Commit Graph

344 Commits

Author SHA1 Message Date
Howard Chu
754cd88819 Add mdb_env_info() to retrieve interesting env params 2012-10-16 16:22:21 -07:00
Howard Chu
7f0771a3ed Merge mdb_stata into mdb_stat 2012-10-16 15:37:29 -07:00
Howard Chu
65d40eb5d2 Make sure increases of mapsize are persisted 2012-10-16 14:53:25 -07:00
Hallvard Furuseth
94965307c8 ITS#7377 Errorcheck Windows calls and thread keys. 2012-10-15 17:04:48 +02:00
Hallvard Furuseth
88b7bb4639 Explicitly unlock the MDB lockfile on Windows. 2012-10-15 17:04:43 +02:00
Hallvard Furuseth
e9d87866c4 mdb_env_open(): Keep mdb_env_set_flags() flags. 2012-10-15 17:04:11 +02:00
Hallvard Furuseth
092f9a5c66 More for mdb_env_copy().
Fix #define _GNU_SOURCE = 1 to match any command line -D_GNU_SOURCE.
.gitignore += mdb_copy.
2012-10-15 17:03:46 +02:00
Hallvard Furuseth
462d9dfd10 Fix last commit.
Add flag MDB_ENV_ACTIVE instead of using the state of me_lfd.
Require MDB_RDONLY for MDB_ROFS.  Swap reader/writer mutex lock.
2012-10-15 17:03:33 +02:00
Howard Chu
e40713b631 Support read access on read-only media 2012-10-13 04:05:00 -07:00
Howard Chu
f037dd2d19 Silence warning in prev commit 2012-10-11 16:47:56 -07:00
Howard Chu
ebb6859ea5 Add mdb_copy for backing up a DB environment 2012-10-11 12:24:01 -07:00
Hallvard Furuseth
756ce8e10c Tweak MDB restrictions 2012-10-03 21:13:21 +02:00
Hallvard Furuseth
a3f33dd20e Fix mdb_open() off-by-one error in maxdbs check 2012-10-03 18:08:56 +02:00
Hallvard Furuseth
38fb8e6eb1 ITS#7377 Wrap fcntl F_SETLK in EINTR loop too.
This can happen even on local filesystems.
2012-10-03 18:06:29 +02:00
Hallvard Furuseth
ab3fea51dc ITS#7377 Catch ftruncate() error 2012-10-03 18:06:29 +02:00
Hallvard Furuseth
2b960613e0 ITS#7377 Document caveats/troubeshooting. 2012-10-03 18:04:07 +02:00
Howard Chu
7333b6bdc9 Fix prev commit
It only worked if the freelist was already non-empty
2012-09-30 10:01:57 -07:00
Howard Chu
cd228b4722 Fix mdb_drop
if it was the only operation in a txn, it would be ignored.
2012-09-30 07:49:30 -07:00
Howard Chu
0c4c6fe72a ITS#7385, check further for bad splits
More for 48ef27b6f5
Our definition of "large" data item needs to be smaller to
handle more cases where a bad split is possible
2012-09-27 05:06:23 -07:00
Howard Chu
ba6c3281a1 Add MDB_GET_CURRENT
return whatever the cursor is currently pointing at
2012-09-19 06:15:09 -07:00
Howard Chu
d13ca0de4f ITS#7394 ignore WRITEMAP if RDONLY 2012-09-18 12:36:54 -07:00
Howard Chu
0a359fb629 More for 48ef27b6f5
page_split with newindex > split_indx
2012-09-17 07:02:41 -07:00
Howard Chu
5ef56b437e More for ab04c50a32
Use explicit MDB_RDONLY flag
2012-09-17 07:01:28 -07:00
Hallvard Furuseth
c67ea9c060 ITS#7377 Catch MDB user errors. 2012-09-17 15:42:15 +02:00
Hallvard Furuseth
c0f3d9b9a8 ITS#7377 Catch MDB setup errors and clean up. 2012-09-17 15:42:15 +02:00
Hallvard Furuseth
fe1b3794de ITS#7364 Limit mdb lock upgrade before sem_unlink.
Do not try shared lock when closing or after error.  Track file lock
state to decide.  Change meaning of *excl to reflect file lock state.
2012-09-17 15:42:15 +02:00
Hallvard Furuseth
c760e536ec ITS#7364 Always sem_unlink() in mdb_env_open().
Drop the sem_unlink() error checks, which could prevent the 2nd
unlink.  Instead use O_EXCL in sem_open().

This makes "open+close the database" the API for trying to clean
away the old semaphores, if they were left behind by a previous run.
2012-09-17 15:42:15 +02:00
Hallvard Furuseth
31be24896b ITS#7377 Wrap sem_wait & file locks in EINTR loops 2012-09-17 15:42:14 +02:00
Hallvard Furuseth
a35f9b2a53 Remove mdb data races. Use (txnid_t)-1 as "no ID".
Avoid race between numreaders++ and reading numreaders at cleanup. Make
the un-mutexed reset of reader table entry, atomic: Reset mr_pid only.

Instead check mr_pid != 0 in mdb_page_alloc()'s scan for readers.
(txnid_t)-1 as "no ID"-mark avoids a check for mr_txnid != 0.
The scan can stop when seeing an old reader.
2012-09-17 15:42:14 +02:00
Hallvard Furuseth
38cc1e96b4 Save pid in MDB_env instead of repeating getpid().
An open MDB environment does not survive or catch fork(),
so repeating getpid() was pointless.
2012-09-17 15:42:14 +02:00
Hallvard Furuseth
20a216fcc9 MDB_WRITEMAP needs no DSYNC descriptor (me_mfd) 2012-09-17 15:42:14 +02:00
Hallvard Furuseth
588a84a5ae ITS#7363 Preprocessor namespace cleanup.
Rename USE_POSIX_SEM to MDB_USE_POSIX_SEM.
Separate MDB_FDATASYNC from MDB_USE_POSIX_SEM.
2012-09-17 15:42:14 +02:00
Howard Chu
8bb10add24 More for ab04c50a32
Fix typos, error code ranges
2012-09-17 06:35:03 -07:00
Howard Chu
acbff5b1ea Add mdb_cursor_renew()
Allow cursors on read-only txns to be reused with later txns.
2012-09-17 04:41:13 -07:00
Howard Chu
076b2b36a9 Shared lib should depend on pthread 2012-09-17 03:48:54 -07:00
Howard Chu
c542442c9f Add MDB_SET_KEY cursor op
Overwrites the passed in key with the DB's key
2012-09-17 02:17:25 -07:00
Howard Chu
ab04c50a32 ITS#7383 More explicit error codes 2012-09-14 06:51:31 -07:00
Howard Chu
48ef27b6f5 ITS#7385 fix mdb_page_split (again)
More for d793594173 ITS#7229
2012-09-12 10:45:51 -07:00
Howard Chu
de77882898 Update comments, bump version
Note read-write vs read-only mmap tradeoffs.
2012-09-07 13:40:07 -07:00
Howard Chu
2e677bcb99 msync should default to fully synchronous 2012-09-06 11:17:25 -07:00
Howard Chu
1c733bdc4e Add msync support for WRITEMAP option 2012-09-05 08:40:17 -07:00
Howard Chu
d8a89464cd Add MDB_WRITEMAP to use a writable mmap 2012-09-05 08:39:46 -07:00
Howard Chu
14a08f877f More for 57b5fbea49
Fix abort of child txn
2012-09-04 06:02:26 -07:00
Hallvard Furuseth
4608cd9552 libmdb: Fix P_INVALID when pgno_t != unsigned long 2012-08-29 06:43:35 +02:00
Hallvard Furuseth
3e3c785034 libmdb: Check sem_open() against SEM_FAILED. 2012-08-29 06:42:52 +02:00
Howard Chu
4bc18a5dd3 In mdb_audit, change assert to fprintf 2012-08-27 14:45:16 -07:00
Howard Chu
85ae5011e7 Fix mdb_audit()
due to a0993354a6
2012-08-26 11:05:45 -07:00
Ondrej Kuznik
bd92b570a8 ITS#7369 mdb_stat: cleanup in case something went wrong. 2012-08-24 05:33:46 -07:00
Howard Chu
a2d1b42e39 ITS#7369 abort txn before exiting 2012-08-24 05:23:55 -07:00
Howard Chu
916995236e ITS#7364 also close the mutexes as needed
Otherwise a process that opens and closes an env more than once
will leak mutex handles. (WIN32/BSD)
2012-08-22 23:01:19 -07:00