Stateful sigs XMSS updates

[ashman-p]

Updated XMSS code to use callbacks for "secure store/update" of secret key after each signature generation.
Reflect these changes in the test app.
Some XMSS tests takes too long and hit the time-out limit.

• Does this PR change the input/output behaviour of a cryptographic algorithm (i.e., does it change known answer test values)? (If so, a version bump will be required from x.y.z to x.(y+1).0.)
• Does this PR change the list of algorithms available -- either adding, removing, or renaming? Does this PR otherwise change an API? (If so, PRs in fully supported downstream projects dependent on these, i.e., oqs-provider and OQS-OpenSSH will also need to be ready for review and merge by the time this is merged.)

[ducnguyen-sb]

thanks @ashman-p , I will integrate the changes into my XMSS code.

[ashman-p]

@ducnguyen-sb ... These are the testcase that I see failures with.
The tests essentially does the following operations...

1.       Key-pair generation
   

2.       Signature generation 
   

3.       Signature verification
   

ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHA2_20/4_256
ERROR: lock thread test Sig Verify 1 failed. :XMSSMT-SHA2_20/4_256
ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHA2_40/4_256
ERROR: lock thread test Sig Verify 1 failed. :XMSSMT-SHA2_40/4_256
ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHA2_40/8_256
ERROR: lock thread test Sig Verify 1 failed. :XMSSMT-SHA2_40/8_256
ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHA2_60/6_256
ERROR: lock thread test Sig Verify 1 failed. :XMSSMT-SHA2_60/6_256
ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHAKE_20/2_256
ERROR: lock thread test Sig Verify 1 failed. :XMSSMT-SHAKE_20/2_256
ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHAKE_20/4_256
ERROR: lock thread test Sig Verify 1 failed. :XMSSMT-SHAKE_20/4_256
ERROR: OQS_SIG_STFL_verify failed. :XMSSMT-SHAKE_40/8_256
ERROR: OQS_SIG_STFL_sign failed. :XMSSMT-SHAKE_60/12_256
ERROR: lock thread test OQS_SIG_STFL_sign failed. :XMSSMT-SHAKE_60/12_256

zsh: segmentation fault tests/test_sig_stfl XMSSMT-SHAKE_40/8_256

The other tests, other than the slow keypair generation, passes.

[baentsch]

@ashman-p @ducnguyen-sb Pardon the question from the "side lines": Why are you merging code that CI already shows as failing? The error list written up above is exactly visible as such by clicking at any of the failing pre-merge CI runs, e.g., this. Or does this statement

Some XMSS tests takes too long and hit the time-out limit.

make you consider CI results as not relevant/trustworthy for this code? Allow me to posit the goal that any code destined for "main" has "full feature coverage" CI that not only passes but does that also within the 1h. Extending this period makes contributions (of anything) "unpleasant" (and is bad to the environment by way of too many CPU cycles/power/CO2 consumed).

@ashman-p You once asked how to separate between "productive" and "experimental"/"research" code: I think this is one such area: Stuff that cannot be fully tested in a standard CI environment should stay in a separate "experimental" branch. IMO we could also further relax other CI tasks in such new branch (like style or memory checks). Maybe something to discuss in an upcoming OQS meeting.

[ducnguyen-sb]

Hi @ashman-p, I've fixed all problems mentioned in @ashman-p post.
#1588

@baentsch Sorry for the inconvenience, in a meeting, we realize we are work on the same code, thus, it's easier for him to transfer the work with XMSS codebase to me and I will update the state-sigs from current broken status to working status.