/*
- * Copyright (c) 1982, 1986, 1989 Regents of the University of California.
- * All rights reserved.
+ * Copyright (c) 1982, 1986, 1989, 1993
+ * The Regents of the University of California. All rights reserved.
+ * (c) UNIX System Laboratories, Inc.
+ * All or some portions of this file are derived from material licensed
+ * to the University of California by American Telephone and Telegraph
+ * Co. or Unix System Laboratories, Inc. and are reproduced herein with
+ * the permission of UNIX System Laboratories, Inc.
*
- * Redistribution and use in source and binary forms are permitted
- * provided that the above copyright notice and this paragraph are
- * duplicated in all such forms and that any documentation,
- * advertising materials, and other materials related to such
- * distribution and use acknowledge that the software was developed
- * by the University of California, Berkeley. The name of the
- * University may not be used to endorse or promote products derived
- * from this software without specific prior written permission.
- * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
+ * %sccs.include.redist.c%
*
- * @(#)vfs_vnops.c 7.17 (Berkeley) %G%
+ * @(#)vfs_vnops.c 8.12 (Berkeley) %G%
*/
-#include "param.h"
-#include "systm.h"
-#include "user.h"
-#include "kernel.h"
-#include "file.h"
-#include "stat.h"
-#include "buf.h"
-#include "proc.h"
-#include "uio.h"
-#include "socket.h"
-#include "socketvar.h"
-#include "mount.h"
-#include "vnode.h"
-#include "ioctl.h"
-#include "tty.h"
-#include "tsleep.h"
+#include <sys/param.h>
+#include <sys/systm.h>
+#include <sys/kernel.h>
+#include <sys/file.h>
+#include <sys/stat.h>
+#include <sys/buf.h>
+#include <sys/proc.h>
+#include <sys/mount.h>
+#include <sys/namei.h>
+#include <sys/vnode.h>
+#include <sys/ioctl.h>
+#include <sys/tty.h>
+
+#include <vm/vm.h>
-int vn_read(), vn_write(), vn_ioctl(), vn_select(), vn_close();
struct fileops vnops =
- { vn_read, vn_write, vn_ioctl, vn_select, vn_close };
+ { vn_read, vn_write, vn_ioctl, vn_select, vn_closefile };
/*
* Common code for vnode open operations.
int fmode, cmode;
{
register struct vnode *vp;
+ register struct proc *p = ndp->ni_cnd.cn_proc;
+ register struct ucred *cred = p->p_ucred;
struct vattr vat;
struct vattr *vap = &vat;
int error;
- if (fmode & FCREAT) {
- ndp->ni_nameiop = CREATE | LOCKPARENT | LOCKLEAF;
- if ((fmode & FEXCL) == 0)
- ndp->ni_nameiop |= FOLLOW;
+ if (fmode & O_CREAT) {
+ ndp->ni_cnd.cn_nameiop = CREATE;
+ ndp->ni_cnd.cn_flags = LOCKPARENT | LOCKLEAF;
+ if ((fmode & O_EXCL) == 0)
+ ndp->ni_cnd.cn_flags |= FOLLOW;
if (error = namei(ndp))
return (error);
if (ndp->ni_vp == NULL) {
- vattr_null(vap);
+ VATTR_NULL(vap);
vap->va_type = VREG;
vap->va_mode = cmode;
- if (error = VOP_CREATE(ndp, vap))
+ if (fmode & O_EXCL)
+ vap->va_vaflags |= VA_EXCLUSIVE;
+ VOP_LEASE(ndp->ni_dvp, p, cred, LEASE_WRITE);
+ if (error = VOP_CREATE(ndp->ni_dvp, &ndp->ni_vp,
+ &ndp->ni_cnd, vap))
return (error);
- fmode &= ~FTRUNC;
+ fmode &= ~O_TRUNC;
vp = ndp->ni_vp;
} else {
+ VOP_ABORTOP(ndp->ni_dvp, &ndp->ni_cnd);
if (ndp->ni_dvp == ndp->ni_vp)
vrele(ndp->ni_dvp);
- else if (ndp->ni_dvp != NULL)
+ else
vput(ndp->ni_dvp);
ndp->ni_dvp = NULL;
vp = ndp->ni_vp;
- if (fmode & FEXCL) {
+ if (fmode & O_EXCL) {
error = EEXIST;
goto bad;
}
- fmode &= ~FCREAT;
+ fmode &= ~O_CREAT;
}
} else {
- ndp->ni_nameiop = LOOKUP | FOLLOW | LOCKLEAF;
+ ndp->ni_cnd.cn_nameiop = LOOKUP;
+ ndp->ni_cnd.cn_flags = FOLLOW | LOCKLEAF;
if (error = namei(ndp))
return (error);
vp = ndp->ni_vp;
error = EOPNOTSUPP;
goto bad;
}
- if ((fmode & FCREAT) == 0) {
+ if ((fmode & O_CREAT) == 0) {
if (fmode & FREAD) {
- if (error = VOP_ACCESS(vp, VREAD, ndp->ni_cred))
+ if (error = VOP_ACCESS(vp, VREAD, cred, p))
goto bad;
}
- if (fmode & (FWRITE|FTRUNC)) {
+ if (fmode & (FWRITE | O_TRUNC)) {
if (vp->v_type == VDIR) {
error = EISDIR;
goto bad;
}
if ((error = vn_writechk(vp)) ||
- (error = VOP_ACCESS(vp, VWRITE, ndp->ni_cred)))
+ (error = VOP_ACCESS(vp, VWRITE, cred, p)))
goto bad;
}
}
- if (fmode & FTRUNC) {
- vattr_null(vap);
+ if (fmode & O_TRUNC) {
+ VOP_UNLOCK(vp, 0, p); /* XXX */
+ VOP_LEASE(vp, p, cred, LEASE_WRITE);
+ vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p); /* XXX */
+ VATTR_NULL(vap);
vap->va_size = 0;
- if (error = VOP_SETATTR(vp, vap, ndp->ni_cred))
+ if (error = VOP_SETATTR(vp, vap, cred, p))
goto bad;
}
- VOP_UNLOCK(vp);
- if (setjmp(&u.u_qsave)) {
- if (error == 0)
- error = EINTR;
- return (error);
- }
- error = VOP_OPEN(vp, fmode, ndp->ni_cred);
- if (error)
- vrele(vp);
- return (error);
-
+ if (error = VOP_OPEN(vp, fmode, cred, p))
+ goto bad;
+ if (fmode & FWRITE)
+ vp->v_writecount++;
+ return (0);
bad:
vput(vp);
- return(error);
+ return (error);
}
/*
* unless the file is a socket or a block or character
* device resident on the file system.
*/
- if ((vp->v_mount->m_flag & M_RDONLY) && vp->v_type != VCHR &&
- vp->v_type != VBLK && vp->v_type != VSOCK)
- return (EROFS);
+ if (vp->v_mount->mnt_flag & MNT_RDONLY) {
+ switch (vp->v_type) {
+ case VREG: case VDIR: case VLNK:
+ return (EROFS);
+ }
+ }
/*
* If there's shared text associated with
* the vnode, try to free it up once. If
* we fail, we can't allow writing.
*/
- if (vp->v_flag & VTEXT)
- xrele(vp);
- if (vp->v_flag & VTEXT)
+ if ((vp->v_flag & VTEXT) && !vnode_pager_uncache(vp))
return (ETXTBSY);
return (0);
}
/*
- * Vnode version of rdwri() for calls on file systems.
+ * Vnode close call
*/
-vn_rdwr(rw, vp, base, len, offset, segflg, ioflg, cred, aresid)
+vn_close(vp, flags, cred, p)
+ register struct vnode *vp;
+ int flags;
+ struct ucred *cred;
+ struct proc *p;
+{
+ int error;
+
+ if (flags & FWRITE)
+ vp->v_writecount--;
+ error = VOP_CLOSE(vp, flags, cred, p);
+ vrele(vp);
+ return (error);
+}
+
+/*
+ * Package up an I/O request on a vnode into a uio and do it.
+ */
+vn_rdwr(rw, vp, base, len, offset, segflg, ioflg, cred, aresid, p)
enum uio_rw rw;
struct vnode *vp;
caddr_t base;
int ioflg;
struct ucred *cred;
int *aresid;
+ struct proc *p;
{
struct uio auio;
struct iovec aiov;
int error;
if ((ioflg & IO_NODELOCKED) == 0)
- VOP_LOCK(vp);
+ vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
auio.uio_iov = &aiov;
auio.uio_iovcnt = 1;
aiov.iov_base = base;
auio.uio_offset = offset;
auio.uio_segflg = segflg;
auio.uio_rw = rw;
- if (rw == UIO_READ)
+ auio.uio_procp = p;
+ if (rw == UIO_READ) {
error = VOP_READ(vp, &auio, ioflg, cred);
- else
+ } else {
error = VOP_WRITE(vp, &auio, ioflg, cred);
+ }
if (aresid)
*aresid = auio.uio_resid;
else
if (auio.uio_resid && error == 0)
error = EIO;
if ((ioflg & IO_NODELOCKED) == 0)
- VOP_UNLOCK(vp);
+ VOP_UNLOCK(vp, 0, p);
return (error);
}
+/*
+ * File table vnode read routine.
+ */
vn_read(fp, uio, cred)
struct file *fp;
struct uio *uio;
struct ucred *cred;
{
- register struct vnode *vp = (struct vnode *)fp->f_data;
+ struct vnode *vp = (struct vnode *)fp->f_data;
+ struct proc *p = uio->uio_procp;
int count, error;
- VOP_LOCK(vp);
+ VOP_LEASE(vp, p, cred, LEASE_READ);
+ vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
uio->uio_offset = fp->f_offset;
count = uio->uio_resid;
- error = VOP_READ(vp, uio, (fp->f_flag & FNDELAY) ? IO_NDELAY : 0, cred);
+ error = VOP_READ(vp, uio, (fp->f_flag & FNONBLOCK) ? IO_NDELAY : 0,
+ cred);
fp->f_offset += count - uio->uio_resid;
- VOP_UNLOCK(vp);
+ VOP_UNLOCK(vp, 0, p);
return (error);
}
+/*
+ * File table vnode write routine.
+ */
vn_write(fp, uio, cred)
struct file *fp;
struct uio *uio;
struct ucred *cred;
{
- register struct vnode *vp = (struct vnode *)fp->f_data;
- int count, error, ioflag = 0;
+ struct vnode *vp = (struct vnode *)fp->f_data;
+ struct proc *p = uio->uio_procp;
+ int count, error, ioflag = IO_UNIT;
- if (vp->v_type == VREG && (fp->f_flag & FAPPEND))
+ if (vp->v_type == VREG && (fp->f_flag & O_APPEND))
ioflag |= IO_APPEND;
- if (fp->f_flag & FNDELAY)
+ if (fp->f_flag & FNONBLOCK)
ioflag |= IO_NDELAY;
- VOP_LOCK(vp);
+ if ((fp->f_flag & O_FSYNC) || (vp->v_mount->mnt_flag & MNT_SYNCHRONOUS))
+ ioflag |= IO_SYNC;
+ VOP_LEASE(vp, p, cred, LEASE_WRITE);
+ vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
uio->uio_offset = fp->f_offset;
count = uio->uio_resid;
error = VOP_WRITE(vp, uio, ioflag, cred);
fp->f_offset = uio->uio_offset;
else
fp->f_offset += count - uio->uio_resid;
- VOP_UNLOCK(vp);
+ VOP_UNLOCK(vp, 0, p);
return (error);
}
/*
- * Get stat info for a vnode.
+ * File table vnode stat routine.
*/
-vn_stat(vp, sb)
+vn_stat(vp, sb, p)
struct vnode *vp;
register struct stat *sb;
+ struct proc *p;
{
struct vattr vattr;
register struct vattr *vap;
u_short mode;
vap = &vattr;
- error = VOP_GETATTR(vp, vap, u.u_cred);
+ error = VOP_GETATTR(vp, vap, p->p_ucred, p);
if (error)
return (error);
/*
sb->st_gid = vap->va_gid;
sb->st_rdev = vap->va_rdev;
sb->st_size = vap->va_size;
- sb->st_atime = vap->va_atime.tv_sec;
- sb->st_spare1 = 0;
- sb->st_mtime = vap->va_mtime.tv_sec;
- sb->st_spare2 = 0;
- sb->st_ctime = vap->va_ctime.tv_sec;
- sb->st_spare3 = 0;
+ sb->st_atimespec = vap->va_atime;
+ sb->st_mtimespec = vap->va_mtime;
+ sb->st_ctimespec = vap->va_ctime;
sb->st_blksize = vap->va_blocksize;
sb->st_flags = vap->va_flags;
sb->st_gen = vap->va_gen;
}
/*
- * Vnode ioctl call
+ * File table vnode ioctl routine.
*/
-vn_ioctl(fp, com, data)
+vn_ioctl(fp, com, data, p)
struct file *fp;
- int com;
+ u_long com;
caddr_t data;
+ struct proc *p;
{
register struct vnode *vp = ((struct vnode *)fp->f_data);
struct vattr vattr;
case VREG:
case VDIR:
if (com == FIONREAD) {
- if (error = VOP_GETATTR(vp, &vattr, u.u_cred))
+ if (error = VOP_GETATTR(vp, &vattr, p->p_ucred, p))
return (error);
- *(off_t *)data = vattr.va_size - fp->f_offset;
+ *(int *)data = vattr.va_size - fp->f_offset;
return (0);
}
if (com == FIONBIO || com == FIOASYNC) /* XXX */
case VFIFO:
case VCHR:
case VBLK:
- u.u_r.r_val1 = 0;
- if (setjmp(&u.u_qsave)) {
- if ((u.u_sigintr & sigmask(u.u_procp->p_cursig)) != 0)
- return(EINTR);
- u.u_eosys = RESTARTSYS;
- return (0);
- }
- error = VOP_IOCTL(vp, com, data, fp->f_flag, u.u_cred);
+ error = VOP_IOCTL(vp, com, data, fp->f_flag, p->p_ucred, p);
if (error == 0 && com == TIOCSCTTY) {
- u.u_procp->p_session->s_ttyvp = vp;
+ if (p->p_session->s_ttyvp)
+ vrele(p->p_session->s_ttyvp);
+ p->p_session->s_ttyvp = vp;
VREF(vp);
}
return (error);
}
/*
- * Vnode select call
+ * File table vnode select routine.
*/
-vn_select(fp, which)
+vn_select(fp, which, p)
struct file *fp;
int which;
+ struct proc *p;
{
- return(VOP_SELECT(((struct vnode *)fp->f_data), which, fp->f_flag,
- u.u_cred));
-}
-/*
- * Vnode close call
- */
-vn_close(fp)
- register struct file *fp;
-{
- struct vnode *vp = ((struct vnode *)fp->f_data);
- int error;
-
- if (fp->f_flag & (FSHLOCK|FEXLOCK))
- vn_unlock(fp, FSHLOCK|FEXLOCK);
- /*
- * Must delete vnode reference from this file entry
- * before VOP_CLOSE, so that only other references
- * will prevent close.
- */
- fp->f_data = (caddr_t) 0;
- error = VOP_CLOSE(vp, fp->f_flag, u.u_cred);
- vrele(vp);
- return (error);
+ return (VOP_SELECT(((struct vnode *)fp->f_data), which, fp->f_flag,
+ fp->f_cred, p));
}
/*
- * Place an advisory lock on a vnode.
- * !! THIS IMPLIES THAT ALL STATEFUL FILE SERVERS WILL USE file table entries
+ * Check that the vnode is still valid, and if so
+ * acquire requested lock.
*/
-vn_lock(fp, cmd)
- register struct file *fp;
- int cmd;
-{
- register int priority = PLOCK;
- register struct vnode *vp = (struct vnode *)fp->f_data;
-
- if ((cmd & LOCK_EX) == 0)
- priority += 4;
- if (setjmp(&u.u_qsave)) {
- if ((u.u_sigintr & sigmask(u.u_procp->p_cursig)) != 0)
- return(EINTR);
- u.u_eosys = RESTARTSYS;
- return (0);
- }
- /*
- * If there's a exclusive lock currently applied
- * to the file, then we've gotta wait for the
- * lock with everyone else.
- */
-again:
- while (vp->v_flag & VEXLOCK) {
- /*
- * If we're holding an exclusive
- * lock, then release it.
- */
- if (fp->f_flag & FEXLOCK) {
- vn_unlock(fp, FEXLOCK);
- continue;
- }
- if (cmd & LOCK_NB)
- return (EWOULDBLOCK);
- vp->v_flag |= VLWAIT;
- tsleep((caddr_t)&vp->v_exlockc, priority, SLP_EXLCK, 0);
- }
- if ((cmd & LOCK_EX) && (vp->v_flag & VSHLOCK)) {
- /*
- * Must wait for any shared locks to finish
- * before we try to apply a exclusive lock.
- *
- * If we're holding a shared
- * lock, then release it.
- */
- if (fp->f_flag & FSHLOCK) {
- vn_unlock(fp, FSHLOCK);
- goto again;
- }
- if (cmd & LOCK_NB)
- return (EWOULDBLOCK);
- vp->v_flag |= VLWAIT;
- tsleep((caddr_t)&vp->v_shlockc, PLOCK, SLP_SHLCK, 0);
- goto again;
- }
- if (fp->f_flag & FEXLOCK)
- panic("vn_lock");
- if (cmd & LOCK_EX) {
- cmd &= ~LOCK_SH;
- vp->v_exlockc++;
- vp->v_flag |= VEXLOCK;
- fp->f_flag |= FEXLOCK;
- }
- if ((cmd & LOCK_SH) && (fp->f_flag & FSHLOCK) == 0) {
- vp->v_shlockc++;
- vp->v_flag |= VSHLOCK;
- fp->f_flag |= FSHLOCK;
- }
- return (0);
-}
-
-/*
- * Unlock a file.
- */
-vn_unlock(fp, kind)
- register struct file *fp;
- int kind;
-{
- register struct vnode *vp = (struct vnode *)fp->f_data;
+int
+vn_lock(vp, flags, p)
+ struct vnode *vp;
int flags;
-
- kind &= fp->f_flag;
- if (vp == NULL || kind == 0)
- return;
- flags = vp->v_flag;
- if (kind & FSHLOCK) {
- if ((flags & VSHLOCK) == 0)
- panic("vn_unlock: SHLOCK");
- if (--vp->v_shlockc == 0) {
- vp->v_flag &= ~VSHLOCK;
- if (flags & VLWAIT)
- wakeup((caddr_t)&vp->v_shlockc);
- }
- fp->f_flag &= ~FSHLOCK;
- }
- if (kind & FEXLOCK) {
- if ((flags & VEXLOCK) == 0)
- panic("vn_unlock: EXLOCK");
- if (--vp->v_exlockc == 0) {
- vp->v_flag &= ~(VEXLOCK|VLWAIT);
- if (flags & VLWAIT)
- wakeup((caddr_t)&vp->v_exlockc);
- }
- fp->f_flag &= ~FEXLOCK;
- }
-}
-
-/*
- * vn_fhtovp() - convert a fh to a vnode ptr (optionally locked)
- * - look up fsid in mount list (if not found ret error)
- * - get vp by calling VFS_FHTOVP() macro
- * - if lockflag lock it with VOP_LOCK()
- */
-vn_fhtovp(fhp, lockflag, vpp)
- fhandle_t *fhp;
- int lockflag;
- struct vnode **vpp;
+ struct proc *p;
{
- register struct mount *mp;
-
- if ((mp = getvfs(&fhp->fh_fsid)) == NULL)
- return (ESTALE);
- if (VFS_FHTOVP(mp, &fhp->fh_fid, vpp))
- return (ESTALE);
- if (!lockflag)
- VOP_UNLOCK(*vpp);
- return (0);
-}
-
-/*
- * Noop
- */
-vfs_noop()
-{
-
- return (ENXIO);
+ int error;
+
+ do {
+ if ((flags & LK_INTERLOCK) == 0)
+ simple_lock(&vp->v_interlock);
+ if (vp->v_flag & VXLOCK) {
+ vp->v_flag |= VXWANT;
+ simple_unlock(&vp->v_interlock);
+ tsleep((caddr_t)vp, PINOD, "vn_lock", 0);
+ error = ENOENT;
+ } else {
+ error = VOP_LOCK(vp, flags | LK_INTERLOCK, p);
+ if (error == 0)
+ return (error);
+ }
+ flags &= ~LK_INTERLOCK;
+ } while (flags & LK_RETRY);
+ return (error);
}
/*
- * Null op
+ * File table vnode close routine.
*/
-vfs_nullop()
+vn_closefile(fp, p)
+ struct file *fp;
+ struct proc *p;
{
- return (0);
+ return (vn_close(((struct vnode *)fp->f_data), fp->f_flag,
+ fp->f_cred, p));
}