third_party: Update waf to verison 2.0.23

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Mon Feb 21 10:06:27 UTC 2022 on sn-devel-184

third_party:waf: Print the version of waf at the end of the update script

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

editorconfig: Final newlines are pycodestyle

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

waf: re-add missing readlink test

this was another portability regression that came with the moving to waf

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13631

Signed-off-by: Bjoern Jacke <bj@sernet.de>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Feb 18 23:12:51 UTC 2022 on sn-devel-184

readlink test: inverse return code

We need to return 0 in case readlink is *broken* here - this is because our waf
CHECK_CODE function does only allow generating defines in case the test succeeds

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13631

Signed-off-by: Bjoern Jacke <bj@sernet.de>
Reviewed-by: Jeremy Allison <jra@samba.org>

wscript: s/default/required/ _static_modules for the acl modules

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14974

Signed-off-by: Bjoern Jacke <bjacke@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfs_aixacl: add proper header file

BUG: https://bugzilla.samba.org/show_bug.cgi?id=7239

Signed-off-by: Bjoern Jacke <bjacke@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

acl: fix function arguments for AIX' and Solaris' sys_acl_get_fd()

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14974

Signed-off-by: Bjoern Jacke <bjacke@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3: smbd: Fix our leases code to return the correct error in the non-dynamic share case.

We now return INVALID_PARAMETER when trying to open a
different file with a duplicate lease key on the same
(non-dynamic) share. This will enable us to pass another
Windows test suite leases test.

We now behave the same as Windows10.

Remove knownfail.d/smb2-lease-duplicateopen

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14737

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Mulder <dmulder@suse.com>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Feb 18 20:12:12 UTC 2022 on sn-devel-184

s4: torture: Add new SMB2 lease test test_lease_duplicate_open().

Checks we return INVALID_PARAMETER when trying to open a
different file with a duplicate lease key on the same share.

Checked against Windows10. Currently fails against smbd
so add knownfail.d/smb2-lease-duplicateopen

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14737

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Mulder <dmulder@suse.com>

s4: torture: Add new SMB2 lease test test_lease_duplicate_create().

Checks we return INVALID_PARAMETER when trying to create a
new file with a duplicate lease key on the same share.

Checked against Windows10. Samba already passes this
but we didn't have a test before.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14737

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Mulder <dmulder@suse.com>

source4/torture: Drop unused variable attribute

clang complains:

../../source4/client/client.c:1569:11: error: variable 'attribute' set but not used [-Werror,-Wunused-but-set-variable]
        uint16_t attribute = FILE_ATTRIBUTE_SYSTEM | FILE_ATTRIBUTE_HIDDEN;
                 ^

That is, the variable is initialised and updated but the value is
never used.

Commit 2f377d5101783ed4d8c96a46aaec61895cc7b6ad from 2004 dropped the
use of this variable.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Feb 17 19:06:25 UTC 2022 on sn-devel-184

source4/torture: Drop unused variable mask

clang complains:

../../source4/torture/smb2/notify.c:871:11: error: variable 'mask' set but not used [-Werror,-Wunused-but-set-variable]
        uint32_t mask;
                 ^

That is, the variable is initialised and updated but the value is
never used.

Looks to have been this way since commit
15d93a5d8e21893e1cca5c989dbf97010aae1622 from 2009.  Just drop it.

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

source4/torture: Drop unused variable tdif

clang complains:

../../source4/torture/basic/denytest.c:1805:11: error: variable 'tdif' set but not used [-Werror,-Wunused-but-set-variable]
                int64_t tdif;
                        ^

That is, the variable is initialised and updated but the value is
never used.

Perhaps it is meant to be used in the nearby torture_comment() call,
but it has been this was since commit
cb1cff90f165d82cbbf1dd87e475a1b13984d45e from 2004.  Just drop it.

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

source4/torture: Avoid unused variable

clang complains:

../../source4/torture/basic/delete.c:2342:7: error: variable 'correct' set but not used [-Werror,-Wunused-but-set-variable]
        bool correct = true;
             ^

That is, the variable is initialised and updated but the value is
never used.  Similar functions return this variable, so try that.

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

vfs_not_implemented: do not mark structs with _PUBLIC_

Commit 5d295e41af4e9316aee1b4cf1c3087663b7c06a4 accidentally marked
some structs with _PUBLIC_, which causes clang to complain:

../../source3/modules/vfs_not_implemented.c:594:1: error: attribute 'visibility' is ignored, place it after "struct" to apply attribute to type declaration [-Werror,-Wignored-attributes]
_PUBLIC_
^
../../lib/replace/replace.h:917:33: note: expanded from macro '_PUBLIC_'
                                ^
../../source3/modules/vfs_not_implemented.c:642:1: error: attribute 'visibility' is ignored, place it after "struct" to apply attribute to type declaration [-Werror,-Wignored-attributes]
_PUBLIC_
^
../../lib/replace/replace.h:917:33: note: expanded from macro '_PUBLIC_'
                                ^

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

util: Drop unused variable mask_perms

clang complains:

../../source3/smbd/posix_acls.c:2783:9: error: variable 'mask_perms' set but not used [-Werror,-Wunused-but-set-variable]
        mode_t mask_perms = 0;
               ^

That is, the variable is initialised and updated but the value is
never used.

This potentially points to a bug in commit
f735551b9edef66b152261cf6eb2f29b7b69d65b from 2002.

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

util: Drop unused variable num_received

clang complains:

../../source4/libcli/clilist.c:111:6: error: variable 'num_received' set but not used [-Werror,-Wunused-but-set-variable]
        int num_received = 0;
            ^
../../source4/libcli/clilist.c:268:6: error: variable 'num_received' set but not used [-Werror,-Wunused-but-set-variable]
        int num_received = 0;
            ^

That is, the variable is initialised and updated but the value is
never used.

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

util: Drop unused variable num_chars

clang complains:

../../lib/util/genrand_util.c:99:9: error: variable 'num_chars' set but not used [-Werror,-Wunused-but-set-variable]
        size_t num_chars = 0;
               ^

That is, the variable is initialised and incremented but the value is
never used.

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

smbd: Make strnorm() static to filename.c

The caller in vfs_prealloc was a bit unneeded, and strnorm is only
called here.

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): Thu Feb 17 18:11:18 UTC 2022 on sn-devel-184

vfs: Don't go through strnorm(..., CASE_LOWER)

With a fixed CASE_LOWER we should go directly to the lowerlevel call, this
makes it more obvious to me.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfs: Fix a typo

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Use ISDOT/ISDOTDOT

This is simpler to read for me

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Filter out "::$DATA" for query name information

Make this piece of code robust against having "::$DATA" passed in.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Simplify open_file_ntcreate()

For streams our caller create_file_unixpath() has already taken care
of properly initializing fsp->base_fsp, so we can rely on
fsp_is_alternate_stream() here instead of looking at the file name.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Only open base_fsp for non-"::$DATA" streams

"is_named_stream()" is more what we really mean here. Make this line robust
against callers passing in "::$DATA".

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfs: Use is_named_stream() for checking if we have an ADS

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Use fsp_is_alternate_stream(), we checked for fsp!=NULL above

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Slightly simplify openat_pathref_fsp()

We don't need to look at the full share-relative path to figure out if
we have a stream name, the original smb_fname is sufficient for this.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfs: Use fsp_get_pathref_fd() in aio_pthread

We only use the fd as "dirfd" in openat, so we don't need an I/O fd
here.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

This involves converting cli_print_queue() to NTSTATUS. No caller
looked at the number of jobs returned.

Review with "git show -b", most of the patch is indentation

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfs_gpfs: use linux oplock specific funcions only when available

Signed-off-by: Bjoern Jacke <bj@sernet.de>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Christof Schmitt <cs@samba.org>
Autobuild-Date(master): Wed Feb 16 21:38:12 UTC 2022 on sn-devel-184

dns.idl/dnsp.idl: add missing DNS ressource record types

Signed-off-by: Bjoern Jacke <bjacke@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Feb 16 20:43:55 UTC 2022 on sn-devel-184

dnsp.idl: add missing DNS_RPC_RECORD defines

taken from MSDN 2.2.2.2.5:
https://docs.microsoft.com/en-us/openspecs/windows_protocols/ms-dnsp/ac793981-1c60-43b8-be59-cdbb5c4ecb8a

Signed-off-by: Bjoern Jacke <bjacke@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

docs-xml: Fix idmap_autorid documentation

What we want to avoid:

$ ./bin/testparm -s | grep "idmap config"
        idmap config * : rangesize = 10000
        idmap config * : range = 10000-19999
        idmap config * : backend = autorid

$ ./bin/wbinfo --name-to-sid BUILTIN/Administrators
S-1-5-32-544 SID_ALIAS (4)

$ ./bin/wbinfo --sid-to-gid S-1-5-32-544
10000

$ ./bin/wbinfo --name-to-sid ADDOMAIN/alice
S-1-5-21-4058748110-895691256-3682847423-1107 SID_USER (1)

$ ./bin/wbinfo --sid-to-gid S-1-5-21-984165912-589366285-3903095728-1107
failed to call wbcSidToGid: WBC_ERR_DOMAIN_NOT_FOUND
Could not convert sid S-1-5-21-984165912-589366285-3903095728-1107 to gid

If only one range is configured we are either not able to map users/groups
from our primary *and* the BUILTIN domain. We need at least two ranges to also
cover the BUILTIN domain!

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14967

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Wed Feb 16 17:04:53 UTC 2022 on sn-devel-184

s3:utils: Add a testparm check for idmap autorid

What we want to avoid:

$ ./bin/testparm -s | grep "idmap config"
        idmap config * : rangesize = 10000
        idmap config * : range = 10000-19999
        idmap config * : backend = autorid

$ ./bin/wbinfo --name-to-sid BUILTIN/Administrators
S-1-5-32-544 SID_ALIAS (4)

$ ./bin/wbinfo --sid-to-gid S-1-5-32-544
10000

$ ./bin/wbinfo --name-to-sid ADDOMAIN/alice
S-1-5-21-4058748110-895691256-3682847423-1107 SID_USER (1)

$ ./bin/wbinfo --sid-to-gid S-1-5-21-984165912-589366285-3903095728-1107
failed to call wbcSidToGid: WBC_ERR_DOMAIN_NOT_FOUND
Could not convert sid S-1-5-21-984165912-589366285-3903095728-1107 to gid

If only one range is configured we are either not able to map users/groups
from our primary *and* the BUILTIN domain. We need at least two ranges to also
cover the BUILTIN domain!

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14967

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

s3:winbindd: Add a sanity check for the range

What we want to avoid:

$ ./bin/testparm -s | grep "idmap config"
        idmap config * : rangesize = 10000
        idmap config * : range = 10000-19999
        idmap config * : backend = autorid

$ ./bin/wbinfo --name-to-sid BUILTIN/Administrators
S-1-5-32-544 SID_ALIAS (4)

$ ./bin/wbinfo --sid-to-gid S-1-5-32-544
10000

$ ./bin/wbinfo --name-to-sid ADDOMAIN/alice
S-1-5-21-4058748110-895691256-3682847423-1107 SID_USER (1)

$ ./bin/wbinfo --sid-to-gid S-1-5-21-984165912-589366285-3903095728-1107
failed to call wbcSidToGid: WBC_ERR_DOMAIN_NOT_FOUND
Could not convert sid S-1-5-21-984165912-589366285-3903095728-1107 to gid

If only one range is configured we are either not able to map users/groups
from our primary *and* the BUILTIN domain. We need at least two ranges to also
cover the BUILTIN domain!

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14967

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

bootstrap: Install mold linker on Fedora 35

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Tue Feb 15 12:31:43 UTC 2022 on sn-devel-184

bootstrap: If the mold linker is available prefer it over gold

The gold linker will be deprecated soon. However we got a new linker called
mold:

https://github.com/rui314/mold/

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

builtools: Make abi_gen.sh less prone to errors

The mold linker has more hidden symbols and we would need to filter them out
with nm, where objdump tells us which symbols are actually hidden. So we just
need to filter out whatever is hidden.

The use of awk makes it also easier to get what we want.

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

ctdb-tests: Iterate protocol tests internally

Instead of repeatedly running a test binary.

Run time for these tests reduces from ~90s to ~75s.

When run under valgrind, the run time for protocol_test_001.sh reduces
from ~390s to <1s.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Autobuild-User(master): Amitay Isaacs <amitay@samba.org>
Autobuild-Date(master): Mon Feb 14 04:32:29 UTC 2022 on sn-devel-184

ctdb-tests: Add iteration support for protocol tests

The current method of repeatedly running a binary has huge overhead,
especially with valgrind.

protocol_test_iterate_tag() allows output that is usually used for
hinting where a test failure occurred to be replaced with a tag
stored in a buffer, which is printed on test failure.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-tests: Add a test for stalled node triggering election

A stalled node probably continues to hold the cluster lock, so confirm
elections work in this case.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14958

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Autobuild-User(master): Amitay Isaacs <amitay@samba.org>
Autobuild-Date(master): Mon Feb 14 02:46:01 UTC 2022 on sn-devel-184

ctdb-tests: Factor out functions to detect when generation changes

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14958

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-recoverd: Consistently log start of election

Elections should now be quite rare, so always log when one begins.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14958

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-recoverd: Always send unknown leader broadcast when starting election

This is currently missed when the cluster lock is lost.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14958

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-recoverd: Consistently have caller set election-in-progress

The problem here is that election-in-progress must be set to
potentially avoid restarting the election broadcast timeout in
main_loop(), so this is already done by leader_handler().

Have force_election() set election-in-progress for all election types
and do not bother setting it in cluster_lock_election().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14958

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

ctdb-recoverd: Always cancel election in progress

Election-in-progress is set by unknown leader broadcast, so needs to
be cleared in all cases when election completes.

This was seen in a case where the leader node stalled, so didn't send
leader broadcasts for some time.  The node continued to hold the
cluster lock, so another node could not become leader.  However, after
the node returned to normal it still did not send leader broadcasts
because election-in-progress was never cleared.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14958

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Amitay Isaacs <amitay@gmail.com>

smbd: Simplify smbd_dirptr_lanman2_mode_fn()

Avoid an else, we return in the "true" branch

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): Fri Feb 11 21:53:22 UTC 2022 on sn-devel-184

torture: Align integer types

finfo.stream_info.out.num_streams is declared as "unsigned int"

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

ndrdump: Small simplification

Remove the talloc_steal(), we can allocate on mem_ctx directly

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Use fstrcpy where possible

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Safeguards for getpwuid

Attempt to fix

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14900

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfstest: Align two integer types

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

vfs: Simplify streams_xattr_unlinkat()

It would be a logic error to call rmdir on a stream. This simplifies
the logic a bit.

Signed-off-by: Volker Lendecke <vl@samba.org>

smbd: Use fsp_is_alternate_stream() where an fsp is available

Make it clear that being an alternate data stream handle is much more
a fsp property than a file name property.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Introduce metadata_fsp()

Centralize the pattern

if (fsp->base_fsp != NULL) {
fsp = fsp->base_fsp;
}

with a descriptive name.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Introduce fsp_is_alternate_stream()

To me this is more descriptive than "fsp->base_fsp != NULL". If this
turns out to be a performance problem, I would go and make this a
static inline in smbd/proto.h.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

wafsamba: replace 'echo -n' with printf

This patch makes samba_cross.py compatible with old bash (e.g. 3.2)

Signed-off-by: Sergey V. Lobanov <sergey@lobanov.in>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Feb 11 07:58:57 UTC 2022 on sn-devel-184

s3:modules: Fix virusfilter_vfs_openat

Bug: https://bugzilla.samba.org/show_bug.cgi?id=14971

Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Pair-Programmed-With: Andreas Schneider <asn@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Feb 10 22:09:06 UTC 2022 on sn-devel-184

s3:selftest: Add test for virus scanner

Bug: https://bugzilla.samba.org/show_bug.cgi?id=14971

Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Pair-Programmed-With: Andreas Schneider <asn@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

selftest: Fix trailing whitespace in Samba3.pm

Bug: https://bugzilla.samba.org/show_bug.cgi?id=14971

Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

docs-xml:manpages: Document 'dummy' virusfilter and 'virusfilter:infected files'

Bug: https://bugzilla.samba.org/show_bug.cgi?id=14971

Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

s3:modules: Implement dummy virus scanner that uses filename matching

Bug: https://bugzilla.samba.org/show_bug.cgi?id=14971

Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>

selftest: Do not force -d0 for smbd/nmbd/winbindd

We have the env variable SERVER_LOG_LEVEL which allows you to change
the log level on the command line. If we force -d0 this will not work.

make test TESTS="samba" SERVER_LOG_LEVEL=10

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Only file_free() a self-created fsp in create_file_unixpath()

This fixes a use-after-free in smb_full_audit_create_file() when
calling SMB_VFS_CREATE_FILE with fsp->fsp_name as smb_fname.

create_file_unixpath() has this comment:

 * This is really subtle. If someone passes in an smb_fname
 * where smb_fname actually is taken from fsp->fsp_name, then
 * the lifetime of these objects is meant to be the same.

so it seems legitimate to call CREATE_FILE this way.

When CREATE_FILE runs into an error, create_file_unixpath() does a
file_free, which also takes fsp->fsp_name with
it. smb_full_audit_create_file() wants to log the failure including
the smb_fname after NEXT_CREATE_FILE has exited, but this will then
use the already free'ed data.

Fix by only doing the file_free() on an fsp that
create_file_unixpath() created itself.

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): Thu Feb 10 19:11:33 UTC 2022 on sn-devel-184

smbd: Introduce close_file_smb()

This does almost everything that close_file_free() does, but it leaves
the fsp around.

A normal close_file() now calls fsp_unbind_smb() twice. Functionally
this is not a problem, fsp_unbind_smb() is idempotent. The only
potential performance penalty might come from the loops in
remove_smb2_chained_fsp(), but those only are potentially large with
deeply queued smb2 requests. If that turns out to be a problem, we'll
cope with it later. The alternative would be to split up file_free()
into even more routines and make it more difficult to figure out which
of the "rundown/unbind/free" routines to call in any particular
situation.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Factor out fsp_unbind_smb() from file_free()

For example, remove our entry from smbXsrv_open_global.tdb

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

torture: Add a test to show that full_audit uses a ptr after free

Run vfstest with this vfstest.cmd under valgrind and you'll see what
happens. Exact explanation a few patches further down...

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Simplify the flow in close_file_free()

We are no longer called on base_fsp's in SHUTDOWN_CLOSE. That
simplifies the logic in the common case, we now have a linear flow for
the very often-called close_file()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: No base fsps to close_file_free() from file_close_user()

Same logic as the change for file_close_conn()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Factor out close_file_in_loop() from file_close_conn_fn()

To be reused in file_close_user(). Deliberately a separate commit to
make the previous commit easier to understand.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: No base fsps to close_file_free() from file_close_conn()

close_file_free() needs to handle base fsps specially. This can be
simplified a lot if we pass the the open files a second time in case
we encountered base_fsps that we could not immediately delete.

file_close_conn() is not our hot code path, and also we don't expect
many thousand open files that we need to walk a second time.

A subsequent patch will simplify close_file_free(), the complicated
logic is now in files.c, where it IMHO belongs because
file_set_base_fsp() are here as well.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: NULL out "fsp" in close_file()

Quite a few places already had this in the caller, but not all. Rename
close_file() to close_file_free() appropriately. We'll factor out
close_file_smb() doing only parts of close_file_free() later.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Call file_free() just once in close_file()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Move the call to file_free() out of close_fake_file()

Centralize calling file_free(), but leave close_fake_file() in for API
symmetry reasons.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Move the call to file_free() out of close_normal_file()

Call file_free() just once

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Move the call to file_free() out of close_directory()

Call file_free() just once

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Slightly simplify create_file_unixpath()

Avoid the "needs_fsp_unlink" variable, describe the talloc hierarchy a
bit differently in the comments.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3:winbind: Reduce the level and improve a couple of debug messages

The commit 1d5c546 changed the debug message printed when setting
winbind to offline state and offline logons are disabled from
level 10 to level 0. This message isn't really an error and might
scare some users, e.g. https://bugzilla.suse.com/show_bug.cgi?id=1195573

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Feb  9 20:20:36 UTC 2022 on sn-devel-184

s3/libads: ensure a sockaddr variable is correctly zero initialized

is_zero_addr() doesn't work with addresses that have been zero-initialized.

This fixes the logic added in c863cc2ba34025731a18ac735f714b5b888504da.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14674
MR: https://gitlab.com/samba-team/samba/-/merge_requests/2354

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Feb  8 20:24:12 UTC 2022 on sn-devel-184

s3/libads: simplify storing existing ads->ldap.ss

We just need temporal storage for ads->ldap.ss, no need to store it as a struct
samba_sockaddr.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14674
MR: https://gitlab.com/samba-team/samba/-/merge_requests/2354

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib: Simplify pm_process()

No need to duplicate the fopen/fclose

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): Mon Feb  7 19:58:57 UTC 2022 on sn-devel-184

vfs: Fix a typo

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Pass "dirfsp" and "smb_fname" to fd_open_atomic()

Dereference fsp once instead of four times

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Avoid an "else" in file_set_dosmode()

Review with git show -b

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Fix open_pathref_base_fsp()'s implicit conn_cwd assumption

Opening a stream base file only worked if "dirfsp == conn->cwd_fsp":
We have replaced fsp->fsp_name with the full dirfsp->relative pathname
at the point where open_pathref_base_fsp() is called. In case dirfsp
is already a subdirectory in a share, this breaks because the
open_pathref_base_fsp() uses fsp->fsp_name, not the original
dirfsp-relative one.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Convert SMBC_getatr() to NTSTATUS

This avoids a few calls to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid two calls to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Avoid a call to SMBC_errno()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

smbd: Fix CID 1497981: Null pointer dereferences (REVERSE_INULL)

brown paper bag quality, sorry...

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib: Fix CID 1465285 Double close

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

script/autobuild.py: let nm_grep_symbols ignore __gcov_ symbols

Currently the gcov build currently fails with the following error:

samba-libs: [allshared-no-public-nss_winbind] Running nm ./bin/plugins/libnss_winbind.so.2 | egrep -v ' (__bss_start|_edata|_init|_fini|_end)' | egrep -v ' T _nss_winbind_' |egrep ' [BDGTRVWS] ' && exit 1; exit 0; in '/tmp/samba-testbase/samba-libs/.'
0000000000232458 B __gcov_error_file
0000000000226340 D __gcov_master
000000000001c080 T __gcov_sort_n_vals
00000000002324a0 B __gcov_var
samba-libs: [allshared-no-public-nss_winbind] failed 'nm ./bin/plugins/libnss_winbind.so.2 | egrep -v ' (__bss_start|_edata|_init|_fini|_end)' | egrep -v ' T _nss_winbind_' |egrep ' [BDGTRVWS] ' && exit 1; exit 0;' with status 1

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Sun Feb  6 13:39:09 UTC 2022 on sn-devel-184

bootstrap: Migrate to CentOS8 Stream

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Feb  4 21:11:40 UTC 2022 on sn-devel-184

sharesec: Add SEC_DIR_DELETE_CHILD to CHANGE permissions

Otherwise you can't rename or delete files using CHANGE permissions using
the sharesec or shareacls utility

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): Fri Feb  4 19:36:53 UTC 2022 on sn-devel-184