Implement -Z function-sections=yes|no

[nox]

This lets rustc users tweak whether all functions should be put in their own TEXT section, using whatever default value the target defines if the flag is missing.

I'm having fun experimenting with musl libc and trying to implement the start symbol in Rust, that means avoiding code that requires relocations, and AFAIK putting everything in its own section makes the toolchain generate GOTPCREL relocations for symbols that could use plain old PC-relative addressing (at least on x86_64) if they were all in the same section.

[rust-highfive]

r? @varkor

(rust_highfive has picked a reviewer for you, use r? to override)

[nox]

I implemented it as a flag because AFAIK that particular setting doesn't require libstd etc to be built the same thing, thus it is a bit unfortunate that it requires building a whole new target currently to change the value. Plus, it would make it easier to check on Windows if we can turn back function-sections on by default there if the bugs are gone.

[jyn514]

cc @bjorn3, https://rust-lang.zulipchat.com/#narrow/stream/247081-t-compiler.2Fperformance/topic/-ffunction-sections

[eddyb]

r? @bjorn3 cc @nagisa

[jyn514]

@bors delegate=bjorn3

[bors]

✌️ @bjorn3 can now approve this pull request

[nagisa]

.text is not exactly the most portable name (e.g. machO uses __text AFAIR). I suggest dropping the . or perhaps the whole .text.

[nox]

That's a good point, I removed it.

[nagisa]

@bors r=nagisa,bjorn3

[bors]

📌 Commit 0569422 has been approved by nagisa,bjorn3

[nox]

Any chance I could get a p=1 here? I can't wait to play around with that and confirm my understanding of that stuff, but as I'm switching jobs and computers and whatnot all I'm doing is building stuff for Linux on my macOS computer and I don't understand how to do that with a custom rustc. 😅

[nagisa]

Its the next PR in the bors queue.

[bors]

⌛ Testing commit 0569422 with merge 3dddf6a...

[bors]

☀️ Test successful - checks-actions
Approved by: nagisa,bjorn3
Pushing 3dddf6a to master...