From: Richard Sandiford Date: Wed, 30 Aug 2017 11:16:42 +0000 (+0000) Subject: [47/77] Make subroutines of nonzero_bits operate on scalar_int_mode X-Git-Url: https://git.libre-soc.org/?a=commitdiff_plain;h=22527b59dbb10204626aeb8660a8e262377ac354;p=gcc.git [47/77] Make subroutines of nonzero_bits operate on scalar_int_mode nonzero_bits1 assumed that all bits of a floating-point or vector mode were needed. It seems likely that fixed-point modes should have been handled in the same way. After excluding those, the only remaining modes that are likely to be useful are scalar integer ones. This patch moves the mode class check to nonzero_bits itself, along with the handling of mode == VOIDmode. The subroutines of nonzero_bits can then take scalar_int_modes. 2017-08-30 Richard Sandiford Alan Hayward David Sherwood gcc/ * rtlanal.c (nonzero_bits): Handle VOIDmode here rather than in subroutines. Return the mode mask for non-integer modes. (cached_nonzero_bits): Change the type of the mode parameter to scalar_int_mode. (nonzero_bits1): Likewise. Remove early exit for other mode classes. Handle CONST_INT_P first and then check whether X also has a scalar integer mode. Co-Authored-By: Alan Hayward Co-Authored-By: David Sherwood From-SVN: r251499 --- diff --git a/gcc/ChangeLog b/gcc/ChangeLog index c7e6009858e..08982d847fc 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,15 @@ +2017-08-30 Richard Sandiford + Alan Hayward + David Sherwood + + * rtlanal.c (nonzero_bits): Handle VOIDmode here rather than + in subroutines. Return the mode mask for non-integer modes. + (cached_nonzero_bits): Change the type of the mode parameter + to scalar_int_mode. + (nonzero_bits1): Likewise. Remove early exit for other mode + classes. Handle CONST_INT_P first and then check whether X + also has a scalar integer mode. + 2017-08-30 Richard Sandiford Alan Hayward David Sherwood diff --git a/gcc/rtlanal.c b/gcc/rtlanal.c index dfb17486a47..c8d802f3f34 100644 --- a/gcc/rtlanal.c +++ b/gcc/rtlanal.c @@ -43,10 +43,10 @@ static bool covers_regno_no_parallel_p (const_rtx, unsigned int); static int computed_jump_p_1 (const_rtx); static void parms_set (rtx, const_rtx, void *); -static unsigned HOST_WIDE_INT cached_nonzero_bits (const_rtx, machine_mode, +static unsigned HOST_WIDE_INT cached_nonzero_bits (const_rtx, scalar_int_mode, const_rtx, machine_mode, unsigned HOST_WIDE_INT); -static unsigned HOST_WIDE_INT nonzero_bits1 (const_rtx, machine_mode, +static unsigned HOST_WIDE_INT nonzero_bits1 (const_rtx, scalar_int_mode, const_rtx, machine_mode, unsigned HOST_WIDE_INT); static unsigned int cached_num_sign_bit_copies (const_rtx, machine_mode, const_rtx, @@ -4237,7 +4237,12 @@ default_address_cost (rtx x, machine_mode, addr_space_t, bool speed) unsigned HOST_WIDE_INT nonzero_bits (const_rtx x, machine_mode mode) { - return cached_nonzero_bits (x, mode, NULL_RTX, VOIDmode, 0); + if (mode == VOIDmode) + mode = GET_MODE (x); + scalar_int_mode int_mode; + if (!is_a (mode, &int_mode)) + return GET_MODE_MASK (mode); + return cached_nonzero_bits (x, int_mode, NULL_RTX, VOIDmode, 0); } unsigned int @@ -4281,7 +4286,7 @@ nonzero_bits_binary_arith_p (const_rtx x) identical subexpressions on the first or the second level. */ static unsigned HOST_WIDE_INT -cached_nonzero_bits (const_rtx x, machine_mode mode, const_rtx known_x, +cached_nonzero_bits (const_rtx x, scalar_int_mode mode, const_rtx known_x, machine_mode known_mode, unsigned HOST_WIDE_INT known_ret) { @@ -4334,7 +4339,7 @@ cached_nonzero_bits (const_rtx x, machine_mode mode, const_rtx known_x, an arithmetic operation, we can do better. */ static unsigned HOST_WIDE_INT -nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, +nonzero_bits1 (const_rtx x, scalar_int_mode mode, const_rtx known_x, machine_mode known_mode, unsigned HOST_WIDE_INT known_ret) { @@ -4342,19 +4347,31 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, unsigned HOST_WIDE_INT inner_nz; enum rtx_code code; machine_mode inner_mode; + scalar_int_mode xmode; + unsigned int mode_width = GET_MODE_PRECISION (mode); - /* For floating-point and vector values, assume all bits are needed. */ - if (FLOAT_MODE_P (GET_MODE (x)) || FLOAT_MODE_P (mode) - || VECTOR_MODE_P (GET_MODE (x)) || VECTOR_MODE_P (mode)) + if (CONST_INT_P (x)) + { + if (SHORT_IMMEDIATES_SIGN_EXTEND + && INTVAL (x) > 0 + && mode_width < BITS_PER_WORD + && (UINTVAL (x) & (HOST_WIDE_INT_1U << (mode_width - 1))) != 0) + return UINTVAL (x) | (HOST_WIDE_INT_M1U << mode_width); + + return UINTVAL (x); + } + + if (!is_a (GET_MODE (x), &xmode)) return nonzero; + unsigned int xmode_width = GET_MODE_PRECISION (xmode); /* If X is wider than MODE, use its mode instead. */ - if (GET_MODE_PRECISION (GET_MODE (x)) > mode_width) + if (xmode_width > mode_width) { - mode = GET_MODE (x); + mode = xmode; nonzero = GET_MODE_MASK (mode); - mode_width = GET_MODE_PRECISION (mode); + mode_width = xmode_width; } if (mode_width > HOST_BITS_PER_WIDE_INT) @@ -4370,15 +4387,13 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, not known to be zero. */ if (!WORD_REGISTER_OPERATIONS - && GET_MODE (x) != VOIDmode - && GET_MODE (x) != mode - && GET_MODE_PRECISION (GET_MODE (x)) <= BITS_PER_WORD - && GET_MODE_PRECISION (GET_MODE (x)) <= HOST_BITS_PER_WIDE_INT - && GET_MODE_PRECISION (mode) > GET_MODE_PRECISION (GET_MODE (x))) + && mode_width > xmode_width + && xmode_width <= BITS_PER_WORD + && xmode_width <= HOST_BITS_PER_WIDE_INT) { - nonzero &= cached_nonzero_bits (x, GET_MODE (x), + nonzero &= cached_nonzero_bits (x, xmode, known_x, known_mode, known_ret); - nonzero |= GET_MODE_MASK (mode) & ~GET_MODE_MASK (GET_MODE (x)); + nonzero |= GET_MODE_MASK (mode) & ~GET_MODE_MASK (xmode); return nonzero; } @@ -4395,7 +4410,8 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, we can do this only if the target does not support different pointer or address modes depending on the address space. */ if (target_default_pointer_address_modes_p () - && POINTERS_EXTEND_UNSIGNED && GET_MODE (x) == Pmode + && POINTERS_EXTEND_UNSIGNED + && xmode == Pmode && REG_POINTER (x) && !targetm.have_ptr_extend ()) nonzero &= GET_MODE_MASK (ptr_mode); @@ -4438,22 +4454,12 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, return nonzero_for_hook; } - case CONST_INT: - /* If X is negative in MODE, sign-extend the value. */ - if (SHORT_IMMEDIATES_SIGN_EXTEND && INTVAL (x) > 0 - && mode_width < BITS_PER_WORD - && (UINTVAL (x) & (HOST_WIDE_INT_1U << (mode_width - 1))) - != 0) - return UINTVAL (x) | (HOST_WIDE_INT_M1U << mode_width); - - return UINTVAL (x); - case MEM: /* In many, if not most, RISC machines, reading a byte from memory zeros the rest of the register. Noticing that fact saves a lot of extra zero-extends. */ - if (load_extend_op (GET_MODE (x)) == ZERO_EXTEND) - nonzero &= GET_MODE_MASK (GET_MODE (x)); + if (load_extend_op (xmode) == ZERO_EXTEND) + nonzero &= GET_MODE_MASK (xmode); break; case EQ: case NE: @@ -4470,7 +4476,7 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, operation in, and not the actual operation mode. We can wind up with (subreg:DI (gt:V4HI x y)), and we don't have anything that describes the results of a vector compare. */ - if (GET_MODE_CLASS (GET_MODE (x)) == MODE_INT + if (GET_MODE_CLASS (xmode) == MODE_INT && mode_width <= HOST_BITS_PER_WIDE_INT) nonzero = STORE_FLAG_VALUE; break; @@ -4479,21 +4485,19 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, #if 0 /* Disabled to avoid exponential mutual recursion between nonzero_bits and num_sign_bit_copies. */ - if (num_sign_bit_copies (XEXP (x, 0), GET_MODE (x)) - == GET_MODE_PRECISION (GET_MODE (x))) + if (num_sign_bit_copies (XEXP (x, 0), xmode) == xmode_width) nonzero = 1; #endif - if (GET_MODE_PRECISION (GET_MODE (x)) < mode_width) - nonzero |= (GET_MODE_MASK (mode) & ~GET_MODE_MASK (GET_MODE (x))); + if (xmode_width < mode_width) + nonzero |= (GET_MODE_MASK (mode) & ~GET_MODE_MASK (xmode)); break; case ABS: #if 0 /* Disabled to avoid exponential mutual recursion between nonzero_bits and num_sign_bit_copies. */ - if (num_sign_bit_copies (XEXP (x, 0), GET_MODE (x)) - == GET_MODE_PRECISION (GET_MODE (x))) + if (num_sign_bit_copies (XEXP (x, 0), xmode) == xmode_width) nonzero = 1; #endif break; @@ -4566,7 +4570,7 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, unsigned HOST_WIDE_INT nz1 = cached_nonzero_bits (XEXP (x, 1), mode, known_x, known_mode, known_ret); - int sign_index = GET_MODE_PRECISION (GET_MODE (x)) - 1; + int sign_index = xmode_width - 1; int width0 = floor_log2 (nz0) + 1; int width1 = floor_log2 (nz1) + 1; int low0 = ctz_or_zero (nz0); @@ -4638,8 +4642,8 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, been zero-extended, we know that at least the high-order bits are zero, though others might be too. */ if (SUBREG_PROMOTED_VAR_P (x) && SUBREG_PROMOTED_UNSIGNED_P (x)) - nonzero = GET_MODE_MASK (GET_MODE (x)) - & cached_nonzero_bits (SUBREG_REG (x), GET_MODE (x), + nonzero = GET_MODE_MASK (xmode) + & cached_nonzero_bits (SUBREG_REG (x), xmode, known_x, known_mode, known_ret); /* If the inner mode is a single word for both the host and target @@ -4663,10 +4667,8 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, ? val_signbit_known_set_p (inner_mode, nonzero) : extend_op != ZERO_EXTEND) || (!MEM_P (SUBREG_REG (x)) && !REG_P (SUBREG_REG (x)))) - && GET_MODE_PRECISION (GET_MODE (x)) - > GET_MODE_PRECISION (inner_mode)) - nonzero - |= (GET_MODE_MASK (GET_MODE (x)) & ~GET_MODE_MASK (inner_mode)); + && xmode_width > GET_MODE_PRECISION (inner_mode)) + nonzero |= (GET_MODE_MASK (xmode) & ~GET_MODE_MASK (inner_mode)); } break; @@ -4675,7 +4677,7 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, case ASHIFT: case ROTATE: /* The nonzero bits are in two classes: any bits within MODE - that aren't in GET_MODE (x) are always significant. The rest of the + that aren't in xmode are always significant. The rest of the nonzero bits are those that are significant in the operand of the shift when shifted the appropriate number of bits. This shows that high-order bits are cleared by the right shift and @@ -4683,19 +4685,17 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, if (CONST_INT_P (XEXP (x, 1)) && INTVAL (XEXP (x, 1)) >= 0 && INTVAL (XEXP (x, 1)) < HOST_BITS_PER_WIDE_INT - && INTVAL (XEXP (x, 1)) < GET_MODE_PRECISION (GET_MODE (x))) + && INTVAL (XEXP (x, 1)) < xmode_width) { - machine_mode inner_mode = GET_MODE (x); - unsigned int width = GET_MODE_PRECISION (inner_mode); int count = INTVAL (XEXP (x, 1)); - unsigned HOST_WIDE_INT mode_mask = GET_MODE_MASK (inner_mode); + unsigned HOST_WIDE_INT mode_mask = GET_MODE_MASK (xmode); unsigned HOST_WIDE_INT op_nonzero = cached_nonzero_bits (XEXP (x, 0), mode, known_x, known_mode, known_ret); unsigned HOST_WIDE_INT inner = op_nonzero & mode_mask; unsigned HOST_WIDE_INT outer = 0; - if (mode_width > width) + if (mode_width > xmode_width) outer = (op_nonzero & nonzero & ~mode_mask); if (code == LSHIFTRT) @@ -4707,15 +4707,16 @@ nonzero_bits1 (const_rtx x, machine_mode mode, const_rtx known_x, /* If the sign bit may have been nonzero before the shift, we need to mark all the places it could have been copied to by the shift as possibly nonzero. */ - if (inner & (HOST_WIDE_INT_1U << (width - 1 - count))) - inner |= ((HOST_WIDE_INT_1U << count) - 1) - << (width - count); + if (inner & (HOST_WIDE_INT_1U << (xmode_width - 1 - count))) + inner |= (((HOST_WIDE_INT_1U << count) - 1) + << (xmode_width - count)); } else if (code == ASHIFT) inner <<= count; else - inner = ((inner << (count % width) - | (inner >> (width - (count % width)))) & mode_mask); + inner = ((inner << (count % xmode_width) + | (inner >> (xmode_width - (count % xmode_width)))) + & mode_mask); nonzero &= (outer | inner); }