tree 81326943c6ce063d19c839bde24794a744219fb3
parent 53844eb8bb2a6935a77746772657fea96fa1fc9f
author Rudolf Polzer <rpolzer@google.com> 1763733176 -0800
committer Boringssl LUCI CQ <boringssl-scoped@luci-project-accounts.iam.gserviceaccount.com> 1765798501 -0800

EVP_AEAD: implement sealv/openv for ChaCha20-Poly1305.

Note that for now, the asm optimizations are only available when using
up to 2 plaintext segments or up to 1 ciphertext segment.

Performance note: this seems to cost 14 to 40 CPU cycles per call with
LTO, and 45 to 51 CPU cycles without; no measurable per-byte overhead
though.

Bug: 383343306
Change-Id: I8ebbba3280ac51c200e463b6f71e6e3ed028d450
Reviewed-on: https://boringssl-review.googlesource.com/c/boringssl/+/84507
Reviewed-by: David Benjamin <davidben@google.com>
Commit-Queue: Rudolf Polzer <rpolzer@google.com>
Auto-Submit: Rudolf Polzer <rpolzer@google.com>