samba.git
26 years agoThis is it ! The mega-merge of the JRA_NMBD_REWRITE branch
Jeremy Allison [Sat, 13 Dec 1997 14:16:07 +0000 (14:16 +0000)]
This is it ! The mega-merge of the JRA_NMBD_REWRITE branch
back into the main tree.
For the cvs logs of all the files starting nmbd_*.c, look
in the JRA_NMBD_REWRITE branch. That branch has now been
discontinued.
Jeremy.

26 years agoFixed bug reported by Gert-Jan Vons <Gert-Jan.Vons@ocegr.fr>
Jeremy Allison [Fri, 12 Dec 1997 09:10:01 +0000 (09:10 +0000)]
Fixed bug reported by Gert-Jan Vons <Gert-Jan.Vons@ocegr.fr>
with doing a dir /s into a unix directory ending in a ':'.
Jeremy.

26 years agofixed over char 127 problems with isupper/islower.
Jeremy Allison [Thu, 11 Dec 1997 12:25:01 +0000 (12:25 +0000)]
fixed over char 127 problems with isupper/islower.
Jeremy.

26 years agoWhile working on a general-purpose caching module (out soon), I thought of
Christopher R. Hertel [Thu, 11 Dec 1997 11:44:18 +0000 (11:44 +0000)]
While working on a general-purpose caching module (out soon), I thought of
a better way to handle the node pointer array used in ubi_BinTree.  The
change simplified the code a bigbunch.  It also forced updates to all of
the binary tree modules.  CRH

26 years agokanji.c: Patch from Anders Blomdell <anders.blomdell@control.lth.se>
Jeremy Allison [Thu, 11 Dec 1997 11:07:09 +0000 (11:07 +0000)]
kanji.c: Patch from Anders Blomdell <anders.blomdell@control.lth.se>
for problem with kanji sj_strXXX functions.
Jeremy.

26 years agofix typo in directory name
Herb Lewis [Thu, 4 Dec 1997 19:17:02 +0000 (19:17 +0000)]
fix typo in directory name

26 years agoAdded one more USA consultant.
Jeremy Allison [Thu, 4 Dec 1997 18:39:08 +0000 (18:39 +0000)]
Added one more USA consultant.
Jeremy.

26 years agocatch signals in the async dns daemon and allow it to auto-restart if
Andrew Tridgell [Thu, 4 Dec 1997 10:58:40 +0000 (10:58 +0000)]
catch signals in the async dns daemon and allow it to auto-restart if
necessary

26 years agoallow for zero size reads in asyncdns. These can happen after a signal
Andrew Tridgell [Thu, 4 Dec 1997 08:10:49 +0000 (08:10 +0000)]
allow for zero size reads in asyncdns. These can happen after a signal

26 years agogot rid of the WRAP_MALLOC code - mem_man does it better
Andrew Tridgell [Thu, 4 Dec 1997 07:49:37 +0000 (07:49 +0000)]
got rid of the WRAP_MALLOC code - mem_man does it better

26 years agoadded optonal MEM_MAN code
Andrew Tridgell [Thu, 4 Dec 1997 07:48:06 +0000 (07:48 +0000)]
added optonal MEM_MAN code

26 years agodon't use free and alloc as structure elements
Andrew Tridgell [Thu, 4 Dec 1997 07:41:48 +0000 (07:41 +0000)]
don't use free and alloc as structure elements

26 years agoMakefile.lib does nothing, so remove it
Andrew Tridgell [Wed, 3 Dec 1997 14:17:38 +0000 (14:17 +0000)]
Makefile.lib does nothing, so remove it

26 years agoslight cleanup to the linked list handling
Andrew Tridgell [Wed, 3 Dec 1997 05:54:51 +0000 (05:54 +0000)]
slight cleanup to the linked list handling

26 years agoput the default PRINTCAP_NAME def in includes.h not local.h so that
Andrew Tridgell [Wed, 3 Dec 1997 05:22:24 +0000 (05:22 +0000)]
put the default PRINTCAP_NAME def in includes.h not local.h so that
it can see the SYSV definitions

26 years agoapplied a patch from Norm Jacobs to allow "printcap name = lpstat"
Andrew Tridgell [Wed, 3 Dec 1997 05:08:07 +0000 (05:08 +0000)]
applied a patch from Norm Jacobs to allow "printcap name = lpstat"
to use lpstat to obtain the printer list on systemV systems.

I've now made this the default on all SYSV systems.

Jeremy, you were a little worried about the security of this patch. I
believe it's OK as the user has no control over the options given to
popen() and the pipe is only open for reading.

26 years agochange the "username map" option to allow the user to stop the
Andrew Tridgell [Wed, 3 Dec 1997 04:20:39 +0000 (04:20 +0000)]
change the "username map" option to allow the user to stop the
processing part way through the file if a match is found. If a line
starts with ! and a match is made by that line then processing
stops. This allows better wildcard handling.

(patch from Anselm.Kruis@Physik.Uni-Muenchen.DE)

26 years agoallow local_machine and remote_machine (%L and %m macros) to contain
Andrew Tridgell [Wed, 3 Dec 1997 03:57:29 +0000 (03:57 +0000)]
allow local_machine and remote_machine (%L and %m macros) to contain
spaces

26 years agoloadparm.c: Added lp_we_are_a_wins_server as an alias for lp_wins_support.
Jeremy Allison [Wed, 3 Dec 1997 03:50:05 +0000 (03:50 +0000)]
loadparm.c: Added lp_we_are_a_wins_server as an alias for lp_wins_support.
Used by the new nmbd and just makes more sense :-).
Jeremy.

26 years agoadd a warning if the timezone is not a multiple of 1 minute. This
Andrew Tridgell [Wed, 3 Dec 1997 03:46:43 +0000 (03:46 +0000)]
add a warning if the timezone is not a multiple of 1 minute. This
should catch broken timezone files in slackware linux.

26 years agoallow users to disable the NetWkstaUserLogon call in server level
Andrew Tridgell [Wed, 3 Dec 1997 03:43:05 +0000 (03:43 +0000)]
allow users to disable the NetWkstaUserLogon call in server level
security by changing a setting in local.h or adding it to their
Makefile. See comment in local.h

26 years agomake the "printing" option a per share option rather than global. When
Andrew Tridgell [Wed, 3 Dec 1997 03:37:02 +0000 (03:37 +0000)]
make the "printing" option a per share option rather than global. When
printing to lots of different sorts of remote printers this is useful

26 years agoI'm slowly getting though the todo list :-)
Andrew Tridgell [Wed, 3 Dec 1997 03:21:25 +0000 (03:21 +0000)]
I'm slowly getting though the todo list :-)

This change allows people to select the SYSLOG_FACILITY in local.h, or
add it to the Makefile if they want to.

26 years agohopefully handle "ready and waiting" messages in print queue output a
Andrew Tridgell [Wed, 3 Dec 1997 03:00:49 +0000 (03:00 +0000)]
hopefully handle "ready and waiting" messages in print queue output a
little better

26 years agofix toupper(c) on a already uppercase char and tolower(c) on an
Andrew Tridgell [Wed, 3 Dec 1997 02:27:50 +0000 (02:27 +0000)]
fix toupper(c) on a already uppercase char and tolower(c) on an
already lowercase char when using code pages. Jeremy, can you confirm
that this looks right?  It comes from PR#1992

26 years agofixed the help message for -p (someone complained!)
Andrew Tridgell [Tue, 2 Dec 1997 23:32:03 +0000 (23:32 +0000)]
fixed the help message for -p (someone complained!)

26 years agoadd the null string to SMBsetatr calls
Andrew Tridgell [Tue, 2 Dec 1997 23:30:43 +0000 (23:30 +0000)]
add the null string to SMBsetatr calls

26 years agochanging the comment in find_new_file() to say why a base of 1 is used
Andrew Tridgell [Tue, 2 Dec 1997 23:29:57 +0000 (23:29 +0000)]
changing the comment in find_new_file() to say why a base of 1 is used

26 years agothe default GUEST_ACCOUNT should be here not in smb.h
Andrew Tridgell [Tue, 2 Dec 1997 23:28:14 +0000 (23:28 +0000)]
the default GUEST_ACCOUNT should be here not in smb.h

26 years agoHPUX trusted systems need to use bigcrypt() not crypt()
Andrew Tridgell [Tue, 2 Dec 1997 23:27:40 +0000 (23:27 +0000)]
HPUX trusted systems need to use bigcrypt() not crypt()

26 years agoget rid of some things out of smb.h that are already in local.h
Andrew Tridgell [Tue, 2 Dec 1997 23:26:42 +0000 (23:26 +0000)]
get rid of some things out of smb.h that are already in local.h

26 years agodont try getpeername() when Client isn't initialised
Andrew Tridgell [Tue, 2 Dec 1997 23:24:26 +0000 (23:24 +0000)]
dont try getpeername() when Client isn't initialised

26 years agoDocumented the 'veto oplock files' parameter.
Jeremy Allison [Tue, 2 Dec 1997 19:55:44 +0000 (19:55 +0000)]
Documented the 'veto oplock files' parameter.
Jeremy.

26 years agoasyncdns.c: Removed warning when compiling with -DSYNC_DNS.
Jeremy Allison [Tue, 2 Dec 1997 19:00:18 +0000 (19:00 +0000)]
asyncdns.c: Removed warning when compiling with -DSYNC_DNS.
nameelect.c: Tidied up settings of work->ServerType when unbecoming things.
nmbd.c: Fixed pidFile warning.
server.c: Fixed pidFile warning.
Jeremy.

26 years agofix for systems that don't have PWD environment variable defined
Herb Lewis [Mon, 1 Dec 1997 22:49:09 +0000 (22:49 +0000)]
fix for systems that don't have PWD environment variable defined

26 years agoremove stutter from closing instructions
Herb Lewis [Mon, 1 Dec 1997 15:21:23 +0000 (15:21 +0000)]
remove stutter from closing instructions

26 years agodded error checking to scripts
Herb Lewis [Mon, 1 Dec 1997 14:50:52 +0000 (14:50 +0000)]
dded error checking to scripts

26 years agothe logon script example should use %U.bat not %u.bat
Andrew Tridgell [Sun, 30 Nov 1997 05:46:33 +0000 (05:46 +0000)]
the logon script example should use %U.bat not %u.bat

26 years agoclientgen.c: Added cli_mv() (used in a recent torture test).
Jeremy Allison [Sun, 30 Nov 1997 02:58:34 +0000 (02:58 +0000)]
clientgen.c: Added cli_mv() (used in a recent torture test).
reply.c: Changed reply_open_and_X to split out the oplock
request bits from core and extended and if an oplock was granted only set
the corresponding bit on reply.
server.c: Added code to dynamically allocate i/o buffers in oplock_break
(prevents recursion problems) , also made reset of sent_oplock_break
explicit.
Jeremy.

26 years agodon't display locks for dead processes in smbstatus
Andrew Tridgell [Sat, 29 Nov 1997 13:35:09 +0000 (13:35 +0000)]
don't display locks for dead processes in smbstatus

26 years agouse -1 not 0xffffffff in SIVALS() macros
Andrew Tridgell [Sat, 29 Nov 1997 13:29:13 +0000 (13:29 +0000)]
use -1 not 0xffffffff in SIVALS() macros

use the same process_exists() code on all systems (it's probably
faster anyway)

26 years agoadded a sent_oplock_break element to Files[] as a paranoia check so we
Andrew Tridgell [Sat, 29 Nov 1997 02:40:31 +0000 (02:40 +0000)]
added a sent_oplock_break element to Files[] as a paranoia check so we
can't sent a oplock break twice on the same file.

changed some debug levels in the oplock code to level 0 so we can
track down a bug

zero the returned Files[] entry in find_free_file()

don't try to overcome client bugs in the handling of non-encrypted
passwords if in server level security mode

added paranoid null termination of password buffers

slight change to my ajt_panic() routine

26 years agopass the mailslot name to process_logon_packet()
Andrew Tridgell [Fri, 28 Nov 1997 05:52:31 +0000 (05:52 +0000)]
pass the mailslot name to process_logon_packet()

- fixed process_logon_packet() to always send the reply to/from the
correct netbios name (it should match the query)

- only do a NT style QUERYFORPDC if the query came in on a
NT_LOGON_MAILSLOT

Luke, can you check to see if these changes work in your branch with
NT logons?

26 years agoget_entries is an int not a bool
Andrew Tridgell [Fri, 28 Nov 1997 05:48:48 +0000 (05:48 +0000)]
get_entries is an int not a bool

26 years agopreparing for release of 1.9.18alpha12
Samba Release Account [Fri, 28 Nov 1997 03:46:03 +0000 (03:46 +0000)]
preparing for release of 1.9.18alpha12

26 years agothere was a bug in my oplock bugfix :-)
Andrew Tridgell [Fri, 28 Nov 1997 03:39:04 +0000 (03:39 +0000)]
there was a bug in my oplock bugfix :-)

It's fixed properly now :-)

26 years agofixed a very nasty oplock bug. We could send oplock break requests on
Andrew Tridgell [Fri, 28 Nov 1997 03:30:24 +0000 (03:30 +0000)]
fixed a very nasty oplock bug. We could send oplock break requests on
files that aren't open if the file happened to close while the oplock
was in transit. We would end up sending a oplock break request on
another random file (actually the open file that happened to have the
highest fnum). Then we wouldn't get a response, so smbd would keep
sending and would actually flood the net with an infinite number of
oplock break requests!

26 years agooh no! my favourite hack is broken. You can't do:
Andrew Tridgell [Wed, 26 Nov 1997 23:43:06 +0000 (23:43 +0000)]
oh no! my favourite hack is broken. You can't do:

[%U]
path = %H

to get a new homes area. I wonder when it broke?

Anyway, I've fixed it. We can all rest easy now, knowing that you
really can produce an incomprehensible smb.conf and get it to work.

26 years agoget rid of stat command (it is a hangover from an old experiment)
Andrew Tridgell [Tue, 25 Nov 1997 12:20:55 +0000 (12:20 +0000)]
get rid of stat command (it is a hangover from an old experiment)

26 years agoRemind them to read the man page about needing to be root.
Dan Shearer [Mon, 24 Nov 1997 19:10:40 +0000 (19:10 +0000)]
Remind them to read the man page about needing to be root.

26 years agoadded cli_rmdir and cli_mkdir
Andrew Tridgell [Mon, 24 Nov 1997 13:44:52 +0000 (13:44 +0000)]
added cli_rmdir and cli_mkdir

added test in smbtorture for the server updating the directory modify
time when a file is added to a directory

cleanup in smbtorture so no garbage files are left on the server

26 years agoadded a SMB_QUERY_FILE_ALL_INFO test into smbtorture
Andrew Tridgell [Sun, 23 Nov 1997 07:26:42 +0000 (07:26 +0000)]
added a SMB_QUERY_FILE_ALL_INFO test into smbtorture

W95 doesn't seem to support this call.

26 years agoadded some QPATHINFO and QFILEINFO tests into smbtorture.
Andrew Tridgell [Sun, 23 Nov 1997 05:55:44 +0000 (05:55 +0000)]
added some QPATHINFO and QFILEINFO tests into smbtorture.

This tests for things like midnight access times, sticky create times
and word reversed INFO_STANDARD returns

26 years agotest SMBsetatr as well
Andrew Tridgell [Sun, 23 Nov 1997 03:09:59 +0000 (03:09 +0000)]
test SMBsetatr as well

26 years agominor wsmbconf and cgi changes
Andrew Tridgell [Sun, 23 Nov 1997 02:42:22 +0000 (02:42 +0000)]
minor wsmbconf and cgi changes

26 years agoadded a test for the NT SMBgetatr bug in smbtorture
Andrew Tridgell [Sun, 23 Nov 1997 02:41:22 +0000 (02:41 +0000)]
added a test for the NT SMBgetatr bug in smbtorture

added support for choosing the protocol level in smbtorture (-m option)

use -1 for null date in cli_close()

get the attributes right in cli_open()

26 years agothis new cgi code includes the ability to act as a mini web server,
Andrew Tridgell [Sat, 22 Nov 1997 07:51:23 +0000 (07:51 +0000)]
this new cgi code includes the ability to act as a mini web server,
allowing people to use web configuration of Samba without installing a
web server

26 years agofixed some typecasts of (char *) to (unsigned char *)
Andrew Tridgell [Sat, 22 Nov 1997 07:17:32 +0000 (07:17 +0000)]
fixed some typecasts of (char *) to (unsigned char *)

26 years agothe usual :-)
Andrew Tridgell [Sat, 22 Nov 1997 07:10:22 +0000 (07:10 +0000)]
the usual :-)

26 years agoadded true enumerated types in loadparm.c. Now we don't need all those
Andrew Tridgell [Sat, 22 Nov 1997 07:09:12 +0000 (07:09 +0000)]
added true enumerated types in loadparm.c. Now we don't need all those
"interpret_security" and similar functions. This also means that
testparm produces string values for these enumerated types rather than
an integer. This was done to make the web interface much cleaner.

While I was doing this I found a couple of very nasty bugs in the
parm_table code. Someone had added a handle_character_set() function
that treated the char** pointer for a P_STRING as an integer! This
would have caused memory corruption for anyone using the "character
set" option. The loadparm code is delicate - please be careful about
using it's more esoteric features!

I've also removed the "coding system" global and made it a
P_STRING. The integer value was never used anywhere in Samba.

26 years agoJust fixed a *serious* bug in nameelect.c to do with local
Jeremy Allison [Fri, 21 Nov 1997 21:06:40 +0000 (21:06 +0000)]
Just fixed a *serious* bug in nameelect.c to do with local
master browsers. Once we have become a local master, we would
*NEVER* stop announcing ourselves as such. This has serious
implications for the new nmbd code.
Jeremy.

26 years agoAdded in mods for new smb.conf "remote browse sync" option as posted to
John Terpstra [Fri, 21 Nov 1997 14:01:23 +0000 (14:01 +0000)]
Added in mods for new smb.conf "remote browse sync" option as posted to
the samba-1.9.17 tree moments ago.

26 years agouse LocTimeDiff() not TimeDiff() to ensure that longdate conversion is
Andrew Tridgell [Fri, 21 Nov 1997 04:57:37 +0000 (04:57 +0000)]
use LocTimeDiff() not TimeDiff() to ensure that longdate conversion is
consistent (this is really just a cosmetic change)

26 years agoadded some debug lines to the rename code
Andrew Tridgell [Thu, 20 Nov 1997 01:36:36 +0000 (01:36 +0000)]
added some debug lines to the rename code

26 years agochanged nmblookup to only set recursion_desired in queries if the -R
Andrew Tridgell [Wed, 19 Nov 1997 23:50:02 +0000 (23:50 +0000)]
changed nmblookup to only set recursion_desired in queries if the -R
option is used.

26 years agofixed typo
Andrew Tridgell [Wed, 19 Nov 1997 04:12:37 +0000 (04:12 +0000)]
fixed typo

26 years agoipc.c: Changed reply_trans to use receive_next_smb() to cope
Jeremy Allison [Tue, 18 Nov 1997 23:30:49 +0000 (23:30 +0000)]
ipc.c: Changed reply_trans to use receive_next_smb() to cope
       with local message processing.
reply.c: Added check to reply_lockingX for chain after oplock break.
server.c: Added receive_next_smb().
trans2.c: Changed reply_trans2 to use receive_next_smb() to cope
       with local message processing.

26 years agodon't do the 255.255.255.255 mapping for group names if the name type
Andrew Tridgell [Tue, 18 Nov 1997 04:12:54 +0000 (04:12 +0000)]
don't do the 255.255.255.255 mapping for group names if the name type
is 0x1c. This allows remote logon servers when using a Samba WINS server.

I did some experiments with a real NT WINS server and this behaviour
is much closer to being correct. It will only be totally correct when
we add proper group name support.

26 years agoloadparm.c : Added "veto oplock files" parameter.
Jeremy Allison [Mon, 17 Nov 1997 19:16:38 +0000 (19:16 +0000)]
loadparm.c : Added "veto oplock files" parameter.
make_printerdef.c: Fixed warning.
quotas.c: Fixed irix root errors.
server.c: Fixed oplock reference count bug.
smb.h: Added IS_VETO_OPLOCK_PATH().
Jeremy.

26 years agoadded code to test the cli_NetServerEnum() function in clientgen.c
Andrew Tridgell [Tue, 11 Nov 1997 02:51:20 +0000 (02:51 +0000)]
added code to test the cli_NetServerEnum() function in clientgen.c

26 years agofixed a bug which caused nmbd to core dump. The problem was incorrect
Andrew Tridgell [Tue, 11 Nov 1997 02:38:54 +0000 (02:38 +0000)]
fixed a bug which caused nmbd to core dump.  The problem was incorrect
parameters to cli_NetServerEnum()

26 years agoLeft global_oplock_break set while request to break oplock is
Jeremy Allison [Tue, 11 Nov 1997 01:19:32 +0000 (01:19 +0000)]
Left global_oplock_break set while request to break oplock is
outstanding to client. Done to allay Andrews' fears about
readbraw crossing on the wire :-).
Jeremy.

26 years agoauto-create the locks directory on startup
Andrew Tridgell [Tue, 11 Nov 1997 00:48:42 +0000 (00:48 +0000)]
auto-create the locks directory on startup

26 years agoRemoved pwd_validate.c as no longer in main branch.
Jeremy Allison [Mon, 10 Nov 1997 19:38:17 +0000 (19:38 +0000)]
Removed pwd_validate.c as no longer in main branch.
Jeremy

26 years agoRolled back tree state to 11:59pm 8th November 1997 EST to
Jeremy Allison [Mon, 10 Nov 1997 19:23:17 +0000 (19:23 +0000)]
Rolled back tree state to 11:59pm 8th November 1997 EST to
remove problems.
Jeremy

26 years agobyteorder.h :
Luke Leighton [Sun, 9 Nov 1997 20:42:47 +0000 (20:42 +0000)]
byteorder.h :

typecasting for the debug messages of the data dump.  hm.

#ifdef NTDOMAIN

ntclient.c rpc_pipes/lsaparse.c rpc_pipes/ntclientnet.c rpc_pipes/smbparse.c :

adding "Net Server Password Set" functionality to smbclient.

#endif

26 years agoattempting to mark up 32 bit error codes, needed for NT domains.
Luke Leighton [Sun, 9 Nov 1997 17:30:10 +0000 (17:30 +0000)]
attempting to mark up 32 bit error codes, needed for NT domains.
separated out smb server-mode password validation into a separate file.
added called and calling netbios names to client gen state: referenced
section in rfc1002.txt.
created workstation trust account checking code in ntclient.c

there might be a bug in reply_session_setup_andX.  i indented and added { }
around single-line if statements: the lm password checking code now doesn't
look right (around the GUEST_SESSSETUP bits).  *no code semantics have been
changed by the indentation process*.

26 years agomod_smbpasswd() was junking the smbpasswd file...
Luke Leighton [Sat, 8 Nov 1997 14:23:49 +0000 (14:23 +0000)]
mod_smbpasswd() was junking the smbpasswd file...
uninitialised variable warnings.

26 years agoa few more tests added, including one that tests whether the server
Andrew Tridgell [Sat, 8 Nov 1997 05:33:45 +0000 (05:33 +0000)]
a few more tests added, including one that tests whether the server
understand the full range of byte offsets in locking requests. Samba
doesn't (due to the 31 bit limitation in fcntl locking)

26 years agoadded two more sets of tests to the smbtorture test. The tests I added
Andrew Tridgell [Sat, 8 Nov 1997 04:02:05 +0000 (04:02 +0000)]
added two more sets of tests to the smbtorture test. The tests I added
are ones that I know Samba fails. They are:

  1) correct support for retaining locks over a close (ie. the server
     must not use posix semantics)
  2) support for lock timeouts
  3) the server supports multiple locking contexts on the one SMB
  connection, distinguished by PID.
  4) the server correctly fails overlapping locks made by the same PID (this
     goes against POSIX behaviour, which is why it is tricky to implement)
  5) the server denies unlock requests by an incorrect client PID

I've been discussing with Jeremy ways that we can re-implement the
locking code to handle these correctly. This test code will be useful
to see that we have got it right.

26 years agolocal time discrepancies between the date on my computer and the date on
Luke Leighton [Fri, 7 Nov 1997 03:37:18 +0000 (03:37 +0000)]
local time discrepancies between the date on my computer and the date on
chris' computer (probably about four minutes) meant that chris' commits
didn't get downloaded when i did a cvs update.

therefore, i did a cvs update; make proto; cvs commit, and still got problems.

so, five minutes later, i do a cvs update, and _then_ chris' server.c and
mangle.c mods get downloaded, and i have to do another make proto.

this particular problem has hit us *really* badly in the past, because i
was taking files home (onto a portable that i hadn't reset the time on
since i went to the states) and then recopying them back onto the computer
here.  if i do this, i do a touch *.c *.h; cvs -t update.  *then* a cvs commit.

26 years agoipc.c :
Luke Leighton [Fri, 7 Nov 1997 03:28:43 +0000 (03:28 +0000)]
ipc.c :

added a #define around the alignment thing: it's a way to stop
NetMonitor from decoding your packets!!!!

proto.h :

usual.

reply.c :

added what i believe to be the correct error messages for getting
correct domain joining.

smb.h :

some guesses at good names of the SAMR_XXXX functions.  sorting
out the SAMR_LOOKUP_RIDS function.  this is *not* the same as
the LSA_LOOKUP_RIDS function, unless paul accidentally put it
on the ntlsa pipe by mistake, instead of the samr pipe :-)

rpc_pipes/lsa_hnd.c rpc_pipes/pipe_hnd.c :

moved creation and allocation of unique policy handles into this module.

rpc_pipes/pipesamr.c rpc_pipes/samrparse.c rpc_pipes/smbparse.c :

SAMR_LOOKUP_RIDS is beginning to look _suspiciously_ like the
LSA_LOOKUP_RIDS function.  but i know that there are subtle
discrepancies.

26 years agoModified Files:
Christopher R. Hertel [Fri, 7 Nov 1997 03:06:24 +0000 (03:06 +0000)]
Modified Files:
mangle.c server.c proto.h

mangle.c
  I am planning to replace the mangled_stack array with a proper stack,
  but found many style inconsistencies (no, really).  As you might expect,
  I have standardized on my own preferences.  ;)

  I also found a potential problem in create_mangled_stack (which I've
  renamed as reset_mangled_stack).  If the stack size were passed into
  the function as 0 or less, there was the possibility that the array
  would have been freed twice.  I doubt that this ever happens, but I
  don't like to leave holes.

  Of course, the fix will be irrelevent once I replace the array with
  a linked-list-based stack.

server.c
  Changed the call to create_mangled_stack() to a call to reset_mangled_stack().

proto.h
  Regenerated to match the above changes.  (A real comment!  How unusual!)

26 years agoipc.c: Changing back arbitrary alignment change until I know *why*
Jeremy Allison [Thu, 6 Nov 1997 23:34:51 +0000 (23:34 +0000)]
ipc.c: Changing back arbitrary alignment change until I know *why*
luke changed it.
proto.h: The usual.
uid.c: Fix crash bug when attaching with smbclient -mCORE. A vuid
pointer was being used when it was null.
Jeremy.

26 years agofollowing a cvs error, i am rewriting this monster-commit. with bad grace.
Luke Leighton [Thu, 6 Nov 1997 23:03:58 +0000 (23:03 +0000)]
following a cvs error, i am rewriting this monster-commit.  with bad grace.

Modified Files:
---------------

Makefile:

adding extra files

ipc.c :

send_trans_reply() - alignment issue.  this makes the alignment
the same as that in NT.  this should be looked at by people who
understand the SMB stuff better than i.

api_fd_commands[] - added samr and wkssvc pipes.

loadparm.c :

lp_domain_controller() changed to mean "samba is a domain controller".
it's a "yes/no" parameter, now.  no, it isn't used _anywhere_.

namedbwork.c nameelect.c :

if "domain controller = yes" then add SV_TYPE_DOMAIN_CTRL to the
host _and_ workgroup announcements.  yes, you must do both: nt does.

namelogon.c :

important NETLOGON bug in SAMLOGON request parsing, which may be
the source of some people's problems with logging on to the Samba PDC.

password.c :

get_smbpwnam() renamed to get_smbpwd_entry().

pipes.c :

added samr and wkssvc pipes.

proto.h :

usual.  can we actually _remove_ proto.h from the cvs tree, and
have it as one of the Makefile dependencies, or something?

reply.c :

get_smbpwnam() renamed to get_smbpwd_entry() - also changed response
error code when logging in from a WORKSTATION$ account.  yes, paul
is right: we need to know when to return the right error code, and why.

server.c :

added call to reset_chain_pnum().

#ifdef NTDOMAIN added call to init_lsa_policy_hnd() #endif.  jeremy,
you'd be proud: i did a compile without NTDOMAIN, and caught a link
error for this function.

smb.h :

defines and structures for samr and wkssvc pipes.

smbpass.c :

modified get_smbpwnam() to get_smbpwd_entry() and it now takes
two arguments.  one for the name; if this is null, it looks up
by smb_userid instead.

oh, by the way, smb_userids are actually domain relative ids
(RIDs).  concatenate a RID with the domain SID, and you have
an internet globally unique way of identifying a user.

we're using RIDs in the wrong way....

added mod_smbpwnam() function.  this was based on code in smbpasswd.c

rpc_pipes/lsaparse.c :

added enum trusted domain parsing.  this is incomplete: i need
a packet trace to write it properly.

rpc_pipes/pipe_hnd.c :

added reset_chain_pnum() function.

rpc_pipes/pipenetlog.c :

get_smbpwnam() function renamed to get_smbpwd_entry().

arcfour() issues.

removed capability of get_md4pw() function to automatically add
workstation accounts.  this should either be done using
smbpasswd -add MACHINE$, or by using \PIPE\samr.

rpc_pipes/pipe_util.c :

create_pol_hnd() - creates a unique LSA Policy Handle.  overkill
function: uses a 64 bit sequence number; current unix time and
the smbd pid.

rpc_pipes/smbparse.c :

arcfour() issues.

smb_io_unistr2() should advance by uni_str_len not uni_max_len.

smb_io_smb_hdr_rb() - request bind uses uint16 for the context
id, and uint8 for the num_syntaxes.  oops, i put these both as
uint32s.

Added Files:
------------

rpc_pipes/lsa_hnd.c :

on the samr pipe, allocate and associate an LSA Policy Handle
with a SID.  you receive queries with the LSA Policy Handle,
and have to turn this back into a SID in order to answer the
query...

rpc_pipes/pipesamr.c rpc_pipes/samrparse.c

\PIPE\samr processing.  samr i presume is the SAM Replication pipe.

rpc_pipes/pipewkssvc.c rpc_pipes/wksparse.c

\PIPE\wkssvc processing.  the Workstation Service pipe?

holy cow.

26 years agolocal.h: Added OPLOCK_BREAK_TIMEOUT_FUDGEFACTOR.
Jeremy Allison [Thu, 6 Nov 1997 19:11:16 +0000 (19:11 +0000)]
local.h: Added OPLOCK_BREAK_TIMEOUT_FUDGEFACTOR.
proto.h: The usual.
server.c: Added timestring() messages to oplock logs. Added fix
          to allow file open processing to continue is an oplock
          break message times out.
Jeremy.

26 years agochange from %D to %m/%d/%Y in timestring(). This doesn't really matter
Andrew Tridgell [Thu, 6 Nov 1997 13:45:50 +0000 (13:45 +0000)]
change from %D to %m/%d/%Y in timestring(). This doesn't really matter
as it is only for logging but it will prevent people from writing in
to say that we are not y2k compliant after running auto-diagnostic
tests.

26 years agono asm/signal.h for linux as this is not portable
Andrew Tridgell [Thu, 6 Nov 1997 06:50:21 +0000 (06:50 +0000)]
no asm/signal.h for linux as this is not portable

add auto username from LOGNAME for smbtorture

proto fix in ufc.c

26 years agoChange to always register the WORKGROUP<1e> name as does Win95.
Jeremy Allison [Wed, 5 Nov 1997 22:09:15 +0000 (22:09 +0000)]
Change to always register the WORKGROUP<1e> name as does Win95.
Jeremy.

26 years agoadd NO_SEMUN for HP
Andrew Tridgell [Wed, 5 Nov 1997 12:21:45 +0000 (12:21 +0000)]
add NO_SEMUN for HP

I really wish you could say

#ifndef "struct foo_struct"

in C :-)

26 years agosmb.h :
Luke Leighton [Tue, 4 Nov 1997 20:27:08 +0000 (20:27 +0000)]
smb.h :

LSA_OPENSECRET and three other unknown #defines are commented as not
being the real defines for these!!!

rpc_pipes/pipenetlog.c :

USE_ARCFOUR for Server Password Set.  and then ignore the password (agh!)
this will *fail* to do logins, by the way.  i need to add a routine
to update the smb password (next job).

26 years agoRolling back the files
Jeremy Allison [Mon, 3 Nov 1997 19:24:45 +0000 (19:24 +0000)]
Rolling back the files
loadparm.c : to equivalent to version 1.67
reply.c : to equivalent to version 1.69
server.c : to equivalent to version 1.122
util.c : to equivalent to version 1.98

to remove the incorrect changes.

proto.h: The usual.

rpc_pipes/smbparse.c : Backeting stuff that SHOULD NOT BE IN THE
                       none-NTDOMAIN build !

Jeremy.

26 years agoadded code that _uses_ arcfour. arcfour itself, or anything remotely
Luke Leighton [Mon, 3 Nov 1997 15:47:03 +0000 (15:47 +0000)]
added code that _uses_ arcfour.  arcfour itself, or anything remotely
like it, has *not* been added.

this is the client and server side of the SAM Logon NT and LM OWF password
obfuscation (rc4 with the long-term session key).

26 years agopreparing for release of 1.9.18alpha11
Samba Release Account [Sun, 2 Nov 1997 20:47:35 +0000 (20:47 +0000)]
preparing for release of 1.9.18alpha11

26 years agoChristian Lademann's contribution: new capabilities in smb.conf.
Luke Leighton [Sun, 2 Nov 1997 20:35:20 +0000 (20:35 +0000)]
Christian Lademann's contribution:  new capabilities in smb.conf.

'<' and '|' characters indicate read file and execute command respectively,
and feed the output into the parameter (!!!).

'<$' and '|$' means run standard_sub_basic() on them.

this is going to be fun to document in smb.conf.5....

also, Christian created a new "online" service parameter.  services can
be taken "off-line"....

26 years agoupdated
Luke Leighton [Sun, 2 Nov 1997 19:59:20 +0000 (19:59 +0000)]
updated

26 years agoupdated
Luke Leighton [Sun, 2 Nov 1997 19:47:48 +0000 (19:47 +0000)]
updated

26 years agodocumented "domain hosts allow/deny" parameters
Luke Leighton [Sun, 2 Nov 1997 19:40:07 +0000 (19:40 +0000)]
documented "domain hosts allow/deny" parameters

26 years agoloadparm.c :
Luke Leighton [Sun, 2 Nov 1997 19:27:26 +0000 (19:27 +0000)]
loadparm.c :

added "domain hosts allow" and "domain hosts deny".  these are to be
used to specify which workstations can log in to a samba PDC from.
it is also used to check whether to add an initial machine password
into the smbpasswd database or not

smbpass.c :

added capability to add a machine password to the smbpasswd database.
***** the default uid is zero *****

rpc_pipes/pipenetlog.c :

use of "domain hosts allow/deny" parameters to allow login access.

proto.h :

usual.

26 years agofix some uchar/char conflicts
Andrew Tridgell [Sun, 2 Nov 1997 04:11:05 +0000 (04:11 +0000)]
fix some uchar/char conflicts