gcc/
* config/arm/arm.c (neon_vdup_constant, neon_make_constant): New.
(neon_expand_vector_init): Use them. Also handle non-constant
vectors with identical elements and vectors with only one
non-constant element.
(arm_print_operand): Handle 'y' modifier.
* config/arm/arm-protos.h (neon_make_constant): Declare.
* config/arm/neon.md (neon_vdup_n<mode>): Split into two
patterns. Use VX instead of VDQW for the first one. Allow
a VFP alternative and V32 modes for the second one.
* config/arm/neon.ml (shape_elt): Add Alternatives.
(ops): Use Alternatives for vdup lane instructions.
* config/arm/neon-testgen.ml (analyze_shape): Handle Alternatives.
* config/arm/vec-common.md (mov<mode>): Use neon_make_constant.
gcc/testsuite/
* gcc.target/arm/neon: Regenerate generated tests.
From-SVN: r154094
* gcc.c-torture/execute/ipa-sra-2.c: Use calloc instead of malloc.
* gcc.dg/pr41963.c: Do not include math.h. Declare fabsf manually.
From-SVN: r154067
2009-11-09 Paul Brook <paul@codesourcery.com>
Daniel Jacobowitz <dan@codesourcery.com>
Sandra Loosemore <sandra@codesourcery.com>
gcc/
* doc/extend.texi (Half-Precision): Update wording to reflect
that there are now multiple -mfpu options that enable fp16
hardware support.
* doc/invoke.texi: Update list of ARM -mfpu= options.
* config.gcc: Update ARM --with-fpu option list.
* config/arm/arm.c (all_fpus): Add vfpv3-fp16, vfpv3-d16-fp16,
vfpv3xd and vfpv3xd-fp16.
(use_vfp_abi): New function.
(aapcs_vfp_is_call_or_return_candidate): Avoid double precision regs
when undesirable.
(aapcs_vfp_is_return_candidate, aapcs_vfp_is_call_candidate,
aapcs_vfp_allocate_return_reg): Use use_vfp_abi.
(arm_rtx_costs_1, arm_size_rtx_costs, arm_fastmul_rtx_costs,
arm_9e_rtx_costs): Only expect double-precision operations if the FPU
provides them.
(coproc_secondary_reload_class): Reload HFmode via GENERAL_REGS if no
NEON.
(arm_print_operand): Handle 'p' modifier.
(arm_hard_regno_mode_ok): : Allow HFmode in VFP registers if
TARGET_FP16.
* config/arm/arm.h (TARGET_VFP_SINGLE, TARGET_VFP_DOUBLE): Define.
(TARGET_FP16): Define.
* config/arm/vfp.md: Disable double-precision patterns if the FPU
does not provide them.
(arm_movdi_vfp, thumb2_movdi_vfp): Use fcpys to move
double-precision values on a single-precision FPU.
(movdf_vfp, thumb2_movdf_vfp): Likewise. Use "Dy" for
double-precision constants.
(movhf_vfp_neon): New pattern (was movhf_vfp).
(movhf_vfp): Remove NEON instructions.
* config/arm/constraints.md: Add new "Dy" constraint for
double-precision constants. Update description of "Dv".
* config/arm/arm.md: Disable double-precision patterns if the FPU
does not provide them
Co-Authored-By: Daniel Jacobowitz <dan@codesourcery.com>
Co-Authored-By: Sandra Loosemore <sandra@codesourcery.com>
From-SVN: r154034
PR target/41985
* config/i386/i386.c (get_some_local_dynamic_name): Don't assert
function contains at least one local dynamic name.
(print_operand) <case '&'>: Instead output operand lossage diagnostics
here if that happens.
* gcc.target/i386/pr41985.c: New test.
From-SVN: r154014
2009-11-08 Richard Guenther <rguenther@suse.de>
* tree-ssa-structalias.c (build_succ_graph): Properly make
variables escape if they are stored to anything.
* gcc.dg/torture/pta-escape-1.c: New testcase.
From-SVN: r154010
* gcc-interface/decl.c (make_packable_type): Fix oversight.
(gnat_to_gnu_field): Do not attempt to change the form of the type
if the field requires strict alignment. Always change the form of
the type if the specified size is smaller than its size.
From-SVN: r154009
2009-11-08 Richard Guenther <rguenther@suse.de>
PR rtl-optimization/41928
* loop-invariant.c (free_loop_data): If we didn't allocate
loop data do not try to free it.
* gfortran.dg/pr41928.f90: New testcase.
From-SVN: r154008
PR c++/18451
PR c++/40738
* cp-tree.h (cp_decl_specifier_seq): Add any_type_specifiers_p.
* parser.c (cp_parser_single_declaration): Call
cp_parser_parse_and_diagnose_invalid_type_name here, too.
(cp_parser_parameter_declaration): And here.
(cp_parser_parse_and_diagnose_invalid_type_name): Be
less picky about declarator form. Don't skip to
the end of the block if we're in a declarator.
(cp_parser_decl_specifier_seq): Set any_type_specifiers_p.
(cp_parser_simple_declaration): Check it.
(cp_parser_member_declaration): Likewise.
(cp_parser_diagnose_invalid_type_name): Tweak error message.
(cp_parser_expression_statement): Likewise.
* decl2.c (grokfield): Mention decltype instead of typeof.
From-SVN: r154006
2009-11-07 Richard Guenther <rguenther@suse.de>
* tree-ssa-structalias.c (build_succ_graph): Feed stores
to anything only to variables that can take pointers.
(get_constraint_for_ssa_var): Properly exclude full
variables from expanding.
(first_vi_for_offset): Avoid overflow in arithmetic.
(first_or_preceding_vi_for_offset): Likewise.
(count_num_arguments): Fix implementation.
(gate_ipa_pta): Do not run when not optimizing.
From-SVN: r154002
2009-11-07 David Binderman <dcb314@hotmail.com>
* builtins.c (apply_args_size, expand_builtin_mathfn, expand_powi,
expand_builtin_expect, maybe_emit_sprintf_chk_warning): Remove some
set but not used variables.
* calls.c (flags_from_decl_or_type): Likewise.
* genautomata.c (check_regexp_units_distribution): Likewise.
From-SVN: r154001
