SHA1 is Shattered

jiggawatt · February 2017

Guys

SHA256

just use it

shasum -a 256 filename to get a sum

debate done

bsdguy · February 2017

@ricardo said:

bsdguy said: "Sha1 is broken!!1!³"

Bit cheeky. Finding a collision is fairly fundamental, and Google (in their wisdom) state

We hope that our practical attack against SHA-1 will finally convince the industry that it is urgent to move to safer alternatives such as SHA-256.

The expert you quote recommends the same.

Nope not cheeky but grinningly persiflaging those who run around like headless chicken in panic.

And: Sure, a successful collision attack on any cryptographically used or used in sensitive areas hash is always a serious matter.

However:

Those were naked hashes (no salt)
That attack needed very serious resources. Few enough that not only nsa could run them but still more than script kiddies and even most criminal gangs can command over.
It was a collision attack (not breaking irreversibility).

While I fully agree with Google and, particularly considering what kind kind of computing is available even in mobile telephones nowadays, also strongly advise to upgrade to better hashing ... this is NOT A DESASTER.

Why? Because of where and how Sha1 is used. The most grave problem is Sha1 hashed certs, because this attack means that one can, with considerable resources, create fake certs. That said, I personally know of no CA still using Sha1 (I might be wrong, though; I'm not in any way deep into CAs).
Another collision related problem is on the fly hashing as is often used in tls channels. This danger, however, is very, very low for two reasons: 1) The session will be dead since days once someone managed to fake a single network packet. 2) For that to work one needed to have the session encryption broken in the first place.
And finally passwords. That is somewhat in the middle. On the one hand passwords are usually relatively long lived (i.e. much longer than the attack needs) but on the other hand they are always salted and the attack doesn't address that at all plus to break passwords one needed to break irreversibility (rather collision resistance). So I'm not worried.

ricardo · February 2017

First off, no one is going around like headless chickens in panic, that is you and only you conflating what you disagree with.

Secondly, it's massively dangerous to assume that you know every case where someone would use SHA-1 encryption. You only cite examples you can think of. Quite simply, the correct response is to us an alternate algorithm.

bsdguy · February 2017

@ricardo

Yet another funny guy trying it? Try your mindless bullshitting with someone else. I'm not the right man for that. Unlike you I don't just quote and re-blabber but I actually know what I'm talking about.

No headless chickens in panic? Obviuosly you missed quite some headlines and feel self-important enough to know better than renowned experts who also wrote what I said (albeit more polite).

ricardo · February 2017

@bsdguy, you come across a guy that has X years experience of something that thinks he knows everything.

I'm quoting specifically you. You, projecting that people are naively talking about this subject, and you aren't.

Honestly, someone that says "nothing in C can be scientifically proven" is someone who talks out their arse.

bsdguy · February 2017

@ricardo

"nothing in C can be scientifically proven"

Are you now desperate enough to invent things?

I did not say that. What I said is that C code can not be proven correct as there are ambiguities in the language.
Obviously you don't have the slightest idea what you are talking about and have never looked at the source of a parser, let alone at static verification.

I come around as someone who thinks he knows everything? Well, maybe that's because I do know what I'm talking about and don't have a big snout like you regarding fields I don't know much about?

ricardo · February 2017

bsdguy said: Are you now desperate enough to invent things?

You said so yourself, in an earlier post, muppet. Can you not stand by the stuff you post?

bsdguy said: I did not say that. What I said is that C code can not be proven correct as there are ambiguities in the language. Obviously you don't have the slightest idea what you are talking about and have never looked at the source of a parser, let alone at static verification.

What on earth are you on about? C compiles to machine code, like every other language. Surely it comes down to mathematical axioms and considering that C is a turing complete language, your statement makes fuck all sense, in the context you say it.

I come around as someone who thinks he knows everything? Well, maybe that's because I do know what I'm talking about and don't have a big snout like you regarding fields I don't know much about?

No bother, prissy boy. I am pretty much indifferent until someone plays the big shot... clearly you are at the limits of what you know and making it up as you go along. Talk all you like, disinterested.

bsdguy · February 2017

@ricardo

Wow, doggy gets pissed and tries insulting me. Doesn't work though.

As for my statement: Bring the quote.

"What on earth are you on about? C compiles to machine code, like every other language. Surely it comes down to mathematical axioms and considering that C is a turing complete language, your statement makes fuck all sense, in the context you say it."

A terriers way to say that he doesn't have the slightest clue what he is talking about.

You see, you seem to be used getting your way because you are insisting and stubborn like a terrier. Won't work with me.

I know those social rules, "let live", "let the other side keep their face", and I usually value them and make sure that the other side has a chance to get out cheaply and without losing their face.
You, however, terrier, had plenty chances and stupidly chose to bet on chuzpah and stubborness. You asked for it, terrier.