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.
This only sets up the random number generator callback these days, so
use a different database for that.
(All secrets data in Samba4 is in secrets.ldb)
Andrew Bartlett
This changes the structure being used to convey the current user state
from the netlogon-derived 'netr_SamInfo3' structure to a purpose-built
structure that matches the internals of the Samba auth subsystem and
contains the final group list, as well as the final privilege set and
session key.
These previously had to be re-created on the server side of the pipe
each time.
Andrew Bartlett
Signed-off-by: Andrew Tridgell <tridge@samba.org>
This changes auth_serversupplied_info into the IDL-defined struct
auth_user_info_dc. This then in turn contains a struct
auth_user_info, which is the only part of the structure that is
mainted into the struct session_info.
The idea here is to avoid keeping the incomplete results of the
authentication (such as session keys, lists of SID memberships etc) in
a namespace where it may be confused for the finalised results.
Andrew Barltett
This makes everything reference a server_info->sids list, which is now
a struct dom_sid *, not a struct dom_sid **. This is in keeping with
the other sid lists in the security_token etc.
In the process, I also tidy up the talloc tree (move more structures
under their logical parents) and check for some possible overflows in
situations with a pathological number of sids.
Andrew Bartlett
We previously allocated sockets as direct children of the event
context. That led to crashes if a service called
task_server_terminate(), as it left the socket open and handling
events for a dead protocol.
Making them a child of the task allows the task to terminate and take
all its sockets with it.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
The new waf-based build system now has all the same functionality, and
the old build system has been broken for quite some time.
Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Sun Oct 31 02:01:44 UTC 2010 on sn-devel-104
the spoolss notify test needs to setup the process models before use
Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Sat Oct 30 13:32:09 UTC 2010 on sn-devel-104
this simplifies some of the dependencies between bin/samba and the
various services, making it easier to get a clean depenency tree
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
this also removes the event_context parameter from process model
initialisation. It isn't needed, and is confusing when a process model
init can be called from more than one place, possibly with different
event contexts.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
This includes dom_sid.h and security_token.h and will be moved
to the top level shortly.
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Tue Oct 12 03:35:36 UTC 2010 on sn-devel-104
By setting the event context to use for this operation (only) onto
the krb5_context just before we call that operation, we can try
and emulate the specification of an event context to the actual send_to_kdc()
This eliminates the specification of an event context to many other
cli_credentials calls, and the last use of event_context_find()
Special care is taken to restore the event context in the event of
nesting in the send_to_kdc function.
Andrew Bartlett
applicable.
Allow using both pyembed and pyext, to prevent unresolved symbols.
Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Sun Oct 10 03:54:01 UTC 2010 on sn-devel-104
this converts all callers that use the Samba4 loadparm lp_ calling
convention to use the lpcfg_ prefix.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
This is needed to remove samba specifc symbols from the bundled
ldb, in order to get the ABI right.
metze
Signed-off-by: Andreas Schneider <asn@samba.org>
This allows us to control what groups should be added in what use
cases, and in particular to more carefully control the introduction of
the 'authenticated' group.
In particular, in the 'service_named_pipe' protocol, we do not have
control over the addition of the authenticated users group, so we key
of 'is this user the anonymous SID'.
This also takes more care to allocate the right length ptoken->sids
Andrew Bartlett
multiple event contexts
It is NEVER valid to free an event context that anybody else may have
a reference to, and never normally valid to have two 'live' at once.
We must instead call tevent_re_initialise() to wipe clean an existing
pointer.
Andrew Bartlett
This will fork off exactly one child to handle some task, ensuring
that if it dies or changes global state, that this does not change
everything.
Andrew Bartlett
The auth context was in the past only for NTLM authentication, but we
need a SAM, an event context and and loadparm context for calculating
the local groups too, so re-use that infrustructure we already have in
place.
However, to avoid problems where we may not have an auth_context (in
torture tests, for example), allow a simpler 'session_info' to be
generated, by passing this via an indirection in gensec and an
generate_session_info() function pointer in the struct auth_context.
In the smb_server (for old-style session setups) we need to change the
async context to a new 'struct sesssetup_context'. This allows us to
use the auth_context in processing the authentication reply .
Andrew Bartlett
We need to be able to give sensible error messages when a kerberos
calls fails. This propogates the kerberos error up the stack to the
caller.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
This reverts commit 69d5cea2e5.
This commit causes issues with the RPC server, revert it until we find the
exact issue and possibly have a torture test to avoid it happening again.
Found playing with w2k8r2 and forest trusts.
After looking at the s4 side of the (s)channel :) I found out that it makes
more sense to simply make it use the tdb based code than redo the same changes
done to s3 to simplify the interface.
Ldb is slow, to the point it needs haks to pre-open the db to speed it up, yet
that does not solve the lookup speed, with ldb it is always going to be slower.
Looking through the history it is evident that the schannel database doesn't
really need greate expanadability. And lookups are always done with a single
Key. This seem a perfet fit for tdb while ldb looks unnecessarily complicated.
The schannel database is not really a persistent one. It can be discared during
an upgrade without causing any real issue. all it contains is temproary session
data.
We need to free the lp_ctx after we free the event context, otherwise
the teardown code in the notify backend dies when it tries to use the
iconv_convenience ptr
Fixes bug 7053
This task watches for changes in the list of DCs, and creates a bind9
formatted file that grants update permission to all DCs, plus to the
administration, and machines update for their own names.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
A KCC is a 'Knowledge Consistency Checker', a fancy name for a daemon
that works out who will replicate with who in a AD domain. This
implements an extremely simple KCC task that just wants to replicate
with everyone :-)
Previously we relied on process groups and SIGTERM to ensure that
child tasks died in the standard process model when the parent task
died. This doesn't work when the server is run in interactive mode, as
in that case we don't call become_daemon() and don't get a separate
process group.
The fix is to have a pipe held open by the parent server process, and
inherited by child tasks. If the parent exits then the write side of
the pipe is implicitly closed, which causes an event in the child
tasks that causes them to exit
While testing the use of the standard process model with 'make test' I
found that testing was much slower (by several times) with the
standard model than with the single model. The primary problem was
that each SMB connection would open a new sam ldb context, and all of
those would reload the full AD schema.
The fix is to pre-open the SAM during server startup, before any child
processes are forked. This sets up the global schema context which is
inherited by all connections.
The standard model is still slower at make test than the single model,
but not by nearly as much. I am working on further reducing the gap.
For now implement just a very simple protocol
to negotiate the transport session info.
We may need to pass more info later,
e.g. client ip or name...
metze
The problem was that the parent smbd doesn't have
any event when the process model standard was in use.
Now we always add an fd event for stdin, but may not
ask for any events to trigger.
metze
(This used to be commit fa6c00734c)
The problem here was that with the packet code set to serialise, we
can have multiple packets 'processing' at once, and previously the
second packet (allowed because we are spining on an event context down
the stack) would clear the flag.
Andrew Bartlett
(This used to be commit 3378911124)
specific debug function.
By default do not debug, this is the most appropriate action for a library
as we cannot assume what stderr is use for in the main app.
The main app is responsible to set ev_debug_stderr if they so desire.
(This used to be commit e566a2f308)
NULL represents an anoymous connection, as that's
cheaper than always calling auth_anonymous_session_info(),
That means if the caller wants to use the transport layer
session_info he should call auth_anonymous_session_info()
to create the session info.
metze
(This used to be commit 3777d0307b)
Fix talloc_steal in both cases to steal connected socket (from accept)
onto it's private structure, rather than stealing the bound socket.
Remove termination code from the prefork modal, we want the process to
still stay around, to serve future clients.
Andrew Bartlett and David Disseldorp
(This used to be commit 07590d893e)
To use, run 'smbd -M prefork'
By default, only the smb service is preforked. 4 children are
created, and all listen for new connections. The Linux Kernel 'wake
one' behaviour should ensure that only one is given the oportunity to
accept. We need to look into the ideal number of worker children, as
well as load balancing behaviours.
To change, set:
prefork children : smb = 6
valid service names (smb in this example) match those in 'server services'.
Andrew Bartlett and David Disseldorp
(This used to be commit 35313c0aa3)
