summaryrefslogtreecommitdiffstats
path: root/sys/fs/ntfs
diff options
context:
space:
mode:
Diffstat (limited to 'sys/fs/ntfs')
-rw-r--r--sys/fs/ntfs/ntfs_subr.c2
-rw-r--r--sys/fs/ntfs/ntfs_vfsops.c84
-rw-r--r--sys/fs/ntfs/ntfs_vfsops.h4
-rw-r--r--sys/fs/ntfs/ntfs_vnops.c22
4 files changed, 56 insertions, 56 deletions
diff --git a/sys/fs/ntfs/ntfs_subr.c b/sys/fs/ntfs/ntfs_subr.c
index e1e6006..c215741 100644
--- a/sys/fs/ntfs/ntfs_subr.c
+++ b/sys/fs/ntfs/ntfs_subr.c
@@ -227,7 +227,7 @@ ntfs_ntvattrget(
vget() */
error = ntfs_vgetex(ntmp->ntm_mountp, aalp->al_inumber,
NTFS_A_DATA, NULL, LK_EXCLUSIVE,
- VG_EXT, curproc, &newvp);
+ VG_EXT, curthread, &newvp);
if (error) {
printf("ntfs_ntvattrget: CAN'T VGET INO: %d\n",
aalp->al_inumber);
diff --git a/sys/fs/ntfs/ntfs_vfsops.c b/sys/fs/ntfs/ntfs_vfsops.c
index 26a09b8..19e6925 100644
--- a/sys/fs/ntfs/ntfs_vfsops.c
+++ b/sys/fs/ntfs/ntfs_vfsops.c
@@ -76,36 +76,36 @@ MALLOC_DEFINE(M_NTFSDIR,"NTFS dir", "NTFS dir buffer");
static int ntfs_root __P((struct mount *, struct vnode **));
static int ntfs_statfs __P((struct mount *, struct statfs *,
- struct proc *));
-static int ntfs_unmount __P((struct mount *, int, struct proc *));
+ struct thread *));
+static int ntfs_unmount __P((struct mount *, int, struct thread *));
static int ntfs_vget __P((struct mount *mp, ino_t ino,
struct vnode **vpp));
static int ntfs_mountfs __P((register struct vnode *, struct mount *,
- struct ntfs_args *, struct proc *));
+ struct ntfs_args *, struct thread *));
static int ntfs_vptofh __P((struct vnode *, struct fid *));
static int ntfs_fhtovp __P((struct mount *, struct fid *,
struct vnode **));
#if !defined (__FreeBSD__)
static int ntfs_quotactl __P((struct mount *, int, uid_t, caddr_t,
- struct proc *));
-static int ntfs_start __P((struct mount *, int, struct proc *));
+ struct thread *));
+static int ntfs_start __P((struct mount *, int, struct thread *));
static int ntfs_sync __P((struct mount *, int, struct ucred *,
- struct proc *));
+ struct thread *));
#endif
#if defined(__FreeBSD__)
struct sockaddr;
static int ntfs_mount __P((struct mount *, char *, caddr_t,
- struct nameidata *, struct proc *));
+ struct nameidata *, struct thread *));
static int ntfs_init __P((struct vfsconf *));
#elif defined(__NetBSD__)
static int ntfs_mount __P((struct mount *, const char *, void *,
- struct nameidata *, struct proc *));
+ struct nameidata *, struct thread *));
static void ntfs_init __P((void));
static int ntfs_mountroot __P((void));
static int ntfs_sysctl __P((int *, u_int, void *, size_t *, void *,
- size_t, struct proc *));
+ size_t, struct thread *));
static int ntfs_checkexp __P((struct mount *, struct mbuf *,
int *, struct ucred **));
#endif
@@ -180,7 +180,7 @@ ntfs_mountroot()
args.gid = 0;
args.mode = 0777;
- if ((error = ntfs_mountfs(rootvp, mp, &args, p)) != 0) {
+ if ((error = ntfs_mountfs(rootvp, mp, &args, td)) != 0) {
mp->mnt_op->vfs_refcount--;
vfs_unbusy(mp);
free(mp, M_MOUNT);
@@ -191,7 +191,7 @@ ntfs_mountroot()
mtx_lock(&mountlist_mtx);
TAILQ_INSERT_TAIL(&mountlist, mp, mnt_list);
mtx_unlock(&mountlist_mtx);
- (void)ntfs_statfs(mp, &mp->mnt_stat, p);
+ (void)ntfs_statfs(mp, &mp->mnt_stat, td);
vfs_unbusy(mp);
return (0);
}
@@ -236,7 +236,7 @@ ntfs_mount (
void *data,
#endif
struct nameidata *ndp,
- struct proc *p )
+ struct thread *td )
{
size_t size;
int err = 0;
@@ -266,7 +266,7 @@ ntfs_mount (
/*
* Attempt mount
*/
- if( ( err = ntfs_mountfs(rootvp, mp, &args, p)) != 0) {
+ if( ( err = ntfs_mountfs(rootvp, mp, &args, td)) != 0) {
/* fs specific cleanup (if any)*/
goto error_1;
}
@@ -316,7 +316,7 @@ ntfs_mount (
* Not an update, or updating the name: look up the name
* and verify that it refers to a sensible block device.
*/
- NDINIT(ndp, LOOKUP, FOLLOW, UIO_USERSPACE, args.fspec, p);
+ NDINIT(ndp, LOOKUP, FOLLOW, UIO_USERSPACE, args.fspec, td);
err = namei(ndp);
if (err) {
/* can't get devvp!*/
@@ -385,7 +385,7 @@ ntfs_mount (
&size); /* real size*/
bzero( mp->mnt_stat.f_mntfromname + size, MNAMELEN - size);
- err = ntfs_mountfs(devvp, mp, &args, p);
+ err = ntfs_mountfs(devvp, mp, &args, td);
}
if (err) {
goto error_2;
@@ -400,7 +400,7 @@ dostatfs:
*
* This code is common to root and non-root mounts
*/
- (void)VFS_STATFS(mp, &mp->mnt_stat, p);
+ (void)VFS_STATFS(mp, &mp->mnt_stat, td);
goto success;
@@ -420,11 +420,11 @@ success:
* Common code for mount and mountroot
*/
int
-ntfs_mountfs(devvp, mp, argsp, p)
+ntfs_mountfs(devvp, mp, argsp, td)
register struct vnode *devvp;
struct mount *mp;
struct ntfs_args *argsp;
- struct proc *p;
+ struct thread *td;
{
struct buf *bp;
struct ntfsmount *ntmp;
@@ -449,19 +449,19 @@ ntfs_mountfs(devvp, mp, argsp, p)
if (ncount > 1 && devvp != rootvp)
return (EBUSY);
#if defined(__FreeBSD__)
- VN_LOCK(devvp, LK_EXCLUSIVE | LK_RETRY, p);
- error = vinvalbuf(devvp, V_SAVE, p->p_ucred, p, 0, 0);
- VOP__UNLOCK(devvp, 0, p);
+ VN_LOCK(devvp, LK_EXCLUSIVE | LK_RETRY, td);
+ error = vinvalbuf(devvp, V_SAVE, td->td_proc->p_ucred, td, 0, 0);
+ VOP__UNLOCK(devvp, 0, td);
#else
- error = vinvalbuf(devvp, V_SAVE, p->p_ucred, p, 0, 0);
+ error = vinvalbuf(devvp, V_SAVE, td->td_proc->p_ucred, td, 0, 0);
#endif
if (error)
return (error);
ronly = (mp->mnt_flag & MNT_RDONLY) != 0;
- VN_LOCK(devvp, LK_EXCLUSIVE | LK_RETRY, p);
- error = VOP_OPEN(devvp, ronly ? FREAD : FREAD|FWRITE, FSCRED, p);
- VOP__UNLOCK(devvp, 0, p);
+ VN_LOCK(devvp, LK_EXCLUSIVE | LK_RETRY, td);
+ error = VOP_OPEN(devvp, ronly ? FREAD : FREAD|FWRITE, FSCRED, td);
+ VOP__UNLOCK(devvp, 0, td);
if (error)
return (error);
@@ -624,11 +624,11 @@ out:
#if defined __NetBSD__
/* lock the device vnode before calling VOP_CLOSE() */
- VN_LOCK(devvp, LK_EXCLUSIVE | LK_RETRY, p);
- (void)VOP_CLOSE(devvp, ronly ? FREAD : FREAD|FWRITE, NOCRED, p);
- VOP__UNLOCK(devvp, 0, p);
+ VN_LOCK(devvp, LK_EXCLUSIVE | LK_RETRY, td);
+ (void)VOP_CLOSE(devvp, ronly ? FREAD : FREAD|FWRITE, NOCRED, td);
+ VOP__UNLOCK(devvp, 0, td);
#else
- (void)VOP_CLOSE(devvp, ronly ? FREAD : FREAD|FWRITE, NOCRED, p);
+ (void)VOP_CLOSE(devvp, ronly ? FREAD : FREAD|FWRITE, NOCRED, td);
#endif
return (error);
@@ -639,7 +639,7 @@ static int
ntfs_start (
struct mount *mp,
int flags,
- struct proc *p )
+ struct thread *td )
{
return (0);
}
@@ -649,7 +649,7 @@ static int
ntfs_unmount(
struct mount *mp,
int mntflags,
- struct proc *p)
+ struct thread *td)
{
register struct ntfsmount *ntmp;
int error, ronly = 0, flags, i;
@@ -689,17 +689,17 @@ ntfs_unmount(
if (ntmp->ntm_devvp->v_type != VBAD)
ntmp->ntm_devvp->v_rdev->si_mountpoint = NULL;
- vinvalbuf(ntmp->ntm_devvp, V_SAVE, NOCRED, p, 0, 0);
+ vinvalbuf(ntmp->ntm_devvp, V_SAVE, NOCRED, td, 0, 0);
#if defined(__NetBSD__)
/* lock the device vnode before calling VOP_CLOSE() */
VOP_LOCK(ntmp->ntm_devvp, LK_EXCLUSIVE | LK_RETRY);
error = VOP_CLOSE(ntmp->ntm_devvp, ronly ? FREAD : FREAD|FWRITE,
- NOCRED, p);
- VOP__UNLOCK(ntmp->ntm_devvp, 0, p);
+ NOCRED, td);
+ VOP__UNLOCK(ntmp->ntm_devvp, 0, td);
#else
error = VOP_CLOSE(ntmp->ntm_devvp, ronly ? FREAD : FREAD|FWRITE,
- NOCRED, p);
+ NOCRED, td);
#endif
vrele(ntmp->ntm_devvp);
@@ -744,7 +744,7 @@ ntfs_quotactl (
int cmds,
uid_t uid,
caddr_t arg,
- struct proc *p)
+ struct thread *td)
{
printf("\nntfs_quotactl():\n");
return EOPNOTSUPP;
@@ -787,7 +787,7 @@ static int
ntfs_statfs(
struct mount *mp,
struct statfs *sbp,
- struct proc *p)
+ struct thread *td)
{
struct ntfsmount *ntmp = VFSTONTFS(mp);
u_int64_t mftsize,mftallocated;
@@ -831,7 +831,7 @@ ntfs_sync (
struct mount *mp,
int waitfor,
struct ucred *cred,
- struct proc *p)
+ struct thread *td)
{
/*dprintf(("ntfs_sync():\n"));*/
return (0);
@@ -888,7 +888,7 @@ ntfs_vgetex(
char *attrname,
u_long lkflags,
u_long flags,
- struct proc *p,
+ struct thread *td,
struct vnode **vpp)
{
int error;
@@ -953,7 +953,7 @@ ntfs_vgetex(
}
if (FTOV(fp)) {
- VGET(FTOV(fp), lkflags, p);
+ VGET(FTOV(fp), lkflags, td);
*vpp = FTOV(fp);
ntfs_ntput(ip);
return (0);
@@ -980,7 +980,7 @@ ntfs_vgetex(
ntfs_ntput(ip);
if (lkflags & LK_TYPE_MASK) {
- error = VN_LOCK(vp, lkflags, p);
+ error = VN_LOCK(vp, lkflags, td);
if (error) {
vput(vp);
return (error);
@@ -999,7 +999,7 @@ ntfs_vget(
struct vnode **vpp)
{
return ntfs_vgetex(mp, ino, NTFS_A_DATA, NULL,
- LK_EXCLUSIVE | LK_RETRY, 0, curproc, vpp);
+ LK_EXCLUSIVE | LK_RETRY, 0, curthread, vpp);
}
#if defined(__FreeBSD__)
diff --git a/sys/fs/ntfs/ntfs_vfsops.h b/sys/fs/ntfs/ntfs_vfsops.h
index c05686f..b82877a 100644
--- a/sys/fs/ntfs/ntfs_vfsops.h
+++ b/sys/fs/ntfs/ntfs_vfsops.h
@@ -36,9 +36,9 @@
struct mount;
struct ntfsmount;
-struct proc;
+struct thread;
struct vnode;
int ntfs_vgetex(struct mount *, ino_t, u_int32_t, char *, u_long, u_long,
- struct proc *, struct vnode **);
+ struct thread *, struct vnode **);
int ntfs_calccfree(struct ntfsmount *, cn_t *);
diff --git a/sys/fs/ntfs/ntfs_vnops.c b/sys/fs/ntfs/ntfs_vnops.c
index 5c2297f..baad440 100644
--- a/sys/fs/ntfs/ntfs_vnops.c
+++ b/sys/fs/ntfs/ntfs_vnops.c
@@ -191,7 +191,7 @@ ntfs_getattr(ap)
struct vnode *a_vp;
struct vattr *a_vap;
struct ucred *a_cred;
- struct proc *a_p;
+ struct thread *a_td;
} */ *ap;
{
register struct vnode *vp = ap->a_vp;
@@ -245,7 +245,7 @@ ntfs_inactive(ap)
if (ntfs_prtactive && vp->v_usecount != 0)
vprint("ntfs_inactive: pushing active", vp);
- VOP__UNLOCK(vp, 0, ap->a_p);
+ VOP__UNLOCK(vp, 0, ap->a_td);
/* XXX since we don't support any filesystem changes
* right now, nothing more needs to be done
@@ -423,7 +423,7 @@ ntfs_access(ap)
struct vnode *a_vp;
int a_mode;
struct ucred *a_cred;
- struct proc *a_p;
+ struct thread *a_td;
} */ *ap;
{
struct vnode *vp = ap->a_vp;
@@ -471,7 +471,7 @@ ntfs_open(ap)
struct vnode *a_vp;
int a_mode;
struct ucred *a_cred;
- struct proc *a_p;
+ struct thread *a_td;
} */ *ap;
{
#if NTFS_DEBUG
@@ -500,7 +500,7 @@ ntfs_close(ap)
struct vnode *a_vp;
int a_fflag;
struct ucred *a_cred;
- struct proc *a_p;
+ struct thread *a_td;
} */ *ap;
{
#if NTFS_DEBUG
@@ -674,7 +674,7 @@ ntfs_lookup(ap)
(int)cnp->cn_namelen, cnp->cn_nameptr, cnp->cn_namelen,
dip->i_number, lockparent, wantparent));
- error = VOP_ACCESS(dvp, VEXEC, cred, cnp->cn_proc);
+ error = VOP_ACCESS(dvp, VEXEC, cred, cnp->cn_thread);
if(error)
return (error);
@@ -713,7 +713,7 @@ ntfs_lookup(ap)
if(error)
return (error);
- VOP__UNLOCK(dvp,0,cnp->cn_proc);
+ VOP__UNLOCK(dvp,0,cnp->cn_thread);
cnp->cn_flags |= PDIRUNLOCK;
dprintf(("ntfs_lookup: parentdir: %d\n",
@@ -722,13 +722,13 @@ ntfs_lookup(ap)
vap->va_a_name->n_pnumber,ap->a_vpp);
ntfs_ntvattrrele(vap);
if (error) {
- if (VN_LOCK(dvp,LK_EXCLUSIVE|LK_RETRY,cnp->cn_proc)==0)
+ if (VN_LOCK(dvp,LK_EXCLUSIVE|LK_RETRY,cnp->cn_thread)==0)
cnp->cn_flags &= ~PDIRUNLOCK;
return (error);
}
if (lockparent && (cnp->cn_flags & ISLASTCN)) {
- error = VN_LOCK(dvp, LK_EXCLUSIVE, cnp->cn_proc);
+ error = VN_LOCK(dvp, LK_EXCLUSIVE, cnp->cn_thread);
if (error) {
vput( *(ap->a_vpp) );
return (error);
@@ -746,7 +746,7 @@ ntfs_lookup(ap)
VTONT(*ap->a_vpp)->i_number));
if(!lockparent || !(cnp->cn_flags & ISLASTCN))
- VOP__UNLOCK(dvp, 0, cnp->cn_proc);
+ VOP__UNLOCK(dvp, 0, cnp->cn_thread);
}
if (cnp->cn_flags & MAKEENTRY)
@@ -768,7 +768,7 @@ ntfs_fsync(ap)
struct vnode *a_vp;
struct ucred *a_cred;
int a_waitfor;
- struct proc *a_p;
+ struct thread *a_td;
} */ *ap;
{
return (0);
OpenPOWER on IntegriCloud