diff options
| author | Kyle McMartin <kyle@mcmartin.ca> | 2007-10-18 00:06:26 -0700 |
|---|---|---|
| committer | Kyle McMartin <kyle@shortfin.cabal.ca> | 2007-10-18 00:58:49 -0700 |
| commit | efb80e7e097d0888e59fbbe4ded2ac5a256f556d (patch) | |
| tree | 98a0f2f1514501aeebb1877bfcb5b528491e5ad5 /arch | |
| parent | 6f7d998e94ec7b7f08bd0c72fc05343435d7fa93 (diff) | |
[PARISC] import necessary bits of libgcc.a
Currently we're hacking libs-y to include libgcc.a, but this has
unforeseen consequences since the userspace libgcc is linked with fpregs
enabled. We need the kernel to stop using fpregs in an uncontrolled manner
to implement lazy fpu state saves.
Signed-off-by: Kyle McMartin <kyle@mcmartin.ca>
Diffstat (limited to 'arch')
31 files changed, 4628 insertions, 24 deletions
diff --git a/arch/parisc/Makefile b/arch/parisc/Makefile index f3d0d7c7097..20d327f63d5 100644 --- a/arch/parisc/Makefile +++ b/arch/parisc/Makefile @@ -69,7 +69,7 @@ kernel-y := mm/ kernel/ math-emu/ kernel/init_task.o kernel-$(CONFIG_HPUX) += hpux/ core-y += $(addprefix arch/parisc/, $(kernel-y)) -libs-y += arch/parisc/lib/ `$(CC) -print-libgcc-file-name` +libs-y += arch/parisc/lib/ drivers-$(CONFIG_OPROFILE) += arch/parisc/oprofile/ diff --git a/arch/parisc/kernel/parisc_ksyms.c b/arch/parisc/kernel/parisc_ksyms.c index 7aca704e96f..671ee5b9950 100644 --- a/arch/parisc/kernel/parisc_ksyms.c +++ b/arch/parisc/kernel/parisc_ksyms.c @@ -122,31 +122,9 @@ EXPORT_SYMBOL($$divI_12); EXPORT_SYMBOL($$divI_14); EXPORT_SYMBOL($$divI_15); -extern void __ashrdi3(void); -extern void __ashldi3(void); -extern void __lshrdi3(void); -extern void __muldi3(void); - -EXPORT_SYMBOL(__ashrdi3); -EXPORT_SYMBOL(__ashldi3); -EXPORT_SYMBOL(__lshrdi3); -EXPORT_SYMBOL(__muldi3); - asmlinkage void * __canonicalize_funcptr_for_compare(void *); EXPORT_SYMBOL(__canonicalize_funcptr_for_compare); -#ifdef CONFIG_64BIT -extern void __divdi3(void); -extern void __udivdi3(void); -extern void __umoddi3(void); -extern void __moddi3(void); - -EXPORT_SYMBOL(__divdi3); -EXPORT_SYMBOL(__udivdi3); -EXPORT_SYMBOL(__umoddi3); -EXPORT_SYMBOL(__moddi3); -#endif - #ifndef CONFIG_64BIT extern void $$dyncall(void); EXPORT_SYMBOL($$dyncall); diff --git a/arch/parisc/lib/Makefile b/arch/parisc/lib/Makefile index 5f2e6904d14..7ce406c7daf 100644 --- a/arch/parisc/lib/Makefile +++ b/arch/parisc/lib/Makefile @@ -4,4 +4,4 @@ lib-y := lusercopy.o bitops.o checksum.o io.o memset.o fixup.o memcpy.o -obj-y := iomap.o +obj-y := libgcc/ milli/ iomap.o diff --git a/arch/parisc/lib/libgcc/Makefile b/arch/parisc/lib/libgcc/Makefile new file mode 100644 index 00000000000..b67a85ad9c8 --- /dev/null +++ b/arch/parisc/lib/libgcc/Makefile @@ -0,0 +1,4 @@ +obj-y := __ashldi3.o __ashrdi3.o __clzsi2.o __divdi3.o __divsi3.o \ + __lshrdi3.o __moddi3.o __modsi3.o __udivdi3.o \ + __udivmoddi4.o __udivmodsi4.o __udivsi3.o \ + __umoddi3.o __umodsi3.o __muldi3.o __umulsidi3.o diff --git a/arch/parisc/lib/libgcc/__ashldi3.c b/arch/parisc/lib/libgcc/__ashldi3.c new file mode 100644 index 00000000000..a14a257abb2 --- /dev/null +++ b/arch/parisc/lib/libgcc/__ashldi3.c @@ -0,0 +1,19 @@ +#include "libgcc.h" + +u64 __ashldi3(u64 v, int cnt) +{ + int c = cnt & 31; + u32 vl = (u32) v; + u32 vh = (u32) (v >> 32); + + if (cnt & 32) { + vh = (vl << c); + vl = 0; + } else { + vh = (vh << c) + (vl >> (32 - c)); + vl = (vl << c); + } + + return ((u64) vh << 32) + vl; +} +EXPORT_SYMBOL(__ashldi3); diff --git a/arch/parisc/lib/libgcc/__ashrdi3.c b/arch/parisc/lib/libgcc/__ashrdi3.c new file mode 100644 index 00000000000..8636a5aa4f7 --- /dev/null +++ b/arch/parisc/lib/libgcc/__ashrdi3.c @@ -0,0 +1,19 @@ +#include "libgcc.h" + +u64 __ashrdi3(u64 v, int cnt) +{ + int c = cnt & 31; + u32 vl = (u32) v; + u32 vh = (u32) (v >> 32); + + if (cnt & 32) { + vl = ((s32) vh >> c); + vh = (s32) vh >> 31; + } else { + vl = (vl >> c) + (vh << (32 - c)); + vh = ((s32) vh >> c); + } + + return ((u64) vh << 32) + vl; +} +EXPORT_SYMBOL(__ashrdi3); diff --git a/arch/parisc/lib/libgcc/__clzsi2.c b/arch/parisc/lib/libgcc/__clzsi2.c new file mode 100644 index 00000000000..a7aa2f55a9c --- /dev/null +++ b/arch/parisc/lib/libgcc/__clzsi2.c @@ -0,0 +1,30 @@ +#include "libgcc.h" + +u32 __clzsi2(u32 v) +{ + int p = 31; + + if (v & 0xffff0000) { + p -= 16; + v >>= 16; + } + if (v & 0xff00) { + p -= 8; + v >>= 8; + } + if (v & 0xf0) { + p -= 4; + v >>= 4; + } + if (v & 0xc) { + p -= 2; + v >>= 2; + } + if (v & 0x2) { + p -= 1; + v >>= 1; + } + + return p; +} +EXPORT_SYMBOL(__clzsi2); diff --git a/arch/parisc/lib/libgcc/__divdi3.c b/arch/parisc/lib/libgcc/__divdi3.c new file mode 100644 index 00000000000..f23c6fe2838 --- /dev/null +++ b/arch/parisc/lib/libgcc/__divdi3.c @@ -0,0 +1,23 @@ +#include "libgcc.h" + +s64 __divdi3(s64 num, s64 den) +{ + int minus = 0; + s64 v; + + if (num < 0) { + num = -num; + minus = 1; + } + if (den < 0) { + den = -den; + minus ^= 1; + } + + v = __udivmoddi4(num, den, NULL); + if (minus) + v = -v; + + return v; +} +EXPORT_SYMBOL(__divdi3); diff --git a/arch/parisc/lib/libgcc/__divsi3.c b/arch/parisc/lib/libgcc/__divsi3.c new file mode 100644 index 00000000000..730fb530680 --- /dev/null +++ b/arch/parisc/lib/libgcc/__divsi3.c @@ -0,0 +1,23 @@ +#include "libgcc.h" + +s32 __divsi3(s32 num, s32 den) +{ + int minus = 0; + s32 v; + + if (num < 0) { + num = -num; + minus = 1; + } + if (den < 0) { + den = -den; + minus ^= 1; + } + + v = __udivmodsi4(num, den, NULL); + if (minus) + v = -v; + + return v; +} +EXPORT_SYMBOL(__divsi3); diff --git a/arch/parisc/lib/libgcc/__lshrdi3.c b/arch/parisc/lib/libgcc/__lshrdi3.c new file mode 100644 index 00000000000..4a820708ec5 --- /dev/null +++ b/arch/parisc/lib/libgcc/__lshrdi3.c @@ -0,0 +1,19 @@ +#include "libgcc.h" + +u64 __lshrdi3(u64 v, int cnt) +{ + int c = cnt & 31; + u32 vl = (u32) v; + u32 vh = (u32) (v >> 32); + + if (cnt & 32) { + vl = (vh >> c); + vh = 0; + } else { + vl = (vl >> c) + (vh << (32 - c)); + vh = (vh >> c); + } + + return ((u64) vh << 32) + vl; +} +EXPORT_SYMBOL(__lshrdi3); diff --git a/arch/parisc/lib/libgcc/__moddi3.c b/arch/parisc/lib/libgcc/__moddi3.c new file mode 100644 index 00000000000..ed64bbafc98 --- /dev/null +++ b/arch/parisc/lib/libgcc/__moddi3.c @@ -0,0 +1,23 @@ +#include "libgcc.h" + +s64 __moddi3(s64 num, s64 den) +{ + int minus = 0; + s64 v; + + if (num < 0) { + num = -num; + minus = 1; + } + if (den < 0) { + den = -den; + minus ^= 1; + } + + (void)__udivmoddi4(num, den, (u64 *) & v); + if (minus) + v = -v; + + return v; +} +EXPORT_SYMBOL(__moddi3); diff --git a/arch/parisc/lib/libgcc/__modsi3.c b/arch/parisc/lib/libgcc/__modsi3.c new file mode 100644 index 00000000000..62f773efaee --- /dev/null +++ b/arch/parisc/lib/libgcc/__modsi3.c @@ -0,0 +1,23 @@ +#include "libgcc.h" + +s32 __modsi3(s32 num, s32 den) +{ + int minus = 0; + s32 v; + + if (num < 0) { + num = -num; + minus = 1; + } + if (den < 0) { + den = -den; + minus ^= 1; + } + + (void)__udivmodsi4(num, den, (u32 *) & v); + if (minus) + v = -v; + + return v; +} +EXPORT_SYMBOL(__modsi3); diff --git a/arch/parisc/lib/libgcc/__muldi3.c b/arch/parisc/lib/libgcc/__muldi3.c new file mode 100644 index 00000000000..3308abdd558 --- /dev/null +++ b/arch/parisc/lib/libgcc/__muldi3.c @@ -0,0 +1,22 @@ +#include "libgcc.h" + +union DWunion { + struct { + s32 high; + s32 low; + } s; + s64 ll; +}; + +s64 __muldi3(s64 u, s64 v) +{ + const union DWunion uu = { .ll = u }; + const union DWunion vv = { .ll = v }; + union DWunion w = { .ll = __umulsidi3(uu.s.low, vv.s.low) }; + + w.s.high += ((u32)uu.s.low * (u32)vv.s.high + + (u32)uu.s.high * (u32)vv.s.low); + + return w.ll; +} +EXPORT_SYMBOL(__muldi3); diff --git a/arch/parisc/lib/libgcc/__udivdi3.c b/arch/parisc/lib/libgcc/__udivdi3.c new file mode 100644 index 00000000000..740023d690f --- /dev/null +++ b/arch/parisc/lib/libgcc/__udivdi3.c @@ -0,0 +1,7 @@ +#include "libgcc.h" + +u64 __udivdi3(u64 num, u64 den) +{ + return __udivmoddi4(num, den, NULL); +} +EXPORT_SYMBOL(__udivdi3); diff --git a/arch/parisc/lib/libgcc/__udivmoddi4.c b/arch/parisc/lib/libgcc/__udivmoddi4.c new file mode 100644 index 00000000000..2df0caa5a7d --- /dev/null +++ b/arch/parisc/lib/libgcc/__udivmoddi4.c @@ -0,0 +1,31 @@ +#include "libgcc.h" + +u64 __udivmoddi4(u64 num, u64 den, u64 * rem_p) +{ + u64 quot = 0, qbit = 1; + + if (den == 0) { + BUG(); + } + + /* Left-justify denominator and count shift */ + while ((s64) den >= 0) { + den <<= 1; + qbit <<= 1; + } + + while (qbit) { + if (den <= num) { + num -= den; + quot += qbit; + } + den >>= 1; + qbit >>= 1; + } + + if (rem_p) + *rem_p = num; + + return quot; +} +EXPORT_SYMBOL(__udivmoddi4); diff --git a/arch/parisc/lib/libgcc/__udivmodsi4.c b/arch/parisc/lib/libgcc/__udivmodsi4.c new file mode 100644 index 00000000000..2a2fc28b202 --- /dev/null +++ b/arch/parisc/lib/libgcc/__udivmodsi4.c @@ -0,0 +1,31 @@ +#include "libgcc.h" + +u32 __udivmodsi4(u32 num, u32 den, u32 * rem_p) +{ + u32 quot = 0, qbit = 1; + + if (den == 0) { + BUG(); + } + + /* Left-justify denominator and count shift */ + while ((s32) den >= 0) { + den <<= 1; + qbit <<= 1; + } + + while (qbit) { + if (den <= num) { + num -= den; + quot += qbit; + } + den >>= 1; + qbit >>= 1; + } + + if (rem_p) + *rem_p = num; + + return quot; +} +EXPORT_SYMBOL(__udivmodsi4); diff --git a/arch/parisc/lib/libgcc/__udivsi3.c b/arch/parisc/lib/libgcc/__udivsi3.c new file mode 100644 index 00000000000..756a44164e9 --- /dev/null +++ b/arch/parisc/lib/libgcc/__udivsi3.c @@ -0,0 +1,7 @@ +#include "libgcc.h" + +u32 __udivsi3(u32 num, u32 den) +{ + return __udivmodsi4(num, den, NULL); +} +EXPORT_SYMBOL(__udivsi3); diff --git a/arch/parisc/lib/libgcc/__umoddi3.c b/arch/parisc/lib/libgcc/__umoddi3.c new file mode 100644 index 00000000000..ac744e948bc --- /dev/null +++ b/arch/parisc/lib/libgcc/__umoddi3.c @@ -0,0 +1,10 @@ +#include "libgcc.h" + +u64 __umoddi3(u64 num, u64 den) +{ + u64 v; + + (void)__udivmoddi4(num, den, &v); + return v; +} +EXPORT_SYMBOL(__umoddi3); diff --git a/arch/parisc/lib/libgcc/__umodsi3.c b/arch/parisc/lib/libgcc/__umodsi3.c new file mode 100644 index 00000000000..51f55aa89f9 --- /dev/null +++ b/arch/parisc/lib/libgcc/__umodsi3.c @@ -0,0 +1,10 @@ +#include "libgcc.h" + +u32 __umodsi3(u32 num, u32 den) +{ + u32 v; + + (void)__udivmodsi4(num, den, &v); + return v; +} +EXPORT_SYMBOL(__umodsi3); diff --git a/arch/parisc/lib/libgcc/__umulsidi3.c b/arch/parisc/lib/libgcc/__umulsidi3.c new file mode 100644 index 00000000000..396f669164d --- /dev/null +++ b/arch/parisc/lib/libgcc/__umulsidi3.c @@ -0,0 +1,46 @@ +#include "libgcc.h" + +#define __ll_B ((u32) 1 << (32 / 2)) +#define __ll_lowpart(t) ((u32) (t) & (__ll_B - 1)) +#define __ll_highpart(t) ((u32) (t) >> 16) + +#define umul_ppmm(w1, w0, u, v) \ + do { \ + u32 __x0, __x1, __x2, __x3; \ + u16 __ul, __vl, __uh, __vh; \ + \ + __ul = __ll_lowpart (u); \ + __uh = __ll_highpart (u); \ + __vl = __ll_lowpart (v); \ + __vh = __ll_highpart (v); \ + \ + __x0 = (u32) __ul * __vl; \ + __x1 = (u32) __ul * __vh; \ + __x2 = (u32) __uh * __vl; \ + __x3 = (u32) __uh * __vh; \ + \ + __x1 += __ll_highpart (__x0);/* this can't give carry */ \ + __x1 += __x2; /* but this indeed can */ \ + if (__x1 < __x2) /* did we get it? */ \ + __x3 += __ll_B; /* yes, add it in the proper pos. */ \ + \ + (w1) = __x3 + __ll_highpart (__x1); \ + (w0) = __ll_lowpart (__x1) * __ll_B + __ll_lowpart (__x0); \ + } while (0) + +union DWunion { + struct { + s32 high; + s32 low; + } s; + s64 ll; +}; + +u64 __umulsidi3(u32 u, u32 v) +{ + union DWunion __w; + + umul_ppmm(__w.s.high, __w.s.low, u, v); + + return __w.ll; +} diff --git a/arch/parisc/lib/libgcc/libgcc.h b/arch/parisc/lib/libgcc/libgcc.h new file mode 100644 index 00000000000..5a6f7a510fb --- /dev/null +++ b/arch/parisc/lib/libgcc/libgcc.h @@ -0,0 +1,32 @@ +#ifndef _PA_LIBGCC_H_ +#define _PA_LIBGCC_H_ + +#include <linux/types.h> +#include <linux/module.h> + +/* Cribbed from klibc/libgcc/ */ +u64 __ashldi3(u64 v, int cnt); +u64 __ashrdi3(u64 v, int cnt); + +u32 __clzsi2(u32 v); + +s64 __divdi3(s64 num, s64 den); +s32 __divsi3(s32 num, s32 den); + +u64 __lshrdi3(u64 v, int cnt); + +s64 __moddi3(s64 num, s64 den); +s32 __modsi3(s32 num, s32 den); + +u64 __udivdi3(u64 num, u64 den); +u32 __udivsi3(u32 num, u32 den); + +u64 __udivmoddi4(u64 num, u64 den, u64 * rem_p); +u32 __udivmodsi4(u32 num, u32 den, u32 * rem_p); + +u64 __umulsidi3(u32 u, u32 v); + +u64 __umoddi3(u64 num, u64 den); +u32 __umodsi3(u32 num, u32 den); + +#endif /*_PA_LIBGCC_H_*/ diff --git a/arch/parisc/lib/milli/Makefile b/arch/parisc/lib/milli/Makefile new file mode 100644 index 00000000000..9b24e9b1f3c --- /dev/null +++ b/arch/parisc/lib/milli/Makefile @@ -0,0 +1 @@ +obj-y := dyncall.o divI.o divU.o remI.o remU.o div_const.o mulI.o diff --git a/arch/parisc/lib/milli/divI.S b/arch/parisc/lib/milli/divI.S new file mode 100644 index 00000000000..ac106b7b6f2 --- /dev/null +++ b/arch/parisc/lib/milli/divI.S @@ -0,0 +1,254 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_divI +/* ROUTINES: $$divI, $$divoI + + Single precision divide for signed binary integers. + + The quotient is truncated towards zero. + The sign of the quotient is the XOR of the signs of the dividend and + divisor. + Divide by zero is trapped. + Divide of -2**31 by -1 is trapped for $$divoI but not for $$divI. + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = quotient + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: + . divisor is zero (traps with ADDIT,= 0,25,0) + . dividend==-2**31 and divisor==-1 and routine is $$divoI + . (traps with ADDO 26,25,0) + . Changes memory at the following places: + . NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Branchs to other millicode routines using BE + . $$div_# for # being 2,3,4,5,6,7,8,9,10,12,14,15 + . + . For selected divisors, calls a divide by constant routine written by + . Karl Pettis. Eligible divisors are 1..15 excluding 11 and 13. + . + . The only overflow case is -2**31 divided by -1. + . Both routines return -2**31 but only $$divoI traps. */ + +RDEFINE(temp,r1) +RDEFINE(retreg,ret1) /* r29 */ +RDEFINE(temp1,arg0) + SUBSPA_MILLI_DIV + ATTR_MILLI + .import $$divI_2,millicode + .import $$divI_3,millicode + .import $$divI_4,millicode + .import $$divI_5,millicode + .import $$divI_6,millicode + .import $$divI_7,millicode + .import $$divI_8,millicode + .import $$divI_9,millicode + .import $$divI_10,millicode + .import $$divI_12,millicode + .import $$divI_14,millicode + .import $$divI_15,millicode + .export $$divI,millicode + .export $$divoI,millicode + .proc + .callinfo millicode + .entry +GSYM($$divoI) + comib,=,n -1,arg1,LREF(negative1) /* when divisor == -1 */ +GSYM($$divI) + ldo -1(arg1),temp /* is there at most one bit set ? */ + and,<> arg1,temp,r0 /* if not, don't use power of 2 divide */ + addi,> 0,arg1,r0 /* if divisor > 0, use power of 2 divide */ + b,n LREF(neg_denom) +LSYM(pow2) + addi,>= 0,arg0,retreg /* if numerator is negative, add the */ + add arg0,temp,retreg /* (denominaotr -1) to correct for shifts */ + extru,= arg1,15,16,temp /* test denominator with 0xffff0000 */ + extrs retreg,15,16,retreg /* retreg = retreg >> 16 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 16) */ + ldi 0xcc,temp1 /* setup 0xcc in temp1 */ + extru,= arg1,23,8,temp /* test denominator with 0xff00 */ + extrs retreg,23,24,retreg /* retreg = retreg >> 8 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 8) */ + ldi 0xaa,temp /* setup 0xaa in temp */ + extru,= arg1,27,4,r0 /* test denominator with 0xf0 */ + extrs retreg,27,28,retreg /* retreg = retreg >> 4 */ + and,= arg1,temp1,r0 /* test denominator with 0xcc */ + extrs retreg,29,30,retreg /* retreg = retreg >> 2 */ + and,= arg1,temp,r0 /* test denominator with 0xaa */ + extrs retreg,30,31,retreg /* retreg = retreg >> 1 */ + MILLIRETN +LSYM(neg_denom) + addi,< 0,arg1,r0 /* if arg1 >= 0, it's not power of 2 */ + b,n LREF(regular_seq) + sub r0,arg1,temp /* make denominator positive */ + comb,=,n arg1,temp,LREF(regular_seq) /* test against 0x80000000 and 0 */ + ldo -1(temp),retreg /* is there at most one bit set ? */ + and,= temp,retreg,r0 /* if so, the denominator is power of 2 */ + b,n LREF(regular_seq) + sub r0,arg0,retreg /* negate numerator */ + comb,=,n arg0,retreg,LREF(regular_seq) /* test against 0x80000000 */ + copy retreg,arg0 /* set up arg0, arg1 and temp */ + copy temp,arg1 /* before branching to pow2 */ + b LREF(pow2) + ldo -1(arg1),temp +LSYM(regular_seq) + comib,>>=,n 15,arg1,LREF(small_divisor) + add,>= 0,arg0,retreg /* move dividend, if retreg < 0, */ +LSYM(normal) + subi 0,retreg,retreg /* make it positive */ + sub 0,arg1,temp /* clear carry, */ + /* negate the divisor */ + ds 0,temp,0 /* set V-bit to the comple- */ + /* ment of the divisor sign */ + add retreg,retreg,retreg /* shift msb bit into carry */ + ds r0,arg1,temp /* 1st divide step, if no carry */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 2nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 3rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 4th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 5th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 6th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 7th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 8th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 9th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 10th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 11th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 12th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 13th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 14th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 15th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 16th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 17th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 18th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 19th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 20th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 21st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 22nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 23rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 24th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 25th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 26th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 27th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 28th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 29th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 30th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 31st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 32nd divide step, */ + addc retreg,retreg,retreg /* shift last retreg bit into retreg */ + xor,>= arg0,arg1,0 /* get correct sign of quotient */ + sub 0,retreg,retreg /* based on operand signs */ + MILLIRETN + nop + +LSYM(small_divisor) + +#if defined(CONFIG_64BIT) +/* Clear the upper 32 bits of the arg1 register. We are working with */ +/* small divisors (and 32-bit integers) We must not be mislead */ +/* by "1" bits left in the upper 32 bits. */ + depd %r0,31,32,%r25 +#endif + blr,n arg1,r0 + nop +/* table for divisor == 0,1, ... ,15 */ + addit,= 0,arg1,r0 /* trap if divisor == 0 */ + nop + MILLIRET /* divisor == 1 */ + copy arg0,retreg + MILLI_BEN($$divI_2) /* divisor == 2 */ + nop + MILLI_BEN($$divI_3) /* divisor == 3 */ + nop + MILLI_BEN($$divI_4) /* divisor == 4 */ + nop + MILLI_BEN($$divI_5) /* divisor == 5 */ + nop + MILLI_BEN($$divI_6) /* divisor == 6 */ + nop + MILLI_BEN($$divI_7) /* divisor == 7 */ + nop + MILLI_BEN($$divI_8) /* divisor == 8 */ + nop + MILLI_BEN($$divI_9) /* divisor == 9 */ + nop + MILLI_BEN($$divI_10) /* divisor == 10 */ + nop + b LREF(normal) /* divisor == 11 */ + add,>= 0,arg0,retreg + MILLI_BEN($$divI_12) /* divisor == 12 */ + nop + b LREF(normal) /* divisor == 13 */ + add,>= 0,arg0,retreg + MILLI_BEN($$divI_14) /* divisor == 14 */ + nop + MILLI_BEN($$divI_15) /* divisor == 15 */ + nop + +LSYM(negative1) + sub 0,arg0,retreg /* result is negation of dividend */ + MILLIRET + addo arg0,arg1,r0 /* trap iff dividend==0x80000000 && divisor==-1 */ + .exit + .procend + .end +#endif diff --git a/arch/parisc/lib/milli/divU.S b/arch/parisc/lib/milli/divU.S new file mode 100644 index 00000000000..9287fe2546f --- /dev/null +++ b/arch/parisc/lib/milli/divU.S @@ -0,0 +1,235 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_divU +/* ROUTINE: $$divU + . + . Single precision divide for unsigned integers. + . + . Quotient is truncated towards zero. + . Traps on divide by zero. + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = quotient + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: + . divisor is zero + . Changes memory at the following places: + . NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Does not create a stack frame. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Branchs to other millicode routines using BE: + . $$divU_# for 3,5,6,7,9,10,12,14,15 + . + . For selected small divisors calls the special divide by constant + . routines written by Karl Pettis. These are: 3,5,6,7,9,10,12,14,15. */ + +RDEFINE(temp,r1) +RDEFINE(retreg,ret1) /* r29 */ +RDEFINE(temp1,arg0) + SUBSPA_MILLI_DIV + ATTR_MILLI + .export $$divU,millicode + .import $$divU_3,millicode + .import $$divU_5,millicode + .import $$divU_6,millicode + .import $$divU_7,millicode + .import $$divU_9,millicode + .import $$divU_10,millicode + .import $$divU_12,millicode + .import $$divU_14,millicode + .import $$divU_15,millicode + .proc + .callinfo millicode + .entry +GSYM($$divU) +/* The subtract is not nullified since it does no harm and can be used + by the two cases that branch back to "normal". */ + ldo -1(arg1),temp /* is there at most one bit set ? */ + and,= arg1,temp,r0 /* if so, denominator is power of 2 */ + b LREF(regular_seq) + addit,= 0,arg1,0 /* trap for zero dvr */ + copy arg0,retreg + extru,= arg1,15,16,temp /* test denominator with 0xffff0000 */ + extru retreg,15,16,retreg /* retreg = retreg >> 16 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 16) */ + ldi 0xcc,temp1 /* setup 0xcc in temp1 */ + extru,= arg1,23,8,temp /* test denominator with 0xff00 */ + extru retreg,23,24,retreg /* retreg = retreg >> 8 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 8) */ + ldi 0xaa,temp /* setup 0xaa in temp */ + extru,= arg1,27,4,r0 /* test denominator with 0xf0 */ + extru retreg,27,28,retreg /* retreg = retreg >> 4 */ + and,= arg1,temp1,r0 /* test denominator with 0xcc */ + extru retreg,29,30,retreg /* retreg = retreg >> 2 */ + and,= arg1,temp,r0 /* test denominator with 0xaa */ + extru retreg,30,31,retreg /* retreg = retreg >> 1 */ + MILLIRETN + nop +LSYM(regular_seq) + comib,>= 15,arg1,LREF(special_divisor) + subi 0,arg1,temp /* clear carry, negate the divisor */ + ds r0,temp,r0 /* set V-bit to 1 */ +LSYM(normal) + add arg0,arg0,retreg /* shift msb bit into carry */ + ds r0,arg1,temp /* 1st divide step, if no carry */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 2nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 3rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 4th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 5th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 6th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 7th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 8th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 9th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 10th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 11th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 12th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 13th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 14th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 15th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 16th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 17th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 18th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 19th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 20th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 21st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 22nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 23rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 24th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 25th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 26th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 27th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 28th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 29th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 30th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 31st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 32nd divide step, */ + MILLIRET + addc retreg,retreg,retreg /* shift last retreg bit into retreg */ + +/* Handle the cases where divisor is a small constant or has high bit on. */ +LSYM(special_divisor) +/* blr arg1,r0 */ +/* comib,>,n 0,arg1,LREF(big_divisor) ; nullify previous instruction */ + +/* Pratap 8/13/90. The 815 Stirling chip set has a bug that prevents us from + generating such a blr, comib sequence. A problem in nullification. So I + rewrote this code. */ + +#if defined(CONFIG_64BIT) +/* Clear the upper 32 bits of the arg1 register. We are working with + small divisors (and 32-bit unsigned integers) We must not be mislead + by "1" bits left in the upper 32 bits. */ + depd %r0,31,32,%r25 +#endif + comib,> 0,arg1,LREF(big_divisor) + nop + blr arg1,r0 + nop + +LSYM(zero_divisor) /* this label is here to provide external visibility */ + addit,= 0,arg1,0 /* trap for zero dvr */ + nop + MILLIRET /* divisor == 1 */ + copy arg0,retreg + MILLIRET /* divisor == 2 */ + extru arg0,30,31,retreg + MILLI_BEN($$divU_3) /* divisor == 3 */ + nop + MILLIRET /* divisor == 4 */ + extru arg0,29,30,retreg + MILLI_BEN($$divU_5) /* divisor == 5 */ + nop + MILLI_BEN($$divU_6) /* divisor == 6 */ + nop + MILLI_BEN($$divU_7) /* divisor == 7 */ + nop + MILLIRET /* divisor == 8 */ + extru arg0,28,29,retreg + MILLI_BEN($$divU_9) /* divisor == 9 */ + nop + MILLI_BEN($$divU_10) /* divisor == 10 */ + nop + b LREF(normal) /* divisor == 11 */ + ds r0,temp,r0 /* set V-bit to 1 */ + MILLI_BEN($$divU_12) /* divisor == 12 */ + nop + b LREF(normal) /* divisor == 13 */ + ds r0,temp,r0 /* set V-bit to 1 */ + MILLI_BEN($$divU_14) /* divisor == 14 */ + nop + MILLI_BEN($$divU_15) /* divisor == 15 */ + nop + +/* Handle the case where the high bit is on in the divisor. + Compute: if( dividend>=divisor) quotient=1; else quotient=0; + Note: dividend>==divisor iff dividend-divisor does not borrow + and not borrow iff carry. */ +LSYM(big_divisor) + sub arg0,arg1,r0 + MILLIRET + addc r0,r0,retreg + .exit + .procend + .end +#endif diff --git a/arch/parisc/lib/milli/div_const.S b/arch/parisc/lib/milli/div_const.S new file mode 100644 index 00000000000..dd660076e94 --- /dev/null +++ b/arch/parisc/lib/milli/div_const.S @@ -0,0 +1,682 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_div_const +/* ROUTINE: $$divI_2 + . $$divI_3 $$divU_3 + . $$divI_4 + . $$divI_5 $$divU_5 + . $$divI_6 $$divU_6 + . $$divI_7 $$divU_7 + . $$divI_8 + . $$divI_9 $$divU_9 + . $$divI_10 $$divU_10 + . + . $$divI_12 $$divU_12 + . + . $$divI_14 $$divU_14 + . $$divI_15 $$divU_15 + . $$divI_16 + . $$divI_17 $$divU_17 + . + . Divide by selected constants for single precision binary integers. + + INPUT REGISTERS: + . arg0 == dividend + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = quotient + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: NONE + . Changes memory at the following places: NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Does not create a stack frame. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Calls other millicode routines using mrp: NONE + . Calls other millicode routines: NONE */ + + +/* TRUNCATED DIVISION BY SMALL INTEGERS + + We are interested in q(x) = floor(x/y), where x >= 0 and y > 0 + (with y fixed). + + Let a = floor(z/y), for some choice of z. Note that z will be + chosen so that division by z is cheap. + + Let r be the remainder(z/y). In other words, r = z - ay. + + Now, our method is to choose a value for b such that + + q'(x) = floor((ax+b)/z) + + is equal to q(x) over as large a range of x as possible. If the + two are equal over a sufficiently large range, and if it is easy to + form the product (ax), and it is easy to divide by z, then we can + perform the division much faster than the general division algorithm. + + So, we want the following to be true: + + . For x in the following range: + . + . ky <= x < (k+1)y + . + . implies that + . + . k <= (ax+b)/z < (k+1) + + We want to determine b such that this is true for all k in the + range {0..K} for some maximum K. + + Since (ax+b) is an increasing function of x, we can take each + bound separately to determine the "best" value for b. + + (ax+b)/z < (k+1) implies + + (a((k+1)y-1)+b < (k+1)z implies + + b < a + (k+1)(z-ay) implies + + b < a + (k+1)r + + This needs to be true for all k in the range {0..K}. In + particular, it is true for k = 0 and this leads to a maximum + acceptable value for b. + + b < a+r or b <= a+r-1 + + Taking the other bound, we have + + k <= (ax+b)/z implies + + k <= (aky+b)/z implies + + k(z-ay) <= b implies + + kr <= b + + Clearly, the largest range for k will be achieved by maximizing b, + when r is not zero. When r is zero, then the simplest choice for b + is 0. When r is not 0, set + + . b = a+r-1 + + Now, by construction, q'(x) = floor((ax+b)/z) = q(x) = floor(x/y) + for all x in the range: + + . 0 <= x < (K+1)y + + We need to determine what K is. Of our two bounds, + + . b < a+(k+1)r is satisfied for all k >= 0, by construction. + + The other bound is + + . kr <= b + + This is always true if r = 0. If r is not 0 (the usual case), then + K = floor((a+r-1)/r), is the maximum value for k. + + Therefore, the formula q'(x) = floor((ax+b)/z) yields the correct + answer for q(x) = floor(x/y) when x is in the range + + (0,(K+1)y-1) K = floor((a+r-1)/r) + + To be most useful, we want (K+1)y-1 = (max x) >= 2**32-1 so that + the formula for q'(x) yields the correct value of q(x) for all x + representable by a single word in HPPA. + + We are also constrained in that computing the product (ax), adding + b, and dividing by z must all be done quickly, otherwise we will be + better off going through the general algorithm using the DS + instruction, which uses approximately 70 cycles. + + For each y, there is a choice of z which satisfies the constraints + for (K+1)y >= 2**32. We may not, however, be able to satisfy the + timing constraints for arbitrary y. It seems that z being equal to + a power of 2 or a power of 2 minus 1 is as good as we can do, since + it minimizes the time to do division by z. We want the choice of z + to also result in a value for (a) that minimizes the computation of + the product (ax). This is best achieved if (a) has a regular bit + pattern (so the multiplication can be done with shifts and adds). + The value of (a) also needs to be less than 2**32 so the product is + always guaranteed to fit in 2 words. + + In actual practice, the following should be done: + + 1) For negative x, you should take the absolute value and remember + . the fact so that the result can be negated. This obviously does + . not apply in the unsigned case. + 2) For even y, you should factor out the power of 2 that divides y + . and divide x by it. You can then proceed by dividing by the + . odd factor of y. + + Here is a table of some odd values of y, and corresponding choices + for z which are "good". + + y z r a (hex) max x (hex) + + 3 2**32 1 55555555 100000001 + 5 2**32 1 33333333 100000003 + 7 2**24-1 0 249249 (infinite) + 9 2**24-1 0 1c71c7 (infinite) + 11 2**20-1 0 1745d (infinite) + 13 2**24-1 0 13b13b (infinite) + 15 2**32 1 11111111 10000000d + 17 2**32 1 f0f0f0f 10000000f + + If r is 1, then b = a+r-1 = a. This simplifies the computation + of (ax+b), since you can compute (x+1)(a) instead. If r is 0, + then b = 0 is ok to use which simplifies (ax+b). + + The bit patterns for 55555555, 33333333, and 11111111 are obviously + very regular. The bit patterns for the other values of a above are: + + y (hex) (binary) + + 7 249249 001001001001001001001001 << regular >> + 9 1c71c7 000111000111000111000111 << regular >> + 11 1745d 000000010111010001011101 << irregular >> + 13 13b13b 000100111011000100111011 << irregular >> + + The bit patterns for (a) corresponding to (y) of 11 and 13 may be + too irregular to warrant using this method. + + When z is a power of 2 minus 1, then the division by z is slightly + more complicated, involving an iterative solution. + + The code presented here solves division by 1 through 17, except for + 11 and 13. There are algorithms for both signed and unsigned + quantities given. + + TIMINGS (cycles) + + divisor positive negative unsigned + + . 1 2 2 2 + . 2 4 4 2 + . 3 19 21 19 + . 4 4 4 2 + . 5 18 22 19 + . 6 19 22 19 + . 8 4 4 2 + . 10 18 19 17 + . 12 18 20 18 + . 15 16 18 16 + . 16 4 4 2 + . 17 16 18 16 + + Now, the algorithm for 7, 9, and 14 is an iterative one. That is, + a loop body is executed until the tentative quotient is 0. The + number of times the loop body is executed varies depending on the + dividend, but is never more than two times. If the dividend is + less than the divisor, then the loop body is not executed at all. + Each iteration adds 4 cycles to the timings. + + divisor positive negative unsigned + + . 7 19+4n 20+4n 20+4n n = number of iterations + . 9 21+4n 22+4n 21+4n + . 14 21+4n 22+4n 20+4n + + To give an idea of how the number of iterations varies, here is a + table of dividend versus number of iterations when dividing by 7. + + smallest largest required + dividend dividend iterations + + . 0 6 0 + . 7 0x6ffffff 1 + 0x1000006 0xffffffff 2 + + There is some overlap in the range of numbers requiring 1 and 2 + iterations. */ + +RDEFINE(t2,r1) +RDEFINE(x2,arg0) /* r26 */ +RDEFINE(t1,arg1) /* r25 */ +RDEFINE(x1,ret1) /* r29 */ + + SUBSPA_MILLI_DIV + ATTR_MILLI + + .proc + .callinfo millicode + .entry +/* NONE of these routines require a stack frame + ALL of these routines are unwindable from millicode */ + +GSYM($$divide_by_constant) + .export $$divide_by_constant,millicode +/* Provides a "nice" label for the code covered by the unwind descriptor + for things like gprof. */ + +/* DIVISION BY 2 (shift by 1) */ +GSYM($$divI_2) + .export $$divI_2,millicode + comclr,>= arg0,0,0 + addi 1,arg0,arg0 + MILLIRET + extrs arg0,30,31,ret1 + + +/* DIVISION BY 4 (shift by 2) */ +GSYM($$divI_4) + .export $$divI_4,millicode + comclr,>= arg0,0,0 + addi 3,arg0,arg0 + MILLIRET + extrs arg0,29,30,ret1 + + +/* DIVISION BY 8 (shift by 3) */ +GSYM($$divI_8) + .export $$divI_8,millicode + comclr,>= arg0,0,0 + addi 7,arg0,arg0 + MILLIRET + extrs arg0,28,29,ret1 + +/* DIVISION BY 16 (shift by 4) */ +GSYM($$divI_16) + .export $$divI_16,millicode + comclr,>= arg0,0,0 + addi 15,arg0,arg0 + MILLIRET + extrs arg0,27,28,ret1 + +/**************************************************************************** +* +* DIVISION BY DIVISORS OF FFFFFFFF, and powers of 2 times these +* +* includes 3,5,15,17 and also 6,10,12 +* +****************************************************************************/ + +/* DIVISION BY 3 (use z = 2**32; a = 55555555) */ + +GSYM($$divI_3) + .export $$divI_3,millicode + comb,<,N x2,0,LREF(neg3) + + addi 1,x2,x2 /* this cannot overflow */ + extru x2,1,2,x1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(pos) + addc x1,0,x1 + +LSYM(neg3) + subi 1,x2,x2 /* this cannot overflow */ + extru x2,1,2,x1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(neg) + addc x1,0,x1 + +GSYM($$divU_3) + .export $$divU_3,millicode + addi 1,x2,x2 /* this CAN overflow */ + addc 0,0,x1 + shd x1,x2,30,t1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(pos) + addc x1,t1,x1 + +/* DIVISION BY 5 (use z = 2**32; a = 33333333) */ + +GSYM($$divI_5) + .export $$divI_5,millicode + comb,<,N x2,0,LREF(neg5) + + addi 3,x2,t1 /* this cannot overflow */ + sh1add x2,t1,x2 /* multiply by 3 to get started */ + b LREF(pos) + addc 0,0,x1 + +LSYM(neg5) + sub 0,x2,x2 /* negate x2 */ + addi 1,x2,x2 /* this cannot overflow */ + shd 0,x2,31,x1 /* get top bit (can be 1) */ + sh1add x2,x2,x2 /* multiply by 3 to get started */ + b LREF(neg) + addc x1,0,x1 + +GSYM($$divU_5) + .export $$divU_5,millicode + addi 1,x2,x2 /* this CAN overflow */ + addc 0,0,x1 + shd x1,x2,31,t1 /* multiply by 3 to get started */ + sh1add x2,x2,x2 + b LREF(pos) + addc t1,x1,x1 + +/* DIVISION BY 6 (shift to divide by 2 then divide by 3) */ +GSYM($$divI_6) + .export $$divI_6,millicode + comb,<,N x2,0,LREF(neg6) + extru x2,30,31,x2 /* divide by 2 */ + addi 5,x2,t1 /* compute 5*(x2+1) = 5*x2+5 */ + sh2add x2,t1,x2 /* multiply by 5 to get started */ + b LREF(pos) + addc 0,0,x1 + +LSYM(neg6) + subi 2,x2,x2 /* negate, divide by 2, and add 1 */ + /* negation and adding 1 are done */ + /* at the same time by the SUBI */ + extru x2,30,31,x2 + shd 0,x2,30,x1 + sh2add x2,x2,x2 /* multiply by 5 to get started */ + b LREF(neg) + addc x1,0,x1 + +GSYM($$divU_6) + .export $$divU_6,millicode + extru x2,30,31,x2 /* divide by 2 */ + addi 1,x2,x2 /* cannot carry */ + shd 0,x2,30,x1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(pos) + addc x1,0,x1 + +/* DIVISION BY 10 (shift to divide by 2 then divide by 5) */ +GSYM($$divU_10) + .export $$divU_10,millicode + extru x2,30,31,x2 /* divide by 2 */ + addi 3,x2,t1 /* compute 3*(x2+1) = (3*x2)+3 */ + sh1add x2,t1,x2 /* multiply by 3 to get started */ + addc 0,0,x1 +LSYM(pos) + shd x1,x2,28,t1 /* multiply by 0x11 */ + shd x2,0,28,t2 + add x2,t2,x2 + addc x1,t1,x1 +LSYM(pos_for_17) + shd x1,x2,24,t1 /* multiply by 0x101 */ + shd x2,0,24,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,16,t1 /* multiply by 0x10001 */ + shd x2,0,16,t2 + add x2,t2,x2 + MILLIRET + addc x1,t1,x1 + +GSYM($$divI_10) + .export $$divI_10,millicode + comb,< x2,0,LREF(neg10) + copy 0,x1 + extru x2,30,31,x2 /* divide by 2 */ + addib,TR 1,x2,LREF(pos) /* add 1 (cannot overflow) */ + sh1add x2,x2,x2 /* multiply by 3 to get started */ + +LSYM(neg10) + subi 2,x2,x2 /* negate, divide by 2, and add 1 */ + /* negation and adding 1 are done */ + /* at the same time by the SUBI */ + extru x2,30,31,x2 + sh1add x2,x2,x2 /* multiply by 3 to get started */ +LSYM(neg) + shd x1,x2,28,t1 /* multiply by 0x11 */ + shd x2,0,28,t2 + add x2,t2,x2 + addc x1,t1,x1 +LSYM(neg_for_17) + shd x1,x2,24,t1 /* multiply by 0x101 */ + shd x2,0,24,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,16,t1 /* multiply by 0x10001 */ + shd x2,0,16,t2 + add x2,t2,x2 + addc x1,t1,x1 + MILLIRET + sub 0,x1,x1 + +/* DIVISION BY 12 (shift to divide by 4 then divide by 3) */ +GSYM($$divI_12) + .export $$divI_12,millicode + comb,< x2,0,LREF(neg12) + copy 0,x1 + extru x2,29,30,x2 /* divide by 4 */ + addib,tr 1,x2,LREF(pos) /* compute 5*(x2+1) = 5*x2+5 */ + sh2add x2,x2,x2 /* multiply by 5 to get started */ + +LSYM(neg12) + subi 4,x2,x2 /* negate, divide by 4, and add 1 */ + /* negation and adding 1 are done */ + /* at the same time by the SUBI */ + extru x2,29,30,x2 + b LREF(neg) + sh2add x2,x2,x2 /* multiply by 5 to get started */ + +GSYM($$divU_12) + .export $$divU_12,millicode + extru x2,29,30,x2 /* divide by 4 */ + addi 5,x2,t1 /* cannot carry */ + sh2add x2,t1,x2 /* multiply by 5 to get started */ + b LREF(pos) + addc 0,0,x1 + +/* DIVISION BY 15 (use z = 2**32; a = 11111111) */ +GSYM($$divI_15) + .export $$divI_15,millicode + comb,< x2,0,LREF(neg15) + copy 0,x1 + addib,tr 1,x2,LREF(pos)+4 + shd x1,x2,28,t1 + +LSYM(neg15) + b LREF(neg) + subi 1,x2,x2 + +GSYM($$divU_15) + .export $$divU_15,millicode + addi 1,x2,x2 /* this CAN overflow */ + b LREF(pos) + addc 0,0,x1 + +/* DIVISION BY 17 (use z = 2**32; a = f0f0f0f) */ +GSYM($$divI_17) + .export $$divI_17,millicode + comb,<,n x2,0,LREF(neg17) + addi 1,x2,x2 /* this cannot overflow */ + shd 0,x2,28,t1 /* multiply by 0xf to get started */ + shd x2,0,28,t2 + sub t2,x2,x2 + b LREF(pos_for_17) + subb t1,0,x1 + +LSYM(neg17) + subi 1,x2,x2 /* this cannot overflow */ + shd 0,x2,28,t1 /* multiply by 0xf to get started */ + shd x2,0,28,t2 + sub t2,x2,x2 + b LREF(neg_for_17) + subb t1,0,x1 + +GSYM($$divU_17) + .export $$divU_17,millicode + addi 1,x2,x2 /* this CAN overflow */ + addc 0,0,x1 + shd x1,x2,28,t1 /* multiply by 0xf to get started */ +LSYM(u17) + shd x2,0,28,t2 + sub t2,x2,x2 + b LREF(pos_for_17) + subb t1,x1,x1 + + +/* DIVISION BY DIVISORS OF FFFFFF, and powers of 2 times these + includes 7,9 and also 14 + + + z = 2**24-1 + r = z mod x = 0 + + so choose b = 0 + + Also, in order to divide by z = 2**24-1, we approximate by dividing + by (z+1) = 2**24 (which is easy), and then correcting. + + (ax) = (z+1)q' + r + . = zq' + (q'+r) + + So to compute (ax)/z, compute q' = (ax)/(z+1) and r = (ax) mod (z+1) + Then the true remainder of (ax)/z is (q'+r). Repeat the process + with this new remainder, adding the tentative quotients together, + until a tentative quotient is 0 (and then we are done). There is + one last correction to be done. It is possible that (q'+r) = z. + If so, then (q'+r)/(z+1) = 0 and it looks like we are done. But, + in fact, we need to add 1 more to the quotient. Now, it turns + out that this happens if and only if the original value x is + an exact multiple of y. So, to avoid a three instruction test at + the end, instead use 1 instruction to add 1 to x at the beginning. */ + +/* DIVISION BY 7 (use z = 2**24-1; a = 249249) */ +GSYM($$divI_7) + .export $$divI_7,millicode + comb,<,n x2,0,LREF(neg7) +LSYM(7) + addi 1,x2,x2 /* cannot overflow */ + shd 0,x2,29,x1 + sh3add x2,x2,x2 + addc x1,0,x1 +LSYM(pos7) + shd x1,x2,26,t1 + shd x2,0,26,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,20,t1 + shd x2,0,20,t2 + add x2,t2,x2 + addc x1,t1,t1 + + /* computed <t1,x2>. Now divide it by (2**24 - 1) */ + + copy 0,x1 + shd,= t1,x2,24,t1 /* tentative quotient */ +LSYM(1) + addb,tr t1,x1,LREF(2) /* add to previous quotient */ + extru x2,31,24,x2 /* new remainder (unadjusted) */ + + MILLIRETN + +LSYM(2) + addb,tr t1,x2,LREF(1) /* adjust remainder */ + extru,= x2,7,8,t1 /* new quotient */ + +LSYM(neg7) + subi 1,x2,x2 /* negate x2 and add 1 */ +LSYM(8) + shd 0,x2,29,x1 + sh3add x2,x2,x2 + addc x1,0,x1 + +LSYM(neg7_shift) + shd x1,x2,26,t1 + shd x2,0,26,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,20,t1 + shd x2,0,20,t2 + add x2,t2,x2 + addc x1,t1,t1 + + /* computed <t1,x2>. Now divide it by (2**24 - 1) */ + + copy 0,x1 + shd,= t1,x2,24,t1 /* tentative quotient */ +LSYM(3) + addb,tr t1,x1,LREF(4) /* add to previous quotient */ + extru x2,31,24,x2 /* new remainder (unadjusted) */ + + MILLIRET + sub 0,x1,x1 /* negate result */ + +LSYM(4) + addb,tr t1,x2,LREF(3) /* adjust remainder */ + extru,= x2,7,8,t1 /* new quotient */ + +GSYM($$divU_7) + .export $$divU_7,millicode + addi 1,x2,x2 /* can carry */ + addc 0,0,x1 + shd x1,x2,29,t1 + sh3add x2,x2,x2 + b LREF(pos7) + addc t1,x1,x1 + +/* DIVISION BY 9 (use z = 2**24-1; a = 1c71c7) */ +GSYM($$divI_9) + .export $$divI_9,millicode + comb,<,n x2,0,LREF(neg9) + addi 1,x2,x2 /* cannot overflow */ + shd 0,x2,29,t1 + shd x2,0,29,t2 + sub t2,x2,x2 + b LREF(pos7) + subb t1,0,x1 + +LSYM(neg9) + subi 1,x2,x2 /* negate and add 1 */ + shd 0,x2,29,t1 + shd x2,0,29,t2 + sub t2,x2,x2 + b LREF(neg7_shift) + subb t1,0,x1 + +GSYM($$divU_9) + .export $$divU_9,millicode + addi 1,x2,x2 /* can carry */ + addc 0,0,x1 + shd x1,x2,29,t1 + shd x2,0,29,t2 + sub t2,x2,x2 + b LREF(pos7) + subb t1,x1,x1 + +/* DIVISION BY 14 (shift to divide by 2 then divide by 7) */ +GSYM($$divI_14) + .export $$divI_14,millicode + comb,<,n x2,0,LREF(neg14) +GSYM($$divU_14) + .export $$divU_14,millicode + b LREF(7) /* go to 7 case */ + extru x2,30,31,x2 /* divide by 2 */ + +LSYM(neg14) + subi 2,x2,x2 /* negate (and add 2) */ + b LREF(8) + extru x2,30,31,x2 /* divide by 2 */ + .exit + .procend + .end +#endif diff --git a/arch/parisc/lib/milli/dyncall.S b/arch/parisc/lib/milli/dyncall.S new file mode 100644 index 00000000000..27f9ca558d0 --- /dev/null +++ b/arch/parisc/lib/milli/dyncall.S @@ -0,0 +1,32 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_dyncall + SUBSPA_MILLI + ATTR_DATA +GSYM($$dyncall) + .export $$dyncall,millicode + .proc + .callinfo millicode + .entry + bb,>=,n %r22,30,LREF(1) ; branch if not plabel address + depi 0,31,2,%r22 ; clear the two least significant bits + ldw 4(%r22),%r19 ; load new LTP value + ldw 0(%r22),%r22 ; load address of target +LSYM(1) + bv %r0(%r22) ; branch to the real target + stw %r2,-24(%r30) ; save return address into frame marker + .exit + .procend +#endif diff --git a/arch/parisc/lib/milli/milli.S b/arch/parisc/lib/milli/milli.S new file mode 100644 index 00000000000..47c6cde712e --- /dev/null +++ b/arch/parisc/lib/milli/milli.S @@ -0,0 +1,2071 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#ifdef CONFIG_64BIT + .level 2.0w +#endif + +/* Hardware General Registers. */ +r0: .reg %r0 +r1: .reg %r1 +r2: .reg %r2 +r3: .reg %r3 +r4: .reg %r4 +r5: .reg %r5 +r6: .reg %r6 +r7: .reg %r7 +r8: .reg %r8 +r9: .reg %r9 +r10: .reg %r10 +r11: .reg %r11 +r12: .reg %r12 +r13: .reg %r13 +r14: .reg %r14 +r15: .reg %r15 +r16: .reg %r16 +r17: .reg %r17 +r18: .reg %r18 +r19: .reg %r19 +r20: .reg %r20 +r21: .reg %r21 +r22: .reg %r22 +r23: .reg %r23 +r24: .reg %r24 +r25: .reg %r25 +r26: .reg %r26 +r27: .reg %r27 +r28: .reg %r28 +r29: .reg %r29 +r30: .reg %r30 +r31: .reg %r31 + +/* Hardware Space Registers. */ +sr0: .reg %sr0 +sr1: .reg %sr1 +sr2: .reg %sr2 +sr3: .reg %sr3 +sr4: .reg %sr4 +sr5: .reg %sr5 +sr6: .reg %sr6 +sr7: .reg %sr7 + +/* Hardware Floating Point Registers. */ +fr0: .reg %fr0 +fr1: .reg %fr1 +fr2: .reg %fr2 +fr3: .reg %fr3 +fr4: .reg %fr4 +fr5: .reg %fr5 +fr6: .reg %fr6 +fr7: .reg %fr7 +fr8: .reg %fr8 +fr9: .reg %fr9 +fr10: .reg %fr10 +fr11: .reg %fr11 +fr12: .reg %fr12 +fr13: .reg %fr13 +fr14: .reg %fr14 +fr15: .reg %fr15 + +/* Hardware Control Registers. */ +cr11: .reg %cr11 +sar: .reg %cr11 /* Shift Amount Register */ + +/* Software Architecture General Registers. */ +rp: .reg r2 /* return pointer */ +#ifdef CONFIG_64BIT +mrp: .reg r2 /* millicode return pointer */ +#else +mrp: .reg r31 /* millicode return pointer */ +#endif +ret0: .reg r28 /* return value */ +ret1: .reg r29 /* return value (high part of double) */ +sp: .reg r30 /* stack pointer */ +dp: .reg r27 /* data pointer */ +arg0: .reg r26 /* argument */ +arg1: .reg r25 /* argument or high part of double argument */ +arg2: .reg r24 /* argument */ +arg3: .reg r23 /* argument or high part of double argument */ + +/* Software Architecture Space Registers. */ +/* sr0 ; return link from BLE */ +sret: .reg sr1 /* return value */ +sarg: .reg sr1 /* argument */ +/* sr4 ; PC SPACE tracker */ +/* sr5 ; process private data */ + +/* Frame Offsets (millicode convention!) Used when calling other + millicode routines. Stack unwinding is dependent upon these + definitions. */ +r31_slot: .equ -20 /* "current RP" slot */ +sr0_slot: .equ -16 /* "static link" slot */ +#if defined(CONFIG_64BIT) +mrp_slot: .equ -16 /* "current RP" slot */ +psp_slot: .equ -8 /* "previous SP" slot */ +#else +mrp_slot: .equ -20 /* "current RP" slot (replacing "r31_slot") */ +#endif + + +#define DEFINE(name,value)name: .EQU value +#define RDEFINE(name,value)name: .REG value +#ifdef milliext +#define MILLI_BE(lbl) BE lbl(sr7,r0) +#define MILLI_BEN(lbl) BE,n lbl(sr7,r0) +#define MILLI_BLE(lbl) BLE lbl(sr7,r0) +#define MILLI_BLEN(lbl) BLE,n lbl(sr7,r0) +#define MILLIRETN BE,n 0(sr0,mrp) +#define MILLIRET BE 0(sr0,mrp) +#define MILLI_RETN BE,n 0(sr0,mrp) +#define MILLI_RET BE 0(sr0,mrp) +#else +#define MILLI_BE(lbl) B lbl +#define MILLI_BEN(lbl) B,n lbl +#define MILLI_BLE(lbl) BL lbl,mrp +#define MILLI_BLEN(lbl) BL,n lbl,mrp +#define MILLIRETN BV,n 0(mrp) +#define MILLIRET BV 0(mrp) +#define MILLI_RETN BV,n 0(mrp) +#define MILLI_RET BV 0(mrp) +#endif + +#define CAT(a,b) a##b + +#define SUBSPA_MILLI .section .text +#define SUBSPA_MILLI_DIV .section .text.div,"ax",@progbits! .align 16 +#define SUBSPA_MILLI_MUL .section .text.mul,"ax",@progbits! .align 16 +#define ATTR_MILLI +#define SUBSPA_DATA .section .data +#define ATTR_DATA +#define GLOBAL $global$ +#define GSYM(sym) !sym: +#define LSYM(sym) !CAT(.L,sym:) +#define LREF(sym) CAT(.L,sym) + +#ifdef L_dyncall + SUBSPA_MILLI + ATTR_DATA +GSYM($$dyncall) + .export $$dyncall,millicode + .proc + .callinfo millicode + .entry + bb,>=,n %r22,30,LREF(1) ; branch if not plabel address + depi 0,31,2,%r22 ; clear the two least significant bits + ldw 4(%r22),%r19 ; load new LTP value + ldw 0(%r22),%r22 ; load address of target +LSYM(1) + bv %r0(%r22) ; branch to the real target + stw %r2,-24(%r30) ; save return address into frame marker + .exit + .procend +#endif + +#ifdef L_divI +/* ROUTINES: $$divI, $$divoI + + Single precision divide for signed binary integers. + + The quotient is truncated towards zero. + The sign of the quotient is the XOR of the signs of the dividend and + divisor. + Divide by zero is trapped. + Divide of -2**31 by -1 is trapped for $$divoI but not for $$divI. + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = quotient + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: + . divisor is zero (traps with ADDIT,= 0,25,0) + . dividend==-2**31 and divisor==-1 and routine is $$divoI + . (traps with ADDO 26,25,0) + . Changes memory at the following places: + . NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Branchs to other millicode routines using BE + . $$div_# for # being 2,3,4,5,6,7,8,9,10,12,14,15 + . + . For selected divisors, calls a divide by constant routine written by + . Karl Pettis. Eligible divisors are 1..15 excluding 11 and 13. + . + . The only overflow case is -2**31 divided by -1. + . Both routines return -2**31 but only $$divoI traps. */ + +RDEFINE(temp,r1) +RDEFINE(retreg,ret1) /* r29 */ +RDEFINE(temp1,arg0) + SUBSPA_MILLI_DIV + ATTR_MILLI + .import $$divI_2,millicode + .import $$divI_3,millicode + .import $$divI_4,millicode + .import $$divI_5,millicode + .import $$divI_6,millicode + .import $$divI_7,millicode + .import $$divI_8,millicode + .import $$divI_9,millicode + .import $$divI_10,millicode + .import $$divI_12,millicode + .import $$divI_14,millicode + .import $$divI_15,millicode + .export $$divI,millicode + .export $$divoI,millicode + .proc + .callinfo millicode + .entry +GSYM($$divoI) + comib,=,n -1,arg1,LREF(negative1) /* when divisor == -1 */ +GSYM($$divI) + ldo -1(arg1),temp /* is there at most one bit set ? */ + and,<> arg1,temp,r0 /* if not, don't use power of 2 divide */ + addi,> 0,arg1,r0 /* if divisor > 0, use power of 2 divide */ + b,n LREF(neg_denom) +LSYM(pow2) + addi,>= 0,arg0,retreg /* if numerator is negative, add the */ + add arg0,temp,retreg /* (denominaotr -1) to correct for shifts */ + extru,= arg1,15,16,temp /* test denominator with 0xffff0000 */ + extrs retreg,15,16,retreg /* retreg = retreg >> 16 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 16) */ + ldi 0xcc,temp1 /* setup 0xcc in temp1 */ + extru,= arg1,23,8,temp /* test denominator with 0xff00 */ + extrs retreg,23,24,retreg /* retreg = retreg >> 8 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 8) */ + ldi 0xaa,temp /* setup 0xaa in temp */ + extru,= arg1,27,4,r0 /* test denominator with 0xf0 */ + extrs retreg,27,28,retreg /* retreg = retreg >> 4 */ + and,= arg1,temp1,r0 /* test denominator with 0xcc */ + extrs retreg,29,30,retreg /* retreg = retreg >> 2 */ + and,= arg1,temp,r0 /* test denominator with 0xaa */ + extrs retreg,30,31,retreg /* retreg = retreg >> 1 */ + MILLIRETN +LSYM(neg_denom) + addi,< 0,arg1,r0 /* if arg1 >= 0, it's not power of 2 */ + b,n LREF(regular_seq) + sub r0,arg1,temp /* make denominator positive */ + comb,=,n arg1,temp,LREF(regular_seq) /* test against 0x80000000 and 0 */ + ldo -1(temp),retreg /* is there at most one bit set ? */ + and,= temp,retreg,r0 /* if so, the denominator is power of 2 */ + b,n LREF(regular_seq) + sub r0,arg0,retreg /* negate numerator */ + comb,=,n arg0,retreg,LREF(regular_seq) /* test against 0x80000000 */ + copy retreg,arg0 /* set up arg0, arg1 and temp */ + copy temp,arg1 /* before branching to pow2 */ + b LREF(pow2) + ldo -1(arg1),temp +LSYM(regular_seq) + comib,>>=,n 15,arg1,LREF(small_divisor) + add,>= 0,arg0,retreg /* move dividend, if retreg < 0, */ +LSYM(normal) + subi 0,retreg,retreg /* make it positive */ + sub 0,arg1,temp /* clear carry, */ + /* negate the divisor */ + ds 0,temp,0 /* set V-bit to the comple- */ + /* ment of the divisor sign */ + add retreg,retreg,retreg /* shift msb bit into carry */ + ds r0,arg1,temp /* 1st divide step, if no carry */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 2nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 3rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 4th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 5th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 6th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 7th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 8th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 9th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 10th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 11th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 12th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 13th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 14th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 15th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 16th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 17th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 18th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 19th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 20th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 21st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 22nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 23rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 24th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 25th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 26th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 27th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 28th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 29th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 30th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 31st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 32nd divide step, */ + addc retreg,retreg,retreg /* shift last retreg bit into retreg */ + xor,>= arg0,arg1,0 /* get correct sign of quotient */ + sub 0,retreg,retreg /* based on operand signs */ + MILLIRETN + nop + +LSYM(small_divisor) + +#if defined(CONFIG_64BIT) +/* Clear the upper 32 bits of the arg1 register. We are working with */ +/* small divisors (and 32-bit integers) We must not be mislead */ +/* by "1" bits left in the upper 32 bits. */ + depd %r0,31,32,%r25 +#endif + blr,n arg1,r0 + nop +/* table for divisor == 0,1, ... ,15 */ + addit,= 0,arg1,r0 /* trap if divisor == 0 */ + nop + MILLIRET /* divisor == 1 */ + copy arg0,retreg + MILLI_BEN($$divI_2) /* divisor == 2 */ + nop + MILLI_BEN($$divI_3) /* divisor == 3 */ + nop + MILLI_BEN($$divI_4) /* divisor == 4 */ + nop + MILLI_BEN($$divI_5) /* divisor == 5 */ + nop + MILLI_BEN($$divI_6) /* divisor == 6 */ + nop + MILLI_BEN($$divI_7) /* divisor == 7 */ + nop + MILLI_BEN($$divI_8) /* divisor == 8 */ + nop + MILLI_BEN($$divI_9) /* divisor == 9 */ + nop + MILLI_BEN($$divI_10) /* divisor == 10 */ + nop + b LREF(normal) /* divisor == 11 */ + add,>= 0,arg0,retreg + MILLI_BEN($$divI_12) /* divisor == 12 */ + nop + b LREF(normal) /* divisor == 13 */ + add,>= 0,arg0,retreg + MILLI_BEN($$divI_14) /* divisor == 14 */ + nop + MILLI_BEN($$divI_15) /* divisor == 15 */ + nop + +LSYM(negative1) + sub 0,arg0,retreg /* result is negation of dividend */ + MILLIRET + addo arg0,arg1,r0 /* trap iff dividend==0x80000000 && divisor==-1 */ + .exit + .procend + .end +#endif + +#ifdef L_divU +/* ROUTINE: $$divU + . + . Single precision divide for unsigned integers. + . + . Quotient is truncated towards zero. + . Traps on divide by zero. + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = quotient + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: + . divisor is zero + . Changes memory at the following places: + . NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Does not create a stack frame. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Branchs to other millicode routines using BE: + . $$divU_# for 3,5,6,7,9,10,12,14,15 + . + . For selected small divisors calls the special divide by constant + . routines written by Karl Pettis. These are: 3,5,6,7,9,10,12,14,15. */ + +RDEFINE(temp,r1) +RDEFINE(retreg,ret1) /* r29 */ +RDEFINE(temp1,arg0) + SUBSPA_MILLI_DIV + ATTR_MILLI + .export $$divU,millicode + .import $$divU_3,millicode + .import $$divU_5,millicode + .import $$divU_6,millicode + .import $$divU_7,millicode + .import $$divU_9,millicode + .import $$divU_10,millicode + .import $$divU_12,millicode + .import $$divU_14,millicode + .import $$divU_15,millicode + .proc + .callinfo millicode + .entry +GSYM($$divU) +/* The subtract is not nullified since it does no harm and can be used + by the two cases that branch back to "normal". */ + ldo -1(arg1),temp /* is there at most one bit set ? */ + and,= arg1,temp,r0 /* if so, denominator is power of 2 */ + b LREF(regular_seq) + addit,= 0,arg1,0 /* trap for zero dvr */ + copy arg0,retreg + extru,= arg1,15,16,temp /* test denominator with 0xffff0000 */ + extru retreg,15,16,retreg /* retreg = retreg >> 16 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 16) */ + ldi 0xcc,temp1 /* setup 0xcc in temp1 */ + extru,= arg1,23,8,temp /* test denominator with 0xff00 */ + extru retreg,23,24,retreg /* retreg = retreg >> 8 */ + or arg1,temp,arg1 /* arg1 = arg1 | (arg1 >> 8) */ + ldi 0xaa,temp /* setup 0xaa in temp */ + extru,= arg1,27,4,r0 /* test denominator with 0xf0 */ + extru retreg,27,28,retreg /* retreg = retreg >> 4 */ + and,= arg1,temp1,r0 /* test denominator with 0xcc */ + extru retreg,29,30,retreg /* retreg = retreg >> 2 */ + and,= arg1,temp,r0 /* test denominator with 0xaa */ + extru retreg,30,31,retreg /* retreg = retreg >> 1 */ + MILLIRETN + nop +LSYM(regular_seq) + comib,>= 15,arg1,LREF(special_divisor) + subi 0,arg1,temp /* clear carry, negate the divisor */ + ds r0,temp,r0 /* set V-bit to 1 */ +LSYM(normal) + add arg0,arg0,retreg /* shift msb bit into carry */ + ds r0,arg1,temp /* 1st divide step, if no carry */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 2nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 3rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 4th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 5th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 6th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 7th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 8th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 9th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 10th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 11th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 12th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 13th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 14th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 15th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 16th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 17th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 18th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 19th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 20th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 21st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 22nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 23rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 24th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 25th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 26th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 27th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 28th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 29th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 30th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 31st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds temp,arg1,temp /* 32nd divide step, */ + MILLIRET + addc retreg,retreg,retreg /* shift last retreg bit into retreg */ + +/* Handle the cases where divisor is a small constant or has high bit on. */ +LSYM(special_divisor) +/* blr arg1,r0 */ +/* comib,>,n 0,arg1,LREF(big_divisor) ; nullify previous instruction */ + +/* Pratap 8/13/90. The 815 Stirling chip set has a bug that prevents us from + generating such a blr, comib sequence. A problem in nullification. So I + rewrote this code. */ + +#if defined(CONFIG_64BIT) +/* Clear the upper 32 bits of the arg1 register. We are working with + small divisors (and 32-bit unsigned integers) We must not be mislead + by "1" bits left in the upper 32 bits. */ + depd %r0,31,32,%r25 +#endif + comib,> 0,arg1,LREF(big_divisor) + nop + blr arg1,r0 + nop + +LSYM(zero_divisor) /* this label is here to provide external visibility */ + addit,= 0,arg1,0 /* trap for zero dvr */ + nop + MILLIRET /* divisor == 1 */ + copy arg0,retreg + MILLIRET /* divisor == 2 */ + extru arg0,30,31,retreg + MILLI_BEN($$divU_3) /* divisor == 3 */ + nop + MILLIRET /* divisor == 4 */ + extru arg0,29,30,retreg + MILLI_BEN($$divU_5) /* divisor == 5 */ + nop + MILLI_BEN($$divU_6) /* divisor == 6 */ + nop + MILLI_BEN($$divU_7) /* divisor == 7 */ + nop + MILLIRET /* divisor == 8 */ + extru arg0,28,29,retreg + MILLI_BEN($$divU_9) /* divisor == 9 */ + nop + MILLI_BEN($$divU_10) /* divisor == 10 */ + nop + b LREF(normal) /* divisor == 11 */ + ds r0,temp,r0 /* set V-bit to 1 */ + MILLI_BEN($$divU_12) /* divisor == 12 */ + nop + b LREF(normal) /* divisor == 13 */ + ds r0,temp,r0 /* set V-bit to 1 */ + MILLI_BEN($$divU_14) /* divisor == 14 */ + nop + MILLI_BEN($$divU_15) /* divisor == 15 */ + nop + +/* Handle the case where the high bit is on in the divisor. + Compute: if( dividend>=divisor) quotient=1; else quotient=0; + Note: dividend>==divisor iff dividend-divisor does not borrow + and not borrow iff carry. */ +LSYM(big_divisor) + sub arg0,arg1,r0 + MILLIRET + addc r0,r0,retreg + .exit + .procend + .end +#endif + +#ifdef L_remI +/* ROUTINE: $$remI + + DESCRIPTION: + . $$remI returns the remainder of the division of two signed 32-bit + . integers. The sign of the remainder is the same as the sign of + . the dividend. + + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = destroyed + . arg1 = destroyed + . ret1 = remainder + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: DIVIDE BY ZERO + . Changes memory at the following places: NONE + + PERMISSIBLE CONTEXT: + . Unwindable + . Does not create a stack frame + . Is usable for internal or external microcode + + DISCUSSION: + . Calls other millicode routines via mrp: NONE + . Calls other millicode routines: NONE */ + +RDEFINE(tmp,r1) +RDEFINE(retreg,ret1) + + SUBSPA_MILLI + ATTR_MILLI + .proc + .callinfo millicode + .entry +GSYM($$remI) +GSYM($$remoI) + .export $$remI,MILLICODE + .export $$remoI,MILLICODE + ldo -1(arg1),tmp /* is there at most one bit set ? */ + and,<> arg1,tmp,r0 /* if not, don't use power of 2 */ + addi,> 0,arg1,r0 /* if denominator > 0, use power */ + /* of 2 */ + b,n LREF(neg_denom) +LSYM(pow2) + comb,>,n 0,arg0,LREF(neg_num) /* is numerator < 0 ? */ + and arg0,tmp,retreg /* get the result */ + MILLIRETN +LSYM(neg_num) + subi 0,arg0,arg0 /* negate numerator */ + and arg0,tmp,retreg /* get the result */ + subi 0,retreg,retreg /* negate result */ + MILLIRETN +LSYM(neg_denom) + addi,< 0,arg1,r0 /* if arg1 >= 0, it's not power */ + /* of 2 */ + b,n LREF(regular_seq) + sub r0,arg1,tmp /* make denominator positive */ + comb,=,n arg1,tmp,LREF(regular_seq) /* test against 0x80000000 and 0 */ + ldo -1(tmp),retreg /* is there at most one bit set ? */ + and,= tmp,retreg,r0 /* if not, go to regular_seq */ + b,n LREF(regular_seq) + comb,>,n 0,arg0,LREF(neg_num_2) /* if arg0 < 0, negate it */ + and arg0,retreg,retreg + MILLIRETN +LSYM(neg_num_2) + subi 0,arg0,tmp /* test against 0x80000000 */ + and tmp,retreg,retreg + subi 0,retreg,retreg + MILLIRETN +LSYM(regular_seq) + addit,= 0,arg1,0 /* trap if div by zero */ + add,>= 0,arg0,retreg /* move dividend, if retreg < 0, */ + sub 0,retreg,retreg /* make it positive */ + sub 0,arg1, tmp /* clear carry, */ + /* negate the divisor */ + ds 0, tmp,0 /* set V-bit to the comple- */ + /* ment of the divisor sign */ + or 0,0, tmp /* clear tmp */ + add retreg,retreg,retreg /* shift msb bit into carry */ + ds tmp,arg1, tmp /* 1st divide step, if no carry */ + /* out, msb of quotient = 0 */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ +LSYM(t1) + ds tmp,arg1, tmp /* 2nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 3rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 4th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 5th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 6th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 7th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 8th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 9th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 10th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 11th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 12th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 13th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 14th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 15th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 16th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 17th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 18th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 19th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 20th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 21st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 22nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 23rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 24th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 25th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 26th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 27th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 28th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 29th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 30th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 31st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 32nd divide step, */ + addc retreg,retreg,retreg /* shift last bit into retreg */ + movb,>=,n tmp,retreg,LREF(finish) /* branch if pos. tmp */ + add,< arg1,0,0 /* if arg1 > 0, add arg1 */ + add,tr tmp,arg1,retreg /* for correcting remainder tmp */ + sub tmp,arg1,retreg /* else add absolute value arg1 */ +LSYM(finish) + add,>= arg0,0,0 /* set sign of remainder */ + sub 0,retreg,retreg /* to sign of dividend */ + MILLIRET + nop + .exit + .procend +#ifdef milliext + .origin 0x00000200 +#endif + .end +#endif + +#ifdef L_remU +/* ROUTINE: $$remU + . Single precision divide for remainder with unsigned binary integers. + . + . The remainder must be dividend-(dividend/divisor)*divisor. + . Divide by zero is trapped. + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = remainder + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: DIVIDE BY ZERO + . Changes memory at the following places: NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Does not create a stack frame. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Calls other millicode routines using mrp: NONE + . Calls other millicode routines: NONE */ + + +RDEFINE(temp,r1) +RDEFINE(rmndr,ret1) /* r29 */ + SUBSPA_MILLI + ATTR_MILLI + .export $$remU,millicode + .proc + .callinfo millicode + .entry +GSYM($$remU) + ldo -1(arg1),temp /* is there at most one bit set ? */ + and,= arg1,temp,r0 /* if not, don't use power of 2 */ + b LREF(regular_seq) + addit,= 0,arg1,r0 /* trap on div by zero */ + and arg0,temp,rmndr /* get the result for power of 2 */ + MILLIRETN +LSYM(regular_seq) + comib,>=,n 0,arg1,LREF(special_case) + subi 0,arg1,rmndr /* clear carry, negate the divisor */ + ds r0,rmndr,r0 /* set V-bit to 1 */ + add arg0,arg0,temp /* shift msb bit into carry */ + ds r0,arg1,rmndr /* 1st divide step, if no carry */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 2nd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 3rd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 4th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 5th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 6th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 7th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 8th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 9th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 10th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 11th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 12th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 13th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 14th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 15th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 16th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 17th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 18th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 19th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 20th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 21st divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 22nd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 23rd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 24th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 25th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 26th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 27th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 28th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 29th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 30th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 31st divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 32nd divide step, */ + comiclr,<= 0,rmndr,r0 + add rmndr,arg1,rmndr /* correction */ + MILLIRETN + nop + +/* Putting >= on the last DS and deleting COMICLR does not work! */ +LSYM(special_case) + sub,>>= arg0,arg1,rmndr + copy arg0,rmndr + MILLIRETN + nop + .exit + .procend + .end +#endif + +#ifdef L_div_const +/* ROUTINE: $$divI_2 + . $$divI_3 $$divU_3 + . $$divI_4 + . $$divI_5 $$divU_5 + . $$divI_6 $$divU_6 + . $$divI_7 $$divU_7 + . $$divI_8 + . $$divI_9 $$divU_9 + . $$divI_10 $$divU_10 + . + . $$divI_12 $$divU_12 + . + . $$divI_14 $$divU_14 + . $$divI_15 $$divU_15 + . $$divI_16 + . $$divI_17 $$divU_17 + . + . Divide by selected constants for single precision binary integers. + + INPUT REGISTERS: + . arg0 == dividend + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = quotient + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: NONE + . Changes memory at the following places: NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Does not create a stack frame. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Calls other millicode routines using mrp: NONE + . Calls other millicode routines: NONE */ + + +/* TRUNCATED DIVISION BY SMALL INTEGERS + + We are interested in q(x) = floor(x/y), where x >= 0 and y > 0 + (with y fixed). + + Let a = floor(z/y), for some choice of z. Note that z will be + chosen so that division by z is cheap. + + Let r be the remainder(z/y). In other words, r = z - ay. + + Now, our method is to choose a value for b such that + + q'(x) = floor((ax+b)/z) + + is equal to q(x) over as large a range of x as possible. If the + two are equal over a sufficiently large range, and if it is easy to + form the product (ax), and it is easy to divide by z, then we can + perform the division much faster than the general division algorithm. + + So, we want the following to be true: + + . For x in the following range: + . + . ky <= x < (k+1)y + . + . implies that + . + . k <= (ax+b)/z < (k+1) + + We want to determine b such that this is true for all k in the + range {0..K} for some maximum K. + + Since (ax+b) is an increasing function of x, we can take each + bound separately to determine the "best" value for b. + + (ax+b)/z < (k+1) implies + + (a((k+1)y-1)+b < (k+1)z implies + + b < a + (k+1)(z-ay) implies + + b < a + (k+1)r + + This needs to be true for all k in the range {0..K}. In + particular, it is true for k = 0 and this leads to a maximum + acceptable value for b. + + b < a+r or b <= a+r-1 + + Taking the other bound, we have + + k <= (ax+b)/z implies + + k <= (aky+b)/z implies + + k(z-ay) <= b implies + + kr <= b + + Clearly, the largest range for k will be achieved by maximizing b, + when r is not zero. When r is zero, then the simplest choice for b + is 0. When r is not 0, set + + . b = a+r-1 + + Now, by construction, q'(x) = floor((ax+b)/z) = q(x) = floor(x/y) + for all x in the range: + + . 0 <= x < (K+1)y + + We need to determine what K is. Of our two bounds, + + . b < a+(k+1)r is satisfied for all k >= 0, by construction. + + The other bound is + + . kr <= b + + This is always true if r = 0. If r is not 0 (the usual case), then + K = floor((a+r-1)/r), is the maximum value for k. + + Therefore, the formula q'(x) = floor((ax+b)/z) yields the correct + answer for q(x) = floor(x/y) when x is in the range + + (0,(K+1)y-1) K = floor((a+r-1)/r) + + To be most useful, we want (K+1)y-1 = (max x) >= 2**32-1 so that + the formula for q'(x) yields the correct value of q(x) for all x + representable by a single word in HPPA. + + We are also constrained in that computing the product (ax), adding + b, and dividing by z must all be done quickly, otherwise we will be + better off going through the general algorithm using the DS + instruction, which uses approximately 70 cycles. + + For each y, there is a choice of z which satisfies the constraints + for (K+1)y >= 2**32. We may not, however, be able to satisfy the + timing constraints for arbitrary y. It seems that z being equal to + a power of 2 or a power of 2 minus 1 is as good as we can do, since + it minimizes the time to do division by z. We want the choice of z + to also result in a value for (a) that minimizes the computation of + the product (ax). This is best achieved if (a) has a regular bit + pattern (so the multiplication can be done with shifts and adds). + The value of (a) also needs to be less than 2**32 so the product is + always guaranteed to fit in 2 words. + + In actual practice, the following should be done: + + 1) For negative x, you should take the absolute value and remember + . the fact so that the result can be negated. This obviously does + . not apply in the unsigned case. + 2) For even y, you should factor out the power of 2 that divides y + . and divide x by it. You can then proceed by dividing by the + . odd factor of y. + + Here is a table of some odd values of y, and corresponding choices + for z which are "good". + + y z r a (hex) max x (hex) + + 3 2**32 1 55555555 100000001 + 5 2**32 1 33333333 100000003 + 7 2**24-1 0 249249 (infinite) + 9 2**24-1 0 1c71c7 (infinite) + 11 2**20-1 0 1745d (infinite) + 13 2**24-1 0 13b13b (infinite) + 15 2**32 1 11111111 10000000d + 17 2**32 1 f0f0f0f 10000000f + + If r is 1, then b = a+r-1 = a. This simplifies the computation + of (ax+b), since you can compute (x+1)(a) instead. If r is 0, + then b = 0 is ok to use which simplifies (ax+b). + + The bit patterns for 55555555, 33333333, and 11111111 are obviously + very regular. The bit patterns for the other values of a above are: + + y (hex) (binary) + + 7 249249 001001001001001001001001 << regular >> + 9 1c71c7 000111000111000111000111 << regular >> + 11 1745d 000000010111010001011101 << irregular >> + 13 13b13b 000100111011000100111011 << irregular >> + + The bit patterns for (a) corresponding to (y) of 11 and 13 may be + too irregular to warrant using this method. + + When z is a power of 2 minus 1, then the division by z is slightly + more complicated, involving an iterative solution. + + The code presented here solves division by 1 through 17, except for + 11 and 13. There are algorithms for both signed and unsigned + quantities given. + + TIMINGS (cycles) + + divisor positive negative unsigned + + . 1 2 2 2 + . 2 4 4 2 + . 3 19 21 19 + . 4 4 4 2 + . 5 18 22 19 + . 6 19 22 19 + . 8 4 4 2 + . 10 18 19 17 + . 12 18 20 18 + . 15 16 18 16 + . 16 4 4 2 + . 17 16 18 16 + + Now, the algorithm for 7, 9, and 14 is an iterative one. That is, + a loop body is executed until the tentative quotient is 0. The + number of times the loop body is executed varies depending on the + dividend, but is never more than two times. If the dividend is + less than the divisor, then the loop body is not executed at all. + Each iteration adds 4 cycles to the timings. + + divisor positive negative unsigned + + . 7 19+4n 20+4n 20+4n n = number of iterations + . 9 21+4n 22+4n 21+4n + . 14 21+4n 22+4n 20+4n + + To give an idea of how the number of iterations varies, here is a + table of dividend versus number of iterations when dividing by 7. + + smallest largest required + dividend dividend iterations + + . 0 6 0 + . 7 0x6ffffff 1 + 0x1000006 0xffffffff 2 + + There is some overlap in the range of numbers requiring 1 and 2 + iterations. */ + +RDEFINE(t2,r1) +RDEFINE(x2,arg0) /* r26 */ +RDEFINE(t1,arg1) /* r25 */ +RDEFINE(x1,ret1) /* r29 */ + + SUBSPA_MILLI_DIV + ATTR_MILLI + + .proc + .callinfo millicode + .entry +/* NONE of these routines require a stack frame + ALL of these routines are unwindable from millicode */ + +GSYM($$divide_by_constant) + .export $$divide_by_constant,millicode +/* Provides a "nice" label for the code covered by the unwind descriptor + for things like gprof. */ + +/* DIVISION BY 2 (shift by 1) */ +GSYM($$divI_2) + .export $$divI_2,millicode + comclr,>= arg0,0,0 + addi 1,arg0,arg0 + MILLIRET + extrs arg0,30,31,ret1 + + +/* DIVISION BY 4 (shift by 2) */ +GSYM($$divI_4) + .export $$divI_4,millicode + comclr,>= arg0,0,0 + addi 3,arg0,arg0 + MILLIRET + extrs arg0,29,30,ret1 + + +/* DIVISION BY 8 (shift by 3) */ +GSYM($$divI_8) + .export $$divI_8,millicode + comclr,>= arg0,0,0 + addi 7,arg0,arg0 + MILLIRET + extrs arg0,28,29,ret1 + +/* DIVISION BY 16 (shift by 4) */ +GSYM($$divI_16) + .export $$divI_16,millicode + comclr,>= arg0,0,0 + addi 15,arg0,arg0 + MILLIRET + extrs arg0,27,28,ret1 + +/**************************************************************************** +* +* DIVISION BY DIVISORS OF FFFFFFFF, and powers of 2 times these +* +* includes 3,5,15,17 and also 6,10,12 +* +****************************************************************************/ + +/* DIVISION BY 3 (use z = 2**32; a = 55555555) */ + +GSYM($$divI_3) + .export $$divI_3,millicode + comb,<,N x2,0,LREF(neg3) + + addi 1,x2,x2 /* this cannot overflow */ + extru x2,1,2,x1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(pos) + addc x1,0,x1 + +LSYM(neg3) + subi 1,x2,x2 /* this cannot overflow */ + extru x2,1,2,x1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(neg) + addc x1,0,x1 + +GSYM($$divU_3) + .export $$divU_3,millicode + addi 1,x2,x2 /* this CAN overflow */ + addc 0,0,x1 + shd x1,x2,30,t1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(pos) + addc x1,t1,x1 + +/* DIVISION BY 5 (use z = 2**32; a = 33333333) */ + +GSYM($$divI_5) + .export $$divI_5,millicode + comb,<,N x2,0,LREF(neg5) + + addi 3,x2,t1 /* this cannot overflow */ + sh1add x2,t1,x2 /* multiply by 3 to get started */ + b LREF(pos) + addc 0,0,x1 + +LSYM(neg5) + sub 0,x2,x2 /* negate x2 */ + addi 1,x2,x2 /* this cannot overflow */ + shd 0,x2,31,x1 /* get top bit (can be 1) */ + sh1add x2,x2,x2 /* multiply by 3 to get started */ + b LREF(neg) + addc x1,0,x1 + +GSYM($$divU_5) + .export $$divU_5,millicode + addi 1,x2,x2 /* this CAN overflow */ + addc 0,0,x1 + shd x1,x2,31,t1 /* multiply by 3 to get started */ + sh1add x2,x2,x2 + b LREF(pos) + addc t1,x1,x1 + +/* DIVISION BY 6 (shift to divide by 2 then divide by 3) */ +GSYM($$divI_6) + .export $$divI_6,millicode + comb,<,N x2,0,LREF(neg6) + extru x2,30,31,x2 /* divide by 2 */ + addi 5,x2,t1 /* compute 5*(x2+1) = 5*x2+5 */ + sh2add x2,t1,x2 /* multiply by 5 to get started */ + b LREF(pos) + addc 0,0,x1 + +LSYM(neg6) + subi 2,x2,x2 /* negate, divide by 2, and add 1 */ + /* negation and adding 1 are done */ + /* at the same time by the SUBI */ + extru x2,30,31,x2 + shd 0,x2,30,x1 + sh2add x2,x2,x2 /* multiply by 5 to get started */ + b LREF(neg) + addc x1,0,x1 + +GSYM($$divU_6) + .export $$divU_6,millicode + extru x2,30,31,x2 /* divide by 2 */ + addi 1,x2,x2 /* cannot carry */ + shd 0,x2,30,x1 /* multiply by 5 to get started */ + sh2add x2,x2,x2 + b LREF(pos) + addc x1,0,x1 + +/* DIVISION BY 10 (shift to divide by 2 then divide by 5) */ +GSYM($$divU_10) + .export $$divU_10,millicode + extru x2,30,31,x2 /* divide by 2 */ + addi 3,x2,t1 /* compute 3*(x2+1) = (3*x2)+3 */ + sh1add x2,t1,x2 /* multiply by 3 to get started */ + addc 0,0,x1 +LSYM(pos) + shd x1,x2,28,t1 /* multiply by 0x11 */ + shd x2,0,28,t2 + add x2,t2,x2 + addc x1,t1,x1 +LSYM(pos_for_17) + shd x1,x2,24,t1 /* multiply by 0x101 */ + shd x2,0,24,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,16,t1 /* multiply by 0x10001 */ + shd x2,0,16,t2 + add x2,t2,x2 + MILLIRET + addc x1,t1,x1 + +GSYM($$divI_10) + .export $$divI_10,millicode + comb,< x2,0,LREF(neg10) + copy 0,x1 + extru x2,30,31,x2 /* divide by 2 */ + addib,TR 1,x2,LREF(pos) /* add 1 (cannot overflow) */ + sh1add x2,x2,x2 /* multiply by 3 to get started */ + +LSYM(neg10) + subi 2,x2,x2 /* negate, divide by 2, and add 1 */ + /* negation and adding 1 are done */ + /* at the same time by the SUBI */ + extru x2,30,31,x2 + sh1add x2,x2,x2 /* multiply by 3 to get started */ +LSYM(neg) + shd x1,x2,28,t1 /* multiply by 0x11 */ + shd x2,0,28,t2 + add x2,t2,x2 + addc x1,t1,x1 +LSYM(neg_for_17) + shd x1,x2,24,t1 /* multiply by 0x101 */ + shd x2,0,24,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,16,t1 /* multiply by 0x10001 */ + shd x2,0,16,t2 + add x2,t2,x2 + addc x1,t1,x1 + MILLIRET + sub 0,x1,x1 + +/* DIVISION BY 12 (shift to divide by 4 then divide by 3) */ +GSYM($$divI_12) + .export $$divI_12,millicode + comb,< x2,0,LREF(neg12) + copy 0,x1 + extru x2,29,30,x2 /* divide by 4 */ + addib,tr 1,x2,LREF(pos) /* compute 5*(x2+1) = 5*x2+5 */ + sh2add x2,x2,x2 /* multiply by 5 to get started */ + +LSYM(neg12) + subi 4,x2,x2 /* negate, divide by 4, and add 1 */ + /* negation and adding 1 are done */ + /* at the same time by the SUBI */ + extru x2,29,30,x2 + b LREF(neg) + sh2add x2,x2,x2 /* multiply by 5 to get started */ + +GSYM($$divU_12) + .export $$divU_12,millicode + extru x2,29,30,x2 /* divide by 4 */ + addi 5,x2,t1 /* cannot carry */ + sh2add x2,t1,x2 /* multiply by 5 to get started */ + b LREF(pos) + addc 0,0,x1 + +/* DIVISION BY 15 (use z = 2**32; a = 11111111) */ +GSYM($$divI_15) + .export $$divI_15,millicode + comb,< x2,0,LREF(neg15) + copy 0,x1 + addib,tr 1,x2,LREF(pos)+4 + shd x1,x2,28,t1 + +LSYM(neg15) + b LREF(neg) + subi 1,x2,x2 + +GSYM($$divU_15) + .export $$divU_15,millicode + addi 1,x2,x2 /* this CAN overflow */ + b LREF(pos) + addc 0,0,x1 + +/* DIVISION BY 17 (use z = 2**32; a = f0f0f0f) */ +GSYM($$divI_17) + .export $$divI_17,millicode + comb,<,n x2,0,LREF(neg17) + addi 1,x2,x2 /* this cannot overflow */ + shd 0,x2,28,t1 /* multiply by 0xf to get started */ + shd x2,0,28,t2 + sub t2,x2,x2 + b LREF(pos_for_17) + subb t1,0,x1 + +LSYM(neg17) + subi 1,x2,x2 /* this cannot overflow */ + shd 0,x2,28,t1 /* multiply by 0xf to get started */ + shd x2,0,28,t2 + sub t2,x2,x2 + b LREF(neg_for_17) + subb t1,0,x1 + +GSYM($$divU_17) + .export $$divU_17,millicode + addi 1,x2,x2 /* this CAN overflow */ + addc 0,0,x1 + shd x1,x2,28,t1 /* multiply by 0xf to get started */ +LSYM(u17) + shd x2,0,28,t2 + sub t2,x2,x2 + b LREF(pos_for_17) + subb t1,x1,x1 + + +/* DIVISION BY DIVISORS OF FFFFFF, and powers of 2 times these + includes 7,9 and also 14 + + + z = 2**24-1 + r = z mod x = 0 + + so choose b = 0 + + Also, in order to divide by z = 2**24-1, we approximate by dividing + by (z+1) = 2**24 (which is easy), and then correcting. + + (ax) = (z+1)q' + r + . = zq' + (q'+r) + + So to compute (ax)/z, compute q' = (ax)/(z+1) and r = (ax) mod (z+1) + Then the true remainder of (ax)/z is (q'+r). Repeat the process + with this new remainder, adding the tentative quotients together, + until a tentative quotient is 0 (and then we are done). There is + one last correction to be done. It is possible that (q'+r) = z. + If so, then (q'+r)/(z+1) = 0 and it looks like we are done. But, + in fact, we need to add 1 more to the quotient. Now, it turns + out that this happens if and only if the original value x is + an exact multiple of y. So, to avoid a three instruction test at + the end, instead use 1 instruction to add 1 to x at the beginning. */ + +/* DIVISION BY 7 (use z = 2**24-1; a = 249249) */ +GSYM($$divI_7) + .export $$divI_7,millicode + comb,<,n x2,0,LREF(neg7) +LSYM(7) + addi 1,x2,x2 /* cannot overflow */ + shd 0,x2,29,x1 + sh3add x2,x2,x2 + addc x1,0,x1 +LSYM(pos7) + shd x1,x2,26,t1 + shd x2,0,26,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,20,t1 + shd x2,0,20,t2 + add x2,t2,x2 + addc x1,t1,t1 + + /* computed <t1,x2>. Now divide it by (2**24 - 1) */ + + copy 0,x1 + shd,= t1,x2,24,t1 /* tentative quotient */ +LSYM(1) + addb,tr t1,x1,LREF(2) /* add to previous quotient */ + extru x2,31,24,x2 /* new remainder (unadjusted) */ + + MILLIRETN + +LSYM(2) + addb,tr t1,x2,LREF(1) /* adjust remainder */ + extru,= x2,7,8,t1 /* new quotient */ + +LSYM(neg7) + subi 1,x2,x2 /* negate x2 and add 1 */ +LSYM(8) + shd 0,x2,29,x1 + sh3add x2,x2,x2 + addc x1,0,x1 + +LSYM(neg7_shift) + shd x1,x2,26,t1 + shd x2,0,26,t2 + add x2,t2,x2 + addc x1,t1,x1 + + shd x1,x2,20,t1 + shd x2,0,20,t2 + add x2,t2,x2 + addc x1,t1,t1 + + /* computed <t1,x2>. Now divide it by (2**24 - 1) */ + + copy 0,x1 + shd,= t1,x2,24,t1 /* tentative quotient */ +LSYM(3) + addb,tr t1,x1,LREF(4) /* add to previous quotient */ + extru x2,31,24,x2 /* new remainder (unadjusted) */ + + MILLIRET + sub 0,x1,x1 /* negate result */ + +LSYM(4) + addb,tr t1,x2,LREF(3) /* adjust remainder */ + extru,= x2,7,8,t1 /* new quotient */ + +GSYM($$divU_7) + .export $$divU_7,millicode + addi 1,x2,x2 /* can carry */ + addc 0,0,x1 + shd x1,x2,29,t1 + sh3add x2,x2,x2 + b LREF(pos7) + addc t1,x1,x1 + +/* DIVISION BY 9 (use z = 2**24-1; a = 1c71c7) */ +GSYM($$divI_9) + .export $$divI_9,millicode + comb,<,n x2,0,LREF(neg9) + addi 1,x2,x2 /* cannot overflow */ + shd 0,x2,29,t1 + shd x2,0,29,t2 + sub t2,x2,x2 + b LREF(pos7) + subb t1,0,x1 + +LSYM(neg9) + subi 1,x2,x2 /* negate and add 1 */ + shd 0,x2,29,t1 + shd x2,0,29,t2 + sub t2,x2,x2 + b LREF(neg7_shift) + subb t1,0,x1 + +GSYM($$divU_9) + .export $$divU_9,millicode + addi 1,x2,x2 /* can carry */ + addc 0,0,x1 + shd x1,x2,29,t1 + shd x2,0,29,t2 + sub t2,x2,x2 + b LREF(pos7) + subb t1,x1,x1 + +/* DIVISION BY 14 (shift to divide by 2 then divide by 7) */ +GSYM($$divI_14) + .export $$divI_14,millicode + comb,<,n x2,0,LREF(neg14) +GSYM($$divU_14) + .export $$divU_14,millicode + b LREF(7) /* go to 7 case */ + extru x2,30,31,x2 /* divide by 2 */ + +LSYM(neg14) + subi 2,x2,x2 /* negate (and add 2) */ + b LREF(8) + extru x2,30,31,x2 /* divide by 2 */ + .exit + .procend + .end +#endif + +#ifdef L_mulI +/* VERSION "@(#)$$mulI $ Revision: 12.4 $ $ Date: 94/03/17 17:18:51 $" */ +/****************************************************************************** +This routine is used on PA2.0 processors when gcc -mno-fpregs is used + +ROUTINE: $$mulI + + +DESCRIPTION: + + $$mulI multiplies two single word integers, giving a single + word result. + + +INPUT REGISTERS: + + arg0 = Operand 1 + arg1 = Operand 2 + r31 == return pc + sr0 == return space when called externally + + +OUTPUT REGISTERS: + + arg0 = undefined + arg1 = undefined + ret1 = result + +OTHER REGISTERS AFFECTED: + + r1 = undefined + +SIDE EFFECTS: + + Causes a trap under the following conditions: NONE + Changes memory at the following places: NONE + +PERMISSIBLE CONTEXT: + + Unwindable + Does not create a stack frame + Is usable for internal or external microcode + +DISCUSSION: + + Calls other millicode routines via mrp: NONE + Calls other millicode routines: NONE + +***************************************************************************/ + + +#define a0 %arg0 +#define a1 %arg1 +#define t0 %r1 +#define r %ret1 + +#define a0__128a0 zdep a0,24,25,a0 +#define a0__256a0 zdep a0,23,24,a0 +#define a1_ne_0_b_l0 comb,<> a1,0,LREF(l0) +#define a1_ne_0_b_l1 comb,<> a1,0,LREF(l1) +#define a1_ne_0_b_l2 comb,<> a1,0,LREF(l2) +#define b_n_ret_t0 b,n LREF(ret_t0) +#define b_e_shift b LREF(e_shift) +#define b_e_t0ma0 b LREF(e_t0ma0) +#define b_e_t0 b LREF(e_t0) +#define b_e_t0a0 b LREF(e_t0a0) +#define b_e_t02a0 b LREF(e_t02a0) +#define b_e_t04a0 b LREF(e_t04a0) +#define b_e_2t0 b LREF(e_2t0) +#define b_e_2t0a0 b LREF(e_2t0a0) +#define b_e_2t04a0 b LREF(e2t04a0) +#define b_e_3t0 b LREF(e_3t0) +#define b_e_4t0 b LREF(e_4t0) +#define b_e_4t0a0 b LREF(e_4t0a0) +#define b_e_4t08a0 b LREF(e4t08a0) +#define b_e_5t0 b LREF(e_5t0) +#define b_e_8t0 b LREF(e_8t0) +#define b_e_8t0a0 b LREF(e_8t0a0) +#define r__r_a0 add r,a0,r +#define r__r_2a0 sh1add a0,r,r +#define r__r_4a0 sh2add a0,r,r +#define r__r_8a0 sh3add a0,r,r +#define r__r_t0 add r,t0,r +#define r__r_2t0 sh1add t0,r,r +#define r__r_4t0 sh2add t0,r,r +#define r__r_8t0 sh3add t0,r,r +#define t0__3a0 sh1add a0,a0,t0 +#define t0__4a0 sh2add a0,0,t0 +#define t0__5a0 sh2add a0,a0,t0 +#define t0__8a0 sh3add a0,0,t0 +#define t0__9a0 sh3add a0,a0,t0 +#define t0__16a0 zdep a0,27,28,t0 +#define t0__32a0 zdep a0,26,27,t0 +#define t0__64a0 zdep a0,25,26,t0 +#define t0__128a0 zdep a0,24,25,t0 +#define t0__t0ma0 sub t0,a0,t0 +#define t0__t0_a0 add t0,a0,t0 +#define t0__t0_2a0 sh1add a0,t0,t0 +#define t0__t0_4a0 sh2add a0,t0,t0 +#define t0__t0_8a0 sh3add a0,t0,t0 +#define t0__2t0_a0 sh1add t0,a0,t0 +#define t0__3t0 sh1add t0,t0,t0 +#define t0__4t0 sh2add t0,0,t0 +#define t0__4t0_a0 sh2add t0,a0,t0 +#define t0__5t0 sh2add t0,t0,t0 +#define t0__8t0 sh3add t0,0,t0 +#define t0__8t0_a0 sh3add t0,a0,t0 +#define t0__9t0 sh3add t0,t0,t0 +#define t0__16t0 zdep t0,27,28,t0 +#define t0__32t0 zdep t0,26,27,t0 +#define t0__256a0 zdep a0,23,24,t0 + + + SUBSPA_MILLI + ATTR_MILLI + .align 16 + .proc + .callinfo millicode + .export $$mulI,millicode +GSYM($$mulI) + combt,<<= a1,a0,LREF(l4) /* swap args if unsigned a1>a0 */ + copy 0,r /* zero out the result */ + xor a0,a1,a0 /* swap a0 & a1 using the */ + xor a0,a1,a1 /* old xor trick */ + xor a0,a1,a0 +LSYM(l4) + combt,<= 0,a0,LREF(l3) /* if a0>=0 then proceed like unsigned */ + zdep a1,30,8,t0 /* t0 = (a1&0xff)<<1 ********* */ + sub,> 0,a1,t0 /* otherwise negate both and */ + combt,<=,n a0,t0,LREF(l2) /* swap back if |a0|<|a1| */ + sub 0,a0,a1 + movb,tr,n t0,a0,LREF(l2) /* 10th inst. */ + +LSYM(l0) r__r_t0 /* add in this partial product */ +LSYM(l1) a0__256a0 /* a0 <<= 8 ****************** */ +LSYM(l2) zdep a1,30,8,t0 /* t0 = (a1&0xff)<<1 ********* */ +LSYM(l3) blr t0,0 /* case on these 8 bits ****** */ + extru a1,23,24,a1 /* a1 >>= 8 ****************** */ + +/*16 insts before this. */ +/* a0 <<= 8 ************************** */ +LSYM(x0) a1_ne_0_b_l2 ! a0__256a0 ! MILLIRETN ! nop +LSYM(x1) a1_ne_0_b_l1 ! r__r_a0 ! MILLIRETN ! nop +LSYM(x2) a1_ne_0_b_l1 ! r__r_2a0 ! MILLIRETN ! nop +LSYM(x3) a1_ne_0_b_l0 ! t0__3a0 ! MILLIRET ! r__r_t0 +LSYM(x4) a1_ne_0_b_l1 ! r__r_4a0 ! MILLIRETN ! nop +LSYM(x5) a1_ne_0_b_l0 ! t0__5a0 ! MILLIRET ! r__r_t0 +LSYM(x6) t0__3a0 ! a1_ne_0_b_l1 ! r__r_2t0 ! MILLIRETN +LSYM(x7) t0__3a0 ! a1_ne_0_b_l0 ! r__r_4a0 ! b_n_ret_t0 +LSYM(x8) a1_ne_0_b_l1 ! r__r_8a0 ! MILLIRETN ! nop +LSYM(x9) a1_ne_0_b_l0 ! t0__9a0 ! MILLIRET ! r__r_t0 +LSYM(x10) t0__5a0 ! a1_ne_0_b_l1 ! r__r_2t0 ! MILLIRETN +LSYM(x11) t0__3a0 ! a1_ne_0_b_l0 ! r__r_8a0 ! b_n_ret_t0 +LSYM(x12) t0__3a0 ! a1_ne_0_b_l1 ! r__r_4t0 ! MILLIRETN +LSYM(x13) t0__5a0 ! a1_ne_0_b_l0 ! r__r_8a0 ! b_n_ret_t0 +LSYM(x14) t0__3a0 ! t0__2t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x15) t0__5a0 ! a1_ne_0_b_l0 ! t0__3t0 ! b_n_ret_t0 +LSYM(x16) t0__16a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x17) t0__9a0 ! a1_ne_0_b_l0 ! t0__t0_8a0 ! b_n_ret_t0 +LSYM(x18) t0__9a0 ! a1_ne_0_b_l1 ! r__r_2t0 ! MILLIRETN +LSYM(x19) t0__9a0 ! a1_ne_0_b_l0 ! t0__2t0_a0 ! b_n_ret_t0 +LSYM(x20) t0__5a0 ! a1_ne_0_b_l1 ! r__r_4t0 ! MILLIRETN +LSYM(x21) t0__5a0 ! a1_ne_0_b_l0 ! t0__4t0_a0 ! b_n_ret_t0 +LSYM(x22) t0__5a0 ! t0__2t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x23) t0__5a0 ! t0__2t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x24) t0__3a0 ! a1_ne_0_b_l1 ! r__r_8t0 ! MILLIRETN +LSYM(x25) t0__5a0 ! a1_ne_0_b_l0 ! t0__5t0 ! b_n_ret_t0 +LSYM(x26) t0__3a0 ! t0__4t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x27) t0__3a0 ! a1_ne_0_b_l0 ! t0__9t0 ! b_n_ret_t0 +LSYM(x28) t0__3a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x29) t0__3a0 ! t0__2t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x30) t0__5a0 ! t0__3t0 ! b_e_shift ! r__r_2t0 +LSYM(x31) t0__32a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +LSYM(x32) t0__32a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x33) t0__8a0 ! a1_ne_0_b_l0 ! t0__4t0_a0 ! b_n_ret_t0 +LSYM(x34) t0__16a0 ! t0__t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x35) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__t0_8a0 +LSYM(x36) t0__9a0 ! a1_ne_0_b_l1 ! r__r_4t0 ! MILLIRETN +LSYM(x37) t0__9a0 ! a1_ne_0_b_l0 ! t0__4t0_a0 ! b_n_ret_t0 +LSYM(x38) t0__9a0 ! t0__2t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x39) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x40) t0__5a0 ! a1_ne_0_b_l1 ! r__r_8t0 ! MILLIRETN +LSYM(x41) t0__5a0 ! a1_ne_0_b_l0 ! t0__8t0_a0 ! b_n_ret_t0 +LSYM(x42) t0__5a0 ! t0__4t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x43) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x44) t0__5a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x45) t0__9a0 ! a1_ne_0_b_l0 ! t0__5t0 ! b_n_ret_t0 +LSYM(x46) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__t0_a0 +LSYM(x47) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__t0_2a0 +LSYM(x48) t0__3a0 ! a1_ne_0_b_l0 ! t0__16t0 ! b_n_ret_t0 +LSYM(x49) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__t0_4a0 +LSYM(x50) t0__5a0 ! t0__5t0 ! b_e_shift ! r__r_2t0 +LSYM(x51) t0__9a0 ! t0__t0_8a0 ! b_e_t0 ! t0__3t0 +LSYM(x52) t0__3a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x53) t0__3a0 ! t0__4t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x54) t0__9a0 ! t0__3t0 ! b_e_shift ! r__r_2t0 +LSYM(x55) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x56) t0__3a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x57) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x58) t0__3a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x59) t0__9a0 ! t0__2t0_a0 ! b_e_t02a0 ! t0__3t0 +LSYM(x60) t0__5a0 ! t0__3t0 ! b_e_shift ! r__r_4t0 +LSYM(x61) t0__5a0 ! t0__3t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x62) t0__32a0 ! t0__t0ma0 ! b_e_shift ! r__r_2t0 +LSYM(x63) t0__64a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +LSYM(x64) t0__64a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x65) t0__8a0 ! a1_ne_0_b_l0 ! t0__8t0_a0 ! b_n_ret_t0 +LSYM(x66) t0__32a0 ! t0__t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x67) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x68) t0__8a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x69) t0__8a0 ! t0__2t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x70) t0__64a0 ! t0__t0_4a0 ! b_e_t0 ! t0__t0_2a0 +LSYM(x71) t0__9a0 ! t0__8t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x72) t0__9a0 ! a1_ne_0_b_l1 ! r__r_8t0 ! MILLIRETN +LSYM(x73) t0__9a0 ! t0__8t0_a0 ! b_e_shift ! r__r_t0 +LSYM(x74) t0__9a0 ! t0__4t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x75) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x76) t0__9a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x77) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x78) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x79) t0__16a0 ! t0__5t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x80) t0__16a0 ! t0__5t0 ! b_e_shift ! r__r_t0 +LSYM(x81) t0__9a0 ! t0__9t0 ! b_e_shift ! r__r_t0 +LSYM(x82) t0__5a0 ! t0__8t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x83) t0__5a0 ! t0__8t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x84) t0__5a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x85) t0__8a0 ! t0__2t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x86) t0__5a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x87) t0__9a0 ! t0__9t0 ! b_e_t02a0 ! t0__t0_4a0 +LSYM(x88) t0__5a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x89) t0__5a0 ! t0__2t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x90) t0__9a0 ! t0__5t0 ! b_e_shift ! r__r_2t0 +LSYM(x91) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x92) t0__5a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x93) t0__32a0 ! t0__t0ma0 ! b_e_t0 ! t0__3t0 +LSYM(x94) t0__9a0 ! t0__5t0 ! b_e_2t0 ! t0__t0_2a0 +LSYM(x95) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x96) t0__8a0 ! t0__3t0 ! b_e_shift ! r__r_4t0 +LSYM(x97) t0__8a0 ! t0__3t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x98) t0__32a0 ! t0__3t0 ! b_e_t0 ! t0__t0_2a0 +LSYM(x99) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x100) t0__5a0 ! t0__5t0 ! b_e_shift ! r__r_4t0 +LSYM(x101) t0__5a0 ! t0__5t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x102) t0__32a0 ! t0__t0_2a0 ! b_e_t0 ! t0__3t0 +LSYM(x103) t0__5a0 ! t0__5t0 ! b_e_t02a0 ! t0__4t0_a0 +LSYM(x104) t0__3a0 ! t0__4t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x105) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x106) t0__3a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x107) t0__9a0 ! t0__t0_4a0 ! b_e_t02a0 ! t0__8t0_a0 +LSYM(x108) t0__9a0 ! t0__3t0 ! b_e_shift ! r__r_4t0 +LSYM(x109) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x110) t0__9a0 ! t0__3t0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x111) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x112) t0__3a0 ! t0__2t0_a0 ! b_e_t0 ! t0__16t0 +LSYM(x113) t0__9a0 ! t0__4t0_a0 ! b_e_t02a0 ! t0__3t0 +LSYM(x114) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__3t0 +LSYM(x115) t0__9a0 ! t0__2t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x116) t0__3a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__4t0_a0 +LSYM(x117) t0__3a0 ! t0__4t0_a0 ! b_e_t0 ! t0__9t0 +LSYM(x118) t0__3a0 ! t0__4t0_a0 ! b_e_t0a0 ! t0__9t0 +LSYM(x119) t0__3a0 ! t0__4t0_a0 ! b_e_t02a0 ! t0__9t0 +LSYM(x120) t0__5a0 ! t0__3t0 ! b_e_shift ! r__r_8t0 +LSYM(x121) t0__5a0 ! t0__3t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x122) t0__5a0 ! t0__3t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x123) t0__5a0 ! t0__8t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x124) t0__32a0 ! t0__t0ma0 ! b_e_shift ! r__r_4t0 +LSYM(x125) t0__5a0 ! t0__5t0 ! b_e_t0 ! t0__5t0 +LSYM(x126) t0__64a0 ! t0__t0ma0 ! b_e_shift ! r__r_2t0 +LSYM(x127) t0__128a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +LSYM(x128) t0__128a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x129) t0__128a0 ! a1_ne_0_b_l0 ! t0__t0_a0 ! b_n_ret_t0 +LSYM(x130) t0__64a0 ! t0__t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x131) t0__8a0 ! t0__8t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x132) t0__8a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x133) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x134) t0__8a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x135) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__3t0 +LSYM(x136) t0__8a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x137) t0__8a0 ! t0__2t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x138) t0__8a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x139) t0__8a0 ! t0__2t0_a0 ! b_e_2t0a0 ! t0__4t0_a0 +LSYM(x140) t0__3a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__5t0 +LSYM(x141) t0__8a0 ! t0__2t0_a0 ! b_e_4t0a0 ! t0__2t0_a0 +LSYM(x142) t0__9a0 ! t0__8t0 ! b_e_2t0 ! t0__t0ma0 +LSYM(x143) t0__16a0 ! t0__9t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x144) t0__9a0 ! t0__8t0 ! b_e_shift ! r__r_2t0 +LSYM(x145) t0__9a0 ! t0__8t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x146) t0__9a0 ! t0__8t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x147) t0__9a0 ! t0__8t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x148) t0__9a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x149) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x150) t0__9a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x151) t0__9a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__2t0_a0 +LSYM(x152) t0__9a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x153) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x154) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x155) t0__32a0 ! t0__t0ma0 ! b_e_t0 ! t0__5t0 +LSYM(x156) t0__9a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x157) t0__32a0 ! t0__t0ma0 ! b_e_t02a0 ! t0__5t0 +LSYM(x158) t0__16a0 ! t0__5t0 ! b_e_2t0 ! t0__t0ma0 +LSYM(x159) t0__32a0 ! t0__5t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x160) t0__5a0 ! t0__4t0 ! b_e_shift ! r__r_8t0 +LSYM(x161) t0__8a0 ! t0__5t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x162) t0__9a0 ! t0__9t0 ! b_e_shift ! r__r_2t0 +LSYM(x163) t0__9a0 ! t0__9t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x164) t0__5a0 ! t0__8t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x165) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x166) t0__5a0 ! t0__8t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x167) t0__5a0 ! t0__8t0_a0 ! b_e_2t0a0 ! t0__2t0_a0 +LSYM(x168) t0__5a0 ! t0__4t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x169) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x170) t0__32a0 ! t0__t0_2a0 ! b_e_t0 ! t0__5t0 +LSYM(x171) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__9t0 +LSYM(x172) t0__5a0 ! t0__4t0_a0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x173) t0__9a0 ! t0__2t0_a0 ! b_e_t02a0 ! t0__9t0 +LSYM(x174) t0__32a0 ! t0__t0_2a0 ! b_e_t04a0 ! t0__5t0 +LSYM(x175) t0__8a0 ! t0__2t0_a0 ! b_e_5t0 ! t0__2t0_a0 +LSYM(x176) t0__5a0 ! t0__4t0_a0 ! b_e_8t0 ! t0__t0_a0 +LSYM(x177) t0__5a0 ! t0__4t0_a0 ! b_e_8t0a0 ! t0__t0_a0 +LSYM(x178) t0__5a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__8t0_a0 +LSYM(x179) t0__5a0 ! t0__2t0_a0 ! b_e_2t0a0 ! t0__8t0_a0 +LSYM(x180) t0__9a0 ! t0__5t0 ! b_e_shift ! r__r_4t0 +LSYM(x181) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x182) t0__9a0 ! t0__5t0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x183) t0__9a0 ! t0__5t0 ! b_e_2t0a0 ! t0__2t0_a0 +LSYM(x184) t0__5a0 ! t0__9t0 ! b_e_4t0 ! t0__t0_a0 +LSYM(x185) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x186) t0__32a0 ! t0__t0ma0 ! b_e_2t0 ! t0__3t0 +LSYM(x187) t0__9a0 ! t0__4t0_a0 ! b_e_t02a0 ! t0__5t0 +LSYM(x188) t0__9a0 ! t0__5t0 ! b_e_4t0 ! t0__t0_2a0 +LSYM(x189) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__9t0 +LSYM(x190) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__5t0 +LSYM(x191) t0__64a0 ! t0__3t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x192) t0__8a0 ! t0__3t0 ! b_e_shift ! r__r_8t0 +LSYM(x193) t0__8a0 ! t0__3t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x194) t0__8a0 ! t0__3t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x195) t0__8a0 ! t0__8t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x196) t0__8a0 ! t0__3t0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x197) t0__8a0 ! t0__3t0 ! b_e_4t0a0 ! t0__2t0_a0 +LSYM(x198) t0__64a0 ! t0__t0_2a0 ! b_e_t0 ! t0__3t0 +LSYM(x199) t0__8a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x200) t0__5a0 ! t0__5t0 ! b_e_shift ! r__r_8t0 +LSYM(x201) t0__5a0 ! t0__5t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x202) t0__5a0 ! t0__5t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x203) t0__5a0 ! t0__5t0 ! b_e_2t0a0 ! t0__4t0_a0 +LSYM(x204) t0__8a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__3t0 +LSYM(x205) t0__5a0 ! t0__8t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x206) t0__64a0 ! t0__t0_4a0 ! b_e_t02a0 ! t0__3t0 +LSYM(x207) t0__8a0 ! t0__2t0_a0 ! b_e_3t0 ! t0__4t0_a0 +LSYM(x208) t0__5a0 ! t0__5t0 ! b_e_8t0 ! t0__t0_a0 +LSYM(x209) t0__5a0 ! t0__5t0 ! b_e_8t0a0 ! t0__t0_a0 +LSYM(x210) t0__5a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__5t0 +LSYM(x211) t0__5a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__5t0 +LSYM(x212) t0__3a0 ! t0__4t0_a0 ! b_e_4t0 ! t0__4t0_a0 +LSYM(x213) t0__3a0 ! t0__4t0_a0 ! b_e_4t0a0 ! t0__4t0_a0 +LSYM(x214) t0__9a0 ! t0__t0_4a0 ! b_e_2t04a0 ! t0__8t0_a0 +LSYM(x215) t0__5a0 ! t0__4t0_a0 ! b_e_5t0 ! t0__2t0_a0 +LSYM(x216) t0__9a0 ! t0__3t0 ! b_e_shift ! r__r_8t0 +LSYM(x217) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x218) t0__9a0 ! t0__3t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x219) t0__9a0 ! t0__8t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x220) t0__3a0 ! t0__9t0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x221) t0__3a0 ! t0__9t0 ! b_e_4t0a0 ! t0__2t0_a0 +LSYM(x222) t0__9a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__3t0 +LSYM(x223) t0__9a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x224) t0__9a0 ! t0__3t0 ! b_e_8t0 ! t0__t0_a0 +LSYM(x225) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__5t0 +LSYM(x226) t0__3a0 ! t0__2t0_a0 ! b_e_t02a0 ! t0__32t0 +LSYM(x227) t0__9a0 ! t0__5t0 ! b_e_t02a0 ! t0__5t0 +LSYM(x228) t0__9a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__3t0 +LSYM(x229) t0__9a0 ! t0__2t0_a0 ! b_e_4t0a0 ! t0__3t0 +LSYM(x230) t0__9a0 ! t0__5t0 ! b_e_5t0 ! t0__t0_a0 +LSYM(x231) t0__9a0 ! t0__2t0_a0 ! b_e_3t0 ! t0__4t0_a0 +LSYM(x232) t0__3a0 ! t0__2t0_a0 ! b_e_8t0 ! t0__4t0_a0 +LSYM(x233) t0__3a0 ! t0__2t0_a0 ! b_e_8t0a0 ! t0__4t0_a0 +LSYM(x234) t0__3a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__9t0 +LSYM(x235) t0__3a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__9t0 +LSYM(x236) t0__9a0 ! t0__2t0_a0 ! b_e_4t08a0 ! t0__3t0 +LSYM(x237) t0__16a0 ! t0__5t0 ! b_e_3t0 ! t0__t0ma0 +LSYM(x238) t0__3a0 ! t0__4t0_a0 ! b_e_2t04a0 ! t0__9t0 +LSYM(x239) t0__16a0 ! t0__5t0 ! b_e_t0ma0 ! t0__3t0 +LSYM(x240) t0__9a0 ! t0__t0_a0 ! b_e_8t0 ! t0__3t0 +LSYM(x241) t0__9a0 ! t0__t0_a0 ! b_e_8t0a0 ! t0__3t0 +LSYM(x242) t0__5a0 ! t0__3t0 ! b_e_2t0 ! t0__8t0_a0 +LSYM(x243) t0__9a0 ! t0__9t0 ! b_e_t0 ! t0__3t0 +LSYM(x244) t0__5a0 ! t0__3t0 ! b_e_4t0 ! t0__4t0_a0 +LSYM(x245) t0__8a0 ! t0__3t0 ! b_e_5t0 ! t0__2t0_a0 +LSYM(x246) t0__5a0 ! t0__8t0_a0 ! b_e_2t0 ! t0__3t0 +LSYM(x247) t0__5a0 ! t0__8t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x248) t0__32a0 ! t0__t0ma0 ! b_e_shift ! r__r_8t0 +LSYM(x249) t0__32a0 ! t0__t0ma0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x250) t0__5a0 ! t0__5t0 ! b_e_2t0 ! t0__5t0 +LSYM(x251) t0__5a0 ! t0__5t0 ! b_e_2t0a0 ! t0__5t0 +LSYM(x252) t0__64a0 ! t0__t0ma0 ! b_e_shift ! r__r_4t0 +LSYM(x253) t0__64a0 ! t0__t0ma0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x254) t0__128a0 ! t0__t0ma0 ! b_e_shift ! r__r_2t0 +LSYM(x255) t0__256a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +/*1040 insts before this. */ +LSYM(ret_t0) MILLIRET +LSYM(e_t0) r__r_t0 +LSYM(e_shift) a1_ne_0_b_l2 + a0__256a0 /* a0 <<= 8 *********** */ + MILLIRETN +LSYM(e_t0ma0) a1_ne_0_b_l0 + t0__t0ma0 + MILLIRET + r__r_t0 +LSYM(e_t0a0) a1_ne_0_b_l0 + t0__t0_a0 + MILLIRET + r__r_t0 +LSYM(e_t02a0) a1_ne_0_b_l0 + t0__t0_2a0 + MILLIRET + r__r_t0 +LSYM(e_t04a0) a1_ne_0_b_l0 + t0__t0_4a0 + MILLIRET + r__r_t0 +LSYM(e_2t0) a1_ne_0_b_l1 + r__r_2t0 + MILLIRETN +LSYM(e_2t0a0) a1_ne_0_b_l0 + t0__2t0_a0 + MILLIRET + r__r_t0 +LSYM(e2t04a0) t0__t0_2a0 + a1_ne_0_b_l1 + r__r_2t0 + MILLIRETN +LSYM(e_3t0) a1_ne_0_b_l0 + t0__3t0 + MILLIRET + r__r_t0 +LSYM(e_4t0) a1_ne_0_b_l1 + r__r_4t0 + MILLIRETN +LSYM(e_4t0a0) a1_ne_0_b_l0 + t0__4t0_a0 + MILLIRET + r__r_t0 +LSYM(e4t08a0) t0__t0_2a0 + a1_ne_0_b_l1 + r__r_4t0 + MILLIRETN +LSYM(e_5t0) a1_ne_0_b_l0 + t0__5t0 + MILLIRET + r__r_t0 +LSYM(e_8t0) a1_ne_0_b_l1 + r__r_8t0 + MILLIRETN +LSYM(e_8t0a0) a1_ne_0_b_l0 + t0__8t0_a0 + MILLIRET + r__r_t0 + + .procend + .end +#endif diff --git a/arch/parisc/lib/milli/milli.h b/arch/parisc/lib/milli/milli.h new file mode 100644 index 00000000000..19ac79f336d --- /dev/null +++ b/arch/parisc/lib/milli/milli.h @@ -0,0 +1,165 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#ifndef _PA_MILLI_H_ +#define _PA_MILLI_H_ + +#define L_dyncall +#define L_divI +#define L_divU +#define L_remI +#define L_remU +#define L_div_const +#define L_mulI + +#ifdef CONFIG_64BIT + .level 2.0w +#endif + +/* Hardware General Registers. */ +r0: .reg %r0 +r1: .reg %r1 +r2: .reg %r2 +r3: .reg %r3 +r4: .reg %r4 +r5: .reg %r5 +r6: .reg %r6 +r7: .reg %r7 +r8: .reg %r8 +r9: .reg %r9 +r10: .reg %r10 +r11: .reg %r11 +r12: .reg %r12 +r13: .reg %r13 +r14: .reg %r14 +r15: .reg %r15 +r16: .reg %r16 +r17: .reg %r17 +r18: .reg %r18 +r19: .reg %r19 +r20: .reg %r20 +r21: .reg %r21 +r22: .reg %r22 +r23: .reg %r23 +r24: .reg %r24 +r25: .reg %r25 +r26: .reg %r26 +r27: .reg %r27 +r28: .reg %r28 +r29: .reg %r29 +r30: .reg %r30 +r31: .reg %r31 + +/* Hardware Space Registers. */ +sr0: .reg %sr0 +sr1: .reg %sr1 +sr2: .reg %sr2 +sr3: .reg %sr3 +sr4: .reg %sr4 +sr5: .reg %sr5 +sr6: .reg %sr6 +sr7: .reg %sr7 + +/* Hardware Floating Point Registers. */ +fr0: .reg %fr0 +fr1: .reg %fr1 +fr2: .reg %fr2 +fr3: .reg %fr3 +fr4: .reg %fr4 +fr5: .reg %fr5 +fr6: .reg %fr6 +fr7: .reg %fr7 +fr8: .reg %fr8 +fr9: .reg %fr9 +fr10: .reg %fr10 +fr11: .reg %fr11 +fr12: .reg %fr12 +fr13: .reg %fr13 +fr14: .reg %fr14 +fr15: .reg %fr15 + +/* Hardware Control Registers. */ +cr11: .reg %cr11 +sar: .reg %cr11 /* Shift Amount Register */ + +/* Software Architecture General Registers. */ +rp: .reg r2 /* return pointer */ +#ifdef CONFIG_64BIT +mrp: .reg r2 /* millicode return pointer */ +#else +mrp: .reg r31 /* millicode return pointer */ +#endif +ret0: .reg r28 /* return value */ +ret1: .reg r29 /* return value (high part of double) */ +sp: .reg r30 /* stack pointer */ +dp: .reg r27 /* data pointer */ +arg0: .reg r26 /* argument */ +arg1: .reg r25 /* argument or high part of double argument */ +arg2: .reg r24 /* argument */ +arg3: .reg r23 /* argument or high part of double argument */ + +/* Software Architecture Space Registers. */ +/* sr0 ; return link from BLE */ +sret: .reg sr1 /* return value */ +sarg: .reg sr1 /* argument */ +/* sr4 ; PC SPACE tracker */ +/* sr5 ; process private data */ + +/* Frame Offsets (millicode convention!) Used when calling other + millicode routines. Stack unwinding is dependent upon these + definitions. */ +r31_slot: .equ -20 /* "current RP" slot */ +sr0_slot: .equ -16 /* "static link" slot */ +#if defined(CONFIG_64BIT) +mrp_slot: .equ -16 /* "current RP" slot */ +psp_slot: .equ -8 /* "previous SP" slot */ +#else +mrp_slot: .equ -20 /* "current RP" slot (replacing "r31_slot") */ +#endif + + +#define DEFINE(name,value)name: .EQU value +#define RDEFINE(name,value)name: .REG value +#ifdef milliext +#define MILLI_BE(lbl) BE lbl(sr7,r0) +#define MILLI_BEN(lbl) BE,n lbl(sr7,r0) +#define MILLI_BLE(lbl) BLE lbl(sr7,r0) +#define MILLI_BLEN(lbl) BLE,n lbl(sr7,r0) +#define MILLIRETN BE,n 0(sr0,mrp) +#define MILLIRET BE 0(sr0,mrp) +#define MILLI_RETN BE,n 0(sr0,mrp) +#define MILLI_RET BE 0(sr0,mrp) +#else +#define MILLI_BE(lbl) B lbl +#define MILLI_BEN(lbl) B,n lbl +#define MILLI_BLE(lbl) BL lbl,mrp +#define MILLI_BLEN(lbl) BL,n lbl,mrp +#define MILLIRETN BV,n 0(mrp) +#define MILLIRET BV 0(mrp) +#define MILLI_RETN BV,n 0(mrp) +#define MILLI_RET BV 0(mrp) +#endif + +#define CAT(a,b) a##b + +#define SUBSPA_MILLI .section .text +#define SUBSPA_MILLI_DIV .section .text.div,"ax",@progbits! .align 16 +#define SUBSPA_MILLI_MUL .section .text.mul,"ax",@progbits! .align 16 +#define ATTR_MILLI +#define SUBSPA_DATA .section .data +#define ATTR_DATA +#define GLOBAL $global$ +#define GSYM(sym) !sym: +#define LSYM(sym) !CAT(.L,sym:) +#define LREF(sym) CAT(.L,sym) + +#endif /*_PA_MILLI_H_*/ diff --git a/arch/parisc/lib/milli/mulI.S b/arch/parisc/lib/milli/mulI.S new file mode 100644 index 00000000000..4c7e0c36d15 --- /dev/null +++ b/arch/parisc/lib/milli/mulI.S @@ -0,0 +1,474 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_mulI +/* VERSION "@(#)$$mulI $ Revision: 12.4 $ $ Date: 94/03/17 17:18:51 $" */ +/****************************************************************************** +This routine is used on PA2.0 processors when gcc -mno-fpregs is used + +ROUTINE: $$mulI + + +DESCRIPTION: + + $$mulI multiplies two single word integers, giving a single + word result. + + +INPUT REGISTERS: + + arg0 = Operand 1 + arg1 = Operand 2 + r31 == return pc + sr0 == return space when called externally + + +OUTPUT REGISTERS: + + arg0 = undefined + arg1 = undefined + ret1 = result + +OTHER REGISTERS AFFECTED: + + r1 = undefined + +SIDE EFFECTS: + + Causes a trap under the following conditions: NONE + Changes memory at the following places: NONE + +PERMISSIBLE CONTEXT: + + Unwindable + Does not create a stack frame + Is usable for internal or external microcode + +DISCUSSION: + + Calls other millicode routines via mrp: NONE + Calls other millicode routines: NONE + +***************************************************************************/ + + +#define a0 %arg0 +#define a1 %arg1 +#define t0 %r1 +#define r %ret1 + +#define a0__128a0 zdep a0,24,25,a0 +#define a0__256a0 zdep a0,23,24,a0 +#define a1_ne_0_b_l0 comb,<> a1,0,LREF(l0) +#define a1_ne_0_b_l1 comb,<> a1,0,LREF(l1) +#define a1_ne_0_b_l2 comb,<> a1,0,LREF(l2) +#define b_n_ret_t0 b,n LREF(ret_t0) +#define b_e_shift b LREF(e_shift) +#define b_e_t0ma0 b LREF(e_t0ma0) +#define b_e_t0 b LREF(e_t0) +#define b_e_t0a0 b LREF(e_t0a0) +#define b_e_t02a0 b LREF(e_t02a0) +#define b_e_t04a0 b LREF(e_t04a0) +#define b_e_2t0 b LREF(e_2t0) +#define b_e_2t0a0 b LREF(e_2t0a0) +#define b_e_2t04a0 b LREF(e2t04a0) +#define b_e_3t0 b LREF(e_3t0) +#define b_e_4t0 b LREF(e_4t0) +#define b_e_4t0a0 b LREF(e_4t0a0) +#define b_e_4t08a0 b LREF(e4t08a0) +#define b_e_5t0 b LREF(e_5t0) +#define b_e_8t0 b LREF(e_8t0) +#define b_e_8t0a0 b LREF(e_8t0a0) +#define r__r_a0 add r,a0,r +#define r__r_2a0 sh1add a0,r,r +#define r__r_4a0 sh2add a0,r,r +#define r__r_8a0 sh3add a0,r,r +#define r__r_t0 add r,t0,r +#define r__r_2t0 sh1add t0,r,r +#define r__r_4t0 sh2add t0,r,r +#define r__r_8t0 sh3add t0,r,r +#define t0__3a0 sh1add a0,a0,t0 +#define t0__4a0 sh2add a0,0,t0 +#define t0__5a0 sh2add a0,a0,t0 +#define t0__8a0 sh3add a0,0,t0 +#define t0__9a0 sh3add a0,a0,t0 +#define t0__16a0 zdep a0,27,28,t0 +#define t0__32a0 zdep a0,26,27,t0 +#define t0__64a0 zdep a0,25,26,t0 +#define t0__128a0 zdep a0,24,25,t0 +#define t0__t0ma0 sub t0,a0,t0 +#define t0__t0_a0 add t0,a0,t0 +#define t0__t0_2a0 sh1add a0,t0,t0 +#define t0__t0_4a0 sh2add a0,t0,t0 +#define t0__t0_8a0 sh3add a0,t0,t0 +#define t0__2t0_a0 sh1add t0,a0,t0 +#define t0__3t0 sh1add t0,t0,t0 +#define t0__4t0 sh2add t0,0,t0 +#define t0__4t0_a0 sh2add t0,a0,t0 +#define t0__5t0 sh2add t0,t0,t0 +#define t0__8t0 sh3add t0,0,t0 +#define t0__8t0_a0 sh3add t0,a0,t0 +#define t0__9t0 sh3add t0,t0,t0 +#define t0__16t0 zdep t0,27,28,t0 +#define t0__32t0 zdep t0,26,27,t0 +#define t0__256a0 zdep a0,23,24,t0 + + + SUBSPA_MILLI + ATTR_MILLI + .align 16 + .proc + .callinfo millicode + .export $$mulI,millicode +GSYM($$mulI) + combt,<<= a1,a0,LREF(l4) /* swap args if unsigned a1>a0 */ + copy 0,r /* zero out the result */ + xor a0,a1,a0 /* swap a0 & a1 using the */ + xor a0,a1,a1 /* old xor trick */ + xor a0,a1,a0 +LSYM(l4) + combt,<= 0,a0,LREF(l3) /* if a0>=0 then proceed like unsigned */ + zdep a1,30,8,t0 /* t0 = (a1&0xff)<<1 ********* */ + sub,> 0,a1,t0 /* otherwise negate both and */ + combt,<=,n a0,t0,LREF(l2) /* swap back if |a0|<|a1| */ + sub 0,a0,a1 + movb,tr,n t0,a0,LREF(l2) /* 10th inst. */ + +LSYM(l0) r__r_t0 /* add in this partial product */ +LSYM(l1) a0__256a0 /* a0 <<= 8 ****************** */ +LSYM(l2) zdep a1,30,8,t0 /* t0 = (a1&0xff)<<1 ********* */ +LSYM(l3) blr t0,0 /* case on these 8 bits ****** */ + extru a1,23,24,a1 /* a1 >>= 8 ****************** */ + +/*16 insts before this. */ +/* a0 <<= 8 ************************** */ +LSYM(x0) a1_ne_0_b_l2 ! a0__256a0 ! MILLIRETN ! nop +LSYM(x1) a1_ne_0_b_l1 ! r__r_a0 ! MILLIRETN ! nop +LSYM(x2) a1_ne_0_b_l1 ! r__r_2a0 ! MILLIRETN ! nop +LSYM(x3) a1_ne_0_b_l0 ! t0__3a0 ! MILLIRET ! r__r_t0 +LSYM(x4) a1_ne_0_b_l1 ! r__r_4a0 ! MILLIRETN ! nop +LSYM(x5) a1_ne_0_b_l0 ! t0__5a0 ! MILLIRET ! r__r_t0 +LSYM(x6) t0__3a0 ! a1_ne_0_b_l1 ! r__r_2t0 ! MILLIRETN +LSYM(x7) t0__3a0 ! a1_ne_0_b_l0 ! r__r_4a0 ! b_n_ret_t0 +LSYM(x8) a1_ne_0_b_l1 ! r__r_8a0 ! MILLIRETN ! nop +LSYM(x9) a1_ne_0_b_l0 ! t0__9a0 ! MILLIRET ! r__r_t0 +LSYM(x10) t0__5a0 ! a1_ne_0_b_l1 ! r__r_2t0 ! MILLIRETN +LSYM(x11) t0__3a0 ! a1_ne_0_b_l0 ! r__r_8a0 ! b_n_ret_t0 +LSYM(x12) t0__3a0 ! a1_ne_0_b_l1 ! r__r_4t0 ! MILLIRETN +LSYM(x13) t0__5a0 ! a1_ne_0_b_l0 ! r__r_8a0 ! b_n_ret_t0 +LSYM(x14) t0__3a0 ! t0__2t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x15) t0__5a0 ! a1_ne_0_b_l0 ! t0__3t0 ! b_n_ret_t0 +LSYM(x16) t0__16a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x17) t0__9a0 ! a1_ne_0_b_l0 ! t0__t0_8a0 ! b_n_ret_t0 +LSYM(x18) t0__9a0 ! a1_ne_0_b_l1 ! r__r_2t0 ! MILLIRETN +LSYM(x19) t0__9a0 ! a1_ne_0_b_l0 ! t0__2t0_a0 ! b_n_ret_t0 +LSYM(x20) t0__5a0 ! a1_ne_0_b_l1 ! r__r_4t0 ! MILLIRETN +LSYM(x21) t0__5a0 ! a1_ne_0_b_l0 ! t0__4t0_a0 ! b_n_ret_t0 +LSYM(x22) t0__5a0 ! t0__2t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x23) t0__5a0 ! t0__2t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x24) t0__3a0 ! a1_ne_0_b_l1 ! r__r_8t0 ! MILLIRETN +LSYM(x25) t0__5a0 ! a1_ne_0_b_l0 ! t0__5t0 ! b_n_ret_t0 +LSYM(x26) t0__3a0 ! t0__4t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x27) t0__3a0 ! a1_ne_0_b_l0 ! t0__9t0 ! b_n_ret_t0 +LSYM(x28) t0__3a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x29) t0__3a0 ! t0__2t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x30) t0__5a0 ! t0__3t0 ! b_e_shift ! r__r_2t0 +LSYM(x31) t0__32a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +LSYM(x32) t0__32a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x33) t0__8a0 ! a1_ne_0_b_l0 ! t0__4t0_a0 ! b_n_ret_t0 +LSYM(x34) t0__16a0 ! t0__t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x35) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__t0_8a0 +LSYM(x36) t0__9a0 ! a1_ne_0_b_l1 ! r__r_4t0 ! MILLIRETN +LSYM(x37) t0__9a0 ! a1_ne_0_b_l0 ! t0__4t0_a0 ! b_n_ret_t0 +LSYM(x38) t0__9a0 ! t0__2t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x39) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x40) t0__5a0 ! a1_ne_0_b_l1 ! r__r_8t0 ! MILLIRETN +LSYM(x41) t0__5a0 ! a1_ne_0_b_l0 ! t0__8t0_a0 ! b_n_ret_t0 +LSYM(x42) t0__5a0 ! t0__4t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x43) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x44) t0__5a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x45) t0__9a0 ! a1_ne_0_b_l0 ! t0__5t0 ! b_n_ret_t0 +LSYM(x46) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__t0_a0 +LSYM(x47) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__t0_2a0 +LSYM(x48) t0__3a0 ! a1_ne_0_b_l0 ! t0__16t0 ! b_n_ret_t0 +LSYM(x49) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__t0_4a0 +LSYM(x50) t0__5a0 ! t0__5t0 ! b_e_shift ! r__r_2t0 +LSYM(x51) t0__9a0 ! t0__t0_8a0 ! b_e_t0 ! t0__3t0 +LSYM(x52) t0__3a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x53) t0__3a0 ! t0__4t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x54) t0__9a0 ! t0__3t0 ! b_e_shift ! r__r_2t0 +LSYM(x55) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x56) t0__3a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x57) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x58) t0__3a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x59) t0__9a0 ! t0__2t0_a0 ! b_e_t02a0 ! t0__3t0 +LSYM(x60) t0__5a0 ! t0__3t0 ! b_e_shift ! r__r_4t0 +LSYM(x61) t0__5a0 ! t0__3t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x62) t0__32a0 ! t0__t0ma0 ! b_e_shift ! r__r_2t0 +LSYM(x63) t0__64a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +LSYM(x64) t0__64a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x65) t0__8a0 ! a1_ne_0_b_l0 ! t0__8t0_a0 ! b_n_ret_t0 +LSYM(x66) t0__32a0 ! t0__t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x67) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x68) t0__8a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x69) t0__8a0 ! t0__2t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x70) t0__64a0 ! t0__t0_4a0 ! b_e_t0 ! t0__t0_2a0 +LSYM(x71) t0__9a0 ! t0__8t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x72) t0__9a0 ! a1_ne_0_b_l1 ! r__r_8t0 ! MILLIRETN +LSYM(x73) t0__9a0 ! t0__8t0_a0 ! b_e_shift ! r__r_t0 +LSYM(x74) t0__9a0 ! t0__4t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x75) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x76) t0__9a0 ! t0__2t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x77) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x78) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x79) t0__16a0 ! t0__5t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x80) t0__16a0 ! t0__5t0 ! b_e_shift ! r__r_t0 +LSYM(x81) t0__9a0 ! t0__9t0 ! b_e_shift ! r__r_t0 +LSYM(x82) t0__5a0 ! t0__8t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x83) t0__5a0 ! t0__8t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x84) t0__5a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x85) t0__8a0 ! t0__2t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x86) t0__5a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x87) t0__9a0 ! t0__9t0 ! b_e_t02a0 ! t0__t0_4a0 +LSYM(x88) t0__5a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x89) t0__5a0 ! t0__2t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x90) t0__9a0 ! t0__5t0 ! b_e_shift ! r__r_2t0 +LSYM(x91) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x92) t0__5a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x93) t0__32a0 ! t0__t0ma0 ! b_e_t0 ! t0__3t0 +LSYM(x94) t0__9a0 ! t0__5t0 ! b_e_2t0 ! t0__t0_2a0 +LSYM(x95) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x96) t0__8a0 ! t0__3t0 ! b_e_shift ! r__r_4t0 +LSYM(x97) t0__8a0 ! t0__3t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x98) t0__32a0 ! t0__3t0 ! b_e_t0 ! t0__t0_2a0 +LSYM(x99) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x100) t0__5a0 ! t0__5t0 ! b_e_shift ! r__r_4t0 +LSYM(x101) t0__5a0 ! t0__5t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x102) t0__32a0 ! t0__t0_2a0 ! b_e_t0 ! t0__3t0 +LSYM(x103) t0__5a0 ! t0__5t0 ! b_e_t02a0 ! t0__4t0_a0 +LSYM(x104) t0__3a0 ! t0__4t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x105) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x106) t0__3a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x107) t0__9a0 ! t0__t0_4a0 ! b_e_t02a0 ! t0__8t0_a0 +LSYM(x108) t0__9a0 ! t0__3t0 ! b_e_shift ! r__r_4t0 +LSYM(x109) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x110) t0__9a0 ! t0__3t0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x111) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x112) t0__3a0 ! t0__2t0_a0 ! b_e_t0 ! t0__16t0 +LSYM(x113) t0__9a0 ! t0__4t0_a0 ! b_e_t02a0 ! t0__3t0 +LSYM(x114) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__3t0 +LSYM(x115) t0__9a0 ! t0__2t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x116) t0__3a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__4t0_a0 +LSYM(x117) t0__3a0 ! t0__4t0_a0 ! b_e_t0 ! t0__9t0 +LSYM(x118) t0__3a0 ! t0__4t0_a0 ! b_e_t0a0 ! t0__9t0 +LSYM(x119) t0__3a0 ! t0__4t0_a0 ! b_e_t02a0 ! t0__9t0 +LSYM(x120) t0__5a0 ! t0__3t0 ! b_e_shift ! r__r_8t0 +LSYM(x121) t0__5a0 ! t0__3t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x122) t0__5a0 ! t0__3t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x123) t0__5a0 ! t0__8t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x124) t0__32a0 ! t0__t0ma0 ! b_e_shift ! r__r_4t0 +LSYM(x125) t0__5a0 ! t0__5t0 ! b_e_t0 ! t0__5t0 +LSYM(x126) t0__64a0 ! t0__t0ma0 ! b_e_shift ! r__r_2t0 +LSYM(x127) t0__128a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +LSYM(x128) t0__128a0 ! a1_ne_0_b_l1 ! r__r_t0 ! MILLIRETN +LSYM(x129) t0__128a0 ! a1_ne_0_b_l0 ! t0__t0_a0 ! b_n_ret_t0 +LSYM(x130) t0__64a0 ! t0__t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x131) t0__8a0 ! t0__8t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x132) t0__8a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x133) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x134) t0__8a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x135) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__3t0 +LSYM(x136) t0__8a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x137) t0__8a0 ! t0__2t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x138) t0__8a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x139) t0__8a0 ! t0__2t0_a0 ! b_e_2t0a0 ! t0__4t0_a0 +LSYM(x140) t0__3a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__5t0 +LSYM(x141) t0__8a0 ! t0__2t0_a0 ! b_e_4t0a0 ! t0__2t0_a0 +LSYM(x142) t0__9a0 ! t0__8t0 ! b_e_2t0 ! t0__t0ma0 +LSYM(x143) t0__16a0 ! t0__9t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x144) t0__9a0 ! t0__8t0 ! b_e_shift ! r__r_2t0 +LSYM(x145) t0__9a0 ! t0__8t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x146) t0__9a0 ! t0__8t0_a0 ! b_e_shift ! r__r_2t0 +LSYM(x147) t0__9a0 ! t0__8t0_a0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x148) t0__9a0 ! t0__4t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x149) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x150) t0__9a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x151) t0__9a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__2t0_a0 +LSYM(x152) t0__9a0 ! t0__2t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x153) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x154) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x155) t0__32a0 ! t0__t0ma0 ! b_e_t0 ! t0__5t0 +LSYM(x156) t0__9a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x157) t0__32a0 ! t0__t0ma0 ! b_e_t02a0 ! t0__5t0 +LSYM(x158) t0__16a0 ! t0__5t0 ! b_e_2t0 ! t0__t0ma0 +LSYM(x159) t0__32a0 ! t0__5t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x160) t0__5a0 ! t0__4t0 ! b_e_shift ! r__r_8t0 +LSYM(x161) t0__8a0 ! t0__5t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x162) t0__9a0 ! t0__9t0 ! b_e_shift ! r__r_2t0 +LSYM(x163) t0__9a0 ! t0__9t0 ! b_e_t0 ! t0__2t0_a0 +LSYM(x164) t0__5a0 ! t0__8t0_a0 ! b_e_shift ! r__r_4t0 +LSYM(x165) t0__8a0 ! t0__4t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x166) t0__5a0 ! t0__8t0_a0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x167) t0__5a0 ! t0__8t0_a0 ! b_e_2t0a0 ! t0__2t0_a0 +LSYM(x168) t0__5a0 ! t0__4t0_a0 ! b_e_shift ! r__r_8t0 +LSYM(x169) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x170) t0__32a0 ! t0__t0_2a0 ! b_e_t0 ! t0__5t0 +LSYM(x171) t0__9a0 ! t0__2t0_a0 ! b_e_t0 ! t0__9t0 +LSYM(x172) t0__5a0 ! t0__4t0_a0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x173) t0__9a0 ! t0__2t0_a0 ! b_e_t02a0 ! t0__9t0 +LSYM(x174) t0__32a0 ! t0__t0_2a0 ! b_e_t04a0 ! t0__5t0 +LSYM(x175) t0__8a0 ! t0__2t0_a0 ! b_e_5t0 ! t0__2t0_a0 +LSYM(x176) t0__5a0 ! t0__4t0_a0 ! b_e_8t0 ! t0__t0_a0 +LSYM(x177) t0__5a0 ! t0__4t0_a0 ! b_e_8t0a0 ! t0__t0_a0 +LSYM(x178) t0__5a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__8t0_a0 +LSYM(x179) t0__5a0 ! t0__2t0_a0 ! b_e_2t0a0 ! t0__8t0_a0 +LSYM(x180) t0__9a0 ! t0__5t0 ! b_e_shift ! r__r_4t0 +LSYM(x181) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x182) t0__9a0 ! t0__5t0 ! b_e_2t0 ! t0__2t0_a0 +LSYM(x183) t0__9a0 ! t0__5t0 ! b_e_2t0a0 ! t0__2t0_a0 +LSYM(x184) t0__5a0 ! t0__9t0 ! b_e_4t0 ! t0__t0_a0 +LSYM(x185) t0__9a0 ! t0__4t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x186) t0__32a0 ! t0__t0ma0 ! b_e_2t0 ! t0__3t0 +LSYM(x187) t0__9a0 ! t0__4t0_a0 ! b_e_t02a0 ! t0__5t0 +LSYM(x188) t0__9a0 ! t0__5t0 ! b_e_4t0 ! t0__t0_2a0 +LSYM(x189) t0__5a0 ! t0__4t0_a0 ! b_e_t0 ! t0__9t0 +LSYM(x190) t0__9a0 ! t0__2t0_a0 ! b_e_2t0 ! t0__5t0 +LSYM(x191) t0__64a0 ! t0__3t0 ! b_e_t0 ! t0__t0ma0 +LSYM(x192) t0__8a0 ! t0__3t0 ! b_e_shift ! r__r_8t0 +LSYM(x193) t0__8a0 ! t0__3t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x194) t0__8a0 ! t0__3t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x195) t0__8a0 ! t0__8t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x196) t0__8a0 ! t0__3t0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x197) t0__8a0 ! t0__3t0 ! b_e_4t0a0 ! t0__2t0_a0 +LSYM(x198) t0__64a0 ! t0__t0_2a0 ! b_e_t0 ! t0__3t0 +LSYM(x199) t0__8a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x200) t0__5a0 ! t0__5t0 ! b_e_shift ! r__r_8t0 +LSYM(x201) t0__5a0 ! t0__5t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x202) t0__5a0 ! t0__5t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x203) t0__5a0 ! t0__5t0 ! b_e_2t0a0 ! t0__4t0_a0 +LSYM(x204) t0__8a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__3t0 +LSYM(x205) t0__5a0 ! t0__8t0_a0 ! b_e_t0 ! t0__5t0 +LSYM(x206) t0__64a0 ! t0__t0_4a0 ! b_e_t02a0 ! t0__3t0 +LSYM(x207) t0__8a0 ! t0__2t0_a0 ! b_e_3t0 ! t0__4t0_a0 +LSYM(x208) t0__5a0 ! t0__5t0 ! b_e_8t0 ! t0__t0_a0 +LSYM(x209) t0__5a0 ! t0__5t0 ! b_e_8t0a0 ! t0__t0_a0 +LSYM(x210) t0__5a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__5t0 +LSYM(x211) t0__5a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__5t0 +LSYM(x212) t0__3a0 ! t0__4t0_a0 ! b_e_4t0 ! t0__4t0_a0 +LSYM(x213) t0__3a0 ! t0__4t0_a0 ! b_e_4t0a0 ! t0__4t0_a0 +LSYM(x214) t0__9a0 ! t0__t0_4a0 ! b_e_2t04a0 ! t0__8t0_a0 +LSYM(x215) t0__5a0 ! t0__4t0_a0 ! b_e_5t0 ! t0__2t0_a0 +LSYM(x216) t0__9a0 ! t0__3t0 ! b_e_shift ! r__r_8t0 +LSYM(x217) t0__9a0 ! t0__3t0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x218) t0__9a0 ! t0__3t0 ! b_e_2t0 ! t0__4t0_a0 +LSYM(x219) t0__9a0 ! t0__8t0_a0 ! b_e_t0 ! t0__3t0 +LSYM(x220) t0__3a0 ! t0__9t0 ! b_e_4t0 ! t0__2t0_a0 +LSYM(x221) t0__3a0 ! t0__9t0 ! b_e_4t0a0 ! t0__2t0_a0 +LSYM(x222) t0__9a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__3t0 +LSYM(x223) t0__9a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x224) t0__9a0 ! t0__3t0 ! b_e_8t0 ! t0__t0_a0 +LSYM(x225) t0__9a0 ! t0__5t0 ! b_e_t0 ! t0__5t0 +LSYM(x226) t0__3a0 ! t0__2t0_a0 ! b_e_t02a0 ! t0__32t0 +LSYM(x227) t0__9a0 ! t0__5t0 ! b_e_t02a0 ! t0__5t0 +LSYM(x228) t0__9a0 ! t0__2t0_a0 ! b_e_4t0 ! t0__3t0 +LSYM(x229) t0__9a0 ! t0__2t0_a0 ! b_e_4t0a0 ! t0__3t0 +LSYM(x230) t0__9a0 ! t0__5t0 ! b_e_5t0 ! t0__t0_a0 +LSYM(x231) t0__9a0 ! t0__2t0_a0 ! b_e_3t0 ! t0__4t0_a0 +LSYM(x232) t0__3a0 ! t0__2t0_a0 ! b_e_8t0 ! t0__4t0_a0 +LSYM(x233) t0__3a0 ! t0__2t0_a0 ! b_e_8t0a0 ! t0__4t0_a0 +LSYM(x234) t0__3a0 ! t0__4t0_a0 ! b_e_2t0 ! t0__9t0 +LSYM(x235) t0__3a0 ! t0__4t0_a0 ! b_e_2t0a0 ! t0__9t0 +LSYM(x236) t0__9a0 ! t0__2t0_a0 ! b_e_4t08a0 ! t0__3t0 +LSYM(x237) t0__16a0 ! t0__5t0 ! b_e_3t0 ! t0__t0ma0 +LSYM(x238) t0__3a0 ! t0__4t0_a0 ! b_e_2t04a0 ! t0__9t0 +LSYM(x239) t0__16a0 ! t0__5t0 ! b_e_t0ma0 ! t0__3t0 +LSYM(x240) t0__9a0 ! t0__t0_a0 ! b_e_8t0 ! t0__3t0 +LSYM(x241) t0__9a0 ! t0__t0_a0 ! b_e_8t0a0 ! t0__3t0 +LSYM(x242) t0__5a0 ! t0__3t0 ! b_e_2t0 ! t0__8t0_a0 +LSYM(x243) t0__9a0 ! t0__9t0 ! b_e_t0 ! t0__3t0 +LSYM(x244) t0__5a0 ! t0__3t0 ! b_e_4t0 ! t0__4t0_a0 +LSYM(x245) t0__8a0 ! t0__3t0 ! b_e_5t0 ! t0__2t0_a0 +LSYM(x246) t0__5a0 ! t0__8t0_a0 ! b_e_2t0 ! t0__3t0 +LSYM(x247) t0__5a0 ! t0__8t0_a0 ! b_e_2t0a0 ! t0__3t0 +LSYM(x248) t0__32a0 ! t0__t0ma0 ! b_e_shift ! r__r_8t0 +LSYM(x249) t0__32a0 ! t0__t0ma0 ! b_e_t0 ! t0__8t0_a0 +LSYM(x250) t0__5a0 ! t0__5t0 ! b_e_2t0 ! t0__5t0 +LSYM(x251) t0__5a0 ! t0__5t0 ! b_e_2t0a0 ! t0__5t0 +LSYM(x252) t0__64a0 ! t0__t0ma0 ! b_e_shift ! r__r_4t0 +LSYM(x253) t0__64a0 ! t0__t0ma0 ! b_e_t0 ! t0__4t0_a0 +LSYM(x254) t0__128a0 ! t0__t0ma0 ! b_e_shift ! r__r_2t0 +LSYM(x255) t0__256a0 ! a1_ne_0_b_l0 ! t0__t0ma0 ! b_n_ret_t0 +/*1040 insts before this. */ +LSYM(ret_t0) MILLIRET +LSYM(e_t0) r__r_t0 +LSYM(e_shift) a1_ne_0_b_l2 + a0__256a0 /* a0 <<= 8 *********** */ + MILLIRETN +LSYM(e_t0ma0) a1_ne_0_b_l0 + t0__t0ma0 + MILLIRET + r__r_t0 +LSYM(e_t0a0) a1_ne_0_b_l0 + t0__t0_a0 + MILLIRET + r__r_t0 +LSYM(e_t02a0) a1_ne_0_b_l0 + t0__t0_2a0 + MILLIRET + r__r_t0 +LSYM(e_t04a0) a1_ne_0_b_l0 + t0__t0_4a0 + MILLIRET + r__r_t0 +LSYM(e_2t0) a1_ne_0_b_l1 + r__r_2t0 + MILLIRETN +LSYM(e_2t0a0) a1_ne_0_b_l0 + t0__2t0_a0 + MILLIRET + r__r_t0 +LSYM(e2t04a0) t0__t0_2a0 + a1_ne_0_b_l1 + r__r_2t0 + MILLIRETN +LSYM(e_3t0) a1_ne_0_b_l0 + t0__3t0 + MILLIRET + r__r_t0 +LSYM(e_4t0) a1_ne_0_b_l1 + r__r_4t0 + MILLIRETN +LSYM(e_4t0a0) a1_ne_0_b_l0 + t0__4t0_a0 + MILLIRET + r__r_t0 +LSYM(e4t08a0) t0__t0_2a0 + a1_ne_0_b_l1 + r__r_4t0 + MILLIRETN +LSYM(e_5t0) a1_ne_0_b_l0 + t0__5t0 + MILLIRET + r__r_t0 +LSYM(e_8t0) a1_ne_0_b_l1 + r__r_8t0 + MILLIRETN +LSYM(e_8t0a0) a1_ne_0_b_l0 + t0__8t0_a0 + MILLIRET + r__r_t0 + + .procend + .end +#endif diff --git a/arch/parisc/lib/milli/remI.S b/arch/parisc/lib/milli/remI.S new file mode 100644 index 00000000000..63bc094471e --- /dev/null +++ b/arch/parisc/lib/milli/remI.S @@ -0,0 +1,185 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_remI +/* ROUTINE: $$remI + + DESCRIPTION: + . $$remI returns the remainder of the division of two signed 32-bit + . integers. The sign of the remainder is the same as the sign of + . the dividend. + + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = destroyed + . arg1 = destroyed + . ret1 = remainder + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: DIVIDE BY ZERO + . Changes memory at the following places: NONE + + PERMISSIBLE CONTEXT: + . Unwindable + . Does not create a stack frame + . Is usable for internal or external microcode + + DISCUSSION: + . Calls other millicode routines via mrp: NONE + . Calls other millicode routines: NONE */ + +RDEFINE(tmp,r1) +RDEFINE(retreg,ret1) + + SUBSPA_MILLI + ATTR_MILLI + .proc + .callinfo millicode + .entry +GSYM($$remI) +GSYM($$remoI) + .export $$remI,MILLICODE + .export $$remoI,MILLICODE + ldo -1(arg1),tmp /* is there at most one bit set ? */ + and,<> arg1,tmp,r0 /* if not, don't use power of 2 */ + addi,> 0,arg1,r0 /* if denominator > 0, use power */ + /* of 2 */ + b,n LREF(neg_denom) +LSYM(pow2) + comb,>,n 0,arg0,LREF(neg_num) /* is numerator < 0 ? */ + and arg0,tmp,retreg /* get the result */ + MILLIRETN +LSYM(neg_num) + subi 0,arg0,arg0 /* negate numerator */ + and arg0,tmp,retreg /* get the result */ + subi 0,retreg,retreg /* negate result */ + MILLIRETN +LSYM(neg_denom) + addi,< 0,arg1,r0 /* if arg1 >= 0, it's not power */ + /* of 2 */ + b,n LREF(regular_seq) + sub r0,arg1,tmp /* make denominator positive */ + comb,=,n arg1,tmp,LREF(regular_seq) /* test against 0x80000000 and 0 */ + ldo -1(tmp),retreg /* is there at most one bit set ? */ + and,= tmp,retreg,r0 /* if not, go to regular_seq */ + b,n LREF(regular_seq) + comb,>,n 0,arg0,LREF(neg_num_2) /* if arg0 < 0, negate it */ + and arg0,retreg,retreg + MILLIRETN +LSYM(neg_num_2) + subi 0,arg0,tmp /* test against 0x80000000 */ + and tmp,retreg,retreg + subi 0,retreg,retreg + MILLIRETN +LSYM(regular_seq) + addit,= 0,arg1,0 /* trap if div by zero */ + add,>= 0,arg0,retreg /* move dividend, if retreg < 0, */ + sub 0,retreg,retreg /* make it positive */ + sub 0,arg1, tmp /* clear carry, */ + /* negate the divisor */ + ds 0, tmp,0 /* set V-bit to the comple- */ + /* ment of the divisor sign */ + or 0,0, tmp /* clear tmp */ + add retreg,retreg,retreg /* shift msb bit into carry */ + ds tmp,arg1, tmp /* 1st divide step, if no carry */ + /* out, msb of quotient = 0 */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ +LSYM(t1) + ds tmp,arg1, tmp /* 2nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 3rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 4th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 5th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 6th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 7th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 8th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 9th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 10th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 11th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 12th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 13th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 14th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 15th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 16th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 17th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 18th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 19th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 20th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 21st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 22nd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 23rd divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 24th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 25th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 26th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 27th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 28th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 29th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 30th divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 31st divide step */ + addc retreg,retreg,retreg /* shift retreg with/into carry */ + ds tmp,arg1, tmp /* 32nd divide step, */ + addc retreg,retreg,retreg /* shift last bit into retreg */ + movb,>=,n tmp,retreg,LREF(finish) /* branch if pos. tmp */ + add,< arg1,0,0 /* if arg1 > 0, add arg1 */ + add,tr tmp,arg1,retreg /* for correcting remainder tmp */ + sub tmp,arg1,retreg /* else add absolute value arg1 */ +LSYM(finish) + add,>= arg0,0,0 /* set sign of remainder */ + sub 0,retreg,retreg /* to sign of dividend */ + MILLIRET + nop + .exit + .procend +#ifdef milliext + .origin 0x00000200 +#endif + .end +#endif diff --git a/arch/parisc/lib/milli/remU.S b/arch/parisc/lib/milli/remU.S new file mode 100644 index 00000000000..c0a2d6e247c --- /dev/null +++ b/arch/parisc/lib/milli/remU.S @@ -0,0 +1,148 @@ +/* 32 and 64-bit millicode, original author Hewlett-Packard + adapted for gcc by Paul Bame <bame@debian.org> + and Alan Modra <alan@linuxcare.com.au>. + + Copyright 2001, 2002, 2003 Free Software Foundation, Inc. + + This file is part of GCC and is released under the terms of + of the GNU General Public License as published by the Free Software + Foundation; either version 2, or (at your option) any later version. + See the file COPYING in the top-level GCC source directory for a copy + of the license. */ + +#include "milli.h" + +#ifdef L_remU +/* ROUTINE: $$remU + . Single precision divide for remainder with unsigned binary integers. + . + . The remainder must be dividend-(dividend/divisor)*divisor. + . Divide by zero is trapped. + + INPUT REGISTERS: + . arg0 == dividend + . arg1 == divisor + . mrp == return pc + . sr0 == return space when called externally + + OUTPUT REGISTERS: + . arg0 = undefined + . arg1 = undefined + . ret1 = remainder + + OTHER REGISTERS AFFECTED: + . r1 = undefined + + SIDE EFFECTS: + . Causes a trap under the following conditions: DIVIDE BY ZERO + . Changes memory at the following places: NONE + + PERMISSIBLE CONTEXT: + . Unwindable. + . Does not create a stack frame. + . Suitable for internal or external millicode. + . Assumes the special millicode register conventions. + + DISCUSSION: + . Calls other millicode routines using mrp: NONE + . Calls other millicode routines: NONE */ + + +RDEFINE(temp,r1) +RDEFINE(rmndr,ret1) /* r29 */ + SUBSPA_MILLI + ATTR_MILLI + .export $$remU,millicode + .proc + .callinfo millicode + .entry +GSYM($$remU) + ldo -1(arg1),temp /* is there at most one bit set ? */ + and,= arg1,temp,r0 /* if not, don't use power of 2 */ + b LREF(regular_seq) + addit,= 0,arg1,r0 /* trap on div by zero */ + and arg0,temp,rmndr /* get the result for power of 2 */ + MILLIRETN +LSYM(regular_seq) + comib,>=,n 0,arg1,LREF(special_case) + subi 0,arg1,rmndr /* clear carry, negate the divisor */ + ds r0,rmndr,r0 /* set V-bit to 1 */ + add arg0,arg0,temp /* shift msb bit into carry */ + ds r0,arg1,rmndr /* 1st divide step, if no carry */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 2nd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 3rd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 4th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 5th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 6th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 7th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 8th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 9th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 10th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 11th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 12th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 13th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 14th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 15th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 16th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 17th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 18th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 19th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 20th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 21st divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 22nd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 23rd divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 24th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 25th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 26th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 27th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 28th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 29th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 30th divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 31st divide step */ + addc temp,temp,temp /* shift temp with/into carry */ + ds rmndr,arg1,rmndr /* 32nd divide step, */ + comiclr,<= 0,rmndr,r0 + add rmndr,arg1,rmndr /* correction */ + MILLIRETN + nop + +/* Putting >= on the last DS and deleting COMICLR does not work! */ +LSYM(special_case) + sub,>>= arg0,arg1,rmndr + copy arg0,rmndr + MILLIRETN + nop + .exit + .procend + .end +#endif |