IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
* Remove unused select backend
* Fix a race condition in tevent_threaded_schedule_immediate()
(bug #13130)
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Mon Nov 13 18:02:46 CET 2017 on sn-devel-144
We can't rely on tctx to exist after we unlocked the mutex. It took a
while, but this does lead to data corruption. If *tctx is replaced with
something where tctx->wakeup_fd points to a real, existing file
descriptor, we're screwed. And by screwed, this means file corruption
on disk.
Again. I am not tall enough for this business.
http://bholley.net/blog/2015/must-be-this-tall-to-write-multi-threaded-code.html
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13130
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Nov 11 03:20:09 CET 2017 on sn-devel-144
This ensures that we correctly detect HAVE_IFACE_GETIFADDRS
et al, which are based on a "build the source" style test.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13087
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
The env and endptr pointers need to be dereferenced, but that is not
enough: we don't really want to regard an empty string (*env == '\0')
as a valid number.
Found by GCC 8.0.0 20170705 (experimental).
[2095/4103] Compiling lib/nss_wrapper/nss_wrapper.c
../lib/nss_wrapper/nss_wrapper.c: In function "nwrap_init":
../lib/nss_wrapper/nss_wrapper.c:1571:13: warning: comparison between pointer and zero character constant [-Wpointer-compare]
if (((env != '\0') && (endptr == '\0')) ||
^~
../lib/nss_wrapper/nss_wrapper.c:1571:9: note: did you mean to dereference the pointer?
if (((env != '\0') && (endptr == '\0')) ||
^
../lib/nss_wrapper/nss_wrapper.c:1571:33: warning: comparison between pointer and zero character constant [-Wpointer-compare]
if (((env != '\0') && (endptr == '\0')) ||
^~
../lib/nss_wrapper/nss_wrapper.c:1571:26: note: did you mean to dereference the pointer?
if (((env != '\0') && (endptr == '\0')) ||
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Oct 19 16:42:17 CEST 2017 on sn-devel-144
Add a pre fork process model to bound the number processes forked by
samba. Currently workers are only pre-forked for the ldap server, all
the other services have pre-fork support disabled.
When pre-fork support is disabled a new process is started for each
service, and requests are processed by that process.
This commit partially reverts commit
b5be45c453.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
PyErr_NewExceptionWithDoc() isn't available in Python 2.6 so it can
be used only in higher versions of Python.
Signed-off-by: Lumir Balhar <lbalhar@redhat.com>
Reviewed-by: Andrew Bartlet <abartlet@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
All supported versions of Heimal already have KRB5_ADDRESS_INET6,
so there's no need for an explicit check.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13079
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Björn Jacke <bjacke@samba.org>
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13062
Pair-programmed-with: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Sep 30 12:14:57 CEST 2017 on sn-devel-144
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13062
Pair-programmed-with: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13062
Pair-programmed-with: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
Ensure the state desctructor calls tfork_destroy to reap the waiter and
worker processes. Otherwise we leave the waiter process as a zombie
behind us as we never call waitpid on it in case of a timeout
or talloc_free() from the caller.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13062
Pair-programmed-with: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
On a system that has older lttng-ust development headers installed that
do not yet provide the tracef api, configure fails with:
Checking for library lttng-ust : yes
ERROR: Target 'lttng-ust' in directory /samba/lib/util re-defined as EMPTY - was SYSLIB
Fix the initialization order in waf, to not redefine the lttng-ust
target
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Display a meaningful process title for the waiter process, now that smbd is
calling setproctitle_init() and this value will be displayed in ps
output.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Tue Sep 26 04:38:59 CEST 2017 on sn-devel-144
Call setproctitle_init() in main which suppresses the
"samba: setproctitle not initialized, please either call
setproctitle_init() or link against libbsd-ctor."
messages, but more importantly it displays meaningful details in ps
output.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=9816
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
* GUID Index support.
NOTE: When activated by setting @IDXGUID in the @INDEXLIST dn, all
entries in the DB are re-keyed in a way that is NOT visible to
ldb 1.2.2 and earlier. To re-key back to the previous format, remove
the @IDXGUID attribute from @INDEXLIST using ldb 1.2.2 or later.
(ldb 1.2.2 can re-key, but not otherwise read, the new DB format).
* Give LDB_ERR_CONSTRAINT_VIOLATION, not LDB_ERR_ENTRY_ALREADY_EXISTS
when a duplicate value is detected in a unique index
* Print status information during a > 10,000 entry re-index
(as this can be slow)
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Sat Sep 23 01:24:19 CEST 2017 on sn-devel-144
This avoids loading any second index for these cases.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
The python TestCase API will keep a reference to the test object until the end
of the tests, long after we need the actual LDB or the fd.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
An ldb context is valid without a backing file for tests of ldb.Message and ldb.MessageElement
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
This is not actually a great test, as the filter would
fail to match these anyway, but it at least checks the
codepath is safe.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
A re-index of 10,000 entries is slow enough and rare enought that we can
justify the message being at LDB_DEBUG_WARNING as otherwise the administrator
will be sure the "lockup" was one.
The default for ldb is to print LDB_DEBUG_WARNING in comand-line tools
and the default for Samba is to log it at level 2.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The ltdb->cache->attribute_indexes test is not correct with the GUID index mode
so for consistency remove it. This will make re-index on a large un-indexed
database slower, but that is better than making the wrong choice on a large
GUID-indexed database.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows easy testing of our unique index code and behaivour from python
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This was (unintentionally) disabled by
6ef6182554 in 2006.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This is important with the GUID index, as a DN lookup is much more common now.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The casefolding of a DN is one of the more expensive and pointless things in LDB
operation. The ldb_dn abstraction works hard to avoid duplicating this work, but
we can work harder to save that information.
Here we copy in the DN, that has been casefolded already for the index,
and keep that as the returned DN, after stripping any extended components.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This ensures we cover the case where the DN does not match the DB exactly
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This should ensure that the upper or lower case the user chooses does not impact
on the filtering, at least for database that have checkBaseOnSearch set.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
We will never have more results than is in either list or list2.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This change ensures we walk the short list and look up into the longer of the two lists.
ltdb_dn_list_find_val() will do a binary search for the GUID case.
Before GUID indexes this was O(n*m), now it is O(n*log(m)).
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This only works if we have the GUID index format, as otherwise these are unsorted.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows the binary search to still operate on the list, even after
a or operator in the search expression
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows us to merge the lists finding common values.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The LDB_ERR_ENTRY_ALREADY_EXISTS error code is detected in repl_meta_data as indicating
that the DN exists, and that a conflict record should be created.
This is really a constraint violation, not a duplicate record.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will give us errors earlier if the index code becomes broken
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This is unlikely, but when it happens it will be really painful to debug.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows all the previous patches to be enabled.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This connects the GUID based index records to GUID based TDB keys.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This avoids re-checking the fetched DN against the scope
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will allow a common point to parse index records into a TDB key,
allowing them to be a GUID or DN in the future
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This would be pointless and we no longer query for it.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The objectGUID (or similar) is already the record key, there is
no need to index it to itself.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
When we have the full ldb_message we can read the objectGUID as the TDB key
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows the optional use of GUID based TDB key.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This follows modern Samba coding style where memory
returned is allocated on a supplied memory context.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows us to format the TDB key as DN=GUID=f7c953ee-cf9c-433f-b423-21ce04d09591
and so be compatible with an un-indexed search and a re-index with an old ldb.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows us to use the ldif_write function later to create a string GUID for the TDB key.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The @IDXGUID attribute in the @INDEXLIST will be objectGUID
in Samba.
The @IDX_DN_GUID attribute in the @INDEXLIST will be GUID
in Samba.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
With the GUID index option, the values in the index result list may
not be a DN but the objectGUID. We look up the @IDXDN index with the
case-folded DN to get that if required.
We re-use the code from the SCOPE_BASE search do avoid duplication
and for reliablity.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
This will allow the code to be re-used for storing the DN->GUID index
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will allow use of a GUID TDB key in the future. While ltdb_search_base()
might be marginally slower than tdb_exists(), no allocation is done for the
attributes or DN, and renmaes are not a very common operation.
This allows a check if the target DN exists even when the direct DN -> key
link is broken.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The GUID index case can not directly use ltdb_key_dn() and tdb_exists() to
show that a records exists.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows, when enabled, the index record to contain (say) the objectGUID, not the DN
of the record.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This function is used to find an existing index value and this
change allows it to find the value by GUID rather than by DN once
the GUID index is enabled.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows the objectGUID, rather than the DN, to be the index key
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will make it easier to delete records with the GUID TDB key
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Nothing reads these currently, but we should refuse to load a mixed up index
in the future
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This packing should be more efficient to read than the ldb_pack format.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This replaces dn_list_cmp() with functions that do not attempt to
to care about string termination. All index values are case sensitive
and correctly length-bound already, even for a DN index
so just use a length check and memcmp()
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This would totally break our index scheme if this could be modified.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will be used to determine if we are in GUID index mode
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will make it easier to switch the GUID index mode on and off
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This allows us to slowly split out the tdb key in the DB from being the DN
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will allow changing to a GUID tdb key in the future
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This will in time allow us to generate a TDB key from
the msg, eg from an objectGUID.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
This avoids an extra DB lookup for the base, when that is the only
record we will return, and avoids going into the index code for
a base search, as that won't work for special DNs once the GUID
index mode is enabled.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
* Bug #13033 LDB open with LDB_FLG_RDONLY can cause the database
to fail to open
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13033
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
As the kernel is no longer enforcing the read-only DB
add some tests.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13033
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
We support opening and LDB multiple times in a process, but do not support this in tdb.
As we can open the ldb with different flags, we must ensure a later read-write
open is possible.
Additionally, a read-only TDB will refuse the all-record lock, preventing
the ldb from even loading.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13033
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
This is a serious condition, and should be logged.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13033
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
The ltdb_lock_read() routine did not return an LDB error code, but -1.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13033
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Free the fde in the event handler to prevent the event triggering again
While not strictly necessary in this case, this code serves as an
example of the usage of tfork.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13037
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Sep 16 23:50:27 CEST 2017 on sn-devel-144
Make closing of the event_fd the global responsibility of the
parent process if it called tfork_event_fd().
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13037
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
The previous design relied on only calling close() of the status pipe.
We now write a single 0 byte to the status FD as well as closing it in the
parent process. Both of these operations typically trigger a read
event on the other end of the FD, held in the waiter process (the child).
The child process blocks on the status FD, until it becomes readable.
However if there is a sibling process that was launched after the waiter
process they also will hold the status FD open and the status FD would,
until this change, never become readable to the waiter process (the child).
This caused the waiter process (child) not to exit and the parent process
to hang in tfork_status() while expecting the waitpid() to return.
That is, file descriptors are essentially global variables copied
to children in the process tree. The last child that (unwittingly) holds
the file descriptor open is the one that needs to trigger the close() this
code previously depended on.
Without this change, there is no notification of process death until
all these unrelated children exit for their own reasons.
We can write up to 4K (PIPE_BUF) into this pipe before blocking,
but we only write one byte. Additionally sys_write() refuses to block.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13037
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Add tests to ensure that:
- The event_fd becomes readable once the worker process has terminated
- That the event_fd is not closed by the tfork code.
- If this is done in tevent code and the event fde has not been
freed, "Bad talloc magic value - " errors can result.
- That the status call does not block if the parent process launches
more than one child process.
- The status file descriptor for a child is passed to the
subsequent children. These processes hold the FD open, so that
closing the fd does not make the read end go readable, and the
process calling status blocks.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13037
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
select() is no longer useful on modern systems.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Sat Sep 16 08:35:39 CEST 2017 on sn-devel-144
The commits c615ebed6e3d273a682806b952d543e834e5630d^..f19ab5d334e3fb15761fb009e5de876dfc6ea785
replaced Str[n]CaseCmp() by str[n]casecmp_m().
The logic we had in str[n]casecmp_w() used to compare
the upper cased as well as the lower cased versions of the
characters and returned the difference between the lower cased versions.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13018
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Sep 15 02:23:29 CEST 2017 on sn-devel-144
Signed-off-by: Justin Maggard <jmaggard@netgear.com>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Sep 8 06:26:52 CEST 2017 on sn-devel-144
This is used in the client and in the server
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
* Bug #13017: Add ldb_ldif_message_redacted_string() to allow debug
of redacted log messages, avoiding showing secret values
* Bug #13015: Allow re-index of newer databases with binary GUID TDB keys
(this officially removes support for re-index of the original
pack format 0, rather than simply segfaulting).
* Avoid memory allocation and so make modify of records in ldb_tdb faster
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
This is designed to be a drop in replacement for
ldb_ldif_message_string() while better protecting privacy.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13017
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
These are not found on any AD DC, and would segfault previous LDB
versions.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13015
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
If backported, this allows old ldb versions to full-search and re-index newer databases
and in current code allows the transition to and from a GUID or incrementing ID based index
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13016
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
We want to rename the objects, then scan looking for the index values.
This avoids a DB modify during the index scan traverse (the index values
are actually added to an in-memory TDB, written in prepare_commit()).
This allows us to remove the "this might already exist" case in the
index handling, we now know that the entry did not exist in the index
before we add it.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13015
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
The keys may not always be a null terminated string, they could well
be a binary GUID in a future revision, for efficiency..
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13016
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Add --accel-aes=[none|intelaesni] to select.
Default is none.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13008
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Causes:
AES_set_encrypt_key()
AES_set_decrypt_key()
AES_encrypt()
AES_decrypt()
to probe for the Intel AES instructions at runtime (only once)
and then call the hardware implementations if so, otherwise
fall back to the software implementations.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13008
Based on original work by Justin Maggard <jmaggard@netgear.com>
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Rename the old struct aes_key as an intermediate struct aes_key_rj
and wrap it in a union so we can chose an alternate aes_key struct
when using Intel AES hardware.
Rename the original software implementations of:
AES_set_encrypt_key()
AES_set_decrypt_key()
AES_encrypt()
AES_decrypt()
by adding an _rj on the end, and call them via a wrapper
function.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13008
Based on original work by Justin Maggard <jmaggard@netgear.com>
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This allows to us to have restricted access to the directory by the group
'named' which bind is a member of.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12957
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlet <abartlet@samba.org>
glibc's pthread_cond_wait(&c, &m) increments m.__data.__nusers, making
pthread_mutex_destroy return EBUSY. Thus we can't allow any thread waiting for
a job across a fork. Also, the state of the condvar itself is unclear across a
fork. Right now to me it looks like an initialized but unused condvar can be
used in the child. Busy worker threads don't cause any trouble here, they don't
hold mutexes or condvars. Also, they can't reach the condvar because _prepare
holds all mutexes.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13006
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Wed Aug 30 14:58:32 CEST 2017 on sn-devel-144
This avoids duplicate code and allows us to use the allocation-avoiding
LDB_UNPACK_DATA_FLAG_NO_DATA_ALLOC flag.
We can not use LDB_UNPACK_DATA_FLAG_NO_VALUES_ALLOC as el2->values
is talloc_realloc()ed in the routine.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Tue Aug 29 11:13:50 CEST 2017 on sn-devel-144
This will allow us to avoid a full unpack in situations where we just want to confirm
if the DN exists
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
* Add protection against EINTR.
* Truncate the file after expand failure, ENOSPC
* Use posix_fallocate() to expand the file
* Fix GCC compiler warnings
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Aug 24 21:17:48 CEST 2017 on sn-devel-144
This should be significantly faster than pwriting.
openbsd doesn't have posix_fallocate, so we do need the fallback. Also, it
might have weird failure modes, so we keep the old code in place except for
posix_fallocate returning success or ENOSPC.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Aug 24 05:38:49 CEST 2017 on sn-devel-144
More README.Coding, but I need "ret" in the next commit as well :-)
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Without this it's very easy to create virtually huge files: ftruncate expands a
file, the pwrites fail with ENOSPC, thus the write fails. The next writer runs
into the same situation, and ftruncate-expands the file even further. tdb_check
will then spend ages reading the 4GB of zeros byte by byte.
Here we hold the freelist lock or are inside a transaction, so it is safe to
cut the file again. Nobody can have used the space that we have tried to
allocate, so we can't have any stray pointers corrupting the database.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This checks for posix_fallocate unless we are sitting on an ancient glibc.
With this we don't need HAVE_BROKEN_POSIX_FALLOCATE anymore,
HAVE_POSIX_FALLOCATE will only be defined if we have a valid [g]libc.
./configure tested on Debian, FreeBSD (which does have posix_fallocate) and
OpenBSD (which does not have posix_fallocate). Also tested with changing the
not have an old-enough glibc around. All did the right thing.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Due to the non-fixable bug in the BUCKET macro tdbtool list printed some
other hash chainlist, not the freelist.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12888
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The following C program demonstrates the issue:
#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <stdbool.h>
#include <string.h>
#include <unistd.h>
#include <errno.h>
#include <dirent.h>
#include <sys/types.h>
int main(int argc, char **argv)
{
int hash = -1;
int tsize_signed = 10;
unsigned int tsize_unsigned = 10;
int bucket;
#define BUCKET(hash, tsize) ((hash) % (tsize))
bucket = BUCKET(hash, tsize_unsigned);
printf("hash [%d] tsize [%d] bucket [%d]\n", hash, tsize_unsigned, bucket);
bucket = BUCKET(hash, tsize_signed);
printf("hash [%d] tsize [%d] bucket [%d]\n", hash, tsize_signed, bucket);
return 0;
}
Output:
$ ./tmp
hash [-1] tsize [10] bucket [5]
hash [-1] tsize [10] bucket [-1]
The first version is what the current BUCKET() macro does. As a result
we lock the hashtable chain in bucket 5, NOT the freelist.
-1 is sign converted to an unsigned int 4294967295 and
4294967295 % 10 = 5.
As all callers will lock the same wrong list consistently locking is
still consistent.
Stumpled across this when looking at the output of `tdbtool DB list`
which always printed some other hashchain and not the freelist.
The freelist bucket offset computation doesn't use the BUCKET macro in
freelist.c (directly or indirectly) when working on the freelist, it
just directly uses the FREELIST_TOP define, so this problem only affects
tdbtool list.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The variable stores the hashtable bucket, not the hash. No change in
behaviour.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This is another level of indentation, but it took me a while staring at the
if-condition to find that "locked" was assigned the result of "==0", not the
return value of tdb_nest_lock().
Best viewed with "git show -b".
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Log failure and exit if fork() or setsid() fails.
Leave the logic in the non-setsid() code as it is. This is probably
meant to fall through on failure of either opening /dev/tty or
ioctl(). Documentation for the ioctl() failure case is far from
clear.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Aug 17 11:48:32 CEST 2017 on sn-devel-144
Switch to using DBG_ERR(), wrap logging/sd_notifyf() lines.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Andreas Schneider <asn@samba.org>
no_process_group -> no_session, name -> daemon, drop _PUBLIC_.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Andreas Schneider <asn@samba.org>
Rename argument no_process_group to no_session to describe what it
actually does. Consistently use "daemon" for name of daemon argument.
Add documentation.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Andreas Schneider <asn@samba.org>
This fixes a GCC warning.
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Aug 10 02:26:09 CEST 2017 on sn-devel-144
This fixes CID 1415704
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Aug 3 15:06:34 CEST 2017 on sn-devel-144
Uses the core of CTDB's create_pidfile_context() for
pidfile_path_create(). pidfile_fd_close() is a subset of CTDB's
pidfile_context_destructor().
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
This causes a race. If 2 callers to pidfile_create() both a find a
stale PID file using pidfile_pid(). The 1st may then return to
pidfile_create() and create a new PID file, which can then be unlinked
by the 2nd caller.
Consequently, PID file creation can not depend on creating the file,
so drop O_EXCL from the call to open().
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
The pytalloc-util dependency logic in lib/talloc/wscript on a
standalone build checks for pytalloc-util in a manner that will
fail if bundling is disabled, this causes issues on
--disable-python builds of ldb, tevent, and samba.
This patch restructures the logic to skip checks if python
is disabled, instead just setting the temporary state variable
'using_system_pytalloc_util' to False
Successfully tested patch on ldb-1.1.31 and above, tevent-0.9.33,
and samba-4.7_rc3
Signed-off-by: Ian Stakenvicius <axs@gentoo.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
ldb modules are not (yet) unloaded and are only loaded once (there is a check
that makes sure of this). Allocate off the NULL context. We never want this
to be freed until process shutdown. If eventually we add the ability to
unload ldb modules we can add a deregister function that walks and frees the list.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12932
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
With a proper implementation this enables modifications without
having to allocate a record. In really performance sensitive code
paths this matters.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This will allow dbwrap_do_locked to check the order without talloc
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Convert all implementors of dbwrap_store to a storev-style call
by using the dbwrap_merge_dbufs call
For dbwrap_tdb, this matches tdb_storev.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
* Bug #12882: Do not install _ldb_text.py if we have system libldb
* Use libraries from build dir for testsuite
* Bug #12900: Fix index out of bound in ldb_msg_find_common_values
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Jul 22 03:46:25 CEST 2017 on sn-devel-144
* make tevent_req_print() more robust against crashes
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
We have the same information available under req->internal.private_type.
This way it's possible to call tevent_req_print() after
tevent_req_received() was called.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Jul 7 20:10:37 CEST 2017 on sn-devel-144
cmocka unit test failed on i386
[==========] Running 2 test(s).
[ RUN ] test_ldb_msg_find_duplicate_val
[ OK ] test_ldb_msg_find_duplicate_val
[ RUN ] test_ldb_msg_find_common_values
[ FAILED ] test_ldb_msg_find_common_values
[==========] 2 test(s) run.
[ ERROR ] --- 0x14 != 0
[ LINE ] --- ../tests/ldb_msg.c:266: error: Failure!
[ PASSED ] 1 test(s).
[ FAILED ] 1 test(s), listed below:
[ FAILED ] test_ldb_msg_find_common_values
1 FAILED TEST(S)
But we were just lucky on other platforms because there is
index out of bound according to valgrind error.
==3298== Invalid read of size 4
==3298== at 0x486FCF6: ldb_val_cmp (ldb_msg.c:95)
==3298== by 0x486FCF6: ldb_msg_find_common_values (ldb_msg.c:266)
==3298== by 0x109A3D: test_ldb_msg_find_common_values (ldb_msg.c:265)
==3298== by 0x48E7490: ??? (in /usr/lib/libcmocka.so.0.4.1)
==3298== by 0x48E7EB0: _cmocka_run_group_tests (in /usr/lib/libcmocka.so.0.4.1)
==3298== by 0x1089B7: main (ldb_msg.c:352)
==3298== Address 0x4b07734 is 4 bytes after a block of size 48 alloc'd
==3298== at 0x483223E: malloc (vg_replace_malloc.c:299)
==3298== by 0x4907AA7: _talloc_array (in /usr/lib/libtalloc.so.2.1.9)
==3298== by 0x486FBF8: ldb_msg_find_common_values (ldb_msg.c:245)
==3298== by 0x109A3D: test_ldb_msg_find_common_values (ldb_msg.c:265)
==3298== by 0x48E7490: ??? (in /usr/lib/libcmocka.so.0.4.1)
==3298== by 0x48E7EB0: _cmocka_run_group_tests (in /usr/lib/libcmocka.so.0.4.1)
==3298== by 0x1089B7: main (ldb_msg.c:352)
Signed-off-by: Lukas Slebodnik <lslebodn@redhat.com>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
_ldb_text.py is installed as part of the ldb package and also if you
compile Samba with the system ldb version. This way we have have the
file twice in the same location and run into file confilcts.
This has already been fixed some time ago:
60dc26bfe1
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12882
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
There was a failure when tests were executed after after extracting
ldb tarball.
sh$ make -j8 check
WAF_MAKE=1 PATH=buildtools/bin:../../buildtools/bin:$PATH waf test
ldbadd: error while loading shared libraries: libldb.so.1: cannot open shared object file: No such file or directory
cat: write error: Broken pipe
Traceback (most recent call last):
File "tests/python/api.py", line 10, in <module>
import ldb
ImportError: libldb.so.1: cannot open shared object file: No such file or directory
Traceback (most recent call last):
File "tests/python/api.py", line 10, in <module>
import ldb
ImportError: libpyldb-util.so.1: cannot open shared object file: No such file or directory
bin/ldb_tdb_mod_op_test: error while loading shared libraries: libldb.so.1: cannot open shared object file: No such file or directory
testsuite returned 1
Signed-off-by: Lukas Slebodnik <lslebodn@redhat.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Signed-off-by: Alexander Bokovoy <ab@samba.org>
make check failed in case of tarball because test_magic_differs.sh
is in top level directory and not in sub-directory lib/talloc
sh: ./lib/talloc/test_magic_differs.sh: No such file or directory
magic differs test returned 127
Signed-off-by: Lukas Slebodnik <lslebodn@redhat.com>
Reviewed-by: Andrew Bartlet <abartlet@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
There was a failure when tests were executed after after extracting
talloc tarball.
sh$ make -j8 check
WAF_MAKE=1 PATH=buildtools/bin:../../buildtools/bin:$PATH waf test
bin/talloc_testsuite: error while loading shared libraries: libtalloc.so.2: cannot open shared object file: No such file or directory
sh: ./lib/talloc/test_magic_differs.sh: No such file or directory
Traceback (most recent call last):
File "test_pytalloc.py", line 11, in <module>
import talloc
ImportError: libtalloc.so.2: cannot open shared object file: No such file or directory
Signed-off-by: Lukas Slebodnik <lslebodn@redhat.com>
Reviewed-by: Andrew Bartlet <abartlet@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
The ntlm auth parameter is expanded to more clearly describe the
role of each option, and to allow the new mode that permits MSCHAPv2
(as declared by the client over the NETLOGON protocol) while
still banning NTLMv1.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12252
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Based on a patch by Mantas Mikulėnas <mantas@utenos-kolegija.lt>:
Commit 0b500d413c ("Added MSV1_0_ALLOW_MSVCHAPV2 flag to ntlm_auth")
added the --allow-mschapv2 option, but didn't implement checking for it
server-side. This implements such checking.
Additionally, Samba now disables NTLMv1 authentication by default for
security reasons. To avoid having to re-enable it globally, 'ntlm auth'
becomes an enum and a new setting is added to allow only MSCHAPv2.
Signed-off-by: Mantas Mikulėnas <mantas@utenos-kolegija.lt>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
No change in behaviour, this just ensures stdout and stderror are
logged with log level 0.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This makes it possible to use samba_runcmd_send() in processes like smbd
that install a SIGCHLD handler that reaps all terminated children.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This function is a solution to the problem of fork() requiring special
preperations in the caller to handle SIGCHLD signals and to reap the
child by wait()ing for it.
Instead, tfork provides a pollable file descriptor. The caller gets the
file descriptor by calling tfork_event_fd() on the handle returned from
tfork_create() and the caller can then get the status of the child
with a call to tfork_status().
tfork avoids raising SIGCHLD signals in the caller by installing a
temporary SIGCHLD handler from inside tfork_create() and tfork_status().
The termination signal of other child processes not created with tfork()
is forwarded to the existing signal handler if any.
There's one thing this thing can't protect us against and that is if a
process installs a SIGCHLD handler from one thread while another thread
is running inside tfork_create() or tfork_status() and the signal
handler doesn't forward signals for exitted childs it didn't fork, ie
our childs.
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
* handle one more LDB_FLAG_INTERNAL_DISABLE_SINGLE_VALUE_CHECK
case in ldb_tdb
* fix ldb_tdb locking (performance) problems
* fix ldb_tdb search inconsistencies by adding
read_[un]lock() hooks to the module stack
(bug #12858)
* add cmocka based tests for the locking issues
* ldb_version.h provides LDB_VERSION_{MAJOR,MINOR,RELEASE} defines
* protect ldb_modules.h from being used by Samba < 4.7
Note: that this release (as well as 1.1.30 and 1.1.31)
may cause problems for older applications, e.g. Samba
See https://bugzilla.samba.org/show_bug.cgi?id=12859
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Samba versions before 4.7 are incompatible with the read_[un]lock()
behaviour introduced into ldb.
This makes sure older Samba versions fail to compile against
ldb >= 1.2.0.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12859
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This indicates what feature set Samba assumes from the used
libldb from the system.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12859
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The new ldb whole-db lock behaviour now allows this test
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
(ldb releases have been made while this patch set was in train)
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
We must hold locks not just for the duration of each search, but for the whole search
as our module stack may make multiple search requests to build up the whole result.
This is explains a number of replication and read corruption issues in Samba
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
This allows Samba to provide a consistent view of the DB
despite the use of multiple databases via the partitions module
and over multiple callbacks via a module stack.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This will be used to implement read locking in ldb_tdb
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Currently, a lock is not held against modifications once the final
record is returned via a callback, so modifications can be made
during the DONE callback. This makes it hard to write modules
that interpert an ldb search result and do further processing
so will change in the future to allow the full search to be
atomic.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
A modify or rename during a search must not cause a search to change
output, and attributes having an index should in particular not see
any change in behaviour in this respect
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This is required to pass the test in future, because
otherwise the clean up will fail because we hold locks.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
If we do not do this, then we never take the all record lock, and instead do a lock
for every record as we go, which is very slow during a large search
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
* allow tdb_traverse_read before tdb_transaction[_prepare]_commit()
* Improve documentation for tdb_transaction_start()
* Add new function tdb_transaction_active()
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This will allow callers to avoid their own reference counting of transactions.
Additionally, this will always line up with the acutal transaction state, even
in the error cases where tdb can cancel the transaction
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
It now references the TDB_ALLOW_NESTING and TDB_DISALLOW_NESTING flags
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This restores the original intent of tdb_traverse_read() in
7dd31288a7
This is needed to avoid a deadlock with tdb_lockall() and the
transaction start, as ldb_tdb should take the allrecord lock during a
search (which calls tdb_traverse), and can otherwise deadlock against
a transaction starting in another process
We add a test to show that a transaction can now start while a read
traverse is in progress
This allows more operations to happen in parallel. The blocking point
is moved to the prepare commit.
This in turn permits a roughly doubling of unindexed search
performance, because currently ldb_tdb omits to take the lock due to
an unrelated bug, but taking the allrecord lock triggers the
above-mentioned deadlock.
This behaviour was added in 251aaafe3a for
Solaris 10 in 2005. But the run-fcntl-deadlock test works also on Solaris 10,
see https://lists.samba.org/archive/samba-technical/2017-April/119876.html.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
We need to migrate all pathname based VFS calls to use a struct
to finish modernising the VFS with extra timestamp and flags parameters.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Richard Sharpe <realrichardsharpe@gmail.com>
Using gcc 4.3.2:
cc1: error: unrecognized command line option "-Wno-format-truncation"
Signed-off-by: Björn Baumbach <bb@sernet.de>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
* Fix mutex locking in tevent_threaded_context_destructor().
* Fix a memleak on FreeBSD.
* Re-init threading in tevent_re_initialise().
* Include the finish location in tevent_req_default_print().
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Jun 22 17:17:33 CEST 2017 on sn-devel-144
It's verify useful when debugging code without a debugger to
be able to use tevent_req_print() in DEBUG statements.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
This has stopped working ages ago. The idea is clear, but if someone
wants to revive it, I think it needs a completely fresh start.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The race is easily reproduced by adding a poll(NULL,0,10) in between the two
pthread_mutex_unlock calls in _tevent_threaded_schedule_immediate.
Before 1828011317, the main thread was signalled only after the helper
had already unlocked event_ctx_mutex.
Full explaination follows:
-----------------------------------------------------------------
Inside _tevent_threaded_schedule_immediate() we have:
476 ret = pthread_mutex_unlock(&ev->scheduled_mutex);
477 if (ret != 0) {
478 abort();
479 }
HERE!!!!
481 ret = pthread_mutex_unlock(&tctx->event_ctx_mutex);
482 if (ret != 0) {
483 abort();
484 }
At the HERE!!! point, what happens is tevent_common_threaded_activate_immediate(),
which is blocked on ev->scheduled_mutex, get released and does:
514 while (ev->scheduled_immediates != NULL) {
515 struct tevent_immediate *im = ev->scheduled_immediates;
516 DLIST_REMOVE(ev->scheduled_immediates, im);
517 DLIST_ADD_END(ev->immediate_events, im);
518 }
- making an immediate event ready to be scheduled.
This then returns into epoll_event_loop_once(), which then calls:
910 if (ev->immediate_events &&
911 tevent_common_loop_immediate(ev)) {
912 return 0;
913 }
which causes the immediate event to fire. This immediate
event is the pthread job terminate event, which was previously
set up in pthreadpool_tevent_job_signal() by:
198 if (state->tctx != NULL) {
199 /* with HAVE_PTHREAD */
200 tevent_threaded_schedule_immediate(state->tctx, state->im,
201 pthreadpool_tevent_job_done,
202 state);
So we now call pthreadpool_tevent_job_done() - which does:
225 TALLOC_FREE(state->tctx);
calling tevent_threaded_context_destructor():
384 ret = pthread_mutex_destroy(&tctx->event_ctx_mutex); <---------------- BOOM returns an error !
385 if (ret != 0) {
386 abort();
387 }
as we haven't gotten to line 481 above (the line after
HERE!!!!) so the tctx->event_ctx_mutex is still
locked when we try to destroy it.
So doing an additional:
ret = pthread_mutex_lock(&tctx->event_ctx_mutex);
ret = pthread_mutex_unlock(&tctx->event_ctx_mutex);
(error checking elided) forces tevent_threaded_context_destructor()
to wait until tctx->event_ctx_mutex is unlocked before it locks/unlocks
and then is guaranteed safe to destroy.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This makes the behaviour much more robust, particularly with forest child
domains over one-way forest trusts.
Sadly we don't support this kind of setup with our current ADDC, so
there's no way to have automated tests for this behaviour, but
at least we know it doesn't break any existing tests.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=8630
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
repl_meta_data knows whether linked attributes are appropriately
[un-]duplicated, and this is how it tells ldb_tdb that.
Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Pair-programmed-with: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
* Add efficient function to find duplicate values in ldb messages
(this makes large multi-valued attributes in ldb_tdb more efficient)
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
ldb backends need to make sure they are not adding duplicate values to
multi-valued attributes in ADD and MODIFY operations. Until now they
have done this inefficiently using nested loops. Here we add common
functions that deal with large numbers of values in O(n log n) time,
but continue to use the simple methods for small numbers of values.
These functions take a struct ldb_context pointer and an options flag
arguments, although the ldb is not used, and only one bit of the
options has meaning. This is to allow further patches to switch on
schema-aware comparisons.
This entails an ABI jump to add the two new functions.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
