Adrian Stanciu

@paulidale This pull request includes a parallel implementation of the AES-CFB128 decryption. I would appreciate it if you could review the changes, the performance, and provide feedback. Improvements should be...

> Is there any chance you could get this peer reviewed by other intel folks? Thanks for your time! Yes, it will be reviewed internally too. I will mark it...

Pending internal review by Intel folks.

> Were block sizes of 32 considered? i.e using aesenc ymm ? In the current implementation the processing has 4 stages: - process sets of 16 blocks until fewer than...

> it looks like the last commits are not relevant to this pr? The `tooling` commit adds support for older assemblers not supporting AVX-512. The `32-bit` commit attempts to fix...

> Note also that I am not sure that the CI runs any of this code.. Which is why I started doing PR #27109 One way to find out is...

> Did you consider leaving the partial block support as c code?.. and just call the assembler to do full block operations. That would seem to be simpler for what...

> Looks like https://www.nasm.us is down currently.. Pending https://github.com/openssl/openssl/pull/27462

> These get auto squashed when the final merge happens. I have only force pushed to rebase over `master`. All the feedback was applied in new commits.

> Maybe a comment that links to this diagram might be helpful for someone trying to understand the decryption code.. Do you think this might be too wide to be...