4 * The contents of this file are subject to the terms of the
5 * Common Development and Distribution License (the "License").
6 * You may not use this file except in compliance with the License.
8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 * or http://www.opensolaris.org/os/licensing.
10 * See the License for the specific language governing permissions
11 * and limitations under the License.
13 * When distributing Covered Code, include this CDDL HEADER in each
14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 * If applicable, add the following below this CDDL HEADER, with the
16 * fields enclosed by brackets "[]" replaced with your own identifying
17 * information: Portions Copyright [yyyy] [name of copyright owner]
22 * Copyright 2010 Sun Microsystems, Inc. All rights reserved.
23 * Use is subject to license terms.
26 #include <sys/types.h>
27 #include <sys/param.h>
28 #include <sys/systm.h>
29 #include <sys/sysmacros.h>
30 #include <sys/cmn_err.h>
32 #include <sys/thread.h>
34 #include <sys/fcntl.h>
36 #include <sys/fs/zfs.h>
37 #include <sys/zfs_znode.h>
38 #include <sys/zfs_dir.h>
39 #include <sys/zfs_acl.h>
40 #include <sys/zfs_fuid.h>
43 #include <sys/byteorder.h>
47 #include <sys/atomic.h>
51 * Functions to replay ZFS intent log (ZIL) records
52 * The functions are called through a function vector (zfs_replay_vector)
53 * which is indexed by the transaction type.
57 zfs_init_vattr(vattr_t *vap, uint64_t mask, uint64_t mode,
58 uint64_t uid, uint64_t gid, uint64_t rdev, uint64_t nodeid)
60 bzero(vap, sizeof (*vap));
61 vap->va_mask = (uint_t)mask;
62 vap->va_type = IFTOVT(mode);
63 vap->va_mode = mode & MODEMASK;
64 vap->va_uid = (uid_t)(IS_EPHEMERAL(uid)) ? -1 : uid;
65 vap->va_gid = (gid_t)(IS_EPHEMERAL(gid)) ? -1 : gid;
66 vap->va_rdev = zfs_cmpldev(rdev);
67 vap->va_nodeid = nodeid;
72 zfs_replay_error(zfsvfs_t *zfsvfs, lr_t *lr, boolean_t byteswap)
78 zfs_replay_xvattr(lr_attr_t *lrattr, xvattr_t *xvap)
80 xoptattr_t *xoap = NULL;
87 xvap->xva_vattr.va_mask |= AT_XVATTR;
88 if ((xoap = xva_getxoptattr(xvap)) == NULL) {
89 xvap->xva_vattr.va_mask &= ~AT_XVATTR; /* shouldn't happen */
93 ASSERT(lrattr->lr_attr_masksize == xvap->xva_mapsize);
95 bitmap = &lrattr->lr_attr_bitmap;
96 for (i = 0; i != lrattr->lr_attr_masksize; i++, bitmap++)
97 xvap->xva_reqattrmap[i] = *bitmap;
99 attrs = (uint64_t *)(lrattr + lrattr->lr_attr_masksize - 1);
101 scanstamp = (caddr_t)(crtime + 2);
103 if (XVA_ISSET_REQ(xvap, XAT_HIDDEN))
104 xoap->xoa_hidden = ((*attrs & XAT0_HIDDEN) != 0);
105 if (XVA_ISSET_REQ(xvap, XAT_SYSTEM))
106 xoap->xoa_system = ((*attrs & XAT0_SYSTEM) != 0);
107 if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE))
108 xoap->xoa_archive = ((*attrs & XAT0_ARCHIVE) != 0);
109 if (XVA_ISSET_REQ(xvap, XAT_READONLY))
110 xoap->xoa_readonly = ((*attrs & XAT0_READONLY) != 0);
111 if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE))
112 xoap->xoa_immutable = ((*attrs & XAT0_IMMUTABLE) != 0);
113 if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK))
114 xoap->xoa_nounlink = ((*attrs & XAT0_NOUNLINK) != 0);
115 if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY))
116 xoap->xoa_appendonly = ((*attrs & XAT0_APPENDONLY) != 0);
117 if (XVA_ISSET_REQ(xvap, XAT_NODUMP))
118 xoap->xoa_nodump = ((*attrs & XAT0_NODUMP) != 0);
119 if (XVA_ISSET_REQ(xvap, XAT_OPAQUE))
120 xoap->xoa_opaque = ((*attrs & XAT0_OPAQUE) != 0);
121 if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED))
122 xoap->xoa_av_modified = ((*attrs & XAT0_AV_MODIFIED) != 0);
123 if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED))
124 xoap->xoa_av_quarantined =
125 ((*attrs & XAT0_AV_QUARANTINED) != 0);
126 if (XVA_ISSET_REQ(xvap, XAT_CREATETIME))
127 ZFS_TIME_DECODE(&xoap->xoa_createtime, crtime);
128 if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP))
129 bcopy(scanstamp, xoap->xoa_av_scanstamp, AV_SCANSTAMP_SZ);
130 if (XVA_ISSET_REQ(xvap, XAT_REPARSE))
131 xoap->xoa_reparse = ((*attrs & XAT0_REPARSE) != 0);
135 zfs_replay_domain_cnt(uint64_t uid, uint64_t gid)
141 uid_idx = FUID_INDEX(uid);
142 gid_idx = FUID_INDEX(gid);
145 if (gid_idx > 0 && gid_idx != uid_idx)
152 zfs_replay_fuid_domain_common(zfs_fuid_info_t *fuid_infop, void *start,
157 for (i = 0; i != domcnt; i++) {
158 fuid_infop->z_domain_table[i] = start;
159 start = (caddr_t)start + strlen(start) + 1;
166 * Set the uid/gid in the fuid_info structure.
169 zfs_replay_fuid_ugid(zfs_fuid_info_t *fuid_infop, uint64_t uid, uint64_t gid)
172 * If owner or group are log specific FUIDs then slurp up
173 * domain information and build zfs_fuid_info_t
175 if (IS_EPHEMERAL(uid))
176 fuid_infop->z_fuid_owner = uid;
178 if (IS_EPHEMERAL(gid))
179 fuid_infop->z_fuid_group = gid;
183 * Load fuid domains into fuid_info_t
185 static zfs_fuid_info_t *
186 zfs_replay_fuid_domain(void *buf, void **end, uint64_t uid, uint64_t gid)
190 zfs_fuid_info_t *fuid_infop;
192 fuid_infop = zfs_fuid_info_alloc();
194 domcnt = zfs_replay_domain_cnt(uid, gid);
199 fuid_infop->z_domain_table =
200 kmem_zalloc(domcnt * sizeof (char **), KM_SLEEP);
202 zfs_replay_fuid_ugid(fuid_infop, uid, gid);
204 fuid_infop->z_domain_cnt = domcnt;
205 *end = zfs_replay_fuid_domain_common(fuid_infop, buf, domcnt);
210 * load zfs_fuid_t's and fuid_domains into fuid_info_t
212 static zfs_fuid_info_t *
213 zfs_replay_fuids(void *start, void **end, int idcnt, int domcnt, uint64_t uid,
216 uint64_t *log_fuid = (uint64_t *)start;
217 zfs_fuid_info_t *fuid_infop;
220 fuid_infop = zfs_fuid_info_alloc();
221 fuid_infop->z_domain_cnt = domcnt;
223 fuid_infop->z_domain_table =
224 kmem_zalloc(domcnt * sizeof (char **), KM_SLEEP);
226 for (i = 0; i != idcnt; i++) {
229 zfuid = kmem_alloc(sizeof (zfs_fuid_t), KM_SLEEP);
230 zfuid->z_logfuid = *log_fuid;
233 list_insert_tail(&fuid_infop->z_fuids, zfuid);
237 zfs_replay_fuid_ugid(fuid_infop, uid, gid);
239 *end = zfs_replay_fuid_domain_common(fuid_infop, log_fuid, domcnt);
244 zfs_replay_swap_attrs(lr_attr_t *lrattr)
246 /* swap the lr_attr structure */
247 byteswap_uint32_array(lrattr, sizeof (*lrattr));
248 /* swap the bitmap */
249 byteswap_uint32_array(lrattr + 1, (lrattr->lr_attr_masksize - 1) *
251 /* swap the attributes, create time + 64 bit word for attributes */
252 byteswap_uint64_array((caddr_t)(lrattr + 1) + (sizeof (uint32_t) *
253 (lrattr->lr_attr_masksize - 1)), 3 * sizeof (uint64_t));
257 * Replay file create with optional ACL, xvattr information as well
258 * as option FUID information.
261 zfs_replay_create_acl(zfsvfs_t *zfsvfs,
262 lr_acl_create_t *lracl, boolean_t byteswap)
264 char *name = NULL; /* location determined later */
265 lr_create_t *lr = (lr_create_t *)lracl;
270 vsecattr_t vsec = { 0 };
278 txtype = (lr->lr_common.lrc_txtype & ~TX_CI);
280 byteswap_uint64_array(lracl, sizeof (*lracl));
281 if (txtype == TX_CREATE_ACL_ATTR ||
282 txtype == TX_MKDIR_ACL_ATTR) {
283 lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
284 zfs_replay_swap_attrs(lrattr);
285 xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
288 aclstart = (caddr_t)(lracl + 1) + xvatlen;
289 zfs_ace_byteswap(aclstart, lracl->lr_acl_bytes, B_FALSE);
291 if (lracl->lr_fuidcnt) {
292 byteswap_uint64_array((caddr_t)aclstart +
293 ZIL_ACE_LENGTH(lracl->lr_acl_bytes),
294 lracl->lr_fuidcnt * sizeof (uint64_t));
298 if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
302 zfs_init_vattr(&xva.xva_vattr, AT_TYPE | AT_MODE | AT_UID | AT_GID,
303 lr->lr_mode, lr->lr_uid, lr->lr_gid, lr->lr_rdev, lr->lr_foid);
306 * All forms of zfs create (create, mkdir, mkxattrdir, symlink)
307 * eventually end up in zfs_mknode(), which assigns the object's
308 * creation time and generation number. The generic VOP_CREATE()
309 * doesn't have either concept, so we smuggle the values inside
310 * the vattr's otherwise unused va_ctime and va_nblocks fields.
312 ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_crtime);
313 xva.xva_vattr.va_nblocks = lr->lr_gen;
315 error = dmu_object_info(zfsvfs->z_os, lr->lr_foid, NULL);
319 if (lr->lr_common.lrc_txtype & TX_CI)
323 aclstart = (caddr_t)(lracl + 1);
324 fuidstart = (caddr_t)aclstart +
325 ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
326 zfsvfs->z_fuid_replay = zfs_replay_fuids(fuidstart,
327 (void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
328 lr->lr_uid, lr->lr_gid);
330 case TX_CREATE_ACL_ATTR:
332 lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
333 xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
334 xva.xva_vattr.va_mask |= AT_XVATTR;
335 zfs_replay_xvattr(lrattr, &xva);
337 vsec.vsa_mask = VSA_ACE | VSA_ACE_ACLFLAGS;
338 vsec.vsa_aclentp = (caddr_t)(lracl + 1) + xvatlen;
339 vsec.vsa_aclcnt = lracl->lr_aclcnt;
340 vsec.vsa_aclentsz = lracl->lr_acl_bytes;
341 vsec.vsa_aclflags = lracl->lr_acl_flags;
342 if (zfsvfs->z_fuid_replay == NULL) {
343 fuidstart = (caddr_t)(lracl + 1) + xvatlen +
344 ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
345 zfsvfs->z_fuid_replay =
346 zfs_replay_fuids(fuidstart,
347 (void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
348 lr->lr_uid, lr->lr_gid);
351 error = VOP_CREATE(ZTOV(dzp), name, &xva.xva_vattr,
352 0, 0, &vp, kcred, vflg, NULL, &vsec);
355 aclstart = (caddr_t)(lracl + 1);
356 fuidstart = (caddr_t)aclstart +
357 ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
358 zfsvfs->z_fuid_replay = zfs_replay_fuids(fuidstart,
359 (void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
360 lr->lr_uid, lr->lr_gid);
362 case TX_MKDIR_ACL_ATTR:
364 lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
365 xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
366 zfs_replay_xvattr(lrattr, &xva);
368 vsec.vsa_mask = VSA_ACE | VSA_ACE_ACLFLAGS;
369 vsec.vsa_aclentp = (caddr_t)(lracl + 1) + xvatlen;
370 vsec.vsa_aclcnt = lracl->lr_aclcnt;
371 vsec.vsa_aclentsz = lracl->lr_acl_bytes;
372 vsec.vsa_aclflags = lracl->lr_acl_flags;
373 if (zfsvfs->z_fuid_replay == NULL) {
374 fuidstart = (caddr_t)(lracl + 1) + xvatlen +
375 ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
376 zfsvfs->z_fuid_replay =
377 zfs_replay_fuids(fuidstart,
378 (void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
379 lr->lr_uid, lr->lr_gid);
381 error = VOP_MKDIR(ZTOV(dzp), name, &xva.xva_vattr,
382 &vp, kcred, NULL, vflg, &vsec);
389 if (error == 0 && vp != NULL)
394 if (zfsvfs->z_fuid_replay)
395 zfs_fuid_info_free(zfsvfs->z_fuid_replay);
396 zfsvfs->z_fuid_replay = NULL;
402 zfs_replay_create(zfsvfs_t *zfsvfs, lr_create_t *lr, boolean_t byteswap)
404 char *name = NULL; /* location determined later */
405 char *link; /* symlink content follows name */
410 size_t lrsize = sizeof (lr_create_t);
417 txtype = (lr->lr_common.lrc_txtype & ~TX_CI);
419 byteswap_uint64_array(lr, sizeof (*lr));
420 if (txtype == TX_CREATE_ATTR || txtype == TX_MKDIR_ATTR)
421 zfs_replay_swap_attrs((lr_attr_t *)(lr + 1));
425 if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
429 zfs_init_vattr(&xva.xva_vattr, AT_TYPE | AT_MODE | AT_UID | AT_GID,
430 lr->lr_mode, lr->lr_uid, lr->lr_gid, lr->lr_rdev, lr->lr_foid);
433 * All forms of zfs create (create, mkdir, mkxattrdir, symlink)
434 * eventually end up in zfs_mknode(), which assigns the object's
435 * creation time and generation number. The generic VOP_CREATE()
436 * doesn't have either concept, so we smuggle the values inside
437 * the vattr's otherwise unused va_ctime and va_nblocks fields.
439 ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_crtime);
440 xva.xva_vattr.va_nblocks = lr->lr_gen;
442 error = dmu_object_info(zfsvfs->z_os, lr->lr_foid, NULL);
446 if (lr->lr_common.lrc_txtype & TX_CI)
450 * Symlinks don't have fuid info, and CIFS never creates
453 * The _ATTR versions will grab the fuid info in their subcases.
455 if ((int)lr->lr_common.lrc_txtype != TX_SYMLINK &&
456 (int)lr->lr_common.lrc_txtype != TX_MKDIR_ATTR &&
457 (int)lr->lr_common.lrc_txtype != TX_CREATE_ATTR) {
459 zfsvfs->z_fuid_replay =
460 zfs_replay_fuid_domain(start, &start,
461 lr->lr_uid, lr->lr_gid);
466 lrattr = (lr_attr_t *)(caddr_t)(lr + 1);
467 xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
468 zfs_replay_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), &xva);
469 start = (caddr_t)(lr + 1) + xvatlen;
470 zfsvfs->z_fuid_replay =
471 zfs_replay_fuid_domain(start, &start,
472 lr->lr_uid, lr->lr_gid);
473 name = (char *)start;
478 name = (char *)start;
480 error = VOP_CREATE(ZTOV(dzp), name, &xva.xva_vattr,
481 0, 0, &vp, kcred, vflg, NULL, NULL);
484 lrattr = (lr_attr_t *)(caddr_t)(lr + 1);
485 xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
486 zfs_replay_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), &xva);
487 start = (caddr_t)(lr + 1) + xvatlen;
488 zfsvfs->z_fuid_replay =
489 zfs_replay_fuid_domain(start, &start,
490 lr->lr_uid, lr->lr_gid);
491 name = (char *)start;
496 name = (char *)(lr + 1);
498 error = VOP_MKDIR(ZTOV(dzp), name, &xva.xva_vattr,
499 &vp, kcred, NULL, vflg, NULL);
502 error = zfs_make_xattrdir(dzp, &xva.xva_vattr, &vp, kcred);
505 name = (char *)(lr + 1);
506 link = name + strlen(name) + 1;
507 error = VOP_SYMLINK(ZTOV(dzp), name, &xva.xva_vattr,
508 link, kcred, NULL, vflg);
515 if (error == 0 && vp != NULL)
520 if (zfsvfs->z_fuid_replay)
521 zfs_fuid_info_free(zfsvfs->z_fuid_replay);
522 zfsvfs->z_fuid_replay = NULL;
527 zfs_replay_remove(zfsvfs_t *zfsvfs, lr_remove_t *lr, boolean_t byteswap)
529 char *name = (char *)(lr + 1); /* name follows lr_remove_t */
535 byteswap_uint64_array(lr, sizeof (*lr));
537 if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
540 if (lr->lr_common.lrc_txtype & TX_CI)
543 switch ((int)lr->lr_common.lrc_txtype) {
545 error = VOP_REMOVE(ZTOV(dzp), name, kcred, NULL, vflg);
548 error = VOP_RMDIR(ZTOV(dzp), name, NULL, kcred, NULL, vflg);
560 zfs_replay_link(zfsvfs_t *zfsvfs, lr_link_t *lr, boolean_t byteswap)
562 char *name = (char *)(lr + 1); /* name follows lr_link_t */
568 byteswap_uint64_array(lr, sizeof (*lr));
570 if ((error = zfs_zget(zfsvfs, lr->lr_doid, &dzp)) != 0)
573 if ((error = zfs_zget(zfsvfs, lr->lr_link_obj, &zp)) != 0) {
578 if (lr->lr_common.lrc_txtype & TX_CI)
581 error = VOP_LINK(ZTOV(dzp), ZTOV(zp), name, kcred, NULL, vflg);
590 zfs_replay_rename(zfsvfs_t *zfsvfs, lr_rename_t *lr, boolean_t byteswap)
592 char *sname = (char *)(lr + 1); /* sname and tname follow lr_rename_t */
593 char *tname = sname + strlen(sname) + 1;
594 znode_t *sdzp, *tdzp;
599 byteswap_uint64_array(lr, sizeof (*lr));
601 if ((error = zfs_zget(zfsvfs, lr->lr_sdoid, &sdzp)) != 0)
604 if ((error = zfs_zget(zfsvfs, lr->lr_tdoid, &tdzp)) != 0) {
609 if (lr->lr_common.lrc_txtype & TX_CI)
612 error = VOP_RENAME(ZTOV(sdzp), sname, ZTOV(tdzp), tname, kcred,
622 zfs_replay_write(zfsvfs_t *zfsvfs, lr_write_t *lr, boolean_t byteswap)
624 char *data = (char *)(lr + 1); /* data follows lr_write_t */
628 uint64_t orig_eof, eod, offset, length;
631 byteswap_uint64_array(lr, sizeof (*lr));
633 if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0) {
635 * As we can log writes out of order, it's possible the
636 * file has been removed. In this case just drop the write
637 * and return success.
644 offset = lr->lr_offset;
645 length = lr->lr_length;
646 eod = offset + length; /* end of data for this write */
648 orig_eof = zp->z_size;
650 /* If it's a dmu_sync() block, write the whole block */
651 if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
652 uint64_t blocksize = BP_GET_LSIZE(&lr->lr_blkptr);
653 if (length < blocksize) {
654 offset -= offset % blocksize;
659 error = vn_rdwr(UIO_WRITE, ZTOV(zp), data, length, offset,
660 UIO_SYSSPACE, 0, RLIM64_INFINITY, kcred, &resid);
663 * This may be a write from a dmu_sync() for a whole block,
664 * and may extend beyond the current end of the file.
665 * We can't just replay what was written for this TX_WRITE as
666 * a future TX_WRITE2 may extend the eof and the data for that
667 * write needs to be there. So we write the whole block and
670 if (orig_eof < zp->z_size) /* file length grew ? */
679 * TX_WRITE2 are only generated when dmu_sync() returns EALREADY
680 * meaning the pool block is already being synced. So now that we always write
681 * out full blocks, all we have to do is expand the eof if
685 zfs_replay_write2(zfsvfs_t *zfsvfs, lr_write_t *lr, boolean_t byteswap)
692 byteswap_uint64_array(lr, sizeof (*lr));
694 if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
697 end = lr->lr_offset + lr->lr_length;
698 if (end > zp->z_size) {
699 ASSERT3U(end - zp->z_size, <, zp->z_blksz);
709 zfs_replay_truncate(zfsvfs_t *zfsvfs, lr_truncate_t *lr, boolean_t byteswap)
716 byteswap_uint64_array(lr, sizeof (*lr));
718 if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
721 bzero(&fl, sizeof (fl));
724 fl.l_start = lr->lr_offset;
725 fl.l_len = lr->lr_length;
727 error = VOP_SPACE(ZTOV(zp), F_FREESP, &fl, FWRITE | FOFFMAX,
728 lr->lr_offset, kcred, NULL);
736 zfs_replay_setattr(zfsvfs_t *zfsvfs, lr_setattr_t *lr, boolean_t byteswap)
740 vattr_t *vap = &xva.xva_vattr;
746 byteswap_uint64_array(lr, sizeof (*lr));
748 if ((lr->lr_mask & AT_XVATTR) &&
749 zfsvfs->z_version >= ZPL_VERSION_INITIAL)
750 zfs_replay_swap_attrs((lr_attr_t *)(lr + 1));
753 if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
756 zfs_init_vattr(vap, lr->lr_mask, lr->lr_mode,
757 lr->lr_uid, lr->lr_gid, 0, lr->lr_foid);
759 vap->va_size = lr->lr_size;
760 ZFS_TIME_DECODE(&vap->va_atime, lr->lr_atime);
761 ZFS_TIME_DECODE(&vap->va_mtime, lr->lr_mtime);
764 * Fill in xvattr_t portions if necessary.
767 start = (lr_setattr_t *)(lr + 1);
768 if (vap->va_mask & AT_XVATTR) {
769 zfs_replay_xvattr((lr_attr_t *)start, &xva);
770 start = (caddr_t)start +
771 ZIL_XVAT_SIZE(((lr_attr_t *)start)->lr_attr_masksize);
773 xva.xva_vattr.va_mask &= ~AT_XVATTR;
775 zfsvfs->z_fuid_replay = zfs_replay_fuid_domain(start, &start,
776 lr->lr_uid, lr->lr_gid);
778 error = VOP_SETATTR(ZTOV(zp), vap, 0, kcred, NULL);
780 zfs_fuid_info_free(zfsvfs->z_fuid_replay);
781 zfsvfs->z_fuid_replay = NULL;
788 zfs_replay_acl_v0(zfsvfs_t *zfsvfs, lr_acl_v0_t *lr, boolean_t byteswap)
790 ace_t *ace = (ace_t *)(lr + 1); /* ace array follows lr_acl_t */
796 byteswap_uint64_array(lr, sizeof (*lr));
797 zfs_oldace_byteswap(ace, lr->lr_aclcnt);
800 if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
803 bzero(&vsa, sizeof (vsa));
804 vsa.vsa_mask = VSA_ACE | VSA_ACECNT;
805 vsa.vsa_aclcnt = lr->lr_aclcnt;
806 vsa.vsa_aclentsz = sizeof (ace_t) * vsa.vsa_aclcnt;
807 vsa.vsa_aclflags = 0;
808 vsa.vsa_aclentp = ace;
810 error = VOP_SETSECATTR(ZTOV(zp), &vsa, 0, kcred, NULL);
818 * Replaying ACLs is complicated by FUID support.
819 * The log record may contain some optional data
820 * to be used for replaying FUID's. These pieces
821 * are the actual FUIDs that were created initially.
822 * The FUID table index may no longer be valid and
823 * during zfs_create() a new index may be assigned.
824 * Because of this the log will contain the original
825 * doman+rid in order to create a new FUID.
827 * The individual ACEs may contain an ephemeral uid/gid which is no
828 * longer valid and will need to be replaced with an actual FUID.
832 zfs_replay_acl(zfsvfs_t *zfsvfs, lr_acl_t *lr, boolean_t byteswap)
834 ace_t *ace = (ace_t *)(lr + 1);
840 byteswap_uint64_array(lr, sizeof (*lr));
841 zfs_ace_byteswap(ace, lr->lr_acl_bytes, B_FALSE);
842 if (lr->lr_fuidcnt) {
843 byteswap_uint64_array((caddr_t)ace +
844 ZIL_ACE_LENGTH(lr->lr_acl_bytes),
845 lr->lr_fuidcnt * sizeof (uint64_t));
849 if ((error = zfs_zget(zfsvfs, lr->lr_foid, &zp)) != 0)
852 bzero(&vsa, sizeof (vsa));
853 vsa.vsa_mask = VSA_ACE | VSA_ACECNT | VSA_ACE_ACLFLAGS;
854 vsa.vsa_aclcnt = lr->lr_aclcnt;
855 vsa.vsa_aclentp = ace;
856 vsa.vsa_aclentsz = lr->lr_acl_bytes;
857 vsa.vsa_aclflags = lr->lr_acl_flags;
859 if (lr->lr_fuidcnt) {
860 void *fuidstart = (caddr_t)ace +
861 ZIL_ACE_LENGTH(lr->lr_acl_bytes);
863 zfsvfs->z_fuid_replay =
864 zfs_replay_fuids(fuidstart, &fuidstart,
865 lr->lr_fuidcnt, lr->lr_domcnt, 0, 0);
868 error = VOP_SETSECATTR(ZTOV(zp), &vsa, 0, kcred, NULL);
870 if (zfsvfs->z_fuid_replay)
871 zfs_fuid_info_free(zfsvfs->z_fuid_replay);
873 zfsvfs->z_fuid_replay = NULL;
880 * Callback vectors for replaying records
882 zil_replay_func_t *zfs_replay_vector[TX_MAX_TYPE] = {
883 zfs_replay_error, /* 0 no such transaction type */
884 zfs_replay_create, /* TX_CREATE */
885 zfs_replay_create, /* TX_MKDIR */
886 zfs_replay_create, /* TX_MKXATTR */
887 zfs_replay_create, /* TX_SYMLINK */
888 zfs_replay_remove, /* TX_REMOVE */
889 zfs_replay_remove, /* TX_RMDIR */
890 zfs_replay_link, /* TX_LINK */
891 zfs_replay_rename, /* TX_RENAME */
892 zfs_replay_write, /* TX_WRITE */
893 zfs_replay_truncate, /* TX_TRUNCATE */
894 zfs_replay_setattr, /* TX_SETATTR */
895 zfs_replay_acl_v0, /* TX_ACL_V0 */
896 zfs_replay_acl, /* TX_ACL */
897 zfs_replay_create_acl, /* TX_CREATE_ACL */
898 zfs_replay_create, /* TX_CREATE_ATTR */
899 zfs_replay_create_acl, /* TX_CREATE_ACL_ATTR */
900 zfs_replay_create_acl, /* TX_MKDIR_ACL */
901 zfs_replay_create, /* TX_MKDIR_ATTR */
902 zfs_replay_create_acl, /* TX_MKDIR_ACL_ATTR */
903 zfs_replay_write2, /* TX_WRITE2 */