pre-release

rename variable to reduce confusion

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!
+
+

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) 2004-2007 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 = doc include man 
+SUBDIRS = doc include man scripts
 
 ifeq ("@INTL@", "yes")
   SUBDIRS += po
@@ -68,3 +68,6 @@ cscope.out: tools
 	@CSCOPE_CMD@ -b -R
 all: cscope.out
 endif
+
+check: all
+	$(MAKE) -C test all

VERSION

@@ -1 +1 @@
-2.02.08-cvs (2006-07-17)
+2.02.37-cvs (2008-06-06)

WHATS_NEW

@@ -1,5 +1,561 @@
-Version 2.02.08 - 
+Version 2.02.37 - 6th June 2008
+===============================
+  Make clvmd-cman use a hash rather than an array for node updown info.
+  Correct config file line numbers in messages when parsing comments.
+  Drop cached metadata when renaming a VG.
+  Allow for vginfo changing during _vg_read.
+  Decode numbers in clvmd debugging output.
+  Add missing deactivation after activation failure in lvcreate -Zy.
+  When activating, if precommitted metadata is still cached, assume it's live.
+  When removing LV symlinks, skip any where the VG name is not determined.
+  Drop metadata cache if update fails in vg_revert or vg_commit.
+  Avoid spurious duplicate VG messages referring to VGs that are gone.
+  Drop dev_name_confirmed error message to debug level.
+  Fix setpriority error message to signed int.
+  Temporarily disable dmeventd mirror monitoring during lvchange --resync.
+  Refactor some vginfo manipulation code.
+  Add assertions to trap deprecated P_ and V_ lock usage.
+  Add missing mutex around clvmd lvmcache_drop_metadata library call.
+  Fix uninitialised mutex in clvmd if all daemons are not running at startup.
+  Avoid using DLM locks with LCK_CACHE type P_ lock requests.
+  When asked to drop cached committed VG metadata, invalidate cached PV labels.
+  Drop metadata cache before writing precommitted metadata instead of after.
+  Don't touch /dev in vgrename if activation is disabled.
+
+Version 2.02.36 - 29th April 2008
+=================================
+  Fix fsadm.sh to work with older blockdev, blkid & readlink binaries.
+  Fix lvresize to pass new size to fsadm when extending device.
+  Remove unused struct in clvmd-openais, and use correct node count.
+  Fix nodes list in clvmd-openais, and allow for broadcast messages.
+  Exclude VG_GLOBAL from internal concurrent VG lock counter.
+  Fix vgsplit internal counting of snapshot LVs.
+  Fix vgmerge snapshot_count when source VG contains snapshots.
+  Simplify clvmd-openais by using non-async saLckResourceLock.
+  Fix internal LV counter when a snapshot is removed.
+  Fix metadata corruption writing lvm1-formatted metadata with snapshots.
+  Fix lvconvert -m0 allocatable space check.
+
+Version 2.02.35 - 15th April 2008
+=================================
+  Drop cached VG metadata before and after committing changes to it.
+  Rename P_global to P_#global.
+  Don't attempt remote metadata backups of non-clustered VGs. (2.02.29)
+  Don't store fid in VG metadata cache to avoid clvmd segfault. (2.02.34)
+  Update vgsplit tests to verify loosening of active LV restriction.
+  Update vgsplit to only restrict split with active LVs involved in split.
+  Add lv_is_active() to determine whether an lv is active.
+
+Version 2.02.34 - 10th April 2008
+=================================
+  Improve preferred_names lvm.conf example.
+  Fix vgdisplay 'Cur LV' field to match lvdisplay output.
+  Fix lv_count report field to exclude hidden LVs.
+  Add vg_is_clustered() helper function.
+  Fix vgsplit to only move hidden 'snapshotN' LVs when necessary.
+  Update vgsplit tests for lvnames on the cmdline.
+  Update vgsplit man page to reflect lvnames on the cmdline.
+  Update vgsplit to take "-n LogicalVolumeName" on the cmdline.
+  Use clustered mirror log with pvmove in clustered VGs, if available.
+  Fix some pvmove error status codes.
+  Fix vgsplit error paths to release vg_to lock.
+  Indicate whether or not VG is clustered in vgcreate log message.
+  Mention default --clustered setting in vgcreate man page.
+  Add config file overrides to clvmd when it reads the active LVs list.
+  Fix vgreduce to use vg_split_mdas to check sufficient mdas remain.
+  Add (empty) orphan VGs to lvmcache during initialisation.
+  Fix orphan VG name used for format_pool.
+  Create a fid for internal orphan VGs.
+  Update lvmcache VG lock state for all locking types now.
+  Fix output if overriding command_names on cmdline.
+  Add detection of clustered mirror log capability.
+  Add check to vg_commit() ensuring VG lock held before writing new VG metadata.
+  Add validation of LV name to pvmove -n.
+  Make clvmd refresh the context correctly when lvm.conf is updated.
+  Add some basic internal VG lock validation.
+  Add per-command flags to control which commands use the VG metadata cache.
+  Fix vgsplit locking of new VG (2.02.30).
+  Avoid erroneous vgsplit error message for new VG. (2.02.29)
+  Suppress duplicate message when lvresize fails because of invalid vgname.
+  Cache VG metadata internally while VG lock is held.
+  Fix redundant lvresize message if vg doesn't exist.
+  Fix another allocation bug with clvmd and large node IDs.
+  Add find_lv_in_lv_list() and find_pv_in_pv_list().
+  Fix uninitialised variable in clvmd that could cause odd hangs.
+  Add vgmerge tests.
+  Add pvseg_is_allocated() for identifying a PV segment allocated to a LV.
+  Add list_move() for moving elements from one list to another.
+  Add 'is_reserved_lvname()' for identifying hidden LVs.
+  Correct command name in lvmdiskscan man page.
+  clvmd no longer crashes if it sees nodeids over 50.
+  Fix potential deadlock in clvmd thread handling.
+  Refactor text format initialisation into _init_text_import.
+  Escape double quotes and backslashes in external metadata and config data.
+  Add functions for escaping double quotes in strings.
+  Rename count_chars_len to count_chars.
+  Use return_0 in a couple more places.
+  Correct a function name typo in _line_append error message.
+  Include limits.h in clvmd so it compiles with newer headers.
+  Add VirtIO disks (virtblk) to filters.
+  Fix resetting of MIRROR_IMAGE and VISIBLE_LV after removal of LV. (2.02.30)
+  Fix remove_layer_from_lv to empty the LV before removing it. (2.02.30)
+  Add missing no-longer-used segs_using_this_lv test to check_lv_segments.
+  Remove redundant non-NULL tests before calling free in clvmd.c.
+  Avoid a compiler warning: make is_orphan's parameter const.
+  Fix lvconvert detection of mirror conversion in progress. (2.02.30)
+  Avoid automatic lvconvert polldaemon invocation when -R specified. (2.02.30)
+  Fix 'pvs -a' to detect VGs of PVs without metadata areas.
+  Divide up internal orphan volume group by format type.
+  Update usage message for clvmd.
+  Fix clvmd man page not to print <br> and clarified debug options.
+  Fix lvresize to support /dev/mapper prefix in the LV name.
+  Fix unfilled parameter passed to fsadm from lvresize.
+  Update fsadm to call lvresize if the partition size differs (with option -l).
+  Fix fsadm to support VG/LV names.
+
+Version 2.02.33 - 31st January 2008
+===================================
+  Fix mirror log name construction during lvconvert. (2.02.30)
+  Make monitor_dev_for_events recurse through the stack of LVs.
+  Clean up some more compiler warnings.
+  Some whitespace tidy-ups.
+  Use stack return macros throughout.
+  Rely upon internally-cached PV labels while corresponding VG lock is held.
+
+Version 2.02.32 - 29th January 2008
+===================================
+  Fix two check_lv_segments error messages to show whole segment.
+  Refactor mirror log attachment code.
+  Fix internal metadata corruption in lvchange --resync. (2.02.30)
+  Fix new parameter validation in vgsplit and test mode. (2.02.30)
+  Remove redundant cnxman-socket.h file from clvmd directory.
+  Fix pvs, vgs, lvs error exit status on some error paths.
+
+Version 2.02.31 - 19th January 2008
+===================================
+  Fix lvcreate --nosync not to wait for non-happening sync. (2.02.30)
+  Add very_verbose lvconvert messages.
+  Avoid readahead error message with default setting of lvcreate -M1. (2.02.29)
+
+Version 2.02.30 - 17th January 2008
+===================================
+  Set default readahead to twice maximium stripe size.
+  Reinstate VG extent size and stripe size defaults (halved). (2.02.29)
+  Add lists of stacked LV segments using each LV to the internal metadata.
+  Change vgsplit -l (for unimplemented --list) into --maxlogicalvolumes.
+  Fix process_all_pvs to detect non-orphans with no MDAs correctly.
+  Don't use block_on_error with mirror targets version 1.12 and above.
+  Update vgsplit to accept vgcreate options when new VG is destination.
+  Update vgsplit to accept existing VG as destination.
+  lvconvert waits for completion of initial sync by default.
+  Refactor vgcreate for parameter validation and add tests.
+  Add new convert_lv field to lvs output.
+  Print warning when lvm tools are running as non-root.
+  Add snapshot dmeventd library (enables dmeventd snapshot monitoring).
+  Prevent pvcreate from overwriting MDA-less PVs belonging to active VGs.
+  Fix a segfault if using pvs with --all argument. (2.02.29)
+  Update --uuid argument description in man pages.
+  Fix vgreduce PV list processing not to process every PV in the VG. (2.02.29)
+  Extend lvconvert to use polldaemon.
+  Add support for stacked mirrors.
+  Major restructuring of pvmove and lvconvert layer manipulation code.
+  Replace tools/fsadm with scripts/fsadm.sh.
+  Append fields to report/pvsegs_cols_verbose.
+  Permit LV segment fields with PV segment reports.
+  Add seg_start_pe and seg_pe_ranges to reports.
+
+Version 2.02.29 - 5th December 2007
+===================================
+  Make clvmd backup vg metadata on remote nodes.
+  Refactor pvmove allocation code.
+  Decode cluster locking state in log message.
+  Change file locking state messages from debug to very verbose.
+  Fix --addtag to drop @ prefix from name.
+  Stop clvmd going haywire if a pre_function fails.
+  Convert some vg_reads into vg_lock_and_reads.
+  Avoid nested vg_reads when processing PVs in VGs and fix associated locking.
+  Accept sizes with --readahead argument.
+  Store size arguments as sectors internally.
+  Attempt to remove incomplete LVs with lvcreate zeroing/activation problems.
+  Add read_ahead activation code.
+  Add activation/readahead configuration option and FMT_RESTRICTED_READAHEAD.
+  Extend readahead arg to accept "auto" and "none".
+  Add lv_read_ahead and lv_kernel_read_ahead fields to reports and lvdisplay.
+  Prevent lvconvert -s from using same LV as origin and snapshot.
+  Fix human-readable output of odd numbers of sectors.
+  Add pv_mda_free and vg_mda_free fields to reports for raw text format.
+  Add LVM2 version to 'Generated by' comment in metadata.
+  Show 'not usable' space when PV is too large for device in pvdisplay.
+  Ignore and fix up any excessive device size found in metadata.
+  Fix error message when fixing up PV size in lvm2 metadata (2.02.11).
+  Fix orphan-related locking in pvdisplay and pvs.
+  Fix missing VG unlocks in some pvchange error paths.
+  Add some missing validation of VG names.
+  Rename validate_vg_name() to validate_new_vg_name().
+  Change orphan lock to VG_ORPHANS.
+  Change format1 to use ORPHAN as orphan VG name.
+  Convert pvchange, pvdisplay, pvscan to use is_orphan()
+  Add is_orphan_vg() and change all hard-coded checks to use it.
+  Detect md superblocks version 1.0, 1.1 and 1.2.
+  Add _alloc_pv() and _free_pv() from _pv_create() code and fix error paths.
+  Add pv_dev_name() to access PV device name.
+  Add const attributes to pv accessor functions.
+  Refactor vg_add_snapshot() and lv_create_empty().
+  Handle new sysfs subsystem/block/devices directory structure.
+  Run test with LVM_SYSTEM_DIR pointing to private root and /dev dirs.
+  Fix a bug in lvm_dump.sh checks for lvm/dmsetup binaries.
+  Fix underquotations in lvm_dump.sh.
+  Refactor lvcreate stripe and mirror parameter validation.
+  Print --help output to stdout, not stderr.
+  After a cmdline processing error, don't print help text but suggest --help.
+  Add %PVS extents option to lvresize, lvextend, and lvcreate.
+  Add 'make check' to run tests in new subdirectory 'test'.
+  Moved the obsolete test subdirectory to old-tests.
+  Cope with relative paths in configure --with-dmdir.
+  Remove no-longer-correct restrictions on PV arg count with stripes/mirrors.
+  Fix strdup memory leak in str_list_dup().
+  Link with -lpthread when static SELinux libraries require that.
+  Detect command line PE values that exceed their 32-bit range.
+  Include strerror string in dev_open_flags' stat failure message.
+  Move guts of pvresize into library.
+  Avoid error when --corelog is provided without --mirrorlog. (2.02.28)
+  Correct --mirrorlog argument name in man pages (not --log).
+  Clear MIRROR_NOTSYNCED LV flag when converting from mirror to linear.
+  Modify lvremove to prompt for removal if LV active on other cluster nodes.
+  Add '-f' to vgremove to force removal of VG even if LVs exist.
+
+Version 2.02.28 - 24th August 2007
+==================================
+  Fix clvmd logging so you can get lvm-level debugging out of it.
+  Introduce VG_GLOBAL lock type for vgscan/pvscan to trigger clvmd -R.
+  Change locking_flags from int to uint32_t.
+  Fix clvmd -R, so it fully refreshes the caches.
+  Change lvconvert_mirrors to use mirror segtype not striped.
+  Fix lvconvert_mirrors detection of number of existing mirrors.
+  Clean up numerous compiler warnings that appeared in recent releases.
+  Remove several unused parameters from _allocate().
+  Only permit --force, --verbose and --debug arguments to be repeated.
+  Fix inconsistent licence notices: executables are GPLv2; libraries LGPLv2.1.
+  Move guts of vgremove and lvremove into library, including yes_no_prompt.
+  Allow clvmd debug to be turned on in a running daemon using clvmd -d [-C].
+  Update to use autoconf 2.61, while still supporting 2.57.
+  Add more cluster info to lvmdump.
+  Add further const attributes throughout.
+  Add support for renaming mirrored LVs.
+  Factor out core of lvrename() to library function.
+  Add --mirrorlog argument to specify log type for mirrors.
+  Don't attempt to monitor devices if their creation failed in _lv_activate.
+  Don't leak a file descriptor in fcntl_lock_file() when fcntl fails.
+  Replace create_dir with dm_create_dir.
+  Detect stream write failure reliably with lvm_fclose using dm_fclose.
+  Fix clvmd if compiled with gulm support. (2.02.26)
+  Fix lvdisplay man page to say LV size is reported in sectors, not KB.
+  Add vg_lock_and_read() external library function.
+  Fix loading of persistent cache if cache_dir is used. (2.02.23)
+  Reduce _compare_paths lstat error message from log_error to log_very_verbose.
+  Create util.h with last_path_component replacing strdup + basename.
+  Use gcc's printf attribute wherever possible.
+  In _line_append, use "sizeof buf - 1" rather than equivalent "4095".
+  Introduce is_same_inode macro, now including a comparison of st_dev.
+  Don't leak a file descriptor in _lock_file() when flock fails.
+  Add SUN's LDOM virtual block device (vdisk) and ps3disk to filters.
+  Split metadata-external.h out from metadata.h for the tools to use.
+
+Version 2.02.27 - 17th July 2007
 ================================
+  Fix snapshot cow area deactivation if origin is not active. (2.02.13)
+  Fix configure libdevmapper.h check when --with-dmdir is used.
+  Turn _add_pv_to_vg() into external library function add_pv_to_vg().
+  Add pv_by_path() external library function.
+  Tidy clvmd-openais of redundant bits, and improve an error report.
+  Cope with find_seg_by_le() failure in check_lv_segments().
+  Call dev_iter_destroy() if _process_all_devs() is interrupted by sigint.
+  Add vg_mda_count and pv_mda_count columns to reports.
+  Fix dumpconfig to use log_print instead of stdout directly.
+  Remove unused parameter 'fid' from _add_pv_to_vg.
+  Add kernel and device-mapper targets versions to lvmdump.
+  Replace BSD (r)index with C89 str(r)chr.
+  Handle vgsplit of an entire VG as a vgrename.
+  Reinitialise internal lvmdiskscan variables when called repeatedly.
+  Fix missing lvm_shell symbol in lvm2cmd library. (2.02.23)
+  Add vg_status function and clean up vg->status in tools directory.
+  Add --ignoremonitoring to disable all dmeventd interaction.
+  Remove get_ prefix from get_pv_* functions.
+  clvmd-openais now uses cpg_local_get() to get nodeid, rather than Clm.
+  Print warnings to stderr instead of stdout.
+
+Version 2.02.26 - 15th June 2007
+================================
+  Update vgcfgrestore man page.
+  Allow keyboard interrupt during user prompts when appropriate.
+  Remove unused clvmd system-lv code.
+  Replace many physical_volume struct dereferences with new get_pv_* functions.
+  Suppress a benign compile-time warning.
+  Convert find_pv_in_vg_by_uuid and pv_create to use PV handles.
+  Add wrappers to some functions in preparation for external LVM library.
+  Add -f to vgcfgrestore to list metadata backup files.
+  Add vg_check_status to consolidate vg status checks and error messages.
+  Add pvdisplay --maps implementation.
+  Remove unsupported LVM1 options from vgcfgrestore man page.
+  Update vgcfgrestore man page to show mandatory VG name.
+  Update vgrename man page to include UUID and be consistent with lvrename.
+  Add (experimental) OpenAIS support to clvmd.
+  Fix deactivation code to follow dependencies and remove symlinks.
+  Fix and clarify vgsplit error messages.
+  Fix a segfault in device_is_usable() if a device has no table.
+  Add some more debug messages to clvmd startup.
+  Misc clvmd cleanups.
+
+Version 2.02.25 - 27th April 2007
+=================================
+  Fix get_config_uint64() to read a 64-bit value not a 32-bit one.
+  Add -Wformat-security and change one fprintf() to fputs().
+  Move regex functions into libdevmapper.
+  Change some #include lines to search only standard system directories.
+  Add devices/preferred_names config regex list for displayed device names.
+  Free a temporary dir string in fcntl_lock_file() after use.
+  Fix a dm_pool_destroy() in matcher_create().
+  Introduce goto_bad macro.
+  Fix warnings on x86_64 involving ptrdiff_t in log_error messages.
+  Update pvck to include text metadata area and record detection.
+  Add support functions for token counting in config file extracts.
+  Update pvck to read labels on disk, with --labelsector parameter.
+  Add count_chars and count_chars_len functions.
+  Add /sys/block listings to lvm_dump.sh.
+  Make lvm_dump.sh list /dev recursively.
+  Fix thread race in clvmd.
+  Add scan_sector param to label_read and _find_labeller.
+  Make clvmd cope with quorum devices.
+  Add extra internal error checking to clvmd.
+  Add dev_read_circular.
+  Add pvck command stub.
+  Update lists of attribute characters in man pages.
+  Change cling alloc policy attribute character from 'C' to l'.
+  Fix creation and conversion of mirrors with tags.
+  Fix vgsplit for lvm1 format (set and validate VG name in PVs metadata).
+  Split metadata areas in vgsplit properly.
+
+Version 2.02.24 - 19th March 2007
+=================================
+  Fix processing of exit status in init scripts
+  Fix vgremove to require at least one vg argument.
+  Fix reading of striped LVs in LVM1 format.
+  Flag nolocking as clustered so clvmd startup sees clustered LVs. (2.02.10)
+  Add a few missing pieces of vgname command line validation.
+  Support the /dev/mapper prefix on most command lines.
+
+Version 2.02.23 - 8th March 2007
+================================
+  Fix vgrename active LV check to ignore differing vgids.
+  Remove no-longer-used uuid_out parameter from activation info functions.
+  Fix two more segfaults if an empty config file section encountered.
+  Move .cache file into a new /etc/lvm/cache directory by default.
+  Add devices/cache_dir & devices/cache_file_prefix, deprecating devices/cache.
+  Create directory in fcntl_lock_file() if required.
+  Exclude readline support from lvm.static.
+  Fix a leak in a reporting error path (2.02.19).
+
+Version 2.02.22 - 13th February 2007
+====================================
+  Correct -b and -P on a couple of man pages.
+  Add global/units to example.conf.
+  Fix loading of segment_libraries.
+  If a PV reappears after it was removed from its VG, make it an orphan.
+  Don't update metadata automatically if VGIDs don't match.
+  Fix some vgreduce --removemissing command line validation.
+
+Version 2.02.21 - 30th January 2007
+===================================
+  Add warning to lvm2_monitoring_init_rhel4 if attempting to stop monitoring.
+  Fix vgsplit to handle mirrors.
+  Reorder fields in reporting field definitions.
+  Fix vgs to treat args as VGs even when PV fields are displayed.
+  Fix md signature check to handle both endiannesses.
+
+Version 2.02.20 - 25th January 2007
+===================================
+  dmeventd mirror sets ignore_suspended_devices and avoids scanning mirrors.
+  Add devices/ignore_suspended_devices to ignore suspended dm devices.
+  Add some missing close() and fclose() return code checks.
+  Fix exit statuses of reporting tools (2.02.19).
+  Add init script for dmeventd monitoring.
+  lvm.static no longer interacts with dmeventd unless explicitly asked to.
+  Add field definitions to report help text.
+  Remove unnecessary cmd arg from target_*monitor_events().
+  Add private variable to dmeventd shared library interface.
+  Long-lived processes write out persistent dev cache in refresh_toolcontext().
+  Fix refresh_toolcontext() always to wipe persistent device filter cache.
+  Add is_long_lived to toolcontext.
+  Add --clustered to man pages.
+  Streamline dm_report_field_* interface.
+  Change remaining dmeventd terminology 'register' to 'monitor'.
+  Update reporting man pages.
+  No longer necessary to specify alignment type for report fields.
+
+Version 2.02.19 - 17th January 2007
+===================================
+  Fix a segfault if an empty config file section encountered.
+  Move basic reporting functions into libdevmapper.
+  Fix partition table processing after sparc changes (2.02.16).
+  Fix cmdline PE range processing segfault (2.02.13).
+  Some libdevmapper-event interface changes.
+  Report dmeventd mirror monitoring status.
+  Fix dmeventd mirror status line processing.
+
+Version 2.02.18 - 11th January 2007
+===================================
+  Revised libdevmapper-event interface for dmeventd.
+  Remove dmeventd mirror status line word limit.
+  Use CFLAGS when linking so mixed sparc builds can supply -m64.
+  Prevent permission changes on active mirrors.
+  Print warning instead of error message if lvconvert cannot zero volume.
+  Add snapshot options to lvconvert man page.
+  dumpconfig accepts a list of configuration variables to display.
+  Change dumpconfig to use --file to redirect output to a file.
+  Avoid vgreduce error when mirror code removes the log LV.
+  Remove 3 redundant AC_MSG_RESULTs from configure.in.
+  Free memory in _raw_read_mda_header() error paths.
+  Fix ambiguous vgsplit error message for split LV.
+  Fix lvextend man page typo.
+  Add configure --with-dmdir to compile against a device-mapper source tree.
+  Use no flush suspending for mirrors.
+  Add dmeventd_mirror register_mutex, tidy initialisation & add memlock.
+  Fix create mirror with name longer than 22 chars.
+  Fix some activate.c prototypes when compiled without devmapper.
+  Fix dmeventd mirror to cope if monitored device disappears.
+
+Version 2.02.17 - 14th December 2006
+====================================
+  Add missing pvremove error message when device doesn't exist.
+  When lvconvert allocates a mirror log, respect parallel area constraints.
+  Use loop to iterate through the now-ordered policy list in _allocate().
+  Check for failure to allocate just the mirror log.
+  Introduce calc_area_multiple().
+  Support mirror log allocation when there is only one PV: area_count now 0.
+  Fix detection of smallest area in _alloc_parallel_area() for cling policy.
+  Add manpage entry for clvmd -T
+  Fix gulm operation of clvmd, including a hang when doing lvchange -aey
+  Fix hang in clvmd if a pre-command failed.
+
+Version 2.02.16 - 1st December 2006
+===================================
+  Fix VG clustered read locks to use PR not CR.
+  Adjust some alignments for ia64/sparc.
+  Fix mirror segment removal to use temporary error segment.
+  Always compile debug logging into clvmd.
+  Add startup timeout to RHEL4 clvmd startup script.
+  Add -T (startup timeout) switch to clvmd.
+  Improve lvm_dump.sh robustness.
+  Update lvm2create_initrd to support gentoo.
+
+Version 2.02.15 - 21st November 2006
+====================================
+  Fix clvmd_init_rhel4 line truncation (2.02.14).
+  Install lvmdump by default.
+  Fix check for snapshot module when activating snapshot.
+  Fix pvremove error path for case when PV is in use.
+  Warn if certain duplicate config file entries are seen.
+  Enhance lvm_dump.sh for sysreport integration and add man page.
+  Fix --autobackup argument which could never disable backups.
+  Fix a label_verify error path.
+
+Version 2.02.14 - 10th November 2006
+====================================
+  Fix adjusted_mirror_region_size() to handle 64-bit size.
+  Add some missing bounds checks on 32-bit extent counters.
+  Add Petabyte and Exabyte support.
+  Fix lvcreate error message when 0 extents requested.
+  lvremove man page: volumes must be cluster inactive before being removed.
+  Protect .cache manipulations with fcntl locking.
+  Change .cache timestamp comparisons to use ctime.
+  Fix mirror log LV writing to set all bits in whole LV.
+  Fix clustered VG detection and default runlevels in clvmd_init_rhel4.
+  Fix high-level free space check for partial allocations.
+
+Version 2.02.13 - 27th October 2006
+===================================
+  Add couple of missing files to tools/Makefile CLEAN_TARGETS.
+  When adding snapshot leave cow LV mapped device active after zeroing.
+  Fix a clvmd debug message.
+  Add dev_flush() to set_lv().
+  Add lvchange --resync.
+  Perform high-level free space check before each allocation attempt.
+  Don't allow a node to remove an LV that's exclusively active on anther node.
+  Cope if same PV is included more than once in cmdline PE range list.
+  Set PV size to current device size if it is found to be zero.
+  Add segment parameter to target_present functions.
+
+Version 2.02.12 - 16th October 2006
+===================================
+  Fix pvdisplay to use vg_read() for non-orphans.
+  Fall back to internal locking if external locking lib is missing or fails.
+  Retain activation state after changing LV minor number with --force.
+  Propagate clustered flag in vgsplit and require resizeable flag.
+
+Version 2.02.11 - 12th October 2006
+===================================
+  Add clvmd function to return the cluster name. not used by LVM yet.
+  Add cling allocation policy.
+  Change _check_contiguous() to use _for_each_pv().
+  Extend _for_each_pv() to allow termination without error.
+  Abstract _is_contiguous().
+  Remove duplicated pv arg from _check_contiguous().
+  Accept regionsize with lvconvert.
+  Add report columns with underscore before field names ending 'size'.
+  Correct regionsize default on lvcreate man page (MB).
+  Fix clvmd bug that could cause it to die when a node with a long name crashed.
+  Add device size to text metadata.
+  Fix format_text mda_setup pv->size and pv_setup pe_count calculations.
+  Fix _for_each_pv() for mirror with core log.
+  Add lvm_dump.sh script to create a tarball of debugging info from a system.
+  Capture error messages in clvmd and pass them back to the user.
+  Remove unused #defines from filter-md.c.
+  Make clvmd restart init script wait until clvmd has died before starting it.
+  Add -R to clvmd which tells running clvmds to reload their device cache.
+  Add LV column to reports listing kernel modules needed for activation.
+  Show available fields if report given invalid field. (e.g. lvs -o list)
+  Add timestamp functions with --disable-realtime configure option.
+  Add %VG, %LV and %FREE suffices to lvcreate/lvresize --extents arg.
+  Fix two potential NULL pointer derefs in error cases in vg_read().
+  Separate --enable-cluster from locking lib options in lvmconf.sh.
+  Add a missing comma in lvcreate man page.
+
+Version 2.02.10 - 19th September 2006
+=====================================
+  Fix lvconvert mirror change case detection logic.
+  Fix mirror log detachment so it correctly becomes a standalone LV.
+  Extend _check_contiguous() to detect single-area LVs.
+  Include mirror log (untested) in _for_each_pv() processing.
+  Use MIRROR_LOG_SIZE constant.
+  Remove struct seg_pvs from _for_each_pv() to generalise.
+  Avoid adding duplicates to list of parallel PVs to avoid.
+  Fix several incorrect comparisons in parallel area avoidance code.
+  Fix segment lengths when flattening existing parallel areas.
+  Log existing parallel areas prior to allocation.
+  Fix mirror log creation when activation disabled.
+  Don't attempt automatic recovery without proper locking.
+  When using local file locking, skip clustered VGs.
+  Add fallback_to_clustered_locking and fallback_to_local_locking parameters.
+  lvm.static uses built-in cluster locking instead of external locking.
+  Don't attempt to load shared libraries if built statically.
+  Change default locking_lib to liblvm2clusterlock.so.
+  Add skip_dev_dir() to process command line VGs.
+  Stop clvmd complaining about nodes that have left the cluster.
+  Move lvm_snprintf(), split_words() and split_dm_name() into libdevmapper.
+  Add lvconvert man page.
+  Add mirror options to man pages.
+  Prevent mirror renames.
+  Move CMDLIB code into separate file and record whether static build.
+
+Version 2.02.09 - 17th August 2006
+==================================
+  Fix PE_ALIGN for pagesize over 32KB.
+  Separate out LVM1_PE_ALIGN and pe_align().
+  Add lvm_getpagesize wrapper.
+  Add --maxphysicalvolumes to vgchange.
+
+Version 2.02.08 - 15th August 2006
+==================================
   Add checks for duplicate LV name, lvid and PV id before writing metadata.
   Report all sanity check failures, not just the first.
   Fix missing lockfs on first snapshot creation.
@@ -630,7 +1186,7 @@ Some bug fixes & minor enhancements, including:
 
 You need to update libdevmapper before using 'vgmknodes' or 'vgscan --mknodes'.
 If your root filesystem is on an LV, you should run one of those two
-commands to fix up the special files in /dev in your real root filesystem 
+commands to fix up the special files in /dev in your real root filesystem
 after finishing with your initrd.  Also, remember you can use
 'vgchange --ignorelockingfailure' on your initrd if the tool fails because
 it can't write a lock file to a read-only filesystem.
@@ -640,7 +1196,7 @@ Wednesday 30th April 2003
 A pvmove implementation is now available for the new metadata format.
 
 When running a command that allocates space (e.g. lvcreate), you can now
-restrict not only which disk(s) may be used but also the Physical Extents 
+restrict not only which disk(s) may be used but also the Physical Extents
 on those disks.  e.g. lvcreate -L 10 vg1 /dev/hda6:1000-2000:3000-4000
 
 
@@ -652,12 +1208,12 @@ The new format of LVM metadata is ready for you to test!
   It's more compact and supports transactional changes and replication.
   Should things go wrong on a system, it's human-readable (and editable).
 
-Please report any problems you find to the mailing list, 
+Please report any problems you find to the mailing list,
 linux-lvm@sistina.com.  The software has NOT yet been thoroughly
 tested and so quite possibly there'll still be some bugs in it.
 Be aware of the disclaimer in the COPYING file.
 
-While testing, we recommend turning logging on in the configuration file 
+While testing, we recommend turning logging on in the configuration file
 to provide us with diagnostic information:
   log {
         file="/tmp/lvm2.log"
@@ -668,7 +1224,7 @@ to provide us with diagnostic information:
 You should schedule regular backups of your configuration file and
 metadata backups and archives (normally kept under /etc/lvm).
 
-Please read docs/example.conf and "man lvm.conf" to find out more about 
+Please read docs/example.conf and "man lvm.conf" to find out more about
 the configuration file.
 
 To convert an existing volume group called vg1 to the new format using
@@ -697,7 +1253,7 @@ first segment could have 3 stripes while the second segment has just 2.
 LVM2 maintains a backup of the current metadata for each volume group
 in /etc/lvm/backup, and puts copies of previous versions in
 /etc/lvm/archive.  "vgcfgbackup" and "vgcfgrestore" can be used to
-create and restore from these files.  If you fully understand what 
+create and restore from these files.  If you fully understand what
 you're doing, metadata can be changed by editing a copy of a current
 backup file and using vgcfgrestore to reload it.
 
@@ -714,8 +1270,8 @@ What's not finished?
 The internal cache.  If you turn on debugging output you'll see lots of
 repeated messages, many of which will eventually get optimised out.
 
---test sometimes causes a command to fail (e.g. vgconvert --test) even 
-though the real command would work: again, fixing this is waiting for 
+--test sometimes causes a command to fail (e.g. vgconvert --test) even
+though the real command would work: again, fixing this is waiting for
 the work on the cache.
 
 Several of the tools do not yet contain the logic to handle full
@@ -728,4 +1284,3 @@ Display output.  Some metadata information cannot yet be displayed.
 
 Recovery tools to salvage "lost" metadata directly from the disks:
 but we hope the new format will mean such tools are hardly ever needed!
-

WHATS_NEW_DM

@@ -1,5 +1,138 @@
-Version 1.02.09 - 15 Aug 2006
+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.
@@ -202,4 +335,3 @@ 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 Free Software Foundation,
+#   Inc.
 
-timestamp='2003-06-17'
+timestamp='2006-09-20'
 
 # 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.
 #
@@ -70,7 +72,7 @@ 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
+Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005
 Free Software Foundation, Inc.
 
 This is free software; see the source for copying conditions.  There is NO
@@ -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 \
 	| 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 \
 	| 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-* \
 	| 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-* \
@@ -316,30 +340,36 @@ 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-*)
 		;;
@@ -359,6 +389,9 @@ case $basic_machine in
 		basic_machine=a29k-amd
 		os=-udi
 		;;
+    	abacus)
+		basic_machine=abacus-unknown
+		;;
 	adobe68k)
 		basic_machine=m68010-adobe
 		os=-scout
@@ -376,6 +409,9 @@ 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
@@ -435,12 +471,27 @@ case $basic_machine in
 		basic_machine=j90-cray
 		os=-unicos
 		;;
+	craynv)
+		basic_machine=craynv-cray
+		os=-unicosmp
+		;;
+	cr16c)
+		basic_machine=cr16c-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 +514,10 @@ 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
@@ -641,10 +696,6 @@ 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 +708,9 @@ 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 +779,6 @@ case $basic_machine in
 	np1)
 		basic_machine=np1-gould
 		;;
-	nv1)
-		basic_machine=nv1-cray
-		os=-unicosmp
-		;;
 	nsr-tandem)
 		basic_machine=nsr-tandem
 		;;
@@ -736,9 +786,12 @@ 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
@@ -765,6 +818,12 @@ 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 +880,10 @@ 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 +910,10 @@ case $basic_machine in
 	sb1el)
 		basic_machine=mipsisa64sb1el-unknown
 		;;
+	sde)
+		basic_machine=mipsisa32-sde
+		os=-elf
+		;;
 	sei)
 		basic_machine=mips-sei
 		os=-seiux
@@ -960,6 +1027,10 @@ 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 +1074,10 @@ 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 +1108,9 @@ case $basic_machine in
 	romp)
 		basic_machine=romp-ibm
 		;;
+	mmix)
+		basic_machine=mmix-knuth
+		;;
 	rs6000)
 		basic_machine=rs6000-ibm
 		;;
@@ -1049,13 +1127,10 @@ 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 +1203,23 @@ 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*)
 	# Remember, each alternative MUST END IN *, to match a version number.
 		;;
 	-qnx*)
@@ -1158,12 +1237,15 @@ 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 +1258,9 @@ case $os in
 	-opened*)
 		os=-openedition
 		;;
+        -os400*)
+		os=-os400
+		;;
 	-wince*)
 		os=-wince
 		;;
@@ -1197,6 +1282,9 @@ case $os in
 	-atheos*)
 		os=-atheos
 		;;
+	-syllable*)
+		os=-syllable
+		;;
 	-386bsd)
 		os=-bsd
 		;;
@@ -1219,6 +1307,9 @@ case $os in
 	-sinix*)
 		os=-sysv4
 		;;
+        -tpf*)
+		os=-tpf
+		;;
 	-triton*)
 		os=-sysv3
 		;;
@@ -1255,6 +1346,9 @@ case $os in
 	-kaos*)
 		os=-kaos
 		;;
+	-zvmoe)
+		os=-zvmoe
+		;;
 	-none)
 		;;
 	*)
@@ -1277,6 +1371,12 @@ 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 +1386,9 @@ 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
@@ -1332,9 +1432,15 @@ case $basic_machine in
 	*-be)
 		os=-beos
 		;;
+	*-haiku)
+		os=-haiku
+		;;
 	*-ibm)
 		os=-aix
 		;;
+    	*-knuth)
+		os=-mmixware
+		;;
 	*-wec)
 		os=-proelf
 		;;
@@ -1467,9 +1573,15 @@ case $basic_machine in
 			-mvs* | -opened*)
 				vendor=ibm
 				;;
+			-os400*)
+				vendor=ibm
+				;;
 			-ptx*)
 				vendor=sequent
 				;;
+			-tpf*)
+				vendor=ibm
+				;;
 			-vxsim* | -vxworks* | -windiss*)
 				vendor=wrs
 				;;
@@ -1494,7 +1606,7 @@ 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 +1,8 @@
 ##
 ## Copyright (C) 2000-2004 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 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
@@ -13,21 +13,22 @@
 ## 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 "$host_os" in
 	linux*)
@@ -42,6 +43,7 @@ case "$host_os" in
 		DEVMAPPER=yes
 		ODIRECT=yes
 		SELINUX=yes
+		REALTIME=yes
 		CLUSTER=internal
 		FSADM=no ;;
 	darwin*)
@@ -56,6 +58,7 @@ case "$host_os" in
 		DEVMAPPER=yes
 		ODIRECT=no
 		SELINUX=no
+		REALTIME=no
 		CLUSTER=none
 		FSADM=no ;;
 esac
@@ -96,7 +99,7 @@ AC_TYPE_OFF_T
 AC_TYPE_PID_T
 AC_TYPE_SIZE_T
 AC_TYPE_MODE_T
-AC_STRUCT_ST_RDEV
+AC_CHECK_MEMBERS([struct stat.st_rdev])
 AC_STRUCT_TM
 
 ################################################################################
@@ -282,6 +285,13 @@ AC_ARG_ENABLE(selinux, [  --disable-selinux       Disable selinux support],
 SELINUX=$enableval)
 AC_MSG_RESULT($SELINUX)
 
+################################################################################
+dnl -- Disable realtime clock support
+AC_MSG_CHECKING(whether to enable realtime support)
+AC_ARG_ENABLE(realtime, [  --disable-realtime      Disable realtime clock support],
+REALTIME=$enableval)
+AC_MSG_RESULT($REALTIME)
+
 ################################################################################
 dnl -- Build cluster LVM daemon
 AC_MSG_CHECKING(whether to build cluster LVM daemon)
@@ -351,13 +361,9 @@ AC_ARG_ENABLE(cmdlib, [  --enable-cmdlib         Build shared command library],
 CMDLIB=$enableval, CMDLIB=no)
 AC_MSG_RESULT($CMDLIB)
 
-if test x$CMDLIB = xyes; then
-	AC_DEFINE([CMDLIB], 1, [Define to 1 to build the shared command library.])
-fi
-
 ################################################################################
 dnl -- Enable fsadm
-AC_MSG_CHECKING(whether to build fsadm)
+AC_MSG_CHECKING(whether to install fsadm)
 AC_ARG_ENABLE(fsadm, [  --enable-fsadm          Enable fsadm],
 FSADM=$enableval)
 AC_MSG_RESULT($FSADM)
@@ -433,17 +439,13 @@ fi
 ################################################################################
 dnl -- Check for selinux
 if test x$SELINUX = xyes; then
-	AC_MSG_CHECKING(for sepol_check_context function)
 	AC_CHECK_LIB(sepol, sepol_check_context, HAVE_SEPOL=yes, HAVE_SEPOL=no)
-	AC_MSG_RESULT($HAVE_SEPOL)
 
 	if test x$HAVE_SEPOL = xyes; then
 		LIBS="-lsepol $LIBS"
 	fi
 
-	AC_MSG_CHECKING(for is_selinux_enabled function)
 	AC_CHECK_LIB(selinux, is_selinux_enabled, HAVE_SELINUX=yes, HAVE_SELINUX=no)
-	AC_MSG_RESULT($HAVE_SELINUX)
 
 	if test x$HAVE_SELINUX = xyes; then
 		AC_DEFINE([HAVE_SELINUX], 1, [Define to 1 to include support for selinux.])
@@ -451,6 +453,31 @@ if test x$SELINUX = xyes; then
 	else
 		AC_MSG_WARN(Disabling selinux)
 	fi
+
+	# With --enable-static_link and selinux enabled, linking lvm.static
+	# 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
 
 ################################################################################
@@ -506,6 +533,17 @@ AC_ARG_WITH(staticdir,
   	    [ STATICDIR="$withval" ],
  	    [ STATICDIR='${exec_prefix}/sbin' ])
 
+AC_ARG_WITH(dmdir,
+	    [  --with-dmdir=DIR        Build against device-mapper source tree in DIR],
+  	    [ DMDIR="$withval" CPPFLAGS="$CPPFLAGS -I$DMDIR/include"],
+	    [ DMDIR= ])
+
+# Convert a relative dir name to absolute.
+case $DMDIR in
+  /*) ;;
+  *) DMDIR="`pwd`/$DMDIR" ;;
+esac
+
 ################################################################################
 dnl -- Ensure additional headers required
 if test x$READLINE = xyes; then
@@ -520,11 +558,6 @@ if test x$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))
@@ -582,11 +615,13 @@ AC_SUBST(DEBUG)
 AC_SUBST(DEVMAPPER)
 AC_SUBST(HAVE_LIBDL)
 AC_SUBST(HAVE_SELINUX)
+AC_SUBST(HAVE_REALTIME)
 AC_SUBST(CMDLIB)
 AC_SUBST(MSGFMT)
 AC_SUBST(LOCALEDIR)
 AC_SUBST(CONFDIR)
 AC_SUBST(STATICDIR)
+AC_SUBST(DMDIR)
 AC_SUBST(INTL_PACKAGE)
 AC_SUBST(INTL)
 AC_SUBST(CLVMD)
@@ -595,36 +630,35 @@ AC_SUBST(FSADM)
 AC_SUBST(DMEVENTD)
 AC_SUBST(CFLOW_CMD)
 AC_SUBST(CSCOPE_CMD)
+AC_SUBST([LIB_PTHREAD])
 
 ################################################################################
-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
+dmeventd/Makefile
+dmeventd/mirror/Makefile
+dmeventd/snapshot/Makefile
+doc/Makefile
+include/Makefile
+lib/Makefile
+lib/format1/Makefile
+lib/format_pool/Makefile
+lib/locking/Makefile
+lib/mirror/Makefile
+lib/snapshot/Makefile
+test/Makefile
+man/Makefile
+po/Makefile
+scripts/Makefile
+tools/Makefile
+tools/version.h
+])
+AC_OUTPUT
 
 if test x$ODIRECT != xyes; then
   AC_MSG_WARN(Warning: O_DIRECT disabled: low-memory pvmove may lock up)

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

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
@@ -19,7 +19,7 @@ SOURCES = \
 	clvmd-command.c  \
 	clvmd.c          \
 	lvm-functions.c  \
-	system-lv.c
+	refresh_clvmd.c
 
 ifeq ("@CLVMD@", "gulm")
 	GULM = yes
@@ -29,9 +29,16 @@ ifeq ("@CLVMD@", "cman")
 	CMAN = yes
 endif
 
+ifeq ("@CLVMD@", "openais")
+	OPENAIS = yes
+	GULM = no
+	CMAN = no
+endif
+
 ifeq ("@CLVMD@", "all")
 	GULM = yes
 	CMAN = yes
+	OPENAIS = no
 endif
 
 ifeq ("@DEBUG@", "yes")
@@ -50,6 +57,12 @@ ifeq ("$(CMAN)", "yes")
 	DEFS += -DUSE_CMAN
 endif
 
+ifeq ("$(OPENAIS)", "yes")
+	SOURCES += clvmd-openais.c
+	LMLIBS += -lSaLck -lcpg
+	DEFS += -DUSE_OPENAIS
+endif
+
 TARGETS = \
 	clvmd
 
@@ -70,7 +83,8 @@ INSTALL_TARGETS = \
 	install_clvmd
 
 clvmd: $(OBJECTS) $(top_srcdir)/lib/liblvm.a
-	$(CC) -o clvmd $(OBJECTS) $(LDFLAGS) $(LVMLIBS) $(LMLIBS) $(LIBS)
+	$(CC) -o clvmd $(OBJECTS) $(CFLAGS) $(LDFLAGS) \
+		$(LVMLIBS) $(LMLIBS) $(LIBS)
 
 .PHONY: install_clvmd
 

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.
  *
@@ -63,4 +63,9 @@ static const char CLVMD_SOCKNAME[] = "\0clvmd";
 #define CLVMD_CMD_LOCK_LV           50
 #define CLVMD_CMD_LOCK_VG           51
 
+/* Misc functions */
+#define CLVMD_CMD_REFRESH	    40
+#define CLVMD_CMD_GET_CLUSTERNAME   41
+#define CLVMD_CMD_SET_DEBUG	    42
+#define CLVMD_CMD_VG_BACKUP	    43
 #endif

daemons/clvmd/clvmd-cman.c

@@ -36,29 +36,34 @@
 #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 "clvmd.h"
 #include "lvm-functions.h"
 
 #define LOCKSPACE_NAME "clvmd"
 
+struct clvmd_node
+{
+	struct cman_node *node;
+	int clvmd_up;
+};
+
 static int num_nodes;
 static struct cman_node *nodes = NULL;
 static struct cman_node this_node;
 static int count_nodes; /* size of allocated nodes array */
-static 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,
@@ -72,12 +77,15 @@ struct lock_wait {
 
 static int _init_cluster(void)
 {
+	node_updown_hash = dm_hash_create(100);
+
 	/* Open the cluster communication socket */
 	c_handle = cman_init(NULL);
 	if (!c_handle) {
 		syslog(LOG_ERR, "Can't open cluster manager socket: %m");
 		return -1;
 	}
+	DEBUGLOG("Connected to CMAN\n");
 
 	if (cman_start_recv_data(c_handle, data_callback, CLUSTER_PORT_CLVMD)) {
 		syslog(LOG_ERR, "Can't bind cluster socket: %m");
@@ -93,6 +101,8 @@ static int _init_cluster(void)
 	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) {
@@ -100,7 +110,7 @@ static int _init_cluster(void)
 		return -1;
 	}
 	dlm_ls_pthread_init(lockspace);
-
+	DEBUGLOG("DLM initialisation complete\n");
 	return 0;
 }
 
@@ -121,14 +131,15 @@ static int _get_num_nodes()
 
 	/* return number of ACTIVE nodes */
 	for (i=0; i<num_nodes; i++) {
-		if (nodes[i].cn_member)
+		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)
 {
 	int nodeid = 0;
 
@@ -137,7 +148,7 @@ static int _cluster_send_message(void *buf, int msglen, char *csid, const char *
 
 	if (cman_send_data(c_handle, buf, msglen, 0, CLUSTER_PORT_CLVMD, nodeid) <= 0)
 	{
-			log_error(errtext);
+		log_error("%s", errtext);
 	}
 	return msglen;
 }
@@ -152,25 +163,29 @@ static void _get_our_csid(char *csid)
 
 /* Call a callback routine for each node is that known (down means not running a clvmd) */
 static int _cluster_do_node_callback(struct local_client *client,
-			     void (*callback) (struct local_client *, char *,
-					       int))
+				     void (*callback) (struct local_client *,
+						       const char *,
+						       int))
 {
 	int i;
 	int somedown = 0;
 
 	for (i = 0; i < _get_num_nodes(); i++) {
-		callback(client, (char *)&nodes[i].cn_nodeid, node_updown[nodes[i].cn_nodeid]);
-		if (!node_updown[nodes[i].cn_nodeid])
-			somedown = -1;
+		if (nodes[i].cn_member && 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 */
+/* Process OOB messages from the cluster socket */
 static void event_callback(cman_handle_t handle, void *private, int reason, int arg)
 {
-	char namebuf[MAX_CLUSTER_NAME_LEN];
+	char namebuf[MAX_CLUSTER_MEMBER_NAME_LEN];
 
 	switch (reason) {
         case CMAN_REASON_PORTCLOSED:
@@ -178,7 +193,7 @@ static void event_callback(cman_handle_t handle, void *private, int reason, int
 		log_notice("clvmd on node %s has died\n", namebuf);
 		DEBUGLOG("Got port closed message, removing node %s\n", namebuf);
 
-		node_updown[arg] = 0;
+		dm_hash_insert_binary(node_updown_hash, (char *)&arg, sizeof(int), (void *)0);
 		break;
 
 	case CMAN_REASON_STATECHANGE:
@@ -204,8 +219,9 @@ static void event_callback(cman_handle_t handle, void *private, int reason, int
 }
 
 static struct local_client *cman_client;
-static int _cluster_fd_callback(struct local_client *fd, char *buf, int len, char *csid,
-			struct local_client **new_client)
+static int _cluster_fd_callback(struct local_client *fd, char *buf, int len,
+				const char *csid,
+				struct local_client **new_client)
 {
 
 	/* Save this for data_callback */
@@ -227,27 +243,12 @@ static void data_callback(cman_handle_t handle, void *private,
 	process_message(cman_client, buf, len, (char *)&nodeid);
 }
 
-static void _add_up_node(char *csid)
+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);
 }
 
@@ -281,7 +282,12 @@ static void count_clvmds_running(void)
 	int i;
 
 	for (i = 0; i < num_nodes; i++) {
-		node_updown[nodes[i].cn_nodeid] = is_listening(nodes[i].cn_nodeid);
+		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);
 	}
 }
 
@@ -290,6 +296,8 @@ static void get_members()
 {
 	int retnodes;
 	int status;
+	int i;
+	int high_nodeid = 0;
 
 	num_nodes = cman_get_node_count(c_handle);
 	if (num_nodes == -1) {
@@ -297,38 +305,37 @@ static void get_members()
 		return;
 	}
 
-	        /* Not enough room for new nodes list ? */
-	        if (num_nodes > count_nodes && nodes) {
-			free(nodes);
-			nodes = NULL;
-		}
+	/* Not enough room for new nodes list ? */
+	if (num_nodes > count_nodes && nodes) {
+		free(nodes);
+		nodes = NULL;
+	}
 
-		if (nodes == NULL) {
-		        count_nodes = num_nodes + 10; /* Overallocate a little */
+	if (nodes == NULL) {
+		count_nodes = num_nodes + 10; /* Overallocate a little */
 		nodes = malloc(count_nodes * sizeof(struct cman_node));
-			if (!nodes) {
-			        log_error("Unable to allocate nodes array\n");
-				exit(5);
-			}
-		}
-
-	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));
+		if (!nodes) {
+			log_error("Unable to allocate nodes array\n");
+			exit(5);
 		}
 	}
 
+	status = cman_get_nodes(c_handle, count_nodes, &retnodes, nodes);
+	if (status < 0) {
+		log_error("Unable to get node details");
+		exit(6);
+	}
+
+	/* Get the highest nodeid */
+	for (i=0; i<retnodes; i++) {
+		if (nodes[i].cn_nodeid > high_nodeid)
+			high_nodeid = nodes[i].cn_nodeid;
+	}
+}
+
+
 /* Convert a node name to a CSID */
-static int _csid_from_name(char *csid, char *name)
+static int _csid_from_name(char *csid, const char *name)
 {
 	int i;
 
@@ -342,7 +349,7 @@ 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;
 
@@ -374,7 +381,7 @@ 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;
 
@@ -468,6 +475,18 @@ static int _sync_unlock(const char *resource /* UNUSED */, int lockid)
 
 }
 
+static int _get_cluster_name(char *buf, int buflen)
+{
+	cman_cluster_t cluster_info;
+	int status;
+
+	status = cman_get_cluster(c_handle, &cluster_info);
+	if (!status) {
+		strncpy(buf, cluster_info.ci_name, buflen);
+	}
+	return status;
+}
+
 static struct cluster_ops _cluster_cman_ops = {
 	.cluster_init_completed   = _cluster_init_completed,
 	.cluster_send_message     = _cluster_send_message,
@@ -481,6 +500,7 @@ 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

@@ -64,8 +64,9 @@
 #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"
@@ -73,7 +74,9 @@
 #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 */
@@ -93,18 +96,34 @@ 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 +135,30 @@ 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_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:
+		lvm_do_backup(&args[2]);
+		break;
+
 	default:
 		/* Won't get here because command is validated in pre_command */
 		break;
@@ -129,7 +166,7 @@ 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 +216,16 @@ static int lock_vg(struct local_client *client)
 	    dm_hash_remove(lock_hash, lockname);
     }
     else {
-
+	/* 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 +249,15 @@ 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 +267,12 @@ 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:
+		break;
+
 	default:
 		log_error("Unknown command %d received\n", header->cmd);
 		status = EINVAL;
@@ -249,6 +300,7 @@ int do_post_command(struct local_client *client)
 		break;
 
 	case CLVMD_CMD_LOCK_VG:
+	case CLVMD_CMD_VG_BACKUP:
 		/* 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);
 
 };
@@ -69,6 +75,23 @@ struct cluster_ops *init_gulm_cluster(void);
 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
 
 
 #endif

daemons/clvmd/clvmd-gulm.c

@@ -40,9 +40,10 @@
 #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"
@@ -51,7 +52,6 @@
 #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 +86,12 @@ 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 */
@@ -278,7 +278,7 @@ 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 +288,7 @@ 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 +582,7 @@ 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 +598,7 @@ 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 +622,7 @@ 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;
 
@@ -661,7 +661,7 @@ 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;
@@ -730,7 +730,7 @@ 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 +828,7 @@ 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 +865,7 @@ 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 +965,20 @@ 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 +995,7 @@ 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 +1,701 @@
+/******************************************************************************
+*******************************************************************************
+**
+**  Copyright (C) 2007 Red Hat, Inc. All rights reserved.
+**
+*******************************************************************************
+******************************************************************************/
+
+/* This provides the interface between clvmd and OpenAIS as the cluster
+ * and lock manager.
+ *
+ */
+
+#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 "list.h"
+#include "locking.h"
+#include "log.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_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");
+	unlock_all();
+
+	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_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_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;
+
+	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);
+	}
+	return 0;
+}
+
+/* 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) 2004-2008 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
@@ -30,11 +30,11 @@
 #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"
@@ -42,20 +42,118 @@
 
 /* LVM2 headers */
 #include "toolcontext.h"
+#include "lvmcache.h"
 #include "log.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;
 };
 
+#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 " : "",
+		flags & LCK_MIRROR_NOSYNC_MODE	  ? "MIRROR_NOSYNC " : "",
+		flags & LCK_DMEVENTD_MONITOR_MODE ? "DMEVENTD_MONITOR " : "");
+
+	return buf;
+}
+
+char *get_last_lvm_error()
+{
+	return last_error;
+}
+
 /* Return the mode a lock is currently held at (or -1 if not held) */
 static int get_current_lock(char *resource)
 {
@@ -201,12 +299,21 @@ 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)
@@ -228,7 +335,7 @@ 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 */
 	}
 
@@ -252,7 +359,7 @@ 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) {
@@ -293,13 +400,15 @@ 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));
 
+	pthread_mutex_lock(&lvm_lock);
 	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.");
+			pthread_mutex_unlock(&lvm_lock);
 			return EINVAL;
 		}
 	}
@@ -310,8 +419,8 @@ int do_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
 	if (lock_flags & LCK_MIRROR_NOSYNC_MODE)
 		init_mirror_in_sync(1);
 
-	if (!(lock_flags & LCK_DMEVENTD_REGISTER_MODE))
-		init_dmeventd_register(0);
+	if (!(lock_flags & LCK_DMEVENTD_MONITOR_MODE))
+		init_dmeventd_monitor(0);
 
 	switch (command) {
 	case LCK_LV_EXCLUSIVE:
@@ -320,11 +429,15 @@ 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:
@@ -347,11 +460,12 @@ int do_lock_lv(unsigned char command, unsigned char lock_flags, char *resource)
 	if (lock_flags & LCK_MIRROR_NOSYNC_MODE)
 		init_mirror_in_sync(0);
 
-	if (!(lock_flags & LCK_DMEVENTD_REGISTER_MODE))
-		init_dmeventd_register(DEFAULT_DMEVENTD_MONITOR);
+	if (!(lock_flags & LCK_DMEVENTD_MONITOR_MODE))
+		init_dmeventd_monitor(DEFAULT_DMEVENTD_MONITOR);
 
 	/* 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;
@@ -365,8 +479,8 @@ 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;
@@ -378,20 +492,25 @@ 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) {
@@ -406,8 +525,8 @@ 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;
 
@@ -416,6 +535,19 @@ 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");
+
+	ret = refresh_toolcontext(cmd);
+	init_full_scan_done(0);
+	lvmcache_label_scan(cmd, 2);
+
+	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
@@ -426,9 +558,10 @@ 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;
@@ -451,7 +584,19 @@ 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);
 }
 
 /*
@@ -466,7 +611,7 @@ 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)
@@ -501,10 +646,34 @@ 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()
 {
@@ -534,19 +703,44 @@ 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);
+}
+
+/* Backups up the LVM metadata if it's changed */
+void lvm_do_backup(const char *vgname)
+{
+	struct volume_group * vg;
+	int consistent = 0;
+
+	DEBUGLOG("Triggering backup of VG metadata for %s. suspended=%d\n", vgname, suspended);
+
+	vg = vg_read(cmd, vgname, NULL /*vgid*/, &consistent);
+	if (vg) {
+		if (consistent)
+			check_current_backup(vg);
+	}
+	else {
+		log_error("Error backing up metadata, can't find VG for group %s", vgname);
+	}
 }
 
 /* Called to initialise the LVM context of the daemon */
 int init_lvm(int using_gulm)
 {
-	if (!(cmd = create_toolcontext(NULL))) {
+	if (!(cmd = create_toolcontext(NULL, 0, 1))) {
 		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);
+	init_debug(cmd->current_settings.debug);
+	init_verbose(cmd->current_settings.verbose + VERBOSE_BASE_LEVEL);
+	set_activation(cmd->current_settings.activation);
+	archive_enable(cmd, cmd->current_settings.archive);
+	backup_enable(cmd, cmd->current_settings.backup);
+	cmd->cmd_line = (char *)"clvmd";
 
 	/* Check lvm.conf is setup for cluster-LVM */
 	check_config();
@@ -557,5 +751,8 @@ int init_lvm(int using_gulm)
 
 	get_initial_state();
 
+	/* Trap log messages so we can pass them back to the user */
+	init_log_fn(lvm2_log_fn);
+
 	return 1;
 }

daemons/clvmd/lvm-functions.h

@@ -24,12 +24,15 @@ extern int do_lock_lv(unsigned char lock_cmd, unsigned char lock_flags,
 		      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 init_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 +1,364 @@
+/*
+ * 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
+ *
+ */
+
+#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

@@ -34,8 +34,8 @@
 #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"
@@ -92,41 +92,37 @@ 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 +177,7 @@ 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 +311,7 @@ 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 +362,7 @@ 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 +390,7 @@ 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 +461,7 @@ 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 +489,7 @@ 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.
 #
@@ -15,8 +15,7 @@ srcdir = @srcdir@
 top_srcdir = @top_srcdir@
 VPATH = @srcdir@
 
-SOURCES = libdevmapper-event.c \
-	  dmeventd.c
+SOURCES = libdevmapper-event.c
 
 LIB_STATIC = libdevmapper-event.a
 
@@ -26,12 +25,20 @@ else
   LIB_SHARED = libdevmapper-event.so
 endif
 
+TARGETS = dmeventd
+CLEAN_TARGETS = dmeventd.o
+
 include ../make.tmpl
 
+LDFLAGS += -ldl -ldevmapper -lpthread
 CLDFLAGS += -ldl -ldevmapper -lpthread
 
+dmeventd: $(LIB_SHARED) dmeventd.o
+	$(CC) -o $@ dmeventd.o $(CFLAGS) $(LDFLAGS) \
+	-L. -ldevmapper-event $(LIBS) -rdynamic
+
 .PHONY: install_dynamic install_static install_include \
-	install_pkgconfig
+	install_pkgconfig install_dmeventd
 
 INSTALL_TYPE = install_dynamic
 
@@ -43,7 +50,7 @@ ifeq ("@PKGCONFIG@", "yes")
   INSTALL_TYPE += install_pkgconfig
 endif
 
-install: $(INSTALL_TYPE) install_include
+install: $(INSTALL_TYPE) install_include install_dmeventd
 
 install_include:
 	$(INSTALL) -D $(OWNER) $(GROUP) -m 444 libdevmapper-event.h \
@@ -55,6 +62,9 @@ install_dynamic: libdevmapper-event.$(LIB_SUFFIX)
 	$(LN_S) -f libdevmapper-event.$(LIB_SUFFIX).$(LIB_VERSION) \
 		$(libdir)/libdevmapper-event.$(LIB_SUFFIX)
 
+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

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/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

@@ -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

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>
@@ -26,83 +25,109 @@
 #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
 
-static pthread_mutex_t _lock = PTHREAD_MUTEX_INITIALIZER;
+/*
+ * 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;
 
-/* FIXME: We may need to lock around operations to these */
-static int register_count = 0;
-static struct dm_pool *mem_pool = NULL;
+/*
+ * 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)
 {
 	if (!strncmp(format, "WARNING: ", 9) && (level < 5))
 		syslog(LOG_CRIT, "%s", format);
@@ -113,69 +138,56 @@ static void _temporary_log_fn(int level, const char *file,
 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, "vgreduce --config devices{ignore_suspended_devices=1} --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;
+		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;
+	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);
 
-	if (pthread_mutex_trylock(&_lock)) {
+	if (pthread_mutex_trylock(&_event_mutex)) {
 		syslog(LOG_NOTICE, "Another thread is handling an event.  Waiting...");
-		pthread_mutex_lock(&_lock);
+		pthread_mutex_lock(&_event_mutex);
 	}
-	/* 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 (!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;
@@ -192,6 +204,7 @@ 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
@@ -203,52 +216,74 @@ 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(&_lock);
+	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 mirror device, %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)
-{
-	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/.exported_symbols

@@ -0,0 +1,3 @@
+process_event
+register_device
+unregister_device

daemons/dmeventd/plugins/snapshot/Makefile.in

@@ -0,0 +1,36 @@
+#
+# Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
+# Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
+#
+# This file is part of the 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
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+VPATH = @srcdir@
+
+INCLUDES += -I${top_srcdir}/tools
+CLDFLAGS += -L${top_srcdir}/tools -ldevmapper -llvm2cmd
+
+SOURCES = dmeventd_snapshot.c
+
+ifeq ("@LIB_SUFFIX@","dylib")
+  LIB_SHARED = libdevmapper-event-lvm2snapshot.dylib
+else
+  LIB_SHARED = libdevmapper-event-lvm2snapshot.so
+endif
+
+include $(top_srcdir)/make.tmpl
+
+install: libdevmapper-event-lvm2snapshot.$(LIB_SUFFIX)
+	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< \
+		$(libdir)/$<.$(LIB_VERSION)
+	$(LN_S) -f $<.$(LIB_VERSION) $(libdir)/$<
+

daemons/dmeventd/plugins/snapshot/dmeventd_snapshot.c

@@ -0,0 +1,232 @@
+/*
+ * 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);
+}
+
+/* send unregister command to itself */
+static void _unregister_self(struct dm_task *dmt)
+{
+	const char *name = dm_task_get_name(dmt);
+	struct dm_event_handler *dmevh;
+
+	if (!(dmevh = dm_event_handler_create()))
+		return;
+
+	if (dm_event_handler_set_dev_name(dmevh, name))
+		goto fail;
+
+	dm_event_handler_set_event_mask(dmevh, DM_EVENT_ALL_ERRORS|DM_EVENT_TIMEOUT);
+	dm_event_unregister_handler(dmevh);
+fail:
+	dm_event_handler_destroy(dmevh);
+}
+
+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);
+		_unregister_self(dmt);
+		*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;
+}

dmeventd/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
 

dmeventd/mirror/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

dmeventd/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>
@@ -26,83 +25,109 @@
 #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
 
-static pthread_mutex_t _lock = PTHREAD_MUTEX_INITIALIZER;
+/*
+ * 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;
 
-/* FIXME: We may need to lock around operations to these */
-static int register_count = 0;
-static struct dm_pool *mem_pool = NULL;
+/*
+ * 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)
 {
 	if (!strncmp(format, "WARNING: ", 9) && (level < 5))
 		syslog(LOG_CRIT, "%s", format);
@@ -113,69 +138,56 @@ static void _temporary_log_fn(int level, const char *file,
 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, "vgreduce --config devices{ignore_suspended_devices=1} --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;
+		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;
+	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);
 
-	if (pthread_mutex_trylock(&_lock)) {
+	if (pthread_mutex_trylock(&_event_mutex)) {
 		syslog(LOG_NOTICE, "Another thread is handling an event.  Waiting...");
-		pthread_mutex_lock(&_lock);
+		pthread_mutex_lock(&_event_mutex);
 	}
-	/* 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 (!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;
@@ -192,6 +204,7 @@ 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
@@ -203,52 +216,74 @@ 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(&_lock);
+	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 mirror device, %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)
-{
-	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;
+}

dmeventd/snapshot/.exported_symbols

@@ -0,0 +1,3 @@
+process_event
+register_device
+unregister_device

dmeventd/snapshot/Makefile.in

@@ -0,0 +1,36 @@
+#
+# Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
+# Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved.
+#
+# This file is part of the 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
+
+srcdir = @srcdir@
+top_srcdir = @top_srcdir@
+VPATH = @srcdir@
+
+INCLUDES += -I${top_srcdir}/tools
+CLDFLAGS += -L${top_srcdir}/tools -ldevmapper -llvm2cmd
+
+SOURCES = dmeventd_snapshot.c
+
+ifeq ("@LIB_SUFFIX@","dylib")
+  LIB_SHARED = libdevmapper-event-lvm2snapshot.dylib
+else
+  LIB_SHARED = libdevmapper-event-lvm2snapshot.so
+endif
+
+include $(top_srcdir)/make.tmpl
+
+install: libdevmapper-event-lvm2snapshot.$(LIB_SUFFIX)
+	$(INSTALL) -D $(OWNER) $(GROUP) -m 555 $(STRIP) $< \
+		$(libdir)/$<.$(LIB_VERSION)
+	$(LN_S) -f $<.$(LIB_VERSION) $(libdir)/$<
+

dmeventd/snapshot/dmeventd_snapshot.c

@@ -0,0 +1,232 @@
+/*
+ * 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);
+}
+
+/* send unregister command to itself */
+static void _unregister_self(struct dm_task *dmt)
+{
+	const char *name = dm_task_get_name(dmt);
+	struct dm_event_handler *dmevh;
+
+	if (!(dmevh = dm_event_handler_create()))
+		return;
+
+	if (dm_event_handler_set_dev_name(dmevh, name))
+		goto fail;
+
+	dm_event_handler_set_event_mask(dmevh, DM_EVENT_ALL_ERRORS|DM_EVENT_TIMEOUT);
+	dm_event_unregister_handler(dmevh);
+fail:
+	dm_event_handler_destroy(dmevh);
+}
+
+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);
+		_unregister_self(dmt);
+		*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;
+}

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

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
@@ -79,6 +92,12 @@ devices {
     # software RAID (md) devices by looking for md superblocks.
     # 1 enables; 0 disables.
     md_component_detection = 1
+
+    # 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
@@ -182,6 +201,9 @@ 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 +229,26 @@ global {
     # Location of proc filesystem
     proc = "/proc"
 
-    # Type of locking to use. Defaults to file-based locking (1).
+    # Type of locking to use. Defaults to local file-based locking (1).
     # Turn locking off by setting to 0 (dangerous: risks metadata corruption
     # if LVM2 commands get run concurrently).
+    # Type 2 uses the external shared library locking_library.
+    # Type 3 uses built-in clustered locking.
     locking_type = 1
 
+    # If using external locking (type 2) and initialisation fails,
+    # with this set to 1 an attempt will be made to use the built-in
+    # clustered locking.
+    # If you are using a customised locking_library you should set this to 0.
+    fallback_to_clustered_locking = 1
+
+    # If an attempt to initialise type 2 or type 3 locking failed, perhaps
+    # because cluster components such as clvmd are not running, with this set
+    # to 1 an attempt will be made to use local file-based locking (type 1).
+    # If this succeeds, only commands against local volume groups will proceed.
+    # Volume Groups marked as clustered will be ignored.
+    fallback_to_local_locking = 1
+
     # Local non-LV directory that holds file-based locks while commands are
     # in progress.  A directory like /tmp that may get wiped on reboot is OK.
     locking_dir = "/var/lock/lvm"
@@ -223,6 +260,9 @@ global {
 
     # Search this directory first for shared libraries.
     #   library_dir = "/lib"
+
+    # The external locking library to load if locking_type is set to 2.
+    #   locking_library = "liblvm2clusterlock.so"
 }
 
 activation {
@@ -252,6 +292,12 @@ activation {
     # 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.
@@ -339,10 +385,20 @@ activation {
 # 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.
-    #
+    # "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

@@ -30,6 +30,7 @@
 ../lib/log/log.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
@@ -37,12 +38,14 @@
 ../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-string.h
+../lib/misc/lvm-wrappers.h
 ../lib/misc/sharedlib.h
-../lib/regex/matcher.h
 ../lib/report/report.h
 ../lib/uuid/uuid.h
 ../po/pogen.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

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) 2004-2007 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
@@ -78,10 +78,10 @@ SOURCES =\
 	misc/lvm-exec.c \
 	misc/lvm-file.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 \

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) 2004-2007 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
  */
@@ -39,7 +39,7 @@ int lvm1_present(struct cmd_context *cmd)
 {
 	char path[PATH_MAX];
 
-	if (lvm_snprintf(path, sizeof(path), "%s/lvm/global", cmd->proc_dir)
+	if (dm_snprintf(path, sizeof(path), "%s/lvm/global", cmd->proc_dir)
 	    < 0) {
 		log_error("LVM1 proc global snprintf failed");
 		return 0;
@@ -51,6 +51,66 @@ int lvm1_present(struct cmd_context *cmd)
 		return 0;
 }
 
+int list_segment_modules(struct dm_pool *mem, const struct lv_segment *seg,
+			 struct list *modules)
+{
+	unsigned int s;
+	struct lv_segment *seg2, *snap_seg;
+	struct list *snh;
+
+	if (seg->segtype->ops->modules_needed &&
+	    !seg->segtype->ops->modules_needed(mem, seg, modules)) {
+		log_error("module string allocation failed");
+		return 0;
+	}
+
+	if (lv_is_origin(seg->lv))
+		list_iterate(snh, &seg->lv->snapshot_segs)
+			if (!list_lv_modules(mem,
+					     list_struct_base(snh,
+							      struct lv_segment,
+							      origin_list)->cow,
+					     modules))
+				return_0;
+
+	if (lv_is_cow(seg->lv)) {
+		snap_seg = find_cow(seg->lv);
+		if (snap_seg->segtype->ops->modules_needed &&
+		    !snap_seg->segtype->ops->modules_needed(mem, snap_seg,
+							    modules)) {
+			log_error("snap_seg module string allocation failed");
+			return 0;
+		}
+	}
+
+	for (s = 0; s < seg->area_count; s++) {
+		switch (seg_type(seg, s)) {
+		case AREA_LV:
+			seg2 = find_seg_by_le(seg_lv(seg, s), seg_le(seg, s));
+			if (seg2 && !list_segment_modules(mem, seg2, modules))
+				return_0;
+			break;
+		case AREA_PV:
+		case AREA_UNASSIGNED:
+			;
+		}
+	}
+
+	return 1;
+}
+
+int list_lv_modules(struct dm_pool *mem, const struct logical_volume *lv,
+		    struct list *modules)
+{
+	struct lv_segment *seg;
+
+	list_iterate_items(seg, &lv->segments)
+		if (!list_segment_modules(mem, seg, modules))
+			return_0;
+
+	return 1;
+}
+
 #ifndef DEVMAPPER_SUPPORT
 void set_activation(int act)
 {
@@ -81,17 +141,17 @@ int target_version(const char *target_name, uint32_t *maj,
 {
 	return 0;
 }
-int target_present(const char *target_name)
+int target_present(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;
 }
@@ -108,6 +168,10 @@ 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;
@@ -151,7 +215,7 @@ 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,
+int pv_uses_vg(struct physical_volume *pv,
 	       struct volume_group *vg)
 {
 	return 0;
@@ -181,7 +245,7 @@ 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.");
 }
 
@@ -249,7 +313,7 @@ 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;
@@ -257,9 +321,9 @@ 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;
 		}
@@ -327,12 +391,26 @@ int target_version(const char *target_name, uint32_t *maj,
 	return r;
 }
 
+int module_present(const char *target_name)
+{
+	int ret = 0;
+#ifdef MODPROBE_CMD
+	char module[128];
+
+	if (dm_snprintf(module, sizeof(module), "dm-%s", target_name) < 0) {
+		log_error("module_present module name too long: %s",
+			  target_name);
+		return 0;
+	}
+
+	ret = exec_cmd(MODPROBE_CMD, module, "", "");
+#endif
+	return ret;
+}
+
 int target_present(const char *target_name, int use_modprobe)
 {
 	uint32_t maj, min, patchlevel;
-#ifdef MODPROBE_CMD
-	char module[128];
-#endif
 
 	if (!activation())
 		return 0;
@@ -342,14 +420,7 @@ 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(target_name))
 			return_0;
 	}
 #endif
@@ -361,21 +432,24 @@ 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;
 	}
 
@@ -388,25 +462,27 @@ 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);
 }
 
 /*
@@ -439,10 +515,17 @@ int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv,
 	struct dev_manager *dm;
 	struct lvinfo info;
 
+	/* If mirrored LV is temporarily shrinked to 1 area (= linear),
+	 * it should be considered in-sync. */
+	if (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)
@@ -459,11 +542,12 @@ int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv,
 	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;
 	}
@@ -475,7 +559,7 @@ 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;
 	}
@@ -547,7 +631,7 @@ static int _lv_suspend_lv(struct logical_volume *lv, int lockfs)
  * 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;
@@ -557,15 +641,25 @@ int lvs_in_vg_activated(struct volume_group *vg)
 
 	list_iterate_items(lvl, &vg->lvs) {
 		if (lvl->lv->status & VISIBLE_LV)
-			count += (_lv_active(vg->cmd, lvl->lv) == 1);
+			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())
@@ -580,50 +674,157 @@ int lvs_in_vg_opened(struct volume_group *vg)
 }
 
 /*
- * register_dev_for_events
- *
- * This function uses proper error codes (but breaks convention)
- * to return:
- *      -1 on error
- *       0 if the lv's targets don't do event [un]registration
- *       0 if the lv is already [un]registered -- FIXME: not implemented
- *       1 if the lv had a segment which was [un]registered
- *
- * Returns: -1 on error
+ * 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 register_dev_for_events(struct cmd_context *cmd,
-			    struct logical_volume *lv, int do_reg)
+int lv_is_active(struct logical_volume *lv)
+{
+	if (_lv_active(lv->vg->cmd, lv, 0))
+		return 1;
+
+	if (!vg_is_clustered(lv->vg))
+		return 0;
+
+	/*
+	 * FIXME: Cluster does not report per-node LV activation status.
+	 * Currently the best we can do is try exclusive local activation.
+	 * If that succeeds, we know the LV is not active elsewhere in the
+	 * cluster.
+	 */
+	if (activate_lv_excl(lv->vg->cmd, lv)) {
+		deactivate_lv(lv->vg->cmd, lv);
+		return 0;
+	}
+
+	/*
+	 * 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 r = 0;
-	struct list *tmp;
+	int i, pending = 0, monitored;
+	int r = 1;
+	struct list *tmp, *snh, *snht;
 	struct lv_segment *seg;
-	int (*reg) (struct lv_segment *, int events);
+	int (*monitor_fn) (struct lv_segment *s, int e);
+	uint32_t s;
 
-	if (do_reg && !dmeventd_register_mode())
+	/* 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)) {
+		list_iterate_safe(snh, snht, &lv->snapshot_segs)
+			if (!monitor_dev_for_events(cmd, list_struct_base(snh,
+				    struct lv_segment, origin_list)->cow, monitor))
+				r = 0;
+		return r;
+	}
+
 	list_iterate(tmp, &lv->segments) {
 		seg = list_item(tmp, struct lv_segment);
 
-		reg = NULL;
-
-		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 (!reg)
-			continue;
-
-		/* FIXME specify events */
-		if (!reg(seg, 0)) {
-			stack;
-			return -1;
+		/* 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;
+			}
 		}
 
-		r = 1;
+		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;
@@ -654,7 +855,7 @@ static int _lv_suspend(struct cmd_context *cmd, const char *lvid_s,
 		return 1;
 	}
 
-	if (!lv_info(cmd, lv, &info, 0))
+	if (!lv_info(cmd, lv, &info, 0, 0))
 		return_0;
 
 	if (!info.exists || info.suspended)
@@ -668,7 +869,7 @@ static int _lv_suspend(struct cmd_context *cmd, const char *lvid_s,
 		}
 	}
 
-	if (register_dev_for_events(cmd, lv, 0) != 1)
+	if (!monitor_dev_for_events(cmd, lv, 0))
 		/* FIXME Consider aborting here */
 		stack;
 
@@ -714,7 +915,7 @@ static int _lv_resume(struct cmd_context *cmd, const char *lvid_s,
 		return 1;
 	}
 
-	if (!lv_info(cmd, lv, &info, 0))
+	if (!lv_info(cmd, lv, &info, 0, 0))
 		return_0;
 
 	if (!info.exists || !info.suspended)
@@ -726,7 +927,7 @@ static int _lv_resume(struct cmd_context *cmd, const char *lvid_s,
 	memlock_dec();
 	fs_unlock();
 
-	if (register_dev_for_events(cmd, lv, 1) != 1)
+	if (!monitor_dev_for_events(cmd, lv, 1))
 		stack;
 
 	return 1;
@@ -760,7 +961,7 @@ int lv_deactivate(struct cmd_context *cmd, const char *lvid_s)
 		return 1;
 	}
 
-	if (!lv_info(cmd, lv, &info, 1))
+	if (!lv_info(cmd, lv, &info, 1, 0))
 		return_0;
 
 	if (!info.exists)
@@ -772,7 +973,7 @@ int lv_deactivate(struct cmd_context *cmd, const char *lvid_s)
 		return 0;
 	}
 
-	if (register_dev_for_events(cmd, lv, 0) != 1)
+	if (!monitor_dev_for_events(cmd, lv, 0))
 		stack;
 
 	memlock_inc();
@@ -792,7 +993,11 @@ int lv_activation_filter(struct cmd_context *cmd, const char *lvid_s,
 	if (!activation())
 		goto activate;
 
-	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
+	/*
+	 * If we're activating and precommitted metadata is still cached,
+	 * we assume it matches the new live metadata.
+	 */
+	if (!(lv = lv_from_lvid(cmd, lvid_s, 1)))
 		return 0;
 
 	if (!_passes_activation_filter(cmd, lv)) {
@@ -817,7 +1022,11 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
 	if (!activation())
 		return 1;
 
-	if (!(lv = lv_from_lvid(cmd, lvid_s, 0)))
+	/*
+	 * If we're activating and precommitted metadata is still cached,
+	 * we assume it matches the new live metadata.
+	 */
+	if (!(lv = lv_from_lvid(cmd, lvid_s, 1)))
 		return 0;
 
 	if (filter && !_passes_activation_filter(cmd, lv)) {
@@ -831,7 +1040,7 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
 		return 1;
 	}
 
-	if (!lv_info(cmd, lv, &info, 0))
+	if (!lv_info(cmd, lv, &info, 0, 0))
 		return_0;
 
 	if (info.exists && !info.suspended && info.live_table)
@@ -845,7 +1054,7 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
 	memlock_dec();
 	fs_unlock();
 
-	if (!register_dev_for_events(cmd, lv, 1) != 1)
+	if (r && !monitor_dev_for_events(cmd, lv, 1))
 		stack;
 
 	return r;
@@ -874,7 +1083,7 @@ 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)

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);
 
@@ -36,9 +40,14 @@ int driver_version(char *version, size_t size);
 int library_version(char *version, size_t size);
 int lvm1_present(struct cmd_context *cmd);
 
+int module_present(const char *target_name);
 int target_present(const char *target_name, int use_modprobe);
 int target_version(const char *target_name, uint32_t *maj,
                    uint32_t *min, uint32_t *patchlevel);
+int list_segment_modules(struct dm_pool *mem, const struct lv_segment *seg,
+			 struct list *modules);
+int list_lv_modules(struct dm_pool *mem, const struct logical_volume *lv,
+		    struct list *modules);
 
 void activation_release(void);
 void activation_exit(void);
@@ -58,9 +67,9 @@ 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.
@@ -79,10 +88,12 @@ int lv_mirror_percent(struct cmd_context *cmd, struct logical_volume *lv,
  * 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 register_dev_for_events(struct cmd_context *cmd,
+int monitor_dev_for_events(struct cmd_context *cmd,
 			    struct logical_volume *lv, int do_reg);
 
 /*
@@ -91,4 +102,9 @@ int register_dev_for_events(struct cmd_context *cmd,
 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>
@@ -99,10 +100,8 @@ static struct dm_task *_setup_task(const char *name, const char *uuid,
 {
 	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);
@@ -117,20 +116,17 @@ static struct dm_task *_setup_task(const char *name, const char *uuid,
 }
 
 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)
 {
 	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)))
+		return_0;
 
 	if (!with_open_count)
 		if (!dm_task_no_open_count(dmt))
@@ -142,11 +138,64 @@ 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) {
+		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(dmt, MAJOR(dev)) || !dm_task_set_minor(dmt, MINOR(dev)))
+		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:
@@ -155,30 +204,32 @@ 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) &&
 	    	    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) &&
 			 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);
 
 	return 0;
 }
 
 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;
 
@@ -187,8 +238,8 @@ 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 */
@@ -204,10 +255,8 @@ 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)))
+		return_0;
 
 	if (!dm_task_no_open_count(dmt))
 		log_error("Failed to disable open_count");
@@ -293,10 +342,8 @@ 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)))
+		return_0;
 
 	if (!dm_task_no_open_count(dmt))
 		log_error("Failed to disable open_count");
@@ -388,15 +435,11 @@ 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_alloc(mem, sizeof(*dm))))
+		goto_bad;
 
 	dm->cmd = cmd;
 	dm->mem = mem;
@@ -408,10 +451,8 @@ struct dev_manager *dev_manager_create(struct cmd_context *cmd,
 	}
 	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;
 
@@ -458,10 +499,8 @@ 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 ? */
 
@@ -493,10 +532,8 @@ 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;
 }
@@ -523,9 +560,9 @@ 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)
@@ -592,11 +629,11 @@ 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;
+	}
 
 	if (info.exists && !dm_tree_add_dev(dtree, info.major, info.minor)) {
 		log_error("Failed to add device (%" PRIu32 ":%" PRIu32") to dtree",
@@ -640,31 +677,25 @@ 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;
-		}
+		if (!_add_lv_to_dtree(dm, dtree, 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)
 		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;
 }
@@ -782,12 +813,19 @@ static int _add_segment_to_dtree(struct dev_manager *dm,
 {
 	uint32_t s;
 	struct 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, 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;
 	}
 
@@ -798,7 +836,7 @@ 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;
 		}
@@ -841,6 +879,9 @@ 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;
@@ -887,11 +928,25 @@ 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)
+			max_stripe_size = seg->stripe_size;
 	}
 
+	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;
+		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.
  */
@@ -912,10 +967,10 @@ 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;
-        		}
+			if (!dm_split_lvm_name(dm->mem, lvlayer->old_name, &vgname, &lvname, &layer)) {
+				log_error("_create_lv_symlinks: Couldn't split up old device name %s", lvlayer->old_name);
+				return 0;
+			}
 			fs_rename_lv(lvlayer->lv, name, lvname);
 		} else if (!dev_manager_lv_mknodes(lvlayer->lv))
 			r = 0;
@@ -924,6 +979,35 @@ static int _create_lv_symlinks(struct dev_manager *dm, struct dm_tree_node *root
 	return r;
 }
 
+/*
+ * 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;
@@ -938,10 +1022,10 @@ static int _clean_tree(struct dev_manager *dm, struct dm_tree_node *root)
 		if (!(uuid = dm_tree_node_get_uuid(child)))
 			continue;
 
-        	if (!split_dm_name(dm->mem, name, &vgname, &lvname, &layer)) {
-                	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)
@@ -983,9 +1067,13 @@ static int _tree_action(struct dev_manager *dm, struct logical_volume *lv, actio
  		/* 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:
 		dm_tree_skip_lockfs(root);
+		if ((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;
@@ -1065,7 +1153,7 @@ int dev_manager_device_uses_vg(struct device *dev,
 {
 	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())) {

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
  */
@@ -40,7 +40,8 @@ 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,
 				 const struct logical_volume *lv,
 				 float *percent);

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:
@@ -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;
 	}
 
@@ -338,6 +330,11 @@ int fs_del_lv(const struct logical_volume *lv)
 		      "", "");
 }
 
+int fs_del_lv_byname(const char *dev_dir, const char *vg_name, const char *lv_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_name)
 {

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
  */
@@ -25,6 +25,7 @@
  */
 int fs_add_lv(const struct logical_volume *lv, const char *dev);
 int fs_del_lv(const struct logical_volume *lv);
+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_name);
 void fs_unlock(void);

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
  */

lib/cache/lvmcache.c

@@ -1,35 +1,40 @@
 /*
- * 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) 2004-2008 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
- *
  */
 
 #include "lib.h"
 #include "lvmcache.h"
 #include "toolcontext.h"
 #include "dev-cache.h"
+#include "locking.h"
 #include "metadata.h"
 #include "filter.h"
 #include "memlock.h"
 #include "str_list.h"
+#include "format-text.h"
+#include "format_pool.h"
+#include "format1.h"
 
 static struct dm_hash_table *_pvid_hash = NULL;
 static struct dm_hash_table *_vgid_hash = NULL;
 static struct dm_hash_table *_vgname_hash = NULL;
 static struct dm_hash_table *_lock_hash = NULL;
 static struct list _vginfos;
+static int _scanning_in_progress = 0;
 static int _has_scanned = 0;
 static int _vgs_locked = 0;
+static int _vg_global_lock_held = 0;	/* Global lock held when cache wiped? */
 
 int lvmcache_init(void)
 {
@@ -47,9 +52,124 @@ int lvmcache_init(void)
 	if (!(_lock_hash = dm_hash_create(128)))
 		return 0;
 
+	if (_vg_global_lock_held)
+		lvmcache_lock_vgname(VG_GLOBAL, 0);
+
 	return 1;
 }
 
+/* Volume Group metadata cache functions */
+static void _free_cached_vgmetadata(struct lvmcache_vginfo *vginfo)
+{
+	if (!vginfo || !vginfo->vgmetadata)
+		return;
+
+	dm_free(vginfo->vgmetadata);
+
+	vginfo->vgmetadata = NULL;
+
+	log_debug("Metadata cache: VG %s wiped.", vginfo->vgname);
+}
+
+static void _store_metadata(struct lvmcache_vginfo *vginfo,
+			    struct volume_group *vg, unsigned precommitted)
+{
+	int size;
+
+	if (vginfo->vgmetadata)
+		_free_cached_vgmetadata(vginfo);
+
+	if (!(size = export_vg_to_buffer(vg, &vginfo->vgmetadata))) {
+		stack;
+		return;
+	}
+
+	vginfo->precommitted = precommitted;
+
+	log_debug("Metadata cache: VG %s stored (%d bytes%s).", vginfo->vgname,
+		  size, precommitted ? ", precommitted" : "");
+}
+
+static void _update_cache_info_lock_state(struct lvmcache_info *info,
+					  int locked,
+					  int *cached_vgmetadata_valid)
+{
+	int was_locked = (info->status & CACHE_LOCKED) ? 1 : 0;
+
+	/*
+	 * Cache becomes invalid whenever lock state changes
+	 */
+	if (was_locked != locked) {
+		info->status |= CACHE_INVALID;
+		*cached_vgmetadata_valid = 0;
+	}
+
+	if (locked)
+		info->status |= CACHE_LOCKED;
+	else
+		info->status &= ~CACHE_LOCKED;
+}
+
+static void _update_cache_vginfo_lock_state(struct lvmcache_vginfo *vginfo,
+					    int locked)
+{
+	struct lvmcache_info *info;
+	int cached_vgmetadata_valid = 1;
+
+	list_iterate_items(info, &vginfo->infos)
+		_update_cache_info_lock_state(info, locked,
+					      &cached_vgmetadata_valid);
+
+	if (!cached_vgmetadata_valid)
+		_free_cached_vgmetadata(vginfo);
+}
+
+static void _update_cache_lock_state(const char *vgname, int locked)
+{
+	struct lvmcache_vginfo *vginfo;
+
+	if (!(vginfo = vginfo_from_vgname(vgname, NULL)))
+		return;
+
+	_update_cache_vginfo_lock_state(vginfo, locked);
+}
+
+static void _drop_metadata(const char *vgname)
+{
+	struct lvmcache_vginfo *vginfo;
+	struct lvmcache_info *info;
+
+	if (!(vginfo = vginfo_from_vgname(vgname, NULL)))
+		return;
+
+	/*
+	 * Invalidate cached PV labels.
+	 * If cached precommitted metadata exists that means we
+	 * already invalidated the PV labels (before caching it)
+	 * and we must not do it again.
+	 */
+
+	if (!vginfo->precommitted)
+		list_iterate_items(info, &vginfo->infos)
+			info->status |= CACHE_INVALID;
+
+	_free_cached_vgmetadata(vginfo);
+}
+
+void lvmcache_drop_metadata(const char *vgname)
+{
+	/* For VG_ORPHANS, we need to invalidate all labels on orphan PVs. */
+	if (!strcmp(vgname, VG_ORPHANS)) {
+		_drop_metadata(FMT_TEXT_ORPHAN_VG_NAME);
+		_drop_metadata(FMT_LVM1_ORPHAN_VG_NAME);
+		_drop_metadata(FMT_POOL_ORPHAN_VG_NAME);
+
+		/* Indicate that PVs could now be missing from the cache */
+		init_full_scan_done(0);
+	} else
+		_drop_metadata(vgname);
+}
+
 void lvmcache_lock_vgname(const char *vgname, int read_only __attribute((unused)))
 {
 	if (!_lock_hash && !lvmcache_init()) {
@@ -57,10 +177,17 @@ void lvmcache_lock_vgname(const char *vgname, int read_only __attribute((unused)
 		return;
 	}
 
+	if (dm_hash_lookup(_lock_hash, vgname))
+		log_error("Internal error: Nested locking attempted on VG %s.",
+			  vgname);
+		
 	if (!dm_hash_insert(_lock_hash, vgname, (void *) 1))
 		log_error("Cache locking failure for %s", vgname);
 
-	_vgs_locked++;
+	_update_cache_lock_state(vgname, 1);
+
+	if (strcmp(vgname, VG_GLOBAL))
+		_vgs_locked++;
 }
 
 int vgname_is_locked(const char *vgname)
@@ -73,11 +200,16 @@ int vgname_is_locked(const char *vgname)
 
 void lvmcache_unlock_vgname(const char *vgname)
 {
-	/* FIXME: Clear all CACHE_LOCKED flags in this vg */
+	if (!dm_hash_lookup(_lock_hash, vgname))
+		log_error("Internal error: Attempt to unlock unlocked VG %s.",
+			  vgname);
+
+	_update_cache_lock_state(vgname, 0);
+
 	dm_hash_remove(_lock_hash, vgname);
 
 	/* FIXME Do this per-VG */
-	if (!--_vgs_locked)
+	if (strcmp(vgname, VG_GLOBAL) && !--_vgs_locked)
 		dev_close_all();
 }
 
@@ -86,11 +218,34 @@ int vgs_locked(void)
 	return _vgs_locked;
 }
 
+static void _vginfo_attach_info(struct lvmcache_vginfo *vginfo,
+				struct lvmcache_info *info)
+{
+	if (!vginfo)
+		return;
+
+	info->vginfo = vginfo;
+	list_add(&vginfo->infos, &info->list);
+}
+
+static void _vginfo_detach_info(struct lvmcache_info *info)
+{
+	if (!list_empty(&info->list)) {
+		list_del(&info->list);
+		list_init(&info->list);
+	}
+
+	info->vginfo = NULL;
+}
+
 /* If vgid supplied, require a match. */
 struct lvmcache_vginfo *vginfo_from_vgname(const char *vgname, const char *vgid)
 {
 	struct lvmcache_vginfo *vginfo;
 
+	if (!vgname)
+		return vginfo_from_vgid(vgid);
+
 	if (!_vgname_hash)
 		return NULL;
 
@@ -98,7 +253,7 @@ struct lvmcache_vginfo *vginfo_from_vgname(const char *vgname, const char *vgid)
 		return NULL;
 
 	if (vgid)
-		do 
+		do
 			if (!strncmp(vgid, vginfo->vgid, ID_LEN))
 				return vginfo;
 		while ((vginfo = vginfo->next));
@@ -114,7 +269,7 @@ const struct format_type *fmt_from_vgname(const char *vgname, const char *vgid)
 	struct list *devh, *tmp;
 	struct list devs;
 	struct device_list *devl;
-	char vgid_found[ID_LEN + 1];
+	char vgid_found[ID_LEN + 1] __attribute((aligned(8)));
 
 	if (!(vginfo = vginfo_from_vgname(vgname, vgid)))
 		return NULL;
@@ -135,7 +290,7 @@ const struct format_type *fmt_from_vgname(const char *vgname, const char *vgid)
 
 	list_iterate_safe(devh, tmp, &devs) {
 		devl = list_item(devh, struct device_list);
-		label_read(devl->dev, &label);
+		label_read(devl->dev, &label, UINT64_C(0));
 		list_del(&devl->list);
 		dm_free(devl);
 	}
@@ -151,7 +306,7 @@ const struct format_type *fmt_from_vgname(const char *vgname, const char *vgid)
 struct lvmcache_vginfo *vginfo_from_vgid(const char *vgid)
 {
 	struct lvmcache_vginfo *vginfo;
-	char id[ID_LEN + 1];
+	char id[ID_LEN + 1] __attribute((aligned(8)));
 
 	if (!_vgid_hash || !vgid)
 		return NULL;
@@ -171,9 +326,6 @@ const char *vgname_from_vgid(struct dm_pool *mem, const char *vgid)
 	struct lvmcache_vginfo *vginfo;
 	const char *vgname = NULL;
 
-	if (!*vgid)
-		vgname = ORPHAN;
-
 	if ((vginfo = vginfo_from_vgid(vgid)))
 		vgname = vginfo->vgname;
 
@@ -183,10 +335,59 @@ const char *vgname_from_vgid(struct dm_pool *mem, const char *vgid)
 	return vgname;
 }
 
-struct lvmcache_info *info_from_pvid(const char *pvid)
+static int _info_is_valid(struct lvmcache_info *info)
+{
+	if (info->status & CACHE_INVALID)
+		return 0;
+
+	/*
+	 * The caller must hold the VG lock to manipulate metadata.
+	 * In a cluster, remote nodes sometimes read metadata in the
+	 * knowledge that the controlling node is holding the lock.
+	 * So if the VG appears to be unlocked here, it should be safe
+	 * to use the cached value.
+	 */
+	if (info->vginfo && !vgname_is_locked(info->vginfo->vgname))
+		return 1;
+
+	if (!(info->status & CACHE_LOCKED))
+		return 0;
+
+	return 1;
+}
+
+static int _vginfo_is_valid(struct lvmcache_vginfo *vginfo)
 {
 	struct lvmcache_info *info;
-	char id[ID_LEN + 1];
+
+	/* Invalid if any info is invalid */
+	list_iterate_items(info, &vginfo->infos)
+		if (!_info_is_valid(info))
+			return 0;
+
+	return 1;
+}
+
+/* vginfo is invalid if it does not contain at least one valid info */
+static int _vginfo_is_invalid(struct lvmcache_vginfo *vginfo)
+{
+	struct lvmcache_info *info;
+
+	list_iterate_items(info, &vginfo->infos)
+		if (_info_is_valid(info))
+			return 0;
+
+	return 1;
+}
+
+/*
+ * If valid_only is set, data will only be returned if the cached data is
+ * known still to be valid.
+ */
+struct lvmcache_info *info_from_pvid(const char *pvid, int valid_only)
+{
+	struct lvmcache_info *info;
+	char id[ID_LEN + 1] __attribute((aligned(8)));
 
 	if (!_pvid_hash || !pvid)
 		return NULL;
@@ -197,6 +398,9 @@ struct lvmcache_info *info_from_pvid(const char *pvid)
 	if (!(info = dm_hash_lookup(_pvid_hash, id)))
 		return NULL;
 
+	if (valid_only && !_info_is_valid(info))
+		return NULL;
+
 	return info;
 }
 
@@ -205,7 +409,7 @@ static void _rescan_entry(struct lvmcache_info *info)
 	struct label *label;
 
 	if (info->status & CACHE_INVALID)
-		label_read(info->dev, &label);
+		label_read(info->dev, &label, UINT64_C(0));
 }
 
 static int _scan_invalid(void)
@@ -222,7 +426,6 @@ int lvmcache_label_scan(struct cmd_context *cmd, int full_scan)
 	struct device *dev;
 	struct format_type *fmt;
 
-	static int _scanning_in_progress = 0;
 	int r = 0;
 
 	/* Avoid recursion when a PVID can't be found! */
@@ -247,7 +450,7 @@ int lvmcache_label_scan(struct cmd_context *cmd, int full_scan)
 	}
 
 	while ((dev = dev_iter_get(iter)))
-		label_read(dev, &label);
+		label_read(dev, &label, UINT64_C(0));
 
 	dev_iter_destroy(iter);
 
@@ -267,6 +470,51 @@ int lvmcache_label_scan(struct cmd_context *cmd, int full_scan)
 	return r;
 }
 
+struct volume_group *lvmcache_get_vg(const char *vgid, unsigned precommitted)
+{
+	struct lvmcache_vginfo *vginfo;
+	struct volume_group *vg;
+	struct format_instance *fid;
+
+	if (!vgid || !(vginfo = vginfo_from_vgid(vgid)) || !vginfo->vgmetadata)
+		return NULL;
+
+	if (!_vginfo_is_valid(vginfo))
+		return NULL;
+
+	/*
+	 * Don't return cached data if either:
+	 * (i)  precommitted metadata is requested but we don't have it cached
+	 *      - caller should read it off disk;
+	 * (ii) live metadata is requested but we have precommitted metadata cached
+	 *      and no devices are suspended so caller may read it off disk.
+	 *
+	 * If live metadata is requested but we have precommitted metadata cached
+	 * and devices are suspended, we assume this precommitted metadata has
+	 * already been preloaded and committed so it's OK to return it as live.
+	 * Note that we do not clear the PRECOMMITTED flag.
+	 */
+	if ((precommitted && !vginfo->precommitted) ||
+	    (!precommitted && vginfo->precommitted && !memlock()))
+		return NULL;
+
+	if (!(fid =  vginfo->fmt->ops->create_instance(vginfo->fmt,
+						       vginfo->vgname,
+						       vgid, NULL)))
+		return_NULL;
+
+	if (!(vg = import_vg_from_buffer(vginfo->vgmetadata, fid)) ||
+	    !vg_validate(vg)) {
+		_free_cached_vgmetadata(vginfo);
+		return_NULL;
+	}
+
+	log_debug("Using cached %smetadata for VG %s.",
+		  vginfo->precommitted ? "pre-committed" : "", vginfo->vgname);
+
+	return vg;
+}
+
 struct list *lvmcache_get_vgids(struct cmd_context *cmd, int full_scan)
 {
 	struct list *vgids;
@@ -280,7 +528,7 @@ struct list *lvmcache_get_vgids(struct cmd_context *cmd, int full_scan)
 	}
 
 	list_iterate_items(vginfo, &_vginfos) {
-		if (!str_list_add(cmd->mem, vgids, 
+		if (!str_list_add(cmd->mem, vgids,
 				  dm_pool_strdup(cmd->mem, vginfo->vgid))) {
 			log_error("strlist allocation failed");
 			return NULL;
@@ -303,7 +551,7 @@ struct list *lvmcache_get_vgnames(struct cmd_context *cmd, int full_scan)
 	}
 
 	list_iterate_items(vginfo, &_vginfos) {
-		if (!str_list_add(cmd->mem, vgnames, 
+		if (!str_list_add(cmd->mem, vgnames,
 				  dm_pool_strdup(cmd->mem, vginfo->vgname))) {
 			log_error("strlist allocation failed");
 			return NULL;
@@ -329,7 +577,7 @@ struct list *lvmcache_get_pvids(struct cmd_context *cmd, const char *vgname,
 		return pvids;
 
 	list_iterate_items(info, &vginfo->infos) {
-		if (!str_list_add(cmd->mem, pvids, 
+		if (!str_list_add(cmd->mem, pvids,
 				  dm_pool_strdup(cmd->mem, info->dev->pvid))) {
 			log_error("strlist allocation failed");
 			return NULL;
@@ -345,8 +593,8 @@ struct device *device_from_pvid(struct cmd_context *cmd, struct id *pvid)
 	struct lvmcache_info *info;
 
 	/* Already cached ? */
-	if ((info = info_from_pvid((char *) pvid))) {
-		if (label_read(info->dev, &label)) {
+	if ((info = info_from_pvid((char *) pvid, 0))) {
+		if (label_read(info->dev, &label, UINT64_C(0))) {
 			info = (struct lvmcache_info *) label->info;
 			if (id_equal(pvid, (struct id *) &info->dev->pvid))
 				return info->dev;
@@ -356,8 +604,8 @@ struct device *device_from_pvid(struct cmd_context *cmd, struct id *pvid)
 	lvmcache_label_scan(cmd, 0);
 
 	/* Try again */
-	if ((info = info_from_pvid((char *) pvid))) {
-		if (label_read(info->dev, &label)) {
+	if ((info = info_from_pvid((char *) pvid, 0))) {
+		if (label_read(info->dev, &label, UINT64_C(0))) {
 			info = (struct lvmcache_info *) label->info;
 			if (id_equal(pvid, (struct id *) &info->dev->pvid))
 				return info->dev;
@@ -370,8 +618,8 @@ struct device *device_from_pvid(struct cmd_context *cmd, struct id *pvid)
 	lvmcache_label_scan(cmd, 2);
 
 	/* Try again */
-	if ((info = info_from_pvid((char *) pvid))) {
-		if (label_read(info->dev, &label)) {
+	if ((info = info_from_pvid((char *) pvid, 0))) {
+		if (label_read(info->dev, &label, UINT64_C(0))) {
 			info = (struct lvmcache_info *) label->info;
 			if (id_equal(pvid, (struct id *) &info->dev->pvid))
 				return info->dev;
@@ -381,34 +629,54 @@ struct device *device_from_pvid(struct cmd_context *cmd, struct id *pvid)
 	return NULL;
 }
 
-static int _drop_vginfo(struct lvmcache_info *info)
+static int _free_vginfo(struct lvmcache_vginfo *vginfo)
 {
-	if (!list_empty(&info->list)) {
-		list_del(&info->list);
-		list_init(&info->list);
+	int r = 1;
+
+	_free_cached_vgmetadata(vginfo);
+
+	if (vginfo_from_vgname(vginfo->vgname, NULL) == vginfo) {
+		dm_hash_remove(_vgname_hash, vginfo->vgname);
+		if (vginfo->next && !dm_hash_insert(_vgname_hash, vginfo->vgname,
+						    vginfo->next)) {
+			log_error("_vgname_hash re-insertion for %s failed",
+				  vginfo->vgname);
+			r = 0;
+		}
 	}
 
-	if (info->vginfo && list_empty(&info->vginfo->infos)) {
-		dm_hash_remove(_vgname_hash, info->vginfo->vgname);
-		if (info->vginfo->next) {
-                	if (!dm_hash_insert(_vgname_hash, info->vginfo->vgname, info->vginfo->next)) {
-                        	log_error("vg hash re-insertion failed: %s",
-					  info->vginfo->vgname);
-                        	return 0;
-			}
-                }
+	if (vginfo->vgname)
+		dm_free(vginfo->vgname);
 
-		if (info->vginfo->vgname)
-			dm_free(info->vginfo->vgname);
-		if (info->vginfo->creation_host)
-			dm_free(info->vginfo->creation_host);
-		if (*info->vginfo->vgid)
-			dm_hash_remove(_vgid_hash, info->vginfo->vgid);
-		list_del(&info->vginfo->list);
-		dm_free(info->vginfo);
-	}
+	if (vginfo->creation_host)
+		dm_free(vginfo->creation_host);
 
-	info->vginfo = NULL;
+	if (*vginfo->vgid && _vgid_hash &&
+	    vginfo_from_vgid(vginfo->vgid) == vginfo)
+		dm_hash_remove(_vgid_hash, vginfo->vgid);
+
+	list_del(&vginfo->list);
+
+	dm_free(vginfo);
+
+	return r;
+}
+
+/*
+ * vginfo must be info->vginfo unless info is NULL
+ */
+static int _drop_vginfo(struct lvmcache_info *info, struct lvmcache_vginfo *vginfo)
+{
+	if (info)
+		_vginfo_detach_info(info);
+
+	/* vginfo still referenced? */
+	if (!vginfo || is_orphan_vg(vginfo->vgname) ||
+	    !list_empty(&vginfo->infos))
+		return 1;
+
+	if (!_free_vginfo(vginfo))
+		return_0;
 
 	return 1;
 }
@@ -419,10 +687,10 @@ void lvmcache_del(struct lvmcache_info *info)
 	if (info->dev->pvid[0] && _pvid_hash)
 		dm_hash_remove(_pvid_hash, info->dev->pvid);
 
-	_drop_vginfo(info);
+	_drop_vginfo(info, info->vginfo);
 
 	info->label->labeller->ops->destroy_label(info->label->labeller,
-						info->label); 
+						info->label);
 	dm_free(info);
 
 	return;
@@ -444,29 +712,36 @@ static int _lvmcache_update_pvid(struct lvmcache_info *info, const char *pvid)
 	return 1;
 }
 
-static int _lvmcache_update_vgid(struct lvmcache_info *info, const char *vgid)
+/*
+ * vginfo must be info->vginfo unless info is NULL (orphans)
+ */
+static int _lvmcache_update_vgid(struct lvmcache_info *info,
+				 struct lvmcache_vginfo *vginfo,
+				 const char *vgid)
 {
-	if (!vgid || !info->vginfo ||
-	    !strncmp(info->vginfo->vgid, vgid, ID_LEN))
+	if (!vgid || !vginfo ||
+	    !strncmp(vginfo->vgid, vgid, ID_LEN))
 		return 1;
 
-	if (info->vginfo && *info->vginfo->vgid)
-		dm_hash_remove(_vgid_hash, info->vginfo->vgid);
+	if (vginfo && *vginfo->vgid)
+		dm_hash_remove(_vgid_hash, vginfo->vgid);
 	if (!vgid) {
-		log_debug("lvmcache: %s: clearing VGID", dev_name(info->dev));
+		log_debug("lvmcache: %s: clearing VGID", info ? dev_name(info->dev) : vginfo->vgname);
 		return 1;
 	}
 
-	strncpy(info->vginfo->vgid, vgid, ID_LEN);
-	info->vginfo->vgid[ID_LEN] = '\0';
-	if (!dm_hash_insert(_vgid_hash, info->vginfo->vgid, info->vginfo)) {
+	strncpy(vginfo->vgid, vgid, ID_LEN);
+	vginfo->vgid[ID_LEN] = '\0';
+	if (!dm_hash_insert(_vgid_hash, vginfo->vgid, vginfo)) {
 		log_error("_lvmcache_update: vgid hash insertion failed: %s",
-			  info->vginfo->vgid);
+			  vginfo->vgid);
 		return 0;
 	}
 
-	log_debug("lvmcache: %s: setting %s VGID to %s", dev_name(info->dev),
-		  info->vginfo->vgname, info->vginfo->vgid);
+	if (!is_orphan_vg(vginfo->vgname))
+		log_debug("lvmcache: %s: setting %s VGID to %s",
+			  dev_name(info->dev), vginfo->vgname,
+			  vginfo->vgid);
 
 	return 1;
 }
@@ -476,7 +751,8 @@ static int _insert_vginfo(struct lvmcache_vginfo *new_vginfo, const char *vgid,
 			  struct lvmcache_vginfo *primary_vginfo)
 {
 	struct lvmcache_vginfo *last_vginfo = primary_vginfo;
-	char uuid_primary[64], uuid_new[64];
+	char uuid_primary[64] __attribute((aligned(8)));
+	char uuid_new[64] __attribute((aligned(8)));
 	int use_new = 0;
 	
 	/* Pre-existing VG takes precedence. Unexported VG takes precedence. */
@@ -556,25 +832,19 @@ static int _insert_vginfo(struct lvmcache_vginfo *new_vginfo, const char *vgid,
 
 static int _lvmcache_update_vgname(struct lvmcache_info *info,
 				   const char *vgname, const char *vgid,
-				   uint32_t vgstatus, const char *creation_host)
+				   uint32_t vgstatus, const char *creation_host,
+				   const struct format_type *fmt)
 {
-	struct lvmcache_vginfo *vginfo, *primary_vginfo;
+	struct lvmcache_vginfo *vginfo, *primary_vginfo, *orphan_vginfo;
+	struct lvmcache_info *info2, *info3;
 	// struct lvmcache_vginfo  *old_vginfo, *next;
 
-	/* If vgname is NULL and we don't already have a vgname, 
-	 * assume ORPHAN - we want every entry to have a vginfo
-	 * attached for scanning reasons.
-	 */
-	if (!vgname && !info->vginfo) {
-		vgname = ORPHAN;
-		vgid = ORPHAN;
-	}
-
-	if (!vgname || (info->vginfo && !strcmp(info->vginfo->vgname, vgname)))
+	if (!vgname || (info && info->vginfo && !strcmp(info->vginfo->vgname, vgname)))
 		return 1;
 
 	/* Remove existing vginfo entry */
-	_drop_vginfo(info);
+	if (info)
+		_drop_vginfo(info, info->vginfo);
 
 	/* Get existing vginfo or create new one */
 	if (!(vginfo = vginfo_from_vgname(vgname, vgid))) {
@@ -587,9 +857,9 @@ static int _lvmcache_update_vgname(struct lvmcache_info *info,
 				dm_hash_remove(_vgname_hash, old_vginfo->vgname);
 				if (old_vginfo->next) {
 					if (!dm_hash_insert(_vgname_hash, old_vginfo->vgname, old_vginfo->next)) {
-                        			log_error("vg hash re-insertion failed: %s",
+						log_error("vg hash re-insertion failed: %s",
 							  old_vginfo->vgname);
-                        			return 0;
+						return 0;
 					}
 				}
 			} else do {
@@ -606,11 +876,11 @@ static int _lvmcache_update_vgname(struct lvmcache_info *info,
 				return 0;
 			}
 
-			// Rename so can assume new name does not already exist 
+			// Rename so can assume new name does not already exist
 			if (!dm_hash_insert(_vgname_hash, vginfo->vgname, vginfo->next)) {
 				log_error("vg hash re-insertion failed: %s",
 					  vginfo->vgname);
-                      		return 0;
+		      		return 0;
 			}
 		} else {
 ***/
@@ -625,7 +895,24 @@ static int _lvmcache_update_vgname(struct lvmcache_info *info,
 			return 0;
 		}
 		list_init(&vginfo->infos);
-			primary_vginfo = vginfo_from_vgname(vgname, NULL);
+
+		/*
+		 * If we're scanning and there's an invalidated entry, remove it.
+		 * Otherwise we risk bogus warnings of duplicate VGs.
+		 */
+		while ((primary_vginfo = vginfo_from_vgname(vgname, NULL)) &&
+		       _scanning_in_progress && _vginfo_is_invalid(primary_vginfo))
+			list_iterate_items_safe(info2, info3, &primary_vginfo->infos) {
+				orphan_vginfo = vginfo_from_vgname(primary_vginfo->fmt->orphan_vg_name, NULL);
+				_drop_vginfo(info2, primary_vginfo);	
+				_vginfo_attach_info(orphan_vginfo, info2);
+				log_debug("lvmcache: %s: now in VG %s%s%s%s",
+					  dev_name(info2->dev),
+					  vgname, orphan_vginfo->vgid[0] ? " (" : "",
+					  orphan_vginfo->vgid[0] ? orphan_vginfo->vgid : "",
+					  orphan_vginfo->vgid[0] ? ")" : "");
+		}
+
 		if (!_insert_vginfo(vginfo, vgid, vgstatus, creation_host,
 				    primary_vginfo)) {
 			dm_free(vginfo->vgname);
@@ -633,7 +920,7 @@ static int _lvmcache_update_vgname(struct lvmcache_info *info,
 			return 0;
 		}
 		/* Ensure orphans appear last on list_iterate */
-		if (!*vgname)
+		if (is_orphan_vg(vgname))
 			list_add(&_vginfos, &vginfo->list);
 		else
 			list_add_h(&_vginfos, &vginfo->list);
@@ -642,17 +929,24 @@ static int _lvmcache_update_vgname(struct lvmcache_info *info,
 ***/
 	}
 
-	info->vginfo = vginfo;
-	list_add(&vginfo->infos, &info->list);
+	if (info)
+		_vginfo_attach_info(vginfo, info);
+	else if (!_lvmcache_update_vgid(NULL, vginfo, vgid)) /* Orphans */
+		return_0;
+
+	_update_cache_vginfo_lock_state(vginfo, vgname_is_locked(vgname));
 
 	/* FIXME Check consistency of list! */
-	vginfo->fmt = info->fmt;
+	vginfo->fmt = fmt;
 
-	log_debug("lvmcache: %s: now %s%s%s%s%s", dev_name(info->dev),
-		  *vgname ? "in VG " : "orphaned", vgname,
-		  vginfo->vgid[0] ? " (" : "",
-		  vginfo->vgid[0] ? vginfo->vgid : "",
-		  vginfo->vgid[0] ? ")" : "");
+	if (info)
+		log_debug("lvmcache: %s: now in VG %s%s%s%s",
+			  dev_name(info->dev),
+			  vgname, vginfo->vgid[0] ? " (" : "",
+			  vginfo->vgid[0] ? vginfo->vgid : "",
+			  vginfo->vgid[0] ? ")" : "");
+	else
+		log_debug("lvmcache: initialised VG %s", vgname);
 
 	return 1;
 }
@@ -692,37 +986,60 @@ static int _lvmcache_update_vgstatus(struct lvmcache_info *info, uint32_t vgstat
 	return 1;
 }
 
+int lvmcache_add_orphan_vginfo(const char *vgname, struct format_type *fmt)
+{
+	if (!_lock_hash && !lvmcache_init()) {
+		log_error("Internal cache initialisation failed");
+		return 0;
+	}
+
+	return _lvmcache_update_vgname(NULL, vgname, vgname, 0, "", fmt);
+}
+
 int lvmcache_update_vgname_and_id(struct lvmcache_info *info,
 				  const char *vgname, const char *vgid,
 				  uint32_t vgstatus, const char *creation_host)
 {
+	if (!vgname && !info->vginfo) {
+		log_error("Internal error: NULL vgname handed to cache");
+		/* FIXME Remove this */
+		vgname = info->fmt->orphan_vg_name;
+		vgid = vgname;
+	}
+
 	if (!_lvmcache_update_vgname(info, vgname, vgid, vgstatus,
-				     creation_host) ||
-	    !_lvmcache_update_vgid(info, vgid) ||
+				     creation_host, info->fmt) ||
+	    !_lvmcache_update_vgid(info, info->vginfo, vgid) ||
 	    !_lvmcache_update_vgstatus(info, vgstatus, creation_host))
 		return_0;
 
 	return 1;
 }
 
-int lvmcache_update_vg(struct volume_group *vg)
+int lvmcache_update_vg(struct volume_group *vg, unsigned precommitted)
 {
 	struct pv_list *pvl;
 	struct lvmcache_info *info;
-	char pvid_s[ID_LEN + 1];
+	struct lvmcache_vginfo *vginfo;
+	char pvid_s[ID_LEN + 1] __attribute((aligned(8)));
 
 	pvid_s[sizeof(pvid_s) - 1] = '\0';
 
 	list_iterate_items(pvl, &vg->pvs) {
 		strncpy(pvid_s, (char *) &pvl->pv->id, sizeof(pvid_s) - 1);
 		/* FIXME Could pvl->pv->dev->pvid ever be different? */
-		if ((info = info_from_pvid(pvid_s)) &&
+		if ((info = info_from_pvid(pvid_s, 0)) &&
 		    !lvmcache_update_vgname_and_id(info, vg->name,
 						   (char *) &vg->id,
 						   vg->status, NULL))
 			return_0;
 	}
 
+	/* store text representation of vg to cache */
+	if (vg->cmd->current_settings.cache_vgmetadata &&
+	    (vginfo = vginfo_from_vgname(vg->name, NULL)))
+		_store_metadata(vginfo, vg, precommitted);
+
 	return 1;
 }
 
@@ -733,7 +1050,7 @@ struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
 {
 	struct label *label;
 	struct lvmcache_info *existing, *info;
-	char pvid_s[ID_LEN + 1];
+	char pvid_s[ID_LEN + 1] __attribute((aligned(8)));
 
 	if (!_vgname_hash && !lvmcache_init()) {
 		log_error("Internal cache initialisation failed");
@@ -743,12 +1060,10 @@ struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
 	strncpy(pvid_s, pvid, sizeof(pvid_s));
 	pvid_s[sizeof(pvid_s) - 1] = '\0';
 
-	if (!(existing = info_from_pvid(pvid_s)) &&
-	    !(existing = info_from_pvid(dev->pvid))) {
-		if (!(label = label_create(labeller))) {
-			stack;
-			return NULL;
-		}
+	if (!(existing = info_from_pvid(pvid_s, 0)) &&
+	    !(existing = info_from_pvid(dev->pvid, 0))) {
+		if (!(label = label_create(labeller)))
+			return_NULL;
 		if (!(info = dm_malloc(sizeof(*info)))) {
 			log_error("lvmcache_info allocation failed");
 			label_destroy(label);
@@ -804,11 +1119,9 @@ struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
 		/* Has labeller changed? */
 		if (info->label->labeller != labeller) {
 			label_destroy(info->label);
-			if (!(info->label = label_create(labeller))) {
+			if (!(info->label = label_create(labeller)))
 				/* FIXME leaves info without label! */
-				stack;
-				return NULL;
-			}
+				return_NULL;
 			info->label->info = info;
 		}
 		label = info->label;
@@ -840,8 +1153,7 @@ struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
 
 static void _lvmcache_destroy_entry(struct lvmcache_info *info)
 {
-	if (!list_empty(&info->list))
-		list_del(&info->list);
+	_vginfo_detach_info(info);
 	strcpy(info->dev->pvid, "");
 	label_destroy(info->label);
 	dm_free(info);
@@ -853,21 +1165,30 @@ static void _lvmcache_destroy_vgnamelist(struct lvmcache_vginfo *vginfo)
 
 	do {
 		next = vginfo->next;
-		if (vginfo->vgname)
-			dm_free(vginfo->vgname);
-		if (vginfo->creation_host)
-			dm_free(vginfo->creation_host);
-		dm_free(vginfo);
+		if (!_free_vginfo(vginfo))
+			stack;
 	} while ((vginfo = next));
 }
 
-static void _lvmcache_destroy_lockname(int present __attribute((unused)))
+static void _lvmcache_destroy_lockname(struct dm_hash_node *n)
 {
-	/* Nothing to do */
+	char *vgname;
+
+	if (!dm_hash_get_data(_lock_hash, n))
+		return;
+
+	vgname = dm_hash_get_key(_lock_hash, n);
+
+	if (!strcmp(vgname, VG_GLOBAL))
+		_vg_global_lock_held = 1;
+	else
+		log_error("Internal error: Volume Group %s was not unlocked",
+			  dm_hash_get_key(_lock_hash, n));
 }
 
-void lvmcache_destroy(void)
+void lvmcache_destroy(struct cmd_context *cmd, int retain_orphans)
 {
+	struct dm_hash_node *n;
 	log_verbose("Wiping internal VG cache");
 
 	_has_scanned = 0;
@@ -891,10 +1212,16 @@ void lvmcache_destroy(void)
 	}
 
 	if (_lock_hash) {
-		dm_hash_iter(_lock_hash, (dm_hash_iterate_fn) _lvmcache_destroy_lockname);
+		dm_hash_iterate(n, _lock_hash)
+			_lvmcache_destroy_lockname(n);
 		dm_hash_destroy(_lock_hash);
 		_lock_hash = NULL;
 	}
 
+	if (!list_empty(&_vginfos))
+		log_error("Internal error: _vginfos list should be empty");
 	list_init(&_vginfos);
+
+	if (retain_orphans)
+		init_lvmcache_orphans(cmd);
 }

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
@@ -44,6 +44,8 @@ struct lvmcache_vginfo {
 	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 */
@@ -60,7 +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 */
@@ -71,13 +73,14 @@ struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
 				   struct device *dev,
 				   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_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);
+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);
@@ -87,7 +90,7 @@ 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);
@@ -105,4 +108,8 @@ struct list *lvmcache_get_vgids(struct cmd_context *cmd, int full_scan);
 struct 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.c

@@ -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) 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
- *
  */
 
 #include "lib.h"
@@ -59,15 +58,13 @@
 #  include <malloc.h>
 #endif
 
-static FILE *_log;
-
 static int _get_env_vars(struct cmd_context *cmd)
 {
 	const char *e;
 
 	/* Set to "" to avoid using any system directory */
 	if ((e = getenv("LVM_SYSTEM_DIR"))) {
-		if (lvm_snprintf(cmd->sys_dir, sizeof(cmd->sys_dir),
+		if (dm_snprintf(cmd->sys_dir, sizeof(cmd->sys_dir),
 				 "%s", e) < 0) {
 			log_error("LVM_SYSTEM_DIR environment variable "
 				  "is too long.");
@@ -156,6 +153,7 @@ static void _init_logging(struct cmd_context *cmd)
 static int _process_config(struct cmd_context *cmd)
 {
 	mode_t old_umask;
+	const char *read_ahead;
 
 	/* umask */
 	cmd->default_settings.umask = find_config_tree_int(cmd,
@@ -167,7 +165,7 @@ static int _process_config(struct cmd_context *cmd)
 		log_verbose("Set umask to %04o", cmd->default_settings.umask);
 
 	/* dev dir */
-	if (lvm_snprintf(cmd->dev_dir, sizeof(cmd->dev_dir), "%s/",
+	if (dm_snprintf(cmd->dev_dir, sizeof(cmd->dev_dir), "%s/",
 			 find_config_tree_str(cmd, "devices/dir",
 					 DEFAULT_DEV_DIR)) < 0) {
 		log_error("Device directory given in config file too long");
@@ -178,7 +176,7 @@ static int _process_config(struct cmd_context *cmd)
 #endif
 
 	/* proc dir */
-	if (lvm_snprintf(cmd->proc_dir, sizeof(cmd->proc_dir), "%s",
+	if (dm_snprintf(cmd->proc_dir, sizeof(cmd->proc_dir), "%s",
 			 find_config_tree_str(cmd, "global/proc",
 					 DEFAULT_PROC_DIR)) < 0) {
 		log_error("Device directory given in config file too long");
@@ -186,7 +184,7 @@ static int _process_config(struct cmd_context *cmd)
 	}
 
 	if (*cmd->proc_dir && !dir_exists(cmd->proc_dir)) {
-		log_error("Warning: proc dir %s not found - some checks will be bypassed",
+		log_error("WARNING: proc dir %s not found - some checks will be bypassed",
 			  cmd->proc_dir);
 		cmd->proc_dir[0] = '\0';
 	}
@@ -210,6 +208,16 @@ static int _process_config(struct cmd_context *cmd)
 		return 0;
 	}
 
+	read_ahead = find_config_tree_str(cmd, "activation/readahead", DEFAULT_READ_AHEAD);
+	if (!strcasecmp(read_ahead, "auto"))
+		cmd->default_settings.read_ahead = DM_READ_AHEAD_AUTO;
+	else if (!strcasecmp(read_ahead, "none"))
+		cmd->default_settings.read_ahead = DM_READ_AHEAD_NONE;
+	else {
+		log_error("Invalid readahead specification");
+		return 0;
+	}
+
 	return 1;
 }
 
@@ -274,10 +282,8 @@ static int _init_tags(struct cmd_context *cmd, struct config_tree *cft)
 	if (!cmd->hosttags && find_config_int(cft->root, "tags/hosttags",
 					      DEFAULT_HOSTTAGS)) {
 		/* FIXME Strip out invalid chars: only A-Za-z0-9_+.- */
-		if (!_set_tag(cmd, cmd->hostname)) {
-			stack;
-			return 0;
-		}
+		if (!_set_tag(cmd, cmd->hostname))
+			return_0;
 		cmd->hosttags = 1;
 	}
 
@@ -293,17 +299,13 @@ static int _init_tags(struct cmd_context *cmd, struct config_tree *cft)
 		}
 		if (cn->child) {
 			passes = 0;
-			if (!_check_host_filters(cmd, cn->child, &passes)) {
-				stack;
-				return 0;
-			}
+			if (!_check_host_filters(cmd, cn->child, &passes))
+				return_0;
 			if (!passes)
 				continue;
 		}
-		if (!_set_tag(cmd, tag)) {
-			stack;
-			return 0;
-		}
+		if (!_set_tag(cmd, tag))
+			return_0;
 	}
 
 	return 1;
@@ -319,7 +321,7 @@ static int _load_config_file(struct cmd_context *cmd, const char *tag)
 	if (*tag)
 		filler = "_";
 
-	if (lvm_snprintf(config_file, sizeof(config_file), "%s/lvm%s%s.conf",
+	if (dm_snprintf(config_file, sizeof(config_file), "%s/lvm%s%s.conf",
 			 cmd->sys_dir, filler, tag) < 0) {
 		log_error("LVM_SYSTEM_DIR or tag was too long");
 		return 0;
@@ -330,7 +332,7 @@ static int _load_config_file(struct cmd_context *cmd, const char *tag)
 		return 0;
 	}
 
-	if (!(cfl->cft = create_config_tree(config_file))) {
+	if (!(cfl->cft = create_config_tree(config_file, 0))) {
 		log_error("config_tree allocation failed");
 		return 0;
 	}
@@ -370,17 +372,15 @@ static int _init_lvm_conf(struct cmd_context *cmd)
 {
 	/* No config file if LVM_SYSTEM_DIR is empty */
 	if (!*cmd->sys_dir) {
-		if (!(cmd->cft = create_config_tree(NULL))) {
+		if (!(cmd->cft = create_config_tree(NULL, 0))) {
 			log_error("Failed to create config tree");
 			return 0;
 		}
 		return 1;
 	}
 
-	if (!_load_config_file(cmd, "")) {
-		stack;
-		return 0;
-	}
+	if (!_load_config_file(cmd, ""))
+		return_0;
 
 	return 1;
 }
@@ -392,11 +392,8 @@ static int _init_tag_configs(struct cmd_context *cmd)
 
 	/* Tag list may grow while inside this loop */
 	list_iterate_items(sl, &cmd->tags) {
-		if (!_load_config_file(cmd, sl->str)) {
-			stack;
-			return 0;
-		}
-
+		if (!_load_config_file(cmd, sl->str))
+			return_0;
 	}
 
 	return 1;
@@ -408,7 +405,7 @@ static int _merge_config_files(struct cmd_context *cmd)
 
 	/* Replace temporary duplicate copy of lvm.conf */
 	if (cmd->cft->root) {
-		if (!(cmd->cft = create_config_tree(NULL))) {
+		if (!(cmd->cft = create_config_tree(NULL, 0))) {
 			log_error("Failed to create config tree");
 			return 0;
 		}
@@ -416,10 +413,8 @@ static int _merge_config_files(struct cmd_context *cmd)
 
 	list_iterate_items(cfl, &cmd->config_files) {
 		/* Merge all config trees into cmd->cft using merge/tag rules */
-		if (!merge_config_tree(cmd, cmd->cft, cfl->cft)) {
-			stack;
-			return 0;
-		}
+		if (!merge_config_tree(cmd, cmd->cft, cfl->cft))
+			return_0;
 	}
 
 	return 1;
@@ -467,10 +462,8 @@ static int _init_dev_cache(struct cmd_context *cmd)
 	const struct config_node *cn;
 	struct config_value *cv;
 
-	if (!dev_cache_init()) {
-		stack;
-		return 0;
-	}
+	if (!dev_cache_init(cmd))
+		return_0;
 
 	if (!(cn = find_config_tree_node(cmd, "devices/scan"))) {
 		if (!dev_cache_add_dir("/dev")) {
@@ -536,7 +529,7 @@ static struct dev_filter *_init_filter_components(struct cmd_context *cmd)
 
 	/*
 	 * sysfs filter. Only available on 2.6 kernels.  Non-critical.
-	 * Listed first because it's very efficient at eliminating 
+	 * Listed first because it's very efficient at eliminating
 	 * unavailable devices.
 	 */
 	if (find_config_tree_bool(cmd, "devices/sysfs_scan",
@@ -575,9 +568,9 @@ static struct dev_filter *_init_filter_components(struct cmd_context *cmd)
 	    filters[0] : composite_filter_create(nr_filt, filters);
 }
 
-static int _init_filters(struct cmd_context *cmd)
+static int _init_filters(struct cmd_context *cmd, unsigned load_persistent_cache)
 {
-	const char *dev_cache;
+	const char *dev_cache = NULL, *cache_dir, *cache_file_prefix;
 	struct dev_filter *f3, *f4;
 	struct stat st;
 	char cache_file[PATH_MAX];
@@ -587,15 +580,37 @@ static int _init_filters(struct cmd_context *cmd)
 	if (!(f3 = _init_filter_components(cmd)))
 		return 0;
 
-	if (lvm_snprintf(cache_file, sizeof(cache_file),
-			 "%s/.cache", cmd->sys_dir) < 0) {
-		log_error("Persistent cache filename too long ('%s/.cache').",
-			  cmd->sys_dir);
+	init_ignore_suspended_devices(find_config_tree_int(cmd,
+	    "devices/ignore_suspended_devices", DEFAULT_IGNORE_SUSPENDED_DEVICES));
+
+	/*
+	 * If 'cache_dir' or 'cache_file_prefix' is set, ignore 'cache'.
+	 */
+	cache_dir = find_config_tree_str(cmd, "devices/cache_dir", NULL);
+	cache_file_prefix = find_config_tree_str(cmd, "devices/cache_file_prefix", NULL);
+
+	if (cache_dir || cache_file_prefix) {
+		if (dm_snprintf(cache_file, sizeof(cache_file),
+		    "%s%s%s/%s.cache",
+		    cache_dir ? "" : cmd->sys_dir,
+		    cache_dir ? "" : "/",
+		    cache_dir ? : DEFAULT_CACHE_SUBDIR,
+		    cache_file_prefix ? : DEFAULT_CACHE_FILE_PREFIX) < 0) {
+			log_error("Persistent cache filename too long.");
+			return 0;
+		}
+	} else if (!(dev_cache = find_config_tree_str(cmd, "devices/cache", NULL)) &&
+		   (dm_snprintf(cache_file, sizeof(cache_file),
+				"%s/%s/%s.cache",
+				cmd->sys_dir, DEFAULT_CACHE_SUBDIR,
+				DEFAULT_CACHE_FILE_PREFIX) < 0)) {
+		log_error("Persistent cache filename too long.");
 		return 0;
 	}
 
-	dev_cache = find_config_tree_str(cmd, "devices/cache",
-				    cache_file);
+	if (!dev_cache)
+		dev_cache = cache_file;
+
 	if (!(f4 = persistent_filter_create(f3, dev_cache))) {
 		log_error("Failed to create persistent device filter");
 		return 0;
@@ -608,9 +623,14 @@ static int _init_filters(struct cmd_context *cmd)
 	if (!*cmd->sys_dir)
 		cmd->dump_filter = 0;
 
-	if (!stat(dev_cache, &st) &&
-	    (st.st_mtime > config_file_timestamp(cmd->cft)) &&
-	    !persistent_filter_load(f4))
+	/*
+	 * Only load persistent filter device cache on startup if it is newer
+	 * than the config file and this is not a long-lived process.
+	 */
+	if (load_persistent_cache && !cmd->is_long_lived &&
+	    !stat(dev_cache, &st) &&
+	    (st.st_ctime > config_file_timestamp(cmd->cft)) &&
+	    !persistent_filter_load(f4, NULL))
 		log_verbose("Failed to load existing device cache from %s",
 			    dev_cache);
 
@@ -646,8 +666,9 @@ static int _init_formats(struct cmd_context *cmd)
 #endif
 
 #ifdef HAVE_LIBDL
-	/* Load any formats in shared libs */
-	if ((cn = find_config_tree_node(cmd, "global/format_libraries"))) {
+	/* Load any formats in shared libs if not static */
+	if (!cmd->is_static &&
+	    (cn = find_config_tree_node(cmd, "global/format_libraries"))) {
 
 		struct config_value *cv;
 		struct format_type *(*init_format_fn) (struct cmd_context *);
@@ -660,10 +681,8 @@ static int _init_formats(struct cmd_context *cmd)
 				return 0;
 			}
 			if (!(lib = load_shared_library(cmd, cv->v.str,
-							"format", 0))) {
-				stack;
-				return 0;
-			}
+							"format", 0)))
+				return_0;
 
 			if (!(init_format_fn = dlsym(lib, "init_format"))) {
 				log_error("Shared library %s does not contain "
@@ -702,6 +721,17 @@ static int _init_formats(struct cmd_context *cmd)
 	return 0;
 }
 
+int init_lvmcache_orphans(struct cmd_context *cmd)
+{
+	struct format_type *fmt;
+
+	list_iterate_items(fmt, &cmd->formats)
+		if (!lvmcache_add_orphan_vginfo(fmt->orphan_vg_name, fmt))
+			return_0;
+
+	return 1;
+}
+
 static int _init_segtypes(struct cmd_context *cmd)
 {
 	struct segment_type *segtype;
@@ -740,13 +770,13 @@ static int _init_segtypes(struct cmd_context *cmd)
 #endif
 
 #ifdef HAVE_LIBDL
-	/* Load any formats in shared libs */
-	if ((cn = find_config_tree_node(cmd, "global/segment_libraries"))) {
+	/* Load any formats in shared libs unless static */
+	if (!cmd->is_static &&
+	    (cn = find_config_tree_node(cmd, "global/segment_libraries"))) {
 
 		struct config_value *cv;
 		struct segment_type *(*init_segtype_fn) (struct cmd_context *);
 		void *lib;
-		struct list *sgtl, *tmp;
 		struct segment_type *segtype2;
 
 		for (cv = cn->v; cv; cv = cv->next) {
@@ -756,10 +786,8 @@ static int _init_segtypes(struct cmd_context *cmd)
 				return 0;
 			}
 			if (!(lib = load_shared_library(cmd, cv->v.str,
-							"segment type", 0))) {
-				stack;
-				return 0;
-			}
+							"segment type", 0)))
+				return_0;
 
 			if (!(init_segtype_fn = dlsym(lib, "init_segtype"))) {
 				log_error("Shared library %s does not contain "
@@ -773,18 +801,16 @@ static int _init_segtypes(struct cmd_context *cmd)
 			segtype->library = lib;
 			list_add(&cmd->segtypes, &segtype->list);
 
-			list_iterate_safe(sgtl, tmp, &cmd->segtypes) {
-				segtype2 = list_item(sgtl, struct segment_type);
-				if (!strcmp(segtype2->name, segtype->name)) {
-					log_error("Duplicate segment type %s: "
-						  "unloading shared library %s",
-						  segtype->name, cv->v.str);
-					list_del(&segtype->list);
-					segtype->ops->destroy(segtype);
-					dlclose(lib);
-					break;
-				}
-
+			list_iterate_items(segtype2, &cmd->segtypes) {
+				if ((segtype == segtype2) ||
+				     strcmp(segtype2->name, segtype->name))
+					continue;
+				log_error("Duplicate segment type %s: "
+					  "unloading shared library %s",
+					  segtype->name, cv->v.str);
+				list_del(&segtype->list);
+				segtype->ops->destroy(segtype);
+				dlclose(lib);
 			}
 		}
 	}
@@ -839,7 +865,7 @@ static int _init_backup(struct cmd_context *cmd)
 	min = (uint32_t) find_config_tree_int(cmd, "backup/retain_min",
 					 DEFAULT_ARCHIVE_NUMBER);
 
-	if (lvm_snprintf
+	if (dm_snprintf
 	    (default_dir, sizeof(default_dir), "%s/%s", cmd->sys_dir,
 	     DEFAULT_ARCHIVE_SUBDIR) == -1) {
 		log_err("Couldn't create default archive path '%s/%s'.",
@@ -860,7 +886,7 @@ static int _init_backup(struct cmd_context *cmd)
 	    find_config_tree_bool(cmd, "backup/backup",
 			     DEFAULT_BACKUP_ENABLED);
 
-	if (lvm_snprintf
+	if (dm_snprintf
 	    (default_dir, sizeof(default_dir), "%s/%s", cmd->sys_dir,
 	     DEFAULT_BACKUP_SUBDIR) == -1) {
 		log_err("Couldn't create default backup path '%s/%s'.",
@@ -879,7 +905,8 @@ static int _init_backup(struct cmd_context *cmd)
 }
 
 /* Entry point */
-struct cmd_context *create_toolcontext(struct arg *the_args)
+struct cmd_context *create_toolcontext(struct arg *the_args, unsigned is_static,
+				       unsigned is_long_lived)
 {
 	struct cmd_context *cmd;
 
@@ -902,6 +929,8 @@ struct cmd_context *create_toolcontext(struct arg *the_args)
 	}
 	memset(cmd, 0, sizeof(*cmd));
 	cmd->args = the_args;
+	cmd->is_static = is_static;
+	cmd->is_long_lived = is_long_lived;
 	cmd->hosttags = 0;
 	list_init(&cmd->formats);
 	list_init(&cmd->segtypes);
@@ -914,7 +943,7 @@ struct cmd_context *create_toolcontext(struct arg *the_args)
 		goto error;
 
 	/* Create system directory if it doesn't already exist */
-	if (*cmd->sys_dir && !create_dir(cmd->sys_dir)) {
+	if (*cmd->sys_dir && !dm_create_dir(cmd->sys_dir)) {
 		log_error("Failed to create LVM2 system dir for metadata backups, config "
 			  "files and internal cache.");
 		log_error("Set environment variable LVM_SYSTEM_DIR to alternative location "
@@ -950,7 +979,7 @@ struct cmd_context *create_toolcontext(struct arg *the_args)
 	if (!_init_dev_cache(cmd))
 		goto error;
 
-	if (!_init_filters(cmd))
+	if (!_init_filters(cmd, 1))
 		goto error;
 
 	if (!(cmd->mem = dm_pool_create("command", 4 * 1024))) {
@@ -963,12 +992,16 @@ struct cmd_context *create_toolcontext(struct arg *the_args)
 	if (!_init_formats(cmd))
 		goto error;
 
+	if (!init_lvmcache_orphans(cmd))
+		goto error;
+
 	if (!_init_segtypes(cmd))
 		goto error;
 
 	if (!_init_backup(cmd))
 		goto error;
 
+	cmd->default_settings.cache_vgmetadata = 1;
 	cmd->current_settings = cmd->default_settings;
 
 	cmd->config_valid = 1;
@@ -1019,13 +1052,13 @@ int refresh_toolcontext(struct cmd_context *cmd)
 {
 	log_verbose("Reloading config files");
 
-	if (cmd->config_valid) {
-		if (cmd->dump_filter)
-			persistent_filter_dump(cmd->filter);
-	}
+	/*
+	 * Don't update the persistent filter cache as we will
+	 * perform a full rescan.
+	 */
 
 	activation_release();
-	lvmcache_destroy();
+	lvmcache_destroy(cmd, 0);
 	label_exit();
 	_destroy_segtypes(&cmd->segtypes);
 	_destroy_formats(&cmd->formats);
@@ -1061,15 +1094,25 @@ int refresh_toolcontext(struct cmd_context *cmd)
 	if (!_init_dev_cache(cmd))
 		return 0;
 
-	if (!_init_filters(cmd))
+	if (!_init_filters(cmd, 0))
 		return 0;
 
 	if (!_init_formats(cmd))
 		return 0;
 
+	if (!init_lvmcache_orphans(cmd))
+		return 0;
+
 	if (!_init_segtypes(cmd))
 		return 0;
 
+	/*
+	 * If we are a long-lived process, write out the updated persistent
+	 * device cache for the benefit of short-lived processes.
+	 */
+	if (cmd->is_long_lived && cmd->dump_filter)
+		persistent_filter_dump(cmd->filter);
+
 	cmd->config_valid = 1;
 	return 1;
 }
@@ -1081,7 +1124,7 @@ void destroy_toolcontext(struct cmd_context *cmd)
 
 	archive_exit(cmd);
 	backup_exit(cmd);
-	lvmcache_destroy();
+	lvmcache_destroy(cmd, 0);
 	label_exit();
 	_destroy_segtypes(&cmd->segtypes);
 	_destroy_formats(&cmd->formats);
@@ -1097,8 +1140,4 @@ void destroy_toolcontext(struct cmd_context *cmd)
 	activation_exit();
 	fin_log();
 	fin_syslog();
-
-	if (_log)
-		fclose(_log);
-
 }

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) 2004-2007 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
  */
@@ -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;
@@ -64,6 +66,8 @@ struct cmd_context {
 	struct command *command;
 	struct arg *args;
 	char **argv;
+	unsigned is_static;	/* Static binary? */
+	unsigned is_long_lived;	/* Optimises persistent_filter handling */
 
 	struct dev_filter *filter;
 	int dump_filter;	/* Dump filter when exiting? */
@@ -87,9 +91,10 @@ struct cmd_context {
 	char proc_dir[PATH_MAX];
 };
 
-struct cmd_context *create_toolcontext(struct arg *the_args);
+struct cmd_context *create_toolcontext(struct arg *the_args, unsigned is_static, unsigned is_long_lived);
 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,
@@ -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 byte %d (line %d): unexpected token", p->tb - p->fb + 1, 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);
@@ -95,7 +109,7 @@ 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);
@@ -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 +138,12 @@ 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)
@@ -136,14 +157,14 @@ static int _parse_config_file(struct parser *p, struct config_tree *cft)
 	return 1;
 }
 
-struct config_tree *create_config_tree_from_string(struct cmd_context *cmd,
+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)))
+	if (!(cft = create_config_tree(NULL, 0)))
 		return_NULL;
 
 	c = (struct cs *) cft;
@@ -174,12 +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;
+	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 */
@@ -187,7 +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);
@@ -197,22 +216,12 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 		}
 		p->fb = p->fb + mmap_offset;
 	} else {
-		if (!(buf = dm_malloc(size + size2))) {
-			stack;
-			return 0;
-		}
-		if (!dev_read(dev, (uint64_t) offset, size, buf)) {
-			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,
-				      buf + size)) {
-				log_error("Circular read from %s failed",
-					  dev_name(dev));
-				goto out;
-			}
-		}
 		p->fb = buf;
 	}
 
@@ -225,10 +234,8 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 
 	p->fe = p->fb + size + size2;
 
-	if (!_parse_config_file(p, cft)) {
-		stack;
-		goto out;
-	}
+	if (!_parse_config_file(p, cft))
+		goto_out;
 
 	r = 1;
 
@@ -250,7 +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)) {
@@ -272,22 +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;
 }
@@ -331,7 +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:
@@ -339,32 +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;
@@ -377,58 +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;
+			if (!_line_start(outline))
+				return_0;
+			_write_config(n->child, 0, outline, level + 1);
+			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;
 }
 
@@ -439,10 +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;
@@ -455,27 +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;
@@ -486,10 +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;
@@ -502,10 +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;
@@ -544,7 +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;
 
@@ -558,16 +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 byte %d (line %d): expected a value", p->tb - p->fb + 1, 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;
@@ -604,12 +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;
@@ -635,7 +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) &&
@@ -689,8 +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;
 	}
@@ -699,11 +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)) {
@@ -747,10 +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;
@@ -763,6 +858,7 @@ 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 */
@@ -773,22 +869,30 @@ static 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;
 }
 
 static struct config_node *_find_first_config_node(const struct config_node *cn1,
@@ -819,7 +923,7 @@ static const char *_find_config_str(const struct config_node *cn1,
 	const struct config_node *n = _find_first_config_node(cn1, cn2, path);
 
 	/* Empty strings are ignored */
-	if ((n && n->v->type == CFG_STRING) && (*n->v->v.str)) {
+	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;
 	}
@@ -836,25 +940,26 @@ const char *find_config_str(const struct config_node *cn,
 	return _find_config_str(cn, NULL, path, fail);
 }
 
-static int _find_config_int(const struct config_node *cn1,
-			    const struct config_node *cn2,
-			    const char *path, int fail)
+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->type == CFG_INT) {
-		log_very_verbose("Setting %s to %d", path, n->v->v.i);
+	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;
 }
 
 int find_config_int(const struct config_node *cn, const char *path, int fail)
 {
-	return _find_config_int(cn, NULL, path, fail);
+	/* FIXME Add log_error message on overflow */
+	return (int) _find_config_int64(cn, NULL, path, (int64_t) fail);
 }
 
 static float _find_config_float(const struct config_node *cn1,
@@ -863,7 +968,7 @@ static float _find_config_float(const struct config_node *cn1,
 {
 	const struct config_node *n = _find_first_config_node(cn1, cn2, path);
 
-	if (n && n->v->type == CFG_FLOAT) {
+	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;
 	}
@@ -882,25 +987,26 @@ float find_config_float(const struct config_node *cn, const char *path,
 }
 
 struct config_node *find_config_tree_node(struct cmd_context *cmd,
-                                          const char *path)
+					  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)
+				 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)
+			 int fail)
 {
-	return _find_config_int(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path, 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)
+			     float fail)
 {
 	return _find_config_float(cmd->cft_override ? cmd->cft_override->root : NULL, cmd->cft->root, path, fail);
 }
@@ -988,7 +1094,6 @@ 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;
 }
@@ -1118,3 +1223,61 @@ 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 == 0))
+		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;
@@ -53,7 +53,7 @@ struct config_tree_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);
@@ -65,7 +65,8 @@ int read_config_fd(struct config_tree *cft, struct device *dev,
 		   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,
@@ -107,4 +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
@@ -30,10 +32,14 @@
 #define DEFAULT_PROC_DIR "/proc"
 #define DEFAULT_SYSFS_SCAN 1
 #define DEFAULT_MD_COMPONENT_DETECTION 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"
@@ -57,6 +63,8 @@
 #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
@@ -94,19 +102,20 @@
 #define DEFAULT_REP_ALIGNED 1
 #define DEFAULT_REP_BUFFERED 1
 #define DEFAULT_REP_HEADINGS 1
+#define DEFAULT_REP_PREFIXES 0
 #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"

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;
@@ -102,7 +101,7 @@ 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;
 }
@@ -114,7 +113,7 @@ static struct node *_left(struct node *n)
 	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

@@ -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
  */
@@ -65,10 +65,19 @@ void list_del(struct list *elem)
 	elem->p->n = elem->n;
 }
 
+/*
+ * Remove an element from existing list and insert before 'head'.
+ */
+void list_move(struct list *head, struct list *elem)
+{
+        list_del(elem);
+        list_add(head, elem);
+}
+
 /*
  * Is the list empty?
  */
-int list_empty(struct list *head)
+int list_empty(const struct list *head)
 {
 	return head->n == head;
 }
@@ -76,7 +85,7 @@ int list_empty(struct list *head)
 /*
  * Is this the first element of the list?
  */
-int list_start(struct list *head, struct list *elem)
+int list_start(const struct list *head, const struct list *elem)
 {
 	return elem->p == head;
 }
@@ -84,7 +93,7 @@ int list_start(struct list *head, struct list *elem)
 /*
  * Is this the last element of the list?
  */
-int list_end(struct list *head, struct list *elem)
+int list_end(const struct list *head, const struct list *elem)
 {
 	return elem->n == head;
 }
@@ -92,7 +101,7 @@ int list_end(struct list *head, struct list *elem)
 /*
  * Return first element of the list or NULL if empty
  */
-struct list *list_first(struct list *head)
+struct list *list_first(const struct list *head)
 {
 	return (list_empty(head) ? NULL : head->n);
 }
@@ -100,7 +109,7 @@ struct list *list_first(struct list *head)
 /*
  * Return last element of the list or NULL if empty
  */
-struct list *list_last(struct list *head)
+struct list *list_last(const struct list *head)
 {
 	return (list_empty(head) ? NULL : head->p);
 }
@@ -108,7 +117,7 @@ struct list *list_last(struct list *head)
 /*
  * Return the previous element of the list, or NULL if we've reached the start.
  */
-struct list *list_prev(struct list *head, struct list *elem)
+struct list *list_prev(const struct list *head, const struct list *elem)
 {
 	return (list_start(head, elem) ? NULL : elem->p);
 }
@@ -116,7 +125,7 @@ struct list *list_prev(struct list *head, struct list *elem)
 /*
  * Return the next element of the list, or NULL if we've reached the end.
  */
-struct list *list_next(struct list *head, struct list *elem)
+struct list *list_next(const struct list *head, const struct list *elem)
 {
 	return (list_end(head, elem) ? NULL : elem->n);
 }

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
  */
@@ -54,40 +54,45 @@ void list_add_h(struct list *head, struct list *elem);
  */
 void list_del(struct list *elem);
 
+/*
+ * Remove an element from existing list and insert before 'head'.
+ */
+void list_move(struct list *head, struct list *elem);
+
 /*
  * Is the list empty?
  */
-int list_empty(struct list *head);
+int list_empty(const struct list *head);
 
 /*
  * Is this the first element of the list?
  */
-int list_start(struct list *head, struct list *elem);
+int list_start(const struct list *head, const struct list *elem);
 
 /*
  * Is this the last element of the list?
  */
-int list_end(struct list *head, struct list *elem);
+int list_end(const struct list *head, const struct list *elem);
 
 /*
  * Return first element of the list or NULL if empty
  */
-struct list *list_first(struct list *head);
+struct list *list_first(const struct list *head);
 
 /*
  * Return last element of the list or NULL if empty
  */
-struct list *list_last(struct list *head);
+struct list *list_last(const struct list *head);
 
 /*
  * Return the previous element of the list, or NULL if we've reached the start.
  */
-struct list *list_prev(struct list *head, struct list *elem);
+struct list *list_prev(const struct list *head, const struct list *elem);
 
 /*
  * Return the next element of the list, or NULL if we've reached the end.
  */
-struct list *list_next(struct list *head, struct list *elem);
+struct list *list_next(const struct list *head, const struct list *elem);
 
 /*
  * Given the address v of an instance of 'struct list' called 'head' 

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
  */
@@ -23,6 +23,7 @@
 
 /* Define some portable printing types */
 #define PRIsize_t "zu"
+#define PRIptrdiff_t "td"
 
 struct str_list {
 	struct list list;

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
  */
@@ -20,10 +20,8 @@ struct list *str_list_create(struct dm_pool *mem)
 {
 	struct list *sl;
 
-	if (!(sl = dm_pool_alloc(mem, sizeof(struct list)))) {
-		stack;
-		return NULL;
-	}
+	if (!(sl = dm_pool_alloc(mem, sizeof(struct list))))
+		return_NULL;
 
 	list_init(sl);
 
@@ -34,19 +32,15 @@ int str_list_add(struct dm_pool *mem, struct 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);
@@ -66,17 +60,16 @@ int str_list_del(struct list *sll, const char *str)
 	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 list *sllnew,
+		 const struct list *sllold)
 {
 	struct str_list *sl;
 
 	list_init(sllnew);
 
 	list_iterate_items(sl, sllold) {
-		if (!str_list_add(mem, sllnew, strdup(sl->str))) {
-			stack;
-			return 0;
-		}
+		if (!str_list_add(mem, sllnew, dm_pool_strdup(mem, sl->str)))
+			return_0;
 	}
 
 	return 1;
@@ -85,7 +78,7 @@ 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 list *sll, const char *str)
 {
 	struct str_list *sl;
 
@@ -99,7 +92,7 @@ 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 list *sll, const struct list *sll2)
 {
 	struct str_list *sl;
 
@@ -113,7 +106,7 @@ 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 list *sll, const struct list *sll2)
 {
 	struct str_list *sl;
 

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
  */
@@ -19,9 +19,10 @@
 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);
+int str_list_match_item(const struct list *sll, const char *str);
+int str_list_match_list(const struct list *sll, const struct list *sll2);
+int str_list_lists_equal(const struct list *sll, const struct list *sll2);
+int str_list_dup(struct dm_pool *mem, struct list *sllnew,
+		 const struct list *sllold);
 
 #endif

lib/device/dev-cache.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 +19,7 @@
 #include "btree.h"
 #include "filter.h"
 #include "filter-persistent.h"
+#include "toolcontext.h"
 
 #include <unistd.h>
 #include <sys/param.h>
@@ -38,6 +39,7 @@ static struct {
 	struct dm_pool *mem;
 	struct dm_hash_table *names;
 	struct btree *devices;
+	struct dm_regex *preferred_names_matcher;
 
 	int has_scanned;
 	struct list dirs;
@@ -129,15 +131,52 @@ static struct device *_dev_create(dev_t d)
 	return dev;
 }
 
+void dev_set_preferred_name(struct str_list *sl, struct device *dev)
+{
+	/*
+	 * Don't interfere with ordering specified in config file.
+	 */
+	if (_cache.preferred_names_matcher)
+		return;
+
+	log_debug("%s: New preferred name", sl->str);
+	list_del(&sl->list);
+	list_add_h(&dev->aliases, &sl->list);
+}
+
 /* Return 1 if we prefer path1 else return 0 */
 static int _compare_paths(const char *path0, const char *path1)
 {
 	int slash0 = 0, slash1 = 0;
+	int m0, m1;
 	const char *p;
 	char p0[PATH_MAX], p1[PATH_MAX];
 	char *s0, *s1;
 	struct stat stat0, stat1;
 
+	/*
+	 * FIXME Better to compare patterns one-at-a-time against all names.
+	 */
+	if (_cache.preferred_names_matcher) {
+		m0 = dm_regex_match(_cache.preferred_names_matcher, path0);
+		m1 = dm_regex_match(_cache.preferred_names_matcher, path1);
+
+		if (m0 != m1) {
+			if (m0 < 0)
+				return 1;
+			if (m1 < 0)
+				return 0;
+			if (m0 < m1)
+				return 1;
+			if (m1 < m0)
+				return 0;
+		}
+	}
+
+	/*
+	 * Built-in rules.
+	 */
+
 	/* Return the path with fewer slashes */
 	for (p = path0; p++; p = (const char *) strchr(p, '/'))
 		slash0++;
@@ -166,11 +205,11 @@ static int _compare_paths(const char *path0, const char *path1)
 			*s1 = '\0';
 		}
 		if (lstat(p0, &stat0)) {
-			log_sys_error("lstat", p0);
+			log_sys_very_verbose("lstat", p0);
 			return 1;
 		}
 		if (lstat(p1, &stat1)) {
-			log_sys_error("lstat", p1);
+			log_sys_very_verbose("lstat", p1);
 			return 0;
 		}
 		if (S_ISLNK(stat0.st_mode) && !S_ISLNK(stat1.st_mode))
@@ -197,10 +236,8 @@ static int _add_alias(struct device *dev, const char *path)
 	const char *oldpath;
 	int prefer_old = 1;
 
-	if (!sl) {
-		stack;
-		return 0;
-	}
+	if (!sl)
+		return_0;
 
 	/* Is name already there? */
 	list_iterate_items(strl, &dev->aliases) {
@@ -210,10 +247,8 @@ static int _add_alias(struct device *dev, const char *path)
 		}
 	}
 
-	if (!(sl->str = dm_pool_strdup(_cache.mem, path))) {
-		stack;
-		return 0;
-	}
+	if (!(sl->str = dm_pool_strdup(_cache.mem, path)))
+		return_0;
 
 	if (!list_empty(&dev->aliases)) {
 		oldpath = list_item(dev->aliases.n, struct str_list)->str;
@@ -255,14 +290,10 @@ static int _insert_dev(const char *path, dev_t d)
 						   (uint32_t) d))) {
 		/* create new device */
 		if (loopfile) {
-			if (!(dev = dev_create_file(path, NULL, NULL, 0))) {
-				stack;
-				return 0;
-			}
-		} else if (!(dev = _dev_create(d))) {
-			stack;
-			return 0;
-		}
+			if (!(dev = dev_create_file(path, NULL, NULL, 0)))
+				return_0;
+		} else if (!(dev = _dev_create(d)))
+			return_0;
 
 		if (!(btree_insert(_cache.devices, (uint32_t) d, dev))) {
 			log_err("Couldn't insert device into binary tree.");
@@ -330,10 +361,8 @@ static int _insert_dir(const char *dir)
 				continue;
 			}
 
-			if (!(path = _join(dir, dirent[n]->d_name))) {
-				stack;
-				return 0;
-			}
+			if (!(path = _join(dir, dirent[n]->d_name)))
+				return_0;
 
 			_collapse_slashes(path);
 			r &= _insert(path, 1);
@@ -361,10 +390,8 @@ static int _insert_file(const char *path)
 		return 0;
 	}
 
-	if (!_insert_dev(path, 0)) {
-		stack;
-		return 0;
-	}
+	if (!_insert_dev(path, 0))
+		return_0;
 
 	return 1;
 }
@@ -400,10 +427,8 @@ static int _insert(const char *path, int rec)
 			return 0;
 		}
 
-		if (!_insert_dev(path, info.st_rdev)) {
-			stack;
-			return 0;
-		}
+		if (!_insert_dev(path, info.st_rdev))
+			return_0;
 
 		r = 1;
 	}
@@ -441,21 +466,76 @@ void dev_cache_scan(int do_scan)
 		_full_scan(1);
 }
 
-int dev_cache_init(void)
+static int _init_preferred_names(struct cmd_context *cmd)
+{
+	const struct config_node *cn;
+	struct config_value *v;
+	struct dm_pool *scratch = NULL;
+	char **regex;
+	unsigned count = 0;
+	int i, r = 0;
+
+	_cache.preferred_names_matcher = NULL;
+
+	if (!(cn = find_config_tree_node(cmd, "devices/preferred_names")) ||
+	    cn->v->type == CFG_EMPTY_ARRAY) {
+		log_very_verbose("devices/preferred_names not found in config file: "
+				 "using built-in preferences");
+		return 1;
+	}
+
+	for (v = cn->v; v; v = v->next) {
+		if (v->type != CFG_STRING) {
+			log_error("preferred_names patterns must be enclosed in quotes");
+			return 0;
+		}
+
+		count++;
+	}
+
+	if (!(scratch = dm_pool_create("preferred device name matcher", 1024)))
+		return_0;
+
+	if (!(regex = dm_pool_alloc(scratch, sizeof(*regex) * count))) {
+		log_error("Failed to allocate preferred device name "
+			  "pattern list.");
+		goto out;
+	}
+
+	for (v = cn->v, i = count - 1; v; v = v->next, i--) {
+		if (!(regex[i] = dm_pool_strdup(scratch, v->v.str))) {
+			log_error("Failed to allocate a preferred device name "
+				  "pattern.");
+			goto out;
+		}
+	}
+
+	if (!(_cache.preferred_names_matcher =
+		dm_regex_create(_cache.mem,(const char **) regex, count))) {
+		log_error("Preferred device name pattern matcher creation failed.");
+		goto out;
+	}
+
+	r = 1;
+
+out:
+	dm_pool_destroy(scratch);
+
+	return r;
+}
+
+int dev_cache_init(struct cmd_context *cmd)
 {
 	_cache.names = NULL;
 	_cache.has_scanned = 0;
 
-	if (!(_cache.mem = dm_pool_create("dev_cache", 10 * 1024))) {
-		stack;
-		return 0;
-	}
+	if (!(_cache.mem = dm_pool_create("dev_cache", 10 * 1024)))
+		return_0;
 
 	if (!(_cache.names = dm_hash_create(128))) {
-		stack;
 		dm_pool_destroy(_cache.mem);
 		_cache.mem = 0;
-		return 0;
+		return_0;
 	}
 
 	if (!(_cache.devices = btree_create(_cache.mem))) {
@@ -466,6 +546,9 @@ int dev_cache_init(void)
 	list_init(&_cache.dirs);
 	list_init(&_cache.files);
 
+	if (!_init_preferred_names(cmd))
+		goto_bad;
+
 	return 1;
 
       bad:
@@ -489,6 +572,9 @@ void dev_cache_exit(void)
 	if (_cache.names)
 		_check_for_open_devices();
 
+	if (_cache.preferred_names_matcher)
+		_cache.preferred_names_matcher = NULL;
+
 	if (_cache.mem) {
 		dm_pool_destroy(_cache.mem);
 		_cache.mem = NULL;
@@ -601,7 +687,8 @@ const char *dev_name_confirmed(struct device *dev, int quiet)
 			continue;
 		}
 
-		log_error("Aborting - please provide new pathname for what "
+		/* Scanning issues this inappropriately sometimes. */
+		log_debug("Aborting - please provide new pathname for what "
 			  "used to be %s", name);
 		return NULL;
 	}

lib/device/dev-cache.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
  */
@@ -30,7 +30,8 @@ struct dev_filter {
 /*
  * The global device cache.
  */
-int dev_cache_init(void);
+struct cmd_context;
+int dev_cache_init(struct cmd_context *cmd);
 void dev_cache_exit(void);
 
 /* Trigger(1) or avoid(0) a scan */
@@ -41,6 +42,8 @@ int dev_cache_add_dir(const char *path);
 int dev_cache_add_loopfile(const char *path);
 struct device *dev_cache_get(const char *name, struct dev_filter *f);
 
+void dev_set_preferred_name(struct str_list *sl, struct device *dev);
+
 /*
  * Object for iterating through the cache.
  */

lib/device/dev-io.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) 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
  */
@@ -170,13 +170,11 @@ static int _aligned_io(struct device_area *where, void *buffer,
 	struct device_area widened;
 
 	if (!(where->dev->flags & DEV_REGULAR) &&
-	    !_get_block_size(where->dev, &block_size)) {
-		stack;
-		return 0;
-	}
+	    !_get_block_size(where->dev, &block_size))
+		return_0;
 
 	if (!block_size)
-		block_size = getpagesize();
+		block_size = lvm_getpagesize();
 
 	_widen_region(block_size, where, &widened);
 
@@ -200,10 +198,8 @@ static int _aligned_io(struct device_area *where, void *buffer,
 
 	/* channel the io through the bounce buffer */
 	if (!_io(&widened, bounce, 0)) {
-		if (!should_write) {
-			stack;
-			return 0;
-		}
+		if (!should_write)
+			return_0;
 		/* FIXME pre-extend the file */
 		memset(bounce, '\n', widened.size);
 	}
@@ -292,11 +288,14 @@ int dev_get_sectsize(struct device *dev, uint32_t *size)
 
 	if (ioctl(fd, BLKSSZGET, &s) < 0) {
 		log_sys_error("ioctl BLKSSZGET", name);
-		close(fd);
+		if (close(fd))
+			log_sys_error("close", name);
 		return 0;
 	}
 
-	close(fd);
+	if (close(fd))
+		log_sys_error("close", name);
+
 	*size = (uint32_t) s;
 
 	log_very_verbose("%s: sector size is %" PRIu32 " bytes", name, *size);
@@ -337,7 +336,7 @@ int dev_open_flags(struct device *dev, int flags, int direct, int quiet)
 
 		if (dev->open_count && !need_excl) {
 			/* FIXME Ensure we never get here */
-			log_debug("WARNING: %s already opened read-only", 
+			log_debug("WARNING: %s already opened read-only",
 				  dev_name(dev));
 			dev->open_count++;
 		}
@@ -351,15 +350,18 @@ int dev_open_flags(struct device *dev, int flags, int direct, int quiet)
 
 	if (dev->flags & DEV_REGULAR)
 		name = dev_name(dev);
-	else if (!(name = dev_name_confirmed(dev, quiet))) {
-		stack;
-		return 0;
-	}
+	else if (!(name = dev_name_confirmed(dev, quiet)))
+		return_0;
 
-	if (!(dev->flags & DEV_REGULAR) &&
-	    ((stat(name, &buf) < 0) || (buf.st_rdev != dev->dev))) {
-		log_error("%s: stat failed: Has device name changed?", name);
-		return 0;
+	if (!(dev->flags & DEV_REGULAR)) {
+		if (stat(name, &buf) < 0) {
+			log_sys_error("%s: stat failed", name);
+			return 0;
+		}
+		if (buf.st_rdev != dev->dev) {
+			log_error("%s: device changed", name);
+			return 0;
+		}
 	}
 
 #ifdef O_DIRECT_SUPPORT
@@ -514,8 +516,8 @@ static int _dev_close(struct device *dev, int immediate)
 
 	/* Close unless device is known to belong to a locked VG */
 	if (immediate ||
-	    (dev->open_count < 1 && 
-	     (!(info = info_from_pvid(dev->pvid)) ||
+	    (dev->open_count < 1 &&
+	     (!(info = info_from_pvid(dev->pvid, 0)) ||
 	      !info->vginfo ||
 	      !vgname_is_locked(info->vginfo->vgname))))
 		_close(dev);
@@ -549,10 +551,8 @@ int dev_read(struct device *dev, uint64_t offset, size_t len, void *buffer)
 {
 	struct device_area where;
 
-	if (!dev->open_count) {
-		stack;
-		return 0;
-	}
+	if (!dev->open_count)
+		return_0;
 
 	where.dev = dev;
 	where.start = offset;
@@ -561,6 +561,35 @@ int dev_read(struct device *dev, uint64_t offset, size_t len, void *buffer)
 	return _aligned_io(&where, buffer, 0);
 }
 
+/*
+ * Read from 'dev' into 'buf', possibly in 2 distinct regions, denoted
+ * by (offset,len) and (offset2,len2).  Thus, the total size of
+ * 'buf' should be len+len2.
+ */
+int dev_read_circular(struct device *dev, uint64_t offset, size_t len,
+		      uint64_t offset2, size_t len2, void *buf)
+{
+	if (!dev_read(dev, offset, len, buf)) {
+		log_error("Read from %s failed", dev_name(dev));
+		return 0;
+	}
+
+	/*
+	 * The second region is optional, and allows for
+	 * a circular buffer on the device.
+	 */
+	if (!len2)
+		return 1;
+
+	if (!dev_read(dev, offset2, len2, buf + len)) {
+		log_error("Circular read from %s failed",
+			  dev_name(dev));
+		return 0;
+	}
+
+	return 1;
+}
+
 /* FIXME If O_DIRECT can't extend file, dev_extend first; dev_truncate after.
  *       But fails if concurrent processes writing
  */
@@ -570,10 +599,8 @@ int dev_append(struct device *dev, size_t len, void *buffer)
 {
 	int r;
 
-	if (!dev->open_count) {
-		stack;
-		return 0;
-	}
+	if (!dev->open_count)
+		return_0;
 
 	r = dev_write(dev, dev->end, len, buffer);
 	dev->end += (uint64_t) len;
@@ -588,10 +615,8 @@ int dev_write(struct device *dev, uint64_t offset, size_t len, void *buffer)
 {
 	struct device_area where;
 
-	if (!dev->open_count) {
-		stack;
-		return 0;
-	}
+	if (!dev->open_count)
+		return_0;
 
 	where.dev = dev;
 	where.start = offset;
@@ -605,12 +630,10 @@ int dev_write(struct device *dev, uint64_t offset, size_t len, void *buffer)
 int dev_set(struct device *dev, uint64_t offset, size_t len, int value)
 {
 	size_t s;
-	char buffer[4096];
+	char buffer[4096] __attribute((aligned(8)));
 
-	if (!dev_open(dev)) {
-		stack;
-		return 0;
-	}
+	if (!dev_open(dev))
+		return_0;
 
 	if ((offset % SECTOR_SIZE) || (len % SECTOR_SIZE))
 		log_debug("Wiping %s at %" PRIu64 " length %" PRIsize_t,

lib/device/dev-md.c

@@ -1,6 +1,6 @@
 /*
  * Copyright (C) 2004 Luca Berra
- * 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.
  *
@@ -17,25 +17,65 @@
 #include "metadata.h"
 #include "xlate.h"
 
+#ifdef linux
+
 /* Lifted from <linux/raid/md_p.h> because of difficulty including it */
 
 #define MD_SB_MAGIC 0xa92b4efc
-#define MD_RESERVED_BYTES (64 * 1024)
+#define MD_RESERVED_BYTES (64 * 1024ULL)
 #define MD_RESERVED_SECTORS (MD_RESERVED_BYTES / 512)
 #define MD_NEW_SIZE_SECTORS(x) ((x & ~(MD_RESERVED_SECTORS - 1)) \
 				- MD_RESERVED_SECTORS)
 
+static int _dev_has_md_magic(struct device *dev, uint64_t sb_offset)
+{
+	uint32_t md_magic;
+
+	/* Version 1 is little endian; version 0.90.0 is machine endian */
+	if (dev_read(dev, sb_offset, sizeof(uint32_t), &md_magic) &&
+	    ((md_magic == xlate32(MD_SB_MAGIC)) ||
+	     (md_magic == MD_SB_MAGIC)))
+		return 1;
+
+	return 0;
+}
+
+/*
+ * Calculate the position of the superblock.
+ * It is always aligned to a 4K boundary and
+ * depending on minor_version, it can be:
+ * 0: At least 8K, but less than 12K, from end of device
+ * 1: At start of device
+ * 2: 4K from start of device.
+ */
+static uint64_t _v1_sb_offset(uint64_t size, unsigned minor_version)
+{
+	uint64_t sb_offset;
+
+	switch(minor_version) {
+	case 0:
+		sb_offset = (size - 8 * 2) & ~(4 * 2 - 1ULL);
+		break;
+	case 1:
+		sb_offset = 0;
+		break;
+	case 2:
+		sb_offset = 4 * 2;
+		break;
+	}
+	sb_offset <<= SECTOR_SHIFT;
+
+	return sb_offset;
+}
+
 /*
  * Returns -1 on error
  */
 int dev_is_md(struct device *dev, uint64_t *sb)
 {
-	int ret = 0;
-
-#ifdef linux
-
+	int ret = 1;
+	unsigned minor = 0;
 	uint64_t size, sb_offset;
-	uint32_t md_magic;
 
 	if (!dev_get_size(dev, &size)) {
 		stack;
@@ -50,20 +90,37 @@ int dev_is_md(struct device *dev, uint64_t *sb)
 		return -1;
 	}
 
-	sb_offset = MD_NEW_SIZE_SECTORS(size) << SECTOR_SHIFT;
-
 	/* Check if it is an md component device. */
-	if (dev_read(dev, sb_offset, sizeof(uint32_t), &md_magic) &&
-	    (md_magic == xlate32(MD_SB_MAGIC))) {
-		if (sb)
-			*sb = sb_offset;
-		ret = 1;
-	}
+	/* Version 0.90.0 */
+	sb_offset = MD_NEW_SIZE_SECTORS(size) << SECTOR_SHIFT;
+	if (_dev_has_md_magic(dev, sb_offset))
+		goto out;
 
+	/* Version 1, try v1.0 -> v1.2 */
+	do {
+		sb_offset = _v1_sb_offset(size, minor);
+		if (_dev_has_md_magic(dev, sb_offset))
+			goto out;
+	} while (++minor <= 2);
+
+	ret = 0;
+
+out:
 	if (!dev_close(dev))
 		stack;
 
-#endif
+	if (ret && sb)
+		*sb = sb_offset;
+
 	return ret;
 }
 
+#else
+
+int dev_is_md(struct device *dev __attribute((unused)),
+	      uint64_t *sb __attribute((unused)))
+{
+	return 0;
+}
+
+#endif

lib/device/device.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) 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
  */
@@ -27,16 +27,16 @@
 #define PART_OFFSET UINT64_C(0x1BE)
 
 struct partition {
-        uint8_t boot_ind;
-        uint8_t head;
-        uint8_t sector;
-        uint8_t cyl;
-        uint8_t sys_ind;	/* partition type */
-        uint8_t end_head;
-        uint8_t end_sector;
-        uint8_t end_cyl;
-        uint32_t start_sect;
-        uint32_t nr_sects;
+	uint8_t boot_ind;
+	uint8_t head;
+	uint8_t sector;
+	uint8_t cyl;
+	uint8_t sys_ind;	/* partition type */
+	uint8_t end_head;
+	uint8_t end_sector;
+	uint8_t end_cyl;
+	uint32_t start_sect;
+	uint32_t nr_sects;
 } __attribute__((packed));
 
 static int _is_partitionable(struct device *dev)
@@ -53,7 +53,7 @@ static int _has_partition_table(struct device *dev)
 {
 	int ret = 0;
 	unsigned p;
-	uint8_t buf[SECTOR_SIZE];
+	uint16_t buf[SECTOR_SIZE/sizeof(uint16_t)];
 	uint16_t *part_magic;
 	struct partition *part;
 
@@ -62,17 +62,15 @@ static int _has_partition_table(struct device *dev)
 		return -1;
 	}
 
-	if (!dev_read(dev, UINT64_C(0), sizeof(buf), &buf)) {
-		stack;
-		goto out;
-	}
+	if (!dev_read(dev, UINT64_C(0), sizeof(buf), &buf))
+		goto_out;
 
 	/* FIXME Check for other types of partition table too */
 
 	/* Check for msdos partition table */
-	part_magic = (uint16_t *)(buf + PART_MAGIC_OFFSET);
+	part_magic = buf + PART_MAGIC_OFFSET/sizeof(buf[0]);
 	if ((*part_magic == xlate16(PART_MAGIC))) {
-		part = (struct partition *) (buf + PART_OFFSET);
+		part = (struct partition *) (buf + PART_OFFSET/sizeof(buf[0]));
 		for (p = 0; p < 4; p++, part++) {
 			/* Table is invalid if boot indicator not 0 or 0x80 */
 			if ((part->boot_ind & 0x7f)) {

lib/device/device.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
  */
@@ -78,6 +78,8 @@ int dev_fd(struct device *dev);
 const char *dev_name(const struct device *dev);
 
 int dev_read(struct device *dev, uint64_t offset, size_t len, void *buffer);
+int dev_read_circular(struct device *dev, uint64_t offset, size_t len,
+		      uint64_t offset2, size_t len2, void *buf);
 int dev_write(struct device *dev, uint64_t offset, size_t len, void *buffer);
 int dev_append(struct device *dev, size_t len, void *buffer);
 int dev_set(struct device *dev, uint64_t offset, size_t len, int value);

lib/display/display.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) 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
  */
@@ -30,6 +30,7 @@ static struct {
 } _policies[] = {
 	{
 	ALLOC_CONTIGUOUS, "contiguous"}, {
+	ALLOC_CLING, "cling"}, {
 	ALLOC_NORMAL, "normal"}, {
 	ALLOC_ANYWHERE, "anywhere"}, {
 	ALLOC_INHERIT, "inherit"}
@@ -81,6 +82,12 @@ uint64_t units_to_bytes(const char *units, char *unit_type)
 	case 't':
 		v *= KILO * KILO * KILO * KILO;
 		break;
+	case 'p':
+		v *= KILO * KILO * KILO * KILO * KILO;
+		break;
+	case 'e':
+		v *= KILO * KILO * KILO * KILO * KILO * KILO;
+		break;
 #undef KILO
 #define KILO UINT64_C(1000)
 	case 'K':
@@ -95,6 +102,12 @@ uint64_t units_to_bytes(const char *units, char *unit_type)
 	case 'T':
 		v *= KILO * KILO * KILO * KILO;
 		break;
+	case 'P':
+		v *= KILO * KILO * KILO * KILO * KILO;
+		break;
+	case 'E':
+		v *= KILO * KILO * KILO * KILO * KILO * KILO;
+		break;
 #undef KILO
 	default:
 		return 0;
@@ -134,7 +147,8 @@ alloc_policy_t get_alloc_from_string(const char *str)
 }
 
 /* Size supplied in sectors */
-static const char *_display_size(struct cmd_context *cmd, uint64_t size, size_len_t sl)
+static const char *_display_size(const struct cmd_context *cmd,
+				 uint64_t size, size_len_t sl)
 {
 	int s;
 	int suffix = 1, precision;
@@ -142,6 +156,8 @@ static const char *_display_size(struct cmd_context *cmd, uint64_t size, size_le
 	uint64_t units = UINT64_C(1024);
 	char *size_buf = NULL;
 	const char *size_str[][3] = {
+		{" Exabyte", " EB", "E"},
+		{" Petabyte", " PB", "P"},
 		{" Terabyte", " TB", "T"},
 		{" Gigabyte", " GB", "G"},
 		{" Megabyte", " MB", "M"},
@@ -160,7 +176,7 @@ static const char *_display_size(struct cmd_context *cmd, uint64_t size, size_le
 
 	suffix = cmd->current_settings.suffix;
 
-	for (s = 0; s < 8; s++)
+	for (s = 0; s < 10; s++)
 		if (toupper((int) cmd->current_settings.unit_type) ==
 		    *size_str[s][2])
 			break;
@@ -170,17 +186,17 @@ static const char *_display_size(struct cmd_context *cmd, uint64_t size, size_le
 		return size_buf;
 	}
 
-	if (s < 8) {
+	size *= UINT64_C(512);
+
+	if (s < 10)
 		byte = cmd->current_settings.unit_factor;
-		size *= UINT64_C(512);
-	} else {
-		size /= 2;
+	else {
 		suffix = 1;
 		if (cmd->current_settings.unit_type == 'H')
 			units = UINT64_C(1000);
 		else
 			units = UINT64_C(1024);
-		byte = units * units * units;
+		byte = units * units * units * units * units * units;
 		s = 0;
 		while (size_str[s] && size < byte)
 			s++, byte /= units;
@@ -202,24 +218,24 @@ static const char *_display_size(struct cmd_context *cmd, uint64_t size, size_le
 	return size_buf;
 }
 
-const char *display_size_long(struct cmd_context *cmd, uint64_t size)
+const char *display_size_long(const struct cmd_context *cmd, uint64_t size)
 {
 	return _display_size(cmd, size, SIZE_LONG);
 }
 
-const char *display_size_units(struct cmd_context *cmd, uint64_t size)
+const char *display_size_units(const struct cmd_context *cmd, uint64_t size)
 {
 	return _display_size(cmd, size, SIZE_UNIT);
 }
 
-const char *display_size(struct cmd_context *cmd, uint64_t size)
+const char *display_size(const struct cmd_context *cmd, uint64_t size)
 {
 	return _display_size(cmd, size, SIZE_SHORT);
 }
 
-void pvdisplay_colons(struct physical_volume *pv)
+void pvdisplay_colons(const struct physical_volume *pv)
 {
-	char uuid[64];
+	char uuid[64] __attribute((aligned(8)));
 
 	if (!pv)
 		return;
@@ -230,7 +246,7 @@ void pvdisplay_colons(struct physical_volume *pv)
 	}
 
 	log_print("%s:%s:%" PRIu64 ":-1:%u:%u:-1:%" PRIu32 ":%u:%u:%u:%s",
-		  dev_name(pv->dev), pv->vg_name, pv->size,
+		  pv_dev_name(pv), pv->vg_name, pv->size,
 		  /* FIXME pv->pv_number, Derive or remove? */
 		  pv->status,	/* FIXME Support old or new format here? */
 		  pv->status & ALLOCATABLE_PV,	/* FIXME remove? */
@@ -243,14 +259,43 @@ void pvdisplay_colons(struct physical_volume *pv)
 	return;
 }
 
+void pvdisplay_segments(const struct physical_volume *pv)
+{
+	const struct pv_segment *pvseg;
+
+	if (pv->pe_size)
+		log_print("--- Physical Segments ---");
+
+	list_iterate_items(pvseg, &pv->segments) {
+		log_print("Physical extent %u to %u:",
+			  pvseg->pe, pvseg->pe + pvseg->len - 1);
+
+		if (pvseg_is_allocated(pvseg)) {
+			log_print("  Logical volume\t%s%s/%s",
+				  pvseg->lvseg->lv->vg->cmd->dev_dir,
+				  pvseg->lvseg->lv->vg->name,
+				  pvseg->lvseg->lv->name);
+			log_print("  Logical extents\t%d to %d",
+				  pvseg->lvseg->le, pvseg->lvseg->le +
+				  pvseg->lvseg->len - 1);
+		} else
+			log_print("  FREE");
+	}
+
+	log_print(" ");
+	return;
+}
+
 /* FIXME Include label fields */
-void pvdisplay_full(struct cmd_context *cmd, struct physical_volume *pv,
+void pvdisplay_full(const struct cmd_context *cmd,
+		    const struct physical_volume *pv,
 		    void *handle __attribute((unused)))
 {
-	char uuid[64];
+	char uuid[64] __attribute((aligned(8)));
 	const char *size;
 
 	uint32_t pe_free;
+	uint64_t data_size, pvsize, unusable;
 
 	if (!pv)
 		return;
@@ -261,23 +306,25 @@ void pvdisplay_full(struct cmd_context *cmd, struct physical_volume *pv,
 	}
 
 	log_print("--- %sPhysical volume ---", pv->pe_size ? "" : "NEW ");
-	log_print("PV Name               %s", dev_name(pv->dev));
-	log_print("VG Name               %s%s", pv->vg_name,
+	log_print("PV Name               %s", pv_dev_name(pv));
+	log_print("VG Name               %s%s",
+		  is_orphan(pv) ? "" : pv->vg_name,
 		  pv->status & EXPORTED_VG ? " (exported)" : "");
 
-	size = display_size(cmd, (uint64_t) pv->size);
-	if (pv->pe_size && pv->pe_count) {
+	data_size = (uint64_t) pv->pe_count * pv->pe_size;
+	if (pv->size > data_size + pv->pe_start) {
+		pvsize = pv->size;
+		unusable = pvsize - data_size;
+	} else {
+		pvsize = data_size + pv->pe_start;
+		unusable = pvsize - pv->size;
+	}
 
-/******** FIXME display LVM on-disk data size
-		size2 = display_size(cmd, pv->size);
-********/
-
-		log_print("PV Size               %s" " / not usable %s",	/*  [LVM: %s]", */
-			  size,
-			  display_size(cmd, (pv->size -
-				       (uint64_t) pv->pe_count * pv->pe_size)));
-
-	} else
+	size = display_size(cmd, pvsize);
+	if (data_size)
+		log_print("PV Size               %s / not usable %s",	/*  [LVM: %s]", */
+			  size, display_size(cmd, unusable));
+	else
 		log_print("PV Size               %s", size);
 
 	/* PV number not part of LVM2 design
@@ -304,22 +351,20 @@ void pvdisplay_full(struct cmd_context *cmd, struct physical_volume *pv,
 	return;
 }
 
-int pvdisplay_short(struct cmd_context *cmd __attribute((unused)),
-		    struct volume_group *vg __attribute((unused)),
-		    struct physical_volume *pv,
+int pvdisplay_short(const struct cmd_context *cmd __attribute((unused)),
+		    const struct volume_group *vg __attribute((unused)),
+		    const struct physical_volume *pv,
 		    void *handle __attribute((unused)))
 {
-	char uuid[64];
+	char uuid[64] __attribute((aligned(8)));
 
 	if (!pv)
 		return 0;
 
-	if (!id_write_format(&pv->id, uuid, sizeof(uuid))) {
-		stack;
-		return 0;
-	}
+	if (!id_write_format(&pv->id, uuid, sizeof(uuid)))
+		return_0;
 
-	log_print("PV Name               %s     ", dev_name(pv->dev));
+	log_print("PV Name               %s     ", pv_dev_name(pv));
 	/* FIXME  pv->pv_number); */
 	log_print("PV UUID               %s", *uuid ? uuid : "none");
 	log_print("PV Status             %sallocatable",
@@ -331,11 +376,11 @@ int pvdisplay_short(struct cmd_context *cmd __attribute((unused)),
 	return 0;
 }
 
-void lvdisplay_colons(struct logical_volume *lv)
+void lvdisplay_colons(const struct logical_volume *lv)
 {
 	int inkernel;
 	struct lvinfo info;
-	inkernel = lv_info(lv->vg->cmd, lv, &info, 1) && info.exists;
+	inkernel = lv_info(lv->vg->cmd, lv, &info, 1, 0) && info.exists;
 
 	log_print("%s%s/%s:%s:%d:%d:-1:%d:%" PRIu64 ":%d:-1:%d:%d:%d:%d",
 		  lv->vg->cmd->dev_dir,
@@ -351,21 +396,20 @@ void lvdisplay_colons(struct logical_volume *lv)
 	return;
 }
 
-int lvdisplay_full(struct cmd_context *cmd, struct logical_volume *lv,
+int lvdisplay_full(struct cmd_context *cmd,
+		   const struct logical_volume *lv,
 		   void *handle __attribute((unused)))
 {
 	struct lvinfo info;
 	int inkernel, snap_active = 0;
-	char uuid[64];
+	char uuid[64] __attribute((aligned(8)));
 	struct lv_segment *snap_seg = NULL;
 	float snap_percent;	/* fused, fsize; */
 
-	if (!id_write_format(&lv->lvid.id[1], uuid, sizeof(uuid))) {
-		stack;
-		return 0;
-	}
+	if (!id_write_format(&lv->lvid.id[1], uuid, sizeof(uuid)))
+		return_0;
 
-	inkernel = lv_info(cmd, lv, &info, 1) && info.exists;
+	inkernel = lv_info(cmd, lv, &info, 1, 1) && info.exists;
 
 	log_print("--- Logical volume ---");
 
@@ -446,7 +490,15 @@ int lvdisplay_full(struct cmd_context *cmd, struct logical_volume *lv,
 ***********/
 
 	log_print("Allocation             %s", get_alloc_string(lv->alloc));
-	log_print("Read ahead sectors     %u", lv->read_ahead);
+	if (lv->read_ahead == DM_READ_AHEAD_AUTO)
+		log_print("Read ahead sectors     auto");
+	else if (lv->read_ahead == DM_READ_AHEAD_NONE)
+		log_print("Read ahead sectors     0");
+	else
+		log_print("Read ahead sectors     %u", lv->read_ahead);
+
+	if (inkernel && lv->read_ahead != info.read_ahead)
+		log_print("- currently set to     %u", info.read_ahead);
 
 	if (lv->status & FIXED_MINOR) {
 		if (lv->major >= 0)
@@ -470,7 +522,7 @@ void display_stripe(const struct lv_segment *seg, uint32_t s, const char *pre)
 		/* FIXME Re-check the conditions for 'Missing' */
 		log_print("%sPhysical volume\t%s", pre,
 			  seg_pv(seg, s) ?
-			  dev_name(seg_dev(seg, s)) :
+			  pv_dev_name(seg_pv(seg, s)) :
 			    "Missing");
 
 		if (seg_pv(seg, s))
@@ -493,9 +545,9 @@ void display_stripe(const struct lv_segment *seg, uint32_t s, const char *pre)
 	}
 }
 
-int lvdisplay_segments(struct logical_volume *lv)
+int lvdisplay_segments(const struct logical_volume *lv)
 {
-	struct lv_segment *seg;
+	const struct lv_segment *seg;
 
 	log_print("--- Segments ---");
 
@@ -513,16 +565,18 @@ int lvdisplay_segments(struct logical_volume *lv)
 	return 1;
 }
 
-void vgdisplay_extents(struct volume_group *vg __attribute((unused)))
+void vgdisplay_extents(const struct volume_group *vg __attribute((unused)))
 {
 	return;
 }
 
-void vgdisplay_full(struct volume_group *vg)
+void vgdisplay_full(const struct volume_group *vg)
 {
 	uint32_t access;
 	uint32_t active_pvs;
-	char uuid[64];
+	uint32_t lv_count = 0;
+	struct lv_list *lvl;
+	char uuid[64] __attribute((aligned(8)));
 
 	if (vg->status & PARTIAL_VG)
 		active_pvs = list_size(&vg->pvs);
@@ -550,13 +604,18 @@ void vgdisplay_full(struct volume_group *vg)
 	/* vg number not part of LVM2 design
 	   log_print ("VG #                  %u\n", vg->vg_number);
 	 */
-	if (vg->status & CLUSTERED) {
+	if (vg_is_clustered(vg)) {
 		log_print("Clustered             yes");
 		log_print("Shared                %s",
 			  vg->status & SHARED ? "yes" : "no");
 	}
+
+	list_iterate_items(lvl, &vg->lvs)
+		if (lv_is_visible(lvl->lv) && !(lvl->lv->status & SNAPSHOT))
+			lv_count++;
+
 	log_print("MAX LV                %u", vg->max_lv);
-	log_print("Cur LV                %u", vg->lv_count + vg->snapshot_count);
+	log_print("Cur LV                %u", lv_count);
 	log_print("Open LV               %u", lvs_in_vg_opened(vg));
 /****** FIXME Max LV Size
       log_print ( "MAX LV Size           %s",
@@ -597,17 +656,23 @@ void vgdisplay_full(struct volume_group *vg)
 	return;
 }
 
-void vgdisplay_colons(struct volume_group *vg)
+void vgdisplay_colons(const struct volume_group *vg)
 {
 	uint32_t active_pvs;
+	uint32_t lv_count;
+	struct lv_list *lvl;
 	const char *access;
-	char uuid[64];
+	char uuid[64] __attribute((aligned(8)));
 
 	if (vg->status & PARTIAL_VG)
 		active_pvs = list_size(&vg->pvs);
 	else
 		active_pvs = vg->pv_count;
 
+	list_iterate_items(lvl, &vg->lvs)
+		if (lv_is_visible(lvl->lv) && !(lvl->lv->status & SNAPSHOT))
+			lv_count++;
+
 	switch (vg->status & (LVM_READ | LVM_WRITE)) {
 		case LVM_READ | LVM_WRITE:
 			access = "r/w";
@@ -643,13 +708,13 @@ void vgdisplay_colons(struct volume_group *vg)
 		(uint64_t) vg->extent_count * (vg->extent_size / 2),
 		vg->extent_size / 2,
 		vg->extent_count,
-		vg->extent_count - vg->free_count, 
+		vg->extent_count - vg->free_count,
 		vg->free_count,
 		uuid[0] ? uuid : "none");
 	return;
 }
 
-void vgdisplay_short(struct volume_group *vg)
+void vgdisplay_short(const struct volume_group *vg)
 {
 	log_print("\"%s\" %-9s [%-9s used / %s free]", vg->name,
 /********* FIXME if "open" print "/used" else print "/idle"???  ******/
@@ -663,21 +728,52 @@ void vgdisplay_short(struct volume_group *vg)
 	return;
 }
 
-void display_formats(struct cmd_context *cmd)
+void display_formats(const struct cmd_context *cmd)
 {
-	struct format_type *fmt;
+	const struct format_type *fmt;
 
 	list_iterate_items(fmt, &cmd->formats) {
 		log_print("%s", fmt->name);
 	}
 }
 
-void display_segtypes(struct cmd_context *cmd)
+void display_segtypes(const struct cmd_context *cmd)
 {
-	struct segment_type *segtype;
+	const struct segment_type *segtype;
 
 	list_iterate_items(segtype, &cmd->segtypes) {
 		log_print("%s", segtype->name);
 	}
 }
 
+char yes_no_prompt(const char *prompt, ...)
+{
+	int c = 0, ret = 0;
+	va_list ap;
+
+	sigint_allow();
+	do {
+		if (c == '\n' || !c) {
+			va_start(ap, prompt);
+			vprintf(prompt, ap);
+			va_end(ap);
+		}
+
+		if ((c = getchar()) == EOF) {
+			ret = 'n';
+			break;
+		}
+
+		c = tolower(c);
+		if ((c == 'y') || (c == 'n'))
+			ret = c;
+	} while (!ret || c != '\n');
+
+	sigint_restore();
+
+	if (c != '\n')
+		printf("\n");
+
+	return ret;
+}
+

lib/display/display.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
  */
@@ -16,38 +16,42 @@
 #ifndef _LVM_DISPLAY_H
 #define _LVM_DISPLAY_H
 
-#include "metadata.h"
+#include "metadata-exported.h"
+#include "locking.h"
 
 #include <stdint.h>
 
 uint64_t units_to_bytes(const char *units, char *unit_type);
 
 /* Specify size in KB */
-const char *display_size(struct cmd_context *cmd, uint64_t size);
-const char *display_size_long(struct cmd_context *cmd, uint64_t size);
-const char *display_size_units(struct cmd_context *cmd, uint64_t size);
+const char *display_size(const struct cmd_context *cmd, uint64_t size);
+const char *display_size_long(const struct cmd_context *cmd, uint64_t size);
+const char *display_size_units(const struct cmd_context *cmd, uint64_t size);
 
 char *display_uuid(char *uuidstr);
 void display_stripe(const struct lv_segment *seg, uint32_t s, const char *pre);
 
-void pvdisplay_colons(struct physical_volume *pv);
-void pvdisplay_full(struct cmd_context *cmd, struct physical_volume *pv,
+void pvdisplay_colons(const struct physical_volume *pv);
+void pvdisplay_segments(const struct physical_volume *pv);
+void pvdisplay_full(const struct cmd_context *cmd,
+		    const struct physical_volume *pv,
 		    void *handle);
-int pvdisplay_short(struct cmd_context *cmd, struct volume_group *vg,
-		    struct physical_volume *pv, void *handle);
+int pvdisplay_short(const struct cmd_context *cmd,
+		    const struct volume_group *vg,
+		    const struct physical_volume *pv, void *handle);
 
-void lvdisplay_colons(struct logical_volume *lv);
-int lvdisplay_segments(struct logical_volume *lv);
-int lvdisplay_full(struct cmd_context *cmd, struct logical_volume *lv,
+void lvdisplay_colons(const struct logical_volume *lv);
+int lvdisplay_segments(const struct logical_volume *lv);
+int lvdisplay_full(struct cmd_context *cmd, const struct logical_volume *lv,
 		   void *handle);
 
-void vgdisplay_extents(struct volume_group *vg);
-void vgdisplay_full(struct volume_group *vg);
-void vgdisplay_colons(struct volume_group *vg);
-void vgdisplay_short(struct volume_group *vg);
+void vgdisplay_extents(const struct volume_group *vg);
+void vgdisplay_full(const struct volume_group *vg);
+void vgdisplay_colons(const struct volume_group *vg);
+void vgdisplay_short(const struct volume_group *vg);
 
-void display_formats(struct cmd_context *cmd);
-void display_segtypes(struct cmd_context *cmd);
+void display_formats(const struct cmd_context *cmd);
+void display_segtypes(const struct cmd_context *cmd);
 
 /*
  * Allocation policy display conversion routines.
@@ -55,4 +59,6 @@ void display_segtypes(struct cmd_context *cmd);
 const char *get_alloc_string(alloc_policy_t alloc);
 alloc_policy_t get_alloc_from_string(const char *str);
 
+char yes_no_prompt(const char *prompt, ...);
+
 #endif

lib/error/errseg.c

@@ -1,13 +1,13 @@
 /*
- * 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
  */
@@ -23,6 +23,8 @@
 #include "targets.h"
 #include "lvm-string.h"
 #include "activate.h"
+#include "str_list.h"
+#include "metadata.h"
 
 static const char *_errseg_name(const struct lv_segment *seg)
 {
@@ -49,7 +51,8 @@ static int _errseg_add_target_line(struct dev_manager *dm __attribute((unused)),
 	return dm_tree_node_add_error_target(node, len);
 }
 
-static int _errseg_target_present(void)
+static int _errseg_target_present(const struct lv_segment *seg __attribute((unused)),
+				  unsigned *attributes __attribute((unused)))
 {
 	static int _errseg_checked = 0;
 	static int _errseg_present = 0;
@@ -64,6 +67,18 @@ static int _errseg_target_present(void)
 }
 #endif
 
+static int _errseg_modules_needed(struct dm_pool *mem,
+				  const struct lv_segment *seg __attribute((unused)),
+				  struct list *modules)
+{
+	if (!str_list_add(mem, modules, "error")) {
+		log_error("error module string list allocation failed");
+		return 0;
+	}
+
+	return 1;
+}
+
 static void _errseg_destroy(const struct segment_type *segtype)
 {
 	dm_free((void *)segtype);
@@ -76,6 +91,7 @@ static struct segtype_handler _error_ops = {
 	.add_target_line = _errseg_add_target_line,
 	.target_present = _errseg_target_present,
 #endif
+	.modules_needed = _errseg_modules_needed,
 	.destroy = _errseg_destroy,
 };
 

lib/filters/filter-composite.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) 2004-2006 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
  */
@@ -50,10 +50,8 @@ struct dev_filter *composite_filter_create(int n, struct dev_filter **filters)
 {
 	struct dev_filter **filters_copy, *cft;
 
-	if (!filters) {
-		stack;
-		return NULL;
-	}
+	if (!filters)
+		return_NULL;
 
 	if (!(filters_copy = dm_malloc(sizeof(*filters) * (n + 1)))) {
 		log_error("composite filters allocation failed");

lib/filters/filter-composite.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/filters/filter-md.c

@@ -1,5 +1,6 @@
 /*
  * Copyright (C) 2004 Luca Berra
+ * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved.
  *
  * This file is part of LVM2.
  *
@@ -18,15 +19,8 @@
 
 #ifdef linux
 
-/* Lifted from <linux/raid/md_p.h> because of difficulty including it */
-
-#define MD_SB_MAGIC 0xa92b4efc
-#define MD_RESERVED_BYTES (64 * 1024)
-#define MD_RESERVED_SECTORS (MD_RESERVED_BYTES / 512)
-#define MD_NEW_SIZE_SECTORS(x) ((x & ~(MD_RESERVED_SECTORS - 1)) \
-				- MD_RESERVED_SECTORS)
-
-static int _ignore_md(struct dev_filter *f, struct device *dev)
+static int _ignore_md(struct dev_filter *f __attribute((unused)),
+		      struct device *dev)
 {
 	int ret;
 	

lib/filters/filter-persistent.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
  */
@@ -17,6 +17,8 @@
 #include "config.h"
 #include "dev-cache.h"
 #include "filter-persistent.h"
+#include "lvm-file.h"
+#include "lvm-string.h"
 
 #include <sys/stat.h>
 #include <fcntl.h>
@@ -26,11 +28,12 @@ struct pfilter {
 	char *file;
 	struct dm_hash_table *devices;
 	struct dev_filter *real;
+	time_t ctime;
 };
 
 /*
- * entries in the table can be in one of these
- * states.
+ * The hash table holds one of these two states
+ * against each entry.
  */
 #define PF_BAD_DEVICE ((void *) 1)
 #define PF_GOOD_DEVICE ((void *) 2)
@@ -40,10 +43,8 @@ static int _init_hash(struct pfilter *pf)
 	if (pf->devices)
 		dm_hash_destroy(pf->devices);
 
-	if (!(pf->devices = dm_hash_create(128))) {
-		stack;
-		return 0;
-	}
+	if (!(pf->devices = dm_hash_create(128)))
+		return_0;
 
 	return 1;
 }
@@ -93,22 +94,26 @@ static int _read_array(struct pfilter *pf, struct config_tree *cft,
 	return 1;
 }
 
-int persistent_filter_load(struct dev_filter *f)
+int persistent_filter_load(struct dev_filter *f, struct config_tree **cft_out)
 {
 	struct pfilter *pf = (struct pfilter *) f->private;
-
-	int r = 0;
 	struct config_tree *cft;
+	struct stat info;
+	int r = 0;
 
-	if (!(cft = create_config_tree(pf->file))) {
-		stack;
-		return 0;
+	if (!stat(pf->file, &info))
+		pf->ctime = info.st_ctime;
+	else {
+		log_very_verbose("%s: stat failed: %s", pf->file,
+				 strerror(errno));
+		return_0;
 	}
 
-	if (!read_config_file(cft)) {
-		stack;
-		goto out;
-	}
+	if (!(cft = create_config_tree(pf->file, 1)))
+		return_0;
+
+	if (!read_config_file(cft))
+		goto_out;
 
 	_read_array(pf, cft, "persistent_filter_cache/valid_devices",
 		    PF_GOOD_DEVICE);
@@ -126,7 +131,10 @@ int persistent_filter_load(struct dev_filter *f)
 	log_very_verbose("Loaded persistent filter cache from %s", pf->file);
 
       out:
-	destroy_config_tree(cft);
+	if (r && cft_out)
+		*cft_out = cft;
+	else
+		destroy_config_tree(cft);
 	return r;
 }
 
@@ -135,6 +143,7 @@ static void _write_array(struct pfilter *pf, FILE *fp, const char *path,
 {
 	void *d;
 	int first = 1;
+	char *buf, *str;
 	struct dm_hash_node *n;
 
 	for (n = dm_hash_get_first(pf->devices); n;
@@ -151,7 +160,13 @@ static void _write_array(struct pfilter *pf, FILE *fp, const char *path,
 			first = 0;
 		}
 
-		fprintf(fp, "\t\t\"%s\"", dm_hash_get_key(pf->devices, n));
+		str = dm_hash_get_key(pf->devices, n);
+		if (!(buf = alloca(escaped_len(str)))) {
+			log_error("persistent filter device path stack "
+				  "allocation failed");
+			return;
+		}
+		fprintf(fp, "\t\t\"%s\"", escape_double_quotes(buf, str));
 	}
 
 	if (!first)
@@ -163,8 +178,12 @@ static void _write_array(struct pfilter *pf, FILE *fp, const char *path,
 int persistent_filter_dump(struct dev_filter *f)
 {
 	struct pfilter *pf = (struct pfilter *) f->private;
-
+	char *tmp_file;
+	struct stat info, info2;
+	struct config_tree *cft = NULL;
 	FILE *fp;
+	int lockfd;
+	int r = 0;
 
 	if (!dm_hash_get_num_entries(pf->devices)) {
 		log_very_verbose("Internal persistent device cache empty "
@@ -179,11 +198,43 @@ int persistent_filter_dump(struct dev_filter *f)
 
 	log_very_verbose("Dumping persistent device cache to %s", pf->file);
 
-	fp = fopen(pf->file, "w");
-	if (!fp) {
-		if (errno != EROFS)
-			log_sys_error("fopen", pf->file);
-		return 0;
+	while (1) {
+		if ((lockfd = fcntl_lock_file(pf->file, F_WRLCK, 0)) < 0)
+			return_0;
+
+		/*
+		 * Ensure we locked the file we expected
+		 */
+		if (fstat(lockfd, &info)) {
+			log_sys_error("fstat", pf->file);
+			goto out;
+		}
+		if (stat(pf->file, &info2)) {
+			log_sys_error("stat", pf->file);
+			goto out;
+		}
+
+		if (is_same_inode(info, info2))
+			break;
+	
+		fcntl_unlock_file(lockfd);
+	}
+
+	/*
+	 * If file contents changed since we loaded it, merge new contents
+	 */
+	if (info.st_ctime != pf->ctime)
+		/* Keep cft open to avoid losing lock */
+		persistent_filter_load(f, &cft);
+
+	tmp_file = alloca(strlen(pf->file) + 5);
+	sprintf(tmp_file, "%s.tmp", pf->file);
+
+	if (!(fp = fopen(tmp_file, "w"))) {
+		/* EACCES has been reported over NFS */
+		if (errno != EROFS && errno != EACCES)
+			log_sys_error("fopen", tmp_file);
+		goto out;
 	}
 
 	fprintf(fp, "# This file is automatically maintained by lvm.\n\n");
@@ -194,8 +245,22 @@ int persistent_filter_dump(struct dev_filter *f)
 	/* _write_array(pf, fp, "invalid_devices", PF_BAD_DEVICE); */
 
 	fprintf(fp, "}\n");
-	fclose(fp);
-	return 1;
+	if (lvm_fclose(fp, tmp_file))
+		goto_out;
+
+	if (rename(tmp_file, pf->file))
+		log_error("%s: rename to %s failed: %s", tmp_file, pf->file,
+			  strerror(errno));
+
+	r = 1;
+
+out:
+	fcntl_unlock_file(lockfd);
+
+	if (cft)
+		destroy_config_tree(cft);
+
+	return r;
 }
 
 static int _lookup_p(struct dev_filter *f, struct device *dev)
@@ -234,16 +299,13 @@ struct dev_filter *persistent_filter_create(struct dev_filter *real,
 	struct pfilter *pf;
 	struct dev_filter *f = NULL;
 
-	if (!(pf = dm_malloc(sizeof(*pf)))) {
-		stack;
-		return NULL;
-	}
+	if (!(pf = dm_malloc(sizeof(*pf))))
+		return_NULL;
 	memset(pf, 0, sizeof(*pf));
 
-	if (!(pf->file = dm_malloc(strlen(file) + 1))) {
-		stack;
-		goto bad;
-	}
+	if (!(pf->file = dm_malloc(strlen(file) + 1)))
+		goto_bad;
+
 	strcpy(pf->file, file);
 	pf->real = real;
 
@@ -252,10 +314,8 @@ struct dev_filter *persistent_filter_create(struct dev_filter *real,
 		goto bad;
 	}
 
-	if (!(f = dm_malloc(sizeof(*f)))) {
-		stack;
-		goto bad;
-	}
+	if (!(f = dm_malloc(sizeof(*f))))
+		goto_bad;
 
 	f->passes_filter = _lookup_p;
 	f->destroy = _persistent_destroy;

lib/filters/filter-persistent.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) 2004-2006 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
  */
@@ -22,7 +22,7 @@ struct dev_filter *persistent_filter_create(struct dev_filter *f,
 					    const char *file);
 
 int persistent_filter_wipe(struct dev_filter *f);
-int persistent_filter_load(struct dev_filter *f);
+int persistent_filter_load(struct dev_filter *f, struct config_tree **cft_out);
 int persistent_filter_dump(struct dev_filter *f);
 
 #endif

lib/filters/filter-regex.c

@@ -1,27 +1,26 @@
 /*
- * 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
  */
 
 #include "lib.h"
 #include "filter-regex.h"
-#include "matcher.h"
 #include "device.h"
 
 struct rfilter {
 	struct dm_pool *mem;
 	dm_bitset_t accept;
-	struct matcher *engine;
+	struct dm_regex *engine;
 };
 
 static int _extract_pattern(struct dm_pool *mem, const char *pat,
@@ -71,10 +70,8 @@ static int _extract_pattern(struct dm_pool *mem, const char *pat,
 	/*
 	 * copy the regex
 	 */
-	if (!(r = dm_pool_strdup(mem, pat))) {
-		stack;
-		return 0;
-	}
+	if (!(r = dm_pool_strdup(mem, pat)))
+		return_0;
 
 	/*
 	 * trim the trailing character, having checked it's sep.
@@ -98,18 +95,15 @@ static int _build_matcher(struct rfilter *rf, struct config_value *val)
 	unsigned count = 0;
 	int i, r = 0;
 
-	if (!(scratch = dm_pool_create("filter matcher", 1024))) {
-		stack;
-		return 0;
-	}
+	if (!(scratch = dm_pool_create("filter dm_regex", 1024)))
+		return_0;
 
 	/*
 	 * count how many patterns we have.
 	 */
 	for (v = val; v; v = v->next) {
-
 		if (v->type != CFG_STRING) {
-			log_info("filter patterns must be enclosed in quotes");
+			log_error("filter patterns must be enclosed in quotes");
 			goto out;
 		}
 
@@ -119,10 +113,8 @@ static int _build_matcher(struct rfilter *rf, struct config_value *val)
 	/*
 	 * allocate space for them
 	 */
-	if (!(regex = dm_pool_alloc(scratch, sizeof(*regex) * count))) {
-		stack;
-		goto out;
-	}
+	if (!(regex = dm_pool_alloc(scratch, sizeof(*regex) * count)))
+		goto_out;
 
 	/*
 	 * create the accept/reject bitset
@@ -136,15 +128,15 @@ static int _build_matcher(struct rfilter *rf, struct config_value *val)
 	 */
 	for (v = val, i = count - 1; v; v = v->next, i--)
 		if (!_extract_pattern(scratch, v->v.str, regex, rf->accept, i)) {
-			log_info("invalid filter pattern");
+			log_error("invalid filter pattern");
 			goto out;
 		}
 
 	/*
 	 * build the matcher.
 	 */
-	if (!(rf->engine = matcher_create(rf->mem, (const char **) regex,
-					  count)))
+	if (!(rf->engine = dm_regex_create(rf->mem, (const char **) regex,
+					   count)))
 		stack;
 	r = 1;
 
@@ -160,17 +152,12 @@ static int _accept_p(struct dev_filter *f, struct device *dev)
 	struct str_list *sl;
 
 	list_iterate_items(sl, &dev->aliases) {
-		m = matcher_run(rf->engine, sl->str);
+		m = dm_regex_match(rf->engine, sl->str);
 
 		if (m >= 0) {
 			if (dm_bit(rf->accept, m)) {
-
-				if (!first) {
-					log_debug("%s: New preferred name",
-						  sl->str);
-					list_del(&sl->list);
-					list_add_h(&dev->aliases, &sl->list);
-				}
+				if (!first)
+					dev_set_preferred_name(sl, dev);
 
 				return 1;
 			}
@@ -203,27 +190,19 @@ struct dev_filter *regex_filter_create(struct config_value *patterns)
 	struct rfilter *rf;
 	struct dev_filter *f;
 
-	if (!mem) {
-		stack;
-		return NULL;
-	}
+	if (!mem)
+		return_NULL;
 
-	if (!(rf = dm_pool_alloc(mem, sizeof(*rf)))) {
-		stack;
-		goto bad;
-	}
+	if (!(rf = dm_pool_alloc(mem, sizeof(*rf))))
+		goto_bad;
 
 	rf->mem = mem;
 
-	if (!_build_matcher(rf, patterns)) {
-		stack;
-		goto bad;
-	}
+	if (!_build_matcher(rf, patterns))
+		goto_bad;
 
-	if (!(f = dm_pool_zalloc(mem, sizeof(*f)))) {
-		stack;
-		goto bad;
-	}
+	if (!(f = dm_pool_zalloc(mem, sizeof(*f))))
+		goto_bad;
 
 	f->passes_filter = _accept_p;
 	f->destroy = _regex_destroy;

lib/filters/filter-regex.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/filters/filter-sysfs.c

@@ -1,13 +1,13 @@
 /*
- * 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
  */
@@ -20,19 +20,21 @@
 
 #include <dirent.h>
 
-static int _locate_sysfs_blocks(const char *proc, char *path, size_t len)
+static int _locate_sysfs_blocks(const char *proc, char *path, size_t len,
+				unsigned *sysfs_depth)
 {
 	char proc_mounts[PATH_MAX];
-	int r = 0;
 	FILE *fp;
 	char *split[4], buffer[PATH_MAX + 16];
+	const char *sys_mnt = NULL;
+	struct stat info;
 
 	if (!*proc) {
 		log_verbose("No proc filesystem found: skipping sysfs filter");
 		return 0;
 	}
 		
-	if (lvm_snprintf(proc_mounts, sizeof(proc_mounts),
+	if (dm_snprintf(proc_mounts, sizeof(proc_mounts),
 			 "%s/mounts", proc) < 0) {
 		log_error("Failed to create /proc/mounts string");
 		return 0;
@@ -44,18 +46,80 @@ static int _locate_sysfs_blocks(const char *proc, char *path, size_t len)
 	}
 
 	while (fgets(buffer, sizeof(buffer), fp)) {
-		if (split_words(buffer, 4, split) == 4 &&
+		if (dm_split_words(buffer, 4, 0, split) == 4 &&
 		    !strcmp(split[2], "sysfs")) {
-			if (lvm_snprintf(path, len, "%s/%s", split[1],
-					 "block") >= 0) {
-				r = 1;
-			}
+			sys_mnt = split[1];
 			break;
 		}
 	}
 
-	fclose(fp);
-	return r;
+	if (fclose(fp))
+		log_sys_error("fclose", proc_mounts);
+
+	if (!sys_mnt) {
+		log_error("Failed to find sysfs mount point");
+		return 0;
+	}
+
+	/*
+	 * unified classification directory for all kernel subsystems
+	 *
+	 * /sys/subsystem/block/devices
+	 * |-- sda -> ../../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda
+	 * |-- sda1 -> ../../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda1
+	 *  `-- sr0 -> ../../../devices/pci0000:00/0000:00:1f.2/host1/target1:0:0/1:0:0:0/block/sr0
+	 *
+	 */
+	if (dm_snprintf(path, len, "%s/%s", sys_mnt,
+			"subsystem/block/devices") >= 0) {
+		if (!stat(path, &info)) {
+			*sysfs_depth = 0;
+			return 1;
+		}
+	}
+
+	/*
+	 * block subsystem as a class
+	 *
+	 * /sys/class/block
+	 * |-- sda -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda
+	 * |-- sda1 -> ../../devices/pci0000:00/0000:00:1f.2/host0/target0:0:0/0:0:0:0/block/sda/sda1
+	 *  `-- sr0 -> ../../devices/pci0000:00/0000:00:1f.2/host1/target1:0:0/1:0:0:0/block/sr0
+	 *
+	 */
+	if (dm_snprintf(path, len, "%s/%s", sys_mnt, "class/block") >= 0) {
+		if (!stat(path, &info)) {
+			*sysfs_depth = 0;
+			return 1;
+		}
+	}
+
+	/*
+	 * old block subsystem layout with nested directories
+	 *
+	 * /sys/block/
+	 * |-- sda
+	 * |   |-- capability
+	 * |   |-- dev
+	 * ...
+	 * |   |-- sda1
+	 * |   |   |-- dev
+	 * ...
+	 * |
+	 * `-- sr0
+	 *     |-- capability
+	 *     |-- dev
+	 * ...
+	 *
+	 */
+	if (dm_snprintf(path, len, "%s/%s", sys_mnt, "block") >= 0) {
+		if (!stat(path, &info)) {
+			*sysfs_depth = 1;
+			return 1;
+		}
+	}
+
+	return 0;
 }
 
 /*----------------------------------------------------------------
@@ -70,11 +134,14 @@ struct entry {
 struct dev_set {
 	struct dm_pool *mem;
 	const char *sys_block;
+	unsigned sysfs_depth;
 	int initialised;
 	struct entry *slots[SET_BUCKETS];
 };
 
-static struct dev_set *_dev_set_create(struct dm_pool *mem, const char *sys_block)
+static struct dev_set *_dev_set_create(struct dm_pool *mem,
+				       const char *sys_block,
+				       unsigned sysfs_depth)
 {
 	struct dev_set *ds;
 
@@ -83,6 +150,7 @@ static struct dev_set *_dev_set_create(struct dm_pool *mem, const char *sys_bloc
 
 	ds->mem = mem;
 	ds->sys_block = dm_pool_strdup(mem, sys_block);
+	ds->sysfs_depth = sysfs_depth;
 	ds->initialised = 0;
 
 	return ds;
@@ -156,7 +224,9 @@ static int _read_dev(const char *file, dev_t *result)
 	}
 
 	r = _parse_dev(file, fp, result);
-	fclose(fp);
+
+	if (fclose(fp))
+		log_sys_error("fclose", file);
 
 	return r;
 }
@@ -164,68 +234,59 @@ static int _read_dev(const char *file, dev_t *result)
 /*
  * Recurse through sysfs directories, inserting any devs found.
  */
-static int _read_devs(struct dev_set *ds, const char *dir)
+static int _read_devs(struct dev_set *ds, const char *dir, unsigned sysfs_depth)
 {
-        struct dirent *d;
-        DIR *dr;
-	unsigned char dtype;
+	struct dirent *d;
+	DIR *dr;
 	struct stat info;
 	char path[PATH_MAX];
+	char file[PATH_MAX];
 	dev_t dev = { 0 };
 	int r = 1;
 
-        if (!(dr = opendir(dir))) {
-                log_sys_error("opendir", dir);
-                return 0;
-        }
+	if (!(dr = opendir(dir))) {
+		log_sys_error("opendir", dir);
+		return 0;
+	}
 
-        while ((d = readdir(dr))) {
-                if (!strcmp(d->d_name, ".") || !strcmp(d->d_name, ".."))
+	while ((d = readdir(dr))) {
+		if (!strcmp(d->d_name, ".") || !strcmp(d->d_name, ".."))
 			continue;
 
-		if (lvm_snprintf(path, sizeof(path), "%s/%s", dir,
+		if (dm_snprintf(path, sizeof(path), "%s/%s", dir,
 				 d->d_name) < 0) {
 			log_error("sysfs path name too long: %s in %s",
 				  d->d_name, dir);
 			continue;
 		}
 
-		dtype = d->d_type;
-
-		if (dtype == DT_UNKNOWN) {
-			if (lstat(path, &info) >= 0) {
-				if (S_ISLNK(info.st_mode))
-					dtype = DT_LNK;
-				else if (S_ISDIR(info.st_mode))
-					dtype = DT_DIR;
-				else if (S_ISREG(info.st_mode))
-					dtype = DT_REG;
-			}
+		/* devices have a "dev" file */
+		if (dm_snprintf(file, sizeof(file), "%s/dev", path) < 0) {
+			log_error("sysfs path name too long: %s in %s",
+				  d->d_name, dir);
+			continue;
 		}
 
-		if (dtype == DT_DIR) {
-			if (!_read_devs(ds, path)) {
-				r = 0;
-				break;
-			}
-		}
+		if (!stat(file, &info)) {
+			/* recurse if we found a device and expect subdirs */
+			if (sysfs_depth)
+				_read_devs(ds, path, sysfs_depth - 1);
 
-		if ((dtype == DT_REG && !strcmp(d->d_name, "dev")))
-			if (!_read_dev(path, &dev) || !_set_insert(ds, dev)) {
-				r = 0;
-				break;
-			}
+			/* add the device we have found */
+			if (_read_dev(file, &dev))
+				_set_insert(ds, dev);
+		}
 	}
 
-        if (closedir(dr))
-                log_sys_error("closedir", dir);
+	if (closedir(dr))
+		log_sys_error("closedir", dir);
 
 	return r;
 }
 
 static int _init_devs(struct dev_set *ds)
 {
-	if (!_read_devs(ds, ds->sys_block)) {
+	if (!_read_devs(ds, ds->sys_block, ds->sysfs_depth)) {
 		ds->initialised = -1;
 		return 0;
 	}
@@ -263,11 +324,12 @@ static void _destroy(struct dev_filter *f)
 struct dev_filter *sysfs_filter_create(const char *proc)
 {
 	char sys_block[PATH_MAX];
+	unsigned sysfs_depth;
 	struct dm_pool *mem;
 	struct dev_set *ds;
 	struct dev_filter *f;
 
-	if (!_locate_sysfs_blocks(proc, sys_block, sizeof(sys_block)))
+	if (!_locate_sysfs_blocks(proc, sys_block, sizeof(sys_block), &sysfs_depth))
 		return NULL;
 
 	if (!(mem = dm_pool_create("sysfs", 256))) {
@@ -275,15 +337,13 @@ struct dev_filter *sysfs_filter_create(const char *proc)
 		return NULL;
 	}
 
-	if (!(ds = _dev_set_create(mem, sys_block))) {
+	if (!(ds = _dev_set_create(mem, sys_block, sysfs_depth))) {
 		log_error("sysfs dev_set creation failed");
 		goto bad;
 	}
 
-	if (!(f = dm_pool_zalloc(mem, sizeof(*f)))) {
-		stack;
-		goto bad;
-	}
+	if (!(f = dm_pool_zalloc(mem, sizeof(*f))))
+		goto_bad;
 
 	f->passes_filter = _accept_p;
 	f->destroy = _destroy;
@@ -297,7 +357,7 @@ struct dev_filter *sysfs_filter_create(const char *proc)
 
 #else
 
-struct dev_filter *sysfs_filter_create(const char *proc)
+struct dev_filter *sysfs_filter_create(const char *proc __attribute((unused)))
 {
 	return NULL;
 }

lib/filters/filter-sysfs.h

@@ -5,9 +5,9 @@
  *
  * 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/filters/filter.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) 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 +19,7 @@
 #include "lvm-string.h"
 #include "config.h"
 #include "metadata.h"
+#include "activate.h"
 
 #include <dirent.h>
 #include <unistd.h>
@@ -37,6 +38,7 @@ typedef struct {
 } device_info_t;
 
 static int _md_major = -1;
+static int _device_mapper_major = -1;
 
 int md_major(void)
 {
@@ -73,10 +75,13 @@ static const device_info_t device_info[] = {
 	{"aoe", 16},		/* ATA over Ethernet */
 	{"device-mapper", 1},	/* Other mapped devices */
 	{"xvd", 16},		/* Xen virtual block device */
+	{"vdisk", 8},		/* SUN's LDOM virtual block device */
+	{"ps3disk", 16},	/* PlayStation 3 internal disk */
+	{"virtblk", 8},		/* VirtIO disk */
 	{NULL, 0}
 };
 
-static int _passes_lvm_type_device_filter(struct dev_filter *f,
+static int _passes_lvm_type_device_filter(struct dev_filter *f __attribute((unused)),
 					  struct device *dev)
 {
 	const char *name = dev_name(dev);
@@ -90,6 +95,13 @@ static int _passes_lvm_type_device_filter(struct dev_filter *f,
 		return 0;
 	}
 
+	/* Skip suspended devices */
+	if (MAJOR(dev->dev) == _device_mapper_major &&
+	    ignore_suspended_devices() && !device_is_usable(dev->dev)) {
+		log_debug("%s: Skipping: Suspended dm device", name);
+		return 0;
+	}
+
 	/* Check it's accessible */
 	if (!dev_open_flags(dev, O_RDONLY, 0, 1)) {
 		log_debug("%s: Skipping: open failed", name);
@@ -145,7 +157,7 @@ static int _scan_proc_dev(const char *proc, const struct config_node *cn)
 	/* All types unrecognised initially */
 	memset(_max_partitions_by_major, 0, sizeof(int) * NUMBER_OF_MAJORS);
 
-	if (lvm_snprintf(proc_devices, sizeof(proc_devices),
+	if (dm_snprintf(proc_devices, sizeof(proc_devices),
 			 "%s/devices", proc) < 0) {
 		log_error("Failed to create /proc/devices string");
 		return 0;
@@ -182,10 +194,14 @@ static int _scan_proc_dev(const char *proc, const struct config_node *cn)
 		if (!strncmp("md", line + i, 2) && isspace(*(line + i + 2)))
 			_md_major = line_maj;
 
+		/* Look for device-mapper device */
+		/* FIXME Cope with multiple majors */
+		if (!strncmp("device-mapper", line + i, 13) && isspace(*(line + i + 13)))
+			_device_mapper_major = line_maj;
+
 		/* Go through the valid device names and if there is a
 		   match store max number of partitions */
 		for (j = 0; device_info[j].name != NULL; j++) {
-
 			dev_len = strlen(device_info[j].name);
 			if (dev_len <= strlen(line + i) &&
 			    !strncmp(device_info[j].name, line + i, dev_len) &&
@@ -204,7 +220,8 @@ static int _scan_proc_dev(const char *proc, const struct config_node *cn)
 			if (cv->type != CFG_STRING) {
 				log_error("Expecting string in devices/types "
 					  "in config file");
-				fclose(pd);
+				if (fclose(pd))
+					log_sys_error("fclose", proc_devices);
 				return 0;
 			}
 			dev_len = strlen(cv->v.str);
@@ -214,14 +231,16 @@ static int _scan_proc_dev(const char *proc, const struct config_node *cn)
 				log_error("Max partition count missing for %s "
 					  "in devices/types in config file",
 					  name);
-				fclose(pd);
+				if (fclose(pd))
+					log_sys_error("fclose", proc_devices);
 				return 0;
 			}
 			if (!cv->v.i) {
 				log_error("Zero partition count invalid for "
 					  "%s in devices/types in config file",
 					  name);
-				fclose(pd);
+				if (fclose(pd))
+					log_sys_error("fclose", proc_devices);
 				return 0;
 			}
 			if (dev_len <= strlen(line + i) &&
@@ -232,7 +251,10 @@ static int _scan_proc_dev(const char *proc, const struct config_node *cn)
 			}
 		}
 	}
-	fclose(pd);
+
+	if (fclose(pd))
+		log_sys_error("fclose", proc_devices);
+
 	return 1;
 }
 
@@ -256,9 +278,8 @@ struct dev_filter *lvm_type_filter_create(const char *proc,
 	f->private = NULL;
 
 	if (!_scan_proc_dev(proc, cn)) {
-		stack;
 		dm_free(f);
-		return NULL;
+		return_NULL;
 	}
 
 	return f;

lib/filters/filter.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
  */

lib/format1/Makefile.in

@@ -2,7 +2,7 @@
 # Copyright (C) 2002-2004 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

lib/format1/disk-rep.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
  */
@@ -21,7 +21,6 @@
 
 #include <fcntl.h>
 
-#define fail do {stack; return 0;} while(0)
 #define xx16(v) disk->v = xlate16(disk->v)
 #define xx32(v) disk->v = xlate32(disk->v)
 #define xx64(v) disk->v = xlate64(disk->v)
@@ -134,15 +133,15 @@ static int _munge_formats(struct pv_disk *pvd)
 		return 0;
 	}
 
-        /* UUID too long? */
-        if (pvd->pv_uuid[ID_LEN]) {
+	/* UUID too long? */
+	if (pvd->pv_uuid[ID_LEN]) {
 		/* Retain ID_LEN chars from end */
-                for (e = ID_LEN; e < sizeof(pvd->pv_uuid); e++) {
-                        if (!pvd->pv_uuid[e]) {
-                                e--;
-                                break;
-                        }
-                }
+		for (e = ID_LEN; e < sizeof(pvd->pv_uuid); e++) {
+			if (!pvd->pv_uuid[e]) {
+				e--;
+				break;
+			}
+		}
 		for (b = 0; b < ID_LEN; b++) {
 			pvd->pv_uuid[b] = pvd->pv_uuid[++e - ID_LEN];
 			/* FIXME Remove all invalid chars */
@@ -150,7 +149,7 @@ static int _munge_formats(struct pv_disk *pvd)
 				pvd->pv_uuid[b] = '#';
 		}
 		memset(&pvd->pv_uuid[ID_LEN], 0, sizeof(pvd->pv_uuid) - ID_LEN);
-        }
+	}
 
 	/* If UUID is missing, create one */
 	if (pvd->pv_uuid[0] == '\0') {
@@ -161,8 +160,8 @@ static int _munge_formats(struct pv_disk *pvd)
 	return 1;
 }
 
-/* 
- * If exported, remove "PV_EXP" from end of VG name 
+/*
+ * If exported, remove "PV_EXP" from end of VG name
  */
 static void _munge_exported_vg(struct pv_disk *pvd)
 {
@@ -178,8 +177,8 @@ static void _munge_exported_vg(struct pv_disk *pvd)
 	s = sizeof(EXPORTED_TAG);
 	if (!strncmp((char *)pvd->vg_name + l - s + 1, EXPORTED_TAG, s)) {
 		pvd->vg_name[l - s + 1] = '\0';
-                pvd->pv_status |= VG_EXPORTED;
-        }
+		pvd->pv_status |= VG_EXPORTED;
+	}
 }
 
 int munge_pvd(struct device *dev, struct pv_disk *pvd)
@@ -218,7 +217,7 @@ static int _read_pvd(struct device *dev, struct pv_disk *pvd)
 static int _read_lvd(struct device *dev, uint64_t pos, struct lv_disk *disk)
 {
 	if (!dev_read(dev, pos, sizeof(*disk), disk))
-		fail;
+		return_0;
 
 	_xlate_lvd(disk);
 
@@ -230,12 +229,12 @@ int read_vgd(struct device *dev, struct vg_disk *vgd, struct pv_disk *pvd)
 	uint64_t pos = pvd->vg_on_disk.base;
 
 	if (!dev_read(dev, pos, sizeof(*vgd), vgd))
-		fail;
+		return_0;
 
 	_xlate_vgd(vgd);
 
 	if ((vgd->lv_max > MAX_LV) || (vgd->pv_max > MAX_PV))
-		fail;
+		return_0;
 		
 	/* If UUID is missing, create one */
 	if (vgd->vg_uuid[0] == '\0')
@@ -248,16 +247,16 @@ static int _read_uuids(struct disk_list *data)
 {
 	unsigned num_read = 0;
 	struct uuid_list *ul;
-	char buffer[NAME_LEN];
+	char buffer[NAME_LEN] __attribute((aligned(8)));
 	uint64_t pos = data->pvd.pv_uuidlist_on_disk.base;
 	uint64_t end = pos + data->pvd.pv_uuidlist_on_disk.size;
 
 	while (pos < end && num_read < data->vgd.pv_cur) {
 		if (!dev_read(data->dev, pos, sizeof(buffer), buffer))
-			fail;
+			return_0;
 
 		if (!(ul = dm_pool_alloc(data->mem, sizeof(*ul))))
-			fail;
+			return_0;
 
 		memcpy(ul->uuid, buffer, NAME_LEN);
 		ul->uuid[NAME_LEN - 1] = '\0';
@@ -288,10 +287,10 @@ static int _read_lvs(struct disk_list *data)
 		ll = dm_pool_alloc(data->mem, sizeof(*ll));
 
 		if (!ll)
-			fail;
+			return_0;
 
 		if (!_read_lvd(data->dev, pos, &ll->lvd))
-			fail;
+			return_0;
 
 		if (!_check_lvd(&ll->lvd))
 			continue;
@@ -310,10 +309,10 @@ static int _read_extents(struct disk_list *data)
 	uint64_t pos = data->pvd.pe_on_disk.base;
 
 	if (!extents)
-		fail;
+		return_0;
 
 	if (!dev_read(data->dev, pos, len, extents))
-		fail;
+		return_0;
 
 	_xlate_extents(extents, data->pvd.pe_total);
 	data->extents = extents;
@@ -327,10 +326,11 @@ static void __update_lvmcache(const struct format_type *fmt,
 			      unsigned exported)
 {
 	struct lvmcache_info *info;
+	const char *vgname = *((char *)dl->pvd.vg_name) ?
+			     (char *)dl->pvd.vg_name : fmt->orphan_vg_name;
 
 	if (!(info = lvmcache_add(fmt->labeller, (char *)dl->pvd.pv_uuid, dev,
-				  (char *)dl->pvd.vg_name, vgid,
-				  exported ? EXPORTED_VG : 0))) {
+				  vgname, vgid, exported ? EXPORTED_VG : 0))) {
 		stack;
 		return;
 	}
@@ -347,20 +347,16 @@ static struct disk_list *__read_disk(const struct format_type *fmt,
 	struct disk_list *dl = dm_pool_zalloc(mem, sizeof(*dl));
 	const char *name = dev_name(dev);
 
-	if (!dl) {
-		stack;
-		return NULL;
-	}
+	if (!dl)
+		return_NULL;
 
 	dl->dev = dev;
 	dl->mem = mem;
 	list_init(&dl->uuids);
 	list_init(&dl->lvds);
 
-	if (!_read_pvd(dev, &dl->pvd)) {
-		stack;
-		goto bad;
-	}
+	if (!_read_pvd(dev, &dl->pvd))
+		goto_bad;
 
 	/*
 	 * is it an orphan ?
@@ -368,20 +364,20 @@ static struct disk_list *__read_disk(const struct format_type *fmt,
 	if (!*dl->pvd.vg_name) {
 		log_very_verbose("%s is not a member of any format1 VG", name);
 
-		__update_lvmcache(fmt, dl, dev, NULL, 0);
+		__update_lvmcache(fmt, dl, dev, fmt->orphan_vg_name, 0);
 		return (vg_name) ? NULL : dl;
 	}
 
 	if (!read_vgd(dl->dev, &dl->vgd, &dl->pvd)) {
 		log_error("Failed to read VG data from PV (%s)", name);
-		__update_lvmcache(fmt, dl, dev, NULL, 0);
+		__update_lvmcache(fmt, dl, dev, fmt->orphan_vg_name, 0);
 		goto bad;
 	}
 
 	if (vg_name && strcmp(vg_name, (char *)dl->pvd.vg_name)) {
 		log_very_verbose("%s is not a member of the VG %s",
 				 name, vg_name);
-		__update_lvmcache(fmt, dl, dev, NULL, 0);
+		__update_lvmcache(fmt, dl, dev, fmt->orphan_vg_name, 0);
 		goto bad;
 	}
 
@@ -417,19 +413,17 @@ static struct disk_list *__read_disk(const struct format_type *fmt,
 struct disk_list *read_disk(const struct format_type *fmt, struct device *dev,
 			    struct dm_pool *mem, const char *vg_name)
 {
-	struct disk_list *r;
+	struct disk_list *dl;
 
-	if (!dev_open(dev)) {
-		stack;
-		return NULL;
-	}
+	if (!dev_open(dev))
+		return_NULL;
 
-	r = __read_disk(fmt, dev, mem, vg_name);
+	dl = __read_disk(fmt, dev, mem, vg_name);
 
 	if (!dev_close(dev))
 		stack;
 
-	return r;
+	return dl;
 }
 
 static void _add_pv_to_list(struct list *head, struct disk_list *data)
@@ -482,7 +476,7 @@ int read_pvs_in_vg(const struct format_type *fmt, const char *vg_name,
 		}
 
 		/* Did we find the whole VG? */
-		if (!vg_name || !*vg_name ||
+		if (!vg_name || is_orphan_vg(vg_name) ||
 		    (data && *data->pvd.vg_name &&
 		     list_size(head) == data->vgd.pv_cur))
 			return 1;
@@ -521,7 +515,7 @@ static int _write_vgd(struct disk_list *data)
 
 	_xlate_vgd(vgd);
 	if (!dev_write(data->dev, pos, sizeof(*vgd), vgd))
-		fail;
+		return_0;
 
 	_xlate_vgd(vgd);
 
@@ -546,7 +540,7 @@ static int _write_uuids(struct disk_list *data)
 			  pos, NAME_LEN);
 
 		if (!dev_write(data->dev, pos, NAME_LEN, ul->uuid))
-			fail;
+			return_0;
 
 		pos += NAME_LEN;
 	}
@@ -562,7 +556,7 @@ static int _write_lvd(struct device *dev, uint64_t pos, struct lv_disk *disk)
 
 	_xlate_lvd(disk);
 	if (!dev_write(dev, pos, sizeof(*disk), disk))
-		fail;
+		return_0;
 
 	_xlate_lvd(disk);
 
@@ -590,7 +584,7 @@ static int _write_lvs(struct disk_list *data)
 		}
 
 		if (!_write_lvd(data->dev, pos + offset, &ll->lvd))
-			fail;
+			return_0;
 	}
 
 	return 1;
@@ -608,7 +602,7 @@ static int _write_extents(struct disk_list *data)
 
 	_xlate_extents(extents, data->pvd.pe_total);
 	if (!dev_write(data->dev, pos, len, extents))
-		fail;
+		return_0;
 
 	_xlate_extents(extents, data->pvd.pe_total);
 
@@ -645,7 +639,7 @@ static int _write_pvd(struct disk_list *data)
 	_xlate_pvd((struct pv_disk *) buf);
 	if (!dev_write(data->dev, pos, size, buf)) {
 		dm_free(buf);
-		fail;
+		return_0;
 	}
 
 	dm_free(buf);
@@ -655,7 +649,7 @@ static int _write_pvd(struct disk_list *data)
 /*
  * assumes the device has been opened.
  */
-static int __write_all_pvd(const struct format_type *fmt,
+static int __write_all_pvd(const struct format_type *fmt __attribute((unused)),
 			   struct disk_list *data)
 {
 	const char *pv_name = dev_name(data->dev);
@@ -709,10 +703,8 @@ static int _write_all_pvd(const struct format_type *fmt, struct disk_list *data)
 {
 	int r;
 
-	if (!dev_open(data->dev)) {
-		stack;
-		return 0;
-	}
+	if (!dev_open(data->dev))
+		return_0;
 
 	r = __write_all_pvd(fmt, data);
 
@@ -733,7 +725,7 @@ int write_disks(const struct format_type *fmt, struct list *pvs)
 
 	list_iterate_items(dl, pvs) {
 		if (!(_write_all_pvd(fmt, dl)))
-			fail;
+			return_0;
 
 		log_very_verbose("Successfully wrote data to %s",
 				 dev_name(dl->dev));

lib/format1/disk-rep.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) 2004-2006 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
  */
@@ -148,7 +148,7 @@ struct pe_disk {
 
 struct uuid_list {
 	struct list list;
-	char uuid[NAME_LEN];
+	char uuid[NAME_LEN] __attribute((aligned(8)));
 };
 
 struct lvd_list {
@@ -161,17 +161,18 @@ struct disk_list {
 	struct dm_pool *mem;
 	struct device *dev;
 
-	struct pv_disk pvd;
-	struct vg_disk vgd;
-	struct list uuids;
-	struct list lvds;
-	struct pe_disk *extents;
+	struct pv_disk pvd __attribute((aligned(8)));
+	struct vg_disk vgd __attribute((aligned(8)));
+	struct list uuids __attribute((aligned(8)));
+	struct list lvds __attribute((aligned(8)));
+	struct pe_disk *extents __attribute((aligned(8)));
 };
 
 /*
  * Layout constants.
  */
 #define METADATA_ALIGN 4096UL
+#define LVM1_PE_ALIGN (65536UL >> SECTOR_SHIFT)      /* PE alignment */
 
 #define	METADATA_BASE 0UL
 #define	PV_SIZE 1024UL
@@ -202,8 +203,8 @@ int write_disks(const struct format_type *fmt, struct list *pvds);
  * Functions to translate to between disk and in
  * core structures.
  */
-int import_pv(struct dm_pool *mem, struct device *dev,
-	      struct volume_group *vg,
+int import_pv(const struct format_type *fmt, struct dm_pool *mem,
+	      struct device *dev, struct volume_group *vg,
 	      struct physical_volume *pv, struct pv_disk *pvd,
 	      struct vg_disk *vgd);
 int export_pv(struct cmd_context *cmd, struct dm_pool *mem,

lib/format1/format1.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
  */
@@ -18,13 +18,10 @@
 #include "limits.h"
 #include "display.h"
 #include "toolcontext.h"
-#include "lvmcache.h"
 #include "lvm1-label.h"
 #include "format1.h"
 #include "segtype.h"
 
-#define FMT_LVM1_NAME "lvm1"
-
 /* VG consistency checks */
 static int _check_vgs(struct list *pvs, int *partial)
 {
@@ -131,10 +128,10 @@ static struct volume_group *_build_vg(struct format_instance *fid,
 	int partial;
 
 	if (!vg)
-		goto bad;
+		goto_bad;
 
 	if (list_empty(pvs))
-		goto bad;
+		goto_bad;
 
 	memset(vg, 0, sizeof(*vg));
 
@@ -146,60 +143,53 @@ static struct volume_group *_build_vg(struct format_instance *fid,
 	list_init(&vg->tags);
 
 	if (!_check_vgs(pvs, &partial))
-		goto bad;
+		goto_bad;
 
 	dl = list_item(pvs->n, struct disk_list);
 
 	if (!import_vg(mem, vg, dl, partial))
-		goto bad;
+		goto_bad;
 
 	if (!import_pvs(fid->fmt, mem, vg, pvs, &vg->pvs, &vg->pv_count))
-		goto bad;
+		goto_bad;
 
 	if (!import_lvs(mem, vg, pvs))
-		goto bad;
+		goto_bad;
 
 	if (!import_extents(fid->fmt->cmd, vg, pvs))
-		goto bad;
+		goto_bad;
 
 	if (!import_snapshots(mem, vg, pvs))
-		goto bad;
+		goto_bad;
 
 	return vg;
 
       bad:
-	stack;
 	dm_pool_free(mem, vg);
 	return NULL;
 }
 
 static struct volume_group *_format1_vg_read(struct format_instance *fid,
 				     const char *vg_name,
-				     struct metadata_area *mda)
+				     struct metadata_area *mda __attribute((unused)))
 {
 	struct dm_pool *mem = dm_pool_create("lvm1 vg_read", 1024 * 10);
 	struct list pvs;
 	struct volume_group *vg = NULL;
 	list_init(&pvs);
 
-	if (!mem) {
-		stack;
-		return NULL;
-	}
+	if (!mem)
+		return_NULL;
 
 	/* Strip dev_dir if present */
 	vg_name = strip_dir(vg_name, fid->fmt->cmd->dev_dir);
 
 	if (!read_pvs_in_vg
-	    (fid->fmt, vg_name, fid->fmt->cmd->filter, mem, &pvs)) {
-		stack;
-		goto bad;
-	}
+	    (fid->fmt, vg_name, fid->fmt->cmd->filter, mem, &pvs))
+		goto_bad;
 
-	if (!(vg = _build_vg(fid, &pvs))) {
-		stack;
-		goto bad;
-	}
+	if (!(vg = _build_vg(fid, &pvs)))
+		goto_bad;
 
       bad:
 	dm_pool_destroy(mem);
@@ -213,10 +203,8 @@ static struct disk_list *_flatten_pv(struct format_instance *fid,
 {
 	struct disk_list *dl = dm_pool_alloc(mem, sizeof(*dl));
 
-	if (!dl) {
-		stack;
-		return NULL;
-	}
+	if (!dl)
+		return_NULL;
 
 	dl->mem = mem;
 	dl->dev = pv->dev;
@@ -228,9 +216,8 @@ static struct disk_list *_flatten_pv(struct format_instance *fid,
 	    !export_vg(&dl->vgd, vg) ||
 	    !export_uuids(dl, vg) ||
 	    !export_lvs(dl, vg, pv, dev_dir) || !calculate_layout(dl)) {
-		stack;
 		dm_pool_free(mem, dl);
-		return NULL;
+		return_NULL;
 	}
 
 	return dl;
@@ -245,10 +232,8 @@ static int _flatten_vg(struct format_instance *fid, struct dm_pool *mem,
 	struct disk_list *data;
 
 	list_iterate_items(pvl, &vg->pvs) {
-		if (!(data = _flatten_pv(fid, mem, vg, pvl->pv, dev_dir))) {
-			stack;
-			return 0;
-		}
+		if (!(data = _flatten_pv(fid, mem, vg, pvl->pv, dev_dir)))
+			return_0;
 
 		list_add(pvds, &data->list);
 	}
@@ -256,25 +241,21 @@ static int _flatten_vg(struct format_instance *fid, struct dm_pool *mem,
 	export_numbers(pvds, vg);
 	export_pv_act(pvds);
 
-	if (!export_vg_number(fid, pvds, vg->name, filter)) {
-		stack;
-		return 0;
-	}
+	if (!export_vg_number(fid, pvds, vg->name, filter))
+		return_0;
 
 	return 1;
 }
 
 static int _format1_vg_write(struct format_instance *fid, struct volume_group *vg,
-		     struct metadata_area *mda)
+		     struct metadata_area *mda __attribute((unused)))
 {
 	struct dm_pool *mem = dm_pool_create("lvm1 vg_write", 1024 * 10);
 	struct list pvds;
 	int r = 0;
 
-	if (!mem) {
-		stack;
-		return 0;
-	}
+	if (!mem)
+		return_0;
 
 	list_init(&pvds);
 
@@ -282,13 +263,13 @@ static int _format1_vg_write(struct format_instance *fid, struct volume_group *v
 			 fid->fmt->cmd->filter) &&
 	     write_disks(fid->fmt, &pvds));
 
-	lvmcache_update_vg(vg);
+	lvmcache_update_vg(vg, 0);
 	dm_pool_destroy(mem);
 	return r;
 }
 
 static int _format1_pv_read(const struct format_type *fmt, const char *pv_name,
-		    struct physical_volume *pv, struct list *mdas)
+		    struct physical_volume *pv, struct list *mdas __attribute((unused)))
 {
 	struct dm_pool *mem = dm_pool_create("lvm1 pv_read", 1024);
 	struct disk_list *dl;
@@ -297,25 +278,17 @@ static int _format1_pv_read(const struct format_type *fmt, const char *pv_name,
 
 	log_very_verbose("Reading physical volume data %s from disk", pv_name);
 
-	if (!mem) {
-		stack;
-		return 0;
-	}
+	if (!mem)
+		return_0;
 
-	if (!(dev = dev_cache_get(pv_name, fmt->cmd->filter))) {
-		stack;
-		goto out;
-	}
+	if (!(dev = dev_cache_get(pv_name, fmt->cmd->filter)))
+		goto_out;
 
-	if (!(dl = read_disk(fmt, dev, mem, NULL))) {
-		stack;
-		goto out;
-	}
+	if (!(dl = read_disk(fmt, dev, mem, NULL)))
+		goto_out;
 
-	if (!import_pv(fmt->cmd->mem, dl->dev, NULL, pv, &dl->pvd, &dl->vgd)) {
-		stack;
-		goto out;
-	}
+	if (!import_pv(fmt, fmt->cmd->mem, dl->dev, NULL, pv, &dl->pvd, &dl->vgd))
+		goto_out;
 
 	pv->fmt = fmt;
 
@@ -329,9 +302,9 @@ static int _format1_pv_read(const struct format_type *fmt, const char *pv_name,
 static int _format1_pv_setup(const struct format_type *fmt,
 		     uint64_t pe_start, uint32_t extent_count,
 		     uint32_t extent_size,
-		     int pvmetadatacopies,
-		     uint64_t pvmetadatasize, struct list *mdas,
-		     struct physical_volume *pv, struct volume_group *vg)
+		     int pvmetadatacopies __attribute((unused)),
+		     uint64_t pvmetadatasize __attribute((unused)), struct list *mdas __attribute((unused)),
+		     struct physical_volume *pv, struct volume_group *vg __attribute((unused)))
 {
 	if (pv->size > MAX_PV_SIZE)
 		pv->size--;
@@ -348,10 +321,8 @@ static int _format1_pv_setup(const struct format_type *fmt,
 	/*
 	 * This works out pe_start and pe_count.
 	 */
-	if (!calculate_extent_count(pv, extent_size, extent_count, pe_start)) {
-		stack;
-		return 0;
-	}
+	if (!calculate_extent_count(pv, extent_size, extent_count, pe_start))
+		return_0;
 
 	/* Retain existing extent locations exactly */
 	if (((pe_start || extent_count) && (pe_start != pv->pe_start)) ||
@@ -385,7 +356,7 @@ static int _format1_lv_setup(struct format_instance *fid, struct logical_volume
 }
 
 static int _format1_pv_write(const struct format_type *fmt, struct physical_volume *pv,
-		     struct list *mdas, int64_t sector)
+		     struct list *mdas __attribute((unused)), int64_t sector __attribute((unused)))
 {
 	struct dm_pool *mem;
 	struct disk_list *dl;
@@ -394,10 +365,8 @@ static int _format1_pv_write(const struct format_type *fmt, struct physical_volu
 	struct lvmcache_info *info;
 
 	if (!(info = lvmcache_add(fmt->labeller, (char *) &pv->id, pv->dev,
-				  pv->vg_name, NULL, 0))) {
-		stack;
-		return 0;
-	}
+				  pv->vg_name, NULL, 0)))
+		return_0;
 	label = info->label;
 	info->device_size = pv->size << SECTOR_SHIFT;
 	info->fmt = fmt;
@@ -408,36 +377,29 @@ static int _format1_pv_write(const struct format_type *fmt, struct physical_volu
 
 	/* Ensure any residual PE structure is gone */
 	pv->pe_size = pv->pe_count = 0;
-	pv->pe_start = PE_ALIGN;
+	pv->pe_start = LVM1_PE_ALIGN;
 
-	if (!(mem = dm_pool_create("lvm1 pv_write", 1024))) {
-		stack;
-		return 0;
-	}
+	if (!(mem = dm_pool_create("lvm1 pv_write", 1024)))
+		return_0;
+
+	if (!(dl = dm_pool_alloc(mem, sizeof(*dl))))
+		goto_bad;
 
-	if (!(dl = dm_pool_alloc(mem, sizeof(*dl)))) {
-		stack;
-		goto bad;
-	}
 	dl->mem = mem;
 	dl->dev = pv->dev;
 
-	if (!export_pv(fmt->cmd, mem, NULL, &dl->pvd, pv)) {
-		stack;
-		goto bad;
-	}
+	if (!export_pv(fmt->cmd, mem, NULL, &dl->pvd, pv))
+		goto_bad;
 
 	/* must be set to be able to zero gap after PV structure in
 	   dev_write in order to make other disk tools happy */
 	dl->pvd.pv_on_disk.base = METADATA_BASE;
 	dl->pvd.pv_on_disk.size = PV_SIZE;
-	dl->pvd.pe_on_disk.base = PE_ALIGN << SECTOR_SHIFT;
+	dl->pvd.pe_on_disk.base = LVM1_PE_ALIGN << SECTOR_SHIFT;
 
 	list_add(&pvs, &dl->list);
-	if (!write_disks(fmt, &pvs)) {
-		stack;
-		goto bad;
-	}
+	if (!write_disks(fmt, &pvs))
+		goto_bad;
 
 	dm_pool_destroy(mem);
 	return 1;
@@ -479,13 +441,11 @@ static int _format1_vg_setup(struct format_instance *fid, struct volume_group *v
 	return 1;
 }
 
-static int _format1_segtype_supported(struct format_instance *fid, 
+static int _format1_segtype_supported(struct format_instance *fid __attribute((unused)),
 				      const struct segment_type *segtype)
 {
-	if (!(segtype->flags & SEG_FORMAT1_SUPPORT)) {
-		stack;
-		return 0;
-	}
+	if (!(segtype->flags & SEG_FORMAT1_SUPPORT))
+		return_0;
 
 	return 1;
 }
@@ -496,26 +456,23 @@ static struct metadata_area_ops _metadata_format1_ops = {
 };
 
 static struct format_instance *_format1_create_instance(const struct format_type *fmt,
-						const char *vgname,
-						const char *vgid,
-						void *private)
+						const char *vgname __attribute((unused)),
+						const char *vgid __attribute((unused)),
+						void *private __attribute((unused)))
 {
 	struct format_instance *fid;
 	struct metadata_area *mda;
 
-	if (!(fid = dm_pool_alloc(fmt->cmd->mem, sizeof(*fid)))) {
-		stack;
-		return NULL;
-	}
+	if (!(fid = dm_pool_alloc(fmt->cmd->mem, sizeof(*fid))))
+		return_NULL;
 
 	fid->fmt = fmt;
 	list_init(&fid->metadata_areas);
 
 	/* Define a NULL metadata area */
 	if (!(mda = dm_pool_alloc(fmt->cmd->mem, sizeof(*mda)))) {
-		stack;
 		dm_pool_free(fmt->cmd->mem, fid);
-		return NULL;
+		return_NULL;
 	}
 
 	mda->ops = &_metadata_format1_ops;
@@ -525,7 +482,7 @@ static struct format_instance *_format1_create_instance(const struct format_type
 	return fid;
 }
 
-static void _format1_destroy_instance(struct format_instance *fid)
+static void _format1_destroy_instance(struct format_instance *fid __attribute((unused)))
 {
 	return;
 }
@@ -556,16 +513,16 @@ struct format_type *init_format(struct cmd_context *cmd)
 {
 	struct format_type *fmt = dm_malloc(sizeof(*fmt));
 
-	if (!fmt) {
-		stack;
-		return NULL;
-	}
+	if (!fmt)
+		return_NULL;
 
 	fmt->cmd = cmd;
 	fmt->ops = &_format1_ops;
 	fmt->name = FMT_LVM1_NAME;
 	fmt->alias = NULL;
-	fmt->features = FMT_RESTRICTED_LVIDS | FMT_ORPHAN_ALLOCATABLE;
+	fmt->orphan_vg_name = FMT_LVM1_ORPHAN_VG_NAME;
+	fmt->features = FMT_RESTRICTED_LVIDS | FMT_ORPHAN_ALLOCATABLE |
+			FMT_RESTRICTED_READAHEAD;
 	fmt->private = NULL;
 
 	if (!(fmt->labeller = lvm1_labeller_create(fmt))) {