Pseudofsize procfs(5).

20 files changed, 483 insertions, 2044 deletions

diff --git a/sys/conf/files b/sys/conf/files
index 7c2407d..6238eb3 100644
--- a/sys/conf/files
+++ b/sys/conf/files
@@ -644,19 +644,18 @@ fs/nwfs/nwfs_vfsops.c		optional nwfs
 fs/nwfs/nwfs_vnops.c		optional nwfs
 fs/portalfs/portal_vfsops.c	optional portalfs
 fs/portalfs/portal_vnops.c	optional portalfs
+fs/procfs/procfs.c	optional procfs
 fs/procfs/procfs_ctl.c	optional procfs
 fs/procfs/procfs_dbregs.c	optional procfs
 fs/procfs/procfs_fpregs.c	optional procfs
+fs/procfs/procfs_ioctl.c	optional procfs
 fs/procfs/procfs_map.c	optional procfs
 fs/procfs/procfs_mem.c	optional procfs
 fs/procfs/procfs_note.c	optional procfs
 fs/procfs/procfs_regs.c	optional procfs
 fs/procfs/procfs_rlimit.c	optional procfs
 fs/procfs/procfs_status.c	optional procfs
-fs/procfs/procfs_subr.c	optional procfs
 fs/procfs/procfs_type.c	optional procfs
-fs/procfs/procfs_vfsops.c	optional procfs
-fs/procfs/procfs_vnops.c	optional procfs
 fs/pseudofs/pseudofs.c		optional pseudofs
 fs/pseudofs/pseudofs_fileno.c	optional pseudofs
 fs/pseudofs/pseudofs_vncache.c	optional pseudofs
diff --git a/sys/fs/procfs/procfs.c b/sys/fs/procfs/procfs.c
new file mode 100644
index 0000000..e4883d4
--- /dev/null
+++ b/sys/fs/procfs/procfs.c
@@ -0,0 +1,198 @@
+/*
+ * Copyright (c) 2001 Dag-Erling Smørgrav
+ * 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_vfsops.c	8.7 (Berkeley) 5/10/95
+ *
+ * $FreeBSD$
+ */
+
+#include <sys/param.h>
+#include <sys/queue.h>
+#include <sys/exec.h>
+#include <sys/lock.h>
+#include <sys/kernel.h>
+#include <sys/malloc.h>
+#include <sys/mount.h>
+#include <sys/mutex.h>
+#include <sys/proc.h>
+#include <sys/sbuf.h>
+#include <sys/sysproto.h>
+#include <sys/systm.h>
+#include <sys/vnode.h>
+
+#include <vm/vm.h>
+#include <vm/pmap.h>
+#include <vm/vm_param.h>
+
+#include <fs/pseudofs/pseudofs.h>
+#include <fs/procfs/procfs.h>
+
+/*
+ * Filler function for proc/pid/self
+ */
+static int
+procfs_doprocfile(PFS_FILL_ARGS)
+{
+	char *fullpath = "unknown";
+	char *freepath = NULL;
+
+	vn_fullpath(td, td->td_proc->p_textvp, &fullpath, &freepath);
+	sbuf_printf(sb, "%s", fullpath);
+	if (freepath)
+		free(freepath, M_TEMP);
+	return (0);
+}
+
+/*
+ * Filler function for proc/curproc
+ */
+static int
+procfs_docurproc(PFS_FILL_ARGS)
+{
+	sbuf_printf(sb, "%ld", (long)td->td_proc->p_pid);
+	return (0);
+}
+
+/*
+ * Adjust mode for some nodes that need it
+ */
+int
+procfs_attr(PFS_ATTR_ARGS)
+{
+	/* XXX inefficient, split into separate functions */
+	if (strcmp(pn->pn_name, "ctl") == 0 ||
+	    strcmp(pn->pn_name, "note") == 0 ||
+	    strcmp(pn->pn_name, "notepg") == 0)
+		vap->va_mode = 0200;
+	else if (strcmp(pn->pn_name, "mem") == 0 ||
+	    strcmp(pn->pn_name, "regs") == 0 ||
+	    strcmp(pn->pn_name, "dbregs") == 0 ||
+	    strcmp(pn->pn_name, "fpregs") == 0)
+		vap->va_mode = 0600;
+
+	/* p is locked by caller */
+	vap->va_uid = p->p_ucred->cr_uid;
+	vap->va_gid = p->p_ucred->cr_gid;
+	
+	return (0);
+}
+
+/*
+ * Visibility: some files only exist for non-system processes
+ * Non-static because linprocfs uses it.
+ */
+int
+procfs_notsystem(PFS_VIS_ARGS)
+{
+	return ((p->p_flag & P_SYSTEM) == 0);
+}
+
+/*
+ * Visibility: some files are only visible to process that can debug
+ * the target process.
+ */
+int
+procfs_candebug(PFS_VIS_ARGS)
+{
+	return ((p->p_flag & P_SYSTEM) == 0 &&
+	    p_candebug(td->td_proc, p) == 0);
+}
+
+/*
+ * Constructor
+ */
+static int
+procfs_init(PFS_INIT_ARGS)
+{
+	struct pfs_node *root;
+	struct pfs_node *dir;
+	struct pfs_node *node;
+
+	root = pi->pi_root;
+
+	pfs_create_link(root, "curproc", &procfs_docurproc,
+	    NULL, NULL, 0);
+	
+	dir = pfs_create_dir(root, "pid",
+	    &procfs_attr, NULL, PFS_PROCDEP);
+	pfs_create_file(dir, "cmdline", &procfs_doproccmdline,
+	    NULL, NULL, PFS_RD);
+	pfs_create_file(dir, "ctl", &procfs_doprocctl,
+	    &procfs_attr, NULL, PFS_WR);
+	pfs_create_file(dir, "dbregs", &procfs_doprocdbregs,
+	    &procfs_attr, &procfs_candebug, PFS_RDWR|PFS_RAW);
+	pfs_create_file(dir, "etype", &procfs_doproctype,
+	    NULL, NULL, PFS_RD);
+	pfs_create_file(dir, "fpregs", &procfs_doprocfpregs,
+	    &procfs_attr, &procfs_candebug, PFS_RDWR|PFS_RAW);
+	pfs_create_file(dir, "map", &procfs_doprocmap,
+	    NULL, &procfs_notsystem, PFS_RD);
+	node = pfs_create_file(dir, "mem", &procfs_doprocmem,
+	    &procfs_attr, &procfs_candebug, PFS_RDWR|PFS_RAW);
+#ifdef PROCFS_DEBUG
+	node->pn_ioctl = &procfs_ioctl;
+	node->pn_close = &procfs_close;
+#endif
+	pfs_create_file(dir, "note", &procfs_doprocnote,
+	    &procfs_attr, &procfs_candebug, PFS_WR);
+	pfs_create_file(dir, "notepg", &procfs_doprocnote,
+	    &procfs_attr, &procfs_candebug, PFS_WR);
+	pfs_create_file(dir, "regs", &procfs_doprocregs,
+	    &procfs_attr, &procfs_candebug, PFS_RDWR|PFS_RAW);
+	pfs_create_file(dir, "rlimit", &procfs_doprocrlimit,
+	    NULL, NULL, PFS_RD);
+	pfs_create_file(dir, "status", &procfs_doprocstatus,
+	    NULL, NULL, PFS_RD);
+	
+	pfs_create_link(dir, "file", &procfs_doprocfile,
+	    NULL, procfs_notsystem, 0);
+		    
+	return (0);
+}
+
+/*
+ * Destructor
+ */
+static int
+procfs_uninit(PFS_INIT_ARGS)
+{
+
+	/* nothing to do, pseudofs will GC */
+	return (0);
+}
+
+PSEUDOFS(procfs, 1);
diff --git a/sys/fs/procfs/procfs.h b/sys/fs/procfs/procfs.h
index 67af4be..059cab8 100644
--- a/sys/fs/procfs/procfs.h
+++ b/sys/fs/procfs/procfs.h
@@ -40,109 +40,30 @@
  * $FreeBSD$
  */
 
-/*
- * The different types of node in a procfs filesystem
- */
-typedef enum {
-	Proot,		/* the filesystem root */
-	Pcurproc,	/* symbolic link for curproc */
-	Pproc,		/* a process-specific sub-directory */
-	Pfile,		/* the executable file */
-	Pmem,		/* the process's memory image */
-	Pregs,		/* the process's register set */
-	Pfpregs,	/* the process's FP register set */
-	Pdbregs,	/* the process's debug register set */
-	Pctl,		/* process control */
-	Pstatus,	/* process status */
-	Pnote,		/* process notifier */
-	Pnotepg,	/* process group notifier */
-	Pmap,		/* memory map */
-	Ptype,		/* executable type */
-	Pcmdline,	/* command line */
-	Prlimit		/* resource limits */
-} pfstype;
-
-/*
- * control data for the proc file system.
- */
-struct pfsnode {
-	struct pfsnode	*pfs_next;	/* next on list */
-	struct vnode	*pfs_vnode;	/* vnode associated with this pfsnode */
-	pfstype		pfs_type;	/* type of procfs node */
-	pid_t		pfs_pid;	/* associated process */
-	u_short		pfs_mode;	/* mode bits for stat() */
-	u_long		pfs_flags;	/* open flags */
-	u_long		pfs_fileno;	/* unique file id */
-	pid_t		pfs_lockowner;	/* pfs lock owner */
-};
-
-#define PROCFS_NOTELEN	64	/* max length of a note (/proc/$pid/note) */
-#define PROCFS_CTLLEN 	8	/* max length of a ctl msg (/proc/$pid/ctl */
-#define PROCFS_NAMELEN 	8	/* max length of a filename component */
-
-/*
- * Kernel stuff follows
- */
 #ifdef _KERNEL
-#define CNEQ(cnp, s, len) \
-	 ((cnp)->cn_namelen == (len) && \
-	  (bcmp((s), (cnp)->cn_nameptr, (len)) == 0))
-
-#define PROCFS_FILENO(pid, type) \
-	(((type) < Pproc) ? \
-			((type) + 2) : \
-			((((pid)+1) << 4) + ((int) (type))))
-
-/*
- * Convert between pfsnode vnode
- */
-#define VTOPFS(vp)	((struct pfsnode *)(vp)->v_data)
-#define PFSTOV(pfs)	((pfs)->pfs_vnode)
-
-typedef struct vfs_namemap vfs_namemap_t;
-struct vfs_namemap {
-	const char *nm_name;
-	int nm_val;
-};
-
-int vfs_getuserstr __P((struct uio *, char *, int *));
-vfs_namemap_t *vfs_findname __P((vfs_namemap_t *, char *, int));
-
-/* <machine/reg.h> */
-struct reg;
-struct fpreg;
-struct dbreg;
-
-#define PFIND(pid) (pfind(pid))
 
-void procfs_exit __P((struct proc *));
-int procfs_freevp __P((struct vnode *));
-int procfs_allocvp __P((struct mount *, struct vnode **, long, pfstype));
-int procfs_donote __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_doregs __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_dofpregs __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_dodbregs __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_domem __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_doctl __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_dostatus __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_domap __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_dotype __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_docmdline __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
-int procfs_dorlimit __P((struct proc *, struct proc *, struct pfsnode *pfsp, struct uio *uio));
+int	 procfs_doproccmdline(PFS_FILL_ARGS);
+int	 procfs_doprocctl(PFS_FILL_ARGS);
+int	 procfs_doprocdbregs(PFS_FILL_ARGS);
+int	 procfs_doprocfpregs(PFS_FILL_ARGS);
+int	 procfs_doprocmap(PFS_FILL_ARGS);
+int	 procfs_doprocmem(PFS_FILL_ARGS);
+int	 procfs_doprocnote(PFS_FILL_ARGS);
+int	 procfs_doprocregs(PFS_FILL_ARGS);
+int	 procfs_doprocrlimit(PFS_FILL_ARGS);
+int	 procfs_doprocstatus(PFS_FILL_ARGS);
+int	 procfs_doproctype(PFS_FILL_ARGS);
+int	 procfs_ioctl(PFS_IOCTL_ARGS);
+int	 procfs_close(PFS_CLOSE_ARGS);
 
-/* functions to check whether or not files should be displayed */
-int procfs_validfile __P((struct thread *));
-int procfs_validfpregs __P((struct thread *));
-int procfs_validregs __P((struct thread *));
-int procfs_validdbregs __P((struct thread *));
-int procfs_validmap __P((struct thread *));
-int procfs_validtype __P((struct thread *));
+/* Return 1 if process has special kernel digging privileges */
+int	 procfs_kmemaccess(struct proc *);
 
-#define PROCFS_LOCKED	0x01
-#define PROCFS_WANT	0x02
+/* Attributes */
+int	 procfs_attr(PFS_ATTR_ARGS);
 
-extern vop_t **procfs_vnodeop_p;
+/* Visbility */
+int	 procfs_notsystem(PFS_VIS_ARGS);
+int	 procfs_candebug(PFS_VIS_ARGS);
 
-int	procfs_root __P((struct mount *, struct vnode **));
-int	procfs_rw __P((struct vop_read_args *));
 #endif /* _KERNEL */
diff --git a/sys/fs/procfs/procfs_ctl.c b/sys/fs/procfs/procfs_ctl.c
index 2dcae24..d4c37fe 100644
--- a/sys/fs/procfs/procfs_ctl.c
+++ b/sys/fs/procfs/procfs_ctl.c
@@ -44,10 +44,12 @@
 #include <sys/mutex.h>
 #include <sys/proc.h>
 #include <sys/ptrace.h>
+#include <sys/sbuf.h>
 #include <sys/signalvar.h>
 #include <sys/sx.h>
-#include <sys/vnode.h>
+#include <sys/uio.h>
 
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 #include <vm/vm.h>
@@ -67,7 +69,12 @@
 #define PROCFS_CTL_RUN		4
 #define PROCFS_CTL_WAIT		5
 
-static vfs_namemap_t ctlnames[] = {
+struct namemap {
+	const char *nm_name;
+	int nm_val;
+};
+
+static struct namemap ctlnames[] = {
 	/* special /proc commands */
 	{ "attach",	PROCFS_CTL_ATTACH },
 	{ "detach",	PROCFS_CTL_DETACH },
@@ -77,7 +84,7 @@ static vfs_namemap_t ctlnames[] = {
 	{ 0 },
 };
 
-static vfs_namemap_t signames[] = {
+static struct namemap signames[] = {
 	/* regular signal names */
 	{ "hup",	SIGHUP },	{ "int",	SIGINT },
 	{ "quit",	SIGQUIT },	{ "ill",	SIGILL },
@@ -101,10 +108,7 @@ static vfs_namemap_t signames[] = {
 static int	procfs_control __P((struct proc *curp, struct proc *p, int op));
 
 static int
-procfs_control(curp, p, op)
-	struct proc *curp;
-	struct proc *p;
-	int op;
+procfs_control(struct proc *curp, struct proc *p, int op)
 {
 	int error = 0;
 
@@ -240,7 +244,6 @@ out:
 	 * Step.  Let the target process execute a single instruction.
 	 */
 	case PROCFS_CTL_STEP:
-		_PHOLD(p);
 		PROC_UNLOCK(p);
 		error = proc_sstep(&p->p_thread); /* XXXKSE */
 		PRELE(p);
@@ -301,26 +304,26 @@ out:
 	return (0);
 }
 
+static struct namemap *
+findname(struct namemap *nm, char *buf, int buflen)
+{
+
+	for (; nm->nm_name; nm++)
+		if (bcmp(buf, nm->nm_name, buflen+1) == 0)
+			return (nm);
+
+	return (0);
+}
+
 int
-procfs_doctl(curp, p, pfs, uio)
-	struct proc *curp;
-	struct pfsnode *pfs;
-	struct uio *uio;
-	struct proc *p;
+procfs_doprocctl(PFS_FILL_ARGS)
 {
-	int xlen;
 	int error;
-	char msg[PROCFS_CTLLEN+1];
-	vfs_namemap_t *nm;
+	struct namemap *nm;
 
-	if (uio->uio_rw != UIO_WRITE)
+	if (uio == NULL || uio->uio_rw != UIO_WRITE)
 		return (EOPNOTSUPP);
 
-	xlen = PROCFS_CTLLEN;
-	error = vfs_getuserstr(uio, msg, &xlen);
-	if (error)
-		return (error);
-
 	/*
 	 * Map signal names into signal generation
 	 * or debug control.  Unknown commands and/or signals
@@ -332,15 +335,19 @@ procfs_doctl(curp, p, pfs, uio)
 	 */
 	error = EOPNOTSUPP;
 
-	nm = vfs_findname(ctlnames, msg, xlen);
+	sbuf_trim(sb);
+	sbuf_finish(sb);
+	nm = findname(ctlnames, sbuf_data(sb), sbuf_len(sb));
 	if (nm) {
-		error = procfs_control(curp, p, nm->nm_val);
+		printf("procfs: got a %s command\n", sbuf_data(sb));
+		error = procfs_control(td->td_proc, p, nm->nm_val);
 	} else {
-		nm = vfs_findname(signames, msg, xlen);
+		nm = findname(signames, sbuf_data(sb), sbuf_len(sb));
 		if (nm) {
+			printf("procfs: got a sig%s\n", sbuf_data(sb));
 			PROC_LOCK(p);
 			mtx_lock_spin(&sched_lock);
-			if (TRACE_WAIT_P(curp, p)) {
+			if (TRACE_WAIT_P(td->td_proc, p)) {
 				p->p_xstat = nm->nm_val;
 #ifdef FIX_SSTEP
 				FIX_SSTEP(&p->p_thread);   /* XXXKSE */
diff --git a/sys/fs/procfs/procfs_dbregs.c b/sys/fs/procfs/procfs_dbregs.c
index dad4025..8977955 100644
--- a/sys/fs/procfs/procfs_dbregs.c
+++ b/sys/fs/procfs/procfs_dbregs.c
@@ -49,27 +49,22 @@
 #include <sys/mutex.h>
 #include <sys/proc.h>
 #include <sys/ptrace.h>
-#include <sys/vnode.h>
+#include <sys/uio.h>
 
 #include <machine/reg.h>
 
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
-#include <vm/vm.h>
-
 int
-procfs_dodbregs(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocdbregs(PFS_FILL_ARGS)
 {
 	int error;
 	struct dbreg r;
 	char *kv;
 	int kl;
 
-	if (p_candebug(curp, p))
+	if (p_candebug(td->td_proc, p) != 0)
 		return (EPERM);
 	kl = sizeof(r);
 	kv = (char *) &r;
@@ -79,8 +74,6 @@ procfs_dodbregs(curp, p, pfs, uio)
 	if (kl > uio->uio_resid)
 		kl = uio->uio_resid;
 
-	PHOLD(p);
-
 	if (kl < 0)
 		error = EINVAL;
 	else
@@ -93,15 +86,7 @@ procfs_dodbregs(curp, p, pfs, uio)
 		else
 			error = proc_write_dbregs(&p->p_thread, &r); /* XXXKSE */
 	}
-	PRELE(p);
 
 	uio->uio_offset = 0;
 	return (error);
 }
-
-int
-procfs_validdbregs(struct thread *td)
-{
-
-	return ((td->td_proc->p_flag & P_SYSTEM) == 0);
-}
diff --git a/sys/fs/procfs/procfs_fpregs.c b/sys/fs/procfs/procfs_fpregs.c
index 99ed221..23be296 100644
--- a/sys/fs/procfs/procfs_fpregs.c
+++ b/sys/fs/procfs/procfs_fpregs.c
@@ -46,27 +46,22 @@
 #include <sys/mutex.h>
 #include <sys/proc.h>
 #include <sys/ptrace.h>
-#include <sys/vnode.h>
+#include <sys/uio.h>
 
 #include <machine/reg.h>
 
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
-#include <vm/vm.h>
-
 int
-procfs_dofpregs(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocfpregs(PFS_FILL_ARGS)
 {
 	int error;
 	struct fpreg r;
 	char *kv;
 	int kl;
 
-	if (p_candebug(curp, p))
+	if (p_candebug(td->td_proc, p))
 		return EPERM;
 	kl = sizeof(r);
 	kv = (char *) &r;
@@ -95,10 +90,3 @@ procfs_dofpregs(curp, p, pfs, uio)
 	uio->uio_offset = 0;
 	return (error);
 }
-
-int
-procfs_validfpregs(struct thread *td)
-{
-
-	return (( td->td_proc->p_flag & P_SYSTEM) == 0);
-}
diff --git a/sys/fs/procfs/procfs_ioctl.c b/sys/fs/procfs/procfs_ioctl.c
new file mode 100644
index 0000000..b5cec14
--- /dev/null
+++ b/sys/fs/procfs/procfs_ioctl.c
@@ -0,0 +1,132 @@
+/*-
+ * Copyright (c) 2001 Dag-Erling Coïdan Smørgrav
+ * All rights reserved.
+ *
+ * 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
+ *    in this position and unchanged.
+ * 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. The name of the author may not be used to endorse or promote products
+ *    derived from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
+ *
+ *      $FreeBSD$
+ */
+
+#include <sys/param.h>
+#include <sys/lock.h>
+#include <sys/mutex.h>
+#include <sys/pioctl.h>
+#include <sys/proc.h>
+#include <sys/signalvar.h>
+#include <sys/systm.h>
+
+#include <fs/pseudofs/pseudofs.h>
+#include <fs/procfs/procfs.h>
+
+#ifdef PROCFS_DEBUG
+/*
+ * Process ioctls
+ */
+int
+procfs_ioctl(PFS_IOCTL_ARGS)
+{
+	struct procfs_status *ps;
+	int error, sig;
+
+	PROC_LOCK(p);
+	error = 0;
+	switch (cmd) {
+	case PIOCBIS:
+		p->p_stops |= *(unsigned int *)data;
+		break;
+	case PIOCBIC:
+		p->p_stops &= ~*(unsigned int *)data;
+		break;
+	case PIOCSFL:
+		/* ignore */
+		break;
+	case PIOCGFL:
+		*(unsigned int *)data = 0; /* nope */
+		break;
+	case PIOCWAIT:
+		while (p->p_step == 0) {
+			/* sleep until p stops */
+			error = msleep(&p->p_stype, &p->p_mtx,
+			    PWAIT|PCATCH, "pioctl", 0);
+			if (error != 0)
+				break;
+		}
+		/* fall through to PIOCSTATUS */
+	case PIOCSTATUS:
+		ps = (struct procfs_status *)data;
+		ps->state = (p->p_step == 0);
+		ps->flags = 0; /* nope */
+		ps->events = p->p_stops;
+		ps->why = p->p_step ? p->p_stype : 0;
+		ps->val = p->p_step ? p->p_xstat : 0;
+		break;
+	case PIOCCONT:
+		if (p->p_step)
+			break;
+		sig = *(int *)data;
+		if (!_SIG_VALID(sig)) {
+			error = EINVAL;
+			break;
+		}
+#if 0
+		mtx_lock_spin(&sched_lock);
+		p->p_step = 0;
+		if (p->p_stat == SSTOP) {
+			p->p_xstat = sig;
+			setrunnable(&p->p_thread);
+			mtx_unlock_spin(&sched_lock);
+		} else {
+			mtx_unlock_spin(&sched_lock);
+			if (sig)
+				psignal(p, sig);
+		}
+#else
+		if (sig)
+			psignal(p, sig);
+		wakeup(&p->p_step);
+#endif
+		break;
+	default:
+		error = (ENOTTY);
+	}
+	PROC_UNLOCK(p);
+	
+	return (error);
+}
+
+/*
+ * Clean up on last close
+ */
+int
+procfs_close(PFS_CLOSE_ARGS)
+{
+	if (p != NULL && (p->p_pfsflags & PF_LINGER) == 0) {
+		p->p_pfsflags = 0;
+		p->p_stops = 0;
+		p->p_step = 0;
+		wakeup(&p->p_step);
+	}
+	return (0);
+}
+#endif
diff --git a/sys/fs/procfs/procfs_map.c b/sys/fs/procfs/procfs_map.c
index 945639a..cae1fc4 100644
--- a/sys/fs/procfs/procfs_map.c
+++ b/sys/fs/procfs/procfs_map.c
@@ -44,8 +44,9 @@
 #include <sys/lock.h>
 #include <sys/mutex.h>
 #include <sys/proc.h>
-#include <sys/vnode.h>
+#include <sys/uio.h>
 
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 #include <vm/vm.h>
@@ -68,11 +69,7 @@
  * can try a bigger buffer.
  */
 int
-procfs_domap(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocmap(PFS_FILL_ARGS)
 {
 	int len;
 	int error;
@@ -90,7 +87,7 @@ procfs_domap(curp, p, pfs, uio)
 		return (0);
 	
 	error = 0;
-	if (map != &curproc->p_vmspace->vm_map)
+	if (map != &curthread->td_proc->p_vmspace->vm_map)
 		vm_map_lock_read(map);
 	for (entry = map->header.next;
 		((uio->uio_resid > 0) && (entry != &map->header));
@@ -118,23 +115,22 @@ procfs_domap(curp, p, pfs, uio)
 			addr += PAGE_SIZE;
 		}
 
-		for( lobj = tobj = obj; tobj; tobj = tobj->backing_object)
+		for (lobj = tobj = obj; tobj; tobj = tobj->backing_object)
 			lobj = tobj;
 
 		if (lobj) {
 			switch(lobj->type) {
-
-default:
-case OBJT_DEFAULT:
+			default:
+			case OBJT_DEFAULT:
 				type = "default";
 				break;
-case OBJT_VNODE:
+			case OBJT_VNODE:
 				type = "vnode";
 				break;
-case OBJT_SWAP:
+			case OBJT_SWAP:
 				type = "swap";
 				break;
-case OBJT_DEVICE:
+			case OBJT_DEVICE:
 				type = "device";
 				break;
 			}
@@ -148,13 +144,12 @@ case OBJT_DEVICE:
 			ref_count = 0;
 			shadow_count = 0;
 		}
-			
 
 		/*
 		 * format:
 		 *  start, end, resident, private resident, cow, access, type.
 		 */
-		snprintf(mebuffer, sizeof(mebuffer),
+		snprintf(mebuffer, sizeof mebuffer,
 		    "0x%lx 0x%lx %d %d %p %s%s%s %d %d 0x%x %s %s %s\n",
 			(u_long)entry->start, (u_long)entry->end,
 			resident, privateresident, obj,
@@ -175,14 +170,8 @@ case OBJT_DEVICE:
 		if (error)
 			break;
 	}
-	if (map != &curproc->p_vmspace->vm_map)
+	if (map != &curthread->td_proc->p_vmspace->vm_map)
 		vm_map_unlock_read(map);
 	
-	return error;
-}
-
-int
-procfs_validmap(struct thread *td)
-{
-	return ((td->td_proc->p_flag & P_SYSTEM) == 0);
+	return (error);
 }
diff --git a/sys/fs/procfs/procfs_mem.c b/sys/fs/procfs/procfs_mem.c
index 95c91f2..5e51cab 100644
--- a/sys/fs/procfs/procfs_mem.c
+++ b/sys/fs/procfs/procfs_mem.c
@@ -41,14 +41,13 @@
  */
 
 #include <sys/param.h>
-#include <sys/systm.h>
 #include <sys/lock.h>
 #include <sys/mutex.h>
 #include <sys/proc.h>
 #include <sys/ptrace.h>
-#include <sys/user.h>
-#include <sys/vnode.h>
+#include <sys/uio.h>
 
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 /*
@@ -58,20 +57,17 @@
  * from the kernel address space.
  */
 int
-procfs_domem(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocmem(PFS_FILL_ARGS)
 {
 	int error;
 
 	if (uio->uio_resid == 0)
 		return (0);
 
-	error = p_candebug(curp, p);
+	error = p_candebug(td->td_proc, p);
 	if (error)
 		return (error);
+	error = proc_rwmem(p, uio);
 
-	return (proc_rwmem(p, uio));
+	return (error);
 }
diff --git a/sys/fs/procfs/procfs_note.c b/sys/fs/procfs/procfs_note.c
index c669654..41f32b0 100644
--- a/sys/fs/procfs/procfs_note.c
+++ b/sys/fs/procfs/procfs_note.c
@@ -40,28 +40,18 @@
  */
 
 #include <sys/param.h>
-#include <sys/vnode.h>
+#include <sys/lock.h>
+#include <sys/mutex.h>
+#include <sys/sbuf.h>
+
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 int
-procfs_donote(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocnote(PFS_FILL_ARGS)
 {
-	int xlen;
-	int error;
-	char note[PROCFS_NOTELEN+1];
-
-	if (uio->uio_rw != UIO_WRITE)
-		return (EINVAL);
-
-	xlen = PROCFS_NOTELEN;
-	error = vfs_getuserstr(uio, note, &xlen);
-	if (error)
-		return (error);
-
+	sbuf_trim(sb);
+	sbuf_finish(sb);
 	/* send to process's notify function */
 	return (EOPNOTSUPP);
 }
diff --git a/sys/fs/procfs/procfs_regs.c b/sys/fs/procfs/procfs_regs.c
index 8488f13..623eb45 100644
--- a/sys/fs/procfs/procfs_regs.c
+++ b/sys/fs/procfs/procfs_regs.c
@@ -46,28 +46,22 @@
 #include <sys/mutex.h>
 #include <sys/proc.h>
 #include <sys/ptrace.h>
-#include <sys/vnode.h>
+#include <sys/uio.h>
 
 #include <machine/reg.h>
 
-#include <vm/vm.h>
-#include <vm/vm_extern.h>
-
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 int
-procfs_doregs(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocregs(PFS_FILL_ARGS)
 {
 	int error;
 	struct reg r;
 	char *kv;
 	int kl;
 
-	if (p_candebug(curp, p))
+	if (p_candebug(td->td_proc, p))
 		return EPERM;
 	kl = sizeof(r);
 	kv = (char *) &r;
@@ -96,10 +90,3 @@ procfs_doregs(curp, p, pfs, uio)
 	uio->uio_offset = 0;
 	return (error);
 }
-
-int
-procfs_validregs(struct thread *td)
-{
-
-	return ((td->td_proc->p_flag & P_SYSTEM) == 0);
-}
diff --git a/sys/fs/procfs/procfs_rlimit.c b/sys/fs/procfs/procfs_rlimit.c
index f74874d..aecfb07 100644
--- a/sys/fs/procfs/procfs_rlimit.c
+++ b/sys/fs/procfs/procfs_rlimit.c
@@ -48,31 +48,18 @@
 #include <sys/param.h>
 #include <sys/systm.h>
 #include <sys/proc.h>
-#include <sys/vnode.h>
 #include <sys/resourcevar.h>
 #include <sys/resource.h>
-#include <sys/types.h>
+#include <sys/sbuf.h>
+
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 
 int
-procfs_dorlimit(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocrlimit(PFS_FILL_ARGS)
 {
-	char *ps;
 	int i;
-	int xlen;
-	int error;
-	char psbuf[512];		/* XXX - conservative */
-
-	if (uio->uio_rw != UIO_READ)
-		return (EOPNOTSUPP);
-
-
-	ps = psbuf;
 
 	for (i = 0; i < RLIM_NLIMITS; i++) {
 
@@ -80,7 +67,7 @@ procfs_dorlimit(curp, p, pfs, uio)
 		 * Add the rlimit ident
 		 */
 
-		ps += sprintf(ps, "%s ", rlimit_ident[i]);
+		sbuf_printf(sb, "%s ", rlimit_ident[i]);
 
 		/* 
 		 * Replace RLIM_INFINITY with -1 in the string
@@ -91,9 +78,9 @@ procfs_dorlimit(curp, p, pfs, uio)
 		 */
 
 		if (p->p_rlimit[i].rlim_cur == RLIM_INFINITY) {
-			ps += sprintf(ps, "-1 ");
+			sbuf_printf(sb, "-1 ");
 		} else {
-			ps += sprintf(ps, "%llu ",
+			sbuf_printf(sb, "%llu ",
 				(unsigned long long)p->p_rlimit[i].rlim_cur);
 		}
 
@@ -102,27 +89,13 @@ procfs_dorlimit(curp, p, pfs, uio)
 		 */
 
 		if (p->p_rlimit[i].rlim_max == RLIM_INFINITY) {
-			ps += sprintf(ps, "-1\n");
+			sbuf_printf(sb, "-1\n");
 		} else {
-			ps += sprintf(ps, "%llu\n",
+			sbuf_printf(sb, "%llu\n",
 				(unsigned long long)p->p_rlimit[i].rlim_max);
 		}
 	}
 
-	/*
-	 * This logic is rather tasty - but its from procfs_status.c, so
-	 * I guess I'll use it here.
-	 */
-
-	xlen = ps - psbuf;
-	xlen -= uio->uio_offset;
-	ps = psbuf + uio->uio_offset;
-	xlen = imin(xlen, uio->uio_resid);
-	if (xlen <= 0)
-		error = 0;
-	else
-		error = uiomove(ps, xlen, uio);
-
-	return (error);
+	return (0);
 }
 
diff --git a/sys/fs/procfs/procfs_status.c b/sys/fs/procfs/procfs_status.c
index 6dd34c7..44772d3 100644
--- a/sys/fs/procfs/procfs_status.c
+++ b/sys/fs/procfs/procfs_status.c
@@ -49,36 +49,26 @@
 #include <sys/malloc.h>
 #include <sys/proc.h>
 #include <sys/resourcevar.h>
+#include <sys/sbuf.h>
 #include <sys/tty.h>
-#include <sys/vnode.h>
 
 #include <vm/vm.h>
 #include <vm/pmap.h>
 #include <vm/vm_param.h>
 
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
-#define DOCHECK() do { if (ps >= psbuf+sizeof(psbuf)) goto bailout; } while (0)
 int
-procfs_dostatus(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doprocstatus(PFS_FILL_ARGS)
 {
 	struct session *sess;
 	struct tty *tp;
 	struct ucred *cr;
-	char *ps, *pc;
+	char *pc;
 	char *sep;
 	int pid, ppid, pgid, sid;
 	int i;
-	int xlen;
-	int error;
-	char psbuf[256];	/* XXX - conservative */
-
-	if (uio->uio_rw != UIO_READ)
-		return (EOPNOTSUPP);
 
 	pid = p->p_pid;
 	PROC_LOCK(p);
@@ -91,45 +81,31 @@ procfs_dostatus(curp, p, pfs, uio)
 /* comm pid ppid pgid sid maj,min ctty,sldr start ut st wmsg 
                                 euid ruid rgid,egid,groups[1 .. NGROUPS]
 */
-	KASSERT(sizeof(psbuf) > MAXCOMLEN,
-			("Too short buffer for new MAXCOMLEN"));
 
-	ps = psbuf;
 	pc = p->p_comm;
-	xlen = strlen(p->p_comm);
 	do {
 		if (*pc < 33 || *pc > 126 || *pc == '\\')
-			ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, "\\%03o",
-			    *pc);
+			sbuf_printf(sb, "\\%03o", *pc);
 		else
-			*ps++ = *pc;
-		DOCHECK();
-	} while (++pc < p->p_comm + xlen);
-	ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-	    " %d %d %d %d ", pid, ppid, pgid, sid);
-	DOCHECK();
+			sbuf_putc(sb, *pc);
+	} while (*++pc);
+	sbuf_printf(sb, " %d %d %d %d ", pid, ppid, pgid, sid);
 	if ((p->p_flag&P_CONTROLT) && (tp = sess->s_ttyp))
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-		    "%d,%d ", major(tp->t_dev), minor(tp->t_dev));
+		sbuf_printf(sb, "%d,%d ", major(tp->t_dev), minor(tp->t_dev));
 	else
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-		    "%d,%d ", -1, -1);
-	DOCHECK();
+		sbuf_printf(sb, "%d,%d ", -1, -1);
 
 	sep = "";
 	if (sess->s_ttyvp) {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, "%sctty", sep);
+		sbuf_printf(sb, "%sctty", sep);
 		sep = ",";
-		DOCHECK();
 	}
 	if (SESS_LEADER(p)) {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, "%ssldr", sep);
+		sbuf_printf(sb, "%ssldr", sep);
 		sep = ",";
-		DOCHECK();
 	}
 	if (*sep != ',') {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, "noflags");
-		DOCHECK();
+		sbuf_printf(sb, "noflags");
 	}
 
 	mtx_lock_spin(&sched_lock);
@@ -138,92 +114,56 @@ procfs_dostatus(curp, p, pfs, uio)
 
 		calcru(p, &ut, &st, (struct timeval *) NULL);
 		mtx_unlock_spin(&sched_lock);
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-		    " %lld,%ld %ld,%ld %ld,%ld",
+		sbuf_printf(sb, " %lld,%ld %ld,%ld %ld,%ld",
 		    (long long)p->p_stats->p_start.tv_sec,
 		    p->p_stats->p_start.tv_usec,
-		    (long)ut.tv_sec, ut.tv_usec,
-		    (long)st.tv_sec, st.tv_usec);
+		    ut.tv_sec, ut.tv_usec,
+		    st.tv_sec, st.tv_usec);
 	} else {
 		mtx_unlock_spin(&sched_lock);
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-		    " -1,-1 -1,-1 -1,-1");
+		sbuf_printf(sb, " -1,-1 -1,-1 -1,-1");
 	}
-	DOCHECK();
 
 	if (p->p_flag & P_KSES) {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, " %s",
-			"-kse- ");
+		sbuf_printf(sb, " %s", "-kse- ");
 	} else {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, " %s",
+		sbuf_printf(sb, " %s",
 			(p->p_thread.td_wchan && p->p_thread.td_wmesg) ?
 			    p->p_thread.td_wmesg : "nochan");
 	}
-	DOCHECK();
 
 	cr = p->p_ucred;
 
-	ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, " %lu %lu %lu", 
+	sbuf_printf(sb, " %lu %lu %lu",
 		(u_long)cr->cr_uid,
 		(u_long)cr->cr_ruid,
 		(u_long)cr->cr_rgid);
-	DOCHECK();
 
 	/* egid (cr->cr_svgid) is equal to cr_ngroups[0] 
 	   see also getegid(2) in /sys/kern/kern_prot.c */
 
 	for (i = 0; i < cr->cr_ngroups; i++) {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-		    ",%lu", (u_long)cr->cr_groups[i]);
-		DOCHECK();
+		sbuf_printf(sb, ",%lu", (u_long)cr->cr_groups[i]);
 	}
 
 	if (jailed(p->p_ucred)) {
 		mtx_lock(&p->p_ucred->cr_prison->pr_mtx);
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps,
-		    " %s", p->p_ucred->cr_prison->pr_host);
+		sbuf_printf(sb, " %s", p->p_ucred->cr_prison->pr_host);
 		mtx_unlock(&p->p_ucred->cr_prison->pr_mtx);
 	} else {
-		ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, " -");
+		sbuf_printf(sb, " -");
 	}
-	DOCHECK();
-	ps += snprintf(ps, psbuf + sizeof(psbuf) - ps, "\n");
-	DOCHECK();
-
-	xlen = ps - psbuf;
-	xlen -= uio->uio_offset;
-	ps = psbuf + uio->uio_offset;
-	xlen = imin(xlen, uio->uio_resid);
-	if (xlen <= 0)
-		error = 0;
-	else
-		error = uiomove(ps, xlen, uio);
-
-	return (error);
+	sbuf_printf(sb, "\n");
 
-bailout:
-	return (ENOMEM);
+	return (0);
 }
 
 int
-procfs_docmdline(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doproccmdline(PFS_FILL_ARGS)
 {
-	char *ps;
-	int xlen;
-	int error;
-	char *buf, *bp;
-	int buflen;
 	struct ps_strings pstr;
-	int i;
-	size_t bytes_left, done;
+	int error, i;
 
-	if (uio->uio_rw != UIO_READ)
-		return (EOPNOTSUPP);
-	
 	/*
 	 * If we are using the ps/cmdline caching, use that.  Otherwise
 	 * revert back to the old way which only implements full cmdline
@@ -234,47 +174,19 @@ procfs_docmdline(curp, p, pfs, uio)
 	 * Linux behaviour is to return zero-length in this case.
 	 */
 
-	if (p->p_args && (ps_argsopen || !p_cansee(curp, p))) {
-		bp = p->p_args->ar_args;
-		buflen = p->p_args->ar_length;
-		buf = 0;
-	} else if (p != curp) {
-		bp = p->p_comm;
-		buflen = MAXCOMLEN;
-		buf = 0;
+	if (p->p_args && (ps_argsopen || !p_cansee(td->td_proc, p))) {
+		sbuf_bcpy(sb, p->p_args->ar_args, p->p_args->ar_length);
+	} else if (p != td->td_proc) {
+		sbuf_printf(sb, "%.*s", MAXCOMLEN, p->p_comm);
 	} else {
-		buflen = 256;
-		MALLOC(buf, char *, buflen + 1, M_TEMP, M_WAITOK);
-		bp = buf;
-		ps = buf;
 		error = copyin((void*)PS_STRINGS, &pstr, sizeof(pstr));
-		if (error) {
-			FREE(buf, M_TEMP);
+		if (error)
 			return (error);
+		for (i = 0; i < pstr.ps_nargvstr; i++) {
+			sbuf_copyin(sb, pstr.ps_argvstr[i], 0);
+			sbuf_printf(sb, "%c", '\0');
 		}
-		bytes_left = buflen;
-		for (i = 0; bytes_left && (i < pstr.ps_nargvstr); i++) {
-			error = copyinstr(pstr.ps_argvstr[i], ps,
-					  bytes_left, &done);
-			/* If too long or malformed, just truncate */
-			if (error) {
-				error = 0;
-				break;
-			}
-			ps += done;
-			bytes_left -= done;
-		}
-		buflen = ps - buf;
 	}
 
-	buflen -= uio->uio_offset;
-	ps = bp + uio->uio_offset;
-	xlen = min(buflen, uio->uio_resid);
-	if (xlen <= 0)
-		error = 0;
-	else
-		error = uiomove(ps, xlen, uio);
-	if (buf)
-		FREE(buf, M_TEMP);
-	return (error);
+	return (0);
 }
diff --git a/sys/fs/procfs/procfs_subr.c b/sys/fs/procfs/procfs_subr.c
deleted file mode 100644
index a07cedc..0000000
--- a/sys/fs/procfs/procfs_subr.c
+++ /dev/null
@@ -1,413 +0,0 @@
-/*
- * 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
- *
- * $FreeBSD$
- */
-
-#include <sys/param.h>
-#include <sys/systm.h>
-#include <sys/lock.h>
-#include <sys/malloc.h>
-#include <sys/mount.h>
-#include <sys/mutex.h>
-#include <sys/proc.h>
-#include <sys/socket.h>
-#include <sys/sysctl.h>
-#include <sys/vnode.h>
-
-#include <fs/procfs/procfs.h>
-
-static struct pfsnode *pfshead;
-static int pfsvplock;
-
-/*
- * allocate a pfsnode/vnode pair.  the vnode is
- * referenced, but not 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.  this is done because there may be
- * zero references to it and so it needs to removed from
- * the vnode free list.
- */
-int
-procfs_allocvp(mp, vpp, pid, pfs_type)
-	struct mount *mp;
-	struct vnode **vpp;
-	long pid;
-	pfstype pfs_type;
-{
-	struct thread *td = curthread;	/* XXX */
-	struct pfsnode *pfs;
-	struct vnode *vp;
-	struct pfsnode **pp;
-	int error;
-
-loop:
-	for (pfs = pfshead; pfs != 0; pfs = pfs->pfs_next) {
-		vp = PFSTOV(pfs);
-		if (pfs->pfs_pid == pid &&
-		    pfs->pfs_type == pfs_type &&
-		    vp->v_mount == mp) {
-			if (vget(vp, 0, td))
-				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, PINOD, "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.
-	 */
-	MALLOC(pfs, struct pfsnode *, sizeof(struct pfsnode), M_TEMP, M_WAITOK);
-
-	if ((error = getnewvnode(VT_PROCFS, mp, procfs_vnodeop_p, vpp)) != 0) {
-		FREE(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 */
-	for (pp = &pfshead; *pp; pp = &(*pp)->pfs_next)
-		continue;
-	*pp = pfs;
-
-out:
-	pfsvplock &= ~PROCFS_LOCKED;
-
-	if (pfsvplock & PROCFS_WANT) {
-		pfsvplock &= ~PROCFS_WANT;
-		wakeup((caddr_t) &pfsvplock);
-	}
-
-	return (error);
-}
-
-int
-procfs_freevp(vp)
-	struct vnode *vp;
-{
-	struct pfsnode **pfspp;
-	struct pfsnode *pfs = VTOPFS(vp);
-
-	for (pfspp = &pfshead; *pfspp != 0; pfspp = &(*pfspp)->pfs_next) {
-		if (*pfspp == pfs) {
-			*pfspp = pfs->pfs_next;
-			break;
-		}
-	}
-
-	FREE(vp->v_data, M_TEMP);
-	vp->v_data = 0;
-	return (0);
-}
-
-int
-procfs_rw(ap)
-	struct vop_read_args *ap;
-{
-	struct vnode *vp = ap->a_vp;
-	struct uio *uio = ap->a_uio;
-	struct proc *curp = uio->uio_td->td_proc;
-	struct pfsnode *pfs = VTOPFS(vp);
-	struct proc *p;
-	int rtval;
-
-	p = PFIND(pfs->pfs_pid);
-	if (p == NULL)
-		return (EINVAL);
-	PROC_UNLOCK(p);
-
-	mp_fixme("pfs_lockowner needs a lock");
-	while (pfs->pfs_lockowner) {
-		tsleep(&pfs->pfs_lockowner, PRIBIO, "pfslck", 0);
-	}
-	pfs->pfs_lockowner = curproc->p_pid;
-
-	switch (pfs->pfs_type) {
-	case Pnote:
-	case Pnotepg:
-		rtval = procfs_donote(curp, p, pfs, uio);
-		break;
-
-	case Pregs:
-		rtval = procfs_doregs(curp, p, pfs, uio);
-		break;
-
-	case Pfpregs:
-		rtval = procfs_dofpregs(curp, p, pfs, uio);
-		break;
-
-        case Pdbregs:
-                rtval = procfs_dodbregs(curp, p, pfs, uio);
-                break;
-
-	case Pctl:
-		rtval = procfs_doctl(curp, p, pfs, uio);
-		break;
-
-	case Pstatus:
-		rtval = procfs_dostatus(curp, p, pfs, uio);
-		break;
-
-	case Pmap:
-		rtval = procfs_domap(curp, p, pfs, uio);
-		break;
-
-	case Pmem:
-		rtval = procfs_domem(curp, p, pfs, uio);
-		break;
-
-	case Ptype:
-		rtval = procfs_dotype(curp, p, pfs, uio);
-		break;
-
-	case Pcmdline:
-		rtval = procfs_docmdline(curp, p, pfs, uio);
-		break;
-
-	case Prlimit:
-		rtval = procfs_dorlimit(curp, p, 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(uio, buf, buflenp)
-	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(nm, buf, buflen)
-	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 proc *p)
-{
-	struct pfsnode *pfs;
-	pid_t pid = p->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.
-	 *
-	 */
-	pfs = pfshead;
-	while (pfs) {
-		if (pfs->pfs_pid == pid) {
-			vgone(PFSTOV(pfs));
-			pfs = pfshead;
-		} else
-			pfs = pfs->pfs_next;
-	}
-}
diff --git a/sys/fs/procfs/procfs_type.c b/sys/fs/procfs/procfs_type.c
index 31cbcf0..f164b67 100644
--- a/sys/fs/procfs/procfs_type.c
+++ b/sys/fs/procfs/procfs_type.c
@@ -40,45 +40,22 @@
 #include <sys/param.h>
 #include <sys/systm.h>
 #include <sys/proc.h>
+#include <sys/sbuf.h>
 #include <sys/sysent.h>
-#include <sys/vnode.h>
+
+#include <fs/pseudofs/pseudofs.h>
 #include <fs/procfs/procfs.h>
 
 int
-procfs_dotype(curp, p, pfs, uio)
-	struct proc *curp;
-	struct proc *p;
-	struct pfsnode *pfs;
-	struct uio *uio;
+procfs_doproctype(PFS_FILL_ARGS)
 {
-	int len;
-	int error;
-	/*
-	 * buffer for emulation type
-	 */
-	char mebuffer[256];
-	char *none = "Not Available";
-
-	if (uio->uio_rw != UIO_READ)
-		return (EOPNOTSUPP);
-
-	if (uio->uio_offset != 0)
-		return (0);
+	static const char *none = "Not Available";
 
 	if (p && p->p_sysent && p->p_sysent->sv_name) {
-		len = strlen(p->p_sysent->sv_name);
-		bcopy(p->p_sysent->sv_name, mebuffer, len);
+		sbuf_printf(sb, p->p_sysent->sv_name);
 	} else {
-		len = strlen(none);
-		bcopy(none, mebuffer, len);
+		sbuf_printf(sb, none);
 	}
-	mebuffer[len++] = '\n';
-	error = uiomove(mebuffer, len, uio);
-	return error;
-}
-
-int
-procfs_validtype(struct thread *td)
-{
-	return ((td->td_proc->p_flag & P_SYSTEM) == 0);
+	sbuf_putc(sb, '\n');
+	return (0);
 }
diff --git a/sys/fs/procfs/procfs_vfsops.c b/sys/fs/procfs/procfs_vfsops.c
deleted file mode 100644
index 2442022..0000000
--- a/sys/fs/procfs/procfs_vfsops.c
+++ /dev/null
@@ -1,177 +0,0 @@
-/*
- * 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_vfsops.c	8.7 (Berkeley) 5/10/95
- *
- * $FreeBSD$
- */
-
-/*
- * procfs VFS interface
- */
-
-#include <sys/param.h>
-#include <sys/systm.h>
-#include <sys/kernel.h>
-#include <sys/proc.h>
-#include <sys/mount.h>
-#include <sys/vnode.h>
-#include <fs/procfs/procfs.h>
-
-static int	procfs_mount __P((struct mount *mp, char *path, caddr_t data,
-				  struct nameidata *ndp, struct thread *td));
-static int	procfs_statfs __P((struct mount *mp, struct statfs *sbp,
-				   struct thread *td));
-static int	procfs_unmount __P((struct mount *mp, int mntflags,
-				    struct thread *td));
-
-/*
- * VFS Operations.
- *
- * mount system call
- */
-/* ARGSUSED */
-static int
-procfs_mount(mp, path, data, ndp, td)
-	struct mount *mp;
-	char *path;
-	caddr_t data;
-	struct nameidata *ndp;
-	struct thread *td;
-{
-	size_t size;
-	int error;
-
-	if (mp->mnt_flag & MNT_UPDATE)
-		return (EOPNOTSUPP);
-
-	if (mp->mnt_vfc->vfc_refcount == 1 && (error = at_exit(procfs_exit))) {
-		printf("procfs:  cannot register procfs_exit with at_exit\n");
-		return(error);
-	}
-
-	mp->mnt_flag |= MNT_LOCAL;
-	mp->mnt_data = 0;
-	vfs_getnewfsid(mp);
-
-	size = sizeof("procfs") - 1;
-	bcopy("procfs", mp->mnt_stat.f_mntfromname, size);
-	bzero(mp->mnt_stat.f_mntfromname + size, MNAMELEN - size);
-	(void)procfs_statfs(mp, &mp->mnt_stat, td);
-
-	return (0);
-}
-
-/*
- * unmount system call
- */
-static int
-procfs_unmount(mp, mntflags, td)
-	struct mount *mp;
-	int mntflags;
-	struct thread *td;
-{
-	int error;
-	int flags = 0;
-
-	if (mntflags & MNT_FORCE)
-		flags |= FORCECLOSE;
-
-	error = vflush(mp, 0, flags);
-	if (error)
-		return (error);
-
-	if (mp->mnt_vfc->vfc_refcount == 1)
-		rm_at_exit(procfs_exit);
-
-	return (0);
-}
-
-int
-procfs_root(mp, vpp)
-	struct mount *mp;
-	struct vnode **vpp;
-{
-
-	return (procfs_allocvp(mp, vpp, 0, Proot));
-}
-
-/*
- * Get file system statistics.
- */
-static int
-procfs_statfs(mp, sbp, td)
-	struct mount *mp;
-	struct statfs *sbp;
-	struct thread *td;
-{
-	sbp->f_bsize = PAGE_SIZE;
-	sbp->f_iosize = PAGE_SIZE;
-	sbp->f_blocks = 1;	/* avoid divide by zero in some df's */
-	sbp->f_bfree = 0;
-	sbp->f_bavail = 0;
-	sbp->f_files = maxproc;			/* approx */
-	sbp->f_ffree = maxproc - nprocs;	/* approx */
-
-	if (sbp != &mp->mnt_stat) {
-		sbp->f_type = mp->mnt_vfc->vfc_typenum;
-		bcopy(&mp->mnt_stat.f_fsid, &sbp->f_fsid, sizeof(sbp->f_fsid));
-		bcopy(mp->mnt_stat.f_mntonname, sbp->f_mntonname, MNAMELEN);
-		bcopy(mp->mnt_stat.f_mntfromname, sbp->f_mntfromname, MNAMELEN);
-	}
-
-	return (0);
-}
-
-static struct vfsops procfs_vfsops = {
-	procfs_mount,
-	vfs_stdstart,
-	procfs_unmount,
-	procfs_root,
-	vfs_stdquotactl,
-	procfs_statfs,
-	vfs_stdsync,
-	vfs_stdvget,
-	vfs_stdfhtovp,
-	vfs_stdcheckexp,
-	vfs_stdvptofh,
-	vfs_stdinit,
-	vfs_stduninit,
-	vfs_stdextattrctl,
-};
-
-VFS_SET(procfs_vfsops, procfs, VFCF_SYNTHETIC);
-MODULE_VERSION(procfs, 1);
diff --git a/sys/fs/procfs/procfs_vnops.c b/sys/fs/procfs/procfs_vnops.c
deleted file mode 100644
index cd168aa..0000000
--- a/sys/fs/procfs/procfs_vnops.c
+++ /dev/null
@@ -1,1023 +0,0 @@
-/*
- * Copyright (c) 1993, 1995 Jan-Simon Pendry
- * Copyright (c) 1993, 1995
- *	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	8.18 (Berkeley) 5/21/95
- *
- * $FreeBSD$
- */
-
-/*
- * procfs vnode interface
- */
-
-#include <sys/param.h>
-#include <sys/systm.h>
-#include <sys/dirent.h>
-#include <sys/fcntl.h>
-#include <sys/kernel.h>
-#include <sys/lock.h>
-#include <sys/malloc.h>
-#include <sys/mount.h>
-#include <sys/mutex.h>
-#include <sys/namei.h>
-#include <sys/pioctl.h>
-#include <sys/proc.h>
-#include <sys/signalvar.h>
-#include <sys/socket.h>
-#include <sys/sx.h>
-#include <sys/time.h>
-#include <sys/uio.h>
-#include <sys/vnode.h>
-
-#include <machine/reg.h>
-
-#include <vm/vm_zone.h>
-
-#include <fs/procfs/procfs.h>
-
-static int	procfs_access __P((struct vop_access_args *));
-static int	procfs_badop __P((void));
-static int	procfs_close __P((struct vop_close_args *));
-static int	procfs_getattr __P((struct vop_getattr_args *));
-static int	procfs_ioctl __P((struct vop_ioctl_args *));
-static int	procfs_lookup __P((struct vop_lookup_args *));
-static int	procfs_open __P((struct vop_open_args *));
-static int	procfs_print __P((struct vop_print_args *));
-static int	procfs_readdir __P((struct vop_readdir_args *));
-static int	procfs_readlink __P((struct vop_readlink_args *));
-static int	procfs_reclaim __P((struct vop_reclaim_args *));
-static int	procfs_setattr __P((struct vop_setattr_args *));
-
-/*
- * 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 proc_target {
-	u_char	pt_type;
-	u_char	pt_namlen;
-	char	*pt_name;
-	pfstype	pt_pfstype;
-	int	(*pt_valid) __P((struct thread *p));
-} proc_targets[] = {
-#define N(s) sizeof(s)-1, s
-	/*	  name		type		validp */
-	{ DT_DIR, N("."),	Pproc,		NULL },
-	{ DT_DIR, N(".."),	Proot,		NULL },
-	{ DT_REG, N("mem"),	Pmem,		NULL },
-	{ DT_REG, N("regs"),	Pregs,		procfs_validregs },
-	{ DT_REG, N("fpregs"),	Pfpregs,	procfs_validfpregs },
-	{ DT_REG, N("dbregs"),	Pdbregs,	procfs_validdbregs },
-	{ DT_REG, N("ctl"),	Pctl,		NULL },
-	{ DT_REG, N("status"),	Pstatus,	NULL },
-	{ DT_REG, N("note"),	Pnote,		NULL },
-	{ DT_REG, N("notepg"),	Pnotepg,	NULL },
-	{ DT_REG, N("map"), 	Pmap,		procfs_validmap },
-	{ DT_REG, N("etype"),	Ptype,		procfs_validtype },
-	{ DT_REG, N("cmdline"),	Pcmdline,	NULL },
-	{ DT_REG, N("rlimit"),	Prlimit,	NULL },
-	{ DT_LNK, N("file"),	Pfile,		NULL },
-#undef N
-};
-static const int nproc_targets = sizeof(proc_targets) / sizeof(proc_targets[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.
- */
-static int
-procfs_open(ap)
-	struct vop_open_args /* {
-		struct vnode *a_vp;
-		int  a_mode;
-		struct ucred *a_cred;
-		struct thread *a_td;
-	} */ *ap;
-{
-	struct pfsnode *pfs = VTOPFS(ap->a_vp);
-	struct proc *p1, *p2;
-	int error = 0;
-
-	p2 = PFIND(pfs->pfs_pid);
-	if (p2 == NULL)
-		return (ENOENT);
-	if (pfs->pfs_pid && p_cansee(ap->a_td->td_proc, p2)) {
-		error = ENOENT;
-		goto out;
-	}
-
-	switch (pfs->pfs_type) {
-	case Pmem:
-		if (((pfs->pfs_flags & FWRITE) && (ap->a_mode & O_EXCL)) ||
-		    ((pfs->pfs_flags & O_EXCL) && (ap->a_mode & FWRITE))) {
-			error = EBUSY;
-			goto out;
-		}
-
-		p1 = ap->a_td->td_proc;
-		error = p_candebug(p1, p2);
-		if (error)
-			goto out;
-
-		if (ap->a_mode & FWRITE)
-			pfs->pfs_flags = ap->a_mode & (FWRITE|O_EXCL);
-
-	default:
-		break;
-	}
-out:
-	PROC_UNLOCK(p2);
-	return (error);
-}
-
-/*
- * 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).
- */
-static int
-procfs_close(ap)
-	struct vop_close_args /* {
-		struct vnode *a_vp;
-		int  a_fflag;
-		struct ucred *a_cred;
-		struct thread *a_td;
-	} */ *ap;
-{
-	struct pfsnode *pfs = VTOPFS(ap->a_vp);
-	struct proc *p;
-
-	switch (pfs->pfs_type) {
-	case Pmem:
-		if ((ap->a_fflag & FWRITE) && (pfs->pfs_flags & O_EXCL))
-			pfs->pfs_flags &= ~(FWRITE|O_EXCL);
-		/*
-		 * If this is the last close, then it checks to see if
-		 * the target process has PF_LINGER set in p_pfsflags,
-		 * if this is *not* the case, then the process' stop flags
-		 * are cleared, and the process is woken up.  This is
-		 * to help prevent the case where a process has been
-		 * told to stop on an event, but then the requesting process
-		 * has gone away or forgotten about it.
-		 */
-		if ((ap->a_vp->v_usecount < 2) && (p = pfind(pfs->pfs_pid))) {
-			if (!(p->p_pfsflags & PF_LINGER)) {
-				p->p_stops = 0;
-				p->p_step = 0;
-				wakeup(&p->p_step);
-			}
-			PROC_UNLOCK(p);
-		}
-		break;
-	default:
-		break;
-	}
-
-	return (0);
-}
-
-/*
- * do an ioctl operation on a pfsnode (vp).
- * (vp) is not locked on entry or exit.
- */
-static int
-procfs_ioctl(ap)
-	struct vop_ioctl_args *ap;
-{
-	struct pfsnode *pfs = VTOPFS(ap->a_vp);
-	struct proc *procp, *p;
-	int error;
-	int signo;
-	struct procfs_status *psp;
-	unsigned char flags;
-
-	p = ap->a_td->td_proc;
-	procp = pfind(pfs->pfs_pid);
-	if (procp == NULL) {
-		return ENOTTY;
-	}
-
-	if ((error = p_candebug(p, procp))) {
-		PROC_UNLOCK(procp);
-		return (error == ESRCH ? ENOENT : error);
-	}
-
-	switch (ap->a_command) {
-	case PIOCBIS:
-	  procp->p_stops |= *(unsigned int*)ap->a_data;
-	  break;
-	case PIOCBIC:
-	  procp->p_stops &= ~*(unsigned int*)ap->a_data;
-	  break;
-	case PIOCSFL:
-	  /*
-	   * NFLAGS is "non-suser_xxx flags" -- currently, only
-	   * PFS_ISUGID ("ignore set u/g id");
-	   */
-#define NFLAGS	(PF_ISUGID)
-	  flags = (unsigned char)*(unsigned int*)ap->a_data;
-	  if (flags & NFLAGS && (error = suser(p))) {
-	    PROC_UNLOCK(procp);
-	    return error;
-	  }
-	  procp->p_pfsflags = flags;
-	  break;
-	case PIOCGFL:
-	  *(unsigned int*)ap->a_data = (unsigned int)procp->p_pfsflags;
-	  break;
-	case PIOCSTATUS:
-	  psp = (struct procfs_status *)ap->a_data;
-	  psp->state = (procp->p_step == 0);
-	  psp->flags = procp->p_pfsflags;
-	  psp->events = procp->p_stops;
-	  if (procp->p_step) {
-	    psp->why = procp->p_stype;
-	    psp->val = procp->p_xstat;
-	  } else {
-	    psp->why = psp->val = 0;	/* Not defined values */
-	  }
-	  break;
-	case PIOCWAIT:
-	  psp = (struct procfs_status *)ap->a_data;
-	  if (procp->p_step == 0) {
-	    error = msleep(&procp->p_stype, &procp->p_mtx, PWAIT | PCATCH,
-	      "piocwait", 0);
-	    if (error) {
-	      PROC_UNLOCK(procp);
-	      return error;
-	    }
-	  }
-	  psp->state = 1;	/* It stopped */
-	  psp->flags = procp->p_pfsflags;
-	  psp->events = procp->p_stops;
-	  psp->why = procp->p_stype;	/* why it stopped */
-	  psp->val = procp->p_xstat;	/* any extra info */
-	  break;
-	case PIOCCONT:	/* Restart a proc */
-	  if (procp->p_step == 0) {
-	    PROC_UNLOCK(procp);
-	    return EINVAL;	/* Can only start a stopped process */
-	  }
-	  if ((signo = *(int*)ap->a_data) != 0) {
-	    if (signo >= NSIG || signo <= 0) {
-	      PROC_UNLOCK(procp);
-	      return EINVAL;
-	    }
-	    psignal(procp, signo);
-	  }
-	  procp->p_step = 0;
-	  wakeup(&procp->p_step);
-	  break;
-	default:
-	  PROC_UNLOCK(procp);
-	  return (ENOTTY);
-	}
-	PROC_UNLOCK(procp);
-	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.
- */
-static int
-procfs_reclaim(ap)
-	struct vop_reclaim_args /* {
-		struct vnode *a_vp;
-	} */ *ap;
-{
-
-	return (procfs_freevp(ap->a_vp));
-}
-
-/*
- * _print is used for debugging.
- * just print a readable description
- * of (vp).
- */
-static int
-procfs_print(ap)
-	struct vop_print_args /* {
-		struct vnode *a_vp;
-	} */ *ap;
-{
-	struct pfsnode *pfs = VTOPFS(ap->a_vp);
-
-	printf("tag VT_PROCFS, type %d, pid %ld, mode %x, flags %lx\n",
-	    pfs->pfs_type, (long)pfs->pfs_pid, pfs->pfs_mode, pfs->pfs_flags);
-	return (0);
-}
-
-/*
- * generic entry point for unsupported operations
- */
-static int
-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.
- */
-static int
-procfs_getattr(ap)
-	struct vop_getattr_args /* {
-		struct vnode *a_vp;
-		struct vattr *a_vap;
-		struct ucred *a_cred;
-		struct thread *a_td;
-	} */ *ap;
-{
-	struct pfsnode *pfs = VTOPFS(ap->a_vp);
-	struct vattr *vap = ap->a_vap;
-	struct proc *procp;
-	int error;
-
-	/*
-	 * First make sure that the process and its credentials 
-	 * still exist.
-	 */
-	switch (pfs->pfs_type) {
-	case Proot:
-	case Pcurproc:
-		procp = NULL;
-		break;
-
-	default:
-		procp = PFIND(pfs->pfs_pid);
-		if (procp == NULL)
-			return (ENOENT);
-		if (procp->p_ucred == NULL) {
-			PROC_UNLOCK(procp);
-			return (ENOENT);
-		}
-
-		if (p_cansee(ap->a_td->td_proc, procp)) {
-			PROC_UNLOCK(procp);
-			return (ENOENT);
-		}
-		PROC_UNLOCK(procp);
-	}
-
-	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;
-	vap->va_fsid = ap->a_vp->v_mount->mnt_stat.f_fsid.val[0];
-
-	/*
-	 * 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.
-	 */
-	nanotime(&vap->va_ctime);
-	vap->va_atime = vap->va_mtime = vap->va_ctime;
-
-	/*
-	 * 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 Pctl:
-	case Pregs:
-	case Pfpregs:
-	case Pdbregs:
-	case Pmem:
-		PROC_LOCK(procp);
-		if (procp->p_flag & P_SUGID)
-			vap->va_mode &= ~((VREAD|VWRITE)|
-					  ((VREAD|VWRITE)>>3)|
-					  ((VREAD|VWRITE)>>6));
-		PROC_UNLOCK(procp);
-		break;
-	default:
-		break;
-	}
-
-	/*
-	 * 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.
-	 */
-
-	vap->va_nlink = 1;
-	if (procp) {
-		PROC_LOCK(procp);
-		vap->va_uid = procp->p_ucred->cr_uid;
-		vap->va_gid = procp->p_ucred->cr_gid;
-		PROC_UNLOCK(procp);
-	}
-
-	switch (pfs->pfs_type) {
-	case Proot:
-		/*
-		 * Set nlink to 1 to tell fts(3) we don't actually know.
-		 */
-		vap->va_nlink = 1;
-		vap->va_uid = 0;
-		vap->va_gid = 0;
-		vap->va_size = vap->va_bytes = DEV_BSIZE;
-		break;
-
-	case Pcurproc: {
-		char buf[16];		/* should be enough */
-		vap->va_uid = 0;
-		vap->va_gid = 0;
-		vap->va_size = vap->va_bytes =
-		    snprintf(buf, sizeof(buf), "%ld", (long)curproc->p_pid);
-		break;
-	}
-
-	case Pproc:
-		vap->va_nlink = nproc_targets;
-		vap->va_size = vap->va_bytes = DEV_BSIZE;
-		break;
-
-	case Pfile: {
-		char *fullpath, *freepath;
-		error = textvp_fullpath(procp, &fullpath, &freepath);
-		if (error == 0) {
-			vap->va_size = strlen(fullpath);
-			free(freepath, M_TEMP);
-		} else {
-			vap->va_size = sizeof("unknown") - 1;
-			error = 0;
-		}
-		vap->va_bytes = vap->va_size;
-		break;
-	}
-
-	case Pmem:
-		/*
-		 * If we denied owner access earlier, then we have to
-		 * change the owner to root - otherwise 'ps' and friends
-		 * will break even though they are setgid kmem. *SIGH*
-		 * XXX: ps and friends are no longer setgid kmem, why
-		 * is this needed?
-		 */
-		PROC_LOCK(procp);
-		if (procp->p_flag & P_SUGID)
-			vap->va_uid = 0;
-		else
-			vap->va_uid = procp->p_ucred->cr_uid;
-		PROC_UNLOCK(procp);
-		break;
-
-	case Pregs:
-		vap->va_bytes = vap->va_size = sizeof(struct reg);
-		break;
-
-	case Pfpregs:
-		vap->va_bytes = vap->va_size = sizeof(struct fpreg);
-		break;
-
-        case Pdbregs:
-                vap->va_bytes = vap->va_size = sizeof(struct dbreg);
-                break;
-
-	case Ptype:
-	case Pmap:
-	case Pctl:
-	case Pstatus:
-	case Pnote:
-	case Pnotepg:
-	case Pcmdline:
-	case Prlimit:
-		break;
-
-	default:
-		panic("procfs_getattr");
-	}
-
-	return (error);
-}
-
-static int
-procfs_setattr(ap)
-	struct vop_setattr_args /* {
-		struct vnode *a_vp;
-		struct vattr *a_vap;
-		struct ucred *a_cred;
-		struct thread *a_td;
-	} */ *ap;
-{
-
-	if (ap->a_vap->va_flags != VNOVAL)
-		return (EOPNOTSUPP);
-
-	/*
-	 * just fake out attribute setting
-	 * it's not good to generate an error
-	 * return, otherwise things like creat()
-	 * will fail when they try to set the
-	 * file length to 0.  worse, this means
-	 * that echo $note > /proc/$pid/note will fail.
-	 */
-
-	return (0);
-}
-
-/*
- * implement access checking.
- *
- * actually, the check for super-user is slightly
- * broken since it will allow read access to write-only
- * objects.  this doesn't cause any particular trouble
- * but does mean that the i/o entry points need to check
- * that the operation really does make sense.
- */
-static int
-procfs_access(ap)
-	struct vop_access_args /* {
-		struct vnode *a_vp;
-		int a_mode;
-		struct ucred *a_cred;
-		struct thread *a_td;
-	} */ *ap;
-{
-	struct pfsnode *pfs = VTOPFS(ap->a_vp);
-	struct vnode *vp = ap->a_vp;
-	struct proc *procp;
-	struct vattr *vap;
-	struct vattr vattr;
-	int error;
-
-	switch (pfs->pfs_type) {
-	case Proot:
-       	case Pcurproc:
-		break;
-	default:
-		procp = PFIND(pfs->pfs_pid);
-		if (procp == NULL)
-			return (ENOENT);
-		if (p_cansee(ap->a_td->td_proc, procp)) {
-			PROC_UNLOCK(procp);
-			return (ENOENT);
-		}
-		PROC_UNLOCK(procp);
-        }
-
-	vap = &vattr;
-	error = VOP_GETATTR(vp, vap, ap->a_cred, ap->a_td);
-	if (error)
-		return (error);
-
-	return (vaccess(vp->v_type, vap->va_mode, vap->va_uid, vap->va_gid,
-	    ap->a_mode, ap->a_cred, NULL));
-}
-
-/*
- * lookup.  this is incredibly complicated in the
- * general case, however for most pseudo-filesystems
- * very little needs to be done.
- *
- * unless you want to get a migraine, just make sure your
- * filesystem doesn't do any locking of its own.  otherwise
- * read and inwardly digest ufs_lookup().
- */
-static int
-procfs_lookup(ap)
-	struct vop_lookup_args /* {
-		struct vnode * a_dvp;
-		struct vnode ** a_vpp;
-		struct componentname * a_cnp;
-	} */ *ap;
-{
-	struct componentname *cnp = ap->a_cnp;
-	struct vnode **vpp = ap->a_vpp;
-	struct vnode *dvp = ap->a_dvp;
-	char *pname = cnp->cn_nameptr;
-	struct proc *curp = cnp->cn_thread->td_proc;
-	struct proc_target *pt;
-	pid_t pid;
-	struct pfsnode *pfs;
-	struct proc *p;
-	int i;
-	struct thread *td;
-
-	*vpp = NULL;
-
-	if (cnp->cn_nameiop == DELETE || cnp->cn_nameiop == RENAME ||
-	    cnp->cn_nameiop == CREATE)
-		return (EROFS);
-
-	if (cnp->cn_namelen == 1 && *pname == '.') {
-		*vpp = dvp;
-		VREF(dvp);
-		/* vn_lock(dvp, LK_EXCLUSIVE | LK_RETRY, curp); */
-		return (0);
-	}
-
-	pfs = VTOPFS(dvp);
-	switch (pfs->pfs_type) {
-	case Proot:
-		if (cnp->cn_flags & ISDOTDOT)
-			return (EIO);
-
-		if (CNEQ(cnp, "curproc", 7))
-			return (procfs_allocvp(dvp->v_mount, vpp, 0, Pcurproc));
-
-		pid = atopid(pname, cnp->cn_namelen);
-		if (pid == NO_PID)
-			break;
-
-		p = PFIND(pid);
-		if (p == NULL)
-			break;
-
-		if (p_cansee(curp, p)) {
-			PROC_UNLOCK(p);
-			break;
-		}
-		PROC_UNLOCK(p);
-
-		return (procfs_allocvp(dvp->v_mount, vpp, pid, Pproc));
-
-	case Pproc:
-		if (cnp->cn_flags & ISDOTDOT)
-			return (procfs_root(dvp->v_mount, vpp));
-
-		p = PFIND(pfs->pfs_pid);
-		td = &p->p_thread;		/* XXXKSE */
-		if (p == NULL)
-			break;
-
-		for (pt = proc_targets, i = 0; i < nproc_targets; pt++, i++) {
-			if (cnp->cn_namelen == pt->pt_namlen &&
-			    bcmp(pt->pt_name, pname, cnp->cn_namelen) == 0 &&
-			    (pt->pt_valid == NULL || (*pt->pt_valid)(td)))
-				goto found;
-		}
-		PROC_UNLOCK(p);
-		break;
-	found:
-		PROC_UNLOCK(p);
-		return (procfs_allocvp(dvp->v_mount, vpp, pfs->pfs_pid,
-		    pt->pt_pfstype));
-
-	default:
-		return (ENOTDIR);
-	}
-
-	return (cnp->cn_nameiop == LOOKUP ? ENOENT : EROFS);
-}
-
-/*
- * Does this process have a text file?
- */
-int
-procfs_validfile(td)
-	struct thread *td;
-{
-
-	return (td->td_proc->p_textvp != NULLVP);
-}
-
-/*
- * readdir() returns directory entries from pfsnode (vp).
- *
- * We generate just one directory entry at a time, as it would probably
- * not pay off to buffer several entries locally to save uiomove calls.
- */
-static int
-procfs_readdir(ap)
-	struct vop_readdir_args /* {
-		struct vnode *a_vp;
-		struct uio *a_uio;
-		struct ucred *a_cred;
-		int *a_eofflag;
-		int *a_ncookies;
-		u_long **a_cookies;
-	} */ *ap;
-{
-	struct uio *uio = ap->a_uio;
-	struct dirent d;
-	struct dirent *dp = &d;
-	struct pfsnode *pfs;
-	int count, error, i, off;
-	static u_int delen;
-	struct thread *td;
-
-	if (!delen) {
-
-		d.d_namlen = PROCFS_NAMELEN;
-		delen = GENERIC_DIRSIZ(&d);
-	}
-
-	pfs = VTOPFS(ap->a_vp);
-
-	off = (int)uio->uio_offset;
-	if (off != uio->uio_offset || off < 0 || 
-	    off % delen != 0 || uio->uio_resid < delen)
-		return (EINVAL);
-
-	error = 0;
-	count = 0;
-	i = off / delen;
-
-	switch (pfs->pfs_type) {
-	/*
-	 * this is for the process-specific sub-directories.
-	 * all that is needed to is copy out all the entries
-	 * from the procent[] table (top of this file).
-	 */
-	case Pproc: {
-		struct proc *p;
-		struct proc_target *pt;
-
-		p = PFIND(pfs->pfs_pid);
-		td = &p->p_thread;	/* XXXKSE */
-		if (p == NULL)
-			break;
-		if (p_cansee(curthread->td_proc, p)) {
-			PROC_UNLOCK(p);
-			break;
-		}
-
-		for (pt = &proc_targets[i];
-		     uio->uio_resid >= delen && i < nproc_targets; pt++, i++) {
-			if (pt->pt_valid && (*pt->pt_valid)(td) == 0)
-				continue;
-
-			dp->d_reclen = delen;
-			dp->d_fileno = PROCFS_FILENO(pfs->pfs_pid, pt->pt_pfstype);
-			dp->d_namlen = pt->pt_namlen;
-			bcopy(pt->pt_name, dp->d_name, pt->pt_namlen + 1);
-			dp->d_type = pt->pt_type;
-
-			PROC_UNLOCK(p);
-			if ((error = uiomove((caddr_t)dp, delen, uio)) != 0)
-				break;
-			PROC_LOCK(p);
-		}
-		PROC_UNLOCK(p);
-
-	    	break;
-	    }
-
-	/*
-	 * this is for the root of the procfs filesystem
-	 * what is needed is a special entry for "curproc"
-	 * followed by an entry for each process on allproc
-#ifdef PROCFS_ZOMBIE
-	 * and zombproc.
-#endif
-	 */
-
-	case Proot: {
-#ifdef PROCFS_ZOMBIE
-		int doingzomb = 0;
-#endif
-		int pcnt = 0;
-		struct proc *p;
-
-		sx_slock(&allproc_lock);
-		p = LIST_FIRST(&allproc);
-		for (; p && uio->uio_resid >= delen; i++, pcnt++) {
-			bzero((char *) dp, delen);
-			dp->d_reclen = delen;
-
-			switch (i) {
-			case 0:		/* `.' */
-			case 1:		/* `..' */
-				dp->d_fileno = PROCFS_FILENO(0, Proot);
-				dp->d_namlen = i + 1;
-				bcopy("..", dp->d_name, dp->d_namlen);
-				dp->d_name[i + 1] = '\0';
-				dp->d_type = DT_DIR;
-				break;
-
-			case 2:
-				dp->d_fileno = PROCFS_FILENO(0, Pcurproc);
-				dp->d_namlen = 7;
-				bcopy("curproc", dp->d_name, 8);
-				dp->d_type = DT_LNK;
-				break;
-
-			default:
-				while (pcnt < i) {
-					p = LIST_NEXT(p, p_list);
-					if (p == NULL)
-						goto done;
-					if (p_cansee(curthread->td_proc, p))
-						continue;
-					pcnt++;
-				}
-				while (p_cansee(curthread->td_proc, p)) {
-					p = LIST_NEXT(p, p_list);
-					if (p == NULL)
-						goto done;
-				}
-				dp->d_fileno = PROCFS_FILENO(p->p_pid, Pproc);
-				dp->d_namlen = sprintf(dp->d_name, "%ld",
-				    (long)p->p_pid);
-				dp->d_type = DT_DIR;
-				p = LIST_NEXT(p, p_list);
-				break;
-			}
-
-			if ((error = uiomove((caddr_t)dp, delen, uio)) != 0)
-				break;
-		}
-	done:
-
-#ifdef PROCFS_ZOMBIE
-		if (p == NULL && doingzomb == 0) {
-			doingzomb = 1;
-			p = LIST_FIRST(&zombproc);
-			goto again;
-		}
-#endif
-
-		sx_sunlock(&allproc_lock);
-		break;
-
-	    }
-
-	default:
-		error = ENOTDIR;
-		break;
-	}
-
-	uio->uio_offset = i * delen;
-
-	return (error);
-}
-
-/*
- * readlink reads the link of `curproc' or `file'
- */
-static int
-procfs_readlink(ap)
-	struct vop_readlink_args *ap;
-{
-	char buf[16];		/* should be enough */
-	struct proc *procp;
-	struct vnode *vp = ap->a_vp;
-	struct pfsnode *pfs = VTOPFS(vp);
-	char *fullpath, *freepath;
-	int error, len;
-
-	switch (pfs->pfs_type) {
-	case Pcurproc:
-		if (pfs->pfs_fileno != PROCFS_FILENO(0, Pcurproc))
-			return (EINVAL);
-
-		len = snprintf(buf, sizeof(buf), "%ld", (long)curproc->p_pid);
-
-		return (uiomove(buf, len, ap->a_uio));
-	/*
-	 * There _should_ be no way for an entire process to disappear
-	 * from under us...
-	 */
-	case Pfile:
-		procp = PFIND(pfs->pfs_pid);
-		if (procp == NULL || procp->p_ucred == NULL) {
-			if (procp != NULL)
-				PROC_UNLOCK(procp);
-			printf("procfs_readlink: pid %d disappeared\n",
-			    pfs->pfs_pid);
-			return (uiomove("unknown", sizeof("unknown") - 1,
-			    ap->a_uio));
-		}
-		PROC_UNLOCK(procp);
-		error = textvp_fullpath(procp, &fullpath, &freepath);
-		if (error != 0)
-			return (uiomove("unknown", sizeof("unknown") - 1,
-			    ap->a_uio));
-		error = uiomove(fullpath, strlen(fullpath), ap->a_uio);
-		free(freepath, M_TEMP);
-		return (error);
-	default:
-		return (EINVAL);
-	}
-}
-
-/*
- * convert decimal ascii to pid_t
- */
-static pid_t
-atopid(b, len)
-	const char *b;
-	u_int len;
-{
-	pid_t p = 0;
-
-	while (len--) {
-		char c = *b++;
-		if (c < '0' || c > '9')
-			return (NO_PID);
-		p = 10 * p + (c - '0');
-		if (p > PID_MAX)
-			return (NO_PID);
-	}
-
-	return (p);
-}
-
-/*
- * procfs vnode operations.
- */
-vop_t **procfs_vnodeop_p;
-static struct vnodeopv_entry_desc procfs_vnodeop_entries[] = {
-	{ &vop_default_desc,		(vop_t *) vop_defaultop },
-	{ &vop_access_desc,		(vop_t *) procfs_access },
-	{ &vop_advlock_desc,		(vop_t *) procfs_badop },
-	{ &vop_close_desc,		(vop_t *) procfs_close },
-	{ &vop_create_desc,		(vop_t *) procfs_badop },
-	{ &vop_getattr_desc,		(vop_t *) procfs_getattr },
-	{ &vop_link_desc,		(vop_t *) procfs_badop },
-	{ &vop_lookup_desc,		(vop_t *) procfs_lookup },
-	{ &vop_mkdir_desc,		(vop_t *) procfs_badop },
-	{ &vop_mknod_desc,		(vop_t *) procfs_badop },
-	{ &vop_open_desc,		(vop_t *) procfs_open },
-	{ &vop_pathconf_desc,		(vop_t *) vop_stdpathconf },
-	{ &vop_print_desc,		(vop_t *) procfs_print },
-	{ &vop_read_desc,		(vop_t *) procfs_rw },
-	{ &vop_readdir_desc,		(vop_t *) procfs_readdir },
-	{ &vop_readlink_desc,		(vop_t *) procfs_readlink },
-	{ &vop_reclaim_desc,		(vop_t *) procfs_reclaim },
-	{ &vop_remove_desc,		(vop_t *) procfs_badop },
-	{ &vop_rename_desc,		(vop_t *) procfs_badop },
-	{ &vop_rmdir_desc,		(vop_t *) procfs_badop },
-	{ &vop_setattr_desc,		(vop_t *) procfs_setattr },
-	{ &vop_symlink_desc,		(vop_t *) procfs_badop },
-	{ &vop_write_desc,		(vop_t *) procfs_rw },
-	{ &vop_ioctl_desc,		(vop_t *) procfs_ioctl },
-	{ NULL, NULL }
-};
-static struct vnodeopv_desc procfs_vnodeop_opv_desc =
-	{ &procfs_vnodeop_p, procfs_vnodeop_entries };
-
-VNODEOP_SET(procfs_vnodeop_opv_desc);
diff --git a/sys/modules/Makefile b/sys/modules/Makefile
index 6d92802..e6660b2 100644
--- a/sys/modules/Makefile
+++ b/sys/modules/Makefile
@@ -68,7 +68,6 @@ SUBDIR=	3dfx \
 	ppbus \
 	ppi \
 	pps \
-	procfs \
 	${_random} \
 	rl \
 	rp \
diff --git a/sys/modules/fs/Makefile b/sys/modules/fs/Makefile
index 49244f3..bedaa5a 100644
--- a/sys/modules/fs/Makefile
+++ b/sys/modules/fs/Makefile
@@ -2,6 +2,7 @@
 
 SUBDIR =
 SUBDIR += linprocfs
+SUBDIR += procfs
 SUBDIR += pseudofs
 
 .include <bsd.subdir.mk>
diff --git a/sys/modules/procfs/Makefile b/sys/modules/fs/procfs/Makefile
index 3d9338e..ffb2ae7 100644
--- a/sys/modules/procfs/Makefile
+++ b/sys/modules/fs/procfs/Makefile
@@ -1,22 +1,20 @@
 # $FreeBSD$
 
-.PATH: ${.CURDIR}/../../fs/procfs
+.PATH: ${.CURDIR}/../../../fs/procfs
 
 KMOD=	procfs
 SRCS=	vnode_if.h \
 	procfs_ctl.c \
 	procfs_dbregs.c \
 	procfs_fpregs.c \
+	procfs_ioctl.c \
 	procfs_map.c \
 	procfs_mem.c \
 	procfs_note.c \
 	procfs_regs.c \
 	procfs_rlimit.c \
 	procfs_status.c \
-	procfs_subr.c \
 	procfs_type.c \
-	procfs_vfsops.c \
-	procfs_vnops.c
-NOMAN=
+	procfs.c
 
 .include <bsd.kmod.mk>