14 years agor23299: Fix the build for !WITH_SENDFILE.
James Peach [Fri, 1 Jun 2007 19:34:08 +0000 (19:34 +0000)]
r23299: Fix the build for !WITH_SENDFILE.

14 years agor23297: This introduces the winbind:ads parameter which defaults to True. Setting it
Volker Lendecke [Fri, 1 Jun 2007 18:39:50 +0000 (18:39 +0000)]
r23297: This introduces the winbind:ads parameter which defaults to True. Setting it
to False makes winbind use RPC and not LDAP methods to connect to the DCs,
even when it figured out they are AD.

14 years agor23295: Apply proto_exists and bin/.dummy dependencies consistently for all
James Peach [Fri, 1 Jun 2007 16:06:40 +0000 (16:06 +0000)]
r23295: Apply proto_exists and bin/.dummy dependencies consistently for all
binaries.  Fix vfs_zfsacl build rule in SAMBA_3_0.

14 years agor23291: Undo the somewhat naive change of r23279:
Michael Adam [Fri, 1 Jun 2007 12:24:57 +0000 (12:24 +0000)]
r23291: Undo the somewhat naive change of r23279:
The clear text presentaion of the sid in the ldap expression
does work with w2k3 but not with w2k....

Thanks to Guenther for advising me of this issue.


14 years agor23290: Fix another small and stupid but severe typo.
Michael Adam [Fri, 1 Jun 2007 12:18:16 +0000 (12:18 +0000)]
r23290: Fix another small and stupid but severe typo.
Hopfully, I have finally got this right... :-)


14 years agor23287: Use talloc_move instead of talloc_steal as this is what I really
Michael Adam [Fri, 1 Jun 2007 10:10:57 +0000 (10:10 +0000)]
r23287: Use talloc_move instead of talloc_steal as this is what I really
wanted to do.


14 years agor23284: Oh what a nasty typo! This gave me some headache,
Michael Adam [Thu, 31 May 2007 23:58:56 +0000 (23:58 +0000)]
r23284: Oh what a nasty typo! This gave me some headache,
with talloc randomly failing.

Hey, shouldn't TALLOC_ARRAY _not_ return NULL when
requested to allocate an array with zero entries? :-)


14 years agor23283: Use a temporary talloc context in ads:lookup_groupmem.
Michael Adam [Thu, 31 May 2007 23:55:37 +0000 (23:55 +0000)]
r23283: Use a temporary talloc context in ads:lookup_groupmem.
And clean up unused stuff at the end.
Daringly, I use talloc_steal at some point, where it
appears natural to me.


14 years agor23279: Replace occurrence of sid_binstring inside lookup_groupmem
Michael Adam [Thu, 31 May 2007 22:02:49 +0000 (22:02 +0000)]
r23279: Replace occurrence of sid_binstring inside lookup_groupmem
by sid_string_static.

14 years agor23274: merge CloseEventlog() pidl conversion from 3.0.26 && fix a few init call...
Gerald Carter [Thu, 31 May 2007 17:59:04 +0000 (17:59 +0000)]
r23274: merge CloseEventlog() pidl conversion from 3.0.26 && fix a few init call renames for svcctl in the previous commit

14 years agor23272: Fix main event loop - reviewed by Volker.
Jeremy Allison [Thu, 31 May 2007 17:26:30 +0000 (17:26 +0000)]
r23272: Fix main event loop - reviewed by Volker.

14 years agor23271: merge service control pidl change for CloseServiceHandle() from SAMBA_3_0_26
Gerald Carter [Thu, 31 May 2007 17:24:30 +0000 (17:24 +0000)]
r23271: merge service control pidl change for CloseServiceHandle() from SAMBA_3_0_26

14 years agor23269: Regenerate svcctl files after IDL fixes
Gerald Carter [Thu, 31 May 2007 15:46:01 +0000 (15:46 +0000)]
r23269: Regenerate svcctl files after IDL fixes

14 years agor23268: Rename some variables in the svcctl.idl to prevent
Gerald Carter [Thu, 31 May 2007 15:43:40 +0000 (15:43 +0000)]
r23268: Rename some variables in the svcctl.idl to prevent
redeclaration compile failures in the generated client

14 years agor23263: Remove an unused variable -- Fix Coverity ID 358
Volker Lendecke [Thu, 31 May 2007 09:23:51 +0000 (09:23 +0000)]
r23263: Remove an unused variable -- Fix Coverity ID 358

14 years agor23253: Add some debugging output.
Michael Adam [Wed, 30 May 2007 22:40:26 +0000 (22:40 +0000)]
r23253: Add some debugging output.

14 years agor23252: Complete the reworking of the ads lookup_groupmem function
Michael Adam [Wed, 30 May 2007 22:04:42 +0000 (22:04 +0000)]
r23252: Complete the reworking of the ads lookup_groupmem function
started in r23070, r23072, r23073, r23078, r23081 and r23082:

After retrieving the list of sids with the extended dn
ldap query, instead of passing all sids to the lsa_lookup_sids
call, now while extracting the sids from the extended dn member
entries, we first try to lookup the sid from cache and only pass
the sids that were not in cache to the lsa_lookup_sids call.


14 years agor23251: whoops! Fix compile error
Gerald Carter [Wed, 30 May 2007 21:55:49 +0000 (21:55 +0000)]
r23251: whoops!  Fix compile error

14 years agor23249: another sync from 3.0.26 for the echo work to rpcclient
Gerald Carter [Wed, 30 May 2007 21:46:03 +0000 (21:46 +0000)]
r23249: another sync from 3.0.26 for the echo work to rpcclient

14 years agor23248: Merge echo pipe implementation chanegs from SAMBA_3_0_26
Gerald Carter [Wed, 30 May 2007 21:43:48 +0000 (21:43 +0000)]
r23248: Merge echo pipe implementation chanegs from SAMBA_3_0_26
just to stay in sink.  This was more or less just for me
to play with.

14 years agor23244: Fix loop with nscd and NSS recusive calls.
Gerald Carter [Wed, 30 May 2007 19:47:35 +0000 (19:47 +0000)]
r23244: Fix loop with nscd and NSS recusive calls.

> Here's the problem I hit:
> getgrnam("foo") -> nscd -> NSS -> winbindd ->
>   winbindd_passdb.c:nam_to_sid() -> lookup_global_sam_name() ->
>   getgrnam("foo") -> nscd -> ....
> This is in the SAMBA_3_0 specifically but in theory could happen
> SAMBA_3_0_25 (or 26) for an unknown group.
> The attached patch passes down enough state for the
> name_to_sid() call to be able to determine the originating
> winbindd cmd that came into the parent.  So we can avoid
> making more NSS calls if the original call came in trough NSS
> so we don't deadlock ?  But you should still service
> lookupname() calls which are needed for example when
> doing the token access checks for a "valid groups" from
> smb.conf.
> I've got this in testing now.  The problem has shown up with the
> DsProvider on OS X and with nscd on SOlaris and Linux.

14 years agor23242: BUG 4671: Fix Makefile syntax error when SMBMOUNT_PROGS is not defined
Gerald Carter [Wed, 30 May 2007 14:53:10 +0000 (14:53 +0000)]
r23242: BUG 4671: Fix Makefile syntax error when SMBMOUNT_PROGS is not defined

14 years agor23241: In preparation for the cluster messaging import the parent smbd needs to
Volker Lendecke [Wed, 30 May 2007 13:41:38 +0000 (13:41 +0000)]
r23241: In preparation for the cluster messaging import the parent smbd needs to
respond to events.c style events.

14 years agor23236: Another bad merge: Correctly free and unlock the session record in
Volker Lendecke [Wed, 30 May 2007 07:02:40 +0000 (07:02 +0000)]
r23236: Another bad merge: Correctly free and unlock the session record in
session_claim. Jerry, this fixes the hanging smbstatus.

Sorry for that,


14 years agor23228: Merge cleanup to the gpfs module from Tridge. Also potentially disable
Volker Lendecke [Tue, 29 May 2007 19:54:26 +0000 (19:54 +0000)]
r23228: Merge cleanup to the gpfs module from Tridge. Also potentially disable
gpfs share modes in special situations. This might be split up in
several modules later.

14 years agor23226: Make the "debug prefix timestamp" output a bit more readable by making
Volker Lendecke [Tue, 29 May 2007 19:36:13 +0000 (19:36 +0000)]
r23226: Make the "debug prefix timestamp" output a bit more readable by making
the debug level alway at least 2 digits

14 years agor23225: Attached find a patch that makes use of NetSamLogonEx in
Volker Lendecke [Tue, 29 May 2007 19:31:57 +0000 (19:31 +0000)]
r23225: Attached find a patch that makes use of NetSamLogonEx in
winbind. With this and W2k3 DCs around it is possible to use
more than one winbind on the same machine account, because
NetSamLogonEx does not use the credentials chain.

I added the flag domain->can_do_samlogon_ex because this
only works against W2k3 and with schannel. The theory is to
try if we're AD and have schannel, and fall back to
NetSamLogon if this fails. can_do_samlogon_ex is thus a
protection against multiple failures.

Only checking into 3_0, this needs more review before going
into a production release.

Feel free to comment :-)

14 years agor23224: Move map_nt_error_from_unix to lib/errmap_unix.c. This function is
Volker Lendecke [Tue, 29 May 2007 19:09:38 +0000 (19:09 +0000)]
r23224: Move map_nt_error_from_unix to lib/errmap_unix.c. This function is
useful even in binaries that don't link in libsmb

14 years agor23223: Convert a leftover direct tdb access to sessionid.tdb
Volker Lendecke [Tue, 29 May 2007 18:41:16 +0000 (18:41 +0000)]
r23223: Convert a leftover direct tdb access to sessionid.tdb

14 years agor23222: share_mode_forall() is only for counting and status display stuff, so
Volker Lendecke [Tue, 29 May 2007 18:32:49 +0000 (18:32 +0000)]
r23222: share_mode_forall() is only for counting and status display stuff, so
traverse_read is enough here

14 years agor23221: Fix a bad merge, avoid a memory leak and tdb lock problem
Volker Lendecke [Tue, 29 May 2007 18:14:54 +0000 (18:14 +0000)]
r23221: Fix a bad merge, avoid a memory leak and tdb lock problem

14 years agor23220: Add traverse_read to dbwrap
Volker Lendecke [Tue, 29 May 2007 18:04:38 +0000 (18:04 +0000)]
r23220: Add traverse_read to dbwrap

14 years agor23219: Two warnings
Volker Lendecke [Tue, 29 May 2007 17:57:52 +0000 (17:57 +0000)]
r23219: Two warnings

14 years agor23210: Very funny, we thought to use netr_GetDcName (e.g. in winbind) but were using
Günther Deschner [Tue, 29 May 2007 15:27:13 +0000 (15:27 +0000)]
r23210: Very funny, we thought to use netr_GetDcName (e.g. in winbind) but were using
netr_GetDcAnyName all the time (which is the correct thing to do).

Fix the naming and opcode mixup in all branches.


14 years agor23207: merge net help test fix from SAMBA_3_0_26
Gerald Carter [Tue, 29 May 2007 14:53:46 +0000 (14:53 +0000)]
r23207: merge net help test fix from SAMBA_3_0_26

14 years agor23206: remove unreachable code
Gerald Carter [Tue, 29 May 2007 14:53:13 +0000 (14:53 +0000)]
r23206: remove unreachable code

14 years agor23204: Add MSG_SMB_BRL_VALIDATE. Tridge, this is a bit different from your bzr
Volker Lendecke [Tue, 29 May 2007 14:49:19 +0000 (14:49 +0000)]
r23204: Add MSG_SMB_BRL_VALIDATE. Tridge, this is a bit different from your bzr
branch, please check if it fulfils your needs.

Two changes: The validation is not done inside the brlock.c traverse_fn,
it's done as a separate routine.

Secondly, this patch does not call the checker routines in smbcontrol
directly but depends on a running smbd.

14 years agor23195: Add void *private_data to brl_forall
Volker Lendecke [Tue, 29 May 2007 13:26:44 +0000 (13:26 +0000)]
r23195: Add void *private_data to brl_forall

14 years agor23194: cherry pick two fixes from SAMBA_3_0_26
Gerald Carter [Tue, 29 May 2007 13:20:40 +0000 (13:20 +0000)]
r23194: cherry pick two fixes from SAMBA_3_0_26

* strptime() failure check
* make legcacy sid/uid/gid calls static

14 years agor23192: Remove fallback to looking up group mappings by the
Gerald Carter [Tue, 29 May 2007 13:18:57 +0000 (13:18 +0000)]
r23192: Remove fallback to looking up group mappings by the
Unix name after discussion with Simo.

14 years agor23186: Checkin on behalf of Tridge:
Volker Lendecke [Tue, 29 May 2007 10:50:25 +0000 (10:50 +0000)]
r23186: Checkin on behalf of Tridge:

Change notify.tdb to use dbwrap

14 years agor23185: Try to fix the IRIX build, also add the forgotten file_id.c in .26
Volker Lendecke [Tue, 29 May 2007 10:48:42 +0000 (10:48 +0000)]
r23185: Try to fix the IRIX build, also add the forgotten file_id.c in .26

14 years agor23184: Checkin for Tridge: Add a speed test to tdbtool
Volker Lendecke [Tue, 29 May 2007 10:15:01 +0000 (10:15 +0000)]
r23184: Checkin for Tridge: Add a speed test to tdbtool

14 years agor23183: Check in a change made by Tridge:
Volker Lendecke [Tue, 29 May 2007 09:30:34 +0000 (09:30 +0000)]
r23183: Check in a change made by Tridge:

This replaces the internal explicit dev/ino file id representation by a
"struct file_id". This is necessary as cluster file systems and NFS
don't necessarily assign the same device number to the shared file
system. With this structure in place we can now easily add different
schemes to map a file to a unique 64-bit device node.

Jeremy, you might note that I did not change the external interface of


14 years agor23173: Convert sessionid.tdb to ctdb. The 3.0.26 patch is a bit larger because
Volker Lendecke [Mon, 28 May 2007 13:33:54 +0000 (13:33 +0000)]
r23173: Convert sessionid.tdb to ctdb. The 3.0.26 patch is a bit larger because
it brings across the tdb-based list_sessions

14 years agor23172: Change shutdown_other_smbds to use connections_traverse instead of
Volker Lendecke [Mon, 28 May 2007 12:39:39 +0000 (12:39 +0000)]
r23172: Change shutdown_other_smbds to use connections_traverse instead of

14 years agor23171: Convert connections.tdb to dbwrap
Volker Lendecke [Mon, 28 May 2007 11:38:42 +0000 (11:38 +0000)]
r23171: Convert connections.tdb to dbwrap

14 years agor23170: Add map_nt_error_from_tdb()
Volker Lendecke [Mon, 28 May 2007 11:08:58 +0000 (11:08 +0000)]
r23170: Add map_nt_error_from_tdb()

14 years agor23169: Merge 535 of Tridge's bzr tree:
Volker Lendecke [Sun, 27 May 2007 17:12:08 +0000 (17:12 +0000)]
r23169: Merge 535 of Tridge's bzr tree:

  store the right data after cleaning lock records. This fixes
  RAW-BENCH-LOCK after a recovery on a cluster

14 years agor23168: Move the lp_max_connections() into service.c.
Volker Lendecke [Sun, 27 May 2007 16:34:49 +0000 (16:34 +0000)]
r23168: Move the lp_max_connections() into service.c.

14 years agor23167: Remove an unused parameter
Volker Lendecke [Sun, 27 May 2007 16:22:12 +0000 (16:22 +0000)]
r23167: Remove an unused parameter

14 years agor23166: Bring's iniparser copy in sync with the upstream version 2.17.
Lars Müller [Sun, 27 May 2007 15:58:19 +0000 (15:58 +0000)]
r23166: Bring's iniparser copy in sync with the upstream version 2.17.

14 years agor23165: Convert brlock.tdb to dbwrap
Volker Lendecke [Sun, 27 May 2007 10:35:14 +0000 (10:35 +0000)]
r23165: Convert brlock.tdb to dbwrap

14 years agor23164: Minor cosmetic checkin
Volker Lendecke [Sun, 27 May 2007 10:25:23 +0000 (10:25 +0000)]
r23164: Minor cosmetic checkin

14 years agor23163: Convert locking.tdb to dbwrap
Volker Lendecke [Sun, 27 May 2007 09:57:05 +0000 (09:57 +0000)]
r23163: Convert locking.tdb to dbwrap

14 years agor23162: Make use of messaging_send_buf
Volker Lendecke [Sun, 27 May 2007 09:29:26 +0000 (09:29 +0000)]
r23162: Make use of messaging_send_buf

14 years agor23161: Add TDB_VOLATILE as open_flag to activate the per-hashchain dead record
Volker Lendecke [Sun, 27 May 2007 09:22:11 +0000 (09:22 +0000)]
r23161: Add TDB_VOLATILE as open_flag to activate the per-hashchain dead record

14 years agor23156: Install the smb_share_modes.h header file.
Lars Müller [Sat, 26 May 2007 13:33:56 +0000 (13:33 +0000)]
r23156: Install the smb_share_modes.h header file.

14 years agor23155: Rename INSTALLPERMS to INSTALLPERMS_BIN to differenciate clearer from
Lars Müller [Sat, 26 May 2007 13:30:45 +0000 (13:30 +0000)]
r23155: Rename INSTALLPERMS to INSTALLPERMS_BIN to differenciate clearer from

14 years agor23154: Add INSTALLPERMS_DATA set to 0644 to install the header files with this
Lars Müller [Sat, 26 May 2007 13:24:56 +0000 (13:24 +0000)]
r23154: Add INSTALLPERMS_DATA set to 0644 to install the header files with this
permission set instead of 0755.

14 years agor23151: Re-arrange the logic so it looks identical to the
Jeremy Allison [Sat, 26 May 2007 01:26:09 +0000 (01:26 +0000)]
r23151: Re-arrange the logic so it looks identical to the
"file existed" case above. This has no functional
change but is easier to see that it's doing the
same thing.

14 years agor23150: Fix Samba3 in the build farm again. In the case where the
Jeremy Allison [Sat, 26 May 2007 01:19:51 +0000 (01:19 +0000)]
r23150: Fix Samba3 in the build farm again. In the case where the
file was being created and we go into the race condition check,
we were testing for compatible open modes, but were not breaking
oplocks in the same way as if the file existed. This meant that
we weren't going into the code path that sets fsp->oplock = FAKE_LEVEL_II_OPLOCK
if the client didn't ask for an oplock on file create. We need
to do this as internally it's what Windows does.

14 years agor23148: Fix old old bug in cli_smbwrite() (not incrementing
Jeremy Allison [Fri, 25 May 2007 23:50:35 +0000 (23:50 +0000)]
r23148: Fix old old bug in cli_smbwrite() (not incrementing
data being sent). Patch from

14 years agor23147: Patch #4566 from jacob berkman <>. Pass password data...
Jeremy Allison [Fri, 25 May 2007 22:04:03 +0000 (22:04 +0000)]
r23147: Patch #4566 from jacob berkman <>. Pass password data to krb5_prompter.

14 years agor23128: Fix typo.
Günther Deschner [Thu, 24 May 2007 23:17:24 +0000 (23:17 +0000)]
r23128: Fix typo.


14 years agor23127: Fill in some more netlogon dsgetdcname flavours (netr_DsRGetDCNameEx,
Günther Deschner [Thu, 24 May 2007 23:11:11 +0000 (23:11 +0000)]
r23127: Fill in some more netlogon dsgetdcname flavours (netr_DsRGetDCNameEx,
netr_DsRGetDCNameEx2) and add new ds request and reply flags, also add some
more WERROR codes.


14 years agor23122: Fix bug #4655. We marshall the enumdataex correctly,
Jeremy Allison [Thu, 24 May 2007 20:40:46 +0000 (20:40 +0000)]
r23122: Fix bug #4655. We marshall the enumdataex correctly,
but W2K3 doesn't follow our rules when sending data to
us. Ensure we look for the data at the correct offsets
when reading the data.
Too late for 3.0.25a - don't merge.

14 years agor23121: Fix Bug #2727 and let pam_smbpass at least link and dlopen correctly again.
Günther Deschner [Thu, 24 May 2007 20:26:07 +0000 (20:26 +0000)]
r23121: Fix Bug #2727 and let pam_smbpass at least link and dlopen correctly again.

Thanks to Bartlomiej Solarz-Niesluchowski <>.


14 years agor23117: Factor out local messaging.
Volker Lendecke [Thu, 24 May 2007 14:47:24 +0000 (14:47 +0000)]
r23117: Factor out local messaging.

This removes message_block / message_unblock. I've talked to Jeremy and
Günther, giving them my reasons why I believe they have no effect.
Neither could come up with a counter-argument, so they go :-)

14 years agor23116: Fix typo, found by Karolin Seeger
Volker Lendecke [Thu, 24 May 2007 13:49:10 +0000 (13:49 +0000)]
r23116: Fix typo, found by Karolin Seeger

14 years agor23112: Trim down the message.c API slightly: The messages_pending_for_pid is now
Volker Lendecke [Thu, 24 May 2007 11:09:37 +0000 (11:09 +0000)]
r23112: Trim down the message.c API slightly: The messages_pending_for_pid is now
replaced by MSG_FLAG_LOWPRIORITY or'ed into the msg_type. To enable this,
changed the msg_type definitions to hexadecimal.

This way we could theoretically add the MSG_FLAG_NODUPLICATES again, but I
would rather not do this, because that one is racy and can't be guaranteed at

14 years agor23109: Redo the change from r22772 I accidentially undid in r23108.
Michael Adam [Thu, 24 May 2007 09:06:15 +0000 (09:06 +0000)]
r23109: Redo the change from r22772 I accidentially undid in r23108.


14 years agor23108: Add defines of RTLD_NOW and RTLD_GLOBAL to replace.h.
Michael Adam [Thu, 24 May 2007 07:15:36 +0000 (07:15 +0000)]
r23108: Add defines of RTLD_NOW and RTLD_GLOBAL to replace.h.
Fixing build of ldb for instance on some systems without dlfcn.h.

14 years agor23107: Fix renames on file descriptors that are supposed to overwrite the
Jeremy Allison [Thu, 24 May 2007 01:57:02 +0000 (01:57 +0000)]
r23107: Fix renames on file descriptors that are supposed to overwrite the
target. Needs merging for 3.0.25a (sorry).

14 years agor23106: Use lchown for symlinks.
Jeremy Allison [Thu, 24 May 2007 00:13:24 +0000 (00:13 +0000)]
r23106: Use lchown for symlinks.

14 years agor23105: Add lchown to the vfs layer. We need this in the POSIX code.
Jeremy Allison [Wed, 23 May 2007 23:55:12 +0000 (23:55 +0000)]
r23105: Add lchown to the vfs layer. We need this in the POSIX code.

14 years agor23100: Implement the delete on close semantics I've just tested for
Jeremy Allison [Wed, 23 May 2007 21:32:10 +0000 (21:32 +0000)]
r23100: Implement the delete on close semantics I've just tested for
in Samba4 smbtorture. Fix rename on an open file handle.
Needed for 3.0.25a.

14 years agor23098: Fix the case of the #include statements so this module can be built on a
James Peach [Wed, 23 May 2007 20:48:53 +0000 (20:48 +0000)]
r23098: Fix the case of the #include statements so this module can be built on a
case-sensitive filesystem.

14 years agor23097: Add a new configure option to disable building SWAT (it is still built by
James Peach [Wed, 23 May 2007 20:46:07 +0000 (20:46 +0000)]
r23097: Add a new configure option to disable building SWAT (it is still built by

14 years agor23096: Make the lock failure message prettier.
James Peach [Wed, 23 May 2007 20:38:14 +0000 (20:38 +0000)]
r23096: Make the lock failure message prettier.

14 years agor23095: Support systems that have their PAM headers in /usr/include/pam.
James Peach [Wed, 23 May 2007 20:31:28 +0000 (20:31 +0000)]
r23095: Support systems that have their PAM headers in /usr/include/pam.

14 years agor23094: Update the iconv detection to handle cross-compiles better. If we are
James Peach [Wed, 23 May 2007 19:57:13 +0000 (19:57 +0000)]
r23094: Update the iconv detection to handle cross-compiles better. If we are
cross-compiling, we accept the first iconv library that we find (with a

14 years agor23093: Allow modules with the same name but different types.
James Peach [Wed, 23 May 2007 19:10:04 +0000 (19:10 +0000)]
r23093: Allow modules with the same name but different types.

14 years agor23091: Add rpccli_netlogon_sam_network_logon_ex, fix its parsing. This does not
Volker Lendecke [Wed, 23 May 2007 15:17:49 +0000 (15:17 +0000)]
r23091: Add rpccli_netlogon_sam_network_logon_ex, fix its parsing. This does not
use the credential chain and only works over netlogon, but it would
allow multiple outstanding auth requests for a single workstation

14 years agor23088: Fix rename for cifsfs client. This may be needed for
Jeremy Allison [Tue, 22 May 2007 23:56:47 +0000 (23:56 +0000)]
r23088: Fix rename for cifsfs client. This may be needed for

14 years agor23087: Fix POSIX setfilepathinfo to use lstat, not stat.
Jeremy Allison [Tue, 22 May 2007 22:35:13 +0000 (22:35 +0000)]
r23087: Fix POSIX setfilepathinfo to use lstat, not stat.
Still missing lchown (will add this for 3.0.26).
Don't merge for 3.0.25a - possibly 3.0.25b (if it

14 years agor23082: Ooops - that portion of the last checkin was by accident.
Michael Adam [Tue, 22 May 2007 21:17:31 +0000 (21:17 +0000)]
r23082: Ooops - that portion of the last checkin was by accident.
Before comitting that, I have to verify that it does not
break anything...


14 years agor23081: Reorder the lsa_lookup_sids functions so that the order makes more sense...
Michael Adam [Tue, 22 May 2007 21:04:56 +0000 (21:04 +0000)]
r23081: Reorder the lsa_lookup_sids functions so that the order makes more sense...

14 years agor23080: Fix bug #4637 - we hads missed some cases where
Jeremy Allison [Tue, 22 May 2007 20:20:01 +0000 (20:20 +0000)]
r23080: Fix bug #4637 - we hads missed some cases where
we were calling PRS_ALLOC_MEM with zero count.

14 years agor23078: Don't handle return code NT_STATUS_NONE_MAPPED from lookup sids
Michael Adam [Tue, 22 May 2007 15:13:32 +0000 (15:13 +0000)]
r23078: Don't handle return code NT_STATUS_NONE_MAPPED from lookup sids
as an error. (This is purely cosmetic here, issuing a success
message at the end.)

14 years agor23077: Minor cleanup
Volker Lendecke [Tue, 22 May 2007 14:31:26 +0000 (14:31 +0000)]
r23077: Minor cleanup

14 years agor23075: more duplicate code blocks from bad merge
Gerald Carter [Tue, 22 May 2007 13:55:31 +0000 (13:55 +0000)]
r23075: more duplicate code blocks from bad merge

14 years agor23074: Remove duplicate code blocks from bad merge
Gerald Carter [Tue, 22 May 2007 13:43:34 +0000 (13:43 +0000)]
r23074: Remove duplicate code blocks from bad merge

14 years agor23073: In the internal rpccli_lsa_lookup_sids_noalloc(), use a temporary
Michael Adam [Tue, 22 May 2007 13:40:01 +0000 (13:40 +0000)]
r23073: In the internal rpccli_lsa_lookup_sids_noalloc(), use a temporary
talloc context for use with the actual rpc query and response.
So the the parent context does not get flooded by the posslibly
large amount of response data (when looking up a lot of sids).

Note: It not possible, to simply use the names and domains arrays
as talloc contexts for the talloc_strdup calls, because from
rpccli_lsa_lookup_sids_all, this is called with names + offset
and domains + offset where names and domains are talloced
arraye for the complete response.

14 years agor23072: In winbindd_ads.c:lookup_groupmem, replace the bottleneck
Michael Adam [Tue, 22 May 2007 12:49:41 +0000 (12:49 +0000)]
r23072: In winbindd_ads.c:lookup_groupmem, replace the bottleneck
dn_lookup loop by a rpccli_lsa_lookupsids_all (see r23070)
call. This replaces one ldap search per member sid by one
rpc call per 1000 sids. This greatly speeds up groupmem
lookups for groups with lots of users.

Since the loop in lookup_groupmem was the only use of dn_lookup,
the function is removed.


14 years agor23070: The lsa rpc lookup sids call has a maximum number of SIDS to be
Michael Adam [Tue, 22 May 2007 11:30:35 +0000 (11:30 +0000)]
r23070: The lsa rpc lookup sids call has a maximum number of SIDS to be
looked up at one time. This limit is at 20480 for w2k3.
Our rpccli_lsa_lookup_sids function ignores this limit, so when
we give the server too long a list of SIDs, then we will get
nothing back. Since typically rpccli_lsa_lookup_sids is given
one SID  (or a small number of SIDS), this did not do harm
up to now. But since I want to use lsa_lookup_sids in a subsequent
modification to winbindd_ads.c:lookup_groupmem to get rid of
a vast number of dn_lookup calls to the server, I had to make
sure we do it correctly.

I have added a function rpccli_lsa_lookup_sids_all function
that has the same prototype but internally splits the list
of SIDs up into hunks of a (conservative, hard coded) 1000
SIDs each for a first go.

If this approach is agreed upon, the new function could replace
the original rpccli_lsa_lookup_sids function.


14 years agor23068: Fix some uninitialized variables in the solarisacl module
Volker Lendecke [Tue, 22 May 2007 10:05:56 +0000 (10:05 +0000)]
r23068: Fix some uninitialized variables in the solarisacl module

14 years agor23055: Rewrite messages.c to use auto-generated marshalling in the tdb. I'm
Volker Lendecke [Mon, 21 May 2007 22:17:13 +0000 (22:17 +0000)]
r23055: Rewrite messages.c to use auto-generated marshalling in the tdb. I'm
doing this because for the clustering the marshalling is needed in more
than one place, so I wanted a decent routine to marshall a message_rec
struct which was not there before.

Tridge, this seems about the same speed as it used to be before, the
librpc/ndr overhead in my tests was under the noise.


14 years agor23054: Move the check for the lookup_domain of S-1-22-{1,2} before the
Gerald Carter [Mon, 21 May 2007 21:33:51 +0000 (21:33 +0000)]
r23054: Move the check for the lookup_domain of S-1-22-{1,2} before the
check for IS_DC.  Otherwise we will for example fail to lookup a
sid of S-1-22-1-780 because it has no valid struct winbindd_domain*
in the list.   Thanks to Simo for the catch.

14 years agor23051: sid_to_[ug]id fixes for smbd
Simo Sorce [Mon, 21 May 2007 20:51:15 +0000 (20:51 +0000)]
r23051: sid_to_[ug]id fixes for smbd