PR target/25731
* config.gcc (hppa*-*-linux*, hppa[12]*-*-hpux10*, hppa*64*-*-hpux11*,
hppa[12]*-*-hpux11*): Override default shared libgcc version for both
sjlj and dwarf2 exception handling.
* pa/t-hpux-shlib (SHLIB_SOVERSION): New make variable.
Rework to allow overriding SHLIB_EXT and SHLIB_SOVERSION.
* pa/pa.c (function_value): Treat complex and vector types as
aggregates.
(function_arg): Likewise. Only pass scalar floats in the floating
point argument registers.
* pa/t-slibgcc-dwarf-ver: New file.
* pa/t-slibgcc-sjlj-ver: New file.
* pa/t-slibgcc-elf-ver: Delete file.
From-SVN: r109894
* regrename.c (regrename_optimize, copyprop_hardreg_forward):
Make them static.
* rtl.h: Remove the prototypes for regrename_optimize and
copyprop_hardreg_forward.
From-SVN: r109891
2006-01-18 Perry Smith <pedz@easesoftware.net>
PR libstdc++/25823
PR libstdc++/25824
* libsupc++/eh_alloc.cc: Fix return type of memset declaration.
* libsupc++/eh_globals.cc: If !_GLIBCXX_HOSTED declare malloc and free.
2006-01-18 Paolo Carlini <pcarlini@suse.de>
* include/ext/pb_assoc/detail/value_type_adapter/
value_type_adapter.hpp: Include <tr1/type_traits>.
* include/ext/pb_assoc/detail/value_type_adapter/
it_value_type_traits.hpp (it_value_type_traits_<>::value_type_holder):
Use tr1::aligned_storage and tr1::alignment_of.
(it_value_type_traits_<>::buf_t): Remove.
(it_value_type_traits_<>::make_valid, recast): Adjust.
From-SVN: r109883
* config/alpha/alpha.c (alpha_split_tfmode_pair): Rename into
alpha_split_tmode_pair. Add 'mode' and 'fixup_overlap' arguments.
Test against the appropriate null constant for the mode.
If 'fixup_overlap' is true, swap the operands if they overlap.
(alpha_split_tfmode_frobsign): Adjust call to alpha_split_tfmode_pair.
* config/alpha/alpha-protos.h (alpha_split_tfmode_pair): Rename into
alpha_split_tmode_pair and adjust for above change.
* config/alpha/alpha.md (movtf_internal): Adjust call to
alpha_split_tfmode_pair and rely on it to swap the operands.
(movti_internal): New insn and post-reload splitter.
(movti): New expander.
From-SVN: r109882
2006-01-17 Andreas Jaeger <aj@suse.de>
Janis Johnson <janis187@us.ibm.com>
* g++.dg/abi/mangle25.C: Run always on ia64 but only on x86/x86-64
in 64-bit.
* gcc.dg/torture/fp-int-convert-float128-timode.c: Likewise.
* gcc.dg/torture/fp-int-convert-float128.c: Likewise.
From-SVN: r109881
* regmove.c (regmove_optimize, combine_stack_adjustments):
Make them static.
* rtl.h: Remove the prototypes for regmove_optimize and
combine_stack_adjustments.
From-SVN: r109875
2006-01-17 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libgfortran/25697
* io/transfer.c (us_read): Detect end of file condition from previous
operations and bail out (no need to pre-position).
From-SVN: r109870
* config/i386/i386.c (COSTS_N_BYTES): New macro.
(size_cost): Encode costs in bytes so that a reg-reg addition is
equivalent to COSTS_N_INSNS(1).
(override_options): Use the size_cost.branch_cost as the default
ix86_branch_cost when optimizing for size.
From-SVN: r109867
2006-01-17 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libgfortran/25631
* io/transfer.c (formatted_transfer_scalar): Adjust pending_spaces and
skips so that TL works correctly when no bytes_used yet.
From-SVN: r109858
* Makefile.in (D32PBIT_FUNCS): New.
(D64PBIT_FUNCS, D128PBIT_FUNCS): Likewise.
(libgcc.mk): Set D32PBIT, D64PBIT, D128PBIT, D32PBIT_FUNCS,
D64PBIT_FUNCS and D128PBIT_FUNCS.
(LIBGCC_DEPS): Include $(D32PBIT), $(D64PBIT), $(D128PBIT).
* mklibgcc.in Bring in the DFP support code if D32PBIT, D64PBIT or
D128PBIT are set.
(decnumber_dep): Define.
* doc/libgcc.texi (Decimal float library routines): New node.
* config/dfp-bit.h, config/dfp-bit.c: New files.
From-SVN: r109856
* expr.c (emit_move_change_mode): Always adjust addresses, not
just during reload. Copy replacements only during reload.
(emit_move_insn_1): Move MODE_DECIMAL_FLOAT modes by invoking
emit_move_via_integer.
From-SVN: r109845
2006-01-17 Shantonu Sen <ssen@opendarwin.org>
* Makefile.in (check-%, check-consistency): Use $${srcdir}
instead of $(srcdir) and ${srcdir}.
From-SVN: r109840
PR target/592
PR middle-end/11135
* arm.h (struct machine_function): Add pic_reg.
* arm.c (arm_pic_register): Make unsigned.
(arm_override_options): Only set arm_pic_register if
TARGET_SINGLE_PIC_BASE.
(use_return_insn): Only test for a pic register if it is fixed.
(arm_compute_save_reg0_reg12_mask): Likewise.
(thumb_compute_save_reg_mask): Likewise.
(legitimate_pic_operand): Factor out some known invariants.
(legitimize_pic_address): If we don't have a fixed pic register,
then set up a pseudo in the function entry sequence. Handle the
pic base being in a pseudo.
(arm_load_pic_register): Handle the pic register being in a pseudo.
(arm_expand_prologue): Only set up the pic register if it is fixed.
(thumb_expand_prologue): Likewise.
* arm.md (pic_load_addr_based): Handle the pic base being a pseudo.
(pic_load_addr_based_insn): Likewise.
(builtin_setjmp_receiver): Don't restore the pic base if it isn't
fixed.
From-SVN: r109839
2006-01-17 Andreas Jaeger <aj@suse.de>
Janis Johnson <janis187@us.ibm.com>
* g++.dg/abi/mangle25.C: Run always on ia64 but only on
* x86/x86-64
in 64-bit.
* gcc.dg/torture/fp-int-convert-float128-timode.c: Likewise.
* gcc.dg/torture/fp-int-convert-float128.c: Likewise.
Co-Authored-By: Janis Johnson <janis187@us.ibm.com>
From-SVN: r109832
PR c/25682
* c-typeck.c (build_unary_op): Fold offsetof-like expressions
even when the pointer is not NULL.
cp/
* decl.c (compute_array_index_type): After issuing not an integral
constant-expression error, set size to 1 to avoid ICEs later on.
testsuite/
* gcc.dg/pr25682.c: New test.
* g++.dg/parse/array-size2.C: New test.
From-SVN: r109812
./: * common.opt (ftoplevel-reorder): New option.
* cgraph.c (cgraph_asm_nodes): New global variable.
(cgraph_asm_last_node): New static variable.
(cgraph_order): New global variable.
(cgraph_create_node): Set new order field.
(cgraph_varpool_node): Likewise.
(decide_is_variable_needed): Return true if not
flag_toplevel_reorder.
(cgraph_add_asm_node): New function.
* cgraph.h (struct cgraph_node): Add order field.
(struct cgraph_varpool_node): Add order field.
(struct cgraph_asm_node): Define.
(cgraph_asm_nodes, cgraph_order): Declare.
(cgraph_add_asm_node): Declare.
* cgraphunit.c (cgraph_varpool_assemble_decl): New static
function.
(cgraph_varpool_assemble_pending_decls): Call it.
(cgraph_output_pending_asms): New static function.
(cgraph_finalize_compilation_unit): Call it.
(struct cgraph_order_sort): Define.
(cgraph_output_in_order): New static function.
(cgraph_optimize): Call cgraph_output_pending_asms. Add code for
!flag_toplevel_reorder case.
* c-parser.c: Include "cgraph.h".
(c_parser_asm_definition): Call cgraph_add_asm_node rather than
assemble_asm.
* Makefile.in (CRTSTUFF_CFLAGS): Use -fno-toplevel-reorder rather
than -fno-unit-at-a-time.
* doc/invoke.texi (Option Summary): Mention
-fno-toplevel-reorder.
(Optimize Options): Document -fno-toplevel-reorder. Mention it in
-funit-at-a-time documentation.
cp/:
* parser.c: Include "cgraph.h".
(cp_parser_asm_definition): Call cgraph_add_asm_node rather than
assemble_asm.
From-SVN: r109811
