Annotation of OpenXM_contrib2/asir2000/asm/ddM.c, Revision 1.3
1.2 noro 1: /*
2: * Copyright (c) 1994-2000 FUJITSU LABORATORIES LIMITED
3: * All rights reserved.
4: *
5: * FUJITSU LABORATORIES LIMITED ("FLL") hereby grants you a limited,
6: * non-exclusive and royalty-free license to use, copy, modify and
7: * redistribute, solely for non-commercial and non-profit purposes, the
8: * computer program, "Risa/Asir" ("SOFTWARE"), subject to the terms and
9: * conditions of this Agreement. For the avoidance of doubt, you acquire
10: * only a limited right to use the SOFTWARE hereunder, and FLL or any
11: * third party developer retains all rights, including but not limited to
12: * copyrights, in and to the SOFTWARE.
13: *
14: * (1) FLL does not grant you a license in any way for commercial
15: * purposes. You may use the SOFTWARE only for non-commercial and
16: * non-profit purposes only, such as academic, research and internal
17: * business use.
18: * (2) The SOFTWARE is protected by the Copyright Law of Japan and
19: * international copyright treaties. If you make copies of the SOFTWARE,
20: * with or without modification, as permitted hereunder, you shall affix
21: * to all such copies of the SOFTWARE the above copyright notice.
22: * (3) An explicit reference to this SOFTWARE and its copyright owner
23: * shall be made on your publication or presentation in any form of the
24: * results obtained by use of the SOFTWARE.
25: * (4) In the event that you modify the SOFTWARE, you shall notify FLL by
1.3 ! noro 26: * e-mail at risa-admin@sec.flab.fujitsu.co.jp of the detailed specification
1.2 noro 27: * for such modification or the source code of the modified part of the
28: * SOFTWARE.
29: *
30: * THE SOFTWARE IS PROVIDED AS IS WITHOUT ANY WARRANTY OF ANY KIND. FLL
31: * MAKES ABSOLUTELY NO WARRANTIES, EXPRESSED, IMPLIED OR STATUTORY, AND
32: * EXPRESSLY DISCLAIMS ANY IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS
33: * FOR A PARTICULAR PURPOSE OR NONINFRINGEMENT OF THIRD PARTIES'
34: * RIGHTS. NO FLL DEALER, AGENT, EMPLOYEES IS AUTHORIZED TO MAKE ANY
35: * MODIFICATIONS, EXTENSIONS, OR ADDITIONS TO THIS WARRANTY.
36: * UNDER NO CIRCUMSTANCES AND UNDER NO LEGAL THEORY, TORT, CONTRACT,
37: * OR OTHERWISE, SHALL FLL BE LIABLE TO YOU OR ANY OTHER PERSON FOR ANY
38: * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE OR CONSEQUENTIAL
39: * DAMAGES OF ANY CHARACTER, INCLUDING, WITHOUT LIMITATION, DAMAGES
40: * ARISING OUT OF OR RELATING TO THE SOFTWARE OR THIS AGREEMENT, DAMAGES
41: * FOR LOSS OF GOODWILL, WORK STOPPAGE, OR LOSS OF DATA, OR FOR ANY
42: * DAMAGES, EVEN IF FLL SHALL HAVE BEEN INFORMED OF THE POSSIBILITY OF
43: * SUCH DAMAGES, OR FOR ANY CLAIM BY ANY OTHER PARTY. EVEN IF A PART
44: * OF THE SOFTWARE HAS BEEN DEVELOPED BY A THIRD PARTY, THE THIRD PARTY
45: * DEVELOPER SHALL HAVE NO LIABILITY IN CONNECTION WITH THE USE,
46: * PERFORMANCE OR NON-PERFORMANCE OF THE SOFTWARE.
47: *
1.3 ! noro 48: * $OpenXM: OpenXM_contrib2/asir2000/asm/ddM.c,v 1.2 2000/08/21 08:31:17 noro Exp $
1.2 noro 49: */
1.1 noro 50: #include "ca.h"
51: #include "base.h"
52: #include "inline.h"
53:
54: void ksquareummain(int,UM,UM);
55: void kmulummain(int,UM,UM,UM);
56: void c_copyum(UM,int,int *);
57: void copyum(UM,UM);
58: void extractum(UM,int,int,UM);
59: void ksquareum(int,UM,UM);
60: void kmulum(int,UM,UM,UM);
61:
62: /*
63: * mod is declared as 'int', because several xxxum functions contains signed
64: * integer addition/subtraction. So mod should be less than 2^31.
65: */
66:
67: void mulum(mod,p1,p2,pr)
68: int mod;
69: UM p1,p2,pr;
70: {
71: int *pc1,*pcr;
72: int *c1,*c2,*cr;
73: unsigned int mul;
74: int i,j,d1,d2;
75:
76: if ( ( (d1 = DEG(p1)) < 0) || ( (d2 = DEG(p2)) < 0 ) ) {
77: DEG(pr) = -1;
78: return;
79: }
80: c1 = COEF(p1); c2 = COEF(p2); cr = COEF(pr);
81: bzero((char *)cr,(int)((d1+d2+1)*sizeof(int)));
82: for ( i = 0; i <= d2; i++, cr++ )
83: if ( mul = *c2++ )
84: for ( j = 0, pc1 = c1, pcr = cr; j <= d1; j++, pc1++, pcr++ ) {
85: DMAR(*pc1,mul,*pcr,mod,*pcr)
86: }
87: DEG(pr) = d1 + d2;
88: }
89:
90: void mulsum(mod,p,n,pr)
91: int mod,n;
92: UM p,pr;
93: {
94: int *sp,*dp;
95: int i;
96:
97: for ( i = DEG(pr) = DEG(p), sp = COEF(p)+i, dp = COEF(pr)+i;
98: i >= 0; i--, dp--, sp-- ) {
99: DMAR(*sp,n,0,mod,*dp)
100: }
101: }
102:
103: int divum(mod,p1,p2,pq)
104: int mod;
105: UM p1,p2,pq;
106: {
107: int *pc1,*pct;
108: int *c1,*c2,*ct;
109: unsigned int inv,hd,tmp;
110: int i,j, d1,d2,dd;
111:
112: if ( (d1 = DEG(p1)) < (d2 = DEG(p2)) ) {
113: DEG(pq) = -1;
114: return d1;
115: }
116: c1 = COEF(p1); c2 = COEF(p2); dd = d1-d2;
117: if ( ( hd = c2[d2] ) != 1 ) {
118: inv = invm(hd,mod);
119: for ( pc1 = c2 + d2; pc1 >= c2; pc1-- ) {
120: DMAR(*pc1,inv,0,mod,*pc1)
121: }
122: } else
123: inv = 1;
124: for ( i = dd, ct = c1+d1; i >= 0; i-- )
125: if ( tmp = *ct-- ) {
126: tmp = mod - tmp;
127: for ( j = d2-1, pct = ct, pc1 = c2+j; j >= 0; j--, pct--, pc1-- ) {
128: DMAR(*pc1,tmp,*pct,mod,*pct)
129: }
130: }
131: if ( inv != 1 ) {
132: for ( pc1 = c1+d2, pct = c1+d1; pc1 <= pct; pc1++ ) {
133: DMAR(*pc1,inv,0,mod,*pc1)
134: }
135: for ( pc1 = c2, pct = c2+d2, inv = hd; pc1 <= pct; pc1++ ) {
136: DMAR(*pc1,inv,0,mod,*pc1)
137: }
138: }
139: for ( i = d2-1, pc1 = c1+i; i >= 0 && !(*pc1); pc1--, i-- );
140: for ( DEG(pq) = j = dd, pc1 = c1+d1, pct = COEF(pq)+j; j >= 0; j-- )
141: *pct-- = *pc1--;
142: return i;
143: }
144:
145: void diffum(mod,f,fd)
146: int mod;
147: UM f,fd;
148: {
149: int *dp,*sp;
150: int i;
151: UL ltmp;
152:
153: for ( i = DEG(f), dp = COEF(fd)+i-1, sp = COEF(f)+i;
154: i >= 1; i--, dp--, sp-- ) {
155: DMAR(*sp,i,0,mod,*dp)
156: }
157: degum(fd,DEG(f) - 1);
158: }
159:
160: unsigned int pwrm(mod,a,n)
161: int mod,a;
162: int n;
163: {
164: unsigned int s,t;
165:
166: if ( !n )
167: return 1;
168: else if ( n == 1 )
169: return a;
170: else {
171: t = pwrm(mod,a,n/2);
172: DMAR(t,t,0,mod,s)
173: if ( n % 2 ) {
174: DMAR(s,a,0,mod,t)
175: return t;
176: } else
177: return s;
178: }
179: }
180:
181: unsigned int invm(s,mod)
182: unsigned int s;
183: int mod;
184: {
185: unsigned int r,a2,q;
186: unsigned int f1,f2,a1;
187:
188: for ( f1 = s, f2 = mod, a1 = 1, a2 = 0; ; ) {
189: q = f1/f2; r = f1 - f2*q; f1 = f2;
190: if ( !(f2 = r) )
191: break;
192: DMAR(a2,q,0,mod,r)
193: /* r = ( a1 - r + mod ) % mod; */
194: if ( a1 >= r )
195: r = a1 - r;
196: else {
197: r = (mod - r) + a1;
198: }
199: a1 = a2; a2 = r;
200: }
201: /* return( ( a2 >= 0 ? a2 : a2 + mod ) ); */
202: return a2;
203: }
204:
205: unsigned int rem(n,m)
206: N n;
207: unsigned int m;
208: {
209: unsigned int *x;
210: unsigned int t,r;
211: int i;
212:
213: if ( !n )
214: return 0;
215: for ( i = PL(n)-1, x = BD(n)+i, r = 0; i >= 0; i--, x-- ) {
216: #if defined(sparc)
217: r = dsar(m,r,*x);
218: #else
219: DSAB(m,r,*x,t,r)
220: #endif
221: }
222: return r;
223: }
224:
225: #ifndef sparc
226: void addpadic(mod,n,n1,n2)
227: int mod;
228: int n;
229: unsigned int *n1,*n2;
230: {
231: unsigned int carry,tmp;
232: int i;
233:
234: for ( i = 0, carry = 0; i < n; i++ ) {
235: tmp = *n1++ + *n2 + carry;
236: DQR(tmp,mod,carry,*n2++)
237: /*
238: carry = tmp / mod;
239: *n2++ = tmp - ( carry * mod );
240: */
241: }
242: }
243: #endif
244:
245: void mulpadic(mod,n,n1,n2,nr)
246: int mod;
247: int n;
248: unsigned int *n1;
249: unsigned int *n2,*nr;
250: {
251: unsigned int *pn1,*pnr;
252: unsigned int carry,mul;
253: int i,j;
254:
255: bzero((char *)nr,(int)(n*sizeof(int)));
256: for ( j = 0; j < n; j++, n2++, nr++ )
257: if ( mul = *n2 )
258: for ( i = j, carry = 0, pn1 = n1, pnr = nr;
259: i < n; i++, pn1++, pnr++ ) {
260: carry += *pnr;
261: DMAB(mod,*pn1,mul,carry,carry,*pnr)
262: }
263: }
264:
265: extern up_kara_mag;
266:
267: void kmulum(mod,n1,n2,nr)
268: UM n1,n2,nr;
269: {
270: UM n,t,s,m,carry;
271: int d,d1,d2,len,i,l;
272: unsigned int *r,*r0;
273:
274: if ( !n1 || !n2 ) {
275: nr->d = -1; return;
276: }
277: d1 = DEG(n1)+1; d2 = DEG(n2)+1;
278: if ( (d1 < up_kara_mag) || (d2 < up_kara_mag) ) {
279: mulum(mod,n1,n2,nr); return;
280: }
281: if ( d1 < d2 ) {
282: n = n1; n1 = n2; n2 = n;
283: d = d1; d1 = d2; d2 = d;
284: }
285: if ( d2 > (d1+1)/2 ) {
286: kmulummain(mod,n1,n2,nr); return;
287: }
288: d = (d1/d2)+((d1%d2)!=0);
289: len = (d+1)*d2;
290: r0 = (unsigned int *)ALLOCA(len*sizeof(int));
291: bzero((char *)r0,len*sizeof(int));
292: m = W_UMALLOC(d2+1);
293: carry = W_UMALLOC(d2+1);
294: t = W_UMALLOC(d1+d2+1);
295: s = W_UMALLOC(d1+d2+1);
296: for ( carry = 0, i = 0, r = r0; i < d; i++, r += d2 ) {
297: extractum(n1,i*d2,d2,m);
298: if ( m ) {
299: kmulum(mod,m,n2,t);
300: addum(mod,t,carry,s);
301: c_copyum(s,d2,r);
302: extractum(s,d2,d2,carry);
303: } else {
304: c_copyum(carry,d2,r);
305: carry = 0;
306: }
307: }
308: c_copyum(carry,d2,r);
309: for ( l = len - 1; !r0[l]; l-- );
310: l++;
311: DEG(nr) = l-1;
312: bcopy((char *)r0,(char *)COEF(nr),l*sizeof(int));
313: }
314:
315: void ksquareum(mod,n1,nr)
316: int mod;
317: UM n1,nr;
318: {
319: int d1;
320:
321: if ( !n1 ) {
322: nr->d = -1; return;
323: }
324: d1 = DEG(n1)+1;
325: if ( (d1 < up_kara_mag) ) {
326: pwrum(mod,n1,2,nr); return;
327: }
328: ksquareummain(mod,n1,nr);
329: }
330:
331: void extractum(n,index,len,nr)
332: UM n;
333: int index,len;
334: UM nr;
335: {
336: int *m;
337: int l;
338:
339: if ( !n ) {
340: nr->d = -1; return;
341: }
342: m = COEF(n)+index;
343: if ( (l = (DEG(n)+1)-index) >= len ) {
344: for ( l = len - 1; (l >= 0) && !m[l]; l-- );
345: l++;
346: }
347: if ( l <= 0 ) {
348: nr->d = -1; return;
349: } else {
350: DEG(nr) = l-1;
351: bcopy((char *)m,(char *)COEF(nr),l*sizeof(Q));
352: }
353: }
354:
355: void copyum(n1,n2)
356: UM n1,n2;
357: {
358: n2->d = n1->d;
359: bcopy((char *)n1->c,(char *)n2->c,(n1->d+1)*sizeof(int));
360: }
361:
362: void c_copyum(n,len,p)
363: UM n;
364: int len;
365: int *p;
366: {
367: if ( n )
368: bcopy((char *)COEF(n),(char *)p,MIN((DEG(n)+1),len)*sizeof(int));
369: }
370:
371: void kmulummain(mod,n1,n2,nr)
372: int mod;
373: UM n1,n2,nr;
374: {
375: int d1,d2,h,len;
376: UM n1lo,n1hi,n2lo,n2hi,hi,lo,mid1,mid2,mid,s1,s2,t1,t2;
377:
378: d1 = DEG(n1)+1; d2 = DEG(n2)+1; h = (d1+1)/2;
379: n1lo = W_UMALLOC(d1+1); n1hi = W_UMALLOC(d1+1);
380: n2lo = W_UMALLOC(d2+1); n2hi = W_UMALLOC(d2+1);
381: lo = W_UMALLOC(d1+d2+1); hi = W_UMALLOC(d1+d2+1);
382: mid1 = W_UMALLOC(d1+d2+1); mid2 = W_UMALLOC(d1+d2+1);
383: mid = W_UMALLOC(d1+d2+1);
384: s1 = W_UMALLOC(d1+d2+1); s2 = W_UMALLOC(d1+d2+1);
385: extractum(n1,0,h,n1lo); extractum(n1,h,d1-h,n1hi);
386: extractum(n2,0,h,n2lo); extractum(n2,h,d2-h,n2hi);
387: kmulum(mod,n1hi,n2hi,hi); kmulum(mod,n1lo,n2lo,lo);
388: len = DEG(hi)+1+2*h; t1 = W_UMALLOC(len-1); DEG(t1) = len-1;
389: bzero((char *)COEF(t1),len*sizeof(int));
390: if ( lo )
391: bcopy((char *)COEF(lo),(char *)COEF(t1),(DEG(lo)+1)*sizeof(int));
392: if ( hi )
393: bcopy((char *)COEF(hi),(char *)(COEF(t1)+2*h),(DEG(hi)+1)*sizeof(int));
394:
395: addum(mod,hi,lo,mid1);
396: subum(mod,n1hi,n1lo,s1); subum(mod,n2lo,n2hi,s2); kmulum(mod,s1,s2,mid2);
397: addum(mod,mid1,mid2,mid);
398: if ( mid ) {
399: len = DEG(mid)+1+h; t2 = W_UMALLOC(len-1); DEG(t2) = len-1;
400: bzero((char *)COEF(t2),len*sizeof(int));
401: bcopy((char *)COEF(mid),(char *)(COEF(t2)+h),(DEG(mid)+1)*sizeof(int));
402: addum(mod,t1,t2,nr);
403: } else
404: copyum(t1,nr);
405: }
406:
407: void ksquareummain(mod,n1,nr)
408: int mod;
409: UM n1,nr;
410: {
411: int d1,h,len;
412: UM n1lo,n1hi,hi,lo,mid1,mid2,mid,s1,t1,t2;
413:
414: d1 = DEG(n1)+1; h = (d1+1)/2;
415: n1lo = W_UMALLOC(d1+1); n1hi = W_UMALLOC(d1+1);
416: lo = W_UMALLOC(2*d1+1); hi = W_UMALLOC(2*d1+1);
417: mid1 = W_UMALLOC(2*d1+1); mid2 = W_UMALLOC(2*d1+1);
418: mid = W_UMALLOC(2*d1+1);
419: s1 = W_UMALLOC(2*d1+1);
420: extractum(n1,0,h,n1lo); extractum(n1,h,d1-h,n1hi);
421: ksquareum(mod,n1hi,hi); ksquareum(mod,n1lo,lo);
422: len = DEG(hi)+1+2*h; t1 = W_UMALLOC(len-1); DEG(t1) = len-1;
423: bzero((char *)COEF(t1),len*sizeof(int));
424: if ( lo )
425: bcopy((char *)COEF(lo),(char *)COEF(t1),(DEG(lo)+1)*sizeof(int));
426: if ( hi )
427: bcopy((char *)COEF(hi),(char *)(COEF(t1)+2*h),(DEG(hi)+1)*sizeof(int));
428:
429: addum(mod,hi,lo,mid1);
430: subum(mod,n1hi,n1lo,s1); ksquareum(mod,s1,mid2);
431: subum(mod,mid1,mid2,mid);
432: if ( mid ) {
433: len = DEG(mid)+1+h; t2 = W_UMALLOC(len-1); DEG(t2) = len-1;
434: bzero((char *)COEF(t2),len*sizeof(int));
435: bcopy((char *)COEF(mid),(char *)(COEF(t2)+h),(DEG(mid)+1)*sizeof(int));
436: addum(mod,t1,t2,nr);
437: } else
438: copyum(t1,nr);
439: }
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>