[PATCH] Fix bug 13465
Jeremy Allison
jra at samba.org
Tue Sep 11 15:53:57 UTC 2018
On Tue, Sep 11, 2018 at 04:02:15PM +0200, Ralph Böhme wrote:
> On Mon, Sep 10, 2018 at 10:35:32AM -0700, Jeremy Allison wrote:
> > On Mon, Sep 10, 2018 at 07:31:11AM +0200, Ralph Böhme wrote:
> > > On Thu, Sep 06, 2018 at 05:54:36PM +0200, Ralph Böhme wrote:
> > > > On Thu, Aug 30, 2018 at 03:31:42PM -0700, Christof Schmitt wrote:
> > > > > Yes, i made the change, together with the other requested changes
> > > > > (commit message update and fixup of pointer initialization). See
> > > > > attached updated patches.
> > > >
> > > > happy. Pushed. Thanks!
> > >
> > > here's a follow-up patch for s3/smbtorture.
> > >
> > > Please review & push if happy.
> >
> > Nice catch, thanks. Pushed.
>
> ...and probably failed (hang) in autobuild due to a hanging torture test
> (LOCAL-MESSAGING-READ4) as pointed out by Volker in a private mail
> conversation.
Sigh. And it looked so obviously right too :-).
> I briefly looked into the test: it creates two processes with fork() and
> then initializes messaging contexts in the parent and the child. Looks like
> this somehow interferes with the global messaging context which may point at
> a bug in messaging. Calling messaging_reinit() on the global messaging
> context did not help. I'd have to take a closer look but this has to wait
> til after IO-Lab and SDC.
>
> Attached patch paves over the issue, not sure we if we wanna go into that
> direction for now. Thoughts?
>
> I started an autobuild on gitlab with the new patch:
> https://gitlab.com/samba-team/devel/samba/pipelines/29916243
>
> Please don't push the patch as is, it's lacking my signed-off for a reason... :)
I'll take a look, but this stuff is obviously too
subtle for me :-(.
Jeremy.
> --
> Ralph Boehme, Samba Team https://samba.org/
> Samba Developer, SerNet GmbH https://sernet.de/en/samba/
> GPG Key Fingerprint: FAE2 C608 8A24 2520 51C5
> 59E4 AA1E 9B71 2639 9E46
> From a777417bb7dde3accd3d8840d61d641013eefd56 Mon Sep 17 00:00:00 2001
> From: Ralph Boehme <slow at samba.org>
> Date: Sun, 9 Sep 2018 21:54:04 +0200
> Subject: [PATCH] s3:smbtorture: Use cmdline_messaging_context
>
> Initialize the messaging context through cmdline_messaging_context to
> allow access to config in clustered Samba.
>
> Bug: https://bugzilla.samba.org/show_bug.cgi?id=13465
> ---
> source3/torture/torture.c | 8 ++++++++
> source3/wscript_build | 1 +
> 2 files changed, 9 insertions(+)
>
> diff --git a/source3/torture/torture.c b/source3/torture/torture.c
> index 098560968dd..e2639422ede 100644
> --- a/source3/torture/torture.c
> +++ b/source3/torture/torture.c
> @@ -45,6 +45,7 @@
> #include "lib/util/base64.h"
> #include "lib/util/time.h"
> #include "lib/crypto/md5.h"
> +#include "cmdline_contexts.h"
>
> extern char *optarg;
> extern int optind;
> @@ -11992,7 +11993,14 @@ static void usage(void)
> set_dyn_CONFIGFILE(getenv("SMB_CONF_PATH"));
> }
> }
> + cmdline_messaging_context(get_dyn_CONFIGFILE());
> lp_load_global(get_dyn_CONFIGFILE());
> + /*
> + * Free the global messaging context otherwise certain messaging related
> + * tests (LOCAL-MESSAGING-READ4) fail (hang) for strange reasons.
> + */
> + cmdline_messaging_context_free();
> +
> load_interfaces();
>
> if (argc < 2) {
> diff --git a/source3/wscript_build b/source3/wscript_build
> index 50d675cf3fc..bb58fa810a1 100644
> --- a/source3/wscript_build
> +++ b/source3/wscript_build
> @@ -1210,6 +1210,7 @@ bld.SAMBA3_BINARY('smbtorture' + bld.env.suffix3,
> idmap
> IDMAP_TDB_COMMON
> samba-cluster-support
> + cmdline_contexts
> ''',
> cflags='-DWINBINDD_SOCKET_DIR=\"%s\"' % bld.env.WINBINDD_SOCKET_DIR,
> install=False)
> --
> 2.13.6
>
More information about the samba-technical
mailing list