On Wed, 2008-09-17 at 14:02 -0700, Hongwei Sun wrote:
> Andrew,
> 
>   After running Samba RPC-PAC test, analyzing network trace  and
> reviewing its source code, we think that we found the problem in the
> Sambatorture  implementation.   In the loop of setting message type
> from 0x00 to 0xFF, the test program sends the exactly same
> PAC_Validate buffer for each call.  This can be observed from the
> network trace.  Then we confirmed that in ndr_push_PAC_Validate(),
> which marshals the PAC_Validate structure,  message type is always set
> to NETLOGON_GENERIC_KRB5_PAC_VALIDATE (0x3).  That explains why
> Microsoft servers always return NT_STATUS_OK for all the calls in your
> test.

That makes much more sense. 

>   We also found that the other tests(wrong length, corrupted data, bad
> signature etc)  performed by Smbtorture failed as expected.
> 
>   Please let us know if what we found is correct.

Thank-you very much for your analysis, particularly the level to which
you were willing to dig into this.   I'll fix the test to avoid the
pointless looping :-)

Andrew Bartlett

-- 
Andrew Bartlett
http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org
Samba Developer, Red Hat Inc.                  http://redhat.com

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
cifs-protocol mailing list
[email protected]
https://lists.samba.org/mailman/listinfo/cifs-protocol

Reply via email to