* fold-const.c (fold): An equality comparison of a non-weak object
against zero has a known result. Similarly an equality comparison
of the address of two non-weak, unaliased symbols has a known result.
* ggc-page.c (struct page_entry): New field PREV.
(ggc_alloc): Update PREV field appropriately.
(sweep_pages): Likewise.
(ggc_free): Likewise. Use PREV field rather than loop to
improve ggc_free performance.
cp/
* init.c (build_vec_delete_1): Convert 2nd argument to NE_EXPR to
the proper type.
From-SVN: r78713
* fold-const.c (fold): Revert 2004-02-25 change. Use the original
operands to build a tree with swapped operands.
* expr.c (expand_expr_real) <MAX_EXPR>: Consistently use the
'unsignedp' predicate to specify the signedness.
Co-Authored-By: Roger Sayle <roger@eyesopen.com>
From-SVN: r78558
PR c/14092
* gcc.dg/pr14092-1.c: New testcase.
* fold-const.c (fold) <NEGATE_EXPR>: Convert result of
negate_expr back to the original type.
From-SVN: r77602
* fold-const.c (fold) <NOP_EXPR>: Use the original type conversion
tree code rather than call fold_convert, which doesn't specify a
default floating point to integer conversion.
* gcc.c-torture/compile/20040209-1.c: New test case.
From-SVN: r77535
PR middle-end/13696
* fold-const.c (fold_convert): New function to provide type
conversion to the middle-end without using convert.
(negate_expr, associate_trees, size_diffop, omit_one_operand,
operand_equal_for_comparison_p, pedantic_omit_one_operand,
invert_truthvalue, optimize_bit_field_compare, range_binop,
decode_field_reference, make_range, build_range_check, unextend,
fold_truthop, extract_muldiv_1, fold_mathfn_compare,
fold_binary_op_with_conditional_arg, fold_inf_compare,
fold_single_bit_test, fold, multiple_of_p): Replace all calls to
convert with calls to fold_convert.
* g++.dg/opt/fold1.C: New test case.
From-SVN: r77464
* fold-const.c (negate_expr_p, negate_expr): Optimize -(A+B) into
either (-A)-B or (-B)-A, if A or B is easily negated respectively.
(fold) <MINUS_EXPR>: Optimize (A*C) - (B*C) -> (A-B)*C for both
integer types and floating point with unsafe_math_optimizations.
Add similar optimization for (A*C1) - (A*C2) -> A*(C1-C2).
Optimize A - B as A + (-B), if B is easily negated.
From-SVN: r77453
* tree.c (commutative_tree_code, associative_tree_code): New
functions.
(iterative_hash_expr): Use commutative_tree_code.
* tree.h (commutative_tree_code, associative_tree_code): Declare.
* fold-const.c (operand_equal_p): Use commutative_tree_code
rather than inlining the commutativity check.
(fold): Likewise.
Co-Authored-By: Roger Sayle <roger@eyesopen.com>
From-SVN: r77152
* fold-const.c (fold_binary_op_with_conditional_arg): Only
build a COMPOUND_EXPR if 'arg' is really a SAVE_EXPR.
Co-Authored-By: Olivier Hainque <hainque@act-europe.fr>
From-SVN: r76434
* fold-const.c (fold_convert): Rename to fold_convert_const.
(fold_convert_const): Change arguments to take a tree_code,
a type and the operand/expression to be converted. Return
NULL_TREE if no simplification is possible. Add support for
FIX_CEIL_EXPR and FIX_FLOOR_EXPR in addition to FIX_TRUNC_EXPR.
(fold): Handle FIX_CEIL_EXPR and FIX_FLOOR_EXPR.
Adjust call to fold_convert to match new fold_convert_const.
Avoid modifying the tree passed to fold in-place.
From-SVN: r76237
* fold-const.c (fold) <COND_EXPR>: Don't require strict type
equality, instead just prevent replacing a COND_EXPR of non-void
type by one of its operands of void type.
From-SVN: r75239
PR fortran/12632
* fold-const.c (fold) <COND_EXPR>: Don't fold a constant condition,
if the type of the selected branch doesn't match its' parent.
* com.c (ffecom_subscript_check_): Take as an extra argument the
(possibly NULL) decl of the array. Don't create unnecessary tree
nodes if the array index is known to be safe at compile-time.
If the array index is unsafe, force the array decl into memory to
avoid RTL expansion problems.
(ffecom_array_ref_): Update calls to ffecom_subscript_check_.
(ffecom_char_args_x_): Likewise.
* g77.dg/12632.f: New test case.
From-SVN: r75203
* fold-const.c (negate_mathfn_p): New function to determine whether
a built-in mathematical function is sign preserving, f(-x) == -f(x).
Add support for BUILT_IN_ASIN, BUILT_IN_ASINF and BUILT_IN_ASINL.
(tree_swap_operands_p): Change API to take an additional argument
indicating that the swapped operands evaluate in reverse order.
Canonicalize VAR_DECLs and PARM_DECLs last if we can, i.e. neither
operand side-effects or we don't care about flag_evaluation_order.
(reorder_operands_p): New function to check whether its safe to
evaluate the given operands in reverse order.
(negate_expr_p): We can always negate integer constants unless
we honor -ftrapv and the signed type would overflow. Only allow
-(A-B) into B-A if reorder_operands_p says that its OK. Allow
negation of COMPLEX_CST if both real and imaginary parts can be
negated. Allow negation through floating point extensions and
sign-preserving built-in functions.
(negate_expr): Move the code to negate integers from "fold" to
here. Always negate integer constants unless we honor -ftrapv
and the signed type would overflow. Always negate real constants
unless we honor -ftrapping-math. Only convert -(A-B) into B-A
if allowed by reorder_operands_p. Add support for COMPLEX_CST.
Optimize negation through floating point extensions and
sign-preserving built-in functions (as defined by negate_mathfn_p).
(fold): Adjust calls to tree_swap_operands_p.
(fold <NEGATE_EXPR>): Move the remaining negation optimizations
to negate_expr_p/negate_expr.
(fold <MINUS_EXPR>): Use reorder_operands_p to check whether we're
allowed to convert (-A) - B into (-B) - A.
From-SVN: r72381
PR optimization/9325, PR java/6391
* fold-const.c (fold_convert): For floating point to integer
conversions, return the maximum/minimum representable integer
value if the real constant overflows the destination type.
* tree.c (real_value_from_int_cst): Allow the type to be NULL,
meaning don't truncate the result to a floating point mode.
Simplify the logic by calling real_from_integer directly.
* simplify-rtx.c (simplify_unary_operation): Implement the
same semantics for folding floating point to integer conversions
in RTL.
* gcc.c-torture/execute/20031003-1.c: New test case.
From-SVN: r72079
* fold-const.c (tree_swap_operands_p): New function to determine
the prefered ordering of operands.
(fold): Numerous clean-ups. Use tree_swap_operands_p when swapping
operands to commutative, comparison or ternary operators. Replace
uses of TREE_SET_CODE with recursive call to fold. Remove duplicate
transformation of A ? B : C into !A ? C : B.
From-SVN: r71438
* builtins.c (real_dconstp, fold_builtin_logarithm,
fold_builtin_exponent): New, split out from fold_builtin. Also
generalize to add log2, log10, exp2 and exp10/pow10 equivalents.
* emit-rtl.c (dconst3, dconst10, dconstthird): New.
(init_emit_once): Initialize new dconsts, use ARRAY_SIZE in lieu
of hardcoded array size.
* fold-const.c (fold): Add cases for exp2, exp10 and pow10.
(tree_expr_nonnegative_p): Likewise.
* real.h (dconst3, dconst10, dconstthird): New.
testsuite:
* gcc.dg/torture/builtin-explog-1.c: New testcase.
From-SVN: r71252
PR middle-end/12002
* tree.h (SCALAR_FLOAT_TYPE_P, COMPLEX_FLOAT_TYPE_P): New macros.
(FLOAT_TYPE_P): Define in terms of these two new macros.
* fold-const.c (fold <PLUS_EXPR>): Don't convert x+x into x*2.0
for complex floating point types.
* g77.f-torture/compile/12002.f: New test case.
From-SVN: r70821
* fold-const.c (fold <MULT_EXPR>): Optimize (C1/X)*C2 into
(C1*C2)/X when unsafe math optimizations are allowed.
(fold <RDIV_EXPR>): Optimize C1/(X*C2) into (C1/C2)/X with unsafe
math optimizations. Minor code clean-ups. Recursively call
fold when constructing sub-expressions.
* gcc.dg/20030826-1.c: New test case.
From-SVN: r70807
PR middle-end/11984
* fold-const.c (fold <PLUS_EXPR>): Check for integer constant
operands before calling tree_int_cst_lt when performing associative
transformations.
* gcc.dg/20030820-1.c: New test case.
From-SVN: r70618
* fold-const.c (negate_expr_p): MULT_EXPRs and RDIV_EXPRs are easy
to negate if either operand is easy to negate, if we don't care
about sign-dependent rounding.
(negate_expr): Make the logic to negate a REAL_CST explicit.
Attempt to negate a MULT_EXPR or RDIV_EXPR by negating an operand
that's easy to negate, if we don't honor sign-dependent rounding.
(fold <MULT_EXPR>): Optimize -A * B as A * -B if B is easy to
negate, and the symmetric A * -B as -A * B if A is easy to negate.
(fold <RDIV_EXPR>): Likewise, optimize -A/B and C/-D as A/-B and
-C/D if B and C are cheap to negate. Add an explicit rule to
optimize X/-1.0 as -X when we don't care about signaling NaNs.
From-SVN: r70455
PR middle-end/11771
* fold-const.c (negate_expr_p <MINUS_EXPR>): Change to match the
logic in negate_expr, i.e. we don't invert (A-B) for floating
point types unless flag_unsafe_math_optimizations.
* gcc.c-torture/compile/20030804-1.c: New test case.
From-SVN: r70159
* fold-const.c (fold <PLUS_EXPR>): Transform x+x into x*2.0.
Optimize x*c+x and x+x*c into x*(c+1) and x*c1+x*c2 into x*(c1+c2)
for floating point expressions with -ffast-math.
(fold <MULT_EXPR>): Don't transform x*2.0 into x+x.
* expmed.c (expand_mult): Wrap long line. Expand x*2.0 as x+x.
* gcc.dg/20030804-1.c: New test case.
From-SVN: r70158
* fold-const.c (fold <MULT_EXPR>): Optimize both x*pow(x,c) and
pow(x,c)*x as pow(x,c+1) for constant values c. Optimize x*x
as pow(x,2.0) when the latter will be expanded back into x*x.
(fold <RDIV_EXPR>): Optimize pow(x,c)/x as pow(x,c-1).
* builtins.c (expand_builtin_pow): Ignore flag_errno_math as
pow can never set errno when used with an integer exponent.
Always use expand_powi when exponent is -1, 0, 1 or 2.
(fold_builtin): Don't rewrite pow(x,2.0) as x*x nor pow(x,-2.0)
as 1.0/(x*x). This avoids unbounded recursion as we now prefer
the pow forms of these expressions.
* gcc.dg/builtins-27.c: New test case.
From-SVN: r70030
* fold-const.c (const_binop): Avoid performing the FP operation at
compile-time, if either operand is NaN and we honor signaling NaNs,
or if we're dividing by zero and either flag_trapping_math is set
or the desired mode doesn't support infinities.
(fold_initializer): New function to fold an expression ignoring any
potential run-time exceptions or traps.
* tree.h (fold_initializer): Prototype here.
* c-typeck.c (build_binary_op): Move to the end of the file so
that intializer_stack is in scope. If constructing an initializer,
i.e. when initializer_stack is not NULL, use fold_initializer to
fold expressions.
* simplify-rtx.c (simplify_binary_operation): Likewise, avoid
performing FP operations at compile-time, if they would raise an
exception at run-time.
From-SVN: r69533
PR c/11449
* fold-const.c (sign_bit_p): Return EXP if VAL is the sign bit
of HOST_WIDE_INT.
(fold_single_bit_test): If sign_bit_p() fails, assume that the
bit being tested is not a sign bit.
PR c/11449
* gcc.c-torture/compile/20030707-1.c: New.
* config/h8300/h8300.md (a peephole2): New.
From-SVN: r69184
* expr.c (do_store_flag): Pass in the correct result type
when calling fold_single_bit_test.
* fold-const.c (fold_single_bit_test): Use result_type for the
result when folding a sign bit test.
From-SVN: r68940
* real.c (real_trunc, real_floor, real_ceil): New functions
to implement trunc, floor and ceil respectively.
* real.h (real_trunc, real_floor, real_ceil): Prototype here.
* builtins.c (integer_valued_real_p): New function to test if
a floating point expression has an integer valued result.
(fold_trunc_transparent_mathfn): Optimize foo(foo(x)) as
foo(x) where foo is an integer rounding function. Similarly,
optimize foo(bar(x)) as bar(x), and foo((double)(int)x) as
(double)(int)x when both foo and bar are integer rounding
functions and we don't need to honor errno.
(fold_builtin_trunc, fold_builtin_floor, fold_builtin_ceil):
New functions to fold trunc, floor and ceil.
(fold_builtin): Use fold_builtin_trunc to fold BUILT_IN_TRUNC*,
fold_builtin_floor to fold BUILT_IN_FLOOR* and fold_builtin_ceil
to fold BUILT_IN_CEIL*.
* fold-const.c (tree_expr_nonnegative_p): Handle FLOAT_EXPR and
the remaining integer rounding functions.
* gcc.dg/builtins-25.c: New testcase.
* gcc.dg/builtins-26.c: New testcase.
From-SVN: r68903
PR target/10700
* fold-const.c (extract_muldiv_1): There's nothing that can be done
if the expression is a SAVE_EXPR.
* gcc.c-torture/compile/20030703-1.c: New test case.
From-SVN: r68877
* expr.c (do_store_flag): Remove special case folding for
single bit tests. Instead call back into the commonized folder
routine.
* fold-const.c (fold_single_bit_test): New function, mostly
extracted from do_store_flag, with an additional case extracted
from fold.
(fold): Call fold_single_bit_test appropriately.
* tree.h (fold_single_bit_test): Prototype.
From-SVN: r68867
PR optimization/11210
* expr.c (handled_component_p) [NOP_EXPR]: Add ??? note
about the behaviour with regard to bitfields.
* fold-const (decode_field_reference): Record outermost type in
case the expression is a NOP. Strip all NOPs. Set the signedness
to that of the outermost type (if any) when the bitsize is equal
to the size of the type.
From-SVN: r68823
* fold-const.c (operand_equal_p): Consider two calls to "const"
functions with identical non-volatile arguments to be equal.
Consider the FUNCTION_DECL for the "__builtin_foo" form of a
built-in function to be equal to the "foo" form.
Co-Authored-By: Jeff Law <law@redhat.com>
From-SVN: r68011
PR middle-end/168
* fold-const.c (tree_expr_nonnegative_p): Handle addition
and multiplication of zero extensions, floating point division,
and integer<->fp, fp<->fp and zero extension conversions.
The built-in ceil and floor functions preserve signedness.
* gcc.dg/20030612-1.c: New test case.
From-SVN: r67850
* fold-const.c (tree_expr_nonnegative_p): Add support for
floating point constants, addition and multiplication.
* gcc.dg/builtins-21.c: New test case.
From-SVN: r67828
* builtins.c (fold_builtin): Optimize cos(-x) as cos(x).
* fold-const.c (fold <NEGATE_EXPR>): Convert -f(x) into f(-x)
when x is easily negated and f is sin, tan or atan.
(fold <MULT_EXPR>): Optimize tan(x)*cos(x) and cos(x)*tan(x) as
sin(x) with flag_unsafe_math_optimizations.
(fold <RDIV_EXPR>): With flag_unsafe_math_optimizations fold
sin(x)/cos(x) as tan(x) and cos(x)/sin(x) as 1.0/tan(x).
* gcc.dg/builtins-20.c: New test case.
From-SVN: r67701
* fold-const.c (fold <EQ_EXPR>): Don't fold x == x only if x
is a floating point type *and* we currently honor NaNs.
(fold <NE_EXPR>): Likewise.
From-SVN: r67700
* fold-const.c (fold <ABS_EXPR>): Re-fold the result of folding
fabs(-x) into fabs(x). Use tree_expr_nonnegative_p to determine
when the ABS_EXPR (fabs or abs) is not required.
(tree_expr_nonnegative_p): Move the logic that sqrt and exp are
always nonnegative from fold to here. Additionally, cabs and fabs
are always non-negative, and pow and atan are non-negative if
their first argument is non-negative.
* builtins.c (fold_builtin_cabs): New function to fold cabs{,f,l}.
Evaluate cabs of a constant at compile-time. Convert cabs of a
non-complex argument into fabs. Convert cabs(z) into
sqrt(z.r*z.r + z.i*z.i) at the tree-level with -ffast-math or
-funsafe-math-optimizations or -ffast-math.
(fold_builtin): Convert BUILT_IN_FABS{,F,L} into an ABS_EXPR.
Fold BUILT_IN_CABS{,F,L} using fold_builtin_cabs.
* gcc.dg/builtins-2.c: Add some more tests.
* gcc.dg/builtins-18.c: New test case.
* gcc.dg/builtins-19.c: New test case.
From-SVN: r67541
* flags.h (flag_wrapv): New flag controlling overflow semantics.
* toplev.c (flag_wrapv): Declare the variable with default false.
(lang_independent_options): New option "-fwrapv" to set the above.
* fold-const.c (extract_muldiv_1): Disable optimization of (2*x)/2
as x, when signed arithmetic overflow wraps around.
(fold): Optimize "-A - B" as "-B - A" if overflow wraps around.
* loop.c (basic_induction_var): Ignore BIVs that rely on undefined
overflow when flag_wrapv is true.
* java/lang.c (java_init_options): Prescribe wrap-around two's
complement arithmetic overflow by setting flag_wrapv.
* doc/invoke.texi: Document new -fwrapv command line option.
* doc/c-tree.texi: Mention that the overflow semantics of
NEGATE_EXPR, PLUS_EXPR, MINUS_EXPR and MULT_EXPR is dependent
upon both flag_wrapv and flag_trapv.
* gcc.dg/fwrapv-1.c: New test case.
* gcc.dg/fwrapv-2.c: New test case.
* libjava.lang/Overflow.java: New test.
* libjava.lang/Overflow.out: New file.
From-SVN: r67270
* real.c (real_maxval): New function to return the largest finite
value representable in a given mode (i.e. FLT_MAX and DBL_MAX).
* real.h (real_maxval): Prototype here.
* fold-const.c (fold_inf_compare): Transform comparisons against
+-Infinity into comparisons against DBL_MAX (or equivalent).
* gcc.c-torture/execute/ieee/inf-2.c: New test case.
From-SVN: r67112
* fold-const.c (fold, case PLUS_EXPR and case MULT_EXPR): Restore
a number of conversions required for type consistency and previously
stripped off by STRIP_NOPS.
From-SVN: r65913
* tree.c (skip_simple_arithmetics_at, saved_expr_p): New functions.
(save_expr): Replace loop by call to skip_simple_arithmetics_at.
* tree.h: Add prototypes for the two new functions.
* fold-const.c (fold_binary_op_with_conditional_arg): Replace test
updates introduced in the previous revision by call to saved_expr_p.
* stor-layout.c (put_pending_size): Use skip_simple_arithmetics_at.
From-SVN: r65702
* fold-const.c (fold): Transform (c1 - x) cmp c2, where cmp is a
comparison operation and c1/c2 are floating point constants into
x swap(cmp) (c1 - c2).
* gcc.dg/20030414-2.c: New test case.
From-SVN: r65584
* fold-const.c (fold): Use "fold" following build in more places.
Optimize sqrt(x)*sqrt(x) as x, pow(x,y)*pow(z,y) as pow(x*z,y),
pow(x,y)*pow(x,z) as pow(x,y+z) and x/pow(y,z) as x*pow(y,-z).
* gcc.dg/builtins-11.c: New test case.
From-SVN: r65387
* fold-const.c (fold_inf_compare): New function to simplify FP
comparisons against +Infinity or -Infinity.
(fold): Optimize floating point comparisons against Infs and NaNs.
* gcc.c-torture/execute/ieee/fp-cmp-6.c: New test case.
* gcc.c-torture/execute/ieee/fp-cmp-7.c: New test case.
From-SVN: r64945
PR c/8224
* fold-const.c (extract_muldiv_1): Don't pass through type conversions
when signedness changes for division or modulus.
PR c/8224
* gcc.dg/20030323-1.c: New test.
From-SVN: r64760
* fold-const.c (fold_mathfn_compare): New function to simplify
comparisons against built-in math functions. Fold comparisons
of sqrt against constants.
(fold): Call fold_mathfn_compare when appropriate.
* gcc.dg/builtins-6.c: New test case.
From-SVN: r64619
* fold-const.c (omit_one_operand): No longer static.
* tree.h (omit_one_operand): Prototype here.
(div_and_round_double): Keep fold-const.c prototypes together.
* builtins.c (builtin_mathfn_code): Handle binary built-in
funtions, such as "pow" and "atan2".
(fold_builtin): Optimize both pow(x,0.0) and pow(1.0,y) to 1.0.
Simplify optimizations using "type" the builtin's return type.
* gcc.dg/builtins-5.c: New test case.
From-SVN: r63273
* fold-const.c (negate_expr_p): New function to determine whether
an expression can be negated cheaply.
(fold) [MINUS_EXPR]: Use it to determine whether to transform
-A - B into -B - A for floating point types.
From-SVN: r63059
* fold-const.c (extract_muldiv_1): Rename from extract_muldiv;
rearrange mult arguments for less recursion.
(extract_muldiv): New. Prevent runaway recursion.
Co-Authored-By: Richard Henderson <rth@redhat.com>
From-SVN: r62963
PR f/9258
* global.c (struct allocno): Add no_stack_reg.
(global_conflicts): Set no_stack_reg.
(find_reg): Use it.
* convert.c (convert_to_real): Fold - and abs only when profitable.
* fold-const.c (fold): Fold truncates in - and abs.
* gcc.c-torture/compile/20030115-1.c: New test.
* gcc.dg/i386-fpcvt-1.c: New test.
* gcc.dg/i386-fpcvt-2.c: New test.
From-SVN: r61329
* fold-const.c (fold) [COND_EXPR]: Avoid NOP_EXPRs better.
* integrate.c (copy_decl_for_inlining): Don't clear the rtl for
static/external decls.
cp/
* call.c (build_conditional_expr): Stabilize lvalues properly.
* cvt.c (ocp_convert): Don't build NOP_EXPRs of class type.
* tree.c (lvalue_p_1): Don't allow sloppy NOP_EXPRs as lvalues.
Don't allow CALL_EXPR or VA_ARG_EXPR, either.
* call.c (convert_like_real): Call decl_constant_value for an
IDENTITY_CONV even if there are no more conversions.
* cvt.c (build_up_reference): Don't push unnamed temps.
* decl2.c (do_namespace_alias): Namespace aliases are DECL_EXTERNAL.
* dump.c (cp_dump_tree): Don't try to dump class-specific fields
for a backend struct.
* except.c (wrap_cleanups_r, build_throw): Make
MUST_NOT_THROW_EXPRs void.
* init.c (expand_default_init): Update to handle MUST_NOT_THROW_EXPR.
* init.c (build_vec_delete_1): Pre-evaluate the base address.
* init.c (get_temp_regvar): Simplify logic.
* tree.c (cp_copy_res_decl_for_inlining): Only do debug tweaks if
our replacement is a decl.
From-SVN: r60851
PR c/8639
* fold-const.c (extract_muldiv): Don't propagate division unless
both arguments are multiples of C.
* gcc.c-torture/execute/20021119-1.c: New.
From-SVN: r59466
2002-11-05 Andrew Haley <aph@redhat.com>
* fold-const.c (fold): Don't transform (a0 op compound(a1,a2))
to (compound(a1,a0 op a2)) if a0 or a1 have side effects.
From-SVN: r58822
* fold-const.c (fold_binary_op_with_conditional_arg): Improve
handling of cases where one or both branches of the conditional
have void type, i.e. throw an exception or don't return.
(fold): Only apply (and undo) type conversion to the non-void
branches of a COND_EXPR.
* f/com.c (ffecom_subscript_check_): Cast the failure branch
of the bounds check COND_EXPR to void, to indicate noreturn.
(ffe_truthvalue_conversion): Only apply truth value conversion
to the non-void branches of a COND_EXPR.
From-SVN: r58661
PR c++/7209
* fold_const.c (fold_binary_op_with_conditional_arg): Always
build compound_expr if we used save_expr.
testsuite:
* g++.dg/expr/cond1.C: New test.
From-SVN: r58405
* fold-const.c (make_range): Only narrow to signed range if
the signed range is smaller than the unsigned range.
* gcc.c-torture/execute/20020911-1.c: New.
From-SVN: r57050
* fold-const.c (associate_trees): Only optimize NEGATE_EXPR in one
of the operands into MINUS_EXPR if code is PLUS_EXPR.
* gcc.c-torture/execute/20020805-1.c: New test.
From-SVN: r56058
* Makefile.in: rtlanal.o now depends upon real.h.
* flags.h [flag_signaling_nans]: New flag.
[HONOR_SNANS]: New macro.
* toplev.c [flag_signaling_nans]: Initialize to false.
(f_options): Add processing for "-fsignaling-nans".
(set_fast_math_flags): Clear flag_signaling_nans with -ffast-math.
(process_options): flag_signaling_nans implies flag_trapping_math.
* c-common.c (cb_register_builtins): Define __SUPPORT_SNAN__
when -fsignaling-nans. First step to implementing WG14's N965.
* fold-const.c (fold) [MULT_EXPR]: Conditionalize transforming
1.0 * x into x, and -1.0 * x into -x on !HONOR_SNANS.
[RDIV_EXPR]: Conditionalize x/1.0 into x on !HONOR_SNANS.
* simplify-rtx.c (simplify_relational_operation): Conditionalize
transforming abs(x) < 0.0 into false on !HONOR_SNANS.
* rtlanal.c: #include real.c for TARGET_FLOAT_FORMAT definitions
required by HONOR_SNANS. (may_trap_p): Floating point DIV, MOD,
UDIV, UMOD, GE, GT, LE, LT and COMPARE may always trap with
-fsignaling_nans. EQ and NE only trap for flag_signaling_nans
not flag_trapping_math (i.e. HONOR_SNANS but not HONOR_NANS).
* doc/invoke.texi: Document new -fsignaling-nans compiler option.
From-SVN: r55804
PR opt/4046
* fold-const.c (fold) [COND_EXPR]: Simplify A ? 0 : 1 to !A,
A ? B : 0 to A && B and A ? B : 1 into !A || B if both A and
B are truth values.
From-SVN: r55153
* fold-const.c (comparison_to_compcode): New function to convert
an comparison TREE CODE into a bit-based representation.
(compcode_to_comparison): New function to convert from this bit
based representation back to a comparison TREE CODE.
(fold_truthop): Simplify (x<y) && (x==y) and related composite
comparisons.
* gcc.c-tortuture/execute/compare-1.c: New test case.
* gcc.c-tortuture/execute/compare-2.c: New test case.
* gcc.c-tortuture/execute/compare-3.c: New test case.
From-SVN: r54647
* convert.c (convert_to_integer) [LSHIFT_EXPR]: Don't pass
the truncation down when the target type is signed.
[trunc1]: Use unsigned arithmetic for LSHIFT_EXPR.
* fold-const.c (extract_muldiv) [NOP_EXPR]: Don't pass through
the conversion if the target type is a smaller type.
From-SVN: r54627
* fold-const.c (fold) [EQ_EXPR]: Place both integer and real
constants last in comparisons. Optimize (x+1.0)>0.0 into the
equivalent x > -1.0 when -ffast-math.
* gcc.dg/20020607-2.c: New test case.
From-SVN: r54356
* tree.c (real_minus_onep): New function to test for -1.0.
* fold-const.c (fold) [MULT_EXPR]: Optimize -1.0*x into -x.
* gcc.dg/fnegate-1.c: New test case.
From-SVN: r54149
* tree.h: Forward-declare struct realvaluetype.
(struct tree_real_cst): Point to the REAL_VALUE_TYPE, do not
contain it.
(TREE_REAL_CST_PTR): New accessor.
(TREE_REAL_CST): Update.
* real.h: Include machmode.h.
(realvaluetype): Make it struct realvaluetype, not a typedef.
(build_real): Prototype here.
* tree.c: Include real.h.
(build_real): Allocate the REAL_VALUE_TYPE as a separate
object in GC memory, set TREE_REAL_CST_PTR to point to it.
(build_real_from_int_cst): Use build_real.
* ggc-common.c (ggc_mark_trees): Mark TREE_REAL_CST_PTR of a
REAL_CST.
* builtins.c, c-common.c, c-lex.c, dwarf2out.c, expr.c,
fold-const.c, print-tree.c, real.c, cp/mangle.c, cp/tree.c,
f/bld.c, f/com.c, f/expr.c, f/target.c, java/decl.c,
java/jcf-parse.c, java/parse.y, java/typeck.c: Include real.h.
* Makefile.in, cp/Make-lang.in, f/Make-lang.in,
java/Make-lang.in: Update dependency lists.
From-SVN: r53959
2002-05-15 Jakub Jelinek <jakub@redhat.com>
* fold-const.c (fold): Fix a typo.
2002-05-15 Eric Botcazou <ebotcazou@multimania.com>
* fold-const.c (fold) [LT_EXPR]: Move the transformation of a
comparison against the highest or lowest integer value before
the 'X >= CST to X > (CST - 1)' and 'X < CST to X <= (CST - 1)'
transformation and that of an unsigned comparison against 0
right after.
From-SVN: r53493
PR opt/3995
* fold-const.c (sign_bit_p): New function.
(fold) [EQ_EXPR]: Use this to convert (A & C) == 0 into A >= 0 and
(A & C) != 0 into A < 0, when constant C is the sign bit of A's type.
Reapply fold when converting (A & C) == C into (A & C) != 0.
(fold_binary_op_with_conditional_arg): Fix typo in comment.
testsuite/
* gcc.c-torture/execute/20020506-1.c: New test case.
From-SVN: r53241
* fold-const.c (lshift-double): Cast the high word to an unsigned
HOST_WIDE_INT when extracting sign bit to avoid compiler warning.
(div_and_round_double): Cast carry to a signed HOST_WIDE_INT to
avoid compiler warning. (fold): Remove redundant code from
BIT_AND_EXPR as integer operands are canonicalized to be arg1.
From-SVN: r53228
2002-04-18 David S. Miller <davem@redhat.com>
* config/sparc/sparc.h (BRANCH_COST): Define.
* fold-const.c (BRANCH_COST): Don't provide default here, expr.h
does it.
From-SVN: r52500
* fold-const.c (fold) [NOP_EXPR]: Convert (T)(x&c) into ((T)x&(T)c)
for integer constant c (if x has unsigned type or sign bit is not
set in c). This folds the zero/sign extension into the bit-wise and
operation.
* gcc.c-torture/compile/20020415-1.c: New.
Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
From-SVN: r52465
* toplev.c: Don't include setjmp.h. Kill float_handler_set,
float_handled, float_handler, float_signal, set_float_handler,
and do_float_handler. Set handler for SIGFPE to crash_signal.
* toplev.h: Don't prototype do_float_handler.
* c-lex.c: Fold parse_float into lex_number. Make warning
about portability of hex float constants more informative, and
don't issue it on top of a syntax error.
* fold-const.c: Fold const_binop_1 and fold_convert_1 into
their callers.
* real.h: Define REAL_VALUE_ABS here...
* simplify-rtx.c: ... not here. Fold check_fold_consts,
simplify_unary_real, simplify_binary_real, and
simplify_binary_is2orm1 into their callers.
* tree.c: Fold build_real_from_int_cst_1 into caller.
* doc/tm.texi: Document REAL_VALUE_ABS and REAL_VALUE_NEGATIVE.
* tsystem.h: Include float.h here...
* libgcc2.c: ... not here.
java:
* lex.c: Change java_perform_atof to take normal parameters
instead of a pointer to a parameter block. Call it directly
from java_lex.
testsuite:
* gcc.dg/c90-hexfloat-1.c: Adjust error regexps.
From-SVN: r51336
* real.h: Don't define REAL_INFINITY or REAL_IS_NOT_DOUBLE.
Always make REAL_VALUE_TYPE a struct containing an array of
HOST_WIDE_INT, not a double. Tidy up the code deciding how
big it is. Don't declare or use union real_extract.
* emit-rtl.c (init_emit_once), varasm.c (immed_real_const_1,
decode_rtx_const, output_constant_pool), config/a29k/a29k.c
(print_operand), config/arm/arm.c (output_move_double),
config/arm/arm.md (consttable_4, consttable_8),
config/romp/romp.c (output_fpops), config/s390/s390.h
(ASM_OUTPUT_SPECIAL_POOL_ENTRY), config/xtensa/xtensa.c
(xtensa_output_literal): Don't use union real_extract.
* config/dsp16xx/dsp16xx.c (print_operand), config/i860/i860.c
(sfmode_constant_to_ulong), config/ns32k/merlin.h
(PRINT_OPERAND), config/ns32k/ns32k.c (print_operand),
config/pdp11/pdp11.h (PRINT_OPERAND), config/we32k/we32k.h
(PRINT_OPERAND): Don't use local version of union
real_extract.
* config/convex/convex.c (check_float_value), config/vax/vax.c
(vax_float_literal), config/m88k/m88k.md (divdf3),
config/dsp16xx/dsp16xx.md (fixuns_trunchfhi2),
config/pdp11/pdp11.c (output_move_quad): Don't do host
arithmetic on target floating point quantities.
* config/a29k/a29k.md, config/dsp16xx/dsp16xx.c
(output_dsp16xx_float_const): Don't test HOST_FLOAT_FORMAT.
* fold-const.c (fold), simplify-rtx.c (simplify_binary_real):
Use MODE_HAS_INFINITIES rather than #ifdef REAL_INFINITY.
* real.c (earith): Test INFINITY rather than REAL_INFINITY;
NANS implies INFINITY, so can drop #ifdef NANS inside #ifndef
INFINITY.
* print-rtl.c (print_rtx): Disable code which needs
floating-point emulator.
* libgcc2.c: Include float.h and use DBL_MANT_DIG,
FLT_MANT_DIG, to define DF_SIZE and SF_SIZE, rather than
depending on HOST_FLOAT_FORMAT to be defined properly.
* ch/grant.c, cp/error.c: Always use REAL_VALUE_TO_DECIMAL;
don't test REAL_IS_NOT_DOUBLE.
* config/1750a/1750a.c (get_double, float_label): Delete.
(print_operand): Delete huge commented-out chunk. Use
REAL_VALUE_TO_DECIMAL.
* config/1750a/1750a-protos.h: Delete prototypes of deleted
functions.
* config/convex/convex.h: Always set TARGET_FLOAT_FORMAT to
IEEE_FLOAT_FORMAT.
* config/i370/i370.h (PRINT_OPERAND [TARGET_HLASM version]):
Use REAL_VALUE_TO_DECIMAL as ELF version does.
* config/m88k/m88k.c (real_power_of_2_operand,
legitimize_operand): Take the REAL_VALUE_TYPE and/or union
real_extract out of the union; run the input through
REAL_VALUE_TO_TARGET_DOUBLE, then plug the pair of longwords
from that into the union.
* config/pdp11/pdp11.c (output_move_double): Rearrange
parentheses to make automatic indenter happy.
* doc/tm.texi (Cross-compilation): Rename node to "Floating
Point" and rewrite to describe current situation. Also adjust
documentation of REAL_VALUE_TO_TARGET_SINGLE and friends to
match code.
* doc/rtl.texi: Adjust cross reference.
From-SVN: r51210
* defaults.h (MODE_HAS_NANS, MODE_HAS_INFINITIES): New.
(MODE_HAS_SIGNED_ZEROS, MODE_HAS_SIGN_DEPENDENT_ROUNDING): New.
* flags.h (HONOR_NANS, HONOR_INFINITIES, HONOR_SIGNED_ZEROS): New.
(HONOR_SIGN_DEPENDENT_ROUNDING): New.
* builtins.c (expand_builtin_mathfn): Use HONOR_NANS.
* c-common.c (truthvalue_conversion): Reduce x - y != 0 to x != y
unless x and y could be infinite.
(expand_unordered_cmp): New, mostly split from expand_tree_builtin.
Check that the common type of both arguments is a real, even for
targets without unordered comparisons. Allow an integer argument
to be compared against a real.
(expand_tree_builtin): Use expand_unordered_cmp.
* combine.c (combine_simplify_rtx): Use the new HONOR_... macros.
* cse.c (fold_rtx): Likewise. Fix indentation.
* fold-const.c (fold_real_zero_addition_p): New.
(fold): Use it, and the new HONOR_... macros.
* ifcvt.c (noce_try_minmax): Use the new HONOR_... macros.
* jump.c (reversed_comparison_code_parts): After searching for
the true comparison mode, use HONOR_NANS to decide whether it
can be safely reversed.
(reverse_condition_maybe_unordered): Remove IEEE check.
* simplify-rtx.c (simplify_binary_operation): Use the new macros
to decide which simplifications are valid. Allow the following
simplifications for IEEE: (-a + b) to (b - a), (a + -b) to (a - b),
and (a - -b) to (a + b).
(simplify_relational_operation): Use HONOR_NANS.
* doc/tm.texi: Document the MODE_HAS_... macros.
From-SVN: r50401
* tree.def (VIEW_CONVERT_EXPR): New tree code.
* tree.h (handled_component_p): New declaration.
* alias.c (handled_component_p): Move function from here ...
* expr.c (handled_component_p): ... to here and non longer static.
(is_zeros_p, case VIEW_CONVERT_EXPR): New case.
(store_field): Remove unneeded MEM_COPY_ATTRIBUTES call.
(get_inner_reference): Handle VIEW_CONVERT_EXPR.
(expand_expr, case VAR_DECL): Clean up handling of re-layout case.
(expand_expr, case VIEW_CONVERT_EXPR): New case.
(expand_expr, cse ADDR_EXPR): No need to copy if already
BIGGEST_ALIGNMENT.
* fold-const.c (fold, case VIEW_CONVERT_EXPR): New case.
* varasm.c (initializer_constant_value_p, case VIEW_CONVERT_EXPR): New.
(output_constant): Handle VIEW_CONVERT_EXPR.
* dwarf2out.c (loc-descriptor_from_tree, add_bound_info):
Add new case for VIEW_CONVERT_EXPR.
From-SVN: r47179
* c-common.c (signed_or_unsigned_type): Handle machine mode types
which have no corresponding C type.
* fold_const.c (target_isinf,target_isnan,target_negative): Update
the representation of 64-bit doubles to work with 64-bit ints.
From-SVN: r45479
* toplev.c (set_float_handler): Make static.
* toplev.h: Don't prototype set_float_handler.
* simplify-rtx.c: Don't include setjmp.h.
(simplify_unary_real, simplify_binary_real, simplify_binary_is2orm1):
New functions.
(simplify_unary_operation, simplify_binary_operation): Use them,
via do_float_handler.
* fold-const.c: Don't include setjmp.h.
(exact_real_inverse_1): New function.
(exact_real_inverse): Use it, via do_float_handler.
* varasm.c: Don't include setjmp.h.
(assemble_real_1): New function.
(assemble_real): Use it, via do_float_handler.
Call internal_error if we get a trap here.
* c-parse.in, cse.c, cselib.c, ch/lex.c, config/i386/i386.c,
config/pj/pj.c, config/s390/s390.c: Don't include setjmp.h.
* java/lex.h: Don't include setjmp.h. Don't define
SET_FLOAT_HANDLER or prototype set_float_handler.
From-SVN: r44815
* integrate.c (copy_insn_list): handle
NOTE_INSN_DELETED_LABEL, as if it were CODE_LABEL.
(copy_rtx_and_substitute): Likewise; behave to NOTE_INSN_DELETED_LABEL
identically as to CODE_LABEL.
* fold-const (fold): Convert A/B/C to A/(B*C) and
A/(B/C) to (A/B)*C
* except.c (connect_post_landing_pads): Delete the RESX insns.
From-SVN: r44193
* fold-const.c (fold): Use first_rtl_op rather than
TREE_CODE_LENGTH when looping over parameters of a tree.
Correct strange grouping in test for evaluated SAVE_EXPR.
From-SVN: r41437
* fold-const.c (fold): Before optimizing unsigned comparison with
0x7fffffffU, make sure arg0 is integral type.
* gcc.c-torture/execute/20010329-1.c: New test.
From-SVN: r41000
* fold-const.c (extract_muldiv) [case PLUS_EXPR]: If not MULT_EXPR,
check if either operand is divisible by C.
(multiple_of_p): Handle LSHIFT_EXPR with small constant shift.
If type is signed, consider negative numbers as well.
* gcc.c-torture/execute/20010222-1.c: New test.
From-SVN: r39996
* fold-const.c (fold_binary_op_with_conditional_arg): New
function, split out from ...
(fold): ... here.
* tree.def (COND_EXPR): Document the use of VOID_TYPE for
conditional arms that throw exceptions.
From-SVN: r39822
* stor-layout.c (is_pending_size, put_pending_size): New functions.
(variable_size): Call put_pending_size.
* tree.h (is_pending_size, put_pending_size): Add prototypes.
* fold-const.c (extract_muldiv): If SAVE_EXPR is on the pending
sizes list, put newly created SAVE_EXPR there as well.
* gcc.c-torture/execute/20010209-1.c: New test.
* config/ia64/ia64.c (last_group): Only 2 entries are needed.
(errata_find_address_regs): load_group has only 2 entries.
(errata_emit_nops): Likewise. shladd is not problematic.
Clear last_group if nop was emitted.
(fixup_errata): load_group has only 2 entries.
Optimize.
From-SVN: r39663
* fold-const.c (split_tree): Don't assume a constant isn't splittable.
(fold): Don't assume a constant isn't foldable.
* tree.c (build): Set TREE_CONSTANT for an expression with no
side-effects and constant operands.
(build1): Likewise.
From-SVN: r39585
* fold-const.c (fold): When folding a CONJ_EXPR of a COMPLEX_CST,
use TREE_REALPART and TREE_IMAGPART instead of TREE_OPERAND.
testsuite:
* gcc.c-torture/compile/20001222-1.x: Remove.
From-SVN: r38696
* fold-const.c (force_fit_type): Unsigned values can overflow
if they are sizetype.
(int_const_binop): Don't use cache if overflows.
* cp/typeck.c (dubious_conversion_warning): Suppress if
TYPE_IS_SIZETYPE.
From-SVN: r36970
* calls.c (ECF_SP_DEPRESSED): New macro.
(calls_function_1): Treat calling sp-depressed function as alloca.
(emit_call_1): Don't adjust SP if calling sp-depressed function.
(expand_call): Set ECF_SP_DEPRESSED if TYPE_RETURNS_STACK_DEPRESSED.
If sp-depressed, ensure block saves and restores SP.
* fold-const.c (extract_muldiv): Only check TYPE_IS_SIZETYPE
for INTEGER_TYPE.
* function.c (keep_stack_depressed): New function.
(thread_prologue_and_epilogue_insns): Call it.
* print-tree.c (print_node): Use HOST_WIDE_INT_PRINT_UNSIGNED
to print DECL_OFFSET_ALIGN.
Print no-force-blk and transparent-union flags properly.
* stmt.c (expand_goto_internal): Don't restore stack if last block
and function returns with sp depressed.
(fixup_gotos): Likewise.
(save_stack_pointer): New function, from code in expand_decl.
(expand_decl): Call new function.
* tree.h (TYPE_IS_SIZETYPE): Call INTEGER_TYPE_CHECK.
(TYPE_RETURNS_STACK_DEPRESSED): New macro.
(save_stack_pointer): New declaration.
From-SVN: r35734
2000-08-04 Andreas Schwab <schwab@suse.de>
* cppmain.c (cb_def_pragma): Skip the first two tokens from the
token list, which are always `#' and `pragma'.
2000-08-04 Zack Weinberg <zack@wolery.cumb.org>
* tree.c (tree_expr_nonnegative_p): Move to...
* fold-const.c: ... here. Also handle BIND_EXPR and RTL_EXPR.
(rtl_expr_nonnegative_p): New.
* tree.h: Add prototype for rtl_expr_nonnegative_p.
* real.h (CONST_DOUBLE_LOW, CONST_DOUBLE_HIGH, CONST_DOUBLE_MEM,
CONST_DOUBLE_CHAIN: Move to...
* rtl.h: ...here. Use XCINT/XCEXP.
* Makefile.in: Remove toplev.o from OBJS. Add rule to make
libbackend.a; add libbackend.a to STAGESTUFF. Add BACKEND
variable. Use BACKEND when linking cc1, not OBJS. Add BACKEND
to VOL_FILES.
* objc/Make-lang.in (cc1obj): Link with $(BACKEND).
ch:
* Make-lang.in (cc1chill): Depend on $(BACKEND), not stamp-objlist.
* Makefile.in (cc1chill): Link with $(BACKEND). Define BACKEND,
eliminate C_OBJS (was commented out), OBJS, OBJDEPS.
cp:
* Make-lang.in (cc1plus): Depend on $(BACKEND), not stamp-objlist.
* Makefile.in: Add C_OBJS, BACKEND; delete OBJS, OBJDEPS.
(cc1plus): Link with $(BACKEND) and $(C_OBJS).
f:
* Make-lang.in (f771): Depend on $(BACKEND), not stamp-objlist.
* Makefile.in: Add BACKEND; delete OBJS, OBJDEPS.
(f771): Link with $(BACKEND).
java:
* Make-lang.in (jc1, jv-scan): Depend on $(BACKEND), not stamp-objlist.
* Makefile.in: Add BACKEND; delete OBJS, OBJDEPS.
(jc1): Link with $(BACKEND).
(jv-scan): Depend on version.o, not all of $(OBJS) or $(BACKEND).
From-SVN: r35501
2000-07-08 Toon Moene <toon@moene.indiv.nluug.nl>
* fold-const.c (extract_muldiv) case PLUS_EXPR, MINUS_EXPR:
Check whether c divides op1 exactly if operation is not
multiplication.
From-SVN: r34928
(extract_muldiv): When constructing a multiplier/divisor, do not expect
const_binop to correctly determine if overflow has occured, so check explicitly.
From-SVN: r34135
* Makefile.in ($(srcdir)/c-parse.y: c-parse.in): Enclose the whole
message in quotes. Otherwise, IBM's make program treats the '#' as the
start of a comment and ignores the remainder of the line.
* c-lex.c (yylex): Change for EBCDIC, lower case characters preceed
upper case.
* cccp.c (initialize_char_syntax): Allow for holes in EBCDIC.
* cexp.y (initialize_random_junk): Likewise.
* cppfiles.c (find_include_file): Cast alloca return value.
* cppinit.c (initialize_standard_includes): Likewise.
* cpplib.c (cpp_define, cpp_undef): Likewise.
* defaults.h (ASM_OUTPUT_ASCII): Use ISPRINT.
* final.c (output_asm_insn): Allow for holes in EBCDIC.
* fold-const.c (CHARMASK): New.
(real_hex_to_f): Use it.
* real.c (CHARMASK): New.
(etoasc, asctoeg): Use it.
(asctoeg): EBCDIC lower case characters preceed upper case.
From-SVN: r33192
* expr.c (store_constructor): SIZE now signed.
For EXPR_SIZE, don't evaluate size; just needed if constant.
* fold-const.c (fold): Fix a number of cases when folded tree is
wrong type.
* function.c (flush_addressof): Reenable.
* tree.h (flush_addressof): Likewise.
* toplev.c (independent_decode_option): Look at strings_processed.
* config/alpha/alpha.h (MINIMUM_ATOMIC_ALIGNMENT): Cast to unsigned.
From-SVN: r32783
modify sizes to be unsigned and use HOST_WIDE_INT.
* alias.c (reg_known_value_size): Now unsigned.
* c-typeck.c (build_unary_op, case ADDR_EXPR): Use byte_position.
(really_start_incremental_init): Use bitsize_zero_node.
(push_init_level, pop_init_level, output_init_element): Likewise.
Use bitsize_unit_node and bitsize_one_node.
(output_pending_init_elements, process_init_element): Likewise.
* combine.c (combine_max_regno, reg_sign_bit_copies): Now unsigned.
(make_extraction): Position and length HOST_WIDE_INT and unsigned
HOST_WIDE_INT, respectively.
(get_pos_from_mask): Passed in value is unsigned HOST_WIDE_INT.
(num_sign_bit_copies): Returns unsigned.
BITWIDTH now unsigned; rework arithmetic.
Remove recursive call from arg to MAX.
(combine_instructions, init_reg_last_arrays): NREGS now unsigned.
(setup_incoming_promotions, can_combine_p, try_combine, simplify_set):
REGNO now unsigned.
(set_nonzero_bit_and_sign_copies): NUM now unsigned.
(find_split_point, expand_compound_operation, make_extraction): LEN
now unsigned HOST_WIDE_INT, POS now HOST_WIDE_INT.
(make_field_assignment): Likewise.
(combine_simplify_rtx): Add cast.
(expand_compound_operation): MODEWIDTH now unsigned; rework arithmetic.
(force_to_mode): WIDTH now unsigned; add cast.
(if_then_else_cond): SIZE now unsigned.
(nonzero_bits): MODE_WIDTH, RESULT_WIDTH, and WIDTH now unsigned.
(extended_count): Now returns unsigned.
(simplify_shift_const): COUNT unsigned; arg is now INPUT_COUNT.
Add SIGNED_COUNT variable; MODE_WORDS and FIRST_COUNT now unsigned.
(simplify_comparison): MODE_WIDTH now unsigned.
(update_table_tick): REGNO and ENDREGNO now unsigned; new var R.
(mark_used_regs_combine): Likewise; rework arithmetic.
(record_value_for_reg): REGNO, ENDREGNO, and I now unsigned.
(record_dead_and_set_regs, reg_dead_at_p, distribute_notes): Likewise.
(record_promoted_value): REGNO now unsigned.
(get_last_value_validate): REGNO, ENDREGNO, and J now unsigned.
(get_last_value): REGNO now unsigned.
(use_crosses_set_p): REGNO and ENDREGNO now unsigned.
(reg_dead_regno, reg_dead_endregno): Now unsigned.
(remove_death): Arg REGNO now unsigned.
(move_deaths): REGNO, DEADREGNO, DEADEND, OUREND, and I now unsigned.
(reg_bitfield_target_p): REGNO, REGNO, ENDREGNO, and ENDTREGNO
now unsigned.
* convert.c (convert_to_integer): INPREC and OUTPREC now unsigned.
* cse.c (struct qty_table_elem): FIRST_REG and LAST_REG now unsigned.
(struct cse_reg_info): REGNO now unsigned.
(cached_regno): Now unsigned.
(REGNO_QTY_VALID_P): Add cast.
(make_new_qty, make_regs_eqv, delete_reg_eqiv): Regno args unsigned.
(remove_invalid_regs): Likewise.
(remove_invalid_subreg_refs): Likewise; arg WORD also unsigned
as are variables END and I.
(get_cse_reg_info, insert): Likewise.
(mention_regs, invalidate_for_call): REGNO, ENDREGNO, and I unsigned.
(canon_hash): Likewise.
(insert_regs, lookup_for_remove): REGNO now unsigned.
(invalidate): REGNO, ENDREGNO, TREGNO, and TENDREGNO now unsigned.
New variable RN.
* dbxout.c (dbxout_parms, dbxout_reg_parms): Don't check for REGNO < 0.
* dwarf2out.c (dwarf2ou_frame_debug_expr): Remove cast.
* emit-rtl.c (subreg_realpart_p): Add cast.
(operand_subword): Arg I is now unsigned as is var PARTWORDS.
(operand_subword_force): Arg I is now unsigned.
* except.c (eh_regs): Variable I is now unsigned.
* explow.c (hard_function_value): BYTES is unsigned HOST_WIDE_INT.
* expmed.c (store_fixed_bit_field): Position is HOST_WIDE_INT;
length is unsigned HOST_WIDE_INT; likewise for internal variables.
(store_split_bit_field, extract_fixed_bit_field): Likewise.
(extract_split_bit_field, store_bit_field, extract_bit_field):
Likewise.
* expr.c (store_constructor_fields, store_constructor, store_field):
Positions are HOST_WIDE_INT and lengths are unsigned HOST_WIDE_INT.
(expand_assignment, expand_expr, expand_expr_unaligned): Likewise.
(do_jump): Likewise.
(move_by_pieces, move_by_pieces_ninsns, clear_by_pieces):
MAX_SIZE is now unsigned.
(emit_group_load): BYTEPOS is HOST_WIDE_INT; BYTELEN is unsigned.
(emit_group_store): Likewise.
(emit_move_insn): I now unsigned.
(store_constructor): Use host_integerp, tree_low_cst, and
bitsize_unit_node.
(get_inner_reference): Return bitpos and bitsize as HOST_WIDE_INT.
Rework all calculations to use trees and new fields.
* expr.h (promoted_input_arg): Regno now unsigned.
(store_bit_field, extract_bit_field): Adjust types of pos and size.
(mark_seen_cases): Arg is HOST_WIDE_INT.
* flow.c (verify_wide_reg_1): REGNO now unsigned.
* fold-const.c (decode_field_reference): Size and pos HOST_WIDE_INT;
precisions and alignments are unsigned.
(optimize_bit_field_compare, fold_truthop): Likewise.
(int_const_binop): Adjust threshold for size_int_type_wide call.
(fold_convert): Likewise.
(size_int_type_wide): Make table larger and fix thinko that only
had half of table used.
(all_ones_mask_p, fold): Precisions are unsigned.
* function.c (put_reg_info_stack): REGNO is unsigned.
(instantiate_decl): Size is HOST_WIDE_INT.
(instantiate_virtual_regs): I is unsigned.
(assign_parms): REGNO, REGNOI, and REGNOR are unsigned.
(promoted_input_arg): REGNO is unsigned.
* function.h (struct function): x_max_parm_reg is now unsigned.
* gcse.c (max_gcse_regno): Now unsigned.
(struct null_pointer_info): min_reg and max_reg now unsigned.
(lookup_set, next_set): REGNO arg now unsigned.
(compute_hash_table): REGNO and I now unsigned.
(handle_avail_expr): regnum_for_replacing now unsigned.
(cprop_insn): REGNO now unsigned.
(delete_null_pointer_checks_1): BLOCK_REG now pointer to unsigned.
* ggc-common.c (ggc_mark_tree_children, case FIELD_DECL): New case.
* global.c (set_preference): SRC_REGNO, DEST_REGNO, and I now unsigned.
* hard-reg-set.h (reg_class_size): Now unsigned.
* integrate.c (mark_stores): LAST_REG and I now unsigned; new UREGNO.
* jump.c (mark_modified_reg): I now unsigned; add cast.
(rtx_equal_for_thread_p): Add cast.
* loop.c (max_reg_before_loop): Now unsigned.
(struct_movable): REGNO now unsigned.
(try_copy_prop): REGNO arg unsigned.
(regs_match_p): XN and YN now unsigned.
(consec_sets_invariant_p, maybe_eliminate_biv): REGNO now unsigned.
(strength_reduce): Likewise; NREGS also unsigned.
(first_increment_giv, last_increment_giv unsigned): Now unsigned.
* loop.h (struct iv_class): REGNO now unsigned.
(max_reg_before_loop, first_increment_giv, last_increment_giv):
Now unsigned.
* machmode.h (mode_size, mode_unit_size): Now unsigned.
(mode_for_size, smallest_mode_for_size): Pass size as unsigned.
* optabs.c (expand_binop): I and NWORDS now unsigned.
(expand_unop): I now unsigned.
* print-tree.c (print_node): Don't print DECL_FIELD_BITPOS, but do
print DECL_FIELD_OFFSET and DECL_FIELD_BIT_OFFSET.
* real.c (significand_size): Now returns unsigned.
* real.h (significand_size): Likewise.
* regclass.c (reg_class_size): Now unsigned.
(choose_hard_reg_mode): Both operands now unsigned.
(record_reg_classes): REGNO and NR now unsigned.
(reg_scan): NREGS now unsigned.
(reg_scan_update): old_max_regno now unsigned.
(reg_scan_mark_refs): Arg MIN_REGNO and var REGNO now unsigned.
* reload.c (find_valid_class): BEST_SIZE now unsigned.
(find_dummy_reload): REGNO, NWORDS, and I now unsigned.
(hard_reg_set_here_p): Args BEG_REGNO and END_REGNO now unsigned.
Likewise for variable R.
(refers_to_regno_for_reload_p): Args REGNO and END_REGNO now unsigned,
as are variables INNER_REGNO and INNER_ENDREGNO; add new variable R.
(find_equiv_reg): Add casts.
(regno_clobbered_p): Arg REGNO now unsigned.
* reload.h (struct reload): NREGS now unsigned.
(refers_to_regno_for_reload_p): Regno args are unsigned.
(regno_clobbered_p): Likewise.
* reload1.c (reg_max_ref_width, spill_stack_slot_width): Now unsigned.
(compute_use_by_pseudos): REGNO now unsigned.
(find_reg): I and J now unsigned, new variable K, and change loop
variables accordingly; THIS_NREGS now unsigned.
(alter_reg): INHERENT_SIZE and TOTAL_SIZE now unsigned.
(spill_hard_reg): REGNO arg now unsigned; add casts.
(forget_old_reloads_1): REGNO, NR, and I now unsigned.
(mark_reload_reg_in_use): Arg REGNO and vars NREGS and I now unsigned.
(clear_reload_reg_in_use): Arg REGNO and vars NREGS, START_REGNO,
END_REGNO, CONFLICT_START, and CONFLICT_END now unsigned.
(reload_reg_free_p, reload_reg_reaches_end_p): Arg REGNO now unsigned.
(choose_reload_regs): MAX_GROUP_SIZE now unsigned.
(emit_reload_insns): REGNO now unsigned.
(reload_cse_move2add): Add cast.
(move2add_note_store): REGNO and I now unsigned; new variable ENDREGNO
and rework loop.
* resource.c (mark_referenced_resources, mark_set_resources): New
variable R; REGNO and LAST_REGNO now unsigned.
(mark_target_live_regs): J and REGNO now unsigned.
* rtl.c (mode_size, mode_unit_size): Now unsigned.
* rtl.h (union rtunion_def): New field rtuint.
(XCUINT): New macro.
(ADDRESSOF_REGNO, REGNO, SUBREG_WORD): New XCUINT.
(operand_subword, operand_subword_force): Word number is unsigned.
(choose_hard_reg_mode): Operands are unsigned.
(refers_to-regno_p, dead_or_set_regno_p): Regno arg is unsigned.
(find_regno_note, find_regno_fusage, replace_regs): Likewise.
(regno_use_in, combine_instructions, remove_death): Likewise.
(reg_scan, reg_scan_update): Likewise.
(extended_count): Return is unsigned.
* rtlanal.c (refers_to_regno_p): Args REGNO and ENDREGNO and vars I,
INNER_REGNO, and INNER_ENDREGNO now unsigned; new variable X_REGNO.
(reg_overlap_mentioned_p): REGNO and ENDREGNO now unsigned.
(reg_set_last_first_regno, reg_set_last_last_regno): Now unsigned.
(reg_reg_last_1): FIRS and LAST now unsigned.
(dead_or_set_p): REGNO, LAST_REGNO, and I now unsigned.
(dead_or_set_regno_p): Arg TEST_REGNO and vars REGNO and ENDREGNO
now unsigned.
(find_regno_note, regno_use_in): Arg REGNO now unsigned.
(find_regno_fusage): Likewise; also var REGNOTE now unsigned.
(find_reg_fusage): Variables REGNO, END_REGNO, and I now unsigned.
(replace_regs): Arg NREGS now unsigned.
* sdbout.c (sdbout_parms, sdbout_reg_parms): Don't check REGNO < 0.
* simplify-rtx.c (simplify_unary_operation): WIDTH now unsigned.
(simplify_binary_operation): Likewise.
(cselib_invalidate_regno): Arg REGNO and variables ENDREGNO, I, and
THIS_LAST now unsigned.
(cselib_record_set): Add cast.
* ssa.c (ssa_max_reg_num): Now unsigned.
(rename_block): REGNO now unsigned.
* stmt.c (expand_return): Bit positions unsigned HOST_WIDE_INT;
sizes now unsigned.
(all_cases_count): Just return -1 not -2.
COUNT, MINVAL, and LASTVAL now HOST_WIDE_INT.
Rework tests to use trees whenever possible.
Use host_integerp and tree_low_cst.
(mark_seen_cases): COUNT arg now HOST_WIDE_INT;
Likewise variable NEXT_NODE_OFFSET; XLO now unsigned.
(check_for_full_enumeration_handing): BYTES_NEEDED, I to HOST_WIDE_INT.
* stor-layout.c (mode_for_size): SIZE arg now unsigned.
(smallest_mode_for_size): Likewise.
(layout_decl): Simplify handing of a specified DECL_SIZE_UNIT.
KNOWN_ALIGN is now an alignment, so simplify code.
Don't turn off DECL_BIT_FIELD if field is BLKmode, but not type.
(start_record_layout): Renamed from new_record_layout_info.
Update to new fields.
(debug_rli, normalize_rli, rli_size_unit_so_far, rli_size_so_far):
New functions.
(place_union_field): Renamed from layout_union_field.
Update to use new fields in rli.
(place_field): Renamed from layout_field.
Major rewrite to use new fields in rli; pass alignment to layout_decl.
(finalize_record_size): Rework to use new fields in rli and handle
union.
(compute_record_mode): Rework to simplify and to use new DECL fields.
(finalize_type_size): Make rounding more consistent.
(finish_union_layout): Deleted.
(layout_type, case VOID_TYPE): Don't set TYPE_SIZE_UNIT either.
(layout_type, case RECORD_TYPE): Call new function names.
(initialize_sizetypes): Set TYPE_IS_SIZETYPE.
(set_sizetype): Set TYPE_IS_SIZETYPE earlier.
(get_best_mode): UNIT is now unsigned; remove casts.
* tree.c (bit_position): Compute from new fields.
(byte_position, int_byte_position): New functions.
(print_type_hash_statistics): Cast to remove warning.
(build_range_type): Use host_integerp and tree_low_cst to try to hash.
(build_index_type): Likewise; make subtype of sizetype.
(build_index_2_type): Pass sizetype to build_range_type.
(build_common_tree_nodes): Use size_int and bitsize_int to
initialize nodes; add bitsize_{zero,one,unit}_node.
* tree.h (DECL_FIELD_CONTEXT): Use FIELD_DECL_CHECK.
(DECL_BIT_FIELD_TYPE, DECL_QUALIFIER, DECL_FCONTEXT): Likewise.
(DECL_PACKED, DECL_BIT_FIELD): Likewise.
(DECL_FIELD_BITPOS): Deleted.
(DECL_FIELD_OFFSET, DECL_FIELD_BIT_OFFSET): New fields.
(DECL_RESULT, DECL_SAVED_INSNS): Use FUNCTION_DECL_CHECK.
(DECL_FRAME_SIZE, DECL_FUNCTION_CODE, DECL_NO_STATIC_CHAIN): Likewise.
(DECL_INLINE, DECL_BUILT_IN_NONANSI, DECL_IS_MALLOC): Likewise.
(DECL_BUILT_IN_CLASS, DECL_STATIC_CONSTRUCTOR): Likewise.
(DECL_STATIC_DESTRUCTOR, DECL_NO_CHECK_MEMORY_USAGE): Likewise.
(DECL_NO_INSTRUMENT_FUNCTION_ENTRY_EXIT, DECL_NO_LIMIT_STACK) Likewise.
(DECL_ORIGINAL_TYPE, TYPE_DECL_SUPPRESS_DEBUG): Use TYPE_DECL_CHECK.
(DECL_ARG_TYPE_AS_WRITEN, DECL_ARG_TYPE): Use PARM_DECL_CHECK.
(DECL_INCOMING_RTL, DECL_TRANSPARENT_UNION): Likewise.
(DECL_ALIGN): Adjust to new field in union.
(DECL_OFFSET_ALIGN): New field.
(DECL_ERROR_ISSUED, DECL_TOO_LATE): Use LABEL_DECL_CHECK.
(DECL_IN_TEXT_SECTION): Use VAR_DECL_CHECK.
(union tree_decl): Add struct for both aligns.
(enum tree_index): Add TI_BITSIZE_{ZERO,ONE,UNIT}.
(bitsize_zero_node, bitsize_one_node, bitsize_unit_node): Added.
(struct record_layout_info): Rework fields to have offset
alignment and byte and bit position.
(start_record_layout, place_field): Renamed from old names.
(rli_size_so_far, rli_size_unit_so_far, normalize_rli): New decls.
(byte_position, int_byte_position): Likewise.
(get_inner_reference): Change types of position and length.
* unroll.c (unroll_loop): New variable R; use for some loops.
MAX_LOCAL_REGNUM and MAXREGNUM now unsigned.
(calculate_giv_inc): Arg REGNO now unsigned.
(copy_loop_body): REGNO and SRC_REGNO now unsigned.
* varasm.c (assemble_variable): Clean up handling of size using
host_integerp and tree_low_cst.
(decode_addr_const): Use byte, not bit, position.
(output_constructor): bitpos and offsets are HOST_WIDE_INT;
use tree_low_cst and int_bit_position.
* objc/objc-act.c (build_ivar_list_initializer): Use byte_position.
* ch/actions.c (check_missing_cases): BYTES_NEEDED is HOST_WIDE_INT.
* ch/typeck.c (expand_constant_to_buffer): Use int_byte_position.
(extract_constant_from_buffer): Likewise.
* cp/class.c (build_vbase_pointer_fields): layout_field now
place_field.
(get_vfield_offset): Use byte_position.
(set_rtti_entry): Set OFFSET to ssizetype zero.
(get_binfo_offset_as_int): Deleted.
(dfs_record_base_offsets): Use tree_low_cst.
(dfs_search_base_offsets): Likewise.
(layout_nonempty_base_or_field): Reflect changes in RLI format
and call byte_position.
(layout_empty_base): Convert offset to ssizetype.
(build_base_field): use rli_size_unit_so_far.
(dfs_propagate_binfo_offsets): Do computation in proper type.
(layout_virtual_bases): Pass ssizetype to propagate_binfo_offsets.
(layout_class_type): Reflect changes in RLI names and fields.
(finish_struct_1): Set DECL_FIELD_OFFSET.
* cp/dump.c (dequeue_and_dump): Call bit_position.
* cp/expr.c (cplus_expand_constant): Use byte_position.
* cp/rtti.c (expand_class_desc): Use bitsize_one_node.
* cp/typeck.c (build_component_addr): Use byte_position and don't
special case for zero offset.
* f/com.c (ffecom_tree_canonize_ptr_): Use bitsize_zero_node.
(ffecom_tree_canonize_ref_): Likewise.
* java/class.c (make_field_value): Use byte_position.
* java/expr.c (JAVA_ARRAY_LENGTH_OFFSET): Use byte_position.
(java_array_data_offset): Likewise.
* java/java-tree.h (MAYBE_CREATE_TYPE_TYPE_LANG_SPECIFIC): Add case to
bzero call.
From-SVN: r32742
* tree.h (INT_CST_LT, INT_CST_LT_UNSIGNED): Remove unneeded casts.
(struct tree_int_cst): int_cst_low is now unsigned HOST_WIDE_INT.
(attribute_hash_list, type_hash_canon): hashcode is now unsigned.
(type_hash_lookup, type_hash_add, type_hash_list): Likewise.
(min_precision): Result is unsignd.
(add_double, neg_double, mul_double): Low word is unsigned.
(lshift_double, rshift_double, lrotate_double): Likewise.
(rrotate_double, div_and_round_double): Likewise.
(tree_floor_log2, compare_tree_int): New functions.
(preserve_rtl_expr_temps): New declaration.
* c-common.c (declare_hidden_char_array): Use compare_tree_int.
(decl_attributes): Use tree_log2 to find alignment.
Check for TREE_INT_CST_HIGH for format args.
(min_precision): Now unsigned.
Use tree_floor_log2.
(truthvalue_conversion): Delete long-disabled code.
* c-decl.c (finish_struct): Clean up tests on field width.
(finish_function): Use compare_tree_int.
* c-pragma.c (handle_pragma_token): Use tree_log2 for alignment.
* c-typeck.c (comptypes): Use tree_int_cst_equal.
(default_conversion, digest_init): Use compare_tree_int.
(build_binary_op): Use integer_all_onesp and compare_tree_int.
Fix type errors in forming masks.
* calls.c (initialize_argument_information): Use compare_tree_int.
* dbxout.c (dbxout_type): Cast TREE_INT_CST_LOW to HOST_WIDE_INT.
* except.c (expand_eh_region_start_tree): Use compare_tree_int.
* expr.c (is_zeros_p, case INTEGER_CST): Use integer_zerop.
(store_field): Use compare_tree_int.
(expand_expr, case CONSTRUCTOR): Use TYPE_SIZE_UNIT.
(expand_expr, case ARRAY_REF): Use compare_tree_int.
(do_jump, case BIT_AND_EXPR): Use tree_floor_log2.
(do_store_flag): Use compare_tree_int.
* fold-const.c (encode, decode): Low part is always unsigned.
(force_fit_type, add_double, neg_double, mul_double): Likewise.
(lshift_double, rshift_double, lrotate_double): Likewise.
(rrotate_double, div_and_round_double, int_const_binop): Likewise.
(fold_convert): Use compare_tree_int.
(operand_equal_p, case INTEGER_CST): Use tree_int_cst_equal.
(invert_truthvalue, case INTEGER_CST): Likewise.
(fold): Use compare_tree_int; add casts for unsigned TREE_INT_CST_LOW.
* mkdeps.c (deps_dummy_targets): Make I unsigned.
* rtl.h (add_double, neg_double, mul_double): Low words are unsigned.
(lshift_double, rshift_double, lrotate_double, rrotate_double):
Likewise.
* stmt.c (expand_decl): Use compare_tree_int and mode_for_size_tree.
(expand_end_case): Use compare_tree_int.
(estimate_case_costs): Cast TREE_INT_CST_LOW to HOST_WIDE_INT.
* stor-layout.c (mode_for_size_tree): Use compare_tree_int.
(layout_decl): Likewise.
(layout_record, layout_union): Make sizes unsigned.
(layout_type, case VOID_TYPE): TYPE_SIZE must be bitsizetype.
(layout_type, case QUAL_UNION_TYPE): Use compare_tree_int.
* tree.c (struct type_hash): hashcode is unsigned.
(build_type_attribute_variant, type_hash_list): Likewise.
(type_hash_lookup, type_hash_add, type_hash_canon): Likewise.
(attribute_hash_list, build_array_type, build_method_type): Likewise.
(build_complex_type): Likewise.
(real_value_from_int_cst): Remove unneeded casts.
(integer_all_onesp): Add casts.
(tree_floor_log2, compare_tree_int): New functions.
(build_index_type): Use tree_int_cst_sgn.
* varasm.c (assemble_variable): Use compare_tree_int.
* ch/actions.c (chill_convert_for_assignment): INDEX is unsigned
HOST_WIDE_INT.
* ch/ch-tree.h (DECL_NESTING_LEVEL): Use TREE_INT_CST_HIGH
since unsigned.
* ch/except.c (chill_handle_on_labels): ALTERNATIVE is unsigned.
Use compare_tree_int.
(expand_goto_except_cleanup): Likewise.
* cp/class.c (dfs_modify_vtables): I is now unsigned.
(check_bitfield_decl): Use tree_int_cst_sgn and compare_tree_int.
(build_base_field): Add casts of TREE_INT_CST_LOW to HOST_WIDE_INT.
* cp/error.c (dump_expr): Cast TREE_INT_CST_HIGH to unsigned.
* cp/init.c (build_vec_init): Cast TREE_INT_CST_LOW to HOST_WIDE_INT.
* cp/method.c (build_overload_int): Cast TREE_INT_CST_HIGH to unsigned.
* cp/typeck.c (build_binary_op, case TRUNC_DIV_EXPR):
Call integer_all_onesp.
* cp/typeck2.c (process_init_constructor): Use compare_tree_int.
* f/com.c (ffecom_f2c_set_lio_code_): Use compare_tree_int.
(ffecom_sym_transform_, ffecom_transform_common_): Likewise.
(ffecom_transform_equiv_): Likewise.
* java/decl.c (emit_init_test_initialization): Mark KEY as unused.
* java/expr.c (build_newarray): Cast TREE_INT_CST_LOW to HOST_WIDE_INT.
(build_anewarray): Likewise.
* java/parse.y (patch_newarray): Likewise.
* java/parse.c: Regenerated.
From-SVN: r32383
* c-common.c (c_common_nodes_and_builtins): Make sizetype_endlink
reference the language-equivalent of sizetype.
* c-typeck.c (comptypes): Treat sizetype like its language equivalent.
* fold-const.c (size_binop, size_diffop): Put back checks.
* gcse.c (dump_hash_table): Fix minor error in last change.
* stor-layout.c (set_sizetype): Set TYPE_DOMAIN of sizetype.
Clear TYPE_{NEXT,MAIN}_VARIANT and TYPE_{POINTER,REFERENCE}_to of
all sizetypes.
* cp/typeck.c (comptypes): Treat sizetype like its language equivalent.
From-SVN: r32291
* builtins.c (c_strlen): Use size_diffop and return ssizetype value.
(expand_builtin_strcpy): Pass correct type to size_binop.
(expand_builtin_strcmp): Likewise.
Clean up conditional structure.
* c-decl.c (init_decl_processing): Don't call set_sizetype twice.
(complete_array_type): Don't use size_binop for MAXINDEX.
* c-typeck.c (c_sizeof): Use size_one_node and TYPE_SIZE_UNIT.
(c_sizeof_nowarn, c_size_in_bytes): Likewise.
(c_alignof): Use size_one_node.
(build_unary_op): Pass arg of proper type to size_binop.
(really_start_incremental_init, push_init_level): Use sizetype for
constructor{,_bit,_unfilled}_index.
(pop_init_label, output_init_element): Likewise.
(output_pending_init_elements, process_init_element): Likewise.
* calls.c (compute_argument_block_size): Field VAR is ssizetype.
* expr.c (store_expr): Use size_int.
(store_constructor): Use proper types for size_binop args.
(get_inner_reference, expand_expr, case ARRAY_REF): Likewise.
(expand_expr_unaligned): Likewise.
(string_contant): Return object of sizetype.
* expr.h (SUB_PARM_SIZE): Call size_diffop and pass proper types.
(ARGS_SIZE_RTX): Call ARGS_SIZE_TREE.
(ARGS_SIZE_TREE): Pass proper types to size_binop.
* fold-const.c (int_const_binop): Refine when size_int is called.
(fold_convert): Likewise.
(size_int_wide): Rework to take KIND as arg, only take low order
bits, handle new sizetype_tab datatype, and chain entries in
size_table.
(size_int_type_wide): New function.
(size_binop): Validate types of arguments.
(ssize_binop): Deleted.
(size_diffop): New function.
(extract_muldiv): Only fold division into multiplication for sizetypes.
* function.c (assign_parms): Use size_diffop and make sure
VAR field is of ssizetype; also pass proper type to size_binop.
(locate_and_pad_parm, pad_to_arg_alignment): Likewise.
(round_down): Deleted from here.
* store-layout.c (sizetype_tab): Now an array.
(sizetype_set, early_root_list): New variables.
(variable_size): Use size_one_node.
(round_up): Pass proper type to size_binop.
(round_down): Moved to here and corrected as above.
(layout_record): Pass proper arg types to size_binop.
(layout_type): Likewise.
If sizetype_set is zero, record the type just laid out.
(make_unsigned_type): Don't call set_sizetype;
(make_signed_type): Likewise; also, call fixup_signed_type.
(initialize_sizetypes): New function.
(set_sizetype): Make copy of types, set TYPE_IS_SIZETYPE, and
set name of bitsizetype to "bit_size_type".
Fix up type of sizes of all types made before call.
* tm.texi (ROUND_TYPE_SIZE_UNIT): New macro.
* tree.c (fix_sizetype): Deleted.
(build_common_tree_nodes): Call initialize_sizetypes.
(build_common_tree_nodes_2): Don't call fix_sizetype.
* tree.h (TYPE_IS_SIZETYPE): New macro.
(initialize_sizetype): New declaration.
(enum size_type_kind): New type.
(struct sizetype_tab): Deleted.
(sizetype_tab): Now array; adjust sizetype macros.
(size_diffop, size_int_type_wide): New functions.
(size_int_wide): Change number of args and type; access macros changed.
(ssize_int, sbitsize_int): New macros.
* config/i960/i960.h (ROUND_TYPE_SIZE): Use size_int.
(ROUND_TYPE_SIZE_UNIT): New macro.
* ch/actions.c (chill_convert_for_assignment): Don't use size_binop
for things that aren't sizes.
(expand_varying_length_assignment): Likewise.
* ch/convert.c (digest_array_tuple, convert): Likewise.
* ch/typeck.c (build_chill_slice, smash_dummy_type): Likewise.
(build_chill_slice_with_range): Likewise.
(build_chill_slice_with_length): Likewise.
(build_array_from_set): Adjust types for size_binop.
* ch/expr.c (build_concat_expr, build_chill_repetition_op): Likewise.
(build_chill_sizeof): Use TYPE_SIZE_UNIT.
* ch/tree.c (build_string_type): Pass proper type to size_binop.
* cp/class.c (dfs_build_vtable_offset_vtbl_entries): Don't use
size_binop on things that are not sizes; ssize_binop deleted.
Call size_diffop when appropriate.
(dfs_build_vcall_offset_vtbl_entries): Likewise.
(build_primary_vtable, build_secondary_vtable): Likewise.
(dfs_set_offset_for_unshared_vbases, dfs_modify_vtables): Likewise.
Variable I is HOST_WIDE_INT.
(get_vfield_offset): Pass proper types to size_binop.
(size_extra_vtbl_entries, layout_virtual_bases): Likewise.
(finish_struct_1): Likewise.
(skip_rtti_stuff): Arg N is now pointer to signed.
(layout_class_type): Use size_zero_node.
* cp/cp-tree.h (skip_rtti_stuff): Arg N is pointer to signed.
* cp/cvt.c (cp_convert_to_pointer): Pass proper types to size_binop.
* cp/decl.c (complete_arry_type): Pass proper types to size_binop.
(xref_basetypes): BINFO_OFFSET is sizetype.
* cp/error.c (dump_expr): Don't use size_binop non-sizes.
* cp/expr.c (cplus_expand_constant): Pass proper types to size_binop.
* cp/init.c (construct_virtual_bases): Fix type error.
(build_vec_delete_1): Pass proper type to size_binop and don't
fold result.
* cp/lex.c (cp_make_lang_type): BINFO_OFFSET is sizetype.
* cp/rtti.c (get_base_offset): Pass proper type to size_binop.
* cp/search.c (dfs_find_vbases): Fix type error.
(expand_upcast_fixups): Arg to skip_rtti_stuff is pointer to signed.
(dfs_get_vbase_types): BINFO_OFFSET is sizetype.
* cp/tree.c (debug_binfo): Variable N is signed.
Use HOST_WIDE_INT_PRINT_DEC.
* cp/typeck.c (comptypes): sizetype is same as equivalent integer type.
(c_sizeof, c_sizeof_nowarn, expr_sizeof): Use TYPE_SIZE_UNIT,
size_one_node and size_zero_node.
(c_alignof): Use size_one_node.
(build_component_addr): Pass proper types to size_binop.
(expand_ptrmemfunc_cst): Don't use size_binop on non-sizes.
* f/com.c (ffecom_arrayref_): Convert args to size_binop to proper
type.
(ffecom_tree_canonize_ptr_): Don't use size_binop for non-sizes.
(ffecom_tree_canonize_ref_): Likewise.
(type_for_mode): Handle TImode.
* f/ste.c (ffeste_io_dofio_, ffeste_io_douio_): Use TYPE_SIZE_UNIT.
(ffeste_io_ciclist_): Likewise.
* java/expr.c (build_java_ret): Pass proper type to size_binop.
From-SVN: r32225
* c-common.c (decl_attributes): Set DECL_SIZE_UNIT.
* c-decl.c (duplicate_decls, finish_enum): Likewise.
(finish_decl): Remove -Wlarger-than code from here.
* flags.h (id_clash_len): Now int.
(larger_than_size): Now HOST_WIDE_INT.
* fold-const.c (size_int_wide): No more HIGH parm; NUMBER is signed.
Clean up checking to see if in table.
(make_bit_field_ref): Remove extra parm to bitsize_int.
* ggc-common.c (ggc_mark_tree_children): Mark DECL_SIZE_UNIT.
* print-tree.c (print_node): Print DECL_SIZE_UNIT and TYPE_SIZE_UNIT.
* stmt.c (expand_decl): Use DECL_SIZE_UNIT for stack checking size
and for computing size of decl.
* stor-layout.c (layout_decl): Set DECL_SIZE_UNIT.
Move -Wlarger-than code to here.
(layout_record): Remove extra arg to bitsize_int.
Set TYPE_BINFO_SIZE_UNIT.
(layout_union): Remove extra arg to bitsize_int.
Use proper type for size of QUAL_UNION.
(layout_type): Remove extra arg to bitsize_int.
* toplev.c (id_clash_len): Now int.
(larger_than_size): Now HOST_WIDE_INT.
(decode_W_option): Clean up id-clash and larger-than- cases.
* tree.c (get_identifier, maybe_get_identifier): Remove unneeded casts.
(expr_align, case FUNCTION_DECL): DECL_ALIGN is not defined.
* tree.h (BINFO_SIZE_UNIT, TYPE_BINFO_SIZE_UNIT, DECL_SIZE_UNIT): New.
(struct tree_decl): New field size_unit.
(size_int_wide): No HIGH operand; NUMBER is now signed.
(size_int_2): Deleted.
(size_int, bitsize_int): Don't use it and rework args.
* varasm.c (assemble_variable, output_constructor): Use DECL_SIZE_UNIT.
* ch/decl.c (layout_enum): Set DECL_SIZE_UNIT.
* ch/satisfy.c (safe_satisfy_decl): Likewise.
* cp/class.c (build_primary_vtable, layout_vtable_decl): Likewise.
(avoid_overlap, build_base_field): Likewise.
(build_base_field, build_base_fields, is_empty_class):
Test DECL_SIZE with integer_zero.
(layout_class_type): Set CLASSTYPE_SIZE_UNIT.
* cp/cp-tree.h (struct lang_type): New field size_unit.
(CLASSTYPE_SIZE_UNIT): New macro.
* cp/decl.c (init_decl_processing): Set DECL_SIZE_UNIT.
(cp_finish_decl): Delete -Wlarger-than processing.
* cp/optimize.c (remap_decl): Walk DECL_SIZE_UNIT.
* cp/pt.c (tsubst_decl): Set DECL_SIZE_UNIT.
* cp/tree.c (make_binfo): binfo vector is one entry longer.
(walk_tree): Walk DECL_SIZE_UNIT.
* f/com.c (ffecom_sym_transform): Use DECL_SIZE_UNIT.
(ffecom_transform_common_, ffecom_transform_equiv_): Likewise.
(duplicate_decls): Likewise.
(ffecom_tree_canonize_ptr_): Delete extra arg to bitsize_int.
(finish_decl): Delete -Wlarger-than processing.
* java/class.c (build_class_ref, push_super_field): Set DECL_SIZE_UNIT.
* java/constants.c (build_constants_constructor): Likewise.
From-SVN: r32068
* alias.c: Minor reformatting.
* flow.c: Likewise.
* regs.h: Likewise.
* stor-layout.c: Likewise.
* fold-const.c: Likewise.
(OVERFLOW_SUM_SIGN): Renamed from overflow_sum_sign.
(struct cb_args, const_binop_1, const_binop): Pass type of arg,
not arg itself.
(size_int_wide): Cache nodes even if garbage collecting.
(twoval_comparison_p): Reenable SAVE_EXPR case if operand
of SAVE_EXPR has no side effects.
* cse.c: Move a comment.
* tree.c: Minor reformatting.
(int_size_in_bytes): Return -1 if constant overflows.
* reload.c (combine_reloads): Do nothing if no output reload
From-SVN: r31017
* fold-const.c (extract_muldiv, case {MIN,MAX}_EXPR): Reverse
operation if C is negative.
(extract_muldiv, case SAVE_EXPR): Supresss if arg has side effects.
(extract_muldiv, case {PLUS,MINUS}_EXPR): Don't apply distributive
law for some divisions if constant is negative and change other
divisions to the opposite rounding.
From-SVN: r30714
* fold-const.c (negate_expr, associate_trees, extract_muldiv): New.
(split_tree): Completely rework to make more general.
(make_range, fold): Call negate_expr.
(fold, case NEGATE_EXPR): Simplify -(a-b) is -ffast-math.
(fold, associate): Call new split_tree and associate_trees.
(fold, case MULT_EXPR, case *_{DIV,MOD}_EXPR): Call extract_muldiv.
From-SVN: r30673
Fri Oct 29 15:25:07 1999 Arnaud Charlet <charlet@ACT-Europe.FR>
* gcov.c (DIR_SEPARATOR): Provide default.
(output_data): Add test for MS-DOS format absolute filename.
(fancy_abort): Correct program name.
(open_files): Open all files in binary mode.
* libgcc2.c (__bb_exit_func): Likewise.
* profile.c (init_branch_prob): Specify binary when opening files.
* flags.h (flag_unwind_tables): New decl.
* toplev.c (flag_unwind_table): New definition.
(f_options): Add -funwind-tables.
(decode_g_option): Clarify warning when unknown -g option is given.
(rest_of_compilation): If inside an inlined external function,
pretend we are just being declared.
* dwarf2out.c (dwarf2out_do_frame): Check -funwind_tables.
(dwarf2out_frame_finish): Likewise.
Fri Oct 29 06:32:44 1999 Geoffrey Keating <geoffk@cygnus.com>
* flow.c (propagate_block): When the last reference to a label
before an ADDR_VEC is deleted because the reference is a dead
store, delete the ADDR_VEC.
Fri Oct 29 07:44:26 1999 Vasco Pedro <vp@di.fct.unl.pt>
* fold-const.c (merge_ranges): In not in0, but in1, handle
upper bounds equal like subset case.
Thu Oct 28 19:22:24 1999 Douglas Rupp <rupp@gnat.com>
* dbxout.c (dbxout_parms): Generate a second stabs line for parameters
passed in a register but moved to the stack.
Thu Oct 28 19:12:57 1999 Sam Tardieu <tardieu@act-europe.fr>
* gcc.c (pass_exit_codes, greatest_status): New variables.
(struct option_map): Add entry for "--pass-exit-codes".
(execute): Update greatest_status if error.
(display_help): Add documentation for -pass-exit-codes.
(process_command): Handle -pass-exit-codes.
(main): Look at pass_exit_codes and greatest_status on call to exit.
Thu Oct 28 18:06:50 1999 Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
* reload.c (find_reloads): Refine test for no input reload
case to not includes reloads emitted after insn.
* function.c (find_temp_slots_from_address): Handle sum involving
a register that points to a temp slot.
(update_temp_slot_address): Make recursive call if both old and
new are PLUS with a common operand.
* calls.c (expand_call): Mark temp slot for result as having
address taken.
* rtlanal.c (reg_referenced_p, case IF_THEN_ELSE): New case.
* gcc.c (process_command): Add standard_exec_prefix with "GCC"
component as well as "BINUTILS".
* integrate.h (copy_rtx_and_substitute): New arg, FOR_LHS.
* integrate.c (copy_rtx_and_substitute): Likewise.
(expand_inline_function, integrate_parm_decls, integrate_decl_tree):
All callers changed.
* unroll.c (inital_reg_note_copy, copy_loop_body): Likewise.
* dbxout.c (dbxout_type, case INTEGER_TYPE_NODE): If can use
gdb extensions, write size of type; also be more consistent
in using references when this is a subtype.
* pa.md (extv, extzv, insv): Use define_expand to reject constant
that is out of range.
* loop.c (unknown_constant_address_altered): New variable.
(prescan_loop): Initialize it.
(note_addr_stored): Set it for RTX_UNCHANGING_P MEM.
(invariant_p, case MEM): Remove handling for volatile and readony;
check new variable if readonly.
(check_dbra_loop): Chdeck unknown_constant_address_altered.
* cse.c (canon_hash, case MEM): Do not record if BLKmode.
(addr_affects_sp_p): Removed from note_mem_written and only
define #ifdef AUTO_INC_DEC.
* alpha.c (input_operand, case ADDRESSOF): Treat as REG.
* regclass.c (record_reg_classes): Properly handle register move
directions.
* varasm.c (initializer_constant_valid_p, case MINUS_EXPR):
Don't think valid if both operands are invalid.
(struct constant_descriptor): New field RTL.
(mark_const_hash_entry): Mark it.
(record_constant{,_rtx}): Initialize it.
(output_constant_def): Allocate RTL in permanent obstack and
save in table.
({record,compare}_constant_1): Modes must match for
CONSTRUCTOR of ARRAY_TYPE.
* c-common.h (initializer_constant_valid_p): Delete decl from here.
* output.h (initializer_constant_valid_p): Move decl to here.
* c-common.c (initializer_constant_valid_p): Delete function from here.
* varasm.c (initializer_constant_valid_p): Move function to here.
* tree.h (STRIP_SIGN_NOPS): New macro.
* fold-const.c (optimize_minmax_comparison): New function.
(invert_truthvalue, case WITH_RECORD_EXPR): New case.
(fold): Use STRIP_SIGN_NOPS instead of STRIP_TYPE_NOPS.
(fold, case EQ_EXPR): Call optimize_minmax_comparison and add
cases with ABS_EXPR, NEGATE_EXPR, PLUS_EXPR, MINUS_EXPR, and
widening conversions.
(fold, case LE_EXPR): Rework changing unsigned to signed comparisons
to look at size of mode, not precision of type; also add missing cases.
(optimize_bit_field_compare, decode_field_reference): Don't try to
optimize COMPONENT_REF of a PLACEHOLDER_EXPR.
* dwarf2out.c (ctype.h): Include.
(dwarf2out_set_demangle_name_func): New function.
(size_of_line_info): Deleted.
(output_line_info): Compute size of line info table from difference
of labels.
(base_type_die, add_name_attribute): Call demangle function, if any.
(field_byte_offset): Use bits per word for variable length fields.
(gen_array_type_die): Add array name.
(gen_subprogram_die): Ignore DECL_INLINE if -fno-inline.
(dwarf2out_add_library_unit_info): New function.
* explow.c (set_stack_check_libfunc): New function.
(stack_check_libfunc): New static variable.
(probe_stack_range): Allow front-end to set up a libfunc to call.
* combine.c (simplify_comparison): When making comparison in wider
mode, check for having commuted an AND and a SUBREG.
(contains_muldiv): New function.
(try_combine): Call it when dividing a PARALLEL.
(simplify_rtx, case TRUNCATE): Don't remove for umulsi3_highpart.
(simplify_comparison, case ASHIFTRT): Recognize sign-extension of
a PLUS.
(record_value_for_reg): If TEM is a binary operation with two CLOBBERs,
use one of the CLOBBERs instead.
(if_then_else_cond): If comparing against zero, just return thing
being compared.
* optabs.c (expand_abs): If machine has MAX, ABS (x) is MAX (x, -x).
Don't generate shifts and subtract if have conditional arithmetic.
* rtl.h (delete_barrier): New declaration.
* jump.c (jump_optimize): Set up to handle conditional call.
In conditional arithmetic case, handle CALL_INSN followed by a BARRIER.
(delete_barrier): New function.
* rtl.c (read_rtx): Call fatal if bad RTL code; check for bad mode.
* recog.c (nonmemory_operand): Accept ADDRESSOF.
* tree.c (build_type_attribute_variant): Push to obstack of
ttype around type_hash_canon call.
* expr.c (placeholder_list): Move decl to file scope.
(expand_expr): Don't force access to volatile just because its
address is taken.
If ignoring reference operations, just expand the operands.
(expand_expr, case COMPONENT_REF): Propagate
EXPAND_CONST_ADDRESS to recursive call when expanding inner.
Refine test for using bitfield operations vs pointer punning.
(expand_expr, case CONVERT_EXPR): If converting to
BLKmode UNION_TYPE from BLKmode, just return inner object.
Use proper mode in store_field call.
Properly set sizes of object to store and total size in store_field
call for convert to union.
(expand_expr, case ARRAY_REF): If OP0 is in a register, put it in
memory (like for ADDR_EXPR). Also, don't put constant in register if
we'll want it in memory.
(readonly_fields_p): New function.
(expand_expr, case INDIRECT_REF): Call it if LHS.
(expand_assignment): Handle a RESULT_DECL where
DECL_RTL is a PARALLEL.
(do_jump, case WITH_RECORD_EXPR): New case.
(get_inner_reference): Always go inside a CONVERT_EXPR
and NOP_EXPR if both modes are the same.
(store_field): Use bitfield operations if size of bitsize is not same
as size of RHS's type.
Check for bitpos not a multiple of alignment in BLKmode case.
Do block move in largest possible alignment.
(store_constructor): Set BITSIZE to -1 for variable size and properly
in case of array of BLKmode.
(expand_expr_unaligned): New function.
(do_compare_and_jump): Call it.
* mips/iris5.h (SWITCHES_NEED_SPACES): New macro.
* collect2.c (main): Only allow -ofoo if SWITCHES_NEED_SPACES
does not include 'o'.
* function.c (instantiate_virtual_regs_1, case SET): Handle case where
both SET_DEST and SET_SRC reference a virtual register.
(gen_mem_addressof): Copy RTX_UNCHANGING_P from new REG to old REG.
* integrate.c (expand_inline_function): Handle case of setting
virtual stack vars register (from built in setjmp); when parameter
lives in memory, expand virtual_{stack_vars,incoming_args}_rtx early.
(subst_constant): Add new parm, MEMONLY.
(expand_inline_function, integrate_parm_decls): Pass new parm.
(integrate_decl_tree): Likewise.
(copy_rtx_and_substitute, case MEM): Do copy RTX_UNCHANGING_P.
(try_constants): Call subst_constants twice, with MEMONLY 0 and 1.
(copy_rtx_and_substitute, case SET): Add explicit calls to
copy_rtx_and_substitute for both sides.
* stmt.c (expand_asm_operands): Don't use TREE_STRING_LENGTH for
constraints.
(pushcase{,_range}): Convert to NOMINAL_TYPE after checking for
within INDEX_TYPE, instead of before.
(fixup_gotos): Use f->target_rtl, not the next insn,
since latter may be from a later fixup.
(expand_value_return): Correctly convert VAL when promoting function
return; support RETURN_REG being a PARALLEL.
(expand_return): When checking for result in regs and having
cleanup, consider PARALLEL in DECL_RTL as being in regs.
From-SVN: r30299
* fold-const.c (fold): Move bit_rotate code to the EXPR_PLUS case,
falltrought to assocate code.
Convert XOR to OR in code like (a&c1)^(a&c2) where c1 and c2 don't have
bits in common.
* combine.c (simplify_logical): Convert XOR to IOR if operands have
no bits in common; remove XOR to ROTATE conversion.
From-SVN: r29998
* Makefile.in (ggc-simple.o): Depend on varray.h.
(rtl.o): Depend on ggc.h.
(genattrtab.o): Depend on ggc.h.
(print-tree.o): Likewise.
(fold-const.o): Likewise.
* emit-rtl.c (sequence_element_free_list): Remove, and all references.
(make_insn_raw): Don't cache insns when GC'ing.
(emit_insn_before): Likewise.
(emit_insn_after): Likewise.
(emit_insn): Likewise.
(start_sequence): Use xmalloc to allocate the sequence_stack.
(end_sequence): Add free to free it.
(gen_sequence): Don't cache insns when GC'ing.
(clear_emit_caches): Don't use sequence_element_free_list.
(init_emit): Use xcalloc, not xmalloc+bzero.
* fold-const.c (size_int_wide): Kill the cache, when GC'ing.
* function.c (pop_function_context_from): Use free to free the
fixup_var_refs_queue.
(put_reg_into_stack): Allocate it with xmalloc.
* genattrtab.c: Include ggc.h.
(operate_exp): Don't use obstack_free when GC'ing.
(simplify_cond): Likewise.
(simplify_text_exp): Likewise.
(optimize_attrs): Likewise.
* gengentrtl.c (gendef): Use ggc_alloc_rtx to allocate RTL, when
GC'ing.
(gencode): Generate a #include for ggc.h.
* ggc-callbacks.c (ggc_p): Define it to zero.
* ggc-none.c (ggc_p): Likewise.
* ggc-simple.c: Include varray.h.
(ggc_mark_tree_varray): New function.
(ggc_add_tree_varray_root): Likewise.
(ggc_mark_tree_varray_ptr): Likewise.
* ggc.h (ggc_p): Declare.
(varray_head_tag): Likewise.
(ggc_add_tree_varray_root): Declare.
* print-tree.c (print_node): Don't check for TREE_PERMANENT
inconsistencies when GC'ing.
* rtl.c: Include ggc.h.
(rtvec_alloc): Use ggc_alloc_rtvec when GC'ing.
(rtx_alloc): Use ggc_alloc_rtx when GC'ing.
(rtx_free): Don't call obstack_free when GC'ing.
* toplev.c (rest_of_compilation): Call ggc_collect after every
pass, if GC'ing.
* tree.c (push_obstacks): Do nothing, if GC'ing.
(pop_obstacks_nochange): Likewise.
(pop_obstacks): Likewise.
(make_node): Use ggc_alloc_tree when GC'ing.
(copy_node): Likewise.
(get_identifier): Use ggc_alloc_string when GC'ing.
(build_string): Likewise.
(make_tree_vec): Use ggc_alloc_tree when GC'ing.
(tree_cons): Likewise.
(build1): Likewise.
(type_hash_canon): Don't call obstack_free when GC'ing.
Co-Authored-By: Bernd Schmidt <bernds@cygnus.co.uk>
Co-Authored-By: Mark Mitchell <mark@codesourcery.com>
From-SVN: r29125
1999-08-27 00:27 -0700 Zack Weinberg <zack@bitmover.com>
* c-typeck.c (c_expand_start_case): Return immediately if exp
is an ERROR_MARK.
* fold-const.c (operand_equal_p): Return immediately if arg1
or arg0 are ERROR_MARKs.
* stor-layout.c (layout_type [case RECORD_TYPE]): Ignore
fields of type ERROR_MARK when calculating if the record can
go in a register.
From-SVN: r28923
* fold-const.c (fold): Reassociate (+ (+ (* a b) c) (* d e))
as (+ (+ (* a b) (* d e)) c). Factor a common power-of-two
multiplicand out of (+ (* a b) (* c d)).
From-SVN: r28829
* fold-const.c (fold_truthop): Verify that the lhs and rhs are
in the same bit position when optimizing bitfield references
which have the same mask.
From-SVN: r25993
* fold-const.c (make_range): If orig_type is unset, set it as soon
as we know the type. Remove now unnecessary set of orig_type for
conversions.
Co-Authored-By: Jeffrey A Law <law@cygnus.com>
From-SVN: r25959
* fold-const.c (exact_real_inverse): Move variable `float_error'
into the scope where it is used.
(const_binop_1): New static function.
(cb_args): New struct.
(const_binop): Use them in call to `do_float_handler'.
(fold_convert_1): New static function.
(fc_args): New struct.
(fold_convert): Use them in call to `do_float_handler'.
From-SVN: r25776
* c-common.c (declare_function_name): Declare predefinied variable
`__func__'.
* c-decl.c (flag_isoc9x): Set to 1 by default.
(c_decode_option): Handle -std= option. Remove -flang-isoc9x.
(grokdeclarator): Always emit warning about implicit int for ISO C 9x.
* c-parse.in: Allow constructors in ISO C 9x.
Rewrite designator list handling.
Allow [*] parameters.
Don't warn about comma at end of enum definition for ISO C 9x.
* cccp.c (c9x): New variable.
(rest_extension): New variable.
(print_help): Document new -std= option.
(main): Recognize -std= option. Set c9x appropriately.
(create_definition): Recognize ISO C 9x vararg macros.
* gcc.c (default_compilers): Adjust specs for -std options.
(option_map): Add --std.
(display_help): Document -std.
* toplev.c (documented_lang_options): Add -std and remove
-flang-isoc9x.
* c-lex.c (yylex): Recognize hex FP constants and call REAL_VALUE_ATOF
or REAL_VALUE_HTOF based on base of the constants.
* fold-const.c (real_hex_to_f): New function. Replacement function
for hex FP conversion if REAL_ARITHMETIC is not defined.
* real.c (asctoeg): Add handling of hex FP constants.
* real.h: Define REAL_VALUE_HTOF if necessary using ereal_atof or
real_hex_to_f.
Co-Authored-By: Richard Henderson <rth@cygnus.com>
Co-Authored-By: Stephen L Moshier <moshier@world.std.com>
From-SVN: r24049
* Makefile.in (gencheck.o): Depend on gansidecl.h.
* c-common.c (print_char_table): Add missing initializers.
(scan_char_table): Likewise.
(time_char_table): Likewise.
* c-decl.c (c_decode_option): Mark parameter `argc' with
ATTRIBUTE_UNUSED.
(declare_parm_level): Mark parameter `definition_flag' with
ATTRIBUTE_UNUSED.
* c-lex.c (readescape): Use `(unsigned)1' in shift.
(yylex): Likewise. Cast `sizeof' to an (int) when comparing
against one.
* calls.c (store_one_arg): Remove unused parameter `fndecl'. All
callers changed.
(emit_call_1): Mark parameters `fndecl' and `funtype' with
ATTRIBUTE_UNUSED.
(expand_call): Cast result of MIN() to (unsigned int) when
comparing against an unsigned value.
* cccp.c (pcfinclude): Remove unused parameter `limit'. All
callers changed.
(make_definition): Remove unused parameter `op'. All callers
changed.
(create_definition): Cast REST_EXTENSION_LENGTH to (long) when
comparing against the result of pointer arithmetic.
* config/mips/mips.h (FUNCTION_ARG_BOUNDARY): Cast to (unsigned)
when comparing against one.
* dwarf2out.c (dwarf2out_frame_debug): Cast REGNO() and
HARD_FRAME_POINTER_REGNUM to (unsigned) when comparing against
one.
(output_die): Move variable `i' into the scope in which it is
used. Change its type to `unsigned'.
(output_die): Cast the result of `strlen' to (int) when passing it
to ASM_OUTPUT_ASCII().
(output_pubnames): Likewise.
(output_line_info): Likewise.
* emit-rtl.c (global_rtl): Add missing initializers.
* explow.c (promote_mode): Mark parameter `for_call' with
ATTRIBUTE_UNUSED.
* expmed.c (expand_shift): Cast the result of GET_MODE_BITSIZE to
`unsigned HOST_WIDE_INT' when comparing against one.
(synth_mult): Change type of variable `cost' to int.
(emit_store_flag): Use `(unsigned HOST_WIDE_INT) 1' in shift.
* expr.c (copy_blkmode_from_reg): Cast BITS_PER_WORD to (unsigned)
when comparing against one.
(get_inner_reference): Change variable `alignment' to unsigned.
(expand_expr): Cast the result of GET_MODE_ALIGNMENT to (unsigned
int) when comparing against one.
(expand_builtin_setjmp): Change type of variable `i' to size_t.
* fold-const.c (div_and_round_double): Cast BASE to
(HOST_WIDE_INT) when comparing against one.
* gencheck.c: Include gansidecl.h.
(main): Mark parameter `argv' with ATTRIBUTE_UNUSED.
* optabs.c (gen_cond_trap): Mark parameters `code', `op2' and
`tcode' with ATTRIBUTE_UNUSED.
* real.c (edivm): Cast constant value to (unsigned long) in
expression compared against an unsigned value.
* stmt.c (expand_return): Cast BITS_PER_WORD to (unsigned) when
comparing against one.
(expand_end_case): Cast CASE_VALUES_THRESHOLD to (unsigned int)
when comparing against one.
* stor-layout.c (mode_for_size): Cast MAX_FIXED_MODE_SIZE to
(unsigned int) when comparing against one. Likewise for
GET_MODE_BITSIZE.
(smallest_mode_for_size): Likewise.
(save_storage_status): Mark parameter `p' with ATTRIBUTE_UNUSED.
(restore_storage_status): Likewise.
* toplev.c (debug_args): Add missing initializer.
(f_options): Spelling correction. Add missing initializers.
(documented_lang_options): Likewise.
(debug_end_source_file): Mark parameter `lineno' with
ATTRIBUTE_UNUSED.
* tree.c (valid_machine_attribute): Mark parameters `attr_args',
`decl' and `type' with ATTRIBUTE_UNUSED.
* varasm.c (decode_reg_name): Cast `sizeof' expression to (int)
when comparing against one.
(assemble_variable): Mark parameter `top_level' with
ATTRIBUTE_UNUSED.
(assemble_external_libcall): Mark parameter `fun' with
ATTRIBUTE_UNUSED.
(output_constant_pool): Mark parameters `fnname' and `fndecl' with
ATTRIBUTE_UNUSED.
From-SVN: r23054
* Makefile.in (cplus-dep.o): Use cplus-dem.c from libiberty.
* cplus-dem.c: Delete.
* Makefile.in (fold-const.o): depend on $(RTL_H).
* fold-const.c: Include rtl.h to get the prototype for
`set_identifier_local_value'.
* loop.c (express_from_1): Remove unused variable `tmp'.
(combine_givs): cast the first argument of bzero to char *.
* toplev.c (display_help): Remove unused variable `looking_for_start'.
Co-Authored-By: Jeffrey A Law <law@cygnus.com>
From-SVN: r21785
* expr.c (safe_from_p): Change code to ERROR_MARK only when not
accessing nodes.
* toplev.c (display_help): Add braces to shut up warnings.
* fold-const.c (non_lvalue): Don't deal with null pointer
constants here.
(fold, case COMPOUND_EXPR): Wrap a constant 0 in a NOP_EXPR.
From-SVN: r21698
* loop.h (struct induction): Add no_const_addval.
* loop.c (the_movables, reg_address_cost): New variables.
(init_loop): Init reg_address_cost.
(loop_optimize): Call end_alias_analysis.
(scan_loop): Init the_movables.
(record_giv): Init induction->no_const_addval.
(basic_induction_var) [PLUS]: Use rtx_equal_p instead of ==.
[REG]: Rearrange loop search test to catch more cases.
(general_induction_var): Return success not benefit; take an extra
argument for that. Change all callers.
(simplify_giv_expr) [PLUS]: Always combine invariants. Use sge_plus.
[MULT]: Use rtx_equal_p instead of ==. Combine simple invariants.
[default]: Search the_movables for additional combinations.
(sge_plus_constant, sge_plus): New functions.
(express_from_1): New function.
(express_from): Always define. Rewrite using express_from_1.
(combine_givs_p): Handle more cases. Ignore address cost.
(cmp_combine_givs_stats): New function.
(combine_givs_used_once, combine_givs_benefit_from): New functions.
(combine_givs): Rewrite to do best-fit combination.
* fold-const.c (operand_equal_p): Handle RTL_EXPR.
(fold): Do a complete (A*C)+(B*C) association check.
From-SVN: r21263
* c-common.c (truthvalue_conversion): Protect side effects in the
expression when splitting a complex value.
* fold-const.c (fold): Likewise.
From-SVN: r20580
* gencheck.c: Remove redundant stdio.h include. Add a definition
of xmalloc for when we are forced to link with alloca.o.
* reload1.c (reload_reg_free_for_value_p): Use `(unsigned)1'
instead of `1U'.
* fold-const.c (constant_boolean_node): Make definition static to
match the prototype.
From-SVN: r20261
* fold-const.c (fold, case EQ_EXPR): When folding VAR++ == CONST
or VAR-- == CONST construct a proper mask if VAR is a bitfield.
Cope with CONST being out of range for the bitfield.
From-SVN: r20198
* tree.h (sizetype_tab[2], sbitsizetype, ubitsizetype): Merge all
of these into a single struct, with additional [us]sizetype entries.
* stor-layout.c (set_sizetype): Initialize [us]sizetype.
* fold-const.c (size_int_wide): Don't rely on sizetype_tab being
an array.
From-SVN: r18994
* varasm.c (make_decl_rtl): Update the DECL_ASSEMBLER_NAME for a
entity in a local scope.
* fold-const.c (fold): Call truthvalue_conversion for values which
are folded to boolean type.
From-SVN: r18929
Wed Mar 18 13:46:07 1998 Richard Kenner <kenner@vlsi1.ultra.nyu.edu>
* fold-const.c (operand_equal_for_comparison_p): See if equal
when nop conversions are removed.
From-SVN: r18672
* fold-const.c: Include "system.h" to get stdlib.h and stdio.h.
(lshift_double): Add parentheses around + or - inside shift.
(rshift_double): Likewise.
(size_int_wide): Explicitly set type of `bit_p' to `int'.
* Makefile.in (fold-const.o): Depend on system.h.
From-SVN: r18057
* tree.c (first_rtl_op): New fn.
(unsave_expr_now): Use it.
* print-tree.c (print_node): Likewise.
* tree.c (has_cleanups): New fn.
* fold-const.c (fold, case CLEANUP_POINT_EXPR): Use it. Be more
conservative about pushing the cleanup point down.
* tree.h: Declare them.
From-SVN: r18023
* expr.c (get_inner_reference): Use sbitsizetype for type sizes.
* fold-const.c (size_int): Replace with
(size_int_wide).
(make_bit_field_ref): Use bitsize_int for bit position.
* stor-layout.c (sizetype): Delete.
(sizetype_tab, sbitsizetype, ubitsizetype): Declare.
(layout_record, layout_union, layout_type):
Use bitsize_int for bit size.
(set_sizetype): New function.
(make_signed_type, make_unsigned_type): Use it.
* c-decl.c (init_decl_processing): Likewise.
* tree.h (size_int): Don't delcare, #define.
(size_int_wide, sizetype_tab, sbitsize, ubitsize): Declare.
(set_sizetype): Declare.
(bitsize_int, size_int_2, BITS_PER_UNIT_LOG, sizetype, bitsizetype):
Define.
* c-typeck.c (c_sizeof, c_sizeof_nowarn, c_size_in_bytes):
Convert result to sizetype.
(really_start_incremental_init, push_init_level):
Set type of constructor_bit_index to sbitsizetype.
(push_init_level): Use unsigned arithmetic to determine padding.
(output_init_element): Likewise.
From-SVN: r17577
* real.h (REAL_VALUES_IDENTICAL): New macro.
* expr.c (is_zeros_p): Don't consider -0.0 to be all zeros.
* fold-const.c (operand_equal_p): Don't consider -0.0 to be
identical to 0.0.
* tree.c (simple_cst_equal): Don't consider -0.0 to have the
same tree structure as 0.0.
* varasm.c (immed_real_const_1): Use new REAL_VALUES_IDENTICAL
macro instead of doing it by hand.
From-SVN: r16472
* fold-const.c (fold_truthop): When changing a one-bit comparison
against zero into a comparison against mask, do a proper sign
extension.
From-SVN: r16448
* fold-const.c (fold): For ((a * C1) / C3) or (((a * C1) + C2) / C3)
optimizations, look inside dividend to determine if the expression
can be simplified by using EXACT_DIV_EXPR.
From-SVN: r16216
* fold-const.c (make_range): Correctly handle cases of converting
from unsigned to signed type.
* fold-const.c (merge_ranges): Make sure that if one range is subset
of another, it will always be the second range. Correct (+,-) case to
account for this.
Brought over from gcc2; fixes sparc bug.
From-SVN: r15678
* pa.md (reload_peepholes): Make sure operand is a REG before
examining REGNO. Allow general registers too.
Fixes sporatic c-torture failure.
Remove last change to fold-const.c and c-decl.c
From-SVN: r15000
* c-decl.c (grokdeclarator): If array index or size calculations
overflow, issue an error.
* fold-const.c (int_const_binop): New static function.
(const_binop, size_binop): Call it.
Brought over from the fsf.
From-SVN: r14967
* version.c: Bump for new snapshot.
* local-alloc.c (update_equiv_regs): All the target to reject
promotion of some REG_EQUAL to REG_EQUIV notes.
* pa.h (DONT_RECORD_EQUIVALENCE): Define.
Fixes some c-torture failures, also improves generated code.
* pa.c (secondary_reload_class): (mem (mem ... )) does not need
secondary reloads.
Fixes 094.fpppp/twldrv.f abort.
* pa.c (hppa_builtin_saveregs): Emit a blockage insn after the
store of the argument registers.
Fixes c-torture failure.
* fold-const.c (multiple_of_p): New function.
(fold): Turn some cases of *_DIV_EXPR into EXACT_DIV_EXPR.
One of the performance patches from the g77 folks.
From-SVN: r14915
(operand_equal_p): Rework to consider two expressions that have
embedded identical SAVE_EXPRs as equivalent; also handle some more
cases.
From-SVN: r13385
(const_binop): Rework to only make constants in
one place; always use size_int if small enough and of sizetype.
(size_int): Call force_fit_type.
(fold): Avoid ever changing type of input tree.
From-SVN: r13377
(make_range): Properly decide when to get TREE_TYPE of arg0.
Handle EXP being an INTEGER_CST at end.
(fold_range_test): Handle return of 0 from make_range.
(fold, case TRUTH_AND_EXPR): Handle first arg of 0.
(fold, case TRUTH_OR_EXPR): Handle first arg of 1.
From-SVN: r13163
instead merge just with zero and convert a range with no upper bound to
opposite one with no lower bound.
(merge_ranges): In (+,-) case, don't treat subset specially if the upper
bounds are the same; fix typo in others case here.
From-SVN: r13090
(fold, shift and rotate): Don't call tree_int_cst_sgn on non-integer.
(fold, case LROTATE_EXPR): If constant count, convert to RROTATE_EXPR.
(fold, case RROTATE_EXPR): Sometimes commute logical op with rotate.
Delete pair of counteracting shifts.
From-SVN: r9425
(decode_field_reference): Don't check TREE_CODE of EXP; let
get_inner_reference decide if have reference.
Allow no bit reference if have AND_MASK.
(all_ones_mask_p): Use tree_int_cst_equal, not operand_equal_p.
(unextend): New function.
(fold_truthop): For constant cases, use new function, rework
conversion, and warn if comparison can never be true.
From-SVN: r9099
(fold, case CONVERT_EXPR): Don't remove an integer conversion when
outer is FP unless it doesn't change signedness; rework code in this
area to be clearer.
From-SVN: r8724
(fold): If moved conversion into COND_EXPR, leave it there if it is an
integral conversion to word or narrower.
(fold, case NOP_EXPR): Ignore conversion to same type as operand.
(fold, case COND_EXPR): Make third operand the simplest.
From-SVN: r8467
(encode, decode): Use 4 HOST_WIDE_INTs for encoded
value with HOST_BITS_PER_WIDE_INT/2 bits in each.
(LOWPART, HIGHPART): New macros.
(BASE): Move definition outside of div_and_round_double.
(add_double, mul_double, lshift_double, rshift_double): Rewrite.
(lrotate_double): Use LOWPART, HIGHPART, and BASE.
(rrotate_double): Likewise.
(div_and_round_double): Major changes to code for general case.
Now it actually produces non-garbage results for large operands.
(div_and_round_double): Simplify condition for special code used when
divisor < BASE.
(const_binop): Delete special cases for multiplying by 0, 1, 2, 4, 8.
(fold, case *_DIV_EXPR): Don't try to optimize for overflow.
From-SVN: r7473
(fold, case RDIV_EXPR): Split case; ignore division by 1.
If -ffast-math, convert to multply by reciprocal.
(fold, case *_DIV_EXPR): Simplify A/C1/C2.
From-SVN: r6780
(real_value_truncate): Don't use its own handler.
(const_binop): Set overflow on FP error and call force_fit_type.
(fold_convert): Likewise.
From-SVN: r6768
(fold): Use it to see when we should make a TRUTH_{AND,OR}_EXPR.
(fold, case TRUTH_{AND,OR,ANDIF,ORIF}_EXPR): Use distributive law on
expressions such as (A && B) || (A && C).
From-SVN: r6352
(fold, case EQ_EXPR, NE_EXPR): If COMPLEX_TYPE and at least one arg is
a COMPLEX_EXPR, split into a logical operation on the real and
imaginary parts.
From-SVN: r6313
(fold, case EQ_EXPR): Re-enable converting signed MOD operations to
unsigned when inside comparisons against zero, but only when the
second operand of the MOD is an integral power of two.
From-SVN: r5166
