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.
instead make the normal composite_done() and composite_error()
functions automatically trigger a delayed callback if the caller has
had no opportunity to setup a async callback
this removes one of the common mistakes in writing a composite function
(This used to be commit f9413ce792ded682e05134b66d433eeec293e6f1)
Tridge, if you have the time, you might want to look at a problem I'm having
with unix domain stream sockets. From a comment in this commit:
/* Using composite_trigger_error here causes problems with the client
* socket. Linux 2.6.8 gives me a ECONNRESET on the next read after
* writing the reply when I don't wait the 100 milliseconds. */
This is in winbind/wb_cmd_userdomgroups.c:93.
The problem I have is that I can not *immediately* send an error reply to the
client because the next receive fails. Waiting 100 milliseconds helps. It
might also be a problem with epoll(), I don't really know.
I'd appreciate if you took a brief look at this, maybe I'm doing something
wrong.
Thanks,
Volker
(This used to be commit 3e535cce743710a68a4264e4f66e9c0c4d6770c6)
* rename the composite helper functions from comp_* to composite_*
* Move the lsa initialization to wb_connect_lsa.c
* Equip smb_composite_connect with a fallback_to_anonymous
The latter two simplify wb_init_domain.c quite a bit.
Volker
(This used to be commit deb127e04ea01ae93394da5ebffb39d81caeb6d9)
Initialize a domain structure properly. Excerpt from wb_init_domain.c:
/*
* Initialize a domain:
*
* - With schannel credentials, try to open the SMB connection with the machine
* creds. Fall back to anonymous.
*
* - If we have schannel creds, do the auth2 and open the schannel'ed netlogon
* pipe.
*
* - Open LSA. If we have machine creds, try to open with ntlmssp. Fall back
* to schannel and then to anon bind.
*
* - With queryinfopolicy, verify that we're talking to the right domain
*
* A bit complex, but with all the combinations I think it's the best we can
* get. NT4, W2k3SP1 and W2k all have different combinations, but in the end we
* have a signed&sealed lsa connection on all of them.
*
* Is this overkill? In particular the authenticated SMB connection seems a
* bit overkill, given that we do schannel for netlogon and ntlmssp for
* lsa later on w2k3, the others don't do this anyway.
*/
Thanks to Jeremy for his detective work, and to the Samba4 team for providing
such a great infrastructure.
Next step is to connect to SAM. Do it via LDAP if we can, fall back to samr
with all we have.
Volker
(This used to be commit 3e69fdc07cd76b4bc01b032148609ee4b59b8be7)
the events code replaces a destructor to one that returns allways -1
while it's calling the event handler
- we don't need the composite and winsrepl specific fixes any more
- this also fixes the problem with smbcli, dcerpc, cldap, ldap and nbt
request timeouts
metze
(This used to be commit 495996cfc49a1c6eefde6ff04fc75e0739be3aab)
stuff.
- don't use SMBCLI_REQUEST_* state's in the genreic composite stuff
- move monitor_fn to libnet.
NOTE: I have maybe found some bugs, in code that is dirrectly in DONE or ERROR
state in the _send() function. I haven't fixed this bugs in this
commit! We may need some composite_trigger_*() functions or so.
And maybe some other generic helper functions...
metze
(This used to be commit 4527815a0a9b96e460f301cb1f0c0b3964c166fc)
I still have issues with Win2k3 SP1, and Samba4 doesn't pass it's own
test for the moment, but I'm working on these issues :-)
This required a change to the credentials API, so that the special
case for NTLM logins using a principal was indeed handled as a
special, not general case.
Also don't set the realm from a ccache, as then it overrides --option=realm=.
Andrew Bartlett
(This used to be commit 194e8f07c0cb4685797c5a7a074577c62dfdebe3)
to cause an event to happen immediately. This allows metzes patch for
recognising IPs in resolve_name() to work, and also allows us to
remove some of the other code where we currently do specific checks
for is_ipaddress().
(This used to be commit 9cc000d868e1257ef6429f6f6f1f9d3c28ca330f)
Kerberos CCACHE into the system.
This again allows the use of the system ccache when no username is
specified, and brings more code in common between gensec_krb5 and
gensec_gssapi.
It also has a side-effect that may (or may not) be expected: If there
is a ccache, even if it is not used (perhaps the remote server didn't
want kerberos), it will change the default username.
Andrew Bartlett
(This used to be commit 6202267f6ec1446d6bd11d1d37d05a977bc8d315)
requirements, and for better error reporting.
In particular, the composite session setup (extended security/SPNEGO)
code now returns errors, rather than NT_STATUS_NO_MEMORY. This is
seen particularly when GENSEC fails to start.
The tighter interface rules apply to NTLMSSP, which must be called
exactly the right number of times. This is to match some of our other
less-tested modules, where adding flexablity is harder. (and this is
security code, so let's just get it right). As such, the DCE/RPC and
LDAP clients have been updated.
Andrew Bartlett
(This used to be commit 134550cf752b9edad66c3368750bfb4bbd9d55d1)
enum type has now been replaced with unsigned 32-bit field and
message data is passed as void pointer.
This allows various extension implementers to plug their monitor
messages in more easily.
rafal
(This used to be commit 4a6ab58133a59d3da3209b3e46c2a8cf848d25e7)
Session Setup code.
Add a mem_ctx argument to a few of the NTLMv2 support functions, and
add smb.conf options to control client NTLMv2 behaviour.
Andrew Bartlett
(This used to be commit 3f35cdb218a3dae08a05e77452ca9f73716ceb28)
event_context for the socket_connect() call, so that when things that
use dcerpc are running alongside anything else it doesn't block the
whole process during a connect.
Then of course I needed to change any code that created a dcerpc
connection (such as the auth code) to also take an event context, and
anything that called that and so on .... thus the size of the patch.
There were 3 places where I punted:
- abartlet wanted me to add a gensec_set_event_context() call
instead of adding it to the gensec init calls. Andrew, my
apologies for not doing this. I didn't do it as adding a new
parameter allowed me to catch all the callers with the
compiler. Now that its done, we could go back and use
gensec_set_event_context()
- the ejs code calls auth initialisation, which means it should pass
in the event context from the web server. I punted on that. Needs fixing.
- I used a NULL event context in dcom_get_pipe(). This is equivalent
to what we did already, but should be fixed to use a callers event
context. Jelmer, can you think of a clean way to do that?
I also cleaned up a couple of things:
- libnet_context_destroy() makes no sense. I removed it.
- removed some unused vars in various places
(This used to be commit 3a3025485bdb8f600ab528c0b4b4eef0c65e3fc9)
- qfsinfo (query file system information)
- appendacl (append an ACL to existing file's security descriptor and get new
full ACL)
The second one also includes an improvement to security descriptor handling
which allows to copy security descriptor. Written by Peter Novodvorsky
<peter.novodvorsky@ru.ibm.com>
Both functions have corresponding torture tests added. Tested under valgrind and
work against Samba 4 and Windows XP.
ToDo: document composite call creation process in prog_guide.txt
(This used to be commit 441cff62ac75ed16851ce7b8daf9d03eb4c3ec79)
client. The issue was actually a cut-and-paste bug, I was filling in
the .old not the .nt1 part of the union.
I've also removed the 'error checks' - I'll shortly document the API
for the credentials code to clarify that it will always return a
pointer here, except in cases of programmer error.
Tridge: I hope this is OK.
Andrew Bartlett
(This used to be commit 6439de9ec8c8d24197ea69dc337473e54c8b36b8)
GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'.
GENSEC now no longer has it's own handling of 'set username' etc,
instead it uses cli_credentials calls.
In order to link the credentails code right though Samba, a lot of
interfaces have changed to remove 'username, domain, password'
arguments, and these have been replaced with a single 'struct
cli_credentials'.
In the session setup code, a new parameter 'workgroup' contains the
client/server current workgroup, which seems unrelated to the
authentication exchange (it was being filled in from the auth info).
This allows in particular kerberos to only call back for passwords
when it actually needs to perform the kinit.
The kerberos code has been modified not to use the SPNEGO provided
'principal name' (in the mechListMIC), but to instead use the name the
host was connected to as. This better matches Microsoft behaviour,
is more secure and allows better use of standard kerberos functions.
To achieve this, I made changes to our socket code so that the
hostname (before name resolution) is now recorded on the socket.
In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now
in libcli/auth/schannel.c, and it looks much more like a standard
GENSEC module. The actual sign/seal code moved to
libcli/auth/schannel_sign.c in a previous commit.
The schannel credentails structure is now merged with the rest of the
credentails, as many of the values (username, workstation, domain)
where already present there. This makes handling this in a generic
manner much easier, as there is no longer a custom entry-point.
The auth_domain module continues to be developed, but is now just as
functional as auth_winbind. The changes here are consequential to the
schannel changes.
The only removed function at this point is the RPC-LOGIN test
(simulating the load of a WinXP login), which needs much more work to
clean it up (it contains copies of too much code from all over the
torture suite, and I havn't been able to penetrate its 'structure').
Andrew Bartlett
(This used to be commit 2301a4b38a21aa60917973451687063d83d18d66)
should be NT_STATUS_HAVE_NO_MEMORY(state->loadfile) instead of
NT_STATUS_NOT_OK_RETURN(status).
(This used to be commit eb57a587889611bcf39d75d4e15b627f36899a53)
make it possible to add optimisations to the events code such as
keeping the next timed event in a sorted list, and using epoll for
file descriptor events.
I also removed the loop events code, as it wasn't being used anywhere,
and changed timed events to always be one-shot (as adding a new timed
event in the event handler is so easy to do if needed)
(This used to be commit d7b4b6de51342a65bf46fce772d313f92f8d73d3)
rename the core structure to composite_context and the wait routine to
composite_wait() (suggestion from metze)
(This used to be commit cf11d05e35179c2c3e51c5ab370cd0a3fb15f24a)
free the connection context. This left a whole lot of state hanging
around and didn't give the memory to the caller properly
(This used to be commit 3e13e1d526563d91cb2342ae68455e54eb49a9bd)
handle the inverted memory hierarchy that a normal session
establishment gave. The inverted hierarchy came from that fact that
you first establish a socket, then a transport, then a session and
finally a tree. That leads to the socket being at the top of the
memory hierarchy and the tree at the bottom, which makes no sense from
the users point of view, as they want to be able to free the tree and
have everything disappear.
The core problem was that the libcli interface didn't distinguish
between establishing a primary context and a secondary context. If you
establish a 2nd session on a transport then you want the transport to
be referenced by the session, whereas if you establish a primary
session then you want the transport to be a child of the session.
To fix this I have added "parent_ctx" and "primary" arguments to the
libcli intialisation functions. This makes using the library much
easier, and gives us a memory hierarchy that makes much more sense.
I was prompted to do this by a bug in the cifs backend, which was
caused by the socket not being properly torn down on a disconnect due
to the inverted memory hierarchy.
(This used to be commit 5e8fd5f70178992e249805c2e1ddafaf6840739b)
