]> sourceware.org Git - lvm2.git/log
lvm2.git
12 days agocleanup: more typos in WHATS_NEW
zkabelac [Fri, 30 Aug 2024 17:47:00 +0000 (17:47 +0000)]
cleanup: more typos in WHATS_NEW

12 days agocleanup: fix typos
Tobias Stoeckmann [Fri, 13 Sep 2024 16:33:45 +0000 (18:33 +0200)]
cleanup: fix typos

Typos found with codespell.

Signed-off-by: Tobias Stoeckmann <tobias@stoeckmann.org>
12 days agomemlock: use value of 0 to disable memory locking
Zdenek Kabelac [Tue, 24 Sep 2024 14:04:53 +0000 (16:04 +0200)]
memlock: use value of 0 to disable memory locking

In cases user is sure he is not using his 'rootfs' or 'swap' on LVs
managed with his command - it possible to completely bypass pinning
process to RAM which may eventually slightly speedup command execution,
(however at the risk the process can be eventually delayed by swapping).
Basicaly use this only at your risk...

TODO: add some dmeventd support for this.

13 days agovdo: depcreate unused settings
Zdenek Kabelac [Fri, 6 Sep 2024 09:39:17 +0000 (11:39 +0200)]
vdo: depcreate unused settings

13 days agolvmlockd: use lvmlock LV size 1470849140
David Teigland [Wed, 25 Sep 2024 21:18:32 +0000 (16:18 -0500)]
lvmlockd: use lvmlock LV size

Previously, lvmlockd detected the end of the lvmlock LV
by doing i/o to it until an i/o error was returned.
This triggered sanlock warning messages, so use the LV
size to avoid accessing beyond the end of the device.

Previously, every lvcreate would refresh the lvmlock LV
in case another machine had extended it.  This involves
a lot of unnecessary work in most cases, so now compare
the LV size and device size to detect when a refresh is
needed.

2 weeks agolvmlockd: fix previous thin locking fix 1465886237
David Teigland [Mon, 23 Sep 2024 22:12:03 +0000 (17:12 -0500)]
lvmlockd: fix previous thin locking fix

Restore the original lvremove locking for non-thin LVs
that were changed in the thin locking fix.

2 weeks agovg: remove unused hostnames hash table 1465773143
David Teigland [Mon, 23 Sep 2024 20:04:53 +0000 (15:04 -0500)]
vg: remove unused hostnames hash table

The hash table of lv creation hostnames was not used for anything,
so remove it.

2 weeks agolvmlockd: fix locking for thin 1465738792
David Teigland [Mon, 23 Sep 2024 19:46:29 +0000 (14:46 -0500)]
lvmlockd: fix locking for thin

lvremove of a thin lv while the pool is inactive would
leave the pool locked but inactive.

lvcreate of a thin snapshot while the pool is inactive
would leave the pool locked but inactive.

lvcreate of a thin lv could activate the pool to check
a threshold before the pool lock was acquired in lvmlockd.

2 weeks agotests: remove superfluous -a option for df used in lvresize-fs.sh 1460603502
Peter Rajnoha [Thu, 19 Sep 2024 13:12:44 +0000 (15:12 +0200)]
tests: remove superfluous -a option for df used in lvresize-fs.sh

4 weeks agometadata: use lv_hash in segment-specific metadata parsing 1447913946
David Teigland [Tue, 10 Sep 2024 16:51:15 +0000 (11:51 -0500)]
metadata: use lv_hash in segment-specific metadata parsing

The lv_hash wasn't being passed to the seg-specific text import
functions, so they were doing many find_lv() calls which consumes
a lot of time when there are many LVs in the metadata.

4 weeks agoargs: readonly description update 1443255555
David Teigland [Fri, 6 Sep 2024 13:44:06 +0000 (08:44 -0500)]
args: readonly description update

Include reference to --permission r.

4 weeks agoWHATS_NEW_DM: update 12/head 1441387489 1441427841
Peter Rajnoha [Thu, 5 Sep 2024 08:34:43 +0000 (10:34 +0200)]
WHATS_NEW_DM: update

4 weeks agolibdm: do not fail if GETVAL semctl fails for udev sync inc and dec
Peter Rajnoha [Mon, 12 Aug 2024 12:31:19 +0000 (14:31 +0200)]
libdm: do not fail if GETVAL semctl fails for udev sync inc and dec

While performing udev sync semaphore's inc/dec operation, we use the
result from GETVAL semctl just to print a debug message with current
value of that sempahore, nothing else.

If the GETVAL fails for whetever reason while the actual inc/dec
completes successfully, just log a warning message about the GETVAL
(and print the debug messages without the actual semaphore value)
and return success for the inc/dec operation as a whole.

4 weeks agolibdm: clean up udev sync semaphore on fail path during its creation
Peter Rajnoha [Mon, 12 Aug 2024 12:16:32 +0000 (14:16 +0200)]
libdm: clean up udev sync semaphore on fail path during its creation

Clean up udev sync semaphore on fail path during its creation, otherwise
the caller will have no handle returned to clean it up itself and the
semaphore will keep staying in the system. The only way to clean it up
would be to call `dmsetup udevcomplete_all` which would destroy all
udev sync semaphores, not just the failed one, which we don't want.

4 weeks agolibdm: add 'cookie create/inc/dec' log prefix if GETVAL fails for udev sync ops
Peter Rajnoha [Mon, 12 Aug 2024 12:01:25 +0000 (14:01 +0200)]
libdm: add 'cookie create/inc/dec' log prefix if GETVAL fails for udev sync ops

The same message is printed while performing create/inc/dec operation and
the GETVAL semctl fails. Add a prefix so we know exactly in which of
these functions the issue actually happened.

5 weeks agoscripts: Fix ConditionPathExists in unit files 1439886932
Marian Csontos [Wed, 4 Sep 2024 12:40:58 +0000 (14:40 +0200)]
scripts: Fix ConditionPathExists in unit files

5 weeks agoWHATS_NEW: update 1433864118
Zdenek Kabelac [Fri, 30 Aug 2024 14:44:50 +0000 (16:44 +0200)]
WHATS_NEW: update

5 weeks agoconfigure: autoreconf
Zdenek Kabelac [Thu, 29 Aug 2024 20:24:27 +0000 (22:24 +0200)]
configure: autoreconf

5 weeks agomake generate
Zdenek Kabelac [Thu, 29 Aug 2024 20:21:45 +0000 (22:21 +0200)]
make generate

5 weeks agocleanup: typos in configure.ac
Zdenek Kabelac [Thu, 29 Aug 2024 20:33:33 +0000 (22:33 +0200)]
cleanup: typos in configure.ac

5 weeks agocleanup: typos in test comments
Zdenek Kabelac [Fri, 30 Aug 2024 10:10:35 +0000 (12:10 +0200)]
cleanup: typos in test comments

5 weeks agocleanup: typos in test logging
Zdenek Kabelac [Fri, 30 Aug 2024 10:10:26 +0000 (12:10 +0200)]
cleanup: typos in test logging

5 weeks agocleanup: typos in doc
Zdenek Kabelac [Thu, 29 Aug 2024 21:30:33 +0000 (23:30 +0200)]
cleanup: typos in doc

5 weeks agocleanup: typos man pages
Zdenek Kabelac [Thu, 29 Aug 2024 20:05:12 +0000 (22:05 +0200)]
cleanup: typos man pages

5 weeks agocleanup: typos in WHATS_NEW
Zdenek Kabelac [Thu, 29 Aug 2024 22:12:36 +0000 (00:12 +0200)]
cleanup: typos in WHATS_NEW

5 weeks agocleanup: typos in comments
Zdenek Kabelac [Thu, 29 Aug 2024 21:05:41 +0000 (23:05 +0200)]
cleanup: typos in comments

Collection of typos in code comments.
Should have no runtime effect.

5 weeks agocleanup: typos in logging
Zdenek Kabelac [Thu, 29 Aug 2024 21:06:04 +0000 (23:06 +0200)]
cleanup: typos in logging

Fixes various typos in printed/logged messages.

5 weeks agocleanup: typo dirctory to directory
Zdenek Kabelac [Thu, 29 Aug 2024 21:25:10 +0000 (23:25 +0200)]
cleanup: typo dirctory to directory

Rename internal lm_idm_scsi_find_block_directory() function.

5 weeks agocleanup: typo premable to preamble
zkabelac [Thu, 29 Aug 2024 14:18:24 +0000 (14:18 +0000)]
cleanup: typo premable to preamble

Correct preable, seems even unused.

5 weeks agolibdm: _DOESNT_ to _DOES_NOT_
zkabelac [Thu, 29 Aug 2024 14:30:33 +0000 (14:30 +0000)]
libdm: _DOESNT_ to _DOES_NOT_

Rename internal define DM_NAME_LIST_FLAG_DOESNT_HAVE_UUID
to DM_NAME_LIST_FLAG_DOES_NOT_HAVE_UUID
(currently unused)

5 weeks agolvmlockd: retrive to retrieve
Zdenek Kabelac [Thu, 29 Aug 2024 20:07:20 +0000 (22:07 +0200)]
lvmlockd: retrive to retrieve

Replace retrive with retrieve for internal function.

5 weeks agolibdaemon: daemonise to daemonize
Zdenek Kabelac [Thu, 29 Aug 2024 20:49:24 +0000 (22:49 +0200)]
libdaemon: daemonise to daemonize

Replace daemonise() with daemonize().

5 weeks agolvconvert: use pvmove as part of renamed LV name
zkabelac [Thu, 29 Aug 2024 09:12:41 +0000 (09:12 +0000)]
lvconvert: use pvmove as part of renamed LV name

LV with pvmove_ prefix is not allowed to be created by user
so bigger chance our selected name will never exist.

TODO: probably add code to get generic unused LV name...

5 weeks agolibdm: check for queue name
Zdenek Kabelac [Thu, 29 Aug 2024 21:05:52 +0000 (23:05 +0200)]
libdm: check for queue name

Fix typo and check for queue string within sysfs kernel name.

5 weeks agointegrity: typo interleave_sectors
Zdenek Kabelac [Thu, 29 Aug 2024 19:33:13 +0000 (21:33 +0200)]
integrity: typo  interleave_sectors

Correct emitted parameter name for interleave_sectors for DM table line.

5 weeks agomake: generate 1433637790
Peter Rajnoha [Fri, 30 Aug 2024 12:17:24 +0000 (14:17 +0200)]
make: generate

5 weeks agoman: lvmreport: add note about log/command_log_selection="all"
Peter Rajnoha [Fri, 30 Aug 2024 12:14:02 +0000 (14:14 +0200)]
man: lvmreport: add note about log/command_log_selection="all"

5 weeks agoconfig: update description for log/command_log_selection
Peter Rajnoha [Fri, 30 Aug 2024 12:11:39 +0000 (14:11 +0200)]
config: update description for log/command_log_selection

5 weeks agolvresize: ignore given stripe and stripe size args for raid0 1432115355
Peter Rajnoha [Thu, 29 Aug 2024 11:57:52 +0000 (13:57 +0200)]
lvresize: ignore given stripe and stripe size args for raid0

6 weeks agotoollib: do not accept duplicate device name on pvcreate input 1430114117
Peter Rajnoha [Wed, 28 Aug 2024 08:35:18 +0000 (10:35 +0200)]
toollib: do not accept duplicate device name on pvcreate input

6 weeks agopost-release 1424917128
Marian Csontos [Fri, 23 Aug 2024 14:02:19 +0000 (16:02 +0200)]
post-release

6 weeks agopre-release 1424917303 v2_03_26
Marian Csontos [Fri, 23 Aug 2024 14:02:19 +0000 (16:02 +0200)]
pre-release

6 weeks agolvmlockd: increase dump buffer size 1423920664
David Teigland [Thu, 22 Aug 2024 21:23:07 +0000 (16:23 -0500)]
lvmlockd: increase dump buffer size

6 weeks agoWHATS_NEW: update 1423283245
Zdenek Kabelac [Thu, 22 Aug 2024 12:40:07 +0000 (14:40 +0200)]
WHATS_NEW: update

6 weeks agomake: generate
Zdenek Kabelac [Thu, 22 Aug 2024 12:39:04 +0000 (14:39 +0200)]
make: generate

6 weeks agogcc: single type initializers
Zdenek Kabelac [Thu, 22 Aug 2024 12:45:35 +0000 (14:45 +0200)]
gcc: single type initializers

Older gcc doesn't really like complex types (buffer, struct) to be
initialized without extra {} around such type.
So pick any other 'single type' var from a struct and set it to 0,
rest will do the compiler without emitting a warning.

6 weeks agocleanup: drop extra space
Zdenek Kabelac [Thu, 22 Aug 2024 12:38:40 +0000 (14:38 +0200)]
cleanup: drop extra space

6 weeks agotests: check pvmove on single PV
Zdenek Kabelac [Thu, 22 Aug 2024 12:34:58 +0000 (14:34 +0200)]
tests: check pvmove on single PV

6 weeks agovalidation: support empty lv segments
Zdenek Kabelac [Thu, 22 Aug 2024 12:27:10 +0000 (14:27 +0200)]
validation: support empty lv segments

Revert 373372c8ab3749bc76ced37cec04b00aae6e5979 and instead update
our validation code to handle LVs with empty segment - currently
we should need this only for pvmove operation, thus such LV should
have name  'pvmove%u'.

This fixes a problem where user tried i.e. pvmove on a VG with single
PV - as reported: https://github.com/lvmteam/lvm2/issues/148

Reported-by: bob@redhat.com
8 weeks agoWHATS_NEW: update 1413295909
Peter Rajnoha [Wed, 14 Aug 2024 13:30:30 +0000 (15:30 +0200)]
WHATS_NEW: update

8 weeks agotools: vgscan: accept --refresh option 1413267322
Peter Rajnoha [Wed, 14 Aug 2024 13:12:59 +0000 (15:12 +0200)]
tools: vgscan: accept --refresh option

8 weeks agotools: vgscan: accept --refresh option 1413262980
Peter Rajnoha [Wed, 14 Aug 2024 13:09:50 +0000 (15:09 +0200)]
tools: vgscan: accept --refresh option

8 weeks agotools: args: update description for --refresh
Peter Rajnoha [Wed, 14 Aug 2024 12:59:07 +0000 (14:59 +0200)]
tools: args: update description for --refresh

8 weeks agotools: vgmknodes: use display_lvname 1413186411
Peter Rajnoha [Wed, 14 Aug 2024 12:20:55 +0000 (14:20 +0200)]
tools: vgmknodes: use display_lvname

8 weeks agotools: vgmknodes: handle /dev/mapper content without LV refresh only if udev not... 1413141350
Peter Rajnoha [Wed, 14 Aug 2024 11:53:17 +0000 (13:53 +0200)]
tools: vgmknodes: handle /dev/mapper content without LV refresh only if udev not running

8 weeks agotools: vgmknodes: log error message if LV refresh fails
Peter Rajnoha [Wed, 14 Aug 2024 11:50:22 +0000 (13:50 +0200)]
tools: vgmknodes: log error message if LV refresh fails

8 weeks agotools: vgmknodes: wait for udev first with --refresh, then check /dev 1413077022
Peter Rajnoha [Wed, 14 Aug 2024 11:07:25 +0000 (13:07 +0200)]
tools: vgmknodes: wait for udev first with --refresh, then check /dev

8 weeks agoman: vgmknodes: add note about udev and --refresh 1412959631
Peter Rajnoha [Wed, 14 Aug 2024 09:51:34 +0000 (11:51 +0200)]
man: vgmknodes: add note about udev and --refresh

8 weeks agogitignore: update 1411258796
Peter Rajnoha [Tue, 13 Aug 2024 08:42:17 +0000 (10:42 +0200)]
gitignore: update

The .cache and compile_commands.json is used by popular source crawling and
indexing clang tools which in turn may be integrated with source code editors.

We may reuse the .cache directory for for other caches and temporary
files.

The /doc/.ikiwiki and /public are related to the ikiwiki.

2 months agomake: generate 1407456797
Zdenek Kabelac [Fri, 9 Aug 2024 14:34:22 +0000 (16:34 +0200)]
make: generate

2 months agoRevert "tests: use put_time" 1407371944
Zdenek Kabelac [Fri, 9 Aug 2024 13:37:59 +0000 (15:37 +0200)]
Revert "tests: use put_time"

This reverts commit b4af2571dbee9dab5717bb166ed63df5ae63974b.
std::put_time is a relatively new feature so breaks builds on older
systems.

2 months agotests: add some udev waits 1407347563
Zdenek Kabelac [Mon, 5 Aug 2024 15:01:13 +0000 (17:01 +0200)]
tests: add some udev waits

Avoid possible udev race - since dmsetup create is
not using the same cookie logic as lvm2 commands,
try to avoid racing on some systems with udev scanning.

2 months agotests: use put_time
Zdenek Kabelac [Fri, 9 Aug 2024 13:00:14 +0000 (15:00 +0200)]
tests: use put_time

2 months agotests: search for delimiter in bounded buffer
Zdenek Kabelac [Fri, 9 Aug 2024 11:13:03 +0000 (13:13 +0200)]
tests: search for delimiter in bounded buffer

2 months agocov: annotate known case
Zdenek Kabelac [Mon, 29 Jul 2024 15:29:08 +0000 (17:29 +0200)]
cov: annotate known case

2 months agocov: validate list emptiness
Zdenek Kabelac [Mon, 29 Jul 2024 14:53:28 +0000 (16:53 +0200)]
cov: validate list emptiness

Emit internal error for empty list.

2 months agocov: validate altdev list is not empty
Zdenek Kabelac [Mon, 29 Jul 2024 14:00:27 +0000 (16:00 +0200)]
cov: validate altdev list is not empty

2 months agocov: check for dev_iter_create result
Zdenek Kabelac [Mon, 29 Jul 2024 14:50:41 +0000 (16:50 +0200)]
cov: check for dev_iter_create result

dev_iter_create() may fail in malloc so check
for its error code.

2 months agocov: drop unused header inclusion
Zdenek Kabelac [Mon, 29 Jul 2024 13:50:59 +0000 (15:50 +0200)]
cov: drop unused header inclusion

2 months agocov: avoid expression overflow
Zdenek Kabelac [Fri, 9 Aug 2024 11:06:01 +0000 (13:06 +0200)]
cov: avoid expression overflow

Use 64bit math to multiply extent_size.

2 months agointegrity: add --integritysettings for tuning 1404928847
David Teigland [Mon, 5 Aug 2024 18:20:58 +0000 (13:20 -0500)]
integrity: add --integritysettings for tuning

The option can be used in multiple ways (like --cachesettings):

--integritysettings key=val
--integritysettings 'key1=val1 key2=val2'
--integritysettings key1=val1 --integritysettings key2=val2

Use with lvcreate or lvconvert when integrity is first enabled
to configure:
journal_sectors
journal_watermark
commit_time
bitmap_flush_interval
allow_discards

Use with lvchange to configure (only while inactive):
journal_watermark
commit_time
bitmap_flush_interval
allow_discards

lvchange --integritysettings "" clears any previously configured
settings, so dm-integrity will use its own defaults.

lvs -a -o integritysettings displays configured settings.

2 months agodevices: fix dev_dm_uuid 1397422518
David Teigland [Thu, 1 Aug 2024 16:30:10 +0000 (11:30 -0500)]
devices: fix dev_dm_uuid

If a non-dm device is passed, the dm code doesn't fail and returns
some dm uuid.

2 months agoWHATS_NEW: update 1393308876
Peter Rajnoha [Tue, 30 Jul 2024 08:32:25 +0000 (10:32 +0200)]
WHATS_NEW: update

2 months agotests: add tests for autoswitching to JSON format for log messages 11/head 1385984161 1386311139
Peter Rajnoha [Mon, 3 Jun 2024 12:54:02 +0000 (14:54 +0200)]
tests: add tests for autoswitching to JSON format for log messages

2 months agomake: generate
Peter Rajnoha [Wed, 20 Mar 2024 14:01:45 +0000 (15:01 +0100)]
make: generate

2 months agoreporter: simplify checking output format setting in report_format_init
Peter Rajnoha [Wed, 20 Mar 2024 12:58:04 +0000 (13:58 +0100)]
reporter: simplify checking output format setting in report_format_init

2 months agoconfig_settings: fix typo
Peter Rajnoha [Wed, 20 Mar 2024 14:00:47 +0000 (15:00 +0100)]
config_settings: fix typo

2 months agoconfig: use default log/command_log_report=1 for json/json_std output format
Peter Rajnoha [Wed, 20 Mar 2024 11:41:57 +0000 (12:41 +0100)]
config: use default log/command_log_report=1 for json/json_std output format

log/command_log_report config setting defaults to 1 now if json or json_std
output format is used (either by setting report/output_format config
setting or using --reportformat cmd line arg).

This means that if we use json/json_std output format, the command log
messages are then part of the json output too, not interleaved as
unstructured text mixed with the json output.

If log/command_log_report is set explicitly in the config, then we still
respect that, no matter what output format is used currently. In this
case, users can still separate and redirect the output by using
LVM_OUT_FD, LVM_ERR_FD and LVM_REPORT_FD so that the different types
do not interleave with the json/json_std output.

2 months agoWHATS_NEW 1380264211
Heinz Mauelshagen [Fri, 19 Jul 2024 11:28:30 +0000 (13:28 +0200)]
WHATS_NEW

2 months agolv_manip: avoid unreleased memory pool(s) message on RAID extend 1378848191
Heinz Mauelshagen [Wed, 17 Jul 2024 15:08:20 +0000 (17:08 +0200)]
lv_manip: avoid unreleased memory pool(s) message on RAID extend

In case of different PV sizes in a VG, the lvm2 allocator falls short
to define extended segments resiliently asked for 100%FREE RaidLV extension
and a RAID distinct allocation check fails.  Fix is to release a memory pool
on the resulting error path.

Until the lvm2 allocator gets enhanced (WIP) to do such complex (and other)
allocations proper, a workaround is to extend a RaidLV to any free space on
its already allocated PVs by defining those PVs on the lvextend command line
then iteratively run further such lvextend commands to extend it to its
final intended size.  Mind, this may be a non-trivial extension interation.

2 months agopages: Update release note tag and timestamp 1371117541
Marian Csontos [Fri, 12 Jul 2024 10:50:20 +0000 (12:50 +0200)]
pages: Update release note tag and timestamp

2 months agopost-release 1371114350
Marian Csontos [Fri, 12 Jul 2024 10:42:39 +0000 (12:42 +0200)]
post-release

2 months agopre-release 1371114594 v2_03_25
Marian Csontos [Fri, 12 Jul 2024 10:42:39 +0000 (12:42 +0200)]
pre-release

3 months agoRevert "dev-cache: move global variables into cmd struct" 1365196031
David Teigland [Mon, 8 Jul 2024 20:32:41 +0000 (15:32 -0500)]
Revert "dev-cache: move global variables into cmd struct"

This reverts commit 77be3250d926263db0d4daa83a4102078601bac2.

3 months agoRevert "bcache: move from global variable to cmd struct"
David Teigland [Mon, 8 Jul 2024 20:32:35 +0000 (15:32 -0500)]
Revert "bcache: move from global variable to cmd struct"

This reverts commit 5bf4efbab5f2c51f0ca1763f47a9aa227dfb915c.

3 months agobcache: move from global variable to cmd struct 1359609056
David Teigland [Wed, 3 Jul 2024 17:58:18 +0000 (12:58 -0500)]
bcache: move from global variable to cmd struct

3 months agodev-cache: move global variables into cmd struct 1359579497
David Teigland [Tue, 2 Jul 2024 21:43:34 +0000 (16:43 -0500)]
dev-cache: move global variables into cmd struct

The cmd struct is now required in many more functions, and
it's added as a function arg for most direct dev-cache function
calls.  The cmd struct is added to struct device (dev->cmd) so
that it can be accessed in many other cases where dev-cache
functions are being called from places where getting the cmd
struct is too difficult.

3 months agoconfig_file: remove unused code 1358190087
David Teigland [Tue, 2 Jul 2024 21:24:12 +0000 (16:24 -0500)]
config_file: remove unused code

config_file_read() is only used for regular files, not devices.

3 months agodevices: isolate fake device setup for config file
David Teigland [Tue, 2 Jul 2024 21:07:01 +0000 (16:07 -0500)]
devices: isolate fake device setup for config file

Freeing dev-cache from dumb fake device usage.

3 months agodevices: isolate fake dev use for mirror checks
David Teigland [Tue, 2 Jul 2024 20:07:43 +0000 (15:07 -0500)]
devices: isolate fake dev use for mirror checks

Freeing dev-cache from dumb fake device usage.

3 months agolvmlockd: fix compilation with disabled lvmlockd 1357676816
Zdenek Kabelac [Tue, 2 Jul 2024 14:58:03 +0000 (16:58 +0200)]
lvmlockd: fix compilation with disabled lvmlockd

With updated functionality for lvmlockd some prototypes
needs to be updated for builds without lvmlockd.

3 months agodevices: simpler dm devs cache naming 1356174681
David Teigland [Mon, 1 Jul 2024 16:50:59 +0000 (11:50 -0500)]
devices: simpler dm devs cache naming

The dm devs cache is separate from the ordinary dev cache,
so give the function names distinct prefixes, using
"dm_devs_cache" to prefix dm devs cache functions.

3 months agodevices: redo bcache invalidation of PVs on LV
David Teigland [Mon, 1 Jul 2024 16:30:51 +0000 (11:30 -0500)]
devices: redo bcache invalidation of PVs on LV

3 months agodevices: update dm uuid cache naming
David Teigland [Mon, 1 Jul 2024 15:50:53 +0000 (10:50 -0500)]
devices: update dm uuid cache naming

replace "dm uuid" with "dm devs" in case it begins
caching more than dm uuids.

3 months agoRevert "devices: simpler bcache invalidation of PVs on LVs"
David Teigland [Mon, 1 Jul 2024 15:58:23 +0000 (10:58 -0500)]
Revert "devices: simpler bcache invalidation of PVs on LVs"

This reverts commit 200b05cec0b75ebd1c6860ca7b2233b390cc2024.

Doesn't work to depend on lv name.

3 months agodoc: release notes for 2.03.25 1355843695
Marian Csontos [Mon, 1 Jul 2024 13:33:00 +0000 (15:33 +0200)]
doc: release notes for 2.03.25

3 months agoWHATS_NEW: update
Marian Csontos [Mon, 1 Jul 2024 13:30:40 +0000 (15:30 +0200)]
WHATS_NEW: update

3 months agodevices: simpler bcache invalidation of PVs on LVs 1353285574
David Teigland [Fri, 28 Jun 2024 23:04:25 +0000 (18:04 -0500)]
devices: simpler bcache invalidation of PVs on LVs

When a PV is stacked on an LV, the PV needs to be
dropped from bcache before the LV is processed.
The LV can be found in dev-cache using its name
rather than the devno.

3 months agodev-cache: unify dm uuid cache struct lifetimes
David Teigland [Fri, 28 Jun 2024 22:57:00 +0000 (17:57 -0500)]
dev-cache: unify dm uuid cache struct lifetimes

The list of dm devs was in the cmd struct and had a
different lifetime than the radix trees referencing
those dm devs.  Now the list and radix trees are
created and destroyed together.

3 months agodevices: update function name to dm_device_is_usable 1353107968
David Teigland [Fri, 28 Jun 2024 18:56:12 +0000 (13:56 -0500)]
devices: update function name to dm_device_is_usable

This page took 0.074783 seconds and 5 git commands to generate.