1
0
mirror of https://github.com/samba-team/samba.git synced 2025-01-26 10:04:02 +03:00

648 Commits

Author SHA1 Message Date
Michael Adam
741e902888 s3:registry: streamline reg_deletekey() somewhat.
use W_ERROR_... macros and separate assignments from checks

Michael
2009-02-26 13:22:55 +01:00
Michael Adam
4b444e4ac4 s3:registry: use create_reg_subkey() in reg_createkey().
Instead of calling store_reg_keys()

On my box, importing 2000 shares into an empty smbconf key
now takes some 32 seconds.

Michael
2009-02-26 13:22:55 +01:00
Michael Adam
d73dafb5fe s3:registry: implement create_subkey for the smbconf backend
by delegating to the db backend

Michael
2009-02-26 13:22:54 +01:00
Michael Adam
20e40d3b7d s3:registry: implement create_subkey for the db backend.
Michael
2009-02-26 13:22:54 +01:00
Michael Adam
602bfeb360 s3:registry: add a create_subkey method to the backend ops.
This is to provide a more atomic means of adding a subkey of a key.

Michael
2009-02-26 13:22:54 +01:00
Michael Adam
0b22f8b99c s3:registry: hash the list of subkeys in the regsubkey_ctr
This removes many loops over all the arrays
(from regsubkey_ctr_key_exists) and thus
reduces "net conf drop" from 1m55 to 48seconds
and "net conf import" from 1m55 to 58 seconds
for 2000 shares on my box.

Michael
2009-02-26 13:22:54 +01:00
Michael Adam
149d94dd8f s3:registry: remove definition of regsubkey_ctr from the surface.
All access is now through accessor functions in reg_objects.c
This allows for performance tuning under the hood in the next step.

Michael
2009-02-26 13:22:53 +01:00
Michael Adam
1ce0035abb s3:registry: use regsubkey_ctr_init() in reg_eventlog.c
Michael
2009-02-26 13:22:53 +01:00
Michael Adam
69be6f5c1d s3:registry: use regsubkey_ctr_get_seqnum() in reg_backend_db.c
Michael
2009-02-26 13:22:53 +01:00
Michael Adam
8a74b535c8 s3:registry: use regsubkey_ctr_set_seqnum() in reg_backend_db.c
Michael
2009-02-26 13:22:53 +01:00
Michael Adam
3c15d053f5 s3:registry: use regsubkey_ctr_init() in reg_backend_db.c
instead of using talloc directly.

Michael
2009-02-26 13:22:53 +01:00
Michael Adam
ea2b74090d s3:registry: use regsubkey_ctr_init() in reg_api.c
instead of using talloc on struct regsubkey_ctr.

Michael
2009-02-26 13:22:52 +01:00
Michael Adam
cd8bfd3a84 s3:registry: add regsubkey_ctr_get_seqnum() to hide implementation
Michael
2009-02-26 13:22:52 +01:00
Michael Adam
63ed47c9e4 s3:registry: add regsubkey_ctr_set_seqnum to hide implementation from caller.
Michael
2009-02-26 13:22:52 +01:00
Michael Adam
763f41f39c s3:registry: add a regsubkey_ctr_init function for allocating a regsubkey_ctr
Michael
2009-02-26 13:22:51 +01:00
Michael Adam
060abd7e38 s3:registry: fix a comment
Michael
2009-02-26 13:22:51 +01:00
Michael Adam
63b576424d s3:registry: don't directly access key->subkeys->subkeys[] in reg_backend_db.c
Use the reg_objects accessor regsubkey_ctr_specific_key() instead.

Michael
2009-02-26 13:22:51 +01:00
Michael Adam
8c1e4034fe s3:registry: don't directly access key->subkeys->subkeys[] in reg_api.c
Use the reg_objects accessor regsubkey_ctr_specific_key() instead.

Michael
2009-02-26 13:22:51 +01:00
Michael Adam
62b53507d2 s3:registry: don't directly access key->subkeys->num_subkeys in reg_backend_db.
Use the reg_objects api instead.

Michael
2009-02-26 13:22:51 +01:00
Michael Adam
e05b7a8627 s3:registry: don't directly access key->subkeys->num_subkeys in reg_api.c
Use the reg_objects api instead.

Michael
2009-02-26 13:22:51 +01:00
Michael Adam
58fc61217d s3:registry: replace typedef "REGSUBKEY_CTR" by "struct regsubkey_ctr"
This paves the way for hiding the typedef and the implementation from the
surface.

Michael
2009-02-26 13:22:51 +01:00
Michael Adam
b56f449d6b Revert "Fix a O(n^2) algorithm in regdb_fetch_keys()"
This reverts commit a13f065bad0f4d21a67e68b743f17f45bf0a4691.

This fix is reverted, because the speedup is going to move
further down into reg_objects.c. The unsorted list of subkey names
is going to be indexed: This O(n^2) search bites us in more places.
This re-establishes the abstraction of reg_objects.c.

Michael
2009-02-26 13:22:51 +01:00
Michael Adam
9f97674ef7 s3:registry: wrap deletekey_recursive in one big transaction.
This speeds up "net conf drop" with 2000 shares on my box from
4m40s to 1m50s, leaving virtually only cpu load.

Michael
2009-02-26 11:05:23 +01:00
Michael Adam
ba0e944c46 s3:registry: provide transaction_start|commit|cancel fns for the registry tdb
Michael
2009-02-26 11:05:22 +01:00
Volker Lendecke
170830c008 Wrap creating the sorted subkey cache in a transaction
Signed-off-by: Michael Adam <obnox@samba.org>
2009-02-26 11:05:22 +01:00
Volker Lendecke
5c91cdcc47 Add a comment describing the sorted subkeys
Signed-off-by: Michael Adam <obnox@samba.org>
2009-02-26 11:05:22 +01:00
Volker Lendecke
1b1aac412c Add sorted subkey cache
On my box this gets net conf list for 1000 records down to .1 seconds

Signed-off-by: Michael Adam <obnox@samba.org>
2009-02-26 11:05:22 +01:00
Jeremy Allison
0b62cdbed0 Quieten down a boatload of shadowed variable warnings on Solaris.
Makes real problems easier to spot.
Jeremy.
2009-02-23 15:03:29 -08:00
Jeremy Allison
2ca48d3740 Fix printf type warning.
Jeremy.
2009-02-19 13:11:36 -08:00
Volker Lendecke
a13f065bad Fix a O(n^2) algorithm in regdb_fetch_keys() 2009-02-19 20:06:29 +01:00
Volker Lendecke
d0ee90ace9 Fix a buffer handling bug when adding lots of registry keys
This is *ancient*... From 2002, and nobody noticed until someone added lots of
shares using net conf... :-)
2009-02-19 14:26:50 +01:00
Volker Lendecke
b222de4eb3 Remove a pointless NULL check 2009-02-14 22:01:03 +01:00
Volker Lendecke
4e34c0ceca Fix some nonempty blank lines 2009-02-14 22:01:03 +01:00
Michael Adam
57a290e942 s3:registry: always provision the registry, not only in smbd.
This will make tools like "net registry" also fill the registry
with the basic hives and keys.

Michael
2009-01-28 12:06:26 +01:00
Jelmer Vernooij
08259c1c52 Add iconv_convenience argument to size functions. 2009-01-01 04:45:33 +01:00
Jelmer Vernooij
ddcab787c4 Rename dos_errstr() to win_errstr() for consistency with Samba 4. 2008-11-01 17:19:26 +01:00
Jeremy Allison
8344e94574 Unify se_access_check with the S4 code. Will make
calculation of SEC_FLAG_MAXIMUM_ALLOWED much easier
for files.
Jeremy.
2008-10-31 10:51:45 -07:00
Jelmer Vernooij
2d89b52be8 Use separate make variables for libutil and libcrypto. 2008-10-18 14:28:40 +02:00
Jelmer Vernooij
4746f79d50 Use {u,}int64_t instead of SMB_BIG_{U,}INT. 2008-10-14 01:59:36 +02:00
Jeremy Allison
e5692d4cbe Remove SEC_ACCESS. It's a uint32_t.
Jeremy.
2008-10-09 09:49:03 -07:00
Andrew Tridgell
0f41961e4f first cut at adding full transactions for ctdb to samba3
(This used to be commit f91a3e0f7b7737c1d0667cd961ea950e2b93e592)
2008-08-13 11:54:08 +02:00
Michael Adam
6e16022d62 registry: use _bystring wrappers to dbwrap_trans_(store|delete).
Michael
(This used to be commit 103ce6c9e94ce74e616fe922f2584fd46ae1f3f8)
2008-08-05 23:40:26 +02:00
Michael Adam
a8ae5228f5 registry: bump debug level to 1 in regdb_init() if storing version string fails.
Higher level callers should take care of level 0 messages.

Michael
(This used to be commit b5829e04eaf2408858b34f53b849aaf8b969a925)
2008-07-15 14:25:00 +02:00
Michael Adam
0586d981a7 registry: bump debug level to 1 in regdb_init() if opening the registry fails.
Higher level callers take care of level 0 messages or more drastic measures.

Michael
(This used to be commit 111d802b5c0dfdf556f736b6c53df74e077a6238)
2008-07-15 14:22:55 +02:00
Michael Adam
f587e0310c registry: fix logic in deleting subkeys record in regdb_fetch_keys().
Don't cancel on NT_STATUS_NOT_FOUND error from dbwrap_delete_bystring().
So deletion of an "incomlete" registry key, i.e. one with an entry in
the list of subkeys of its parent key but not a subkey list of its own,
works again.

Michael
(This used to be commit 75be2116ac2589aaf69038a4115197f40e4b16a5)
2008-07-09 10:43:25 +02:00
Michael Adam
ad2e1757e7 registry: improve logic for deleting value records in regdb_store_keys().
Don't ignore all errors from dbwrap_delete_bystring() but
only NT_STATUS_NOT_FOUND.

Michael
(This used to be commit d7ec9b2d52d1eddd98eba222f723fb6cdff4541f)
2008-07-09 10:43:25 +02:00
Michael Adam
70d8e57c17 registry: fetch regdb_fetch_keys() when a key exists but not its subkey-record.
According to the new policy a key (that is not a base key) exists,
iff it exists in the subkey list of its parent key.

Usually this subkeylist is present, but in a transaction-less
dbwrap backend (ctdb), a failing write can leave an "incomplete"
key without its own subkeylist-record. (Otherwise such an
incomplete key can be generated with e.g. tdbtool.)
For such a key net registry enumerate (e.g.) would fail.
This commit fixes this behaviour of regdb_fetch_keys().

Michael
(This used to be commit f329aaf0452cc9bbad9fb6f67dac00bf8d1ef128)
2008-07-08 17:10:01 +02:00
Michael Adam
f7d39653f5 registry: add removal of secdesc of a key to regdb_store_keys().
Michael
(This used to be commit a01d8fe5e89ca9620d23198b11e5ae605a97e5a6)
2008-05-08 18:29:10 +02:00
Michael Adam
1f279f866f registry: add comments to removing lists in regdb_store_keys().
Michael
(This used to be commit 3e89217e7edea5d7ec707de065d38b1921ed09af)
2008-05-08 18:29:10 +02:00
Michael Adam
9f7653bfa1 registry: change order of deleting value and subkey lists in store_keys().
Next step in the plan to make store_keys safer without transactions.

Michael
(This used to be commit f7c8718dead63964481954a86427b8c05d15452c)
2008-05-08 18:29:10 +02:00