For crypto unit tests, it is generally better to use a pseudo-random RNG, with a known (but not constant) seed, so that potential failures can be diagnosed more easily. E.g., the seed could be logged in the test output.

Yes, we can create a host test specific lib if no better ways.

 

Regards,

Yi

 

From: Yao, Jiewen <jiewen.yao@intel.com>
Sent: Friday, June 14, 2024 3:20 PM
To: Li, Yi1 <yi1.li@intel.com>; devel@edk2.groups.io
Cc: Hou, Wenxing <wenxing.hou@intel.com>; Kinney, Michael D <michael.d.kinney@intel.com>; Pedro Falcato <pedro.falcato@gmail.com>; Ard Biesheuvel <ardb+tianocore@kernel.org>
Subject: RE: CryptoPkg host test broken due to smoketest for RDRAND

 

Can we use a host test specific RngLib?

 

 

 

From: Li, Yi1 <yi1.li@intel.com>
Sent: Friday, June 14, 2024 3:08 PM
To: devel@edk2.groups.io
Cc: Hou, Wenxing <wenxing.hou@intel.com>; Yao, Jiewen <jiewen.yao@intel.com>; Kinney, Michael D <michael.d.kinney@intel.com>; Pedro Falcato <pedro.falcato@gmail.com>; Ard Biesheuvel <ardb+tianocore@kernel.org>
Subject: CryptoPkg host test broken due to smoketest for RDRAND

 

All crypto host tests which consumed randlib broken due to:

https://github.com/tianocore/edk2/pull/5714

Not sure why this issue not reported  by CI when merge this PR.

 

The reason is that the ```BaseRngLibConstructor``` of rnglib is not called in host test, so ```mRdRandSupported``` is not enabled.

Then the Crypto API calls ```GetRandomNumber*``` will fail.

GetRandomNumber64 (

  OUT     UINT64  *Rand

  )

{

  ……

  if (!ArchIsRngSupported ()) {

    return FALSE;

  }

 

Is there a way to let unit test host to call the constructors correctly?

 

Regards,

Yi