xref: /freebsd/sys/geom/geom_ccd.c (revision 9768746ba83efa02837c5b9c66348db6e900208f)
1 /*-
2  * SPDX-License-Identifier: (BSD-2-Clause-NetBSD AND BSD-3-Clause)
3  *
4  * Copyright (c) 2003 Poul-Henning Kamp.
5  * Copyright (c) 1996, 1997 The NetBSD Foundation, Inc.
6  * All rights reserved.
7  *
8  * This code is derived from software contributed to The NetBSD Foundation
9  * by Jason R. Thorpe.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions
13  * are met:
14  * 1. Redistributions of source code must retain the above copyright
15  *    notice, this list of conditions and the following disclaimer.
16  * 2. Redistributions in binary form must reproduce the above copyright
17  *    notice, this list of conditions and the following disclaimer in the
18  *    documentation and/or other materials provided with the distribution.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
21  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
22  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
23  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
24  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
25  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
26  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
27  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
28  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
29  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
30  * POSSIBILITY OF SUCH DAMAGE.
31  *
32  * $NetBSD: ccd.c,v 1.22 1995/12/08 19:13:26 thorpej Exp $
33  */
34 
35 /*-
36  * Copyright (c) 1988 University of Utah.
37  * Copyright (c) 1990, 1993
38  *	The Regents of the University of California.  All rights reserved.
39  *
40  * This code is derived from software contributed to Berkeley by
41  * the Systems Programming Group of the University of Utah Computer
42  * Science Department.
43  *
44  * Redistribution and use in source and binary forms, with or without
45  * modification, are permitted provided that the following conditions
46  * are met:
47  * 1. Redistributions of source code must retain the above copyright
48  *    notice, this list of conditions and the following disclaimer.
49  * 2. Redistributions in binary form must reproduce the above copyright
50  *    notice, this list of conditions and the following disclaimer in the
51  *    documentation and/or other materials provided with the distribution.
52  * 3. Neither the name of the University nor the names of its contributors
53  *    may be used to endorse or promote products derived from this software
54  *    without specific prior written permission.
55  *
56  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
57  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
58  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
59  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
60  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
61  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
62  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
63  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
64  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
65  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
66  * SUCH DAMAGE.
67  *
68  * from: Utah $Hdr: cd.c 1.6 90/11/28$
69  *
70  *	@(#)cd.c	8.2 (Berkeley) 11/16/93
71  */
72 
73 /*
74  * Dynamic configuration and disklabel support by:
75  *	Jason R. Thorpe <thorpej@nas.nasa.gov>
76  *	Numerical Aerodynamic Simulation Facility
77  *	Mail Stop 258-6
78  *	NASA Ames Research Center
79  *	Moffett Field, CA 94035
80  */
81 
82 #include <sys/cdefs.h>
83 __FBSDID("$FreeBSD$");
84 
85 #include <sys/param.h>
86 #include <sys/systm.h>
87 #include <sys/kernel.h>
88 #include <sys/module.h>
89 #include <sys/bio.h>
90 #include <sys/malloc.h>
91 #include <sys/sbuf.h>
92 #include <geom/geom.h>
93 
94 /*
95  * Number of blocks to untouched in front of a component partition.
96  * This is to avoid violating its disklabel area when it starts at the
97  * beginning of the slice.
98  */
99 #if !defined(CCD_OFFSET)
100 #define CCD_OFFSET 16
101 #endif
102 
103 /* sc_flags */
104 #define CCDF_UNIFORM	0x02	/* use LCCD of sizes for uniform interleave */
105 #define CCDF_MIRROR	0x04	/* use mirroring */
106 #define CCDF_NO_OFFSET	0x08	/* do not leave space in front */
107 #define CCDF_LINUX	0x10	/* use Linux compatibility mode */
108 
109 /* Mask of user-settable ccd flags. */
110 #define CCDF_USERMASK	(CCDF_UNIFORM|CCDF_MIRROR)
111 
112 /*
113  * Interleave description table.
114  * Computed at boot time to speed irregular-interleave lookups.
115  * The idea is that we interleave in "groups".  First we interleave
116  * evenly over all component disks up to the size of the smallest
117  * component (the first group), then we interleave evenly over all
118  * remaining disks up to the size of the next-smallest (second group),
119  * and so on.
120  *
121  * Each table entry describes the interleave characteristics of one
122  * of these groups.  For example if a concatenated disk consisted of
123  * three components of 5, 3, and 7 DEV_BSIZE blocks interleaved at
124  * DEV_BSIZE (1), the table would have three entries:
125  *
126  *	ndisk	startblk	startoff	dev
127  *	3	0		0		0, 1, 2
128  *	2	9		3		0, 2
129  *	1	13		5		2
130  *	0	-		-		-
131  *
132  * which says that the first nine blocks (0-8) are interleaved over
133  * 3 disks (0, 1, 2) starting at block offset 0 on any component disk,
134  * the next 4 blocks (9-12) are interleaved over 2 disks (0, 2) starting
135  * at component block 3, and the remaining blocks (13-14) are on disk
136  * 2 starting at offset 5.
137  */
138 struct ccdiinfo {
139 	int	ii_ndisk;	/* # of disks range is interleaved over */
140 	daddr_t	ii_startblk;	/* starting scaled block # for range */
141 	daddr_t	ii_startoff;	/* starting component offset (block #) */
142 	int	*ii_index;	/* ordered list of components in range */
143 };
144 
145 /*
146  * Component info table.
147  * Describes a single component of a concatenated disk.
148  */
149 struct ccdcinfo {
150 	daddr_t		ci_size; 		/* size */
151 	struct g_provider *ci_provider;		/* provider */
152 	struct g_consumer *ci_consumer;		/* consumer */
153 };
154 
155 /*
156  * A concatenated disk is described by this structure.
157  */
158 
159 struct ccd_s {
160 	LIST_ENTRY(ccd_s) list;
161 
162 	int		 sc_unit;		/* logical unit number */
163 	int		 sc_flags;		/* flags */
164 	daddr_t		 sc_size;		/* size of ccd */
165 	int		 sc_ileave;		/* interleave */
166 	u_int		 sc_ndisks;		/* number of components */
167 	struct ccdcinfo	 *sc_cinfo;		/* component info */
168 	struct ccdiinfo	 *sc_itable;		/* interleave table */
169 	uint32_t	 sc_secsize;		/* # bytes per sector */
170 	int		 sc_pick;		/* side of mirror picked */
171 	daddr_t		 sc_blk[2];		/* mirror localization */
172 	uint32_t	 sc_offset;		/* actual offset used */
173 };
174 
175 static g_start_t g_ccd_start;
176 static void ccdiodone(struct bio *bp);
177 static void ccdinterleave(struct ccd_s *);
178 static int ccdinit(struct gctl_req *req, struct ccd_s *);
179 static int ccdbuffer(struct bio **ret, struct ccd_s *,
180 		      struct bio *, daddr_t, caddr_t, long);
181 
182 static void
183 g_ccd_orphan(struct g_consumer *cp)
184 {
185 	/*
186 	 * XXX: We don't do anything here.  It is not obvious
187 	 * XXX: what DTRT would be, so we do what the previous
188 	 * XXX: code did: ignore it and let the user cope.
189 	 */
190 }
191 
192 static int
193 g_ccd_access(struct g_provider *pp, int dr, int dw, int de)
194 {
195 	struct g_geom *gp;
196 	struct g_consumer *cp1, *cp2;
197 	int error;
198 
199 	de += dr;
200 	de += dw;
201 
202 	gp = pp->geom;
203 	error = ENXIO;
204 	LIST_FOREACH(cp1, &gp->consumer, consumer) {
205 		error = g_access(cp1, dr, dw, de);
206 		if (error) {
207 			LIST_FOREACH(cp2, &gp->consumer, consumer) {
208 				if (cp1 == cp2)
209 					break;
210 				g_access(cp2, -dr, -dw, -de);
211 			}
212 			break;
213 		}
214 	}
215 	return (error);
216 }
217 
218 /*
219  * Free the softc and its substructures.
220  */
221 static void
222 g_ccd_freesc(struct ccd_s *sc)
223 {
224 	struct ccdiinfo *ii;
225 
226 	g_free(sc->sc_cinfo);
227 	if (sc->sc_itable != NULL) {
228 		for (ii = sc->sc_itable; ii->ii_ndisk > 0; ii++)
229 			g_free(ii->ii_index);
230 		g_free(sc->sc_itable);
231 	}
232 	g_free(sc);
233 }
234 
235 static int
236 ccdinit(struct gctl_req *req, struct ccd_s *cs)
237 {
238 	struct ccdcinfo *ci;
239 	daddr_t size;
240 	int ix;
241 	daddr_t minsize;
242 	int maxsecsize;
243 	off_t mediasize;
244 	u_int sectorsize;
245 
246 	cs->sc_size = 0;
247 
248 	maxsecsize = 0;
249 	minsize = 0;
250 
251 	if (cs->sc_flags & CCDF_LINUX) {
252 		cs->sc_offset = 0;
253 		cs->sc_ileave *= 2;
254 		if (cs->sc_flags & CCDF_MIRROR && cs->sc_ndisks != 2)
255 			gctl_error(req, "Mirror mode for Linux raids is "
256 			                "only supported with 2 devices");
257 	} else {
258 		if (cs->sc_flags & CCDF_NO_OFFSET)
259 			cs->sc_offset = 0;
260 		else
261 			cs->sc_offset = CCD_OFFSET;
262 	}
263 	for (ix = 0; ix < cs->sc_ndisks; ix++) {
264 		ci = &cs->sc_cinfo[ix];
265 
266 		mediasize = ci->ci_provider->mediasize;
267 		sectorsize = ci->ci_provider->sectorsize;
268 		if (sectorsize > maxsecsize)
269 			maxsecsize = sectorsize;
270 		size = mediasize / DEV_BSIZE - cs->sc_offset;
271 
272 		/* Truncate to interleave boundary */
273 
274 		if (cs->sc_ileave > 1)
275 			size -= size % cs->sc_ileave;
276 
277 		if (size == 0) {
278 			gctl_error(req, "Component %s has effective size zero",
279 			    ci->ci_provider->name);
280 			return(ENODEV);
281 		}
282 
283 		if (minsize == 0 || size < minsize)
284 			minsize = size;
285 		ci->ci_size = size;
286 		cs->sc_size += size;
287 	}
288 
289 	/*
290 	 * Don't allow the interleave to be smaller than
291 	 * the biggest component sector.
292 	 */
293 	if ((cs->sc_ileave > 0) &&
294 	    (cs->sc_ileave < (maxsecsize / DEV_BSIZE))) {
295 		gctl_error(req, "Interleave to small for sector size");
296 		return(EINVAL);
297 	}
298 
299 	/*
300 	 * If uniform interleave is desired set all sizes to that of
301 	 * the smallest component.  This will guarantee that a single
302 	 * interleave table is generated.
303 	 *
304 	 * Lost space must be taken into account when calculating the
305 	 * overall size.  Half the space is lost when CCDF_MIRROR is
306 	 * specified.
307 	 */
308 	if (cs->sc_flags & CCDF_UNIFORM) {
309 		for (ix = 0; ix < cs->sc_ndisks; ix++) {
310 			ci = &cs->sc_cinfo[ix];
311 			ci->ci_size = minsize;
312 		}
313 		cs->sc_size = cs->sc_ndisks * minsize;
314 	}
315 
316 	if (cs->sc_flags & CCDF_MIRROR) {
317 		/*
318 		 * Check to see if an even number of components
319 		 * have been specified.  The interleave must also
320 		 * be non-zero in order for us to be able to
321 		 * guarantee the topology.
322 		 */
323 		if (cs->sc_ndisks % 2) {
324 			gctl_error(req,
325 			      "Mirroring requires an even number of disks");
326 			return(EINVAL);
327 		}
328 		if (cs->sc_ileave == 0) {
329 			gctl_error(req,
330 			     "An interleave must be specified when mirroring");
331 			return(EINVAL);
332 		}
333 		cs->sc_size = (cs->sc_ndisks/2) * minsize;
334 	}
335 
336 	/*
337 	 * Construct the interleave table.
338 	 */
339 	ccdinterleave(cs);
340 
341 	/*
342 	 * Create pseudo-geometry based on 1MB cylinders.  It's
343 	 * pretty close.
344 	 */
345 	cs->sc_secsize = maxsecsize;
346 
347 	return (0);
348 }
349 
350 static void
351 ccdinterleave(struct ccd_s *cs)
352 {
353 	struct ccdcinfo *ci, *smallci;
354 	struct ccdiinfo *ii;
355 	daddr_t bn, lbn;
356 	int ix;
357 	daddr_t size;
358 
359 	/*
360 	 * Allocate an interleave table.  The worst case occurs when each
361 	 * of N disks is of a different size, resulting in N interleave
362 	 * tables.
363 	 *
364 	 * Chances are this is too big, but we don't care.
365 	 */
366 	size = (cs->sc_ndisks + 1) * sizeof(struct ccdiinfo);
367 	cs->sc_itable = g_malloc(size, M_WAITOK | M_ZERO);
368 
369 	/*
370 	 * Trivial case: no interleave (actually interleave of disk size).
371 	 * Each table entry represents a single component in its entirety.
372 	 *
373 	 * An interleave of 0 may not be used with a mirror setup.
374 	 */
375 	if (cs->sc_ileave == 0) {
376 		bn = 0;
377 		ii = cs->sc_itable;
378 
379 		for (ix = 0; ix < cs->sc_ndisks; ix++) {
380 			/* Allocate space for ii_index. */
381 			ii->ii_index = g_malloc(sizeof(int), M_WAITOK);
382 			ii->ii_ndisk = 1;
383 			ii->ii_startblk = bn;
384 			ii->ii_startoff = 0;
385 			ii->ii_index[0] = ix;
386 			bn += cs->sc_cinfo[ix].ci_size;
387 			ii++;
388 		}
389 		ii->ii_ndisk = 0;
390 		return;
391 	}
392 
393 	/*
394 	 * The following isn't fast or pretty; it doesn't have to be.
395 	 */
396 	size = 0;
397 	bn = lbn = 0;
398 	for (ii = cs->sc_itable; ; ii++) {
399 		/*
400 		 * Allocate space for ii_index.  We might allocate more then
401 		 * we use.
402 		 */
403 		ii->ii_index = g_malloc((sizeof(int) * cs->sc_ndisks),
404 		    M_WAITOK);
405 
406 		/*
407 		 * Locate the smallest of the remaining components
408 		 */
409 		smallci = NULL;
410 		for (ci = cs->sc_cinfo; ci < &cs->sc_cinfo[cs->sc_ndisks];
411 		    ci++) {
412 			if (ci->ci_size > size &&
413 			    (smallci == NULL ||
414 			     ci->ci_size < smallci->ci_size)) {
415 				smallci = ci;
416 			}
417 		}
418 
419 		/*
420 		 * Nobody left, all done
421 		 */
422 		if (smallci == NULL) {
423 			ii->ii_ndisk = 0;
424 			g_free(ii->ii_index);
425 			ii->ii_index = NULL;
426 			break;
427 		}
428 
429 		/*
430 		 * Record starting logical block using an sc_ileave blocksize.
431 		 */
432 		ii->ii_startblk = bn / cs->sc_ileave;
433 
434 		/*
435 		 * Record starting component block using an sc_ileave
436 		 * blocksize.  This value is relative to the beginning of
437 		 * a component disk.
438 		 */
439 		ii->ii_startoff = lbn;
440 
441 		/*
442 		 * Determine how many disks take part in this interleave
443 		 * and record their indices.
444 		 */
445 		ix = 0;
446 		for (ci = cs->sc_cinfo;
447 		    ci < &cs->sc_cinfo[cs->sc_ndisks]; ci++) {
448 			if (ci->ci_size >= smallci->ci_size) {
449 				ii->ii_index[ix++] = ci - cs->sc_cinfo;
450 			}
451 		}
452 		ii->ii_ndisk = ix;
453 		bn += ix * (smallci->ci_size - size);
454 		lbn = smallci->ci_size / cs->sc_ileave;
455 		size = smallci->ci_size;
456 	}
457 }
458 
459 static void
460 g_ccd_start(struct bio *bp)
461 {
462 	long bcount, rcount;
463 	struct bio *cbp[2];
464 	caddr_t addr;
465 	daddr_t bn;
466 	int err;
467 	struct ccd_s *cs;
468 
469 	cs = bp->bio_to->geom->softc;
470 
471 	/*
472 	 * Block all GETATTR requests, we wouldn't know which of our
473 	 * subdevices we should ship it off to.
474 	 * XXX: this may not be the right policy.
475 	 */
476 	if(bp->bio_cmd == BIO_GETATTR) {
477 		g_io_deliver(bp, EINVAL);
478 		return;
479 	}
480 
481 	/*
482 	 * Translate the partition-relative block number to an absolute.
483 	 */
484 	bn = bp->bio_offset / cs->sc_secsize;
485 
486 	/*
487 	 * Allocate component buffers and fire off the requests
488 	 */
489 	addr = bp->bio_data;
490 	for (bcount = bp->bio_length; bcount > 0; bcount -= rcount) {
491 		err = ccdbuffer(cbp, cs, bp, bn, addr, bcount);
492 		if (err) {
493 			bp->bio_completed += bcount;
494 			if (bp->bio_error == 0)
495 				bp->bio_error = err;
496 			if (bp->bio_completed == bp->bio_length)
497 				g_io_deliver(bp, bp->bio_error);
498 			return;
499 		}
500 		rcount = cbp[0]->bio_length;
501 
502 		if (cs->sc_flags & CCDF_MIRROR) {
503 			/*
504 			 * Mirroring.  Writes go to both disks, reads are
505 			 * taken from whichever disk seems most appropriate.
506 			 *
507 			 * We attempt to localize reads to the disk whos arm
508 			 * is nearest the read request.  We ignore seeks due
509 			 * to writes when making this determination and we
510 			 * also try to avoid hogging.
511 			 */
512 			if (cbp[0]->bio_cmd != BIO_READ) {
513 				g_io_request(cbp[0], cbp[0]->bio_from);
514 				g_io_request(cbp[1], cbp[1]->bio_from);
515 			} else {
516 				int pick = cs->sc_pick;
517 				daddr_t range = cs->sc_size / 16;
518 
519 				if (bn < cs->sc_blk[pick] - range ||
520 				    bn > cs->sc_blk[pick] + range
521 				) {
522 					cs->sc_pick = pick = 1 - pick;
523 				}
524 				cs->sc_blk[pick] = bn + btodb(rcount);
525 				g_io_request(cbp[pick], cbp[pick]->bio_from);
526 			}
527 		} else {
528 			/*
529 			 * Not mirroring
530 			 */
531 			g_io_request(cbp[0], cbp[0]->bio_from);
532 		}
533 		bn += btodb(rcount);
534 		addr += rcount;
535 	}
536 }
537 
538 /*
539  * Build a component buffer header.
540  */
541 static int
542 ccdbuffer(struct bio **cb, struct ccd_s *cs, struct bio *bp, daddr_t bn, caddr_t addr, long bcount)
543 {
544 	struct ccdcinfo *ci, *ci2 = NULL;
545 	struct bio *cbp;
546 	daddr_t cbn, cboff;
547 	off_t cbc;
548 
549 	/*
550 	 * Determine which component bn falls in.
551 	 */
552 	cbn = bn;
553 	cboff = 0;
554 
555 	if (cs->sc_ileave == 0) {
556 		/*
557 		 * Serially concatenated and neither a mirror nor a parity
558 		 * config.  This is a special case.
559 		 */
560 		daddr_t sblk;
561 
562 		sblk = 0;
563 		for (ci = cs->sc_cinfo; cbn >= sblk + ci->ci_size; ci++)
564 			sblk += ci->ci_size;
565 		cbn -= sblk;
566 	} else {
567 		struct ccdiinfo *ii;
568 		int ccdisk, off;
569 
570 		/*
571 		 * Calculate cbn, the logical superblock (sc_ileave chunks),
572 		 * and cboff, a normal block offset (DEV_BSIZE chunks) relative
573 		 * to cbn.
574 		 */
575 		cboff = cbn % cs->sc_ileave;	/* DEV_BSIZE gran */
576 		cbn = cbn / cs->sc_ileave;	/* DEV_BSIZE * ileave gran */
577 
578 		/*
579 		 * Figure out which interleave table to use.
580 		 */
581 		for (ii = cs->sc_itable; ii->ii_ndisk; ii++) {
582 			if (ii->ii_startblk > cbn)
583 				break;
584 		}
585 		ii--;
586 
587 		/*
588 		 * off is the logical superblock relative to the beginning
589 		 * of this interleave block.
590 		 */
591 		off = cbn - ii->ii_startblk;
592 
593 		/*
594 		 * We must calculate which disk component to use (ccdisk),
595 		 * and recalculate cbn to be the superblock relative to
596 		 * the beginning of the component.  This is typically done by
597 		 * adding 'off' and ii->ii_startoff together.  However, 'off'
598 		 * must typically be divided by the number of components in
599 		 * this interleave array to be properly convert it from a
600 		 * CCD-relative logical superblock number to a
601 		 * component-relative superblock number.
602 		 */
603 		if (ii->ii_ndisk == 1) {
604 			/*
605 			 * When we have just one disk, it can't be a mirror
606 			 * or a parity config.
607 			 */
608 			ccdisk = ii->ii_index[0];
609 			cbn = ii->ii_startoff + off;
610 		} else {
611 			if (cs->sc_flags & CCDF_MIRROR) {
612 				/*
613 				 * We have forced a uniform mapping, resulting
614 				 * in a single interleave array.  We double
615 				 * up on the first half of the available
616 				 * components and our mirror is in the second
617 				 * half.  This only works with a single
618 				 * interleave array because doubling up
619 				 * doubles the number of sectors, so there
620 				 * cannot be another interleave array because
621 				 * the next interleave array's calculations
622 				 * would be off.
623 				 */
624 				int ndisk2 = ii->ii_ndisk / 2;
625 				ccdisk = ii->ii_index[off % ndisk2];
626 				cbn = ii->ii_startoff + off / ndisk2;
627 				ci2 = &cs->sc_cinfo[ccdisk + ndisk2];
628 			} else {
629 				ccdisk = ii->ii_index[off % ii->ii_ndisk];
630 				cbn = ii->ii_startoff + off / ii->ii_ndisk;
631 			}
632 		}
633 
634 		ci = &cs->sc_cinfo[ccdisk];
635 
636 		/*
637 		 * Convert cbn from a superblock to a normal block so it
638 		 * can be used to calculate (along with cboff) the normal
639 		 * block index into this particular disk.
640 		 */
641 		cbn *= cs->sc_ileave;
642 	}
643 
644 	/*
645 	 * Fill in the component buf structure.
646 	 */
647 	cbp = g_clone_bio(bp);
648 	if (cbp == NULL)
649 		return (ENOMEM);
650 	cbp->bio_done = g_std_done;
651 	cbp->bio_offset = dbtob(cbn + cboff + cs->sc_offset);
652 	cbp->bio_data = addr;
653 	if (cs->sc_ileave == 0)
654               cbc = dbtob((off_t)(ci->ci_size - cbn));
655 	else
656               cbc = dbtob((off_t)(cs->sc_ileave - cboff));
657 	cbp->bio_length = (cbc < bcount) ? cbc : bcount;
658 
659 	cbp->bio_from = ci->ci_consumer;
660 	cb[0] = cbp;
661 
662 	if (cs->sc_flags & CCDF_MIRROR) {
663 		cbp = g_clone_bio(bp);
664 		if (cbp == NULL)
665 			return (ENOMEM);
666 		cbp->bio_done = cb[0]->bio_done = ccdiodone;
667 		cbp->bio_offset = cb[0]->bio_offset;
668 		cbp->bio_data = cb[0]->bio_data;
669 		cbp->bio_length = cb[0]->bio_length;
670 		cbp->bio_from = ci2->ci_consumer;
671 		cbp->bio_caller1 = cb[0];
672 		cb[0]->bio_caller1 = cbp;
673 		cb[1] = cbp;
674 	}
675 	return (0);
676 }
677 
678 /*
679  * Called only for mirrored operations.
680  */
681 static void
682 ccdiodone(struct bio *cbp)
683 {
684 	struct bio *mbp, *pbp;
685 
686 	mbp = cbp->bio_caller1;
687 	pbp = cbp->bio_parent;
688 
689 	if (pbp->bio_cmd == BIO_READ) {
690 		if (cbp->bio_error == 0) {
691 			/* We will not be needing the partner bio */
692 			if (mbp != NULL) {
693 				pbp->bio_inbed++;
694 				g_destroy_bio(mbp);
695 			}
696 			g_std_done(cbp);
697 			return;
698 		}
699 		if (mbp != NULL) {
700 			/* Try partner the bio instead */
701 			mbp->bio_caller1 = NULL;
702 			pbp->bio_inbed++;
703 			g_destroy_bio(cbp);
704 			g_io_request(mbp, mbp->bio_from);
705 			/*
706 			 * XXX: If this comes back OK, we should actually
707 			 * try to write the good data on the failed mirror
708 			 */
709 			return;
710 		}
711 		g_std_done(cbp);
712 		return;
713 	}
714 	if (mbp != NULL) {
715 		mbp->bio_caller1 = NULL;
716 		pbp->bio_inbed++;
717 		if (cbp->bio_error != 0 && pbp->bio_error == 0)
718 			pbp->bio_error = cbp->bio_error;
719 		g_destroy_bio(cbp);
720 		return;
721 	}
722 	g_std_done(cbp);
723 }
724 
725 static void
726 g_ccd_create(struct gctl_req *req, struct g_class *mp)
727 {
728 	int *unit, *ileave, *nprovider;
729 	struct g_geom *gp;
730 	struct g_consumer *cp;
731 	struct g_provider *pp;
732 	struct ccd_s *sc;
733 	struct sbuf *sb;
734 	char buf[20];
735 	int i, error;
736 
737 	g_topology_assert();
738 	unit = gctl_get_paraml(req, "unit", sizeof (*unit));
739 	if (unit == NULL) {
740 		gctl_error(req, "unit parameter not given");
741 		return;
742 	}
743 	ileave = gctl_get_paraml(req, "ileave", sizeof (*ileave));
744 	if (ileave == NULL) {
745 		gctl_error(req, "ileave parameter not given");
746 		return;
747 	}
748 	nprovider = gctl_get_paraml(req, "nprovider", sizeof (*nprovider));
749 	if (nprovider == NULL) {
750 		gctl_error(req, "nprovider parameter not given");
751 		return;
752 	}
753 
754 	/* Check for duplicate unit */
755 	LIST_FOREACH(gp, &mp->geom, geom) {
756 		sc = gp->softc;
757 		if (sc != NULL && sc->sc_unit == *unit) {
758 			gctl_error(req, "Unit %d already configured", *unit);
759 			return;
760 		}
761 	}
762 
763 	if (*nprovider <= 0) {
764 		gctl_error(req, "Bogus nprovider argument (= %d)", *nprovider);
765 		return;
766 	}
767 
768 	/* Check all providers are valid */
769 	for (i = 0; i < *nprovider; i++) {
770 		snprintf(buf, sizeof(buf), "provider%d", i);
771 		pp = gctl_get_provider(req, buf);
772 		if (pp == NULL)
773 			return;
774 	}
775 
776 	gp = g_new_geomf(mp, "ccd%d", *unit);
777 	sc = g_malloc(sizeof *sc, M_WAITOK | M_ZERO);
778 	gp->softc = sc;
779 	sc->sc_ndisks = *nprovider;
780 
781 	/* Allocate space for the component info. */
782 	sc->sc_cinfo = g_malloc(sc->sc_ndisks * sizeof(struct ccdcinfo),
783 	    M_WAITOK | M_ZERO);
784 
785 	/* Create consumers and attach to all providers */
786 	for (i = 0; i < *nprovider; i++) {
787 		snprintf(buf, sizeof(buf), "provider%d", i);
788 		pp = gctl_get_provider(req, buf);
789 		cp = g_new_consumer(gp);
790 		error = g_attach(cp, pp);
791 		KASSERT(error == 0, ("attach to %s failed", pp->name));
792 		sc->sc_cinfo[i].ci_consumer = cp;
793 		sc->sc_cinfo[i].ci_provider = pp;
794 	}
795 
796 	sc->sc_unit = *unit;
797 	sc->sc_ileave = *ileave;
798 
799 	if (gctl_get_param(req, "no_offset", NULL))
800 		sc->sc_flags |= CCDF_NO_OFFSET;
801 	if (gctl_get_param(req, "linux", NULL))
802 		sc->sc_flags |= CCDF_LINUX;
803 
804 	if (gctl_get_param(req, "uniform", NULL))
805 		sc->sc_flags |= CCDF_UNIFORM;
806 	if (gctl_get_param(req, "mirror", NULL))
807 		sc->sc_flags |= CCDF_MIRROR;
808 
809 	if (sc->sc_ileave == 0 && (sc->sc_flags & CCDF_MIRROR)) {
810 		printf("%s: disabling mirror, interleave is 0\n", gp->name);
811 		sc->sc_flags &= ~(CCDF_MIRROR);
812 	}
813 
814 	if ((sc->sc_flags & CCDF_MIRROR) && !(sc->sc_flags & CCDF_UNIFORM)) {
815 		printf("%s: mirror/parity forces uniform flag\n", gp->name);
816 		sc->sc_flags |= CCDF_UNIFORM;
817 	}
818 
819 	error = ccdinit(req, sc);
820 	if (error != 0) {
821 		g_ccd_freesc(sc);
822 		gp->softc = NULL;
823 		g_wither_geom(gp, ENXIO);
824 		return;
825 	}
826 
827 	pp = g_new_providerf(gp, "%s", gp->name);
828 	pp->mediasize = sc->sc_size * (off_t)sc->sc_secsize;
829 	pp->sectorsize = sc->sc_secsize;
830 	g_error_provider(pp, 0);
831 
832 	sb = sbuf_new_auto();
833 	sbuf_printf(sb, "ccd%d: %d components ", sc->sc_unit, *nprovider);
834 	for (i = 0; i < *nprovider; i++) {
835 		sbuf_printf(sb, "%s%s",
836 		    i == 0 ? "(" : ", ",
837 		    sc->sc_cinfo[i].ci_provider->name);
838 	}
839 	sbuf_printf(sb, "), %jd blocks ", (off_t)pp->mediasize / DEV_BSIZE);
840 	if (sc->sc_ileave != 0)
841 		sbuf_printf(sb, "interleaved at %d blocks\n",
842 			sc->sc_ileave);
843 	else
844 		sbuf_printf(sb, "concatenated\n");
845 	sbuf_finish(sb);
846 	gctl_set_param_err(req, "output", sbuf_data(sb), sbuf_len(sb) + 1);
847 	sbuf_delete(sb);
848 }
849 
850 static int
851 g_ccd_destroy_geom(struct gctl_req *req, struct g_class *mp, struct g_geom *gp)
852 {
853 	struct g_provider *pp;
854 	struct ccd_s *sc;
855 
856 	g_topology_assert();
857 	sc = gp->softc;
858 	pp = LIST_FIRST(&gp->provider);
859 	if (sc == NULL || pp == NULL)
860 		return (EBUSY);
861 	if (pp->acr != 0 || pp->acw != 0 || pp->ace != 0) {
862 		gctl_error(req, "%s is open(r%dw%de%d)", gp->name,
863 		    pp->acr, pp->acw, pp->ace);
864 		return (EBUSY);
865 	}
866 	g_ccd_freesc(sc);
867 	gp->softc = NULL;
868 	g_wither_geom(gp, ENXIO);
869 	return (0);
870 }
871 
872 static void
873 g_ccd_list(struct gctl_req *req, struct g_class *mp)
874 {
875 	struct sbuf *sb;
876 	struct ccd_s *cs;
877 	struct g_geom *gp;
878 	int i, unit, *up;
879 
880 	up = gctl_get_paraml(req, "unit", sizeof (*up));
881 	if (up == NULL) {
882 		gctl_error(req, "unit parameter not given");
883 		return;
884 	}
885 	unit = *up;
886 	sb = sbuf_new_auto();
887 	LIST_FOREACH(gp, &mp->geom, geom) {
888 		cs = gp->softc;
889 		if (cs == NULL || (unit >= 0 && unit != cs->sc_unit))
890 			continue;
891 		sbuf_printf(sb, "ccd%d\t\t%d\t%d\t",
892 		    cs->sc_unit, cs->sc_ileave, cs->sc_flags & CCDF_USERMASK);
893 
894 		for (i = 0; i < cs->sc_ndisks; ++i) {
895 			sbuf_printf(sb, "%s/dev/%s", i == 0 ? "" : " ",
896 			    cs->sc_cinfo[i].ci_provider->name);
897 		}
898 		sbuf_printf(sb, "\n");
899 	}
900 	sbuf_finish(sb);
901 	gctl_set_param_err(req, "output", sbuf_data(sb), sbuf_len(sb) + 1);
902 	sbuf_delete(sb);
903 }
904 
905 static void
906 g_ccd_config(struct gctl_req *req, struct g_class *mp, char const *verb)
907 {
908 	struct g_geom *gp;
909 
910 	g_topology_assert();
911 	if (!strcmp(verb, "create geom")) {
912 		g_ccd_create(req, mp);
913 	} else if (!strcmp(verb, "destroy geom")) {
914 		gp = gctl_get_geom(req, mp, "geom");
915 		if (gp != NULL)
916 			g_ccd_destroy_geom(req, mp, gp);
917 	} else if (!strcmp(verb, "list")) {
918 		g_ccd_list(req, mp);
919 	} else {
920 		gctl_error(req, "unknown verb");
921 	}
922 }
923 
924 static struct g_class g_ccd_class = {
925 	.name = "CCD",
926 	.version = G_VERSION,
927 	.ctlreq = g_ccd_config,
928 	.destroy_geom = g_ccd_destroy_geom,
929 	.start = g_ccd_start,
930 	.orphan = g_ccd_orphan,
931 	.access = g_ccd_access,
932 };
933 
934 DECLARE_GEOM_CLASS(g_ccd_class, g_ccd);
935 MODULE_VERSION(geom_ccd, 0);
936