Have arrays' drop_glue just unsize and call the slice version

[scottmcm]

It's silly to emit two loops (because of the drop ladder -- just one in panic=abort) for every array length that's dropped when we can just polymorphize to the slice version.

Built atop #154327 to avoid conflicts later, so draft for now.

r? @WaffleLapkin

[rustbot]

WaffleLapkin is not on the review rotation at the moment.
They may take a while to respond.

[scottmcm]

View changes since the review

Compare nightly https://rust.godbolt.org/z/5Wv1q86ja with a loop in each of the three.

[WaffleLapkin]

💭 does it make sense to keep the array case separate for N <= 1 or something?

[scottmcm]

My understanding is that LLVM will very effectively inline those -- just like how we don't special case the array iterator for N ≤ 1 and still use the polymorphic version.

Does make me ponder if we should have a separate general "hey, there's no Drop so we can delegate to the one for the one field" kind of check, but I'd leave all these for a different PR if that's ok.

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[scottmcm]

@bors try @rust-timer queue

[scottmcm]

does it make sense to keep the array case separate for N <= 1 or something?

Oh, also for N == 0 it's detected as !needs_drop, so it gets a fully empty body.

[rust-bors]

☀️ Try build successful (CI)
Build commit: 69592c2 (69592c2a37914f5bee8cabac44cc5f28707bd578, parent: 4ddd4538a881317c622ed674b08300b8fc8dabdd)

[rust-timer]

Finished benchmarking commit (69592c2): comparison URL.

Overall result: ❌✅ regressions and improvements - please read:

Benchmarking means the PR may be perf-sensitive. It's automatically marked not fit for rolling up. Overriding is possible but disadvised: it risks changing compiler perf.

Next, please: If you can, justify the regressions found in this try perf run in writing along with @rustbot label: +perf-regression-triaged. If not, fix the regressions and do another perf run. Neutral or positive results will clear the label automatically.

@bors rollup=never
@rustbot label: -S-waiting-on-perf +perf-regression

Instruction count

Our most reliable metric. Used to determine the overall result above. However, even this metric can be noisy.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.1%	[0.1%, 0.1%]	2
Improvements ✅ (primary)	-0.4%	[-0.4%, -0.4%]	1
Improvements ✅ (secondary)	-0.2%	[-0.4%, -0.0%]	5
All ❌✅ (primary)	-0.4%	[-0.4%, -0.4%]	1

Max RSS (memory usage)

Results (primary -1.3%, secondary -0.6%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	2.0%	[2.0%, 2.0%]	1
Regressions ❌ (secondary)	1.5%	[0.5%, 2.4%]	6
Improvements ✅ (primary)	-2.9%	[-4.0%, -1.9%]	2
Improvements ✅ (secondary)	-2.2%	[-3.9%, -0.9%]	8
All ❌✅ (primary)	-1.3%	[-4.0%, 2.0%]	3

Cycles

Results (primary -2.3%, secondary 1.1%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	3.1%	[3.1%, 3.1%]	1
Regressions ❌ (secondary)	1.3%	[0.6%, 4.6%]	7
Improvements ✅ (primary)	-7.7%	[-7.7%, -7.7%]	1
Improvements ✅ (secondary)	-0.5%	[-0.5%, -0.5%]	1
All ❌✅ (primary)	-2.3%	[-7.7%, 3.1%]	2

Binary size

Results (primary 0.3%)

A less reliable metric. May be of interest, but not used to determine the overall result above.

	mean	range	count
Regressions ❌ (primary)	0.6%	[0.2%, 1.1%]	3
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-0.6%, -0.6%]	1
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.3%	[-0.6%, 1.1%]	4

Bootstrap: 499.861s -> 496.397s (-0.69%)
Artifact size: 394.98 MiB -> 397.13 MiB (0.54%)