NetBSD/sys/nfs/nfs_vnops.c

1909 lines
48 KiB
C

/*
* Copyright (c) 1989 The Regents of the University of California.
* All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Rick Macklem at The University of Guelph.
*
* 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: @(#)nfs_vnops.c 7.60 (Berkeley) 5/24/91
* $Id: nfs_vnops.c,v 1.6 1993/06/03 01:12:44 cgd Exp $
*/
/*
* vnode op calls for sun nfs version 2
*/
#include "param.h"
#include "proc.h"
#include "kernel.h"
#include "systm.h"
#include "mount.h"
#include "buf.h"
#include "malloc.h"
#include "mbuf.h"
#include "conf.h"
#include "namei.h"
#include "vnode.h"
#include "specdev.h"
#include "fifo.h"
#include "../ufs/quota.h"
#include "../ufs/inode.h"
#include "../ufs/dir.h"
#include "nfsv2.h"
#include "nfs.h"
#include "nfsnode.h"
#include "nfsmount.h"
#include "xdr_subs.h"
#include "nfsm_subs.h"
#include "nfsiom.h"
/* Defs */
#define TRUE 1
#define FALSE 0
/*
* Global vfs data structures for nfs
*/
struct vnodeops nfsv2_vnodeops = {
nfs_lookup, /* lookup */
nfs_create, /* create */
nfs_mknod, /* mknod */
nfs_open, /* open */
nfs_close, /* close */
nfs_access, /* access */
nfs_getattr, /* getattr */
nfs_setattr, /* setattr */
nfs_read, /* read */
nfs_write, /* write */
nfs_ioctl, /* ioctl */
nfs_select, /* select */
nfs_mmap, /* mmap */
nfs_fsync, /* fsync */
nfs_seek, /* seek */
nfs_remove, /* remove */
nfs_link, /* link */
nfs_rename, /* rename */
nfs_mkdir, /* mkdir */
nfs_rmdir, /* rmdir */
nfs_symlink, /* symlink */
nfs_readdir, /* readdir */
nfs_readlink, /* readlink */
nfs_abortop, /* abortop */
nfs_inactive, /* inactive */
nfs_reclaim, /* reclaim */
nfs_lock, /* lock */
nfs_unlock, /* unlock */
nfs_bmap, /* bmap */
nfs_strategy, /* strategy */
nfs_print, /* print */
nfs_islocked, /* islocked */
nfs_advlock, /* advlock */
};
/*
* Special device vnode ops
*/
struct vnodeops spec_nfsv2nodeops = {
spec_lookup, /* lookup */
spec_create, /* create */
spec_mknod, /* mknod */
spec_open, /* open */
spec_close, /* close */
nfs_access, /* access */
nfs_getattr, /* getattr */
nfs_setattr, /* setattr */
spec_read, /* read */
spec_write, /* write */
spec_ioctl, /* ioctl */
spec_select, /* select */
spec_mmap, /* mmap */
spec_fsync, /* fsync */
spec_seek, /* seek */
spec_remove, /* remove */
spec_link, /* link */
spec_rename, /* rename */
spec_mkdir, /* mkdir */
spec_rmdir, /* rmdir */
spec_symlink, /* symlink */
spec_readdir, /* readdir */
spec_readlink, /* readlink */
spec_abortop, /* abortop */
nfs_inactive, /* inactive */
nfs_reclaim, /* reclaim */
nfs_lock, /* lock */
nfs_unlock, /* unlock */
spec_bmap, /* bmap */
spec_strategy, /* strategy */
nfs_print, /* print */
nfs_islocked, /* islocked */
spec_advlock, /* advlock */
};
#ifdef FIFO
struct vnodeops fifo_nfsv2nodeops = {
fifo_lookup, /* lookup */
fifo_create, /* create */
fifo_mknod, /* mknod */
fifo_open, /* open */
fifo_close, /* close */
nfs_access, /* access */
nfs_getattr, /* getattr */
nfs_setattr, /* setattr */
fifo_read, /* read */
fifo_write, /* write */
fifo_ioctl, /* ioctl */
fifo_select, /* select */
fifo_mmap, /* mmap */
fifo_fsync, /* fsync */
fifo_seek, /* seek */
fifo_remove, /* remove */
fifo_link, /* link */
fifo_rename, /* rename */
fifo_mkdir, /* mkdir */
fifo_rmdir, /* rmdir */
fifo_symlink, /* symlink */
fifo_readdir, /* readdir */
fifo_readlink, /* readlink */
fifo_abortop, /* abortop */
nfs_inactive, /* inactive */
nfs_reclaim, /* reclaim */
nfs_lock, /* lock */
nfs_unlock, /* unlock */
fifo_bmap, /* bmap */
fifo_badop, /* strategy */
nfs_print, /* print */
nfs_islocked, /* islocked */
fifo_advlock, /* advlock */
};
#endif /* FIFO */
/*
* Global vars
*/
extern u_long nfs_procids[NFS_NPROCS];
extern u_long nfs_prog, nfs_vers;
extern char nfsiobuf[MAXPHYS+NBPG];
struct buf nfs_bqueue; /* Queue head for nfsiod's */
struct proc *nfs_iodwant[NFS_MAXASYNCDAEMON];
enum vtype ntov_type[7] = { VNON, VREG, VDIR, VBLK, VCHR, VLNK, VNON };
int nfs_numasync = 0;
/*
* nfs null call from vfs.
*/
nfs_null(vp, cred, p)
struct vnode *vp;
struct ucred *cred;
struct proc *p;
{
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb;
nfsm_reqhead(nfs_procids[NFSPROC_NULL], cred, 0);
nfsm_request(vp, NFSPROC_NULL, p, 0);
nfsm_reqdone;
return (error);
}
/*
* nfs access vnode op.
* Essentially just get vattr and then imitate iaccess()
*/
nfs_access(vp, mode, cred, p)
struct vnode *vp;
int mode;
register struct ucred *cred;
struct proc *p;
{
register struct vattr *vap;
register gid_t *gp;
struct vattr vattr;
register int i;
int error;
/*
* If you're the super-user,
* you always get access.
*/
if (cred->cr_uid == 0)
return (0);
vap = &vattr;
if (error = nfs_dogetattr(vp, vap, cred, 0, p))
return (error);
/*
* Access check is based on only one of owner, group, public.
* If not owner, then check group. If not a member of the
* group, then check public access.
*/
if (cred->cr_uid != vap->va_uid) {
mode >>= 3;
gp = cred->cr_groups;
for (i = 0; i < cred->cr_ngroups; i++, gp++)
if (vap->va_gid == *gp)
goto found;
mode >>= 3;
found:
;
}
if ((vap->va_mode & mode) == mode)
return (0);
return (EACCES);
}
/*
* nfs open vnode op
* Just check to see if the type is ok
*/
/* ARGSUSED */
nfs_open(vp, mode, cred, p)
struct vnode *vp;
int mode;
struct ucred *cred;
struct proc *p;
{
register enum vtype vtyp;
vtyp = vp->v_type;
if (vtyp == VREG || vtyp == VDIR || vtyp == VLNK)
return (0);
else
return (EACCES);
}
/*
* nfs close vnode op
* For reg files, invalidate any buffer cache entries.
*/
/* ARGSUSED */
nfs_close(vp, fflags, cred, p)
register struct vnode *vp;
int fflags;
struct ucred *cred;
struct proc *p;
{
register struct nfsnode *np = VTONFS(vp);
int error = 0;
if (vp->v_type == VREG && (np->n_flag & NMODIFIED)) {
nfs_lock(vp);
np->n_flag &= ~NMODIFIED;
vinvalbuf(vp, TRUE);
np->n_attrstamp = 0;
if (np->n_flag & NWRITEERR) {
np->n_flag &= ~NWRITEERR;
error = np->n_error;
}
nfs_unlock(vp);
}
return (error);
}
/*
* nfs getattr call from vfs.
*/
nfs_getattr(vp, vap, cred, p)
register struct vnode *vp;
struct vattr *vap;
struct ucred *cred;
struct proc *p;
{
return (nfs_dogetattr(vp, vap, cred, 0, p));
}
nfs_dogetattr(vp, vap, cred, tryhard, p)
register struct vnode *vp;
struct vattr *vap;
struct ucred *cred;
int tryhard;
struct proc *p;
{
register caddr_t cp;
register long t1;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
/* First look in the cache.. */
if (nfs_getattrcache(vp, vap) == 0)
return (0);
nfsstats.rpccnt[NFSPROC_GETATTR]++;
nfsm_reqhead(nfs_procids[NFSPROC_GETATTR], cred, NFSX_FH);
nfsm_fhtom(vp);
nfsm_request(vp, NFSPROC_GETATTR, p, tryhard);
nfsm_loadattr(vp, vap);
nfsm_reqdone;
return (error);
}
/*
* nfs setattr call.
*/
nfs_setattr(vp, vap, cred, p)
register struct vnode *vp;
register struct vattr *vap;
struct ucred *cred;
struct proc *p;
{
register struct nfsv2_sattr *sp;
register caddr_t cp;
register long t1;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
struct nfsnode *np;
nfsstats.rpccnt[NFSPROC_SETATTR]++;
nfsm_reqhead(nfs_procids[NFSPROC_SETATTR], cred, NFSX_FH+NFSX_SATTR);
nfsm_fhtom(vp);
nfsm_build(sp, struct nfsv2_sattr *, NFSX_SATTR);
if (vap->va_mode == 0xffff)
sp->sa_mode = VNOVAL;
else
sp->sa_mode = vtonfs_mode(vp->v_type, vap->va_mode);
if (vap->va_uid == 0xffff)
sp->sa_uid = VNOVAL;
else
sp->sa_uid = txdr_unsigned(vap->va_uid);
if (vap->va_gid == 0xffff)
sp->sa_gid = VNOVAL;
else
sp->sa_gid = txdr_unsigned(vap->va_gid);
sp->sa_size = txdr_unsigned(vap->va_size);
/* jfw@ksr.com 6/2/93 */
#if 0 /* bad assumption; Suns (at least) make full use of usec field */
sp->sa_atime.tv_sec = txdr_unsigned(vap->va_atime.tv_sec);
sp->sa_atime.tv_usec = txdr_unsigned(vap->va_flags);
#else
txdr_time(&vap->va_atime, &sp->sa_atime);
#endif
txdr_time(&vap->va_mtime, &sp->sa_mtime);
if (vap->va_size != VNOVAL || vap->va_mtime.tv_sec != VNOVAL ||
vap->va_atime.tv_sec != VNOVAL) {
np = VTONFS(vp);
if (np->n_flag & NMODIFIED) {
np->n_flag &= ~NMODIFIED;
if (vap->va_size == 0)
vinvalbuf(vp, FALSE);
else
vinvalbuf(vp, TRUE);
np->n_attrstamp = 0;
}
}
nfsm_request(vp, NFSPROC_SETATTR, p, 1);
nfsm_loadattr(vp, (struct vattr *)0);
/* should we fill in any vap fields ?? */
nfsm_reqdone;
return (error);
}
/*
* nfs lookup call, one step at a time...
* First look in cache
* If not found, unlock the directory nfsnode and do the rpc
*/
nfs_lookup(vp, ndp, p)
register struct vnode *vp;
register struct nameidata *ndp;
struct proc *p;
{
register struct vnode *vdp;
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos, cp2;
u_long xid;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
struct vnode *newvp;
long len;
nfsv2fh_t *fhp;
struct nfsnode *np;
int lockparent, wantparent, flag, error = 0;
ndp->ni_dvp = vp;
ndp->ni_vp = NULL;
if (vp->v_type != VDIR)
return (ENOTDIR);
lockparent = ndp->ni_nameiop & LOCKPARENT;
flag = ndp->ni_nameiop & OPMASK;
wantparent = ndp->ni_nameiop & (LOCKPARENT|WANTPARENT);
if ((error = cache_lookup(ndp)) && error != ENOENT) {
struct vattr vattr;
int vpid;
vdp = ndp->ni_vp;
vpid = vdp->v_id;
/*
* See the comment starting `Step through' in ufs/ufs_lookup.c
* for an explanation of the locking protocol
*/
if (vp == vdp) {
VREF(vdp);
error = 0;
} else if (ndp->ni_isdotdot) {
nfs_unlock(vp);
error = vget(vdp);
if (!error && lockparent && *ndp->ni_next == '\0')
nfs_lock(vp);
} else {
error = vget(vdp);
if (!lockparent || error || *ndp->ni_next != '\0')
nfs_unlock(vp);
}
if (!error) {
if (vpid == vdp->v_id) {
if (!nfs_dogetattr(vdp, &vattr, ndp->ni_cred, 0, p)&&
vattr.va_ctime.tv_sec == VTONFS(vdp)->n_ctime) {
nfsstats.lookupcache_hits++;
if (flag != LOOKUP && *ndp->ni_next == 0)
ndp->ni_nameiop |= SAVENAME;
return (0);
}
cache_purge(vdp);
}
nfs_nput(vdp);
if (lockparent && vdp != vp && *ndp->ni_next == '\0')
nfs_unlock(vp);
}
ndp->ni_vp = NULLVP;
} else
nfs_unlock(vp);
error = 0;
nfsstats.lookupcache_misses++;
nfsstats.rpccnt[NFSPROC_LOOKUP]++;
len = ndp->ni_namelen;
nfsm_reqhead(nfs_procids[NFSPROC_LOOKUP], ndp->ni_cred, NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(len));
nfsm_fhtom(vp);
nfsm_strtom(ndp->ni_ptr, len, NFS_MAXNAMLEN);
nfsm_request(vp, NFSPROC_LOOKUP, p, 0);
nfsmout:
if (error) {
if (lockparent || (flag != CREATE && flag != RENAME) ||
*ndp->ni_next != 0)
nfs_lock(vp);
if (flag != LOOKUP && *ndp->ni_next == 0)
ndp->ni_nameiop |= SAVENAME;
return (error);
}
nfsm_disect(fhp,nfsv2fh_t *,NFSX_FH);
/*
* Handle DELETE and RENAME cases...
*/
if (flag == DELETE && *ndp->ni_next == 0) {
if (!bcmp(VTONFS(vp)->n_fh.fh_bytes, (caddr_t)fhp, NFSX_FH)) {
VREF(vp);
newvp = vp;
np = VTONFS(vp);
} else {
if (error = nfs_nget(vp->v_mount, fhp, &np)) {
nfs_lock(vp);
m_freem(mrep);
return (error);
}
newvp = NFSTOV(np);
}
if (error =
nfs_loadattrcache(&newvp, &md, &dpos, (struct vattr *)0)) {
nfs_lock(vp);
if (newvp != vp)
nfs_nput(newvp);
else
vrele(vp);
m_freem(mrep);
return (error);
}
ndp->ni_vp = newvp;
if (lockparent || vp == newvp)
nfs_lock(vp);
m_freem(mrep);
ndp->ni_nameiop |= SAVENAME;
return (0);
}
if (flag == RENAME && wantparent && *ndp->ni_next == 0) {
if (!bcmp(VTONFS(vp)->n_fh.fh_bytes, (caddr_t)fhp, NFSX_FH)) {
nfs_lock(vp);
m_freem(mrep);
return (EISDIR);
}
if (error = nfs_nget(vp->v_mount, fhp, &np)) {
nfs_lock(vp);
m_freem(mrep);
return (error);
}
newvp = NFSTOV(np);
if (error =
nfs_loadattrcache(&newvp, &md, &dpos, (struct vattr *)0)) {
nfs_lock(vp);
nfs_nput(newvp);
m_freem(mrep);
return (error);
}
ndp->ni_vp = newvp;
if (lockparent)
nfs_lock(vp);
m_freem(mrep);
ndp->ni_nameiop |= SAVENAME;
return (0);
}
if (!bcmp(VTONFS(vp)->n_fh.fh_bytes, (caddr_t)fhp, NFSX_FH)) {
VREF(vp);
newvp = vp;
np = VTONFS(vp);
} else if (ndp->ni_isdotdot) {
if (error = nfs_nget(vp->v_mount, fhp, &np)) {
nfs_lock(vp);
m_freem(mrep);
return (error);
}
newvp = NFSTOV(np);
} else {
if (error = nfs_nget(vp->v_mount, fhp, &np)) {
nfs_lock(vp);
m_freem(mrep);
return (error);
}
newvp = NFSTOV(np);
}
if (error = nfs_loadattrcache(&newvp, &md, &dpos, (struct vattr *)0)) {
nfs_lock(vp);
if (newvp != vp)
nfs_nput(newvp);
else
vrele(vp);
m_freem(mrep);
return (error);
}
m_freem(mrep);
if (vp == newvp || (lockparent && *ndp->ni_next == '\0'))
nfs_lock(vp);
ndp->ni_vp = newvp;
if (flag != LOOKUP && *ndp->ni_next == 0)
ndp->ni_nameiop |= SAVENAME;
if (error == 0 && ndp->ni_makeentry) {
np->n_ctime = np->n_vattr.va_ctime.tv_sec;
cache_enter(ndp);
}
return (error);
}
/*
* nfs read call.
* Just call nfs_bioread() to do the work.
*/
nfs_read(vp, uiop, ioflag, cred)
register struct vnode *vp;
struct uio *uiop;
int ioflag;
struct ucred *cred;
{
if (vp->v_type != VREG)
return (EPERM);
return (nfs_bioread(vp, uiop, ioflag, cred));
}
/*
* nfs readlink call
*/
nfs_readlink(vp, uiop, cred)
struct vnode *vp;
struct uio *uiop;
struct ucred *cred;
{
if (vp->v_type != VLNK)
return (EPERM);
return (nfs_bioread(vp, uiop, 0, cred));
}
/*
* Do a readlink rpc.
* Called by nfs_doio() from below the buffer cache.
*/
nfs_readlinkrpc(vp, uiop, cred)
register struct vnode *vp;
struct uio *uiop;
struct ucred *cred;
{
register u_long *tl;
register caddr_t cp;
register long t1;
caddr_t bpos, dpos, cp2;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
long len;
nfsstats.rpccnt[NFSPROC_READLINK]++;
nfsm_reqhead(nfs_procids[NFSPROC_READLINK], cred, NFSX_FH);
nfsm_fhtom(vp);
nfsm_request(vp, NFSPROC_READLINK, uiop->uio_procp, 0);
nfsm_strsiz(len, NFS_MAXPATHLEN);
nfsm_mtouio(uiop, len);
nfsm_reqdone;
return (error);
}
/*
* nfs read rpc call
* Ditto above
*/
nfs_readrpc(vp, uiop, cred)
register struct vnode *vp;
struct uio *uiop;
struct ucred *cred;
{
register u_long *tl;
register caddr_t cp;
register long t1;
caddr_t bpos, dpos, cp2;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
struct nfsmount *nmp;
long len, retlen, tsiz;
nmp = VFSTONFS(vp->v_mount);
tsiz = uiop->uio_resid;
while (tsiz > 0) {
nfsstats.rpccnt[NFSPROC_READ]++;
len = (tsiz > nmp->nm_rsize) ? nmp->nm_rsize : tsiz;
nfsm_reqhead(nfs_procids[NFSPROC_READ], cred, NFSX_FH+NFSX_UNSIGNED*3);
nfsm_fhtom(vp);
nfsm_build(tl, u_long *, NFSX_UNSIGNED*3);
*tl++ = txdr_unsigned(uiop->uio_offset);
*tl++ = txdr_unsigned(len);
*tl = 0;
nfsm_request(vp, NFSPROC_READ, uiop->uio_procp, 1);
nfsm_loadattr(vp, (struct vattr *)0);
nfsm_strsiz(retlen, nmp->nm_rsize);
nfsm_mtouio(uiop, retlen);
m_freem(mrep);
if (retlen < len)
tsiz = 0;
else
tsiz -= len;
}
nfsmout:
return (error);
}
/*
* nfs write call
*/
nfs_writerpc(vp, uiop, cred)
register struct vnode *vp;
struct uio *uiop;
struct ucred *cred;
{
register u_long *tl;
register caddr_t cp;
register long t1;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
struct nfsmount *nmp;
long len, tsiz;
nmp = VFSTONFS(vp->v_mount);
tsiz = uiop->uio_resid;
while (tsiz > 0) {
nfsstats.rpccnt[NFSPROC_WRITE]++;
len = (tsiz > nmp->nm_wsize) ? nmp->nm_wsize : tsiz;
nfsm_reqhead(nfs_procids[NFSPROC_WRITE], cred,
NFSX_FH+NFSX_UNSIGNED*4);
nfsm_fhtom(vp);
nfsm_build(tl, u_long *, NFSX_UNSIGNED*4);
*(tl+1) = txdr_unsigned(uiop->uio_offset);
*(tl+3) = txdr_unsigned(len);
nfsm_uiotom(uiop, len);
nfsm_request(vp, NFSPROC_WRITE, uiop->uio_procp, 1);
nfsm_loadattr(vp, (struct vattr *)0);
m_freem(mrep);
tsiz -= len;
}
nfsmout:
return (error);
}
/*
* nfs mknod call
* This is a kludge. Use a create rpc but with the IFMT bits of the mode
* set to specify the file type and the size field for rdev.
*/
/* ARGSUSED */
nfs_mknod(ndp, vap, cred, p)
struct nameidata *ndp;
struct ucred *cred;
register struct vattr *vap;
struct proc *p;
{
register struct nfsv2_sattr *sp;
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
u_long rdev;
if (vap->va_type == VCHR || vap->va_type == VBLK)
rdev = txdr_unsigned(vap->va_rdev);
#ifdef FIFO
else if (vap->va_type == VFIFO)
rdev = 0xffffffff;
#endif /* FIFO */
else {
VOP_ABORTOP(ndp);
vput(ndp->ni_dvp);
return (EOPNOTSUPP);
}
nfsstats.rpccnt[NFSPROC_CREATE]++;
nfsm_reqhead(nfs_procids[NFSPROC_CREATE], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(ndp->ni_namelen)+NFSX_SATTR);
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, ndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_build(sp, struct nfsv2_sattr *, NFSX_SATTR);
sp->sa_mode = vtonfs_mode(vap->va_type, vap->va_mode);
sp->sa_uid = txdr_unsigned(ndp->ni_cred->cr_uid);
sp->sa_gid = txdr_unsigned(ndp->ni_cred->cr_gid);
sp->sa_size = rdev;
/* or should these be VNOVAL ?? */
txdr_time(&vap->va_atime, &sp->sa_atime);
txdr_time(&vap->va_mtime, &sp->sa_mtime);
nfsm_request(ndp->ni_dvp, NFSPROC_CREATE, p, 1);
nfsm_reqdone;
FREE(ndp->ni_pnbuf, M_NAMEI);
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
nfs_nput(ndp->ni_dvp);
return (error);
}
/*
* nfs file create call
*/
nfs_create(ndp, vap, p)
register struct nameidata *ndp;
register struct vattr *vap;
struct proc *p;
{
register struct nfsv2_sattr *sp;
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos, cp2;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
nfsstats.rpccnt[NFSPROC_CREATE]++;
nfsm_reqhead(nfs_procids[NFSPROC_CREATE], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(ndp->ni_namelen)+NFSX_SATTR);
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, ndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_build(sp, struct nfsv2_sattr *, NFSX_SATTR);
sp->sa_mode = vtonfs_mode(vap->va_type, vap->va_mode);
sp->sa_uid = txdr_unsigned(ndp->ni_cred->cr_uid);
sp->sa_gid = txdr_unsigned(ndp->ni_cred->cr_gid);
sp->sa_size = txdr_unsigned(0);
/* or should these be VNOVAL ?? */
txdr_time(&vap->va_atime, &sp->sa_atime);
txdr_time(&vap->va_mtime, &sp->sa_mtime);
nfsm_request(ndp->ni_dvp, NFSPROC_CREATE, p, 1);
nfsm_mtofh(ndp->ni_dvp, ndp->ni_vp);
nfsm_reqdone;
FREE(ndp->ni_pnbuf, M_NAMEI);
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
nfs_nput(ndp->ni_dvp);
return (error);
}
/*
* nfs file remove call
* To try and make nfs semantics closer to ufs semantics, a file that has
* other processes using the vnode is renamed instead of removed and then
* removed later on the last close.
* - If v_usecount > 1
* If a rename is not already in the works
* call nfs_sillyrename() to set it up
* else
* do the remove rpc
*/
nfs_remove(ndp, p)
register struct nameidata *ndp;
struct proc *p;
{
register struct vnode *vp = ndp->ni_vp;
register struct nfsnode *np = VTONFS(ndp->ni_vp);
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
if (vp->v_usecount > 1) {
if (!np->n_sillyrename)
error = nfs_sillyrename(ndp, p);
} else {
nfsstats.rpccnt[NFSPROC_REMOVE]++;
nfsm_reqhead(nfs_procids[NFSPROC_REMOVE], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(ndp->ni_namelen));
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, ndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_request(ndp->ni_dvp, NFSPROC_REMOVE, p, 1);
nfsm_reqdone;
FREE(ndp->ni_pnbuf, M_NAMEI);
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
/*
* Kludge City: If the first reply to the remove rpc is lost..
* the reply to the retransmitted request will be ENOENT
* since the file was in fact removed
* Therefore, we cheat and return success.
*/
if (error == ENOENT)
error = 0;
}
np->n_attrstamp = 0;
if (ndp->ni_dvp == vp)
vrele(vp);
else
nfs_nput(ndp->ni_dvp);
nfs_nput(vp);
return (error);
}
/*
* nfs file remove rpc called from nfs_inactive
*/
nfs_removeit(sp, p)
register struct sillyrename *sp;
struct proc *p;
{
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
nfsstats.rpccnt[NFSPROC_REMOVE]++;
nfsm_reqhead(nfs_procids[NFSPROC_REMOVE], sp->s_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(sp->s_namlen));
nfsm_fhtom(sp->s_dvp);
nfsm_strtom(sp->s_name, sp->s_namlen, NFS_MAXNAMLEN);
nfsm_request(sp->s_dvp, NFSPROC_REMOVE, p, 1);
nfsm_reqdone;
VTONFS(sp->s_dvp)->n_flag |= NMODIFIED;
return (error);
}
/*
* nfs file rename call
*/
nfs_rename(sndp, tndp, p)
register struct nameidata *sndp, *tndp;
struct proc *p;
{
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
nfsstats.rpccnt[NFSPROC_RENAME]++;
nfsm_reqhead(nfs_procids[NFSPROC_RENAME], tndp->ni_cred,
(NFSX_FH+NFSX_UNSIGNED)*2+nfsm_rndup(sndp->ni_namelen) +
nfsm_rndup(tndp->ni_namelen)); /* or sndp->ni_cred?*/
nfsm_fhtom(sndp->ni_dvp);
nfsm_strtom(sndp->ni_ptr, sndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_fhtom(tndp->ni_dvp);
nfsm_strtom(tndp->ni_ptr, tndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_request(sndp->ni_dvp, NFSPROC_RENAME, p, 1);
nfsm_reqdone;
VTONFS(sndp->ni_dvp)->n_flag |= NMODIFIED;
VTONFS(tndp->ni_dvp)->n_flag |= NMODIFIED;
if (sndp->ni_vp->v_type == VDIR) {
if (tndp->ni_vp != NULL && tndp->ni_vp->v_type == VDIR)
cache_purge(tndp->ni_dvp);
cache_purge(sndp->ni_dvp);
}
if (tndp->ni_dvp == tndp->ni_vp)
vrele(tndp->ni_dvp);
else
vput(tndp->ni_dvp);
if (tndp->ni_vp)
vput(tndp->ni_vp);
vrele(sndp->ni_dvp);
vrele(sndp->ni_vp);
/*
* Kludge: Map ENOENT => 0 assuming that it is a reply to a retry.
*/
if (error == ENOENT)
error = 0;
return (error);
}
/*
* nfs file rename rpc called from nfs_remove() above
*/
nfs_renameit(sndp, sp, p)
register struct nameidata *sndp;
register struct sillyrename *sp;
struct proc *p;
{
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
nfsstats.rpccnt[NFSPROC_RENAME]++;
nfsm_reqhead(nfs_procids[NFSPROC_RENAME], sp->s_cred,
(NFSX_FH+NFSX_UNSIGNED)*2+nfsm_rndup(sndp->ni_namelen) +
nfsm_rndup(sp->s_namlen)); /* or sndp->ni_cred?*/
nfsm_fhtom(sndp->ni_dvp);
nfsm_strtom(sndp->ni_ptr, sndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_fhtom(sp->s_dvp);
nfsm_strtom(sp->s_name, sp->s_namlen, NFS_MAXNAMLEN);
nfsm_request(sndp->ni_dvp, NFSPROC_RENAME, p, 1);
nfsm_reqdone;
FREE(sndp->ni_pnbuf, M_NAMEI);
VTONFS(sndp->ni_dvp)->n_flag |= NMODIFIED;
VTONFS(sp->s_dvp)->n_flag |= NMODIFIED;
return (error);
}
/*
* nfs hard link create call
*/
nfs_link(vp, ndp, p)
register struct vnode *vp;
register struct nameidata *ndp;
struct proc *p;
{
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
if (ndp->ni_dvp != vp)
nfs_lock(vp);
nfsstats.rpccnt[NFSPROC_LINK]++;
nfsm_reqhead(nfs_procids[NFSPROC_LINK], ndp->ni_cred,
NFSX_FH*2+NFSX_UNSIGNED+nfsm_rndup(ndp->ni_namelen));
nfsm_fhtom(vp);
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, ndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_request(vp, NFSPROC_LINK, p, 1);
nfsm_reqdone;
FREE(ndp->ni_pnbuf, M_NAMEI);
VTONFS(vp)->n_attrstamp = 0;
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
if (ndp->ni_dvp != vp)
nfs_unlock(vp);
nfs_nput(ndp->ni_dvp);
/*
* Kludge: Map EEXIST => 0 assuming that it is a reply to a retry.
*/
if (error == EEXIST)
error = 0;
return (error);
}
/*
* nfs symbolic link create call
*/
nfs_symlink(ndp, vap, nm, p)
struct nameidata *ndp;
struct vattr *vap;
char *nm; /* is this the path ?? */
struct proc *p;
{
register struct nfsv2_sattr *sp;
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
nfsstats.rpccnt[NFSPROC_SYMLINK]++;
nfsm_reqhead(nfs_procids[NFSPROC_SYMLINK], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(ndp->ni_namelen)+NFSX_UNSIGNED);
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, ndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_strtom(nm, strlen(nm), NFS_MAXPATHLEN);
nfsm_build(sp, struct nfsv2_sattr *, NFSX_SATTR);
sp->sa_mode = vtonfs_mode(VLNK, vap->va_mode);
sp->sa_uid = txdr_unsigned(ndp->ni_cred->cr_uid);
sp->sa_gid = txdr_unsigned(ndp->ni_cred->cr_gid);
sp->sa_size = txdr_unsigned(VNOVAL);
txdr_time(&vap->va_atime, &sp->sa_atime); /* or VNOVAL ?? */
txdr_time(&vap->va_mtime, &sp->sa_mtime); /* or VNOVAL ?? */
nfsm_request(ndp->ni_dvp, NFSPROC_SYMLINK, p, 1);
nfsm_reqdone;
FREE(ndp->ni_pnbuf, M_NAMEI);
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
nfs_nput(ndp->ni_dvp);
/*
* Kludge: Map EEXIST => 0 assuming that it is a reply to a retry.
*/
if (error == EEXIST)
error = 0;
return (error);
}
/*
* nfs make dir call
*/
nfs_mkdir(ndp, vap, p)
register struct nameidata *ndp;
struct vattr *vap;
struct proc *p;
{
register struct nfsv2_sattr *sp;
register u_long *tl;
register caddr_t cp;
register long t1, t2;
register int len;
caddr_t bpos, dpos, cp2;
u_long xid;
int error = 0, firsttry = 1;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
len = ndp->ni_namelen;
nfsstats.rpccnt[NFSPROC_MKDIR]++;
nfsm_reqhead(nfs_procids[NFSPROC_MKDIR], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(len)+NFSX_SATTR);
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, len, NFS_MAXNAMLEN);
nfsm_build(sp, struct nfsv2_sattr *, NFSX_SATTR);
sp->sa_mode = vtonfs_mode(VDIR, vap->va_mode);
sp->sa_uid = txdr_unsigned(ndp->ni_cred->cr_uid);
sp->sa_gid = txdr_unsigned(ndp->ni_cred->cr_gid);
sp->sa_size = txdr_unsigned(VNOVAL);
txdr_time(&vap->va_atime, &sp->sa_atime); /* or VNOVAL ?? */
txdr_time(&vap->va_mtime, &sp->sa_mtime); /* or VNOVAL ?? */
nfsm_request(ndp->ni_dvp, NFSPROC_MKDIR, p, 1);
nfsm_mtofh(ndp->ni_dvp, ndp->ni_vp);
nfsm_reqdone;
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
/*
* Kludge: Map EEXIST => 0 assuming that you have a reply to a retry
* if we can succeed in looking up the directory.
* "firsttry" is necessary since the macros may "goto nfsmout" which
* is above the if on errors. (Ugh)
*/
if (error == EEXIST && firsttry) {
firsttry = 0;
error = 0;
nfsstats.rpccnt[NFSPROC_LOOKUP]++;
ndp->ni_vp = NULL;
nfsm_reqhead(nfs_procids[NFSPROC_LOOKUP], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(len));
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, len, NFS_MAXNAMLEN);
nfsm_request(ndp->ni_dvp, NFSPROC_LOOKUP, p, 1);
nfsm_mtofh(ndp->ni_dvp, ndp->ni_vp);
if (ndp->ni_vp->v_type != VDIR) {
vput(ndp->ni_vp);
error = EEXIST;
}
m_freem(mrep);
}
FREE(ndp->ni_pnbuf, M_NAMEI);
nfs_nput(ndp->ni_dvp);
return (error);
}
/*
* nfs remove directory call
*/
nfs_rmdir(ndp, p)
register struct nameidata *ndp;
struct proc *p;
{
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
if (ndp->ni_dvp == ndp->ni_vp) {
vrele(ndp->ni_dvp);
nfs_nput(ndp->ni_dvp);
return (EINVAL);
}
nfsstats.rpccnt[NFSPROC_RMDIR]++;
nfsm_reqhead(nfs_procids[NFSPROC_RMDIR], ndp->ni_cred,
NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(ndp->ni_namelen));
nfsm_fhtom(ndp->ni_dvp);
nfsm_strtom(ndp->ni_ptr, ndp->ni_namelen, NFS_MAXNAMLEN);
nfsm_request(ndp->ni_dvp, NFSPROC_RMDIR, p, 1);
nfsm_reqdone;
FREE(ndp->ni_pnbuf, M_NAMEI);
VTONFS(ndp->ni_dvp)->n_flag |= NMODIFIED;
cache_purge(ndp->ni_dvp);
cache_purge(ndp->ni_vp);
nfs_nput(ndp->ni_vp);
nfs_nput(ndp->ni_dvp);
/*
* Kludge: Map ENOENT => 0 assuming that you have a reply to a retry.
*/
if (error == ENOENT)
error = 0;
return (error);
}
/*
* nfs readdir call
* Although cookie is defined as opaque, I translate it to/from net byte
* order so that it looks more sensible. This appears consistent with the
* Ultrix implementation of NFS.
*/
nfs_readdir(vp, uiop, cred, eofflagp)
register struct vnode *vp;
struct uio *uiop;
struct ucred *cred;
int *eofflagp;
{
register struct nfsnode *np = VTONFS(vp);
int tresid, error;
struct vattr vattr;
if (vp->v_type != VDIR)
return (EPERM);
/*
* First, check for hit on the EOF offset cache
*/
if (uiop->uio_offset != 0 && uiop->uio_offset == np->n_direofoffset &&
(np->n_flag & NMODIFIED) == 0 &&
nfs_dogetattr(vp, &vattr, cred, 0, uiop->uio_procp) == 0 &&
np->n_mtime == vattr.va_mtime.tv_sec) {
*eofflagp = 1;
nfsstats.direofcache_hits++;
return (0);
}
/*
* Call nfs_bioread() to do the real work.
*/
tresid = uiop->uio_resid;
error = nfs_bioread(vp, uiop, 0, cred);
if (!error && uiop->uio_resid == tresid) {
*eofflagp = 1;
nfsstats.direofcache_misses++;
} else
*eofflagp = 0;
return (error);
}
/*
* Readdir rpc call.
* Called from below the buffer cache by nfs_doio().
*/
nfs_readdirrpc(vp, uiop, cred)
register struct vnode *vp;
struct uio *uiop;
struct ucred *cred;
{
register long len;
register struct direct *dp;
register u_long *tl;
register caddr_t cp;
register long t1;
long tlen, lastlen;
caddr_t bpos, dpos, cp2;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
struct mbuf *md2;
caddr_t dpos2;
int siz;
int more_dirs = 1;
off_t off, savoff;
struct direct *savdp;
struct nfsmount *nmp;
struct nfsnode *np = VTONFS(vp);
long tresid;
nmp = VFSTONFS(vp->v_mount);
tresid = uiop->uio_resid;
/*
* Loop around doing readdir rpc's of size uio_resid or nm_rsize,
* whichever is smaller, truncated to a multiple of NFS_DIRBLKSIZ.
* The stopping criteria is EOF or buffer full.
*/
while (more_dirs && uiop->uio_resid >= NFS_DIRBLKSIZ) {
nfsstats.rpccnt[NFSPROC_READDIR]++;
nfsm_reqhead(nfs_procids[NFSPROC_READDIR], cred, xid);
nfsm_fhtom(vp);
nfsm_build(tl, u_long *, 2*NFSX_UNSIGNED);
*tl++ = txdr_unsigned(uiop->uio_offset);
*tl = txdr_unsigned(((uiop->uio_resid > nmp->nm_rsize) ?
nmp->nm_rsize : uiop->uio_resid) & ~(NFS_DIRBLKSIZ-1));
nfsm_request(vp, NFSPROC_READDIR, uiop->uio_procp, 0);
siz = 0;
nfsm_disect(tl, u_long *, NFSX_UNSIGNED);
more_dirs = fxdr_unsigned(int, *tl);
/* Save the position so that we can do nfsm_mtouio() later */
dpos2 = dpos;
md2 = md;
/* loop thru the dir entries, doctoring them to 4bsd form */
off = uiop->uio_offset;
#ifdef lint
dp = (struct direct *)0;
#endif /* lint */
while (more_dirs && siz < uiop->uio_resid) {
savoff = off; /* Hold onto offset and dp */
savdp = dp;
nfsm_disecton(tl, u_long *, 2*NFSX_UNSIGNED);
dp = (struct direct *)tl;
dp->d_ino = fxdr_unsigned(u_long, *tl++);
len = fxdr_unsigned(int, *tl);
if (len <= 0 || len > NFS_MAXNAMLEN) {
error = EBADRPC;
m_freem(mrep);
goto nfsmout;
}
dp->d_namlen = (u_short)len;
nfsm_adv(len); /* Point past name */
tlen = nfsm_rndup(len);
/*
* This should not be necessary, but some servers have
* broken XDR such that these bytes are not null filled.
*/
if (tlen != len) {
*dpos = '\0'; /* Null-terminate */
nfsm_adv(tlen - len);
len = tlen;
}
nfsm_disecton(tl, u_long *, 2*NFSX_UNSIGNED);
off = fxdr_unsigned(off_t, *tl);
*tl++ = 0; /* Ensures null termination of name */
more_dirs = fxdr_unsigned(int, *tl);
dp->d_reclen = len+4*NFSX_UNSIGNED;
siz += dp->d_reclen;
}
/*
* If at end of rpc data, get the eof boolean
*/
if (!more_dirs) {
nfsm_disecton(tl, u_long *, NFSX_UNSIGNED);
more_dirs = (fxdr_unsigned(int, *tl) == 0);
/*
* If at EOF, cache directory offset
*/
if (!more_dirs)
np->n_direofoffset = off;
}
/*
* If there is too much to fit in the data buffer, use savoff and
* savdp to trim off the last record.
* --> we are not at eof
*/
if (siz > uiop->uio_resid) {
off = savoff;
siz -= dp->d_reclen;
dp = savdp;
more_dirs = 0; /* Paranoia */
}
if (siz > 0) {
lastlen = dp->d_reclen;
md = md2;
dpos = dpos2;
nfsm_mtouio(uiop, siz);
uiop->uio_offset = off;
} else
more_dirs = 0; /* Ugh, never happens, but in case.. */
m_freem(mrep);
}
/*
* Fill last record, iff any, out to a multiple of NFS_DIRBLKSIZ
* by increasing d_reclen for the last record.
*/
if (uiop->uio_resid < tresid) {
len = uiop->uio_resid & (NFS_DIRBLKSIZ - 1);
if (len > 0) {
dp = (struct direct *)
(uiop->uio_iov->iov_base - lastlen);
dp->d_reclen += len;
uiop->uio_iov->iov_base += len;
uiop->uio_iov->iov_len -= len;
uiop->uio_resid -= len;
}
}
nfsmout:
return (error);
}
static char hextoasc[] = "0123456789abcdef";
/*
* Silly rename. To make the NFS filesystem that is stateless look a little
* more like the "ufs" a remove of an active vnode is translated to a rename
* to a funny looking filename that is removed by nfs_inactive on the
* nfsnode. There is the potential for another process on a different client
* to create the same funny name between the nfs_lookitup() fails and the
* nfs_rename() completes, but...
*/
nfs_sillyrename(ndp, p)
register struct nameidata *ndp;
struct proc *p;
{
register struct nfsnode *np;
register struct sillyrename *sp;
int error;
short pid;
np = VTONFS(ndp->ni_dvp);
cache_purge(ndp->ni_dvp);
MALLOC(sp, struct sillyrename *, sizeof (struct sillyrename),
M_NFSREQ, M_WAITOK);
bcopy((caddr_t)&np->n_fh, (caddr_t)&sp->s_fh, NFSX_FH);
np = VTONFS(ndp->ni_vp);
sp->s_cred = crdup(ndp->ni_cred);
sp->s_dvp = ndp->ni_dvp;
VREF(sp->s_dvp);
/* Fudge together a funny name */
pid = p->p_pid;
bcopy(".nfsAxxxx4.4", sp->s_name, 13);
sp->s_namlen = 12;
sp->s_name[8] = hextoasc[pid & 0xf];
sp->s_name[7] = hextoasc[(pid >> 4) & 0xf];
sp->s_name[6] = hextoasc[(pid >> 8) & 0xf];
sp->s_name[5] = hextoasc[(pid >> 12) & 0xf];
/* Try lookitups until we get one that isn't there */
while (nfs_lookitup(sp, (nfsv2fh_t *)0, p) == 0) {
sp->s_name[4]++;
if (sp->s_name[4] > 'z') {
error = EINVAL;
goto bad;
}
}
if (error = nfs_renameit(ndp, sp, p))
goto bad;
nfs_lookitup(sp, &np->n_fh, p);
np->n_sillyrename = sp;
return (0);
bad:
vrele(sp->s_dvp);
crfree(sp->s_cred);
free((caddr_t)sp, M_NFSREQ);
return (error);
}
/*
* Look up a file name for silly rename stuff.
* Just like nfs_lookup() except that it doesn't load returned values
* into the nfsnode table.
* If fhp != NULL it copies the returned file handle out
*/
nfs_lookitup(sp, fhp, p)
register struct sillyrename *sp;
nfsv2fh_t *fhp;
struct proc *p;
{
register struct vnode *vp = sp->s_dvp;
register u_long *tl;
register caddr_t cp;
register long t1, t2;
caddr_t bpos, dpos, cp2;
u_long xid;
int error = 0;
struct mbuf *mreq, *mrep, *md, *mb, *mb2;
long len;
nfsstats.rpccnt[NFSPROC_LOOKUP]++;
len = sp->s_namlen;
nfsm_reqhead(nfs_procids[NFSPROC_LOOKUP], sp->s_cred, NFSX_FH+NFSX_UNSIGNED+nfsm_rndup(len));
nfsm_fhtom(vp);
nfsm_strtom(sp->s_name, len, NFS_MAXNAMLEN);
nfsm_request(vp, NFSPROC_LOOKUP, p, 1);
if (fhp != NULL) {
nfsm_disect(cp, caddr_t, NFSX_FH);
bcopy(cp, (caddr_t)fhp, NFSX_FH);
}
nfsm_reqdone;
return (error);
}
/*
* Kludge City..
* - make nfs_bmap() essentially a no-op that does no translation
* - do nfs_strategy() by faking physical I/O with nfs_readrpc/nfs_writerpc
* after mapping the physical addresses into Kernel Virtual space in the
* nfsiobuf area.
* (Maybe I could use the process's page mapping, but I was concerned that
* Kernel Write might not be enabled and also figured copyout() would do
* a lot more work than bcopy() and also it currently happens in the
* context of the swapper process (2).
*/
nfs_bmap(vp, bn, vpp, bnp)
struct vnode *vp;
daddr_t bn;
struct vnode **vpp;
daddr_t *bnp;
{
if (vpp != NULL)
*vpp = vp;
if (bnp != NULL)
*bnp = bn * btodb(vp->v_mount->mnt_stat.f_bsize);
return (0);
}
/*
* Strategy routine for phys. i/o
* If the biod's are running, queue a request
* otherwise just call nfs_doio() to get it done
*/
nfs_strategy(bp)
register struct buf *bp;
{
register struct buf *dp;
register int i;
int error = 0;
int fnd = 0;
/*
* Set b_proc. It seems a bit silly to do it here, but since bread()
* doesn't set it, I will.
* Set b_proc == NULL for asynchronous ops, since these may still
* be hanging about after the process terminates.
*/
if ((bp->b_flags & B_PHYS) == 0) {
if (bp->b_flags & B_ASYNC)
bp->b_proc = (struct proc *)0;
else
bp->b_proc = curproc;
}
/*
* If the op is asynchronous and an i/o daemon is waiting
* queue the request, wake it up and wait for completion
* otherwise just do it ourselves.
*/
if ((bp->b_flags & B_ASYNC) == 0 || nfs_numasync == 0)
return (nfs_doio(bp));
for (i = 0; i < NFS_MAXASYNCDAEMON; i++) {
if (nfs_iodwant[i]) {
dp = &nfs_bqueue;
if (dp->b_actf == NULL) {
dp->b_actl = bp;
bp->b_actf = dp;
} else {
dp->b_actf->b_actl = bp;
bp->b_actf = dp->b_actf;
}
dp->b_actf = bp;
bp->b_actl = dp;
fnd++;
wakeup((caddr_t)&nfs_iodwant[i]);
break;
}
}
if (!fnd)
error = nfs_doio(bp);
return (error);
}
/*
* Fun and games with i/o
* Essentially play ubasetup() and disk interrupt service routine by
* mapping the data buffer into kernel virtual space and doing the
* nfs read or write rpc's from it.
* If the nfsiod's are not running, this is just called from nfs_strategy(),
* otherwise it is called by the nfsiods to do what would normally be
* partially disk interrupt driven.
*/
nfs_doio(bp)
register struct buf *bp;
{
register struct uio *uiop;
register struct vnode *vp;
struct nfsnode *np;
struct ucred *cr;
int error;
struct uio uio;
struct iovec io;
#if !defined(hp300) && !defined(i386)
register struct pte *pte, *ppte;
register caddr_t vaddr;
int npf, npf2;
int reg, o;
caddr_t vbase;
unsigned v;
#endif
vp = bp->b_vp;
np = VTONFS(vp);
uiop = &uio;
uiop->uio_iov = &io;
uiop->uio_iovcnt = 1;
uiop->uio_segflg = UIO_SYSSPACE;
uiop->uio_procp = (struct proc *)0;
/*
* For phys i/o, map the b_addr into kernel virtual space using
* the Nfsiomap pte's
* Also, add a temporary b_rcred for reading using the process's uid
* and a guess at a group
*/
if (bp->b_flags & B_PHYS) {
if (bp->b_flags & B_DIRTY)
uiop->uio_procp = pageproc;
cr = crcopy(uiop->uio_procp->p_ucred);
/* mapping was already done by vmapbuf */
io.iov_base = bp->b_un.b_addr;
/*
* And do the i/o rpc
*/
io.iov_len = uiop->uio_resid = bp->b_bcount;
uiop->uio_offset = bp->b_blkno * DEV_BSIZE;
if (bp->b_flags & B_READ) {
uiop->uio_rw = UIO_READ;
nfsstats.read_physios++;
bp->b_error = error = nfs_readrpc(vp, uiop, cr);
(void) vnode_pager_uncache(vp);
} else {
uiop->uio_rw = UIO_WRITE;
nfsstats.write_physios++;
bp->b_error = error = nfs_writerpc(vp, uiop, cr);
}
/*
* Finally, release pte's used by physical i/o
*/
crfree(cr);
} else {
if (bp->b_flags & B_READ) {
io.iov_len = uiop->uio_resid = bp->b_bcount;
io.iov_base = bp->b_un.b_addr;
uiop->uio_rw = UIO_READ;
switch (vp->v_type) {
case VREG:
uiop->uio_offset = bp->b_blkno * DEV_BSIZE;
nfsstats.read_bios++;
error = nfs_readrpc(vp, uiop, bp->b_rcred);
break;
case VLNK:
uiop->uio_offset = 0;
nfsstats.readlink_bios++;
error = nfs_readlinkrpc(vp, uiop, bp->b_rcred);
break;
case VDIR:
uiop->uio_offset = bp->b_lblkno;
nfsstats.readdir_bios++;
error = nfs_readdirrpc(vp, uiop, bp->b_rcred);
/*
* Save offset cookie in b_blkno.
*/
bp->b_blkno = uiop->uio_offset;
break;
};
bp->b_error = error;
} else {
io.iov_len = uiop->uio_resid = bp->b_dirtyend
- bp->b_dirtyoff;
uiop->uio_offset = (bp->b_blkno * DEV_BSIZE)
+ bp->b_dirtyoff;
io.iov_base = bp->b_un.b_addr + bp->b_dirtyoff;
uiop->uio_rw = UIO_WRITE;
nfsstats.write_bios++;
bp->b_error = error = nfs_writerpc(vp, uiop,
bp->b_wcred);
if (error) {
np->n_error = error;
np->n_flag |= NWRITEERR;
}
bp->b_dirtyoff = bp->b_dirtyend = 0;
}
}
if (error)
bp->b_flags |= B_ERROR;
bp->b_resid = uiop->uio_resid;
biodone(bp);
return (error);
}
/*
* Mmap a file
*
* NB Currently unsupported.
*/
/* ARGSUSED */
nfs_mmap(vp, fflags, cred, p)
struct vnode *vp;
int fflags;
struct ucred *cred;
struct proc *p;
{
return (EINVAL);
}
/*
* Flush all the blocks associated with a vnode.
* Walk through the buffer pool and push any dirty pages
* associated with the vnode.
*/
/* ARGSUSED */
nfs_fsync(vp, fflags, cred, waitfor, p)
register struct vnode *vp;
int fflags;
struct ucred *cred;
int waitfor;
struct proc *p;
{
register struct nfsnode *np = VTONFS(vp);
int error = 0;
if (np->n_flag & NMODIFIED) {
np->n_flag &= ~NMODIFIED;
vflushbuf(vp, waitfor == MNT_WAIT ? B_SYNC : 0);
}
if (!error && (np->n_flag & NWRITEERR))
error = np->n_error;
return (error);
}
/*
* NFS advisory byte-level locks.
* Currently unsupported.
*/
nfs_advlock(vp, id, op, fl, flags)
struct vnode *vp;
caddr_t id;
int op;
struct flock *fl;
int flags;
{
register struct nfsnode *np = VTONFS(vp);
return (lf_advlock(&(np->n_lockf), np->n_size, id, op, fl, flags));
}
/*
* Print out the contents of an nfsnode.
*/
nfs_print(vp)
struct vnode *vp;
{
register struct nfsnode *np = VTONFS(vp);
printf("tag VT_NFS, fileid %d fsid 0x%x",
np->n_vattr.va_fileid, np->n_vattr.va_fsid);
#ifdef FIFO
if (vp->v_type == VFIFO)
fifo_printinfo(vp);
#endif /* FIFO */
printf("%s\n", (np->n_flag & NLOCKED) ? " (LOCKED)" : "");
if (np->n_lockholder == 0)
return;
printf("\towner pid %d", np->n_lockholder);
if (np->n_lockwaiter)
printf(" waiting pid %d", np->n_lockwaiter);
printf("\n");
}
/*
* Attribute cache routines.
* nfs_loadattrcache() - loads or updates the cache contents from attributes
* that are on the mbuf list
* nfs_getattrcache() - returns valid attributes if found in cache, returns
* error otherwise
*/
/*
* Load the attribute cache (that lives in the nfsnode entry) with
* the values on the mbuf list and
* Iff vap not NULL
* copy the attributes to *vaper
*/
nfs_loadattrcache(vpp, mdp, dposp, vaper)
struct vnode **vpp;
struct mbuf **mdp;
caddr_t *dposp;
struct vattr *vaper;
{
register struct vnode *vp = *vpp;
register struct vattr *vap;
register struct nfsv2_fattr *fp;
extern struct vnodeops spec_nfsv2nodeops;
register struct nfsnode *np;
register long t1;
caddr_t dpos, cp2;
int error = 0;
struct mbuf *md;
enum vtype type;
u_short mode;
long rdev;
struct timeval mtime;
struct vnode *nvp;
md = *mdp;
dpos = *dposp;
t1 = (mtod(md, caddr_t)+md->m_len)-dpos;
if (error = nfsm_disct(&md, &dpos, NFSX_FATTR, t1, TRUE, &cp2))
return (error);
fp = (struct nfsv2_fattr *)cp2;
type = nfstov_type(fp->fa_type);
mode = fxdr_unsigned(u_short, fp->fa_mode);
if (type == VNON)
type = IFTOVT(mode);
rdev = fxdr_unsigned(long, fp->fa_rdev);
fxdr_time(&fp->fa_mtime, &mtime);
/*
* If v_type == VNON it is a new node, so fill in the v_type,
* n_mtime fields. Check to see if it represents a special
* device, and if so, check for a possible alias. Once the
* correct vnode has been obtained, fill in the rest of the
* information.
*/
np = VTONFS(vp);
if (vp->v_type == VNON) {
if (type == VCHR && rdev == 0xffffffff)
vp->v_type = type = VFIFO;
else
vp->v_type = type;
if (vp->v_type == VFIFO) {
#ifdef FIFO
extern struct vnodeops fifo_nfsv2nodeops;
vp->v_op = &fifo_nfsv2nodeops;
#else
return (EOPNOTSUPP);
#endif /* FIFO */
}
if (vp->v_type == VCHR || vp->v_type == VBLK) {
vp->v_op = &spec_nfsv2nodeops;
if (nvp = checkalias(vp, (dev_t)rdev, vp->v_mount)) {
/*
* Reinitialize aliased node.
*/
np = VTONFS(nvp);
np->n_vnode = nvp;
np->n_flag = 0;
nfs_lock(nvp);
bcopy((caddr_t)&VTONFS(vp)->n_fh,
(caddr_t)&np->n_fh, NFSX_FH);
insque(np, nfs_hash(&np->n_fh));
np->n_attrstamp = 0;
np->n_sillyrename = (struct sillyrename *)0;
/*
* Discard unneeded vnode and update actual one
*/
vput(vp);
*vpp = nvp;
}
}
np->n_mtime = mtime.tv_sec;
}
vap = &np->n_vattr;
vap->va_type = type;
vap->va_mode = (mode & 07777);
vap->va_nlink = fxdr_unsigned(u_short, fp->fa_nlink);
vap->va_uid = fxdr_unsigned(uid_t, fp->fa_uid);
vap->va_gid = fxdr_unsigned(gid_t, fp->fa_gid);
vap->va_size = fxdr_unsigned(u_long, fp->fa_size);
if ((np->n_flag & NMODIFIED) == 0 || vap->va_size > np->n_size) {
np->n_size = vap->va_size;
vnode_pager_setsize(vp, np->n_size);
}
vap->va_size_rsv = 0;
vap->va_blocksize = fxdr_unsigned(long, fp->fa_blocksize);
vap->va_rdev = (dev_t)rdev;
vap->va_bytes = fxdr_unsigned(long, fp->fa_blocks) * NFS_FABLKSIZE;
vap->va_bytes_rsv = 0;
vap->va_fsid = vp->v_mount->mnt_stat.f_fsid.val[0];
vap->va_fileid = fxdr_unsigned(long, fp->fa_fileid);
/* jfw@ksr.com 6/2/93 */
#if 0 /* bad assumption; Suns make full (and obvious) use of .usec fields */
vap->va_atime.tv_sec = fxdr_unsigned(long, fp->fa_atime.tv_sec);
vap->va_atime.tv_usec = 0;
vap->va_flags = fxdr_unsigned(u_long, fp->fa_atime.tv_usec);
vap->va_mtime = mtime;
vap->va_ctime.tv_sec = fxdr_unsigned(long, fp->fa_ctime.tv_sec);
vap->va_ctime.tv_usec = 0;
vap->va_gen = fxdr_unsigned(u_long, fp->fa_ctime.tv_usec);
#else
fxdr_time(&fp->fa_atime, &vap->va_atime);
vap->va_mtime = mtime;
fxdr_time(&fp->fa_ctime, &vap->va_ctime);
vap->va_gen = 0; /* can reliably learn nothing about this via NFS. */
vap->va_flags = 0; /* can reliably learn nothing about this via NFS. */
#endif
np->n_attrstamp = time.tv_sec;
*dposp = dpos;
*mdp = md;
if (vaper != NULL) {
bcopy((caddr_t)vap, (caddr_t)vaper, sizeof(*vap));
if ((np->n_flag & NMODIFIED) && (np->n_size > vap->va_size))
vaper->va_size = np->n_size;
}
return (0);
}
/*
* Check the time stamp
* If the cache is valid, copy contents to *vap and return 0
* otherwise return an error
*/
nfs_getattrcache(vp, vap)
register struct vnode *vp;
struct vattr *vap;
{
register struct nfsnode *np;
np = VTONFS(vp);
if ((time.tv_sec-np->n_attrstamp) < NFS_ATTRTIMEO) {
nfsstats.attrcache_hits++;
bcopy((caddr_t)&np->n_vattr,(caddr_t)vap,sizeof(struct vattr));
if ((np->n_flag & NMODIFIED) == 0) {
np->n_size = vap->va_size;
vnode_pager_setsize(vp, np->n_size);
} else if (np->n_size > vap->va_size)
vap->va_size = np->n_size;
return (0);
} else {
nfsstats.attrcache_misses++;
return (ENOENT);
}
}