[PATCH v6 05/13] ARC: math soft float support

[Adhemerval Zanella]

On 22/04/2020 22:41, Vineet Gupta via Libc-alpha wrote:
> Signed-off-by: Vineet Gupta <vgupta@synopsys.com>

As prior patch we do not use DCO, but rather copyright assignment.

Looks ok in general, with some comments below.

> ---
>  sysdeps/arc/bits/fenv.h                   | 77 +++++++++++++++++++++++
>  sysdeps/arc/nofpu/Implies                 |  1 +
>  sysdeps/arc/nofpu/math-tests-exceptions.h | 27 ++++++++
>  sysdeps/arc/nofpu/math-tests-rounding.h   | 27 ++++++++
>  sysdeps/arc/sfp-machine.h                 | 73 +++++++++++++++++++++
>  5 files changed, 205 insertions(+)
>  create mode 100644 sysdeps/arc/bits/fenv.h
>  create mode 100644 sysdeps/arc/nofpu/Implies
>  create mode 100644 sysdeps/arc/nofpu/math-tests-exceptions.h
>  create mode 100644 sysdeps/arc/nofpu/math-tests-rounding.h
>  create mode 100644 sysdeps/arc/sfp-machine.h
> 
> diff --git a/sysdeps/arc/bits/fenv.h b/sysdeps/arc/bits/fenv.h
> new file mode 100644
> index 000000000000..80afa50db9c6
> --- /dev/null
> +++ b/sysdeps/arc/bits/fenv.h
> @@ -0,0 +1,77 @@

Missing one line description.

> +/* Copyright (C) 2012-2020 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library.  If not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#ifndef _FENV_H
> +# error "Never use <bits/fenv.h> directly; include <fenv.h> instead."
> +#endif
> +
> +enum
> +  {
> +    FE_INVALID   =
> +# define FE_INVALID	(0x01)
> +      FE_INVALID,
> +    FE_DIVBYZERO =
> +# define FE_DIVBYZERO	(0x02)
> +      FE_DIVBYZERO,
> +    FE_OVERFLOW  =
> +# define FE_OVERFLOW	(0x04)
> +      FE_OVERFLOW,
> +    FE_UNDERFLOW =
> +# define FE_UNDERFLOW	(0x08)
> +      FE_UNDERFLOW,
> +    FE_INEXACT   =
> +# define FE_INEXACT	(0x10)
> +      FE_INEXACT
> +  };
> +
> +# define FE_ALL_EXCEPT \
> +	(FE_INVALID | FE_DIVBYZERO | FE_OVERFLOW | FE_UNDERFLOW | FE_INEXACT)
> +
> +enum
> +  {
> +    FE_TOWARDZERO =
> +# define FE_TOWARDZERO	(0x0)
> +      FE_TOWARDZERO,
> +    FE_TONEAREST  =
> +# define FE_TONEAREST	(0x1)	/* default */
> +      FE_TONEAREST,
> +    FE_UPWARD     =
> +# define FE_UPWARD	(0x2)
> +      FE_UPWARD,
> +    FE_DOWNWARD   =
> +# define FE_DOWNWARD	(0x3)
> +      FE_DOWNWARD
> +  };
> +
> +typedef unsigned int fexcept_t;
> +
> +typedef struct
> +{
> +  unsigned int __fpcr;
> +  unsigned int __fpsr;
> +} fenv_t;
> +
> +/* If the default argument is used we use this value.  */
> +#define FE_DFL_ENV	((const fenv_t *) -1)
> +
> +#if __GLIBC_USE (IEC_60559_BFP_EXT)
> +/* Type representing floating-point control modes.  */
> +typedef unsigned int femode_t;
> +
> +/* Default floating-point control modes.  */
> +# define FE_DFL_MODE	((const femode_t *) -1L)
> +#endif

Ok.

> diff --git a/sysdeps/arc/nofpu/Implies b/sysdeps/arc/nofpu/Implies
> new file mode 100644
> index 000000000000..abcbadb25f22
> --- /dev/null
> +++ b/sysdeps/arc/nofpu/Implies
> @@ -0,0 +1 @@
> +ieee754/soft-fp

Ok.

> diff --git a/sysdeps/arc/nofpu/math-tests-exceptions.h b/sysdeps/arc/nofpu/math-tests-exceptions.h
> new file mode 100644
> index 000000000000..2ec58e088ebf
> --- /dev/null
> +++ b/sysdeps/arc/nofpu/math-tests-exceptions.h
> @@ -0,0 +1,27 @@
> +/* Configuration for math tests. exceptions support ARC version.

Double space after period and capitalize 'exceptions'. 

> +   Copyright (C) 2017-2020 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#ifndef ARC_NOFPU_MATH_TESTS_EXCEPTIONS_H
> +#define ARC_NOFPU_MATH_TESTS_EXCEPTIONS_H 1
> +
> +/* Soft-float does not support exceptions.  */
> +#define EXCEPTION_TESTS_float		0
> +#define EXCEPTION_TESTS_double		0
> +#define EXCEPTION_TESTS_long_double	0
> +
> +#endif

Ok.

> diff --git a/sysdeps/arc/nofpu/math-tests-rounding.h b/sysdeps/arc/nofpu/math-tests-rounding.h
> new file mode 100644
> index 000000000000..6e5376cb35b5
> --- /dev/null
> +++ b/sysdeps/arc/nofpu/math-tests-rounding.h
> @@ -0,0 +1,27 @@
> +/* Configuration for math tests: rounding mode support.  ARC version.
> +   Copyright (C) 2017-2020 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#ifndef ARC_NOFPU_MATH_TESTS_ROUNDING_H
> +#define ARC_NOFPU_MATH_TESTS_ROUNDING_H 1
> +
> +/* Soft-float only supports to-nearest rounding mode.  */
> +#define ROUNDING_TESTS_float(MODE)		((MODE) == FE_TONEAREST)
> +#define ROUNDING_TESTS_double(MODE)		((MODE) == FE_TONEAREST)
> +#define ROUNDING_TESTS_long_double(MODE)	((MODE) == FE_TONEAREST)
> +
> +#endif

Ok.

> diff --git a/sysdeps/arc/sfp-machine.h b/sysdeps/arc/sfp-machine.h
> new file mode 100644
> index 000000000000..b2bfd33e944c
> --- /dev/null
> +++ b/sysdeps/arc/sfp-machine.h
> @@ -0,0 +1,73 @@
> +/* Machine-dependent software floating-point definitions.  ARC version.
> +   Copyright (C) 2004-2020 Free Software Foundation, Inc.
> +   This file is part of the GNU C Library.
> +   Contributed by Richard Henderson (rth@cygnus.com),
> +		  Jakub Jelinek (jj@ultra.linux.cz) and
> +		  David S. Miller (davem@redhat.com).

We don't add 'Contributed by' anymore in new files, but I am not sure
how should we proceed with code that might be shared with gcc.

> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library.  If not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +
> +#define _FP_W_TYPE_SIZE		32
> +#define _FP_W_TYPE		unsigned long
> +#define _FP_WS_TYPE		signed long
> +#define _FP_I_TYPE		long
> +
> +#define _FP_MUL_MEAT_S(R,X,Y)				\
> +  _FP_MUL_MEAT_1_wide(_FP_WFRACBITS_S,R,X,Y,umul_ppmm)
> +#define _FP_MUL_MEAT_D(R,X,Y)				\
> +  _FP_MUL_MEAT_2_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
> +#define _FP_MUL_MEAT_Q(R,X,Y)				\
> +  _FP_MUL_MEAT_4_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
> +
> +#define _FP_MUL_MEAT_DW_S(R,X,Y)				\
> +  _FP_MUL_MEAT_DW_1_wide(_FP_WFRACBITS_S,R,X,Y,umul_ppmm)
> +#define _FP_MUL_MEAT_DW_D(R,X,Y)				\
> +  _FP_MUL_MEAT_DW_2_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
> +#define _FP_MUL_MEAT_DW_Q(R,X,Y)				\
> +  _FP_MUL_MEAT_DW_4_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
> +
> +#define _FP_DIV_MEAT_S(R,X,Y)	_FP_DIV_MEAT_1_loop(S,R,X,Y)
> +#define _FP_DIV_MEAT_D(R,X,Y)	_FP_DIV_MEAT_2_udiv(D,R,X,Y)
> +#define _FP_DIV_MEAT_Q(R,X,Y)	_FP_DIV_MEAT_4_udiv(Q,R,X,Y)
> +
> +#define _FP_NANFRAC_S		((_FP_QNANBIT_S << 1) - 1)
> +#define _FP_NANFRAC_D		((_FP_QNANBIT_D << 1) - 1), -1
> +#define _FP_NANFRAC_Q		((_FP_QNANBIT_Q << 1) - 1), -1, -1, -1
> +#define _FP_NANSIGN_S		0
> +#define _FP_NANSIGN_D		0
> +#define _FP_NANSIGN_Q		0
> +
> +#define _FP_KEEPNANFRACP 1
> +#define _FP_QNANNEGATEDP 0
> +
> +/* This is arbitrarily taken from the PowerPC version.  */
> +#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP)			\
> +  do {								\
> +    if ((_FP_FRAC_HIGH_RAW_##fs(X) & _FP_QNANBIT_##fs)		\
> +	&& !(_FP_FRAC_HIGH_RAW_##fs(Y) & _FP_QNANBIT_##fs))	\
> +      {								\
> +	R##_s = Y##_s;						\
> +	_FP_FRAC_COPY_##wc(R,Y);				\
> +      }								\
> +    else							\
> +      {								\
> +	R##_s = X##_s;						\
> +	_FP_FRAC_COPY_##wc(R,X);				\
> +      }								\
> +    R##_c = FP_CLS_NAN;						\
> +  } while (0)
> +
> +#define _FP_TININESS_AFTER_ROUNDING 1
> 

Ok.