Arthur Cohen [Thu, 24 Aug 2023 15:49:42 +0000 (17:49 +0200)]
nr2.0: Add new ImmutableNameResolutionCtx class.
gcc/rust/ChangeLog:
* Make-lang.in: Compile it.
* resolve/rust-immutable-name-resolution-context.cc: New file.
* resolve/rust-immutable-name-resolution-context.h: New file.
jjasmine [Sun, 25 Feb 2024 11:10:37 +0000 (03:10 -0800)]
Split up rust-macro-builtins.cc
Fixes issue #2855
gcc/rust/ChangeLog:
* Make-lang.in: add new .o builds for new .cc files
* expand/rust-cfg-strip.h (RUST_CFG_STRIP_H): Add include guards
for rust-cfg-strip
* expand/rust-macro-builtins.cc (make_macro_path_str): moved to new respective files
(make_token): moved to new respective files
(make_string): moved to new respective files
(macro_end_token): moved to new respective files
(try_extract_string_literal_from_fragment): moved to new respective files
(try_expand_many_expr): moved to new respective files
(parse_single_string_literal): moved to new respective files
(source_relative_path): moved to new respective files
(load_file_bytes): moved to new respective files
(MacroBuiltin::assert_handler): moved to new respective files
(MacroBuiltin::file_handler): moved to new respective files
(MacroBuiltin::column_handler): moved to new respective files
(MacroBuiltin::include_bytes_handler): moved to new respective files
(MacroBuiltin::include_str_handler): moved to new respective files
(MacroBuiltin::compile_error_handler): moved to new respective files
(MacroBuiltin::concat_handler): moved to new respective files
(MacroBuiltin::env_handler): moved to new respective files
(MacroBuiltin::cfg_handler): moved to new respective files
(MacroBuiltin::include_handler): moved to new respective files
(MacroBuiltin::line_handler): moved to new respective files
(MacroBuiltin::stringify_handler): moved to new respective files
(struct FormatArgsInput): moved to new respective files
(struct FormatArgsParseError): moved to new respective files
(format_args_parse_arguments): moved to new respective files
(MacroBuiltin::format_args_handler): moved to new respective files
* expand/rust-macro-builtins.h (builtin_macro_from_string):
merge tl::optional from master
* expand/rust-macro-builtins-asm.cc: New file.
* expand/rust-macro-builtins-format-args.cc: New file.
* expand/rust-macro-builtins-helpers.cc: New file.
* expand/rust-macro-builtins-helpers.h: New file.
* expand/rust-macro-builtins-include.cc: New file.
* expand/rust-macro-builtins-location.cc: New file.
* expand/rust-macro-builtins-log-debug.cc: New file.
* expand/rust-macro-builtins-test-bench.cc: New file.
* expand/rust-macro-builtins-trait.cc: New file.
* expand/rust-macro-builtins-utility.cc: New file.
Arthur Cohen [Thu, 7 Mar 2024 13:57:54 +0000 (14:57 +0100)]
format-args: Only pass the format string to the parser.
This fixes an issue we had where the generated code ended with more static
pieces than its rustc counterpart.
gcc/rust/ChangeLog:
* expand/rust-macro-builtins.cc (struct FormatArgsInput): Store the format_str
as a string instead of an AST::Expr.
(format_args_parse_arguments): Transform format_expr into a format string
properly - add note for handling eager macro invocations later on.
(MacroBuiltin::format_args_handler): Parse the correct input, append
newline to format_str if necessary.
Arthur Cohen [Thu, 29 Feb 2024 12:01:32 +0000 (13:01 +0100)]
format-args: Add base for expanding FormatArgs nodes
gcc/rust/ChangeLog:
* Make-lang.in: Add new object.
* hir/rust-ast-lower-expr.cc (ASTLoweringExpr::visit): Remove calls to
FormatArgsLowering.
* expand/rust-expand-format-args.cc: New file.
* expand/rust-expand-format-args.h: New file.
Merge "macro: Use MacroInvocation's node_id in ExternalItem constructor", to
avoid bootstrap failure:
[...]
In file included from [...]/source-gcc/gcc/rust/ast/rust-expr.h:6,
from [...]/source-gcc/gcc/rust/ast/rust-ast-full.h:24,
from [...]/source-gcc/gcc/rust/ast/rust-ast.cc:24:
In copy constructor ‘Rust::AST::MacroInvocation::MacroInvocation(const Rust::AST::MacroInvocation&)’,
inlined from ‘Rust::AST::MacroInvocation* Rust::AST::MacroInvocation::clone_macro_invocation_impl() const’ at [...]/source-gcc/gcc/rust/ast/rust-\
macro.h:798:38:
[...]/source-gcc/gcc/rust/ast/rust-macro.h:734:39: error: ‘*(Rust::AST::MacroInvocation*)<unknown>.Rust::AST::MacroInvocation::Rust::AST::ExprWithout\
Block.Rust::AST::ExprWithoutBlock::Rust::AST::Expr.Rust::AST::Expr::node_id’ is used uninitialized [-Werror=uninitialized]
734 | builtin_kind (other.builtin_kind)
| ^
cc1plus: all warnings being treated as errors
make[3]: *** [[...]/source-gcc/gcc/rust/Make-lang.in:423: rust/rust-ast.o] Error 1
[...]
In file included from [...]/source-gcc/gcc/rust/ast/rust-expr.h:6,
from [...]/source-gcc/gcc/rust/ast/rust-item.h:27,
from [...]/source-gcc/gcc/rust/parse/rust-parse.h:20,
from [...]/source-gcc/gcc/rust/expand/rust-macro-expand.h:24,
from [...]/source-gcc/gcc/rust/expand/rust-macro-expand.cc:19:
In copy constructor ‘Rust::AST::MacroInvocation::MacroInvocation(const Rust::AST::MacroInvocation&)’,
inlined from ‘Rust::AST::MacroInvocation* Rust::AST::MacroInvocation::clone_macro_invocation_impl() const’ at [...]/source-gcc/gcc/rust/ast/rust-\
macro.h:798:38:
[...]/source-gcc/gcc/rust/ast/rust-macro.h:734:39: error: ‘*(Rust::AST::MacroInvocation*)<unknown>.Rust::AST::MacroInvocation::Rust::AST::ExprWithout\
Block.Rust::AST::ExprWithoutBlock::Rust::AST::Expr.Rust::AST::Expr::node_id’ is used uninitialized [-Werror=uninitialized]
734 | builtin_kind (other.builtin_kind)
| ^
cc1plus: all warnings being treated as errors
make[3]: *** [[...]/source-gcc/gcc/rust/Make-lang.in:433: rust/rust-macro-expand.o] Error 1
[...]
Richard Biener [Mon, 27 Nov 2023 09:20:02 +0000 (10:20 +0100)]
tree-optimization/112653 - PTA and return
The following separates the escape solution for return stmts not
only during points-to solving but also for later querying. This
requires adjusting the points-to-global tests to include escapes
through returns. Technically the patch replaces the existing
post-processing which computes the transitive closure of the
returned value solution by a proper artificial variable with
transitive closure constraints. Instead of adding the solution
to escaped we track it separately.
PR tree-optimization/112653
* gimple-ssa.h (gimple_df): Add escaped_return solution.
* tree-ssa.cc (init_tree_ssa): Reset it.
(delete_tree_ssa): Likewise.
* tree-ssa-structalias.cc (escaped_return_id): New.
(find_func_aliases): Handle non-IPA return stmts by
adding to ESCAPED_RETURN.
(set_uids_in_ptset): Adjust HEAP escaping to also cover
escapes through return.
(init_base_vars): Initialize ESCAPED_RETURN.
(compute_points_to_sets): Replace ESCAPED post-processing
with recording the ESCAPED_RETURN solution.
* tree-ssa-alias.cc (ref_may_alias_global_p_1): Check
the ESCAPED_RETUNR solution.
(dump_alias_info): Dump it.
* cfgexpand.cc (update_alias_info_with_stack_vars): Update it.
* ipa-icf.cc (sem_item_optimizer::fixup_points_to_sets):
Likewise.
* tree-inline.cc (expand_call_inline): Reset it.
* tree-parloops.cc (parallelize_loops): Likewise.
* tree-sra.cc (maybe_add_sra_candidate): Check it.
Martin Jambor [Fri, 24 Nov 2023 16:32:35 +0000 (17:32 +0100)]
sra: SRA of non-escaped aggregates passed by reference to calls
PR109849 shows that a loop that heavily pushes and pops from a stack
implemented by a C++ std::vec results in slow code, mainly because the
vector structure is not split by SRA and so we end up in many loads
and stores into it. This is because it is passed by reference
to (re)allocation methods and so needs to live in memory, even though
it does not escape from them and so we could SRA it if we
re-constructed it before the call and then separated it to distinct
replacements afterwards.
This patch does exactly that, first relaxing the selection of
candidates to also include those which are addressable but do not
escape and then adding code to deal with the calls. The
micro-benchmark that is also the (scan-dump) testcase in this patch
runs twice as fast with it than with current trunk. Honza measured
its effect on the libjxl benchmark and it almost closes the
performance gap between Clang and GCC while not requiring excessive
inlining and thus code growth.
The patch disallows creation of replacements for such aggregates which
are also accessed with a precision smaller than their size because I
have observed that this led to excessive zero-extending of data
leading to slow-downs of perlbench (on some CPUs). Apart from this
case I have not noticed any regressions, at least not so far.
Gimple call argument flags can tell if an argument is unused (and then
we do not need to generate any statements for it) or if it is not
written to and then we do not need to generate statements loading
replacements from the original aggregate after the call statement.
Unfortunately, we cannot symmetrically use flags that an aggregate is
not read because to avoid re-constructing the aggregate before the
call because flags don't tell which what parts of aggregates were not
written to, so we load all replacements, and so all need to have the
correct value before the call.
This version of the patch also takes care to avoid attempts to modify
abnormal edges, something which was missing in the previosu version.
gcc/ChangeLog:
2023-11-23 Martin Jambor <mjambor@suse.cz>
PR middle-end/109849
* tree-sra.cc (passed_by_ref_in_call): New.
(sra_initialize): Allocate passed_by_ref_in_call.
(sra_deinitialize): Free passed_by_ref_in_call.
(create_access): Add decl pool candidates only if they are not
already candidates.
(build_access_from_expr_1): Bail out on ADDR_EXPRs.
(build_access_from_call_arg): New function.
(asm_visit_addr): Rename to scan_visit_addr, change the
disqualification dump message.
(scan_function): Check taken addresses for all non-call statements,
including phi nodes. Process all call arguments, including the static
chain, build_access_from_call_arg.
(maybe_add_sra_candidate): Relax need_to_live_in_memory check to allow
non-escaped local variables.
(sort_and_splice_var_accesses): Disallow smaller-than-precision
replacements for aggregates passed by reference to functions.
(sra_modify_expr): Use a separate stmt iterator for adding satements
before the processed statement and after it.
(enum out_edge_check): New type.
(abnormal_edge_after_stmt_p): New function.
(sra_modify_call_arg): New function.
(sra_modify_assign): Adjust calls to sra_modify_expr.
(sra_modify_function_body): Likewise, use sra_modify_call_arg to
process call arguments, including the static chain.
This kind of double indirection is pointless and prone to error. This
commit change the api of all getters from the AST to use references
directly instead of references to unique pointers.
Replace reference to unique pointer with reference
Reference to unique pointers are a known anti pattern, only the element
shall be taken by reference instead of the whole wrapper.
gcc/rust/ChangeLog:
* ast/rust-item.h: Change getter function prototype to return a
reference directly instead of a reference to the wrapper type.
* checks/errors/rust-ast-validation.cc (ASTValidation::visit): Fix
the code to accept references instead.
* hir/rust-ast-lower-base.cc (ASTLoweringBase::lower_self): Change
function implementation to return a reference.
* hir/rust-ast-lower-base.h: Accept a reference instead of a unique
pointer reference.
* resolve/rust-ast-resolve-item.cc (ResolveItem::visit): Adapt the code
to a reference instead of a unique pointer.
Thomas Schwinge [Tue, 12 Mar 2024 08:36:43 +0000 (09:36 +0100)]
Placate clang-format re 'gcc/rust/backend/rust-tree.cc'
Reformat the upstream GCC commit f4a2ae2338962208b8039f154f5912402e94c378
"Change MODE_BITSIZE to MODE_PRECISION for MODE_VECTOR_BOOL" change to
'gcc/rust/backend/rust-tree.cc' to clang-format's liking.
Jakub Jelinek [Wed, 18 Oct 2023 15:01:26 +0000 (17:01 +0200)]
cse: Workaround GCC < 5 bug in cse_insn [PR111852]
Before the r5-3834 commit for PR63362, GCC 4.8-4.9 refuses to compile
cse.cc which contains a variable with rtx_def type, because rtx_def
contains a union with poly_uint16 element. poly_int template has
defaulted default constructor and a variadic template constructor which
could have empty parameter pack. GCC < 5 treated it as non-trivially
constructible class and deleted rtunion and rtx_def default constructors.
For the cse_insn purposes, all we need is a variable with size and alignment
of rtx_def, not necessarily rtx_def itself, which we then memset to 0 and
fill in like rtx is normally allocated from heap, so this patch for
GCC_VERSION < 5000 uses an unsigned char array of the right size/alignment.
2023-10-18 Jakub Jelinek <jakub@redhat.com>
PR bootstrap/111852
* cse.cc (cse_insn): Add workaround for GCC 4.8-4.9, instead of
using rtx_def type for memory_extend_buf, use unsigned char
arrayy with size of rtx_def and its alignment.
projects / thirdparty / gcc.git / log
