Fuse i32.{and,or, xor} + [i32.eqz] + br_if Wasm instructions

[Robbepop]

Closes #791.

Added a new fuse benchmark that shows the performance improvements gained in this PR. The improvements in this artificial .wat file are 25% compared to the current master branch which already implements branch+cmp fusion. On master each iteration executes 5 wasmi instructions whereas this PR gets it down to just 4 explaining the 25% performance improvement. Generally the gains are low for non-artificial tests such as tiny_keccak or coremark.wasm because they do not profit a lot from this op-code fusion in their hot loops it seems.

TODO

• Tests for i32.{and,or,xor}[_imm16] + i32.eqz fusion.
• Tests for i32.{and,or,xor}[_imm16] + br_if fusion.
• Tests for i32.{and,or,xor}[_imm16] + i32.eqz + br_if fusion.

[codecov-commenter]

Codecov Report

Attention: 171 lines in your changes are missing coverage. Please review.

Comparison is base (ea3a29c) 81.29% compared to head (47a8421) 81.07%.

Files	Patch %	Lines
crates/wasmi/src/engine/regmach/translator/mod.rs	43.03%	90 Missing ⚠️
crates/wasmi/src/engine/regmach/executor/instrs.rs	0.00%	24 Missing ⚠️
...mi/src/engine/regmach/translator/visit_register.rs	0.00%	21 Missing ⚠️
...smi/src/engine/regmach/translator/instr_encoder.rs	55.00%	18 Missing ⚠️
.../wasmi/src/engine/regmach/translator/result_mut.rs	5.55%	17 Missing ⚠️
...rates/wasmi/src/engine/regmach/translator/visit.rs	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #796      +/-   ##
==========================================
- Coverage   81.29%   81.07%   -0.23%     
==========================================
  Files         273      273              
  Lines       25085    25204     +119     
==========================================
+ Hits        20394    20435      +41     
- Misses       4691     4769      +78     

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

[paritytech-cicd-pr]

BENCHMARKS

	NATIVE			WASMTIME
BENCHMARK	MASTER	PR	DIFF	MASTER	PR	DIFF	WASMTIME OVERHEAD
execute/ bare_call_0	1.55ms	1.58ms	🔴 2.35%	1.11ms	1.10ms	🟢 -1.42%	🟢 -31%
execute/ bare_call_0/typed	1.25ms	1.17ms	🟢 -6.61%	801.13µs	802.82µs	⚪ 0.21%	🟢 -31%
execute/ bare_call_1	1.60ms	1.64ms	🔴 2.72%	1.26ms	1.32ms	🔴 4.78%	🟢 -19%
execute/ bare_call_16	2.64ms	2.55ms	🔴 -2.69%	3.53ms	3.50ms	🔴 -0.61%	🟢 37%
execute/ bare_call_16/typed	1.56ms	1.68ms	🔴 8.25%	1.94ms	1.98ms	🔴 2.48%	🟢 18%
execute/ bare_call_1/typed	1.29ms	1.41ms	🔴 8.90%	981.24µs	983.40µs	⚪ 0.34%	🟢 -30%
execute/ bare_call_4	1.77ms	1.85ms	🔴 4.10%	1.67ms	1.64ms	🟢 -1.58%	🟢 -11%
execute/ bare_call_4/typed	1.23ms	1.25ms	🔴 2.35%	1.08ms	1.08ms	⚪ -0.49%	🟢 -14%
execute/ br_table	1.64ms	1.35ms	🟢 -17.18%	1.17ms	1.18ms	⚪ 0.84%	🟢 -13%
execute/ count_until	546.24µs	545.95µs	⚪ -0.04%	1.69ms	1.70ms	⚪ 0.12%	🔴 211%
execute/ factorial_iterative	321.57µs	322.19µs	⚪ 0.15%	806.55µs	793.65µs	🟢 -1.47%	🔴 146%
execute/ factorial_recursive	497.78µs	490.71µs	🟢 -1.41%	968.06µs	974.06µs	⚪ 0.57%	🟡 98%
execute/ fibonacci_iter	1.40ms	1.40ms	⚪ 0.33%	3.73ms	3.81ms	🔴 1.84%	🔴 172%
execute/ fibonacci_rec	3.99ms	3.97ms	⚪ -0.31%	8.59ms	8.59ms	⚪ 0.03%	🔴 116%
execute/ fibonacci_tail	857.41µs	855.59µs	⚪ -0.21%	2.19ms	2.19ms	⚪ 0.11%	🔴 156%
execute/ global_bump	747.14µs	740.43µs	⚪ -0.92%	2.19ms	2.74ms	🔴 25.31%	🔴 270%
execute/ global_const	659.38µs	660.60µs	⚪ 0.20%	2.45ms	2.44ms	⚪ -0.43%	🔴 269%
execute/ host_calls	36.14µs	37.71µs	⚪ 1.12%	39.84µs	39.70µs	⚪ -0.29%	🟢 5%
execute/ memory_fill	1.15ms	1.22ms	🔴 5.64%	3.32ms	3.33ms	⚪ 0.30%	🔴 174%
execute/ memory_sum	1.19ms	1.18ms	⚪ -0.67%	3.28ms	3.30ms	⚪ 0.45%	🔴 179%
execute/ memory_vec_add	2.35ms	2.34ms	⚪ -0.08%	7.41ms	7.40ms	⚪ -0.10%	🔴 216%
execute/ recursive_is_even	664.02µs	661.74µs	⚪ -0.36%	1.47ms	1.45ms	🟢 -1.25%	🔴 119%
execute/ recursive_ok	94.41µs	94.49µs	⚪ 0.15%	200.02µs	201.34µs	⚪ 0.70%	🔴 113%
execute/ recursive_scan	129.20µs	129.56µs	⚪ 0.27%	284.81µs	285.81µs	⚪ 0.37%	🔴 121%
execute/ recursive_trap	8.81µs	8.85µs	⚪ 0.44%	20.85µs	20.84µs	⚪ -0.06%	🔴 136%
execute/ regex_redux	459.00µs	460.97µs	⚪ 0.45%	1.23ms	1.24ms	⚪ 0.67%	🔴 169%
execute/ rev_complement	425.75µs	434.94µs	🔴 2.19%	1.14ms	1.15ms	⚪ 0.53%	🔴 164%
execute/ tiny_keccak	320.98µs	324.18µs	⚪ -0.25%	1.10ms	1.09ms	⚪ -0.83%	🔴 238%
execute/ trunc_f2i	737.41µs	733.28µs	⚪ -0.54%	1.71ms	1.71ms	⚪ -0.01%	🔴 134%
instantiate/ wasm_kernel	53.13µs	56.85µs	🔴 5.67%	59.86µs	56.55µs	🟢 -5.93%	🟢 -1%
translate/ erc1155	208.80µs	205.82µs	🟢 -1.42%	375.50µs	360.05µs	🟢 -4.31%	🟡 75%
translate/ erc20	103.18µs	103.57µs	⚪ 0.45%	181.46µs	174.37µs	🟢 -3.85%	🟡 68%
translate/ erc721	146.60µs	145.35µs	⚪ -0.88%	262.24µs	251.88µs	🟢 -4.05%	🟡 73%
translate/ spidermonkey	64.27ms	64.25ms	⚪ -0.20%	0.00ns	0.00ns	🟢 -3.63%	🟢 -100%
translate/ wasm_kernel	4.26ms	4.23ms	⚪ -0.46%	6.85ms	6.60ms	🟢 -3.63%	🟡 56%

Link to pipeline