Schneier on Security

Clive Robinson • September 22, 2018 3:04 PM

@ Albert,

The “Escape to the last bustop south story” goes hand in hand with this story,

https://www.counterpunch.org/2018/09/21/laquan-mcdonald-is-being-tried-for-his-own-racist-murder/

The point is it does not matter what minority you are in, even if it’s the majority, the 1% see everything that way.

For them probability has no meaning, they throw four dice and they all come up six for a count of twenty four[1], that’s due to their hard work, you get all one’s for a count of six, that is your failing as a “waste of human flesh”.

There is an old joke that is unfortunatly true, because they rig the game the same way.

A small person’s car breaks down in the big persons bad neighborhood. Trying to escape they get caught by the local big person blue gang enforcers who throw them in a cage. After a while the small person is draged into a room in near dargness where a blue lieutenant is sitting behind a table lit by the only light in the room. The small peron is restrained in another chair with just their left arm free. In the small pool of light on the table is a pair of dice and a deep voice says “Throw”. The small person though scared says “no, I want to know what’s going on?” The deep woice says “Throw the dice, ifvyou don’t get double six we beat you and throw you in the garbage”. The small person after a few smacks to the head finally throws the dice and to his supprise gets double six and asks “What happens now?” The deep voice says “You get to throw again”.

That is the way neo-liberals think… As for their authoritarian “blue gang” followers they are generaly to dumb to realise that when that “minority that is the majority” turns as history shows it always does. The first targets of the “minority that are the majority” will be the foot soldiers of the “blue gang”. At which point their glorious leaders and the 1% will start humming that old Peter Paul and Mary tune with those famous words, “I’m leaving on a jet plane, don’t know when I’ll be back again”…

[1] Not many people realise that whilst the odds of throwing four sixes or four one’s is a little under one in thirteen hundred, less realise that the face value count “N” or spot value follows closely to a normal distribution. Even though it only goes from 6 to 24. Normalisation (0-18) then doing a X^(N^Y) if you get X and Y right gives you a curve not to disimilar to the “real asset” -v- population distribution…

Clive Robinson • September 23, 2018 7:33 AM

@ Rartio,

… and again the most basic arithmetic turns out to be juuuuuust out of reach.

I was tired but can’t sleep as I’m unwell again, and wanted to keep things as brief as I could.

But your comment is too cryptic even for normal times. If you disagree with something say so? If you don’t understand something then say so?[0]. But just to try to keep it short for the sake of other readers.

To try and shorten the potential for a long back and forth, I’ll expand on what I said.

Most here understand four “fair dice” are dice each having a flat probabability, that is each face is equiprobable at 1 in 6 of conming up when thrown.

Further the reason to use four dice as opposed to one thrown four times is to get independent results (ie Bernoulli trials) as well as faster results.

Thus four dice or four throws of one dice gives 6^4 = 1296 possibilities, which is just under 1300 which many say as “thirteen hundred”.

However unless you can tell the dice appart many throws will appear the same as other throws even though the actual throws are actually different. Which also has advantages as it helps reduce potential bias.

That said not every one has read Dr Donald Knuth’s multi-volume “Art of Computer Programing” –hence the semi famous Bill Gates comment– or the sometimes obscure or difficult to find primary sources the volumes are based on.

In the volume and chapter on Random Numbers[1], how to generator them, (RNG) and make use of their output effectively, you will find a bit on the approximation of a normal distribution with values from a flat distribution RNG.

Put simply if you add four or more Bernoulli trial results together, over many such operations the valurs you get form a fair enough aproximation to the normal distribution curve for many purposes[2]. Providing the RNG is “sufficient”, even though fully determanistic, you will get this aproximation to the normal distribution with successive outputs. As the normal distribution pops up in more places than you could ever look this approximation is quite usefull to engineers, scientists and those doing “applied” subjects as study or as an endeavor.

With regards “curve fitting” that is a subject that has many books addressing it but with regards power law curves f(x)=c^x, fx=x^c, f(x,y)=x^y they are often a good aproximation to what goes on in natural processes with time provided they do not become bounded. As an example the rules for cooking meat in an oven of X minutes per pound plus a pound for the oven is a linearized aproximation to a power law curve over a limited range. So are many engineering “cheats”, such as the “5CR” point etc., as it can save on the cost of log/lin or log/log graph paper or the use of log tables or slide rules or the time involved in using them.

As for “real wealth” or assets that do not devalue as money does and it’s distribution across the population due to the simple fact of the way the 1%’s behave. Their aquisition follows the exponential power law[3] as the number of “real assets” is effectively fixed[4].

[0] We’ve been through the use of maths approximations in engineering befor, likewise with Shakespear and his influance on the English language. Unfortunatly each time has taken up a lot of blog space in giving you a one to one helping hand. Other people have likewise helped you as well. Unfortunatly it does not sit well with other blog readers as often it gets to tangental or long winded. Further the use of cryptic comments just makes things a whole lot worse.

[1] Volume 2 –the one that starts with the Shakespear Hamlet quote of “O dear Ophelia…”— In chapter 3 “Random Numbers” that starts with a John Gay quote about “… Keep probability in view.”. Dr Knuth provides a number of ways to convert a flat distribution RNG output into other distributions. As part of this he points out that the formular used to generate the continuous normal distribution curve is not realy amenable to being inverted thus aproximations of the required accuracy be sort as they are much more efficient in terms of both time and memory.

[2] You can see this fairly easily, it you use one “fair dice” after a while you will find that each face value comes up around one sixth of the time and the value average to 3.5 as the “law of large numbers” indicates it should. However if instead you take two throws –or Bernoulli trials– and write down the sum of the spot values you get 2-12 value range with the thirty six combinations giving you a triangular distribution, which supprises most people the first time they see it. If you use three values it starts to round off towards a “bell curve” shape at the peak. With four throws starting to bring the tails up. With each additional throw / value improving not just the number of ranges under the curve it makes better tails. In most cases six Bernoulli trials are sufficient. It’s one of the reasons Monte Carlo simulations used in engineering and finance can find or aproximate answers to questions that whilst determanistic in principle are not realy practical using other methods[1].

[3] There is a relitively simple way to produce an exponential curve on graph paper. Pick a finishing or maximum value D then starting at a the first increment mark N make a point that is a fixed percentage P of the finishing value D. At the next increment make a new N which is the percentage P of the finishing value D minus the previous N value. The N values get fractionally closer to D but each subsequent fraction is less so it never quite reaches D.

[4] As exemplified by the “Mark Twain” famous advice, “Buy land, they’re not making it anymore.”. In human terms currently our total “real assets” are defined by the physicality of our world hence for practical purposes is fixed. It’s only energy that comes from the stars and ideas that are not bound in human terms.

Clive Robinson • September 23, 2018 11:42 AM

@ The usual suspects,

For a while now the subject of US Diplomats –CIA staffers– and families in Canada, China and Cuba suffering decidedly odd causless symptoms has been discussed on this blog.

Various things have been suggested and history trotted out that might explain what could be “Fourth Generation Warfare” being practiced.

But if it is “Warefare” what is the motive and money trail, of the “Who benifits Why and Where?” question.

The people who have done most to investigate “Non Leathal Weapons” historically is the US and several other Western “White Anglo Saxon Protestant” (WASP) nations such as the Leads in the Many-Eyes nations such as Australia, Canada, Germany, UK and the likes of France and Israel.

More recently however other’s have joind the club such as Super powers like China and Russia and a number of US and other National Defence Contractor Corporates.

Put simply it’s not in Cuba’s interests be it social, political or economic. Nor is it in most US citizens interests. Which provides possibly a distinguisher.

As I’ve mentioned before there are certain people in the US trying overly hard to restart the cold war with either China or Russia and the US MSM have played along as the “Usefull Idiot” to certainly the Russia-fobia and to a lesser extent China-fobia.

Thus even journalists are now starting to ask “Who in America benifits?” and it’s not difficult to work out exactly where the finger is pointing.

https://www.counterpunch.org/2018/09/21/us-harasses-cuba-amid-mysterious-circumstances/

To be honest it makes only a modicum more sense, than the other potential super power candidates playing at “proxie war” even though they have the technological capabilities. Which not meaning to be nasty the Cuban’s realy don’t have sufficient excess GDP to divert to such a weapons program, especially as they have little or no need for such weapons.

Any way as with many other things in the “Great Game” we’ll have to keep refining our theories, probably to only have them dashed by some other equally improbable event.

But I must admit I could warm to the idea it was certain US Corporates it has that delightfull 1960’s feel of “The Billion Dollar Brain”[1] about it.

[1] Written by Len Deighton in the mid 1960’s. In the book(s) the main protagonist is not named. However in the films a name was needed, it was desired to be the antithesis of “James Bond” so a boring name was sort. According to Michael Cain, due to a significant social gaff and the rememnerance of some dull school kid the name Harry Palmer was arived at. The brain of the book title was a computer that was programed to be an artificially intelligent war stratagist for an oil billionaire. Who had decided to go to war with “the commies”. So bang upto date then 😉

Clive Robinson • September 24, 2018 8:00 AM

@ Bruce and the usual suspects,

New Tor attack via AI called DeepCorr,

https://arxiv.org/pdf/1808.07285.pdf

Tor is vulnerable to trafic analysis as people have been saying for a while now. In fact a number of attacks have used “flow correlation attacks” already, what DeepCorr brings to the party is substantially improved deanonymisation through the use of “deep learning” AI.

When you think about it many privacy / security attacks would br considerably improved with deep learning, however untill recently deep learning has been hampered by the way can implement it on general purpose Turing Engines, something Google has been addressing in their new CPU design ad I mentioned a few days ago.

Whilst I’ve no doubt all “low latency mix nets” will suffer from this attack there are other architectures for anonymity networks. Some increase latency and add padding and others use fixed rate sending and padding other use not just padding but store and forward to move real traffic at a slower rate and use it in prefrence to padding thus increasing bandwidth utilisation especially when fixed.

As I’ve noted before Tor’s biggest failing aside from not having any anti traffic analysis methods is that the end points are outside of the network, thus making correlation between an input and output possible and without the anti-traffic-analysis measures effectively trivial…

It’s why I’ve indicated that clients and servers be inside the neywork snd fixed rate signalling as well as padding be used.

But is that sufficient?

For sometime now I’ve been doubtfull which is why I proposed a variation on the Fleet Broadcast method be used ontop of a link re-encrypting mix network with the clients and servers integrated as fixed rate nodes.

Clive Robinson • September 24, 2018 8:27 AM

@ bttb, Alejandro, ALL,

Getting Google out of chrome,

https://github.com/Eloston/ungoogled-chromium/blob/master/README.md

I know nothing further about it than what it says in the readme.

Clive Robinson • September 25, 2018 10:05 AM

@ echo,

This is why I fully reserve the right to be sceptical and wait for all the evidence.

You and me both.

Though I hate “circumstantial evidence” as I realy do not consider it anything other than an untested supposition on potential coincidence.

I’m not that keen on “forensic evidence” either, it argues backwards from effect to just one of many probable causes, as history shows “forensic practitioners” get it wrong oh so many times, especially if they work for the FBI.

That is I do not believe in applying the laws of probability in evidence when “intelligent agents” are involved. Because not only can the forensic examiners be wrong, they can be misled and worse willfully make statments they must know to be wrong. Sometimes this is due to Dunning-Kruger, sometimes cognative-bias, sometimes peer-preasure and other times due to what is best described as malicious-acts. All of which apply equally well to those who leave the evidenve to ne found.

Let’s put it this way if you are going to commit a crime and you do not want to face any consequences your basic options as an “intelligent agent” are,

1, Make it not be a crime.
2, Remove or nullify evidence.
3, Place evidence that misleads.
4, Place evidence to frame another.
5, Get another to commit the crime.

As increasing numbers are pointing out, there are around twenty untimely deaths of prominent Russian’s that have been put down by over streached underfunded police forcess as natural / accident. Which kind of makes a statistical anomaly out of them if treated as a group rather than separately.

But it’s not just the police, there is a dirty little secret about “tox-reports” which is the tests they are based on only pick up a tiny tiny tiny fraction of poisons… Some poisons they do not test for can be found on supermarket shelves, believe it or not in the food isles as well as the cleaning products. Oh then there are “party makers” and a number of other easily available stuff in garden centers. Nature is both bountiful and deadly in tooth and claw.

Clive Robinson • September 25, 2018 1:45 PM

@ Hmm,

So circumstantial and forensic evidence is out… you want a full confession we must assume?

Is not what I said is it…

What I am saying is that it needs to be treated not just with way more caution but with a great deal more skepticism than most juries ever do. Often because they are often effectively lied to by supposed experts, who treat the whole thing as a game to more prestige etc.

Have a look at the history of supposed “pour patterns” for claiming arson, they were nothing of the sort, ordinary fires would produce them for quiye easily explainable scientific reasons.

Likewise “cocain in banknotes” as a sign you are either a dealer or a habitual user, even though hair and other tests on the person show otherwise.

Or how about “metal mix” in bullets being indicative of individual batches thus tracable via purchase. Turned out to be yet another of the FBI’s long list of fake science.

The list goes on and on and on, it’s why I’m skeptical about both circumstantial and forensic evidence, and say it should be treated with care.

Which brings me onto your statment of,

I’m “skeptical” some will ever be satisfied by the facts, ridiculous as they are in this case.
There is as yet ZERO evidence these suspects are falsely accused, unless you’ve got some.
ZERO.

Actually you are wrong with that statment, all we currently know in the public domain is an accusation from the UK backed up by a little photographic information, and some statments from two Russian’s who were not under oath.

Neither is actually “evidence” in the accepted sense, nor is it something we can “test” as evidence. What might or might not exist in the UK police / CPS and in Russia we have as yet no knowledge, and the chances are that is exactly the way it will remaine.

But lets go through the little we have been told by various journalists with regards the Salisbury visit by the two Russian’s. Firstly I would expect that their visas would have been issued at a similar time if they had not been to England before and were planning a holiday here. Much as I would expect a husband and wife to obtained their visas at the same time if they were going to somewhere new and exotic etc on holiday for the first time.

As for passport numbers again if they had put in their applications at the same time to the same office then yes it’s likely they would have numbers in the same range. But also it’s posible that they could be from a “special batch” or a batch issued to a very infrequently used agency, so toss a coin on that.

However I know people with sequentially issued passports, they have lived abroad for many years, thus go to the consulate in the country they live in to get their passports renewed as it’s quite a journy the whole family renew at the same time and place.

But also there is another aspect to consider, if people on holiday staying in the same room for various reasons got their room turned over by thieves they to could end up with sequentialy numbered replacments from the consulate, to get home etc.

Thus there are a number of reasons such things could happen, what we as members of the public don’t have is the information as to which it is and what the probabilities are as nobody has said yet…

This is exactly the same issue as I have with forensics, we have an effect and we are trying to argue back to a cause, without actually knowing how many potential causes there are or what the probabilities are.

Somebody may well say, which is why I frequently say to wait and see what becomes available.

It’s the sort of thing you would expect to get cleared up in court by cross questioning etc. Then when you have their explanation you can decide on if it’s true, improbable to some degree, or false.

There was a recent case in the UK where a reasonably famous person was accused of assult and was charged. They went through several months of being pilloried by the media and being told not only were they a disgrace to the country but should loose there lively hood etc etc.

They were by law not alowed to say anything in public, which just made things worse. Come the trial a whole load of facts came out and they were found not only to be innocent but actually have been trying to prevent a serious crime from happening. Did the media apologize did they heck theyveither lied or ignored it…

So pardon me for taking the cautious approach but I’ve good reason to be, because a one sided series of ever odder statments is not a reasoned argument it’s usually a free for all as talking heads line up to out do each other on supposition that either does not or can not for good and proper reasons be challenged. The media is the modern equivalent of a bear pit where the bear is chained and hobbled and the dogs way to numerous.

Clive Robinson • September 25, 2018 4:36 PM

@ Hmm,

You do this all the time. You say “there are infinite doubts”

Here we go again with your very silly “strawman” argument, claiming things that I have not said.

It always comes down to this with you, you are at best a fantasist.

As I’ve frequently said “wait for the evidence”, we don’t actually have anything we can actually “test”.

What I may or may not think, I’ve actually not said, but you are determined to make the world think I have with “infinite doubts” in the above just being yet another example.

Thus the question arises as to what personality defect is it that you have that makes you do this endlessly? Even after you have said you are going to desist from your previous behaviours…

Enough, I see no reason to continue wasting space on this thread and upseting other readers.

Clive Robinson • September 26, 2018 6:25 AM

Riemann hypothesis solved?

Under the title,

Skepticism surrounds renowned mathematician’s attempted proof of 160-year-old hypothesis

The following article is presented,

https://www.sciencemag.org/news/2018/09/skepticism-surrounds-renowned-mathematician-s-attempted-proof-160-year-old-hypothesis

Which appears to be a compleat downer on Michael Atiyah.

That said people are waiting to see the compleat documentation on the alledged proof. What we know of it appears to be somewhat tangential and even simplistic. But the hypothesis has stood for a century and a half thus a more conventional maths approach has had a century and a half to come forward, so conventional might not be expected to work.

So as normal we will have to wait for the evidence, then test it or atleast let others test it 😉

But whilst solving the hypothesis might have a big effect on mathmatics, the reality as @Bruce notes is it’s not likely to have much of an effect on Crypto.

Clive Robinson • September 26, 2018 12:29 PM

@ bttb,

With regards,

And [Bannon] said, “Well, I have a very easy answer for you. Our side, we go for the head wound. Your side, you like to have pillow fights.

It was once pointed out to me when I was in a less northerly part of the US many many years ago that,

If the hound won’t stop chewing the table leg, you have three choices, sell the table, sell the hound or take it out the back and bury it.

It’s time some people considered the third option with the likes of those who give Bannon and co house room… After all why bother with “head wounds” when all you needs is a shovel and with just a twitch, then twenty to forty minutes and job done, the average IQ in the US has gone up a fraction…

Clive Robinson • September 26, 2018 4:46 PM

@ Ratio,

Uh-oh.

Yup

If this bit is true,

<

ul>The suspect using the cover identity of “Ruslan Boshirov” is in fact Colonel Anatoliy Chepiga, a highly decorated GRU officer bestowed with Russia’s highest state award, Hero of the Russian Federation.

And it can be checked independently to a satisfactory level, which it appears to have been with,

… multiple sources familiar with the person and/or the investigation have confirmed the suspect’s identity.

Then Russia has “over played” it’s hand, and been caught in a verifiable lie. Which is actually the sort of evidence prosecuters like even better than confessions. Especially as in effect it’s direct verification of a crime in it’s owm right (traveling under false documentation, unless Russia comes up with the equivalent of a “deed poll” as a legal deed of name change, which they could do, if their legal system alows this). With the potential for “fraud” charges as well, then there are false declarations at a boarder and no doubt a small heap of other charges that would now meet the requirments to procead to court.

I guess the real / original identity of the second person is the next most likely thing to come out. Unless the UK releases something else.

The problem is it does not realy move us forward to a court case as it’s unlikely that either person will come out of Russia any time soon. Which brings up the question of DNA / fingerprints and does the UK have them and will Interpol “Red Notices” stay open.

Unlike Russia the UK does not have legislation in place for extra territorial executions. Nor does the UK have “snatch” legislation even for it’s own citizens, it uses the International court system for such actions and trials.

But as the article notes Putin has slipped up as well, which is unusual to say the least of it. Not that it realy matters much in the big scheme of things.

As I said it will be intetesting to see what bit of evidence comes out next. We know that the two can be placed in the vicinity, but it’s not “publicly known” yet if the UK authorities can place the pair at the door stop or not, or at the place the alleged nerve agent container “scent bottle” was dumped. Whilst we have a “smoking gun” it would be nice to have finger print and DNA evidence on it or wrappings etc. (which there might be but has not been made public).

Clive Robinson • September 27, 2018 5:39 AM

@ Nemtsov fired first / Hmm?,

How’s the foot covering itchy?

“but medics quickly established they had not been the victims of a nerve agent attack.”

And how did they do this?

They looked for “evidence” which they then “tested” and found that contrary to what was being portrayed, the couple were not poisoned by nerve agent.

What is it I say people should do, “Wait for the evidence and test it”…

So you try to use the fact that people have done this, as an argument I’m wrong… Hmm where’s the logic in that?

But worse some idiot jumped in with both feet first at a red top online site and then had to cover up their rash and embarrassing action by removing the article from public viewing…

Just one reason I say “wait for the evidence and test it”. Which again confirms my view point is valid, and your “go off half cocked” with accusations is not.

Now you appear to be congenitaly incapable of understanding this and lead a fantasy life or you are a troll.

Either way your behaviour can and has been seen by many to be not just “rash” (Trump is still in office) but “embarrassingly” so (list of quotes without meaning, newspaper articles that can be shown to be wrong etc).

What you appear to fail to understand is that “method and process” are not the same as “belief” and have frequently conflated them.

Thus you go off at best imprudently half cocked, effectively tilting at windmills. You then compound it by going into denial and failure to realise you got things grossely wrong. You repeatedly made claims that were against method and process (treason) and after much pushing I gave you a considered belief that President Trump would still be in the White House come midterm elections. As the mid terms have approached and Trumps is still in office, you have thus throw up spurious or strawman argument. And when that fails you resort to childish behaviour.

You get repeatedly warned but your behaviour changes not a jot. So you resort to “sock puppetry” in the vain hope it will go unnoticed by @Moderator. But you change neither your basic writing style or argument style.

It would be amusing but for the fact it is a Denial of Service attack on this blog, and has been ever since President Trump manipulated the weakness of a two party political system and gained the Republican nomination. Which eventually gave him his place in the White House due to the faillings of what many Americans have claimed is “the most democratic of democratic processes”. Which I would suggest from your behaviour you now vehemently disagree with.

So rather than wasting your time trolling, why not go out and actually do something to change the US Political Process?

But you won’t will you? Because you don’t have “the right stuff” do you?..

I suspect you will now try to blaim someone else who disagres with your fragile fantasies, for your errant behaviour. But it should by now be clear it’s all you, and you alone. Based I suspect, because you cherish your fantasy, and find any disagrement an afront to your apparent narcistic ego.

Clive Robinson • September 27, 2018 10:26 AM

@ echo,

The case is much more than this single item and we don’t have enough certainty to rush to judgment on the rest of the allegations.

Oh I agree, with the propaganda we have seen today –which I fully expected for obvious reasons– contrary to what others have said it’s desirable to have as much of a chain of evidence as we can get.

People just pointing and saying “It must be ‘im wot dun it” does not cut the mustard when it comes to getting justice for not just the victims of what happened in Salisbury, but also getting the other cases where Russian’s in the UK have died unexpectedly and the police have for a number of reasons failed to investigate the deaths properly.

I suspect the only way it will happen is via “the domino effect” where one investigation provides evidence etc for the next and so on. Provided of course we can stop certain political and administrative persons running inyerferance, which unfortunatly we have both witnessed on more than one occasion.

Clive Robinson • September 27, 2018 4:11 PM

@ echo,

With regards,

“I don’t feel an affiliation with the state because it means to kill us – it exists to disempower us,”

Whilst I feel sorry for what he has experienced, it is unfortunatly not just people from the Caribbean who suffer this. I know and have helped disabled people who get treated almost as manikins, with just about everybody talking to me and not to the person, who is sitting next to me. One interviewer actually said to me “Do they have problems with going to the toilet?” and they did not appear to comprehend that the person was quite capable of answering for herself. But more importantly it was a totally inapropriate question to ask…

It is as you note endemic in certain UK agencies and probably many more that I’ve not experianced first hand. Worse it appears to be so ingrained and to such a level that I suspect those doing it are nolonger aware of just what they are doing, other than going through the motions of a process.

Back in the 1970’s and 80’s the Police especially the Met were accused of ingrained behaviour and it was eventually called “canteen culture” as an attempt to some how “minimise” it in the press, and to make it sound as though it was a “lower ranks only” issue. When in fact it became quite clear it went right up the hierarchy, only those at the top were a little more circumspect in how and when they voiced it.

Clive Robinson • September 28, 2018 1:29 AM

@ Hmm,

Is that actually true? I find it hard to fathom there’s no work-around.

It’s certainly true in the UK Benifits system, if you go in with a “claimant” the first question they ask after your name but sometimes before is “NI Number?”.

They will if you push hard enough for several minutes reluctantly and with a very “put upon air” go and look it up at their computer which if they can type takes at most thirty seconds…

The only time I’ve seen it not asked was when an individual came in smelling of rotting excrement, compleatly disheveled, with matted hair and dirt on their hands and face. They got taken as an immediate priority to an interview room next to the “special exit”. Shortly after the person that had interviewed them came back to their desk got a departmental issued aerosol can of insecticide out of their draw and sprayed themselves with it. Then offer the can to her colleague at the next desk making a comment that included the phrase “flea riddeled” which stuck in my mind as it should have been “flea ridden”.

Clive Robinson • September 28, 2018 6:50 AM

@ Wesley Parish,

There must be easier ways to protest room service, methinks …

I posted a couple of links about wierd things about octopuses the other day. One of them has a story about some kept in captivity where a researcher feed them some not so fresh squid… One one significant eye contact with the researcher and then very deliberatly shoved the squid down the drain in the tank…

@ Ratio,

There are supposadly two photographs of the same person to my eyes they look more different than I would expect with age alone. But human face recognition outside ethnicity is natoriously poor…

However they are both face on, so someone could upload them to one of those Social Media sites that use facial recognition to see what the match coincidence is. At one point Apples software alledgedly could pick out relatives with ease without being prompted.

Might be worth a laugh to see what the results are.

Clive Robinson • September 28, 2018 7:56 AM

@ Thoth,

… how to do 64-bit word circular rotate left (ROTL64) on a 8-bit MCU with only 8-bit math

There are three forms of “logical” shift, those that fill with zeros, those that fill from the carry flag and those that “rotate within the word”[1].

Clearly you do not want the inword rotate. Likewise you don’t want the fill with zeros. Which leaves the fill from carry.

In principle to do a one bit rotate to the left on a multiword unsigned integer object you mask off the most significant bit of the most significant word into the carry flag then do a Logical Left Shift through Carry (LLStC) instruction from the least significant word through every word in in monotonicaly increasing word order including the most significanf word.

However you say “only 8bit math”, when it comes to shifting integers you have to distinguish between “unsigned” and “signed”. Usually –but not always– the most significant bit of the most significant word is used as a sign bit in signed arithmetic, and the shift functions for that word and that word alone are usually refered to as the “Arithmetic Shift” instrunctions like ALS for left shift and ARS for right shift. Often the “through carry” is implicit for ARS output but not the input because the “sign bit” is usually left as it was and it is also shifted in to the LSB-1 bit.

So you need to carefully check any “arithmetic” shift instruction for it’s exact functionality.

However for multiword integers signed or unsigned all the other words are not signed so the ordinary logical shift instructions apply to them. So I would expect the LLStC and LRStC to be on the instruction list as well as the ALS and ASR instructions.

If you only have ALS and ASR with their akward sign bit corrections then you hit the problem of having to mask off bits do a reverse “pre-shift” and then shift in the required direction and then mask of any bits that are caught up in the sign bit manipulation…

It’s way easier to describe with the assistance of pictures not just words only…

[1] Although an 8bit CPU is usually –but not always– “byte wide” I’m going to mostly use the generic “word” as it applies to all CPU’s from 4bit Bit Slice chips all the way up to 128bit or greater graphics procrssors. Provided the CPU aloes “bus width” integer or signed integer operations.

Clive Robinson • September 28, 2018 8:24 AM

@ Ratio, Thoth,

You have a 64-bit word in memory and want to ROTL up to 63 (not 64) places

You have to be carefull although ROT[0] should be equall to ROT[64] and do nothing, either can be used for “special purposes” in some CPU’s because some alow negative values in such as ROT[-1] etc and it then depends on if it is a two’s complement or one’s complement number system, the latter having both a +0 and a -0…

Which is as scary as it sounds, most are blissfully unaware of this because they just do not get to meet that sort of hardware any longer except in highly specialised systems (which even I stopped developing more than a decade ago). Unfortunatly it’s only when they try to do crypto or similar that they find out that such things in CPUs because they alow for reduced gate count on chip, thus less real estate or more likely faster execution.

Worse not all data sheets include this “minor issue” untill you get to the signed NDA stage. As otherwise it might discourage people from using it (beware smart card NDAs)… However unlike the silicon the NDA “is for ever” so you can not run down the street pulling out your hair screaming “Never again will I use XXXY’s products ever again” either…

Clive Robinson • September 28, 2018 9:58 AM

@ Thoth,

As @Ratio has pointed out the fast way to rotate is to first “word” swap as those are just simple reads and writes with one word held in a tempory location. Then do the appropriate rotate up or down which should be at most half the number of bits in a word. So even if you are shifting left, it may be faster to shift left on word to many then bit shift right to get back to the desired value.

However you have to make a choice between increased code size thus ROM storage and slower executing code in each loop but many less loops used and small code less ROM used but faster executing code loops but sometimes many more loops used.

If however total execution speed is of most concern and ROM size is not an issue, you can have two blocks of code, one designed for small shifts one for large.

If you are only going to shift up by one bit, you can do a clear carry then shift with just ADDwC, INC and Branch on carry instructions in the array then a final INC if required to wrap the rotated bit back in at the bottom. On some CPUs without decent shift instructions you can end up using a few less CPU cycles.

You can also do it as a fixed period trick in a similar way to a barrel shifter but it’s rarely worth it in software, as on average it takes longer…

As you’ve probably gathered indepth consultation of the data sheet will alow you to calculate which way uses the least number of CPU cycles and sometimes code.

Oh and on more modern 8bit CPUs avoid “Memory bit set” and “memory bit clear” instructions and also be suspicious of the equivalent in registers. Because on many occasions the microcode is fairly inefficient RTL and does an AND mask and OR set rather than just the one you need. The only real advantage is in effect a “hidden register” thus if you are running short of registers…

It’s all this sort of stuff high level coders have “abstracted away” in return for some realy quite bad inefficiencies. Especially with complex pointer arithmetic in which way to many people try to be “clever” and make WOC that is never ever going to get ported to a different OS let alone hardware architecture.

Clive Robinson • September 28, 2018 5:26 PM

@ Hmm,

With regards LoJax what suprises me is how long it has taken to appear.

@Nick P, RobertT and my self discussed malwarecin Flash ROM in quite some depth on a number of occasions including the questions arising about BadBIOS. And then there was Lenovo’s consumer level laptops etc ariving with malware built into the BIOS you could not easily remove.

The method and pieces of code were all waiting to be bolted together.

After that it was always a question of when not if it was found in the wild, I’m just suprised by how long it’s taken…

Clive Robinson • September 29, 2018 10:37 AM

@ Ratio, vas pup,

one of the “skeptics” will suddenly “remember” another piece of evidence or some question or other we really, REALLY, REALLY *MUST* deal with before we can say anything about who the perpetrators are.

Some what over dramatic me thinks, an attempt at oratory not logic.

The reality is law is about method and processes, not belief. Policing on belief goes by the more common name of “vigilantism”, with the “He’s black, the bible says black is the sign of evil, hang him quick”…

The method of law not emotive belief is reasoned legislation, that is based on over a thousand years of experience. From years of “trial by combat” where skill with weapons got replaced with the far nastier skill with words or oratory, with all it’s “fake evidence” potential.

To balance this oratory a reliance on the interpretation of physical facts became part of the process, but even this had problems due to eye witnesses who earned a living by giving false testimony who because of the way they advertised their services were called “strawmen”. Yes we still have problems but we are I hope still moving in the right ditection.

However the oratory “fake news” stage is the situation we are now in with the Salisbury nerve agent poisonings, and we need evidence to move us past it…

Publically we have credible suspects but why? Well a variation of Oratory, that is we have been given cherry picked information and words by the UK authorities, could it be “fake news”? There could be political advantages in it as we have seen with the US. Pit simply we actualy do not know, which is very important. Because what ever it is it has made many people jump the method and process of justice and just “believe” the persons in the pictures are “Them wot dun it”.

But the reality is at best all we have been given by the UK authorities is “chery picked” thus weak circumstantial evidence. But as you should remember I said in response to your posting at the time it was the first evidence. That is like any journey the method is the map but the process has a first step which we hope is in the right direction.

Now we have an interesting state the next evidence that we have publically is Putin and two persons appearing on Russian television making claims. Is it “fake news” again we didn’t know but it’s fairly certain most Russian’s believed or wanted to believe it was not fake news.

Since then we’ve had an investagative news source saying that one of the people on the program was a GRU agent that had been given Russia’s highest medal personally by Putin. Has this been verified, apparently so to some level. Though now not unexpectedly there are others in Russia claiming the opposite, for various reasons. Which is the “fake news”?

In reality What you might chose to “believe” is immaterial as the next piece of news or evidence might change it. Which is just one reason why you need conformatory independent evidence for “beyond reasonable doubt” for the required method and process… Otherwise you are no better than a vigilante working themselves up for a “hanging party”.

Predictably Russia is now claiming what has been said by the investagative journalists is “fake news” with out unsuprisingly offering any “evidence”, this was very much to be expected, as Russia under Putin has a history of doing this over and over ad nausium.

Because at this stage the Russians are “playing to the home audience” and trying to kick the issue into the long grass, so situation normal for politics not just Russian politics.

The only way you can defeat “fake news” tactics by either side is not to work by “belief” but “evidence” produced by time tested “process” from time tested “method”. The down side of being a member of the public is the governments involved these days rarely hand out evidence, as it’s not in their interest to do so.

It realy is that simple, when you take the surface politics out of it, a choice. Historicaly a belief driven oratory system which we would describe as a “witch trial” which in reality is a legitimised “vigilantism” for entertainment. Or a time tested and refined “method” and it’s “processes” that have over a millennium or more given us the closest thing to “justice” we have.

Which do you personaly want the entertainment of a much discredited medieval belief system that was compleatly abused by people with personal agenders. Or a system that atleast gives an opportunity to seek out the truth even though it has faults, and needs improvment.

Then put yourself in the position of the accused that is you are the person facing the “hanging party” or the considered opinion of your peers evaluating as best they can what factual evidence there is?

If however you were to ask me not what my reasond opinion which is “wait for the evidence” but to take a little wager on what I think most likely, then my answer might supprise you.

It is that there have been quite a few untimely deaths of Russian’s in the UK who Putin and his friends would like to be rid of. For whatever reason the UK Police have put them down to natural causes and accidents. Taken in issolation each untimely death might have been an accident or natural causes. However collectively, I would guess somebody is getting away repeatedly with murder. Which if you check I’ve already said a number of times on this blog.

So the thought in my mind is if it is Putin “Why make it obvious?” that is what has he to gain by two obvious assasinations in just a few days one in Salisbury and one in New Malden just a few days appart?

So far we’ve heard next to nothing about the New Malden murder why?

Now a problem arises, were there two independent assasination teams or one?

Which raises the question of why two teams? If you are getting away with murder up in the range of twenty plus in a couple of years, why use more than just one team? Multiple teams means multiple tongues that could wag… Which would be undesirable from a security asspect.

So whilst I’ve a belief Russia under Putin has been getting away with multiple assasinations in the UK, I also believe the UK government are hiding quite a bit as well, for political reasons. Especially as the woman who is now the UK Prime Minister was directly responsible for the UK Police Forces at the time the assasinations were taking place, and she was being a very busy beaver slashing resources not just to the Police but also to the forensic resources they are very much dependent on.

Who do you think the British Public would behave if say the Evening Standard ran an article under the banner of “PM May responsible for 20 murderers escaping Justice” or more punchy?

Hence my thoughts about “Fake News” from both the Russian and UK governments…

I suspect the key that will unlock the answers is the New Malden assasination, which suprise suprise we are not hearing anything…

Clive Robinson • September 29, 2018 10:26 PM

@ Hmm,

I didn’t realize it was directly borrowing the CODE from Lojack, I thought just the vector…

Well as everyone knows, code reuse is the way the software industry is supposed to get efficiency 😉

Actually the reuse of not just old vectors but old code as well in malware is something we are starting to see more of.

The implication is somebody is not doing their job somewhere. Which if you think about it for a little while is rather worrying.

Look at it this way the software vendor is at the root of a tree of responsability the leaf or end nodes are the individual machines. The further down the tree of responsability the somebody who is not doing their job is the smaller the number of viable targets under their responsability.

For it to be viable to reuse code in malware, the larger a target population is the more viable it is to do.

So at the bottom of the tree of responsability an attacker would need a very large number of “somebodies” not doing their job, but at the very top of the tree just one somebody is enough.

One implication of this is that the software vendor is in effect “re-opening closed vectors”… If that is the case then the question of “by Accident, by negligence, or by intent?” arises.

A similar question applies to nodes further down the tree. For instance we have the example of the decision by the UK Government Minister in charge of the NHS chosing not to pay for extended support for Win XP. Thus knowingly and for reasons of “political mantra” leaving tens of thousands of computers that it was not possible to upgrade but could not be replaced for the same reasons vulnerable…

All of which is rather worrying.

Clive Robinson • September 30, 2018 1:03 AM

@ Hmm,

Going in and out of business is too cheap to be serious about security, bottom line I think.

From the security POV yes, and quality likewise.

But there is a counter argument about “inovation”. You will find people arguing that keeping the cost of trying and failing low as well as not stigmatizing failure alows inovation to not just succeed but happen frequently thus providing significant growth etc.

The two POV’s appear to be in opposition, thus is there a middle ground or alternative.

Honest answer I don’t know, but I think it should be a high priority for society that we seek to find one.