Re: mac mini build broken was Re: [SCM] Samba Shared Repository - branch master updated

2010-12-02 Thread Love Hörnquist Åstrand
Matthieu,

SLIST_ENTRY should be defined by mechqueue.h or sys/queue.h, without cpp 
output it hard to determine what went wrong.

Love


2 dec 2010 kl. 03.45 skrev Matthieu Patou:

 Hi Andrews,
 
 I highly suspect the new import of heimdal to be the root cause of the 
 breakage on mac mini can one of you look on it ?
 
 Thanks.
 
 Matthieu.
 On 01/12/2010 09:49, Andrew Tridgell wrote:
 The branch, master has been updated
via  b7172e7 s4-drs: cope with invalid NTDS DNs from DsReplicaInfo()
via  00ecbdb wintest: cope with w2k3 form of dcdiag output
via  06fd5b7 wintest Move stopping of BIND into a new step
via  c5bea98 s4:heimdal: import lorikeet-heimdal-201012010201 (commit 
 81fe27bcc0148d410ca4617f8759b9df1a5e935c)
   from  9c84f98 wintest: make command matching case insensitive by 
 default
 
 http://gitweb.samba.org/?p=samba.git;a=shortlog;h=master
 
 
 - Log -
 commit b7172e7a71b152be687fe5045565c8cd99a73a18
 Author: Andrew Tridgelltri...@samba.org
 Date:   Wed Dec 1 16:40:17 2010 +1100
 
 s4-drs: cope with invalid NTDS DNs from DsReplicaInfo()
 
 w2k3 sometimes returns a deleted DN
 
 Autobuild-User: Andrew Tridgelltri...@samba.org
 Autobuild-Date: Wed Dec  1 07:48:19 CET 2010 on sn-devel-104
 
 commit 00ecbdbbd13ba191400c6f4185df2dd8e72d1459
 Author: Andrew Tridgelltri...@samba.org
 Date:   Wed Dec 1 16:34:16 2010 +1100
 
 wintest: cope with w2k3 form of dcdiag output
 
 commit 06fd5b70037728800cfeb2d1989ad8e851f604df
 Author: Andrew Bartlettabart...@samba.org
 Date:   Wed Dec 1 15:34:19 2010 +1100
 
 wintest Move stopping of BIND into a new step
 
 We must run this early, to ensure that BIND isn't alive to write to
 the zone file after provision has cleaned it up.
 
 Andrew Bartlett
 
 commit c5bea98ddb2f7967df572160f639da3cba381a87
 Author: Andrew Bartlettabart...@samba.org
 Date:   Mon Nov 29 11:24:08 2010 +1100
 
 s4:heimdal: import lorikeet-heimdal-201012010201 (commit 
 81fe27bcc0148d410ca4617f8759b9df1a5e935c)
 
 ---
 
 Summary of changes:
  source4/heimdal/base/baselocl.h   |   52 +-
  source4/heimdal/base/heimbase.c   |   15 +-
  source4/heimdal/cf/make-proto.pl  |1 +
  source4/heimdal/kdc/default_config.c  |3 +-
  source4/heimdal/kdc/kaserver.c|  955 
 
  source4/heimdal/kdc/misc.c|   21 +-
  source4/heimdal/kdc/process.c |   77 --
  source4/heimdal/kuser/kinit.c |   14 +-
  source4/heimdal/kuser/kuser_locl.h|3 +
  source4/heimdal/lib/asn1/asn1parse.c  |  724 +
  source4/heimdal/lib/asn1/asn1parse.y  |4 +
  source4/heimdal/lib/asn1/gen_template.c   |4 +-
  source4/heimdal/lib/com_err/com_err.h |   20 -
  source4/heimdal/lib/com_err/com_right.h   |   35 +-
  source4/heimdal/lib/com_err/error.c   |8 +-
  source4/heimdal/lib/com_err/lex.c |   83 +--
  source4/heimdal/lib/com_err/lex.h |2 +-
  source4/heimdal/lib/com_err/lex.l |5 +-
  source4/heimdal/lib/com_err/parse.c   |  265 ---
  source4/heimdal/lib/com_err/parse.y   |5 +-
  source4/heimdal/lib/gssapi/gssapi/gssapi.h|   85 ++-
  source4/heimdal/lib/gssapi/gssapi/gssapi_krb5.h   |   73 --
  source4/heimdal/lib/gssapi/gssapi/gssapi_oid.h|  231 +
  source4/heimdal/lib/gssapi/gssapi_mech.h  |   80 ++-
  source4/heimdal/lib/gssapi/krb5/acquire_cred.c|   14 +-
  source4/heimdal/lib/gssapi/krb5/external.c|  238 +++---
  source4/heimdal/lib/gssapi/krb5/set_cred_option.c |9 -
  source4/heimdal/lib/gssapi/mech/gss_mech_switch.c |2 +-
  source4/heimdal/lib/gssapi/mech/gss_mo.c  |  464 ++
  source4/heimdal/lib/gssapi/mech/gss_oid.c |  253 ++
  source4/heimdal/lib/gssapi/mech/gss_oid_equal.c   |2 +-
  source4/heimdal/lib/gssapi/mech/gss_oid_to_str.c  |   31 +
  source4/heimdal/lib/gssapi/mech/gss_wrap.c|   15 +
  source4/heimdal/lib/gssapi/mech/mech_locl.h   |   15 +-
  source4/heimdal/lib/gssapi/spnego/external.c  |   49 +-
  source4/heimdal/lib/hcrypto/rsa-ltm.c |3 +-
  source4/heimdal/lib/hcrypto/validate.c|1 -
  source4/heimdal/lib/hdb/db.c  |2 +-
  source4/heimdal/lib/hdb/hdb-keytab.c  |   10 +-
  source4/heimdal/lib/hdb/hdb.c |2 +-
  source4/heimdal/lib/hdb/hdb.h |   22 +-
  source4/heimdal/lib/hdb/keytab.c  |   15 +-
  source4/heimdal/lib/hdb/ndbm.c|2 +-
  source4/heimdal/lib/hx509/sel-gram.c  |  248 ---
  source4

Re: [Samba] Broken support for Smart Card Logon in Windows 2003 and XP

2010-10-21 Thread Love Hörnquist Åstrand

17 okt 2010 kl. 20.31 skrev Николай Домуховский:

 2010/10/7 Love Hörnquist Åstrand l...@kth.se:
 
 6 okt 2010 kl. 02:49 skrev Michael Wood:
 
 hx509_cms_create_signed function and
 
 make sigctx.cmsidflag always equal CMS_ID_NAME)
 
 I think this failed because you are looking at enveloped data and not signed
 data. try patching fill_CMSIdentifier() in hx509_cms_envelope_1() instead.
 Love
 
 
 Thanks, Love.
 I've tried patching hx509_cms_ebvelope_1() but it didn't help.
 But now, I'm think, I've found real issue:
 XP box include in KRB5_AS_REQ only one supported digest algorithm:
 md5withRSAEncryption (1.2.840.113549.1.1.4) (and this is only
 supported algorithm for XP, 2000 and 2003 - this is written in secrion
 2.2 of MS-PKCA).
 But response from Samba (I found a way to decrypt it!!!) contains
 digital signature made with sha512WithRSAEncryptions (in fact it is
 rather hard to understand openssl ans1parse output, but fact that
 there is no md5withRSAEncryption signature). So it looks like some bug
 in Heimdal code - I will investigate it further and try to locate
 exact place, where wrong signature formed, but maybe you already know
 answer...
 
 
 P.S. If you need I can send trafic capture files and decrypted KDC
 answers (both form Windows DC and from Samba).


You can probably change the code in kdc/pkinit.c around 870 that sets up the 
supported cms types it will use,

if you use hx509_signature_rsa_with_md5() and hx509_signature_md5() instead of 
SHA1 it might work for you.

Love


-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/options/samba

Re: [Samba] Broken support for Smart Card Logon in Windows 2003 and XP

2010-10-18 Thread Love Hörnquist Åstrand

6 okt 2010 kl. 02:49 skrev Michael Wood:

 hx509_cms_create_signed function and
 make sigctx.cmsidflag always equal CMS_ID_NAME)

I think this failed because you are looking at enveloped data and not signed 
data. try patching fill_CMSIdentifier() in hx509_cms_envelope_1() instead.

Love


-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/options/samba

Re: svn commit: lorikeet r583 - in trunk/heimdal/lib/hdb: .

2006-10-20 Thread Love Hörnquist Åstrand

+ret = (*db-hdb_fetch)(context, db, principal,
+  HDB_F_DECRYPT||


is that really correct using '||'?


No, that correct, its wrong :)

Love




[Samba] Remote Downlevel Document Hangs in XP Pro Print Queue

2005-03-23 Thread Flash Love
I used SAMBA, CUPS, hplip, and linuxprinting.org documentation to made all 
configuration changes necessary to successfully print from FC3 print server 
using samba-3.0.10-1.fc3 for a HP PSC 1350 SAMBA windoww share on a XPPro SP2 
PC. 

I created the FC3 print queue using the CUPS web interface-Device=Windows
Printer via 
SAMBA-URI=smb/DOMAIN/HOST/Sharename-Mode/Driver=raw-Model/Driver=raw. 

Under CUPS only configuration, when testing, the test file enters the XP print 
queue as remote downlevel document, the printer goes into maintenance mode 
and the print queue hangs while displaying refreshing intermittently. The 
only way to clear the document is to turn the printer off and reboot the 
XPPro PC.

Under CUPS+HPLIP configuration using 
Mode/Driver=HP-Model/Driver=HP_DeskJet3320 and 
URI=hp:/net/deskjet_3320?ip=XP Host IP address or URI=socket://XP Host:9100 
the system  complaints Open device failed.

How do I resolve the SAMBA Window Share printing problem?

Flash
-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/listinfo/samba


Re: Compiler waring with heimdal-0.5.1

2003-03-03 Thread Love
[EMAIL PROTECTED] writes:

 Anybody else seen this ?

 This is brokenness in the Heimdal header files, not in Samba.
 The Heimdal developers need to compile with more warnings set
 in gcc.

What more then 

-Wall -Wmissing-prototypes -Wpointer-arith -Wbad-function-cast
-Wmissing-declarations -Wnested-externs

?

It a mistake to make it `open', it originates from RFC2744 and was copied
verbose from the spec. I'll make sure it get fixed to next release of
heimdal.

Love


Re: auto-detecting krb versions for build

2003-03-01 Thread Love
Jim McDonough [EMAIL PROTECTED] writes:

 As some of you may have noticed, the UL builds on the farm don't work.  The
 culprit is the pre- 0.5 heimdal that is installed, as it doesn't have
 AP_OPTS_USE_SUBKEY, which is needed to do password changes.  Unless someone
 knows how to accomplish password changes to a win KDC without it, we need
 to be able to detect the package and release of kerberos installed.
 Heimdal before 0.5 doesn't have it.  I'm not sure how far you have to go
 back in MIT to not find it.

Heimdal before 0.5 didn't have AP_OPTS_USE_SUBKEY, however I think it
allways generated and used a subkey anyway, so defining it to 0 should make
it work.

 krb5-config takes a --version option and prints something like:
 heimdal 0.4d
 $Id: krb5-config.in,v 1.8 2001/01/29 06:56:51 assar Exp $

 or in the case of MIT:
 Kerberos 5 release 0.0.0

 Yes, 0.0.0, but I'm using a dev version.   In any case, I'd like to take
 that output, and if it has heimdal, make a decision based on the
 number...since I'm not a wiz (more like a whiz) at these sorts of string
 cutting/comparing in shells...anyone want to tell me how to do it?

`expr expr1 : expr2` or case statement is usully what I use.

Love


Re: samba + w2k + kerberos + trusted realm

2003-03-01 Thread Love
Luke Howard [EMAIL PROTECTED] writes:

What is it that limit samba to root ? When I use samba with afs beeing root
will certenly not help samba access files, what else do samba need.

 SAMBA does need to bind to privileged ports.

Ok, sure. Now concentrate on the other issues :)

Its not that simple as my patch it since samba breaks itself after a couple
of hours and the key seem to change. Dunno what key, if its the in memory
key or the key in the (ad/kdc) database.

Love


Re: Samba as a gateway to OpenAFS

2002-05-29 Thread Love

Andrew Bartlett [EMAIL PROTECTED] writes:

 1. Get rid of AFS's need for plaintext passwords.
[]
  Ah, of course credential forwarding/proxying would be a requirement for
  making this work without giving the gateway special privileges; I'd
  completely overlooked that.  I'm afraid I don't know the answer, though.
  Perhaps someone currently doing Samba 3.0 work has run into this and can
  say?
 
 I see no reason why this would not be possible.  We would need to do a
 little bit of work on the smbd side of things, but credential forwarding
 is pretty standard.  This assumes either a AD domain, or Samba modified
 to correctlly function with krb5 but without AD (which also implies
 windows clients joined to such a domain).

So, so how do you tell the client to forward creds to the fileserver, and
can you chose want creds you want to forward ?
 
Love




Re: Samba as a gateway to OpenAFS

2002-05-29 Thread Love

Andrew Bartlett [EMAIL PROTECTED] writes:
   I see no reason why this would not be possible.  We would need to do a
   little bit of work on the smbd side of things, but credential forwarding
   is pretty standard.  This assumes either a AD domain, or Samba modified
   to correctlly function with krb5 but without AD (which also implies
   windows clients joined to such a domain).
  
  So, so how do you tell the client to forward creds to the fileserver, and
  can you chose want creds you want to forward ?
 
 This assumes krb5, where this is all quite standard.  

Sure, but the question is if there any provision make for this within
SMB/CIFS ? Doing it out of band make it quite a lot harder, since the all
(smb) clients need to have a program that forward their creds to the
server.

Love




Re: Samba as a gateway to OpenAFS

2002-05-28 Thread Love

Steve Langasek [EMAIL PROTECTED] writes:

  To re-phrase, I am trying to:
 
  1. Get rid of AFS's need for plaintext passwords.
  2. Establish a registration mechanism for new samba users and those that
 change their passwords.
  3. Turn on encrypted password support.
 
  The patches that will give you AFS support with plaintext turned on can be
  found at www.ualberta.ca/~sholstea
 
  The routines that will allow me to turn on encrypted pasword support for
  AFS users are still under developement.
 
 Unfortunately, my interest in this is strictly academic, since my
 current employer doesn't use AFS and won't any time soon, either.
 Nevertheless, I'm quite pleased to see development in this area.  I
 assume that as a large university, you have a need for supporting old
 Windows clients that precludes a pure Active Directory+AFS style of
 integration (NT password hashes only)?

Is there credtioal forwardation i smb/cifs or is there a need to send that
out of band ?
 
The solution I've been using is giving the samba gateway priveliges into
the afs-space (by storing the afs KeyFile on the gateway and cooking cred's
on the fly).

 I'd be tickled pink if someone were actually implementing a Samba-AFS
 gateway using pure Kerberos 5, but AIUI there's still quite a lot of
 work involved in getting OpenAFS to use /anything/ other than DES.

Its not talking des, its using fcrypt. And yes there is work in progress to
make it talk something better then fcrypt, and no, its not that hard.

Love