- separate submap for zio buffers
- different allocation strategies (best fit, first fit, next fit)
- per-cpu caching of vm allocations (similarly to uma)
- 2^N freelists similar to our phys pages allocator
AndriyGapon/AvgRandomNotes (last edited 2016-07-21T11:05:48+0000 by KubilayKocak)