Sebastian Pop [Thu, 21 Jun 2007 17:06:05 +0000 (17:06 +0000)]
re PR tree-optimization/19590 (IVs with the same evolution not eliminated)
PR tree-optimization/19590
* tree-vrp.c (adjust_range_with_scev): Set the range when the result
of scev is a constant.
* gcc/testsuite/gcc.dg/tree-ssa/pr19590.c: New.
Jakub Jelinek [Thu, 21 Jun 2007 12:11:00 +0000 (14:11 +0200)]
re PR middle-end/32362 (ICE: in lookup_decl_in_outer_ctx, at omp-low.c:1508)
PR middle-end/32362
* omp-low.c (lookup_decl_in_outer_ctx): Don't ICE if t is NULL,
but decl is a global var, instead return decl.
* gimplify.c (gimplify_adjust_omp_clauses_1): Add shared clauses
even for is_global_var decls, if they are private in some outer
context.
* testsuite/libgomp.c/pr32362-1.c: New test.
* testsuite/libgomp.c/pr32362-2.c: New test.
* testsuite/libgomp.c/pr32362-3.c: New test.
* gnu/classpath/jdwp/value/StringValue.java
(StringValue): Tag of StringValue is STRING not OBJECT.
(write): String values are written to the wire as tag byte
and object ID, not JdwpString.
Keith Seitz [Wed, 20 Jun 2007 20:30:34 +0000 (20:30 +0000)]
Update from classpath trunk:
* gnu/classpath/jdwp/processor/ClassTypeCommandSet.java
(executeInvokeMethod): No need to use ValueFactory any more;
MethodResult.getReturnedValue now returns a Value.
(executeNewInstance): Double-check that return result is
an ObjectValue; throw JdwpInternalErrorException if it is not.
(invokeMethod): Method IDs come from VMMethod, not VMIdManager.
Arguments are Values not Objects.
Use ValueFactory to create arguments.
Pass invocation options to VMVirtualMachine.executeMethod.
Don't do any thread suspend/resume work: VMVM.executeMethod
will take care of it.
* gnu/classpath/jdwp/processor/ObjectReferenceCommandSet.java
(executeInvokeMethod): Method IDs come from VMMethod, not
VMIdManager.
Arguments should be Values instead of Objects.
Use ValueFactory to create Values.
Remove specific option handling and pass options to
VMVirtualMachine.executeMethod.
Remove thread suspension.
Use MethodResult.getReturnedValue to get method's result.
* gnu/classpath/jdwp/util/MethodResult.java
(returnedValue): Change type to Value.
(thrownException): Change type to Throwable.
(resType): Remove.
(MethodResult): New constructor.
(setReturnedValue): Remove.
(SetThrownException): Remove.
(getResultType): Remove.
(setResultType): Remove.
* gnu/classpath/jdwp/value/ObjectValue.java (getValue):
New method.
* vm/reference/gnu/classpath/jdwp/VMVirtualMachine.java
(executeMethod): Replace "nonVirtual" parameter with more
generic "options" parameter.
Replace java.lang.reflect.Method parameter with VMMethod.
Replace Object[] parameter with Value[] parameter.
PR target/32335
* config/m32c/m32c.c (m32c_emit_epilogue): Use new HImode epilogue
for TARGET_A16.
* config/m32c/prologue.md (epilogue_exitd_16): New.
(epilogue_reit_16): New.
(epilogue_exitd): Rename to epilogue_exitd_24.
(epilogue_reit): Rename to epilogue_reit_24.
PR target/32335
* config/m32c/m32c.c (m32c_emit_epilogue): Use new HImode epilogue
for TARGET_A16.
* config/m32c/prologue.md (epilogue_exitd_16): New.
(epilogue_reit_16): New.
(epilogue_exitd): Rename to epilogue_exitd_24.
(epilogue_reit): Rename to epilogue_reit_24.
Seongbae Park [Wed, 20 Jun 2007 18:22:28 +0000 (18:22 +0000)]
dbgcnt.def (global_alloc_at_func, [...]): New counters.
2007-06-20 Seongbae Park <seongbae.park@gmail.com>
Maxim Kuvyrkov <mkuvyrkov@ispras.ru>
* dbgcnt.def (global_alloc_at_func, global_alloc_at_reg):
New counters.
* haifa-sched.c (queue_to_ready): Don't requeue next insn
if dbg_cnt (sched_insn) reaches the limit.
(choose_ready): New parameter INSN_PTR and new return value.
(schedule_block): Handle dbg_cnt (sched_insn). Handle
the new return value from choose_ready.
* global.c (global_aloc): New dbgcnt global_alloc_at_reg.
(rest_of_handle_global_alloc): New global_alloc_at_func.
Adam Nemet [Wed, 20 Jun 2007 18:19:10 +0000 (18:19 +0000)]
re PR tree-optimization/25737 (ACATS c974001c974013 hang with struct aliasing)
PR tree-optimization/25737
* tree.h (struct tree_struct_field_tag): Add new field alias_set.
(SFT_NONADDRESSABLE_P, SFT_ALIAS_SET): New macros.
* tree-flow.h (struct fieldoff): Add new field alias_set.
* tree-ssa-structalias.c (push_fields_onto_fieldstack): Add new
argument addressable_type. Set alias_set of fieldoff.
* tree-ssa-alias.c (create_sft): Add new argument alias_set.
(create_overlap_variables_for): Pass alias_set from fieldoff to
create_sft.
* alias.c (get_alias_set): Use alias_set from SFT if set.
Jakub Jelinek [Wed, 20 Jun 2007 06:39:53 +0000 (08:39 +0200)]
re PR middle-end/31959 (ICE in expand_builtin_expect, at builtins.c:5112)
PR middle-end/31959
* builtins.c: Include diagnostic.h.
(expand_builtin_expect): Make gcc_assert more permissive.
* Makefile.in (builtins.o): Depend on $(DIAGNOSTIC_H).
* df.h (DF_FIRST_OPTIONAL_PROBLEM): Removed.
(struct df_problem.free_blocks_on_set_blocks): New field.
(struct dataflow.optional_p): New field.
(df_bb_regno_last_use_find, df_insn_regno_def_p): Removed.
(df_live_set_all_dirty): New function.
* df-scan.c (df_scan_alloc): Initialize optional_p.
(problem_SCAN): Initialize free_blocks_on_set_blocks.
* df-core.c (df_set_blocks): Removed use of
DF_FIRST_OPTIONAL_PROBLEM. Now uses
df_problem.free_blocks_on_set_blocks to determine which blocks are
recycled.
(df_remove_problem): Removed use of DF_FIRST_OPTIONAL_PROBLEM.
(df_finish_pass): Removed use of DF_FIRST_OPTIONAL_PROBLEM. Now
uses dataflow.optional_p to determine if problem should be
deleted.
(rest_of_handle_df_initialize): Only start live problem if
-02 or above.
(df_bb_regno_last_use_find, df_insn_regno_def_p): Removed.
* df-problems.c (df_ru_alloc, df_rd_alloc, df_lr_alloc,
df_live_alloc, df_urec_alloc, df_note_alloc): set optional_p.
(problem_RU, problem_RD, problem_LR, problem_UREC, problem_CHAIN,
problem_NOTE): Initialize free_blocks_on_set_blocks.
(df_lr_bb_local_compute): Recompute luids if df_live problem is
not active.
(df_live_set_all_dirty, df_note_alloc): New function.
* regrename.c (merge_overlapping_regs): Change DF_LIVE_* to
df_get_live_*.
* sched_ebb.c (compute_jump_reg_dependencies): Ditto.
* postreload.c (reload_combine): Ditto.
* cse.c (cse_extended_basic_block): Ditto.
* regmove.c (mark_flags_life_zones): Ditto.
* rtlfactoring.c (split_blocks_after_seqs, split_pattern_seq,
erase_matching_seqs): Ditto.
* bt-load.c (compute_defs_uses_and_gen): Ditto.
* integrate (allocate_initial_values): Ditto.
* combine.c (reg_dead_at_p): Ditto.
* resource.c (mark_target_live_regs): Ditto.
* sched-rgn.c (check_live_1, update_live_1): Ditto.
* config/sh/sh.c (find_r0_life_regions): Ditto.
* global.c (rest_of_handle_global_alloc): Only add back df_live
for -O > 1.
* local-alloc.c (rest_of_handle_local_alloc): Only remove
df_live for -O > 1.
* ifcvt.c (dead_or_predicable): Change DF_LIVE_* to
df_get_live_*.
(if_convert): Make sure df_live is there at -O == 1.
(pass_if_after_combine): Cleanup flags.
* init-regs.c (initialize_uninitialized_regs): Make sure df_live
is there at -O == 1.
Sebastian Pop [Tue, 19 Jun 2007 18:35:39 +0000 (18:35 +0000)]
re PR tree-optimization/32367 (internal compiler error: in build_polynomial_chrec, at tree-chrec.h:113)
PR tree-optimization/32367
* tree-chrec.h (build_polynomial_chrec): Verify that the left hand side
of the chrec has no evolution in that loop.
* testsuite/gcc.dg/tree-ssa/pr32367.c: New.
Bob Wilson [Tue, 19 Jun 2007 17:58:13 +0000 (17:58 +0000)]
xtensa.c: Include "df.h".
* config/xtensa/xtensa.c: Include "df.h".
(xtensa_builtin_saveregs): Use adjust_address instead of change_address.
(xtensa_va_start): Invoke make_tree with sizetype for
expand_builtin_saveregs and then convert the result to a pointer.
Use POINTER_PLUS_EXPR. Use size_int instead of build_int_cst.
(xtensa_gimplify_va_arg_expr): Use size_int instead of build_int_cst.
Subtract argument size from index value as integers and then use
POINTER_PLUS_EXPR to add the result to the array address.
PR target/32335
* config/m32c/m32c.c: Include dataflow header file.
(m32c_emit_prologue): Adjust for prologue insn change.
* config/m32c/prologue.md (prologue_enter_16): Only modify SP_REGNO
once inside a PARALLEL. Assume frame size passed in operand 0
includes space to save the fb register.
(prologue_enter_24): Likewise.
(epilogue_exitd): Only modify SP_REGNO once inside a PARALLEL.
Keith Seitz [Tue, 19 Jun 2007 00:10:10 +0000 (00:10 +0000)]
Breakpoint.java: Make abstract.
* gnu/gcj/jvmti/Breakpoint.java: Make abstract.
(method): Change from private to protected.
(location): Likewise.
(Breakpoint): Change argument list to take only integer type.
Add default constructor.
(initialize_native): Renamed to ...
(_save_insn): ... this to make function more explicit.
(execute): New method.
* gnu/gcj/jvmti/Breakpoint.h: Regenerate.
* gnu/gcj/jvmti/natBreakpoint.cc (initialize_native): Rename to...
(_save_insn): ... this.
(install): Save the original instruction.
* gnu/gcj/jvmti/NormalBreakpoint.java: New file.
* gnu/gcj/jvmti/NormalBreakpoint.h: New file.
* gnu/gcj/jvmti/natNormalBreakpoint.cc: New file.
* gnu/gcj/jvmti/BreakpointManager.java (newBreakpoint):
Instantiate a NormalBreakpoint instead of Breakpoint.
* interpret-run.cc (insn_breakpoint): Remove breakpoint actions
and call Breakpoint.execute to do them.
* classpath/lib/gnu/gcj/jvmti/Breakpoint.class: Regenerate.
* classpath/lib/gnu/gcj/jvmti/BreakpointManager.class: Likewise.
* classpath/lib/gnu/gcj/jvmti/NormalBreakpoint.class: New file.
* sources.am: Regenerate.
* Makefile.am (nat_source_files): Add natNormalBreakpoint.cc.
* Makefile.in: Regenerated.
Paul Thomas [Mon, 18 Jun 2007 23:04:28 +0000 (23:04 +0000)]
re PR fortran/20863 ([4.2 only] Pointer problems in PURE procedures)
2007-06-19 Paul Thomas <pault@gcc.gnu.org>
PR fortran/20863
PR fortran/20082
* resolve.c (resolve_code): Use gfc_impure_variable as a
condition for rejecting derived types with pointers, in pure
procedures.
(gfc_impure_variable): Add test for dummy arguments of pure
procedures; any for functions and INTENT_IN for subroutines.
PR fortran/32236
* data.c (gfc_assign_data_value): Change the ICE on an array
reference initializer not being an array into an error and
clear init to prevent a repetition of the error.
2007-06-19 Paul Thomas <pault@gcc.gnu.org>
PR fortran/20863
PR fortran/20082
* gfortran.dg/impure_assignment_2.f90 : New test.
PR fortran/32236
* gfortran.dg/data_initialized_2.f90 : New test.
* gfortran.dg/equiv_7.f90 : Test for endianess and call the
appropriate version of 'dmach'.
Simon Baldwin [Mon, 18 Jun 2007 22:09:14 +0000 (22:09 +0000)]
re PR c++/31923 (g++ accepts a storage-class-specifier on a template explicit specialization)
gcc/cp/ChangeLog
2007-06-15 Simon Baldwin <simonb@google.com>
PR c++/31923
* parser.c (cp_parser_single_declaration): Added check for storage
class other than sc_none in parsed declaration, and a flag to indicate
if the call is part of an explicit template specialization parse.
* (cp_parser_explicit_specialization): Specialization check flag added
to call to cp_parser_single_declaration(), set true.
* (cp_parser_template_declaration_after_export): Specialization check
flag added to call to cp_parser_single_declaration(), set false.
* pt.c (check_explicit_specialization): Added code to copy visiblity
and linkage from the templated function to the explicit specialization.
gcc/testsuite/ChangeLog
2007-06-15 Simon Baldwin <simonb@google.com>
Uros Bizjak [Mon, 18 Jun 2007 08:30:47 +0000 (10:30 +0200)]
re PR tree-optimization/32383 (ICE with reciprocals and -ffast-math)
PR tree-optimization/32383
* targhooks.c (default_builtin_reciprocal): Add new bool argument.
* targhooks.h (default_builtin_reciprocal): Update prototype.
* target.h (struct gcc_target): Update builtin_reciprocal.
* doc/tm.texi (TARGET_BUILTIN_RECIPROCAL): Update description.
* tree-ssa-math-opts (execute_cse_reciprocals): Skip statements
where arg1 is not SSA_NAME. Pass true to targetm.builtin_reciprocal
when fndecl is in BUILT_IN_MD class.
(execute_convert_to_rsqrt): Ditto.
* config/i386/i386.c (ix86_builtin_reciprocal): Update for new bool
argument. Convert IX86_BUILTIN_SQRTPS code only when md_fn is true.
Convert BUILT_IN_SQRTF code only when md_fn is false.
testsuite/ChangeLog:
PR tree-optimization/32383
* testsuite/g++.dg/opt/pr32383.C: New test.
PR middle-end/32349
* modulo-sched (generate_reg_moves): Added rescan parameter and if
this is true, rescan insn being modified.
(sms_schedule): Added rescan parameter.
(rest_of_handle_sms): Moved freeing of dominance info to before
getting out of cfg_layout.
* config/m68k/predicates.md (const_call_operand): Adjust comment.
(const_sibcall_operand): New.
(sibcall_operand): Use it.
* config/m68k/m68k.c (FL_FOR_isa_c): Not ISA_B compatible.
(m68k_isas): ISAC does not imply FPU or EMAC.
(override_options): Add ISA_C logic for symbolic jump & call.
Uros Bizjak [Sat, 16 Jun 2007 09:52:48 +0000 (11:52 +0200)]
re PR middle-end/31723 (Use reciprocal and reciprocal square root with -ffast-math)
PR middle-end/31723
* hooks.c (hook_tree_tree_bool_null): New hook.
* hooks.h (hook_tree_tree_bool_null): Add prototype.
* tree-pass.h (pass_convert_to_rsqrt): Declare.
* passes.c (init_optimization_passes): Add pass_convert_to_rsqrt.
* tree-ssa-math-opts.c (execute_cse_reciprocals): Scan for a/func(b)
and convert it to reciprocal a*rfunc(b).
(execute_convert_to_rsqrt): New function.
(gate_convert_to_rsqrt): New function.
(pass_convert_to_rsqrt): New pass definition.
* target.h (struct gcc_target): Add builtin_reciprocal.
* target-def.h (TARGET_BUILTIN_RECIPROCAL): New define.
(TARGET_INITIALIZER): Initialize builtin_reciprocal with
TARGET_BUILTIN_RECIPROCAL.
* doc/tm.texi (TARGET_BUILTIN_RECIPROCAL): Document.
* config/i386/i386.h (TARGET_RECIP): New define.
* config/i386/i386.md (divsf3): Expand by calling ix86_emit_swdivsf
for TARGET_SSE_MATH and TARGET_RECIP when
flag_unsafe_math_optimizations is set and not optimizing for size.
(*rcpsf2_sse): New insn pattern.
(*rsqrtsf2_sse): Ditto.
(rsqrtsf2): New expander. Expand by calling ix86_emit_swsqrtsf
for TARGET_SSE_MATH and TARGET_RECIP when
flag_unsafe_math_optimizations is set and not optimizing for size.
(sqrt<mode>2): Expand SFmode operands by calling ix86_emit_swsqrtsf
for TARGET_SSE_MATH and TARGET_RECIP when
flag_unsafe_math_optimizations is set and not optimizing for size.
* config/i386/sse.md (divv4sf): Expand by calling ix86_emit_swdivsf
for TARGET_SSE_MATH and TARGET_RECIP when
flag_unsafe_math_optimizations is set and not optimizing for size.
(*sse_rsqrtv4sf2): Do not export.
(sqrtv4sf2): Ditto.
(sse_rsqrtv4sf2): New expander. Expand by calling ix86_emit_swsqrtsf
for TARGET_SSE_MATH and TARGET_RECIP when
flag_unsafe_math_optimizations is set and not optimizing for size.
(sqrtv4sf2): Ditto.
* config/i386/i386.opt (mrecip): New option.
* config/i386/i386-protos.h (ix86_emit_swdivsf): Declare.
(ix86_emit_swsqrtsf): Ditto.
* config/i386/i386.c (IX86_BUILTIN_RSQRTF): New constant.
(ix86_init_mmx_sse_builtins): __builtin_ia32_rsqrtf: New
builtin definition.
(ix86_expand_builtin): Expand IX86_BUILTIN_RSQRTF using
ix86_expand_unop1_builtin.
(ix86_emit_swdivsf): New function.
(ix86_emit_swsqrtsf): Ditto.
(ix86_builtin_reciprocal): New function.
(TARGET_BUILTIN_RECIPROCAL): Use it.
(ix86_vectorize_builtin_conversion): Rename from
ix86_builtin_conversion.
(TARGET_VECTORIZE_BUILTIN_CONVERSION): Use renamed function.
* doc/invoke.texi (Machine Dependent Options): Add -mrecip to
"i386 and x86_64 Options" section.
(Intel 386 and AMD x86_64 Options): Document -mrecip.
Andrew Pinski [Sat, 16 Jun 2007 05:42:36 +0000 (22:42 -0700)]
[multiple changes]
2007-06-15 Andrew Pinski <andrew_pinski@playstation.sony.com>
Zdenek Dvorak <dvorakz@suse.cz>
Richard Guenther <rguenther@suse.de>
Kaz Kojima <kkojima@gcc.gnu.org>
* tree-vrp.c (compare_values_warnv): Convert val2 to
the type of val1.
(extract_range_from_assert): Create
POINTER_PLUS_EXPR for pointer types.
(extract_range_from_binary_expr): Handle
only POINTER_PLUS_EXPR, MIN_EXPR, and MAX_EXPR
for pointer types.
* doc/c-tree.texi (POINTER_PLUS_EXPR): Document.
* tree-ssa-loop-niter.c (split_to_var_and_offset): Handle
POINTER_PLUS_EXPR as PLUS_EXPR.
(number_of_iterations_lt_to_ne):
For pointer types, use sizetype when
creating MINUS_EXPR/PLUS_EXPRs.
(assert_loop_rolls_lt): For pointer types, use sizetype when
creating MINUS_EXPR/PLUS_EXPRs.
(number_of_iterations_le): Likewise.
(expand_simple_operations): POINTER_PLUS_EXPR are simple also.
(derive_constant_upper_bound): Handle POINTER_PLUS_EXPR just
like PLUS_EXPR and MINUS_EXPR.
* tree-pretty-print.c (dump_generic_node): Handle
POINTER_PLUS_EXPR.
(op_prio): Likewise.
(op_symbol_1): Likewise.
* optabs.c (optab_for_tree_code): Likewise.
* tree-ssa-loop-manip.c (create_iv): Handle pointer base
specially.
* tree-tailcall.c (process_assignment): Mention
POINTER_PLUS_EXPR in a TODO comment.
* tree.c (build2_stat): Assert when trying to use PLUS_EXPR or
MINUS_EXPR with a pointer. Also assert for POINTER_PLUS_EXPR
not used with a pointer and an integer type.
* tree-scalar-evolution.c (add_to_evolution_1): Convert the
increment using chrec_convert_rhs instead of chrec_convert.
(follow_ssa_edge_in_rhs): Handle POINTER_PLUS_EXPR like
PLUS_EXPR except for the right hand side's type will be
sizetype.
(interpret_rhs_modify_stmt): Handle POINTER_PLUS_EXPR.
(fold_used_pointer_cast): Kill.
(pointer_offset_p): Kill.
(fold_used_pointer): Kill.
(pointer_used_p): Kill.
(analyze_scalar_evolution_1 <case GIMPLE_MODIFY_STMT>): Don't
call fold_used_pointer.
(instantiate_parameters_1): Convert the increment
using chrec_convert_rhs instead of chrec_convert.
Handle POINTER_PLUS_EXPR as PLUS_EXPR.
* builtins.c (get_pointer_alignment): Handle POINTER_PLUS_EXPR
instead of PLUS_EXPR.
(expand_builtin_strcat): Create a POINTER_PLUS_EXPR instead of
PLUS_EXPR for pointers.
(std_gimplify_va_arg_expr): Likewise.
(fold_builtin_memory_op): Likewise.
(fold_builtin_strstr): Likewise.
(fold_builtin_strchr): Likewise.
(fold_builtin_strrchr): Likewise.
(fold_builtin_strpbrk): Likewise.
(expand_builtin_memory_chk): Likewise.
(fold_builtin_memory_chk): Likewise.
(std_expand_builtin_va_start): Use
sizetype for the call to make_tree and then convert
to the pointer type.
(fold_builtin_memchr): Use POINTER_PLUS_EXPR
instead of PLUS_EXPR for adding to a pointer.
(std_gimplify_va_arg_expr): Use fold_build2 for
the creating of POINTER_PLUS_EXPR. For the BIT_AND_EXPR, cast
the operands to sizetype first and then cast the BIT_AND_EXPR
back to the pointer type.
* fold-const.c (build_range_check): Handle pointer types
specially.
(extract_array_ref): Look for POINTER_PLUS_EXPR instead
of PLUS_EXPR's. Make sure the offset is converted to
sizetype.
(try_move_mult_to_index): Strip the NOPs from the offset.
Remove code argument and replace all uses with PLUS_EXPR.
(fold_to_nonsharp_ineq_using_bound): Handle pointer types
specially. Don't use a pointer type for MINUS_EXPR.
(fold_unary): Handle for (T1)(X op Y),
only p+ as that is the only as that can be handled for
binary operators now.
(fold_binary <case POINTER_PLUS_EXPR>): Add folding of
POINTER_PLUS_EXPR.
<case PLUS_EXPR>: Add folding of PTR+INT into
PTR p+ INT.
Don't call try_move_mult_to_index.
<case MINUS_EXPR>: Fold (PTR0 p+ A) - (PTR1 p+ B)
into (PTR0 - PTR1) + (A - B). Fold (PTR0 p+ A) - PTR1 into
(PTR0 - PTR1) + A iff (PTR0 - PTR1) simplifies.
Don't call try_move_mult_to_index.
(tree_expr_nonnegative_warnv_p): Handle POINTER_PLUS_EXPR.
(tree_expr_nonzero_p): Likewise.
(fold_indirect_ref_1): Look at POINTER_PLUS_EXPR instead
of PLUS_EXPR for the complex expression folding.
* tree-chrec.c (chrec_fold_plus_poly_poly): If the
first chrec is a pointer type, then the second should
be sizetype and not the first's type.
For POINTER_PLUS_EXPR, use a different right hand side type.
Handle POINTER_PLUS_EXPR like PLUS_EXPR.
(chrec_fold_plus_1): For POINTER_PLUS_EXPR, use a
different right hand side type.
Handle POINTER_PLUS_EXPR like PLUS_EXPR.
(chrec_fold_plus): For pointer types, use POINTER_PLUS_EXPR
instead of PLUS_EXPR.
When either operand is zero, convert the other operand.
(chrec_apply): Use chrec_convert_rhs
on the argument x instead of chrec_convert.
(reset_evolution_in_loop): For pointer types, the new_evol
should be sizetype.
(convert_affine_scev): For POINTER_PLUS_EXPR, use a
different right hand side type.
Handle POINTER_PLUS_EXPR like PLUS_EXPR.
(chrec_convert_rhs): New function.
(chrec_convert_aggressive): For POINTER_PLUS_EXPR, use a
different right hand side type.
Handle POINTER_PLUS_EXPR like PLUS_EXPR.
* tree-chrec.h (chrec_convert_rhs): New prototype.
(build_polynomial_chrec): For pointer types, the right hand
* tree-ssa-ccp.c (maybe_fold_stmt_indirect): Look for
POINTER_PLUS_EXPR instead of PLUS_EXPR's.
Remove subtraction case as it is always addition now.
Make sure the offset is converted to sizetype.
(fold_stmt_r): Don't handle PLUS_EXPR/MINUS_EXPR specially.
Handle POINTER_PLUS_EXPR like PLUS_EXPR was handled before.
* tree-ssa-loop-ivopts.c (determine_base_object): Abort for
PLUS_EXPR in pointer type.
Handle POINTER_PLUS_EXPR.
(tree_to_aff_combination): Likewise.
(force_expr_to_var_cost): Likewise.
(force_expr_to_var_cost): Likewise. Create a POINTER_PLUS_EXPR
instead of PLUS_EXPR for pointers.
* c-format.c (check_format_arg): Handle POINTER_PLUS_EXPR
instead of PLUS_EXPR of pointer types.
* tree-stdarg.c (va_list_counter_bump): Handle POINTER_PLUS_EXPR
as PLUS_EXPR.
(check_va_list_escapes): Likewise.
(check_all_va_list_escapes): Likewise.
* dwarf2out.c (loc_descriptor_from_tree_1):
Handle POINT_PLUS_EXPR as a PLUS_EXPR.
* expr.c (expand_expr_real_1): Handle POINTER_PLUS_EXPR.
(string_constant): Likewise.
* tree-ssa-address.c (tree_mem_ref_addr): When adding
the offset to the base, use POINTER_PLUS_EXPR.
(add_to_parts): Convert the index to sizetype.
(create_mem_ref): Create A POINTER_PLUS_EXPR for the one case.
* matrix-reorg.c (collect_data_for_malloc_call): Stmt
will now only be either INDIRECT_REF and POINTER_PLUS_EXPR.
Offset only holds something for PLUS_EXPR.
(ssa_accessed_in_tree): Handle POINTER_PLUS_EXPR just as
a PLUS_EXPR.
(analyze_transpose): POINTER_PLUS_EXPR will only show up now
and not PLUS_EXPR.
(analyze_accesses_for_modify_stmt): Likewise.
Remove comment about the type being integral type as it is
wrong now.
(can_calculate_expr_before_stmt): Handle POINTER_PLUS_EXPR as
PLUS_EXPR.
(transform_access_sites): POINTER_PLUS_EXPR will only show up now
and not PLUS_EXPR.
Correct the type which the artimentic is done in (is now
sizetype).
Reindent one loop.
* tree-data-ref.c (split_constant_offset): Handle
POINTER_PLUS_EXPR
* tree-affine.c (tree_to_aff_combination): Likewise.
* c-typeck.c (build_unary_op): For pointers create the increment
as a sizetype. Create a POINTER_PLUS_EXPR instead of PLUS_EXPR
for pointers.
* gimplify.c (gimplify_self_mod_expr): Create a
POINTER_PLUS_EXPR instead of PLUS_EXPR for pointers.
(gimplify_omp_atomic_fetch_op): Handle POINTER_PLUS_EXPR.
* tree.def (POINTER_PLUS_EXPR): New tree code.
* tree-predcom.c (ref_at_iteration): If we have a pointer
type do the multiplication in sizetype.
* tree-mudflap.c (mf_xform_derefs_1): Create a
POINTER_PLUS_EXPR instead of PLUS_EXPR for pointers.
* tree-ssa-forwprop.c
(forward_propagate_addr_into_variable_array_index):
Don't expect there to be a cast for the index as that
does not exist anymore.
(forward_propagate_addr_expr_1): Check for POINTER_PLUS_EXPR
instead of PLUS_EXPR.
Don't check for the first operand of the POINTER_PLUS_EXPR
was the index as it cannot be.
Call forward_propagate_addr_into_variable_array_index with
the SSA_NAME instead of the statement.
* varasm.c (const_hash_1): Handle POINTER_PLUS_EXPR.
(compare_constant): Likewise.
(copy_constant): Likewise.
(compute_reloc_for_constant): Likewise.
(output_addressed_constants): Likewise.
(initializer_constant_valid_p): Likewise.
* tree-ssa.c (tree_ssa_useless_type_conversion_1):
Convert the MIN/MAX of the inner type to the outer
type before comparing them.
* tree-ssa-loop-prefetch.c (idx_analyze_ref): Handle
POINTER_PLUS_EXPR instead of PLUS_EXPR.
(issue_prefetch_ref): Create a POINTER_PLUS_EXPR instead
of PLUS_EXPR for pointers.
* tree-inline.c (estimate_num_insns_1): Handle
POINTER_PLUS_EXPR.
* tree-vect-transform.c (vect_create_addr_base_for_vector_ref):
Create a POINTER_PLUS_EXPR instead of PLUS_EXPR for pointers.
(bump_vector_ptr): Create a POINTER_PLUS_EXPR
instead of PLUS_EXPR for the pointer increment statement.
(vect_update_ivs_after_vectorizer): For pointer types, create
POINTER_PLUS_EXPR instead of PLUS_EXPR and also create
MULT_EXPR in sizetype.
(vect_gen_niters_for_prolog_loop): Add a cast when creating
byte_misalign.
* tree-object-size.c (plus_expr_object_size): Handle
POINTER_PLUS_EXPR instead of PLUS_EXPR. Removing all the extra
code which is trying to figure out which side is a pointer and
is the index.
(check_for_plus_in_loops_1): Likewise.
(check_for_plus_in_loops): Likewise.
* c-common.c (pointer_int_sum): Create a
POINTER_PLUS_EXPR instead of PLUS_EXPR for pointers.
* tree-ssa-structalias.c (handle_ptr_arith): Handle
only POINTER_PLUS_EXPR. Removing all the extra
code which is trying to figure out which side is a pointer and
is the index.
* tree-cfg.c (verify_expr): Add extra checking for pointers and
PLUS_EXPR and MINUS_EXPR.
Also add checking to make sure the operands of POINTER_PLUS_EXPR
are correct.
* config/frv/frv.c (frv_expand_builtin_va_start): Use sizetype
with make_tree, instead of a pointer type.
* config/s390/s390.c (s390_va_start): Use POINTER_PLUS_EXPR
for pointers instead of PLUS_EXPR.
(s390_gimplify_va_arg): Likewise.
* config/spu/spu.c (spu_va_start): Create POINTER_PLUS_EXPR
instead of PLUS_EXPR when doing addition on pointer
types. Use sizetype for the second operand.
(spu_gimplify_va_arg_expr): Likewise.
* config/sparc/sparc.c (sparc_gimplify_va_arg): Use
POINTER_PLUS_EXPR instead of PLUS_EXPR when the operand was
a pointer. Don't create a BIT_AND_EXPR for pointer types.
* config/i386/i386.c (ix86_va_start): Use POINTER_PLUS_EXPR
for the pointer addition and also use size_int/sizetype
for the offset.
(ix86_gimplify_va_arg): Likewise.
Perform BIT_AND_EXPR on sizetype arguments.
* config/sh/sh.c (sh_va_start): Call make_tree with sizetype
and convert its result to a pointer type. Use POINTER_PLUS_EXPR
for the pointer additions and also use size_int for the offsets.
(sh_gimplify_va_arg_expr): Use POINTER_PLUS_EXPR for the pointer
additions and also use size_int for the offsets. Perform
BIT_AND_EXPR on sizetype arguments.
* config/ia64/ia64.c (ia64_gimplify_va_arg): Use
POINTER_PLUS_EXPR for pointers and create the
BIT_AND_EXPR in sizetype.
* config/rs6000/rs6000.c (rs6000_va_start): Use POINTER_PLUS_EXPR
instead of PLUS_EXPR for pointer addition.
(rs6000_va_start): Likewise.
Also use sizetype for the offset.
* config/pa/pa.c (reloc_needed): Handle POINTER_PLUS_EXPR
as PLUS_EXPR/MINUS_EXPR.
(hppa_gimplify_va_arg_expr): Don't create MINUS_EXPR or
PLUS_EXPR for pointers, instead use POINTER_PLUS_EXPR.
Don't use BIT_AND_EXPR on a pointer type, convert the
expression to sizetype first.
* config/mips/mips.c (mips_va_start): Use POINTER_PLUS_EXPR
for pointers.
(mips_gimplify_va_arg_expr): Likewise.
Don't create BIT_AND_EXPR in a pointer type.
2007-06-15 Andrew Pinski <andrew_pinski@playstation.sony.com>
* trans-intrinsic.c (gfc_conv_intrinsic_repeat): Use
POINTER_PLUS_EXPR instead of PLUS_EXPR for pointer addition.
* trans-expr.c (gfc_trans_string_copy): Create
POINTER_PLUS_EXPR instead of a PLUS_EXPR
for pointer types.
2007-06-15 Andrew Pinski <andrew_pinski@playstation.sony.com>
* typeck.c (build_binary_op): For templates build the
expression in pieces to avoid the assert in build2_stat.
(get_member_function_from_ptrfunc):
Change over to using POINTER_PLUS_EXPR and convert
the second operand to sizetype.
* typeck2.c (build_m_component_ref): Likewise.
* init.c (expand_virtual_init): Create a POINTER_PLUS_EXPR
instead of PLUS_EXPR for pointers.
(build_new_1): Likewise.
(build_vec_delete_1): Likewise.
(build_vec_delete): Likewise.
* class.c (build_base_path): Likewise.
(build_base_path): Likewise.
(convert_to_base_statically): Likewise.
(fixed_type_or_null): Handle POINTER_PLUS_EXPR.
(get_vtbl_decl_for_binfo): Handle POINTER_PLUS_EXPR
instead of PLUS_EXPR.
(dfs_accumulate_vtbl_inits): Create a POINTER_PLUS_EXPR
instead of PLUS_EXPR for pointers.
* call.c (build_special_member_call): Likewise.
* rtti.c (build_headof): Likewise.
Use sizetype instead of ptrdiff_type_node.
(tinfo_base_init): Create a POINTER_PLUS_EXPR
instead of PLUS_EXPR for pointers.
* except.c (expand_start_catch_block): Do a
NEGATIVE and then a POINTER_PLUS_EXPR instead
of a MINUS_EXPR.
* cp-gimplify.c (cxx_omp_clause_apply_fn): Convert
PLUS_EXPR on pointer types over to use
POINTER_PLUS_EXPR and remove the conversion
to the pointer types.
* method.c (thunk_adjust): Use POINTER_PLUS_EXPR for
adding to a pointer type. Use size_int instead of
ssize_int. Convert the index to sizetype before
adding it to the pointer.
2007-06-15 Andrew Pinski <andrew_pinski@playstation.sony.com>
* trans.c (Attribute_to_gnu): When subtracting an
offset from a pointer, use POINTER_PLUS_EXPR with
NEGATE_EXPR instead of MINUS_EXPR.
(gnat_to_gnu): Likewise.
* utils.c (convert): When converting between
thin pointers, use POINTER_PLUS_EXPR and sizetype
for the offset.
* utils2.c (known_alignment): POINTER_PLUS_EXPR
have the same semantics as PLUS_EXPR for alignment.
(build_binary_op): Add support for the semantics of
POINTER_PLUS_EXPR's operands.
When adding an offset to a pointer, use POINTER_PLUS_EXPR.
2007-06-15 Andrew Pinski <andrew_pinski@playstation.sony.com>
* class.c (make_class_data): Build the index in sizetype.
Use POINTER_PLUS_EXPR instead of PLUS_EXPR when
adding to a pointer type.
(build_symbol_entry): Likewise.
* expr.c (build_java_arrayaccess): Likewise.
(build_field_ref): Likewise.
(build_known_method_ref): Likewise.
(build_invokevirtual): Likewise.
* except.c (build_exception_object_ref): Do a
NEGATIVE and then a POINTER_PLUS_EXPR instead
of a MINUS_EXPR.
2007-06-15 Andrew Pinski <andrew_pinski@playstation.sony.com>
PR tree-opt/32225
* gcc.c-torture/compile/20070605-1.c: New test.
* gcc.c-torture/compile/20070603-1.c: New testcase.
* gcc.c-torture/compile/20070603-2.c: New testcase.
PR tree-opt/32167
* gcc.c-torture/compile/20070531-2.c: New test.
PR tree-opt/32144
* gcc.c-torture/compile/20070529-1.c: New test.
PR tree-opt/32145
* gcc.c-torture/compile/20070529-2.c: New test.
PR tree-opt/32015
* gcc.c-torture/compile/20070520-1.c: New test.
* g++.dg/ext/java-1.C: New test.
* gcc.dg/vect/vect-106.c: We are now able to vectorize two
loops instead of one. Remove the "can't determine dependence"
check.
* gcc.dg/tree-ssa/20030815-1.c: Remove testcase which is no longer
needed as the cast is gone in the first place.
* gcc.dg/max-1.c: Change local variable a to be a global one.
* gcc.dg/tree-ssa/ssa-pre-8.c: Update testcase since we don't
have a cast which is PREd.
Mark Mitchell [Fri, 15 Jun 2007 16:58:29 +0000 (16:58 +0000)]
cp-tree.h (DECL_VAR_MARKED_P): Remove.
* cp-tree.h (DECL_VAR_MARKED_P): Remove.
(DECL_ANON_UNION_VAR_P): New macro.
* class.c (fixed_type_or_null): Tidy. Use a hash table, rather
than DECL_VAR_MARKED_P, to keep track of which variables we have
seen.
* decl.c (redeclaration_error_message): Complain about redeclaring
anonymous union members at namespace scope.
* decl2.c (build_anon_union_vars): Set DECL_ANON_UNION_VAR_P.
* g++.dg/lookup/anon6.C: New test.