Index: head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ctldir.c =================================================================== --- head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ctldir.c (revision 301274) +++ head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ctldir.c (revision 301275) @@ -1,1770 +1,1762 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2013 by Delphix. All rights reserved. * Copyright 2015, OmniTI Computer Consulting, Inc. All rights reserved. */ /* * ZFS control directory (a.k.a. ".zfs") * * This directory provides a common location for all ZFS meta-objects. * Currently, this is only the 'snapshot' directory, but this may expand in the * future. The elements are built using the GFS primitives, as the hierarchy * does not actually exist on disk. * * For 'snapshot', we don't want to have all snapshots always mounted, because * this would take up a huge amount of space in /etc/mnttab. We have three * types of objects: * * ctldir ------> snapshotdir -------> snapshot * | * | * V * mounted fs * * The 'snapshot' node contains just enough information to lookup '..' and act * as a mountpoint for the snapshot. Whenever we lookup a specific snapshot, we * perform an automount of the underlying filesystem and return the * corresponding vnode. * * All mounts are handled automatically by the kernel, but unmounts are * (currently) handled from user land. The main reason is that there is no * reliable way to auto-unmount the filesystem when it's "no longer in use". * When the user unmounts a filesystem, we call zfsctl_unmount(), which * unmounts any snapshots within the snapshot directory. * * The '.zfs', '.zfs/snapshot', and all directories created under * '.zfs/snapshot' (ie: '.zfs/snapshot/') are all GFS nodes and * share the same vfs_t as the head filesystem (what '.zfs' lives under). * * File systems mounted ontop of the GFS nodes '.zfs/snapshot/' * (ie: snapshots) are ZFS nodes and have their own unique vfs_t. * However, vnodes within these mounted on file systems have their v_vfsp * fields set to the head filesystem to make NFS happy (see * zfsctl_snapdir_lookup()). We VFS_HOLD the head filesystem's vfs_t * so that it cannot be freed until all snapshots have been unmounted. */ #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_namecheck.h" typedef struct zfsctl_node { gfs_dir_t zc_gfs_private; uint64_t zc_id; timestruc_t zc_cmtime; /* ctime and mtime, always the same */ } zfsctl_node_t; typedef struct zfsctl_snapdir { zfsctl_node_t sd_node; kmutex_t sd_lock; avl_tree_t sd_snaps; } zfsctl_snapdir_t; typedef struct { char *se_name; vnode_t *se_root; avl_node_t se_node; } zfs_snapentry_t; static int snapentry_compare(const void *a, const void *b) { const zfs_snapentry_t *sa = a; const zfs_snapentry_t *sb = b; int ret = strcmp(sa->se_name, sb->se_name); if (ret < 0) return (-1); else if (ret > 0) return (1); else return (0); } #ifdef illumos vnodeops_t *zfsctl_ops_root; vnodeops_t *zfsctl_ops_snapdir; vnodeops_t *zfsctl_ops_snapshot; vnodeops_t *zfsctl_ops_shares; vnodeops_t *zfsctl_ops_shares_dir; static const fs_operation_def_t zfsctl_tops_root[]; static const fs_operation_def_t zfsctl_tops_snapdir[]; static const fs_operation_def_t zfsctl_tops_snapshot[]; static const fs_operation_def_t zfsctl_tops_shares[]; #else static struct vop_vector zfsctl_ops_root; static struct vop_vector zfsctl_ops_snapdir; static struct vop_vector zfsctl_ops_snapshot; static struct vop_vector zfsctl_ops_shares; static struct vop_vector zfsctl_ops_shares_dir; #endif static vnode_t *zfsctl_mknode_snapdir(vnode_t *); static vnode_t *zfsctl_mknode_shares(vnode_t *); static vnode_t *zfsctl_snapshot_mknode(vnode_t *, uint64_t objset); static int zfsctl_unmount_snap(zfs_snapentry_t *, int, cred_t *); #ifdef illumos static gfs_opsvec_t zfsctl_opsvec[] = { { ".zfs", zfsctl_tops_root, &zfsctl_ops_root }, { ".zfs/snapshot", zfsctl_tops_snapdir, &zfsctl_ops_snapdir }, { ".zfs/snapshot/vnode", zfsctl_tops_snapshot, &zfsctl_ops_snapshot }, { ".zfs/shares", zfsctl_tops_shares, &zfsctl_ops_shares_dir }, { ".zfs/shares/vnode", zfsctl_tops_shares, &zfsctl_ops_shares }, { NULL } }; #endif /* * Root directory elements. We only have two entries * snapshot and shares. */ static gfs_dirent_t zfsctl_root_entries[] = { { "snapshot", zfsctl_mknode_snapdir, GFS_CACHE_VNODE }, { "shares", zfsctl_mknode_shares, GFS_CACHE_VNODE }, { NULL } }; /* include . and .. in the calculation */ #define NROOT_ENTRIES ((sizeof (zfsctl_root_entries) / \ sizeof (gfs_dirent_t)) + 1) /* * Initialize the various GFS pieces we'll need to create and manipulate .zfs * directories. This is called from the ZFS init routine, and initializes the * vnode ops vectors that we'll be using. */ void zfsctl_init(void) { #ifdef illumos VERIFY(gfs_make_opsvec(zfsctl_opsvec) == 0); #endif } void zfsctl_fini(void) { #ifdef illumos /* * Remove vfsctl vnode ops */ if (zfsctl_ops_root) vn_freevnodeops(zfsctl_ops_root); if (zfsctl_ops_snapdir) vn_freevnodeops(zfsctl_ops_snapdir); if (zfsctl_ops_snapshot) vn_freevnodeops(zfsctl_ops_snapshot); if (zfsctl_ops_shares) vn_freevnodeops(zfsctl_ops_shares); if (zfsctl_ops_shares_dir) vn_freevnodeops(zfsctl_ops_shares_dir); zfsctl_ops_root = NULL; zfsctl_ops_snapdir = NULL; zfsctl_ops_snapshot = NULL; zfsctl_ops_shares = NULL; zfsctl_ops_shares_dir = NULL; #endif /* illumos */ } boolean_t zfsctl_is_node(vnode_t *vp) { return (vn_matchops(vp, zfsctl_ops_root) || vn_matchops(vp, zfsctl_ops_snapdir) || vn_matchops(vp, zfsctl_ops_snapshot) || vn_matchops(vp, zfsctl_ops_shares) || vn_matchops(vp, zfsctl_ops_shares_dir)); } /* * Return the inode number associated with the 'snapshot' or * 'shares' directory. */ /* ARGSUSED */ static ino64_t zfsctl_root_inode_cb(vnode_t *vp, int index) { zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; ASSERT(index < 2); if (index == 0) return (ZFSCTL_INO_SNAPDIR); return (zfsvfs->z_shares_dir); } /* * Create the '.zfs' directory. This directory is cached as part of the VFS * structure. This results in a hold on the vfs_t. The code in zfs_umount() * therefore checks against a vfs_count of 2 instead of 1. This reference * is removed when the ctldir is destroyed in the unmount. */ void zfsctl_create(zfsvfs_t *zfsvfs) { vnode_t *vp, *rvp; zfsctl_node_t *zcp; uint64_t crtime[2]; ASSERT(zfsvfs->z_ctldir == NULL); vp = gfs_root_create(sizeof (zfsctl_node_t), zfsvfs->z_vfs, &zfsctl_ops_root, ZFSCTL_INO_ROOT, zfsctl_root_entries, zfsctl_root_inode_cb, MAXNAMELEN, NULL, NULL); zcp = vp->v_data; zcp->zc_id = ZFSCTL_INO_ROOT; VERIFY(VFS_ROOT(zfsvfs->z_vfs, LK_EXCLUSIVE, &rvp) == 0); VERIFY(0 == sa_lookup(VTOZ(rvp)->z_sa_hdl, SA_ZPL_CRTIME(zfsvfs), &crtime, sizeof (crtime))); ZFS_TIME_DECODE(&zcp->zc_cmtime, crtime); VN_URELE(rvp); /* * We're only faking the fact that we have a root of a filesystem for * the sake of the GFS interfaces. Undo the flag manipulation it did * for us. */ vp->v_vflag &= ~VV_ROOT; zfsvfs->z_ctldir = vp; VOP_UNLOCK(vp, 0); } /* * Destroy the '.zfs' directory. Only called when the filesystem is unmounted. * There might still be more references if we were force unmounted, but only * new zfs_inactive() calls can occur and they don't reference .zfs */ void zfsctl_destroy(zfsvfs_t *zfsvfs) { VN_RELE(zfsvfs->z_ctldir); zfsvfs->z_ctldir = NULL; } /* * Given a root znode, retrieve the associated .zfs directory. * Add a hold to the vnode and return it. */ vnode_t * zfsctl_root(znode_t *zp) { ASSERT(zfs_has_ctldir(zp)); VN_HOLD(zp->z_zfsvfs->z_ctldir); return (zp->z_zfsvfs->z_ctldir); } static int zfsctl_common_print(ap) struct vop_print_args /* { struct vnode *a_vp; } */ *ap; { vnode_t *vp = ap->a_vp; gfs_file_t *fp = vp->v_data; printf(" parent = %p\n", fp->gfs_parent); printf(" type = %d\n", fp->gfs_type); printf(" index = %d\n", fp->gfs_index); printf(" ino = %ju\n", (uintmax_t)fp->gfs_ino); return (0); } /* * Common open routine. Disallow any write access. */ /* ARGSUSED */ static int zfsctl_common_open(struct vop_open_args *ap) { int flags = ap->a_mode; if (flags & FWRITE) return (SET_ERROR(EACCES)); return (0); } /* * Common close routine. Nothing to do here. */ /* ARGSUSED */ static int zfsctl_common_close(struct vop_close_args *ap) { return (0); } /* * Common access routine. Disallow writes. */ /* ARGSUSED */ static int zfsctl_common_access(ap) struct vop_access_args /* { struct vnode *a_vp; accmode_t a_accmode; struct ucred *a_cred; struct thread *a_td; } */ *ap; { accmode_t accmode = ap->a_accmode; #ifdef TODO if (flags & V_ACE_MASK) { if (accmode & ACE_ALL_WRITE_PERMS) return (SET_ERROR(EACCES)); } else { #endif if (accmode & VWRITE) return (SET_ERROR(EACCES)); #ifdef TODO } #endif return (0); } /* * Common getattr function. Fill in basic information. */ static void zfsctl_common_getattr(vnode_t *vp, vattr_t *vap) { timestruc_t now; vap->va_uid = 0; vap->va_gid = 0; vap->va_rdev = 0; /* * We are a purely virtual object, so we have no * blocksize or allocated blocks. */ vap->va_blksize = 0; vap->va_nblocks = 0; vap->va_seq = 0; vap->va_fsid = vp->v_mount->mnt_stat.f_fsid.val[0]; vap->va_mode = S_IRUSR | S_IXUSR | S_IRGRP | S_IXGRP | S_IROTH | S_IXOTH; vap->va_type = VDIR; /* * We live in the now (for atime). */ gethrestime(&now); vap->va_atime = now; /* FreeBSD: Reset chflags(2) flags. */ vap->va_flags = 0; } /*ARGSUSED*/ static int zfsctl_common_fid(ap) struct vop_fid_args /* { struct vnode *a_vp; struct fid *a_fid; } */ *ap; { vnode_t *vp = ap->a_vp; fid_t *fidp = (void *)ap->a_fid; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; zfsctl_node_t *zcp = vp->v_data; uint64_t object = zcp->zc_id; zfid_short_t *zfid; int i; ZFS_ENTER(zfsvfs); #ifdef illumos if (fidp->fid_len < SHORT_FID_LEN) { fidp->fid_len = SHORT_FID_LEN; ZFS_EXIT(zfsvfs); return (SET_ERROR(ENOSPC)); } #endif zfid = (zfid_short_t *)fidp; zfid->zf_len = SHORT_FID_LEN; for (i = 0; i < sizeof (zfid->zf_object); i++) zfid->zf_object[i] = (uint8_t)(object >> (8 * i)); /* .zfs znodes always have a generation number of 0 */ for (i = 0; i < sizeof (zfid->zf_gen); i++) zfid->zf_gen[i] = 0; ZFS_EXIT(zfsvfs); return (0); } /*ARGSUSED*/ static int zfsctl_shares_fid(ap) struct vop_fid_args /* { struct vnode *a_vp; struct fid *a_fid; } */ *ap; { vnode_t *vp = ap->a_vp; fid_t *fidp = (void *)ap->a_fid; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; znode_t *dzp; int error; ZFS_ENTER(zfsvfs); if (zfsvfs->z_shares_dir == 0) { ZFS_EXIT(zfsvfs); return (SET_ERROR(ENOTSUP)); } if ((error = zfs_zget(zfsvfs, zfsvfs->z_shares_dir, &dzp)) == 0) { error = VOP_FID(ZTOV(dzp), fidp); VN_RELE(ZTOV(dzp)); } ZFS_EXIT(zfsvfs); return (error); } /* * .zfs inode namespace * * We need to generate unique inode numbers for all files and directories * within the .zfs pseudo-filesystem. We use the following scheme: * * ENTRY ZFSCTL_INODE * .zfs 1 * .zfs/snapshot 2 * .zfs/snapshot/ objectid(snap) */ #define ZFSCTL_INO_SNAP(id) (id) /* * Get root directory attributes. */ /* ARGSUSED */ static int zfsctl_root_getattr(ap) struct vop_getattr_args /* { struct vnode *a_vp; struct vattr *a_vap; struct ucred *a_cred; } */ *ap; { struct vnode *vp = ap->a_vp; struct vattr *vap = ap->a_vap; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; zfsctl_node_t *zcp = vp->v_data; ZFS_ENTER(zfsvfs); vap->va_nodeid = ZFSCTL_INO_ROOT; vap->va_nlink = vap->va_size = NROOT_ENTRIES; vap->va_mtime = vap->va_ctime = zcp->zc_cmtime; vap->va_birthtime = vap->va_ctime; zfsctl_common_getattr(vp, vap); ZFS_EXIT(zfsvfs); return (0); } /* * Special case the handling of "..". */ /* ARGSUSED */ int zfsctl_root_lookup(vnode_t *dvp, char *nm, vnode_t **vpp, pathname_t *pnp, int flags, vnode_t *rdir, cred_t *cr, caller_context_t *ct, int *direntflags, pathname_t *realpnp) { zfsvfs_t *zfsvfs = dvp->v_vfsp->vfs_data; int err; /* * No extended attributes allowed under .zfs */ if (flags & LOOKUP_XATTR) return (SET_ERROR(EINVAL)); ZFS_ENTER(zfsvfs); if (strcmp(nm, "..") == 0) { #ifdef illumos err = VFS_ROOT(dvp->v_vfsp, LK_EXCLUSIVE, vpp); #else /* * NB: can not use VFS_ROOT here as it would acquire * the vnode lock of the parent (root) vnode while * holding the child's (.zfs) lock. */ znode_t *rootzp; err = zfs_zget(zfsvfs, zfsvfs->z_root, &rootzp); if (err == 0) *vpp = ZTOV(rootzp); #endif } else { err = gfs_vop_lookup(dvp, nm, vpp, pnp, flags, rdir, cr, ct, direntflags, realpnp); } ZFS_EXIT(zfsvfs); return (err); } static int zfsctl_freebsd_root_lookup(ap) struct vop_lookup_args /* { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; } */ *ap; { vnode_t *dvp = ap->a_dvp; vnode_t **vpp = ap->a_vpp; cred_t *cr = ap->a_cnp->cn_cred; int flags = ap->a_cnp->cn_flags; int lkflags = ap->a_cnp->cn_lkflags; int nameiop = ap->a_cnp->cn_nameiop; char nm[NAME_MAX + 1]; int err; if ((flags & ISLASTCN) && (nameiop == RENAME || nameiop == CREATE)) return (EOPNOTSUPP); ASSERT(ap->a_cnp->cn_namelen < sizeof(nm)); strlcpy(nm, ap->a_cnp->cn_nameptr, ap->a_cnp->cn_namelen + 1); relookup: err = zfsctl_root_lookup(dvp, nm, vpp, NULL, 0, NULL, cr, NULL, NULL, NULL); if (err == 0 && (nm[0] != '.' || nm[1] != '\0')) { if (flags & ISDOTDOT) { VOP_UNLOCK(dvp, 0); err = vn_lock(*vpp, lkflags); if (err != 0) { vrele(*vpp); *vpp = NULL; } vn_lock(dvp, LK_EXCLUSIVE | LK_RETRY); } else { err = vn_lock(*vpp, LK_EXCLUSIVE); if (err != 0) { VERIFY3S(err, ==, ENOENT); goto relookup; } } } return (err); } static int zfsctl_root_print(ap) struct vop_print_args /* { struct vnode *a_vp; } */ *ap; { printf(" .zfs node\n"); zfsctl_common_print(ap); return (0); } #ifdef illumos static int zfsctl_pathconf(vnode_t *vp, int cmd, ulong_t *valp, cred_t *cr, caller_context_t *ct) { /* * We only care about ACL_ENABLED so that libsec can * display ACL correctly and not default to POSIX draft. */ if (cmd == _PC_ACL_ENABLED) { *valp = _ACL_ACE_ENABLED; return (0); } return (fs_pathconf(vp, cmd, valp, cr, ct)); } #endif /* illumos */ #ifdef illumos static const fs_operation_def_t zfsctl_tops_root[] = { { VOPNAME_OPEN, { .vop_open = zfsctl_common_open } }, { VOPNAME_CLOSE, { .vop_close = zfsctl_common_close } }, { VOPNAME_IOCTL, { .error = fs_inval } }, { VOPNAME_GETATTR, { .vop_getattr = zfsctl_root_getattr } }, { VOPNAME_ACCESS, { .vop_access = zfsctl_common_access } }, { VOPNAME_READDIR, { .vop_readdir = gfs_vop_readdir } }, { VOPNAME_LOOKUP, { .vop_lookup = zfsctl_root_lookup } }, { VOPNAME_SEEK, { .vop_seek = fs_seek } }, { VOPNAME_INACTIVE, { .vop_inactive = gfs_vop_inactive } }, { VOPNAME_PATHCONF, { .vop_pathconf = zfsctl_pathconf } }, { VOPNAME_FID, { .vop_fid = zfsctl_common_fid } }, { NULL } }; #endif /* illumos */ static struct vop_vector zfsctl_ops_root = { .vop_default = &default_vnodeops, .vop_open = zfsctl_common_open, .vop_close = zfsctl_common_close, .vop_ioctl = VOP_EINVAL, .vop_getattr = zfsctl_root_getattr, .vop_access = zfsctl_common_access, .vop_readdir = gfs_vop_readdir, .vop_lookup = zfsctl_freebsd_root_lookup, .vop_inactive = VOP_NULL, .vop_reclaim = gfs_vop_reclaim, #ifdef TODO .vop_pathconf = zfsctl_pathconf, #endif .vop_fid = zfsctl_common_fid, .vop_print = zfsctl_root_print, }; /* * Gets the full dataset name that corresponds to the given snapshot name * Example: * zfsctl_snapshot_zname("snap1") -> "mypool/myfs@snap1" */ static int zfsctl_snapshot_zname(vnode_t *vp, const char *name, int len, char *zname) { objset_t *os = ((zfsvfs_t *)((vp)->v_vfsp->vfs_data))->z_os; if (zfs_component_namecheck(name, NULL, NULL) != 0) return (SET_ERROR(EILSEQ)); dmu_objset_name(os, zname); if (strlen(zname) + 1 + strlen(name) >= len) return (SET_ERROR(ENAMETOOLONG)); (void) strcat(zname, "@"); (void) strcat(zname, name); return (0); } static int zfsctl_unmount_snap(zfs_snapentry_t *sep, int fflags, cred_t *cr) { vnode_t *svp = sep->se_root; int error; ASSERT(vn_ismntpt(svp)); /* this will be dropped by dounmount() */ if ((error = vn_vfswlock(svp)) != 0) return (error); #ifdef illumos VN_HOLD(svp); error = dounmount(vn_mountedvfs(svp), fflags, cr); if (error) { VN_RELE(svp); return (error); } /* * We can't use VN_RELE(), as that will try to invoke * zfsctl_snapdir_inactive(), which would cause us to destroy * the sd_lock mutex held by our caller. */ ASSERT(svp->v_count == 1); gfs_vop_reclaim(svp, cr, NULL); kmem_free(sep->se_name, strlen(sep->se_name) + 1); kmem_free(sep, sizeof (zfs_snapentry_t)); return (0); #else vfs_ref(vn_mountedvfs(svp)); return (dounmount(vn_mountedvfs(svp), fflags, curthread)); #endif } #ifdef illumos static void zfsctl_rename_snap(zfsctl_snapdir_t *sdp, zfs_snapentry_t *sep, const char *nm) { avl_index_t where; vfs_t *vfsp; refstr_t *pathref; char newpath[MAXNAMELEN]; char *tail; ASSERT(MUTEX_HELD(&sdp->sd_lock)); ASSERT(sep != NULL); vfsp = vn_mountedvfs(sep->se_root); ASSERT(vfsp != NULL); vfs_lock_wait(vfsp); /* * Change the name in the AVL tree. */ avl_remove(&sdp->sd_snaps, sep); kmem_free(sep->se_name, strlen(sep->se_name) + 1); sep->se_name = kmem_alloc(strlen(nm) + 1, KM_SLEEP); (void) strcpy(sep->se_name, nm); VERIFY(avl_find(&sdp->sd_snaps, sep, &where) == NULL); avl_insert(&sdp->sd_snaps, sep, where); /* * Change the current mountpoint info: * - update the tail of the mntpoint path * - update the tail of the resource path */ pathref = vfs_getmntpoint(vfsp); (void) strncpy(newpath, refstr_value(pathref), sizeof (newpath)); VERIFY((tail = strrchr(newpath, '/')) != NULL); *(tail+1) = '\0'; ASSERT3U(strlen(newpath) + strlen(nm), <, sizeof (newpath)); (void) strcat(newpath, nm); refstr_rele(pathref); vfs_setmntpoint(vfsp, newpath, 0); pathref = vfs_getresource(vfsp); (void) strncpy(newpath, refstr_value(pathref), sizeof (newpath)); VERIFY((tail = strrchr(newpath, '@')) != NULL); *(tail+1) = '\0'; ASSERT3U(strlen(newpath) + strlen(nm), <, sizeof (newpath)); (void) strcat(newpath, nm); refstr_rele(pathref); vfs_setresource(vfsp, newpath, 0); vfs_unlock(vfsp); } #endif /* illumos */ #ifdef illumos /*ARGSUSED*/ static int zfsctl_snapdir_rename(vnode_t *sdvp, char *snm, vnode_t *tdvp, char *tnm, cred_t *cr, caller_context_t *ct, int flags) { zfsctl_snapdir_t *sdp = sdvp->v_data; zfs_snapentry_t search, *sep; zfsvfs_t *zfsvfs; avl_index_t where; char from[MAXNAMELEN], to[MAXNAMELEN]; char real[MAXNAMELEN], fsname[MAXNAMELEN]; int err; zfsvfs = sdvp->v_vfsp->vfs_data; ZFS_ENTER(zfsvfs); if ((flags & FIGNORECASE) || zfsvfs->z_case == ZFS_CASE_INSENSITIVE) { err = dmu_snapshot_realname(zfsvfs->z_os, snm, real, MAXNAMELEN, NULL); if (err == 0) { snm = real; } else if (err != ENOTSUP) { ZFS_EXIT(zfsvfs); return (err); } } ZFS_EXIT(zfsvfs); dmu_objset_name(zfsvfs->z_os, fsname); err = zfsctl_snapshot_zname(sdvp, snm, MAXNAMELEN, from); if (err == 0) err = zfsctl_snapshot_zname(tdvp, tnm, MAXNAMELEN, to); if (err == 0) err = zfs_secpolicy_rename_perms(from, to, cr); if (err != 0) return (err); /* * Cannot move snapshots out of the snapdir. */ if (sdvp != tdvp) return (SET_ERROR(EINVAL)); if (strcmp(snm, tnm) == 0) return (0); mutex_enter(&sdp->sd_lock); search.se_name = (char *)snm; if ((sep = avl_find(&sdp->sd_snaps, &search, &where)) == NULL) { mutex_exit(&sdp->sd_lock); return (SET_ERROR(ENOENT)); } err = dsl_dataset_rename_snapshot(fsname, snm, tnm, 0); if (err == 0) zfsctl_rename_snap(sdp, sep, tnm); mutex_exit(&sdp->sd_lock); return (err); } #endif /* illumos */ #ifdef illumos /* ARGSUSED */ static int zfsctl_snapdir_remove(vnode_t *dvp, char *name, vnode_t *cwd, cred_t *cr, caller_context_t *ct, int flags) { zfsctl_snapdir_t *sdp = dvp->v_data; zfs_snapentry_t *sep; zfs_snapentry_t search; zfsvfs_t *zfsvfs; char snapname[MAXNAMELEN]; char real[MAXNAMELEN]; int err; zfsvfs = dvp->v_vfsp->vfs_data; ZFS_ENTER(zfsvfs); if ((flags & FIGNORECASE) || zfsvfs->z_case == ZFS_CASE_INSENSITIVE) { err = dmu_snapshot_realname(zfsvfs->z_os, name, real, MAXNAMELEN, NULL); if (err == 0) { name = real; } else if (err != ENOTSUP) { ZFS_EXIT(zfsvfs); return (err); } } ZFS_EXIT(zfsvfs); err = zfsctl_snapshot_zname(dvp, name, MAXNAMELEN, snapname); if (err == 0) err = zfs_secpolicy_destroy_perms(snapname, cr); if (err != 0) return (err); mutex_enter(&sdp->sd_lock); search.se_name = name; sep = avl_find(&sdp->sd_snaps, &search, NULL); if (sep) { avl_remove(&sdp->sd_snaps, sep); err = zfsctl_unmount_snap(sep, MS_FORCE, cr); if (err != 0) avl_add(&sdp->sd_snaps, sep); else err = dsl_destroy_snapshot(snapname, B_FALSE); } else { err = SET_ERROR(ENOENT); } mutex_exit(&sdp->sd_lock); return (err); } #endif /* illumos */ /* * This creates a snapshot under '.zfs/snapshot'. */ /* ARGSUSED */ static int zfsctl_snapdir_mkdir(vnode_t *dvp, char *dirname, vattr_t *vap, vnode_t **vpp, cred_t *cr, caller_context_t *cc, int flags, vsecattr_t *vsecp) { zfsvfs_t *zfsvfs = dvp->v_vfsp->vfs_data; char name[MAXNAMELEN]; int err; static enum symfollow follow = NO_FOLLOW; static enum uio_seg seg = UIO_SYSSPACE; if (zfs_component_namecheck(dirname, NULL, NULL) != 0) return (SET_ERROR(EILSEQ)); dmu_objset_name(zfsvfs->z_os, name); *vpp = NULL; err = zfs_secpolicy_snapshot_perms(name, cr); if (err != 0) return (err); if (err == 0) { err = dmu_objset_snapshot_one(name, dirname); if (err != 0) return (err); err = lookupnameat(dirname, seg, follow, NULL, vpp, dvp); } return (err); } static int zfsctl_freebsd_snapdir_mkdir(ap) struct vop_mkdir_args /* { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; struct vattr *a_vap; } */ *ap; { ASSERT(ap->a_cnp->cn_flags & SAVENAME); return (zfsctl_snapdir_mkdir(ap->a_dvp, ap->a_cnp->cn_nameptr, NULL, ap->a_vpp, ap->a_cnp->cn_cred, NULL, 0, NULL)); } /* * Lookup entry point for the 'snapshot' directory. Try to open the * snapshot if it exist, creating the pseudo filesystem vnode as necessary. * Perform a mount of the associated dataset on top of the vnode. */ /* ARGSUSED */ int zfsctl_snapdir_lookup(ap) struct vop_lookup_args /* { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; } */ *ap; { vnode_t *dvp = ap->a_dvp; vnode_t **vpp = ap->a_vpp; struct componentname *cnp = ap->a_cnp; char nm[NAME_MAX + 1]; zfsctl_snapdir_t *sdp = dvp->v_data; objset_t *snap; char snapname[MAXNAMELEN]; char real[MAXNAMELEN]; char *mountpoint; zfs_snapentry_t *sep, search; size_t mountpoint_len; avl_index_t where; zfsvfs_t *zfsvfs = dvp->v_vfsp->vfs_data; int err; int ltype, flags = 0; /* * No extended attributes allowed under .zfs */ if (flags & LOOKUP_XATTR) return (SET_ERROR(EINVAL)); ASSERT(ap->a_cnp->cn_namelen < sizeof(nm)); strlcpy(nm, ap->a_cnp->cn_nameptr, ap->a_cnp->cn_namelen + 1); ASSERT(dvp->v_type == VDIR); *vpp = NULL; /* * If we get a recursive call, that means we got called * from the domount() code while it was trying to look up the * spec (which looks like a local path for zfs). We need to * add some flag to domount() to tell it not to do this lookup. */ if (MUTEX_HELD(&sdp->sd_lock)) return (SET_ERROR(ENOENT)); ZFS_ENTER(zfsvfs); if (gfs_lookup_dot(vpp, dvp, zfsvfs->z_ctldir, nm) == 0) { if (nm[0] == '.' && nm[1] == '.' && nm[2] =='\0') { VOP_UNLOCK(dvp, 0); VERIFY0(vn_lock(*vpp, LK_EXCLUSIVE)); VERIFY0(vn_lock(dvp, LK_EXCLUSIVE)); } ZFS_EXIT(zfsvfs); return (0); } if (flags & FIGNORECASE) { boolean_t conflict = B_FALSE; err = dmu_snapshot_realname(zfsvfs->z_os, nm, real, MAXNAMELEN, &conflict); if (err == 0) { strlcpy(nm, real, sizeof(nm)); } else if (err != ENOTSUP) { ZFS_EXIT(zfsvfs); return (err); } #if 0 if (realpnp) (void) strlcpy(realpnp->pn_buf, nm, realpnp->pn_bufsize); if (conflict && direntflags) *direntflags = ED_CASE_CONFLICT; #endif } relookup: mutex_enter(&sdp->sd_lock); search.se_name = (char *)nm; if ((sep = avl_find(&sdp->sd_snaps, &search, &where)) != NULL) { *vpp = sep->se_root; VN_HOLD(*vpp); err = traverse(vpp, LK_EXCLUSIVE | LK_RETRY); if (err != 0) { *vpp = NULL; } else if (*vpp == sep->se_root) { /* * The snapshot was unmounted behind our backs, * try to remount it. */ VERIFY(zfsctl_snapshot_zname(dvp, nm, MAXNAMELEN, snapname) == 0); goto domount; - } else { - /* - * VROOT was set during the traverse call. We need - * to clear it since we're pretending to be part - * of our parent's vfs. - */ - (*vpp)->v_flag &= ~VROOT; } mutex_exit(&sdp->sd_lock); ZFS_EXIT(zfsvfs); return (err); } /* * The requested snapshot is not currently mounted, look it up. */ err = zfsctl_snapshot_zname(dvp, nm, MAXNAMELEN, snapname); if (err != 0) { mutex_exit(&sdp->sd_lock); ZFS_EXIT(zfsvfs); /* * handle "ls *" or "?" in a graceful manner, * forcing EILSEQ to ENOENT. * Since shell ultimately passes "*" or "?" as name to lookup */ return (err == EILSEQ ? ENOENT : err); } if (dmu_objset_hold(snapname, FTAG, &snap) != 0) { mutex_exit(&sdp->sd_lock); #ifdef illumos ZFS_EXIT(zfsvfs); return (SET_ERROR(ENOENT)); #else /* !illumos */ /* Translate errors and add SAVENAME when needed. */ if ((cnp->cn_flags & ISLASTCN) && cnp->cn_nameiop == CREATE) { err = EJUSTRETURN; cnp->cn_flags |= SAVENAME; } else { err = SET_ERROR(ENOENT); } ZFS_EXIT(zfsvfs); return (err); #endif /* illumos */ } sep = kmem_alloc(sizeof (zfs_snapentry_t), KM_SLEEP); sep->se_name = kmem_alloc(strlen(nm) + 1, KM_SLEEP); (void) strcpy(sep->se_name, nm); *vpp = sep->se_root = zfsctl_snapshot_mknode(dvp, dmu_objset_id(snap)); avl_insert(&sdp->sd_snaps, sep, where); dmu_objset_rele(snap, FTAG); domount: mountpoint_len = strlen(dvp->v_vfsp->mnt_stat.f_mntonname) + strlen("/" ZFS_CTLDIR_NAME "/snapshot/") + strlen(nm) + 1; mountpoint = kmem_alloc(mountpoint_len, KM_SLEEP); (void) snprintf(mountpoint, mountpoint_len, "%s/" ZFS_CTLDIR_NAME "/snapshot/%s", dvp->v_vfsp->mnt_stat.f_mntonname, nm); mutex_exit(&sdp->sd_lock); /* * The vnode may get reclaimed between dropping sd_lock and * getting the vnode lock. * */ err = vn_lock(*vpp, LK_EXCLUSIVE); if (err == ENOENT) goto relookup; VERIFY0(err); err = mount_snapshot(curthread, vpp, "zfs", mountpoint, snapname, 0); kmem_free(mountpoint, mountpoint_len); if (err == 0) { /* * Fix up the root vnode mounted on .zfs/snapshot/. * * This is where we lie about our v_vfsp in order to * make .zfs/snapshot/ accessible over NFS * without requiring manual mounts of . */ ASSERT(VTOZ(*vpp)->z_zfsvfs != zfsvfs); VTOZ(*vpp)->z_zfsvfs->z_parent = zfsvfs; - (*vpp)->v_flag &= ~VROOT; } ZFS_EXIT(zfsvfs); #ifdef illumos /* * If we had an error, drop our hold on the vnode and * zfsctl_snapshot_inactive() will clean up. */ if (err != 0) { VN_RELE(*vpp); *vpp = NULL; } #else if (err != 0) *vpp = NULL; #endif return (err); } /* ARGSUSED */ int zfsctl_shares_lookup(ap) struct vop_lookup_args /* { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; } */ *ap; { vnode_t *dvp = ap->a_dvp; vnode_t **vpp = ap->a_vpp; struct componentname *cnp = ap->a_cnp; zfsvfs_t *zfsvfs = dvp->v_vfsp->vfs_data; char nm[NAME_MAX + 1]; znode_t *dzp; int error; ZFS_ENTER(zfsvfs); ASSERT(cnp->cn_namelen < sizeof(nm)); strlcpy(nm, cnp->cn_nameptr, cnp->cn_namelen + 1); if (gfs_lookup_dot(vpp, dvp, zfsvfs->z_ctldir, nm) == 0) { if (nm[0] == '.' && nm[1] == '.' && nm[2] =='\0') { VOP_UNLOCK(dvp, 0); VERIFY0(vn_lock(*vpp, LK_EXCLUSIVE)); VERIFY0(vn_lock(dvp, LK_EXCLUSIVE)); } ZFS_EXIT(zfsvfs); return (0); } if (zfsvfs->z_shares_dir == 0) { ZFS_EXIT(zfsvfs); return (SET_ERROR(ENOTSUP)); } if ((error = zfs_zget(zfsvfs, zfsvfs->z_shares_dir, &dzp)) == 0) { error = VOP_LOOKUP(ZTOV(dzp), vpp, cnp); VN_RELE(ZTOV(dzp)); } ZFS_EXIT(zfsvfs); return (error); } /* ARGSUSED */ static int zfsctl_snapdir_readdir_cb(vnode_t *vp, void *dp, int *eofp, offset_t *offp, offset_t *nextp, void *data, int flags) { zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; char snapname[MAXNAMELEN]; uint64_t id, cookie; boolean_t case_conflict; int error; ZFS_ENTER(zfsvfs); cookie = *offp; dsl_pool_config_enter(dmu_objset_pool(zfsvfs->z_os), FTAG); error = dmu_snapshot_list_next(zfsvfs->z_os, MAXNAMELEN, snapname, &id, &cookie, &case_conflict); dsl_pool_config_exit(dmu_objset_pool(zfsvfs->z_os), FTAG); if (error) { ZFS_EXIT(zfsvfs); if (error == ENOENT) { *eofp = 1; return (0); } return (error); } if (flags & V_RDDIR_ENTFLAGS) { edirent_t *eodp = dp; (void) strcpy(eodp->ed_name, snapname); eodp->ed_ino = ZFSCTL_INO_SNAP(id); eodp->ed_eflags = case_conflict ? ED_CASE_CONFLICT : 0; } else { struct dirent64 *odp = dp; (void) strcpy(odp->d_name, snapname); odp->d_ino = ZFSCTL_INO_SNAP(id); } *nextp = cookie; ZFS_EXIT(zfsvfs); return (0); } /* ARGSUSED */ static int zfsctl_shares_readdir(ap) struct vop_readdir_args /* { struct vnode *a_vp; struct uio *a_uio; struct ucred *a_cred; int *a_eofflag; int *a_ncookies; u_long **a_cookies; } */ *ap; { vnode_t *vp = ap->a_vp; uio_t *uiop = ap->a_uio; cred_t *cr = ap->a_cred; int *eofp = ap->a_eofflag; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; znode_t *dzp; int error; ZFS_ENTER(zfsvfs); if (zfsvfs->z_shares_dir == 0) { ZFS_EXIT(zfsvfs); return (SET_ERROR(ENOTSUP)); } if ((error = zfs_zget(zfsvfs, zfsvfs->z_shares_dir, &dzp)) == 0) { vn_lock(ZTOV(dzp), LK_SHARED | LK_RETRY); error = VOP_READDIR(ZTOV(dzp), uiop, cr, eofp, ap->a_ncookies, ap->a_cookies); VN_URELE(ZTOV(dzp)); } else { *eofp = 1; error = SET_ERROR(ENOENT); } ZFS_EXIT(zfsvfs); return (error); } /* * pvp is the '.zfs' directory (zfsctl_node_t). * * Creates vp, which is '.zfs/snapshot' (zfsctl_snapdir_t). * * This function is the callback to create a GFS vnode for '.zfs/snapshot' * when a lookup is performed on .zfs for "snapshot". */ vnode_t * zfsctl_mknode_snapdir(vnode_t *pvp) { vnode_t *vp; zfsctl_snapdir_t *sdp; vp = gfs_dir_create(sizeof (zfsctl_snapdir_t), pvp, pvp->v_vfsp, &zfsctl_ops_snapdir, NULL, NULL, MAXNAMELEN, zfsctl_snapdir_readdir_cb, NULL); sdp = vp->v_data; sdp->sd_node.zc_id = ZFSCTL_INO_SNAPDIR; sdp->sd_node.zc_cmtime = ((zfsctl_node_t *)pvp->v_data)->zc_cmtime; mutex_init(&sdp->sd_lock, NULL, MUTEX_DEFAULT, NULL); avl_create(&sdp->sd_snaps, snapentry_compare, sizeof (zfs_snapentry_t), offsetof(zfs_snapentry_t, se_node)); VOP_UNLOCK(vp, 0); return (vp); } vnode_t * zfsctl_mknode_shares(vnode_t *pvp) { vnode_t *vp; zfsctl_node_t *sdp; vp = gfs_dir_create(sizeof (zfsctl_node_t), pvp, pvp->v_vfsp, &zfsctl_ops_shares, NULL, NULL, MAXNAMELEN, NULL, NULL); sdp = vp->v_data; sdp->zc_cmtime = ((zfsctl_node_t *)pvp->v_data)->zc_cmtime; VOP_UNLOCK(vp, 0); return (vp); } /* ARGSUSED */ static int zfsctl_shares_getattr(ap) struct vop_getattr_args /* { struct vnode *a_vp; struct vattr *a_vap; struct ucred *a_cred; struct thread *a_td; } */ *ap; { vnode_t *vp = ap->a_vp; vattr_t *vap = ap->a_vap; cred_t *cr = ap->a_cred; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; znode_t *dzp; int error; ZFS_ENTER(zfsvfs); if (zfsvfs->z_shares_dir == 0) { ZFS_EXIT(zfsvfs); return (SET_ERROR(ENOTSUP)); } if ((error = zfs_zget(zfsvfs, zfsvfs->z_shares_dir, &dzp)) == 0) { vn_lock(ZTOV(dzp), LK_SHARED | LK_RETRY); error = VOP_GETATTR(ZTOV(dzp), vap, cr); VN_URELE(ZTOV(dzp)); } ZFS_EXIT(zfsvfs); return (error); } /* ARGSUSED */ static int zfsctl_snapdir_getattr(ap) struct vop_getattr_args /* { struct vnode *a_vp; struct vattr *a_vap; struct ucred *a_cred; } */ *ap; { vnode_t *vp = ap->a_vp; vattr_t *vap = ap->a_vap; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; zfsctl_snapdir_t *sdp = vp->v_data; ZFS_ENTER(zfsvfs); zfsctl_common_getattr(vp, vap); vap->va_nodeid = gfs_file_inode(vp); vap->va_nlink = vap->va_size = avl_numnodes(&sdp->sd_snaps) + 2; vap->va_ctime = vap->va_mtime = dmu_objset_snap_cmtime(zfsvfs->z_os); vap->va_birthtime = vap->va_ctime; ZFS_EXIT(zfsvfs); return (0); } /* ARGSUSED */ static int zfsctl_snapdir_reclaim(ap) struct vop_reclaim_args /* { struct vnode *a_vp; struct thread *a_td; } */ *ap; { vnode_t *vp = ap->a_vp; zfsctl_snapdir_t *sdp = vp->v_data; zfs_snapentry_t *sep; ASSERT(avl_numnodes(&sdp->sd_snaps) == 0); mutex_destroy(&sdp->sd_lock); avl_destroy(&sdp->sd_snaps); gfs_vop_reclaim(ap); return (0); } static int zfsctl_shares_print(ap) struct vop_print_args /* { struct vnode *a_vp; } */ *ap; { printf(" .zfs/shares node\n"); zfsctl_common_print(ap); return (0); } static int zfsctl_snapdir_print(ap) struct vop_print_args /* { struct vnode *a_vp; } */ *ap; { vnode_t *vp = ap->a_vp; zfsctl_snapdir_t *sdp = vp->v_data; printf(" .zfs/snapshot node\n"); printf(" number of children = %lu\n", avl_numnodes(&sdp->sd_snaps)); zfsctl_common_print(ap); return (0); } #ifdef illumos static const fs_operation_def_t zfsctl_tops_snapdir[] = { { VOPNAME_OPEN, { .vop_open = zfsctl_common_open } }, { VOPNAME_CLOSE, { .vop_close = zfsctl_common_close } }, { VOPNAME_IOCTL, { .error = fs_inval } }, { VOPNAME_GETATTR, { .vop_getattr = zfsctl_snapdir_getattr } }, { VOPNAME_ACCESS, { .vop_access = zfsctl_common_access } }, { VOPNAME_RENAME, { .vop_rename = zfsctl_snapdir_rename } }, { VOPNAME_RMDIR, { .vop_rmdir = zfsctl_snapdir_remove } }, { VOPNAME_MKDIR, { .vop_mkdir = zfsctl_snapdir_mkdir } }, { VOPNAME_READDIR, { .vop_readdir = gfs_vop_readdir } }, { VOPNAME_LOOKUP, { .vop_lookup = zfsctl_snapdir_lookup } }, { VOPNAME_SEEK, { .vop_seek = fs_seek } }, { VOPNAME_INACTIVE, { .vop_inactive = zfsctl_snapdir_inactive } }, { VOPNAME_FID, { .vop_fid = zfsctl_common_fid } }, { NULL } }; static const fs_operation_def_t zfsctl_tops_shares[] = { { VOPNAME_OPEN, { .vop_open = zfsctl_common_open } }, { VOPNAME_CLOSE, { .vop_close = zfsctl_common_close } }, { VOPNAME_IOCTL, { .error = fs_inval } }, { VOPNAME_GETATTR, { .vop_getattr = zfsctl_shares_getattr } }, { VOPNAME_ACCESS, { .vop_access = zfsctl_common_access } }, { VOPNAME_READDIR, { .vop_readdir = zfsctl_shares_readdir } }, { VOPNAME_LOOKUP, { .vop_lookup = zfsctl_shares_lookup } }, { VOPNAME_SEEK, { .vop_seek = fs_seek } }, { VOPNAME_INACTIVE, { .vop_inactive = gfs_vop_inactive } }, { VOPNAME_FID, { .vop_fid = zfsctl_shares_fid } }, { NULL } }; #else /* !illumos */ static struct vop_vector zfsctl_ops_snapdir = { .vop_default = &default_vnodeops, .vop_open = zfsctl_common_open, .vop_close = zfsctl_common_close, .vop_ioctl = VOP_EINVAL, .vop_getattr = zfsctl_snapdir_getattr, .vop_access = zfsctl_common_access, .vop_mkdir = zfsctl_freebsd_snapdir_mkdir, .vop_readdir = gfs_vop_readdir, .vop_lookup = zfsctl_snapdir_lookup, .vop_inactive = VOP_NULL, .vop_reclaim = zfsctl_snapdir_reclaim, .vop_fid = zfsctl_common_fid, .vop_print = zfsctl_snapdir_print, }; static struct vop_vector zfsctl_ops_shares = { .vop_default = &default_vnodeops, .vop_open = zfsctl_common_open, .vop_close = zfsctl_common_close, .vop_ioctl = VOP_EINVAL, .vop_getattr = zfsctl_shares_getattr, .vop_access = zfsctl_common_access, .vop_readdir = zfsctl_shares_readdir, .vop_lookup = zfsctl_shares_lookup, .vop_inactive = VOP_NULL, .vop_reclaim = gfs_vop_reclaim, .vop_fid = zfsctl_shares_fid, .vop_print = zfsctl_shares_print, }; #endif /* illumos */ /* * pvp is the GFS vnode '.zfs/snapshot'. * * This creates a GFS node under '.zfs/snapshot' representing each * snapshot. This newly created GFS node is what we mount snapshot * vfs_t's ontop of. */ static vnode_t * zfsctl_snapshot_mknode(vnode_t *pvp, uint64_t objset) { vnode_t *vp; zfsctl_node_t *zcp; vp = gfs_dir_create(sizeof (zfsctl_node_t), pvp, pvp->v_vfsp, &zfsctl_ops_snapshot, NULL, NULL, MAXNAMELEN, NULL, NULL); zcp = vp->v_data; zcp->zc_id = objset; VOP_UNLOCK(vp, 0); return (vp); } static int zfsctl_snapshot_inactive(ap) struct vop_inactive_args /* { struct vnode *a_vp; struct thread *a_td; } */ *ap; { vnode_t *vp = ap->a_vp; vrecycle(vp); return (0); } static int zfsctl_snapshot_reclaim(ap) struct vop_reclaim_args /* { struct vnode *a_vp; struct thread *a_td; } */ *ap; { vnode_t *vp = ap->a_vp; cred_t *cr = ap->a_td->td_ucred; zfsctl_snapdir_t *sdp; zfs_snapentry_t *sep, *next; int locked; vnode_t *dvp; VERIFY(gfs_dir_lookup(vp, "..", &dvp, cr, 0, NULL, NULL) == 0); sdp = dvp->v_data; /* this may already have been unmounted */ if (sdp == NULL) { VN_RELE(dvp); return (0); } if (!(locked = MUTEX_HELD(&sdp->sd_lock))) mutex_enter(&sdp->sd_lock); ASSERT(!vn_ismntpt(vp)); sep = avl_first(&sdp->sd_snaps); while (sep != NULL) { next = AVL_NEXT(&sdp->sd_snaps, sep); if (sep->se_root == vp) { avl_remove(&sdp->sd_snaps, sep); kmem_free(sep->se_name, strlen(sep->se_name) + 1); kmem_free(sep, sizeof (zfs_snapentry_t)); break; } sep = next; } ASSERT(sep != NULL); if (!locked) mutex_exit(&sdp->sd_lock); VN_RELE(dvp); /* * Dispose of the vnode for the snapshot mount point. * This is safe to do because once this entry has been removed * from the AVL tree, it can't be found again, so cannot become * "active". If we lookup the same name again we will end up * creating a new vnode. */ gfs_vop_reclaim(ap); return (0); } static int zfsctl_snapshot_vptocnp(struct vop_vptocnp_args *ap) { zfsvfs_t *zfsvfs = ap->a_vp->v_vfsp->vfs_data; vnode_t *dvp, *vp; zfsctl_snapdir_t *sdp; zfs_snapentry_t *sep; int error; ASSERT(zfsvfs->z_ctldir != NULL); error = zfsctl_root_lookup(zfsvfs->z_ctldir, "snapshot", &dvp, NULL, 0, NULL, kcred, NULL, NULL, NULL); if (error != 0) return (error); sdp = dvp->v_data; mutex_enter(&sdp->sd_lock); sep = avl_first(&sdp->sd_snaps); while (sep != NULL) { vp = sep->se_root; if (vp == ap->a_vp) break; sep = AVL_NEXT(&sdp->sd_snaps, sep); } if (sep == NULL) { mutex_exit(&sdp->sd_lock); error = ENOENT; } else { size_t len; len = strlen(sep->se_name); *ap->a_buflen -= len; bcopy(sep->se_name, ap->a_buf + *ap->a_buflen, len); mutex_exit(&sdp->sd_lock); vref(dvp); *ap->a_vpp = dvp; } VN_RELE(dvp); return (error); } static int zfsctl_snaphot_print(ap) struct vop_print_args /* { struct vnode *a_vp; } */ *ap; { vnode_t *vp = ap->a_vp; zfsctl_node_t *zcp = vp->v_data; printf(" .zfs/snapshot/ node\n"); printf(" id = %ju\n", (uintmax_t)zcp->zc_id); zfsctl_common_print(ap); return (0); } /* * These VP's should never see the light of day. They should always * be covered. */ static struct vop_vector zfsctl_ops_snapshot = { .vop_default = &default_vnodeops, .vop_inactive = zfsctl_snapshot_inactive, .vop_reclaim = zfsctl_snapshot_reclaim, .vop_vptocnp = zfsctl_snapshot_vptocnp, .vop_print = zfsctl_snaphot_print, }; int zfsctl_lookup_objset(vfs_t *vfsp, uint64_t objsetid, zfsvfs_t **zfsvfsp) { zfsvfs_t *zfsvfs = vfsp->vfs_data; vnode_t *dvp, *vp; zfsctl_snapdir_t *sdp; zfsctl_node_t *zcp; zfs_snapentry_t *sep; int error; ASSERT(zfsvfs->z_ctldir != NULL); error = zfsctl_root_lookup(zfsvfs->z_ctldir, "snapshot", &dvp, NULL, 0, NULL, kcred, NULL, NULL, NULL); if (error != 0) return (error); sdp = dvp->v_data; mutex_enter(&sdp->sd_lock); sep = avl_first(&sdp->sd_snaps); while (sep != NULL) { vp = sep->se_root; zcp = vp->v_data; if (zcp->zc_id == objsetid) break; sep = AVL_NEXT(&sdp->sd_snaps, sep); } if (sep != NULL) { VN_HOLD(vp); /* * Return the mounted root rather than the covered mount point. * Takes the GFS vnode at .zfs/snapshot/ * and returns the ZFS vnode mounted on top of the GFS node. * This ZFS vnode is the root of the vfs for objset 'objsetid'. */ error = traverse(&vp, LK_SHARED | LK_RETRY); if (error == 0) { if (vp == sep->se_root) { VN_RELE(vp); /* release covered vp */ error = SET_ERROR(EINVAL); } else { *zfsvfsp = VTOZ(vp)->z_zfsvfs; VN_URELE(vp); /* put snapshot's root vp */ } } mutex_exit(&sdp->sd_lock); } else { error = SET_ERROR(EINVAL); mutex_exit(&sdp->sd_lock); } VN_RELE(dvp); return (error); } /* * Unmount any snapshots for the given filesystem. This is called from * zfs_umount() - if we have a ctldir, then go through and unmount all the * snapshots. */ int zfsctl_umount_snapshots(vfs_t *vfsp, int fflags, cred_t *cr) { zfsvfs_t *zfsvfs = vfsp->vfs_data; vnode_t *dvp; zfsctl_snapdir_t *sdp; zfs_snapentry_t *sep, *next; int error; ASSERT(zfsvfs->z_ctldir != NULL); error = zfsctl_root_lookup(zfsvfs->z_ctldir, "snapshot", &dvp, NULL, 0, NULL, cr, NULL, NULL, NULL); if (error != 0) return (error); sdp = dvp->v_data; mutex_enter(&sdp->sd_lock); sep = avl_first(&sdp->sd_snaps); while (sep != NULL) { next = AVL_NEXT(&sdp->sd_snaps, sep); /* * If this snapshot is not mounted, then it must * have just been unmounted by somebody else, and * will be cleaned up by zfsctl_snapdir_inactive(). */ if (vn_ismntpt(sep->se_root)) { error = zfsctl_unmount_snap(sep, fflags, cr); if (error) { avl_index_t where; /* * Before reinserting snapshot to the tree, * check if it was actually removed. For example * when snapshot mount point is busy, we will * have an error here, but there will be no need * to reinsert snapshot. */ if (avl_find(&sdp->sd_snaps, sep, &where) == NULL) avl_insert(&sdp->sd_snaps, sep, where); break; } } sep = next; } mutex_exit(&sdp->sd_lock); VN_RELE(dvp); return (error); } Index: head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c =================================================================== --- head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c (revision 301274) +++ head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c (revision 301275) @@ -1,2201 +1,2202 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2014 by Delphix. All rights reserved. * Copyright (c) 2014 Integros [integros.com] */ /* Portions Copyright 2007 Jeremy Teo */ /* Portions Copyright 2011 Martin Matuska */ #ifdef _KERNEL #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #endif /* _KERNEL */ #include #include #include #include #include #include #include #include #include #include #include "zfs_prop.h" #include "zfs_comutil.h" /* Used by fstat(1). */ SYSCTL_INT(_debug_sizeof, OID_AUTO, znode, CTLFLAG_RD, SYSCTL_NULL_INT_PTR, sizeof(znode_t), "sizeof(znode_t)"); /* * Define ZNODE_STATS to turn on statistic gathering. By default, it is only * turned on when DEBUG is also defined. */ #ifdef DEBUG #define ZNODE_STATS #endif /* DEBUG */ #ifdef ZNODE_STATS #define ZNODE_STAT_ADD(stat) ((stat)++) #else #define ZNODE_STAT_ADD(stat) /* nothing */ #endif /* ZNODE_STATS */ /* * Functions needed for userland (ie: libzpool) are not put under * #ifdef_KERNEL; the rest of the functions have dependencies * (such as VFS logic) that will not compile easily in userland. */ #ifdef _KERNEL /* * Needed to close a small window in zfs_znode_move() that allows the zfsvfs to * be freed before it can be safely accessed. */ krwlock_t zfsvfs_lock; static kmem_cache_t *znode_cache = NULL; /*ARGSUSED*/ static void znode_evict_error(dmu_buf_t *dbuf, void *user_ptr) { /* * We should never drop all dbuf refs without first clearing * the eviction callback. */ panic("evicting znode %p\n", user_ptr); } extern struct vop_vector zfs_vnodeops; extern struct vop_vector zfs_fifoops; extern struct vop_vector zfs_shareops; static int zfs_znode_cache_constructor(void *buf, void *arg, int kmflags) { znode_t *zp = buf; POINTER_INVALIDATE(&zp->z_zfsvfs); list_link_init(&zp->z_link_node); mutex_init(&zp->z_lock, NULL, MUTEX_DEFAULT, NULL); rw_init(&zp->z_parent_lock, NULL, RW_DEFAULT, NULL); rw_init(&zp->z_name_lock, NULL, RW_DEFAULT, NULL); mutex_init(&zp->z_acl_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zp->z_range_lock, NULL, MUTEX_DEFAULT, NULL); avl_create(&zp->z_range_avl, zfs_range_compare, sizeof (rl_t), offsetof(rl_t, r_node)); zp->z_dirlocks = NULL; zp->z_acl_cached = NULL; zp->z_vnode = NULL; zp->z_moved = 0; return (0); } /*ARGSUSED*/ static void zfs_znode_cache_destructor(void *buf, void *arg) { znode_t *zp = buf; ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs)); ASSERT(ZTOV(zp) == NULL); vn_free(ZTOV(zp)); ASSERT(!list_link_active(&zp->z_link_node)); mutex_destroy(&zp->z_lock); rw_destroy(&zp->z_parent_lock); rw_destroy(&zp->z_name_lock); mutex_destroy(&zp->z_acl_lock); avl_destroy(&zp->z_range_avl); mutex_destroy(&zp->z_range_lock); ASSERT(zp->z_dirlocks == NULL); ASSERT(zp->z_acl_cached == NULL); } #ifdef ZNODE_STATS static struct { uint64_t zms_zfsvfs_invalid; uint64_t zms_zfsvfs_recheck1; uint64_t zms_zfsvfs_unmounted; uint64_t zms_zfsvfs_recheck2; uint64_t zms_obj_held; uint64_t zms_vnode_locked; uint64_t zms_not_only_dnlc; } znode_move_stats; #endif /* ZNODE_STATS */ #ifdef illumos static void zfs_znode_move_impl(znode_t *ozp, znode_t *nzp) { vnode_t *vp; /* Copy fields. */ nzp->z_zfsvfs = ozp->z_zfsvfs; /* Swap vnodes. */ vp = nzp->z_vnode; nzp->z_vnode = ozp->z_vnode; ozp->z_vnode = vp; /* let destructor free the overwritten vnode */ ZTOV(ozp)->v_data = ozp; ZTOV(nzp)->v_data = nzp; nzp->z_id = ozp->z_id; ASSERT(ozp->z_dirlocks == NULL); /* znode not in use */ ASSERT(avl_numnodes(&ozp->z_range_avl) == 0); nzp->z_unlinked = ozp->z_unlinked; nzp->z_atime_dirty = ozp->z_atime_dirty; nzp->z_zn_prefetch = ozp->z_zn_prefetch; nzp->z_blksz = ozp->z_blksz; nzp->z_seq = ozp->z_seq; nzp->z_mapcnt = ozp->z_mapcnt; nzp->z_gen = ozp->z_gen; nzp->z_sync_cnt = ozp->z_sync_cnt; nzp->z_is_sa = ozp->z_is_sa; nzp->z_sa_hdl = ozp->z_sa_hdl; bcopy(ozp->z_atime, nzp->z_atime, sizeof (uint64_t) * 2); nzp->z_links = ozp->z_links; nzp->z_size = ozp->z_size; nzp->z_pflags = ozp->z_pflags; nzp->z_uid = ozp->z_uid; nzp->z_gid = ozp->z_gid; nzp->z_mode = ozp->z_mode; /* * Since this is just an idle znode and kmem is already dealing with * memory pressure, release any cached ACL. */ if (ozp->z_acl_cached) { zfs_acl_free(ozp->z_acl_cached); ozp->z_acl_cached = NULL; } sa_set_userp(nzp->z_sa_hdl, nzp); /* * Invalidate the original znode by clearing fields that provide a * pointer back to the znode. Set the low bit of the vfs pointer to * ensure that zfs_znode_move() recognizes the znode as invalid in any * subsequent callback. */ ozp->z_sa_hdl = NULL; POINTER_INVALIDATE(&ozp->z_zfsvfs); /* * Mark the znode. */ nzp->z_moved = 1; ozp->z_moved = (uint8_t)-1; } /*ARGSUSED*/ static kmem_cbrc_t zfs_znode_move(void *buf, void *newbuf, size_t size, void *arg) { znode_t *ozp = buf, *nzp = newbuf; zfsvfs_t *zfsvfs; vnode_t *vp; /* * The znode is on the file system's list of known znodes if the vfs * pointer is valid. We set the low bit of the vfs pointer when freeing * the znode to invalidate it, and the memory patterns written by kmem * (baddcafe and deadbeef) set at least one of the two low bits. A newly * created znode sets the vfs pointer last of all to indicate that the * znode is known and in a valid state to be moved by this function. */ zfsvfs = ozp->z_zfsvfs; if (!POINTER_IS_VALID(zfsvfs)) { ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_invalid); return (KMEM_CBRC_DONT_KNOW); } /* * Close a small window in which it's possible that the filesystem could * be unmounted and freed, and zfsvfs, though valid in the previous * statement, could point to unrelated memory by the time we try to * prevent the filesystem from being unmounted. */ rw_enter(&zfsvfs_lock, RW_WRITER); if (zfsvfs != ozp->z_zfsvfs) { rw_exit(&zfsvfs_lock); ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_recheck1); return (KMEM_CBRC_DONT_KNOW); } /* * If the znode is still valid, then so is the file system. We know that * no valid file system can be freed while we hold zfsvfs_lock, so we * can safely ensure that the filesystem is not and will not be * unmounted. The next statement is equivalent to ZFS_ENTER(). */ rrm_enter(&zfsvfs->z_teardown_lock, RW_READER, FTAG); if (zfsvfs->z_unmounted) { ZFS_EXIT(zfsvfs); rw_exit(&zfsvfs_lock); ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_unmounted); return (KMEM_CBRC_DONT_KNOW); } rw_exit(&zfsvfs_lock); mutex_enter(&zfsvfs->z_znodes_lock); /* * Recheck the vfs pointer in case the znode was removed just before * acquiring the lock. */ if (zfsvfs != ozp->z_zfsvfs) { mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_recheck2); return (KMEM_CBRC_DONT_KNOW); } /* * At this point we know that as long as we hold z_znodes_lock, the * znode cannot be freed and fields within the znode can be safely * accessed. Now, prevent a race with zfs_zget(). */ if (ZFS_OBJ_HOLD_TRYENTER(zfsvfs, ozp->z_id) == 0) { mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_obj_held); return (KMEM_CBRC_LATER); } vp = ZTOV(ozp); if (mutex_tryenter(&vp->v_lock) == 0) { ZFS_OBJ_HOLD_EXIT(zfsvfs, ozp->z_id); mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_vnode_locked); return (KMEM_CBRC_LATER); } /* Only move znodes that are referenced _only_ by the DNLC. */ if (vp->v_count != 1 || !vn_in_dnlc(vp)) { mutex_exit(&vp->v_lock); ZFS_OBJ_HOLD_EXIT(zfsvfs, ozp->z_id); mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_not_only_dnlc); return (KMEM_CBRC_LATER); } /* * The znode is known and in a valid state to move. We're holding the * locks needed to execute the critical section. */ zfs_znode_move_impl(ozp, nzp); mutex_exit(&vp->v_lock); ZFS_OBJ_HOLD_EXIT(zfsvfs, ozp->z_id); list_link_replace(&ozp->z_link_node, &nzp->z_link_node); mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); return (KMEM_CBRC_YES); } #endif /* illumos */ void zfs_znode_init(void) { /* * Initialize zcache */ rw_init(&zfsvfs_lock, NULL, RW_DEFAULT, NULL); ASSERT(znode_cache == NULL); znode_cache = kmem_cache_create("zfs_znode_cache", sizeof (znode_t), 0, zfs_znode_cache_constructor, zfs_znode_cache_destructor, NULL, NULL, NULL, 0); kmem_cache_set_move(znode_cache, zfs_znode_move); } void zfs_znode_fini(void) { #ifdef illumos /* * Cleanup vfs & vnode ops */ zfs_remove_op_tables(); #endif /* * Cleanup zcache */ if (znode_cache) kmem_cache_destroy(znode_cache); znode_cache = NULL; rw_destroy(&zfsvfs_lock); } #ifdef illumos struct vnodeops *zfs_dvnodeops; struct vnodeops *zfs_fvnodeops; struct vnodeops *zfs_symvnodeops; struct vnodeops *zfs_xdvnodeops; struct vnodeops *zfs_evnodeops; struct vnodeops *zfs_sharevnodeops; void zfs_remove_op_tables() { /* * Remove vfs ops */ ASSERT(zfsfstype); (void) vfs_freevfsops_by_type(zfsfstype); zfsfstype = 0; /* * Remove vnode ops */ if (zfs_dvnodeops) vn_freevnodeops(zfs_dvnodeops); if (zfs_fvnodeops) vn_freevnodeops(zfs_fvnodeops); if (zfs_symvnodeops) vn_freevnodeops(zfs_symvnodeops); if (zfs_xdvnodeops) vn_freevnodeops(zfs_xdvnodeops); if (zfs_evnodeops) vn_freevnodeops(zfs_evnodeops); if (zfs_sharevnodeops) vn_freevnodeops(zfs_sharevnodeops); zfs_dvnodeops = NULL; zfs_fvnodeops = NULL; zfs_symvnodeops = NULL; zfs_xdvnodeops = NULL; zfs_evnodeops = NULL; zfs_sharevnodeops = NULL; } extern const fs_operation_def_t zfs_dvnodeops_template[]; extern const fs_operation_def_t zfs_fvnodeops_template[]; extern const fs_operation_def_t zfs_xdvnodeops_template[]; extern const fs_operation_def_t zfs_symvnodeops_template[]; extern const fs_operation_def_t zfs_evnodeops_template[]; extern const fs_operation_def_t zfs_sharevnodeops_template[]; int zfs_create_op_tables() { int error; /* * zfs_dvnodeops can be set if mod_remove() calls mod_installfs() * due to a failure to remove the the 2nd modlinkage (zfs_modldrv). * In this case we just return as the ops vectors are already set up. */ if (zfs_dvnodeops) return (0); error = vn_make_ops(MNTTYPE_ZFS, zfs_dvnodeops_template, &zfs_dvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_fvnodeops_template, &zfs_fvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_symvnodeops_template, &zfs_symvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_xdvnodeops_template, &zfs_xdvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_evnodeops_template, &zfs_evnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_sharevnodeops_template, &zfs_sharevnodeops); return (error); } #endif /* illumos */ int zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx) { zfs_acl_ids_t acl_ids; vattr_t vattr; znode_t *sharezp; znode_t *zp; int error; vattr.va_mask = AT_MODE|AT_UID|AT_GID|AT_TYPE; vattr.va_type = VDIR; vattr.va_mode = S_IFDIR|0555; vattr.va_uid = crgetuid(kcred); vattr.va_gid = crgetgid(kcred); sharezp = kmem_cache_alloc(znode_cache, KM_SLEEP); ASSERT(!POINTER_IS_VALID(sharezp->z_zfsvfs)); sharezp->z_moved = 0; sharezp->z_unlinked = 0; sharezp->z_atime_dirty = 0; sharezp->z_zfsvfs = zfsvfs; sharezp->z_is_sa = zfsvfs->z_use_sa; VERIFY(0 == zfs_acl_ids_create(sharezp, IS_ROOT_NODE, &vattr, kcred, NULL, &acl_ids)); zfs_mknode(sharezp, &vattr, tx, kcred, IS_ROOT_NODE, &zp, &acl_ids); ASSERT3P(zp, ==, sharezp); POINTER_INVALIDATE(&sharezp->z_zfsvfs); error = zap_add(zfsvfs->z_os, MASTER_NODE_OBJ, ZFS_SHARES_DIR, 8, 1, &sharezp->z_id, tx); zfsvfs->z_shares_dir = sharezp->z_id; zfs_acl_ids_free(&acl_ids); sa_handle_destroy(sharezp->z_sa_hdl); kmem_cache_free(znode_cache, sharezp); return (error); } /* * define a couple of values we need available * for both 64 and 32 bit environments. */ #ifndef NBITSMINOR64 #define NBITSMINOR64 32 #endif #ifndef MAXMAJ64 #define MAXMAJ64 0xffffffffUL #endif #ifndef MAXMIN64 #define MAXMIN64 0xffffffffUL #endif /* * Create special expldev for ZFS private use. * Can't use standard expldev since it doesn't do * what we want. The standard expldev() takes a * dev32_t in LP64 and expands it to a long dev_t. * We need an interface that takes a dev32_t in ILP32 * and expands it to a long dev_t. */ static uint64_t zfs_expldev(dev_t dev) { return (((uint64_t)major(dev) << NBITSMINOR64) | minor(dev)); } /* * Special cmpldev for ZFS private use. * Can't use standard cmpldev since it takes * a long dev_t and compresses it to dev32_t in * LP64. We need to do a compaction of a long dev_t * to a dev32_t in ILP32. */ dev_t zfs_cmpldev(uint64_t dev) { return (makedev((dev >> NBITSMINOR64), (dev & MAXMIN64))); } static void zfs_znode_sa_init(zfsvfs_t *zfsvfs, znode_t *zp, dmu_buf_t *db, dmu_object_type_t obj_type, sa_handle_t *sa_hdl) { ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs) || (zfsvfs == zp->z_zfsvfs)); ASSERT(MUTEX_HELD(ZFS_OBJ_MUTEX(zfsvfs, zp->z_id))); mutex_enter(&zp->z_lock); ASSERT(zp->z_sa_hdl == NULL); ASSERT(zp->z_acl_cached == NULL); if (sa_hdl == NULL) { VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, zp, SA_HDL_SHARED, &zp->z_sa_hdl)); } else { zp->z_sa_hdl = sa_hdl; sa_set_userp(sa_hdl, zp); } zp->z_is_sa = (obj_type == DMU_OT_SA) ? B_TRUE : B_FALSE; /* - * Slap on VROOT if we are the root znode + * Slap on VROOT if we are the root znode unless we are the root + * node of a snapshot mounted under .zfs. */ - if (zp->z_id == zfsvfs->z_root) + if (zp->z_id == zfsvfs->z_root && zfsvfs->z_parent == zfsvfs) ZTOV(zp)->v_flag |= VROOT; mutex_exit(&zp->z_lock); vn_exists(ZTOV(zp)); } void zfs_znode_dmu_fini(znode_t *zp) { ASSERT(MUTEX_HELD(ZFS_OBJ_MUTEX(zp->z_zfsvfs, zp->z_id)) || zp->z_unlinked || RW_WRITE_HELD(&zp->z_zfsvfs->z_teardown_inactive_lock)); sa_handle_destroy(zp->z_sa_hdl); zp->z_sa_hdl = NULL; } static void zfs_vnode_forget(vnode_t *vp) { /* copied from insmntque_stddtr */ vp->v_data = NULL; vp->v_op = &dead_vnodeops; vgone(vp); vput(vp); } /* * Construct a new znode/vnode and intialize. * * This does not do a call to dmu_set_user() that is * up to the caller to do, in case you don't want to * return the znode */ static znode_t * zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz, dmu_object_type_t obj_type, sa_handle_t *hdl) { znode_t *zp; vnode_t *vp; uint64_t mode; uint64_t parent; sa_bulk_attr_t bulk[9]; int count = 0; int error; zp = kmem_cache_alloc(znode_cache, KM_SLEEP); KASSERT(curthread->td_vp_reserv > 0, ("zfs_znode_alloc: getnewvnode without any vnodes reserved")); error = getnewvnode("zfs", zfsvfs->z_parent->z_vfs, &zfs_vnodeops, &vp); if (error != 0) { kmem_cache_free(znode_cache, zp); return (NULL); } zp->z_vnode = vp; vp->v_data = zp; ASSERT(zp->z_dirlocks == NULL); ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs)); zp->z_moved = 0; /* * Defer setting z_zfsvfs until the znode is ready to be a candidate for * the zfs_znode_move() callback. */ zp->z_sa_hdl = NULL; zp->z_unlinked = 0; zp->z_atime_dirty = 0; zp->z_mapcnt = 0; zp->z_id = db->db_object; zp->z_blksz = blksz; zp->z_seq = 0x7A4653; zp->z_sync_cnt = 0; vp = ZTOV(zp); zfs_znode_sa_init(zfsvfs, zp, db, obj_type, hdl); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GEN(zfsvfs), NULL, &zp->z_gen, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL, &zp->z_size, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL, &zp->z_links, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PARENT(zfsvfs), NULL, &parent, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &zp->z_atime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &zp->z_uid, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &zp->z_gid, 8); if (sa_bulk_lookup(zp->z_sa_hdl, bulk, count) != 0 || zp->z_gen == 0) { if (hdl == NULL) sa_handle_destroy(zp->z_sa_hdl); zfs_vnode_forget(vp); zp->z_vnode = NULL; kmem_cache_free(znode_cache, zp); return (NULL); } zp->z_mode = mode; vp->v_type = IFTOVT((mode_t)mode); switch (vp->v_type) { case VDIR: zp->z_zn_prefetch = B_TRUE; /* z_prefetch default is enabled */ break; #ifdef illumos case VBLK: case VCHR: { uint64_t rdev; VERIFY(sa_lookup(zp->z_sa_hdl, SA_ZPL_RDEV(zfsvfs), &rdev, sizeof (rdev)) == 0); vp->v_rdev = zfs_cmpldev(rdev); } break; #endif case VFIFO: #ifdef illumos case VSOCK: case VDOOR: #endif vp->v_op = &zfs_fifoops; break; case VREG: if (parent == zfsvfs->z_shares_dir) { ASSERT(zp->z_uid == 0 && zp->z_gid == 0); vp->v_op = &zfs_shareops; } break; #ifdef illumos case VLNK: vn_setops(vp, zfs_symvnodeops); break; default: vn_setops(vp, zfs_evnodeops); break; #endif } mutex_enter(&zfsvfs->z_znodes_lock); list_insert_tail(&zfsvfs->z_all_znodes, zp); membar_producer(); /* * Everything else must be valid before assigning z_zfsvfs makes the * znode eligible for zfs_znode_move(). */ zp->z_zfsvfs = zfsvfs; mutex_exit(&zfsvfs->z_znodes_lock); /* * Acquire vnode lock before making it available to the world. */ vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); VN_LOCK_AREC(vp); if (vp->v_type != VFIFO) VN_LOCK_ASHARE(vp); #ifdef illumos VFS_HOLD(zfsvfs->z_vfs); #endif return (zp); } static uint64_t empty_xattr; static uint64_t pad[4]; static zfs_acl_phys_t acl_phys; /* * Create a new DMU object to hold a zfs znode. * * IN: dzp - parent directory for new znode * vap - file attributes for new znode * tx - dmu transaction id for zap operations * cr - credentials of caller * flag - flags: * IS_ROOT_NODE - new object will be root * IS_XATTR - new object is an attribute * bonuslen - length of bonus buffer * setaclp - File/Dir initial ACL * fuidp - Tracks fuid allocation. * * OUT: zpp - allocated znode * */ void zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr, uint_t flag, znode_t **zpp, zfs_acl_ids_t *acl_ids) { uint64_t crtime[2], atime[2], mtime[2], ctime[2]; uint64_t mode, size, links, parent, pflags; uint64_t dzp_pflags = 0; uint64_t rdev = 0; zfsvfs_t *zfsvfs = dzp->z_zfsvfs; dmu_buf_t *db; timestruc_t now; uint64_t gen, obj; int err; int bonuslen; sa_handle_t *sa_hdl; dmu_object_type_t obj_type; sa_bulk_attr_t sa_attrs[ZPL_END]; int cnt = 0; zfs_acl_locator_cb_t locate = { 0 }; ASSERT(vap && (vap->va_mask & (AT_TYPE|AT_MODE)) == (AT_TYPE|AT_MODE)); if (zfsvfs->z_replay) { obj = vap->va_nodeid; now = vap->va_ctime; /* see zfs_replay_create() */ gen = vap->va_nblocks; /* ditto */ } else { obj = 0; vfs_timestamp(&now); gen = dmu_tx_get_txg(tx); } obj_type = zfsvfs->z_use_sa ? DMU_OT_SA : DMU_OT_ZNODE; bonuslen = (obj_type == DMU_OT_SA) ? DN_MAX_BONUSLEN : ZFS_OLD_ZNODE_PHYS_SIZE; /* * Create a new DMU object. */ /* * There's currently no mechanism for pre-reading the blocks that will * be needed to allocate a new object, so we accept the small chance * that there will be an i/o error and we will fail one of the * assertions below. */ if (vap->va_type == VDIR) { if (zfsvfs->z_replay) { VERIFY0(zap_create_claim_norm(zfsvfs->z_os, obj, zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS, obj_type, bonuslen, tx)); } else { obj = zap_create_norm(zfsvfs->z_os, zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS, obj_type, bonuslen, tx); } } else { if (zfsvfs->z_replay) { VERIFY0(dmu_object_claim(zfsvfs->z_os, obj, DMU_OT_PLAIN_FILE_CONTENTS, 0, obj_type, bonuslen, tx)); } else { obj = dmu_object_alloc(zfsvfs->z_os, DMU_OT_PLAIN_FILE_CONTENTS, 0, obj_type, bonuslen, tx); } } ZFS_OBJ_HOLD_ENTER(zfsvfs, obj); VERIFY(0 == sa_buf_hold(zfsvfs->z_os, obj, NULL, &db)); /* * If this is the root, fix up the half-initialized parent pointer * to reference the just-allocated physical data area. */ if (flag & IS_ROOT_NODE) { dzp->z_id = obj; } else { dzp_pflags = dzp->z_pflags; } /* * If parent is an xattr, so am I. */ if (dzp_pflags & ZFS_XATTR) { flag |= IS_XATTR; } if (zfsvfs->z_use_fuids) pflags = ZFS_ARCHIVE | ZFS_AV_MODIFIED; else pflags = 0; if (vap->va_type == VDIR) { size = 2; /* contents ("." and "..") */ links = (flag & (IS_ROOT_NODE | IS_XATTR)) ? 2 : 1; } else { size = links = 0; } if (vap->va_type == VBLK || vap->va_type == VCHR) { rdev = zfs_expldev(vap->va_rdev); } parent = dzp->z_id; mode = acl_ids->z_mode; if (flag & IS_XATTR) pflags |= ZFS_XATTR; /* * No execs denied will be deterimed when zfs_mode_compute() is called. */ pflags |= acl_ids->z_aclp->z_hints & (ZFS_ACL_TRIVIAL|ZFS_INHERIT_ACE|ZFS_ACL_AUTO_INHERIT| ZFS_ACL_DEFAULTED|ZFS_ACL_PROTECTED); ZFS_TIME_ENCODE(&now, crtime); ZFS_TIME_ENCODE(&now, ctime); if (vap->va_mask & AT_ATIME) { ZFS_TIME_ENCODE(&vap->va_atime, atime); } else { ZFS_TIME_ENCODE(&now, atime); } if (vap->va_mask & AT_MTIME) { ZFS_TIME_ENCODE(&vap->va_mtime, mtime); } else { ZFS_TIME_ENCODE(&now, mtime); } /* Now add in all of the "SA" attributes */ VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, NULL, SA_HDL_SHARED, &sa_hdl)); /* * Setup the array of attributes to be replaced/set on the new file * * order for DMU_OT_ZNODE is critical since it needs to be constructed * in the old znode_phys_t format. Don't change this ordering */ if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zfsvfs), NULL, &crtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zfsvfs), NULL, &gen, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zfsvfs), NULL, &size, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zfsvfs), NULL, &parent, 8); } else { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zfsvfs), NULL, &size, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zfsvfs), NULL, &gen, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zfsvfs), NULL, &acl_ids->z_fuid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zfsvfs), NULL, &acl_ids->z_fgid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zfsvfs), NULL, &parent, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zfsvfs), NULL, &pflags, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zfsvfs), NULL, &crtime, 16); } SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_LINKS(zfsvfs), NULL, &links, 8); if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_XATTR(zfsvfs), NULL, &empty_xattr, 8); } if (obj_type == DMU_OT_ZNODE || (vap->va_type == VBLK || vap->va_type == VCHR)) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_RDEV(zfsvfs), NULL, &rdev, 8); } if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zfsvfs), NULL, &pflags, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zfsvfs), NULL, &acl_ids->z_fuid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zfsvfs), NULL, &acl_ids->z_fgid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PAD(zfsvfs), NULL, pad, sizeof (uint64_t) * 4); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ZNODE_ACL(zfsvfs), NULL, &acl_phys, sizeof (zfs_acl_phys_t)); } else if (acl_ids->z_aclp->z_version >= ZFS_ACL_VERSION_FUID) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_DACL_COUNT(zfsvfs), NULL, &acl_ids->z_aclp->z_acl_count, 8); locate.cb_aclp = acl_ids->z_aclp; SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_DACL_ACES(zfsvfs), zfs_acl_data_locator, &locate, acl_ids->z_aclp->z_acl_bytes); mode = zfs_mode_compute(mode, acl_ids->z_aclp, &pflags, acl_ids->z_fuid, acl_ids->z_fgid); } VERIFY(sa_replace_all_by_template(sa_hdl, sa_attrs, cnt, tx) == 0); if (!(flag & IS_ROOT_NODE)) { *zpp = zfs_znode_alloc(zfsvfs, db, 0, obj_type, sa_hdl); ASSERT(*zpp != NULL); } else { /* * If we are creating the root node, the "parent" we * passed in is the znode for the root. */ *zpp = dzp; (*zpp)->z_sa_hdl = sa_hdl; } (*zpp)->z_pflags = pflags; (*zpp)->z_mode = mode; if (vap->va_mask & AT_XVATTR) zfs_xvattr_set(*zpp, (xvattr_t *)vap, tx); if (obj_type == DMU_OT_ZNODE || acl_ids->z_aclp->z_version < ZFS_ACL_VERSION_FUID) { VERIFY0(zfs_aclset_common(*zpp, acl_ids->z_aclp, cr, tx)); } if (!(flag & IS_ROOT_NODE)) { vnode_t *vp; vp = ZTOV(*zpp); vp->v_vflag |= VV_FORCEINSMQ; err = insmntque(vp, zfsvfs->z_vfs); vp->v_vflag &= ~VV_FORCEINSMQ; KASSERT(err == 0, ("insmntque() failed: error %d", err)); } ZFS_OBJ_HOLD_EXIT(zfsvfs, obj); } /* * Update in-core attributes. It is assumed the caller will be doing an * sa_bulk_update to push the changes out. */ void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx) { xoptattr_t *xoap; xoap = xva_getxoptattr(xvap); ASSERT(xoap); if (XVA_ISSET_REQ(xvap, XAT_CREATETIME)) { uint64_t times[2]; ZFS_TIME_ENCODE(&xoap->xoa_createtime, times); (void) sa_update(zp->z_sa_hdl, SA_ZPL_CRTIME(zp->z_zfsvfs), ×, sizeof (times), tx); XVA_SET_RTN(xvap, XAT_CREATETIME); } if (XVA_ISSET_REQ(xvap, XAT_READONLY)) { ZFS_ATTR_SET(zp, ZFS_READONLY, xoap->xoa_readonly, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_READONLY); } if (XVA_ISSET_REQ(xvap, XAT_HIDDEN)) { ZFS_ATTR_SET(zp, ZFS_HIDDEN, xoap->xoa_hidden, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_HIDDEN); } if (XVA_ISSET_REQ(xvap, XAT_SYSTEM)) { ZFS_ATTR_SET(zp, ZFS_SYSTEM, xoap->xoa_system, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_SYSTEM); } if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE)) { ZFS_ATTR_SET(zp, ZFS_ARCHIVE, xoap->xoa_archive, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_ARCHIVE); } if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) { ZFS_ATTR_SET(zp, ZFS_IMMUTABLE, xoap->xoa_immutable, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_IMMUTABLE); } if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) { ZFS_ATTR_SET(zp, ZFS_NOUNLINK, xoap->xoa_nounlink, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_NOUNLINK); } if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) { ZFS_ATTR_SET(zp, ZFS_APPENDONLY, xoap->xoa_appendonly, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_APPENDONLY); } if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) { ZFS_ATTR_SET(zp, ZFS_NODUMP, xoap->xoa_nodump, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_NODUMP); } if (XVA_ISSET_REQ(xvap, XAT_OPAQUE)) { ZFS_ATTR_SET(zp, ZFS_OPAQUE, xoap->xoa_opaque, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_OPAQUE); } if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) { ZFS_ATTR_SET(zp, ZFS_AV_QUARANTINED, xoap->xoa_av_quarantined, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_AV_QUARANTINED); } if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) { ZFS_ATTR_SET(zp, ZFS_AV_MODIFIED, xoap->xoa_av_modified, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_AV_MODIFIED); } if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) { zfs_sa_set_scanstamp(zp, xvap, tx); XVA_SET_RTN(xvap, XAT_AV_SCANSTAMP); } if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) { ZFS_ATTR_SET(zp, ZFS_REPARSE, xoap->xoa_reparse, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_REPARSE); } if (XVA_ISSET_REQ(xvap, XAT_OFFLINE)) { ZFS_ATTR_SET(zp, ZFS_OFFLINE, xoap->xoa_offline, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_OFFLINE); } if (XVA_ISSET_REQ(xvap, XAT_SPARSE)) { ZFS_ATTR_SET(zp, ZFS_SPARSE, xoap->xoa_sparse, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_SPARSE); } } int zfs_zget(zfsvfs_t *zfsvfs, uint64_t obj_num, znode_t **zpp) { dmu_object_info_t doi; dmu_buf_t *db; znode_t *zp; vnode_t *vp; sa_handle_t *hdl; struct thread *td; int locked; int err; td = curthread; getnewvnode_reserve(1); again: *zpp = NULL; ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num); err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db); if (err) { ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); getnewvnode_drop_reserve(); return (err); } dmu_object_info_from_db(db, &doi); if (doi.doi_bonus_type != DMU_OT_SA && (doi.doi_bonus_type != DMU_OT_ZNODE || (doi.doi_bonus_type == DMU_OT_ZNODE && doi.doi_bonus_size < sizeof (znode_phys_t)))) { sa_buf_rele(db, NULL); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); #ifdef __FreeBSD__ getnewvnode_drop_reserve(); #endif return (SET_ERROR(EINVAL)); } hdl = dmu_buf_get_user(db); if (hdl != NULL) { zp = sa_get_userdata(hdl); /* * Since "SA" does immediate eviction we * should never find a sa handle that doesn't * know about the znode. */ ASSERT3P(zp, !=, NULL); mutex_enter(&zp->z_lock); ASSERT3U(zp->z_id, ==, obj_num); if (zp->z_unlinked) { err = SET_ERROR(ENOENT); } else { vp = ZTOV(zp); *zpp = zp; err = 0; } /* Don't let the vnode disappear after ZFS_OBJ_HOLD_EXIT. */ if (err == 0) VN_HOLD(vp); mutex_exit(&zp->z_lock); sa_buf_rele(db, NULL); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); if (err == 0) { locked = VOP_ISLOCKED(vp); VI_LOCK(vp); if ((vp->v_iflag & VI_DOOMED) != 0 && locked != LK_EXCLUSIVE) { /* * The vnode is doomed and this thread doesn't * hold the exclusive lock on it, so the vnode * must be being reclaimed by another thread. * Otherwise the doomed vnode is being reclaimed * by this thread and zfs_zget is called from * ZIL internals. */ VI_UNLOCK(vp); VN_RELE(vp); goto again; } VI_UNLOCK(vp); } getnewvnode_drop_reserve(); return (err); } /* * Not found create new znode/vnode * but only if file exists. * * There is a small window where zfs_vget() could * find this object while a file create is still in * progress. This is checked for in zfs_znode_alloc() * * if zfs_znode_alloc() fails it will drop the hold on the * bonus buffer. */ zp = zfs_znode_alloc(zfsvfs, db, doi.doi_data_block_size, doi.doi_bonus_type, NULL); if (zp == NULL) { err = SET_ERROR(ENOENT); } else { *zpp = zp; } if (err == 0) { vnode_t *vp = ZTOV(zp); err = insmntque(vp, zfsvfs->z_vfs); if (err == 0) { vp->v_hash = obj_num; VOP_UNLOCK(vp, 0); } else { zp->z_vnode = NULL; zfs_znode_dmu_fini(zp); zfs_znode_free(zp); *zpp = NULL; } } ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); getnewvnode_drop_reserve(); return (err); } int zfs_rezget(znode_t *zp) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; dmu_object_info_t doi; dmu_buf_t *db; vnode_t *vp; uint64_t obj_num = zp->z_id; uint64_t mode, size; sa_bulk_attr_t bulk[8]; int err; int count = 0; uint64_t gen; ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num); mutex_enter(&zp->z_acl_lock); if (zp->z_acl_cached) { zfs_acl_free(zp->z_acl_cached); zp->z_acl_cached = NULL; } mutex_exit(&zp->z_acl_lock); ASSERT(zp->z_sa_hdl == NULL); err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db); if (err) { ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (err); } dmu_object_info_from_db(db, &doi); if (doi.doi_bonus_type != DMU_OT_SA && (doi.doi_bonus_type != DMU_OT_ZNODE || (doi.doi_bonus_type == DMU_OT_ZNODE && doi.doi_bonus_size < sizeof (znode_phys_t)))) { sa_buf_rele(db, NULL); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EINVAL)); } zfs_znode_sa_init(zfsvfs, zp, db, doi.doi_bonus_type, NULL); size = zp->z_size; /* reload cached values */ SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GEN(zfsvfs), NULL, &gen, sizeof (gen)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL, &zp->z_size, sizeof (zp->z_size)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL, &zp->z_links, sizeof (zp->z_links)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, sizeof (zp->z_pflags)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &zp->z_atime, sizeof (zp->z_atime)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &zp->z_uid, sizeof (zp->z_uid)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &zp->z_gid, sizeof (zp->z_gid)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &mode, sizeof (mode)); if (sa_bulk_lookup(zp->z_sa_hdl, bulk, count)) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EIO)); } zp->z_mode = mode; if (gen != zp->z_gen) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EIO)); } /* * It is highly improbable but still quite possible that two * objects in different datasets are created with the same * object numbers and in transaction groups with the same * numbers. znodes corresponding to those objects would * have the same z_id and z_gen, but their other attributes * may be different. * zfs recv -F may replace one of such objects with the other. * As a result file properties recorded in the replaced * object's vnode may no longer match the received object's * properties. At present the only cached property is the * files type recorded in v_type. * So, handle this case by leaving the old vnode and znode * disassociated from the actual object. A new vnode and a * znode will be created if the object is accessed * (e.g. via a look-up). The old vnode and znode will be * recycled when the last vnode reference is dropped. */ vp = ZTOV(zp); if (vp->v_type != IFTOVT((mode_t)zp->z_mode)) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (EIO); } zp->z_unlinked = (zp->z_links == 0); zp->z_blksz = doi.doi_data_block_size; vn_pages_remove(vp, 0, 0); if (zp->z_size != size) vnode_pager_setsize(vp, zp->z_size); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (0); } void zfs_znode_delete(znode_t *zp, dmu_tx_t *tx) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; objset_t *os = zfsvfs->z_os; uint64_t obj = zp->z_id; uint64_t acl_obj = zfs_external_acl(zp); ZFS_OBJ_HOLD_ENTER(zfsvfs, obj); if (acl_obj) { VERIFY(!zp->z_is_sa); VERIFY(0 == dmu_object_free(os, acl_obj, tx)); } VERIFY(0 == dmu_object_free(os, obj, tx)); zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj); zfs_znode_free(zp); } void zfs_zinactive(znode_t *zp) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; uint64_t z_id = zp->z_id; ASSERT(zp->z_sa_hdl); /* * Don't allow a zfs_zget() while were trying to release this znode */ ZFS_OBJ_HOLD_ENTER(zfsvfs, z_id); mutex_enter(&zp->z_lock); /* * If this was the last reference to a file with no links, * remove the file from the file system. */ if (zp->z_unlinked) { mutex_exit(&zp->z_lock); ZFS_OBJ_HOLD_EXIT(zfsvfs, z_id); zfs_rmnode(zp); return; } mutex_exit(&zp->z_lock); zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, z_id); zfs_znode_free(zp); } void zfs_znode_free(znode_t *zp) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; ASSERT(zp->z_sa_hdl == NULL); zp->z_vnode = NULL; mutex_enter(&zfsvfs->z_znodes_lock); POINTER_INVALIDATE(&zp->z_zfsvfs); list_remove(&zfsvfs->z_all_znodes, zp); mutex_exit(&zfsvfs->z_znodes_lock); if (zp->z_acl_cached) { zfs_acl_free(zp->z_acl_cached); zp->z_acl_cached = NULL; } kmem_cache_free(znode_cache, zp); #ifdef illumos VFS_RELE(zfsvfs->z_vfs); #endif } void zfs_tstamp_update_setup(znode_t *zp, uint_t flag, uint64_t mtime[2], uint64_t ctime[2], boolean_t have_tx) { timestruc_t now; vfs_timestamp(&now); if (have_tx) { /* will sa_bulk_update happen really soon? */ zp->z_atime_dirty = 0; zp->z_seq++; } else { zp->z_atime_dirty = 1; } if (flag & AT_ATIME) { ZFS_TIME_ENCODE(&now, zp->z_atime); } if (flag & AT_MTIME) { ZFS_TIME_ENCODE(&now, mtime); if (zp->z_zfsvfs->z_use_fuids) { zp->z_pflags |= (ZFS_ARCHIVE | ZFS_AV_MODIFIED); } } if (flag & AT_CTIME) { ZFS_TIME_ENCODE(&now, ctime); if (zp->z_zfsvfs->z_use_fuids) zp->z_pflags |= ZFS_ARCHIVE; } } /* * Grow the block size for a file. * * IN: zp - znode of file to free data in. * size - requested block size * tx - open transaction. * * NOTE: this function assumes that the znode is write locked. */ void zfs_grow_blocksize(znode_t *zp, uint64_t size, dmu_tx_t *tx) { int error; u_longlong_t dummy; if (size <= zp->z_blksz) return; /* * If the file size is already greater than the current blocksize, * we will not grow. If there is more than one block in a file, * the blocksize cannot change. */ if (zp->z_blksz && zp->z_size > zp->z_blksz) return; error = dmu_object_set_blocksize(zp->z_zfsvfs->z_os, zp->z_id, size, 0, tx); if (error == ENOTSUP) return; ASSERT0(error); /* What blocksize did we actually get? */ dmu_object_size_from_db(sa_get_db(zp->z_sa_hdl), &zp->z_blksz, &dummy); } #ifdef illumos /* * This is a dummy interface used when pvn_vplist_dirty() should *not* * be calling back into the fs for a putpage(). E.g.: when truncating * a file, the pages being "thrown away* don't need to be written out. */ /* ARGSUSED */ static int zfs_no_putpage(vnode_t *vp, page_t *pp, u_offset_t *offp, size_t *lenp, int flags, cred_t *cr) { ASSERT(0); return (0); } #endif /* * Increase the file length * * IN: zp - znode of file to free data in. * end - new end-of-file * * RETURN: 0 on success, error code on failure */ static int zfs_extend(znode_t *zp, uint64_t end) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; dmu_tx_t *tx; rl_t *rl; uint64_t newblksz; int error; /* * We will change zp_size, lock the whole file. */ rl = zfs_range_lock(zp, 0, UINT64_MAX, RL_WRITER); /* * Nothing to do if file already at desired length. */ if (end <= zp->z_size) { zfs_range_unlock(rl); return (0); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); if (end > zp->z_blksz && (!ISP2(zp->z_blksz) || zp->z_blksz < zfsvfs->z_max_blksz)) { /* * We are growing the file past the current block size. */ if (zp->z_blksz > zp->z_zfsvfs->z_max_blksz) { /* * File's blocksize is already larger than the * "recordsize" property. Only let it grow to * the next power of 2. */ ASSERT(!ISP2(zp->z_blksz)); newblksz = MIN(end, 1 << highbit64(zp->z_blksz)); } else { newblksz = MIN(end, zp->z_zfsvfs->z_max_blksz); } dmu_tx_hold_write(tx, zp->z_id, 0, newblksz); } else { newblksz = 0; } error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_range_unlock(rl); return (error); } if (newblksz) zfs_grow_blocksize(zp, newblksz, tx); zp->z_size = end; VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zp->z_zfsvfs), &zp->z_size, sizeof (zp->z_size), tx)); vnode_pager_setsize(ZTOV(zp), end); zfs_range_unlock(rl); dmu_tx_commit(tx); return (0); } /* * Free space in a file. * * IN: zp - znode of file to free data in. * off - start of section to free. * len - length of section to free. * * RETURN: 0 on success, error code on failure */ static int zfs_free_range(znode_t *zp, uint64_t off, uint64_t len) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; rl_t *rl; int error; /* * Lock the range being freed. */ rl = zfs_range_lock(zp, off, len, RL_WRITER); /* * Nothing to do if file already at desired length. */ if (off >= zp->z_size) { zfs_range_unlock(rl); return (0); } if (off + len > zp->z_size) len = zp->z_size - off; error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, off, len); if (error == 0) { /* * In FreeBSD we cannot free block in the middle of a file, * but only at the end of a file, so this code path should * never happen. */ vnode_pager_setsize(ZTOV(zp), off); } zfs_range_unlock(rl); return (error); } /* * Truncate a file * * IN: zp - znode of file to free data in. * end - new end-of-file. * * RETURN: 0 on success, error code on failure */ static int zfs_trunc(znode_t *zp, uint64_t end) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; vnode_t *vp = ZTOV(zp); dmu_tx_t *tx; rl_t *rl; int error; sa_bulk_attr_t bulk[2]; int count = 0; /* * We will change zp_size, lock the whole file. */ rl = zfs_range_lock(zp, 0, UINT64_MAX, RL_WRITER); /* * Nothing to do if file already at desired length. */ if (end >= zp->z_size) { zfs_range_unlock(rl); return (0); } error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, end, -1); if (error) { zfs_range_unlock(rl); return (error); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_range_unlock(rl); return (error); } zp->z_size = end; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL, &zp->z_size, sizeof (zp->z_size)); if (end == 0) { zp->z_pflags &= ~ZFS_SPARSE; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); } VERIFY(sa_bulk_update(zp->z_sa_hdl, bulk, count, tx) == 0); dmu_tx_commit(tx); /* * Clear any mapped pages in the truncated region. This has to * happen outside of the transaction to avoid the possibility of * a deadlock with someone trying to push a page that we are * about to invalidate. */ vnode_pager_setsize(vp, end); zfs_range_unlock(rl); return (0); } /* * Free space in a file * * IN: zp - znode of file to free data in. * off - start of range * len - end of range (0 => EOF) * flag - current file open mode flags. * log - TRUE if this action should be logged * * RETURN: 0 on success, error code on failure */ int zfs_freesp(znode_t *zp, uint64_t off, uint64_t len, int flag, boolean_t log) { vnode_t *vp = ZTOV(zp); dmu_tx_t *tx; zfsvfs_t *zfsvfs = zp->z_zfsvfs; zilog_t *zilog = zfsvfs->z_log; uint64_t mode; uint64_t mtime[2], ctime[2]; sa_bulk_attr_t bulk[3]; int count = 0; int error; if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_MODE(zfsvfs), &mode, sizeof (mode))) != 0) return (error); if (off > zp->z_size) { error = zfs_extend(zp, off+len); if (error == 0 && log) goto log; else return (error); } /* * Check for any locks in the region to be freed. */ if (MANDLOCK(vp, (mode_t)mode)) { uint64_t length = (len ? len : zp->z_size - off); if (error = chklock(vp, FWRITE, off, length, flag, NULL)) return (error); } if (len == 0) { error = zfs_trunc(zp, off); } else { if ((error = zfs_free_range(zp, off, len)) == 0 && off + len > zp->z_size) error = zfs_extend(zp, off+len); } if (error || !log) return (error); log: tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); return (error); } SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, mtime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, ctime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime, B_TRUE); error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx); ASSERT(error == 0); zfs_log_truncate(zilog, tx, TX_TRUNCATE, zp, off, len); dmu_tx_commit(tx); return (0); } void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) { uint64_t moid, obj, sa_obj, version; uint64_t sense = ZFS_CASE_SENSITIVE; uint64_t norm = 0; nvpair_t *elem; int error; int i; znode_t *rootzp = NULL; zfsvfs_t *zfsvfs; vattr_t vattr; znode_t *zp; zfs_acl_ids_t acl_ids; /* * First attempt to create master node. */ /* * In an empty objset, there are no blocks to read and thus * there can be no i/o errors (which we assert below). */ moid = MASTER_NODE_OBJ; error = zap_create_claim(os, moid, DMU_OT_MASTER_NODE, DMU_OT_NONE, 0, tx); ASSERT(error == 0); /* * Set starting attributes. */ version = zfs_zpl_version_map(spa_version(dmu_objset_spa(os))); elem = NULL; while ((elem = nvlist_next_nvpair(zplprops, elem)) != NULL) { /* For the moment we expect all zpl props to be uint64_ts */ uint64_t val; char *name; ASSERT(nvpair_type(elem) == DATA_TYPE_UINT64); VERIFY(nvpair_value_uint64(elem, &val) == 0); name = nvpair_name(elem); if (strcmp(name, zfs_prop_to_name(ZFS_PROP_VERSION)) == 0) { if (val < version) version = val; } else { error = zap_update(os, moid, name, 8, 1, &val, tx); } ASSERT(error == 0); if (strcmp(name, zfs_prop_to_name(ZFS_PROP_NORMALIZE)) == 0) norm = val; else if (strcmp(name, zfs_prop_to_name(ZFS_PROP_CASE)) == 0) sense = val; } ASSERT(version != 0); error = zap_update(os, moid, ZPL_VERSION_STR, 8, 1, &version, tx); /* * Create zap object used for SA attribute registration */ if (version >= ZPL_VERSION_SA) { sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE, DMU_OT_NONE, 0, tx); error = zap_add(os, moid, ZFS_SA_ATTRS, 8, 1, &sa_obj, tx); ASSERT(error == 0); } else { sa_obj = 0; } /* * Create a delete queue. */ obj = zap_create(os, DMU_OT_UNLINKED_SET, DMU_OT_NONE, 0, tx); error = zap_add(os, moid, ZFS_UNLINKED_SET, 8, 1, &obj, tx); ASSERT(error == 0); /* * Create root znode. Create minimal znode/vnode/zfsvfs * to allow zfs_mknode to work. */ VATTR_NULL(&vattr); vattr.va_mask = AT_MODE|AT_UID|AT_GID|AT_TYPE; vattr.va_type = VDIR; vattr.va_mode = S_IFDIR|0755; vattr.va_uid = crgetuid(cr); vattr.va_gid = crgetgid(cr); zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP); rootzp = kmem_cache_alloc(znode_cache, KM_SLEEP); ASSERT(!POINTER_IS_VALID(rootzp->z_zfsvfs)); rootzp->z_moved = 0; rootzp->z_unlinked = 0; rootzp->z_atime_dirty = 0; rootzp->z_is_sa = USE_SA(version, os); zfsvfs->z_os = os; zfsvfs->z_parent = zfsvfs; zfsvfs->z_version = version; zfsvfs->z_use_fuids = USE_FUIDS(version, os); zfsvfs->z_use_sa = USE_SA(version, os); zfsvfs->z_norm = norm; error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END, &zfsvfs->z_attr_table); ASSERT(error == 0); /* * Fold case on file systems that are always or sometimes case * insensitive. */ if (sense == ZFS_CASE_INSENSITIVE || sense == ZFS_CASE_MIXED) zfsvfs->z_norm |= U8_TEXTPREP_TOUPPER; mutex_init(&zfsvfs->z_znodes_lock, NULL, MUTEX_DEFAULT, NULL); list_create(&zfsvfs->z_all_znodes, sizeof (znode_t), offsetof(znode_t, z_link_node)); for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_init(&zfsvfs->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL); rootzp->z_zfsvfs = zfsvfs; VERIFY(0 == zfs_acl_ids_create(rootzp, IS_ROOT_NODE, &vattr, cr, NULL, &acl_ids)); zfs_mknode(rootzp, &vattr, tx, cr, IS_ROOT_NODE, &zp, &acl_ids); ASSERT3P(zp, ==, rootzp); error = zap_add(os, moid, ZFS_ROOT_OBJ, 8, 1, &rootzp->z_id, tx); ASSERT(error == 0); zfs_acl_ids_free(&acl_ids); POINTER_INVALIDATE(&rootzp->z_zfsvfs); sa_handle_destroy(rootzp->z_sa_hdl); kmem_cache_free(znode_cache, rootzp); /* * Create shares directory */ error = zfs_create_share_dir(zfsvfs, tx); ASSERT(error == 0); for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_destroy(&zfsvfs->z_hold_mtx[i]); kmem_free(zfsvfs, sizeof (zfsvfs_t)); } #endif /* _KERNEL */ static int zfs_sa_setup(objset_t *osp, sa_attr_type_t **sa_table) { uint64_t sa_obj = 0; int error; error = zap_lookup(osp, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj); if (error != 0 && error != ENOENT) return (error); error = sa_setup(osp, sa_obj, zfs_attr_table, ZPL_END, sa_table); return (error); } static int zfs_grab_sa_handle(objset_t *osp, uint64_t obj, sa_handle_t **hdlp, dmu_buf_t **db, void *tag) { dmu_object_info_t doi; int error; if ((error = sa_buf_hold(osp, obj, tag, db)) != 0) return (error); dmu_object_info_from_db(*db, &doi); if ((doi.doi_bonus_type != DMU_OT_SA && doi.doi_bonus_type != DMU_OT_ZNODE) || doi.doi_bonus_type == DMU_OT_ZNODE && doi.doi_bonus_size < sizeof (znode_phys_t)) { sa_buf_rele(*db, tag); return (SET_ERROR(ENOTSUP)); } error = sa_handle_get(osp, obj, NULL, SA_HDL_PRIVATE, hdlp); if (error != 0) { sa_buf_rele(*db, tag); return (error); } return (0); } void zfs_release_sa_handle(sa_handle_t *hdl, dmu_buf_t *db, void *tag) { sa_handle_destroy(hdl); sa_buf_rele(db, tag); } /* * Given an object number, return its parent object number and whether * or not the object is an extended attribute directory. */ static int zfs_obj_to_pobj(objset_t *osp, sa_handle_t *hdl, sa_attr_type_t *sa_table, uint64_t *pobjp, int *is_xattrdir) { uint64_t parent; uint64_t pflags; uint64_t mode; uint64_t parent_mode; sa_bulk_attr_t bulk[3]; sa_handle_t *sa_hdl; dmu_buf_t *sa_db; int count = 0; int error; SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_PARENT], NULL, &parent, sizeof (parent)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_FLAGS], NULL, &pflags, sizeof (pflags)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_MODE], NULL, &mode, sizeof (mode)); if ((error = sa_bulk_lookup(hdl, bulk, count)) != 0) return (error); /* * When a link is removed its parent pointer is not changed and will * be invalid. There are two cases where a link is removed but the * file stays around, when it goes to the delete queue and when there * are additional links. */ error = zfs_grab_sa_handle(osp, parent, &sa_hdl, &sa_db, FTAG); if (error != 0) return (error); error = sa_lookup(sa_hdl, ZPL_MODE, &parent_mode, sizeof (parent_mode)); zfs_release_sa_handle(sa_hdl, sa_db, FTAG); if (error != 0) return (error); *is_xattrdir = ((pflags & ZFS_XATTR) != 0) && S_ISDIR(mode); /* * Extended attributes can be applied to files, directories, etc. * Otherwise the parent must be a directory. */ if (!*is_xattrdir && !S_ISDIR(parent_mode)) return (SET_ERROR(EINVAL)); *pobjp = parent; return (0); } /* * Given an object number, return some zpl level statistics */ static int zfs_obj_to_stats_impl(sa_handle_t *hdl, sa_attr_type_t *sa_table, zfs_stat_t *sb) { sa_bulk_attr_t bulk[4]; int count = 0; SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_MODE], NULL, &sb->zs_mode, sizeof (sb->zs_mode)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_GEN], NULL, &sb->zs_gen, sizeof (sb->zs_gen)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_LINKS], NULL, &sb->zs_links, sizeof (sb->zs_links)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_CTIME], NULL, &sb->zs_ctime, sizeof (sb->zs_ctime)); return (sa_bulk_lookup(hdl, bulk, count)); } static int zfs_obj_to_path_impl(objset_t *osp, uint64_t obj, sa_handle_t *hdl, sa_attr_type_t *sa_table, char *buf, int len) { sa_handle_t *sa_hdl; sa_handle_t *prevhdl = NULL; dmu_buf_t *prevdb = NULL; dmu_buf_t *sa_db = NULL; char *path = buf + len - 1; int error; *path = '\0'; sa_hdl = hdl; for (;;) { uint64_t pobj; char component[MAXNAMELEN + 2]; size_t complen; int is_xattrdir; if (prevdb) zfs_release_sa_handle(prevhdl, prevdb, FTAG); if ((error = zfs_obj_to_pobj(osp, sa_hdl, sa_table, &pobj, &is_xattrdir)) != 0) break; if (pobj == obj) { if (path[0] != '/') *--path = '/'; break; } component[0] = '/'; if (is_xattrdir) { (void) sprintf(component + 1, ""); } else { error = zap_value_search(osp, pobj, obj, ZFS_DIRENT_OBJ(-1ULL), component + 1); if (error != 0) break; } complen = strlen(component); path -= complen; ASSERT(path >= buf); bcopy(component, path, complen); obj = pobj; if (sa_hdl != hdl) { prevhdl = sa_hdl; prevdb = sa_db; } error = zfs_grab_sa_handle(osp, obj, &sa_hdl, &sa_db, FTAG); if (error != 0) { sa_hdl = prevhdl; sa_db = prevdb; break; } } if (sa_hdl != NULL && sa_hdl != hdl) { ASSERT(sa_db != NULL); zfs_release_sa_handle(sa_hdl, sa_db, FTAG); } if (error == 0) (void) memmove(buf, path, buf + len - path); return (error); } int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len) { sa_attr_type_t *sa_table; sa_handle_t *hdl; dmu_buf_t *db; int error; error = zfs_sa_setup(osp, &sa_table); if (error != 0) return (error); error = zfs_grab_sa_handle(osp, obj, &hdl, &db, FTAG); if (error != 0) return (error); error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len); zfs_release_sa_handle(hdl, db, FTAG); return (error); } int zfs_obj_to_stats(objset_t *osp, uint64_t obj, zfs_stat_t *sb, char *buf, int len) { char *path = buf + len - 1; sa_attr_type_t *sa_table; sa_handle_t *hdl; dmu_buf_t *db; int error; *path = '\0'; error = zfs_sa_setup(osp, &sa_table); if (error != 0) return (error); error = zfs_grab_sa_handle(osp, obj, &hdl, &db, FTAG); if (error != 0) return (error); error = zfs_obj_to_stats_impl(hdl, sa_table, sb); if (error != 0) { zfs_release_sa_handle(hdl, db, FTAG); return (error); } error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len); zfs_release_sa_handle(hdl, db, FTAG); return (error); }