On Tue 21-04-20 14:47:13, Vlastimil Babka wrote: [...] > In that case the generic implementation should be just reverted to be simple and > not assume any (micro)architectural details. If any architecture wants an > optimized version they could add it opt-in, and justifify it by using real > workloads, not microbenchmarks. Exactly! And I have requested that already with no response. So can we simply go and make the current implementation x86 specific and reintroduce the previous one as a default? -- Michal Hocko SUSE Labs
