Age | Commit message (Collapse) | Author | Files | Lines |
|
License and to point readers at the COPYING3 file and the FSF's license web page.
From-SVN: r126948
|
|
* doc/rtl.texi (Machine Modes): Document QQ, HQ, SQ, DQ, TQ,
UQQ, UHQ, USQ, UDQ, UTQ, HA, SA, DA, TA, UHA, USA, UDA, and UTAmodes.
Document MODE_FRACT, MODE_UFRACT, MODE_ACCUM, MODE_UACCUM.
Document GET_MODE_IBIT, and GET_MODE_FBIT.
* machmode.h (VECTOR_MODE_P): Test MODE_VECTOR_FRACT,
MODE_VECTOR_UFRACT, MODE_VECTOR_ACCUM, MODE_VECTOR_UACCUM.
(SCALAR_FRACT_MODE_P, SCALAR_UFRACT_MODE_P, ALL_SCALAR_FRACT_MODE_P,
SCALAR_ACCUM_MODE_P, SCALAR_UACCUM_MODE_P, ALL_SCALAR_ACCUM_MODE_P,
SIGNED_SCALAR_FIXED_POINT_MODE_P, UNSIGNED_SCALAR_FIXED_POINT_MODE_P,
ALL_SCALAR_FIXED_POINT_MODE_P, FRACT_MODE_P, UFRACT_MODE_P,
ALL_FRACT_MODE_P, ACCUM_MODE_P, UACCUM_MODE_P, ALL_ACCUM_MODE_P,
SIGNED_FIXED_POINT_MODE_P, UNSIGNED_FIXED_POINT_MODE_P,
ALL_FIXED_POINT_MODE_P): New define.
(CLASS_HAS_WIDER_MODES_P): Test MODE_FRACT, MODE_UFRACT, MODE_ACCUM,
MODE_UACCUM.
(GET_MODE_IBIT, GET_MODE_FBIT): New define.
* mode-classes.def (MODE_FRACT, MODE_UFRACT, MODE_ACCUM, MODE_UACCUM,
MODE_VECTOR_FRACT, MODE_VECTOR_UFRACT, MODE_VECTOR_ACCUM,
MODE_VECTOR_UACCUM): New mode classes.
* machmode.def: Document FRACT_MODE, UFRACT_MODE, ACCUM_MODE,
UACCUM_MODE, ADJUST_IBIT, and ADJUST_FBIT.
Add QQ, HQ, SQ, DQ, TQ, UQQ, UHQ, USQ, UDQ, UTQ, HA, SA, DA, TA, UHA,
USA, UDA, and UTA.
* genmodes.c (struct mode_data): Add ibit and fbit fields.
(blank_mode): Initialize ibit and fbit.
(adj_ibit, adj_fbit): New to adjust ibit and fbit.
(vector_class): Handle MODE_FRACT, MODE_UFRACT, MODE_ACCUM, MODE_UACCUM.
(new_adjust): Change required_class to required_class_from and
required_class_to for testing within a range.
(complete_mode): Handle MODE_FRACT, MODE_UFRACT, MODE_ACCUM,
MODE_UACCUM, MODE_VECTOR_FRACT, MODE_VECTOR_UFRACT, MODE_VECTOR_ACCUM,
MODE_VECTOR_UACCUM.
(FRACT_MODE, UFRACT_MODE, ACCUM_MODE, UACCUM_MODE): New define.
(make_fixed_point_mode): New.
(_ADD_ADJUST): Change C to C1 and C2.
(ADJUST_BYTESIZE, ADJUST_ALIGNMENT, ADJUST_FLOAT_FORMAT): Change to
use a range for machine classes.
(ADJUST_IBIT, ADJUST_FBIT): New.
(emit_insn_modes_h): Output defines of CONST_MODE_IBIT and
CONST_MODE_FBIT.
(emit_mode_adjustments): Handle MODE_VECTOR_FRACT, MODE_VECTOR_UFRACT,
MODE_VECTOR_ACCUM, MODE_VECTOR_UACCUM.
Emit adjustment for ibit and fbit.
(emit_mode_ibit, emit_mode_fbit): New.
(emit_insn_modes_c): Add emit_mode_ibit and emit_mode_fbit.
From-SVN: r126007
|
|
2005-12-02 Jon Grimm <jgrimm2@us.ibm.com>
Janis Johnson <janis187@us.ibm.com>
David Edelsohn <dje@watson.ibm.com>
Ben Elliston <bje@au.ibm.com>
* dfp.h, dfp.c: New files.
* Makefile.in (DECNUM, DECNUMINC, LIBDECNUMBER): New variables.
(DECNUM_H): Likewise.
(LIBDEPS, LIBS, BACKEND): Append $(LIBDECNUMBER).
(INCLUDES): Append $(DECNUMINC).
(OBJS-common): Add dfp.o.
(dfp.o): New rule.
* real.h (EXP_BITS): Pinch one bit to ..
(struct real_value): Add decimal field.
(real_format): Change table size, update documentation.
(REAL_MODE_FORMAT): Update for to handle float, decimal float.
(real_from_string3): Declare.
(decimal_single_format): Declare.
(decimal_double_format): Declare.
(decimal_quad_format): Declare.
(REAL_VALUE_TO_TARGET_DECIMAL32): New.
(REAL_VALUE_TO_TARGET_DECIMAL64): New.
(REAL_VALUE_TO_TARGET_DECIMAL128): New.
* real.c: Include dfp.h.
(normalize): Early return for decimal floats.
(do_add): Zero decimal field.
(do_compare): Call do_decimal_compare for decimal floats.
(do_fix_trunc): Likewise, call decimal_do_fix_trunc.
(real_arithmetic): Call decimal_real_arithmetic for decimal
floating point operands.
(real_identical): If a and b are of differing radix, return false.
(real_to_integer): Call decimal_real_to_integer if the value is a
decimal float.
(real_to_integer2): Likewise, call decimal_real_to_integer2.
(real_to_decimal): Likewise, call decimal_real_to_decimal.
(real_to_hexadecimal): Place "N/A" in the return string for
decimal float.
(real_from_string3): New variant, given a mode.
(real_maxval): Use decimal_real_maxval for decimal floats.
(round_for_format): Use decimal_round_for_format for decimals.
(real_convert): Use decimal_real_convert where appropriate.
(significand_size): Handle base 10.
(encode_decimal_single, decode_decimal_single,
encode_decimal_double, decode_decimal_double, encode_decimal_quad,
decode_decimal_quad): New functions.
(decimal_single_format): New.
(decimal_double_format): New.
(decimal_quad_format): New.
* machmode.def: Add SD, DD and TD decimal floating point modes.
* machmode.h (FLOAT_MODE_P, SCALAR_FLOAT_MODE_P, MODES_WIDEN_P):
Include MODE_DECIMAL_FLOAT.
(DECIMAL_FLOAT_MODE_P): New.
* mode-classes.def (MODE_DECIMAL_FLOAT): New mode class.
* genmodes.c (struct mode_data): Add counter field.
(struct mode_data): Update comment for format.
(blank_mode): Initialise counter field.
(new_mode): Increment counter field for each mode defined.
(complete_mode): Handle MODE_DECIMAL_FLOAT, update check for mode
using a format.
(make_complex_modes): Handle modes containing `D'.
(DECIMAL_FLOAT_MODE, FRACTIONAL_DECIMAL_FLOAT_MODE): New.
(make_decimal_float_mode): New.
(reset_float_format): Handle MODE_DECIMAL_FLOAT.
(cmp_modes): Compare counter field if other characteristics
similar.
(emit_real_format_for_mode): Support formats for decimal floats.
* doc/rtl.texi (Machine Modes): Document SD, DD and TDmodes.
Document MODE_DECIMAL_FLOAT.
Co-Authored-By: Ben Elliston <bje@au.ibm.com>
Co-Authored-By: David Edelsohn <dje@watson.ibm.com>
Co-Authored-By: Janis Johnson <janis187@us.ibm.com>
From-SVN: r107861
|
|
From-SVN: r101317
|
|
* attribs.c, c-pragma.c, caller-save.c, cfghooks.h,
coverage.c, cselib.h, domwalk.c, domwalk.h, errors.c,
errors.h, gcov-dump.c, gcov-io.c, gcov-io.h, gen-protos.c,
genattrtab.h, genextract.c, gthr-win32.h, insn-notes.def,
integrate.c, lambda-mat.c, lambda.h, libgcov.c, local-alloc.c,
machmode.def, mips-tfile.c, params.c, pretty-print.c,
print-rtl.c, protoize.c, regmove.c, sched-vis.c, tree-chrec.h,
tree-data-ref.h, vec.h, config/darwin-c.c, config/sol2-c.c,
config/sol2.c, config/arm/arm-cores.def, config/arm/cirrus.md,
config/arm/symbian.h, config/c4x/c4x.c, config/c4x/c4x.h,
config/i386/cygming.h, config/i386/djgpp.h,
config/i386/lynx.h, config/i386/netware.c,
config/i386/winnt.c, config/ia64/ia64-c.c,
config/iq2000/iq2000.c, config/m32r/little.h,
config/m68k/m68k-protos.h, config/m68k/m68k.h,
config/m68k/m68k.md, config/mcore/mcore.c,
config/mcore/mcore.h, config/mmix/mmix.c, config/mmix/mmix.md,
config/mn10300/mn10300-protos.h, config/mn10300/mn10300.c,
config/mn10300/mn10300.h, config/ns32k/netbsd.h,
config/ns32k/ns32k.c, config/ns32k/ns32k.h,
config/pa/pa-hpux11.h, config/pdp11/pdp11.c,
config/pdp11/pdp11.h, config/rs6000/darwin.h,
config/rs6000/default64.h, config/rs6000/rs6000-c.c,
config/s390/2064.md, config/s390/2084.md,
config/s390/s390-modes.def, config/s390/s390-protos.h,
config/s390/tpf.h, config/sh/sh.h, config/sh/symbian.c,
config/stormy16/stormy16.c, config/vax/vax-protos.h,
config/vax/vax.c, config/vax/vax.h,
config/xtensa/lib1funcs.asm, config/xtensa/xtensa.md: Update
copyright.
From-SVN: r98914
|
|
* cgraphunit.c, dbxout.c, flow.c, gcse.c, gimplify.c,
lambda-code.c, loop.c, machmode.def, mips-tfile.c,
modulo-sched.c, passes.c, postreload-gcse.c, tree-eh.c,
tree-ssa-ccp.c, varasm.c, config/frv/frv.c, config/frv/frv.h,
config/frv/frv.md, config/i386/i386.c, config/i386/i386.h,
config/i386/i386.md, config/rs6000/predicates.md,
config/rs6000/rs6000.c, config/s390/fixdfdi.h,
config/s390/s390.c, config/stormy16/stormy16.c,
config/stormy16/stormy16.md, config/vax/vax.md: Fix comment
typos.
From-SVN: r97362
|
|
* bitmap.c, bitmap.h, builtin-attrs.def, cfglayout.h,
collect2.h, conflict.c, coretypes.h, coverage.h, errors.h,
gcc.h, gcov-dump.c, gen-protos.c, gencheck.c, gencodes.c,
genconstants.c, gengenrtl.c, genmodes.c, genpeep.c,
gensupport.h, graph.h, gthr-gnat.c, gthr-gnat.h, gthr-posix.c,
gthr-posix.h, hard-reg-set.h, hwint.h, integrate.h, libgcc2.c,
lists.c, machmode.def, mips-tdump.c, opts.h, params.c,
predict.def, predict.h, protoize.c, reload.h, resource.h,
rtl-error.c, sbitmap.c, sbitmap.h, sreal.c, tlink.c,
tree-dump.h, tree-inline.h, unwind-dw2-fde.c, unwind-dw2.c,
unwind-sjlj.c, value-prof.h: Update copyright.
From-SVN: r89097
|
|
2004-07-22 Paolo Bonzini <bonzini@gnu.org>
* machmode.def: Remove vector modes.
* config/alpha/alpha-modes.def: Add supported vector modes.
* config/arm/arm-modes.def: Likewise.
* config/frv/frv-modes.def: Likewise.
* config/i386/i386-modes.def: Likewise.
* config/rs6000/rs6000-modes.def: Likewise.
* config/sh/sh-modes.def: Likewise.
From-SVN: r85076
|
|
* genmodes.c: Change the word "bitsize" to "precision" throughout.
* machmode.def: Likewise.
* machmode.h (GET_MODE_SIZE): Cast value to unsigned short.
(GET_MODE_BITSIZE): Define as GET_MODE_SIZE * BITS_PER_UNIT.
(GET_MODE_PRECISION): New macro.
(mode_bitsize): Renamed mode_precision.
* stor-layout.c (mode_for_size, smallest_mode_for_size):
Use GET_MODE_PRECISION; clarify comments.
ada:
* misc.c (fp_prec_to_size, fp_size_to_prec): Use GET_MODE_PRECISION
and update for changed meaning of GET_MODE_BITSIZE.
From-SVN: r73295
|
|
* genmodes.c: Include hashtab.h.
(modes_by_name, hash_mode, eq_mode, struct mode_adjust)
(adj_bytesize, adj_alignment, adj_format, new_adjust)
(_ADD_ADJUST, ADJUST_BYTESIZE, ADJUST_ALIGNMENT, ADJUST_FORMAT)
(print_maybe_const_decl, emit_mode_adjustments): New.
(known_modes): Rename to modes.
(find_mode): Kill class argument; look up in hash table.
(new_mode): Insert into hash table also.
(new_adjust): New.
(reset_float_format, make_partial_integer_mode)
(make_vector_mode): Tweak error reporting.
(reset_float_format): Correct type of fourth argument.
(emit_insn_modes_h): Add #defines to help make mode_size,
mode_base_align, and real_format_for_mode conditionally const.
(emit_mode_size, emit_mode_base_align): Use print_maybe_const_decl.
(emit_real_format_for_mode): Likewise, but temporarily disabled.
(emit_insn_modes_c): Call emit_mode_adjustments.
(main): Initialize modes_by_name.
* Makefile.in: Update dependencies.
* machmode.def: Document EXPR arguments and new ADJUST_* statements.
* machmode.h: Use CONST_MODE_SIZE and CONST_MODE_BASE_ALIGN in
declarations of mode_size and mode_base_align. Declare
init_adjust_machine_modes.
* toplev.c (backend_init): Call init_adjust_machine_modes.
From-SVN: r72534
|
|
min-insn-modes.o.
* Makefile.in (BUILD_RTL): Replace $(BUILD_PREFIX)insn-modes.o
with min-insn-modes.o.
(STAGESTUFF): Add min-insn-modes.c.
(genobjs): Add genmodes.o.
(print-rtl.o, print-rtl1.o): Depend on $(TM_P_H).
(insn-modes.o): Depend on $(TM_H) not $(GTM_H); also real.h.
(min-insn-modes.c, min-insn-modes.o): New rules.
(s-modes): Also generate min-insn-modes.c.
($(BUILD_PREFIX_1)insn-modes.o): Kill.
* genmodes.c (struct mode_data): Add format field.
(blank_mode, validate_mode, complete_mode): Update to match.
(make_scalar_mode): Separate into make_int_mode and make_float_mode.
(_SCALAR_MODE): Kill.
(FLOAT_MODE, FRACTIONAL_FLOAT_MODE): Add format argument.
(emit_insn_modes_c_header): Adjust.
(emit_min_insn_modes_c_header, emit_real_format_for_mode)
(emit_min_insn_modes_c): New functions.
(emit_insn_modes_c): Call emit_real_format_for_mode.
(main): Add -m option to generate min-insn-modes.c.
* machmode.h: Update documentation. Add format argument to
all uses of FLOAT_MODE.
* real.c: Don't define real_format_for_mode here.
* dwarfout.c: Move default definition of PRINT_REG...
* defaults.h: ...here.
* print-rtl.c: Include tm_p.h.
(DEBUG_PRINT_REG, DEBUG_REGISTER_NAMES, debug_reg_names, reg_names):
Kill.
(print_rtx): Use PRINT_REG, not DEBUG_PRINT_REG. But surround
this entire block with #ifndef GENERATOR_FILE.
* regclass.c: Unconditionally define reg_names.
* config/mips/mips.h, config/rs6000/rs6000.h, config/sh/sh.h
Don't define DEBUG_REGISTER_NAMES.
* config/rs6000/darwin.h: Don't use DEBUG_REGISTER_NAMES in
redefinition of REGISTER_NAMES.
* config/i386/i386.h: Don't define DEBUG_PRINT_REG.
* combine.c: Change all preprocessor conditionals on
EXTRA_CC_MODES to use SELECT_CC_MODE instead; rearrange a bit
for clarity.
* genopinit.c: Remove mention of EXTRA_CC_MODES in comment.
* configure.in: Don't define EXTRA_CC_MODES.
* configure, config.in: Regenerate.
* doc/tm.texi: Remove documentation of EXTRA_CC_MODES.
* config/arc/arc.c, config/m32r/m32r.c, config/sparc/sparc.c:
May assume that GET_MODE_CLASS is accurate for extra CC modes
at all times.
* config/i860/i860.h (INIT_CUMULATIVE_ARGS): Pass correct
number of arguments to aggregate_value_p.
* genmodes.c (RESET_FLOAT_FORMAT, reset_float_format): New.
* machmode.def: Explain ARCH-modes.def. Document
RESET_FLOAT_FORMAT. Improve commentary on various mode
clusters. Do not define OI, PQI, PHI, PSI, PDI, QF, HF, TQF,
XF, or TF modes here. Remove backward-compatibility
definition of CC.
* config/alpha/alpha-modes.def: New file; define TF mode.
* config/arc/arm-modes.def: Define XF mode.
* config/c4x/c4x-modes.def: Define QF and HF modes. Unset
float format for SF and DF modes.
* config/dsp16xx/dsp16xx-modes.def: New file; define HF mode.
* config/i386/i386-modes.def: Define XF and TF modes.
* config/i960/i960-modes.def: Define TF mode.
* config/ia64/ia64-modes.def: Define TF and OI modes.
* config/m68k/m68k-modes.def: New file; define XF mode.
* config/mips/mips-modes.def: New file; define TF mode, reset
formats for SF and DF modes.
* config/pa/pa-modes.def: Define TF mode.
* config/rs6000/rs6000.c: Define TF and PSI modes.
* config/s390/s390-modes.def: Define OI mode.
* config/sh/sh-modes.def: New file; define PSI mode.
* config/sparc/sparc-modes.def: Define TF mode.
* config/vax/vax-modes.def: New file; reset formats for SF and
DF modes.
* config/c4x/c4x.c (c4x_override_options): No need to mess
with real_format_for_mode or set REAL_MODE_FORMATs.
(c4x_immed_int_constant): Don't apply GET_MODE_CLASS to rtx
variable.
* config/i386/i386.c (override_options): No need to set
REAL_MODE_FORMATs here.
* config/i960/i960.c (i960_initialize): Likewise.
* config/m68k/m68k.c (m68k_override_options): Likewise.
* config/ia64/ia64.c (ia64_override_options): Set REAL_MODE_FORMAT
for TFmode only if not the default.
* config/mips/mips.c (override_options): Likewise.
* config/vax/vax.c (override_optionms): Set REAL_MODE_FORMAT for
DFmode only if not the default.
* config/i370/i370.h (RET_REG): Don't consider TFmode.
* config/m68hc11/m68hc11.c (print_operand): Don't consider XFmode.
* config/dsp16xx/dsp16xx.c (hard_regno_mode_ok): #if 0 out use
of modes that don't appear anywhere in the machine description.
* config/arc/arc-modes.def, config/arm/arm-modes.def
* config/c4x/c4x-modes.def, config/frv/frv-modes.def
* config/i386/i386-modes.def, config/i960/i960-modes.def
* config/ia64/ia64-modes.def, config/mmix/mmix-modes.def
* config/pa/pa-modes.def, config/pdp11/pdp11-modes.def
* config/rs6000/rs6000-modes.def, config/s390/s390-modes.def
* config/sparc/sparc-modes.def: Convert to new style for
declaring extra CC modes.
From-SVN: r72440
|
|
* genmodes.c, mode-classes.def: New files.
* machmode.def: Rewritten to genmodes.c interface.
* Makefile.in (extra_modes_file): New substitution variable.
(MACHMODE_H): No longer includes machmode.def or
@extra_modes_file@; instead, mode-classes.def and insn-modes.h.
(BUILD_RTL): Add $(BUILD_PREFIX)insn-modes.o.
(OBJS-common): Add insn-modes.o.
(STAGESTUFF): Add insn-modes.c, insn-modes.h, s-modes, and
genmodes$(build_exeext).
(insn-modes.o, insn-modes.c, insn-modes.h, s-modes, genmodes.o,
genmodes$(build_exeext), $(BUILD_PREFIX_1)insn-modes.o): New targets.
(s-genrtl): Don't depend on $(RTL_BASE_H).
(gengenrtl.o): Don't depend on coretypes.h, $(GTM_H), real.h,
or $(RTL_BASE_H); just rtl.def.
* gengenrtl.c: Don't include coretypes.h, tm.h, rtl.h, or
real.h. Give fake definition of CONST_DOUBLE_FORMAT and
substitute definition of NUM_RTX_CODE. Add casts to avoid
warnings.
* machmode.h: Include insn-modes.h, not machmode.def. Include
mode-classes.def to define enum mode_class. Tweak definitions
of GET_MODE_CLASS, GET_MODE_SIZE, GET_MODE_BITSIZE, GET_MODE_MASK,
GET_MODE_INNER, GET_MODE_WIDER_MODE, GET_CLASS_NARROWEST_MODE.
(inner_mode_array): Renamed mode_inner.
(mode_base_align): New.
* rtl.c (mode_name, mode_class, mode_bitsize, mode_size,
mode_unit_size, mode_wider_mode, mode_mask_array,
inner_mode_array, class_narrowest_mode): Delete definitions.
* stor-layout.c (get_mode_alignment): Use mode_base_align.
* real.h: Use MIN_MODE_FLOAT and MAX_MODE_FLOAT, not QFmode
and TFmode, in real_format_for_mode and REAL_MODE_FORMAT.
* config/ip2k/ip2k.h, config/iq2000/iq2000.h:
No need to define BITS_PER_UNIT.
From-SVN: r72313
|
|
ISO C.
* builtin-attrs.def, builtin-attrs.def, builtins.c, cpplex.c,
cpplib.c, gencheck.c, gengenrtl.c, machmode.def, protoize.c: Don't
use macros from "symcat.h", instead rely on ISO C.
* system.h: Don't include "symcat.h".
* configure.in (AC_C_STRINGIZE): Delete.
* config.in, configure: Regenerate.
From-SVN: r68035
|
|
From-SVN: r60174
|
|
* machmode.def (V1DImode): New mode. A single element vector.
* tree.h (TI_UV1DI_TYPE, TI_V1DI_TYPE): New tree_index enums.
(unsigned_V1DI_type_node, V1D1_type_node): New type nodes.
* tree.c (build_common_tree_nodes_2): Build
unsigned_V1DI_type_node and V1D1_type_node.
* c-common.c (c_common_type_for_mode): Return
unsigned_V1DI_type_node or V1D1_type_node for V1DImode.
* rtl.c (class_narrowest_): Start integer vector nodes with V1DImode.
From-SVN: r57248
|
|
* Makefile.in (MACHMODE_H): Add @extra_modes_file@.
* configure.in: If $srcdir/config/${cpu_type}/${cpu_type}-modes.def
exists, substitute its pathname as @extra_modes_file@, define
EXTRA_MODES_FILE to be an appropriate string to #include it
with, and define EXTRA_CC_MODES to 1.
* machmode.def: Update comments. Include EXTRA_MODES_FILE if
it's defined. Get rid of redundancy in calling sequence for
CC; don't use it to define CCmode, to avoid a warning.
* libgcc2.c: Include symcat.h for the sake of machmode.def.
* arc-modes.def, arm-modes.def, c4x-modes.def, i386-modes.def,
i960-modes.def, ia64-modes.def, m88k-modes.def, mmix-modes.def,
pa-modes.def, pdp11-modes.def, rs6000-modes.def, sparc-modes.def:
New files.
* arc.h, arm.h, c4x.h, i386.h, i960.h, ia64.h, m88k.h, mmix.h,
pa.h, pdp11.h, rs6000.h, sparc.h: Don't define EXTRA_CC_MODES.
* doc/sourcebuild.texi, doc/tm.texi: Document new scheme for
defining extra CC modes.
From-SVN: r54478
|
|
returning 1 byte values)
PR optimization/5742
* machmode.def: Add inner mode field to complex modes.
* config/mips/mips.c (mips_function_value): Always define. Add
new argument to handle libcalls.
* config/mips/mips.h (LIBCALL_VALUE): Use mips_function_value.
(FUNCTION_VALUE): Likewise.
* config/mips/abi64.h (FUNCTION_VALUE): Remove.
* config/mips/mips-protos.h: Update.
From-SVN: r51250
|
|
PR bootstrap/4195
* genrecog.c (maybe_both_true_mode): Remove.
(maybe_both_true_2, write_switch): Revert 2001-07-17 changes.
* machmode.def (Pmode): Likewise.
From-SVN: r51115
|
|
2002-02-19 Aldy Hernandez <aldyh@redhat.com>
* config/i386/i386.md ("mmx_uavgv8qi3"): Use const_vector.
("mmx_uavgv4hi3"): Same.
("pmulhrwv4hi3"): Same.
* tree-inline.c (walk_tree): Handle vectors.
* c-common.c (constant_expression_warning): Handle vectors.
(overflow_warning): Same.
* sched-deps.c (sched_analyze_2): Handle vectors.
* rtlanal.c (rtx_unstable_p): Handle vectors.
(rtx_varies_p): Same.
(count_occurrences): Same.
(regs_set_between_p): Same.
(modified_between_p): Same.
(modified_in_p): Same.
(volatile_insn_p): Same.
(volatile_refs_p): Same.
(side_effects_p): Same.
(may_trap_p): Same.
(inequality_comparisons_p): Same.
(replace_regs): Same.
(computed_jump_p_1): Same.
* rtl.c (DEF_MACHMODE): Change all definitions to accept 8th
argument.
(inner_mode_array): New.
(copy_rtx): Handle vectors.
(copy_most_rtx): Same.
(rtx_equal_p): Same.
(get_mode_alignment): Adjust for vectors.
* resource.c (mark_referenced_resources): Handle vectors.
(mark_set_resources): Same.
* reload1.c (eliminate_regs): Handle vectors.
(elimination_effects): Same.
(scan_paradoxical_subregs): Same.
* reload.c (subst_reg_equivs): Handle vectors.
* regrename.c (scan_rtx): Handle vectors.
* regclass.c (reg_scan_mark_refs): Handle vectors.
* recog.c (find_single_use_1): Handle vectors.
* local-alloc.c (equiv_init_varies_p): Handle vectors.
(contains_replace_regs): Same.
(memref_referenced_p): Same.
* integrate.c (copy_rtx_and_substitute): Handle vectors.
(subst_constants): Same.
* genattrtab.c (attr_copy_rtx): Handle vectors.
(encode_units_mask): Same.
(clear_struct_flag): Same.
(count_sub_rtxs): Same.
* gcse.c (want_to_gcse_p): Handle vectors.
(oprs_unchanged_p): Same.
(hash_expr_1): Same.
(oprs_not_set_p): Same.
(expr_killed_p): Same.
(compute_transp): Same.
(store_ops_ok): Same.
* function.c (purge_addressof_1): Do not allow paradoxical subregs
of vectors.
(fixup_var_refs_1): Same.
(instantiate_virtual_regs_1): Same.
* fold-const.c (operand_equal_p): Handle vectors.
(fold): Same.
(rtl_expr_nonnegative_p): Same.
* flow.c (mark_used_regs): Handle vectors.
* df.c (df_uses_record): Handle vectors.
* cselib.c (cselib_subst_to_values): Handle vectors.
(cselib_mem_conflict_p): Same.
(hash_rtx): Same.
* cse.c (canon_reg): Handle vectors.
(fold_rt): Same.
(cse_process_notes): Same.
(count_reg_usage): Same.
(canon_hash): Same.
* alias.c (nonlocal_mentioned_p): Add case for CONST_VECTOR.
* combine.c (mark_used_regs_combine): Add case for CONST_VECTOR.
* emit-rtl.c (init_emit_once): Generate const0_rtx for vectors.
(gen_rtx): Handle CONST_VECTOR.
(gen_const_vector_0): New.
(copy_rtx_if_shared): CONST_VECTORs can be shared.
(reset_used_flags): Same.
(copy_insn_1): Same.
(initializer_constant_valid_p): Handle VECTOR_CST.
* doc/c-tree.texi (Expression trees): Document VECTOR_CST.
* doc/rtl.texi (Constants): Document const_vector.
(CONST0_RTX): Update for vectors.
(RTL sharing): Same.
* print-tree.c (print_node): Add case for VECTOR_CST.
* tree.h (TREE_VECTOR_CST_ELTS): New.
(struct tree_vector): New.
(union tree_node): Add vector node.
(build_vector): Add prototype.
* tree.def (VECTOR_CST): New.
* tree.c (build_vector): New.
* expmed.c (make_tree): Handle CONST_VECTOR.
* rtl.h (CONSTANT_P): CONST_VECTORs are constants too.
(CONST_VECTOR_ELT): New.
(CONST_VECTOR_NUNITS): New.
* machmode.h (GET_MODE_INNER): New.
(DEF_MACHMODE): Accept 8th arg.
* machmode.def: Add 8th argument for vector inner mode.
Add inner vector modes for vectors.
* rtl.def (VEC_CONST): Remove.
(CONST_VECTOR): New.
* expr.c (clear_storage): Allow vectors.
(is_zeros_p): Handle VECTOR_CST.
* varasm.c (output_constant_pool): Handle vectors.
(rtx_const): Add veclo and vechi fields.
(kind): Add RTX_VECTOR.
(decode_rtx_const): Add case for vector.
* config/rs6000/rs6000-protos.h: Add zero_constant.
* config/rs6000/rs6000.c (rs6000_emit_move): Handle vector
constants. Force easy vector constants into memory.
(easy_vector_constant): New.
(emit_easy_vector_constant): New.
(rs6000_legitimize_reload_address): Do not generate bad reloads on
darwin.
* config/rs6000/rs6000.md ("altivec_lvx"): Reflect what
instruction does.
("altivec_lvxl"): Same.
(altivec_lvebx): Same.
(altivec_lvehx): Same.
(altivec_lvewx): Same.
("*movv4si_const0"): New.
("*movv4sf_const0"): New.
("*movv8hi_const0"): New.
("*movv16qi_const0"): New.
From-SVN: r49853
|
|
2002-02-09 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (TARGET_CANNOT_MODIFY_JUMPS_P): Define to...
(sh_cannot_modify_jumps_p): New function.
2002-02-05 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (TARGET_MS_BITFIELD_LAYOUT_P): Define to...
(sh_ms_bitfield_layout_p): New function.
2002-02-04 Alexandre Oliva <aoliva@redhat.com>
Zack Weinberg <zack@codesourcery.com>
* config/sh/sh.h (TRAMPOLINE_ADJUST_ADDRESS): Use
expand_simple_binop instead of expand_binop.
2002-02-03 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (OVERRIDE_OPTIONS) [! TARGET_SH5]: Disable
use of .quad and .uaquad.
* config/sh/sh.c (TARGET_ASM_UNALIGNED_DI_OP,
TARGET_ASM_ALIGNED_DI_OP): Add comment pointing to the above.
2002-01-24 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (movdi_const, movdi_const_32bit,
movdi_const_16bit): Make sure all CONSTs have modes.
(sym2PIC): Ditto, but by adjusting all callers.
* config/sh/sh.c (calc_live_regs) [TARGET_SHCOMPACT]: Set pr_live
if the prologue calls the SHmedia argument decoder or register
saver.
2002-01-24 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (TARGET_ASM_UNALIGNED_DI_OP): Define.
(TARGET_ASM_ALIGNED_DI_OP): Likewise.
(sh_expand_epilogue): Don't emit USE of return target register.
(prepare_move_operands): Legitimize DImode PIC addresses.
(sh_media_register_for_return): Skip tr0, used to initialize the
PIC register.
(sh_expand_prologue): Remove explicit USE of return register.
(nonpic_symbol_mentioned_p): PC is non-PIC. Don't recurse in
CONST_DOUBLEs. UNSPEC_GOTPLT is PIC.
* config/sh/sh.h (ASM_OUTPUT_DOUBLE_INT): Removed, obsolete.
(OVERRIDE_OPTIONS): Don't disable PIC on SH5.
(EXTRA_CONSTRAINT_S): Use MOVI_SHORI_BASE_OPERAND_P instead of
EXTRA_CONSTRAINT_T.
(GOT_ENTRY_P, GOTPLT_ENTRY_P, GOTOFF_P, PIC_ADDR_P): New.
(MOVI_SHORI_BASE_OPERAND_P): New.
(NON_PIC_REFERENCE_P, PIC_REFERENCE_P): New.
(EXTRA_CONSTRAINT_T): Define in terms of them.
(OUTPUT_ADDR_CONST_EXTRA): Handle UNSPEC_GOTPLT.
* config/sh/sh.md (movsi_media, movsi_media_nofpu,
movdi_media, movdi_media_nofpu): Add SIBCALL_REGS class to
alternatives supporting TARGET_REGS.
(UNSPEC_GOTPLT): New constant.
(movdi split): Move incrementing of LABEL_NUSES...
(movdi_const, movdi_const_32bit): Here. Use
MOVI_SHORI_BASE_OPERAND_P instead of EXTRA_CONSTRAINT_T.
(movdi_const_16bit): New.
(call, call_value) [flag_pic]: Use GOTPLT.
(call_pop, call_value_pop): New expands.
(call_pop_compact, call_pop_rettramp): New insns.
(call_value_pop_compact, call_value_pop_rettramp): New insns.
(sibcall) [flag_pic]: Use GOT.
(builtint_setjmp_receiver): Remove bogus, unused expand.
(GOTaddr2picreg): Implement for SHcompact and SHmedia.
(*pt, *ptb, ptrel): New insns.
(sym2GOT): Handle DImode GOT.
(sym2GOTPLT, symGOTPLT2reg): New expands.
(sym2PIC): New expand.
(shcompact_return_tramp): Use GOTPLT to return trampoline.
(shcompact_return_tramp_i): Use return register explicitly.
* config/sh/sh.h (OVERRIDE_OPTIONS) [TARGET_SHMEDIA]: Don't
disable flag_reorder_blocks.
2002-01-19 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (sibcall_compact): Reorder return, uses and
clobbers, for clarity.
(sibcall_epilogue) [TARGET_SHCOMPACT]: Mark saving and
restoring of r0 in macl as MAYBE_DEAD.
2002-01-18 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (LONG_DOUBLE_TYPE_SIZE): Define.
* config/sh/sh.md (movv4sf_i, movv16sf_i): Fix uses of
alter_subreg all over.
(jump) [TARGET_SHMEDIA]: FAIL to create new jumps after
reload, instead of emitting instructions that would require
reloading.
(casesi_load_media): Add missing modes.
2001-11-09 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (sh_expand_prologue): Mark the PIC register
as used if the argument decoder is called.
2001-08-28 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (udivsi3, divsi3): Load libcall symbol name in
Pmode, then extend it to DImode if necessary.
2001-08-28 Stephen Clarke <Stephen.Clarke@st.com>
* config/sh/sh.h (LEGITIMATE_CONSTANT_P): Don't accept DFmode
constants in FPU-enabled SHmedia, let them be loaded from memory.
2001-08-28 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (cmpeqdi_media, cmpgtdi_media, cmpgtudi_media):
Adjust whitespace in assembly output templates.
2001-08-28 Stephen Clarke <Stephen.Clarke@st.com>
* config/sh/sh.md (movdicc_false, movdicc_true, movdicc): Adjust
mode of if_then_else.
2001-08-04 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh64.h (CPP_DEFAULT_CPU_SPEC): Override definition in
sh.h.
2001-07-26 Andrew Haley <aph@cambridge.redhat.com>
Joern Rennecke <amylaar@redhat.com>
* config/sh/sh64.h (CPP_DEFAULT_CPU_SPEC): New.
(SUBTARGET_CPP_PTR_SPEC): New.
(SUBTARGET_CPP_SPEC): Remove.
2001-07-06 Chandrakala Chavva <cchavva@redhat.com>
* config/sh/sh.md (movsf_media_nofpu+1, movdf_media_nofpu+1):
Fix typo in previous checkin.
2001-07-11 Chandrakala Chavva <cchavva@redhat.com>
* config/sh/sh.h (MODES_TIEABLE_P): Fix redact indentations.
2001-07-10 Chandrakala Chavva <cchavva@cygnus.com>
Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (MODES_TIEABLE_P): Don't tie modes wider than
what single FP register can hold for SHmedia target.
2001-07-06 Chandrakala Chavva <cchavva@redhat.com>
Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (movsf_media_nofpu+1, movdf_media_nofpu+1):
Do not split into SUBREG.
2001-06-14 Alexandre Oliva <aoliva@redhat.com>
* config/sh/ushmedia.h, config/sh/sshmedia.h: Updated signatures
and added new functions as specified in SH5 ABI r9.
2001-06-04 Alexandre Oliva <aoliva@redhat.com>
* config/sh/lib1funcs.asm (GCC_nested_trampoline): Align to an
8-byte boundary.
2001-06-03 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (dump_table): Add const0_rtx in calls of
gen_consttable_4 and gen_consttable_8. Emit multiple labels
and consttable_window_ends.
2001-06-03 Graham Stott <grahams@redhat,com>
* config/sh/sh.md (movdi split): Remove unused variable last_insn.
2001-05-16 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (print_operand): Handle floating-point pair,
vector and matrix registers.
* config/sh/sh.h (REGISTER_MOVE_COST): Take floating-pointer
vector modes into account.
* config/sh/sh.md (movv2sf): Split move between registers into
movdf.
(movv4sf, movv16sf): Introduce insns that get split only after
reload.
* config/sh/shmedia.h: Fix Copyright dates.
* config/sh/ushmedia.h: Likewise. Move loop counter
declarations into conditionals that uses them.
(sh_media_FVADD_S, sh_media_FVSUB_S): Fix off-by-one error in
loop boundary.
* config/sh/sshmedia.h: Fix Copyright dates.
(sh_media_PUTCFG): Fix constraints.
2001-05-12 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (TARGET_PTRMEMFUNC_VBIT_LOCATION): Define to
ptrmemfunc_vbit_in_delta for SH5.
2001-05-08 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (TARGET_SWITCHES): Document -m5-*.
* invoke.texi: Likewise.
2001-04-14 Alexandre Oliva <aoliva@redhat.com>
* config/sh/lib1funcs.asm (GCC_push_shmedia_regs,
GCC_push_shmedia_regs_nofpu, GCC_pop_shmedia_regs,
GCC_pop_shmedia_regs_nofpu): New global symbols.
* config/sh/t-sh64 (LIB1ASMFUNCS): Add them.
* config/sh/sh.h (SHMEDIA_REGS_STACK_ADJUST): New macro.
* config/sh/sh.c (calc_live_regs): Account for PR's saving in
compact function with nonlocal labels.
(sh_expand_prologue) [SHcompact]: Push SHmedia regs if needed.
(sh_expand_epilogue) [SHcompact]: Pop them when appropriate.
(initial_elimination_offset): Account for their stack space.
* config/sh/sh.md (shmedia_save_restore_regs_compact): New insn.
* config/sh/sh.md (movsi_media, movsi_media_nofpu, movqi_media,
movhi_media, movdi_media, movdi_media_nofpu, movdf_media,
movdf_media_nofpu, movsf_media, movsf_media_nofpu): Require at
least one of the operands to be a register.
(movv2sf): Likewise. Renamed to movv2sf_i.
(movdi, movdf, movv2sf, movv4sf, movv16sf, movsf):
prepare_move_operands() before emitting SHmedia insns.
2001-04-03 Alexandre Oliva <aoliva@redhat.com>
* config/sh/crti.asm (init, fini) [__SH5__ && ! __SHMEDIA__]:
Don't save nor initialize r12. Don't mis-align the stack.
Pad the code with a nop.
* config/sh/crti.asm: Don't restore r12. Don't mis-align the
stack.
2001-03-13 Alexandre Oliva <aoliva@redhat.com>
* gcc/longlong.h (__umulsidi3, count_leading_zeros)
[__SHMEDIA__]: Implement.
2001-03-11 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md: Set latency of `pt' closer to reality.
(movsi_media, movsi_media_nofpu, movdi_media, movdi_media_nofpu,
movdf_media, movdf_media_nofpu, movsf_media, movsf_media_nofpu):
Set move, load and store type attributes.
* config/sh/sh.c (sh_loop_align) [TARGET_SH5]: Set to 3.
* config/sh/sh.h (OVERRIDE_OPTIONS) [TARGET_SH5]: Disable
profiling.
* config/sh/sh.h (PROMOTE_MODE): Sign-extend SImode to DImode.
* config/sh/sh-protos.h (sh_media_register_for_return): Declare.
* config/sh/sh.c (sh_media_register_for_return): New function.
(sh_expand_prologue) [TARGET_SHMEDIA]: Copy r18 to an available
branch-target register.
(sh_expand_epilogue) [TARGET_SHMEDIA]: Explicitly USE it.
* config/sh/sh.md (return_media_i): Use any call-clobbered
branch-target register.
(return_media): If r18 wasn't copied in the prologue, copy it
here.
* config/sh/sh.h (CONDITIONAL_REGISTER_USAGE) [TARGET_SHMEDIA]:
Clear class FP0_REGS.
* config/sh/sh64.h (LINK_SPEC): Removed incorrect default copied
from elf.h.
2001-03-08 DJ Delorie <dj@redhat.com>
* config/sh/sh.h (OVERRIDE_OPTIONS): Disable relaxing for SHMEDIA.
2001-02-09 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (sibcall_compact): Set fp_mode to single.
2001-02-07 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (INT_ASM_OP) [SHMEDIA64]: Use `.quad'.
2001-02-03 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (INIT_CUMULATIVE_ARGS): Compute size of BLKmode
return value correctly for call_cookie.
2001-02-01 Alexandre Oliva <aoliva@redhat.com>
* config/sh/crt1.asm (start): Modified so as to call
___setup_argv_and_call_main.
2001-01-26 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (FUNCTION_ARG_ADVANCE): Don't count stack_regs in
SHmedia mode.
2001-01-20 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (STRIP_DATALABEL_ENCODING): New macro.
(STRIP_NAME_ENCODING): Use it.
(ASM_OUTPUT_LABELREF): Likewise. Don't call assemble_name().
2001-01-19 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (sgeu) [! SHMEDIA]: Fix invocation of
prepare_scc_operands().
* config/sh/sh.h (SH_DATALABEL_ENCODING): Change to "#"...
(DATALABEL_SYMNAME_P): ... so that we don't need memcmp here.
2001-01-17 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (STRIP_NAME_ENCODING): Strip leading `*'.
2001-01-13 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (shcompact_incoming_args): Use R0_REG.
* config/sh/sh.md (R7_REG, R8_REG, R9_REG): Define as constants,
used in shcompact_incoming_args.
* config/sh/sh.c (sh_expand_epilogue): Fix thinko in previous
change.
* config/sh/crt1.asm (start) [SH5]: Switch to single-precision
mode.
* config/sh/lib1funcs.asm (sdivsi3_i4, udivsi3_i4, set_fpscr):
Adjust accordingly.
* config/sh/sh.c (sh_expand_prologue, sh_expand_epilogue):
Simplify. Adjust. Add sanity check.
* config/sh/sh.h (TARGET_SWITCHES) [5-compact]: Set
FPU_SINGLE_BIT.
* config/sh/sh.md (udivsi3_i4_single, divsi3_i4_single): Match
TARGET_SHCOMPACT.
(udivsi3, divsi3): Use them.
(force_mode_for_call): New insn.
(call, call_value, sibcall_value): Emit it before SHcompact
calls.
2001-01-11 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (call, call_value, sibcall): Make sure the
call cookie is non-NULL before taking its value.
2001-01-10 Alexandre Oliva <aoliva@redhat.com>
* config.gcc (sh64): Set target_requires_64bit_host_wide_int.
2001-01-09 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (shcompact_incoming_args): Set argument memory
block.
* config/sh/sh.h (STATIC_CHAIN_REGNUM) [SH5]: Use r1.
* config/sh/sh.c (sh_expand_prologue) [SH5]: Use r0 as
temporary for stack adjusts. Use MACL and MACH to pass
arguments to shcompact_incoming_args.
* config/sh/sh.md (shcompact_incoming_args): Adjust. Don't
clobber r1.
* config/sh/lib1funcs.asm (shcompact_incoming_args): Likewise.
(nested_trampoline): Load static chain address into r1.
* config/sh/sh.md (movdi_media splits): Fix sign-extension.
2001-01-07 Alexandre Oliva <aoliva@redhat.com
* config/sh/sh.c (fpul_operand) [SHMEDIA]: Just call
fp_arith_reg_operand().
2001-01-06 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (casesi): Sign-extend the first two operands,
and use signed compares for them.
* config/sh/sh.c (dump_table): Don't emit 8-byte constants after
4-byte ones. Instead, inter-leave them, maintaining the 8-byte
ones properly aligned.
(find_barrier): Account for extra alignment needed for 8-byte wide
constants.
(machine_dependent_reorg): Require a label for the second 4-byte
constant after an 8-byte one.
* config/sh/lib1funcs.asm (sdivsi3): Fix typo in yesterday's
change.
2001-01-05 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (machine_dependent_reorg) [SHCOMPACT]: Reset
last_float when switching float modes.
* config/sh/sh.md (movdf) [SH5]: Don't use stack-pointer
auto-increment for general-purpose registers.
* config/sh/lib1funcs.asm (sdivsi3) [SHMEDIA]: Sign-extend the
result.
* config/sh/sh.c (sh_expand_prologue) [SH5]: Use r1 as temporary
for stack adjust.
* config/sh/sh.c (sh_builtin_saveregs): Support using all
registers for varargs.
2001-01-01 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (FUNCTION_ARG_ADVANCE): Simplify.
* config/sh/sh.h (CALL_COOKIE_STACKSEQ,
CALL_COOKIE_STACKSEQ_SHIFT, CALL_COOKIE_STACKSEQ_GET): New macros.
(CALL_COOKIE_INT_REG_SHIFT): Adjust.
(FUNCTION_ARG_ADVANCE): Use SHCOMPACT_FORCE_ON_STACK. Adjust
call_cookie accordingly.
(FUNCTION_ARG): Test SHCOMPACT_FORCE_ON_STACK.
(SHCOMPACT_BYREF): Likewise.
(SHCOMPACT_FORCE_ON_STACK): New macro.
* config/sh/sh.c (sh_expand_prologue): Use new call_cookie format.
(sh_builtin_saveregs): Likewise.
* config/sh/lib1funcs.asm (shcompact_call_trampoline,
shcompact_incoming_args): Use new shift values. Support
sequences of consecutive and non-consecutive pushes/pops.
* config/sh/sh.md (return): Don't explicitly use PR_REG.
2001-01-05 Hans-Peter Nilsson <hpn@cygnus.com>
* config/sh/sh.h (TEXT_SECTION): Define.
* config/sh/elf.h (ASM_FILE_START): Output TEXT_SECTION_ASM_OP.
2001-01-05 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (INIT_CUMULATIVE_LIBCALL_ARGS): New macro.
* config/sh/sh.h (BASE_RETURN_VALUE_REG): Use FP regs for
return values on FPU-enabled SHmedia.
(FUNCTION_VALUE_REGNO_P): Mark FIRST_FP_RET_REG as used on
FPU-enabled SHmedia.
(INIT_CUMULATIVE_ARGS): Set up return trampoline only if
value is returned in a non-FP reg and is not returned by
reference.
* config/sh/sh.md (shcompact_return_tramp_i): Change type to
jump_ind.
2000-01-04 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (SH_MIN_ALIGN_FOR_CALLEE_COPY): New.
(FUNCTION_ARG_CALLEE_COPIES): Require argument to be
quad-aligned to be passed by callee-copy reference.
2001-01-03 Alexandre Oliva <aoliva@redhat.com>
* config/sh/elf.h (MAX_WCHAR_TYPE_SIZE): Define.
* config/sh/sh64.h (MAX_WCHAR_TYPE_SIZE): Undefine.
2001-01-02 Alexandre Oliva <aoliva@redhat.com>
* config/sh/lib1funcs.asm (shcompact_call_trampoline): Fix error in
copying low-numbered FP regs to r7 and r8.
* config/sh/sh.h (FUNCTION_ARG_ADVANCE): Don't request copying of
FP regs to general-purpose regs only if the copy was passed on the
stack.
* config/sh/lib1funcs.asm (shcompact_call_trampoline): Fix typo in
copying FP reg to r9.
* config/sh/sh.h (FUNCTION_ARG_ADVANCE): Use trampoline to
copy FP regs to general-purpose regs only in outgoing calls.
* config/sh/sh.md (movdf_media, movsf_media): Revert incorrect
change from 2000-10-30. Adjust for 64-bit (or 32-bit)
HOST_WIDE_INT.
* config/sh/sh.h (struct sh_args): Document all fields.
(FUNCTION_OK_FOR_SIBCALL): Functions that receive arguments
passed partially on the stack should not consider making
sibcalls.
* config/sh/sh.h (FUNCTION_ARG_ADVANCE): Add byref regs to
stack_regs only for incoming calls. When passing FP args,
make sure there are FP regs available before modifying
call_cookie.
(SHCOMPACT_BYREF): Pass double args in general-purpose
registers by reference.
2000-12-30 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (FUNCTION_OK_FOR_SIBCALL) [SHCOMPACT]: Don't
attempt to generate sibcalls if the caller got any arguments
by reference.
* config/sh/lib1funcs.asm (set_fpscr) [SH5]: Default to double.
* config/sh/sh.c (dump_table) [SHCOMPACT]: Align DImode and DFmode
to 8-byte boundaries.
* config/sh/sh.md (shcompact_preserve_incoming_args): New insn.
* config/sh/sh.h (CALL_COOKIE_INT_REG_GET): New macro.
* config/sh/sh.c (sh_expand_prologue): Preserve args that will be
stored in the stack.
* config/sh/lib1funcs.asm (ct_main_table, ia_main_table): Arrange
for the offsets to have the ISA bit set.
(shcompact_call_trampoline): Document. Swap r0 and r1, to match
invocation. Use beq instead of bgt to mark end of sequence of
loads.
(shcompact_incoming_args): Fix store of r2. Use beq instead of
bgt to mark end of sequence of stores.
* config/sh/sh.c (arith_operand): Don't check whether
CONST_OK_FOR_J for now.
* config/sh/sh.md (movdf_media, movsf_media): Use HOST_WIDE_INT
instead of long for conversion.
2000-12-29 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.c (print_operand_address): Convert INTVAL to int
before passing it to fprintf.
2000-12-28 Alexandre Oliva <aoliva@redhat.com>
* config/sh/crt1.asm (start): Reset SR.FD, to enable the FP unit.
Call set_fpscr before reading/writing SR.
* config/sh/crt1.asm (start): Set SR.SZ and SR.PR, but not SR.FR.
Call set_fpscr.
* config/sh/lib1funcs.asm: Add `.align 2' directives before
SHmedia code.
(FMOVD_WORKS): Define on SH5 with FPU.
(set_fpscr): Define on SH5. Remove separate _fpscr_values
setting.
* config/sh/t-sh64 (LIB1ASMFUNCS): Add _set_fpscr instead of
_fpscr_values.
2000-12-28 Hans-Peter Nilsson <hpn@cygnus.com>
* config/sh/lib1funcs.asm (ct_main_table): Align contents to even
address.
(ia_main_table): Ditto.
2000-12-27 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.h (MAX_WCHAR_TYPE_SIZE): Don't define.
* config/sh/sh64.h (WCHAR_TYPE, WCHAR_TYPE_SIZE): Reinstate
the definitions from sh.h.
* config/sh/sh.h (PTRDIFF_TYPE): Define as conditional on
TARGET_SH5.
(SUBTARGET_CPP_SPEC): Arrange for __PTRDIFF_TYPE__ to be defined.
* config/sh/elf.h (PTRDIFF_TYPE): Likewise.
* config/sh/sh64.h (SUBTARGET_CPP_SPEC): Likewise.
2000-12-26 Alexandre Oliva <aoliva@redhat.com>
* config/sh/sh.md (movdi_media split): Don't add REG_LABEL notes.
Increment LABEL_NUSES.
From-SVN: r49630
|
|
2001-11-03 Aldy Hernandez <aldyh@redhat.com>
* machmode.def: Add documentation for the seventh argument in
vector modes.
From-SVN: r46753
|
|
* builtins.c (apply_args_size): Handle vector arguments.
(apply_result_size): Same.
* machmode.def: Set WIDER mode field for all vector entries.
From-SVN: r46718
|
|
* Makefile.in, alias.c, basic-block.h, bb-reorder.c, bitmap.c,
bitmap.h, builtin-types.def, builtins.c, builtins.def,
c-aux-info.c, c-common.c, c-common.def, c-common.h,
c-convert.c, c-decl.c, c-dump.c, c-dump.h, c-errors.c,
c-format.c, c-lang.c, c-lex.c, c-lex.h, c-parse.in,
c-pragma.c, c-pragma.h, c-semantics.c, c-tree.h, c-typeck.c,
caller-save.c, calls.c, collect2.c, collect2.h, combine.c,
conditions.h, config.gcc, configure.frag, configure.in,
conflict.c, convert.c, convert.h, cppspec.c, crtstuff.c,
cse.c, cselib.c, cselib.h, dbxout.c, dbxout.h, defaults.h,
dependence.c, df.c, df.h, diagnostic.c, diagnostic.h,
doloop.c, dominance.c, dwarf.h, dwarf2.h, dwarf2asm.c,
dwarf2asm.h, dwarf2out.c, dwarf2out.h, dwarfout.c,
emit-rtl.c, errors.c, errors.h, except.c, except.h,
exgettext, explow.c, expmed.c, expr.c, expr.h, final.c,
fixproto, flags.h, flow.c, fold-const.c, fp-test.c,
function.c, function.h, gbl-ctors.h, gcc.c, gcc.h, gcc.hlp,
gccspec.c, gcov-io.h, gcse.c, genattr.c, genattrtab.c,
gencheck.c, gencodes.c, genconfig.c, genemit.c,
genextract.c, genflags.c, gengenrtl.c, genmultilib,
genopinit.c, genoutput.c, genpeep.c, genrecog.c,
gensupport.c, gensupport.h, ggc-callbacks.c, ggc-common.c,
ggc-none.c, ggc-page.c, ggc-simple.c, ggc.h, global.c,
graph.c, graph.h, gthr-aix.h, gthr-dce.h, gthr-posix.h,
gthr-rtems.h, gthr-single.h, gthr-solaris.h, gthr-vxworks.h,
gthr-win32.h, gthr.h, haifa-sched.c, halfpic.c, halfpic.h,
hard-reg-set.h, hwint.h, ifcvt.c, input.h, insn-addr.h,
integrate.c, integrate.h, jump.c, lcm.c, libgcc2.c,
libgcc2.h, lists.c, local-alloc.c, loop.c, loop.h,
machmode.def, machmode.h, main.c, mbchar.c, mbchar.h,
mips-tdump.c, mips-tfile.c, mklibgcc.in, mkmap-flat.awk,
mkmap-symver.awk, optabs.c, output.h, params.c, params.def,
params.h, predict.c, predict.def, predict.h, prefix.c,
prefix.h, print-rtl.c, print-tree.c, profile.c, protoize.c,
read-rtl.c, real.c, real.h, recog.c, recog.h, reg-stack.c,
regclass.c, regmove.c, regrename.c, regs.h, reload.c,
reload.h, reload1.c, reorg.c, resource.c, resource.h, rtl.c,
rtl.def, rtl.h, rtlanal.c, sbitmap.c, sbitmap.h,
sched-deps.c, sched-ebb.c, sched-int.h, sched-rgn.c,
sched-vis.c, sdbout.c, sdbout.h, sibcall.c, simplify-rtx.c,
ssa-ccp.c, ssa-dce.c, ssa.c, ssa.h, stmt.c, stor-layout.c,
stringpool.c, system.h, timevar.c, timevar.def, timevar.h,
tlink.c, toplev.c, toplev.h, tree.c, tree.def, tree.h,
tsystem.h, unroll.c, unwind-dw2-fde.c, unwind-dw2-fde.h,
unwind-dw2.c, unwind-pe.h, unwind-sjlj.c, unwind.h,
unwind.inc, varasm.c, varray.c, varray.h, xcoffout.c,
xcoffout.h: replace "GNU CC" with "GCC".
From-SVN: r45105
|
|
* machmode.def (Pmode): Redefine if GENERATOR_FILE.
* genrecog.c (maybe_both_true_mode): New.
(maybe_both_true_2): Use it.
(write_switch): Don't put Pmode in a switch.
* rtl.c (mode arrays): Don't explicitly size them.
From-SVN: r44029
|
|
BITS_PER_UNIT != 8.
* machmode.def: Use correct definition for BITSIZE on targets where
BITS_PER_UNIT != 8.
From-SVN: r38085
|
|
From-SVN: r37198
|
|
From-SVN: r36886
|
|
* machmode.def: Add BImode. Add a column for bitsize.
* machmode.h (DEF_MACHMODE): Adjust for extra column.
(GET_MODE_BITSIZE): Use it.
* rtl.c (DEF_MACHMODE): Adjust for extra column.
(mode_bitsize): New.
(mode_mask_array): Use bitsize.
* combine.c (combine_simplify_rtx): Require inner and outer
modes to match on nonzero_bits optimizations.
From-SVN: r36501
|
|
From-SVN: r34677
|
|
CONST_DOUBLE...
1999-08-27 13:27 -0700 Zack Weinberg <zack@bitmover.com>
* rtl.c: Define CONST_DOUBLE_FORMAT to the appropriate format
for a CONST_DOUBLE, at compile time. Initialize rtx_length
and class_narrowest_mode at compile time. Kill init_rtl.
Mark rtx_length, mode_class, mode_size, mode_unit_size,
mode_wider_mode, mode_mask_array, class_narrowest_mode, and
rtx_format as const. Kill all references to EXTRA_CC_MODES or
EXTRA_CC_NAMES.
* rtl.def (CONST_DOUBLE): Use CONST_DOUBLE_FORMAT macro for
format.
* rtl.h: Declare rtx_length and rtx_format as const.
* machmode.def: Define CC(). Use CC() to define CCmode. If
EXTRA_CC_MODES is defined, expand it here.
* machmode.h: Declare mode_class, mode_size, mode_unit_size,
mode_wider_mode, mode_mask_array, and class_narrowest_mode as
const. Kill all references to EXTRA_CC_MODES.
* toplev.c: Don't prototype or call init_rtl.
* optabs.c: Don't call init_mov_optab.
* genemit.c: Don't generate init_mov_optab. Don't call
init_rtl.
* gengenrtl.c: Duplicate calculation of CONST_DOUBLE_FORMAT
here.
* genattr.c, genattrtab.c, gencodes.c, genconfig.c,
genextract.c, genflags.c, genopinit.c, genoutput.c, genpeep.c,
genrecog.c: Don't call init_rtl.
* arc.h, arm.h, c4x.h, i386.h, i960.h, m88k.h, pa.h, pdp11.h,
rs6000.h, sparc.h: Don't define EXTRA_CC_NAMES. Use CC() in
definition of EXTRA_CC_MODES.
* md.texi: Kill ref to EXTRA_CC_NAMES.
* tm.texi: Document new way to define EXTRA_CC_MODES.
* genrecog.c: Do not look up the name of a define_split.
(Unrelated bugfix.)
From-SVN: r28937
|
|
From-SVN: r19601
|
|
for details.
* haifa-sched.c: Mirror recent changes from gcc2.
From-SVN: r18984
|
|
From-SVN: r18980
|
|
From-SVN: r9961
|
|
From-SVN: r8160
|
|
From-SVN: r7732
|
|
From-SVN: r4803
|
|
From-SVN: r3643
|
|
From-SVN: r3543
|
|
From-SVN: r3128
|
|
From-SVN: r26
|