introduce a fast-path for the hash disk job

[arvidn]

This patch reads entire pieces at a time when hashing them, unless we already have some mixed in blocks already in the cache, in which case we use the existing (slow-path) that reads one block at a time

[codecov-io]

Codecov Report

Merging #3112 into RC_1_1 will decrease coverage by <.01%.
The diff coverage is 74.72%.

@@            Coverage Diff             @@
##           RC_1_1    #3112      +/-   ##
==========================================
- Coverage   66.59%   66.59%   -0.01%     
==========================================
  Files         376      376              
  Lines       61234    61266      +32     
  Branches     9749     9754       +5     
==========================================
+ Hits        40777    40798      +21     
+ Misses      16015    16007       -8     
- Partials     4442     4461      +19

Impacted Files	Coverage Δ
src/disk_io_thread.cpp	65.86% <74.72%> (+0.05%)	⬆️
include/libtorrent/peer_connection.hpp	76.99% <0%> (-2.66%)	⬇️
src/close_reason.cpp	18.88% <0%> (-2.23%)	⬇️
src/utp_stream.cpp	74.09% <0%> (-0.8%)	⬇️
src/choker.cpp	33.33% <0%> (-0.7%)	⬇️
src/kademlia/routing_table.cpp	65.63% <0%> (-0.67%)	⬇️
src/escape_string.cpp	78.81% <0%> (-0.5%)	⬇️
src/torrent.cpp	53.6% <0%> (-0.37%)	⬇️
src/ut_pex.cpp	59.31% <0%> (-0.35%)	⬇️
src/piece_picker.cpp	88.79% <0%> (-0.07%)	⬇️
... and 11 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4ae262d...7af1178. Read the comment docs.