1
0
mirror of git://sourceware.org/git/lvm2.git synced 2025-09-24 21:44:22 +03:00

Compare commits

..

48 Commits

Author SHA1 Message Date
David Teigland
124b515d9f makefile: fixes and improvements 2017-02-09 10:47:41 -06:00
David Teigland
e00ebdc934 man: break long lines that would wrap
When all the required command line elements would go
beyond 80 chars and wrap to a new line, insert a line
break and indent, and continue the required elements
on a second line.
2017-02-08 16:52:26 -06:00
David Teigland
60d26a4e42 man: add examples from previous man pages 2017-02-08 16:10:45 -06:00
David Teigland
dce38761ac man: use .end suffix instead of .notes 2017-02-08 12:59:22 -06:00
David Teigland
14cc164262 man: fix generation of builtin commands 2017-02-08 11:52:15 -06:00
David Teigland
0d929d5a3d commands: remove debugging output
Remove the command id from the error messages.
2017-02-08 10:56:07 -06:00
David Teigland
c8945374d5 commands: rename structure fields
remove "line" from command_line_id/command_line_enum
2017-02-08 10:33:16 -06:00
David Teigland
1f67df932d args: use arg parsing function for region size
Consolidate the validation of the region size arg
in a new arg parsing function.
2017-02-08 09:57:13 -06:00
David Teigland
1bef87796f lvconvert: remove code for changing region size
from the generic raid type conversion code.
2017-02-07 15:01:30 -06:00
David Teigland
cfbbb4bee9 lvconvert: add command to change region size of a raid LV 2017-02-07 13:27:58 -06:00
David Teigland
058b53cdf4 vgchange: fix uint32 parsing of logicalvolume arg 2017-02-07 10:38:30 -06:00
David Teigland
2b8019fd43 commands: recognize raid variations 2017-02-07 10:38:30 -06:00
David Teigland
c213aa76c4 commands: move command def parsing into lvm binary 2017-02-07 10:38:30 -06:00
David Teigland
15e69fae19 lvconvert: remove unused code
For "split" which is not an alias for splitmirrors.
2017-02-07 10:38:30 -06:00
David Teigland
fe0ad1157d args: split is a synonym for splitcache
also tidy the other synonyms
2017-02-07 10:38:30 -06:00
David Teigland
88a2b1dc37 man: add vgsplit notes
from previous man page
2017-02-07 10:38:30 -06:00
David Teigland
03970369c5 man: add vgimportclone notes
from previous man page
2017-02-07 10:38:30 -06:00
David Teigland
c2147bda74 man: add vgimport notes
from previous man page
2017-02-07 10:38:30 -06:00
David Teigland
a61bc922a9 man: add vgexport notes
from previous man page
2017-02-07 10:38:30 -06:00
David Teigland
d41a2e6f30 man: add pvscan notes
from previous man page
2017-02-07 10:38:30 -06:00
David Teigland
55d6e9ebb7 man: add pvmove notes
Copied from the previous man page.
2017-02-07 10:38:30 -06:00
David Teigland
0a61ea395f man: lvmlockd/clvmd notes about activation
activation details about lvmlockd and clvmd that were
previously in the generic lvchange -a section are now
moved to these man pages.
2017-02-07 10:38:30 -06:00
David Teigland
e61f84cd1f generate man pages 2017-02-07 10:38:30 -06:00
David Teigland
ca8e344041 man: lvmthin updates
Some minor changes to some of the command syntaxes
to use more standard forms.
2017-02-07 10:38:30 -06:00
David Teigland
14838f584f ccmd: split into multiple files 2017-02-07 10:38:30 -06:00
David Teigland
6033e894e2 command struct: remove command name refs
Change run time access to the command_name struct
cmd->cname instead of indirectly through
cmd->command->cname. This removes the two run time
fields from struct command.
2017-02-07 10:38:30 -06:00
David Teigland
9b4c2027bf command.h comment tidying 2017-02-07 10:38:30 -06:00
David Teigland
61e0925844 lvm shell: clear argv for each command 2017-02-07 10:38:30 -06:00
David Teigland
214f48aa96 help: accept positional args
lvm help <commandname> ...
2017-02-07 10:38:30 -06:00
David Teigland
29ee87987c fix lvmcmdline warning
declaration of ‘usage’ shadows a globa
2017-02-07 10:38:30 -06:00
David Teigland
2391324196 Makefile: clean up create-command parts 2017-02-07 10:38:30 -06:00
David Teigland
e8230719ee man lvm: remove options
all options are now included in commands
2017-02-07 10:38:30 -06:00
David Teigland
2c5f6b8a02 args: add man page descriptions 2017-02-07 10:38:30 -06:00
David Teigland
fe3b2f5ead args: use uint32 arg for maxphysicalvolumes 2017-02-07 10:38:30 -06:00
David Teigland
91a1273d56 lvconvert: remove unused code 2017-02-07 10:38:30 -06:00
David Teigland
5507307b5f lvconvert: use command defs for raid/mirror types
All lvconvert functionality has been moved out of the
previous monolithic lvconvert code, except conversions
related to raid/mirror/striped/linear.  This switches
that remaining code to be based on command defs, and
standard process_each_lv arg processing.  This final
switch results in quite a bit of dead code that is
also removed.
2017-02-07 10:38:30 -06:00
David Teigland
91b71a8b02 tests: use swapmetadata
and some other pool/cache/thin related changes
2017-02-07 10:38:30 -06:00
David Teigland
3057fe9fec lvconvert: use command defs for mergemirrors
and route the generic --merge to one of the
specific merge functions
2017-02-07 10:38:30 -06:00
David Teigland
258027d1b1 toollib: find VG name in option values when needed 2017-02-07 10:38:30 -06:00
David Teigland
2da0f6110d lvconvert: use command defs for thin/cache/pool creation
Everything related to thin and cache.
2017-02-07 10:38:30 -06:00
David Teigland
09d6229350 lvconvert: add startpoll command using command def
This is a new explicit version of 'lvconvert LV'
which has been an obscure way of triggering polling
to be restarted on an LV that was previously converted.
2017-02-07 10:38:30 -06:00
David Teigland
6312e0872c lvconvert: snapshot: use command definitions
Lift all the snapshot utilities (merge, split, combine)
out of the monolithic lvconvert implementation, using
the command definitions.  The old code associated with
these commands is now unused and will be removed separately.
2017-02-07 10:38:30 -06:00
David Teigland
c752f4e1eb lvconvert: remove unused calls for repair and replace
repair and replace are no longer called from the
monolithic lvconvert code, so remove the unused code.
2017-02-07 10:38:30 -06:00
David Teigland
cd36b1c698 lvconvert: repair and replace: use command definitions
This lifts the lvconvert --repair and --replace commands
out of the monolithic lvconvert implementation.  The
previous calls into repair/replace can no longer be
reached and will be removed in a separate commit.
2017-02-07 10:36:18 -06:00
David Teigland
310d9a09fd lvchange: make use of command definitions
Reorganize the lvchange code to take advantage of
the command definition, and remove the validation
that is done by the command definintion rules.
2017-02-07 10:36:18 -06:00
David Teigland
81e84556b5 process_each_lv: add check_single_lv function
The new check_single_lv() function is called prior to the
existing process_single_lv().  If the check function returns 0,
the LV will not be processed.

The check_single_lv function is meant to be a standard method
to validate the combination of specific command + specific LV,
and decide if the combination is allowed.  The check_single
function can be used by anything that calls process_each_lv.

As commands are migrated to take advantage of command
definitions, each command definition gets its own entry
point which calls process_each for itself, passing a
pair of check_single/process_single functions which can
be specific to the narrowly defined command def.
2017-02-07 10:36:18 -06:00
David Teigland
c903d0e4b3 commands: new method for defining commands
. Define a prototype for every lvm command.
. Match every user command with one definition.
. Generate help text and man pages from them.

The new file command-lines.in defines a prototype for every
unique lvm command.  A unique lvm command is a unique
combination of: command name + required option args +
required positional args.  Each of these prototypes also
includes the optional option args and optional positional
args that the command will accept, a description, and a
unique string ID for the definition.  Any valid command
will match one of the prototypes.

Here's an example of the lvresize command definitions from
command-lines.in, there are three unique lvresize commands:

lvresize --size SizeMB LV
OO: --alloc Alloc, --autobackup Bool, --force,
--nofsck, --nosync, --noudevsync, --reportformat String, --resizefs,
--stripes Number, --stripesize SizeKB, --poolmetadatasize SizeMB
OP: PV ...
ID: lvresize_by_size
DESC: Resize an LV by a specified size.

lvresize LV PV ...
OO: --alloc Alloc, --autobackup Bool, --force,
--nofsck, --nosync, --noudevsync,
--reportformat String, --resizefs, --stripes Number, --stripesize SizeKB
ID: lvresize_by_pv
DESC: Resize an LV by specified PV extents.
FLAGS: SECONDARY_SYNTAX

lvresize --poolmetadatasize SizeMB LV_thinpool
OO: --alloc Alloc, --autobackup Bool, --force,
--nofsck, --nosync, --noudevsync,
--reportformat String, --stripes Number, --stripesize SizeKB
OP: PV ...
ID: lvresize_pool_metadata_by_size
DESC: Resize a pool metadata SubLV by a specified size.

The three commands have separate definitions because they have
different required parameters.  Required parameters are specified
on the first line of the definition.  Optional options are
listed after OO, and optional positional args are listed after OP.

This data is used to generate corresponding command definition
structures for lvm in command-lines.h.  usage/help output is also
auto generated, so it is always in sync with the definitions.

Every user-entered command is compared against the set of
command structures, and matched with one.  An error is
reported if an entered command does not have the required
parameters for any definition.  The closest match is printed
as a suggestion, and running lvresize --help will display
the usage for each possible lvresize command.

The prototype syntax used for help/man output includes
required --option and positional args on the first line,
and optional --option and positional args enclosed in [ ]
on subsequent lines.

  command_name <required_opt_args> <required_pos_args>
          [ <optional_opt_args> ]
          [ <optional_pos_args> ]

Command definitions that are not to be advertised/suggested
have the flag SECONDARY_SYNTAX.  These commands will not be
printed in the normal help output.

Man page prototypes are also generated from the same original
command definitions, and are always in sync with the code
and help text.

Very early in command execution, a matching command definition
is found.  lvm then knows the operation being done, and that
the provided args conform to the definition.  This will allow
lots of ad hoc checking/validation to be removed throughout
the code.

Each command definition can also be routed to a specific
function to implement it.  The function is associated with
an enum value for the command definition (generated from
the ID string.)  These per-command-definition implementation
functions have not yet been created, so all commands
currently fall back to the existing per-command-name
implementation functions.

Using per-command-definition functions will allow lots of
code to be removed which tries to figure out what the
command is meant to do.  This is currently based on ad hoc
and complicated option analysis.  When using the new
functions, what the command is doing is already known
from the associated command definition.
2017-02-07 10:36:18 -06:00
David Teigland
2da7b10988 lvmlockd: test mode doesn't work
The --test option is not yet compatible with shared VGs
because changes are made in lvmlockd that cannot be
reversed or faked.
2017-02-07 10:36:18 -06:00
741 changed files with 14649 additions and 54987 deletions

View File

@@ -59,8 +59,6 @@ liblvm: lib
daemons: lib libdaemon tools
tools: lib libdaemon device-mapper
po: tools daemons
man: tools
all_man: tools
scripts: liblvm libdm
lib.device-mapper: include.device-mapper
@@ -100,7 +98,7 @@ CLEAN_DIRS += autom4te.cache
check check_system check_cluster check_local check_lvmetad check_lvmpolld check_lvmlockd_test check_lvmlockd_dlm check_lvmlockd_sanlock unit: all
$(MAKE) -C test $(@)
conf.generate man.generate: tools
conf.generate: tools
# how to use parenthesis in makefiles
leftparen:=(
@@ -130,9 +128,8 @@ rpm: dist
$(top_srcdir)/spec/source.inc >$(rpmbuilddir)/SOURCES/source.inc
rpmbuild -v --define "_topdir $(rpmbuilddir)" -ba $(top_srcdir)/spec/lvm2.spec
generate: conf.generate man.generate
generate: conf.generate
$(MAKE) -C conf generate
$(MAKE) -C man generate
all_man:
$(MAKE) -C man all_man

18
README
View File

@@ -6,17 +6,11 @@ Installation instructions are in INSTALL.
There is no warranty - see COPYING and COPYING.LIB.
Tarballs are available from:
ftp://sourceware.org/pub/lvm2/
ftp://sources.redhat.com/pub/lvm2/
https://github.com/lvmteam/lvm2/releases
The source code is stored in git:
https://sourceware.org/git/?p=lvm2.git
git clone git://sourceware.org/git/lvm2.git
mirrored to:
https://github.com/lvmteam/lvm2
git clone https://github.com/lvmteam/lvm2.git
git clone git@github.com:lvmteam/lvm2.git
http://git.fedorahosted.org/git/lvm2.git
git clone git://git.fedorahosted.org/git/lvm2.git
Mailing list for general discussion related to LVM2:
linux-lvm@redhat.com
@@ -34,14 +28,6 @@ and multipath-tools:
dm-devel@redhat.com
Subscribe from https://www.redhat.com/mailman/listinfo/dm-devel
Website:
https://sourceware.org/lvm2/
Report upstream bugs at:
https://bugzilla.redhat.com/enter_bug.cgi?product=LVM%20and%20device-mapper
or open issues at:
https://github.com/lvmteam/lvm2/issues
The source code repository used until 7th June 2012 is accessible here:
http://sources.redhat.com/cgi-bin/cvsweb.cgi/LVM2/?cvsroot=lvm2.

View File

@@ -1 +1 @@
2.02.177(2)-git (2017-11-03)
2.02.169(2)-git (2016-11-30)

View File

@@ -1 +1 @@
1.02.146-git (2017-11-03)
1.02.138-git (2016-11-30)

183
WHATS_NEW
View File

@@ -1,183 +1,6 @@
Version 2.02.177 -
====================================
Fix memleak of string buffer when vgcfgbackup runs in secure mode.
Do not print error when clvmd cannot find running clvmd.
Prevent start of new merge of snapshot if origin is already being merged.
Version 2.02.176 - 3rd November 2017
====================================
Keep Install section only in lvm2-{lvmetad,lvmpolld}.socket systemd unit.
Fix segfault in lvm_pv_remove in liblvm. (2.02.173)
Do not allow storing VG metadata with LV without any segment.
Fix printed message when thin snapshot was already merged.
Remove created spare LV when creation of thin-pool failed.
Avoid reading ignored metadata when mda gets used again.
Fix detection of moved PVs in vgsplit. (2.02.175)
Ignore --stripes/--stripesize on RAID takeover
Improve used paths for generated systemd units and init shells.
Disallow creation of snapshot of mirror/raid subLV (was never supported).
Fix regression in more advanced vgname extraction in lvconvert (2.02.169).
Allow lvcreate to be used for caching of _tdata LV.
Avoid internal error when resizing cache type _tdata LV (not yet supported).
Show original converted names when lvconverting LV to pool volume.
Move lib code used only by liblvm into metadata-liblvm.c.
Distinguish between device not found and excluded by filter.
Monitor external origin LVs.
Remove the replicator code, including configure --with-replicators.
Allow lvcreate --type mirror to work with 100%FREE.
Improve selection of resource name for complex volume activation lock.
Avoid cutting first character of resource name for activation lock.
Support for encrypted devices in fsadm.
Improve thin pool overprovisioning and repair warning messages.
Fix incorrect adjustment of region size on striped RaidLVs.
Version 2.02.175 - 6th October 2017
===================================
Use --help with blockdev when checking for --getsize64 support in fsadm.
Dump lvmdbusd debug information with SIGUSR1.
Fix metadata corruption in vgsplit and vgmerge intermediate states.
Add PV_MOVED_VG PV status flag to mark PVs moving between VGs.
Fix lvmdbus hang and recognise unknown VG correctly.
Improve error messages when command rules fail.
Require LV name with pvmove in a shared VG.
Allow shared active mirror LVs with lvmlockd, dlm, and cmirrord.
Support lvconvert --repair with cache and cachepool volumes.
lvconvert --repair respects --poolmetadataspare option.
Mark that we don't plan to develop liblvm2app and python bindings any further.
Fix thin pool creation in shared VG. (2.02.173)
Version 2.02.174 - 13th September 2017
======================================
Prevent raid1 split with trackchanges in a shared VG.
Avoid double unlocking of client & lockspace mutexes in lvmlockd.
Fix leaking of file descriptor for non-blocking filebased locking.
Fix check for 2nd mda at end of disk fits if using pvcreate --restorefile.
Use maximum metadataarea size that fits with pvcreate --restorefile.
Always clear cached bootloaderarea when wiping label e.g. in pvcreate.
Disallow --bootloaderareasize with pvcreate --restorefile.
Fix lvmlockd check for running lock managers during lock adoption.
Add --withgeneralpreamble and --withlocalpreamble to lvmconfig.
Improve makefiles' linking.
Fix some paths in generated makefiles to respected configured settings.
Add warning when creating thin-pool with zeroing and chunk size >= 512KiB.
Introduce exit code 4 EINIT_FAILED to replace -1 when initialisation fails.
Add synchronization points with udev during reshape of raid LVs.
Version 2.02.173 - 20th July 2017
=================================
Add synchronization points with udev during conversion of raid LVs.
Improve --size args validation and report more detailed error message.
Initialize debugging mutex before any debug message in clvmd.
Log error instead of warn when noticing connection problem with lvmetad.
Fix memory leak in lvmetad when working with duplicates.
Remove restrictions on reshaping open and clustered raid devices.
Add incompatible data_offset to raid metadata to fix reshape activation.
Accept 'lvm -h' and 'lvm --help' as well as 'lvm help' for help.
Suppress error message from accept() on clean lvmetad shutdown.
Tidy clvmd client list processing and fix segfaults.
Protect clvmd debug log messages with mutex and add client id.
Fix shellcheck reported issues for script files.
Version 2.02.172 - 28th June 2017
=================================
Add missing NULL to argv array when spliting cmdline arguments.
Add display_percent helper function for printing percent values.
lvconvert --repair handles failing raid legs (present but marked 'D'ead).
Do not lvdisplay --maps unset settings of cache pool.
Fix lvdisplay --maps for cache pool without policy settings.
Support aborting of flushing cache LV.
Reenable conversion of data and metadata thin-pool volumes to raid.
Improve raid status reporting with lvs.
No longer necessary to '--force' a repair for RAID1.
Linear to RAID1 upconverts now use "recover" sync action, not "resync".
Improve lvcreate --cachepool arg validation.
Limit maximum size of thin-pool for specific chunk size.
Print a warning about in-use PVs with no VG using them.
Disable automatic clearing of PVs that look like in-use orphans.
Cache format2 flag is now using segment name type field.
Support storing status flags via segtype name field.
Stop using '--yes' mode when fsadm runs without terminal.
Extend validation of filesystems resized by fsadm.
Enhance lvconvert automatic settings of possible (raid) LV types.
Allow lvchange to change properties on a thin pool data sub LV.
Fix lvcreate extent percentage calculation for mirrors.
Don't reinstate still-missing devices when correcting inconsistent metadata.
Properly handle subshell return codes in fsadm.
Disallow cachepool creation with policy cleaner and mode writeback.
Version 2.02.171 - 3rd May 2017
===============================
Fix memory warnings by using mempools for command definition processing.
Fix running commands from a script file.
Add pvcreate prompt when device size doesn't match setphysicalvolumesize.
lvconvert - preserve region size on raid1 image count changes
Adjust pvresize/pvcreate messages and prompt if underlying dev size differs.
raid - sanely handle insufficient space on takeover.
Fix configure --enable-notify-dbus status message.
Change configure option name prefix from --enable-lockd to --enable-lvmlockd.
lvcreate - raise mirror/raid default regionsize to 2MiB
Add missing configurable prefix to configuration file installation directory.
Version 2.02.170 - 13th April 2017
==================================
Introduce global/fsadm_executable to make fsadm path configurable.
Look for limited thin pool metadata size when using 16G metadata.
Add lvconvert pool creation rule disallowing options with poolmetadata.
Fix lvconvert when the same LV is incorrectly reused in options.
Fix lvconvert VG name validation in option values.
Fix missing lvmlockd LV locks in lvchange and lvconvert.
Fix dmeventd setup for lvchange --poll.
Fix use of --poll and --monitor with lvchange and vgchange.
Disallow lvconvert of hidden LV to a pool.
Ignore --partial option when not used for activation.
Allow --activationmode option with lvchange --refresh.
Better message on lvconvert --regionsize
Allow valid lvconvert --regionsize change
Add raid10 alias raid10_near
Handle insufficient PVs on lvconvert takeover
Fix SIGINT blocking to prevent corrupted metadata
Fix systemd unit existence check for lvmconf --services --startstopservices.
Check and use PATH_MAX buffers when creating vgrename device paths.
Version 2.02.169 - 28th March 2017
==================================
Automatically decide whether '-' in a man page is a hyphen or a minus sign.
Add build-time configuration command line to 'lvm version' output.
Handle known table line parameter order change in specific raid target vsns.
Conditionally reject raid convert to striped/raid0* after reshape.
Ensure raid6 upconversion restrictions.
Adjust mirror & raid dmeventd plugins for new lvconvert --repair behaviour.
Disable lvmetad when lvconvert --repair is run.
Remove obsolete lvmchange binary - convert to built-in command.
Show more information for cached volumes in lvdisplay [-m].
Add option for lvcreate/lvconvert --cachemetadataformat auto|1|2.
Support cache segment with configurable metadata format.
Add allocation/cache_metadata_format profilable settings.
Use function cache_set_params() for both lvcreate and lvconvert.
Skip rounding on cache chunk size boudary when create cache LV.
Improve cache_set_params support for chunk_size selection.
Fix metadata profile allocation/cache_[mode|policy] setting.
Fix missing support for using allocation/cache_pool_chunk_size setting.
Upstream git moved to https://sourceware.org/git/?p=lvm2
Support conversion of raid type, stripesize and number of disks
Reject writemostly/writebehind in lvchange during resynchronization.
Deactivate active origin first before removal for improved workflow.
Fix regression of accepting both --type and -m with lvresize. (2.02.158)
Add lvconvert --swapmetadata, new specific way to swap pool metadata LVs.
Add lvconvert --startpoll, new specific way to start polling conversions.
Add lvconvert --mergethin, new specific way to merge thin snapshots.
Add lvconvert --mergemirrors, new specific way to merge split mirrors.
Add lvconvert --mergesnapshot, new specific way to combine cow LVs.
Split up lvconvert code based on command definitions.
Split up lvchange code based on command definitions.
Generate help output and man pages from command definitions.
Verify all command line items against command definition.
Match every command run to one command definition.
Specify every allowed command definition/syntax in command-lines.in.
Add extra memory page when limiting pthread stack size in clvmd.
Support striped/raid0* <-> raid10_near conversions.
Support shrinking of RaidLVs.
Support region size changes on existing RaidLVs.
Version 2.02.169 -
=====================================
Support region size changes on existing RaidLVs
Avoid parallel usage of cpg_mcast_joined() in clvmd with corosync.
Support raid6_{ls,rs,la,ra}_6 segment types and conversions from/to it.
Support raid6_n_6 segment type and conversions from/to it.

View File

@@ -1,78 +1,16 @@
Version 1.02.146 -
====================================
Suppress integrity encryption keys in 'table' output unless --showkeys supplied.
Version 1.02.145 - 3rd November 2017
====================================
Keep Install section only in dm-event.socket systemd unit.
Issue a specific error with dmsetup status if device is unknown.
Fix RT_LIBS reference in generated libdevmapper.pc for pkg-config
Version 1.02.144 - 6th October 2017
===================================
Schedule exit when received SIGTERM in dmeventd.
Also try to unmount /boot on blkdeactivate -u if on top of supported device.
Use blkdeactivate -r wait in blk-availability systemd service/initscript.
Add blkdeactivate -r wait option to wait for MD resync/recovery/reshape.
Fix blkdeactivate regression with failing DM/MD devs deactivation (1.02.142).
Fix typo in blkdeactivate's '--{dm,lvm,mpath}options' option name.
Correct return value testing when get reserved values for reporting.
Take -S with dmsetup suspend/resume/clear/wipe_table/remove/deps/status/table.
Version 1.02.143 - 13th September 2017
======================================
Restore umask when creation of node fails.
Add --concise to dmsetup create for many devices with tables in one command.
Accept minor number without major in library when it knows dm major number.
Introduce single-line concise table output format: dmsetup table --concise
Version 1.02.142 - 20th July 2017
=================================
Create /dev/disk/by-part{uuid,label} and gpt-auto-root symlinks with udev.
Version 1.02.141 - 28th June 2017
=================================
Fix reusing of dm_task structure for status reading (used by dmeventd).
Add dm_percent_to_round_float for adjusted percentage rounding.
Reset array with dead rimage devices once raid gets in sync.
Drop unneeded --config option from raid dmeventd plugin.
dm_get_status_raid() handle better some incosistent md statuses.
Accept truncated files in calls to dm_stats_update_regions_from_fd().
Restore Warning by 5% increment when thin-pool is over 80% (1.02.138).
Version 1.02.140 - 3rd May 2017
===============================
Add missing configure --enable-dmfilemapd status message and fix --disable.
Version 1.02.139 - 13th April 2017
==================================
Fix assignment in _target_version() when dm task can't run.
Flush stdout on each iteration when using --count or --interval.
Show detailed error message when execvp fails while starting dmfilemapd.
Fix segmentation fault when dmfilemapd is run with no arguments.
Numerous minor dmfilemapd fixes from coverity.
Version 1.02.138 - 28th March 2017
==================================
Support additional raid5/6 configurations.
Provide dm_tree_node_add_cache_target@base compatible symbol.
Support DM_CACHE_FEATURE_METADATA2, new cache metadata format 2.
Improve code to handle mode mask for cache nodes.
Cache status check for passthrough also require trailing space.
Add extra memory page when limiting pthread stack size in dmeventd.
Avoids immediate resume when preloaded device is smaller.
Do not suppress kernel key description in dmsetup table output for dm-crypt.
Version 1.02.138 -
=====================================
Support configurable command executed from dmeventd thin plugin.
Support new R|r human readable units output format.
Thin dmeventd plugin reacts faster on lvextend failure path with umount.
Add dm_stats_bind_from_fd() to bind a stats handle from a file descriptor.
Do not try call callback when reverting activation on error path.
Fix file mapping for extents with physically adjacent extents in dmstats.
Fix file mapping for extents with physically adjacent extents.
Validation vsnprintf result in runtime translate of dm_log (1.02.136).
Separate filemap extent allocation from region table in dmstats.
Fix segmentation fault when filemap region creation fails in dmstats.
Fix performance of region cleanup for failed filemap creation in dmstats.
Fix very slow region deletion with many regions in dmstats.
Separate filemap extent allocation from region table.
Fix segmentation fault when filemap region creation fails.
Fix performance of region cleanup for failed filemap creation.
Fix very slow region deletion with many regions.
Version 1.02.137 - 30th November 2016
=====================================

View File

@@ -32,8 +32,8 @@ include $(top_builddir)/make.tmpl
.PHONY: install_conf install_localconf install_profiles
generate:
LD_LIBRARY_PATH=$(top_builddir)/libdm:$(LD_LIBRARY_PATH) $(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withgeneralpreamble --withcomments --ignorelocal --withspaces > example.conf.in
LD_LIBRARY_PATH=$(top_builddir)/libdm:$(LD_LIBRARY_PATH) $(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withlocalpreamble --withcomments --withspaces local > lvmlocal.conf.in
(cat $(top_srcdir)/conf/example.conf.base && LD_LIBRARY_PATH=$(top_builddir)/libdm:$(LD_LIBRARY_PATH) $(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withcomments --ignorelocal --withspaces) > example.conf.in
(cat $(top_srcdir)/conf/lvmlocal.conf.base && LD_LIBRARY_PATH=$(top_builddir)/libdm:$(LD_LIBRARY_PATH) $(top_builddir)/tools/lvm dumpconfig --type default --unconfigured --withcomments --withspaces local) > lvmlocal.conf.in
install_conf: $(CONFSRC)
@if [ ! -e $(confdir)/$(CONFDEST) ]; then \
@@ -48,8 +48,8 @@ install_localconf: $(CONFLOCAL)
fi
install_profiles: $(PROFILES)
$(INSTALL_DIR) $(profiledir)
$(INSTALL_DATA) $(PROFILES) $(profiledir)/
$(INSTALL_DIR) $(DESTDIR)$(DEFAULT_PROFILE_DIR)
$(INSTALL_DATA) $(PROFILES) $(DESTDIR)$(DEFAULT_PROFILE_DIR)/
install_lvm2: install_conf install_localconf install_profiles

View File

@@ -9,6 +9,6 @@ allocation {
cache_mode = "writethrough"
cache_policy = "smq"
cache_settings {
# currently no settings for "smq" policy
# currently no settins for "smq" policy
}
}

23
conf/example.conf.base Normal file
View File

@@ -0,0 +1,23 @@
# This is an example configuration file for the LVM2 system.
# It contains the default settings that would be used if there was no
# @DEFAULT_SYS_DIR@/lvm.conf file.
#
# Refer to 'man lvm.conf' for further information including the file layout.
#
# Refer to 'man lvm.conf' for information about how settings configured in
# this file are combined with built-in values and command line options to
# arrive at the final values used by LVM.
#
# Refer to 'man lvmconfig' for information about displaying the built-in
# and configured values used by LVM.
#
# If a default value is set in this file (not commented out), then a
# new version of LVM using this file will continue using that value,
# even if the new version of LVM changes the built-in default value.
#
# To put this file in a different directory and override @DEFAULT_SYS_DIR@ set
# the environment variable LVM_SYSTEM_DIR before running the tools.
#
# N.B. Take care that each setting only appears once if uncommenting
# example settings in this file.

View File

@@ -379,9 +379,8 @@ allocation {
# Configuration option allocation/raid_stripe_all_devices.
# Stripe across all PVs when RAID stripes are not specified.
# If enabled, all PVs in the VG or on the command line are used for
# raid0/4/5/6/10 when the command does not specify the number of
# stripes to use.
# If enabled, all PVs in the VG or on the command line are used for raid0/4/5/6/10
# when the command does not specify the number of stripes to use.
# This was the default behaviour until release 2.02.162.
# This configuration option has an automatic default value.
# raid_stripe_all_devices = 0
@@ -390,17 +389,6 @@ allocation {
# Cache pool metadata and data will always use different PVs.
cache_pool_metadata_require_separate_pvs = 0
# Configuration option allocation/cache_metadata_format.
# Sets default metadata format for new cache.
#
# Accepted values:
# 0 Automatically detected best available format
# 1 Original format
# 2 Improved 2nd. generation format
#
# This configuration option has an automatic default value.
# cache_metadata_format = 0
# Configuration option allocation/cache_mode.
# The default cache mode used for new cache.
#
@@ -417,7 +405,7 @@ allocation {
# Configuration option allocation/cache_policy.
# The default cache policy used for new cache volume.
# Since kernel 4.2 the default policy is smq (Stochastic multiqueue),
# Since kernel 4.2 the default policy is smq (Stochastic multique),
# otherwise the older mq (Multiqueue) policy is selected.
# This configuration option does not have a default value defined.
@@ -940,7 +928,7 @@ global {
use_lvmetad = @DEFAULT_USE_LVMETAD@
# Configuration option global/lvmetad_update_wait_time.
# Number of seconds a command will wait for lvmetad update to finish.
# The number of seconds a command will wait for lvmetad update to finish.
# After waiting for this period, a command will not use lvmetad, and
# will revert to disk scanning.
# This configuration option has an automatic default value.
@@ -1025,7 +1013,7 @@ global {
# Configuration option global/cache_disabled_features.
# Features to not use in the cache driver.
# This can be helpful for testing, or to avoid using a feature that is
# causing problems. Features include: policy_mq, policy_smq, metadata2.
# causing problems. Features include: policy_mq, policy_smq.
#
# Example
# cache_disabled_features = [ "policy_smq" ]
@@ -1070,12 +1058,6 @@ global {
# This configuration option has an automatic default value.
# cache_repair_options = [ "" ]
# Configuration option global/fsadm_executable.
# The full path to the fsadm command.
# LVM uses this command to help with lvresize -r operations.
# This configuration option has an automatic default value.
# fsadm_executable = "@FSADM_PATH@"
# Configuration option global/system_id_source.
# The method LVM uses to set the local system ID.
# Volume Groups can also be given a system ID (by vgcreate, vgchange,
@@ -1295,10 +1277,9 @@ activation {
# Configuration option activation/raid_region_size.
# Size in KiB of each raid or mirror synchronization region.
# The clean/dirty state of data is tracked for each region.
# The value is rounded down to a power of two if necessary, and
# is ignored if it is not a multiple of the machine memory page size.
raid_region_size = 2048
# For raid or mirror segment types, this is the amount of data that is
# copied at once when initializing, or moved at once by pvmove.
raid_region_size = 512
# Configuration option activation/error_when_full.
# Return errors if a thin pool runs out of space.
@@ -2073,7 +2054,6 @@ dmeventd {
# or metadata volume gets above 50%.
# Command which starts with 'lvm ' prefix is internal lvm command.
# You can write your own handler to customise behaviour in more details.
# User handler is specified with the full path starting with '/'.
# This configuration option has an automatic default value.
# thin_command = "lvm lvextend --use-policies"

19
conf/lvmlocal.conf.base Normal file
View File

@@ -0,0 +1,19 @@
# This is a local configuration file template for the LVM2 system
# which should be installed as @DEFAULT_SYS_DIR@/lvmlocal.conf .
#
# Refer to 'man lvm.conf' for information about the file layout.
#
# To put this file in a different directory and override
# @DEFAULT_SYS_DIR@ set the environment variable LVM_SYSTEM_DIR before
# running the tools.
#
# The lvmlocal.conf file is normally expected to contain only the
# "local" section which contains settings that should not be shared or
# repeated among different hosts. (But if other sections are present,
# they *will* get processed. Settings in this file override equivalent
# ones in lvm.conf and are in turn overridden by ones in any enabled
# lvm_<tag>.conf files.)
#
# Please take care that each setting only appears once if uncommenting
# example settings in this file and never copy this file between hosts.

695
configure vendored

File diff suppressed because it is too large Load Diff

View File

@@ -15,7 +15,6 @@ AC_PREREQ(2.69)
################################################################################
dnl -- Process this file with autoconf to produce a configure script.
AC_INIT
CONFIGURE_LINE="$0 $@"
AC_CONFIG_SRCDIR([lib/device/dev-cache.h])
AC_CONFIG_HEADERS([include/configure.h])
@@ -31,7 +30,6 @@ AS_IF([test -z "$CFLAGS"], [COPTIMISE_FLAG="-O2"])
case "$host_os" in
linux*)
CLDFLAGS="$CLDFLAGS -Wl,--version-script,.export.sym"
# equivalent to -rdynamic
ELDFLAGS="-Wl,--export-dynamic"
# FIXME Generate list and use --dynamic-list=.dlopen.sym
CLDWHOLEARCHIVE="-Wl,-whole-archive"
@@ -39,7 +37,6 @@ case "$host_os" in
LDDEPS="$LDDEPS .export.sym"
LIB_SUFFIX=so
DEVMAPPER=yes
AIO=yes
BUILD_LVMETAD=no
BUILD_LVMPOLLD=no
LOCKDSANLOCK=no
@@ -59,7 +56,6 @@ case "$host_os" in
CLDNOWHOLEARCHIVE=
LIB_SUFFIX=dylib
DEVMAPPER=yes
AIO=no
ODIRECT=no
DM_IOCTLS=no
SELINUX=no
@@ -87,12 +83,9 @@ AC_PROG_LN_S
AC_PROG_MAKE_SET
AC_PROG_MKDIR_P
AC_PROG_RANLIB
AC_CHECK_TOOL(AR, ar)
AC_PATH_TOOL(CFLOW_CMD, cflow)
AC_PATH_TOOL(CSCOPE_CMD, cscope)
AC_PATH_TOOL(CHMOD, chmod)
AC_PATH_TOOL(WC, wc)
AC_PATH_TOOL(SORT, sort)
################################################################################
dnl -- Check for header files.
@@ -110,7 +103,7 @@ AC_CHECK_HEADERS([assert.h ctype.h dirent.h errno.h fcntl.h float.h \
sys/time.h sys/types.h sys/utsname.h sys/wait.h time.h \
unistd.h], , [AC_MSG_ERROR(bailing out)])
AC_CHECK_HEADERS(termios.h sys/statvfs.h sys/timerfd.h sys/vfs.h linux/magic.h linux/fiemap.h)
AC_CHECK_HEADERS(termios.h sys/statvfs.h sys/timerfd.h linux/magic.h linux/fiemap.h)
case "$host_os" in
linux*)
@@ -125,7 +118,6 @@ AC_C_CONST
AC_C_INLINE
AC_CHECK_MEMBERS([struct stat.st_rdev])
AC_CHECK_TYPES([ptrdiff_t])
AC_STRUCT_ST_BLOCKS
AC_STRUCT_TM
AC_TYPE_OFF_T
AC_TYPE_PID_T
@@ -193,15 +185,9 @@ AC_SUBST(HAVE_FULL_RELRO)
################################################################################
dnl -- Prefix is /usr by default, the exec_prefix default is setup later
AC_PREFIX_DEFAULT(/usr)
################################################################################
dnl -- Clear default exec_prefix - install into /sbin rather than /usr/sbin
test "$exec_prefix" = NONE -a "$prefix" = NONE && exec_prefix=""
test "x$prefix" = xNONE && prefix=$ac_default_prefix
# Let make expand exec_prefix.
test "x$exec_prefix" = xNONE && exec_prefix='${prefix}'
if test "$prefix" = NONE; then
datarootdir=${ac_default_prefix}/share
fi
################################################################################
dnl -- Setup the ownership of the files
@@ -416,6 +402,22 @@ AC_DEFINE_UNQUOTED([DEFAULT_RAID10_SEGTYPE], ["$DEFAULT_RAID10_SEGTYPE"],
[Default segtype used for raid10 volumes.])
################################################################################
dnl -- asynchronous volume replicator inclusion type
AC_MSG_CHECKING(whether to include replicators)
AC_ARG_WITH(replicators,
AC_HELP_STRING([--with-replicators=TYPE],
[replicator support: internal/shared/none [none]]),
REPLICATORS=$withval, REPLICATORS=none)
AC_MSG_RESULT($REPLICATORS)
case "$REPLICATORS" in
none|shared) ;;
internal) AC_DEFINE([REPLICATOR_INTERNAL], 1,
[Define to 1 to include built-in support for replicators.]) ;;
*) AC_MSG_ERROR([--with-replicators parameter invalid ($REPLICATORS)]) ;;
esac
AC_ARG_WITH(default-sparse-segtype,
AC_HELP_STRING([--with-default-sparse-segtype=TYPE],
[default sparse segtype: thin/snapshot [thin]]),
@@ -611,10 +613,6 @@ case "$CACHE" in
CACHE_CHECK_VERSION_WARN=y
CACHE_CHECK_NEEDS_CHECK=no
fi
if test "$CACHE_CHECK_VSN_MINOR" -lt 7 ; then
AC_MSG_WARN([$CACHE_CHECK_CMD: Old version "$CACHE_CHECK_VSN" does not support new cache format V2])
CACHE_CHECK_VERSION_WARN=y
fi
fi
fi
# Empty means a config way to ignore cache dumping
@@ -668,9 +666,11 @@ AC_DEFINE_UNQUOTED([CACHE_RESTORE_CMD], ["$CACHE_RESTORE_CMD"],
################################################################################
dnl -- Disable readline
AC_MSG_CHECKING(whether to enable readline)
AC_ARG_ENABLE([readline],
AC_HELP_STRING([--disable-readline], [disable readline support]),
READLINE=$enableval, READLINE=maybe)
AC_MSG_RESULT($READLINE)
################################################################################
dnl -- Disable realtime clock support
@@ -1123,24 +1123,6 @@ if test "$DEVMAPPER" = yes; then
AC_DEFINE([DEVMAPPER_SUPPORT], 1, [Define to 1 to enable LVM2 device-mapper interaction.])
fi
################################################################################
dnl -- Disable aio
AC_MSG_CHECKING(whether to use aio)
AC_ARG_ENABLE(aio,
AC_HELP_STRING([--disable-aio],
[disable async i/o]),
AIO=$enableval)
AC_MSG_RESULT($AIO)
if test "$AIO" = yes; then
AC_CHECK_LIB(aio, io_setup,
[AC_DEFINE([AIO_SUPPORT], 1, [Define to 1 if aio is available.])
AIO_LIBS="-laio"
AIO_SUPPORT=yes],
[AIO_LIBS=
AIO_SUPPORT=no ])
fi
################################################################################
dnl -- Build lvmetad
AC_MSG_CHECKING(whether to build LVMetaD)
@@ -1164,10 +1146,10 @@ AC_MSG_RESULT($BUILD_LVMPOLLD)
################################################################################
BUILD_LVMLOCKD=no
dnl -- Build lvmlockdsanlock
AC_MSG_CHECKING(whether to build lvmlockdsanlock)
AC_ARG_ENABLE(lvmlockd-sanlock,
AC_HELP_STRING([--enable-lvmlockd-sanlock],
dnl -- Build lockdsanlock
AC_MSG_CHECKING(whether to build lockdsanlock)
AC_ARG_ENABLE(lockd-sanlock,
AC_HELP_STRING([--enable-lockd-sanlock],
[enable the LVM lock daemon using sanlock]),
LOCKDSANLOCK=$enableval)
AC_MSG_RESULT($LOCKDSANLOCK)
@@ -1182,10 +1164,10 @@ if test "$BUILD_LOCKDSANLOCK" = yes; then
fi
################################################################################
dnl -- Build lvmlockddlm
AC_MSG_CHECKING(whether to build lvmlockddlm)
AC_ARG_ENABLE(lvmlockd-dlm,
AC_HELP_STRING([--enable-lvmlockd-dlm],
dnl -- Build lockddlm
AC_MSG_CHECKING(whether to build lockddlm)
AC_ARG_ENABLE(lockd-dlm,
AC_HELP_STRING([--enable-lockd-dlm],
[enable the LVM lock daemon using dlm]),
LOCKDDLM=$enableval)
AC_MSG_RESULT($LOCKDDLM)
@@ -1287,80 +1269,75 @@ fi
AC_DEFINE_UNQUOTED(DEFAULT_USE_LVMPOLLD, [$DEFAULT_USE_LVMPOLLD],
[Use lvmpolld by default.])
################################################################################
dnl -- Check dmfilemapd
AC_MSG_CHECKING(whether to build dmfilemapd)
AC_ARG_ENABLE(dmfilemapd, AC_HELP_STRING([--enable-dmfilemapd],
[enable the dmstats filemap daemon]),
BUILD_DMFILEMAPD=$enableval, BUILD_DMFILEMAPD=no)
AC_MSG_RESULT($BUILD_DMFILEMAPD)
AC_DEFINE([DMFILEMAPD], $BUILD_DMFILEMAPD, [Define to 1 to enable the device-mapper filemap daemon.])
dnl -- dmfilemapd requires FIEMAP
if test "$BUILD_DMFILEMAPD" = yes; then
AC_CHECK_HEADER([linux/fiemap.h], , [AC_MSG_ERROR(--enable-dmfilemapd requires fiemap.h)])
fi
################################################################################
dnl -- Build notifydbus
AC_MSG_CHECKING(whether to build notifydbus)
AC_ARG_ENABLE(notify-dbus,
AC_HELP_STRING([--enable-notify-dbus],
[enable LVM notification using dbus]),
NOTIFYDBUS_SUPPORT=$enableval, NOTIFYDBUS_SUPPORT=no)
AC_MSG_RESULT($NOTIFYDBUS_SUPPORT)
NOTIFYDBUS=$enableval)
AC_MSG_RESULT($NOTIFYDBUS)
if test "$NOTIFYDBUS_SUPPORT" = yes; then
BUILD_NOTIFYDBUS=$NOTIFYDBUS
if test "$BUILD_NOTIFYDBUS" = yes; then
AC_DEFINE([NOTIFYDBUS_SUPPORT], 1, [Define to 1 to include code that uses dbus notification.])
SYSTEMD_LIBS="-lsystemd"
LIBS="-lsystemd $LIBS"
fi
################################################################################
dnl -- Look for dbus libraries
if test "$NOTIFYDBUS_SUPPORT" = yes; then
if test "$BUILD_NOTIFYDBUS" = yes; then
PKG_CHECK_MODULES(NOTIFY_DBUS, systemd >= 221, [HAVE_NOTIFY_DBUS=yes], $bailout)
fi
################################################################################
dnl -- Enable blkid wiping functionality
AC_MSG_CHECKING(whether to enable libblkid detection of signatures when wiping)
AC_ARG_ENABLE(blkid_wiping,
AC_HELP_STRING([--disable-blkid_wiping],
[disable libblkid detection of signatures when wiping and use native code instead]),
BLKID_WIPING=$enableval, BLKID_WIPING=maybe)
AC_MSG_RESULT($BLKID_WIPING)
DEFAULT_USE_BLKID_WIPING=0
if test "$BLKID_WIPING" != no; then
pkg_config_init
PKG_CHECK_MODULES(BLKID, blkid >= 2.24,
[ BLKID_WIPING=yes
BLKID_PC="blkid"
DEFAULT_USE_BLKID_WIPING=1
AC_DEFINE([BLKID_WIPING_SUPPORT], 1, [Define to 1 to use libblkid detection of signatures when wiping.])
], [if test "$BLKID_WIPING" = maybe; then
[test "$BLKID_WIPING" = maybe && BLKID_WIPING=yes],
[if test "$BLKID_WIPING" = maybe; then
BLKID_WIPING=no
else
AC_MSG_ERROR([bailing out... blkid library >= 2.24 is required])
fi])
if test "$BLKID_WIPING" = yes; then
BLKID_PC="blkid"
DEFAULT_USE_BLKID_WIPING=1
AC_DEFINE([BLKID_WIPING_SUPPORT], 1, [Define to 1 to use libblkid detection of signatures when wiping.])
else
DEFAULT_USE_BLKID_WIPING=0
fi
else
DEFAULT_USE_BLKID_WIPING=0
fi
AC_MSG_CHECKING([whether to enable libblkid detection of signatures when wiping])
AC_MSG_RESULT($BLKID_WIPING)
AC_DEFINE_UNQUOTED(DEFAULT_USE_BLKID_WIPING, [$DEFAULT_USE_BLKID_WIPING],
[Use blkid wiping by default.])
################################################################################
dnl -- Enable udev-systemd protocol to instantiate a service for background jobs
dnl -- Requires systemd version 205 at least (including support for systemd-run)
AC_MSG_CHECKING(whether to use udev-systemd protocol for jobs in background)
AC_ARG_ENABLE(udev-systemd-background-jobs,
AC_HELP_STRING([--disable-udev-systemd-background-jobs],
[disable udev-systemd protocol to instantiate a service for background job]),
UDEV_SYSTEMD_BACKGROUND_JOBS=$enableval,
UDEV_SYSTEMD_BACKGROUND_JOBS=maybe)
AC_MSG_RESULT($UDEV_SYSTEMD_BACKGROUND_JOBS)
if test "$UDEV_SYSTEMD_BACKGROUND_JOBS" != no; then
pkg_config_init
PKG_CHECK_MODULES(SYSTEMD, systemd >= 205,
[UDEV_SYSTEMD_BACKGROUND_JOBS=yes],
[test "$UDEV_SYSTEMD_BACKGROUND_JOBS" = maybe && UDEV_SYSTEMD_BACKGROUND_JOBS=yes],
[if test "$UDEV_SYSTEMD_BACKGROUND_JOBS" = maybe; then
UDEV_SYSTEMD_BACKGROUND_JOBS=no
else
@@ -1368,9 +1345,6 @@ if test "$UDEV_SYSTEMD_BACKGROUND_JOBS" != no; then
fi])
fi
AC_MSG_CHECKING(whether to use udev-systemd protocol for jobs in background)
AC_MSG_RESULT($UDEV_SYSTEMD_BACKGROUND_JOBS)
################################################################################
dnl -- Enable udev synchronisation
AC_MSG_CHECKING(whether to enable synchronisation with udev processing)
@@ -1474,8 +1448,6 @@ AC_SUBST([LVM2APP_LIB])
test "$APPLIB" = yes \
&& LVM2APP_LIB=-llvm2app \
|| LVM2APP_LIB=
AS_IF([test "$APPLIB"],
[AC_MSG_WARN([Python bindings are deprecated. Use D-Bus API])])
################################################################################
dnl -- Enable cmdlib
@@ -1496,8 +1468,6 @@ AC_ARG_ENABLE(dbus-service,
AC_HELP_STRING([--enable-dbus-service], [install D-Bus support]),
BUILD_LVMDBUSD=$enableval, BUILD_LVMDBUSD=no)
AC_MSG_RESULT($BUILD_LVMDBUSD)
AS_IF([test "$NOTIFYDBUS_SUPPORT" = yes && test "BUILD_LVMDBUSD" = yes],
[AC_MSG_WARN([Building D-Bus support without D-Bus notifications.])])
################################################################################
dnl -- Enable Python liblvm2app bindings
@@ -1560,7 +1530,6 @@ if test "$BUILD_LVMDBUSD" = yes; then
fi
if test "$PYTHON_BINDINGS" = yes -o "$PYTHON2_BINDINGS" = yes -o "$PYTHON3_BINDINGS" = yes; then
AC_MSG_WARN([Python bindings are deprecated. Use D-Bus API])
test "$APPLIB" != yes && AC_MSG_ERROR([Python_bindings require --enable-applib])
fi
@@ -1596,11 +1565,13 @@ dnl -- enable dmeventd handling
AC_MSG_CHECKING(whether to use dmeventd)
AC_ARG_ENABLE(dmeventd, AC_HELP_STRING([--enable-dmeventd],
[enable the device-mapper event daemon]),
BUILD_DMEVENTD=$enableval, BUILD_DMEVENTD=no)
AC_MSG_RESULT($BUILD_DMEVENTD)
DMEVENTD=$enableval)
AC_MSG_RESULT($DMEVENTD)
BUILD_DMEVENTD=$DMEVENTD
dnl -- dmeventd currently requires internal mirror support
if test "$BUILD_DMEVENTD" = yes; then
if test "$DMEVENTD" = yes; then
if test "$MIRRORS" != internal; then
AC_MSG_ERROR([--enable-dmeventd currently requires --with-mirrors=internal])
fi
@@ -1624,6 +1595,10 @@ AC_CHECK_LIB(c, canonicalize_file_name,
AC_DEFINE([HAVE_CANONICALIZE_FILE_NAME], 1,
[Define to 1 if canonicalize_file_name is available.]))
################################################################################
dnl -- Clear default exec_prefix - install into /sbin rather than /usr/sbin
test "$exec_prefix" = NONE -a "$prefix" = NONE && exec_prefix=""
################################################################################
dnl -- Check for dlopen
AC_CHECK_LIB(dl, dlopen,
@@ -1680,16 +1655,13 @@ fi
################################################################################
dnl -- Check for realtime clock support
RT_LIBS=
HAVE_REALTIME=no
if test "$REALTIME" = yes; then
AC_CHECK_FUNCS([clock_gettime], HAVE_REALTIME=yes)
AS_IF([test "$HAVE_REALTIME" != yes], [ # try again with -lrt
AC_CHECK_LIB([rt], [clock_gettime], RT_LIBS="-lrt"; HAVE_REALTIME=yes)])
AC_CHECK_LIB(rt, clock_gettime, HAVE_REALTIME=yes, HAVE_REALTIME=no)
if test "$HAVE_REALTIME" = yes; then
AC_DEFINE([HAVE_REALTIME], 1, [Define to 1 to include support for realtime clock.])
LIBS="-lrt $LIBS"
RT_LIB="-lrt"
else
AC_MSG_WARN(Disabling realtime clock)
fi
@@ -1733,7 +1705,6 @@ Note: (n)curses also seems to work as a substitute for termcap. This was
AC_DEFINE([READLINE_SUPPORT], 1,
[Define to 1 to include the LVM readline shell.])
dnl -- Try only with -lreadline and check for different symbol
READLINE=yes
LIBS=$lvm_saved_libs
AC_CHECK_LIB([readline], [rl_line_buffer],
[ READLINE_LIBS="-lreadline" ], [
@@ -1840,16 +1811,13 @@ dnl -- Ensure additional headers required
if test "$READLINE" = yes; then
AC_CHECK_HEADERS(readline/readline.h readline/history.h,,hard_bailout)
fi
AC_MSG_CHECKING(whether to enable readline)
AC_MSG_RESULT($READLINE)
if test "$BUILD_CMIRRORD" = yes; then
AC_CHECK_FUNCS(atexit,,hard_bailout)
fi
if test "$BUILD_LVMLOCKD" = yes; then
AS_IF([test "$HAVE_REALTIME" != yes], [AC_MSG_ERROR([Realtime clock support is mandatory for lvmlockd.])])
AC_CHECK_FUNCS(strtoull,,hard_bailout)
AC_CHECK_FUNCS(clock_gettime strtoull,,hard_bailout)
fi
if test "$BUILD_LVMPOLLD" = yes; then
@@ -1869,7 +1837,7 @@ if test "$CLUSTER" != none; then
AC_CHECK_FUNCS(socket,,hard_bailout)
fi
if test "$BUILD_DMEVENTD" = yes; then
if test "$DMEVENTD" = yes; then
AC_CHECK_HEADERS(arpa/inet.h,,hard_bailout)
fi
@@ -1885,10 +1853,6 @@ if test "$UDEV_SYNC" = yes; then
AC_CHECK_HEADERS(sys/ipc.h sys/sem.h,,hard_bailout)
fi
if test "$BUILD_DMFILEMAPD" = yes; then
AC_CHECK_HEADERS([sys/inotify.h],,hard_bailout)
fi
################################################################################
AC_PATH_TOOL(MODPROBE_CMD, modprobe)
@@ -1896,19 +1860,18 @@ if test -n "$MODPROBE_CMD"; then
AC_DEFINE_UNQUOTED([MODPROBE_CMD], ["$MODPROBE_CMD"], [The path to 'modprobe', if available.])
fi
SYSCONFDIR="$(eval echo $(eval echo $sysconfdir))"
SBINDIR="$(eval echo $(eval echo $sbindir))"
LVM_PATH="$SBINDIR/lvm"
lvm_exec_prefix=$exec_prefix
test "$lvm_exec_prefix" = NONE && lvm_exec_prefix=$prefix
test "$lvm_exec_prefix" = NONE && lvm_exec_prefix=$ac_default_prefix
LVM_PATH="$lvm_exec_prefix/sbin/lvm"
AC_DEFINE_UNQUOTED(LVM_PATH, ["$LVM_PATH"], [Path to lvm binary.])
USRSBINDIR="$(eval echo $(eval echo $usrsbindir))"
CLVMD_PATH="$USRSBINDIR/clvmd"
clvmd_prefix=$ac_default_prefix
test "$prefix" != NONE && clvmd_prefix=$prefix
CLVMD_PATH="$clvmd_prefix/sbin/clvmd"
AC_DEFINE_UNQUOTED(CLVMD_PATH, ["$CLVMD_PATH"], [Path to clvmd binary.])
FSADM_PATH="$SBINDIR/fsadm"
AC_DEFINE_UNQUOTED(FSADM_PATH, ["$FSADM_PATH"], [Path to fsadm binary.])
################################################################################
dnl -- dmeventd pidfile and executable path
if test "$BUILD_DMEVENTD" = yes; then
@@ -1926,7 +1889,7 @@ if test "$BUILD_DMEVENTD" = yes; then
AC_HELP_STRING([--with-dmeventd-path=PATH],
[dmeventd path [EPREFIX/sbin/dmeventd]]),
DMEVENTD_PATH=$withval,
DMEVENTD_PATH="$SBINDIR/dmeventd")
DMEVENTD_PATH="$lvm_exec_prefix/sbin/dmeventd")
AC_DEFINE_UNQUOTED(DMEVENTD_PATH, ["$DMEVENTD_PATH"],
[Path to dmeventd binary.])
fi
@@ -1969,17 +1932,13 @@ AC_ARG_WITH(default-cache-subdir,
AC_DEFINE_UNQUOTED(DEFAULT_CACHE_SUBDIR, ["$DEFAULT_CACHE_SUBDIR"],
[Name of default metadata cache subdirectory.])
# Select default system locking dir, prefer /run/lock over /var/lock
DEFAULT_SYS_LOCK_DIR="$RUN_DIR/lock"
test -d "$DEFAULT_SYS_LOCK_DIR" || DEFAULT_SYS_LOCK_DIR="/var/lock"
# Support configurable locking subdir for lvm
AC_ARG_WITH(default-locking-dir,
AC_HELP_STRING([--with-default-locking-dir=DIR],
[default locking directory [autodetect_lock_dir/lvm]]),
DEFAULT_LOCK_DIR=$withval,
[AC_MSG_CHECKING(for default lock directory)
DEFAULT_LOCK_DIR="$DEFAULT_SYS_LOCK_DIR/lvm"
DEFAULT_LOCK_DIR="$RUN_DIR/lock/lvm"
test -d "$RUN_DIR/lock" || DEFAULT_LOCK_DIR="/var/lock/lvm"
AC_MSG_RESULT($DEFAULT_LOCK_DIR)])
AC_DEFINE_UNQUOTED(DEFAULT_LOCK_DIR, ["$DEFAULT_LOCK_DIR"],
[Name of default locking directory.])
@@ -2021,8 +1980,6 @@ LVM_MINOR=`echo "$VER" | $AWK -F '.' '{print $2}'`
LVM_PATCHLEVEL=`echo "$VER" | $AWK -F '[[(.]]' '{print $3}'`
LVM_LIBAPI=`echo "$VER" | $AWK -F '[[()]]' '{print $2}'`
AC_DEFINE_UNQUOTED(LVM_CONFIGURE_LINE, "$CONFIGURE_LINE", [configure command line used])
################################################################################
AC_SUBST(APPLIB)
AC_SUBST(AWK)
@@ -2035,7 +1992,7 @@ AC_SUBST(BUILD_LVMPOLLD)
AC_SUBST(BUILD_LVMLOCKD)
AC_SUBST(BUILD_LOCKDSANLOCK)
AC_SUBST(BUILD_LOCKDDLM)
AC_SUBST(BUILD_DMFILEMAPD)
AC_SUBST(BUILD_NOTIFYDBUS)
AC_SUBST(CACHE)
AC_SUBST(CFLAGS)
AC_SUBST(CFLOW_CMD)
@@ -2074,22 +2031,19 @@ AC_SUBST(DEFAULT_RAID10_SEGTYPE)
AC_SUBST(DEFAULT_RUN_DIR)
AC_SUBST(DEFAULT_SPARSE_SEGTYPE)
AC_SUBST(DEFAULT_SYS_DIR)
AC_SUBST(DEFAULT_SYS_LOCK_DIR)
AC_SUBST(DEFAULT_USE_BLKID_WIPING)
AC_SUBST(DEFAULT_USE_LVMETAD)
AC_SUBST(DEFAULT_USE_LVMPOLLD)
AC_SUBST(DEFAULT_USE_LVMLOCKD)
AC_SUBST(DEVMAPPER)
AC_SUBST(AIO)
AC_SUBST(DLM_CFLAGS)
AC_SUBST(DLM_LIBS)
AC_SUBST(DL_LIBS)
AC_SUBST(AIO_LIBS)
AC_SUBST(DMEVENTD)
AC_SUBST(DMEVENTD_PATH)
AC_SUBST(DM_LIB_PATCHLEVEL)
AC_SUBST(ELDFLAGS)
AC_SUBST(FSADM)
AC_SUBST(FSADM_PATH)
AC_SUBST(BLKDEACTIVATE)
AC_SUBST(HAVE_LIBDL)
AC_SUBST(HAVE_REALTIME)
@@ -2134,18 +2088,15 @@ AC_SUBST(PYTHON3DIR)
AC_SUBST(QUORUM_CFLAGS)
AC_SUBST(QUORUM_LIBS)
AC_SUBST(RAID)
AC_SUBST(RT_LIBS)
AC_SUBST(RT_LIB)
AC_SUBST(READLINE_LIBS)
AC_SUBST(REPLICATORS)
AC_SUBST(SACKPT_CFLAGS)
AC_SUBST(SACKPT_LIBS)
AC_SUBST(SALCK_CFLAGS)
AC_SUBST(SALCK_LIBS)
AC_SUBST(SBINDIR)
AC_SUBST(SELINUX_LIBS)
AC_SUBST(SELINUX_PC)
AC_SUBST(SYSCONFDIR)
AC_SUBST(SYSTEMD_LIBS)
AC_SUBST(SNAPSHOTS)
AC_SUBST(STATICDIR)
AC_SUBST(STATIC_LINK)
@@ -2167,7 +2118,6 @@ AC_SUBST(UDEV_SYSTEMD_BACKGROUND_JOBS)
AC_SUBST(UDEV_RULE_EXEC_DETECTION)
AC_SUBST(UDEV_HAS_BUILTIN_BLKID)
AC_SUBST(USE_TRACKING)
AC_SUBST(USRSBINDIR)
AC_SUBST(VALGRIND_POOL)
AC_SUBST(WRITE_INSTALL)
AC_SUBST(DMEVENTD_PIDFILE)
@@ -2206,7 +2156,6 @@ daemons/dmeventd/plugins/raid/Makefile
daemons/dmeventd/plugins/mirror/Makefile
daemons/dmeventd/plugins/snapshot/Makefile
daemons/dmeventd/plugins/thin/Makefile
daemons/dmfilemapd/Makefile
daemons/lvmdbusd/Makefile
daemons/lvmdbusd/path.py
daemons/lvmetad/Makefile
@@ -2224,6 +2173,7 @@ lib/format1/Makefile
lib/format_pool/Makefile
lib/locking/Makefile
lib/mirror/Makefile
lib/replicator/Makefile
include/lvm-version.h
lib/raid/Makefile
lib/snapshot/Makefile
@@ -2282,14 +2232,10 @@ AS_IF([test -n "$THIN_CONFIGURE_WARN"],
[AC_MSG_WARN([Support for thin provisioning is limited since some thin provisioning tools are missing!])])
AS_IF([test -n "$THIN_CHECK_VERSION_WARN"],
[AC_MSG_WARN([You should also install latest thin_check vsn 0.7.0 (or later) for lvm2 thin provisioning])])
[AC_MSG_WARN([You should also install thin_check vsn 0.3.2 (or later) to use lvm2 thin provisioning])])
AS_IF([test -n "$CACHE_CONFIGURE_WARN"],
[AC_MSG_WARN([Support for cache is limited since some cache tools are missing!])])
AS_IF([test -n "$CACHE_CHECK_VERSION_WARN"],
[AC_MSG_WARN([You should install latest cache_check vsn 0.7.0 to use lvm2 cache metadata format 2])])
AS_IF([test "$ODIRECT" != yes],
[AC_MSG_WARN([O_DIRECT disabled: low-memory pvmove may lock up])])

View File

@@ -41,19 +41,6 @@ struct lv_segment *last_seg(const struct logical_volume *lv)
return ((struct lv_segment **)lv)[0];
}
const char *find_config_tree_str(struct cmd_context *cmd, int id, struct profile *profile)
{
return "STRING";
}
struct logical_volume *origin_from_cow(const struct logical_volume *lv)
{
if (lv)
return lv;
__coverity_panic__();
}
/* simple_memccpy() from glibc */
void *memccpy(void *dest, const void *src, int c, size_t n)
{
@@ -84,17 +71,6 @@ void model_FD_ZERO(void *fdset)
((long*)fdset)[i] = 0;
}
/* Resent Coverity reports quite weird errors... */
int *__errno_location(void)
{
}
const unsigned short **__ctype_b_loc (void)
{
}
/*
* Added extra pointer check to not need these models,
* for now just keep then in file

View File

@@ -48,12 +48,8 @@ ifeq ("@BUILD_LVMDBUSD@", "yes")
SUBDIRS += lvmdbusd
endif
ifeq ("@BUILD_DMFILEMAPD@", "yes")
SUBDIRS += dmfilemapd
endif
ifeq ($(MAKECMDGOALS),distclean)
SUBDIRS = clvmd cmirrord dmeventd lvmetad lvmpolld lvmlockd lvmdbusd dmfilemapd
SUBDIRS = clvmd cmirrord dmeventd lvmetad lvmpolld lvmlockd lvmdbusd
endif
include $(top_builddir)/make.tmpl

View File

@@ -31,9 +31,9 @@ SALCK_LIBS = @SALCK_LIBS@
SALCK_CFLAGS = @SALCK_CFLAGS@
SOURCES = \
clvmd-command.c\
clvmd.c\
lvm-functions.c\
clvmd-command.c \
clvmd.c \
lvm-functions.c \
refresh_clvmd.c
ifneq (,$(findstring cman,, "@CLVMD@,"))
@@ -72,21 +72,26 @@ endif
TARGETS = \
clvmd
LVMLIBS = $(LVMINTERNAL_LIBS)
ifeq ("@DMEVENTD@", "yes")
LVMLIBS += -ldevmapper-event
endif
include $(top_builddir)/make.tmpl
LIBS += $(LVMINTERNAL_LIBS) -ldevmapper $(PTHREAD_LIBS)
CFLAGS += -fno-strict-aliasing $(EXTRA_EXEC_CFLAGS)
LVMLIBS += -ldevmapper
LIBS += $(PTHREAD_LIBS)
ifeq ("@AIO@", "yes")
LIBS += $(AIO_LIBS)
endif
CFLAGS += -fno-strict-aliasing $(EXTRA_EXEC_CFLAGS)
LDFLAGS += $(EXTRA_EXEC_LDFLAGS)
INSTALL_TARGETS = \
install_clvmd
clvmd: $(OBJECTS) $(top_builddir)/lib/liblvm-internal.a
$(CC) $(CFLAGS) $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) \
-o clvmd $(OBJECTS) $(LMLIBS) $(LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) -o clvmd $(OBJECTS) \
$(LVMLIBS) $(LMLIBS) $(LIBS)
.PHONY: install_clvmd

View File

@@ -171,10 +171,8 @@ 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) {
if (*buf)
*retlen = dm_snprintf(*buf, buflen, "%s", strerror(status)) + 1;
else
*retlen = 0;
*retlen = 1 + ((*buf) ? dm_snprintf(*buf, buflen, "%s",
strerror(status)) : -1);
}
return status;
@@ -208,7 +206,7 @@ static int lock_vg(struct local_client *client)
lock_mode = ((int) lock_cmd & LCK_TYPE_MASK);
/* lock_flags = args[1]; */
lockname = &args[2];
DEBUGLOG("(%p) doing PRE command LOCK_VG '%s' at %x\n", client, lockname, lock_cmd);
DEBUGLOG("doing PRE command LOCK_VG '%s' at %x (client=%p)\n", lockname, lock_cmd, client);
if (lock_mode == LCK_UNLOCK) {
if (!(lkid = (int) (long) dm_hash_lookup(lock_hash, lockname)))
@@ -325,7 +323,7 @@ void cmd_client_cleanup(struct local_client *client)
int lkid;
char *lockname;
DEBUGLOG("(%p) Client thread cleanup\n", client);
DEBUGLOG("Client thread cleanup (%p)\n", client);
if (!client->bits.localsock.private)
return;
@@ -334,7 +332,7 @@ void cmd_client_cleanup(struct local_client *client)
dm_hash_iterate(v, lock_hash) {
lkid = (int)(long)dm_hash_get_data(lock_hash, v);
lockname = dm_hash_get_key(lock_hash, v);
DEBUGLOG("(%p) Cleanup: Unlocking lock %s %x\n", client, lockname, lkid);
DEBUGLOG("Cleanup (%p): Unlocking lock %s %x\n", client, lockname, lkid);
(void) sync_unlock(lockname, lkid);
}

View File

@@ -425,6 +425,8 @@ static void _add_up_node(const char *csid)
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 */

View File

@@ -58,7 +58,6 @@
/* Head of the fd list. Also contains
the cluster_socket details */
static struct local_client local_client_head;
static int _local_client_count = 0;
static unsigned short global_xid = 0; /* Last transaction ID issued */
@@ -69,37 +68,6 @@ static unsigned max_csid_len;
static unsigned max_cluster_message;
static unsigned max_cluster_member_name_len;
static void _add_client(struct local_client *new_client, struct local_client *existing_client)
{
_local_client_count++;
DEBUGLOG("(%p) Adding listener for fd %d. (Now %d monitored fds.)\n", new_client, new_client->fd, _local_client_count);
new_client->next = existing_client->next;
existing_client->next = new_client;
}
int add_client(struct local_client *new_client)
{
_add_client(new_client, &local_client_head);
return 0;
}
/* Returns 0 if delfd is found and removed from list */
static int _del_client(struct local_client *delfd)
{
struct local_client *lastfd, *thisfd;
for (lastfd = &local_client_head; (thisfd = lastfd->next); lastfd = thisfd)
if (thisfd == delfd) {
DEBUGLOG("(%p) Removing listener for fd %d\n", thisfd, thisfd->fd);
lastfd->next = delfd->next;
_local_client_count--;
return 0;
}
return 1;
}
/* Structure of items on the LVM thread list */
struct lvm_thread_cmd {
struct dm_list list;
@@ -124,7 +92,6 @@ static const size_t STACK_SIZE = 128 * 1024;
static pthread_attr_t stack_attr;
static int lvm_thread_exit = 0;
static pthread_mutex_t lvm_thread_mutex;
static pthread_mutex_t _debuglog_mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t lvm_thread_cond;
static pthread_barrier_t lvm_start_barrier;
static struct dm_list lvm_cmd_head;
@@ -251,17 +218,14 @@ void debuglog(const char *fmt, ...)
switch (clvmd_get_debug()) {
case DEBUG_STDERR:
pthread_mutex_lock(&_debuglog_mutex);
va_start(ap,fmt);
time(&P);
fprintf(stderr, "CLVMD[%x]: %.15s ", (int)pthread_self(), ctime_r(&P, buf_ctime) + 4);
vfprintf(stderr, fmt, ap);
va_end(ap);
fflush(stderr);
pthread_mutex_unlock(&_debuglog_mutex);
break;
case DEBUG_SYSLOG:
pthread_mutex_lock(&_debuglog_mutex);
if (!syslog_init) {
openlog("clvmd", LOG_PID, LOG_DAEMON);
syslog_init = 1;
@@ -270,7 +234,6 @@ void debuglog(const char *fmt, ...)
va_start(ap,fmt);
vsyslog(LOG_DEBUG, fmt, ap);
va_end(ap);
pthread_mutex_unlock(&_debuglog_mutex);
break;
case DEBUG_OFF:
break;
@@ -554,7 +517,7 @@ int main(int argc, char *argv[])
/* Initialise the LVM thread variables */
dm_list_init(&lvm_cmd_head);
if (pthread_attr_init(&stack_attr) ||
pthread_attr_setstacksize(&stack_attr, STACK_SIZE + getpagesize())) {
pthread_attr_setstacksize(&stack_attr, STACK_SIZE)) {
log_sys_error("pthread_attr_init", "");
exit(1);
}
@@ -621,7 +584,6 @@ int main(int argc, char *argv[])
local_client_head.fd = clops->get_main_cluster_fd();
local_client_head.type = CLUSTER_MAIN_SOCK;
local_client_head.callback = clops->cluster_fd_callback;
_local_client_count++;
/* Add the local socket to the list */
if (!(newfd = dm_zalloc(sizeof(struct local_client)))) {
@@ -632,14 +594,14 @@ int main(int argc, char *argv[])
newfd->fd = local_sock;
newfd->type = LOCAL_RENDEZVOUS;
newfd->callback = local_rendezvous_callback;
(void) add_client(newfd);
newfd->next = local_client_head.next;
local_client_head.next = newfd;
/* This needs to be started after cluster initialisation
as it may need to take out locks */
DEBUGLOG("Starting LVM thread\n");
DEBUGLOG("(%p) Main cluster socket fd %d with local socket %d (%p)\n",
&local_client_head, local_client_head.fd, newfd->fd, newfd);
DEBUGLOG("Main cluster socket fd %d (%p) with local socket %d (%p)\n",
local_client_head.fd, &local_client_head, newfd->fd, newfd);
/* Don't let anyone else to do work until we are started */
if (pthread_create(&lvm_thread, &stack_attr, lvm_thread_fn, &lvm_params)) {
@@ -675,7 +637,6 @@ int main(int argc, char *argv[])
while ((delfd = local_client_head.next)) {
local_client_head.next = delfd->next;
_local_client_count--;
/* Failing cleanup_zombie leaks... */
if (delfd->type == LOCAL_SOCK && !cleanup_zombie(delfd))
cmd_client_cleanup(delfd); /* calls sync_unlock */
@@ -737,13 +698,13 @@ static int local_rendezvous_callback(struct local_client *thisfd, char *buf,
pthread_mutex_init(&newfd->bits.localsock.mutex, NULL);
if (fcntl(client_fd, F_SETFD, 1))
DEBUGLOG("(%p) Setting CLOEXEC on client fd %d failed: %s\n", thisfd, client_fd, strerror(errno));
DEBUGLOG("Setting CLOEXEC on client fd failed: %s\n", strerror(errno));
newfd->fd = client_fd;
newfd->type = LOCAL_SOCK;
newfd->callback = local_sock_callback;
newfd->bits.localsock.all_success = 1;
DEBUGLOG("(%p) Got new connection on fd %d\n", newfd, newfd->fd);
DEBUGLOG("Got new connection on fd %d (%p)\n", newfd->fd, newfd);
*new_client = newfd;
}
return 1;
@@ -765,8 +726,8 @@ static int local_pipe_callback(struct local_client *thisfd, char *buf,
if (len == sizeof(int))
memcpy(&status, buffer, sizeof(int));
DEBUGLOG("(%p) Read on pipe %d, %d bytes, status %d\n",
thisfd, thisfd->fd, len, status);
DEBUGLOG("Read on pipe %d, %d bytes, status %d\n",
thisfd->fd, len, status);
/* EOF on pipe or an error, close it */
if (len <= 0) {
@@ -789,11 +750,11 @@ static int local_pipe_callback(struct local_client *thisfd, char *buf,
}
return -1;
} else {
DEBUGLOG("(%p) Background routine status was %d, sock_client %p\n",
thisfd, status, sock_client);
DEBUGLOG("Background routine status was %d, sock_client (%p)\n",
status, sock_client);
/* But has the client gone away ?? */
if (!sock_client) {
DEBUGLOG("(%p) Got pipe response for dead client, ignoring it\n", thisfd);
DEBUGLOG("Got pipe response for dead client, ignoring it\n");
} else {
/* If error then just return that code */
if (status)
@@ -833,7 +794,7 @@ static void timedout_callback(struct local_client *client, const char *csid,
return;
clops->name_from_csid(csid, nodename);
DEBUGLOG("(%p) Checking for a reply from %s\n", client, nodename);
DEBUGLOG("Checking for a reply from %s\n", nodename);
pthread_mutex_lock(&client->bits.localsock.mutex);
reply = client->bits.localsock.replies;
@@ -843,7 +804,7 @@ static void timedout_callback(struct local_client *client, const char *csid,
pthread_mutex_unlock(&client->bits.localsock.mutex);
if (!reply) {
DEBUGLOG("(%p) Node %s timed-out\n", client, nodename);
DEBUGLOG("Node %s timed-out\n", nodename);
add_reply_to_list(client, ETIMEDOUT, csid,
"Command timed out", 18);
}
@@ -858,7 +819,7 @@ static void timedout_callback(struct local_client *client, const char *csid,
*/
static void request_timed_out(struct local_client *client)
{
DEBUGLOG("(%p) Request timed-out. padding\n", client);
DEBUGLOG("Request timed-out. padding\n");
clops->cluster_do_node_callback(client, timedout_callback);
if (!client->bits.localsock.threadid)
@@ -892,11 +853,13 @@ static void main_loop(int cmd_timeout)
while (!quit) {
fd_set in;
int select_status;
struct local_client *thisfd, *nextfd;
struct local_client *thisfd;
struct timeval tv = { cmd_timeout, 0 };
int quorate = clops->is_quorate();
int client_count = 0;
int max_fd = 0;
struct local_client *lastfd = &local_client_head;
struct local_client *nextfd = local_client_head.next;
/* Wait on the cluster FD and all local sockets/pipes */
local_client_head.fd = clops->get_main_cluster_fd();
@@ -912,22 +875,21 @@ static void main_loop(int cmd_timeout)
fprintf(stderr, "WARNING: Your cluster may freeze up if the number of clvmd file descriptors (%d) exceeds %d.\n", max_fd + 1, FD_SETSIZE);
}
for (thisfd = &local_client_head; thisfd; thisfd = nextfd) {
nextfd = thisfd->next;
for (thisfd = &local_client_head; thisfd; thisfd = nextfd, nextfd = thisfd ? thisfd->next : NULL) {
if (thisfd->removeme && !cleanup_zombie(thisfd)) {
/* cleanup_zombie might have removed the next list element */
nextfd = thisfd->next;
(void) _del_client(thisfd);
DEBUGLOG("(%p) removeme set with %d monitored fds remaining\n", thisfd, _local_client_count);
struct local_client *free_fd = thisfd;
lastfd->next = nextfd;
DEBUGLOG("removeme set for %p with %d monitored fds remaining\n", free_fd, client_count - 1);
/* Queue cleanup, this also frees the client struct */
add_to_lvmqueue(thisfd, NULL, 0, NULL);
add_to_lvmqueue(free_fd, NULL, 0, NULL);
continue;
}
lastfd = thisfd;
if (thisfd->removeme)
continue;
@@ -977,15 +939,16 @@ static void main_loop(int cmd_timeout)
type == CLUSTER_INTERNAL)
goto closedown;
DEBUGLOG("(%p) ret == %d, errno = %d. removing client\n",
thisfd, ret, errno);
DEBUGLOG("ret == %d, errno = %d. removing client\n",
ret, errno);
thisfd->removeme = 1;
continue;
}
/* New client...simply add it to the list */
if (newfd) {
_add_client(newfd, thisfd);
newfd->next = thisfd->next;
thisfd->next = newfd;
thisfd = newfd;
}
}
@@ -1003,8 +966,8 @@ static void main_loop(int cmd_timeout)
thisfd->bits.localsock.expected_replies !=
thisfd->bits.localsock.num_replies) {
/* Send timed out message + replies we already have */
DEBUGLOG("Request to client %p timed-out (send: %ld, now: %ld)\n",
thisfd, thisfd->bits.localsock.sent_time, the_time);
DEBUGLOG("Request timed-out (send: %ld, now: %ld)\n",
thisfd->bits.localsock.sent_time, the_time);
thisfd->bits.localsock.all_success = 0;
@@ -1105,31 +1068,31 @@ static void be_daemon(int timeout)
break;
default: /* Parent */
(void) close(devnull);
(void) close(child_pipe[1]);
wait_for_child(child_pipe[0], timeout); /* noreturn */
wait_for_child(child_pipe[0], timeout);
}
/* Detach ourself from the calling environment */
if ((dup2(devnull, STDIN_FILENO) == -1) ||
(dup2(devnull, STDOUT_FILENO) == -1) ||
(dup2(devnull, STDERR_FILENO) == -1)) {
if (close(0) || close(1) || close(2)) {
perror("Error closing terminal FDs");
exit(4);
}
setsid();
if (dup2(devnull, 0) < 0 || dup2(devnull, 1) < 0
|| dup2(devnull, 2) < 0) {
perror("Error setting terminal FDs to /dev/null");
log_error("Error setting terminal FDs to /dev/null: %m");
exit(5);
}
if ((devnull > STDERR_FILENO) && close(devnull)) {
log_sys_error("close", "/dev/null");
exit(7);
}
if (chdir("/")) {
log_error("Error setting current directory to /: %m");
exit(6);
}
setsid();
}
static int verify_message(char *buf, int len)
@@ -1216,8 +1179,8 @@ static int cleanup_zombie(struct local_client *thisfd)
if (!thisfd->bits.localsock.cleanup_needed)
return 0;
DEBUGLOG("(%p) EOF on local socket %d: inprogress=%d\n",
thisfd, thisfd->fd, thisfd->bits.localsock.in_progress);
DEBUGLOG("EOF on local socket: inprogress=%d\n",
thisfd->bits.localsock.in_progress);
if ((pipe_client = thisfd->bits.localsock.pipe_client))
pipe_client = pipe_client->bits.pipe.client;
@@ -1239,7 +1202,7 @@ static int cleanup_zombie(struct local_client *thisfd)
/* Kill the subthread & free resources */
if (thisfd->bits.localsock.threadid) {
DEBUGLOG("(%p) Waiting for pre&post thread\n", pipe_client);
DEBUGLOG("Waiting for pre&post thread (%p)\n", pipe_client);
pthread_mutex_lock(&thisfd->bits.localsock.mutex);
thisfd->bits.localsock.state = PRE_COMMAND;
thisfd->bits.localsock.finished = 1;
@@ -1250,22 +1213,26 @@ static int cleanup_zombie(struct local_client *thisfd)
(void **) &status)))
log_sys_error("pthread_join", "");
DEBUGLOG("(%p) Joined pre&post thread\n", pipe_client);
DEBUGLOG("Joined pre&post thread\n");
thisfd->bits.localsock.threadid = 0;
/* Remove the pipe client */
if (thisfd->bits.localsock.pipe_client) {
struct local_client *delfd = thisfd->bits.localsock.pipe_client;
struct local_client *delfd;
struct local_client *lastfd;
(void) close(delfd->fd); /* Close pipe */
(void) close(thisfd->bits.localsock.pipe_client->fd); /* Close pipe */
(void) close(thisfd->bits.localsock.pipe);
/* Remove pipe client */
if (!_del_client(delfd)) {
dm_free(delfd);
thisfd->bits.localsock.pipe_client = NULL;
}
for (lastfd = &local_client_head; (delfd = lastfd->next); lastfd = delfd)
if (thisfd->bits.localsock.pipe_client == delfd) {
thisfd->bits.localsock.pipe_client = NULL;
lastfd->next = delfd->next;
dm_free(delfd);
break;
}
}
}
@@ -1296,7 +1263,7 @@ static int read_from_local_sock(struct local_client *thisfd)
if (len == -1 && errno == EINTR)
return 1;
DEBUGLOG("(%p) Read on local socket %d, len = %d\n", thisfd, thisfd->fd, len);
DEBUGLOG("Read on local socket %d, len = %d\n", thisfd->fd, len);
if (len && verify_message(buffer, len) < 0) {
log_error("read_from_local_sock from %d len %d bad verify.",
@@ -1370,15 +1337,15 @@ static int read_from_local_sock(struct local_client *thisfd)
char *argptr = inheader->node + strlen(inheader->node) + 1;
while (missing_len > 0) {
DEBUGLOG("(%p) got %d bytes, need another %d (total %d)\n",
thisfd, argslen, missing_len, inheader->arglen);
DEBUGLOG("got %d bytes, need another %d (total %d)\n",
argslen, missing_len, inheader->arglen);
len = read(thisfd->fd, argptr + argslen, missing_len);
if (len == -1 && errno == EINTR)
continue;
if (len <= 0) {
/* EOF or error on socket */
DEBUGLOG("(%p) EOF on local socket\n", thisfd);
DEBUGLOG("EOF on local socket\n");
dm_free(thisfd->bits.localsock.cmd);
thisfd->bits.localsock.cmd = NULL;
return 0;
@@ -1406,7 +1373,7 @@ static int read_from_local_sock(struct local_client *thisfd)
.status = ENOENT
};
DEBUGLOG("(%p) Unknown node: '%s'\n", thisfd, inheader->node);
DEBUGLOG("Unknown node: '%s'\n", inheader->node);
send_message(&reply, sizeof(reply), our_csid, thisfd->fd,
"Error sending ENOENT reply to local user");
thisfd->bits.localsock.expected_replies = 0;
@@ -1432,7 +1399,7 @@ static int read_from_local_sock(struct local_client *thisfd)
.status = EBUSY
};
DEBUGLOG("(%p) Creating pipe failed: %s\n", thisfd, strerror(errno));
DEBUGLOG("Creating pipe failed: %s\n", strerror(errno));
send_message(&reply, sizeof(reply), our_csid, thisfd->fd,
"Error sending EBUSY reply to local user");
return len;
@@ -1452,7 +1419,7 @@ static int read_from_local_sock(struct local_client *thisfd)
return len;
}
DEBUGLOG("(%p) Creating pipe, [%d, %d]\n", thisfd, comms_pipe[0], comms_pipe[1]);
DEBUGLOG("Creating pipe, [%d, %d]\n", comms_pipe[0], comms_pipe[1]);
if (fcntl(comms_pipe[0], F_SETFD, 1))
DEBUGLOG("setting CLOEXEC on pipe[0] failed: %s\n", strerror(errno));
@@ -1463,8 +1430,8 @@ static int read_from_local_sock(struct local_client *thisfd)
newfd->type = THREAD_PIPE;
newfd->callback = local_pipe_callback;
newfd->bits.pipe.client = thisfd;
_add_client(newfd, thisfd);
newfd->next = thisfd->next;
thisfd->next = newfd;
/* Store a cross link to the pipe */
thisfd->bits.localsock.pipe_client = newfd;
@@ -1477,10 +1444,10 @@ static int read_from_local_sock(struct local_client *thisfd)
thisfd->bits.localsock.in_progress = TRUE;
thisfd->bits.localsock.state = PRE_COMMAND;
thisfd->bits.localsock.cleanup_needed = 1;
DEBUGLOG("(%p) Creating pre&post thread for pipe fd %d\n", newfd, newfd->fd);
DEBUGLOG("Creating pre&post thread for pipe fd %d (%p)\n", newfd->fd, newfd);
status = pthread_create(&thisfd->bits.localsock.threadid,
&stack_attr, pre_and_post_thread, thisfd);
DEBUGLOG("(%p) Created pre&post thread, state = %d\n", newfd, status);
DEBUGLOG("Created pre&post thread, state = %d\n", status);
return len;
}
@@ -1488,6 +1455,13 @@ static int read_from_local_sock(struct local_client *thisfd)
/* Add a file descriptor from the cluster or comms interface to
our list of FDs for select
*/
int add_client(struct local_client *new_client)
{
new_client->next = local_client_head.next;
local_client_head.next = new_client;
return 0;
}
/* Called when the pre-command has completed successfully - we
now execute the real command on all the requested nodes */
@@ -1498,8 +1472,8 @@ static int distribute_command(struct local_client *thisfd)
int len = thisfd->bits.localsock.cmd_len;
thisfd->xid = global_xid++;
DEBUGLOG("(%p) distribute command: XID = %d, flags=0x%x (%s%s)\n",
thisfd, thisfd->xid, inheader->flags,
DEBUGLOG("distribute command: XID = %d, flags=0x%x (%s%s)\n",
thisfd->xid, inheader->flags,
(inheader->flags & CLVMD_FLAG_LOCAL) ? "LOCAL" : "",
(inheader->flags & CLVMD_FLAG_REMOTE) ? "REMOTE" : "");
@@ -1521,7 +1495,7 @@ static int distribute_command(struct local_client *thisfd)
*/
add_to_lvmqueue(thisfd, inheader, len, NULL);
DEBUGLOG("(%p) Sending message to all cluster nodes\n", thisfd);
DEBUGLOG("Sending message to all cluster nodes\n");
inheader->xid = thisfd->xid;
send_message(inheader, len, NULL, -1,
"Error forwarding message to cluster");
@@ -1540,11 +1514,11 @@ static int distribute_command(struct local_client *thisfd)
/* Are we the requested node ?? */
if (memcmp(csid, our_csid, max_csid_len) == 0) {
DEBUGLOG("(%p) Doing command on local node only\n", thisfd);
DEBUGLOG("Doing command on local node only\n");
add_to_lvmqueue(thisfd, inheader, len, NULL);
} else {
DEBUGLOG("(%p) Sending message to single node: %s\n",
thisfd, inheader->node);
DEBUGLOG("Sending message to single node: %s\n",
inheader->node);
inheader->xid = thisfd->xid;
send_message(inheader, len, csid, -1,
"Error forwarding message to cluster node");
@@ -1555,7 +1529,7 @@ static int distribute_command(struct local_client *thisfd)
thisfd->bits.localsock.in_progress = TRUE;
thisfd->bits.localsock.expected_replies = 1;
thisfd->bits.localsock.num_replies = 0;
DEBUGLOG("(%p) Doing command explicitly on local node only\n", thisfd);
DEBUGLOG("Doing command explicitly on local node only\n");
add_to_lvmqueue(thisfd, inheader, len, NULL);
}
@@ -1681,7 +1655,7 @@ static void add_reply_to_list(struct local_client *client, int status,
reply->status = status;
clops->name_from_csid(csid, reply->node);
DEBUGLOG("(%p) Reply from node %s: %d bytes\n", client, reply->node, len);
DEBUGLOG("Reply from node %s: %d bytes\n", reply->node, len);
if (len > 0) {
if (!(reply->replymsg = dm_malloc(len)))
@@ -1708,8 +1682,8 @@ static void add_reply_to_list(struct local_client *client, int status,
client->bits.localsock.state = POST_COMMAND;
pthread_cond_signal(&client->bits.localsock.cond);
}
DEBUGLOG("(%p) Got %d replies, expecting: %d\n",
client, client->bits.localsock.num_replies,
DEBUGLOG("Got %d replies, expecting: %d\n",
client->bits.localsock.num_replies,
client->bits.localsock.expected_replies);
}
pthread_mutex_unlock(&client->bits.localsock.mutex);
@@ -1724,7 +1698,7 @@ static __attribute__ ((noreturn)) void *pre_and_post_thread(void *arg)
sigset_t ss;
int pipe_fd = client->bits.localsock.pipe;
DEBUGLOG("(%p) Pre&post thread pipe fd %d\n", client, pipe_fd);
DEBUGLOG("Pre&post thread (%p), pipe fd %d\n", client, pipe_fd);
pthread_mutex_lock(&client->bits.localsock.mutex);
/* Ignore SIGUSR1 (handled by master process) but enable
@@ -1744,7 +1718,7 @@ static __attribute__ ((noreturn)) void *pre_and_post_thread(void *arg)
if ((status = do_pre_command(client)))
client->bits.localsock.all_success = 0;
DEBUGLOG("(%p) Pre&post thread writes status %d down to pipe fd %d\n",
DEBUGLOG("Pre&post thread (%p) writes status %d down to pipe fd %d\n",
client, status, pipe_fd);
/* Tell the parent process we have finished this bit */
@@ -1762,13 +1736,13 @@ static __attribute__ ((noreturn)) void *pre_and_post_thread(void *arg)
/* We may need to wait for the condition variable before running the post command */
if (client->bits.localsock.state != POST_COMMAND &&
!client->bits.localsock.finished) {
DEBUGLOG("(%p) Pre&post thread waiting to do post command, state = %d\n",
DEBUGLOG("Pre&post thread (%p) waiting to do post command, state = %d\n",
client, client->bits.localsock.state);
pthread_cond_wait(&client->bits.localsock.cond,
&client->bits.localsock.mutex);
}
DEBUGLOG("(%p) Pre&post thread got post command condition...\n", client);
DEBUGLOG("Pre&post thread (%p) got post command condition...\n", client);
/* POST function must always run, even if the client aborts */
status = 0;
@@ -1782,15 +1756,15 @@ static __attribute__ ((noreturn)) void *pre_and_post_thread(void *arg)
next_pre:
if (client->bits.localsock.state != PRE_COMMAND &&
!client->bits.localsock.finished) {
DEBUGLOG("(%p) Pre&post thread waiting for next pre command\n", client);
DEBUGLOG("Pre&post thread (%p) waiting for next pre command\n", client);
pthread_cond_wait(&client->bits.localsock.cond,
&client->bits.localsock.mutex);
}
DEBUGLOG("(%p) Pre&post thread got pre command condition...\n", client);
DEBUGLOG("Pre&post thread (%p) got pre command condition...\n", client);
}
pthread_mutex_unlock(&client->bits.localsock.mutex);
DEBUGLOG("(%p) Pre&post thread finished\n", client);
DEBUGLOG("Pre&post thread (%p) finished\n", client);
pthread_exit(NULL);
}
@@ -1808,8 +1782,8 @@ static int process_local_command(struct clvm_header *msg, int msglen,
if (!(replybuf = dm_malloc(max_cluster_message)))
return -1;
DEBUGLOG("(%p) process_local_command: %s msg=%p, msglen =%d\n",
client, decode_cmd(msg->cmd), msg, msglen);
DEBUGLOG("process_local_command: %s msg=%p, msglen =%d, client=%p\n",
decode_cmd(msg->cmd), msg, msglen, client);
/* If remote flag is set, just set a successful status code. */
if (msg->flags & CLVMD_FLAG_REMOTE)
@@ -1824,8 +1798,8 @@ static int process_local_command(struct clvm_header *msg, int msglen,
if (xid == client->xid)
add_reply_to_list(client, status, our_csid, replybuf, replylen);
else
DEBUGLOG("(%p) Local command took too long, discarding xid %d, current is %d\n",
client, xid, client->xid);
DEBUGLOG("Local command took too long, discarding xid %d, current is %d\n",
xid, client->xid);
dm_free(replybuf);
@@ -1867,7 +1841,7 @@ static void send_local_reply(struct local_client *client, int status, int fd)
char *ptr;
int message_len = 0;
DEBUGLOG("(%p) Send local reply\n", client);
DEBUGLOG("Send local reply\n");
/* Work out the total size of the reply */
while (thisreply) {
@@ -1884,7 +1858,7 @@ static void send_local_reply(struct local_client *client, int status, int fd)
/* Add in the size of our header */
message_len = message_len + sizeof(struct clvm_header);
if (!(replybuf = dm_malloc(message_len))) {
DEBUGLOG("(%p) Memory allocation fails\n", client);
DEBUGLOG("Memory allocation fails\n");
return;
}
@@ -2013,7 +1987,6 @@ static int send_message(void *buf, int msglen, const char *csid, int fd,
(void) nanosleep (&delay, &remtime);
continue;
}
DEBUGLOG("%s", errtext);
log_error("%s", errtext);
break;
}
@@ -2027,7 +2000,7 @@ static int process_work_item(struct lvm_thread_cmd *cmd)
{
/* If msg is NULL then this is a cleanup request */
if (cmd->msg == NULL) {
DEBUGLOG("(%p) process_work_item: free\n", cmd->client);
DEBUGLOG("process_work_item: free %p\n", cmd->client);
cmd_client_cleanup(cmd->client);
pthread_mutex_destroy(&cmd->client->bits.localsock.mutex);
pthread_cond_destroy(&cmd->client->bits.localsock.cond);
@@ -2036,11 +2009,11 @@ static int process_work_item(struct lvm_thread_cmd *cmd)
}
if (!cmd->remote) {
DEBUGLOG("(%p) process_work_item: local\n", cmd->client);
DEBUGLOG("process_work_item: local\n");
process_local_command(cmd->msg, cmd->msglen, cmd->client,
cmd->xid);
} else {
DEBUGLOG("(%p) process_work_item: remote\n", cmd->client);
DEBUGLOG("process_work_item: remote\n");
process_remote_command(cmd->msg, cmd->msglen, cmd->client->fd,
cmd->csid);
}
@@ -2134,8 +2107,8 @@ static int add_to_lvmqueue(struct local_client *client, struct clvm_header *msg,
} else
cmd->remote = 0;
DEBUGLOG("(%p) add_to_lvmqueue: cmd=%p, msg=%p, len=%d, csid=%p, xid=%d\n",
client, cmd, msg, msglen, csid, cmd->xid);
DEBUGLOG("add_to_lvmqueue: cmd=%p. client=%p, msg=%p, len=%d, csid=%p, xid=%d\n",
cmd, client, msg, msglen, csid, cmd->xid);
pthread_mutex_lock(&lvm_thread_mutex);
if (lvm_thread_exit) {
pthread_mutex_unlock(&lvm_thread_mutex);
@@ -2151,14 +2124,6 @@ static int add_to_lvmqueue(struct local_client *client, struct clvm_header *msg,
}
/* Return 0 if we can talk to an existing clvmd */
/*
* FIXME:
*
* This function returns only -1 or 0, but there are
* different levels of errors, some of them should stop
* further execution of clvmd thus another state is needed
* and some error message need to be only informational.
*/
static int check_local_clvmd(void)
{
int local_socket;
@@ -2178,11 +2143,7 @@ static int check_local_clvmd(void)
if (connect(local_socket,(struct sockaddr *) &sockaddr,
sizeof(sockaddr))) {
/* connection failure is expected state */
if (errno == ENOENT)
log_sys_debug("connect", "local socket");
else
log_sys_error("connect", "local socket");
log_sys_error("connect", "local socket");
ret = -1;
}
@@ -2283,8 +2244,7 @@ static void check_all_callback(struct local_client *client, const char *csid,
If not, returns -1 and prints out a list of errant nodes */
static int check_all_clvmds_running(struct local_client *client)
{
DEBUGLOG("(%p) check_all_clvmds_running\n", client);
DEBUGLOG("check_all_clvmds_running\n");
return clops->cluster_do_node_callback(client, check_all_callback);
}
@@ -2323,11 +2283,13 @@ static void ntoh_clvm(struct clvm_header *hdr)
static void sigusr2_handler(int sig)
{
DEBUGLOG("SIGUSR2 received\n");
return;
}
static void sigterm_handler(int sig)
{
quit = 1;
return;
}
static void sighup_handler(int sig)

View File

@@ -661,8 +661,6 @@ int do_refresh_cache(void)
return -1;
}
cmd->use_aio = 0;
cmd->use_scan_cache = 0;
init_full_scan_done(0);
init_ignore_suspended_devices(1);
lvmcache_force_next_label_scan();
@@ -922,8 +920,6 @@ int init_clvm(struct dm_hash_table *excl_uuid)
/* Check lvm.conf is setup for cluster-LVM */
check_config();
init_ignore_suspended_devices(1);
cmd->use_aio = 0;
cmd->use_scan_cache = 0;
/* Trap log messages so we can pass them back to the user */
init_log_fn(lvm2_log_fn);

View File

@@ -29,7 +29,7 @@ include $(top_builddir)/make.tmpl
LIBS += -ldevmapper
LMLIBS += $(CPG_LIBS) $(SACKPT_LIBS)
CFLAGS += $(CPG_CFLAGS) $(SACKPT_CFLAGS) $(EXTRA_EXEC_CFLAGS)
LDFLAGS += $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
LDFLAGS += $(EXTRA_EXEC_LDFLAGS)
cmirrord: $(OBJECTS) $(top_builddir)/lib/liblvm-internal.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) \

View File

@@ -182,7 +182,7 @@ int cluster_send(struct clog_request *rq)
}
/*
* Once the request heads for the cluster, the luid loses
* Once the request heads for the cluster, the luid looses
* all its meaning.
*/
rq->u_rq.luid = 0;

View File

@@ -377,7 +377,7 @@ static int _clog_ctr(char *uuid, uint64_t luid,
uint32_t block_on_error = 0;
int disk_log;
char disk_path[PATH_MAX];
char disk_path[128];
int unlink_path = 0;
long page_size;
int pages;

View File

@@ -56,16 +56,18 @@ include $(top_builddir)/make.tmpl
all: device-mapper
device-mapper: $(TARGETS)
LIBS += -ldevmapper
LVMLIBS += -ldevmapper-event $(PTHREAD_LIBS)
CFLAGS_dmeventd.o += $(EXTRA_EXEC_CFLAGS)
LIBS += -ldevmapper $(PTHREAD_LIBS)
dmeventd: $(LIB_SHARED) dmeventd.o
$(CC) $(CFLAGS) -L. $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) dmeventd.o \
-o $@ $(DL_LIBS) $(DMEVENT_LIBS) $(LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) -L. -o $@ dmeventd.o \
$(DL_LIBS) $(LVMLIBS) $(LIBS) -rdynamic
dmeventd.static: $(LIB_STATIC) dmeventd.o $(interfacebuilddir)/libdevmapper.a
$(CC) $(CFLAGS) $(LDFLAGS) -static -L. -L$(interfacebuilddir) dmeventd.o \
-o $@ $(DL_LIBS) $(DMEVENT_LIBS) $(LIBS) $(STATIC_LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) $(ELDFLAGS) -static -L. -L$(interfacebuilddir) -o $@ \
dmeventd.o $(DL_LIBS) $(LVMLIBS) $(LIBS) $(STATIC_LIBS)
ifeq ("@PKGCONFIG@", "yes")
INSTALL_LIB_TARGETS += install_pkgconfig

View File

@@ -62,8 +62,6 @@
#include <syslog.h>
#define DM_SIGNALED_EXIT 1
#define DM_SCHEDULED_EXIT 2
static volatile sig_atomic_t _exit_now = 0; /* set to '1' when signal is given to exit */
/* List (un)link macros. */
@@ -470,7 +468,7 @@ static int _pthread_create_smallstack(pthread_t *t, void *(*fun)(void *), void *
/*
* We use a smaller stack since it gets preallocated in its entirety
*/
pthread_attr_setstacksize(&attr, THREAD_STACK_SIZE + getpagesize());
pthread_attr_setstacksize(&attr, THREAD_STACK_SIZE);
/*
* If no-one will be waiting, we need to detach.
@@ -865,7 +863,6 @@ static int _event_wait(struct thread_status *thread)
* This is so that you can break out of waiting on an event,
* either for a timeout event, or to cancel the thread.
*/
sigemptyset(&old);
sigemptyset(&set);
sigaddset(&set, SIGALRM);
if (pthread_sigmask(SIG_UNBLOCK, &set, &old) != 0) {
@@ -1753,7 +1750,7 @@ static void _init_thread_signals(void)
*/
static void _exit_handler(int sig __attribute__((unused)))
{
_exit_now = DM_SIGNALED_EXIT;
_exit_now = 1;
}
#ifdef __linux__
@@ -2251,14 +2248,11 @@ int main(int argc, char *argv[])
for (;;) {
if (_idle_since) {
if (_exit_now) {
if (_exit_now == DM_SCHEDULED_EXIT)
break; /* Only prints shutdown message */
log_info("dmeventd detected break while being idle "
"for %ld second(s), exiting.",
(long) (time(NULL) - _idle_since));
break;
}
if (idle_exit_timeout) {
} else if (idle_exit_timeout) {
now = time(NULL);
if (now < _idle_since)
_idle_since = now; /* clock change? */
@@ -2269,14 +2263,15 @@ int main(int argc, char *argv[])
break;
}
}
} else if (_exit_now == DM_SIGNALED_EXIT) {
_exit_now = DM_SCHEDULED_EXIT;
} else if (_exit_now) {
_exit_now = 0;
/*
* When '_exit_now' is set, signal has been received,
* but can not simply exit unless all
* threads are done processing.
*/
log_info("dmeventd received break, scheduling exit.");
log_warn("WARNING: There are still devices being monitored.");
log_warn("WARNING: Refusing to exit.");
}
_process_request(&fifos);
_cleanup_unused_threads();

View File

@@ -250,9 +250,10 @@ static int _daemon_read(struct dm_event_fifos *fifos,
if (ret < 0) {
if ((errno == EINTR) || (errno == EAGAIN))
continue;
log_error("Unable to read from event server.");
return 0;
else {
log_error("Unable to read from event server.");
return 0;
}
}
bytes += ret;
@@ -328,9 +329,10 @@ static int _daemon_write(struct dm_event_fifos *fifos,
if (ret < 0) {
if ((errno == EINTR) || (errno == EAGAIN))
continue;
log_error("Unable to talk to event daemon.");
return 0;
else {
log_error("Unable to talk to event daemon.");
return 0;
}
}
bytes += ret;
@@ -452,8 +454,7 @@ static int _start_daemon(char *dmeventd_path, struct dm_event_fifos *fifos)
if (close(fifos->client))
log_sys_debug("close", fifos->client_path);
return 1;
}
if (errno != ENXIO && errno != ENOENT) {
} else if (errno != ENXIO && errno != ENOENT) {
/* problem */
log_sys_error("open", fifos->client_path);
return 0;

View File

@@ -1,5 +1,5 @@
/*
* Copyright (C) 2005-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2005-2015 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -25,6 +25,7 @@
struct dso_state {
struct dm_pool *mem;
char cmd_lvscan[512];
char cmd_lvconvert[512];
};
@@ -98,8 +99,12 @@ static int _get_mirror_event(struct dso_state *state, char *params)
return r;
}
static int _remove_failed_devices(const char *cmd_lvconvert, const char *device)
static int _remove_failed_devices(const char *cmd_lvscan, const char *cmd_lvconvert,
const char *device)
{
if (!dmeventd_lvm2_run_with_lock(cmd_lvscan))
log_warn("WARNING: Re-scan of mirrored device %s failed.", device);
/* if repair goes OK, report success even if lvscan has failed */
if (!dmeventd_lvm2_run_with_lock(cmd_lvconvert)) {
log_error("Repair of mirrored device %s failed.", device);
@@ -146,7 +151,9 @@ void process_event(struct dm_task *dmt,
break;
case ME_FAILURE:
log_error("Device failure in %s.", device);
if (!_remove_failed_devices(state->cmd_lvconvert, device))
if (!_remove_failed_devices(state->cmd_lvscan,
state->cmd_lvconvert,
device))
/* FIXME Why are all the error return codes unused? Get rid of them? */
log_error("Failed to remove faulty devices in %s.",
device);
@@ -176,10 +183,17 @@ int register_device(const char *device,
if (!dmeventd_lvm2_init_with_pool("mirror_state", state))
goto_bad;
/* CANNOT use --config as this disables cached content */
if (!dmeventd_lvm2_command(state->mem, state->cmd_lvconvert, sizeof(state->cmd_lvconvert),
"lvconvert --repair --use-policies", device))
if (!dmeventd_lvm2_command(state->mem, state->cmd_lvscan, sizeof(state->cmd_lvscan),
"lvscan --cache", device)) {
dmeventd_lvm2_exit_with_pool(state);
goto_bad;
}
if (!dmeventd_lvm2_command(state->mem, state->cmd_lvconvert, sizeof(state->cmd_lvconvert),
"lvconvert --repair --use-policies", device)) {
dmeventd_lvm2_exit_with_pool(state);
goto_bad;
}
*user = state;
@@ -189,9 +203,6 @@ int register_device(const char *device,
bad:
log_error("Failed to monitor mirror %s.", device);
if (state)
dmeventd_lvm2_exit_with_pool(state);
return 0;
}

View File

@@ -1,5 +1,5 @@
/*
* Copyright (C) 2005-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2005-2016 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -22,6 +22,7 @@
struct dso_state {
struct dm_pool *mem;
char cmd_lvscan[512];
char cmd_lvconvert[512];
uint64_t raid_devs[RAID_DEVS_ELEMS];
int failed;
@@ -58,23 +59,6 @@ static int _process_raid_event(struct dso_state *state, char *params, const char
dead = 1;
}
/*
* if we are converting from non-RAID to RAID (e.g. linear -> raid1)
* and too many original devices die, such that we cannot continue
* the "recover" operation, the sync action will go to "idle", the
* unsynced devs will remain at 'a', and the original devices will
* NOT SWITCH TO 'D', but will remain at 'A' - hoping to be revived.
*
* This is simply the way the kernel works...
*/
if (!strcmp(status->sync_action, "idle") &&
(status->dev_health[0] == 'a') &&
(status->insync_regions < status->total_regions)) {
log_error("Primary sources for new RAID, %s, have failed.",
device);
dead = 1; /* run it through LVM repair */
}
if (dead) {
if (status->insync_regions < status->total_regions) {
if (!state->warned) {
@@ -90,6 +74,8 @@ static int _process_raid_event(struct dso_state *state, char *params, const char
goto out; /* already reported */
state->failed = 1;
if (!dmeventd_lvm2_run_with_lock(state->cmd_lvscan))
log_warn("WARNING: Re-scan of RAID device %s failed.", device);
/* if repair goes OK, report success even if lvscan has failed */
if (!dmeventd_lvm2_run_with_lock(state->cmd_lvconvert)) {
@@ -98,8 +84,6 @@ static int _process_raid_event(struct dso_state *state, char *params, const char
}
} else {
state->failed = 0;
if (status->insync_regions == status->total_regions)
memset(&state->raid_devs, 0, sizeof(state->raid_devs));
log_info("%s array, %s, is %s in-sync.",
status->raid_type, device,
(status->insync_regions == status->total_regions) ? "now" : "not");
@@ -152,9 +136,14 @@ int register_device(const char *device,
if (!dmeventd_lvm2_init_with_pool("raid_state", state))
goto_bad;
if (!dmeventd_lvm2_command(state->mem, state->cmd_lvconvert, sizeof(state->cmd_lvconvert),
"lvconvert --repair --use-policies", device))
if (!dmeventd_lvm2_command(state->mem, state->cmd_lvscan, sizeof(state->cmd_lvscan),
"lvscan --cache", device) ||
!dmeventd_lvm2_command(state->mem, state->cmd_lvconvert, sizeof(state->cmd_lvconvert),
"lvconvert --config devices{ignore_suspended_devices=1} "
"--repair --use-policies", device)) {
dmeventd_lvm2_exit_with_pool(state);
goto_bad;
}
*user = state;
@@ -164,9 +153,6 @@ int register_device(const char *device,
bad:
log_error("Failed to monitor RAID %s.", device);
if (state)
dmeventd_lvm2_exit_with_pool(state);
return 0;
}

View File

@@ -231,7 +231,7 @@ void process_event(struct dm_task *dmt,
if (percent >= WARNING_THRESH) /* Print a warning to syslog. */
log_warn("WARNING: Snapshot %s is now %.2f%% full.",
device, dm_percent_to_round_float(percent, 2));
device, dm_percent_to_float(percent));
/* Try to extend the snapshot, in accord with user-set policies */
if (!_extend(state->cmd_lvextend))
@@ -254,8 +254,10 @@ int register_device(const char *device,
if (!dmeventd_lvm2_command(state->mem, state->cmd_lvextend,
sizeof(state->cmd_lvextend),
"lvextend --use-policies", device))
"lvextend --use-policies", device)) {
dmeventd_lvm2_exit_with_pool(state);
goto_bad;
}
state->percent_check = CHECK_MINIMUM;
*user = state;
@@ -266,9 +268,6 @@ int register_device(const char *device,
bad:
log_error("Failed to monitor snapshot %s.", device);
if (state)
dmeventd_lvm2_exit_with_pool(state);
return 0;
}

View File

@@ -18,6 +18,7 @@
#include <sys/wait.h>
#include <stdarg.h>
#include <pthread.h>
/* TODO - move this mountinfo code into library to be reusable */
#ifdef __linux__
@@ -47,8 +48,10 @@ struct dso_state {
struct dm_pool *mem;
int metadata_percent_check;
int metadata_percent;
int metadata_warn_once;
int data_percent_check;
int data_percent;
int data_warn_once;
uint64_t known_metadata_size;
uint64_t known_data_size;
unsigned fails;
@@ -56,8 +59,8 @@ struct dso_state {
int restore_sigset;
sigset_t old_sigset;
pid_t pid;
char *argv[3];
char *cmd_str;
char **argv;
char cmd_str[1024];
};
DM_EVENT_LOG_FN("thin")
@@ -83,7 +86,7 @@ static int _run_command(struct dso_state *state)
} else {
/* For an error event it's for a user to check status and decide */
env[1] = NULL;
log_debug("Error event processing.");
log_debug("Error event processing");
}
log_verbose("Executing command: %s", state->cmd_str);
@@ -113,7 +116,7 @@ static int _use_policy(struct dm_task *dmt, struct dso_state *state)
#if THIN_DEBUG
log_debug("dmeventd executes: %s.", state->cmd_str);
#endif
if (state->argv[0])
if (state->argv)
return _run_command(state);
if (!dmeventd_lvm2_run_with_lock(state->cmd_str)) {
@@ -172,8 +175,8 @@ void process_event(struct dm_task *dmt,
#if THIN_DEBUG
log_debug("Watch for tp-data:%.2f%% tp-metadata:%.2f%%.",
dm_percent_to_round_float(state->data_percent_check, 2),
dm_percent_to_round_float(state->metadata_percent_check, 2));
dm_percent_to_float(state->data_percent_check),
dm_percent_to_float(state->metadata_percent_check));
#endif
if (!_wait_for_pid(state)) {
log_warn("WARNING: Skipping event, child %d is still running (%s).",
@@ -251,10 +254,11 @@ void process_event(struct dm_task *dmt,
* action is called for: >50%, >55% ... >95%, 100%
*/
state->metadata_percent = dm_make_percent(tps->used_metadata_blocks, tps->total_metadata_blocks);
if ((state->metadata_percent > WARNING_THRESH) &&
(state->metadata_percent > state->metadata_percent_check))
if (state->metadata_percent <= WARNING_THRESH)
state->metadata_warn_once = 0; /* Dropped bellow threshold, reset warn once */
else if (!state->metadata_warn_once++) /* Warn once when raised above threshold */
log_warn("WARNING: Thin pool %s metadata is now %.2f%% full.",
device, dm_percent_to_round_float(state->metadata_percent, 2));
device, dm_percent_to_float(state->metadata_percent));
if (state->metadata_percent > CHECK_MINIMUM) {
/* Run action when usage raised more than CHECK_STEP since the last time */
if (state->metadata_percent > state->metadata_percent_check)
@@ -266,10 +270,11 @@ void process_event(struct dm_task *dmt,
state->metadata_percent_check = CHECK_MINIMUM;
state->data_percent = dm_make_percent(tps->used_data_blocks, tps->total_data_blocks);
if ((state->data_percent > WARNING_THRESH) &&
(state->data_percent > state->data_percent_check))
if (state->data_percent <= WARNING_THRESH)
state->data_warn_once = 0;
else if (!state->data_warn_once++)
log_warn("WARNING: Thin pool %s data is now %.2f%% full.",
device, dm_percent_to_round_float(state->data_percent, 2));
device, dm_percent_to_float(state->data_percent));
if (state->data_percent > CHECK_MINIMUM) {
/* Run action when usage raised more than CHECK_STEP since the last time */
if (state->data_percent > state->data_percent_check)
@@ -348,41 +353,34 @@ int register_device(const char *device,
void **user)
{
struct dso_state *state;
int maxcmd;
char *str;
char cmd_str[PATH_MAX + 128 + 2]; /* cmd ' ' vg/lv \0 */
if (!dmeventd_lvm2_init_with_pool("thin_pool_state", state))
goto_bad;
if (!dmeventd_lvm2_command(state->mem, cmd_str, sizeof(cmd_str),
"_dmeventd_thin_command", device))
if (!dmeventd_lvm2_command(state->mem, state->cmd_str,
sizeof(state->cmd_str),
"_dmeventd_thin_command", device)) {
dmeventd_lvm2_exit_with_pool(state);
goto_bad;
}
if (strncmp(cmd_str, "lvm ", 4) == 0) {
if (!(state->cmd_str = dm_pool_strdup(state->mem, cmd_str + 4))) {
log_error("Failed to copy lvm command.");
goto bad;
}
} else if (cmd_str[0] == '/') {
if (!(state->cmd_str = dm_pool_strdup(state->mem, cmd_str))) {
log_error("Failed to copy thin command.");
if (strncmp(state->cmd_str, "lvm ", 4)) {
maxcmd = 2; /* space for last NULL element */
for (str = state->cmd_str; *str; str++)
if (*str == ' ')
maxcmd++;
if (!(str = dm_pool_strdup(state->mem, state->cmd_str)) ||
!(state->argv = dm_pool_zalloc(state->mem, maxcmd * sizeof(char *)))) {
log_error("Failed to allocate memory for command.");
goto bad;
}
/* Find last space before 'vg/lv' */
if (!(str = strrchr(state->cmd_str, ' ')))
goto inval;
if (!(state->argv[0] = dm_pool_strndup(state->mem, state->cmd_str,
str - state->cmd_str))) {
log_error("Failed to copy command.");
goto bad;
}
state->argv[1] = str + 1; /* 1 argument - vg/lv */
dm_split_words(str, maxcmd - 1, 0, state->argv);
_init_thread_signals(state);
} else /* Unuspported command format */
goto inval;
} else
memmove(state->cmd_str, state->cmd_str + 4, strlen(state->cmd_str + 4) + 1);
state->pid = -1;
*user = state;
@@ -390,14 +388,9 @@ int register_device(const char *device,
log_info("Monitoring thin pool %s.", device);
return 1;
inval:
log_error("Invalid command for monitoring: %s.", cmd_str);
bad:
log_error("Failed to monitor thin pool %s.", device);
if (state)
dmeventd_lvm2_exit_with_pool(state);
return 0;
}

View File

@@ -1 +0,0 @@
dmfilemapd

View File

@@ -1,66 +0,0 @@
#
# Copyright (C) 2016 Red Hat, Inc. All rights reserved.
#
# This file is part of the device-mapper userspace tools.
#
# This copyrighted material is made available to anyone wishing to use,
# modify, copy, or redistribute it subject to the terms and conditions
# of the GNU Lesser General Public License v.2.1.
#
# You should have received a copy of the GNU Lesser General Public License
# along with this program; if not, write to the Free Software Foundation,
# Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
srcdir = @srcdir@
top_srcdir = @top_srcdir@
top_builddir = @top_builddir@
SOURCES = dmfilemapd.c
TARGETS = dmfilemapd
.PHONY: install_dmfilemapd install_dmfilemapd_static
INSTALL_DMFILEMAPD_TARGETS = install_dmfilemapd_dynamic
CLEAN_TARGETS = dmfilemapd.static
CFLOW_LIST = $(SOURCES)
CFLOW_LIST_TARGET = $(LIB_NAME).cflow
CFLOW_TARGET = dmfilemapd
include $(top_builddir)/make.tmpl
all: device-mapper
device-mapper: $(TARGETS)
CFLAGS_dmfilemapd.o += $(EXTRA_EXEC_CFLAGS)
LIBS += -ldevmapper
dmfilemapd: $(LIB_SHARED) dmfilemapd.o
$(CC) $(CFLAGS) $(LDFLAGS) $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS) \
-o $@ dmfilemapd.o $(DL_LIBS) $(LIBS)
dmfilemapd.static: $(LIB_STATIC) dmfilemapd.o $(interfacebuilddir)/libdevmapper.a
$(CC) $(CFLAGS) $(LDFLAGS) $(ELDFLAGS) -static -L$(interfacebuilddir) \
-o $@ dmfilemapd.o $(DL_LIBS) $(LIBS) $(STATIC_LIBS)
ifneq ("$(CFLOW_CMD)", "")
CFLOW_SOURCES = $(addprefix $(srcdir)/, $(SOURCES))
-include $(top_builddir)/libdm/libdevmapper.cflow
-include $(top_builddir)/lib/liblvm-internal.cflow
-include $(top_builddir)/lib/liblvm2cmd.cflow
-include $(top_builddir)/daemons/dmfilemapd/$(LIB_NAME).cflow
endif
install_dmfilemapd_dynamic: dmfilemapd
$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)
install_dmfilemapd_static: dmfilemapd.static
$(INSTALL_PROGRAM) -D $< $(staticdir)/$(<F)
install_dmfilemapd: $(INSTALL_DMFILEMAPD_TARGETS)
install: install_dmfilemapd
install_device-mapper: install_dmfilemapd

View File

@@ -1,834 +0,0 @@
/*
* Copyright (C) 2016 Red Hat, Inc. All rights reserved.
*
* This file is part of the device-mapper userspace tools.
*
* It includes tree drawing code based on pstree: http://psmisc.sourceforge.net/
*
* This copyrighted material is made available to anyone wishing to use,
* modify, copy, or redistribute it subject to the terms and conditions
* of the GNU General Public License v.2.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "tool.h"
#include "dm-logging.h"
#include "defaults.h"
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/inotify.h>
#include <dirent.h>
#include <ctype.h>
#ifdef __linux__
# include "kdev_t.h"
#else
# define MAJOR(x) major((x))
# define MINOR(x) minor((x))
# define MKDEV(x,y) makedev((x),(y))
#endif
/* limit to two updates/sec */
#define FILEMAPD_WAIT_USECS 500000
/* how long to wait for unlinked files */
#define FILEMAPD_NOFILE_WAIT_USECS 100000
#define FILEMAPD_NOFILE_WAIT_TRIES 10
struct filemap_monitor {
dm_filemapd_mode_t mode;
const char *program_id;
uint64_t group_id;
char *path;
int fd;
int inotify_fd;
int inotify_watch_fd;
/* monitoring heuristics */
int64_t blocks; /* allocated blocks, from stat.st_blocks */
uint64_t nr_regions;
int deleted;
};
static int _foreground;
static int _verbose;
const char *const _usage = "dmfilemapd <fd> <group_id> <abs_path> <mode> "
"[<foreground>[<log_level>]]";
/*
* Daemon logging. By default, all messages are thrown away: messages
* are only written to the terminal if the daemon is run in the foreground.
*/
__attribute__((format(printf, 5, 0)))
static void _dmfilemapd_log_line(int level,
const char *file __attribute__((unused)),
int line __attribute__((unused)),
int dm_errno_or_class,
const char *f, va_list ap)
{
static int _abort_on_internal_errors = -1;
FILE *out = log_stderr(level) ? stderr : stdout;
level = log_level(level);
if (level <= _LOG_WARN || _verbose) {
if (level < _LOG_WARN)
out = stderr;
vfprintf(out, f, ap);
fputc('\n', out);
}
if (_abort_on_internal_errors < 0)
/* Set when env DM_ABORT_ON_INTERNAL_ERRORS is not "0" */
_abort_on_internal_errors =
strcmp(getenv("DM_ABORT_ON_INTERNAL_ERRORS") ? : "0", "0");
if (_abort_on_internal_errors &&
!strncmp(f, INTERNAL_ERROR, sizeof(INTERNAL_ERROR) - 1))
abort();
}
__attribute__((format(printf, 5, 6)))
static void _dmfilemapd_log_with_errno(int level,
const char *file, int line,
int dm_errno_or_class,
const char *f, ...)
{
va_list ap;
va_start(ap, f);
_dmfilemapd_log_line(level, file, line, dm_errno_or_class, f, ap);
va_end(ap);
}
/*
* Only used for reporting errors before daemonise().
*/
__attribute__((format(printf, 1, 2)))
static void _early_log(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
vfprintf(stderr, fmt, ap);
fputc('\n', stderr);
va_end(ap);
}
static void _setup_logging(void)
{
dm_log_init_verbose(_verbose - 1);
dm_log_with_errno_init(_dmfilemapd_log_with_errno);
}
#define PROC_FD_DELETED_STR "(deleted)"
/*
* Scan the /proc/<pid>/fd directory for pid and check for an fd
* symlink whose contents match path.
*/
static int _is_open_in_pid(pid_t pid, const char *path)
{
char deleted_path[PATH_MAX + sizeof(PROC_FD_DELETED_STR)];
struct dirent *pid_dp = NULL;
char path_buf[PATH_MAX];
char link_buf[PATH_MAX];
DIR *pid_d = NULL;
ssize_t len;
if (pid == getpid())
return 0;
if (dm_snprintf(path_buf, sizeof(path_buf),
DEFAULT_PROC_DIR "%d/fd", pid) < 0) {
log_error("Could not format pid path.");
return 0;
}
/*
* Test for the kernel 'file (deleted)' form when scanning.
*/
if (dm_snprintf(deleted_path, sizeof(deleted_path), "%s %s",
path, PROC_FD_DELETED_STR) < 0) {
log_error("Could not format check path.");
return 0;
}
pid_d = opendir(path_buf);
if (!pid_d) {
log_error("Could not open proc path: %s.", path_buf);
return 0;
}
while ((pid_dp = readdir(pid_d)) != NULL) {
if (pid_dp->d_name[0] == '.')
continue;
if ((len = readlinkat(dirfd(pid_d), pid_dp->d_name, link_buf,
sizeof(link_buf))) < 0) {
log_error("readlink failed for " DEFAULT_PROC_DIR
"/%d/fd/.", pid);
goto bad;
}
link_buf[len] = '\0';
if (!strcmp(deleted_path, link_buf)) {
if (closedir(pid_d))
log_sys_error("closedir", path_buf);
return 1;
}
}
bad:
if (closedir(pid_d))
log_sys_error("closedir", path_buf);
return 0;
}
/*
* Attempt to determine whether a file is open by any process by
* scanning symbolic links in /proc/<pid>/fd.
*
* This is a heuristic since it cannot guarantee to detect brief
* access in all cases: a process that opens and then closes the
* file rapidly may never be seen by the scan.
*
* The method will also give false-positives if a process exists
* that has a deleted file open that had the same path, but a
* different inode number, to the file being monitored.
*
* For this reason the daemon only uses _is_open() for unlinked
* files when the mode is DM_FILEMAPD_FOLLOW_INODE, since these
* files can no longer be newly opened by processes.
*
* In this situation !is_open(path) provides an indication that
* the daemon should shut down: the file has been unlinked from
* the file system and we appear to hold the final reference.
*/
static int _is_open(const char *path)
{
struct dirent *proc_dp = NULL;
DIR *proc_d = NULL;
pid_t pid;
proc_d = opendir(DEFAULT_PROC_DIR);
if (!proc_d)
return 0;
while ((proc_dp = readdir(proc_d)) != NULL) {
if (!isdigit(proc_dp->d_name[0]))
continue;
errno = 0;
pid = (pid_t) strtol(proc_dp->d_name, NULL, 10);
if (errno || !pid)
continue;
if (_is_open_in_pid(pid, path)) {
if (closedir(proc_d))
log_sys_error("closedir", DEFAULT_PROC_DIR);
return 1;
}
}
if (closedir(proc_d))
log_sys_error("closedir", DEFAULT_PROC_DIR);
return 0;
}
static void _filemap_monitor_wait(uint64_t usecs)
{
if (_verbose) {
if (usecs == FILEMAPD_WAIT_USECS)
log_very_verbose("Waiting for check interval");
if (usecs == FILEMAPD_NOFILE_WAIT_USECS)
log_very_verbose("Waiting for unlinked path");
}
usleep((useconds_t) usecs);
}
static int _parse_args(int argc, char **argv, struct filemap_monitor *fm)
{
char *endptr;
/* we don't care what is in argv[0]. */
argc--;
argv++;
if (argc < 5) {
_early_log("Wrong number of arguments.");
_early_log("usage: %s", _usage);
return 0;
}
/*
* We don't know the true nr_regions at daemon start time,
* and it is not worth a dm_stats_list()/group walk to count:
* we can assume that there is at least one region or the
* daemon would not have been started.
*
* A correct value will be obtained following the first update
* of the group's regions.
*/
fm->nr_regions = 1;
/* parse <fd> */
errno = 0;
fm->fd = (int) strtol(argv[0], &endptr, 10);
if (errno || *endptr) {
_early_log("Could not parse file descriptor: %s", argv[0]);
return 0;
}
argc--;
argv++;
/* parse <group_id> */
errno = 0;
fm->group_id = strtoull(argv[0], &endptr, 10);
if (*endptr || errno) {
_early_log("Could not parse group identifier: %s", argv[0]);
return 0;
}
argc--;
argv++;
/* parse <path> */
if (!argv[0] || !strlen(argv[0])) {
_early_log("Path argument is required.");
return 0;
}
if (*argv[0] != '/') {
_early_log("Path argument must specify an absolute path.");
return 0;
}
fm->path = dm_strdup(argv[0]);
if (!fm->path) {
_early_log("Could not allocate memory for path argument.");
return 0;
}
argc--;
argv++;
/* parse <mode> */
if (!argv[0] || !strlen(argv[0])) {
_early_log("Mode argument is required.");
return 0;
}
fm->mode = dm_filemapd_mode_from_string(argv[0]);
if (fm->mode == DM_FILEMAPD_FOLLOW_NONE)
return 0;
argc--;
argv++;
/* parse [<foreground>[<verbose>]] */
if (argc) {
errno = 0;
_foreground = (int) strtol(argv[0], &endptr, 10);
if (errno || *endptr) {
_early_log("Could not parse debug argument: %s.",
argv[0]);
return 0;
}
argc--;
argv++;
if (argc) {
errno = 0;
_verbose = (int) strtol(argv[0], &endptr, 10);
if (errno || *endptr) {
_early_log("Could not parse verbose "
"argument: %s", argv[0]);
return 0;
}
if (_verbose < 0 || _verbose > 3) {
_early_log("Verbose argument out of range: %d.",
_verbose);
return 0;
}
}
}
return 1;
}
static int _filemap_fd_update_blocks(struct filemap_monitor *fm)
{
struct stat buf;
if (fm->fd < 0) {
log_error("Filemap fd is not open.");
return 0;
}
if (fstat(fm->fd, &buf)) {
log_error("Failed to fstat filemap file descriptor.");
return 0;
}
fm->blocks = buf.st_blocks;
return 1;
}
static int _filemap_fd_check_changed(struct filemap_monitor *fm)
{
int64_t old_blocks;
old_blocks = fm->blocks;
if (!_filemap_fd_update_blocks(fm))
return -1;
return (fm->blocks != old_blocks);
}
static void _filemap_monitor_close_fd(struct filemap_monitor *fm)
{
if (close(fm->fd))
log_error("Error closing file descriptor.");
fm->fd = -1;
}
static void _filemap_monitor_end_notify(struct filemap_monitor *fm)
{
inotify_rm_watch(fm->inotify_fd, fm->inotify_watch_fd);
}
static int _filemap_monitor_set_notify(struct filemap_monitor *fm)
{
int inotify_fd, watch_fd;
/*
* Set IN_NONBLOCK since we do not want to block in event read()
* calls. Do not set IN_CLOEXEC as dmfilemapd is single-threaded
* and does not fork or exec.
*/
if ((inotify_fd = inotify_init1(IN_NONBLOCK)) < 0) {
log_sys_error("inotify_init1", "IN_NONBLOCK");
return 0;
}
if ((watch_fd = inotify_add_watch(inotify_fd, fm->path,
IN_MODIFY | IN_DELETE_SELF)) < 0) {
log_sys_error("inotify_add_watch", fm->path);
return 0;
}
fm->inotify_fd = inotify_fd;
fm->inotify_watch_fd = watch_fd;
return 1;
}
static int _filemap_monitor_reopen_fd(struct filemap_monitor *fm)
{
int tries = FILEMAPD_NOFILE_WAIT_TRIES;
/*
* In DM_FILEMAPD_FOLLOW_PATH mode, inotify watches must be
* re-established whenever the file at the watched path is
* changed.
*
* FIXME: stat file and skip if inode is unchanged.
*/
if (fm->fd > 0)
log_error("Filemap file descriptor already open.");
while ((fm->fd < 0) && --tries)
if (((fm->fd = open(fm->path, O_RDONLY)) < 0) && tries)
_filemap_monitor_wait(FILEMAPD_NOFILE_WAIT_USECS);
if (!tries && (fm->fd < 0)) {
log_error("Could not re-open file descriptor.");
return 0;
}
return _filemap_monitor_set_notify(fm);
}
static int _filemap_monitor_get_events(struct filemap_monitor *fm)
{
/* alignment as per man(7) inotify */
char buf[sizeof(struct inotify_event) + NAME_MAX + 1]
__attribute__ ((aligned(__alignof__(struct inotify_event))));
struct inotify_event *event;
int check = 0;
ssize_t len;
char *ptr;
/*
* Close the file descriptor for the file being monitored here
* when mode=path: this will allow the inode to be de-allocated,
* and an IN_DELETE_SELF event generated in the case that the
* daemon is holding the last open reference to the file.
*/
if (fm->mode == DM_FILEMAPD_FOLLOW_PATH) {
_filemap_monitor_end_notify(fm);
_filemap_monitor_close_fd(fm);
}
len = read(fm->inotify_fd, (void *) &buf, sizeof(buf));
/* no events to read? */
if (len < 0 && (errno == EAGAIN))
goto out;
/* interrupted by signal? */
if (len < 0 && (errno == EINTR))
goto out;
if (len < 0)
return -1;
if (!len)
goto out;
for (ptr = buf; ptr < buf + len; ptr += sizeof(*event) + event->len) {
event = (struct inotify_event *) ptr;
if (event->mask & IN_DELETE_SELF)
fm->deleted = 1;
if (event->mask & IN_MODIFY)
check = 1;
/*
* Event IN_IGNORED is generated when a file has been deleted
* and IN_DELETE_SELF generated, and indicates that the file
* watch has been automatically removed.
*
* This can only happen for the DM_FILEMAPD_FOLLOW_PATH mode,
* since inotify IN_DELETE events are generated at the time
* the inode is destroyed: DM_FILEMAPD_FOLLOW_INODE will hold
* the file descriptor open, meaning that the event will not
* be generated until after the daemon closes the file.
*
* The event is ignored here since inotify monitoring will
* be reestablished (or the daemon will terminate) following
* deletion of a DM_FILEMAPD_FOLLOW_PATH monitored file.
*/
if (event->mask & IN_IGNORED)
log_very_verbose("Inotify watch removed: IN_IGNORED "
"in event->mask");
}
out:
/*
* Re-open file descriptor if required and log disposition.
*/
if (fm->mode == DM_FILEMAPD_FOLLOW_PATH)
if (!_filemap_monitor_reopen_fd(fm))
return -1;
log_very_verbose("exiting _filemap_monitor_get_events() with "
"deleted=%d, check=%d", fm->deleted, check);
return check;
}
static void _filemap_monitor_destroy(struct filemap_monitor *fm)
{
if (fm->fd > 0) {
_filemap_monitor_end_notify(fm);
_filemap_monitor_close_fd(fm);
}
dm_free((void *) fm->program_id);
dm_free(fm->path);
}
static int _filemap_monitor_check_same_file(int fd1, int fd2)
{
struct stat buf1, buf2;
if ((fd1 < 0) || (fd2 < 0))
return 0;
if (fstat(fd1, &buf1)) {
log_error("Failed to fstat file descriptor %d", fd1);
return -1;
}
if (fstat(fd2, &buf2)) {
log_error("Failed to fstat file descriptor %d", fd2);
return -1;
}
return ((buf1.st_dev == buf2.st_dev) && (buf1.st_ino == buf2.st_ino));
}
static int _filemap_monitor_check_file_unlinked(struct filemap_monitor *fm)
{
char path_buf[PATH_MAX];
char link_buf[PATH_MAX];
int same, fd;
ssize_t len;
fm->deleted = 0;
same = 0;
if ((fd = open(fm->path, O_RDONLY)) < 0)
goto check_unlinked;
same = _filemap_monitor_check_same_file(fm->fd, fd);
if (close(fd))
log_error("Error closing fd %d", fd);
if (same < 0)
return 0;
if (same)
return 1;
check_unlinked:
/*
* The file has been unlinked from its original location: test
* whether it is still reachable in the filesystem, or if it is
* unlinked and anonymous.
*/
if (dm_snprintf(path_buf, sizeof(path_buf), DEFAULT_PROC_DIR
"/%d/fd/%d", getpid(), fm->fd) < 0) {
log_error("Could not format pid path.");
return 0;
}
if ((len = readlink(path_buf, link_buf, sizeof(link_buf) - 1)) < 0) {
log_error("readlink failed for " DEFAULT_PROC_DIR "/%d/fd/%d.",
getpid(), fm->fd);
return 0;
}
link_buf[len] = '\0';
/*
* Try to re-open the file, from the path now reported in /proc/pid/fd.
*/
if ((fd = open(link_buf, O_RDONLY)) < 0)
fm->deleted = 1;
else
same = _filemap_monitor_check_same_file(fm->fd, fd);
if ((fd >= 0) && close(fd))
log_error("Error closing fd %d", fd);
if (same < 0)
return 0;
/* Should not happen with normal /proc. */
if ((fd > 0) && !same) {
log_error("File descriptor mismatch: %d and %s (read from %s) "
"are not the same file!", fm->fd, link_buf, path_buf);
return 0;
}
return 1;
}
static int _daemonise(struct filemap_monitor *fm)
{
pid_t pid = 0, sid;
int fd;
if (!(sid = setsid())) {
_early_log("setsid failed.");
return 0;
}
if ((pid = fork()) < 0) {
_early_log("Failed to fork daemon process.");
return 0;
}
if (pid > 0) {
if (_verbose)
_early_log("Started dmfilemapd with pid=%d", pid);
exit(0);
}
if (chdir("/")) {
_early_log("Failed to change directory.");
return 0;
}
if (!_verbose) {
if (close(STDIN_FILENO))
_early_log("Error closing stdin");
if (close(STDOUT_FILENO))
_early_log("Error closing stdout");
if (close(STDERR_FILENO))
_early_log("Error closing stderr");
if ((open("/dev/null", O_RDONLY) < 0) ||
(open("/dev/null", O_WRONLY) < 0) ||
(open("/dev/null", O_WRONLY) < 0)) {
_early_log("Error opening stdio streams.");
return 0;
}
}
/* TODO: Use libdaemon/server/daemon-server.c _daemonise() */
for (fd = (int) sysconf(_SC_OPEN_MAX) - 1; fd > STDERR_FILENO; fd--) {
if (fd == fm->fd)
continue;
(void) close(fd);
}
return 1;
}
static int _update_regions(struct dm_stats *dms, struct filemap_monitor *fm)
{
uint64_t *regions = NULL, *region, nr_regions = 0;
regions = dm_stats_update_regions_from_fd(dms, fm->fd, fm->group_id);
if (!regions) {
log_error("Failed to update filemap regions for group_id="
FMTu64 ".", fm->group_id);
return 0;
}
for (region = regions; *region != DM_STATS_REGIONS_ALL; region++)
nr_regions++;
if (!nr_regions)
log_warn("File contains no extents: exiting.");
if (nr_regions && (regions[0] != fm->group_id)) {
log_warn("group_id changed from " FMTu64 " to " FMTu64,
fm->group_id, regions[0]);
fm->group_id = regions[0];
}
dm_free(regions);
fm->nr_regions = nr_regions;
return 1;
}
static int _dmfilemapd(struct filemap_monitor *fm)
{
int running = 1, check = 0, open = 0;
const char *program_id;
struct dm_stats *dms;
/*
* The correct program_id is retrieved from the group leader
* following the call to dm_stats_list().
*/
if (!(dms = dm_stats_create(NULL)))
goto_bad;
if (!dm_stats_bind_from_fd(dms, fm->fd)) {
log_error("Could not bind dm_stats handle to file descriptor "
"%d", fm->fd);
goto bad;
}
if (!_filemap_monitor_set_notify(fm))
goto bad;
if (!_filemap_fd_update_blocks(fm))
goto bad;
if (!dm_stats_list(dms, DM_STATS_ALL_PROGRAMS)) {
log_error("Failed to list stats handle.");
goto bad;
}
/*
* Take the program_id for new regions (created by calls to
* dm_stats_update_regions_from_fd()) from the value used by
* the group leader.
*/
program_id = dm_stats_get_region_program_id(dms, fm->group_id);
if (program_id)
fm->program_id = dm_strdup(program_id);
else
fm->program_id = NULL;
dm_stats_set_program_id(dms, 1, program_id);
do {
if (!dm_stats_group_present(dms, fm->group_id)) {
log_info("Filemap group removed: exiting.");
running = 0;
continue;
}
if ((check = _filemap_monitor_get_events(fm)) < 0)
goto bad;
if (!check)
goto wait;
if ((check = _filemap_fd_check_changed(fm)) < 0)
goto bad;
if (check && !_update_regions(dms, fm))
goto bad;
running = !!fm->nr_regions;
if (!running)
continue;
wait:
_filemap_monitor_wait(FILEMAPD_WAIT_USECS);
/* mode=inode termination condions */
if (fm->mode == DM_FILEMAPD_FOLLOW_INODE) {
if (!_filemap_monitor_check_file_unlinked(fm))
goto bad;
if (fm->deleted && !(open = _is_open(fm->path))) {
log_info("File unlinked and closed: exiting.");
running = 0;
} else if (fm->deleted && open)
log_verbose("File unlinked and open: "
"continuing.");
}
if (!dm_stats_list(dms, NULL)) {
log_error("Failed to list stats handle.");
goto bad;
}
} while (running);
_filemap_monitor_destroy(fm);
dm_stats_destroy(dms);
return 0;
bad:
_filemap_monitor_destroy(fm);
dm_stats_destroy(dms);
log_error("Exiting");
return 1;
}
static const char * _mode_names[] = {
"inode",
"path"
};
/*
* dmfilemapd <fd> <group_id> <path> <mode> [<foreground>[<log_level>]]
*/
int main(int argc, char **argv)
{
struct filemap_monitor fm;
memset(&fm, 0, sizeof(fm));
if (!_parse_args(argc, argv, &fm)) {
dm_free(fm.path);
return 1;
}
_setup_logging();
log_info("Starting dmfilemapd with fd=%d, group_id=" FMTu64 " "
"mode=%s, path=%s", fm.fd, fm.group_id,
_mode_names[fm.mode], fm.path);
if (!_foreground && !_daemonise(&fm))
return 1;
return _dmfilemapd(&fm);
}

View File

@@ -38,7 +38,7 @@ class AutomatedProperties(dbus.service.Object):
props = {}
for i in self.interface():
props[i] = AutomatedProperties._get_all_prop(self, i)
props[i] = self.GetAll(i)
return self._ap_o_path, props
@@ -65,52 +65,31 @@ class AutomatedProperties(dbus.service.Object):
return self._ap_interface
@staticmethod
def _get_prop(obj, interface_name, property_name):
value = getattr(obj, property_name)
# Properties
# noinspection PyUnusedLocal
@dbus.service.method(dbus_interface=dbus.PROPERTIES_IFACE,
in_signature='ss', out_signature='v')
def Get(self, interface_name, property_name):
value = getattr(self, property_name)
# Note: If we get an exception in this handler we won't know about it,
# only the side effect of no returned value!
log_debug('Get (%s), type (%s), value(%s)' %
(property_name, str(type(value)), str(value)))
return value
# Properties
# noinspection PyUnusedLocal
@dbus.service.method(dbus_interface=dbus.PROPERTIES_IFACE,
in_signature='ss', out_signature='v',
async_callbacks=('cb', 'cbe'))
def Get(self, interface_name, property_name, cb, cbe):
# Note: If we get an exception in this handler we won't know about it,
# only the side effect of no returned value!
r = cfg.create_request_entry(
-1, AutomatedProperties._get_prop,
(self, interface_name, property_name),
cb, cbe, False)
cfg.worker_q.put(r)
@staticmethod
def _get_all_prop(obj, interface_name):
if interface_name in obj.interface(True):
in_signature='s', out_signature='a{sv}')
def GetAll(self, interface_name):
if interface_name in self.interface(True):
# Using introspection, lets build this dynamically
properties = get_properties(obj)
properties = get_properties(self)
if interface_name in properties:
return properties[interface_name][1]
return {}
raise dbus.exceptions.DBusException(
obj._ap_interface,
self._ap_interface,
'The object %s does not implement the %s interface'
% (obj.__class__, interface_name))
@dbus.service.method(dbus_interface=dbus.PROPERTIES_IFACE,
in_signature='s', out_signature='a{sv}',
async_callbacks=('cb', 'cbe'))
def GetAll(self, interface_name, cb, cbe):
r = cfg.create_request_entry(
-1, AutomatedProperties._get_all_prop,
(self, interface_name),
cb, cbe, False)
cfg.worker_q.put(r)
% (self.__class__, interface_name))
@dbus.service.method(dbus_interface=dbus.PROPERTIES_IFACE,
in_signature='ssv')

View File

@@ -9,13 +9,11 @@
import subprocess
from . import cfg
from .cmdhandler import options_to_cli_args, LvmExecutionMeta
from .cmdhandler import options_to_cli_args
import dbus
from .utils import pv_range_append, pv_dest_ranges, log_error, log_debug,\
add_no_notify
from .utils import pv_range_append, pv_dest_ranges, log_error, log_debug
import os
import threading
import time
def pv_move_lv_cmd(move_options, lv_full_name,
@@ -44,15 +42,6 @@ def _move_merge(interface_name, command, job_state):
# the command always as we will be getting periodic output from them on
# the status of the long running operation.
command.insert(0, cfg.LVM_CMD)
# Instruct lvm to not register an event with us
command = add_no_notify(command)
#(self, start, ended, cmd, ec, stdout_txt, stderr_txt)
meta = LvmExecutionMeta(time.time(), 0, command, -1000, None, None)
cfg.blackbox.add(meta)
process = subprocess.Popen(command, stdout=subprocess.PIPE,
env=os.environ,
stderr=subprocess.PIPE, close_fds=True)
@@ -70,21 +59,12 @@ def _move_merge(interface_name, command, job_state):
(device, ignore, percentage) = line_str.split(':')
job_state.Percent = round(
float(percentage.strip()[:-1]), 1)
# While the move is in progress we need to periodically update
# the state to reflect where everything is at.
cfg.load()
except ValueError:
log_error("Trying to parse percentage which failed for %s" %
line_str)
out = process.communicate()
with meta.lock:
meta.ended = time.time()
meta.ec = process.returncode
meta.stderr_txt = out[1]
if process.returncode == 0:
job_state.Percent = 100
else:
@@ -158,6 +138,5 @@ def _run_cmd(req):
def cmd_runner(request):
t = threading.Thread(target=_run_cmd, args=(request,),
name="cmd_runner %s" % str(request.method))
t = threading.Thread(target=_run_cmd, args=(request,))
t.start()

View File

@@ -26,7 +26,7 @@ bus = None
args = None
# Set to true if we are depending on external events for updates
got_external_event = False
ee = False
# Shared state variable across all processes
run = multiprocessing.Value('i', 1)
@@ -84,6 +84,3 @@ db = None
# lvm flight recorder
blackbox = None
# RequestEntry ctor
create_request_entry = None

View File

@@ -16,7 +16,7 @@ import traceback
import os
from lvmdbusd import cfg
from lvmdbusd.utils import pv_dest_ranges, log_debug, log_error, add_no_notify
from lvmdbusd.utils import pv_dest_ranges, log_debug, log_error
from lvmdbusd.lvm_shell_proxy import LVMShellProxy
try:
@@ -37,7 +37,6 @@ cmd_lock = threading.RLock()
class LvmExecutionMeta(object):
def __init__(self, start, ended, cmd, ec, stdout_txt, stderr_txt):
self.lock = threading.RLock()
self.start = start
self.ended = ended
self.cmd = cmd
@@ -46,13 +45,12 @@ class LvmExecutionMeta(object):
self.stderr_txt = stderr_txt
def __str__(self):
with self.lock:
return "EC= %d for %s\n" \
"STARTED: %f, ENDED: %f\n" \
"STDOUT=%s\n" \
"STDERR=%s\n" % \
(self.ec, str(self.cmd), self.start, self.ended, self.stdout_txt,
self.stderr_txt)
return "EC= %d for %s\n" \
"STARTED: %f, ENDED: %f\n" \
"STDOUT=%s\n" \
"STDERR=%s\n" % \
(self.ec, str(self.cmd), self.start, self.ended, self.stdout_txt,
self.stderr_txt)
class LvmFlightRecorder(object):
@@ -95,7 +93,6 @@ def call_lvm(command, debug=False):
# Prepend the full lvm executable so that we can run different versions
# in different locations on the same box
command.insert(0, cfg.LVM_CMD)
command = add_no_notify(command)
process = Popen(command, stdout=PIPE, stderr=PIPE, close_fds=True,
env=os.environ)
@@ -281,7 +278,7 @@ def vg_lv_create(vg_name, create_options, name, size_bytes, pv_dests):
cmd = ['lvcreate']
cmd.extend(options_to_cli_args(create_options))
cmd.extend(['--size', str(size_bytes) + 'B'])
cmd.extend(['--name', name, vg_name, '--yes'])
cmd.extend(['--name', name, vg_name])
pv_dest_ranges(cmd, pv_dests)
return call(cmd)
@@ -306,8 +303,6 @@ def _vg_lv_create_common_cmd(create_options, size_bytes, thin_pool):
cmd.extend(['--size', str(size_bytes) + 'B'])
else:
cmd.extend(['--thin', '--size', str(size_bytes) + 'B'])
cmd.extend(['--yes'])
return cmd
@@ -344,7 +339,7 @@ def _vg_lv_create_raid(vg_name, create_options, name, raid_type, size_bytes,
if stripe_size_kb != 0:
cmd.extend(['--stripesize', str(stripe_size_kb)])
cmd.extend(['--name', name, vg_name, '--yes'])
cmd.extend(['--name', name, vg_name])
return call(cmd)
@@ -365,7 +360,7 @@ def vg_lv_create_mirror(
cmd.extend(['--type', 'mirror'])
cmd.extend(['--mirrors', str(num_copies)])
cmd.extend(['--size', str(size_bytes) + 'B'])
cmd.extend(['--name', name, vg_name, '--yes'])
cmd.extend(['--name', name, vg_name])
return call(cmd)
@@ -419,7 +414,7 @@ def lv_lv_create(lv_full_name, create_options, name, size_bytes):
cmd = ['lvcreate']
cmd.extend(options_to_cli_args(create_options))
cmd.extend(['--virtualsize', str(size_bytes) + 'B', '-T'])
cmd.extend(['--name', name, lv_full_name, '--yes'])
cmd.extend(['--name', name, lv_full_name])
return call(cmd)
@@ -555,7 +550,7 @@ def pv_resize(device, size_bytes, create_options):
cmd.extend(options_to_cli_args(create_options))
if size_bytes != 0:
cmd.extend(['--yes', '--setphysicalvolumesize', str(size_bytes) + 'B'])
cmd.extend(['--setphysicalvolumesize', str(size_bytes) + 'B'])
cmd.extend([device])
return call(cmd)
@@ -620,10 +615,10 @@ def vg_reduce(vg_name, missing, pv_devices, reduce_options):
cmd = ['vgreduce']
cmd.extend(options_to_cli_args(reduce_options))
if len(pv_devices) == 0:
cmd.append('--all')
if missing:
cmd.append('--removemissing')
elif len(pv_devices) == 0:
cmd.append('--all')
cmd.append(vg_name)
cmd.extend(pv_devices)

View File

@@ -82,10 +82,10 @@ class StateUpdate(object):
@staticmethod
def update_thread(obj):
queued_requests = []
while cfg.run.value != 0:
# noinspection PyBroadException
try:
queued_requests = []
refresh = True
emit_signal = True
cache_refresh = True
@@ -96,7 +96,7 @@ class StateUpdate(object):
wait = not obj.deferred
obj.deferred = False
if len(queued_requests) == 0 and wait:
if wait:
queued_requests.append(obj.queue.get(True, 2))
# Ok we have one or the deferred queue has some,
@@ -131,17 +131,11 @@ class StateUpdate(object):
for i in queued_requests:
i.set_result(num_changes)
# Only clear out the requests after we have given them a result
# otherwise we can orphan the waiting threads and they never
# wake up if we get an exception
queued_requests = []
except queue.Empty:
pass
except Exception:
st = traceback.format_exc()
log_error("update_thread exception: \n%s" % st)
cfg.blackbox.dump()
def __init__(self):
self.lock = threading.RLock()
@@ -152,8 +146,7 @@ class StateUpdate(object):
load(refresh=False, emit_signal=False, need_main_thread=False)
self.thread = threading.Thread(target=StateUpdate.update_thread,
args=(self,),
name="StateUpdate.update_thread")
args=(self,))
def load(self, refresh=True, emit_signal=True, cache_refresh=True,
log=True, need_main_thread=True):

View File

@@ -8,7 +8,7 @@
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from .automatedproperties import AutomatedProperties
from .utils import job_obj_path_generate, mt_async_call
from .utils import job_obj_path_generate, mt_async_result, mt_run_no_wait
from . import cfg
from .cfg import JOB_INTERFACE
import dbus
@@ -30,7 +30,7 @@ class WaitingClient(object):
# Remove ourselves from waiting client
wc.job_state.remove_waiting_client(wc)
wc.timer_id = -1
mt_async_call(wc.cb, wc.job_state.Complete)
mt_async_result(wc.cb, wc.job_state.Complete)
wc.job_state = None
def __init__(self, job_state, tmo, cb, cbe):
@@ -55,7 +55,7 @@ class WaitingClient(object):
GLib.source_remove(self.timer_id)
self.timer_id = -1
mt_async_call(self.cb, self.job_state.Complete)
mt_async_result(self.cb, self.job_state.Complete)
self.job_state = None
@@ -188,7 +188,7 @@ class Job(AutomatedProperties):
@Complete.setter
def Complete(self, value):
self.state.Complete = value
mt_async_call(Job._signal_complete, self)
mt_run_no_wait(Job._signal_complete, self)
@property
def GetError(self):

View File

@@ -29,7 +29,7 @@ except ImportError:
from lvmdbusd.cfg import LVM_CMD
from lvmdbusd.utils import log_debug, log_error, add_no_notify
from lvmdbusd.utils import log_debug, log_error
SHELL_PROMPT = "lvm> "
@@ -206,8 +206,6 @@ class LVMShellProxy(object):
self.lvm_shell.returncode,
"Underlying lvm shell process is not present!")
argv = add_no_notify(argv)
# create the command string
cmd = " ".join(_quote_arg(arg) for arg in argv)
cmd += "\n"

View File

@@ -30,7 +30,6 @@ import argparse
import os
import sys
from .cmdhandler import LvmFlightRecorder
from .request import RequestEntry
class Lvm(objectmanager.ObjectManager):
@@ -63,24 +62,6 @@ def check_bb_size(value):
return v
def install_signal_handlers():
# Because of the glib main loop stuff the python signal handler code is
# apparently not usable and we need to use the glib calls instead
signal_add = None
if hasattr(GLib, 'unix_signal_add'):
signal_add = GLib.unix_signal_add
elif hasattr(GLib, 'unix_signal_add_full'):
signal_add = GLib.unix_signal_add_full
if signal_add:
signal_add(GLib.PRIORITY_HIGH, signal.SIGHUP, utils.handler, signal.SIGHUP)
signal_add(GLib.PRIORITY_HIGH, signal.SIGINT, utils.handler, signal.SIGINT)
signal_add(GLib.PRIORITY_HIGH, signal.SIGUSR1, utils.handler, signal.SIGUSR1)
else:
log_error("GLib.unix_signal_[add|add_full] are NOT available!")
def main():
start = time.time()
# Add simple command line handling
@@ -116,7 +97,6 @@ def main():
os.environ["LC_ALL"] = "C"
cfg.args = parser.parse_args()
cfg.create_request_entry = RequestEntry
# We create a flight recorder in cmdhandler too, but we replace it here
# as the user may be specifying a different size. The default one in
@@ -130,7 +110,12 @@ def main():
# List of threads that we start up
thread_list = []
install_signal_handlers()
# Install signal handlers
for s in [signal.SIGHUP, signal.SIGINT]:
try:
signal.signal(s, utils.handler)
except RuntimeError:
pass
dbus.mainloop.glib.DBusGMainLoop(set_as_default=True)
dbus.mainloop.glib.threads_init()
@@ -151,8 +136,7 @@ def main():
# Using a thread to process requests, we cannot hang the dbus library
# thread that is handling the dbus interface
thread_list.append(threading.Thread(target=process_request,
name='process_request'))
thread_list.append(threading.Thread(target=process_request))
# Have a single thread handling updating lvm and the dbus model so we
# don't have multiple threads doing this as the same time
@@ -160,6 +144,7 @@ def main():
thread_list.append(updater.thread)
cfg.load = updater.load
cfg.event = updater.event
cfg.loop = GLib.MainLoop()
@@ -190,7 +175,5 @@ def main():
for thread in thread_list:
thread.join()
except KeyboardInterrupt:
# If we are unable to register signal handler, we will end up here when
# the service gets a ^C or a kill -2 <parent pid>
utils.handler(signal.SIGINT)
utils.handler(signal.SIGINT, None)
return 0

View File

@@ -6,6 +6,7 @@
#
# You should have received a copy of the GNU General Public License
# along with this program. If not, see <http://www.gnu.org/licenses/>.
from .automatedproperties import AutomatedProperties
from . import utils
@@ -47,7 +48,7 @@ class Manager(AutomatedProperties):
pv = cfg.om.get_object_path_by_uuid_lvm_id(device, device)
if pv:
raise dbus.exceptions.DBusException(
MANAGER_INTERFACE, "PV %s Already exists!" % device)
MANAGER_INTERFACE, "PV Already exists!")
rc, out, err = cmdhandler.pv_create(create_options, [device])
Manager.handle_execute(rc, out, err)
@@ -131,28 +132,11 @@ class Manager(AutomatedProperties):
r = RequestEntry(-1, Manager._refresh, (), cb, cbe, False)
cfg.worker_q.put(r)
@dbus.service.method(
dbus_interface=MANAGER_INTERFACE)
def FlightRecorderDump(self):
"""
Dump the flight recorder to syslog
"""
cfg.blackbox.dump()
@staticmethod
def _lookup_by_lvm_id(key):
p = cfg.om.get_object_path_by_uuid_lvm_id(key, key)
if not p:
p = '/'
utils.log_debug('LookUpByLvmId: key = %s, result = %s' % (key, p))
return p
@dbus.service.method(
dbus_interface=MANAGER_INTERFACE,
in_signature='s',
out_signature='o',
async_callbacks=('cb', 'cbe'))
def LookUpByLvmId(self, key, cb, cbe):
out_signature='o')
def LookUpByLvmId(self, key):
"""
Given a lvm id in one of the forms:
@@ -166,8 +150,10 @@ class Manager(AutomatedProperties):
:param key: The lookup value
:return: Return the object path. If object not found you will get '/'
"""
r = RequestEntry(-1, Manager._lookup_by_lvm_id, (key,), cb, cbe, False)
cfg.worker_q.put(r)
p = cfg.om.get_object_path_by_uuid_lvm_id(key, key)
if p:
return p
return '/'
@staticmethod
def _use_lvm_shell(yes_no):
@@ -183,33 +169,25 @@ class Manager(AutomatedProperties):
:param yes_no:
:param cb: dbus python call back parameter, not client visible
:param cbe: dbus python error call back parameter, not client visible
:return: Boolean
:return: Nothing
"""
r = RequestEntry(-1, Manager._use_lvm_shell, (yes_no,), cb, cbe, False)
cfg.worker_q.put(r)
@staticmethod
def _external_event(command):
utils.log_debug("Processing _external_event= %s" % command,
'bg_black', 'fg_orange')
cfg.load()
@dbus.service.method(
dbus_interface=MANAGER_INTERFACE,
in_signature='s', out_signature='i')
def ExternalEvent(self, command):
utils.log_debug("ExternalEvent %s" % command)
# If a user didn't explicitly specify udev, we will turn it off now.
if not cfg.args.use_udev:
if udevwatch.remove():
utils.log_debug("ExternalEvent received, disabling "
"udev monitoring")
# We are dependent on external events now to stay current!
cfg.got_external_event = True
r = RequestEntry(
-1, Manager._external_event, (command,), None, None, False)
cfg.worker_q.put(r)
cfg.ee = True
utils.log_debug("ExternalEvent %s" % command)
cfg.event()
return dbus.Int32(0)
@staticmethod

View File

@@ -32,12 +32,14 @@ class ObjectManager(AutomatedProperties):
self._id_to_object_path = {}
self.rlock = threading.RLock()
@staticmethod
def _get_managed_objects(obj):
with obj.rlock:
@dbus.service.method(
dbus_interface="org.freedesktop.DBus.ObjectManager",
out_signature='a{oa{sa{sv}}}')
def GetManagedObjects(self):
with self.rlock:
rc = {}
try:
for k, v in list(obj._objects.items()):
for k, v in list(self._objects.items()):
path, props = v[0].emit_data()
rc[path] = props
except Exception:
@@ -45,14 +47,6 @@ class ObjectManager(AutomatedProperties):
sys.exit(1)
return rc
@dbus.service.method(
dbus_interface="org.freedesktop.DBus.ObjectManager",
out_signature='a{oa{sa{sv}}}', async_callbacks=('cb', 'cbe'))
def GetManagedObjects(self, cb, cbe):
r = cfg.create_request_entry(-1, ObjectManager._get_managed_objects,
(self, ), cb, cbe, False)
cfg.worker_q.put(r)
def locked(self):
"""
If some external code need to run across a number of different
@@ -223,9 +217,8 @@ class ObjectManager(AutomatedProperties):
:param lvm_id: The lvm identifier
"""
with self.rlock:
lookup_rc = self._id_lookup(lvm_id)
if lookup_rc:
return self.get_object_by_path(lookup_rc)
if lvm_id in self._id_to_object_path:
return self.get_object_by_path(self._id_to_object_path[lvm_id])
return None
def get_object_path_by_lvm_id(self, lvm_id):
@@ -235,9 +228,8 @@ class ObjectManager(AutomatedProperties):
:return: Object path or '/' if not found
"""
with self.rlock:
lookup_rc = self._id_lookup(lvm_id)
if lookup_rc:
return lookup_rc
if lvm_id in self._id_to_object_path:
return self._id_to_object_path[lvm_id]
return '/'
def _uuid_verify(self, path, uuid, lvm_id):

View File

@@ -79,9 +79,7 @@ class PvState(State):
self.lv = self._lv_object_list(vg_name)
# It's possible to have a vg_name and no uuid with the main example
# being when the vg_name == '[unknown]'
if vg_uuid and vg_name:
if vg_name:
self.vg_path = cfg.om.get_object_path_by_uuid_lvm_id(
vg_uuid, vg_name, vg_obj_path_generate)
else:

View File

@@ -13,7 +13,7 @@ from gi.repository import GLib
from .job import Job
from . import cfg
import traceback
from .utils import log_error, mt_async_call
from .utils import log_error, mt_async_result
class RequestEntry(object):
@@ -116,9 +116,9 @@ class RequestEntry(object):
if error_rc == 0:
if self.cb:
if self._return_tuple:
mt_async_call(self.cb, (result, '/'))
mt_async_result(self.cb, (result, '/'))
else:
mt_async_call(self.cb, result)
mt_async_result(self.cb, result)
else:
if self.cb_error:
if not error_exception:
@@ -129,7 +129,7 @@ class RequestEntry(object):
else:
error_exception = Exception(error_msg)
mt_async_call(self.cb_error, error_exception)
mt_async_result(self.cb_error, error_exception)
else:
# We have a job and it's complete, indicate that it's done.
self._job.Complete = True

View File

@@ -10,41 +10,10 @@
import pyudev
import threading
from . import cfg
from .request import RequestEntry
from . import utils
observer = None
observer_lock = threading.RLock()
_udev_lock = threading.RLock()
_udev_count = 0
def udev_add():
global _udev_count
with _udev_lock:
if _udev_count == 0:
_udev_count += 1
# Place this on the queue so any other operations will sequence
# behind it
r = RequestEntry(
-1, _udev_event, (), None, None, False)
cfg.worker_q.put(r)
def udev_complete():
global _udev_count
with _udev_lock:
if _udev_count > 0:
_udev_count -= 1
def _udev_event():
utils.log_debug("Processing udev event")
udev_complete()
cfg.load()
# noinspection PyUnusedLocal
def filter_event(action, device):
@@ -68,7 +37,7 @@ def filter_event(action, device):
refresh = True
if refresh:
udev_add()
cfg.event()
def add():

View File

@@ -20,8 +20,7 @@ from lvmdbusd import cfg
# noinspection PyUnresolvedReferences
from gi.repository import GLib
import threading
import traceback
import signal
STDOUT_TTY = os.isatty(sys.stdout.fileno())
@@ -282,47 +281,12 @@ def log_error(msg, *attributes):
_common_log(msg, *attributes)
def dump_threads_stackframe():
ident_to_name = {}
for thread_object in threading.enumerate():
ident_to_name[thread_object.ident] = thread_object
stacks = []
for thread_ident, frame in sys._current_frames().items():
stack = traceback.format_list(traceback.extract_stack(frame))
# There is a possibility that a thread gets created after we have
# enumerated all threads, so this lookup table may be incomplete, so
# account for this
if thread_ident in ident_to_name:
thread_name = ident_to_name[thread_ident].name
else:
thread_name = "unknown"
stacks.append("Thread: %s" % (thread_name))
stacks.append("".join(stack))
log_error("Dumping thread stack frames!\n" + "\n".join(stacks))
# noinspection PyUnusedLocal
def handler(signum):
try:
if signum == signal.SIGUSR1:
dump_threads_stackframe()
else:
cfg.run.value = 0
log_debug('Exiting daemon with signal %d' % signum)
if cfg.loop is not None:
cfg.loop.quit()
except:
st = traceback.format_exc()
log_error("signal handler: exception (logged, not reported!) \n %s" % st)
# It's important we report that we handled the exception for the exception
# handler to continue to work, especially for signal 10 (SIGUSR1)
return True
def handler(signum, frame):
cfg.run.value = 0
log_debug('Signal handler called with signal %d' % signum)
if cfg.loop is not None:
cfg.loop.quit()
def pv_obj_path_generate():
@@ -535,62 +499,27 @@ def validate_tag(interface, tag):
% (tag, _ALLOWABLE_TAG_CH))
def add_no_notify(cmdline):
"""
Given a command line to execute we will see if `--config` is present, if it
is we will add the global/notify_dbus=0 to it, otherwise we will append it
to the end of the list.
:param: cmdline: The command line to inspect
:type: cmdline: list
:return: cmdline with notify_dbus config option present
:rtype: list
"""
# Only after we have seen an external event will be disable lvm from sending
# us one when we call lvm
if cfg.got_external_event:
if 'help' in cmdline:
return cmdline
if '--config' in cmdline:
for i, arg in enumerate(cmdline):
if arg == '--config':
if len(cmdline) <= i+1:
raise dbus.exceptions.DBusException("Missing value for --config option.")
cmdline[i+1] += " global/notify_dbus=0"
break
else:
cmdline.extend(['--config', 'global/notify_dbus=0'])
return cmdline
# The methods below which start with mt_* are used to execute the desired code
# on the the main thread of execution to alleviate any issues the dbus-python
# library with regards to multi-threaded access. Essentially, we are trying to
# ensure all dbus library interaction is done from the same thread!
def _async_handler(call_back, parameters):
params_str = ", ".join(str(x) for x in parameters)
log_debug('Main thread execution, callback = %s, parameters = (%s)' %
(str(call_back), params_str))
try:
if parameters:
call_back(*parameters)
else:
call_back()
except:
st = traceback.format_exc()
log_error("mt_async_call: exception (logged, not reported!) \n %s" % st)
def _async_result(call_back, results):
log_debug('Results = %s' % str(results))
call_back(results)
# Execute the function on the main thread with the provided parameters, do
# not return *any* value or wait for the execution to complete!
def mt_async_call(function_call_back, *parameters):
GLib.idle_add(_async_handler, function_call_back, parameters)
# Return result in main thread
def mt_async_result(call_back, results):
GLib.idle_add(_async_result, call_back, results)
# Take the supplied function and run it on the main thread and not wait for
# a result!
def mt_run_no_wait(function, param):
GLib.idle_add(function, param)
# Run the supplied function and arguments on the main thread and wait for them
# to complete while allowing the ability to get the return value too.
#
@@ -610,7 +539,6 @@ class MThreadRunner(object):
def __init__(self, function, *args):
self.f = function
self.rc = None
self.exception = None
self.args = args
self.function_complete = False
self.cond = threading.Condition(threading.Lock())
@@ -620,21 +548,13 @@ class MThreadRunner(object):
with self.cond:
if not self.function_complete:
self.cond.wait()
if self.exception:
raise self.exception
return self.rc
def _run(self):
try:
if self.args:
self.rc = self.f(*self.args)
else:
self.rc = self.f()
except BaseException as be:
self.exception = be
st = traceback.format_exc()
log_error("MThreadRunner: exception \n %s" % st)
log_error("Exception will be raised in calling thread!")
if len(self.args):
self.rc = self.f(*self.args)
else:
self.rc = self.f()
def _remove_objects(dbus_objects_rm):

View File

@@ -16,7 +16,7 @@ top_srcdir = @top_srcdir@
top_builddir = @top_builddir@
SOURCES = lvmetad-core.c
SOURCES2 = lvmetactl.c
SOURCES2 = testclient.c
TARGETS = lvmetad lvmetactl
@@ -28,19 +28,22 @@ CFLOW_TARGET = lvmetad
include $(top_builddir)/make.tmpl
CFLAGS_lvmetactl.o += $(EXTRA_EXEC_CFLAGS)
CFLAGS_lvmetad-core.o += $(EXTRA_EXEC_CFLAGS)
INCLUDES += -I$(top_srcdir)/libdaemon/server
LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
LIBS += $(RT_LIBS) $(DAEMON_LIBS) -ldevmapper $(PTHREAD_LIBS)
LVMLIBS = -ldaemonserver $(LVMINTERNAL_LIBS) -ldevmapper
LIBS += $(PTHREAD_LIBS)
LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS)
CLDFLAGS += -L$(top_builddir)/libdaemon/server
CFLAGS += $(EXTRA_EXEC_CFLAGS)
lvmetad: $(OBJECTS) $(top_builddir)/libdaemon/client/libdaemonclient.a \
$(top_builddir)/libdaemon/server/libdaemonserver.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) -ldaemonserver $(LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LVMLIBS) $(LIBS)
lvmetactl: lvmetactl.o $(top_builddir)/libdaemon/client/libdaemonclient.a \
$(top_builddir)/libdaemon/server/libdaemonserver.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ lvmetactl.o $(LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ lvmetactl.o $(LVMLIBS)
CLEAN_TARGETS += lvmetactl.o

View File

@@ -25,7 +25,6 @@
#define LVMETAD_DISABLE_REASON_LVM1 "LVM1"
#define LVMETAD_DISABLE_REASON_DUPLICATES "DUPLICATES"
#define LVMETAD_DISABLE_REASON_VGRESTORE "VGRESTORE"
#define LVMETAD_DISABLE_REASON_REPAIR "REPAIR"
struct volume_group;

View File

@@ -203,9 +203,8 @@ struct vg_info {
#define GLFL_DISABLE_REASON_LVM1 0x00000008
#define GLFL_DISABLE_REASON_DUPLICATES 0x00000010
#define GLFL_DISABLE_REASON_VGRESTORE 0x00000020
#define GLFL_DISABLE_REASON_REPAIR 0x00000040
#define GLFL_DISABLE_REASON_ALL (GLFL_DISABLE_REASON_DIRECT | GLFL_DISABLE_REASON_REPAIR | GLFL_DISABLE_REASON_LVM1 | GLFL_DISABLE_REASON_DUPLICATES | GLFL_DISABLE_REASON_VGRESTORE)
#define GLFL_DISABLE_REASON_ALL (GLFL_DISABLE_REASON_DIRECT | GLFL_DISABLE_REASON_LVM1 | GLFL_DISABLE_REASON_DUPLICATES | GLFL_DISABLE_REASON_VGRESTORE)
#define VGFL_INVALID 0x00000001
@@ -258,21 +257,6 @@ static void destroy_metadata_hashes(lvmetad_state *s)
dm_hash_iterate(n, s->pvid_to_pvmeta)
dm_config_destroy(dm_hash_get_data(s->pvid_to_pvmeta, n));
dm_hash_iterate(n, s->vgid_to_vgname)
dm_free(dm_hash_get_data(s->vgid_to_vgname, n));
dm_hash_iterate(n, s->vgname_to_vgid)
dm_free(dm_hash_get_data(s->vgname_to_vgid, n));
dm_hash_iterate(n, s->vgid_to_info)
dm_free(dm_hash_get_data(s->vgid_to_info, n));
dm_hash_iterate(n, s->device_to_pvid)
dm_free(dm_hash_get_data(s->device_to_pvid, n));
dm_hash_iterate(n, s->pvid_to_vgid)
dm_free(dm_hash_get_data(s->pvid_to_vgid, n));
dm_hash_destroy(s->pvid_to_pvmeta);
dm_hash_destroy(s->vgid_to_metadata);
dm_hash_destroy(s->vgid_to_vgname);
@@ -808,8 +792,7 @@ static int _update_pvid_to_vgid(lvmetad_state *s, struct dm_config_tree *vg,
if ((mode == REMOVE_EMPTY) && vgid_old) {
/* This copies the vgid_old string, doesn't reference it. */
if ((dm_hash_lookup(to_check, vgid_old) != (void*) 1) &&
!dm_hash_insert(to_check, vgid_old, (void*) 1)) {
if (!dm_hash_insert(to_check, vgid_old, (void*) 1)) {
ERROR(s, "update_pvid_to_vgid out of memory for hash insert vgid_old %s", vgid_old);
goto abort_daemon;
}
@@ -885,13 +868,16 @@ static int remove_metadata(lvmetad_state *s, const char *vgid, int update_pvids)
/* free the unmapped data */
if (info_lookup)
dm_free(info_lookup);
if (meta_lookup)
dm_config_destroy(meta_lookup);
if (name_lookup)
dm_free(name_lookup);
if (outdated_pvs_lookup)
dm_config_destroy(outdated_pvs_lookup);
dm_free(info_lookup);
dm_free(name_lookup);
dm_free(vgid_lookup);
if (vgid_lookup)
dm_free(vgid_lookup);
return 1;
}
@@ -1218,8 +1204,10 @@ static int _update_metadata_add_new(lvmetad_state *s, const char *new_name, cons
out:
out_free:
if (!new_name_dup || !new_vgid_dup || abort_daemon) {
dm_free(new_name_dup);
dm_free(new_vgid_dup);
if (new_name_dup)
dm_free(new_name_dup);
if (new_vgid_dup)
dm_free(new_vgid_dup);
ERROR(s, "lvmetad could not be updated and is aborting.");
exit(EXIT_FAILURE);
}
@@ -1809,7 +1797,8 @@ static response pv_gone(lvmetad_state *s, request r)
}
dm_config_destroy(pvmeta);
dm_free(old_pvid);
if (old_pvid)
dm_free(old_pvid);
return daemon_reply_simple("OK", NULL );
}
@@ -1922,7 +1911,7 @@ static response pv_found(lvmetad_state *s, request r)
const char *arg_pvid = NULL;
const char *arg_pvid_lookup = NULL;
const char *new_pvid = NULL;
char *new_pvid_dup = NULL;
const char *new_pvid_dup = NULL;
const char *arg_name = NULL;
const char *arg_vgid = NULL;
const char *arg_vgid_lookup = NULL;
@@ -2085,7 +2074,7 @@ static response pv_found(lvmetad_state *s, request r)
if (!(new_pvid_dup = dm_strdup(new_pvid)))
goto nomem_free1;
if (!dm_hash_insert_binary(s->device_to_pvid, &new_device, sizeof(new_device), new_pvid_dup))
if (!dm_hash_insert_binary(s->device_to_pvid, &new_device, sizeof(new_device), (char *)new_pvid_dup))
goto nomem_free2;
if (!dm_hash_insert(s->pvid_to_pvmeta, new_pvid, new_pvmeta))
@@ -2121,8 +2110,6 @@ static response pv_found(lvmetad_state *s, request r)
DEBUGLOG(s, "pv_found ignore duplicate device %" PRIu64 " of existing PV for pvid %s",
arg_device, arg_pvid);
dm_config_destroy(new_pvmeta);
/* device_to_pvid no longer references prev_pvid_lookup */
dm_free((void*)prev_pvid_on_dev);
s->flags |= GLFL_DISABLE;
s->flags |= GLFL_DISABLE_REASON_DUPLICATES;
return reply_fail("Ignore duplicate PV");
@@ -2133,7 +2120,7 @@ static response pv_found(lvmetad_state *s, request r)
if (!(new_pvid_dup = dm_strdup(new_pvid)))
goto nomem_free1;
if (!dm_hash_insert_binary(s->device_to_pvid, &arg_device, sizeof(arg_device), new_pvid_dup))
if (!dm_hash_insert_binary(s->device_to_pvid, &arg_device, sizeof(arg_device), (char *)new_pvid_dup))
goto nomem_free2;
if (!dm_hash_insert(s->pvid_to_pvmeta, new_pvid, new_pvmeta))
@@ -2233,7 +2220,8 @@ static response pv_found(lvmetad_state *s, request r)
}
/* This was unhashed from device_to_pvid above. */
dm_free((void *)prev_pvid_on_dev);
if (prev_pvid_on_dev)
dm_free((void *)prev_pvid_on_dev);
return daemon_reply_simple("OK",
"status = %s", vg_status,
@@ -2245,7 +2233,7 @@ static response pv_found(lvmetad_state *s, request r)
NULL);
nomem_free2:
dm_free(new_pvid_dup);
dm_free((char *)new_pvid_dup);
nomem_free1:
dm_config_destroy(new_pvmeta);
nomem:
@@ -2367,8 +2355,6 @@ static response set_global_info(lvmetad_state *s, request r)
if ((reason = daemon_request_str(r, "disable_reason", NULL))) {
if (strstr(reason, LVMETAD_DISABLE_REASON_DIRECT))
reason_flags |= GLFL_DISABLE_REASON_DIRECT;
if (strstr(reason, LVMETAD_DISABLE_REASON_REPAIR))
reason_flags |= GLFL_DISABLE_REASON_REPAIR;
if (strstr(reason, LVMETAD_DISABLE_REASON_LVM1))
reason_flags |= GLFL_DISABLE_REASON_LVM1;
if (strstr(reason, LVMETAD_DISABLE_REASON_DUPLICATES))
@@ -2432,9 +2418,8 @@ static response get_global_info(lvmetad_state *s, request r)
pid = (int)daemon_request_int(r, "pid", 0);
if (s->flags & GLFL_DISABLE) {
snprintf(reason, REASON_BUF_SIZE - 1, "%s%s%s%s%s",
snprintf(reason, REASON_BUF_SIZE - 1, "%s%s%s%s",
(s->flags & GLFL_DISABLE_REASON_DIRECT) ? LVMETAD_DISABLE_REASON_DIRECT "," : "",
(s->flags & GLFL_DISABLE_REASON_REPAIR) ? LVMETAD_DISABLE_REASON_REPAIR "," : "",
(s->flags & GLFL_DISABLE_REASON_LVM1) ? LVMETAD_DISABLE_REASON_LVM1 "," : "",
(s->flags & GLFL_DISABLE_REASON_DUPLICATES) ? LVMETAD_DISABLE_REASON_DUPLICATES "," : "",
(s->flags & GLFL_DISABLE_REASON_VGRESTORE) ? LVMETAD_DISABLE_REASON_VGRESTORE "," : "");
@@ -2572,12 +2557,14 @@ static void _dump_pairs(struct buffer *buf, struct dm_hash_table *ht, const char
dm_hash_iterate(n, ht) {
const char *key = dm_hash_get_key(ht, n),
*val = dm_hash_get_data(ht, n);
buffer_append(buf, " ");
if (int_key)
(void) dm_asprintf(&append, " %d = \"%s\"\n", *(const int*)key, val);
(void) dm_asprintf(&append, "%d = \"%s\"", *(const int*)key, val);
else
(void) dm_asprintf(&append, " %s = \"%s\"\n", key, val);
(void) dm_asprintf(&append, "%s = \"%s\"", key, val);
if (append)
buffer_append(buf, append);
buffer_append(buf, "\n");
dm_free(append);
}
buffer_append(buf, "}\n");
@@ -2595,9 +2582,11 @@ static void _dump_info_version(struct buffer *buf, struct dm_hash_table *ht, con
while (n) {
const char *key = dm_hash_get_key(ht, n);
info = dm_hash_get_data(ht, n);
(void) dm_asprintf(&append, " %s = %lld\n", key, (long long)info->external_version);
buffer_append(buf, " ");
(void) dm_asprintf(&append, "%s = %lld", key, (long long)info->external_version);
if (append)
buffer_append(buf, append);
buffer_append(buf, "\n");
dm_free(append);
n = dm_hash_get_next(ht, n);
}
@@ -2616,9 +2605,11 @@ static void _dump_info_flags(struct buffer *buf, struct dm_hash_table *ht, const
while (n) {
const char *key = dm_hash_get_key(ht, n);
info = dm_hash_get_data(ht, n);
(void) dm_asprintf(&append, " %s = %llx\n", key, (long long)info->flags);
buffer_append(buf, " ");
(void) dm_asprintf(&append, "%s = %llx", key, (long long)info->flags);
if (append)
buffer_append(buf, append);
buffer_append(buf, "\n");
dm_free(append);
n = dm_hash_get_next(ht, n);
}

View File

@@ -19,12 +19,10 @@ SOURCES = lvmlockd-core.c
ifeq ("@BUILD_LOCKDSANLOCK@", "yes")
SOURCES += lvmlockd-sanlock.c
LOCK_LIBS += -lsanlock_client
endif
ifeq ("@BUILD_LOCKDDLM@", "yes")
SOURCES += lvmlockd-dlm.c
LOCK_LIBS += -ldlm_lt
endif
TARGETS = lvmlockd lvmlockctl
@@ -33,17 +31,29 @@ TARGETS = lvmlockd lvmlockctl
include $(top_builddir)/make.tmpl
CFLAGS += $(EXTRA_EXEC_CFLAGS)
INCLUDES += -I$(top_srcdir)/libdaemon/server
LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
LIBS += $(RT_LIBS) $(DAEMON_LIBS) -ldevmapper $(PTHREAD_LIBS)
LVMLIBS = -ldaemonserver $(LVMINTERNAL_LIBS) -ldevmapper
LIBS += $(PTHREAD_LIBS)
ifeq ("@BUILD_LOCKDSANLOCK@", "yes")
LIBS += -lsanlock_client
endif
ifeq ("@BUILD_LOCKDDLM@", "yes")
LIBS += -ldlm_lt
endif
LDFLAGS += -L$(top_builddir)/libdaemon/server
CLDFLAGS += -L$(top_builddir)/libdaemon/server
lvmlockd: $(OBJECTS) $(top_builddir)/libdaemon/client/libdaemonclient.a \
$(top_builddir)/libdaemon/server/libdaemonserver.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LOCK_LIBS) -ldaemonserver $(LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LVMLIBS) $(LIBS)
lvmlockctl: lvmlockctl.o $(top_builddir)/libdaemon/client/libdaemonclient.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ lvmlockctl.o $(LIBS)
lvmlockctl: lvmlockctl.o $(top_builddir)/libdaemon/client/libdaemonclient.a \
$(top_builddir)/libdaemon/server/libdaemonserver.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ lvmlockctl.o $(LVMLIBS)
install_lvmlockd: lvmlockd
$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)

View File

@@ -379,7 +379,7 @@ static int setup_dump_socket(void)
rv = bind(s, (struct sockaddr *) &dump_addr, dump_addrlen);
if (rv < 0) {
rv = -errno;
if (close(s))
if (!close(s))
log_error("failed to close dump socket");
return rv;
}

View File

@@ -48,6 +48,5 @@ static inline void lvmlockd_close(daemon_handle h)
#define EVGKILLED 217 /* sanlock lost access to leases and VG is killed. */
#define ELOCKIO 218 /* sanlock io errors during lock op, may be transient. */
#define EREMOVED 219
#define EDEVOPEN 220 /* sanlock failed to open lvmlock LV */
#endif /* _LVM_LVMLOCKD_CLIENT_H */

View File

@@ -19,7 +19,6 @@
#include "lvm-version.h"
#include "lvmetad-client.h"
#include "lvmlockd-client.h"
#include "dm-ioctl.h" /* for DM_UUID_LEN */
/* #include <assert.h> */
#include <errno.h>
@@ -2652,16 +2651,10 @@ out_act:
ls->drop_vg = drop_vg;
if (ls->lm_type == LD_LM_DLM && !strcmp(ls->name, gl_lsname_dlm))
global_dlm_lockspace_exists = 0;
/*
* Avoid a name collision of the same lockspace is added again before
* this thread is cleaned up. We just set ls->name to a "junk" value
* for the short period until the struct is freed. We could make it
* blank or fill it with garbage, but instead set it to REM:<name>
* to make it easier to follow progress of freeing is via log_debug.
*/
dm_strncpy(tmp_name, ls->name, sizeof(tmp_name));
snprintf(ls->name, sizeof(ls->name), "REM:%s", tmp_name);
/* Avoid a name collision of the same lockspace is added again before this thread is cleaned up. */
memset(tmp_name, 0, sizeof(tmp_name));
snprintf(tmp_name, MAX_NAME, "REM:%s", ls->name);
memcpy(ls->name, tmp_name, MAX_NAME);
pthread_mutex_unlock(&lockspaces_mutex);
/* worker_thread will join this thread, and free the ls */
@@ -3310,6 +3303,7 @@ static int work_init_lv(struct action *act)
lm_type = ls->lm_type;
memcpy(vg_args, ls->vg_args, MAX_ARGS);
free_offset = ls->free_lock_offset;
ls->free_lock_offset = 0;
}
pthread_mutex_unlock(&lockspaces_mutex);
@@ -3539,15 +3533,11 @@ static int setup_worker_thread(void)
static void close_worker_thread(void)
{
int perrno;
pthread_mutex_lock(&worker_mutex);
worker_stop = 1;
pthread_cond_signal(&worker_cond);
pthread_mutex_unlock(&worker_mutex);
if ((perrno = pthread_join(worker_thread, NULL)))
log_error("pthread_join worker_thread error %d", perrno);
pthread_join(worker_thread, NULL);
}
/* client_mutex is locked */
@@ -3676,17 +3666,7 @@ static int client_send_result(struct client *cl, struct action *act)
if (!gl_lsname_dlm[0])
strcat(result_flags, "NO_GL_LS,");
} else {
int found_lm = 0;
if (lm_support_dlm() && lm_is_running_dlm())
found_lm++;
if (lm_support_sanlock() && lm_is_running_sanlock())
found_lm++;
if (!found_lm)
strcat(result_flags, "NO_GL_LS,NO_LM");
else
strcat(result_flags, "NO_GL_LS");
strcat(result_flags, "NO_GL_LS,");
}
}
@@ -3783,8 +3763,7 @@ static int client_send_result(struct client *cl, struct action *act)
if (dump_fd >= 0) {
/* To avoid deadlock, send data here after the reply. */
send_dump_buf(dump_fd, dump_len);
if (close(dump_fd))
log_error("failed to close dump socket %d", dump_fd);
close(dump_fd);
}
return rv;
@@ -3857,9 +3836,8 @@ static int add_lock_action(struct action *act)
pthread_mutex_lock(&lockspaces_mutex);
if (ls_name[0])
ls = find_lockspace_name(ls_name);
pthread_mutex_unlock(&lockspaces_mutex);
if (!ls) {
pthread_mutex_unlock(&lockspaces_mutex);
if (act->op == LD_OP_UPDATE && act->rt == LD_RT_VG) {
log_debug("lockspace \"%s\" not found ignored for vg update", ls_name);
return -ENOLS;
@@ -4776,8 +4754,8 @@ static void *client_thread_main(void *arg_in)
} else {
pthread_mutex_unlock(&cl->mutex);
}
} else
pthread_mutex_unlock(&client_mutex);
}
pthread_mutex_unlock(&client_mutex);
}
out:
return NULL;
@@ -4801,15 +4779,11 @@ static int setup_client_thread(void)
static void close_client_thread(void)
{
int perrno;
pthread_mutex_lock(&client_mutex);
client_stop = 1;
pthread_cond_signal(&client_cond);
pthread_mutex_unlock(&client_mutex);
if ((perrno = pthread_join(client_thread, NULL)))
log_error("pthread_join client_thread error %d", perrno);
pthread_join(client_thread, NULL);
}
/*
@@ -4933,10 +4907,14 @@ static int get_lockd_vgs(struct list_head *vg_lockd)
continue;
for (lv_cn = md_cn->child; lv_cn; lv_cn = lv_cn->sib) {
snprintf(find_str_path, PATH_MAX, "%s/lock_type", lv_cn->key);
lock_type = dm_config_find_str(lv_cn, find_str_path, NULL);
if (!lock_type)
continue;
snprintf(find_str_path, PATH_MAX, "%s/lock_args", lv_cn->key);
lock_args = dm_config_find_str(lv_cn, find_str_path, NULL);
if (!lock_args)
continue;
snprintf(find_str_path, PATH_MAX, "%s/id", lv_cn->key);
lv_uuid = dm_config_find_str(lv_cn, find_str_path, NULL);
@@ -4982,7 +4960,7 @@ out:
return rv;
}
static char _dm_uuid[DM_UUID_LEN];
static char _dm_uuid[64];
static char *get_dm_uuid(char *dm_name)
{
@@ -5201,17 +5179,20 @@ static void adopt_locks(void)
* Get list of lockspaces from lock managers.
* Get list of VGs from lvmetad with a lockd type.
* Get list of active lockd type LVs from /dev.
*
* ECONNREFUSED means the lock manager is not running.
* This is expected for at least one of them.
*/
if (lm_support_dlm() && lm_is_running_dlm()) {
if (lm_support_dlm()) {
rv = lm_get_lockspaces_dlm(&ls_found);
if (rv < 0)
if ((rv < 0) && (rv != -ECONNREFUSED))
goto fail;
}
if (lm_support_sanlock() && lm_is_running_sanlock()) {
if (lm_support_sanlock()) {
rv = lm_get_lockspaces_sanlock(&ls_found);
if (rv < 0)
if ((rv < 0) && (rv != -ECONNREFUSED))
goto fail;
}
@@ -5288,7 +5269,7 @@ static void adopt_locks(void)
list_for_each_entry_safe(ls1, l1safe, &ls_found, list) {
/* The dlm global lockspace is special and doesn't match a VG. */
if ((ls1->lm_type == LD_LM_DLM) && !strcmp(ls1->name, gl_lsname_dlm)) {
if (!strcmp(ls1->name, gl_lsname_dlm)) {
list_del(&ls1->list);
free(ls1);
continue;

View File

@@ -224,10 +224,7 @@ static int lock_lv_offset_from_args(char *lv_args, uint64_t *lock_lv_offset)
if (rv < 0)
return rv;
errno = 0;
*lock_lv_offset = strtoull(offset_str, NULL, 10);
if (errno)
return -1;
return 0;
}
@@ -356,19 +353,12 @@ int lm_init_vg_sanlock(char *ls_name, char *vg_name, uint32_t flags, char *vg_ar
log_debug("sanlock daemon version %08x proto %08x",
daemon_version, daemon_proto);
rv = sanlock_align(&disk);
if (rv <= 0) {
if (rv == -EACCES) {
log_error("S %s init_vg_san sanlock error -EACCES: no permission to access %s",
ls_name, disk.path);
return -EDEVOPEN;
} else {
log_error("S %s init_vg_san sanlock error %d trying to get align size of %s",
ls_name, rv, disk.path);
return -EARGS;
}
} else
align_size = rv;
align_size = sanlock_align(&disk);
if (align_size <= 0) {
log_error("S %s init_vg_san bad disk align size %d %s",
ls_name, align_size, disk.path);
return -EARGS;
}
strncpy(ss.name, ls_name, SANLK_NAME_LEN);
memcpy(ss.host_id_disk.path, disk.path, SANLK_PATH_LEN);
@@ -945,9 +935,7 @@ int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset)
struct lm_sanlock *lms = (struct lm_sanlock *)ls->lm_data;
struct sanlk_resourced rd;
uint64_t offset;
uint64_t start_offset;
int rv;
int round = 0;
if (daemon_test) {
*free_offset = (1048576 * LV_LOCK_BEGIN) + (1048576 * (daemon_test_lv_count + 1));
@@ -960,22 +948,9 @@ int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset)
rd.rs.num_disks = 1;
strncpy(rd.rs.disks[0].path, lms->ss.host_id_disk.path, SANLK_PATH_LEN-1);
if (ls->free_lock_offset)
offset = ls->free_lock_offset;
else
offset = lms->align_size * LV_LOCK_BEGIN;
start_offset = offset;
offset = lms->align_size * LV_LOCK_BEGIN;
while (1) {
if (offset >= start_offset && round) {
/* This indicates the all space are allocated. */
log_debug("S %s init_lv_san read back to start offset %llu",
ls->name, (unsigned long long)offset);
rv = -EMSGSIZE;
return rv;
}
rd.rs.disks[0].offset = offset;
memset(rd.rs.name, 0, SANLK_NAME_LEN);
@@ -985,14 +960,7 @@ int lm_find_free_lock_sanlock(struct lockspace *ls, uint64_t *free_offset)
/* This indicates the end of the device is reached. */
log_debug("S %s find_free_lock_san read limit offset %llu",
ls->name, (unsigned long long)offset);
/* remember the NO SPACE offset, if no free area left,
* search from this offset after extend */
*free_offset = offset;
offset = lms->align_size * LV_LOCK_BEGIN;
round = 1;
continue;
return -EMSGSIZE;
}
/*

View File

@@ -27,14 +27,18 @@ CFLOW_TARGET = lvmpolld
include $(top_builddir)/make.tmpl
CFLAGS += $(EXTRA_EXEC_CFLAGS)
INCLUDES += -I$(top_srcdir)/libdaemon/server
LDFLAGS += -L$(top_builddir)/libdaemon/server $(EXTRA_EXEC_LDFLAGS) $(ELDFLAGS)
LIBS += $(DAEMON_LIBS) -ldaemonserver -ldevmapper $(PTHREAD_LIBS)
LVMLIBS = -ldaemonserver $(LVMINTERNAL_LIBS) -ldevmapper
LIBS += $(PTHREAD_LIBS)
LDFLAGS += -L$(top_builddir)/libdaemon/server $(DAEMON_LDFLAGS)
CLDFLAGS += -L$(top_builddir)/libdaemon/server
CFLAGS += $(DAEMON_CFLAGS)
lvmpolld: $(OBJECTS) $(top_builddir)/libdaemon/client/libdaemonclient.a \
$(top_builddir)/libdaemon/server/libdaemonserver.a
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LIBS)
$(CC) $(CFLAGS) $(LDFLAGS) -o $@ $(OBJECTS) $(LVMLIBS) $(LIBS)
install_lvmpolld: lvmpolld
$(INSTALL_PROGRAM) -D $< $(sbindir)/$(<F)

View File

@@ -19,12 +19,10 @@
#define MIN_ARGV_SIZE 8
static const char *const polling_ops[] = {
[PVMOVE] = LVMPD_REQ_PVMOVE,
[CONVERT] = LVMPD_REQ_CONVERT,
[MERGE] = LVMPD_REQ_MERGE,
[MERGE_THIN] = LVMPD_REQ_MERGE_THIN
};
static const char *const const polling_ops[] = { [PVMOVE] = LVMPD_REQ_PVMOVE,
[CONVERT] = LVMPD_REQ_CONVERT,
[MERGE] = LVMPD_REQ_MERGE,
[MERGE_THIN] = LVMPD_REQ_MERGE_THIN };
const char *polling_op(enum poll_type type)
{

View File

@@ -207,10 +207,6 @@ Optional feature arguments are:
block, then the cache block is invalidated.
To enable passthrough mode the cache must be clean.
metadata2 : use version 2 of the metadata. This stores the dirty bits
in a separate btree, which improves speed of shutting
down the cache.
A policy called 'default' is always registered. This is an alias for
the policy we currently think is giving best all round performance.
@@ -290,7 +286,7 @@ message, which takes an arbitrary number of cblock ranges. Each cblock
range's end value is "one past the end", meaning 5-10 expresses a range
of values from 5 to 9. Each cblock must be expressed as a decimal
value, in the future a variant message that takes cblock ranges
expressed in hexadecimal may be needed to better support efficient
expressed in hexidecimal may be needed to better support efficient
invalidation of larger caches. The cache must be in passthrough mode
when invalidate_cblocks is used.

View File

@@ -11,57 +11,23 @@ Parameters: <cipher> <key> <iv_offset> <device path> \
<offset> [<#opt_params> <opt_params>]
<cipher>
Encryption cipher, encryption mode and Initial Vector (IV) generator.
The cipher specifications format is:
cipher[:keycount]-chainmode-ivmode[:ivopts]
Encryption cipher and an optional IV generation mode.
(In format cipher[:keycount]-chainmode-ivmode[:ivopts]).
Examples:
des
aes-cbc-essiv:sha256
aes-xts-plain64
serpent-xts-plain64
twofish-ecb
Cipher format also supports direct specification with kernel crypt API
format (selected by capi: prefix). The IV specification is the same
as for the first format type.
This format is mainly used for specification of authenticated modes.
The crypto API cipher specifications format is:
capi:cipher_api_spec-ivmode[:ivopts]
Examples:
capi:cbc(aes)-essiv:sha256
capi:xts(aes)-plain64
Examples of authenticated modes:
capi:gcm(aes)-random
capi:authenc(hmac(sha256),xts(aes))-random
capi:rfc7539(chacha20,poly1305)-random
The /proc/crypto contains a list of curently loaded crypto modes.
/proc/crypto contains supported crypto modes
<key>
Key used for encryption. It is encoded either as a hexadecimal number
or it can be passed as <key_string> prefixed with single colon
character (':') for keys residing in kernel keyring service.
Key used for encryption. It is encoded as a hexadecimal number.
You can only use key sizes that are valid for the selected cipher
in combination with the selected iv mode.
Note that for some iv modes the key string can contain additional
keys (for example IV seed) so the key contains more parts concatenated
into a single string.
<key_string>
The kernel keyring key is identified by string in following format:
<key_size>:<key_type>:<key_description>.
<key_size>
The encryption key size in bytes. The kernel key payload size must match
the value passed in <key_size>.
<key_type>
Either 'logon' or 'user' kernel key type.
<key_description>
The kernel keyring key description crypt target should look for
when loading key of <key_type>.
<keycount>
Multi-key compatibility mode. You can define <keycount> keys and
then sectors are encrypted according to their offsets (sector 0 uses key0;
@@ -110,32 +76,6 @@ submit_from_crypt_cpus
thread because it benefits CFQ to have writes submitted using the
same context.
integrity:<bytes>:<type>
The device requires additional <bytes> metadata per-sector stored
in per-bio integrity structure. This metadata must by provided
by underlying dm-integrity target.
The <type> can be "none" if metadata is used only for persistent IV.
For Authenticated Encryption with Additional Data (AEAD)
the <type> is "aead". An AEAD mode additionally calculates and verifies
integrity for the encrypted device. The additional space is then
used for storing authentication tag (and persistent IV if needed).
sector_size:<bytes>
Use <bytes> as the encryption unit instead of 512 bytes sectors.
This option can be in range 512 - 4096 bytes and must be power of two.
Virtual device will announce this size as a minimal IO and logical sector.
iv_large_sectors
IV generators will use sector number counted in <sector_size> units
instead of default 512 bytes sectors.
For example, if <sector_size> is 4096 bytes, plain64 IV for the second
sector will be 8 (without flag) and 1 if iv_large_sectors is present.
The <iv_offset> must be multiple of <sector_size> (in 512 bytes units)
if this flag is specified.
Example scripts
===============
LUKS (Linux Unified Key Setup) is now the preferred way to set up disk
@@ -145,13 +85,7 @@ https://gitlab.com/cryptsetup/cryptsetup
[[
#!/bin/sh
# Create a crypt device using dmsetup
dmsetup create crypt1 --table "0 `blockdev --getsz $1` crypt aes-cbc-essiv:sha256 babebabebabebabebabebabebabebabe 0 $1 0"
]]
[[
#!/bin/sh
# Create a crypt device using dmsetup when encryption key is stored in keyring service
dmsetup create crypt2 --table "0 `blockdev --getsize $1` crypt aes-cbc-essiv:sha256 :32:logon:my_prefix:my_key 0 $1 0"
dmsetup create crypt1 --table "0 `blockdev --getsize $1` crypt aes-cbc-essiv:sha256 babebabebabebabebabebabebabebabe 0 $1 0"
]]
[[

View File

@@ -16,12 +16,12 @@ Example scripts
[[
#!/bin/sh
# Create device delaying rw operation for 500ms
echo "0 `blockdev --getsz $1` delay $1 0 500" | dmsetup create delayed
echo "0 `blockdev --getsize $1` delay $1 0 500" | dmsetup create delayed
]]
[[
#!/bin/sh
# Create device delaying only write operation for 500ms and
# splitting reads and writes to different devices $1 $2
echo "0 `blockdev --getsz $1` delay $1 0 0 $2 0 500" | dmsetup create delayed
echo "0 `blockdev --getsize $1` delay $1 0 0 $2 0 500" | dmsetup create delayed
]]

View File

@@ -42,7 +42,7 @@ Optional feature parameters:
<direction>: Either 'r' to corrupt reads or 'w' to corrupt writes.
'w' is incompatible with drop_writes.
<value>: The value (from 0-255) to write.
<flags>: Perform the replacement only if bio->bi_opf has all the
<flags>: Perform the replacement only if bio->bi_rw has all the
selected flags set.
Examples:

View File

@@ -1,199 +0,0 @@
The dm-integrity target emulates a block device that has additional
per-sector tags that can be used for storing integrity information.
A general problem with storing integrity tags with every sector is that
writing the sector and the integrity tag must be atomic - i.e. in case of
crash, either both sector and integrity tag or none of them is written.
To guarantee write atomicity, the dm-integrity target uses journal, it
writes sector data and integrity tags into a journal, commits the journal
and then copies the data and integrity tags to their respective location.
The dm-integrity target can be used with the dm-crypt target - in this
situation the dm-crypt target creates the integrity data and passes them
to the dm-integrity target via bio_integrity_payload attached to the bio.
In this mode, the dm-crypt and dm-integrity targets provide authenticated
disk encryption - if the attacker modifies the encrypted device, an I/O
error is returned instead of random data.
The dm-integrity target can also be used as a standalone target, in this
mode it calculates and verifies the integrity tag internally. In this
mode, the dm-integrity target can be used to detect silent data
corruption on the disk or in the I/O path.
When loading the target for the first time, the kernel driver will format
the device. But it will only format the device if the superblock contains
zeroes. If the superblock is neither valid nor zeroed, the dm-integrity
target can't be loaded.
To use the target for the first time:
1. overwrite the superblock with zeroes
2. load the dm-integrity target with one-sector size, the kernel driver
will format the device
3. unload the dm-integrity target
4. read the "provided_data_sectors" value from the superblock
5. load the dm-integrity target with the the target size
"provided_data_sectors"
6. if you want to use dm-integrity with dm-crypt, load the dm-crypt target
with the size "provided_data_sectors"
Target arguments:
1. the underlying block device
2. the number of reserved sector at the beginning of the device - the
dm-integrity won't read of write these sectors
3. the size of the integrity tag (if "-" is used, the size is taken from
the internal-hash algorithm)
4. mode:
D - direct writes (without journal) - in this mode, journaling is
not used and data sectors and integrity tags are written
separately. In case of crash, it is possible that the data
and integrity tag doesn't match.
J - journaled writes - data and integrity tags are written to the
journal and atomicity is guaranteed. In case of crash,
either both data and tag or none of them are written. The
journaled mode degrades write throughput twice because the
data have to be written twice.
R - recovery mode - in this mode, journal is not replayed,
checksums are not checked and writes to the device are not
allowed. This mode is useful for data recovery if the
device cannot be activated in any of the other standard
modes.
5. the number of additional arguments
Additional arguments:
journal_sectors:number
The size of journal, this argument is used only if formatting the
device. If the device is already formatted, the value from the
superblock is used.
interleave_sectors:number
The number of interleaved sectors. This values is rounded down to
a power of two. If the device is already formatted, the value from
the superblock is used.
buffer_sectors:number
The number of sectors in one buffer. The value is rounded down to
a power of two.
The tag area is accessed using buffers, the buffer size is
configurable. The large buffer size means that the I/O size will
be larger, but there could be less I/Os issued.
journal_watermark:number
The journal watermark in percents. When the size of the journal
exceeds this watermark, the thread that flushes the journal will
be started.
commit_time:number
Commit time in milliseconds. When this time passes, the journal is
written. The journal is also written immediatelly if the FLUSH
request is received.
internal_hash:algorithm(:key) (the key is optional)
Use internal hash or crc.
When this argument is used, the dm-integrity target won't accept
integrity tags from the upper target, but it will automatically
generate and verify the integrity tags.
You can use a crc algorithm (such as crc32), then integrity target
will protect the data against accidental corruption.
You can also use a hmac algorithm (for example
"hmac(sha256):0123456789abcdef"), in this mode it will provide
cryptographic authentication of the data without encryption.
When this argument is not used, the integrity tags are accepted
from an upper layer target, such as dm-crypt. The upper layer
target should check the validity of the integrity tags.
journal_crypt:algorithm(:key) (the key is optional)
Encrypt the journal using given algorithm to make sure that the
attacker can't read the journal. You can use a block cipher here
(such as "cbc(aes)") or a stream cipher (for example "chacha20",
"salsa20", "ctr(aes)" or "ecb(arc4)").
The journal contains history of last writes to the block device,
an attacker reading the journal could see the last sector nubmers
that were written. From the sector numbers, the attacker can infer
the size of files that were written. To protect against this
situation, you can encrypt the journal.
journal_mac:algorithm(:key) (the key is optional)
Protect sector numbers in the journal from accidental or malicious
modification. To protect against accidental modification, use a
crc algorithm, to protect against malicious modification, use a
hmac algorithm with a key.
This option is not needed when using internal-hash because in this
mode, the integrity of journal entries is checked when replaying
the journal. Thus, modified sector number would be detected at
this stage.
block_size:number
The size of a data block in bytes. The larger the block size the
less overhead there is for per-block integrity metadata.
Supported values are 512, 1024, 2048 and 4096 bytes. If not
specified the default block size is 512 bytes.
The journal mode (D/J), buffer_sectors, journal_watermark, commit_time can
be changed when reloading the target (load an inactive table and swap the
tables with suspend and resume). The other arguments should not be changed
when reloading the target because the layout of disk data depend on them
and the reloaded target would be non-functional.
The layout of the formatted block device:
* reserved sectors (they are not used by this target, they can be used for
storing LUKS metadata or for other purpose), the size of the reserved
area is specified in the target arguments
* superblock (4kiB)
* magic string - identifies that the device was formatted
* version
* log2(interleave sectors)
* integrity tag size
* the number of journal sections
* provided data sectors - the number of sectors that this target
provides (i.e. the size of the device minus the size of all
metadata and padding). The user of this target should not send
bios that access data beyond the "provided data sectors" limit.
* flags - a flag is set if journal_mac is used
* journal
The journal is divided into sections, each section contains:
* metadata area (4kiB), it contains journal entries
every journal entry contains:
* logical sector (specifies where the data and tag should
be written)
* last 8 bytes of data
* integrity tag (the size is specified in the superblock)
every metadata sector ends with
* mac (8-bytes), all the macs in 8 metadata sectors form a
64-byte value. It is used to store hmac of sector
numbers in the journal section, to protect against a
possibility that the attacker tampers with sector
numbers in the journal.
* commit id
* data area (the size is variable; it depends on how many journal
entries fit into the metadata area)
every sector in the data area contains:
* data (504 bytes of data, the last 8 bytes are stored in
the journal entry)
* commit id
To test if the whole journal section was written correctly, every
512-byte sector of the journal ends with 8-byte commit id. If the
commit id matches on all sectors in a journal section, then it is
assumed that the section was written correctly. If the commit id
doesn't match, the section was written partially and it should not
be replayed.
* one or more runs of interleaved tags and data. Each run contains:
* tag area - it contains integrity tags. There is one tag for each
sector in the data area
* data area - it contains data sectors. The number of data sectors
in one run must be a power of two. log2 of this value is stored
in the superblock.

View File

@@ -16,15 +16,15 @@ Example scripts
[[
#!/bin/sh
# Create an identity mapping for a device
echo "0 `blockdev --getsz $1` linear $1 0" | dmsetup create identity
echo "0 `blockdev --getsize $1` linear $1 0" | dmsetup create identity
]]
[[
#!/bin/sh
# Join 2 devices together
size1=`blockdev --getsz $1`
size2=`blockdev --getsz $2`
size1=`blockdev --getsize $1`
size2=`blockdev --getsize $2`
echo "0 $size1 linear $1 0
$size1 $size2 linear $2 0" | dmsetup create joined
]]
@@ -44,7 +44,7 @@ if (!defined($dev)) {
die("Please specify a device.\n");
}
my $dev_size = `blockdev --getsz $dev`;
my $dev_size = `blockdev --getsize $dev`;
my $extents = int($dev_size / $extent_size) -
(($dev_size % $extent_size) ? 1 : 0);

View File

@@ -14,14 +14,14 @@ Log Ordering
We log things in order of completion once we are sure the write is no longer in
cache. This means that normal WRITE requests are not actually logged until the
next REQ_PREFLUSH request. This is to make it easier for userspace to replay
the log in a way that correlates to what is on disk and not what is in cache,
to make it easier to detect improper waiting/flushing.
next REQ_FLUSH request. This is to make it easier for userspace to replay the
log in a way that correlates to what is on disk and not what is in cache, to
make it easier to detect improper waiting/flushing.
This works by attaching all WRITE requests to a list once the write completes.
Once we see a REQ_PREFLUSH request we splice this list onto the request and once
Once we see a REQ_FLUSH request we splice this list onto the request and once
the FLUSH request completes we log all of the WRITEs and then the FLUSH. Only
completed WRITEs, at the time the REQ_PREFLUSH is issued, are added in order to
completed WRITEs, at the time the REQ_FLUSH is issued, are added in order to
simulate the worst case scenario with regard to power failures. Consider the
following example (W means write, C means complete):

View File

@@ -14,12 +14,8 @@ The target is named "raid" and it accepts the following parameters:
<#raid_devs> <metadata_dev0> <dev0> [.. <metadata_devN> <devN>]
<raid_type>:
raid0 RAID0 striping (no resilience)
raid1 RAID1 mirroring
raid4 RAID4 with dedicated last parity disk
raid5_n RAID5 with dedicated last parity disk supporting takeover
Same as raid4
-Transitory layout
raid4 RAID4 dedicated parity disk
raid5_la RAID5 left asymmetric
- rotating parity 0 with data continuation
raid5_ra RAID5 right asymmetric
@@ -34,19 +30,7 @@ The target is named "raid" and it accepts the following parameters:
- rotating parity N (right-to-left) with data restart
raid6_nc RAID6 N continue
- rotating parity N (right-to-left) with data continuation
raid6_n_6 RAID6 with dedicate parity disks
- parity and Q-syndrome on the last 2 disks;
layout for takeover from/to raid4/raid5_n
raid6_la_6 Same as "raid_la" plus dedicated last Q-syndrome disk
- layout for takeover from raid5_la from/to raid6
raid6_ra_6 Same as "raid5_ra" dedicated last Q-syndrome disk
- layout for takeover from raid5_ra from/to raid6
raid6_ls_6 Same as "raid5_ls" dedicated last Q-syndrome disk
- layout for takeover from raid5_ls from/to raid6
raid6_rs_6 Same as "raid5_rs" dedicated last Q-syndrome disk
- layout for takeover from raid5_rs from/to raid6
raid10 Various RAID10 inspired algorithms chosen by additional params
(see raid10_format and raid10_copies below)
- RAID10: Striped Mirrors (aka 'Striping on top of mirrors')
- RAID1E: Integrated Adjacent Stripe Mirroring
- RAID1E: Integrated Offset Stripe Mirroring
@@ -132,57 +116,10 @@ The target is named "raid" and it accepts the following parameters:
Here we see layouts closely akin to 'RAID1E - Integrated
Offset Stripe Mirroring'.
[delta_disks <N>]
The delta_disks option value (-251 < N < +251) triggers
device removal (negative value) or device addition (positive
value) to any reshape supporting raid levels 4/5/6 and 10.
RAID levels 4/5/6 allow for addition of devices (metadata
and data device tuple), raid10_near and raid10_offset only
allow for device addition. raid10_far does not support any
reshaping at all.
A minimum of devices have to be kept to enforce resilience,
which is 3 devices for raid4/5 and 4 devices for raid6.
[data_offset <sectors>]
This option value defines the offset into each data device
where the data starts. This is used to provide out-of-place
reshaping space to avoid writing over data whilst
changing the layout of stripes, hence an interruption/crash
may happen at any time without the risk of losing data.
E.g. when adding devices to an existing raid set during
forward reshaping, the out-of-place space will be allocated
at the beginning of each raid device. The kernel raid4/5/6/10
MD personalities supporting such device addition will read the data from
the existing first stripes (those with smaller number of stripes)
starting at data_offset to fill up a new stripe with the larger
number of stripes, calculate the redundancy blocks (CRC/Q-syndrome)
and write that new stripe to offset 0. Same will be applied to all
N-1 other new stripes. This out-of-place scheme is used to change
the RAID type (i.e. the allocation algorithm) as well, e.g.
changing from raid5_ls to raid5_n.
[journal_dev <dev>]
This option adds a journal device to raid4/5/6 raid sets and
uses it to close the 'write hole' caused by the non-atomic updates
to the component devices which can cause data loss during recovery.
The journal device is used as writethrough thus causing writes to
be throttled versus non-journaled raid4/5/6 sets.
Takeover/reshape is not possible with a raid4/5/6 journal device;
it has to be deconfigured before requesting these.
[journal_mode <mode>]
This option sets the caching mode on journaled raid4/5/6 raid sets
(see 'journal_dev <dev>' above) to 'writethrough' or 'writeback'.
If 'writeback' is selected the journal device has to be resilient
and must not suffer from the 'write hole' problem itself (e.g. use
raid1 or raid10) to avoid a single point of failure.
<#raid_devs>: The number of devices composing the array.
Each device consists of two entries. The first is the device
containing the metadata (if any); the second is the one containing the
data. A Maximum of 64 metadata/data device entries are supported
up to target version 1.8.0.
1.9.0 supports up to 253 which is enforced by the used MD kernel runtime.
data.
If a drive has failed or is missing at creation time, a '-' can be
given for both the metadata and data drives for a given position.
@@ -258,14 +195,6 @@ recovery. Here is a fuller description of the individual fields:
in RAID1/10 or wrong parity values found in RAID4/5/6.
This value is valid only after a "check" of the array
is performed. A healthy array has a 'mismatch_cnt' of 0.
<data_offset> The current data offset to the start of the user data on
each component device of a raid set (see the respective
raid parameter to support out-of-place reshaping).
<journal_char> 'A' - active write-through journal device.
'a' - active write-back journal device.
'D' - dead journal device.
'-' - no journal device.
Message Interface
-----------------
@@ -278,6 +207,7 @@ include:
"recover"- Initiate/continue a recover process.
"check" - Initiate a check (i.e. a "scrub") of the array.
"repair" - Initiate a repair of the array.
"reshape"- Currently unsupported (-EINVAL).
Discard Support
@@ -327,19 +257,3 @@ Version History
1.5.2 'mismatch_cnt' is zero unless [last_]sync_action is "check".
1.6.0 Add discard support (and devices_handle_discard_safely module param).
1.7.0 Add support for MD RAID0 mappings.
1.8.0 Explicitly check for compatible flags in the superblock metadata
and reject to start the raid set if any are set by a newer
target version, thus avoiding data corruption on a raid set
with a reshape in progress.
1.9.0 Add support for RAID level takeover/reshape/region size
and set size reduction.
1.9.1 Fix activation of existing RAID 4/10 mapped devices
1.9.2 Don't emit '- -' on the status table line in case the constructor
fails reading a superblock. Correctly emit 'maj:min1 maj:min2' and
'D' on the status line. If '- -' is passed into the constructor, emit
'- -' on the table line and '-' as the status line health character.
1.10.0 Add support for raid4/5/6 journal device
1.10.1 Fix data corruption on reshape request
1.11.0 Fix table line argument order
(wrong raid10_copies/raid10_format sequence)
1.11.1 Add raid4/5/6 journal write-back support via journal_mode option

View File

@@ -37,9 +37,9 @@ if (!$num_devs) {
die("Specify at least one device\n");
}
$min_dev_size = `blockdev --getsz $devs[0]`;
$min_dev_size = `blockdev --getsize $devs[0]`;
for ($i = 1; $i < $num_devs; $i++) {
my $this_size = `blockdev --getsz $devs[$i]`;
my $this_size = `blockdev --getsize $devs[$i]`;
$min_dev_size = ($min_dev_size < $this_size) ?
$min_dev_size : $this_size;
}

View File

@@ -123,7 +123,7 @@ Assume that you have volumes vg1/switch0 vg1/switch1 vg1/switch2 with
the same size.
Create a switch device with 64kB region size:
dmsetup create switch --table "0 `blockdev --getsz /dev/vg1/switch0`
dmsetup create switch --table "0 `blockdev --getsize /dev/vg1/switch0`
switch 3 128 0 /dev/vg1/switch0 0 /dev/vg1/switch1 0 /dev/vg1/switch2 0"
Set mappings for the first 7 entries to point to devices switch0, switch1,

View File

@@ -1,144 +0,0 @@
dm-zoned
========
The dm-zoned device mapper target exposes a zoned block device (ZBC and
ZAC compliant devices) as a regular block device without any write
pattern constraints. In effect, it implements a drive-managed zoned
block device which hides from the user (a file system or an application
doing raw block device accesses) the sequential write constraints of
host-managed zoned block devices and can mitigate the potential
device-side performance degradation due to excessive random writes on
host-aware zoned block devices.
For a more detailed description of the zoned block device models and
their constraints see (for SCSI devices):
http://www.t10.org/drafts.htm#ZBC_Family
and (for ATA devices):
http://www.t13.org/Documents/UploadedDocuments/docs2015/di537r05-Zoned_Device_ATA_Command_Set_ZAC.pdf
The dm-zoned implementation is simple and minimizes system overhead (CPU
and memory usage as well as storage capacity loss). For a 10TB
host-managed disk with 256 MB zones, dm-zoned memory usage per disk
instance is at most 4.5 MB and as little as 5 zones will be used
internally for storing metadata and performaing reclaim operations.
dm-zoned target devices are formatted and checked using the dmzadm
utility available at:
https://github.com/hgst/dm-zoned-tools
Algorithm
=========
dm-zoned implements an on-disk buffering scheme to handle non-sequential
write accesses to the sequential zones of a zoned block device.
Conventional zones are used for caching as well as for storing internal
metadata.
The zones of the device are separated into 2 types:
1) Metadata zones: these are conventional zones used to store metadata.
Metadata zones are not reported as useable capacity to the user.
2) Data zones: all remaining zones, the vast majority of which will be
sequential zones used exclusively to store user data. The conventional
zones of the device may be used also for buffering user random writes.
Data in these zones may be directly mapped to the conventional zone, but
later moved to a sequential zone so that the conventional zone can be
reused for buffering incoming random writes.
dm-zoned exposes a logical device with a sector size of 4096 bytes,
irrespective of the physical sector size of the backend zoned block
device being used. This allows reducing the amount of metadata needed to
manage valid blocks (blocks written).
The on-disk metadata format is as follows:
1) The first block of the first conventional zone found contains the
super block which describes the on disk amount and position of metadata
blocks.
2) Following the super block, a set of blocks is used to describe the
mapping of the logical device blocks. The mapping is done per chunk of
blocks, with the chunk size equal to the zoned block device size. The
mapping table is indexed by chunk number and each mapping entry
indicates the zone number of the device storing the chunk of data. Each
mapping entry may also indicate if the zone number of a conventional
zone used to buffer random modification to the data zone.
3) A set of blocks used to store bitmaps indicating the validity of
blocks in the data zones follows the mapping table. A valid block is
defined as a block that was written and not discarded. For a buffered
data chunk, a block is always valid only in the data zone mapping the
chunk or in the buffer zone of the chunk.
For a logical chunk mapped to a conventional zone, all write operations
are processed by directly writing to the zone. If the mapping zone is a
sequential zone, the write operation is processed directly only if the
write offset within the logical chunk is equal to the write pointer
offset within of the sequential data zone (i.e. the write operation is
aligned on the zone write pointer). Otherwise, write operations are
processed indirectly using a buffer zone. In that case, an unused
conventional zone is allocated and assigned to the chunk being
accessed. Writing a block to the buffer zone of a chunk will
automatically invalidate the same block in the sequential zone mapping
the chunk. If all blocks of the sequential zone become invalid, the zone
is freed and the chunk buffer zone becomes the primary zone mapping the
chunk, resulting in native random write performance similar to a regular
block device.
Read operations are processed according to the block validity
information provided by the bitmaps. Valid blocks are read either from
the sequential zone mapping a chunk, or if the chunk is buffered, from
the buffer zone assigned. If the accessed chunk has no mapping, or the
accessed blocks are invalid, the read buffer is zeroed and the read
operation terminated.
After some time, the limited number of convnetional zones available may
be exhausted (all used to map chunks or buffer sequential zones) and
unaligned writes to unbuffered chunks become impossible. To avoid this
situation, a reclaim process regularly scans used conventional zones and
tries to reclaim the least recently used zones by copying the valid
blocks of the buffer zone to a free sequential zone. Once the copy
completes, the chunk mapping is updated to point to the sequential zone
and the buffer zone freed for reuse.
Metadata Protection
===================
To protect metadata against corruption in case of sudden power loss or
system crash, 2 sets of metadata zones are used. One set, the primary
set, is used as the main metadata region, while the secondary set is
used as a staging area. Modified metadata is first written to the
secondary set and validated by updating the super block in the secondary
set, a generation counter is used to indicate that this set contains the
newest metadata. Once this operation completes, in place of metadata
block updates can be done in the primary metadata set. This ensures that
one of the set is always consistent (all modifications committed or none
at all). Flush operations are used as a commit point. Upon reception of
a flush request, metadata modification activity is temporarily blocked
(for both incoming BIO processing and reclaim process) and all dirty
metadata blocks are staged and updated. Normal operation is then
resumed. Flushing metadata thus only temporarily delays write and
discard requests. Read requests can be processed concurrently while
metadata flush is being executed.
Usage
=====
A zoned block device must first be formatted using the dmzadm tool. This
will analyze the device zone configuration, determine where to place the
metadata sets on the device and initialize the metadata sets.
Ex:
dmzadm --format /dev/sdxx
For a formatted device, the target can be created normally with the
dmsetup utility. The only parameter that dm-zoned requires is the
underlying zoned block device name. Ex:
echo "0 `blockdev --getsize ${dev}` zoned ${dev}" | dmsetup create dmz-`basename ${dev}`

View File

@@ -1,246 +0,0 @@
LVM disk reading
Reading disks happens in two phases. The first is a discovery phase,
which determines what's on the disks. The second is a working phase,
which does a particular job for the command.
Phase 1: Discovery
------------------
Read all the disks on the system to find out:
- What are the LVM devices?
- What VG's exist on those devices?
This phase is called "label scan" (although it reads and scans everything,
not just the label.) It stores the information it discovers (what LVM
devices exist, and what VGs exist on them) in lvmcache. The devs/VGs info
in lvmcache is the starting point for phase two.
Phase 1 in outline:
For each device:
a. Read the first <N> KB of the device. (N is configurable.)
b. Look for the lvm label_header in the first four sectors,
if none exists, it's not an lvm device, so quit looking at it.
(By default, label_header is in the second sector.)
c. Look at the pv_header, which follows the label_header.
This tells us the location of VG metadata on the device.
There can be 0, 1 or 2 copies of VG metadata. The first
is always at the start of the device, the second (if used)
is at the end.
d. Look at the first mda_header (location came from pv_header
in the previous step). This is by default in sector 8,
4096 bytes from the start of the device. This tells us the
location of the actual VG metadata text.
e. Look at the first copy of the text VG metadata (location came
from mda_header in the previous step). This is by default
in sector 9, 4608 bytes from the start of the device.
The VG metadata is only partially analyzed to create a basic
summary of the VG.
f. Store an "info" entry in lvmcache for this device,
indicating that it is an lvm device, and store a "vginfo"
entry in lvmcache indicating the name of the VG seen
in the metadata in step e.
g. If the pv_header in step c shows a second mda_header
location at the end of the device, then read that as
in step d, and repeat steps e-f for it.
At the end of phase 1, lvmcache will have a list of devices
that belong to LVM, and a list of VG names that exist on
those devices. Each device (info struct) is associated
with the VG (vginfo struct) it is used in.
If the number <N> of KB read in step (a) was large enough, then
all the structs/metadata needed in steps b-e will be found
in the data buffer returned by a. If a particular struct
or metadata needed in steps b-e are located outside the range
of the initial read, then those steps need to issue their own
read at the necessary location to get that bit of data.
(The optional second mda_header and VG metadata in step g
is located at the end of the device, and will always require
an additional read.)
Phase 1 in code:
The most relevant functions are listed for each step in the outline.
lvmcache_label_scan()
label_scan()
_label_scan_async()
for each dev: dev = dev_iter_get(iter) ...
a. _label_read_async_start()
b. _label_read_data_process()
_find_label_header()
c. _label_read_data_process()
ops->read()
_text_read()
d. _read_mda_header_and_metadata()
raw_read_mda_header()
e. _read_mda_header_and_metadata()
read_metadata_location()
text_read_metadata_summary()
config_file_read_fd()
ops->read_vgsummary()
_read_vgsummary()
f. _text_read(): lvmcache_add()
[adds this device to list of lvm devices]
_read_mda_header_and_metadata(): lvmcache_update_vgname_and_id()
[adds the VG name to list of VGs]
Phase 1 in log messages:
For each device:
Scanning data from all devs async
a. Reading sectors from device <dev>
b. Parsing label and data from device <dev>
d. Copying mda header sector from <dev> ...
or if the mda_header needs to be read from disk:
Reading mda header sector from <dev> ...
e. Copying metadata summary for <dev> ...
or if the metadata needs to be read from disk:
Reading metadata summary for <dev> ...
f. lvmcache <dev> ...
Phase 2: Work
-------------
This phase carries out the operation requested by the command that was
run.
Whereas the first phase is based on iterating through each device on the
system, this phase is based on iterating through each VG name. The list
of VG names comes from phase 1, which stored the list in lvmcache to be
used by phase 2.
Some commands may need to iterate through all VG names, while others may
need to iterate through just one or two.
This phase includes locking each VG as work is done on it, so that two
commands do not interfere with each other.
Phase 2 in outline:
For each VG name:
a. Lock the VG.
b. Repeat the phase 1 scan steps for each device (PV) in this VG.
The phase 1 information in lvmcache may have changed because no VG lock
was held during phase 1. So, repeat the phase 1 steps, but only for the
devices in this VG.
c. Get the list of on-disk metadata locations for this VG.
Phase 1 created this list in lvmcache to be used here. At this
point we copy it out of lvmcache. In the simple/common case,
this is a list of devices in the VG. But, some devices may
have 0 or 2 metadata locations instead of the default 1, so it
is not always equal to the list of devices. We want to read
every copy of the metadata for this VG.
d. For each metadata location on each device in the VG
(the list from the previous step):
1) Look at the mda_header. The location of the mda_header was saved
in the lvmcache info struct by phase 1 (where it came from the
pv_header.) The mda_header tells us where the text VG metadata is
located.
2) Look at the text VG metadata. The location came from mda_header
in the previous step. The VG metadata is fully analyzed and used
to create an in-memory 'struct volume_group'.
Copying or reading the mda_header and VG metadata in steps d.1 and d.2
follow the same model as in phase 1: if the data read in scan step 2.b
covered these areas, then data is simply copied out of the buffer from
step 2.b, otherwise new reads are done.
e. Compare the copies of VG metadata that were found in each location.
If some copies are older, choose the newest one to use, and update
any older copies.
f. Update details about the devices/VG in lvmcache.
g. Pass the 'vg' struct to the command-specific code to work with.
Phase 2 in code:
The most relevant functions are listed for each step in the outline.
For each VG name:
process_each_vg()
a. vg_read()
lock_vol()
b. vg_read()
lvmcache_label_rescan_vg()
[insert phase 1 steps a-f]
c. vg_read()
create_instance()
_text_create_text_instance()
_create_vg_text_instance()
lvmcache_fid_add_mdas_vg()
[Copies mda locations from info->mdas where it was saved
by phase 1, into fid->metadata_areas_in_use. This is
the key connection between phase 1 and phase 2.]
d. dm_list_iterate_items(mda, &fid->metadata_areas_in_use)
d1. ops->vg_read()
_vg_read_raw()
raw_read_mda_header()
d2. _vg_read_raw()
text_read_metadata()
config_file_read_fd()
ops->read_vg()
_read_vg()
Phase 2 in log messages:
For each VG name:
Processing VG <name>
Reading VG <name>
b. Reading VG rereading labels for <name>
Scanning data from devs async
[insert log messages from phase 1 steps a-f]
Scanned data from <N> devs async
For each mda on each <dev> in the VG:
d. Reading VG <name> from <dev>
d.1. Copying|Reading mda header sector from <dev> ...
d.2. Copying|Reading metadata from <dev> ...

View File

@@ -1,8 +1,5 @@
/* include/configure.h.in. Generated from configure.in by autoheader. */
/* Define to 1 if aio is available. */
#undef AIO_SUPPORT
/* Define to 1 to use libblkid detection of signatures when wiping. */
#undef BLKID_WIPING_SUPPORT
@@ -130,9 +127,6 @@
/* Path to dmeventd pidfile. */
#undef DMEVENTD_PIDFILE
/* Define to 1 to enable the device-mapper filemap daemon. */
#undef DMFILEMAPD
/* Define to enable compat protocol */
#undef DM_COMPAT
@@ -151,9 +145,6 @@
/* Library version */
#undef DM_LIB_VERSION
/* Path to fsadm binary. */
#undef FSADM_PATH
/* Define to 1 if you have the `alarm' function. */
#undef HAVE_ALARM
@@ -497,9 +488,6 @@
/* Define to 1 if you have the <sys/file.h> header file. */
#undef HAVE_SYS_FILE_H
/* Define to 1 if you have the <sys/inotify.h> header file. */
#undef HAVE_SYS_INOTIFY_H
/* Define to 1 if you have the <sys/ioctl.h> header file. */
#undef HAVE_SYS_IOCTL_H
@@ -635,9 +623,6 @@
/* Define to 1 to include code that uses lvmpolld. */
#undef LVMPOLLD_SUPPORT
/* configure command line used */
#undef LVM_CONFIGURE_LINE
/* Path to lvm binary. */
#undef LVM_PATH
@@ -688,6 +673,9 @@
/* Define to 1 to include the LVM readline shell. */
#undef READLINE_SUPPORT
/* Define to 1 to include built-in support for replicators. */
#undef REPLICATOR_INTERNAL
/* Define as the return type of signal handlers (`int' or `void'). */
#undef RETSIGTYPE

View File

@@ -36,6 +36,10 @@ ifeq ("@RAID@", "shared")
SUBDIRS += raid
endif
ifeq ("@REPLICATORS@", "shared")
SUBDIRS += replicator
endif
ifeq ("@THIN@", "shared")
SUBDIRS += thin
endif
@@ -44,10 +48,6 @@ ifeq ("@CACHE@", "shared")
SUBDIRS += cache_segtype
endif
ifeq ("@CLUSTER@", "shared")
SUBDIRS += locking
endif
SOURCES =\
activate/activate.c \
cache/lvmcache.c \
@@ -96,13 +96,13 @@ SOURCES =\
metadata/lv_manip.c \
metadata/merge.c \
metadata/metadata.c \
metadata/metadata-liblvm.c \
metadata/mirror.c \
metadata/pool_manip.c \
metadata/pv.c \
metadata/pv_manip.c \
metadata/pv_map.c \
metadata/raid_manip.c \
metadata/replicator_manip.c \
metadata/segtype.c \
metadata/snapshot_manip.c \
metadata/thin_manip.c \
@@ -149,6 +149,10 @@ ifeq ("@CLUSTER@", "internal")
SOURCES += locking/cluster_locking.c
endif
ifeq ("@CLUSTER@", "shared")
SUBDIRS += locking
endif
ifeq ("@SNAPSHOTS@", "internal")
SOURCES += snapshot/snapshot.c
endif
@@ -161,6 +165,10 @@ ifeq ("@RAID@", "internal")
SOURCES += raid/raid.c
endif
ifeq ("@REPLICATORS@", "internal")
SOURCES += replicator/replicator.c
endif
ifeq ("@THIN@", "internal")
SOURCES += thin/thin.c
endif
@@ -196,6 +204,11 @@ ifeq ("@BUILD_LVMLOCKD@", "yes")
locking/lvmlockd.c
endif
ifeq ("@DMEVENTD@", "yes")
CLDFLAGS += -L$(top_builddir)/daemons/dmeventd
LIBS += -ldevmapper-event
endif
LIB_NAME = liblvm-internal
LIB_STATIC = $(LIB_NAME).a
@@ -207,6 +220,7 @@ ifeq ($(MAKECMDGOALS),distclean)
mirror \
notify \
raid \
replicator \
thin \
cache_segtype \
locking
@@ -215,10 +229,10 @@ endif
CFLOW_LIST = $(SOURCES)
CFLOW_LIST_TARGET = $(LIB_NAME).cflow
PROGS_CFLAGS = $(BLKID_CFLAGS) $(UDEV_CFLAGS)
include $(top_builddir)/make.tmpl
CFLAGS += $(BLKID_CFLAGS) $(UDEV_CFLAGS) $(VALGRIND_CFLAGS)
$(SUBDIRS): $(LIB_STATIC)
CLEAN_TARGETS += misc/configure.h misc/lvm-version.h

View File

@@ -1,6 +1,6 @@
/*
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
* Copyright (C) 2004-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2004-2016 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -28,7 +28,6 @@
#include "config.h"
#include "segtype.h"
#include "sharedlib.h"
#include "lvmcache.h"
#include <limits.h>
#include <fcntl.h>
@@ -151,15 +150,15 @@ static int _lv_passes_volumes_filter(struct cmd_context *cmd, const struct logic
|| str_list_match_list(&cmd->tags,
&lv->vg->tags, NULL))
return 1;
continue;
else
continue;
}
/* If supplied tag matches LV or VG tag, activate */
if (str_list_match_item(&lv->tags, str) ||
str_list_match_item(&lv->vg->tags, str))
return 1;
continue;
else
continue;
}
/* If supplied name is vgname[/lvname] */
@@ -273,18 +272,10 @@ int lv_raid_percent(const struct logical_volume *lv, dm_percent_t *percent)
{
return 0;
}
int lv_raid_data_offset(const struct logical_volume *lv, uint64_t *data_offset)
{
return 0;
}
int lv_raid_dev_health(const struct logical_volume *lv, char **dev_health)
{
return 0;
}
int lv_raid_dev_count(const struct logical_volume *lv, uint32_t *dev_cnt)
{
return 0;
}
int lv_raid_mismatch_count(const struct logical_volume *lv, uint64_t *cnt)
{
return 0;
@@ -782,8 +773,7 @@ int lv_info_with_seg_status(struct cmd_context *cmd,
if (lv_is_used_cache_pool(lv)) {
/* INFO is not set as cache-pool cannot be active.
* STATUS is collected from cache LV */
if (!(lv_seg = get_only_segment_using_this_lv(lv)))
return_0;
lv_seg = get_only_segment_using_this_lv(lv);
(void) _lv_info(cmd, lv_seg->lv, 1, NULL, lv_seg, &status->seg_status, 0, 0);
return 1;
}
@@ -798,18 +788,14 @@ int lv_info_with_seg_status(struct cmd_context *cmd,
status->info.exists = 0; /* So pool LV is not active */
}
return 1;
}
if (lv_is_external_origin(lv)) {
} else if (lv_is_external_origin(lv)) {
if (!_lv_info(cmd, lv, 0, &status->info, NULL, NULL,
with_open_count, with_read_ahead))
return_0;
(void) _lv_info(cmd, lv, 1, NULL, lv_seg, &status->seg_status, 0, 0);
return 1;
}
if (lv_is_origin(lv)) {
} else if (lv_is_origin(lv)) {
/* Query segment status for 'layered' (-real) device most of the time,
* only for merging snapshot, query its progress.
* TODO: single LV may need couple status to be exposed at once....
@@ -826,9 +812,7 @@ int lv_info_with_seg_status(struct cmd_context *cmd,
/* Grab STATUS from layered -real */
(void) _lv_info(cmd, lv, 1, NULL, lv_seg, &status->seg_status, 0, 0);
return 1;
}
if (lv_is_cow(lv)) {
} else if (lv_is_cow(lv)) {
if (lv_is_merging_cow(lv)) {
olv = origin_from_cow(lv);
@@ -843,6 +827,7 @@ int lv_info_with_seg_status(struct cmd_context *cmd,
* When merge is in progress, query merging origin LV instead.
* COW volume is already mapped as error target in this case.
*/
status->lv = olv;
return 1;
}
@@ -999,30 +984,6 @@ int lv_raid_percent(const struct logical_volume *lv, dm_percent_t *percent)
return lv_mirror_percent(lv->vg->cmd, lv, 0, percent, NULL);
}
int lv_raid_data_offset(const struct logical_volume *lv, uint64_t *data_offset)
{
int r;
struct dev_manager *dm;
struct dm_status_raid *status;
if (!lv_info(lv->vg->cmd, lv, 0, NULL, 0, 0))
return 0;
log_debug_activation("Checking raid data offset and dev sectors for LV %s/%s",
lv->vg->name, lv->name);
if (!(dm = dev_manager_create(lv->vg->cmd, lv->vg->name, 1)))
return_0;
if (!(r = dev_manager_raid_status(dm, lv, &status)))
stack;
*data_offset = status->data_offset;
dev_manager_destroy(dm);
return r;
}
int lv_raid_dev_health(const struct logical_volume *lv, char **dev_health)
{
int r;
@@ -1052,32 +1013,6 @@ int lv_raid_dev_health(const struct logical_volume *lv, char **dev_health)
return r;
}
int lv_raid_dev_count(const struct logical_volume *lv, uint32_t *dev_cnt)
{
struct dev_manager *dm;
struct dm_status_raid *status;
*dev_cnt = 0;
if (!lv_info(lv->vg->cmd, lv, 0, NULL, 0, 0))
return 0;
log_debug_activation("Checking raid device count for LV %s/%s",
lv->vg->name, lv->name);
if (!(dm = dev_manager_create(lv->vg->cmd, lv->vg->name, 1)))
return_0;
if (!dev_manager_raid_status(dm, lv, &status)) {
dev_manager_destroy(dm);
return_0;
}
*dev_cnt = status->dev_count;
dev_manager_destroy(dm);
return 1;
}
int lv_raid_mismatch_count(const struct logical_volume *lv, uint64_t *cnt)
{
struct dev_manager *dm;
@@ -1708,7 +1643,7 @@ static char *_build_target_uuid(struct cmd_context *cmd, const struct logical_vo
if (lv_is_thin_pool(lv))
layer = "tpool"; /* Monitor "tpool" for the "thin pool". */
else if (lv_is_origin(lv) || lv_is_external_origin(lv))
else if (lv_is_origin(lv))
layer = "real"; /* Monitor "real" for "snapshot-origin". */
else
layer = NULL;
@@ -1856,15 +1791,12 @@ int monitor_dev_for_events(struct cmd_context *cmd, const struct logical_volume
* However in case command would have crashed, such LV is
* left unmonitored and may potentially require dmeventd.
*/
if (lv_is_cache_pool_data(lv) || lv_is_cache_pool_metadata(lv)) {
if (!(seg = find_pool_seg(first_seg(lv))))
return_0;
if (!lv_is_used_cache_pool(seg->lv)) {
log_debug_activation("Skipping %smonitor of %s.%s",
(monitor) ? "" : "un", display_lvname(lv),
(monitor) ? " Cache pool activation for clearing only." : "");
return 1;
}
if ((lv_is_cache_pool_data(lv) || lv_is_cache_pool_metadata(lv)) &&
!lv_is_used_cache_pool((find_pool_seg(first_seg(lv))->lv))) {
log_debug_activation("Skipping %smonitor of %s.%s",
(monitor) ? "" : "un", display_lvname(lv),
(monitor) ? " Cache pool activation for clearing only." : "");
return 1;
}
/*
@@ -1950,13 +1882,6 @@ int monitor_dev_for_events(struct cmd_context *cmd, const struct logical_volume
r = 0;
}
if (seg->external_lv &&
!monitor_dev_for_events(cmd, seg->external_lv,
(!monitor) ? laopts : NULL, monitor)) {
stack;
r = 0;
}
if (seg->metadata_lv &&
!monitor_dev_for_events(cmd, seg->metadata_lv, NULL, monitor)) {
stack;
@@ -2023,13 +1948,16 @@ int monitor_dev_for_events(struct cmd_context *cmd, const struct logical_volume
/* Check [un]monitor results */
/* Try a couple times if pending, but not forever... */
for (i = 0;; i++) {
for (i = 0; i < 40; i++) {
pending = 0;
monitored = seg->segtype->ops->target_monitored(seg, &pending);
if (!pending || i >= 40)
if (pending ||
(!monitored && monitor) ||
(monitored && !monitor))
log_very_verbose("%s %smonitoring still pending: waiting...",
display_lvname(lv), monitor ? "" : "un");
else
break;
log_very_verbose("%s %smonitoring still pending: waiting...",
display_lvname(lv), monitor ? "" : "un");
usleep(10000 * i);
}
@@ -2124,17 +2052,6 @@ static int _lv_suspend(struct cmd_context *cmd, const char *lvid_s,
if (!lv_info(cmd, lv, laopts->origin_only, &info, 0, 0))
goto_out;
/*
* Save old and new (current and precommitted) versions of the
* VG metadata for lv_resume() to use, since lv_resume can't
* read metadata given that devices are suspended. lv_resume()
* will resume LVs using the old/current metadata if the vg_commit
* did happen (or failed), and it will resume LVs using the
* new/precommitted metadata if the vg_commit succeeded.
*/
lvmcache_save_suspended_vg(lv->vg, 0);
lvmcache_save_suspended_vg(lv_pre->vg, 1);
if (!info.exists || info.suspended) {
if (!error_if_not_suspended) {
r = 1;
@@ -2144,6 +2061,9 @@ static int _lv_suspend(struct cmd_context *cmd, const char *lvid_s,
goto out;
}
if (!lv_read_replicator_vgs(lv))
goto_out;
lv_calculate_readahead(lv, NULL);
/*
@@ -2265,8 +2185,10 @@ static int _lv_suspend(struct cmd_context *cmd, const char *lvid_s,
out:
if (lv_pre_to_free)
release_vg(lv_pre_to_free->vg);
if (lv_to_free)
if (lv_to_free) {
lv_release_replicator_vgs(lv_to_free);
release_vg(lv_to_free->vg);
}
return r;
}
@@ -2293,54 +2215,15 @@ static int _lv_resume(struct cmd_context *cmd, const char *lvid_s,
struct lv_activate_opts *laopts, int error_if_not_active,
const struct logical_volume *lv)
{
struct volume_group *vg = NULL;
struct logical_volume *lv_found = NULL;
const union lvid *lvid;
const char *vgid;
const struct logical_volume *lv_to_free = NULL;
struct lvinfo info;
int r = 0;
if (!activation())
return 1;
/*
* When called in clvmd, lvid_s is set and lv is not. We need to
* get the VG metadata without reading disks because devs are
* suspended. lv_suspend() saved old and new VG metadata for us
* to use here. If vg_commit() happened, lvmcache_get_suspended_vg
* will return the new metadata for us to use in resuming LVs.
* If vg_commit() did not happen, lvmcache_get_suspended_vg
* returns the old metadata which we use to resume LVs.
*/
if (!lv && lvid_s) {
lvid = (const union lvid *) lvid_s;
vgid = (const char *)lvid->id[0].uuid;
if ((vg = lvmcache_get_suspended_vg(vgid))) {
log_debug_activation("Resuming LVID %s found saved vg seqno %d %s", lvid_s, vg->seqno, vg->name);
if ((lv_found = find_lv_in_vg_by_lvid(vg, lvid))) {
log_debug_activation("Resuming LVID %s found saved LV %s", lvid_s, display_lvname(lv_found));
lv = lv_found;
} else
log_debug_activation("Resuming LVID %s did not find saved LV", lvid_s);
} else
log_debug_activation("Resuming LVID %s did not find saved VG", lvid_s);
/*
* resume must have been called without a preceding suspend,
* so we need to read the vg.
*/
if (!lv) {
log_debug_activation("Resuming LVID %s reading VG", lvid_s);
if (!(lv_found = lv_from_lvid(cmd, lvid_s, 0))) {
log_debug_activation("Resuming LVID %s failed to read VG", lvid_s);
goto out;
}
lv = lv_found;
}
}
if (!lv && !(lv_to_free = lv = lv_from_lvid(cmd, lvid_s, 0)))
goto_out;
if (!lv_is_origin(lv) && !lv_is_thin_volume(lv) && !lv_is_thin_pool(lv))
laopts->origin_only = 0;
@@ -2385,6 +2268,9 @@ static int _lv_resume(struct cmd_context *cmd, const char *lvid_s,
r = 1;
out:
if (lv_to_free)
release_vg(lv_to_free->vg);
return r;
}
@@ -2486,6 +2372,9 @@ int lv_deactivate(struct cmd_context *cmd, const char *lvid_s, const struct logi
goto_out;
}
if (!lv_read_replicator_vgs(lv))
goto_out;
if (!monitor_dev_for_events(cmd, lv, &laopts, 0))
stack;
@@ -2500,8 +2389,10 @@ int lv_deactivate(struct cmd_context *cmd, const char *lvid_s, const struct logi
r = 0;
}
out:
if (lv_to_free)
if (lv_to_free) {
lv_release_replicator_vgs(lv_to_free);
release_vg(lv_to_free->vg);
}
return r;
}
@@ -2511,10 +2402,6 @@ int lv_activation_filter(struct cmd_context *cmd, const char *lvid_s,
int *activate_lv, const struct logical_volume *lv)
{
const struct logical_volume *lv_to_free = NULL;
struct volume_group *vg = NULL;
struct logical_volume *lv_found = NULL;
const union lvid *lvid;
const char *vgid;
int r = 0;
if (!activation()) {
@@ -2522,24 +2409,6 @@ int lv_activation_filter(struct cmd_context *cmd, const char *lvid_s,
return 1;
}
/*
* This function is called while devices are suspended,
* so try to use the copy of the vg that was saved in
* lv_suspend.
*/
if (!lv && lvid_s) {
lvid = (const union lvid *) lvid_s;
vgid = (const char *)lvid->id[0].uuid;
if ((vg = lvmcache_get_suspended_vg(vgid))) {
log_debug_activation("activation_filter for %s found saved VG seqno %d %s", lvid_s, vg->seqno, vg->name);
if ((lv_found = find_lv_in_vg_by_lvid(vg, lvid))) {
log_debug_activation("activation_filter for %s found saved LV %s", lvid_s, display_lvname(lv_found));
lv = lv_found;
}
}
}
if (!lv && !(lv_to_free = lv = lv_from_lvid(cmd, lvid_s, 0)))
goto_out;
@@ -2638,6 +2507,9 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
goto out;
}
if (!lv_read_replicator_vgs(lv))
goto_out;
lv_calculate_readahead(lv, NULL);
critical_section_inc(cmd, "activating");
@@ -2649,8 +2521,10 @@ static int _lv_activate(struct cmd_context *cmd, const char *lvid_s,
stack;
out:
if (lv_to_free)
if (lv_to_free) {
lv_release_replicator_vgs(lv_to_free);
release_vg(lv_to_free->vg);
}
return r;
}

View File

@@ -168,13 +168,11 @@ int lv_snapshot_percent(const struct logical_volume *lv, dm_percent_t *percent);
int lv_mirror_percent(struct cmd_context *cmd, const struct logical_volume *lv,
int wait, dm_percent_t *percent, uint32_t *event_nr);
int lv_raid_percent(const struct logical_volume *lv, dm_percent_t *percent);
int lv_raid_dev_count(const struct logical_volume *lv, uint32_t *dev_cnt);
int lv_raid_data_offset(const struct logical_volume *lv, uint64_t *data_offset);
int lv_raid_dev_health(const struct logical_volume *lv, char **dev_health);
int lv_raid_mismatch_count(const struct logical_volume *lv, uint64_t *cnt);
int lv_raid_sync_action(const struct logical_volume *lv, char **sync_action);
int lv_raid_message(const struct logical_volume *lv, const char *msg);
int lv_cache_status(const struct logical_volume *cache_lv,
int lv_cache_status(const struct logical_volume *lv,
struct lv_status_cache **status);
int lv_thin_pool_percent(const struct logical_volume *lv, int metadata,
dm_percent_t *percent);
@@ -202,12 +200,12 @@ int lv_has_target_type(struct dm_pool *mem, const struct logical_volume *lv,
const char *layer, const char *target_type);
int monitor_dev_for_events(struct cmd_context *cmd, const struct logical_volume *lv,
const struct lv_activate_opts *laopts, int monitor);
const struct lv_activate_opts *laopts, int do_reg);
#ifdef DMEVENTD
# include "libdevmapper-event.h"
char *get_monitor_dso_path(struct cmd_context *cmd, const char *libpath);
int target_registered_with_dmeventd(struct cmd_context *cmd, const char *dso,
int target_registered_with_dmeventd(struct cmd_context *cmd, const char *libpath,
const struct logical_volume *lv, int *pending);
int target_register_events(struct cmd_context *cmd, const char *dso, const struct logical_volume *lv,
int evmask __attribute__((unused)), int set, int timeout);

View File

@@ -1,6 +1,6 @@
/*
* Copyright (C) 2002-2004 Sistina Software, Inc. All rights reserved.
* Copyright (C) 2004-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2004-2016 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -214,14 +214,6 @@ typedef enum {
STATUS, /* DM_DEVICE_STATUS ioctl */
} info_type_t;
/* Return length of segment depending on type and reshape_len */
static uint32_t _seg_len(const struct lv_segment *seg)
{
uint32_t reshape_len = seg_is_raid(seg) ? ((seg->area_count - seg->segtype->parity_devs) * seg->reshape_len) : 0;
return seg->len - reshape_len;
}
static int _info_run(const char *dlid, struct dm_info *dminfo,
uint32_t *read_ahead,
struct lv_seg_status *seg_status,
@@ -258,12 +250,7 @@ static int _info_run(const char *dlid, struct dm_info *dminfo,
if (seg_status && dminfo->exists) {
start = length = seg_status->seg->lv->vg->extent_size;
start *= seg_status->seg->le;
length *= _seg_len(seg_status->seg);
/* Uses max DM_THIN_MAX_METADATA_SIZE sectors for metadata device */
if (lv_is_thin_pool_metadata(seg_status->seg->lv) &&
(length > DM_THIN_MAX_METADATA_SIZE))
length = DM_THIN_MAX_METADATA_SIZE;
length *= seg_status->seg->len;
do {
target = dm_get_next_target(dmt, target, &target_start,
@@ -275,8 +262,7 @@ static int _info_run(const char *dlid, struct dm_info *dminfo,
target_params = NULL; /* Marking this target_params unusable */
} while (target);
if (!target_name ||
!_get_segment_status_from_target_params(target_name, target_params, seg_status))
if (!_get_segment_status_from_target_params(target_name, target_params, seg_status))
stack;
}
@@ -1038,8 +1024,7 @@ static int _percent_run(struct dev_manager *dm, const char *name,
goto_out;
}
log_debug_activation("LV percent: %s",
display_percent(dm->cmd, *overall_percent));
log_debug_activation("LV percent: %.2f", dm_percent_to_float(*overall_percent));
r = 1;
out:
@@ -1056,11 +1041,10 @@ static int _percent(struct dev_manager *dm, const char *name, const char *dlid,
if (_percent_run(dm, NULL, dlid, target_type, wait, lv, percent,
event_nr, fail_if_percent_unsupported))
return 1;
if (_original_uuid_format_check_required(dm->cmd) &&
_percent_run(dm, NULL, dlid + sizeof(UUID_PREFIX) - 1,
target_type, wait, lv, percent,
event_nr, fail_if_percent_unsupported))
else if (_original_uuid_format_check_required(dm->cmd) &&
_percent_run(dm, NULL, dlid + sizeof(UUID_PREFIX) - 1,
target_type, wait, lv, percent,
event_nr, fail_if_percent_unsupported))
return 1;
}
@@ -1324,13 +1308,14 @@ int dev_manager_raid_message(struct dev_manager *dm,
return 0;
}
/* These are the supported RAID messages for dm-raid v1.9.0 */
/* These are the supported RAID messages for dm-raid v1.5.0 */
if (strcmp(msg, "idle") &&
strcmp(msg, "frozen") &&
strcmp(msg, "resync") &&
strcmp(msg, "recover") &&
strcmp(msg, "check") &&
strcmp(msg, "repair")) {
strcmp(msg, "repair") &&
strcmp(msg, "reshape")) {
log_error(INTERNAL_ERROR "Unknown RAID message: %s.", msg);
return 0;
}
@@ -1774,6 +1759,84 @@ static int _add_dev_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
return 1;
}
/*
* Add replicator devices
*
* Using _add_dev_to_dtree() directly instead of _add_lv_to_dtree()
* to avoid extra checks with extensions.
*/
static int _add_partial_replicator_to_dtree(struct dev_manager *dm,
struct dm_tree *dtree,
const struct logical_volume *lv)
{
struct logical_volume *rlv = first_seg(lv)->replicator;
struct replicator_device *rdev;
struct replicator_site *rsite;
struct dm_tree_node *rep_node, *rdev_node;
const char *uuid;
if (!lv_is_active_replicator_dev(lv)) {
if (!_add_dev_to_dtree(dm, dtree, lv->rdevice->lv,
NULL))
return_0;
return 1;
}
/* Add _rlog and replicator device */
if (!_add_dev_to_dtree(dm, dtree, first_seg(rlv)->rlog_lv, NULL))
return_0;
if (!_add_dev_to_dtree(dm, dtree, rlv, NULL))
return_0;
if (!(uuid = build_dm_uuid(dm->mem, rlv, NULL)))
return_0;
rep_node = dm_tree_find_node_by_uuid(dtree, uuid);
/* Add all related devices for replicator */
dm_list_iterate_items(rsite, &rlv->rsites)
dm_list_iterate_items(rdev, &rsite->rdevices) {
if (rsite->state == REPLICATOR_STATE_ACTIVE) {
/* Add _rimage LV */
if (!_add_dev_to_dtree(dm, dtree, rdev->lv, NULL))
return_0;
/* Add replicator-dev LV, except of the already added one */
if ((lv != rdev->replicator_dev->lv) &&
!_add_dev_to_dtree(dm, dtree,
rdev->replicator_dev->lv, NULL))
return_0;
/* If replicator exists - try connect existing heads */
if (rep_node) {
uuid = build_dm_uuid(dm->mem,
rdev->replicator_dev->lv,
NULL);
if (!uuid)
return_0;
rdev_node = dm_tree_find_node_by_uuid(dtree, uuid);
if (rdev_node)
dm_tree_node_set_presuspend_node(rdev_node,
rep_node);
}
}
if (!rdev->rsite->vg_name)
continue;
if (!_add_dev_to_dtree(dm, dtree, rdev->lv, NULL))
return_0;
if (rdev->slog &&
!_add_dev_to_dtree(dm, dtree, rdev->slog, NULL))
return_0;
}
return 1;
}
struct pool_cb_data {
struct dev_manager *dm;
const struct logical_volume *pool_lv;
@@ -1787,7 +1850,7 @@ struct pool_cb_data {
static int _pool_callback(struct dm_tree_node *node,
dm_node_callback_t type, void *cb_data)
{
int ret, status = 0, fd;
int ret, status, fd;
const struct dm_config_node *cn;
const struct dm_config_value *cv;
const struct pool_cb_data *data = cb_data;
@@ -2078,6 +2141,11 @@ static int _add_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
}
}
/* Adding LV head of replicator adds all other related devs */
if (lv_is_replicator_dev(lv) &&
!_add_partial_replicator_to_dtree(dm, dtree, lv))
return_0;
/* Add any LVs used by segments in this LV */
dm_list_iterate_items(seg, &lv->segments) {
if (seg->external_lv && dm->track_external_lv_deps &&
@@ -2146,7 +2214,7 @@ static char *_add_error_or_zero_device(struct dev_manager *dm, struct dm_tree *d
struct lv_segment *seg_i;
struct dm_info info;
int segno = -1, i = 0;
uint64_t size = (uint64_t) _seg_len(seg) * seg->lv->vg->extent_size;
uint64_t size = (uint64_t) seg->len * seg->lv->vg->extent_size;
dm_list_iterate_items(seg_i, &seg->lv->segments) {
if (seg == seg_i) {
@@ -2432,7 +2500,7 @@ static int _add_target_to_dtree(struct dev_manager *dm,
return seg->segtype->ops->add_target_line(dm, dm->mem, dm->cmd,
&dm->target_state, seg,
laopts, dnode,
extent_size * _seg_len(seg),
extent_size * seg->len,
&dm->pvmove_mirror_count);
}
@@ -2441,6 +2509,64 @@ static int _add_new_lv_to_dtree(struct dev_manager *dm, struct dm_tree *dtree,
struct lv_activate_opts *laopts,
const char *layer);
/* Add all replicators' LVs */
static int _add_replicator_dev_target_to_dtree(struct dev_manager *dm,
struct dm_tree *dtree,
struct lv_segment *seg,
struct lv_activate_opts *laopts)
{
struct replicator_device *rdev;
struct replicator_site *rsite;
/* For inactive replicator add linear mapping */
if (!lv_is_active_replicator_dev(seg->lv)) {
if (!_add_new_lv_to_dtree(dm, dtree, seg->lv->rdevice->lv, laopts, NULL))
return_0;
return 1;
}
/* Add rlog and replicator nodes */
if (!seg->replicator ||
!first_seg(seg->replicator)->rlog_lv ||
!_add_new_lv_to_dtree(dm, dtree,
first_seg(seg->replicator)->rlog_lv,
laopts, NULL) ||
!_add_new_lv_to_dtree(dm, dtree, seg->replicator, laopts, NULL))
return_0;
/* Activation of one replicator_dev node activates all other nodes */
dm_list_iterate_items(rsite, &seg->replicator->rsites) {
dm_list_iterate_items(rdev, &rsite->rdevices) {
if (rdev->lv &&
!_add_new_lv_to_dtree(dm, dtree, rdev->lv,
laopts, NULL))
return_0;
if (rdev->slog &&
!_add_new_lv_to_dtree(dm, dtree, rdev->slog,
laopts, NULL))
return_0;
}
}
/* Add remaining replicator-dev nodes in the second loop
* to avoid multiple retries for inserting all elements */
dm_list_iterate_items(rsite, &seg->replicator->rsites) {
if (rsite->state != REPLICATOR_STATE_ACTIVE)
continue;
dm_list_iterate_items(rdev, &rsite->rdevices) {
if (rdev->replicator_dev->lv == seg->lv)
continue;
if (!rdev->replicator_dev->lv ||
!_add_new_lv_to_dtree(dm, dtree,
rdev->replicator_dev->lv,
laopts, NULL))
return_0;
}
}
return 1;
}
static int _add_new_external_lv_to_dtree(struct dev_manager *dm,
struct dm_tree *dtree,
struct logical_volume *external_lv,
@@ -2541,6 +2667,11 @@ static int _add_segment_to_dtree(struct dev_manager *dm,
lv_layer(seg->pool_lv)))
return_0;
if (seg_is_replicator_dev(seg)) {
if (!_add_replicator_dev_target_to_dtree(dm, dtree, seg, laopts))
return_0;
}
/* Add any LVs used by this segment */
for (s = 0; s < seg->area_count; ++s) {
if ((seg_type(seg, s) == AREA_LV) &&
@@ -2562,7 +2693,7 @@ static int _add_segment_to_dtree(struct dev_manager *dm,
/* Replace target and all its used devs with error mapping */
log_debug_activation("Using error for pending delete %s.",
display_lvname(seg->lv));
if (!dm_tree_node_add_error_target(dnode, (uint64_t)seg->lv->vg->extent_size * _seg_len(seg)))
if (!dm_tree_node_add_error_target(dnode, (uint64_t)seg->lv->vg->extent_size * seg->len))
return_0;
} else if (!_add_target_to_dtree(dm, dnode, seg, laopts))
return_0;
@@ -3034,6 +3165,7 @@ static int _tree_action(struct dev_manager *dm, const struct logical_volume *lv,
log_error(INTERNAL_ERROR "_tree_action: Action %u not supported.", action);
goto out;
}
r = 1;
out:

View File

@@ -186,11 +186,11 @@ static int _mk_link(const char *dev_dir, const char *vg_name,
!stat(lv_path, &buf)) {
if (buf_lp.st_rdev == buf.st_rdev)
return 1;
log_warn("Symlink %s that should have been "
"created by udev does not have "
"correct target. Falling back to "
"direct link creation", lv_path);
else
log_warn("Symlink %s that should have been "
"created by udev does not have "
"correct target. Falling back to "
"direct link creation", lv_path);
} else
log_warn("Symlink %s that should have been "
"created by udev could not be checked "
@@ -239,9 +239,7 @@ static int _rm_link(const char *dev_dir, const char *vg_name,
return 1;
log_sys_error("lstat", lv_path);
return 0;
}
if (dm_udev_get_sync_support() && udev_checking() && check_udev)
} else if (dm_udev_get_sync_support() && udev_checking() && check_udev)
log_warn("The link %s should have been removed by udev "
"but it is still present. Falling back to "
"direct link removal.", lv_path);
@@ -480,9 +478,9 @@ int fs_rename_lv(const struct logical_volume *lv, const char *dev,
_fs_op(FS_ADD, lv->vg->cmd->dev_dir, lv->vg->name,
lv->name, dev, "", lv->vg->cmd->current_settings.udev_rules));
}
return _fs_op(FS_RENAME, lv->vg->cmd->dev_dir, lv->vg->name, lv->name,
dev, old_lvname, lv->vg->cmd->current_settings.udev_rules);
else
return _fs_op(FS_RENAME, lv->vg->cmd->dev_dir, lv->vg->name, lv->name,
dev, old_lvname, lv->vg->cmd->current_settings.udev_rules);
}
void fs_unlock(void)

709
lib/cache/lvmcache.c vendored
View File

@@ -63,42 +63,15 @@ struct lvmcache_vginfo {
char *lock_type;
uint32_t mda_checksum;
size_t mda_size;
int independent_metadata_location; /* metadata read from independent areas */
/*
* The following are not related to lvmcache or vginfo,
* but are borrowing the vginfo to store the data.
*
* suspended_vg_* are used only by clvmd suspend/resume.
* In suspend, both old (current) and new (precommitted)
* metadata is saved. (Each in three forms: buffer, cft,
* and vg). In resume, if the vg was committed
* (suspended_vg_committed is set), then LVs are resumed
* using the new metadata, but if the vg wasn't committed,
* then LVs are resumed using the old metadata.
*
* suspended_vg_committed is set to 1 when clvmd gets
* LCK_VG_COMMIT from vg_commit().
*
* These fields are only used between suspend and resume
* in clvmd, and should never be used in any other way.
* The contents of this data are never changed. This
* data does not really belong in lvmcache, it's unrelated
* to lvmcache or vginfo, but it's just a convenient place
* for clvmd to stash the VG between suspend and resume
* (since the same caller isn't present to pass the VG to
* both suspend and resume in the case of clvmd.)
*
* This data is not really a "cache" of the VG, it is just
* a location to pass the VG between suspend and resume.
*/
int suspended_vg_committed;
char *suspended_vg_old_buf;
struct dm_config_tree *suspended_vg_old_cft;
struct volume_group *suspended_vg_old;
char *suspended_vg_new_buf;
struct dm_config_tree *suspended_vg_new_cft;
struct volume_group *suspended_vg_new;
size_t vgmetadata_size;
char *vgmetadata; /* Copy of VG metadata as format_text string */
struct dm_config_tree *cft; /* Config tree created from vgmetadata */
/* Lifetime is directly tied to vgmetadata */
struct volume_group *cached_vg;
unsigned holders;
unsigned vg_use_count; /* Counter of vg reusage */
unsigned precommitted; /* Is vgmetadata live or precommitted? */
unsigned cached_vg_invalidated; /* Signal to regenerate cached_vg */
};
static struct dm_hash_table *_pvid_hash = NULL;
@@ -165,7 +138,73 @@ void lvmcache_seed_infos_from_lvmetad(struct cmd_context *cmd)
_has_scanned = 1;
}
static void _update_cache_info_lock_state(struct lvmcache_info *info, int locked)
/* 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;
/* Release also cached config tree */
if (vginfo->cft) {
dm_config_destroy(vginfo->cft);
vginfo->cft = NULL;
}
log_debug_cache("Metadata cache: VG %s wiped.", vginfo->vgname);
release_vg(vginfo->cached_vg);
}
/*
* Cache VG metadata against the vginfo with matching vgid.
*/
static void _store_metadata(struct volume_group *vg, unsigned precommitted)
{
char uuid[64] __attribute__((aligned(8)));
struct lvmcache_vginfo *vginfo;
char *data;
size_t size;
if (!(vginfo = lvmcache_vginfo_from_vgid((const char *)&vg->id))) {
stack;
return;
}
if (!(size = export_vg_to_buffer(vg, &data))) {
stack;
_free_cached_vgmetadata(vginfo);
return;
}
/* Avoid reparsing of the same data string */
if (vginfo->vgmetadata && vginfo->vgmetadata_size == size &&
strcmp(vginfo->vgmetadata, data) == 0)
dm_free(data);
else {
_free_cached_vgmetadata(vginfo);
vginfo->vgmetadata_size = size;
vginfo->vgmetadata = data;
}
vginfo->precommitted = precommitted;
if (!id_write_format((const struct id *)vginfo->vgid, uuid, sizeof(uuid))) {
stack;
return;
}
log_debug_cache("Metadata cache: VG %s (%s) stored (%" PRIsize_t " bytes%s).",
vginfo->vgname, uuid, 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;
@@ -173,8 +212,10 @@ static void _update_cache_info_lock_state(struct lvmcache_info *info, int locked
* Cache becomes invalid whenever lock state changes unless
* exclusive VG_GLOBAL is held (i.e. while scanning).
*/
if (!lvmcache_vgname_is_locked(VG_GLOBAL) && (was_locked != locked))
if (!lvmcache_vgname_is_locked(VG_GLOBAL) && (was_locked != locked)) {
info->status |= CACHE_INVALID;
*cached_vgmetadata_valid = 0;
}
if (locked)
info->status |= CACHE_LOCKED;
@@ -186,9 +227,14 @@ static void _update_cache_vginfo_lock_state(struct lvmcache_vginfo *vginfo,
int locked)
{
struct lvmcache_info *info;
int cached_vgmetadata_valid = 1;
dm_list_iterate_items(info, &vginfo->infos)
_update_cache_info_lock_state(info, locked);
_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)
@@ -201,35 +247,6 @@ static void _update_cache_lock_state(const char *vgname, int locked)
_update_cache_vginfo_lock_state(vginfo, locked);
}
static void _suspended_vg_free(struct lvmcache_vginfo *vginfo, int free_old, int free_new)
{
if (free_old) {
if (vginfo->suspended_vg_old_buf)
dm_free(vginfo->suspended_vg_old_buf);
if (vginfo->suspended_vg_old_cft)
dm_config_destroy(vginfo->suspended_vg_old_cft);
if (vginfo->suspended_vg_old)
release_vg(vginfo->suspended_vg_old);
vginfo->suspended_vg_old_buf = NULL;
vginfo->suspended_vg_old_cft = NULL;
vginfo->suspended_vg_old = NULL;
}
if (free_new) {
if (vginfo->suspended_vg_new_buf)
dm_free(vginfo->suspended_vg_new_buf);
if (vginfo->suspended_vg_new_cft)
dm_config_destroy(vginfo->suspended_vg_new_cft);
if (vginfo->suspended_vg_new)
release_vg(vginfo->suspended_vg_new);
vginfo->suspended_vg_new_buf = NULL;
vginfo->suspended_vg_new_cft = NULL;
vginfo->suspended_vg_new = NULL;
}
}
static void _drop_metadata(const char *vgname, int drop_precommitted)
{
struct lvmcache_vginfo *vginfo;
@@ -238,98 +255,25 @@ static void _drop_metadata(const char *vgname, int drop_precommitted)
if (!(vginfo = lvmcache_vginfo_from_vgname(vgname, NULL)))
return;
if (drop_precommitted)
/*
* 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 (!drop_precommitted && vginfo->precommitted && !vginfo->vgmetadata)
log_error(INTERNAL_ERROR "metadata commit (or revert) missing before "
"dropping metadata from cache.");
if (drop_precommitted || !vginfo->precommitted)
dm_list_iterate_items(info, &vginfo->infos)
info->status |= CACHE_INVALID;
_free_cached_vgmetadata(vginfo);
/* VG revert */
if (drop_precommitted)
_suspended_vg_free(vginfo, 0, 1);
else
_suspended_vg_free(vginfo, 1, 1);
}
void lvmcache_save_suspended_vg(struct volume_group *vg, int precommitted)
{
struct lvmcache_vginfo *vginfo;
struct format_instance *fid;
struct format_instance_ctx fic;
struct volume_group *susp_vg = NULL;
struct dm_config_tree *susp_cft = NULL;
char *susp_buf = NULL;
size_t size;
int new = precommitted;
int old = !precommitted;
if (!(vginfo = lvmcache_vginfo_from_vgid((const char *)&vg->id)))
goto_bad;
/* already saved */
if (old && vginfo->suspended_vg_old &&
(vginfo->suspended_vg_old->seqno == vg->seqno))
return;
/* already saved */
if (new && vginfo->suspended_vg_new &&
(vginfo->suspended_vg_new->seqno == vg->seqno))
return;
_suspended_vg_free(vginfo, old, new);
if (!(size = export_vg_to_buffer(vg, &susp_buf)))
goto_bad;
fic.type = FMT_INSTANCE_MDAS | FMT_INSTANCE_AUX_MDAS;
fic.context.vg_ref.vg_name = vginfo->vgname;
fic.context.vg_ref.vg_id = vginfo->vgid;
if (!(fid = vginfo->fmt->ops->create_instance(vginfo->fmt, &fic)))
goto_bad;
if (!(susp_cft = config_tree_from_string_without_dup_node_check(susp_buf)))
goto_bad;
if (!(susp_vg = import_vg_from_config_tree(susp_cft, fid)))
goto_bad;
if (old) {
vginfo->suspended_vg_old_buf = susp_buf;
vginfo->suspended_vg_old_cft = susp_cft;
vginfo->suspended_vg_old = susp_vg;
log_debug_cache("lvmcache saved suspended vg old seqno %d %s", vg->seqno, vg->name);
} else {
vginfo->suspended_vg_new_buf = susp_buf;
vginfo->suspended_vg_new_cft = susp_cft;
vginfo->suspended_vg_new = susp_vg;
log_debug_cache("lvmcache saved suspended vg new seqno %d %s", vg->seqno, vg->name);
}
return;
bad:
_suspended_vg_free(vginfo, old, new);
log_debug_cache("lvmcache failed to save suspended pre %d vg %s", precommitted, vg->name);
}
struct volume_group *lvmcache_get_suspended_vg(const char *vgid)
{
struct lvmcache_vginfo *vginfo;
if (!(vginfo = lvmcache_vginfo_from_vgid(vgid)))
return_NULL;
if (vginfo->suspended_vg_committed)
return vginfo->suspended_vg_new;
else
return vginfo->suspended_vg_old;
}
void lvmcache_drop_suspended_vg(struct volume_group *vg)
{
struct lvmcache_vginfo *vginfo;
if (!(vginfo = lvmcache_vginfo_from_vgid((const char *)&vg->id)))
return;
_suspended_vg_free(vginfo, 1, 1);
vginfo->precommitted = 0;
}
/*
@@ -344,7 +288,11 @@ void lvmcache_commit_metadata(const char *vgname)
if (!(vginfo = lvmcache_vginfo_from_vgname(vgname, NULL)))
return;
vginfo->suspended_vg_committed = 1;
if (vginfo->precommitted) {
log_debug_cache("Precommitted metadata cache: VG %s upgraded to committed.",
vginfo->vgname);
vginfo->precommitted = 0;
}
}
void lvmcache_drop_metadata(const char *vgname, int drop_precommitted)
@@ -594,6 +542,7 @@ const struct format_type *lvmcache_fmt_from_vgname(struct cmd_context *cmd,
{
struct lvmcache_vginfo *vginfo;
struct lvmcache_info *info;
struct label *label;
struct dm_list *devh, *tmp;
struct dm_list devs;
struct device_list *devl;
@@ -638,7 +587,7 @@ const struct format_type *lvmcache_fmt_from_vgname(struct cmd_context *cmd,
dm_list_iterate_safe(devh, tmp, &devs) {
devl = dm_list_item(devh, struct device_list);
label_read(devl->dev, NULL, UINT64_C(0));
(void) label_read(devl->dev, &label, UINT64_C(0));
dm_list_del(&devl->list);
dm_free(devl);
}
@@ -726,6 +675,18 @@ static int _info_is_valid(struct lvmcache_info *info)
return 1;
}
static int _vginfo_is_valid(struct lvmcache_vginfo *vginfo)
{
struct lvmcache_info *info;
/* Invalid if any info is invalid */
dm_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)
{
@@ -791,7 +752,7 @@ char *lvmcache_vgname_from_pvid(struct cmd_context *cmd, const char *pvid)
struct lvmcache_info *info;
char *vgname;
if (!lvmcache_device_from_pvid(cmd, (const struct id *)pvid, NULL)) {
if (!lvmcache_device_from_pvid(cmd, (const struct id *)pvid, NULL, NULL)) {
log_error("Couldn't find device with uuid %s.", pvid);
return NULL;
}
@@ -807,42 +768,19 @@ char *lvmcache_vgname_from_pvid(struct cmd_context *cmd, const char *pvid)
return vgname;
}
/*
* FIXME: get rid of the CACHE_INVALID state and rescanning
* infos with that flag. The code should just know which devices
* need scanning and when.
*/
static int _label_scan_invalid(struct cmd_context *cmd)
static void _rescan_entry(struct lvmcache_info *info)
{
struct dm_list devs;
struct dm_hash_node *n;
struct device_list *devl;
struct lvmcache_info *info;
int dev_count = 0;
int ret;
struct label *label;
dm_list_init(&devs);
if (info->status & CACHE_INVALID)
(void) label_read(info->dev, &label, UINT64_C(0));
}
dm_hash_iterate(n, _pvid_hash) {
if (!(info = dm_hash_get_data(_pvid_hash, n)))
continue;
static int _scan_invalid(void)
{
dm_hash_iter(_pvid_hash, (dm_hash_iterate_fn) _rescan_entry);
if (!(info->status & CACHE_INVALID))
continue;
if (!(devl = dm_pool_zalloc(cmd->mem, sizeof(*devl))))
return_0;
devl->dev = info->dev;
dm_list_add(&devs, &devl->list);
dev_count++;
}
log_debug_cache("Scanning %d devs with invalid info.", dev_count);
ret = label_scan_devs(cmd, &devs);
return ret;
return 1;
}
/*
@@ -1157,89 +1095,17 @@ next:
goto next;
}
/*
* The initial label_scan at the start of the command is done without
* holding VG locks. Then for each VG identified during the label_scan,
* vg_read(vgname) is called while holding the VG lock. The labels
* and metadata on this VG's devices could have changed between the
* initial unlocked label_scan and the current vg_read(). So, we reread
* the labels/metadata for each device in the VG now that we hold the
* lock, and use this for processing the VG.
*
* FIXME: In some cases, the data read by label_scan may be fine, and not
* need to be reread here. e.g. a reporting command, possibly with a
* special option, could skip this second reread. Or, we could look
* at the VG seqno in each copy of the metadata read in the first label
* scan, and if they all match, consider it good enough to use for
* reporting without rereading it. (A command modifying the VG would
* always want to reread while the lock is held before modifying.)
*
* A label scan is ultimately creating associations between devices
* and VGs so that when vg_read wants to get VG metadata, it knows
* which devices to read. In the special case where VG metadata is
* stored in files on the file system (configured in lvm.conf), the
* vginfo->independent_metadata_location flag is set during label scan.
* When we get here to rescan, we are revalidating the device to VG
* mapping from label scan by repeating the label scan on a subset of
* devices. If we see independent_metadata_location is set from the
* initial label scan, we know that there is nothing to do because
* there is no device to VG mapping to revalidate, since the VG metadata
* comes directly from files.
*/
int lvmcache_label_rescan_vg(struct cmd_context *cmd, const char *vgname, const char *vgid)
{
struct dm_list devs;
struct device_list *devl;
struct lvmcache_vginfo *vginfo;
struct lvmcache_info *info;
if (lvmetad_used())
return 1;
dm_list_init(&devs);
if (!(vginfo = lvmcache_vginfo_from_vgname(vgname, vgid)))
return_0;
/*
* When the VG metadata is from an independent location,
* then rescanning the devices in the VG won't find the
* metadata, and will destroy the vginfo/info associations
* that were created during label scan when the
* independent locations were read.
*/
if (vginfo->independent_metadata_location)
return 1;
dm_list_iterate_items(info, &vginfo->infos) {
if (!(devl = dm_malloc(sizeof(*devl)))) {
log_error("device_list element allocation failed");
return 0;
}
devl->dev = info->dev;
dm_list_add(&devs, &devl->list);
}
label_scan_devs(cmd, &devs);
/*
* TODO: grab vginfo again, and compare vginfo->infos
* to what was found above before rereading labels.
* If there are any info->devs now that were not in the
* first devs list, then do label_read on those also.
*/
return 1;
}
int lvmcache_label_scan(struct cmd_context *cmd)
{
struct dm_list del_cache_devs;
struct dm_list add_cache_devs;
struct lvmcache_info *info;
struct device_list *devl;
struct label *label;
struct dev_iter *iter;
struct device *dev;
struct format_type *fmt;
int dev_count = 0;
int r = 0;
@@ -1257,40 +1123,34 @@ int lvmcache_label_scan(struct cmd_context *cmd)
goto out;
}
/*
* Scan devices whose info struct has the INVALID flag set.
* When scanning has read the pv_header, mda_header and
* mda locations, it will clear the INVALID flag (via
* lvmcache_make_valid).
*/
if (_has_scanned && !_force_label_scan) {
r = _label_scan_invalid(cmd);
r = _scan_invalid();
goto out;
}
if (_force_label_scan && (cmd->full_filter && !cmd->full_filter->use_count) && !refresh_filters(cmd))
goto_out;
if (!cmd->full_filter) {
log_error("label scan is missing full filter");
if (!cmd->full_filter || !(iter = dev_iter_create(cmd->full_filter, _force_label_scan))) {
log_error("dev_iter creation failed");
goto out;
}
log_very_verbose("Scanning device labels");
/*
* Duplicates found during this label scan are added to _found_duplicate_devs().
*/
_destroy_duplicate_device_list(&_found_duplicate_devs);
/*
* Do the actual scanning. This populates lvmcache
* with infos/vginfos based on reading headers from
* each device, and a vg summary from each mda.
*
* Note that this will *skip* scanning a device if
* an info struct already exists in lvmcache for
* the device.
*/
label_scan(cmd);
while ((dev = dev_iter_get(iter))) {
(void) label_read(dev, &label, UINT64_C(0));
dev_count++;
}
dev_iter_destroy(iter);
log_very_verbose("Scanned %d device labels", dev_count);
/*
* _choose_preferred_devs() returns:
@@ -1324,7 +1184,7 @@ int lvmcache_label_scan(struct cmd_context *cmd)
dm_list_iterate_items(devl, &add_cache_devs) {
log_debug_cache("Rescan preferred device %s for lvmcache", dev_name(devl->dev));
label_read(devl->dev, NULL, UINT64_C(0));
(void) label_read(devl->dev, &label, UINT64_C(0));
}
dm_list_splice(&_unused_duplicate_devs, &del_cache_devs);
@@ -1356,12 +1216,129 @@ int lvmcache_label_scan(struct cmd_context *cmd)
return r;
}
struct volume_group *lvmcache_get_vg(struct cmd_context *cmd, const char *vgname,
const char *vgid, unsigned precommitted)
{
struct lvmcache_vginfo *vginfo;
struct volume_group *vg = NULL;
struct format_instance *fid;
struct format_instance_ctx fic;
/*
* We currently do not store precommitted metadata in lvmetad at
* all. This means that any request for precommitted metadata is served
* using the classic scanning mechanics, and read from disk or from
* lvmcache.
*/
if (lvmetad_used() && !precommitted) {
/* Still serve the locally cached VG if available */
if (vgid && (vginfo = lvmcache_vginfo_from_vgid(vgid)) &&
vginfo->vgmetadata && (vg = vginfo->cached_vg))
goto out;
return lvmetad_vg_lookup(cmd, vgname, vgid);
}
if (!vgid || !(vginfo = lvmcache_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 && !critical_section()))
return NULL;
/* Use already-cached VG struct when available */
if ((vg = vginfo->cached_vg) && !vginfo->cached_vg_invalidated)
goto out;
release_vg(vginfo->cached_vg);
fic.type = FMT_INSTANCE_MDAS | FMT_INSTANCE_AUX_MDAS;
fic.context.vg_ref.vg_name = vginfo->vgname;
fic.context.vg_ref.vg_id = vgid;
if (!(fid = vginfo->fmt->ops->create_instance(vginfo->fmt, &fic)))
return_NULL;
/* Build config tree from vgmetadata, if not yet cached */
if (!vginfo->cft &&
!(vginfo->cft =
config_tree_from_string_without_dup_node_check(vginfo->vgmetadata)))
goto_bad;
if (!(vg = import_vg_from_config_tree(vginfo->cft, fid)))
goto_bad;
/* Cache VG struct for reuse */
vginfo->cached_vg = vg;
vginfo->holders = 1;
vginfo->vg_use_count = 0;
vginfo->cached_vg_invalidated = 0;
vg->vginfo = vginfo;
if (!dm_pool_lock(vg->vgmem, detect_internal_vg_cache_corruption()))
goto_bad;
out:
vginfo->holders++;
vginfo->vg_use_count++;
log_debug_cache("Using cached %smetadata for VG %s with %u holder(s).",
vginfo->precommitted ? "pre-committed " : "",
vginfo->vgname, vginfo->holders);
return vg;
bad:
_free_cached_vgmetadata(vginfo);
return NULL;
}
// #if 0
int lvmcache_vginfo_holders_dec_and_test_for_zero(struct lvmcache_vginfo *vginfo)
{
log_debug_cache("VG %s decrementing %d holder(s) at %p.",
vginfo->cached_vg->name, vginfo->holders, vginfo->cached_vg);
if (--vginfo->holders)
return 0;
if (vginfo->vg_use_count > 1)
log_debug_cache("VG %s reused %d times.",
vginfo->cached_vg->name, vginfo->vg_use_count);
/* Debug perform crc check only when it's been used more then once */
if (!dm_pool_unlock(vginfo->cached_vg->vgmem,
detect_internal_vg_cache_corruption() &&
(vginfo->vg_use_count > 1)))
stack;
vginfo->cached_vg->vginfo = NULL;
vginfo->cached_vg = NULL;
return 1;
}
// #endif
int lvmcache_get_vgnameids(struct cmd_context *cmd, int include_internal,
struct dm_list *vgnameids)
{
struct vgnameid_list *vgnl;
struct lvmcache_vginfo *vginfo;
lvmcache_label_scan(cmd);
dm_list_iterate_items(vginfo, &_vginfos) {
if (!include_internal && is_orphan_vg(vginfo->vgname))
continue;
@@ -1466,45 +1443,61 @@ struct dm_list *lvmcache_get_pvids(struct cmd_context *cmd, const char *vgname,
return pvids;
}
int lvmcache_get_vg_devs(struct cmd_context *cmd,
struct lvmcache_vginfo *vginfo,
struct dm_list *devs)
{
struct lvmcache_info *info;
struct device_list *devl;
dm_list_iterate_items(info, &vginfo->infos) {
if (!(devl = dm_pool_zalloc(cmd->mem, sizeof(*devl))))
return_0;
devl->dev = info->dev;
dm_list_add(devs, &devl->list);
}
return 1;
}
static struct device *_device_from_pvid(const struct id *pvid, uint64_t *label_sector)
static struct device *_device_from_pvid(const struct id *pvid,
uint64_t *label_sector)
{
struct lvmcache_info *info;
struct label *label;
if ((info = lvmcache_info_from_pvid((const char *) pvid, NULL, 0))) {
if (info->label && label_sector)
*label_sector = info->label->sector;
return info->dev;
}
if (lvmetad_used()) {
if (info->label && label_sector)
*label_sector = info->label->sector;
return info->dev;
}
if (label_read(info->dev, &label, UINT64_C(0))) {
info = (struct lvmcache_info *) label->info;
if (id_equal(pvid, (struct id *) &info->dev->pvid)) {
if (label_sector)
*label_sector = label->sector;
return info->dev;
}
}
}
return NULL;
}
struct device *lvmcache_device_from_pvid(struct cmd_context *cmd, const struct id *pvid, uint64_t *label_sector)
struct device *lvmcache_device_from_pvid(struct cmd_context *cmd, const struct id *pvid,
unsigned *scan_done_once, uint64_t *label_sector)
{
struct device *dev;
/* Already cached ? */
dev = _device_from_pvid(pvid, label_sector);
if (dev)
return dev;
lvmcache_label_scan(cmd);
/* Try again */
dev = _device_from_pvid(pvid, label_sector);
if (dev)
return dev;
if (critical_section() || (scan_done_once && *scan_done_once))
return NULL;
lvmcache_force_next_label_scan();
lvmcache_label_scan(cmd);
if (scan_done_once)
*scan_done_once = 1;
/* Try again */
dev = _device_from_pvid(pvid, label_sector);
if (dev)
return dev;
log_debug_devs("No device with uuid %s.", (const char *)pvid);
return NULL;
}
@@ -1512,6 +1505,7 @@ const char *lvmcache_pvid_from_devname(struct cmd_context *cmd,
const char *devname)
{
struct device *dev;
struct label *label;
if (!(dev = dev_cache_get(devname, cmd->filter))) {
log_error("%s: Couldn't find device. Check your filters?",
@@ -1519,7 +1513,7 @@ const char *lvmcache_pvid_from_devname(struct cmd_context *cmd,
return NULL;
}
if (!label_read(dev, NULL, UINT64_C(0)))
if (!(label_read(dev, &label, UINT64_C(0))))
return NULL;
return dev->pvid;
@@ -1541,6 +1535,8 @@ static int _free_vginfo(struct lvmcache_vginfo *vginfo)
struct lvmcache_vginfo *primary_vginfo, *vginfo2;
int r = 1;
_free_cached_vgmetadata(vginfo);
vginfo2 = primary_vginfo = lvmcache_vginfo_from_vgname(vginfo->vgname, NULL);
if (vginfo == primary_vginfo) {
@@ -1563,7 +1559,6 @@ static int _free_vginfo(struct lvmcache_vginfo *vginfo)
dm_free(vginfo->system_id);
dm_free(vginfo->vgname);
dm_free(vginfo->creation_host);
_suspended_vg_free(vginfo, 1, 1);
if (*vginfo->vgid && _vgid_hash &&
lvmcache_vginfo_from_vgid(vginfo->vgid) == vginfo)
@@ -1605,6 +1600,8 @@ void lvmcache_del(struct lvmcache_info *info)
info->label->labeller->ops->destroy_label(info->label->labeller,
info->label);
dm_free(info);
return;
}
/*
@@ -1901,7 +1898,8 @@ static int _lvmcache_update_vgstatus(struct lvmcache_info *info, uint32_t vgstat
info->vginfo->creation_host))
goto set_lock_type;
dm_free(info->vginfo->creation_host);
if (info->vginfo->creation_host)
dm_free(info->vginfo->creation_host);
if (!(info->vginfo->creation_host = dm_strdup(creation_host))) {
log_error("cache creation host alloc failed for %s.",
@@ -1920,7 +1918,8 @@ set_lock_type:
if (info->vginfo->lock_type && !strcmp(lock_type, info->vginfo->lock_type))
goto set_system_id;
dm_free(info->vginfo->lock_type);
if (info->vginfo->lock_type)
dm_free(info->vginfo->lock_type);
if (!(info->vginfo->lock_type = dm_strdup(lock_type))) {
log_error("cache lock_type alloc failed for %s", lock_type);
@@ -1938,7 +1937,8 @@ set_system_id:
if (info->vginfo->system_id && !strcmp(system_id, info->vginfo->system_id))
goto out;
dm_free(info->vginfo->system_id);
if (info->vginfo->system_id)
dm_free(info->vginfo->system_id);
if (!(info->vginfo->system_id = dm_strdup(system_id))) {
log_error("cache system_id alloc failed for %s", system_id);
@@ -2002,6 +2002,12 @@ int lvmcache_update_vgname_and_id(struct lvmcache_info *info, struct lvmcache_vg
!is_orphan_vg(info->vginfo->vgname) && critical_section())
return 1;
/* If making a PV into an orphan, any cached VG metadata may become
* invalid, incorrectly still referencing device structs.
* (Example: pvcreate -ff) */
if (is_orphan_vg(vgname) && info->vginfo && !is_orphan_vg(info->vginfo->vgname))
info->vginfo->cached_vg_invalidated = 1;
/* If moving PV from orphan to real VG, always mark it valid */
if (!is_orphan_vg(vgname))
info->status &= ~CACHE_INVALID;
@@ -2039,6 +2045,10 @@ int lvmcache_update_vg(struct volume_group *vg, unsigned precommitted)
return_0;
}
/* store text representation of vg to cache */
if (vg->cmd->current_settings.cache_vgmetadata)
_store_metadata(vg, precommitted);
return 1;
}
@@ -2372,29 +2382,56 @@ int lvmcache_fid_add_mdas_vg(struct lvmcache_vginfo *vginfo, struct format_insta
return 1;
}
int lvmcache_populate_pv_fields(struct lvmcache_info *info,
struct volume_group *vg,
struct physical_volume *pv)
static int _get_pv_if_in_vg(struct lvmcache_info *info,
struct physical_volume *pv)
{
struct data_area_list *da;
if (!info->label) {
log_error("No cached label for orphan PV %s", pv_dev_name(pv));
return 0;
char vgname[NAME_LEN + 1];
char vgid[ID_LEN + 1];
if (info->vginfo && info->vginfo->vgname &&
!is_orphan_vg(info->vginfo->vgname)) {
/*
* get_pv_from_vg_by_id() may call
* lvmcache_label_scan() and drop cached
* vginfo so make a local copy of string.
*/
(void) dm_strncpy(vgname, info->vginfo->vgname, sizeof(vgname));
memcpy(vgid, info->vginfo->vgid, sizeof(vgid));
if (get_pv_from_vg_by_id(info->fmt, vgname, vgid,
info->dev->pvid, pv))
return 1;
}
pv->label_sector = info->label->sector;
return 0;
}
int lvmcache_populate_pv_fields(struct lvmcache_info *info,
struct physical_volume *pv,
int scan_label_only)
{
struct data_area_list *da;
/* Have we already cached vgname? */
if (!scan_label_only && _get_pv_if_in_vg(info, pv))
return 1;
/* Perform full scan (just the first time) and try again */
if (!scan_label_only && !critical_section() && !full_scan_done()) {
lvmcache_force_next_label_scan();
lvmcache_label_scan(info->fmt->cmd);
if (_get_pv_if_in_vg(info, pv))
return 1;
}
/* Orphan */
pv->dev = info->dev;
pv->fmt = info->fmt;
pv->size = info->device_size >> SECTOR_SHIFT;
pv->vg_name = FMT_TEXT_ORPHAN_VG_NAME;
memcpy(&pv->id, &info->dev->pvid, sizeof(pv->id));
if (!pv->size) {
log_error("PV %s size is zero.", dev_name(info->dev));
return 0;
}
/* Currently only support exactly one data area */
if (dm_list_size(&info->das) != 1) {
log_error("Must be exactly one data area (found %d) on PV %s",
@@ -2575,10 +2612,6 @@ struct label *lvmcache_get_label(struct lvmcache_info *info) {
return info->label;
}
/*
* After label_scan reads pv_header, mda_header and mda locations
* from a PV, it clears the INVALID flag.
*/
void lvmcache_make_valid(struct lvmcache_info *info) {
info->status &= ~CACHE_INVALID;
}
@@ -2634,14 +2667,6 @@ int lvmcache_vgid_is_cached(const char *vgid) {
return 1;
}
void lvmcache_set_independent_location(const char *vgname)
{
struct lvmcache_vginfo *vginfo;
if ((vginfo = lvmcache_vginfo_from_vgname(vgname, NULL)))
vginfo->independent_metadata_location = 1;
}
/*
* Return true iff it is impossible to find out from this info alone whether the
* PV in question is or is not an orphan.

24
lib/cache/lvmcache.h vendored
View File

@@ -74,7 +74,6 @@ void lvmcache_destroy(struct cmd_context *cmd, int retain_orphans, int reset);
*/
void lvmcache_force_next_label_scan(void);
int lvmcache_label_scan(struct cmd_context *cmd);
int lvmcache_label_rescan_vg(struct cmd_context *cmd, const char *vgname, const char *vgid);
/* Add/delete a device */
struct lvmcache_info *lvmcache_add(struct labeller *labeller, const char *pvid,
@@ -106,8 +105,10 @@ struct lvmcache_vginfo *lvmcache_vginfo_from_vgid(const char *vgid);
struct lvmcache_info *lvmcache_info_from_pvid(const char *pvid, struct device *dev, int valid_only);
const char *lvmcache_vgname_from_vgid(struct dm_pool *mem, const char *vgid);
const char *lvmcache_vgid_from_vgname(struct cmd_context *cmd, const char *vgname);
struct device *lvmcache_device_from_pvid(struct cmd_context *cmd, const struct id *pvid, uint64_t *label_sector);
const char *lvmcache_pvid_from_devname(struct cmd_context *cmd, const char *devname);
struct device *lvmcache_device_from_pvid(struct cmd_context *cmd, const struct id *pvid,
unsigned *scan_done_once, uint64_t *label_sector);
const char *lvmcache_pvid_from_devname(struct cmd_context *cmd,
const char *devname);
char *lvmcache_vgname_from_pvid(struct cmd_context *cmd, const char *pvid);
const char *lvmcache_vgname_from_info(struct lvmcache_info *info);
const struct format_type *lvmcache_fmt_from_info(struct lvmcache_info *info);
@@ -133,6 +134,9 @@ int lvmcache_get_vgnameids(struct cmd_context *cmd, int include_internal,
struct dm_list *lvmcache_get_pvids(struct cmd_context *cmd, const char *vgname,
const char *vgid);
/* Returns cached volume group metadata. */
struct volume_group *lvmcache_get_vg(struct cmd_context *cmd, const char *vgname,
const char *vgid, unsigned precommitted);
void lvmcache_drop_metadata(const char *vgname, int drop_precommitted);
void lvmcache_commit_metadata(const char *vgname);
@@ -142,8 +146,8 @@ int lvmcache_fid_add_mdas(struct lvmcache_info *info, struct format_instance *fi
int lvmcache_fid_add_mdas_pv(struct lvmcache_info *info, struct format_instance *fid);
int lvmcache_fid_add_mdas_vg(struct lvmcache_vginfo *vginfo, struct format_instance *fid);
int lvmcache_populate_pv_fields(struct lvmcache_info *info,
struct volume_group *vg,
struct physical_volume *pv);
struct physical_volume *pv,
int scan_label_only);
int lvmcache_check_format(struct lvmcache_info *info, const struct format_type *fmt);
void lvmcache_del_mdas(struct lvmcache_info *info);
void lvmcache_del_das(struct lvmcache_info *info);
@@ -177,7 +181,7 @@ int lvmcache_foreach_ba(struct lvmcache_info *info,
int (*fun)(struct disk_locn *, void *),
void *baton);
int lvmcache_foreach_pv(struct lvmcache_vginfo *vginfo,
int lvmcache_foreach_pv(struct lvmcache_vginfo *vg,
int (*fun)(struct lvmcache_info *, void *), void * baton);
uint64_t lvmcache_device_size(struct lvmcache_info *info);
@@ -211,12 +215,4 @@ void lvmcache_remove_unchosen_duplicate(struct device *dev);
int lvmcache_pvid_in_unchosen_duplicates(const char *pvid);
void lvmcache_save_suspended_vg(struct volume_group *vg, int precommitted);
struct volume_group *lvmcache_get_suspended_vg(const char *vgid);
void lvmcache_drop_suspended_vg(struct volume_group *vg);
int lvmcache_get_vg_devs(struct cmd_context *cmd,
struct lvmcache_vginfo *vginfo,
struct dm_list *devs);
void lvmcache_set_independent_location(const char *vgname);
#endif

515
lib/cache/lvmetad.c vendored
View File

@@ -39,7 +39,7 @@ static int64_t _lvmetad_update_timeout;
static int _found_lvm1_metadata = 0;
static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct volume_group *vg, const char *vgid, struct format_type *fmt);
static struct volume_group *lvmetad_pvscan_vg(struct cmd_context *cmd, struct volume_group *vg);
static uint64_t _monotonic_seconds(void)
{
@@ -66,7 +66,7 @@ static int _log_debug_inequality(const char *name, struct dm_config_node *a, str
log_debug_lvmetad("VG %s metadata inequality at %s / %s: %s / %s",
name, a->key, b->key, av->v.str, bv->v.str);
else if (a->v->type == DM_CFG_INT && b->v->type == DM_CFG_INT)
log_debug_lvmetad("VG %s metadata inequality at %s / %s: " FMTd64 " / " FMTd64,
log_debug_lvmetad("VG %s metadata inequality at %s / %s: " FMTi64 " / " FMTi64,
name, a->key, b->key, av->v.i, bv->v.i);
else
log_debug_lvmetad("VG %s metadata inequality at %s / %s: type %d / type %d",
@@ -145,14 +145,13 @@ int lvmetad_connect(struct cmd_context *cmd)
_lvmetad_use = 1;
_lvmetad_cmd = cmd;
return 1;
} else {
log_debug_lvmetad("Failed to connect to lvmetad: %s", strerror(_lvmetad.error));
_lvmetad_connected = 0;
_lvmetad_use = 0;
_lvmetad_cmd = NULL;
return 0;
}
log_debug_lvmetad("Failed to connect to lvmetad: %s", strerror(_lvmetad.error));
_lvmetad_connected = 0;
_lvmetad_use = 0;
_lvmetad_cmd = NULL;
return 0;
}
int lvmetad_used(void)
@@ -663,7 +662,7 @@ static int _lvmetad_handle_reply(daemon_reply reply, const char *id, const char
}
if (reply.error) {
log_error("lvmetad cannot be used due to error: %s", strerror(reply.error));
log_warn("WARNING: lvmetad cannot be used due to error: %s", strerror(reply.error));
goto fail;
}
@@ -1090,17 +1089,14 @@ struct volume_group *lvmetad_vg_lookup(struct cmd_context *cmd, const char *vgna
* invalidated the cached vg.
*/
if (rescan) {
if (!(vg2 = _lvmetad_pvscan_vg(cmd, vg, vgid, fmt))) {
if (!(vg2 = lvmetad_pvscan_vg(cmd, vg))) {
log_debug_lvmetad("VG %s from lvmetad not found during rescan.", vgname);
fid = NULL;
release_vg(vg);
vg = NULL;
goto out;
}
fid->ref_count++;
release_vg(vg);
fid->ref_count--;
fmt->ops->destroy_instance(fid);
vg = vg2;
fid = vg2->fid;
}
@@ -1108,14 +1104,14 @@ struct volume_group *lvmetad_vg_lookup(struct cmd_context *cmd, const char *vgna
dm_list_iterate_items(pvl, &vg->pvs) {
if (!_pv_update_struct_pv(pvl->pv, fid)) {
vg = NULL;
goto_out; /* FIXME: use an error path that disables lvmetad */
goto_out; /* FIXME error path */
}
}
dm_list_iterate_items(pvl, &vg->pvs_outdated) {
if (!_pv_update_struct_pv(pvl->pv, fid)) {
vg = NULL;
goto_out; /* FIXME: use an error path that disables lvmetad */
goto_out; /* FIXME error path */
}
}
@@ -1284,7 +1280,7 @@ int lvmetad_vg_update_finish(struct volume_group *vg)
if (pvl->pv->dev && !lvmetad_pv_found(vg->cmd, &pvl->pv->id, pvl->pv->dev,
vgu->fid ? vgu->fid->fmt : pvl->pv->fmt,
pvl->pv->label_sector, NULL, NULL, NULL))
return_0;
return 0;
}
vg->lvmetad_update_pending = 0;
@@ -1308,7 +1304,7 @@ int lvmetad_vg_remove_pending(struct volume_group *vg)
reply = _lvmetad_send(vg->cmd, "set_vg_info",
"name = %s", vg->name,
"uuid = %s", uuid,
"version = %"PRId64, (int64_t)0,
"version = %d", 0,
NULL);
if (!_lvmetad_handle_reply(reply, "set_vg_info", vg->name, NULL)) {
@@ -1759,7 +1755,6 @@ int lvmetad_pv_gone_by_dev(struct device *dev)
*/
struct _lvmetad_pvscan_baton {
struct cmd_context *cmd;
struct volume_group *vg;
struct format_instance *fid;
};
@@ -1767,14 +1762,10 @@ struct _lvmetad_pvscan_baton {
static int _lvmetad_pvscan_single(struct metadata_area *mda, void *baton)
{
struct _lvmetad_pvscan_baton *b = baton;
struct device *mda_dev = mda_get_device(mda);
struct label_read_data *ld;
struct volume_group *vg;
ld = get_label_read_data(b->cmd, mda_dev);
if (mda_is_ignored(mda) ||
!(vg = mda->ops->vg_read(b->fid, "", mda, ld, NULL, NULL)))
!(vg = mda->ops->vg_read(b->fid, "", mda, NULL, NULL, 1)))
return 1;
/* FIXME Also ensure contents match etc. */
@@ -1786,37 +1777,6 @@ static int _lvmetad_pvscan_single(struct metadata_area *mda, void *baton)
return 1;
}
/*
* FIXME: handle errors and do proper comparison of metadata from each area
* like vg_read and fall back to real vg_read from disk if there's any problem.
*/
static int _lvmetad_pvscan_vg_single(struct metadata_area *mda, void *baton)
{
struct _lvmetad_pvscan_baton *b = baton;
struct device *mda_dev = mda_get_device(mda);
struct label_read_data *ld;
struct volume_group *vg = NULL;
if (mda_is_ignored(mda))
return 1;
ld = get_label_read_data(b->cmd, mda_dev);
if (!(vg = mda->ops->vg_read(b->fid, "", mda, ld, NULL, NULL)))
return 1;
if (!b->vg)
b->vg = vg;
else if (vg->seqno > b->vg->seqno) {
release_vg(b->vg);
b->vg = vg;
} else
release_vg(vg);
return 1;
}
/*
* The lock manager may detect that the vg cached in lvmetad is out of date,
* due to something like an lvcreate from another host.
@@ -1826,41 +1786,41 @@ static int _lvmetad_pvscan_vg_single(struct metadata_area *mda, void *baton)
* the VG, and that PV may have been reused for another VG.
*/
static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct volume_group *vg,
const char *vgid, struct format_type *fmt)
static struct volume_group *lvmetad_pvscan_vg(struct cmd_context *cmd, struct volume_group *vg)
{
char pvid_s[ID_LEN + 1] __attribute__((aligned(8)));
char uuid[64] __attribute__((aligned(8)));
struct label *label;
struct volume_group *vg_ret = NULL;
struct dm_config_tree *vgmeta_ret = NULL;
struct dm_config_tree *vgmeta;
struct pv_list *pvl, *pvl_new;
struct device_list *devl, *devlsafe;
struct device_list *devl, *devl_new, *devlsafe;
struct dm_list pvs_scan;
struct dm_list pvs_drop;
struct lvmcache_vginfo *vginfo = NULL;
struct dm_list pvs_new;
struct lvmcache_info *info = NULL;
struct format_instance *fid;
struct format_instance_ctx fic = { .type = 0 };
struct _lvmetad_pvscan_baton baton;
struct volume_group *save_vg;
struct dm_config_tree *save_meta;
struct device *save_dev = NULL;
uint32_t save_seqno = 0;
int found_new_pvs = 0;
int retried_reads = 0;
int missing_devs = 0;
int check_new_pvs = 0;
int found;
save_vg = NULL;
save_meta = NULL;
save_dev = NULL;
save_seqno = 0;
dm_list_init(&pvs_scan);
dm_list_init(&pvs_drop);
dm_list_init(&pvs_new);
log_debug_lvmetad("Rescan VG %s to update lvmetad (seqno %u).", vg->name, vg->seqno);
log_debug_lvmetad("Rescanning VG %s (seqno %u).", vg->name, vg->seqno);
/*
* Make sure this command knows about all PVs from lvmetad.
* Another host may have added a PV to the VG, and some
* commands do not always populate their lvmcache with
* all devs from lvmetad, so they would fail to find
* the new PV when scanning the VG. So make sure this
* command knows about all PVs from lvmetad.
*/
lvmcache_seed_infos_from_lvmetad(cmd);
@@ -1875,111 +1835,54 @@ static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct v
dm_list_add(&pvs_scan, &devl->list);
}
/*
* Rescan labels/metadata only from devs that we previously
* saw in the VG. If we find below that there are new PVs
* in the VG, we'll have to rescan all devices to find which
* device(s) are now being used.
*/
log_debug_lvmetad("Rescan VG %s scanning data from devs in previous metadata.", vg->name);
label_scan_devs(cmd, &pvs_scan);
scan_more:
/*
* Check if any pvs_scan entries are no longer PVs.
* In that case, label_read/_find_label_header will have
* found no label_header, and would have dropped the
* info struct for the device from lvmcache. So, if
* we look up the info struct here and don't find it,
* we can infer it's no longer a PV.
*
* FIXME: we should record specific results from the
* label_read and then check specifically for whatever
* result means "no label was found", rather than going
* about this indirectly via the lvmcache side effects.
*/
dm_list_iterate_items_safe(devl, devlsafe, &pvs_scan) {
if (!(info = lvmcache_info_from_pvid(devl->dev->pvid, devl->dev, 0))) {
/* Another host removed this PV from the VG. */
log_debug_lvmetad("Rescan VG %s from %s dropping dev (no label).",
vg->name, dev_name(devl->dev));
dm_list_move(&pvs_drop, &devl->list);
}
}
fic.type = FMT_INSTANCE_MDAS | FMT_INSTANCE_AUX_MDAS;
fic.context.vg_ref.vg_name = vg->name;
fic.context.vg_ref.vg_id = vgid;
retry_reads:
if (!(fid = fmt->ops->create_instance(fmt, &fic))) {
/* FIXME: are there only internal reasons for failures here? */
log_error("Reading VG %s failed to create format instance.", vg->name);
return NULL;
}
/* FIXME: not sure if this is necessary */
fid->ref_count++;
baton.fid = fid;
baton.cmd = cmd;
/*
* FIXME: this vg_read path does not have the ability to repair
* any problems with the VG, e.g. VG on one dev has an older
* seqno. When vg_read() is reworked, we need to fall back
* to using that from here (and vg_read's from lvmetad) when
* there is a problem. Perhaps by disabling lvmetad when a
* VG problem is detected, causing commands to fully fall
* back to disk, which will repair the VG. Then lvmetad can
* be repopulated and re-enabled (possibly automatically.)
*/
/*
* Do a low level vg_read on each dev, verify the vg returned
* from metadata on each device is for the VG being read
* (the PV may have been removed from the VG being read and
* added to a different one), and return this vg to the caller
* as the current vg to use.
*
* The label scan above will have saved in lvmcache which
* vg each device is used in, so we could figure that part
* out without doing the vg_read.
* Run the equivalent of lvmetad_pvscan_single on each dev in the VG.
*/
dm_list_iterate_items_safe(devl, devlsafe, &pvs_scan) {
if (!devl->dev)
continue;
log_debug_lvmetad("Rescan VG %s getting metadata from %s.",
vg->name, dev_name(devl->dev));
log_debug_lvmetad("Rescan VG %s scanning %s.", vg->name, dev_name(devl->dev));
if (!label_read(devl->dev, &label, 0)) {
/* Another host removed this PV from the VG. */
log_debug_lvmetad("Rescan VG %s found %s was removed.", vg->name, dev_name(devl->dev));
if ((info = lvmcache_info_from_pvid(devl->dev->pvid, NULL, 0)))
lvmcache_del(info);
/*
* The info struct for this dev knows what and where
* the mdas are for this dev (the label scan saved
* the mda locations for this dev on the lvmcache info struct).
*/
if (!(info = lvmcache_info_from_pvid(devl->dev->pvid, devl->dev, 0))) {
log_debug_lvmetad("Rescan VG %s from %s dropping dev (no info).",
vg->name, dev_name(devl->dev));
dm_list_move(&pvs_drop, &devl->list);
continue;
}
info = (struct lvmcache_info *) label->info;
baton.vg = NULL;
baton.fid = lvmcache_fmt(info)->ops->create_instance(lvmcache_fmt(info), &fic);
if (!baton.fid)
return_NULL;
if (baton.fid->fmt->features & FMT_OBSOLETE) {
log_debug_lvmetad("Ignoring obsolete format on PV %s in VG %s.", dev_name(devl->dev), vg->name);
lvmcache_fmt(info)->ops->destroy_instance(baton.fid);
dm_list_move(&pvs_drop, &devl->list);
continue;
}
/*
* Read VG metadata from this dev's mdas.
*/
lvmcache_foreach_mda(info, _lvmetad_pvscan_vg_single, &baton);
lvmcache_foreach_mda(info, _lvmetad_pvscan_single, &baton);
/*
* The PV may have been removed from the VG by another host
* since we last read the VG.
*/
if (!baton.vg) {
log_debug_lvmetad("Rescan VG %s from %s dropping dev (no metadata).",
vg->name, dev_name(devl->dev));
log_debug_lvmetad("Rescan VG %s did not find %s.", vg->name, dev_name(devl->dev));
lvmcache_fmt(info)->ops->destroy_instance(baton.fid);
dm_list_move(&pvs_drop, &devl->list);
continue;
}
@@ -1989,15 +1892,10 @@ static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct v
* different VG since we last read the VG.
*/
if (strcmp(baton.vg->name, vg->name)) {
log_debug_lvmetad("Rescan VG %s from %s dropping dev (other VG %s).",
vg->name, dev_name(devl->dev), baton.vg->name);
release_vg(baton.vg);
continue;
}
if (!(vgmeta = export_vg_to_config_tree(baton.vg))) {
log_error("VG export to config tree failed");
log_debug_lvmetad("Rescan VG %s found different VG %s on PV %s.",
vg->name, baton.vg->name, dev_name(devl->dev));
release_vg(baton.vg);
dm_list_move(&pvs_drop, &devl->list);
continue;
}
@@ -2007,35 +1905,20 @@ static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct v
* read from each other dev.
*/
if (save_vg && (save_seqno != baton.vg->seqno)) {
/* FIXME: fall back to vg_read to correct this. */
log_warn("WARNING: inconsistent metadata for VG %s on devices %s seqno %u and %s seqno %u.",
vg->name, dev_name(save_dev), save_seqno,
dev_name(devl->dev), baton.vg->seqno);
log_warn("WARNING: temporarily disable lvmetad to repair metadata.");
if (!save_seqno)
save_seqno = baton.vg->seqno;
/* Use the most recent */
if (save_seqno < baton.vg->seqno) {
release_vg(save_vg);
dm_config_destroy(save_meta);
save_vg = baton.vg;
save_meta = vgmeta;
save_seqno = baton.vg->seqno;
save_dev = devl->dev;
} else {
release_vg(baton.vg);
dm_config_destroy(vgmeta);
}
continue;
if (!(vgmeta = export_vg_to_config_tree(baton.vg))) {
log_error("VG export to config tree failed");
release_vg(baton.vg);
return NULL;
}
if (!save_vg) {
save_vg = baton.vg;
save_meta = vgmeta;
save_seqno = baton.vg->seqno;
if (!vgmeta_ret) {
vgmeta_ret = vgmeta;
save_dev = devl->dev;
} else {
struct dm_config_node *meta1 = save_meta->root;
struct dm_config_node *meta1 = vgmeta_ret->root;
struct dm_config_node *meta2 = vgmeta->root;
struct dm_config_node *sib1 = meta1->sib;
struct dm_config_node *sib2 = meta2->sib;
@@ -2060,128 +1943,73 @@ static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct v
meta2->sib = NULL;
if (compare_config(meta1, meta2)) {
/* FIXME: fall back to vg_read to correct this. */
log_warn("WARNING: inconsistent metadata for VG %s on devices %s seqno %u and %s seqno %u.",
vg->name, dev_name(save_dev), save_seqno,
dev_name(devl->dev), baton.vg->seqno);
log_warn("WARNING: temporarily disable lvmetad to repair metadata.");
log_error("VG %s metadata comparison failed for device %s vs %s",
vg->name, dev_name(devl->dev), save_dev ? dev_name(save_dev) : "none");
_log_debug_inequality(vg->name, save_meta->root, vgmeta->root);
_log_debug_inequality(vg->name, vgmeta_ret->root, vgmeta->root);
meta1->sib = sib1;
meta2->sib = sib2;
/* no right choice, just use the previous copy */
release_vg(baton.vg);
dm_config_destroy(vgmeta);
dm_config_destroy(vgmeta_ret);
release_vg(baton.vg);
return NULL;
}
meta1->sib = sib1;
meta2->sib = sib2;
release_vg(baton.vg);
dm_config_destroy(vgmeta);
}
}
/* FIXME: see above */
fid->ref_count--;
/*
* Look for any new PVs in the VG metadata that were not in our
* previous version of the VG.
*
* (Don't look for new PVs after a rescan and retry.)
*/
found_new_pvs = 0;
if (save_vg && !retried_reads) {
dm_list_iterate_items(pvl_new, &save_vg->pvs) {
found = 0;
dm_list_iterate_items(pvl, &vg->pvs) {
if (pvl_new->pv->dev != pvl->pv->dev)
continue;
found = 1;
break;
}
/*
* PV in new VG metadata not found in old VG metadata.
* There's a good chance we don't know about this new
* PV or what device it's on; a label scan is needed
* of all devices so we know which device the VG is
* now using.
*/
if (!found) {
found_new_pvs++;
strncpy(pvid_s, (char *) &pvl_new->pv->id, sizeof(pvid_s) - 1);
if (!id_write_format((const struct id *)&pvid_s, uuid, sizeof(uuid)))
stack;
log_debug_lvmetad("Rescan VG %s found new PV %s.", vg->name, uuid);
}
}
}
if (!save_vg && retried_reads) {
log_error("VG %s not found after rescanning devices.", vg->name);
goto out;
}
/*
* Do a full rescan of devices, then look up which devices the
* scan found for this VG name, and select those devices to
* read metadata from in the loop above (rather than the list
* of devices we created from our last copy of the vg metadata.)
*
* Case 1: VG we knew is no longer on any of the devices we knew it
* to be on (save_vg is NULL, which means the metadata wasn't found
* when reading mdas on each of the initial pvs_scan devices).
* Rescan all devs and then retry reading metadata from the devs that
* the scan finds associated with this VG.
*
* Case 2: VG has new PVs but we don't know what devices they are
* so rescan all devs and then retry reading metadata from the devs
* that the scan finds associated with this VG.
*
* (N.B. after a retry, we don't check for found_new_pvs.)
*/
if (!save_vg || found_new_pvs) {
if (!save_vg)
log_debug_lvmetad("Rescan VG %s did not find VG on previous devs.", vg->name);
if (found_new_pvs)
log_debug_lvmetad("Rescan VG %s scanning all devs to find new PVs.", vg->name);
label_scan_force(cmd);
if (!(vginfo = lvmcache_vginfo_from_vgname(vg->name, NULL))) {
log_error("VG %s vg info not found after rescanning devices.", vg->name);
goto out;
}
/*
* Set pvs_scan to devs that the label scan found
* in the VG and retry the metadata reading loop.
* Look for any new PVs in the VG metadata that were not in our
* previous version of the VG. Add them to pvs_new to be
* scanned in this loop just like the old PVs.
*/
dm_list_init(&pvs_scan);
if (!lvmcache_get_vg_devs(cmd, vginfo, &pvs_scan)) {
log_error("VG %s info devs not found after rescanning devices.", vg->name);
goto out;
if (!check_new_pvs) {
check_new_pvs = 1;
dm_list_iterate_items(pvl_new, &baton.vg->pvs) {
found = 0;
dm_list_iterate_items(pvl, &vg->pvs) {
if (pvl_new->pv->dev != pvl->pv->dev)
continue;
found = 1;
break;
}
if (found)
continue;
if (!pvl_new->pv->dev) {
strncpy(pvid_s, (char *) &pvl_new->pv->id, sizeof(pvid_s) - 1);
if (!id_write_format((const struct id *)&pvid_s, uuid, sizeof(uuid)))
stack;
log_error("Device not found for PV %s in VG %s", uuid, vg->name);
missing_devs++;
continue;
}
if (!(devl_new = dm_pool_zalloc(cmd->mem, sizeof(*devl_new))))
return_NULL;
devl_new->dev = pvl_new->pv->dev;
dm_list_add(&pvs_new, &devl_new->list);
log_debug_lvmetad("Rescan VG %s found %s was added.", vg->name, dev_name(devl_new->dev));
}
}
log_debug_lvmetad("Rescan VG %s has %d PVs after label scan.",
vg->name, dm_list_size(&pvs_scan));
release_vg(baton.vg);
}
if (save_vg)
release_vg(save_vg);
if (save_meta)
dm_config_destroy(save_meta);
save_vg = NULL;
save_meta = NULL;
save_dev = NULL;
save_seqno = 0;
found_new_pvs = 0;
retried_reads = 1;
goto retry_reads;
/*
* Do the same scanning above for any new PVs.
*/
if (!dm_list_empty(&pvs_new)) {
dm_list_init(&pvs_scan);
dm_list_splice(&pvs_scan, &pvs_new);
dm_list_init(&pvs_new);
log_debug_lvmetad("Rescan VG %s found new PVs to scan.", vg->name);
goto scan_more;
}
if (missing_devs) {
if (vgmeta_ret)
dm_config_destroy(vgmeta_ret);
return_NULL;
}
/*
@@ -2190,50 +2018,52 @@ static struct volume_group *_lvmetad_pvscan_vg(struct cmd_context *cmd, struct v
dm_list_iterate_items(devl, &pvs_drop) {
if (!devl->dev)
continue;
log_debug_lvmetad("Rescan VG %s removing %s from lvmetad.", vg->name, dev_name(devl->dev));
if (!lvmetad_pv_gone_by_dev(devl->dev)) {
/* FIXME: use an error path that disables lvmetad */
log_error("Failed to remove %s from lvmetad.", dev_name(devl->dev));
}
log_debug_lvmetad("Rescan VG %s dropping %s.", vg->name, dev_name(devl->dev));
if (!lvmetad_pv_gone_by_dev(devl->dev))
return_NULL;
}
/*
* Update lvmetad with the newly read version of the VG.
* When the seqno is unchanged the cached VG can be left.
* Update the VG in lvmetad.
*/
if (save_vg && (save_seqno != vg->seqno)) {
dm_list_iterate_items(devl, &pvs_scan) {
if (!devl->dev)
continue;
log_debug_lvmetad("Rescan VG %s removing %s from lvmetad to replace.",
vg->name, dev_name(devl->dev));
if (!lvmetad_pv_gone_by_dev(devl->dev)) {
/* FIXME: use an error path that disables lvmetad */
log_error("Failed to remove %s from lvmetad.", dev_name(devl->dev));
}
if (vgmeta_ret) {
fid = lvmcache_fmt(info)->ops->create_instance(lvmcache_fmt(info), &fic);
if (!(vg_ret = import_vg_from_config_tree(vgmeta_ret, fid))) {
log_error("VG import from config tree failed");
lvmcache_fmt(info)->ops->destroy_instance(fid);
goto out;
}
log_debug_lvmetad("Rescan VG %s updating lvmetad from seqno %u to seqno %u.",
vg->name, vg->seqno, save_seqno);
/*
* If this vg_update fails the cached metadata in
* lvmetad will remain invalid.
* Update lvmetad with the newly read version of the VG.
* When the seqno is unchanged the cached VG can be left.
*/
save_vg->lvmetad_update_pending = 1;
if (!lvmetad_vg_update_finish(save_vg)) {
/* FIXME: use an error path that disables lvmetad */
log_error("Failed to update lvmetad with new VG meta");
if (save_seqno != vg->seqno) {
dm_list_iterate_items(devl, &pvs_scan) {
if (!devl->dev)
continue;
log_debug_lvmetad("Rescan VG %s dropping to replace %s.", vg->name, dev_name(devl->dev));
if (!lvmetad_pv_gone_by_dev(devl->dev))
return_NULL;
}
log_debug_lvmetad("Rescan VG %s updating lvmetad from seqno %u to seqno %u.",
vg->name, vg->seqno, save_seqno);
/*
* If this vg_update fails the cached metadata in
* lvmetad will remain invalid.
*/
vg_ret->lvmetad_update_pending = 1;
if (!lvmetad_vg_update_finish(vg_ret))
log_error("Failed to update lvmetad with new VG meta");
}
dm_config_destroy(vgmeta_ret);
}
out:
if (!save_vg && fid)
fmt->ops->destroy_instance(fid);
if (save_meta)
dm_config_destroy(save_meta);
if (save_vg)
log_debug_lvmetad("Rescan VG %s done (new seqno %u).", save_vg->name, save_vg->seqno);
return save_vg;
if (vg_ret)
log_debug_lvmetad("Rescan VG %s done (seqno %u).", vg_ret->name, vg_ret->seqno);
return vg_ret;
}
int lvmetad_pvscan_single(struct cmd_context *cmd, struct device *dev,
@@ -2243,12 +2073,9 @@ int lvmetad_pvscan_single(struct cmd_context *cmd, struct device *dev,
struct label *label;
struct lvmcache_info *info;
struct _lvmetad_pvscan_baton baton;
const struct format_type *fmt;
/* Create a dummy instance. */
struct format_instance_ctx fic = { .type = 0 };
log_debug_lvmetad("Scan metadata from dev %s", dev_name(dev));
if (!lvmetad_used()) {
log_error("Cannot proceed since lvmetad is not active.");
return 0;
@@ -2259,31 +2086,23 @@ int lvmetad_pvscan_single(struct cmd_context *cmd, struct device *dev,
return 1;
}
if (!(info = lvmcache_info_from_pvid(dev->pvid, dev, 0))) {
log_print_unless_silent("No PV info found on %s for PVID %s.", dev_name(dev), dev->pvid);
if (!label_read(dev, &label, 0)) {
log_print_unless_silent("No PV label found on %s.", dev_name(dev));
if (!lvmetad_pv_gone_by_dev(dev))
goto_bad;
return 1;
}
if (!(label = lvmcache_get_label(info))) {
log_print_unless_silent("No PV label found for %s.", dev_name(dev));
if (!lvmetad_pv_gone_by_dev(dev))
goto_bad;
return 1;
}
info = (struct lvmcache_info *) label->info;
fmt = lvmcache_fmt(info);
baton.cmd = cmd;
baton.vg = NULL;
baton.fid = fmt->ops->create_instance(fmt, &fic);
baton.fid = lvmcache_fmt(info)->ops->create_instance(lvmcache_fmt(info), &fic);
if (!baton.fid)
goto_bad;
if (fmt->features & FMT_OBSOLETE) {
fmt->ops->destroy_instance(baton.fid);
if (baton.fid->fmt->features & FMT_OBSOLETE) {
lvmcache_fmt(info)->ops->destroy_instance(baton.fid);
log_warn("WARNING: Disabling lvmetad cache which does not support obsolete (lvm1) metadata.");
lvmetad_set_disabled(cmd, LVMETAD_DISABLE_REASON_LVM1);
_found_lvm1_metadata = 1;
@@ -2297,9 +2116,9 @@ int lvmetad_pvscan_single(struct cmd_context *cmd, struct device *dev,
lvmcache_foreach_mda(info, _lvmetad_pvscan_single, &baton);
if (!baton.vg)
fmt->ops->destroy_instance(baton.fid);
lvmcache_fmt(info)->ops->destroy_instance(baton.fid);
if (!lvmetad_pv_found(cmd, (const struct id *) &dev->pvid, dev, fmt,
if (!lvmetad_pv_found(cmd, (const struct id *) &dev->pvid, dev, lvmcache_fmt(info),
label->sector, baton.vg, found_vgnames, changed_vgnames)) {
release_vg(baton.vg);
goto_bad;
@@ -2365,13 +2184,6 @@ int lvmetad_pvscan_all_devs(struct cmd_context *cmd, int do_wait)
replacing_other_update = 1;
}
label_scan(cmd);
if (lvmcache_found_duplicate_pvs()) {
log_warn("WARNING: Scan found duplicate PVs.");
return 0;
}
log_verbose("Scanning all devices to update lvmetad.");
if (!(iter = dev_iter_create(cmd->lvmetad_filter, 1))) {
@@ -2742,8 +2554,6 @@ void lvmetad_validate_global_cache(struct cmd_context *cmd, int force)
*/
_lvmetad_get_pv_cache_list(cmd, &pvc_before);
log_debug_lvmetad("Rescan all devices to validate global cache.");
/*
* Update the local lvmetad cache so it correctly reflects any
* changes made on remote hosts. (It's possible that this command
@@ -2812,7 +2622,7 @@ void lvmetad_validate_global_cache(struct cmd_context *cmd, int force)
_update_changed_pvs_in_udev(cmd, &pvc_before, &pvc_after);
}
log_debug_lvmetad("Rescanned all devices");
log_debug_lvmetad("Validating global lvmetad cache finished");
}
int lvmetad_vg_is_foreign(struct cmd_context *cmd, const char *vgname, const char *vgid)
@@ -3064,9 +2874,6 @@ int lvmetad_is_disabled(struct cmd_context *cmd, const char **reason)
} else if (strstr(reply_reason, LVMETAD_DISABLE_REASON_DIRECT)) {
*reason = "the disable flag was set directly";
} else if (strstr(reply_reason, LVMETAD_DISABLE_REASON_REPAIR)) {
*reason = "a repair command was run";
} else if (strstr(reply_reason, LVMETAD_DISABLE_REASON_LVM1)) {
*reason = "LVM1 metadata was found";

View File

@@ -36,38 +36,6 @@ static unsigned _feature_mask;
log_error(t " segment %s of logical volume %s.", ## p, \
dm_config_parent_name(sn), seg->lv->name), 0;
static int _cache_out_line(const char *line, void *_f)
{
log_print(" Setting\t\t%s", line);
return 1;
}
static void _cache_display(const struct lv_segment *seg)
{
const struct dm_config_node *n;
const struct lv_segment *pool_seg =
seg_is_cache_pool(seg) ? seg : first_seg(seg->pool_lv);
log_print(" Chunk size\t\t%s",
display_size(seg->lv->vg->cmd, pool_seg->chunk_size));
if (pool_seg->cache_metadata_format != CACHE_METADATA_FORMAT_UNSELECTED)
log_print(" Metadata format\t%u", pool_seg->cache_metadata_format);
if (pool_seg->cache_mode != CACHE_MODE_UNSELECTED)
log_print(" Mode\t\t%s", get_cache_mode_name(pool_seg));
if (pool_seg->policy_name)
log_print(" Policy\t\t%s", pool_seg->policy_name);
if (pool_seg->policy_settings &&
(n = pool_seg->policy_settings->child))
dm_config_write_node(n, _cache_out_line, NULL);
log_print(" ");
}
/*
* When older metadata are loaded without newer settings,
* set then to default settings (the one that could have been
@@ -84,13 +52,7 @@ static void _fix_missing_defaults(struct lv_segment *cpool_seg)
cpool_seg->policy_name);
}
if (cpool_seg->cache_metadata_format == CACHE_METADATA_FORMAT_UNSELECTED) {
cpool_seg->cache_metadata_format = CACHE_METADATA_FORMAT_1;
log_verbose("Cache pool %s uses implicit metadata format %u.",
display_lvname(cpool_seg->lv), cpool_seg->cache_metadata_format);
}
if (cpool_seg->cache_mode == CACHE_MODE_UNSELECTED) {
if (cpool_seg->cache_mode == CACHE_MODE_UNDEFINED) {
cpool_seg->cache_mode = CACHE_MODE_WHEN_MISSING;
log_verbose("Cache pool %s is missing cache mode, using %s.",
display_lvname(cpool_seg->lv),
@@ -145,16 +107,6 @@ static int _cache_pool_text_import(struct lv_segment *seg,
return SEG_LOG_ERROR("Failed to duplicate policy in");
}
if (dm_config_has_node(sn, "metadata_format")) {
if (!dm_config_get_uint32(sn, "metadata_format", &seg->cache_metadata_format) ||
((seg->cache_metadata_format != CACHE_METADATA_FORMAT_1) &&
(seg->cache_metadata_format != CACHE_METADATA_FORMAT_2)))
return SEG_LOG_ERROR("Unknown cache metadata format %u number in",
seg->cache_metadata_format);
if (seg->cache_metadata_format == CACHE_METADATA_FORMAT_2)
seg->lv->status |= LV_METADATA_FORMAT;
}
/*
* Read in policy args:
* policy_settings {
@@ -212,31 +164,12 @@ static int _cache_pool_text_export(const struct lv_segment *seg,
outf(f, "metadata = \"%s\"", seg->metadata_lv->name);
outf(f, "chunk_size = %" PRIu32, seg->chunk_size);
switch (seg->cache_metadata_format) {
case CACHE_METADATA_FORMAT_UNSELECTED:
/* Unselected format is not printed */
break;
case CACHE_METADATA_FORMAT_1:
/* If format 1 was already specified with cache pool, store it,
* otherwise format gets stored when LV is cached.
* NB: format 1 could be lost anytime, it's a default format.
* Older lvm2 tool can easily drop it.
*/
case CACHE_METADATA_FORMAT_2: /* more in future ? */
outf(f, "metadata_format = " FMTu32, seg->cache_metadata_format);
break;
default:
log_error(INTERNAL_ERROR "LV %s is using unknown cache metadada format %u.",
display_lvname(seg->lv), seg->cache_metadata_format);
return 0;
}
/*
* Cache pool used by a cache LV holds data. Not ideal,
* but not worth to break backward compatibility, by shifting
* content to cache segment
*/
if (seg->cache_mode != CACHE_MODE_UNSELECTED) {
if (seg->cache_mode != CACHE_MODE_UNDEFINED) {
if (!(cache_mode = get_cache_mode_name(seg)))
return_0;
outf(f, "cache_mode = \"%s\"", cache_mode);
@@ -265,39 +198,6 @@ static void _destroy(struct segment_type *segtype)
}
#ifdef DEVMAPPER_SUPPORT
/*
* Parse and look for kernel symbol in /proc/kallsyms
* this could be our only change to figure out there is
* cache policy symbol already in the monolithic kernel
* where 'modprobe dm-cache-smq' will simply not work
*/
static int _lookup_kallsyms(const char *symbol)
{
static const char _syms[] = "/proc/kallsyms";
int ret = 0;
char *line = NULL;
size_t len;
FILE *s;
if (!(s = fopen(_syms, "r")))
log_sys_debug("fopen", _syms);
else {
while (getline(&line, &len, s) != -1)
if (strstr(line, symbol)) {
ret = 1; /* Found symbol */
log_debug("Found kernel symbol%s.", symbol); /* space is in symbol */
break;
}
free(line);
if (fclose(s))
log_sys_debug("fclose", _syms);
}
return ret;
}
static int _target_present(struct cmd_context *cmd,
const struct lv_segment *seg __attribute__((unused)),
unsigned *attributes __attribute__((unused)))
@@ -310,15 +210,13 @@ static int _target_present(struct cmd_context *cmd,
unsigned cache_alias;
const char feature[12];
const char module[12]; /* check dm-%s */
const char ksymbol[12]; /* check for kernel symbol */
const char *aliasing;
} _features[] = {
{ 1, 10, CACHE_FEATURE_METADATA2, 0, "metadata2" },
/* Assumption: cache >=1.9 always aliases MQ policy */
{ 1, 9, CACHE_FEATURE_POLICY_SMQ, CACHE_FEATURE_POLICY_MQ, "policy_smq", "cache-smq",
" smq_exit", " and aliases cache-mq" },
{ 1, 8, CACHE_FEATURE_POLICY_SMQ, 0, "policy_smq", "cache-smq", " smq_exit" },
{ 1, 3, CACHE_FEATURE_POLICY_MQ, 0, "policy_mq", "cache-mq", " mq_init" },
" and aliases cache-mq" },
{ 1, 8, CACHE_FEATURE_POLICY_SMQ, 0, "policy_smq", "cache-smq" },
{ 1, 3, CACHE_FEATURE_POLICY_MQ, 0, "policy_mq", "cache-mq" },
};
static const char _lvmconf[] = "global/cache_disabled_features";
static unsigned _attrs = 0;
@@ -352,20 +250,9 @@ static int _target_present(struct cmd_context *cmd,
for (i = 0; i < DM_ARRAY_SIZE(_features); ++i) {
if (_attrs & _features[i].cache_feature)
continue; /* already present */
if (!_features[i].module[0]) {
if ((maj > _features[i].maj) ||
(maj == _features[i].maj && min >= _features[i].min)) {
log_debug_activation("Cache supports %s.",
_features[i].feature);
_attrs |= _features[i].cache_feature;
}
continue;
}
if (((maj > _features[i].maj) ||
(maj == _features[i].maj && min >= _features[i].min)) &&
((_features[i].ksymbol[0] && _lookup_kallsyms(_features[i].ksymbol)) ||
module_present(cmd, _features[i].module))) {
module_present(cmd, _features[i].module)) {
log_debug_activation("Cache policy %s is available%s.",
_features[i].module,
_features[i].aliasing ? : "");
@@ -423,7 +310,6 @@ static int _modules_needed(struct dm_pool *mem,
#endif /* DEVMAPPER_SUPPORT */
static struct segtype_handler _cache_pool_ops = {
.display = _cache_display,
.text_import = _cache_pool_text_import,
.text_import_area_count = _cache_pool_text_import_area_count,
.text_export = _cache_pool_text_export,
@@ -513,7 +399,6 @@ static int _cache_add_target_line(struct dev_manager *dm,
struct lv_segment *cache_pool_seg;
char *metadata_uuid, *data_uuid, *origin_uuid;
uint64_t feature_flags = 0;
unsigned attr;
if (!seg->pool_lv || !seg_is_cache(seg)) {
log_error(INTERNAL_ERROR "Passed segment is not cache.");
@@ -541,26 +426,6 @@ static int _cache_add_target_line(struct dev_manager *dm,
break;
}
switch (cache_pool_seg->cache_metadata_format) {
case CACHE_METADATA_FORMAT_1: break;
case CACHE_METADATA_FORMAT_2:
if (!_target_present(cmd, NULL, &attr))
return_0;
if (!(attr & CACHE_FEATURE_METADATA2)) {
log_error("LV %s has metadata format %u unsuported by kernel.",
display_lvname(seg->lv), cache_pool_seg->cache_metadata_format);
return 0;
}
feature_flags |= DM_CACHE_FEATURE_METADATA2;
log_debug_activation("Using metadata2 format for %s.", display_lvname(seg->lv));
break;
default:
log_error(INTERNAL_ERROR "LV %s has unknown metadata format %u.",
display_lvname(seg->lv), cache_pool_seg->cache_metadata_format);
return 0;
}
if (!(metadata_uuid = build_dm_uuid(mem, cache_pool_seg->metadata_lv, NULL)))
return_0;
@@ -587,7 +452,6 @@ static int _cache_add_target_line(struct dev_manager *dm,
#endif /* DEVMAPPER_SUPPORT */
static struct segtype_handler _cache_ops = {
.display = _cache_display,
.text_import = _cache_text_import,
.text_import_area_count = _cache_text_import_area_count,
.text_export = _cache_text_export,

View File

@@ -54,7 +54,7 @@
# include <malloc.h>
#endif
static const size_t _linebuffer_size = 4096;
static const size_t linebuffer_size = 4096;
/*
* Copy the input string, removing invalid characters.
@@ -474,12 +474,10 @@ bad:
int process_profilable_config(struct cmd_context *cmd)
{
const char *units;
if (!(cmd->default_settings.unit_factor =
dm_units_to_factor(units = find_config_tree_str(cmd, global_units_CFG, NULL),
dm_units_to_factor(find_config_tree_str(cmd, global_units_CFG, NULL),
&cmd->default_settings.unit_type, 1, NULL))) {
log_error("Unrecognised configuration setting for global/units: %s", units);
log_error("Invalid units specification");
return 0;
}
@@ -542,8 +540,6 @@ static int _process_config(struct cmd_context *cmd)
const struct dm_config_value *cv;
int64_t pv_min_kb;
int udev_disabled = 0;
int scan_size_kb;
int use_scan_cache;
char sysfs_dir[PATH_MAX];
if (!_check_config(cmd))
@@ -627,34 +623,6 @@ static int _process_config(struct cmd_context *cmd)
cmd->default_settings.udev_sync = udev_disabled ? 0 :
find_config_tree_bool(cmd, activation_udev_sync_CFG, NULL);
use_scan_cache = find_config_tree_bool(cmd, devices_scan_cache_CFG, NULL);
#ifdef AIO_SUPPORT
cmd->use_aio = find_config_tree_bool(cmd, devices_scan_async_CFG, NULL);
if (cmd->use_aio) {
cmd->use_scan_cache = 1;
if (!use_scan_cache)
log_warn("WARNING: Ignoring scan_cache setting because of scan_async.");
} else
cmd->use_scan_cache = use_scan_cache;
#else
cmd->use_scan_cache = use_scan_cache;
cmd->use_aio = 0;
if (find_config_tree_bool(cmd, devices_scan_async_CFG, NULL))
log_warn("WARNING: Ignoring scan_async setting, no async I/O support.");
#endif
scan_size_kb = find_config_tree_int(cmd, devices_scan_size_CFG, NULL);
if (!scan_size_kb || (scan_size_kb < 0) || (scan_size_kb % 4)) {
log_warn("WARNING: Ignoring invalid scan_size %d KB, using default %u KB.",
scan_size_kb, DEFAULT_SCAN_SIZE_KB);
log_warn("scan_size has units of KB and must be a multiple of 4 KB.");
scan_size_kb = DEFAULT_SCAN_SIZE_KB;
}
cmd->default_settings.scan_size_kb = scan_size_kb;
/*
* Set udev_fallback lazily on first use since it requires
* checking DM driver version which is an extra ioctl!
@@ -715,6 +683,9 @@ static int _process_config(struct cmd_context *cmd)
if (find_config_tree_bool(cmd, report_two_word_unknown_device_CFG, NULL))
init_unknown_device_name("unknown device");
init_detect_internal_vg_cache_corruption
(find_config_tree_bool(cmd, global_detect_internal_vg_cache_corruption_CFG, NULL));
if (!_init_system_id(cmd))
return_0;
@@ -1528,6 +1499,11 @@ static int _init_segtypes(struct cmd_context *cmd)
dm_list_add(&cmd->segtypes, &segtype->list);
}
#ifdef REPLICATOR_INTERNAL
if (!init_replicator_segtype(cmd, &seglib))
return 0;
#endif
#ifdef RAID_INTERNAL
if (!init_raid_segtypes(cmd, &seglib))
return 0;
@@ -1900,7 +1876,7 @@ struct cmd_context *create_toolcontext(unsigned is_long_lived,
/* Set in/out stream buffering before glibc */
if (set_buffering) {
/* Allocate 2 buffers */
if (!(cmd->linebuffer = dm_malloc(2 * _linebuffer_size))) {
if (!(cmd->linebuffer = dm_malloc(2 * linebuffer_size))) {
log_error("Failed to allocate line buffer.");
goto out;
}
@@ -1911,7 +1887,7 @@ struct cmd_context *create_toolcontext(unsigned is_long_lived,
(flags & O_ACCMODE) != O_WRONLY) {
if (!reopen_standard_stream(&stdin, "r"))
goto_out;
if (setvbuf(stdin, cmd->linebuffer, _IOLBF, _linebuffer_size)) {
if (setvbuf(stdin, cmd->linebuffer, _IOLBF, linebuffer_size)) {
log_sys_error("setvbuf", "");
goto out;
}
@@ -1922,8 +1898,8 @@ struct cmd_context *create_toolcontext(unsigned is_long_lived,
(flags & O_ACCMODE) != O_RDONLY) {
if (!reopen_standard_stream(&stdout, "w"))
goto_out;
if (setvbuf(stdout, cmd->linebuffer + _linebuffer_size,
_IOLBF, _linebuffer_size)) {
if (setvbuf(stdout, cmd->linebuffer + linebuffer_size,
_IOLBF, linebuffer_size)) {
log_sys_error("setvbuf", "");
goto out;
}
@@ -2023,6 +1999,7 @@ struct cmd_context *create_toolcontext(unsigned is_long_lived,
if (set_filters && !init_filters(cmd, 1))
goto_out;
cmd->default_settings.cache_vgmetadata = 1;
cmd->current_settings = cmd->default_settings;
cmd->initialized.config = 1;
@@ -2032,6 +2009,7 @@ out:
cmd = NULL;
}
return cmd;
}
@@ -2252,8 +2230,6 @@ void destroy_toolcontext(struct cmd_context *cmd)
!cmd->filter->dump(cmd->filter, 1))
stack;
label_scan_destroy(cmd);
archive_exit(cmd);
backup_exit(cmd);
lvmcache_destroy(cmd, 0, 0);

View File

@@ -39,7 +39,7 @@ struct config_info {
int udev_rules;
int udev_sync;
int udev_fallback;
int scan_size_kb;
int cache_vgmetadata;
const char *msg_prefix;
const char *fmt_name;
uint64_t unit_factor;
@@ -160,13 +160,9 @@ struct cmd_context {
unsigned lockd_vg_rescan:1;
unsigned lockd_vg_default_sh:1;
unsigned lockd_vg_enforce_sh:1;
unsigned lockd_lv_sh:1;
unsigned vg_notify:1;
unsigned lv_notify:1;
unsigned pv_notify:1;
unsigned use_scan_cache:1;
unsigned use_aio:1;
unsigned pvscan_cache_single:1;
/*
* Filtering.

View File

@@ -65,11 +65,11 @@ struct config_source {
* Map each ID to respective definition of the configuration item.
*/
static struct cfg_def_item _cfg_def_items[CFG_COUNT + 1] = {
#define cfg_section(id, name, parent, flags, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_SECTION, {0}, (flags), since_version, {0}, deprecated_since_version, deprecation_comment, comment},
#define cfg(id, name, parent, flags, type, default_value, since_version, unconfigured_value, deprecated_since_version, deprecation_comment, comment) {id, parent, name, type, {.v_##type = (default_value)}, (flags), since_version, {.v_UNCONFIGURED = (unconfigured_value)}, deprecated_since_version, deprecation_comment, comment},
#define cfg_runtime(id, name, parent, flags, type, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, type, {.fn_##type = get_default_##id}, (flags) | CFG_DEFAULT_RUN_TIME, since_version, {.fn_UNCONFIGURED = get_default_unconfigured_##id}, deprecated_since_version, (deprecation_comment), comment},
#define cfg_array(id, name, parent, flags, types, default_value, since_version, unconfigured_value, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_ARRAY | (types), {.v_CFG_TYPE_STRING = (default_value)}, (flags), (since_version), {.v_UNCONFIGURED = (unconfigured_value)}, deprecated_since_version, deprecation_comment, comment},
#define cfg_array_runtime(id, name, parent, flags, types, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_ARRAY | (types), {.fn_CFG_TYPE_STRING = get_default_##id}, (flags) | CFG_DEFAULT_RUN_TIME, (since_version), {.fn_UNCONFIGURED = get_default_unconfigured_##id}, deprecated_since_version, deprecation_comment, comment},
#define cfg_section(id, name, parent, flags, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_SECTION, {0}, flags, since_version, {0}, deprecated_since_version, deprecation_comment, comment},
#define cfg(id, name, parent, flags, type, default_value, since_version, unconfigured_value, deprecated_since_version, deprecation_comment, comment) {id, parent, name, type, {.v_##type = default_value}, flags, since_version, {.v_UNCONFIGURED = unconfigured_value}, deprecated_since_version, deprecation_comment, comment},
#define cfg_runtime(id, name, parent, flags, type, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, type, {.fn_##type = get_default_##id}, flags | CFG_DEFAULT_RUN_TIME, since_version, {.fn_UNCONFIGURED = get_default_unconfigured_##id}, deprecated_since_version, deprecation_comment, comment},
#define cfg_array(id, name, parent, flags, types, default_value, since_version, unconfigured_value, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_ARRAY | types, {.v_CFG_TYPE_STRING = default_value}, flags, since_version, {.v_UNCONFIGURED = unconfigured_value}, deprecated_since_version, deprecation_comment, comment},
#define cfg_array_runtime(id, name, parent, flags, types, since_version, deprecated_since_version, deprecation_comment, comment) {id, parent, name, CFG_TYPE_ARRAY | types, {.fn_CFG_TYPE_STRING = get_default_##id}, flags | CFG_DEFAULT_RUN_TIME, since_version, {.fn_UNCONFIGURED = get_default_unconfigured_##id}, deprecated_since_version, deprecation_comment, comment},
#include "config_settings.h"
#undef cfg_section
#undef cfg
@@ -481,8 +481,7 @@ int override_config_tree_from_profile(struct cmd_context *cmd,
if (profile->source == CONFIG_PROFILE_COMMAND)
return _override_config_tree_from_command_profile(cmd, profile);
if (profile->source == CONFIG_PROFILE_METADATA)
else if (profile->source == CONFIG_PROFILE_METADATA)
return _override_config_tree_from_metadata_profile(cmd, profile);
log_error(INTERNAL_ERROR "override_config_tree_from_profile: incorrect profile source type");
@@ -494,7 +493,7 @@ int override_config_tree_from_profile(struct cmd_context *cmd,
* and function avoids parsing of mda into config tree which
* remains unmodified and should not be used.
*/
int config_file_read_fd(struct dm_config_tree *cft, struct device *dev, char *buf_async,
int config_file_read_fd(struct dm_config_tree *cft, struct device *dev,
off_t offset, size_t size, off_t offset2, size_t size2,
checksum_fn_t checksum_fn, uint32_t checksum,
int checksum_only, int no_dup_node_check)
@@ -517,18 +516,7 @@ int config_file_read_fd(struct dm_config_tree *cft, struct device *dev, char *bu
if (!(dev->flags & DEV_REGULAR) || size2)
use_mmap = 0;
if (buf_async) {
if (!(buf = dm_malloc(size + size2))) {
log_error("Failed to allocate circular buffer.");
return 0;
}
memcpy(buf, buf_async + offset, size);
if (size2)
memcpy(buf + size, buf_async + offset2, size2);
fb = buf;
} else if (use_mmap) {
if (use_mmap) {
mmap_offset = offset % lvm_getpagesize();
/* memory map the file */
fb = mmap((caddr_t) 0, size + mmap_offset, PROT_READ,
@@ -543,7 +531,6 @@ int config_file_read_fd(struct dm_config_tree *cft, struct device *dev, char *bu
log_error("Failed to allocate circular buffer.");
return 0;
}
if (!dev_read_circular(dev, (uint64_t) offset, size,
(uint64_t) offset2, size2, buf)) {
goto out;
@@ -554,7 +541,7 @@ int config_file_read_fd(struct dm_config_tree *cft, struct device *dev, char *bu
if (checksum_fn && checksum !=
(checksum_fn(checksum_fn(INITIAL_CRC, (const uint8_t *)fb, size),
(const uint8_t *)(fb + size), size2))) {
log_error("%s: Checksum error at offset %" PRIu64, dev_name(dev), (uint64_t) offset);
log_error("%s: Checksum error", dev_name(dev));
goto out;
}
@@ -613,7 +600,7 @@ int config_file_read(struct dm_config_tree *cft)
}
}
r = config_file_read_fd(cft, cf->dev, NULL, 0, (size_t) info.st_size, 0, 0,
r = config_file_read_fd(cft, cf->dev, 0, (size_t) info.st_size, 0, 0,
(checksum_fn_t) NULL, 0, 0, 0);
if (!cf->keep_open) {
@@ -632,9 +619,9 @@ struct timespec config_file_timestamp(struct dm_config_tree *cft)
}
#define cfg_def_get_item_p(id) (&_cfg_def_items[id])
#define cfg_def_get_default_unconfigured_value_hint(cmd,item) (((item)->flags & CFG_DEFAULT_RUN_TIME) ? (item)->default_unconfigured_value.fn_UNCONFIGURED(cmd) : (item)->default_unconfigured_value.v_UNCONFIGURED)
#define cfg_def_get_default_value_hint(cmd,item,type,profile) (((item)->flags & CFG_DEFAULT_RUN_TIME) ? (item)->default_value.fn_##type(cmd,profile) : (item)->default_value.v_##type)
#define cfg_def_get_default_value(cmd,item,type,profile) ((item)->flags & CFG_DEFAULT_UNDEFINED ? 0 : cfg_def_get_default_value_hint(cmd,item,type,profile))
#define cfg_def_get_default_unconfigured_value_hint(cmd,item) ((item->flags & CFG_DEFAULT_RUN_TIME) ? item->default_unconfigured_value.fn_UNCONFIGURED(cmd) : item->default_unconfigured_value.v_UNCONFIGURED)
#define cfg_def_get_default_value_hint(cmd,item,type,profile) ((item->flags & CFG_DEFAULT_RUN_TIME) ? item->default_value.fn_##type(cmd,profile) : item->default_value.v_##type)
#define cfg_def_get_default_value(cmd,item,type,profile) (item->flags & CFG_DEFAULT_UNDEFINED ? 0 : cfg_def_get_default_value_hint(cmd,item,type,profile))
static int _cfg_def_make_path(char *buf, size_t buf_size, int id, cfg_def_item_t *item, int xlate)
{
@@ -755,16 +742,14 @@ static struct dm_config_value *_get_def_array_values(struct cmd_context *cmd,
switch (toupper(token[0])) {
case 'I':
case 'B':
errno = 0;
v->v.i = strtoll(token + 1, &r, 10);
if (errno || *r)
if (*r)
goto bad;
v->type = DM_CFG_INT;
break;
case 'F':
errno = 0;
v->v.f = strtod(token + 1, &r);
if (errno || *r)
if (*r)
goto bad;
v->type = DM_CFG_FLOAT;
break;
@@ -1892,12 +1877,6 @@ int config_write(struct dm_config_tree *cft,
}
log_verbose("Dumping configuration to %s", file);
if (tree_spec->withgeneralpreamble)
fprintf(baton.fp, CFG_PREAMBLE_GENERAL);
if (tree_spec->withlocalpreamble)
fprintf(baton.fp, CFG_PREAMBLE_LOCAL);
if (!argc) {
if (!dm_config_write_node_out(cft->root, &_out_spec, &baton)) {
log_error("Failure while writing to %s", file);
@@ -2463,13 +2442,21 @@ const char *get_default_activation_mirror_image_fault_policy_CFG(struct cmd_cont
int get_default_allocation_thin_pool_chunk_size_CFG(struct cmd_context *cmd, struct profile *profile)
{
const char *str;
uint32_t chunk_size;
int chunk_size_calc_method;
if (!get_default_allocation_thin_pool_chunk_size(cmd, profile, &chunk_size,
&chunk_size_calc_method)) {
stack; /* Ignore this error, never happens... */
if (!(str = find_config_tree_str(cmd, allocation_thin_pool_chunk_size_policy_CFG, profile))) {
log_error(INTERNAL_ERROR "Cannot find configuration.");
return 0;
}
if (!strcasecmp(str, "generic"))
chunk_size = DEFAULT_THIN_POOL_CHUNK_SIZE * 2;
else if (!strcasecmp(str, "performance"))
chunk_size = DEFAULT_THIN_POOL_CHUNK_SIZE_PERFORMANCE * 2;
else {
log_error("Thin pool chunk size calculation policy \"%s\" is unrecognised.", str);
return 0;
}
return (int) chunk_size;

View File

@@ -141,7 +141,6 @@ typedef struct cfg_def_item {
uint16_t deprecated_since_version; /* version since this item is deprecated */
const char *deprecation_comment; /* comment about reasons for deprecation and settings that supersede this one */
const char *comment; /* comment */
const char *file_premable; /* comment text to use at the start of the file */
} cfg_def_item_t;
/* configuration definition tree types */
@@ -174,8 +173,6 @@ struct config_def_tree_spec {
unsigned withversions:1; /* include versions */
unsigned withspaces:1; /* add more spaces in output for better readability */
unsigned unconfigured:1; /* use unconfigured path strings */
unsigned withgeneralpreamble:1; /* include preamble for a general config file */
unsigned withlocalpreamble:1; /* include preamble for a local config file */
uint8_t *check_status; /* status of last tree check (currently needed for CFG_DEF_TREE_MISSING only) */
};
@@ -239,7 +236,7 @@ config_source_t config_get_source_type(struct dm_config_tree *cft);
typedef uint32_t (*checksum_fn_t) (uint32_t initial, const uint8_t *buf, uint32_t size);
struct dm_config_tree *config_open(config_source_t source, const char *filename, int keep_open);
int config_file_read_fd(struct dm_config_tree *cft, struct device *dev, char *buf_async,
int config_file_read_fd(struct dm_config_tree *cft, struct device *dev,
off_t offset, size_t size, off_t offset2, size_t size2,
checksum_fn_t checksum_fn, uint32_t checksum,
int skip_parse, int no_dup_node_check);

View File

@@ -121,30 +121,6 @@
cfg_section(root_CFG_SECTION, "(root)", root_CFG_SECTION, 0, vsn(0, 0, 0), 0, NULL, NULL)
#define CFG_PREAMBLE_GENERAL \
"# This is an example configuration file for the LVM2 system.\n" \
"# It contains the default settings that would be used if there was no\n" \
"# @DEFAULT_SYS_DIR@/lvm.conf file.\n" \
"#\n" \
"# Refer to 'man lvm.conf' for further information including the file layout.\n" \
"#\n" \
"# Refer to 'man lvm.conf' for information about how settings configured in\n" \
"# this file are combined with built-in values and command line options to\n" \
"# arrive at the final values used by LVM.\n" \
"#\n" \
"# Refer to 'man lvmconfig' for information about displaying the built-in\n" \
"# and configured values used by LVM.\n" \
"#\n" \
"# If a default value is set in this file (not commented out), then a\n" \
"# new version of LVM using this file will continue using that value,\n" \
"# even if the new version of LVM changes the built-in default value.\n" \
"#\n" \
"# To put this file in a different directory and override @DEFAULT_SYS_DIR@ set\n" \
"# the environment variable LVM_SYSTEM_DIR before running the tools.\n" \
"#\n" \
"# N.B. Take care that each setting only appears once if uncommenting\n" \
"# example settings in this file.\n\n"
cfg_section(config_CFG_SECTION, "config", root_CFG_SECTION, 0, vsn(2, 2, 99), 0, NULL,
"How LVM configuration settings are handled.\n")
@@ -185,26 +161,6 @@ cfg_section(tags_CFG_SECTION, "tags", root_CFG_SECTION, CFG_DEFAULT_COMMENTED, v
cfg_section(local_CFG_SECTION, "local", root_CFG_SECTION, 0, vsn(2, 2, 117), 0, NULL,
"LVM settings that are specific to the local host.\n")
#define CFG_PREAMBLE_LOCAL \
"# This is a local configuration file template for the LVM2 system\n" \
"# which should be installed as @DEFAULT_SYS_DIR@/lvmlocal.conf .\n" \
"#\n" \
"# Refer to 'man lvm.conf' for information about the file layout.\n" \
"#\n" \
"# To put this file in a different directory and override\n" \
"# @DEFAULT_SYS_DIR@ set the environment variable LVM_SYSTEM_DIR before\n" \
"# running the tools.\n" \
"#\n" \
"# The lvmlocal.conf file is normally expected to contain only the\n" \
"# \"local\" section which contains settings that should not be shared or\n" \
"# repeated among different hosts. (But if other sections are present,\n" \
"# they *will* get processed. Settings in this file override equivalent\n" \
"# ones in lvm.conf and are in turn overridden by ones in any enabled\n" \
"# lvm_<tag>.conf files.)\n" \
"#\n" \
"# Please take care that each setting only appears once if uncommenting\n" \
"# example settings in this file and never copy this file between hosts.\n\n"
cfg(config_checks_CFG, "checks", config_CFG_SECTION, 0, CFG_TYPE_BOOL, 1, vsn(2, 2, 99), NULL, 0, NULL,
"If enabled, any LVM configuration mismatch is reported.\n"
"This implies checking that the configuration key is understood by\n"
@@ -457,30 +413,6 @@ cfg(devices_allow_changes_with_duplicate_pvs_CFG, "allow_changes_with_duplicate_
"Enabling this setting allows the VG to be used as usual even with\n"
"uncertain devices.\n")
cfg(devices_scan_async_CFG, "scan_async", devices_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_BOOL, DEFAULT_SCAN_ASYNC, vsn(2, 2, 176), NULL, 0, NULL,
"Use async I/O to read headers and metadata from disks in parallel.\n")
cfg(devices_scan_cache_CFG, "scan_cache", devices_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_BOOL, DEFAULT_SCAN_CACHE, vsn(2, 2, 176), NULL, 0, NULL,
"Cache the disk blocks that a command scans from each PV.\n"
"When scanning disks for LVM headers and metadata, scan_size KiB\n"
"are initially read from each disk. These scanned blocks can be\n"
"cached by the command to avoid rereading them from disk.\n"
"The scan_cache is required when scan_async is enabled.\n")
cfg(devices_scan_size_CFG, "scan_size", devices_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_INT, DEFAULT_SCAN_SIZE_KB, vsn(2, 2, 176), NULL, 0, NULL,
"Number of KiB to read from each disk when scanning disks.\n"
"The initial scan size is intended to cover all the headers\n"
"and metadata that LVM places at the start of each disk so\n"
"that a single read operation can retrieve them all.\n"
"Any headers or metadata that lie beyond this size require\n"
"an additional disk read. Must be a multiple of 4KiB.\n")
cfg(devices_async_events_CFG, "async_events", devices_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_INT, DEFAULT_ASYNC_EVENTS, vsn(2, 2, 176), NULL, 0, NULL,
"Max number of concurrent async reads when scanning disks.\n"
"Up to this many disks can be read concurrently when scanning\n"
"disks with async I/O. This setting may be limited by the system\n"
"aio configuration. This should not exceed the open files limit.\n")
cfg_array(allocation_cling_tag_list_CFG, "cling_tag_list", allocation_CFG_SECTION, CFG_DEFAULT_UNDEFINED, CFG_TYPE_STRING, NULL, vsn(2, 2, 77), NULL, 0, NULL,
"Advise LVM which PVs to use when searching for new space.\n"
"When searching for free space to extend an LV, the 'cling' allocation\n"
@@ -537,28 +469,18 @@ cfg(allocation_mirror_logs_require_separate_pvs_CFG, "mirror_logs_require_separa
cfg(allocation_raid_stripe_all_devices_CFG, "raid_stripe_all_devices", allocation_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_BOOL, DEFAULT_ALLOCATION_STRIPE_ALL_DEVICES, vsn(2, 2, 162), NULL, 0, NULL,
"Stripe across all PVs when RAID stripes are not specified.\n"
"If enabled, all PVs in the VG or on the command line are used for\n"
"raid0/4/5/6/10 when the command does not specify the number of\n"
"stripes to use.\n"
"If enabled, all PVs in the VG or on the command line are used for raid0/4/5/6/10\n"
"when the command does not specify the number of stripes to use.\n"
"This was the default behaviour until release 2.02.162.\n")
cfg(allocation_cache_pool_metadata_require_separate_pvs_CFG, "cache_pool_metadata_require_separate_pvs", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA, CFG_TYPE_BOOL, DEFAULT_CACHE_POOL_METADATA_REQUIRE_SEPARATE_PVS, vsn(2, 2, 106), NULL, 0, NULL,
cfg(allocation_cache_pool_metadata_require_separate_pvs_CFG, "cache_pool_metadata_require_separate_pvs", allocation_CFG_SECTION, 0, CFG_TYPE_BOOL, DEFAULT_CACHE_POOL_METADATA_REQUIRE_SEPARATE_PVS, vsn(2, 2, 106), NULL, 0, NULL,
"Cache pool metadata and data will always use different PVs.\n")
cfg(allocation_cache_pool_cachemode_CFG, "cache_pool_cachemode", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_CACHE_MODE, vsn(2, 2, 113), NULL, vsn(2, 2, 128),
cfg(allocation_cache_pool_cachemode_CFG, "cache_pool_cachemode", allocation_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_CACHE_MODE, vsn(2, 2, 113), NULL, vsn(2, 2, 128),
"This has been replaced by the allocation/cache_mode setting.\n",
"Cache mode.\n")
cfg(allocation_cache_metadata_format_CFG, "cache_metadata_format", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_COMMENTED, CFG_TYPE_INT, DEFAULT_CACHE_METADATA_FORMAT, vsn(2, 2, 169), NULL, 0, NULL,
"Sets default metadata format for new cache.\n"
"#\n"
"Accepted values:\n"
" 0 Automatically detected best available format\n"
" 1 Original format\n"
" 2 Improved 2nd. generation format\n"
"#\n")
cfg(allocation_cache_mode_CFG, "cache_mode", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_CACHE_MODE, vsn(2, 2, 128), NULL, 0, NULL,
cfg(allocation_cache_mode_CFG, "cache_mode", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_CACHE_MODE, vsn(2, 2, 128), NULL, 0, NULL,
"The default cache mode used for new cache.\n"
"#\n"
"Accepted values:\n"
@@ -570,20 +492,20 @@ cfg(allocation_cache_mode_CFG, "cache_mode", allocation_CFG_SECTION, CFG_PROFILA
"#\n"
"This setting replaces allocation/cache_pool_cachemode.\n")
cfg(allocation_cache_policy_CFG, "cache_policy", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_UNDEFINED, CFG_TYPE_STRING, 0, vsn(2, 2, 128), NULL, 0, NULL,
cfg(allocation_cache_policy_CFG, "cache_policy", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_DEFAULT_UNDEFINED, CFG_TYPE_STRING, 0, vsn(2, 2, 128), NULL, 0, NULL,
"The default cache policy used for new cache volume.\n"
"Since kernel 4.2 the default policy is smq (Stochastic multiqueue),\n"
"Since kernel 4.2 the default policy is smq (Stochastic multique),\n"
"otherwise the older mq (Multiqueue) policy is selected.\n")
cfg_section(allocation_cache_settings_CFG_SECTION, "cache_settings", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_COMMENTED, vsn(2, 2, 128), 0, NULL,
cfg_section(allocation_cache_settings_CFG_SECTION, "cache_settings", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_DEFAULT_COMMENTED, vsn(2, 2, 128), 0, NULL,
"Settings for the cache policy.\n"
"See documentation for individual cache policies for more info.\n")
cfg_section(policy_settings_CFG_SUBSECTION, "policy_settings", allocation_cache_settings_CFG_SECTION, CFG_NAME_VARIABLE | CFG_SECTION_NO_CHECK | CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_COMMENTED, vsn(2, 2, 128), 0, NULL,
cfg_section(policy_settings_CFG_SUBSECTION, "policy_settings", allocation_cache_settings_CFG_SECTION, CFG_NAME_VARIABLE | CFG_SECTION_NO_CHECK | CFG_PROFILABLE | CFG_DEFAULT_COMMENTED, vsn(2, 2, 128), 0, NULL,
"Replace this subsection name with a policy name.\n"
"Multiple subsections for different policies can be created.\n")
cfg_runtime(allocation_cache_pool_chunk_size_CFG, "cache_pool_chunk_size", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_UNDEFINED, CFG_TYPE_INT, vsn(2, 2, 106), 0, NULL,
cfg_runtime(allocation_cache_pool_chunk_size_CFG, "cache_pool_chunk_size", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_DEFAULT_UNDEFINED, CFG_TYPE_INT, vsn(2, 2, 106), 0, NULL,
"The minimal chunk size in KiB for cache pool volumes.\n"
"Using a chunk_size that is too large can result in wasteful use of\n"
"the cache, where small reads and writes can cause large sections of\n"
@@ -594,7 +516,7 @@ cfg_runtime(allocation_cache_pool_chunk_size_CFG, "cache_pool_chunk_size", alloc
"on the smaller end of the spectrum. Supported values range from\n"
"32KiB to 1GiB in multiples of 32.\n")
cfg(allocation_cache_pool_max_chunks_CFG, "cache_pool_max_chunks", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_PROFILABLE_METADATA | CFG_DEFAULT_UNDEFINED, CFG_TYPE_INT, 0, vsn(2, 2, 165), NULL, 0, NULL,
cfg(allocation_cache_pool_max_chunks_CFG, "cache_pool_max_chunks", allocation_CFG_SECTION, CFG_PROFILABLE | CFG_DEFAULT_UNDEFINED, CFG_TYPE_INT, 0, vsn(2, 2, 165), NULL, 0, NULL,
"The maximum number of chunks in a cache pool.\n"
"For cache target v1.9 the recommended maximumm is 1000000 chunks.\n"
"Using cache pool with more chunks may degrade cache performance.\n")
@@ -892,8 +814,11 @@ cfg(global_abort_on_internal_errors_CFG, "abort_on_internal_errors", global_CFG_
"Treat any internal errors as fatal errors, aborting the process that\n"
"encountered the internal error. Please only enable for debugging.\n")
cfg(global_detect_internal_vg_cache_corruption_CFG, "detect_internal_vg_cache_corruption", global_CFG_SECTION, 0, CFG_TYPE_BOOL, 0, vsn(2, 2, 96), NULL, vsn(2, 2, 174), NULL,
"No longer used.\n")
cfg(global_detect_internal_vg_cache_corruption_CFG, "detect_internal_vg_cache_corruption", global_CFG_SECTION, 0, CFG_TYPE_BOOL, DEFAULT_DETECT_INTERNAL_VG_CACHE_CORRUPTION, vsn(2, 2, 96), NULL, 0, NULL,
"Internal verification of VG structures.\n"
"Check if CRC matches when a parsed VG is used multiple times. This\n"
"is useful to catch unexpected changes to cached VG structures.\n"
"Please only enable for debugging.\n")
cfg(global_metadata_read_only_CFG, "metadata_read_only", global_CFG_SECTION, 0, CFG_TYPE_BOOL, DEFAULT_METADATA_READ_ONLY, vsn(2, 2, 75), NULL, 0, NULL,
"No operations that change on-disk metadata are permitted.\n"
@@ -1000,7 +925,7 @@ cfg(global_use_lvmetad_CFG, "use_lvmetad", global_CFG_SECTION, 0, CFG_TYPE_BOOL,
"devices/global_filter.\n")
cfg(global_lvmetad_update_wait_time_CFG, "lvmetad_update_wait_time", global_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_INT, DEFAULT_LVMETAD_UPDATE_WAIT_TIME, vsn(2, 2, 151), NULL, 0, NULL,
"Number of seconds a command will wait for lvmetad update to finish.\n"
"The number of seconds a command will wait for lvmetad update to finish.\n"
"After waiting for this period, a command will not use lvmetad, and\n"
"will revert to disk scanning.\n")
@@ -1067,7 +992,7 @@ cfg_array(global_thin_disabled_features_CFG, "thin_disabled_features", global_CF
cfg_array(global_cache_disabled_features_CFG, "cache_disabled_features", global_CFG_SECTION, CFG_ALLOW_EMPTY | CFG_DEFAULT_UNDEFINED, CFG_TYPE_STRING, NULL, vsn(2, 2, 128), NULL, 0, NULL,
"Features to not use in the cache driver.\n"
"This can be helpful for testing, or to avoid using a feature that is\n"
"causing problems. Features include: policy_mq, policy_smq, metadata2.\n"
"causing problems. Features include: policy_mq, policy_smq.\n"
"#\n"
"Example\n"
"cache_disabled_features = [ \"policy_smq\" ]\n"
@@ -1101,10 +1026,6 @@ cfg_array(global_cache_check_options_CFG, "cache_check_options", global_CFG_SECT
cfg_array(global_cache_repair_options_CFG, "cache_repair_options", global_CFG_SECTION, CFG_ALLOW_EMPTY | CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_CACHE_REPAIR_OPTIONS_CONFIG, vsn(2, 2, 108), NULL, 0, NULL,
"List of options passed to the cache_repair command.\n")
cfg(global_fsadm_executable_CFG, "fsadm_executable", global_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_FSADM_PATH, vsn(2, 2, 170), "@FSADM_PATH@", 0, NULL,
"The full path to the fsadm command.\n"
"LVM uses this command to help with lvresize -r operations.\n")
cfg(global_system_id_source_CFG, "system_id_source", global_CFG_SECTION, 0, CFG_TYPE_STRING, DEFAULT_SYSTEM_ID_SOURCE, vsn(2, 2, 117), NULL, 0, NULL,
"The method LVM uses to set the local system ID.\n"
"Volume Groups can also be given a system ID (by vgcreate, vgchange,\n"
@@ -1943,9 +1864,7 @@ cfg(dmeventd_thin_command_CFG, "thin_command", dmeventd_CFG_SECTION, CFG_DEFAULT
"The plugin runs command with each 5% increment when thin-pool data volume\n"
"or metadata volume gets above 50%.\n"
"Command which starts with 'lvm ' prefix is internal lvm command.\n"
"You can write your own handler to customise behaviour in more details.\n"
"User handler is specified with the full path starting with '/'.\n")
/* TODO: systemd service handler */
"You can write your own handler to customise behaviour in more details.\n")
cfg(dmeventd_executable_CFG, "executable", dmeventd_CFG_SECTION, CFG_DEFAULT_COMMENTED, CFG_TYPE_STRING, DEFAULT_DMEVENTD_PATH, vsn(2, 2, 73), "@DMEVENTD_PATH@", 0, NULL,
"The full path to the dmeventd binary.\n")

View File

@@ -71,7 +71,7 @@
* FIXME: Increase these to 64 and further to the MD maximum
* once the SubLVs split and name shift got enhanced
*/
#define DEFAULT_RAID1_MAX_IMAGES 64
#define DEFAULT_RAID1_MAX_IMAGES 10
#define DEFAULT_RAID_MAX_IMAGES 64
#define DEFAULT_ALLOCATION_STRIPE_ALL_DEVICES 0 /* Don't stripe across all devices if not -i/--stripes given */
@@ -104,9 +104,9 @@
#define DEFAULT_THIN_REPAIR_OPTION1 ""
#define DEFAULT_THIN_REPAIR_OPTIONS_CONFIG "#S" DEFAULT_THIN_REPAIR_OPTION1
#define DEFAULT_THIN_POOL_METADATA_REQUIRE_SEPARATE_PVS 0
#define DEFAULT_THIN_POOL_MAX_METADATA_SIZE (DM_THIN_MAX_METADATA_SIZE / 2) /* KB */
#define DEFAULT_THIN_POOL_MAX_METADATA_SIZE (16 * 1024 * 1024) /* KB */
#define DEFAULT_THIN_POOL_MIN_METADATA_SIZE 2048 /* KB */
#define DEFAULT_THIN_POOL_OPTIMAL_METADATA_SIZE (128 * 1024) /* KB */
#define DEFAULT_THIN_POOL_OPTIMAL_SIZE (128 * 1024 * 1024) /* KB */
#define DEFAULT_THIN_POOL_CHUNK_SIZE_POLICY "generic"
#define DEFAULT_THIN_POOL_CHUNK_SIZE 64 /* KB */
#define DEFAULT_THIN_POOL_CHUNK_SIZE_PERFORMANCE 512 /* KB */
@@ -132,11 +132,8 @@
#define DEFAULT_CACHE_POOL_MIN_METADATA_SIZE 2048 /* KB */
#define DEFAULT_CACHE_POOL_MAX_METADATA_SIZE (16 * 1024 * 1024) /* KB */
#define DEFAULT_CACHE_POLICY "mq"
#define DEFAULT_CACHE_METADATA_FORMAT CACHE_METADATA_FORMAT_UNSELECTED /* Autodetect */
#define DEFAULT_CACHE_MODE "writethrough"
#define DEFAULT_FSADM_PATH FSADM_PATH
#define DEFAULT_UMASK 0077
#define DEFAULT_FORMAT "lvm2"
@@ -179,6 +176,7 @@
#define DEFAULT_LOGLEVEL 0
#define DEFAULT_INDENT 1
#define DEFAULT_ABORT_ON_INTERNAL_ERRORS 0
#define DEFAULT_DETECT_INTERNAL_VG_CACHE_CORRUPTION 0
#define DEFAULT_UNITS "r"
#define DEFAULT_SUFFIX 1
#define DEFAULT_HOSTTAGS 0
@@ -201,7 +199,7 @@
#define DEFAULT_ACTIVATION_MODE "degraded"
#define DEFAULT_USE_LINEAR_TARGET 1
#define DEFAULT_STRIPE_FILLER "error"
#define DEFAULT_RAID_REGION_SIZE 2048 /* KB */
#define DEFAULT_RAID_REGION_SIZE 512 /* KB */
#define DEFAULT_INTERVAL 15
#define DEFAULT_MAX_HISTORY 100
@@ -266,9 +264,4 @@
#define DEFAULT_THIN_POOL_AUTOEXTEND_THRESHOLD 100
#define DEFAULT_THIN_POOL_AUTOEXTEND_PERCENT 20
#define DEFAULT_ASYNC_EVENTS 100
#define DEFAULT_SCAN_ASYNC 1
#define DEFAULT_SCAN_CACHE 1
#define DEFAULT_SCAN_SIZE_KB 128
#endif /* _LVM_DEFAULTS_H */

View File

@@ -320,8 +320,8 @@ static int _compare_paths(const char *path0, const char *path1)
/* ASCII comparison */
if (strcmp(path0, path1) < 0)
return 0;
return 1;
else
return 1;
}
static int _add_alias(struct device *dev, const char *path)
@@ -706,12 +706,6 @@ static int _insert_dev(const char *path, dev_t d)
}
}
if (dm_hash_lookup(_cache.names, path) == dev) {
/* Hash already has matching entry present */
log_debug("Path already cached %s.", path);
return 1;
}
if (!(path_copy = dm_pool_strdup(_cache.mem, path))) {
log_error("Failed to duplicate path string.");
return 0;
@@ -898,10 +892,10 @@ int dev_cache_index_devs(void)
if (errno == ENOENT) {
sysfs_has_dev_block = 0;
return 1;
} else {
log_sys_error("stat", path);
return 0;
}
log_sys_error("stat", path);
return 0;
}
} else if (!sysfs_has_dev_block)
return 1;
@@ -939,20 +933,12 @@ static int _insert_udev_dir(struct udev *udev, const char *dir)
struct udev_device *device;
int r = 1;
if (!(udev_enum = udev_enumerate_new(udev))) {
log_error("Failed to udev_enumerate_new.");
return 0;
}
if (!(udev_enum = udev_enumerate_new(udev)))
goto bad;
if (udev_enumerate_add_match_subsystem(udev_enum, "block")) {
log_error("Failed to udev_enumerate_add_match_subsystem.");
goto out;
}
if (udev_enumerate_scan_devices(udev_enum)) {
log_error("Failed to udev_enumerate_scan_devices.");
goto out;
}
if (udev_enumerate_add_match_subsystem(udev_enum, "block") ||
udev_enumerate_scan_devices(udev_enum))
goto bad;
/*
* Report any missing information as "log_very_verbose" only, do not
@@ -989,10 +975,13 @@ static int _insert_udev_dir(struct udev *udev, const char *dir)
udev_device_unref(device);
}
out:
udev_enumerate_unref(udev_enum);
return r;
bad:
log_error("Failed to enumerate udev device list.");
udev_enumerate_unref(udev_enum);
return 0;
}
static void _insert_dirs(struct dm_list *dirs)
@@ -1086,8 +1075,6 @@ static void _full_scan(int dev_scan)
if (_cache.has_scanned && !dev_scan)
return;
log_debug_devs("Adding device paths to dev cache");
_insert_dirs(&_cache.dirs);
(void) dev_cache_index_devs();
@@ -1097,8 +1084,6 @@ static void _full_scan(int dev_scan)
_cache.has_scanned = 1;
init_full_scan_done(1);
log_debug_devs("Added %d device paths to dev cache", dm_hash_get_num_entries(_cache.names));
}
int dev_cache_has_scanned(void)
@@ -1379,19 +1364,6 @@ const char *dev_name_confirmed(struct device *dev, int quiet)
return dev_name(dev);
}
/* Provide a custom reason when a device is ignored */
const char *dev_cache_filtered_reason(const char *name)
{
const char *reason = "not found";
struct device *d = (struct device *) dm_hash_lookup(_cache.names, name);
if (d)
/* FIXME Record which filter caused the exclusion */
reason = "excluded by a filter";
return reason;
}
struct device *dev_cache_get(const char *name, struct dev_filter *f)
{
struct stat buf;

View File

@@ -55,7 +55,6 @@ int dev_cache_add_dir(const char *path);
int dev_cache_add_loopfile(const char *path);
__attribute__((nonnull(1)))
struct device *dev_cache_get(const char *name, struct dev_filter *f);
const char *dev_cache_filtered_reason(const char *name);
// TODO
struct device *dev_cache_get_by_devt(dev_t device, struct dev_filter *f);

View File

@@ -405,8 +405,8 @@ int dev_get_size(struct device *dev, uint64_t *size)
if ((dev->flags & DEV_REGULAR))
return _dev_get_size_file(dev, size);
return _dev_get_size_dev(dev, size);
else
return _dev_get_size_dev(dev, size);
}
int dev_get_read_ahead(struct device *dev, uint32_t *read_ahead)
@@ -827,324 +827,3 @@ int dev_set(struct device *dev, uint64_t offset, size_t len, int value)
return (len == 0);
}
#ifdef AIO_SUPPORT
/*
* io_setup() wrapper:
* async_event_count is the max number of concurrent async
* i/os, i.e. the number of devices that can be read at once
*
* max_io_alloc_count: max number of aio structs to allocate,
* each with a buf_len size buffer.
*
* max_buf_alloc_bytes: max number of bytes to use for buffers
* attached to all aio structs; each aio struct gets a
* buf_len size buffer.
*
* When only max_io_alloc_count is set, it is used directly.
*
* When only max_buf_alloc_bytes is set, the number of aio
* structs is determined by this number divided by buf_len.
*
* When both are set, max_io_alloc_count is reduced, if needed,
* to whatever value max_buf_alloc_bytes would allow.
*
* When both are zero, there is no limit on the number of aio
* structs. If allocation fails for an aio struct or its buffer,
* the code should revert to synchronous io.
*/
struct dev_async_context *dev_async_context_setup(unsigned async_event_count,
unsigned max_io_alloc_count,
unsigned max_buf_alloc_bytes,
int buf_len)
{
struct dev_async_context *ac;
unsigned nr_events = DEFAULT_ASYNC_EVENTS;
int count;
int error;
if (async_event_count)
nr_events = async_event_count;
if (!(ac = malloc(sizeof(struct dev_async_context))))
return_0;
memset(ac, 0, sizeof(struct dev_async_context));
dm_list_init(&ac->unused_ios);
error = io_setup(nr_events, &ac->aio_ctx);
if (error < 0) {
log_warn("WARNING: async io setup error %d with %u events.", error, nr_events);
free(ac);
return_0;
}
if (!max_io_alloc_count && !max_buf_alloc_bytes)
count = 0;
else if (!max_io_alloc_count && max_buf_alloc_bytes)
count = max_buf_alloc_bytes / buf_len;
else if (max_io_alloc_count && max_buf_alloc_bytes) {
if (max_io_alloc_count * buf_len > max_buf_alloc_bytes)
count = max_buf_alloc_bytes / buf_len;
} else
count = max_io_alloc_count;
ac->max_ios = count;
return ac;
}
void dev_async_context_destroy(struct dev_async_context *ac)
{
io_destroy(ac->aio_ctx);
free(ac);
}
static struct dev_async_io *_async_io_alloc(int buf_len)
{
struct dev_async_io *aio;
char *buf;
char **p_buf;
/*
* mem pool doesn't seem to work for this, probably because
* of the memalign that follows.
*/
if (!(aio = malloc(sizeof(struct dev_async_io))))
return_0;
memset(aio, 0, sizeof(struct dev_async_io));
buf = NULL;
p_buf = &buf;
if (posix_memalign((void *)p_buf, getpagesize(), buf_len)) {
free(aio);
return_NULL;
}
memset(buf, 0, buf_len);
aio->buf = buf;
aio->buf_len = buf_len;
return aio;
}
static void _async_io_free(struct dev_async_io *aio)
{
if (aio->buf)
free(aio->buf);
free(aio);
}
int dev_async_alloc_ios(struct dev_async_context *ac, int num, int buf_len, int *available)
{
struct dev_async_io *aio;
int count;
int i;
/* FIXME: check if num wants more pre allocated? */
if (!dm_list_empty(&ac->unused_ios))
return 1;
/*
* When no limit is used and no pre-alloc number is set,
* then no ios are allocated up front, but the are
* allocated as needed in get().
*/
if (!ac->max_ios && !num) {
*available = 0;
return 1;
}
if (num && !ac->max_ios)
count = num;
else if (!num && ac->max_ios)
count = ac->max_ios;
else if (num > ac->max_ios)
count = ac->max_ios;
else if (num < ac->max_ios)
count = num;
else
count = ac->max_ios;
for (i = 0; i < count; i++) {
if (!(aio = _async_io_alloc(buf_len))) {
ac->num_ios = i;
*available = i;
return 1;
}
dm_list_add(&ac->unused_ios, &aio->list);
}
ac->num_ios = count;
*available = count;
return 1;
}
void dev_async_free_ios(struct dev_async_context *ac)
{
struct dev_async_io *aio, *aio2;
dm_list_iterate_items_safe(aio, aio2, &ac->unused_ios) {
dm_list_del(&aio->list);
_async_io_free(aio);
}
}
struct dev_async_io *dev_async_io_get(struct dev_async_context *ac, int buf_len)
{
struct dm_list *aio_item;
struct dev_async_io *aio;
if (!(aio_item = dm_list_first(&ac->unused_ios)))
goto alloc_new;
aio = dm_list_item(aio_item, struct dev_async_io);
dm_list_del(&aio->list);
return aio;
alloc_new:
/* alloc on demand if there is no max or we have used less than max */
if (!ac->max_ios || (ac->num_ios < ac->max_ios)) {
if ((aio = _async_io_alloc(buf_len))) {
ac->num_ios++;
return aio;
}
}
return NULL;
}
void dev_async_io_put(struct dev_async_context *ac, struct dev_async_io *aio)
{
if (!ac)
_async_io_free(aio);
else {
memset(aio->buf, 0, aio->buf_len);
aio->dev = NULL;
aio->len = 0;
aio->done = 0;
aio->result = 0;
dm_list_add(&ac->unused_ios, &aio->list);
}
}
/* io_submit() wrapper */
int dev_async_read_submit(struct dev_async_context *ac, struct dev_async_io *aio,
struct device *dev, uint32_t len, uint64_t offset, int *nospace)
{
struct iocb *iocb = &aio->iocb;
int error;
*nospace = 0;
if (len > aio->buf_len)
return_0;
aio->len = len;
iocb->data = aio;
iocb->aio_fildes = dev_fd(dev);
iocb->aio_lio_opcode = IO_CMD_PREAD;
iocb->u.c.buf = aio->buf;
iocb->u.c.nbytes = len;
iocb->u.c.offset = offset;
error = io_submit(ac->aio_ctx, 1, &iocb);
if (error == -EAGAIN)
*nospace = 1;
if (error < 0)
return 0;
return 1;
}
/* io_getevents() wrapper */
int dev_async_getevents(struct dev_async_context *ac, int wait_count, struct timespec *timeout,
int *done_count)
{
int wait_nr;
int rv;
int i;
*done_count = 0;
retry:
memset(&ac->events, 0, sizeof(ac->events));
if (wait_count >= MAX_GET_EVENTS)
wait_nr = MAX_GET_EVENTS;
else
wait_nr = wait_count;
rv = io_getevents(ac->aio_ctx, 1, wait_nr, (struct io_event *)&ac->events, timeout);
if (rv == -EINTR)
goto retry;
if (rv < 0)
return 0;
if (!rv)
return 1;
for (i = 0; i < rv; i++) {
struct iocb *iocb = ac->events[i].obj;
struct dev_async_io *aio = iocb->data;
aio->result = ac->events[i].res;
aio->done = 1;
}
*done_count = rv;
return 1;
}
#else /* AIO_SUPPORT */
struct dev_async_context *dev_async_context_setup(unsigned async_event_count,
unsigned max_io_alloc_count,
unsigned max_buf_alloc_bytes,
int buf_len)
{
return NULL;
}
void dev_async_context_destroy(struct dev_async_context *ac)
{
}
int dev_async_alloc_ios(struct dev_async_context *ac, int num, int buf_len, int *available)
{
return 0;
}
void dev_async_free_ios(struct dev_async_context *ac)
{
}
struct dev_async_io *dev_async_io_get(struct dev_async_context *ac, int buf_len)
{
return NULL;
}
void dev_async_io_put(struct dev_async_context *ac, struct dev_async_io *aio)
{
}
int dev_async_read_submit(struct dev_async_context *ac, struct dev_async_io *aio,
struct device *dev, uint32_t len, uint64_t offset, int *nospace)
{
return 0;
}
int dev_async_getevents(struct dev_async_context *ac, int wait_count, struct timespec *timeout,
int *done_count)
{
return 0;
}
#endif /* AIO_SUPPORT */

View File

@@ -28,7 +28,7 @@
#define MD_SB_MAGIC 0xa92b4efc
#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)) \
#define MD_NEW_SIZE_SECTORS(x) ((x & ~(MD_RESERVED_SECTORS - 1)) \
- MD_RESERVED_SECTORS)
#define MD_MAX_SYSFS_SIZE 64
@@ -261,7 +261,8 @@ out:
/*
* Retrieve chunk size from md device using sysfs.
*/
static unsigned long _dev_md_chunk_size(struct dev_types *dt, struct device *dev)
static unsigned long dev_md_chunk_size(struct dev_types *dt,
struct device *dev)
{
const char *attribute = "chunk_size";
unsigned long chunk_size_bytes = 0UL;
@@ -279,7 +280,7 @@ static unsigned long _dev_md_chunk_size(struct dev_types *dt, struct device *dev
/*
* Retrieve level from md device using sysfs.
*/
static int _dev_md_level(struct dev_types *dt, struct device *dev)
static int dev_md_level(struct dev_types *dt, struct device *dev)
{
char level_string[MD_MAX_SYSFS_SIZE];
const char *attribute = "level";
@@ -302,7 +303,7 @@ static int _dev_md_level(struct dev_types *dt, struct device *dev)
/*
* Retrieve raid_disks from md device using sysfs.
*/
static int _dev_md_raid_disks(struct dev_types *dt, struct device *dev)
static int dev_md_raid_disks(struct dev_types *dt, struct device *dev)
{
const char *attribute = "raid_disks";
int raid_disks = 0;
@@ -326,15 +327,15 @@ unsigned long dev_md_stripe_width(struct dev_types *dt, struct device *dev)
unsigned long stripe_width_sectors = 0UL;
int level, raid_disks, data_disks;
chunk_size_sectors = _dev_md_chunk_size(dt, dev);
chunk_size_sectors = dev_md_chunk_size(dt, dev);
if (!chunk_size_sectors)
return 0;
level = _dev_md_level(dt, dev);
level = dev_md_level(dt, dev);
if (level < 0)
return 0;
raid_disks = _dev_md_raid_disks(dt, dev);
raid_disks = dev_md_raid_disks(dt, dev);
if (!raid_disks)
return 0;

View File

@@ -20,7 +20,8 @@
#define MAX_PAGESIZE (64 * 1024)
#define SIGNATURE_SIZE 10
static int _swap_detect_signature(const char *buf)
static int
_swap_detect_signature(const char *buf)
{
if (memcmp(buf, "SWAP-SPACE", 10) == 0 ||
memcmp(buf, "SWAPSPACE2", 10) == 0)

View File

@@ -76,7 +76,7 @@ struct dev_types *create_dev_types(const char *proc_dir,
i++;
/* If it's not a number it may be name of section */
line_maj = atoi(line + i);
line_maj = atoi(((char *) (line + i)));
if (line_maj < 0 || line_maj >= NUMBER_OF_MAJORS) {
/*
@@ -615,38 +615,38 @@ static int _blkid_wipe(blkid_probe probe, struct device *dev, const char *name,
if (force < DONT_PROMPT) {
log_error(MSG_FAILED_SIG_OFFSET, type, name);
return 0;
} else {
log_error("WARNING: " MSG_FAILED_SIG_OFFSET MSG_WIPING_SKIPPED, type, name);
return 2;
}
log_error("WARNING: " MSG_FAILED_SIG_OFFSET MSG_WIPING_SKIPPED, type, name);
return 2;
}
if (blkid_probe_lookup_value(probe, "SBMAGIC", &magic, &len)) {
if (force < DONT_PROMPT) {
log_error(MSG_FAILED_SIG_LENGTH, type, name);
return 0;
} else {
log_warn("WARNING: " MSG_FAILED_SIG_LENGTH MSG_WIPING_SKIPPED, type, name);
return 2;
}
log_warn("WARNING: " MSG_FAILED_SIG_LENGTH MSG_WIPING_SKIPPED, type, name);
return 2;
}
} else if (!blkid_probe_lookup_value(probe, "PTTYPE", &type, NULL)) {
if (blkid_probe_lookup_value(probe, "PTMAGIC_OFFSET", &offset, NULL)) {
if (force < DONT_PROMPT) {
log_error(MSG_FAILED_SIG_OFFSET, type, name);
return 0;
} else {
log_warn("WARNING: " MSG_FAILED_SIG_OFFSET MSG_WIPING_SKIPPED, type, name);
return 2;
}
log_warn("WARNING: " MSG_FAILED_SIG_OFFSET MSG_WIPING_SKIPPED, type, name);
return 2;
}
if (blkid_probe_lookup_value(probe, "PTMAGIC", &magic, &len)) {
if (force < DONT_PROMPT) {
log_error(MSG_FAILED_SIG_LENGTH, type, name);
return 0;
} else {
log_warn("WARNING: " MSG_FAILED_SIG_LENGTH MSG_WIPING_SKIPPED, type, name);
return 2;
}
log_warn("WARNING: " MSG_FAILED_SIG_LENGTH MSG_WIPING_SKIPPED, type, name);
return 2;
}
usage = "partition table";
} else

View File

@@ -19,7 +19,6 @@
#include "uuid.h"
#include <fcntl.h>
#include <libaio.h>
#define DEV_ACCESSED_W 0x00000001 /* Device written to? */
#define DEV_REGULAR 0x00000002 /* Regular file? */
@@ -91,32 +90,6 @@ struct device_area {
uint64_t size; /* Bytes */
};
/*
* We'll collect the results of this many async reads
* in one system call. It shouldn't matter much what
* number is used here.
*/
#define MAX_GET_EVENTS 16
struct dev_async_context {
io_context_t aio_ctx;
struct io_event events[MAX_GET_EVENTS]; /* for processing completions */
struct dm_list unused_ios; /* unused/available aio structcs */
int num_ios; /* number of allocated aio structs */
int max_ios; /* max number of aio structs to allocate */
};
struct dev_async_io {
struct dm_list list;
struct iocb iocb;
struct device *dev;
char *buf;
uint32_t buf_len; /* size of buf */
uint32_t len; /* size of submitted io */
int done;
int result;
};
/*
* Support for external device info.
*/
@@ -171,27 +144,4 @@ void dev_destroy_file(struct device *dev);
/* Return a valid device name from the alias list; NULL otherwise */
const char *dev_name_confirmed(struct device *dev, int quiet);
struct dev_async_context *dev_async_context_setup(unsigned async_event_count,
unsigned max_io_alloc_count,
unsigned max_buf_alloc_bytes,
int buf_len);
void dev_async_context_destroy(struct dev_async_context *ac);
/* allocate aio structs (with buffers), up to the max specified during context setup */
int dev_async_alloc_ios(struct dev_async_context *ac, int num, int buf_len, int *available);
/* free aio structs (and buffers) */
void dev_async_free_ios(struct dev_async_context *ac);
/* get an available aio struct (with buffer) */
struct dev_async_io *dev_async_io_get(struct dev_async_context *ac, int buf_len);
/* make an aio struct (with buffer) available for use (by another get) */
void dev_async_io_put(struct dev_async_context *ac, struct dev_async_io *aio);
int dev_async_read_submit(struct dev_async_context *ac, struct dev_async_io *aio,
struct device *dev, uint32_t len, uint64_t offset, int *nospace);
int dev_async_getevents(struct dev_async_context *ac, int wait_count, struct timespec *timeout,
int *done_count);
#endif

View File

@@ -1,6 +1,6 @@
/*
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
* Copyright (C) 2004-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -152,30 +152,6 @@ const char *display_lvname(const struct logical_volume *lv)
return name;
}
/* Display percentage with (TODO) configurable precision */
const char *display_percent(struct cmd_context *cmd, dm_percent_t percent)
{
char *buf;
int r;
/* Reusing same ring buffer we use for displaying LV names */
if ((cmd->display_lvname_idx + NAME_LEN) >= sizeof((cmd->display_buffer)))
cmd->display_lvname_idx = 0;
buf = cmd->display_buffer + cmd->display_lvname_idx;
/* TODO: Make configurable hardcoded 2 digits */
r = dm_snprintf(buf, NAME_LEN, "%.2f", dm_percent_to_round_float(percent, 2));
if (r < 0) {
log_error("Percentage %d does not fit.", percent);
return NULL;
}
cmd->display_lvname_idx += r + 1;
return buf;
}
/* Size supplied in sectors */
static const char *_display_size(const struct cmd_context *cmd,
uint64_t size, dm_size_suffix_t suffix_type)
@@ -410,7 +386,6 @@ int lvdisplay_full(struct cmd_context *cmd,
dm_percent_t thin_data_percent, thin_metadata_percent;
int thin_active = 0;
dm_percent_t thin_percent;
struct lv_status_cache *cache_status = NULL;
if (lv_is_historical(lv))
return _lvdisplay_historical_full(cmd, lv);
@@ -516,19 +491,6 @@ int lvdisplay_full(struct cmd_context *cmd,
seg = first_seg(lv);
log_print("LV Pool metadata %s", seg->metadata_lv->name);
log_print("LV Pool data %s", seg_lv(seg, 0)->name);
} else if (lv_is_cache_origin(lv)) {
if ((seg = get_only_segment_using_this_lv(lv)))
log_print("LV origin of Cache LV %s", seg->lv->name);
} else if (lv_is_cache(lv)) {
seg = first_seg(lv);
if (inkernel && !lv_cache_status(lv, &cache_status))
return_0;
log_print("LV Cache pool name %s", seg->pool_lv->name);
log_print("LV Cache origin name %s", seg_lv(seg, 0)->name);
} else if (lv_is_cache_pool(lv)) {
seg = first_seg(lv);
log_print("LV Pool metadata %s", seg->metadata_lv->name);
log_print("LV Pool data %s", seg_lv(seg, 0)->name);
}
if (inkernel && info.suspended)
@@ -548,38 +510,17 @@ int lvdisplay_full(struct cmd_context *cmd,
display_size(cmd,
snap_seg ? snap_seg->origin->size : lv->size));
if (cache_status) {
log_print("Cache used blocks %s%%",
display_percent(cmd, cache_status->data_usage));
log_print("Cache metadata blocks %s%%",
display_percent(cmd, cache_status->metadata_usage));
log_print("Cache dirty blocks %s%%",
display_percent(cmd, cache_status->dirty_usage));
log_print("Cache read hits/misses " FMTu64 " / " FMTu64,
cache_status->cache->read_hits,
cache_status->cache->read_misses);
log_print("Cache wrt hits/misses " FMTu64 " / " FMTu64,
cache_status->cache->write_hits,
cache_status->cache->write_misses);
log_print("Cache demotions " FMTu64,
cache_status->cache->demotions);
log_print("Cache promotions " FMTu64,
cache_status->cache->promotions);
dm_pool_destroy(cache_status->mem);
}
if (thin_data_active)
log_print("Allocated pool data %s%%",
display_percent(cmd, thin_data_percent));
log_print("Allocated pool data %.2f%%",
dm_percent_to_float(thin_data_percent));
if (thin_metadata_active)
log_print("Allocated metadata %s%%",
display_percent(cmd, thin_metadata_percent));
log_print("Allocated metadata %.2f%%",
dm_percent_to_float(thin_metadata_percent));
if (thin_active)
log_print("Mapped size %s%%",
display_percent(cmd, thin_percent));
log_print("Mapped size %.2f%%",
dm_percent_to_float(thin_percent));
log_print("Current LE %u",
snap_seg ? snap_seg->origin->le_count : lv->le_count);
@@ -590,8 +531,8 @@ int lvdisplay_full(struct cmd_context *cmd,
log_print("COW-table LE %u", lv->le_count);
if (snap_active)
log_print("Allocated to snapshot %s%%",
display_percent(cmd, snap_percent));
log_print("Allocated to snapshot %.2f%%",
dm_percent_to_float(snap_percent));
log_print("Snapshot chunk size %s",
display_size(cmd, (uint64_t) snap_seg->chunk_size));

View File

@@ -1,6 +1,6 @@
/*
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
* Copyright (C) 2004-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -24,8 +24,6 @@
const char *display_lvname(const struct logical_volume *lv);
const char *display_percent(struct cmd_context *cmd, dm_percent_t percent);
/* Specify size in KB */
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);

View File

@@ -274,9 +274,8 @@ static int _accept_p(struct dev_filter *f, struct device *dev)
if (!_set_lookup(ds, dev->dev)) {
log_debug_devs("%s: Skipping (sysfs)", dev_name(dev));
return 0;
}
return 1;
} else
return 1;
}
static void _destroy(struct dev_filter *f)

View File

@@ -180,9 +180,9 @@ out:
static struct volume_group *_format1_vg_read(struct format_instance *fid,
const char *vg_name,
struct metadata_area *mda __attribute__((unused)),
struct label_read_data *ld __attribute__((unused)),
struct cached_vg_fmtdata **vg_fmtdata __attribute__((unused)),
unsigned *use_previous_vg __attribute__((unused)))
unsigned *use_previous_vg __attribute__((unused)),
int single_device __attribute__((unused)))
{
struct volume_group *vg;
struct disk_list *dl;

View File

@@ -128,8 +128,8 @@ int import_pv(const struct format_type *fmt, struct dm_pool *mem,
int generate_lvm1_system_id(struct cmd_context *cmd, char *s, const char *prefix)
{
if (dm_snprintf(s, NAME_LEN, "%s%s" FMTu64,
prefix, cmd->hostname, (uint64_t)time(NULL)) < 0) {
if (dm_snprintf(s, NAME_LEN, "%s%s%lu",
prefix, cmd->hostname, time(NULL)) < 0) {
log_error("Generated LVM1 format system_id too long");
return 0;
}

View File

@@ -1,6 +1,6 @@
/*
* Copyright (C) 2001-2004 Sistina Software, Inc. All rights reserved.
* Copyright (C) 2004-2017 Red Hat, Inc. All rights reserved.
* Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved.
*
* This file is part of LVM2.
*
@@ -134,35 +134,37 @@ static int _fill_maps(struct dm_hash_table *maps, struct volume_group *vg,
if (lv_num == UNMAPPED_EXTENT)
continue;
lv_num--;
lvm = lvms[lv_num];
else {
lv_num--;
lvm = lvms[lv_num];
if (!lvm) {
log_error("Invalid LV in extent map "
"(PV %s, PE %" PRIu32
", LV %" PRIu32
", LE %" PRIu32 ")",
dev_name(pv->dev), i,
lv_num, e[i].le_num);
return 0;
if (!lvm) {
log_error("Invalid LV in extent map "
"(PV %s, PE %" PRIu32
", LV %" PRIu32
", LE %" PRIu32 ")",
dev_name(pv->dev), i,
lv_num, e[i].le_num);
return 0;
}
le = e[i].le_num;
if (le >= lvm->lv->le_count) {
log_error("logical extent number "
"out of bounds");
return 0;
}
if (lvm->map[le].pv) {
log_error("logical extent (%u) "
"already mapped.", le);
return 0;
}
lvm->map[le].pv = pv;
lvm->map[le].pe = i;
}
le = e[i].le_num;
if (le >= lvm->lv->le_count) {
log_error("logical extent number "
"out of bounds");
return 0;
}
if (lvm->map[le].pv) {
log_error("logical extent (%u) "
"already mapped.", le);
return 0;
}
lvm->map[le].pv = pv;
lvm->map[le].pe = i;
}
}
@@ -223,8 +225,8 @@ static int _read_linear(struct cmd_context *cmd, struct lv_map *lvm)
while (le < lvm->lv->le_count) {
len = _area_length(lvm, le);
if (!(seg = alloc_lv_segment(segtype, lvm->lv, le, len, 0, 0, 0,
NULL, 1, len, 0, 0, 0, 0, NULL))) {
if (!(seg = alloc_lv_segment(segtype, lvm->lv, le, len, 0, 0,
NULL, 1, len, 0, 0, 0, NULL))) {
log_error("Failed to allocate linear segment.");
return 0;
}
@@ -295,10 +297,10 @@ static int _read_stripes(struct cmd_context *cmd, struct lv_map *lvm)
if (!(seg = alloc_lv_segment(segtype, lvm->lv,
lvm->stripes * first_area_le,
lvm->stripes * area_len, 0,
lvm->stripes * area_len,
0, lvm->stripe_size, NULL,
lvm->stripes,
area_len, 0, 0, 0, 0, NULL))) {
area_len, 0, 0, 0, NULL))) {
log_error("Failed to allocate striped segment.");
return 0;
}

View File

@@ -55,7 +55,6 @@ static int _lvm1_write(struct label *label __attribute__((unused)), void *buf __
}
static int _lvm1_read(struct labeller *l, struct device *dev, void *buf,
struct label_read_data *ld,
struct label **label)
{
struct pv_disk *pvd = (struct pv_disk *) buf;

View File

@@ -101,9 +101,9 @@ static int _check_usp(const char *vgname, struct user_subpool *usp, int sp_count
static struct volume_group *_pool_vg_read(struct format_instance *fid,
const char *vg_name,
struct metadata_area *mda __attribute__((unused)),
struct label_read_data *ld __attribute__((unused)),
struct cached_vg_fmtdata **vg_fmtdata __attribute__((unused)),
unsigned *use_previous_vg __attribute__((unused)))
unsigned *use_previous_vg __attribute__((unused)),
int single_device __attribute__((unused)))
{
struct volume_group *vg;
struct user_subpool *usp;

Some files were not shown because too many files have changed in this diff Show More