perf: use padded BitReader fast path in codestream sections

[hjanuschka]

This adds a padded BitReader constructor and uses it for codestream sections that already have complete buffered data. It keeps incremental/incomplete paths unchanged while avoiding slow refill handling on complete section reads.

[github-actions]

Benchmark @ a3453d2

MULTI-FILE BENCHMARK RESULTS (4 files)
  CPU architecture: x86_64
  WARNING: System appears noisy: high system load (2.03). Results may be unreliable.
Statistics:
  Confidence:               99.0%
  Max relative error:        3.0%

Comparing: e883140e (Base) vs 3593c52f (PR)

File	Base (MP/s)	PR (MP/s)	Δ%
bike.jxl	24.503	24.125	-1.54% ±2.6%
green_queen_modular_e3.jxl	7.926	7.756	-2.14% ±0.6%
green_queen_vardct_e3.jxl	23.945	23.884	-0.25% ±1.5%
sunset_logo.jxl	2.797	2.893	+3.43% ±0.8%