fix crypto test failures #5775
fix crypto test failures #5775
Conversation
kraxel
commented
Jun 14, 2024
kraxel
commented
- CryptoPkg/Test: call ProcessLibraryConstructorList
- MdePkg/X86UnitTestHost: set rdrand cpuid bit
Needed to properly initialize BaseRngLib. Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Set the rdrand feature bit when faking cpuid for host test cases. Needed to make the CryptoPkg test cases work. Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
kraxel
force-pushed
the
devel/rdrand-crypto-test
branch
from
330f8d9
to
2979709
Compare
| RetEcx = 0; | ||
| switch (Index) { | ||
| case 1: | ||
| RetEcx |= BIT30; /* RdRand */ |
There was a problem hiding this comment.
I don't think it is right fix to hardcode the value.
I recommend calling CPUID to get real value.
There was a problem hiding this comment.
Before merging #5714 the situation was exactly the same, the host test simply assumed rdrand support is there without actually checking. So, yes, it's not ideal, but it also is not worse than it used to be.
Note that proper rdrand checking needs more than just calling CPUID to figure the real value, the test cases also need an update to either skip tests or fallback to another RNG source in case rdrand is not there.
Given that rdrand is roughly a one decade old (2012+ for intel, 2015+ for amd) and using rdrand without checking apparently was not much of an problem before I'm not convinced this would be worth the effort.
