Lame. Both protection mechanisms have more effective and general alternatives.
This tweet is unavailable
2
1
7
For anti-ROP: Why does this not take an immediate operand to indicate type signature etc.? Like @grsecurity’s RAP but presumably faster.
5
2
8
It could also have the advantage of being compatible with execute-only code, since it wouldn't need to read the code.
1
3
RAP's C++ virtual function checking must be more complicated than one ID check so CET would still be weaker with that.
1
it's exactly the same as normal fptr checks just at a different offset for the ancestor method hash.

Jun 14, 2016 · 8:25 PM UTC