Content Installation

Actually, you're presentation at 25C3 made more sense about altering the data so the first byte is null.

Does the strcmp from the Wii check lengths as well? I would assume so, and if it does, you still need to get the second byte on the real hash to equal 0.

Is that really what Datel did? Hmmm, maybe they found an easy way to get the 2nd byte of the hash to 0. Otherwise it wouldn't make sense that they just wouldn't try and get the first byte.

By the way, what was the Nintendo's fix to the fakesigning bug?
Is there still exists a silly mistake like the initial twilight hack removal code?
Did they actually replace the strcmp with memcmp?

Quote
WiiCrazy
By the way, what was the Nintendo's fix to the fakesigning bug?
Is there still exists a silly mistake like the initial twilight hack removal code?
Did they actually replace the strcmp with memcmp?

How else would they fix it? It's not like they need to write any new code or anything--there's nothing to screw up here.

They did some updates to the RSA checks. Yes, it's using memcmp now and they are also checking the padding now. It looks like a real secure RSA implementation this time and I don't think we're able to exploit anything there :(

Quote
tona

Quote
WiiCrazy
By the way, what was the Nintendo's fix to the fakesigning bug?
Is there still exists a silly mistake like the initial twilight hack removal code?
Did they actually replace the strcmp with memcmp?

How else would they fix it? It's not like they need to write any new code or anything--there's nothing to screw up here.

Maybe sampling some fake signatures and just adding a few extra checks for them... I don't know, if I did know some arm assembly then in the first place I wouldn't have asked that question...

What we are talking about is the last step of a huge process... maybe the easisest part of it... I can't see any reason they might still fail in the other parts of that big picture...

A thought just occured to me. Please stop me if you see me go wrong.

1. The Wii has a private and public key
2. Nintendo have a private and public key
3. Nintendo signs content with their private key
4. The Wii validates signatures against Nintendo's public key

This means that Nintendo's public key must be registered within the Wii.

Would it not be possible to add a new "trusted" public key from a keyset we generate to the Wii's system? Then we could sign our own content properly.

IIRC the wii's public key is in OTP memory so it can not be reprogrammed.

That's not what I meant. Let me try and explain a different way.

From what I remember of PGP, you create a keyset (Public and Private). You distribute the public key to your friends, and you receive their public keys. You add your friends' public keys into a "key pool" so you can easily encrypt stuff for them and validate their messages.

Now I'm working on the assumption that the Wii has a "key pool", and it has Nintendo's public key listed in it (in order to validate incoming code signing). Could we not add a Homebrew public key to this pool so it would then validate code signed by either Nintendo or Homebrew.

It's not that you want to remove or edit Nintendo's public key (because you'd still want to be able to play official games), so regardless where this is stored doesn't matter.


I did have concerns on the piracy side of this, but considering they can already pirate everything anyway, I thought I'd ask.



Edited 1 time(s). Last edit at 01/04/2009 08:55PM by whodares.

You can do two things with RSA:
* Encrypt data with your public key and decrypt it with your private key.
* Encrypt data with your private key and decrypt it with your public key.
Signatures are basically hashes encrypted with the private key.
Nintendo's public key is stored in boot1, boot2 and every IOS. You could easily modify it given that you can install titles on the NAND. But why would you want to that then?
You can already run code on both CPUs. Adding a homebrew key would just make things way more complicated. It's easier to just patch out the signature check or make it always return true (1 byte patch) .
And you can't just magically add a new key when you can't install titles on the NAND. It therefore doesn't help you to run your own code.



Edited 1 time(s). Last edit at 01/05/2009 12:28AM by svpe.

Put more simply: The Wii doesn't have a keyring of public keys it can verify against. It has one key. Adding support for multiple keys is harder than removing the signature check.

Quote
bushing
Put more simply: The Wii doesn't have a keyring of public keys it can verify against. It has one key. Adding support for multiple keys is harder than removing the signature check.

Fair enough. Was just an idea :-)