Improve performance for extremely small I/O workloads

Hermes performs well at 4KB, but not so well at 300 bytes -- which is the case for certain workload types we are encountering.

1. Make a lock-free hashmap supporting both inserts and deletes
2. Improve shared memory allocator performance by leveraging lock-free queues for a small cache
3. Allow clients to read data directly from the SHM hashmap

Start by mapping Hermes’s shared-memory allocator, SHM hashmap, and I/O paths before evaluating the three proposed performance changes. Use the 300-byte workload as the baseline; the work is done when inserts and deletes are lock-free, small allocations use the proposed cache, and clients can read directly from the SHM hashmap with improved performance.

Written by the indexing model from the issue text.