Refactor local.set register preservation

[Robbepop]

Closes #766

This fixes a bug in preservation of local.set for local variables that have been pushed multiple times onto the stack upon preservation and implements recycling of preservation slots to reduce register pressure.

[paritytech-cicd-pr]

BENCHMARKS

	NATIVE			WASMTIME
BENCHMARK	MASTER	PR	DIFF	MASTER	PR	DIFF	WASMTIME OVERHEAD
execute/ bare_call_0	1.66ms	1.55ms	🟢 -6.45%	1.13ms	1.08ms	🟢 -4.58%	🟢 -31%
execute/ bare_call_0/typed	1.17ms	1.18ms	⚪ 1.03%	788.55µs	783.17µs	⚪ -1.00%	🟢 -34%
execute/ bare_call_1	1.74ms	1.61ms	🟢 -7.39%	1.30ms	1.26ms	🔴 -3.23%	🟢 -22%
execute/ bare_call_16	2.59ms	2.53ms	🟢 -2.03%	3.44ms	3.39ms	🟢 -1.53%	🟢 34%
execute/ bare_call_16/typed	1.58ms	1.56ms	🟢 -1.57%	1.90ms	1.82ms	🟢 -4.71%	🟢 17%
execute/ bare_call_1/typed	1.32ms	1.27ms	🟢 -3.40%	1.04ms	1.14ms	🔴 8.94%	🟢 -10%
execute/ bare_call_4	1.78ms	1.78ms	🟢 -0.02%	1.72ms	1.63ms	🟢 -5.26%	🟢 -8%
execute/ bare_call_4/typed	1.32ms	1.22ms	🟢 -7.43%	1.13ms	1.06ms	🟢 -6.41%	🟢 -14%
execute/ br_table	1.40ms	1.38ms	⚪ -1.72%	1.19ms	1.19ms	⚪ 0.13%	🟢 -14%
execute/ count_until	575.38µs	620.58µs	🔴 7.89%	1.29ms	1.30ms	⚪ 0.48%	🔴 110%
execute/ factorial_iterative	319.88µs	321.56µs	⚪ 0.47%	576.40µs	575.42µs	⚪ -0.06%	🟡 79%
execute/ factorial_recursive	495.43µs	493.96µs	⚪ -0.25%	759.04µs	753.17µs	⚪ -0.84%	🟡 52%
execute/ fibonacci_iter	1.66ms	1.40ms	🟢 -15.98%	2.78ms	2.79ms	⚪ 0.28%	🟡 99%
execute/ fibonacci_rec	3.97ms	3.97ms	⚪ -0.09%	6.68ms	6.68ms	⚪ 0.05%	🟡 68%
execute/ fibonacci_tail	863.98µs	857.10µs	⚪ -0.79%	1.64ms	1.64ms	⚪ 0.19%	🟡 91%
execute/ global_bump	730.53µs	747.67µs	🔴 2.48%	1.66ms	1.69ms	⚪ 1.18%	🔴 126%
execute/ global_const	681.35µs	659.01µs	🟢 -3.38%	1.34ms	1.35ms	⚪ 0.17%	🔴 104%
execute/ host_calls	37.17µs	37.13µs	⚪ 0.20%	39.08µs	38.43µs	🟢 -1.70%	🟢 3%
execute/ memory_fill	1.19ms	1.15ms	🟢 -3.37%	2.45ms	2.45ms	⚪ -0.02%	🔴 113%
execute/ memory_sum	1.13ms	1.14ms	⚪ 0.08%	2.40ms	2.41ms	⚪ 0.31%	🔴 112%
execute/ memory_vec_add	2.35ms	2.35ms	⚪ -0.06%	5.29ms	5.28ms	⚪ 1.23%	🔴 124%
execute/ recursive_is_even	667.75µs	664.61µs	⚪ -0.48%	1.12ms	1.12ms	⚪ 0.35%	🟡 69%
execute/ recursive_ok	93.52µs	95.13µs	🔴 1.64%	169.79µs	169.59µs	⚪ -0.20%	🟡 78%
execute/ recursive_scan	129.88µs	129.94µs	⚪ 0.08%	227.89µs	227.76µs	⚪ -0.11%	🟡 75%
execute/ recursive_trap	8.70µs	8.86µs	🔴 1.82%	17.14µs	17.06µs	⚪ -0.48%	🟡 93%
execute/ regex_redux	455.96µs	461.26µs	🔴 1.29%	890.35µs	895.53µs	⚪ 0.60%	🟡 94%
execute/ rev_complement	425.14µs	422.53µs	⚪ -0.75%	863.55µs	869.91µs	⚪ 0.58%	🔴 106%
execute/ tiny_keccak	320.34µs	333.02µs	🔴 3.59%	672.81µs	671.23µs	⚪ -0.14%	🔴 102%
execute/ trunc_f2i	727.79µs	738.30µs	🔴 1.60%	1.70ms	1.71ms	⚪ 0.18%	🔴 131%
instantiate/ wasm_kernel	55.16µs	54.85µs	⚪ 0.20%	57.96µs	55.71µs	🟢 -4.47%	🟢 2%
translate/ erc1155	209.56µs	208.27µs	⚪ -0.80%	356.46µs	359.20µs	⚪ 0.60%	🟡 72%
translate/ erc20	104.09µs	104.38µs	⚪ 0.26%	173.21µs	174.06µs	⚪ 0.54%	🟡 67%
translate/ erc721	146.65µs	146.38µs	⚪ -0.20%	250.15µs	252.41µs	⚪ 0.92%	🟡 72%
translate/ spidermonkey	64.67ms	64.41ms	⚪ -0.53%	0.00ns	0.00ns	⚪ 0.73%	🟢 -100%
translate/ wasm_kernel	4.32ms	4.25ms	🟢 -1.60%	6.50ms	6.51ms	⚪ -0.08%	🟡 53%

Link to pipeline