docs/smbclient: document -Tcn

BUG:https://bugzilla.samba.org/show_bug.cgi?id=2352
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

man net: include --json option

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

net idmap check: correct spelling of --force

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/scripting/smbstatus: approach py3 compatibility

It is still unused and untested, probably not working.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

pytests/subunitrun: not usefully executable

it only defines classes and fiddles with signals.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/tests/dsdb_schema: not usefully executable as script

You could run the script, but it wouldn't do anything.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/script/rodcdns: str type doesn't need decoding

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

wintest: py3 telnet.sendline() doesn't like string with ^Z

Try using bytes instead.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

script/bisect-test: fix for py3

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/scripting/autoidl: another py3 incompatible except

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

repl_md: Avoid dropping cross-partition links

Cross-partition links could still be dropped if GET_TGT was already
previously set for the replication.

This was due to a slight error in the order of logic. We never want to
ignore cross-partition links (regardless of whether the TARGETS_UPTODATE
/GET_TGT flag is set). We should only be returning early in the
GET_TGT case if the objects are both in the same partition.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14022
RN: When the AD domain contained a linked attribute that spanned
partitions, DRS replication could drop the link. This dropped link could
then result in subtle differences in behaviour between DCs, as some DCs
would have the link and others wouldn't. When this issue occurred, the
dropped link would be logged in a warning message:
 "<target-dn> is Unknown but up to date. Ignoring link from <source-dn>"
This issue would not always occur - it depended a lot on the database
contents. Typically, it would only potentially occur when joining a new
DC to the domain (doing an ldapcmp after the join would also highlight
the problem, if it occurred). This issue has now been resolved.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: Add getncchanges test for cross-partition links + TGT

This adds a test-case to highlight a bug in the client side GetNCChanges
handling.

These tests mostly exercise the server-side behaviour of sending the
GetNCChanges, however, there's a bug in the client-side code when we try
to handle a missing cross-partition link target *in combination* with
the GET_TGT flag already having been set.

The test is exercising the client-side code by using the 'samba-tool drs
replicate' command. By adding a one-way link to a deleted target object,
we force the client code to retry with the GET_TGT flag set.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14022

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

WHATSNEW.txt: samba-tool --backend-size-parameter

Document the --backend-store-size option added to the samba-tool sub
commands:
 * domain provision
 * domain join
 * domain dcpromo
 * drs clone-dc-database

And the removal os the join subdomain option.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Gary Lockyer <gary@samba.org>
Autobuild-Date(master): Tue Jul  2 04:19:29 UTC 2019 on sn-devel-184

samba-tool domain join: remove the subdomain option

Remove the sub domain option from join, as it currently does not work.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

join subdomain: changes for --backend-store-size

Pass the 'backend_store_size' parameter through to DCJoinContext.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-tool domain dcpromo: add --backend-store-size option

Add a new "samba-tool domain dcpromo" option "backend-store-size".
This allows the lmdb map size to be set during a promotion, instead of
hard-wiring it to 8Gb.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-tool dcpromo tests: add --backend-store-size option

Add a new "samba-tool domain dcpromo" option "backend-store-size".
This allows the lmdb map size to be set during a promotion, instead of
hard-wiring it to 8Gb.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-tool clone-dc-database: Add --backend-store-size option

Add a new "samba-tool drs clone-dc-database" option "backend-store-size".
This allows the lmdb map size to be set during a clone, instead of
hard-wiring it to 8Gb.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>'
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

clone-dc-database tests: Add --backend-store-size option

Add a new "samba-tool drs clone-dc-database" option "backend-store-size".
This allows the lmdb map size to be set during a clone, instead of
hard-wiring it to 8Gb.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>'
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-tool domain join: Add --backend-store-size option

Add a new "samba-tool domain join" option "backend-store-size".
This allows the lmdb map size to be set during a provision, instead of
hard-wiring it to 8Gb.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

domain join tests: Add --backend-store-size option.

Tests for the new "samba-tool domain join" option
"backend-store-size". This allows the lmdb map size to be set during a
provision, instead of hard-wiring it to 8Gb.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

provision: Add --backend-store-size option

Add a new "samba-tool domain provision" option "backend-store-size".
This allows the lmdb map size to be set during a provision, instead of
hard-wiring it to 8Gb

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

lib ldb ldb_mdb: Pass the lmdb map size as an ldb option

Allow the lmdb map size to be specified in the ldb option
"lmdb_env_size".

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb: Release ldb 2.0.4

* copy the ldb_options passed to ldb_connect onto the ldb_context,
  making them more generally available.

* fix index buffering.
  As a performance enhancement the indexes are cached in memory during a
  transaction, and written to disk as part of the prepare commit. The
  indexes could become corrupt in the event of a failed operation.

* fix read beyond buffer
  Calling the "ldb_parse_tree" function with a filter consisting of
  exactly a single space (" ") would trigger a read beyond the input
  buffer.
  BUG: https://bugzilla.samba.org/show_bug.cgi?id=13900

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

lib ldb: save a copy of the options on the context

Copy the options supplied to to ldb_connect, and place them on the
ldb_context. This allows backend options i.e. lmbd map size to be passed
cleanly from the callers.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

provision: If --targetdir has been specified then we must always reset posix:eadb and xattr_tdb:file

This is required because otherwise this may have been set previously by another
part of selftest thanks to the global loadparm and the running of multiple
samba-tool tests within the same python process.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

samba-tool domain provision: add lmdb database size option

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-tool: Make the 'bytes' option type avaiable

Allow samba tool to use the custom bytes option type.
  Option("--size", type="bytes", metavar="SIZE")
To allow the input of file and memory sizes using unit suffixes i.e.  2Gb,
4KiB ...

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

python getopt: Add bytes option type

Add a new option type to the python command line options.
  Option("--size", type="bytes", metavar="SIZE")
To allow the input of file and memory sizes using unit suffixes i.e.
2Gb, 4KiB ...

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

provision tests: Add --backend-store-size option.

Tests for the new "samba-tool domain provision" option
"backend-store-size".  This allows the lmdb map size to be set during a
provision, instead of hard-wiring it to 8Gb

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: add a test that itime is not set when setting DOS attrs

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): Mon Jul  1 23:21:07 UTC 2019 on sn-devel-184

vfs_fruit: change default for "fruit:zero_file_id" to false

Now that we have reliable File-IDs, change the default for the option
"zero_file_id" to false.

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

vfs_fruit: make "fruit:zero_file_id" a per share option

Now that File-ID calculation goes through the VFS, we can nicely make a
per-share option out of it.

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

vfs_fruit: move zero file-id to vfs_fruit

Now that this stuff goes through the VFS, let's do it right. :)

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

s3:smbd: set file_id in fetch_dos_mode_done()

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

s3:smbd: use stored file_id in SMB_VFS_FS_FILE_ID()

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

s3:vfs: move get_fs_file_id to vfs_default

This makes get_fs_file_id() the default implementation of
SMB_VFS_FS_FILE_ID(). No change in behaviour.

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

s3:vfs: add SMB_VFS_FS_FILE_ID()

Not yet used, that comes next.

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

s3:smbd: rename get_FileIndex() to get_fs_file_id()

Also rename all variable used to store the result of the renamed
get_fs_file_id() in the callers.

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

s3:smbd: return inode number, not FileIndex for UNIX query info level

get_FileIndex() returns a stable and unique numerical identifier when "store dos
attributes" is enabled. It is NOT the same as a file's inode number which used
as the key for locking.tdb.

For POSIX clients we should return the inode numbers in the POSIX query info
level and the FileIndex in the corresponding SMB protocol fields (eg SMB2 QFID
create context).

This fix lets test "smbtorture_s3.plain.CLEANUP3" test work with get_FileIndex()
returning something different then the file's inode number.

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

s3:smbd: use xattr_DosInfo4 in set_ea_dos_attribute()

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

s3:smbd: parse xattr_DosInfo4 in parse_dos_attribute_blob()

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

s3:smbd: mark itime non-calculated if we created a file or dir

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

s3/lib: add make_file_index_from_itime()

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

s3/lib: add update_stat_ex_file_id()

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

s3/lib: add update_stat_ex_itime()

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

s3: add st_ex_file_id to struct stat_ex

st_ex_file_id is an immutable, never reused numeric identifier for objects in a
filesystem.

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

s3: add st_ex_itime to struct stat_ex

st_ex_itime is an immutable original birth time aka instantiation time. Set when
a file is created, never changes thereafter. May not be set by the client.

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

s3: convert struct stat_ex st_ex_calculated_birthtime bool to flags

Subsequent commits will add more flags, this paves the way.

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

s3: remove unused st_ex_mask from struct stat_ex

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

idl: add xattr_DosInfo4 to xattr_DosInfo in xattr.idl

Adding itime, removing unused ea_size, size, alloc_size and change_time.

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

vfs_catia: pass stat info to synthetic_smb_fname()

This doesn't cause visible damage in vanilla Samba, but would affect downstream
consumers that add additional fields to struct smb_filename.

For the same reason there's no test.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14015
RN: Ensure vfs_catia passes stat info to stacked VFS modules

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

s3:winbind: Add support for storing KRB5 credential in KCM

This can store crentiials in the Kerberos Credential Manager e.g.
provided by sssd.

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 Jul  1 19:22:02 UTC 2019 on sn-devel-184

lib tdb: memcmp ubsan warning

Fix the ubsan warning

lib/tdb/common/tdb.c:184:9: runtime error: null pointer passed as
argument 2, which is declared to never be null"

memcmp call now guarded by a length check.

memcmp returns zero when called with a zero length parameter.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Mon Jul  1 14:50:54 UTC 2019 on sn-devel-184

lib/krb5_wrap: clang: Fix warning: Null pointer passed as an argument

Fixes:

lib/krb5_wrap/krb5_samba.c:3241:3: warning: Null pointer passed as an argument to a 'nonnull' parameter <--[clang]
                memcpy(gss_cksum + 28, in_data->data, orig_length);
                ^

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

lib/krb5_wrap: clang: Fix warning: Call to function 'mktemp' is insecure

Fixes:

lib/krb5_wrap/krb5_samba.c:2012:2: warning: Call to function 'mktemp' is insecure as it always creates or uses insecure temporary file.  Use 'mkstemp' instead <--[clang]
        mktemp(tmp_name);
        ^~~~~~

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

lib/util: clang: Fix warning: Value stored to 'ret' is never read warning

Fixes:

lib/util/server_id_db.c:181:3: warning: Value stored to 'ret' is never read <--[clang]
                ret = tdb_store(tdb, key, talloc_tdb_data(ids), TDB_MODIFY);
                ^     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

lib/ldb/common: clang: Fix Value stored to 'ret' is never read warning

Fixes:

lib/ldb/common/ldb.c:1091:3: warning: Value stored to 'ret' is never read <--[clang]
                ret = 0;
                ^     ~
1 warning generated.

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

lib/tdb/common: Fix Array access results in a null pointer dereference

Fixes;

lib/tdb/common/transaction.c:613:7: warning: Array access (via field 'blocks') results in a null pointer dereference <--[clang]
                if (tdb->transaction->blocks[i] != NULL) {
                    ^
1 warning generated.

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

lib/tdb/common: Fix warning: Null pointer passed as argument to param

Fixes:

lib/tdb/common/rescue.c:299:2: warning: Null pointer passed as an argument to a 'nonnull' parameter <--[clang]
        qsort(found.arr, found.num, sizeof(found.arr[0]), cmp_key);
        ^     ~~~~~~~~~

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

lib/util: Fix Value stored to 'ret' is never read warning

Fixes:

lib/util/tfork.c:260:3: warning: Value stored to 'ret' is never read <--[clang]

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

util: Fix signed/unsigned comparisons by casting

One case needs a variable declared, so it can be compared to -1 and
then cast to size_t for comparison.

Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Mon Jul  1 08:00:29 UTC 2019 on sn-devel-184

util: Fix signed/unsigned comparisons by declaring as size_t

I may be missing something subtle but I can't see a reason for
declaring these as ssize_t.

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

util: Fix signed/unsigned comparisons by declaring as size_t

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

tests-util: Adding test to verify "allow no conversion" flag

The internal string conversion routines smb_strtoul(l) return
an error if the provided string could not be converted to an integer.
This can be the case if the string is empty or if it starts with non-numeric
characters which cannot be converted.
The standard C library, however, does allow this and simply returns 0 as the
converted value.
If this behaviour is wanted, it can be enabled by using
the "SMB_STR_ALLOW_NO_CONVERSION" flag.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sun Jun 30 12:47:24 UTC 2019 on sn-devel-184

tests-util: Adding test to verify "full-string-conversion" flag

The standard string to integer conversion routines stop at the first
character which cannot be converted to a number.
However, if such a character is found, it is not considered an error.
With the flag "SMB_STR_FULL_STR_CONV" enabled, an error will be returned
if the string could not be converted entirely.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

tests-util: Adding test to verify "allow-negative" flag

The standard string to integer conversion routines allow strings
with a leading "-" to indicate a negative number.
However, the returned value is always an unsigned value representing
the bit-pattern of this negative value.
Typically, this behaviour is NOT wanted and therefore the standard
behavior of the internal smb_strtoul(l) return an erros in such situations.
It can be enabled though by using the flag SMB_STR_ALLOW_NEGATIVE.
This test verifies the correct processing.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

util: cleanup API change for strtoul(l) wrappers

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

source4: Update all consumers of strtoul_err(), strtoull_err() to new API

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

source3: Update all consumers of strtoul_err(), strtoull_err() to new API

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

nsswitch: Update all consumers of strtoul_err(), strtoull_err() to new API

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

libcli: Update all consumers of strtoul_err(), strtoull_err() to new API

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

lib: Update all consumers of strtoul_err(), strtoull_err() to new API

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

ctdb: Update all consumers of strtoul_err(), strtoull_err() to new API

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

lib: Add capability to enable standard glibc behaviour for string to int conversion

Adding two addtl. flags SAMBA_STR_ALLOW_NO_CONVERSION and SAMBA_STR_GLIBC_STANDARD
for the wrappers strtoul_err() and strtoull_err() providing the possibility
to get standard glibc behaviour for string to integer conversion.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

lib: Add check for full string consumption when converting string to int

Some callers want to have the entire string being used for a
string to integer conversion, otherwise flag an error.
This is possible by providing the SAMBA_STR_FULL_STR_CONV flag.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

lib: Prepare for strtoul_err(), strtoull_err() API change

In order to still be bisectable when changing the API for the wrappers
strtoul_err() and strtoull_err() some preparations need to be performed.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

lib: Add flag definitions to control the internal string to int conversion routines

The following flags are defined intially

    SMB_STR_STANDARD # raise error if negative or non-numeric
    SMB_STR_ALLOW_NEGATIVE # allow strings with a leading "-"
    SMB_STR_FULL_STR_CONV # entire string must be converted
    SMB_STR_ALLOW_NO_CONVERSION # allow empty strings or non-numeric
    SMB_STR_GLIBC_STANDARD # act exactly as the standard glibc strtoul

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

tests-util: Adding test to verify "no-conversion" detection

The standard string to integer conversion routines return zero
if a string was to be converted which did not reflect a number.
It is not flag'ed as an error.
The wrapper functions strtoul_err() and strtoull_err() are expected
to exactly do this.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

tests-util: Adding test to verify negative "number" detection

Verify that a string representing a negative number is throwing an error.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

tests-util: Adding strtoul(l)_err() test leaving errno untouched

The wrapper functions strtoul_err() and strtoull_err() trigger
other functions/routines which modify errno.
However, callers of those wrapper functions expect errno to be unchanged.
This test verifies the expectation.

Signed-off-by: Swen Schillig <swen@linux.ibm.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>

s3: smbd: Make open_directory() always open a fd.

As we never use kernel oplocks on directory handles,
there is no reason not to always open file descriptors (no
more "stat" opens on directories).

Preparing to have SMB1search use real directory
opens.

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

.gitlab-ci.yml: Increase resources for samba-ad-dc-backup job

This job seems to need more resources, possibly after
b6b5b5fe355fee2a4096e9214831cb88c7a2a4c6 and
565341baf537c9b3145a882d6a23ad4f155c0036

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Jun 28 18:48:49 UTC 2019 on sn-devel-184

vfs_preopen: TALLOC_FREE(fde) before closing the underlying fd

Without that we might get wrong stuff out of epoll

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Fri Jun 28 16:42:42 UTC 2019 on sn-devel-184

vfs_preopen: Fix an uninitialized variable read

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>

vfs_preopen: Fix for O_NOFOLLOW

Since 4301505d977449d core smbd code passes O_NOFOLLOW together with
[O_RDONLY|O_RDWR] as flags. This breaks activating vfs_preopen, we
need to look at *just* the access mode.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>

smbd: Fix broken brlock for clustering

This should have been in f11c5887f4fb4b766, sorry. We now always need
TDB_SEQNUM on brlock.tdb.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>

vfs:glusterfs_fuse: treat ENOATTR as ENOENT

The original implementation of the virtual xattr get_real_filename
in gluster was misusing the ENOENT errno as the authoritative anwer
that the file/dir that we were asking the real filename for does not
exist. But since the getxattr call is done on the parent directory,
this is a violation of the getxattr API which uses ENOENT for the
case that the file/dir that the getxattr call is done against does
not exist.

Now after a recent regression for fuse-mount re-exports due to
gluster mapping ENOENT to ESTALE in the fuse-bridge, the gluster
implementation is changed to more correctly return ENOATTR if the
requested file does not exist.

This patch changes the glusterfs_fuse vfs module to treat ENOATTR as
ENOENT to be fully functional again with latest gluster.

- Without this patch, samba against a new gluster will work correctly,
  but the get_real_filename optimization for a non-existing entry
  is lost.

- With this patch, Samba will not work correctly any more against
  very old gluster servers: Those (correctly) returned ENOATTR
  always, which Samba originally interpreted as EOPNOTSUPP, triggering
  the expensive directory scan. With this patch, ENOATTR is
  interpreted as ENOENT, the authoritative answer that the requested
  entry does not exist, which is wrong unless it really does not exist.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Fri Jun 28 12:52:03 UTC 2019 on sn-devel-184

vfs:glusterfs: treat ENOATTR as ENOENT

The original implementation of the virtual xattr get_real_filename
in gluster was misusing the ENOENT errno as the authoritative anwer
that the file/dir that we were asking the real filename for does not
exist. But since the getxattr call is done on the parent directory,
this is a violation of the getxattr API which uses ENOENT for the
case that the file/dir that the getxattr call is done against does
not exist.

Now after a recent regression for fuse-mount re-exports due to
gluster mapping ENOENT to ESTALE in the fuse-bridge, the gluster
implementation is changed to more correctly return ENOATTR if the
requested file does not exist.

This patch changes the glusterfs vfs module to treat ENOATTR as ENOENT
to be fully functional again with latest gluster.

- Without this patch, samba against a new gluster will work correctly,
  but the get_real_filename optimization for a non-existing entry
  is lost.

- With this patch, Samba will not work correctly any more against
  very old gluster servers: Those (correctly) returned ENOATTR
  always, which Samba originally interpreted as EOPNOTSUPP, triggering
  the expensive directory scan. With this patch, ENOATTR is
  interpreted as ENOENT, the authoritative answer that the requested
  entry does not exist, which is wrong unless it really does not exist.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>

s3:notifyd: Handle sigup in notifyd to reparse smb.conf

At present, SIGHUP is blocked on notifyd. So, if parent smbd
is started with "log level" 10 in smb.conf, and later changed
to 0, the SIGHUP will not change the log level to 0 in notify
smbd process and it will keep printing verbose logs in the
corresponding log files.

Proposed fix is to write a SIGHUP handler for notifyd and set
it to reload services.

Reviewed-by: Hemanth Thummala <hemanth.thummala@nutanix.com>
Signed-off-by: Shyamsunder Rathi <shyam.rathi@nutanix.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by:Volker Lendecke <vl@samba.org>

Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Jun 27 17:57:12 UTC 2019 on sn-devel-184

s4:torture: Use GnuTLS RC4 in rpc forest_trust test

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Jun 27 14:11:39 UTC 2019 on sn-devel-184

s4:tortue: Use GnuTLS RC4 in rpc lsa test

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

nsswitch: Use GnuTLS RC4 in wbclient test

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:rpc_server: Use GnuTLS RC4 in lsa server

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s3:utils: Use GnuTLS RC4 in npc_rpc_trust

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:rpc_server: Use GnuTLS RC4 in lsa endpoint

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s3:rpc_client: Use C99 inititializer in dcerpc_samr_chgpasswd_user()

This also cleans up after using them.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

auth:ntlmssp: Use GnuTLS RC4 in ntlmssp server

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

auth:gensec: Return NTSTATUS for netsec_do_seal()

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

auth:gensec: Use GnuTLS RC4 in netsec_do_seal()

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>