Use protected visibility when LLD feature is enabled and enable it when building rustc

[davidlattimore]

rust-lang/compiler-team#782

I wasn't sure about having two commits in a PR, but I figured, at least initially it might make sense to discuss these commits together. Happy to squash, or move the second commit to a separate PR.

I contemplated trying to enable protected visibility for more cases when LLD will be used other than just -Zlinker-features=+lld, but that would be more a complex change that probably still wouldn't cover all cases when LLD is used, so went with the simplest option of just checking if the linker-feature is enabled.

r? lqd

[tgross35]

I wasn't sure about having two commits in a PR, but I figured, at least initially it might make sense to discuss these commits together. Happy to squash, or move the second commit to a separate PR.

Multiple atomic commits in a single PR are totally fine, that's easier to review and better for history anyway (the way you have this PR is ideal imo). We definitely don't want a bunch of back and forth WIP-style commits that make changes just to undo them or fix things up, but there is no 1PR=1commit rule like LLVM.

[nikic]

@bors try @rust-timer queue

[bors]

⌛ Trying commit 1e1bf2b with merge 6ad0952...

[bors]

☀️ Try build successful - checks-actions
Build commit: 6ad0952 (6ad095273a92c251fb0c1697fd745ab6caf21d2c)

[rust-timer]

Finished benchmarking commit (6ad0952): comparison URL.

Overall result: ✅ improvements - no action needed

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

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

Instruction count

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.6%	[-34.2%, -0.2%]	212
Improvements ✅ (secondary)	-4.6%	[-29.9%, -0.2%]	225
All ❌✅ (primary)	-1.6%	[-34.2%, -0.2%]	212

Max RSS (memory usage)

Results (primary -2.2%, secondary -3.2%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-2.2%	[-6.5%, -0.5%]	146
Improvements ✅ (secondary)	-3.2%	[-7.0%, -0.6%]	151
All ❌✅ (primary)	-2.2%	[-6.5%, -0.5%]	146

Cycles

Results (primary -3.3%, secondary -7.4%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-3.3%	[-24.2%, -0.6%]	53
Improvements ✅ (secondary)	-7.4%	[-22.6%, -1.9%]	94
All ❌✅ (primary)	-3.3%	[-24.2%, -0.6%]	53

Binary size

Results (primary -0.3%, secondary -0.7%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.3%	[-0.8%, -0.1%]	11
Improvements ✅ (secondary)	-0.7%	[-0.8%, -0.5%]	37
All ❌✅ (primary)	-0.3%	[-0.8%, -0.1%]	11

Bootstrap: 783.66s -> 778.727s (-0.63%)
Artifact size: 332.16 MiB -> 331.51 MiB (-0.20%)