xref: /netbsd-src/usr.bin/sort/fsort.c (revision de1dfb1250df962f1ff3a011772cf58e605aed11)
1 /*	$NetBSD: fsort.c,v 1.30 2004/02/15 11:54:17 jdolecek Exp $	*/
2 
3 /*-
4  * Copyright (c) 2000-2003 The NetBSD Foundation, Inc.
5  * All rights reserved.
6  *
7  * This code is derived from software contributed to The NetBSD Foundation
8  * by Ben Harris and Jaromir Dolecek.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. All advertising materials mentioning features or use of this software
19  *    must display the following acknowledgement:
20  *        This product includes software developed by the NetBSD
21  *        Foundation, Inc. and its contributors.
22  * 4. Neither the name of The NetBSD Foundation nor the names of its
23  *    contributors may be used to endorse or promote products derived
24  *    from this software without specific prior written permission.
25  *
26  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
27  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
28  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
29  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
30  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
31  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
32  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
33  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
34  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
35  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
36  * POSSIBILITY OF SUCH DAMAGE.
37  */
38 
39 /*-
40  * Copyright (c) 1993
41  *	The Regents of the University of California.  All rights reserved.
42  *
43  * This code is derived from software contributed to Berkeley by
44  * Peter McIlroy.
45  *
46  * Redistribution and use in source and binary forms, with or without
47  * modification, are permitted provided that the following conditions
48  * are met:
49  * 1. Redistributions of source code must retain the above copyright
50  *    notice, this list of conditions and the following disclaimer.
51  * 2. Redistributions in binary form must reproduce the above copyright
52  *    notice, this list of conditions and the following disclaimer in the
53  *    documentation and/or other materials provided with the distribution.
54  * 3. Neither the name of the University nor the names of its contributors
55  *    may be used to endorse or promote products derived from this software
56  *    without specific prior written permission.
57  *
58  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
59  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
60  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
61  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
62  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
63  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
64  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
65  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
66  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
67  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
68  * SUCH DAMAGE.
69  */
70 
71 /*
72  * Read in the next bin.  If it fits in one segment sort it;
73  * otherwise refine it by segment deeper by one character,
74  * and try again on smaller bins.  Sort the final bin at this level
75  * of recursion to keep the head of fstack at 0.
76  * After PANIC passes, abort to merge sort.
77  */
78 #include "sort.h"
79 #include "fsort.h"
80 
81 #ifndef lint
82 __RCSID("$NetBSD: fsort.c,v 1.30 2004/02/15 11:54:17 jdolecek Exp $");
83 __SCCSID("@(#)fsort.c	8.1 (Berkeley) 6/6/93");
84 #endif /* not lint */
85 
86 #include <stdlib.h>
87 #include <string.h>
88 
89 static const u_char **keylist = 0;
90 u_char *buffer = 0, *linebuf = 0;
91 size_t bufsize = DEFBUFSIZE;
92 size_t linebuf_size;
93 #define FSORTMAX 4
94 int PANIC = FSORTMAX;
95 
96 struct tempfile fstack[MAXFCT];
97 #define MSTART		(MAXFCT - MERGE_FNUM)
98 #define	CHECKFSTACK(n)					\
99 	if (n >= MAXFCT)				\
100 		errx(2, "fstack: too many temporary files; use -H or sort in pieces")
101 
102 #define SALIGN(n) ((n+sizeof(length_t)-1) & ~(sizeof(length_t)-1))
103 
104 void
105 fsort(binno, depth, top, filelist, nfiles, outfp, ftbl)
106 	int binno, depth, top;
107 	struct filelist *filelist;
108 	int nfiles;
109 	FILE *outfp;
110 	struct field *ftbl;
111 {
112 	const u_char **keypos;
113 	u_char *bufend;
114 	u_char *weights;
115 	int ntfiles, mfct = 0, total, i, maxb, lastb, panic = 0;
116 	int c, nelem, base;
117 	long sizes [NBINS + 1];
118 	get_func_t get;
119 	struct recheader *crec;
120 	struct field tfield[2];
121 	FILE *prevfp, *tailfp[FSORTMAX + 1];
122 	u_char *nbuffer;
123 
124 	memset(tailfp, 0, sizeof(tailfp));
125 	prevfp = outfp;
126 	memset(tfield, 0, sizeof(tfield));
127 	if (ftbl[0].flags & R)
128 		tfield[0].weights = Rascii;
129 	else
130 		tfield[0].weights = ascii;
131 	tfield[0].icol.num = 1;
132 	weights = ftbl[0].weights;
133 	if (!buffer) {
134 		buffer = malloc(bufsize);
135 		keylist = malloc(MAXNUM * sizeof(u_char *));
136 		memset(keylist, 0, MAXNUM * sizeof(u_char *));
137 		if (!SINGL_FLD) {
138 			linebuf_size = DEFLLEN;
139 			if ((linebuf = malloc(linebuf_size)) == NULL)
140 				errx(2, "cannot allocate memory");
141 		}
142 	}
143 	bufend = buffer + bufsize;
144 	if (binno >= 0) {
145 		base = top + nfiles;
146 		get = getnext;
147 	} else {
148 		base = 0;
149 		if (SINGL_FLD)
150 			get = makeline;
151 		else
152 			get = makekey;
153 	}
154 	for (;;) {
155 		memset(sizes, 0, sizeof(sizes));
156 		c = ntfiles = 0;
157 		if (binno == weights[REC_D] &&
158 		    !(SINGL_FLD && ftbl[0].flags & F)) {	/* pop */
159 			rd_append(weights[REC_D], top,
160 			    nfiles, prevfp, buffer, bufend);
161 			break;
162 		} else if (binno == weights[REC_D]) {
163 			depth = 0;		/* start over on flat weights */
164 			ftbl = tfield;
165 			weights = ftbl[0].weights;
166 		}
167 		while (c != EOF) {
168 			keypos = keylist;
169 			nelem = 0;
170 			crec = (RECHEADER *) buffer;
171 
172 		   do_read:
173 			while ((c = get(binno, top, filelist, nfiles, crec,
174 			    bufend, ftbl)) == 0) {
175 				*keypos++ = crec->data + depth;
176 				if (++nelem == MAXNUM) {
177 					c = BUFFEND;
178 					break;
179 				}
180 				crec =(RECHEADER *)((char *) crec +
181 				    SALIGN(crec->length) + sizeof(TRECHEADER));
182 			}
183 
184 			if (c == BUFFEND && nelem < MAXNUM
185 			    && bufsize < MAXBUFSIZE) {
186 				const u_char **keyp;
187 				u_char *oldb = buffer;
188 
189 				/* buffer was too small for data, allocate
190 				 * bigger buffer */
191 				nbuffer = realloc(buffer, bufsize * 2);
192 				if (!nbuffer) {
193 					err(2, "failed to realloc buffer to %lu bytes",
194 						(unsigned long) bufsize * 2);
195 				}
196 				buffer = nbuffer;
197 				bufsize *= 2;
198 				bufend = buffer + bufsize;
199 
200 				/* patch up keylist[] */
201 				for (keyp = &keypos[-1]; keyp >= keylist; keyp--)
202 					*keyp = buffer + (*keyp - oldb);
203 
204 				crec = (RECHEADER *) (buffer + ((u_char *)crec - oldb));
205 				goto do_read;
206 			}
207 
208 			if (c != BUFFEND && !ntfiles && !mfct) {
209 				/* do not push */
210 				continue;
211 			}
212 
213 			/* push */
214 			if (panic >= PANIC) {
215 				fstack[MSTART + mfct].fp = ftmp();
216 				if ((stable_sort)
217 					? sradixsort(keylist, nelem,
218 						weights, REC_D)
219 					: radixsort(keylist, nelem,
220 						weights, REC_D) )
221 					err(2, NULL);
222 				append(keylist, nelem, depth,
223 				    fstack[MSTART + mfct].fp, putrec,
224 				    ftbl);
225 				mfct++;
226 				/* reduce number of open files */
227 				if (mfct == MERGE_FNUM ||(c == EOF && ntfiles)) {
228 					/*
229 					 * Only copy extra incomplete crec
230 					 * data if there are any.
231 					 */
232 					int nodata = (bufend >= (u_char *)crec
233 					    && bufend <= crec->data);
234 					size_t sz=0;
235 					u_char *tmpbuf=NULL;
236 
237 					if (!nodata) {
238 						sz = bufend - crec->data;
239 						tmpbuf = malloc(sz);
240 						memmove(tmpbuf, crec->data, sz);
241 					}
242 
243 					CHECKFSTACK(base + ntfiles);
244 					fstack[base + ntfiles].fp = ftmp();
245 					fmerge(0, MSTART, filelist,
246 					    mfct, geteasy,
247 					    fstack[base + ntfiles].fp,
248 					    putrec, ftbl);
249 					ntfiles++;
250 					mfct = 0;
251 
252 					if (!nodata) {
253 						memmove(crec->data, tmpbuf, sz);
254 						free(tmpbuf);
255 					}
256 				}
257 			} else {
258 				CHECKFSTACK(base + ntfiles);
259 				fstack[base + ntfiles].fp = ftmp();
260 				onepass(keylist, depth, nelem, sizes,
261 				    weights, fstack[base + ntfiles].fp);
262 				ntfiles++;
263 			}
264 		}
265 		if (!ntfiles && !mfct) {	/* everything in memory--pop */
266 			if (nelem > 1
267 			    && ((stable_sort)
268 				? sradixsort(keylist, nelem, weights, REC_D)
269 				: radixsort(keylist, nelem, weights, REC_D) ))
270 				err(2, NULL);
271 			if (nelem > 0)
272 			  append(keylist, nelem, depth, outfp, putline, ftbl);
273 			break;					/* pop */
274 		}
275 		if (panic >= PANIC) {
276 			if (!ntfiles)
277 				fmerge(0, MSTART, filelist, mfct, geteasy,
278 				    outfp, putline, ftbl);
279 			else
280 				fmerge(0, base, filelist, ntfiles, geteasy,
281 				    outfp, putline, ftbl);
282 			break;
283 
284 		}
285 		total = maxb = lastb = 0;	/* find if one bin dominates */
286 		for (i = 0; i < NBINS; i++)
287 			if (sizes[i]) {
288 				if (sizes[i] > sizes[maxb])
289 					maxb = i;
290 				lastb = i;
291 				total += sizes[i];
292 			}
293 		if (sizes[maxb] < max((total / 2) , BUFSIZE))
294 			maxb = lastb;	/* otherwise pop after last bin */
295 		fstack[base].lastb = lastb;
296 		fstack[base].maxb = maxb;
297 
298 		/* start refining next level. */
299 		getnext(-1, base, NULL, ntfiles, crec, bufend, 0); /* rewind */
300 		for (i = 0; i < maxb; i++) {
301 			if (!sizes[i])	/* bin empty; step ahead file offset */
302 				getnext(i, base, NULL,ntfiles, crec, bufend, 0);
303 			else
304 				fsort(i, depth + 1, base, filelist, ntfiles,
305 					outfp, ftbl);
306 		}
307 
308 		get = getnext;
309 
310 		if (lastb != maxb) {
311 			if (prevfp != outfp)
312 				tailfp[panic] = prevfp;
313 			prevfp = ftmp();
314 			for (i = maxb + 1; i <= lastb; i++)
315 				if (!sizes[i])
316 					getnext(i, base, NULL, ntfiles, crec,
317 					    bufend,0);
318 				else
319 					fsort(i, depth + 1, base, filelist,
320 					    ntfiles, prevfp, ftbl);
321 		}
322 
323 		/* sort biggest (or last) bin at this level */
324 		depth++;
325 		panic++;
326 		binno = maxb;
327 		top = base;
328 		nfiles = ntfiles;		/* so overwrite them */
329 	}
330 	if (prevfp != outfp) {
331 		concat(outfp, prevfp);
332 		fclose(prevfp);
333 	}
334 	for (i = panic; i >= 0; --i)
335 		if (tailfp[i]) {
336 			concat(outfp, tailfp[i]);
337 			fclose(tailfp[i]);
338 		}
339 
340 	/* If on top level, free our structures */
341 	if (depth == 0) {
342 		free(keylist), keylist = NULL;
343 		free(buffer), buffer = NULL;
344 	}
345 }
346 
347 /*
348  * This is one pass of radix exchange, dumping the bins to disk.
349  */
350 #define swap(a, b, t) t = a, a = b, b = t
351 void
352 onepass(a, depth, n, sizes, tr, fp)
353 	const u_char **a;
354 	int depth;
355 	long n, sizes[];
356 	u_char *tr;
357 	FILE *fp;
358 {
359 	size_t tsizes[NBINS + 1];
360 	const u_char **bin[257], ***bp, ***bpmax, **top[256], ***tp;
361 	static int histo[256];
362 	int *hp;
363 	int c;
364 	const u_char **an, *t, **aj;
365 	const u_char **ak, *r;
366 
367 	memset(tsizes, 0, sizeof(tsizes));
368 	depth += sizeof(TRECHEADER);
369 	an = &a[n];
370 	for (ak = a; ak < an; ak++) {
371 		histo[c = tr[**ak]]++;
372 		tsizes[c] += ((const RECHEADER *) (*ak -= depth))->length;
373 	}
374 
375 	bin[0] = a;
376 	bpmax = bin + 256;
377 	tp = top, hp = histo;
378 	for (bp = bin; bp < bpmax; bp++) {
379 		*tp++ = *(bp + 1) = *bp + (c = *hp);
380 		*hp++ = 0;
381 		if (c <= 1)
382 			continue;
383 	}
384 	for (aj = a; aj < an; *aj = r, aj = bin[c + 1])
385 		for (r = *aj; aj < (ak = --top[c = tr[r[depth]]]) ;)
386 			swap(*ak, r, t);
387 
388 	for (ak = a, c = 0; c < 256; c++) {
389 		an = bin[c + 1];
390 		n = an - ak;
391 		tsizes[c] += n * sizeof(TRECHEADER);
392 		/* tell getnext how many elements in this bin, this segment. */
393 		EWRITE(&tsizes[c], sizeof(size_t), 1, fp);
394 		sizes[c] += tsizes[c];
395 		for (; ak < an; ++ak)
396 			putrec((const RECHEADER *) *ak, fp);
397 	}
398 }
399