Use SYSREF to reference count struct vnode. v_usecount is now

[dragonfly.git] / sys / vfs / procfs / procfs_subr.c

/*
 * 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_subr.c	8.6 (Berkeley) 5/14/95
 *
 * $FreeBSD: src/sys/miscfs/procfs/procfs_subr.c,v 1.26.2.3 2002/02/18 21:28:04 des Exp $
 * $DragonFly: src/sys/vfs/procfs/procfs_subr.c,v 1.17 2007/05/06 19:23:35 dillon Exp $
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sysctl.h>
#include <sys/proc.h>
#include <sys/mount.h>
#include <sys/vnode.h>
#include <sys/malloc.h>

#include <vfs/procfs/procfs.h>

#define PFS_HSIZE	256
#define PFS_HMASK	(PFS_HSIZE - 1)

static struct pfsnode *pfshead[PFS_HSIZE];
static int pfsvplock;

#define PFSHASH(pid)	&pfshead[(pid) & PFS_HMASK]

/*
 * Allocate a pfsnode/vnode pair.  If no error occurs the returned vnode
 * will be referenced and exclusively locked.
 *
 * The pid, pfs_type, and mount point uniquely identify a pfsnode.
 * The mount point is needed because someone might mount this filesystem
 * twice.
 *
 * All pfsnodes are maintained on a singly-linked list.  new nodes are
 * only allocated when they cannot be found on this list.  entries on
 * the list are removed when the vfs reclaim entry is called.
 *
 * A single lock is kept for the entire list.  this is needed because the
 * getnewvnode() function can block waiting for a vnode to become free,
 * in which case there may be more than one process trying to get the same
 * vnode.  this lock is only taken if we are going to call getnewvnode, 
 * since the kernel itself is single-threaded.
 *
 * If an entry is found on the list, then call vget() to take a reference
 * and obtain the lock.  This will properly re-reference the vnode if it
 * had gotten onto the free list.
 */
int
procfs_allocvp(struct mount *mp, struct vnode **vpp, long pid, pfstype pfs_type)
{
	struct pfsnode *pfs;
	struct vnode *vp;
	struct pfsnode **pp;
	int error;

	pp = PFSHASH(pid);
loop:
	for (pfs = *pp; pfs; pfs = pfs->pfs_next) {
		if (pfs->pfs_pid == pid && pfs->pfs_type == pfs_type &&
		    PFSTOV(pfs)->v_mount == mp) {
			vp = PFSTOV(pfs);
			if (vget(vp, LK_EXCLUSIVE))
				goto loop;

			/*
			 * Make sure the vnode is still in the cache after
			 * getting the interlock to avoid racing a free.
			 */
			for (pfs = *pp; pfs; pfs = pfs->pfs_next) {
				if (PFSTOV(pfs) == vp &&
				    pfs->pfs_pid == pid && 
				    pfs->pfs_type == pfs_type &&
				    PFSTOV(pfs)->v_mount == mp) {
					break;
				}
			}
			if (pfs == NULL || PFSTOV(pfs) != vp) {
				vput(vp);
				goto loop;

			}
			*vpp = vp;
			return (0);
		}
	}

	/*
	 * otherwise lock the vp list while we call getnewvnode
	 * since that can block.
	 */
	if (pfsvplock & PROCFS_LOCKED) {
		pfsvplock |= PROCFS_WANT;
		(void) tsleep((caddr_t) &pfsvplock, 0, "pfsavp", 0);
		goto loop;
	}
	pfsvplock |= PROCFS_LOCKED;

	/*
	 * Do the MALLOC before the getnewvnode since doing so afterward
	 * might cause a bogus v_data pointer to get dereferenced
	 * elsewhere if MALLOC should block.
	 *
	 * XXX this may not matter anymore since getnewvnode now returns
	 * a VX locked vnode.
	 */
	MALLOC(pfs, struct pfsnode *, sizeof(struct pfsnode), M_TEMP, M_WAITOK);

	error = getnewvnode(VT_PROCFS, mp, vpp, 0, 0);
	if (error) {
		kfree(pfs, M_TEMP);
		goto out;
	}
	vp = *vpp;

	vp->v_data = pfs;

	pfs->pfs_next = 0;
	pfs->pfs_pid = (pid_t) pid;
	pfs->pfs_type = pfs_type;
	pfs->pfs_vnode = vp;
	pfs->pfs_flags = 0;
	pfs->pfs_lockowner = 0;
	pfs->pfs_fileno = PROCFS_FILENO(pid, pfs_type);

	switch (pfs_type) {
	case Proot:	/* /proc = dr-xr-xr-x */
		pfs->pfs_mode = (VREAD|VEXEC) |
				(VREAD|VEXEC) >> 3 |
				(VREAD|VEXEC) >> 6;
		vp->v_type = VDIR;
		vp->v_flag = VROOT;
		break;

	case Pcurproc:	/* /proc/curproc = lr--r--r-- */
		pfs->pfs_mode = (VREAD) |
				(VREAD >> 3) |
				(VREAD >> 6);
		vp->v_type = VLNK;
		break;

	case Pproc:
		pfs->pfs_mode = (VREAD|VEXEC) |
				(VREAD|VEXEC) >> 3 |
				(VREAD|VEXEC) >> 6;
		vp->v_type = VDIR;
		break;

	case Pfile:
		pfs->pfs_mode = (VREAD|VEXEC) |
				(VREAD|VEXEC) >> 3 |
				(VREAD|VEXEC) >> 6;
		vp->v_type = VLNK;
		break;

	case Pmem:
		pfs->pfs_mode = (VREAD|VWRITE);
		vp->v_type = VREG;
		break;

	case Pregs:
	case Pfpregs:
	case Pdbregs:
		pfs->pfs_mode = (VREAD|VWRITE);
		vp->v_type = VREG;
		break;

	case Pctl:
	case Pnote:
	case Pnotepg:
		pfs->pfs_mode = (VWRITE);
		vp->v_type = VREG;
		break;

	case Ptype:
	case Pmap:
	case Pstatus:
	case Pcmdline:
	case Prlimit:
		pfs->pfs_mode = (VREAD) |
				(VREAD >> 3) |
				(VREAD >> 6);
		vp->v_type = VREG;
		break;

	default:
		panic("procfs_allocvp");
	}

	/* add to procfs vnode list */
	pfs->pfs_next = *pp;
	*pp = pfs;

out:
	pfsvplock &= ~PROCFS_LOCKED;

	if (pfsvplock & PROCFS_WANT) {
		pfsvplock &= ~PROCFS_WANT;
		wakeup((caddr_t) &pfsvplock);
	}

	return (error);
}

int
procfs_freevp(struct vnode *vp)
{
	struct pfsnode **pfspp;
	struct pfsnode *pfs;

	pfs = VTOPFS(vp);
	vp->v_data = NULL;

	pfspp = PFSHASH(pfs->pfs_pid);
	while (*pfspp != pfs && *pfspp)
		pfspp = &(*pfspp)->pfs_next;
	KKASSERT(*pfspp);
	*pfspp = pfs->pfs_next;
	pfs->pfs_next = NULL;
	kfree(pfs, M_TEMP);
	return (0);
}

int
procfs_rw(struct vop_read_args *ap)
{
	struct vnode *vp = ap->a_vp;
	struct uio *uio = ap->a_uio;
	struct thread *curtd = uio->uio_td;
	struct proc *curp;
	struct pfsnode *pfs = VTOPFS(vp);
	struct proc *p;
	struct lwp *lp;
	int rtval;

	if (curtd == NULL)
		return (EINVAL);
	if ((curp = curtd->td_proc) == NULL)	/* XXX */
		return (EINVAL);

	p = PFIND(pfs->pfs_pid);
	if (p == NULL)
		return (EINVAL);
	if (p->p_pid == 1 && securelevel > 0 && uio->uio_rw == UIO_WRITE)
		return (EACCES);
	/* XXX lwp */
	lp = FIRST_LWP_IN_PROC(p);

	while (pfs->pfs_lockowner) {
		tsleep(&pfs->pfs_lockowner, 0, "pfslck", 0);
	}
	pfs->pfs_lockowner = curproc->p_pid;

	switch (pfs->pfs_type) {
	case Pnote:
	case Pnotepg:
		rtval = procfs_donote(curp, lp, pfs, uio);
		break;

	case Pregs:
		rtval = procfs_doregs(curp, lp, pfs, uio);
		break;

	case Pfpregs:
		rtval = procfs_dofpregs(curp, lp, pfs, uio);
		break;

        case Pdbregs:
                rtval = procfs_dodbregs(curp, lp, pfs, uio);
                break;

	case Pctl:
		rtval = procfs_doctl(curp, lp, pfs, uio);
		break;

	case Pstatus:
		rtval = procfs_dostatus(curp, lp, pfs, uio);
		break;

	case Pmap:
		rtval = procfs_domap(curp, lp, pfs, uio);
		break;

	case Pmem:
		rtval = procfs_domem(curp, lp, pfs, uio);
		break;

	case Ptype:
		rtval = procfs_dotype(curp, lp, pfs, uio);
		break;

	case Pcmdline:
		rtval = procfs_docmdline(curp, lp, pfs, uio);
		break;

	case Prlimit:
		rtval = procfs_dorlimit(curp, lp, pfs, uio);
		break;

	default:
		rtval = EOPNOTSUPP;
		break;
	}
	pfs->pfs_lockowner = 0;
	wakeup(&pfs->pfs_lockowner);
	return rtval;
}

/*
 * Get a string from userland into (buf).  Strip a trailing
 * nl character (to allow easy access from the shell).
 * The buffer should be *buflenp + 1 chars long.  vfs_getuserstr
 * will automatically add a nul char at the end.
 *
 * Returns 0 on success or the following errors
 *
 * EINVAL:    file offset is non-zero.
 * EMSGSIZE:  message is longer than kernel buffer
 * EFAULT:    user i/o buffer is not addressable
 */
int
vfs_getuserstr(struct uio *uio, char *buf, int *buflenp)
{
	int xlen;
	int error;

	if (uio->uio_offset != 0)
		return (EINVAL);

	xlen = *buflenp;

	/* must be able to read the whole string in one go */
	if (xlen < uio->uio_resid)
		return (EMSGSIZE);
	xlen = uio->uio_resid;

	if ((error = uiomove(buf, xlen, uio)) != 0)
		return (error);

	/* allow multiple writes without seeks */
	uio->uio_offset = 0;

	/* cleanup string and remove trailing newline */
	buf[xlen] = '\0';
	xlen = strlen(buf);
	if (xlen > 0 && buf[xlen-1] == '\n')
		buf[--xlen] = '\0';
	*buflenp = xlen;

	return (0);
}

vfs_namemap_t *
vfs_findname(vfs_namemap_t *nm, char *buf, int buflen)
{

	for (; nm->nm_name; nm++)
		if (bcmp(buf, nm->nm_name, buflen+1) == 0)
			return (nm);

	return (0);
}

void
procfs_exit(struct thread *td)
{
	struct pfsnode *pfs;
	struct vnode *vp;
	pid_t pid;

	KKASSERT(td->td_proc);
	pid = td->td_proc->p_pid;

	/*
	 * The reason for this loop is not obvious -- basicly,
	 * procfs_freevp(), which is called via vgone() (eventually),
	 * removes the specified procfs node from the pfshead list.
	 * It does this by *pfsp = pfs->pfs_next, meaning that it
	 * overwrites the node.  So when we do pfs = pfs->next, we
	 * end up skipping the node that replaces the one that was
	 * vgone'd.  Since it may have been the last one on the list,
	 * it may also have been set to null -- but *our* pfs pointer,
	 * here, doesn't see this.  So the loop starts from the beginning
	 * again.
	 *
	 * This is not a for() loop because the final event
	 * would be "pfs = pfs->pfs_next"; in the case where
	 * pfs is set to pfshead again, that would mean that
	 * pfshead is skipped over.
	 *
	 */
again:
	pfs = *PFSHASH(pid);
	while (pfs) {
		if (pfs->pfs_pid == pid) {
			vp = PFSTOV(pfs);
			vx_lock(vp);
			vgone_vxlocked(vp);
			vx_unlock(vp);
			goto again;
		}
		pfs = pfs->pfs_next;
	}
}