Jason Merrill [Thu, 20 Mar 2025 16:57:15 +0000 (12:57 -0400)]
ipa: target clone and mangling alias [PR114992]
Since the mangling of the second lambda changed (previously we counted all
lambdas, now we only count lambdas with the same signature), we
generate_mangling_alias for handler<lambda2> for backward compatibility.
Since handler is COMDAT, resolve_alias puts the alias in the same comdat
group as handler itself. Then create_dispatcher_calls tries to add the
alias to the same comdat group as the dispatcher, but it's already in a
same_comdat_group, so we ICE.
It seems like we're just missing a remove_from_same_comdat_group before
add_to_same_comdat_group.
PR c++/114992
gcc/ChangeLog:
* multiple_target.cc (create_dispatcher_calls):
remove_from_same_comdat_group before add_to_same_comdat_group.
Paul Thomas [Fri, 21 Mar 2025 16:20:21 +0000 (16:20 +0000)]
Fortran: Implement the F2018 reduce intrinsic [PR85836]
2025-03-21 Paul Thomas <pault@gcc.gnu.org>
gcc/fortran
PR fortran/85836
* check.cc (get_ul_from_cst_cl): New function used in
check_operation.
(check_operation): New function used in check_reduce and
check_co_reduce.
(gfc_check_co_reduce): Use it.
(gfc_check_reduce): New function.
(gfc_check_rename): Add prototype for intrinsic with 6 arguments.
* gfortran.h : Add isym id for reduce and prototype for f6.
* intrinsic.cc (do_check): Add another argument expression and use
it in the call to the six argument specific check.
(add_sym_6): New function.
(add_functions): Add the discription of the reduce intrinsic and
add it to the intrinsic list.
* intrinsic.h : Add prototypes for gfc_check_reduce and
gfc_resolve_reduce.
* iresolve.cc (generate_reduce_op_wrapper): Generate a wrapper
subroutine for the 'operation' function to enable the library
implementation to be type agnostic and use pointer arithmetic
throughout.
(gfc_resolve_reduce): New function.
* trans-expr.cc (gfc_conv_procedure_call): Add flag for scalar
reduce. Generate a return variable 'sr' for scalar reduce, pass its
address to the library function and return it as the scalar result.
* trans-intrinsic.cc (gfc_conv_intrinsic_function): Array valued
reduce is called in same way as reshape. Fall through for call to
the scalar version.
gcc/testsuite/
PR fortran/85836
* gfortran.dg/reduce_1.f90: New test
* gfortran.dg/reduce_2.f90: New test
libgfortran/
PR libfortran/85836
* Makefile.am : Add reduce.c
* Makefile.in : Regenerated
* gfortran.map : Add _gfortran_reduce, _gfortran_reduce_scalar,
_gfortran_reduce_c and _gfortran_reduce_scalar_c to the list.
* intrinsics/reduce.c (reduce, reduce_scalar, reduce_c,
reduce_scalar_c): New functions and prototypes
Richard Earnshaw [Fri, 21 Mar 2025 15:20:03 +0000 (15:20 +0000)]
arm: testsuite: make unaligned-memcpy-*.c executable tests [PR91614]
These tests have been looking for a very specific instruction sequence
which has the tendency to be fairly unstable as a result. But what is
more interesting is that the the tests must not contain instructions
that can't be used for unaligned data, and whether or not the copy is
executed correctly.
So make these tests executable and scan the assembler only to confirm
the absence of instructions that must not be used when the data is not
aligned.
These tests also used to be restricted to targets that support
unaligned accesses (because you get very different code otherwise).
But now we've made the tests executable and to check for the absence
of problem instructions, just falling back to memcpy *is* an
acceptable implementation. So remove the requirement for unaligned
accesses.
gcc/testsuite:
PR target/91614
* gcc.target/arm/unaligned-memcpy-1.c: Make the test executable.
Only scan for the absence of instructions that cannot access
misaligned data. Remove constraint of having unaligned accesses.
* gcc.target/arm/unaligned-memcpy-2.c: Likewise.
* gcc.target/arm/unaligned-memcpy-3.c: Likewise.
* gcc.target/arm/unaligned-memcpy-4.c: Likewise.
This test is designed to check that if one of the operands is
aligned (but the other isn't) we expand to a sensible sequence and
bypass most of the overhead of doing a memcpy. But on targets without
unaligned accessess, we still end up calling memcpy. It's then a
lottery as to whether the prologue and epilogue code, plus the
set-up for the memcpy itself, generate instructions that match the
scan patterns.
Since in those cases we're not actually testing what the test is looking
for anyway, just skip the test on strict-alignment targets.
Thomas Schwinge [Wed, 19 Mar 2025 11:18:26 +0000 (12:18 +0100)]
C++: Adjust implicit '__cxa_bad_cast' prototype to reality
In 2001 Subversion r40924 (Git commit 52a11cbfcf0cfb32628b6953588b6af4037ac0b6)
"IA-64 ABI Exception Handling", '__cxa_bad_cast' changed from 'void *' to
'void' return type:
--- libstdc++-v3/libsupc++/exception_support.cc
+++ /dev/null
@@ -1,388 +0,0 @@
-[...]
-// Helpers for rtti. Although these don't return, we give them return types so
-// that the type system is not broken.
-extern "C" void *
-__cxa_bad_cast ()
-{
- [...]
-}
-[...]
..., which is in conflict with the library code with 'void' return type:
// BEGIN GLOBAL FUNCTION DECL: __cxa_bad_cast
.visible .func __cxa_bad_cast;
// BEGIN GLOBAL FUNCTION DEF: __cxa_bad_cast
.visible .func __cxa_bad_cast
{
[...]
}
..., and we thus get execution test FAIL for 'g++.dg/rtti/dyncast2.C':
error : Prototype doesn't match for '__cxa_bad_cast' in 'input file 7 at offset 51437', first defined in 'input file 7 at offset 51437'
nvptx-run: cuLinkAddData failed: device kernel image is invalid (CUDA_ERROR_INVALID_SOURCE, 300)
With this patched, we get the expected:
// BEGIN GLOBAL FUNCTION DECL: __cxa_bad_cast
-.extern .func (.param .u64 %value_out) __cxa_bad_cast;
+.extern .func __cxa_bad_cast;
... next to '-malias' variant: commit a1865fd33897bc6c6e0109df0a12ee73ce386315
"Add 'g++.target/nvptx/alias-g++.dg_init_dtor2-1.C'", to document what we're
doing to '-mno-alias'.
Tobias Burnus [Fri, 21 Mar 2025 12:54:49 +0000 (13:54 +0100)]
testsuite/lib/libgomp.exp: compile with -fdiagnostics-plain-output
libgomp.exp added -fno-diagnostics-show-caret and -fdiagnostics-color=never
as 'additional_flags' for compilation. However, it turned out that this now
is insufficient as the [...] part of diagnostics have a hyperlink URL.
Solution: Use the -fdiagnostics-plain-output flag instead, added in commit r11-2701-g129a1319c0ab73. This flag currently implies the following flags:
-fno-diagnostics-show-caret
-fno-diagnostics-show-line-numbers
-fdiagnostics-color=never
-fdiagnostics-urls=never
-fdiagnostics-path-format=separate-events
-fdiagnostics-text-art-charset=none
-fno-diagnostics-show-event-links
libgomp/ChangeLog:
* testsuite/lib/libgomp.exp (libgomp_init): Add
-fdiagnostics-plain-output to additional_flags; remove
-fno-diagnostics-show-caret and -fdiagnostics-color=never.
Arthur Cohen [Fri, 3 Jan 2025 15:14:45 +0000 (15:14 +0000)]
gccrs: derive(Clone): Implement derive clone for enum struct variants
gcc/rust/ChangeLog:
* expand/rust-derive-clone.cc (DeriveClone::clone_enum_struct): New function for deriving
enum struct variants.
(DeriveClone::visit_enum): Call into the new function.
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude:
* rust/compile/derive_clone_enum1.rs: New test.
* rust/compile/derive_clone_enum2.rs: New test.
* rust/compile/derive_clone_enum3.rs: New test.
* rust/execute/torture/derive_clone_enum1.rs: New test.
Arthur Cohen [Thu, 26 Dec 2024 21:46:03 +0000 (21:46 +0000)]
gccrs: ast: Refactor how lang item paths are handled.
Lang item typepaths were not handled properly, and required a complete overhaul.
All old classes that concerned lang item paths are now modified to use a simpler
version of `AST::LangItemPath`, which has been removed. TypePath segments can now
be lang items, as this is requied for having generic lang item paths such as
PhantomData<T>.
Arthur Cohen [Tue, 23 Jan 2024 16:19:31 +0000 (17:19 +0100)]
gccrs: Fix scan-gimple testcases on LE platforms.
gcc/testsuite/ChangeLog:
* rust/compile/macros/builtin/eager1.rs: Switch to scan-assembler directive as the
GIMPLE dump does not contain strings on LE.
* rust/compile/macros/builtin/recurse2.rs: Likewise.
Owen Avery [Fri, 15 Nov 2024 00:57:42 +0000 (19:57 -0500)]
gccrs: Add ForeverStackStore
ForeverStackStore is meant to partially unify the internal states of
per-namespace ForeverStack instances. This commit does not contain
modifications to ForeverStack which would allow it to rely on a
ForeverStackStore to store nodes, but a future commit should address
this.
gcc/rust/ChangeLog:
* Make-lang.in: Handle rust-forever-stack.cc.
* resolve/rust-forever-stack.h
(class ForeverStackStore): Add.
* resolve/rust-forever-stack.cc: New file, based on
rust-forever-stack.hxx.
liushuyu [Mon, 2 Dec 2024 21:52:35 +0000 (14:52 -0700)]
gccrs: add two more tests to test try-catch (unwind) code generation
gcc/testsuite/ChangeLog:
* rust/compile/try-catch-unwind-old.rs: add a test to test the older
try intrinsics from plain old Rust to v1.78.0
* rust/compile/try-catch-unwind-new.rs: add a test to test the newer
catch_unwind instrinsics since Rust v1.78.0
liushuyu [Mon, 2 Dec 2024 21:24:04 +0000 (14:24 -0700)]
gccrs: rust/intrinsic: add new "catch_unwind" variant of API
gcc/rust/ChangeLog:
* backend/rust-compile-intrinsic.cc: add the new `catch_unwind` variant
of the `try` intrinsic: this variant can be seen on Rust 1.78+
and returns `()` instead of `i32`.
Philip Herron [Thu, 9 Jan 2025 16:47:47 +0000 (16:47 +0000)]
gccrs: match arms are a LUB
Unify rules are not the same as coercion rules. The coercion of ! is
allowed to any type but not for a unify site which is different.
Match arms are another least upper bound coercion.
gcc/rust/ChangeLog:
* backend/rust-compile-expr.cc (CompileExpr::visit): implement coercion
* typecheck/rust-hir-type-check-expr.cc (TypeCheckExpr::visit): this is an LUB
* typecheck/rust-unify.cc (UnifyRules::go): remove unify ! coercion
Signed-off-by: Philip Herron <herron.philip@googlemail.com>
Philip Herron [Tue, 7 Jan 2025 18:15:37 +0000 (18:15 +0000)]
gccrs: cleanup our enum type layout to be closer to rustc
This changes our enum type layout so for example:
enum Foo {
A,
B,
C(char),
D { x: i32, y: i32 },
}
Used to get layed out like this in gccrs:
union {
struct A { int RUST$ENUM$DISR; };
struct B { int RUST$ENUM$DISR; };
struct C { int RUST$ENUM$DISR; char __0; };
struct D { int RUST$ENUM$DISR; i64 x; i64 y; };
}
This has some issues notably with the constexpr because this is just a
giant union it means its not simple to constify what enum variant we are
looking at because the discriminant is a mess.
This now gets layed out as:
struct {
int RUST$ENUM$DISR;
union {
struct A { };
struct B { };
struct C { char __0; };
struct D { i64 x; i64 y; };
} payload;
}
This layout is much cleaner and allows for our constexpr to work properly.
gcc/rust/ChangeLog:
* backend/rust-compile-expr.cc (CompileExpr::visit): new layout
* backend/rust-compile-pattern.cc (CompilePatternCheckExpr::visit): likewise
(CompilePatternBindings::visit): likewise
* backend/rust-compile-resolve-path.cc: likewise
* backend/rust-compile-type.cc (TyTyResolveCompile::visit): implement new layout
* rust-gcc.cc (constructor_expression): get rid of useless assert
Signed-off-by: Philip Herron <herron.philip@googlemail.com>
Owen Avery [Tue, 7 Jan 2025 19:03:13 +0000 (14:03 -0500)]
gccrs: nr2.0: Resolve Self inside impl blocks
gcc/rust/ChangeLog:
* resolve/rust-toplevel-name-resolver-2.0.cc
(TopLevel::visit): Insert a definition for Self when visiting
InherentImpl and TraitImpl instances.
* resolve/rust-toplevel-name-resolver-2.0.h
(TopLevel::visit): Add visitors for InherentImpl and TraitImpl.
Philip Herron [Thu, 19 Dec 2024 16:43:49 +0000 (16:43 +0000)]
gccrs: add support for lang_item eq and PartialEq trait
The Eq and Partial Ord are very similar to the operator overloads
we support for add/sub/etc... but they differ in that usually the
function call name matches the name of the lang item. This time
we need to have support to send in a new path for the method call
on the lang item we want instead of just the name of the lang item.
NOTE: this test case doesnt work correctly yet we need to support
the derive of partial eq on enums to generate the correct comparison
code for that.
Fixes Rust-GCC#3302
gcc/rust/ChangeLog:
* backend/rust-compile-expr.cc (CompileExpr::visit): handle partial_eq possible call
* backend/rust-compile-expr.h: handle case where lang item calls differ from name
* hir/tree/rust-hir-expr.cc (OperatorExprMeta::OperatorExprMeta): new helper
* hir/tree/rust-hir-expr.h: likewise
* typecheck/rust-hir-type-check-expr.cc (TypeCheckExpr::visit): handle partial_eq
(TypeCheckExpr::resolve_operator_overload): likewise
* typecheck/rust-hir-type-check-expr.h: likewise
* util/rust-lang-item.cc (LangItem::ComparisonToLangItem): map comparison to lang item
(LangItem::ComparisonToSegment): likewise
* util/rust-lang-item.h: new lang items PartialOrd and Eq
* util/rust-operators.h (enum class): likewise
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude: nr2 cant handle this
* rust/compile/cmp1.rs: New test.
Signed-off-by: Philip Herron <herron.philip@googlemail.com>
Owen Avery [Thu, 2 Jan 2025 10:55:38 +0000 (05:55 -0500)]
gccrs: nr2.0: Improve default, top-level, and late resolvers
gcc/rust/ChangeLog:
* resolve/rust-default-resolver.cc
(DefaultResolver::visit): Make sure to scope visitation of the
children of type definition items.
* resolve/rust-default-resolver.h
(DefaultResolver::visit): Add overrides for TupleStruct, Union,
and TypeAlias.
* resolve/rust-late-name-resolver-2.0.cc
(Late::visit): Remove override for Enum.
* resolve/rust-late-name-resolver-2.0.h
(Late::visit): Likewise.
* resolve/rust-toplevel-name-resolver-2.0.cc
(TopLevel::visit): Rely more on DefaultResolver::visit.
* resolve/rust-toplevel-name-resolver-2.0.h
(TopLevel::visit): Remove override for BlockExpr.
Arthur Cohen [Sat, 21 Dec 2024 22:56:52 +0000 (22:56 +0000)]
gccrs: ast: Add new Kind enums for more precise downcasting
This commit adds things like Item::Kind, Expr::Kind, etc, and implements the associated `get_*_kind` functions.
It also removes the more generic AST::Kind enum we were using, which was incomplete and painful to use.
Arthur Cohen [Wed, 25 Dec 2024 17:55:09 +0000 (17:55 +0000)]
gccrs: lower: Correctly lower parenthesized types
This is useful for handling multiple trait bounds, and required for better handling of auto traits.
gcc/rust/ChangeLog:
* hir/rust-ast-lower-type.cc (ASTLoweringType::visit): Add implementation for
ParenthesizedType.
* hir/rust-ast-lower-type.h: Declare that new visitor.
Liam Naddell [Tue, 17 Dec 2024 16:48:03 +0000 (11:48 -0500)]
gccrs: Fix NR2.0 compiler ICE caused by Generics in Enums
gcc/rust/ChangeLog:
* resolve/rust-late-name-resolver-2.0.cc:
Change the late name resolver to enter proper lexical scope during typechecking
* resolve/rust-late-name-resolver-2.0.h:
Add needed prototype to header
* resolve/rust-toplevel-name-resolver-2.0.cc:
Add generic parameters to enum's scoped RIB to allow for proper name resolution on types.
gcc/testsuite/ChangeLog:
* rust/compile/issue-3304.rs:
Add small test for generics+enums combination for NR2.0
Arthur Cohen [Wed, 18 Dec 2024 11:20:27 +0000 (12:20 +0100)]
gccrs: lang-items: Collect trait functions that are lang items
gcc/rust/ChangeLog:
* ast/rust-collect-lang-items.cc (CollectLangItems::visit): Add visitor for collecting
functions that might be lang items.
* ast/rust-collect-lang-items.h: Likewise.
Arthur Cohen [Wed, 25 Dec 2024 11:03:44 +0000 (11:03 +0000)]
gccrs: resolve: Name resolve trait bounds properly
gcc/rust/ChangeLog:
* resolve/rust-ast-resolve-type.cc (ResolveTypeToCanonicalPath::visit): Resolve additional
trait bounds.
* resolve/rust-late-name-resolver-2.0.cc (Late::visit): Error out properly on unresolved
type-path instead of crashing.
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude: Exclude additional-trait-bounds2 for different error message.
* rust/compile/additional-trait-bounds1.rs: New test.
* rust/compile/additional-trait-bounds2.rs: New test.
* rust/compile/additional-trait-bounds2nr2.rs: New test.
Philip Herron [Fri, 13 Dec 2024 15:51:55 +0000 (15:51 +0000)]
gccrs: implement the TuplePattern and use it for function patterns
In order to handle the tuple pattern of: fn test ((x _) : (i32, i32)) -> i32 { x }
we need to recognize that ABI wise this function still takes a tuple as the parameter
to this function its just how we can address the "pattern" of the tuple changes.
So reall if this was C it would look like:
void test (struct tuple_type __prameter)
{
return __parameter.0
}
The code here reuses our existing pattern code so that we generate these implicit
bindings of the paramter with a field access so any time x is referenced it's really
just emplacing __parameter.0 for the field access into the struct which is a tuple.
Fixes Rust-GCC#2847
gcc/rust/ChangeLog:
* backend/rust-compile-fnparam.cc (CompileFnParam::visit): compile tuple patterns
(CompileSelfParam::compile): update return type
(CompileFnParam::create_tmp_param_var): return Bvariable not tree to stop ICE
* backend/rust-compile-fnparam.h: update prototype
* backend/rust-compile-pattern.cc (CompilePatternBindings::visit): implement TuplePattern
* backend/rust-compile-pattern.h: update prototype
gcc/testsuite/ChangeLog:
* rust/compile/issue-2847.rs: New test.
Signed-off-by: Philip Herron <herron.philip@googlemail.com>
Dhruv Chawla [Wed, 19 Mar 2025 16:34:09 +0000 (09:34 -0700)]
aarch64: Add support for -mcpu=olympus
This adds support for the NVIDIA Olympus core to the AArch64 backend. The
initial patch does not add any special tuning decisions, and those may come
later.
Bootstrapped and tested on aarch64-none-linux-gnu.
gcc/ChangeLog:
* config/aarch64/aarch64-cores.def (olympus): New entry.
* config/aarch64/aarch64-tune.md: Regenerate.
* doc/invoke.texi (AArch64 Options): Document the above.
Arthur Cohen [Wed, 6 Nov 2024 15:32:35 +0000 (16:32 +0100)]
gccrs: hir: Start adapting visitors to accept multiple kinds of Paths
gcc/rust/ChangeLog:
* ast/rust-item.h: Add new method to specifically get a type-path.
* ast/rust-path.cc (LangItemPath::as_string): Implement properly.
* hir/rust-ast-lower-type.cc (ASTLowerTypePath::translate): Adapt
visitor to use the new LangItemPath.
* hir/rust-ast-lower-type.h: Likewise.
* resolve/rust-ast-resolve-item.cc (ResolveItem::visit): Likewise.
* resolve/rust-ast-resolve-type.h: Likewise.
* resolve/rust-forever-stack.h
(ForeverStack::find_starting_point): Use type
'std::reference_wrapper<Node> &' instead of 'Node &' for
parameter starting_point.
* resolve/rust-forever-stack.hxx
(ForeverStack::find_starting_point): Likewise.
(ForeverStack::resolve_path): Handle change to
ForeverStack::find_starting_point.
Arthur Cohen [Mon, 25 Nov 2024 14:05:32 +0000 (15:05 +0100)]
gccrs: lang-items: Store NodeId mappings for lang items
gcc/rust/ChangeLog:
* util/rust-hir-map.h: Keep a NodeId mappings for lang items.
* util/rust-hir-map.cc (Mappings::insert_lang_item_node): New function.
(Mappings::lookup_lang_item_node): Likewise.
Philip Herron [Mon, 2 Dec 2024 16:17:54 +0000 (16:17 +0000)]
gccrs: constant evaluation like these are coercion sites
The code here was wrongly assuming the decl type from the folding of the
expression would be the type of the constant decl. This is not the case for
unsized coercions for slices, where the expression here is a reference to
an array then we require the coercion to fix the result up to the expected
type.
Fixes Rust-GCC#1525
gcc/rust/ChangeLog:
* backend/rust-compile-base.cc: apply coercion site to result
* backend/rust-compile-base.h: update prototype
* backend/rust-compile-implitem.cc (CompileTraitItem::visit): send in coercion info
* backend/rust-compile-item.cc (CompileItem::visit): likewise
gcc/testsuite/ChangeLog:
* rust/compile/issue-1525.rs: New test.
Signed-off-by: Philip Herron <herron.philip@googlemail.com>
Philip Herron [Mon, 2 Dec 2024 14:10:11 +0000 (14:10 +0000)]
gccrs: Remove bad assertion in name resolution
This was a handy debug assertion but only works for valid rust code. This
needs to handle the case where the type is not resolved which is a valid
case.
projects / thirdparty / gcc.git / log
