Richard Biener <richard.guent...@gmail.com> writes:
> On Fri, Aug 18, 2017 at 10:10 AM, Richard Sandiford
> <richard.sandif...@linaro.org> wrote:
>> There are several places that test whether:
>>
>> TYPE_PRECISION (t) == GET_MODE_PRECISION (TYPE_MODE (t))
>>
>> for some integer type T. With SVE variable-length modes, this would
>> need to become:
>>
>> TYPE_PRECISION (t) == GET_MODE_PRECISION (SCALAR_TYPE_MODE (t))
>>
>> (or SCALAR_INT_TYPE_MODE, it doesn't matter which in this case).
>> But rather than add the "SCALAR_" everywhere, it seemed neater to
>> introduce a new helper function that tests whether T is an integral
>> type that has the same number of bits as its underlying mode. This
>> patch does that, calling it full_integral_type_p.
>>
>> It isn't possible to use TYPE_MODE in tree.h because vector_type_mode
>> is defined in stor-layout.h, so for now the function accesses the mode
>> field directly. After the 77-patch machine_mode series (thanks again
>> Jeff for the reviews) it would use SCALAR_TYPE_MODE instead.
>>
>> Of the changes that didn't previously have an INTEGRAL_TYPE_P check:
>>
>> - for fold_single_bit_test_into_sign_test it is obvious from the
>> integer_foop tests that this is restricted to integral types.
>>
>> - vect_recog_vector_vector_shift_pattern is inherently restricted
>> to integral types.
>>
>> - the register_edge_assert_for_2 hunk is dominated by:
>>
>> TREE_CODE (val) == INTEGER_CST
>>
>> - the ubsan_instrument_shift hunk is preceded by an early exit:
>>
>> if (!INTEGRAL_TYPE_P (type0))
>> return NULL_TREE;
>>
>> - the second and third match.pd hunks are from:
>>
>> /* Fold (X << C1) & C2 into (X << C1) & (C2 | ((1 << C1) - 1))
>> (X >> C1) & C2 into (X >> C1) & (C2 | ~((type) -1 >> C1))
>> if the new mask might be further optimized. */
>>
>> I'm a bit confused about:
>>
>> /* Try to fold (type) X op CST -> (type) (X op ((type-x) CST))
>> when profitable.
>> For bitwise binary operations apply operand conversions to the
>> binary operation result instead of to the operands. This allows
>> to combine successive conversions and bitwise binary operations.
>> We combine the above two cases by using a conditional convert. */
>> (for bitop (bit_and bit_ior bit_xor)
>> (simplify
>> (bitop (convert @0) (convert? @1))
>> (if (((TREE_CODE (@1) == INTEGER_CST
>> && INTEGRAL_TYPE_P (TREE_TYPE (@0))
>> && int_fits_type_p (@1, TREE_TYPE (@0)))
>> || types_match (@0, @1))
>> /* ??? This transform conflicts with fold-const.c doing
>> Convert (T)(x & c) into (T)x & (T)c, if c is an integer
>> constants (if x has signed type, the sign bit cannot be set
>> in c). This folds extension into the BIT_AND_EXPR.
>> Restrict it to GIMPLE to avoid endless recursions. */
>> && (bitop != BIT_AND_EXPR || GIMPLE)
>> && (/* That's a good idea if the conversion widens the operand, thus
>> after hoisting the conversion the operation will be narrower.
>> */
>> TYPE_PRECISION (TREE_TYPE (@0)) < TYPE_PRECISION (type)
>> /* It's also a good idea if the conversion is to a non-integer
>> mode. */
>> || GET_MODE_CLASS (TYPE_MODE (type)) != MODE_INT
>> /* Or if the precision of TO is not the same as the precision
>> of its mode. */
>> || TYPE_PRECISION (type) != GET_MODE_PRECISION (TYPE_MODE
>> (type))))
>> (convert (bitop @0 (convert @1))))))
>>
>> though. The "INTEGRAL_TYPE_P (TREE_TYPE (@0))" suggests that we can't
>> rely on @0 and @1 being integral (although conversions from float would
>> use FLOAT_EXPR), but then what is:
>
> bit_and is valid on POINTER_TYPE and vector integer types
>
>>
>> /* It's also a good idea if the conversion is to a non-integer
>> mode. */
>> || GET_MODE_CLASS (TYPE_MODE (type)) != MODE_INT
>>
>> letting through? MODE_PARTIAL_INT maybe, but that's a sort of integer
>> mode too. MODE_COMPLEX_INT or MODE_VECTOR_INT? I thought for those
>> it would be better to apply the scalar rules to the element type.
>
> I suppose extra caution ;) I think I have seen BLKmode for not naturally
> aligned integer types at least on strict-align targets? The code is
> a copy from original code in tree-ssa-forwprop.c.
>
>> Either way, having allowed all non-INT modes, using full_integral_type_p
>> for the remaining condition seems correct.
>>
>> If the feeling is that this isn't a useful abstraction, I can just update
>> each site individually to cope with variable-sized modes.
>
> I think "full_integral_type_p" is a name from which I cannot infer
> its meaning. Maybe type_has_mode_precision_p? Or
> type_matches_mode_p?
type_has_mode_precision_p sounds good. With that name I guess it
should be written to cope with all types (even those with variable-
width modes), so I think we'd need to continue using TYPE_MODE.
The VECTOR_MODE_P check should get optimised away in most of
the cases touched by the patch though.
Would it be OK to move the declaration of vector_type_mode to tree.h
so that type_has_mode_precision_p can be inline?
> Does TYPE_PRECISION == GET_MODE_PRECISION
> imply TYPE_SIZE == GET_MODE_BITSIZE btw?
Good question :-) You'd know better than me. The only case I can
think of is if we ever tried to use BImode for Fortran logicals
(regardless of kind). But I guess we wouldn't do that?
Thanks,
Richard
>
> Richard.
>
>> Tested on aarch64-linux-gnu and x86_64-linux-gnu. OK to install?
>>
>> Richard
>>
>>
>> 2017-08-18 Richard Sandiford <richard.sandif...@linaro.org>
>> Alan Hayward <alan.hayw...@arm.com>
>> David Sherwood <david.sherw...@arm.com>
>>
>> gcc/
>> * tree.h (scalar_type_is_full_p): New function.
>> (full_integral_type_p): Likewise.
>> * fold-const.c (fold_single_bit_test_into_sign_test): Likewise.
>> * match.pd: Likewise.
>> * tree-ssa-forwprop.c (simplify_rotate): Likewise.
>> * tree-vect-patterns.c (vect_recog_vector_vector_shift_pattern)
>> (vect_recog_mult_pattern, vect_recog_divmod_pattern): Likewise.
>> (adjust_bool_pattern): Likewise.
>> * tree-vrp.c (register_edge_assert_for_2): Likewise.
>> * ubsan.c (instrument_si_overflow): Likewise.
>>
>> gcc/c-family/
>> * c-ubsan.c (ubsan_instrument_shift): Use full_integral_type_p.
>>
>> diff --git a/gcc/c-family/c-ubsan.c b/gcc/c-family/c-ubsan.c
>> index b1386db..20f78e7 100644
>> --- a/gcc/c-family/c-ubsan.c
>> +++ b/gcc/c-family/c-ubsan.c
>> @@ -131,8 +131,8 @@ ubsan_instrument_shift (location_t loc, enum tree_code
>> code,
>>
>> /* If this is not a signed operation, don't perform overflow checks.
>> Also punt on bit-fields. */
>> - if (TYPE_OVERFLOW_WRAPS (type0)
>> - || GET_MODE_BITSIZE (TYPE_MODE (type0)) != TYPE_PRECISION (type0)
>> + if (!full_integral_type_p (type0)
>> + || TYPE_OVERFLOW_WRAPS (type0)
>> || !sanitize_flags_p (SANITIZE_SHIFT_BASE))
>> ;
>>
>> diff --git a/gcc/fold-const.c b/gcc/fold-const.c
>> index 0a5b168..1985a14 100644
>> --- a/gcc/fold-const.c
>> +++ b/gcc/fold-const.c
>> @@ -6672,8 +6672,7 @@ fold_single_bit_test_into_sign_test (location_t loc,
>> if (arg00 != NULL_TREE
>> /* This is only a win if casting to a signed type is cheap,
>> i.e. when arg00's type is not a partial mode. */
>> - && TYPE_PRECISION (TREE_TYPE (arg00))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (arg00))))
>> + && full_integral_type_p (TREE_TYPE (arg00)))
>> {
>> tree stype = signed_type_for (TREE_TYPE (arg00));
>> return fold_build2_loc (loc, code == EQ_EXPR ? GE_EXPR : LT_EXPR,
>> diff --git a/gcc/match.pd b/gcc/match.pd
>> index 0e36f46..9ad9930 100644
>> --- a/gcc/match.pd
>> +++ b/gcc/match.pd
>> @@ -992,7 +992,7 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
>> || GET_MODE_CLASS (TYPE_MODE (type)) != MODE_INT
>> /* Or if the precision of TO is not the same as the precision
>> of its mode. */
>> - || TYPE_PRECISION (type) != GET_MODE_PRECISION (TYPE_MODE
>> (type))))
>> + || !full_integral_type_p (type)))
>> (convert (bitop @0 (convert @1))))))
>>
>> (for bitop (bit_and bit_ior)
>> @@ -1920,8 +1920,7 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
>> if (shift == LSHIFT_EXPR)
>> zerobits = ((HOST_WIDE_INT_1U << shiftc) - 1);
>> else if (shift == RSHIFT_EXPR
>> - && (TYPE_PRECISION (shift_type)
>> - == GET_MODE_PRECISION (TYPE_MODE (shift_type))))
>> + && full_integral_type_p (shift_type))
>> {
>> prec = TYPE_PRECISION (TREE_TYPE (@3));
>> tree arg00 = @0;
>> @@ -1931,8 +1930,7 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
>> && TYPE_UNSIGNED (TREE_TYPE (@0)))
>> {
>> tree inner_type = TREE_TYPE (@0);
>> - if ((TYPE_PRECISION (inner_type)
>> - == GET_MODE_PRECISION (TYPE_MODE (inner_type)))
>> + if (full_integral_type_p (inner_type)
>> && TYPE_PRECISION (inner_type) < prec)
>> {
>> prec = TYPE_PRECISION (inner_type);
>> @@ -3225,9 +3223,7 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
>> ncmp (ge lt)
>> (simplify
>> (cmp (bit_and (convert?@2 @0) integer_pow2p@1) integer_zerop)
>> - (if (INTEGRAL_TYPE_P (TREE_TYPE (@0))
>> - && (TYPE_PRECISION (TREE_TYPE (@0))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (@0))))
>> + (if (full_integral_type_p (TREE_TYPE (@0))
>> && element_precision (@2) >= element_precision (@0)
>> && wi::only_sign_bit_p (@1, element_precision (@0)))
>> (with { tree stype = signed_type_for (TREE_TYPE (@0)); }
>> @@ -4021,19 +4017,13 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
>> (for op (plus minus)
>> (simplify
>> (convert (op:s (convert@2 @0) (convert?@3 @1)))
>> - (if (INTEGRAL_TYPE_P (type)
>> - /* We check for type compatibility between @0 and @1 below,
>> - so there's no need to check that @1/@3 are integral types. */
>> - && INTEGRAL_TYPE_P (TREE_TYPE (@0))
>> - && INTEGRAL_TYPE_P (TREE_TYPE (@2))
>> + (if (INTEGRAL_TYPE_P (TREE_TYPE (@2))
>> /* The precision of the type of each operand must match the
>> precision of the mode of each operand, similarly for the
>> result. */
>> - && (TYPE_PRECISION (TREE_TYPE (@0))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (@0))))
>> - && (TYPE_PRECISION (TREE_TYPE (@1))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (@1))))
>> - && TYPE_PRECISION (type) == GET_MODE_PRECISION (TYPE_MODE (type))
>> + && full_integral_type_p (TREE_TYPE (@0))
>> + && full_integral_type_p (TREE_TYPE (@1))
>> + && full_integral_type_p (type)
>> /* The inner conversion must be a widening conversion. */
>> && TYPE_PRECISION (TREE_TYPE (@2)) > TYPE_PRECISION (TREE_TYPE (@0))
>> && types_match (@0, type)
>> @@ -4055,19 +4045,13 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT)
>> (for op (minus plus)
>> (simplify
>> (bit_and (op:s (convert@2 @0) (convert@3 @1)) INTEGER_CST@4)
>> - (if (INTEGRAL_TYPE_P (type)
>> - /* We check for type compatibility between @0 and @1 below,
>> - so there's no need to check that @1/@3 are integral types. */
>> - && INTEGRAL_TYPE_P (TREE_TYPE (@0))
>> - && INTEGRAL_TYPE_P (TREE_TYPE (@2))
>> + (if (INTEGRAL_TYPE_P (TREE_TYPE (@2))
>> /* The precision of the type of each operand must match the
>> precision of the mode of each operand, similarly for the
>> result. */
>> - && (TYPE_PRECISION (TREE_TYPE (@0))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (@0))))
>> - && (TYPE_PRECISION (TREE_TYPE (@1))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (@1))))
>> - && TYPE_PRECISION (type) == GET_MODE_PRECISION (TYPE_MODE (type))
>> + && full_integral_type_p (TREE_TYPE (@0))
>> + && full_integral_type_p (TREE_TYPE (@1))
>> + && full_integral_type_p (type)
>> /* The inner conversion must be a widening conversion. */
>> && TYPE_PRECISION (TREE_TYPE (@2)) > TYPE_PRECISION (TREE_TYPE (@0))
>> && types_match (@0, @1)
>> diff --git a/gcc/tree-ssa-forwprop.c b/gcc/tree-ssa-forwprop.c
>> index 5719b99..20d5c86 100644
>> --- a/gcc/tree-ssa-forwprop.c
>> +++ b/gcc/tree-ssa-forwprop.c
>> @@ -1528,8 +1528,7 @@ simplify_rotate (gimple_stmt_iterator *gsi)
>>
>> /* Only create rotates in complete modes. Other cases are not
>> expanded properly. */
>> - if (!INTEGRAL_TYPE_P (rtype)
>> - || TYPE_PRECISION (rtype) != GET_MODE_PRECISION (TYPE_MODE (rtype)))
>> + if (!full_integral_type_p (rtype))
>> return false;
>>
>> for (i = 0; i < 2; i++)
>> @@ -1606,11 +1605,9 @@ simplify_rotate (gimple_stmt_iterator *gsi)
>> defcodefor_name (def_arg2[i], &cdef_code[i],
>> &cdef_arg1[i], &cdef_arg2[i]);
>> if (CONVERT_EXPR_CODE_P (cdef_code[i])
>> - && INTEGRAL_TYPE_P (TREE_TYPE (cdef_arg1[i]))
>> + && full_integral_type_p (TREE_TYPE (cdef_arg1[i]))
>> && TYPE_PRECISION (TREE_TYPE (cdef_arg1[i]))
>> - > floor_log2 (TYPE_PRECISION (rtype))
>> - && TYPE_PRECISION (TREE_TYPE (cdef_arg1[i]))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE
>> (cdef_arg1[i]))))
>> + > floor_log2 (TYPE_PRECISION (rtype)))
>> {
>> def_arg2_alt[i] = cdef_arg1[i];
>> defcodefor_name (def_arg2_alt[i], &cdef_code[i],
>> @@ -1636,11 +1633,9 @@ simplify_rotate (gimple_stmt_iterator *gsi)
>> }
>> defcodefor_name (cdef_arg2[i], &code, &tem, NULL);
>> if (CONVERT_EXPR_CODE_P (code)
>> - && INTEGRAL_TYPE_P (TREE_TYPE (tem))
>> + && full_integral_type_p (TREE_TYPE (tem))
>> && TYPE_PRECISION (TREE_TYPE (tem))
>> > floor_log2 (TYPE_PRECISION (rtype))
>> - && TYPE_PRECISION (TREE_TYPE (tem))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (tem)))
>> && (tem == def_arg2[1 - i]
>> || tem == def_arg2_alt[1 - i]))
>> {
>> @@ -1664,11 +1659,9 @@ simplify_rotate (gimple_stmt_iterator *gsi)
>>
>> defcodefor_name (cdef_arg1[i], &code, &tem, NULL);
>> if (CONVERT_EXPR_CODE_P (code)
>> - && INTEGRAL_TYPE_P (TREE_TYPE (tem))
>> + && full_integral_type_p (TREE_TYPE (tem))
>> && TYPE_PRECISION (TREE_TYPE (tem))
>> - > floor_log2 (TYPE_PRECISION (rtype))
>> - && TYPE_PRECISION (TREE_TYPE (tem))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (tem))))
>> + > floor_log2 (TYPE_PRECISION (rtype)))
>> defcodefor_name (tem, &code, &tem, NULL);
>>
>> if (code == NEGATE_EXPR)
>> @@ -1680,11 +1673,9 @@ simplify_rotate (gimple_stmt_iterator *gsi)
>> }
>> defcodefor_name (tem, &code, &tem, NULL);
>> if (CONVERT_EXPR_CODE_P (code)
>> - && INTEGRAL_TYPE_P (TREE_TYPE (tem))
>> + && full_integral_type_p (TREE_TYPE (tem))
>> && TYPE_PRECISION (TREE_TYPE (tem))
>> > floor_log2 (TYPE_PRECISION (rtype))
>> - && TYPE_PRECISION (TREE_TYPE (tem))
>> - == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (tem)))
>> && (tem == def_arg2[1 - i]
>> || tem == def_arg2_alt[1 - i]))
>> {
>> diff --git a/gcc/tree-vect-patterns.c b/gcc/tree-vect-patterns.c
>> index 17d1083..a96f784 100644
>> --- a/gcc/tree-vect-patterns.c
>> +++ b/gcc/tree-vect-patterns.c
>> @@ -2067,8 +2067,7 @@ vect_recog_vector_vector_shift_pattern (vec<gimple *>
>> *stmts,
>> if (TREE_CODE (oprnd0) != SSA_NAME
>> || TREE_CODE (oprnd1) != SSA_NAME
>> || TYPE_MODE (TREE_TYPE (oprnd0)) == TYPE_MODE (TREE_TYPE (oprnd1))
>> - || TYPE_PRECISION (TREE_TYPE (oprnd1))
>> - != GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (oprnd1)))
>> + || !full_integral_type_p (TREE_TYPE (oprnd1))
>> || TYPE_PRECISION (TREE_TYPE (lhs))
>> != TYPE_PRECISION (TREE_TYPE (oprnd0)))
>> return NULL;
>> @@ -2469,8 +2468,7 @@ vect_recog_mult_pattern (vec<gimple *> *stmts,
>>
>> if (TREE_CODE (oprnd0) != SSA_NAME
>> || TREE_CODE (oprnd1) != INTEGER_CST
>> - || !INTEGRAL_TYPE_P (itype)
>> - || TYPE_PRECISION (itype) != GET_MODE_PRECISION (TYPE_MODE (itype)))
>> + || !full_integral_type_p (itype))
>> return NULL;
>>
>> vectype = get_vectype_for_scalar_type (itype);
>> @@ -2584,8 +2582,7 @@ vect_recog_divmod_pattern (vec<gimple *> *stmts,
>> itype = TREE_TYPE (oprnd0);
>> if (TREE_CODE (oprnd0) != SSA_NAME
>> || TREE_CODE (oprnd1) != INTEGER_CST
>> - || TREE_CODE (itype) != INTEGER_TYPE
>> - || TYPE_PRECISION (itype) != GET_MODE_PRECISION (TYPE_MODE (itype)))
>> + || !full_integral_type_p (itype))
>> return NULL;
>>
>> vectype = get_vectype_for_scalar_type (itype);
>> @@ -3385,9 +3382,8 @@ adjust_bool_pattern (tree var, tree out_type,
>> do_compare:
>> gcc_assert (TREE_CODE_CLASS (rhs_code) == tcc_comparison);
>> if (TREE_CODE (TREE_TYPE (rhs1)) != INTEGER_TYPE
>> - || !TYPE_UNSIGNED (TREE_TYPE (rhs1))
>> - || (TYPE_PRECISION (TREE_TYPE (rhs1))
>> - != GET_MODE_BITSIZE (TYPE_MODE (TREE_TYPE (rhs1)))))
>> + || !full_integral_type_p (TREE_TYPE (rhs1))
>> + || !TYPE_UNSIGNED (TREE_TYPE (rhs1)))
>> {
>> machine_mode mode = TYPE_MODE (TREE_TYPE (rhs1));
>> itype
>> diff --git a/gcc/tree-vrp.c b/gcc/tree-vrp.c
>> index 657a8d1..cfa9a97 100644
>> --- a/gcc/tree-vrp.c
>> +++ b/gcc/tree-vrp.c
>> @@ -5245,7 +5245,7 @@ register_edge_assert_for_2 (tree name, edge e,
>> && tree_fits_uhwi_p (cst2)
>> && INTEGRAL_TYPE_P (TREE_TYPE (name2))
>> && IN_RANGE (tree_to_uhwi (cst2), 1, prec - 1)
>> - && prec == GET_MODE_PRECISION (TYPE_MODE (TREE_TYPE (val))))
>> + && full_integral_type_p (TREE_TYPE (val)))
>> {
>> mask = wi::mask (tree_to_uhwi (cst2), false, prec);
>> val2 = fold_binary (LSHIFT_EXPR, TREE_TYPE (val), val, cst2);
>> diff --git a/gcc/tree.h b/gcc/tree.h
>> index 46debc1..237f234 100644
>> --- a/gcc/tree.h
>> +++ b/gcc/tree.h
>> @@ -5394,4 +5394,25 @@ struct builtin_structptr_type
>> const char *str;
>> };
>> extern const builtin_structptr_type builtin_structptr_types[6];
>> +
>> +/* Return true if the mode underlying scalar type T has the same number
>> + of bits as T does. Examples of when this is false include bitfields
>> + that are narrower than the mode that contains them. */
>> +
>> +inline bool
>> +scalar_type_is_full_p (const_tree t)
>> +{
>> + return (GET_MODE_PRECISION (TYPE_CHECK (t)->type_common.mode)
>> + == TYPE_PRECISION (t));
>> +}
>> +
>> +/* Return true if T is an integral type that has the same number of bits
>> + as its underlying mode. */
>> +
>> +inline bool
>> +full_integral_type_p (const_tree t)
>> +{
>> + return INTEGRAL_TYPE_P (t) && scalar_type_is_full_p (t);
>> +}
>> +
>> #endif /* GCC_TREE_H */
>> diff --git a/gcc/ubsan.c b/gcc/ubsan.c
>> index 49e38fa..40f5f3e 100644
>> --- a/gcc/ubsan.c
>> +++ b/gcc/ubsan.c
>> @@ -1582,9 +1582,8 @@ instrument_si_overflow (gimple_stmt_iterator gsi)
>>
>> /* If this is not a signed operation, don't instrument anything here.
>> Also punt on bit-fields. */
>> - if (!INTEGRAL_TYPE_P (lhsinner)
>> - || TYPE_OVERFLOW_WRAPS (lhsinner)
>> - || GET_MODE_BITSIZE (TYPE_MODE (lhsinner)) != TYPE_PRECISION
>> (lhsinner))
>> + if (!full_integral_type_p (lhsinner)
>> + || TYPE_OVERFLOW_WRAPS (lhsinner))
>> return;
>>
>> switch (code)
