| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
| |
'release-alpha-1-0'.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
2nd info level at the container, but sam_set_userinfo2 *does*.
secondly, it _looks_ like there are two revisions for SAM database
entries, where some unknown trash is _not_ placed in the structure
(at the end: unknown_6 and padding4) if the revision number (unknown_5)
is zero.
have to keep an eye on this one over the next year or so, to make sure
all cases are caught.
third, it _looks_ like if you do a password set that the account disabled
acb_info bit is automatically unset, but i cannot confirm that right
now, i'll just have to wing-it (in the next commit)
|
|
|
|
|
|
|
|
|
|
|
| |
GROUP, it is processed *differently*, there is a domain-name string in it,
so i skip 22 bytes and _then_ read the version number etc.
this must go in 2.0.x and cvs main (oh, and i'm not going to do it, i'm
not "trusted" enough to do the job "correctly").
this will be the only notification that this is needed in 2.0.x and cvs
main branches.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Date: Fri, 17 Mar 2000 16:02:28 +0100
From: Elrond <Elrond@Wunder-Nett.org>
To: Luke Kenneth Casson Leighton <lkcl@samba.org>
Subject: Re: SVC_UNKNOWN_3
On Fri, Mar 17, 2000 at 08:27:49AM +1100, Luke Kenneth Casson Leighton wrote:
> it's probably a delete service :-) :-)
>
> check your box!!!
[...]
Hi Luke,
You pass in a Service-Control-Manager-Handle, not a
service-handle, so it can't be service-delete. It could of
course be scman-del... which would be quite crazy.
(includind the fact, that this is done by srvmgr, which
hasn't any means to add/del a service...)
But it was all my fault. I copied cli_svc.c:svc_close()
over to svc_unknown_3... and guess what... I changed
everything except the Opcode...
After that I was more successful: You pass in a scman_hnd,
and get back some other handle...I don't know yet, what it
is good for... I added an appropiate server-side
implementation and srvmgr.exe got a bit further, but now it
calls opcode 11 on a service-handle. Maybe more soon.
(I'm pressing the "Startup"-button in Services in srvmgr)
The patch contains:
- complete server and client-side implementation of
SVC_UNKNOWN_3 (yet no idea, what it is, but we have it)
- --with-privatedir=... from HEAD is now completely in TNG
- merged some things from HEAD into lib/time.c
PLEASE DON'T run indent on lib/time.c !
[lkcl: okie]
- dito for param/loadparm.c
PLEASE DON'T run indent on param/loadparm.c !
[lkcl: okie]
Elrond
|
| |
|
|
|
|
|
| |
added to *inside* smb_io_unistr2(). this is the _only_ place where
alignment rules are broken.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Date: Thu, 16 Mar 2000 16:29:53 +0100
From: Elrond <Elrond@Wunder-Nett.org>
To: Luke Kenneth Casson Leighton <lkcl@samba.org>
Subject: SVC_UNKNOWN_3
Hi Luke,
I played a bit with srvmgr around and it did a opnum=3 on
svcctl, so I tried to decode the request. For me it looked
lit it only consists of a sc_man_hnd. So I added some
parsing code and an "svcunk3" to rpcclient, so I could ask
the NT-box for the answer. The answer were 20 NUL-bytes. I
guess, this is a closed handle and a status. Also the
following svc_close(sc_man_hnd) seemed to fail somehow (I
didn't realy see the fault-pdu or so... but I did not see
an svc_io_r_close_hnd either.)
[lkcl: i suspect that this is a "delete service" call,
which is why the close fails because the handle
is alread closed!]
All this looks to me like opnum=3 is something like
SVC_CLOSE2 or so.
Could you apply the patch and verify that?
I also did the following:
- modified configure.in, so it is more like HEAD (still lot
to do)
- fixed some issues in groupdb/builtinunix.c:
When I created aliases with domain alias map, all these
aliases also existed in the Builtin domain, so usrmgr
showed them all doubled. (might be interesting for
HEAD...)
- made global_sid_S_1_1_0 static, one should use
global_sid_everyone instead.
- continued my share_get_info-effort by writing *_free_*
for all the fun. Yet have to use them.
Elrond
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
responsible for some of the password set failure problems: treating
the wrong field as an acb_info would result in rejection of password
sets.
2) in attempting to track down server-password-set problems i noticed
that the "security = domain" mode sets the password incorrectly.
|
|
|
|
| |
way round (BDC instead of PDC, PDC instead of BDC).
|
| |
|
|
|
|
| |
message from level 0 to level 1.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Date: Tue, 14 Mar 2000 19:23:04 +0100
From: Elrond <Elrond@Wunder-Nett.org>
To: Luke Kenneth Casson Leighton <lkcl@samba.org>
Subject: Re: _lsa_lookup_names
On Tue, Mar 14, 2000 at 08:20:26AM +1100, Luke Kenneth Casson Leighton wrote:
> elrond, this function isn't used, yet.
I did the conversion this morning and the diff is
appended.
[lkcl: thanks elrond]
I also added some more handle-names. Hope someone else
likes them. ;)
[lkcl: me! me! :)]
Elrond
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Date: Mon, 13 Mar 2000 18:26:00 +0100
From: Elrond <Elrond@Wunder-Nett.org>
To: Luke Kenneth Casson Leighton <lkcl@samba.org>
Subject: more cleanup
- removed most of the old --with-lib*-stuff, we're using
libtool for this
- srv_samr_passdb.c: Added some handle-names for debugging
(helped me somehow to track netlogond-probs)
- made a whole bunch of functions in display_srv.c and
display_sam.c
- Added a bunch of error-checking in lsarpcd, it looks like
it tries to check/change the trust-accoutn password as a
normal user (or even guest)
[lkcl: checks can't hurt: _lsa_open_secret is called, direct, not
client-side marshalling wrapper, lsa_open_secret, etc]
Elrond
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Date: Mon, 13 Mar 2000 18:16:51 +0100
From: Elrond <Elrond@Wunder-Nett.org>
To: Luke Kenneth Casson Leighton <lkcl@samba.org>
Subject: netlogond: crash and ggroup membership
Hi Luke,
netlogond crashes, when my workstation tries to change its
trust-account password. I tracked this to some point and
fixed those points. Since my workstation only tries to
change the pw every 15 minutes, testing is quite
timeconsuming and I didn't yet test my last fix, that I did
in this area.
After some funny errors (most notable: root is in Domain
Admins, but I couldn't change any thing in the
network-settings), I noticed, that netlogond didn't
transmit _any_ group memberships, when I logged into the
workstation, I found this and fixed it.
Could you please take a look at the following functions:
- direct_samr_userinfo
- _net_srv_pwset
I modified both, but I'm not too sure, I did it the right
way.
[lkcl: looks good, elrond. explains why groups didn't get set up!]
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Date: Sat, 11 Mar 2000 17:57:54 +0100
From: Elrond <Elrond@Wunder-Nett.org>
To: Luke Kenneth Casson Leighton <lkcl@samba.org>
Subject: more cleanup, global_sid_builtin
Hi Luke,
- more cleanup on parse_samr.c
- micro-merge from HEAD for Makefile
- use smb_io_rpc_uuid-return-value in smb_io_pol_hnd
- replaced nearly all &global_sid_S_1_5_20 by
global_sid_builtin (inspired from a post by tim, he
didn't apply this patch)
- closing an lsa-handle in rpcclient->createuser only, if
also opened
|
| |
|
|
|
|
|
| |
"invalid users = root" will stop any msrpc daemons working, including
being able to add users to the sam database (!!!! not good!)
|
| |
|
| |
|
| |
|
|
|
|
| |
on msrpc daemons.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- finaly fixed %U (at least for me), it was one line in
passdb/sampass.c
- started to merge some configure things from HEAD (not
that we currently realy need it)
- added some debugging to lsarpcd (policy-handles), so
maybe we'll get those mysterious crashes with ldap;
made some functions static
- prs_set_offset() (from J.F.) didn't prs_grow enough. I
don't know, if he needs prs_set_offset to actually grow
any buffer (for marshalling), but maybe it helps
- parse_samr.c cleanup:
+ rewrote most of the memory-things to use g_new
+ removed useless prs_align after smb_io_pol_hnd()
+ converted more memcpy(&q->pol, pol, sizeof()) to
q->pol = *pol;
this all made parse_samr.c about 100 lines smaller.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
| |
do a net use \\samba-tng\share /user:tnguser
you had to do net use \\samba-tng\share /user:samba-tngdomainname\tnguser
i solved this by putting "if domain name is unrecognised, use local SAM"
where domain name, by default on standalone workstations, is the name
of the workstation itself.
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
off sun's cascade box. HOORAY!
good suggestion, elrond.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
a different byte order from nt is VERY useful, it can be used to
identify the size of structure members because the byte order will
be swapped, in comparative traces.
- auth_resp in NETLOGON is a parameter to net_r_sam_logon, NOT part of
the NET_USER_INFO struct.
- domain_client_validate was grabbing the secret $MACHINE.ACC, then
the unicode string needed to be byte-swapped before being passwd
to nt_owf_genW.
|
|
|
|
|
|
| |
2) logon to local workstation when it is a member of a domain is also
allowed, by making domain_client_validate accept our own localservername
as the domain name (a la MYSERVERNAME\user)
|
|
|
|
| |
support but no neg_flags used, yet.
|
| |
|