The first phase of crowd-funded audit of TrueCrypt has turned up several vulnerabilities, but nothing particularly amiss and certainly nothing that looks like a backdoor. iSEC Partners, which was contracted to carry out the audit by the Open Crypto Audit Project (OCAP), found 11 vulnerabilities in the full disk and file …
This is extremely good news for anyone who used Truecrypt to protect assets.
The problems found are comparatively minor, easily fixable, and indicative of competent people writing code just outside of their core competencies. That last issue happens on any project anywhere.
Phase 2 will look at the actual encryption code. Unless that is found to be seriously flawed, the insights found from the review will ensure that Truecrypt 8.0 is rock solid and approved by experts.
they could look at OpenSSL once they are finished here...
Look at OpenSSL!! Ha! Have you SEEN the code?!
I thought this too.
Then I actually spent an idle couple of hours trying to make sense of some of the OpenSSL code.
Now admittedly, my C is a bit rusty but I wouldn't wish the auditing of the clusterfuck of preprocessor directives and godawful coding style that is the OpenSSL codebase on my worst enemy.
If an audit of OpenSSL was 'crowdfunded' (gods, I hate that term) then you'd actually need *two* funds - one to actually pay the auditors to do their thing and another for an extensive course of therapy for the poor sods afterwards.
if there are "backdoors" to be discovered, I would expect to find them in the cryptographic code. since this has yet to be audited, it's too early to declare that truecrypt is kosher.
No you wouldn't. The crypto routines are implemented form a set of standard published algorithms. It is relatively easy to prove that encrypting string X with key Y in truecrypt produces the same output as anyone else's AES
There might be mistakes, there might be bugs which leave memory around and help you guess the keys, there might even be backdoors (if you are truly paranoid) in the original design of the algorithms that the NSA and Bruce put in there and the worlds cryptographers haven't spotted.
But backdoors in the code that allow an NSA passwd to decrypt anything are going to be in the keyhandling. The big concern fro truecrypt users was that on windows you need to run a signed driver which you can't (easily) build yourself - so you have no idea if what you were running was what they claimed.
The problems found are comparatively minor, easily fixable, and indicative of competent people writing code just outside of their core competencies.
Well... some of these errors (the messing with different data types - signed & unsigned integers etc) does not look to my (granted, inexperienced) eye as something a competent programmer would do.
However, yes, the problems do seem fairly minor (once again, no C dev here so I admit my relative ignorance - though I must commend the report writers on their clear explanations)
Phew thank god
My "collection" is safe from the girlfriend although she does have trouble using the PC at the best of times. Maybe it is a cunning façade.
Of course, the real problem is that this audit only took place years after the software has been available. This audit should have taken place on a beta version of the software, and then should be undertaken on a regular basis to ensure no security bugs are introduced with new versions.
... of course recognise the 'good practice' basis of your comment, given that TrueCrypt is free at point of consumption, who exactly would pay for the type and level of rigour you suggest?
Not in any way intended as a criticism, more a comment on certain limitations imposed on a product such as TrueCrypt and the developers behind it.
As The_Idiot says, where are they going to get the money from to do that on a regular basis? They would have to start charging for it, which would put many people off. Or a rich sponsor.
Doing a Kickstarter is a one of, because it was in the news. That won't work, if you are doing this every year or so.
Where are they going to get the money from? No idea. But it's funny how the 'having no money' issue constantly arises when you give your software away for free, and doesn't arise as often with companies that do not give their software away for free.
They could just create a new kickstarter, saying we want to audit X software package, anyone interested? Donate now! For popular packages and libraries they will get donations.
We are currently having security auditing, including source code done on one of our products. It ain't cheap!
Here's hoping that iSEC Partners are not already subject to a gagging order preventing them from disclosing that Truecrypt is riddled with NSA backdoors!
Indeed, given the environment, why contract an American security firm?
8. MainThreadProc() integer overflow
Is it just me, or does '8. MainThreadProc() integer overflow' sound extremely similar to the heartbleed issue?
Is there an overlap of developers between truecrypt and openssl?
Re: 8. MainThreadProc() integer overflow
Integer overflow is very different from buffer overflow (and to be pedantic, heartbleed is a buffer overread issue). It can cause mathematical issues (e.g., for a signed byte, 100 + 100 = -56), but it's not easy to turn that into a security flaw.
...worse than a colonoscopy
"Probe's nearly all the way in ... no backdoor hit yet"
At least with the colonoscopy the doctors find your backdoor *before* inserting the probe.
What was the thing they found named "_NSAKEY"?
Oh, sorry. My mistake. NSAKEY was found not in TrueCrypt but in Windows itself.
Don't trust iSEC or NCC Group audit
Why can't TrueCrypt be audited by a company that is not based in the UK or in the US - i.e. potentially subject to GCHQ or the NSA leaning on them? iSEC is part of NCC Group - a UK company. I would not trust their audit, not until it is also audited by a company based in a land far-far-away from UK or US.
Re: Don't trust iSEC or NCC Group audit
Probably because any company NOT in bed with the NSA or GCHQ is in bed with someone else. IOW, it's pick your poison.
- Game Theory The agony and ecstasy of SteamOS: WHERE ARE MY GAMES?
- Intel's Raspberry Pi rival Galileo can now run Windows
- Hello, police, El Reg here. Are we a bunch of terrorists now?
- Microsoft and HTC are M8s again: New One mobe sports WinPhone
- Worstall on Wednesday Wall Street woes: Oh noes, tech titans aren't using bankers