Add #[inline] to trivial AsRef/AsMut impls

[erikdesjardins]

These appeared uninlined in some perf runs, but they're trivial.

r? @ghost

[nikic]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[bors]

⌛ Trying commit 4194d75 with merge cf7838d4076c630a0ff2ad5a8acf6ee148956056...

[bors]

☀️ Try build successful - checks-actions
Build commit: cf7838d4076c630a0ff2ad5a8acf6ee148956056 (cf7838d4076c630a0ff2ad5a8acf6ee148956056)

[rust-timer]

Queued cf7838d4076c630a0ff2ad5a8acf6ee148956056 with parent d981633, future comparison URL.

[rust-timer]

Finished benchmarking commit (cf7838d4076c630a0ff2ad5a8acf6ee148956056): comparison url.

Summary: This benchmark run shows 13 relevant improvements 🎉 to instruction counts.

• Arithmetic mean of relevant improvements: -0.7%
• Largest improvement in instruction counts: -0.9% on incr-unchanged builds of externs debug

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

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 led to changes in compiler perf.

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

[Kobzol]

It's interesting that inline helps even on generic methods.

[erikdesjardins]

A small but consistent win.

r? rust-lang/compiler

[erikdesjardins]

It's interesting that inline helps even on generic methods.

I was wondering if someone would comment that :P

It is true that generic functions are codegen'd in downstream crates (without -Zshare-generics), which enables inlining.

But #[inline] also adds the inlinehint LLVM attribute, which increases the inlining cost threshold, making LLVM more willing to inline: https://github.com/llvm/llvm-project/blob/a524a12231efa6cae94843fabdf0ec4fdd42d65d/llvm/lib/Analysis/InlineCost.cpp#L1837-L1841

[michaelwoerister]

It is true that generic functions are codegen'd in downstream crates (without -Zshare-generics), which enables inlining.

That's right. One thing to keep in mind is that downstream crates are often compiled with multiple codegen units, so even though a generic function will have its own downstream instantiation, it might still not be available for inlining in all except one of the codegen units. On the other hand ThinLTO takes care of this pretty reliably for small functions, so I suspect the additional inlinehint to make the difference here.

Since this is in libcore: r? rust-lang/libs

[dtolnay]

Thanks!

[dtolnay]

@bors r+

[bors]

📌 Commit 4194d75 has been approved by dtolnay

[bors]

⌛ Testing commit 4194d75 with merge 90f79882992b98ddba920ab7da0175a63df55346...

[rust-log-analyzer]

A job failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

[bors]

💔 Test failed - checks-actions

[dtolnay]

x86_64-apple-2 failed with no indication why.

@bors retry

[bors]

⌛ Testing commit 4194d75 with merge 2947c05ec6c442c1790f86fe0aa04d3815362b9a...

[bors]

💔 Test failed - checks-actions

[rust-log-analyzer]

A job failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

[ehuss]

@bors retry

dist-x86_64-apple-alt was cancelled after running 1.5 hours.

[bors]

⌛ Testing commit 4194d75 with merge f2661cf...

[bors]

☀️ Test successful - checks-actions
Approved by: dtolnay
Pushing f2661cf to master...

[rust-timer]

Finished benchmarking commit (f2661cf): comparison url.

Summary: This benchmark run did not return any relevant results.

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

@rustbot label: -perf-regression