int ext2_setattr(inode_t *inode, file_attributes_t *attr) {
struct stat s;
if (getattr_inode((ext2_fs_instance_t*)inode->i_instance, inode->i_ino, &s) == 0) {
if (attr->mask & ATTR_UID) {
s.st_uid = attr->stbuf.st_uid;
}
if (attr->mask & ATTR_GID) {
s.st_gid = attr->stbuf.st_gid;
}
if (attr->mask & ATTR_MODE) {
s.st_mode = attr->stbuf.st_mode;
}
if (attr->mask & ATTR_ATIME) {
s.st_atime = attr->stbuf.st_atime;
}
if (attr->mask & ATTR_MTIME) {
s.st_mtime = attr->stbuf.st_mtime;
}
if (attr->mask & ATTR_CTIME) {
s.st_ctime = attr->stbuf.st_ctime;
}
if (attr->mask & ATTR_SIZE && attr->ia_size != inode->i_size) {
ext2_truncate(inode, attr->ia_size);
}
setattr_inode((ext2_fs_instance_t*)inode->i_instance, inode->i_ino, &s);
return 0;
}
return -1;
}
/*
* discard preallocated blocks
*/
int
ext2_inactive(struct vop_inactive_args *ap)
{
struct vnode *vp = ap->a_vp;
struct inode *ip = VTOI(vp);
struct thread *td = ap->a_td;
int mode, error = 0;
/*
* Ignore inodes related to stale file handles.
*/
if (ip->i_mode == 0)
goto out;
if (ip->i_nlink <= 0) {
error = ext2_truncate(vp, (off_t)0, 0, NOCRED, td);
ip->i_rdev = 0;
mode = ip->i_mode;
ip->i_mode = 0;
ip->i_flag |= IN_CHANGE | IN_UPDATE;
ext2_vfree(vp, ip->i_number, mode);
}
if (ip->i_flag & (IN_ACCESS | IN_CHANGE | IN_MODIFIED | IN_UPDATE))
ext2_update(vp, 0);
out:
/*
* If we are done with the inode, reclaim it
* so that it can be reused immediately.
*/
if (ip->i_mode == 0)
vrecycle(vp);
return (error);
}
dentry_t* ext2_lookup(struct _fs_instance_t *instance, struct _dentry_t* dentry, const char * name) {
int flags = 0; // XXX
int inode = getinode_from_name((ext2_fs_instance_t*)instance, dentry->d_inode->i_ino, name);
if (inode == -ENOENT && (flags & O_CREAT)) {
//inode = ext2_mknod2((ext2_fs_instance_t*)instance, name, 00644 | 0x8000, 0); //FIXME!
} else if (inode <= 0) {
return NULL;
} else if (flags & O_EXCL && flags & O_CREAT) {
return NULL;
}
struct ext2_inode *einode = read_inode((ext2_fs_instance_t*)instance, inode);
dentry_t *d =(dentry_t*) kmalloc(sizeof(dentry_t),0);
char *n = strdup(name);
d->d_name = (const char*)n;
d->d_pdentry = dentry;
d->d_inode = (inode_t*)kmalloc(sizeof(inode_t),0);
ext2inode_2_inode(d->d_inode, instance, inode, einode);
d->d_inode->i_count = 0;
if (flags & O_TRUNC) {
ext2_truncate(d->d_inode, 0);
}
return d;
}
dentry_t* ext2_lookup(struct _fs_instance_t *instance, struct _dentry_t* dentry, const char * name) {
int flags = 0; // XXX
int inode = getinode_from_name((ext2_fs_instance_t*)instance, dentry->d_inode->i_ino, name);
if (inode == -ENOENT && (flags & O_CREAT)) {
//inode = ext2_mknod2((ext2_fs_instance_t*)instance, name, 00644 | 0x8000, 0); //FIXME!
// Plusieurs choses : getinode_from_name retourne ENOTDIR et pas ENOENT,
// flags vaut 0, et ce mknod est réalisé plus haut, peut-être qu'on devrait
// supprimer ce cas tout simplement ?
} else if (inode <= 0) {
return NULL;
} else if (flags & O_EXCL && flags & O_CREAT) {
return NULL;
}
struct ext2_inode *einode = read_inode((ext2_fs_instance_t*)instance, inode);
dentry_t *d = kmalloc(sizeof(dentry_t));
char *n = strdup(name);
d->d_name = (const char*)n;
d->d_pdentry = dentry;
d->d_inode = kmalloc(sizeof(inode_t));
ext2inode_2_inode(d->d_inode, instance, inode, einode);
d->d_inode->i_count = 0;
if (flags & O_TRUNC) {
ext2_truncate(d->d_inode, 0);
}
return d;
}
/*
* Called at the last iput() if i_nlink is zero.
*/
void ext2_delete_inode (struct inode * inode)
{
if (is_bad_inode(inode))
goto no_delete;
EXT2_I(inode)->i_dtime = get_seconds();
mark_inode_dirty(inode);
ext2_update_inode(inode, inode_needs_sync(inode));
inode->i_size = 0;
if (inode->i_blocks)
ext2_truncate (inode);
ext2_free_inode (inode);
return;
no_delete:
clear_inode(inode); /* We must guarantee clearing of inode... */
}
/*
* Called at the last iput() if i_nlink is zero.
*/
void ext2_delete_inode (struct inode * inode)
{
lock_kernel();
if (is_bad_inode(inode))
goto no_delete;
inode->u.ext2_i.i_dtime = CURRENT_TIME;
mark_inode_dirty(inode);
ext2_update_inode(inode, IS_SYNC(inode));
inode->i_size = 0;
if (inode->i_blocks)
ext2_truncate (inode);
ext2_free_inode (inode);
unlock_kernel();
return;
no_delete:
unlock_kernel();
clear_inode(inode); /* We must guarantee clearing of inode... */
}
/*
* Rename system call.
* rename("foo", "bar");
* is essentially
* unlink("bar");
* link("foo", "bar");
* unlink("foo");
* but ``atomically''. Can't do full commit without saving state in the
* inode on disk which isn't feasible at this time. Best we can do is
* always guarantee the target exists.
*
* Basic algorithm is:
*
* 1) Bump link count on source while we're linking it to the
* target. This also ensure the inode won't be deleted out
* from underneath us while we work (it may be truncated by
* a concurrent `trunc' or `open' for creation).
* 2) Link source to destination. If destination already exists,
* delete it first.
* 3) Unlink source reference to inode if still around. If a
* directory was moved and the parent of the destination
* is different from the source, patch the ".." entry in the
* directory.
*/
static int
ext2_rename(struct vop_rename_args *ap)
{
struct vnode *tvp = ap->a_tvp;
struct vnode *tdvp = ap->a_tdvp;
struct vnode *fvp = ap->a_fvp;
struct vnode *fdvp = ap->a_fdvp;
struct componentname *tcnp = ap->a_tcnp;
struct componentname *fcnp = ap->a_fcnp;
struct inode *ip, *xp, *dp;
struct dirtemplate dirbuf;
int doingdirectory = 0, oldparent = 0, newparent = 0;
int error = 0;
u_char namlen;
#ifdef DIAGNOSTIC
if ((tcnp->cn_flags & HASBUF) == 0 ||
(fcnp->cn_flags & HASBUF) == 0)
panic("ext2_rename: no name");
#endif
/*
* Check for cross-device rename.
*/
if ((fvp->v_mount != tdvp->v_mount) ||
(tvp && (fvp->v_mount != tvp->v_mount))) {
error = EXDEV;
abortit:
if (tdvp == tvp)
vrele(tdvp);
else
vput(tdvp);
if (tvp)
vput(tvp);
vrele(fdvp);
vrele(fvp);
return (error);
}
if (tvp && ((VTOI(tvp)->i_flags & (NOUNLINK | IMMUTABLE | APPEND)) ||
(VTOI(tdvp)->i_flags & APPEND))) {
error = EPERM;
goto abortit;
}
/*
* Renaming a file to itself has no effect. The upper layers should
* not call us in that case. Temporarily just warn if they do.
*/
if (fvp == tvp) {
printf("ext2_rename: fvp == tvp (can't happen)\n");
error = 0;
goto abortit;
}
if ((error = vn_lock(fvp, LK_EXCLUSIVE)) != 0)
goto abortit;
dp = VTOI(fdvp);
ip = VTOI(fvp);
if (ip->i_nlink >= EXT2_LINK_MAX) {
VOP_UNLOCK(fvp, 0);
error = EMLINK;
goto abortit;
}
if ((ip->i_flags & (NOUNLINK | IMMUTABLE | APPEND))
|| (dp->i_flags & APPEND)) {
VOP_UNLOCK(fvp, 0);
error = EPERM;
goto abortit;
}
if ((ip->i_mode & IFMT) == IFDIR) {
/*
* Avoid ".", "..", and aliases of "." for obvious reasons.
*/
if ((fcnp->cn_namelen == 1 && fcnp->cn_nameptr[0] == '.') ||
dp == ip || (fcnp->cn_flags | tcnp->cn_flags) & ISDOTDOT ||
(ip->i_flag & IN_RENAME)) {
VOP_UNLOCK(fvp, 0);
error = EINVAL;
goto abortit;
}
ip->i_flag |= IN_RENAME;
oldparent = dp->i_number;
doingdirectory++;
}
vrele(fdvp);
/*
* When the target exists, both the directory
* and target vnodes are returned locked.
*/
dp = VTOI(tdvp);
xp = NULL;
if (tvp)
xp = VTOI(tvp);
/*
* 1) Bump link count while we're moving stuff
* around. If we crash somewhere before
* completing our work, the link count
* may be wrong, but correctable.
*/
ip->i_nlink++;
ip->i_flag |= IN_CHANGE;
if ((error = ext2_update(fvp, !DOINGASYNC(fvp))) != 0) {
VOP_UNLOCK(fvp, 0);
goto bad;
}
/*
* If ".." must be changed (ie the directory gets a new
* parent) then the source directory must not be in the
* directory hierarchy above the target, as this would
* orphan everything below the source directory. Also
* the user must have write permission in the source so
* as to be able to change "..". We must repeat the call
* to namei, as the parent directory is unlocked by the
* call to checkpath().
*/
error = VOP_ACCESS(fvp, VWRITE, tcnp->cn_cred, tcnp->cn_thread);
VOP_UNLOCK(fvp, 0);
if (oldparent != dp->i_number)
newparent = dp->i_number;
if (doingdirectory && newparent) {
if (error) /* write access check above */
goto bad;
if (xp != NULL)
vput(tvp);
error = ext2_checkpath(ip, dp, tcnp->cn_cred);
if (error)
goto out;
VREF(tdvp);
error = relookup(tdvp, &tvp, tcnp);
if (error)
goto out;
vrele(tdvp);
dp = VTOI(tdvp);
xp = NULL;
if (tvp)
xp = VTOI(tvp);
}
/*
* 2) If target doesn't exist, link the target
* to the source and unlink the source.
* Otherwise, rewrite the target directory
* entry to reference the source inode and
* expunge the original entry's existence.
*/
if (xp == NULL) {
if (dp->i_devvp != ip->i_devvp)
panic("ext2_rename: EXDEV");
/*
* Account for ".." in new directory.
* When source and destination have the same
* parent we don't fool with the link count.
*/
if (doingdirectory && newparent) {
if ((nlink_t)dp->i_nlink >= EXT2_LINK_MAX) {
error = EMLINK;
goto bad;
}
dp->i_nlink++;
dp->i_flag |= IN_CHANGE;
error = ext2_update(tdvp, !DOINGASYNC(tdvp));
if (error)
goto bad;
}
error = ext2_direnter(ip, tdvp, tcnp);
if (error) {
if (doingdirectory && newparent) {
dp->i_nlink--;
dp->i_flag |= IN_CHANGE;
(void)ext2_update(tdvp, 1);
}
goto bad;
}
vput(tdvp);
} else {
if (xp->i_devvp != dp->i_devvp || xp->i_devvp != ip->i_devvp)
panic("ext2_rename: EXDEV");
/*
* Short circuit rename(foo, foo).
*/
if (xp->i_number == ip->i_number)
panic("ext2_rename: same file");
/*
* If the parent directory is "sticky", then the user must
* own the parent directory, or the destination of the rename,
* otherwise the destination may not be changed (except by
* root). This implements append-only directories.
*/
if ((dp->i_mode & S_ISTXT) && tcnp->cn_cred->cr_uid != 0 &&
tcnp->cn_cred->cr_uid != dp->i_uid &&
xp->i_uid != tcnp->cn_cred->cr_uid) {
error = EPERM;
goto bad;
}
/*
* Target must be empty if a directory and have no links
* to it. Also, ensure source and target are compatible
* (both directories, or both not directories).
*/
if ((xp->i_mode&IFMT) == IFDIR) {
if (! ext2_dirempty(xp, dp->i_number, tcnp->cn_cred) ||
xp->i_nlink > 2) {
error = ENOTEMPTY;
goto bad;
}
if (!doingdirectory) {
error = ENOTDIR;
goto bad;
}
cache_purge(tdvp);
} else if (doingdirectory) {
error = EISDIR;
goto bad;
}
error = ext2_dirrewrite(dp, ip, tcnp);
if (error)
goto bad;
/*
* If the target directory is in the same
* directory as the source directory,
* decrement the link count on the parent
* of the target directory.
*/
if (doingdirectory && !newparent) {
dp->i_nlink--;
dp->i_flag |= IN_CHANGE;
}
vput(tdvp);
/*
* Adjust the link count of the target to
* reflect the dirrewrite above. If this is
* a directory it is empty and there are
* no links to it, so we can squash the inode and
* any space associated with it. We disallowed
* renaming over top of a directory with links to
* it above, as the remaining link would point to
* a directory without "." or ".." entries.
*/
xp->i_nlink--;
if (doingdirectory) {
if (--xp->i_nlink != 0)
panic("ext2_rename: linked directory");
error = ext2_truncate(tvp, (off_t)0, IO_SYNC,
tcnp->cn_cred, tcnp->cn_thread);
}
xp->i_flag |= IN_CHANGE;
vput(tvp);
xp = NULL;
}
/*
* 3) Unlink the source.
*/
fcnp->cn_flags &= ~MODMASK;
fcnp->cn_flags |= LOCKPARENT | LOCKLEAF;
VREF(fdvp);
error = relookup(fdvp, &fvp, fcnp);
if (error == 0)
vrele(fdvp);
if (fvp != NULL) {
xp = VTOI(fvp);
dp = VTOI(fdvp);
} else {
/*
* From name has disappeared.
*/
if (doingdirectory)
panic("ext2_rename: lost dir entry");
vrele(ap->a_fvp);
return (0);
}
/*
* Ensure that the directory entry still exists and has not
* changed while the new name has been entered. If the source is
* a file then the entry may have been unlinked or renamed. In
* either case there is no further work to be done. If the source
* is a directory then it cannot have been rmdir'ed; its link
* count of three would cause a rmdir to fail with ENOTEMPTY.
* The IN_RENAME flag ensures that it cannot be moved by another
* rename.
*/
if (xp != ip) {
if (doingdirectory)
panic("ext2_rename: lost dir entry");
} else {
/*
* If the source is a directory with a
* new parent, the link count of the old
* parent directory must be decremented
* and ".." set to point to the new parent.
*/
if (doingdirectory && newparent) {
dp->i_nlink--;
dp->i_flag |= IN_CHANGE;
error = vn_rdwr(UIO_READ, fvp, (caddr_t)&dirbuf,
sizeof(struct dirtemplate), (off_t)0,
UIO_SYSSPACE, IO_NODELOCKED | IO_NOMACCHECK,
tcnp->cn_cred, NOCRED, NULL, NULL);
if (error == 0) {
/* Like ufs little-endian: */
namlen = dirbuf.dotdot_type;
if (namlen != 2 ||
dirbuf.dotdot_name[0] != '.' ||
dirbuf.dotdot_name[1] != '.') {
ext2_dirbad(xp, (doff_t)12,
"rename: mangled dir");
} else {
dirbuf.dotdot_ino = newparent;
(void) vn_rdwr(UIO_WRITE, fvp,
(caddr_t)&dirbuf,
sizeof(struct dirtemplate),
(off_t)0, UIO_SYSSPACE,
IO_NODELOCKED | IO_SYNC |
IO_NOMACCHECK, tcnp->cn_cred,
NOCRED, NULL, NULL);
cache_purge(fdvp);
}
}
}
error = ext2_dirremove(fdvp, fcnp);
if (!error) {
xp->i_nlink--;
xp->i_flag |= IN_CHANGE;
}
xp->i_flag &= ~IN_RENAME;
}
if (dp)
vput(fdvp);
if (xp)
vput(fvp);
vrele(ap->a_fvp);
return (error);
bad:
if (xp)
vput(ITOV(xp));
vput(ITOV(dp));
out:
if (doingdirectory)
ip->i_flag &= ~IN_RENAME;
if (vn_lock(fvp, LK_EXCLUSIVE) == 0) {
ip->i_nlink--;
ip->i_flag |= IN_CHANGE;
ip->i_flag &= ~IN_RENAME;
vput(fvp);
} else
vrele(fvp);
return (error);
}
/*
* Set attribute vnode op. called from several syscalls
*/
static int
ext2_setattr(struct vop_setattr_args *ap)
{
struct vattr *vap = ap->a_vap;
struct vnode *vp = ap->a_vp;
struct inode *ip = VTOI(vp);
struct ucred *cred = ap->a_cred;
struct thread *td = curthread;
int error;
/*
* Check for unsettable attributes.
*/
if ((vap->va_type != VNON) || (vap->va_nlink != VNOVAL) ||
(vap->va_fsid != VNOVAL) || (vap->va_fileid != VNOVAL) ||
(vap->va_blocksize != VNOVAL) || (vap->va_rdev != VNOVAL) ||
((int)vap->va_bytes != VNOVAL) || (vap->va_gen != VNOVAL)) {
return (EINVAL);
}
if (vap->va_flags != VNOVAL) {
/* Disallow flags not supported by ext2fs. */
if(vap->va_flags & ~(SF_APPEND | SF_IMMUTABLE | UF_NODUMP))
return (EOPNOTSUPP);
if (vp->v_mount->mnt_flag & MNT_RDONLY)
return (EROFS);
/*
* Callers may only modify the file flags on objects they
* have VADMIN rights for.
*/
if ((error = VOP_ACCESS(vp, VADMIN, cred, td)))
return (error);
/*
* Unprivileged processes and privileged processes in
* jail() are not permitted to unset system flags, or
* modify flags if any system flags are set.
* Privileged non-jail processes may not modify system flags
* if securelevel > 0 and any existing system flags are set.
*/
if (!priv_check_cred(cred, PRIV_VFS_SYSFLAGS, 0)) {
if (ip->i_flags & (SF_IMMUTABLE | SF_APPEND)) {
error = securelevel_gt(cred, 0);
if (error)
return (error);
}
} else {
if (ip->i_flags & (SF_IMMUTABLE | SF_APPEND) ||
((vap->va_flags ^ ip->i_flags) & SF_SETTABLE))
return (EPERM);
}
ip->i_flags = vap->va_flags;
ip->i_flag |= IN_CHANGE;
if (ip->i_flags & (IMMUTABLE | APPEND))
return (0);
}
if (ip->i_flags & (IMMUTABLE | APPEND))
return (EPERM);
/*
* Go through the fields and update iff not VNOVAL.
*/
if (vap->va_uid != (uid_t)VNOVAL || vap->va_gid != (gid_t)VNOVAL) {
if (vp->v_mount->mnt_flag & MNT_RDONLY)
return (EROFS);
if ((error = ext2_chown(vp, vap->va_uid, vap->va_gid, cred,
td)) != 0)
return (error);
}
if (vap->va_size != VNOVAL) {
/*
* Disallow write attempts on read-only file systems;
* unless the file is a socket, fifo, or a block or
* character device resident on the file system.
*/
switch (vp->v_type) {
case VDIR:
return (EISDIR);
case VLNK:
case VREG:
if (vp->v_mount->mnt_flag & MNT_RDONLY)
return (EROFS);
break;
default:
break;
}
if ((error = ext2_truncate(vp, vap->va_size, 0, cred, td)) != 0)
return (error);
}
if (vap->va_atime.tv_sec != VNOVAL || vap->va_mtime.tv_sec != VNOVAL) {
if (vp->v_mount->mnt_flag & MNT_RDONLY)
return (EROFS);
/*
* From utimes(2):
* If times is NULL, ... The caller must be the owner of
* the file, have permission to write the file, or be the
* super-user.
* If times is non-NULL, ... The caller must be the owner of
* the file or be the super-user.
*/
if ((error = VOP_ACCESS(vp, VADMIN, cred, td)) &&
((vap->va_vaflags & VA_UTIMES_NULL) == 0 ||
(error = VOP_ACCESS(vp, VWRITE, cred, td))))
return (error);
if (vap->va_atime.tv_sec != VNOVAL)
ip->i_flag |= IN_ACCESS;
if (vap->va_mtime.tv_sec != VNOVAL)
ip->i_flag |= IN_CHANGE | IN_UPDATE;
ext2_itimes(vp);
if (vap->va_atime.tv_sec != VNOVAL) {
ip->i_atime = vap->va_atime.tv_sec;
ip->i_atimensec = vap->va_atime.tv_nsec;
}
if (vap->va_mtime.tv_sec != VNOVAL) {
ip->i_mtime = vap->va_mtime.tv_sec;
ip->i_mtimensec = vap->va_mtime.tv_nsec;
}
ip->i_birthtime = vap->va_birthtime.tv_sec;
ip->i_birthnsec = vap->va_birthtime.tv_nsec;
error = ext2_update(vp, 0);
if (error)
return (error);
}
error = 0;
if (vap->va_mode != (mode_t)VNOVAL) {
if (vp->v_mount->mnt_flag & MNT_RDONLY)
return (EROFS);
error = ext2_chmod(vp, (int)vap->va_mode, cred, td);
}
return (error);
}
/*
* Vnode op for writing.
*/
static int
ext2_write(struct vop_write_args *ap)
{
struct vnode *vp;
struct uio *uio;
struct inode *ip;
struct m_ext2fs *fs;
struct buf *bp;
daddr_t lbn;
off_t osize;
int blkoffset, error, flags, ioflag, resid, size, seqcount, xfersize;
ioflag = ap->a_ioflag;
uio = ap->a_uio;
vp = ap->a_vp;
seqcount = ioflag >> IO_SEQSHIFT;
ip = VTOI(vp);
#ifdef INVARIANTS
if (uio->uio_rw != UIO_WRITE)
panic("%s: mode", "ext2_write");
#endif
switch (vp->v_type) {
case VREG:
if (ioflag & IO_APPEND)
uio->uio_offset = ip->i_size;
if ((ip->i_flags & APPEND) && uio->uio_offset != ip->i_size)
return (EPERM);
/* FALLTHROUGH */
case VLNK:
break;
case VDIR:
/* XXX differs from ffs -- this is called from ext2_mkdir(). */
if ((ioflag & IO_SYNC) == 0)
panic("ext2_write: nonsync dir write");
break;
default:
panic("ext2_write: type %p %d (%jd,%jd)", (void *)vp,
vp->v_type, (intmax_t)uio->uio_offset,
(intmax_t)uio->uio_resid);
}
KASSERT(uio->uio_resid >= 0, ("ext2_write: uio->uio_resid < 0"));
KASSERT(uio->uio_offset >= 0, ("ext2_write: uio->uio_offset < 0"));
fs = ip->i_e2fs;
if ((uoff_t)uio->uio_offset + uio->uio_resid > fs->e2fs_maxfilesize)
return (EFBIG);
/*
* Maybe this should be above the vnode op call, but so long as
* file servers have no limits, I don't think it matters.
*/
if (vn_rlimit_fsize(vp, uio, uio->uio_td))
return (EFBIG);
resid = uio->uio_resid;
osize = ip->i_size;
if (seqcount > BA_SEQMAX)
flags = BA_SEQMAX << BA_SEQSHIFT;
else
flags = seqcount << BA_SEQSHIFT;
if ((ioflag & IO_SYNC) && !DOINGASYNC(vp))
flags |= IO_SYNC;
for (error = 0; uio->uio_resid > 0;) {
lbn = lblkno(fs, uio->uio_offset);
blkoffset = blkoff(fs, uio->uio_offset);
xfersize = fs->e2fs_fsize - blkoffset;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
if (uio->uio_offset + xfersize > ip->i_size)
vnode_pager_setsize(vp, uio->uio_offset + xfersize);
/*
* We must perform a read-before-write if the transfer size
* does not cover the entire buffer.
*/
if (fs->e2fs_bsize > xfersize)
flags |= BA_CLRBUF;
else
flags &= ~BA_CLRBUF;
error = ext2_balloc(ip, lbn, blkoffset + xfersize,
ap->a_cred, &bp, flags);
if (error != 0)
break;
/*
* If the buffer is not valid and we did not clear garbage
* out above, we have to do so here even though the write
* covers the entire buffer in order to avoid a mmap()/write
* race where another process may see the garbage prior to
* the uiomove() for a write replacing it.
*/
if ((bp->b_flags & B_CACHE) == 0 && fs->e2fs_bsize <= xfersize)
vfs_bio_clrbuf(bp);
if ((ioflag & (IO_SYNC|IO_INVAL)) == (IO_SYNC|IO_INVAL))
bp->b_flags |= B_NOCACHE;
if (uio->uio_offset + xfersize > ip->i_size)
ip->i_size = uio->uio_offset + xfersize;
size = blksize(fs, ip, lbn) - bp->b_resid;
if (size < xfersize)
xfersize = size;
error =
uiomove((char *)bp->b_data + blkoffset, (int)xfersize, uio);
if (ioflag & (IO_VMIO|IO_DIRECT)) {
bp->b_flags |= B_RELBUF;
}
/*
* If IO_SYNC each buffer is written synchronously. Otherwise
* if we have a severe page deficiency write the buffer
* asynchronously. Otherwise try to cluster, and if that
* doesn't do it then either do an async write (if O_DIRECT),
* or a delayed write (if not).
*/
if (ioflag & IO_SYNC) {
(void)bwrite(bp);
} else if (vm_page_count_severe() ||
buf_dirty_count_severe() ||
(ioflag & IO_ASYNC)) {
bp->b_flags |= B_CLUSTEROK;
bawrite(bp);
} else if (xfersize + blkoffset == fs->e2fs_fsize) {
if ((vp->v_mount->mnt_flag & MNT_NOCLUSTERW) == 0) {
bp->b_flags |= B_CLUSTEROK;
cluster_write(vp, bp, ip->i_size, seqcount, 0);
} else {
bawrite(bp);
}
} else if (ioflag & IO_DIRECT) {
bp->b_flags |= B_CLUSTEROK;
bawrite(bp);
} else {
bp->b_flags |= B_CLUSTEROK;
bdwrite(bp);
}
if (error || xfersize == 0)
break;
}
/*
* If we successfully wrote any data, and we are not the superuser
* we clear the setuid and setgid bits as a precaution against
* tampering.
*/
if ((ip->i_mode & (ISUID | ISGID)) && resid > uio->uio_resid &&
ap->a_cred) {
if (priv_check_cred(ap->a_cred, PRIV_VFS_RETAINSUGID, 0))
ip->i_mode &= ~(ISUID | ISGID);
}
if (error) {
if (ioflag & IO_UNIT) {
(void)ext2_truncate(vp, osize,
ioflag & IO_SYNC, ap->a_cred, uio->uio_td);
uio->uio_offset -= resid - uio->uio_resid;
uio->uio_resid = resid;
}
}
if (uio->uio_resid != resid) {
ip->i_flag |= IN_CHANGE | IN_UPDATE;
if (ioflag & IO_SYNC)
error = ext2_update(vp, 1);
}
return (error);
}
/*
* Rmdir system call.
*/
static int
ext2_rmdir(struct vop_rmdir_args *ap)
{
struct vnode *vp = ap->a_vp;
struct vnode *dvp = ap->a_dvp;
struct componentname *cnp = ap->a_cnp;
struct inode *ip, *dp;
int error;
ip = VTOI(vp);
dp = VTOI(dvp);
/*
* Verify the directory is empty (and valid).
* (Rmdir ".." won't be valid since
* ".." will contain a reference to
* the current directory and thus be
* non-empty.)
*/
error = 0;
if (ip->i_nlink != 2 || !ext2_dirempty(ip, dp->i_number, cnp->cn_cred)) {
error = ENOTEMPTY;
goto out;
}
if ((dp->i_flags & APPEND)
|| (ip->i_flags & (NOUNLINK | IMMUTABLE | APPEND))) {
error = EPERM;
goto out;
}
/*
* Delete reference to directory before purging
* inode. If we crash in between, the directory
* will be reattached to lost+found,
*/
error = ext2_dirremove(dvp, cnp);
if (error)
goto out;
dp->i_nlink--;
dp->i_flag |= IN_CHANGE;
cache_purge(dvp);
VOP_UNLOCK(dvp, 0);
/*
* Truncate inode. The only stuff left
* in the directory is "." and "..". The
* "." reference is inconsequential since
* we're quashing it. The ".." reference
* has already been adjusted above. We've
* removed the "." reference and the reference
* in the parent directory, but there may be
* other hard links so decrement by 2 and
* worry about them later.
*/
ip->i_nlink -= 2;
error = ext2_truncate(vp, (off_t)0, IO_SYNC, cnp->cn_cred,
cnp->cn_thread);
cache_purge(ITOV(ip));
if (vn_lock(dvp, LK_EXCLUSIVE | LK_NOWAIT) != 0) {
VOP_UNLOCK(vp, 0);
vn_lock(dvp, LK_EXCLUSIVE | LK_RETRY);
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
}
out:
return (error);
}
int
xfs_write_file(xfs_inode_t *xip, struct uio *uio, int ioflag)
{
struct buf *bp;
//struct thread *td;
daddr_t lbn;
off_t osize = 0;
off_t offset= 0;
int blkoffset, error, resid, xfersize;
int fsblocksize;
int seqcount;
xfs_iomap_t iomap;
int maps = 1;
xfs_vnode_t *xvp = XFS_ITOV(xip);
struct vnode *vp = xvp->v_vnode;
xfs_mount_t *mp = (&xip->i_iocore)->io_mount;
seqcount = ioflag >> IO_SEQSHIFT;
memset(&iomap,0,sizeof(xfs_iomap_t));
/*
* Maybe this should be above the vnode op call, but so long as
* file servers have no limits, I don't think it matters.
*/
#if 0
td = uio->uio_td;
if (vp->v_type == VREG && td != NULL) {
PROC_LOCK(td->td_proc);
if (uio->uio_offset + uio->uio_resid >
lim_cur(td->td_proc, RLIMIT_FSIZE)) {
psignal(td->td_proc, SIGXFSZ);
PROC_UNLOCK(td->td_proc);
return (EFBIG);
}
PROC_UNLOCK(td->td_proc);
}
#endif
resid = uio->uio_resid;
offset = uio->uio_offset;
osize = xip->i_d.di_size;
/* xfs bmap wants bytes for both offset and size */
XVOP_BMAP(xvp,
uio->uio_offset,
uio->uio_resid,
BMAPI_WRITE|BMAPI_DIRECT,
&iomap, &maps, error);
if(error) {
printf("XVOP_BMAP failed\n");
goto error;
}
for (error = 0; uio->uio_resid > 0;) {
lbn = XFS_B_TO_FSBT(mp, offset);
blkoffset = XFS_B_FSB_OFFSET(mp, offset);
xfersize = mp->m_sb.sb_blocksize - blkoffset;
fsblocksize = mp->m_sb.sb_blocksize;
if (uio->uio_resid < xfersize)
xfersize = uio->uio_resid;
/*
* getblk sets buf by blkno * bo->bo_bsize
* bo_bsize is set from the mnt point fsize
* so we call getblk in the case using fsblocks
* not basic blocks
*/
bp = getblk(vp, lbn, fsblocksize, 0, 0, 0);
if(!bp) {
printf("getblk failed\n");
error = EINVAL;
break;
}
if (!(bp->b_flags & B_CACHE) && fsblocksize > xfersize)
vfs_bio_clrbuf(bp);
if (offset + xfersize > xip->i_d.di_size) {
xip->i_d.di_size = offset + xfersize;
vnode_pager_setsize(vp, offset + fsblocksize);
}
/* move the offset for the next itteration of the loop */
offset += xfersize;
error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio);
if ((ioflag & IO_VMIO) &&
(LIST_FIRST(&bp->b_dep) == NULL)) /* in ext2fs? */
bp->b_flags |= B_RELBUF;
/* force to full direct for now */
bp->b_flags |= B_DIRECT;
/* and sync ... the delay path is not pushing data out */
ioflag |= IO_SYNC;
if (ioflag & IO_SYNC) {
(void)bwrite(bp);
} else if (0 /* RMC xfersize + blkoffset == fs->s_frag_size */) {
if ((vp->v_mount->mnt_flag & MNT_NOCLUSTERW) == 0) {
bp->b_flags |= B_CLUSTEROK;
cluster_write(vp, bp, osize, seqcount);
} else {
bawrite(bp);
}
} else {
bp->b_flags |= B_CLUSTEROK;
bdwrite(bp);
}
if (error || xfersize == 0)
break;
}
/*
* If we successfully wrote any data, and we are not the superuser
* we clear the setuid and setgid bits as a precaution against
* tampering.
*/
#if 0
if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0)
ip->i_mode &= ~(ISUID | ISGID);
#endif
if (error) {
if (ioflag & IO_UNIT) {
#if 0
(void)ext2_truncate(vp, osize,
ioflag & IO_SYNC, ap->a_cred, uio->uio_td);
#endif
uio->uio_offset -= resid - uio->uio_resid;
uio->uio_resid = resid;
}
} else if (resid > uio->uio_resid && (ioflag & IO_SYNC)) {
/* Update the vnode here? */
}
error:
return error;
}
