abi: add a rust-preserve-none calling convention
This is the conceptual opposite of the rust-cold calling convention and is particularly useful in combination with the new `explicit_tail_calls` feature.
For relatively tight loops implemented with tail calling (`become`) each of the function with the regular calling convention is still responsible for restoring the initial value of the preserved registers. So it is not unusual to end up with a situation where each step in the tail call loop is spilling and reloading registers, along the lines of:
foo:
push r12
; do things
pop r12
jmp next_step
This adds up quickly, especially when most of the clobberable registers are already used to pass arguments or other uses.
I was thinking of making the name of this ABI a little less LLVM-derived and more like a conceptual inverse of `rust-cold`, but could not come with a great name (`rust-cold` is itself not a great name: cold in what context? from which perspective? is it supposed to mean that the function is rarely called?)
This is the conceptual opposite of the rust-cold calling convention and
is particularly useful in combination with the new `explicit_tail_calls`
feature.
For relatively tight loops implemented with tail calling (`become`) each
of the function with the regular calling convention is still responsible
for restoring the initial value of the preserved registers. So it is not
unusual to end up with a situation where each step in the tail call loop
is spilling and reloading registers, along the lines of:
foo:
push r12
; do things
pop r12
jmp next_step
This adds up quickly, especially when most of the clobberable registers
are already used to pass arguments or other uses.
I was thinking of making the name of this ABI a little less LLVM-derived
and more like a conceptual inverse of `rust-cold`, but could not come
with a great name (`rust-cold` is itself not a great name: cold in what
context? from which perspective? is it supposed to mean that the
function is rarely called?)
Add checks for gpu-kernel calling conv
The `gpu-kernel` calling convention has several restrictions that were not enforced by the compiler until now.
Add the following restrictions:
1. Cannot be async
2. Cannot be called
3. Cannot return values, return type must be `()` or `!`
4. Arguments should be simple, i.e. passed by value. More complicated types can work when you know what you are doing, but it is rather unintuitive, one needs to know ABI/compiler internals.
5. Export name should be unmangled, either through `no_mangle` or `export_name`. Kernels are searched by name on the CPU side, having a mangled name makes it hard to find and probably almost always unintentional.
Tracking issue: rust-lang/rust#135467
amdgpu target tracking issue: rust-lang/rust#135024
``@workingjubilee,`` these should be all the restrictions we talked about a year ago.
cc ``@RDambrosio016`` ``@kjetilkjeka`` for nvptx
The `gpu-kernel` calling convention has several restrictions that were
not enforced by the compiler until now.
Add the following restrictions:
1. Cannot be async
2. Cannot be called
3. Cannot return values, return type must be `()` or `!`
4. Arguments should be primitives, i.e. passed by value. More complicated
types can work when you know what you are doing, but it is rather
unintuitive, one needs to know ABI/compiler internals.
5. Export name should be unmangled, either through `no_mangle` or
`export_name`. Kernels are searched by name on the CPU side, having
a mangled name makes it hard to find and probably almost always
unintentional.
All usages of `memory_index` start by calling `invert_bijective_mapping`, so
storing the inverted mapping directly saves some work and simplifies the code.
The `fmt::Debug` impl for `TyAndLayout<'a, Ty>'` requires `fmt::Display`
on the `Ty` parameter. In `ArgAbi`, `TyAndLayout`'s Ty` is instantiated
with a parameter that implements `TyAbiInterface`. `TyAbiInterface`
only required `fmt::Debug` be implemented on `Self`, not `fmt::Display`,
which meant that it wasn't actually possible to debug print `ArgAbi`.
Introduces `BackendRepr::ScalableVector` corresponding to scalable
vector types annotated with `repr(scalable)` which lowers to a scalable
vector type in LLVM.
Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>
Extend parsing of `ReprOptions` with `rustc_scalable_vector(N)` which
optionally accepts a single literal integral value - the base multiple of
lanes that are in a scalable vector. Can only be applied to structs.
Co-authored-by: Jamie Cunliffe <Jamie.Cunliffe@arm.com>
Unify the configuration of the compiler docs
Previously it was rather inconsistent which crates got the rust logo and which didn't and setting html_root_url was forgotten in many cases.
FCW for repr(C) enums whose discriminant values do not fit into a c_int or c_uint
Context: https://github.com/rust-lang/rust/issues/124403
The current behavior of repr(C) enums is as follows:
- The discriminant values are interpreted as const expressions of type `isize`
- We compute the smallest size that can hold all discriminant values
- The target spec contains the smallest size for repr(C) enums
- We take the larger of these two sizes
Unfortunately, this doesn't always match what C compilers do. In particular, MSVC seems to *always* give enums a size of 4 bytes, whereas the algorithm above will give enums a size of up to 8 bytes on 64bit targets. Here's an example enum affected by this:
```
// We give this size 4 on 32bit targets (with a warning since the discriminant is wrapped to fit an isize)
// and size 8 on 64bit targets.
#[repr(C)]
enum OverflowingEnum {
A = 9223372036854775807, // i64::MAX
}
// MSVC always gives this size 4 (without any warning).
// GCC always gives it size 8 (without any warning).
// Godbolt: https://godbolt.org/z/P49MaYvMd
enum overflowing_enum {
OVERFLOWING_ENUM_A = 9223372036854775807,
};
```
If we look at the C standard, then up until C20, there was no official support enums without an explicit underlying type and with discriminants that do not fit an `int`. With C23, this has changed: now enums have to grow automatically if there is an integer type that can hold all their discriminants. MSVC does not implement this part of C23.
Furthermore, Rust fundamentally cannot implement this (without major changes)! Enum discriminants work fundamentally different in Rust and C:
- In Rust, every enum has a discriminant type entirely determined by its repr flags, and then the discriminant values must be const expressions of that type. For repr(C), that type is `isize`. So from the outset we interpret 9223372036854775807 as an isize literal and never give it a chance to be stored in a bigger type. If the discriminant is given as a literal without type annotation, it gets wrapped implicitly with a warning; otherwise the user has to write `as isize` explicitly and thus trigger the wrapping. Later, we can then decide to make the *tag* that stores the discriminant smaller than the discriminant type if all discriminant values fit into a smaller type, but those values have allready all been made to fit an `isize` so nothing bigger than `isize` could ever come out of this. That makes the behavior of 32bit GCC impossible for us to match.
- In C, things flow the other way around: every discriminant value has a type determined entirely by its constant expression, and then the type for the enum is determined based on that. IOW, the expression can have *any type* a priori, different variants can even use a different type, and then the compiler is supposed to look at the resulting *values* (presumably as mathematical integers) and find a type that can hold them all. For the example above, 9223372036854775807 is a signed integer, so the compiler looks for the smallest signed type that can hold it, which is `long long`, and then uses that to compute the size of the enum (at least that's what C23 says should happen and GCC does this correctly).
Realistically I think the best we can do is to not attempt to support C23 enums, and to require repr(C) enums to satisfy the C20 requirements: all discriminants must fit into a c_int. So that's what this PR implements, by adding a FCW for enums with discriminants that do not fit into `c_int`. As a slight extension, we do *not* lint enums where all discriminants fit into a `c_uint` (i.e. `unsigned int`): while C20 does (in my reading) not allow this, and C23 does not prescribe the size of such an enum, this seems to behave consistently across compilers (giving the enum the size of an `unsigned int`). IOW, the lint fires whenever our layout algorithm would make the enum larger than an `int`, irrespective of whether we pick a signed or unsigned discriminant. This extension was added because [crater found](https://github.com/rust-lang/rust/pull/147017#issuecomment-3357077199) multiple cases of such enums across the ecosystem.
Note that it is impossible to trigger this FCW on targets where isize and c_int are the same size (i.e., the typical 32bit target): since we interpret discriminant values as isize, by the time we look at them, they have already been wrapped. However, we have an existing lint (overflowing_literals) that should notify people when this kind of wrapping occurs implicitly. Also, 64bit targets are much more common. On the other hand, even on 64bit targets it is possible to fall into the same trap by writing a literal that is so big that it does not fit into isize, gets wrapped (triggering overflowing_literals), and the wrapped value fits into c_int. Furthermore, overflowing_literals is just a lint, so if it occurs in a dependency you won't notice. (Arguably there is also a more general problem here: for literals of type `usize`/`isize`, it is fairly easy to write code that only triggers `overflowing_literals` on 32bit targets, and to never see that lint if one develops on a 64bit target.)
Specifically, the above example triggers the FCW on 64bit targets, but on 32bit targets we get this err-by-default lint instead (which will be hidden if it occurs in a dependency):
```
error: literal out of range for `isize`
--> $DIR/repr-c-big-discriminant1.rs:16:9
|
LL | A = 9223372036854775807,
| ^^^^^^^^^^^^^^^^^^^
|
= note: the literal `9223372036854775807` does not fit into the type `isize` whose range is `-2147483648..=2147483647`
= note: `#[deny(overflowing_literals)]` on by default
```
Also see the tests added by this PR.
This isn't perfect, but so far I don't think I have seen a better option. In https://github.com/rust-lang/rust/pull/146504 I tried adjusting our enum logic to make the size of the example enum above actually match what C compilers do, but that's a massive breaking change since we have to change the expected type of the discriminant expression from `isize` to `i64` or even `i128` -- so that seems like a no-go. To improve the lint we could analyze things on the HIR level and specifically catch "repr(C) enums with discriminants defined as literals that are too big", but that would have to be on top of the lint in this PR I think since we'd still want to also always check the actually evaluated value (which we can't always determined on the HIR level).
Cc `@workingjubilee` `@CAD97`
remove explicit deref of AbiAlign for most methods
Much of the compiler calls functions on Align projected from AbiAlign. AbiAlign impls Deref to its inner Align, so we can simplify these away. Also, it will minimize disruption when AbiAlign is removed.
For now, preserve usages that might resolve to PartialOrd or PartialEq, as those have odd inference.
Much of the compiler calls functions on Align projected from AbiAlign.
AbiAlign impls Deref to its inner Align, so we can simplify these away.
Also, it will minimize disruption when AbiAlign is removed.
For now, preserve usages that might resolve to PartialOrd or PartialEq,
as those have odd inference.
stabilize c-style varargs for sysv64, win64, efiapi, aapcs
This has been split up so the PR now only contains the extended_varargs_abi_support stabilization; "system" has been moved to https://github.com/rust-lang/rust/pull/145954.
**Previous (combined) PR description:**
This stabilizes extern block declarations of variadic functions with the system, sysv64, win64, efiapi, aapcs ABIs. This corresponds to the extended_varargs_abi_support and extern_system_varargs feature gates.
The feature gates were split up since it seemed like there might be further discussion needed for what exactly "system" ABI variadic functions should do, but a [consensus](https://github.com/rust-lang/rust/issues/136946#issuecomment-2967847553) has meanwhile been reached: they shall behave like "C" functions. IOW, the ABI of a "system" function is (bold part is new in this PR):
- "stdcall" for win32 targets **for non-variadic functions**
- "C" for everything else
This had been previously stabilized *without FCP* in https://github.com/rust-lang/rust/pull/116161, which got reverted in https://github.com/rust-lang/rust/pull/136897. There was also a "fun" race condition involved with the system ABI being [added](https://github.com/rust-lang/rust/pull/119587) to the list of variadic-supporting ABIs between the creation and merge of rust-lang/rust#116161.
There was a question raised [here](https://github.com/rust-lang/rust/pull/116161#issuecomment-1983829513) whether t-lang even needs to be involved for a change like this. Not sure if that has meanwhile been clarified? The behavior of the "system" ABI (a Rust-specific ABI) definitely feels like t-lang territory to me.
Fixesrust-lang/rust#100189
Cc `@rust-lang/lang`
# Stabilization report
> ## General design
> ### What is the RFC for this feature and what changes have occurred to the user-facing design since the RFC was finalized?
AFAIK there is no RFC. The tracking issues are
- https://github.com/rust-lang/rust/issues/100189
- https://github.com/rust-lang/rust/issues/136946
> ### What behavior are we committing to that has been controversial? Summarize the major arguments pro/con.
The only controversial point is whether "system" ABI functions should support variadics.
- Pro: This allows crates like windows-rs to consistently use "system", see e.g. https://github.com/microsoft/windows-rs/issues/3626.
- Cons: `@workingjubilee` had some implementation concerns, but I think those have been [resolved](https://github.com/rust-lang/rust/issues/136946#issuecomment-2967847553). EDIT: turns out Jubilee still has concerns (she mentioned that in a DM); I'll let her express those.
Note that "system" is already a magic ABI we introduced to "do the right thing". This just makes it do the right thing in more cases. In particular, it means that on Windows one can almost always just do
```rust
extern "system" {
// put all the things here
}
```
and it'll do the right thing, rather than having to split imports into non-varargs and varargs, with the varargs in a separate `extern "C"` block (and risking accidentally putting a non-vararg there).
(I am saying "almost" always because some Windows API functions actually use cdecl, not stdcall, on x86. Those of course need to go in `extern "C"` blocks.)
> ### Are there extensions to this feature that remain unstable? How do we know that we are not accidentally committing to those?
Actually defining variadic functions in Rust remains unstable, under the [c_variadic feature gate](https://github.com/rust-lang/rust/issues/44930).
> ## Has a Call for Testing period been conducted? If so, what feedback was received?
>
> Does any OSS nightly users use this feature? For instance, a useful indication might be "search <grep.app> for `#![feature(FEATURE_NAME)]` and had `N` results".
There was no call for testing.
A search brings up https://github.com/rust-osdev/uefi-rs/blob/main/uefi-raw/src/table/boot.rs using this for "efiapi". This doesn't seem widely used, but it is an "obvious" gap in our support for c-variadics.
> ## Implementation quality
All rustc does here is forward the ABI to LLVM so there's lot a lot to say here...
> ### Summarize the major parts of the implementation and provide links into the code (or to PRs)
>
> An example for async closures: <https://rustc-dev-guide.rust-lang.org/coroutine-closures.html>.
The check for allowed variadic ABIs is [here](9c870d30e2/compiler/rustc_hir_analysis/src/lib.rs (L109-L126)).
The special handling of "system" is [here](c24914ec83/compiler/rustc_target/src/spec/abi_map.rs (L82-L85)).
> ### Summarize existing test coverage of this feature
>
> Consider what the "edges" of this feature are. We're particularly interested in seeing tests that assure us about exactly what nearby things we're not stabilizing.
>
> Within each test, include a comment at the top describing the purpose of the test and what set of invariants it intends to demonstrate. This is a great help to those reviewing the tests at stabilization time.
>
> - What does the test coverage landscape for this feature look like?
> - Tests for compiler errors when you use the feature wrongly or make mistakes?
> - Tests for the feature itself:
> - Limits of the feature (so failing compilation)
> - Exercises of edge cases of the feature
> - Tests that checks the feature works as expected (where applicable, `//@ run-pass`).
> - Are there any intentional gaps in test coverage?
>
> Link to test folders or individual tests (ui/codegen/assembly/run-make tests, etc.).
Prior PRs add a codegen test for all ABIs and tests actually calling extern variadic functions for sysv64 and win64:
- https://github.com/rust-lang/rust/pull/144359
- https://github.com/rust-lang/rust/pull/144379
We don't have a way of executing uefi target code in the test suite, so it's unclear how to fully test efiapi. aapcs could probably be done? (But note that we have hardly an such actually-calling-functions tests for ABI things, we almost entirely rely on codegen tests.)
The test ensuring that we do *not* stabilize *defining* c-variadic functions is `tests/ui/feature-gates/feature-gate-c_variadic.rs`.
> ### What outstanding bugs in the issue tracker involve this feature? Are they stabilization-blocking?
None that I am aware of.
> ### What FIXMEs are still in the code for that feature and why is it ok to leave them there?
None that I am aware of.
> ### Summarize contributors to the feature by name for recognition and assuredness that people involved in the feature agree with stabilization
`@Soveu` added sysv64, win64, efiapi, aapcs to the list of ABIs that allow variadics, `@beepster4096` added system. `@workingjubilee` recently refactored the ABI handling in the compiler, also affecting this feature.
> ### Which tools need to be adjusted to support this feature. Has this work been done?
>
> Consider rustdoc, clippy, rust-analyzer, rustfmt, rustup, docs.rs.
Maybe RA needs to be taught about the new allowed ABIs? No idea how precisely they mirror what exactly rustc accepts and rejects here.
> ## Type system and execution rules
> ### What compilation-time checks are done that are needed to prevent undefined behavior?
>
> (Be sure to link to tests demonstrating that these tests are being done.)
Nothing new here, this just expands the existing support for calling variadic functions to more ABIs.
> ### Does the feature's implementation need checks to prevent UB or is it sound by default and needs opt in in places to perform the dangerous/unsafe operations? If it is not sound by default, what is the rationale?
Nothing new here, this just expands the existing support for calling variadic functions to more ABIs.
> ### Can users use this feature to introduce undefined behavior, or use this feature to break the abstraction of Rust and expose the underlying assembly-level implementation? (Describe.)
Nothing new here, this just expands the existing support for calling variadic functions to more ABIs.
> ### What updates are needed to the reference/specification? (link to PRs when they exist)
- https://github.com/rust-lang/reference/pull/1936
> ## Common interactions
> ### Does this feature introduce new expressions and can they produce temporaries? What are the lifetimes of those temporaries?
No.
> ### What other unstable features may be exposed by this feature?
None.
This was done in #145740 and #145947. It is causing problems for people
using r-a on anything that uses the rustc-dev rustup package, e.g. Miri,
clippy.
This repository has lots of submodules and subtrees and various
different projects are carved out of pieces of it. It seems like
`[workspace.dependencies]` will just be more trouble than it's worth.
Remove TmpLayout in layout_of_enum
09a3846 from <https://github.com/rust-lang/rust/pull/103693> made LayoutData be owned instead of interned in `Variants::Multiple::variants`[^1], so there's no need for `TmpLayout` in layout_of_enum anymore, and we can just store the variants' layouts directly in the prospective `LayoutData`s' `variants` fields.
This should have no effect on semantics or layout.
(written as part of rust-lang/rust#145337 but not related to the layout optimizations in that PR)
[^1]: see line 1154 of `compiler/rustc_target/src/abi/mod.rs` in the linked commit; `Variants::Multiple::variants` effectively changed from `IndexVec<.., Layout<'tcx>>` to `IndexVec<.., LayoutData>` where the `LayoutData`s are not interned as `Layout`s (`LayoutData` was at the time called `LayoutS`)
`rustc_errors` depends on numerous crates, solely to implement its
`IntoDiagArg` trait on types from those crates. Many crates depend on
`rustc_errors`, and it's on the critical path.
We can't swap things around to make all of those crates depend on
`rustc_errors` instead, because `rustc_errors` would end up in
dependency cycles.
Instead, move `IntoDiagArg` into `rustc_error_messages`, which has far
fewer dependencies, and then have most of these crates depend on
`rustc_error_messages`.
This allows `rustc_errors` to drop dependencies on several crates,
including the large `rustc_target`.
(This doesn't fully reduce dependency chains yet, as `rustc_errors`
still depends on `rustc_hir` which depends on `rustc_target`. That will
get fixed in a subsequent commit.)
Remove dead code and extend test coverage and diagnostics around it
I was staring a bit at the `dont_niche_optimize_enum` variable and figured out that part of it is dead code (at least today it is). I changed the diagnostic and test around the code that makes that part dead code, so everything that makes removing that code sound is visible in this PR
We lost the following comment during refactorings:
The current code for niche-filling relies on variant indices instead of actual discriminants, so enums with explicit discriminants (RFC 2363) would misbehave.
For example, transmuting between `bool` and `Ordering` doesn't need two `assume`s because one range is a superset of the other.
Multiple are still used for things like `char` <-> `NonZero<u32>`, which overlap but where neither fully contains the other.
Allow custom default address spaces and parse `p-` specifications in the datalayout string
Some targets, such as CHERI, use as default an address space different from the "normal" default address space `0` (in the case of CHERI, [200 is used](https://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-877.pdf)). Currently, `rustc` does not allow to specify custom address spaces and does not take into consideration [`p-` specifications in the datalayout string](https://llvm.org/docs/LangRef.html#langref-datalayout).
This patch tries to mitigate these problems by allowing targets to define a custom default address space (while keeping the default value to address space `0`) and adding the code to parse the `p-` specifications in `rustc_abi`. The main changes are that `TargetDataLayout` now uses functions to refer to pointer-related informations, instead of having specific fields for the size and alignment of pointers in the default address space; furthermore, the two `pointer_size` and `pointer_align` fields in `TargetDataLayout` are replaced with an `FxHashMap` that holds info for all the possible address spaces, as parsed by the `p-` specifications.
The potential performance drawbacks of not having ad-hoc fields for the default address space will be tested in this PR's CI run.
r? workingjubilee
