Re: Bug found in smbclient

[Richard Sharpe]

On Mon, 13 May 2002, Jay Ts wrote:

> Hi,
> 
> Just in case I'm the first to have noticed this:
> 
> $ smbclient //server/share  -c "ls" -Tc
> added interface ip=172.16.1.3 bcast=172.16.1.255 nmask=255.255.255.0
> Segmentation fault

Well, it happens to me as well. Did you manage to get a traceback from the 
core dump?
 
> This happened when out of curiosity, I wondered if
> it was possible to execute some smbclient commands,
> such as a cd to a subdirectory, before running the
> -Tc tar backup of it.
> 
> The "server" and "share" strings in the above command
> can be anything, and the command with -c can also be
> a "cd dir" command. Just a plain "cd" with no arguments
> resulted in odd behavior (the cd command ran, then
> smbclient hung, without doing a tar backup). I haven't
> tried any other smbclient commands with -c.
> 
> Jay Ts
> [EMAIL PROTECTED]
> 

-- 
Regards
-
Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED], 
[EMAIL PROTECTED]

Re: Adding Recursion Available bit to name_query and node_status_query

[Richard Sharpe]

On Mon, 13 May 2002, Christopher R. Hertel wrote:

> Sorry.  Just to be clear, the older version of Samba that I'm running
> returns RA TRUE *if* it is running as a WINS server.
> 
> I don't know what a newer version of Samba does, but I doubt this has
> changed and I really should check to see what W/NT and W2K do if they are
> running the WINS server.

Well, Samba 2,2,3pre that is not a WINS server sets RA, as does Win2K that 
is not a WINS server, but so to does WinNT that is a DC and WINS server 
and Win2K that is a WINS server,

Looke like all and sundry set the bit!

Regards
-
Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED], 
[EMAIL PROTECTED]

Re: Adding Recursion Available bit to name_query and node_status_query

[Richard Sharpe]

On Mon, 13 May 2002, Christopher R. Hertel wrote:

> Sorry.  Just to be clear, the older version of Samba that I'm running
> returns RA TRUE *if* it is running as a WINS server.
> 
> I don't know what a newer version of Samba does, but I doubt this has
> changed and I really should check to see what W/NT and W2K do if they are
> running the WINS server.

OK, I have added support for these flags to name_query, and have modified 
all places that use it, I believe.

However, I have not modified nmblookup.c to print the results out as yet, 
as it requires some thought, but those who want to can experiment.

Regards
-
Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED], 
[EMAIL PROTECTED]

Samba 2.2.4 & daemontools

[Michael Handler]

I'm trying to run smbd and nmbd under Dan Bernstein's supervise &
svscan process control programs from his daemontools package
(http://cr.yp.to/daemontools.html). Jos Backus has raised this issue
in the past, re: the -i "interactive" flag and making smbd and nmbd
create their own process groups, but unfortunately the -i flag isn't
sufficient to make smbd run correctly under daemontools, as it makes
smbd only answer the first connection comes in. Additionally, the
-i flag forces smbd and nmbd to log to stdout, which may not be
desireable in all situations.

Attached please find a set of patches against Samba 2.2.4 to make
the following changes:

* smbd, nmbd, and winbindd now all have -S and -F options. The -S flag
  means log to stdout. The -F flag makes the daemons run in the
  foreground properly for use under daemontools, e.g. all operations
  are normal, setsid(2) is called for process group creation, but the
  fork(2)/exit pair and the closing of fd's 0, 1, and 2 (for rsh) are
  skipped.

* when setup_logging() selects stdout to write all logs to, it sets
  that filedescriptor to "unbuffered" to ensure that logs are written out
  promptly, especially under multilog (from daemontools).

Samba team members, if this code and changes are acceptable to you,
I would appreciate being told what work I need to do in order to
ensure that this functionality is rolled forward into Samba 3.0 and
forward, and what documentation patches I should submit for both
2.0 and 3.0.

If this code or changes are unacceptable for any reason, please don't
hesitate to let me know so we can begin a dialogue about the proper way
to implement this functionality into Samba.

Thank you all very much for your excellent work on Samba over the years.

--michael

-- 
[EMAIL PROTECTED] (michael handler)   washington, dc


diff -x configure -x *.in -ur /home/handler/src/samba-2.2.4/source/include/proto.h 
work/samba-2.2.4/source/include/proto.h
--- /home/handler/src/samba-2.2.4/source/include/proto.hThu May  2 21:03:05 
2002
+++ work/samba-2.2.4/source/include/proto.h Mon May 13 20:51:26 2002
@@ -1125,7 +1125,7 @@
ssize_t (*write_fn)(int, const void *, 
size_t));
 SMB_OFF_T transfer_file(int infd,int outfd,SMB_OFF_T n);
 void msleep(unsigned int t);
-void become_daemon(void);
+void become_daemon(BOOL fork);
 BOOL yesno(char *p);
 void *Realloc(void *p,size_t size);
 void safe_free(void *p);
diff -x configure -x *.in -ur /home/handler/src/samba-2.2.4/source/lib/debug.c 
work/samba-2.2.4/source/lib/debug.c
--- /home/handler/src/samba-2.2.4/source/lib/debug.cThu May  2 21:03:07 2002
+++ work/samba-2.2.4/source/lib/debug.c Mon May 13 22:36:00 2002
@@ -291,6 +291,7 @@
if (interactive) {
stdout_logging = True;
dbf = stdout;
+   setbuf(stdout, NULL);
}
 #ifdef WITH_SYSLOG
else {
diff -x configure -x *.in -ur /home/handler/src/samba-2.2.4/source/lib/util.c 
work/samba-2.2.4/source/lib/util.c
--- /home/handler/src/samba-2.2.4/source/lib/util.c Thu May  2 21:03:11 2002
+++ work/samba-2.2.4/source/lib/util.c  Mon May 13 20:49:52 2002
@@ -607,10 +607,12 @@
  Become a daemon, discarding the controlling terminal.
 /
 
-void become_daemon(void)
+void become_daemon(BOOL fork)
 {
-   if (sys_fork()) {
-   _exit(0);
+   if (fork) {
+   if (sys_fork()) {
+   _exit(0);
+   }
}
 
   /* detach from the terminal */
@@ -626,8 +628,10 @@
}
 #endif /* HAVE_SETSID */
 
-   /* Close fd's 0,1,2. Needed if started by rsh */
-   close_low_fds();
+   if (fork) {
+   /* Close fd's 0,1,2. Needed if started by rsh */
+   close_low_fds();
+   }
 }
 
 /
diff -x configure -x *.in -ur /home/handler/src/samba-2.2.4/source/nmbd/nmbd.c 
work/samba-2.2.4/source/nmbd/nmbd.c
--- /home/handler/src/samba-2.2.4/source/nmbd/nmbd.cThu May  2 21:03:17 2002
+++ work/samba-2.2.4/source/nmbd/nmbd.c Mon May 13 23:09:47 2002
@@ -643,9 +643,11 @@
 static void usage(char *pname)
 {
 
-  printf( "Usage: %s [-DaiohV] [-H lmhosts file] [-d debuglevel] [-l log 
basename]\n", pname );
+  printf( "Usage: %s [-DFSaiohV] [-H lmhosts file] [-d debuglevel] [-l log 
+basename]\n", pname );
   printf( "   [-n name] [-p port] [-s configuration file]\n" );
   printf( "\t-DBecome a daemon (default)\n" );
+  printf( "\t-FRun daemon in foreground (for daemontools, etc)\n" 
+);
+  printf( "\t-SLog to stdout\n" );
   printf( "\t-aAppend to log file (default)\n" );
   printf( "\t-iRun interactive (not a daemon)\n" );
   printf( "\t-oOverwrite log file, don't append\

Re: 2.2.2 Bug in unix_to_nt_time() + Fix

[Jordan Russell]

Jeremy Allison wrote:
> On Sat, Feb 02, 2002 at 03:58:51AM -0600, Jordan Russell wrote:
> > Today I came across a significant bug in time.c's unix_to_nt_time()
function
> > that causes file time stamps to be 1 hour off when a file was last
modified
> > within hours of Daylight Savings Time going into effect.
>
> This is probably right, and I'll apply it to HEAD - but
> too risky to add to 2.2.3 at this stage. 2.2.3-pre has
> been used in production in a number of places and I don't
> want to peturb the time code at this point.

It does not appear that the fix made its way into CVS. The bug -
http://lists.samba.org/pipermail/samba-technical/2002-February/034464.html -
still exists in 2.2.4.

Attached is the latest version of my one-line patch, which changes the
incorrect LocTimeDiff call into a TimeDiff call. I would greatly appreciate
it if the fix were included in 2.2.5.

Thanks,
Jordan Russell



samba-2.2.3a-timefix.patch
Description: Binary data

Re: Adding Recursion Available bit to name_query and node_status_query

[Christopher R . Hertel]

Sorry.  Just to be clear, the older version of Samba that I'm running
returns RA TRUE *if* it is running as a WINS server.

I don't know what a newer version of Samba does, but I doubt this has
changed and I really should check to see what W/NT and W2K do if they are
running the WINS server.

Chris -)-

On Mon, May 13, 2002 at 09:10:09PM -0500, Christopher R.Hertel wrote:
> > On Mon, 13 May 2002, Christopher R. Hertel wrote:
> :
> > > RA - Recursion Available.
> > >  Should only be set in messages which come from the NBNS (WINS)
> > >  server itself.  Note that a machine that is running WINS will
> > >  also have a local name table.  If the reply comes from the
> > >  local name table, then RA should be clear even if the same node
> > >  is running the NBNS (WINS) server.
> 
> Testing against an older version of Samba, this doesn't appear to be the
> way things work.  I would like to test against a Windows box, but I don't
> have one that will do the job just now.  The older version of Samba that
> I run on a machine in the basement returns RA TRUE even if the name is
> from the local name table.
> 
> I really would like to see what W/NT and/or W2K do...
> Richard's fixes should make that easier.
> 
> Chris -)-

-- 
Samba Team -- http://www.samba.org/ -)-   Christopher R. Hertel
jCIFS Team -- http://jcifs.samba.org/   -)-   ubiqx development, uninq.
ubiqx Team -- http://www.ubiqx.org/ -)-   [EMAIL PROTECTED]
OnLineBook -- http://ubiqx.org/cifs/-)-   [EMAIL PROTECTED]

Re: Adding Recursion Available bit to name_query and node_status_query

[Christopher R . Hertel]

> On Mon, 13 May 2002, Christopher R. Hertel wrote:
:
> > RA - Recursion Available.
> >  Should only be set in messages which come from the NBNS (WINS)
> >  server itself.  Note that a machine that is running WINS will
> >  also have a local name table.  If the reply comes from the
> >  local name table, then RA should be clear even if the same node
> >  is running the NBNS (WINS) server.

Testing against an older version of Samba, this doesn't appear to be the
way things work.  I would like to test against a Windows box, but I don't
have one that will do the job just now.  The older version of Samba that
I run on a machine in the basement returns RA TRUE even if the name is
from the local name table.

I really would like to see what W/NT and/or W2K do...
Richard's fixes should make that easier.

Chris -)-

-- 
Samba Team -- http://www.samba.org/ -)-   Christopher R. Hertel
jCIFS Team -- http://jcifs.samba.org/   -)-   ubiqx development, uninq.
ubiqx Team -- http://www.ubiqx.org/ -)-   [EMAIL PROTECTED]
OnLineBook -- http://ubiqx.org/cifs/-)-   [EMAIL PROTECTED]

sorry!

[Tristan Ball]

Sorry about the large attachment everyone, I tried canceling the send as
soon as I noticed how large the attachment was, but I obviously wasn't
fast enough.

Apologies,
T.

Too Many Open Files on 2.2.4???

[Tristan Ball]

Guys, I'm getting the following in some of my logs:

[2002/05/14 08:36:21, 0] lib/debug.c:reopen_logs(349)
  Unable to open new log file /var/log/samba/log.dancer: Too many open
files
[2002/05/14 08:36:21, 0] lib/username.c:map_username(130)
  can't open username map /etc/samba/smb.user.map. Error Too many open
files

However I can't see how I'd be anyware near the limit of file
discriptors. I've used plimit, ulimit, and MAX_OPEN_FILES in local.h to
up every limit I can find, and it's still happening!

Host is solaris 2.8, samba 2.2.4 from cvs. Clients are Win nt/2k,
running clearcase 4.2.

If I pick one of the samba processes thats given log errors, and run
pfiles on it, It shows 1303 open files, but an rlimit of 2.

I've just upped the log level in samba, and the interesting thing that I
see is that a successfull open call contains a relatively small number
of open files (<500), but a very high fsp->fnum (>9000). Is this
important??? :-)

I have some level 3 logs, but due to the clients making a massive number
of TRANSACT2_QFILEINFO, a 2 minute snapshot of the logs, containing a
both a good open call, and then a "too many files" call is 54K
compressed, a little to large to attach to the list!.

I'm still looking into this, but if anyone has any ideas, I'd love to
hear them!

Thanks,
T

Too Many Open Files on 2.2.4???

[Tristan Ball]

Guys, I'm getting the following in some of my logs:

[2002/05/14 08:36:21, 0] lib/debug.c:reopen_logs(349)
  Unable to open new log file /var/log/samba/log.dancer: Too many open
files
[2002/05/14 08:36:21, 0] lib/username.c:map_username(130)
  can't open username map /etc/samba/smb.user.map. Error Too many open
files

However I can't see how I'd be anyware near the limit of file
discriptors. I've used plimit, ulimit, and MAX_OPEN_FILES in local.h to
up every limit I can find, and it's still happening!

Host is solaris 2.8, samba 2.2.4 from cvs. Clients are Win nt/2k,
running clearcase 4.2.

If I pick one of the samba processes thats given log errors, and run
pfiles on it, It shows 1303 open files, but an rlimit of 2.

I've just upped the log level in samba, and the interesting thing that I
see is that a successfull open call contains a relatively small number
of open files, but a very high fsp->fnum. Is this important??? :-)

I've attached a snippet of a level 3 log, from the last successfull open
to a "too many open" error. Apologies for the binary attachment!

I'm still looking into this, but if anyone has any ideas, I'd love to
hear them!

Thanks,
T



Log1.zip
Description: Zip compressed data

Re: Dumped Packets?

[Tim Potter]

On Mon, May 13, 2002 at 03:50:41PM -0500, root wrote:

> I was feeling rather underinformed one day, and got rather bent out of
> shape about it, so I turned my Samba debug level up to 100, because I
> can. I got the info I needed at the time, but today I noticed raw SMB
> packets being dumped into /tmp, each in an individual binary
> file. There are rather a lot of them in there. This could be useful,
> in some testing scenarios, if they could be converted from the binary
> format they are in.
> 
> Is there some tool which will dump these packets in human readable
> form? Can I cat them all together in some way and look at them with
> Ethereal?

You might be able to import them into a capture file somehow.  Ethereal has
a library for reading and writing capture files of various formats which
you could use for this.

I had a better idea the other day.  There should be enough information in
a level 10 debug log tyo completely reconstruct a capture file.  So you
should be able to dummy up ethernet/ip/tcp headers and insert this into
a capture file and you have an instant ethereal trace!


Tim.

Re: winbindd uid and gid range assumptions

[Andrew Bartlett]

Mike Gerdts wrote:
> 
> It seems as though there was an assumption that users of winbindd would
> have switched entirely to NT domain authentication or that they would
> have distinct range of UIDs/GIDs for /etc/passwd (or NIS) and NT domain.

> Is there any good reason to not just let get_id_from_sid() in
> nsswitch/winbindd_idmap.c do all the checking?  It seems as though the
> optimization achieved by the multitude of checks is minimal and greatly
> reduces the chances integration possibilities between winbindd and
> traditional unix authentication.

Moving over the socket is a very expencive operation, particularly
compared to a simple if statement.  Also, where we know that a uid is
local, we need to check with code that winbind isn't linked to - the
passdb backend.

But yes, we need to deal with things like getting the uid from the SFU
LDAP schema, so this may well change in the future. 

Andrew Bartlett

-- 
Andrew Bartlett [EMAIL PROTECTED]
Manager, Authentication Subsystems, Samba Team  [EMAIL PROTECTED]
Student Network Administrator, Hawker College   [EMAIL PROTECTED]
http://samba.org http://build.samba.org http://hawkerc.net

Re: session exec ???

[Andrew Bartlett]

Kevin Wheatley wrote:
> 
> Hi,
> 
> I was wondering if anything ever came of the session exec suggestion
> from earlier in the year ?

Not yet.

Andrew,

-- 
Andrew Bartlett [EMAIL PROTECTED]
Manager, Authentication Subsystems, Samba Team  [EMAIL PROTECTED]
Student Network Administrator, Hawker College   [EMAIL PROTECTED]
http://samba.org http://build.samba.org http://hawkerc.net

Re: Adding Recursion Available bit to name_query and node_status_query

[Christopher R. Hertel]

Richard Sharpe wrote:
:
> These would be nice, except that I have found that these bits are
> all separate BOOLs in the decoded structure, instead of being a
> bit-field.
> 
> I will have to re-assemble them :-(

Eee.  Yuck.

-- 
Christopher R. Hertel -)-   University of Minnesota
[EMAIL PROTECTED]  Networking and Telecommunications Services

Re: Adding Recursion Available bit to name_query and node_status_query

[Richard Sharpe]

On Mon, 13 May 2002, Christopher R. Hertel wrote:

> 
> 1) It would be nice to see the flags on the outgoing packet as well
>as the repl[y|ies].
> 
> 2) It might make sense produce this output on debug level 1, rather
>than by default.  Or perhaps incoming flags at -d1 and both sets
>at -d2.
> 
> 3) It would also be nice to see the RCode value in any error message
>received (we may already provide the diagnostics).
> 
> For those just tuning in from home, here's a rundown of the flag
> values we are talking about:
> 
> R  - Reply bit.
>  This should always be set on packets received by nmblookup.  It
>  would be an error if it wasn't.
> 
> OpCode - Operation Code.
>  Always 0x00 for queries (which is what nmblookup sends).
> 
> AA - Authoritative Answer.
>  Likewise.  I *think* that the AA flag is a hold-over from the
>  DNS system (the basis for the NBT Name Service).
> 
>  I would love to see this reported, however, because I am
>  wondering whether WINS Proxies (does anyone still use those?)
>  leave this bit clear.
> 
> TC - Truncation flag.
>  No one has, to my knowledge, ever implemented truncated Name
>  Service packets.  According to the RFCs, if you get a packet
>  with the TC bit set, the you can negotiate the use of TCP.
>  (Yes, that would be TCP on port 137.  No one ever implemented
>  that, as far as I know.)  I believe that all implementations
>  ignore this bit (though I've not tested to see what Windows
>  does if this is set).  I would be very surprised to see this
>  bit set.
> 
> RD - Recursion Desired.
>  Only has meaning in request messages.  Used to indicate that
>  the NBNS (WINS) server should be consulted.
> 
>  There is a lengthy discussion of all of this in section 3.4.3.2
>  of my docs.  The RD bit is used heavily to distinguish between
>  requests, queries, and demands.  In reply messages, however, its
>  state is (probably) not significant at all.  My guess is that it
>  will typically match the state of the RD bit in the request.
> 
> RA - Recursion Available.
>  Should only be set in messages which come from the NBNS (WINS)
>  server itself.  Note that a machine that is running WINS will
>  also have a local name table.  If the reply comes from the
>  local name table, then RA should be clear even if the same node
>  is running the NBNS (WINS) server.
> 
> B  - Broadcast bit.
>  Set when the packet is being sent as a broadcast.  Interesting
>  to play with this bit, particularly in conjunction with the RD
>  bit.
> 
> RCode - Result code.
>  0 for success, else an error code.

These would be nice, except that I have found that these bits are all 
separate BOOLs in the decoded structure, instead of being a bit-field.

I will have to re-assemble them :-(

Regards
-
Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED], 
[EMAIL PROTECTED]

Re: winbind UID, GID assignment

[Andrew Bartlett]

Mike Gerdts wrote:
> 
> On Mon, 2002-05-13 at 07:35, Andrew Bartlett wrote:
> > The username and domain may not be known at the time a mapping is
> > required.  Thats the easy bit - we might not even know if it is a uid or
> > gid!
> 
> Have you taken a look at the patch that I created?  If not, please take
> a look at http://www.cae.wisc.edu/~gerdts/samba/ and let me know if the
> changes to nsswitch/winbindd_idmap.c address the issue of not knowing
> whether the SID is a user or group (or other..) SID.

The problem is that we may not be able to resolve the SID at this
time.   We can't even assume that the DC is even contactable in some
situtations, let alone that it is alive and responding to requests in a
timely manner.

Andrew Bartlett

-- 
Andrew Bartlett [EMAIL PROTECTED]
Manager, Authentication Subsystems, Samba Team  [EMAIL PROTECTED]
Student Network Administrator, Hawker College   [EMAIL PROTECTED]
http://samba.org http://build.samba.org http://hawkerc.net

Dumped Packets?

[root]

I was feeling rather underinformed one day, and got rather bent out of
shape about it, so I turned my Samba debug level up to 100, because I
can. I got the info I needed at the time, but today I noticed raw SMB
packets being dumped into /tmp, each in an individual binary
file. There are rather a lot of them in there. This could be useful,
in some testing scenarios, if they could be converted from the binary
format they are in.

Is there some tool which will dump these packets in human readable
form? Can I cat them all together in some way and look at them with
Ethereal?

-- 
Andrew C. Esh  mail:[EMAIL PROTECTED]
Tricord Systems, Inc.
2905 Northwest Blvd., Suite 20   763-557-9005 (main)
Plymouth, MN 55441-2644 USA  763-551-6418 (direct)
http://www.tricord.com - Tricord Home Page

Re: winbind authentication

[Juergen Hasch]

Hi Tim,

Am Montag, 13. Mai 2002 02:38 schrieb Tim Potter:
> On Sun, May 12, 2002 at 10:26:41AM +0200, Juergen Hasch wrote:
> > this problem applies to 2.2.4 and partly to head, too.
> >
> > I need to set a username when connecting to a Win2K server using winbind.
> > The documentiation states "wbinfo -A user%password" would do the trick.
>
> Hi - I've checked this fix in to various branches.

thanks. What is the reason you didn't change the variable "username" in the
secrets_store(SECRET_AUTH_USER,...) call to "user" ?

Now SECRETS_AUTH_USER is still set to "username%password" instead of 
"username" alone.

...Juergen

Re: winbind UID, GID assignment

[Mike Gerdts]

On Mon, 2002-05-13 at 07:35, Andrew Bartlett wrote:
> The username and domain may not be known at the time a mapping is
> required.  Thats the easy bit - we might not even know if it is a uid or
> gid!

Have you taken a look at the patch that I created?  If not, please take
a look at http://www.cae.wisc.edu/~gerdts/samba/ and let me know if the
changes to nsswitch/winbindd_idmap.c address the issue of not knowing
whether the SID is a user or group (or other..) SID.

> There is more to this whole mess than meets the eye, but yes, we do need
> to allow an aribtary SID->uid/gid, uid/gid -> SID backend system.

Would you be open to following a structure like the VFS uses now?  The
existing functions could go into the default_idmap_ops structure and
plugins could replace any or all of those ops.  I am quite motivated to
provide a patch that does this.

Mike 

Hot wo get a files_struc tin VFS

[Claudia Moroder]

Hello,

I am triing to write a VFS that uncompresses compressed
read only files in the background.

I have the following problem: in VFS open returns a int,
but all the other functions want a files_struct.

How can I fill up cleanly this struct. Is there a
standard function to use ?

Thanks

Andreas

-
This mail sent through IMP: http://web.horde.org/imp/

winbindd uid and gid range assumptions

[Mike Gerdts]

It seems as though there was an assumption that users of winbindd would
have switched entirely to NT domain authentication or that they would
have distinct range of UIDs/GIDs for /etc/passwd (or NIS) and NT domain.

The most obvious sign of this is the peppering of statements like the
following throughout various files:

/* Bug out if the gid isn't in the winbind range */

if ((state->request.data.gid < server_state.gid_low) ||
(state->request.data.gid > server_state.gid_high)) {
return WINBINDD_ERROR;
}

Also, smbd/uid.c does a similar check before calling
winbind_[ug]id_to_sid():

if (lp_winbind_uid(&low, &high) && uid >= low && uid <= high) {
if (winbind_uid_to_sid(psid, uid)) {

DEBUG(10,("uid_to_sid: winbindd %u -> %s\n",
(unsigned int)uid, sid_to_string(sid, psid)));

return psid;
}
}

Is there any good reason to not just let get_id_from_sid() in
nsswitch/winbindd_idmap.c do all the checking?  It seems as though the
optimization achieved by the multitude of checks is minimal and greatly
reduces the chances integration possibilities between winbindd and
traditional unix authentication.

Mike

Re: A IE 6.0 patch

[rajesh pancholi]

 What is IE 6.0 Patch ? I don't know U  ? Is htis software ?
 
  samba-technical <[EMAIL PROTECTED]> wrote: 
Hi,This is a IE 6.0 patchI wish you would like it. Do You Yahoo!?
LAUNCH - Your Yahoo! Music Experience

Re: Adding Recursion Available bit to name_query and node_status_query

[Christopher R. Hertel]

Richard Sharpe wrote:
:
> > > Any objections?
> >
> > Nope.
> 
> Actually, I was thinking that it would be useful to return the entire
> set of flags.
> 
> This might be better than returning just one of them.
> 
> Any further thoughts?

Well, you've got me thinking (always a dangerous thing to do).  There
are three things that came to mind:

1) It would be nice to see the flags on the outgoing packet as well
   as the repl[y|ies].

2) It might make sense produce this output on debug level 1, rather
   than by default.  Or perhaps incoming flags at -d1 and both sets
   at -d2.

3) It would also be nice to see the RCode value in any error message
   received (we may already provide the diagnostics).

For those just tuning in from home, here's a rundown of the flag
values we are talking about:

R  - Reply bit.
 This should always be set on packets received by nmblookup.  It
 would be an error if it wasn't.

OpCode - Operation Code.
 Always 0x00 for queries (which is what nmblookup sends).

AA - Authoritative Answer.
 Likewise.  I *think* that the AA flag is a hold-over from the
 DNS system (the basis for the NBT Name Service).

 I would love to see this reported, however, because I am
 wondering whether WINS Proxies (does anyone still use those?)
 leave this bit clear.

TC - Truncation flag.
 No one has, to my knowledge, ever implemented truncated Name
 Service packets.  According to the RFCs, if you get a packet
 with the TC bit set, the you can negotiate the use of TCP.
 (Yes, that would be TCP on port 137.  No one ever implemented
 that, as far as I know.)  I believe that all implementations
 ignore this bit (though I've not tested to see what Windows
 does if this is set).  I would be very surprised to see this
 bit set.

RD - Recursion Desired.
 Only has meaning in request messages.  Used to indicate that
 the NBNS (WINS) server should be consulted.

 There is a lengthy discussion of all of this in section 3.4.3.2
 of my docs.  The RD bit is used heavily to distinguish between
 requests, queries, and demands.  In reply messages, however, its
 state is (probably) not significant at all.  My guess is that it
 will typically match the state of the RD bit in the request.

RA - Recursion Available.
 Should only be set in messages which come from the NBNS (WINS)
 server itself.  Note that a machine that is running WINS will
 also have a local name table.  If the reply comes from the
 local name table, then RA should be clear even if the same node
 is running the NBNS (WINS) server.

B  - Broadcast bit.
 Set when the packet is being sent as a broadcast.  Interesting
 to play with this bit, particularly in conjunction with the RD
 bit.

RCode - Result code.
 0 for success, else an error code.

Chris -)-

-- 
Christopher R. Hertel -)-   University of Minnesota
[EMAIL PROTECTED]  Networking and Telecommunications Services

Bug in samba 2.2.4??

[Ariel Mella]

im using the samba rpm for redhat from the ftp.samba.org
i upgrade it to a copuple installations that i have and in all the upgrades
i notice the problem with smbstatus.
i copy a backup of /var/cache/samba/*.tdb and all the things work now.
i see this problem in the mailing list before related to smbd not creating
the connections.tdb

Re: Bug found in smbclient

[Jay Ts]

Jay Ts wrote:
> 
> $ smbclient //server/share  -c "ls" -Tc
> added interface ip=172.16.1.3 bcast=172.16.1.255 nmask=255.255.255.0
> Segmentation fault

BTW, I should have mentioned that I'm running Samba 2.2.4
on Red Hat 7.1 here.

Jay Ts
[EMAIL PROTECTED]

Re: Adding Recursion Available bit to name_query and node_status_query

[Richard Sharpe]

On Mon, 13 May 2002, Gerald Carter wrote:

> On Mon, 13 May 2002, Richard Sharpe wrote:
> 
> > Hi,
> > 
> > I would like to add an additional flag to both name_query and 
> > node_status_query.
> > 
> > This flag will return the value of the Recursion Available flag in the 
> > nmb packet header.
> > 
> > I am sure some people will find it useful.
> > 
> > Any objections?
> 
> Nope.

Actually, I was thinking that it would be useful to return the entire set 
of flags. 

This might be better than returning just one of them.

Any further thoughts?

Regards
-
Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED], 
[EMAIL PROTECTED]

Autoreply: Welcome to my hometown

[filtered_mail]

A message submitted by you for handling by the NCSL email server
contained a prohibited file attachment type and was not delivered
to the intended recipient. We regret that security considerations
prohibit us from processing this transaction.

Re: compile samba-2.2.4 on Solaris 7

[Gerald Carter]

On Tue, 7 May 2002, Tian-xiong Lu wrote:

> Just tried compling samba-2.2.4 on a Solaris 7 box with two configure options:
> --with-libsmbclient  --with-acl-support
> 
> When compiling I got a warning message like this:
> In function "sys_readdir"
> lib/system.c 331 Warning: return from incompatible pointer type.

Am looking into it.

> Also, a lot of messages like this:
> gcc: unrecognized option '-KPIC'

Already fixed.





cheers, jerry
 -
 Hewlett-Packard http://www.hp.com
 SAMBA Team   http://www.samba.org
 --http://www.plainjoe.org
 "Sam's Teach Yourself Samba in 24 Hours" 2ed.  ISBN 0-672-32269-2
 --"I never saved anything for the swim back." Ethan Hawk in Gattaca--

Re: Problems building 2.2.4 on Solaris 7

[Gerald Carter]

On Mon, 6 May 2002, Nicholas Brealey wrote:

> Linking shared library bin/smbwrapper.so
> ld: fatal: library -lpopt: not found
> ld: fatal: File processing errors. No output written to bin/smbwrapper.so
> *** Error code 1
> make: Fatal error: Command failed for target `bin/smbwrapper.so'

> 
> should be changed to:
> 
> bin/smbwrapper.@SHLIBEXT@: $(PICOBJS) bin/.dummy 
> @echo Linking shared library $@ 
> @$(SHLD) @LDSHFLAGS@ -o $@ $(PICOBJS) $(LDFLAGS) $(LIBS) \ 
> @SONAMEFLAG@`basename $@` 

Got it.

> bin/pam_smbpass.@SHLIBEXT@: $(PAM_SMBPASS_OBJ) bin/.dummy 
> @echo Linking shared library $@ 
> $(SHLD) @LDSHFLAGS@ -symbolic -o $@ $(PAM_SMBPASS_OBJ) $(LDFLAGS) -lpam
> $(DYNEXP) $(LIBS) -lc \ 
> @SONAMEFLAG@`basename $@` 
> 
> also looks wrong because the Sun linker does not have a -symbolic option.

Already fixed this one.


> Linking bin/smbpasswd
> cc: Warning: illegal option -dc
> cc: Warning: illegal option -dp
> 
> Deleting the line
> 
> DYNEXP="-dc -dp" 
> from configure.in fixes the problem.

Got it.

Thanks for the heads up.



cheers, jerry
 -
 Hewlett-Packard http://www.hp.com
 SAMBA Team   http://www.samba.org
 --http://www.plainjoe.org
 "Sam's Teach Yourself Samba in 24 Hours" 2ed.  ISBN 0-672-32269-2
 --"I never saved anything for the swim back." Ethan Hawk in Gattaca--

Re: CVS 2_2 Fails to Compile on Solaris - rpc_parse/parse_spoolss.c

[Richard Bollinger]

Builds OK now.  Thanks!

- Original Message - 
From: "Gerald Carter" <[EMAIL PROTECTED]>
To: "Richard Bollinger" <[EMAIL PROTECTED]>
Cc: "Samba Technical" <[EMAIL PROTECTED]>
Sent: Monday, May 13, 2002 9:43 AM
Subject: Re: CVS 2_2 Fails to Compile on Solaris - rpc_parse/parse_spoolss.c


> On Mon, 13 May 2002, Richard Bollinger wrote:
> 
> 
> I'm about to check in a fix right now.  10 minutes. :-)
> 
> cheers, jerry
>  -

Re: Bug found in smbclient

[David Collier-Brown]

Jay Ts wrote:
> Just in case I'm the first to have noticed this:
> 
> $ smbclient //server/share  -c "ls" -Tc
> added interface ip=172.16.1.3 bcast=172.16.1.255 nmask=255.255.255.0
> Segmentation fault
> 
> This happened when out of curiosity, I wondered if
> it was possible to execute some smbclient commands,
> such as a cd to a subdirectory, before running the
> -Tc tar backup of it.
> 
> The "server" and "share" strings in the above command
> can be anything, and the command with -c can also be
> a "cd dir" command. Just a plain "cd" with no arguments
> resulted in odd behavior (the cd command ran, then
> smbclient hung, without doing a tar backup). I haven't
> tried any other smbclient commands with -c.

I just tried it with a Solaris system, 
and got exactly the same result.

truss -f -v uname said:
3267:   so_socket(2, 1, 0, "", 1)   = 3
3267:   ioctl(3, 0xC0086914, 0xFFBED478)= 0
3267:   ioctl(3, 0xC020690D, 0xFFBEB498)= 0
3267:   ioctl(3, 0xC0206911, 0xFFBEB498)= 0
3267:   ioctl(3, 0xC0206919, 0xFFBEB498)= 0
3267:   ioctl(3, 0xC020690D, 0xFFBEB478)= 0
3267:   ioctl(3, 0xC0206911, 0xFFBEB478)= 0
3267:   ioctl(3, 0xC0206919, 0xFFBEB478)= 0
3267:   close(3)= 0
3267:   write(1, " a d d e d   i n t e r f".., 72)  = 72
3267:   umask(0)= 02
3267:   umask(02)   = 0
3267:   uname(0xFFBED908)   = 1
3267:   sys=SunOS nod=elsbeth rel=5.8 ver=Generic_108528-09
mch=sun4u
3267:   Incurred fault #6, FLTBOUNDS  %pc = 0x000393DC
3267: siginfo: SIGSEGV SEGV_MAPERR addr=0x0D60
3267:   Received signal #11, SIGSEGV [default]
3267: siginfo: SIGSEGV SEGV_MAPERR addr=0x0D60
3267:   *** process killed ***


--dave
-- 
David Collier-Brown,   | Always do right. This will gratify 
Performance & Engineering  | some people and astonish the rest.
Americas Customer Engineering, |  -- Mark Twain
(905) 415-2849 | [EMAIL PROTECTED]

Re: CVS 2_2 Fails to Compile on Solaris - rpc_parse/parse_spoolss.c

[Gerald Carter]

On Mon, 13 May 2002, Richard Bollinger wrote:

> Looks like most recent change to rpc_parse/parse_spoolss.c includes a structure 
>initializer just a
> bit too fancy for Sun's compiler...  On sparc-sun-solaris2.8 using cc: WorkShop 
>Compilers 5.0
> 98/12/15 C 5.0, I get following build failures from today's 2_2 CVS:
> 
> Compiling rpc_parse/parse_spoolss.c
> "rpc_parse/parse_spoolss.c", line 559: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 560: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 561: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 562: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 563: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 564: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 565: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 566: non-constant initializer: op "+"
> "rpc_parse/parse_spoolss.c", line 2009: cannot recover from previous errors
> cc: acomp failed for rpc_parse/parse_spoolss.c
> *** Error code 2
> make: Fatal error: Command failed for target `rpc_parse/parse_spoolss.o'

I'm about to check in a fix right now.  10 minutes. :-)







cheers, jerry
 -
 Hewlett-Packard http://www.hp.com
 SAMBA Team   http://www.samba.org
 --http://www.plainjoe.org
 "Sam's Teach Yourself Samba in 24 Hours" 2ed.  ISBN 0-672-32269-2
 --"I never saved anything for the swim back." Ethan Hawk in Gattaca--

CVS 2_2 Fails to Compile on Solaris - rpc_parse/parse_spoolss.c

[Richard Bollinger]

Looks like most recent change to rpc_parse/parse_spoolss.c includes a structure 
initializer just a
bit too fancy for Sun's compiler...  On sparc-sun-solaris2.8 using cc: WorkShop 
Compilers 5.0
98/12/15 C 5.0, I get following build failures from today's 2_2 CVS:

Compiling rpc_parse/parse_spoolss.c
"rpc_parse/parse_spoolss.c", line 559: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 560: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 561: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 562: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 563: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 564: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 565: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 566: non-constant initializer: op "+"
"rpc_parse/parse_spoolss.c", line 2009: cannot recover from previous errors
cc: acomp failed for rpc_parse/parse_spoolss.c
*** Error code 2
make: Fatal error: Command failed for target `rpc_parse/parse_spoolss.o'

Thanks, Rich Bollinger

Re: 'getent group' only returns a few groups - Solaris 8

[Gerald Carter]

On Sat, 11 May 2002, Travis Freeland wrote:

> In order to get samba 2.2.4 working correctly under solaris (where we 
> rely heavily on NIS unix groups) I had to make a couple of changes to 
> username.c (basically take it back to the 2.0 style unix group checking 
> code .. and take out the winbind group check which was failing because 
> we dont run winbind).
> 
> I'm assuming that those who are testing samba and ok'ing a production 
> release are only testing a winbind implementation?

I've got a moderately sized Solaris site running the Samba 2.2 code
in user mode security (PDC for NT  & 9x).  I've not heard of any 
complaints from them.  All UNIX user accounts are maintained in 
NIS with local smbpasswd files.

What problem did you see?









cheers, jerry
 -
 Hewlett-Packard http://www.hp.com
 SAMBA Team   http://www.samba.org
 --http://www.plainjoe.org
 "Sam's Teach Yourself Samba in 24 Hours" 2ed.  ISBN 0-672-32269-2
 --"I never saved anything for the swim back." Ethan Hawk in Gattaca--

Re: Adding Recursion Available bit to name_query and node_status_query

[Gerald Carter]

On Mon, 13 May 2002, Richard Sharpe wrote:

> Hi,
> 
> I would like to add an additional flag to both name_query and 
> node_status_query.
> 
> This flag will return the value of the Recursion Available flag in the 
> nmb packet header.
> 
> I am sure some people will find it useful.
> 
> Any objections?

Nope.





jerry
 -
 Hewlett-Packard http://www.hp.com
 SAMBA Team   http://www.samba.org
 --http://www.plainjoe.org
 "Sam's Teach Yourself Samba in 24 Hours" 2ed.  ISBN 0-672-32269-2
 --"I never saved anything for the swim back." Ethan Hawk in Gattaca--

Re: Make error found in Samba 2.2.2

[Simo Sorce]

please can you avoid sending html formatted email and proprietary
unreadable documents on this list?

plain text mails and text documents are the best way if you want people
to take care of your requests.

Simo.

On Mon, 2002-05-13 at 13:15, MAKRO BADILLA TAMSI, Abigail wrote:
> >  <> 
> > Hi! I hope you could help me with the make fatal error which I received.
> > Attached is the document listing the errors found during the end of the
> > make command. Due to this the make install command also produces the same
> > error.
> > 
> > I installed Samba 2.2.2 in a server with Solaris 5.8 on Intel OS and NCR
> > S20 Pentium III-450 MHz server. 
> > 
> > I hope you can help me. Thank you very much. 
> > 
> > ;; 
> > Abigail B. Tamsi 
> > System Administrator, Information Technology Dept. 
> > Pilipinas Makro, Inc. 
> > 
> > Direct Line: (632) 876-3188 
> > Trunk Line: (632) 842-6771 
> > Fax No.: (632) 842-0761 
> > Email: [EMAIL PROTECTED] 
> > ;; 
> > 
> > 
> > 
-- 
Simo Sorce - [EMAIL PROTECTED]
Xsec s.r.l.
via Durando 10 Ed. G - 20158 - Milano
tel. +39 02 2399 7130 - fax: +39 02 700 442 399



signature.asc
Description: This is a digitally signed message part

Re: winbind UID, GID assignment

[Andrew Bartlett]

Mike Gerdts wrote:
> 
> I have just started testing of winbindd, to make it so that NT users may
> or may not have UNIX accounts.  I have run into two problems.  I am
> fishing for solutions that work for me and others.
> 
> 1.  UNIX UID/GID assigment
> 
> The incremental UNIX UID and GID assigment method seems great for
> standalone samba servers or standalone Unix workstations.  When you
> introduce NFS it becomes a mess.  To address this, I have hacked up
> winbindd to get non-cached UIDs from an Oracle database.  Thus, our
> Samba server no longer needs to use NIS, Unix accounts do not need to be
> created in advance, and if a Unix account is created at some time, the
> UID is consistent with the UID that the user will get at some time.
> 
> Would it be useful to make the UID/GID miss code allow for a generic
> plugin?  I am thinking of something along the lines of
> 
> BOOL local_lookup_uid(const char *domain, const char *user, uid_t *uid)
> BOOL local_lookup_gid(const char *domain, const char *user, gid_t *gid)

The username and domain may not be known at the time a mapping is
required.  Thats the easy bit - we might not even know if it is a uid or
gid!

There is more to this whole mess than meets the eye, but yes, we do need
to allow an aribtary SID->uid/gid, uid/gid -> SID backend system.

Andrew Bartlett

-- 
Andrew Bartlett [EMAIL PROTECTED]
Manager, Authentication Subsystems, Samba Team  [EMAIL PROTECTED]
Student Network Administrator, Hawker College   [EMAIL PROTECTED]
http://samba.org http://build.samba.org http://hawkerc.net

(no subject)

[Raymundo, Joseph]

>  We're using Redhat Linux 6.2 and Samba 2.0.7 for our file server.  It is driven 
>by an Intel Pentium III 900Mhz processor, 128MB PC133 DIMM, and RAID1 (40GB 5400rpm 
>IDE).  We have an average of 30-40 active connections.  Users workstation is mostly 
>Intel Pentium III 900Mhz with 128MB memory and connected via 10mbps Ethernet.
>
File access is not that heavy, but the files can be quite large, 100Kb to 30Mb.  A 
software being used by production directly accessed hi-res images in TIFF format or 
.tif file.  We also have postscript files that are being extracted from the server to 
a local PC for distilling.  The average size is about 100-200MB and sometimes goes up 
to 1.5GB.
>
 > Users are complaining that the server is going very slow whenever a large file 
is taken out or during peak hours where a lot of users are accessing the server all at 
the same time.  Based on the result of the "top" command, I have only 3.5 to 3MB left 
on my memory.  What I'm doing is just restart the samba and everything goes back to 
normal.
>
 > The problem is that this is happening almost once or at least twice everyday.  
We have another file server with Redhat 7.2 and Samba 2.2.1 same processor, but have a 
RAID5, 5400rpm IDE hard disk and 256MB DIMM PC133 memory.  This too is having the same 
problems as the first one.
>
 > Is 128MB enough for our file server or is there something wrong with the 
configuration of our Linux? Is RAID really adding that much load to the CPU usage, I/O 
operation and memory considering that the disks are controlled by the motherboard 
and/or processor? Do you have any hardware recommendations or softwares needed to be 
install on our samba files server?

Regards,

Joseph S. Raymundo
SPS SYSTEMS ENGINEER
SPI TECHNOLOGIES INC.
(632) 855-8756
(632) 853-2773/4 (fax)
URL : http:www.spitech.com/  
Email : mailto:[EMAIL PROTECTED]  

Make error found in Samba 2.2.2

[MAKRO BADILLA TAMSI, Abigail]

Title: Make error found in Samba 2.2.2





 <> 
Hi! I hope you could help me with the make fatal error which I received. Attached is the document listing the errors found during the end of the make command. Due to this the make install command also produces the same error.

I installed Samba 2.2.2 in a server with Solaris 5.8 on Intel OS and NCR S20 Pentium III-450 MHz server. 


I hope you can help me. Thank you very much. 


;; 
Abigail B. Tamsi 
System Administrator, Information Technology Dept. 
Pilipinas Makro, Inc. 


Direct Line: (632) 876-3188 
Trunk Line: (632) 842-6771 
Fax No.: (632) 842-0761 
Email: [EMAIL PROTECTED] 
;; 







samba error.doc
Description: MS-Word document

[È«º¸]³õÄ¥ ¼ö ¾ø´Â Âù½º 100¼ÛÀÌ ¹«·á ²É¹è´Þ

[ÀÎÆÄÁ¸]

Re: compile samba-2.2.4 on Solaris 7

[Tian-xiong Lu]

Yes, I am using 2.2.4 that I got from ftp site a week ago and that was
released May 2, 2002.
Will get the latest CVS.

Tian.
- Original Message -
From: "Richard Sharpe" <[EMAIL PROTECTED]>
To: "Tian-xiong Lu" <[EMAIL PROTECTED]>
Cc: "Gerald Carter" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Monday, May 13, 2002 5:59 PM
Subject: Re: compile samba-2.2.4 on Solaris 7


> On Mon, 13 May 2002, Tian-xiong Lu wrote:
>
> > Hi Richard,
> >
> > Tried, but both config.cache and autoconf do not exit in the source
> > directory.
>
> That means you do not have autoconf installed. Was it 2.2.4 you were
> playing with or the latest CVS?
>
> If it is 2.2.4, then there will be no changes, because Jerry's changes
> were applied to CVS and are not in 2.2.4.
>
> > Tian.
> > - Original Message -
> > From: "Richard Sharpe" <[EMAIL PROTECTED]>
> > To: "Tian-xiong Lu" <[EMAIL PROTECTED]>
> > Cc: "Gerald Carter" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
> > Sent: Monday, May 13, 2002 5:07 PM
> > Subject: Re: compile samba-2.2.4 on Solaris 7
> >
> >
> > > On Mon, 13 May 2002, Tian-xiong Lu wrote:
> > >
> > > > Applied patches in
ftp://ftp.samba.org:/pub/jerry/samba-2.2.4-patches/
> > > > and re-compiled.
> > > > Warning "nsswitch/winbind_nss_config.h:135: warning: `ZERO_STRUCTP'
> > > > redefined"
> > > > disappeared. But I still got -KPIC rather than -fPIC, so that had to
> > > > manually edit Makefile.
> > > > Seems the patch for configure.in did not work.
> > >
> > > It is possible that configure has not been rebuild from configure.in.
> > >
> > > Can you try something for us:
> > >
> > >   rm config.cache
> > >   autoconf
> > >   ./configure
> > >   make
> > >
> > > --
> > > Regards
> > > -
> > > Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED],
> > > [EMAIL PROTECTED]
> > >
> >
>
> --
> Regards
> -
> Richard Sharpe, [EMAIL PROTECTED], [EMAIL PROTECTED],
> [EMAIL PROTECTED]
>