Share the naked asm impl between cg_ssa and cg_clif #134232
Conversation
|
cc @folkertdev |
|
r? @chenyukang rustbot has assigned @chenyukang. Use |
rustbot
added
S-waiting-on-review
This comment has been minimized.
This comment has been minimized.
bjorn3
force-pushed
the
naked_asm_improvements
branch
2 times, most recently
from
b8677a2 to
d6b3c06
Compare
This comment has been minimized.
This comment has been minimized.
bjorn3
force-pushed
the
naked_asm_improvements
branch
from
d6b3c06 to
0dd4292
Compare
|
Some changes occurred in compiler/rustc_codegen_gcc |
bjorn3
added
A-inline-assembly
| if let Visibility::Hidden = item_data.visibility { | ||
| writeln!(begin, ".hidden {asm_name}").unwrap(); | ||
| match item_data.visibility { | ||
| Visibility::Default => {} | ||
| Visibility::Protected => writeln!(begin, ".protected {asm_name}").unwrap(), | ||
| Visibility::Hidden => writeln!(begin, ".hidden {asm_name}").unwrap(), |
There was a problem hiding this comment.
I think we discussed documenting https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/mono/enum.Visibility.html a while ago, now might be a good time for that? (you had some ideas for that I believe)?
Also I don't think .hidden and .protected are ever hit by tests, is there any way we could reach them?
There was a problem hiding this comment.
I think we discussed documenting https://doc.rust-lang.org/nightly/nightly-rustc/rustc_middle/mir/mono/enum.Visibility.html a while ago, now might be a good time for that? (you had some ideas for that I believe)?
Will open a separate PR for that. Edit: #134261
Also I don't think .hidden and .protected are ever hit by tests, is there any way we could reach them?
You can set the visibility for all symbols using -Zdefault-visibility. Also I'm not sure if using .protected or .hidden in case of internal linkage will cause the symbol to be exported anyway, but that is a pre-existing issue.
There was a problem hiding this comment.
yeah that's fine, we just want to make sure that naked functions behave like standard functions
|
☔ The latest upstream changes (presumably #134381) made this pull request unmergeable. Please resolve the merge conflicts. |
bjorn3
force-pushed
the
naked_asm_improvements
branch
from
42ae211 to
1358f53
Compare
bjorn3
force-pushed
the
naked_asm_improvements
branch
from
1358f53 to
dcf8b87
Compare
|
r? compiler |
|
☔ The latest upstream changes (presumably #135959) made this pull request unmergeable. Please resolve the merge conflicts. |
While LLVM is rather permissive in this regards, some other codegen backends demand that once you declare a function for definition you actually define contents of the function, which doesn't happen for naked functions as we actually generate assembly for them.
This allows it to be reused by codegen backends that don't use cg_ssa like cg_clif.
bjorn3
force-pushed
the
naked_asm_improvements
branch
from
dcf8b87 to
0b43964
Compare
This was introduced in #128004.