Jeremy Allison [Mon, 24 May 2021 23:09:24 +0000 (16:09 -0700)]
s3: smbd: Get a parent pathref in create_file_unixpath().
Not yet used.
We will be passing this down to open_directory() and
open_file_ntcreate() and using it within create_file_unixpath()
as all of these functions need a parent pathref to check parent
ACLs etc.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Jeremy Allison [Mon, 7 Jun 2021 21:45:24 +0000 (14:45 -0700)]
s3: smbd: smbd_check_access_rights_fsp(), for a symlink handle just check the handle bits.
For the pathname verison of this function smbd_check_access_rights()
we return the st_mode bits turned into an NT ACL for a symlink.
For a symlink the mode bits are always 'lrwxrwxrwx' which means
smbd_check_access_rights() version always returned NT_STATUS_OK
for any access rights requested on a symlink.
For smbd_check_access_rights_fsp() to a symlink use the handle
access bits as this is a better representation of the access
allowed.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Jeremy Allison [Fri, 4 Jun 2021 00:05:51 +0000 (17:05 -0700)]
s3: smbd: Subtle change to semantics needed for smbd_check_access_rights_fsp().
smbd_check_access_rights() is checking permissions on a file *before*
open, so getting ACCESS_DENIED and mapping to NT_STATUS_OK when
reading the security descriptor is fine, as if we really don't have
access the open will fail.
smbd_check_access_rights_fsp() takes place *after* the open (pathref
or otherwise), so being unable to get the security decriptor should
be reported back to the caller and not mapped to NT_STATUS_OK.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Jeremy Allison [Fri, 4 Jun 2021 19:32:34 +0000 (12:32 -0700)]
s3: smbd: Cleanup - fix the comment for dptr_SearchDir().
It doesn't and never has skipped VETO or unreadable files.
It's only used in call_trans2findnext() in the SMB1 code
given a name already returned to the client to find a
resume position. Even if the client gave us a name that
it had never been given to it (client bug or malicious
client) we'd just start the search from the wrong position,
which doesn't cause problems (we still check for VETO
or unreadable before returning any names).
Worst thing that actually happens is the client messes
up their own search, which is what you'd expect from
giving an incorrect resume name.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Jeremy Allison [Tue, 8 Jun 2021 03:55:44 +0000 (20:55 -0700)]
s3: smbd: In open_file(), remove post-open check for opening a directory.
Now we do an early check for opening a directory, this code
could never have been triggered.
The only case we need to consider now is when the filesystem
object exists as a file when we enter open_file() and another
smbd removes and then re-creates the object as a directory before we
call reopen_from_fsp(). In that case, we will open the object,
and come back out to open_file_ntcreate(), where the race
condition detection code that calls check_same_dev_ino()
will catch the case and error out the client request.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Jeremy Allison [Tue, 8 Jun 2021 16:53:18 +0000 (18:53 +0200)]
smbd: fix pathref unlinking in create_file_unixpath()
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.
This is commonly the case from an SMB1 path-based call
(eg call_trans2qfilepathinfo()) where we use the pathref fsp
(smb_fname->fsp) as the handle. In this case we must not unlink smb_fname->fsp
from it's owner.
Pair-Programmed-With: Ralph Boehme <slow@samba.org> Signed-off-by: Jeremy Allison <jra@samba.org> Signed-off-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Tue Jun 8 20:44:41 UTC 2021 on sn-devel-184
lib:cmdline: Use getprogname() to avoid possible issues with setproctitle()
Reviewed-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Tue Jun 8 14:58:58 UTC 2021 on sn-devel-184
The NULL assert is not really required, it will crash nicely if that's
not fulfilled.
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 Jun 4 17:34:06 UTC 2021 on sn-devel-184
Volker Lendecke [Thu, 17 Dec 2020 11:36:42 +0000 (12:36 +0100)]
smbd: Simplify share_mode_entry_do()
Looking at the logic of "ha[d|ve]_share_mode_entry" d->modified=true
can only happen if we remove the last share mode entry . Make this
more explicit, avoid booleans.
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
had a bug where it carefully set unlink_flags but then didn't use them in the
following SMB_VFS_UNLINKAT() call.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Jun 3 20:23:38 UTC 2021 on sn-devel-184
Jeremy Allison [Wed, 26 May 2021 18:51:38 +0000 (11:51 -0700)]
s3: smbd: Remove parent_smb_fname(), no longer used.
Moved into the VFS as SMB_VFS_PARENT_PATHNAME() to
allow modules to process the returned parent dirname.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Wed Jun 2 06:30:36 UTC 2021 on sn-devel-184
Jeremy Allison [Wed, 26 May 2021 17:39:43 +0000 (10:39 -0700)]
s3: VFS: Add SMB_VFS_PARENT_PATHNAME().
Not yet used.
Default is NTSTATUS version of parent_smb_fname(). Now
to replace all users of parent_smb_fname() with
SMB_VFS_PARENT_PATHNAME() and then remove parent_smb_fname().
Needed due to snapdirseverywhere code in vfs_shadow_copy2.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Andrew Bartlett [Wed, 19 May 2021 01:57:21 +0000 (13:57 +1200)]
selftest: standardise and shorten winbind socket name
The full path to the winbindd socket must fit within a struct sockaddr_un and this helps us work
where this is quite deep on the server.
Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Wed Jun 2 04:46:39 UTC 2021 on sn-devel-184
Andrew Bartlett [Sat, 22 May 2021 06:40:13 +0000 (18:40 +1200)]
pidl: Avoid leaving array_size NDR tokens around
In many cases these can and should be consumed as soon as
they are used.
This is not a complete fix, we don't clean up the array_size
token after using it split between an NDR_SCALARS and
an NDR_BUFFERS pass, but it is much better than it was
and helps the winbind case with a large number of groups
(eg 100,000) as otherwise we hit the 65535 NDR token limit.
(This is an arbitary Samba-only limit to avoid DoS conditions)
Andrew Bartlett [Fri, 21 May 2021 20:46:26 +0000 (08:46 +1200)]
spoolss: Avoid indirection via ndr_get_array_size()
This is set in the call just above and otherwise we will (in the next commit)
need an intermediate variable once we need to check error codes from
ndr_get_array_size().
Douglas Bagnall [Wed, 26 May 2021 21:46:02 +0000 (09:46 +1200)]
samba-tool dns zoneoptions: timestamp manipulation options
There was a bug in Samba before 4.9 that marked all records intended
to be static with a current timestamp, and all records intended to be
dynamic with a zero timestamp. This was exactly the opposite of
correct behaviour.
It follows that a domain which has been upgraded past 4.9, but on
which aging is not enabled, records intended to be static will have a
timestamp from before the upgrade date (unless their nodes have
suffered a DNS update, which due to another bug, will change the
timestmap). The following command will make these truly static:
$ samba-tool dns zoneoptions --mark-old-records-static=2018-07-23 -U...
where '2018-07-23' should be replaced by the approximate date of the
upgrade beyond 4.9.
It seems riskier making blanket conversions of static records into
dynamic records, but there are sometimes useful patterns in the names
given to machines that we can exploit. For example, if there is a
group of machines with names like 'desktop-123' that are all supposed
to using dynamic DNS, the adminstrator can go
$ samba-tool dns zoneoptions --mark-records-dynamic-regex='desktop-\d+'
and there's a --mark-records-static-regex for symmetry.
These options are deliberately long and cumbersome to type, so people
have a chance to think before they get to the end. We also introduce a
'--dry-run' (or '-n') option so they can inspect the likely results
before going ahead.
*NOTE* ageing will still not work properly after this commit, due to
other bugs that will be fixed in other commits.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Douglas Bagnall [Fri, 28 May 2021 04:54:33 +0000 (16:54 +1200)]
pytest:samba-tool dns: more robust clean-up
If setUp() fails (and here we have a big .setUp), .tearDown is not run,
and that can leave the zone undeleted, breaking all the other tests who
expect to be able to recreate it.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Douglas Bagnall [Wed, 26 May 2021 21:45:18 +0000 (09:45 +1200)]
samba-tool dns: move dns_record_match to dnsserver.py
This function is used here and in tests, but the tests should not be
importing things from netcmd.dns, which is really supposed to be UI
code. So we move to a common place.
the only difference is the function raises DNSParseError instead of
CommandError, and netcmd.dns has to catch and wrap that.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Joseph Sutton [Thu, 6 May 2021 00:05:26 +0000 (12:05 +1200)]
pidl: Handle assigning to an inline array from Python
When obtaining a reference to items in an assigned-from list, ensure
that we do not try to use the first element of the inline array as a
talloc context, but instead use the talloc context associated with the
Python object.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri May 28 09:50:02 UTC 2021 on sn-devel-184
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
Martin Schwenke [Mon, 4 Jan 2021 02:35:11 +0000 (13:35 +1100)]
ctdb-scripts: Simplify logic in debug_via_proc_locks()
The path of the TDB is known, so calculate the file ID (device number
+ inode number) from it and use this to directly filter /proc/locks to
find processes holding locks.
Signed-off-by: Martin Schwenke <martin@meltin.net> Reviewed-by: Amitay Isaacs <amitay@gmail.com>
projects / thirdparty / samba.git / log
