summaryrefslogtreecommitdiff
path: root/source4
AgeCommit message (Collapse)AuthorFilesLines
2008-12-18s4:ntvfs_generic: pass RAW-OPEN againStefan Metzmacher1-0/+2
This is the samba4 version of commit 9391dad85d08bb0939f4db1472c6cf063ebea892 metze
2008-12-18RAW-CHKPATH: not every server behaves like samba3...Stefan Metzmacher1-1/+1
metze
2008-12-18BASE-LOCK: fix LOCK5Stefan Metzmacher1-4/+12
We need to close all 3 file handle to make the unlink work. metze
2008-12-18Handle different failure modes when we wipe the db in provisionAndrew Bartlett1-7/+8
We didn't handle the mode where we can't load the main sam.ldb due to the modules being 'wrong', and when we did remove the file, we didn't wipe the partitions.
2008-12-18Parse options (and open the database) before starting transactionsAndrew Bartlett1-2/+2
2008-12-18Print error strings when transactions fail in ldb toolsAndrew Bartlett3-6/+6
2008-12-18Rename ldb index pointer wrapper functionAndrew Bartlett1-8/+8
This also asserts that it is used only for index records. Andrew Bartlett
2008-12-17Samba3 smbd now passes test BATCH23 (with the fix for bug #5979), only ↵Jeremy Allison1-4/+0
BATCH22 left to fix. Jeremy.
2008-12-17s4: smbtorture, fix the build of netlogon test after idl change.Günther Deschner1-2/+2
Guenther
2008-12-17s4: fix smbtorture build after idl change.Günther Deschner1-4/+13
Guenther
2008-12-17s4:libcli/resolve: specify the port for the resulting socket_addressesStefan Metzmacher8-9/+25
metze
2008-12-17s4:libcli/resolve: optionally return the name that belongs to the returned ↵Stefan Metzmacher8-23/+82
address E.g. this helps for DNS CNAME and SRV results. metze
2008-12-17s4:libcli/resolve: pass down flags to the resolver backendsStefan Metzmacher8-8/+29
metze
2008-12-17s4:libcli/resolve: remove all backend specific sync functionsStefan Metzmacher5-85/+0
metze
2008-12-17s4:libcli/resolve: let the "host" module use the dns_ex.c codeStefan Metzmacher1-183/+11
That means we now return all ip addresses instead of just the first one. metze
2008-12-17s4:libcli/resolve: add getaddrinfo()/dns_looup() resolvingStefan Metzmacher2-0/+485
This "dns_ex" module provides flexible lookup methods for dns lookups. The getaddrinfo() part looks at /etc/hosts and dns. As it handles CNAME replies badly we fallback to use dns_lookup(name, "A"). The dns_lookup() makes DNS SRV lookups possible. This module is not a real resolve module, it's just a generic helper as the nbtlist.c code is. The next step will be that the "host" module will use the dns_ex.c code. metze
2008-12-17s4:libcli/resolve: don't use __RESOLVE_H__ it might be used by system ↵Stefan Metzmacher1-3/+3
headers too metze
2008-12-17s4:lib/socket: we need to lookup the #20 netbios name when we connect to a ↵Stefan Metzmacher2-2/+2
remote server metze
2008-12-17s4:lib/socket: remove unused codeStefan Metzmacher1-22/+0
metze
2008-12-17s4:headermap: dom_sid.h was renamed to server_id.hStefan Metzmacher2-3/+3
metze
2008-12-17s4:headermap: we need the pathes for gen_ndr headers with and without ../Stefan Metzmacher1-25/+52
This should fix the OpenChange build metze
2008-12-17s4:lib/tevent: add lib/events/ compat and let things compileStefan Metzmacher8-3/+18
metze
2008-12-17s4:lib/events: move to toplevel directory as lib/tevent/Stefan Metzmacher30-7911/+0
This commit will not compile on its own. metze
2008-12-17s4: fix LIBEVENTS dependencies and use more forward declarationsStefan Metzmacher22-16/+23
We should only include events.h where we really need it and prefer forward declarations of 'struct event_context' metze
2008-12-17s4:ldb: add some python tests for extended dnsAndrew Bartlett1-27/+60
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:dsdb: remove normalise moduleAndrew Bartlett2-206/+0
The extended_dn_out module provides the functionality now. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:provision: use extended_dn_out_ldb or extended_dn_out_dereference ↵Andrew Bartlett2-4/+7
depending on the backend This just changes the existing stratagy of loading different modules for the OpenLDAP backend to also include extended_dn_out_* When we provision the OpenLDAP backend, we make sure to include the 'deref' overlay (which must be made available by the OpenLDAP build) Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:dsdb: split extended_dn into extended_dn_in, extended_dn_out and ↵Andrew Bartlett7-673/+1517
extended_dn_store. By splitting the module, the extended_dn_in and extended_dn_store moudles can use extended_dn_out to actually get the extended DN. This avoids code duplication. The extended_dn_out module also contains a client implementation of the OpenLDAP dereference control (draft-masarati-ldap-deref-00). This also introduces a new control 'DSDB_CONTROL_DN_STORAGE_FORMAT_OID' to ask the extended_dn_out module to return whatever the 'storage format' is. This allows us to work with both OpenLDAP (which performs a dereference at run time) and LDB (which stores the GUID and SID on disk). Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:dsdb: Make the linked_attributes module set an extended dnAndrew Bartlett1-192/+325
This means that linked attributes will always have the same case form as the actaul entry, as we search for that entry. We then also use the GUID and SID found on that entry to fill in the extended DN on disk. Andrew Bartlett Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:rootdse: fix the logic to indentify a rootdse searchAndrew Bartlett1-2/+1
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldb: make it possible to return per entry controlsAndrew Bartlett25-27/+62
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:setup: fix cut-n-paste error Builtin-Domain => Samba4-Local-DomainAndrew Bartlett1-1/+1
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:setup: don't set objectCategory: CN=Domain-DNS,${SCHEMADN}Andrew Bartlett1-3/+0
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:torture: add ldb testsAndrew Bartlett3-1/+751
These tests are for both the new extended DN functionality (and were vital in finding bugs during implementation) and for the normal DN parsing and comparison routines. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldap_server: return the extended dn to the LDAP client if availableAndrew Bartlett1-1/+16
This uses an early peek at the extended_dn_control (in the request) to see what output format to use. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldb-samba: register samba specific extended dn handlersAndrew Bartlett1-51/+138
This provides the two extended DN handlers for the GUID and SID types, and makes the parsing more strict (where possible, it uses ndr_pull_struct_blob_all(), to cause an error if trailing data is found). Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:samldb: make use of dom_sid_split_rid()Andrew Bartlett1-4/+3
Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:samldb: improve error stringsAndrew Bartlett1-6/+8
When things go wrong with LDB, this routine seems to be particularly sensitive to it. This extra debugging should help the next poor soul who breaks LDB. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldb.i: hang the dn on the NULL context as the python destructor will free itAndrew Bartlett1-0/+2
This fixes a bug in the ldb.i python wrapper, that showed up under valgrind. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldb: use try to print the extended dn in the ldif outputAndrew Bartlett1-2/+4
This allows searches with the extended DN control to still print the extended DN in ldif output (it would otherwise be parsed and hidden in the structure). Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:dsdb: add support for DSDB_OPENLDAP_DEREFERENCE_CONTROLAndrew Bartlett2-0/+137
Encode and decode the OpenLDAP dereference control (draft-masarati-ldap-deref-00) At this time, the ldb_controls infrustructure does not handle request and reply controls having different formats, so this is purely the client implementation (ie, there is no decode of the client->server packet, and no encode of the server->client packet). Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:libcli/ldap: split out a ldap_decode_attribs_bare() functionAndrew Bartlett1-8/+18
The OpenLDAP dereference control (draft-masarati-ldap-deref-00) uses an attribute list, as found in the search reply, but without one enclosing ASN1_SEQUENCE(0) This allows the dereference control parsing code to use this as a helper function. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldb_ildap: try to pass extended DNs to the serverAndrew Bartlett1-5/+5
Whenever we pass a DN to the LDAP server, we now use ldb_dn_get_extended_linearized(). This allows us to send the extended DN if set, and therefore allows searches of the form '<GUID=aaa45ea0-94cd-45e9-8753-abe455d9a8f1>'. We actually use the '0' format (GUID=aaa45ea094cd45e98753abe455d9a8f1) because it is more widely supported (by Win2k in particular). Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17s4:ldb: add infrastructure for extended dn handlersAndrew Bartlett4-60/+524
This introduces a new set of pluggable syntax, for use on the extended DN, and uses them when parsing the DN. If the DN appears to be in the extended form, we no longer return the full DN 'as is', but only return the normal part from ldb_dn_get_linearized(). When validating/parsing the DN we validate not only the format of the DN, but also the contents of the GUID or SID (to ensure they are plausable). We also have functions to set and get the extended components on the DN. For now, extended_dn_get_linearized() returns a newly constructed and allocated string each time. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17Add hint to use passwordAttributes in @KLUDGE_ACL in futureAndrew Bartlett1-1/+2
This module is not used at the moment, but if we do use it again, we should try to avoid duplicate lists. Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17Make greater use of 'GUID_from_data_blob'Andrew Bartlett1-40/+11
This avoids accidentily running off the end of a string, and uses a single 'guess which type of GUID I have' algorithm. Andrew Bartlett Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-17Fix sequence number generation against OpenLDAPAndrew Bartlett1-0/+8
It seems that in 2deeb99fff1a90c79ba1927e1a069362e250a63c adding the partition control to this request was missed out. Andrew Bartlett Signed-off-by: Stefan Metzmacher <metze@samba.org>
2008-12-16Fix header installation after move of gen_ndr headers.Jelmer Vernooij2-29/+29
2008-12-17Merge branch 'master' of ssh://git.samba.org/data/git/sambaAndrew Tridgell43-5211/+242
2008-12-17an experimental patch for fixing ldb bloatAndrew Tridgell3-9/+352
ldb indexing can cause huge files, and huge memory usage. This experiment allows us to keep indexes in memory during a transaction, then to write the indexes to disk when the transaction completes. The result is that the db is much smaller (we have seen improvements of about 100x in file size) and memory usage during large transactions is also greatly reduced Note that this patch uses the unusual strategy of putting pointers into a ldb (and thus into a tdb). This works because the pointers are only there during a transaction, so the pointers are not exposed to any other users of the database. The pointers allow us to avoid some really bad allocation problems with tdb record allocation during the re-indexing.