On Fri, May 16, 2025 at 05:23:03PM +0800, Herbert Xu wrote: > On Thu, May 15, 2025 at 12:35:29PM -0700, Eric Biggers wrote: > > > > That seems backwards. The shash one should be the regular one and ahash should > > be special. > > That's how it was in v3 but because of the switch to ahash in > testmgr So don't do that. > > Still lacks any explanation for why this even matters. > > I've explained it many times before. The point is so that you > can fallback from async to sync at any point in time by exporting > the async hash state and importing it into the sync fallback that's > now allocated for every async ahash. So how come this hasn't been a problem until now? > > As usual, missing a base-commit. (Use the --base option to 'git format-patch') > > It's based on cryptodev. Which is a moving target. You still need to use base-commit. - Eric
