rustc: Handle #[inline(always)] at -O0

[alexcrichton]

This commit updates the handling of #[inline(always)] functions at -O0 to
ensure that it's always inlined regardless of the number of codegen units used.

Closes #45201

[alexcrichton]

r? @michaelwoerister

[kennytm]

[01:05:17] ---- [codegen-units] codegen-units/partitioning/local-inlining-but-not-all.rs stdout ----
[01:05:17] 	
[01:05:17] The following items were assigned to wrong codegen units:
[01:05:17] 
[01:05:17] fn local_inlining_but_not_all::inline[0]::inlined_function[0]
[01:05:17]   expected: local_inlining_but_not_all-inline[External] 
[01:05:17]   actual:   local_inlining_but_not_all-user1[Internal] local_inlining_but_not_all-user2[Internal] 
[01:05:17] 
[01:05:17] thread '[codegen-units] codegen-units/partitioning/local-inlining-but-not-all.rs' panicked at 'explicit panic', /checkout/src/tools/compiletest/src/runtest.rs:1919:12
[01:05:17] note: Run with `RUST_BACKTRACE=1` for a backtrace.
[01:05:17] 
[01:05:17] 
[01:05:17] failures:
[01:05:17]     [codegen-units] codegen-units/partitioning/local-inlining-but-not-all.rs
[01:05:17] 
[01:05:17] test result: �[31mFAILED�(B�[m. 32 passed; 1 failed; 3 ignored; 0 measured; 0 filtered out

[alexcrichton]

ping r? @michaelwoerister

[michaelwoerister]

I somehow thought I had already approved this ...

@bors r+

[bors]

📌 Commit 5d415e8 has been approved by michaelwoerister

[bors]

⌛ Testing commit 5d415e8 with merge ba4e8d7...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: michaelwoerister
Pushing ba4e8d7 to master...