[thirdparty/glibc.git] / sysdeps / sparc / sparc32 / sfp-machine.h

/* Machine-dependent software floating-point definitions.
   Sparc userland (_Q_*) version.
   Copyright (C) 1997-2019 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).

   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/>.  */

#include <fpu_control.h>
#include <stdlib.h>

#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_DIV_MEAT_S(R,X,Y)	_FP_DIV_MEAT_1_udiv(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

/* If one NaN is signaling and the other is not,
 * we choose that one, otherwise we choose X.
 */
/* For _Qp_* and _Q_*, this should prefer X, for
 * CPU instruction emulation this should prefer Y.
 * (see SPAMv9 B.2.2 section).
 */
#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)

/* Some assembly to speed things up. */
#define __FP_FRAC_ADD_3(r2,r1,r0,x2,x1,x0,y2,y1,y0)			\
  __asm__ ("addcc %r7,%8,%2\n\
	    addxcc %r5,%6,%1\n\
	    addx %r3,%4,%0"						\
	   : "=r" ((USItype)(r2)),					\
	     "=&r" ((USItype)(r1)),					\
	     "=&r" ((USItype)(r0))					\
	   : "%rJ" ((USItype)(x2)),					\
	     "rI" ((USItype)(y2)),					\
	     "%rJ" ((USItype)(x1)),					\
	     "rI" ((USItype)(y1)),					\
	     "%rJ" ((USItype)(x0)),					\
	     "rI" ((USItype)(y0))					\
	   : "cc")

#define __FP_FRAC_SUB_3(r2,r1,r0,x2,x1,x0,y2,y1,y0)			\
  __asm__ ("subcc %r7,%8,%2\n\
	    subxcc %r5,%6,%1\n\
	    subx %r3,%4,%0"						\
	   : "=r" ((USItype)(r2)),					\
	     "=&r" ((USItype)(r1)),					\
	     "=&r" ((USItype)(r0))					\
	   : "%rJ" ((USItype)(x2)),					\
	     "rI" ((USItype)(y2)),					\
	     "%rJ" ((USItype)(x1)),					\
	     "rI" ((USItype)(y1)),					\
	     "%rJ" ((USItype)(x0)),					\
	     "rI" ((USItype)(y0))					\
	   : "cc")

#define __FP_FRAC_ADD_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0)		\
  do {									\
    /* We need to fool gcc,  as we need to pass more than 10		\
       input/outputs.  */						\
    register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2");		\
    __asm__ __volatile__ ("\
	    addcc %r8,%9,%1\n\
	    addxcc %r6,%7,%0\n\
	    addxcc %r4,%5,%%g2\n\
	    addx %r2,%3,%%g1"						\
	   : "=&r" ((USItype)(r1)),					\
	     "=&r" ((USItype)(r0))					\
	   : "%rJ" ((USItype)(x3)),					\
	     "rI" ((USItype)(y3)),					\
	     "%rJ" ((USItype)(x2)),					\
	     "rI" ((USItype)(y2)),					\
	     "%rJ" ((USItype)(x1)),					\
	     "rI" ((USItype)(y1)),					\
	     "%rJ" ((USItype)(x0)),					\
	     "rI" ((USItype)(y0))					\
	   : "cc", "g1", "g2");						\
    __asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2));			\
    r3 = _t1; r2 = _t2;							\
  } while (0)

#define __FP_FRAC_SUB_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0)		\
  do {									\
    /* We need to fool gcc,  as we need to pass more than 10		\
       input/outputs.  */						\
    register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2");		\
    __asm__ __volatile__ ("\
	    subcc %r8,%9,%1\n\
	    subxcc %r6,%7,%0\n\
	    subxcc %r4,%5,%%g2\n\
	    subx %r2,%3,%%g1"						\
	   : "=&r" ((USItype)(r1)),					\
	     "=&r" ((USItype)(r0))					\
	   : "%rJ" ((USItype)(x3)),					\
	     "rI" ((USItype)(y3)),					\
	     "%rJ" ((USItype)(x2)),					\
	     "rI" ((USItype)(y2)),					\
	     "%rJ" ((USItype)(x1)),					\
	     "rI" ((USItype)(y1)),					\
	     "%rJ" ((USItype)(x0)),					\
	     "rI" ((USItype)(y0))					\
	   : "cc", "g1", "g2");						\
    __asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2));			\
    r3 = _t1; r2 = _t2;							\
  } while (0)

#define __FP_FRAC_DEC_3(x2,x1,x0,y2,y1,y0) __FP_FRAC_SUB_3(x2,x1,x0,x2,x1,x0,y2,y1,y0)

#define __FP_FRAC_DEC_4(x3,x2,x1,x0,y3,y2,y1,y0) __FP_FRAC_SUB_4(x3,x2,x1,x0,x3,x2,x1,x0,y3,y2,y1,y0)

#define __FP_FRAC_ADDI_4(x3,x2,x1,x0,i)					\
  __asm__ ("addcc %3,%4,%3\n\
	    addxcc %2,%%g0,%2\n\
	    addxcc %1,%%g0,%1\n\
	    addx %0,%%g0,%0"						\
	   : "=&r" ((USItype)(x3)),					\
	     "=&r" ((USItype)(x2)),					\
	     "=&r" ((USItype)(x1)),					\
	     "=&r" ((USItype)(x0))					\
	   : "rI" ((USItype)(i)),					\
	     "0" ((USItype)(x3)),					\
	     "1" ((USItype)(x2)),					\
	     "2" ((USItype)(x1)),					\
	     "3" ((USItype)(x0))					\
	   : "cc")

/* Obtain the current rounding mode. */
#ifndef FP_ROUNDMODE
#define FP_ROUNDMODE	((_fcw >> 30) & 0x3)
#endif

/* Exception flags. */
#define FP_EX_INVALID		(1 << 4)
#define FP_EX_OVERFLOW		(1 << 3)
#define FP_EX_UNDERFLOW		(1 << 2)
#define FP_EX_DIVZERO		(1 << 1)
#define FP_EX_INEXACT		(1 << 0)

#define _FP_TININESS_AFTER_ROUNDING 0

#define _FP_DECL_EX \
  fpu_control_t _fcw __attribute__ ((unused)) = (FP_RND_NEAREST << 30)

#define FP_INIT_ROUNDMODE					\
do {								\
  _FPU_GETCW(_fcw);						\
} while (0)

#define FP_TRAPPING_EXCEPTIONS ((_fcw >> 23) & 0x1f)
#define FP_INHIBIT_RESULTS ((_fcw >> 23) & _fex)

/* Simulate exceptions using double arithmetics. */
extern void ___Q_simulate_exceptions(int exc);

#define FP_HANDLE_EXCEPTIONS					\
do {								\
  if (!_fex)							\
    {								\
      /* This is the common case, so we do it inline.		\
       * We need to clear cexc bits if any.			\
       */							\
      extern unsigned long long ___Q_zero;			\
      __asm__ __volatile__("ldd [%0], %%f30\n\t"		\
			   "faddd %%f30, %%f30, %%f30"		\
			   : : "r" (&___Q_zero) : "f30");	\
    }								\
  else								\
    ___Q_simulate_exceptions (_fex);			        \
} while (0)
Commit	Line	Data
d876f532 UD	1	/* Machine-dependent software floating-point definitions.
d876f532 UD	2	Sparc userland (_Q_*) version.
04277e02	3	Copyright (C) 1997-2019 Free Software Foundation, Inc.
d876f532 UD	4	This file is part of the GNU C Library.
	5	Contributed by Richard Henderson (rth@cygnus.com),
	6	Jakub Jelinek (jj@ultra.linux.cz) and
	7	David S. Miller (davem@redhat.com).
	8
	9	The GNU C Library is free software; you can redistribute it and/or
41bdb6e2 AJ	10	modify it under the terms of the GNU Lesser General Public
	11	License as published by the Free Software Foundation; either
	12	version 2.1 of the License, or (at your option) any later version.
d876f532 UD	13
	14	The GNU C Library is distributed in the hope that it will be useful,
	15	but WITHOUT ANY WARRANTY; without even the implied warranty of
	16	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
41bdb6e2	17	Lesser General Public License for more details.
d876f532	18
41bdb6e2	19	You should have received a copy of the GNU Lesser General Public
59ba27a6	20	License along with the GNU C Library; if not, see
5a82c748	21	<https://www.gnu.org/licenses/>. */
62497f9c	22
d876f532	23	#include <fpu_control.h>
c6251f03	24	#include <stdlib.h>
d876f532 UD	25
	26	#define _FP_W_TYPE_SIZE 32
	27	#define _FP_W_TYPE unsigned long
	28	#define _FP_WS_TYPE signed long
	29	#define _FP_I_TYPE long
	30
	31	#define _FP_MUL_MEAT_S(R,X,Y) \
	32	_FP_MUL_MEAT_1_wide(_FP_WFRACBITS_S,R,X,Y,umul_ppmm)
	33	#define _FP_MUL_MEAT_D(R,X,Y) \
	34	_FP_MUL_MEAT_2_wide(_FP_WFRACBITS_D,R,X,Y,umul_ppmm)
	35	#define _FP_MUL_MEAT_Q(R,X,Y) \
	36	_FP_MUL_MEAT_4_wide(_FP_WFRACBITS_Q,R,X,Y,umul_ppmm)
	37
	38	#define _FP_DIV_MEAT_S(R,X,Y) _FP_DIV_MEAT_1_udiv(S,R,X,Y)
	39	#define _FP_DIV_MEAT_D(R,X,Y) _FP_DIV_MEAT_2_udiv(D,R,X,Y)
	40	#define _FP_DIV_MEAT_Q(R,X,Y) _FP_DIV_MEAT_4_udiv(Q,R,X,Y)
	41
	42	#define _FP_NANFRAC_S ((_FP_QNANBIT_S << 1) - 1)
	43	#define _FP_NANFRAC_D ((_FP_QNANBIT_D << 1) - 1), -1
	44	#define _FP_NANFRAC_Q ((_FP_QNANBIT_Q << 1) - 1), -1, -1, -1
	45	#define _FP_NANSIGN_S 0
	46	#define _FP_NANSIGN_D 0
	47	#define _FP_NANSIGN_Q 0
	48
	49	#define _FP_KEEPNANFRACP 1
2848b105	50	#define _FP_QNANNEGATEDP 0
d876f532 UD	51
	52	/* If one NaN is signaling and the other is not,
	53	* we choose that one, otherwise we choose X.
	54	*/
	55	/* For _Qp_* and _Q_*, this should prefer X, for
	56	* CPU instruction emulation this should prefer Y.
	57	* (see SPAMv9 B.2.2 section).
	58	*/
	59	#define _FP_CHOOSENAN(fs, wc, R, X, Y, OP) \
	60	do { \
	61	if ((_FP_FRAC_HIGH_RAW_##fs(X) & _FP_QNANBIT_##fs) \
	62	&& !(_FP_FRAC_HIGH_RAW_##fs(Y) & _FP_QNANBIT_##fs)) \
	63	{ \
	64	R##_s = Y##_s; \
	65	_FP_FRAC_COPY_##wc(R,Y); \
	66	} \
	67	else \
	68	{ \
	69	R##_s = X##_s; \
	70	_FP_FRAC_COPY_##wc(R,X); \
	71	} \
	72	R##_c = FP_CLS_NAN; \
	73	} while (0)
	74
	75	/* Some assembly to speed things up. */
	76	#define __FP_FRAC_ADD_3(r2,r1,r0,x2,x1,x0,y2,y1,y0) \
62497f9c UD	77	__asm__ ("addcc %r7,%8,%2\n\
62497f9c UD	78	addxcc %r5,%6,%1\n\
d876f532 UD	79	addx %r3,%4,%0" \
	80	: "=r" ((USItype)(r2)), \
	81	"=&r" ((USItype)(r1)), \
	82	"=&r" ((USItype)(r0)) \
	83	: "%rJ" ((USItype)(x2)), \
	84	"rI" ((USItype)(y2)), \
	85	"%rJ" ((USItype)(x1)), \
	86	"rI" ((USItype)(y1)), \
	87	"%rJ" ((USItype)(x0)), \
	88	"rI" ((USItype)(y0)) \
	89	: "cc")
	90
	91	#define __FP_FRAC_SUB_3(r2,r1,r0,x2,x1,x0,y2,y1,y0) \
62497f9c UD	92	__asm__ ("subcc %r7,%8,%2\n\
62497f9c UD	93	subxcc %r5,%6,%1\n\
d876f532 UD	94	subx %r3,%4,%0" \
	95	: "=r" ((USItype)(r2)), \
	96	"=&r" ((USItype)(r1)), \
	97	"=&r" ((USItype)(r0)) \
	98	: "%rJ" ((USItype)(x2)), \
	99	"rI" ((USItype)(y2)), \
	100	"%rJ" ((USItype)(x1)), \
	101	"rI" ((USItype)(y1)), \
	102	"%rJ" ((USItype)(x0)), \
	103	"rI" ((USItype)(y0)) \
	104	: "cc")
	105
	106	#define __FP_FRAC_ADD_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0) \
	107	do { \
	108	/* We need to fool gcc, as we need to pass more than 10 \
	109	input/outputs. */ \
	110	register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2"); \
62497f9c UD	111	__asm__ __volatile__ ("\
	112	addcc %r8,%9,%1\n\
	113	addxcc %r6,%7,%0\n\
	114	addxcc %r4,%5,%%g2\n\
d876f532 UD	115	addx %r2,%3,%%g1" \
	116	: "=&r" ((USItype)(r1)), \
	117	"=&r" ((USItype)(r0)) \
	118	: "%rJ" ((USItype)(x3)), \
	119	"rI" ((USItype)(y3)), \
	120	"%rJ" ((USItype)(x2)), \
	121	"rI" ((USItype)(y2)), \
	122	"%rJ" ((USItype)(x1)), \
	123	"rI" ((USItype)(y1)), \
	124	"%rJ" ((USItype)(x0)), \
	125	"rI" ((USItype)(y0)) \
	126	: "cc", "g1", "g2"); \
	127	__asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2)); \
	128	r3 = _t1; r2 = _t2; \
	129	} while (0)
	130
	131	#define __FP_FRAC_SUB_4(r3,r2,r1,r0,x3,x2,x1,x0,y3,y2,y1,y0) \
	132	do { \
	133	/* We need to fool gcc, as we need to pass more than 10 \
	134	input/outputs. */ \
	135	register USItype _t1 __asm__ ("g1"), _t2 __asm__ ("g2"); \
62497f9c UD	136	__asm__ __volatile__ ("\
	137	subcc %r8,%9,%1\n\
	138	subxcc %r6,%7,%0\n\
	139	subxcc %r4,%5,%%g2\n\
d876f532 UD	140	subx %r2,%3,%%g1" \
	141	: "=&r" ((USItype)(r1)), \
	142	"=&r" ((USItype)(r0)) \
	143	: "%rJ" ((USItype)(x3)), \
	144	"rI" ((USItype)(y3)), \
	145	"%rJ" ((USItype)(x2)), \
	146	"rI" ((USItype)(y2)), \
	147	"%rJ" ((USItype)(x1)), \
	148	"rI" ((USItype)(y1)), \
	149	"%rJ" ((USItype)(x0)), \
	150	"rI" ((USItype)(y0)) \
	151	: "cc", "g1", "g2"); \
	152	__asm__ __volatile__ ("" : "=r" (_t1), "=r" (_t2)); \
	153	r3 = _t1; r2 = _t2; \
	154	} while (0)
	155
	156	#define __FP_FRAC_DEC_3(x2,x1,x0,y2,y1,y0) __FP_FRAC_SUB_3(x2,x1,x0,x2,x1,x0,y2,y1,y0)
	157
	158	#define __FP_FRAC_DEC_4(x3,x2,x1,x0,y3,y2,y1,y0) __FP_FRAC_SUB_4(x3,x2,x1,x0,x3,x2,x1,x0,y3,y2,y1,y0)
	159
	160	#define __FP_FRAC_ADDI_4(x3,x2,x1,x0,i) \
62497f9c UD	161	__asm__ ("addcc %3,%4,%3\n\
	162	addxcc %2,%%g0,%2\n\
	163	addxcc %1,%%g0,%1\n\
d876f532 UD	164	addx %0,%%g0,%0" \
	165	: "=&r" ((USItype)(x3)), \
	166	"=&r" ((USItype)(x2)), \
	167	"=&r" ((USItype)(x1)), \
	168	"=&r" ((USItype)(x0)) \
	169	: "rI" ((USItype)(i)), \
	170	"0" ((USItype)(x3)), \
	171	"1" ((USItype)(x2)), \
	172	"2" ((USItype)(x1)), \
	173	"3" ((USItype)(x0)) \
	174	: "cc")
	175
	176	/* Obtain the current rounding mode. */
	177	#ifndef FP_ROUNDMODE
	178	#define FP_ROUNDMODE ((_fcw >> 30) & 0x3)
	179	#endif
	180
	181	/* Exception flags. */
	182	#define FP_EX_INVALID (1 << 4)
	183	#define FP_EX_OVERFLOW (1 << 3)
	184	#define FP_EX_UNDERFLOW (1 << 2)
	185	#define FP_EX_DIVZERO (1 << 1)
	186	#define FP_EX_INEXACT (1 << 0)
	187
ace614b8 JM	188	#define _FP_TININESS_AFTER_ROUNDING 0
ace614b8 JM	189
d66ef399 DM	190	#define _FP_DECL_EX \
d66ef399 DM	191	fpu_control_t _fcw __attribute__ ((unused)) = (FP_RND_NEAREST << 30)
d876f532 UD	192
	193	#define FP_INIT_ROUNDMODE \
	194	do { \
	195	_FPU_GETCW(_fcw); \
	196	} while (0)
	197
ae251b0b	198	#define FP_TRAPPING_EXCEPTIONS ((_fcw >> 23) & 0x1f)
d66ef399 DM	199	#define FP_INHIBIT_RESULTS ((_fcw >> 23) & _fex)
d66ef399 DM	200
d876f532	201	/* Simulate exceptions using double arithmetics. */
d66ef399	202	extern void ___Q_simulate_exceptions(int exc);
d876f532 UD	203
	204	#define FP_HANDLE_EXCEPTIONS \
	205	do { \
	206	if (!_fex) \
	207	{ \
	208	/* This is the common case, so we do it inline. \
	209	* We need to clear cexc bits if any. \
	210	*/ \
d66ef399 DM	211	extern unsigned long long ___Q_zero; \
	212	__asm__ __volatile__("ldd [%0], %%f30\n\t" \
	213	"faddd %%f30, %%f30, %%f30" \
	214	: : "r" (&___Q_zero) : "f30"); \
d876f532 UD	215	} \
d876f532 UD	216	else \
c6251f03	217	___Q_simulate_exceptions (_fex); \
d876f532	218	} while (0)