| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The following patch implements C++23 P1774R8 - Portable assumptions
paper, by introducing support for [[assume (cond)]]; attribute for C++.
In addition to that the patch adds [[gnu::assume (cond)]]; and
__attribute__((assume (cond))); support to both C and C++.
As described in C++23, the attribute argument is conditional-expression
rather than the usual assignment-expression for attribute arguments,
the condition is contextually converted to bool (for C truthvalue conversion
is done on it) and is never evaluated at runtime.
For C++ constant expression evaluation, I only check the simplest conditions
for undefined behavior, because otherwise I'd need to undo changes to
*ctx->global which happened during the evaluation (but I believe the spec
allows that and we can further improve later).
The patch uses a new internal function, .ASSUME, to hold the condition
in the FEs. At gimplification time, if the condition is simple/without
side-effects, it is gimplified as if (cond) ; else __builtin_unreachable ();
and otherwise for now dropped on the floor. The intent is to incrementally
outline the conditions into separate artificial functions and use
.ASSUME further to tell the ranger and perhaps other optimization passes
about the assumptions, as detailed in the PR.
When implementing it, I found that assume entry hasn't been added to
https://eel.is/c++draft/cpp.cond#6
Jonathan said he'll file a NB comment about it, this patch assumes it
has been added into the table as 202207L when the paper has been voted in.
With the attributes for both C/C++, I'd say we don't need to add
__builtin_assume with similar purpose, especially when __builtin_assume
in LLVM is just weird. It is strange for side-effects in function call's
argument not to be evaluated, and LLVM in that case (annoyingly) warns
and ignores the side-effects (but doesn't do then anything with it),
if there are no side-effects, it will work like our
if (!cond) __builtin_unreachable ();
2022-10-06 Jakub Jelinek <jakub@redhat.com>
PR c++/106654
gcc/
* internal-fn.def (ASSUME): New internal function.
* internal-fn.h (expand_ASSUME): Declare.
* internal-fn.cc (expand_ASSUME): Define.
* gimplify.cc (gimplify_call_expr): Gimplify IFN_ASSUME.
* fold-const.h (simple_condition_p): Declare.
* fold-const.cc (simple_operand_p_2): Rename to ...
(simple_condition_p): ... this. Remove forward declaration.
No longer static. Adjust function comment and fix a typo in it.
Adjust recursive call.
(simple_operand_p): Adjust function comment.
(fold_truth_andor): Adjust simple_operand_p_2 callers to call
simple_condition_p.
* doc/extend.texi: Document assume attribute. Move fallthrough
attribute example to its section.
gcc/c-family/
* c-attribs.cc (handle_assume_attribute): New function.
(c_common_attribute_table): Add entry for assume attribute.
* c-lex.cc (c_common_has_attribute): Handle
__have_cpp_attribute (assume).
gcc/c/
* c-parser.cc (handle_assume_attribute): New function.
(c_parser_declaration_or_fndef): Handle assume attribute.
(c_parser_attribute_arguments): Add assume_attr argument,
if true, parse first argument as conditional expression.
(c_parser_gnu_attribute, c_parser_std_attribute): Adjust
c_parser_attribute_arguments callers.
(c_parser_statement_after_labels) <case RID_ATTRIBUTE>: Handle
assume attribute.
gcc/cp/
* cp-tree.h (process_stmt_assume_attribute): Implement C++23
P1774R8 - Portable assumptions. Declare.
(diagnose_failing_condition): Declare.
(find_failing_clause): Likewise.
* parser.cc (assume_attr): New enumerator.
(cp_parser_parenthesized_expression_list): Handle assume_attr.
Remove identifier variable, for id_attr push the identifier into
expression_list right away instead of inserting it before all the
others at the end.
(cp_parser_conditional_expression): New function.
(cp_parser_constant_expression): Use it.
(cp_parser_statement): Handle assume attribute.
(cp_parser_expression_statement): Likewise.
(cp_parser_gnu_attribute_list): Use assume_attr for assume
attribute.
(cp_parser_std_attribute): Likewise. Handle standard assume
attribute like gnu::assume.
* cp-gimplify.cc (process_stmt_assume_attribute): New function.
* constexpr.cc: Include fold-const.h.
(find_failing_clause_r, find_failing_clause): New functions,
moved from semantics.cc with ctx argument added and if non-NULL,
call cxx_eval_constant_expression rather than fold_non_dependent_expr.
(cxx_eval_internal_function): Handle IFN_ASSUME.
(potential_constant_expression_1): Likewise.
* pt.cc (tsubst_copy_and_build): Likewise.
* semantics.cc (diagnose_failing_condition): New function.
(find_failing_clause_r, find_failing_clause): Moved to constexpr.cc.
(finish_static_assert): Use it. Add auto_diagnostic_group.
gcc/testsuite/
* gcc.dg/attr-assume-1.c: New test.
* gcc.dg/attr-assume-2.c: New test.
* gcc.dg/attr-assume-3.c: New test.
* g++.dg/cpp2a/feat-cxx2a.C: Add colon to C++20 features
comment, add C++20 attributes comment and move C++20
new features after the attributes before them.
* g++.dg/cpp23/feat-cxx2b.C: Likewise. Test
__has_cpp_attribute(assume).
* g++.dg/cpp23/attr-assume1.C: New test.
* g++.dg/cpp23/attr-assume2.C: New test.
* g++.dg/cpp23/attr-assume3.C: New test.
* g++.dg/cpp23/attr-assume4.C: New test.
|
|
For BImode get_narrowest_mode evaluates to QImode but BImode < QImode.
Thus FOR_EACH_MODE_UNTIL never reaches BImode and iterates until OImode
for which no wider mode exists so we end up with VOIDmode and fail.
Fixed by adding a size guard so we effectively skip BImode.
gcc/ChangeLog:
PR rtl-optimization/107088
* cselib.cc (new_cselib_val): Skip BImode while keeping track of
subvalue relations.
|
|
We recently agreed that setting a range of NAN should instead set
UNDEFINED for -ffinite-math-only. This patch makes that change to
frange::set_nan() directly. Also, calling frange::update_nan() will now
be a nop for !HONOR_NANS.
Doing this in the setters simplifies everywhere we set NANs, as it keeps
us from introducing NANs by mistake.
gcc/ChangeLog:
* value-range.cc (frange::set): Call set_nan unconditionally.
(range_tests_nan): Adjust tests.
(range_tests_signed_zeros): Same.
(range_tests_floats): Same.
* value-range.h (frange::update_nan): Guard with HONOR_NANS.
(frange::set_nan): Set undefined if !HONOR_NANS.
|
|
The uses of finite_operands_p removed are guarded by a call to
finite_operands_p already.
gcc/ChangeLog:
* range-op-float.cc (foperator_lt::fold_range): Remove extra check
to finite_operands_p.
(foperator_le::fold_range): Same.
(foperator_gt::fold_range): Same.
(foperator_ge::fold_range): Same.
|
|
gcc/ChangeLog:
* value-range-pretty-print.cc (vrange_printer::print_real_value):
Avoid printing INF and NAN twice.
|
|
|
|
2022-10-05 Segher Boessenkool <segher@kernel.crashing.org>
* config/rs6000/constraints.md (wD): Delete.
* doc/md.texi (Machine Constraints): Adjust.
|
|
Extracting the left and right halfs of a vector are entirely different
operations. Things are simpler if they are separate define_insns, and
it is easy to get rid of the "wD" constraint use then.
This also give the variant that is a no-op copy its own alternative, of
length 0 (and this, cost 0, making it more likely RA will choose it.
2022-10-05 Segher Boessenkool <segher@kernel.crashing.org>
* config/rs6000/vsx.md (vsx_extract_<mode>): Replace define_insn by a
define_expand. Split the contents to...
(*vsx_extract_<mode>_0): ... this. Rewrite.
(*vsx_extract_<mode>_1): ... and this. Rewrite.
|
|
We can use "n" instead of "wD" if we simply test the value of the
integer constant directly.
2022-10-05 Segher Boessenkool <segher@kernel.crashing.org>
* config/rs6000/vsx.md (*vsx_extract_<mode>_store): Use "n" instead of
"wD" constraint.
|
|
PR analyzer/107158 reports an ICE when using
-fanalyzer -fanalyzer-call-summaries
on a particular source file.
It turns out I just fixed this ICE in r13-3094-g6832c95c0e1a58.
This followup patch adds a somewhat reduced reproducer as a regression
test. Unfortunately, although the ICE is fixed, there are two false
positives from -Wanalyzer-malloc-leak on the test case, so I'm going to
use PR analyzer/107158 for tracking those false positives.
gcc/testsuite/ChangeLog:
PR analyzer/107158
* gcc.dg/analyzer/call-summaries-pr107158.c: New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
gcc/analyzer/ChangeLog:
* analysis-plan.cc: Simplify includes.
* analyzer-pass.cc: Likewise.
* analyzer-selftests.cc: Likewise.
* analyzer.cc: Likewise.
* analyzer.h: Add includes of "json.h" and "tristate.h".
* call-info.cc: Simplify includes.
* call-string.cc: Likewise.
* call-summary.cc: Likewise.
* checker-path.cc: Likewise.
* complexity.cc: Likewise.
* constraint-manager.cc: Likewise.
* diagnostic-manager.cc: Likewise.
* engine.cc: Likewise.
* feasible-graph.cc: Likewise.
* known-function-manager.cc: Likewise.
* pending-diagnostic.cc: Likewise.
* program-point.cc: Likewise.
* program-state.cc: Likewise.
* region-model-asm.cc: Likewise.
* region-model-impl-calls.cc: Likewise.
* region-model-manager.cc: Likewise.
* region-model-reachability.cc: Likewise.
* region-model.cc: Likewise.
* region-model.h: Include "selftest.h".
* region.cc: Simplify includes.
* sm-fd.cc: Likewise.
* sm-file.cc: Likewise.
* sm-malloc.cc: Likewise.
* sm-pattern-test.cc: Likewise.
* sm-sensitive.cc: Likewise.
* sm-signal.cc: Likewise.
* sm-taint.cc: Likewise.
* sm.cc: Likewise.
* state-purge.cc: Likewise.
* store.cc: Likewise.
* store.h: Likewise.
* supergraph.cc: Likewise.
* svalue.cc: Likewise.
* svalue.h: Likewise.
* trimmed-graph.cc: Likewise.
* varargs.cc: Likewise.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
This doesn't fix the various false positives seen with
-fanalyzer-call-summaries on PR 107060, but stops it crashing at -O2.
gcc/analyzer/ChangeLog:
PR analyzer/107060
* call-summary.cc
(call_summary_replay::convert_svalue_from_summary_1): Handle NULL
results from convert_svalue_from_summary in SK_UNARY_OP and
SK_BIN_OP.
* engine.cc (impl_region_model_context::on_unknown_change): Bail
out on svalues that can't have associated state.
* region-model-impl-calls.cc
(region_model::impl_call_analyzer_get_unknown_ptr): New.
* region-model.cc (region_model::on_stmt_pre): Handle
"__analyzer_get_unknown_ptr".
* region-model.h
(region_model::impl_call_analyzer_get_unknown_ptr): New decl.
* store.cc (store::replay_call_summary_cluster): Avoid trying to
create binding clusters for base regions that shouldn't have them.
gcc/ChangeLog:
PR analyzer/107060
* doc/analyzer.texi (__analyzer_get_unknown_ptr): Document.
gcc/testsuite/ChangeLog:
PR analyzer/107060
* gcc.dg/analyzer/analyzer-decls.h (__analyzer_get_unknown_ptr):
New decl.
* gcc.dg/analyzer/call-summaries-2.c
(test_summarized_writes_param_to_ptr_unknown): New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
libgomp/ChangeLog:
* libgomp.texi (OpenMP 5.1 Impl. Status): Mark 'assume' as 'Y'.
gcc/fortran/ChangeLog:
* dump-parse-tree.cc (show_omp_assumes): New.
(show_omp_clauses, show_namespace): Call it.
(show_omp_node, show_code_node): Handle OpenMP ASSUME.
* gfortran.h (enum gfc_statement): Add ST_OMP_ASSUME,
ST_OMP_END_ASSUME, ST_OMP_ASSUMES and ST_NOTHING.
(gfc_exec_op): Add EXEC_OMP_ASSUME.
(gfc_omp_assumptions): New struct.
(gfc_get_omp_assumptions): New XCNEW #define.
(gfc_omp_clauses, gfc_namespace): Add assume member.
(gfc_resolve_omp_assumptions): New prototype.
* match.h (gfc_match_omp_assume, gfc_match_omp_assumes): New.
* openmp.cc (omp_code_to_statement): Forward declare.
(enum gfc_omp_directive_kind, struct gfc_omp_directive): New.
(gfc_free_omp_clauses): Free assume member and its struct data.
(enum omp_mask2): Add OMP_CLAUSE_ASSUMPTIONS.
(gfc_omp_absent_contains_clause): New.
(gfc_match_omp_clauses): Call it; optionally use passed
omp_clauses argument.
(omp_verify_merge_absent_contains, gfc_match_omp_assume,
gfc_match_omp_assumes, gfc_resolve_omp_assumptions): New.
(resolve_omp_clauses): Call the latter.
(gfc_resolve_omp_directive, omp_code_to_statement): Handle
EXEC_OMP_ASSUME.
* parse.cc (decode_omp_directive): Parse OpenMP ASSUME(S).
(next_statement, parse_executable, parse_omp_structured_block):
Handle ST_OMP_ASSUME.
(case_omp_decl): Add ST_OMP_ASSUMES.
(gfc_ascii_statement): Handle Assumes, optional return
string without '!$OMP '/'!$ACC ' prefix.
* parse.h (gfc_ascii_statement): Add optional bool arg to prototype.
* resolve.cc (gfc_resolve_blocks, gfc_resolve_code): Add
EXEC_OMP_ASSUME.
(gfc_resolve): Resolve ASSUMES directive.
* symbol.cc (gfc_free_namespace): Free omp_assumes member.
* st.cc (gfc_free_statement): Handle EXEC_OMP_ASSUME.
* trans-openmp.cc (gfc_trans_omp_directive): Likewise.
* trans.cc (trans_code): Likewise.
gcc/testsuite/ChangeLog:
* gfortran.dg/gomp/assume-1.f90: New test.
* gfortran.dg/gomp/assume-2.f90: New test.
* gfortran.dg/gomp/assumes-1.f90: New test.
* gfortran.dg/gomp/assumes-2.f90: New test.
|
|
I was wondering how lvalue_kind handles VIEW_CONVERT_EXPR; in cases where
the type actually changes, it should have the same prvalue->xvalue effect as
ARRAY_REF, since we need to materialize a temporary to get an object we can
reinterpret as another type.
Currently this only fires on builtin-shufflevector-3.c, where we use
VIEW_CONVERT_EXPR to reinterpret a vector as an array.
REALPART_EXPR and IMAGPART_EXPR should also be treated like COMPONENT_REF.
PREINCREMENT_EXPR and PREDECREMENT_EXPR should only be applied to glvalues,
but if for some reason they were applied to a prvalue this would be correct.
TRY_CATCH_EXPR around a prvalue is also questionable, but this is the right
handling.
gcc/cp/ChangeLog:
* tree.cc (lvalue_kind) [VIEW_CONVERT_EXPR]: Change prvalue to
xvalue.
|
|
gcc/ChangeLog:
* config.gcc: Add riscv_vector.h.
* config/riscv/riscv-builtins.cc: Add RVV builtin types support.
* config/riscv/riscv-c.cc (riscv_pragma_intrinsic): New function.
(riscv_register_pragmas): Ditto.
* config/riscv/riscv-protos.h (riscv_register_pragmas): Ditto.
(init_builtins): Move declaration from riscv-vector-builtins.h to riscv-protos.h.
(mangle_builtin_type): Ditto.
(verify_type_context): Ditto.
(handle_pragma_vector): New function.
* config/riscv/riscv-vector-builtins.cc (GTY): New variable.
(register_vector_type): New function.
(init_builtins): Add RVV builtin types support.
(handle_pragma_vector): New function.
* config/riscv/riscv-vector-builtins.h (GCC_RISCV_V_BUILTINS_H): Change
name according to file name.
(GCC_RISCV_VECTOR_BUILTINS_H): Ditto.
(init_builtins): Remove declaration in riscv-vector-builtins.h.
(mangle_builtin_type): Ditto.
(verify_type_context): Ditto.
* config/riscv/riscv.cc: Adjust for RVV builtin types support.
* config/riscv/riscv.h (REGISTER_TARGET_PRAGMAS): New macro.
* config/riscv/t-riscv: Remove redundant file including.
* config/riscv/riscv_vector.h: New file.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/pragma-1.c: New test.
* gcc.target/riscv/rvv/base/pragma-2.c: New test.
* gcc.target/riscv/rvv/base/pragma-3.c: New test.
* gcc.target/riscv/rvv/base/user-1.c: New test.
* gcc.target/riscv/rvv/base/user-2.c: New test.
* gcc.target/riscv/rvv/base/user-3.c: New test.
* gcc.target/riscv/rvv/base/user-4.c: New test.
* gcc.target/riscv/rvv/base/user-5.c: New test.
* gcc.target/riscv/rvv/base/user-6.c: New test.
* gcc.target/riscv/rvv/base/vread_csr.c: New test.
* gcc.target/riscv/rvv/base/vwrite_csr.c: New test.
|
|
gcc/ChangeLog:
* range-op.cc (operator_cast::fold_range): Handle truncating casts
for nonzero masks.
|
|
PR c/107156
gcc/ChangeLog:
* attribs.h (lookup_attribute_by_prefix): Support the attribute
starting with underscore (_Noreturn).
|
|
popcount.
PR tree-optimization/107052
gcc/ChangeLog:
* gimple-range-op.cc (cfn_popcount::fold_range): Take into account
nonzero bit mask.
|
|
Track nonzero masks through a cast in range-ops.
PR tree-optimization/107052
gcc/ChangeLog:
* range-op.cc (operator_cast::fold_range): Set nonzero mask.
|
|
It comes from a discrepancy between get_offset_range, which uses a signed
type, and handle_array_ref, which uses an unsigned one, to do computations.
gcc/
PR tree-optimization/106698
* pointer-query.cc (handle_array_ref): Fix handling of low bound.
gcc/testsuite/
* gnat.dg/lto26.adb: New test.
* gnat.dg/lto26_pkg1.ads, gnat.dg/lto26_pkg1.adb: New helper.
* gnat.dg/lto26_pkg2.ads, gnat.dg/lto26_pkg2.adb: Likewise.
|
|
Fixes:
gcc/analyzer/call-summary.h:103:13: warning: private field 'm_called_fn' is not used [-Wunused-private-field]
gcc/analyzer/engine.cc:1631:24: warning: unused parameter 'uncertainty' [-Wunused-parameter]
gcc/analyzer/ChangeLog:
* call-summary.cc (call_summary_replay::call_summary_replay):
Remove unused variable and arguments.
* call-summary.h: Likewise.
* engine.cc (exploded_node::on_stmt): Likewise.
(exploded_node::replay_call_summaries): Likewise.
(exploded_node::replay_call_summary): Likewise.
* exploded-graph.h (class exploded_node): Likewise.
|
|
PR tree-optimization/106679
gcc/testsuite/ChangeLog:
* gcc.dg/tree-prof/cmpsf-1.c: Mark as a known limitation.
|
|
After moving the testglue in commit 9d503515cee, the jump to exit and
abort is too far for the 'b' instruction on Cortex-M0. As most of the
C code would generate a 'bl' instruction instead of a 'b'
instruction, lets do the same for the inline assembler.
The error seen without this patch:
/tmp/cccCRiCl.o: in function `main':
stack-protector-1.c:(.text+0x4e): relocation truncated to fit: R_ARM_THM_JUMP11 against symbol `__wrap_exit' defined in .text section in gcc_tg.o
stack-protector-1.c:(.text+0x50): relocation truncated to fit: R_ARM_THM_JUMP11 against symbol `__wrap_abort' defined in .text section in gcc_tg.o
collect2: error: ld returned 1 exit status
gcc/testsuite/ChangeLog:
* gcc.target/arm/stack-protector-1.c: Use 'bl' instead of 'b'
instruction.
* gcc.target/arm/stack-protector-3.c: Likewise.
Co-Authored-By: Yvan ROUX <yvan.roux@foss.st.com>
Signed-off-by: Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
|
|
When the mapper can't be executed, Windows report the error like:
.../bad-mapper-1.C: error: failed CreateProcess mapper 'this-will-not-work'
On Linux, the same error is reported this way:
.../bad-mapper-1.C: error: failed execvp mapper 'this-will-not-work'
This patch allows both output forms to be accepted.
Patch has been verified on Windows and Linux.
gcc/testsuite:
* g++.dg/modules/bad-mapper-1.C: Also accept CreateProcess.
Co-Authored-By: Yvan ROUX <yvan.roux@foss.st.com>
Signed-off-by: Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
Signed-off-by: Jonathan Yong <10walls@gmail.com>
|
|
When running the DejaGNU testsuite on a toolchain built for native
Windows, the path /dev/null can't be used to open a stream to void.
On native Windows, the resource is instead named "nul".
The error would look like this:
c:/arm-11.3.rel1/bin/../lib/gcc/arm-none-eabi/11.3.1/../../../../arm-none-eabi/bin/ld.exe: cannot find @/dev/null: No such file or directory
Patch has been verified on Windows and Linux.
gcc/testsuite:
* gcc.misc-tests/outputs.exp: Use "@nul" for Windows,
"@/dev/null" for other environments.
Co-Authored-By: Yvan ROUX <yvan.roux@foss.st.com>
Signed-off-by: Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
Signed-off-by: Jonathan Yong <10walls@gmail.com>
|
|
Came across this deprecated symbol when looking around for
-mexplicit-relocs handling in code
Signed-off-by: Vineet Gupta <vineetg@rivosinc.com>
gcc/ChangeLog:
* config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins):
Remove __riscv_cmodel_pic, that deprecated in last version.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/predef-1.c: Remove __riscv_cmodel_pic check.
* gcc.target/riscv/predef-2.c: Ditto.
* gcc.target/riscv/predef-3.c: Ditto.
* gcc.target/riscv/predef-4.c: Ditto.
* gcc.target/riscv/predef-5.c: Ditto.
* gcc.target/riscv/predef-6.c: Ditto.
* gcc.target/riscv/predef-7.c: Ditto.
* gcc.target/riscv/predef-8.c: Ditto.
|
|
With -fanalyzer-call-summaries the analyzer canl attempt to summarize
the effects of some function calls at their call site, rather than
simulate the call directly, which can avoid big slowdowns during
analysis.
Previously, this summarization was extremely simplistic: no attempt
was made to update sm-state, and region_model::update_for_call_summary
would simply set the return value of the function to UNKNOWN, and assume
the function had no side effects.
This patch implements less simplistic summarizations: it tracks each
possible return enode from the called function, and attempts to generate
a successor enode from the callsite for each that have compatible
conditions, mapping state changes in the summary to state changes
at the callsite. It also implements the beginnings of heuristics for
generating user-facing descriptions of a summary e.g.
"when 'foo' returns NULL"
versus:
"when 'foo' returns a heap-allocated buffer"
This still has some bugs, but much more accurately tracks the effects
of a call, and so is an improvement; it should only have an effect
when -fanalyzer-call-summaries is enabled.
As before, -fanalyzer-call-summaries is disabled by default in
analyzer.opt (but enabled by default in the test suite).
gcc/ChangeLog:
PR analyzer/107072
* Makefile.in (ANALYZER_OBJS): Add analyzer/call-summary.o.
gcc/analyzer/ChangeLog:
PR analyzer/107072
* analyzer-logging.h: Include "diagnostic-core.h".
* analyzer.h: Include "function.h".
(class call_summary): New forward decl.
(class call_summary_replay): New forward decl.
(struct per_function_data): New forward decl.
(struct interesting_t): New forward decl.
(custom_edge_info::update_state): New vfunc.
* call-info.cc (custom_edge_info::update_state): New.
* call-summary.cc: New file.
* call-summary.h: New file.
* constraint-manager.cc: Include "analyzer/call-summary.h".
(class replay_fact_visitor): New.
(constraint_manager::replay_call_summary): New.
* constraint-manager.h (constraint_manager::replay_call_summary):
New.
* engine.cc: Include "analyzer/call-summary.h".
(exploded_node::on_stmt): Handle call summaries.
(class call_summary_edge_info): New.
(exploded_node::replay_call_summaries): New.
(exploded_node::replay_call_summary): New.
(per_function_data::~per_function_data): New.
(per_function_data::add_call_summary): Move here from header and
reimplement.
(exploded_graph::process_node): Call update_state rather than
update_model when handling bifurcation
(viz_callgraph_node::dump_dot): Use a regular label rather
than an HTML table; add summaries to dump.
* exploded-graph.h: Include "alloc-pool.h", "fibonacci_heap.h",
"supergraph.h", "sbitmap.h", "shortest-paths.h", "analyzer/sm.h",
"analyzer/program-state.h", and "analyzer/diagnostic-manager.h".
(exploded_node::replay_call_summaries): New decl.
(exploded_node::replay_call_summary): New decl.
(per_function_data::~per_function_data): New decl.
(per_function_data::add_call_summary): Move implemention from
header.
(per_function_data::m_summaries): Update type of element.
* known-function-manager.h: Include "analyzer/analyzer-logging.h".
* program-point.h: Include "pretty-print.h" and
"analyzer/call-string.h".
* program-state.cc: Include "analyzer/call-summary.h".
(sm_state_map::replay_call_summary): New.
(program_state::replay_call_summary): New.
* program-state.h (sm_state_map::replay_call_summary): New decl.
(program_state::replay_call_summary): New decl.
* region-model-manager.cc
(region_model_manager::get_or_create_asm_output_svalue): New
overload.
* region-model-manager.h
(region_model_manager::get_or_create_asm_output_svalue): New
overload decl.
* region-model.cc: Include "analyzer/call-summary.h".
(region_model::maybe_update_for_edge): Remove call to
region_model::update_for_call_summary on
SUPEREDGE_INTRAPROCEDURAL_CALL.
(region_model::update_for_call_summary): Delete.
(region_model::replay_call_summary): New.
* region-model.h (region_model::replay_call_summary): New decl.
(region_model::update_for_call_summary): Delete decl.
* store.cc: Include "analyzer/call-summary.h".
(store::replay_call_summary): New.
(store::replay_call_summary_cluster): New.
* store.h: Include "tristate.h".
(is_a_helper <const ana::concrete_binding *>::test): New.
(store::replay_call_summary): New decl.
(store::replay_call_summary_cluster): New decl.
* supergraph.cc (get_ultimate_function_for_cgraph_edge): Remove
"static" from decl.
(supergraph_call_edge): Make stmt param const.
* supergraph.h: Include "ordered-hash-map.h", "cfg.h",
"basic-block.h", "gimple.h", "gimple-iterator.h", and "digraph.h".
(supergraph_call_edge): Make stmt param const.
(get_ultimate_function_for_cgraph_edge): New decl.
* svalue.cc (compound_svalue::compound_svalue): Assert that we're
not nesting compound_svalues.
* svalue.h: Include "json.h", "analyzer/store.h", and
"analyzer/program-point.h".
(asm_output_svalue::get_num_outputs): New accessor.
gcc/testsuite/ChangeLog:
PR analyzer/107072
* gcc.dg/analyzer/call-summaries-2.c: New test.
* gcc.dg/analyzer/call-summaries-3.c: New test.
* gcc.dg/analyzer/call-summaries-asm-x86.c: New test.
* gcc.dg/analyzer/call-summaries-malloc.c: New test.
* gcc.dg/analyzer/call-summaries-pr107072.c: New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
gcc/analyzer/ChangeLog:
* region-model.h: Include "analyzer/region-model-manager.h"
(class region_model_manager): Move decl to...
* region-model-manager.h: ...this new file.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
gcc/analyzer/ChangeLog:
* region-model-manager.cc
(region_model_manager::maybe_fold_unaryop): Fold -(-(VAL)) to VAL.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
Enabling work towrads better call summarization.
gcc/analyzer/ChangeLog:
* region-model-manager.cc
(region_model_manager::get_or_create_widening_svalue): Use a
function_point rather than a program_point.
* region-model.cc (selftest::test_widening_constraints): Likewise.
* region-model.h
(region_model_manager::get_or_create_widening_svalue): Likewise.
(model_merger::get_function_point): New.
* svalue.cc (svalue::can_merge_p): Use a function_point rather
than a program_point.
(svalue::can_merge_p): Likewise.
* svalue.h (widening_svalue::key_t): Likewise.
(widening_svalue::widening_svalue): Likewise.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
|
|
In the testcase the elaboration of the array init that happens at genericize
time was getting the location info for the end of the function; fixed by
doing the expansion at the location of the original expression.
PR c++/107154
gcc/cp/ChangeLog:
* cp-gimplify.cc (cp_genericize_init_expr): Use iloc_sentinel.
(cp_genericize_target_expr): Likewise.
gcc/testsuite/ChangeLog:
* g++.dg/debug/dwarf2/lineno-array1.C: New test.
|
|
I've discovered a problem with the way we handle scoped attributes. For
declaration or type attributes for attributes we don't know anything about
we just don't add them to the declarations or types, so later in the FEs and
middle-end it is fine to use lookup_attribute etc. which just check the
attribute name and not namespace because non-standard non-GNU attributes
just won't show there. But in the case of attributes on statements, nothing
has filtered out the unknown attributes, so with my earlier assume
attribute patch e.g. c-c++-common/Wno-attributes-6.c test failed because
it uses:
[[vendor::assume(1 + 1 == 2)]];
with -Wno-attributes=vendor::assume and lookup_attribute ("assume", )
finds such attribute and handled it that way.
So, for those cases, this patch introduces lookup_attribute and
remove_attribute overloads which specify also the namespace.
I think the fallthrough, hot, cold, likely, unlikely attribute handling
will need to use the new APIs too, so that we don't handle
msft::fallthrough attribute as something we'd know.
2022-10-04 Jakub Jelinek <jakub@redhat.com>
* attribs.h (remove_attribute): Declare overload with additional
attr_ns argument.
(private_lookup_attribute): Declare overload with additional
attr_ns and attr_ns_len arguments.
(lookup_attribute): New overload with additional attr_ns argument.
* attribs.cc (remove_attribute): New overload with additional
attr_ns argument.
(private_lookup_attribute): New overload with additional
attr_ns and attr_ns_len arguments.
|
|
In these spots, the error/error_at has some inform afterwards which are
explanation part of the same diagnostics, so should be tied with
auto_diagnostic_group with it.
2022-10-04 Jakub Jelinek <jakub@redhat.com>
* attribs.cc (handle_ignored_attributes_option, decl_attributes,
common_function_versions): Use auto_diagnostic_group.
|
|
The assert removed by this patch was there to keep users from passing
masks of incompatible types. The self tests are passing host wide
ints down (set_nonzero_bits (-1)), which seem to be 32 bits, whereas
some embedded targets have integer_type_node's of 16-bits. This is
causing problems in m32c-elf, among others.
I suppose there's no harm in passing a 32-bit mask, because
set_nonzero_bits calls wide_int::from() to convert the mask to the
appropriate type. So we can remove the assert.
gcc/ChangeLog:
* value-range.cc (irange::set_nonzero_bits): Remove assert.
|
|
This is apparently needed since we include cp-trait.def from cp-tree.h
(in order to define the cp_trait_kind enum), as with operators.def.
PR c++/107136
gcc/cp/ChangeLog:
* Make-lang.in (CP_PLUGIN_HEADERS): Add cp-trait.def.
|
|
https://github.com/ARM-software/acle/pull/199 adds a new feature
macro for RCPC, for use in things like inline assembly. This patch
adds the associated support to GCC.
Also, RCPC is required for Armv8.3-A and later, but the armv8.3-a
entry didn't include it. This was probably harmless in practice
since GCC simply ignored the extension until now. (The GAS
definition is OK.)
gcc/
* config/aarch64/aarch64.h (AARCH64_ISA_RCPC): New macro.
* config/aarch64/aarch64-arches.def (armv8.3-a): Include RCPC.
* config/aarch64/aarch64-cores.def (thunderx3t110, zeus, neoverse-v1)
(neoverse-512tvb, saphira): Remove RCPC from these Armv8.3-A+ cores.
* config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Define
__ARM_FEATURE_RCPC when appropriate.
gcc/testsuite/
* gcc.target/aarch64/pragma_cpp_predefs_1.c: Add RCPC tests.
|
|
Split off from the 'Fortran: Add OpenMP's assume(s) directives' patch.
gcc/
* doc/invoke.texi (-fopenmp): Mention C++ attribut syntax.
(-fopenmp-simd): Likewise; update permitted directives.
gcc/fortran/
* parse.cc (decode_omp_directive): Handle '(end) loop' and 'scan'
also with -fopenmp-simd.
gcc/testsuite/
* gfortran.dg/gomp/openmp-simd-7.f90: New test.
|
|
gcc/
* doc/install.texi (Specific): Add missing items to bullet list.
(amdgcn): Update LLVM requirements, use version not date for newlib.
(nvptx): Use version not git hash for newlib.
|
|
The following patch adds support for the begin declare target construct,
which is another spelling for declare target construct without clauses
(where it needs paired end declare target), but unlike that one accepts
clauses.
This is an OpenMP 5.1 feature, implemented with 5.2 clarification because
in 5.1 we had a restriction in the declare target chapter shared by
declare target and begin declare target that if there are any clauses
specified at least one of them needs to be to or link. But that
was of course meant just for declare target and not begin declare target,
because begin declare target doesn't even allow to/link/enter clauses.
In addition to that, the patch also makes device_type clause duplication
an error (as stated in 5.1) and similarly makes declare target with
just device_type clause an error rather than warning.
What this patch doesn't do is:
1) OpenMP 5.1 also added an indirect clause, we don't support that
neither on declare target nor begin declare target
and I couldn't find it in our features pages (neither libgomp.texi
nor web)
2) I think device_type(nohost)/device_type(host) support can't work for
variables (in 5.0 it only talked about procedures so this could be
also thought as 5.1 feature that we should just add to the list
and implement)
3) I don't see any use of the "omp declare target nohost" attribute, so
I'm not sure if device_type(nohost) works at all
2022-10-04 Jakub Jelinek <jakub@redhat.com>
gcc/c-family/
* c-omp.cc (c_omp_directives): Uncomment begin declare target
entry.
gcc/c/
* c-lang.h (struct c_omp_declare_target_attr): New type.
(current_omp_declare_target_attribute): Change type from
int to vec<c_omp_declare_target_attr, va_gc> *.
* c-parser.cc (c_parser_translation_unit): Adjust for that change.
If last pushed directive was begin declare target, use different
wording and simplify format strings for easier translations.
(c_parser_omp_clause_device_type): Uncomment
check_no_duplicate_clause call.
(c_parser_omp_declare_target): Adjust for the
current_omp_declare_target_attribute type change, push { -1 }.
Use error_at rather than warning_at for declare target with
only device_type clauses.
(OMP_BEGIN_DECLARE_TARGET_CLAUSE_MASK): Define.
(c_parser_omp_begin): Add begin declare target support.
(c_parser_omp_end): Adjust for the
current_omp_declare_target_attribute type change, adjust
diagnostics wording and simplify format strings for easier
translations.
* c-decl.cc (current_omp_declare_target_attribute): Change type from
int to vec<c_omp_declare_target_attr, va_gc> *.
(c_decl_attributes): Adjust for the
current_omp_declare_target_attribute type change. If device_type
was present on begin declare target, add "omp declare target host"
and/or "omp declare target nohost" attributes.
gcc/cp/
* cp-tree.h (struct omp_declare_target_attr): Rename to ...
(cp_omp_declare_target_attr): ... this. Add device_type member.
(omp_begin_assumes_data): Rename to ...
(cp_omp_begin_assumes_data): ... this.
(struct saved_scope): Change types of omp_declare_target_attribute
and omp_begin_assumes.
* parser.cc (cp_parser_omp_clause_device_type): Uncomment
check_no_duplicate_clause call.
(cp_parser_omp_all_clauses): Fix up pasto, c_name for OMP_CLAUSE_LINK
should be "link" rather than "to".
(cp_parser_omp_declare_target): Adjust for omp_declare_target_attr
to cp_omp_declare_target_attr changes, push -1 as device_type. Use
error_at rather than warning_at for declare target with only
device_type clauses.
(OMP_BEGIN_DECLARE_TARGET_CLAUSE_MASK): Define.
(cp_parser_omp_begin): Add begin declare target support. Adjust
for omp_begin_assumes_data to cp_omp_begin_assumes_data change.
(cp_parser_omp_end): Adjust for the
omp_declare_target_attr to cp_omp_declare_target_attr and
omp_begin_assumes_data to cp_omp_begin_assumes_data type changes,
adjust diagnostics wording and simplify format strings for easier
translations.
* semantics.cc (finish_translation_unit): Likewise.
* decl2.cc (cplus_decl_attributes): If device_type was present on
begin declare target, add "omp declare target host" and/or
"omp declare target nohost" attributes.
gcc/testsuite/
* c-c++-common/gomp/declare-target-4.c: Move tests that are now
rejected into declare-target-7.c.
* c-c++-common/gomp/declare-target-6.c: Adjust expected diagnostics.
* c-c++-common/gomp/declare-target-7.c: New test.
* c-c++-common/gomp/begin-declare-target-1.c: New test.
* c-c++-common/gomp/begin-declare-target-2.c: New test.
* c-c++-common/gomp/begin-declare-target-3.c: New test.
* c-c++-common/gomp/begin-declare-target-4.c: New test.
* g++.dg/gomp/attrs-9.C: Add begin declare target tests.
* g++.dg/gomp/attrs-18.C: New test.
libgomp/
* libgomp.texi (Support begin/end declare target syntax in C/C++):
Mark as implemented.
|
|
The reason the nonzero mask was kept in a tree was basically inertia,
as everything in irange is a tree. However, there's no need to keep
it in a tree, as the conversions to and from wide ints are very
annoying. That, plus special casing NULL masks to be -1 is prone
to error.
I have not only rewritten all the uses to assume a wide int, but
have corrected a few places where we weren't propagating the masks, or
rather pessimizing them to -1. This will become more important in
upcoming patches where we make better use of the masks.
Performance testing shows a trivial improvement in VRP, as things like
irange::contains_p() are tied to a tree. Ughh, can't wait for trees in
iranges to go away.
gcc/ChangeLog:
* value-range-storage.cc (irange_storage_slot::set_irange): Remove
special case.
* value-range.cc (irange::irange_set): Adjust for nonzero mask
being a wide int.
(irange::irange_set_anti_range): Same.
(irange::set): Same.
(irange::verify_range): Same.
(irange::legacy_equal_p): Same.
(irange::operator==): Same.
(irange::contains_p): Same.
(irange::legacy_intersect): Same.
(irange::legacy_union): Same.
(irange::irange_single_pair_union): Call union_nonzero_bits.
(irange::irange_union): Same.
(irange::irange_intersect): Call intersect_nonzero_bits.
(irange::intersect): Adjust for nonzero mask being a wide int.
(irange::invert): Same.
(irange::set_nonzero_bits): Same.
(irange::get_nonzero_bits_from_range): New.
(irange::set_range_from_nonzero_bits): New.
(irange::get_nonzero_bits): Adjust for nonzero mask being a wide
int.
(irange::intersect_nonzero_bits): Same.
(irange::union_nonzero_bits): Same.
(range_tests_nonzero_bits): Remove test.
* value-range.h (irange::varying_compatible_p): Adjust for nonzero
mask being a wide int.
(gt_ggc_mx): Same.
(gt_pch_nx): Same.
(irange::set_undefined): Same.
(irange::set_varying): Same.
(irange::normalize_kind): Same.
|
|
__builtin_popcount and __builtin_ffs were sharing the same range-ops
entry, but the nonzero mask optimization is not valid for ffs.
Separate them out into two entries.
PR tree-optimization/107130
gcc/ChangeLog:
* gimple-range-op.cc (class cfn_popcount): Call op_cfn_ffs.
(class cfn_ffs): New.
(gimple_range_op_handler::maybe_builtin_call): Separate out
CASE_CFN_FFS into its own case.
gcc/testsuite/ChangeLog:
* gcc.dg/tree-ssa/pr107130.c: New test.
|
|
This PR related to _Pragma locations and diagnostic pragmas was fixed by a
combination of r10-325 and r13-1596. Add missing test coverage.
gcc/testsuite/ChangeLog:
PR c/91669
* c-c++-common/pr91669.c: New test.
|
|
|
|
i386-builtin-types.inc is included indirectly via i386-builtins.h
into 4 files: i386.cc i386-builtins.cc i386-expand.cc i386-features.cc
Only i386.cc dependency was present in gcc/config/t-i386 makefile.
As a result parallel builds occasionally fail as:
g++ ... -o i386-builtins.o ... ../../gcc-13-20220911/gcc/config/i386/i386-builtins.cc
In file included from ../../gcc-13-20220911/gcc/config/i386/i386-builtins.cc:92:
../../gcc-13-20220911/gcc/config/i386/i386-builtins.h:25:10:
fatal error: i386-builtin-types.inc: No such file or directory
25 | #include "i386-builtin-types.inc"
| ^~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [../../gcc-13-20220911/gcc/config/i386/t-i386:54: i386-builtins.o]
Error 1 shuffle=1663349189
gcc/
* config/i386/t-i386: Add build-time dependencies against
i386-builtin-types.inc to i386-builtins.o, i386-expand.o,
i386-features.o.
|
|
The cmse-15.c testcase fails at -Os because ICF means that we
generate
secure3:
b secure1
which is OK, but does not match the currently expected
secure3:
...
bx r[0-3]
gcc/testsuite/ChangeLog:
* gcc.target/arm/cmse/cmse-15.c: Align with -Os improvements.
Co-Authored-By: Yvan ROUX <yvan.roux@foss.st.com>
Signed-off-by: Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
|
|
On Fri, Sep 30, 2022 at 04:39:25PM -0400, Jason Merrill wrote:
> > --- gcc/cp/decl.cc.jj 2022-09-22 00:14:55.478599363 +0200
> > +++ gcc/cp/decl.cc 2022-09-22 00:24:01.121178256 +0200
> > @@ -223,6 +223,7 @@ struct GTY((for_user)) named_label_entry
> > bool in_transaction_scope;
> > bool in_constexpr_if;
> > bool in_consteval_if;
> > + bool in_assume;
>
> I think it would be better to reject jumps into statement-expressions like
> the C front-end.
Ok, here is a self-contained patch that does that.
2022-10-03 Jakub Jelinek <jakub@redhat.com>
* cp-tree.h (BCS_STMT_EXPR): New enumerator.
* name-lookup.h (enum scope_kind): Add sk_stmt_expr.
* name-lookup.cc (begin_scope): Handle sk_stmt_expr like sk_block.
* semantics.cc (begin_compound_stmt): For BCS_STMT_EXPR use
sk_stmt_expr.
* parser.cc (cp_parser_statement_expr): Use BCS_STMT_EXPR instead of
BCS_NORMAL.
* decl.cc (struct named_label_entry): Add in_stmt_expr.
(poplevel_named_label_1): Handle sk_stmt_expr.
(check_previous_goto_1): Diagnose entering of statement expression.
(check_goto): Likewise.
* g++.dg/ext/stmtexpr24.C: New test.
|
|
* sv.po: Update.
|
|
... to match the trait's canonical spelling __is_same instead
of its alternative spelling __is_same_as.
gcc/c-family/ChangeLog:
* c-common.cc (c_common_reswords): Use RID_IS_SAME instead of
RID_IS_SAME_AS.
gcc/cp/ChangeLog:
* constraint.cc (diagnose_trait_expr): Use CPTK_IS_SAME instead
of CPTK_IS_SAME_AS.
* cp-trait.def (IS_SAME_AS): Rename to ...
(IS_SAME): ... this.
* pt.cc (alias_ctad_tweaks): Use CPTK_IS_SAME instead of
CPTK_IS_SAME_AS.
* semantics.cc (trait_expr_value): Likewise.
(finish_trait_expr): Likewise.
|
|
Add a vector length parameter needed by amdgcn without breaking aarch64.
All amdgcn vector masks are DImode, regardless of vector length, so we can't
tell what length is implied simply from the operator mode. (Even if we used
different integer modes there's no mode small enough to differenciate a 2 or
4 lane mask). Without knowing the intended length we end up using a mask with
too many lanes enabled, which leads to undefined behaviour..
The extra operand is not added for vector mask types so AArch64 does not need
to be adjusted.
gcc/ChangeLog:
* config/gcn/gcn-valu.md (while_ultsidi): Limit mask length using
operand 3.
* doc/md.texi (while_ult): Document new operand 3 usage.
* internal-fn.cc (expand_while_optab_fn): Set operand 3 when lhs_type
maps to a non-vector mode.
|
