[$] Two new block I/O schedulers for 4.12

The multiqueue block layer subsystem,introduced in 2013, was a necessary step for the kernel to scale to the fasteststorage devices on large systems. The implementation in current kernels isincomplete, though, in that it lacks an I/O scheduler designed to work withmultiqueue devices. That gap iscurrently set to be closed in the 4.12 development cycle when the kernelwill probably get not just one, but two new multiqueue I/O schedulers.