patch-2.2.18 linux/arch/m68k/math-emu/fp_util.S

Next file: linux/arch/m68k/math-emu/multi_arith.h
Previous file: linux/arch/m68k/math-emu/fp_trig.h
Back to the patch index
Back to the overall index

diff -u --new-file --recursive --exclude-from /usr/src/exclude v2.2.17/arch/m68k/math-emu/fp_util.S linux/arch/m68k/math-emu/fp_util.S
@@ -0,0 +1,1454 @@
+/*
+ * fp_util.S
+ *
+ * Copyright Roman Zippel, 1997.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, and the entire permission notice in its entirety,
+ *    including the disclaimer of warranties.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. The name of the author may not be used to endorse or promote
+ *    products derived from this software without specific prior
+ *    written permission.
+ *
+ * ALTERNATIVELY, this product may be distributed under the terms of
+ * the GNU Public License, in which case the provisions of the GPL are
+ * required INSTEAD OF the above restrictions.  (This clause is
+ * necessary due to a potential bad interaction between the GPL and
+ * the restrictions contained in a BSD-style copyright.)
+ *
+ * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
+ * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
+ * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+ * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "fp_emu.h"
+
+/*
+ * Here are lots of conversion and normalization functions mainly
+ * used by fp_scan.S
+ * Note that these functions are optimized for "normal" numbers,
+ * these are handled first and exit as fast as possible, this is
+ * especially important for fp_normalize_ext/fp_conv_ext2ext, as
+ * it's called very often.
+ * The register usage is optimized for fp_scan.S and which register
+ * is currently at that time unused, be careful if you want change
+ * something here. %d0 and %d1 is always usable, sometimes %d2 (or
+ * only the lower half) most function have to return the %a0
+ * unmodified, so that the caller can immediatly reuse it.
+ */
+
+	.globl	fp_ill, fp_end
+
+	| exits from fp_scan:
+	| illegal instruction
+fp_ill:
+	printf	,"fp_illegal\n"
+	rts
+	| completed instruction
+fp_end:
+	tst.l	(TASK_MM-8,%a2)
+	jmi	1f
+	tst.l	(TASK_MM-4,%a2)
+	jmi	1f
+	tst.l	(TASK_MM,%a2)
+	jpl	2f
+1:	printf	,"oops:%p,%p,%p\n",3,%a2@(TASK_MM-8),%a2@(TASK_MM-4),%a2@(TASK_MM)
+2:	clr.l	%d0
+	rts
+
+	.globl	fp_conv_long2ext, fp_conv_single2ext
+	.globl	fp_conv_double2ext, fp_conv_ext2ext
+	.globl	fp_normalize_ext, fp_normalize_double
+	.globl	fp_normalize_single, fp_normalize_single_fast
+	.globl	fp_conv_ext2double, fp_conv_ext2single
+	.globl	fp_conv_ext2long, fp_conv_ext2short
+	.globl	fp_conv_ext2byte
+	.globl	fp_finalrounding_single, fp_finalrounding_single_fast
+	.globl	fp_finalrounding_double
+	.globl	fp_finalrounding, fp_finaltest, fp_final
+
+/*
+ * First several conversion functions from a source operand
+ * into the extended format. Note, that only fp_conv_ext2ext
+ * normalizes the number and is always called after the other
+ * conversion functions, which only move the information into
+ * fp_ext structure.
+ */
+
+	| fp_conv_long2ext:
+	|
+	| args:	%d0 = source (32-bit long)
+	|	%a0 = destination (ptr to struct fp_ext)
+
+fp_conv_long2ext:
+	printf	PCONV,"l2e: %p -> %p(",2,%d0,%a0
+	clr.l	%d1			| sign defaults to zero
+	tst.l	%d0
+	jeq	fp_l2e_zero		| is source zero?
+	jpl	1f			| positive?
+	moveq	#1,%d1
+	neg.l	%d0
+1:	swap	%d1
+	move.w	#0x3fff+31,%d1
+	move.l	%d1,(%a0)+		| set sign / exp
+	move.l	%d0,(%a0)+		| set mantissa
+	clr.l	(%a0)
+	subq.l	#8,%a0			| restore %a0
+	printx	PCONV,%a0@
+	printf	PCONV,")\n"
+	rts
+	| source is zero
+fp_l2e_zero:
+	clr.l	(%a0)+
+	clr.l	(%a0)+
+	clr.l	(%a0)
+	subq.l	#8,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,")\n"
+	rts
+
+	| fp_conv_single2ext
+	| args:	%d0 = source (single-precision fp value)
+	|	%a0 = dest (struct fp_ext *)
+
+fp_conv_single2ext:
+	printf	PCONV,"s2e: %p -> %p(",2,%d0,%a0
+	move.l	%d0,%d1
+	lsl.l	#8,%d0			| shift mantissa
+	lsr.l	#8,%d1			| exponent / sign
+	lsr.l	#7,%d1
+	lsr.w	#8,%d1
+	jeq	fp_s2e_small		| zero / denormal?
+	cmp.w	#0xff,%d1		| NaN / Inf?
+	jeq	fp_s2e_large
+	bset	#31,%d0			| set explizit bit
+	add.w	#0x3fff-0x7f,%d1	| re-bias the exponent.
+9:	move.l	%d1,(%a0)+		| fp_ext.sign, fp_ext.exp
+	move.l	%d0,(%a0)+		| high lword of fp_ext.mant
+	clr.l	(%a0)			| low lword = 0
+	subq.l	#8,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,")\n"
+	rts
+	| zeros and denormalized
+fp_s2e_small:
+	| exponent is zero, so explizit bit is already zero too
+	tst.l	%d0
+	jeq	9b
+	move.w	#0x4000-0x7f,%d1
+	jra	9b
+	| infinities and NAN
+fp_s2e_large:
+	bclr	#31,%d0			| clear explizit bit
+	move.w	#0x7fff,%d1
+	jra	9b
+
+fp_conv_double2ext:
+#ifdef FPU_EMU_DEBUG
+	getuser.l %a1@(0),%d0,fp_err_ua2,%a1
+	getuser.l %a1@(4),%d1,fp_err_ua2,%a1
+	printf	PCONV,"d2e: %p%p -> %p(",3,%d0,%d1,%a0
+#endif
+	getuser.l (%a1)+,%d0,fp_err_ua2,%a1
+	move.l	%d0,%d1
+	lsl.l	#8,%d0			| shift high mantissa
+	lsl.l	#3,%d0
+	lsr.l	#8,%d1			| exponent / sign
+	lsr.l	#7,%d1
+	lsr.w	#5,%d1
+	jeq	fp_d2e_small		| zero / denormal?
+	cmp.w	#0x7ff,%d1		| NaN / Inf?
+	jeq	fp_d2e_large
+	bset	#31,%d0			| set explizit bit
+	add.w	#0x3fff-0x3ff,%d1	| re-bias the exponent.
+9:	move.l	%d1,(%a0)+		| fp_ext.sign, fp_ext.exp
+	move.l	%d0,(%a0)+
+	getuser.l (%a1)+,%d0,fp_err_ua2,%a1
+	move.l	%d0,%d1
+	lsl.l	#8,%d0
+	lsl.l	#3,%d0
+	move.l	%d0,(%a0)
+	moveq	#21,%d0
+	lsr.l	%d0,%d1
+	or.l	%d1,-(%a0)
+	subq.l	#4,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,")\n"
+	rts
+	| zeros and denormalized
+fp_d2e_small:
+	| exponent is zero, so explizit bit is already zero too
+	tst.l	%d0
+	jeq	9b
+	move.w	#0x4000-0x3ff,%d1
+	jra	9b
+	| infinities and NAN
+fp_d2e_large:
+	bclr	#31,%d0			| clear explizit bit
+	move.w	#0x7fff,%d1
+	jra	9b
+
+	| fp_conv_ext2ext:
+	| originally used to get longdouble from userspace, now it's
+	| called before arithmetic operations to make sure the number
+	| is normalized [maybe rename it?].
+	| args:	%a0 = dest (struct fp_ext *)
+	| returns 0 in %d0 for a NaN, otherwise 1
+
+fp_conv_ext2ext:
+	printf	PCONV,"e2e: %p(",1,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,"), "
+	move.l	(%a0)+,%d0
+	cmp.w	#0x7fff,%d0		| Inf / NaN?
+	jeq	fp_e2e_large
+	move.l	(%a0),%d0
+	jpl	fp_e2e_small		| zero / denorm?
+	| The high bit is set, so normalization is irrelevant.
+fp_e2e_checkround:
+	subq.l	#4,%a0
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+	move.b	(%a0),%d0
+	jne	fp_e2e_round
+#endif
+	printf	PCONV,"%p(",1,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,")\n"
+	moveq	#1,%d0
+	rts
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+fp_e2e_round:
+	fp_set_sr FPSR_EXC_INEX2
+	clr.b	(%a0)
+	move.w	(FPD_RND,FPDATA),%d2
+	jne	fp_e2e_roundother	| %d2 == 0, round to nearest
+	tst.b	%d0			| test guard bit
+	jpl	9f			| zero is closer
+	btst	#0,(11,%a0)		| test lsb bit
+	jne	fp_e2e_doroundup	| round to infinity
+	lsl.b	#1,%d0			| check low bits
+	jeq	9f			| round to zero
+fp_e2e_doroundup:
+	addq.l	#1,(8,%a0)
+	jcc	9f
+	addq.l	#1,(4,%a0)
+	jcc	9f
+	move.w	#0x8000,(4,%a0)
+	addq.w	#1,(2,%a0)
+9:	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+fp_e2e_roundother:
+	subq.w	#2,%d2
+	jcs	9b			| %d2 < 2, round to zero
+	jhi	1f			| %d2 > 2, round to +infinity
+	tst.b	(1,%a0)			| to -inf
+	jne	fp_e2e_doroundup	| negative, round to infinity
+	jra	9b			| positive, round to zero
+1:	tst.b	(1,%a0)			| to +inf
+	jeq	fp_e2e_doroundup	| positive, round to infinity
+	jra	9b			| negative, round to zero
+#endif
+	| zeros and subnormals:
+	| try to normalize these anyway.
+fp_e2e_small:
+	jne	fp_e2e_small1		| high lword zero?
+	move.l	(4,%a0),%d0
+	jne	fp_e2e_small2
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+	clr.l	%d0
+	move.b	(-4,%a0),%d0
+	jne	fp_e2e_small3
+#endif
+	| Genuine zero.
+	clr.w	-(%a0)
+	subq.l	#2,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	moveq	#1,%d0
+	rts
+	| definitely subnormal, need to shift all 64 bits
+fp_e2e_small1:
+	bfffo	%d0{#0,#32},%d1
+	move.w	-(%a0),%d2
+	sub.w	%d1,%d2
+	jcc	1f
+	| Pathologically small, denormalize.
+	add.w	%d2,%d1
+	clr.w	%d2
+1:	move.w	%d2,(%a0)+
+	move.w	%d1,%d2
+	jeq	fp_e2e_checkround
+	| fancy 64-bit double-shift begins here
+	lsl.l	%d2,%d0
+	move.l	%d0,(%a0)+
+	move.l	(%a0),%d0
+	move.l	%d0,%d1
+	lsl.l	%d2,%d0
+	move.l	%d0,(%a0)
+	neg.w	%d2
+	and.w	#0x1f,%d2
+	lsr.l	%d2,%d1
+	or.l	%d1,-(%a0)
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+fp_e2e_extra1:
+	clr.l	%d0
+	move.b	(-4,%a0),%d0
+	neg.w	%d2
+	add.w	#24,%d2
+	jcc	1f
+	clr.b	(-4,%a0)
+	lsl.l	%d2,%d0
+	or.l	%d0,(4,%a0)
+	jra	fp_e2e_checkround
+1:	addq.w	#8,%d2
+	lsl.l	%d2,%d0
+	move.b	%d0,(-4,%a0)
+	lsr.l	#8,%d0
+	or.l	%d0,(4,%a0)
+#endif
+	jra	fp_e2e_checkround
+	| pathologically small subnormal
+fp_e2e_small2:
+	bfffo	%d0{#0,#32},%d1
+	add.w	#32,%d1
+	move.w	-(%a0),%d2
+	sub.w	%d1,%d2
+	jcc	1f
+	| Beyond pathologically small, denormalize.
+	add.w	%d2,%d1
+	clr.w	%d2
+1:	move.w	%d2,(%a0)+
+	ext.l	%d1
+	jeq	fp_e2e_checkround
+	clr.l	(4,%a0)
+	sub.w	#32,%d2
+	jcs	1f
+	lsl.l	%d1,%d0			| lower lword needs only to be shifted
+	move.l	%d0,(%a0)		| into the higher lword
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+	clr.l	%d0
+	move.b	(-4,%a0),%d0
+	clr.b	(-4,%a0)
+	neg.w	%d1
+	add.w	#32,%d1
+	bfins	%d0,(%a0){%d1,#8}
+#endif
+	jra	fp_e2e_checkround
+1:	neg.w	%d1			| lower lword is splitted between
+	bfins	%d0,(%a0){%d1,#32}	| higher and lower lword
+#ifndef CONFIG_FPU_EMU_EXTRAPREC
+	jra	fp_e2e_checkround
+#else
+	move.w	%d1,%d2
+	jra	fp_e2e_extra1
+	| These are extremely small numbers, that will mostly end up as zero
+	| anyway, so this is only important for correct rounding.
+fp_e2e_small3:
+	bfffo	%d0{#24,#8},%d1
+	add.w	#40,%d1
+	move.w	-(%a0),%d2
+	sub.w	%d1,%d2
+	jcc	1f
+	| Pathologically small, denormalize.
+	add.w	%d2,%d1
+	clr.w	%d2
+1:	move.w	%d2,(%a0)+
+	ext.l	%d1
+	jeq	fp_e2e_checkround
+	cmp.w	#8,%d1
+	jcs	2f
+1:	clr.b	(-4,%a0)
+	sub.w	#64,%d1
+	jcs	1f
+	add.w	#24,%d1
+	lsl.l	%d1,%d0
+	move.l	%d0,(%a0)
+	jra	fp_e2e_checkround
+1:	neg.w	%d1
+	bfins	%d0,(%a0){%d1,#8}
+	jra	fp_e2e_checkround
+2:	lsl.l	%d1,%d0
+	move.b	%d0,(-4,%a0)
+	lsr.l	#8,%d0
+	move.b	%d0,(7,%a0)
+	jra	fp_e2e_checkround
+#endif
+1:	move.l	%d0,%d1			| lower lword is splitted between
+	lsl.l	%d2,%d0			| higher and lower lword
+	move.l	%d0,(%a0)
+	move.l	%d1,%d0
+	neg.w	%d2
+	add.w	#32,%d2
+	lsr.l	%d2,%d0
+	move.l	%d0,-(%a0)
+	jra	fp_e2e_checkround
+	| Infinities and NaNs
+fp_e2e_large:
+	move.l	(%a0)+,%d0
+	jne	3f
+1:	tst.l	(%a0)
+	jne	4f
+	moveq	#1,%d0
+2:	subq.l	#8,%a0
+	printf	PCONV,"%p(",1,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,")\n"
+	rts
+	| we have maybe a NaN, shift off the highest bit
+3:	lsl.l	#1,%d0
+	jeq	1b
+	| we have a NaN, clear the return value
+4:	clrl	%d0
+	jra	2b
+
+
+/*
+ * Normalization functions.  Call these on the output of general
+ * FP operators, and before any conversion into the destination
+ * formats. fp_normalize_ext has always to be called first, the
+ * following conversion functions expect an already normalized
+ * number.
+ */
+
+	| fp_normalize_ext:
+	| normalize an extended in extended (unpacked) format, basically
+	| it does the same as fp_conv_ext2ext, additionally it also does
+	| the necessary postprocessing checks.
+	| args:	%a0 (struct fp_ext *)
+	| NOTE: it does _not_ modify %a0/%a1 and the upper word of %d2
+
+fp_normalize_ext:
+	printf	PNORM,"ne: %p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,"), "
+	move.l	(%a0)+,%d0
+	cmp.w	#0x7fff,%d0		| Inf / NaN?
+	jeq	fp_ne_large
+	move.l	(%a0),%d0
+	jpl	fp_ne_small		| zero / denorm?
+	| The high bit is set, so normalization is irrelevant.
+fp_ne_checkround:
+	subq.l	#4,%a0
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+	move.b	(%a0),%d0
+	jne	fp_ne_round
+#endif
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+fp_ne_round:
+	fp_set_sr FPSR_EXC_INEX2
+	clr.b	(%a0)
+	move.w	(FPD_RND,FPDATA),%d2
+	jne	fp_ne_roundother	| %d2 == 0, round to nearest
+	tst.b	%d0			| test guard bit
+	jpl	9f			| zero is closer
+	btst	#0,(11,%a0)		| test lsb bit
+	jne	fp_ne_doroundup		| round to infinity
+	lsl.b	#1,%d0			| check low bits
+	jeq	9f			| round to zero
+fp_ne_doroundup:
+	addq.l	#1,(8,%a0)
+	jcc	9f
+	addq.l	#1,(4,%a0)
+	jcc	9f
+	addq.w	#1,(2,%a0)
+	move.w	#0x8000,(4,%a0)
+9:	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+fp_ne_roundother:
+	subq.w	#2,%d2
+	jcs	9b			| %d2 < 2, round to zero
+	jhi	1f			| %d2 > 2, round to +infinity
+	tst.b	(1,%a0)			| to -inf
+	jne	fp_ne_doroundup		| negative, round to infinity
+	jra	9b			| positive, round to zero
+1:	tst.b	(1,%a0)			| to +inf
+	jeq	fp_ne_doroundup		| positive, round to infinity
+	jra	9b			| negative, round to zero
+#endif
+	| Zeros and subnormal numbers
+	| These are probably merely subnormal, rather than "denormalized"
+	|  numbers, so we will try to make them normal again.
+fp_ne_small:
+	jne	fp_ne_small1		| high lword zero?
+	move.l	(4,%a0),%d0
+	jne	fp_ne_small2
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+	clr.l	%d0
+	move.b	(-4,%a0),%d0
+	jne	fp_ne_small3
+#endif
+	| Genuine zero.
+	clr.w	-(%a0)
+	subq.l	#2,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+	| Subnormal.
+fp_ne_small1:
+	bfffo	%d0{#0,#32},%d1
+	move.w	-(%a0),%d2
+	sub.w	%d1,%d2
+	jcc	1f
+	| Pathologically small, denormalize.
+	add.w	%d2,%d1
+	clr.w	%d2
+	fp_set_sr FPSR_EXC_UNFL
+1:	move.w	%d2,(%a0)+
+	move.w	%d1,%d2
+	jeq	fp_ne_checkround
+	| This is exactly the same 64-bit double shift as seen above.
+	lsl.l	%d2,%d0
+	move.l	%d0,(%a0)+
+	move.l	(%a0),%d0
+	move.l	%d0,%d1
+	lsl.l	%d2,%d0
+	move.l	%d0,(%a0)
+	neg.w	%d2
+	and.w	#0x1f,%d2
+	lsr.l	%d2,%d1
+	or.l	%d1,-(%a0)
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+fp_ne_extra1:
+	clr.l	%d0
+	move.b	(-4,%a0),%d0
+	neg.w	%d2
+	add.w	#24,%d2
+	jcc	1f
+	clr.b	(-4,%a0)
+	lsl.l	%d2,%d0
+	or.l	%d0,(4,%a0)
+	jra	fp_ne_checkround
+1:	addq.w	#8,%d2
+	lsl.l	%d2,%d0
+	move.b	%d0,(-4,%a0)
+	lsr.l	#8,%d0
+	or.l	%d0,(4,%a0)
+#endif
+	jra	fp_ne_checkround
+	| May or may not be subnormal, if so, only 32 bits to shift.
+fp_ne_small2:
+	bfffo	%d0{#0,#32},%d1
+	add.w	#32,%d1
+	move.w	-(%a0),%d2
+	sub.w	%d1,%d2
+	jcc	1f
+	| Beyond pathologically small, denormalize.
+	add.w	%d2,%d1
+	clr.w	%d2
+	fp_set_sr FPSR_EXC_UNFL
+1:	move.w	%d2,(%a0)+
+	ext.l	%d1
+	jeq	fp_ne_checkround
+	clr.l	(4,%a0)
+	sub.w	#32,%d1
+	jcs	1f
+	lsl.l	%d1,%d0			| lower lword needs only to be shifted
+	move.l	%d0,(%a0)		| into the higher lword
+#ifdef CONFIG_FPU_EMU_EXTRAPREC
+	clr.l	%d0
+	move.b	(-4,%a0),%d0
+	clr.b	(-4,%a0)
+	neg.w	%d1
+	add.w	#32,%d1
+	bfins	%d0,(%a0){%d1,#8}
+#endif
+	jra	fp_ne_checkround
+1:	neg.w	%d1			| lower lword is splitted between
+	bfins	%d0,(%a0){%d1,#32}	| higher and lower lword
+#ifndef CONFIG_FPU_EMU_EXTRAPREC
+	jra	fp_ne_checkround
+#else
+	move.w	%d1,%d2
+	jra	fp_ne_extra1
+	| These are extremely small numbers, that will mostly end up as zero
+	| anyway, so this is only important for correct rounding.
+fp_ne_small3:
+	bfffo	%d0{#24,#8},%d1
+	add.w	#40,%d1
+	move.w	-(%a0),%d2
+	sub.w	%d1,%d2
+	jcc	1f
+	| Pathologically small, denormalize.
+	add.w	%d2,%d1
+	clr.w	%d2
+1:	move.w	%d2,(%a0)+
+	ext.l	%d1
+	jeq	fp_ne_checkround
+	cmp.w	#8,%d1
+	jcs	2f
+1:	clr.b	(-4,%a0)
+	sub.w	#64,%d1
+	jcs	1f
+	add.w	#24,%d1
+	lsl.l	%d1,%d0
+	move.l	%d0,(%a0)
+	jra	fp_ne_checkround
+1:	neg.w	%d1
+	bfins	%d0,(%a0){%d1,#8}
+	jra	fp_ne_checkround
+2:	lsl.l	%d1,%d0
+	move.b	%d0,(-4,%a0)
+	lsr.l	#8,%d0
+	move.b	%d0,(7,%a0)
+	jra	fp_ne_checkround
+#endif
+	| Infinities and NaNs, again, same as above.
+fp_ne_large:
+	move.l	(%a0)+,%d0
+	jne	3f
+1:	tst.l	(%a0)
+	jne	4f
+2:	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+	| we have maybe a NaN, shift off the highest bit
+3:	move.l	%d0,%d1
+	lsl.l	#1,%d1
+	jne	4f
+	clr.l	(-4,%a0)
+	jra	1b
+	| we have a NaN, test if it is signaling
+4:	bset	#30,%d0
+	jne	2b
+	fp_set_sr FPSR_EXC_SNAN
+	move.l	%d0,(-4,%a0)
+	jra	2b
+
+	| these next two do rounding as per the IEEE standard.
+	| values for the rounding modes appear to be:
+	| 0:	Round to nearest
+	| 1:	Round to zero
+	| 2:	Round to -Infinity
+	| 3:	Round to +Infinity
+	| both functions expect that fp_normalize was already
+	| called (and extended argument is already normalized
+	| as far as possible), these are used if there is different
+	| rounding precision is selected and before converting
+	| into single/double
+
+	| fp_normalize_double:
+	| normalize an extended with double (52-bit) precision
+	| args:	 %a0 (struct fp_ext *)
+
+fp_normalize_double:
+	printf	PNORM,"nd: %p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,"), "
+	move.l	(%a0)+,%d2
+	tst.w	%d2
+	jeq	fp_nd_zero		| zero / denormalized
+	cmp.w	#0x7fff,%d2
+	jeq	fp_nd_huge		| NaN / infinitive.
+	sub.w	#0x4000-0x3ff,%d2	| will the exponent fit?
+	jcs	fp_nd_small		| too small.
+	cmp.w	#0x7fe,%d2
+	jcc	fp_nd_large		| too big.
+	addq.l	#4,%a0
+	move.l	(%a0),%d0		| low lword of mantissa
+	| now, round off the low 11 bits.
+fp_nd_round:
+	moveq	#21,%d1
+	lsl.l	%d1,%d0			| keep 11 low bits.
+	jne	fp_nd_checkround	| Are they non-zero?
+	| nothing to do here
+9:	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+	| Be careful with the X bit! It contains the lsb
+	| from the shift above, it is needed for round to nearest.
+fp_nd_checkround:
+	fp_set_sr FPSR_EXC_INEX2	| INEX2 bit
+	and.w	#0xf800,(2,%a0)		| clear bits 0-10
+	move.w	(FPD_RND,FPDATA),%d2	| rounding mode
+	jne	2f			| %d2 == 0, round to nearest
+	tst.l	%d0			| test guard bit
+	jpl	9b			| zero is closer
+	| here we test the X bit by adding it to %d2
+	clr.w	%d2			| first set z bit, addx only clears it
+	addx.w	%d2,%d2			| test lsb bit
+	| IEEE754-specified "round to even" behaviour.  If the guard
+	| bit is set, then the number is odd, so rounding works like
+	| in grade-school arithmetic (i.e. 1.5 rounds to 2.0)
+	| Otherwise, an equal distance rounds towards zero, so as not
+	| to produce an odd number.  This is strange, but it is what
+	| the standard says.
+	jne	fp_nd_doroundup		| round to infinity
+	lsl.l	#1,%d0			| check low bits
+	jeq	9b			| round to zero
+fp_nd_doroundup:
+	| round (the mantissa, that is) towards infinity
+	add.l	#0x800,(%a0)
+	jcc	9b			| no overflow, good.
+	addq.l	#1,-(%a0)		| extend to high lword
+	jcc	1f			| no overflow, good.
+	| Yow! we have managed to overflow the mantissa.  Since this
+	| only happens when %d1 was 0xfffff800, it is now zero, so
+	| reset the high bit, and increment the exponent.
+	move.w	#0x8000,(%a0)
+	addq.w	#1,-(%a0)
+	cmp.w	#0x43ff,(%a0)+		| exponent now overflown?
+	jeq	fp_nd_large		| yes, so make it infinity.
+1:	subq.l	#4,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+2:	subq.w	#2,%d2
+	jcs	9b			| %d2 < 2, round to zero
+	jhi	3f			| %d2 > 2, round to +infinity
+	| Round to +Inf or -Inf.  High word of %d2 contains the
+	| sign of the number, by the way.
+	swap	%d2			| to -inf
+	tst.b	%d2
+	jne	fp_nd_doroundup		| negative, round to infinity
+	jra	9b			| positive, round to zero
+3:	swap	%d2			| to +inf
+	tst.b	%d2
+	jeq	fp_nd_doroundup		| positive, round to infinity
+	jra	9b			| negative, round to zero
+	| Exponent underflow.  Try to make a denormal, and set it to
+	| the smallest possible fraction if this fails.
+fp_nd_small:
+	fp_set_sr FPSR_EXC_UNFL		| set UNFL bit
+	move.w	#0x3c01,(-2,%a0)	| 2**-1022
+	neg.w	%d2			| degree of underflow
+	cmp.w	#32,%d2			| single or double shift?
+	jcc	1f
+	| Again, another 64-bit double shift.
+	move.l	(%a0),%d0
+	move.l	%d0,%d1
+	lsr.l	%d2,%d0
+	move.l	%d0,(%a0)+
+	move.l	(%a0),%d0
+	lsr.l	%d2,%d0
+	neg.w	%d2
+	add.w	#32,%d2
+	lsl.l	%d2,%d1
+	or.l	%d1,%d0
+	move.l	(%a0),%d1
+	move.l	%d0,(%a0)
+	| Check to see if we shifted off any significant bits
+	lsl.l	%d2,%d1
+	jeq	fp_nd_round		| Nope, round.
+	bset	#0,%d0			| Yes, so set the "sticky bit".
+	jra	fp_nd_round		| Now, round.
+	| Another 64-bit single shift and store
+1:	sub.w	#32,%d2
+	cmp.w	#32,%d2			| Do we really need to shift?
+	jcc	2f			| No, the number is too small.
+	move.l	(%a0),%d0
+	clr.l	(%a0)+
+	move.l	%d0,%d1
+	lsr.l	%d2,%d0
+	neg.w	%d2
+	add.w	#32,%d2
+	| Again, check to see if we shifted off any significant bits.
+	tst.l	(%a0)
+	jeq	1f
+	bset	#0,%d0			| Sticky bit.
+1:	move.l	%d0,(%a0)
+	lsl.l	%d2,%d1
+	jeq	fp_nd_round
+	bset	#0,%d0
+	jra	fp_nd_round
+	| Sorry, the number is just too small.
+2:	clr.l	(%a0)+
+	clr.l	(%a0)
+	moveq	#1,%d0			| Smallest possible fraction,
+	jra	fp_nd_round		| round as desired.
+	| zero and denormalized
+fp_nd_zero:
+	tst.l	(%a0)+
+	jne	1f
+	tst.l	(%a0)
+	jne	1f
+	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts				| zero.  nothing to do.
+	| These are not merely subnormal numbers, but true denormals,
+	| i.e. pathologically small (exponent is 2**-16383) numbers.
+	| It is clearly impossible for even a normal extended number
+	| with that exponent to fit into double precision, so just
+	| write these ones off as "too darn small".
+1:	fp_set_sr FPSR_EXC_UNFL		| Set UNFL bit
+	clr.l	(%a0)
+	clr.l	-(%a0)
+	move.w	#0x3c01,-(%a0)		| i.e. 2**-1022
+	addq.l	#6,%a0
+	moveq	#1,%d0
+	jra	fp_nd_round		| round.
+	| Exponent overflow.  Just call it infinity.
+fp_nd_large:
+	move.w	#0x7ff,%d0
+	and.w	(6,%a0),%d0
+	jeq	1f
+	fp_set_sr FPSR_EXC_INEX2
+1:	fp_set_sr FPSR_EXC_OVFL
+	move.w	(FPD_RND,FPDATA),%d2
+	jne	3f			| %d2 = 0 round to nearest
+1:	move.w	#0x7fff,(-2,%a0)
+	clr.l	(%a0)+
+	clr.l	(%a0)
+2:	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+3:	subq.w	#2,%d2
+	jcs	5f			| %d2 < 2, round to zero
+	jhi	4f			| %d2 > 2, round to +infinity
+	tst.b	(-3,%a0)		| to -inf
+	jne	1b
+	jra	5f
+4:	tst.b	(-3,%a0)		| to +inf
+	jeq	1b
+5:	move.w	#0x43fe,(-2,%a0)
+	moveq	#-1,%d0
+	move.l	%d0,(%a0)+
+	move.w	#0xf800,%d0
+	move.l	%d0,(%a0)
+	jra	2b
+	| Infinities or NaNs
+fp_nd_huge:
+	subq.l	#4,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+
+	| fp_normalize_single:
+	| normalize an extended with single (23-bit) precision
+	| args:	 %a0 (struct fp_ext *)
+
+fp_normalize_single:
+	printf	PNORM,"ns: %p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,") "
+	addq.l	#2,%a0
+	move.w	(%a0)+,%d2
+	jeq	fp_ns_zero		| zero / denormalized
+	cmp.w	#0x7fff,%d2
+	jeq	fp_ns_huge		| NaN / infinitive.
+	sub.w	#0x4000-0x7f,%d2	| will the exponent fit?
+	jcs	fp_ns_small		| too small.
+	cmp.w	#0xfe,%d2
+	jcc	fp_ns_large		| too big.
+	move.l	(%a0)+,%d0		| get high lword of mantissa
+fp_ns_round:
+	tst.l	(%a0)			| check the low lword
+	jeq	1f
+	| Set a sticky bit if it is non-zero.  This should only
+	| affect the rounding in what would otherwise be equal-
+	| distance situations, which is what we want it to do.
+	bset	#0,%d0
+1:	clr.l	(%a0)			| zap it from memory.
+	| now, round off the low 8 bits of the hi lword.
+	tst.b	%d0			| 8 low bits.
+	jne	fp_ns_checkround	| Are they non-zero?
+	| nothing to do here
+	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+fp_ns_checkround:
+	fp_set_sr FPSR_EXC_INEX2	| INEX2 bit
+	clr.b	-(%a0)			| clear low byte of high lword
+	subq.l	#3,%a0
+	move.w	(FPD_RND,FPDATA),%d2	| rounding mode
+	jne	2f			| %d2 == 0, round to nearest
+	tst.b	%d0			| test guard bit
+	jpl	9f			| zero is closer
+	btst	#8,%d0			| test lsb bit
+	| round to even behaviour, see above.
+	jne	fp_ns_doroundup		| round to infinity
+	lsl.b	#1,%d0			| check low bits
+	jeq	9f			| round to zero
+fp_ns_doroundup:
+	| round (the mantissa, that is) towards infinity
+	add.l	#0x100,(%a0)
+	jcc	9f			| no overflow, good.
+	| Overflow.  This means that the %d1 was 0xffffff00, so it
+	| is now zero.  We will set the mantissa to reflect this, and
+	| increment the exponent (checking for overflow there too)
+	move.w	#0x8000,(%a0)
+	addq.w	#1,-(%a0)
+	cmp.w	#0x407f,(%a0)+		| exponent now overflown?
+	jeq	fp_ns_large		| yes, so make it infinity.
+9:	subq.l	#4,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+	| check nondefault rounding modes
+2:	subq.w	#2,%d2
+	jcs	9b			| %d2 < 2, round to zero
+	jhi	3f			| %d2 > 2, round to +infinity
+	tst.b	(-3,%a0)		| to -inf
+	jne	fp_ns_doroundup		| negative, round to infinity
+	jra	9b			| positive, round to zero
+3:	tst.b	(-3,%a0)		| to +inf
+	jeq	fp_ns_doroundup		| positive, round to infinity
+	jra	9b			| negative, round to zero
+	| Exponent underflow.  Try to make a denormal, and set it to
+	| the smallest possible fraction if this fails.
+fp_ns_small:
+	fp_set_sr FPSR_EXC_UNFL		| set UNFL bit
+	move.w	#0x3f81,(-2,%a0)	| 2**-126
+	neg.w	%d2			| degree of underflow
+	cmp.w	#32,%d2			| single or double shift?
+	jcc	2f
+	| a 32-bit shift.
+	move.l	(%a0),%d0
+	move.l	%d0,%d1
+	lsr.l	%d2,%d0
+	move.l	%d0,(%a0)+
+	| Check to see if we shifted off any significant bits.
+	neg.w	%d2
+	add.w	#32,%d2
+	lsl.l	%d2,%d1
+	jeq	1f
+	bset	#0,%d0			| Sticky bit.
+	| Check the lower lword
+1:	tst.l	(%a0)
+	jeq	fp_ns_round
+	clr	(%a0)
+	bset	#0,%d0			| Sticky bit.
+	jra	fp_ns_round
+	| Sorry, the number is just too small.
+2:	clr.l	(%a0)+
+	clr.l	(%a0)
+	moveq	#1,%d0			| Smallest possible fraction,
+	jra	fp_ns_round		| round as desired.
+	| Exponent overflow.  Just call it infinity.
+fp_ns_large:
+	tst.b	(3,%a0)
+	jeq	1f
+	fp_set_sr FPSR_EXC_INEX2
+1:	fp_set_sr FPSR_EXC_OVFL
+	move.w	(FPD_RND,FPDATA),%d2
+	jne	3f			| %d2 = 0 round to nearest
+1:	move.w	#0x7fff,(-2,%a0)
+	clr.l	(%a0)+
+	clr.l	(%a0)
+2:	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+3:	subq.w	#2,%d2
+	jcs	5f			| %d2 < 2, round to zero
+	jhi	4f			| %d2 > 2, round to +infinity
+	tst.b	(-3,%a0)		| to -inf
+	jne	1b
+	jra	5f
+4:	tst.b	(-3,%a0)		| to +inf
+	jeq	1b
+5:	move.w	#0x407e,(-2,%a0)
+	move.l	#0xffffff00,(%a0)+
+	clr.l	(%a0)
+	jra	2b
+	| zero and denormalized
+fp_ns_zero:
+	tst.l	(%a0)+
+	jne	1f
+	tst.l	(%a0)
+	jne	1f
+	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts				| zero.  nothing to do.
+	| These are not merely subnormal numbers, but true denormals,
+	| i.e. pathologically small (exponent is 2**-16383) numbers.
+	| It is clearly impossible for even a normal extended number
+	| with that exponent to fit into single precision, so just
+	| write these ones off as "too darn small".
+1:	fp_set_sr FPSR_EXC_UNFL		| Set UNFL bit
+	clr.l	(%a0)
+	clr.l	-(%a0)
+	move.w	#0x3f81,-(%a0)		| i.e. 2**-126
+	addq.l	#6,%a0
+	moveq	#1,%d0
+	jra	fp_ns_round		| round.
+	| Infinities or NaNs
+fp_ns_huge:
+	subq.l	#4,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+
+	| fp_normalize_single_fast:
+	| normalize an extended with single (23-bit) precision
+	| this is only used by fsgldiv/fsgdlmul, where the
+	| operand is not completly normalized.
+	| args:	 %a0 (struct fp_ext *)
+
+fp_normalize_single_fast:
+	printf	PNORM,"nsf: %p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,") "
+	addq.l	#2,%a0
+	move.w	(%a0)+,%d2
+	cmp.w	#0x7fff,%d2
+	jeq	fp_nsf_huge		| NaN / infinitive.
+	move.l	(%a0)+,%d0		| get high lword of mantissa
+fp_nsf_round:
+	tst.l	(%a0)			| check the low lword
+	jeq	1f
+	| Set a sticky bit if it is non-zero.  This should only
+	| affect the rounding in what would otherwise be equal-
+	| distance situations, which is what we want it to do.
+	bset	#0,%d0
+1:	clr.l	(%a0)			| zap it from memory.
+	| now, round off the low 8 bits of the hi lword.
+	tst.b	%d0			| 8 low bits.
+	jne	fp_nsf_checkround	| Are they non-zero?
+	| nothing to do here
+	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+fp_nsf_checkround:
+	fp_set_sr FPSR_EXC_INEX2	| INEX2 bit
+	clr.b	-(%a0)			| clear low byte of high lword
+	subq.l	#3,%a0
+	move.w	(FPD_RND,FPDATA),%d2	| rounding mode
+	jne	2f			| %d2 == 0, round to nearest
+	tst.b	%d0			| test guard bit
+	jpl	9f			| zero is closer
+	btst	#8,%d0			| test lsb bit
+	| round to even behaviour, see above.
+	jne	fp_nsf_doroundup		| round to infinity
+	lsl.b	#1,%d0			| check low bits
+	jeq	9f			| round to zero
+fp_nsf_doroundup:
+	| round (the mantissa, that is) towards infinity
+	add.l	#0x100,(%a0)
+	jcc	9f			| no overflow, good.
+	| Overflow.  This means that the %d1 was 0xffffff00, so it
+	| is now zero.  We will set the mantissa to reflect this, and
+	| increment the exponent (checking for overflow there too)
+	move.w	#0x8000,(%a0)
+	addq.w	#1,-(%a0)
+	cmp.w	#0x407f,(%a0)+		| exponent now overflown?
+	jeq	fp_nsf_large		| yes, so make it infinity.
+9:	subq.l	#4,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+	| check nondefault rounding modes
+2:	subq.w	#2,%d2
+	jcs	9b			| %d2 < 2, round to zero
+	jhi	3f			| %d2 > 2, round to +infinity
+	tst.b	(-3,%a0)		| to -inf
+	jne	fp_nsf_doroundup	| negative, round to infinity
+	jra	9b			| positive, round to zero
+3:	tst.b	(-3,%a0)		| to +inf
+	jeq	fp_nsf_doroundup		| positive, round to infinity
+	jra	9b			| negative, round to zero
+	| Exponent overflow.  Just call it infinity.
+fp_nsf_large:
+	tst.b	(3,%a0)
+	jeq	1f
+	fp_set_sr FPSR_EXC_INEX2
+1:	fp_set_sr FPSR_EXC_OVFL
+	move.w	(FPD_RND,FPDATA),%d2
+	jne	3f			| %d2 = 0 round to nearest
+1:	move.w	#0x7fff,(-2,%a0)
+	clr.l	(%a0)+
+	clr.l	(%a0)
+2:	subq.l	#8,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+3:	subq.w	#2,%d2
+	jcs	5f			| %d2 < 2, round to zero
+	jhi	4f			| %d2 > 2, round to +infinity
+	tst.b	(-3,%a0)		| to -inf
+	jne	1b
+	jra	5f
+4:	tst.b	(-3,%a0)		| to +inf
+	jeq	1b
+5:	move.w	#0x407e,(-2,%a0)
+	move.l	#0xffffff00,(%a0)+
+	clr.l	(%a0)
+	jra	2b
+	| Infinities or NaNs
+fp_nsf_huge:
+	subq.l	#4,%a0
+	printf	PNORM,"%p(",1,%a0
+	printx	PNORM,%a0@
+	printf	PNORM,")\n"
+	rts
+
+	| conv_ext2int (macro):
+	| Generates a subroutine that converts an extended value to an
+	| integer of a given size, again, with the appropriate type of
+	| rounding.
+
+	| Macro arguments:
+	| s:	size, as given in an assembly instruction.
+	| b:	number of bits in that size.
+
+	| Subroutine arguments:
+	| %a0:	source (struct fp_ext *)
+
+	| Returns the integer in %d0 (like it should)
+
+.macro conv_ext2int s,b
+	.set	inf,(1<<(\b-1))-1	| i.e. MAXINT
+	printf	PCONV,"e2i%d: %p(",2,#\b,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,") "
+	addq.l	#2,%a0
+	move.w	(%a0)+,%d2		| exponent
+	jeq	fp_e2i_zero\b		| zero / denorm (== 0, here)
+	cmp.w	#0x7fff,%d2
+	jeq	fp_e2i_huge\b		| Inf / NaN
+	sub.w	#0x3ffe,%d2
+	jcs	fp_e2i_small\b
+	cmp.w	#\b,%d2
+	jhi	fp_e2i_large\b
+	move.l	(%a0),%d0
+	move.l	%d0,%d1
+	lsl.l	%d2,%d1
+	jne	fp_e2i_round\b
+	tst.l	(4,%a0)
+	jne	fp_e2i_round\b
+	neg.w	%d2
+	add.w	#32,%d2
+	lsr.l	%d2,%d0
+9:	tst.w	(-4,%a0)
+	jne	1f
+	tst.\s	%d0
+	jmi	fp_e2i_large\b
+	printf	PCONV,"-> %p\n",1,%d0
+	rts
+1:	neg.\s	%d0
+	jeq	1f
+	jpl	fp_e2i_large\b
+1:	printf	PCONV,"-> %p\n",1,%d0
+	rts
+fp_e2i_round\b:
+	fp_set_sr FPSR_EXC_INEX2	| INEX2 bit
+	neg.w	%d2
+	add.w	#32,%d2
+	.if	\b>16
+	jeq	5f
+	.endif
+	lsr.l	%d2,%d0
+	move.w	(FPD_RND,FPDATA),%d2	| rounding mode
+	jne	2f			| %d2 == 0, round to nearest
+	tst.l	%d1			| test guard bit
+	jpl	9b			| zero is closer
+	btst	%d2,%d0			| test lsb bit (%d2 still 0)
+	jne	fp_e2i_doroundup\b
+	lsl.l	#1,%d1			| check low bits
+	jne	fp_e2i_doroundup\b
+	tst.l	(4,%a0)
+	jeq	9b
+fp_e2i_doroundup\b:
+	addq.l	#1,%d0
+	jra	9b
+	| check nondefault rounding modes
+2:	subq.w	#2,%d2
+	jcs	9b			| %d2 < 2, round to zero
+	jhi	3f			| %d2 > 2, round to +infinity
+	tst.w	(-4,%a0)		| to -inf
+	jne	fp_e2i_doroundup\b	| negative, round to infinity
+	jra	9b			| positive, round to zero
+3:	tst.w	(-4,%a0)		| to +inf
+	jeq	fp_e2i_doroundup\b	| positive, round to infinity
+	jra	9b	| negative, round to zero
+	| we are only want -2**127 get correctly rounded here,
+	| since the guard bit is in the lower lword.
+	| everything else ends up anyway as overflow.
+	.if	\b>16
+5:	move.w	(FPD_RND,FPDATA),%d2	| rounding mode
+	jne	2b			| %d2 == 0, round to nearest
+	move.l	(4,%a0),%d1		| test guard bit
+	jpl	9b			| zero is closer
+	lsl.l	#1,%d1			| check low bits
+	jne	fp_e2i_doroundup\b
+	jra	9b
+	.endif
+fp_e2i_zero\b:
+	clr.l	%d0
+	tst.l	(%a0)+
+	jne	1f
+	tst.l	(%a0)
+	jeq	3f
+1:	subq.l	#4,%a0
+	fp_clr_sr FPSR_EXC_UNFL		| fp_normalize_ext has set this bit
+fp_e2i_small\b:
+	fp_set_sr FPSR_EXC_INEX2
+	clr.l	%d0
+	move.w	(FPD_RND,FPDATA),%d2	| rounding mode
+	subq.w	#2,%d2
+	jcs	3f			| %d2 < 2, round to nearest/zero
+	jhi	2f			| %d2 > 2, round to +infinity
+	tst.w	(-4,%a0)		| to -inf
+	jeq	3f
+	subq.\s	#1,%d0
+	jra	3f
+2:	tst.w	(-4,%a0)		| to +inf
+	jne	3f
+	addq.\s	#1,%d0
+3:	printf	PCONV,"-> %p\n",1,%d0
+	rts
+fp_e2i_large\b:
+	fp_set_sr FPSR_EXC_OPERR
+	move.\s	#inf,%d0
+	tst.w	(-4,%a0)
+	jeq	1f
+	addq.\s	#1,%d0
+1:	printf	PCONV,"-> %p\n",1,%d0
+	rts
+fp_e2i_huge\b:
+	move.\s	(%a0),%d0
+	tst.l	(%a0)
+	jne	1f
+	tst.l	(%a0)
+	jeq	fp_e2i_large\b
+	| fp_normalize_ext has set this bit already
+	| and made the number nonsignaling
+1:	fp_tst_sr FPSR_EXC_SNAN
+	jne	1f
+	fp_set_sr FPSR_EXC_OPERR
+1:	printf	PCONV,"-> %p\n",1,%d0
+	rts
+.endm
+
+fp_conv_ext2long:
+	conv_ext2int l,32
+
+fp_conv_ext2short:
+	conv_ext2int w,16
+
+fp_conv_ext2byte:
+	conv_ext2int b,8
+
+fp_conv_ext2double:
+	jsr	fp_normalize_double
+	printf	PCONV,"e2d: %p(",1,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,"), "
+	move.l	(%a0)+,%d2
+	cmp.w	#0x7fff,%d2
+	jne	1f
+	move.w	#0x7ff,%d2
+	move.l	(%a0)+,%d0
+	jra	2f
+1:	sub.w	#0x3fff-0x3ff,%d2
+	move.l	(%a0)+,%d0
+	jmi	2f
+	clr.w	%d2
+2:	lsl.w	#5,%d2
+	lsl.l	#7,%d2
+	lsl.l	#8,%d2
+	move.l	%d0,%d1
+	lsl.l	#1,%d0
+	lsr.l	#4,%d0
+	lsr.l	#8,%d0
+	or.l	%d2,%d0
+	putuser.l %d0,(%a1)+,fp_err_ua2,%a1
+	moveq	#21,%d0
+	lsl.l	%d0,%d1
+	move.l	(%a0),%d0
+	lsr.l	#4,%d0
+	lsr.l	#7,%d0
+	or.l	%d1,%d0
+	putuser.l %d0,(%a1),fp_err_ua2,%a1
+#ifdef FPU_EMU_DEBUG
+	getuser.l %a1@(-4),%d0,fp_err_ua2,%a1
+	getuser.l %a1@(0),%d1,fp_err_ua2,%a1
+	printf	PCONV,"%p(%08x%08x)\n",3,%a1,%d0,%d1
+#endif
+	rts
+
+fp_conv_ext2single:
+	jsr	fp_normalize_single
+	printf	PCONV,"e2s: %p(",1,%a0
+	printx	PCONV,%a0@
+	printf	PCONV,"), "
+	move.l	(%a0)+,%d1
+	cmp.w	#0x7fff,%d1
+	jne	1f
+	move.w	#0xff,%d1
+	move.l	(%a0)+,%d0
+	jra	2f
+1:	sub.w	#0x3fff-0x7f,%d1
+	move.l	(%a0)+,%d0
+	jmi	2f
+	clr.w	%d1
+2:	lsl.w	#8,%d1
+	lsl.l	#7,%d1
+	lsl.l	#8,%d1
+	bclr	#31,%d0
+	lsr.l	#8,%d0
+	or.l	%d1,%d0
+	printf	PCONV,"%08x\n",1,%d0
+	rts
+
+	| special return addresses for instr that
+	| encode the rounding precision in the opcode
+	| (e.g. fsmove,fdmove)
+
+fp_finalrounding_single:
+	addq.l	#8,%sp
+	jsr	fp_normalize_ext
+	jsr	fp_normalize_single
+	jra	fp_finaltest
+
+fp_finalrounding_single_fast:
+	addq.l	#8,%sp 
+	jsr	fp_normalize_ext
+	jsr	fp_normalize_single_fast
+	jra	fp_finaltest
+
+fp_finalrounding_double:
+	addq.l	#8,%sp
+	jsr	fp_normalize_ext
+	jsr	fp_normalize_double
+	jra	fp_finaltest
+
+	| fp_finaltest:
+	| set the emulated status register based on the outcome of an
+	| emulated instruction.
+
+fp_finalrounding:
+	addq.l	#8,%sp
+|	printf	,"f: %p\n",1,%a0
+	jsr	fp_normalize_ext
+	move.w	(FPD_PREC,FPDATA),%d0
+	subq.w	#1,%d0
+	jcs	fp_finaltest
+	jne	1f
+	jsr	fp_normalize_single
+	jra	2f
+1:	jsr	fp_normalize_double
+2:|	printf	,"f: %p\n",1,%a0
+fp_finaltest:
+	| First, we do some of the obvious tests for the exception
+	| status byte and condition code bytes of fp_sr here, so that
+	| they do not have to be handled individually by every
+	| emulated instruction.
+	clr.l	%d0
+	addq.l	#1,%a0
+	tst.b	(%a0)+			| sign
+	jeq	1f
+	bset	#FPSR_CC_NEG-24,%d0	| N bit
+1:	cmp.w	#0x7fff,(%a0)+		| exponent
+	jeq	2f
+	| test for zero
+	moveq	#FPSR_CC_Z-24,%d1
+	tst.l	(%a0)+
+	jne	9f
+	tst.l	(%a0)
+	jne	9f
+	jra	8f
+	| infinitiv and NAN
+2:	moveq	#FPSR_CC_NAN-24,%d1
+	move.l	(%a0)+,%d2
+	lsl.l	#1,%d2			| ignore high bit
+	jne	8f
+	tst.l	(%a0)
+	jne	8f
+	moveq	#FPSR_CC_INF-24,%d1
+8:	bset	%d1,%d0
+9:	move.b	%d0,(FPD_FPSR+0,FPDATA)	| set condition test result
+	| move instructions enter here
+	| Here, we test things in the exception status byte, and set
+	| other things in the accrued exception byte accordingly.
+	| Emulated instructions can set various things in the former,
+	| as defined in fp_emu.h.
+fp_final:
+	move.l	(FPD_FPSR,FPDATA),%d0
+#if 0
+	btst	#FPSR_EXC_SNAN,%d0	| EXC_SNAN
+	jne	1f
+	btst	#FPSR_EXC_OPERR,%d0	| EXC_OPERR
+	jeq	2f
+1:	bset	#FPSR_AEXC_IOP,%d0	| set IOP bit
+2:	btst	#FPSR_EXC_OVFL,%d0	| EXC_OVFL
+	jeq	1f
+	bset	#FPSR_AEXC_OVFL,%d0	| set OVFL bit
+1:	btst	#FPSR_EXC_UNFL,%d0	| EXC_UNFL
+	jeq	1f
+	btst	#FPSR_EXC_INEX2,%d0	| EXC_INEX2
+	jeq	1f
+	bset	#FPSR_AEXC_UNFL,%d0	| set UNFL bit
+1:	btst	#FPSR_EXC_DZ,%d0	| EXC_INEX1
+	jeq	1f
+	bset	#FPSR_AEXC_DZ,%d0	| set DZ bit
+1:	btst	#FPSR_EXC_OVFL,%d0	| EXC_OVFL
+	jne	1f
+	btst	#FPSR_EXC_INEX2,%d0	| EXC_INEX2
+	jne	1f
+	btst	#FPSR_EXC_INEX1,%d0	| EXC_INEX1
+	jeq	2f
+1:	bset	#FPSR_AEXC_INEX,%d0	| set INEX bit
+2:	move.l	%d0,(FPD_FPSR,FPDATA)
+#else
+	| same as above, greatly optimized, but untested (yet)
+	move.l	%d0,%d2
+	lsr.l	#5,%d0
+	move.l	%d0,%d1
+	lsr.l	#4,%d1
+	or.l	%d0,%d1
+	and.b	#0x08,%d1
+	move.l	%d2,%d0
+	lsr.l	#6,%d0
+	or.l	%d1,%d0
+	move.l	%d2,%d1
+	lsr.l	#4,%d1
+	or.b	#0xdf,%d1
+	and.b	%d1,%d0
+	move.l	%d2,%d1
+	lsr.l	#7,%d1
+	and.b	#0x80,%d1
+	or.b	%d1,%d0
+	and.b	#0xf8,%d0
+	or.b	%d0,%d2
+	move.l	%d2,(FPD_FPSR,FPDATA)
+#endif
+	move.b	(FPD_FPSR+2,FPDATA),%d0
+	and.b	(FPD_FPCR+2,FPDATA),%d0
+	jeq	1f
+	printf	,"send signal!!!\n"
+1:	jra	fp_end

FUNET's LINUX-ADM group, linux-adm@nic.funet.fi
TCL-scripts by Sam Shen (who was at: slshen@lbl.gov)