Badlock regression fixes

[Christian Ambach]

> I've pushed
> https://git.samba.org/?p=metze/samba/wip.git;a=shortlog;h=refs/heads/master3-tmp1
> to autobuild after review from Andreas and Günther.

I am afraid those have brought new regressions.
Look at Bug 11910. It's reproducible when having an AD DC with default
server signing setting and attempting to connect with smbclient and -U%.
This seems to falsely activate SMB signing chcks on the server side
where it shouldn't.

It does not show up in our automated testing as the DC that we provision
uses server signing = auto (not the default).
We could apply a patch like the one that I have attached, but I am not
sure which testcases that will break.

Regards,
Christian

-------------- next part --------------
From e292dbd6165ebcf4f9a02e0e4559ba855979aea7 Mon Sep 17 00:00:00 2001
From: Christian Ambach <ambi at samba.org>
Date: Thu, 12 May 2016 16:45:23 +0200
Subject: [PATCH] selftest: use standard signing parameters in test env

Signed-off-by: Christian Ambach <ambi at samba.org>
---
 selftest/target/Samba4.pm | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/selftest/target/Samba4.pm b/selftest/target/Samba4.pm
index 6daa6b2..0801a86 100755
--- a/selftest/target/Samba4.pm
+++ b/selftest/target/Samba4.pm
@@ -570,7 +570,7 @@ sub provision_raw_step1($$)
 
 	# remove this again, when our smb2 client library
 	# supports signin on compound related requests
-	server signing = on
+	# server signing = on
 
         idmap_ldb:use rfc2307=yes
 	winbind enum users = yes
@@ -1638,7 +1638,7 @@ sub provision_ad_dc($$)
 
 	max protocol = SMB3
 	read only = no
-	server signing = auto
+	# server signing = auto
 
 	smbd:sharedelay = 100000
 	smbd:writetimeupdatedelay = 500000
-- 
1.9.1