vn.c

来自「基于组件方式开发操作系统的OSKIT源代码」· C语言 代码 · 共 686 行 · 第 1/2 页

/*
 * Copyright (c) 1988 University of Utah.
 * Copyright (c) 1990, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * the Systems Programming Group of the University of Utah Computer
 * Science Department.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 * from: Utah Hdr: vn.c 1.13 94/04/02
 *
 *	from: @(#)vn.c	8.6 (Berkeley) 4/1/94
 *	$Id: vn.c,v 1.71.2.1 1999/03/13 00:32:36 julian Exp $
 */

/*
 * Vnode disk driver.
 *
 * Block/character interface to a vnode.  Allows one to treat a file
 * as a disk (e.g. build a filesystem in it, mount it, etc.).
 *
 * NOTE 1: This uses the VOP_BMAP/VOP_STRATEGY interface to the vnode
 * instead of a simple VOP_RDWR.  We do this to avoid distorting the
 * local buffer cache.
 *
 * NOTE 2: There is a security issue involved with this driver.
 * Once mounted all access to the contents of the "mapped" file via
 * the special file is controlled by the permissions on the special
 * file, the protection of the mapped file is ignored (effectively,
 * by using root credentials in all transactions).
 *
 * NOTE 3: Doesn't interact with leases, should it?
 */
#include "vn.h"
#if NVN > 0

/* default is to have 8 VN's */
#if NVN < 8
#undef NVN
#define	NVN	8
#endif

#include "opt_devfs.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/namei.h>
#include <sys/proc.h>
#include <sys/buf.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/vnode.h>
#include <sys/fcntl.h>
#include <sys/disklabel.h>
#include <sys/diskslice.h>
#include <sys/stat.h>
#include <sys/conf.h>
#ifdef DEVFS
#include <sys/devfsext.h>
#endif /*DEVFS*/
#include <miscfs/specfs/specdev.h>
#include <sys/vnioctl.h>

static	d_ioctl_t	vnioctl;
static	d_open_t	vnopen;
static	d_read_t	vnread;
static	d_write_t	vnwrite;
static	d_close_t	vnclose;
static	d_dump_t	vndump;
static	d_psize_t	vnsize;
static	d_strategy_t	vnstrategy;

#define CDEV_MAJOR 43
#define BDEV_MAJOR 15


static struct cdevsw vn_cdevsw = {
	  vnopen,	vnclose,	vnread,		vnwrite,
	  vnioctl,	nostop,		nullreset,	nodevtotty,
	  seltrue,	nommap,		vnstrategy,	"vn",
	  NULL,		-1,		vndump,		vnsize,
	  D_DISK|D_NOCLUSTERRW,	0,	-1 };


#define	vnunit(dev)	dkunit(dev)

#define	getvnbuf()	\
	((struct buf *)malloc(sizeof(struct buf), M_DEVBUF, M_WAITOK))

#define putvnbuf(bp)	\
	free((caddr_t)(bp), M_DEVBUF)

struct vn_softc {
	int		 sc_flags;	/* flags */
	size_t		 sc_size;	/* size of vn */
	struct diskslices *sc_slices;
	struct vnode	*sc_vp;		/* vnode */
	struct ucred	*sc_cred;	/* credentials */
	int		 sc_maxactive;	/* max # of active requests */
	struct buf	 sc_tab;	/* transfer queue */
	u_long		 sc_options;	/* options */
#ifdef DEVFS
	void		*r_devfs_token;
	void		*devfs_token;
#endif
};

/* sc_flags */
#define VNF_INITED	0x01

static struct vn_softc *vn_softc[NVN];
static u_long	vn_options;

#define IFOPT(vn,opt) if (((vn)->sc_options|vn_options) & (opt))

static void	vniodone (struct buf *bp);
static int	vnsetcred (struct vn_softc *vn, struct ucred *cred);
static void	vnshutdown (int, void *);
static void	vnclear (struct vn_softc *vn);

static	int
vnclose(dev_t dev, int flags, int mode, struct proc *p)
{
	struct vn_softc *vn = vn_softc[vnunit(dev)];

	IFOPT(vn, VN_LABELS)
		if (vn->sc_slices != NULL)
			dsclose(dev, mode, vn->sc_slices);
	return (0);
}

static	int
vnopen(dev_t dev, int flags, int mode, struct proc *p)
{
	int unit = vnunit(dev);
	struct vn_softc *vn;

	if (unit >= NVN) {
		if (vn_options & VN_FOLLOW)
			printf("vnopen(0x%lx, 0x%x, 0x%x, %p)\n",
			    (u_long)dev, flags, mode, (void *)p);
		return(ENOENT);
	}

	vn = vn_softc[unit];
	if (!vn) {
		vn = malloc(sizeof *vn, M_DEVBUF, M_WAITOK);
		if (!vn)
			return (ENOMEM);
		bzero(vn, sizeof *vn);
		vn_softc[unit] = vn;
	}

	IFOPT(vn, VN_FOLLOW)
		printf("vnopen(0x%lx, 0x%x, 0x%x, %p)\n",
		    (u_long)dev, flags, mode, (void *)p);

	IFOPT(vn, VN_LABELS) {
		if (vn->sc_flags & VNF_INITED) {
			struct disklabel label;

			/* Build label for whole disk. */
			bzero(&label, sizeof label);
			label.d_secsize = DEV_BSIZE;
			label.d_nsectors = 32;
			label.d_ntracks = 64;
			label.d_ncylinders = vn->sc_size / (32 * 64);
			label.d_secpercyl = 32 * 64;
			label.d_secperunit =
					label.d_partitions[RAW_PART].p_size =
					vn->sc_size;

			return (dsopen("vn", dev, mode, 0, &vn->sc_slices,
				       &label, vnstrategy, (ds_setgeom_t *)NULL,
				       &vn_cdevsw));
		}
		if (dkslice(dev) != WHOLE_DISK_SLICE ||
		    dkpart(dev) != RAW_PART ||
		    mode != S_IFCHR)
			return (ENXIO);
	}
	return(0);
}

static int
vnread(dev_t dev, struct uio *uio, int ioflag)
{
	return (physio(vnstrategy, NULL, dev, 1, minphys, uio));
}

static int
vnwrite(dev_t dev, struct uio *uio, int ioflag)
{
	return (physio(vnstrategy, NULL, dev, 0, minphys, uio));
}

/*
 * this code does I/O calls through the appropriate VOP entry point...
 * unless a swap_pager I/O request is being done.  This strategy (-))
 * allows for coherency with mmap except in the case of paging.  This
 * is necessary, because the VOP calls use lots of memory (and actually
 * are not extremely efficient -- but we want to keep semantics correct),
 * and the pageout daemon gets really unhappy (and so does the rest of the
 * system) when it runs out of memory.
 */
static	void
vnstrategy(struct buf *bp)
{
	int unit = vnunit(bp->b_dev);
	register struct vn_softc *vn = vn_softc[unit];
	register daddr_t bn;
	int error;
	int isvplocked = 0;
	long sz;
	struct uio auio;
	struct iovec aiov;

	IFOPT(vn, VN_DEBUG)
		printf("vnstrategy(%p): unit %d\n", bp, unit);

	if ((vn->sc_flags & VNF_INITED) == 0) {
		bp->b_error = ENXIO;
		bp->b_flags |= B_ERROR;
		biodone(bp);
		return;
	}
	IFOPT(vn, VN_LABELS) {
		bp->b_resid = bp->b_bcount;/* XXX best place to set this? */
		if (vn->sc_slices != NULL && dscheck(bp, vn->sc_slices) <= 0) {
			biodone(bp);
			return;
		}
		bn = bp->b_pblkno;
		bp->b_resid = bp->b_bcount;/* XXX best place to set this? */
	} else {
		bn = bp->b_blkno;
		sz = howmany(bp->b_bcount, DEV_BSIZE);
		bp->b_resid = bp->b_bcount;
		if (bn < 0 || bn + sz > vn->sc_size) {
			if (bn != vn->sc_size) {
				bp->b_error = EINVAL;
				bp->b_flags |= B_ERROR;
			}
			biodone(bp);
			return;
		}
	}

	if( (bp->b_flags & B_PAGING) == 0) {
		aiov.iov_base = bp->b_data;
		aiov.iov_len = bp->b_bcount;
		auio.uio_iov = &aiov;
		auio.uio_iovcnt = 1;
		auio.uio_offset = dbtob(bn);
		auio.uio_segflg = UIO_SYSSPACE;
		if( bp->b_flags & B_READ)
			auio.uio_rw = UIO_READ;
		else
			auio.uio_rw = UIO_WRITE;
		auio.uio_resid = bp->b_bcount;
		auio.uio_procp = curproc;
		if (!VOP_ISLOCKED(vn->sc_vp)) {
			isvplocked = 1;
			vn_lock(vn->sc_vp, LK_EXCLUSIVE | LK_RETRY, curproc);
		}
		if( bp->b_flags & B_READ)
			error = VOP_READ(vn->sc_vp, &auio, 0, vn->sc_cred);
		else
			error = VOP_WRITE(vn->sc_vp, &auio, 0, vn->sc_cred);
		if (isvplocked) {
			VOP_UNLOCK(vn->sc_vp, 0, curproc);
			isvplocked = 0;
		}
		bp->b_resid = auio.uio_resid;

		if( error )
			bp->b_flags |= B_ERROR;
		biodone(bp);
	} else {
		long bsize, resid;
		off_t byten;
		int flags;
		caddr_t addr;
		struct buf *nbp;

		nbp = getvnbuf();
		bzero(nbp, sizeof(struct buf));
		LIST_INIT(&nbp->b_dep);
		byten = dbtob(bn);
		bsize = vn->sc_vp->v_mount->mnt_stat.f_iosize;
		addr = bp->b_data;
		flags = bp->b_flags | B_CALL;
		for (resid = bp->b_resid; resid; ) {
			struct vnode *vp;
			daddr_t nbn;
			int off, s, nra;

			nra = 0;
			if (!VOP_ISLOCKED(vn->sc_vp)) {
				isvplocked = 1;
				vn_lock(vn->sc_vp, LK_EXCLUSIVE | LK_RETRY, curproc);
			}
			error = VOP_BMAP(vn->sc_vp, (daddr_t)(byten / bsize),
					 &vp, &nbn, &nra, NULL);
			if (isvplocked) {
				VOP_UNLOCK(vn->sc_vp, 0, curproc);
				isvplocked = 0;
			}
			if (error == 0 && nbn == -1)
				error = EIO;

			IFOPT(vn, VN_DONTCLUSTER)
				nra = 0;