Drivers signed with grandfathered certs & Win10 1809

NTDEV
21

There are justifiable legal liability and traceability aspects to that process that make total sense to me.

Well, I don’t really know what you mean by “legal liability and traceability aspects”, but I would rather suggest checking the following thread

https://community.osr.com/discussion/173161/why-is-signing-drivers-such-a-mess

As you are going to see with your own eyes, a mere suggestion that a certificate holder should be liable for the damages that drivers signed with their certificate may cause immediately results in being branded as Stalin’s fan ( certainly, not in Mr.Kyler’s style, but still)…

https://community.osr.com/discussion/comment/173297#Comment_173297

Furthermore, according to the same comment, the worst thing that may possibly happen to the certificate owner is the certificate getting revoked by MSFT.

In other words, legal liability does not really seem to be anywhere in sight, does it…

I do a fair amount of work in the scientific, embedded, and industrial control world. I can tell you that quite a lot of them
have moved to Linux rather than trust Windows 10.

Well, AFAIK, Windows has never been particularly popular in the above mentioned domains. Linux has always been the OS of choice down there, and it had worked this way long before the advent of Windows10.

I think that the very fact of being free is, probably, the most significant factor that leads to choosing Linux over the proprietary OSes. For example, consider an HPC cluster with hundreds of thousands of separate nodes. If you use the proprietary OS every single node is going to count as a separate installation, significantly increasing the overall costs of the project. The situation with the embedded development is pretty much the same - the additional cost of the proprietary OS may have a very negative effect on price/cost ratio and overall marketability of the target device…

Therefore, I think the reasons for this may, probably, be mainly of an economical nature, rather than of a technical one…

Anton Bassov

22

anton_bassov wrote:

Well, I don’t really know what you mean by “legal liability and traceability aspects”,

Yes, you do.

but I would rather suggest checking the following thread
https://community.osr.com/discussion/173161/why-is-signing-drivers-such-a-mess
As you are going to see with your own eyes, a mere suggestion that a certificate holder should be liable for the damages that drivers signed with their certificate may cause immediately results in being branded as Stalin’s fan …

You do have a unique way of bringing out the best in people.

In other words, legal liability does not really seem to be anywhere in sight, does it…

Peter is ancient and wise, but he is not a lawyer, nor does he speak for
Microsoft.  If you sign a driver that causes physical damage or injury,
you can bet your ass that you will be drawn into court, and your
liability is going to be determined in part by the chain of trust that
leads to you.  There is no doubt in my mind that this was one of the
stronger justifications for the implementation of KMCS.

23

You do have a unique way of bringing out the best in people.

See below…

If you sign a driver that causes physical damage or injury, you can bet your ass that you will be drawn into court, and your
liability is going to be determined in part by the chain of trust that leads to you. There is no doubt in my mind that this
was one of the stronger justifications for the implementation of KMCS.

Well, and you seem to have an unique,exceptional ability to stuff the incredible amount of nonsense into a short statement
(sorry, but you have asked for it yourself, did not you) …

To begin with, anything that may cause a physical damage, death or injury (medical equipment, avionics systems, industrial control systems, etc,etc,etc) is a subject to stringent safety standards , regulations and certifications. These systems tend to be closed and tightly-coupled,
with all the hardware and software components known in advance. In highly unlikely scenario of the system in question running Windows,
this is the very last place anywhere in the Observable Universe where one may encounter a potentially rogue driver of unknown origin,
i.e. the problem that driver signing is meant to deal with. Therefore, driver signing has absolutely nothing to do with the safety concerns.

The most likely place where a driver of this kind may be encountered is “average Joe’s” personal system . Furthermore, this kind of driver is more than likely to be installed programatically, rather than upon the user’s decision. The is particularly true if our “Joe” runs his system as an Admin on regular basis, which is still a common occurrence in the Windows world. The damages that this kind of driver may cause are
more than likely to be of economic nature (money stolen from the bank account; files encrypted by “ransomware”; etc).

This is what driver signing may,indeed, THEORETICALLY prevent, but, in actually, it does not for the reasons stated below

Now let’s proceed to the legal part. Everything that you say about the possible legal ramifications may apply only as long as we are speaking about a legitimate business with the offices in the US (or at least somewhere in the civilised world). However, not all jurisdictions are like that. Some of them are “not-so-civilised”, so that the rules that are generally accepted in the Western world do not necessarily apply down there (put an outsourcing location of your choice here). Some of them are particularly suitable for registering potentially dubious companies that exist only as PO Boxes (put an offshore location of your choice here).

Someone who is about to sign a malicious driver is, apparently, going to use the certificate issued to some entity that is registered
in a “safe” jurisdiction of either of above mentioned types, right. Therefore, my question is " If a driver signed by XYZ had caused truly disastrous damages, in which particular jurisdiction are you going to sue the certificate holder, especially if it happens to be extremely short-lived (which is a usual case with the companies set up for the nefarious purposes) ???"

As you can see it yourself, for all the annoyance that it causes, driver signing does not really seem to solve anything, does it…

Anton Bassov

24

On Mar 19, 2019, at 5:09 PM, anton_bassov wrote:
> As you can see it yourself, for all the annoyance that it causes, driver signing does not really seem to solve anything, does it…

Yes, it does. You keep setting up your own arguments and knocking them down, but you haven’t really established anything.
—
Tim Roberts, timr@probo.com
Providenza & Boekelheide, Inc.

25

Yes, it does. You keep setting up your own arguments and knocking them down, but you haven’t really established anything.

Fair enough - if you believe that driver signing is going to prevent injuries and bring the attackers to justice, it is your right to believe so. Therefore, I am not going to try convincing you otherwise - judging from your latest post, it seems to be pretty much the same thing as getting into an intelligent argument with someone who believes that the Earth is flat and about 7000 years old…

Anton Bassov