Slashdot Mirror
Playstation 3 Code Signing Cracked For Good
ReportedlyWorking writes "It appears that Sony's PS3 has been fatally compromised. At the Chaos Communication Congress in Berlin, a team named 'fail0verflow' revealed that they had calculated the Private Keys, which would let them or anyone else generate signed software for the PS3. Additionally, they also claim to have a method of jailbreaking the PS3 without the use of a Dongle, which is the current method. If all these statements are true, this opens the door to custom firmware, and homebrew software. Assuming that Sony doesn't take radical action and invalidate their private keys, this could mean that Jailbreaking is viable on all PS3, regardless of their firmware! From the article: 'Approximately a half hour in, the team revealed their new PS3 secrets, the moment we all were waiting for. One of the major highlights here was, dongle-less jailbreaking by overflowing the bootup NOR flash, giving complete control over the system. The other major feat, was calculating the public private keys (due to botched security), giving users the ability to sign their own SELFs. Following this, the team declared Sony's security to be EPIC FAIL!'"
"Following this, the team declared Sony's security to be EPIC FAIL!"
Is it really necessary for everybody to talk like complete dicks nowadays?
Epic Fail? WTF?
How many years has it taken to crack the PS3?
I'd say that Sony has done a remarkable job.
Does it go on forever?
It's a bit late to invalidate private keys.
My understanding is that every PS3 game is signed with those keys. Therefore, invalidating them through a firmware update would mean that every PS3 game to date will no longer work.
While I wouldn't put it past Sony to try this, this would result in not only massive lawsuits, but also would be a massive PR blunder.
Having said that, there could in theory be some sort of additional key telling what date a disc was signed, but even if that were true, it would be trivial to work around.
GLaDOS for President 2016! "Well here we are again. It's always such a pleasure." -- GLaDOS, 2011
From the blurb:
'Approximately a half hour in, the team revealed their new PS3 secrets, the moment we all were waiting for. One of the major highlights here was, dongle-less jailbreaking by overflowing the bootup NOR flash, giving complete control over the system.
Ok, the PS3 was launched on November 11, 2006. Today's date is December 29, 2010. That means that it took over four years to be broken.
Compared to DVD and Blu-Ray, that is actually pretty darn good.
Weaselmancer
rediculous.
That was intentionally done for game developers.
It is impressive indeed. Though I do note that it didn't completly resist attack for four years. It just took for years to be completly, irrepairably and conveniently broken. There have been wayst o break the PS3s DRM for years, but their complexity put the beyond the ability of all but the most technologically capable users. With the code-signing cracked, it's as simple as burning an ISO.
How did Sony fuck that one up?
It was my(admittedly layman's) understanding that a public/private key crypto implementation, assuming it isn't deeply flawed, using key lengths suited to the computational capacities of PDP-8s, or otherwise totally fucked, was mathematically secure against anything other than a profound breakthrough in prime factorization algorithms, an unbelievable advance in computational power, or an insider leaking your private key.
With stuffy like HDCP, it was understood that serious tradeoffs were made in order to make the crypto cheap and fast enough that any POS $200 monitor should be able to decode an encrypted bitstream fast enough to handle the demands of uncompressed digital monitor connections. The weaknesses just came with the territory.
With something like the PS3, though, they have serious computing power available, and were dealing with a straightforward case of "verify that the code signed with private key X has indeed been thus signed, and not modified since, using public key Y, from which private key X is essentially not computable". Virtually every real-world use of cryptography depends on the ability to do that without disclosing your private key(save by malicious insider/hacker attack).
What did Sony do wrong? Obviously, they could do nothing about a suitably well-equipped hacker physically modifying a PS3 to stop it from verifying at all, or to always return "yup, all good" regardless of the verification outcome; similarly, a firmware bug could allow the same outcome without the expense of physical modification; but how could it be that they would have to put anything in their client(no matter how well hidden by hardware obfuscation/TPMs/smarcards/whatever) that could be used to compute their private key? Isn't a public key, which is a totally safe piece of data to disclose, all you need to verify whether or not something has been signed with the matching private key?
I admit that I don't have a deep understanding of this stuff; but it seems like this is the equivalent of "Hey, possession of the list of trusted CAs and their public keys has allowed a hacker with a copy of firefox to compute Verisign's root signing keys!".
How did Sony fuck up such that this story is not the biggest breakthrough in cryptoanalysis since frequency analysis?
I wanted to commit a PS3 to biomedical research on a project of MY choosing, as well as play LEGIT games but that was taken because ... well it doesn't matter as it's too late now.
Ok, the PS3 was launched on November 11, 2006. Today's date is December 29, 2010. That means that it took over four years to be broken.
Compared to DVD and Blu-Ray, that is actually pretty darn good.
I was at the presentation in Berlin today. They did bring up this exact point.
Their counter argument was that people don't take into consideration that the console did support homebrew until Sony declared they'd drop that. The argument for that action was they'd save money not having to support it for their then-new PS3 Slim models, which turned out to be bullshit after hackers discovered that the Slim (with some hacking) could actually run the same Linux distros as the PS3 Fat. They then disabled OtherOS on the PS3 Fat, too.
This was 12 months ago (can't cite a source other than the slides), making it take only 12 months of actual effort for it to get cracked, as opposed to other (closed) platforms where the homebrew hacking efforts begin at day 0.
That's true. And Sony have been boasting of having the toughest DRM of all consoles.
However, it only took half a year from removing Linux support, and in that short period have had many partially successful attacks against it. Before, while they had the Linux support, such stories were remarkably rarer.
Many critics meant that the continued security of the console was partially because they allowed linux to run on it, and so many of the talented people had no reason to look closer at it. Since PS3, after four year of "DRM cracking almost never heard of" have now gone to "Completely broken" in just over half a year's time, I think they have a point there.
It's not that it was that much more secure, it's just that most of the really talented people had no reason to look into it.
It's The Golden Rule: "He who has the gold makes the rules."
Folks toss about the phrase "Epic Fail" far too loosely. Here's what a real Epic Fail looks like:
The DRM code has a bug that, when a certain condition happens (time passes, specially-formulated packet received, etc.), it overclocks the CPU to the point that it catches on fire. Within minutes of the event, most of the millions of PS3s in the wild have set peoples' homes ablaze.
As a result, thousands die and the insurance industry collapses. Anarchy reigns, so there's nobody to enforce copyright anymore and the original DRM is rendered irrelevant.
THAT is an epic fail.
"Time flies like an arrow; fruit flies like a banana." --Groucho Marx
From @fail0verflow:
"we only started looking at the ps3 after otheros was killed."
and
"our goal is to have linux running on all existing PS3 consoles, whatever their firmware versions."
If Sony would have left OtherOS alone, they wouldn't be in this predicament.
They don't have Sony's signing key, from what I've read. What they have is a flaw in the key generation process, which allows them to generate valid signed packages without the private key. In fact, here's the video from the conference itself: http://www.youtube.com/watch?v=GPjd6gHY6A4
No, GP was right. The exact signing key used by Sony may be derived from the public components of their ECDSA signatures. Not something close; not something equivalent.
So does this mean a hypervisor free linux is around the corner? I may change my stance on buying a PS3.
Sometimes, life itself is sarcasm...
In other words, Sony has just gone and proved that the only DRM that remains unhacked is the kind that nobody cares to hack. See also: SACD.
SACD is cracked. Or at least worked around enough so that it doesn't matter.
There are two hacks for SACD:
1) Physical modification of various players to extract the PCM audio after conversion from DSD, this approach is a few years old now.
2) The widespread crack of HDCP enabled extraction of the original DSD audio from any HDMI equipped SACD player.
There are plenty of SACD rips floating around the net
When information is power, privacy is freedom.
Which then means any ps3 not connected to the internet cannot play new games. That would be epic fail.
I hope XBMC will be ported to it now.
ayottesoftware.com
And that helps:
people who bought their PS3's before Sony manages to rush a new firmware image through the factory, and who hold back their online updates before Sony manages to rush a new one through the update system. Remember, if they can update the signing keys, they can also update the key checking code, so there's no reason the second key has to be as easily compromised as the first.
Anyone who can emancipate their PS3 in this (presumably) short window of time is gonna be able to keep their PS3 well-stocked with spoofed updates from this day forward. But this doesn't break all PS3 security forever.
DRM: Terminator crops for your mind!
That's all I want, badly, very badly.
It's half the reason I got the PS3 when I did, XBMC was in the early stages of PS3 support, however the idiots at Sony blocked the GPU acceleration for the video so the team abandoned it once the 3D loophole was closed in linux. I don't know the full term, something along the lines of a hypervisor.
Then they closed off linux all together.
I love it as a gaming machine but I wish it could match my Xbox1. The Ps3 hardware is amazing, XBMC would be brilliant on it.
Unless they can get every publisher to send the hashes for every version of every game they have sent to the CD press, some people will find their games broken
But Sony already possesses them - they had to sign them in the first place! Either that or they entrusted all those publishers with with their private signing key. Which would be a terrible idea.
DRM: Terminator crops for your mind!
The PS3 was being attacked well before OtherOS removal. When linux was available the graphics on the machine were limited to virtualization. The race was on too crack the 7 locked down SPUs. Were people successful? Mostly no, but that doesnt mean attempts havent been made. If i remember correctly, Geohot's intention was to gain access to the cores. They just happened to find an exploit to give them keys to the kingdom
Removing linux definitely brought the talent out of the woodwork, but it did not start a war
I did consider that possibility myself, but I don't think it can be done perfectly. They can include a list of hashes for all the big games, but think how many games there are - and then they have slight variations by version, by region, and so on. Unless they can get every publisher to send the hashes for every version of every game they have sent to the CD press, some people will find their games broken. Sony might consider that a price worth paying.
It is almost certain that the process of signing the games includes a hash generation. Usually the way these things work is that you hash the entire image, then encrypt the hash with the private key. Mainly because hashing is orders of magnitude faster than encryption. So even if Sony didn't archive a copy of every game they signed due to laziness or lack of process or they ran out of shelf-space, it would have been trivial to archive a copy of every hash that they signed.
When information is power, privacy is freedom.
> Do they really have Sony's signing key?
Yes, we have most of their signing private keys.
This is exactly the only possible fix. It is, however, technically quite hard to pull of for a number of reasons. I'm not at all certain that Sony will do that. They need to build a hash list of every version of every game, package, downloadable cotent, deal with shop versions and stuff like that, etc...
Assuming they don't botch signing with the new key, no, we don't. The code running on the PS3 is perfectly fine (the signature verification, that is; the rest of the security is a clusterfuck). So is the way the signature is implemented. The screwup is in Sony's signer code. If they fix that and only issue safe signatures from now on, we can't compute new keys.
But because we can downgrade and due to the oracle attack on the secure SPE, this will likely not gain them much.
Sony cannot permanently regain any existing PS3 with a firmware update (nor can they fix this hole trivially at all, including in new manufactured units). They can make it harder for you to install a hacked firmware on a PS3, but as of today every manufactured PS3 is vulnerable to a modchip (NOR/NAND flasher) forever.
That 512MB RAM with proper coding acts very much like 2-4GB of DDR2.
Size isn't all that matters, you know. The bandwidth alone is insane.
As for outdated? The newest AMD/nVidia GPUs are just now touching the PS3's theoretical performance of 2TFLOPS.
The hypervisor slowed things down. Without it, and with direct access to the other locked SPE and full RSX access, the system in itself has the capability to perform on par with current PC systems of high-end gaming spec.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
Now all they have to do is crack the "having to buy an overpriced piece of proprietary hardware that merely replicates what the PC I own can already do" part of the equation.
Crack that, and I'll be all set. Otherwise I'm not spending several hundred dollars to buy a box to take up more space simply to play software that my PC would be able to play if it weren't for someone's desire to complete control and every last dollar.
Change from a hardware/software company to a software company, and I'll use your product.
This space available.
Nintendo had a nifty solution for the old Gameboy(/color) - code wasn't signed, but games did need to have [...] the Nintendo logo
Typography is not copyrightable, and a U.S. trademark cannot be used as an ersatz copyright or patent. See Dastar v. Fox, and especially Sega v. Accolade.
...
everyone seems to see it as a fail on behalf of Sony . Isn't this IBM's Cell at fault ?
The Epic Fail, exposing Sony's private key, had nothing to do with the IBM Cell processor. In fact the flaw was not in any of the PS3 software. It was a mistake in the program used to sign software approved to run on the PS3. That program presumably runs only on some highly guarded server in the bowels of Sony. It could have been fixed by adding one line of code, a call to random number generator to generate a new random value for each signature. Even a crappy random number generator would probably have resisted attack. All that was needed was keeping attackers from finding two different signatures that used the same "random" number. You have to go back to the Venona NSA exploit in the Cold War to find an example of a large organization screwing up what should have been an unbreakable cipher system.