[cifs-protocol] Running AD Family test suites

Nathan Manis nmanis at microsoft.com
Wed Sep 21 15:19:47 UTC 2016


Hi Sam,

Thank you for the e-mail.   Someone from the test suite team will be reviewing to assist further.

Thanks,
Nathan


-----Original Message-----
From: Garming Sam [mailto:garming at catalyst.net.nz] 
Sent: Wednesday, September 21, 2016 1:10 AM
To: Interoperability Documentation Help <dochelp at microsoft.com>
Cc: cifs-protocol at lists.samba.org
Subject: Running AD Family test suites

CC: cifs-protocol.


On 21/09/16 17:07, Garming Sam wrote:
> Hi,
>
> In preparation for the upcoming IO Lab, I have been attempting to run 
> the AD Protocol Family test suites (from
> https://connect.microsoft.com/site216/Downloads) to identify any 
> glaring flaws.
>
> I was wondering about if there was any documentation or guidance on 
> how to get Samba (or a non-Windows AD) running these tests. For 
> instance there was a non-documented enum value for OS version, NonWin 
> which bypassed some checks.
>
> Despite passing it down, it is not always respected. In the version of 
> the tests distributed, AD_LDAPModelAdapter.cs omits the isWindows 
> attribute in the cleanup code causing all the LDAP tests to fail 
> against Samba. Adding the extra argument appears to resolve this immediate issue.
>
>
> Another crucial failure I've seen so far is the inability to perform a 
> DrsBind to Samba.
>
> e.g. in
> Microsoft.Protocols.TestSuites.ActiveDirectory.Drsr.DRSContextHandle.D
> RSR_DRSBind_UnBind
>
> Samba aborts on null padding checks, claiming that the bind was in 
> non-DCE style mode (which network traces appear to support). Adding 
> the flag doesn't help as the DsBind packet appears malformed in other ways.
> I tried switching it to the native RPC library but it gives me the 
> following error message:
>
> attempt to read/write 1 bytes behind allocated memory region
>
>
> This is despite the fact that across the wire, the bind appears to 
> have succeeded correctly.
>
> Any help would be appreciated.
>
>
> Cheers,
>
> Garming
>



More information about the cifs-protocol mailing list