tree 3336507f2149255c139bff966a55011df533c241
parent a1403cba9d015eed381bd2eced84ad399665d245
author David Benjamin <davidben@google.com> 1693245711 -0400
committer Boringssl LUCI CQ <boringssl-scoped@luci-project-accounts.iam.gserviceaccount.com> 1693251240 +0000

Fix RBP-based unwind in fiat assembly

RBP pointed 8 bytes off of where it should be. I've left the RSP offsets
alone, though it does mean they're shifted by 8 from what they
previously were. Per Andres, the new version of CryptOpt will generate
an RBP-compatible prolog, but for now I've just fixed it up by hand.
(This part was already hand-written.)

Change-Id: I23720e76affff6fae46b8f85b0a509380ccc8bc0
Reviewed-on: https://boringssl-review.googlesource.com/c/boringssl/+/62805
Commit-Queue: David Benjamin <davidben@google.com>
Reviewed-by: Adam Langley <agl@google.com>
Auto-Submit: David Benjamin <davidben@google.com>