
 				error_msg = self.get_error_msg()
--- a/daemons/lvmdbusd/lvmdb.py.in
+++ b/daemons/lvmdbusd/lvmdb.py.in
@@ -1,4 +1,4 @@
-#!@PYTHON3@
+#!/usr/bin/env python3

 # Copyright (C) 2015-2016 Red Hat, Inc. All rights reserved.
 #
@@ -141,22 +141,13 @@ class DataStore(object):

 	@staticmethod
 	def _parse_vgs(_vgs):
-		vgs = sorted(_vgs, key=lambda vk: vk['vg_uuid'])
+		vgs = sorted(_vgs, key=lambda vk: vk['vg_name'])

 		c_vgs = OrderedDict()
 		c_lookup = {}

 		for i in vgs:
-			vg_name = i['vg_name']
-
-			# Lvm allows duplicate vg names.  When this occurs, each subsequent
-			# matching VG name will be called vg_name:vg_uuid.  Note: ':' is an
-			# invalid character for lvm VG names
-			if vg_name in c_lookup:
-				vg_name = "%s:%s" % (vg_name, i['vg_uuid'])
-				i['vg_name'] = vg_name
-
-			c_lookup[vg_name] = i['vg_uuid']
+			c_lookup[i['vg_name']] = i['vg_uuid']
 			DataStore._insert_record(c_vgs, i['vg_uuid'], i, [])

 		return c_vgs, c_lookup
@@ -171,22 +162,13 @@ class DataStore(object):
 				tmp_vg.extend(r['vg'])

 		# Sort for consistent output, however this is optional
-		vgs = sorted(tmp_vg, key=lambda vk: vk['vg_uuid'])
+		vgs = sorted(tmp_vg, key=lambda vk: vk['vg_name'])

 		c_vgs = OrderedDict()
 		c_lookup = {}

 		for i in vgs:
-			vg_name = i['vg_name']
-
-			# Lvm allows duplicate vg names.  When this occurs, each subsequent
-			# matching VG name will be called vg_name:vg_uuid.  Note: ':' is an
-			# invalid character for lvm VG names
-			if vg_name in c_lookup:
-				vg_name = "%s:%s" % (vg_name, i['vg_uuid'])
-				i['vg_name'] = vg_name
-
-			c_lookup[vg_name] = i['vg_uuid']
+			c_lookup[i['vg_name']] = i['vg_uuid']
 			c_vgs[i['vg_uuid']] = i

 		return c_vgs, c_lookup
@@ -539,10 +521,6 @@ if __name__ == "__main__":
 	for v in ds.vgs.values():
 		pp.pprint(v)

-	print("VG name to UUID")
-	for k, v in ds.vg_name_to_uuid.items():
-		print("%s: %s" % (k, v))
-
 	print("LVS")
 	for v in ds.lvs.values():
 		pp.pprint(v)
--- a/daemons/lvmdbusd/lvmdbusd.in
+++ b/daemons/lvmdbusd/lvmdbusd.in
@@ -1,4 +1,4 @@
-#!@PYTHON3@
+#!/usr/bin/env python3

 # Copyright (C) 2015-2016 Red Hat, Inc. All rights reserved.
 #
--- a/daemons/lvmdbusd/manager.py
+++ b/daemons/lvmdbusd/manager.py
@@ -164,8 +164,6 @@ class Manager(AutomatedProperties):
 		return the object path in O(1) time.

 		:param key: The lookup value
-		:param cb:	dbus python call back parameter, not client visible
-		:param cbe:	dbus python error call back parameter, not client visible
 		:return: Return the object path.  If object not found you will get '/'
 		"""
 		r = RequestEntry(-1, Manager._lookup_by_lvm_id, (key,), cb, cbe, False)
--- a/daemons/lvmdbusd/objectmanager.py
+++ b/daemons/lvmdbusd/objectmanager.py
@@ -189,8 +189,8 @@ class ObjectManager(AutomatedProperties):
 			path = dbus_object.dbus_object_path()
 			interfaces = dbus_object.interface()

-			# print('UN-Registering object path %s for %s' %
-			#		(path, dbus_object.lvm_id))
+			# print 'UN-Registering object path %s for %s' % \
+			#      (path, dbus_object.lvm_id)

 			self._lookup_remove(path)

@@ -240,19 +240,39 @@ class ObjectManager(AutomatedProperties):
 				return lookup_rc
 			return '/'

-	def _id_verify(self, path, uuid, lvm_id):
+	def _uuid_verify(self, path, uuid, lvm_id):
 		"""
-		Ensure our lookups are correct
+		Ensure uuid is present for a successful lvm_id lookup
 		NOTE: Internal call, assumes under object manager lock
 		:param path: 		Path to object we looked up
-		:param uuid: 		uuid lookup
-		:param lvm_id:		lvm_id lookup
+		:param uuid: 		lvm uuid to verify
+		:param lvm_id:		lvm_id used to find object
 		:return: None
 		"""
-		# There is no durable non-changeable name in lvm
+		# This gets called when we found an object based on lvm_id, ensure
+		# uuid is correct too, as they can change. There is no durable
+		# non-changeable name in lvm
 		if lvm_id != uuid:
-			obj = self.get_object_by_path(path)
-			self._lookup_add(obj, path, lvm_id, uuid)
+			if uuid and uuid not in self._id_to_object_path:
+				obj = self.get_object_by_path(path)
+				self._lookup_add(obj, path, lvm_id, uuid)
+
+	def _lvm_id_verify(self, path, uuid, lvm_id):
+		"""
+		Ensure lvm_id is present for a successful uuid lookup
+		NOTE: Internal call, assumes under object manager lock
+		:param path: 		Path to object we looked up
+		:param uuid: 		uuid used to find object
+		:param lvm_id:		lvm_id to verify
+		:return: None
+		"""
+		# This gets called when we found an object based on uuid, ensure
+		# lvm_id is correct too, as they can change.  There is no durable
+		# non-changeable name in lvm
+		if lvm_id != uuid:
+			if lvm_id and lvm_id not in self._id_to_object_path:
+				obj = self.get_object_by_path(path)
+				self._lookup_add(obj, path, lvm_id, uuid)

 	def _id_lookup(self, the_id):
 		path = None
@@ -319,22 +339,22 @@ class ObjectManager(AutomatedProperties):
 				# Lets check for the uuid first
 				path = self._id_lookup(uuid)
 				if path:
-					# Ensure table lookups are correct
-					self._id_verify(path, uuid, lvm_id)
+					# Verify the lvm_id is sane
+					self._lvm_id_verify(path, uuid, lvm_id)
 				else:
 					# Unable to find by UUID, lets lookup by lvm_id
 					path = self._id_lookup(lvm_id)
 					if path:
-						# Ensure table lookups are correct
-						self._id_verify(path, uuid, lvm_id)
+						# Verify the uuid is sane
+						self._uuid_verify(path, uuid, lvm_id)
 					else:
 						# We have exhausted all lookups, let's create if we can
 						if path_create:
 							path = path_create()
 							self._lookup_add(None, path, lvm_id, uuid)

-			# print('get_object_path_by_lvm_id(%s, %s, %s): return %s' %
-			#	(uuid, lvm_id, str(path_create), path))
+			# print('get_object_path_by_lvm_id(%s, %s, %s, %s: return %s' %
+			# 	   (uuid, lvm_id, str(path_create), str(gen_new), path))

 			return path

--- a/daemons/lvmdbusd/vg.py
+++ b/daemons/lvmdbusd/vg.py
@@ -52,23 +52,18 @@ def load_vgs(vg_specific=None, object_path=None, refresh=False,

 # noinspection PyPep8Naming,PyUnresolvedReferences,PyUnusedLocal
 class VgState(State):
-
-	@property
-	def internal_name(self):
-		return self.Name
-
 	@property
 	def lvm_id(self):
-		return self.internal_name
+		return self.Name

 	def identifiers(self):
-		return (self.Uuid, self.internal_name)
+		return (self.Uuid, self.Name)

 	def _lv_paths_build(self):
 		rc = []
 		for lv in cfg.db.lvs_in_vg(self.Uuid):
 			(lv_name, meta, lv_uuid) = lv
-			full_name = "%s/%s" % (self.internal_name, lv_name)
+			full_name = "%s/%s" % (self.Name, lv_name)

 			gen = utils.lv_object_path_method(lv_name, meta)

@@ -97,7 +92,7 @@ class VgState(State):
 	def create_dbus_object(self, path):
 		if not path:
 			path = cfg.om.get_object_path_by_uuid_lvm_id(
-				self.Uuid, self.internal_name, vg_obj_path_generate)
+				self.Uuid, self.Name, vg_obj_path_generate)
 		return Vg(path, self)

 	# noinspection PyMethodMayBeStatic
@@ -107,6 +102,7 @@ class VgState(State):

 # noinspection PyPep8Naming
@utils.dbus_property(VG_INTERFACE, 'Uuid', 's')
+@utils.dbus_property(VG_INTERFACE, 'Name', 's')
@utils.dbus_property(VG_INTERFACE, 'Fmt', 's')
@utils.dbus_property(VG_INTERFACE, 'SizeBytes', 't', 0)
@utils.dbus_property(VG_INTERFACE, 'FreeBytes', 't', 0)
@@ -139,7 +135,6 @@ class Vg(AutomatedProperties):
 	_AllocNormal_meta = ('b', VG_INTERFACE)
 	_AllocAnywhere_meta = ('b', VG_INTERFACE)
 	_Clustered_meta = ('b', VG_INTERFACE)
-	_Name_meta = ('s', VG_INTERFACE)

 	# noinspection PyUnusedLocal,PyPep8Naming
 	def __init__(self, object_path, object_state):
@@ -177,7 +172,7 @@ class Vg(AutomatedProperties):
 		# Make sure we have a dbus object representing it
 		Vg.validate_dbus_object(uuid, vg_name)
 		rc, out, err = cmdhandler.vg_rename(
-			uuid, new_name, rename_options)
+			vg_name, new_name, rename_options)
 		Vg.handle_execute(rc, out, err)
 		return '/'

@@ -221,7 +216,7 @@ class Vg(AutomatedProperties):
 	# TODO: This should be broken into a number of different methods
 	# instead of having one method that takes a hash for parameters.  Some of
 	# the changes that vgchange does works on entire system, not just a
-	# specific vg, thus that should be in the Manager interface.
+	# specfic vg, thus that should be in the Manager interface.
 	@dbus.service.method(
 		dbus_interface=VG_INTERFACE,
 		in_signature='ia{sv}',
@@ -734,12 +729,6 @@ class Vg(AutomatedProperties):
 				cb, cbe, return_tuple=False)
 		cfg.worker_q.put(r)

-	@property
-	def Name(self):
-		if ':' in self.state.Name:
-			return self.state.Name.split(':')[0]
-		return self.state.Name
-
 	@property
 	def Tags(self):
 		return utils.parse_tags(self.state.tags)
--- a/daemons/lvmetad/.gitignore
+++ b/daemons/lvmetad/.gitignore
@@ -0,0 +1,2 @@
+lvmetad
+lvmetactl
--- a/daemons/lvmetad/Makefile.in
+++ b/daemons/lvmetad/Makefile.in
@@ -0,0 +1,62 @@
+#
+# Copyright (C) 2011-2012 Red Hat, Inc.
+#
+# 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 Lesser General Public License v.2.1.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program; if not, write to the Free Software Foundation,
+# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+top_builddir = @top_builddir@
+
+SOURCES = lvmetad-core.c
+SOURCES2 = lvmetactl.c
+
+TARGETS = lvmetad lvmetactl
+
+.PHONY: install_lvmetad
+
+CFLOW_LIST = $(SOURCES)
+CFLOW_LIST_TARGET = $(LIB_NAME).cflow
+CFLOW_TARGET = lvmetad
+
+include $(top_builddir)/make.tmpl
+
+CFLAGS_lvmetactl.o += $(EXTRA_EXEC_CFLAGS)
+CFLAGS_lvmetad-core.o += $(EXTRA_EXEC_CFLAGS)
+INCLUDES += -I$(top_srcdir)/libdaemon/server
+LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
+LIBS += $(RT_LIBS) $(DAEMON_LIBS) -ldevmapper $(PTHREAD_LIBS)
+
+lvmetad: $(OBJECTS) $(top_builddir)/libdaemon/client/libdaemonclient.a \
+		    $(top_builddir)/libdaemon/server/libdaemonserver.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) -ldaemonserver $(LIBS)
+
+lvmetactl: lvmetactl.o $(top_builddir)/libdaemon/client/libdaemonclient.a \
+	$(top_builddir)/libdaemon/server/libdaemonserver.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ lvmetactl.o $(LIBS)
+
+CLEAN_TARGETS += lvmetactl.o
+
+# TODO: No idea. No idea how to test either.
+#ifneq ("$(CFLOW_CMD)", "")
+#CFLOW_SOURCES = $(addprefix $(srcdir)/, $(SOURCES))
+#-include $(top_builddir)/libdm/libdevmapper.cflow
+#-include $(top_builddir)/lib/liblvm-internal.cflow
+#-include $(top_builddir)/lib/liblvm2cmd.cflow
+#-include $(top_builddir)/daemons/dmeventd/$(LIB_NAME).cflow
+#-include $(top_builddir)/daemons/dmeventd/plugins/mirror/$(LIB_NAME)-lvm2mirror.cflow
+#endif
+
+install_lvmetad: lvmetad
+	$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
+
+install_lvm2: install_lvmetad
+
+install: install_lvm2
--- a/daemons/lvmetad/lvmetactl.c
+++ b/daemons/lvmetad/lvmetactl.c
@@ -0,0 +1,249 @@
+/*
+ * Copyright (C) 2014 Red Hat, Inc.
+ *
+ * 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 Lesser General Public License v.2.1.
+ */
+
+#include "tool.h"
+
+#include "lvmetad-client.h"
+
+daemon_handle h;
+
+static void print_reply(daemon_reply reply)
+{
+	const char *a = daemon_reply_str(reply, "response", NULL);
+	const char *b = daemon_reply_str(reply, "status", NULL);
+	const char *c = daemon_reply_str(reply, "reason", NULL);
+
+	printf("response \"%s\" status \"%s\" reason \"%s\"\n",
+	       a ? a : "", b ? b : "", c ? c : "");
+}
+
+int main(int argc, char **argv)
+{
+	daemon_reply reply;
+	char *cmd;
+	char *uuid;
+	char *name;
+	int val;
+	int ver;
+
+	if (argc < 2) {
+		printf("lvmetactl dump\n");
+		printf("lvmetactl pv_list\n");
+		printf("lvmetactl vg_list\n");
+		printf("lvmetactl get_global_info\n");
+		printf("lvmetactl vg_lookup_name <name>\n");
+		printf("lvmetactl vg_lookup_uuid <uuid>\n");
+		printf("lvmetactl pv_lookup_uuid <uuid>\n");
+		printf("lvmetactl set_global_invalid 0|1\n");
+		printf("lvmetactl set_global_disable 0|1\n");
+		printf("lvmetactl set_vg_version <uuid> <name> <version>\n");
+		printf("lvmetactl vg_lock_type <uuid>\n");
+		return -1;
+	}
+
+	cmd = argv[1];
+
+	h = lvmetad_open(NULL);
+
+	if (!strcmp(cmd, "dump")) {
+		reply = daemon_send_simple(h, "dump",
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else if (!strcmp(cmd, "pv_list")) {
+		reply = daemon_send_simple(h, "pv_list",
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else if (!strcmp(cmd, "vg_list")) {
+		reply = daemon_send_simple(h, "vg_list",
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else if (!strcmp(cmd, "get_global_info")) {
+		reply = daemon_send_simple(h, "get_global_info",
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else if (!strcmp(cmd, "set_global_invalid")) {
+		if (argc < 3) {
+			printf("set_global_invalid 0|1\n");
+			return -1;
+		}
+		val = atoi(argv[2]);
+
+		reply = daemon_send_simple(h, "set_global_info",
+					   "global_invalid = " FMTd64, (int64_t) val,
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		print_reply(reply);
+
+	} else if (!strcmp(cmd, "set_global_disable")) {
+		if (argc < 3) {
+			printf("set_global_disable 0|1\n");
+			return -1;
+		}
+		val = atoi(argv[2]);
+
+		reply = daemon_send_simple(h, "set_global_info",
+					   "global_disable = " FMTd64, (int64_t) val,
+					   "disable_reason = %s", LVMETAD_DISABLE_REASON_DIRECT,
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		print_reply(reply);
+
+	} else if (!strcmp(cmd, "set_vg_version")) {
+		if (argc < 5) {
+			printf("set_vg_version <uuid> <name> <ver>\n");
+			return -1;
+		}
+		uuid = argv[2];
+		name = argv[3];
+		ver = atoi(argv[4]);
+
+		if ((strlen(uuid) == 1) && (uuid[0] == '-'))
+			uuid = NULL;
+		if ((strlen(name) == 1) && (name[0] == '-'))
+			name = NULL;
+
+		if (uuid && name) {
+			reply = daemon_send_simple(h, "set_vg_info",
+						   "uuid = %s", uuid,
+						   "name = %s", name,
+						   "version = " FMTd64, (int64_t) ver,
+						   "token = %s", "skip",
+						   "pid = " FMTd64, (int64_t)getpid(),
+						   "cmd = %s", "lvmetactl",
+						   NULL);
+		} else if (uuid) {
+			reply = daemon_send_simple(h, "set_vg_info",
+						   "uuid = %s", uuid,
+						   "version = " FMTd64, (int64_t) ver,
+						   "token = %s", "skip",
+						   "pid = " FMTd64, (int64_t)getpid(),
+						   "cmd = %s", "lvmetactl",
+						   NULL);
+		} else if (name) {
+			reply = daemon_send_simple(h, "set_vg_info",
+						   "name = %s", name,
+						   "version = " FMTd64, (int64_t) ver,
+						   "token = %s", "skip",
+						   "pid = " FMTd64, (int64_t)getpid(),
+						   "cmd = %s", "lvmetactl",
+						   NULL);
+		} else {
+			printf("name or uuid required\n");
+			return -1;
+		}
+
+		print_reply(reply);
+
+	} else if (!strcmp(cmd, "vg_lookup_name")) {
+		if (argc < 3) {
+			printf("vg_lookup_name <name>\n");
+			return -1;
+		}
+		name = argv[2];
+
+		reply = daemon_send_simple(h, "vg_lookup",
+					   "name = %s", name,
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else if (!strcmp(cmd, "vg_lookup_uuid")) {
+		if (argc < 3) {
+			printf("vg_lookup_uuid <uuid>\n");
+			return -1;
+		}
+		uuid = argv[2];
+
+		reply = daemon_send_simple(h, "vg_lookup",
+					   "uuid = %s", uuid,
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else if (!strcmp(cmd, "vg_lock_type")) {
+		struct dm_config_node *metadata;
+		const char *lock_type;
+
+		if (argc < 3) {
+			printf("vg_lock_type <uuid>\n");
+			return -1;
+		}
+		uuid = argv[2];
+
+		reply = daemon_send_simple(h, "vg_lookup",
+					   "uuid = %s", uuid,
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		/* printf("%s\n", reply.buffer.mem); */
+
+		metadata = dm_config_find_node(reply.cft->root, "metadata");
+		if (!metadata) {
+			printf("no metadata\n");
+			goto out;
+		}
+
+		lock_type = dm_config_find_str(metadata, "metadata/lock_type", NULL);
+		if (!lock_type) {
+			printf("no lock_type\n");
+			goto out;
+		}
+		printf("lock_type %s\n", lock_type);
+
+	} else if (!strcmp(cmd, "pv_lookup_uuid")) {
+		if (argc < 3) {
+			printf("pv_lookup_uuid <uuid>\n");
+			return -1;
+		}
+		uuid = argv[2];
+
+		reply = daemon_send_simple(h, "pv_lookup",
+					   "uuid = %s", uuid,
+					   "token = %s", "skip",
+					   "pid = " FMTd64, (int64_t)getpid(),
+					   "cmd = %s", "lvmetactl",
+					   NULL);
+		printf("%s\n", reply.buffer.mem);
+
+	} else {
+		printf("unknown command\n");
+		goto out_close;
+	}
+out:
+	daemon_reply_destroy(reply);
+out_close:
+	daemon_close(h);
+	return 0;
+}
--- a/daemons/lvmetad/lvmetad-client.h
+++ b/daemons/lvmetad/lvmetad-client.h
@@ -0,0 +1,91 @@
+/*
+ * Copyright (C) 2011-2012 Red Hat, Inc.
+ *
+ * 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 Lesser General Public License v.2.1.
+ *
+ * You should have received a copy of the GNU Lesser General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#ifndef _LVM_LVMETAD_CLIENT_H
+#define _LVM_LVMETAD_CLIENT_H
+
+#include "daemon-client.h"
+
+#define LVMETAD_SOCKET DEFAULT_RUN_DIR "/lvmetad.socket"
+
+#define LVMETAD_TOKEN_UPDATE_IN_PROGRESS "update in progress"
+
+#define LVMETAD_DISABLE_REASON_DIRECT		"DIRECT"
+#define LVMETAD_DISABLE_REASON_LVM1		"LVM1"
+#define LVMETAD_DISABLE_REASON_DUPLICATES	"DUPLICATES"
+#define LVMETAD_DISABLE_REASON_VGRESTORE	"VGRESTORE"
+#define LVMETAD_DISABLE_REASON_REPAIR		"REPAIR"
+
+struct volume_group;
+
+/* Different types of replies we may get from lvmetad. */
+
+typedef struct {
+	daemon_reply r;
+	const char **uuids; /* NULL terminated array */
+} lvmetad_uuidlist;
+
+typedef struct {
+	daemon_reply r;
+	struct dm_config_tree *cft;
+} lvmetad_vg;
+
+/* Get a list of VG UUIDs that match a given VG name. */
+lvmetad_uuidlist lvmetad_lookup_vgname(daemon_handle h, const char *name);
+
+/* Get the metadata of a single VG, identified by UUID. */
+lvmetad_vg lvmetad_get_vg(daemon_handle h, const char *uuid);
+
+/*
+ * Add and remove PVs on demand. Udev-driven systems will use this interface
+ * instead of scanning.
+ */
+daemon_reply lvmetad_add_pv(daemon_handle h, const char *pv_uuid, const char *mda_content);
+daemon_reply lvmetad_remove_pv(daemon_handle h, const char *pv_uuid);
+
+/* Trigger a full disk scan, throwing away all caches. XXX do we eventually want
+ * this? Probably not yet, anyway.
+ *     daemon_reply lvmetad_rescan(daemon_handle h);
+ */
+
+/*
+ * Update the version of metadata of a volume group. The VG has to be locked for
+ * writing for this, and the VG metadata here has to match whatever has been
+ * written to the disk (under this lock). This initially avoids the requirement
+ * for lvmetad to write to disk (in later revisions, lvmetad_supersede_vg may
+ * also do the writing, or we probably add another function to do that).
+ */
+daemon_reply lvmetad_supersede_vg(daemon_handle h, struct volume_group *vg);
+
+/* Wrappers to open/close connection */
+
+static inline daemon_handle lvmetad_open(const char *socket)
+{
+	daemon_info lvmetad_info = {
+		.path = "lvmetad",
+		.socket = socket ?: LVMETAD_SOCKET,
+		.protocol = "lvmetad",
+		.protocol_version = 1,
+		.autostart = 0
+	};
+
+	return daemon_open(lvmetad_info);
+}
+
+static inline void lvmetad_close(daemon_handle h)
+{
+	return daemon_close(h);
+}
+
+#endif
--- a/daemons/lvmetad/lvmetad-core.c
+++ b/daemons/lvmetad/lvmetad-core.c
--- a/daemons/lvmetad/test.sh
+++ b/daemons/lvmetad/test.sh
@@ -0,0 +1,16 @@
+#!/bin/bash
+
+export LD_LIBRARY_PATH="$1"
+
+test -n "$2" && {
+    rm -f /var/run/lvmetad.{socket,pid}
+    chmod +rx lvmetad
+    valgrind ./lvmetad -f &
+    PID=$!
+    sleep 1
+    ./testclient
+    kill $PID
+    exit 0
+}
+
+sudo ./test.sh "$1" .
--- a/daemons/lvmetad/testclient.c
+++ b/daemons/lvmetad/testclient.c
@@ -0,0 +1,147 @@
+/*
+ * Copyright (C) 2011-2014 Red Hat, Inc.
+ *
+ * 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+#include "tool.h"
+
+#include "lvmetad-client.h"
+#include "label.h"
+#include "lvmcache.h"
+#include "metadata.h"
+
+const char *uuid1 = "abcd-efgh";
+const char *uuid2 = "bbcd-efgh";
+const char *vgid = "yada-yada";
+const char *uuid3 = "cbcd-efgh";
+
+const char *metadata2 = "{\n"
+	"id = \"yada-yada\"\n"
+	"seqno = 15\n"
+	"status = [\"READ\", \"WRITE\"]\n"
+	"flags = []\n"
+	"extent_size = 8192\n"
+	"physical_volumes {\n"
+	"    pv0 {\n"
+	"        id = \"abcd-efgh\"\n"
+	"    }\n"
+	"    pv1 {\n"
+	"        id = \"bbcd-efgh\"\n"
+	"    }\n"
+	"    pv2 {\n"
+	"        id = \"cbcd-efgh\"\n"
+	"    }\n"
+	"}\n"
+	"}\n";
+
+void _handle_reply(daemon_reply reply) {
+	const char *repl = daemon_reply_str(reply, "response", NULL);
+	const char *status = daemon_reply_str(reply, "status", NULL);
+	const char *vgid = daemon_reply_str(reply, "vgid", NULL);
+
+	fprintf(stderr, "[C] REPLY: %s\n", repl);
+	if (!strcmp(repl, "failed"))
+		fprintf(stderr, "[C] REASON: %s\n", daemon_reply_str(reply, "reason", "unknown"));
+	if (vgid)
+		fprintf(stderr, "[C] VGID: %s\n", vgid);
+	if (status)
+		fprintf(stderr, "[C] STATUS: %s\n", status);
+	daemon_reply_destroy(reply);
+}
+
+void _pv_add(daemon_handle h, const char *uuid, const char *metadata)
+{
+	daemon_reply reply = daemon_send_simple(h, "pv_add", "uuid = %s", uuid,
+						             "metadata = %b", metadata,
+						             NULL);
+	_handle_reply(reply);
+}
+
+int scan(daemon_handle h, char *fn) {
+	struct device *dev = dev_cache_get(fn, NULL);
+
+	struct label *label;
+	if (!label_read(dev, &label, 0)) {
+		fprintf(stderr, "[C] no label found on %s\n", fn);
+		return;
+	}
+
+	char uuid[64];
+	if (!id_write_format(dev->pvid, uuid, 64)) {
+		fprintf(stderr, "[C] Failed to format PV UUID for %s", dev_name(dev));
+		return;
+	}
+	fprintf(stderr, "[C] found PV: %s\n", uuid);
+	struct lvmcache_info *info = (struct lvmcache_info *) label->info;
+	struct physical_volume pv = { 0, };
+
+	if (!(info->fmt->ops->pv_read(info->fmt, dev_name(dev), &pv, 0))) {
+		fprintf(stderr, "[C] Failed to read PV %s", dev_name(dev));
+		return;
+	}
+
+	struct format_instance_ctx fic;
+	struct format_instance *fid = info->fmt->ops->create_instance(info->fmt, &fic);
+	struct metadata_area *mda;
+	struct volume_group *vg = NULL;
+	dm_list_iterate_items(mda, &info->mdas) {
+		struct volume_group *this = mda->ops->vg_read(fid, "", mda);
+		if (this && !vg || this->seqno > vg->seqno)
+			vg = this;
+	}
+	if (vg) {
+		char *buf = NULL;
+		/* TODO. This is not entirely correct, since export_vg_to_buffer
+		 * adds trailing garbage to the buffer. We may need to use
+		 * export_vg_to_config_tree and format the buffer ourselves. It
+		 * does, however, work for now, since the garbage is well
+		 * formatted and has no conflicting keys with the rest of the
+		 * request.  */
+		export_vg_to_buffer(vg, &buf);
+		daemon_reply reply =
+			daemon_send_simple(h, "pv_add", "uuid = %s", uuid,
+					      "metadata = %b", strchr(buf, '{'),
+					      NULL);
+		_handle_reply(reply);
+	}
+}
+
+void _dump_vg(daemon_handle h, const char *uuid)
+{
+	daemon_reply reply = daemon_send_simple(h, "vg_by_uuid", "uuid = %s", uuid, NULL);
+	fprintf(stderr, "[C] reply buffer: %s\n", reply.buffer);
+	daemon_reply_destroy(reply);
+}
+
+int main(int argc, char **argv) {
+	daemon_handle h = lvmetad_open();
+	/* FIXME Missing error path */
+
+	if (argc > 1) {
+		int i;
+		struct cmd_context *cmd = create_toolcontext(0, NULL, 0, 0, 1, 1);
+		for (i = 1; i < argc; ++i) {
+			const char *uuid = NULL;
+			scan(h, argv[i]);
+		}
+		destroy_toolcontext(cmd);
+		/* FIXME Missing lvmetad_close() */
+		return 0;
+	}
+
+	_pv_add(h, uuid1, NULL);
+	_pv_add(h, uuid2, metadata2);
+	_dump_vg(h, vgid);
+	_pv_add(h, uuid3, NULL);
+
+	daemon_close(h);	/* FIXME lvmetad_close? */
+	return 0;
+}
--- a/daemons/lvmlockd/Makefile.in
+++ b/daemons/lvmlockd/Makefile.in
@@ -15,8 +15,6 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 top_builddir = @top_builddir@

-USE_SD_NOTIFY=yes
-
 SOURCES = lvmlockd-core.c

 ifeq ("@BUILD_LOCKDSANLOCK@", "yes")
@@ -27,37 +25,31 @@ endif
 ifeq ("@BUILD_LOCKDDLM@", "yes")
  SOURCES += lvmlockd-dlm.c
  LOCK_LIBS += -ldlm_lt
-  LOCK_LIBS += -ldlmcontrol
 endif

-SOURCES2 = lvmlockctl.c
-
 TARGETS = lvmlockd lvmlockctl

 .PHONY: install_lvmlockd

 include $(top_builddir)/make.tmpl

-CFLAGS += $(EXTRA_EXEC_CFLAGS) $(SYSTEMD_CFLAGS)
+CFLAGS += $(EXTRA_EXEC_CFLAGS)
 INCLUDES += -I$(top_srcdir)/libdaemon/server
-LDFLAGS += $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
-LIBS += $(PTHREAD_LIBS) $(SYSTEMD_LIBS)
+LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
+LIBS += $(RT_LIBS) $(DAEMON_LIBS) -ldevmapper $(PTHREAD_LIBS)

-lvmlockd: $(OBJECTS) $(top_builddir)/libdaemon/server/libdaemonserver.a $(INTERNAL_LIBS)
-	@echo "    [CC] $@"
-	$(Q) $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $+ $(LOCK_LIBS) $(LIBS)
+lvmlockd: $(OBJECTS) $(top_builddir)/libdaemon/client/libdaemonclient.a \
+		    $(top_builddir)/libdaemon/server/libdaemonserver.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LOCK_LIBS) -ldaemonserver $(LIBS)

-lvmlockctl: lvmlockctl.o $(INTERNAL_LIBS)
-	@echo "    [CC] $@"
-	$(Q) $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $+ $(LIBS)
+lvmlockctl: lvmlockctl.o $(top_builddir)/libdaemon/client/libdaemonclient.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ lvmlockctl.o $(LIBS)

 install_lvmlockd: lvmlockd
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
+	$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)

 install_lvmlockctl: lvmlockctl
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
+	$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)

 install_lvm2: install_lvmlockd install_lvmlockctl

--- a/daemons/lvmlockd/lvmlockctl.c
+++ b/daemons/lvmlockd/lvmlockctl.c
@@ -8,9 +8,9 @@
 * of the GNU Lesser General Public License v.2.1.
 */

-#include "tools/tool.h"
+#include "tool.h"

-#include "daemons/lvmlockd/lvmlockd-client.h"
+#include "lvmlockd-client.h"

 #include <stddef.h>
 #include <getopt.h>
@@ -24,7 +24,7 @@
 static int quit = 0;
 static int info = 0;
 static int dump = 0;
-static int wait_opt = 1;
+static int wait_opt = 0;
 static int force_opt = 0;
 static int kill_vg = 0;
 static int drop_vg = 0;
--- a/daemons/lvmlockd/lvmlockd-client.h
+++ b/daemons/lvmlockd/lvmlockd-client.h
@@ -11,7 +11,7 @@
 #ifndef _LVM_LVMLOCKD_CLIENT_H
 #define _LVM_LVMLOCKD_CLIENT_H

-#include "libdaemon/client/daemon-client.h"
+#include "daemon-client.h"

 #define LVMLOCKD_SOCKET DEFAULT_RUN_DIR "/lvmlockd.socket"

@@ -49,6 +49,5 @@ static inline void lvmlockd_close(daemon_handle h)
 #define ELOCKIO   218 /* sanlock io errors during lock op, may be transient. */
 #define EREMOVED  219
 #define EDEVOPEN  220 /* sanlock failed to open lvmlock LV */
-#define ELMERR    221

 #endif	/* _LVM_LVMLOCKD_CLIENT_H */
--- a/daemons/lvmlockd/lvmlockd-core.c
+++ b/daemons/lvmlockd/lvmlockd-core.c
@@ -8,13 +8,18 @@
 * of the GNU Lesser General Public License v.2.1.
 */

-#include "tools/tool.h"
+#define _XOPEN_SOURCE 500  /* pthread */
+#define _ISOC99_SOURCE
+#define _REENTRANT

-#include "libdaemon/client/daemon-io.h"
+#include "tool.h"
+
+#include "daemon-io.h"
 #include "daemon-server.h"
 #include "lvm-version.h"
-#include "daemons/lvmlockd/lvmlockd-client.h"
-#include "device_mapper/misc/dm-ioctl.h"
+#include "lvmetad-client.h"
+#include "lvmlockd-client.h"
+#include "dm-ioctl.h" /* for DM_UUID_LEN */

 /* #include <assert.h> */
 #include <errno.h>
@@ -31,10 +36,6 @@
 #include <sys/utsname.h>
 #include <sys/un.h>

-#ifdef NOTIFYDBUS_SUPPORT
-#include <systemd/sd-daemon.h>
-#endif
-
 #define EXTERN
 #include "lvmlockd-internal.h"

@@ -143,6 +144,10 @@ static const int lvmlockd_protocol_version = 1;
 static int daemon_quit;
 static int adopt_opt;

+static daemon_handle lvmetad_handle;
+static pthread_mutex_t lvmetad_mutex;
+static int lvmetad_connected;
+
 /*
 * We use a separate socket for dumping daemon info.
 * This will not interfere with normal operations, and allows
@@ -405,11 +410,12 @@ struct lockspace *alloc_lockspace(void)
 {
 	struct lockspace *ls;

-	if (!(ls = zalloc(sizeof(struct lockspace)))) {
+	if (!(ls = malloc(sizeof(struct lockspace)))) {
 		log_error("out of memory for lockspace");
 		return NULL;
 	}

+	memset(ls, 0, sizeof(struct lockspace));
 	INIT_LIST_HEAD(&ls->actions);
 	INIT_LIST_HEAD(&ls->resources);
 	pthread_mutex_init(&ls->mutex, NULL);
@@ -502,10 +508,6 @@ static struct lock *alloc_lock(void)

 static void free_action(struct action *act)
 {
-	if (act->path) {
-		free(act->path);
-		act->path = NULL;
-	}
 	pthread_mutex_lock(&unused_struct_mutex);
 	if (unused_action_count >= MAX_UNUSED_ACTION) {
 		free(act);
@@ -729,8 +731,6 @@ static const char *op_str(int x)
 		return "rename_final";
 	case LD_OP_RUNNING_LM:
 		return "running_lm";
-	case LD_OP_QUERY_LOCK:
-		return "query_lock";
 	case LD_OP_FIND_FREE_LOCK:
 		return "find_free_lock";
 	case LD_OP_KILL_VG:
@@ -743,8 +743,6 @@ static const char *op_str(int x)
 		return "dump_info";
 	case LD_OP_BUSY:
 		return "busy";
-	case LD_OP_REFRESH_LV:
-		return "refresh_lv";
 	default:
 		return "op_unknown";
 	};
@@ -932,12 +930,12 @@ static void lm_rem_resource(struct lockspace *ls, struct resource *r)
 		lm_rem_resource_sanlock(ls, r);
 }

-static int lm_find_free_lock(struct lockspace *ls, uint64_t *free_offset, int *sector_size, int *align_size)
+static int lm_find_free_lock(struct lockspace *ls, uint64_t *free_offset)
 {
 	if (ls->lm_type == LD_LM_DLM)
 		return 0;
 	else if (ls->lm_type == LD_LM_SANLOCK)
-		return lm_find_free_lock_sanlock(ls, free_offset, sector_size, align_size);
+		return lm_find_free_lock_sanlock(ls, free_offset);
 	return -1;
 }

@@ -1011,6 +1009,52 @@ static void add_work_action(struct action *act)
 	pthread_mutex_unlock(&worker_mutex);
 }

+static daemon_reply send_lvmetad(const char *id, ...)
+{
+	daemon_reply reply;
+	va_list ap;
+	int retries = 0;
+	int err;
+
+	va_start(ap, id);
+
+	/*
+	 * mutex is used because all threads share a single
+	 * lvmetad connection/handle.
+	 */
+	pthread_mutex_lock(&lvmetad_mutex);
+retry:
+	if (!lvmetad_connected) {
+		lvmetad_handle = lvmetad_open(NULL);
+		if (lvmetad_handle.error || lvmetad_handle.socket_fd < 0) {
+			err = lvmetad_handle.error ?: lvmetad_handle.socket_fd;
+			pthread_mutex_unlock(&lvmetad_mutex);
+			log_error("lvmetad_open reconnect error %d", err);
+			memset(&reply, 0, sizeof(reply));
+			reply.error = err;
+			va_end(ap);
+			return reply;
+		} else {
+			log_debug("lvmetad reconnected");
+			lvmetad_connected = 1;
+		}
+	}
+
+	reply = daemon_send_simple_v(lvmetad_handle, id, ap);
+
+	/* lvmetad may have been restarted */
+	if ((reply.error == ECONNRESET) && (retries < 2)) {
+		daemon_close(lvmetad_handle);
+		lvmetad_connected = 0;
+		retries++;
+		goto retry;
+	}
+	pthread_mutex_unlock(&lvmetad_mutex);
+
+	va_end(ap);
+	return reply;
+}
+
 static int res_lock(struct lockspace *ls, struct resource *r, struct action *act, int *retry)
 {
 	struct lock *lk;
@@ -1206,18 +1250,6 @@ static int res_lock(struct lockspace *ls, struct resource *r, struct action *act
 		rv = -EREMOVED;
 	}

-	/*
-	 * lvmetad is no longer used, but the infrastructure for
-	 * distributed cache validation remains.  The points
-	 * where vg or global cache state would be invalidated
-	 * remain below and log_debug messages point out where
-	 * they would occur.
-	 *
-	 * The comments related to "lvmetad" remain because they
-	 * describe how some other local cache like lvmetad would
-	 * be invalidated here.
-	 */
-
 	/*
 	 * r is vglk: tell lvmetad to set the vg invalid
 	 * flag, and provide the new r_version.  If lvmetad finds
@@ -1233,22 +1265,44 @@ static int res_lock(struct lockspace *ls, struct resource *r, struct action *act
 	 * caches, and tell lvmetad to set global invalid to 0.
 	 */

-	/*
-	 * lvmetad not running:
-	 * Even if we have not previously found lvmetad running,
-	 * we attempt to connect and invalidate in case it has
-	 * been started while lvmlockd is running.  We don't
-	 * want to allow lvmetad to be used with invalid data if
-	 * it happens to be enabled and started after lvmlockd.
-	 */
-
 	if (inval_meta && (r->type == LD_RT_VG)) {
-		log_debug("S %s R %s res_lock invalidate vg state version %u",
+		daemon_reply reply;
+		char *uuid;
+
+		log_debug("S %s R %s res_lock set lvmetad vg version %u",
 			  ls->name, r->name, new_version);
+	
+		if (!ls->vg_uuid[0] || !strcmp(ls->vg_uuid, "none"))
+			uuid = (char *)"none";
+		else
+			uuid = ls->vg_uuid;
+
+		reply = send_lvmetad("set_vg_info",
+				     "token = %s", "skip",
+				     "uuid = %s", uuid,
+				     "name = %s", ls->vg_name,
+				     "version = " FMTd64, (int64_t)new_version,
+				     NULL);
+
+		if (reply.error || strcmp(daemon_reply_str(reply, "response", ""), "OK"))
+			log_error("set_vg_info in lvmetad failed %d", reply.error);
+		daemon_reply_destroy(reply);
 	}

 	if (inval_meta && (r->type == LD_RT_GL)) {
-		log_debug("S %s R %s res_lock invalidate global state", ls->name, r->name);
+		daemon_reply reply;
+
+		log_debug("S %s R %s res_lock set lvmetad global invalid",
+			  ls->name, r->name);
+
+		reply = send_lvmetad("set_global_info",
+				     "token = %s", "skip",
+				     "global_invalid = " FMTd64, INT64_C(1),
+				     NULL);
+
+		if (reply.error || strcmp(daemon_reply_str(reply, "response", ""), "OK"))
+			log_error("set_global_info in lvmetad failed %d", reply.error);
+		daemon_reply_destroy(reply);
 	}

 	/*
@@ -1335,11 +1389,12 @@ static int res_convert(struct lockspace *ls, struct resource *r,
 	}

 	rv = lm_convert(ls, r, act->mode, act, r_version);
-
-	log_debug("S %s R %s res_convert rv %d", ls->name, r->name, rv);
-
-	if (rv < 0)
+	if (rv < 0) {
+		log_error("S %s R %s res_convert lm error %d", ls->name, r->name, rv);
 		return rv;
+	}
+
+	log_debug("S %s R %s res_convert lm done", ls->name, r->name);

 	if (lk->mode == LD_LK_EX && act->mode == LD_LK_SH) {
 		r->sh_count = 1;
@@ -1822,9 +1877,9 @@ static void res_process(struct lockspace *ls, struct resource *r,
 			add_client_result(act);
 		} else {
 			/* persistent lock is sh, transient request is ex */
+			/* FIXME: can we remove this case? do a convert here? */
 			log_debug("res_process %s existing persistent lock new transient", r->name);
 			r->last_client_id = act->client_id;
-			act->flags |= LD_AF_SH_EXISTS;
 			act->result = -EEXIST;
 			list_del(&act->list);
 			add_client_result(act);
@@ -2204,7 +2259,6 @@ static int process_op_during_kill(struct action *act)
 	case LD_OP_UPDATE:
 	case LD_OP_RENAME_BEFORE:
 	case LD_OP_RENAME_FINAL:
-	case LD_OP_QUERY_LOCK:
 	case LD_OP_FIND_FREE_LOCK:
 		return 0;
 	};
@@ -2230,7 +2284,7 @@ static void *lockspace_thread_main(void *arg_in)
 	struct action *act_op_free = NULL;
 	struct list_head tmp_act;
 	struct list_head act_close;
-	char tmp_name[MAX_NAME+5];
+	char tmp_name[MAX_NAME+1];
 	int free_vg = 0;
 	int drop_vg = 0;
 	int error = 0;
@@ -2429,31 +2483,13 @@ static void *lockspace_thread_main(void *arg_in)
 				break;
 			}

-			if (act->op == LD_OP_QUERY_LOCK) {
-				r = find_resource_act(ls, act, 0);
-				if (!r)
-					act->result = -ENOENT;
-				else {
-					act->result = 0;
-					act->mode = r->mode;
-				}
-				list_del(&act->list);
-				add_client_result(act);
-				continue;
-			}
-
 			if (act->op == LD_OP_FIND_FREE_LOCK && act->rt == LD_RT_VG) {
 				uint64_t free_offset = 0;
-				int sector_size = 0;
-				int align_size = 0;
-
 				log_debug("S %s find free lock", ls->name);
-				rv = lm_find_free_lock(ls, &free_offset, &sector_size, &align_size);
-				log_debug("S %s find free lock %d offset %llu sector_size %d align_size %d",
-					  ls->name, rv, (unsigned long long)free_offset, sector_size, align_size);
+				rv = lm_find_free_lock(ls, &free_offset);
+				log_debug("S %s find free lock %d offset %llu",
+					  ls->name, rv, (unsigned long long)free_offset);
 				ls->free_lock_offset = free_offset;
-				ls->free_lock_sector_size = sector_size;
-				ls->free_lock_align_size = align_size;
 				list_del(&act->list);
 				act->result = rv;
 				add_client_result(act);
@@ -2624,10 +2660,8 @@ out_act:
 	 * blank or fill it with garbage, but instead set it to REM:<name>
 	 * to make it easier to follow progress of freeing is via log_debug.
 	 */
-	memset(tmp_name, 0, sizeof(tmp_name));
-	memcpy(tmp_name, "REM:", 4);
-	strncpy(tmp_name+4, ls->name, sizeof(tmp_name)-4);
-	memcpy(ls->name, tmp_name, sizeof(ls->name));
+	dm_strncpy(tmp_name, ls->name, sizeof(tmp_name));
+	snprintf(ls->name, sizeof(ls->name), "REM:%s", tmp_name);
 	pthread_mutex_unlock(&lockspaces_mutex);

 	/* worker_thread will join this thread, and free the ls */
@@ -2767,9 +2801,6 @@ static int add_lockspace_thread(const char *ls_name,
 		if (ls2->thread_stop) {
 			log_debug("add_lockspace_thread %s exists and stopping", ls->name);
 			rv = -EAGAIN;
-		} else if (!ls2->create_fail && !ls2->create_done) {
-			log_debug("add_lockspace_thread %s exists and starting", ls->name);
-			rv = -ESTARTING;
 		} else {
 			log_debug("add_lockspace_thread %s exists", ls->name);
 			rv = -EEXIST;
@@ -3011,7 +3042,7 @@ static int count_lockspace_starting(uint32_t client_id)

 	pthread_mutex_lock(&lockspaces_mutex);
 	list_for_each_entry(ls, &lockspaces, list) {
-		if (client_id && (ls->start_client_id != client_id))
+		if (ls->start_client_id != client_id)
 			continue;

 		if (!ls->create_done && !ls->create_fail) {
@@ -3264,8 +3295,6 @@ static int work_init_lv(struct action *act)
 	char vg_args[MAX_ARGS+1];
 	char lv_args[MAX_ARGS+1];
 	uint64_t free_offset = 0;
-	int sector_size = 0;
-	int align_size = 0;
 	int lm_type = 0;
 	int rv = 0;

@@ -3281,8 +3310,6 @@ static int work_init_lv(struct action *act)
 		lm_type = ls->lm_type;
 		memcpy(vg_args, ls->vg_args, MAX_ARGS);
 		free_offset = ls->free_lock_offset;
-		sector_size = ls->free_lock_sector_size;
-		align_size = ls->free_lock_align_size;
 	}
 	pthread_mutex_unlock(&lockspaces_mutex);

@@ -3299,7 +3326,7 @@ static int work_init_lv(struct action *act)

 	if (lm_type == LD_LM_SANLOCK) {
 		rv = lm_init_lv_sanlock(ls_name, act->vg_name, act->lv_uuid,
-					vg_args, lv_args, sector_size, align_size, free_offset);
+					vg_args, lv_args, free_offset);

 		memcpy(act->lv_args, lv_args, MAX_ARGS);
 		return rv;
@@ -3416,7 +3443,7 @@ static void *worker_thread_main(void *arg_in)
 			add_client_result(act);

 		} else if (act->op == LD_OP_START_WAIT) {
-			act->result = count_lockspace_starting(0);
+			act->result = count_lockspace_starting(act->client_id);
 			if (!act->result)
 				add_client_result(act);
 			else
@@ -3429,15 +3456,6 @@ static void *worker_thread_main(void *arg_in)
 			else
 				list_add(&act->list, &delayed_list);

-		} else if (act->op == LD_OP_REFRESH_LV) {
-			log_debug("work refresh_lv %s %s", act->lv_uuid, act->path);
-			rv = lm_refresh_lv_start_dlm(act);
-			if (rv < 0) {
-				act->result = rv;
-				add_client_result(act);
-			} else
-				list_add(&act->list, &delayed_list);
-
 		} else {
 			log_error("work unknown op %d", act->op);
 			act->result = -EINVAL;
@@ -3459,7 +3477,7 @@ static void *worker_thread_main(void *arg_in)
 		list_for_each_entry_safe(act, safe, &delayed_list, list) {
 			if (act->op == LD_OP_START_WAIT) {
 				log_debug("work delayed start_wait for client %u", act->client_id);
-				act->result = count_lockspace_starting(0);
+				act->result = count_lockspace_starting(act->client_id);
 				if (!act->result) {
 					list_del(&act->list);
 					add_client_result(act);
@@ -3473,19 +3491,6 @@ static void *worker_thread_main(void *arg_in)
 					act->result = 0;
 					add_client_result(act);
 				}
-
-			} else if (act->op == LD_OP_REFRESH_LV) {
-				log_debug("work delayed refresh_lv");
-				rv = lm_refresh_lv_check_dlm(act);
-				if (!rv) {
-					list_del(&act->list);
-					act->result = 0;
-					add_client_result(act);
-				} else if ((rv < 0) && (rv != -EAGAIN)) {
-					list_del(&act->list);
-					act->result = rv;
-					add_client_result(act);
-				}
 			}
 		}

@@ -3691,9 +3696,6 @@ static int client_send_result(struct client *cl, struct action *act)
 	if ((act->flags & LD_AF_WARN_GL_REMOVED) || gl_vg_removed)
 		strcat(result_flags, "WARN_GL_REMOVED,");
 	
-	if (act->flags & LD_AF_SH_EXISTS)
-		strcat(result_flags, "SH_EXISTS,");
-
 	if (act->op == LD_OP_INIT) {
 		/*
 		 * init is a special case where lock args need
@@ -3722,20 +3724,6 @@ static int client_send_result(struct client *cl, struct action *act)
 					  "result_flags = %s", result_flags[0] ? result_flags : "none",
 					  NULL);

-	} else if (act->op == LD_OP_QUERY_LOCK) {
-
-		log_debug("send %s[%d] cl %u %s %s rv %d mode %d",
-			  cl->name[0] ? cl->name : "client", cl->pid, cl->id,
-			  op_str(act->op), rt_str(act->rt),
-			  act->result, act->mode);
-
-		res = daemon_reply_simple("OK",
-					  "op = " FMTd64, (int64_t)act->op,
-					  "op_result = " FMTd64, (int64_t) act->result,
-					  "lock_type = %s", lm_str(act->lm_type),
-					  "mode = %s", mode_str(act->mode),
-					  NULL);
-
 	} else if (act->op == LD_OP_DUMP_LOG || act->op == LD_OP_DUMP_INFO) {
 		/*
 		 * lvmlockctl creates the unix socket then asks us to write to it.
@@ -4066,16 +4054,6 @@ static int str_to_op_rt(const char *req_name, int *op, int *rt)
 		*rt = 0;
 		return 0;
 	}
-	if (!strcmp(req_name, "query_lock_vg")) {
-		*op = LD_OP_QUERY_LOCK;
-		*rt = LD_RT_VG;
-		return 0;
-	}
-	if (!strcmp(req_name, "query_lock_lv")) {
-		*op = LD_OP_QUERY_LOCK;
-		*rt = LD_RT_LV;
-		return 0;
-	}
 	if (!strcmp(req_name, "find_free_lock")) {
 		*op = LD_OP_FIND_FREE_LOCK;
 		*rt = LD_RT_VG;
@@ -4091,11 +4069,6 @@ static int str_to_op_rt(const char *req_name, int *op, int *rt)
 		*rt = LD_RT_VG;
 		return 0;
 	}
-	if (!strcmp(req_name, "refresh_lv")) {
-		*op = LD_OP_REFRESH_LV;
-		*rt = 0;
-		return 0;
-	}
 out:
 	return -1;
 }
@@ -4457,7 +4430,6 @@ static void client_recv_action(struct client *cl)
 	const char *vg_name;
 	const char *vg_uuid;
 	const char *vg_sysid;
-	const char *path;
 	const char *str;
 	int64_t val;
 	uint32_t opts = 0;
@@ -4544,7 +4516,6 @@ static void client_recv_action(struct client *cl)
 	opts = str_to_opts(str);
 	str = daemon_request_str(req, "vg_lock_type", NULL);
 	lm = str_to_lm(str);
-	path = daemon_request_str(req, "path", NULL);

 	if (cl_pid && cl_pid != cl->pid)
 		log_error("client recv bad message pid %d client %d", cl_pid, cl->pid);
@@ -4577,9 +4548,6 @@ static void client_recv_action(struct client *cl)
 	act->flags = opts;
 	act->lm_type = lm;

-	if (path)
-		act->path = strdup(path);
-
 	if (vg_name && strcmp(vg_name, "none"))
 		strncpy(act->vg_name, vg_name, MAX_NAME);

@@ -4656,7 +4624,6 @@ static void client_recv_action(struct client *cl)
 	case LD_OP_STOP_ALL:
 	case LD_OP_RENAME_FINAL:
 	case LD_OP_RUNNING_LM:
-	case LD_OP_REFRESH_LV:
 		add_work_action(act);
 		rv = 0;
 		break;
@@ -4666,7 +4633,6 @@ static void client_recv_action(struct client *cl)
 	case LD_OP_DISABLE:
 	case LD_OP_FREE:
 	case LD_OP_RENAME_BEFORE:
-	case LD_OP_QUERY_LOCK:
 	case LD_OP_FIND_FREE_LOCK:
 	case LD_OP_KILL_VG:
 	case LD_OP_DROP_VG:
@@ -4847,7 +4813,7 @@ static void close_client_thread(void)
 }

 /*
- * Get a list of all VGs with a lockd type (sanlock|dlm).
+ * Get a list of all VGs with a lockd type (sanlock|dlm) from lvmetad.
 * We'll match this list against a list of existing lockspaces that are
 * found in the lock manager.
 *
@@ -4858,9 +4824,6 @@ static void close_client_thread(void)

 static int get_lockd_vgs(struct list_head *vg_lockd)
 {
-	/* FIXME: get VGs some other way */
-	return -1;
-#if 0
 	struct list_head update_vgs;
 	daemon_reply reply;
 	struct dm_config_node *cn;
@@ -5017,7 +4980,6 @@ out:
 	}

 	return rv;
-#endif
 }

 static char _dm_uuid[DM_UUID_LEN];
@@ -5292,7 +5254,7 @@ static void adopt_locks(void)
 		gl_use_sanlock = 1;

 	list_for_each_entry(ls, &vg_lockd, list) {
-		log_debug("adopt vg %s lock_type %s lock_args %s",
+		log_debug("adopt lvmetad vg %s lock_type %s lock_args %s",
 			  ls->vg_name, lm_str(ls->lm_type), ls->vg_args);

 		list_for_each_entry(r, &ls->resources, list)
@@ -5357,7 +5319,7 @@ static void adopt_locks(void)
 		/*
 		 * LS in ls_found, not in vg_lockd.
 		 * An lvm lockspace found in the lock manager has no
-		 * corresponding VG.  This shouldn't usually
+		 * corresponding VG in lvmetad.  This shouldn't usually
 		 * happen, but it's possible the VG could have been removed
 		 * while the orphaned lockspace from it was still around.
 		 * Report an error and leave the ls in the lm alone.
@@ -5372,7 +5334,7 @@ static void adopt_locks(void)

 	/*
 	 * LS in vg_lockd, not in ls_found.
-	 * lockd vgs that do not have an existing lockspace.
+	 * lockd vgs from lvmetad that do not have an existing lockspace.
 	 * This wouldn't be unusual; we just skip the vg.
 	 * But, if the vg has active lvs, then it should have had locks
 	 * and a lockspace.  Should we attempt to join the lockspace and
@@ -5424,6 +5386,8 @@ static void adopt_locks(void)
 		memcpy(act->vg_args, ls->vg_args, MAX_ARGS);
 		act->host_id = ls->host_id;

+		/* set act->version from lvmetad data? */
+
 		log_debug("adopt add %s vg lockspace %s", lm_str(act->lm_type), act->vg_name);

 		rv = add_lockspace_thread(ls->name, act->vg_name, act->vg_uuid,
@@ -5882,9 +5846,12 @@ static int main_loop(daemon_state *ds_arg)
 	setup_worker_thread();
 	setup_restart();

-#ifdef USE_SD_NOTIFY
-	sd_notify(0, "READY=1");
-#endif
+	pthread_mutex_init(&lvmetad_mutex, NULL);
+	lvmetad_handle = lvmetad_open(NULL);
+	if (lvmetad_handle.error || lvmetad_handle.socket_fd < 0)
+		log_error("lvmetad_open error %d", lvmetad_handle.error);
+	else
+		lvmetad_connected = 1;

 	/*
 	 * Attempt to rejoin lockspaces and adopt locks from a previous
@@ -6007,6 +5974,7 @@ static int main_loop(daemon_state *ds_arg)
 	close_worker_thread();
 	close_client_thread();
 	closelog();
+	daemon_close(lvmetad_handle);
 	return 1; /* libdaemon uses 1 for success */
 }

--- a/daemons/lvmlockd/lvmlockd-dlm.c
+++ b/daemons/lvmlockd/lvmlockd-dlm.c
@@ -11,20 +11,19 @@
 #define _XOPEN_SOURCE 500  /* pthread */
 #define _ISOC99_SOURCE

-#include "tools/tool.h"
+#include "tool.h"

 #include "daemon-server.h"
-#include "lib/mm/xlate.h"
+#include "xlate.h"

 #include "lvmlockd-internal.h"
-#include "daemons/lvmlockd/lvmlockd-client.h"
+#include "lvmlockd-client.h"

 /*
 * Using synchronous _wait dlm apis so do not define _REENTRANT and
 * link with non-threaded version of library, libdlm_lt.
 */
 #include "libdlm.h"
-#include "libdlmcontrol.h"

 #include <stddef.h>
 #include <poll.h>
@@ -128,18 +127,16 @@ static int read_cluster_name(char *clustername)
 	return 0;
 }

-#define MAX_VERSION 16
-
 int lm_init_vg_dlm(char *ls_name, char *vg_name, uint32_t flags, char *vg_args)
 {
 	char clustername[MAX_ARGS+1];
-	char lock_args_version[MAX_VERSION+1];
+	char lock_args_version[MAX_ARGS+1];
 	int rv;

 	memset(clustername, 0, sizeof(clustername));
 	memset(lock_args_version, 0, sizeof(lock_args_version));

-	snprintf(lock_args_version, MAX_VERSION, "%u.%u.%u",
+	snprintf(lock_args_version, MAX_ARGS, "%u.%u.%u",
 		 VG_LOCK_ARGS_MAJOR, VG_LOCK_ARGS_MINOR, VG_LOCK_ARGS_PATCH);

 	rv = read_cluster_name(clustername);
@@ -151,9 +148,7 @@ int lm_init_vg_dlm(char *ls_name, char *vg_name, uint32_t flags, char *vg_args)
 		return -EARGS;
 	}

-	rv = snprintf(vg_args, MAX_ARGS, "%s:%s", lock_args_version, clustername);
-	if (rv >= MAX_ARGS)
-		log_debug("init_vg_dlm vg_args may be too long %d %s", rv, vg_args);
+	snprintf(vg_args, MAX_ARGS, "%s:%s", lock_args_version, clustername);
 	rv = 0;

 	log_debug("init_vg_dlm done %s vg_args %s", ls_name, vg_args);
@@ -277,9 +272,10 @@ static int lm_add_resource_dlm(struct lockspace *ls, struct resource *r, int wit
 	int rv;

 	if (r->type == LD_RT_GL || r->type == LD_RT_VG) {
-		buf = zalloc(sizeof(struct val_blk) + DLM_LVB_LEN);
+		buf = malloc(sizeof(struct val_blk) + DLM_LVB_LEN);
 		if (!buf)
 			return -ENOMEM;
+		memset(buf, 0, sizeof(struct val_blk) + DLM_LVB_LEN);

 		rdd->vb = (struct val_blk *)buf;
 		rdd->lksb.sb_lvbptr = buf + sizeof(struct val_blk);
@@ -512,7 +508,7 @@ lockrv:
 	}
 	if (rv < 0) {
 		log_error("S %s R %s lock_dlm acquire error %d errno %d", ls->name, r->name, rv, errno);
-		return -ELMERR;
+		return rv;
 	}

 	if (rdd->vb) {
@@ -585,7 +581,6 @@ int lm_convert_dlm(struct lockspace *ls, struct resource *r,
 	}
 	if (rv < 0) {
 		log_error("S %s R %s convert_dlm error %d", ls->name, r->name, rv);
-		rv = -ELMERR;
 	}
 	return rv;
 }
@@ -659,7 +654,6 @@ int lm_unlock_dlm(struct lockspace *ls, struct resource *r,
 			      0, NULL, NULL, NULL);
 	if (rv < 0) {
 		log_error("S %s R %s unlock_dlm error %d", ls->name, r->name, rv);
-		rv = -ELMERR;
 	}

 	return rv;
@@ -703,7 +697,7 @@ int lm_hosts_dlm(struct lockspace *ls, int notify)
 		return 0;

 	memset(ls_nodes_path, 0, sizeof(ls_nodes_path));
-	snprintf(ls_nodes_path, PATH_MAX, "%s/%s/nodes",
+	snprintf(ls_nodes_path, PATH_MAX-1, "%s/%s/nodes",
 		 DLM_LOCKSPACES_PATH, ls->name);

 	if (!(ls_dir = opendir(ls_nodes_path)))
@@ -781,107 +775,3 @@ int lm_is_running_dlm(void)
 	return 1;
 }

-#ifdef LOCKDDLM_CONTROL_SUPPORT
-
-int lm_refresh_lv_start_dlm(struct action *act)
-{
-	char path[PATH_MAX];
-	char command[DLMC_RUN_COMMAND_LEN];
-	char run_uuid[DLMC_RUN_UUID_LEN];
-	char *p, *vgname, *lvname;
-	int rv;
-
-	/* split /dev/vgname/lvname into vgname and lvname strings */
-	strncpy(path, act->path, strlen(act->path));
-
-	/* skip past dev */
-	p = strchr(path + 1, '/');
-
-	/* skip past slashes */
-	while (*p == '/')
-		p++;
-
-	/* start of vgname */
-	vgname = p;
-
-	/* skip past vgname */
-	while (*p != '/')
-		p++;
-
-	/* terminate vgname */
-	*p = '\0';
-	p++;
-
-	/* skip past slashes */
-	while (*p == '/')
-		p++;
-
-	lvname = p;
-
-	memset(command, 0, sizeof(command));
-	memset(run_uuid, 0, sizeof(run_uuid));
-
-	/* todo: add --readonly */
-
-	snprintf(command, DLMC_RUN_COMMAND_LEN,
-		 "lvm lvchange --refresh --partial --nolocking %s/%s",
-		 vgname, lvname);
-
-	rv = dlmc_run_start(command, strlen(command), 0,
-			    DLMC_FLAG_RUN_START_NODE_NONE,
-			    run_uuid);
-	if (rv < 0) {
-		log_debug("refresh_lv run_start error %d", rv);
-		return rv;
-	}
-
-	log_debug("refresh_lv run_start %s", run_uuid);
-
-	/* Bit of a hack here, we don't need path once started,
-	   but we do need to save the run_uuid somewhere, so just
-	   replace the path with the uuid. */
-
-	free(act->path);
-	act->path = strdup(run_uuid);
-	return 0;
-}
-
-int lm_refresh_lv_check_dlm(struct action *act)
-{
-	uint32_t check_status = 0;
-	int rv;
-
-	/* NB act->path was replaced with run_uuid */
-
-	rv = dlmc_run_check(act->path, strlen(act->path), 0,
-			    DLMC_FLAG_RUN_CHECK_CLEAR,
-			    &check_status);
-	if (rv < 0) {
-		log_debug("refresh_lv check error %d", rv);
-		return rv;
-	}
-
-	log_debug("refresh_lv check %s status %x", act->path, check_status);
-
-	if (!(check_status & DLMC_RUN_STATUS_DONE))
-		return -EAGAIN;
-
-	if (check_status & DLMC_RUN_STATUS_FAILED)
-		return -1;
-
-	return 0;
-}
-
-#else /* LOCKDDLM_CONTROL_SUPPORT */
-
-int lm_refresh_lv_start_dlm(struct action *act)
-{
-	return 0;
-}
-
-int lm_refresh_lv_check_dlm(struct action *act)
-{
-	return 0;
-}
-
-#endif /* LOCKDDLM_CONTROL_SUPPORT */
--- a/daemons/lvmlockd/lvmlockd-internal.h
+++ b/daemons/lvmlockd/lvmlockd-internal.h
@@ -53,8 +53,6 @@ enum {
 	LD_OP_KILL_VG,
 	LD_OP_DROP_VG,
 	LD_OP_BUSY,
-	LD_OP_QUERY_LOCK,
-	LD_OP_REFRESH_LV,
 };

 /* resource types */
@@ -107,7 +105,6 @@ struct client {
 #define LD_AF_WARN_GL_REMOVED	   0x00020000
 #define LD_AF_LV_LOCK              0x00040000
 #define LD_AF_LV_UNLOCK            0x00080000
-#define LD_AF_SH_EXISTS            0x00100000

 /*
 * Number of times to repeat a lock request after
@@ -130,7 +127,6 @@ struct action {
 	int max_retries;
 	int result;
 	int lm_rv;			/* return value from lm_ function */
-	char *path;
 	char vg_uuid[64];
 	char vg_name[MAX_NAME+1];
 	char lv_name[MAX_NAME+1];
@@ -178,9 +174,7 @@ struct lockspace {
 	int8_t lm_type;			/* lock manager: LM_DLM, LM_SANLOCK */
 	void *lm_data;
 	uint64_t host_id;
-	uint64_t free_lock_offset;	/* for sanlock, start search for free lock here */
-	int free_lock_sector_size;	/* for sanlock */
-	int free_lock_align_size;	/* for sanlock */
+	uint64_t free_lock_offset;	/* start search for free lock here */

 	uint32_t start_client_id;	/* client_id that started the lockspace */
 	pthread_t thread;		/* makes synchronous lock requests */
@@ -393,8 +387,6 @@ int lm_get_lockspaces_dlm(struct list_head *ls_rejoin);
 int lm_data_size_dlm(void);
 int lm_is_running_dlm(void);
 int lm_hosts_dlm(struct lockspace *ls, int notify);
-int lm_refresh_lv_start_dlm(struct action *act);
-int lm_refresh_lv_check_dlm(struct action *act);

 static inline int lm_support_dlm(void)
 {
@@ -471,22 +463,12 @@ static inline int lm_hosts_dlm(struct lockspace *ls, int notify)
 	return 0;
 }

-static inline int lm_refresh_lv_start_dlm(struct action *act)
-{
-	return 0;
-}
-
-static inline int lm_refresh_lv_check_dlm(struct action *act)
-{
-	return 0;
-}
-
 #endif /* dlm support */

 #ifdef LOCKDSANLOCK_SUPPORT

 int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_args);
-int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name, char *vg_args, char *lv_args, int sector_size, int align_size, uint64_t free_offset);
+int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name, char *vg_args, char *lv_args, uint64_t free_offset);
 int lm_free_lv_sanlock(struct lockspace *ls, struct resource *r);
 int lm_rename_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_args);
 int lm_prepare_lockspace_sanlock(struct lockspace *ls);
@@ -506,7 +488,7 @@ int lm_gl_is_enabled(struct lockspace *ls);
 int lm_get_lockspaces_sanlock(struct list_head *ls_rejoin);
 int lm_data_size_sanlock(void);
 int lm_is_running_sanlock(void);
-int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset, int *sector_size, int *align_size);
+int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset);

 static inline int lm_support_sanlock(void)
 {
@@ -520,7 +502,7 @@ static inline int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flag
 	return -1;
 }

-static inline int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name, char *vg_args, char *lv_args, int sector_size, int align_size, uint64_t free_offset)
+static inline int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name, char *vg_args, char *lv_args, uint64_t free_offset)
 {
 	return -1;
 }
@@ -608,7 +590,7 @@ static inline int lm_is_running_sanlock(void)
 	return 0;
 }

-static inline int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset, int *sector_size, int *align_size)
+static inline int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset)
 {
 	return -1;
 }
--- a/daemons/lvmlockd/lvmlockd-sanlock.c
+++ b/daemons/lvmlockd/lvmlockd-sanlock.c
@@ -11,42 +11,23 @@
 #define _XOPEN_SOURCE 500  /* pthread */
 #define _ISOC99_SOURCE

-#include "tools/tool.h"
+#include "tool.h"

 #include "daemon-server.h"
-#include "lib/mm/xlate.h"
+#include "xlate.h"

 #include "lvmlockd-internal.h"
-#include "daemons/lvmlockd/lvmlockd-client.h"
+#include "lvmlockd-client.h"

 #include "sanlock.h"
 #include "sanlock_rv.h"
 #include "sanlock_admin.h"
 #include "sanlock_resource.h"

-/* FIXME: these are copied from sanlock.h only until
-   an updated version of sanlock is available with them. */
-#define SANLK_RES_ALIGN1M       0x00000010
-#define SANLK_RES_ALIGN2M       0x00000020
-#define SANLK_RES_ALIGN4M       0x00000040
-#define SANLK_RES_ALIGN8M       0x00000080
-#define SANLK_RES_SECTOR512     0x00000100
-#define SANLK_RES_SECTOR4K      0x00000200
-#define SANLK_LSF_ALIGN1M       0x00000010
-#define SANLK_LSF_ALIGN2M       0x00000020
-#define SANLK_LSF_ALIGN4M       0x00000040
-#define SANLK_LSF_ALIGN8M       0x00000080
-#define SANLK_LSF_SECTOR512     0x00000100
-#define SANLK_LSF_SECTOR4K      0x00000200
-
 #include <stddef.h>
 #include <poll.h>
 #include <errno.h>
 #include <syslog.h>
-#include <blkid/blkid.h>
-#include <sys/sysmacros.h>
-
-#define ONE_MB 1048576

 /*
 -------------------------------------------------------------------------------
@@ -158,7 +139,6 @@ release all the leases for the VG.

 struct lm_sanlock {
 	struct sanlk_lockspace ss;
-	int sector_size;
 	int align_size;
 	int sock; /* sanlock daemon connection */
 };
@@ -221,6 +201,7 @@ int lm_data_size_sanlock(void)
 * ...
 */

+#define LS_BEGIN 0
 #define GL_LOCK_BEGIN UINT64_C(65)
 #define VG_LOCK_BEGIN UINT64_C(66)
 #define LV_LOCK_BEGIN UINT64_C(67)
@@ -307,191 +288,12 @@ static int read_host_id_file(void)
 		}
 	}
 	if (fclose(file))
-		log_debug("Failed to fclose host id file %s (%s).",
-			  daemon_host_id_file, strerror(errno));
+		log_error("failed to close host id file %s", daemon_host_id_file);
 out:
 	log_debug("host_id %d from %s", host_id, daemon_host_id_file);
 	return host_id;
 }

-/* Prepare valid /dev/mapper/vgname-lvname with all the mangling */
-static int build_dm_path(char *path, size_t path_len,
-			 const char *vg_name, const char *lv_name)
-{
-	struct dm_pool *mem;
-	char *dm_name;
-	int rv = 0;
-
-	if (!(mem = dm_pool_create("namepool", 1024))) {
-		log_error("Failed to create mempool.");
-		return -ENOMEM;
-	}
-
-	if (!(dm_name = dm_build_dm_name(mem, vg_name, lv_name, NULL))) {
-		log_error("Failed to build dm name for %s/%s.", vg_name, lv_name);
-		rv = -EINVAL;
-		goto fail;
-	}
-
-	if ((dm_snprintf(path, path_len, "%s/%s", dm_dir(), dm_name) < 0)) {
-		log_error("Failed to create path %s/%s.", dm_dir(), dm_name);
-		rv = -EINVAL;
-	}
-
-fail:
-	dm_pool_destroy(mem);
-
-	return rv;
-}
-
-static void _read_sysfs_size(dev_t devno, const char *name, unsigned int *val)
-{
-	char path[PATH_MAX];
-	char buf[32];
-	FILE *fp;
-	size_t len;
-
-	snprintf(path, sizeof(path), "/sys/dev/block/%d:%d/queue/%s",
-		 (int)major(devno), (int)minor(devno), name);
-
-	if (!(fp = fopen(path, "r")))
-		return;
-
-	if (!fgets(buf, sizeof(buf), fp))
-		goto out;
-
-	if ((len = strlen(buf)) && buf[len - 1] == '\n')
-		buf[--len] = '\0';
-
-	if (strlen(buf))
-		*val = atoi(buf);
-out:
-	if (fclose(fp))
-		log_debug("Failed to fclose host id file %s (%s).", path, strerror(errno));
-
-}
-
-/* Select sector/align size for a new VG based on what the device reports for
-   sector size of the lvmlock LV. */
-
-static int get_sizes_device(char *path, int *sector_size, int *align_size)
-{
-	unsigned int physical_block_size = 0;
-	unsigned int logical_block_size = 0;
-	struct stat st;
-	int rv;
-
-	rv = stat(path, &st);
-	if (rv < 0) {
-		log_error("Failed to stat device to get block size %s %d", path, errno);
-		return -1;
-	}
-
-	_read_sysfs_size(st.st_rdev, "physical_block_size", &physical_block_size);
-	_read_sysfs_size(st.st_rdev, "logical_block_size", &logical_block_size);
-
-	if ((physical_block_size == 512) && (logical_block_size == 512)) {
-		*sector_size = 512;
-		*align_size = ONE_MB;
-		return 0;
-	}
-
-	if ((physical_block_size == 4096) && (logical_block_size == 4096)) {
-		*sector_size = 4096;
-		*align_size = 8 * ONE_MB;
-		return 0;
-	}
-
-	if (physical_block_size && (physical_block_size != 512) && (physical_block_size != 4096)) {
-		log_warn("WARNING: invalid block sizes physical %u logical %u for %s",
-			 physical_block_size, logical_block_size, path);
-		physical_block_size = 0;
-	}
-
-	if (logical_block_size && (logical_block_size != 512) && (logical_block_size != 4096)) {
-		log_warn("WARNING: invalid block sizes physical %u logical %u for %s",
-			 physical_block_size, logical_block_size, path);
-		logical_block_size = 0;
-	}
-
-	if (!physical_block_size && !logical_block_size) {
-		log_error("Failed to get a block size for %s", path);
-		return -1;
-	}
-
-	if (!physical_block_size || !logical_block_size) {
-		log_warn("WARNING: incomplete block size information physical %u logical %u for %s",
-			 physical_block_size, logical_block_size, path);
-		if (!physical_block_size)
-			physical_block_size = logical_block_size;
-		if (!logical_block_size)
-			logical_block_size = physical_block_size;
-	}
-
-	if ((logical_block_size == 4096) && (physical_block_size == 512)) {
-		log_warn("WARNING: mixed block sizes physical %u logical %u (using 4096) for %s",
-			 physical_block_size, logical_block_size, path);
-		*sector_size = 4096;
-		*align_size = 8 * ONE_MB;
-		return 0;
-	}
-
-	if ((physical_block_size == 4096) && (logical_block_size == 512)) {
-		log_warn("WARNING: mixed block sizes physical %u logical %u (using 4096) for %s",
-			 physical_block_size, logical_block_size, path);
-		*sector_size = 4096;
-		*align_size = 8 * ONE_MB;
-		return 0;
-	}
-
-	if (physical_block_size == 512) {
-		*sector_size = 512;
-		*align_size = ONE_MB;
-		return 0;
-	}
-
-	if (physical_block_size == 4096) {
-		*sector_size = 4096;
-		*align_size = 8 * ONE_MB;
-		return 0;
-	}
-
-	log_error("Failed to get a block size for %s", path);
-	return -1;
-}
-
-
-/* Get the sector/align sizes that were used to create an existing VG.
-   sanlock encoded this in the lockspace/resource structs on disk. */
-
-static int get_sizes_lockspace(char *path, int *sector_size, int *align_size)
-{
-	struct sanlk_lockspace ss;
-	uint32_t io_timeout = 0;
-	int rv;
-
-	memset(&ss, 0, sizeof(ss));
-	memcpy(ss.host_id_disk.path, path, SANLK_PATH_LEN);
-	ss.host_id_disk.offset = 0;
-
-	rv = sanlock_read_lockspace(&ss, 0, &io_timeout);
-	if (rv < 0) {
-		log_error("get_sizes_lockspace %s error %d", path, rv);
-		return rv;
-	}
-
-	if ((ss.flags & SANLK_LSF_SECTOR4K) && (ss.flags & SANLK_LSF_ALIGN8M)) {
-		*sector_size = 4096;
-		*align_size = 8 * ONE_MB;
-	} else if ((ss.flags & SANLK_LSF_SECTOR512) && (ss.flags & SANLK_LSF_ALIGN1M)) {
-		*sector_size = 512;
-		*align_size = ONE_MB;
-	}
-
-	log_debug("get_sizes_lockspace found %d %d", *sector_size, *align_size);
-	return 0;
-}
-
 /*
 * vgcreate
 *
@@ -500,21 +302,18 @@ static int get_sizes_lockspace(char *path, int *sector_size, int *align_size)
 * version and lv name, and returns the real lock_args in vg_args.
 */

-#define MAX_VERSION 16
-
 int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_args)
 {
 	struct sanlk_lockspace ss;
 	struct sanlk_resourced rd;
 	struct sanlk_disk disk;
 	char lock_lv_name[MAX_ARGS+1];
-	char lock_args_version[MAX_VERSION+1];
+	char lock_args_version[MAX_ARGS+1];
 	const char *gl_name = NULL;
 	uint32_t daemon_version;
 	uint32_t daemon_proto;
 	uint64_t offset;
-	int sector_size = 0;
-	int align_size = 0;
+	int align_size;
 	int i, rv;

 	memset(&ss, 0, sizeof(ss));
@@ -528,7 +327,7 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 		return -EARGS;
 	}

-	snprintf(lock_args_version, MAX_VERSION, "%u.%u.%u",
+	snprintf(lock_args_version, MAX_ARGS, "%u.%u.%u",
 		 VG_LOCK_ARGS_MAJOR, VG_LOCK_ARGS_MINOR, VG_LOCK_ARGS_PATCH);

 	/* see comment above about input vg_args being only lock_lv_name */
@@ -537,17 +336,14 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 	if (strlen(lock_lv_name) + strlen(lock_args_version) + 2 > MAX_ARGS)
 		return -EARGS;

-	if ((rv = build_dm_path(disk.path, SANLK_PATH_LEN, vg_name, lock_lv_name)))
-		return rv;
+	snprintf(disk.path, SANLK_PATH_LEN-1, "/dev/mapper/%s-%s", vg_name, lock_lv_name);

 	log_debug("S %s init_vg_san path %s", ls_name, disk.path);

 	if (daemon_test) {
 		if (!gl_lsname_sanlock[0])
 			strncpy(gl_lsname_sanlock, ls_name, MAX_NAME);
-		rv = snprintf(vg_args, MAX_ARGS, "%s:%s", lock_args_version, lock_lv_name);
-		if (rv >= MAX_ARGS)
-			log_debug("init_vg_san vg_args may be too long %d %s", rv, vg_args);
+		snprintf(vg_args, MAX_ARGS, "%s:%s", lock_args_version, lock_lv_name);
 		return 0;
 	}

@@ -560,25 +356,23 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 	log_debug("sanlock daemon version %08x proto %08x",
 		  daemon_version, daemon_proto);

-	/* Nothing formatted on disk yet, use what the device reports. */
-	rv = get_sizes_device(disk.path, &sector_size, &align_size);
-	if (rv < 0) {
+	rv = sanlock_align(&disk);
+	if (rv <= 0) {
 		if (rv == -EACCES) {
 			log_error("S %s init_vg_san sanlock error -EACCES: no permission to access %s",
 				  ls_name, disk.path);
 			return -EDEVOPEN;
 		} else {
-			log_error("S %s init_vg_san sanlock error %d trying to get sector/align size of %s",
+			log_error("S %s init_vg_san sanlock error %d trying to get align size of %s",
 				  ls_name, rv, disk.path);
 			return -EARGS;
 		}
-	}
+	} else
+		align_size = rv;

 	strncpy(ss.name, ls_name, SANLK_NAME_LEN);
 	memcpy(ss.host_id_disk.path, disk.path, SANLK_PATH_LEN);
-	ss.host_id_disk.offset = 0;
-	ss.flags = (sector_size == 4096) ? (SANLK_LSF_SECTOR4K | SANLK_LSF_ALIGN8M) :
-					   (SANLK_LSF_SECTOR512 | SANLK_LSF_ALIGN1M);
+	ss.host_id_disk.offset = LS_BEGIN * align_size;

 	rv = sanlock_write_lockspace(&ss, 0, 0, sanlock_io_timeout);
 	if (rv < 0) {
@@ -611,8 +405,6 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 	memcpy(rd.rs.disks[0].path, disk.path, SANLK_PATH_LEN);
 	rd.rs.disks[0].offset = align_size * GL_LOCK_BEGIN;
 	rd.rs.num_disks = 1;
-	rd.rs.flags = (sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-					      (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 	rv = sanlock_write_resource(&rd.rs, 0, 0, 0);
 	if (rv < 0) {
@@ -626,8 +418,6 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 	memcpy(rd.rs.disks[0].path, disk.path, SANLK_PATH_LEN);
 	rd.rs.disks[0].offset = align_size * VG_LOCK_BEGIN;
 	rd.rs.num_disks = 1;
-	rd.rs.flags = (sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-					      (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 	rv = sanlock_write_resource(&rd.rs, 0, 0, 0);
 	if (rv < 0) {
@@ -639,9 +429,7 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 	if (!strcmp(gl_name, R_NAME_GL))
 		strncpy(gl_lsname_sanlock, ls_name, MAX_NAME);
 
-	rv = snprintf(vg_args, MAX_ARGS, "%s:%s", lock_args_version, lock_lv_name);
-	if (rv >= MAX_ARGS)
-		log_debug("init_vg_san vg_args may be too long %d %s", rv, vg_args);
+	snprintf(vg_args, MAX_ARGS, "%s:%s", lock_args_version, lock_lv_name);

 	log_debug("S %s init_vg_san done vg_args %s", ls_name, vg_args);

@@ -653,8 +441,6 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar

 	memset(&rd, 0, sizeof(rd));
 	rd.rs.num_disks = 1;
-	rd.rs.flags = (sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-					      (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);
 	memcpy(rd.rs.disks[0].path, disk.path, SANLK_PATH_LEN);
 	strncpy(rd.rs.lockspace_name, ls_name, SANLK_NAME_LEN);
 	strcpy(rd.rs.name, "#unused");
@@ -693,13 +479,13 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
 */

 int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name,
-		       char *vg_args, char *lv_args,
-		       int sector_size, int align_size, uint64_t free_offset)
+		       char *vg_args, char *lv_args, uint64_t free_offset)
 {
 	struct sanlk_resourced rd;
 	char lock_lv_name[MAX_ARGS+1];
-	char lock_args_version[MAX_VERSION+1];
+	char lock_args_version[MAX_ARGS+1];
 	uint64_t offset;
+	int align_size;
 	int rv;

 	memset(&rd, 0, sizeof(rd));
@@ -713,11 +499,11 @@ int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name,
 		return rv;
 	}

-	snprintf(lock_args_version, MAX_VERSION, "%u.%u.%u",
+	snprintf(lock_args_version, MAX_ARGS, "%u.%u.%u",
 		 LV_LOCK_ARGS_MAJOR, LV_LOCK_ARGS_MINOR, LV_LOCK_ARGS_PATCH);

 	if (daemon_test) {
-		align_size = ONE_MB;
+		align_size = 1048576;
 		snprintf(lv_args, MAX_ARGS, "%s:%llu",
 			 lock_args_version,
 			 (unsigned long long)((align_size * LV_LOCK_BEGIN) + (align_size * daemon_test_lv_count)));
@@ -727,38 +513,14 @@ int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name,

 	strncpy(rd.rs.lockspace_name, ls_name, SANLK_NAME_LEN);
 	rd.rs.num_disks = 1;
-	if ((rv = build_dm_path(rd.rs.disks[0].path, SANLK_PATH_LEN, vg_name, lock_lv_name)))
-		return rv;
+	snprintf(rd.rs.disks[0].path, SANLK_PATH_LEN-1, "/dev/mapper/%s-%s", vg_name, lock_lv_name);

-	/*
-	 * These should not usually be zero, maybe only the first time this function is called?
-	 * We need to use the same sector/align sizes that are already being used.
-	 */
-	if (!sector_size || !align_size) {
-		rv = get_sizes_lockspace(rd.rs.disks[0].path, &sector_size, &align_size);
-		if (rv < 0) {
-			log_error("S %s init_lv_san read_lockspace error %d %s",
-				  ls_name, rv, rd.rs.disks[0].path);
-			return rv;
-		}
-
-		if (sector_size)
-			log_debug("S %s init_lv_san found ls sector_size %d align_size %d", ls_name, sector_size, align_size);
-		else {
-			/* use the old method */
-			align_size = sanlock_align(&rd.rs.disks[0]);
-			if (align_size <= 0) {
-				log_error("S %s init_lv_san align error %d", ls_name, align_size);
-				return -EINVAL;
-			}
-			sector_size = (align_size == ONE_MB) ? 512 : 4096;
-			log_debug("S %s init_lv_san found old sector_size %d align_size %d", ls_name, sector_size, align_size);
-		}
+	align_size = sanlock_align(&rd.rs.disks[0]);
+	if (align_size <= 0) {
+		log_error("S %s init_lv_san align error %d", ls_name, align_size);
+		return -EINVAL;
 	}

-	rd.rs.flags = (sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-					      (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);
-
 	if (free_offset)
 		offset = free_offset;
 	else
@@ -801,8 +563,6 @@ int lm_init_lv_sanlock(char *ls_name, char *vg_name, char *lv_name,
 				  ls_name, lv_name, (unsigned long long)offset);

 			strncpy(rd.rs.name, lv_name, SANLK_NAME_LEN);
-			rd.rs.flags = (sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-							      (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 			rv = sanlock_write_resource(&rd.rs, 0, 0, 0);
 			if (!rv) {
@@ -834,8 +594,7 @@ int lm_rename_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_
 	char lock_lv_name[MAX_ARGS+1];
 	uint64_t offset;
 	uint32_t io_timeout;
-	int sector_size = 0;
-	int align_size = 0;
+	int align_size;
 	int i, rv;

 	memset(&disk, 0, sizeof(disk));
@@ -853,8 +612,7 @@ int lm_rename_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_
 		return rv;
 	}

-	if ((rv = build_dm_path(disk.path, SANLK_PATH_LEN, vg_name, lock_lv_name)))
-		return rv;
+	snprintf(disk.path, SANLK_PATH_LEN-1, "/dev/mapper/%s-%s", vg_name, lock_lv_name);

 	log_debug("S %s rename_vg_san path %s", ls_name, disk.path);

@@ -864,13 +622,20 @@ int lm_rename_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_
 	/* FIXME: device is not always ready for us here */
 	sleep(1);

+	align_size = sanlock_align(&disk);
+	if (align_size <= 0) {
+		log_error("S %s rename_vg_san bad align size %d %s",
+			  ls_name, align_size, disk.path);
+		return -EINVAL;
+	}
+
 	/*
 	 * Lockspace
 	 */

 	memset(&ss, 0, sizeof(ss));
 	memcpy(ss.host_id_disk.path, disk.path, SANLK_PATH_LEN);
-	ss.host_id_disk.offset = 0;
+	ss.host_id_disk.offset = LS_BEGIN * align_size;

 	rv = sanlock_read_lockspace(&ss, 0, &io_timeout);
 	if (rv < 0) {
@@ -879,26 +644,6 @@ int lm_rename_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_
 		return rv;
 	}

-	if ((ss.flags & SANLK_LSF_SECTOR4K) && (ss.flags & SANLK_LSF_ALIGN8M)) {
-		sector_size = 4096;
-		align_size = 8 * ONE_MB;
-	} else if ((ss.flags & SANLK_LSF_SECTOR512) && (ss.flags & SANLK_LSF_ALIGN1M)) {
-		sector_size = 512;
-		align_size = ONE_MB;
-	} else {
-		/* use the old method */
-		align_size = sanlock_align(&ss.host_id_disk);
-		if (align_size <= 0) {
-			log_error("S %s rename_vg_san unknown sector/align size for %s",
-				 ls_name, ss.host_id_disk.path);
-			return -1;
-		}
-		sector_size = (align_size == ONE_MB) ? 512 : 4096;
-	}
-
-	if (!sector_size || !align_size)
-		return -1;
-
 	strncpy(ss.name, ls_name, SANLK_NAME_LEN);

 	rv = sanlock_write_lockspace(&ss, 0, 0, sanlock_io_timeout);
@@ -1052,11 +797,6 @@ int lm_ex_disable_gl_sanlock(struct lockspace *ls)
 	rd1.rs.num_disks = 1;
 	strncpy(rd1.rs.disks[0].path, lms->ss.host_id_disk.path, SANLK_PATH_LEN-1);
 	rd1.rs.disks[0].offset = lms->align_size * GL_LOCK_BEGIN;
-	
-	rd1.rs.flags = (lms->sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-						    (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);
-	rd2.rs.flags = (lms->sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-						    (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 	rv = sanlock_acquire(lms->sock, -1, 0, 1, &rs1, NULL);
 	if (rv < 0) {
@@ -1118,8 +858,6 @@ int lm_able_gl_sanlock(struct lockspace *ls, int enable)
 	rd.rs.num_disks = 1;
 	strncpy(rd.rs.disks[0].path, lms->ss.host_id_disk.path, SANLK_PATH_LEN-1);
 	rd.rs.disks[0].offset = lms->align_size * GL_LOCK_BEGIN;
-	rd.rs.flags = (lms->sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-						   (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 	rv = sanlock_write_resource(&rd.rs, 0, 0, 0);
 	if (rv < 0) {
@@ -1165,8 +903,7 @@ static int gl_is_enabled(struct lockspace *ls, struct lm_sanlock *lms)

 	rv = sanlock_read_resource(&rd.rs, 0);
 	if (rv < 0) {
-		log_error("gl_is_enabled read_resource align_size %d offset %llu error %d",
-			  lms->align_size, (unsigned long long)offset, rv);
+		log_error("gl_is_enabled read_resource error %d", rv);
 		return rv;
 	}

@@ -1203,7 +940,7 @@ int lm_gl_is_enabled(struct lockspace *ls)
 * been disabled.)
 */

-int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset, int *sector_size, int *align_size)
+int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset)
 {
 	struct lm_sanlock *lms = (struct lm_sanlock *)ls->lm_data;
 	struct sanlk_resourced rd;
@@ -1213,22 +950,15 @@ int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset, int *
 	int round = 0;

 	if (daemon_test) {
-		*free_offset = (ONE_MB * LV_LOCK_BEGIN) + (ONE_MB * (daemon_test_lv_count + 1));
-		*sector_size = 512;
-		*align_size = ONE_MB;
+		*free_offset = (1048576 * LV_LOCK_BEGIN) + (1048576 * (daemon_test_lv_count + 1));
 		return 0;
 	}

-	*sector_size = lms->sector_size;
-	*align_size = lms->align_size;
-
 	memset(&rd, 0, sizeof(rd));

 	strncpy(rd.rs.lockspace_name, ls->name, SANLK_NAME_LEN);
 	rd.rs.num_disks = 1;
 	strncpy(rd.rs.disks[0].path, lms->ss.host_id_disk.path, SANLK_PATH_LEN-1);
-	rd.rs.flags = (lms->sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) :
-						   (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 	if (ls->free_lock_offset)
 		offset = ls->free_lock_offset;
@@ -1328,8 +1058,6 @@ int lm_prepare_lockspace_sanlock(struct lockspace *ls)
 	char disk_path[SANLK_PATH_LEN];
 	char killpath[SANLK_PATH_LEN];
 	char killargs[SANLK_PATH_LEN];
-	int sector_size = 0;
-	int align_size = 0;
 	int gl_found;
 	int ret, rv;

@@ -1341,10 +1069,10 @@ int lm_prepare_lockspace_sanlock(struct lockspace *ls)
 	 * and appending "lockctl" to get /path/to/lvmlockctl.
 	 */
 	memset(killpath, 0, sizeof(killpath));
-	snprintf(killpath, SANLK_PATH_LEN, "%slockctl", LVM_PATH);
+	snprintf(killpath, SANLK_PATH_LEN - 1, "%slockctl", LVM_PATH);

 	memset(killargs, 0, sizeof(killargs));
-	snprintf(killargs, SANLK_PATH_LEN, "--kill %s", ls->vg_name);
+	snprintf(killargs, SANLK_PATH_LEN - 1, "--kill %s", ls->vg_name);

 	rv = check_args_version(ls->vg_args, VG_LOCK_ARGS_MAJOR);
 	if (rv < 0) {
@@ -1360,8 +1088,8 @@ int lm_prepare_lockspace_sanlock(struct lockspace *ls)
 		goto fail;
 	}

-	if ((ret = build_dm_path(disk_path, SANLK_PATH_LEN, ls->vg_name, lock_lv_name)))
-		goto fail;
+	snprintf(disk_path, SANLK_PATH_LEN-1, "/dev/mapper/%s-%s",
+		 ls->vg_name, lock_lv_name);

 	/*
 	 * When a vg is started, the internal sanlock lv should be
@@ -1399,7 +1127,7 @@ int lm_prepare_lockspace_sanlock(struct lockspace *ls)
 		goto fail;
 	}

-	lms = zalloc(sizeof(struct lm_sanlock));
+	lms = malloc(sizeof(struct lm_sanlock));
 	if (!lms) {
 		ret = -ENOMEM;
 		goto fail;
@@ -1408,6 +1136,7 @@ int lm_prepare_lockspace_sanlock(struct lockspace *ls)
 	memset(lsname, 0, sizeof(lsname));
 	strncpy(lsname, ls->name, SANLK_NAME_LEN);

+	memset(lms, 0, sizeof(struct lm_sanlock));
 	memcpy(lms->ss.name, lsname, SANLK_NAME_LEN);
 	lms->ss.host_id_disk.offset = 0;
 	lms->ss.host_id = ls->host_id;
@@ -1445,34 +1174,13 @@ int lm_prepare_lockspace_sanlock(struct lockspace *ls)
 		goto fail;
 	}

-	rv = get_sizes_lockspace(disk_path, &sector_size, &align_size);
-	if (rv < 0) {
-		log_error("S %s prepare_lockspace_san cannot get sector/align sizes %d", lsname, rv);
+	lms->align_size = sanlock_align(&lms->ss.host_id_disk);
+	if (lms->align_size <= 0) {
+		log_error("S %s prepare_lockspace_san align error %d", lsname, lms->align_size);
 		ret = -EMANAGER;
 		goto fail;
 	}

-	if (!sector_size) {
-		log_debug("S %s prepare_lockspace_san using old size method", lsname);
-		/* use the old method */
-		align_size = sanlock_align(&lms->ss.host_id_disk);
-		if (align_size <= 0) {
-			log_error("S %s prepare_lockspace_san align error %d", lsname, align_size);
-			ret = -EINVAL;
-			goto fail;
-		}
-		sector_size = (align_size == ONE_MB) ? 512 : 4096;
-		log_debug("S %s prepare_lockspace_san found old sector_size %d align_size %d", lsname, sector_size, align_size);
-	}
-
-	log_debug("S %s prepare_lockspace_san sizes %d %d", lsname, sector_size, align_size);
-
-	lms->align_size = align_size;
-	lms->sector_size = sector_size;
-
-	lms->ss.flags = (sector_size == 4096) ? (SANLK_LSF_SECTOR4K | SANLK_LSF_ALIGN8M) :
-						(SANLK_LSF_SECTOR512 | SANLK_LSF_ALIGN1M);
-
 	gl_found = gl_is_enabled(ls, lms);
 	if (gl_found < 0) {
 		log_error("S %s prepare_lockspace_san gl_enabled error %d", lsname, gl_found);
@@ -1610,7 +1318,6 @@ static int lm_add_resource_sanlock(struct lockspace *ls, struct resource *r)
 	strncpy(rds->rs.name, r->name, SANLK_NAME_LEN);
 	rds->rs.num_disks = 1;
 	memcpy(rds->rs.disks[0].path, lms->ss.host_id_disk.path, SANLK_PATH_LEN);
-	rds->rs.flags = (lms->sector_size == 4096) ? (SANLK_RES_SECTOR4K | SANLK_RES_ALIGN8M) : (SANLK_RES_SECTOR512 | SANLK_RES_ALIGN1M);

 	if (r->type == LD_RT_GL)
 		rds->rs.disks[0].offset = GL_LOCK_BEGIN * lms->align_size;
@@ -1620,9 +1327,10 @@ static int lm_add_resource_sanlock(struct lockspace *ls, struct resource *r)
 	/* LD_RT_LV offset is set in each lm_lock call from lv_args. */

 	if (r->type == LD_RT_GL || r->type == LD_RT_VG) {
-		rds->vb = zalloc(sizeof(struct val_blk));
+		rds->vb = malloc(sizeof(struct val_blk));
 		if (!rds->vb)
 			return -ENOMEM;
+		memset(rds->vb, 0, sizeof(struct val_blk));
 	}

 	return 0;
@@ -1752,12 +1460,6 @@ int lm_lock_sanlock(struct lockspace *ls, struct resource *r, int ld_mode,

 	rv = sanlock_acquire(lms->sock, -1, flags, 1, &rs, &opt);

-	/*
-	 * errors: translate the sanlock error number to an lvmlockd error.
-	 * We don't want to return an sanlock-specific error number from
-	 * this function to code that doesn't recognize sanlock error numbers.
-	 */
-
 	if (rv == -EAGAIN) {
 		/*
 		 * It appears that sanlock_acquire returns EAGAIN when we request
@@ -1826,26 +1528,6 @@ int lm_lock_sanlock(struct lockspace *ls, struct resource *r, int ld_mode,
 		return -EAGAIN;
 	}

-	if (rv == SANLK_AIO_TIMEOUT) {
-		/*
-		 * sanlock got an i/o timeout when trying to acquire the
-		 * lease on disk.
-		 */
-		log_debug("S %s R %s lock_san acquire mode %d rv %d", ls->name, r->name, ld_mode, rv);
-		*retry = 0;
-		return -EAGAIN;
-	}
-
-	if (rv == SANLK_DBLOCK_LVER || rv == SANLK_DBLOCK_MBAL) {
-		/*
-		 * There was contention with another host for the lease,
-		 * and we lost.
-		 */
-		log_debug("S %s R %s lock_san acquire mode %d rv %d", ls->name, r->name, ld_mode, rv);
-		*retry = 0;
-		return -EAGAIN;
-	}
-
 	if (rv == SANLK_ACQUIRE_OWNED_RETRY) {
 		/*
 		 * The lock is held by a failed host, and will eventually
@@ -1896,25 +1578,15 @@ int lm_lock_sanlock(struct lockspace *ls, struct resource *r, int ld_mode,
 		if (rv == -ENOSPC)
 			rv = -ELOCKIO;

-		/*
-		 * generic error number for sanlock errors that we are not
-		 * catching above.
-		 */
-		return -ELMERR;
+		return rv;
 	}

-	/*
-	 * sanlock acquire success (rv 0)
-	 */
-
 	if (rds->vb) {
 		rv = sanlock_get_lvb(0, rs, (char *)&vb, sizeof(vb));
 		if (rv < 0) {
 			log_error("S %s R %s lock_san get_lvb error %d", ls->name, r->name, rv);
 			memset(rds->vb, 0, sizeof(struct val_blk));
 			memset(vb_out, 0, sizeof(struct val_blk));
-			/* the lock is still acquired, the vb values considered invalid */
-			rv = 0;
 			goto out;
 		}

@@ -1967,7 +1639,6 @@ int lm_convert_sanlock(struct lockspace *ls, struct resource *r,
 		if (rv < 0) {
 			log_error("S %s R %s convert_san set_lvb error %d",
 				  ls->name, r->name, rv);
-			return -ELMERR;
 		}
 	}

@@ -1980,35 +1651,14 @@ int lm_convert_sanlock(struct lockspace *ls, struct resource *r,
 	if (daemon_test)
 		return 0;

-	/*
-	 * Don't block waiting for a failed lease to expire since it causes
-	 * sanlock_convert to block for a long time, which would prevent this
-	 * thread from processing other lock requests.
-	 *
-	 * FIXME: SANLK_CONVERT_OWNER_NOWAIT is the same as SANLK_ACQUIRE_OWNER_NOWAIT.
-	 * Change to use the CONVERT define when the latest sanlock version has it.
-	 */
-	flags |= SANLK_ACQUIRE_OWNER_NOWAIT;
-
 	rv = sanlock_convert(lms->sock, -1, flags, rs);
-	if (!rv)
-		return 0;
-
-	switch (rv) {
-	case -EAGAIN:
-	case SANLK_ACQUIRE_IDLIVE:
-	case SANLK_ACQUIRE_OWNED:
-	case SANLK_ACQUIRE_OWNED_RETRY:
-	case SANLK_ACQUIRE_OTHER:
-	case SANLK_AIO_TIMEOUT:
-	case SANLK_DBLOCK_LVER:
-	case SANLK_DBLOCK_MBAL:
-		/* expected errors from known/normal cases like lock contention or io timeouts */
-		log_debug("S %s R %s convert_san error %d", ls->name, r->name, rv);
+	if (rv == -EAGAIN) {
+		/* FIXME: When could this happen?  Should something different be done? */
+		log_error("S %s R %s convert_san EAGAIN", ls->name, r->name);
 		return -EAGAIN;
-	default:
+	}
+	if (rv < 0) {
 		log_error("S %s R %s convert_san convert error %d", ls->name, r->name, rv);
-		rv = -ELMERR;
 	}

 	return rv;
@@ -2045,7 +1695,6 @@ static int release_rename(struct lockspace *ls, struct resource *r)
 	rv = sanlock_release(lms->sock, -1, SANLK_REL_RENAME, 2, res_args);
 	if (rv < 0) {
 		log_error("S %s R %s unlock_san release rename error %d", ls->name, r->name, rv);
-		rv = -ELMERR;
 	}

 	free(res_args);
@@ -2102,7 +1751,6 @@ int lm_unlock_sanlock(struct lockspace *ls, struct resource *r,
 		if (rv < 0) {
 			log_error("S %s R %s unlock_san set_lvb error %d",
 				  ls->name, r->name, rv);
-			return -ELMERR;
 		}
 	}

@@ -2119,20 +1767,10 @@ int lm_unlock_sanlock(struct lockspace *ls, struct resource *r,
 	if (rv < 0)
 		log_error("S %s R %s unlock_san release error %d", ls->name, r->name, rv);

-	/*
-	 * sanlock may return an error here if it fails to release the lease on
-	 * disk because of an io timeout.  But, sanlock will continue trying to
-	 * release the lease after this call returns.  We shouldn't return an
-	 * error here which would result in lvmlockd-core keeping the lock
-	 * around.  By releasing the lock in lvmlockd-core at this point,
-	 * lvmlockd may send another acquire request to lvmlockd.  If sanlock
-	 * has not been able to release the previous instance of the lock yet,
-	 * then it will return an error for the new request.  But, acquiring a
-	 * new lock is able o fail gracefully, until sanlock is finally able to
-	 * release the old lock.
-	 */
+	if (rv == -EIO)
+		rv = -ELOCKIO;

-	return 0;
+	return rv;
 }

 int lm_hosts_sanlock(struct lockspace *ls, int notify)
--- a/daemons/lvmpolld/Makefile.in
+++ b/daemons/lvmpolld/Makefile.in
@@ -29,16 +29,15 @@ include $(top_builddir)/make.tmpl

 CFLAGS += $(EXTRA_EXEC_CFLAGS)
 INCLUDES += -I$(top_srcdir)/libdaemon/server
-LDFLAGS += $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
-LIBS += $(DAEMON_LIBS) $(PTHREAD_LIBS)
+LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
+LIBS += $(DAEMON_LIBS) -ldaemonserver -ldevmapper $(PTHREAD_LIBS)

-lvmpolld: $(OBJECTS) $(top_builddir)/libdaemon/server/libdaemonserver.a $(INTERNAL_LIBS)
-	@echo "    [CC] $@"
-	$(Q) $(CC) $(CFLAGS) $(LDFLAGS) -o $@ $+ $(LIBS)
+lvmpolld: $(OBJECTS) $(top_builddir)/libdaemon/client/libdaemonclient.a \
+		    $(top_builddir)/libdaemon/server/libdaemonserver.a
+	$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LIBS)

 install_lvmpolld: lvmpolld
-	@echo "    [INSTALL] $<"
-	$(Q) $(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
+	$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)

 install_lvm2: install_lvmpolld

--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
David Teigland	b783ef5d0c	lvmcache: simplify metadata cache The copy of VG metadata stored in lvmcache was not being used in general. It pretended to be a generic VG metadata cache, but was not being used except for clvmd activation. There it was used to avoid reading from disk while devices were suspended, i.e. in resume. This removes the code that attempted to make this look like a generic metadata cache, and replaces with with something narrowly targetted to what it's actually used for. This is a way of passing the VG from suspend to resume in clvmd. Since in the case of clvmd one caller can't simply pass the same VG to both suspend and resume, suspend needs to stash the VG somewhere that resume can grab it from. (resume doesn't want to read it from disk since devices are suspended.) The lvmcache vginfo struct is used as a convenient place to stash the VG to pass it from suspend to resume, even though it isn't related to the lvmcache or vginfo. These suspended_vg* vginfo fields should not be used or touched anywhere else, they are only to be used for passing the VG data from suspend to resume in clvmd. The VG data being passed between suspend and resume is never modified, and will only exist in the brief period between suspend and resume in clvmd. suspend has both old (current) and new (precommitted) copies of the VG metadata. It stashes both of these in the vginfo prior to suspending devices. When vg_commit is successful, it sets a flag in vginfo as before, signaling the transition from old to new metadata. resume grabs the VG stashed by suspend. If the vg_commit happened, it grabs the new VG, and if the vg_commit didn't happen it grabs the old VG. The VG is then used to resume LVs. This isolates clvmd-specific code and usage from the normal lvm vg_read code, making the code simpler and the behavior easier to verify. Sequence of operations: - lv_suspend() has both vg_old and vg_new and stashes a copy of each onto the vginfo: lvmcache_save_suspended_vg(vg_old); lvmcache_save_suspended_vg(vg_new); - vg_commit() happens, which causes all clvmd instances to call lvmcache_commit_metadata(vg). A flag is set in the vginfo indicating the transition from the old to new VG: vginfo->suspended_vg_committed = 1; - lv_resume() needs either vg_old or vg_new to use in resuming LVs. It doesn't want to read the VG from disk since devices are suspended, so it gets the VG stashed by lv_suspend: vg = lvmcache_get_suspended_vg(vgid); If the vg_commit did not happen, suspended_vg_committed will not be set, and in this case, lvmcache_get_suspended_vg() will return the old VG instead of the new VG, and it will resume LVs based on the old metadata.	2017-11-01 16:39:35 -05:00
David Teigland	d02c5ec33e	vgcreate: improve the use of label_scan The old code was doing unnecessary label scans when checking to see if the new VG name exists. A single label_scan is sufficient if it is done after the new VG lock is held.	2017-11-01 16:39:35 -05:00
David Teigland	f1228fda71	lvmetad: use new label_scan for update from pvscan Take advantage of the common implementation with aio and reduced disk reads.	2017-11-01 16:39:35 -05:00
David Teigland	15c4d305e1	lvmetad: use new label_scan for update from lvmlockd When lvmlockd indicates that the lvmetad cache is out of date because of changes by another node, lvmetad_pvscan_vg() rescans the devices in the VG to update lvmetad. Use the new label_scan in this function to use the common code and take advantage of the new aio and reduced reads.	2017-11-01 16:39:35 -05:00
David Teigland	e1120c1ca5	label_scan/vg_read: use label_read_data to avoid disk reads The new label_scan() function reads a large buffer of data from the start of the disk, and saves it so that multiple structs can be read from it. Previously, only the label_header was read from this buffer, and the code which needed data structures that immediately followed the label_header would read those from disk separately. This created a large number of small, unnecessary disk reads. In each place that the two read paths (label_scan and vg_read) need to read data from disk, first check if that data is already available from the label_read_data buffer, and if so just copy it from the buffer instead of reading from disk. Code changes ------------ - passing the label_read_data struct down through both read paths to make it available. - before every disk read, first check if the location and size of the desired piece of data exists fully in the label_read_data buffer, and if so copy it from there. Otherwise, use the existing code to read the data from disk. - adding some log_error messages on existing error paths that were already being updated for the reasons above. - using similar naming for parallel functions on the two parallel read paths that are being updated above. label_scan path calls: read_metadata_location_summary, text_read_metadata_summary vg_read path calls: read_metadata_location_vg, text_read_metadata_file Previously, those functions were named: label_scan path calls: vgname_from_mda, text_vgsummary_import vg_read path calls: _find_vg_rlocn, text_vg_import_fd I/O changes ----------- In the label_scan path, the following data is either copied from label_read_data or read from disk for each PV: - label_header and pv_header - mda_header (in _raw_read_mda_header) - vg metadata name (in read_metadata_location_summary) - vg metadata (in config_file_read_fd) Total of 4 reads per PV in the label_scan path. In the vg_read path, the following data is either copied from label_read_data or read from disk for each PV: - mda_header (in _raw_read_mda_header) - vg metadata name (in read_metadata_location_vg) - vg metadata (in config_file_read_fd) Total of 3 reads per PV in the vg_read path. For a common read/reporting command, each PV will be: - read by the command's initial lvmcache_label_scan() - read by lvmcache_label_rescan_vg() at the start of vg_read() - read by vg_read() Previously, this would cause 11 synchronous disk reads per PV: 4 from lvmcache_label_scan(), 4 from lvmcache_label_rescan_vg() and 3 from vg_read(). With this commit's optimization, there are now 2 async disk reads per PV: 1 from lvmcache_label_scan() and 1 from lvmcache_label_rescan_vg(). When a second mda is used on a PV, it is located at the end of the PV. This second mda and copy of metadata will not be found in the label_read_data buffer, and will always require separate disk reads.	2017-11-01 16:39:35 -05:00
David Teigland	6d51d8be53	independent metadata areas: fix bogus code Fix mixing bitwise & and logical && which was always 1 in any case.	2017-11-01 16:39:35 -05:00
David Teigland	b24750e5ff	label_scan: fix independent metadata areas This fixes the use of lvmcache_label_rescan_vg() in the previous commit for the special case of independent metadata areas. label scan is about discovering VG name to device associations using information from disks, but devices in VGs with independent metadata areas have no information on disk, so the label scan does nothing for these VGs/devices. With independent metadata areas, only the VG metadata found in files is used. This metadata is found and read in vg_read in the processing phase. lvmcache_label_rescan_vg() drops lvmcache info for the VG devices before repeating the label scan on them. In the case of independent metadata areas, there is no metadata on devices, so the label scan of the devices will find nothing, so will not recreate the necessary vginfo/info data in lvmcache for the VG. Fix this by setting a flag in the lvmcache vginfo struct indicating that the VG uses independent metadata areas, and label rescanning should be skipped. In the case of independent metadata areas, it is the metadata processing in the vg_read phase that sets up the lvmcache vginfo/info information, and label scan has no role.	2017-11-01 16:39:35 -05:00
David Teigland	3166d67041	label_scan: move to start of command LVM's general design for scanning/reading of metadata from disks is that a command begins with a discovery phase, called "label scan", in which it discovers which devices belong to lvm, what VGs exist on those devices, and which devices are associated with each VG. After this comes the processing phase, which is based around processing specific VGs. In this phase, lvm acquires a lock on the VG, and rescans the devices associated with that VG, i.e. it repeats the label scan steps on the devices in the VG in case something has changed between the initial label scan and taking the VG lock. This ensures that the command is processing the lastest, unchanging data on disk. This commit moves the location of these label scans to make them clearer and avoid unnecessary repeated calls to them. Previously, the initial label scan was called as a side effect from various utility functions. This would lead to it being called unnecessarily. It is an expensive operation, and should only be called when necessary. Also, this is a primary step in the function of the command, and as such it should be called prominently at the top level of command processing, not as a hidden side effect of a utility function. lvm knows exactly where and when the label scan needs to be done. Because of this, move the label scan calls from the internal functions to the top level of processing. Other specific instances of lvmcache_label_scan() are still called unnecessarily or unclearly by specific commands that do not use the common process_each functions. These will be improved in future commits. During the processing phase, rescanning labels for devices in a VG needs to be done after the VG lock is acquired in case things have changed since the initial label scan. This was being done by way of rescanning devices that had the INVALID flag set in lvmcache. This usually approximated the right set of devices, but it was not exact, and obfuscated the real requirement. Correct this by using a new function that rescans the devices in the VG: lvmcache_label_rescan_vg(). Apart from being inexact, the rescanning was extremely well hidden. _vg_read() would call ->create_instance(), _text_create_text_instance(), _create_vg_text_instance() which would call lvmcache_label_scan() which would call _scan_invalid() which repeats the label scan on devices flagged INVALID. lvmcache_label_rescan_vg() is now called prominently by _vg_read() directly.	2017-11-01 16:39:34 -05:00
David Teigland	0b2ce69613	label_scan: call new label_scan from lvmcache_label_scan To do label scanning, lvm code calls lvmcache_label_scan(). Change lvmcache_label_scan() to use the new label_scan() which can use async io, rather than implementing its own dev iter loop and calling the synchronous label_read() on each device. Also add lvmcache_label_rescan_vg() which calls the new label_scan_devs() which does label scanning on only the specified devices. This is for a subsequent commit and is not yet used.	2017-11-01 16:39:34 -05:00
David Teigland	5d8e42e924	label_scan: add new implementation for async and sync This adds the implementation without using it in the code. The code still calls label_read() on each individual device to do scanning. Calling the new label_scan() will use async io if async io is enabled in config settings. If not enabled, or if async io fails, label_scan() will fall back to using synchronous io. If only some aio ops fail, the code will attempt to perform synchronous io on just the ios that failed. Uses linux native aio system calls, not the posix wrappers which are messier and may not have all the latest linux capabilities. Internally, the same functionality is used before: - iterate through each visible device on the system, provided from from dev-cache - call _find_label_header on the dev to find the sector containing the label_header - call _text_read to look at the pv_header and mda locations after the pv_header - for each mda location, read the mda_header and the vg metadata - add info/vginfo structs to lvmcache which associate the device name (info) with the VG name (vginfo) so that vg_read can know which devices to read for a given VG name The new label scanning issues a "large" read beginning at the start of the device, where large is configurable, but intended to cover all the labels/headers/metadata that is located at the start of the device. This large data buffer from each device is saved in a global list using a new 'label_read_data' struct. Currently, this buffer is only used to find the label_header from the first four sectors of the device. In subsequent commits, other functions that read other structs/metadata will first try to find that data in the saved label_read_data buffer. In most common cases, the data they need can simply be copied out of the existing buffer, and they can avoid issuing another disk read to get it.	2017-11-01 16:39:30 -05:00
David Teigland	70d973675c	command: add settings to enable async io There are config settings to enable aio, and to configure the concurrency and read size.	2017-10-31 14:57:55 -05:00
David Teigland	a41ffd52d6	io: add low level async io support The interface consists of: - A context struct, one for the entire command. - An io struct, one per io operation (read). - dev_async_context_setup() creates an aio context. - dev_async_context_destroy() destroys an aio context. - dev_async_alloc_ios() allocates a specified number of io structs, along with an associated buffer for the data. - dev_async_free_ios() frees all the allocated io structs+buffers. - dev_async_io_get() gets an available io struct from those allocated in alloc_ios. If none are available, it will allocate a new io struct if under limit. - dev_async_io_put() puts a used io struct back into the set of unused io structs, making it available for get. - dev_async_read_submit() start an async read io. - dev_async_getevents() collect async io completions.	2017-10-27 10:11:17 -05:00