io_uring: wire up bio allocation cache

io_uring: wire up bio allocation cache

Initialize a bio allocation cache, and mark it as being used for
IOPOLL. We could use it for non-polled IO as well, but it'd need some
locking and probably would negate much of the win in that case.

We start with IOPOLL, as completions are locked by the ctx lock anyway.
So no further locking is needed there.

This brings an IOPOLL gen2 Optane QD=128 workload from ~3.0M IOPS to
~3.25M IOPS.

Signed-off-by: Jens Axboe <axboe@kernel.dk>