The Secure Enclave is essentially a hardware security module, in more general terms. The only thing that leaves its boundaries are the results of crypto operations, not the parameters that went into calculating them.
I'm going on and on about this because I see no way in which this problem is not down-boil-able to brute-forcing the password the user puts in.
Actually Apple admits this much! They can build a work around! What stops the three-letter-agency from building it?
There must at some point be a complex user entered passphrase if you want to be safe. This can be a fingerprint of course but there is always the 4 letter password/passphrase that is the weak point.
I could be completely wrong, so far I'm not convinced I am.
i believe that 'Hardware Key' refers to the UID. if this is the case, then learning the UID + user passcode gives you the root keys, from which you can decrypt the remaining key hierarchy until you reach the decrypted files.
re: brute forcing. they are AES 256 bit keys. good luck.
How about disabling the wipe signal by cutting some interconnects in the silicon?
Is the function by which the AES keys are calculated from the password really completely inaccessible?
The secure enclave must still give it's UID under some circumstances? This still does not appear to be immune to hardware hacking.
Moreover, this UID can also be brute forced imo, when the memory chip and secure enclave are physically separated. Whatever is needed to de-encrypt the data must be brute force-able, especially when the memory is separated from the wipe-all-data initiator which does not seem to be impossible if you know the chip design well enough?