*
* CDDL HEADER END
*/
+
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2013, 2016 by Delphix. All rights reserved.
+ * Copyright 2017 Nexenta Systems, Inc.
*/
#include <sys/types.h>
#include <sys/policy.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
+#include <sys/zfs_vnops.h>
#include <sys/fs/zfs.h>
#include "fs/fs_subr.h"
#include <sys/zap.h>
* of names after deciding which is the appropriate lookup interface.
*/
static int
-zfs_match_find(zfsvfs_t *zfsvfs, znode_t *dzp, char *name, boolean_t exact,
+zfs_match_find(zfsvfs_t *zfsvfs, znode_t *dzp, char *name, matchtype_t mt,
boolean_t update, int *deflags, pathname_t *rpnp, uint64_t *zoid)
{
+ boolean_t conflict = B_FALSE;
int error;
if (zfsvfs->z_norm) {
- matchtype_t mt = MT_FIRST;
- boolean_t conflict = B_FALSE;
size_t bufsz = 0;
char *buf = NULL;
buf = rpnp->pn_buf;
bufsz = rpnp->pn_bufsize;
}
- if (exact)
- mt = MT_EXACT;
+
/*
* In the non-mixed case we only expect there would ever
* be one match, but we need to use the normalizing lookup.
*/
error = zap_lookup_norm(zfsvfs->z_os, dzp->z_id, name, 8, 1,
zoid, mt, buf, bufsz, &conflict);
- if (!error && deflags)
- *deflags = conflict ? ED_CASE_CONFLICT : 0;
} else {
error = zap_lookup(zfsvfs->z_os, dzp->z_id, name, 8, 1, zoid);
}
+
+ /*
+ * Allow multiple entries provided the first entry is
+ * the object id. Non-zpl consumers may safely make
+ * use of the additional space.
+ *
+ * XXX: This should be a feature flag for compatibility
+ */
+ if (error == EOVERFLOW)
+ error = 0;
+
+ if (zfsvfs->z_norm && !error && deflags)
+ *deflags = conflict ? ED_CASE_CONFLICT : 0;
+
*zoid = ZFS_DIRENT_OBJ(*zoid);
+#ifdef HAVE_DNLC
if (error == ENOENT && update)
- dnlc_update(ZTOV(dzp), name, DNLC_NO_VNODE);
+ dnlc_update(ZTOI(dzp), name, DNLC_NO_VNODE);
+#endif /* HAVE_DNLC */
return (error);
}
zfs_dirent_lock(zfs_dirlock_t **dlpp, znode_t *dzp, char *name, znode_t **zpp,
int flag, int *direntflags, pathname_t *realpnp)
{
- zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zfs_dirlock_t *dl;
boolean_t update;
- boolean_t exact;
+ matchtype_t mt = 0;
uint64_t zoid;
+#ifdef HAVE_DNLC
vnode_t *vp = NULL;
+#endif /* HAVE_DNLC */
int error = 0;
int cmpflags;
/*
* Verify that we are not trying to lock '.', '..', or '.zfs'
*/
- if (name[0] == '.' &&
- (name[1] == '\0' || (name[1] == '.' && name[2] == '\0')) ||
- zfs_has_ctldir(dzp) && strcmp(name, ZFS_CTLDIR_NAME) == 0)
- return (EEXIST);
+ if ((name[0] == '.' &&
+ (name[1] == '\0' || (name[1] == '.' && name[2] == '\0'))) ||
+ (zfs_has_ctldir(dzp) && strcmp(name, ZFS_CTLDIR_NAME) == 0))
+ return (SET_ERROR(EEXIST));
/*
* Case sensitivity and normalization preferences are set when
*/
/*
- * Decide if exact matches should be requested when performing
- * a zap lookup on file systems supporting case-insensitive
- * access.
+ * When matching we may need to normalize & change case according to
+ * FS settings.
+ *
+ * Note that a normalized match is necessary for a case insensitive
+ * filesystem when the lookup request is not exact because normalization
+ * can fold case independent of normalizing code point sequences.
+ *
+ * See the table above zfs_dropname().
*/
- exact =
- ((zfsvfs->z_case == ZFS_CASE_INSENSITIVE) && (flag & ZCIEXACT)) ||
- ((zfsvfs->z_case == ZFS_CASE_MIXED) && !(flag & ZCILOOK));
+ if (zfsvfs->z_norm != 0) {
+ mt = MT_NORMALIZE;
+
+ /*
+ * Determine if the match needs to honor the case specified in
+ * lookup, and if so keep track of that so that during
+ * normalization we don't fold case.
+ */
+ if ((zfsvfs->z_case == ZFS_CASE_INSENSITIVE &&
+ (flag & ZCIEXACT)) ||
+ (zfsvfs->z_case == ZFS_CASE_MIXED && !(flag & ZCILOOK))) {
+ mt |= MT_MATCH_CASE;
+ }
+ }
/*
* Only look in or update the DNLC if we are looking for the
* case for performance improvement?
*/
update = !zfsvfs->z_norm ||
- ((zfsvfs->z_case == ZFS_CASE_MIXED) &&
+ (zfsvfs->z_case == ZFS_CASE_MIXED &&
!(zfsvfs->z_norm & ~U8_TEXTPREP_TOUPPER) && !(flag & ZCILOOK));
/*
mutex_enter(&dzp->z_lock);
for (;;) {
- if (dzp->z_unlinked) {
+ if (dzp->z_unlinked && !(flag & ZXATTR)) {
mutex_exit(&dzp->z_lock);
if (!(flag & ZHAVELOCK))
rw_exit(&dzp->z_name_lock);
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
}
for (dl = dzp->z_dirlocks; dl != NULL; dl = dl->dl_next) {
if ((u8_strcmp(name, dl->dl_name, 0, cmpflags,
mutex_exit(&dzp->z_lock);
if (!(flag & ZHAVELOCK))
rw_exit(&dzp->z_name_lock);
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
}
if (dl == NULL) {
/*
* dl_name in case the first thread goes away before we do.
* Note that we initialize the new name before storing its
* pointer into dl_name, because the first thread may load
- * dl->dl_name at any time. He'll either see the old value,
- * which is his, or the new shared copy; either is OK.
+ * dl->dl_name at any time. It'll either see the old value,
+ * which belongs to it, or the new shared copy; either is OK.
*/
dl->dl_namesize = strlen(dl->dl_name) + 1;
name = kmem_alloc(dl->dl_namesize, KM_SLEEP);
error = sa_lookup(dzp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &zoid,
sizeof (zoid));
if (error == 0)
- error = (zoid == 0 ? ENOENT : 0);
+ error = (zoid == 0 ? SET_ERROR(ENOENT) : 0);
} else {
+#ifdef HAVE_DNLC
if (update)
- vp = dnlc_lookup(ZTOV(dzp), name);
+ vp = dnlc_lookup(ZTOI(dzp), name);
if (vp == DNLC_NO_VNODE) {
- VN_RELE(vp);
- error = ENOENT;
+ iput(vp);
+ error = SET_ERROR(ENOENT);
} else if (vp) {
if (flag & ZNEW) {
zfs_dirent_unlock(dl);
- VN_RELE(vp);
- return (EEXIST);
+ iput(vp);
+ return (SET_ERROR(EEXIST));
}
*dlpp = dl;
*zpp = VTOZ(vp);
return (0);
} else {
- error = zfs_match_find(zfsvfs, dzp, name, exact,
+ error = zfs_match_find(zfsvfs, dzp, name, mt,
update, direntflags, realpnp, &zoid);
}
+#else
+ error = zfs_match_find(zfsvfs, dzp, name, mt,
+ update, direntflags, realpnp, &zoid);
+#endif /* HAVE_DNLC */
}
if (error) {
if (error != ENOENT || (flag & ZEXISTS)) {
} else {
if (flag & ZNEW) {
zfs_dirent_unlock(dl);
- return (EEXIST);
+ return (SET_ERROR(EEXIST));
}
error = zfs_zget(zfsvfs, zoid, zpp);
if (error) {
zfs_dirent_unlock(dl);
return (error);
}
+#ifdef HAVE_DNLC
if (!(flag & ZXATTR) && update)
- dnlc_update(ZTOV(dzp), name, ZTOV(*zpp));
+ dnlc_update(ZTOI(dzp), name, ZTOI(*zpp));
+#endif /* HAVE_DNLC */
}
*dlpp = dl;
* special pseudo-directory.
*/
int
-zfs_dirlook(znode_t *dzp, char *name, vnode_t **vpp, int flags,
+zfs_dirlook(znode_t *dzp, char *name, struct inode **ipp, int flags,
int *deflg, pathname_t *rpnp)
{
zfs_dirlock_t *dl;
uint64_t parent;
if (name[0] == 0 || (name[0] == '.' && name[1] == 0)) {
- *vpp = ZTOV(dzp);
- VN_HOLD(*vpp);
+ *ipp = ZTOI(dzp);
+ igrab(*ipp);
} else if (name[0] == '.' && name[1] == '.' && name[2] == 0) {
- zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(dzp);
/*
* If we are a snapshot mounted under .zfs, return
- * the vp for the snapshot directory.
+ * the inode pointer for the snapshot directory.
*/
if ((error = sa_lookup(dzp->z_sa_hdl,
SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent))) != 0)
return (error);
+
if (parent == dzp->z_id && zfsvfs->z_parent != zfsvfs) {
error = zfsctl_root_lookup(zfsvfs->z_parent->z_ctldir,
- "snapshot", vpp, NULL, 0, NULL, kcred,
- NULL, NULL, NULL);
+ "snapshot", ipp, 0, kcred, NULL, NULL);
return (error);
}
rw_enter(&dzp->z_parent_lock, RW_READER);
error = zfs_zget(zfsvfs, parent, &zp);
if (error == 0)
- *vpp = ZTOV(zp);
+ *ipp = ZTOI(zp);
rw_exit(&dzp->z_parent_lock);
} else if (zfs_has_ctldir(dzp) && strcmp(name, ZFS_CTLDIR_NAME) == 0) {
- *vpp = zfsctl_root(dzp);
+ *ipp = zfsctl_root(dzp);
} else {
int zf;
error = zfs_dirent_lock(&dl, dzp, name, &zp, zf, deflg, rpnp);
if (error == 0) {
- *vpp = ZTOV(zp);
+ *ipp = ZTOI(zp);
zfs_dirent_unlock(dl);
dzp->z_zn_prefetch = B_TRUE; /* enable prefetching */
}
void
zfs_unlinked_add(znode_t *zp, dmu_tx_t *tx)
{
- zfsvfs_t *zfsvfs = zp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(zp);
ASSERT(zp->z_unlinked);
- ASSERT(zp->z_links == 0);
+ ASSERT(ZTOI(zp)->i_nlink == 0);
VERIFY3U(0, ==,
zap_add_int(zfsvfs->z_os, zfsvfs->z_unlinkedobj, zp->z_id, tx));
int error;
/*
- * Interate over the contents of the unlinked set.
+ * Iterate over the contents of the unlinked set.
*/
for (zap_cursor_init(&zc, zfsvfs->z_os, zfsvfs->z_unlinkedobj);
zap_cursor_retrieve(&zc, &zap) == 0;
continue;
zp->z_unlinked = B_TRUE;
- VN_RELE(ZTOV(zp));
+ iput(ZTOI(zp));
}
zap_cursor_fini(&zc);
}
zap_attribute_t zap;
znode_t *xzp;
dmu_tx_t *tx;
- zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(dzp);
zfs_dirlock_t dl;
int skipped = 0;
int error;
continue;
}
- ASSERT((ZTOV(xzp)->v_type == VREG) ||
- (ZTOV(xzp)->v_type == VLNK));
+ ASSERT(S_ISREG(ZTOI(xzp)->i_mode) ||
+ S_ISLNK(ZTOI(xzp)->i_mode));
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE);
dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL);
/* Is this really needed ? */
zfs_sa_upgrade_txholds(tx, xzp);
+ dmu_tx_mark_netfree(tx);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
- VN_RELE(ZTOV(xzp));
+ zfs_iput_async(ZTOI(xzp));
skipped += 1;
continue;
}
skipped += 1;
dmu_tx_commit(tx);
- VN_RELE(ZTOV(xzp));
+ zfs_iput_async(ZTOI(xzp));
}
zap_cursor_fini(&zc);
if (error != ENOENT)
void
zfs_rmnode(znode_t *zp)
{
- zfsvfs_t *zfsvfs = zp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(zp);
objset_t *os = zfsvfs->z_os;
znode_t *xzp = NULL;
dmu_tx_t *tx;
uint64_t acl_obj;
uint64_t xattr_obj;
+ uint64_t links;
int error;
- ASSERT(zp->z_links == 0);
- ASSERT(ZTOV(zp)->v_count == 0);
+ ASSERT(ZTOI(zp)->i_nlink == 0);
+ ASSERT(atomic_read(&ZTOI(zp)->i_count) == 0);
/*
* If this is an attribute directory, purge its contents.
*/
- if (ZTOV(zp)->v_type == VDIR && (zp->z_pflags & ZFS_XATTR)) {
+ if (S_ISDIR(ZTOI(zp)->i_mode) && (zp->z_pflags & ZFS_XATTR)) {
if (zfs_purgedir(zp) != 0) {
/*
* Not enough space to delete some xattrs.
* Leave it in the unlinked set.
*/
zfs_znode_dmu_fini(zp);
- zfs_znode_free(zp);
+
return;
}
}
/*
- * Free up all the data in the file.
+ * Free up all the data in the file. We don't do this for directories
+ * because we need truncate and remove to be in the same tx, like in
+ * zfs_znode_delete(). Otherwise, if we crash here we'll end up with
+ * an inconsistent truncated zap object in the delete queue. Note a
+ * truncated file is harmless since it only contains user data.
*/
- error = dmu_free_long_range(os, zp->z_id, 0, DMU_OBJECT_END);
- if (error) {
- /*
- * Not enough space. Leave the file in the unlinked set.
- */
- zfs_znode_dmu_fini(zp);
- zfs_znode_free(zp);
- return;
+ if (S_ISREG(ZTOI(zp)->i_mode)) {
+ error = dmu_free_long_range(os, zp->z_id, 0, DMU_OBJECT_END);
+ if (error) {
+ /*
+ * Not enough space or we were interrupted by unmount.
+ * Leave the file in the unlinked set.
+ */
+ zfs_znode_dmu_fini(zp);
+ return;
+ }
}
/*
*/
dmu_tx_abort(tx);
zfs_znode_dmu_fini(zp);
- zfs_znode_free(zp);
goto out;
}
ASSERT(error == 0);
mutex_enter(&xzp->z_lock);
xzp->z_unlinked = B_TRUE; /* mark xzp for deletion */
- xzp->z_links = 0; /* no more links to it */
+ clear_nlink(ZTOI(xzp)); /* no more links to it */
+ links = 0;
VERIFY(0 == sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs),
- &xzp->z_links, sizeof (xzp->z_links), tx));
+ &links, sizeof (links), tx));
mutex_exit(&xzp->z_lock);
zfs_unlinked_add(xzp, tx);
}
dmu_tx_commit(tx);
out:
if (xzp)
- VN_RELE(ZTOV(xzp));
+ zfs_iput_async(ZTOI(xzp));
}
static uint64_t
{
uint64_t de = zp->z_id;
- if (zp->z_zfsvfs->z_version >= ZPL_VERSION_DIRENT_TYPE)
+ if (ZTOZSB(zp)->z_version >= ZPL_VERSION_DIRENT_TYPE)
de |= IFTODT(mode) << 60;
return (de);
}
/*
- * Link zp into dl. Can only fail if zp has been unlinked.
+ * Link zp into dl. Can fail in the following cases :
+ * - if zp has been unlinked.
+ * - if the number of entries with the same hash (aka. colliding entries)
+ * exceed the capacity of a leaf-block of fatzap and splitting of the
+ * leaf-block does not help.
*/
int
zfs_link_create(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag)
{
znode_t *dzp = dl->dl_dzp;
- zfsvfs_t *zfsvfs = zp->z_zfsvfs;
- vnode_t *vp = ZTOV(zp);
+ zfsvfs_t *zfsvfs = ZTOZSB(zp);
uint64_t value;
- int zp_is_dir = (vp->v_type == VDIR);
+ int zp_is_dir = S_ISDIR(ZTOI(zp)->i_mode);
sa_bulk_attr_t bulk[5];
uint64_t mtime[2], ctime[2];
+ uint64_t links;
int count = 0;
int error;
if (zp->z_unlinked) { /* no new links to unlinked zp */
ASSERT(!(flag & (ZNEW | ZEXISTS)));
mutex_exit(&zp->z_lock);
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
}
- zp->z_links++;
- SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL,
- &zp->z_links, sizeof (zp->z_links));
+ if (!(flag & ZNEW)) {
+ /*
+ * ZNEW nodes come from zfs_mknode() where the link
+ * count has already been initialised
+ */
+ inc_nlink(ZTOI(zp));
+ links = ZTOI(zp)->i_nlink;
+ SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
+ NULL, &links, sizeof (links));
+ }
+ }
+ value = zfs_dirent(zp, zp->z_mode);
+ error = zap_add(ZTOZSB(zp)->z_os, dzp->z_id, dl->dl_name, 8, 1,
+ &value, tx);
+
+ /*
+ * zap_add could fail to add the entry if it exceeds the capacity of the
+ * leaf-block and zap_leaf_split() failed to help.
+ * The caller of this routine is responsible for failing the transaction
+ * which will rollback the SA updates done above.
+ */
+ if (error != 0) {
+ if (!(flag & ZRENAMING) && !(flag & ZNEW))
+ drop_nlink(ZTOI(zp));
+ mutex_exit(&zp->z_lock);
+ return (error);
}
+
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PARENT(zfsvfs), NULL,
&dzp->z_id, sizeof (dzp->z_id));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
ctime, sizeof (ctime));
zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime,
- ctime, B_TRUE);
+ ctime);
}
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
mutex_enter(&dzp->z_lock);
dzp->z_size++;
- dzp->z_links += zp_is_dir;
+ if (zp_is_dir)
+ inc_nlink(ZTOI(dzp));
+ links = ZTOI(dzp)->i_nlink;
count = 0;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL,
&dzp->z_size, sizeof (dzp->z_size));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL,
- &dzp->z_links, sizeof (dzp->z_links));
+ &links, sizeof (links));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL,
mtime, sizeof (mtime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL,
ctime, sizeof (ctime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL,
&dzp->z_pflags, sizeof (dzp->z_pflags));
- zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime, B_TRUE);
+ zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(dzp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
mutex_exit(&dzp->z_lock);
- value = zfs_dirent(zp, zp->z_mode);
- error = zap_add(zp->z_zfsvfs->z_os, dzp->z_id, dl->dl_name,
- 8, 1, &value, tx);
- ASSERT(error == 0);
-
- dnlc_update(ZTOV(dzp), dl->dl_name, vp);
-
return (0);
}
+/*
+ * The match type in the code for this function should conform to:
+ *
+ * ------------------------------------------------------------------------
+ * fs type | z_norm | lookup type | match type
+ * ---------|-------------|-------------|----------------------------------
+ * CS !norm | 0 | 0 | 0 (exact)
+ * CS norm | formX | 0 | MT_NORMALIZE
+ * CI !norm | upper | !ZCIEXACT | MT_NORMALIZE
+ * CI !norm | upper | ZCIEXACT | MT_NORMALIZE | MT_MATCH_CASE
+ * CI norm | upper|formX | !ZCIEXACT | MT_NORMALIZE
+ * CI norm | upper|formX | ZCIEXACT | MT_NORMALIZE | MT_MATCH_CASE
+ * CM !norm | upper | !ZCILOOK | MT_NORMALIZE | MT_MATCH_CASE
+ * CM !norm | upper | ZCILOOK | MT_NORMALIZE
+ * CM norm | upper|formX | !ZCILOOK | MT_NORMALIZE | MT_MATCH_CASE
+ * CM norm | upper|formX | ZCILOOK | MT_NORMALIZE
+ *
+ * Abbreviations:
+ * CS = Case Sensitive, CI = Case Insensitive, CM = Case Mixed
+ * upper = case folding set by fs type on creation (U8_TEXTPREP_TOUPPER)
+ * formX = unicode normalization form set on fs creation
+ */
static int
zfs_dropname(zfs_dirlock_t *dl, znode_t *zp, znode_t *dzp, dmu_tx_t *tx,
int flag)
{
int error;
- if (zp->z_zfsvfs->z_norm) {
- if (((zp->z_zfsvfs->z_case == ZFS_CASE_INSENSITIVE) &&
+ if (ZTOZSB(zp)->z_norm) {
+ matchtype_t mt = MT_NORMALIZE;
+
+ if ((ZTOZSB(zp)->z_case == ZFS_CASE_INSENSITIVE &&
(flag & ZCIEXACT)) ||
- ((zp->z_zfsvfs->z_case == ZFS_CASE_MIXED) &&
- !(flag & ZCILOOK)))
- error = zap_remove_norm(zp->z_zfsvfs->z_os,
- dzp->z_id, dl->dl_name, MT_EXACT, tx);
- else
- error = zap_remove_norm(zp->z_zfsvfs->z_os,
- dzp->z_id, dl->dl_name, MT_FIRST, tx);
+ (ZTOZSB(zp)->z_case == ZFS_CASE_MIXED &&
+ !(flag & ZCILOOK))) {
+ mt |= MT_MATCH_CASE;
+ }
+
+ error = zap_remove_norm(ZTOZSB(zp)->z_os, dzp->z_id,
+ dl->dl_name, mt, tx);
} else {
- error = zap_remove(zp->z_zfsvfs->z_os,
- dzp->z_id, dl->dl_name, tx);
+ error = zap_remove(ZTOZSB(zp)->z_os, dzp->z_id, dl->dl_name,
+ tx);
}
return (error);
}
/*
- * Unlink zp from dl, and mark zp for deletion if this was the last link.
- * Can fail if zp is a mount point (EBUSY) or a non-empty directory (EEXIST).
+ * Unlink zp from dl, and mark zp for deletion if this was the last link. Can
+ * fail if zp is a mount point (EBUSY) or a non-empty directory (ENOTEMPTY).
* If 'unlinkedp' is NULL, we put unlinked znodes on the unlinked list.
* If it's non-NULL, we use it to indicate whether the znode needs deletion,
* and it's the caller's job to do it.
*/
int
zfs_link_destroy(zfs_dirlock_t *dl, znode_t *zp, dmu_tx_t *tx, int flag,
- boolean_t *unlinkedp)
+ boolean_t *unlinkedp)
{
znode_t *dzp = dl->dl_dzp;
- zfsvfs_t *zfsvfs = dzp->z_zfsvfs;
- vnode_t *vp = ZTOV(zp);
- int zp_is_dir = (vp->v_type == VDIR);
+ zfsvfs_t *zfsvfs = ZTOZSB(dzp);
+ int zp_is_dir = S_ISDIR(ZTOI(zp)->i_mode);
boolean_t unlinked = B_FALSE;
sa_bulk_attr_t bulk[5];
uint64_t mtime[2], ctime[2];
+ uint64_t links;
int count = 0;
int error;
- dnlc_remove(ZTOV(dzp), dl->dl_name);
+#ifdef HAVE_DNLC
+ dnlc_remove(ZTOI(dzp), dl->dl_name);
+#endif /* HAVE_DNLC */
if (!(flag & ZRENAMING)) {
- if (vn_vfswlock(vp)) /* prevent new mounts on zp */
- return (EBUSY);
-
- if (vn_ismntpt(vp)) { /* don't remove mount point */
- vn_vfsunlock(vp);
- return (EBUSY);
- }
-
mutex_enter(&zp->z_lock);
if (zp_is_dir && !zfs_dirempty(zp)) {
mutex_exit(&zp->z_lock);
- vn_vfsunlock(vp);
- return (EEXIST);
+ return (SET_ERROR(ENOTEMPTY));
}
/*
error = zfs_dropname(dl, zp, dzp, tx, flag);
if (error != 0) {
mutex_exit(&zp->z_lock);
- vn_vfsunlock(vp);
return (error);
}
- if (zp->z_links <= zp_is_dir) {
- zfs_panic_recover("zfs: link count on %s is %u, "
- "should be at least %u",
- zp->z_vnode->v_path ? zp->z_vnode->v_path :
- "<unknown>", (int)zp->z_links,
- zp_is_dir + 1);
- zp->z_links = zp_is_dir + 1;
+ if (ZTOI(zp)->i_nlink <= zp_is_dir) {
+ zfs_panic_recover("zfs: link count on %lu is %u, "
+ "should be at least %u", zp->z_id,
+ (int)ZTOI(zp)->i_nlink, zp_is_dir + 1);
+ set_nlink(ZTOI(zp), zp_is_dir + 1);
}
- if (--zp->z_links == zp_is_dir) {
+ drop_nlink(ZTOI(zp));
+ if (ZTOI(zp)->i_nlink == zp_is_dir) {
zp->z_unlinked = B_TRUE;
- zp->z_links = 0;
+ clear_nlink(ZTOI(zp));
unlinked = B_TRUE;
} else {
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs),
NULL, &ctime, sizeof (ctime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &zp->z_pflags, sizeof (zp->z_pflags));
- zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime, ctime,
- B_TRUE);
+ zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime,
+ ctime);
}
+ links = ZTOI(zp)->i_nlink;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
- NULL, &zp->z_links, sizeof (zp->z_links));
+ NULL, &links, sizeof (links));
error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx);
count = 0;
ASSERT(error == 0);
mutex_exit(&zp->z_lock);
- vn_vfsunlock(vp);
} else {
error = zfs_dropname(dl, zp, dzp, tx, flag);
if (error != 0)
mutex_enter(&dzp->z_lock);
dzp->z_size--; /* one dirent removed */
- dzp->z_links -= zp_is_dir; /* ".." link from zp */
+ if (zp_is_dir)
+ drop_nlink(ZTOI(dzp)); /* ".." link from zp */
+ links = ZTOI(dzp)->i_nlink;
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs),
- NULL, &dzp->z_links, sizeof (dzp->z_links));
+ NULL, &links, sizeof (links));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs),
NULL, &dzp->z_size, sizeof (dzp->z_size));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs),
NULL, mtime, sizeof (mtime));
SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs),
NULL, &dzp->z_pflags, sizeof (dzp->z_pflags));
- zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime, B_TRUE);
+ zfs_tstamp_update_setup(dzp, CONTENT_MODIFIED, mtime, ctime);
error = sa_bulk_update(dzp->z_sa_hdl, bulk, count, tx);
ASSERT(error == 0);
mutex_exit(&dzp->z_lock);
* Indicate whether the directory is empty. Works with or without z_lock
* held, but can only be consider a hint in the latter case. Returns true
* if only "." and ".." remain and there's no work in progress.
+ *
+ * The internal ZAP size, rather than zp->z_size, needs to be checked since
+ * some consumers (Lustre) do not strictly maintain an accurate SA_ZPL_SIZE.
*/
boolean_t
zfs_dirempty(znode_t *dzp)
{
- return (dzp->z_size == 2 && dzp->z_dirlocks == 0);
+ zfsvfs_t *zfsvfs = ZTOZSB(dzp);
+ uint64_t count;
+ int error;
+
+ if (dzp->z_dirlocks != NULL)
+ return (B_FALSE);
+
+ error = zap_count(zfsvfs->z_os, dzp->z_id, &count);
+ if (error != 0 || count != 0)
+ return (B_FALSE);
+
+ return (B_TRUE);
}
int
-zfs_make_xattrdir(znode_t *zp, vattr_t *vap, vnode_t **xvpp, cred_t *cr)
+zfs_make_xattrdir(znode_t *zp, vattr_t *vap, struct inode **xipp, cred_t *cr)
{
- zfsvfs_t *zfsvfs = zp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(zp);
znode_t *xzp;
dmu_tx_t *tx;
int error;
zfs_acl_ids_t acl_ids;
boolean_t fuid_dirtied;
+#ifdef DEBUG
uint64_t parent;
+#endif
- *xvpp = NULL;
+ *xipp = NULL;
- if (error = zfs_zaccess(zp, ACE_WRITE_NAMED_ATTRS, 0, B_FALSE, cr))
+ if ((error = zfs_zaccess(zp, ACE_WRITE_NAMED_ATTRS, 0, B_FALSE, cr)))
return (error);
if ((error = zfs_acl_ids_create(zp, IS_XATTR, vap, cr, NULL,
&acl_ids)) != 0)
return (error);
- if (zfs_acl_ids_overquota(zfsvfs, &acl_ids)) {
+ if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zp->z_projid)) {
zfs_acl_ids_free(&acl_ids);
- return (EDQUOT);
+ return (SET_ERROR(EDQUOT));
}
-top:
tx = dmu_tx_create(zfsvfs->z_os);
dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes +
ZFS_SA_BASE_ATTR_SIZE);
fuid_dirtied = zfsvfs->z_fuid_dirty;
if (fuid_dirtied)
zfs_fuid_txhold(zfsvfs, tx);
- error = dmu_tx_assign(tx, TXG_NOWAIT);
+ error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
- if (error == ERESTART) {
- dmu_tx_wait(tx);
- dmu_tx_abort(tx);
- goto top;
- }
zfs_acl_ids_free(&acl_ids);
dmu_tx_abort(tx);
return (error);
VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xzp->z_id,
sizeof (xzp->z_id), tx));
- (void) zfs_log_create(zfsvfs->z_log, tx, TX_MKXATTR, zp,
- xzp, "", NULL, acl_ids.z_fuidp, vap);
+ if (!zp->z_unlinked)
+ (void) zfs_log_create(zfsvfs->z_log, tx, TX_MKXATTR, zp,
+ xzp, "", NULL, acl_ids.z_fuidp, vap);
zfs_acl_ids_free(&acl_ids);
dmu_tx_commit(tx);
- *xvpp = ZTOV(xzp);
+ *xipp = ZTOI(xzp);
return (0);
}
* cr - credentials of caller
* flags - flags from the VOP_LOOKUP call
*
- * OUT: xzpp - pointer to extended attribute znode
+ * OUT: xipp - pointer to extended attribute znode
*
* RETURN: 0 on success
* error number on failure
*/
int
-zfs_get_xattrdir(znode_t *zp, vnode_t **xvpp, cred_t *cr, int flags)
+zfs_get_xattrdir(znode_t *zp, struct inode **xipp, cred_t *cr, int flags)
{
- zfsvfs_t *zfsvfs = zp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(zp);
znode_t *xzp;
zfs_dirlock_t *dl;
vattr_t va;
return (error);
if (xzp != NULL) {
- *xvpp = ZTOV(xzp);
+ *xipp = ZTOI(xzp);
zfs_dirent_unlock(dl);
return (0);
}
-
if (!(flags & CREATE_XATTR_DIR)) {
zfs_dirent_unlock(dl);
- return (ENOENT);
+ return (SET_ERROR(ENOENT));
}
- if (zfsvfs->z_vfs->vfs_flag & VFS_RDONLY) {
+ if (zfs_is_readonly(zfsvfs)) {
zfs_dirent_unlock(dl);
- return (EROFS);
+ return (SET_ERROR(EROFS));
}
/*
* Once in a directory the ability to read/write attributes
* is controlled by the permissions on the attribute file.
*/
- va.va_mask = AT_TYPE | AT_MODE | AT_UID | AT_GID;
- va.va_type = VDIR;
+ va.va_mask = ATTR_MODE | ATTR_UID | ATTR_GID;
va.va_mode = S_IFDIR | S_ISVTX | 0777;
zfs_fuid_map_ids(zp, cr, &va.va_uid, &va.va_gid);
- error = zfs_make_xattrdir(zp, &va, xvpp, cr);
+ va.va_dentry = NULL;
+ error = zfs_make_xattrdir(zp, &va, xipp, cr);
zfs_dirent_unlock(dl);
if (error == ERESTART) {
*
* you own the directory,
* you own the entry,
- * the entry is a plain file and you have write access,
+ * you have write access to the entry,
* or you are privileged (checked in secpolicy...).
*
* The function returns 0 if remove access is granted.
int
zfs_sticky_remove_access(znode_t *zdp, znode_t *zp, cred_t *cr)
{
- uid_t uid;
+ uid_t uid;
uid_t downer;
uid_t fowner;
- zfsvfs_t *zfsvfs = zdp->z_zfsvfs;
+ zfsvfs_t *zfsvfs = ZTOZSB(zdp);
- if (zdp->z_zfsvfs->z_replay)
+ if (zfsvfs->z_replay)
return (0);
if ((zdp->z_mode & S_ISVTX) == 0)
return (0);
- downer = zfs_fuid_map_id(zfsvfs, zdp->z_uid, cr, ZFS_OWNER);
- fowner = zfs_fuid_map_id(zfsvfs, zp->z_uid, cr, ZFS_OWNER);
+ downer = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(ZTOI(zdp)->i_uid),
+ cr, ZFS_OWNER);
+ fowner = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(ZTOI(zp)->i_uid),
+ cr, ZFS_OWNER);
if ((uid = crgetuid(cr)) == downer || uid == fowner ||
- (ZTOV(zp)->v_type == VREG &&
- zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr) == 0))
+ zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr) == 0)
return (0);
else
return (secpolicy_vnode_remove(cr));