RE: Junkware stamped with 'Certified for Vista' logo?

NTDEV
1

I think that the important issue here is Vista certification. Forget about
StarForce.

I agree completely with Anton. If certification should be meaningful,
there must be a way
to cancel certification also. If such cancelations were made public, it
would also show to users
that certification did really mean something.

Kurt Nyström

Piece of nonsense? Biggest publisher throwing away protection system b/c
users complain about frequent system instability? And what you think is
SENSE in such a case?

See, I really don’t want you to make a mess here. The only thing I’m
interested in - how to complain about broken driver. If you don’t know
the answer you don’t really need to work as a free public advocate for.
StarForce here. Actually on their place can be ANY company.

I thought about certification working in such a way:

  1. “Company A” releases some software and grabs “OK to install”
    certificate from the “Authority B”. 2) Somebody (probably even you)
    finds a serious bug in the “Company A” driver and complains to
    “Authority B”. 3) “B” makes “A” certificate void thus pushes “Company A”
    to fix the broken driver. 4) “Company A” applies the fix and
    re-certifies.
  2. Everybody is happy.

For now we get only 1) from the list above working. We can see the
driver logo’d but it really means nothing. What sense in such a case in
such a certification?

That’s the point… And not really StarForce blacklisting our driver
having nothing to do with the DVD protection breaking at all (just like
MSiSCSI). And like MS I don’t really care. Server software we write and
games don’t mix well.

Finally you understand?

-a

-----Original Message-----
From: xxxxx@lists.osr.com
[mailto:xxxxx@lists.osr.com] On Behalf Of
xxxxx@hotmail.com Sent: Friday, July 13, 2007 4:17 AM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev] Junkware stamped with “Certified for Vista” logo?

> Hope other game publishers will follow Ubisoft…

> http://www.glop.org/starforce/

Here is a link to one more piece of nonsense…

[begin quote]

Starforce is a software copy protection tool installed by PC game
publishers, which is designed to prevent the casual copying of retail
CDROM applications. It installs as a hidden device driver, without the
end-user’s knowledge or consent.

[end quote]

If StarForce’s driver is “hidden” from the user, why does it need MSFT
certification, in the first place???

But, compared to the “masterpiece” below, it is just absolutely nothing.

[begin quote]

Moreover, the Starforce drivers, installed on your system, grant ring 0
(system level) privileges to any code under the ring 3 (user level)
privileges. Thus, any virus or trojan can get OS privileges and totally
control your system. Since Windows 2000, the Windows line security and
stability got enhanced by separating those privileges, but with the
Starforce drivers, the old system holes and instabilities are back and
any program (or virus) can reach the core of your system by using the
Starforce drivers as a backdoor.

[end quote]

Therefore, writers of the above doc believe that:

  1. No version of Windows prior to W2K made a distinction between
    privileged and non-privileged code

  2. StarForce driver elevates app’s privilege level to that of Ring 0
    code (I can imagine the system’s reaction if privileged code segment
    makes GUI-related calls)

Anton(Kolomyeytsev), can you provide more or less serious links in order
to back up your claims - the ones you have provided in so far are really
funny for system-level developers, although they may be convincing for
technically ignorant PC users…

Anton Bassov

Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

To unsubscribe, visit the List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer

Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

To unsubscribe, visit the List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer

2

> If certification should be meaningful, there must be a way

to cancel certification also. If such cancelations were made public, it would also show to users
that certification did really mean something.

What MSFT certification actually means is that a driver that has passed it does not use any “unsupported” techniques and adhers to MSFT guidelines - nothing more than that. No matter how you look at it, you just cannot certify intentions of its writer, simply because “good” and “malicious” are subjective human terms, rather than objective technical ones. If you want to hide files, deny privileged users access to the registry keys or screw up operations of third-party software, it all can get done by “supported” means. Therefore, as long as you don’t violate MSFT guidelines, you are eligible for certification, no matter what your driver actually does. Certainly, this raises questions about the very purpose of such certification, but don’t forget that the very idea of *any*certification is all about meeting some certain formal criteria and impressing people with it, rather than indicating the objective reality…

Anton Bassov

3

Unfortunately you’re absolutely correct. So the question is still opened - who needs certifications if
driver logo means basically nothing :slight_smile:

Anton

-----Original Message-----
From: xxxxx@lists.osr.com [mailto:xxxxx@lists.osr.com] On Behalf Of xxxxx@hotmail.com
Sent: Sunday, July 15, 2007 6:42 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev] RE: Junkware stamped with ‘Certified for Vista’ logo?

If certification should be meaningful, there must be a way
to cancel certification also. If such cancelations were made public, it would also show to users
that certification did really mean something.

What MSFT certification actually means is that a driver that has passed it does not use any “unsupported” techniques and adhers to MSFT guidelines - nothing more than that. No matter how you look at it, you just cannot certify intentions of its writer, simply because “good” and “malicious” are subjective human terms, rather than objective technical ones. If you want to hide files, deny privileged users access to the registry keys or screw up operations of third-party software, it all can get done by “supported” means. Therefore, as long as you don’t violate MSFT guidelines, you are eligible for certification, no matter what your driver actually does. Certainly, this raises questions about the very purpose of such certification, but don’t forget that the very idea of *any*certification is all about meeting some certain formal criteria and impressing people with it, rather than indicating the objective reality…

Anton Bassov

Questions? First check the Kernel Driver FAQ at http://www.osronline.com/article.cfm?id=256

To unsubscribe, visit the List Server section of OSR Online at http://www.osronline.com/page.cfm?name=ListServer

4

> So the question is still opened - who needs certifications if driver logo means basically nothing :slight_smile:

Actually, the question does not even need an answer, because it must be obvious anyway…

Don’t forget that, just like any certification, MSFT certification is mostly related to marketing and not to the actual technical issues. All marketing tricks are intended for, softly speaking,“not-so-bright individuals”, and these “individuals” tend to have unhealthy obsession with the formal parameters of any description, be it a logo, a certificate, a diploma, a title, etc. Therefore, MSFT logo is always guaranteed to be in the high demand - even if it means nothing…

Anton Bassov

5

In the final analysis certification does not truly matter. Instead, it is strictly the reputation of the driver vendor in the eyes of the end-user that matters.

There are a lot of vendors that produce high-quality (or reasonable-quality) drivers that do not bother with logo signing. These vendors develop and test to their own internal standards and will use Microsoft tools or their own tools to insure their driver quality.

What can Microsoft do to help insure driver quality?

1.) Provide good driver development documentation.
2.) Provide good driver samples.
3.) Provide and encourage driver development training (e.g., DevCon/WinHEC).
4.) Provide tools to assist driver developers in developing quality drivers (e.g., PREfast, Verifier, other WDK tooks).

What sort of “tests” can they provide?

5.) WLK (heavy sigh…)

They can provide tests that provide limited assurance that some minimum requirements have been met under some limited conditions.

Right now I’m writing a WDF driver that uses the Windows Filtering Platform (WFP). Microsoft doesn’t have a clue as to what the functionality of that driver will be. All they can do is to provide the simple “Driver Reliability Test” that insures primarily that the driver can respond to a lot PnP requests and fend off bogus I/O calls. That is a modest step towards “reliability”.

Throughout the test my driver will operate in “passthrough mode” doing almost nothing. It is only when the network has a special configuration (e.g., routers), a companion control application is run and test applications begin transferring very specific network traffic that my driver will actually begin doing its real job. Microsoft can have no prior knowledge of this aspect of my driver can they? They can’t have a canned test that exercises these custom features can they?

I don’t see how Microsoft could do much more to “test” my driver (within practical cost constraints…).

We can make suggestions to them on improving things.

For example, on Vista Common Scenario I/O test for network devices uses only IPv6. That’s not smart, IMHO, because some network devices simply do not support IPv6. On XP the situation is reversed; Only IPv4 is tested - not IPv6. There should be some configurability here.

The current WLK is geared towards “certification” (I think) and is difficult (read “impossible”) for me to use in daily testing. I think that focus on 1.) through 4.) (above) by Microsoft would be helpful. I don’t see how focus on 5.) can be helpful except in a few cases that I am not interested in anyway.

Regards,

Thomas F. Divine

-----Original Message-----
From: xxxxx@lists.osr.com [mailto:bounce-293772-
xxxxx@lists.osr.com] On Behalf Of xxxxx@hotmail.com
Sent: Sunday, July 15, 2007 3:20 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev] RE: Junkware stamped with ‘Certified for Vista’
logo?

> So the question is still opened - who needs certifications if driver
logo means basically nothing :slight_smile:

Actually, the question does not even need an answer, because it must
be obvious anyway…

Don’t forget that, just like any certification, MSFT certification is
mostly related to marketing and not to the actual technical issues. All
marketing tricks are intended for, softly speaking,“not-so-bright
individuals”, and these “individuals” tend to have unhealthy obsession
with the formal parameters of any description, be it a logo, a
certificate, a diploma, a title, etc. Therefore, MSFT logo is always
guaranteed to be in the high demand - even if it means nothing…

Anton Bassov

Questions? First check the Kernel Driver FAQ at
http://www.osronline.com/article.cfm?id=256

To unsubscribe, visit the List Server section of OSR Online at
http://www.osronline.com/page.cfm?name=ListServer

6

Thomas,

What can Microsoft do to help insure driver quality?

Unfortunately, quality seems to be a relative concept. In its conventional sense, high-quality driver is
by its very definition meant, among other things, to be compatible with all other kernel-mode components. However, the very purpose of some drivers is to screw up operations of other drivers,
and the more efficient they are in pursuit of their goals, the higher their quality is from the perspective of their writers, although the community of driver writers in general tends to have exactly the opposite opinion on the subject.

According to the description of StarForce driver’s internals that the OP has provided on another thread, StarForce seems to be an example of such driver - it *intentionally* screws up operations of emulators just to make sure data can be read only from the physical and not from the virtual storage device (it does so by loop-wise DPC queuing, which, in terms of negative impact on the system performance, is almost the same thing as spinning in a loop at DPC level).

Is this driver of high or low quality??? As you can see, your answer depends on which side you happen to be - as a publisher who wants to protect his software, you are really happy about it(unless users just refuse to buy your software because of such protection) , but as a driver writer who has to worry about the possible presence of this “masterpiece” on the machine where your driver runs, you find it pretty annoying…

Anton Bassov