On 11/20/08 18:34, MC wrote:
>> If you truly believe this is a valuable mode of operation, please
>> feel free to submit an ARC case.
> 
> That is not a reasonable request.  I'm not sure you are in touch
 > with users and potential users. I'm hardcore enough to read this
 > obscure forum, and even hardcore enough to know what you mean
 > when you say ARC, but I'm not hardcore enough to know how or why
 > I should introduce myself into your development system just to
 > request a one sentence feature.

I suspect Bill is intimately familiar with the process.

> So given that I am so experienced
 > yet am so far detached from your request, how do you know you
 > aren't missing a larger number of people who are even further
 > detached from this thread on this little forum?

The process would be to submit a request for enhancement (RFE)
change request (CR).

As this change would affect an external interface and would
have a security impact, it would have to be submitted to the
architectural review committee (ARC) for approval.

> Theoretical situation: I run an internet cafe with freely accessible
> internet access and a cifs share with some read-only files.
> Why can't I use solaris as my server?  Because solaris doesn't
 > let me do that.

I am absolutely certain that Solaris will not stop to you sharing
read-only files over a CIFS share just because you run an
Internet cafe.  I think that might be considered discrimination.

 > I'm not filing the request because I'm not invested
> enough in solaris because it doesn't do what I need.

If you have no interest in filing an RFE, it seems the discussion
is moot.

> "File an ARC case" LOL.  I don't think you're serious when you
 > tell a user to do that.

I think I said that such a request would have to be presented to
the ARC, and that I suspected the ARC would need more justification
than: someone found it inconvenient/too hard/painful to configure
systems on their home network for user authentication.  It might
be a pain but that's not the point.  If such as case was presented,
I suspect the ARC would probably focus on the security issues and
whoever was presenting it would have to provide sufficient assurance
that it would not present a security risk.  The fact that the request
came from someone who wants to use it on a home network is not
relevant.  The wider implications of null sessions and anonymous
access need to be considered.

For example, I believe that null sessions were disabled on Windows
2000 but were re-enabled on Windows XP and Windows Server 2003
after AnonymousLogon was removed from Everyone and an
EveryoneIncludesAnonymous registry value was added, which is also
disabled by default.

On Windows 2000, RestrictAnonymous disallowed null sessions by
default.

On Windows XP and Windows Server 2003, RestrictAnonymous controls
whether or not AnonymousLogon can enumerate SAM accounts and shares,
and EveryoneIncludesAnonymous is used to avoid unanticipated access
to files via the Everyone group.

That's just one example of the wider ramifications of changing the
authentication policy.  There are ripple effects and I think the
ARC would want reassured that security had been considered from a
range of perspectives and that everyone felt suitably comfortable
that it would not present an unacceptable risk.

 > I think the decision is made and you're just saying "no".

I think I'm ambivalent on supporting null sessions and that a decision
would be the outcome of the review process.  Perhaps we can agree to
disagree on what I'm saying.

Alan

_______________________________________________
cifs-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/cifs-discuss

Reply via email to