5 ms·

Given that the CDM is sandboxed, does this mean that it should be significantly easier to recover encryption keys from it given that you could effectively run t

Given that the CDM is sandboxed, does this mean that it should be significantly easier to recover encryption keys from it given that you could effectively run the CDM inside of a tracing emulator?

adrusi 11y ago
Perhaps. I don't think anyone who can think of that attack vector would really care because they'd also understand that their DRM will be broken as soon as anyone really wants the data it "protects" because it's a futile endeavor.
nekitamo 11y ago
The encryption keys would probably be whiteboxed (http://www.whiteboxcrypto.com/ http://www.whiteboxcrypto.com/) and thus you would have to break the whitebox scheme in order to recover the unprotected keys.
- Nursie 11y ago
  That is kinda interesting but.... "This means that an attacker can easily ... use any kind of attack tool such as IDA Pro, debuggers, emulators, etc." Attack tools?!
  - Aissen 11y ago
    I like calling those "Debug tools".
  - dmm 11y ago
    """ Dan had had a classmate in software, Frank Martucci, who had obtained an illicit debugging tool, and used it to skip over the copyright monitor code when reading books. But he had told too many friends about it, and one of them turned him in to the SPA for a reward (students deep in debt were easily tempted into betrayal). In 2047, Frank was in prison, not for pirate reading, but for possessing a debugger. """" -Richard Stallman in The Right to Read http://www.gnu.org/philosophy/right-to-read.en.html http://www.gnu.org/philosophy/right-to-read.en.html
    - joe563323 11y ago
      How True. What an insight. Fuck Intel.
- chongli 11y ago
  This seems silly. So why doesn't the attacker just rip out the whole of the decryption function from the binary instead of just the key? What's stopping an attacker from re-running that function in another program?
  - yk 11y ago
    Nothing white box crypto is just an academic exercise.
TD-Linux 11y ago
Not any easier than other browsers, such as Chrome's Widevine, for example.
pdkl95 11y ago
Today? It should be fairly easy to recover anything you want with a debugger, given sufficient time and effort. In future versions once SGX is available (>= Intel Skylake)? Good luck extracting data from an encrypted enclave. This is, after all, the intended purpose[1] for SGX - to create the "trusted computing" (DRM) that Microsoft has wanted for the last decade[2]. It's not like these new instructions are for the end-user; Intel has to authorize[3] you binary before it can be loaded into an enclave. Anybody using the CDM - even sandboxed - is helping to create that future. If this isn't fought and rejected now, yet another battle in the War On General Purpose Computers will be lost. [1] https://software.intel.com/en-us/blogs/2013/09/26/protecting-application-secrets-with-intel-sgx https://software.intel.com/en-us/blogs/2013/09/26/protecting... [2] "Palladium" / "Next-Generation Secure Computing Base" in the early 2000s [3] https://jbeekman.nl/blog/2015/10/intel-has-full-control-over-sgx/ https://jbeekman.nl/blog/2015/10/intel-has-full-control-over...
- MichaelGG 11y ago
  If this is true, fuck Intel. SGX had tons of promise. For example, you could run a provably secure Bitcoin mixer. Or any SaaS. MSR even published something to that effect (using SGX to run custom Hadoop jobs on their servers, without being able to see/modify your code). Why would they gimp their platform so hard? It's not like enabling arbitrary use of SGX prevents DRM from using it.
- walterbell 11y ago
  > Intel has to authorize[3] you binary before it can be loaded into an enclave. Wouldn't that make Intel liable for the actions of the binary? They would have to manage an application review process for any binary on any operating system, which sounds like an app store nightmare writ large. Not to mention revocation and repeating the entire process for security updates to software in the enclave. Could different governments require Intel to impose different blacklists on enclave software authors? It's hard to see why they Intel would open themselves to such complex scenarios, unless it's a trial balloon that precedes a federated model or some other method of distributing key authority to multiple app stores and/or the hardware owner.

Given that the CDM is sandboxed, does this mean that it should be significantly easier to recover encryption keys from it given that you could effectively run the CDM inside of a tracing emulator?

adrusi 11y ago

Perhaps. I don't think anyone who can think of that attack vector would really care because they'd also understand that their DRM will be broken as soon as anyone really wants the data it "protects" because it's a futile endeavor.

nekitamo 11y ago

The encryption keys would probably be whiteboxed (http://www.whiteboxcrypto.com/ http://www.whiteboxcrypto.com/) and thus you would have to break the whitebox scheme in order to recover the unprotected keys.

Nursie 11y ago
That is kinda interesting but.... "This means that an attacker can easily ... use any kind of attack tool such as IDA Pro, debuggers, emulators, etc." Attack tools?!
- Aissen 11y ago
  I like calling those "Debug tools".
- dmm 11y ago
  """ Dan had had a classmate in software, Frank Martucci, who had obtained an illicit debugging tool, and used it to skip over the copyright monitor code when reading books. But he had told too many friends about it, and one of them turned him in to the SPA for a reward (students deep in debt were easily tempted into betrayal). In 2047, Frank was in prison, not for pirate reading, but for possessing a debugger. """" -Richard Stallman in The Right to Read http://www.gnu.org/philosophy/right-to-read.en.html http://www.gnu.org/philosophy/right-to-read.en.html
  - joe563323 11y ago
    How True. What an insight. Fuck Intel.
chongli 11y ago
This seems silly. So why doesn't the attacker just rip out the whole of the decryption function from the binary instead of just the key? What's stopping an attacker from re-running that function in another program?
- yk 11y ago
  Nothing white box crypto is just an academic exercise.

TD-Linux 11y ago

Not any easier than other browsers, such as Chrome's Widevine, for example.

pdkl95 11y ago

Today? It should be fairly easy to recover anything you want with a debugger, given sufficient time and effort. In future versions once SGX is available (>= Intel Skylake)? Good luck extracting data from an encrypted enclave. This is, after all, the intended purpose[1] for SGX - to create the "trusted computing" (DRM) that Microsoft has wanted for the last decade[2]. It's not like these new instructions are for the end-user; Intel has to authorize[3] you binary before it can be loaded into an enclave. Anybody using the CDM - even sandboxed - is helping to create that future. If this isn't fought and rejected now, yet another battle in the War On General Purpose Computers will be lost. [1] https://software.intel.com/en-us/blogs/2013/09/26/protecting-application-secrets-with-intel-sgx https://software.intel.com/en-us/blogs/2013/09/26/protecting... [2] "Palladium" / "Next-Generation Secure Computing Base" in the early 2000s [3] https://jbeekman.nl/blog/2015/10/intel-has-full-control-over-sgx/ https://jbeekman.nl/blog/2015/10/intel-has-full-control-over...

MichaelGG 11y ago
If this is true, fuck Intel. SGX had tons of promise. For example, you could run a provably secure Bitcoin mixer. Or any SaaS. MSR even published something to that effect (using SGX to run custom Hadoop jobs on their servers, without being able to see/modify your code). Why would they gimp their platform so hard? It's not like enabling arbitrary use of SGX prevents DRM from using it.
walterbell 11y ago
> Intel has to authorize[3] you binary before it can be loaded into an enclave. Wouldn't that make Intel liable for the actions of the binary? They would have to manage an application review process for any binary on any operating system, which sounds like an app store nightmare writ large. Not to mention revocation and repeating the entire process for security updates to software in the enclave. Could different governments require Intel to impose different blacklists on enclave software authors? It's hard to see why they Intel would open themselves to such complex scenarios, unless it's a trial balloon that precedes a federated model or some other method of distributing key authority to multiple app stores and/or the hardware owner.