Greg Silverman | 6 Feb 00:43 2016

Indexing Services

Is it possible for the Linux server running samba to provide Windows or Mac
with the index files they use, so they will not generate indexes themselves
by reading all files in the share?

Can the samba share host a search service that clients will use, to avoid
indexing, themselves?


Greg Silverman
SF Bay Area

Günther Deschner | 5 Feb 15:07 2016

[PATCHES] fix some build warnings


after building samba with gcc6, here are some first build warning fixes.

Please review & push,


Günther Deschner                    GPG-ID: 8EE11688
Red Hat                         gdeschner <at>
Samba Team                              gd <at>
Volker Lendecke | 5 Feb 12:43 2016

[PATCH] pdb: Fix segfault in pdb_ldap for missing gecos


This is the patch from which to me looks
entirely correct. I've taken the liberty to add the signed-off-by the
original author. Not sure this is right, feel free to remove that
again. Anyway, it's got my review.

Second review appreciated!




SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen, mailto:kontakt <at>
From 691feb51483ceb1803a25ea1dac5c155b3b5506a Mon Sep 17 00:00:00 2001
From: Luca Olivetti <luca <at>>
Date: Fri, 5 Feb 2016 12:02:51 +0100
Subject: [PATCH] pdb: Fix segfault in pdb_ldap for missing gecos

Signed-off-by: Luca Olivetti <luca <at>>
Reviewed-by: Volker Lendecke <vl <at>>
(Continue reading)

Michael Adam | 5 Feb 12:12 2016

[PATCH] clean type argument from dlinklist

I just stumbled across the fact that the
type argument was left unused for a few functions
in a change back in 2010, to avoid a bigger diff
and allow fo easier roll-back in case of problems.
Now 6 years later I think we can say the change
did not introduce a lot of problems, and remove
the type arg...

Review appreciated.

Thanks - Michael
From 134932ae6d52285de18464a7ebe6a554f3c465ab Mon Sep 17 00:00:00 2001
From: Michael Adam <obnox <at>>
Date: Fri, 5 Feb 2016 11:32:18 +0100
Subject: [PATCH 01/12] dlist: remove unneeded type argument from

Signed-off-by: Michael Adam <obnox <at>>
 ctdb/common/ctdb_io.c                              |  2 +-
 ctdb/server/ctdb_call.c                            |  4 ++--
 ctdb/server/ctdb_daemon.c                          |  2 +-
 ctdb/server/ctdb_lock.c                            |  8 ++++----
 ctdb/server/ctdb_logging.c                         |  2 +-
 lib/torture/torture.c                              | 14 +++++++-------
 lib/util/dlinklist.h                               |  5 ++---
 lib/util/tests/dlinklist.c                         |  4 ++--
 libcli/cldap/cldap.c                               |  2 +-
(Continue reading)

Richard Sharpe | 5 Feb 03:00 2016

Is this a valid talloc region I am seeing

Hi folks,

As part of my investigation of some memory issues, I have looked at a
memory dump, and towards the end of the heap (most recently added
part) I found the following striung of bytes

         770c 15e8

This, when converted to the correct order, looks like the talloc magic
with flags 0x07, indicating that the memory is now free (among other

Then, looking back a bit I find that the location that did the freeing
is: 0x7fcd49339980 and looking at that with gdb I find that it is:

(gdb) p (char *)0x7fcd49339980
$1 = 0x7fcd49339980 "../source3/smbd/smb2_server.c:3432"

But if I look at that location in the source we are using it is:

        DEBUG(10,("smbd_smb2_request idx[%d] of %d vectors\n",
                 req->current_idx, req->in.vector_count));

But that should have done nothing because the debug level is 0.
Moreover, I cannot see anywhere in the DEBUG code that it is doing a
talloc with that location as the type.

I cannot understand what is going on at this stage. Can anyone suggest
what is going on?

(Continue reading)

Douglas Bagnall | 4 Feb 21:55 2016

Fwd: [Fwd: Checking in]

hi Garming,

Does this look like a reasonable summary of where we're at?:

[Issue 25052] Windows clients do not update DNS after changing their
hostname as a domain member

 * We are looking at it

[Issue 25053] Replication fails when object moved across OUs

 * Can't reproduce, paused

[Issue 25054] Samba 4.1.12 malforms DNS TXT lookups over 255 chars

 * more or less fixed?

[Issue 25055] Implement LDAP virtual list view (VLV) request control
(OID 2.16.840.1.113730.3.4.9)

 * nearly there

[Issue 25056] Support 2 or more DNS forwarders for Samba

 * Barely started, pending 255 char thing?

[Issue 25871] Revive LDAP Recursive Search patches

 * Done, pushed upstream and back-ported but not in any actual release?

(Continue reading)

Herb Lewis | 4 Feb 18:31 2016

bug in smbcacls

In source3/lib/util_sd.c function parse_ace there is the following
line that I think is incorrect

if (sscanf(p, "%u/%u/%u", &atype, &aflags, &amask) == 3 &&

in the old samba 3.6 code it used to read

if (sscanf(p, "%i/%i/%i", &atype, &aflags, &amask) == 3 &&

and according to the man page for sscanf the %u is only for decimal
integers while %i also checks for base 16 and base 8 as well. Was there
a reason this was changed so that hex numbers are no longer allowed?

Ralph Boehme | 4 Feb 18:20 2016

[PATCH] Patch for bug 11714: Listening for network interface changes doesn't work



winbindd is watching for network interface address changes by virtue
of our addrchange_send/recv() API that uses netlink to query the
kernel for this information.


[2016/02/04 11:43:31.723123, 10, pid=17692, effective(0, 0), real(0,0)] ../source3/lib/addrchange.c:154(addrchange_done)
  tdgram_recvfrom_recv returned Operation not supported
[2016/02/04 11:43:31.723176, 10, pid=17692, effective(0, 0), real(0,0), class=winbind] ../source3/winbindd/winbindd.c:1486(winbindd_addr_changed)
  addrchange_recv failed: NT_STATUS_NOT_SUPPORTED, stop listening

strace shows a failing ioctl(fd, FIONREAD, ...) returning
EOPNOTSUPP. We use the ioctl to determine the amount of pending data
so we can allocate an appropriate buffer.

This is a regression introduced in 0d161e42d9aeb155eae2b04eccec497b21de8029.

What can we do?

We could use MSG_PEEK, but that's not portable and we'd have to call
recvmsg() twice every time the fd is reported readable.

As a band aid, allocate a laaaarge buffer and watch out for MSG_TRUNC
after receiving.

We could also revert 0d161e42d9aeb155eae2b04eccec497b21de8029 and fix
(Continue reading)

Andreas Schneider | 4 Feb 10:27 2016

[PATCHSET] Add MIT KDC kdb driver for Samba


the attached patchset adds a mit_samba a shim layer between Samba and KDB and 
a KDB driver for the MIT KDC.

Sumit has reviewed the patchset already and we fixed two small issues. We run 
Coverity on the code and it didn't find an issue.

## master-mit-kdc-ok...origin/master [ahead 29];a=shortlog;h=refs/heads/master-mit-kdc-ok

Please review and pushing the code is appreciated.


	-- andreas


Andreas Schneider                   GPG-ID: CC014E3D
Samba Team                             asn <at>
Attachment (master-mit-kdc.patch): text/x-patch, 129 KiB
Martin Schwenke | 4 Feb 10:28 2016

[PATCH] Clean up util_strlist.c and separate out str_list_make_v3() into a separate file

I really want to use str_list_make() in CTDB instead of reinventing
the wheel.  However, because of str_list_make_v3() I would need to
drag in 1/2 the world.  So split, clean-up and move util_strlist.c to
samba-util-core so I can use it in CTDB.

Please review and push...

peace & happiness,
Attachment (ctdb.patch): text/x-patch, 17 KiB
Hemanth Thummala | 3 Feb 19:53 2016

[PATCH] Possible memory leak in loadparm.

Hi Andreas,

Thanks for reviewing the changes. 
I have created a bug and included the details in the patch now.


On 2/2/16, 11:03 PM, "Andreas Schneider" <asn <at>> wrote:

>On Tuesday 02 February 2016 19:38:57 Hemanth Thummala wrote:
>> Hi folks,
>> We are using samba 4.3 stack. We ran into some kind of memory leak issues
>> recently. To detect those leaks, we have run samba under valgrind. We were
>> able to detect and fix one of the leaks specific to us(not in samba).
>> talloc_free_children(ServicePtrs[idx]) is used in free_service_byindex().
>> According to its documentation, talloc_free_children() will only cleanup
>> its child contexts and not the actual one.
>> Instead of using NULL we can use "ServicePtrs” as context for allocating
>> individual records. Valgrind is happy after these changes. No leak reports
>> are seen now.
>Hi Hemanth,
>thanks for your contributions. The patch looks good for me. Could you please 
(Continue reading)