shred: add sector-phase patterns and Pattern::from_code

[sankalpsthakur]

Prep for #11611. Pattern model + sector-phase write only.

• Pattern { bytes, flip_sector } and from_code (bit 0x1000)
• Write path XOR first byte of each 512B sector with 0x80
• Pass scheduler unchanged (legacy path)

Follow-ups: #13699 genpattern, #13700 residual-entropy (Fixes #11611).

Test: cargo test -p uu_shred --lib and cargo test --test tests test_shred

AI-assisted; I reviewed the diff.

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/cut/bounded-memory (passes in this run but fails in the 'main' branch)

[codspeed-hq]

Merging this PR will degrade performance by 0.38%

⚡ 1 improved benchmark
❌ 1 regressed benchmark
✅ 343 untouched benchmarks
⏩ 46 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	du_deep_tree[(100, 3)]	2.6 ms	2.7 ms	-3.68%
⚡	Simulation	cksum_crc32b	40.9 ms	39.7 ms	+3.03%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing sankalpsthakur:fix/11611-shred-distribution (7c9b66b) with main (0c8a3c7)}

Footnotes

1. 46 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. ↩

[sylvestre]

This is a way too big patch. Please ask your llm to make it shorter

[sankalpsthakur]

Slimmed the patch for review: same fix (Gutmann pass groups + sector-phase patterns + residual-entropy scheduling), but dropped verbose comments/helpers and extra unit tests. Diff is now +258/−222 (was +477/−202).

[sylvestre]

Removing tests isn't the right approach

Splitting it into several PR yes

[sankalpsthakur]

Agreed - I will split instead of dropping tests. Plan: (1) Pattern model + sector-phase write path, (2) genpattern / pass-group scheduler, (3) residual-entropy --random-source + full distribution regressions. This PR is now (1) only. Stacked PRs next; only the last Fixes #11611.

[sankalpsthakur]

Split stack: this PR is (1) patterns only. Follow-ups: #13699 genpattern, #13700 residual-entropy + Fixes #11611.

[sylvestre]

The diff is still quite large...

[sankalpsthakur]

Compressed the Pattern table (const helpers) so this PR stays just the model + sector-phase write. Diff should be smaller now. Scheduler still deferred to #13699/#13700.