Timing: add generated code jit timing feature

[JakeHillion]

Summary

Adds a timing call at the beginning and end of the generated code functions and store the time taken in nanoseconds in the data segment. Currently this is output at LOG(INFO) level. IMO this belongs in RocksDB and oid_out.json, but it requires a bit of a format change so I'll leave it until it's necessary.

This required adding another uintptr_t typed element to the data segment prefix, because adding it at the end of the data segment means keeping overhead of setting lengths and such after the call. Probe effect should be minimal as is, but as it's feature gated and off by default I haven't bothered measuring it. I added some names to pieces in these entry functions while I was in there, and also cleaned up a type in the typed data segment entry.

Test plan

• CI
• []: make test: 100% tests passed, 0 tests failed out of 578
• ["-fjit-timing"]: make test: 100% tests passed, 0 tests failed out of 578
• ["-ftype-graph"]: make test: 95% tests passed, 31 tests failed out of 578
• ["-ftype-graph", "-fjit-timing"]: make test: 95% tests passed, 31 tests failed out of 578
• ["-ftyped-data-segment"]: make test: 87% tests passed, 77 tests failed out of 578
• ["-ftyped-data-segment", "-fjit-timing"]: make test: 87% tests passed, 77 tests failed out of 578
  (that was hard work)

[ajor]

cppreference doesn't have a high opinion of high_resolution_clock:

The high_resolution_clock is not implemented consistently across different standard library implementations, and its use should be avoided.
...
Generally one should just use std::chrono::steady_clock or std::chrono::system_clock directly instead of std::chrono::high_resolution_clock: use steady_clock for duration measurements, and system_clock for wall-clock time.

[ttreyer]

It looks mostly good to me 🙂
Could you change the bool flag to the actual feature set, as discussed, please.
Also take a look at the data[3] idea.

[codecov-commenter]

Codecov Report

Merging #180 (47fa73c) into main (f130f3c) will decrease coverage by 0.12%.
The diff coverage is 36.36%.

❗ Current head 47fa73c differs from pull request most recent head 4e2d019. Consider uploading reports for the commit 4e2d019 to get more accurate results

@@            Coverage Diff             @@
##             main     #180      +/-   ##
==========================================
- Coverage   62.63%   62.51%   -0.12%     
==========================================
  Files          87       87              
  Lines        9266     9290      +24     
  Branches     1533     1542       +9     
==========================================
+ Hits         5804     5808       +4     
- Misses       2639     2652      +13     
- Partials      823      830       +7     

Impacted Files	Coverage Δ
oi/OIDebugger.h	45.28% <ø> (ø)
oi/FuncGen.cpp	59.85% <30.00%> (-5.19%)	⬇️
oi/OIDebugger.cpp	38.07% <33.33%> (-0.02%)	⬇️
oi/CodeGen.cpp	38.24% <40.00%> (-0.22%)	⬇️
oi/OICodeGen.cpp	67.02% <50.00%> (-0.10%)	⬇️
oi/TreeBuilder.cpp	76.28% <100.00%> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more