Optimize histogram generation + fix masked histograms for 2-band images + add test coverage

[akx]

Follows up on #9050
Follows up on #9054
Follows up on #9752 (comment)

This applies similar hoist/restrict optimizations as the earlier optimization PRs.
While doing that, I noticed the masked/non-masked loops were curiously different, and it looks like #9054 was never applied to the masked case.

I also noticed the masked path for histograms didn't have actual test coverage (they were incidentally covered by e.g. autocontrast), so now they do.

[codspeed]

Merging this PR will improve performance by 66.4%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 2 improved benchmarks
✅ 607 untouched benchmarks
⏩ 338 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_histogram_masked[1237x811-L]	13.3 ms	7.1 ms	+88.85%
⚡	test_histogram_masked[1237x811-LA]	21.4 ms	14.6 ms	+46.62%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing akx:histo-opt (349c6ca) with main (26ddd05)}

Footnotes

1. 338 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩