vm_mmap.c

open bsd vm device design

/*
 * Copyright (c) 1988 University of Utah.
 * Copyright (c) 1991, 1993
 *	The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * the Systems Programming Group of the University of Utah Computer
 * Science Department.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 * from: Utah $Hdr: vm_mmap.c 1.6 91/10/21$
 *
 *	@(#)vm_mmap.c	8.10 (Berkeley) 2/19/95
 */

/*
 * Mapped file (mmap) interface to VM
 */

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/filedesc.h>
#include <sys/resourcevar.h>
#include <sys/proc.h>
#include <sys/vnode.h>
#include <sys/file.h>
#include <sys/mman.h>
#include <sys/conf.h>

#include <sys/mount.h>
#include <sys/syscallargs.h>

#include <miscfs/specfs/specdev.h>

#include <vm/vm.h>
#include <vm/vm_pager.h>
#include <vm/vm_prot.h>

#ifdef DEBUG
int mmapdebug = 0;
#define MDB_FOLLOW	0x01
#define MDB_SYNC	0x02
#define MDB_MAPIT	0x04
#endif

/* ARGSUSED */
int
sbrk(p, uap, retval)
	struct proc *p;
	struct sbrk_args /* {
		syscallarg(int) incr;
	} */ *uap;
	register_t *retval;
{

	/* Not yet implemented */
	return (EOPNOTSUPP);
}

/* ARGSUSED */
int
sstk(p, uap, retval)
	struct proc *p;
	struct sstk_args /* {
		syscallarg(int) incr;
	} */ *uap;
	register_t *retval;
{

	/* Not yet implemented */
	return (EOPNOTSUPP);
}

#if defined(COMPAT_43) || defined(COMPAT_SUNOS)
/* ARGSUSED */
int
compat_43_getpagesize(p, uap, retval)
	struct proc *p;
	void *uap;
	register_t *retval;
{

	*retval = PAGE_SIZE;
	return (0);
}
#endif /* COMPAT_43 || COMPAT_SUNOS */

#ifdef COMPAT_43
int
compat_43_mmap(p, uap, retval)
	struct proc *p;
	register struct compat_43_mmap_args /* {
		syscallarg(caddr_t) addr;
		syscallarg(int) len;
		syscallarg(int) prot;
		syscallarg(int) flags;
		syscallarg(int) fd;
		syscallarg(long) pos;
	} */ *uap;
	register_t *retval;
{
	struct mmap_args /* {
		syscallarg(caddr_t) addr;
		syscallarg(size_t) len;
		syscallarg(int) prot;
		syscallarg(int) flags;
		syscallarg(int) fd;
		syscallarg(long) pad;
		syscallarg(off_t) pos;
	} */ nargs;
	static const char cvtbsdprot[8] = {
		0,
		PROT_EXEC,
		PROT_WRITE,
		PROT_EXEC|PROT_WRITE,
		PROT_READ,
		PROT_EXEC|PROT_READ,
		PROT_WRITE|PROT_READ,
		PROT_EXEC|PROT_WRITE|PROT_READ,
	};
#define	OMAP_ANON	0x0002
#define	OMAP_COPY	0x0020
#define	OMAP_SHARED	0x0010
#define	OMAP_FIXED	0x0100
#define	OMAP_INHERIT	0x0800

	SCARG(&nargs, addr) = SCARG(uap, addr);
	SCARG(&nargs, len) = SCARG(uap, len);
	SCARG(&nargs, prot) = cvtbsdprot[SCARG(uap, prot)&0x7];
	SCARG(&nargs, flags) = 0;
	if (SCARG(uap, flags) & OMAP_ANON)
		SCARG(&nargs, flags) |= MAP_ANON;
	if (SCARG(uap, flags) & OMAP_COPY)
		SCARG(&nargs, flags) |= MAP_COPY;
	if (SCARG(uap, flags) & OMAP_SHARED)
		SCARG(&nargs, flags) |= MAP_SHARED;
	else
		SCARG(&nargs, flags) |= MAP_PRIVATE;
	if (SCARG(uap, flags) & OMAP_FIXED)
		SCARG(&nargs, flags) |= MAP_FIXED;
	if (SCARG(uap, flags) & OMAP_INHERIT)
		SCARG(&nargs, flags) |= MAP_INHERIT;
	SCARG(&nargs, fd) = SCARG(uap, fd);
	SCARG(&nargs, pos) = SCARG(uap, pos);
	return (mmap(p, &nargs, retval));
}
#endif

int
mmap(p, uap, retval)
	struct proc *p;
	register struct mmap_args /* {
		syscallarg(caddr_t) addr;
		syscallarg(size_t) len;
		syscallarg(int) prot;
		syscallarg(int) flags;
		syscallarg(int) fd;
		syscallarg(long) pad;
		syscallarg(off_t) pos;
	} */ *uap;
	register_t *retval;
{
	register struct filedesc *fdp = p->p_fd;
	register struct file *fp;
	struct vnode *vp;
	vm_offset_t addr, pos;
	vm_size_t size;
	vm_prot_t prot, maxprot;
	caddr_t handle;
	int flags, error;

	prot = SCARG(uap, prot) & VM_PROT_ALL;
	flags = SCARG(uap, flags);
	pos = SCARG(uap, pos);
#ifdef DEBUG
	if (mmapdebug & MDB_FOLLOW)
		printf("mmap(%d): addr %x len %x pro %x flg %x fd %d pos %x\n",
		       p->p_pid, SCARG(uap, addr), SCARG(uap, len), prot,
		       flags, SCARG(uap, fd), pos);
#endif
	/*
	 * Address (if FIXED) must be page aligned.
	 * Size is implicitly rounded to a page boundary.
	 *
	 * XXX most (all?) vendors require that the file offset be
	 * page aligned as well.  However, we already have applications
	 * (e.g. nlist) that rely on unrestricted alignment.  Since we
	 * support it, let it happen.
	 */
	addr = (vm_offset_t) SCARG(uap, addr);
	if (((flags & MAP_FIXED) && (addr & PAGE_MASK)) ||
#if 0
	    ((flags & MAP_ANON) == 0 && (pos & PAGE_MASK)) ||
#endif
	    (ssize_t)SCARG(uap, len) < 0 || ((flags & MAP_ANON) && SCARG(uap, fd) != -1))
		return (EINVAL);
	size = (vm_size_t) round_page(SCARG(uap, len));
	/*
	 * Check for illegal addresses.  Watch out for address wrap...
	 * Note that VM_*_ADDRESS are not constants due to casts (argh).
	 */
	if (flags & MAP_FIXED) {
		if (VM_MAXUSER_ADDRESS > 0 && addr + size >= VM_MAXUSER_ADDRESS)
			return (EINVAL);
		if (VM_MIN_ADDRESS > 0 && addr < VM_MIN_ADDRESS)
			return (EINVAL);
		if (addr > addr + size)
			return (EINVAL);
	}
	/*
	 * XXX for non-fixed mappings where no hint is provided or
	 * the hint would fall in the potential heap space,
	 * place it after the end of the largest possible heap.
	 *
	 * There should really be a pmap call to determine a reasonable
	 * location.
	 */
	else if (addr < round_page(p->p_vmspace->vm_daddr + MAXDSIZ))
		addr = round_page(p->p_vmspace->vm_daddr + MAXDSIZ);
	if (flags & MAP_ANON) {
		/*
		 * Mapping blank space is trivial.
		 */
		handle = NULL;
		maxprot = VM_PROT_ALL;
		pos = 0;
	} else {
		/*
		 * Mapping file, get fp for validation.
		 * Obtain vnode and make sure it is of appropriate type.
		 */
		if (((unsigned)SCARG(uap, fd)) >= fdp->fd_nfiles ||
		    (fp = fdp->fd_ofiles[SCARG(uap, fd)]) == NULL)
			return (EBADF);
		if (fp->f_type != DTYPE_VNODE)
			return (EINVAL);
		vp = (struct vnode *)fp->f_data;
		if (vp->v_type != VREG && vp->v_type != VCHR)
			return (EINVAL);
		/*
		 * XXX hack to handle use of /dev/zero to map anon
		 * memory (ala SunOS).
		 */
		if (vp->v_type == VCHR && iszerodev(vp->v_rdev)) {
			handle = NULL;
			maxprot = VM_PROT_ALL;
			flags |= MAP_ANON;
		} else {
			/*
			 * Ensure that file and memory protections are
			 * compatible.  Note that we only worry about
			 * writability if mapping is shared; in this case,
			 * current and max prot are dictated by the open file.
			 * XXX use the vnode instead?  Problem is: what
			 * credentials do we use for determination?
			 * What if proc does a setuid?
			 */
			maxprot = VM_PROT_EXECUTE;	/* ??? */
			if (fp->f_flag & FREAD)
				maxprot |= VM_PROT_READ;
			else if (prot & PROT_READ)
				return (EACCES);
			if (flags & MAP_SHARED) {
				if (fp->f_flag & FWRITE)
					maxprot |= VM_PROT_WRITE;
				else if (prot & PROT_WRITE)
					return (EACCES);
			} else
				maxprot |= VM_PROT_WRITE;
			handle = (caddr_t)vp;
		}
	}
	error = vm_mmap(&p->p_vmspace->vm_map, &addr, size, prot, maxprot,
	    flags, handle, pos);
	if (error == 0)
		*retval = (register_t)addr;
	return (error);
}

int
msync(p, uap, retval)
	struct proc *p;
	struct msync_args /* {
		syscallarg(caddr_t) addr;
		syscallarg(int) len;
	} */ *uap;
	register_t *retval;
{
	vm_offset_t addr;
	vm_size_t size;
	vm_map_t map;
	int rv;
	boolean_t syncio, invalidate;

#ifdef DEBUG
	if (mmapdebug & (MDB_FOLLOW|MDB_SYNC))
		printf("msync(%d): addr %x len %x\n",
		       p->p_pid, SCARG(uap, addr), SCARG(uap, len));
#endif
	if (((vm_offset_t)SCARG(uap, addr) & PAGE_MASK) ||
	    SCARG(uap, addr) + SCARG(uap, len) < SCARG(uap, addr))
		return (EINVAL);
	map = &p->p_vmspace->vm_map;
	addr = (vm_offset_t)SCARG(uap, addr);
	size = (vm_size_t)SCARG(uap, len);
	/*
	 * XXX Gak!  If size is zero we are supposed to sync "all modified
	 * pages with the region containing addr".  Unfortunately, we
	 * don't really keep track of individual mmaps so we approximate
	 * by flushing the range of the map entry containing addr.
	 * This can be incorrect if the region splits or is coalesced
	 * with a neighbor.
	 */
	if (size == 0) {
		vm_map_entry_t entry;

		vm_map_lock_read(map);
		rv = vm_map_lookup_entry(map, addr, &entry);
		vm_map_unlock_read(map);
		if (!rv)
			return (EINVAL);
		addr = entry->start;
		size = entry->end - entry->start;
	}
#ifdef DEBUG
	if (mmapdebug & MDB_SYNC)
		printf("msync: cleaning/flushing address range [%x-%x)\n",
		       addr, addr+size);
#endif
	/*
	 * Could pass this in as a third flag argument to implement
	 * Sun's MS_ASYNC.
	 */
	syncio = TRUE;
	/*
	 * XXX bummer, gotta flush all cached pages to ensure
	 * consistency with the file system cache.  Otherwise, we could
	 * pass this in to implement Sun's MS_INVALIDATE.
	 */
	invalidate = TRUE;
	/*
	 * Clean the pages and interpret the return value.
	 */
	rv = vm_map_clean(map, addr, addr+size, syncio, invalidate);
	switch (rv) {
	case KERN_SUCCESS:
		break;
	case KERN_INVALID_ADDRESS:
		return (EINVAL);	/* Sun returns ENOMEM? */
	case KERN_FAILURE:
		return (EIO);
	default:
		return (EINVAL);
	}
	return (0);
}

int
munmap(p, uap, retval)
	register struct proc *p;
	register struct munmap_args /* {
		syscallarg(caddr_t) addr;
		syscallarg(int) len;
	} */ *uap;
	register_t *retval;
{
	vm_offset_t addr;
	vm_size_t size;
	vm_map_t map;

#ifdef DEBUG
	if (mmapdebug & MDB_FOLLOW)
		printf("munmap(%d): addr %x len %x\n",
		       p->p_pid, SCARG(uap, addr), SCARG(uap, len));
#endif

	addr = (vm_offset_t) SCARG(uap, addr);
	if ((addr & PAGE_MASK) || SCARG(uap, len) < 0)
		return(EINVAL);
	size = (vm_size_t) round_page(SCARG(uap, len));
	if (size == 0)
		return(0);
	/*
	 * Check for illegal addresses.  Watch out for address wrap...
	 * Note that VM_*_ADDRESS are not constants due to casts (argh).
	 */
	if (VM_MAXUSER_ADDRESS > 0 && addr + size >= VM_MAXUSER_ADDRESS)
		return (EINVAL);
	if (VM_MIN_ADDRESS > 0 && addr < VM_MIN_ADDRESS)
		return (EINVAL);
	if (addr > addr + size)
		return (EINVAL);
	map = &p->p_vmspace->vm_map;
	/*
	 * Make sure entire range is allocated.
	 * XXX this seemed overly restrictive, so we relaxed it.