IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
independent socket library.
this is not used, but compiled currently
there're maybe some api changes later...
metze
(This used to be commit de4447d7a5)
but do not support SPNEGO (such as XP, when not joined to a domain).
This is triggered by the presense or lack of a security blob in the
negprot reply.
Andrew Bartlett
(This used to be commit 99f7a38c07)
NTLM sign
NTLM sign+seal
NTLM2 sign
NTLM2 sign+seal
and all of the above both with and without key exchange
the NTLM2 seal case is ugly and involves an extra data copy, which
some API changes in gensec or the ndr layer might avoid in future.
(This used to be commit fce7a4218b)
'authenticated' connections.
Fix kerberos session key issues - we need to call the
routine for extracting the session key, not just read the cache.
Andrew Bartlett
(This used to be commit b80d849b6b)
These fixes aim particularly at allowing PAC-less logins, as I don't
yet generate a PAC in the lorikeet-heimdal KDC.
This is for the benifit of a Kerbeors-enabled domain join, which seems
to be progressing quite well!
Andrew Bartlett
(This used to be commit f5a381094d)
This means that 'require NTLMv2 session security' now works for RPC
pipe signing. We don't yet have sealing, but it can't be much further.
This is almost all tridge's code, munged into a form that can work
with the GENSEC API.
This commit also includes more lsakey fixes - that key is used for all
DCE-RPC level authenticated connections, even over CIFS/ncacn_np.
No doubt I missed something, but I'm going to get some sleep :-)
Andrew Bartlett
(This used to be commit a1fe175eec)
I had previously thought this was unnecessary, as windows doesn't use
standards compliant UTF-16, and for filesystem operations treats bytes
as UCS-2, but Bjoern Jacke has pointed out to me that this means we
don't correctly store extended UTF-16 characters as UTF-8 on
disk. This can be seen with (for example) the gothic characters with
codepoints above 64k.
This commit also adds a LOCAL-ICONV torture test that tests the first
1 million codepoints against the system iconv library, and tests 5
million random UTF-16LE buffers for identical error handling to the
system iconv library.
the lib/iconv.c changes need backporting to samba3
(This used to be commit 756f28ac95)
Argl. I could never get the naming right. Having the most significant byte at
the lowest memory address is big endian, at least according to the google
search for 'big endian'....
Volker
(This used to be commit bc4c188362)
The bug (found by tridge) is that Win2k3 is being tighter about the
NTLMSSP flags. If we don't negotiate sealing, we can't use it.
We now have a way to indicate to the GENSEC implementation mechanisms
what things we want for a connection.
Andrew Bartlett
(This used to be commit 86f61568ea)
now that talloc_free() doesn't need to take a context ptr, there is no
reason we can't use talloc everywhere that we currently use malloc().
(This used to be commit a2ad77fb3a)
an SPNEGO login from WinXP at least).
talloc_asprintf_append() lost an argument, but because TALLOC_CTX is
now a void*, this was not picked up by the compiler.
I've tested the login (asn1), but not the registry/gtk changes.
Andrew Bartlett
(This used to be commit 4294be4405)
It simplifies our structure handling a lot, making the code shorter
and easier to understand. Look at the diff carefully and see if you
can understand it. If you're still confused then please ask.
(This used to be commit 03c341aca7)
This version does the following:
1) talloc_free(), talloc_realloc() and talloc_steal() lose their
(redundent) first arguments
2) you can use _any_ talloc pointer as a talloc context to allocate
more memory. This allows you to create complex data structures
where the top level structure is the logical parent of the next
level down, and those are the parents of the level below
that. Then destroy either the lot with a single talloc_free() or
destroy any sub-part with a talloc_free() of that part
3) you can name any pointer. Use talloc_named() which is just like
talloc() but takes the printf style name argument as well as the
parent context and the size.
The whole thing ends up being a very simple piece of code, although
some of the pointer walking gets hairy.
So far, I'm just using the new talloc() like the old one. The next
step is to actually take advantage of the new interface
properly. Expect some new commits soon that simplify some common
coding styles in samba4 by using the new talloc().
(This used to be commit e35bb094c5)
- print the received security_descriptor in the smbclient "acl" command
- make sure we zero the alignment data in nttrans packet sends
(This used to be commit 8925b8b219)
handle connections using the IP as the server name, while not trying
for NBT name resolution on names like "192" and "192.168.1.2".
also removed the ip address argument to smbcli_socket_connect() as it
isn't used and doesn't really make sense.
(This used to be commit 2ce4028842)
ALL_INFO queryfileinfo level. It is useful having it here as many
non-Microsoft servers implement it this way, which breaks just about
all the torture tests, so when testing against these broken systems
just change this one #if line and recompile smbtorture.
(This used to be commit cd8887293e)
there're some cleanups needed and we need to verify the PAC correctly
and create the auth_session_info correctly...
metze
(This used to be commit d8fe497097)
all ldb functions has been renamed to ldap_ as we don't
really want to include ldb functions here, let's keep ldap
and ldb separate.
(This used to be commit f9d7b731c9)
- use lp_maxprotocol() in the libcli/raw/ negotiate code, so we obey
the smb.conf "max protocol" option
- better handling of -M option in masktest
(This used to be commit 8685a584c9)
- make cli_list_new() use the last_name continue method rather than
the trans2 findnext continue flag, as the continue flag is broken on
win2003 (win2003 sometimes misses up to 1/3 of all files in a directory)
(This used to be commit daa9648b3f)
allow tests for 'unwrapped' krb5, allowed by Win2k3.
SPENGO changes, trying to get the logic right (when and what
sub-mechanisms to wrap).
Andrew Bartlett
(This used to be commit 8a0f7bf5e2)
Up to now the client code has had an async API, and operated
asynchronously at the packet level, but was not truly async in that it
assumed that it could always write to the socket and when a partial
packet came in that it could block waiting for the rest of the packet.
This change makes the SMB client library full async, by adding a
separate outgoing packet queue, using non-blocking socket IO and
having a input buffer that can fill asynchonously until the full
packet has arrived.
The main complexity was in dealing with the events structure when
using the CIFS proxy backend. In that case the same events structure
needs to be used in both the client library and the main smbd server,
so that when the client library is waiting for a reply that the main
server keeps processing packets. This required some changes in the
events library code.
Next step is to make the generated rpc client code use these new
capabilities.
(This used to be commit 96bf4da3ed)
e.g. we now have 'union smb_mkdir' and 'enum smb_mkdir_level' in sync
we may should also rename 'RAW_MKDIR_*' -> 'SMB_MKDIR_*'
metze
(This used to be commit 0bb50dcf1c)
- This causes our client and server code to use the same core code,
with the same debugs etc.
- In turn, this will allow the 'mandetory/fallback' signing algorithms
to be shared, and only written once.
Updates to the SPNEGO code
- Don't wrap an empty token to the server, if we are actually already finished.
Andrew Bartlett
(This used to be commit 35b83eb329)
Rework our random number generation system.
On systems with /dev/urandom, this avoids a change to secrets.tdb for every fork().
For other systems, we now only re-seed after a fork, and on startup.
No need to do it per-operation. This removes the 'need_reseed'
parameter from generate_random_buffer().
This also requires that we start the secrets subsystem, as that is
where the reseed value is stored, for systems without /dev/urandom.
In order to aviod identical streams in forked children, the random
state is re-initialised after the fork(), at the same point were we do
that to the tdbs.
Andrew Bartlett
(This used to be commit b97d3cb2ef)
- We can now connect to hosts that follow the SPNEGO RFC, and *do not*
give us their principal name in the mechListMIC.
- The client code now remembers the hostname it connects to
- We now kinit for a user, if there is not valid ticket already
- Re-introduce clock skew compensation
TODO:
- See if the username in the ccache matches the username specified
- Use a private ccache, rather then the global one, for a 'new' kinit
- Determine 'default' usernames.
- The default for Krb5 is the one in the ccache, then $USER
- For NTLMSSP, it's just $USER
Andrew Bartlett
(This used to be commit de5da66939)
- Spelling - it's SPNEGO, not SPENGO
- SMB signing - Krb5 logins are now correctly signed
- SPNEGO - Changes to always tell GENSEC about incoming packets, empty or not.
Andrew Bartlett
(This used to be commit cea578d6f3)
The session key in the client is wrong, we don't do signing/sealing
and we are sending raw Kerberos, not GSSAPI.
But it's a start, and if we continue to have to call Krb5 directly,
this will be the basis.
I also intend to provide an alternate implementation, using just
GSSAPI.
Andrew Bartlett
(This used to be commit eb0dd4a821)
switch, rather than a series of if statements.
Also start to use the GENSEC subcontexts, and add some comments
explaining some of the 'odd' logic in parts.
I'll probably break these out into subfunctions soon.
Thanks to metze for getting me to do this :-)
Andrew Bartlett
(This used to be commit 73e03596d3)
- Add the concept of a 'subcontext' into gensec, so that the spengo
code doesn't have to figure out how to make one.
(A subcontext inherits the username, domain, password (or callback)
from the main context).
- Add comments to some other routines, and explain a bit about what
the various 'start' functions are for.
Andrew Bartlett
(This used to be commit 7aedbfbdd9)
w2k3 can handle in a single request. With the samba3 client rpc libs I can do
about 21000 SIDs in a single request. test_many_LookupSIDs with 10000 SIDs
fails on the subsequent request with a NET_WRITE_FAULT. Maybe the Samba4 DCE
people want to take a look at this -- I don't see the problem.
Bug fix: SID components should be treated as unsigned when parsing
Volker
(This used to be commit 8c997a2ad2)
This implements gensec for Samba's server side, and brings gensec up
to the standards of a full subsystem.
This means that use of the subsystem is by gensec_* functions, not
function pointers in structures (this is internal). This causes
changes in all the existing gensec users.
Our RPC server no longer contains it's own generalised security
scheme, and now calls gensec directly.
Gensec has also taken over the role of auth/auth_ntlmssp.c
An important part of gensec, is the output of the 'session_info'
struct. This is now reference counted, so that we can correctly free
it when a pipe is closed, no matter if it was inherited, or created by
per-pipe authentication.
The schannel code is reworked, to be in the same file for client and
server.
ntlm_auth is reworked to use gensec.
The major problem with this code is the way it relies on subsystem
auto-initialisation. The primary reason for this commit now.is to
allow these problems to be looked at, and fixed.
There are problems with the new code:
- I've tested it with smbtorture, but currently don't have VMware and
valgrind working (this I'll fix soon).
- The SPNEGO code is client-only at this point.
- We still do not do kerberos.
Andrew Bartlett
(This used to be commit 07fd885fd4)
This layer is used for DCERPC security, as well as ntlm_auth at this
time. It expect things like SASL and the CIFS layer to use it as
well.
The particular purpose of this layer is to introduce SPENGO, which
needs generic access to the actual implementation mechanisms.
Schannel, due to it's 'interesting' setup properties is in GENSEC, but
is only in the RPC code.
Andrew Bartlett
(This used to be commit 902af49006)
I have moved the SPNEGO and Kerberos code into libcli/auth, and intend
to refactor them into the same format as NTLMSSP.
Andrew Bartlett
(This used to be commit 58da78a746)
- implement key weakening
- don't create large 'hashes' when we only want a key (signing subkeys)
- make more useful debugs.
NTLM2 is still off by default, till I figure out how to do NTLM2 signing.
Andrew Bartlett
(This used to be commit 079c265485)
reply also initialise the LM session key, when we have it (was failing
because the auth code was setting it's length wrong).
Andrew Bartlett
(This used to be commit de97d9df22)
Changes:
- Check for a valid 'pipe_state' in netr_ServerAuthenticate3 before
we dereference it
- removes the expansionroom[7] in the netr_SamInfo* structs to 7
individual elements.
- renames netr_SamInfo -> netr_SamInfo2
netr_SamInfo2 -> netr_SamInfo3
- Having the thing we always called an 'info3' being 'netr_SamInfo2'
was just too confusing.
- Expand and fill in extra details about users from the SAM, into
the server_info, for processing into the SamLogon reply.
- Add a dum_sid_dup() function to duplicate a struct dom_sid
The SamLogon code currently does not return supplementary groups, and is
only tested with Samba4 smbtorture.
Andrew Bartlett
(This used to be commit 6c92563b79)
multiple torture tests to temporarily join a domain
- fixed a session key size problem
- added a schannel test suite
- allow schannel to work with ncacn_ip_tcp
(This used to be commit 36f05e4d57)
names rather than our crazy naming scheme. So DES is now called
des_crypt() rather than smbhash()
- added the code from the solution of the ADS crypto challenge that
allows Samba to correctly handle a 128 bit session key in all of the
netr_ServerAuthenticateX() varients. A huge thanks to Luke Howard
from PADL for solving this one!
- restructured the server side rpc authentication to allow for other
than NTLMSSP sign and seal. This commit just adds the structure, the
next commit will add schannel server side support.
- added 128 bit session key support to our client side code, and
testing against w2k3 with smbtorture. Works well.
(This used to be commit 729b2f41c9)
WinXP sees us as an ADS server.
Unfortunately WinXP also uses a set of negotiate_flags that we don't
support yet. Some crypto work needed.
(This used to be commit 2d740b6570)
- added start of QueryDomainInfo in samr server
"net rpc info" from samba3 now works against a samba4 server. I
suspect join will work fairly soon.
(This used to be commit 0a2c6a1062)
Samba's NTLMSSP code is now fully talloc based, which should go a long
way to cleaning up the memory leaks in this code. This also avoids a
lot of extra copies of data, as we now allocate the 'return' blobs on
a caller-supplied context.
I have also been doing a lot of work towards NTLM2 signing and
sealing. I have this working for sealing, but not for the verifier
(MD5 integrity check on the stream) which is still incorrect.
(I can aim a rpcecho sinkdata from a Win2k3 box to my server, and the
data arrives intact, but the signature check fails. It does however
match the test values I have...).
The new torture test is cludged in - when we get a unit test suite
back, I'll happliy put it in the 'right' place....
Andrew Bartlett
(This used to be commit 399e2e2b11)
structures. This was suggested by metze recently.
I checked on the build farm and all the machines we have support 64
bit ints, and support the LL suffix for 64 bit constants. I suspect
some won't support strtoll() and related functions, so we will
probably need replacements for those.
(This used to be commit 9a9244a1c6)
This involves allowing the password set code in samdb to take an
already hashed password, and some fixes to our torture code.
Andrew Bartlett
(This used to be commit f9f581b580)
Currently this only authentiates the machine, not real users.
As a consequence of running the Samba4 NETLOGON test against Samba4, I
found a number of issues in the SAMR server, which I have addressed.
There are more templates in the provison.ldif for this reason.
I also added some debug to our credentials code, and fixed some bugs
in the auth_sam module.
The static buffer in generate_random_string() bit me badly, so I
removed it in favor of a talloc based system.
Andrew Bartlett
(This used to be commit 94624e519b)
Now that all session keys are DATA_BLOBs, fix the callers.
This assumes some things about the behaviour of certain crypto
algorithms, without the ability to test it on session keys != 16 bytes
in length. We will just need to retest when we get the KRB5 support
in (DES keys are 8 bytes).
Andrew Bartlett
(This used to be commit e4355a7ec1)
- This required using NETLOGON_NEG_AUTH2_FLAGS for the
SetupCredentials2 negotiation flags, which is what Samba3 does,
because otherwise the server uses different crypto.
- This tests the returned session keys, which we decrypt.
- Update the Samba4 notion of a 'session key' to be a DATA_BLOB in
most places.
- Fix session key code to return NT_STATUS_NO_SESSION_KEY if none is
available.
- Remove a useless argument to SMBsesskeygen_ntv1
- move netr_CredentialState from the .idl to the new credentials.h
Andrew Bartlett
(This used to be commit 44f8b5b53e)
The work here is trying to get the LM_KEY option for NLTMSSP
operating, however until that functions properly, it is now controlled
by some new smb.conf options, defaulting off.
Andrew Bartlett
(This used to be commit c63eb35b45)
Not all the auth code is merged - only those parts that are actually
being used in Samba4.
There is a lot more work to do in the NTLMSSP area, and I hope to
develop that work here. There is a start on this here - splitting
NTLMSSP into two parts that my operate in an async fashion (before and
after the actual authentication)
Andrew Bartlett
(This used to be commit 5876c78806)
request (a dead socket). I discovered this when testing against Sun's
PC-NetLink.
cleaned up the naming of some of the samr requests
add IDL and test code for samr_QueryGroupMember(),
samr_SetMemberAttributesOfGroup() and samr_Shutdown(). (actually, I
didn't leave the samr_Shutdown() test in, as its fatal to windows
servers due to doing exactly what it says it does).
(This used to be commit 925bc2622c)
- completed the IDL and test code for the various set user password
mechanisms in samr. Three password mechanisms are now working, the
UserInfo24 method, the OemChangePasswordUser2() method (which only
sets the LM password) and the ChangePasswordUser2() method which sets
both the LM and NT passwords.
- updated some crypto routines to support the password change tests
(This used to be commit 051efa2abf)
samr_AddMultipleMembersToAlias(),
samr_RemoveMultipleMembersFromAlias(), samr_OemChangePasswordUser2(),
and samr_ChangePasswordUser2()
The password change functions don't actually work yet (but should
soon). At this stage I have just completed the IDL for them. Next step
is to get the hash verifiers right and the torture test should be able
to do password changes.
(This used to be commit 849d0d314a)
rare thing, a non-length string (ie. not a WIRE_STRING) but a null
terminated char string. There wasn't a good interface to pull that
out of a blob (all the string interfaces assumed WIRE_STRINGS). Added
a new one, only used for this call. Sucks, I know - but the alternatives
suck more. Added tests for some of the unix info returned.
Jeremy.
(This used to be commit 4d0ed04c54)
added ldbedit, a _really_ useful command
added ldbadd, ldbdel, ldbsearch and ldbmodify to build
solved lots of timezone issues, we now pass the torture tests with
client and server in different zones
fixed several build issues
I know this breaks the no-LDAP build. Wait till I arrive in San Jose for that
fix.
(This used to be commit af34710d4d)
- always use the 14 word writex varient even for small transfers as
long as large offsets are negotiated (this matches windows
behaviour)
- make sure we fill in the top 16 bits of the count for large writex
calls
(This used to be commit 9ea20d0c9a)
and schannel are both instances of possible security modules
- added schannel sign and sign/seal support to the dcerpc client
code. You select it with binding options of "schannel,sign" or
"schannel,seal".
(This used to be commit 05db0b9d94)
smbtorture to use the new interface.
Part 2 will be to eliminate cli_state from smbtorture as this is now
the only place where it is used.
(This used to be commit db1cc96af6)
This adds support for bigendian rpc in the client. I have installed
SUN pcnetlink locally and am using it to test the samba4 rpc
code. This allows us to easily find places where we have stuffed up
the types (such as 2 uint16 versus a uint32), as testing both
big-endian and little-endian easily shows which is correct. I have now
used this to fix several bugs like that in the samba4 IDL.
In order to make this work I also had to redefine a GUID as a true
structure, not a blob. From the pcnetlink wire it is clear that it is
indeed defined as a structure (the byte order changes). This required
changing lots of Samba code to use a GUID as a structure.
I also had to fix the if_version code in dcerpc syntax IDs, as it
turns out they are a single uint32 not two uint16s.
The big-endian support is a bit ugly at the moment, and breaks the
layering in some places. More work is needed, especially on the server
side.
(This used to be commit bb1af644a5)
yet, but at least the request is understood by w2k3
Also modified pidl to allow multiple branches in a union to have the
same element. This is used in netlogon.
(This used to be commit 983c0e9683)
machine account password.
* neater handling on value() options in IDL. The auto-print code
will now display the right value so you don't need to initialise
it in your C code
(This used to be commit 3dd978b12b)
I would like the netlogon test suite to eventually do a new domain
join using a fake workstation name, then remove itself afterwards, but
for now I'm assuming we are already joined to the domain when the
testsuite runs. This means you need to use the Samba3 net command to
do a join before running RPC-NETLOGON
(This used to be commit 8c7a9446a0)
* moved ntlmssp code into libcli/auth/, and updated to latest ntlmssp
code from samba3 (thanks Andrew! the new interface is great)
* added signing/ntlmssp support in the dcerpc code
* added a dcerpc_auth.c module for the various dcerpc auth mechanisms
(This used to be commit c18c9b5585)
files around, but also added the first bits of auto-generated code for
the lsa pipe.
I haven't updated the Makefile to call pidl yet, so for now the code
was cut-and-pasted into librpc/ndr/ndr_lsa.c manually
(This used to be commit 6b222d3b65)
- added SMBD_LISTEN_BACKLOG in local.h
- added the beginnings of a ndr/rpc parsing framework for Samba4. It
currently correctly parses security descriptors for the nttrans
QUERY_SECDESC call, but I hope it will become a reasonable framework
that an idl based generator can work to
(This used to be commit 9bf904fc34)
cope with servers that return bogus (too large) values in max_xmit
cope with a couple more error conditions in RAW-SFILEINFO
better startup time heuristics in NBENCH
(This used to be commit 89f7261ba5)
fields in level 261 and level 262 of directory search are, plus the
names of the levels
the unknown fields are a 64bit unique file id, and match the 64 bit
number from the internal_information qfileinfo level
(This used to be commit b69f54eb02)
