s3-winbindd: provide one wcache_open() function for all tdb opens

Guenther

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15963

Signed-off-by: Guenther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

s3-winbindd: make initialize_winbindd_cache() static

Guenther

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15963

Signed-off-by: Guenther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

s3-winbind: make wcache_store_seqnum static

Guenther

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15963

Signed-off-by: Guenther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

s3-winbindd: Fix winbind NDR caching.

All of winbindd's core caching relies on NDR entries. Those entries can
not be stored in winbindd_cache.tdb via wcache_store_ndr() as long as
there is no SEQNUM entry present in the cache.

Guenther

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15963

Signed-off-by: Guenther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

s3-selftest: add tests for winbindd_cache.tdb sanity

Guenther

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15963

Signed-off-by: Guenther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

selftest: use common and simpler code to read config.h

This also removes some garbage variables from these module's namespaces.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Thu Dec  4 23:54:18 UTC 2025 on atb-devel-224

s3:selftest:tests.py: remove unused imports

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

vfs_ceph_new: proper failure-handling in chdir and getwd

In case of an unlikely failure of SMB_VFS_HANDLE_GET_DATA macro, bail
out using 'goto' in order to ensure that START_PROFILE_X is properly
matched by END_PROFILE_X.

Signed-off-by: Shachar Sharon <ssharon@redhat.com>
Reviewed-by: John Mulligan <jmulligan@redhat.com>
Reviewed-by: Anoop C S <anoopcs@samba.org>
Autobuild-User(master): Anoop C S <anoopcs@samba.org>
Autobuild-Date(master): Thu Dec  4 10:02:55 UTC 2025 on atb-devel-224

Georgian translation for tools

Signed-off-by: Ekaterine Papava <papava.e@gtu.ge>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Wed Dec  3 10:54:19 UTC 2025 on atb-devel-224

Revert "ldb: User hexchars_upper from replace.h"

This reverts commit 542cf01bfe530a83dfbc8a606d182c0a5a622059.

We shouldn't put a hard requirement for libreplace in libldb! We do not need
libreplace on Linux until we start using hexbytes_upper.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15961

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

s3:libads: Set udp_preference_limit = 0 for MIT Kerberos

This option enable TCP connection before UDP, when sending a message to
the KDC.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

s3:libads: Set a request timeout for Kerberos requests

Without this, libkrb5 can wait indefinitely after creating a TCP
connection. This means winbind is stuck forever till it is restarted.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15955

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

wscript: Fix double quoting for USING_SYSTEM_MITKRB5

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

s4:selftest: move tests that want to be last, last

These tests want to run last because they do a dbcheck to ensure no
earlier test caused corruption. There are also some tests that want to
run second last.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Thu Nov 27 22:40:02 UTC 2025 on atb-devel-224

smbd: reshuffle requested/granted lease level logic of Directory Leases

This moves the logic to adjust the lease levels for Directory Leases to the same
place where we do it for other cases.

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Nov 27 10:04:47 UTC 2025 on atb-devel-224

smbd: distinguish "no record" from "bad record" in smbXsrv_open_global_lookup()

Allows callers to distinguish between "no record" and "bad record" when
smbXsrv_open_global_parse_record() fails.

No change in behaviour, but might be useful in the future.

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>

net/serverid: initialize state->open_db

Fixes a crash that happens when smbXsrv_open_global.tdb contains only replay
records.

  #0  0x00007ffff70a718a in dbwrap_exists (db=0x0, key=...) at ../../lib/dbwrap/dbwrap.c:347
  #1  0x00000000002dc7f2 in wipedbs_traverse_replay_records (rec=0x7fffffffd180, wipedbs_state=0x362d70)
      at ../../source3/utils/net_serverid.c:710
  #2  0x00007ffff70aa932 in db_rbt_traverse_internal (db=0x37e940, f=0x2dc749 <wipedbs_traverse_replay_records>, private_data=0x362d70,
      count=0x7fffffffd220, rw=false) at ../../lib/dbwrap/dbwrap_rbt.c:472
  #3  0x00007ffff70aaa0e in db_rbt_traverse_read (db=0x37e940, f=0x2dc749 <wipedbs_traverse_replay_records>, private_data=0x362d70)
      at ../../lib/dbwrap/dbwrap_rbt.c:499
  #4  0x00007ffff70a7433 in dbwrap_traverse_read (db=0x37e940, f=0x2dc749 <wipedbs_traverse_replay_records>, private_data=0x362d70, count=0x0)
      at ../../lib/dbwrap/dbwrap.c:437
  #5  0x00000000002dcedf in net_serverid_wipedbs (c=0x34c0d0, argc=0, argv=0x365380) at ../../source3/utils/net_serverid.c:840
  #6  0x00000000002c9675 in net_run_function (c=0x34c0d0, argc=1, argv=0x365378, whoami=0x258c96 "net tdb smbXsrv", table=0x7fffffffd360)
      at ../../source3/utils/net_util.c:451
  #7  0x00000000002e6e2c in net_tdb_smbXsrv (c=0x34c0d0, argc=1, argv=0x365378) at ../../source3/utils/net_tdb.c:105
  #8  0x00000000002c9675 in net_run_function (c=0x34c0d0, argc=2, argv=0x365370, whoami=0x258d45 "net tdb", table=0x7fffffffd440)
      at ../../source3/utils/net_util.c:451
  #9  0x00000000002e6f04 in net_tdb (c=0x34c0d0, argc=2, argv=0x365370) at ../../source3/utils/net_tdb.c:128
  #10 0x00000000002c9675 in net_run_function (c=0x34c0d0, argc=3, argv=0x365368, whoami=0x22ca4b "net", table=0x323640 <net_func>)
      at ../../source3/utils/net_util.c:451
  #11 0x0000000000277227 in main (argc=6, argv=0x7fffffffe2a8)
      at ../../source3/utils/net.c:1481

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>

smbXsrvdump: catch exception from unparsable records

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>

smbtorture: fix a test rundown

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>

vfs_gpfs: Fix CID 1666430 Control flow issues (DEADCODE)

Restore the saved errno properly and record it with vfs_aio_state once
vfs_gpfs_get_winattrs_helper() returns for async code path.

Signed-off-by: Anoop C S <anoopcs@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Wed Nov 26 16:48:00 UTC 2025 on atb-devel-224

vfs_ceph_new: use vfs_ceph_iget/vfs_ceph_iput in vfs_ceph_disk_free

Currently inode of root is using to report stats that works fine for share of root volume.
But for subvolume share it reports incorrect information. Hence choose the inode based on path,
so that it will report stats information correctly.

Bug:https://bugzilla.samba.org/show_bug.cgi?id=15954

Signed-off-by: Shweta Sodani <ssodani@redhat.com>
Reviewed-by: Xavi Hernandez <xhernandez@redhat.com>
Reviewed-by: Anoop C S <anoopcs@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Mon Nov 24 18:03:08 UTC 2025 on atb-devel-224

s4:dsdb: Do not declare cm_print_error()

This is part of the cmocka.h header file.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Nov 24 11:28:08 UTC 2025 on atb-devel-224

libcli:ldap: Do not declare cm_print_error()

This is part of the cmocka.h header file.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>

tdb: Fix parse_hex during `tdbtool storehex`

Fixes: fd0561279
During `tdbtool storehex`, tdbtool check whether the input
string in hex format.

However, during the check, the index is never moved forward,
resulting in checking beyond the valid input string.

This patch fix the issue by checking the valid string

Signed-off-by: Lin Liu <lin.liu01@citrix.com>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Jennifer Sutton <jsutton@samba.org>
Autobuild-Date(master): Thu Nov 20 22:29:03 UTC 2025 on atb-devel-224

selftest: move knownfail.d/krb5-no-dollar to expectedfail.d

If these fail to fail, there is a problem.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

autobuild: run ntvfs krb5 tests on MIT build

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: avoid reusing a variable name

fallback_principal was used for two different uses: a copy of the
original principal from which to derive values, and a new principal
which has the '$' appended on the account name. We might as well be
clear and an optimising compiler won't see the difference.

Whether we actually need a temporary principal as opposed to using the
one that was passed in is a separate question.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: do not match principal + '$' if smb.conf says not to

With this patch we honour

 kdc name match implicit dollar without canonicalization = no

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc:principal lookup will soon succeed on ad_dc_ntvfs

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4/torture:kdc-canon understands no-implicit-dollar setting

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytests: krb5 raw tests use TestCase.get_server_param()

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytest: krb5 alias tests: expect no machine$ match on ad_dc_ntvfs

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytest:krb5: ms-kile-client tests notice lack of dollar matching

In the ad_dc_ntvfs environment.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytest: krb5 tests remember implicit dollar option

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytests: add TestCase.get_server_param() method

This makes it easier to get a loadparm value from the server smb.conf
(rather than the client smb.conf), so you can alter test behaviour
accordingly.

The class._server_lp attribute is lazily loaded when first needed.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytest:krb5: print error names on error

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytest:krb5: errcode errors include names

Before:

> AssertionError: 6 not found in (20,)

After:

> AssertionError: 6 not found in (20,) : KDC_ERR_C_PRINCIPAL_UNKNOWN not in ['KDC_ERR_TGT_REVOKED']

Useful for people who don't know the codes off by heart.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytests: rename KRB_ERR_TKT_NYV as KDC_ERR_TKT_NYV

to be consistent with all the others.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

tests: run krb5.kdc tests on ad_dc_ntvfs without implicit dollar match

that is, with "kdc name match with implicit dollar = no"

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

loadparm: add "kdc name match implicit dollar without canonicalization"

This does nothing yet.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: allocate fallback realm later, closer to use

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: do not fallback to "$$" if user is "$"

or from "" to "$", though I am not sure it is easy to get this far
with an empty account name.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: flatten samba_kdc_lookup_client dollar fallback

There is a single linear successful path through the 'num_comp == 1'
and various NULL checks, but it is written in a branchy fashion as if
you could skip some portion.

git diff -b is probably useful.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: avoid a leak on error

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc: improve a comment

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:torture:kdc-canon: test each combination only once

These tests exhaustively combinations of binary options. With
492d9f083dc23aff2c1fa12e21765861df1c1b38 ("s4:torture: Remove netbios
realm and lowercase realm tests") we removed some test flags, reducing
the number of flags to 8, so there are 256 combinations. But we test
every bit combination of TEST_ALL which was 10 bits (0x3ff), and each
test was run 4 times ignoring the 0x4 and 0x10 bits.

So we compact the flags into 8 bits and run each one once.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

pytest: test auth.user_session with principals

This tests authsam_get_session_info_principal() which tests
sam_get_results_principal() which tests crack_user_principal_name().

sam_get_results_principal() is also used in samba_kdc_lookup_client(),
and we are sort of testing on behalf of that.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:cracknames: initialise a string variable

because later we go

ret = krb5_unparse_name_flags([...],
      &unparsed_name_short);
if (ret) {
free(unparsed_name_short);
return WERR_NOT_ENOUGH_MEMORY;
}

which is bad if a krb5_unparse_name_flags() errors without setting
unparsed_name_short -- not that I see that happening in MIT or Heimdal.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

docs-xml:smb.conf: fix a sentence

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s3:ntlm_auth: Fix typo in error message (protools -> protocols)

Signed-off-by: Frediano Ziglio <freddy77@gmail.com>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Anoop C S <anoopcs@samba.org>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Wed Nov 19 19:38:52 UTC 2025 on atb-devel-224

ctdb: Fix ctdb startup with inconsistent cluster lock settings

ctdb_shutdown_sequence() normally exits. When we end up here, it is
because we have received a reclock callback twice. We can't handle
that, we have already removed "state", which would be referenced deep
in run_start_recovery_event() returning here another time.

The bug is triggered since b84fbd7b3fedc998 introduced a nested event
loop, making ctdb_shutdown_sequence() return into
start_recovery_reclock_callback() due to multiple reclock checks being
triggered somehow (not sure exactly how, but we should not crash under
any circumstance).

Reproducer: Run one ctdb daemon with cluster lock set, try to start
another one without cluster lock set.

Bug: https://bugzilla.samba.org/show_bug.cgi?id=15950
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Wed Nov 19 03:04:13 UTC 2025 on atb-devel-224

mdssvc: support a wider range of years [0000,9999] in $time.iso

Most importantly use strtoll to allow negative numbers and use a filed width
with %Y in strftime() to parse years with less then four digits.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15947

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Wed Nov 19 01:59:34 UTC 2025 on atb-devel-224

tests/krb5: Construct signed_attrs correctly

signed_attrs is supposed to be a list of key‐value pairs, but we forgot the
values. Because the field was not constructed correctly, the pyasn1 encoder
simply stripped it out.

Also properly separate the signature algorithm and digest algorithms.

Signed-off-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Wed Nov 19 00:32:31 UTC 2025 on atb-devel-224

tests/krb5: Add TD_CMS_DIGEST_ALGORITHMS constant

Signed-off-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

wscript: Fix help text spelling

Signed-off-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

python:tests: Correct test class name

Signed-off-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

vfs_fruit: psd->dacl can be NULL, use orig_num_aces

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15926

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Tue Nov 18 10:13:44 UTC 2025 on atb-devel-224

vfs:streams_xattr: Allow VFS_OPEN_HOW_RESOLVE_NO_XDEV

The open function returns a fake fd. Extended attributes will be stored by
vfs_xattr_tdb or vfs_default.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Samuel Cabrero <scabrero@samba.org>
Autobuild-Date(master): Tue Nov 18 09:08:38 UTC 2025 on atb-devel-224

vfs:fruit: Allow RESOLVE_NO_XDEV flag

For stream opens, it returns a fake fd. The streams will be stored by
vfs_streams_depot or vfs_streams_xattr.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs:streams_depot: Allow VFS_OPEN_HOW_RESOLVE_NO_XDEV flag

The flag is passed down the modules stack.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs:shadow_copy2: Allow RESOLVE_NO_XDEV flag

This module updates the path and calls the next VFS module.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs:glusterfs: Allow VFS_OPEN_HOW_RESOLVE_NO_XDEV

Don't return ENOSYS if the flag is set. It will be ignored as does not make
sense in this module.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs:ceph_new: Allow VFS_OPEN_HOW_RESOLVE_NO_XDEV flag

Don't return ENOSYS if the flag is set. It will be ignored,
does not make sense in a ceph filesystem.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs:ceph: Allow VFS_OPEN_HOW_RESOLVE_NO_XDEV flag

Don't return ENOSYS if the flag is set. It will be ignored,
does not make sense in a ceph virtual filesystem.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs:aio_pthread: Handle VFS_OPEN_HOW_RESOLVE_NO_XDEV flag

This module uses openat() instead of openat2() so the flag won't be used and
automounts might not be triggered.

Disable flag usage for subsequent opens and return an error to callers to warn
the user and retry without the flag.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Fix crossing direct automounter mount points

The workaround implemented in commit ac7a16f9cc4bd97ef546d1b7b02605991000d0f9
to trigger automounts does not work for direct automounts (either with
systemd-automount or autofs daemon).

In direct automounts the mount point is a real directory instead of a "ghost"
directory so when turning the O_PATH handle into a real one through
/proc/self/fd/<fdnum> openat() does not return ENOENT, it returs a fd referring
to the mount point without triggering the mount.

To trigger the mount first we have to know when we are crossing mount points
by using the RESOLVE_NO_XDEV flag in open_how.resolve, then we can check with
fstatfs() the .f_type and fallback to a path-based open for automounts or
retry without RESOLVE_NO_XDEV otherwise.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Refactor reopen_from_fsp(), factor out pathref based

Best viewed ignoring white space changes

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Refactor reopen_from_fsp(), factor out automounter mountpoint check

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Refactor reopen_from_fsp(), factor out name based reopen

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs: Pass the RESOLVE_NO_XDEV from upper layers to openat2() syscall

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

selftest/Samba3: nt4_dc* use vfs_default:VFS_OPEN_HOW_RESOLVE_NO_XDEV=no

From 076c22fbd7ecbf22dbfeb1711609f07fd42f88b0, we should always test the
code path without openat2 being available, even if the kernel supports it.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs: Use RESOLVE_NO_XDEV by default on all shares

Enable the flag by default on all shares, it will be automatically
disabled if the system does not support openat2().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs: Bump VFS interface version after adding VFS_OPEN_HOW_RESOLVE_NO_XDEV

If a module was compiled with an older version it won't handle the new flag.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

vfs: Add VFS_OPEN_HOW_RESOLVE_NO_XDEV flag

It disallows traversal of mount points during path resolution, including bind
mounts.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15805

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

ci: Don't run on private rackspace runner

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Nov 17 09:37:30 UTC 2025 on atb-devel-224

smbd: Change open_rootdir_pathref_fsp() to return a smb_filename

Simpler and safer to close in the caller, and
openat_pathref_fsp_rootdir() can use
openat_pathref_fsp_simple_openat(). De-duplicate some logic.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Factor out openat_pathref_fsp_simple_openat()

open_rootdir_pathref_fsp() and openat_pathref_fsp_dot() serve very
similar purposes. Avoid code duplication, this is to be used in
open_rootdir_pathref_fsp() next.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Fix a comment

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Make fsp_set_gen_id() static

Only called from files.c

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Make fsp_new() return a files_struct

There's only the ENOMEM failure condition

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Make fsp_set_smb_fname() return bool

There's only the ENOMEM failure condition

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Make fsp_attach_smb_fname() return bool

There's only the ENOMEM failure condition

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Make fsp_smb_fname_link() return bool

There's only the ENOMEM failure condition

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

smbd: Make file_name_hash() return bool

There's only the ENOMEM failure condition

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

printing: Fix coverity issue CID#1669074 (Unchecked return value)

A new coverity issue (Unchecked return value) is flagged as a regression
due to commit#e9a7dce599e.
This could have been present prior to comit#e9a7dce599e, however got missed.

This change fixes this issue by checking return value.

Signed-off-by: Vinit Agnihotri <vagnihot@redhat.com>
Reviewed-by: Anoop C S <anoopcs@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Anoop C S <anoopcs@samba.org>
Autobuild-Date(master): Mon Nov 17 08:31:24 UTC 2025 on atb-devel-224

python:tests: Take INTERDOMAIN_TRUST into account

If we run e.g. fl2008rc2 env before we run this test, we have a trust account
for this domain.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Nov 14 19:23:38 UTC 2025 on atb-devel-224

python:tests: Add debug output to test_query_filter_enum()

This test runs relatively late in 'make test'. It finds some accounts which are
neither a computer nor a user account and complains that the overall account
count doesn't match.

Add some debug so we can find out more about the accounts.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

python:tests: Use random users for domain_auth_silo.py

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

python:tests: Use random users for user_auth_silo.py

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

python:tests: Use random users for user_auth_policy.py

We should not work on alice, bob and joe. We should use random user so we can
clean up afterwards. If we don't do this the dbcheck tests will fails with:

UNEXPECTED(failure): samba4.blackbox.dbcheck(ad_dc).dbcheck(ad_dc:local)
REASON: Exception: Exception: Checking 21449 objects
WARNING: target DN is deleted for msDS-AssignedAuthNPolicy in object CN=alice,CN=Users,DC=addom,DC=samba,DC=example,DC=com - <GUID=66bd1f51-084f-4259-a769-efa59adb6e31>;<RMD_ADDTIM
E=134051822550000000>;<RMD_CHANGETIME=134051822550000000>;<RMD_FLAGS=1>;<RMD_INVOCID=b5064b02-c91e-40c0-bc1f-a92a0105a810>;<RMD_LOCAL_USN=14061>;<RMD_ORIGINATING_USN=14061>;<RMD_VE
RSION=2>;CN=User Policy,CN=AuthN Policies,CN=AuthN Policy Configuration,CN=Services,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com
Target GUID points at deleted DN 'CN=User Policy\\0ADEL:66bd1f51-084f-4259-a769-efa59adb6e31,CN=Deleted Objects,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com'
Not removing
WARNING: target DN is deleted for msDS-AssignedAuthNPolicySilo in object CN=alice,CN=Users,DC=addom,DC=samba,DC=example,DC=com - <GUID=287d6c3d-bea8-4c06-bdf8-7d4b579bc0e9>;<RMD_AD
DTIME=134051822560000000>;<RMD_CHANGETIME=134051822560000000>;<RMD_FLAGS=1>;<RMD_INVOCID=b5064b02-c91e-40c0-bc1f-a92a0105a810>;<RMD_LOCAL_USN=14081>;<RMD_ORIGINATING_USN=14081>;<RM
D_VERSION=2>;CN=Developers,CN=AuthN Silos,CN=AuthN Policy Configuration,CN=Services,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com
Target GUID points at deleted DN 'CN=Developers\\0ADEL:287d6c3d-bea8-4c06-bdf8-7d4b579bc0e9,CN=Deleted Objects,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com'
Not removing
WARNING: target DN is deleted for msDS-AssignedAuthNPolicy in object CN=bob,CN=Users,DC=addom,DC=samba,DC=example,DC=com - <GUID=66bd1f51-084f-4259-a769-efa59adb6e31>;<RMD_ADDTIME=
134051822550000000>;<RMD_CHANGETIME=134051822550000000>;<RMD_FLAGS=1>;<RMD_INVOCID=b5064b02-c91e-40c0-bc1f-a92a0105a810>;<RMD_LOCAL_USN=14065>;<RMD_ORIGINATING_USN=14065>;<RMD_VERS
ION=4>;CN=User Policy,CN=AuthN Policies,CN=AuthN Policy Configuration,CN=Services,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com
Target GUID points at deleted DN 'CN=User Policy\\0ADEL:66bd1f51-084f-4259-a769-efa59adb6e31,CN=Deleted Objects,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com'
Not removing
WARNING: target DN is deleted for msDS-AssignedAuthNPolicySilo in object CN=bob,CN=Users,DC=addom,DC=samba,DC=example,DC=com - <GUID=1a1e5cdf-b92e-4a80-bc35-cccad8e9f865>;<RMD_ADDT
IME=134051822560000000>;<RMD_CHANGETIME=134051822560000000>;<RMD_FLAGS=1>;<RMD_INVOCID=b5064b02-c91e-40c0-bc1f-a92a0105a810>;<RMD_LOCAL_USN=14085>;<RMD_ORIGINATING_USN=14085>;<RMD_
VERSION=4>;CN=QA,CN=AuthN Silos,CN=AuthN Policy Configuration,CN=Services,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com
Target GUID points at deleted DN 'CN=QA\\0ADEL:1a1e5cdf-b92e-4a80-bc35-cccad8e9f865,CN=Deleted Objects,CN=Configuration,DC=addom,DC=samba,DC=example,DC=com'
Not removing
Checked 21449 objects (4 errors)

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

selftest: Create libdir in setup_clusteredmember()

This fixes:

can't open st/clusteredmember/lib/krb5.conf at selftest/target/Samba.pm line 328.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

selftest: Add blackbox claims test to knownfail_mit_kdc

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>

bootstrap: Remove libtracker-sparql-2

This is not used anymore.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Nov 14 16:34:48 UTC 2025 on atb-devel-224

s3:mdssvc: Remove support for tracker sparql support

This only support v2 of tracker sparql which is only avilable on the Ubuntu
20.04 runner. All newer distributions provide only tracker sparql v3.

Also the tevent glib stopped working on Fedora 43.

Time to remove it.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>

docs-xml: Improve the samba-bgqd manpage

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15809

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Anoop C S <anoopcs@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Nov 14 15:05:11 UTC 2025 on atb-devel-224

s3:printing: Load the shares for [printers] in samba-bgqd

One of the main functions of bgqd is:

        delete_and_reload_printers_full()

It isn't able to do its work, if we don't load the shares. Normally bgqd was
forked from smbd and this loaded the shares. But with the introduction of
samba-dcerpcd it is a standalone service now. As a standalone service it is
responsible to load the shares if it needs to work on them.

The following message is printed if delete_and_reload_printers_full() tries to
do its job:

[2025/10/23 09:57:27,  7, pid=41935, effective(0, 0), real(0, 0)] ../../source3/param/loadparm.c:4419(lp_servicenumber)
  lp_servicenumber: couldn't find printers
[2025/10/23 09:57:27,  7, pid=41935, effective(0, 0), real(0, 0)] ../../source3/param/loadparm.c:4419(lp_servicenumber)
  lp_servicenumber: couldn't find printers

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15936

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Anoop C S <anoopcs@samba.org>

s3:smbd: Fix segfault in smb2_reply

This is triggered by: make test TESTS="samba3.smb2.streams.simpleserver

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15948

Pair-Programmed-With: Ralph Boehme <slow@samba.org>
Signed-off-by: Andreas Schneider <asn@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>

s4:kdc always include the PAC

Set the heimdal always_include_pac configuration flag, based on the samba
kdc always include pac option

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Thu Nov 13 23:12:55 UTC 2025 on atb-devel-224

s4:kdc:tests: support "kdc always generate pac"

Update the tests to check the "kdc always generate pac" configuration and
expect the presence of a PAC accordingly.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>

s4:kdc:test add tgs tests to fl2008r2dc env

Add kdc_tgs tests to the fl2008r2dc test environment, to ensure that they are
run with "kdc always generate pac" set to no

Note: This required updating known_fail_mit/kdc_tgs to handle the
      expected failures for the fl2008r2dc environment when run against
      the MIT kdc

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Jennifer Sutton <jennifersutton@catalyst.net.nz>