Add #[inline] to shift and wrapping arithmetic ops

[cong-or]

Went through the shift and wrapping arithmetic paths and added #[inline] where it was missing — mostly thin wrappers and small loop bodies in Uint, Int, and Limb.

Stuck with #[inline] over #[inline(always)] so LLVM still gets the final say.

Ran benchmarks before/after on default codegen-units:

Benchmark	Change
shl, U2048	-5.5%
shl_vartime large, U2048	-12.7%
shr, U2048	-10.9%
shr_vartime large, U2048	-12.3%
shr_vartime_wide, U2048	-11.4%
wrapping_square, U256	-18%
widening_mul, I2048	-18%

Still working through the rest of the ops flagged in #981, will follow up separately.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.99%. Comparing base (9ba2ef6) to head (984209d).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1229   +/-   ##
=======================================
  Coverage   87.99%   87.99%           
=======================================
  Files         184      184           
  Lines       21244    21244           
=======================================
  Hits        18693    18693           
  Misses       2551     2551           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.