Timing attack - stupid questions thread.

[Hack_Bird]

  Great work!

  Next week I get my 360 back and going to "mod" it for sure ...
and then wait <or try myself> find a damm exploit

[ace]

Niiiiiiice! 

[ForSwitch]

great work Karl
www.hardstore.com for all your infectus needs

...and spend $45 to ship a $33 chip to the States...no thanks.

I share sentinel0's sentiment.  Shame there are no distributors in the states.

[lasonnette]

yeah, what do you expect as a price for express delivery? try regular shipping method

[sentinel0]

ForSwitch if you go to the reseller link on the infetus page mrmodchips has them the cheapest 10$ shipping to the states

[tumba]

Great work! Awesome really!

Congratulations

[mrblack1134]

Great work robinsod!

Now what happens next? MS pushes a new kernel, changes memcmp()'s implementation and blows another efuse? I guess we're fair to assume that the next kernel they push on live will plug that hole (as they were pretty prompt to fix the hv exploit).

Are we ok if R6T3 is gone or should we make sure not to upgrade to any new kernel?

Again, great work.

[Protonus]

Are we ok if R6T3 is gone

That's something I'm wondering too.. should I remove this while I have my 360 open soon?  I plan on getting an infectus chip today in preperation of being ready to do this mod.

[rgh2468]

awesome work kudos to all involved!!!!!!now for my stupid question.As I understand it this attack relies on the memcmp comparing byte-wise not dword.I have kernel 5759 is byte-wise still used in this kernel?again this is great news. Maybe we will see homebrew on the 360 after all.Hell Yeah!!!!!!!!!!!

[arnezami]

Just to be clear, the timing attack will allow you to downgrade to 2.0.1888. You can then upgrade to 4532 & run the KK sploit and obtain your CPU keys. You should be able to replace the original CB after the upgrade (this needs to be confirmed) and then the only "clue" to what happened is that you may have 1 or 2 more burned eFuses for the HV/Kernel version you are running

Since every upgrade burns 1 more fuse wouldn't it be possible to simply skip an upgrade or better: upgrade yourself (by encrypting an updated fw from another xbox with your own keys/kv) to avoid having "too many" burned fuses? Sure the fuse burn by 4532 is unavoidable (not counting the resistor removal) but doing the downgrade attack just before a new kernel version would do the trick since you can prevent the additional fuse to be burned when upgrading yourself. MS wouldn't notice a thing especially when using a dual flash.

Regards,

arnezami

[SeventhSon]

Now what happens next? MS pushes a new kernel, changes memcmp()'s implementation and blows another efuse? I guess we're fair to assume that the next kernel they push on live will plug that hole (as they were pretty prompt to fix the hv exploit).

Fortunately we've got some warning. So if anybody thinks that he/she will be able to pull off the timing attack hack when the end-user schematics and code goes public (i.e. if you can solder a basic circuit together and generally have some clue), then they should hold off any upgrades until either a) it's confirmed that the upgrade hasn't fixed the memcmp function or b) they can implement the timing hack and get their fuse data.

Kev

[SeventhSon]

Also congrats Arnezami. Nice idea ;-)

[arnezami]

Also congrats Arnezami. Nice idea ;-)

Thanks SeventhSon .

Just to clear up:

MS cannot fix this problem by simply changing the memcmp function in a future kernel version. Thats not gonna help them. The weakness is that the byte-wise memcmp function is in the 1888 kernel/bootloader (and they cannot change that one anymore of course).

There is some other stuff they could do, but I'm pretty sure making a backup of your flash now (link) will allow people to downgrade at any time in the future no matter what they throw at us.

Of course a new kernel version should still be checked for "the other things" they could do to prevent downgrading for those that do not have a backup of their fw yet...

Regards,

arnezami

[xordef]

Well for a start all new xboxes will come with the latest memcpy-patched dash from the factory - with no possibility to corrupt and thus boot an earlier version. That change should not take more than a month from now, or if they understood that this hack would work (I'm amazed as to the sheer incompetence in their crypto team. It's crypto 101 to randomize the response time when performing true/false calculations on passwords) they might be closer than that to do it.

[Geremia]

always good reading on this forum, much respect to all 

btw, i saw that modchip.it has infectus at -30% discount (till 27 august) and ships worldwide

[Sc0rpion]

I came to this thread expecting to see a Timex holding a Rolex in a half nelson!
Last time I rely on google for my fetish fix

while i'm in here great work everyone

[TheSpecialist]

Again, congratulations to Arnezami for his great idea. I love it when a plan comes together 

Now people can finally get serious about homebrew. hopefully Team XBMC will reconsider their Linux idea and code an XBMC that can be booted directly.

[Ced2911]

now we need sound under linux

[tmbinc]

sure, microsoft can change the 2BL, and burn a fuse (of the fuseline 2) so that an old 2BL doesn't work anymore...

But that means that we have 10 mio. potentially  exploitable boxes out there *at the moment*. No more excuses for people to not being able to get an exploitable box. It's different than last time. If you update now, it's your fault, not anyone else's. You have been warned.

btw, i disagree with xordef. If you design a security solution, you take first define what kind of attacks you want to be resistive against. It seems that some kind of hardware attacks (including timing attacks) weren't part of this. For example, the "beq"/"bne" in the signature check is a single instruction, not protected against hardware attacks like glitching. If you would want to be (more) resistive against that kind of things, that's no problem: just compare each data twice or so, making successful glitch attacks much less probable. but it wasn't designed that way.

The problem wasn't the implementation, the problem was a consciously taken design decision. A bad one, as it seems.

[arnezami]

Again, congratulations to Arnezami for his great idea. I love it when a plan comes together 

Now people can finally get serious about homebrew. hopefully Team XBMC will reconsider their Linux idea and code an XBMC that can be booted directly.

Thanks a lot for believing in it.

I would also like to personally thank Geremia for his major contribution to this attack. He has made many of us (well at least me) really believe in the practicallity of the attack when measuring the time differences with his so called "cheap USB LA". His results were earth shattering. Anyway. He seems to have a feel to be at the right place at the right time...

Thank you Geremia . Again I might add.

Also to SeventhSon for (among other things) RE-ing the CB-auth algo. And there are others too (like tmbinc for sticking with the idea based on its merits). I think its fair to say XBH has some hard working and smart people. This is a good place and the mods take good care.

Regards,

arnezami