gmane.network.samba.internals

Re: [PATCH] winbind: Print error code on connection error in ping_dc

Christian Ambach — 2013-05-25T19:50:51

Am 25.05.13 20:35, schrieb Christof Schmitt:

Pushed, thanks!

Cheers,
Christian

Re: [PATCH v6] vfs_glusterfs: Samba VFS module for glusterfs

Anand Avati — 2013-05-25T19:40:51


The patch does have pkg-config integration in configure.in. I'm not sure 
about the waf build or how it works, so I just copied waf changes from 
other patches, and may well be broken :( Any help in fixing the waf 
changes will be helpful!

Feel free to modify the patch to make it waf compatible, or suggest 
changes for me to resubmit -- as you find convenient.

Avati

Re: [Gluster-devel] [PATCH v6] vfs_glusterfs: Samba VFS module forglusterfs

Anand Avati — 2013-05-25T19:06:36

The dependency patches have just got merged.. so 
github.com/gluster/glusterfs master HEAD is good for testing. Let us 
know if you need help in getting RPMs built..

Avati

Re: Fw: [PROPOSAL] To retire autoconf for 4.1

Alexander Bokovoy — 2013-05-25T18:52:00

Majority of what we use is indeed autogenerated. However, we rely on
security and param, and few others which are manually coded. So there is
still amount of work to port to any changed module API.

I think actual issue is to find out what we use in our Python bindings
willingfully from Python module APIs and Python language features that
would change between 2 and 3. Strings are one part but most of other
APIs are hardly influencing complexity of maintaining the ports.
After all, we are not using CObject API so we don't need to migrate to
Capsule, and module initialization part is easy to rewrite.

Also, at Python level (as opposed to currently discussed Python C API) we
also don't need many of specific Python language features. We hardly are
beyond exceptions, iterators and dicts. Majority of new syntax is
backported as back as Python 2.6 which is what we rely on right now.

So I think the issue of maintaining support for both 2.6+ and 3.2+ is
rather question of having someone to come up and do actual work

Re: [PATCH] winbind: Print error code on connection error in ping_dc

Christof Schmitt — 2013-05-25T18:37:53

Yes, the patches are submitted under my personal copyright. I
resubmitted the patch with the "Signed-off-by" tag:
https://lists.samba.org/archive/samba-technical/2013-May/092689.html

[PATCH] winbind: Print error code on connection error in ping_dc

Christof Schmitt — 2013-05-25T18:35:11

For debugging, it is useful to include the error code in the message.

Signed-off-by: Christof Schmitt <christof.schmitt< at >us.ibm.com>
---
 source3/winbindd/winbindd_dual_srv.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/source3/winbindd/winbindd_dual_srv.c b/source3/winbindd/winbindd_dual_srv.c
index 744e251..e23d048 100644
--- a/source3/winbindd/winbindd_dual_srv.c
+++ b/source3/winbindd/winbindd_dual_srv.c
< at >< at > -690,7 +690,8 < at >< at > NTSTATUS _wbint_PingDc(struct pipes_struct *p, struct wbint_PingDc *r)
 status = cm_connect_netlogon(domain, &netlogon_pipe);
 reset_cm_connection_on_error(domain, status);
         if (!NT_STATUS_IS_OK(status)) {
-                DEBUG(3, ("could not open handle to NETLOGON pipe\n"));
+DEBUG(3, ("could not open handle to NETLOGON pipe: %s\n",
+  nt_errstr(status)));
 return status;
         }

Re: samba4 debian pkg build 4.0.7-dev - looking for updated debianpkg rules

C.J. Adams-Collier — 2013-05-25T16:06:04

git-buildpackage
On May 25, 2013 6:23 AM, "Horst Venzke-Fa Remsnet Ltd" <support< at >remsnet.de>
wrote:

Re: Fw: [PROPOSAL] To retire autoconf for 4.1

Jelmer Vernooij — 2013-05-25T15:11:22

See http://mail.python.org/pipermail/python-dev/2010-May/099971.html
It looks like there will be support for 2.7 at least until Mid 2016.

Yeah, I would expect the same - it's the default strategy for this kind of
transition. Ubuntu has plans to make Python3 the only Python on the desktop CD
for 14.04. https://wiki.ubuntu.com/Python/3

Cheers,

Jelmer

Fw: [PROPOSAL] To retire autoconf for 4.1

2013-05-25T14:24:21

If we go back to what triggered the discussion between Python 2 vs Python 
3

http://wiki.python.org/moin/Python2orPython3
"Python 2.x is the status quo, The final 2.x version 2.7 release is out, 
with a statement of extended support for this end-of-life release. "

Maybe you should ask the Python community when they expect to end the 
extended support for Python 2.7.

As for the Linux distributions, I expect they might decide to provide 
Python 3 instead Python 2 as their base Python package, but still provide 
Python 2.7 as an optional package until the Python community stop 
supporting it.
If they don't, they're not helping themself to keep their user base.


Best regards,

Yannick Bergeron
450 534-7711
yaberger< at >ca.ibm.com
Advisory IT Specialist

Never say never, say "it depends" / Ne jamais dire jamais, dites "ca 
dépend"

samba4 debian pkg build 4.0.7-dev - looking for updated debian pkgrules

Horst Venzke-Fa Remsnet Ltd — 2013-05-25T13:23:19

Hello Samba Fellows,

Got samba 4.0.7 RUN on arm6 ( raspberry PI ) , source via latest GIT .

~ #:/usr/src/GIT/samba4/samba-v4-0-test# bin/smbd --version
Version 4.0.7-DEVELOPERBUILD
~ #:/usr/src/GIT/samba4/samba-v4-0-test# uname -a
Linux pdc 3.6.11+ #456 PREEMPT Mon May 20 17:42:15 BST 2013 armv6l GNU/Linux

That seems to contains no debian pkg rules , thus i can´t yet build an deb pkg (yet).

Was read & follow out https://wiki.samba.org/index.php/Samba4/Debian , but end up with :

./debian/rules get-orig-source
dh get-orig-source --with python2
dh: Unknown sequence get-orig-source (choose from: binary binary-arch binary-indep build build-arch build-indep clean install install-arch install-indep)
make: *** [get-orig-source] Fehler 255

Looks like that the git://git.debian.org/pkg-samba/samba4 samba4-unstable not work with Current latest 4.0.7-development.

Deb pkg build on samba4 work seems allready in progress .. i.e http://lists.alioth.debian.org/pipermail/pkg-samba-maint/2013-May/014051.html
( have´nt

[PATCH 2/2] idl: Add DNS HINFO record support

Kai Blin — 2013-05-25T12:05:22

Not sure if we need this, bug samba.org specifies it, so this can cause ANY queries
against samba.org to fail.

Signed-off-by: Kai Blin <kai< at >samba.org>
---
 librpc/idl/dns.idl | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/librpc/idl/dns.idl b/librpc/idl/dns.idl
index 7d29278..28dddad 100644
--- a/librpc/idl/dns.idl
+++ b/librpc/idl/dns.idl
< at >< at > -151,6 +151,13 < at >< at > interface dns
 } dns_soa_record;
 
 typedef [public] struct {
+[value(strlen(cpu))] uint8 cpu_length;
+[charset(DOS)] uint8 cpu[cpu_length];
+[value(strlen(os))] uint8 os_length;
+[charset(DOS)] uint8 os[os_length];
+} dns_hinfo_record;
+
+typedef [public] struct {
 uint16     preference;
 dns_string exchange;
 } dns_mx_record;
< at >< at > -218,6 +225,7 < at >< at > interface dns
 [case(DNS_QTYPE_CNAME)] dns_string       cname_record;
 [case(DNS_QTYPE_SOA)]   dns_soa_record   soa_record;
 [case(DNS_QTYPE_PTR)]   dns_string       ptr_record;
+[case(DNS_QTYPE_HINFO)] dns_hinfo_record  hinfo_record;
 [case(DNS_QTYPE_MX)]    dns_mx_re

[PATCH 1/2] idl: Add DNS RP record support as per RFC 1183

Kai Blin — 2013-05-25T12:05:21

Not sure if we need this, bug samba.org specifies it, so this could cause ANY
queries against samba.org to fail.

Signed-off-by: Kai Blin <kai< at >samba.org>
---
 librpc/idl/dns.idl | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/librpc/idl/dns.idl b/librpc/idl/dns.idl
index 984f2b3..7d29278 100644
--- a/librpc/idl/dns.idl
+++ b/librpc/idl/dns.idl
< at >< at > -161,6 +161,12 < at >< at > interface dns
 } dns_txt_record;
 
 typedef [public] struct {
+dns_string mbox;
+dns_string txt;
+
+} dns_rp_record;
+
+typedef [public] struct {
 uint16     priority;
 uint16     weight;
 uint16     port;
< at >< at > -214,6 +220,7 < at >< at > interface dns
 [case(DNS_QTYPE_PTR)]   dns_string       ptr_record;
 [case(DNS_QTYPE_MX)]    dns_mx_record    mx_record;
 [case(DNS_QTYPE_TXT)]dns_txt_record   txt_record;
+[case(DNS_QTYPE_RP)]dns_rp_record    rp_record;
 [case(DNS_QTYPE_AAAA)]ipv6address ipv6_record;
 [case(DNS_QTYPE_SRV)]dns_srv_record srv_record;
 [case(DNS_QTYPE_TSIG)]dns_tsig_record tsig_record;

[PATCH 0/2] dns: Add support for more DNS RR types

Kai Blin — 2013-05-25T12:05:20

A bug reproducer I recently dealt with used dig < at >samba -t any samba.org as
reproducer. This failed on the internal server because the RP and HINFO RR
types were missing. This adds these types and makes dig -t any samba.org work.

Kai Blin (2):
  idl: Add DNS RP record support as per RFC 1183
  idl: Add DNS HINFO record support

 librpc/idl/dns.idl | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

Re: [PATCH v6] vfs_glusterfs: Samba VFS module for glusterfs

Andrew Bartlett — 2013-05-25T11:23:29


I installed glusterfs-devel on my Fedora 18 system, but didn't get
api/glfs.h anywhere.  It also seems unlikely that such header name would
be in the default search path, so I'm wondering should we be using some
pkg-config search to find where that would be?

I wanted to at least compile this locally before I put my review
blessing on it, which I was otherwise ready to do, as my other concerns
are addressed. 

Thanks,

Andrew Bartlett

[PATCH 2/2] dns: Support larger queries when asking forwarder

Kai Blin — 2013-05-25T11:21:33

This should fix bug #9632

Signed-off-by: Kai Blin <kai< at >samba.org>
---
 source4/dns_server/dns_query.c  | 13 +++++++++++++
 source4/dns_server/dns_server.c |  2 ++
 source4/dns_server/dns_server.h |  4 ++++
 source4/dns_server/dns_utils.c  | 21 +++++++++++++++++++++
 4 files changed, 40 insertions(+)

diff --git a/source4/dns_server/dns_query.c b/source4/dns_server/dns_query.c
index 4ad14b9..5414e1d 100644
--- a/source4/dns_server/dns_query.c
+++ b/source4/dns_server/dns_query.c
< at >< at > -138,14 +138,17 < at >< at > struct ask_forwarder_state {
 static void ask_forwarder_done(struct tevent_req *subreq);
 
 static struct tevent_req *ask_forwarder_send(
+struct dns_server *dns,
 TALLOC_CTX *mem_ctx, struct tevent_context *ev,
 const char *forwarder, struct dns_name_question *question)
 {
 struct tevent_req *req, *subreq;
 struct ask_forwarder_state *state;
+struct dns_res_rec *options;
 struct dns_name_packet out_packet = { 0, };
 DATA_BLOB out_blob;
 enum ndr_err_code ndr_err;
+WERROR werr;
 
 req = tevent_req_cre

[PATCH 1/2] idl: Add support for parsing OPT records

Kai Blin — 2013-05-25T11:21:32

Signed-off-by: Kai Blin <kai< at >samba.org>
---
 librpc/idl/dns.idl | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/librpc/idl/dns.idl b/librpc/idl/dns.idl
index 984f2b3..5b86f7e 100644
--- a/librpc/idl/dns.idl
+++ b/librpc/idl/dns.idl
< at >< at > -106,6 +106,7 < at >< at > interface dns
 DNS_QTYPE_ATMA   = 0x0022,
 DNS_QTYPE_NAPTR  = 0x0023,
 DNS_QTYPE_DNAME  = 0x0027,
+DNS_QTYPE_OPT    = 0x0029,
 DNS_QTYPE_DS     = 0x002B,
 DNS_QTYPE_RRSIG  = 0x002E,
 DNS_QTYPE_NSEC   = 0x002F,
< at >< at > -168,6 +169,12 < at >< at > interface dns
 } dns_srv_record;
 
 typedef [public] struct {
+uint16     option_code;
+uint16     option_length;
+uint8      option_data[option_length];
+} dns_opt_record;
+
+typedef [public] struct {
 dns_string     algorithm;
 uint32         inception;
 uint32         expiration;
< at >< at > -216,6 +223,7 < at >< at > interface dns
 [case(DNS_QTYPE_TXT)]dns_txt_record   txt_record;
 [case(DNS_QTYPE_AAAA)]ipv6address ipv6_record;
 [case(DNS_QTYPE_SRV)]dns_srv_record srv_record;
+[case(DNS_QTYPE_OP

[PATCH 0/2] Support DNS UDP packets larger than 512 bytes

Kai Blin — 2013-05-25T11:21:31

By supporting EDNS0 OPT packets as specified in RFC 2671, we can ask our
forwarder for larger DNS UDP packets without having to requery via TCP.
The first patch adds the parser support, the second patch makes the
forwarder code use the OPT record, fixing bug #9632.

Kai Blin (2):
  idl: Add support for parsing OPT records
  dns: Support larger queries when asking forwarder

 librpc/idl/dns.idl              |  8 ++++++++
 source4/dns_server/dns_query.c  | 13 +++++++++++++
 source4/dns_server/dns_server.c |  2 ++
 source4/dns_server/dns_server.h |  4 ++++
 source4/dns_server/dns_utils.c  | 21 +++++++++++++++++++++
 5 files changed, 48 insertions(+)

Re: [PATCH] winbind: Print error code on connection error in ping_dc

Michael Adam — 2013-05-25T11:03:42

I also review+ and would push,
but before pushing, since is currently in discussion:

Christof: Can we add your "signed-off-by" tag?
In the future, could you post patches with sign-off?

I assume that your patches are submitted under personal
copyright. correct?

Cheers - Michael

On 2013-05-24 at 17:05 -0700, Jeremy Allison wrote:

[PATCH v6] vfs_glusterfs: Samba VFS module for glusterfs

Anand Avati — 2013-05-25T09:30:46

Implement a Samba VFS plugin for glusterfs based on gluster's gfapi.
This is a "bottom" vfs plugin (not something to be stacked on top of
another module), and translates (most) calls into closest actions
on gfapi.

Signed-off-by: Anand Avati <avati< at >redhat.com>
---
 source3/Makefile.in             |    5 +
 source3/configure.in            |   11 +
 source3/modules/vfs_glusterfs.c | 1493 +++++++++++++++++++++++++++++++++++++++
 source3/modules/wscript_build   |    9 +
 source3/wscript                 |    6 +
 5 files changed, 1524 insertions(+), 0 deletions(-)
 create mode 100644 source3/modules/vfs_glusterfs.c

diff --git a/source3/Makefile.in b/source3/Makefile.in
index efe0694..a12887d 100644
--- a/source3/Makefile.in
+++ b/source3/Makefile.in
< at >< at > -915,6 +915,7 < at >< at > VFS_TIME_AUDIT_OBJ = modules/vfs_time_audit.o
 VFS_MEDIA_HARMONY_OBJ = modules/vfs_media_harmony.o
 VFS_BTRFS_OBJ = modules/vfs_btrfs.o
 VFS_CEPH_OBJ = modules/vfs_ceph.o
+VFS_GLUSTERFS_OBJ = modules/vfs_glusterfs.o
 
 PAM_ERRORS_OBJ = ../libcli/a

Re: [PATCH] Optimisation for readdir using fstatat.

Stefan (metze) Metzmacher — 2013-05-25T07:13:48

Hi Jeremy,


Here's a typo HAVE_fstatat => HAVE_FSTATAT.

But also notice that the additional conf.DEFINE() calls are not needed.

conf.CHECK_FUNCS() calls conf.CHECK_FUNC() which uses define='HAVE_%s' %
f.upper()
as the default define.

metze

Re: [PATCH 3/3 v4] vfs_glusterfs: Samba VFS module for glusterfs

Simo — 2013-05-25T01:42:43

look for dlinklist.h


Ok, maybe just add a comment then that this will be supported in the 
near future.


Ok, it will definitely make a big difference.


As far as I can see in the code the errno is ignored.
A -1 will cause a NT_STATUS_SHARING_VIOLATION

Maybe we should patch the vfs to check the errno value and aout disable 
kernel share modes if ENOTSUP is returned ...


Gluster makes this available on both trusted. and user. namespace ? I am 
confused.


Yes, this is probably best.


I am not sure I understand how this works, if samba is 'trusted' does it 
mean samba is responsible to perform access control ? Because samba 
doesn't, it defers access control to the kernel.

Does gluster use geteuid()/getegid()/getgroups() to find out what are 
the current ids ?

Simo.