Skip to content

R
riscv-gcc-1
Switch branch/tag
  1. 07 Sep, 2007 8 commits
    • regrename.c (copyprop_hardreg_forward_1): New variable next. · 62a4a967 
      2007-09-05  Laurynas Biveinis  <laurynas.biveinis@gmail.com>

	* regrename.c (copyprop_hardreg_forward_1): New variable next. Use
	FOR_BB_INSNS_SAFE instead of for loop.
	* cse.c (cse_extended_basic_block): Likewise.
	* postreload.c (reload_cse_regs_1): New variable next. Make sure
	that the for loop does not invoke NEXT_INSN on a deleted insn.
	* function.c (instantiate_virtual_regs): Likewise.
	* lower-subreg.c (remove_retval_note): Likewise.
	(decompose_multiword_subregs): Use FOR_BB_INSNS_SAFE instead of
	FOR_BB_INSNS.
	* emit-rtl.c (remove_insn): Set NEXT_INSN and PREV_INSN to NULL on
	a deleted insn.
	* cfgrtl.c (delete_insn): Set JUMP_LABEL to NULL on a deleted
	insn, if it's a jump.
	(try_redirect_by_replacing_jump): New variable jump_p. Call
	tablejump_p before delete_insn_chain.
	* reload1.c (reload): New variable next. Make sure that the for
	loop does not invoke NEXT_INSN on a deleted insn.
	(fixup_eh_region_note): Make the loop terminate if i becomes NULL.
	(delete_output_reload): New variable prev. Make sure the the for
	loops do not invoke PREV_INSN on a deleted insn.

From-SVN: r128224
      Laurynas Biveinis committed
    • cgraphbuild.c (rebuild_cgraph_edges): Export. · 917948d3 
      	* cgraphbuild.c (rebuild_cgraph_edges): Export.
	* cgraph.h (rebuild_cgraph_edges): Declare.
	* tree-pass.h (pass_expand_omp_ssa): New.
	* omp-low.c (find_omp_clause): Export.
	(copy_var_decl): Split from omp_copy_decl_2.
	(build_omp_barrier): Return the call to emit instead of emitting
	it directly.
	(lower_rec_input_clauses, expand_omp_single): Gimplify the result of
	build_omp_barrier.
	(extract_omp_for_data, expand_parallel_call, expand_omp_parallel,
	expand_omp_for_generic, expand_omp_for_static_nochunk,
	expand_omp_for_static_chunk, expand_omp_for, expand_omp_sections):
	Adapted to work on SSA form.
	(execute_expand_omp): Do not invalidate dominance information.
	(gate_expand_omp): Do not run with -fopenmp-ssa flag.
	(gate_expand_omp_ssa, pass_expand_omp_ssa): New.
	* gimplify.c (gimplify_omp_for): Ensure that the control variable is
	a gimple_reg.
	(force_gimple_operand): Allow gimplifying code expressions without
	value.
	* tree-predcom.c (mark_virtual_ops_for_renaming): Handle phi nodes.
	* common.opt (fopenmp-ssa): New.
	* tree-flow.h (find_omp_clause, copy_var_decl): Declare.
	* Makefile.in (tree-cfg.o): Add TREE_INLINE_H dependency.
	* tree-cfg.c: Include tree-inline.h.
	(struct move_stmt_d): Replace vars_to_remove by vars_map field.
	(replace_by_duplicate_decl, replace_ssa_name,
	mark_virtual_ops_in_region): New functions.
	(move_stmt_r, move_block_to_fn, move_sese_region_to_fn): Adapted
	to work on SSA form.
	* passes.c (init_optimization_passes): Add pass_expand_omp_ssa pass.
	* tree-ssa-operands.c (get_expr_operands): Handle operands of OMP
	constructs.

From-SVN: r128223
      Zdenek Dvorak committed
    • re PR libstdc++/29367 (pb_ds hash containers vs. _GLIBCXX_DEBUG) · e054e6fb 
      
2007-09-06  Benjamin Kosnik  <bkoz@redhat.com>
	
	PR libstdc++/29367
	* include/ext/pb_ds/detail/map_debug_base.hpp: From this..
	* include/ext/pb_ds/detail/debug_map_base.hpp: ...to this.
	Rename map_debug_base to debug_map_base, add std::pair output
	extractors, change duplicate map_debug_base typedef in containers
	to debug_base. Use m_store_extra_indicator for m_store_hash_indicator.
	* include/ext/pb_ds/*: Likewise.
	* include/Makefile.am: Same.
	* include/Makefile.in: Regenerate.

From-SVN: r128222
      Benjamin Kosnik committed
    • re PR libstdc++/29367 (pb_ds hash containers vs. _GLIBCXX_DEBUG) · 551fe1a2 
      2007-09-06  Benjamin Kosnik  <bkoz@redhat.com>
	
	PR libstdc++/29367
	* include/ext/pb_ds/detail/map_debug_base.hpp: From this..
	* include/ext/pb_ds/detail/debug_map_base.hpp: ...to this.
	Rename map_debug_base to debug_map_base, add std::pair output
	extractors, change duplicate map_debug_base typedef in containers
	to debug_base. Use m_store_extra_indicator for m_store_hash_indicator.
	* include/ext/pb_ds/*: Likewise.
	* include/Makefile.am: Same.
	* include/Makefile.in: Regenerate.

From-SVN: r128221
      Benjamin Kosnik committed
    • stl_vector.h (_Vector_base<>::_M_allocate): Do not call _M_impl.allocate when __n == 0. · ccd04b9f 
      2007-09-06  Paolo Carlini  <pcarlini@suse.de>

	* include/bits/stl_vector.h (_Vector_base<>::_M_allocate):
	Do not call _M_impl.allocate when __n == 0.
	* testsuite/23_containers/vector/zero_sized_allocations.cc: New.

From-SVN: r128220
      Paolo Carlini committed
    • tree-loop-linear.c: Include obstack.h. · b9dd78fa 
      2007-09-06  Laurynas Biveinis  <laurynas.biveinis@gmail.com>

	* tree-loop-linear.c: Include obstack.h.
	(linear_transform_loops): New obstack lambda_obstack.
	Initialize it, pass it to gcc_loopnest_to_lambda_loopnest,
	lambda_loopnest_transform, lambda_loopnest_to_gcc_loopnest calls
	and free afterwards.
	* lambda.h (struct obstack): New forward declaration.
	(lambda_linear_expression_new): New parameter of type struct
	obstack *.
	(lambda_loopnest_new): Likewise.
	(lambda_loopnest_transform): Likewise.
	(lambda_body_vector_new): Likewise.
	(lambda_body_vector_compute_new): Likewise.
	(gcc_loopnest_to_lambda_loopnest): Likewise.
	(lambda_loopnest_to_gcc_loopnest): Likewise.
	* lambda-code.c: Include obstack.h.
	(lambda_lattice_new): New parameter lambda_obstack.  Use it for
	allocation of ret.
	(lambda_body_vector_new): Likewise.
	(lambda_linear_expression_new): Likewise.
	(lambda_lattice_new): Likewise.
	(lambda_loopnest_new): Likewise.  Additionally use obstack to
	allocate LN_LOOPS(ret).
	(lambda_lattice_compute_base): New parameter lambda_obstack.  Pass
	it to lambda_lattice_new.
	(lambda_body_vector_compute_new): New parameter lambda_obstack.
	Pass it to lambda_body_vector_new.
	(lambda_lattice_compute_base): New paramater lambda_obstack.  Pass
	it to lambda_lattice_new.
	(compute_nest_using_fourier_motzkin): New parameter lambda_obstack.
	Pass it to lambda_loopnest_new, lambda_linear_expression_new.
	(lambda_compute_target_space): Likewise.
	(lambda_compute_auxillary_space): New parameter lambda_obstack.
	Pass it to lambda_lattice_compute_base and
	compute_nest_using_fourieer_motzkin.
	(lambda_loopnest_transform): New parameter lambda_obstack.  Pass
	it to lambda_lattice_compute_base, lambda_lattice_auxillary_space
	and lambda_lattice_compute_target_space.
	(gcc_tree_to_linear_expression): Nex parameter lambda_obstack.
	Pass it to lambda_linear_expression_new.
	(gcc_loop_to_lambda_loop): New parameter lambda_obstack.  Pass it
	to gcc_tree_to_linear_expression.
	(gcc_loopnest_to_lambda_loopnest): New parameter lambda_obstack.
	Pass it to gcc_loop_to_lambda_loop and lambda_loopnest_new.
	(lambda_loopnest_to_gcc_loopnest): New parameter lambda_obstack.
	Pass it to lambda_body_vector_new and
	lambda_body_vector_compute_new.
	* Makefile.in (tree-loop-linear.o): Add $(OBSTACK_H) dependency.
	(lambda-code.o): Likewise.

From-SVN: r128219
      Laurynas Biveinis committed
    • stdfix.h: New file. · 0f996086 
      	* ginclude/stdfix.h: New file.
	* Makefile.in (USER_H): Add $(srcdir)/ginclude/stdfix.h.
	(convert.o): Add dependence on fixed-value.h.
	* c-convert.c (convert): Support FIXED_POINT_TYPE.
	* c-cppbuiltin.c (builtin_define_fixed_point_constants): New function
	to define fixed-point constants.
	(c_cpp_builtins): Define fixed-point constants.
	* convert.c (fixed-value.h): New include.
	(convert_to_real): Update comment to include fixed-point.
	Support FIXED_POINT_TYPE.
	(convert_to_integer): Update comment to include fixed-point.
	Support FIXED_POINT_TYPE.
	(convert_to_complex): Support FIXED_POINT_TYPE.
	(convert_to_fixed): New function.
	* convert.h (convert_to_fixed): Declare.
	* genopinit.c: Add comment about $Q for only fixed-point modes.
	(optabs): Add fract_optab, fractuns_optab, satfract_optab,
	satfractuns_optab, add_optab, ssadd_optab, usadd_optab, sub_optab,
	sssub_optab, ussub_optab, smul_optab, ssmul_optab, usmul_optab,
	ssmadd_widen_optab, usmadd_widen_optab, ssdiv_optab, udiv_optab,
	usdiv_optab, ssashl_optab, usashl_optab, neg_optab, ssneg_optab,
	usneg_optab for fixed-point modes.
	(gen_insn): Add force_fixed to track the $Q format for all fixed-point
	modes.
	* optabs.c (optab_for_tree_code): For *DIV_EXPR, LSHIFT_EXPR,
	PLUS_EXPR, MINUS_EXPR, MULT_EXPR, NEGATE_EXPR, return signed or
	unsigned saturation optabs, when type is saturating.
	(shift_optab_p): Return true for SS_ASHIFT or US_ASHIFT.
	(expand_fixed_convert): New function.
	(gen_fixed_libfunc, gen_signed_fixed_libfunc,
	gen_unsigned_fixed_libfunc, gen_int_fp_fixed_libfunc,
	gen_int_fp_signed_fixed_libfunc, gen_int_fixed_libfunc,
	gen_int_signed_fixed_libfunc, gen_int_unsigned_fixed_libfunc,
	gen_fract_conv_libfunc, gen_fractuns_conv_libfunc,
	gen_satfract_conv_libfunc, gen_satfractuns_conv_libfunc): New
	functions.
	(init_optabs): Initialize ssadd_optab, usadd_optab, sssub_optab,
	ussub_optab, ssmul_optab, usmul_optab, ssmadd_widen_optab,
	usmadd_widen_optab, ssmsub_widen_optab, usmsub_widen_optab,
	ssdiv_optab, usdiv_optab, ssashl_optab, usashl_optab, ssneg_optab,
	usneg_optab, fract_optab, fractuns_optab, satfract_optab,
	satfractuns_optab.
	Initialize fixed-point libraries, including add, ssadd, usadd, sub,
	sssub, ussub, mul, ssmul, usmul, div, ssdiv, udiv, usdiv, ashl,
	ssashl, usashl, ashr, lshr, neg, ssneg, usneg, cmp, fract, satfract,
	fractuns, satfractuns.
	* optabs.h (enum optab_index): Add OTI_ssadd, OTI_usadd, OTI_sssub,
	OTI_ussub, OTI_ssmul, OTI_usmul, OTI_ssdiv, OTI_usdiv, OTI_ssneg,
	OTI_usneg, OTI_ssashl, OTI_usashl, OTI_ssmadd_widen, OTI_usmadd_widen,
	OTI_ssmsub_widen,  OTI_usmsub_widen.
	(ssadd_optab, usadd_optab, sssub_optab, ussub_optab, ssmul_optab,
	usmul_optab, ssdiv_optab, usdiv_optab, ssneg_optab, usneg_optab,
	ssashl_optab, usashl_optab, ssmadd_widen_optab, usmadd_widen_optab,
	umsub_widen_optab, usmsub_widen_optab): Define.
	(enum convert_optab_index): Add COI_fract, COI_fractuns, COI_satfract,
	COI_satfractuns.
	(fract_optab, fractuns_optab, satfract_optab, satfractuns_optab):
	Define.
	(expand_fixed_convert): Declare.
	* expr.c (convert_move): Support the move of fixed-point modes.
	(emit_move_insn_1): Handle fixed-point mode to move via integer.
	(categorize_ctor_elements_1): Handle FIXED_CST.
	(count_type_elements): Handle FIXED_POINT_TYPE.
	(expand_expr_real_1): For VECTOR_CST, check MODE_VECTOR_FRACT,
	MODE_VECTOR_UFRACT, MODE_VECTOR_ACCUM, MODE_VECTOR_UACCUM.
	Support FIXED_CST.
	For PLUS_EXPR and MINUS_EXPR, support saturating and non-saturating
	multiply and add/subtract for fixed-point types.
	For MULT_EXPR, *DIV_EXPR, *SHIFT_EXPR, if the mode if a fixed-point
	mode, we jump to binop directly.
	Support FIXED_CONVERT_EXPR.
	(do_store_flag): Check FIXED_CST to put a constant second.
	(vector_mode_valid_p): Handle MODE_VECTOR_FRACT,
	MODE_VECTOR_UFRACT, MODE_VECTOR_ACCUM, MODE_VECTOR_UACCUM.
	(const_vector_from_tree): Support FIXED_CST.
	* doc/extend.texi (Fixed-Point): New node.
	* doc/md.texi (ssadd, usadd, sssub, ussub, ssmul, usmul, ssdiv, usdiv,
	ssmadd, usmadd, ssmsub, usmsub, ssashl, usashl, ssneg, usneg, fract,
	satfract, fractuns, satfractuns): Document them.

From-SVN: r128218
      Chao-ying Fu committed
    • Daily bump. · 58cfe698 
      From-SVN: r128215
      GCC Administrator committed
  3. 06 Sep, 2007 32 commits