1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-21 13:34:40 +03:00
Commit Graph

17820 Commits

Author SHA1 Message Date
Zdenek Kabelac
377e326156 cov: ensuring buffer is zeroed 2021-09-21 21:03:47 +02:00
Zdenek Kabelac
8c44597820 gcc-fanalyzer: zallocate memory for clean buffer 2021-09-21 21:03:47 +02:00
Zdenek Kabelac
4ef8f7cde6 gcc-fanalyzer: remove unneded zeroing
Since we zallocate array here, we don't need to set to NULL
and clean gcc warning about strdup leakage.
2021-09-21 21:03:47 +02:00
Zdenek Kabelac
178df1c8f4 gcc-fanalyzer: helping understand answer is not null
gcc can't easily see pointer 'answer' is only being checked once 'ret > 0'.
2021-09-21 21:03:47 +02:00
Zdenek Kabelac
462ac47003 clang: check for buffer existance
Make obvious to analyzer we can't get NULL here.
2021-09-21 21:03:47 +02:00
Zdenek Kabelac
bb6fd1a15f shellcheck: warning enhancements 2021-09-21 21:03:47 +02:00
Zdenek Kabelac
5bd99d55b0 shellcheck: tests fixes
Correcting some shell problems.
2021-09-21 21:03:47 +02:00
David Teigland
130993bab1 lvmcache: follow pv_id/pvid naming convention
name the variable pv_id_arg to follow the convention:
. "pvid" variable is null-terminated string
. "pv_id" variable is non-null-terminated struct
2021-09-20 09:35:47 -05:00
Zdenek Kabelac
b1f33a4000 tests: len always initialized 2021-09-20 16:24:08 +02:00
Zdenek Kabelac
133686ebc7 gcc: avoid name colision with function pv_id() from metadata.h 2021-09-20 16:12:10 +02:00
Zdenek Kabelac
2e990d1fec gcc: avoid warns on missing braces with older version 2021-09-20 16:11:49 +02:00
Zdenek Kabelac
1a8a33ae9b tests: a bit more fancy way to initialized struct
Older g++ is not smart enough to figure this out, so add a bit more push.
2021-09-20 16:01:22 +02:00
Zdenek Kabelac
4c0354fe0f cleanup: better wording
Follow-up advice of native speaker.
2021-09-20 15:41:14 +02:00
Zdenek Kabelac
1b52451996 cov: mask some warning about leakage
Mask false-positive leak report.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
548c69f581 cov: hide reports from optarg being NULL
It's basically irrelavant which value we assing to optarg,
since it's set by getopt() function, but Coverity tool
is incorrectly reporting possibly dereference of NULL.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
efaab93491 cov: ignore syscall 2021-09-20 14:26:09 +02:00
Zdenek Kabelac
055406bb39 cov: validate syscall result 2021-09-20 14:26:09 +02:00
Zdenek Kabelac
9721f3e0ff cov: test for fixture being NULL pointer
Explicit check for teoretical NULL pointer passed as fixture.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
3f946bfce7 cov: capture runtime exception 2021-09-20 14:26:09 +02:00
Zdenek Kabelac
d5c82b0e73 cov: better defined constructors
Ensure C++ initializes all struct members.
Also some deterministic buffer initialization.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
ebd150366f cov: add explicit NULL pointer check
Make obvious to coverity strcmp() is not getting NULL pointer.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
752a9ec4b4 cov: make read buffer obviously zero ended
Coverity doesn't get the idea from memset() here.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
6c87e98ee3 cov: check for possible NULL segtype
Although likely impossible to ever miss ERROR segtype,
make analyzer hapier.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
bfc6fe2bac cov: ensure data is not derefenced
Ensure all hypothetical paths cannot dereference NULL *(u->data).
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
3e7e9cc5e6 cov: handle better leak reports
Rewrite code so the leaks are eliminated from Coverity.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
825e49a244 cov: use different variable name
Better for Coverity to see no change for possibly opened handle.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
376e707232 cov: better handle reallocation of adapt_file
If the option would be given more then once.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
ce907bcd26 cov: handle teoretical sysconf failure
sysconf() may also return -1 although rather theoretically.
Default to 4K when such case would happen.
Also in function call it just once and keep as static variable.
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
f410035181 cov: check fid instance is created 2021-09-20 14:26:09 +02:00
Zdenek Kabelac
e1840dd9e0 cov: add error path for daemon_request_extend
Check for errors in daemon_request_extend().
2021-09-20 14:26:09 +02:00
Zdenek Kabelac
6427803252 cov: pvck validates setup_device return value 2021-09-20 14:18:13 +02:00
Zdenek Kabelac
b6e9892bc6 tests: use single funtion to get random value 2021-09-20 14:18:13 +02:00
Zdenek Kabelac
a428856a21 tests: fix test of read buffer
Patch 7aba7fe68b incorrectly
converted  SECTOR_SIZE * BLOCK_SIZE_SECTORS to sizeof(f->data).
2021-09-20 14:18:13 +02:00
Zdenek Kabelac
becffe4567 gcc-fanalyzer: drop zeroing of pointers
This explicit pointer zeroing is not needed since
we use 'zalloced' buffers.
2021-09-20 14:18:13 +02:00
Zdenek Kabelac
909401226c gcc-fanalyzer: rewrite _fetch_string
Slightly more optimal code, that is less confusing analyzers.
2021-09-20 14:18:13 +02:00
Zdenek Kabelac
30b188857c gcc-fanalyzer: some explicit NULL checks for tests
Testing code being happier with these extra checks...
2021-09-20 14:17:46 +02:00
Zdenek Kabelac
9512638656 gcc-fanalyzer: better losetup error path
Fix leaking memory on error path on _process_losetup_switches().
However real impact is close to zero as dm-loop target isn't really used.
2021-09-20 13:58:57 +02:00
Zdenek Kabelac
3e21c8524e gcc-fanalyzer: add extra check for origin_from_cow
Make analyzer work easier with explicit check for internal error.
2021-09-20 13:58:57 +02:00
Zdenek Kabelac
93d565eda9 gcc-fanalyzer: use zeroed memory
Analyzer here was rather confused about possiblity of loosing previously
assigned device pointers - fixed by passing zero initialize memory
before first assign.
2021-09-20 13:58:57 +02:00
Zdenek Kabelac
5126ac7c3a gcc-fanalyzer: explicit test null not pass
Make analyzer explicitelly aware we can't get NULL here.
2021-09-20 10:51:30 +02:00
Zdenek Kabelac
e15d1e53c1 gcc-fanalyzer: notices 'du' could be NULL
Ensure only nonNULL 'du' pointer is dereference altough the comment
to the last assign 'du' pointer already suggest 'NULL' case should not happen.
So just being explicit.

mer du
2021-09-20 10:51:30 +02:00
Zdenek Kabelac
834f17bb33 clang: ensure new_len is at least 1 2021-09-20 10:51:30 +02:00
Zdenek Kabelac
0d9f3dbd95 clang: add explicit test for VG
Clang can't figure out connection to 'skip' variable.
2021-09-20 10:51:30 +02:00
Zdenek Kabelac
48163e32f1 lvmlockd: better error path
Set ->num to correct num of pointers.
and fixes regression from previous clang fix
9c5d2874a7.
2021-09-20 10:49:56 +02:00
Zdenek Kabelac
a0bbdfba7f cleanup: typo fix 2021-09-17 22:42:41 +02:00
Zdenek Kabelac
3817392d87 lvm_import_vdo: --dry-run automatically verbose
Dry-run is way more usable when it's verbose.
2021-09-17 16:49:17 +02:00
Zdenek Kabelac
944e560e96 debug: change message level
These message do not cause command error - switch to debug.
2021-09-15 15:24:56 +02:00
Zdenek Kabelac
ed59b50512 cleanup: free already does NULL check
Drop extra check for NULL before calling free().
Also drop zeroing data that will be free().
2021-09-15 15:24:56 +02:00
Zdenek Kabelac
73087c2c0f clang: free on arg parsing
If the parameter would be given twice, the 1st. alloc memleaks.
2021-09-15 15:24:56 +02:00
Zdenek Kabelac
dd5f8b3f8c clang: keep metaname initialized
Never access uninitialized metaname buffer.
2021-09-15 15:24:56 +02:00