There are now four static/shared scenarios that can happen for the
supported LLVM versions:
- 3.9+: By default use `llvm-config --link-static`
- 3.9+ and `--enable-llvm-link-shared`: Use `--link-shared` instead.
- 3.8: Use `llvm-config --shared-mode` and go with its answer.
- 3.7: Just assume static, maintaining the status quo.
The makefiles and `mklldeps.py` called `llvm-config --shared-mode` to
find out if LLVM defaulted to shared or static libraries, and just went
with that. But under rustbuild, `librustc_llvm/build.rs` was assuming
that LLVM should be static, and even forcing `--link-static` for 3.9+.
Now that build script also uses `--shared-mode` to learn the default,
which should work better for pre-3.9 configured for dynamic linking, as
it wasn't possible back then to choose differently via `llvm-config`.
Further, the configure script now has a new `--enable-llvm-link-shared`
option, which allows one to manually override `--link-shared` on 3.9+
instead of forcing static.
rustc: Flag all builtins functions as hidden
When compiling compiler-rt you typically compile with `-fvisibility=hidden`
which to ensure that all symbols are hidden in shared objects and don't show up
in symbol tables. This is important for these intrinsics being linked in every
crate to ensure that we're not unnecessarily bloating the public ABI of Rust
crates.
This should help allow the compiler-builtins project with Rust-defined builtins
start landing in-tree as well.
to let people experiment with this target out of tree.
The MSP430 architecture is used in 16-bit microcontrollers commonly used
in Digital Signal Processing applications.
When compiling compiler-rt you typically compile with `-fvisibility=hidden`
which to ensure that all symbols are hidden in shared objects and don't show up
in symbol tables. This is important for these intrinsics being linked in every
crate to ensure that we're not unnecessarily bloating the public ABI of Rust
crates.
This should help allow the compiler-builtins project with Rust-defined builtins
start landing in-tree as well.
The `Linkage` enum in librustc_llvm got out of sync with the version in LLVM and it caused two variants of the #[linkage=""] attribute to break.
This adds the functions `LLVMRustGetLinkage` and `LLVMRustSetLinkage` which convert between the Rust Linkage enum and the LLVM one, which should stop this from breaking every time LLVM changes it.
Fixes#33992
A new target, `s390x-unknown-linux-gnu`, has been added to the compiler
and can be used to build no_core/no_std Rust programs.
Known limitations:
- librustc_trans/cabi_s390x.rs is missing. This means no support for
`extern "C" fn`.
- No support for this arch in libc. This means std can be cross compiled
for this target.
Macro expansions produce code tagged with debug locations that are completely different from the surrounding expressions. This wrecks havoc on debugger's ability the step over source lines.
In order to have a good line stepping behavior in debugger, we overwrite debug locations of macro expansions with that of the outermost expansion site.
LLVM upgrade
As discussed in https://internals.rust-lang.org/t/need-help-with-emscripten-port/3154/46 I'm trying to update the used LLVM checkout in Rust.
I basically took @shepmaster's code and applied it on top (though I did the commits manually, the [original commits have better descriptions](https://github.com/rust-lang/rust/compare/master...avr-rust:avr-support).
With these changes I was able to build rustc. `make check` throws one last error on `run-pass/issue-28950.rs`. Output: https://gist.github.com/badboy/bcdd3bbde260860b6159aa49070a9052
I took the metadata changes as is and they seem to work, though it now uses the module in another step. I'm not sure if this is the best and correct way.
Things to do:
* [x] ~~Make `run-pass/issue-28950.rs` pass~~ unrelated
* [x] Find out how the `PositionIndependentExecutable` setting is now used
* [x] Is the `llvm::legacy` still the right way to do these things?
cc @brson @alexcrichton
This is to pull in changes to support ARM MUSL targets.
This change also commits a couple of other cargo-generated changes
to other dependencies in the various Cargo.toml files.
This is a spiritual succesor to #34268/8531d581, in which we replaced a
number of matches of None to the unit value with `if let` conditionals
where it was judged that this made for clearer/simpler code (as would be
recommended by Manishearth/rust-clippy's `single_match` lint). The same
rationale applies to matches of None to the empty block.
Compute `target_feature` from LLVM
This is a work-in-progress fix for #31662.
The logic that computes the target features from the command line has been replaced with queries to the `TargetMachine`.
When reuing a definition across codegen units, we obviously cannot use
internal linkage, but using external linkage means that we can end up
with multiple conflicting definitions of a single symbol across
multiple crates. Since the definitions should all be equal
semantically, we can use weak_odr linkage to resolve the situation.
Fixes#32518
We use a 64bit integer to pass the set of attributes that is to be
removed, but the called C function expects a 32bit integer. On most
platforms this doesn't cause any problems other than being unable to
unset some attributes, but on ARM even the lower 32bit aren't handled
correctly because the 64bit value is passed in different registers, so
the C function actually sees random garbage.
So we need to fix the relevant functions to use 32bit integers instead.
Additionally we need an implementation that actually accepts 64bit
integers because some attributes can only be unset that way.
Fixes#32360
`fast` a.k.a UnsafeAlgebra is the flag for enabling all "unsafe"
(according to llvm) float optimizations.
See LangRef for more information http://llvm.org/docs/LangRef.html#fast-math-flags
Providing these operations with less precise associativity rules (for
example) is useful to numerical applications.
For example, the summation loop:
let sum = 0.;
for element in data {
sum += *element;
}
Using the default floating point semantics, this loop expresses the
floats must be added in a sequence, one after another. This constraint
is usually completely unintended, and it means that no autovectorization
is possible.
Unfortunately on i686-pc-windows-gnu LLVM's answer to `--host-target` is
`x86_64-pc-windows-gnu` even though we're building in a 32-bit shell as well as
compiling 32-bit libraries. For now use Cargo's `HOST` environment variable to
determine whether we're doing a cross compilation or not.
Zeroing on-drop seems to work fine. Still thinking about the best way to approach zeroing on-move.
(based on top of the other drop PR; only the last 2 commits are relevant)
