__try __except on high irql

Anton

Your choice of an analogy in physics was poor I think. Physics is a discipline in which it is possible to make discoveries which transform the way that we understand the ‘system’ works. An obvious example is classic physics and the discovery of relativity.

Computer science is not a discipline in which this sort of discovery can be possible because it is not in fact a science per se. There is no search for an abstraction to model the existing physical world using the scientific method of hypothesis, experiment and refine, but rather a concerted program of creating interfaces based on some extant physical reality but on a desired set of behaviours

Of course it is always possible to use discoveries in physics and other sciences to build better machines that can be used to solve problems better faster and cheaper, but there is no certainty that programs designed for the current generation of machines would run on them. When the changes fall within existing abstractions, such as a new generation of fibre optics for an HBA + SAN connection, the programs need not be modified in any way. When the changes fall into the category of breaking existing abstractions, such as a CPU based on quantum effects instead of sequential processing, those new machines don’t break the existing abstractions or transform them in any way, but augment them with new incompatible ones that need to be programmed for explicitly.

To that end, the idea that ‘eventually’ every assumption is specious. What would you do with a machine where assigning a value to a memory location causes side effects in the machine which do not include the

Ability to read the same value from that location later – oh wait lots of control registers work that way ???

Sent from Mailhttps: for Windows 10

________________________________
From: xxxxx@lists.osr.com on behalf of xxxxx@hotmail.com
Sent: Monday, April 23, 2018 1:15:09 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev]__try __except on high irql

> Didn’t you read Mike’s most recent post? He wrote: “That may be possible”.

I guess this is just a classical example of a cultural misunderstanding. Look what Mike said



This sounds (at least to me) like a typical English humourous and indirect assertion of something being very obviously impossible (like, for example, " It may be possible to see a cow flying in a birdlike fashion - please let me know when there is a confirmed observation of it.“).

However, you seem to be coming from a culture that may be known for anything but indirection and sense of humour, so that you take his words literally like " Look- he agrees that cows may fly”.

https://www.newscientist.com/article/2134843-chinese-satellite-beats-distance-rec ord-for-quantum-entanglement/

Well, I don’t know how the whole thing may be possibly related to this discussion. This article seems to be speaking about the quantum cryptography that has absolutely nothing to do with either quantum computing or superluminal communication. It takes advantage of the fact that it is impossible to measure the photon transparently to its entangled twin. As a result, if an eavesdropper manages to intercept the encryption key it is going to become obvious to the communicating parties, so that they are going to choose another key.

AFAIK, unlike quantum computing, quantum cryptography is a relatively mature field with commercial-grade products being available. However, it does not imply superluminal communication in any possible way. After all,I have never heard about the products being advertised like “Want to receive the emails from your deceased auntie? Want to know who wins the World Cup in the year 2178? Our new product makes it possible…(etc)”

Anton Bassov


NTDEV is sponsored by OSR

Visit the list online at: http:

MONTHLY seminars on crash dump analysis, WDF, Windows internals and software drivers!
Details at http:

To unsubscribe, visit the List Server section of OSR Online at http:</http:></http:></http:></https:>

To cut a long story short: I erroneously understood that faster than light information transfer by Quantum Entanglement has very recently become common scientific knowledge proven by Zeilinger and Susskind. After revalidating details now however, this appears to be wrong indeed. Thus my statements and sample above are invalid and to be ignored.

PS: Thanks for not making any funny comments.

Marcel Ruedinger

datronicsoft

> Your choice of an analogy in physics was poor I think.

Agreed…

What I meant to say with this example is that some suggestions of certain “future technologies” may be so ridiculous that the very thought of their feasibility turns the entire foundations that our conventional wisdom and common sense are based upon, upside down. In this particular example, the very suggestion of a superluminal communication device turns the whole cause-effect relationship upside down (at least as far as our modern understanding of the laws of nature is concerned), and turns the whole thing into something reminiscent of the following limerick:

There was a young lady named Bright

Who used to be quicker than light

She went out one day, in a relative way

And came home the previous night

To be honest, it just did not occur to me to think that someone might challenge this assertion and seriously suggest the feasibility of superluminal communication, effectively turning this thread into, probably, the most ridiculous one in NTDEV’s entire history.

The example of a CPU that can dynamically self-reconfigure its own features without resetting itself
falls in the same class of"solutions" for the reason I had earlier explained. Certainly, if you remove “without resetting itself” part, the whole thing, once in a sudden, becomes perfectly feasible and reasonable.

For example, IIRC, Intel CPUs allow you to choose between PIC and APIC mode, as well as optionally enable or disable multicore/multithreading features, via the control registers.However,
these are “sticky” bits that can be modified only once per boot, because changing these options requires making a choice of OS software that cannot be changed without a reboot (i.e.different HALs, in case of Windows machine). Here we will have to deal with exactly the same situation - in order to be functional, the system will have to restart itself with a new set of software options
that correspond to CPU reconfiguration.

What would you do with a machine where assigning a value to a memory location
causes side effects in the machine which do not include the Ability to read the same
value from that location later

Actually, I don’t see anything particularly problematic here, especially taking into account that this is exactly how memory-mapped device registers work - you write to the memory location X
(i.e. a command register), and read from the memory location Y (i.e. a status one). Why do you think it should present any logical contradiction?

The only problematic part here may arise if these are the registers of a superluminal communication device. With such a device, the status register may indicate the result of an operation that is yet to be requested via the command one. Communicating with drivers for these devices is a daunting task as well - not only your IO completion routine may get invoked before you have submitted an IRP to it, but, to make it even worse, it may complete your IRP before you have even had a chance to allocate it…

Anton Bassov

So… the new community forum/mailing-list system should be in place in June.

I, personally, can’t wait.

Peter
OSR
@OSRDrivers

xxxxx@osr.com wrote:

So… the new community forum/mailing-list system should be in place in June.

I, personally, can’t wait.

If only you could use faster-than-light quantum entanglement to make
that happen more quickly…


Tim Roberts, xxxxx@probo.com
Providenza & Boekelheide, Inc.

> So… the new community forum/mailing-list system should be in place in June.

I, personally, can’t wait.

Well, as I had earlier said, one does not really need any superluminal communication device in order to see who gets on the receiving end of "The Hanging Judge"s whip…

Anton Bassov

Marcel,

To cut a long story short: I erroneously understood… Thus my statements and
sample above are invalid and to be ignored.

You seem to be a way too pessimistic. Look what showed up on Yahoo news just a couple of hours after your most recent post

https://www.yahoo.com/finance/news/meet-quantum-blockchain-works-time-144820654.html

Again, they refer to Professor Zeilinger 's work. In fact, I was left just speechless by the following excerpt from one of the papers he had co-authored

The only thing I can tell you for sure that is that, even if your statements on this thread happen to be erroneous, this is definitely neither your fault or the result of your misunderstanding - as we can see it with or own eyes, Professor Zeilinger seems to be,indeed,explicitly claiming to have had found a way of influencing the past.

Therefore, all the fault (if any) lies with Professor Zeilinger, rather than with you - the only thing you did was referring to his claims…

Anton Bassov

Not quite

While it is clearly against all of our normal expectations to change the past, the premise that the present can be influenced by the future isn?t. Extending this premise to a mathematical abstraction which looks at a complete system of ?all time? does not necessarily mean that the past can be altered but merely that there is some kind of connection between events spaced in time.

Like most things in quantum theory, as long as you understand that it is an abstract theory which is useful for describing subatomic behaviours and which equally cannot be true in the universal sense (since it cannot account for many macroscopic observable effects) you will be just fine. And in this context, just like we use words in specific technical meanings, they are doing the same thing in papers like this.

Sent from Mailhttps: for Windows 10

________________________________
From: xxxxx@lists.osr.com on behalf of xxxxx@hotmail.com
Sent: Tuesday, April 24, 2018 3:50:20 PM
To: Windows System Software Devs Interest List
Subject: RE:[ntdev]__try __except on high irql

Marcel,

> To cut a long story short: I erroneously understood… Thus my statements and
>sample above are invalid and to be ignored.

You seem to be a way too pessimistic. Look what showed up on Yahoo news just a couple of hours after your most recent post

https://www.yahoo.com/finance/news/meet-quantum-blockchain-works-time-144820654.html

Again, they refer to Professor Zeilinger 's work. In fact, I was left just speechless by the following excerpt from one of the papers he had co-authored



The only thing I can tell you for sure that is that, even if your statements on this thread happen to be erroneous, this is definitely neither your fault or the result of your misunderstanding - as we can see it with or own eyes, Professor Zeilinger seems to be,indeed,explicitly claiming to have had found a way of influencing the past.

Therefore, all the fault (if any) lies with Professor Zeilinger, rather than with you - the only thing you did was referring to his claims…

Anton Bassov


NTDEV is sponsored by OSR

Visit the list online at: http:

MONTHLY seminars on crash dump analysis, WDF, Windows internals and software drivers!
Details at http:

To unsubscribe, visit the List Server section of OSR Online at http:</http:></http:></http:></https:>

“…useful for describing subatomic behaviours…”, “…it cannot account for any macroscopic observable effects…” (I assume “many” was a typo). Sorry, but I have to disagree again. Just as I disagreed when Anton made a similar statement.

According to my understanding of common scientific knowledge, an upper size limit for quantum effects is neither known nor expected.

It is commonly known that the double slit experiment (*) can even be done with buckyballs (large molecules - spheres which are clearly visible in an electronic microscope - very far away from subatomic). Just google for “double slit experiment and buckyballs” ( e.g. https://medium.com/the-physics-arxiv-blog/physicists-smash-record-for-wave-particle-duality-462c39db8e7b )

(*) Most publicly and commonly known quantum effect involving wave/particle duality.

PS: Peter, this is your house. I don’t want to offend. Please tell me to stop if you consider that it is time to stop and this is leading too far away off topic for too many consecutive posts.

Marcel Ruedinger

datronicsoft

> According to my understanding of common scientific knowledge, an upper size limit for quantum >effects is neither known nor expected.

Indeed, it seems to be the case. You may want to read the articles below for more details

https://www.mnn.com/green-tech/research-innovations/stories/quantum-entanglement-demonstrated-level-visible-naked-eye

https://theconversation.com/experiment-shows-einsteins-quantum-spooky-action-approaches-the-human-scale-95372

OTOH, I am still not sure if it somehow implies the possibility of superluminal communication
(unlike the example below)

PS: Peter, this is your house. I don’t want to offend. Please tell me to stop if you consider
that it is time to stop and this is leading too far away off topic for too many consecutive posts.

Well, I can assure you that, no matter how it all goes, the “end recipient” on the receiving end (pun intended) of "The Hanging Judge"s whip had been appointed before you even had a chance to make your very first post on this thread, effectively taking it into a “funny” direction. This is another example of quantum-level affects taking place in the macroscopic world…

Anton Bassov

I apologize for starting this digression, but I actually think the
discussion is interesting and that qe is both fascinating and
mind-boggling. You really have to rethink your expectations, classic
physics is an emergent property of quantum physics at a macro scale.

Mark Roddy

On Fri, Apr 27, 2018 at 10:17 AM, xxxxx@hotmail.com <
xxxxx@lists.osr.com> wrote:

> According to my understanding of common scientific knowledge, an upper
size limit for quantum >effects is neither known nor expected.

Indeed, it seems to be the case. You may want to read the articles below
for more details

https://www.mnn.com/green-tech/research-innovations/
stories/quantum-entanglement-demonstrated-level-visible-naked-eye

https://theconversation.com/experiment-shows-einsteins-
quantum-spooky-action-approaches-the-human-scale-95372

OTOH, I am still not sure if it somehow implies the possibility of
superluminal communication
(unlike the example below)

> PS: Peter, this is your house. I don’t want to offend. Please tell me to
stop if you consider
> that it is time to stop and this is leading too far away off topic for
too many consecutive posts.

Well, I can assure you that, no matter how it all goes, the “end
recipient” on the receiving end (pun intended) of "The Hanging Judge"s whip
had been appointed before you even had a chance to make your very first
post on this thread, effectively taking it into a “funny” direction. This
is another example of quantum-level affects taking place in the macroscopic
world…

Anton Bassov


NTDEV is sponsored by OSR

Visit the list online at: http:> showlists.cfm?list=ntdev>
>
> MONTHLY seminars on crash dump analysis, WDF, Windows internals and
> software drivers!
> Details at http:
>
> To unsubscribe, visit the List Server section of OSR Online at <
> http://www.osronline.com/page.cfm?name=ListServer&gt;
></http:></http:>

I haven’t seen any STOP signal yet. Thus just one more post:

The real fascinating mind-boggling actually starts when thinking just one step beyond the scope of what is currently reachable or provable by contemporary science…

…according to scientific believe, all the universe was in one point at the time of the Big Bang. Thus “it may be possible” that the whole universe is entangled. Thus things happening here on earth may possibly be influenced by things happening somewhere else in the universe, in a galaxy far, far away…

Marcel Ruedinger

datronicsoft

I don’t know if it can get explained somehow by the quantum mechanics effects, but I have personally observed “a strange phenomenon” - almost the very same moment this thread took an “unexpected” direction Yahoo news started, once in a sudden, spitting out articles about the experiments that claim to confirm the most “unconventional” interpretations of quantum mechanics, and doing so at unprecedented rate. Here is one more link

https://www.mnn.com/green-tech/research-innovations/stories/parallel-worlds-exist-and-interact-with-our-world-say

If it is still not enough, it started providing “unconventional” news concerning the cosmology as well

https://www.space.com/40422-are-white-holes-dark-matter.html

According to this article, so-called “white holes” may actually predate the Big Bang

…according to scientific believe, all the universe was in one point at the time of the Big Bang.

Taking into consideration the above links, the very first questions that get into my head are

  1. Which particular universe are we speaking about - as we can see, there are many of them, and they even are claimed to interact with one another

  2. Did the Big Bang start them all in one go, or it applies only to our particular universe?

  3. How do white holes that predate the Big Bang map into the whole picture?

Anton Bassov

> effectively turning this thread into, probably, the most

ridiculous one in NTDEV’s entire history.

But Mr. Bassov, this way it is still better than the other alternative (discussion of SEH and stack unwinding at high IRQLs…)

The example of a CPU that can dynamically self-reconfigure its own features
without resetting itself
falls in the same class of"solutions" for the reason I had earlier explained.
Certainly, if you remove “without resetting itself” part, the whole thing, once
in a sudden, becomes perfectly feasible and reasonable.

If the baddest problem case here is live migration of VMs to a different CPU - this is now mitigated by containers, AFAIK. Workloads are migrated by applications in containers more often than entire VMs, and decent applications should not depend on low-level CPU details.
Virtual or physical machines (with the OS, drivers and so on) typically are started in a normal way, then containers with apps are migrated among them. Then, no problem.

Regards,
– pa

P.S. and you can now enjoy these quantum effects in the comfort of Visual Studio :slight_smile:
https://code.msdn.microsoft.com/An-Introduction-To-Quantum-62671688

> If the baddest problem case here is live migration of VMs to a different CPU - this

is now mitigated by containers, AFAIK. Workloads are migrated by applications
in containers more often than entire VMs, and decent applications should not
depend on low-level CPU details.

Fair enough - the instructions that we were speaking about tend to be privileged ones, and,hence,
are unavailable to the userland apps anyway…

BTW, once we started speaking about the containers, the very first thing that gets into my head is the thread (unfortunately, I cannot immediately find it) where our “Windows fanboy” (a.k.a.Max) was claiming that lightweight virtualisation mechanisms (i.e. jails, zones and containers under respectively FreeBSD, Solaris and Linux) were totally pointless, and that the only virtualisation mechanism that mattered was a full-blown machine virtualisation. However, as we can see, in actuality, light-weight virtualisation turned into one of the hottest trends, contrary to his beliefs…

https://code.msdn.microsoft.com/An-Introduction-To-Quantum-62671688



OMG - as it turns out, the only thing that I was doing on this thread was promoting a new MSFT technology…

Anton Bassov

Unsupported assertion here. As measured by, what exactly?

It would seem to me that VMotion and Veeam Quick Migration are still quite popular.

Not that it matters. Never mind, the horse is dead.

Peter
OSR
@OSRDrivers

Just seen this thread and wanted to share my understanding. I hope members of the forum can accept this post. The points mentioned below were correct and surprisingly so precise touching the limits of world experienced:

effectively stripping the original statement of its
actual meaning and assigning a totally different one to it

I would summarize the stuff as ‘context’. So we need to interpret a word in its context.

give different
interpretations of exactly the same binary instruction (i.e. before and after
self-reconfiguration)

Sir, as we can see it is due to context. It is possible in technology … if you see elisp where the self modifying characteristics of a programming language is achieved more here. Actually whole philosophy of current thinking about how we are programmed to think has to change in my view since the fundamentals of John Von Neumann architecture of what is separation of memory and processing and also the halting problem are not understood or taught properly. Very few indeed even are aware of Alan Turing Vs John Van Neumann conflict and the implications down to our day to day decisions down to things like from porting ipv4 to ipv6 to caching to cracking bitcoin to thinking beyond AI. The reason I have mentioned all this is to highlight the importance of different way of thinking.

There are other things in this thread like going back in time which reminded me 12th chapter in “Tao of Physics” (it is possible) and same macroscopic object appearing at different places (not just object but living being but I would stop here lest I do not get idea whether to proceed or not in this way here). Thanks.

> It is possible in technology … if you see elisp where the self modifying characteristics

of a programming language is achieved more here.

Don’t forget that LISP is a high-level language that heavily depends upon its runtime library.

It “self-modifying” features are due to the fact that it does not make any distinction between the code and the data, because “everything is a list” , from its perspective. Therefore, if a function returns a list that subsequently gets executed, the whole thing starts looking as if a LISP program is capable of writing its own code. However, it cannot rewrite its own runtime, can it. Therefore, interpretation of its “instructions” does not change in any possible way.

In fact, the same thing is possible at the level of assembly language as well. Although you would not normally expect to encounter something like that at the time when the system is up and running, this is an absolutely standard procedure at the early stages of the boot process when a bootloader loads the code into memory, modifies certain parts (normally the ones that deal with offsets) of it, and then jumps to modified code.

However,how on Earth all the above may be possibly related to CPU that is capable of dynamically reconfiguring itself??? Look - dynamic hardware reconfiguration is a basically different thing that involves modifying the logic cells. Basically, you can think of a logic cell as of a cell with, say, 6 inputs that is configured to produce either 0 or 1 for each of 64 possible combinations of inputs.
Therefore, by reconfiguring these cells you are capable of changing ABSOLUTELY everything
about the target hardware. Assuming that the target hardware implements an instruction set, how can you possibly reconfigure it without upsetting the software that it executes???

Anton Bassov

In principle, I can give simple example of “repne scasb”. It is a single instruction which satisfies the definition of N possible inputs and different output even when the input is same. But the direction of search changes (hardware is reconfigured in a way here) depending on the context set. I would like to remember transmeta approach to hardware and ways to reconfigure processors. I do not know details but I believe their IP has answer.



Just seen this thread and wanted to share my understanding.



In principle, I can give simple example of “repne scasb”. It is a single instruction which satisfies the definition of N possible inputs and different output even when the input is same. But the direction of search changes (hardware is reconfigured in a way here) depending on the context set.

Am I the only one who finds the above combination…ugh,“amusing”, so to say…

I would like to remember transmeta approach to hardware and ways to reconfigure processors.

AFAIK, they relied upon translating x86 instructions to their proprietary IS. As long as you don’t expose your actual IS to the software that you run you may, indeed, change certain implementation details transparently to the software that you run, at least in theory.However, this is
NOT what we are speaking about here,don’t you think…

Anton Bassov