s3:smb2_server: In CCM and GCM mode we can't reuse nonces

Reuse of nonces with AES-CCM and AES-GCM leads to catastrophic failure,
so make sure the server drops the connection if that ever happens.

Bug: https://bugzilla.samba.org/show_bug.cgi?id=11300

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Simo Sorce <simo@redhat.com>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Simo Sorce <idra@samba.org>
Autobuild-Date(master): Fri May 29 22:38:50 CEST 2015 on sn-devel-104

libcli/smb: In CCM and GCM mode we can't reuse nonces

Reuse of nonces with AES-CCM and AES-GCM leads to catastrophic failure,
so make sure the server drops the connection if that ever happens.

Bug: https://bugzilla.samba.org/show_bug.cgi?id=11300

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Simo Sorce <idra@samba.org>

ctdbd_conn: Fix CID 1301580 Explicit null dereferenced

Coverity does not really tell me which code path it sees where we could leave
hdr==NULL, but this check does not hurt.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ira Cooper <ira@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Fri May 29 19:49:30 CEST 2015 on sn-devel-104

KCC: test suite for the graph_utils

This found a few bugs in the tests which were fixed.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri May 29 13:55:54 CEST 2015 on sn-devel-104

KCC: rename 'repsFrom_%s_all' graphs --> 'all-repsFrom_%s' for better sort order

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

KCC: more debugging changes

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

KCC: (doc) explain intrasite max edge count a bit better

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

KCC: remove print statements from kcc_utils

debug noise should not go to stdout

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

KCC: pep8/flake8 fixes for samba_kcc

Also note a couple of unused variables. I am not removing them yet
in case their intended use turns up.

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

KCC: more pep8 for kcc_utils

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

KCC: pep8 pass over graph_utils.py

Using the `flake8` tool, which also spots e.g. unused imports.

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

KCC: add graph tests of robustness against edge and vertex failure

These tests are themselves tested in a later patch.

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

KCC: improve directed_double_ring graph check

The previous test assumed there would be only a double directed ring
but in fact there could be other edges.  In large graphs there are
certain to be more edges.

Now we want to be sure there is a complete ring apart from any other
connections. This is called the Hamiltonian path problem and takes
exponential time in general, so now our test is that it looks *quite*
a lot like a complete ring.

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

KCC: --test-all-reps-from uses same random seed for all DSAs

Otherwise some of the links end up different for each KCC run. That is
expected and proper, but it is confusing.

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

KCC: RODCs are their own bridgeheads

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

KCC: ignore non-IP transports more thoroughly

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

KCC: don't create duplicate DSA objects

load_site() returns the canonical site even if it didn't make it

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

KCC: Add more debugging and fix a comment

It seems I lost my train of thought in that comment.

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

KCC: use 75% fewer lines to assign a Boolean to a variable

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

KCC: A woeful warning comment about the state of our code

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

KCC: Debugging changes -- including DEBUG_FN() function

DEBUG_FN(msg) prefixes the msg with the function name and line no.

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

KCC: Fail earlier if there is no IP transport

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

KCC: graph the result of partial edge reversal

What it shows is we don't ever reverse an edge because we have no
partial replica in our test.

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

KCC: merge copy_output_edges into get_spanning_tree_edges

copy_output_edges() was rearranging the edges, not copying them, and
it wasn't used elsewhere.

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

KCC: move get_spanning_tree_edges out of KCC object

It doesn't use the object parameters, and might be better in another
module (e.g. graph_utils) with the other graph stuff.

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

KCC: slight rewrite for the sake of pep8

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

KCC: remove essentially dead code

As the removed comment noted, the logic goes:

    if partial:           # ~60 lines up
       if not partial:
           ...

and we have kept it there for this long because the spec implies it.

(As a matter of fact I can't see how this entire `if partial` loop does
anything of consequence, given the previous loop didn't exclude the
partial case).

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

KCC: add a warning about repsFRom magic objects

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

KCC: more pep8, using temp variables in places

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

KCC: fix square bracket padding for pep8

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

KCC: deduplicate connection schedule creation

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

KCC: reformat kcc_util object __str__ for pep8

Many lines were too long, which is to a large extent fixed by replacing
`text = text +` with `text +=`.

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

KCC: pep8 conformance

I ran the files through the pep8 command-line tool.

Most changes are for line length, inline comment formatting, adjusting
numbers of blank lines, and the indentation of conditions on if
statements.

This is pretty useless work, but I thought I would have a go with the
pep8 tool, and it came up with a lot of complaints.

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

KCC: raise KCCError instead of vanilla Exception

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

KCC: Adds some comments and rearrange translate_ntdsconn()

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

KCC: remove another needless loop variable

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

KCC: Help RW DCs to ignore RODCs when doing kcc

As far as writable DCs are concerned, RODCs don't even exist. So we make
tables that leave out the RO ones.

An RODC needs to know itself as well as writable DCs, so we add it in
that case.

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

KCC: use less verbose constructions in a few places

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

KCC: produce fewer dot graphs unless --debug is used

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

KCC: avoid (so far harmless) variable name clash

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri May 29 09:45:23 CEST 2015 on sn-devel-104

KCC: colour dot graph vertices as well as edges

It's quicker to see who is RODC.

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

KCC: use connect_table.values(), not .items() where that's best

The keys in dsa.connect_table objects are fairly arbitrary strings
(containing guids) and are of no use except when we're deleting
connections. Looping over the .items() clutters up the namespace and
potentially confuses or clobbers, given the hugeness of some of these
functions.

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

KCC: adjust update_rodc_connection to cope with multiple connections

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

KCC: use a table of bridgeheads to simplify and clarify

Looping over lists to find membership has bitten us enough times.

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

KCC: remove unused enumeration variable

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

KCC: a "should not happen" vanishing DSA is worth a warning

Not just a debug message.

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

KCC: index DSAs by guid and dn_str when loading sites

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

kcc: Change create_connection to properly query existing connections

Originally it didn't correctly detect the right 'from server' continuing
with any connection.

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

KCC: samba_kcc --test-all-reps-from tries to graph repsfrom

In future it may try to verify them too. Right now it reveals that KCC
is broken.

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

KCC: sort vertices by ndr_packed_guid instead of string guid

ndr_packed guids seems to be what windows uses.

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

KCC: changes in debugging messages

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

KCC: warn when we are ignoring non-IP transports

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

KCC: waste less time assigning transport a NULL guid

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

KCC: use a set to track kept connections

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

KCC: Comment about schedule types

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

KCC: add options to forget local and intersite links

The graphs formed using these options are likely to be different from
those you would normally get, which is OK because this is only for
testing.

Also add a helper function to draw a graph from DSA tables.

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

KCC: kcc.my_site_guid was not really used

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

KCC: use more efficient, less polluting dictionary operations

`in d` is more efficient and idiomatic than `in d.keys()`.

`for v in d.values()` is better than `for k, v in d.items()` when `k`
is not used.

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

KCC: truth in documentation

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

KCC: a whole lot of debugging messages

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

KCC: correctly update flags when repsFrom schedule is changed

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

KCC: fix is_bridgehead_failed() according to documentation

Throughout the KCC specification `detectFailedDCs` is documented along
the lines of "true to detect failed DCs", and it gets passed down to
this function. And what do we see here? It is used as a default value
when a stale link is not detected. That is entirely different. So who is
right -- the comments or the pseudo-code?

This commit follows the comments. It works!

   Documentation 1,  Pseudo-code 0

See [MS-ADTS] — v20140502, section 6.2.2.3.4.4, page 569.

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

KCC: set site topo_generator in RO case, for debug clarity

This has no actual effect except that it makes the Site object print with
the site_topo_generator that is actually used (because RODCs are always
their own topology generators and this variable doesn't affect that).

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

KCC: colour dsa dot graphs, and plot site edges

The red links in the dsa graphs are RODC topology.

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

KCC: don't try to connect to a nonexistent bridgehead

If self.get_bridgehead returns None, there are no eligible bridgeheads
over there to connect to.

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

KCC: Don't form self connections

This was a bug found in the original script.

We have found a remote bridgehead to connect to, but we were accidentally
connection to ourselves instead. Oops.

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

KCC: reuse convert_schedule_to_repltimes() for Sitelinks

This means wrenching it out of NTDSConnection, where it didn't really need to be anyway.

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

KCC: fix NTDSConnection.convert_schedule_to_repltimes

There were two bugs: the wrong nibbles were being added, and they weren't actually being saved.

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

KCC: InternalEdge should know its sitelink

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

KCC: use rep.is_default rather than non-existent part.is_default

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

KCC: debugging changes

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

KCC: add colouring for DSA link graphs

Red for RODC, blue otherwise.

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

KCC: Use fake guids in readonly mode

This is necessary because the connections need to look valid so they are
found when unnecesary connections are removed. The usual way is to query
the database for the guid.

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

KCC: beginning of --test-all-reps-from, exposing brokenness

This should produce a list of repsFrom links from all DSAs, which together
should produce a nice graph.

samba_kcc --test-all-reps-from won't actually run, but the problems are
things it found rather than caused.

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

KCC: kcc.list_dsas() returns a list rather than prints it

This will be useful for the --test-all-reps-from option.

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

KCC: intrasite doesn't need a ring with < 3 nodes

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

KCC: intrasite graph only uses local DSAs

In the specs this is implied by the special definition of 'implies'.

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

KCC: post-Kruskal graph test only needs multi-edge-forest

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

KCC: minor formatting changes

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

KCC: move console colours from kcc_utils to graph_utils

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

KCC: pep8 for ldif_utils.py

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

KCC: move ldif import/export functions into their own module

They might be of use elsewhere, and they are easily separable from the
KCC core.

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

KCC: rearrange samba_kcc entry point

For locality, so that it becomes clearer which variables are used
where.

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

KCC: move import to top of file

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

KCC: add an option to list valid DSA dn_strs

This makes testing a bit easier.

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

KCC graph_utils: rename KCCGraphError, GraphError

In case it gets used elsewhere.

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

KCC: try graphing edges in colour for partition type

It doesn't add much.

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

KCC: shift graph plotting and verification into a separate module

These might possibly be useful outside the KCC context, and the don't
rely on the rest of kcc_utils.

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

KCC: improved documentation for verify_graph_directed_double_ring

The actual function is still somewhat broken.

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

KCC: more graph verification: repsFrom

Some of these graphs are wrong headed and broken

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

KCC: the commit that mops up stray comments

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

KCC: add --attempt-live-connections to verify real life links

The spec makes these connections in a couple of places. That makes no
sense in our main testing environment which is a samba database
pretending to be a network.

The default should be changed to True when failed link detection is
implemented.

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

KCC: make --verify graph errors fatal

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

KCC: attempt to use correct verification tests for graphs

Before they were liberally assigned because the failures were as
interesting as the passes.

Also add a stub for the forest_of_rings test.

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

KCC: debug nc_type by name rather than enum number

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

KCC: sitelink graph is undirected

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

KCC: use ndr_pack for connection GUID comparisons

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

KCC: consistently use ndr_pack() for DSA GUID comparisons

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

KCC: Use ndrpack(GUID) sorting for all graph vertex operations

For intrasite rings, DSA GUIDs are sorted by their ndr_pack
representation, not their string/as-if-128-bit-int representation.

Supposing some consistency across KCC, the graph vertex guids might be
compared the same way. But we don't yet know for sure.

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

kcc: reduce brokenness of --import-lidf

It still doesn't combine well with --forced-local-dsa, due it seems to
pervasive fragile cross-dependencies within the ldb system.

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