Bug#473796: [Pkg-openldap-devel] Bug#473796: Bug#473796: TLS fails completely

2008-07-27 Thread Steve Langasek
On Sun, Jul 27, 2008 at 10:22:48PM +0200, Petter Reinholdtsen wrote:
 tags 473796 + patch
 thanks

 Here is what I believe is the correct patch to solve this problem.  It
 is from upstream.  I'm testing this patch in Debian Edu now.

Did you see that this bug is already tagged 'pending'?

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developerhttp://www.debian.org/
[EMAIL PROTECTED] [EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#473796: [Pkg-openldap-devel] Bug#473796: Bug#473796: TLS fails completely

2008-06-30 Thread Quanah Gibson-Mount
--On Sunday, June 29, 2008 1:12 AM -0700 Steve Langasek [EMAIL PROTECTED] 
wrote:



I.e., the TLS SSF is 32.  So no value  32 will ever work.


This suggests to me that the SSF values haven't been properly normalized
for GNUtls.  Doesn't the 128 mean, roughly, a symmetric cipher with
keylength of 128?  Surely the user's TLSCipherSuite
TLS_RSA_AES_256_CBC_SHA1 should satisfy this?


The GnuTLS library is what reports back the SSF value.  It may be 
worthwhile to discuss with them why their values are so low.


--Quanah



--

Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc

Zimbra ::  the leader in open source messaging and collaboration



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#473796: [Pkg-openldap-devel] Bug#473796: Bug#473796: TLS fails completely

2008-06-30 Thread Quanah Gibson-Mount
--On Monday, June 30, 2008 2:26 PM -0700 Quanah Gibson-Mount 
[EMAIL PROTECTED] wrote:



This suggests to me that the SSF values haven't been properly normalized
for GNUtls.  Doesn't the 128 mean, roughly, a symmetric cipher with
keylength of 128?  Surely the user's TLSCipherSuite
TLS_RSA_AES_256_CBC_SHA1 should satisfy this?


The GnuTLS library is what reports back the SSF value.  It may be
worthwhile to discuss with them why their values are so low.


Scratch that, it is an OpenLDAP conversion bug.  I'll file an ITS on it
and report back.


http://www.openldap.org/its/index.cgi/?findid=5585

--Quanah

--

Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc

Zimbra ::  the leader in open source messaging and collaboration



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#473796: [Pkg-openldap-devel] Bug#473796: Bug#473796: TLS fails completely

2008-06-30 Thread Quanah Gibson-Mount
--On Monday, June 30, 2008 2:22 PM -0700 Quanah Gibson-Mount 
[EMAIL PROTECTED] wrote:



--On Sunday, June 29, 2008 1:12 AM -0700 Steve Langasek
[EMAIL PROTECTED] wrote:


I.e., the TLS SSF is 32.  So no value  32 will ever work.


This suggests to me that the SSF values haven't been properly normalized
for GNUtls.  Doesn't the 128 mean, roughly, a symmetric cipher with
keylength of 128?  Surely the user's TLSCipherSuite
TLS_RSA_AES_256_CBC_SHA1 should satisfy this?


The GnuTLS library is what reports back the SSF value.  It may be
worthwhile to discuss with them why their values are so low.


Scratch that, it is an OpenLDAP conversion bug.  I'll file an ITS on it and 
report back.


--Quanah

--

Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc

Zimbra ::  the leader in open source messaging and collaboration



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#473796: [Pkg-openldap-devel] Bug#473796: Bug#473796: TLS fails completely

2008-06-30 Thread Quanah Gibson-Mount
--On Monday, June 30, 2008 2:29 PM -0700 Quanah Gibson-Mount 
[EMAIL PROTECTED] wrote:



--On Monday, June 30, 2008 2:26 PM -0700 Quanah Gibson-Mount
[EMAIL PROTECTED] wrote:


This suggests to me that the SSF values haven't been properly
normalized for GNUtls.  Doesn't the 128 mean, roughly, a symmetric
cipher with keylength of 128?  Surely the user's TLSCipherSuite
TLS_RSA_AES_256_CBC_SHA1 should satisfy this?


The GnuTLS library is what reports back the SSF value.  It may be
worthwhile to discuss with them why their values are so low.


Scratch that, it is an OpenLDAP conversion bug.  I'll file an ITS on it
and report back.


http://www.openldap.org/its/index.cgi/?findid=5585


Fixed:

Update of /repo/OpenLDAP/pkg/ldap/libraries/libldap

Modified Files:
tls.c  1.160 - 1.161

Log Message:
ITS#5585 GnuTLS key strength is in bytes, we expected bits


CVS Web URLs:
 http://www.openldap.org/devel/cvsweb.cgi/libraries/libldap/
   http://www.openldap.org/devel/cvsweb.cgi/libraries/libldap/tls.c

--

Quanah Gibson-Mount
Principal Software Engineer
Zimbra, Inc

Zimbra ::  the leader in open source messaging and collaboration



--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Bug#473796: [Pkg-openldap-devel] Bug#473796: Bug#473796: TLS fails completely

2008-06-29 Thread Steve Langasek
On Thu, Apr 03, 2008 at 12:28:56PM -0700, Quanah Gibson-Mount wrote:
  ok, more testing, more news:

  now with the slapd 2.4.7 package (with gnutls) this seems to force
  client-certs, too. a TLS query without client-cert won't work - but
  commenting the 'security' line out results in working TLS and working
  non-TLS queries.

  The default behavior when TLS is enabled is TLSVerifyClient never;
  2.4.7 did have a bug related to this, but this was resolved in the
  2.4.7-5 package.

  well it seems to me like with gnutls the 'security tls=' value controls
  the tls reqirements, TLSVerifyClient is (more or less?) ignored. but i
  could be missing something ofc...

  all queries done with a server cert and without a client cert:

  security tls=128
  TLSVerifyClient never

  ldapsearch  fails (TLS confidentiality required)
  ldapsearch -ZZ  fails (stronger TLS confidentiality required)

 This will always fail as long as the keystrength of the cert in question is 
 so low.  It states quite clearly in your log:

 conn=0 fd=12 TLS established tls_ssf=32 ssf=32

 I.e., the TLS SSF is 32.  So no value  32 will ever work.

This suggests to me that the SSF values haven't been properly normalized for
GNUtls.  Doesn't the 128 mean, roughly, a symmetric cipher with keylength
of 128?  Surely the user's TLSCipherSuite TLS_RSA_AES_256_CBC_SHA1 should
satisfy this?

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developerhttp://www.debian.org/
[EMAIL PROTECTED] [EMAIL PROTECTED]



-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]