Messages in this thread | | | Date | Wed, 10 Sep 2008 16:33:16 -0700 | From | Mike Travis <> | Subject | Re: [RFC] CPUMASK: proposal for replacing cpumask_t |
| |
Andi Kleen wrote: >> Here's an initial proposal for abstracting cpumask_t to be either > > At least for some cases I don't think you'll get around defining > a "nearby subset of CPUs that can be handled together" type. Handling 1K > objects all the time in one piece is simply not a good idea. > > -Andi
Every time I stop to think about this, the problems with the cpu operators come to mind. Should there be a separate set? Or simply conversion functions to/from a "cpumask_subset" type?
Thanks, Mike
| |