Ah. I misread that schedstate_percpu uses percpu(9)'s fast path, which doesn't exist...
Anyway if it's assumed that cpu is not attached at run-time, assigning struct cpu_data::void *cpu_dtraceinfo at module attachment would be just fine. On Thu, Mar 6, 2014 at 2:08 PM, Matt Thomas <[email protected]> wrote: > > On Mar 5, 2014, at 9:07 PM, Masao Uebayashi <[email protected]> wrote: > >> percpu(9) might help? > > Not so much. It's better to just extend cpu_data if it's a common > component. Much less overhead.
