1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-30 17:18:21 +03:00
Commit Graph

17496 Commits

Author SHA1 Message Date
Wu Guanghao
262008f1ce unit-test: modify for systems with PGSIZE of 64K
This modification supports dynamically obtaining the value of PAGE_SIZE,
which is compatible with systems with PAGE_SIZE of (4K/64K)

Signed-off-by: wuguanghao <wuguanghao3@huawei.com>
2021-03-11 00:56:43 +01:00
Zdenek Kabelac
236a45dfc5 toollib: preserve return code 5 2021-03-11 00:56:43 +01:00
Zdenek Kabelac
4bbaff1fa6 tests: settle after wiping
Add call to settle after wiping device so there is reduced
change to meet watch rule race during tests.
2021-03-11 00:18:01 +01:00
Zdenek Kabelac
2cbe88d6a1 debug: display_lvname 2021-03-11 00:18:01 +01:00
Zdenek Kabelac
a654148b76 gcc: adding const 2021-03-11 00:18:01 +01:00
Zdenek Kabelac
d7bb23dd15 cov: initialize value
Coverity can't understand conditions together, so just keep it quiet.
2021-03-11 00:18:01 +01:00
Zdenek Kabelac
23afad8247 cleanup: use already set vg_complete
Already set to 0.
2021-03-11 00:18:01 +01:00
Zdenek Kabelac
c4f5d93122 cleanup: eliminate unused assign 2021-03-11 00:18:01 +01:00
Zdenek Kabelac
d86e943b80 toollib: use EINIT_FAILED for missing devs
There is really no practical reason to continue running
when we fail on allocation.

It seems we may need further fine frained errors, as for
some error type we simply need to exit ASAP, while
others may still produce usable results.
2021-03-11 00:18:01 +01:00
Zdenek Kabelac
f4543aca15 lvremove: support faster removal of thin-pools
When lvremove/vgremove removes thin volumes with its thin-pool as well,
try to skip any updates of such thin-pool, so when everything properly
deactivates, there is no message send to this thin-pool and whole
thin-pool is removed with a single commit.
2021-03-11 00:18:01 +01:00
Zdenek Kabelac
dac8274701 toollib: prioritize processing thin-pool first
When generating list of processed LV, add thin-pool to the head of the
list, while other LVs are added on tail.

This makes it easier when removing many thin volumes, to recognize easily
when its thin-pool is also supposed to be removed.
2021-03-11 00:18:01 +01:00
Zdenek Kabelac
7a99dac9ad lvconvert: fix error path on writecache removal
After setting ret=0 there was missing 'goto out_release'.
Simplify with setting ret to 0 at front and to 1
at succesful end.
2021-03-10 23:32:12 +01:00
Zdenek Kabelac
127c2fc6e2 lv_check_not_in_use: correct check
Since lv_info() may return 0 without setting info struct,
make the test correct and even more readable.
2021-03-10 23:32:12 +01:00
Marian Csontos
d4293b579d tests: Skip test killing kernel
Fix the pattern and for now skip all RHEL-8.4 kernels

Bug 1916891
2021-03-10 16:31:04 +01:00
Zdenek Kabelac
6097dfb9ad cov: happier with defined extent_size
Set extent_size with != 0 value so there is not division by 0
in any code path.
2021-03-10 01:35:02 +01:00
Zdenek Kabelac
94712e3233 cov: defined flv 2021-03-10 01:35:02 +01:00
Zdenek Kabelac
a66380ea3e cov: ensure settings is set 2021-03-10 01:35:02 +01:00
Zdenek Kabelac
f7912635a0 cov: ensure buffer is initialized 2021-03-10 01:35:02 +01:00
Zdenek Kabelac
8a03675241 cov: variable initialization 2021-03-10 01:34:58 +01:00
Zdenek Kabelac
bee9b5c1d8 cov: mask uninitialized value
Coverity doesn't track ioctl() too well, so let's just make it quiet.
2021-03-10 01:34:27 +01:00
Zdenek Kabelac
74936f53f7 cov: unused value 2021-03-10 01:29:44 +01:00
Zdenek Kabelac
d95c0e977c cov: remove unnecessary headers 2021-03-10 01:29:44 +01:00
Zdenek Kabelac
a6075fe2f2 cov: memleak on error path 2021-03-10 01:29:44 +01:00
Zdenek Kabelac
241c63f7a7 cov: fix iter memleak in vgimportclose
Add missing release of iterator on error path.
2021-03-10 01:29:06 +01:00
Zdenek Kabelac
131ca0eb95 activation: use existing LV as best effort
Returning NULL for lv_committed is basically instant crash,
so instead try with passed LV instead.
It shouldn't matter as this is internall error path anyway,
but coverity should be happier.
2021-03-10 01:29:06 +01:00
Zdenek Kabelac
15c588f931 gcc: match size of holder name to direcnt d_name
As d_name is defined as 256 avoid gcc warning - although
in our case the used size will be much smaller.
2021-03-10 01:28:08 +01:00
Zdenek Kabelac
75037bee5d debug: more tracing
Check result of device_ids_write() and at least provide stack;
2021-03-10 01:27:13 +01:00
Zdenek Kabelac
d01c17ff22 debug: more use of display_lvname 2021-03-10 01:11:52 +01:00
Zdenek Kabelac
7342ab06fc debug: change sys_error to sys_debug
These messages do not cause command error - so changing logging level
to just 'sys_debug' (so visible only with -vvvv)
2021-03-10 01:11:52 +01:00
Zdenek Kabelac
5f7a7af7f2 cleanup: no backtraces needed after log_error
Reduce double backtracing.
2021-03-10 01:11:52 +01:00
Zdenek Kabelac
c05077f4d4 cleanup: easier print of historical lv
Drop unneeded macros.
2021-03-10 01:11:52 +01:00
Zdenek Kabelac
177b63becc backup: set in vg_commit
Another step towards better automatic handling of backup,
and automatically setup needs_backup after commit.

In some next step we should reduce number of backups and takem
then only at the command finish with  vg_committed content.
2021-03-10 01:09:46 +01:00
Zdenek Kabelac
88d04e9173 lvm-file: remove duplication of dm_is_empty_dir 2021-03-10 01:09:14 +01:00
Zdenek Kabelac
413a114cdb thin: correct ptr test for thin snapshot merging
The correct test needs to actually check  'lv->snapshot' is not NULL,
so the 'find_snapshot()' can work.

Test lv_is_snapshot was actually irrelavant for this case.

Also initialize device_id.
2021-03-10 01:09:14 +01:00
Zdenek Kabelac
b2616cd456 device_mapper: keep bigger size for most ioctls
Actually the idea was to keep 2K ioctl only for certain command,
so keep 16 for them.
2021-03-10 01:07:16 +01:00
Zdenek Kabelac
d2a3bfe6ca vg_write: shift check for exported metadata
Since we now immediatelly use created buffer for CFT creation,
the result needs to be validate before such use.
2021-03-10 01:00:18 +01:00
Zdenek Kabelac
17802084c9 bcache: fix incorrect pointer check
With commit b44db5d1a7
needs to check allocated pointer for failed malloc().

Existing check was actually no checking anything so failing
malloc here would result in segfault (although with very
low chance to ever happen).
2021-03-10 00:59:05 +01:00
David Teigland
d5f3239de4 tests: check_devicesfile
make check_devicesfile T=...

(A number of tests should probably get a new
SKIP_WITH_DEVICES_FILE.)
2021-03-08 17:19:26 -06:00
Zdenek Kabelac
843ee943ab lvremove: correct return code
Need to return ECMD_FAILED from toollib code.
Add missing stack traces.
2021-03-08 20:24:04 +01:00
Zdenek Kabelac
3cbb751815 WHATS_NEW: updates 2021-03-08 15:43:27 +01:00
Zdenek Kabelac
6d6e1ae887 cleanup: compare only LV uuid part
Match VG uuid just once per list of all LVs in VG.

TODO: maybe some more efficeint tree or hash could be better here,
but since it's used not so often, the total benefit is not so great,
so ATM just reducing amount of checked bytes.
2021-03-08 15:43:27 +01:00
Zdenek Kabelac
64447e9d9b cleanup: move code
just evaluate later in code path.
2021-03-08 15:43:27 +01:00
Zdenek Kabelac
e5456c259f cleanup: simpler checks first
Minor optimizatoins...
2021-03-08 15:43:27 +01:00
Zdenek Kabelac
f4200acac2 cleanup: on stack structure instead of allocation 2021-03-08 15:43:27 +01:00
Zdenek Kabelac
dac990ae03 cleanup: avoid allocating memory for hash key
Hash always allocates its own copy of the key so avoid
this extra uuid copy.
2021-03-08 15:43:27 +01:00
Zdenek Kabelac
d3cff64408 libdm-config: replace check for 0
No need to call strlen() when checking for zero length string.
2021-03-08 15:43:27 +01:00
Zdenek Kabelac
ca12dae32b hints: keep strings aligned in structure
Preffer aligned string access.
2021-03-08 15:33:15 +01:00
Zdenek Kabelac
ff21723512 hash: replace hash with better function
Add Bob Jenkins hash function to get better working hash function,
which does genarate way less colisions (especially with similar
strings).

For a comparision also a kernel function used in DM kernel is include.
While it's better then our existing one, it's still far worse,
then Bob Jenkins hash.
2021-03-08 15:33:15 +01:00
Zdenek Kabelac
d602837b91 hash: speed up hash tables
Enhance hash perfomance by remembering the computed
hash value for the hashed node - this allows to speedup
lookup of nodes with the hash collision when
different keys map to the same masked hash value.

For the easier use 'num_slots' become 'mask_slots',
so we only add '1' in rare case of need of the original
num_slots value.

Also add statistic counters for hashes and print stats in
debug build (-DDEBUG) on hash wiping.
(so badly performing hash can be optimized.)
2021-03-08 15:33:15 +01:00
Zdenek Kabelac
84679d254f hash: use unsigned size
There is not much point in using 64bit hash size, since we hash
with way less bits anyway. So keep size 32bit.
2021-03-08 15:33:15 +01:00