procfs_vnops.c

来自「早期freebsd实现」· C语言 代码 · 共 815 行 · 第 1/2 页

/*
 * Copyright (c) 1993 Jan-Simon Pendry
 * Copyright (c) 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Jan-Simon Pendry.
 *
 * 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.
 *
 *	@(#)procfs_vnops.c	8.6 (Berkeley) 2/7/94
 *
 * From:
 *	$Id: procfs_vnops.c,v 3.2 1993/12/15 09:40:17 jsp Exp $
 */

/*
 * procfs vnode interface
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/time.h>
#include <sys/kernel.h>
#include <sys/file.h>
#include <sys/proc.h>
#include <sys/vnode.h>
#include <sys/namei.h>
#include <sys/malloc.h>
#include <sys/dirent.h>
#include <sys/resourcevar.h>
#include <miscfs/procfs/procfs.h>
#include <vm/vm.h>	/* for PAGE_SIZE */

/*
 * Vnode Operations.
 *
 */

/*
 * This is a list of the valid names in the
 * process-specific sub-directories.  It is
 * used in procfs_lookup and procfs_readdir
 */
static struct pfsnames {
	u_short	d_namlen;
	char	d_name[PROCFS_NAMELEN];
	pfstype	d_pfstype;
} procent[] = {
#define N(s) sizeof(s)-1, s
	/* namlen, nam, type */
	{  N("file"),	Pfile },
	{  N("mem"),	Pmem },
	{  N("regs"),	Pregs },
	{  N("fpregs"),	Pfpregs },
	{  N("ctl"),	Pctl },
	{  N("status"),	Pstatus },
	{  N("note"),	Pnote },
	{  N("notepg"),	Pnotepg },
#undef N
};
#define Nprocent (sizeof(procent)/sizeof(procent[0]))

static pid_t atopid __P((const char *, u_int));

/*
 * set things up for doing i/o on
 * the pfsnode (vp).  (vp) is locked
 * on entry, and should be left locked
 * on exit.
 *
 * for procfs we don't need to do anything
 * in particular for i/o.  all that is done
 * is to support exclusive open on process
 * memory images.
 */
procfs_open(ap)
	struct vop_open_args *ap;
{
	struct pfsnode *pfs = VTOPFS(ap->a_vp);

	switch (pfs->pfs_type) {
	case Pmem:
		if (PFIND(pfs->pfs_pid) == 0)
			return (ENOENT);	/* was ESRCH, jsp */

		if ((pfs->pfs_flags & FWRITE) && (ap->a_mode & O_EXCL) ||
				(pfs->pfs_flags & O_EXCL) && (ap->a_mode & FWRITE))
			return (EBUSY);


		if (ap->a_mode & FWRITE)
			pfs->pfs_flags = ap->a_mode & (FWRITE|O_EXCL);

		return (0);

	default:
		break;
	}

	return (0);
}

/*
 * close the pfsnode (vp) after doing i/o.
 * (vp) is not locked on entry or exit.
 *
 * nothing to do for procfs other than undo
 * any exclusive open flag (see _open above).
 */
procfs_close(ap)
	struct vop_close_args *ap;
{
	struct pfsnode *pfs = VTOPFS(ap->a_vp);

	switch (pfs->pfs_type) {
	case Pmem:
		if ((ap->a_fflag & FWRITE) && (pfs->pfs_flags & O_EXCL))
			pfs->pfs_flags &= ~(FWRITE|O_EXCL);
		break;
	}

	return (0);
}

/*
 * do an ioctl operation on pfsnode (vp).
 * (vp) is not locked on entry or exit.
 */
procfs_ioctl(ap)
	struct vop_ioctl_args *ap;
{

	return (ENOTTY);
}

/*
 * do block mapping for pfsnode (vp).
 * since we don't use the buffer cache
 * for procfs this function should never
 * be called.  in any case, it's not clear
 * what part of the kernel ever makes use
 * of this function.  for sanity, this is the
 * usual no-op bmap, although returning
 * (EIO) would be a reasonable alternative.
 */
procfs_bmap(ap)
	struct vop_bmap_args *ap;
{

	if (ap->a_vpp != NULL)
		*ap->a_vpp = ap->a_vp;
	if (ap->a_bnp != NULL)
		*ap->a_bnp = ap->a_bn;
	return (0);
}

/*
 * _inactive is called when the pfsnode
 * is vrele'd and the reference count goes
 * to zero.  (vp) will be on the vnode free
 * list, so to get it back vget() must be
 * used.
 *
 * for procfs, check if the process is still
 * alive and if it isn't then just throw away
 * the vnode by calling vgone().  this may
 * be overkill and a waste of time since the
 * chances are that the process will still be
 * there and PFIND is not free.
 *
 * (vp) is not locked on entry or exit.
 */
procfs_inactive(ap)
	struct vop_inactive_args *ap;
{
	struct pfsnode *pfs = VTOPFS(ap->a_vp);

	if (PFIND(pfs->pfs_pid) == 0)
		vgone(ap->a_vp);

	return (0);
}

/*
 * _reclaim is called when getnewvnode()
 * wants to make use of an entry on the vnode
 * free list.  at this time the filesystem needs
 * to free any private data and remove the node
 * from any private lists.
 */
procfs_reclaim(ap)
	struct vop_reclaim_args *ap;
{
	int error;

	error = procfs_freevp(ap->a_vp);
	return (error);
}

/*
 * Return POSIX pathconf information applicable to special devices.
 */
procfs_pathconf(ap)
	struct vop_pathconf_args /* {
		struct vnode *a_vp;
		int a_name;
		int *a_retval;
	} */ *ap;
{

	switch (ap->a_name) {
	case _PC_LINK_MAX:
		*ap->a_retval = LINK_MAX;
		return (0);
	case _PC_MAX_CANON:
		*ap->a_retval = MAX_CANON;
		return (0);
	case _PC_MAX_INPUT:
		*ap->a_retval = MAX_INPUT;
		return (0);
	case _PC_PIPE_BUF:
		*ap->a_retval = PIPE_BUF;
		return (0);
	case _PC_CHOWN_RESTRICTED:
		*ap->a_retval = 1;
		return (0);
	case _PC_VDISABLE:
		*ap->a_retval = _POSIX_VDISABLE;
		return (0);
	default:
		return (EINVAL);
	}
	/* NOTREACHED */
}

/*
 * _print is used for debugging.
 * just print a readable description
 * of (vp).
 */
procfs_print(ap)
	struct vop_print_args *ap;
{
	struct pfsnode *pfs = VTOPFS(ap->a_vp);

	printf("tag VT_PROCFS, pid %d, mode %x, flags %x\n",
		pfs->pfs_pid,
		pfs->pfs_mode, pfs->pfs_flags);
}

/*
 * _abortop is called when operations such as
 * rename and create fail.  this entry is responsible
 * for undoing any side-effects caused by the lookup.
 * this will always include freeing the pathname buffer.
 */
procfs_abortop(ap)
	struct vop_abortop_args *ap;
{

	if ((ap->a_cnp->cn_flags & (HASBUF | SAVESTART)) == HASBUF)
		FREE(ap->a_cnp->cn_pnbuf, M_NAMEI);
	return (0);
}

/*
 * generic entry point for unsupported operations
 */
procfs_badop()
{

	return (EIO);
}

/*
 * Invent attributes for pfsnode (vp) and store
 * them in (vap).
 * Directories lengths are returned as zero since
 * any real length would require the genuine size
 * to be computed, and nothing cares anyway.
 *
 * this is relatively minimal for procfs.
 */
procfs_getattr(ap)
	struct vop_getattr_args *ap;
{
	struct pfsnode *pfs = VTOPFS(ap->a_vp);
	struct vattr *vap = ap->a_vap;
	struct proc *procp;
	int error;

	/* first check the process still exists */
	switch (pfs->pfs_type) {
	case Proot:
		procp = 0;
		break;

	default:
		procp = PFIND(pfs->pfs_pid);
		if (procp == 0)
			return (ENOENT);
	}

	error = 0;

	/* start by zeroing out the attributes */
	VATTR_NULL(vap);

	/* next do all the common fields */
	vap->va_type = ap->a_vp->v_type;
	vap->va_mode = pfs->pfs_mode;
	vap->va_fileid = pfs->pfs_fileno;
	vap->va_flags = 0;
	vap->va_blocksize = PAGE_SIZE;
	vap->va_bytes = vap->va_size = 0;

	/*
	 * If the process has exercised some setuid or setgid
	 * privilege, then rip away read/write permission so
	 * that only root can gain access.
	 */
	switch (pfs->pfs_type) {
	case Pregs:
	case Pfpregs:
	case Pmem:
		if (procp->p_flag & P_SUGID)
			vap->va_mode &= ~((VREAD|VWRITE)|
					  ((VREAD|VWRITE)>>3)|
					  ((VREAD|VWRITE)>>6));
		break;
	}

	/*
	 * Make all times be current TOD.
	 * It would be possible to get the process start
	 * time from the p_stat structure, but there's
	 * no "file creation" time stamp anyway, and the
	 * p_stat structure is not addressible if u. gets
	 * swapped out for that process.
	 */
	microtime(&vap->va_ctime);
	vap->va_atime = vap->va_mtime = vap->va_ctime;

	/*
	 * now do the object specific fields
	 *
	 * The size could be set from struct reg, but it's hardly
	 * worth the trouble, and it puts some (potentially) machine
	 * dependent data into this machine-independent code.  If it
	 * becomes important then this function should break out into
	 * a per-file stat function in the corresponding .c file.
	 */

	switch (pfs->pfs_type) {
	case Proot:
		vap->va_nlink = 2;
		vap->va_uid = 0;
		vap->va_gid = 0;
		break;

	case Pproc:
		vap->va_nlink = 2;
		vap->va_uid = procp->p_ucred->cr_uid;
		vap->va_gid = procp->p_ucred->cr_gid;
		break;

	case Pfile:
		error = EOPNOTSUPP;
		break;

	case Pmem:
		vap->va_nlink = 1;
		vap->va_bytes = vap->va_size =
			ctob(procp->p_vmspace->vm_tsize +
				    procp->p_vmspace->vm_dsize +
				    procp->p_vmspace->vm_ssize);
		vap->va_uid = procp->p_ucred->cr_uid;
		vap->va_gid = procp->p_ucred->cr_gid;
		break;

	case Pregs: