On 2017-07-16 11:59, Richard Henderson wrote: > On 07/16/2017 11:43 AM, Aurelien Jarno wrote: > > Indeed, if the same atomic code is used often it might be better to have > > it cached. That said it's only true for TB that are recognized, as IIRC > > TB with the exclusive lock are not cached. > > At the moment they are not. > > But in Emilio's multi-threaded tcg v2 patch set, posted just today, we > change that by adding a CF_PARALLEL flag to TB->cflags, and incorporating > that into the tb hash.
Thanks for the hint, I'll have a look. Unfortunately I have one week backlog on the mailing list... -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurel...@aurel32.net http://www.aurel32.net
