Home | History | Annotate | Line # | Download | only in dev
ccd.c revision 1.106
      1 /*	$NetBSD: ccd.c,v 1.106 2005/10/15 17:29:11 yamt Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 1996, 1997, 1998, 1999 The NetBSD Foundation, Inc.
      5  * All rights reserved.
      6  *
      7  * This code is derived from software contributed to The NetBSD Foundation
      8  * by Jason R. Thorpe.
      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) 1990, 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  * the Systems Programming Group of the University of Utah Computer
     45  * Science Department.
     46  *
     47  * Redistribution and use in source and binary forms, with or without
     48  * modification, are permitted provided that the following conditions
     49  * are met:
     50  * 1. Redistributions of source code must retain the above copyright
     51  *    notice, this list of conditions and the following disclaimer.
     52  * 2. Redistributions in binary form must reproduce the above copyright
     53  *    notice, this list of conditions and the following disclaimer in the
     54  *    documentation and/or other materials provided with the distribution.
     55  * 3. Neither the name of the University nor the names of its contributors
     56  *    may be used to endorse or promote products derived from this software
     57  *    without specific prior written permission.
     58  *
     59  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     60  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     61  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     62  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     63  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     64  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     65  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     66  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     67  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     68  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     69  * SUCH DAMAGE.
     70  *
     71  * from: Utah $Hdr: cd.c 1.6 90/11/28$
     72  *
     73  *	@(#)cd.c	8.2 (Berkeley) 11/16/93
     74  */
     75 
     76 /*
     77  * Copyright (c) 1988 University of Utah.
     78  *
     79  * This code is derived from software contributed to Berkeley by
     80  * the Systems Programming Group of the University of Utah Computer
     81  * Science Department.
     82  *
     83  * Redistribution and use in source and binary forms, with or without
     84  * modification, are permitted provided that the following conditions
     85  * are met:
     86  * 1. Redistributions of source code must retain the above copyright
     87  *    notice, this list of conditions and the following disclaimer.
     88  * 2. Redistributions in binary form must reproduce the above copyright
     89  *    notice, this list of conditions and the following disclaimer in the
     90  *    documentation and/or other materials provided with the distribution.
     91  * 3. All advertising materials mentioning features or use of this software
     92  *    must display the following acknowledgement:
     93  *	This product includes software developed by the University of
     94  *	California, Berkeley and its contributors.
     95  * 4. Neither the name of the University nor the names of its contributors
     96  *    may be used to endorse or promote products derived from this software
     97  *    without specific prior written permission.
     98  *
     99  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
    100  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
    101  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
    102  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
    103  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
    104  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
    105  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
    106  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
    107  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
    108  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
    109  * SUCH DAMAGE.
    110  *
    111  * from: Utah $Hdr: cd.c 1.6 90/11/28$
    112  *
    113  *	@(#)cd.c	8.2 (Berkeley) 11/16/93
    114  */
    115 
    116 /*
    117  * "Concatenated" disk driver.
    118  *
    119  * Dynamic configuration and disklabel support by:
    120  *	Jason R. Thorpe <thorpej (at) nas.nasa.gov>
    121  *	Numerical Aerodynamic Simulation Facility
    122  *	Mail Stop 258-6
    123  *	NASA Ames Research Center
    124  *	Moffett Field, CA 94035
    125  */
    126 
    127 #include <sys/cdefs.h>
    128 __KERNEL_RCSID(0, "$NetBSD: ccd.c,v 1.106 2005/10/15 17:29:11 yamt Exp $");
    129 
    130 #include <sys/param.h>
    131 #include <sys/systm.h>
    132 #include <sys/proc.h>
    133 #include <sys/errno.h>
    134 #include <sys/buf.h>
    135 #include <sys/bufq.h>
    136 #include <sys/malloc.h>
    137 #include <sys/pool.h>
    138 #include <sys/namei.h>
    139 #include <sys/stat.h>
    140 #include <sys/ioctl.h>
    141 #include <sys/disklabel.h>
    142 #include <sys/device.h>
    143 #include <sys/disk.h>
    144 #include <sys/syslog.h>
    145 #include <sys/fcntl.h>
    146 #include <sys/vnode.h>
    147 #include <sys/conf.h>
    148 #include <sys/lock.h>
    149 #include <sys/queue.h>
    150 
    151 #include <dev/ccdvar.h>
    152 
    153 #if defined(CCDDEBUG) && !defined(DEBUG)
    154 #define DEBUG
    155 #endif
    156 
    157 #ifdef DEBUG
    158 #define CCDB_FOLLOW	0x01
    159 #define CCDB_INIT	0x02
    160 #define CCDB_IO		0x04
    161 #define CCDB_LABEL	0x08
    162 #define CCDB_VNODE	0x10
    163 int ccddebug = 0x00;
    164 #endif
    165 
    166 #define	ccdunit(x)	DISKUNIT(x)
    167 
    168 struct ccdbuf {
    169 	struct buf	cb_buf;		/* new I/O buf */
    170 	struct buf	*cb_obp;	/* ptr. to original I/O buf */
    171 	struct ccd_softc *cb_sc;	/* pointer to ccd softc */
    172 	int		cb_comp;	/* target component */
    173 	SIMPLEQ_ENTRY(ccdbuf) cb_q;	/* fifo of component buffers */
    174 };
    175 
    176 /* component buffer pool */
    177 static struct pool ccd_cbufpool;
    178 
    179 #define	CCD_GETBUF()		pool_get(&ccd_cbufpool, PR_NOWAIT)
    180 #define	CCD_PUTBUF(cbp)		pool_put(&ccd_cbufpool, cbp)
    181 
    182 #define CCDLABELDEV(dev)	\
    183 	(MAKEDISKDEV(major((dev)), ccdunit((dev)), RAW_PART))
    184 
    185 /* called by main() at boot time */
    186 void	ccdattach(int);
    187 
    188 /* called by biodone() at interrupt time */
    189 static void	ccdiodone(struct buf *);
    190 
    191 static void	ccdstart(struct ccd_softc *);
    192 static void	ccdinterleave(struct ccd_softc *);
    193 static void	ccdintr(struct ccd_softc *, struct buf *);
    194 static int	ccdinit(struct ccd_softc *, char **, struct vnode **,
    195 		    struct proc *);
    196 static int	ccdlookup(char *, struct proc *p, struct vnode **);
    197 static struct ccdbuf *ccdbuffer(struct ccd_softc *, struct buf *,
    198 		    daddr_t, caddr_t, long);
    199 static void	ccdgetdefaultlabel(struct ccd_softc *, struct disklabel *);
    200 static void	ccdgetdisklabel(dev_t);
    201 static void	ccdmakedisklabel(struct ccd_softc *);
    202 
    203 static dev_type_open(ccdopen);
    204 static dev_type_close(ccdclose);
    205 static dev_type_read(ccdread);
    206 static dev_type_write(ccdwrite);
    207 static dev_type_ioctl(ccdioctl);
    208 static dev_type_strategy(ccdstrategy);
    209 static dev_type_dump(ccddump);
    210 static dev_type_size(ccdsize);
    211 
    212 const struct bdevsw ccd_bdevsw = {
    213 	ccdopen, ccdclose, ccdstrategy, ccdioctl, ccddump, ccdsize, D_DISK
    214 };
    215 
    216 const struct cdevsw ccd_cdevsw = {
    217 	ccdopen, ccdclose, ccdread, ccdwrite, ccdioctl,
    218 	nostop, notty, nopoll, nommap, nokqfilter, D_DISK
    219 };
    220 
    221 #ifdef DEBUG
    222 static	void printiinfo(struct ccdiinfo *);
    223 #endif
    224 
    225 /* Non-private for the benefit of libkvm. */
    226 struct	ccd_softc *ccd_softc;
    227 int	numccd = 0;
    228 
    229 /*
    230  * Called by main() during pseudo-device attachment.  All we need
    231  * to do is allocate enough space for devices to be configured later.
    232  */
    233 void
    234 ccdattach(int num)
    235 {
    236 	struct ccd_softc *cs;
    237 	int i;
    238 
    239 	if (num <= 0) {
    240 #ifdef DIAGNOSTIC
    241 		panic("ccdattach: count <= 0");
    242 #endif
    243 		return;
    244 	}
    245 
    246 	ccd_softc = (struct ccd_softc *)malloc(num * sizeof(struct ccd_softc),
    247 	    M_DEVBUF, M_NOWAIT|M_ZERO);
    248 	if (ccd_softc == NULL) {
    249 		printf("WARNING: no memory for concatenated disks\n");
    250 		return;
    251 	}
    252 	numccd = num;
    253 
    254 	/* Initialize the component buffer pool. */
    255 	pool_init(&ccd_cbufpool, sizeof(struct ccdbuf), 0,
    256 	    0, 0, "ccdpl", NULL);
    257 
    258 	/* Initialize per-softc structures. */
    259 	for (i = 0; i < num; i++) {
    260 		cs = &ccd_softc[i];
    261 		snprintf(cs->sc_xname, sizeof(cs->sc_xname), "ccd%d", i);
    262 		cs->sc_dkdev.dk_name = cs->sc_xname;	/* XXX */
    263 		lockinit(&cs->sc_lock, PRIBIO, "ccdlk", 0, 0);
    264 		pseudo_disk_init(&cs->sc_dkdev);
    265 	}
    266 }
    267 
    268 static int
    269 ccdinit(struct ccd_softc *cs, char **cpaths, struct vnode **vpp,
    270     struct proc *p)
    271 {
    272 	struct ccdcinfo *ci = NULL;
    273 	size_t size;
    274 	int ix;
    275 	struct vattr va;
    276 	size_t minsize;
    277 	int maxsecsize;
    278 	struct partinfo dpart;
    279 	struct ccdgeom *ccg = &cs->sc_geom;
    280 	char tmppath[MAXPATHLEN];
    281 	int error, path_alloced;
    282 
    283 #ifdef DEBUG
    284 	if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
    285 		printf("%s: ccdinit\n", cs->sc_xname);
    286 #endif
    287 
    288 	/* Allocate space for the component info. */
    289 	cs->sc_cinfo = malloc(cs->sc_nccdisks * sizeof(struct ccdcinfo),
    290 	    M_DEVBUF, M_WAITOK);
    291 
    292 	cs->sc_size = 0;
    293 
    294 	/*
    295 	 * Verify that each component piece exists and record
    296 	 * relevant information about it.
    297 	 */
    298 	maxsecsize = 0;
    299 	minsize = 0;
    300 	for (ix = 0, path_alloced = 0; ix < cs->sc_nccdisks; ix++) {
    301 		ci = &cs->sc_cinfo[ix];
    302 		ci->ci_vp = vpp[ix];
    303 
    304 		/*
    305 		 * Copy in the pathname of the component.
    306 		 */
    307 		memset(tmppath, 0, sizeof(tmppath));	/* sanity */
    308 		error = copyinstr(cpaths[ix], tmppath,
    309 		    MAXPATHLEN, &ci->ci_pathlen);
    310 		if (error) {
    311 #ifdef DEBUG
    312 			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
    313 				printf("%s: can't copy path, error = %d\n",
    314 				    cs->sc_xname, error);
    315 #endif
    316 			goto out;
    317 		}
    318 		ci->ci_path = malloc(ci->ci_pathlen, M_DEVBUF, M_WAITOK);
    319 		memcpy(ci->ci_path, tmppath, ci->ci_pathlen);
    320 		path_alloced++;
    321 
    322 		/*
    323 		 * XXX: Cache the component's dev_t.
    324 		 */
    325 		if ((error = VOP_GETATTR(vpp[ix], &va, p->p_ucred, p)) != 0) {
    326 #ifdef DEBUG
    327 			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
    328 				printf("%s: %s: getattr failed %s = %d\n",
    329 				    cs->sc_xname, ci->ci_path,
    330 				    "error", error);
    331 #endif
    332 			goto out;
    333 		}
    334 		ci->ci_dev = va.va_rdev;
    335 
    336 		/*
    337 		 * Get partition information for the component.
    338 		 */
    339 		error = VOP_IOCTL(vpp[ix], DIOCGPART, &dpart,
    340 		    FREAD, p->p_ucred, p);
    341 		if (error) {
    342 #ifdef DEBUG
    343 			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
    344 				 printf("%s: %s: ioctl failed, error = %d\n",
    345 				     cs->sc_xname, ci->ci_path, error);
    346 #endif
    347 			goto out;
    348 		}
    349 
    350 /*
    351  * This diagnostic test is disabled (for now?) since not all port supports
    352  * on-disk BSD disklabel.
    353  */
    354 #if 0 /* def DIAGNOSTIC */
    355 		/* Check fstype field of component. */
    356 		if (dpart.part->p_fstype != FS_CCD)
    357 			printf("%s: WARNING: %s: fstype %d != FS_CCD\n",
    358 			    cs->sc_xname, ci->ci_path, dpart.part->p_fstype);
    359 #endif
    360 
    361 		/*
    362 		 * Calculate the size, truncating to an interleave
    363 		 * boundary if necessary.
    364 		 */
    365 		maxsecsize =
    366 		    ((dpart.disklab->d_secsize > maxsecsize) ?
    367 		    dpart.disklab->d_secsize : maxsecsize);
    368 		size = dpart.part->p_size;
    369 		if (cs->sc_ileave > 1)
    370 			size -= size % cs->sc_ileave;
    371 
    372 		if (size == 0) {
    373 #ifdef DEBUG
    374 			if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
    375 				printf("%s: %s: size == 0\n",
    376 				    cs->sc_xname, ci->ci_path);
    377 #endif
    378 			error = ENODEV;
    379 			goto out;
    380 		}
    381 
    382 		if (minsize == 0 || size < minsize)
    383 			minsize = size;
    384 		ci->ci_size = size;
    385 		cs->sc_size += size;
    386 	}
    387 
    388 	/*
    389 	 * Don't allow the interleave to be smaller than
    390 	 * the biggest component sector.
    391 	 */
    392 	if ((cs->sc_ileave > 0) &&
    393 	    (cs->sc_ileave < (maxsecsize / DEV_BSIZE))) {
    394 #ifdef DEBUG
    395 		if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
    396 			printf("%s: interleave must be at least %d\n",
    397 			    cs->sc_xname, (maxsecsize / DEV_BSIZE));
    398 #endif
    399 		error = EINVAL;
    400 		goto out;
    401 	}
    402 
    403 	/*
    404 	 * If uniform interleave is desired set all sizes to that of
    405 	 * the smallest component.
    406 	 */
    407 	if (cs->sc_flags & CCDF_UNIFORM) {
    408 		for (ci = cs->sc_cinfo;
    409 		     ci < &cs->sc_cinfo[cs->sc_nccdisks]; ci++)
    410 			ci->ci_size = minsize;
    411 
    412 		cs->sc_size = cs->sc_nccdisks * minsize;
    413 	}
    414 
    415 	/*
    416 	 * Construct the interleave table.
    417 	 */
    418 	ccdinterleave(cs);
    419 
    420 	/*
    421 	 * Create pseudo-geometry based on 1MB cylinders.  It's
    422 	 * pretty close.
    423 	 */
    424 	ccg->ccg_secsize = DEV_BSIZE;
    425 	ccg->ccg_ntracks = 1;
    426 	ccg->ccg_nsectors = 1024 * (1024 / ccg->ccg_secsize);
    427 	ccg->ccg_ncylinders = cs->sc_size / ccg->ccg_nsectors;
    428 
    429 	cs->sc_flags |= CCDF_INITED;
    430 
    431 	return (0);
    432 
    433  out:
    434 	for (ix = 0; ix < path_alloced; ix++)
    435 		free(cs->sc_cinfo[ix].ci_path, M_DEVBUF);
    436 	free(cs->sc_cinfo, M_DEVBUF);
    437 	return (error);
    438 }
    439 
    440 static void
    441 ccdinterleave(struct ccd_softc *cs)
    442 {
    443 	struct ccdcinfo *ci, *smallci;
    444 	struct ccdiinfo *ii;
    445 	daddr_t bn, lbn;
    446 	int ix;
    447 	u_long size;
    448 
    449 #ifdef DEBUG
    450 	if (ccddebug & CCDB_INIT)
    451 		printf("ccdinterleave(%p): ileave %d\n", cs, cs->sc_ileave);
    452 #endif
    453 	/*
    454 	 * Allocate an interleave table.
    455 	 * Chances are this is too big, but we don't care.
    456 	 */
    457 	size = (cs->sc_nccdisks + 1) * sizeof(struct ccdiinfo);
    458 	cs->sc_itable = (struct ccdiinfo *)malloc(size, M_DEVBUF,
    459 	    M_WAITOK|M_ZERO);
    460 
    461 	/*
    462 	 * Trivial case: no interleave (actually interleave of disk size).
    463 	 * Each table entry represents a single component in its entirety.
    464 	 */
    465 	if (cs->sc_ileave == 0) {
    466 		bn = 0;
    467 		ii = cs->sc_itable;
    468 
    469 		for (ix = 0; ix < cs->sc_nccdisks; ix++) {
    470 			/* Allocate space for ii_index. */
    471 			ii->ii_index = malloc(sizeof(int), M_DEVBUF, M_WAITOK);
    472 			ii->ii_ndisk = 1;
    473 			ii->ii_startblk = bn;
    474 			ii->ii_startoff = 0;
    475 			ii->ii_index[0] = ix;
    476 			bn += cs->sc_cinfo[ix].ci_size;
    477 			ii++;
    478 		}
    479 		ii->ii_ndisk = 0;
    480 #ifdef DEBUG
    481 		if (ccddebug & CCDB_INIT)
    482 			printiinfo(cs->sc_itable);
    483 #endif
    484 		return;
    485 	}
    486 
    487 	/*
    488 	 * The following isn't fast or pretty; it doesn't have to be.
    489 	 */
    490 	size = 0;
    491 	bn = lbn = 0;
    492 	for (ii = cs->sc_itable; ; ii++) {
    493 		/* Allocate space for ii_index. */
    494 		ii->ii_index = malloc((sizeof(int) * cs->sc_nccdisks),
    495 		    M_DEVBUF, M_WAITOK);
    496 
    497 		/*
    498 		 * Locate the smallest of the remaining components
    499 		 */
    500 		smallci = NULL;
    501 		for (ci = cs->sc_cinfo;
    502 		     ci < &cs->sc_cinfo[cs->sc_nccdisks]; ci++)
    503 			if (ci->ci_size > size &&
    504 			    (smallci == NULL ||
    505 			     ci->ci_size < smallci->ci_size))
    506 				smallci = ci;
    507 
    508 		/*
    509 		 * Nobody left, all done
    510 		 */
    511 		if (smallci == NULL) {
    512 			ii->ii_ndisk = 0;
    513 			break;
    514 		}
    515 
    516 		/*
    517 		 * Record starting logical block and component offset
    518 		 */
    519 		ii->ii_startblk = bn / cs->sc_ileave;
    520 		ii->ii_startoff = lbn;
    521 
    522 		/*
    523 		 * Determine how many disks take part in this interleave
    524 		 * and record their indices.
    525 		 */
    526 		ix = 0;
    527 		for (ci = cs->sc_cinfo;
    528 		     ci < &cs->sc_cinfo[cs->sc_nccdisks]; ci++)
    529 			if (ci->ci_size >= smallci->ci_size)
    530 				ii->ii_index[ix++] = ci - cs->sc_cinfo;
    531 		ii->ii_ndisk = ix;
    532 		bn += ix * (smallci->ci_size - size);
    533 		lbn = smallci->ci_size / cs->sc_ileave;
    534 		size = smallci->ci_size;
    535 	}
    536 #ifdef DEBUG
    537 	if (ccddebug & CCDB_INIT)
    538 		printiinfo(cs->sc_itable);
    539 #endif
    540 }
    541 
    542 /* ARGSUSED */
    543 static int
    544 ccdopen(dev_t dev, int flags, int fmt, struct proc *p)
    545 {
    546 	int unit = ccdunit(dev);
    547 	struct ccd_softc *cs;
    548 	struct disklabel *lp;
    549 	int error = 0, part, pmask;
    550 
    551 #ifdef DEBUG
    552 	if (ccddebug & CCDB_FOLLOW)
    553 		printf("ccdopen(0x%x, 0x%x)\n", dev, flags);
    554 #endif
    555 	if (unit >= numccd)
    556 		return (ENXIO);
    557 	cs = &ccd_softc[unit];
    558 
    559 	if ((error = lockmgr(&cs->sc_lock, LK_EXCLUSIVE, NULL)) != 0)
    560 		return (error);
    561 
    562 	lp = cs->sc_dkdev.dk_label;
    563 
    564 	part = DISKPART(dev);
    565 	pmask = (1 << part);
    566 
    567 	/*
    568 	 * If we're initialized, check to see if there are any other
    569 	 * open partitions.  If not, then it's safe to update
    570 	 * the in-core disklabel.  Only read the disklabel if it is
    571 	 * not already valid.
    572 	 */
    573 	if ((cs->sc_flags & (CCDF_INITED|CCDF_VLABEL)) == CCDF_INITED &&
    574 	    cs->sc_dkdev.dk_openmask == 0)
    575 		ccdgetdisklabel(dev);
    576 
    577 	/* Check that the partition exists. */
    578 	if (part != RAW_PART) {
    579 		if (((cs->sc_flags & CCDF_INITED) == 0) ||
    580 		    ((part >= lp->d_npartitions) ||
    581 		     (lp->d_partitions[part].p_fstype == FS_UNUSED))) {
    582 			error = ENXIO;
    583 			goto done;
    584 		}
    585 	}
    586 
    587 	/* Prevent our unit from being unconfigured while open. */
    588 	switch (fmt) {
    589 	case S_IFCHR:
    590 		cs->sc_dkdev.dk_copenmask |= pmask;
    591 		break;
    592 
    593 	case S_IFBLK:
    594 		cs->sc_dkdev.dk_bopenmask |= pmask;
    595 		break;
    596 	}
    597 	cs->sc_dkdev.dk_openmask =
    598 	    cs->sc_dkdev.dk_copenmask | cs->sc_dkdev.dk_bopenmask;
    599 
    600  done:
    601 	(void) lockmgr(&cs->sc_lock, LK_RELEASE, NULL);
    602 	return (error);
    603 }
    604 
    605 /* ARGSUSED */
    606 static int
    607 ccdclose(dev_t dev, int flags, int fmt, struct proc *p)
    608 {
    609 	int unit = ccdunit(dev);
    610 	struct ccd_softc *cs;
    611 	int error = 0, part;
    612 
    613 #ifdef DEBUG
    614 	if (ccddebug & CCDB_FOLLOW)
    615 		printf("ccdclose(0x%x, 0x%x)\n", dev, flags);
    616 #endif
    617 
    618 	if (unit >= numccd)
    619 		return (ENXIO);
    620 	cs = &ccd_softc[unit];
    621 
    622 	if ((error = lockmgr(&cs->sc_lock, LK_EXCLUSIVE, NULL)) != 0)
    623 		return (error);
    624 
    625 	part = DISKPART(dev);
    626 
    627 	/* ...that much closer to allowing unconfiguration... */
    628 	switch (fmt) {
    629 	case S_IFCHR:
    630 		cs->sc_dkdev.dk_copenmask &= ~(1 << part);
    631 		break;
    632 
    633 	case S_IFBLK:
    634 		cs->sc_dkdev.dk_bopenmask &= ~(1 << part);
    635 		break;
    636 	}
    637 	cs->sc_dkdev.dk_openmask =
    638 	    cs->sc_dkdev.dk_copenmask | cs->sc_dkdev.dk_bopenmask;
    639 
    640 	if (cs->sc_dkdev.dk_openmask == 0) {
    641 		if ((cs->sc_flags & CCDF_KLABEL) == 0)
    642 			cs->sc_flags &= ~CCDF_VLABEL;
    643 	}
    644 
    645 	(void) lockmgr(&cs->sc_lock, LK_RELEASE, NULL);
    646 	return (0);
    647 }
    648 
    649 static void
    650 ccdstrategy(struct buf *bp)
    651 {
    652 	int unit = ccdunit(bp->b_dev);
    653 	struct ccd_softc *cs = &ccd_softc[unit];
    654 	daddr_t blkno;
    655 	int s;
    656 	int wlabel;
    657 	struct disklabel *lp;
    658 
    659 #ifdef DEBUG
    660 	if (ccddebug & CCDB_FOLLOW)
    661 		printf("ccdstrategy(%p): unit %d\n", bp, unit);
    662 #endif
    663 	if ((cs->sc_flags & CCDF_INITED) == 0) {
    664 #ifdef DEBUG
    665 		if (ccddebug & CCDB_FOLLOW)
    666 			printf("ccdstrategy: unit %d: not inited\n", unit);
    667 #endif
    668 		bp->b_error = ENXIO;
    669 		bp->b_flags |= B_ERROR;
    670 		goto done;
    671 	}
    672 
    673 	/* If it's a nil transfer, wake up the top half now. */
    674 	if (bp->b_bcount == 0)
    675 		goto done;
    676 
    677 	lp = cs->sc_dkdev.dk_label;
    678 
    679 	/*
    680 	 * Do bounds checking and adjust transfer.  If there's an
    681 	 * error, the bounds check will flag that for us.  Convert
    682 	 * the partition relative block number to an absolute.
    683 	 */
    684 	blkno = bp->b_blkno;
    685 	wlabel = cs->sc_flags & (CCDF_WLABEL|CCDF_LABELLING);
    686 	if (DISKPART(bp->b_dev) != RAW_PART) {
    687 		if (bounds_check_with_label(&cs->sc_dkdev, bp, wlabel) <= 0)
    688 			goto done;
    689 		blkno += lp->d_partitions[DISKPART(bp->b_dev)].p_offset;
    690 	}
    691 	bp->b_rawblkno = blkno;
    692 
    693 	/* Place it in the queue and start I/O on the unit. */
    694 	s = splbio();
    695 	BUFQ_PUT(cs->sc_bufq, bp);
    696 	ccdstart(cs);
    697 	splx(s);
    698 	return;
    699 
    700  done:
    701 	bp->b_resid = bp->b_bcount;
    702 	biodone(bp);
    703 }
    704 
    705 static void
    706 ccdstart(struct ccd_softc *cs)
    707 {
    708 	long bcount, rcount;
    709 	struct buf *bp;
    710 	struct ccdbuf *cbp;
    711 	caddr_t addr;
    712 	daddr_t bn;
    713 	SIMPLEQ_HEAD(, ccdbuf) cbufq;
    714 
    715 #ifdef DEBUG
    716 	if (ccddebug & CCDB_FOLLOW)
    717 		printf("ccdstart(%p)\n", cs);
    718 #endif
    719 
    720 	/* See if there is work for us to do. */
    721 	while ((bp = BUFQ_PEEK(cs->sc_bufq)) != NULL) {
    722 		/* Instrumentation. */
    723 		disk_busy(&cs->sc_dkdev);
    724 
    725 		bp->b_resid = bp->b_bcount;
    726 		bn = bp->b_rawblkno;
    727 
    728 		/* Allocate the component buffers. */
    729 		SIMPLEQ_INIT(&cbufq);
    730 		addr = bp->b_data;
    731 		for (bcount = bp->b_bcount; bcount > 0; bcount -= rcount) {
    732 			cbp = ccdbuffer(cs, bp, bn, addr, bcount);
    733 			if (cbp == NULL) {
    734 				/*
    735 				 * Can't allocate a component buffer; just
    736 				 * defer the job until later.
    737 				 *
    738 				 * XXX We might consider a watchdog timer
    739 				 * XXX to make sure we are kicked into action,
    740 				 * XXX or consider a low-water mark for our
    741 				 * XXX component buffer pool.
    742 				 */
    743 				while ((cbp = SIMPLEQ_FIRST(&cbufq)) != NULL) {
    744 					SIMPLEQ_REMOVE_HEAD(&cbufq, cb_q);
    745 					CCD_PUTBUF(cbp);
    746 				}
    747 				disk_unbusy(&cs->sc_dkdev, 0, 0);
    748 				return;
    749 			}
    750 			SIMPLEQ_INSERT_TAIL(&cbufq, cbp, cb_q);
    751 			rcount = cbp->cb_buf.b_bcount;
    752 			bn += btodb(rcount);
    753 			addr += rcount;
    754 		}
    755 
    756 		/* Transfer all set up, remove job from the queue. */
    757 		(void) BUFQ_GET(cs->sc_bufq);
    758 
    759 		/* Now fire off the requests. */
    760 		while ((cbp = SIMPLEQ_FIRST(&cbufq)) != NULL) {
    761 			SIMPLEQ_REMOVE_HEAD(&cbufq, cb_q);
    762 			if ((cbp->cb_buf.b_flags & B_READ) == 0)
    763 				cbp->cb_buf.b_vp->v_numoutput++;
    764 			DEV_STRATEGY(&cbp->cb_buf);
    765 		}
    766 	}
    767 }
    768 
    769 /*
    770  * Build a component buffer header.
    771  */
    772 static struct ccdbuf *
    773 ccdbuffer(struct ccd_softc *cs, struct buf *bp, daddr_t bn, caddr_t addr,
    774     long bcount)
    775 {
    776 	struct ccdcinfo *ci;
    777 	struct ccdbuf *cbp;
    778 	daddr_t cbn, cboff;
    779 	u_int64_t cbc;
    780 	int ccdisk;
    781 
    782 #ifdef DEBUG
    783 	if (ccddebug & CCDB_IO)
    784 		printf("ccdbuffer(%p, %p, %" PRId64 ", %p, %ld)\n",
    785 		       cs, bp, bn, addr, bcount);
    786 #endif
    787 	/*
    788 	 * Determine which component bn falls in.
    789 	 */
    790 	cbn = bn;
    791 	cboff = 0;
    792 
    793 	/*
    794 	 * Serially concatenated
    795 	 */
    796 	if (cs->sc_ileave == 0) {
    797 		daddr_t sblk;
    798 
    799 		sblk = 0;
    800 		for (ccdisk = 0, ci = &cs->sc_cinfo[ccdisk];
    801 		    cbn >= sblk + ci->ci_size;
    802 		    ccdisk++, ci = &cs->sc_cinfo[ccdisk])
    803 			sblk += ci->ci_size;
    804 		cbn -= sblk;
    805 	}
    806 	/*
    807 	 * Interleaved
    808 	 */
    809 	else {
    810 		struct ccdiinfo *ii;
    811 		int off;
    812 
    813 		cboff = cbn % cs->sc_ileave;
    814 		cbn /= cs->sc_ileave;
    815 		for (ii = cs->sc_itable; ii->ii_ndisk; ii++)
    816 			if (ii->ii_startblk > cbn)
    817 				break;
    818 		ii--;
    819 		off = cbn - ii->ii_startblk;
    820 		if (ii->ii_ndisk == 1) {
    821 			ccdisk = ii->ii_index[0];
    822 			cbn = ii->ii_startoff + off;
    823 		} else {
    824 			ccdisk = ii->ii_index[off % ii->ii_ndisk];
    825 			cbn = ii->ii_startoff + off / ii->ii_ndisk;
    826 		}
    827 		cbn *= cs->sc_ileave;
    828 		ci = &cs->sc_cinfo[ccdisk];
    829 	}
    830 
    831 	/*
    832 	 * Fill in the component buf structure.
    833 	 */
    834 	cbp = CCD_GETBUF();
    835 	if (cbp == NULL)
    836 		return (NULL);
    837 	BUF_INIT(&cbp->cb_buf);
    838 	cbp->cb_buf.b_flags = bp->b_flags | B_CALL;
    839 	cbp->cb_buf.b_iodone = ccdiodone;
    840 	cbp->cb_buf.b_proc = bp->b_proc;
    841 	cbp->cb_buf.b_dev = ci->ci_dev;
    842 	cbp->cb_buf.b_blkno = cbn + cboff;
    843 	cbp->cb_buf.b_data = addr;
    844 	cbp->cb_buf.b_vp = ci->ci_vp;
    845 	if (cs->sc_ileave == 0)
    846 		cbc = dbtob((u_int64_t)(ci->ci_size - cbn));
    847 	else
    848 		cbc = dbtob((u_int64_t)(cs->sc_ileave - cboff));
    849 	cbp->cb_buf.b_bcount = cbc < bcount ? cbc : bcount;
    850 
    851 	/*
    852 	 * context for ccdiodone
    853 	 */
    854 	cbp->cb_obp = bp;
    855 	cbp->cb_sc = cs;
    856 	cbp->cb_comp = ccdisk;
    857 
    858 	BIO_COPYPRIO(&cbp->cb_buf, bp);
    859 
    860 #ifdef DEBUG
    861 	if (ccddebug & CCDB_IO)
    862 		printf(" dev 0x%x(u%lu): cbp %p bn %" PRId64 " addr %p"
    863 		       " bcnt %d\n",
    864 		    ci->ci_dev, (unsigned long) (ci-cs->sc_cinfo), cbp,
    865 		    cbp->cb_buf.b_blkno, cbp->cb_buf.b_data,
    866 		    cbp->cb_buf.b_bcount);
    867 #endif
    868 
    869 	return (cbp);
    870 }
    871 
    872 static void
    873 ccdintr(struct ccd_softc *cs, struct buf *bp)
    874 {
    875 
    876 #ifdef DEBUG
    877 	if (ccddebug & CCDB_FOLLOW)
    878 		printf("ccdintr(%p, %p)\n", cs, bp);
    879 #endif
    880 	/*
    881 	 * Request is done for better or worse, wakeup the top half.
    882 	 */
    883 	if (bp->b_flags & B_ERROR)
    884 		bp->b_resid = bp->b_bcount;
    885 	disk_unbusy(&cs->sc_dkdev, (bp->b_bcount - bp->b_resid),
    886 	    (bp->b_flags & B_READ));
    887 	biodone(bp);
    888 }
    889 
    890 /*
    891  * Called at interrupt time.
    892  * Mark the component as done and if all components are done,
    893  * take a ccd interrupt.
    894  */
    895 static void
    896 ccdiodone(struct buf *vbp)
    897 {
    898 	struct ccdbuf *cbp = (struct ccdbuf *) vbp;
    899 	struct buf *bp = cbp->cb_obp;
    900 	struct ccd_softc *cs = cbp->cb_sc;
    901 	int count, s;
    902 
    903 	s = splbio();
    904 #ifdef DEBUG
    905 	if (ccddebug & CCDB_FOLLOW)
    906 		printf("ccdiodone(%p)\n", cbp);
    907 	if (ccddebug & CCDB_IO) {
    908 		printf("ccdiodone: bp %p bcount %d resid %d\n",
    909 		       bp, bp->b_bcount, bp->b_resid);
    910 		printf(" dev 0x%x(u%d), cbp %p bn %" PRId64 " addr %p"
    911 		       " bcnt %d\n",
    912 		       cbp->cb_buf.b_dev, cbp->cb_comp, cbp,
    913 		       cbp->cb_buf.b_blkno, cbp->cb_buf.b_data,
    914 		       cbp->cb_buf.b_bcount);
    915 	}
    916 #endif
    917 
    918 	if (cbp->cb_buf.b_flags & B_ERROR) {
    919 		bp->b_flags |= B_ERROR;
    920 		bp->b_error = cbp->cb_buf.b_error ?
    921 		    cbp->cb_buf.b_error : EIO;
    922 
    923 		printf("%s: error %d on component %d\n",
    924 		       cs->sc_xname, bp->b_error, cbp->cb_comp);
    925 	}
    926 	count = cbp->cb_buf.b_bcount;
    927 	CCD_PUTBUF(cbp);
    928 
    929 	/*
    930 	 * If all done, "interrupt".
    931 	 */
    932 	bp->b_resid -= count;
    933 	if (bp->b_resid < 0)
    934 		panic("ccdiodone: count");
    935 	if (bp->b_resid == 0)
    936 		ccdintr(cs, bp);
    937 	splx(s);
    938 }
    939 
    940 /* ARGSUSED */
    941 static int
    942 ccdread(dev_t dev, struct uio *uio, int flags)
    943 {
    944 	int unit = ccdunit(dev);
    945 	struct ccd_softc *cs;
    946 
    947 #ifdef DEBUG
    948 	if (ccddebug & CCDB_FOLLOW)
    949 		printf("ccdread(0x%x, %p)\n", dev, uio);
    950 #endif
    951 	if (unit >= numccd)
    952 		return (ENXIO);
    953 	cs = &ccd_softc[unit];
    954 
    955 	if ((cs->sc_flags & CCDF_INITED) == 0)
    956 		return (ENXIO);
    957 
    958 	return (physio(ccdstrategy, NULL, dev, B_READ, minphys, uio));
    959 }
    960 
    961 /* ARGSUSED */
    962 static int
    963 ccdwrite(dev_t dev, struct uio *uio, int flags)
    964 {
    965 	int unit = ccdunit(dev);
    966 	struct ccd_softc *cs;
    967 
    968 #ifdef DEBUG
    969 	if (ccddebug & CCDB_FOLLOW)
    970 		printf("ccdwrite(0x%x, %p)\n", dev, uio);
    971 #endif
    972 	if (unit >= numccd)
    973 		return (ENXIO);
    974 	cs = &ccd_softc[unit];
    975 
    976 	if ((cs->sc_flags & CCDF_INITED) == 0)
    977 		return (ENXIO);
    978 
    979 	return (physio(ccdstrategy, NULL, dev, B_WRITE, minphys, uio));
    980 }
    981 
    982 static int
    983 ccdioctl(dev_t dev, u_long cmd, caddr_t data, int flag, struct proc *p)
    984 {
    985 	int unit = ccdunit(dev);
    986 	int s, i, j, lookedup = 0, error;
    987 	int part, pmask;
    988 	struct ccd_softc *cs;
    989 	struct ccd_ioctl *ccio = (struct ccd_ioctl *)data;
    990 	struct ucred *uc;
    991 	char **cpp;
    992 	struct vnode **vpp;
    993 #ifdef __HAVE_OLD_DISKLABEL
    994 	struct disklabel newlabel;
    995 #endif
    996 
    997 	if (unit >= numccd)
    998 		return (ENXIO);
    999 	cs = &ccd_softc[unit];
   1000 
   1001 	/* Must be open for writes for these commands... */
   1002 	switch (cmd) {
   1003 	case CCDIOCSET:
   1004 	case CCDIOCCLR:
   1005 	case DIOCSDINFO:
   1006 	case DIOCWDINFO:
   1007 #ifdef __HAVE_OLD_DISKLABEL
   1008 	case ODIOCSDINFO:
   1009 	case ODIOCWDINFO:
   1010 #endif
   1011 	case DIOCKLABEL:
   1012 	case DIOCWLABEL:
   1013 		if ((flag & FWRITE) == 0)
   1014 			return (EBADF);
   1015 	}
   1016 
   1017 	if ((error = lockmgr(&cs->sc_lock, LK_EXCLUSIVE, NULL)) != 0)
   1018 		return (error);
   1019 
   1020 	/* Must be initialized for these... */
   1021 	switch (cmd) {
   1022 	case CCDIOCCLR:
   1023 	case DIOCGDINFO:
   1024 	case DIOCCACHESYNC:
   1025 	case DIOCSDINFO:
   1026 	case DIOCWDINFO:
   1027 	case DIOCGPART:
   1028 	case DIOCWLABEL:
   1029 	case DIOCKLABEL:
   1030 	case DIOCGDEFLABEL:
   1031 #ifdef __HAVE_OLD_DISKLABEL
   1032 	case ODIOCGDINFO:
   1033 	case ODIOCSDINFO:
   1034 	case ODIOCWDINFO:
   1035 	case ODIOCGDEFLABEL:
   1036 #endif
   1037 		if ((cs->sc_flags & CCDF_INITED) == 0) {
   1038 			error = ENXIO;
   1039 			goto out;
   1040 		}
   1041 	}
   1042 
   1043 	switch (cmd) {
   1044 	case CCDIOCSET:
   1045 		if (cs->sc_flags & CCDF_INITED) {
   1046 			error = EBUSY;
   1047 			goto out;
   1048 		}
   1049 
   1050 		/* Validate the flags. */
   1051 		if ((ccio->ccio_flags & CCDF_USERMASK) != ccio->ccio_flags) {
   1052 			error = EINVAL;
   1053 			goto out;
   1054 		}
   1055 
   1056 		if (ccio->ccio_ndisks > CCD_MAXNDISKS) {
   1057 			error = EINVAL;
   1058 			goto out;
   1059 		}
   1060 
   1061 		/* Fill in some important bits. */
   1062 		cs->sc_ileave = ccio->ccio_ileave;
   1063 		cs->sc_nccdisks = ccio->ccio_ndisks;
   1064 		cs->sc_flags = ccio->ccio_flags & CCDF_USERMASK;
   1065 
   1066 		/*
   1067 		 * Allocate space for and copy in the array of
   1068 		 * componet pathnames and device numbers.
   1069 		 */
   1070 		cpp = malloc(ccio->ccio_ndisks * sizeof(char *),
   1071 		    M_DEVBUF, M_WAITOK);
   1072 		vpp = malloc(ccio->ccio_ndisks * sizeof(struct vnode *),
   1073 		    M_DEVBUF, M_WAITOK);
   1074 
   1075 		error = copyin(ccio->ccio_disks, cpp,
   1076 		    ccio->ccio_ndisks * sizeof(char **));
   1077 		if (error) {
   1078 			free(vpp, M_DEVBUF);
   1079 			free(cpp, M_DEVBUF);
   1080 			goto out;
   1081 		}
   1082 
   1083 #ifdef DEBUG
   1084 		if (ccddebug & CCDB_INIT)
   1085 			for (i = 0; i < ccio->ccio_ndisks; ++i)
   1086 				printf("ccdioctl: component %d: %p\n",
   1087 				    i, cpp[i]);
   1088 #endif
   1089 
   1090 		for (i = 0; i < ccio->ccio_ndisks; ++i) {
   1091 #ifdef DEBUG
   1092 			if (ccddebug & CCDB_INIT)
   1093 				printf("ccdioctl: lookedup = %d\n", lookedup);
   1094 #endif
   1095 			if ((error = ccdlookup(cpp[i], p, &vpp[i])) != 0) {
   1096 				for (j = 0; j < lookedup; ++j)
   1097 					(void)vn_close(vpp[j], FREAD|FWRITE,
   1098 					    p->p_ucred, p);
   1099 				free(vpp, M_DEVBUF);
   1100 				free(cpp, M_DEVBUF);
   1101 				goto out;
   1102 			}
   1103 			++lookedup;
   1104 		}
   1105 
   1106 		/*
   1107 		 * Initialize the ccd.  Fills in the softc for us.
   1108 		 */
   1109 		if ((error = ccdinit(cs, cpp, vpp, p)) != 0) {
   1110 			for (j = 0; j < lookedup; ++j)
   1111 				(void)vn_close(vpp[j], FREAD|FWRITE,
   1112 				    p->p_ucred, p);
   1113 			free(vpp, M_DEVBUF);
   1114 			free(cpp, M_DEVBUF);
   1115 			goto out;
   1116 		}
   1117 
   1118 		/* We can free the temporary variables now. */
   1119 		free(vpp, M_DEVBUF);
   1120 		free(cpp, M_DEVBUF);
   1121 
   1122 		/*
   1123 		 * The ccd has been successfully initialized, so
   1124 		 * we can place it into the array.  Don't try to
   1125 		 * read the disklabel until the disk has been attached,
   1126 		 * because space for the disklabel is allocated
   1127 		 * in disk_attach();
   1128 		 */
   1129 		ccio->ccio_unit = unit;
   1130 		ccio->ccio_size = cs->sc_size;
   1131 
   1132 		bufq_alloc(&cs->sc_bufq, "fcfs", 0);
   1133 
   1134 		/* Attach the disk. */
   1135 		pseudo_disk_attach(&cs->sc_dkdev);
   1136 
   1137 		/* Try and read the disklabel. */
   1138 		ccdgetdisklabel(dev);
   1139 		break;
   1140 
   1141 	case CCDIOCCLR:
   1142 		/*
   1143 		 * Don't unconfigure if any other partitions are open
   1144 		 * or if both the character and block flavors of this
   1145 		 * partition are open.
   1146 		 */
   1147 		part = DISKPART(dev);
   1148 		pmask = (1 << part);
   1149 		if ((cs->sc_dkdev.dk_openmask & ~pmask) ||
   1150 		    ((cs->sc_dkdev.dk_bopenmask & pmask) &&
   1151 		    (cs->sc_dkdev.dk_copenmask & pmask))) {
   1152 			error = EBUSY;
   1153 			goto out;
   1154 		}
   1155 
   1156 		/* Kill off any queued buffers. */
   1157 		s = splbio();
   1158 		bufq_drain(cs->sc_bufq);
   1159 		splx(s);
   1160 
   1161 		bufq_free(cs->sc_bufq);
   1162 
   1163 		/*
   1164 		 * Free ccd_softc information and clear entry.
   1165 		 */
   1166 
   1167 		/* Close the components and free their pathnames. */
   1168 		for (i = 0; i < cs->sc_nccdisks; ++i) {
   1169 			/*
   1170 			 * XXX: this close could potentially fail and
   1171 			 * cause Bad Things.  Maybe we need to force
   1172 			 * the close to happen?
   1173 			 */
   1174 #ifdef DEBUG
   1175 			if (ccddebug & CCDB_VNODE)
   1176 				vprint("CCDIOCCLR: vnode info",
   1177 				    cs->sc_cinfo[i].ci_vp);
   1178 #endif
   1179 			(void)vn_close(cs->sc_cinfo[i].ci_vp, FREAD|FWRITE,
   1180 			    p->p_ucred, p);
   1181 			free(cs->sc_cinfo[i].ci_path, M_DEVBUF);
   1182 		}
   1183 
   1184 		/* Free interleave index. */
   1185 		for (i = 0; cs->sc_itable[i].ii_ndisk; ++i)
   1186 			free(cs->sc_itable[i].ii_index, M_DEVBUF);
   1187 
   1188 		/* Free component info and interleave table. */
   1189 		free(cs->sc_cinfo, M_DEVBUF);
   1190 		free(cs->sc_itable, M_DEVBUF);
   1191 		cs->sc_flags &= ~(CCDF_INITED|CCDF_VLABEL);
   1192 
   1193 		/* Detatch the disk. */
   1194 		pseudo_disk_detach(&cs->sc_dkdev);
   1195 		break;
   1196 
   1197 	case DIOCGDINFO:
   1198 		*(struct disklabel *)data = *(cs->sc_dkdev.dk_label);
   1199 		break;
   1200 #ifdef __HAVE_OLD_DISKLABEL
   1201 	case ODIOCGDINFO:
   1202 		newlabel = *(cs->sc_dkdev.dk_label);
   1203 		if (newlabel.d_npartitions > OLDMAXPARTITIONS)
   1204 			return ENOTTY;
   1205 		memcpy(data, &newlabel, sizeof (struct olddisklabel));
   1206 		break;
   1207 #endif
   1208 
   1209 	case DIOCGPART:
   1210 		((struct partinfo *)data)->disklab = cs->sc_dkdev.dk_label;
   1211 		((struct partinfo *)data)->part =
   1212 		    &cs->sc_dkdev.dk_label->d_partitions[DISKPART(dev)];
   1213 		break;
   1214 
   1215 	case DIOCCACHESYNC:
   1216 		/*
   1217 		 * XXX Do we really need to care about having a writable
   1218 		 * file descriptor here?
   1219 		 */
   1220 		if ((flag & FWRITE) == 0)
   1221 			return (EBADF);
   1222 
   1223 		/*
   1224 		 * We pass this call down to all components and report
   1225 		 * the first error we encounter.
   1226 		 */
   1227 		uc = (p != NULL) ? p->p_ucred : NOCRED;
   1228 		for (error = 0, i = 0; i < cs->sc_nccdisks; i++) {
   1229 			j = VOP_IOCTL(cs->sc_cinfo[i].ci_vp, cmd, data,
   1230 				      flag, uc, p);
   1231 			if (j != 0 && error == 0)
   1232 				error = j;
   1233 		}
   1234 		break;
   1235 
   1236 	case DIOCWDINFO:
   1237 	case DIOCSDINFO:
   1238 #ifdef __HAVE_OLD_DISKLABEL
   1239 	case ODIOCWDINFO:
   1240 	case ODIOCSDINFO:
   1241 #endif
   1242 	{
   1243 		struct disklabel *lp;
   1244 #ifdef __HAVE_OLD_DISKLABEL
   1245 		if (cmd == ODIOCSDINFO || cmd == ODIOCWDINFO) {
   1246 			memset(&newlabel, 0, sizeof newlabel);
   1247 			memcpy(&newlabel, data, sizeof (struct olddisklabel));
   1248 			lp = &newlabel;
   1249 		} else
   1250 #endif
   1251 		lp = (struct disklabel *)data;
   1252 
   1253 		cs->sc_flags |= CCDF_LABELLING;
   1254 
   1255 		error = setdisklabel(cs->sc_dkdev.dk_label,
   1256 		    lp, 0, cs->sc_dkdev.dk_cpulabel);
   1257 		if (error == 0) {
   1258 			if (cmd == DIOCWDINFO
   1259 #ifdef __HAVE_OLD_DISKLABEL
   1260 			    || cmd == ODIOCWDINFO
   1261 #endif
   1262 			   )
   1263 				error = writedisklabel(CCDLABELDEV(dev),
   1264 				    ccdstrategy, cs->sc_dkdev.dk_label,
   1265 				    cs->sc_dkdev.dk_cpulabel);
   1266 		}
   1267 
   1268 		cs->sc_flags &= ~CCDF_LABELLING;
   1269 		break;
   1270 	}
   1271 
   1272 	case DIOCKLABEL:
   1273 		if (*(int *)data != 0)
   1274 			cs->sc_flags |= CCDF_KLABEL;
   1275 		else
   1276 			cs->sc_flags &= ~CCDF_KLABEL;
   1277 		break;
   1278 
   1279 	case DIOCWLABEL:
   1280 		if (*(int *)data != 0)
   1281 			cs->sc_flags |= CCDF_WLABEL;
   1282 		else
   1283 			cs->sc_flags &= ~CCDF_WLABEL;
   1284 		break;
   1285 
   1286 	case DIOCGDEFLABEL:
   1287 		ccdgetdefaultlabel(cs, (struct disklabel *)data);
   1288 		break;
   1289 
   1290 #ifdef __HAVE_OLD_DISKLABEL
   1291 	case ODIOCGDEFLABEL:
   1292 		ccdgetdefaultlabel(cs, &newlabel);
   1293 		if (newlabel.d_npartitions > OLDMAXPARTITIONS)
   1294 			return ENOTTY;
   1295 		memcpy(data, &newlabel, sizeof (struct olddisklabel));
   1296 		break;
   1297 #endif
   1298 
   1299 	default:
   1300 		error = ENOTTY;
   1301 	}
   1302 
   1303  out:
   1304 	(void) lockmgr(&cs->sc_lock, LK_RELEASE, NULL);
   1305 	return (error);
   1306 }
   1307 
   1308 static int
   1309 ccdsize(dev_t dev)
   1310 {
   1311 	struct ccd_softc *cs;
   1312 	struct disklabel *lp;
   1313 	int part, unit, omask, size;
   1314 
   1315 	unit = ccdunit(dev);
   1316 	if (unit >= numccd)
   1317 		return (-1);
   1318 	cs = &ccd_softc[unit];
   1319 
   1320 	if ((cs->sc_flags & CCDF_INITED) == 0)
   1321 		return (-1);
   1322 
   1323 	part = DISKPART(dev);
   1324 	omask = cs->sc_dkdev.dk_openmask & (1 << part);
   1325 	lp = cs->sc_dkdev.dk_label;
   1326 
   1327 	if (omask == 0 && ccdopen(dev, 0, S_IFBLK, curproc))
   1328 		return (-1);
   1329 
   1330 	if (lp->d_partitions[part].p_fstype != FS_SWAP)
   1331 		size = -1;
   1332 	else
   1333 		size = lp->d_partitions[part].p_size *
   1334 		    (lp->d_secsize / DEV_BSIZE);
   1335 
   1336 	if (omask == 0 && ccdclose(dev, 0, S_IFBLK, curproc))
   1337 		return (-1);
   1338 
   1339 	return (size);
   1340 }
   1341 
   1342 static int
   1343 ccddump(dev_t dev, daddr_t blkno, caddr_t va, size_t size)
   1344 {
   1345 
   1346 	/* Not implemented. */
   1347 	return ENXIO;
   1348 }
   1349 
   1350 /*
   1351  * Lookup the provided name in the filesystem.  If the file exists,
   1352  * is a valid block device, and isn't being used by anyone else,
   1353  * set *vpp to the file's vnode.
   1354  */
   1355 static int
   1356 ccdlookup(char *path, struct proc *p, struct vnode **vpp /* result */)
   1357 {
   1358 	struct nameidata nd;
   1359 	struct vnode *vp;
   1360 	struct vattr va;
   1361 	int error;
   1362 
   1363 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, path, p);
   1364 	if ((error = vn_open(&nd, FREAD|FWRITE, 0)) != 0) {
   1365 #ifdef DEBUG
   1366 		if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
   1367 			printf("ccdlookup: vn_open error = %d\n", error);
   1368 #endif
   1369 		return (error);
   1370 	}
   1371 	vp = nd.ni_vp;
   1372 
   1373 	if (vp->v_usecount > 1) {
   1374 		VOP_UNLOCK(vp, 0);
   1375 		(void)vn_close(vp, FREAD|FWRITE, p->p_ucred, p);
   1376 		return (EBUSY);
   1377 	}
   1378 
   1379 	if ((error = VOP_GETATTR(vp, &va, p->p_ucred, p)) != 0) {
   1380 #ifdef DEBUG
   1381 		if (ccddebug & (CCDB_FOLLOW|CCDB_INIT))
   1382 			printf("ccdlookup: getattr error = %d\n", error);
   1383 #endif
   1384 		VOP_UNLOCK(vp, 0);
   1385 		(void)vn_close(vp, FREAD|FWRITE, p->p_ucred, p);
   1386 		return (error);
   1387 	}
   1388 
   1389 	/* XXX: eventually we should handle VREG, too. */
   1390 	if (va.va_type != VBLK) {
   1391 		VOP_UNLOCK(vp, 0);
   1392 		(void)vn_close(vp, FREAD|FWRITE, p->p_ucred, p);
   1393 		return (ENOTBLK);
   1394 	}
   1395 
   1396 #ifdef DEBUG
   1397 	if (ccddebug & CCDB_VNODE)
   1398 		vprint("ccdlookup: vnode info", vp);
   1399 #endif
   1400 
   1401 	VOP_UNLOCK(vp, 0);
   1402 	*vpp = vp;
   1403 	return (0);
   1404 }
   1405 
   1406 static void
   1407 ccdgetdefaultlabel(struct ccd_softc *cs, struct disklabel *lp)
   1408 {
   1409 	struct ccdgeom *ccg = &cs->sc_geom;
   1410 
   1411 	memset(lp, 0, sizeof(*lp));
   1412 
   1413 	lp->d_secperunit = cs->sc_size;
   1414 	lp->d_secsize = ccg->ccg_secsize;
   1415 	lp->d_nsectors = ccg->ccg_nsectors;
   1416 	lp->d_ntracks = ccg->ccg_ntracks;
   1417 	lp->d_ncylinders = ccg->ccg_ncylinders;
   1418 	lp->d_secpercyl = lp->d_ntracks * lp->d_nsectors;
   1419 
   1420 	strncpy(lp->d_typename, "ccd", sizeof(lp->d_typename));
   1421 	lp->d_type = DTYPE_CCD;
   1422 	strncpy(lp->d_packname, "fictitious", sizeof(lp->d_packname));
   1423 	lp->d_rpm = 3600;
   1424 	lp->d_interleave = 1;
   1425 	lp->d_flags = 0;
   1426 
   1427 	lp->d_partitions[RAW_PART].p_offset = 0;
   1428 	lp->d_partitions[RAW_PART].p_size = cs->sc_size;
   1429 	lp->d_partitions[RAW_PART].p_fstype = FS_UNUSED;
   1430 	lp->d_npartitions = RAW_PART + 1;
   1431 
   1432 	lp->d_magic = DISKMAGIC;
   1433 	lp->d_magic2 = DISKMAGIC;
   1434 	lp->d_checksum = dkcksum(cs->sc_dkdev.dk_label);
   1435 }
   1436 
   1437 /*
   1438  * Read the disklabel from the ccd.  If one is not present, fake one
   1439  * up.
   1440  */
   1441 static void
   1442 ccdgetdisklabel(dev_t dev)
   1443 {
   1444 	int unit = ccdunit(dev);
   1445 	struct ccd_softc *cs = &ccd_softc[unit];
   1446 	const char *errstring;
   1447 	struct disklabel *lp = cs->sc_dkdev.dk_label;
   1448 	struct cpu_disklabel *clp = cs->sc_dkdev.dk_cpulabel;
   1449 
   1450 	memset(clp, 0, sizeof(*clp));
   1451 
   1452 	ccdgetdefaultlabel(cs, lp);
   1453 
   1454 	/*
   1455 	 * Call the generic disklabel extraction routine.
   1456 	 */
   1457 	if ((cs->sc_flags & CCDF_NOLABEL) != 0)
   1458 		errstring = "CCDF_NOLABEL set; ignoring on-disk label";
   1459 	else
   1460 		errstring = readdisklabel(CCDLABELDEV(dev), ccdstrategy,
   1461 		    cs->sc_dkdev.dk_label, cs->sc_dkdev.dk_cpulabel);
   1462 	if (errstring)
   1463 		ccdmakedisklabel(cs);
   1464 	else {
   1465 		int i;
   1466 		struct partition *pp;
   1467 
   1468 		/*
   1469 		 * Sanity check whether the found disklabel is valid.
   1470 		 *
   1471 		 * This is necessary since total size of ccd may vary
   1472 		 * when an interleave is changed even though exactly
   1473 		 * same componets are used, and old disklabel may used
   1474 		 * if that is found.
   1475 		 */
   1476 		if (lp->d_secperunit != cs->sc_size)
   1477 			printf("WARNING: %s: "
   1478 			    "total sector size in disklabel (%d) != "
   1479 			    "the size of ccd (%lu)\n", cs->sc_xname,
   1480 			    lp->d_secperunit, (u_long)cs->sc_size);
   1481 		for (i = 0; i < lp->d_npartitions; i++) {
   1482 			pp = &lp->d_partitions[i];
   1483 			if (pp->p_offset + pp->p_size > cs->sc_size)
   1484 				printf("WARNING: %s: end of partition `%c' "
   1485 				    "exceeds the size of ccd (%lu)\n",
   1486 				    cs->sc_xname, 'a' + i, (u_long)cs->sc_size);
   1487 		}
   1488 	}
   1489 
   1490 #ifdef DEBUG
   1491 	/* It's actually extremely common to have unlabeled ccds. */
   1492 	if (ccddebug & CCDB_LABEL)
   1493 		if (errstring != NULL)
   1494 			printf("%s: %s\n", cs->sc_xname, errstring);
   1495 #endif
   1496 
   1497 	/* In-core label now valid. */
   1498 	cs->sc_flags |= CCDF_VLABEL;
   1499 }
   1500 
   1501 /*
   1502  * Take care of things one might want to take care of in the event
   1503  * that a disklabel isn't present.
   1504  */
   1505 static void
   1506 ccdmakedisklabel(struct ccd_softc *cs)
   1507 {
   1508 	struct disklabel *lp = cs->sc_dkdev.dk_label;
   1509 
   1510 	/*
   1511 	 * For historical reasons, if there's no disklabel present
   1512 	 * the raw partition must be marked FS_BSDFFS.
   1513 	 */
   1514 	lp->d_partitions[RAW_PART].p_fstype = FS_BSDFFS;
   1515 
   1516 	strncpy(lp->d_packname, "default label", sizeof(lp->d_packname));
   1517 
   1518 	lp->d_checksum = dkcksum(lp);
   1519 }
   1520 
   1521 #ifdef DEBUG
   1522 static void
   1523 printiinfo(struct ccdiinfo *ii)
   1524 {
   1525 	int ix, i;
   1526 
   1527 	for (ix = 0; ii->ii_ndisk; ix++, ii++) {
   1528 		printf(" itab[%d]: #dk %d sblk %" PRId64 " soff %" PRId64,
   1529 		    ix, ii->ii_ndisk, ii->ii_startblk, ii->ii_startoff);
   1530 		for (i = 0; i < ii->ii_ndisk; i++)
   1531 			printf(" %d", ii->ii_index[i]);
   1532 		printf("\n");
   1533 	}
   1534 }
   1535 #endif
   1536