static inline long do_sys_truncate(const char * path, loff_t length)
{
struct nameidata nd;
struct inode * inode;
int error;
error = -EINVAL;
if (length < 0) /* sorry, but loff_t says... */
goto out;
error = user_path_walk(path, &nd);
if (error)
goto out;
inode = nd.dentry->d_inode;
/* For directories it's -EISDIR, for other non-regulars - -EINVAL */
error = -EISDIR;
if (S_ISDIR(inode->i_mode))
goto dput_and_out;
error = -EINVAL;
if (!S_ISREG(inode->i_mode))
goto dput_and_out;
error = permission(inode,MAY_WRITE);
if (error)
goto dput_and_out;
error = -EROFS;
if (IS_RDONLY(inode))
goto dput_and_out;
error = -EPERM;
if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
goto dput_and_out;
/*
* Make sure that there are no leases.
*/
error = get_lease(inode, FMODE_WRITE);
if (error)
goto dput_and_out;
error = get_write_access(inode);
if (error)
goto dput_and_out;
error = locks_verify_truncate(inode, NULL, length);
if (!error)
{
DQUOT_INIT(inode);
error = do_truncate(nd.dentry, length);
}
put_write_access(inode);
dput_and_out:
path_release(&nd);
out:
return error;
}
/*
* Open an existing file or directory.
* The wflag argument indicates write access.
* N.B. After this call fhp needs an fh_put
*/
int
nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
int wflag, struct file *filp)
{
struct dentry *dentry;
struct inode *inode;
int access, err;
access = wflag? MAY_WRITE : MAY_READ;
err = fh_verify(rqstp, fhp, type, access);
if (err)
goto out;
dentry = fhp->fh_dentry;
inode = dentry->d_inode;
/* Disallow access to files with the append-only bit set or
* with mandatory locking enabled
*/
err = nfserr_perm;
if (IS_APPEND(inode) || IS_ISMNDLK(inode))
goto out;
if (!inode->i_op || !inode->i_op->default_file_ops)
goto out;
if (wflag && (err = get_write_access(inode)) != 0)
goto out_nfserr;
memset(filp, 0, sizeof(*filp));
filp->f_op = inode->i_op->default_file_ops;
filp->f_count = 1;
filp->f_flags = wflag? O_WRONLY : O_RDONLY;
filp->f_mode = wflag? FMODE_WRITE : FMODE_READ;
filp->f_dentry = dentry;
if (wflag)
DQUOT_INIT(inode);
err = 0;
if (filp->f_op && filp->f_op->open) {
err = filp->f_op->open(inode, filp);
if (err) {
if (wflag)
put_write_access(inode);
/* I nearly added put_filp() call here, but this filp
* is really on callers stack frame. -DaveM
*/
filp->f_count--;
}
}
out_nfserr:
if (err)
err = nfserrno(-err);
out:
return err;
}
/*
* Truncate a file.
* The calling routines must make sure to update the ctime
* field and call notify_change.
*
* XXX Nobody calls this thing? -DaveM
* N.B. After this call fhp needs an fh_put
*/
int
nfsd_truncate(struct svc_rqst *rqstp, struct svc_fh *fhp, unsigned long size)
{
struct dentry *dentry;
struct inode *inode;
struct iattr newattrs;
int err;
kernel_cap_t saved_cap;
err = fh_verify(rqstp, fhp, S_IFREG, MAY_WRITE | MAY_TRUNC);
if (err)
goto out;
dentry = fhp->fh_dentry;
inode = dentry->d_inode;
err = get_write_access(inode);
if (err)
goto out_nfserr;
/* Things look sane, lock and do it. */
fh_lock(fhp);
DQUOT_INIT(inode);
newattrs.ia_size = size;
newattrs.ia_valid = ATTR_SIZE | ATTR_CTIME;
if (current->fsuid != 0) {
saved_cap = current->cap_effective;
cap_clear(current->cap_effective);
}
err = notify_change(dentry, &newattrs);
if (current->fsuid != 0)
current->cap_effective = saved_cap;
if (!err) {
vmtruncate(inode, size);
if (inode->i_op && inode->i_op->truncate)
inode->i_op->truncate(inode);
}
put_write_access(inode);
DQUOT_DROP(inode);
fh_unlock(fhp);
out_nfserr:
if (err)
err = nfserrno(-err);
out:
return err;
}
void jfs_delete_inode(struct inode *inode)
{
jfs_info("In jfs_delete_inode, inode = 0x%p", inode);
if (test_cflag(COMMIT_Freewmap, inode))
freeZeroLink(inode);
diFree(inode);
/*
* Free the inode from the quota allocation.
*/
DQUOT_INIT(inode);
DQUOT_FREE_INODE(inode);
DQUOT_DROP(inode);
clear_inode(inode);
}
/* utility function that does setup for reiserfs_new_inode.
** DQUOT_INIT needs lots of credits so it's better to have it
** outside of a transaction, so we had to pull some bits of
** reiserfs_new_inode out into this func.
*/
static int new_inode_init(struct inode *inode, struct inode *dir, int mode) {
/* the quota init calls have to know who to charge the quota to, so
** we have to set uid and gid here
*/
inode->i_uid = current->fsuid;
inode->i_mode = mode;
if (dir->i_mode & S_ISGID) {
inode->i_gid = dir->i_gid;
if (S_ISDIR(mode))
inode->i_mode |= S_ISGID;
} else {
inode->i_gid = current->fsgid;
}
DQUOT_INIT(inode);
return 0 ;
}
/* utility function that does setup for reiserfs_new_inode.
** DQUOT_INIT needs lots of credits so it's better to have it
** outside of a transaction, so we had to pull some bits of
** reiserfs_new_inode out into this func.
*/
static int new_inode_init(struct inode *inode, struct inode *dir, int mode)
{
/* the quota init calls have to know who to charge the quota to, so
** we have to set uid and gid here
*/
inode->i_uid = current->fsuid;
inode->i_mode = mode;
/* Make inode invalid - just in case we are going to drop it before
* the initialization happens */
INODE_PKEY(inode)->k_objectid = 0;
if (dir->i_mode & S_ISGID) {
inode->i_gid = dir->i_gid;
if (S_ISDIR(mode))
inode->i_mode |= S_ISGID;
} else {
inode->i_gid = current->fsgid;
}
DQUOT_INIT(inode);
return 0;
}
void jfs_delete_inode(struct inode *inode)
{
jfs_info("In jfs_delete_inode, inode = 0x%p", inode);
if (!is_bad_inode(inode) &&
(JFS_IP(inode)->fileset == FILESYSTEM_I)) {
truncate_inode_pages(&inode->i_data, 0);
if (test_cflag(COMMIT_Freewmap, inode))
jfs_free_zero_link(inode);
diFree(inode);
/*
* Free the inode from the quota allocation.
*/
DQUOT_INIT(inode);
DQUOT_FREE_INODE(inode);
DQUOT_DROP(inode);
}
clear_inode(inode);
}
static long do_sys_truncate(const char __user *pathname, loff_t length)
{
struct path path;
struct inode *inode;
int error;
error = -EINVAL;
if (length < 0) /* sorry, but loff_t says... */
goto out;
error = user_path(pathname, &path);
if (error)
goto out;
inode = path.dentry->d_inode;
/* For directories it's -EISDIR, for other non-regulars - -EINVAL */
error = -EISDIR;
if (S_ISDIR(inode->i_mode))
goto dput_and_out;
error = -EINVAL;
if (!S_ISREG(inode->i_mode))
goto dput_and_out;
error = mnt_want_write(path.mnt);
if (error)
goto dput_and_out;
error = inode_permission(inode, MAY_WRITE);
if (error)
goto mnt_drop_write_and_out;
error = -EPERM;
if (IS_APPEND(inode))
goto mnt_drop_write_and_out;
error = get_write_access(inode);
if (error)
goto mnt_drop_write_and_out;
/*
* Make sure that there are no leases. get_write_access() protects
* against the truncate racing with a lease-granting setlease().
*/
error = break_lease(inode, FMODE_WRITE);
if (error)
goto put_write_and_out;
error = locks_verify_truncate(inode, NULL, length);
if (!error)
error = security_path_truncate(&path, length, 0);
if (!error) {
DQUOT_INIT(inode);
error = do_truncate(path.dentry, length, 0, NULL);
}
put_write_and_out:
put_write_access(inode);
mnt_drop_write_and_out:
mnt_drop_write(path.mnt);
dput_and_out:
path_put(&path);
out:
return error;
}
/*
* Set various file attributes.
* N.B. After this call fhp needs an fh_put
*/
int
nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap,
int check_guard, time_t guardtime)
{
struct dentry *dentry;
struct inode *inode;
int accmode = MAY_SATTR;
int ftype = 0;
int imode;
int err;
int size_change = 0;
if (iap->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_SIZE))
accmode |= MAY_WRITE|MAY_OWNER_OVERRIDE;
if (iap->ia_valid & ATTR_SIZE)
ftype = S_IFREG;
/* Get inode */
err = fh_verify(rqstp, fhp, ftype, accmode);
if (err || !iap->ia_valid)
goto out;
dentry = fhp->fh_dentry;
inode = dentry->d_inode;
/* NFSv2 does not differentiate between "set-[ac]time-to-now"
* which only requires access, and "set-[ac]time-to-X" which
* requires ownership.
* So if it looks like it might be "set both to the same time which
* is close to now", and if inode_change_ok fails, then we
* convert to "set to now" instead of "set to explicit time"
*
* We only call inode_change_ok as the last test as technically
* it is not an interface that we should be using. It is only
* valid if the filesystem does not define it's own i_op->setattr.
*/
#define BOTH_TIME_SET (ATTR_ATIME_SET | ATTR_MTIME_SET)
#define MAX_TOUCH_TIME_ERROR (30*60)
if ((iap->ia_valid & BOTH_TIME_SET) == BOTH_TIME_SET
&& iap->ia_mtime == iap->ia_atime
) {
/* Looks probable. Now just make sure time is in the right ballpark.
* Solaris, at least, doesn't seem to care what the time request is.
* We require it be within 30 minutes of now.
*/
time_t delta = iap->ia_atime - CURRENT_TIME;
if (delta<0) delta = -delta;
if (delta < MAX_TOUCH_TIME_ERROR &&
inode_change_ok(inode, iap) != 0) {
/* turn off ATTR_[AM]TIME_SET but leave ATTR_[AM]TIME
* this will cause notify_change to set these times to "now"
*/
iap->ia_valid &= ~BOTH_TIME_SET;
}
}
/* The size case is special. It changes the file as well as the attributes. */
if (iap->ia_valid & ATTR_SIZE) {
if (iap->ia_size < inode->i_size) {
err = nfsd_permission(fhp->fh_export, dentry, MAY_TRUNC|MAY_OWNER_OVERRIDE);
if (err)
goto out;
}
/*
* If we are changing the size of the file, then
* we need to break all leases.
*/
err = get_lease(inode, FMODE_WRITE);
if (err)
goto out_nfserr;
err = get_write_access(inode);
if (err)
goto out_nfserr;
err = locks_verify_truncate(inode, NULL, iap->ia_size);
if (err) {
put_write_access(inode);
goto out_nfserr;
}
DQUOT_INIT(inode);
}
imode = inode->i_mode;
if (iap->ia_valid & ATTR_MODE) {
iap->ia_mode &= S_IALLUGO;
imode = iap->ia_mode |= (imode & ~S_IALLUGO);
}
/* Revoke setuid/setgid bit on chown/chgrp */
if ((iap->ia_valid & ATTR_UID) && (imode & S_ISUID)
&& iap->ia_uid != inode->i_uid) {
iap->ia_valid |= ATTR_MODE;
iap->ia_mode = imode &= ~S_ISUID;
}
if ((iap->ia_valid & ATTR_GID) && (imode & S_ISGID)
&& iap->ia_gid != inode->i_gid) {
iap->ia_valid |= ATTR_MODE;
iap->ia_mode = imode &= ~S_ISGID;
}
/* Change the attributes. */
iap->ia_valid |= ATTR_CTIME;
if (iap->ia_valid & ATTR_SIZE) {
fh_lock(fhp);
size_change = 1;
}
err = nfserr_notsync;
if (!check_guard || guardtime == inode->i_ctime) {
err = notify_change(dentry, iap);
err = nfserrno(err);
}
if (size_change) {
fh_unlock(fhp);
put_write_access(inode);
}
if (!err)
if (EX_ISSYNC(fhp->fh_export))
write_inode_now(inode, 1);
out:
return err;
out_nfserr:
err = nfserrno(err);
goto out;
}
static int may_handle_open(struct dentry *dentry, int open_flag)
{
int acc_mode;
int error;
struct inode *inode = dentry->d_inode;
if((open_flag + 1) & O_ACCMODE)
open_flag++;
acc_mode = ACC_MODE(open_flag);
/* O_TRUNC implies we need access checks for write permissions */
if(open_flag & O_TRUNC)
acc_mode |= MAY_WRITE;
/* Allow the LSM permission hook to distinguish append
access from general write access. */
if(open_flag & O_APPEND)
acc_mode |= MAY_APPEND;
if(S_ISDIR(inode->i_mode) && (acc_mode & MAY_WRITE))
return -EISDIR;
error = handle_permission(inode, acc_mode);
if (error)
return error;
if(S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode))
{
open_flag &= ~O_TRUNC;
}
else if(S_ISBLK(inode->i_mode) || S_ISCHR(inode->i_mode))
{
#ifdef IN_KERNEL_CHANGE_NOT_SUPP
if(nd->mnt->mnt_flags & MNT_NODEV)
return -EACCES;
#endif
open_flag &= ~O_TRUNC;
}
else if(IS_RDONLY(inode) && (acc_mode & MAY_WRITE))
return -EROFS;
/*
* An append-only file must be opened in append mode for writing.
*/
if(IS_APPEND(inode))
{
if((open_flag & FMODE_WRITE) && !(open_flag & O_APPEND))
return -EPERM;
if(open_flag & O_TRUNC)
return -EPERM;
}
/* O_NOATIME can only be set by the owner or superuser */
if(open_flag & O_NOATIME)
if(current->fsuid != inode->i_uid && !capable(CAP_FOWNER))
return -EPERM;
/*
* Ensure there are no outstanding leases on the file.
*/
error = break_lease(inode, open_flag);
if(error)
return error;
if(open_flag & O_TRUNC)
{
error = get_write_access(inode);
if(error)
return error;
#ifdef IN_KERNEL_CHANGE_NOT_SUPP
/*
* Refuse to truncate files with mandatory locks held on them.
*/
error = locks_verify_locked(inode);
#endif
if(!error)
{
DQUOT_INIT(inode);
error = handle_truncate(dentry, 0, ATTR_MTIME | ATTR_CTIME);
}
put_write_access(inode);
if(error)
return error;
}
else if(open_flag & FMODE_WRITE)
DQUOT_INIT(inode);
return 0;
}
/*
* NAME: jfs_unlink(dip, dentry)
*
* FUNCTION: remove a link to object <vp> named by <name>
* from parent directory <dvp>
*
* PARAMETER: dip - inode of parent directory
* dentry - dentry of object to be removed
*
* RETURN: errors from subroutines
*
* note:
* temporary file: if one or more processes have the file open
* when the last link is removed, the link will be removed before
* unlink() returns, but the removal of the file contents will be
* postponed until all references to the files are closed.
*
* JFS does NOT support unlink() on directories.
*
*/
static int jfs_unlink(struct inode *dip, struct dentry *dentry)
{
int rc;
tid_t tid; /* transaction id */
struct inode *ip = dentry->d_inode;
ino_t ino;
struct component_name dname; /* object name */
struct inode *iplist[2];
struct tblock *tblk;
s64 new_size = 0;
int commit_flag;
jfs_info("jfs_unlink: dip:0x%p name:%s", dip, dentry->d_name.name);
/* Init inode for quota operations. */
DQUOT_INIT(ip);
if ((rc = get_UCSname(&dname, dentry)))
goto out;
IWRITE_LOCK(ip, RDWRLOCK_NORMAL);
tid = txBegin(dip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dip)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
iplist[0] = dip;
iplist[1] = ip;
/*
* delete the entry of target file from parent directory
*/
ino = ip->i_ino;
if ((rc = dtDelete(tid, dip, &dname, &ino, JFS_REMOVE))) {
jfs_err("jfs_unlink: dtDelete returned %d", rc);
if (rc == -EIO)
txAbort(tid, 1); /* Marks FS Dirty */
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
IWRITE_UNLOCK(ip);
goto out1;
}
ASSERT(ip->i_nlink);
ip->i_ctime = dip->i_ctime = dip->i_mtime = CURRENT_TIME;
mark_inode_dirty(dip);
/* update target's inode */
inode_dec_link_count(ip);
/*
* commit zero link count object
*/
if (ip->i_nlink == 0) {
assert(!test_cflag(COMMIT_Nolink, ip));
/* free block resources */
if ((new_size = commitZeroLink(tid, ip)) < 0) {
txAbort(tid, 1); /* Marks FS Dirty */
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
IWRITE_UNLOCK(ip);
rc = new_size;
goto out1;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = ip;
}
/*
* Incomplete truncate of file data can
* result in timing problems unless we synchronously commit the
* transaction.
*/
if (new_size)
commit_flag = COMMIT_SYNC;
else
commit_flag = 0;
/*
* If xtTruncate was incomplete, commit synchronously to avoid
* timing complications
*/
rc = txCommit(tid, 2, &iplist[0], commit_flag);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
while (new_size && (rc == 0)) {
tid = txBegin(dip->i_sb, 0);
mutex_lock(&JFS_IP(ip)->commit_mutex);
new_size = xtTruncate_pmap(tid, ip, new_size);
if (new_size < 0) {
txAbort(tid, 1); /* Marks FS Dirty */
rc = new_size;
} else
rc = txCommit(tid, 2, &iplist[0], COMMIT_SYNC);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
}
if (ip->i_nlink == 0)
set_cflag(COMMIT_Nolink, ip);
IWRITE_UNLOCK(ip);
/*
* Truncating the directory index table is not guaranteed. It
* may need to be done iteratively
*/
if (test_cflag(COMMIT_Stale, dip)) {
if (dip->i_size > 1)
jfs_truncate_nolock(dip, 0);
clear_cflag(COMMIT_Stale, dip);
}
out1:
free_UCSname(&dname);
out:
jfs_info("jfs_unlink: rc:%d", rc);
return rc;
}
/*
* NAME: jfs_rename
*
* FUNCTION: rename a file or directory
*/
static int jfs_rename(struct inode *old_dir, struct dentry *old_dentry,
struct inode *new_dir, struct dentry *new_dentry)
{
struct btstack btstack;
ino_t ino;
struct component_name new_dname;
struct inode *new_ip;
struct component_name old_dname;
struct inode *old_ip;
int rc;
tid_t tid;
struct tlock *tlck;
struct dt_lock *dtlck;
struct lv *lv;
int ipcount;
struct inode *iplist[4];
struct tblock *tblk;
s64 new_size = 0;
int commit_flag;
jfs_info("jfs_rename: %s %s", old_dentry->d_name.name,
new_dentry->d_name.name);
old_ip = old_dentry->d_inode;
new_ip = new_dentry->d_inode;
if ((rc = get_UCSname(&old_dname, old_dentry)))
goto out1;
if ((rc = get_UCSname(&new_dname, new_dentry)))
goto out2;
/*
* Make sure source inode number is what we think it is
*/
rc = dtSearch(old_dir, &old_dname, &ino, &btstack, JFS_LOOKUP);
if (rc || (ino != old_ip->i_ino)) {
rc = -ENOENT;
goto out3;
}
/*
* Make sure dest inode number (if any) is what we think it is
*/
rc = dtSearch(new_dir, &new_dname, &ino, &btstack, JFS_LOOKUP);
if (rc == 0) {
if ((new_ip == 0) || (ino != new_ip->i_ino)) {
rc = -ESTALE;
goto out3;
}
} else if (rc != -ENOENT)
goto out3;
else if (new_ip) {
/* no entry exists, but one was expected */
rc = -ESTALE;
goto out3;
}
if (S_ISDIR(old_ip->i_mode)) {
if (new_ip) {
if (!dtEmpty(new_ip)) {
rc = -ENOTEMPTY;
goto out3;
}
} else if ((new_dir != old_dir) &&
(new_dir->i_nlink == JFS_LINK_MAX)) {
rc = -EMLINK;
goto out3;
}
} else if (new_ip) {
IWRITE_LOCK(new_ip, RDWRLOCK_NORMAL);
/* Init inode for quota operations. */
DQUOT_INIT(new_ip);
}
/*
* The real work starts here
*/
tid = txBegin(new_dir->i_sb, 0);
/*
* How do we know the locking is safe from deadlocks?
* The vfs does the hard part for us. Any time we are taking nested
* commit_mutexes, the vfs already has i_mutex held on the parent.
* Here, the vfs has already taken i_mutex on both old_dir and new_dir.
*/
mutex_lock_nested(&JFS_IP(new_dir)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(old_ip)->commit_mutex, COMMIT_MUTEX_CHILD);
if (old_dir != new_dir)
mutex_lock_nested(&JFS_IP(old_dir)->commit_mutex,
COMMIT_MUTEX_SECOND_PARENT);
if (new_ip) {
mutex_lock_nested(&JFS_IP(new_ip)->commit_mutex,
COMMIT_MUTEX_VICTIM);
/*
* Change existing directory entry to new inode number
*/
ino = new_ip->i_ino;
rc = dtModify(tid, new_dir, &new_dname, &ino,
old_ip->i_ino, JFS_RENAME);
if (rc)
goto out4;
drop_nlink(new_ip);
if (S_ISDIR(new_ip->i_mode)) {
drop_nlink(new_ip);
if (new_ip->i_nlink) {
mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
if (old_dir != new_dir)
mutex_unlock(&JFS_IP(old_dir)->commit_mutex);
mutex_unlock(&JFS_IP(old_ip)->commit_mutex);
mutex_unlock(&JFS_IP(new_dir)->commit_mutex);
if (!S_ISDIR(old_ip->i_mode) && new_ip)
IWRITE_UNLOCK(new_ip);
jfs_error(new_ip->i_sb,
"jfs_rename: new_ip->i_nlink != 0");
return -EIO;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = new_ip;
} else if (new_ip->i_nlink == 0) {
assert(!test_cflag(COMMIT_Nolink, new_ip));
/* free block resources */
if ((new_size = commitZeroLink(tid, new_ip)) < 0) {
txAbort(tid, 1); /* Marks FS Dirty */
rc = new_size;
goto out4;
}
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = new_ip;
} else {
new_ip->i_ctime = CURRENT_TIME;
mark_inode_dirty(new_ip);
}
} else {
/*
* Add new directory entry
*/
rc = dtSearch(new_dir, &new_dname, &ino, &btstack,
JFS_CREATE);
if (rc) {
jfs_err("jfs_rename didn't expect dtSearch to fail "
"w/rc = %d", rc);
goto out4;
}
ino = old_ip->i_ino;
rc = dtInsert(tid, new_dir, &new_dname, &ino, &btstack);
if (rc) {
if (rc == -EIO)
jfs_err("jfs_rename: dtInsert returned -EIO");
goto out4;
}
if (S_ISDIR(old_ip->i_mode))
inc_nlink(new_dir);
}
/*
* Remove old directory entry
*/
ino = old_ip->i_ino;
rc = dtDelete(tid, old_dir, &old_dname, &ino, JFS_REMOVE);
if (rc) {
jfs_err("jfs_rename did not expect dtDelete to return rc = %d",
rc);
txAbort(tid, 1); /* Marks Filesystem dirty */
goto out4;
}
if (S_ISDIR(old_ip->i_mode)) {
drop_nlink(old_dir);
if (old_dir != new_dir) {
/*
* Change inode number of parent for moved directory
*/
JFS_IP(old_ip)->i_dtroot.header.idotdot =
cpu_to_le32(new_dir->i_ino);
/* Linelock header of dtree */
tlck = txLock(tid, old_ip,
(struct metapage *) &JFS_IP(old_ip)->bxflag,
tlckDTREE | tlckBTROOT | tlckRELINK);
dtlck = (struct dt_lock *) & tlck->lock;
ASSERT(dtlck->index == 0);
lv = & dtlck->lv[0];
lv->offset = 0;
lv->length = 1;
dtlck->index++;
}
}
/*
* Update ctime on changed/moved inodes & mark dirty
*/
old_ip->i_ctime = CURRENT_TIME;
mark_inode_dirty(old_ip);
new_dir->i_ctime = new_dir->i_mtime = current_fs_time(new_dir->i_sb);
mark_inode_dirty(new_dir);
/* Build list of inodes modified by this transaction */
ipcount = 0;
iplist[ipcount++] = old_ip;
if (new_ip)
iplist[ipcount++] = new_ip;
iplist[ipcount++] = old_dir;
if (old_dir != new_dir) {
iplist[ipcount++] = new_dir;
old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME;
mark_inode_dirty(old_dir);
}
/*
* Incomplete truncate of file data can
* result in timing problems unless we synchronously commit the
* transaction.
*/
if (new_size)
commit_flag = COMMIT_SYNC;
else
commit_flag = 0;
rc = txCommit(tid, ipcount, iplist, commit_flag);
out4:
txEnd(tid);
if (new_ip)
mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
if (old_dir != new_dir)
mutex_unlock(&JFS_IP(old_dir)->commit_mutex);
mutex_unlock(&JFS_IP(old_ip)->commit_mutex);
mutex_unlock(&JFS_IP(new_dir)->commit_mutex);
while (new_size && (rc == 0)) {
tid = txBegin(new_ip->i_sb, 0);
mutex_lock(&JFS_IP(new_ip)->commit_mutex);
new_size = xtTruncate_pmap(tid, new_ip, new_size);
if (new_size < 0) {
txAbort(tid, 1);
rc = new_size;
} else
rc = txCommit(tid, 1, &new_ip, COMMIT_SYNC);
txEnd(tid);
mutex_unlock(&JFS_IP(new_ip)->commit_mutex);
}
if (new_ip && (new_ip->i_nlink == 0))
set_cflag(COMMIT_Nolink, new_ip);
out3:
free_UCSname(&new_dname);
out2:
free_UCSname(&old_dname);
out1:
if (new_ip && !S_ISDIR(new_ip->i_mode))
IWRITE_UNLOCK(new_ip);
/*
* Truncating the directory index table is not guaranteed. It
* may need to be done iteratively
*/
if (test_cflag(COMMIT_Stale, old_dir)) {
if (old_dir->i_size > 1)
jfs_truncate_nolock(old_dir, 0);
clear_cflag(COMMIT_Stale, old_dir);
}
jfs_info("jfs_rename: returning %d", rc);
return rc;
}
/*
* Create a hardlink
* N.B. After this call _both_ ffhp and tfhp need an fh_put
*/
int
nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp,
char *fname, int len, struct svc_fh *tfhp)
{
struct dentry *ddir, *dnew, *dold;
struct inode *dirp, *dest;
int err;
err = fh_verify(rqstp, ffhp, S_IFDIR, MAY_CREATE);
if (err)
goto out;
err = fh_verify(rqstp, tfhp, S_IFREG, MAY_NOP);
if (err)
goto out;
err = nfserr_perm;
if (!len)
goto out;
ddir = ffhp->fh_dentry;
dirp = ddir->d_inode;
dnew = lookup_dentry(fname, dget(ddir), 0);
err = PTR_ERR(dnew);
if (IS_ERR(dnew))
goto out_nfserr;
/*
* Lock the parent before checking for existence
*/
err = fh_lock_parent(ffhp, dnew);
if (err)
goto out_dput;
err = nfserr_exist;
if (dnew->d_inode)
goto out_unlock;
dold = tfhp->fh_dentry;
dest = dold->d_inode;
err = nfserr_acces;
if (nfsd_iscovered(ddir, ffhp->fh_export))
goto out_unlock;
/* FIXME: nxdev for NFSv3 */
if (dirp->i_dev != dest->i_dev)
goto out_unlock;
err = nfserr_perm;
if (IS_IMMUTABLE(dest) /* || IS_APPEND(dest) */ )
goto out_unlock;
if (!dirp->i_op || !dirp->i_op->link)
goto out_unlock;
DQUOT_INIT(dirp);
err = dirp->i_op->link(dold, dirp, dnew);
DQUOT_DROP(dirp);
if (!err) {
if (EX_ISSYNC(ffhp->fh_export)) {
write_inode_now(dirp);
write_inode_now(dest);
}
} else
err = nfserrno(-err);
out_unlock:
fh_unlock(ffhp);
out_dput:
dput(dnew);
out:
return err;
out_nfserr:
err = nfserrno(-err);
goto out;
}
/*
* Create a symlink and look up its inode
* N.B. After this call _both_ fhp and resfhp need an fh_put
*/
int
nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp,
char *fname, int flen,
char *path, int plen,
struct svc_fh *resfhp)
{
struct dentry *dentry, *dnew;
struct inode *dirp;
int err;
err = nfserr_noent;
if (!flen || !plen)
goto out;
err = fh_verify(rqstp, fhp, S_IFDIR, MAY_CREATE);
if (err)
goto out;
dentry = fhp->fh_dentry;
err = nfserr_perm;
if (nfsd_iscovered(dentry, fhp->fh_export))
goto out;
dirp = dentry->d_inode;
if (!dirp->i_op || !dirp->i_op->symlink)
goto out;
dnew = lookup_dentry(fname, dget(dentry), 0);
err = PTR_ERR(dnew);
if (IS_ERR(dnew))
goto out_nfserr;
/*
* Lock the parent before checking for existence
*/
err = fh_lock_parent(fhp, dnew);
if (err)
goto out_compose;
err = nfserr_exist;
if (!dnew->d_inode) {
DQUOT_INIT(dirp);
err = dirp->i_op->symlink(dirp, dnew, path);
DQUOT_DROP(dirp);
if (!err) {
if (EX_ISSYNC(fhp->fh_export))
write_inode_now(dirp);
} else
err = nfserrno(-err);
}
fh_unlock(fhp);
/* Compose the fh so the dentry will be freed ... */
out_compose:
fh_compose(resfhp, fhp->fh_export, dnew);
out:
return err;
out_nfserr:
err = nfserrno(-err);
goto out;
}
/*
* Create a file (regular, directory, device, fifo); UNIX sockets
* not yet implemented.
* If the response fh has been verified, the parent directory should
* already be locked. Note that the parent directory is left locked.
*
* N.B. Every call to nfsd_create needs an fh_put for _both_ fhp and resfhp
*/
int
nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp,
char *fname, int flen, struct iattr *iap,
int type, dev_t rdev, struct svc_fh *resfhp)
{
struct dentry *dentry, *dchild;
struct inode *dirp;
nfsd_dirop_t opfunc = NULL;
int err;
err = nfserr_perm;
if (!flen)
goto out;
err = fh_verify(rqstp, fhp, S_IFDIR, MAY_CREATE);
if (err)
goto out;
dentry = fhp->fh_dentry;
dirp = dentry->d_inode;
err = nfserr_notdir;
if(!dirp->i_op || !dirp->i_op->lookup)
goto out;
/*
* Check whether the response file handle has been verified yet.
* If it has, the parent directory should already be locked.
*/
if (!resfhp->fh_dverified) {
dchild = lookup_dentry(fname, dget(dentry), 0);
err = PTR_ERR(dchild);
if (IS_ERR(dchild))
goto out_nfserr;
fh_compose(resfhp, fhp->fh_export, dchild);
/* Lock the parent and check for errors ... */
err = fh_lock_parent(fhp, dchild);
if (err)
goto out;
} else {
dchild = resfhp->fh_dentry;
if (!fhp->fh_locked)
printk(KERN_ERR
"nfsd_create: parent %s/%s not locked!\n",
dentry->d_parent->d_name.name,
dentry->d_name.name);
}
/*
* Make sure the child dentry is still negative ...
*/
err = nfserr_exist;
if (dchild->d_inode) {
printk(KERN_WARNING
"nfsd_create: dentry %s/%s not negative!\n",
dentry->d_name.name, dchild->d_name.name);
goto out;
}
/*
* Get the dir op function pointer.
*/
err = nfserr_perm;
switch (type) {
case S_IFREG:
opfunc = (nfsd_dirop_t) dirp->i_op->create;
break;
case S_IFDIR:
opfunc = (nfsd_dirop_t) dirp->i_op->mkdir;
break;
case S_IFCHR:
case S_IFBLK:
/* The client is _NOT_ required to do security enforcement */
if(!capable(CAP_SYS_ADMIN))
{
err = -EPERM;
goto out;
}
case S_IFIFO:
case S_IFSOCK:
opfunc = dirp->i_op->mknod;
break;
}
if (!opfunc)
goto out;
if (!(iap->ia_valid & ATTR_MODE))
iap->ia_mode = 0;
/*
* Call the dir op function to create the object.
*/
DQUOT_INIT(dirp);
err = opfunc(dirp, dchild, iap->ia_mode, rdev);
DQUOT_DROP(dirp);
if (err < 0)
goto out_nfserr;
if (EX_ISSYNC(fhp->fh_export))
write_inode_now(dirp);
/*
* Update the file handle to get the new inode info.
*/
fh_update(resfhp);
/* Set file attributes. Mode has already been set and
* setting uid/gid works only for root. Irix appears to
* send along the gid when it tries to implement setgid
* directories via NFS.
*/
err = 0;
if ((iap->ia_valid &= (ATTR_UID|ATTR_GID|ATTR_MODE)) != 0)
err = nfsd_setattr(rqstp, resfhp, iap);
out:
return err;
out_nfserr:
err = nfserrno(-err);
goto out;
}
/*
* NOTE! When we get the inode, we're the only people
* that have access to it, and as such there are no
* race conditions we have to worry about. The inode
* is not on the hash-lists, and it cannot be reached
* through the filesystem because the directory entry
* has been deleted earlier.
*
* HOWEVER: we must make sure that we get no aliases,
* which means that we have to call "clear_inode()"
* _before_ we mark the inode not in use in the inode
* bitmaps. Otherwise a newly created file might use
* the same inode number (not actually the same pointer
* though), and then we'd have two inodes sharing the
* same inode number and space on the harddisk.
*/
void ext3_free_inode (handle_t *handle, struct inode * inode)
{
struct super_block * sb = inode->i_sb;
int is_directory;
unsigned long ino;
struct buffer_head *bitmap_bh = NULL;
struct buffer_head *bh2;
unsigned long block_group;
unsigned long bit;
struct ext3_group_desc * gdp;
struct ext3_super_block * es;
struct ext3_sb_info *sbi;
int fatal = 0, err;
if (atomic_read(&inode->i_count) > 1) {
printk ("ext3_free_inode: inode has count=%d\n",
atomic_read(&inode->i_count));
return;
}
if (inode->i_nlink) {
printk ("ext3_free_inode: inode has nlink=%d\n",
inode->i_nlink);
return;
}
if (!sb) {
printk("ext3_free_inode: inode on nonexistent device\n");
return;
}
sbi = EXT3_SB(sb);
ino = inode->i_ino;
ext3_debug ("freeing inode %lu\n", ino);
/*
* Note: we must free any quota before locking the superblock,
* as writing the quota to disk may need the lock as well.
*/
DQUOT_INIT(inode);
ext3_xattr_delete_inode(handle, inode);
DQUOT_FREE_INODE(inode);
DQUOT_DROP(inode);
is_directory = S_ISDIR(inode->i_mode);
/* Do this BEFORE marking the inode not in use or returning an error */
clear_inode (inode);
es = EXT3_SB(sb)->s_es;
if (ino < EXT3_FIRST_INO(sb) || ino > le32_to_cpu(es->s_inodes_count)) {
ext3_error (sb, "ext3_free_inode",
"reserved or nonexistent inode %lu", ino);
goto error_return;
}
block_group = (ino - 1) / EXT3_INODES_PER_GROUP(sb);
bit = (ino - 1) % EXT3_INODES_PER_GROUP(sb);
bitmap_bh = read_inode_bitmap(sb, block_group);
if (!bitmap_bh)
goto error_return;
BUFFER_TRACE(bitmap_bh, "get_write_access");
fatal = ext3_journal_get_write_access(handle, bitmap_bh);
if (fatal)
goto error_return;
/* Ok, now we can actually update the inode bitmaps.. */
if (!ext3_clear_bit_atomic(sb_bgl_lock(sbi, block_group),
bit, bitmap_bh->b_data))
ext3_error (sb, "ext3_free_inode",
"bit already cleared for inode %lu", ino);
else {
gdp = ext3_get_group_desc (sb, block_group, &bh2);
BUFFER_TRACE(bh2, "get_write_access");
fatal = ext3_journal_get_write_access(handle, bh2);
if (fatal) goto error_return;
if (gdp) {
spin_lock(sb_bgl_lock(sbi, block_group));
gdp->bg_free_inodes_count = cpu_to_le16(
le16_to_cpu(gdp->bg_free_inodes_count) + 1);
if (is_directory)
gdp->bg_used_dirs_count = cpu_to_le16(
le16_to_cpu(gdp->bg_used_dirs_count) - 1);
spin_unlock(sb_bgl_lock(sbi, block_group));
percpu_counter_inc(&sbi->s_freeinodes_counter);
if (is_directory)
percpu_counter_dec(&sbi->s_dirs_counter);
}
BUFFER_TRACE(bh2, "call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle, bh2);
if (!fatal) fatal = err;
}
BUFFER_TRACE(bitmap_bh, "call ext3_journal_dirty_metadata");
err = ext3_journal_dirty_metadata(handle, bitmap_bh);
if (!fatal)
fatal = err;
sb->s_dirt = 1;
error_return:
brelse(bitmap_bh);
ext3_std_error(sb, fatal);
}
/*
* Set various file attributes.
* N.B. After this call fhp needs an fh_put
*/
__be32
nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap,
int check_guard, time_t guardtime)
{
struct dentry *dentry;
struct inode *inode;
int accmode = NFSD_MAY_SATTR;
int ftype = 0;
__be32 err;
int host_err;
int size_change = 0;
if (iap->ia_valid & (ATTR_ATIME | ATTR_MTIME | ATTR_SIZE))
accmode |= NFSD_MAY_WRITE|NFSD_MAY_OWNER_OVERRIDE;
if (iap->ia_valid & ATTR_SIZE)
ftype = S_IFREG;
/* Get inode */
err = fh_verify(rqstp, fhp, ftype, accmode);
if (err)
goto out;
dentry = fhp->fh_dentry;
inode = dentry->d_inode;
/* Ignore any mode updates on symlinks */
if (S_ISLNK(inode->i_mode))
iap->ia_valid &= ~ATTR_MODE;
if (!iap->ia_valid)
goto out;
/*
* NFSv2 does not differentiate between "set-[ac]time-to-now"
* which only requires access, and "set-[ac]time-to-X" which
* requires ownership.
* So if it looks like it might be "set both to the same time which
* is close to now", and if inode_change_ok fails, then we
* convert to "set to now" instead of "set to explicit time"
*
* We only call inode_change_ok as the last test as technically
* it is not an interface that we should be using. It is only
* valid if the filesystem does not define it's own i_op->setattr.
*/
#define BOTH_TIME_SET (ATTR_ATIME_SET | ATTR_MTIME_SET)
#define MAX_TOUCH_TIME_ERROR (30*60)
if ((iap->ia_valid & BOTH_TIME_SET) == BOTH_TIME_SET &&
iap->ia_mtime.tv_sec == iap->ia_atime.tv_sec) {
/*
* Looks probable.
*
* Now just make sure time is in the right ballpark.
* Solaris, at least, doesn't seem to care what the time
* request is. We require it be within 30 minutes of now.
*/
time_t delta = iap->ia_atime.tv_sec - get_seconds();
if (delta < 0)
delta = -delta;
if (delta < MAX_TOUCH_TIME_ERROR &&
inode_change_ok(inode, iap) != 0) {
/*
* Turn off ATTR_[AM]TIME_SET but leave ATTR_[AM]TIME.
* This will cause notify_change to set these times
* to "now"
*/
iap->ia_valid &= ~BOTH_TIME_SET;
}
}
/*
* The size case is special.
* It changes the file as well as the attributes.
*/
if (iap->ia_valid & ATTR_SIZE) {
if (iap->ia_size < inode->i_size) {
err = nfsd_permission(rqstp, fhp->fh_export, dentry,
NFSD_MAY_TRUNC|NFSD_MAY_OWNER_OVERRIDE);
if (err)
goto out;
}
/*
* If we are changing the size of the file, then
* we need to break all leases.
*/
host_err = break_lease(inode, FMODE_WRITE | O_NONBLOCK);
if (host_err == -EWOULDBLOCK)
host_err = -ETIMEDOUT;
if (host_err) /* ENOMEM or EWOULDBLOCK */
goto out_nfserr;
host_err = get_write_access(inode);
if (host_err)
goto out_nfserr;
size_change = 1;
host_err = locks_verify_truncate(inode, NULL, iap->ia_size);
if (host_err) {
put_write_access(inode);
goto out_nfserr;
}
DQUOT_INIT(inode);
}
/* sanitize the mode change */
if (iap->ia_valid & ATTR_MODE) {
iap->ia_mode &= S_IALLUGO;
iap->ia_mode |= (inode->i_mode & ~S_IALLUGO);
}
/* Revoke setuid/setgid on chown */
if (((iap->ia_valid & ATTR_UID) && iap->ia_uid != inode->i_uid) ||
((iap->ia_valid & ATTR_GID) && iap->ia_gid != inode->i_gid)) {
iap->ia_valid |= ATTR_KILL_PRIV;
if (iap->ia_valid & ATTR_MODE) {
/* we're setting mode too, just clear the s*id bits */
iap->ia_mode &= ~S_ISUID;
if (iap->ia_mode & S_IXGRP)
iap->ia_mode &= ~S_ISGID;
} else {
/* set ATTR_KILL_* bits and let VFS handle it */
iap->ia_valid |= (ATTR_KILL_SUID | ATTR_KILL_SGID);
}
}
/* Change the attributes. */
iap->ia_valid |= ATTR_CTIME;
err = nfserr_notsync;
if (!check_guard || guardtime == inode->i_ctime.tv_sec) {
fh_lock(fhp);
host_err = notify_change(dentry, iap);
err = nfserrno(host_err);
fh_unlock(fhp);
}
if (size_change)
put_write_access(inode);
if (!err)
if (EX_ISSYNC(fhp->fh_export))
write_inode_now(inode, 1);
out:
return err;
out_nfserr:
err = nfserrno(host_err);
goto out;
}
/*
* NAME: jfs_rmdir(dip, dentry)
*
* FUNCTION: remove a link to child directory
*
* PARAMETER: dip - parent inode
* dentry - child directory dentry
*
* RETURN: -EINVAL - if name is . or ..
* -EINVAL - if . or .. exist but are invalid.
* errors from subroutines
*
* note:
* if other threads have the directory open when the last link
* is removed, the "." and ".." entries, if present, are removed before
* rmdir() returns and no new entries may be created in the directory,
* but the directory is not removed until the last reference to
* the directory is released (cf.unlink() of regular file).
*/
static int jfs_rmdir(struct inode *dip, struct dentry *dentry)
{
int rc;
tid_t tid; /* transaction id */
struct inode *ip = dentry->d_inode;
ino_t ino;
struct component_name dname;
struct inode *iplist[2];
struct tblock *tblk;
jfs_info("jfs_rmdir: dip:0x%p name:%s", dip, dentry->d_name.name);
/* Init inode for quota operations. */
DQUOT_INIT(ip);
/* directory must be empty to be removed */
if (!dtEmpty(ip)) {
rc = -ENOTEMPTY;
goto out;
}
if ((rc = get_UCSname(&dname, dentry))) {
goto out;
}
tid = txBegin(dip->i_sb, 0);
mutex_lock_nested(&JFS_IP(dip)->commit_mutex, COMMIT_MUTEX_PARENT);
mutex_lock_nested(&JFS_IP(ip)->commit_mutex, COMMIT_MUTEX_CHILD);
iplist[0] = dip;
iplist[1] = ip;
tblk = tid_to_tblock(tid);
tblk->xflag |= COMMIT_DELETE;
tblk->u.ip = ip;
/*
* delete the entry of target directory from parent directory
*/
ino = ip->i_ino;
if ((rc = dtDelete(tid, dip, &dname, &ino, JFS_REMOVE))) {
jfs_err("jfs_rmdir: dtDelete returned %d", rc);
if (rc == -EIO)
txAbort(tid, 1);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
goto out2;
}
/* update parent directory's link count corresponding
* to ".." entry of the target directory deleted
*/
dip->i_ctime = dip->i_mtime = CURRENT_TIME;
inode_dec_link_count(dip);
/*
* OS/2 could have created EA and/or ACL
*/
/* free EA from both persistent and working map */
if (JFS_IP(ip)->ea.flag & DXD_EXTENT) {
/* free EA pages */
txEA(tid, ip, &JFS_IP(ip)->ea, NULL);
}
JFS_IP(ip)->ea.flag = 0;
/* free ACL from both persistent and working map */
if (JFS_IP(ip)->acl.flag & DXD_EXTENT) {
/* free ACL pages */
txEA(tid, ip, &JFS_IP(ip)->acl, NULL);
}
JFS_IP(ip)->acl.flag = 0;
/* mark the target directory as deleted */
clear_nlink(ip);
mark_inode_dirty(ip);
rc = txCommit(tid, 2, &iplist[0], 0);
txEnd(tid);
mutex_unlock(&JFS_IP(ip)->commit_mutex);
mutex_unlock(&JFS_IP(dip)->commit_mutex);
/*
* Truncating the directory index table is not guaranteed. It
* may need to be done iteratively
*/
if (test_cflag(COMMIT_Stale, dip)) {
if (dip->i_size > 1)
jfs_truncate_nolock(dip, 0);
clear_cflag(COMMIT_Stale, dip);
}
out2:
free_UCSname(&dname);
out:
jfs_info("jfs_rmdir: rc:%d", rc);
return rc;
}
