xref: /csrg-svn/sys/ufs/lfs/lfs_alloc.c (revision 6531)
1 /*	alloc.c	4.1	82/03/25	*/
2 
3 /* merged into kernel:	@(#)lfs_alloc.c 2.3 04/11/82 */
4 
5 /* last monet version:	alloc.c	4.8	81/03/08	*/
6 
7 #include "../h/param.h"
8 #include "../h/systm.h"
9 #include "../h/mount.h"
10 #include "../h/fs.h"
11 #include "../h/conf.h"
12 #include "../h/buf.h"
13 #include "../h/inode.h"
14 #include "../h/ndir.h"
15 #include "../h/user.h"
16 
17 extern u_long		hashalloc();
18 extern ino_t		ialloccg();
19 extern daddr_t		alloccg();
20 extern daddr_t		alloccgblk();
21 extern daddr_t		fragextend();
22 extern daddr_t		blkpref();
23 extern daddr_t		mapsearch();
24 extern int		inside[], around[];
25 extern unsigned char	*fragtbl[];
26 
27 /*
28  * Allocate a block in the file system.
29  *
30  * The size of the requested block is given, which must be some
31  * multiple of fs_fsize and <= fs_bsize.
32  * A preference may be optionally specified. If a preference is given
33  * the following hierarchy is used to allocate a block:
34  *   1) allocate the requested block.
35  *   2) allocate a rotationally optimal block in the same cylinder.
36  *   3) allocate a block in the same cylinder group.
37  *   4) quadradically rehash into other cylinder groups, until an
38  *      available block is located.
39  * If no block preference is given the following heirarchy is used
40  * to allocate a block:
41  *   1) allocate a block in the cylinder group that contains the
42  *      inode for the file.
43  *   2) quadradically rehash into other cylinder groups, until an
44  *      available block is located.
45  */
46 struct buf *
47 alloc(ip, bpref, size)
48 	register struct inode *ip;
49 	daddr_t bpref;
50 	int size;
51 {
52 	daddr_t bno;
53 	register struct fs *fs;
54 	register struct buf *bp;
55 	int cg;
56 
57 	fs = ip->i_fs;
58 	if ((unsigned)size > fs->fs_bsize || fragoff(fs, size) != 0)
59 		panic("alloc: bad size");
60 	if (size == fs->fs_bsize && fs->fs_cstotal.cs_nbfree == 0)
61 		goto nospace;
62 	if (u.u_uid != 0 &&
63 	    fs->fs_cstotal.cs_nbfree * fs->fs_frag + fs->fs_cstotal.cs_nffree <
64 	      fs->fs_dsize * fs->fs_minfree / 100)
65 		goto nospace;
66 	if (bpref >= fs->fs_size)
67 		bpref = 0;
68 	if (bpref == 0)
69 		cg = itog(fs, ip->i_number);
70 	else
71 		cg = dtog(fs, bpref);
72 	bno = (daddr_t)hashalloc(ip, cg, (long)bpref, size, alloccg);
73 	if (bno == 0)
74 		goto nospace;
75 	bp = getblk(ip->i_dev, fsbtodb(fs, bno), size);
76 	clrbuf(bp);
77 	return (bp);
78 nospace:
79 	fserr(fs, "file system full");
80 	uprintf("\n%s: write failed, file system is full\n", fs->fs_fsmnt);
81 	u.u_error = ENOSPC;
82 	return (NULL);
83 }
84 
85 /*
86  * Reallocate a fragment to a bigger size
87  *
88  * The number and size of the old block is given, and a preference
89  * and new size is also specified. The allocator attempts to extend
90  * the original block. Failing that, the regular block allocator is
91  * invoked to get an appropriate block.
92  */
93 struct buf *
94 realloccg(ip, bprev, bpref, osize, nsize)
95 	register struct inode *ip;
96 	daddr_t bprev, bpref;
97 	int osize, nsize;
98 {
99 	daddr_t bno;
100 	register struct fs *fs;
101 	register struct buf *bp, *obp;
102 	caddr_t cp;
103 	int cg;
104 
105 	fs = ip->i_fs;
106 	if ((unsigned)osize > fs->fs_bsize || fragoff(fs, osize) != 0 ||
107 	    (unsigned)nsize > fs->fs_bsize || fragoff(fs, nsize) != 0)
108 		panic("realloccg: bad size");
109 	if (u.u_uid != 0 &&
110 	    fs->fs_cstotal.cs_nbfree * fs->fs_frag + fs->fs_cstotal.cs_nffree <
111 	      fs->fs_dsize * fs->fs_minfree / 100)
112 		goto nospace;
113 	if (bprev == 0)
114 		panic("realloccg: bad bprev");
115 	cg = dtog(fs, bprev);
116 	bno = fragextend(ip, cg, (long)bprev, osize, nsize);
117 	if (bno != 0) {
118 		bp = bread(ip->i_dev, fsbtodb(fs, bno), osize);
119 		if (bp->b_flags & B_ERROR) {
120 			brelse(bp);
121 			return (NULL);
122 		}
123 		bp->b_bcount = nsize;
124 		blkclr(bp->b_un.b_addr + osize, nsize - osize);
125 		return (bp);
126 	}
127 	if (bpref >= fs->fs_size)
128 		bpref = 0;
129 	bno = (daddr_t)hashalloc(ip, cg, (long)bpref, nsize, alloccg);
130 	if (bno != 0) {
131 		/*
132 		 * make a new copy
133 		 */
134 		obp = bread(ip->i_dev, fsbtodb(fs, bprev), osize);
135 		if (obp->b_flags & B_ERROR) {
136 			brelse(obp);
137 			return (NULL);
138 		}
139 		bp = getblk(ip->i_dev, fsbtodb(fs, bno), nsize);
140 		cp = bp->b_un.b_addr;
141 		bp->b_un.b_addr = obp->b_un.b_addr;
142 		obp->b_un.b_addr = cp;
143 		obp->b_flags |= B_INVAL;
144 		brelse(obp);
145 		fre(ip, bprev, (off_t)osize);
146 		blkclr(bp->b_un.b_addr + osize, nsize - osize);
147 		return (bp);
148 	}
149 nospace:
150 	/*
151 	 * no space available
152 	 */
153 	fserr(fs, "file system full");
154 	uprintf("\n%s: write failed, file system is full\n", fs->fs_fsmnt);
155 	u.u_error = ENOSPC;
156 	return (NULL);
157 }
158 
159 /*
160  * Allocate an inode in the file system.
161  *
162  * A preference may be optionally specified. If a preference is given
163  * the following hierarchy is used to allocate an inode:
164  *   1) allocate the requested inode.
165  *   2) allocate an inode in the same cylinder group.
166  *   3) quadradically rehash into other cylinder groups, until an
167  *      available inode is located.
168  * If no inode preference is given the following heirarchy is used
169  * to allocate an inode:
170  *   1) allocate an inode in cylinder group 0.
171  *   2) quadradically rehash into other cylinder groups, until an
172  *      available inode is located.
173  */
174 struct inode *
175 ialloc(pip, ipref, mode)
176 	register struct inode *pip;
177 	ino_t ipref;
178 	int mode;
179 {
180 	ino_t ino;
181 	register struct fs *fs;
182 	register struct inode *ip;
183 	int cg;
184 
185 	fs = pip->i_fs;
186 	if (fs->fs_cstotal.cs_nifree == 0)
187 		goto noinodes;
188 	if (ipref >= fs->fs_ncg * fs->fs_ipg)
189 		ipref = 0;
190 	cg = itog(fs, ipref);
191 	ino = (ino_t)hashalloc(pip, cg, (long)ipref, mode, ialloccg);
192 	if (ino == 0)
193 		goto noinodes;
194 	ip = iget(pip->i_dev, pip->i_fs, ino);
195 	if (ip == NULL) {
196 		ifree(ip, ino, 0);
197 		return (NULL);
198 	}
199 	if (ip->i_mode)
200 		panic("ialloc: dup alloc");
201 	return (ip);
202 noinodes:
203 	fserr(fs, "out of inodes");
204 	uprintf("\n%s: create/symlink failed, no inodes free\n", fs->fs_fsmnt);
205 	u.u_error = ENOSPC;
206 	return (NULL);
207 }
208 
209 /*
210  * Find a cylinder to place a directory.
211  *
212  * The policy implemented by this algorithm is to select from
213  * among those cylinder groups with above the average number of
214  * free inodes, the one with the smallest number of directories.
215  */
216 dirpref(fs)
217 	register struct fs *fs;
218 {
219 	int cg, minndir, mincg, avgifree;
220 
221 	avgifree = fs->fs_cstotal.cs_nifree / fs->fs_ncg;
222 	minndir = fs->fs_ipg;
223 	mincg = 0;
224 	for (cg = 0; cg < fs->fs_ncg; cg++)
225 		if (fs->fs_cs(fs, cg).cs_ndir < minndir &&
226 		    fs->fs_cs(fs, cg).cs_nifree >= avgifree) {
227 			mincg = cg;
228 			minndir = fs->fs_cs(fs, cg).cs_ndir;
229 		}
230 	return (fs->fs_ipg * mincg);
231 }
232 
233 /*
234  * Select a cylinder to place a large block of data.
235  *
236  * The policy implemented by this algorithm is to maintain a
237  * rotor that sweeps the cylinder groups. When a block is
238  * needed, the rotor is advanced until a cylinder group with
239  * greater than the average number of free blocks is found.
240  */
241 daddr_t
242 blkpref(fs)
243 	register struct fs *fs;
244 {
245 	int cg, avgbfree;
246 
247 	avgbfree = fs->fs_cstotal.cs_nbfree / fs->fs_ncg;
248 	for (cg = fs->fs_cgrotor + 1; cg < fs->fs_ncg; cg++)
249 		if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) {
250 			fs->fs_cgrotor = cg;
251 			return (fs->fs_fpg * cg + fs->fs_frag);
252 		}
253 	for (cg = 0; cg <= fs->fs_cgrotor; cg++)
254 		if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) {
255 			fs->fs_cgrotor = cg;
256 			return (fs->fs_fpg * cg + fs->fs_frag);
257 		}
258 	return (NULL);
259 }
260 
261 /*
262  * Implement the cylinder overflow algorithm.
263  *
264  * The policy implemented by this algorithm is:
265  *   1) allocate the block in its requested cylinder group.
266  *   2) quadradically rehash on the cylinder group number.
267  *   3) brute force search for a free block.
268  */
269 /*VARARGS5*/
270 u_long
271 hashalloc(ip, cg, pref, size, allocator)
272 	struct inode *ip;
273 	int cg;
274 	long pref;
275 	int size;	/* size for data blocks, mode for inodes */
276 	u_long (*allocator)();
277 {
278 	register struct fs *fs;
279 	long result;
280 	int i, icg = cg;
281 
282 	fs = ip->i_fs;
283 	/*
284 	 * 1: preferred cylinder group
285 	 */
286 	result = (*allocator)(ip, cg, pref, size);
287 	if (result)
288 		return (result);
289 	/*
290 	 * 2: quadratic rehash
291 	 */
292 	for (i = 1; i < fs->fs_ncg; i *= 2) {
293 		cg += i;
294 		if (cg >= fs->fs_ncg)
295 			cg -= fs->fs_ncg;
296 		result = (*allocator)(ip, cg, 0, size);
297 		if (result)
298 			return (result);
299 	}
300 	/*
301 	 * 3: brute force search
302 	 */
303 	cg = icg;
304 	for (i = 0; i < fs->fs_ncg; i++) {
305 		result = (*allocator)(ip, cg, 0, size);
306 		if (result)
307 			return (result);
308 		cg++;
309 		if (cg == fs->fs_ncg)
310 			cg = 0;
311 	}
312 	return (NULL);
313 }
314 
315 /*
316  * Determine whether a fragment can be extended.
317  *
318  * Check to see if the necessary fragments are available, and
319  * if they are, allocate them.
320  */
321 daddr_t
322 fragextend(ip, cg, bprev, osize, nsize)
323 	struct inode *ip;
324 	int cg;
325 	long bprev;
326 	int osize, nsize;
327 {
328 	register struct fs *fs;
329 	register struct buf *bp;
330 	register struct cg *cgp;
331 	long bno;
332 	int frags, bbase;
333 	int i;
334 
335 	fs = ip->i_fs;
336 	if (fs->fs_cs(fs, cg).cs_nffree < nsize - osize)
337 		return (NULL);
338 	frags = numfrags(fs, nsize);
339 	bbase = fragoff(fs, bprev);
340 	if (bbase > (bprev + frags - 1) % fs->fs_frag) {
341 		/* cannot extend across a block boundry */
342 		return (NULL);
343 	}
344 	bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), fs->fs_bsize);
345 	cgp = bp->b_un.b_cg;
346 	if (bp->b_flags & B_ERROR || cgp->cg_magic != CG_MAGIC) {
347 		brelse(bp);
348 		return (NULL);
349 	}
350 	bno = dtogd(fs, bprev);
351 	for (i = numfrags(fs, osize); i < frags; i++)
352 		if (isclr(cgp->cg_free, bno + i)) {
353 			brelse(bp);
354 			return (NULL);
355 		}
356 	/*
357 	 * the current fragment can be extended
358 	 * deduct the count on fragment being extended into
359 	 * increase the count on the remaining fragment (if any)
360 	 * allocate the extended piece
361 	 */
362 	for (i = frags; i < fs->fs_frag - bbase; i++)
363 		if (isclr(cgp->cg_free, bno + i))
364 			break;
365 	cgp->cg_frsum[i - numfrags(fs, osize)]--;
366 	if (i != frags)
367 		cgp->cg_frsum[i - frags]++;
368 	for (i = numfrags(fs, osize); i < frags; i++) {
369 		clrbit(cgp->cg_free, bno + i);
370 		cgp->cg_cs.cs_nffree--;
371 		fs->fs_cstotal.cs_nffree--;
372 		fs->fs_cs(fs, cg).cs_nffree--;
373 	}
374 	fs->fs_fmod++;
375 	bdwrite(bp);
376 	return (bprev);
377 }
378 
379 /*
380  * Determine whether a block can be allocated.
381  *
382  * Check to see if a block of the apprpriate size is available,
383  * and if it is, allocate it.
384  */
385 daddr_t
386 alloccg(ip, cg, bpref, size)
387 	struct inode *ip;
388 	int cg;
389 	daddr_t bpref;
390 	int size;
391 {
392 	register struct fs *fs;
393 	register struct buf *bp;
394 	register struct cg *cgp;
395 	int bno, frags;
396 	int allocsiz;
397 	register int i;
398 
399 	fs = ip->i_fs;
400 	if (fs->fs_cs(fs, cg).cs_nbfree == 0 && size == fs->fs_bsize)
401 		return (NULL);
402 	bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), fs->fs_bsize);
403 	cgp = bp->b_un.b_cg;
404 	if (bp->b_flags & B_ERROR || cgp->cg_magic != CG_MAGIC) {
405 		brelse(bp);
406 		return (NULL);
407 	}
408 	if (size == fs->fs_bsize) {
409 		bno = alloccgblk(fs, cgp, bpref);
410 		bdwrite(bp);
411 		return (bno);
412 	}
413 	/*
414 	 * check to see if any fragments are already available
415 	 * allocsiz is the size which will be allocated, hacking
416 	 * it down to a smaller size if necessary
417 	 */
418 	frags = numfrags(fs, size);
419 	for (allocsiz = frags; allocsiz < fs->fs_frag; allocsiz++)
420 		if (cgp->cg_frsum[allocsiz] != 0)
421 			break;
422 	if (allocsiz == fs->fs_frag) {
423 		/*
424 		 * no fragments were available, so a block will be
425 		 * allocated, and hacked up
426 		 */
427 		if (cgp->cg_cs.cs_nbfree == 0) {
428 			brelse(bp);
429 			return (NULL);
430 		}
431 		bno = alloccgblk(fs, cgp, bpref);
432 		bpref = dtogd(fs, bno);
433 		for (i = frags; i < fs->fs_frag; i++)
434 			setbit(cgp->cg_free, bpref + i);
435 		i = fs->fs_frag - frags;
436 		cgp->cg_cs.cs_nffree += i;
437 		fs->fs_cstotal.cs_nffree += i;
438 		fs->fs_cs(fs, cg).cs_nffree += i;
439 		cgp->cg_frsum[i]++;
440 		bdwrite(bp);
441 		return (bno);
442 	}
443 	bno = mapsearch(fs, cgp, bpref, allocsiz);
444 	if (bno == -1)
445 		return (NULL);
446 	for (i = 0; i < frags; i++)
447 		clrbit(cgp->cg_free, bno + i);
448 	cgp->cg_cs.cs_nffree -= frags;
449 	fs->fs_cstotal.cs_nffree -= frags;
450 	fs->fs_cs(fs, cg).cs_nffree -= frags;
451 	cgp->cg_frsum[allocsiz]--;
452 	if (frags != allocsiz)
453 		cgp->cg_frsum[allocsiz - frags]++;
454 	bdwrite(bp);
455 	return (cg * fs->fs_fpg + bno);
456 }
457 
458 /*
459  * Allocate a block in a cylinder group.
460  *
461  * This algorithm implements the following policy:
462  *   1) allocate the requested block.
463  *   2) allocate a rotationally optimal block in the same cylinder.
464  *   3) allocate the next available block on the block rotor for the
465  *      specified cylinder group.
466  * Note that this routine only allocates fs_bsize blocks; these
467  * blocks may be fragmented by the routine that allocates them.
468  */
469 daddr_t
470 alloccgblk(fs, cgp, bpref)
471 	register struct fs *fs;
472 	register struct cg *cgp;
473 	daddr_t bpref;
474 {
475 	daddr_t bno;
476 	int cylno, pos, delta;
477 	short *cylbp;
478 	register int i;
479 
480 	if (bpref == 0) {
481 		bpref = cgp->cg_rotor;
482 		goto norot;
483 	}
484 	bpref &= ~(fs->fs_frag - 1);
485 	bpref = dtogd(fs, bpref);
486 	/*
487 	 * if the requested block is available, use it
488 	 */
489 	if (isblock(fs, cgp->cg_free, bpref/fs->fs_frag)) {
490 		bno = bpref;
491 		goto gotit;
492 	}
493 	/*
494 	 * check for a block available on the same cylinder
495 	 */
496 	cylno = cbtocylno(fs, bpref);
497 	if (cgp->cg_btot[cylno] == 0)
498 		goto norot;
499 	if (fs->fs_cpc == 0) {
500 		/*
501 		 * block layout info is not available, so just have
502 		 * to take any block in this cylinder.
503 		 */
504 		bpref = howmany(fs->fs_spc * cylno, NSPF(fs));
505 		goto norot;
506 	}
507 	/*
508 	 * find a block that is rotationally optimal
509 	 */
510 	cylbp = cgp->cg_b[cylno];
511 	if (fs->fs_rotdelay == 0) {
512 		pos = cbtorpos(fs, bpref);
513 	} else {
514 		/*
515 		 * here we convert ms of delay to frags as:
516 		 * (frags) = (ms) * (rev/sec) * (sect/rev) /
517 		 *	((sect/frag) * (ms/sec))
518 		 * then round up to the next rotational position
519 		 */
520 		bpref += fs->fs_rotdelay * fs->fs_rps * fs->fs_nsect /
521 		    (NSPF(fs) * 1000);
522 		pos = cbtorpos(fs, bpref);
523 		pos = (pos + 1) % NRPOS;
524 	}
525 	/*
526 	 * check the summary information to see if a block is
527 	 * available in the requested cylinder starting at the
528 	 * optimal rotational position and proceeding around.
529 	 */
530 	for (i = pos; i < NRPOS; i++)
531 		if (cylbp[i] > 0)
532 			break;
533 	if (i == NRPOS)
534 		for (i = 0; i < pos; i++)
535 			if (cylbp[i] > 0)
536 				break;
537 	if (cylbp[i] > 0) {
538 		/*
539 		 * found a rotational position, now find the actual
540 		 * block. A panic if none is actually there.
541 		 */
542 		pos = cylno % fs->fs_cpc;
543 		bno = (cylno - pos) * fs->fs_spc / NSPB(fs);
544 		if (fs->fs_postbl[pos][i] == -1)
545 			panic("alloccgblk: cyl groups corrupted");
546 		for (i = fs->fs_postbl[pos][i];; ) {
547 			if (isblock(fs, cgp->cg_free, bno + i)) {
548 				bno = (bno + i) * fs->fs_frag;
549 				goto gotit;
550 			}
551 			delta = fs->fs_rotbl[i];
552 			if (delta <= 0 || delta > MAXBPC - i)
553 				break;
554 			i += delta;
555 		}
556 		panic("alloccgblk: can't find blk in cyl");
557 	}
558 norot:
559 	/*
560 	 * no blocks in the requested cylinder, so take next
561 	 * available one in this cylinder group.
562 	 */
563 	bno = mapsearch(fs, cgp, bpref, fs->fs_frag);
564 	if (bno == -1)
565 		return (NULL);
566 	cgp->cg_rotor = bno;
567 gotit:
568 	clrblock(fs, cgp->cg_free, bno/fs->fs_frag);
569 	cgp->cg_cs.cs_nbfree--;
570 	fs->fs_cstotal.cs_nbfree--;
571 	fs->fs_cs(fs, cgp->cg_cgx).cs_nbfree--;
572 	cylno = cbtocylno(fs, bno);
573 	cgp->cg_b[cylno][cbtorpos(fs, bno)]--;
574 	cgp->cg_btot[cylno]--;
575 	fs->fs_fmod++;
576 	return (cgp->cg_cgx * fs->fs_fpg + bno);
577 }
578 
579 /*
580  * Determine whether an inode can be allocated.
581  *
582  * Check to see if an inode is available, and if it is,
583  * allocate it using the following policy:
584  *   1) allocate the requested inode.
585  *   2) allocate the next available inode after the requested
586  *      inode in the specified cylinder group.
587  */
588 ino_t
589 ialloccg(ip, cg, ipref, mode)
590 	struct inode *ip;
591 	int cg;
592 	daddr_t ipref;
593 	int mode;
594 {
595 	register struct fs *fs;
596 	register struct buf *bp;
597 	register struct cg *cgp;
598 	int i;
599 
600 	fs = ip->i_fs;
601 	if (fs->fs_cs(fs, cg).cs_nifree == 0)
602 		return (NULL);
603 	bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), fs->fs_bsize);
604 	cgp = bp->b_un.b_cg;
605 	if (bp->b_flags & B_ERROR || cgp->cg_magic != CG_MAGIC) {
606 		brelse(bp);
607 		return (NULL);
608 	}
609 	if (ipref) {
610 		ipref %= fs->fs_ipg;
611 		if (isclr(cgp->cg_iused, ipref))
612 			goto gotit;
613 	} else
614 		ipref = cgp->cg_irotor;
615 	for (i = 0; i < fs->fs_ipg; i++) {
616 		ipref++;
617 		if (ipref >= fs->fs_ipg)
618 			ipref = 0;
619 		if (isclr(cgp->cg_iused, ipref)) {
620 			cgp->cg_irotor = ipref;
621 			goto gotit;
622 		}
623 	}
624 	brelse(bp);
625 	return (NULL);
626 gotit:
627 	setbit(cgp->cg_iused, ipref);
628 	cgp->cg_cs.cs_nifree--;
629 	fs->fs_cstotal.cs_nifree--;
630 	fs->fs_cs(fs, cg).cs_nifree--;
631 	fs->fs_fmod++;
632 	if ((mode & IFMT) == IFDIR) {
633 		cgp->cg_cs.cs_ndir++;
634 		fs->fs_cstotal.cs_ndir++;
635 		fs->fs_cs(fs, cg).cs_ndir++;
636 	}
637 	bdwrite(bp);
638 	return (cg * fs->fs_ipg + ipref);
639 }
640 
641 /*
642  * Free a block or fragment.
643  *
644  * The specified block or fragment is placed back in the
645  * free map. If a fragment is deallocated, a possible
646  * block reassembly is checked.
647  */
648 fre(ip, bno, size)
649 	register struct inode *ip;
650 	daddr_t bno;
651 	off_t size;
652 {
653 	register struct fs *fs;
654 	register struct cg *cgp;
655 	register struct buf *bp;
656 	int cg, blk, frags, bbase;
657 	register int i;
658 
659 	fs = ip->i_fs;
660 	if ((unsigned)size > fs->fs_bsize || fragoff(fs, size) != 0)
661 		panic("free: bad size");
662 	cg = dtog(fs, bno);
663 	if (badblock(fs, bno))
664 		return;
665 	bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), fs->fs_bsize);
666 	cgp = bp->b_un.b_cg;
667 	if (bp->b_flags & B_ERROR || cgp->cg_magic != CG_MAGIC) {
668 		brelse(bp);
669 		return;
670 	}
671 	bno = dtogd(fs, bno);
672 	if (size == fs->fs_bsize) {
673 		if (isblock(fs, cgp->cg_free, bno/fs->fs_frag))
674 			panic("free: freeing free block");
675 		setblock(fs, cgp->cg_free, bno/fs->fs_frag);
676 		cgp->cg_cs.cs_nbfree++;
677 		fs->fs_cstotal.cs_nbfree++;
678 		fs->fs_cs(fs, cg).cs_nbfree++;
679 		i = cbtocylno(fs, bno);
680 		cgp->cg_b[i][cbtorpos(fs, bno)]++;
681 		cgp->cg_btot[i]++;
682 	} else {
683 		bbase = bno - (bno % fs->fs_frag);
684 		/*
685 		 * decrement the counts associated with the old frags
686 		 */
687 		blk = blkmap(fs, cgp->cg_free, bbase);
688 		fragacct(fs, blk, cgp->cg_frsum, -1);
689 		/*
690 		 * deallocate the fragment
691 		 */
692 		frags = numfrags(fs, size);
693 		for (i = 0; i < frags; i++) {
694 			if (isset(cgp->cg_free, bno + i))
695 				panic("free: freeing free frag");
696 			setbit(cgp->cg_free, bno + i);
697 		}
698 		cgp->cg_cs.cs_nffree += i;
699 		fs->fs_cstotal.cs_nffree += i;
700 		fs->fs_cs(fs, cg).cs_nffree += i;
701 		/*
702 		 * add back in counts associated with the new frags
703 		 */
704 		blk = blkmap(fs, cgp->cg_free, bbase);
705 		fragacct(fs, blk, cgp->cg_frsum, 1);
706 		/*
707 		 * if a complete block has been reassembled, account for it
708 		 */
709 		if (isblock(fs, cgp->cg_free, bbase / fs->fs_frag)) {
710 			cgp->cg_cs.cs_nffree -= fs->fs_frag;
711 			fs->fs_cstotal.cs_nffree -= fs->fs_frag;
712 			fs->fs_cs(fs, cg).cs_nffree -= fs->fs_frag;
713 			cgp->cg_cs.cs_nbfree++;
714 			fs->fs_cstotal.cs_nbfree++;
715 			fs->fs_cs(fs, cg).cs_nbfree++;
716 			i = cbtocylno(fs, bbase);
717 			cgp->cg_b[i][cbtorpos(fs, bbase)]++;
718 			cgp->cg_btot[i]++;
719 		}
720 	}
721 	fs->fs_fmod++;
722 	bdwrite(bp);
723 }
724 
725 /*
726  * Free an inode.
727  *
728  * The specified inode is placed back in the free map.
729  */
730 ifree(ip, ino, mode)
731 	struct inode *ip;
732 	ino_t ino;
733 	int mode;
734 {
735 	register struct fs *fs;
736 	register struct cg *cgp;
737 	register struct buf *bp;
738 	int cg;
739 
740 	fs = ip->i_fs;
741 	if ((unsigned)ino >= fs->fs_ipg*fs->fs_ncg)
742 		panic("ifree: range");
743 	cg = itog(fs, ino);
744 	bp = bread(ip->i_dev, fsbtodb(fs, cgtod(fs, cg)), fs->fs_bsize);
745 	cgp = bp->b_un.b_cg;
746 	if (bp->b_flags & B_ERROR || cgp->cg_magic != CG_MAGIC) {
747 		brelse(bp);
748 		return;
749 	}
750 	ino %= fs->fs_ipg;
751 	if (isclr(cgp->cg_iused, ino))
752 		panic("ifree: freeing free inode");
753 	clrbit(cgp->cg_iused, ino);
754 	cgp->cg_cs.cs_nifree++;
755 	fs->fs_cstotal.cs_nifree++;
756 	fs->fs_cs(fs, cg).cs_nifree++;
757 	if ((mode & IFMT) == IFDIR) {
758 		cgp->cg_cs.cs_ndir--;
759 		fs->fs_cstotal.cs_ndir--;
760 		fs->fs_cs(fs, cg).cs_ndir--;
761 	}
762 	fs->fs_fmod++;
763 	bdwrite(bp);
764 }
765 
766 /*
767  * Find a block of the specified size in the specified cylinder group.
768  *
769  * It is a panic if a request is made to find a block if none are
770  * available.
771  */
772 daddr_t
773 mapsearch(fs, cgp, bpref, allocsiz)
774 	register struct fs *fs;
775 	register struct cg *cgp;
776 	daddr_t bpref;
777 	int allocsiz;
778 {
779 	daddr_t bno;
780 	int start, len, loc, i;
781 	int blk, field, subfield, pos;
782 
783 	/*
784 	 * find the fragment by searching through the free block
785 	 * map for an appropriate bit pattern
786 	 */
787 	if (bpref)
788 		start = dtogd(fs, bpref) / NBBY;
789 	else
790 		start = cgp->cg_frotor / NBBY;
791 	len = howmany(fs->fs_fpg, NBBY) - start;
792 	loc = scanc(len, &cgp->cg_free[start], fragtbl[fs->fs_frag],
793 		1 << (allocsiz - 1 + (fs->fs_frag % NBBY)));
794 	if (loc == 0) {
795 		len = start + 1;
796 		start = 0;
797 		loc = scanc(len, &cgp->cg_free[start], fragtbl[fs->fs_frag],
798 			1 << (allocsiz - 1 + (fs->fs_frag % NBBY)));
799 		if (loc == 0) {
800 			panic("alloccg: map corrupted");
801 			return (-1);
802 		}
803 	}
804 	bno = (start + len - loc) * NBBY;
805 	cgp->cg_frotor = bno;
806 	/*
807 	 * found the byte in the map
808 	 * sift through the bits to find the selected frag
809 	 */
810 	for (i = bno + NBBY; bno < i; bno += fs->fs_frag) {
811 		blk = blkmap(fs, cgp->cg_free, bno);
812 		blk <<= 1;
813 		field = around[allocsiz];
814 		subfield = inside[allocsiz];
815 		for (pos = 0; pos <= fs->fs_frag - allocsiz; pos++) {
816 			if ((blk & field) == subfield)
817 				return (bno + pos);
818 			field <<= 1;
819 			subfield <<= 1;
820 		}
821 	}
822 	panic("alloccg: block not in map");
823 	return (-1);
824 }
825 
826 /*
827  * Update the frsum fields to reflect addition or deletion
828  * of some frags.
829  */
830 fragacct(fs, fragmap, fraglist, cnt)
831 	struct fs *fs;
832 	int fragmap;
833 	long fraglist[];
834 	int cnt;
835 {
836 	int inblk;
837 	register int field, subfield;
838 	register int siz, pos;
839 
840 	inblk = (int)(fragtbl[fs->fs_frag][fragmap]) << 1;
841 	fragmap <<= 1;
842 	for (siz = 1; siz < fs->fs_frag; siz++) {
843 		if ((inblk & (1 << (siz + (fs->fs_frag % NBBY)))) == 0)
844 			continue;
845 		field = around[siz];
846 		subfield = inside[siz];
847 		for (pos = siz; pos <= fs->fs_frag; pos++) {
848 			if ((fragmap & field) == subfield) {
849 				fraglist[siz] += cnt;
850 				pos += siz;
851 				field <<= siz;
852 				subfield <<= siz;
853 			}
854 			field <<= 1;
855 			subfield <<= 1;
856 		}
857 	}
858 }
859 
860 /*
861  * Check that a specified block number is in range.
862  */
863 badblock(fs, bn)
864 	register struct fs *fs;
865 	daddr_t bn;
866 {
867 
868 	if ((unsigned)bn >= fs->fs_size) {
869 		fserr(fs, "bad block");
870 		return (1);
871 	}
872 	return (0);
873 }
874 
875 /*
876  * Getfs maps a device number into a pointer to the incore super block.
877  *
878  * The algorithm is a linear search through the mount table. A
879  * consistency check of the super block magic number is performed.
880  *
881  * panic: no fs -- the device is not mounted.
882  *	this "cannot happen"
883  */
884 struct fs *
885 getfs(dev)
886 	dev_t dev;
887 {
888 	register struct mount *mp;
889 	register struct fs *fs;
890 
891 	for (mp = &mount[0]; mp < &mount[NMOUNT]; mp++) {
892 		if (mp->m_bufp == NULL || mp->m_dev != dev)
893 			continue;
894 		fs = mp->m_bufp->b_un.b_fs;
895 		if (fs->fs_magic != FS_MAGIC)
896 			panic("getfs: bad magic");
897 		return (fs);
898 	}
899 	panic("getfs: no fs");
900 	return (NULL);
901 }
902 
903 /*
904  * Fserr prints the name of a file system with an error diagnostic.
905  *
906  * The form of the error message is:
907  *	fs: error message
908  */
909 fserr(fs, cp)
910 	struct fs *fs;
911 	char *cp;
912 {
913 
914 	printf("%s: %s\n", fs->fs_fsmnt, cp);
915 }
916 
917 /*
918  * Getfsx returns the index in the file system
919  * table of the specified device.  The swap device
920  * is also assigned a pseudo-index.  The index may
921  * be used as a compressed indication of the location
922  * of a block, recording
923  *	<getfsx(dev),blkno>
924  * rather than
925  *	<dev, blkno>
926  * provided the information need remain valid only
927  * as long as the file system is mounted.
928  */
929 getfsx(dev)
930 	dev_t dev;
931 {
932 	register struct mount *mp;
933 
934 	if (dev == swapdev)
935 		return (MSWAPX);
936 	for(mp = &mount[0]; mp < &mount[NMOUNT]; mp++)
937 		if (mp->m_dev == dev)
938 			return (mp - &mount[0]);
939 	return (-1);
940 }
941 
942 /*
943  * Update is the internal name of 'sync'.  It goes through the disk
944  * queues to initiate sandbagged IO; goes through the inodes to write
945  * modified nodes; and it goes through the mount table to initiate
946  * the writing of the modified super blocks.
947  */
948 update(flag)
949 	int flag;
950 {
951 	register struct inode *ip;
952 	register struct mount *mp;
953 	register struct buf *bp;
954 	struct fs *fs;
955 	int i, blks;
956 
957 	if (updlock)
958 		return;
959 	updlock++;
960 	/*
961 	 * Write back modified superblocks.
962 	 * Consistency check that the superblock
963 	 * of each file system is still in the buffer cache.
964 	 */
965 	for (mp = &mount[0]; mp < &mount[NMOUNT]; mp++) {
966 		if (mp->m_bufp == NULL)
967 			continue;
968 		fs = mp->m_bufp->b_un.b_fs;
969 		if (fs->fs_fmod == 0)
970 			continue;
971 		if (fs->fs_ronly != 0)
972 			panic("update: rofs mod");
973 		bp = getblk(mp->m_dev, SBLOCK, SBSIZE);
974 		if (bp->b_un.b_fs != fs || fs->fs_magic != FS_MAGIC)
975 			panic("update: bad b_fs");
976 		fs->fs_fmod = 0;
977 		fs->fs_time = time;
978 		bwrite(bp);
979 		blks = howmany(fs->fs_cssize, fs->fs_fsize);
980 		for (i = 0; i < blks; i += fs->fs_frag) {
981 			bp = getblk(mp->m_dev,
982 			    fsbtodb(fs, fs->fs_csaddr + i),
983 			    blks - i < fs->fs_frag ?
984 				(blks - i) * fs->fs_fsize :
985 				fs->fs_bsize);
986 			bwrite(bp);
987 		}
988 	}
989 	/*
990 	 * Write back each (modified) inode.
991 	 */
992 	for (ip = inode; ip < inodeNINODE; ip++) {
993 		if ((ip->i_flag & ILOCK) != 0 || ip->i_count == 0)
994 			continue;
995 		ip->i_flag |= ILOCK;
996 		ip->i_count++;
997 		iupdat(ip, &time, &time, 0);
998 		iput(ip);
999 	}
1000 	updlock = 0;
1001 	/*
1002 	 * Force stale buffer cache information to be flushed,
1003 	 * for all devices.
1004 	 */
1005 	bflush(NODEV);
1006 }
1007 
1008 /*
1009  * block operations
1010  *
1011  * check if a block is available
1012  */
1013 isblock(fs, cp, h)
1014 	struct fs *fs;
1015 	unsigned char *cp;
1016 	int h;
1017 {
1018 	unsigned char mask;
1019 
1020 	switch (fs->fs_frag) {
1021 	case 8:
1022 		return (cp[h] == 0xff);
1023 	case 4:
1024 		mask = 0x0f << ((h & 0x1) << 2);
1025 		return ((cp[h >> 1] & mask) == mask);
1026 	case 2:
1027 		mask = 0x03 << ((h & 0x3) << 1);
1028 		return ((cp[h >> 2] & mask) == mask);
1029 	case 1:
1030 		mask = 0x01 << (h & 0x7);
1031 		return ((cp[h >> 3] & mask) == mask);
1032 	default:
1033 		panic("isblock");
1034 		return (NULL);
1035 	}
1036 }
1037 
1038 /*
1039  * take a block out of the map
1040  */
1041 clrblock(fs, cp, h)
1042 	struct fs *fs;
1043 	unsigned char *cp;
1044 	int h;
1045 {
1046 	switch ((fs)->fs_frag) {
1047 	case 8:
1048 		cp[h] = 0;
1049 		return;
1050 	case 4:
1051 		cp[h >> 1] &= ~(0x0f << ((h & 0x1) << 2));
1052 		return;
1053 	case 2:
1054 		cp[h >> 2] &= ~(0x03 << ((h & 0x3) << 1));
1055 		return;
1056 	case 1:
1057 		cp[h >> 3] &= ~(0x01 << (h & 0x7));
1058 		return;
1059 	default:
1060 		panic("clrblock");
1061 		return;
1062 	}
1063 }
1064 
1065 /*
1066  * put a block into the map
1067  */
1068 setblock(fs, cp, h)
1069 	struct fs *fs;
1070 	unsigned char *cp;
1071 	int h;
1072 {
1073 	switch (fs->fs_frag) {
1074 	case 8:
1075 		cp[h] = 0xff;
1076 		return;
1077 	case 4:
1078 		cp[h >> 1] |= (0x0f << ((h & 0x1) << 2));
1079 		return;
1080 	case 2:
1081 		cp[h >> 2] |= (0x03 << ((h & 0x3) << 1));
1082 		return;
1083 	case 1:
1084 		cp[h >> 3] |= (0x01 << (h & 0x7));
1085 		return;
1086 	default:
1087 		panic("setblock");
1088 		return;
1089 	}
1090 }
1091