tree e32bea6154a2eb235591e56aeb2b4cd3a50c1c40
parent 2105f289ad64d1d8b8599555677dc5f9c429c568
author Nathaniel Wesley Filardo <wes.filardo@scisemi.com> 1734972787 +0000
committer Nathaniel Wesley Filardo <VP331RHQ115POU58JFRLKB7OPA0L18E3@cmx.ietfng.org> 1736787813 -0500

Hardware revoker: correct wait for indicated epoch

The commentary says that we wait for the indicated (even) epoch to have
finished, but in practice we were merely waiting for the next one to
begin.  Because has_revocation_finished_for_epoch() had the intended
semantics and is the actual function used to gate release from
quarantine, wait_for_completion()'s incorrect waiting was merely
ineffecient, making malloc_internal spin a bit in the "Quarantine has
enough memory to satisfy allocation" case.