Fix incomplete revert for lvconvert.

Sun May  3 11:40:51 CEST 2009  Petr Rockai <me@mornfall.net>
  * Convert the straight instances of vg_lock_and_read to new vg_read(_for_update).


Author: Petr Rockai <prockai@redhat.com>
Committer: Dave Wysochanski <dwysocha@redhat.com>

COPYING.LIB

@@ -1,14 +1,14 @@
+		  GNU LESSER GENERAL PUBLIC LICENSE
+		       Version 2.1, February 1999
 
-		  GNU LIBRARY GENERAL PUBLIC LICENSE
-		       Version 2, June 1991
-
- Copyright (C) 1991 Free Software Foundation, Inc.
- 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
+ Copyright (C) 1991, 1999 Free Software Foundation, Inc.
+ 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
  Everyone is permitted to copy and distribute verbatim copies
  of this license document, but changing it is not allowed.
 
-[This is the first released version of the library GPL.  It is
- numbered 2 because it goes with version 2 of the ordinary GPL.]
+[This is the first released version of the Lesser GPL.  It also counts
+ as the successor of the GNU Library Public License, version 2, hence
+ the version number 2.1.]
 
 			    Preamble
 
@@ -17,97 +17,109 @@ freedom to share and change it.  By contrast, the GNU General Public
 Licenses are intended to guarantee your freedom to share and change
 free software--to make sure the software is free for all its users.
 
-  This license, the Library General Public License, applies to some
-specially designated Free Software Foundation software, and to any
-other libraries whose authors decide to use it.  You can use it for
-your libraries, too.
+  This license, the Lesser General Public License, applies to some
+specially designated software packages--typically libraries--of the
+Free Software Foundation and other authors who decide to use it.  You
+can use it too, but we suggest you first think carefully about whether
+this license or the ordinary General Public License is the better
+strategy to use in any particular case, based on the explanations below.
 
-  When we speak of free software, we are referring to freedom, not
-price.  Our General Public Licenses are designed to make sure that you
-have the freedom to distribute copies of free software (and charge for
-this service if you wish), that you receive source code or can get it
-if you want it, that you can change the software or use pieces of it
-in new free programs; and that you know you can do these things.
+  When we speak of free software, we are referring to freedom of use,
+not price.  Our General Public Licenses are designed to make sure that
+you have the freedom to distribute copies of free software (and charge
+for this service if you wish); that you receive source code or can get
+it if you want it; that you can change the software and use pieces of
+it in new free programs; and that you are informed that you can do
+these things.
 
   To protect your rights, we need to make restrictions that forbid
-anyone to deny you these rights or to ask you to surrender the rights.
-These restrictions translate to certain responsibilities for you if
-you distribute copies of the library, or if you modify it.
+distributors to deny you these rights or to ask you to surrender these
+rights.  These restrictions translate to certain responsibilities for
+you if you distribute copies of the library or if you modify it.
 
   For example, if you distribute copies of the library, whether gratis
 or for a fee, you must give the recipients all the rights that we gave
 you.  You must make sure that they, too, receive or can get the source
-code.  If you link a program with the library, you must provide
-complete object files to the recipients so that they can relink them
-with the library, after making changes to the library and recompiling
+code.  If you link other code with the library, you must provide
+complete object files to the recipients, so that they can relink them
+with the library after making changes to the library and recompiling
 it.  And you must show them these terms so they know their rights.
 
-  Our method of protecting your rights has two steps: (1) copyright
-the library, and (2) offer you this license which gives you legal
+  We protect your rights with a two-step method: (1) we copyright the
+library, and (2) we offer you this license, which gives you legal
 permission to copy, distribute and/or modify the library.
 
-  Also, for each distributor's protection, we want to make certain
-that everyone understands that there is no warranty for this free
-library.  If the library is modified by someone else and passed on, we
-want its recipients to know that what they have is not the original
-version, so that any problems introduced by others will not reflect on
-the original authors' reputations.
+  To protect each distributor, we want to make it very clear that
+there is no warranty for the free library.  Also, if the library is
+modified by someone else and passed on, the recipients should know
+that what they have is not the original version, so that the original
+author's reputation will not be affected by problems that might be
+introduced by others.
 
-  Finally, any free program is threatened constantly by software
-patents.  We wish to avoid the danger that companies distributing free
-software will individually obtain patent licenses, thus in effect
-transforming the program into proprietary software.  To prevent this,
-we have made it clear that any patent must be licensed for everyone's
-free use or not licensed at all.
+  Finally, software patents pose a constant threat to the existence of
+any free program.  We wish to make sure that a company cannot
+effectively restrict the users of a free program by obtaining a
+restrictive license from a patent holder.  Therefore, we insist that
+any patent license obtained for a version of the library must be
+consistent with the full freedom of use specified in this license.
 
-  Most GNU software, including some libraries, is covered by the ordinary
-GNU General Public License, which was designed for utility programs.  This
-license, the GNU Library General Public License, applies to certain
-designated libraries.  This license is quite different from the ordinary
-one; be sure to read it in full, and don't assume that anything in it is
-the same as in the ordinary license.
+  Most GNU software, including some libraries, is covered by the
+ordinary GNU General Public License.  This license, the GNU Lesser
+General Public License, applies to certain designated libraries, and
+is quite different from the ordinary General Public License.  We use
+this license for certain libraries in order to permit linking those
+libraries into non-free programs.
 
-  The reason we have a separate public license for some libraries is that
-they blur the distinction we usually make between modifying or adding to a
-program and simply using it.  Linking a program with a library, without
-changing the library, is in some sense simply using the library, and is
-analogous to running a utility program or application program.  However, in
-a textual and legal sense, the linked executable is a combined work, a
-derivative of the original library, and the ordinary General Public License
-treats it as such.
+  When a program is linked with a library, whether statically or using
+a shared library, the combination of the two is legally speaking a
+combined work, a derivative of the original library.  The ordinary
+General Public License therefore permits such linking only if the
+entire combination fits its criteria of freedom.  The Lesser General
+Public License permits more lax criteria for linking other code with
+the library.
 
-  Because of this blurred distinction, using the ordinary General
-Public License for libraries did not effectively promote software
-sharing, because most developers did not use the libraries.  We
-concluded that weaker conditions might promote sharing better.
+  We call this license the "Lesser" General Public License because it
+does Less to protect the user's freedom than the ordinary General
+Public License.  It also provides other free software developers Less
+of an advantage over competing non-free programs.  These disadvantages
+are the reason we use the ordinary General Public License for many
+libraries.  However, the Lesser license provides advantages in certain
+special circumstances.
 
-  However, unrestricted linking of non-free programs would deprive the
-users of those programs of all benefit from the free status of the
-libraries themselves.  This Library General Public License is intended to
-permit developers of non-free programs to use free libraries, while
-preserving your freedom as a user of such programs to change the free
-libraries that are incorporated in them.  (We have not seen how to achieve
-this as regards changes in header files, but we have achieved it as regards
-changes in the actual functions of the Library.)  The hope is that this
-will lead to faster development of free libraries.
+  For example, on rare occasions, there may be a special need to
+encourage the widest possible use of a certain library, so that it becomes
+a de-facto standard.  To achieve this, non-free programs must be
+allowed to use the library.  A more frequent case is that a free
+library does the same job as widely used non-free libraries.  In this
+case, there is little to gain by limiting the free library to free
+software only, so we use the Lesser General Public License.
+
+  In other cases, permission to use a particular library in non-free
+programs enables a greater number of people to use a large body of
+free software.  For example, permission to use the GNU C Library in
+non-free programs enables many more people to use the whole GNU
+operating system, as well as its variant, the GNU/Linux operating
+system.
+
+  Although the Lesser General Public License is Less protective of the
+users' freedom, it does ensure that the user of a program that is
+linked with the Library has the freedom and the wherewithal to run
+that program using a modified version of the Library.
 
   The precise terms and conditions for copying, distribution and
 modification follow.  Pay close attention to the difference between a
 "work based on the library" and a "work that uses the library".  The
-former contains code derived from the library, while the latter only
-works together with the library.
-
-  Note that it is possible for a library to be covered by the ordinary
-General Public License rather than by this special one.
+former contains code derived from the library, whereas the latter must
+be combined with the library in order to run.
 
-		  GNU LIBRARY GENERAL PUBLIC LICENSE
+		  GNU LESSER GENERAL PUBLIC LICENSE
    TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
 
-  0. This License Agreement applies to any software library which
-contains a notice placed by the copyright holder or other authorized
-party saying it may be distributed under the terms of this Library
-General Public License (also called "this License").  Each licensee is
-addressed as "you".
+  0. This License Agreement applies to any software library or other
+program which contains a notice placed by the copyright holder or
+other authorized party saying it may be distributed under the terms of
+this Lesser General Public License (also called "this License").
+Each licensee is addressed as "you".
 
   A "library" means a collection of software functions and/or data
 prepared so as to be conveniently linked with application programs
@@ -256,7 +268,7 @@ distribute the object code for the work under the terms of Section 6.
 Any executables containing that work also fall under Section 6,
 whether or not they are linked directly with the Library itself.
 
-  6. As an exception to the Sections above, you may also compile or
+  6. As an exception to the Sections above, you may also combine or
 link a "work that uses the Library" with the Library to produce a
 work containing portions of the Library, and distribute that work
 under terms of your choice, provided that the terms permit
@@ -283,23 +295,31 @@ of these things:
     Library will not necessarily be able to recompile the application
     to use the modified definitions.)
 
-    b) Accompany the work with a written offer, valid for at
+    b) Use a suitable shared library mechanism for linking with the
+    Library.  A suitable mechanism is one that (1) uses at run time a
+    copy of the library already present on the user's computer system,
+    rather than copying library functions into the executable, and (2)
+    will operate properly with a modified version of the library, if
+    the user installs one, as long as the modified version is
+    interface-compatible with the version that the work was made with.
+
+    c) Accompany the work with a written offer, valid for at
     least three years, to give the same user the materials
     specified in Subsection 6a, above, for a charge no more
     than the cost of performing this distribution.
 
-    c) If distribution of the work is made by offering access to copy
+    d) If distribution of the work is made by offering access to copy
     from a designated place, offer equivalent access to copy the above
     specified materials from the same place.
 
-    d) Verify that the user has already received a copy of these
+    e) Verify that the user has already received a copy of these
     materials or that you have already sent this user a copy.
 
   For an executable, the required form of the "work that uses the
 Library" must include any data and utility programs needed for
 reproducing the executable from it.  However, as a special exception,
-the source code distributed need not include anything that is normally
-distributed (in either source or binary form) with the major
+the materials to be distributed need not include anything that is
+normally distributed (in either source or binary form) with the major
 components (compiler, kernel, and so on) of the operating system on
 which the executable runs, unless that component itself accompanies
 the executable.
@@ -348,7 +368,7 @@ Library), the recipient automatically receives a license from the
 original licensor to copy, distribute, link with or modify the Library
 subject to these terms and conditions.  You may not impose any further
 restrictions on the recipients' exercise of the rights granted herein.
-You are not responsible for enforcing compliance by third parties to
+You are not responsible for enforcing compliance by third parties with
 this License.
 
   11. If, as a consequence of a court judgment or allegation of patent
@@ -391,7 +411,7 @@ excluded.  In such case, this License incorporates the limitation as if
 written in the body of this License.
 
   13. The Free Software Foundation may publish revised and/or new
-versions of the Library General Public License from time to time.
+versions of the Lesser General Public License from time to time.
 Such new versions will be similar in spirit to the present version,
 but may differ in detail to address new problems or concerns.
 
@@ -437,7 +457,7 @@ DAMAGES.
 
 		     END OF TERMS AND CONDITIONS
 
-     Appendix: How to Apply These Terms to Your New Libraries
+           How to Apply These Terms to Your New Libraries
 
   If you develop a new library, and you want it to be of the greatest
 possible use to the public, we recommend making it free software that
@@ -454,19 +474,18 @@ convey the exclusion of warranty; and each file should have at least the
     Copyright (C) <year>  <name of author>
 
     This library is free software; you can redistribute it and/or
-    modify it under the terms of the GNU Library General Public
+    modify it under the terms of the GNU Lesser General Public
     License as published by the Free Software Foundation; either
-    version 2 of the License, or (at your option) any later version.
+    version 2.1 of the License, or (at your option) any later version.
 
     This library 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
-    Library General Public License for more details.
+    Lesser General Public License for more details.
 
-    You should have received a copy of the GNU Library General Public
-    License along with this library; if not, write to the Free
-    Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
-    MA 02111-1307, USA
+    You should have received a copy of the GNU Lesser General Public
+    License along with this library; if not, write to the Free Software
+    Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 
 Also add information on how to contact you by electronic and paper mail.
 
@@ -481,3 +500,5 @@ necessary.  Here is a sample; alter the names:
   Ty Coon, President of Vice
 
 That's all there is to it!
+
+

INSTALL

@@ -1,44 +1,30 @@
-LVM2 installation
-=================
+Installation
+============
 
-1) Install device-mapper
-
-   Ensure the device-mapper has been installed on the machine.
-
-   The device-mapper should be in the kernel (look for 'device-mapper'
-   messages in the kernel logs) and /usr/include/libdevmapper.h 
-   and libdevmapper.so should be present.
-
-   The device-mapper is available from:
-     ftp://ftp.sistina.com/pub/LVM2/device-mapper/
-
-
-2) Generate custom makefiles.
+1) Generate custom makefiles.
 
    Run the 'configure' script from the top directory.
 
-   If you wish to use the built-in LVM2 shell and have GNU readline 
-   installed (http://www.gnu.org/directory/readline.html) use:
-     ./configure --enable-readline
-
    If you don't want to include the LVM1 backwards-compatibility code use:
      ./configure --with-lvm1=none 
 
    To separate the LVM1 support into a shared library loaded by lvm.conf use:
      ./configure --with-lvm1=shared
 
+   Use ./configure --help to see other options.
 
-3) Build and install LVM2.
+2) Build and install.
 
-   Run 'make install' from the top directory.
+   Run 'make' from the top directory to build everything you configured.
+   Run 'make install' to build and install everything you configured.
 
+   If you only want the device-mapper libraries and tools use
+   'make device-mapper' or 'make install_device-mapper'.
 
-4) Create a configuration file
+3) If using LVM2, create a configuration file.
 
    The tools will work fine without a configuration file being
    present, but you ought to review the example file in doc/example.conf.
-   For example, specifying the devices that LVM2 is to use can
-   make the tools run more efficiently - and avoid scanning /dev/cdrom!
 
 Please also refer to the WHATS_NEW file and the manual pages for the 
 individual commands.

Makefile.in

@@ -1,8 +1,8 @@
 #
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
-# Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -16,44 +16,102 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
-SUBDIRS = doc include man 
+SUBDIRS = doc include man scripts
 
 ifeq ("@INTL@", "yes")
   SUBDIRS += po
 endif
 
-SUBDIRS += lib tools daemons
+SUBDIRS += lib tools daemons libdm
 
-ifeq ("@DMEVENTD@", "yes")
-  SUBDIRS += dmeventd
+ifeq ("@APPLIB@", "yes")
+  SUBDIRS += liblvm
 endif
 
 ifeq ($(MAKECMDGOALS),distclean)
   SUBDIRS += daemons/clvmd \
-	     dmeventd \
+	     daemons/dmeventd/plugins \
+	     daemons/dmeventd \
 	     lib/format1 \
 	     lib/format_pool \
 	     lib/locking \
 	     lib/mirror \
 	     lib/snapshot \
-	     po \
-	     test/mm test/device test/format1 test/regex test/filters
+	     test/api \
+	     test \
+	     po
+  DISTCLEAN_TARGETS += lib/misc/configure.h lib/misc/lvm-version.h
+  DISTCLEAN_DIRS += lcov_reports*
 endif
 
 include make.tmpl
 
-daemons: lib
-lib: include
-tools: lib
-dmeventd: tools
-po: tools daemons dmeventd
+libdm: include
+lib: libdm
+liblvm: lib
+daemons: lib tools
+tools: lib device-mapper
+po: tools daemons
+
+libdm.device-mapper: include.device-mapper
+daemons.device-mapper: libdm.device-mapper
+tools.device-mapper: libdm.device-mapper
+device-mapper: tools.device-mapper daemons.device-mapper man.device-mapper
 
 ifeq ("@INTL@", "yes")
 lib.pofile: include.pofile
 tools.pofile: lib.pofile
 daemons.pofile: lib.pofile
-dmeventd.pofile: tools.pofile
-po.pofile: tools.pofile daemons.pofile dmeventd.pofile
+po.pofile: tools.pofile daemons.pofile
 pofile: po.pofile
 endif
 
+ifneq ("@CFLOW_CMD@", "")
+tools.cflow: lib.cflow
+cflow: tools.cflow
+endif
+
+ifneq ("@CSCOPE_CMD@", "")
+cscope.out: tools
+	@CSCOPE_CMD@ -b -R
+all: cscope.out
+endif
+
+check: all
+	$(MAKE) -C test all
+
+ifneq ("@LCOV@", "")
+.PHONY: lcov-reset lcov lcov-dated
+
+ifeq ($(MAKECMDGOALS),lcov-dated)
+LCOV_REPORTS_DIR=$(top_srcdir)/lcov_reports-$(shell date +%Y%m%d%k%M%S)
+else
+LCOV_REPORTS_DIR=$(top_srcdir)/lcov_reports
+endif
+
+lcov-reset:
+	$(LCOV) -d $(top_srcdir)/dmeventd --zerocounters
+	$(LCOV) -d $(top_srcdir)/libdm --zerocounters
+	$(LCOV) -d $(top_srcdir)/lib --zerocounters
+	$(LCOV) -d $(top_srcdir)/tools --zerocounters
+
+lcov: all
+	$(RM) -rf $(LCOV_REPORTS_DIR)
+	$(MKDIR_P) $(LCOV_REPORTS_DIR)
+	$(LCOV) -b ${top_srcdir}/libdm -d $(top_srcdir)/libdm -c -o $(LCOV_REPORTS_DIR)/libdm.info
+	$(LCOV) -b $(top_srcdir)/lib -d $(top_srcdir)/lib -c -o $(LCOV_REPORTS_DIR)/lib.info
+	$(LCOV) -b $(top_srcdir)/tools -d $(top_srcdir)/tools -c -o $(LCOV_REPORTS_DIR)/tools.info
+	DMEVENTD_INFO="$(LCOV_REPORTS_DIR)/dmeventd.info" ;\
+	DMEVENTD_INFO_A="-a $$DMEVENTDINFO" ;\
+	$(LCOV) -b $(top_srcdir)/dmeventd -d $(top_srcdir)/dmeventd -c -o $$DMEVENTD_INFO || DMEVENTD_INFO_A="" ;\
+	$(LCOV) $$DMEVENTD_INFO_A -a $(LCOV_REPORTS_DIR)/lib.info \
+		-a $(LCOV_REPORTS_DIR)/libdm.info \
+		-a $(LCOV_REPORTS_DIR)/tools.info \
+		-o $(LCOV_REPORTS_DIR)/lvm.info
+ifneq ("@GENHTML@", "")
+	$(GENHTML) -o $(LCOV_REPORTS_DIR) -p $(top_srcdir) $(LCOV_REPORTS_DIR)/lvm.info
+endif
+
+lcov-dated: lcov
+
+endif

README

@@ -1,23 +1,27 @@
-This directory contains LVM2, the new version of the userland LVM
-tools designed for the new device-mapper for the Linux kernel.
+This tree contains the LVM2 and device-mapper tools and libraries.
 
-The device-mapper needs to be installed before compiling these LVM2 tools.
-
-For more information about LVM2 read the WHATS_NEW file.
+For more information about LVM2 read the changelog in the WHATS_NEW file.
 Installation instructions are in INSTALL.
 
 There is no warranty - see COPYING and COPYING.LIB.
 
 Tarballs are available from:
   ftp://sources.redhat.com/pub/lvm2/
-  ftp://sources.redhat.com/pub/dm/
 
 To access the CVS tree use:
   cvs -d :pserver:cvs@sources.redhat.com:/cvs/lvm2 login
   CVS password: cvs
   cvs -d :pserver:cvs@sources.redhat.com:/cvs/lvm2 co LVM2
 
-Mailing list for discussion/bug reports etc.
+Mailing list for general discussion related to LVM2:
   linux-lvm@redhat.com
   Subscribe from https://www.redhat.com/mailman/listinfo/linux-lvm
 
+Mailing list for LVM2 development, patches and commits:
+  lvm-devel@redhat.com
+  Subscribe from https://www.redhat.com/mailman/listinfo/linux-lvm
+
+Mailing list for device-mapper development, including kernel patches
+and multipath-tools:
+  dm-devel@redhat.com
+  Subscribe from https://www.redhat.com/mailman/listinfo/dm-devel

VERSION

@@ -1 +1 @@
-2.02.02-cvs (2006-02-07)
+2.02.48-cvs (2009-06-26)

VERSION_DM

@@ -0,0 +1 @@
+1.02.33-cvs (2009-06-26)

WHATS_NEW_DM

@@ -1,3 +1,231 @@
+Version 1.02.33 - 26th June 2009
+================================
+  Do not fallback to default major number, provide new function instead (1.02.31)
+  Do not fork daemon when dmeventd cannot be found.
+  Add crypt target handling to libdevmapper node.
+  Add splitname command to dmsetup.
+  Add subsystem, vg_name, lv_name, lv_layer fields to dmsetup reports.
+  Make mempool optional in dm_split_lvm_name().
+
+Version 1.02.32 - 21st May 2009
+===============================
+  Only generate libdevmapper.a when configured to link statically.
+  Export dm_tree_node_size_changed() from libdevmapper.
+  Propagate the table size_changed property up the dm device tree.
+  Detect failure to free memory pools when releasing the library.
+  Fix segfault when getopt processes dmsetup -U, -G and -M options.
+
+Version 1.02.31 - 3rd March 2009
+================================
+  If kernel supports only one dm major number, use in place of any supplied.
+
+Version 1.02.30 - 26th January 2009
+====================================
+  Add "all" field to reports expanding to all fields of report type.
+  Enforce device name length and character limitations in libdm.
+  Replace _dm_snprintf with EMIT_PARAMS macro for creating target lines.
+
+Version 1.02.29 - 10th November 2008
+====================================
+  Merge device-mapper into the LVM2 tree.
+  Split out dm-logging.h from log.h.
+  Use lvm-types.h.
+  Add usrsbindir to configure.
+
+Version 1.02.28 - 18th September 2008
+=====================================
+  Only resume devices in dm_tree_preload_children if size changes.
+  Extend deptree buffers so the largest possible device numbers fit.
+  Generate versioned libdevmapper-event.so.
+  Underline longer report help text headings.
+
+Version 1.02.27 - 25th June 2008
+================================
+  Align struct memblock in dbg_malloc for sparc.
+  Add --unquoted and --rows to dmsetup.
+  Avoid compiler warning about cast in dmsetup.c's OFFSET_OF macro.
+  Fix inverted no_flush debug message.
+  Remove --enable-jobs from configure. (Set at runtime instead.)
+  Bring configure.in and list.h into line with the lvm2 versions.
+
+Version 1.02.26 - 6th June 2008
+===============================
+  Initialise params buffer to empty string in _emit_segment.
+  Skip add_dev_node when ioctls disabled.
+  Make dm_hash_iter safe against deletion.
+  Accept a NULL pointer to dm_free silently.
+  Add tables_loaded, readonly and suspended columns to reports.
+  Add --nameprefixes to dmsetup.
+  Add field name prefix option to reporting functions.
+  Calculate string size within dm_pool_grow_object.
+
+Version 1.02.25 - 10th April 2008
+=================================
+  Remove redundant if-before-free tests.
+  Use log_warn for reporting field help text instead of log_print.
+  Change cluster mirror log type name (s/clustered_/clustered-/)
+
+Version 1.02.24 - 20th December 2007
+====================================
+  Fix deptree to pass new name to _resume_node after a rename.
+  Suppress other node operations if node is deleted.
+  Add node operation stack debug messages.
+  Report error when empty device name passed to readahead functions.
+  Fix minimum readahead debug message.
+
+Version 1.02.23 - 5th December 2007
+===================================
+  Update dm-ioctl.h after removal of compat code.
+  Add readahead support to libdevmapper and dmsetup.
+  Fix double free in a libdevmapper-event error path.
+  Fix configure --with-dmeventd-path substitution.
+  Allow a DM_DEV_DIR environment variable to override /dev in dmsetup.
+  Create a libdevmapper.so.$LIB_VERSION symlink within the build tree.
+  Avoid static link failure with some SELinux libraries that require libpthread.
+  Remove obsolete dmfs code from tree and update INSTALL.
+
+Version 1.02.22 - 21st August 2007
+==================================
+  Fix inconsistent licence notices: executables are GPLv2; libraries LGPLv2.1.
+  Update to use autoconf 2.61, while still supporting 2.57.
+  Avoid repeated dm_task free on some dm_event_get_registered_device errors.
+  Introduce log_sys_* macros from LVM2.
+  Export dm_fclose and dm_create_dir; remove libdm-file.h.
+  Don't log EROFS mkdir failures in _create_dir_recursive (for LVM2).
+  Add fclose wrapper dm_fclose that catches write failures (using ferror).
+
+Version 1.02.21 - 13th July 2007
+================================
+  Introduce _LOG_STDERR to send log_warn() messages to stderr not stdout.
+  Fix dmsetup -o devno string termination. (1.02.20)
+
+Version 1.02.20 - 15th June 2007
+================================
+  Fix default dmsetup report buffering and add --unbuffered.
+  Add tree-based and dependency fields to dmsetup reports.
+
+Version 1.02.19 - 27th April 2007
+=================================
+  Standardise protective include file #defines.
+  Add regex functions to library.
+  Avoid trailing separator in reports when there are hidden sort fields.
+  Fix segfault in 'dmsetup status' without --showkeys against crypt target.
+  Deal with some more compiler warnings.
+  Introduce _add_field() and _is_same_field() to libdm-report.c.
+  Fix some libdevmapper-event and dmeventd memory leaks.
+  Remove unnecessary memset() return value checks.
+  Fix a few leaks in reporting error paths. [1.02.15+]
+
+Version 1.02.18 - 13th February 2007
+====================================
+  Improve dmeventd messaging protocol: drain pipe and tag messages.
+
+Version 1.02.17 - 29th January 2007
+===================================
+  Add recent reporting options to dmsetup man page.
+  Revise some report fields names.
+  Add dmsetup 'help' command and update usage text.
+  Use fixed-size fields in report interface and reorder.
+
+Version 1.02.16 - 25th January 2007
+===================================
+  Add some missing close() and fclose() return value checks.
+  Migrate dmsetup column-based output over to new libdevmapper report framework.
+  Add descriptions to reporting field definitions.
+  Add a dso-private variable to dmeventd dso interface.
+  Add dm_event_handler_[gs]et_timeout functions.
+  Streamline dm_report_field_* interface.
+  Add cmdline debug & version options to dmeventd.
+  Add DM_LIB_VERSION definition to configure.h.
+  Suppress 'Unrecognised field' error if report field is 'help'.
+  Add --separator and --sort to dmsetup (unused).
+  Make alignment flag optional when specifying report fields.
+
+Version 1.02.15 - 17th January 2007
+===================================
+  Add basic reporting functions to libdevmapper.
+  Fix a malloc error path in dmsetup message.
+  More libdevmapper-event interface changes and fixes.
+  Rename dm_saprintf() to dm_asprintf().
+  Report error if NULL pointer is supplied to dm_strdup_aux().
+  Reinstate dm_event_get_registered_device.
+
+Version 1.02.14 - 11th January 2007
+===================================
+  Add dm_saprintf().
+  Use CFLAGS when linking so mixed sparc builds can supply -m64.
+  Add dm_tree_use_no_flush_suspend().
+  Lots of dmevent changes including revised interface.
+  Export dm_basename().
+  Cope with a trailing space when comparing tables prior to possible reload.
+  Fix dmeventd to cope if monitored device disappears.
+
+Version 1.02.13 - 28 Nov 2006
+=============================
+  Update dmsetup man page (setgeometry & message).
+  Fix dmsetup free after getline with debug.
+  Suppress encryption key in 'dmsetup table' output unless --showkeys supplied.
+
+Version 1.02.12 - 13 Oct 2006
+=============================
+  Avoid deptree attempting to suspend a device that's already suspended.
+
+Version 1.02.11 -  12 Oct 2006
+==============================
+  Add suspend noflush support.
+  Add basic dmsetup loop support.
+  Switch dmsetup to use dm_malloc and dm_free.
+
+Version 1.02.10 - 19 Sep 2006
+=============================
+  Add dm_snprintf(), dm_split_words() and dm_split_lvm_name() to libdevmapper.
+  Reorder mm bounds_check code to reduce window for a dmeventd race.
+
+Version 1.02.09 - 15 Aug 2006
+=============================
+  Add --table argument to dmsetup for a one-line table.
+  Abort if errors are found during cmdline option processing.
+  Add lockfs indicator to debug output.
+
+Version 1.02.08 - 17 July 2006
+==============================
+  Append full patch to check in emails.
+  Avoid duplicate dmeventd subdir with 'make distclean'.
+  Update dmsetup man page.
+  Add --force to dmsetup remove* to load error target.
+  dmsetup remove_all also performs mknodes.
+  Don't suppress identical table reloads if permission changes.
+  Fix corelog segment line.
+  Suppress some compiler warnings.
+
+Version 1.02.07 - 11 May 2006
+=============================
+  Add DM_CORELOG flag to dm_tree_node_add_mirror_target().
+  Avoid a dmeventd compiler warning.
+
+Version 1.02.06 - 10 May 2006
+=============================
+  Move DEFS into configure.h.
+  Fix leaks in error paths found by coverity.
+  Remove dmsetup line buffer limitation.
+
+Version 1.02.05 - 19 Apr 2006
+=============================
+  Separate install_include target in makefiles.
+  Separate out DEFS from CFLAGS.
+  Support pkg-config.
+  Check for libsepol.
+
+Version 1.02.04 - 14 Apr 2006
+=============================
+  Bring dmsetup man page up-to-date.
+  Use name-based device refs if kernel doesn't support device number refs.
+  Fix memory leak (struct dm_ioctl) when struct dm_task is reused.
+  If _create_and_load_v4 fails part way through, revert the creation.
+  dmeventd thread/fifo fixes.
+  Add file & line to dm_strdup_aux().
+  Add setgeometry.
+
 Version 1.02.03 - 7 Feb 2006
 ============================
   Add exported functions to set uid, gid and mode.
@@ -157,4 +385,3 @@ Version 1.00.08 - 27 Feb 2004
   Fixed DESTDIR for make install/install_static_lib.
   Updated README/INSTALL to reflect move to sources.redhat.com.
   Updated autoconf files to 2003-06-17.
-

autoconf/config.sub

@@ -1,9 +1,10 @@
 #! /bin/sh
 # Configuration validation subroutine script.
 #   Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
-#   2000, 2001, 2002, 2003 Free Software Foundation, Inc.
+#   2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
+#   Free Software Foundation, Inc.
 
-timestamp='2003-06-17'
+timestamp='2008-01-16'
 
 # This file is (in principle) common to ALL GNU software.
 # The presence of a machine in this file suggests that SOME GNU software
@@ -21,14 +22,15 @@ timestamp='2003-06-17'
 #
 # 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.
-
+# Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA
+# 02110-1301, 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.
 
+
 # Please send patches to <config-patches@gnu.org>.  Submit a context
 # diff and a properly formatted ChangeLog entry.
 #
@@ -70,8 +72,8 @@ Report bugs and patches to <config-patches@gnu.org>."
 version="\
 GNU config.sub ($timestamp)
 
-Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001
-Free Software Foundation, Inc.
+Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
+2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
 
 This is free software; see the source for copying conditions.  There is NO
 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@@ -83,11 +85,11 @@ Try \`$me --help' for more information."
 while test $# -gt 0 ; do
   case $1 in
     --time-stamp | --time* | -t )
-       echo "$timestamp" ; exit 0 ;;
+       echo "$timestamp" ; exit ;;
     --version | -v )
-       echo "$version" ; exit 0 ;;
+       echo "$version" ; exit ;;
     --help | --h* | -h )
-       echo "$usage"; exit 0 ;;
+       echo "$usage"; exit ;;
     -- )     # Stop option processing
        shift; break ;;
     - )	# Use stdin as input.
@@ -99,7 +101,7 @@ while test $# -gt 0 ; do
     *local*)
        # First pass through any local machine types.
        echo $1
-       exit 0;;
+       exit ;;
 
     * )
        break ;;
@@ -118,7 +120,9 @@ esac
 # Here we must recognize all the valid KERNEL-OS combinations.
 maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
 case $maybe_os in
-  nto-qnx* | linux-gnu* | freebsd*-gnu* | netbsd*-gnu* | storm-chaos* | os2-emx* | rtmk-nova*)
+  nto-qnx* | linux-gnu* | linux-dietlibc | linux-newlib* | linux-uclibc* | \
+  uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | knetbsd*-gnu* | netbsd*-gnu* | \
+  storm-chaos* | os2-emx* | rtmk-nova*)
     os=-$maybe_os
     basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
     ;;
@@ -144,7 +148,7 @@ case $os in
 	-convergent* | -ncr* | -news | -32* | -3600* | -3100* | -hitachi* |\
 	-c[123]* | -convex* | -sun | -crds | -omron* | -dg | -ultra | -tti* | \
 	-harris | -dolphin | -highlevel | -gould | -cbm | -ns | -masscomp | \
-	-apple | -axis)
+	-apple | -axis | -knuth | -cray)
 		os=
 		basic_machine=$1
 		;;
@@ -169,6 +173,10 @@ case $os in
 	-hiux*)
 		os=-hiuxwe2
 		;;
+	-sco6)
+		os=-sco5v6
+		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		;;
 	-sco5)
 		os=-sco3.2v5
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
@@ -185,6 +193,10 @@ case $os in
 		# Don't forget version if it is 3.2v4 or newer.
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
 		;;
+	-sco5v6*)
+		# Don't forget version if it is 3.2v4 or newer.
+		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		;;
 	-sco*)
 		os=-sco3.2v2
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
@@ -228,14 +240,17 @@ case $basic_machine in
 	| a29k \
 	| alpha | alphaev[4-8] | alphaev56 | alphaev6[78] | alphapca5[67] \
 	| alpha64 | alpha64ev[4-8] | alpha64ev56 | alpha64ev6[78] | alpha64pca5[67] \
-	| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr \
+	| am33_2.0 \
+	| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr | avr32 \
+	| bfin \
 	| c4x | clipper \
 	| d10v | d30v | dlx | dsp16xx \
-	| fr30 | frv \
+	| fido | fr30 | frv \
 	| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
 	| i370 | i860 | i960 | ia64 \
-	| ip2k \
-	| m32r | m68000 | m68k | m88k | mcore \
+	| ip2k | iq2000 \
+	| m32c | m32r | m32rle | m68000 | m68k | m88k \
+	| maxq | mb | microblaze | mcore | mep \
 	| mips | mipsbe | mipseb | mipsel | mipsle \
 	| mips16 \
 	| mips64 | mips64el \
@@ -244,27 +259,33 @@ case $basic_machine in
 	| mips64vr4100 | mips64vr4100el \
 	| mips64vr4300 | mips64vr4300el \
 	| mips64vr5000 | mips64vr5000el \
+	| mips64vr5900 | mips64vr5900el \
 	| mipsisa32 | mipsisa32el \
 	| mipsisa32r2 | mipsisa32r2el \
 	| mipsisa64 | mipsisa64el \
+	| mipsisa64r2 | mipsisa64r2el \
 	| mipsisa64sb1 | mipsisa64sb1el \
 	| mipsisa64sr71k | mipsisa64sr71kel \
 	| mipstx39 | mipstx39el \
 	| mn10200 | mn10300 \
+	| mt \
 	| msp430 \
+	| nios | nios2 \
 	| ns16k | ns32k \
-	| openrisc | or32 \
+	| or32 \
 	| pdp10 | pdp11 | pj | pjl \
 	| powerpc | powerpc64 | powerpc64le | powerpcle | ppcbe \
 	| pyramid \
-	| sh | sh[1234] | sh[23]e | sh[34]eb | shbe | shle | sh[1234]le | sh3ele \
+	| score \
+	| sh | sh[1234] | sh[24]a | sh[23]e | sh[34]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
 	| sh64 | sh64le \
-	| sparc | sparc64 | sparc86x | sparclet | sparclite | sparcv9 | sparcv9b \
-	| strongarm \
+	| sparc | sparc64 | sparc64b | sparc64v | sparc86x | sparclet | sparclite \
+	| sparcv8 | sparcv9 | sparcv9b | sparcv9v \
+	| spu | strongarm \
 	| tahoe | thumb | tic4x | tic80 | tron \
 	| v850 | v850e \
 	| we32k \
-	| x86 | xscale | xstormy16 | xtensa \
+	| x86 | xc16x | xscale | xscalee[bl] | xstormy16 | xtensa \
 	| z8k)
 		basic_machine=$basic_machine-unknown
 		;;
@@ -275,6 +296,9 @@ case $basic_machine in
 		;;
 	m88110 | m680[12346]0 | m683?2 | m68360 | m5200 | v70 | w65 | z8k)
 		;;
+	ms1)
+		basic_machine=mt-unknown
+		;;
 
 	# We use `pc' rather than `unknown'
 	# because (1) that's what they normally are, and
@@ -294,20 +318,20 @@ case $basic_machine in
 	| alpha64-* | alpha64ev[4-8]-* | alpha64ev56-* | alpha64ev6[78]-* \
 	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* \
 	| arm-*  | armbe-* | armle-* | armeb-* | armv*-* \
-	| avr-* \
-	| bs2000-* \
+	| avr-* | avr32-* \
+	| bfin-* | bs2000-* \
 	| c[123]* | c30-* | [cjt]90-* | c4x-* | c54x-* | c55x-* | c6x-* \
-	| clipper-* | cydra-* \
+	| clipper-* | craynv-* | cydra-* \
 	| d10v-* | d30v-* | dlx-* \
 	| elxsi-* \
-	| f30[01]-* | f700-* | fr30-* | frv-* | fx80-* \
+	| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
 	| h8300-* | h8500-* \
 	| hppa-* | hppa1.[01]-* | hppa2.0-* | hppa2.0[nw]-* | hppa64-* \
 	| i*86-* | i860-* | i960-* | ia64-* \
-	| ip2k-* \
-	| m32r-* \
+	| ip2k-* | iq2000-* \
+	| m32c-* | m32r-* | m32rle-* \
 	| m68000-* | m680[012346]0-* | m68360-* | m683?2-* | m68k-* \
-	| m88110-* | m88k-* | mcore-* \
+	| m88110-* | m88k-* | maxq-* | mcore-* \
 	| mips-* | mipsbe-* | mipseb-* | mipsel-* | mipsle-* \
 	| mips16-* \
 	| mips64-* | mips64el-* \
@@ -316,33 +340,43 @@ case $basic_machine in
 	| mips64vr4100-* | mips64vr4100el-* \
 	| mips64vr4300-* | mips64vr4300el-* \
 	| mips64vr5000-* | mips64vr5000el-* \
+	| mips64vr5900-* | mips64vr5900el-* \
 	| mipsisa32-* | mipsisa32el-* \
 	| mipsisa32r2-* | mipsisa32r2el-* \
 	| mipsisa64-* | mipsisa64el-* \
+	| mipsisa64r2-* | mipsisa64r2el-* \
 	| mipsisa64sb1-* | mipsisa64sb1el-* \
 	| mipsisa64sr71k-* | mipsisa64sr71kel-* \
 	| mipstx39-* | mipstx39el-* \
+	| mmix-* \
+	| mt-* \
 	| msp430-* \
-	| none-* | np1-* | nv1-* | ns16k-* | ns32k-* \
+	| nios-* | nios2-* \
+	| none-* | np1-* | ns16k-* | ns32k-* \
 	| orion-* \
 	| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
 	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* | ppcbe-* \
 	| pyramid-* \
 	| romp-* | rs6000-* \
-	| sh-* | sh[1234]-* | sh[23]e-* | sh[34]eb-* | shbe-* \
+	| sh-* | sh[1234]-* | sh[24]a-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \
 	| shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \
-	| sparc-* | sparc64-* | sparc86x-* | sparclet-* | sparclite-* \
-	| sparcv9-* | sparcv9b-* | strongarm-* | sv1-* | sx?-* \
+	| sparc-* | sparc64-* | sparc64b-* | sparc64v-* | sparc86x-* | sparclet-* \
+	| sparclite-* \
+	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | strongarm-* | sv1-* | sx?-* \
 	| tahoe-* | thumb-* \
 	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* \
 	| tron-* \
 	| v850-* | v850e-* | vax-* \
 	| we32k-* \
-	| x86-* | x86_64-* | xps100-* | xscale-* | xstormy16-* \
-	| xtensa-* \
+	| x86-* | x86_64-* | xc16x-* | xps100-* | xscale-* | xscalee[bl]-* \
+	| xstormy16-* | xtensa*-* \
 	| ymp-* \
 	| z8k-*)
 		;;
+	# Recognize the basic CPU types without company name, with glob match.
+	xtensa*)
+		basic_machine=$basic_machine-unknown
+		;;
 	# Recognize the various machine names and aliases which stand
 	# for a CPU type and a company and sometimes even an OS.
 	386bsd)
@@ -359,6 +393,9 @@ case $basic_machine in
 		basic_machine=a29k-amd
 		os=-udi
 		;;
+    	abacus)
+		basic_machine=abacus-unknown
+		;;
 	adobe68k)
 		basic_machine=m68010-adobe
 		os=-scout
@@ -376,6 +413,9 @@ case $basic_machine in
 	amd64)
 		basic_machine=x86_64-pc
 		;;
+	amd64-*)
+		basic_machine=x86_64-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	amdahl)
 		basic_machine=580-amdahl
 		os=-sysv
@@ -407,6 +447,14 @@ case $basic_machine in
 		basic_machine=ns32k-sequent
 		os=-dynix
 		;;
+	blackfin)
+		basic_machine=bfin-unknown
+		os=-linux
+		;;
+	blackfin-*)
+		basic_machine=bfin-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	c90)
 		basic_machine=c90-cray
 		os=-unicos
@@ -435,12 +483,27 @@ case $basic_machine in
 		basic_machine=j90-cray
 		os=-unicos
 		;;
+	craynv)
+		basic_machine=craynv-cray
+		os=-unicosmp
+		;;
+	cr16)
+		basic_machine=cr16-unknown
+		os=-elf
+		;;
 	crds | unos)
 		basic_machine=m68k-crds
 		;;
+	crisv32 | crisv32-* | etraxfs*)
+		basic_machine=crisv32-axis
+		;;
 	cris | cris-* | etrax*)
 		basic_machine=cris-axis
 		;;
+	crx)
+		basic_machine=crx-unknown
+		os=-elf
+		;;
 	da30 | da30-*)
 		basic_machine=m68k-da30
 		;;
@@ -463,6 +526,10 @@ case $basic_machine in
 		basic_machine=m88k-motorola
 		os=-sysv3
 		;;
+	djgpp)
+		basic_machine=i586-pc
+		os=-msdosdjgpp
+		;;
 	dpx20 | dpx20-*)
 		basic_machine=rs6000-bull
 		os=-bosx
@@ -613,6 +680,14 @@ case $basic_machine in
 		basic_machine=m68k-isi
 		os=-sysv
 		;;
+	m68knommu)
+		basic_machine=m68k-unknown
+		os=-linux
+		;;
+	m68knommu-*)
+		basic_machine=m68k-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	m88k-omron*)
 		basic_machine=m88k-omron
 		;;
@@ -628,6 +703,10 @@ case $basic_machine in
 		basic_machine=i386-pc
 		os=-mingw32
 		;;
+	mingw32ce)
+		basic_machine=arm-unknown
+		os=-mingw32ce
+		;;
 	miniframe)
 		basic_machine=m68000-convergent
 		;;
@@ -641,10 +720,6 @@ case $basic_machine in
 	mips3*)
 		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`-unknown
 		;;
-	mmix*)
-		basic_machine=mmix-knuth
-		os=-mmixware
-		;;
 	monitor)
 		basic_machine=m68k-rom68k
 		os=-coff
@@ -657,6 +732,9 @@ case $basic_machine in
 		basic_machine=i386-pc
 		os=-msdos
 		;;
+	ms1-*)
+		basic_machine=`echo $basic_machine | sed -e 's/ms1-/mt-/'`
+		;;
 	mvs)
 		basic_machine=i370-ibm
 		os=-mvs
@@ -725,10 +803,6 @@ case $basic_machine in
 	np1)
 		basic_machine=np1-gould
 		;;
-	nv1)
-		basic_machine=nv1-cray
-		os=-unicosmp
-		;;
 	nsr-tandem)
 		basic_machine=nsr-tandem
 		;;
@@ -736,9 +810,12 @@ case $basic_machine in
 		basic_machine=hppa1.1-oki
 		os=-proelf
 		;;
-	or32 | or32-*)
+	openrisc | openrisc-*)
 		basic_machine=or32-unknown
-		os=-coff
+		;;
+	os400)
+		basic_machine=powerpc-ibm
+		os=-os400
 		;;
 	OSE68000 | ose68000)
 		basic_machine=m68000-ericsson
@@ -756,6 +833,14 @@ case $basic_machine in
 		basic_machine=i860-intel
 		os=-osf
 		;;
+	parisc)
+		basic_machine=hppa-unknown
+		os=-linux
+		;;
+	parisc-*)
+		basic_machine=hppa-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	pbd)
 		basic_machine=sparc-tti
 		;;
@@ -765,6 +850,12 @@ case $basic_machine in
 	pc532 | pc532-*)
 		basic_machine=ns32k-pc532
 		;;
+	pc98)
+		basic_machine=i386-pc
+		;;
+	pc98-*)
+		basic_machine=i386-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	pentium | p5 | k5 | k6 | nexgen | viac3)
 		basic_machine=i586-pc
 		;;
@@ -821,6 +912,10 @@ case $basic_machine in
 		basic_machine=i586-unknown
 		os=-pw32
 		;;
+	rdos)
+		basic_machine=i386-pc
+		os=-rdos
+		;;
 	rom68k)
 		basic_machine=m68k-rom68k
 		os=-coff
@@ -847,6 +942,10 @@ case $basic_machine in
 	sb1el)
 		basic_machine=mipsisa64sb1el-unknown
 		;;
+	sde)
+		basic_machine=mipsisa32-sde
+		os=-elf
+		;;
 	sei)
 		basic_machine=mips-sei
 		os=-seiux
@@ -858,6 +957,9 @@ case $basic_machine in
 		basic_machine=sh-hitachi
 		os=-hms
 		;;
+	sh5el)
+		basic_machine=sh5le-unknown
+		;;
 	sh64)
 		basic_machine=sh64-unknown
 		;;
@@ -947,6 +1049,10 @@ case $basic_machine in
 		basic_machine=tic6x-unknown
 		os=-coff
 		;;
+	tile*)
+		basic_machine=tile-unknown
+		os=-linux-gnu
+		;;
 	tx39)
 		basic_machine=mipstx39-unknown
 		;;
@@ -960,6 +1066,10 @@ case $basic_machine in
 	tower | tower-32)
 		basic_machine=m68k-ncr
 		;;
+	tpf)
+		basic_machine=s390x-ibm
+		os=-tpf
+		;;
 	udi29k)
 		basic_machine=a29k-amd
 		os=-udi
@@ -1003,6 +1113,10 @@ case $basic_machine in
 		basic_machine=hppa1.1-winbond
 		os=-proelf
 		;;
+	xbox)
+		basic_machine=i686-pc
+		os=-mingw32
+		;;
 	xps | xps100)
 		basic_machine=xps100-honeywell
 		;;
@@ -1033,6 +1147,9 @@ case $basic_machine in
 	romp)
 		basic_machine=romp-ibm
 		;;
+	mmix)
+		basic_machine=mmix-knuth
+		;;
 	rs6000)
 		basic_machine=rs6000-ibm
 		;;
@@ -1049,13 +1166,10 @@ case $basic_machine in
 	we32k)
 		basic_machine=we32k-att
 		;;
-	sh3 | sh4 | sh[34]eb | sh[1234]le | sh[23]ele)
+	sh[1234] | sh[24]a | sh[34]eb | sh[1234]le | sh[23]ele)
 		basic_machine=sh-unknown
 		;;
-	sh64)
-		basic_machine=sh64-unknown
-		;;
-	sparc | sparcv9 | sparcv9b)
+	sparc | sparcv8 | sparcv9 | sparcv9b | sparcv9v)
 		basic_machine=sparc-sun
 		;;
 	cydra)
@@ -1128,19 +1242,23 @@ case $os in
 	      | -aos* \
 	      | -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
 	      | -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
-	      | -hiux* | -386bsd* | -netbsd* | -openbsd* | -freebsd* | -riscix* \
-	      | -lynxos* | -bosx* | -nextstep* | -cxux* | -aout* | -elf* | -oabi* \
+	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
+	      | -openbsd* | -solidbsd* \
+	      | -ekkobsd* | -kfreebsd* | -freebsd* | -riscix* | -lynxos* \
+	      | -bosx* | -nextstep* | -cxux* | -aout* | -elf* | -oabi* \
 	      | -ptx* | -coff* | -ecoff* | -winnt* | -domain* | -vsta* \
 	      | -udi* | -eabi* | -lites* | -ieee* | -go32* | -aux* \
 	      | -chorusos* | -chorusrdb* \
 	      | -cygwin* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
-	      | -mingw32* | -linux-gnu* | -uxpv* | -beos* | -mpeix* | -udk* \
+	      | -mingw32* | -linux-gnu* | -linux-newlib* | -linux-uclibc* \
+	      | -uxpv* | -beos* | -mpeix* | -udk* \
 	      | -interix* | -uwin* | -mks* | -rhapsody* | -darwin* | -opened* \
 	      | -openstep* | -oskit* | -conix* | -pw32* | -nonstopux* \
 	      | -storm-chaos* | -tops10* | -tenex* | -tops20* | -its* \
 	      | -os2* | -vos* | -palmos* | -uclinux* | -nucleus* \
 	      | -morphos* | -superux* | -rtmk* | -rtmk-nova* | -windiss* \
-	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei*)
+	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly* \
+	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops*)
 	# Remember, each alternative MUST END IN *, to match a version number.
 		;;
 	-qnx*)
@@ -1158,12 +1276,15 @@ case $os in
 		os=`echo $os | sed -e 's|nto|nto-qnx|'`
 		;;
 	-sim | -es1800* | -hms* | -xray | -os68k* | -none* | -v88r* \
-	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* \
+	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* | -haiku* \
 	      | -macos* | -mpw* | -magic* | -mmixware* | -mon960* | -lnews*)
 		;;
 	-mac*)
 		os=`echo $os | sed -e 's|mac|macos|'`
 		;;
+	-linux-dietlibc)
+		os=-linux-dietlibc
+		;;
 	-linux*)
 		os=`echo $os | sed -e 's|linux|linux-gnu|'`
 		;;
@@ -1176,6 +1297,9 @@ case $os in
 	-opened*)
 		os=-openedition
 		;;
+        -os400*)
+		os=-os400
+		;;
 	-wince*)
 		os=-wince
 		;;
@@ -1197,6 +1321,9 @@ case $os in
 	-atheos*)
 		os=-atheos
 		;;
+	-syllable*)
+		os=-syllable
+		;;
 	-386bsd)
 		os=-bsd
 		;;
@@ -1219,6 +1346,9 @@ case $os in
 	-sinix*)
 		os=-sysv4
 		;;
+        -tpf*)
+		os=-tpf
+		;;
 	-triton*)
 		os=-sysv3
 		;;
@@ -1255,6 +1385,9 @@ case $os in
 	-kaos*)
 		os=-kaos
 		;;
+	-zvmoe)
+		os=-zvmoe
+		;;
 	-none)
 		;;
 	*)
@@ -1277,6 +1410,12 @@ else
 # system, and we'll never get to this point.
 
 case $basic_machine in
+        score-*)
+		os=-elf
+		;;
+        spu-*)
+		os=-elf
+		;;
 	*-acorn)
 		os=-riscix1.2
 		;;
@@ -1286,9 +1425,9 @@ case $basic_machine in
 	arm*-semi)
 		os=-aout
 		;;
-    c4x-* | tic4x-*)
-        os=-coff
-        ;;
+        c4x-* | tic4x-*)
+        	os=-coff
+		;;
 	# This must come before the *-dec entry.
 	pdp10-*)
 		os=-tops20
@@ -1314,6 +1453,9 @@ case $basic_machine in
 	m68*-cisco)
 		os=-aout
 		;;
+        mep-*)
+		os=-elf
+		;;
 	mips*-cisco)
 		os=-elf
 		;;
@@ -1332,9 +1474,15 @@ case $basic_machine in
 	*-be)
 		os=-beos
 		;;
+	*-haiku)
+		os=-haiku
+		;;
 	*-ibm)
 		os=-aix
 		;;
+    	*-knuth)
+		os=-mmixware
+		;;
 	*-wec)
 		os=-proelf
 		;;
@@ -1467,9 +1615,15 @@ case $basic_machine in
 			-mvs* | -opened*)
 				vendor=ibm
 				;;
+			-os400*)
+				vendor=ibm
+				;;
 			-ptx*)
 				vendor=sequent
 				;;
+			-tpf*)
+				vendor=ibm
+				;;
 			-vxsim* | -vxworks* | -windiss*)
 				vendor=wrs
 				;;
@@ -1494,7 +1648,7 @@ case $basic_machine in
 esac
 
 echo $basic_machine$os
-exit 0
+exit
 
 # Local variables:
 # eval: (add-hook 'write-file-hooks 'time-stamp)

autoconf/install-sh

@@ -1,7 +1,7 @@
 #!/bin/sh
 # install - install a program, script, or datafile
 
-scriptversion=2003-06-13.21
+scriptversion=2006-10-14.15
 
 # This originates from X11R5 (mit/util/scripts/install.sh), which was
 # later released in X11R6 (xc/config/util/install.sh) with the
@@ -39,15 +39,24 @@ scriptversion=2003-06-13.21
 # when there is no Makefile.
 #
 # This script is compatible with the BSD install script, but was written
-# from scratch.  It can only install one file at a time, a restriction
-# shared with many OS's install programs.
+# from scratch.
+
+nl='
+'
+IFS=" ""	$nl"
 
 # set DOITPROG to echo to test this script
 
 # Don't use :- since 4.3BSD and earlier shells don't like it.
 doit="${DOITPROG-}"
+if test -z "$doit"; then
+  doit_exec=exec
+else
+  doit_exec=$doit
+fi
 
-# put in absolute paths if you don't have them in your path; or use env. vars.
+# Put in absolute file names if you don't have them in your path;
+# or use environment vars.
 
 mvprog="${MVPROG-mv}"
 cpprog="${CPPROG-cp}"
@@ -58,10 +67,13 @@ stripprog="${STRIPPROG-strip}"
 rmprog="${RMPROG-rm}"
 mkdirprog="${MKDIRPROG-mkdir}"
 
-transformbasename=
-transform_arg=
-instcmd="$mvprog"
-chmodcmd="$chmodprog 0755"
+posix_glob=
+posix_mkdir=
+
+# Desired mode of installed file.
+mode=0755
+
+chmodcmd=$chmodprog
 chowncmd=
 chgrpcmd=
 stripcmd=
@@ -70,22 +82,27 @@ mvcmd="$mvprog"
 src=
 dst=
 dir_arg=
+dstarg=
+no_target_directory=
 
-usage="Usage: $0 [OPTION]... SRCFILE DSTFILE
-   or: $0 -d DIR1 DIR2...
+usage="Usage: $0 [OPTION]... [-T] SRCFILE DSTFILE
+   or: $0 [OPTION]... SRCFILES... DIRECTORY
+   or: $0 [OPTION]... -t DIRECTORY SRCFILES...
+   or: $0 [OPTION]... -d DIRECTORIES...
 
-In the first form, install SRCFILE to DSTFILE, removing SRCFILE by default.
-In the second, create the directory path DIR.
+In the 1st form, copy SRCFILE to DSTFILE.
+In the 2nd and 3rd, copy all SRCFILES to DIRECTORY.
+In the 4th, create DIRECTORIES.
 
 Options:
--b=TRANSFORMBASENAME
--c         copy source (using $cpprog) instead of moving (using $mvprog).
+-c         (ignored)
 -d         create directories instead of installing files.
--g GROUP   $chgrp installed files to GROUP.
--m MODE    $chmod installed files to MODE.
--o USER    $chown installed files to USER.
--s         strip installed files (using $stripprog).
--t=TRANSFORM
+-g GROUP   $chgrpprog installed files to GROUP.
+-m MODE    $chmodprog installed files to MODE.
+-o USER    $chownprog installed files to USER.
+-s         $stripprog installed files.
+-t DIRECTORY  install into DIRECTORY.
+-T         report an error if DSTFILE is a directory.
 --help     display this help and exit.
 --version  display version info and exit.
 
@@ -93,14 +110,9 @@ Environment variables override the default commands:
   CHGRPPROG CHMODPROG CHOWNPROG CPPROG MKDIRPROG MVPROG RMPROG STRIPPROG
 "
 
-while test -n "$1"; do
+while test $# -ne 0; do
   case $1 in
-    -b=*) transformbasename=`echo $1 | sed 's/-b=//'`
-        shift
-        continue;;
-
-    -c) instcmd=$cpprog
-        shift
+    -c) shift
         continue;;
 
     -d) dir_arg=true
@@ -112,11 +124,17 @@ while test -n "$1"; do
         shift
         continue;;
 
-    --help) echo "$usage"; exit 0;;
+    --help) echo "$usage"; exit $?;;
 
-    -m) chmodcmd="$chmodprog $2"
+    -m) mode=$2
         shift
         shift
+	case $mode in
+	  *' '* | *'	'* | *'
+'*	  | *'*'* | *'?'* | *'['*)
+	    echo "$0: invalid mode: $mode" >&2
+	    exit 1;;
+	esac
         continue;;
 
     -o) chowncmd="$chownprog $2"
@@ -128,155 +146,358 @@ while test -n "$1"; do
         shift
         continue;;
 
-    -t=*) transformarg=`echo $1 | sed 's/-t=//'`
-        shift
-        continue;;
+    -t) dstarg=$2
+	shift
+	shift
+	continue;;
 
-    --version) echo "$0 $scriptversion"; exit 0;;
+    -T) no_target_directory=true
+	shift
+	continue;;
 
-    *)  if test -z "$src"; then
-          src=$1
-        else
-          # this colon is to work around a 386BSD /bin/sh bug
-          :
-          dst=$1
-        fi
-        shift
-        continue;;
+    --version) echo "$0 $scriptversion"; exit $?;;
+
+    --)	shift
+	break;;
+
+    -*)	echo "$0: invalid option: $1" >&2
+	exit 1;;
+
+    *)  break;;
   esac
 done
 
-if test -z "$src"; then
-  echo "$0: no input file specified." >&2
-  exit 1
-fi
-
-if test -n "$dir_arg"; then
-  dst=$src
-  src=
-
-  if test -d "$dst"; then
-    instcmd=:
-    chmodcmd=
-  else
-    instcmd=$mkdirprog
-  fi
-else
-  # Waiting for this to be detected by the "$instcmd $src $dsttmp" command
-  # might cause directories to be created, which would be especially bad
-  # if $src (and thus $dsttmp) contains '*'.
-  if test ! -f "$src" && test ! -d "$src"; then
-    echo "$0: $src does not exist." >&2
-    exit 1
-  fi
-
-  if test -z "$dst"; then
-    echo "$0: no destination specified." >&2
-    exit 1
-  fi
-
-  # If destination is a directory, append the input filename; won't work
-  # if double slashes aren't ignored.
-  if test -d "$dst"; then
-    dst=$dst/`basename "$src"`
-  fi
-fi
-
-## this sed command emulates the dirname command
-dstdir=`echo "$dst" | sed -e 's,[^/]*$,,;s,/$,,;s,^$,.,'`
-
-# Make sure that the destination directory exists.
-# (this part is taken from Noah Friedman's mkinstalldirs script.)
-
-# Skip lots of stat calls in the usual case.
-if test ! -d "$dstdir"; then
-  defaultIFS='
-	'
-  IFS="${IFS-$defaultIFS}"
-
-  oIFS=$IFS
-  # Some sh's can't handle IFS=/ for some reason.
-  IFS='%'
-  set - `echo "$dstdir" | sed -e 's@/@%@g' -e 's@^%@/@'`
-  IFS=$oIFS
-
-  pathcomp=
-
-  while test $# -ne 0 ; do
-    pathcomp=$pathcomp$1
-    shift
-    test -d "$pathcomp" || $mkdirprog "$pathcomp"
-    pathcomp=$pathcomp/
+if test $# -ne 0 && test -z "$dir_arg$dstarg"; then
+  # When -d is used, all remaining arguments are directories to create.
+  # When -t is used, the destination is already specified.
+  # Otherwise, the last argument is the destination.  Remove it from $@.
+  for arg
+  do
+    if test -n "$dstarg"; then
+      # $@ is not empty: it contains at least $arg.
+      set fnord "$@" "$dstarg"
+      shift # fnord
+    fi
+    shift # arg
+    dstarg=$arg
   done
 fi
 
-if test -n "$dir_arg"; then
-  $doit $instcmd "$dst" \
-    && { test -z "$chowncmd" || $doit $chowncmd "$dst"; } \
-    && { test -z "$chgrpcmd" || $doit $chgrpcmd "$dst"; } \
-    && { test -z "$stripcmd" || $doit $stripcmd "$dst"; } \
-    && { test -z "$chmodcmd" || $doit $chmodcmd "$dst"; }
-
-else
-  # If we're going to rename the final executable, determine the name now.
-  if test -z "$transformarg"; then
-    dstfile=`basename "$dst"`
-  else
-    dstfile=`basename "$dst" $transformbasename \
-             | sed $transformarg`$transformbasename
+if test $# -eq 0; then
+  if test -z "$dir_arg"; then
+    echo "$0: no input file specified." >&2
+    exit 1
   fi
+  # It's OK to call `install-sh -d' without argument.
+  # This can happen when creating conditional directories.
+  exit 0
+fi
 
-  # don't allow the sed command to completely eliminate the filename.
-  test -z "$dstfile" && dstfile=`basename "$dst"`
-
-  # Make a couple of temp file names in the proper directory.
-  dsttmp=$dstdir/_inst.$$_
-  rmtmp=$dstdir/_rm.$$_
-
-  # Trap to clean up those temp files at exit.
-  trap 'status=$?; rm -f "$dsttmp" "$rmtmp" && exit $status' 0
+if test -z "$dir_arg"; then
   trap '(exit $?); exit' 1 2 13 15
 
-  # Move or copy the file name to the temp name
-  $doit $instcmd "$src" "$dsttmp" &&
+  # Set umask so as not to create temps with too-generous modes.
+  # However, 'strip' requires both read and write access to temps.
+  case $mode in
+    # Optimize common cases.
+    *644) cp_umask=133;;
+    *755) cp_umask=22;;
 
-  # and set any options; do chmod last to preserve setuid bits.
-  #
-  # If any of these fail, we abort the whole thing.  If we want to
-  # ignore errors from any of these, just make sure not to ignore
-  # errors from the above "$doit $instcmd $src $dsttmp" command.
-  #
-  { test -z "$chowncmd" || $doit $chowncmd "$dsttmp"; } \
-    && { test -z "$chgrpcmd" || $doit $chgrpcmd "$dsttmp"; } \
-    && { test -z "$stripcmd" || $doit $stripcmd "$dsttmp"; } \
-    && { test -z "$chmodcmd" || $doit $chmodcmd "$dsttmp"; } &&
+    *[0-7])
+      if test -z "$stripcmd"; then
+	u_plus_rw=
+      else
+	u_plus_rw='% 200'
+      fi
+      cp_umask=`expr '(' 777 - $mode % 1000 ')' $u_plus_rw`;;
+    *)
+      if test -z "$stripcmd"; then
+	u_plus_rw=
+      else
+	u_plus_rw=,u+rw
+      fi
+      cp_umask=$mode$u_plus_rw;;
+  esac
+fi
 
-  # Now remove or move aside any old file at destination location.  We
-  # try this two ways since rm can't unlink itself on some systems and
-  # the destination file might be busy for other reasons.  In this case,
-  # the final cleanup might fail but the new file should still install
-  # successfully.
-  {
-    if test -f "$dstdir/$dstfile"; then
-      $doit $rmcmd -f "$dstdir/$dstfile" 2>/dev/null \
-      || $doit $mvcmd -f "$dstdir/$dstfile" "$rmtmp" 2>/dev/null \
-      || {
-	  echo "$0: cannot unlink or rename $dstdir/$dstfile" >&2
-	  (exit 1); exit
-      }
-    else
-      :
+for src
+do
+  # Protect names starting with `-'.
+  case $src in
+    -*) src=./$src ;;
+  esac
+
+  if test -n "$dir_arg"; then
+    dst=$src
+    dstdir=$dst
+    test -d "$dstdir"
+    dstdir_status=$?
+  else
+
+    # Waiting for this to be detected by the "$cpprog $src $dsttmp" command
+    # might cause directories to be created, which would be especially bad
+    # if $src (and thus $dsttmp) contains '*'.
+    if test ! -f "$src" && test ! -d "$src"; then
+      echo "$0: $src does not exist." >&2
+      exit 1
     fi
-  } &&
 
-  # Now rename the file to the real destination.
-  $doit $mvcmd "$dsttmp" "$dstdir/$dstfile"
-fi &&
+    if test -z "$dstarg"; then
+      echo "$0: no destination specified." >&2
+      exit 1
+    fi
 
-# The final little trick to "correctly" pass the exit status to the exit trap.
-{
-  (exit 0); exit
-}
+    dst=$dstarg
+    # Protect names starting with `-'.
+    case $dst in
+      -*) dst=./$dst ;;
+    esac
+
+    # If destination is a directory, append the input filename; won't work
+    # if double slashes aren't ignored.
+    if test -d "$dst"; then
+      if test -n "$no_target_directory"; then
+	echo "$0: $dstarg: Is a directory" >&2
+	exit 1
+      fi
+      dstdir=$dst
+      dst=$dstdir/`basename "$src"`
+      dstdir_status=0
+    else
+      # Prefer dirname, but fall back on a substitute if dirname fails.
+      dstdir=`
+	(dirname "$dst") 2>/dev/null ||
+	expr X"$dst" : 'X\(.*[^/]\)//*[^/][^/]*/*$' \| \
+	     X"$dst" : 'X\(//\)[^/]' \| \
+	     X"$dst" : 'X\(//\)$' \| \
+	     X"$dst" : 'X\(/\)' \| . 2>/dev/null ||
+	echo X"$dst" |
+	    sed '/^X\(.*[^/]\)\/\/*[^/][^/]*\/*$/{
+		   s//\1/
+		   q
+		 }
+		 /^X\(\/\/\)[^/].*/{
+		   s//\1/
+		   q
+		 }
+		 /^X\(\/\/\)$/{
+		   s//\1/
+		   q
+		 }
+		 /^X\(\/\).*/{
+		   s//\1/
+		   q
+		 }
+		 s/.*/./; q'
+      `
+
+      test -d "$dstdir"
+      dstdir_status=$?
+    fi
+  fi
+
+  obsolete_mkdir_used=false
+
+  if test $dstdir_status != 0; then
+    case $posix_mkdir in
+      '')
+	# Create intermediate dirs using mode 755 as modified by the umask.
+	# This is like FreeBSD 'install' as of 1997-10-28.
+	umask=`umask`
+	case $stripcmd.$umask in
+	  # Optimize common cases.
+	  *[2367][2367]) mkdir_umask=$umask;;
+	  .*0[02][02] | .[02][02] | .[02]) mkdir_umask=22;;
+
+	  *[0-7])
+	    mkdir_umask=`expr $umask + 22 \
+	      - $umask % 100 % 40 + $umask % 20 \
+	      - $umask % 10 % 4 + $umask % 2
+	    `;;
+	  *) mkdir_umask=$umask,go-w;;
+	esac
+
+	# With -d, create the new directory with the user-specified mode.
+	# Otherwise, rely on $mkdir_umask.
+	if test -n "$dir_arg"; then
+	  mkdir_mode=-m$mode
+	else
+	  mkdir_mode=
+	fi
+
+	posix_mkdir=false
+	case $umask in
+	  *[123567][0-7][0-7])
+	    # POSIX mkdir -p sets u+wx bits regardless of umask, which
+	    # is incompatible with FreeBSD 'install' when (umask & 300) != 0.
+	    ;;
+	  *)
+	    tmpdir=${TMPDIR-/tmp}/ins$RANDOM-$$
+	    trap 'ret=$?; rmdir "$tmpdir/d" "$tmpdir" 2>/dev/null; exit $ret' 0
+
+	    if (umask $mkdir_umask &&
+		exec $mkdirprog $mkdir_mode -p -- "$tmpdir/d") >/dev/null 2>&1
+	    then
+	      if test -z "$dir_arg" || {
+		   # Check for POSIX incompatibilities with -m.
+		   # HP-UX 11.23 and IRIX 6.5 mkdir -m -p sets group- or
+		   # other-writeable bit of parent directory when it shouldn't.
+		   # FreeBSD 6.1 mkdir -m -p sets mode of existing directory.
+		   ls_ld_tmpdir=`ls -ld "$tmpdir"`
+		   case $ls_ld_tmpdir in
+		     d????-?r-*) different_mode=700;;
+		     d????-?--*) different_mode=755;;
+		     *) false;;
+		   esac &&
+		   $mkdirprog -m$different_mode -p -- "$tmpdir" && {
+		     ls_ld_tmpdir_1=`ls -ld "$tmpdir"`
+		     test "$ls_ld_tmpdir" = "$ls_ld_tmpdir_1"
+		   }
+		 }
+	      then posix_mkdir=:
+	      fi
+	      rmdir "$tmpdir/d" "$tmpdir"
+	    else
+	      # Remove any dirs left behind by ancient mkdir implementations.
+	      rmdir ./$mkdir_mode ./-p ./-- 2>/dev/null
+	    fi
+	    trap '' 0;;
+	esac;;
+    esac
+
+    if
+      $posix_mkdir && (
+	umask $mkdir_umask &&
+	$doit_exec $mkdirprog $mkdir_mode -p -- "$dstdir"
+      )
+    then :
+    else
+
+      # The umask is ridiculous, or mkdir does not conform to POSIX,
+      # or it failed possibly due to a race condition.  Create the
+      # directory the slow way, step by step, checking for races as we go.
+
+      case $dstdir in
+	/*) prefix=/ ;;
+	-*) prefix=./ ;;
+	*)  prefix= ;;
+      esac
+
+      case $posix_glob in
+        '')
+	  if (set -f) 2>/dev/null; then
+	    posix_glob=true
+	  else
+	    posix_glob=false
+	  fi ;;
+      esac
+
+      oIFS=$IFS
+      IFS=/
+      $posix_glob && set -f
+      set fnord $dstdir
+      shift
+      $posix_glob && set +f
+      IFS=$oIFS
+
+      prefixes=
+
+      for d
+      do
+	test -z "$d" && continue
+
+	prefix=$prefix$d
+	if test -d "$prefix"; then
+	  prefixes=
+	else
+	  if $posix_mkdir; then
+	    (umask=$mkdir_umask &&
+	     $doit_exec $mkdirprog $mkdir_mode -p -- "$dstdir") && break
+	    # Don't fail if two instances are running concurrently.
+	    test -d "$prefix" || exit 1
+	  else
+	    case $prefix in
+	      *\'*) qprefix=`echo "$prefix" | sed "s/'/'\\\\\\\\''/g"`;;
+	      *) qprefix=$prefix;;
+	    esac
+	    prefixes="$prefixes '$qprefix'"
+	  fi
+	fi
+	prefix=$prefix/
+      done
+
+      if test -n "$prefixes"; then
+	# Don't fail if two instances are running concurrently.
+	(umask $mkdir_umask &&
+	 eval "\$doit_exec \$mkdirprog $prefixes") ||
+	  test -d "$dstdir" || exit 1
+	obsolete_mkdir_used=true
+      fi
+    fi
+  fi
+
+  if test -n "$dir_arg"; then
+    { test -z "$chowncmd" || $doit $chowncmd "$dst"; } &&
+    { test -z "$chgrpcmd" || $doit $chgrpcmd "$dst"; } &&
+    { test "$obsolete_mkdir_used$chowncmd$chgrpcmd" = false ||
+      test -z "$chmodcmd" || $doit $chmodcmd $mode "$dst"; } || exit 1
+  else
+
+    # Make a couple of temp file names in the proper directory.
+    dsttmp=$dstdir/_inst.$$_
+    rmtmp=$dstdir/_rm.$$_
+
+    # Trap to clean up those temp files at exit.
+    trap 'ret=$?; rm -f "$dsttmp" "$rmtmp" && exit $ret' 0
+
+    # Copy the file name to the temp name.
+    (umask $cp_umask && $doit_exec $cpprog "$src" "$dsttmp") &&
+
+    # and set any options; do chmod last to preserve setuid bits.
+    #
+    # If any of these fail, we abort the whole thing.  If we want to
+    # ignore errors from any of these, just make sure not to ignore
+    # errors from the above "$doit $cpprog $src $dsttmp" command.
+    #
+    { test -z "$chowncmd" || $doit $chowncmd "$dsttmp"; } \
+      && { test -z "$chgrpcmd" || $doit $chgrpcmd "$dsttmp"; } \
+      && { test -z "$stripcmd" || $doit $stripcmd "$dsttmp"; } \
+      && { test -z "$chmodcmd" || $doit $chmodcmd $mode "$dsttmp"; } &&
+
+    # Now rename the file to the real destination.
+    { $doit $mvcmd -f "$dsttmp" "$dst" 2>/dev/null \
+      || {
+	   # The rename failed, perhaps because mv can't rename something else
+	   # to itself, or perhaps because mv is so ancient that it does not
+	   # support -f.
+
+	   # Now remove or move aside any old file at destination location.
+	   # We try this two ways since rm can't unlink itself on some
+	   # systems and the destination file might be busy for other
+	   # reasons.  In this case, the final cleanup might fail but the new
+	   # file should still install successfully.
+	   {
+	     if test -f "$dst"; then
+	       $doit $rmcmd -f "$dst" 2>/dev/null \
+	       || { $doit $mvcmd -f "$dst" "$rmtmp" 2>/dev/null \
+		     && { $doit $rmcmd -f "$rmtmp" 2>/dev/null; :; }; }\
+	       || {
+		 echo "$0: cannot unlink or rename $dst" >&2
+		 (exit 1); exit 1
+	       }
+	     else
+	       :
+	     fi
+	   } &&
+
+	   # Now rename the file to the real destination.
+	   $doit $mvcmd "$dsttmp" "$dst"
+	 }
+    } || exit 1
+
+    trap '' 0
+  fi
+done
 
 # Local variables:
 # eval: (add-hook 'write-file-hooks 'time-stamp)

configure.in

@@ -1,8 +1,6 @@
-##
+###############################################################################
 ## Copyright (C) 2000-2004 Sistina Software, Inc. All rights reserved.
-## Copyright (C) 2004 Red Hat, Inc. All rights reserved.
-##
-## This file is part of the LVM2.
+## Copyright (C) 2004-2009 Red Hat, Inc. All rights reserved.
 ##
 ## This copyrighted material is made available to anyone wishing to use,
 ## modify, copy, or redistribute it subject to the terms and conditions
@@ -13,18 +11,20 @@
 ## Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 ################################################################################
 
-AC_PREREQ(2.53)
+AC_PREREQ(2.57)
 ################################################################################
 dnl -- Process this file with autoconf to produce a configure script.
-AC_INIT(lib/device/dev-cache.h)
+AC_INIT
+AC_CONFIG_SRCDIR([lib/device/dev-cache.h])
+AC_CONFIG_HEADERS([lib/misc/configure.h])
 
 ################################################################################
 dnl -- Setup the directory where autoconf has auxilary files
-AC_CONFIG_AUX_DIR(autoconf) 
+AC_CONFIG_AUX_DIR(autoconf)
 
 ################################################################################
 dnl -- Get system type
-AC_CANONICAL_SYSTEM
+AC_CANONICAL_TARGET([])
 
 case "$host_os" in
 	linux*)
@@ -35,46 +35,58 @@ case "$host_os" in
 		CLDNOWHOLEARCHIVE="-Wl,-no-whole-archive"
 		LDDEPS="$LDDEPS .export.sym"
 		LDFLAGS="$LDFLAGS -Wl,--export-dynamic"
-		LIB_SUFFIX="so"
+		LIB_SUFFIX=so
 		DEVMAPPER=yes
 		ODIRECT=yes
+		DM_IOCTLS=yes
 		SELINUX=yes
+		REALTIME=yes
 		CLUSTER=internal
-		FSADM=no ;;
+		FSADM=yes
+		;;
 	darwin*)
 		CFLAGS="$CFLAGS -no-cpp-precomp -fno-common"
 		COPTIMISE_FLAG="-O2"
 		CLDFLAGS="$CLDFLAGS"
 		CLDWHOLEARCHIVE="-all_load"
 		CLDNOWHOLEARCHIVE=
-		LDDEPS="$LDDEPS"
-		LDFLAGS="$LDFLAGS"
-		LIB_SUFFIX="dylib"
+		LIB_SUFFIX=dylib
 		DEVMAPPER=yes
 		ODIRECT=no
+		DM_IOCTLS=no
 		SELINUX=no
+		REALTIME=no
 		CLUSTER=none
-		FSADM=no ;;
+		FSADM=no
+		;;
 esac
 
 ################################################################################
 dnl -- Checks for programs.
+AC_PROG_SED
 AC_PROG_AWK
 AC_PROG_CC
+
+dnl probably no longer needed in 2008, but...
+AC_PROG_GCC_TRADITIONAL
 AC_PROG_INSTALL
 AC_PROG_LN_S
 AC_PROG_MAKE_SET
+AC_PROG_MKDIR_P
 AC_PROG_RANLIB
+AC_PATH_PROG(CFLOW_CMD, cflow)
+AC_PATH_PROG(CSCOPE_CMD, cscope)
 
 ################################################################################
-dnl -- Checks for header files.
+dnl -- Check for header files.
 AC_HEADER_DIRENT
 AC_HEADER_STDC
 AC_HEADER_SYS_WAIT
 AC_HEADER_TIME
 
-AC_CHECK_HEADERS(fcntl.h limits.h locale.h stddef.h syslog.h sys/file.h sys/ioctl.h sys/param.h sys/time.h,,AC_MSG_ERROR(bailing out))
-AC_CHECK_HEADERS(assert.h ctype.h libgen.h signal.h stdio.h sys/mman.h sys/resource.h sys/stat.h sys/types.h sys/utsname.h sys/wait.h time.h,,AC_MSG_ERROR(bailing out))
+AC_CHECK_HEADERS([locale.h stddef.h syslog.h sys/file.h sys/time.h assert.h \
+  libgen.h signal.h sys/mman.h sys/resource.h sys/utsname.h sys/wait.h time.h], ,
+  [AC_MSG_ERROR(bailing out)])
 
 case "$host_os" in
 	linux*)
@@ -83,20 +95,29 @@ case "$host_os" in
 		AC_CHECK_HEADERS(machine/endian.h sys/disk.h,,AC_MSG_ERROR(bailing out)) ;;
 esac
 
+AC_CHECK_HEADERS([ctype.h dirent.h errno.h fcntl.h getopt.h inttypes.h limits.h \
+  stdarg.h stdio.h stdlib.h string.h sys/ioctl.h sys/param.h sys/stat.h \
+  sys/types.h unistd.h], , [AC_MSG_ERROR(bailing out)])
+AC_CHECK_HEADERS(termios.h sys/statvfs.h)
+
 ################################################################################
-dnl -- Checks for typedefs, structures, and compiler characteristics.
+dnl -- Check for typedefs, structures, and compiler characteristics.
 AC_C_CONST
 AC_C_INLINE
+AC_CHECK_MEMBERS([struct stat.st_rdev])
 AC_TYPE_OFF_T
 AC_TYPE_PID_T
+AC_TYPE_SIGNAL
 AC_TYPE_SIZE_T
 AC_TYPE_MODE_T
-AC_STRUCT_ST_RDEV
+AC_CHECK_MEMBERS([struct stat.st_rdev])
 AC_STRUCT_TM
 
 ################################################################################
 dnl -- Check for functions
-AC_CHECK_FUNCS(gethostname getpagesize memset munmap setlocale strcasecmp strchr strdup strncasecmp strerror strrchr strstr strtol strtoul,,AC_MSG_ERROR(bailing out))
+AC_CHECK_FUNCS([gethostname getpagesize memset mkdir rmdir munmap setlocale \
+  strcasecmp strchr strdup strncasecmp strerror strrchr strstr strtol strtoul \
+  uname], , [AC_MSG_ERROR(bailing out)])
 AC_FUNC_ALLOCA
 AC_FUNC_CLOSEDIR_VOID
 AC_FUNC_FORK
@@ -106,22 +127,28 @@ AC_FUNC_MEMCMP
 AC_FUNC_MMAP
 AC_FUNC_STAT
 AC_FUNC_STRTOD
+AC_FUNC_VPRINTF
+
+################################################################################
+dnl -- Enables statically-linked tools
+AC_MSG_CHECKING(whether to use static linking)
+AC_ARG_ENABLE(static_link,
+  [  --enable-static_link    Use this to link the tools to their libraries
+                          statically.  Default is dynamic linking],
+  STATIC_LINK=$enableval, STATIC_LINK=no)
+AC_MSG_RESULT($STATIC_LINK)
 
 ################################################################################
 dnl -- Prefix is /usr by default, the exec_prefix default is setup later
 AC_PREFIX_DEFAULT(/usr)
 
-################################################################################
-dnl -- Parallel make jobs?
-AC_ARG_ENABLE(jobs, [  --enable-jobs=NUM       Number of jobs to run simultaneously], JOBS=-j$enableval, JOBS=-j2)
-
 ################################################################################
 dnl -- Setup the ownership of the files
 AC_MSG_CHECKING(file owner)
 OWNER="root"
 
 AC_ARG_WITH(user,
-  [  --with-user=USER        Set the owner of installed files ],
+  [  --with-user=USER        Set the owner of installed files [[USER=root]] ],
   [ OWNER="$withval" ])
 AC_MSG_RESULT($OWNER)
 
@@ -134,7 +161,7 @@ dnl -- Setup the group ownership of the files
 AC_MSG_CHECKING(group owner)
 GROUP="root"
 AC_ARG_WITH(group,
-  [  --with-group=GROUP      Set the group owner of installed files ],
+  [  --with-group=GROUP      Set the group owner of installed files [[GROUP=root]] ],
   [ GROUP="$withval" ])
 AC_MSG_RESULT($GROUP)
 
@@ -142,6 +169,33 @@ if test x$GROUP != x; then
 	GROUP="-g $GROUP"
 fi
 
+################################################################################
+dnl -- Setup device node ownership
+AC_MSG_CHECKING(device node uid)
+
+AC_ARG_WITH(device-uid,
+  [  --with-device-uid=UID   Set the owner used for new device nodes [[UID=0]] ],
+  [ DM_DEVICE_UID="$withval" ], [ DM_DEVICE_UID="0" ] )
+AC_MSG_RESULT($DM_DEVICE_UID)
+
+################################################################################
+dnl -- Setup device group ownership
+AC_MSG_CHECKING(device node gid)
+
+AC_ARG_WITH(device-gid,
+  [  --with-device-gid=UID   Set the group used for new device nodes [[GID=0]] ],
+  [ DM_DEVICE_GID="$withval" ], [ DM_DEVICE_GID="0" ] )
+AC_MSG_RESULT($DM_DEVICE_GID)
+
+################################################################################
+dnl -- Setup device mode
+AC_MSG_CHECKING(device node mode)
+
+AC_ARG_WITH(device-mode,
+  [  --with-device-mode=MODE Set the mode used for new device nodes [[MODE=0600]] ],
+  [ DM_DEVICE_MODE="$withval" ], [ DM_DEVICE_MODE="0600" ] )
+AC_MSG_RESULT($DM_DEVICE_MODE)
+
 ################################################################################
 dnl -- LVM1 tool fallback option
 AC_MSG_CHECKING(whether to enable lvm1 fallback)
@@ -150,7 +204,7 @@ AC_ARG_ENABLE(lvm1_fallback, [  --enable-lvm1_fallback  Use this to fall back an
 AC_MSG_RESULT($LVM1_FALLBACK)
 
 if test x$LVM1_FALLBACK = xyes; then
-	CFLAGS="$CFLAGS -DLVM1_FALLBACK"
+	AC_DEFINE([LVM1_FALLBACK], 1, [Define to 1 if 'lvm' should fall back to using LVM1 binaries if device-mapper is missing from the kernel])
 fi
 
 ################################################################################
@@ -170,7 +224,7 @@ if [[ "x$LVM1" != xnone -a "x$LVM1" != xinternal -a "x$LVM1" != xshared ]];
 fi;
 
 if test x$LVM1 = xinternal; then
-	CFLAGS="$CFLAGS -DLVM1_INTERNAL"
+	AC_DEFINE([LVM1_INTERNAL], 1, [Define to 1 to include built-in support for LVM1 metadata.])
 fi
 
 ################################################################################
@@ -190,7 +244,7 @@ if [[ "x$POOL" != xnone -a "x$POOL" != xinternal -a "x$POOL" != xshared ]];
 fi;
 
 if test x$POOL = xinternal; then
-	CFLAGS="$CFLAGS -DPOOL_INTERNAL"
+	AC_DEFINE([POOL_INTERNAL], 1, [Define to 1 to include built-in support for GFS pool metadata.])
 fi
 
 ################################################################################
@@ -209,7 +263,7 @@ if [[ "x$CLUSTER" != xnone -a "x$CLUSTER" != xinternal -a "x$CLUSTER" != xshared
 fi;
 
 if test x$CLUSTER = xinternal; then
-	CFLAGS="$CFLAGS -DCLUSTER_LOCKING_INTERNAL"
+	AC_DEFINE([CLUSTER_LOCKING_INTERNAL], 1, [Define to 1 to include built-in support for clustered LVM locking.])
 fi
 
 ################################################################################
@@ -229,7 +283,7 @@ if [[ "x$SNAPSHOTS" != xnone -a "x$SNAPSHOTS" != xinternal -a "x$SNAPSHOTS" != x
 fi;
 
 if test x$SNAPSHOTS = xinternal; then
-	CFLAGS="$CFLAGS -DSNAPSHOT_INTERNAL"
+	AC_DEFINE([SNAPSHOT_INTERNAL], 1, [Define to 1 to include built-in support for snapshots.])
 fi
 
 ################################################################################
@@ -249,39 +303,29 @@ if [[ "x$MIRRORS" != xnone -a "x$MIRRORS" != xinternal -a "x$MIRRORS" != xshared
 fi;
 
 if test x$MIRRORS = xinternal; then
-	CFLAGS="$CFLAGS -DMIRRORED_INTERNAL"
+	AC_DEFINE([MIRRORED_INTERNAL], 1, [Define to 1 to include built-in support for mirrors.])
 fi
 
 ################################################################################
-dnl -- Enables staticly-linked tools
-AC_MSG_CHECKING(whether to use static linking)
-AC_ARG_ENABLE(static_link, [  --enable-static_link    Use this to link the tools to their libraries
-                          statically.  Default is dynamic linking],  STATIC_LINK=$enableval, STATIC_LINK=no)
-AC_MSG_RESULT($STATIC_LINK)
-
-################################################################################
-dnl -- Enable readline
+dnl -- Disable readline
 AC_MSG_CHECKING(whether to enable readline)
-AC_ARG_ENABLE(readline, [  --enable-readline       Enable readline support],
-READLINE=$enableval, READLINE=no)
+AC_ARG_ENABLE([readline],
+  [  --disable-readline      Disable readline support],
+  [READLINE=$enableval], [READLINE=maybe])
 AC_MSG_RESULT($READLINE)
 
-if test x$READLINE = xyes; then
-	CFLAGS="$CFLAGS -DREADLINE_SUPPORT"
-fi
-
 ################################################################################
-dnl -- Disable selinux
-AC_MSG_CHECKING(whether to enable selinux support)
-AC_ARG_ENABLE(selinux, [  --disable-selinux       Disable selinux support],
-SELINUX=$enableval)
-AC_MSG_RESULT($SELINUX)
+dnl -- Disable realtime clock support
+AC_MSG_CHECKING(whether to enable realtime support)
+AC_ARG_ENABLE(realtime, [  --disable-realtime      Disable realtime clock support],
+REALTIME=$enableval)
+AC_MSG_RESULT($REALTIME)
 
 ################################################################################
 dnl -- Build cluster LVM daemon
 AC_MSG_CHECKING(whether to build cluster LVM daemon)
 AC_ARG_WITH(clvmd,
-  [  --with-clvmd=TYPE       Build cluster LVM Daemon: cman/gulm/none/all
+  [  --with-clvmd=TYPE       Build cluster LVM Daemon: cman/gulm/corosync/none/all
                           [TYPE=none] ],
   [ CLVMD="$withval" ],
   [ CLVMD="none" ])
@@ -295,37 +339,79 @@ if  test x$CLVMD != xnone && test x$CLUSTER = xnone; then
 	CLUSTER=internal
 fi
 
+dnl -- Look for corosync libraries if required.
+if [[ "x$CLVMD" = xcorosync -o "x$CLVMD" = xall ]]; then
+	PKG_CHECK_MODULES(QUORUM, libquorum, [],
+		[AC_MSG_RESULT([no pkg for quorum library, using -lquorum]);
+		QUORUM_LIBS="-lquorum"])
+	PKG_CHECK_MODULES(CONFDB, libconfdb, [],
+		[AC_MSG_RESULT([no pkg for confdb library, using -lconfdb]);
+		CONFDB_LIBS="-lconfdb"])
+	PKG_CHECK_MODULES(CPG, libcpg, [],
+		[AC_MSG_RESULT([no pkg for libcpg library, using -lcpg]);
+		CPG_LIBS="-lcpg"])
+fi
+
 ################################################################################
 dnl -- Enable debugging
 AC_MSG_CHECKING(whether to enable debugging)
 AC_ARG_ENABLE(debug,    [  --enable-debug          Enable debugging],
-DEBUG=$enableval, DEBUG=no)
+  DEBUG=$enableval, DEBUG=no)
 AC_MSG_RESULT($DEBUG)
 
 dnl -- Normally turn off optimisation for debug builds
 if test x$DEBUG = xyes; then
 	COPTIMISE_FLAG=
+else
+	CSCOPE_CMD=
 fi
 
 ################################################################################
 dnl -- Override optimisation
 AC_MSG_CHECKING(for C optimisation flag)
 AC_ARG_WITH(optimisation,
-  [  --with-optimisation=OPT C optimisation flag [OPT=-O2] ],
+  [  --with-optimisation=OPT C optimisation flag [[OPT=-O2]] ],
   [ COPTIMISE_FLAG="$withval" ])
 AC_MSG_RESULT($COPTIMISE_FLAG)
 
+################################################################################
+dnl -- Enable profiling
+AC_MSG_CHECKING(whether to gather gcov profiling data)
+AC_ARG_ENABLE(profiling,
+  AC_HELP_STRING(--enable-profiling, [Gather gcov profiling data]),
+  PROFILING=$enableval, PROFILING=no)
+AC_MSG_RESULT($PROFILING)
+
+if test "x$PROFILING" = xyes; then
+  COPTIMISE_FLAG="$COPTIMISE_FLAG -fprofile-arcs -ftest-coverage"
+  AC_PATH_PROG(LCOV, lcov, no)
+  AC_PATH_PROG(GENHTML, genhtml, no)
+  if test "$LCOV" = no -o "$GENHTML" = no ; then
+    AC_MSG_ERROR([lcov and genhtml are required for profiling])
+  fi
+fi
+
 ################################################################################
 dnl -- Disable devmapper
 AC_MSG_CHECKING(whether to use device-mapper)
-AC_ARG_ENABLE(devmapper, [  --disable-devmapper     Disable device-mapper interaction],
+AC_ARG_ENABLE(devmapper, [  --disable-devmapper     Disable LVM2 device-mapper interaction],
 DEVMAPPER=$enableval)
 AC_MSG_RESULT($DEVMAPPER)
 
 if test x$DEVMAPPER = xyes; then
-	CFLAGS="$CFLAGS -DDEVMAPPER_SUPPORT"
+	AC_DEFINE([DEVMAPPER_SUPPORT], 1, [Define to 1 to enable LVM2 device-mapper interaction.])
 fi
 
+################################################################################
+dnl -- Compatibility mode
+AC_ARG_ENABLE(compat,   [  --enable-compat         Enable support for old device-mapper versions],
+  DM_COMPAT=$enableval, DM_COMPAT=no)
+
+################################################################################
+dnl -- Disable ioctl
+AC_ARG_ENABLE(ioctl,   [  --disable-driver        Disable calls to device-mapper in the kernel],
+  DM_IOCTLS=$enableval)
+
 ################################################################################
 dnl -- Disable O_DIRECT
 AC_MSG_CHECKING(whether to enable O_DIRECT)
@@ -334,23 +420,40 @@ ODIRECT=$enableval)
 AC_MSG_RESULT($ODIRECT)
 
 if test x$ODIRECT = xyes; then
-	CFLAGS="$CFLAGS -DO_DIRECT_SUPPORT"
+	AC_DEFINE([O_DIRECT_SUPPORT], 1, [Define to 1 to enable O_DIRECT support.])
 fi
 
+################################################################################
+dnl -- Enable liblvm2app.so
+AC_MSG_CHECKING(whether to build liblvm2app.so application library)
+AC_ARG_ENABLE(applib,
+  [  --enable-applib         Build application library],
+  APPLIB=$enableval, APPLIB=no)
+AC_MSG_RESULT($APPLIB)
+AC_SUBST([LVM2APP_LIB])
+test x$APPLIB = xyes \
+  && LVM2APP_LIB=-llvm2app \
+  || LVM2APP_LIB=
+
 ################################################################################
 dnl -- Enable cmdlib
 AC_MSG_CHECKING(whether to compile liblvm2cmd.so)
 AC_ARG_ENABLE(cmdlib, [  --enable-cmdlib         Build shared command library],
 CMDLIB=$enableval, CMDLIB=no)
 AC_MSG_RESULT($CMDLIB)
+AC_SUBST([LVM2CMD_LIB])
+test x$CMDLIB = xyes \
+  && LVM2CMD_LIB=-llvm2cmd \
+  || LVM2CMD_LIB=
 
-if test x$CMDLIB = xyes; then
-	CFLAGS="$CFLAGS -DCMDLIB"
-fi
+################################################################################
+dnl -- Enable pkg-config
+AC_ARG_ENABLE(pkgconfig,   [  --enable-pkgconfig      Install pkgconfig support],
+  PKGCONFIG=$enableval, PKGCONFIG=no)
 
 ################################################################################
 dnl -- Enable fsadm
-AC_MSG_CHECKING(whether to build fsadm)
+AC_MSG_CHECKING(whether to install fsadm)
 AC_ARG_ENABLE(fsadm, [  --enable-fsadm          Enable fsadm],
 FSADM=$enableval)
 AC_MSG_RESULT($FSADM)
@@ -362,34 +465,52 @@ AC_ARG_ENABLE(dmeventd, [  --enable-dmeventd       Enable the device-mapper even
 DMEVENTD=$enableval)
 AC_MSG_RESULT($DMEVENTD)
 
+BUILD_DMEVENTD=$DMEVENTD
+
 dnl -- dmeventd currently requires internal mirror support
-if test x$DMEVENTD = xyes && test x$MIRRORS != xinternal; then
-AC_MSG_ERROR(
---enable-dmeventd currently requires --with-mirrors=internal
-)
+if test x$DMEVENTD = xyes; then
+   if test x$MIRRORS != xinternal; then
+      AC_MSG_ERROR(
+	--enable-dmeventd currently requires --with-mirrors=internal
+      )
+   fi
+   if test x$CMDLIB = xno; then
+      AC_MSG_ERROR(
+        --enable-dmeventd requires --enable-cmdlib to be used as well
+      )
+   fi
 fi
 
 if test x$DMEVENTD = xyes; then
-	CFLAGS="$CFLAGS -DDMEVENTD"
+	AC_DEFINE([DMEVENTD], 1, [Define to 1 to enable the device-mapper event daemon.])
 fi
+
 ################################################################################
-dnl -- Mess with default exec_prefix
+dnl -- getline included in recent libc
+
+AC_CHECK_LIB(c, getline, AC_DEFINE([HAVE_GETLINE], 1,
+  [Define to 1 if getline is available.]))
+
+################################################################################
+dnl -- canonicalize_file_name included in recent libc
+
+AC_CHECK_LIB(c, canonicalize_file_name,
+  AC_DEFINE([HAVE_CANONICALIZE_FILE_NAME], 1,
+    [Define to 1 if canonicalize_file_name is available.]))
+
+################################################################################
+dnl -- Clear default exec_prefix - install into /sbin rather than /usr/sbin
 if [[ "x$exec_prefix" = xNONE -a "x$prefix" = xNONE ]];
  then  exec_prefix="";
 fi;
 
-################################################################################
-dnl -- Checks for library functions.
-AC_PROG_GCC_TRADITIONAL
-AC_TYPE_SIGNAL
-AC_FUNC_VPRINTF
-AC_CHECK_FUNCS(mkdir rmdir uname,,AC_MSG_ERROR(bailing out))
-
 ################################################################################
 dnl -- Check for termcap (Shamelessly copied from parted 1.4.17)
-if test x$READLINE = xyes; then
-	AC_SEARCH_LIBS(tgetent, ncurses curses termcap termlib, ,
-		AC_MSG_ERROR(
+if test x$READLINE != xno; then
+	AC_SEARCH_LIBS([tgetent], [tinfo ncurses curses termcap termlib],
+	  [tg_found=yes], [tg_found=no])
+	test x$READLINE:$tg_found = xyes:no &&
+	  AC_MSG_ERROR(
 termcap could not be found which is required for the
 --enable-readline option (which is enabled by default).  Either disable readline
 support with --disable-readline or download and install termcap from:
@@ -399,7 +520,6 @@ Note: if you are using precompiled packages you will also need the development
 Note: (n)curses also seems to work as a substitute for termcap.  This was
   not found either - but you could try installing that as well.
 )
-	)
 fi
 
 ################################################################################
@@ -407,7 +527,7 @@ dnl -- Check for dlopen
 AC_CHECK_LIB(dl, dlopen, HAVE_LIBDL=yes, HAVE_LIBDL=no)
 
 if [[ "x$HAVE_LIBDL" = xyes ]]; then
-	CFLAGS="$CFLAGS -DHAVE_LIBDL"
+	AC_DEFINE([HAVE_LIBDL], 1, [Define to 1 if dynamic libraries are available.])
 	LIBS="-ldl $LIBS"
 else
 	HAVE_LIBDL=no
@@ -424,29 +544,69 @@ Features cannot be 'shared' when building statically
 fi
 
 ################################################################################
-dnl -- Check for is_selinux_enabled
+dnl -- Disable selinux
+AC_MSG_CHECKING(whether to enable selinux support)
+AC_ARG_ENABLE(selinux, [  --disable-selinux       Disable selinux support],
+  SELINUX=$enableval)
+AC_MSG_RESULT($SELINUX)
+
+################################################################################
+dnl -- Check for selinux
 if test x$SELINUX = xyes; then
-	AC_MSG_CHECKING(for is_selinux_enabled function)
+	AC_CHECK_LIB(sepol, sepol_check_context, HAVE_SEPOL=yes, HAVE_SEPOL=no)
+
+	if test x$HAVE_SEPOL = xyes; then
+		AC_DEFINE([HAVE_SEPOL], 1,
+		  [Define to 1 if sepol_check_context is available.])
+		LIBS="-lsepol $LIBS"
+	fi
+
 	AC_CHECK_LIB(selinux, is_selinux_enabled, HAVE_SELINUX=yes, HAVE_SELINUX=no)
-	AC_MSG_RESULT($HAVE_SELINUX)
 
 	if test x$HAVE_SELINUX = xyes; then
-		CFLAGS="$CFLAGS -DHAVE_SELINUX"
+		AC_DEFINE([HAVE_SELINUX], 1, [Define to 1 to include support for selinux.])
 		LIBS="-lselinux $LIBS"
 	else
 		AC_MSG_WARN(Disabling selinux)
 	fi
+
+	# With --enable-static_link and selinux enabled, linking
+	# fails on at least Debian unstable due to unsatisfied references
+	# to pthread_mutex_lock and _unlock.  See if we need -lpthread.
+	if test "$STATIC_LINK-$HAVE_SELINUX" = yes-yes; then
+		lvm_saved_libs=$LIBS
+		LIBS="$LIBS -static"
+		AC_SEARCH_LIBS([pthread_mutex_lock], [pthread],
+		  [test "$ac_cv_search_pthread_mutex_lock" = "none required" ||
+				LIB_PTHREAD=-lpthread])
+		LIBS=$lvm_saved_libs
+	fi
+fi
+
+################################################################################
+dnl -- Check for realtime clock support
+if test x$REALTIME = xyes; then
+	AC_CHECK_LIB(rt, clock_gettime, HAVE_REALTIME=yes, HAVE_REALTIME=no)
+
+	if test x$HAVE_REALTIME = xyes; then
+		AC_DEFINE([HAVE_REALTIME], 1, [Define to 1 to include support for realtime clock.])
+		LIBS="-lrt $LIBS"
+	else
+		AC_MSG_WARN(Disabling realtime clock)
+	fi
 fi
 
 ################################################################################
 dnl -- Check for getopt
-AC_CHECK_HEADERS(getopt.h, CFLAGS="$CFLAGS -DHAVE_GETOPTLONG")
+AC_CHECK_HEADERS(getopt.h, AC_DEFINE([HAVE_GETOPTLONG], 1, [Define to 1 if getopt_long is available.]))
 
 ################################################################################
 dnl -- Check for readline (Shamelessly copied from parted 1.4.17)
-if test x$READLINE = xyes; then
-	AC_CHECK_LIB(readline, readline, ,
-		AC_MSG_ERROR(
+if test x$READLINE != xno; then
+	rl_found=yes
+	AC_CHECK_LIB([readline], [readline], , [rl_found=no])
+	test x$READLINE:$rl_found = xyes:no &&
+	  AC_MSG_ERROR(
 GNU Readline could not be found which is required for the
 --enable-readline option (which is enabled by default).  Either disable readline
 support with --disable-readline or download and install readline from:
@@ -454,9 +614,11 @@ support with --disable-readline or download and install readline from:
 Note: if you are using precompiled packages you will also need the development
 package as well (which may be called readline-devel or something similar).
 )
-	)
-	AC_CHECK_FUNC(rl_completion_matches, CFLAGS="$CFLAGS -DHAVE_RL_COMPLETION_MATCHES")
-		
+	if test $rl_found = yes; then
+		AC_CHECK_FUNCS([rl_completion_matches])
+		AC_DEFINE([READLINE_SUPPORT], 1,
+		  [Define to 1 to include the LVM readline shell.])
+	fi
 fi
 
 ################################################################################
@@ -467,6 +629,7 @@ AC_ARG_ENABLE(nls, [  --enable-nls            Enable Native Language Support],
 AC_MSG_RESULT($INTL)
 
 if test x$INTL = xyes; then
+# FIXME - Move this - can be device-mapper too
 	INTL_PACKAGE="lvm2"
 	AC_PATH_PROG(MSGFMT, msgfmt)
 	if [[ "x$MSGFMT" == x ]];
@@ -476,21 +639,31 @@ if test x$INTL = xyes; then
 	fi;
 
 	AC_ARG_WITH(localedir,
-  		    [  --with-localedir=DIR    Translation files in DIR [PREFIX/share/locale]],
-  		    [ LOCALEDIR="$withval" ],
-  		    [ LOCALEDIR='${prefix}/share/locale' ])
+		    [  --with-localedir=DIR    Translation files in DIR [[PREFIX/share/locale]] ],
+		    [ LOCALEDIR="$withval" ],
+		    [ LOCALEDIR='${prefix}/share/locale' ])
 fi
 
 ################################################################################
 AC_ARG_WITH(confdir,
 	    [  --with-confdir=DIR      Configuration files in DIR [/etc]],
-  	    [ CONFDIR="$withval" ],
- 	    [ CONFDIR='/etc' ])
+	    [ CONFDIR="$withval" ],
+	    [ CONFDIR='/etc' ])
 
 AC_ARG_WITH(staticdir,
 	    [  --with-staticdir=DIR    Static binary in DIR [EXEC_PREFIX/sbin]],
-  	    [ STATICDIR="$withval" ],
- 	    [ STATICDIR='${exec_prefix}/sbin' ])
+	    [ STATICDIR="$withval" ],
+	    [ STATICDIR='${exec_prefix}/sbin' ])
+
+AC_ARG_WITH(usrlibdir,
+	    [  --with-usrlibdir=DIR],
+	    [ usrlibdir="$withval"],
+	    [ usrlibdir='${prefix}/lib' ])
+
+AC_ARG_WITH(usrsbindir,
+	    [  --with-usrsbindir=DIR],
+	    [ usrsbindir="$withval"],
+	    [ usrsbindir='${prefix}/sbin' ])
 
 ################################################################################
 dnl -- Ensure additional headers required
@@ -506,11 +679,6 @@ if test x$CLVMD != xnone; then
 	AC_FUNC_SELECT_ARGTYPES
 fi
 
-if test x$FSADM = xyes; then
-	AC_CHECK_HEADERS(fstab.h sys/mount.h sys/vfs.h,,AC_MSG_ERROR(bailing out))
-	AC_CHECK_FUNCS(memmove,,AC_MSG_ERROR(bailing out))
-fi
-
 if test x$CLUSTER != xnone; then
 	AC_CHECK_HEADERS(sys/socket.h sys/un.h,,AC_MSG_ERROR(bailing out))
 	AC_CHECK_FUNCS(socket,,AC_MSG_ERROR(bailing out))
@@ -524,10 +692,6 @@ if test x$INTL = xyes; then
 	AC_CHECK_HEADERS(libintl.h,,AC_MSG_ERROR(bailing out))
 fi
 
-if test x$DEVMAPPER = xyes; then
-	AC_CHECK_HEADERS(libdevmapper.h,,AC_MSG_ERROR(bailing out))
-fi
-
 if test x$HAVE_SELINUX = xyes; then
 	AC_CHECK_HEADERS(selinux/selinux.h,,AC_MSG_ERROR(bailing out))
 fi
@@ -536,90 +700,166 @@ fi
 AC_PATH_PROG(MODPROBE_CMD, modprobe)
 
 if test x$MODPROBE_CMD != x; then
-	CFLAGS="$CFLAGS -DMODPROBE_CMD=\\\"$MODPROBE_CMD\\\""
+	AC_DEFINE_UNQUOTED([MODPROBE_CMD], ["$MODPROBE_CMD"], [The path to 'modprobe', if available.])
 fi
 
 ################################################################################
-if test "-f VERSION"; then
-  LVM_VERSION="\"`cat VERSION`\""
-else
-  LVM_VERSION="Unknown"
+dnl -- dmeventd pidfile and executable path
+AH_TEMPLATE(DMEVENTD_PIDFILE, [Path to dmeventd pidfile.])
+if test "$BUILD_DMEVENTD" = yes; then
+	AC_ARG_WITH(dmeventd-pidfile,
+		    [  --with-dmeventd-pidfile=PATH    dmeventd pidfile [[/var/run/dmeventd.pid]] ],
+		    [ AC_DEFINE_UNQUOTED(DMEVENTD_PIDFILE,"$withval") ],
+		    [ AC_DEFINE_UNQUOTED(DMEVENTD_PIDFILE,"/var/run/dmeventd.pid") ])
+fi
+
+AH_TEMPLATE(DMEVENTD_PATH, [Path to dmeventd binary.])
+if test "$BUILD_DMEVENTD" = yes; then
+        dmeventd_prefix="$exec_prefix"
+        if test "x$dmeventd_prefix" = "xNONE"; then
+                dmeventd_prefix="$prefix"
+        fi
+        if test "x$dmeventd_prefix" = "xNONE"; then
+                dmeventd_prefix=""
+        fi
+	AC_ARG_WITH(dmeventd-path,
+		    [  --with-dmeventd-path=PATH       dmeventd path [[${exec_prefix}/sbin/dmeventd]] ],
+		    [ AC_DEFINE_UNQUOTED(DMEVENTD_PATH,"$withval") ],
+		    [ AC_DEFINE_UNQUOTED(DMEVENTD_PATH,"$dmeventd_prefix/sbin/dmeventd") ])
 fi
 
 ################################################################################
-AC_SUBST(JOBS)
-AC_SUBST(STATIC_LINK)
-AC_SUBST(LVM1)
-AC_SUBST(POOL)
-AC_SUBST(SNAPSHOTS)
-AC_SUBST(MIRRORS)
-AC_SUBST(OWNER)
-AC_SUBST(GROUP)
+dnl -- which kernel interface to use (ioctl only)
+AC_MSG_CHECKING(for kernel interface choice)
+AC_ARG_WITH(interface,
+  [  --with-interface=IFACE  Choose kernel interface (ioctl) [[ioctl]] ],
+  [ interface="$withval" ],
+  [ interface=ioctl ])
+if [[ "x$interface" != xioctl ]];
+then
+  AC_MSG_ERROR(--with-interface=ioctl required. fs no longer supported.)
+fi
+AC_MSG_RESULT($interface)
+
+################################################################################
+DM_LIB_VERSION="\"`cat VERSION_DM 2>/dev/null || echo Unknown`\""
+AC_DEFINE_UNQUOTED(DM_LIB_VERSION, $DM_LIB_VERSION, [Library version])
+
+DM_LIB_PATCHLEVEL=`cat VERSION_DM | $AWK -F '[[-. ]]' '{printf "%s.%s.%s",$1,$2,$3}'`
+
+LVM_VERSION="\"`cat VERSION 2>/dev/null || echo Unknown`\""
+
+VER=`cat VERSION`
+LVM_RELEASE_DATE="\"`echo $VER | $SED 's/.*(//;s/).*//'`\""
+VER=`echo "$VER" | $AWK '{print $1}'`
+LVM_RELEASE="\"`echo "$VER" | $AWK -F '-' '{print $2}'`\""
+VER=`echo "$VER" | $AWK -F '-' '{print $1}'`
+LVM_MAJOR=`echo "$VER" | $AWK -F '.' '{print $1}'`
+LVM_MINOR=`echo "$VER" | $AWK -F '.' '{print $2}'`
+LVM_PATCHLEVEL=`echo "$VER" | $AWK -F '.' '{print $3}'`
+
+################################################################################
+AC_SUBST(APPLIB)
+AC_SUBST(BUILD_DMEVENTD)
 AC_SUBST(CFLAGS)
-AC_SUBST(COPTIMISE_FLAG)
+AC_SUBST(CFLOW_CMD)
 AC_SUBST(CLDFLAGS)
-AC_SUBST(CLDWHOLEARCHIVE)
 AC_SUBST(CLDNOWHOLEARCHIVE)
-AC_SUBST(LDDEPS)
-AC_SUBST(LDFLAGS)
-AC_SUBST(LIB_SUFFIX)
-AC_SUBST(LIBS)
-AC_SUBST(LVM_VERSION)
-AC_SUBST(LVM1_FALLBACK)
+AC_SUBST(CLDWHOLEARCHIVE)
+AC_SUBST(CLUSTER)
+AC_SUBST(CLVMD)
+AC_SUBST(CMDLIB)
+AC_SUBST(CONFDB_CFLAGS)
+AC_SUBST(CONFDB_LIBS)
+AC_SUBST(CONFDIR)
+AC_SUBST(COPTIMISE_FLAG)
+AC_SUBST(CPG_CFLAGS)
+AC_SUBST(CPG_LIBS)
+AC_SUBST(CSCOPE_CMD)
 AC_SUBST(DEBUG)
 AC_SUBST(DEVMAPPER)
-AC_SUBST(HAVE_LIBDL)
-AC_SUBST(HAVE_SELINUX)
-AC_SUBST(CMDLIB)
-AC_SUBST(MSGFMT)
-AC_SUBST(LOCALEDIR)
-AC_SUBST(CONFDIR)
-AC_SUBST(STATICDIR)
-AC_SUBST(INTL_PACKAGE)
-AC_SUBST(INTL)
-AC_SUBST(CLVMD)
-AC_SUBST(CLUSTER)
-AC_SUBST(FSADM)
 AC_SUBST(DMEVENTD)
+AC_SUBST(DM_COMPAT)
+AC_SUBST(DM_DEVICE_GID)
+AC_SUBST(DM_DEVICE_MODE)
+AC_SUBST(DM_DEVICE_UID)
+AC_SUBST(DM_IOCTLS)
+AC_SUBST(DM_LIB_VERSION)
+AC_SUBST(DM_LIB_PATCHLEVEL)
+AC_SUBST(FSADM)
+AC_SUBST(GROUP)
+AC_SUBST(HAVE_LIBDL)
+AC_SUBST(HAVE_REALTIME)
+AC_SUBST(HAVE_SELINUX)
+AC_SUBST(INTL)
+AC_SUBST(INTL_PACKAGE)
+AC_SUBST(JOBS)
+AC_SUBST(LDDEPS)
+AC_SUBST(LIBS)
+AC_SUBST(LIB_SUFFIX)
+AC_SUBST(LOCALEDIR)
+AC_SUBST(LVM1)
+AC_SUBST(LVM1_FALLBACK)
+AC_SUBST(LVM_VERSION)
+AC_SUBST(LVM_MAJOR)
+AC_SUBST(LVM_MINOR)
+AC_SUBST(LVM_PATCHLEVEL)
+AC_SUBST(LVM_RELEASE)
+AC_SUBST(LVM_RELEASE_DATE)
+AC_SUBST(MIRRORS)
+AC_SUBST(MSGFMT)
+AC_SUBST(OWNER)
+AC_SUBST(PKGCONFIG)
+AC_SUBST(POOL)
+AC_SUBST(QUORUM_CFLAGS)
+AC_SUBST(QUORUM_LIBS)
+AC_SUBST(SNAPSHOTS)
+AC_SUBST(STATICDIR)
+AC_SUBST(STATIC_LINK)
+AC_SUBST([LIB_PTHREAD])
+AC_SUBST(interface)
+AC_SUBST(kerneldir)
+AC_SUBST(missingkernel)
+AC_SUBST(kernelvsn)
+AC_SUBST(tmpdir)
+AC_SUBST(usrlibdir)
+AC_SUBST(usrsbindir)
 
 ################################################################################
-dnl -- First and last lines should not contain files to generate in order to 
+dnl -- First and last lines should not contain files to generate in order to
 dnl -- keep utility scripts running properly
-AC_OUTPUT( 								\
-Makefile								\
-make.tmpl								\
-daemons/Makefile							\
-daemons/clvmd/Makefile							\
-dmeventd/Makefile							\
-dmeventd/mirror/Makefile						\
-doc/Makefile								\
-include/Makefile						 	\
-lib/Makefile							 	\
-lib/format1/Makefile						 	\
-lib/format_pool/Makefile						\
-lib/locking/Makefile							\
-lib/mirror/Makefile							\
-lib/snapshot/Makefile							\
-man/Makefile							 	\
-po/Makefile								\
-tools/Makefile							 	\
-tools/version.h								\
-tools/fsadm/Makefile							\
-test/mm/Makefile							\
-test/device/Makefile							\
-test/format1/Makefile							\
-test/regex/Makefile                                                     \
-test/filters/Makefile                                                   \
-)
+AC_CONFIG_FILES([
+Makefile
+make.tmpl
+daemons/Makefile
+daemons/clvmd/Makefile
+daemons/dmeventd/Makefile
+daemons/dmeventd/libdevmapper-event.pc
+daemons/dmeventd/plugins/Makefile
+daemons/dmeventd/plugins/mirror/Makefile
+daemons/dmeventd/plugins/snapshot/Makefile
+doc/Makefile
+include/Makefile
+lib/Makefile
+lib/format1/Makefile
+lib/format_pool/Makefile
+lib/locking/Makefile
+lib/mirror/Makefile
+lib/misc/lvm-version.h
+lib/snapshot/Makefile
+libdm/Makefile
+libdm/libdevmapper.pc
+liblvm/Makefile
+man/Makefile
+po/Makefile
+scripts/clvmd_init_red_hat
+scripts/Makefile
+test/Makefile
+test/api/Makefile
+tools/Makefile
+])
+AC_OUTPUT
 
 if test x$ODIRECT != xyes; then
   AC_MSG_WARN(Warning: O_DIRECT disabled: low-memory pvmove may lock up)
 fi
-
-if test x$FSADM == xyes; then
-  AC_MSG_WARN(fsadm support is untested)
-fi
-
-if test x$DMEVENTD == xyes; then
-  AC_MSG_WARN(dmeventd support is untested)
-fi

daemons/Makefile.in

@@ -1,7 +1,7 @@
 #
 # Copyright (C) 2004 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -15,9 +15,18 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
+.PHONY: dmeventd clvmd
+
 ifneq ("@CLVMD@", "none")
   SUBDIRS = clvmd
 endif
 
+ifeq ("@DMEVENTD@", "yes")
+  SUBDIRS += dmeventd
+endif
+
 include $(top_srcdir)/make.tmpl
 
+ifeq ("@DMEVENTD@", "yes")
+device-mapper: dmeventd.device-mapper
+endif

daemons/clogd/Makefile

@@ -0,0 +1,77 @@
+###############################################################################
+###############################################################################
+##
+##  Copyright (C) 2009 Red Hat, Inc.  All rights reserved.
+##
+##  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.
+##
+###############################################################################
+###############################################################################
+
+SOURCES = clogd.c cluster.c functions.c link_mon.c local.c logging.c
+
+TARGET = $(shell if [ ! -e /usr/include/linux/dm-clog-tfr.h ]; then \
+		echo 'no_clogd_kernel_headers'; \
+	elif [ ! -e /usr/include/linux/ext2_fs.h ]; then \
+		echo 'no_e2fsprogs_devel'; \
+	elif [ ! -e /usr/include/openais/saCkpt.h ]; then \
+		echo 'no_openais_devel'; \
+	else \
+		echo 'clogd'; \
+	fi)
+
+ifneq ($(DEBUG), )
+CFLAGS += -DDEBUG
+endif
+
+ifneq ($(MEMB), )
+CFLAGS += -DMEMB
+endif
+
+ifneq ($(CKPT), )
+CFLAGS += -DCKPT
+endif
+
+ifneq ($(RESEND), )
+CFLAGS += -DRESEND
+endif
+
+CFLAGS += -g
+
+LDFLAGS += $(shell if [ -e /usr/lib64/openais ]; then \
+		echo '-L/usr/lib64/openais -L/usr/lib64'; \
+	else \
+		echo '-L/usr/lib/openais -L/usr/lib'; \
+	fi)
+LDFLAGS += -lcpg -lSaCkpt -lext2fs
+
+all: ${TARGET}
+
+clogd: ${SOURCES}
+	${CC} ${CFLAGS} -o $@ $^ ${LDFLAGS}
+
+no_clogd_kernel_headers:
+	echo "Unable to find clogd kernel headers"
+	exit 1
+
+no_e2fsprogs_devel:
+	echo "Unable to find ext2fs kernel headers."
+	echo "Install 'e2fsprogs-devel'?"
+	exit 1
+
+no_openais_devel:
+	echo "Unable to find openAIS headers."
+	echo "http://sources.redhat.com/cluster/wiki/"
+	exit 1
+
+install: clogd
+	install -d /usr/sbin
+	install clogd /usr/sbin
+
+uninstall:
+	rm /usr/sbin/clogd
+
+clean:
+	rm -f *.o clogd *~

daemons/clogd/clogd.c

@@ -0,0 +1,265 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <string.h>
+#include <errno.h>
+#include <sched.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <sys/stat.h>
+#include <signal.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <linux/types.h>
+#include <sys/socket.h>
+#include <linux/netlink.h>
+#include <linux/dm-clog-tfr.h>
+#include <linux/dm-ioctl.h>
+
+#include "functions.h"
+#include "local.h"
+#include "cluster.h"
+#include "common.h"
+#include "logging.h"
+#include "link_mon.h"
+
+static int exit_now = 0;
+static sigset_t signal_mask;
+static int signal_received;
+
+static void process_signals(void);
+static void daemonize(void);
+static void init_all(void);
+static void cleanup_all(void);
+
+int main(int argc, char *argv[])
+{
+	daemonize();
+
+	init_all();
+
+	/* Parent can now exit, we're ready to handle requests */
+	kill(getppid(), SIGTERM);
+
+	LOG_PRINT("Starting clogd:");
+	LOG_PRINT(" Built: "__DATE__" "__TIME__"\n");
+	LOG_DBG(" Compiled with debugging.");
+
+	while (!exit_now) {
+		links_monitor();
+
+		links_issue_callbacks();
+
+		process_signals();
+	}
+	exit(EXIT_SUCCESS);
+}
+
+/*
+ * parent_exit_handler: exit the parent
+ * @sig: the signal
+ *
+ */
+static void parent_exit_handler(int sig)
+{
+	exit_now = 1;
+}
+
+/*
+ * create_lockfile - create and lock a lock file
+ * @lockfile: location of lock file
+ *
+ * Returns: 0 on success, -1 otherwise
+ */
+static int create_lockfile(char *lockfile)
+{
+	int fd;
+	struct flock lock;
+	char buffer[50];
+
+	if((fd = open(lockfile, O_CREAT | O_WRONLY,
+		      (S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH))) < 0)
+		return -errno;
+
+	lock.l_type = F_WRLCK;
+	lock.l_start = 0;
+	lock.l_whence = SEEK_SET;
+	lock.l_len = 0;
+
+	if (fcntl(fd, F_SETLK, &lock) < 0) {
+		close(fd);
+		return -errno;
+	}
+
+	if (ftruncate(fd, 0) < 0) {
+		close(fd);
+		return -errno;
+	}
+
+	sprintf(buffer, "%d\n", getpid());
+
+	if(write(fd, buffer, strlen(buffer)) < strlen(buffer)){
+		close(fd);
+		unlink(lockfile);
+		return -errno;
+	}
+
+	return 0;
+}
+
+static void sig_handler(int sig)
+{
+	sigaddset(&signal_mask, sig);
+	++signal_received;
+}
+
+static void process_signal(int sig){
+	int r = 0;
+
+	switch(sig) {
+	case SIGINT:
+	case SIGQUIT:
+	case SIGTERM:
+	case SIGHUP:
+		r += log_status();
+		break;
+	case SIGUSR1:
+	case SIGUSR2:
+		log_debug();
+		/*local_debug();*/
+		cluster_debug();
+		return;
+	default:
+		LOG_PRINT("Unknown signal received... ignoring");
+		return;
+	}
+
+	if (!r) {
+		LOG_DBG("No current cluster logs... safe to exit.");
+		cleanup_all();
+		exit(EXIT_SUCCESS);
+	}
+
+	LOG_ERROR("Cluster logs exist.  Refusing to exit.");
+}
+
+static void process_signals(void)
+{
+	int x;
+
+	if (!signal_received)
+		return;
+
+	signal_received = 0;
+
+	for (x = 1; x < _NSIG; x++) {
+		if (sigismember(&signal_mask, x)) {
+			sigdelset(&signal_mask, x);
+			process_signal(x);
+		}
+	}
+}
+
+/*
+ * daemonize
+ *
+ * Performs the steps necessary to become a daemon.
+ */
+static void daemonize(void)
+{
+	int pid;
+	int status;
+
+	signal(SIGTERM, &parent_exit_handler);
+
+	pid = fork();
+
+	if (pid < 0) {
+		LOG_ERROR("Unable to fork()");
+		exit(EXIT_FAILURE);
+	}
+
+	if (pid) {
+		/* Parent waits here for child to get going */
+		while (!waitpid(pid, &status, WNOHANG) && !exit_now);
+		if (exit_now)
+			exit(EXIT_SUCCESS);
+
+		switch (WEXITSTATUS(status)) {
+		case EXIT_LOCKFILE:
+			LOG_ERROR("Failed to create lockfile");
+			LOG_ERROR("Process already running?");
+			break;
+		case EXIT_KERNEL_TFR_SOCKET:
+			LOG_ERROR("Unable to create netlink socket");
+			break;
+		case EXIT_KERNEL_TFR_BIND:
+			LOG_ERROR("Unable to bind to netlink socket");
+			break;
+		case EXIT_KERNEL_TFR_SETSOCKOPT:
+			LOG_ERROR("Unable to setsockopt on netlink socket");
+			break;
+		case EXIT_CLUSTER_CKPT_INIT:
+			LOG_ERROR("Unable to initialize checkpoint service");
+			LOG_ERROR("Has the cluster infrastructure been started?");
+			break;
+		case EXIT_FAILURE:
+			LOG_ERROR("Failed to start: Generic error");
+			break;
+		default:
+			LOG_ERROR("Failed to start: Unknown error");
+			break;
+		}
+		exit(EXIT_FAILURE);
+	}
+
+	setsid();
+	chdir("/");
+	umask(0);
+
+	close(0); close(1); close(2);
+	open("/dev/null", O_RDONLY); /* reopen stdin */
+	open("/dev/null", O_WRONLY); /* reopen stdout */
+	open("/dev/null", O_WRONLY); /* reopen stderr */
+
+	LOG_OPEN("clogd", LOG_PID, LOG_DAEMON);
+
+	if (create_lockfile("/var/run/clogd.pid"))
+		exit(EXIT_LOCKFILE);
+
+	signal(SIGINT, &sig_handler);
+	signal(SIGQUIT, &sig_handler);
+	signal(SIGTERM, &sig_handler);
+	signal(SIGHUP, &sig_handler);
+	signal(SIGPIPE, SIG_IGN);
+	signal(SIGUSR1, &sig_handler);
+	signal(SIGUSR2, &sig_handler);
+	sigemptyset(&signal_mask);
+	signal_received = 0;
+}
+
+/*
+ * init_all
+ *
+ * Initialize modules.  Exit on failure.
+ */
+static void init_all(void)
+{
+	int r;
+
+	if ((r = init_local()) ||
+	    (r = init_cluster())) {
+		exit(r);
+	}
+}
+
+/*
+ * cleanup_all
+ *
+ * Clean up before exiting
+ */
+static void cleanup_all(void)
+{
+	cleanup_local();
+	cleanup_cluster();
+}

daemons/clogd/cluster.h

@@ -0,0 +1,13 @@
+#ifndef __CLUSTER_LOG_CLUSTER_DOT_H__
+#define __CLUSTER_LOG_CLUSTER_DOT_H__
+
+int init_cluster(void);
+void cleanup_cluster(void);
+void cluster_debug(void);
+
+int create_cluster_cpg(char *str);
+int destroy_cluster_cpg(char *str);
+
+int cluster_send(struct clog_tfr *tfr);
+
+#endif /* __CLUSTER_LOG_CLUSTER_DOT_H__ */

daemons/clogd/common.h

@@ -0,0 +1,40 @@
+#ifndef __CLUSTER_LOG_COMMON_DOT_H__
+#define __CLUSTER_LOG_COMMON_DOT_H__
+
+/*
+#define EXIT_SUCCESS 0
+#define EXIT_FAILURE 1
+*/
+
+#define EXIT_LOCKFILE              2
+
+#define EXIT_KERNEL_TFR_SOCKET     3 /* Failed netlink socket create */
+#define EXIT_KERNEL_TFR_BIND       4
+#define EXIT_KERNEL_TFR_SETSOCKOPT 5
+
+#define EXIT_CLUSTER_CKPT_INIT     6 /* Failed to init checkpoint */
+
+#define EXIT_QUEUE_NOMEM           7
+
+/* Located in dm-clog-tfr.h
+#define RQ_TYPE(x) \
+	((x) == DM_CLOG_CTR) ? "DM_CLOG_CTR" : \
+	((x) == DM_CLOG_DTR) ? "DM_CLOG_DTR" : \
+	((x) == DM_CLOG_PRESUSPEND) ? "DM_CLOG_PRESUSPEND" : \
+	((x) == DM_CLOG_POSTSUSPEND) ? "DM_CLOG_POSTSUSPEND" : \
+	((x) == DM_CLOG_RESUME) ? "DM_CLOG_RESUME" : \
+	((x) == DM_CLOG_GET_REGION_SIZE) ? "DM_CLOG_GET_REGION_SIZE" : \
+	((x) == DM_CLOG_IS_CLEAN) ? "DM_CLOG_IS_CLEAN" : \
+	((x) == DM_CLOG_IN_SYNC) ? "DM_CLOG_IN_SYNC" : \
+	((x) == DM_CLOG_FLUSH) ? "DM_CLOG_FLUSH" : \
+	((x) == DM_CLOG_MARK_REGION) ? "DM_CLOG_MARK_REGION" : \
+	((x) == DM_CLOG_CLEAR_REGION) ? "DM_CLOG_CLEAR_REGION" : \
+	((x) == DM_CLOG_GET_RESYNC_WORK) ? "DM_CLOG_GET_RESYNC_WORK" : \
+	((x) == DM_CLOG_SET_REGION_SYNC) ? "DM_CLOG_SET_REGION_SYNC" : \
+	((x) == DM_CLOG_GET_SYNC_COUNT) ? "DM_CLOG_GET_SYNC_COUNT" : \
+	((x) == DM_CLOG_STATUS_INFO) ? "DM_CLOG_STATUS_INFO" : \
+	((x) == DM_CLOG_STATUS_TABLE) ? "DM_CLOG_STATUS_TABLE" : \
+	NULL
+*/
+
+#endif /* __CLUSTER_LOG_COMMON_DOT_H__ */

daemons/clogd/functions.h

@@ -0,0 +1,20 @@
+#ifndef __CLOG_FUNCTIONS_DOT_H__
+#define __CLOG_FUNCTIONS_DOT_H__
+
+#include <linux/dm-clog-tfr.h>
+
+#define LOG_RESUMED   1
+#define LOG_SUSPENDED 2
+
+int local_resume(struct clog_tfr *tfr);
+int cluster_postsuspend(char *);
+
+int do_request(struct clog_tfr *tfr, int server);
+int push_state(const char *uuid, const char *which, char **buf, uint32_t debug_who);
+int pull_state(const char *uuid, const char *which, char *buf, int size);
+
+int log_get_state(struct clog_tfr *tfr);
+int log_status(void);
+void log_debug(void);
+
+#endif /* __CLOG_FUNCTIONS_DOT_H__ */

daemons/clogd/link_mon.c

@@ -0,0 +1,138 @@
+#include <stdlib.h>
+#include <errno.h>
+#include <poll.h>
+
+#include "logging.h"
+
+struct link_callback {
+	int fd;
+	char *name;
+	void *data;
+	int (*callback)(void *data);
+
+	struct link_callback *next;
+};
+
+static int used_pfds = 0;
+static int free_pfds = 0;
+static struct pollfd *pfds = NULL;
+static struct link_callback *callbacks = NULL;
+
+int links_register(int fd, char *name, int (*callback)(void *data), void *data)
+{
+	int i;
+	struct link_callback *lc;
+
+	for (i = 0; i < used_pfds; i++) {
+		if (fd == pfds[i].fd) {
+			LOG_ERROR("links_register: Duplicate file descriptor");
+			return -EINVAL;
+		}
+	}
+
+	lc = malloc(sizeof(*lc));
+	if (!lc)
+		return -ENOMEM;
+
+	lc->fd = fd;
+	lc->name = name;
+	lc->data = data;
+	lc->callback = callback;
+
+	if (!free_pfds) {
+		struct pollfd *tmp;
+		tmp = realloc(pfds, sizeof(struct pollfd) * ((used_pfds*2) + 1));
+		if (!tmp) {
+			free(lc);
+			return -ENOMEM;
+		}
+		
+		pfds = tmp;
+		free_pfds = used_pfds + 1;
+	}
+
+	free_pfds--;
+	pfds[used_pfds].fd = fd;
+	pfds[used_pfds].events = POLLIN;
+	pfds[used_pfds].revents = 0;
+	used_pfds++;
+
+	lc->next = callbacks;
+	callbacks = lc;
+	LOG_DBG("Adding %s/%d", lc->name, lc->fd);
+	LOG_DBG(" used_pfds = %d, free_pfds = %d",
+		used_pfds, free_pfds);
+
+	return 0;
+}
+
+int links_unregister(int fd)
+{
+	int i;
+	struct link_callback *p, *c;
+
+	for (i = 0; i < used_pfds; i++)
+		if (fd == pfds[i].fd) {
+			/* entire struct is copied (overwritten) */
+			pfds[i] = pfds[used_pfds - 1];
+			used_pfds--;
+			free_pfds++;
+		}
+
+	for (p = NULL, c = callbacks; c; p = c, c = c->next)
+		if (fd == c->fd) {
+			LOG_DBG("Freeing up %s/%d", c->name, c->fd);
+			LOG_DBG(" used_pfds = %d, free_pfds = %d",
+				used_pfds, free_pfds);
+			if (p)
+				p->next = c->next;
+			else
+				callbacks = c->next;
+			free(c);
+			break;
+		}
+
+	return 0;
+}
+
+int links_monitor(void)
+{
+	int i, r;
+
+	for (i = 0; i < used_pfds; i++) {
+		pfds[i].revents = 0;
+	}
+
+	r = poll(pfds, used_pfds, -1);
+	if (r <= 0)
+		return r;
+
+	r = 0;
+	/* FIXME: handle POLLHUP */
+	for (i = 0; i < used_pfds; i++)
+		if (pfds[i].revents & POLLIN) {
+			LOG_DBG("Data ready on %d", pfds[i].fd);
+				
+			/* FIXME: Add this back return 1;*/
+			r++;
+		}
+
+	return r;
+}
+
+int links_issue_callbacks(void)
+{
+	int i;
+	struct link_callback *lc;
+
+	for (i = 0; i < used_pfds; i++)
+		if (pfds[i].revents & POLLIN)
+			for (lc = callbacks; lc; lc = lc->next)
+				if (pfds[i].fd == lc->fd) {
+					LOG_DBG("Issuing callback on %s/%d",
+						lc->name, lc->fd);
+					lc->callback(lc->data);
+					break;
+				}
+	return 0;
+}

daemons/clogd/link_mon.h

@@ -0,0 +1,9 @@
+#ifndef __LINK_MON_DOT_H__
+#define __LINK_MON_DOT_H__
+
+int links_register(int fd, char *name, int (*callback)(void *data), void *data);
+int links_unregister(int fd);
+int links_monitor(void);
+int links_issue_callbacks(void);
+
+#endif /* __LINK_MON_DOT_H__ */

daemons/clogd/list.h

@@ -0,0 +1,471 @@
+#ifndef _LINUX_LIST_H
+#define _LINUX_LIST_H
+
+/*
+ * These are non-NULL pointers that will result in page faults
+ * under normal circumstances, used to verify that nobody uses
+ * non-initialized list entries.
+ */
+#define LIST_POISON1  ((void *) 0x00100100)
+#define LIST_POISON2  ((void *) 0x00200200)
+
+/*
+ * Simple doubly linked list implementation.
+ *
+ * Some of the internal functions ("__xxx") are useful when
+ * manipulating whole lists rather than single entries, as
+ * sometimes we already know the next/prev entries and we can
+ * generate better code by using them directly rather than
+ * using the generic single-entry routines.
+ */
+
+struct list_head {
+	struct list_head *next, *prev;
+};
+
+#define LIST_HEAD_INIT(name) { &(name), &(name) }
+
+#define LIST_HEAD(name) \
+	struct list_head name = LIST_HEAD_INIT(name)
+
+#define INIT_LIST_HEAD(ptr) do { \
+	(ptr)->next = (ptr); (ptr)->prev = (ptr); \
+} while (0)
+
+/*
+ * Insert a new entry between two known consecutive entries.
+ *
+ * This is only for internal list manipulation where we know
+ * the prev/next entries already!
+ */
+static inline void __list_add(struct list_head *new,
+			      struct list_head *prev,
+			      struct list_head *next)
+{
+	next->prev = new;
+	new->next = next;
+	new->prev = prev;
+	prev->next = new;
+}
+
+/**
+ * list_add - add a new entry
+ * @new: new entry to be added
+ * @head: list head to add it after
+ *
+ * Insert a new entry after the specified head.
+ * This is good for implementing stacks.
+ */
+static inline void list_add(struct list_head *new, struct list_head *head)
+{
+	__list_add(new, head, head->next);
+}
+
+/**
+ * list_add_tail - add a new entry
+ * @new: new entry to be added
+ * @head: list head to add it before
+ *
+ * Insert a new entry before the specified head.
+ * This is useful for implementing queues.
+ */
+static inline void list_add_tail(struct list_head *new, struct list_head *head)
+{
+	__list_add(new, head->prev, head);
+}
+
+/*
+ * Delete a list entry by making the prev/next entries
+ * point to each other.
+ *
+ * This is only for internal list manipulation where we know
+ * the prev/next entries already!
+ */
+static inline void __list_del(struct list_head * prev, struct list_head * next)
+{
+	next->prev = prev;
+	prev->next = next;
+}
+
+/**
+ * list_del - deletes entry from list.
+ * @entry: the element to delete from the list.
+ * Note: list_empty on entry does not return true after this, the entry is
+ * in an undefined state.
+ */
+static inline void list_del(struct list_head *entry)
+{
+	__list_del(entry->prev, entry->next);
+	entry->next = LIST_POISON1;
+	entry->prev = LIST_POISON2;
+}
+
+/**
+ * list_del_init - deletes entry from list and reinitialize it.
+ * @entry: the element to delete from the list.
+ */
+static inline void list_del_init(struct list_head *entry)
+{
+	__list_del(entry->prev, entry->next);
+	INIT_LIST_HEAD(entry);
+}
+
+/**
+ * list_move - delete from one list and add as another's head
+ * @list: the entry to move
+ * @head: the head that will precede our entry
+ */
+static inline void list_move(struct list_head *list, struct list_head *head)
+{
+        __list_del(list->prev, list->next);
+        list_add(list, head);
+}
+
+/**
+ * list_move_tail - delete from one list and add as another's tail
+ * @list: the entry to move
+ * @head: the head that will follow our entry
+ */
+static inline void list_move_tail(struct list_head *list,
+				  struct list_head *head)
+{
+        __list_del(list->prev, list->next);
+        list_add_tail(list, head);
+}
+
+/**
+ * list_empty - tests whether a list is empty
+ * @head: the list to test.
+ */
+static inline int list_empty(const struct list_head *head)
+{
+	return head->next == head;
+}
+
+/**
+ * list_empty_careful - tests whether a list is
+ * empty _and_ checks that no other CPU might be
+ * in the process of still modifying either member
+ *
+ * NOTE: using list_empty_careful() without synchronization
+ * can only be safe if the only activity that can happen
+ * to the list entry is list_del_init(). Eg. it cannot be used
+ * if another CPU could re-list_add() it.
+ *
+ * @head: the list to test.
+ */
+static inline int list_empty_careful(const struct list_head *head)
+{
+	struct list_head *next = head->next;
+	return (next == head) && (next == head->prev);
+}
+
+static inline void __list_splice(struct list_head *list,
+				 struct list_head *head)
+{
+	struct list_head *first = list->next;
+	struct list_head *last = list->prev;
+	struct list_head *at = head->next;
+
+	first->prev = head;
+	head->next = first;
+
+	last->next = at;
+	at->prev = last;
+}
+
+/**
+ * list_splice - join two lists
+ * @list: the new list to add.
+ * @head: the place to add it in the first list.
+ */
+static inline void list_splice(struct list_head *list, struct list_head *head)
+{
+	if (!list_empty(list))
+		__list_splice(list, head);
+}
+
+/**
+ * list_splice_init - join two lists and reinitialise the emptied list.
+ * @list: the new list to add.
+ * @head: the place to add it in the first list.
+ *
+ * The list at @list is reinitialised
+ */
+static inline void list_splice_init(struct list_head *list,
+				    struct list_head *head)
+{
+	if (!list_empty(list)) {
+		__list_splice(list, head);
+		INIT_LIST_HEAD(list);
+	}
+}
+
+#define offsetof(TYPE, MEMBER) ((size_t) &((TYPE *)0)->MEMBER)
+
+/**
+ * container_of - cast a member of a structure out to the containing structure
+ *
+ * @ptr:        the pointer to the member.
+ * @type:       the type of the container struct this is embedded in.
+ * @member:     the name of the member within the struct.
+ *
+ */
+#define container_of(ptr, type, member) ({                      \
+        const typeof( ((type *)0)->member ) *__mptr = (ptr);    \
+        (type *)( (char *)__mptr - offsetof(type,member) );})
+
+
+/**
+ * list_entry - get the struct for this entry
+ * @ptr:	the &struct list_head pointer.
+ * @type:	the type of the struct this is embedded in.
+ * @member:	the name of the list_struct within the struct.
+ */
+#define list_entry(ptr, type, member) \
+	container_of(ptr, type, member)
+
+/**
+ * list_for_each	-	iterate over a list
+ * @pos:	the &struct list_head to use as a loop counter.
+ * @head:	the head for your list.
+ */
+#define list_for_each(pos, head) \
+	for (pos = (head)->next; prefetch(pos->next), pos != (head); \
+        	pos = pos->next)
+
+/**
+ * __list_for_each	-	iterate over a list
+ * @pos:	the &struct list_head to use as a loop counter.
+ * @head:	the head for your list.
+ *
+ * This variant differs from list_for_each() in that it's the
+ * simplest possible list iteration code, no prefetching is done.
+ * Use this for code that knows the list to be very short (empty
+ * or 1 entry) most of the time.
+ */
+#define __list_for_each(pos, head) \
+	for (pos = (head)->next; pos != (head); pos = pos->next)
+
+/**
+ * list_for_each_prev	-	iterate over a list backwards
+ * @pos:	the &struct list_head to use as a loop counter.
+ * @head:	the head for your list.
+ */
+#define list_for_each_prev(pos, head) \
+	for (pos = (head)->prev; prefetch(pos->prev), pos != (head); \
+        	pos = pos->prev)
+
+/**
+ * list_for_each_safe	-	iterate over a list safe against removal of list entry
+ * @pos:	the &struct list_head to use as a loop counter.
+ * @n:		another &struct list_head to use as temporary storage
+ * @head:	the head for your list.
+ */
+#define list_for_each_safe(pos, n, head) \
+	for (pos = (head)->next, n = pos->next; pos != (head); \
+		pos = n, n = pos->next)
+
+/**
+ * list_for_each_entry	-	iterate over list of given type
+ * @pos:	the type * to use as a loop counter.
+ * @head:	the head for your list.
+ * @member:	the name of the list_struct within the struct.
+ */
+#define list_for_each_entry(pos, head, member)				\
+	for (pos = list_entry((head)->next, typeof(*pos), member);	\
+	     prefetch(pos->member.next), &pos->member != (head); 	\
+	     pos = list_entry(pos->member.next, typeof(*pos), member))
+
+/**
+ * list_for_each_entry_reverse - iterate backwards over list of given type.
+ * @pos:	the type * to use as a loop counter.
+ * @head:	the head for your list.
+ * @member:	the name of the list_struct within the struct.
+ */
+#define list_for_each_entry_reverse(pos, head, member)			\
+	for (pos = list_entry((head)->prev, typeof(*pos), member);	\
+	     prefetch(pos->member.prev), &pos->member != (head); 	\
+	     pos = list_entry(pos->member.prev, typeof(*pos), member))
+
+/**
+ * list_prepare_entry - prepare a pos entry for use as a start point in
+ *			list_for_each_entry_continue
+ * @pos:	the type * to use as a start point
+ * @head:	the head of the list
+ * @member:	the name of the list_struct within the struct.
+ */
+#define list_prepare_entry(pos, head, member) \
+	((pos) ? : list_entry(head, typeof(*pos), member))
+
+/**
+ * list_for_each_entry_continue -	iterate over list of given type
+ *			continuing after existing point
+ * @pos:	the type * to use as a loop counter.
+ * @head:	the head for your list.
+ * @member:	the name of the list_struct within the struct.
+ */
+#define list_for_each_entry_continue(pos, head, member) 		\
+	for (pos = list_entry(pos->member.next, typeof(*pos), member);	\
+	     prefetch(pos->member.next), &pos->member != (head);	\
+	     pos = list_entry(pos->member.next, typeof(*pos), member))
+
+/**
+ * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry
+ * @pos:	the type * to use as a loop counter.
+ * @n:		another type * to use as temporary storage
+ * @head:	the head for your list.
+ * @member:	the name of the list_struct within the struct.
+ */
+#define list_for_each_entry_safe(pos, n, head, member)			\
+	for (pos = list_entry((head)->next, typeof(*pos), member),	\
+		n = list_entry(pos->member.next, typeof(*pos), member);	\
+	     &pos->member != (head); 					\
+	     pos = n, n = list_entry(n->member.next, typeof(*n), member))
+
+/*
+ * Double linked lists with a single pointer list head.
+ * Mostly useful for hash tables where the two pointer list head is
+ * too wasteful.
+ * You lose the ability to access the tail in O(1).
+ */
+
+struct hlist_head {
+	struct hlist_node *first;
+};
+
+struct hlist_node {
+	struct hlist_node *next, **pprev;
+};
+
+#define HLIST_HEAD_INIT { .first = NULL }
+#define HLIST_HEAD(name) struct hlist_head name = {  .first = NULL }
+#define INIT_HLIST_HEAD(ptr) ((ptr)->first = NULL)
+#define INIT_HLIST_NODE(ptr) ((ptr)->next = NULL, (ptr)->pprev = NULL)
+
+static inline int hlist_unhashed(const struct hlist_node *h)
+{
+	return !h->pprev;
+}
+
+static inline int hlist_empty(const struct hlist_head *h)
+{
+	return !h->first;
+}
+
+static inline void __hlist_del(struct hlist_node *n)
+{
+	struct hlist_node *next = n->next;
+	struct hlist_node **pprev = n->pprev;
+	*pprev = next;
+	if (next)
+		next->pprev = pprev;
+}
+
+static inline void hlist_del(struct hlist_node *n)
+{
+	__hlist_del(n);
+	n->next = LIST_POISON1;
+	n->pprev = LIST_POISON2;
+}
+
+static inline void hlist_del_init(struct hlist_node *n)
+{
+	if (n->pprev)  {
+		__hlist_del(n);
+		INIT_HLIST_NODE(n);
+	}
+}
+
+static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h)
+{
+	struct hlist_node *first = h->first;
+	n->next = first;
+	if (first)
+		first->pprev = &n->next;
+	h->first = n;
+	n->pprev = &h->first;
+}
+
+/* next must be != NULL */
+static inline void hlist_add_before(struct hlist_node *n,
+					struct hlist_node *next)
+{
+	n->pprev = next->pprev;
+	n->next = next;
+	next->pprev = &n->next;
+	*(n->pprev) = n;
+}
+
+static inline void hlist_add_after(struct hlist_node *n,
+					struct hlist_node *next)
+{
+	next->next = n->next;
+	n->next = next;
+	next->pprev = &n->next;
+
+	if(next->next)
+		next->next->pprev  = &next->next;
+}
+
+#define hlist_entry(ptr, type, member) container_of(ptr,type,member)
+
+#define hlist_for_each(pos, head) \
+	for (pos = (head)->first; pos && ({ prefetch(pos->next); 1; }); \
+	     pos = pos->next)
+
+#define hlist_for_each_safe(pos, n, head) \
+	for (pos = (head)->first; pos && ({ n = pos->next; 1; }); \
+	     pos = n)
+
+/**
+ * hlist_for_each_entry	- iterate over list of given type
+ * @tpos:	the type * to use as a loop counter.
+ * @pos:	the &struct hlist_node to use as a loop counter.
+ * @head:	the head for your list.
+ * @member:	the name of the hlist_node within the struct.
+ */
+#define hlist_for_each_entry(tpos, pos, head, member)			 \
+	for (pos = (head)->first;					 \
+	     pos && ({ prefetch(pos->next); 1;}) &&			 \
+		({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \
+	     pos = pos->next)
+
+/**
+ * hlist_for_each_entry_continue - iterate over a hlist continuing after existing point
+ * @tpos:	the type * to use as a loop counter.
+ * @pos:	the &struct hlist_node to use as a loop counter.
+ * @member:	the name of the hlist_node within the struct.
+ */
+#define hlist_for_each_entry_continue(tpos, pos, member)		 \
+	for (pos = (pos)->next;						 \
+	     pos && ({ prefetch(pos->next); 1;}) &&			 \
+		({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \
+	     pos = pos->next)
+
+/**
+ * hlist_for_each_entry_from - iterate over a hlist continuing from existing point
+ * @tpos:	the type * to use as a loop counter.
+ * @pos:	the &struct hlist_node to use as a loop counter.
+ * @member:	the name of the hlist_node within the struct.
+ */
+#define hlist_for_each_entry_from(tpos, pos, member)			 \
+	for (; pos && ({ prefetch(pos->next); 1;}) &&			 \
+		({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \
+	     pos = pos->next)
+
+/**
+ * hlist_for_each_entry_safe - iterate over list of given type safe against removal of list entry
+ * @tpos:	the type * to use as a loop counter.
+ * @pos:	the &struct hlist_node to use as a loop counter.
+ * @n:		another &struct hlist_node to use as temporary storage
+ * @head:	the head for your list.
+ * @member:	the name of the hlist_node within the struct.
+ */
+#define hlist_for_each_entry_safe(tpos, pos, n, head, member) 		 \
+	for (pos = (head)->first;					 \
+	     pos && ({ n = pos->next; 1; }) && 				 \
+		({ tpos = hlist_entry(pos, typeof(*tpos), member); 1;}); \
+	     pos = n)
+
+#endif

daemons/clogd/local.c

@@ -0,0 +1,379 @@
+#include <unistd.h>
+#include <errno.h>
+#include <string.h>
+#include <stdint.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <sys/poll.h>
+#include <linux/connector.h>
+#include <linux/netlink.h>
+
+#include "linux/dm-clog-tfr.h"
+#include "functions.h"
+#include "cluster.h"
+#include "common.h"
+#include "logging.h"
+#include "link_mon.h"
+#include "local.h"
+
+static int cn_fd;  /* Connector (netlink) socket fd */
+static char recv_buf[2048];
+
+
+/* FIXME: merge this function with kernel_send_helper */
+static int kernel_ack(uint32_t seq, int error)
+{
+	int r;
+	unsigned char buf[sizeof(struct nlmsghdr) + sizeof(struct cn_msg)];
+	struct nlmsghdr *nlh = (struct nlmsghdr *)buf;
+	struct cn_msg *msg = NLMSG_DATA(nlh);
+
+	if (error < 0) {
+		LOG_ERROR("Programmer error: error codes must be positive");
+		return -EINVAL;
+	}
+
+	memset(buf, 0, sizeof(buf));
+
+	nlh->nlmsg_seq = 0;
+	nlh->nlmsg_pid = getpid();
+	nlh->nlmsg_type = NLMSG_DONE;
+	nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct cn_msg));
+	nlh->nlmsg_flags = 0;
+
+	msg->len = 0;
+	msg->id.idx = 0x4;
+	msg->id.val = 0x1;
+	msg->seq = seq;
+	msg->ack = error;
+
+	r = send(cn_fd, nlh, NLMSG_LENGTH(sizeof(struct cn_msg)), 0);
+	/* FIXME: do better error processing */
+	if (r <= 0)
+		return -EBADE;
+
+	return 0;
+}
+
+
+/*
+ * kernel_recv
+ * @tfr: the newly allocated request from kernel
+ *
+ * Read requests from the kernel and allocate space for the new request.
+ * If there is no request from the kernel, *tfr is NULL.
+ *
+ * This function is not thread safe due to returned stack pointer.  In fact,
+ * the returned pointer must not be in-use when this function is called again.
+ *
+ * Returns: 0 on success, -EXXX on error
+ */
+static int kernel_recv(struct clog_tfr **tfr)
+{
+	int r = 0;
+	int len;
+	struct cn_msg *msg;
+
+	*tfr = NULL;
+	memset(recv_buf, 0, sizeof(recv_buf));
+
+	len = recv(cn_fd, recv_buf, sizeof(recv_buf), 0);
+	if (len < 0) {
+		LOG_ERROR("Failed to recv message from kernel");
+		r = -errno;
+		goto fail;
+	}
+
+	switch (((struct nlmsghdr *)recv_buf)->nlmsg_type) {
+	case NLMSG_ERROR:
+		LOG_ERROR("Unable to recv message from kernel: NLMSG_ERROR");
+		r = -EBADE;
+		goto fail;
+	case NLMSG_DONE:
+		msg = (struct cn_msg *)NLMSG_DATA((struct nlmsghdr *)recv_buf);
+		len -= sizeof(struct nlmsghdr);
+
+		if (len < sizeof(struct cn_msg)) {
+			LOG_ERROR("Incomplete request from kernel received");
+			r = -EBADE;
+			goto fail;
+		}
+
+		if (msg->len > DM_CLOG_TFR_SIZE) {
+			LOG_ERROR("Not enough space to receive kernel request (%d/%d)",
+				  msg->len, DM_CLOG_TFR_SIZE);
+			r = -EBADE;
+			goto fail;
+		}
+
+		if (!msg->len)
+			LOG_ERROR("Zero length message received");
+
+		len -= sizeof(struct cn_msg);
+
+		if (len < msg->len)
+			LOG_ERROR("len = %d, msg->len = %d", len, msg->len);
+
+		msg->data[msg->len] = '\0'; /* Cleaner way to ensure this? */
+		*tfr = (struct clog_tfr *)msg->data;
+
+		if (!(*tfr)->request_type) {
+			LOG_DBG("Bad transmission, requesting resend [%u]", msg->seq);
+			r = -EAGAIN;
+
+			if (kernel_ack(msg->seq, EAGAIN)) {
+				LOG_ERROR("Failed to NACK kernel transmission [%u]",
+					  msg->seq);
+				r = -EBADE;
+			}
+		}
+		break;
+	default:
+		LOG_ERROR("Unknown nlmsg_type");
+		r = -EBADE;
+	}
+
+fail:
+	if (r)
+		*tfr = NULL;
+
+	return (r == -EAGAIN) ? 0 : r;
+}
+
+static int kernel_send_helper(void *data, int out_size)
+{
+	int r;
+	struct nlmsghdr *nlh;
+	struct cn_msg *msg;
+	unsigned char buf[2048];
+
+	memset(buf, 0, sizeof(buf));
+
+	nlh = (struct nlmsghdr *)buf;
+	nlh->nlmsg_seq = 0;  /* FIXME: Is this used? */
+	nlh->nlmsg_pid = getpid();
+	nlh->nlmsg_type = NLMSG_DONE;
+	nlh->nlmsg_len = NLMSG_LENGTH(out_size + sizeof(struct cn_msg));
+	nlh->nlmsg_flags = 0;
+
+	msg = NLMSG_DATA(nlh);
+	memcpy(msg->data, data, out_size);
+	msg->len = out_size;
+	msg->id.idx = 0x4;
+	msg->id.val = 0x1;
+	msg->seq = 0;
+
+	r = send(cn_fd, nlh, NLMSG_LENGTH(out_size + sizeof(struct cn_msg)), 0);
+	/* FIXME: do better error processing */
+	if (r <= 0)
+		return -EBADE;
+
+	return 0;
+}
+
+/*
+ * do_local_work
+ *
+ * Any processing errors are placed in the 'tfr'
+ * structure to be reported back to the kernel.
+ * It may be pointless for this function to
+ * return an int.
+ *
+ * Returns: 0 on success, -EXXX on failure
+ */
+static int do_local_work(void *data)
+{
+	int r;
+	struct clog_tfr *tfr = NULL;
+
+	r = kernel_recv(&tfr);
+	if (r)
+		return r;
+
+	if (!tfr)
+		return 0;
+
+	LOG_DBG("[%s]  Request from kernel received: [%s/%u]",
+		SHORT_UUID(tfr->uuid), RQ_TYPE(tfr->request_type),
+		tfr->seq);
+	switch (tfr->request_type) {
+	case DM_CLOG_CTR:
+	case DM_CLOG_DTR:
+	case DM_CLOG_IN_SYNC:
+	case DM_CLOG_GET_SYNC_COUNT:
+	case DM_CLOG_STATUS_INFO:
+	case DM_CLOG_STATUS_TABLE:
+	case DM_CLOG_PRESUSPEND:
+		/* We do not specify ourselves as server here */
+		r = do_request(tfr, 0);
+		if (r)
+			LOG_DBG("Returning failed request to kernel [%s]",
+				RQ_TYPE(tfr->request_type));
+		r = kernel_send(tfr);
+		if (r)
+			LOG_ERROR("Failed to respond to kernel [%s]",
+				  RQ_TYPE(tfr->request_type));
+			
+		break;
+	case DM_CLOG_RESUME:
+		/*
+		 * Resume is a special case that requires a local
+		 * component to join the CPG, and a cluster component
+		 * to handle the request.
+		 */
+		r = local_resume(tfr);
+		if (r) {
+			LOG_DBG("Returning failed request to kernel [%s]",
+				RQ_TYPE(tfr->request_type));
+			r = kernel_send(tfr);
+			if (r)
+				LOG_ERROR("Failed to respond to kernel [%s]",
+					  RQ_TYPE(tfr->request_type));
+			break;
+		}
+		/* ELSE, fall through */
+	case DM_CLOG_IS_CLEAN:
+	case DM_CLOG_FLUSH:
+	case DM_CLOG_MARK_REGION:
+	case DM_CLOG_GET_RESYNC_WORK:
+	case DM_CLOG_SET_REGION_SYNC:
+	case DM_CLOG_IS_REMOTE_RECOVERING:
+	case DM_CLOG_POSTSUSPEND:
+		r = cluster_send(tfr);
+		if (r) {
+			tfr->data_size = 0;
+			tfr->error = r;
+			kernel_send(tfr);
+		}
+
+		break;
+	case DM_CLOG_CLEAR_REGION:
+		r = kernel_ack(tfr->seq, 0);
+
+		r = cluster_send(tfr);
+		if (r) {
+			/*
+			 * FIXME: store error for delivery on flush
+			 *        This would allow us to optimize MARK_REGION
+			 *        too.
+			 */
+		}
+
+		break;
+	case DM_CLOG_GET_REGION_SIZE:
+	default:
+		LOG_ERROR("Invalid log request received, ignoring.");
+		return 0;
+	}
+
+	if (r && !tfr->error)
+		tfr->error = r;
+
+	return r;
+}
+
+/*
+ * kernel_send
+ * @tfr: result to pass back to kernel
+ *
+ * This function returns the tfr structure
+ * (containing the results) to the kernel.
+ * It then frees the structure.
+ *
+ * WARNING: should the structure be freed if
+ * there is an error?  I vote 'yes'.  If the
+ * kernel doesn't get the response, it should
+ * resend the request.
+ *
+ * Returns: 0 on success, -EXXX on failure
+ */
+int kernel_send(struct clog_tfr *tfr)
+{
+	int r;
+	int size;
+
+	if (!tfr)
+		return -EINVAL;
+
+	size = sizeof(struct clog_tfr) + tfr->data_size;
+
+	if (!tfr->data_size && !tfr->error) {
+		/* An ACK is all that is needed */
+
+		/* FIXME: add ACK code */
+	} else if (size > DM_CLOG_TFR_SIZE) {
+		/*
+		 * If we gotten here, we've already overrun
+		 * our allotted space somewhere.
+		 *
+		 * We must do something, because the kernel
+		 * is waiting for a response.
+		 */
+		LOG_ERROR("Not enough space to respond to server");
+		tfr->error = -ENOSPC;
+		size = sizeof(struct clog_tfr);
+	}
+
+	r = kernel_send_helper(tfr, size);
+	if (r)
+		LOG_ERROR("Failed to send msg to kernel.");
+
+	return r;
+}
+
+/*
+ * init_local
+ *
+ * Initialize kernel communication socket (netlink)
+ *
+ * Returns: 0 on success, values from common.h on failure
+ */
+int init_local(void)
+{
+	int r = 0;
+	int opt;
+	struct sockaddr_nl addr;
+
+	cn_fd = socket(PF_NETLINK, SOCK_DGRAM, NETLINK_CONNECTOR);
+	if (cn_fd < 0)
+		return EXIT_KERNEL_TFR_SOCKET;
+
+	/* memset to fix valgrind complaint */
+	memset(&addr, 0, sizeof(struct sockaddr_nl));
+
+	addr.nl_family = AF_NETLINK;
+	addr.nl_groups = 0x4;
+	addr.nl_pid = 0;
+
+	r = bind(cn_fd, (struct sockaddr *) &addr, sizeof(addr));
+	if (r < 0) {
+		close(cn_fd);
+		return EXIT_KERNEL_TFR_BIND;
+	}
+
+	opt = addr.nl_groups;
+	r = setsockopt(cn_fd, 270, NETLINK_ADD_MEMBERSHIP, &opt, sizeof(opt));
+	if (r) {
+		close(cn_fd);
+		return EXIT_KERNEL_TFR_SETSOCKOPT;
+	}
+
+	/*
+	r = fcntl(cn_fd, F_SETFL, FNDELAY);
+	*/
+
+	links_register(cn_fd, "local", do_local_work, NULL);
+
+	return 0;
+}
+
+/*
+ * cleanup_local
+ *
+ * Clean up before exiting
+ */
+void cleanup_local(void)
+{
+	links_unregister(cn_fd);
+	close(cn_fd);
+}

daemons/clogd/local.h

@@ -0,0 +1,9 @@
+#ifndef __CLUSTER_LOG_LOCAL_DOT_H__
+#define __CLUSTER_LOG_LOCAL_DOT_H__
+
+int init_local(void);
+void cleanup_local(void);
+
+int kernel_send(struct clog_tfr *tfr);
+
+#endif /* __CLUSTER_LOG_LOCAL_DOT_H__ */

daemons/clogd/logging.c

@@ -0,0 +1,25 @@
+#include <syslog.h>
+
+int log_tabbing = 0;
+int log_is_open = 0;
+
+/*
+ * Variables for various conditional logging
+ */
+#ifdef MEMB
+int log_membership_change = 1;
+#else
+int log_membership_change = 0;
+#endif
+
+#ifdef CKPT
+int log_checkpoint = 1;
+#else
+int log_checkpoint = 0;
+#endif
+
+#ifdef RESEND
+int log_resend_requests = 1;
+#else
+int log_resend_requests = 0;
+#endif

daemons/clogd/logging.h

@@ -0,0 +1,81 @@
+#ifndef __CLUSTER_LOG_LOGGING_DOT_H__
+#define __CLUSTER_LOG_LOGGING_DOT_H__
+
+#include <stdio.h>
+#include <syslog.h>
+
+#if (BITS_PER_LONG == 64)
+#define PRIu64 "lu"
+#define PRId64 "ld"
+#define PRIo64 "lo"
+#define PRIx64 "lx"
+#define PRIX64 "lX"
+#define SCNu64 "lu"
+#define SCNd64 "ld"
+#define SCNo64 "lo"
+#define SCNx64 "lx"
+#define SCNX64 "lX"
+#else
+#define PRIu64 "Lu"
+#define PRId64 "Ld"
+#define PRIo64 "Lo"
+#define PRIx64 "Lx"
+#define PRIX64 "LX"
+#define SCNu64 "Lu"
+#define SCNd64 "Ld"
+#define SCNo64 "Lo"
+#define SCNx64 "Lx"
+#define SCNX64 "LX"
+#endif
+
+/* SHORT_UUID - print last 8 chars of a string */
+#define SHORT_UUID(x) (strlen(x) > 8) ? ((x) + (strlen(x) - 8)) : (x)
+
+extern int log_tabbing;
+extern int log_is_open;
+extern int log_membership_change;
+extern int log_checkpoint;
+extern int log_resend_requests;
+
+#define LOG_OPEN(ident, option, facility) do { \
+		openlog(ident, option, facility); \
+		log_is_open = 1;		  \
+	} while (0)
+
+#define LOG_CLOSE(void) do { \
+		log_is_open = 0; \
+		closelog();	 \
+	} while (0)
+
+#define LOG_OUTPUT(level, f, arg...) do {				\
+		int __i;						\
+		char __buffer[16];					\
+		FILE *fp = (level > LOG_NOTICE) ? stderr : stdout;	\
+		if (log_is_open) {					\
+			for (__i = 0; (__i < log_tabbing) && (__i < 15); __i++) \
+				__buffer[__i] = '\t';			\
+			__buffer[__i] = '\0';				\
+			syslog(level, "%s" f "\n", __buffer, ## arg);	\
+		} else {						\
+			for (__i = 0; __i < log_tabbing; __i++)		\
+				fprintf(fp, "\t");			\
+			fprintf(fp, f "\n", ## arg);			\
+		}							\
+	} while (0)
+
+
+#ifdef DEBUG
+#define LOG_DBG(f, arg...) LOG_OUTPUT(LOG_DEBUG, f, ## arg)
+#else /* DEBUG */
+#define LOG_DBG(f, arg...)
+#endif /* DEBUG */
+
+#define LOG_COND(__X, f, arg...) do {\
+		if (__X) { 	     \
+			LOG_OUTPUT(LOG_NOTICE, f, ## arg); \
+		} \
+	} while (0)
+#define LOG_PRINT(f, arg...) LOG_OUTPUT(LOG_NOTICE, f, ## arg)
+#define LOG_ERROR(f, arg...) LOG_OUTPUT(LOG_ERR, f, ## arg)
+
+#endif /* __CLUSTER_LOG_LOGGING_DOT_H__ */

daemons/clvmd/Makefile.in

@@ -1,7 +1,7 @@
 #
 # Copyright (C) 2004 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -15,69 +15,100 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
+QUORUM_LIBS = @QUORUM_LIBS@
+QUORUM_CFLAGS = @QUORUM_CFLAGS@
+CONFDB_LIBS = @CONFDB_LIBS@
+CONFDB_CFLAGS = @CONFDB_CFLAGS@
+CPG_LIBS = @CPG_LIBS@
+CPG_CFLAGS = @CPG_CFLAGS@
+
 SOURCES = \
 	clvmd-command.c  \
 	clvmd.c          \
 	lvm-functions.c  \
-	system-lv.c
+	refresh_clvmd.c
 
-ifeq ("@CLVMD@", "gulm")
+ifneq (,$(findstring gulm,, "@CLVMD@,"))
 	GULM = yes
 endif
 
-ifeq ("@CLVMD@", "cman")
+ifneq (,$(findstring cman,, "@CLVMD@,"))
 	CMAN = yes
 endif
 
-ifeq ("@CLVMD@", "all")
+ifneq (,$(findstring openais,, "@CLVMD@,"))
+	OPENAIS = yes
+endif
+
+ifneq (,$(findstring corosync,, "@CLVMD@,"))
+	COROSYNC = yes
+endif
+
+ifneq (,$(findstring all,, "@CLVMD@,"))
 	GULM = yes
 	CMAN = yes
+	OPENAIS = yes
+	COROSYNC = yes
 endif
 
 ifeq ("@DEBUG@", "yes")
-	CFLAGS += -DDEBUG
+	DEFS += -DDEBUG
 endif
 
 ifeq ("$(GULM)", "yes")
 	SOURCES += clvmd-gulm.c tcp-comms.c
 	LMLIBS += -lccs -lgulm
-	CFLAGS += -DUSE_GULM
+	DEFS += -DUSE_GULM
 endif
 
 ifeq ("$(CMAN)", "yes")
 	SOURCES += clvmd-cman.c
-	LMLIBS += -ldlm
-	CFLAGS += -DUSE_CMAN
+	LMLIBS += -ldlm -lcman
+	DEFS += -DUSE_CMAN
 endif
 
+ifeq ("$(OPENAIS)", "yes")
+	SOURCES += clvmd-openais.c
+	LMLIBS += -lSaLck -lcpg
+	DEFS += -DUSE_OPENAIS
+endif
+
+ifeq ("$(COROSYNC)", "yes")
+	SOURCES += clvmd-corosync.c
+	LMLIBS += $(QUORUM_LIBS) $(CONFDB_LIBS) $(CPG_LIBS) -ldlm
+	CFLAGS += $(QUORUM_CFLAGS) $(CONFDB_CFLAGS) $(CPG_CFLAGS)
+	DEFS += -DUSE_COROSYNC
+endif
+
+
 TARGETS = \
 	clvmd
 
-LVMLIBS = -llvm
+LVMLIBS = -llvm-internal -lpthread
 
 ifeq ("@DMEVENTD@", "yes")
-	LVMLIBS += -ldevmapper-event -lpthread
+	LVMLIBS += -ldevmapper-event
 endif
  
-ifeq ("@DEVMAPPER@", "yes")
-	LVMLIBS += -ldevmapper
-endif
+LVMLIBS += -ldevmapper
 
-CFLAGS += -D_REENTRANT -fno-strict-aliasing
+DEFS += -D_REENTRANT
+CFLAGS += -fno-strict-aliasing
 
 include $(top_srcdir)/make.tmpl
 
 INSTALL_TARGETS = \
 	install_clvmd
 
-clvmd: $(OBJECTS) $(top_srcdir)/lib/liblvm.a
-	$(CC) -o clvmd $(OBJECTS) $(LDFLAGS) $(LVMLIBS) $(LMLIBS) $(LIBS)
+clvmd: $(OBJECTS) $(top_srcdir)/lib/liblvm-internal.a
+	$(CC) -o clvmd $(OBJECTS) $(CFLAGS) $(LDFLAGS) \
+		$(LVMLIBS) $(LMLIBS) $(LIBS)
 
 .PHONY: install_clvmd
 
 install_clvmd: $(TARGETS)
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) clvmd \
-		$(sbindir)/clvmd
+		$(usrsbindir)/clvmd
 
 install: $(INSTALL_TARGETS)
 

daemons/clvmd/clvm.h

@@ -1,6 +1,6 @@
 /*
  * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
@@ -62,5 +62,11 @@ static const char CLVMD_SOCKNAME[] = "\0clvmd";
 /* 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
 #endif

daemons/clvmd/clvmd-cman.c

@@ -17,6 +17,10 @@
  * CMAN communication layer for clvmd.
  */
 
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include <configure.h>
 #include <pthread.h>
 #include <sys/types.h>
 #include <sys/stat.h>
@@ -36,29 +40,38 @@
 #include <fcntl.h>
 #include <getopt.h>
 #include <errno.h>
+#include <libdevmapper.h>
+#include <libdlm.h>
 
 #include "clvmd-comms.h"
 #include "clvm.h"
-#include "libdlm.h"
-#include "log.h"
+#include "lvm-logging.h"
 #include "clvmd.h"
 #include "lvm-functions.h"
 
 #define LOCKSPACE_NAME "clvmd"
 
-static int cluster_sock;
+struct clvmd_node
+{
+	struct cman_node *node;
+	int clvmd_up;
+};
+
 static int num_nodes;
-static struct cl_cluster_node *nodes = NULL;
+static struct cman_node *nodes = NULL;
+static struct cman_node this_node;
 static int count_nodes; /* size of allocated nodes array */
-static int max_updown_nodes = 50;	/* Current size of the allocated array */
-/* Node up/down status, indexed by nodeid */
-static int *node_updown = NULL;
+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(char *csid);
+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;
@@ -68,42 +81,45 @@ struct lock_wait {
 
 static int _init_cluster(void)
 {
-	struct sockaddr_cl saddr;
-	int port = CLUSTER_PORT_CLVMD;
+	node_updown_hash = dm_hash_create(100);
 
 	/* Open the cluster communication socket */
-	cluster_sock = socket(AF_CLUSTER, SOCK_DGRAM, CLPROTO_CLIENT);
-	if (cluster_sock == -1) {
-		/* Don't print an error here because we could be just probing for CMAN */
+	c_handle = cman_init(NULL);
+	if (!c_handle) {
+		syslog(LOG_ERR, "Can't open cluster manager socket: %m");
 		return -1;
 	}
-	/* Set Close-on-exec */
-	fcntl(cluster_sock, F_SETFD, 1);
+	DEBUGLOG("Connected to CMAN\n");
 
-	/* Bind to our port number on the cluster.
-	   Writes to this will block if the cluster loses quorum */
-	saddr.scl_family = AF_CLUSTER;
-	saddr.scl_port = port;
-
-	if (bind
-	    (cluster_sock, (struct sockaddr *) &saddr,
-	     sizeof(struct sockaddr_cl))) {
+	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_create_lockspace(LOCKSPACE_NAME, 0600);
 	if (!lockspace) {
-		syslog(LOG_ERR, "Unable to create lockspace for CLVM: %m");
-		return -1;
+		if (errno == EEXIST) {
+			lockspace = dlm_open_lockspace(LOCKSPACE_NAME);
+		}
+		if (!lockspace) {
+			syslog(LOG_ERR, "Unable to create lockspace for CLVM: %m");
+			return -1;
+		}
 	}
 	dlm_ls_pthread_init(lockspace);
-
+	DEBUGLOG("DLM initialisation complete\n");
 	return 0;
 }
 
@@ -114,195 +130,150 @@ static void _cluster_init_completed(void)
 
 static int _get_main_cluster_fd()
 {
-	return cluster_sock;
+	return cman_get_fd(c_handle);
 }
 
 static int _get_num_nodes()
 {
-	return num_nodes;
+	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(void *buf, int msglen, char *csid, const char *errtext)
+static int _cluster_send_message(const void *buf, int msglen, const char *csid,
+				 const char *errtext)
 {
-	struct iovec iov[2];
-	struct msghdr msg;
-	struct sockaddr_cl saddr;
-	int len = 0;
+	int nodeid = 0;
 
-	msg.msg_control = NULL;
-	msg.msg_controllen = 0;
-	msg.msg_iovlen = 1;
-	msg.msg_iov = iov;
-	msg.msg_flags = 0;
-	iov[0].iov_len = msglen;
-	iov[0].iov_base = buf;
+	if (csid)
+		memcpy(&nodeid, csid, CMAN_MAX_CSID_LEN);
 
-	saddr.scl_family = AF_CLUSTER;
-	saddr.scl_port = CLUSTER_PORT_CLVMD;
-	if (csid) {
-		msg.msg_name = &saddr;
-		msg.msg_namelen = sizeof(saddr);
-		memcpy(&saddr.scl_nodeid, csid, CMAN_MAX_CSID_LEN);
-	} else {		/* Cluster broadcast */
-
-		msg.msg_name = NULL;
-		msg.msg_namelen = 0;
+	if (cman_send_data(c_handle, buf, msglen, 0, CLUSTER_PORT_CLVMD, nodeid) <= 0)
+	{
+		log_error("%s", errtext);
 	}
-
-	do {
-		len = sendmsg(cluster_sock, &msg, 0);
-		if (len < 0 && errno != EAGAIN)
-			log_error(errtext);
-
-	} while (len == -1 && errno == EAGAIN);
-	return len;
+	return msglen;
 }
 
 static void _get_our_csid(char *csid)
 {
-	int i;
-	memset(csid, 0, CMAN_MAX_CSID_LEN);
-
-	for (i = 0; i < num_nodes; i++) {
-		if (nodes[i].us)
-			memcpy(csid, &nodes[i].node_id, CMAN_MAX_CSID_LEN);
+	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 that known (down mean not running a clvmd) */
+/* Call a callback routine for each node is that known (down means not running a clvmd) */
 static int _cluster_do_node_callback(struct local_client *client,
-			     void (*callback) (struct local_client *, char *,
-					       int))
+				     void (*callback) (struct local_client *,
+						       const char *,
+						       int))
 {
 	int i;
 	int somedown = 0;
 
 	for (i = 0; i < _get_num_nodes(); i++) {
-		callback(client, (char *)&nodes[i].node_id, node_updown[nodes[i].node_id]);
-		if (!node_updown[nodes[i].node_id])
-			somedown = -1;
+		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 message from the cluster socket,
-   this currently just means that a node has stopped listening on our port */
-static void process_oob_msg(char *buf, int len, int nodeid)
+/* Process OOB messages from the cluster socket */
+static void event_callback(cman_handle_t handle, void *private, int reason, int arg)
 {
-	char namebuf[256];
-	switch (buf[0]) {
-        case CLUSTER_OOB_MSG_PORTCLOSED:
-		name_from_nodeid(nodeid, namebuf);
-		log_notice("clvmd on node %s has died\n", namebuf);
-		DEBUGLOG("Got OOB message, removing node %s\n", namebuf);
+	char namebuf[MAX_CLUSTER_MEMBER_NAME_LEN];
 
-		node_updown[nodeid] = 0;
+	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 CLUSTER_OOB_MSG_STATECHANGE:
-		DEBUGLOG("Got OOB message, Cluster state change\n");
+	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 OOB message: %d\n", buf[0]);
+		DEBUGLOG("Got unknown event callback message: %d\n", reason);
+		break;
 	}
 }
 
-static int _cluster_fd_callback(struct local_client *client, char *buf, int len, char *csid,
-			struct local_client **new_client)
+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)
 {
-	struct iovec iov[2];
-	struct msghdr msg;
-	struct sockaddr_cl saddr;
+
+	/* Save this for data_callback */
+	cman_client = fd;
 
 	/* We never return a new client */
 	*new_client = NULL;
 
-	msg.msg_control = NULL;
-	msg.msg_controllen = 0;
-	msg.msg_iovlen = 1;
-	msg.msg_iov = iov;
-	msg.msg_name = &saddr;
-	msg.msg_flags = 0;
-	msg.msg_namelen = sizeof(saddr);
-	iov[0].iov_len = len;
-	iov[0].iov_base = buf;
-
-	len = recvmsg(cluster_sock, &msg, MSG_OOB | O_NONBLOCK);
-	if (len < 0 && errno == EAGAIN)
-		return len;
-
-	DEBUGLOG("Read on cluster socket, len = %d\n", len);
-
-	/* A real error */
-	if (len < 0) {
-		log_error("read error on cluster socket: %m");
-		return 0;
-	}
-
-	/* EOF - we have left the cluster */
-	if (len == 0)
-		return 0;
-
-	/* Is it OOB? probably a node gone down */
-	if (msg.msg_flags & MSG_OOB) {
-		process_oob_msg(iov[0].iov_base, len, saddr.scl_nodeid);
-
-		/* Tell the upper layer to ignore this message */
-		len = -1;
-		errno = EAGAIN;
-	}
-	else {
-		memcpy(csid, &saddr.scl_nodeid, sizeof(saddr.scl_nodeid));
-		/* Send it back to clvmd */
-		process_message(client, buf, len, csid);
-	}
-	return len;
+	return cman_dispatch(c_handle, 0);
 }
 
-static void _add_up_node(char *csid)
+
+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);
 
-	if (nodeid >= max_updown_nodes) {
-	        int new_size = nodeid + 10;
-		int *new_updown = realloc(node_updown, new_size);
-
-		if (new_updown) {
-			node_updown = new_updown;
-			max_updown_nodes = new_size;
-			DEBUGLOG("realloced more space for nodes. now %d\n",
-				 max_updown_nodes);
-		} else {
-			log_error
-			    ("Realloc failed. Node status for clvmd will be wrong. quitting\n");
-			exit(999);
-		}
-	}
-	node_updown[nodeid] = 1;
+	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()
 {
-	unlock_all();
+	destroy_lvhash();
 	dlm_release_lockspace(LOCKSPACE_NAME, lockspace, 1);
-	close(cluster_sock);
+	cman_finish(c_handle);
 }
 
 static int is_listening(int nodeid)
 {
-	struct cl_listen_request rq;
 	int status;
 
-	rq.port = CLUSTER_PORT_CLVMD;
-	rq.nodeid = nodeid;
-
 	do {
-		status = ioctl(cluster_sock, SIOCCLUSTER_ISLISTENING, &rq);
+		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 */
@@ -320,67 +291,66 @@ static void count_clvmds_running(void)
 	int i;
 
 	for (i = 0; i < num_nodes; i++) {
-		node_updown[nodes[i].node_id] = is_listening(nodes[i].node_id);
+		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()
 {
-	struct cl_cluster_nodelist nodelist;
+	int retnodes;
+	int status;
+	int i;
+	int high_nodeid = 0;
 
-	num_nodes = ioctl(cluster_sock, SIOCCLUSTER_GETMEMBERS, 0);
+	num_nodes = cman_get_node_count(c_handle);
 	if (num_nodes == -1) {
 		log_error("Unable to get node count");
-	} else {
-	        /* Not enough room for new nodes list ? */
-	        if (num_nodes > count_nodes && nodes) {
-			free(nodes);
-			nodes = NULL;
-		}
+		return;
+	}
 
-		if (nodes == NULL) {
-		        count_nodes = num_nodes + 10; /* Overallocate a little */
-		        nodes = malloc(count_nodes * sizeof(struct cl_cluster_node));
-			if (!nodes) {
-			        log_error("Unable to allocate nodes array\n");
-				exit(5);
-			}
-		}
-		nodelist.max_members = count_nodes;
-		nodelist.nodes = nodes;
+	/* Not enough room for new nodes list ? */
+	if (num_nodes > count_nodes && nodes) {
+		free(nodes);
+		nodes = NULL;
+	}
 
-		num_nodes = ioctl(cluster_sock, SIOCCLUSTER_GETMEMBERS, &nodelist);
-		if (num_nodes <= 0) {
-		        log_error("Unable to get node details");
-			exit(6);
+	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);
 		}
+	}
 
-		/* Sanity check struct */
-		if (nodes[0].size != sizeof(struct cl_cluster_node)) {
-			log_error
-			    ("sizeof(cl_cluster_node) does not match size returned from the kernel: aborting\n");
-			exit(10);
-		}
+	status = cman_get_nodes(c_handle, count_nodes, &retnodes, nodes);
+	if (status < 0) {
+		log_error("Unable to get node details");
+		exit(6);
+	}
 
-		if (node_updown == NULL) {
-			node_updown =
-			    (int *) malloc(sizeof(int) *
-					   max(num_nodes, max_updown_nodes));
-			memset(node_updown, 0,
-			       sizeof(int) * max(num_nodes, max_updown_nodes));
-		}
+	/* 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, char *name)
+static int _csid_from_name(char *csid, const char *name)
 {
 	int i;
 
 	for (i = 0; i < num_nodes; i++) {
-		if (strcmp(name, nodes[i].name) == 0) {
-			memcpy(csid, &nodes[i].node_id, CMAN_MAX_CSID_LEN);
+		if (strcmp(name, nodes[i].cn_name) == 0) {
+			memcpy(csid, &nodes[i].cn_nodeid, CMAN_MAX_CSID_LEN);
 			return 0;
 		}
 	}
@@ -388,13 +358,13 @@ static int _csid_from_name(char *csid, char *name)
 }
 
 /* Convert a CSID to a node name */
-static int _name_from_csid(char *csid, char *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].node_id, CMAN_MAX_CSID_LEN) == 0) {
-			strcpy(name, nodes[i].name);
+		if (memcmp(csid, &nodes[i].cn_nodeid, CMAN_MAX_CSID_LEN) == 0) {
+			strcpy(name, nodes[i].cn_name);
 			return 0;
 		}
 	}
@@ -409,8 +379,8 @@ static int name_from_nodeid(int nodeid, char *name)
 	int i;
 
 	for (i = 0; i < num_nodes; i++) {
-		if (nodeid == nodes[i].node_id) {
-			strcpy(name, nodes[i].name);
+		if (nodeid == nodes[i].cn_nodeid) {
+			strcpy(name, nodes[i].cn_name);
 			return 0;
 		}
 	}
@@ -420,7 +390,7 @@ static int name_from_nodeid(int nodeid, char *name)
 }
 
 /* Convert a CSID to a node ID */
-static int nodeid_from_csid(char *csid)
+static int nodeid_from_csid(const char *csid)
 {
         int nodeid;
 
@@ -431,7 +401,7 @@ static int nodeid_from_csid(char *csid)
 
 static int _is_quorate()
 {
-	return ioctl(cluster_sock, SIOCCLUSTER_ISQUORATE, 0);
+	return cman_is_quorate(c_handle);
 }
 
 static void sync_ast_routine(void *arg)
@@ -514,6 +484,18 @@ static int _sync_unlock(const char *resource /* UNUSED */, int lockid)
 
 }
 
+static int _get_cluster_name(char *buf, int buflen)
+{
+	cman_cluster_t cluster_info;
+	int status;
+
+	status = cman_get_cluster(c_handle, &cluster_info);
+	if (!status) {
+		strncpy(buf, cluster_info.ci_name, buflen);
+	}
+	return status;
+}
+
 static struct cluster_ops _cluster_cman_ops = {
 	.cluster_init_completed   = _cluster_init_completed,
 	.cluster_send_message     = _cluster_send_message,
@@ -527,6 +509,7 @@ static struct cluster_ops _cluster_cman_ops = {
 	.get_our_csid             = _get_our_csid,
 	.add_up_node              = _add_up_node,
 	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
 	.sync_lock                = _sync_lock,
 	.sync_unlock              = _sync_unlock,
 };

daemons/clvmd/clvmd-command.c

@@ -50,6 +50,10 @@
 
 */
 
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include <configure.h>
 #include <pthread.h>
 #include <sys/types.h>
 #include <sys/utsname.h>
@@ -64,16 +68,18 @@
 #include <stddef.h>
 #include <unistd.h>
 #include <errno.h>
+#include <libdevmapper.h>
+#include <libdlm.h>
 
-#include "libdevmapper.h"
-#include "list.h"
 #include "locking.h"
-#include "log.h"
+#include "lvm-logging.h"
 #include "lvm-functions.h"
 #include "clvmd-comms.h"
 #include "clvm.h"
 #include "clvmd.h"
-#include "libdlm.h"
+
+extern debug_t debug;
+extern struct cluster_ops *clops;
 
 /* This is where all the real work happens:
    NOTE: client will be NULL when this is executed on a remote node */
@@ -84,6 +90,7 @@ int do_command(struct local_client *client, struct clvm_header *msg, int msglen,
 	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;
@@ -93,18 +100,34 @@ int do_command(struct local_client *client, struct clvm_header *msg, int msglen,
 		/* Just a test message */
 	case CLVMD_CMD_TEST:
 		if (arglen > buflen) {
+			char *new_buf;
 			buflen = arglen + 200;
-			*buf = realloc(*buf, buflen);
+			new_buf = realloc(*buf, buflen);
+			if (new_buf == NULL) {
+				status = errno;
+				free (*buf);
+			}
+			*buf = new_buf;
+		}
+		if (*buf) {
+			uname(&nodeinfo);
+			*retlen = 1 + snprintf(*buf, buflen,
+					       "TEST from %s: %s v%s",
+					       nodeinfo.nodename, args,
+					       nodeinfo.release);
 		}
-		uname(&nodeinfo);
-		*retlen = 1 + snprintf(*buf, buflen, "TEST from %s: %s v%s",
-				       nodeinfo.nodename, args,
-				       nodeinfo.release);
 		break;
 
 	case CLVMD_CMD_LOCK_VG:
+		lockname = &args[2];
 		/* Check to see if the VG is in use by LVM1 */
-		status = do_check_lvm1(&args[2]);
+		status = do_check_lvm1(lockname);
+		/* P_#global causes a full cache refresh */
+		if (!strcmp(lockname, "P_" VG_GLOBAL))
+			do_refresh_cache();
+		else
+			drop_metadata(lockname + 2);
+
 		break;
 
 	case CLVMD_CMD_LOCK_LV:
@@ -116,12 +139,42 @@ int do_command(struct local_client *client, struct clvm_header *msg, int msglen,
 		/* Replace EIO with something less scary */
 		if (status == EIO) {
 			*retlen =
-			    1 + snprintf(*buf, buflen,
-					 "Internal lvm error, check syslog");
+			    1 + 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 + snprintf(*buf, buflen, "%s", locktype);
+		break;
+
+	case CLVMD_CMD_REFRESH:
+		do_refresh_cache();
+		break;
+
+	case CLVMD_CMD_SET_DEBUG:
+		debug = args[0];
+		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;
@@ -129,7 +182,7 @@ int do_command(struct local_client *client, struct clvm_header *msg, int msglen,
 
 	/* Check the status of the command and return the error text */
 	if (status) {
-		*retlen = 1 + snprintf(*buf, buflen, strerror(status));
+		*retlen = 1 + snprintf(*buf, buflen, "%s", strerror(status));
 	}
 
 	return status;
@@ -179,12 +232,16 @@ static int lock_vg(struct local_client *client)
 	    dm_hash_remove(lock_hash, lockname);
     }
     else {
-
-	status = sync_lock(lockname, (int)lock_cmd, (int)lock_flags, &lkid);
+	/* Read locks need to be PR; other modes get passed through */
+	if ((lock_cmd & LCK_TYPE_MASK) == LCK_READ) {
+	    lock_cmd &= ~LCK_TYPE_MASK;
+	    lock_cmd |= LCK_PREAD;
+	}
+	status = sync_lock(lockname, (int)lock_cmd, (lock_flags & LCK_NONBLOCK) ? LKF_NOQUEUE : 0, &lkid);
 	if (status)
 	    status = errno;
 	else
-	    dm_hash_insert(lock_hash, lockname, (void *)lkid);
+	    dm_hash_insert(lock_hash, lockname, (void *)(long)lkid);
     }
 
     return status;
@@ -208,11 +265,15 @@ int do_pre_command(struct local_client *client)
 	switch (header->cmd) {
 	case CLVMD_CMD_TEST:
 		status = sync_lock("CLVMD_TEST", LKM_EXMODE, 0, &lockid);
-		client->bits.localsock.private = (void *) lockid;
+		client->bits.localsock.private = (void *)(long)lockid;
 		break;
 
 	case CLVMD_CMD_LOCK_VG:
-       	        status = lock_vg(client);
+		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:
@@ -222,6 +283,13 @@ int do_pre_command(struct local_client *client)
 		status = pre_lock_lv(lock_cmd, lock_flags, lockname);
 		break;
 
+	case CLVMD_CMD_REFRESH:
+	case CLVMD_CMD_GET_CLUSTERNAME:
+	case CLVMD_CMD_SET_DEBUG:
+	case CLVMD_CMD_VG_BACKUP:
+	case CLVMD_CMD_LOCK_QUERY:
+		break;
+
 	default:
 		log_error("Unknown command %d received\n", header->cmd);
 		status = EINVAL;
@@ -249,6 +317,8 @@ int do_post_command(struct local_client *client)
 		break;
 
 	case CLVMD_CMD_LOCK_VG:
+	case CLVMD_CMD_VG_BACKUP:
+	case CLVMD_CMD_LOCK_QUERY:
 		/* Nothing to do here */
 		break;
 

daemons/clvmd/clvmd-comms.h

@@ -25,25 +25,31 @@ struct local_client;
 struct cluster_ops {
 	void (*cluster_init_completed) (void);
 
-	int (*cluster_send_message) (void *buf, int msglen, char *csid,
-				const char *errtext);
-	int (*name_from_csid) (char *csid, char *name);
-	int (*csid_from_name) (char *csid, char *name);
+	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,
-			       char *csid, struct local_client **new_client);
+				    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 *,
-						      char *csid, int node_up));
+					 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) (char *csid);
+	void (*add_up_node) (const char *csid);
 	void (*reread_config) (void);
 	void (*cluster_closedown) (void);
 
-	int (*sync_lock) (const char *resource, int mode, int flags, int *lockid);
+	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);
 
 };
@@ -56,15 +62,53 @@ struct cluster_ops *init_gulm_cluster(void);
 #endif
 
 #ifdef USE_CMAN
-#  include "cnxman-socket.h"
+#  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_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 <openais/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
+
 
 #endif

daemons/clvmd/clvmd-corosync.c

@@ -0,0 +1,643 @@
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2009 Red Hat, Inc. All rights reserved.
+**
+*******************************************************************************
+******************************************************************************/
+
+/* This provides the interface between clvmd and corosync/DLM as the cluster
+ * and lock manager.
+ *
+ */
+
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include <configure.h>
+#include <pthread.h>
+#include <sys/types.h>
+#include <sys/utsname.h>
+#include <sys/ioctl.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/file.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <signal.h>
+#include <fcntl.h>
+#include <string.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <unistd.h>
+#include <errno.h>
+#include <utmpx.h>
+#include <syslog.h>
+#include <assert.h>
+#include <libdevmapper.h>
+
+#include <corosync/corotypes.h>
+#include <corosync/cpg.h>
+#include <corosync/quorum.h>
+#include <corosync/confdb.h>
+#include <libdlm.h>
+
+#include "locking.h"
+#include "lvm-logging.h"
+#include "clvm.h"
+#include "clvmd-comms.h"
+#include "lvm-functions.h"
+#include "clvmd.h"
+
+/* Timeout value for several corosync calls */
+#define LOCKSPACE_NAME "clvmd"
+
+static void 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 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 cpg_callbacks = {
+	.cpg_deliver_fn =            cpg_deliver_callback,
+	.cpg_confchg_fn =            cpg_confchg_callback,
+};
+
+quorum_callbacks_t quorum_callbacks = {
+	.quorum_notify_fn = NULL,
+};
+
+struct node_info
+{
+	enum {NODE_UNKNOWN, NODE_DOWN, NODE_UP, 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 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 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;
+	}
+
+	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,
+				COROSYNC_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,
+						COROSYNC_CSID_LEN, ninfo);
+			}
+		}
+		ninfo->state = NODE_CLVMD;
+	}
+
+	num_nodes = member_list_entries;
+}
+
+static int _init_cluster(void)
+{
+	cs_error_t err;
+
+	node_hash = dm_hash_create(100);
+
+	err = cpg_initialize(&cpg_handle,
+			     &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);
+	}
+
+	err = quorum_initialize(&quorum_handle,
+				&quorum_callbacks);
+	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_create_lockspace(LOCKSPACE_NAME, 0600);
+	if (!lockspace) {
+		if (errno == EEXIST) {
+			lockspace = dlm_open_lockspace(LOCKSPACE_NAME);
+		}
+		if (!lockspace) {
+			syslog(LOG_ERR, "Unable to create lockspace for CLVM: %m");
+			quorum_finalize(quorum_handle);
+			return -1;
+		}
+	}
+	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)
+{
+	DEBUGLOG("cluster_closedown\n");
+	destroy_lvhash();
+
+	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()
+{
+	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;
+	int somedown = 0;
+
+	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_DOWN)
+			callback(master_client, csid, ninfo->state == NODE_CLVMD);
+		if (ninfo->state != NODE_CLVMD)
+			somedown = -1;
+	}
+	return somedown;
+}
+
+/* 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()
+{
+	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)
+{
+	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;
+
+	err = cpg_mcast_joined(cpg_handle, CPG_TYPE_AGREED, iov, 2);
+	return cs_to_errno(err);
+}
+
+/*
+ * 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;
+}
+
+static struct cluster_ops _cluster_corosync_ops = {
+	.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;
+}

daemons/clvmd/clvmd-gulm.c

@@ -40,18 +40,17 @@
 #include <utmpx.h>
 #include <syslog.h>
 #include <assert.h>
+#include <libdevmapper.h>
+#include <ccs.h>
+#include <libgulm.h>
 
-#include "libdevmapper.h"
-#include "ccs.h"
-#include "list.h"
 #include "locking.h"
-#include "log.h"
+#include "lvm-logging.h"
 #include "clvm.h"
 #include "clvmd-comms.h"
 #include "lvm-functions.h"
 #include "clvmd.h"
 #include "clvmd-gulm.h"
-#include "libgulm.h"
 
 /* Hash list of nodes in the cluster */
 static struct dm_hash_table *node_hash;
@@ -86,12 +85,12 @@ struct lock_wait
 };
 
 /* Forward */
-static int read_from_core_sock(struct local_client *client, char *buf, int len, char *csid,
+static int read_from_core_sock(struct local_client *client, char *buf, int len, const char *csid,
 			       struct local_client **new_client);
-static int read_from_lock_sock(struct local_client *client, char *buf, int len, char *csid,
+static int read_from_lock_sock(struct local_client *client, char *buf, int len, const char *csid,
 			       struct local_client **new_client);
 static int get_all_cluster_nodes(void);
-static int _csid_from_name(char *csid, char *name);
+static int _csid_from_name(char *csid, const char *name);
 static void _cluster_closedown(void);
 
 /* In tcp-comms.c */
@@ -249,7 +248,7 @@ static void _cluster_closedown(void)
 {
     DEBUGLOG("cluster_closedown\n");
     in_shutdown = 1;
-    unlock_all();
+    destroy_lvhash();
     lg_lock_logout(gulm_if);
     lg_core_logout(gulm_if);
     lg_release(gulm_if);
@@ -278,7 +277,7 @@ static void drop_expired_locks(char *nodename)
 }
 
 
-static int read_from_core_sock(struct local_client *client, char *buf, int len, char *csid,
+static int read_from_core_sock(struct local_client *client, char *buf, int len, const char *csid,
 			       struct local_client **new_client)
 {
     int status;
@@ -288,7 +287,7 @@ static int read_from_core_sock(struct local_client *client, char *buf, int len,
     return status<0 ? status : 1;
 }
 
-static int read_from_lock_sock(struct local_client *client, char *buf, int len, char *csid,
+static int read_from_lock_sock(struct local_client *client, char *buf, int len, const char *csid,
 			       struct local_client **new_client)
 {
     int status;
@@ -582,7 +581,7 @@ int get_next_node_csid(void **context, char *csid)
     return 1;
 }
 
-int gulm_name_from_csid(char *csid, char *name)
+int gulm_name_from_csid(const char *csid, char *name)
 {
     struct node_info *ninfo;
 
@@ -598,7 +597,7 @@ int gulm_name_from_csid(char *csid, char *name)
 }
 
 
-static int _csid_from_name(char *csid, char *name)
+static int _csid_from_name(char *csid, const char *name)
 {
     struct dm_hash_node *hn;
     struct node_info *ninfo;
@@ -622,7 +621,7 @@ static int _get_num_nodes()
 }
 
 /* Node is now known to be running a clvmd */
-void gulm_add_up_node(char *csid)
+void gulm_add_up_node(const char *csid)
 {
     struct node_info *ninfo;
 
@@ -661,10 +660,11 @@ void add_down_node(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 *, char *csid, int node_up))
+				     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)
     {
@@ -686,12 +686,14 @@ static int _cluster_do_node_callback(struct local_client *master_client,
 	    client = dm_hash_lookup_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
 
 	}
+	DEBUGLOG("down_callback2. node %s, state = %d\n", ninfo->name, ninfo->state);
 	if (ninfo->state != NODE_DOWN)
 		callback(master_client, csid, ninfo->state == NODE_CLVMD);
 
-
+	if (ninfo->state != NODE_CLVMD)
+		somedown = -1;
     }
-    return 0;
+    return somedown;
 }
 
 /* Convert gulm error codes to unix errno numbers */
@@ -730,7 +732,7 @@ static int _lock_resource(char *resource, int mode, int flags, int *lockid)
     pthread_mutex_lock(&lwait.mutex);
 
     /* This needs to be converted from DLM/LVM2 value for GULM */
-    if (flags == LCK_NONBLOCK) flags = lg_lock_flag_Try;
+    if (flags & LKF_NOQUEUE) flags = lg_lock_flag_Try;
 
     dm_hash_insert(lock_hash, resource, &lwait);
     DEBUGLOG("lock_resource '%s', flags=%d, mode=%d\n", resource, flags, mode);
@@ -828,6 +830,7 @@ static int _sync_lock(const char *resource, int mode, int flags, int *lockid)
         }
 	break;
 
+    case LCK_PREAD:
     case LCK_READ:
 	status = _lock_resource(lock1, lg_lock_state_Shared, flags, lockid);
 	if (status)
@@ -864,6 +867,7 @@ static int _sync_unlock(const char *resource, int lockid)
     /* The held lock mode is in the lock id */
     assert(lockid == LCK_EXCL ||
 	   lockid == LCK_READ ||
+	   lockid == LCK_PREAD ||
 	   lockid == LCK_WRITE);
 
     status = _unlock_resource(lock1, lockid);
@@ -963,14 +967,20 @@ static int _get_main_cluster_fd(void)
 	return get_main_gulm_cluster_fd();
 }
 
-static int _cluster_fd_callback(struct local_client *fd, char *buf, int len, char *csid, struct local_client **new_client)
+static int _cluster_fd_callback(struct local_client *fd, char *buf, int len, const char *csid, struct local_client **new_client)
 {
 	return cluster_fd_gulm_callback(fd, buf, len, csid, new_client);
 }
 
-static int _cluster_send_message(void *buf, int msglen, char *csid, const char *errtext)
+static int _cluster_send_message(const void *buf, int msglen, const char *csid, const char *errtext)
 {
-	return gulm_cluster_send_message(buf, msglen, csid, errtext);
+	return gulm_cluster_send_message((char *)buf, msglen, csid, errtext);
+}
+
+static int _get_cluster_name(char *buf, int buflen)
+{
+	strncpy(buf, cluster_name, buflen);
+	return 0;
 }
 
 static struct cluster_ops _cluster_gulm_ops = {
@@ -987,6 +997,7 @@ static struct cluster_ops _cluster_gulm_ops = {
 	.add_up_node              = gulm_add_up_node,
 	.reread_config            = _reread_config,
 	.cluster_closedown        = _cluster_closedown,
+	.get_cluster_name         = _get_cluster_name,
 	.sync_lock                = _sync_lock,
 	.sync_unlock              = _sync_unlock,
 };

daemons/clvmd/clvmd-gulm.h

@@ -1,12 +1,13 @@
 
-
+/* DLM constant that clvmd uses as a generic NONBLOCK lock flag */
+#define LKF_NOQUEUE 1
 
 extern int get_next_node_csid(void **context, char *csid);
 extern void add_down_node(char *csid);
 extern int gulm_fd(void);
-extern int get_ip_address(char *node, char *addr);
-extern void tcp_remove_client(char *csid);
-extern int alloc_client(int fd, char *csid, struct local_client **new_client);
+extern int get_ip_address(const char *node, char *addr);
+extern void tcp_remove_client(const char *csid);
+extern int alloc_client(int fd, const char *csid, struct local_client **new_client);
 
-void gulm_add_up_node(char *csid);
-int gulm_name_from_csid(char *csid, char *name);
+void gulm_add_up_node(const char *csid);
+int gulm_name_from_csid(const char *csid, char *name);

daemons/clvmd/clvmd-openais.c

@@ -0,0 +1,707 @@
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2007 Red Hat, Inc. All rights reserved.
+**
+*******************************************************************************
+******************************************************************************/
+
+/* This provides the interface between clvmd and OpenAIS as the cluster
+ * and lock manager.
+ *
+ */
+
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include <configure.h>
+#include <pthread.h>
+#include <sys/types.h>
+#include <sys/utsname.h>
+#include <sys/ioctl.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/file.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <signal.h>
+#include <fcntl.h>
+#include <string.h>
+#include <stddef.h>
+#include <stdint.h>
+#include <unistd.h>
+#include <errno.h>
+#include <utmpx.h>
+#include <syslog.h>
+#include <assert.h>
+#include <libdevmapper.h>
+
+#include <openais/saAis.h>
+#include <openais/saLck.h>
+#include <openais/cpg.h>
+
+#include "locking.h"
+#include "lvm-logging.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 cpg_deliver_callback (cpg_handle_t handle,
+				  struct cpg_name *groupName,
+				  uint32_t nodeid,
+				  uint32_t pid,
+				  void *msg,
+				  int msg_len);
+static void cpg_confchg_callback(cpg_handle_t handle,
+				 struct cpg_name *groupName,
+				 struct cpg_address *member_list, int member_list_entries,
+				 struct cpg_address *left_list, int left_list_entries,
+				 struct cpg_address *joined_list, int 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 cpg_callbacks = {
+	.cpg_deliver_fn =            cpg_deliver_callback,
+	.cpg_confchg_fn =            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);
+
+	client = malloc(sizeof(struct local_client));
+	if (!client)
+	{
+		DEBUGLOG("malloc failed\n");
+		return -1;
+	}
+
+	memset(client, 0, sizeof(struct local_client));
+	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 cpg_deliver_callback (cpg_handle_t handle,
+				  struct cpg_name *groupName,
+				  uint32_t nodeid,
+				  uint32_t pid,
+				  void *msg,
+				  int msg_len)
+{
+	int target_nodeid;
+
+	memcpy(&target_nodeid, msg, OPENAIS_CSID_LEN);
+
+	DEBUGLOG("%u got message from nodeid %d for %d. len %d\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 cpg_confchg_callback(cpg_handle_t handle,
+				 struct cpg_name *groupName,
+				 struct cpg_address *member_list, int member_list_entries,
+				 struct cpg_address *left_list, int left_list_entries,
+				 struct cpg_address *joined_list, int joined_list_entries)
+{
+	int i;
+	struct node_info *ninfo;
+
+	DEBUGLOG("confchg callback. %d joined, %d left, %d 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,
+			     &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, &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)
+{
+	DEBUGLOG("cluster_closedown\n");
+	destroy_lvhash();
+
+	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;
+
+	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;
+	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=%llx\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: %llx\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 = {
+	.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;
+	else
+		return NULL;
+}

daemons/clvmd/clvmd.h

@@ -35,6 +35,8 @@ struct node_reply {
 	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
@@ -76,7 +78,8 @@ struct netsock_bits {
 };
 
 typedef int (*fd_callback_t) (struct local_client * fd, char *buf, int len,
-			      char *csid, struct local_client ** new_client);
+			      const char *csid,
+			      struct local_client ** new_client);
 
 /* One of these for each fd we are listening on */
 struct local_client {
@@ -95,11 +98,7 @@ struct local_client {
 	} bits;
 };
 
-#ifdef DEBUG
-#define DEBUGLOG(fmt, args...) {time_t P; time(&P); fprintf(stderr, "CLVMD[%x]: %.15s ", (int)pthread_self(), ctime(&P)+4 ); fprintf(stderr, fmt, ## args);}
-#else
-#define DEBUGLOG(fmt, args...)
-#endif
+#define DEBUGLOG(fmt, args...) debuglog(fmt, ## args);
 
 #ifndef max
 #define max(a,b) ((a)>(b)?(a):(b))
@@ -116,7 +115,10 @@ 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, char *csid);
+extern void process_message(struct local_client *client, const char *buf,
+			    int len, const char *csid);
+extern void debuglog(const char *fmt, ... )
+  __attribute__ ((format(printf, 1, 2)));
 
 int sync_lock(const char *resource, int mode, int flags, int *lockid);
 int sync_unlock(const char *resource, int lockid);

daemons/clvmd/cnxman-socket.h

@@ -1,226 +0,0 @@
-/******************************************************************************
-*******************************************************************************
-**
-**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
-**  Copyright (C) 2004 Red Hat, Inc.  All rights reserved.
-**
-**  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.
-**
-*******************************************************************************
-******************************************************************************/
-
-/* CMAN socket interface header,
-   may be include by user or kernel code */
-
-#ifndef __CNXMAN_SOCKET_H
-#define __CNXMAN_SOCKET_H
-
-/* A currently unused number. TIPC also uses this number and you're unlikely
-   to be using both.
- */
-#define AF_CLUSTER 30
-#define PF_CLUSTER AF_CLUSTER
-
-/* Protocol(socket) types */
-#define CLPROTO_MASTER 2
-#define CLPROTO_CLIENT 3
-
-/* ioctls -- should register these properly */
-#define SIOCCLUSTER_NOTIFY            _IOW('x', 0x01, int)
-#define SIOCCLUSTER_REMOVENOTIFY      _IO( 'x', 0x02)
-#define SIOCCLUSTER_GETMEMBERS        _IOR('x', 0x03, struct cl_cluster_nodelist)
-#define SIOCCLUSTER_SETEXPECTED_VOTES _IOW('x', 0x04, int)
-#define SIOCCLUSTER_ISQUORATE         _IO( 'x', 0x05)
-#define SIOCCLUSTER_ISLISTENING       _IOW('x', 0x06, struct cl_listen_request)
-#define SIOCCLUSTER_GETALLMEMBERS     _IOR('x', 0x07, struct cl_cluster_nodelist)
-#define SIOCCLUSTER_SET_VOTES         _IOW('x', 0x08, int)
-#define SIOCCLUSTER_GET_VERSION       _IOR('x', 0x09, struct cl_version)
-#define SIOCCLUSTER_SET_VERSION       _IOW('x', 0x0a, struct cl_version)
-#define SIOCCLUSTER_ISACTIVE          _IO( 'x', 0x0b)
-#define SIOCCLUSTER_KILLNODE          _IOW('x', 0x0c, int)
-#define SIOCCLUSTER_GET_JOINCOUNT     _IO( 'x', 0x0d)
-#define SIOCCLUSTER_SERVICE_REGISTER  _IOW('x', 0x0e, char)
-#define SIOCCLUSTER_SERVICE_UNREGISTER _IO('x', 0x0f)
-#define SIOCCLUSTER_SERVICE_JOIN      _IO( 'x', 0x10)
-#define SIOCCLUSTER_SERVICE_LEAVE     _IO( 'x', 0x20)
-#define SIOCCLUSTER_SERVICE_SETSIGNAL _IOW('x', 0x30, int)
-#define SIOCCLUSTER_SERVICE_STARTDONE _IOW('x', 0x40, unsigned int)
-#define SIOCCLUSTER_SERVICE_GETEVENT  _IOR('x', 0x50, struct cl_service_event)
-#define SIOCCLUSTER_SERVICE_GETMEMBERS _IOR('x', 0x60, struct cl_cluster_nodelist)
-#define SIOCCLUSTER_SERVICE_GLOBALID  _IOR('x', 0x70, uint32_t)
-#define SIOCCLUSTER_SERVICE_SETLEVEL  _IOR('x', 0x80, int)
-#define SIOCCLUSTER_GETNODE	      _IOWR('x', 0x90, struct cl_cluster_node)
-#define SIOCCLUSTER_BARRIER           _IOW('x', 0x0a0, struct cl_barrier_info)
-
-/* These were setsockopts */
-#define SIOCCLUSTER_PASS_SOCKET       _IOW('x', 0x0b0, struct cl_passed_sock)
-#define SIOCCLUSTER_SET_NODENAME      _IOW('x', 0x0b1, char *)
-#define SIOCCLUSTER_SET_NODEID        _IOW('x', 0x0b2, int)
-#define SIOCCLUSTER_JOIN_CLUSTER      _IOW('x', 0x0b3, struct cl_join_cluster_info)
-#define SIOCCLUSTER_LEAVE_CLUSTER     _IOW('x', 0x0b4, int)
-
-
-/* Maximum size of a cluster message */
-#define CMAN_MAX_CLUSTER_MESSAGE          1500
-#define CMAN_MAX_CLUSTER_MEMBER_NAME_LEN   255
-#define MAX_BARRIER_NAME_LEN           33
-#define MAX_SA_ADDR_LEN                12
-#define MAX_CLUSTER_NAME_LEN           16
-
-/* Well-known cluster port numbers */
-#define CLUSTER_PORT_MEMBERSHIP  1	/* Mustn't block during cluster
-					 * transitions! */
-#define CLUSTER_PORT_SERVICES    2
-#define CLUSTER_PORT_SYSMAN      10	/* Remote execution daemon */
-#define CLUSTER_PORT_CLVMD       11	/* Cluster LVM daemon */
-#define CLUSTER_PORT_SLM         12	/* LVM SLM (simple lock manager) */
-
-/* Port numbers above this will be blocked when the cluster is inquorate or in
- * transition */
-#define HIGH_PROTECTED_PORT      9
-
-/* Reasons for leaving the cluster */
-#define CLUSTER_LEAVEFLAG_DOWN     0	/* Normal shutdown */
-#define CLUSTER_LEAVEFLAG_KILLED   1
-#define CLUSTER_LEAVEFLAG_PANIC    2
-#define CLUSTER_LEAVEFLAG_REMOVED  3	/* This one can reduce quorum */
-#define CLUSTER_LEAVEFLAG_REJECTED 4	/* Not allowed into the cluster in the
-					 * first place */
-#define CLUSTER_LEAVEFLAG_INCONSISTENT 5	/* Our view of the cluster is
-						 * in a minority */
-#define CLUSTER_LEAVEFLAG_DEAD         6	/* Discovered to be dead */
-#define CLUSTER_LEAVEFLAG_FORCE     0x10	/* Forced by command-line */
-
-/* OOB messages sent to a local socket */
-#define CLUSTER_OOB_MSG_PORTCLOSED  1
-#define CLUSTER_OOB_MSG_STATECHANGE 2
-#define CLUSTER_OOB_MSG_SERVICEEVENT 3
-
-/* Sendmsg flags, these are above the normal sendmsg flags so they don't
- * interfere */
-#define MSG_NOACK     0x010000	/* Don't need an ACK for this message */
-#define MSG_QUEUE     0x020000	/* Queue the message for sending later */
-#define MSG_MULTICAST 0x080000	/* Message was sent to all nodes in the cluster
-				 */
-#define MSG_ALLINT    0x100000	/* Send out of all interfaces */
-#define MSG_REPLYEXP  0x200000	/* Reply is expected */
-
-typedef enum { NODESTATE_REMOTEMEMBER, NODESTATE_JOINING, NODESTATE_MEMBER,
-	       NODESTATE_DEAD } nodestate_t;
-
-
-struct sockaddr_cl {
-	unsigned short scl_family;
-	unsigned char scl_flags;
-	unsigned char scl_port;
-	int           scl_nodeid;
-};
-
-/*
- * This is how we pass the multicast & receive sockets into kernel space.
- */
-struct cl_passed_sock {
-	int fd;			/* FD of master socket to do multicast on */
-	int number;		/* Socket number, to match up recvonly & bcast
-				 * sockets */
-        int multicast;          /* Is it multicast or receive ? */
-};
-
-/* Cluster configuration info passed when we join the cluster */
-struct cl_join_cluster_info {
-	unsigned char votes;
-	unsigned int expected_votes;
-	unsigned int two_node;
-	unsigned int config_version;
-
-        char cluster_name[17];
-};
-
-
-/* This is the structure, per node, returned from the membership ioctl */
-struct cl_cluster_node {
-	unsigned int size;
-	unsigned int node_id;
-	unsigned int us;
-	unsigned int leave_reason;
-	unsigned int incarnation;
-	nodestate_t state;
-	char name[CMAN_MAX_CLUSTER_MEMBER_NAME_LEN];
-	unsigned char votes;
-};
-
-/* The struct passed to the membership ioctls */
-struct cl_cluster_nodelist {
-        uint32_t max_members;
-        struct cl_cluster_node *nodes;
-};
-
-/* Structure passed to SIOCCLUSTER_ISLISTENING */
-struct cl_listen_request {
-	unsigned char port;
-        int           nodeid;
-};
-
-/* A Cluster PORTCLOSED message - received by a local user as an OOB message */
-struct cl_portclosed_oob {
-	unsigned char cmd;	/* CLUSTER_OOB_MSG_PORTCLOSED */
-	unsigned char port;
-};
-
-/* Get all version numbers or set the config version */
-struct cl_version {
-	unsigned int major;
-	unsigned int minor;
-	unsigned int patch;
-	unsigned int config;
-};
-
-/* structure passed to barrier ioctls */
-struct cl_barrier_info {
-	char cmd;
-	char name[MAX_BARRIER_NAME_LEN];
-	unsigned int flags;
-	unsigned long arg;
-};
-
-typedef enum { SERVICE_EVENT_STOP, SERVICE_EVENT_START, SERVICE_EVENT_FINISH,
-		SERVICE_EVENT_LEAVEDONE } service_event_t;
-
-typedef enum { SERVICE_START_FAILED, SERVICE_START_JOIN, SERVICE_START_LEAVE }
-		service_start_t;
-
-struct cl_service_event {
-	service_event_t type;
-	service_start_t start_type;
-	unsigned int event_id;
-	unsigned int last_stop;
-	unsigned int last_start;
-	unsigned int last_finish;
-	unsigned int node_count;
-};
-
-
-/* Commands to the barrier ioctl */
-#define BARRIER_IOCTL_REGISTER 1
-#define BARRIER_IOCTL_CHANGE   2
-#define BARRIER_IOCTL_DELETE   3
-#define BARRIER_IOCTL_WAIT     4
-
-/* Attributes of a barrier - bitmask */
-#define BARRIER_ATTR_AUTODELETE 1
-#define BARRIER_ATTR_MULTISTEP  2
-#define BARRIER_ATTR_MANUAL     4
-#define BARRIER_ATTR_ENABLED    8
-#define BARRIER_ATTR_CALLBACK  16
-
-/* Attribute setting commands */
-#define BARRIER_SETATTR_AUTODELETE 1
-#define BARRIER_SETATTR_MULTISTEP  2
-#define BARRIER_SETATTR_ENABLED    3
-#define BARRIER_SETATTR_NODES      4
-#define BARRIER_SETATTR_CALLBACK   5
-#define BARRIER_SETATTR_TIMEOUT    6
-
-#endif

daemons/clvmd/lvm-functions.c

@@ -1,6 +1,6 @@
 /*
  * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2009 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
@@ -13,6 +13,10 @@
  * Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include <configure.h>
 #include <pthread.h>
 #include <sys/types.h>
 #include <sys/utsname.h>
@@ -30,11 +34,10 @@
 #include <errno.h>
 #include <syslog.h>
 #include <assert.h>
+#include <libdevmapper.h>
+#include <libdlm.h>
 
-#include "libdevmapper.h"
-#include "list.h"
 #include "lvm-types.h"
-#include "libdlm.h"
 #include "clvm.h"
 #include "clvmd-comms.h"
 #include "clvmd.h"
@@ -42,45 +45,192 @@
 
 /* LVM2 headers */
 #include "toolcontext.h"
-#include "log.h"
+#include "lvmcache.h"
+#include "lvm-logging.h"
+#include "lvm-globals.h"
 #include "activate.h"
 #include "locking.h"
+#include "archiver.h"
+#include "defaults.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];
+static int suspended = 0;
 
 struct lv_info {
 	int lock_id;
 	int lock_mode;
 };
 
-/* Return the mode a lock is currently held at (or -1 if not held) */
-static int get_current_lock(char *resource)
+#define LCK_MASK (LCK_TYPE_MASK | LCK_SCOPE_MASK)
+
+static const char *decode_locking_cmd(unsigned char 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"; 
+		break;
+	case LCK_LV: 
+		scope = "LV"; 
+		break;
+	default:
+		scope = "unknown";
+		break;
+	}
+
+	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;
+	}
+
+	sprintf(buf, "0x%x %s (%s|%s%s%s%s%s%s)", cmdl, command, type, scope,
+		cmdl & LCK_NONBLOCK   ? "|NONBLOCK" : "",
+		cmdl & LCK_HOLD       ? "|HOLD" : "",
+		cmdl & LCK_LOCAL      ? "|LOCAL" : "",
+		cmdl & LCK_CLUSTER_VG ? "|CLUSTER_VG" : "",
+		cmdl & LCK_CACHE      ? "|CACHE" : "");
+
+	return buf;
+}
+
+static const char *decode_flags(unsigned char flags)
+{
+	static char buf[128];
+
+	sprintf(buf, "0x%x (%s%s%s)", flags,
+		flags & LCK_PARTIAL_MODE	  ? "PARTIAL_MODE " : "",
+		flags & LCK_MIRROR_NOSYNC_MODE	  ? "MIRROR_NOSYNC " : "",
+		flags & LCK_DMEVENTD_MONITOR_MODE ? "DMEVENTD_MONITOR " : "");
+
+	return buf;
+}
+
+char *get_last_lvm_error()
+{
+	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);
-	if (lvi) {
+
+	return lvi;
+}
+
+static void insert_info(const char *resource, struct lv_info *lvi)
+{
+	pthread_mutex_lock(&lv_hash_lock);
+	dm_hash_insert(lv_hash, resource, lvi);
+	pthread_mutex_unlock(&lv_hash_lock);
+}
+
+static void remove_info(const char *resource)
+{
+	pthread_mutex_lock(&lv_hash_lock);
+	dm_hash_remove(lv_hash, resource);
+	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;
-	} else {
-		return -1;
-	}
+
+	return -1;
+}
+
+
+void init_lvhash()
+{
+	/* Create hash table for keeping LV locks & status */
+	lv_hash = dm_hash_create(100);
+	pthread_mutex_init(&lv_hash_lock, NULL);
+	pthread_mutex_init(&lvm_lock, NULL);
 }
 
 /* Called at shutdown to tidy the lockspace */
-void unlock_all()
+void destroy_lvhash()
 {
 	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) {
-		struct lv_info *lvi = dm_hash_get_data(lv_hash, v);
 
-		sync_unlock(dm_hash_get_key(lv_hash, v), lvi->lock_id);
+	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));
+		free(lvi);
 	}
+
+	dm_hash_destroy(lv_hash);
+	lv_hash = NULL;
+
 	pthread_mutex_unlock(&lv_hash_lock);
 }
 
@@ -93,10 +243,7 @@ int hold_lock(char *resource, int mode, int flags)
 
 	flags &= LKF_NOQUEUE;	/* Only LKF_NOQUEUE is valid here */
 
-	pthread_mutex_lock(&lv_hash_lock);
-	lvi = dm_hash_lookup(lv_hash, resource);
-	pthread_mutex_unlock(&lv_hash_lock);
-	if (lvi) {
+	if ((lvi = lookup_info(resource))) {
 		/* Already exists - convert it */
 		status =
 		    sync_lock(resource, mode, LKF_CONVERT | flags,
@@ -122,11 +269,9 @@ int hold_lock(char *resource, int mode, int flags)
 			free(lvi);
 			DEBUGLOG("hold_lock. lock at %d failed: %s\n", mode,
 				 strerror(errno));
-		} else {
-		        pthread_mutex_lock(&lv_hash_lock);
-			dm_hash_insert(lv_hash, resource, lvi);
-			pthread_mutex_unlock(&lv_hash_lock);
-		}
+		} else
+			insert_info(resource, lvi);
+
 		errno = saved_errno;
 	}
 	return status;
@@ -139,10 +284,7 @@ int hold_unlock(char *resource)
 	int status;
 	int saved_errno;
 
-	pthread_mutex_lock(&lv_hash_lock);
-	lvi = dm_hash_lookup(lv_hash, resource);
-	pthread_mutex_unlock(&lv_hash_lock);
-	if (!lvi) {
+	if (!(lvi = lookup_info(resource))) {
 		DEBUGLOG("hold_unlock, lock not already held\n");
 		return 0;
 	}
@@ -150,9 +292,7 @@ int hold_unlock(char *resource)
 	status = sync_unlock(resource, lvi->lock_id);
 	saved_errno = errno;
 	if (!status) {
-	    	pthread_mutex_lock(&lv_hash_lock);
-		dm_hash_remove(lv_hash, resource);
-		pthread_mutex_unlock(&lv_hash_lock);
+		remove_info(resource);
 		free(lvi);
 	} else {
 		DEBUGLOG("hold_unlock. unlock failed(%d): %s\n", status,
@@ -200,12 +340,21 @@ static int do_activate_lv(char *resource, unsigned char lock_flags, int mode)
 	/* Try to get the lock if it's a clustered volume group */
 	if (lock_flags & LCK_CLUSTER_VG) {
 		status = hold_lock(resource, mode, LKF_NOQUEUE);
-		if (status)
+		if (status) {
+			/* Return an LVM-sensible error for this.
+			 * Forcing EIO makes the upper level return this text
+			 * rather than the strerror text for EAGAIN.
+			 */
+			if (errno == EAGAIN) {
+				sprintf(last_error, "Volume is busy on another node");
+				errno = EIO;
+			}
 			return errno;
+		}
 	}
 
 	/* If it's suspended then resume it */
-	if (!lv_info_by_lvid(cmd, resource, &lvi, 0))
+	if (!lv_info_by_lvid(cmd, resource, &lvi, 0, 0))
 		return EIO;
 
 	if (lvi.suspended)
@@ -227,7 +376,7 @@ static int do_resume_lv(char *resource)
 	/* Is it open ? */
 	oldmode = get_current_lock(resource);
 	if (oldmode == -1) {
-		DEBUGLOG("do_deactivate_lock, lock not already held\n");
+		DEBUGLOG("do_resume_lv, lock not already held\n");
 		return 0;	/* We don't need to do anything */
 	}
 
@@ -251,7 +400,7 @@ static int do_suspend_lv(char *resource)
 	}
 
 	/* Only suspend it if it exists */
-	if (!lv_info_by_lvid(cmd, resource, &lvi, 0))
+	if (!lv_info_by_lvid(cmd, resource, &lvi, 0, 0))
 		return EIO;
 
 	if (lvi.exists) {
@@ -286,23 +435,52 @@ static int do_deactivate_lv(char *resource, unsigned char lock_flags)
 	return 0;
 }
 
+const char *do_lock_query(char *resource)
+{
+	int mode;
+	const char *type = NULL;
+
+	mode = get_current_lock(resource);
+	switch (mode) {
+		case LKM_NLMODE: type = "NL"; break;
+		case LKM_CRMODE: type = "CR"; break;
+		case LKM_CWMODE: type = "CW"; break;
+		case LKM_PRMODE: type = "PR"; break;
+		case LKM_PWMODE: type = "PW"; break;
+		case LKM_EXMODE: type = "EX"; break;
+	}
+
+	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 = 0x%x, flags = %x\n",
-		 resource, command, lock_flags);
+	DEBUGLOG("do_lock_lv: resource '%s', cmd = %s, flags = %s\n",
+		 resource, decode_locking_cmd(command), decode_flags(lock_flags));
 
 	if (!cmd->config_valid || config_files_changed(cmd)) {
 		/* Reinitialise various settings inc. logging, filters */
-		if (!refresh_toolcontext(cmd)) {
+		if (do_refresh_cache()) {
 			log_error("Updated config file invalid. Aborting.");
 			return EINVAL;
 		}
 	}
 
+	pthread_mutex_lock(&lvm_lock);
+	if (lock_flags & LCK_MIRROR_NOSYNC_MODE)
+		init_mirror_in_sync(1);
+
+	if (!(lock_flags & LCK_DMEVENTD_MONITOR_MODE))
+		init_dmeventd_monitor(0);
+
+	cmd->partial_activation = (lock_flags & LCK_PARTIAL_MODE) ? 1 : 0;
+
 	switch (command) {
 	case LCK_LV_EXCLUSIVE:
 		status = do_activate_lv(resource, lock_flags, LKM_EXMODE);
@@ -310,11 +488,15 @@ int do_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
 
 	case LCK_LV_SUSPEND:
 		status = do_suspend_lv(resource);
+		if (!status)
+			suspended++;
 		break;
 
 	case LCK_UNLOCK:
 	case LCK_LV_RESUME:	/* if active */
 		status = do_resume_lv(resource);
+		if (!status)
+			suspended--;
 		break;
 
 	case LCK_LV_ACTIVATE:
@@ -331,8 +513,17 @@ int do_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
 		break;
 	}
 
+	if (lock_flags & LCK_MIRROR_NOSYNC_MODE)
+		init_mirror_in_sync(0);
+
+	if (!(lock_flags & LCK_DMEVENTD_MONITOR_MODE))
+		init_dmeventd_monitor(DEFAULT_DMEVENTD_MONITOR);
+
+	cmd->partial_activation = 0;
+
 	/* clean the pool for another command */
 	dm_pool_empty(cmd->mem);
+	pthread_mutex_unlock(&lvm_lock);
 
 	DEBUGLOG("Command return is %d\n", status);
 	return status;
@@ -346,8 +537,8 @@ int pre_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
 	   before suspending cluster-wide.
 	 */
 	if (command == LCK_LV_SUSPEND) {
-		DEBUGLOG("pre_lock_lv: resource '%s', cmd = 0x%x, flags = %d\n",
-			 resource, command, lock_flags);
+		DEBUGLOG("pre_lock_lv: resource '%s', cmd = %s, flags = %s\n",
+			 resource, decode_locking_cmd(command), decode_flags(lock_flags));
 
 		if (hold_lock(resource, LKM_PWMODE, LKF_NOQUEUE))
 			return errno;
@@ -359,20 +550,25 @@ int pre_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
 int post_lock_lv(unsigned char command, unsigned char lock_flags,
 		 char *resource)
 {
+	int status;
+
 	/* Opposite of above, done on resume after a metadata update */
 	if (command == LCK_LV_RESUME) {
 		int oldmode;
 
 		DEBUGLOG
-		    ("post_lock_lv: resource '%s', cmd = 0x%x, flags = %d\n",
-		     resource, command, lock_flags);
+		    ("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 == LKM_PWMODE) {
 			struct lvinfo lvi;
 
-			if (!lv_info_by_lvid(cmd, resource, &lvi, 0))
+			pthread_mutex_lock(&lvm_lock);
+			status = lv_info_by_lvid(cmd, resource, &lvi, 0, 0);
+			pthread_mutex_unlock(&lvm_lock);
+			if (!status)
 				return EIO;
 
 			if (lvi.exists) {
@@ -387,8 +583,8 @@ int post_lock_lv(unsigned char command, unsigned char lock_flags,
 	return 0;
 }
 
-/* Check if a VG is un use by LVM1 so we don't stomp on it */
-int do_check_lvm1(char *vgname)
+/* 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;
 
@@ -397,6 +593,24 @@ int do_check_lvm1(char *vgname)
 	return status == 1 ? 0 : EBUSY;
 }
 
+int do_refresh_cache()
+{
+	int ret;
+	DEBUGLOG("Refreshing context\n");
+	log_notice("Refreshing context");
+
+	pthread_mutex_lock(&lvm_lock);
+
+	ret = refresh_toolcontext(cmd);
+	init_full_scan_done(0);
+	lvmcache_label_scan(cmd, 2);
+	dm_pool_empty(cmd->mem);
+
+	pthread_mutex_unlock(&lvm_lock);
+
+	return ret==1?0:-1;
+}
+
 
 /* Only called at gulm startup. Drop any leftover VG or P_orphan locks
    that might be hanging around if we died for any reason
@@ -407,9 +621,10 @@ static void drop_vg_locks()
 	char line[255];
 	FILE *vgs =
 	    popen
-	    ("lvm pvs --nolocking --noheadings -o vg_name", "r");
+	    ("lvm pvs  --config 'log{command_names=0 prefix=\"\"}' --nolocking --noheadings -o vg_name", "r");
 
-	sync_unlock("P_orphans", LCK_EXCL);
+	sync_unlock("P_" VG_ORPHANS, LCK_EXCL);
+	sync_unlock("P_" VG_GLOBAL, LCK_EXCL);
 
 	if (!vgs)
 		return;
@@ -432,7 +647,19 @@ static void drop_vg_locks()
 		sync_unlock(vg, LCK_EXCL);
 
 	}
-	fclose(vgs);
+	if (fclose(vgs))
+		DEBUGLOG("vgs fclose failed: %s\n", strerror(errno));
+}
+
+/*
+ * Drop lvmcache metadata
+ */
+void drop_metadata(const char *vgname)
+{
+	DEBUGLOG("Dropping metadata for VG %s\n", vgname);
+	pthread_mutex_lock(&lvm_lock);
+	lvmcache_drop_metadata(vgname);
+	pthread_mutex_unlock(&lvm_lock);
 }
 
 /*
@@ -447,7 +674,7 @@ static void *get_initial_state()
 	char line[255];
 	FILE *lvs =
 	    popen
-	    ("lvm lvs --nolocking --noheadings -o vg_uuid,lv_uuid,lv_attr,vg_attr",
+	    ("lvm lvs  --config 'log{command_names=0 prefix=\"\"}' --nolocking --noheadings -o vg_uuid,lv_uuid,lv_attr,vg_attr",
 	     "r");
 
 	if (!lvs)
@@ -482,16 +709,40 @@ static void *get_initial_state()
 			}
 		}
 	}
-	fclose(lvs);
+	if (fclose(lvs))
+		DEBUGLOG("lvs fclose failed: %s\n", strerror(errno));
 	return NULL;
 }
 
+static void lvm2_log_fn(int level, const char *file, int line,
+			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, "%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()
 {
 	int locking_type;
 
-	locking_type = find_config_int(cmd->cft->root, "global/locking_type", 1);
+	locking_type = find_config_tree_int(cmd, "global/locking_type", 1);
 
 	if (locking_type == 3) /* compiled-in cluster support */
 		return;
@@ -499,7 +750,7 @@ static void check_config()
 	if (locking_type == 2) { /* External library, check name */
 		const char *libname;
 
-		libname = find_config_str(cmd->cft->root, "global/locking_library",
+		libname = find_config_tree_str(cmd, "global/locking_library",
 					  "");
 		if (strstr(libname, "liblvm2clusterlock.so"))
 			return;
@@ -510,24 +761,41 @@ static void check_config()
 	log_error("locking_type not set correctly in lvm.conf, cluster operations will not work.");
 }
 
-void init_lvhash()
+/* Backups up the LVM metadata if it's changed */
+void lvm_do_backup(const char *vgname)
 {
-	/* Create hash table for keeping LV locks & status */
-	lv_hash = dm_hash_create(100);
-	pthread_mutex_init(&lv_hash_lock, NULL);
+	struct volume_group * vg;
+	int consistent = 0;
+
+	DEBUGLOG("Triggering backup of VG metadata for %s. suspended=%d\n", vgname, suspended);
+
+	pthread_mutex_lock(&lvm_lock);
+
+	vg = vg_read_internal(cmd, vgname, NULL /*vgid*/, &consistent);
+
+	if (vg && consistent)
+		check_current_backup(vg);
+	else
+		log_error("Error backing up metadata, can't find VG for group %s", vgname);
+
+	vg_release(vg);
+	dm_pool_empty(cmd->mem);
+
+	pthread_mutex_unlock(&lvm_lock);
 }
 
 /* Called to initialise the LVM context of the daemon */
 int init_lvm(int using_gulm)
 {
-	if (!(cmd = create_toolcontext(NULL))) {
+	if (!(cmd = create_toolcontext(1, NULL))) {
 		log_error("Failed to allocate command context");
 		return 0;
 	}
 
 	/* Use LOG_DAEMON for syslog messages instead of LOG_USER */
 	init_syslog(LOG_DAEMON);
-	init_debug(_LOG_ERR);
+	openlog("clvmd", LOG_PID, LOG_DAEMON);
+	cmd->cmd_line = (char *)"clvmd";
 
 	/* Check lvm.conf is setup for cluster-LVM */
 	check_config();
@@ -538,5 +806,15 @@ int init_lvm(int using_gulm)
 
 	get_initial_state();
 
+	/* Trap log messages so we can pass them back to the user */
+	init_log_fn(lvm2_log_fn);
+
 	return 1;
 }
+
+void destroy_lvm(void)
+{
+	if (cmd)
+		destroy_toolcontext(cmd);
+	cmd = NULL;
+}

daemons/clvmd/lvm-functions.h

@@ -22,14 +22,19 @@ 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(char *vgname);
+extern int do_check_lvm1(const char *vgname);
+extern int do_refresh_cache(void);
 extern int init_lvm(int using_gulm);
+extern void destroy_lvm(void);
 extern void init_lvhash(void);
-
+extern void destroy_lvhash(void);
+extern void lvm_do_backup(const char *vgname);
 extern int hold_unlock(char *resource);
 extern int hold_lock(char *resource, int mode, int flags);
-extern void unlock_all(void);
+extern char *get_last_lvm_error(void);
+extern void drop_metadata(const char *vgname);
 
 #endif

daemons/clvmd/refresh_clvmd.c

@@ -0,0 +1,368 @@
+/*
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+/*
+ * Tell all clvmds in a cluster to refresh their toolcontext
+ *
+ */
+
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
+
+#include <configure.h>
+#include <stddef.h>
+#include <sys/socket.h>
+#include <sys/un.h>
+#include <errno.h>
+#include <unistd.h>
+#include <libdevmapper.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <limits.h>
+
+#include "clvm.h"
+#include "refresh_clvmd.h"
+
+typedef struct lvm_response {
+	char node[255];
+	char *response;
+	int status;
+	int len;
+} lvm_response_t;
+
+/*
+ * This gets stuck at the start of memory we allocate so we
+ * can sanity-check it at deallocation time
+ */
+#define LVM_SIGNATURE 0x434C564D
+
+static int _clvmd_sock = -1;
+
+/* Open connection to the clvm daemon */
+static int _open_local_sock(void)
+{
+	int local_socket;
+	struct sockaddr_un sockaddr;
+
+	/* Open local socket */
+	if ((local_socket = socket(PF_UNIX, SOCK_STREAM, 0)) < 0) {
+		fprintf(stderr, "Local socket creation failed: %s", strerror(errno));
+		return -1;
+	}
+
+	memset(&sockaddr, 0, sizeof(sockaddr));
+	memcpy(sockaddr.sun_path, CLVMD_SOCKNAME, sizeof(CLVMD_SOCKNAME));
+
+	sockaddr.sun_family = AF_UNIX;
+
+	if (connect(local_socket,(struct sockaddr *) &sockaddr,
+		    sizeof(sockaddr))) {
+		int saved_errno = errno;
+
+		fprintf(stderr, "connect() failed on local socket: %s\n",
+			  strerror(errno));
+		if (close(local_socket))
+			return -1;
+
+		errno = saved_errno;
+		return -1;
+	}
+
+	return local_socket;
+}
+
+/* Send a request and return the status */
+static int _send_request(const char *inbuf, int inlen, char **retbuf)
+{
+	char outbuf[PIPE_BUF];
+	struct clvm_header *outheader = (struct clvm_header *) outbuf;
+	int len;
+	int off;
+	int buflen;
+	int err;
+
+	/* Send it to CLVMD */
+ rewrite:
+	if ( (err = write(_clvmd_sock, inbuf, inlen)) != inlen) {
+	        if (err == -1 && errno == EINTR)
+		        goto rewrite;
+		fprintf(stderr, "Error writing data to clvmd: %s", strerror(errno));
+		return 0;
+	}
+
+	/* Get the response */
+ reread:
+	if ((len = read(_clvmd_sock, outbuf, sizeof(struct clvm_header))) < 0) {
+	        if (errno == EINTR)
+		        goto reread;
+		fprintf(stderr, "Error reading data from clvmd: %s", strerror(errno));
+		return 0;
+	}
+
+	if (len == 0) {
+		fprintf(stderr, "EOF reading CLVMD");
+		errno = ENOTCONN;
+		return 0;
+	}
+
+	/* Allocate buffer */
+	buflen = len + outheader->arglen;
+	*retbuf = dm_malloc(buflen);
+	if (!*retbuf) {
+		errno = ENOMEM;
+		return 0;
+	}
+
+	/* Copy the header */
+	memcpy(*retbuf, outbuf, len);
+	outheader = (struct clvm_header *) *retbuf;
+
+	/* Read the returned values */
+	off = 1;		/* we've already read the first byte */
+	while (off <= outheader->arglen && len > 0) {
+		len = read(_clvmd_sock, outheader->args + off,
+			   buflen - off - offsetof(struct clvm_header, args));
+		if (len > 0)
+			off += len;
+	}
+
+	/* Was it an error ? */
+	if (outheader->status != 0) {
+		errno = outheader->status;
+
+		/* Only return an error here if there are no node-specific
+		   errors present in the message that might have more detail */
+		if (!(outheader->flags & CLVMD_FLAG_NODEERRS)) {
+			fprintf(stderr, "cluster request failed: %s\n", strerror(errno));
+			return 0;
+		}
+
+	}
+
+	return 1;
+}
+
+/* Build the structure header and parse-out wildcard node names */
+static void _build_header(struct clvm_header *head, int cmd, const char *node,
+			  int len)
+{
+	head->cmd = cmd;
+	head->status = 0;
+	head->flags = 0;
+	head->clientid = 0;
+	head->arglen = len;
+
+	if (node) {
+		/*
+		 * Allow a couple of special node names:
+		 * "*" for all nodes,
+		 * "." for the local node only
+		 */
+		if (strcmp(node, "*") == 0) {
+			head->node[0] = '\0';
+		} else if (strcmp(node, ".") == 0) {
+			head->node[0] = '\0';
+			head->flags = CLVMD_FLAG_LOCAL;
+		} else
+			strcpy(head->node, node);
+	} else
+		head->node[0] = '\0';
+}
+
+/*
+ * Send a message to a(or all) node(s) in the cluster and wait for replies
+ */
+static int _cluster_request(char cmd, const char *node, void *data, int len,
+			   lvm_response_t ** response, int *num)
+{
+	char outbuf[sizeof(struct clvm_header) + len + strlen(node) + 1];
+	char *inptr;
+	char *retbuf = NULL;
+	int status;
+	int i;
+	int num_responses = 0;
+	struct clvm_header *head = (struct clvm_header *) outbuf;
+	lvm_response_t *rarray;
+
+	*num = 0;
+
+	if (_clvmd_sock == -1)
+		_clvmd_sock = _open_local_sock();
+
+	if (_clvmd_sock == -1)
+		return 0;
+
+	_build_header(head, cmd, node, len);
+	memcpy(head->node + strlen(head->node) + 1, data, len);
+
+	status = _send_request(outbuf, sizeof(struct clvm_header) +
+			      strlen(head->node) + len, &retbuf);
+	if (!status)
+		goto out;
+
+	/* Count the number of responses we got */
+	head = (struct clvm_header *) retbuf;
+	inptr = head->args;
+	while (inptr[0]) {
+		num_responses++;
+		inptr += strlen(inptr) + 1;
+		inptr += sizeof(int);
+		inptr += strlen(inptr) + 1;
+	}
+
+	/*
+	 * Allocate response array.
+	 * With an extra pair of INTs on the front to sanity
+	 * check the pointer when we are given it back to free
+	 */
+	*response = dm_malloc(sizeof(lvm_response_t) * num_responses +
+			    sizeof(int) * 2);
+	if (!*response) {
+		errno = ENOMEM;
+		status = 0;
+		goto out;
+	}
+
+	rarray = *response;
+
+	/* Unpack the response into an lvm_response_t array */
+	inptr = head->args;
+	i = 0;
+	while (inptr[0]) {
+		strcpy(rarray[i].node, inptr);
+		inptr += strlen(inptr) + 1;
+
+		memcpy(&rarray[i].status, inptr, sizeof(int));
+		inptr += sizeof(int);
+
+		rarray[i].response = dm_malloc(strlen(inptr) + 1);
+		if (rarray[i].response == NULL) {
+			/* Free up everything else and return error */
+			int j;
+			for (j = 0; j < i; j++)
+				dm_free(rarray[i].response);
+			free(*response);
+			errno = ENOMEM;
+			status = -1;
+			goto out;
+		}
+
+		strcpy(rarray[i].response, inptr);
+		rarray[i].len = strlen(inptr);
+		inptr += strlen(inptr) + 1;
+		i++;
+	}
+	*num = num_responses;
+	*response = rarray;
+
+      out:
+	if (retbuf)
+		dm_free(retbuf);
+
+	return status;
+}
+
+/* Free reply array */
+static int _cluster_free_request(lvm_response_t * response, int num)
+{
+	int i;
+
+	for (i = 0; i < num; i++) {
+		dm_free(response[i].response);
+	}
+
+	dm_free(response);
+
+	return 1;
+}
+
+int refresh_clvmd()
+{
+	int num_responses;
+	char args[1]; // No args really.
+	lvm_response_t *response;
+	int saved_errno;
+	int status;
+	int i;
+
+	status = _cluster_request(CLVMD_CMD_REFRESH, "*", args, 0, &response, &num_responses);
+
+	/* If any nodes were down then display them and return an error */
+	for (i = 0; i < num_responses; i++) {
+		if (response[i].status == EHOSTDOWN) {
+			fprintf(stderr, "clvmd not running on node %s",
+				  response[i].node);
+			status = 0;
+			errno = response[i].status;
+		} else if (response[i].status) {
+			fprintf(stderr, "Error resetting node %s: %s",
+				  response[i].node,
+				  response[i].response[0] ?
+				  	response[i].response :
+				  	strerror(response[i].status));
+			status = 0;
+			errno = response[i].status;
+		}
+	}
+
+	saved_errno = errno;
+	_cluster_free_request(response, num_responses);
+	errno = saved_errno;
+
+	return status;
+}
+
+int debug_clvmd(int level, int clusterwide)
+{
+	int num_responses;
+	char args[1];
+	const char *nodes;
+	lvm_response_t *response;
+	int saved_errno;
+	int status;
+	int i;
+
+	args[0] = level;
+	if (clusterwide)
+		nodes = "*";
+	else
+		nodes = ".";
+
+	status = _cluster_request(CLVMD_CMD_SET_DEBUG, nodes, args, 1, &response, &num_responses);
+
+	/* If any nodes were down then display them and return an error */
+	for (i = 0; i < num_responses; i++) {
+		if (response[i].status == EHOSTDOWN) {
+			fprintf(stderr, "clvmd not running on node %s",
+				  response[i].node);
+			status = 0;
+			errno = response[i].status;
+		} else if (response[i].status) {
+			fprintf(stderr, "Error 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;
+}

daemons/clvmd/refresh_clvmd.h

@@ -1,6 +1,5 @@
 /*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2007 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
@@ -13,13 +12,7 @@
  * Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
-#ifndef _LVM_MATCHER_H
-#define _LVM_MATCHER_H
 
-struct matcher;
-struct matcher *matcher_create(struct dm_pool *mem, const char **patterns,
-			       unsigned num);
+int refresh_clvmd(void);
+int debug_clvmd(int level, int clusterwide);
 
-int matcher_run(struct matcher *m, const char *begin);
-
-#endif

daemons/clvmd/system-lv.c

@@ -1,371 +0,0 @@
-/*
- * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- */
-
-/* Routines dealing with the System LV */
-
-#include <pthread.h>
-#include <sys/types.h>
-#include <sys/stat.h>
-#include <sys/socket.h>
-#include <sys/uio.h>
-#include <sys/un.h>
-#include <sys/time.h>
-#include <sys/ioctl.h>
-#include <sys/mount.h>
-#include <sys/utsname.h>
-#include <syslog.h>
-#include <netinet/in.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <stddef.h>
-#include <signal.h>
-#include <unistd.h>
-#include <fcntl.h>
-#include <dirent.h>
-#include <errno.h>
-#include <mntent.h>
-
-#include "libdlm.h"
-#include "log.h"
-#include "list.h"
-#include "locking.h"
-#include "system-lv.h"
-#include "clvm.h"
-#include "clvmd-comms.h"
-#include "clvmd.h"
-#ifdef HAVE_CCS
-#include "ccs.h"
-#endif
-
-#define SYSTEM_LV_FILESYSTEM "ext2"
-#define SYSTEM_LV_MOUNTPOINT "/tmp/.clvmd-XXXXXX"
-
-extern char *config_filename(void);
-
-static char system_lv_name[PATH_MAX] = { '\0' };
-static char mount_point[PATH_MAX] = { '\0' };
-static int mounted = 0;
-static int mounted_rw = 0;
-static int lockid;
-static const char *lock_name = "CLVM_SYSTEM_LV";
-
-/* Look in /proc/mounts or (as a last resort) /etc/mtab to
-   see if the system-lv is mounted. If it is mounted and we
-   think it's not then abort because we don't have the right
-   lock status and we don't know what other processes are doing with it.
-
-   Returns 1 for mounted, 0 for not mounted so it matches the condition
-   of the "mounted" static variable above.
-*/
-static int is_really_mounted(void)
-{
-	FILE *mountfile;
-	struct mntent *ment;
-
-	mountfile = setmntent("/proc/mounts", "r");
-	if (!mountfile) {
-		mountfile = setmntent("/etc/mtab", "r");
-		if (!mountfile) {
-			log_error("Unable to open /proc/mounts or /etc/mtab");
-			return -1;
-		}
-	}
-
-	/* Look for system LV name in the file */
-	do {
-		ment = getmntent(mountfile);
-		if (ment) {
-			if (strcmp(ment->mnt_fsname, system_lv_name) == 0) {
-				endmntent(mountfile);
-				return 1;
-			}
-		}
-	}
-	while (ment);
-
-	endmntent(mountfile);
-	return 0;
-}
-
-/* Get the system LV name from the config file */
-static int find_system_lv(void)
-{
-	if (system_lv_name[0] == '\0') {
-#ifdef HAVE_CCS
-		int error;
-		ccs_node_t *ctree;
-
-		/* Read the cluster config file */
-		/* Open the config file */
-		error = open_ccs_file(&ctree, "clvm.ccs");
-		if (error) {
-			perror("reading config file");
-			return -1;
-		}
-
-		strcpy(system_lv_name, find_ccs_str(ctree,
-						    "cluster/systemlv", '/',
-						    "/dev/vg/system_lv"));
-
-		/* Finished with config file */
-		close_ccs_file(ctree);
-#else
-		if (getenv("CLVMD_SYSTEM_LV"))
-			strcpy(system_lv_name, getenv("CLVMD_SYSTEM_LV"));
-		else
-			return -1;
-#endif
-	}
-
-	/* See if it has been mounted outside our control */
-	if (is_really_mounted() != mounted) {
-		log_error
-		    ("The system LV state has been mounted/umounted outside the control of clvmd\n"
-		     "it cannot not be used for cluster communications until this is fixed.\n");
-		return -1;
-	}
-	return 0;
-}
-
-/* No prizes */
-int system_lv_umount(void)
-{
-	if (!mounted)
-		return 0;
-
-	if (umount(mount_point) < 0) {
-		log_error("umount of system LV (%s) failed: %m\n",
-			  system_lv_name);
-		return -1;
-	}
-
-	sync_unlock(lock_name, lockid);
-	mounted = 0;
-
-	/* Remove the mount point */
-	rmdir(mount_point);
-
-	return 0;
-}
-
-int system_lv_mount(int readwrite)
-{
-	int status;
-	int saved_errno;
-	int fd;
-
-	if (find_system_lv()) {
-		errno = EBUSY;
-		return -1;
-	}
-
-	/* Is it already mounted suitably? */
-	if (mounted) {
-		if (!readwrite || (readwrite && mounted_rw)) {
-			return 0;
-		} else {
-			/* Mounted RO and we need RW */
-			if (system_lv_umount() < 0)
-				return -1;
-		}
-	}
-
-	/* Randomize the mount point */
-	strcpy(mount_point, SYSTEM_LV_MOUNTPOINT);
-	fd = mkstemp(mount_point);
-	if (fd < 0) {
-		log_error("mkstemp for system LV mount point failed: %m\n");
-		return -1;
-	}
-
-	/* Race condition here but there's no mkstemp for directories */
-	close(fd);
-	unlink(mount_point);
-	mkdir(mount_point, 0600);
-
-	/* Make sure we have a system-lv lock */
-	status =
-	    sync_lock(lock_name, (readwrite) ? LKM_EXMODE : LKM_CRMODE, 0,
-		      &lockid);
-	if (status < 0)
-		return -1;
-
-	/* Mount it */
-	if (mount(system_lv_name, mount_point, SYSTEM_LV_FILESYSTEM,
-		  MS_MGC_VAL | MS_NOSUID | MS_NODEV | MS_NOEXEC | MS_SYNCHRONOUS
-		  | (readwrite ? 0 : MS_RDONLY), NULL) < 0) {
-		/* mount(2) returns EINVAL if the volume has no FS on it. So, if we want to
-		   write to it we try to make a filesystem in it and retry the mount */
-		if (errno == EINVAL && readwrite) {
-			char cmd[256];
-
-			log_error("Attempting mkfs on system LV device %s\n",
-				  system_lv_name);
-			snprintf(cmd, sizeof(cmd), "/sbin/mkfs -t %s %s",
-				 SYSTEM_LV_FILESYSTEM, system_lv_name);
-			system(cmd);
-
-			if (mount
-			    (system_lv_name, mount_point, SYSTEM_LV_FILESYSTEM,
-			     MS_MGC_VAL | MS_NOSUID | MS_NODEV | MS_NOEXEC |
-			     MS_SYNCHRONOUS | (readwrite ? 0 : MS_RDONLY),
-			     NULL) == 0)
-				goto mounted;
-		}
-
-		saved_errno = errno;
-		log_error("mount of system LV (%s, %s, %s) failed: %m\n",
-			  system_lv_name, mount_point, SYSTEM_LV_FILESYSTEM);
-		sync_unlock(lock_name, lockid);
-		errno = saved_errno;
-		return -1;
-	}
-
-      mounted:
-/* Set the internal flags */
-	mounted = 1;
-	mounted_rw = readwrite;
-
-	return 0;
-}
-
-/* Erase *all* files in the root directory of the system LV.
-   This *MUST* be called with an appropriate lock held!
-   The LV is left mounted RW because it is assumed that the
-   caller wants to write something here after clearing some space */
-int system_lv_eraseall(void)
-{
-	DIR *dir;
-	struct dirent *ent;
-	char fname[PATH_MAX];
-
-	/* Must be mounted R/W */
-	system_lv_mount(1);
-
-	dir = opendir(mount_point);
-	if (!dir)
-		return -1;
-
-	while ((ent = readdir(dir))) {
-		struct stat st;
-		snprintf(fname, sizeof(fname), "%s/%s", mount_point,
-			 ent->d_name);
-
-		if (stat(fname, &st)) {
-			if (S_ISREG(st.st_mode))
-				unlink(fname);
-		}
-	}
-	closedir(dir);
-	return 0;
-}
-
-/* This is a "high-level" routine - it mounts the system LV, writes
-   the data into a file named after this node and then umounts the LV
-   again */
-int system_lv_write_data(char *data, ssize_t len)
-{
-	struct utsname nodeinfo;
-	char fname[PATH_MAX];
-	int outfile;
-	ssize_t thiswrite;
-	ssize_t written;
-
-	if (system_lv_mount(1))
-		return -1;
-
-	/* Build the file name we are goingto use. */
-	uname(&nodeinfo);
-	snprintf(fname, sizeof(fname), "%s/%s", mount_point, nodeinfo.nodename);
-
-	/* Open the file for output */
-	outfile = open(fname, O_RDWR | O_CREAT | O_TRUNC, 0600);
-	if (outfile < 0) {
-		int saved_errno = errno;
-		system_lv_umount();
-		errno = saved_errno;
-		return -1;
-	}
-
-	written = 0;
-	do {
-		thiswrite = write(outfile, data + written, len - written);
-		if (thiswrite > 0)
-			written += thiswrite;
-
-	} while (written < len && thiswrite > 0);
-
-	close(outfile);
-
-	system_lv_umount();
-	return (thiswrite < 0) ? -1 : 0;
-}
-
-/* This is a "high-level" routine - it mounts the system LV, reads
-   the data from a named file and then umounts the LV
-   again */
-int system_lv_read_data(char *fname_base, char *data, ssize_t *len)
-{
-	char fname[PATH_MAX];
-	int outfile;
-	struct stat st;
-	ssize_t filesize;
-	ssize_t thisread;
-	ssize_t readbytes;
-
-	if (system_lv_mount(0))
-		return -1;
-
-	/* Build the file name we are going to use. */
-	snprintf(fname, sizeof(fname), "%s/%s", mount_point, fname_base);
-
-	/* Get the file size and stuff. Actually we only need the file size but
-	   this will also check that the file exists */
-	if (stat(fname, &st) < 0) {
-		int saved_errno = errno;
-
-		log_error("stat of file %s on system LV failed: %m\n", fname);
-		system_lv_umount();
-		errno = saved_errno;
-		return -1;
-	}
-	filesize = st.st_size;
-
-	outfile = open(fname, O_RDONLY);
-	if (outfile < 0) {
-		int saved_errno = errno;
-
-		log_error("open of file %s on system LV failed: %m\n", fname);
-		system_lv_umount();
-		errno = saved_errno;
-		return -1;
-	}
-
-	readbytes = 0;
-	do {
-		thisread =
-		    read(outfile, data + readbytes, filesize - readbytes);
-		if (thisread > 0)
-			readbytes += thisread;
-
-	} while (readbytes < filesize && thisread > 0);
-
-	close(outfile);
-
-	system_lv_umount();
-
-	*len = readbytes;
-	return (thisread < 0) ? -1 : 0;
-}

daemons/clvmd/system-lv.h

@@ -1,30 +0,0 @@
-/*
- * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- */
-
-#ifndef _CLVM_SYSTEM_LV_H
-#define _CLVM_SYSTEM_LV_H
-
-/* Prototypes for System-LV functions */
-
-/* "low-level" functions */
-extern int system_lv_umount(void);
-extern int system_lv_mount(int readwrite);
-extern int system_lv_eraseall(void);
-
-/* "high-level" functions */
-extern int system_lv_write_data(char *data, ssize_t len);
-extern int system_lv_read_data(char *fname_base, char *data, ssize_t *len);
-
-#endif

daemons/clvmd/tcp-comms.c

@@ -13,7 +13,10 @@
    It can also make outgoing connnections to the other clvmd nodes.
 */
 
+#define _GNU_SOURCE
+#define _FILE_OFFSET_BITS 64
 
+#include <configure.h>
 #include <pthread.h>
 #include <sys/types.h>
 #include <sys/utsname.h>
@@ -34,8 +37,8 @@
 #include <syslog.h>
 #include <netdb.h>
 #include <assert.h>
+#include <libdevmapper.h>
 
-#include "libdevmapper.h"
 #include "clvm.h"
 #include "clvmd-comms.h"
 #include "clvmd.h"
@@ -57,7 +60,7 @@ int init_comms(unsigned short port)
     struct sockaddr_in6 addr;
 
     sock_hash = dm_hash_create(100);
-    tcp_port = port ? port : DEFAULT_TCP_PORT;
+    tcp_port = port ? : DEFAULT_TCP_PORT;
 
     listen_fd = socket(AF_INET6, SOCK_STREAM, 0);
 
@@ -92,41 +95,37 @@ int init_comms(unsigned short port)
     return 0;
 }
 
-void tcp_remove_client(char *csid)
- {
+void tcp_remove_client(const char *c_csid)
+{
     struct local_client *client;
+    char csid[GULM_MAX_CSID_LEN];
+    unsigned int i;
+    memcpy(csid, c_csid, sizeof csid);
     DEBUGLOG("tcp_remove_client\n");
 
     /* Don't actually close the socket here - that's the
        job of clvmd.c whch will do the job when it notices the
        other end has gone. We just need to remove the client(s) from
        the hash table so we don't try to use it for sending any more */
-    client = dm_hash_lookup_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
-    if (client)
+    for (i = 0; i < 2; i++)
     {
-	dm_hash_remove_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
-	client->removeme = 1;
-	close(client->fd);
+	client = dm_hash_lookup_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
+	if (client)
+	{
+	    dm_hash_remove_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
+	    client->removeme = 1;
+	    close(client->fd);
+	}
+	/* Look for a mangled one too, on the 2nd iteration. */
+	csid[0] ^= 0x80;
     }
-
-    /* Look for a mangled one too */
-    csid[0] ^= 0x80;
-
-    client = dm_hash_lookup_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
-    if (client)
-    {
-	dm_hash_remove_binary(sock_hash, csid, GULM_MAX_CSID_LEN);
-	client->removeme = 1;
-	close(client->fd);
-    }
-
-    /* Put it back as we found it */
-    csid[0] ^= 0x80;
 }
 
-int alloc_client(int fd, char *csid, struct local_client **new_client)
+int alloc_client(int fd, const char *c_csid, struct local_client **new_client)
 {
     struct local_client *client;
+    char csid[GULM_MAX_CSID_LEN];
+    memcpy(csid, c_csid, sizeof csid);
 
     DEBUGLOG("alloc_client %d csid = %s\n", fd, print_csid(csid));
 
@@ -181,7 +180,7 @@ int get_main_gulm_cluster_fd()
 
 
 /* Read on main comms (listen) socket, accept it */
-int cluster_fd_gulm_callback(struct local_client *fd, char *buf, int len, char *csid,
+int cluster_fd_gulm_callback(struct local_client *fd, char *buf, int len, const char *csid,
 			struct local_client **new_client)
 {
     int newfd;
@@ -315,7 +314,7 @@ static int read_from_tcpsock(struct local_client *client, char *buf, int len, ch
     return status;
 }
 
-int gulm_connect_csid(char *csid, struct local_client **newclient)
+int gulm_connect_csid(const char *csid, struct local_client **newclient)
 {
     int fd;
     struct sockaddr_in6 addr;
@@ -366,7 +365,7 @@ int gulm_connect_csid(char *csid, struct local_client **newclient)
 }
 
 /* Send a message to a known CSID */
-static int tcp_send_message(void *buf, int msglen, unsigned char *csid, const char *errtext)
+static int tcp_send_message(void *buf, int msglen, const char *csid, const char *errtext)
 {
     int status;
     struct local_client *client;
@@ -394,7 +393,7 @@ static int tcp_send_message(void *buf, int msglen, unsigned char *csid, const ch
 }
 
 
-int gulm_cluster_send_message(void *buf, int msglen, char *csid, const char *errtext)
+int gulm_cluster_send_message(void *buf, int msglen, const char *csid, const char *errtext)
 {
     int status=0;
 
@@ -465,7 +464,7 @@ static void map_v4_to_v6(struct in_addr *ip4, struct in6_addr *ip6)
 }
 
 /* Get someone else's IP address from DNS */
-int get_ip_address(char *node, char *addr)
+int get_ip_address(const char *node, char *addr)
 {
     struct hostent *he;
 
@@ -493,7 +492,7 @@ int get_ip_address(char *node, char *addr)
     return 0;
 }
 
-char *print_csid(char *csid)
+char *print_csid(const char *csid)
 {
     static char buf[128];
     int *icsid = (int *)csid;

daemons/clvmd/tcp-comms.h

@@ -5,9 +5,9 @@
 #define GULM_MAX_CLUSTER_MEMBER_NAME_LEN 128
 
 extern int init_comms(unsigned short);
-extern char *print_csid(char *);
+extern char *print_csid(const char *);
 int get_main_gulm_cluster_fd(void);
-int cluster_fd_gulm_callback(struct local_client *fd, char *buf, int len, char *csid, struct local_client **new_client);
-int gulm_cluster_send_message(void *buf, int msglen, char *csid, const char *errtext);
+int cluster_fd_gulm_callback(struct local_client *fd, char *buf, int len, const char *csid, struct local_client **new_client);
+int gulm_cluster_send_message(void *buf, int msglen, const char *csid, const char *errtext);
 void get_our_gulm_csid(char *csid);
-int gulm_connect_csid(char *csid, struct local_client **newclient);
+int gulm_connect_csid(const char *csid, struct local_client **newclient);

daemons/dmeventd/.exported_symbols

@@ -1,5 +1,19 @@
-dm_event_register
-dm_event_unregister
+dm_event_handler_create
+dm_event_handler_destroy
+dm_event_handler_set_dso
+dm_event_handler_set_dev_name
+dm_event_handler_set_uuid
+dm_event_handler_set_major
+dm_event_handler_set_minor
+dm_event_handler_set_event_mask
+dm_event_handler_get_dso
+dm_event_handler_get_devname
+dm_event_handler_get_uuid
+dm_event_handler_get_major
+dm_event_handler_get_minor
+dm_event_handler_get_event_mask
+dm_event_register_handler
+dm_event_unregister_handler
 dm_event_get_registered_device
-dm_event_set_timeout
-dm_event_get_timeout
+dm_event_handler_set_timeout
+dm_event_handler_get_timeout

daemons/dmeventd/Makefile.in

@@ -1,5 +1,5 @@
 #
-# Copyright (C) 2005 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2005-2007 Red Hat, Inc. All rights reserved.
 #
 # This file is part of the device-mapper userspace tools.
 #
@@ -15,22 +15,41 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
-SOURCES = libdevmapper-event.c \
-	  dmeventd.c
+SOURCES = libdevmapper-event.c
 
+ifeq ("@STATIC_LINK@", "yes")
 LIB_STATIC = libdevmapper-event.a
+endif
+LIB_VERSION = $(LIB_VERSION_DM)
 
 ifeq ("@LIB_SUFFIX@","dylib")
   LIB_SHARED = libdevmapper-event.dylib
 else
   LIB_SHARED = libdevmapper-event.so
+  VERSIONED_SHLIB = $(LIB_SHARED).$(LIB_VERSION)
 endif
 
-include ../make.tmpl
+TARGETS = dmeventd
+CLEAN_TARGETS = dmeventd.o
 
+ifneq ($(MAKECMDGOALS),device-mapper)
+  SUBDIRS+=plugins
+endif
+
+include $(top_srcdir)/make.tmpl
+
+all: dmeventd
+device-mapper: dmeventd $(LIB_STATIC)
+
+LDFLAGS += -ldl -ldevmapper -lpthread
 CLDFLAGS += -ldl -ldevmapper -lpthread
 
-.PHONY: install_dynamic install_static
+dmeventd: $(LIB_SHARED) $(VERSIONED_SHLIB) dmeventd.o
+	$(CC) -o $@ dmeventd.o $(CFLAGS) $(LDFLAGS) \
+	-L. -ldevmapper-event $(LIBS) -rdynamic
+
+.PHONY: install_dynamic install_static install_include \
+	install_pkgconfig install_dmeventd
 
 INSTALL_TYPE = install_dynamic
 
@@ -38,18 +57,44 @@ ifeq ("@STATIC_LINK@", "yes")
   INSTALL_TYPE += install_static
 endif
 
-install: $(INSTALL_TYPE)
+ifeq ("@PKGCONFIG@", "yes")
+  INSTALL_TYPE += install_pkgconfig
+endif
+
+install: $(INSTALL_TYPE) install_include install_dmeventd
+
+install_device-mapper: $(INSTALL_TYPE) install_include install_dmeventd
+
+install_include:
+	$(INSTALL) -D $(OWNER) $(GROUP) -m 444 libdevmapper-event.h \
+		$(includedir)/libdevmapper-event.h
 
 install_dynamic: libdevmapper-event.$(LIB_SUFFIX)
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< \
 		$(libdir)/libdevmapper-event.$(LIB_SUFFIX).$(LIB_VERSION)
 	$(LN_S) -f libdevmapper-event.$(LIB_SUFFIX).$(LIB_VERSION) \
 		$(libdir)/libdevmapper-event.$(LIB_SUFFIX)
-	$(INSTALL) -D $(OWNER) $(GROUP) -m 444 libdevmapper-event.h \
-		$(includedir)/libdevmapper-event.h
+
+install_dmeventd: dmeventd
+	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< $(sbindir)/$<
+
+install_pkgconfig:
+	$(INSTALL) -D $(OWNER) $(GROUP) -m 444 libdevmapper-event.pc \
+		$(usrlibdir)/pkgconfig/devmapper-event.pc
 
 install_static: libdevmapper-event.a
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< \
 		$(libdir)/libdevmapper-event.a.$(LIB_VERSION)
 	$(LN_S) -f libdevmapper-event.a.$(LIB_VERSION) $(libdir)/libdevmapper-event.a
 
+$(VERSIONED_SHLIB): $(LIB_SHARED)
+	$(RM) -f $@
+	$(LN_S) $(LIB_SHARED) $@
+
+.PHONY: distclean_lib distclean
+
+distclean_lib: 
+	$(RM) libdevmapper-event.pc
+
+distclean: distclean_lib
+

daemons/dmeventd/dmeventd.h

@@ -1,13 +1,66 @@
+/*
+ * Copyright (C) 2005-2007 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
 #ifndef __DMEVENTD_DOT_H__
 #define __DMEVENTD_DOT_H__
 
+/* FIXME This stuff must be configurable. */
+
+#define	DM_EVENT_DAEMON		"/sbin/dmeventd"
+#define DM_EVENT_LOCKFILE	"/var/lock/dmeventd"
+#define	DM_EVENT_FIFO_CLIENT	"/var/run/dmeventd-client"
+#define	DM_EVENT_FIFO_SERVER	"/var/run/dmeventd-server"
+#define DM_EVENT_PIDFILE	"/var/run/dmeventd.pid"
+
+#define DM_EVENT_DEFAULT_TIMEOUT 10
+
+/* Commands for the daemon passed in the message below. */
+enum dm_event_command {
+	DM_EVENT_CMD_ACTIVE = 1,
+	DM_EVENT_CMD_REGISTER_FOR_EVENT,
+	DM_EVENT_CMD_UNREGISTER_FOR_EVENT,
+	DM_EVENT_CMD_GET_REGISTERED_DEVICE,
+	DM_EVENT_CMD_GET_NEXT_REGISTERED_DEVICE,
+	DM_EVENT_CMD_SET_TIMEOUT,
+	DM_EVENT_CMD_GET_TIMEOUT,
+	DM_EVENT_CMD_HELLO,
+};
+
+/* Message passed between client and daemon. */
+struct dm_event_daemon_message {
+	uint32_t cmd;
+	uint32_t size;
+	char *data;
+};
+
+/* FIXME Is this meant to be exported?  I can't see where the
+   interface uses it. */
+/* Fifos for client/daemon communication. */
+struct dm_event_fifos {
+	int client;
+	int server;
+	const char *client_path;
+	const char *server_path;
+};
+
+/*      EXIT_SUCCESS             0 -- stdlib.h */
+/*      EXIT_FAILURE             1 -- stdlib.h */
 #define EXIT_LOCKFILE_INUSE      2
 #define EXIT_DESC_CLOSE_FAILURE  3
-#define EXIT_OPEN_PID_FAILURE    4
-#define EXIT_FIFO_FAILURE        5
-#define EXIT_CHDIR_FAILURE       6
-
-void dmeventd(void)
-    __attribute((noreturn));
+#define EXIT_DESC_OPEN_FAILURE   4
+#define EXIT_OPEN_PID_FAILURE    5
+#define EXIT_FIFO_FAILURE        6
+#define EXIT_CHDIR_FAILURE       7
 
 #endif /* __DMEVENTD_DOT_H__ */

daemons/dmeventd/libdevmapper-event.h

@@ -1,5 +1,5 @@
 /*
- * Copyright (C) 2005 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2005-2007 Red Hat, Inc. All rights reserved.
  *
  * This file is part of the device-mapper userspace tools.
  *
@@ -23,86 +23,84 @@
 
 #include <stdint.h>
 
-/* FIXME This stuff must be configurable. */
+/*
+ * Event library interface.
+ */
 
-#define	DM_EVENT_DAEMON		"/sbin/dmeventd"
-#define DM_EVENT_LOCKFILE	"/var/lock/dmeventd"
-#define	DM_EVENT_FIFO_CLIENT	"/var/run/dmeventd-client"
-#define	DM_EVENT_FIFO_SERVER	"/var/run/dmeventd-server"
-#define DM_EVENT_PIDFILE	"/var/run/dmeventd.pid"
+enum dm_event_mask {
+	DM_EVENT_SETTINGS_MASK  = 0x0000FF,
+	DM_EVENT_SINGLE		= 0x000001, /* Report multiple errors just once. */
+	DM_EVENT_MULTI		= 0x000002, /* Report all of them. */
 
-#define DM_EVENT_DEFAULT_TIMEOUT 10
+	DM_EVENT_ERROR_MASK     = 0x00FF00,
+	DM_EVENT_SECTOR_ERROR	= 0x000100, /* Failure on a particular sector. */
+	DM_EVENT_DEVICE_ERROR	= 0x000200, /* Device failure. */
+	DM_EVENT_PATH_ERROR	= 0x000400, /* Failure on an io path. */
+	DM_EVENT_ADAPTOR_ERROR	= 0x000800, /* Failure of a host adaptor. */
 
-/* Commands for the daemon passed in the message below. */
-enum dm_event_command {
-	DM_EVENT_CMD_ACTIVE = 1,
-	DM_EVENT_CMD_REGISTER_FOR_EVENT,
-	DM_EVENT_CMD_UNREGISTER_FOR_EVENT,
-	DM_EVENT_CMD_GET_REGISTERED_DEVICE,
-	DM_EVENT_CMD_GET_NEXT_REGISTERED_DEVICE,
-	DM_EVENT_CMD_SET_TIMEOUT,
-	DM_EVENT_CMD_GET_TIMEOUT,
+	DM_EVENT_STATUS_MASK    = 0xFF0000,
+	DM_EVENT_SYNC_STATUS	= 0x010000, /* Mirror synchronization completed/failed. */
+	DM_EVENT_TIMEOUT	= 0x020000, /* Timeout has occured */
+
+	DM_EVENT_REGISTRATION_PENDING = 0x1000000, /* Monitor thread is setting-up/shutting-down */
 };
 
-/* Message passed between client and daemon. */
-struct dm_event_daemon_message {
-	union {
-		unsigned int cmd;	/* FIXME Use fixed size. */
-		int	 status;	/* FIXME Use fixed size. */
-	} opcode;
-	char msg[252];		/* FIXME Why is this 252 ? */
-} __attribute__((packed));	/* FIXME Do this properly! */
+#define DM_EVENT_ALL_ERRORS DM_EVENT_ERROR_MASK
 
-/* FIXME Is this meant to be exported?  I can't see where the interface uses it. */
-/* Fifos for client/daemon communication. */
-struct dm_event_fifos {
-	int client;
-	int server;
-	const char *client_path;
-	const char *server_path;
-};
+struct dm_event_handler;
 
-/* Event type definitions. */
-/* FIXME Use masks to separate the types and provide for extension. */
-enum dm_event_type {
-	DM_EVENT_SINGLE		= 0x01, /* Report multiple errors just once. */
-	DM_EVENT_MULTI		= 0x02, /* Report all of them. */
+struct dm_event_handler *dm_event_handler_create(void);
+void dm_event_handler_destroy(struct dm_event_handler *dmevh);
 
-	DM_EVENT_SECTOR_ERROR	= 0x04, /* Failure on a particular sector. */
-	DM_EVENT_DEVICE_ERROR	= 0x08, /* Device failure. */
-	DM_EVENT_PATH_ERROR	= 0x10, /* Failure on an io path. */
-	DM_EVENT_ADAPTOR_ERROR	= 0x20, /* Failure off a host adaptor. */
+/*
+ * Path of shared library to handle events.
+ *
+ * All of dso, device_name and uuid strings are duplicated, you do not
+ * need to keep the pointers valid after the call succeeds. Thes may
+ * return -ENOMEM though.
+ */
+int dm_event_handler_set_dso(struct dm_event_handler *dmevh, const char *path);
 
-	DM_EVENT_SYNC_STATUS	= 0x40, /* Mirror synchronization completed/failed. */
-	DM_EVENT_TIMEOUT	= 0x80, /* Timeout has occured */
-};
+/*
+ * Identify the device to monitor by exactly one of device_name, uuid or
+ * device number. String arguments are duplicated, see above.
+ */
+int dm_event_handler_set_dev_name(struct dm_event_handler *dmevh, const char *device_name);
 
-/* FIXME Use a mask. */
-#define	DM_EVENT_ALL_ERRORS (DM_EVENT_SECTOR_ERROR | DM_EVENT_DEVICE_ERROR | \
-			     DM_EVENT_PATH_ERROR | DM_EVENT_ADAPTOR_ERROR)
+int dm_event_handler_set_uuid(struct dm_event_handler *dmevh, const char *uuid);
 
-/* Prototypes for event lib interface. */
+void dm_event_handler_set_major(struct dm_event_handler *dmevh, int major);
+void dm_event_handler_set_minor(struct dm_event_handler *dmevh, int minor);
+void dm_event_handler_set_timeout(struct dm_event_handler *dmevh, int timeout);
 
-/* FIXME Replace device with standard name/uuid/devno choice */
-/* Interface changes: 
-   First register a handler, passing in a unique ref for the device. */
-//  int dm_event_register_handler(const char *dso_name, const char *device);
-//  int dm_event_register(const char *dso_name, const char *name, const char *uuid, uint32_t major, uint32_t minor, enum dm_event_type events);
-/* Or (better?) add to task structure and use existing functions - run a task to register/unregister events - we may need to run task withe that with the new event mechanism anyway, then the dso calls just hook in.
-*/
- 
-/* FIXME Missing consts? */
-int dm_event_register(char *dso_name, char *device, enum dm_event_type events);
-int dm_event_unregister(char *dso_name, char *device,
-			enum dm_event_type events);
-int dm_event_get_registered_device(char **dso_name, char **device,
-				   enum dm_event_type *events, int next);
-int dm_event_set_timeout(char *device, uint32_t timeout);
-int dm_event_get_timeout(char *device, uint32_t *timeout);
+/*
+ * Specify mask for events to monitor.
+ */
+void dm_event_handler_set_event_mask(struct dm_event_handler *dmevh,
+				     enum dm_event_mask evmask);
 
-/* Prototypes for DSO interface. */
-void process_event(const char *device, enum dm_event_type event);
-int register_device(const char *device);
-int unregister_device(const char *device);
+const char *dm_event_handler_get_dso(const struct dm_event_handler *dmevh);
+const char *dm_event_handler_get_dev_name(const struct dm_event_handler *dmevh);
+const char *dm_event_handler_get_uuid(const struct dm_event_handler *dmevh);
+int dm_event_handler_get_major(const struct dm_event_handler *dmevh);
+int dm_event_handler_get_minor(const struct dm_event_handler *dmevh);
+int dm_event_handler_get_timeout(const struct dm_event_handler *dmevh);
+enum dm_event_mask dm_event_handler_get_event_mask(const struct dm_event_handler *dmevh);
+
+/* FIXME Review interface (what about this next thing?) */
+int dm_event_get_registered_device(struct dm_event_handler *dmevh, int next);
+
+/*
+ * Initiate monitoring using dmeventd.
+ */
+int dm_event_register_handler(const struct dm_event_handler *dmevh);
+int dm_event_unregister_handler(const struct dm_event_handler *dmevh);
+
+/* Prototypes for DSO interface, see dmeventd.c, struct dso_data for
+   detailed descriptions. */
+void process_event(struct dm_task *dmt, enum dm_event_mask evmask, void **user);
+int register_device(const char *device_name, const char *uuid, int major, int minor, void **user);
+int unregister_device(const char *device_name, const char *uuid, int major,
+		      int minor, void **user);
 
 #endif

daemons/dmeventd/libdevmapper-event.pc.in

@@ -0,0 +1,12 @@
+prefix=@prefix@
+exec_prefix=@exec_prefix@
+libdir=@libdir@
+includedir=@includedir@
+
+Name: devmapper-event
+Description: device-mapper event library
+Version: @DM_LIB_VERSION@
+Requires: devmapper
+Cflags: -I${includedir}
+Libs: -L${libdir} -ldevmapper-event
+Libs.private: -lpthread -ldl

daemons/dmeventd/plugins/Makefile.in

@@ -2,7 +2,7 @@
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
 # Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -16,7 +16,7 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
-SUBDIRS += mirror
+SUBDIRS += mirror snapshot
 
 include $(top_srcdir)/make.tmpl
 

daemons/dmeventd/plugins/mirror/Makefile.in

@@ -1,8 +1,8 @@
 #
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
-# Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2005, 2008 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -16,8 +16,8 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
-CFLAGS += -I${top_srcdir}/tools -I${top_srcdir}/include
-CLDFLAGS += -L${top_srcdir}/tools -ldevmapper -llvm2cmd
+INCLUDES += -I${top_srcdir}/tools
+CLDFLAGS += -L${top_srcdir}/tools -ldevmapper @LVM2CMD_LIB@
 
 SOURCES = dmeventd_mirror.c
 
@@ -27,10 +27,13 @@ else
   LIB_SHARED = libdevmapper-event-lvm2mirror.so
 endif
 
+LIB_VERSION = $(LIB_VERSION_LVM)
+
 include $(top_srcdir)/make.tmpl
 
-install: libdevmapper-event-lvm2mirror.$(LIB_SUFFIX)
+install_lvm2: libdevmapper-event-lvm2mirror.$(LIB_SUFFIX)
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< \
 		$(libdir)/$<.$(LIB_VERSION)
-	$(LN_S) -f $(libdir)/$<.$(LIB_VERSION) $(libdir)/$<
+	$(LN_S) -f $<.$(LIB_VERSION) $(libdir)/$<
 
+install: install_lvm2

daemons/dmeventd/plugins/mirror/dmeventd_mirror.c

@@ -12,11 +12,10 @@
  * Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
 
-#include "libdevmapper.h"
-#include "libdevmapper-event.h"
 #include "lvm2cmd.h"
-#include "lvm-string.h"
 
+#include <libdevmapper.h>
+#include <libdevmapper-event.h>
 #include <errno.h>
 #include <signal.h>
 #include <string.h>
@@ -26,147 +25,175 @@
 #include <unistd.h>
 
 #include <syslog.h> /* FIXME Replace syslog with multilog */
+/* FIXME Missing openlog? */
 
 #define ME_IGNORE    0
 #define ME_INSYNC    1
 #define ME_FAILURE   2
 
-/* FIXME: We may need to lock around operations to these */
-static int register_count = 0;
-static struct dm_pool *mem_pool = NULL;
+/*
+ * register_device() is called first and performs initialisation.
+ * Only one device may be registered or unregistered at a time.
+ */
+static pthread_mutex_t _register_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+/*
+ * Number of active registrations.
+ */
+static int _register_count = 0;
+
+static struct dm_pool *_mem_pool = NULL;
+static void *_lvm_handle = NULL;
+
+/*
+ * Currently only one event can be processed at a time.
+ */
+static pthread_mutex_t _event_mutex = PTHREAD_MUTEX_INITIALIZER;
 
 static int _get_mirror_event(char *params)
 {
-	int i, rtn = ME_INSYNC;
-	int max_args = 30;  /* should support at least 8-way mirrors */
-	char *args[max_args];
+	int i, r = ME_INSYNC;
+	char **args = NULL;
 	char *dev_status_str;
 	char *log_status_str;
 	char *sync_str;
-	char *p;
-	int log_argc, num_devs, num_failures=0;
-
-	if (max_args <= split_words(params, max_args, args)) {
-		syslog(LOG_ERR, "Unable to split mirror parameters: Arg list too long");
-		return -E2BIG;
-	}
+	char *p = NULL;
+	int log_argc, num_devs;
 
 	/*
-	 * Unused:  0 409600 mirror
-	 * Used  :  2 253:4 253:5 400/400 1 AA 3 cluster 253:3 A
-	*/
-	num_devs = atoi(args[0]);
-	dev_status_str = args[3 + num_devs];
-	log_argc = atoi(args[4 + num_devs]);
-	log_status_str = args[4 + num_devs + log_argc];
-	sync_str = args[1 + num_devs];
+	 * dm core parms:	     0 409600 mirror
+	 * Mirror core parms:	     2 253:4 253:5 400/400
+	 * New-style failure params: 1 AA
+	 * New-style log params:     3 cluster 253:3 A
+	 *			 or  3 disk 253:3 A
+	 *			 or  1 core
+	 */
+
+	/* number of devices */
+	if (!dm_split_words(params, 1, 0, &p))
+		goto out_parse;
+
+	if (!(num_devs = atoi(p)))
+		goto out_parse;
+	p += strlen(p) + 1;
+
+	/* devices names + "400/400" + "1 AA" + 1 or 3 log parms + NULL */
+	args = dm_malloc((num_devs + 7) * sizeof(char *));
+	if (!args || dm_split_words(p, num_devs + 7, 0, args) < num_devs + 5)
+		goto out_parse;
+
+	dev_status_str = args[2 + num_devs];
+	log_argc = atoi(args[3 + num_devs]);
+	log_status_str = args[3 + num_devs + log_argc];
+	sync_str = args[num_devs];
 
 	/* Check for bad mirror devices */
-	for (i = 0; i < num_devs; i++) {
+	for (i = 0; i < num_devs; i++)
 		if (dev_status_str[i] == 'D') {
-			syslog(LOG_ERR, "Mirror device, %s, has failed.\n", args[i+1]);
-			num_failures++;
+			syslog(LOG_ERR, "Mirror device, %s, has failed.\n", args[i]);
+			r = ME_FAILURE;
 		}
-	}
 
-	/* Check for bad log device */
-	if (log_status_str[0] == 'D') {
+	/* Check for bad disk log device */
+	if (log_argc > 1 && log_status_str[0] == 'D') {
 		syslog(LOG_ERR, "Log device, %s, has failed.\n",
-		       args[3 + num_devs + log_argc]);
-		num_failures++;
+		       args[2 + num_devs + log_argc]);
+		r = ME_FAILURE;
 	}
 
-	if (num_failures) {
-		rtn = ME_FAILURE;
+	if (r == ME_FAILURE)
 		goto out;
-	}
 
 	p = strstr(sync_str, "/");
 	if (p) {
 		p[0] = '\0';
 		if (strcmp(sync_str, p+1))
-			rtn = ME_IGNORE;
+			r = ME_IGNORE;
 		p[0] = '/';
-	} else {
-		/*
-		 * How the hell did we get this?
-		 * Might mean all our parameters are screwed.
-		 */
-		syslog(LOG_ERR, "Unable to parse sync string.");
-		rtn = ME_IGNORE;
-	}
- out:
-	return rtn;
+	} else
+		goto out_parse;
+
+out:
+	if (args)
+		dm_free(args);
+	return r;
+	
+out_parse:
+	if (args)
+		dm_free(args);
+	syslog(LOG_ERR, "Unable to parse mirror status string.");
+	return ME_IGNORE;
 }
 
-static void _temporary_log_fn(int level, const char *file, int line, const char *format)
+static void _temporary_log_fn(int level, const char *file __attribute((unused)),
+			      int line __attribute((unused)),
+			      const char *format)
 {
-	return;
-	syslog(LOG_DEBUG, "%s", format);
+	if (!strncmp(format, "WARNING: ", 9) && (level < 5))
+		syslog(LOG_CRIT, "%s", format);
+	else
+		syslog(LOG_DEBUG, "%s", format);
 }
 
 static int _remove_failed_devices(const char *device)
 {
 	int r;
-	void *handle;
-	int cmd_size = 256;	/* FIXME Use system restriction */
-	char cmd_str[cmd_size];
+#define CMD_SIZE 256	/* FIXME Use system restriction */
+	char cmd_str[CMD_SIZE];
 	char *vg = NULL, *lv = NULL, *layer = NULL;
 
-	if (strlen(device) > 200)
-		return -ENAMETOOLONG;
+	if (strlen(device) > 200)  /* FIXME Use real restriction */
+		return -ENAMETOOLONG;	/* FIXME These return code distinctions are not used so remove them! */
 
-	if (!split_dm_name(mem_pool, device, &vg, &lv, &layer)) {
+	if (!dm_split_lvm_name(_mem_pool, device, &vg, &lv, &layer)) {
 		syslog(LOG_ERR, "Unable to determine VG name from %s",
 		       device);
-		return -ENOMEM;
+		return -ENOMEM;	/* FIXME Replace with generic error return - reason for failure has already got logged */
 	}
 
 	/* FIXME Is any sanity-checking required on %s? */
-	if (cmd_size <= snprintf(cmd_str, cmd_size, "vgreduce --removemissing %s", vg)) {
+	if (CMD_SIZE <= snprintf(cmd_str, CMD_SIZE, "lvconvert --config devices{ignore_suspended_devices=1} --repair --use-policies %s/%s", vg, lv)) {
 		/* this error should be caught above, but doesn't hurt to check again */
 		syslog(LOG_ERR, "Unable to form LVM command: Device name too long");
-		dm_pool_empty(mem_pool);  /* FIXME: not safe with multiple threads */
-		return -ENAMETOOLONG;
+		dm_pool_empty(_mem_pool);  /* FIXME: not safe with multiple threads */
+		return -ENAMETOOLONG; /* FIXME Replace with generic error return - reason for failure has already got logged */
 	}
 
-	lvm2_log_fn(_temporary_log_fn);
-	handle = lvm2_init();
-	lvm2_log_level(handle, 1);
-	r = lvm2_run(handle, cmd_str);
+	r = lvm2_run(_lvm_handle, cmd_str);
 
-	dm_pool_empty(mem_pool);  /* FIXME: not safe with multiple threads */
-	return (r == 1)? 0: -1;
+	if (r == 1) {
+		snprintf(cmd_str, CMD_SIZE, "vgreduce --removemissing %s", vg);
+		if (lvm2_run(_lvm_handle, cmd_str) != 1)
+			syslog(LOG_ERR, "Unable to remove failed PVs from VG %s", vg);
+	}
+
+	dm_pool_empty(_mem_pool);  /* FIXME: not safe with multiple threads */
+	return (r == 1) ? 0 : -1;
 }
 
-void process_event(const char *device, enum dm_event_type event)
+void process_event(struct dm_task *dmt,
+		   enum dm_event_mask event __attribute((unused)),
+		   void **unused __attribute((unused)))
 {
-	struct dm_task *dmt;
 	void *next = NULL;
 	uint64_t start, length;
 	char *target_type = NULL;
 	char *params;
+	const char *device = dm_task_get_name(dmt);
 
-	/* FIXME Move inside libdevmapper */
-	if (!(dmt = dm_task_create(DM_DEVICE_STATUS))) {
-		syslog(LOG_ERR, "Unable to create dm_task.\n");
-		goto fail;
+	if (pthread_mutex_trylock(&_event_mutex)) {
+		syslog(LOG_NOTICE, "Another thread is handling an event.  Waiting...");
+		pthread_mutex_lock(&_event_mutex);
 	}
-
-	if (!dm_task_set_name(dmt, device)) {
-		syslog(LOG_ERR, "Unable to set device name.\n");
-		goto fail;
-	}
-
-	if (!dm_task_run(dmt)) {
-		syslog(LOG_ERR, "Unable to run task.\n");
-		goto fail;
-	}
-
 	do {
 		next = dm_get_next_target(dmt, next, &start, &length,
 					  &target_type, &params);
 
+		if (!target_type) {
+			syslog(LOG_INFO, "%s mapping lost.\n", device);
+			continue;
+		}
+
 		if (strcmp(target_type, "mirror")) {
 			syslog(LOG_INFO, "%s has unmirrored portion.\n", device);
 			continue;
@@ -183,6 +210,7 @@ void process_event(const char *device, enum dm_event_type event)
 		case ME_FAILURE:
 			syslog(LOG_ERR, "Device failure in %s\n", device);
 			if (_remove_failed_devices(device))
+				/* FIXME Why are all the error return codes unused? Get rid of them? */
 				syslog(LOG_ERR, "Failed to remove faulty devices in %s\n",
 				       device);
 			/* Should check before warning user that device is now linear
@@ -194,53 +222,74 @@ void process_event(const char *device, enum dm_event_type event)
 		case ME_IGNORE:
 			break;
 		default:
+			/* FIXME Provide value then! */
 			syslog(LOG_INFO, "Unknown event received.\n");
 		}
 	} while (next);
 
- fail:
-	if (dmt)
-		dm_task_destroy(dmt);
+	pthread_mutex_unlock(&_event_mutex);
 }
 
-int register_device(const char *device)
+int register_device(const char *device,
+		    const char *uuid __attribute((unused)),
+		    int major __attribute((unused)),
+		    int minor __attribute((unused)),
+		    void **unused __attribute((unused)))
 {
-	syslog(LOG_INFO, "Monitoring %s for events\n", device);
+	int r = 0;
+
+	pthread_mutex_lock(&_register_mutex);
+
+	syslog(LOG_INFO, "Monitoring mirror device %s for events\n", device);
 
 	/*
 	 * Need some space for allocations.  1024 should be more
 	 * than enough for what we need (device mapper name splitting)
 	 */
-	if (!mem_pool)
-		mem_pool = dm_pool_create("mirror_dso", 1024);
+	if (!_mem_pool && !(_mem_pool = dm_pool_create("mirror_dso", 1024)))
+		goto out;
 
-	if (!mem_pool)
-		return 0;
-
-	register_count++;
-
-        return 1;
-}
-
-int unregister_device(const char *device)
-{
-        syslog(LOG_INFO, "Stopped monitoring %s for events\n", device);
-
-	if (!(--register_count)) {
-		dm_pool_destroy(mem_pool);
-		mem_pool = NULL;
+	if (!_lvm_handle) {
+		lvm2_log_fn(_temporary_log_fn);
+		if (!(_lvm_handle = lvm2_init())) {
+			dm_pool_destroy(_mem_pool);
+			_mem_pool = NULL;
+			goto out;
+		}
+		lvm2_log_level(_lvm_handle, LVM2_LOG_SUPPRESS);
+		/* FIXME Temporary: move to dmeventd core */
+		lvm2_run(_lvm_handle, "_memlock_inc");
 	}
 
-        return 1;
+	_register_count++;
+	r = 1;
+
+out:
+	pthread_mutex_unlock(&_register_mutex);
+
+	return r;
 }
 
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
+int unregister_device(const char *device,
+		      const char *uuid __attribute((unused)),
+		      int major __attribute((unused)),
+		      int minor __attribute((unused)),
+		      void **unused __attribute((unused)))
+{
+	pthread_mutex_lock(&_register_mutex);
+
+	syslog(LOG_INFO, "No longer monitoring mirror device %s for events\n",
+	       device);
+
+	if (!--_register_count) {
+		dm_pool_destroy(_mem_pool);
+		_mem_pool = NULL;
+		lvm2_run(_lvm_handle, "_memlock_dec");
+		lvm2_exit(_lvm_handle);
+		_lvm_handle = NULL;
+	}
+
+	pthread_mutex_unlock(&_register_mutex);
+
+	return 1;
+}

daemons/dmeventd/plugins/snapshot/Makefile.in

@@ -1,6 +1,6 @@
 #
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
-# Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
 #
 # This file is part of the LVM2.
 #
@@ -16,21 +16,24 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
-CFLAGS += -I${top_srcdir}/tools -I${top_srcdir}/include
-CLDFLAGS += -L${top_srcdir}/tools -ldevmapper -llvm2cmd
+INCLUDES += -I${top_srcdir}/tools
+CLDFLAGS += -L${top_srcdir}/tools -ldevmapper @LVM2CMD_LIB@
 
-SOURCES = dmeventd_mirror.c
+SOURCES = dmeventd_snapshot.c
 
 ifeq ("@LIB_SUFFIX@","dylib")
-  LIB_SHARED = libdevmapper-event-lvm2mirror.dylib
+  LIB_SHARED = libdevmapper-event-lvm2snapshot.dylib
 else
-  LIB_SHARED = libdevmapper-event-lvm2mirror.so
+  LIB_SHARED = libdevmapper-event-lvm2snapshot.so
 endif
 
+LIB_VERSION = $(LIB_VERSION_LVM)
+
 include $(top_srcdir)/make.tmpl
 
-install: libdevmapper-event-lvm2mirror.$(LIB_SUFFIX)
+install_lvm2: libdevmapper-event-lvm2snapshot.$(LIB_SUFFIX)
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< \
 		$(libdir)/$<.$(LIB_VERSION)
-	$(LN_S) -f $(libdir)/$<.$(LIB_VERSION) $(libdir)/$<
+	$(LN_S) -f $<.$(LIB_VERSION) $(libdir)/$<
 
+install: install_lvm2

daemons/dmeventd/plugins/snapshot/dmeventd_snapshot.c

@@ -0,0 +1,213 @@
+/*
+ * Copyright (C) 2007-2008 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include "libdevmapper.h"
+#include "libdevmapper-event.h"
+#include "lvm2cmd.h"
+#include "lvm-string.h"
+
+#include <errno.h>
+#include <signal.h>
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <pthread.h>
+#include <unistd.h>
+
+#include <syslog.h> /* FIXME Replace syslog with multilog */
+/* FIXME Missing openlog? */
+
+/* First warning when snapshot is 80% full. */
+#define WARNING_THRESH 80
+/* Further warnings at 85%, 90% and 95% fullness. */
+#define WARNING_STEP 5
+
+static pthread_mutex_t _register_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+/*
+ * Number of active registrations.
+ */
+static int _register_count = 0;
+
+static struct dm_pool *_mem_pool = NULL;
+static void *_lvm_handle = NULL;
+
+struct snap_status {
+	int invalid;
+	int used;
+	int max;
+};
+
+/*
+ * Currently only one event can be processed at a time.
+ */
+static pthread_mutex_t _event_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+static void _temporary_log_fn(int level,
+			      const char *file __attribute((unused)),
+			      int line __attribute((unused)),
+			      const char *format)
+{
+	if (!strncmp(format, "WARNING: ", 9) && (level < 5))
+		syslog(LOG_CRIT, "%s", format);
+	else
+		syslog(LOG_DEBUG, "%s", format);
+}
+
+/* FIXME possibly reconcile this with target_percent when we gain
+   access to regular LVM library here. */
+static void _parse_snapshot_params(char *params, struct snap_status *stat)
+{
+	char *p;
+	/*
+	 * xx/xx	-- fractions used/max
+	 * Invalid	-- snapshot invalidated
+	 * Unknown	-- status unknown
+	 */
+	stat->used = stat->max = 0;
+
+	if (!strncmp(params, "Invalid", 7)) {
+		stat->invalid = 1;
+		return;
+	}
+
+	/*
+	 * When we return without setting non-zero max, the parent is
+	 * responsible for reporting errors.
+	 */
+	if (!strncmp(params, "Unknown", 7))
+		return;
+
+	if (!(p = strstr(params, "/")))
+		return;
+
+	*p = '\0';
+	p++;
+
+	stat->used = atoi(params);
+	stat->max = atoi(p);
+}
+
+void process_event(struct dm_task *dmt,
+		   enum dm_event_mask event __attribute((unused)),
+		   void **private)
+{
+	void *next = NULL;
+	uint64_t start, length;
+	char *target_type = NULL;
+	char *params;
+	struct snap_status stat = { 0 };
+	const char *device = dm_task_get_name(dmt);
+	int percent, *percent_warning = (int*)private;
+
+	/* No longer monitoring, waiting for remove */
+	if (!*percent_warning)
+		return;
+
+	if (pthread_mutex_trylock(&_event_mutex)) {
+		syslog(LOG_NOTICE, "Another thread is handling an event.  Waiting...");
+		pthread_mutex_lock(&_event_mutex);
+	}
+
+	dm_get_next_target(dmt, next, &start, &length, &target_type, &params);
+	if (!target_type)
+		goto out;
+
+	_parse_snapshot_params(params, &stat);
+	/*
+	 * If the snapshot has been invalidated or we failed to parse
+	 * the status string. Report the full status string to syslog.
+	 */
+	if (stat.invalid || !stat.max) {
+		syslog(LOG_ERR, "Snapshot %s changed state to: %s\n", device, params);
+		*percent_warning = 0;
+		goto out;
+	}
+
+	percent = 100 * stat.used / stat.max;
+	if (percent >= *percent_warning) {
+		syslog(LOG_WARNING, "Snapshot %s is now %i%% full.\n", device, percent);
+		/* Print warning on the next multiple of WARNING_STEP. */
+		*percent_warning = (percent / WARNING_STEP) * WARNING_STEP + WARNING_STEP;
+	}
+out:
+	pthread_mutex_unlock(&_event_mutex);
+}
+
+int register_device(const char *device,
+		    const char *uuid __attribute((unused)),
+		    int major __attribute((unused)),
+		    int minor __attribute((unused)),
+		    void **private)
+{
+	int r = 0;
+	int *percent_warning = (int*)private;
+
+	pthread_mutex_lock(&_register_mutex);
+
+	/*
+	 * Need some space for allocations.  1024 should be more
+	 * than enough for what we need (device mapper name splitting)
+	 */
+	if (!_mem_pool && !(_mem_pool = dm_pool_create("snapshot_dso", 1024)))
+		goto out;
+
+	*percent_warning = WARNING_THRESH; /* Print warning if snapshot is full */
+
+	if (!_lvm_handle) {
+		lvm2_log_fn(_temporary_log_fn);
+		if (!(_lvm_handle = lvm2_init())) {
+			dm_pool_destroy(_mem_pool);
+			_mem_pool = NULL;
+			goto out;
+		}
+		lvm2_log_level(_lvm_handle, LVM2_LOG_SUPPRESS);
+		/* FIXME Temporary: move to dmeventd core */
+		lvm2_run(_lvm_handle, "_memlock_inc");
+	}
+
+	syslog(LOG_INFO, "Monitoring snapshot %s\n", device);
+
+	_register_count++;
+	r = 1;
+
+out:
+	pthread_mutex_unlock(&_register_mutex);
+
+	return r;
+}
+
+int unregister_device(const char *device,
+		      const char *uuid __attribute((unused)),
+		      int major __attribute((unused)),
+		      int minor __attribute((unused)),
+		      void **unused __attribute((unused)))
+{
+	pthread_mutex_lock(&_register_mutex);
+
+	syslog(LOG_INFO, "No longer monitoring snapshot %s\n",
+	       device);
+
+	if (!--_register_count) {
+		dm_pool_destroy(_mem_pool);
+		_mem_pool = NULL;
+		lvm2_run(_lvm_handle, "_memlock_dec");
+		lvm2_exit(_lvm_handle);
+		_lvm_handle = NULL;
+	}
+
+	pthread_mutex_unlock(&_register_mutex);
+
+	return 1;
+}

debian/README.Debian

@@ -1,6 +0,0 @@
-LVM2 requires the device-mapper kernel module (dm-mod).  This is
-available as a kernel patch for 2.4 (in kernel-patch-device-mapper), and
-is distributed with linux 2.5 and above.  The LVM1 kernel module (lvm-mod)
-will not work with lvm2 packages.  dm-mod and lvm-mod may both be loaded
-in the kernel at the same time with no problems.  Without dm-mod, this
-package is pretty useless.

debian/changelog

@@ -1,82 +0,0 @@
-lvm2 (1.95.15-1) unstable; urgency=low
-  
-  * New upstream release.
-  * Remove undocumented manpage symlinks.
-  * Update description to be more informative.  (Closes: #173499)
-  * Add kernel-patch-device-mapper suggestion.
-  * Update standards version.
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Sun, 16 Feb 2002 04:21:26 -0400
-
-lvm2 (1.95.11-1) unstable; urgency=low
-
-  * New upstream release.  (Closes: #171436)
-  * Removed TODO and INTRO from debian/docs; added WHATS_NEW.
-  * Remove vgcfgrestore.8 undocumented symlink.
-  * Added a README.Debian, mentioning the device-mapper kernel module
-    requirement that lvm2 has.  (Closes: #171674, #163020)
-  * Get rid of debian/conffiles (debhelper's smart enough to figure that out).
-  * debian/copyright fix to appease lintian.
-  * Fix typo in tools/commands.h that caused /usr/sbin/; to be created.
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Mon,  9 Dec 2002 02:51:02 -0400
-
-lvm2 (1.95.10-2) unstable; urgency=low
-
-  * Fix software raid problems by ensuring lvm init script runs after
-    raidtools init script.  (Closes: #152569)
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Tue,  3 Sep 2002 04:05:43 -0400
-
-lvm2 (1.95.10-1) unstable; urgency=low
-
-  * New upstream release (Beta 3.2).
-  * Change all references to /dev/device-mapper/control to
-    /dev/mapper/control.
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Sun,  1 Sep 2002 18:55:12 -0400
-
-lvm2 (0.95.05-3) unstable; urgency=low
-
-  * Get rid of awk dependency in init script.  (Closes: #146257)
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Sun, 12 May 2002 04:39:06 -0500
-
-lvm2 (0.95.05-2) unstable; urgency=low
-
-  * Use ${shlibs:Depends} in Depends.
-  * Get rid of postinst/postrm scripts, use debhelper's init script instead.
-  * Add Conflicts against lvm10, lvm-common.
-  * Fix endian issues on big-endian machines.
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Thu,  2 May 2002 23:53:53 -0500
-
-lvm2 (0.95.05-1) unstable; urgency=low
-
-  * New release (Beta2).
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Thu, 25 Apr 2002 00:37:41 -0500
-
-lvm2 (0.95.04cvs20020306-1) unstable; urgency=low
-
-  * CVS updated.
-  * Convert from debian native package.
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Wed,  6 Mar 2002 00:43:21 -0500
-
-lvm2 (0.95.04cvs20020304) unstable; urgency=low
-
-  * CVS updated.
-  * Enhance init script; create devmapper control device, etc.
-  * Add dmsetup as a suggestion.
-  * Add /etc/lvm/lvm.conf conffile.
-  * Add undocumented(7) for the commands missing manpages.
-  
- -- Andres Salomon <dilinger@mp3revolution.net>  Mon,  4 Mar 2002 04:51:26 -0500
-
-lvm2 (0.95.02cvs20020220) unstable; urgency=low
-
-  * Initial Release.
-
- -- Andres Salomon <dilinger@mp3revolution.net>  Wed, 20 Feb 2002 03:17:25 -0500
-

debian/control

@@ -1,24 +0,0 @@
-Source: lvm2
-Section: admin
-Priority: optional
-Maintainer: Andres Salomon <dilinger@mp3revolution.net>
-Build-Depends: debhelper (>> 3.0.0), libdevmapper-dev (>= 0.96.04), libreadline4-dev
-Standards-Version: 3.5.8.0
-
-Package: lvm2
-Architecture: any
-Depends: ${shlibs:Depends}
-Conflicts: lvm10, lvm-common
-Replaces: lvm10, lvm-common
-Provides: lvm-binaries
-Suggests: dmsetup, kernel-patch-device-mapper
-Description: The Linux Logical Volume Manager
- This is LVM2, the rewrite of The Linux Logical Volume Manager.  LVM
- supports enterprise level volume management of disk and disk subsystems
- by grouping arbitrary disks into volume groups. The total capacity of
- volume groups can be allocated to logical volumes, which are accessed as
- regular block devices.
- .
- LVM2 is currently stable, but has some unimplemented features (most notably,
- pvmove and e2fsadm).  It is not yet recommended for production use.  It is
- backwards-compatible with LVM1 (lvm10), and requires Linux kernel 2.4.

debian/copyright

@@ -1,25 +0,0 @@
-This package was debianized by Andres Salomon <dilinger@mp3revolution.net> on
-Wed, 20 Feb 2002 03:17:25 -0500.
-
-It was downloaded from http://www.sistina.com/products_lvm.htm
-
-Upstream Author: LVM Development Team
-
-Copyright (c) 2001-2002 LVM Development Team
-
-LVM2 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.
-
-LVM2 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
-
-On Debian systems, the full text of the GPL can be found in
-/usr/share/common-licenses/GPL

debian/dirs

@@ -1,4 +0,0 @@
-etc/lvm
-usr/share/man/man5
-usr/share/man/man8
-sbin

debian/docs

@@ -1,5 +0,0 @@
-BUGS
-README
-VERSION
-WHATS_NEW
-doc/*

debian/init.d

@@ -1,64 +0,0 @@
-#! /bin/sh
-#
-# lvm2		This script handles LVM2 initialization/shutdown.
-#
-#		Written by Andres Salomon <dilinger@mp3revolution.net>.
-#
-
-PATH=/sbin:/bin:/usr/sbin:/usr/bin
-NAME=lvm2
-DESC=LVM
-
-test -x /sbin/vgchange || exit 0
-modprobe dm-mod >/dev/null 2>&1
-
-# Create necessary files in /dev for device-mapper
-create_devfiles() {
-	DIR="/dev/mapper"
-	FILE="$DIR/control"
-	major=$(grep "[0-9] misc$" /proc/devices | sed 's/[ ]\+misc//')
-	minor=$(grep "[0-9] device-mapper$" /proc/misc | sed 's/[ ]\+device-mapper//')
-
-	if test ! -d $DIR; then
-		mkdir --mode=755 $DIR >/dev/null 2>&1
-	fi
-
-	if test ! -c $FILE -a ! -z "$minor"; then
-		mknod --mode=600 $FILE c $major $minor >/dev/null 2>&1
-	fi
-}
-
-case "$1" in
-  start)
-	echo -n "Initializing $DESC: "
-	create_devfiles
-	vgchange -a y
-
-#	# Mount all LVM devices
-#	for vg in $( vgchange -a y 2>/dev/null | grep active | awk -F\" '{print $2}' ); do
-#		MTPT=$( grep $vg /etc/fstab | awk '{print $2}' )
-#		mount $MTPT
-#	done
-	echo "$NAME."
-	;;
-  stop)
-	echo -n "Shutting down $DESC: "
-	# We don't really try all that hard to shut it down; far too many
-	# things that can keep it from successfully shutting down.
-	vgchange -a n
-	echo "$NAME."
-	;;
-  restart|force-reload)
-	echo -n "Restarting $DESC: "
-	vgchange -a n
-	sleep 1
-	vgchange -a y
-	echo "$NAME."
-	;;
-  *)
-	echo "Usage: /etc/init.d/$NAME {start|stop|restart|force-reload}" >&2
-	exit 1
-	;;
-esac
-
-exit 0

debian/manpages

@@ -1,26 +0,0 @@
-debian/lvm2/usr/share/man/man5/lvm.conf.5
-debian/lvm2/usr/share/man/man8/lvchange.8
-debian/lvm2/usr/share/man/man8/lvcreate.8
-debian/lvm2/usr/share/man/man8/lvdisplay.8
-debian/lvm2/usr/share/man/man8/lvextend.8
-debian/lvm2/usr/share/man/man8/lvm.8
-debian/lvm2/usr/share/man/man8/lvmchange.8
-debian/lvm2/usr/share/man/man8/lvreduce.8
-debian/lvm2/usr/share/man/man8/lvremove.8
-debian/lvm2/usr/share/man/man8/lvrename.8
-debian/lvm2/usr/share/man/man8/lvscan.8
-debian/lvm2/usr/share/man/man8/pvchange.8
-debian/lvm2/usr/share/man/man8/pvcreate.8
-debian/lvm2/usr/share/man/man8/pvdisplay.8
-debian/lvm2/usr/share/man/man8/pvscan.8
-debian/lvm2/usr/share/man/man8/vgcfgbackup.8
-debian/lvm2/usr/share/man/man8/vgchange.8
-debian/lvm2/usr/share/man/man8/vgck.8
-debian/lvm2/usr/share/man/man8/vgcreate.8
-debian/lvm2/usr/share/man/man8/vgdisplay.8
-debian/lvm2/usr/share/man/man8/vgextend.8
-debian/lvm2/usr/share/man/man8/vgmerge.8
-debian/lvm2/usr/share/man/man8/vgreduce.8
-debian/lvm2/usr/share/man/man8/vgremove.8
-debian/lvm2/usr/share/man/man8/vgrename.8
-debian/lvm2/usr/share/man/man8/vgscan.8

debian/rules

@@ -1,119 +0,0 @@
-#!/usr/bin/make -f
-# Sample debian/rules that uses debhelper. 
-# GNU copyright 1997 by Joey Hess.
-#
-# This version is for a hypothetical package that builds an
-# architecture-dependant package, as well as an architecture-independent
-# package.
-
-# Uncomment this to turn on verbose mode. 
-#export DH_VERBOSE=1
-
-# This is the debhelper compatibility version to use.
-export DH_COMPAT=3
-
-# These are used for cross-compiling and for saving the configure script
-# from having to guess our platform (since we know it already)
-DEB_HOST_GNU_TYPE   ?= $(shell dpkg-architecture -qDEB_HOST_GNU_TYPE)
-DEB_BUILD_GNU_TYPE  ?= $(shell dpkg-architecture -qDEB_BUILD_GNU_TYPE)
-
-ifneq (,$(findstring debug,$(DEB_BUILD_OPTIONS)))
-	CFLAGS += -g
-endif
-ifeq (,$(findstring nostrip,$(DEB_BUILD_OPTIONS)))
-	INSTALL_PROGRAM += -s
-endif
-
-configure: configure-stamp
-configure-stamp:
-	dh_testdir
-	# Add here commands to configure the package.
-	./configure --host=$(DEB_HOST_GNU_TYPE) --build=$(DEB_BUILD_GNU_TYPE) --prefix=/ --mandir=\$${prefix}/usr/share/man --infodir=\$${prefix}/usr/share/info
-
-	touch configure-stamp
-
-build-arch: configure-stamp build-arch-stamp
-build-arch-stamp:
-	dh_testdir
-
-	# Add here command to compile/build the package.
-	$(MAKE)
-
-	touch build-arch-stamp
-
-build-indep: configure-stamp build-indep-stamp
-build-indep-stamp:
-	dh_testdir
-
-	# Add here command to compile/build the arch indep package.
-	# It's ok not to do anything here, if you don't need to build
-	#  anything for this package.
-	#/usr/bin/docbook-to-man debian/lvm2.sgml > lvm2.1
-
-	touch build-indep-stamp
-
-build: build-arch build-indep
-
-clean:
-	dh_testdir
-	dh_testroot
-	rm -f build-stamp configure-stamp
-
-	# Add here commands to clean up after the build process.
-	-$(MAKE) distclean
-	-test -r /usr/share/misc/config.sub && \
-	  cp -f /usr/share/misc/config.sub config.sub
-	-test -r /usr/share/misc/config.guess && \
-	  cp -f /usr/share/misc/config.guess config.guess
-
-
-	dh_clean
-
-install: DH_OPTIONS=
-install: build
-	dh_testdir
-	dh_testroot
-	dh_clean -k
-	dh_installdirs
-
-	# Add here commands to install the package into debian/lvm2.
-	$(MAKE) install prefix=$(CURDIR)/debian/lvm2
-	install -m 0644 doc/example.conf debian/lvm2/etc/lvm/lvm.conf
-
-
-# Build architecture-independent files here.
-# Pass -i to all debhelper commands in this target to reduce clutter.
-binary-indep: build install
-# nada.
-
-# Build architecture-dependent files here.
-binary-arch: build install
-	dh_testdir
-	dh_testroot
-
-#	dh_installdebconf
-	dh_installdocs
-	dh_installexamples
-#	dh_installlogrotate -a
-#	dh_installemacsen -a
-#	dh_installpam -a
-#	dh_installmime -a
-	dh_installinit --update-rcd-params="start 26 S . start 50 0 6 ."
-	dh_installcron
-	dh_installman
-	dh_installinfo 
-	dh_installchangelogs
-	dh_strip
-	dh_link
-	dh_compress
-	dh_fixperms
-	dh_makeshlibs
-	dh_installdeb
-#	dh_perl -a
-	dh_shlibdeps
-	dh_gencontrol
-	dh_md5sums
-	dh_builddeb
-
-binary: binary-indep binary-arch
-.PHONY: build clean binary-indep binary-arch binary install configure

dmeventd/mirror/dmeventd_mirror.c

@@ -1,246 +0,0 @@
-/*
- * Copyright (C) 2005 Red Hat, Inc. All rights reserved.
- *
- * This file is part of LVM2.
- *
- * This copyrighted material is made available to anyone wishing to use,
- * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- */
-
-#include "libdevmapper.h"
-#include "libdevmapper-event.h"
-#include "lvm2cmd.h"
-#include "lvm-string.h"
-
-#include <errno.h>
-#include <signal.h>
-#include <string.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <pthread.h>
-#include <unistd.h>
-
-#include <syslog.h> /* FIXME Replace syslog with multilog */
-
-#define ME_IGNORE    0
-#define ME_INSYNC    1
-#define ME_FAILURE   2
-
-/* FIXME: We may need to lock around operations to these */
-static int register_count = 0;
-static struct dm_pool *mem_pool = NULL;
-
-static int _get_mirror_event(char *params)
-{
-	int i, rtn = ME_INSYNC;
-	int max_args = 30;  /* should support at least 8-way mirrors */
-	char *args[max_args];
-	char *dev_status_str;
-	char *log_status_str;
-	char *sync_str;
-	char *p;
-	int log_argc, num_devs, num_failures=0;
-
-	if (max_args <= split_words(params, max_args, args)) {
-		syslog(LOG_ERR, "Unable to split mirror parameters: Arg list too long");
-		return -E2BIG;
-	}
-
-	/*
-	 * Unused:  0 409600 mirror
-	 * Used  :  2 253:4 253:5 400/400 1 AA 3 cluster 253:3 A
-	*/
-	num_devs = atoi(args[0]);
-	dev_status_str = args[3 + num_devs];
-	log_argc = atoi(args[4 + num_devs]);
-	log_status_str = args[4 + num_devs + log_argc];
-	sync_str = args[1 + num_devs];
-
-	/* Check for bad mirror devices */
-	for (i = 0; i < num_devs; i++) {
-		if (dev_status_str[i] == 'D') {
-			syslog(LOG_ERR, "Mirror device, %s, has failed.\n", args[i+1]);
-			num_failures++;
-		}
-	}
-
-	/* Check for bad log device */
-	if (log_status_str[0] == 'D') {
-		syslog(LOG_ERR, "Log device, %s, has failed.\n",
-		       args[3 + num_devs + log_argc]);
-		num_failures++;
-	}
-
-	if (num_failures) {
-		rtn = ME_FAILURE;
-		goto out;
-	}
-
-	p = strstr(sync_str, "/");
-	if (p) {
-		p[0] = '\0';
-		if (strcmp(sync_str, p+1))
-			rtn = ME_IGNORE;
-		p[0] = '/';
-	} else {
-		/*
-		 * How the hell did we get this?
-		 * Might mean all our parameters are screwed.
-		 */
-		syslog(LOG_ERR, "Unable to parse sync string.");
-		rtn = ME_IGNORE;
-	}
- out:
-	return rtn;
-}
-
-static void _temporary_log_fn(int level, const char *file, int line, const char *format)
-{
-	return;
-	syslog(LOG_DEBUG, "%s", format);
-}
-
-static int _remove_failed_devices(const char *device)
-{
-	int r;
-	void *handle;
-	int cmd_size = 256;	/* FIXME Use system restriction */
-	char cmd_str[cmd_size];
-	char *vg = NULL, *lv = NULL, *layer = NULL;
-
-	if (strlen(device) > 200)
-		return -ENAMETOOLONG;
-
-	if (!split_dm_name(mem_pool, device, &vg, &lv, &layer)) {
-		syslog(LOG_ERR, "Unable to determine VG name from %s",
-		       device);
-		return -ENOMEM;
-	}
-
-	/* FIXME Is any sanity-checking required on %s? */
-	if (cmd_size <= snprintf(cmd_str, cmd_size, "vgreduce --removemissing %s", vg)) {
-		/* this error should be caught above, but doesn't hurt to check again */
-		syslog(LOG_ERR, "Unable to form LVM command: Device name too long");
-		dm_pool_empty(mem_pool);  /* FIXME: not safe with multiple threads */
-		return -ENAMETOOLONG;
-	}
-
-	lvm2_log_fn(_temporary_log_fn);
-	handle = lvm2_init();
-	lvm2_log_level(handle, 1);
-	r = lvm2_run(handle, cmd_str);
-
-	dm_pool_empty(mem_pool);  /* FIXME: not safe with multiple threads */
-	return (r == 1)? 0: -1;
-}
-
-void process_event(const char *device, enum dm_event_type event)
-{
-	struct dm_task *dmt;
-	void *next = NULL;
-	uint64_t start, length;
-	char *target_type = NULL;
-	char *params;
-
-	/* FIXME Move inside libdevmapper */
-	if (!(dmt = dm_task_create(DM_DEVICE_STATUS))) {
-		syslog(LOG_ERR, "Unable to create dm_task.\n");
-		goto fail;
-	}
-
-	if (!dm_task_set_name(dmt, device)) {
-		syslog(LOG_ERR, "Unable to set device name.\n");
-		goto fail;
-	}
-
-	if (!dm_task_run(dmt)) {
-		syslog(LOG_ERR, "Unable to run task.\n");
-		goto fail;
-	}
-
-	do {
-		next = dm_get_next_target(dmt, next, &start, &length,
-					  &target_type, &params);
-
-		if (strcmp(target_type, "mirror")) {
-			syslog(LOG_INFO, "%s has unmirrored portion.\n", device);
-			continue;
-		}
-
-		switch(_get_mirror_event(params)) {
-		case ME_INSYNC:
-			/* FIXME: all we really know is that this
-			   _part_ of the device is in sync
-			   Also, this is not an error
-			*/
-			syslog(LOG_NOTICE, "%s is now in-sync\n", device);
-			break;
-		case ME_FAILURE:
-			syslog(LOG_ERR, "Device failure in %s\n", device);
-			if (_remove_failed_devices(device))
-				syslog(LOG_ERR, "Failed to remove faulty devices in %s\n",
-				       device);
-			/* Should check before warning user that device is now linear
-			else
-				syslog(LOG_NOTICE, "%s is now a linear device.\n",
-					device);
-			*/
-			break;
-		case ME_IGNORE:
-			break;
-		default:
-			syslog(LOG_INFO, "Unknown event received.\n");
-		}
-	} while (next);
-
- fail:
-	if (dmt)
-		dm_task_destroy(dmt);
-}
-
-int register_device(const char *device)
-{
-	syslog(LOG_INFO, "Monitoring %s for events\n", device);
-
-	/*
-	 * Need some space for allocations.  1024 should be more
-	 * than enough for what we need (device mapper name splitting)
-	 */
-	if (!mem_pool)
-		mem_pool = dm_pool_create("mirror_dso", 1024);
-
-	if (!mem_pool)
-		return 0;
-
-	register_count++;
-
-        return 1;
-}
-
-int unregister_device(const char *device)
-{
-        syslog(LOG_INFO, "Stopped monitoring %s for events\n", device);
-
-	if (!(--register_count)) {
-		dm_pool_destroy(mem_pool);
-		mem_pool = NULL;
-	}
-
-        return 1;
-}
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */

doc/Makefile.in

@@ -1,7 +1,7 @@
 #
 # Copyright (C) 2004 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -20,10 +20,11 @@ CONFDEST=lvm.conf
 
 include $(top_srcdir)/make.tmpl
 
-install:
+install_lvm2:
 	@if [ ! -e $(confdir)/$(CONFDEST) ]; then \
 		echo "Installing $(CONFSRC) as $(confdir)/$(CONFDEST)"; \
 		@INSTALL@ -D $(OWNER) $(GROUP) -m 644 $(CONFSRC) \
 			$(confdir)/$(CONFDEST); \
 	fi
 
+install: install_lvm2

doc/example.conf

@@ -19,6 +19,15 @@ devices {
     # to use with LVM2.
     scan = [ "/dev" ]
 
+    # If several entries in the scanned directories correspond to the
+    # same block device and the tools need to display a name for device,
+    # all the pathnames are matched against each item in the following
+    # list of regular expressions in turn and the first match is used.
+    preferred_names = [ ]
+
+    # Try to avoid using undescriptive /dev/dm-N names, if present.
+    # preferred_names = [ "^/dev/mpath/", "^/dev/mapper/mpath", "^/dev/[hs]d" ]
+
     # A filter that tells LVM2 to only use a restricted set of devices.
     # The filter consists of an array of regular expressions.  These
     # expressions can be delimited by a character of your choice, and
@@ -56,10 +65,14 @@ devices {
     # filter = [ "a|^/dev/hda8$|", "r/.*/" ]
 
     # The results of the filtering are cached on disk to avoid
-    # rescanning dud devices (which can take a very long time).  By
-    # default this cache file is hidden in the /etc/lvm directory.
-    # It is safe to delete this file: the tools regenerate it.
-    cache = "/etc/lvm/.cache"
+    # rescanning dud devices (which can take a very long time).
+    # By default this cache is stored in the /etc/lvm/cache directory
+    # in a file called '.cache'.
+    # It is safe to delete the contents: the tools regenerate it.
+    # (The old setting 'cache' is still respected if neither of
+    # these new ones is present.)
+    cache_dir = "/etc/lvm/cache"
+    cache_file_prefix = ""
 
     # You can turn off writing this cache file by setting this to 0.
     write_cache_state = 1
@@ -73,12 +86,29 @@ devices {
     # If sysfs is mounted (2.6 kernels) restrict device scanning to 
     # the block devices it believes are valid.
     # 1 enables; 0 disables.
-    sysfs_scan = 1	
+    sysfs_scan = 1
 
     # By default, LVM2 will ignore devices used as components of
     # software RAID (md) devices by looking for md superblocks.
     # 1 enables; 0 disables.
     md_component_detection = 1
+
+    # By default, if a PV is placed directly upon an md device, LVM2
+    # will align its data blocks with the the chunk_size exposed in sysfs.
+    # 1 enables; 0 disables.
+    md_chunk_alignment = 1
+
+    # Alignment (in KB) of start of data area when creating a new PV.
+    # If a PV is placed directly upon an md device and md_chunk_alignment is
+    # enabled this parameter is ignored.
+    # Set to 0 for the default alignment of 64KB or page size, if larger.
+    data_alignment = 0
+
+    # If, while scanning the system for PVs, LVM2 encounters a device-mapper
+    # device that has its I/O suspended, it waits for it to become accessible.
+    # Set this to 1 to skip such devices.  This should only be needed
+    # in recovery situations.
+    ignore_suspended_devices = 0
 }
 
 # This section that allows you to configure the nature of the
@@ -105,7 +135,7 @@ log {
     # There are 6 syslog-like log levels currently in use - 2 to 7 inclusive.
     # 7 is the most verbose (LOG_DEBUG).
     level = 0
-    
+
     # Format of output messages
     # Whether or not (1 or 0) to indent messages according to their severity
     indent = 1
@@ -151,7 +181,7 @@ backup {
     # Where should archived files go ?
     # Remember to back up this directory regularly!
     archive_dir = "/etc/lvm/archive"
-    
+
     # What is the minimum number of archive files you wish to keep ?
     retain_min = 10
 
@@ -169,7 +199,7 @@ shell {
 
 # Miscellaneous global LVM2 settings
 global {
-    
+
     # The file creation mask for any files and directories created.
     # Interpreted as octal if the first digit is zero.
     umask = 077
@@ -182,6 +212,9 @@ global {
     # command.  Defaults to off.
     test = 0
 
+    # Default value for --units argument
+    units = "h"
+
     # Whether or not to communicate with the kernel device-mapper.
     # Set to 0 if you want to use the tools to manipulate LVM metadata 
     # without activating any logical volumes.
@@ -207,11 +240,26 @@ global {
     # Location of proc filesystem
     proc = "/proc"
 
-    # Type of locking to use. Defaults to file-based locking (1).
+    # Type of locking to use. Defaults to local file-based locking (1).
     # Turn locking off by setting to 0 (dangerous: risks metadata corruption
     # if LVM2 commands get run concurrently).
+    # Type 2 uses the external shared library locking_library.
+    # Type 3 uses built-in clustered locking.
     locking_type = 1
 
+    # If using external locking (type 2) and initialisation fails,
+    # with this set to 1 an attempt will be made to use the built-in
+    # clustered locking.
+    # If you are using a customised locking_library you should set this to 0.
+    fallback_to_clustered_locking = 1
+
+    # If an attempt to initialise type 2 or type 3 locking failed, perhaps
+    # because cluster components such as clvmd are not running, with this set
+    # to 1 an attempt will be made to use local file-based locking (type 1).
+    # If this succeeds, only commands against local volume groups will proceed.
+    # Volume Groups marked as clustered will be ignored.
+    fallback_to_local_locking = 1
+
     # Local non-LV directory that holds file-based locks while commands are
     # in progress.  A directory like /tmp that may get wiped on reboot is OK.
     locking_dir = "/var/lock/lvm"
@@ -223,17 +271,19 @@ global {
 
     # Search this directory first for shared libraries.
     #   library_dir = "/lib"
+
+    # The external locking library to load if locking_type is set to 2.
+    #   locking_library = "liblvm2clusterlock.so"
 }
 
 activation {
-    # Device used in place of missing stripes if activating incomplete volume.
-    # For now, you need to set this up yourself first (e.g. with 'dmsetup')
-    # For example, you could make it return I/O errors using the 'error' 
-    # target or make it return zeros.
-    missing_stripe_filler = "/dev/ioerror"
-
-    # Size (in KB) of each copy operation when mirroring
-    mirror_region_size = 512
+    # How to fill in missing stripes if activating an incomplete volume.
+    # Using "error" will make inaccessible parts of the device return
+    # I/O errors on access.  You can instead use a device path, in which 
+    # case, that device will be used to in place of missing stripes.
+    # But note that using anything other than "error" with mirrored 
+    # or snapshotted volumes is likely to result in data corruption.
+    missing_stripe_filler = "error"
 
     # How much stack (in KB) to reserve for use while devices suspended
     reserved_stack = 256
@@ -251,6 +301,55 @@ activation {
     #   "@*" matches if any tag defined on the host is also set in the LV or VG
     #
     # volume_list = [ "vg1", "vg2/lvol1", "@tag1", "@*" ]
+
+    # Size (in KB) of each copy operation when mirroring
+    mirror_region_size = 512
+
+    # Setting to use when there is no readahead value stored in the metadata.
+    #
+    # "none" - Disable readahead.
+    # "auto" - Use default value chosen by kernel.
+    readahead = "auto"
+
+    # 'mirror_image_fault_policy' and 'mirror_log_fault_policy' define
+    # how a device failure affecting a mirror is handled.
+    # A mirror is composed of mirror images (copies) and a log.
+    # A disk log ensures that a mirror does not need to be re-synced
+    # (all copies made the same) every time a machine reboots or crashes.
+    #
+    # In the event of a failure, the specified policy will be used to determine
+    # what happens. This applies to automatic repairs (when the mirror is being
+    # monitored by dmeventd) and to manual lvconvert --repair when
+    # --use-policies is given.
+    #
+    # "remove" - Simply remove the faulty device and run without it.  If
+    #            the log device fails, the mirror would convert to using
+    #            an in-memory log.  This means the mirror will not
+    #            remember its sync status across crashes/reboots and
+    #            the entire mirror will be re-synced.  If a
+    #            mirror image fails, the mirror will convert to a
+    #            non-mirrored device if there is only one remaining good
+    #            copy.
+    #
+    # "allocate" - Remove the faulty device and try to allocate space on
+    #            a new device to be a replacement for the failed device.
+    #            Using this policy for the log is fast and maintains the
+    #            ability to remember sync state through crashes/reboots.
+    #            Using this policy for a mirror device is slow, as it
+    #            requires the mirror to resynchronize the devices, but it
+    #            will preserve the mirror characteristic of the device.
+    #            This policy acts like "remove" if no suitable device and
+    #            space can be allocated for the replacement.
+    #
+    # "allocate_anywhere" - Not yet implemented. Useful to place the log device
+    #            temporarily on same physical volume as one of the mirror
+    #            images. This policy is not recommended for mirror devices
+    #            since it would break the redundant nature of the mirror. This
+    #            policy acts like "remove" if no suitable device and space can
+    #            be allocated for the replacement.
+
+    mirror_log_fault_policy = "allocate"
+    mirror_device_fault_policy = "remove"
 }
 
 
@@ -291,7 +390,23 @@ activation {
 
 # Event daemon
 #
-#dmeventd {
+# dmeventd {
+    # mirror_library is the library used when monitoring a mirror device.
+    #
+    # "libdevmapper-event-lvm2mirror.so" attempts to recover from
+    # failures.  It removes failed devices from a volume group and
+    # reconfigures a mirror as necessary. If no mirror library is
+    # provided, mirrors are not monitored through dmeventd.
+
     # mirror_library = "libdevmapper-event-lvm2mirror.so"
+
+    # snapshot_library is the library used when monitoring a snapshot device.
+    #
+    # "libdevmapper-event-lvm2snapshot.so" monitors the filling of
+    # snapshots and emits a warning through syslog, when the use of
+    # snapshot exceedes 80%. The warning is repeated when 85%, 90% and
+    # 95% of the snapshot are filled.
+
+    # snapshot_library = "libdevmapper-event-lvm2snapshot.so"
 #}
 

doc/testing.txt

@@ -22,8 +22,8 @@ LVM2 that is running the LV's on my development box.
    }
 
 
-   The important this to note is the devices section which makes sure that
-   only the loopback devices are considered for LVM2 operations.
+   The important thing to note is the devices section which makes sure
+   that only the loopback devices are considered for LVM2 operations.
 
 4) When you want to use this test setup just set the environment
    variable LVM_SYSTEM_DIR to point to your config directory
@@ -39,8 +39,3 @@ LVM2 that is running the LV's on my development box.
 7) Test away.  Make sure that you are explicit about which lvm
    executable you want to execute (eg, ./lvm if you are in
    LVM2/tools).
-
-   
-
-
-

include/.symlinks

@@ -1,4 +1,6 @@
 ../daemons/clvmd/clvm.h
+../daemons/dmeventd/libdevmapper-event.h
+../liblvm/lvm.h
 ../lib/activate/activate.h
 ../lib/activate/targets.h
 ../lib/cache/lvmcache.h
@@ -7,7 +9,6 @@
 ../lib/config/config.h
 ../lib/config/defaults.h
 ../lib/datastruct/btree.h
-../lib/datastruct/list.h
 ../lib/datastruct/lvm-types.h
 ../lib/datastruct/str_list.h
 ../lib/device/dev-cache.h
@@ -28,21 +29,32 @@
 ../lib/label/label.h
 ../lib/locking/locking.h
 ../lib/log/log.h
+../lib/log/lvm-logging.h
 ../lib/metadata/lv_alloc.h
 ../lib/metadata/metadata.h
+../lib/metadata/metadata-exported.h
 ../lib/metadata/pv_alloc.h
 ../lib/metadata/segtype.h
 ../lib/mm/memlock.h
 ../lib/mm/xlate.h
+../lib/misc/configure.h
 ../lib/misc/crc.h
 ../lib/misc/intl.h
+../lib/misc/util.h
+../lib/misc/last-path-component.h
 ../lib/misc/lib.h
 ../lib/misc/lvm-exec.h
 ../lib/misc/lvm-file.h
+../lib/misc/lvm-globals.h
 ../lib/misc/lvm-string.h
+../lib/misc/lvm-version.h
+../lib/misc/lvm-wrappers.h
 ../lib/misc/sharedlib.h
-../lib/regex/matcher.h
 ../lib/report/report.h
 ../lib/uuid/uuid.h
+../libdm/libdevmapper.h
+../libdm/misc/dm-ioctl.h
+../libdm/misc/dm-logging.h
+../libdm/misc/dmlib.h
+../libdm/misc/kdev_t.h
 ../po/pogen.h
-../tools/version.h

include/Makefile.in

@@ -2,7 +2,7 @@
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
 # Copyright (C) 2004 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -20,7 +20,7 @@ VPATH = @srcdir@
 
 LN_S = @LN_S@
 
-.PHONY: clean distclean all install pofile install_cluster
+.PHONY: clean distclean all install pofile install_cluster install_device-mapper
 
 all: .symlinks_created
 
@@ -35,9 +35,17 @@ distclean:
 
 pofile: all
 
+device-mapper: all
+
 clean:
 
 install:
 
 install_cluster:
 
+install_device-mapper:
+
+install_lvm2:
+
+cflow:
+

lib/Makefile.in

@@ -1,8 +1,8 @@
 #
 # Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
-# Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+# Copyright (C) 2004-2009 Red Hat, Inc. All rights reserved.
 #
-# This file is part of the LVM2.
+# 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
@@ -42,6 +42,7 @@ SOURCES =\
 	device/dev-cache.c \
 	device/dev-io.c \
 	device/dev-md.c \
+	device/dev-swap.c \
 	device/device.c \
 	display/display.c \
 	error/errseg.c \
@@ -60,6 +61,7 @@ SOURCES =\
 	format_text/import_vsn1.c \
 	format_text/tags.c \
 	format_text/text_label.c \
+	freeseg/freeseg.c \
 	label/label.c \
 	locking/file_locking.c \
 	locking/locking.c \
@@ -76,11 +78,12 @@ SOURCES =\
 	misc/crc.c \
 	misc/lvm-exec.c \
 	misc/lvm-file.c \
+	misc/lvm-globals.c \
 	misc/lvm-string.c \
+	misc/lvm-wrappers.c \
+	misc/timestamp.c \
+	misc/util.c \
 	mm/memlock.c \
-	regex/matcher.c \
-	regex/parse_rx.c \
-	regex/ttree.c \
 	report/report.c \
 	striped/striped.c \
 	uuid/uuid.c \
@@ -137,9 +140,19 @@ ifeq ("@DMEVENTD@", "yes")
   CLDFLAGS += -ldevmapper-event
 endif
 
-LIB_STATIC = liblvm.a
+LIB_NAME = liblvm-internal
+LIB_STATIC = $(LIB_NAME).a
 
-$(SUBDIRS): $(LIB_STATIC)
+CLEAN_TARGETS += $(LIB_NAME).cflow
 
 include $(top_srcdir)/make.tmpl
 
+$(SUBDIRS): $(LIB_STATIC)
+
+$(LIB_NAME).cflow: $(SOURCES)
+	set -e; (echo -n "SOURCES += "; \
+		 echo $(SOURCES) | \
+		 sed "s/^/ /;s/ / $(top_srcdir)\/lib\//g;s/$$//"; \
+		 ) > $@
+
+cflow: $(LIB_NAME).cflow

lib/activate/activate.c

@@ -1,14 +1,14 @@
 /*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
+ * Copyright (C) 2004-2009 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 General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -39,7 +39,7 @@ int lvm1_present(struct cmd_context *cmd)
 {
 	char path[PATH_MAX];
 
-	if (lvm_snprintf(path, sizeof(path), "%s/lvm/global", cmd->proc_dir)
+	if (dm_snprintf(path, sizeof(path), "%s/lvm/global", cmd->proc_dir)
 	    < 0) {
 		log_error("LVM1 proc global snprintf failed");
 		return 0;
@@ -51,6 +51,66 @@ int lvm1_present(struct cmd_context *cmd)
 		return 0;
 }
 
+int list_segment_modules(struct dm_pool *mem, const struct lv_segment *seg,
+			 struct dm_list *modules)
+{
+	unsigned int s;
+	struct lv_segment *seg2, *snap_seg;
+	struct dm_list *snh;
+
+	if (seg->segtype->ops->modules_needed &&
+	    !seg->segtype->ops->modules_needed(mem, seg, modules)) {
+		log_error("module string allocation failed");
+		return 0;
+	}
+
+	if (lv_is_origin(seg->lv))
+		dm_list_iterate(snh, &seg->lv->snapshot_segs)
+			if (!list_lv_modules(mem,
+					     dm_list_struct_base(snh,
+							      struct lv_segment,
+							      origin_list)->cow,
+					     modules))
+				return_0;
+
+	if (lv_is_cow(seg->lv)) {
+		snap_seg = find_cow(seg->lv);
+		if (snap_seg->segtype->ops->modules_needed &&
+		    !snap_seg->segtype->ops->modules_needed(mem, snap_seg,
+							    modules)) {
+			log_error("snap_seg module string allocation failed");
+			return 0;
+		}
+	}
+
+	for (s = 0; s < seg->area_count; s++) {
+		switch (seg_type(seg, s)) {
+		case AREA_LV:
+			seg2 = find_seg_by_le(seg_lv(seg, s), seg_le(seg, s));
+			if (seg2 && !list_segment_modules(mem, seg2, modules))
+				return_0;
+			break;
+		case AREA_PV:
+		case AREA_UNASSIGNED:
+			;
+		}
+	}
+
+	return 1;
+}
+
+int list_lv_modules(struct dm_pool *mem, const struct logical_volume *lv,
+		    struct dm_list *modules)
+{
+	struct lv_segment *seg;
+
+	dm_list_iterate_items(seg, &lv->segments)
+		if (!list_segment_modules(mem, seg, modules))
+			return_0;
+
+	return 1;
+}
+
 #ifndef DEVMAPPER_SUPPORT
 void set_activation(int act)
 {
@@ -81,26 +141,27 @@ int target_version(const char *target_name, uint32_t *maj,
 {
 	return 0;
 }
-int target_present(const char *target_name)
+int target_present(struct cmd_context *cmd, const char *target_name,
+		   int use_modprobe)
 {
 	return 0;
 }
 int lv_info(struct cmd_context *cmd, const struct logical_volume *lv, struct lvinfo *info,
-	    int with_open_count)
+	    int with_open_count, int with_read_ahead)
 {
 	return 0;
 }
 int lv_info_by_lvid(struct cmd_context *cmd, const char *lvid_s,
-		    struct lvinfo *info, int with_open_count)
+		    struct lvinfo *info, int with_open_count, int with_read_ahead)
 {
 	return 0;
 }
-int lv_snapshot_percent(struct logical_volume *lv, float *percent)
+int lv_snapshot_percent(const struct logical_volume *lv, float *percent)
 {
 	return 0;
 }
-int lv_mirror_percent(struct logical_volume *lv, int wait, float *percent,
-		      uint32_t *event_nr)
+int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv,
+		      int wait, float *percent, uint32_t *event_nr)
 {
 	return 0;
 }
@@ -108,6 +169,10 @@ int lvs_in_vg_activated(struct volume_group *vg)
 {
 	return 0;
 }
+int lvs_in_vg_activated_by_uuid_only(struct volume_group *vg)
+{
+	return 0;
+}
 int lvs_in_vg_opened(struct volume_group *vg)
 {
 	return 0;
@@ -151,12 +216,17 @@ int lv_mknodes(struct cmd_context *cmd, const struct logical_volume *lv)
 	return 1;
 }
 
-int pv_uses_vg(struct cmd_context *cmd, struct physical_volume *pv,
-               struct volume_group *vg)
+int pv_uses_vg(struct physical_volume *pv,
+	       struct volume_group *vg)
 {
 	return 0;
 }
 
+void activation_release(void)
+{
+	return;
+}
+
 void activation_exit(void)
 {
 	return;
@@ -176,7 +246,7 @@ void set_activation(int act)
 		log_verbose("Activation enabled. Device-mapper kernel "
 			    "driver will be used.");
 	else
-		log_print("WARNING: Activation disabled. No device-mapper "
+		log_warn("WARNING: Activation disabled. No device-mapper "
 			  "interaction will be attempted.");
 }
 
@@ -193,9 +263,9 @@ static int _passes_activation_filter(struct cmd_context *cmd,
 	char *str;
 	char path[PATH_MAX];
 
-	if (!(cn = find_config_node(cmd->cft->root, "activation/volume_list"))) {
+	if (!(cn = find_config_tree_node(cmd, "activation/volume_list"))) {
 		/* If no host tags defined, activate */
-		if (list_empty(&cmd->tags))
+		if (dm_list_empty(&cmd->tags))
 			return 1;
 
 		/* If any host tag matches any LV or VG tag, activate */
@@ -244,7 +314,7 @@ static int _passes_activation_filter(struct cmd_context *cmd,
 			else
 				continue;
 		}
-		if (!index(str, '/')) {
+		if (!strchr(str, '/')) {
 			/* vgname supplied */
 			if (!strcmp(str, lv->vg->name))
 				return 1;
@@ -252,9 +322,9 @@ static int _passes_activation_filter(struct cmd_context *cmd,
 				continue;
 		}
 		/* vgname/lvname */
-		if (lvm_snprintf(path, sizeof(path), "%s/%s", lv->vg->name,
+		if (dm_snprintf(path, sizeof(path), "%s/%s", lv->vg->name,
 				 lv->name) < 0) {
-			log_error("lvm_snprintf error from %s/%s", lv->vg->name,
+			log_error("dm_snprintf error from %s/%s", lv->vg->name,
 				  lv->name);
 			continue;
 		}
@@ -322,12 +392,32 @@ int target_version(const char *target_name, uint32_t *maj,
 	return r;
 }
 
-int target_present(const char *target_name, int use_modprobe)
+int module_present(struct cmd_context *cmd, const char *target_name)
 {
-	uint32_t maj, min, patchlevel;
+	int ret = 0;
 #ifdef MODPROBE_CMD
 	char module[128];
+	const char *argv[3];
+
+	if (dm_snprintf(module, sizeof(module), "dm-%s", target_name) < 0) {
+		log_error("module_present module name too long: %s",
+			  target_name);
+		return 0;
+	}
+
+	argv[0] = MODPROBE_CMD;
+	argv[1] = module;
+	argv[2] = NULL;
+
+	ret = exec_cmd(cmd, argv);
 #endif
+	return ret;
+}
+
+int target_present(struct cmd_context *cmd, const char *target_name,
+		   int use_modprobe)
+{
+	uint32_t maj, min, patchlevel;
 
 	if (!activation())
 		return 0;
@@ -337,14 +427,7 @@ int target_present(const char *target_name, int use_modprobe)
 		if (target_version(target_name, &maj, &min, &patchlevel))
 			return 1;
 
-		if (lvm_snprintf(module, sizeof(module), "dm-%s", target_name)
-		    < 0) {
-			log_error("target_present module name too long: %s",
-				  target_name);
-			return 0;
-		}
-
-		if (!exec_cmd(MODPROBE_CMD, module, "", ""))
+		if (!module_present(cmd, target_name))
 			return_0;
 	}
 #endif
@@ -356,21 +439,24 @@ int target_present(const char *target_name, int use_modprobe)
  * Returns 1 if info structure populated, else 0 on failure.
  */
 static int _lv_info(struct cmd_context *cmd, const struct logical_volume *lv, int with_mknodes,
-		    struct lvinfo *info, int with_open_count)
+		    struct lvinfo *info, int with_open_count, int with_read_ahead, unsigned by_uuid_only)
 {
 	struct dm_info dminfo;
-	char *name;
+	char *name = NULL;
 
 	if (!activation())
 		return 0;
 
-	if (!(name = build_dm_name(cmd->mem, lv->vg->name, lv->name, NULL)))
+	if (!by_uuid_only &&
+	    !(name = build_dm_name(cmd->mem, lv->vg->name, lv->name, NULL)))
 		return_0;
 
 	log_debug("Getting device info for %s", name);
 	if (!dev_manager_info(lv->vg->cmd->mem, name, lv, with_mknodes,
-			      with_open_count, &dminfo)) {
-		dm_pool_free(cmd->mem, name);
+			      with_open_count, with_read_ahead, &dminfo,
+			      &info->read_ahead)) {
+		if (name)
+			dm_pool_free(cmd->mem, name);
 		return_0;
 	}
 
@@ -383,31 +469,33 @@ static int _lv_info(struct cmd_context *cmd, const struct logical_volume *lv, in
 	info->live_table = dminfo.live_table;
 	info->inactive_table = dminfo.inactive_table;
 
-	dm_pool_free(cmd->mem, name);
+	if (name)
+		dm_pool_free(cmd->mem, name);
+
 	return 1;
 }
 
 int lv_info(struct cmd_context *cmd, const struct logical_volume *lv, struct lvinfo *info,
-	    int with_open_count)
+	    int with_open_count, int with_read_ahead)
 {
-	return _lv_info(cmd, lv, 0, info, with_open_count);
+	return _lv_info(cmd, lv, 0, info, with_open_count, with_read_ahead, 0);
 }
 
 int lv_info_by_lvid(struct cmd_context *cmd, const char *lvid_s,
-		    struct lvinfo *info, int with_open_count)
+		    struct lvinfo *info, int with_open_count, int with_read_ahead)
 {
 	struct logical_volume *lv;
 
 	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
 		return 0;
 
-	return _lv_info(cmd, lv, 0, info, with_open_count);
+	return _lv_info(cmd, lv, 0, info, with_open_count, with_read_ahead, 0);
 }
 
 /*
  * Returns 1 if percent set, else 0 on failure.
  */
-int lv_snapshot_percent(struct logical_volume *lv, float *percent)
+int lv_snapshot_percent(const struct logical_volume *lv, float *percent)
 {
 	int r;
 	struct dev_manager *dm;
@@ -427,17 +515,24 @@ int lv_snapshot_percent(struct logical_volume *lv, float *percent)
 }
 
 /* FIXME Merge with snapshot_percent */
-int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv, int wait, float *percent,
-		      uint32_t *event_nr)
+int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv,
+		      int wait, float *percent, uint32_t *event_nr)
 {
 	int r;
 	struct dev_manager *dm;
 	struct lvinfo info;
 
+	/* If mirrored LV is temporarily shrinked to 1 area (= linear),
+	 * it should be considered in-sync. */
+	if (dm_list_size(&lv->segments) == 1 && first_seg(lv)->area_count == 1) {
+		*percent = 100.0;
+		return 1;
+	}
+
 	if (!activation())
 		return 0;
 
-	if (!lv_info(cmd, lv, &info, 0))
+	if (!lv_info(cmd, lv, &info, 0, 0))
 		return_0;
 
 	if (!info.exists)
@@ -454,11 +549,12 @@ int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv, int wa
 	return r;
 }
 
-static int _lv_active(struct cmd_context *cmd, struct logical_volume *lv)
+static int _lv_active(struct cmd_context *cmd, struct logical_volume *lv,
+		      unsigned by_uuid_only)
 {
 	struct lvinfo info;
 
-	if (!lv_info(cmd, lv, &info, 0)) {
+	if (!_lv_info(cmd, lv, 0, &info, 0, 0, by_uuid_only)) {
 		stack;
 		return -1;
 	}
@@ -470,7 +566,7 @@ static int _lv_open_count(struct cmd_context *cmd, struct logical_volume *lv)
 {
 	struct lvinfo info;
 
-	if (!lv_info(cmd, lv, &info, 1)) {
+	if (!lv_info(cmd, lv, &info, 1, 0)) {
 		stack;
 		return -1;
 	}
@@ -493,7 +589,7 @@ static int _lv_activate_lv(struct logical_volume *lv)
 	return r;
 }
 
-static int _lv_preload(struct logical_volume *lv)
+static int _lv_preload(struct logical_volume *lv, int *flush_required)
 {
 	int r;
 	struct dev_manager *dm;
@@ -501,7 +597,7 @@ static int _lv_preload(struct logical_volume *lv)
 	if (!(dm = dev_manager_create(lv->vg->cmd, lv->vg->name)))
 		return_0;
 
-	if (!(r = dev_manager_preload(dm, lv)))
+	if (!(r = dev_manager_preload(dm, lv, flush_required)))
 		stack;
 
 	dev_manager_destroy(dm);
@@ -523,7 +619,7 @@ static int _lv_deactivate(struct logical_volume *lv)
 	return r;
 }
 
-static int _lv_suspend_lv(struct logical_volume *lv)
+static int _lv_suspend_lv(struct logical_volume *lv, int lockfs, int flush_required)
 {
 	int r;
 	struct dev_manager *dm;
@@ -531,7 +627,7 @@ static int _lv_suspend_lv(struct logical_volume *lv)
 	if (!(dm = dev_manager_create(lv->vg->cmd, lv->vg->name)))
 		return_0;
 
-	if (!(r = dev_manager_suspend(dm, lv)))
+	if (!(r = dev_manager_suspend(dm, lv, lockfs, flush_required)))
 		stack;
 
 	dev_manager_destroy(dm);
@@ -542,7 +638,7 @@ static int _lv_suspend_lv(struct logical_volume *lv)
  * These two functions return the number of visible LVs in the state,
  * or -1 on error.
  */
-int lvs_in_vg_activated(struct volume_group *vg)
+static int _lvs_in_vg_activated(struct volume_group *vg, unsigned by_uuid_only)
 {
 	struct lv_list *lvl;
 	int count = 0;
@@ -550,108 +646,267 @@ int lvs_in_vg_activated(struct volume_group *vg)
 	if (!activation())
 		return 0;
 
-	list_iterate_items(lvl, &vg->lvs) {
-		if (lvl->lv->status & VISIBLE_LV)
-			count += (_lv_active(vg->cmd, lvl->lv) == 1);
+	dm_list_iterate_items(lvl, &vg->lvs) {
+		if (lv_is_visible(lvl->lv))
+			count += (_lv_active(vg->cmd, lvl->lv, by_uuid_only) == 1);
 	}
 
 	return count;
 }
 
-int lvs_in_vg_opened(struct volume_group *vg)
+int lvs_in_vg_activated_by_uuid_only(struct volume_group *vg)
 {
-	struct lv_list *lvl;
+	return _lvs_in_vg_activated(vg, 1);
+}
+
+int lvs_in_vg_activated(struct volume_group *vg)
+{
+	return _lvs_in_vg_activated(vg, 0);
+}
+
+int lvs_in_vg_opened(const struct volume_group *vg)
+{
+	const struct lv_list *lvl;
 	int count = 0;
 
 	if (!activation())
 		return 0;
 
-	list_iterate_items(lvl, &vg->lvs) {
-		if (lvl->lv->status & VISIBLE_LV)
+	dm_list_iterate_items(lvl, &vg->lvs) {
+		if (lv_is_visible(lvl->lv))
 			count += (_lv_open_count(vg->cmd, lvl->lv) > 0);
 	}
 
 	return count;
 }
 
-static int _register_dev_for_events(struct cmd_context *cmd,
-				    struct logical_volume *lv, int do_reg)
+/*
+ * Determine whether an LV is active locally or in a cluster.
+ * Assumes vg lock held.
+ * Returns:
+ * 0 - not active locally or on any node in cluster
+ * 1 - active either locally or some node in the cluster
+ */
+int lv_is_active(struct logical_volume *lv)
 {
-#ifdef DMEVENTD
-	struct list *tmp;
-	struct lv_segment *seg;
-	int (*reg) (struct dm_pool *mem, struct lv_segment *,
-		    struct config_tree *cft, int events);
+	int ret;
 
-	list_iterate(tmp, &lv->segments) {
-		seg = list_item(tmp, struct lv_segment);
+	if (_lv_active(lv->vg->cmd, lv, 0))
+		return 1;
 
-		reg = NULL;
+	if (!vg_is_clustered(lv->vg))
+		return 0;
 
-		if (do_reg) {
-			if (seg->segtype->ops->target_register_events)
-				reg = seg->segtype->ops->target_register_events;
-		} else if (seg->segtype->ops->target_unregister_events)
-			reg = seg->segtype->ops->target_unregister_events;
+	if ((ret = remote_lock_held(lv->lvid.s)) >= 0)
+		return ret;
 
-		if (reg)
-			/* FIXME specify events */
-			if (!reg(cmd->mem, seg, cmd->cft, 0)) {
-				stack;
-				return 0;
-			}
+	/*
+	 * Old compatibility code if locking doesn't support lock query
+	 * FIXME: check status to not deactivate already activate device
+	 */
+	if (activate_lv_excl(lv->vg->cmd, lv)) {
+		deactivate_lv(lv->vg->cmd, lv);
+		return 0;
 	}
 
-#endif
+	/*
+	 * Exclusive local activation failed so assume it is active elsewhere.
+	 */
 	return 1;
 }
 
+/*
+ * Returns 0 if an attempt to (un)monitor the device failed.
+ * Returns 1 otherwise.
+ */
+int monitor_dev_for_events(struct cmd_context *cmd,
+			    struct logical_volume *lv, int monitor)
+{
+#ifdef DMEVENTD
+	int i, pending = 0, monitored;
+	int r = 1;
+	struct dm_list *tmp, *snh, *snht;
+	struct lv_segment *seg;
+	int (*monitor_fn) (struct lv_segment *s, int e);
+	uint32_t s;
+
+	/* skip dmeventd code altogether */
+	if (dmeventd_monitor_mode() == DMEVENTD_MONITOR_IGNORE)
+		return 1;
+
+	/*
+	 * Nothing to do if dmeventd configured not to be used.
+	 */
+	if (monitor && !dmeventd_monitor_mode())
+		return 1;
+
+	/*
+	 * In case of a snapshot device, we monitor lv->snapshot->lv,
+	 * not the actual LV itself.
+	 */
+	if (lv_is_cow(lv))
+		return monitor_dev_for_events(cmd, lv->snapshot->lv, monitor);
+
+	/*
+	 * In case this LV is a snapshot origin, we instead monitor
+	 * each of its respective snapshots (the origin itself does
+	 * not need to be monitored).
+	 *
+	 * TODO: This may change when snapshots of mirrors are allowed.
+	 */
+	if (lv_is_origin(lv)) {
+		dm_list_iterate_safe(snh, snht, &lv->snapshot_segs)
+			if (!monitor_dev_for_events(cmd, dm_list_struct_base(snh,
+				    struct lv_segment, origin_list)->cow, monitor))
+				r = 0;
+		return r;
+	}
+
+	dm_list_iterate(tmp, &lv->segments) {
+		seg = dm_list_item(tmp, struct lv_segment);
+
+		/* Recurse for AREA_LV */
+		for (s = 0; s < seg->area_count; s++) {
+			if (seg_type(seg, s) != AREA_LV)
+				continue;
+			if (!monitor_dev_for_events(cmd, seg_lv(seg, s),
+						    monitor)) {
+				log_error("Failed to %smonitor %s",
+					  monitor ? "" : "un",
+					  seg_lv(seg, s)->name);
+				r = 0;
+			}
+		}
+
+		if (!seg_monitored(seg) || (seg->status & PVMOVE))
+			continue;
+
+		monitor_fn = NULL;
+
+		/* Check monitoring status */
+		if (seg->segtype->ops->target_monitored)
+			monitored = seg->segtype->ops->target_monitored(seg, &pending);
+		else
+			continue;  /* segtype doesn't support registration */
+
+		/*
+		 * FIXME: We should really try again if pending
+		 */
+		monitored = (pending) ? 0 : monitored;
+
+		if (monitor) {
+			if (monitored)
+				log_verbose("%s/%s already monitored.", lv->vg->name, lv->name);
+			else if (seg->segtype->ops->target_monitor_events)
+				monitor_fn = seg->segtype->ops->target_monitor_events;
+		} else {
+			if (!monitored)
+				log_verbose("%s/%s already not monitored.", lv->vg->name, lv->name);
+			else if (seg->segtype->ops->target_unmonitor_events)
+				monitor_fn = seg->segtype->ops->target_unmonitor_events;
+		}
+
+		/* Do [un]monitor */
+		if (!monitor_fn)
+			continue;
+
+		log_verbose("%sonitoring %s/%s", monitor ? "M" : "Not m", lv->vg->name, lv->name);
+
+		/* FIXME specify events */
+		if (!monitor_fn(seg, 0)) {
+			log_error("%s/%s: %s segment monitoring function failed.",
+				  lv->vg->name, lv->name, seg->segtype->name);
+			return 0;
+		}
+
+		/* Check [un]monitor results */
+		/* Try a couple times if pending, but not forever... */
+		for (i = 0; i < 10; i++) {
+			pending = 0;
+			monitored = seg->segtype->ops->target_monitored(seg, &pending);
+			if (pending ||
+			    (!monitored && monitor) ||
+			    (monitored && !monitor))
+				log_very_verbose("%s/%s %smonitoring still pending: waiting...",
+						 lv->vg->name, lv->name, monitor ? "" : "un");
+			else
+				break;
+			sleep(1);
+		}
+
+		r = (monitored && monitor) || (!monitored && !monitor);
+	}
+
+	return r;
+#else
+	return 1;
+#endif
+}
+
 static int _lv_suspend(struct cmd_context *cmd, const char *lvid_s,
 		       int error_if_not_suspended)
 {
-	struct logical_volume *lv, *lv_pre;
+	struct logical_volume *lv = NULL, *lv_pre = NULL;
 	struct lvinfo info;
+	int r = 0, lockfs = 0, flush_required = 0;
 
 	if (!activation())
 		return 1;
 
 	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
-		return_0;
+		goto_out;
 
 	/* Use precommitted metadata if present */
 	if (!(lv_pre = lv_from_lvid(cmd, lvid_s, 1)))
-		return_0;
+		goto_out;
 
 	if (test_mode()) {
 		_skip("Suspending '%s'.", lv->name);
-		return 1;
+		r = 1;
+		goto out;
 	}
 
-	if (!lv_info(cmd, lv, &info, 0))
-		return_0;
+	if (!lv_info(cmd, lv, &info, 0, 0))
+		goto_out;
 
-	if (!info.exists || info.suspended)
-		return error_if_not_suspended ? 0 : 1;
+	if (!info.exists || info.suspended) {
+		r = error_if_not_suspended ? 0 : 1;
+		goto out;
+	}
+
+	lv_calculate_readahead(lv, NULL);
 
 	/* If VG was precommitted, preload devices for the LV */
 	if ((lv_pre->vg->status & PRECOMMITTED)) {
-		if (!_lv_preload(lv_pre)) {
+		if (!_lv_preload(lv_pre, &flush_required)) {
 			/* FIXME Revert preloading */
-			return_0;
+			goto_out;
 		}
 	}
 
-	if (!_register_dev_for_events(cmd, lv, 0))
+	if (!monitor_dev_for_events(cmd, lv, 0))
+		/* FIXME Consider aborting here */
 		stack;
 
 	memlock_inc();
-	if (!_lv_suspend_lv(lv)) {
+
+	if (lv_is_origin(lv_pre) || lv_is_cow(lv_pre))
+		lockfs = 1;
+
+	if (!_lv_suspend_lv(lv, lockfs, flush_required)) {
 		memlock_dec();
 		fs_unlock();
-		return 0;
+		goto out;
 	}
 
-	return 1;
+	r = 1;
+out:
+	if (lv_pre)
+		vg_release(lv_pre->vg);
+	if (lv)
+		vg_release(lv->vg);
+
+	return r;
 }
 
 /* Returns success if the device is not active */
@@ -670,34 +925,43 @@ static int _lv_resume(struct cmd_context *cmd, const char *lvid_s,
 {
 	struct logical_volume *lv;
 	struct lvinfo info;
+	int r = 0;
 
 	if (!activation())
 		return 1;
 
 	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
-		return 0;
+		goto out;
 
 	if (test_mode()) {
 		_skip("Resuming '%s'.", lv->name);
-		return 1;
+		r = 1;
+		goto out;
 	}
 
-	if (!lv_info(cmd, lv, &info, 0))
-		return_0;
+	if (!lv_info(cmd, lv, &info, 0, 0))
+		goto_out;
 
-	if (!info.exists || !info.suspended)
-		return error_if_not_active ? 0 : 1;
+	if (!info.exists || !info.suspended) {
+		r = error_if_not_active ? 0 : 1;
+		goto out;
+	}
 
 	if (!_lv_activate_lv(lv))
-		return 0;
+		goto out;
 
 	memlock_dec();
 	fs_unlock();
 
-	if (!_register_dev_for_events(cmd, lv, 1))
+	if (!monitor_dev_for_events(cmd, lv, 1))
 		stack;
 
-	return 1;
+	r = 1;
+out:
+	if (lv)
+		vg_release(lv->vg);
+
+	return r;
 }
 
 /* Returns success if the device is not active */
@@ -715,32 +979,37 @@ int lv_deactivate(struct cmd_context *cmd, const char *lvid_s)
 {
 	struct logical_volume *lv;
 	struct lvinfo info;
-	int r;
+	int r = 0;
 
 	if (!activation())
 		return 1;
 
 	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
-		return 0;
+		goto out;
 
 	if (test_mode()) {
 		_skip("Deactivating '%s'.", lv->name);
-		return 1;
+		r = 1;
+		goto out;
 	}
 
-	if (!lv_info(cmd, lv, &info, 1))
-		return_0;
+	if (!lv_info(cmd, lv, &info, 1, 0))
+		goto_out;
 
-	if (!info.exists)
-		return 1;
+	if (!info.exists) {
+		r = 1;
+		goto out;
+	}
 
-	if (info.open_count && (lv->status & VISIBLE_LV)) {
+	if (info.open_count && lv_is_visible(lv)) {
 		log_error("LV %s/%s in use: not deactivating", lv->vg->name,
 			  lv->name);
-		return 0;
+		goto out;
 	}
 
-	if (!_register_dev_for_events(cmd, lv, 0))
+	lv_calculate_readahead(lv, NULL);
+
+	if (!monitor_dev_for_events(cmd, lv, 0))
 		stack;
 
 	memlock_inc();
@@ -748,6 +1017,10 @@ int lv_deactivate(struct cmd_context *cmd, const char *lvid_s)
 	memlock_dec();
 	fs_unlock();
 
+out:
+	if (lv)
+		vg_release(lv->vg);
+
 	return r;
 }
 
@@ -756,23 +1029,28 @@ int lv_activation_filter(struct cmd_context *cmd, const char *lvid_s,
 			 int *activate_lv)
 {
 	struct logical_volume *lv;
+	int r = 0;
 
-	if (!activation())
-		goto activate;
+	if (!activation()) {
+		*activate_lv = 1;
+		return 1;
+	}
 
 	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
-		return 0;
+		goto out;
 
 	if (!_passes_activation_filter(cmd, lv)) {
 		log_verbose("Not activating %s/%s due to config file settings",
 			    lv->vg->name, lv->name);
 		*activate_lv = 0;
-		return 1;
-	}
+	} else
+		*activate_lv = 1;
+	r = 1;
+out:
+	if (lv)
+		vg_release(lv->vg);
 
-      activate:
-	*activate_lv = 1;
-	return 1;
+	return r;
 }
 
 static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
@@ -780,30 +1058,41 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
 {
 	struct logical_volume *lv;
 	struct lvinfo info;
-	int r;
+	int r = 0;
 
 	if (!activation())
 		return 1;
 
 	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
-		return 0;
+		goto out;
 
 	if (filter && !_passes_activation_filter(cmd, lv)) {
 		log_verbose("Not activating %s/%s due to config file settings",
 			    lv->vg->name, lv->name);
-		return 0;
+		goto out;
+	}
+
+	if ((!lv->vg->cmd->partial_activation) && (lv->status & PARTIAL_LV)) {
+		log_error("Refusing activation of partial LV %s. Use --partial to override.",
+			  lv->name);
+		goto_out;
 	}
 
 	if (test_mode()) {
 		_skip("Activating '%s'.", lv->name);
-		return 1;
+		r = 1;
+		goto out;
 	}
 
-	if (!lv_info(cmd, lv, &info, 0))
-		return_0;
+	if (!lv_info(cmd, lv, &info, 0, 0))
+		goto_out;
 
-	if (info.exists && !info.suspended && info.live_table)
-		return 1;
+	if (info.exists && !info.suspended && info.live_table) {
+		r = 1;
+		goto out;
+	}
+
+	lv_calculate_readahead(lv, NULL);
 
 	if (exclusive)
 		lv->status |= ACTIVATE_EXCL;
@@ -813,9 +1102,13 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
 	memlock_dec();
 	fs_unlock();
 
-	if (!_register_dev_for_events(cmd, lv, 1))
+	if (r && !monitor_dev_for_events(cmd, lv, 1))
 		stack;
 
+out:
+	if (lv)
+		vg_release(lv->vg);
+
 	return r;
 }
 
@@ -842,12 +1135,13 @@ int lv_mknodes(struct cmd_context *cmd, const struct logical_volume *lv)
 		return r;
 	}
 
-	if (!_lv_info(cmd, lv, 1, &info, 0))
+	if (!_lv_info(cmd, lv, 1, &info, 0, 0, 0))
 		return_0;
 
-	if (info.exists)
-		r = dev_manager_lv_mknodes(lv);
-	else
+	if (info.exists) {
+		if (lv_is_visible(lv))
+			r = dev_manager_lv_mknodes(lv);
+	} else
 		r = dev_manager_lv_rmnodes(lv);
 
 	fs_unlock();
@@ -859,28 +1153,21 @@ int lv_mknodes(struct cmd_context *cmd, const struct logical_volume *lv)
  * Does PV use VG somewhere in its construction?
  * Returns 1 on failure.
  */
-int pv_uses_vg(struct cmd_context *cmd, struct physical_volume *pv,
-               struct volume_group *vg)
+int pv_uses_vg(struct physical_volume *pv,
+	       struct volume_group *vg)
 {
-	struct dev_manager *dm;
-	int r;
-
 	if (!activation())
 		return 0;
 
 	if (!dm_is_dm_major(MAJOR(pv->dev->dev)))
 		return 0;
 
-	if (!(dm = dev_manager_create(cmd, vg->name))) {
-		stack;
-		return 1;
-	}
+	return dev_manager_device_uses_vg(pv->dev, vg);
+}
 
-	r = dev_manager_device_uses_vg(dm, pv->dev, vg);
-
-	dev_manager_destroy(dm);
-
-	return r;
+void activation_release(void)
+{
+	dev_manager_release();
 }
 
 void activation_exit(void)

lib/activate/activate.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -16,7 +16,7 @@
 #ifndef LVM_ACTIVATE_H
 #define LVM_ACTIVATE_H
 
-#include "metadata.h"
+#include "metadata-exported.h"
 
 struct lvinfo {
 	int exists;
@@ -27,8 +27,12 @@ struct lvinfo {
 	int read_only;
 	int live_table;
 	int inactive_table;
+	uint32_t read_ahead;
 };
 
+/* target attribute flags */
+#define MIRROR_LOG_CLUSTERED	0x00000001U
+
 void set_activation(int activation);
 int activation(void);
 
@@ -36,10 +40,17 @@ int driver_version(char *version, size_t size);
 int library_version(char *version, size_t size);
 int lvm1_present(struct cmd_context *cmd);
 
-int target_present(const char *target_name, int use_modprobe);
+int module_present(struct cmd_context *cmd, const char *target_name);
+int target_present(struct cmd_context *cmd, const char *target_name,
+		   int use_modprobe);
 int target_version(const char *target_name, uint32_t *maj,
                    uint32_t *min, uint32_t *patchlevel);
+int list_segment_modules(struct dm_pool *mem, const struct lv_segment *seg,
+			 struct dm_list *modules);
+int list_lv_modules(struct dm_pool *mem, const struct logical_volume *lv,
+		    struct dm_list *modules);
 
+void activation_release(void);
 void activation_exit(void);
 
 int lv_suspend(struct cmd_context *cmd, const char *lvid_s);
@@ -57,9 +68,9 @@ int lv_mknodes(struct cmd_context *cmd, const struct logical_volume *lv);
  * Returns 1 if info structure has been populated, else 0.
  */
 int lv_info(struct cmd_context *cmd, const struct logical_volume *lv, struct lvinfo *info,
-	    int with_open_count);
+	    int with_open_count, int with_read_ahead);
 int lv_info_by_lvid(struct cmd_context *cmd, const char *lvid_s,
-		    struct lvinfo *info, int with_open_count);
+		    struct lvinfo *info, int with_open_count, int with_read_ahead);
 
 /*
  * Returns 1 if activate_lv has been set: 1 = activate; 0 = don't.
@@ -70,20 +81,31 @@ int lv_activation_filter(struct cmd_context *cmd, const char *lvid_s,
 /*
  * Returns 1 if percent has been set, else 0.
  */
-int lv_snapshot_percent(struct logical_volume *lv, float *percent);
-int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv, int wait, float *percent,
-		      uint32_t *event_nr);
+int lv_snapshot_percent(const struct logical_volume *lv, float *percent);
+int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv,
+		      int wait, float *percent, uint32_t *event_nr);
 
 /*
  * Return number of LVs in the VG that are active.
  */
 int lvs_in_vg_activated(struct volume_group *vg);
-int lvs_in_vg_opened(struct volume_group *vg);
+int lvs_in_vg_activated_by_uuid_only(struct volume_group *vg);
+int lvs_in_vg_opened(const struct volume_group *vg);
+
+int lv_is_active(struct logical_volume *lv);
+
+int monitor_dev_for_events(struct cmd_context *cmd,
+			    struct logical_volume *lv, int do_reg);
 
 /*
  * Returns 1 if PV has a dependency tree that uses anything in VG.
  */
-int pv_uses_vg(struct cmd_context *cmd, struct physical_volume *pv,
+int pv_uses_vg(struct physical_volume *pv,
 	       struct volume_group *vg);
 
+/*
+ * Returns 1 if mapped device is not suspended.
+ */
+int device_is_usable(dev_t dev);
+
 #endif

lib/activate/dev_manager.c

@@ -1,14 +1,14 @@
 /*
- * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
+ * 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -25,6 +25,7 @@
 #include "targets.h"
 #include "config.h"
 #include "filter.h"
+#include "activate.h"
 
 #include <limits.h>
 #include <dirent.h>
@@ -37,6 +38,7 @@ typedef enum {
 	ACTIVATE,
 	DEACTIVATE,
 	SUSPEND,
+	SUSPEND_WITH_LOCKFS,
 	CLEAN
 } action_t;
 
@@ -45,9 +47,9 @@ struct dev_manager {
 
 	struct cmd_context *cmd;
 
-	const char *stripe_filler;
 	void *target_state;
 	uint32_t pvmove_mirror_count;
+	int flush_required;
 
 	char *vg_name;
 };
@@ -57,8 +59,6 @@ struct lv_layer {
 	const char *old_name;
 };
 
-static const char *stripe_filler = NULL;
-
 static char *_build_dlid(struct dm_pool *mem, const char *lvid, const char *layer)
 {
 	char *dlid;
@@ -85,7 +85,7 @@ char *build_dlid(struct dev_manager *dm, const char *lvid, const char *layer)
 	return _build_dlid(dm->mem, lvid, layer);
 }
 
-static inline int _read_only_lv(struct logical_volume *lv)
+static int _read_only_lv(struct logical_volume *lv)
 {
 	return (!(lv->vg->status & LVM_WRITE) || !(lv->status & LVM_WRITE));
 }
@@ -94,14 +94,13 @@ static inline int _read_only_lv(struct logical_volume *lv)
  * Low level device-layer operations.
  */
 static struct dm_task *_setup_task(const char *name, const char *uuid,
-				   uint32_t *event_nr, int task)
+				   uint32_t *event_nr, int task,
+				   uint32_t major, uint32_t minor)
 {
 	struct dm_task *dmt;
 
-	if (!(dmt = dm_task_create(task))) {
-		stack;
-		return NULL;
-	}
+	if (!(dmt = dm_task_create(task)))
+		return_NULL;
 
 	if (name)
 		dm_task_set_name(dmt, name);
@@ -112,24 +111,24 @@ static struct dm_task *_setup_task(const char *name, const char *uuid,
 	if (event_nr)
 		dm_task_set_event_nr(dmt, *event_nr);
 
+	if (major)
+		dm_task_set_major_minor(dmt, major, minor, 1);
+
 	return dmt;
 }
 
 static int _info_run(const char *name, const char *dlid, struct dm_info *info,
-		     int mknodes, int with_open_count, struct dm_pool *mem,
-		     char **uuid_out)
+		     uint32_t *read_ahead, int mknodes, int with_open_count,
+		     int with_read_ahead, uint32_t major, uint32_t minor)
 {
 	int r = 0;
 	struct dm_task *dmt;
-	const char *u;
 	int dmtask;
 
 	dmtask = mknodes ? DM_DEVICE_MKNODES : DM_DEVICE_INFO;
 
-	if (!(dmt = _setup_task(name, dlid, 0, dmtask))) {
-		stack;
-		return 0;
-	}
+	if (!(dmt = _setup_task(name, dlid, 0, dmtask, major, minor)))
+		return_0;
 
 	if (!with_open_count)
 		if (!dm_task_no_open_count(dmt))
@@ -141,11 +140,64 @@ static int _info_run(const char *name, const char *dlid, struct dm_info *info,
 	if (!dm_task_get_info(dmt, info))
 		goto_out;
 
-	if (info->exists && uuid_out) {
-		if (!(u = dm_task_get_uuid(dmt)))
+	if (with_read_ahead && info->exists) {
+		if (!dm_task_get_read_ahead(dmt, read_ahead))
 			goto_out;
-		*uuid_out = dm_pool_strdup(mem, u);
+	} else if (read_ahead)
+		*read_ahead = DM_READ_AHEAD_NONE;
+
+	r = 1;
+
+      out:
+	dm_task_destroy(dmt);
+	return r;
+}
+
+int device_is_usable(dev_t dev)
+{
+	struct dm_task *dmt;
+	struct dm_info info;
+	const char *name;
+	uint64_t start, length;
+	char *target_type = NULL;
+	char *params;
+	void *next = NULL;
+	int r = 0;
+
+	if (!(dmt = dm_task_create(DM_DEVICE_STATUS))) {
+		log_error("Failed to allocate dm_task struct to check dev status");
+		return 0;
 	}
+
+	if (!dm_task_set_major_minor(dmt, MAJOR(dev), MINOR(dev), 1))
+		goto_out;
+
+	if (!dm_task_run(dmt)) {
+		log_error("Failed to get state of mapped device");
+		goto out;
+	}
+
+	if (!dm_task_get_info(dmt, &info))
+		goto_out;
+
+	if (!info.exists || info.suspended)
+		goto out;
+
+	name = dm_task_get_name(dmt);
+
+	/* FIXME Also check for mirror block_on_error and mpath no paths */
+	/* For now, we exclude all mirrors */
+
+	do {
+		next = dm_get_next_target(dmt, next, &start, &length,
+					  &target_type, &params);
+		/* Skip if target type doesn't match */
+		if (target_type && !strcmp(target_type, "mirror"))
+			goto out;
+	} while (next);
+
+	/* FIXME Also check dependencies? */
+
 	r = 1;
 
       out:
@@ -154,30 +206,37 @@ static int _info_run(const char *name, const char *dlid, struct dm_info *info,
 }
 
 static int _info(const char *name, const char *dlid, int mknodes,
-		 int with_open_count, struct dm_info *info,
-		 struct dm_pool *mem, char **uuid_out)
+		 int with_open_count, int with_read_ahead,
+		 struct dm_info *info, uint32_t *read_ahead)
 {
 	if (!mknodes && dlid && *dlid) {
-		if (_info_run(NULL, dlid, info, 0, with_open_count, mem,
-			      uuid_out) &&
+		if (_info_run(NULL, dlid, info, read_ahead, 0, with_open_count,
+			      with_read_ahead, 0, 0) &&
 	    	    info->exists)
 			return 1;
 		else if (_info_run(NULL, dlid + sizeof(UUID_PREFIX) - 1, info,
-				   0, with_open_count, mem, uuid_out) &&
+				   read_ahead, 0, with_open_count,
+				   with_read_ahead, 0, 0) &&
 			 info->exists)
 			return 1;
 	}
 
 	if (name)
-		return _info_run(name, NULL, info, mknodes, with_open_count,
-				 mem, uuid_out);
+		return _info_run(name, NULL, info, read_ahead, mknodes,
+				 with_open_count, with_read_ahead, 0, 0);
 
 	return 0;
 }
 
+static int _info_by_dev(uint32_t major, uint32_t minor, struct dm_info *info)
+{
+	return _info_run(NULL, NULL, info, NULL, 0, 0, 0, major, minor);
+}
+
 int dev_manager_info(struct dm_pool *mem, const char *name,
 		     const struct logical_volume *lv, int with_mknodes,
-		     int with_open_count, struct dm_info *info)
+		     int with_open_count, int with_read_ahead,
+		     struct dm_info *info, uint32_t *read_ahead)
 {
 	const char *dlid;
 
@@ -186,8 +245,8 @@ int dev_manager_info(struct dm_pool *mem, const char *name,
 		return 0;
 	}
 
-	return _info(name, dlid, with_mknodes, with_open_count, info,
-		     NULL, NULL);
+	return _info(name, dlid, with_mknodes, with_open_count, with_read_ahead,
+		     info, read_ahead);
 }
 
 /* FIXME Interface must cope with multiple targets */
@@ -203,10 +262,8 @@ static int _status_run(const char *name, const char *uuid,
 	char *type = NULL;
 	char *params = NULL;
 
-	if (!(dmt = _setup_task(name, uuid, 0, DM_DEVICE_STATUS))) {
-		stack;
-		return 0;
-	}
+	if (!(dmt = _setup_task(name, uuid, 0, DM_DEVICE_STATUS, 0, 0)))
+		return_0;
 
 	if (!dm_task_no_open_count(dmt))
 		log_error("Failed to disable open_count");
@@ -283,7 +340,7 @@ static int _percent_run(struct dev_manager *dm, const char *name,
 	uint64_t start, length;
 	char *type = NULL;
 	char *params = NULL;
-	struct list *segh = &lv->segments;
+	struct dm_list *segh = &lv->segments;
 	struct lv_segment *seg = NULL;
 	struct segment_type *segtype;
 
@@ -292,10 +349,8 @@ static int _percent_run(struct dev_manager *dm, const char *name,
 	*percent = -1;
 
 	if (!(dmt = _setup_task(name, dlid, event_nr,
-				wait ? DM_DEVICE_WAITEVENT : DM_DEVICE_STATUS))) {
-		stack;
-		return 0;
-	}
+				wait ? DM_DEVICE_WAITEVENT : DM_DEVICE_STATUS, 0, 0)))
+		return_0;
 
 	if (!dm_task_no_open_count(dmt))
 		log_error("Failed to disable open_count");
@@ -313,12 +368,12 @@ static int _percent_run(struct dev_manager *dm, const char *name,
 		next = dm_get_next_target(dmt, next, &start, &length, &type,
 					  &params);
 		if (lv) {
-			if (!(segh = list_next(&lv->segments, segh))) {
+			if (!(segh = dm_list_next(&lv->segments, segh))) {
 				log_error("Number of segments in active LV %s "
 					  "does not match metadata", lv->name);
 				goto out;
 			}
-			seg = list_item(segh, struct lv_segment);
+			seg = dm_list_item(segh, struct lv_segment);
 		}
 
 		if (!type || !params || strcmp(type, target_type))
@@ -329,15 +384,14 @@ static int _percent_run(struct dev_manager *dm, const char *name,
 
 		if (segtype->ops->target_percent &&
 		    !segtype->ops->target_percent(&dm->target_state, dm->mem,
-						  dm->cmd->cft, seg, params,
+						  dm->cmd, seg, params,
 						  &total_numerator,
-						  &total_denominator,
-						  percent))
+						  &total_denominator))
 			goto_out;
 
 	} while (next);
 
-	if (lv && (segh = list_next(&lv->segments, segh))) {
+	if (lv && (segh = dm_list_next(&lv->segments, segh))) {
 		log_error("Number of segments in active LV %s does not "
 			  "match metadata", lv->name);
 		goto out;
@@ -345,7 +399,7 @@ static int _percent_run(struct dev_manager *dm, const char *name,
 
 	if (total_denominator)
 		*percent = (float) total_numerator *100 / total_denominator;
-	else if (*percent < 0)
+	else
 		*percent = 100;
 
 	log_debug("LV percent: %f", *percent);
@@ -387,30 +441,17 @@ struct dev_manager *dev_manager_create(struct cmd_context *cmd,
 	struct dm_pool *mem;
 	struct dev_manager *dm;
 
-	if (!(mem = dm_pool_create("dev_manager", 16 * 1024))) {
-		stack;
-		return NULL;
-	}
+	if (!(mem = dm_pool_create("dev_manager", 16 * 1024)))
+		return_NULL;
 
-	if (!(dm = dm_pool_alloc(mem, sizeof(*dm)))) {
-		stack;
-		goto bad;
-	}
+	if (!(dm = dm_pool_zalloc(mem, sizeof(*dm))))
+		goto_bad;
 
 	dm->cmd = cmd;
 	dm->mem = mem;
 
-	if (!stripe_filler) {
-		stripe_filler = find_config_str(cmd->cft->root,
-						"activation/missing_stripe_filler",
-						DEFAULT_STRIPE_FILLER);
-	}
-	dm->stripe_filler = stripe_filler;
-
-	if (!(dm->vg_name = dm_pool_strdup(dm->mem, vg_name))) {
-		stack;
-		goto bad;
-	}
+	if (!(dm->vg_name = dm_pool_strdup(dm->mem, vg_name)))
+		goto_bad;
 
 	dm->target_state = NULL;
 
@@ -426,13 +467,19 @@ void dev_manager_destroy(struct dev_manager *dm)
 	dm_pool_destroy(dm->mem);
 }
 
+void dev_manager_release(void)
+{
+	dm_lib_release();
+}
+
 void dev_manager_exit(void)
 {
 	dm_lib_exit();
 }
 
 int dev_manager_snapshot_percent(struct dev_manager *dm,
-				 struct logical_volume *lv, float *percent)
+				 const struct logical_volume *lv,
+				 float *percent)
 {
 	char *name;
 	const char *dlid;
@@ -451,10 +498,8 @@ int dev_manager_snapshot_percent(struct dev_manager *dm,
 	 */
 	log_debug("Getting device status percentage for %s", name);
 	if (!(_percent(dm, name, dlid, "snapshot", 0, NULL, percent,
-		       NULL))) {
-		stack;
-		return 0;
-	}
+		       NULL)))
+		return_0;
 
 	/* FIXME dm_pool_free ? */
 
@@ -486,10 +531,8 @@ int dev_manager_mirror_percent(struct dev_manager *dm,
 
 	log_debug("Getting device mirror status percentage for %s", name);
 	if (!(_percent(dm, name, dlid, "mirror", wait, lv, percent,
-		       event_nr))) {
-		stack;
-		return 0;
-	}
+		       event_nr)))
+		return_0;
 
 	return 1;
 }
@@ -516,9 +559,9 @@ int dev_manager_mirror_percent(struct dev_manager *dm,
 	log_debug("Getting device info for %s", dl->name);
 
 	/* Rename? */
-		if ((suffix = rindex(dl->dlid + sizeof(UUID_PREFIX) - 1, '-')))
+		if ((suffix = strrchr(dl->dlid + sizeof(UUID_PREFIX) - 1, '-')))
 			suffix++;
-		newname = build_dm_name(dm->mem, dm->vg_name, dl->lv->name,
+		new_name = build_dm_name(dm->mem, dm->vg_name, dl->lv->name,
 					suffix);
 
 static int _belong_to_vg(const char *vgname, const char *name)
@@ -543,7 +586,7 @@ static int _belong_to_vg(const char *vgname, const char *name)
 	old_origin = snap_seg->origin;
 
 	/* Was this the last active snapshot with this origin? */
-	list_iterate_items(lvl, active_head) {
+	dm_list_iterate_items(lvl, active_head) {
 		active = lvl->lv;
 		if ((snap_seg = find_cow(active)) &&
 		    snap_seg->origin == old_origin) {
@@ -577,7 +620,7 @@ static int _add_dev_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 			       struct logical_volume *lv, const char *layer)
 {
 	char *dlid, *name;
-	struct dm_info info;
+	struct dm_info info, info2;
 
 	if (!(name = build_dm_name(dm->mem, lv->vg->name, lv->name, layer)))
 		return_0;
@@ -585,11 +628,35 @@ static int _add_dev_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 	if (!(dlid = build_dlid(dm, lv->lvid.s, layer)))
 		return_0;
 
-        log_debug("Getting device info for %s [%s]", name, dlid);
-        if (!_info(name, dlid, 0, 1, &info, dm->mem, NULL)) {
-                log_error("Failed to get info for %s [%s].", name, dlid);
-                return 0;
-        }
+	log_debug("Getting device info for %s [%s]", name, dlid);
+	if (!_info(name, dlid, 0, 1, 0, &info, NULL)) {
+		log_error("Failed to get info for %s [%s].", name, dlid);
+		return 0;
+	}
+
+	/*
+	 * For top level volumes verify that existing device match
+	 * requested major/minor and that major/minor pair is available for use
+	 */
+	if (!layer && lv->major != -1 && lv->minor != -1) {
+		/*
+		 * FIXME compare info.major with lv->major if multiple major support
+		 */
+		if (info.exists && (info.minor != lv->minor)) {
+			log_error("Volume %s (%" PRIu32 ":%" PRIu32")"
+				  " differs from already active device "
+				  "(%" PRIu32 ":%" PRIu32")",
+				  lv->name, lv->major, lv->minor, info.major, info.minor);
+			return 0;
+		}
+		if (!info.exists && _info_by_dev(lv->major, lv->minor, &info2) &&
+		    info2.exists) {
+			log_error("The requested major:minor pair "
+				  "(%" PRIu32 ":%" PRIu32") is already used",
+				  lv->major, lv->minor);
+			return 0;
+		}
+	}
 
 	if (info.exists && !dm_tree_add_dev(dtree, info.major, info.minor)) {
 		log_error("Failed to add device (%" PRIu32 ":%" PRIu32") to dtree",
@@ -615,13 +682,16 @@ static int _add_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree, struc
 	if (!_add_dev_to_dtree(dm, dtree, lv, "cow"))
 		return_0;
 
+	if (!_add_dev_to_dtree(dm, dtree, lv, "_mlog"))
+		return_0;
+
 	return 1;
 }
 
 static struct dm_tree *_create_partial_dtree(struct dev_manager *dm, struct logical_volume *lv)
 {
 	struct dm_tree *dtree;
-	struct list *snh, *snht;
+	struct dm_list *snh, *snht;
 	struct lv_segment *seg;
 	uint32_t s;
 
@@ -630,37 +700,94 @@ static struct dm_tree *_create_partial_dtree(struct dev_manager *dm, struct logi
 		return NULL;
 	}
 
-	if (!_add_lv_to_dtree(dm, dtree, lv)) {
-		stack;
-		goto fail;
-	}
+	if (!_add_lv_to_dtree(dm, dtree, lv))
+		goto_bad;
 
 	/* Add any snapshots of this LV */
-	list_iterate_safe(snh, snht, &lv->snapshot_segs)
-		if (!_add_lv_to_dtree(dm, dtree, list_struct_base(snh, struct lv_segment, origin_list)->cow)) {
-			stack;
-			goto fail;
-		}
+	dm_list_iterate_safe(snh, snht, &lv->snapshot_segs)
+		if (!_add_lv_to_dtree(dm, dtree, dm_list_struct_base(snh, struct lv_segment, origin_list)->cow))
+			goto_bad;
 
 	/* Add any LVs used by segments in this LV */
-	list_iterate_items(seg, &lv->segments)
+	dm_list_iterate_items(seg, &lv->segments)
 		for (s = 0; s < seg->area_count; s++)
 			if (seg_type(seg, s) == AREA_LV && seg_lv(seg, s)) {
-				if (!_add_lv_to_dtree(dm, dtree, seg_lv(seg, s))) {
-					stack;
-					goto fail;
-				}
+				if (!_add_lv_to_dtree(dm, dtree, seg_lv(seg, s)))
+					goto_bad;
 			}
 
 	return dtree;
 
-fail:
+bad:
 	dm_tree_free(dtree);
 	return NULL;
 }
 
+static char *_add_error_device(struct dev_manager *dm, struct dm_tree *dtree,
+			       struct lv_segment *seg, int s)
+{
+	char *id, *name;
+	char errid[32];
+	struct dm_tree_node *node;
+	struct lv_segment *seg_i;
+	int segno = -1, i = 0;;
+	uint64_t size = seg->len * seg->lv->vg->extent_size;
+
+	dm_list_iterate_items(seg_i, &seg->lv->segments) {
+		if (seg == seg_i)
+			segno = i;
+		++i;
+	}
+
+	if (segno < 0) {
+		log_error("_add_error_device called with bad segment");
+		return_NULL;
+	}
+
+	sprintf(errid, "missing_%d_%d", segno, s);
+
+	if (!(id = build_dlid(dm, seg->lv->lvid.s, errid))) 
+		return_NULL;
+
+	if (!(name = build_dm_name(dm->mem, seg->lv->vg->name,
+				   seg->lv->name, errid)))
+		return_NULL;
+	if (!(node = dm_tree_add_new_dev(dtree, name, id, 0, 0, 0, 0, 0)))
+		return_NULL;
+	if (!dm_tree_node_add_error_target(node, size))
+		return_NULL;
+
+	return id;
+}
+
+static int _add_error_area(struct dev_manager *dm, struct dm_tree_node *node,
+			   struct lv_segment *seg, int s)
+{
+	char *dlid;
+	uint64_t extent_size = seg->lv->vg->extent_size;
+
+	if (!strcmp(dm->cmd->stripe_filler, "error")) {
+		/*
+		 * FIXME, the tree pointer is first field of dm_tree_node, but
+		 * we don't have the struct definition available.
+		 */
+		struct dm_tree **tree = (struct dm_tree **) node;
+		dlid = _add_error_device(dm, *tree, seg, s);
+		if (!dlid)
+			return_0;
+		dm_tree_node_add_target_area(node, NULL, dlid,
+					     extent_size * seg_le(seg, s));
+	} else
+		dm_tree_node_add_target_area(node,
+					     dm->cmd->stripe_filler,
+					     NULL, UINT64_C(0));
+
+	return 1;
+}
+
 int add_areas_line(struct dev_manager *dm, struct lv_segment *seg,
-		   struct dm_tree_node *node, int start_area, int areas)
+		   struct dm_tree_node *node, uint32_t start_area,
+		   uint32_t areas)
 {
 	uint64_t extent_size = seg->lv->vg->extent_size;
 	uint32_t s;
@@ -671,11 +798,10 @@ int add_areas_line(struct dev_manager *dm, struct lv_segment *seg,
 		     (!seg_pvseg(seg, s) ||
 		      !seg_pv(seg, s) ||
 		      !seg_dev(seg, s))) ||
-		    (seg_type(seg, s) == AREA_LV && !seg_lv(seg, s)))
-			dm_tree_node_add_target_area(node,
-							dm->stripe_filler,
-							NULL, 0);
-		else if (seg_type(seg, s) == AREA_PV)
+		    (seg_type(seg, s) == AREA_LV && !seg_lv(seg, s))) {
+			if (!_add_error_area(dm, node, seg, s))
+				return_0;
+		} else if (seg_type(seg, s) == AREA_PV)
 			dm_tree_node_add_target_area(node,
 							dev_name(seg_dev(seg, s)),
 							NULL,
@@ -699,7 +825,6 @@ int add_areas_line(struct dev_manager *dm, struct lv_segment *seg,
 }
 
 static int _add_origin_target_to_dtree(struct dev_manager *dm,
-					 struct dm_tree *dtree,
 					 struct dm_tree_node *dnode,
 					 struct logical_volume *lv)
 {
@@ -715,7 +840,6 @@ static int _add_origin_target_to_dtree(struct dev_manager *dm,
 }
 
 static int _add_snapshot_target_to_dtree(struct dev_manager *dm,
-					   struct dm_tree *dtree,
 					   struct dm_tree_node *dnode,
 					   struct logical_volume *lv)
 {
@@ -744,7 +868,6 @@ static int _add_snapshot_target_to_dtree(struct dev_manager *dm,
 }
 
 static int _add_target_to_dtree(struct dev_manager *dm,
-				  struct dm_tree *dtree,
 				  struct dm_tree_node *dnode,
 				  struct lv_segment *seg)
 {
@@ -756,7 +879,7 @@ static int _add_target_to_dtree(struct dev_manager *dm,
 		return 0;
 	}
 
-	return seg->segtype->ops->add_target_line(dm, dm->mem, dm->cmd->cft,
+	return seg->segtype->ops->add_target_line(dm, dm->mem, dm->cmd,
 						  &dm->target_state, seg,
 						  dnode,
 						  extent_size * seg->len,
@@ -773,13 +896,21 @@ static int _add_segment_to_dtree(struct dev_manager *dm,
 				   const char *layer)
 {
 	uint32_t s;
-	struct list *snh;
+	struct dm_list *snh;
+	struct lv_segment *seg_present;
 
 	/* Ensure required device-mapper targets are loaded */
-	if (seg->segtype->ops->target_present &&
-	    !seg->segtype->ops->target_present()) {
+	seg_present = find_cow(seg->lv) ? : seg;
+
+	log_debug("Checking kernel supports %s segment type for %s%s%s",
+		  seg_present->segtype->name, seg->lv->name,
+		  layer ? "-" : "", layer ? : "");
+
+	if (seg_present->segtype->ops->target_present &&
+	    !seg_present->segtype->ops->target_present(seg_present->lv->vg->cmd,
+						       seg_present, NULL)) {
 		log_error("Can't expand LV %s: %s target support missing "
-			  "from kernel?", seg->lv->name, seg->segtype->name);
+			  "from kernel?", seg->lv->name, seg_present->segtype->name);
 		return 0;
 	}
 
@@ -790,7 +921,7 @@ static int _add_segment_to_dtree(struct dev_manager *dm,
 
 	/* If this is a snapshot origin, add real LV */
 	if (lv_is_origin(seg->lv) && !layer) {
-		if (seg->lv->vg->status & CLUSTERED) {
+		if (vg_is_clustered(seg->lv->vg)) {
 			log_error("Clustered snapshots are not yet supported");
 			return 0;
 		}
@@ -809,18 +940,18 @@ static int _add_segment_to_dtree(struct dev_manager *dm,
 
 	/* Now we've added its dependencies, we can add the target itself */
 	if (lv_is_origin(seg->lv) && !layer) {
-		if (!_add_origin_target_to_dtree(dm, dtree, dnode, seg->lv))
+		if (!_add_origin_target_to_dtree(dm, dnode, seg->lv))
 			return_0;
 	} else if (lv_is_cow(seg->lv) && !layer) {
-		if (!_add_snapshot_target_to_dtree(dm, dtree, dnode, seg->lv))
+		if (!_add_snapshot_target_to_dtree(dm, dnode, seg->lv))
 			return_0;
-	} else if (!_add_target_to_dtree(dm, dtree, dnode, seg))
+	} else if (!_add_target_to_dtree(dm, dnode, seg))
 		return_0;
 
 	if (lv_is_origin(seg->lv) && !layer)
 		/* Add any snapshots of this LV */
-		list_iterate(snh, &seg->lv->snapshot_segs)
-			if (!_add_new_lv_to_dtree(dm, dtree, list_struct_base(snh, struct lv_segment, origin_list)->cow, NULL))
+		dm_list_iterate(snh, &seg->lv->snapshot_segs)
+			if (!_add_new_lv_to_dtree(dm, dtree, dm_list_struct_base(snh, struct lv_segment, origin_list)->cow, NULL))
 				return_0;
 
 	return 1;
@@ -833,6 +964,9 @@ static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 	struct lv_layer *lvlayer;
 	struct dm_tree_node *dnode;
 	char *name, *dlid;
+	uint32_t max_stripe_size = UINT32_C(0);
+	uint32_t read_ahead = lv->read_ahead;
+	uint32_t read_ahead_flags = UINT32_C(0);
 
 	if (!(name = build_dm_name(dm->mem, lv->vg->name, lv->name, layer)))
 		return_0;
@@ -859,10 +993,10 @@ static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 	 * Major/minor settings only apply to the visible layer.
 	 */
 	if (!(dnode = dm_tree_add_new_dev(dtree, name, dlid,
-					     layer ? lv->major : 0,
-					     layer ? lv->minor : 0,
+					     layer ? UINT32_C(0) : (uint32_t) lv->major,
+					     layer ? UINT32_C(0) : (uint32_t) lv->minor,
 					     _read_only_lv(lv),
-					     lv->vg->status & PRECOMMITTED,
+					     (lv->vg->status & PRECOMMITTED) ? 1 : 0,
 					     lvlayer)))
 		return_0;
 
@@ -871,7 +1005,7 @@ static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 
 	/* Create table */
 	dm->pvmove_mirror_count = 0u;
-	list_iterate_items(seg, &lv->segments) {
+	dm_list_iterate_items(seg, &lv->segments) {
 		if (!_add_segment_to_dtree(dm, dtree, dnode, seg, layer))
 			return_0;
 		/* These aren't real segments in the LVM2 metadata */
@@ -879,11 +1013,27 @@ static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
 			break;
 		if (lv_is_cow(lv) && !layer)
 			break;
+		if (max_stripe_size < seg->stripe_size * seg->area_count)
+			max_stripe_size = seg->stripe_size * seg->area_count;
 	}
 
+	if (read_ahead == DM_READ_AHEAD_AUTO) {
+		/* we need RA at least twice a whole stripe - see the comment in md/raid0.c */
+		read_ahead = max_stripe_size * 2;
+		if (!read_ahead)
+			lv_calculate_readahead(lv, &read_ahead);
+		read_ahead_flags = DM_READ_AHEAD_MINIMUM_FLAG;
+	}
+
+	dm_tree_node_set_read_ahead(dnode, read_ahead, read_ahead_flags);
+
 	return 1;
 }
 
+/* FIXME: symlinks should be created/destroyed at the same time
+ * as the kernel devices but we can't do that from within libdevmapper
+ * at present so we must walk the tree twice instead. */
+
 /*
  * Create LV symlinks for children of supplied root node.
  */
@@ -892,7 +1042,8 @@ static int _create_lv_symlinks(struct dev_manager *dm, struct dm_tree_node *root
 	void *handle = NULL;
 	struct dm_tree_node *child;
 	struct lv_layer *lvlayer;
-	char *vgname, *lvname, *layer;
+	char *old_vgname, *old_lvname, *old_layer;
+	char *new_vgname, *new_lvname, *new_layer;
 	const char *name;
 	int r = 1;
 
@@ -904,19 +1055,60 @@ static int _create_lv_symlinks(struct dev_manager *dm, struct dm_tree_node *root
 		name = dm_tree_node_get_name(child);
 
 		if (name && lvlayer->old_name && *lvlayer->old_name && strcmp(name, lvlayer->old_name)) {
-	        	if (!split_dm_name(dm->mem, lvlayer->old_name, &vgname, &lvname, &layer)) {
-                		log_error("_create_lv_symlinks: Couldn't split up old device name %s", lvlayer->old_name);
-                		return 0;
-        		}
-			fs_rename_lv(lvlayer->lv, name, lvname);
-		} else if (!dev_manager_lv_mknodes(lvlayer->lv))
+			if (!dm_split_lvm_name(dm->mem, lvlayer->old_name, &old_vgname, &old_lvname, &old_layer)) {
+				log_error("_create_lv_symlinks: Couldn't split up old device name %s", lvlayer->old_name);
+				return 0;
+			}
+			if (!dm_split_lvm_name(dm->mem, name, &new_vgname, &new_lvname, &new_layer)) {
+				log_error("_create_lv_symlinks: Couldn't split up new device name %s", name);
+				return 0;
+			}
+			if (!fs_rename_lv(lvlayer->lv, name, old_vgname, old_lvname))
+				r = 0;
+			continue;
+		}
+		if (lv_is_visible(lvlayer->lv)) {
+			if (!dev_manager_lv_mknodes(lvlayer->lv))
+				r = 0;
+			continue;
+		}
+		if (!dev_manager_lv_rmnodes(lvlayer->lv))
 			r = 0;
 	}
 
 	return r;
 }
 
-static int _clean_tree(struct dev_manager *dm, struct logical_volume *lv, struct dm_tree_node *root)
+/*
+ * Remove LV symlinks for children of supplied root node.
+ */
+static int _remove_lv_symlinks(struct dev_manager *dm, struct dm_tree_node *root)
+{
+	void *handle = NULL;
+	struct dm_tree_node *child;
+	char *vgname, *lvname, *layer;
+	int r = 1;
+
+	while ((child = dm_tree_next_child(&handle, root, 0))) {
+		if (!dm_split_lvm_name(dm->mem, dm_tree_node_get_name(child), &vgname, &lvname, &layer)) {
+			r = 0;
+			continue;
+		}
+
+		if (!*vgname)
+			continue;
+
+		/* only top level layer has symlinks */
+		if (*layer)
+			continue;
+
+		fs_del_lv_byname(dm->cmd->dev_dir, vgname, lvname);
+	}
+
+	return r;
+}
+
+static int _clean_tree(struct dev_manager *dm, struct dm_tree_node *root)
 {
 	void *handle = NULL;
 	struct dm_tree_node *child;
@@ -930,10 +1122,10 @@ static int _clean_tree(struct dev_manager *dm, struct logical_volume *lv, struct
 		if (!(uuid = dm_tree_node_get_uuid(child)))
 			continue;
 
-        	if (!split_dm_name(dm->mem, name, &vgname, &lvname, &layer)) {
-                	log_error("_clean_tree: Couldn't split up device name %s.", name);
-                	return 0;
-        	}
+		if (!dm_split_lvm_name(dm->mem, name, &vgname, &lvname, &layer)) {
+			log_error("_clean_tree: Couldn't split up device name %s.", name);
+			return 0;
+		}
 
 		/* Not meant to be top level? */
 		if (!*layer)
@@ -968,17 +1160,21 @@ static int _tree_action(struct dev_manager *dm, struct logical_volume *lv, actio
 	switch(action) {
 	case CLEAN:
 		/* Deactivate any unused non-toplevel nodes */
-		if (!_clean_tree(dm, lv, root))
+		if (!_clean_tree(dm, root))
 			goto_out;
 		break;
 	case DEACTIVATE:
  		/* Deactivate LV and all devices it references that nothing else has open. */
 		if (!dm_tree_deactivate_children(root, dlid, ID_LEN + sizeof(UUID_PREFIX) - 1))
 			goto_out;
+		if (!_remove_lv_symlinks(dm, root))
+			log_error("Failed to remove all device symlinks associated with %s.", lv->name);
 		break;
 	case SUSPEND:
-		if (!lv_is_origin(lv) && !lv_is_cow(lv))
-			dm_tree_skip_lockfs(root);
+		dm_tree_skip_lockfs(root);
+		if (!dm->flush_required && (lv->status & MIRRORED) && !(lv->status & PVMOVE))
+			dm_tree_use_no_flush_suspend(root);
+	case SUSPEND_WITH_LOCKFS:
 		if (!dm_tree_suspend_children(root, dlid, ID_LEN + sizeof(UUID_PREFIX) - 1))
 			goto_out;
 		break;
@@ -992,6 +1188,9 @@ static int _tree_action(struct dev_manager *dm, struct logical_volume *lv, actio
 		if (!dm_tree_preload_children(root, dlid, ID_LEN + sizeof(UUID_PREFIX) - 1))
 			goto_out;
 
+		if (dm_tree_node_size_changed(root))
+			dm->flush_required = 1;
+
 		if ((action == ACTIVATE) &&
 		    !dm_tree_activate_children(root, dlid, ID_LEN + sizeof(UUID_PREFIX) - 1))
 			goto_out;
@@ -1022,13 +1221,19 @@ int dev_manager_activate(struct dev_manager *dm, struct logical_volume *lv)
 	return _tree_action(dm, lv, CLEAN);
 }
 
-int dev_manager_preload(struct dev_manager *dm, struct logical_volume *lv)
+int dev_manager_preload(struct dev_manager *dm, struct logical_volume *lv,
+			int *flush_required)
 {
 	/* FIXME Update the pvmove implementation! */
 	if ((lv->status & PVMOVE) || (lv->status & LOCKED))
 		return 1;
 
-	return _tree_action(dm, lv, PRELOAD);
+	if (!_tree_action(dm, lv, PRELOAD))
+		return 0;
+
+	*flush_required = dm->flush_required;
+
+	return 1;
 }
 
 int dev_manager_deactivate(struct dev_manager *dm, struct logical_volume *lv)
@@ -1042,21 +1247,24 @@ int dev_manager_deactivate(struct dev_manager *dm, struct logical_volume *lv)
 	return r;
 }
 
-int dev_manager_suspend(struct dev_manager *dm, struct logical_volume *lv)
+int dev_manager_suspend(struct dev_manager *dm, struct logical_volume *lv,
+			int lockfs, int flush_required)
 {
-	return _tree_action(dm, lv, SUSPEND);
+	dm->flush_required = flush_required;
+
+	return _tree_action(dm, lv, lockfs ? SUSPEND_WITH_LOCKFS : SUSPEND);
 }
 
 /*
  * Does device use VG somewhere in its construction?
  * Returns 1 if uncertain.
  */
-int dev_manager_device_uses_vg(struct dev_manager *dm, struct device *dev,
+int dev_manager_device_uses_vg(struct device *dev,
 			       struct volume_group *vg)
 {
 	struct dm_tree *dtree;
 	struct dm_tree_node *root;
-	char dlid[sizeof(UUID_PREFIX) + sizeof(struct id) - 1];
+	char dlid[sizeof(UUID_PREFIX) + sizeof(struct id) - 1] __attribute((aligned(8)));
 	int r = 1;
 
 	if (!(dtree = dm_tree_create())) {
@@ -1064,7 +1272,7 @@ int dev_manager_device_uses_vg(struct dev_manager *dm, struct device *dev,
 		return r;
 	}
 
-	if (!dm_tree_add_dev(dtree, MAJOR(dev->dev), MINOR(dev->dev))) {
+	if (!dm_tree_add_dev(dtree, (uint32_t) MAJOR(dev->dev), (uint32_t) MINOR(dev->dev))) {
 		log_error("Failed to add device %s (%" PRIu32 ":%" PRIu32") to dtree",
 			  dev_name(dev), (uint32_t) MAJOR(dev->dev), (uint32_t) MINOR(dev->dev));
 		goto out;

lib/activate/dev_manager.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
  * This copyrighted material is made available to anyone wishing to use,
  * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -29,6 +29,7 @@ struct device;
 struct dev_manager *dev_manager_create(struct cmd_context *cmd,
 				       const char *vg_name);
 void dev_manager_destroy(struct dev_manager *dm);
+void dev_manager_release(void);
 void dev_manager_exit(void);
 
 /*
@@ -39,15 +40,19 @@ void dev_manager_exit(void);
  */
 int dev_manager_info(struct dm_pool *mem, const char *name,
 		     const struct logical_volume *lv,
-		     int mknodes, int with_open_count, struct dm_info *info);
+		     int mknodes, int with_open_count, int with_read_ahead,
+		     struct dm_info *info, uint32_t *read_ahead);
 int dev_manager_snapshot_percent(struct dev_manager *dm,
-				 struct logical_volume *lv, float *percent);
+				 const struct logical_volume *lv,
+				 float *percent);
 int dev_manager_mirror_percent(struct dev_manager *dm,
 			       struct logical_volume *lv, int wait,
 			       float *percent, uint32_t *event_nr);
-int dev_manager_suspend(struct dev_manager *dm, struct logical_volume *lv);
+int dev_manager_suspend(struct dev_manager *dm, struct logical_volume *lv,
+			int lockfs, int flush_required);
 int dev_manager_activate(struct dev_manager *dm, struct logical_volume *lv);
-int dev_manager_preload(struct dev_manager *dm, struct logical_volume *lv);
+int dev_manager_preload(struct dev_manager *dm, struct logical_volume *lv,
+			int *flush_required);
 int dev_manager_deactivate(struct dev_manager *dm, struct logical_volume *lv);
 
 int dev_manager_lv_mknodes(const struct logical_volume *lv);
@@ -58,7 +63,7 @@ int dev_manager_lv_rmnodes(const struct logical_volume *lv);
  */
 int dev_manager_execute(struct dev_manager *dm);
 
-int dev_manager_device_uses_vg(struct dev_manager *dm, struct device *dev,
+int dev_manager_device_uses_vg(struct device *dev,
 			       struct volume_group *vg);
 
 #endif

lib/activate/fs.c

@@ -1,14 +1,14 @@
 /*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2001-2004 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -30,7 +30,7 @@ static int _mk_dir(const char *dev_dir, const char *vg_name)
 {
 	char vg_path[PATH_MAX];
 
-	if (lvm_snprintf(vg_path, sizeof(vg_path), "%s%s",
+	if (dm_snprintf(vg_path, sizeof(vg_path), "%s%s",
 			 dev_dir, vg_name) == -1) {
 		log_error("Couldn't construct name of volume "
 			  "group directory.");
@@ -53,7 +53,7 @@ static int _rm_dir(const char *dev_dir, const char *vg_name)
 {
 	char vg_path[PATH_MAX];
 
-	if (lvm_snprintf(vg_path, sizeof(vg_path), "%s%s",
+	if (dm_snprintf(vg_path, sizeof(vg_path), "%s%s",
 			 dev_dir, vg_name) == -1) {
 		log_error("Couldn't construct name of volume "
 			  "group directory.");
@@ -87,7 +87,7 @@ static void _rm_blks(const char *dir)
 		if (!strcmp(name, ".") || !strcmp(name, ".."))
 			continue;
 
-		if (lvm_snprintf(path, sizeof(path), "%s/%s", dir, name) == -1) {
+		if (dm_snprintf(path, sizeof(path), "%s/%s", dir, name) == -1) {
 			log_error("Couldn't create path for %s", name);
 			continue;
 		}
@@ -109,28 +109,28 @@ static int _mk_link(const char *dev_dir, const char *vg_name,
 	char vg_path[PATH_MAX];
 	struct stat buf;
 
-	if (lvm_snprintf(vg_path, sizeof(vg_path), "%s%s",
+	if (dm_snprintf(vg_path, sizeof(vg_path), "%s%s",
 			 dev_dir, vg_name) == -1) {
 		log_error("Couldn't create path for volume group dir %s",
 			  vg_name);
 		return 0;
 	}
 
-	if (lvm_snprintf(lv_path, sizeof(lv_path), "%s/%s", vg_path,
+	if (dm_snprintf(lv_path, sizeof(lv_path), "%s/%s", vg_path,
 			 lv_name) == -1) {
 		log_error("Couldn't create source pathname for "
 			  "logical volume link %s", lv_name);
 		return 0;
 	}
 
-	if (lvm_snprintf(link_path, sizeof(link_path), "%s/%s",
+	if (dm_snprintf(link_path, sizeof(link_path), "%s/%s",
 			 dm_dir(), dev) == -1) {
 		log_error("Couldn't create destination pathname for "
 			  "logical volume link for %s", lv_name);
 		return 0;
 	}
 
-	if (lvm_snprintf(lvm1_group_path, sizeof(lvm1_group_path), "%s/group",
+	if (dm_snprintf(lvm1_group_path, sizeof(lvm1_group_path), "%s/group",
 			 vg_path) == -1) {
 		log_error("Couldn't create pathname for LVM1 group file for %s",
 			  vg_name);
@@ -138,7 +138,7 @@ static int _mk_link(const char *dev_dir, const char *vg_name,
 	}
 
 	/* To reach this point, the VG must have been locked.
-	 * As locking fails if the VG is active under LVM1, it's 
+	 * As locking fails if the VG is active under LVM1, it's
 	 * now safe to remove any LVM1 devices we find here
 	 * (as well as any existing LVM2 symlink). */
 	if (!lstat(lvm1_group_path, &buf)) {
@@ -175,10 +175,8 @@ static int _mk_link(const char *dev_dir, const char *vg_name,
 	}
 
 #ifdef HAVE_SELINUX
-        if (!dm_set_selinux_context(lv_path, S_IFLNK)) {
-                stack;
-                return 0;
-        }
+	if (!dm_set_selinux_context(lv_path, S_IFLNK))
+		return_0;
 #endif
 
 	return 1;
@@ -190,7 +188,7 @@ static int _rm_link(const char *dev_dir, const char *vg_name,
 	struct stat buf;
 	char lv_path[PATH_MAX];
 
-	if (lvm_snprintf(lv_path, sizeof(lv_path), "%s%s/%s",
+	if (dm_snprintf(lv_path, sizeof(lv_path), "%s%s/%s",
 			 dev_dir, vg_name, lv_name) == -1) {
 		log_error("Couldn't determine link pathname.");
 		return 0;
@@ -225,17 +223,13 @@ static int _do_fs_op(fs_op_t type, const char *dev_dir, const char *vg_name,
 	switch (type) {
 	case FS_ADD:
 		if (!_mk_dir(dev_dir, vg_name) ||
-		    !_mk_link(dev_dir, vg_name, lv_name, dev)) {
-			stack;
-			return 0;
-		}
+		    !_mk_link(dev_dir, vg_name, lv_name, dev))
+			return_0;
 		break;
 	case FS_DEL:
 		if (!_rm_link(dev_dir, vg_name, lv_name) ||
-		    !_rm_dir(dev_dir, vg_name)) {
-			stack;
-			return 0;
-		}
+		    !_rm_dir(dev_dir, vg_name))
+			return_0;
 		break;
 		/* FIXME Use rename() */
 	case FS_RENAME:
@@ -249,10 +243,10 @@ static int _do_fs_op(fs_op_t type, const char *dev_dir, const char *vg_name,
 	return 1;
 }
 
-static LIST_INIT(_fs_ops);
+static DM_LIST_INIT(_fs_ops);
 
 struct fs_op_parms {
-	struct list list;
+	struct dm_list list;
 	fs_op_t type;
 	char *dev_dir;
 	char *vg_name;
@@ -292,21 +286,21 @@ static int _stack_fs_op(fs_op_t type, const char *dev_dir, const char *vg_name,
 	_store_str(&pos, &fsp->dev, dev);
 	_store_str(&pos, &fsp->old_lv_name, old_lv_name);
 
-	list_add(&_fs_ops, &fsp->list);
+	dm_list_add(&_fs_ops, &fsp->list);
 
 	return 1;
 }
 
 static void _pop_fs_ops(void)
 {
-	struct list *fsph, *fspht;
+	struct dm_list *fsph, *fspht;
 	struct fs_op_parms *fsp;
 
-	list_iterate_safe(fsph, fspht, &_fs_ops) {
-		fsp = list_item(fsph, struct fs_op_parms);
+	dm_list_iterate_safe(fsph, fspht, &_fs_ops) {
+		fsp = dm_list_item(fsph, struct fs_op_parms);
 		_do_fs_op(fsp->type, fsp->dev_dir, fsp->vg_name, fsp->lv_name,
 			  fsp->dev, fsp->old_lv_name);
-		list_del(&fsp->list);
+		dm_list_del(&fsp->list);
 		dm_free(fsp);
 	}
 }
@@ -316,10 +310,8 @@ static int _fs_op(fs_op_t type, const char *dev_dir, const char *vg_name,
 {
 	if (memlock()) {
 		if (!_stack_fs_op(type, dev_dir, vg_name, lv_name, dev,
-				  old_lv_name)) {
-			stack;
-			return 0;
-		}
+				  old_lv_name))
+			return_0;
 		return 1;
 	}
 
@@ -338,11 +330,22 @@ int fs_del_lv(const struct logical_volume *lv)
 		      "", "");
 }
 
-int fs_rename_lv(struct logical_volume *lv,
-		 const char *dev, const char *old_name)
+int fs_del_lv_byname(const char *dev_dir, const char *vg_name, const char *lv_name)
 {
-	return _fs_op(FS_RENAME, lv->vg->cmd->dev_dir, lv->vg->name, lv->name,
-		      dev, old_name);
+	return _fs_op(FS_DEL, dev_dir, vg_name, lv_name, "", "");
+}
+
+int fs_rename_lv(struct logical_volume *lv, const char *dev, 
+		const char *old_vgname, const char *old_lvname)
+{
+	if (strcmp(old_vgname, lv->vg->name)) {
+		return
+			(_fs_op(FS_DEL, lv->vg->cmd->dev_dir, old_vgname, old_lvname, "", "") &&
+			 _fs_op(FS_ADD, lv->vg->cmd->dev_dir, lv->vg->name, lv->name, dev, ""));
+	}
+	else 
+		return _fs_op(FS_RENAME, lv->vg->cmd->dev_dir, lv->vg->name, lv->name,
+			      dev, old_lvname);
 }
 
 void fs_unlock(void)

lib/activate/fs.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -25,8 +25,9 @@
  */
 int fs_add_lv(const struct logical_volume *lv, const char *dev);
 int fs_del_lv(const struct logical_volume *lv);
-int fs_rename_lv(struct logical_volume *lv,
-		 const char *dev, const char *old_name);
+int fs_del_lv_byname(const char *dev_dir, const char *vg_name, const char *lv_name);
+int fs_rename_lv(struct logical_volume *lv, const char *dev, 
+		const char *old_vgname, const char *old_lvname);
 void fs_unlock(void);
 
 #endif

lib/activate/targets.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2003-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
  * This copyrighted material is made available to anyone wishing to use,
  * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -24,7 +24,7 @@ int compose_areas_line(struct dev_manager *dm, struct lv_segment *seg,
 		       int start_area, int areas);
 
 int add_areas_line(struct dev_manager *dm, struct lv_segment *seg,
-                   struct dm_tree_node *node, int start_area, int areas);
+                   struct dm_tree_node *node, uint32_t start_area, uint32_t areas);
 
 int build_dev_string(struct dev_manager *dm, char *dlid, char *devbuf,
 		     size_t bufsize, const char *desc);

lib/cache/lvmcache.h

@@ -1,17 +1,16 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2008 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
  */
 
 #ifndef _LVM_CACHE_H
@@ -21,7 +20,8 @@
 #include "uuid.h"
 #include "label.h"
 
-#define ORPHAN ""
+#define ORPHAN_PREFIX "#"
+#define ORPHAN_VG_NAME(fmt) ORPHAN_PREFIX "orphans_" fmt
 
 #define CACHE_INVALID	0x00000001
 #define CACHE_LOCKED	0x00000002
@@ -33,19 +33,26 @@ struct cmd_context;
 struct format_type;
 struct volume_group;
 
+/* One per VG */
 struct lvmcache_vginfo {
-	struct list list;	/* Join these vginfos together */
-	struct list infos;	/* List head for lvmcache_infos */
+	struct dm_list list;	/* Join these vginfos together */
+	struct dm_list infos;	/* List head for lvmcache_infos */
 	const struct format_type *fmt;
 	char *vgname;		/* "" == orphan */
+	uint32_t status;
 	char vgid[ID_LEN + 1];
 	char _padding[7];
+	struct lvmcache_vginfo *next; /* Another VG with same name? */
+	char *creation_host;
+	char *vgmetadata;	/* Copy of VG metadata as format_text string */
+	unsigned precommitted;	/* Is vgmetadata live or precommitted? */
 };
 
+/* One per device */
 struct lvmcache_info {
-	struct list list;	/* Join VG members together */
-	struct list mdas;	/* list head for metadata areas */
-	struct list das;	/* list head for data areas */
+	struct dm_list list;	/* Join VG members together */
+	struct dm_list mdas;	/* list head for metadata areas */
+	struct dm_list das;	/* list head for data areas */
 	struct lvmcache_vginfo *vginfo;	/* NULL == unknown */
 	struct label *label;
 	const struct format_type *fmt;
@@ -55,7 +62,7 @@ struct lvmcache_info {
 };
 
 int lvmcache_init(void);
-void lvmcache_destroy(void);
+void lvmcache_destroy(struct cmd_context *cmd, int retain_orphans);
 
 /* Set full_scan to 1 to reread every filtered device label or
  * 2 to rescan /dev for new devices */
@@ -64,27 +71,45 @@ int lvmcache_label_scan(struct cmd_context *cmd, int full_scan);
 /* Add/delete a device */
 struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
 				   struct device *dev,
-				   const char *vgname, const char *vgid);
+				   const char *vgname, const char *vgid,
+				   uint32_t vgstatus);
+int lvmcache_add_orphan_vginfo(const char *vgname, struct format_type *fmt);
 void lvmcache_del(struct lvmcache_info *info);
 
 /* Update things */
-int lvmcache_update_vgname(struct lvmcache_info *info, const char *vgname);
-int lvmcache_update_vg(struct volume_group *vg);
+int lvmcache_update_vgname_and_id(struct lvmcache_info *info,
+				  const char *vgname, const char *vgid,
+				  uint32_t vgstatus, const char *hostname);
+int lvmcache_update_vg(struct volume_group *vg, unsigned precommitted);
 
 void lvmcache_lock_vgname(const char *vgname, int read_only);
 void lvmcache_unlock_vgname(const char *vgname);
 
 /* Queries */
-const struct format_type *fmt_from_vgname(const char *vgname);
-struct lvmcache_vginfo *vginfo_from_vgname(const char *vgname);
+const struct format_type *fmt_from_vgname(const char *vgname, const char *vgid);
+struct lvmcache_vginfo *vginfo_from_vgname(const char *vgname,
+					   const char *vgid);
 struct lvmcache_vginfo *vginfo_from_vgid(const char *vgid);
-struct lvmcache_info *info_from_pvid(const char *pvid);
+struct lvmcache_info *info_from_pvid(const char *pvid, int valid_only);
+const char *vgname_from_vgid(struct dm_pool *mem, const char *vgid);
 struct device *device_from_pvid(struct cmd_context *cmd, struct id *pvid);
 int vgs_locked(void);
 int vgname_is_locked(const char *vgname);
 
 /* Returns list of struct str_lists containing pool-allocated copy of vgnames */
 /* Set full_scan to 1 to reread every filtered device label */
-struct list *lvmcache_get_vgnames(struct cmd_context *cmd, int full_scan);
+struct dm_list *lvmcache_get_vgnames(struct cmd_context *cmd, int full_scan);
+
+/* Returns list of struct str_lists containing pool-allocated copy of vgids */
+/* Set full_scan to 1 to reread every filtered device label */
+struct dm_list *lvmcache_get_vgids(struct cmd_context *cmd, int full_scan);
+
+/* Returns list of struct str_lists containing pool-allocated copy of pvids */
+struct dm_list *lvmcache_get_pvids(struct cmd_context *cmd, const char *vgname,
+				const char *vgid);
+
+/* Returns cached volume group metadata. */
+struct volume_group *lvmcache_get_vg(const char *vgid, unsigned precommitted);
+void lvmcache_drop_metadata(const char *vgname);
 
 #endif

lib/commands/errors.h

@@ -6,9 +6,9 @@
  *
  * 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */

lib/commands/toolcontext.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2009 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 General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -33,6 +33,8 @@ struct config_info {
 	int suffix;
 	int archive;		/* should we archive ? */
 	int backup;		/* should we backup ? */
+	int read_ahead;		/* DM_READ_AHEAD_NONE or _AUTO */
+	int cache_vgmetadata;
 	const char *msg_prefix;
 	struct format_type *fmt;
 	uint64_t unit_factor;
@@ -55,40 +57,52 @@ struct cmd_context {
 	const struct format_type *fmt;	/* Current format to use by default */
 	struct format_type *fmt_backup;	/* Format to use for backups */
 
-	struct list formats;	/* Available formats */
-	struct list segtypes;	/* Available segment types */
+	struct dm_list formats;	/* Available formats */
+	struct dm_list segtypes;	/* Available segment types */
 	const char *hostname;
 	const char *kernel_vsn;
 
+	unsigned rand_seed;
 	char *cmd_line;
 	struct command *command;
-	struct arg *args;
 	char **argv;
+	unsigned is_long_lived:1;	/* Optimises persistent_filter handling */
+	unsigned handles_missing_pvs:1;
+	unsigned partial_activation:1;
 
 	struct dev_filter *filter;
 	int dump_filter;	/* Dump filter when exiting? */
 
-	struct list config_files;
+	struct dm_list config_files;
 	int config_valid;
 	struct config_tree *cft;
+	struct config_tree *cft_override;
 	struct config_info default_settings;
 	struct config_info current_settings;
 
 	struct archive_params *archive_params;
 	struct backup_params *backup_params;
+	const char *stripe_filler;
 
 	/* List of defined tags */
-	struct list tags;
+	struct dm_list tags;
 	int hosttags;
 
-	char sys_dir[PATH_MAX];
+	char system_dir[PATH_MAX];
 	char dev_dir[PATH_MAX];
 	char proc_dir[PATH_MAX];
+	char sysfs_dir[PATH_MAX];
 };
 
-struct cmd_context *create_toolcontext(struct arg *the_args);
+/*
+ * system_dir may be NULL to use the default value.
+ * The environment variable LVM_SYSTEM_DIR always takes precedence.
+ */
+struct cmd_context *create_toolcontext(unsigned is_long_lived,
+				       const char *system_dir);
 void destroy_toolcontext(struct cmd_context *cmd);
 int refresh_toolcontext(struct cmd_context *cmd);
 int config_files_changed(struct cmd_context *cmd);
+int init_lvmcache_orphans(struct cmd_context *cmd);
 
 #endif

lib/config/config.c

@@ -1,14 +1,14 @@
 /*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2001-2004 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -19,6 +19,8 @@
 #include "device.h"
 #include "str_list.h"
 #include "toolcontext.h"
+#include "lvm-string.h"
+#include "lvm-file.h"
 
 #include <sys/stat.h>
 #include <sys/mman.h>
@@ -26,10 +28,14 @@
 #include <fcntl.h>
 #include <ctype.h>
 
+#define SECTION_B_CHAR '{'
+#define SECTION_E_CHAR '}'
+
 enum {
 	TOK_INT,
 	TOK_FLOAT,
-	TOK_STRING,
+	TOK_STRING,		/* Single quotes */
+	TOK_STRING_ESCAPED,	/* Double quotes */
 	TOK_EQ,
 	TOK_SECTION_B,
 	TOK_SECTION_E,
@@ -41,10 +47,10 @@ enum {
 };
 
 struct parser {
-	char *fb, *fe;		/* file limits */
+	const char *fb, *fe;		/* file limits */
 
 	int t;			/* token limits and type */
-	char *tb, *te;
+	const char *tb, *te;
 
 	int fd;			/* descriptor for file being parsed */
 	int line;		/* line number we are on */
@@ -58,6 +64,13 @@ struct cs {
 	time_t timestamp;
 	char *filename;
 	int exists;
+	int keep_open;
+	struct device *dev;
+};
+
+struct output_line {
+	FILE *fp;
+	struct dm_pool *mem;
 };
 
 static void _get_token(struct parser *p, int tok_prev);
@@ -77,7 +90,8 @@ static const int sep = '/';
 
 #define match(t) do {\
    if (!_match_aux(p, (t))) {\
-	log_error("Parse error at line %d: unexpected token", p->line); \
+	log_error("Parse error at byte %" PRIptrdiff_t " (line %d): unexpected token", \
+		  p->tb - p->fb + 1, p->line); \
       return 0;\
    } \
 } while(0);
@@ -95,18 +109,18 @@ static int _tok_match(const char *str, const char *b, const char *e)
 /*
  * public interface
  */
-struct config_tree *create_config_tree(const char *filename)
+struct config_tree *create_config_tree(const char *filename, int keep_open)
 {
 	struct cs *c;
 	struct dm_pool *mem = dm_pool_create("config", 10 * 1024);
 
 	if (!mem) {
-		stack;
+		log_error("Failed to allocate config pool.");
 		return 0;
 	}
 
 	if (!(c = dm_pool_zalloc(mem, sizeof(*c)))) {
-		stack;
+		log_error("Failed to allocate config tree.");
 		dm_pool_destroy(mem);
 		return 0;
 	}
@@ -115,6 +129,8 @@ struct config_tree *create_config_tree(const char *filename)
 	c->cft.root = (struct config_node *) NULL;
 	c->timestamp = 0;
 	c->exists = 0;
+	c->keep_open = keep_open;
+	c->dev = 0;
 	if (filename)
 		c->filename = dm_pool_strdup(c->mem, filename);
 	return &c->cft;
@@ -122,7 +138,52 @@ struct config_tree *create_config_tree(const char *filename)
 
 void destroy_config_tree(struct config_tree *cft)
 {
-	dm_pool_destroy(((struct cs *) cft)->mem);
+	struct cs *c = (struct cs *) cft;
+
+	if (c->dev)
+		dev_close(c->dev);
+
+	dm_pool_destroy(c->mem);
+}
+
+static int _parse_config_file(struct parser *p, struct config_tree *cft)
+{
+	p->tb = p->te = p->fb;
+	p->line = 1;
+	_get_token(p, TOK_SECTION_E);
+	if (!(cft->root = _file(p)))
+		return_0;
+
+	return 1;
+}
+
+struct config_tree *create_config_tree_from_string(struct cmd_context *cmd __attribute((unused)),
+						   const char *config_settings)
+{
+	struct cs *c;
+	struct config_tree *cft;
+	struct parser *p;
+
+	if (!(cft = create_config_tree(NULL, 0)))
+		return_NULL;
+
+	c = (struct cs *) cft;
+	if (!(p = dm_pool_alloc(c->mem, sizeof(*p)))) {
+		log_error("Failed to allocate config tree parser.");
+		destroy_config_tree(cft);
+		return NULL;
+	}
+
+	p->mem = c->mem;
+	p->fb = config_settings;
+	p->fe = config_settings + strlen(config_settings);
+
+	if (!_parse_config_file(p, cft)) {
+		destroy_config_tree(cft);
+		return_NULL;
+	}
+
+	return cft;
 }
 
 int read_config_fd(struct config_tree *cft, struct device *dev,
@@ -134,11 +195,10 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 	int r = 0;
 	int use_mmap = 1;
 	off_t mmap_offset = 0;
+	char *buf = NULL;
 
-	if (!(p = dm_pool_alloc(c->mem, sizeof(*p)))) {
-		stack;
-		return 0;
-	}
+	if (!(p = dm_pool_alloc(c->mem, sizeof(*p))))
+		return_0;
 	p->mem = c->mem;
 
 	/* Only use mmap with regular files */
@@ -146,7 +206,7 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 		use_mmap = 0;
 
 	if (use_mmap) {
-		mmap_offset = offset % getpagesize();
+		mmap_offset = offset % lvm_getpagesize();
 		/* memory map the file */
 		p->fb = mmap((caddr_t) 0, size + mmap_offset, PROT_READ,
 			     MAP_PRIVATE, dev_fd(dev), offset - mmap_offset);
@@ -156,22 +216,13 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 		}
 		p->fb = p->fb + mmap_offset;
 	} else {
-		if (!(p->fb = dm_malloc(size + size2))) {
-			stack;
-			return 0;
-		}
-		if (!dev_read(dev, (uint64_t) offset, size, p->fb)) {
-			log_error("Read from %s failed", dev_name(dev));
+		if (!(buf = dm_malloc(size + size2)))
+			return_0;
+		if (!dev_read_circular(dev, (uint64_t) offset, size,
+				       (uint64_t) offset2, size2, buf)) {
 			goto out;
 		}
-		if (size2) {
-			if (!dev_read(dev, (uint64_t) offset2, size2,
-				      p->fb + size)) {
-				log_error("Circular read from %s failed",
-					  dev_name(dev));
-				goto out;
-			}
-		}
+		p->fb = buf;
 	}
 
 	if (checksum_fn && checksum !=
@@ -183,20 +234,14 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 
 	p->fe = p->fb + size + size2;
 
-	/* parse */
-	p->tb = p->te = p->fb;
-	p->line = 1;
-	_get_token(p, TOK_SECTION_E);
-	if (!(cft->root = _file(p))) {
-		stack;
-		goto out;
-	}
+	if (!_parse_config_file(p, cft))
+		goto_out;
 
 	r = 1;
 
       out:
 	if (!use_mmap)
-		dm_free(p->fb);
+		dm_free(buf);
 	else {
 		/* unmap the file */
 		if (munmap((char *) (p->fb - mmap_offset), size + mmap_offset)) {
@@ -212,7 +257,6 @@ int read_config_file(struct config_tree *cft)
 {
 	struct cs *c = (struct cs *) cft;
 	struct stat info;
-	struct device *dev;
 	int r = 1;
 
 	if (stat(c->filename, &info)) {
@@ -234,22 +278,23 @@ int read_config_file(struct config_tree *cft)
 		return 1;
 	}
 
-	if (!(dev = dev_create_file(c->filename, NULL, NULL, 1))) {
-		stack;
-		return 0;
+	if (!c->dev) {
+		if (!(c->dev = dev_create_file(c->filename, NULL, NULL, 1)))
+			return_0;
+
+		if (!dev_open_flags(c->dev, O_RDONLY, 0, 0))
+			return_0;
 	}
 
-	if (!dev_open_flags(dev, O_RDONLY, 0, 0)) {
-		stack;
-		return 0;
-	}
-
-	r = read_config_fd(cft, dev, 0, (size_t) info.st_size, 0, 0,
+	r = read_config_fd(cft, c->dev, 0, (size_t) info.st_size, 0, 0,
 			   (checksum_fn_t) NULL, 0);
 
-	dev_close(dev);
+	if (!c->keep_open) {
+		dev_close(c->dev);
+		c->dev = 0;
+	}
 
-	c->timestamp = info.st_mtime;
+	c->timestamp = info.st_ctime;
 
 	return r;
 }
@@ -293,7 +338,7 @@ int config_file_changed(struct config_tree *cft)
 	}
 
 	/* Unchanged? */
-	if (c->timestamp == info.st_mtime)
+	if (c->timestamp == info.st_ctime)
 		return 0;
 
       reload:
@@ -301,32 +346,96 @@ int config_file_changed(struct config_tree *cft)
 	return 1;
 }
 
-static void _write_value(FILE *fp, struct config_value *v)
+static int _line_start(struct output_line *outline)
 {
+	if (!dm_pool_begin_object(outline->mem, 128)) {
+		log_error("dm_pool_begin_object failed for config line");
+		return 0;
+	}
+
+	return 1;
+}
+
+static int _line_append(struct output_line *outline, const char *fmt, ...)
+  __attribute__ ((format(printf, 2, 3)));
+static int _line_append(struct output_line *outline, const char *fmt, ...)
+{
+	char buf[4096];
+	va_list ap;
+	int n;
+
+	va_start(ap, fmt);
+	n = vsnprintf(&buf[0], sizeof buf - 1, fmt, ap);
+	if (n < 0 || n > (int) sizeof buf - 1) {
+		log_error("vsnprintf failed for config line");
+		return 0;
+	}
+	va_end(ap);
+
+	if (!dm_pool_grow_object(outline->mem, &buf[0], strlen(buf))) {
+		log_error("dm_pool_grow_object failed for config line");
+		return 0;
+	}
+
+	return 1;
+}
+
+#define line_append(args...) do {if (!_line_append(outline, args)) {return_0;}} while (0)
+
+static int _line_end(struct output_line *outline)
+{
+	const char *line;
+
+	if (!dm_pool_grow_object(outline->mem, "\0", 1)) {
+		log_error("dm_pool_grow_object failed for config line");
+		return 0;
+	}
+
+	line = dm_pool_end_object(outline->mem);
+	if (!outline->fp)
+		log_print("%s", line);
+	else
+		fprintf(outline->fp, "%s\n", line);
+
+	return 1;
+}
+
+static int _write_value(struct output_line *outline, struct config_value *v)
+{
+	char *buf;
+
 	switch (v->type) {
 	case CFG_STRING:
-		fprintf(fp, "\"%s\"", v->v.str);
+		if (!(buf = alloca(escaped_len(v->v.str)))) {
+			log_error("temporary stack allocation for a config "
+				  "string failed");
+			return 0;
+		}
+		line_append("\"%s\"", escape_double_quotes(buf, v->v.str));
 		break;
 
 	case CFG_FLOAT:
-		fprintf(fp, "%f", v->v.r);
+		line_append("%f", v->v.r);
 		break;
 
 	case CFG_INT:
-		fprintf(fp, "%d", v->v.i);
+		line_append("%" PRId64, v->v.i);
 		break;
 
 	case CFG_EMPTY_ARRAY:
-		fprintf(fp, "[]");
+		line_append("[]");
 		break;
 
 	default:
 		log_error("_write_value: Unknown value type: %d", v->type);
 
 	}
+
+	return 1;
 }
 
-static int _write_config(struct config_node *n, FILE *fp, int level)
+static int _write_config(struct config_node *n, int only_one,
+			 struct output_line *outline, int level)
 {
 	char space[MAX_INDENT + 1];
 	int l = (level < MAX_INDENT) ? level : MAX_INDENT;
@@ -339,58 +448,87 @@ static int _write_config(struct config_node *n, FILE *fp, int level)
 		space[i] = '\t';
 	space[i] = '\0';
 
-	while (n) {
-		fprintf(fp, "%s%s", space, n->key);
+	do {
+		if (!_line_start(outline))
+			return_0;
+		line_append("%s%s", space, n->key);
 		if (!n->v) {
 			/* it's a sub section */
-			fprintf(fp, " {\n");
-			_write_config(n->child, fp, level + 1);
-			fprintf(fp, "%s}", space);
+			line_append(" {");
+			if (!_line_end(outline))
+				return_0;
+			_write_config(n->child, 0, outline, level + 1);
+			if (!_line_start(outline))
+				return_0;
+			line_append("%s}", space);
 		} else {
 			/* it's a value */
 			struct config_value *v = n->v;
-			fprintf(fp, "=");
+			line_append("=");
 			if (v->next) {
-				fprintf(fp, "[");
+				line_append("[");
 				while (v) {
-					_write_value(fp, v);
+					if (!_write_value(outline, v))
+						return_0;
 					v = v->next;
 					if (v)
-						fprintf(fp, ", ");
+						line_append(", ");
 				}
-				fprintf(fp, "]");
+				line_append("]");
 			} else
-				_write_value(fp, v);
+				if (!_write_value(outline, v))
+					return_0;
 		}
-		fprintf(fp, "\n");
+		if (!_line_end(outline))
+			return_0;
 		n = n->sib;
-	}
+	} while (n && !only_one);
 	/* FIXME: add error checking */
 	return 1;
 }
 
-int write_config_file(struct config_tree *cft, const char *file)
+int write_config_file(struct config_tree *cft, const char *file,
+		      int argc, char **argv)
 {
+	struct config_node *cn;
 	int r = 1;
-	FILE *fp;
+	struct output_line outline;
+	outline.fp = NULL;
 
-	if (!file) {
-		fp = stdout;
+	if (!file)
 		file = "stdout";
-	} else if (!(fp = fopen(file, "w"))) {
+	else if (!(outline.fp = fopen(file, "w"))) {
 		log_sys_error("open", file);
 		return 0;
 	}
 
+	outline.mem = dm_pool_create("config_line", 1024);
+
 	log_verbose("Dumping configuration to %s", file);
-	if (!_write_config(cft->root, fp, 0)) {
-		log_error("Failure while writing configuration");
+	if (!argc) {
+		if (!_write_config(cft->root, 0, &outline, 0)) {
+			log_error("Failure while writing to %s", file);
+			r = 0;
+		}
+	} else while (argc--) {
+		if ((cn = find_config_node(cft->root, *argv))) {
+			if (!_write_config(cn, 1, &outline, 0)) {
+				log_error("Failure while writing to %s", file);
+				r = 0;
+			}
+		} else {
+			log_error("Configuration node %s not found", *argv);
+			r = 0;
+		}
+		argv++;
+	}
+
+	if (outline.fp && lvm_fclose(outline.fp, file)) {
+		stack;
 		r = 0;
 	}
 
-	if (fp != stdout)
-		fclose(fp);
-
+	dm_pool_destroy(outline.mem);
 	return r;
 }
 
@@ -401,10 +539,8 @@ static struct config_node *_file(struct parser *p)
 {
 	struct config_node *root = NULL, *n, *l = NULL;
 	while (p->t != TOK_EOF) {
-		if (!(n = _section(p))) {
-			stack;
-			return 0;
-		}
+		if (!(n = _section(p)))
+			return_0;
 
 		if (!root)
 			root = n;
@@ -417,27 +553,21 @@ static struct config_node *_file(struct parser *p)
 
 static struct config_node *_section(struct parser *p)
 {
-	/* IDENTIFIER '{' VALUE* '}' */
+	/* IDENTIFIER SECTION_B_CHAR VALUE* SECTION_E_CHAR */
 	struct config_node *root, *n, *l = NULL;
-	if (!(root = _create_node(p))) {
-		stack;
-		return 0;
-	}
+	if (!(root = _create_node(p)))
+		return_0;
 
-	if (!(root->key = _dup_tok(p))) {
-		stack;
-		return 0;
-	}
+	if (!(root->key = _dup_tok(p)))
+		return_0;
 
 	match(TOK_IDENTIFIER);
 
 	if (p->t == TOK_SECTION_B) {
 		match(TOK_SECTION_B);
 		while (p->t != TOK_SECTION_E) {
-			if (!(n = _section(p))) {
-				stack;
-				return 0;
-			}
+			if (!(n = _section(p)))
+				return_0;
 
 			if (!root->child)
 				root->child = n;
@@ -448,10 +578,8 @@ static struct config_node *_section(struct parser *p)
 		match(TOK_SECTION_E);
 	} else {
 		match(TOK_EQ);
-		if (!(root->v = _value(p))) {
-			stack;
-			return 0;
-		}
+		if (!(root->v = _value(p)))
+			return_0;
 	}
 
 	return root;
@@ -464,10 +592,8 @@ static struct config_value *_value(struct parser *p)
 	if (p->t == TOK_ARRAY_B) {
 		match(TOK_ARRAY_B);
 		while (p->t != TOK_ARRAY_E) {
-			if (!(l = _type(p))) {
-				stack;
-				return 0;
-			}
+			if (!(l = _type(p)))
+				return_0;
 
 			if (!h)
 				h = l;
@@ -506,7 +632,7 @@ static struct config_value *_type(struct parser *p)
 	switch (p->t) {
 	case TOK_INT:
 		v->type = CFG_INT;
-		v->v.i = strtol(p->tb, NULL, 0);	/* FIXME: check error */
+		v->v.i = strtoll(p->tb, NULL, 0);	/* FIXME: check error */
 		match(TOK_INT);
 		break;
 
@@ -520,16 +646,26 @@ static struct config_value *_type(struct parser *p)
 		v->type = CFG_STRING;
 
 		p->tb++, p->te--;	/* strip "'s */
-		if (!(v->v.str = _dup_tok(p))) {
-			stack;
-			return 0;
-		}
+		if (!(v->v.str = _dup_tok(p)))
+			return_0;
 		p->te++;
 		match(TOK_STRING);
 		break;
 
+	case TOK_STRING_ESCAPED:
+		v->type = CFG_STRING;
+
+		p->tb++, p->te--;	/* strip "'s */
+		if (!(v->v.str = _dup_tok(p)))
+			return_0;
+		unescape_double_quotes(v->v.str);
+		p->te++;
+		match(TOK_STRING_ESCAPED);
+		break;
+
 	default:
-		log_error("Parse error at line %d: expected a value", p->line);
+		log_error("Parse error at byte %" PRIptrdiff_t " (line %d): expected a value",
+			  p->tb - p->fb + 1, p->line);
 		return 0;
 	}
 	return v;
@@ -566,12 +702,12 @@ static void _get_token(struct parser *p, int tok_prev)
 	p->t = TOK_INT;		/* fudge so the fall through for
 				   floats works */
 	switch (*p->te) {
-	case '{':
+	case SECTION_B_CHAR:
 		p->t = TOK_SECTION_B;
 		p->te++;
 		break;
 
-	case '}':
+	case SECTION_E_CHAR:
 		p->t = TOK_SECTION_E;
 		p->te++;
 		break;
@@ -597,7 +733,7 @@ static void _get_token(struct parser *p, int tok_prev)
 		break;
 
 	case '"':
-		p->t = TOK_STRING;
+		p->t = TOK_STRING_ESCAPED;
 		p->te++;
 		while ((p->te != p->fe) && (*p->te) && (*p->te != '"')) {
 			if ((*p->te == '\\') && (p->te + 1 != p->fe) &&
@@ -651,8 +787,9 @@ static void _get_token(struct parser *p, int tok_prev)
 	default:
 		p->t = TOK_IDENTIFIER;
 		while ((p->te != p->fe) && (*p->te) && !isspace(*p->te) &&
-		       (*p->te != '#') && (*p->te != '=') && (*p->te != '{') &&
-		       (*p->te != '}'))
+		       (*p->te != '#') && (*p->te != '=') &&
+		       (*p->te != SECTION_B_CHAR) &&
+		       (*p->te != SECTION_E_CHAR))
 			p->te++;
 		break;
 	}
@@ -661,11 +798,9 @@ static void _get_token(struct parser *p, int tok_prev)
 static void _eat_space(struct parser *p)
 {
 	while ((p->tb != p->fe) && (*p->tb)) {
-		if (*p->te == '#') {
+		if (*p->te == '#')
 			while ((p->te != p->fe) && (*p->te) && (*p->te != '\n'))
 				p->te++;
-			p->line++;
-		}
 
 		else if (isspace(*p->te)) {
 			while ((p->te != p->fe) && (*p->te) && isspace(*p->te)) {
@@ -688,14 +823,20 @@ static void _eat_space(struct parser *p)
 static struct config_value *_create_value(struct parser *p)
 {
 	struct config_value *v = dm_pool_alloc(p->mem, sizeof(*v));
-	memset(v, 0, sizeof(*v));
+
+	if (v)
+		memset(v, 0, sizeof(*v));
+
 	return v;
 }
 
 static struct config_node *_create_node(struct parser *p)
 {
 	struct config_node *n = dm_pool_alloc(p->mem, sizeof(*n));
-	memset(n, 0, sizeof(*n));
+
+	if (n)
+		memset(n, 0, sizeof(*n));
+
 	return n;
 }
 
@@ -703,10 +844,8 @@ static char *_dup_tok(struct parser *p)
 {
 	size_t len = p->te - p->tb;
 	char *str = dm_pool_alloc(p->mem, len + 1);
-	if (!str) {
-		stack;
-		return 0;
-	}
+	if (!str)
+		return_0;
 	strncpy(str, p->tb, len);
 	str[len] = '\0';
 	return str;
@@ -715,10 +854,11 @@ static char *_dup_tok(struct parser *p)
 /*
  * utility functions
  */
-struct config_node *find_config_node(const struct config_node *cn,
-				     const char *path)
+static struct config_node *_find_config_node(const struct config_node *cn,
+					     const char *path)
 {
 	const char *e;
+	const struct config_node *cn_found = NULL;
 
 	while (cn) {
 		/* trim any leading slashes */
@@ -729,32 +869,62 @@ struct config_node *find_config_node(const struct config_node *cn,
 		for (e = path; *e && (*e != sep); e++) ;
 
 		/* hunt for the node */
+		cn_found = NULL;
 		while (cn) {
-			if (_tok_match(cn->key, path, e))
-				break;
+			if (_tok_match(cn->key, path, e)) {
+				/* Inefficient */
+				if (!cn_found)
+					cn_found = cn;
+				else
+					log_error("WARNING: Ignoring duplicate"
+						  " config node: %s ("
+						  "seeking %s)", cn->key, path);
+			}
 
 			cn = cn->sib;
 		}
 
-		if (cn && *e)
-			cn = cn->child;
+		if (cn_found && *e)
+			cn = cn_found->child;
 		else
 			break;	/* don't move into the last node */
 
 		path = e;
 	}
 
-	return (struct config_node *) cn;
+	return (struct config_node *) cn_found;
 }
 
-const char *find_config_str(const struct config_node *cn,
-			    const char *path, const char *fail)
+static struct config_node *_find_first_config_node(const struct config_node *cn1,
+						   const struct config_node *cn2,
+						   const char *path)
 {
-	const struct config_node *n = find_config_node(cn, path);
+	struct config_node *cn;
 
-	if (n && n->v->type == CFG_STRING) {
-		if (*n->v->v.str)
-			log_very_verbose("Setting %s to %s", path, n->v->v.str);
+	if (cn1 && (cn = _find_config_node(cn1, path)))
+		return cn;
+
+	if (cn2 && (cn = _find_config_node(cn2, path)))
+		return cn;
+
+	return NULL;
+}
+
+struct config_node *find_config_node(const struct config_node *cn,
+				     const char *path)
+{
+	return _find_config_node(cn, path);
+}
+
+static const char *_find_config_str(const struct config_node *cn1,
+				    const struct config_node *cn2,
+				    const char *path, const char *fail)
+{
+	const struct config_node *n = _find_first_config_node(cn1, cn2, path);
+
+	/* Empty strings are ignored */
+	if ((n && n->v && n->v->type == CFG_STRING) && (*n->v->v.str)) {
+		log_very_verbose("Setting %s to %s", path, n->v->v.str);
 		return n->v->v.str;
 	}
 
@@ -764,26 +934,41 @@ const char *find_config_str(const struct config_node *cn,
 	return fail;
 }
 
-int find_config_int(const struct config_node *cn, const char *path, int fail)
+const char *find_config_str(const struct config_node *cn,
+			    const char *path, const char *fail)
 {
-	const struct config_node *n = find_config_node(cn, path);
+	return _find_config_str(cn, NULL, path, fail);
+}
 
-	if (n && n->v->type == CFG_INT) {
-		log_very_verbose("Setting %s to %d", path, n->v->v.i);
+static int64_t _find_config_int64(const struct config_node *cn1,
+				  const struct config_node *cn2,
+				  const char *path, int64_t fail)
+{
+	const struct config_node *n = _find_first_config_node(cn1, cn2, path);
+
+	if (n && n->v && n->v->type == CFG_INT) {
+		log_very_verbose("Setting %s to %" PRId64, path, n->v->v.i);
 		return n->v->v.i;
 	}
 
-	log_very_verbose("%s not found in config: defaulting to %d",
+	log_very_verbose("%s not found in config: defaulting to %" PRId64,
 			 path, fail);
 	return fail;
 }
 
-float find_config_float(const struct config_node *cn, const char *path,
-			float fail)
+int find_config_int(const struct config_node *cn, const char *path, int fail)
 {
-	const struct config_node *n = find_config_node(cn, path);
+	/* FIXME Add log_error message on overflow */
+	return (int) _find_config_int64(cn, NULL, path, (int64_t) fail);
+}
 
-	if (n && n->v->type == CFG_FLOAT) {
+static float _find_config_float(const struct config_node *cn1,
+				const struct config_node *cn2,
+				const char *path, float fail)
+{
+	const struct config_node *n = _find_first_config_node(cn1, cn2, path);
+
+	if (n && n->v && n->v->type == CFG_FLOAT) {
 		log_very_verbose("Setting %s to %f", path, n->v->v.r);
 		return n->v->v.r;
 	}
@@ -795,7 +980,38 @@ float find_config_float(const struct config_node *cn, const char *path,
 
 }
 
-static int _str_in_array(const char *str, const char *values[])
+float find_config_float(const struct config_node *cn, const char *path,
+			float fail)
+{
+	return _find_config_float(cn, NULL, path, fail);
+}
+
+struct config_node *find_config_tree_node(struct cmd_context *cmd,
+					  const char *path)
+{
+	return _find_first_config_node(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path);
+}
+
+const char *find_config_tree_str(struct cmd_context *cmd,
+				 const char *path, const char *fail)
+{
+	return _find_config_str(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path, fail);
+}
+
+int find_config_tree_int(struct cmd_context *cmd, const char *path,
+			 int fail)
+{
+	/* FIXME Add log_error message on overflow */
+	return (int) _find_config_int64(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path, (int64_t) fail);
+}
+
+float find_config_tree_float(struct cmd_context *cmd, const char *path,
+			     float fail)
+{
+	return _find_config_float(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path, fail);
+}
+
+static int _str_in_array(const char *str, const char * const values[])
 {
 	int i;
 
@@ -808,9 +1024,8 @@ static int _str_in_array(const char *str, const char *values[])
 
 static int _str_to_bool(const char *str, int fail)
 {
-	static const char *_true_values[] = { "y", "yes", "on", "true", NULL };
-	static const char *_false_values[] =
-	    { "n", "no", "off", "false", NULL };
+	const char * const _true_values[]  = { "y", "yes", "on", "true", NULL };
+	const char * const _false_values[] = { "n", "no", "off", "false", NULL };
 
 	if (_str_in_array(str, _true_values))
 		return 1;
@@ -821,9 +1036,11 @@ static int _str_to_bool(const char *str, int fail)
 	return fail;
 }
 
-int find_config_bool(const struct config_node *cn, const char *path, int fail)
+static int _find_config_bool(const struct config_node *cn1,
+			     const struct config_node *cn2,
+			     const char *path, int fail)
 {
-	const struct config_node *n = find_config_node(cn, path);
+	const struct config_node *n = _find_first_config_node(cn1, cn2, path);
 	struct config_value *v;
 
 	if (!n)
@@ -842,6 +1059,16 @@ int find_config_bool(const struct config_node *cn, const char *path, int fail)
 	return fail;
 }
 
+int find_config_bool(const struct config_node *cn, const char *path, int fail)
+{
+	return _find_config_bool(cn, NULL, path, fail);
+}
+
+int find_config_tree_bool(struct cmd_context *cmd, const char *path, int fail)
+{
+	return _find_config_bool(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path, fail);
+}
+
 int get_config_uint32(const struct config_node *cn, const char *path,
 		      uint32_t *result)
 {
@@ -866,7 +1093,6 @@ int get_config_uint64(const struct config_node *cn, const char *path,
 	if (!n || !n->v || n->v->type != CFG_INT)
 		return 0;
 
-	/* FIXME Support 64-bit value! */
 	*result = (uint64_t) n->v->v.i;
 	return 1;
 }
@@ -939,7 +1165,7 @@ static void _merge_section(struct config_node *cn1, struct config_node *cn2)
 	}
 }
 
-static int _match_host_tags(struct list *tags, struct config_node *tn)
+static int _match_host_tags(struct dm_list *tags, struct config_node *tn)
 {
 	struct config_value *tv;
 	const char *str;
@@ -996,3 +1222,61 @@ int merge_config_tree(struct cmd_context *cmd, struct config_tree *cft,
 
 	return 1;
 }
+
+/*
+ * Convert a token type to the char it represents.
+ */
+static char _token_type_to_char(int type)
+{
+	switch (type) {
+		case TOK_SECTION_B:
+			return SECTION_B_CHAR;
+		case TOK_SECTION_E:
+			return SECTION_E_CHAR;
+		default:
+			return 0;
+	}
+}
+
+/*
+ * Returns:
+ *  # of 'type' tokens in 'str'.
+ */
+static unsigned _count_tokens(const char *str, unsigned len, int type)
+{
+	char c;
+
+	c = _token_type_to_char(type);
+
+	return count_chars(str, len, c);
+}
+
+/*
+ * Heuristic function to make a quick guess as to whether a text
+ * region probably contains a valid config "section".  (Useful for
+ * scanning areas of the disk for old metadata.)
+ * Config sections contain various tokens, may contain other sections
+ * and strings, and are delimited by begin (type 'TOK_SECTION_B') and
+ * end (type 'TOK_SECTION_E') tokens.  As a quick heuristic, we just
+ * count the number of begin and end tokens, and see if they are
+ * non-zero and the counts match.
+ * Full validation of the section should be done with another function
+ * (for example, read_config_fd).
+ *
+ * Returns:
+ *  0 - probably is not a valid config section
+ *  1 - probably _is_ a valid config section
+ */
+unsigned maybe_config_section(const char *str, unsigned len)
+{
+	int begin_count;
+	int end_count;
+
+	begin_count = _count_tokens(str, len, TOK_SECTION_B);
+	end_count = _count_tokens(str, len, TOK_SECTION_E);
+
+	if (begin_count && end_count && (begin_count == end_count))
+		return 1;
+	else
+		return 0;
+}

lib/config/config.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -31,7 +31,7 @@ enum {
 struct config_value {
 	int type;
 	union {
-		int i;
+		int64_t i;
 		float r;
 		char *str;
 	} v;
@@ -49,21 +49,24 @@ struct config_tree {
 };
 
 struct config_tree_list {
-	struct list list;
+	struct dm_list list;
 	struct config_tree *cft;
 };
 
-struct config_tree *create_config_tree(const char *filename);
+struct config_tree *create_config_tree(const char *filename, int keep_open);
+struct config_tree *create_config_tree_from_string(struct cmd_context *cmd,
+						   const char *config_settings);
 void destroy_config_tree(struct config_tree *cft);
 
-typedef uint32_t (*checksum_fn_t) (uint32_t initial, void *buf, uint32_t size);
+typedef uint32_t (*checksum_fn_t) (uint32_t initial, const void *buf, uint32_t size);
 
 int read_config_fd(struct config_tree *cft, struct device *dev,
 		   off_t offset, size_t size, off_t offset2, size_t size2,
 		   checksum_fn_t checksum_fn, uint32_t checksum);
 
 int read_config_file(struct config_tree *cft);
-int write_config_file(struct config_tree *cft, const char *file);
+int write_config_file(struct config_tree *cft, const char *file,
+		      int argc, char **argv);
 time_t config_file_timestamp(struct config_tree *cft);
 int config_file_changed(struct config_tree *cft);
 int merge_config_tree(struct cmd_context *cmd, struct config_tree *cft,
@@ -71,20 +74,30 @@ int merge_config_tree(struct cmd_context *cmd, struct config_tree *cft,
 
 struct config_node *find_config_node(const struct config_node *cn,
 				     const char *path);
-
 const char *find_config_str(const struct config_node *cn, const char *path,
 			    const char *fail);
-
 int find_config_int(const struct config_node *cn, const char *path, int fail);
-
 float find_config_float(const struct config_node *cn, const char *path,
 			float fail);
 
+/*
+ * These versions check an override tree, if present, first.
+ */
+struct config_node *find_config_tree_node(struct cmd_context *cmd,
+					  const char *path);
+const char *find_config_tree_str(struct cmd_context *cmd,
+				 const char *path, const char *fail);
+int find_config_tree_int(struct cmd_context *cmd, const char *path,
+			 int fail);
+float find_config_tree_float(struct cmd_context *cmd, const char *path,
+			     float fail);
+
 /*
  * Understands (0, ~0), (y, n), (yes, no), (on,
  * off), (true, false).
  */
 int find_config_bool(const struct config_node *cn, const char *path, int fail);
+int find_config_tree_bool(struct cmd_context *cmd, const char *path, int fail);
 
 int get_config_uint32(const struct config_node *cn, const char *path,
 		      uint32_t *result);
@@ -95,4 +108,6 @@ int get_config_uint64(const struct config_node *cn, const char *path,
 int get_config_str(const struct config_node *cn, const char *path,
 		   char **result);
 
+unsigned maybe_config_section(const char *str, unsigned len);
+
 #endif

lib/config/defaults.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -21,6 +21,8 @@
 
 #define DEFAULT_ARCHIVE_SUBDIR "archive"
 #define DEFAULT_BACKUP_SUBDIR "backup"
+#define DEFAULT_CACHE_SUBDIR "cache"
+#define DEFAULT_CACHE_FILE_PREFIX ""
 
 #define DEFAULT_ARCHIVE_DAYS 30
 #define DEFAULT_ARCHIVE_NUMBER 10
@@ -30,11 +32,19 @@
 #define DEFAULT_PROC_DIR "/proc"
 #define DEFAULT_SYSFS_SCAN 1
 #define DEFAULT_MD_COMPONENT_DETECTION 1
+#define DEFAULT_MD_CHUNK_ALIGNMENT 1
+#define DEFAULT_IGNORE_SUSPENDED_DEVICES 1
 
 #define DEFAULT_LOCK_DIR "/var/lock/lvm"
-#define DEFAULT_LOCKING_LIB "lvm2_locking.so"
+#define DEFAULT_LOCKING_LIB "liblvm2clusterlock.so"
+#define DEFAULT_FALLBACK_TO_LOCAL_LOCKING 1
+#define DEFAULT_FALLBACK_TO_CLUSTERED_LOCKING 1
 
+#define DEFAULT_MIRRORLOG "disk"
+#define DEFAULT_MIRROR_LOG_FAULT_POLICY "allocate"
+#define DEFAULT_MIRROR_DEV_FAULT_POLICY "remove"
 #define DEFAULT_DMEVENTD_MIRROR_LIB "libdevmapper-event-lvm2mirror.so"
+#define DEFAULT_DMEVENTD_MONITOR 1
 
 #define DEFAULT_UMASK 0077
 
@@ -54,6 +64,8 @@
 #define DEFAULT_PVMETADATASIZE 255
 #define DEFAULT_PVMETADATACOPIES 1
 #define DEFAULT_LABELSECTOR UINT64_C(1)
+#define DEFAULT_READ_AHEAD "auto"
+#define DEFAULT_EXTENT_SIZE 4096	/* In KB */
 
 #define DEFAULT_MSG_PREFIX "  "
 #define DEFAULT_CMD_NAME 0
@@ -80,7 +92,7 @@
 #  define DEFAULT_ACTIVATION 0
 #endif
 
-#define DEFAULT_STRIPE_FILLER "/dev/ioerror"
+#define DEFAULT_STRIPE_FILLER "error"
 #define DEFAULT_MIRROR_REGION_SIZE 512	/* KB */
 #define DEFAULT_INTERVAL 15
 
@@ -90,20 +102,23 @@
 
 #define DEFAULT_REP_ALIGNED 1
 #define DEFAULT_REP_BUFFERED 1
+#define DEFAULT_REP_COLUMNS_AS_ROWS 0
 #define DEFAULT_REP_HEADINGS 1
+#define DEFAULT_REP_PREFIXES 0
+#define DEFAULT_REP_QUOTED 1
 #define DEFAULT_REP_SEPARATOR " "
 
-#define DEFAULT_LVS_COLS "lv_name,vg_name,lv_attr,lv_size,origin,snap_percent,move_pv,mirror_log,copy_percent"
+#define DEFAULT_LVS_COLS "lv_name,vg_name,lv_attr,lv_size,origin,snap_percent,move_pv,mirror_log,copy_percent,convert_lv"
 #define DEFAULT_VGS_COLS "vg_name,pv_count,lv_count,snap_count,vg_attr,vg_size,vg_free"
 #define DEFAULT_PVS_COLS "pv_name,vg_name,pv_fmt,pv_attr,pv_size,pv_free"
 #define DEFAULT_SEGS_COLS "lv_name,vg_name,lv_attr,stripes,segtype,seg_size"
 #define DEFAULT_PVSEGS_COLS "pv_name,vg_name,pv_fmt,pv_attr,pv_size,pv_free,pvseg_start,pvseg_size"
 
-#define DEFAULT_LVS_COLS_VERB "lv_name,vg_name,seg_count,lv_attr,lv_size,lv_major,lv_minor,lv_kernel_major,lv_kernel_minor,origin,snap_percent,move_pv,copy_percent,mirror_log,lv_uuid"
+#define DEFAULT_LVS_COLS_VERB "lv_name,vg_name,seg_count,lv_attr,lv_size,lv_major,lv_minor,lv_kernel_major,lv_kernel_minor,origin,snap_percent,move_pv,copy_percent,mirror_log,convert_lv,lv_uuid"
 #define DEFAULT_VGS_COLS_VERB "vg_name,vg_attr,vg_extent_size,pv_count,lv_count,snap_count,vg_size,vg_free,vg_uuid"
 #define DEFAULT_PVS_COLS_VERB "pv_name,vg_name,pv_fmt,pv_attr,pv_size,pv_free,dev_size,pv_uuid"
 #define DEFAULT_SEGS_COLS_VERB "lv_name,vg_name,lv_attr,seg_start,seg_size,stripes,segtype,stripesize,chunksize"
-#define DEFAULT_PVSEGS_COLS_VERB "pv_name,vg_name,pv_fmt,pv_attr,pv_size,pv_free,pvseg_start,pvseg_size"
+#define DEFAULT_PVSEGS_COLS_VERB "pv_name,vg_name,pv_fmt,pv_attr,pv_size,pv_free,pvseg_start,pvseg_size,lv_name,seg_start_pe,segtype,seg_pe_ranges"
 
 #define DEFAULT_LVS_SORT "vg_name,lv_name"
 #define DEFAULT_VGS_SORT "vg_name"
@@ -111,4 +126,7 @@
 #define DEFAULT_SEGS_SORT "vg_name,lv_name,seg_start"
 #define DEFAULT_PVSEGS_SORT "pv_name,pvseg_start"
 
+#define DEFAULT_MIRROR_DEVICE_FAULT_POLICY "remove"
+#define DEFAULT_MIRROR_LOG_FAULT_POLICY "allocate"
+
 #endif				/* _LVM_DEFAULTS_H */

lib/datastruct/btree.c

@@ -1,14 +1,14 @@
 /*
- * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2001-2004 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -55,7 +55,8 @@ static uint32_t _shuffle(uint32_t k)
 #endif
 }
 
-static struct node **_lookup(struct node **c, uint32_t key, struct node **p)
+static struct node **_lookup(struct node *const *c, uint32_t key,
+			     struct node **p)
 {
 	*p = NULL;
 	while (*c) {
@@ -70,10 +71,10 @@ static struct node **_lookup(struct node **c, uint32_t key, struct node **p)
 			c = &(*c)->r;
 	}
 
-	return c;
+	return (struct node **)c;
 }
 
-void *btree_lookup(struct btree *t, uint32_t k)
+void *btree_lookup(const struct btree *t, uint32_t k)
 {
 	uint32_t key = _shuffle(k);
 	struct node *p, **c = _lookup(&t->root, key, &p);
@@ -86,10 +87,8 @@ int btree_insert(struct btree *t, uint32_t k, void *data)
 	struct node *p, **c = _lookup(&t->root, key, &p), *n;
 
 	if (!*c) {
-		if (!(n = dm_pool_alloc(t->mem, sizeof(*n)))) {
-			stack;
-			return 0;
-		}
+		if (!(n = dm_pool_alloc(t->mem, sizeof(*n))))
+			return_0;
 
 		n->key = key;
 		n->data = data;
@@ -102,19 +101,19 @@ int btree_insert(struct btree *t, uint32_t k, void *data)
 	return 1;
 }
 
-void *btree_get_data(struct btree_iter *it)
+void *btree_get_data(const struct btree_iter *it)
 {
 	return ((struct node *) it)->data;
 }
 
-static inline struct node *_left(struct node *n)
+static struct node *_left(struct node *n)
 {
 	while (n->l)
 		n = n->l;
 	return n;
 }
 
-struct btree_iter *btree_first(struct btree *t)
+struct btree_iter *btree_first(const struct btree *t)
 {
 	if (!t->root)
 		return NULL;
@@ -122,7 +121,7 @@ struct btree_iter *btree_first(struct btree *t)
 	return (struct btree_iter *) _left(t->root);
 }
 
-struct btree_iter *btree_next(struct btree_iter *it)
+struct btree_iter *btree_next(const struct btree_iter *it)
 {
 	struct node *n = (struct node *) it;
 	uint32_t k = n->key;

lib/datastruct/btree.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
  * This copyrighted material is made available to anyone wishing to use,
  * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -20,13 +20,13 @@ struct btree;
 
 struct btree *btree_create(struct dm_pool *mem);
 
-void *btree_lookup(struct btree *t, uint32_t k);
+void *btree_lookup(const struct btree *t, uint32_t k);
 int btree_insert(struct btree *t, uint32_t k, void *data);
 
 struct btree_iter;
-void *btree_get_data(struct btree_iter *it);
+void *btree_get_data(const struct btree_iter *it);
 
-struct btree_iter *btree_first(struct btree *t);
-struct btree_iter *btree_next(struct btree_iter *it);
+struct btree_iter *btree_first(const struct btree *t);
+struct btree_iter *btree_next(const struct btree_iter *it);
 
 #endif

lib/datastruct/list.c

@@ -0,0 +1,145 @@
+/*
+ * Copyright (C) 2001-2004 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 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+
+#include "lib.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;
+}

lib/datastruct/list.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004 Red Hat, 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -24,147 +24,106 @@
  * The list head's pointers point to the first and the last element.
  */
 
-struct list {
-	struct list *n, *p;
+struct dm_list {
+	struct dm_list *n, *p;
 };
 
 /*
  * Initialise a list before use.
  * The list head's next and previous pointers point back to itself.
  */
-#define LIST_INIT(name)	struct list name = { &(name), &(name) }
-static inline void list_init(struct list *head)
-{
-	head->n = head->p = head;
-}
+#define DM_LIST_INIT(name)	struct dm_list name = { &(name), &(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.
  */
-static inline void list_add(struct list *head, struct list *elem)
-{
-	assert(head->n);
-
-	elem->n = head;
-	elem->p = head->p;
-
-	head->p->n = elem;
-	head->p = elem;
-}
+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.
  */
-static inline void list_add_h(struct list *head, struct list *elem)
-{
-	assert(head->n);
-
-	elem->n = head->n;
-	elem->p = head;
-
-	head->n->p = elem;
-	head->n = elem;
-}
+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.
  */
-static inline void list_del(struct list *elem)
-{
-	elem->n->p = elem->p;
-	elem->p->n = elem->n;
-}
+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);
 
 /*
  * Is the list empty?
  */
-static inline int list_empty(struct list *head)
-{
-	return head->n == head;
-}
+int dm_list_empty(const struct dm_list *head);
 
 /*
  * Is this the first element of the list?
  */
-static inline int list_start(struct list *head, struct list *elem)
-{
-	return elem->p == head;
-}
+int dm_list_start(const struct dm_list *head, const struct dm_list *elem);
 
 /*
  * Is this the last element of the list?
  */
-static inline int list_end(struct list *head, struct list *elem)
-{
-	return elem->n == head;
-}
+int dm_list_end(const struct dm_list *head, const struct dm_list *elem);
 
 /*
  * Return first element of the list or NULL if empty
  */
-static inline struct list *list_first(struct list *head)
-{
-	return (list_empty(head) ? NULL : head->n);
-}
+struct dm_list *dm_list_first(const struct dm_list *head);
 
 /*
  * Return last element of the list or NULL if empty
  */
-static inline struct list *list_last(struct list *head)
-{
-	return (list_empty(head) ? NULL : head->p);
-}
+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.
  */
-static inline struct list *list_prev(struct list *head, struct list *elem)
-{
-	return (list_start(head, elem) ? NULL : elem->p);
-}
+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.
  */
-static inline struct list *list_next(struct list *head, struct list *elem)
-{
-	return (list_end(head, elem) ? NULL : elem->n);
-}
+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 list' called 'head' 
+ * 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 list_struct_base(v, t, head) \
+#define dm_list_struct_base(v, t, head) \
     ((t *)((uintptr_t)(v) - (uintptr_t)&((t *) 0)->head))
 
 /*
- * Given the address v of an instance of 'struct list list' contained in
+ * Given the address v of an instance of 'struct dm_list list' contained in
  * a structure of type t, return the containing structure.
  */
-#define list_item(v, t) list_struct_base((v), t, list)
+#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 struct_field(v, t, e, 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 list_head(v, t, e) struct_field(v, t, e, 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 list_iterate(v, head) \
+#define dm_list_iterate(v, head) \
 	for (v = (head)->n; v != head; v = v->n)
 
 /*
@@ -174,7 +133,7 @@ static inline struct list *list_next(struct list *head, struct list *elem)
  * already-processed elements.
  * If 'start' is 'head' it walks the list backwards.
  */
-#define list_uniterate(v, head, start) \
+#define dm_list_uniterate(v, head, start) \
 	for (v = (start)->p; v != head; v = v->p)
 
 /*
@@ -182,77 +141,68 @@ static inline struct list *list_next(struct list *head, struct list *elem)
  * the way.
  * t must be defined as a temporary variable of the same type as v.
  */
-#define list_iterate_safe(v, t, head) \
+#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 list' variable within the containing structure is 'field'.
+ * The 'struct dm_list' variable within the containing structure is 'field'.
  */
-#define list_iterate_items_gen(v, head, field) \
-	for (v = list_struct_base((head)->n, typeof(*v), 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 = list_struct_base(v->field.n, typeof(*v), field))
+	     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 list list' within the containing structure.
+ * The list should be 'struct dm_list list' within the containing structure.
  */
-#define list_iterate_items(v, head) list_iterate_items_gen(v, (head), list)
+#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 list' variable within the containing structure is 'field'.
+ * 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 list_iterate_items_gen_safe(v, t, head, field) \
-	for (v = list_struct_base((head)->n, typeof(*v), field), \
-	     t = list_struct_base(v->field.n, typeof(*v), field); \
+#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 = list_struct_base(v->field.n, typeof(*v), field))
+	     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 list list' within the containing structure.
+ * 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 list_iterate_items_safe(v, t, head) \
-	list_iterate_items_gen_safe(v, t, (head), list)
+#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 list' variable within the containing structure is 'field'.
+ * The 'struct dm_list' variable within the containing structure is 'field'.
  */
-#define list_iterate_back_items_gen(v, head, field) \
-	for (v = list_struct_base((head)->p, typeof(*v), 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 = list_struct_base(v->field.p, typeof(*v), field))
+	     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 list list' within the containing structure.
+ * The list should be 'struct dm_list list' within the containing structure.
  */
-#define list_iterate_back_items(v, head) list_iterate_back_items_gen(v, (head), list)
+#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.
  */
-static inline unsigned int list_size(const struct list *head)
-{
-	unsigned int s = 0;
-	const struct list *v;
-
-	list_iterate(v, head)
-	    s++;
-
-	return s;
-}
+unsigned int dm_list_size(const struct dm_list *head);
 
 #endif

lib/datastruct/lvm-types.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, 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.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -16,16 +16,16 @@
 #ifndef _LVM_TYPES_H
 #define _LVM_TYPES_H
 
-#include "list.h"
-
 #include <sys/types.h>
 #include <inttypes.h>
 
 /* Define some portable printing types */
 #define PRIsize_t "zu"
+#define PRIptrdiff_t "td"
+#define PRIpid_t PRId32
 
 struct str_list {
-	struct list list;
+	struct dm_list list;
 	const char *str;
 };
 

lib/datastruct/str_list.c

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2003-2004 Sistina Software, Inc. All rights reserved.
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
  * This copyrighted material is made available to anyone wishing to use,
  * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -16,67 +16,60 @@
 #include "lib.h"
 #include "str_list.h"
 
-struct list *str_list_create(struct dm_pool *mem)
+struct dm_list *str_list_create(struct dm_pool *mem)
 {
-	struct list *sl;
+	struct dm_list *sl;
 
-	if (!(sl = dm_pool_alloc(mem, sizeof(struct list)))) {
-		stack;
-		return NULL;
-	}
+	if (!(sl = dm_pool_alloc(mem, sizeof(struct dm_list))))
+		return_NULL;
 
-	list_init(sl);
+	dm_list_init(sl);
 
 	return sl;
 }
 
-int str_list_add(struct dm_pool *mem, struct list *sll, const char *str)
+int str_list_add(struct dm_pool *mem, struct dm_list *sll, const char *str)
 {
 	struct str_list *sln;
 
-	if (!str) {
-		stack;
-		return 0;
-	}
+	if (!str)
+		return_0;
 
 	/* Already in list? */
 	if (str_list_match_item(sll, str))
 		return 1;
 
-	if (!(sln = dm_pool_alloc(mem, sizeof(*sln)))) {
-		stack;
-		return 0;
-	}
+	if (!(sln = dm_pool_alloc(mem, sizeof(*sln))))
+		return_0;
 
 	sln->str = str;
-	list_add(sll, &sln->list);
+	dm_list_add(sll, &sln->list);
 
 	return 1;
 }
 
-int str_list_del(struct list *sll, const char *str)
+int str_list_del(struct dm_list *sll, const char *str)
 {
-	struct list *slh, *slht;
+	struct dm_list *slh, *slht;
 
-	list_iterate_safe(slh, slht, sll) {
-		if (!strcmp(str, list_item(slh, struct str_list)->str))
-			 list_del(slh);
+	dm_list_iterate_safe(slh, slht, sll) {
+		if (!strcmp(str, dm_list_item(slh, struct str_list)->str))
+			 dm_list_del(slh);
 	}
 
 	return 1;
 }
 
-int str_list_dup(struct dm_pool *mem, struct list *sllnew, struct list *sllold)
+int str_list_dup(struct dm_pool *mem, struct dm_list *sllnew,
+		 const struct dm_list *sllold)
 {
 	struct str_list *sl;
 
-	list_init(sllnew);
+	dm_list_init(sllnew);
 
-	list_iterate_items(sl, sllold) {
-		if (!str_list_add(mem, sllnew, strdup(sl->str))) {
-			stack;
-			return 0;
-		}
+	dm_list_iterate_items(sl, sllold) {
+		if (!str_list_add(mem, sllnew, dm_pool_strdup(mem, sl->str)))
+			return_0;
 	}
 
 	return 1;
@@ -85,11 +78,11 @@ int str_list_dup(struct dm_pool *mem, struct list *sllnew, struct list *sllold)
 /*
  * Is item on list?
  */
-int str_list_match_item(struct list *sll, const char *str)
+int str_list_match_item(const struct dm_list *sll, const char *str)
 {
 	struct str_list *sl;
 
-	list_iterate_items(sl, sll)
+	dm_list_iterate_items(sl, sll)
 	    if (!strcmp(str, sl->str))
 		return 1;
 
@@ -99,11 +92,11 @@ int str_list_match_item(struct list *sll, const char *str)
 /*
  * Is at least one item on both lists?
  */
-int str_list_match_list(struct list *sll, struct list *sll2)
+int str_list_match_list(const struct dm_list *sll, const struct dm_list *sll2)
 {
 	struct str_list *sl;
 
-	list_iterate_items(sl, sll)
+	dm_list_iterate_items(sl, sll)
 	    if (str_list_match_item(sll2, sl->str))
 		return 1;
 
@@ -113,14 +106,14 @@ int str_list_match_list(struct list *sll, struct list *sll2)
 /*
  * Do both lists contain the same set of items?
  */
-int str_list_lists_equal(struct list *sll, struct list *sll2)
+int str_list_lists_equal(const struct dm_list *sll, const struct dm_list *sll2)
 {
 	struct str_list *sl;
 
-	if (list_size(sll) != list_size(sll2))
+	if (dm_list_size(sll) != dm_list_size(sll2))
 		return 0;
 
-	list_iterate_items(sl, sll)
+	dm_list_iterate_items(sl, sll)
 	    if (!str_list_match_item(sll2, sl->str))
 		return 0;
 

lib/datastruct/str_list.h

@@ -1,14 +1,14 @@
 /*
  * Copyright (C) 2003-2004 Sistina Software, Inc. All rights reserved.  
- * Copyright (C) 2004 Red Hat, Inc. All rights reserved.
+ * Copyright (C) 2004-2005 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
  * This copyrighted material is made available to anyone wishing to use,
  * modify, copy, or redistribute it subject to the terms and conditions
- * of the GNU General Public License v.2.
+ * of the GNU Lesser General Public License v.2.1.
  *
- * You should have received a copy of the GNU General Public License
+ * 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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  */
@@ -16,12 +16,13 @@
 #ifndef _LVM_STR_LIST_H
 #define _LVM_STR_LIST_H
 
-struct list *str_list_create(struct dm_pool *mem);
-int str_list_add(struct dm_pool *mem, struct list *sll, const char *str);
-int str_list_del(struct list *sll, const char *str);
-int str_list_match_item(struct list *sll, const char *str);
-int str_list_match_list(struct list *sll, struct list *sll2);
-int str_list_lists_equal(struct list *sll, struct list *sll2);
-int str_list_dup(struct dm_pool *mem, struct list *sllnew, struct list *sllold);
+struct dm_list *str_list_create(struct dm_pool *mem);
+int str_list_add(struct dm_pool *mem, struct dm_list *sll, const char *str);
+int str_list_del(struct dm_list *sll, const char *str);
+int str_list_match_item(const struct dm_list *sll, const char *str);
+int str_list_match_list(const struct dm_list *sll, const struct dm_list *sll2);
+int str_list_lists_equal(const struct dm_list *sll, const struct dm_list *sll2);
+int str_list_dup(struct dm_pool *mem, struct dm_list *sllnew,
+		 const struct dm_list *sllold);
 
 #endif