Andrew Bartlett | 1 Feb 06:34 2011
Picon

Re: Proposal: Split libtalloc, libtdb, libtevent and libldb into a separate upstream project

On Tue, 2011-01-11 at 07:37 -0500, Stephen Gallagher wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 01/10/2011 08:03 PM, tridge <at> samba.org wrote:
> > Hi Simo,
> > 
> >  > there are build bugs that need fixing in talloc and tevent before we can
> >  > release.
> > 
> > what sort of build bugs?
> > 
> 
> 1) When building talloc with pytalloc, the build doesn't drop the shared
> object into the correct python_sitearch location:
> https://bugzilla.samba.org/show_bug.cgi?id=7905
> 
> 2) If you build talloc without pytalloc, you can't build libldb, because
> the configure script fails to locate pytalloc (even when using
> - --disable-python) I don't think a bug was filed for this yet; Simo was
> narrowing it down.

As I mention in that bug, the fix for this has caused the current issues
with Samba4 on 64 bit non-debian platforms.  

I hope to find a way to force the install into just lib64 in the
interim. 

Andrew Bartlett

(Continue reading)

Andrew Bartlett | 1 Feb 06:47 2011
Picon

Re: Patch for LDAPS on GC

On Thu, 2011-01-20 at 19:43 +1030, William Brown wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> 
> On 20/01/2011, at 19:40, Stefan (metze) Metzmacher wrote:
> 
> > Hi William,
> > 
> >> Here is the patches to enable LDAPS on 3269 for global catalog. I have tested this as functional also.
> > 
> > Please squash them into one commit, it makes no sense to add broken patches
> > and directly fix them in the next commit.
> 
> My mistake. Andrew Bartlett just helped me with this since I am not very experienced with git yet.

William,

Can you re-mail the git format-patch output as an attachement?  I
thought I had applied it, but it seems instead I only got as far as
trying to fight git am.  

Andrew Bartlett

--

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org
Samba Developer, Cisco Inc.

(Continue reading)

Matthias Dieter Wallnöfer | 1 Feb 08:30 2011
Picon

Re: Patch for LDAPS on GC

Andrew,

I've managed to fix the patch provided on the list. I can apply it if 
you are fine.

Cheers,
Matthias

Andrew Bartlett wrote:
> On Thu, 2011-01-20 at 19:43 +1030, William Brown wrote:
>    
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>>
>> On 20/01/2011, at 19:40, Stefan (metze) Metzmacher wrote:
>>
>>      
>>> Hi William,
>>>
>>>        
>>>> Here is the patches to enable LDAPS on 3269 for global catalog. I have tested this as functional also.
>>>>          
>>> Please squash them into one commit, it makes no sense to add broken patches
>>> and directly fix them in the next commit.
>>>        
>> My mistake. Andrew Bartlett just helped me with this since I am not very experienced with git yet.
>>      
>
> William,
(Continue reading)

Andrew Bartlett | 1 Feb 09:09 2011
Picon

Re: Patch for LDAPS on GC

On Tue, 2011-02-01 at 08:30 +0100, Matthias Dieter Wallnöfer wrote:
> Andrew,
> 
> I've managed to fix the patch provided on the list. I can apply it if 
> you are fine.

That's the idea :-)

--

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org
Samba Developer, Cisco Inc.

Kamen Mazdrashki | 1 Feb 11:48 2011
Picon

Re: [PATCH] Add LoadParm optional parameter for Py_Credentials class constructor

On Thu, Jan 27, 2011 at 10:09, Andrew Bartlett <abartlet <at> samba.org> wrote:
> On Wed, 2011-01-26 at 22:04 +0200, Kamen Mazdrashki wrote:
>> Hi Jelmer,
>>
>> Could you please take a look at following branch:
>> http://git.samba.org/?p=kamenim/samba.git;a=shortlog;h=refs/heads/py-review
>>
>> It is test to exploit a segfault when Py_Credentials object is used without
>> initializing 'domain' and 'workstation' members.
>>
>> After a short discussion with abartlet, he prefers the fix to be
>> implemented in python layer via adding LoadParm parameter
>> for Credentials constructor.
>> This way Py_Credentials instances will get the chance to
>> call cli_credentials_guess() during construction to try to populate
>> cli_credentials structure fields.
>> Basically, most common pattern was:
>>   creds = Credentials()
>>   creds.guess(lp)
>> which now turns into just:
>>   creds = Credentials(lp)
>
> This looks good to me.  Perhaps just put the segfault test at the end of
> the patch stream?
>
Thanks Andrew!
I've moved segfault test at the of the patch -
http://git.samba.org/?p=kamenim/samba.git;a=shortlog;h=refs/heads/py-review
I am waiting for a review from Jelmer in order to push it.

(Continue reading)

Andrew Bartlett | 1 Feb 12:10 2011
Picon

Failure joining Windows 2008R2 RODC to Samba4

Just a heads up.  In my wintest, I can't get Windows 2008R2 to join
Samba4 current GIT (plus heimdal update) as an RODC.

It fails to find a working RW DC.

I have captures and a keytab (attached), and am looking into it, but I
wanted to mention it in case someone else also found the same failure
(so I can eliminate local environmental issues). 

If anyone else wants to look, I would welcome it.  The error appears
around a set of searches over LDAP for DC objects. 

The error on the telnet connection is:
Validating environment and parameters...

A domain controller running Windows Server 2008 or Windows Server 2008
R2 could not be located in this domain. To install a read-only domain
controller, the domain must have a domain controller running Windows
Server 2008 or Windows Server 2008 R2.

Andrew Bartlett
--

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org
Samba Developer, Cisco Inc.

Attachment (secrets.keytab): application/octet-stream, 1700 bytes
Attachment (wintest-fail.cap.gz): application/x-gzip, 73 KiB
Jacky Lam | 1 Feb 02:42 2011

Zero-copy patch

Dear all,

    Does any one have the kernel patch for zero-copy from skb to IO buffer/cache? The performance of Samba using
splice is slowing down because of busy to copy the buffer from socket.

    Thanks.

Jacky

________________________________
IMPORTANT CONFIDENTIALITY NOTICE
This message and any attached documents contain information from ViXS Systems, Inc. and are confidential
and privileged and further subject to any confidentiality agreement between the parties. The
information is intended to be viewed only by the individual(s) or entity(ies) to whom the message is
addressed. If you are not the intended recipient, be aware that reading, disclosing, copying,
distributing or using the contents of this transmission is prohibited. Please notify us immediately if
you have received this transmission in error, and delete this message along with any attached files.

Volker Lendecke | 1 Feb 12:17 2011
Picon

Re: Zero-copy patch

On Mon, Jan 31, 2011 at 08:42:35PM -0500, Jacky Lam wrote:
> Does any one have the kernel patch for zero-copy from skb
> to IO buffer/cache? The performance of Samba using splice
> is slowing down because of busy to copy the buffer from
> socket.

Wait -- If I used 2 splice calls, one from TCP socket to a
pipe and the second from the pipe to a on-disk file it still
does a copy internally?

Volker

--

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen

Volker Lendecke | 1 Feb 12:30 2011
Picon

Re: Zero-copy patch

On Tue, Feb 01, 2011 at 12:17:35PM +0100, Volker Lendecke wrote:
> > Does any one have the kernel patch for zero-copy from skb
> > to IO buffer/cache? The performance of Samba using splice
> > is slowing down because of busy to copy the buffer from
> > socket.
> 
> Wait -- If I used 2 splice calls, one from TCP socket to a
> pipe and the second from the pipe to a on-disk file it still
> does a copy internally?

Hmmm. Found the following comment in relatively recent Linux
kernel source:

 *      - Destination page already exists in the address space and there
 *        are users of it. For that case we have no other option that
 *        copying the data. Tough luck.
 *      - Destination page already exists in the address space, but there
 *        are no users of it. Make sure it's uptodate, then drop it. Fall
 *        through to last case.
 *      - Destination page does not exist, we can add the pipe page to
 *        the page cache and avoid the copy.

So it seems that for normal samba file server use it should
avoid the copy.

Volker

--

-- 
SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
(Continue reading)

Stephen Gallagher | 1 Feb 13:03 2011
Picon

Re: 64 bit install Python


On 01/31/2011 06:51 AM, Andrew Bartlett wrote:
> On Sat, 2011-01-29 at 13:31 -0500, Scott Carradice wrote:
>> When installing samba on a 64 bit machine ( Opensuse 11.3 ) the python
>> portion seems to split up into lib and lib64.  Python seems to be able
>> to use one directory so my fix so far is just to copy all of the lib64
>> directory into the lib directory.  Is there a more correct solution at
>> this time?
> 
> The issue is that installmisc.sh in Samba4 rewrites:
> 
> sys.path.insert(0, "bin/python")
> into
> sys.path.insert(0, "/usr/local/samba/lib/python2.7/site-packages")
> 
> This means that the 64 bit files are now not found on a fresh install.  
> 
> We need to fix installmisc.sh to be written in python itself, or for it
> to be folded into the waf rules.  Then we can create a more sensible
> pattern match for fixing the scripts, that doesn't rely on lowest common
> denominator sed, and can insert both paths. 

One approach that could also work would be to have installmisc.sh do the
following:

PYTHON_SITELIB=`python -c "from distutils.sysconfig import
get_python_lib; print(get_python_lib())"`

PYTHON_SITEARCH=`python -c "from distutils.sysconfig import
get_python_lib; print(get_python_lib(1))"`
(Continue reading)


Gmane