d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
64aefee1b8
freebsd-nq/module/zfs/zfs_replay.c

951 lines
25 KiB
C
Raw Normal View History

Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
/*
* 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
*/
/*
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
ZFS replay transaction error 5 When zfs_replay_write() replays TX_WRITE records from ZIL it calls zpl_write_common() to perform the actual write. zpl_write_common() returns the number of bytes written (similar to write() system call) or an (negative) error. However, the code expects the positive return value to be a residual counter. Thus when zpl_write_common() successfully completes it is mistakenly considered to be a partial write and the error code delivered further. At this point the ZIL processing is aborted with famous "ZFS replay transaction error 5" error message given to the message buffer. The fix is to compare the zpl_write_commmon() return value with the buffer size and flag error only when they disagree. Signed-off-by: Cyril Plisko <cyril.plisko@mountall.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #933
2012-09-13 20:25:15 +00:00
* Copyright (c) 2012 Cyril Plisko. All rights reserved.
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 18:41:28 +00:00
* Copyright (c) 2013 by Delphix. All rights reserved.
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
*/
#include <sys/types.h>
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sysmacros.h>
#include <sys/cmn_err.h>
#include <sys/kmem.h>
#include <sys/thread.h>
#include <sys/file.h>
#include <sys/fcntl.h>
#include <sys/vfs.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_dir.h>
#include <sys/zfs_acl.h>
#include <sys/zfs_fuid.h>
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
#include <sys/zfs_vnops.h>
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
#include <sys/spa.h>
#include <sys/zil.h>
#include <sys/byteorder.h>
#include <sys/stat.h>
#include <sys/mode.h>
#include <sys/acl.h>
#include <sys/atomic.h>
#include <sys/cred.h>
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
#include <sys/zpl.h>
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
/*
* Functions to replay ZFS intent log (ZIL) records
* The functions are called through a function vector (zfs_replay_vector)
* which is indexed by the transaction type.
*/
static void
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
zfs_init_vattr(vattr_t *vap, uint64_t mask, uint64_t mode,
uint64_t uid, uint64_t gid, uint64_t rdev, uint64_t nodeid)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
bzero(vap, sizeof (*vap));
vap->va_mask = (uint_t)mask;
vap->va_type = IFTOVT(mode);
vap->va_mode = mode;
vap->va_uid = (uid_t)(IS_EPHEMERAL(uid)) ? -1 : uid;
vap->va_gid = (gid_t)(IS_EPHEMERAL(gid)) ? -1 : gid;
vap->va_rdev = rdev;
vap->va_nodeid = nodeid;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
}
/* ARGSUSED */
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_error(zfs_sb_t *zsb, lr_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 18:41:28 +00:00
return (SET_ERROR(ENOTSUP));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
}
static void
zfs_replay_xvattr(lr_attr_t *lrattr, xvattr_t *xvap)
{
xoptattr_t *xoap = NULL;
uint64_t *attrs;
uint64_t *crtime;
uint32_t *bitmap;
void *scanstamp;
int i;
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
xvap->xva_vattr.va_mask |= ATTR_XVATTR;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if ((xoap = xva_getxoptattr(xvap)) == NULL) {
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
xvap->xva_vattr.va_mask &= ~ATTR_XVATTR; /* shouldn't happen */
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return;
}
ASSERT(lrattr->lr_attr_masksize == xvap->xva_mapsize);
bitmap = &lrattr->lr_attr_bitmap;
for (i = 0; i != lrattr->lr_attr_masksize; i++, bitmap++)
xvap->xva_reqattrmap[i] = *bitmap;
attrs = (uint64_t *)(lrattr + lrattr->lr_attr_masksize - 1);
crtime = attrs + 1;
scanstamp = (caddr_t)(crtime + 2);
if (XVA_ISSET_REQ(xvap, XAT_HIDDEN))
xoap->xoa_hidden = ((*attrs & XAT0_HIDDEN) != 0);
if (XVA_ISSET_REQ(xvap, XAT_SYSTEM))
xoap->xoa_system = ((*attrs & XAT0_SYSTEM) != 0);
if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE))
xoap->xoa_archive = ((*attrs & XAT0_ARCHIVE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_READONLY))
xoap->xoa_readonly = ((*attrs & XAT0_READONLY) != 0);
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE))
xoap->xoa_immutable = ((*attrs & XAT0_IMMUTABLE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK))
xoap->xoa_nounlink = ((*attrs & XAT0_NOUNLINK) != 0);
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY))
xoap->xoa_appendonly = ((*attrs & XAT0_APPENDONLY) != 0);
if (XVA_ISSET_REQ(xvap, XAT_NODUMP))
xoap->xoa_nodump = ((*attrs & XAT0_NODUMP) != 0);
if (XVA_ISSET_REQ(xvap, XAT_OPAQUE))
xoap->xoa_opaque = ((*attrs & XAT0_OPAQUE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED))
xoap->xoa_av_modified = ((*attrs & XAT0_AV_MODIFIED) != 0);
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED))
xoap->xoa_av_quarantined =
((*attrs & XAT0_AV_QUARANTINED) != 0);
if (XVA_ISSET_REQ(xvap, XAT_CREATETIME))
ZFS_TIME_DECODE(&xoap->xoa_createtime, crtime);
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP))
bcopy(scanstamp, xoap->xoa_av_scanstamp, AV_SCANSTAMP_SZ);
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
if (XVA_ISSET_REQ(xvap, XAT_REPARSE))
xoap->xoa_reparse = ((*attrs & XAT0_REPARSE) != 0);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
if (XVA_ISSET_REQ(xvap, XAT_OFFLINE))
xoap->xoa_offline = ((*attrs & XAT0_OFFLINE) != 0);
if (XVA_ISSET_REQ(xvap, XAT_SPARSE))
xoap->xoa_sparse = ((*attrs & XAT0_SPARSE) != 0);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
}
static int
zfs_replay_domain_cnt(uint64_t uid, uint64_t gid)
{
uint64_t uid_idx;
uint64_t gid_idx;
int domcnt = 0;
uid_idx = FUID_INDEX(uid);
gid_idx = FUID_INDEX(gid);
if (uid_idx)
domcnt++;
if (gid_idx > 0 && gid_idx != uid_idx)
domcnt++;
return (domcnt);
}
static void *
zfs_replay_fuid_domain_common(zfs_fuid_info_t *fuid_infop, void *start,
int domcnt)
{
int i;
for (i = 0; i != domcnt; i++) {
fuid_infop->z_domain_table[i] = start;
start = (caddr_t)start + strlen(start) + 1;
}
return (start);
}
/*
* Set the uid/gid in the fuid_info structure.
*/
static void
zfs_replay_fuid_ugid(zfs_fuid_info_t *fuid_infop, uint64_t uid, uint64_t gid)
{
/*
* If owner or group are log specific FUIDs then slurp up
* domain information and build zfs_fuid_info_t
*/
if (IS_EPHEMERAL(uid))
fuid_infop->z_fuid_owner = uid;
if (IS_EPHEMERAL(gid))
fuid_infop->z_fuid_group = gid;
}
/*
* Load fuid domains into fuid_info_t
*/
static zfs_fuid_info_t *
zfs_replay_fuid_domain(void *buf, void **end, uint64_t uid, uint64_t gid)
{
int domcnt;
zfs_fuid_info_t *fuid_infop;
fuid_infop = zfs_fuid_info_alloc();
domcnt = zfs_replay_domain_cnt(uid, gid);
if (domcnt == 0)
return (fuid_infop);
fuid_infop->z_domain_table =
kmem_zalloc(domcnt * sizeof (char **), KM_SLEEP);
zfs_replay_fuid_ugid(fuid_infop, uid, gid);
fuid_infop->z_domain_cnt = domcnt;
*end = zfs_replay_fuid_domain_common(fuid_infop, buf, domcnt);
return (fuid_infop);
}
/*
* load zfs_fuid_t's and fuid_domains into fuid_info_t
*/
static zfs_fuid_info_t *
zfs_replay_fuids(void *start, void **end, int idcnt, int domcnt, uint64_t uid,
uint64_t gid)
{
uint64_t *log_fuid = (uint64_t *)start;
zfs_fuid_info_t *fuid_infop;
int i;
fuid_infop = zfs_fuid_info_alloc();
fuid_infop->z_domain_cnt = domcnt;
fuid_infop->z_domain_table =
kmem_zalloc(domcnt * sizeof (char **), KM_SLEEP);
for (i = 0; i != idcnt; i++) {
zfs_fuid_t *zfuid;
zfuid = kmem_alloc(sizeof (zfs_fuid_t), KM_SLEEP);
zfuid->z_logfuid = *log_fuid;
zfuid->z_id = -1;
zfuid->z_domidx = 0;
list_insert_tail(&fuid_infop->z_fuids, zfuid);
log_fuid++;
}
zfs_replay_fuid_ugid(fuid_infop, uid, gid);
*end = zfs_replay_fuid_domain_common(fuid_infop, log_fuid, domcnt);
return (fuid_infop);
}
static void
zfs_replay_swap_attrs(lr_attr_t *lrattr)
{
/* swap the lr_attr structure */
byteswap_uint32_array(lrattr, sizeof (*lrattr));
/* swap the bitmap */
byteswap_uint32_array(lrattr + 1, (lrattr->lr_attr_masksize - 1) *
sizeof (uint32_t));
/* swap the attributes, create time + 64 bit word for attributes */
byteswap_uint64_array((caddr_t)(lrattr + 1) + (sizeof (uint32_t) *
(lrattr->lr_attr_masksize - 1)), 3 * sizeof (uint64_t));
}
/*
* Replay file create with optional ACL, xvattr information as well
* as option FUID information.
*/
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_create_acl(zfs_sb_t *zsb, lr_acl_create_t *lracl, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
char *name = NULL; /* location determined later */
lr_create_t *lr = (lr_create_t *)lracl;
znode_t *dzp;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
struct inode *ip = NULL;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
xvattr_t xva;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
int vflg = 0;
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
vsecattr_t vsec = { 0 };
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
lr_attr_t *lrattr;
void *aclstart;
void *fuidstart;
size_t xvatlen = 0;
uint64_t txtype;
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
uint64_t objid;
uint64_t dnodesize;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
int error;
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
txtype = (lr->lr_common.lrc_txtype & ~TX_CI);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if (byteswap) {
byteswap_uint64_array(lracl, sizeof (*lracl));
if (txtype == TX_CREATE_ACL_ATTR ||
txtype == TX_MKDIR_ACL_ATTR) {
lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
zfs_replay_swap_attrs(lrattr);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
}
aclstart = (caddr_t)(lracl + 1) + xvatlen;
zfs_ace_byteswap(aclstart, lracl->lr_acl_bytes, B_FALSE);
/* swap fuids */
if (lracl->lr_fuidcnt) {
byteswap_uint64_array((caddr_t)aclstart +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes),
lracl->lr_fuidcnt * sizeof (uint64_t));
}
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_doid, &dzp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
objid = LR_FOID_GET_OBJ(lr->lr_foid);
dnodesize = LR_FOID_GET_SLOTS(lr->lr_foid) << DNODE_SHIFT;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
xva_init(&xva);
Use Linux ATTR_ versions The AT_ versions of these macros are used on Solaris and while they map to their Linux equivilants the code has been updated to use the ATTR_ versions.
2011-03-03 19:29:15 +00:00
zfs_init_vattr(&xva.xva_vattr, ATTR_MODE | ATTR_UID | ATTR_GID,
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
lr->lr_mode, lr->lr_uid, lr->lr_gid, lr->lr_rdev, objid);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
/*
* All forms of zfs create (create, mkdir, mkxattrdir, symlink)
* eventually end up in zfs_mknode(), which assigns the object's
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
* creation time, generation number, and dnode size. The generic
* zfs_create() has no concept of these attributes, so we smuggle
* the values inside the vattr's otherwise unused va_ctime,
* va_nblocks, and va_fsid fields.
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
*/
ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_crtime);
xva.xva_vattr.va_nblocks = lr->lr_gen;
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
xva.xva_vattr.va_fsid = dnodesize;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = dmu_object_info(zsb->z_os, lr->lr_foid, NULL);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if (error != ENOENT)
goto bail;
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
switch (txtype) {
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
case TX_CREATE_ACL:
aclstart = (caddr_t)(lracl + 1);
fuidstart = (caddr_t)aclstart +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay = zfs_replay_fuids(fuidstart,
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
/*FALLTHROUGH*/
case TX_CREATE_ACL_ATTR:
if (name == NULL) {
lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
xva.xva_vattr.va_mask |= ATTR_XVATTR;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_xvattr(lrattr, &xva);
}
vsec.vsa_mask = VSA_ACE | VSA_ACE_ACLFLAGS;
vsec.vsa_aclentp = (caddr_t)(lracl + 1) + xvatlen;
vsec.vsa_aclcnt = lracl->lr_aclcnt;
vsec.vsa_aclentsz = lracl->lr_acl_bytes;
vsec.vsa_aclflags = lracl->lr_acl_flags;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (zsb->z_fuid_replay == NULL) {
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
fuidstart = (caddr_t)(lracl + 1) + xvatlen +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay =
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_fuids(fuidstart,
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_create(ZTOI(dzp), name, &xva.xva_vattr,
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
0, 0, &ip, kcred, vflg, &vsec);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
case TX_MKDIR_ACL:
aclstart = (caddr_t)(lracl + 1);
fuidstart = (caddr_t)aclstart +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay = zfs_replay_fuids(fuidstart,
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
/*FALLTHROUGH*/
case TX_MKDIR_ACL_ATTR:
if (name == NULL) {
lrattr = (lr_attr_t *)(caddr_t)(lracl + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
zfs_replay_xvattr(lrattr, &xva);
}
vsec.vsa_mask = VSA_ACE | VSA_ACE_ACLFLAGS;
vsec.vsa_aclentp = (caddr_t)(lracl + 1) + xvatlen;
vsec.vsa_aclcnt = lracl->lr_aclcnt;
vsec.vsa_aclentsz = lracl->lr_acl_bytes;
vsec.vsa_aclflags = lracl->lr_acl_flags;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (zsb->z_fuid_replay == NULL) {
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
fuidstart = (caddr_t)(lracl + 1) + xvatlen +
ZIL_ACE_LENGTH(lracl->lr_acl_bytes);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay =
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_fuids(fuidstart,
(void *)&name, lracl->lr_fuidcnt, lracl->lr_domcnt,
lr->lr_uid, lr->lr_gid);
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_mkdir(ZTOI(dzp), name, &xva.xva_vattr,
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
&ip, kcred, vflg, &vsec);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
default:
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 18:41:28 +00:00
error = SET_ERROR(ENOTSUP);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
}
bail:
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (error == 0 && ip != NULL)
iput(ip);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(dzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (zsb->z_fuid_replay)
zfs_fuid_info_free(zsb->z_fuid_replay);
zsb->z_fuid_replay = NULL;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_create(zfs_sb_t *zsb, lr_create_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
char *name = NULL; /* location determined later */
char *link; /* symlink content follows name */
znode_t *dzp;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
struct inode *ip = NULL;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
xvattr_t xva;
int vflg = 0;
size_t lrsize = sizeof (lr_create_t);
lr_attr_t *lrattr;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
void *start;
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
size_t xvatlen;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
uint64_t txtype;
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
uint64_t objid;
uint64_t dnodesize;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
int error;
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
txtype = (lr->lr_common.lrc_txtype & ~TX_CI);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
if (txtype == TX_CREATE_ATTR || txtype == TX_MKDIR_ATTR)
zfs_replay_swap_attrs((lr_attr_t *)(lr + 1));
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_doid, &dzp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
objid = LR_FOID_GET_OBJ(lr->lr_foid);
dnodesize = LR_FOID_GET_SLOTS(lr->lr_foid) << DNODE_SHIFT;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
xva_init(&xva);
Use Linux ATTR_ versions The AT_ versions of these macros are used on Solaris and while they map to their Linux equivilants the code has been updated to use the ATTR_ versions.
2011-03-03 19:29:15 +00:00
zfs_init_vattr(&xva.xva_vattr, ATTR_MODE | ATTR_UID | ATTR_GID,
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
lr->lr_mode, lr->lr_uid, lr->lr_gid, lr->lr_rdev, objid);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
/*
* All forms of zfs create (create, mkdir, mkxattrdir, symlink)
* eventually end up in zfs_mknode(), which assigns the object's
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
* creation time, generation number, and dnode slot count. The
* generic zfs_create() has no concept of these attributes, so
* we smuggle the values inside * the vattr's otherwise unused
* va_ctime, va_nblocks, and va_nlink fields.
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
*/
ZFS_TIME_DECODE(&xva.xva_vattr.va_ctime, lr->lr_crtime);
xva.xva_vattr.va_nblocks = lr->lr_gen;
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
xva.xva_vattr.va_fsid = dnodesize;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 01:25:34 +00:00
error = dmu_object_info(zsb->z_os, objid, NULL);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if (error != ENOENT)
goto out;
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
/*
* Symlinks don't have fuid info, and CIFS never creates
* symlinks.
*
* The _ATTR versions will grab the fuid info in their subcases.
*/
if ((int)lr->lr_common.lrc_txtype != TX_SYMLINK &&
(int)lr->lr_common.lrc_txtype != TX_MKDIR_ATTR &&
(int)lr->lr_common.lrc_txtype != TX_CREATE_ATTR) {
start = (lr + 1);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay =
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
}
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
switch (txtype) {
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
case TX_CREATE_ATTR:
lrattr = (lr_attr_t *)(caddr_t)(lr + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
zfs_replay_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), &xva);
start = (caddr_t)(lr + 1) + xvatlen;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay =
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
name = (char *)start;
/*FALLTHROUGH*/
case TX_CREATE:
if (name == NULL)
name = (char *)start;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_create(ZTOI(dzp), name, &xva.xva_vattr,
0, 0, &ip, kcred, vflg, NULL);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
case TX_MKDIR_ATTR:
lrattr = (lr_attr_t *)(caddr_t)(lr + 1);
xvatlen = ZIL_XVAT_SIZE(lrattr->lr_attr_masksize);
zfs_replay_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), &xva);
start = (caddr_t)(lr + 1) + xvatlen;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay =
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_fuid_domain(start, &start,
lr->lr_uid, lr->lr_gid);
name = (char *)start;
/*FALLTHROUGH*/
case TX_MKDIR:
if (name == NULL)
name = (char *)(lr + 1);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_mkdir(ZTOI(dzp), name, &xva.xva_vattr,
&ip, kcred, vflg, NULL);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
case TX_MKXATTR:
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_make_xattrdir(dzp, &xva.xva_vattr, &ip, kcred);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
case TX_SYMLINK:
name = (char *)(lr + 1);
link = name + strlen(name) + 1;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_symlink(ZTOI(dzp), name, &xva.xva_vattr,
link, &ip, kcred, vflg);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
default:
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 18:41:28 +00:00
error = SET_ERROR(ENOTSUP);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
}
out:
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (error == 0 && ip != NULL)
iput(ip);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(dzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (zsb->z_fuid_replay)
zfs_fuid_info_free(zsb->z_fuid_replay);
zsb->z_fuid_replay = NULL;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_remove(zfs_sb_t *zsb, lr_remove_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
char *name = (char *)(lr + 1); /* name follows lr_remove_t */
znode_t *dzp;
int error;
int vflg = 0;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_doid, &dzp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
switch ((int)lr->lr_common.lrc_txtype) {
case TX_REMOVE:
Add pn_alloc()/pn_free() functions In order to remove the HAVE_PN_UTILS wrappers the pn_alloc() and pn_free() functions must be implemented. The existing illumos implementation were used for this purpose. The `flags` argument which was used in places wrapped by the HAVE_PN_UTILS condition has beed added back to zfs_remove() and zfs_link() functions. This removes a small point of divergence between the ZoL code and upstream. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #4522
2016-04-13 15:55:35 +00:00
error = zfs_remove(ZTOI(dzp), name, kcred, vflg);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
case TX_RMDIR:
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_rmdir(ZTOI(dzp), name, NULL, kcred, vflg);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
break;
default:
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 18:41:28 +00:00
error = SET_ERROR(ENOTSUP);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(dzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_link(zfs_sb_t *zsb, lr_link_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
char *name = (char *)(lr + 1); /* name follows lr_link_t */
znode_t *dzp, *zp;
int error;
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
int vflg = 0;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_doid, &dzp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_link_obj, &zp)) != 0) {
iput(ZTOI(dzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
Add pn_alloc()/pn_free() functions In order to remove the HAVE_PN_UTILS wrappers the pn_alloc() and pn_free() functions must be implemented. The existing illumos implementation were used for this purpose. The `flags` argument which was used in places wrapped by the HAVE_PN_UTILS condition has beed added back to zfs_remove() and zfs_link() functions. This removes a small point of divergence between the ZoL code and upstream. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #4522
2016-04-13 15:55:35 +00:00
error = zfs_link(ZTOI(dzp), ZTOI(zp), name, kcred, vflg);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(zp));
iput(ZTOI(dzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_rename(zfs_sb_t *zsb, lr_rename_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
char *sname = (char *)(lr + 1); /* sname and tname follow lr_rename_t */
char *tname = sname + strlen(sname) + 1;
znode_t *sdzp, *tdzp;
int error;
int vflg = 0;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_sdoid, &sdzp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_tdoid, &tdzp)) != 0) {
iput(ZTOI(sdzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
if (lr->lr_common.lrc_txtype & TX_CI)
vflg |= FIGNORECASE;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_rename(ZTOI(sdzp), sname, ZTOI(tdzp), tname, kcred, vflg);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(tdzp));
iput(ZTOI(sdzp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_write(zfs_sb_t *zsb, lr_write_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
char *data = (char *)(lr + 1); /* data follows lr_write_t */
znode_t *zp;
ZFS replay transaction error 5 When zfs_replay_write() replays TX_WRITE records from ZIL it calls zpl_write_common() to perform the actual write. zpl_write_common() returns the number of bytes written (similar to write() system call) or an (negative) error. However, the code expects the positive return value to be a residual counter. Thus when zpl_write_common() successfully completes it is mistakenly considered to be a partial write and the error code delivered further. At this point the ZIL processing is aborted with famous "ZFS replay transaction error 5" error message given to the message buffer. The fix is to compare the zpl_write_commmon() return value with the buffer size and flag error only when they disagree. Signed-off-by: Cyril Plisko <cyril.plisko@mountall.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #933
2012-09-13 20:25:15 +00:00
int error, written;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
uint64_t eod, offset, length;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_foid, &zp)) != 0) {
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
/*
* As we can log writes out of order, it's possible the
* file has been removed. In this case just drop the write
* and return success.
*/
if (error == ENOENT)
error = 0;
return (error);
}
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
offset = lr->lr_offset;
length = lr->lr_length;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
eod = offset + length; /* end of data for this write */
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
/*
* This may be a write from a dmu_sync() for a whole block,
* and may extend beyond the current end of the file.
* We can't just replay what was written for this TX_WRITE as
* a future TX_WRITE2 may extend the eof and the data for that
* write needs to be there. So we write the whole block and
* reduce the eof. This needs to be done within the single dmu
* transaction created within vn_rdwr -> zfs_write. So a possible
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
* new end of file is passed through in zsb->z_replay_eof
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
*/
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_replay_eof = 0; /* 0 means don't change end of file */
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
/* If it's a dmu_sync() block, write the whole block */
if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
uint64_t blocksize = BP_GET_LSIZE(&lr->lr_blkptr);
if (length < blocksize) {
offset -= offset % blocksize;
length = blocksize;
}
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
if (zp->z_size < eod)
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_replay_eof = eod;
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
}
Linux AIO Support nfsd uses do_readv_writev() to implement fops->read and fops->write. do_readv_writev() will attempt to read/write using fops->aio_read and fops->aio_write, but it will fallback to fops->read and fops->write when AIO is not available. However, the fallback will perform a call for each individual data page. Since our default recordsize is 128KB, sequential operations on NFS will generate 32 DMU transactions where only 1 transaction was needed. That was unnecessary overhead and we implement fops->aio_read and fops->aio_write to eliminate it. ZFS originated in OpenSolaris, where the AIO API is entirely implemented in userland's libc by intelligently mapping them to VOP_WRITE, VOP_READ and VOP_FSYNC. Linux implements AIO inside the kernel itself. Linux filesystems therefore must implement their own AIO logic and nearly all of them implement fops->aio_write synchronously. Consequently, they do not implement aio_fsync(). However, since the ZPL works by mapping Linux's VFS calls to the functions implementing Illumos' VFS operations, we instead implement AIO in the kernel by mapping the operations to the VOP_READ, VOP_WRITE and VOP_FSYNC equivalents. We therefore implement fops->aio_fsync. One might be inclined to make our fops->aio_write implementation synchronous to make software that expects this behavior safe. However, there are several reasons not to do this: 1. Other platforms do not implement aio_write() synchronously and since the majority of userland software using AIO should be cross platform, expectations of synchronous behavior should not be a problem. 2. We would hurt the performance of programs that use POSIX interfaces properly while simultaneously encouraging the creation of more non-compliant software. 3. The broader community concluded that userland software should be patched to properly use POSIX interfaces instead of implementing hacks in filesystems to cater to broken software. This concept is best described as the O_PONIES debate. 4. Making an asynchronous write synchronous is non sequitur. Any software dependent on synchronous aio_write behavior will suffer data loss on ZFSOnLinux in a kernel panic / system failure of at most zfs_txg_timeout seconds, which by default is 5 seconds. This seems like a reasonable consequence of using non-compliant software. It should be noted that this is also a problem in the kernel itself where nfsd does not pass O_SYNC on files opened with it and instead relies on a open()/write()/close() to enforce synchronous behavior when the flush is only guarenteed on last close. Exporting any filesystem that does not implement AIO via NFS risks data loss in the event of a kernel panic / system failure when something else is also accessing the file. Exporting any file system that implements AIO the way this patch does bears similar risk. However, it seems reasonable to forgo crippling our AIO implementation in favor of developing patches to fix this problem in Linux's nfsd for the reasons stated earlier. In the interim, the risk will remain. Failing to implement AIO will not change the problem that nfsd created, so there is no reason for nfsd's mistake to block our implementation of AIO. It also should be noted that `aio_cancel()` will always return `AIO_NOTCANCELED` under this implementation. It is possible to implement aio_cancel by deferring work to taskqs and use `kiocb_set_cancel_fn()` to set a callback function for cancelling work sent to taskqs, but the simpler approach is allowed by the specification: ``` Which operations are cancelable is implementation-defined. ``` http://pubs.opengroup.org/onlinepubs/009695399/functions/aio_cancel.html The only programs on my system that are capable of using `aio_cancel()` are QEMU, beecrypt and fio use it according to a recursive grep of my system's `/usr/src/debug`. That suggests that `aio_cancel()` users are rare. Implementing aio_cancel() is left to a future date when it is clear that there are consumers that benefit from its implementation to justify the work. Lastly, it is important to know that handling of the iovec updates differs between Illumos and Linux in the implementation of read/write. On Linux, it is the VFS' responsibility whle on Illumos, it is the filesystem's responsibility. We take the intermediate solution of copying the iovec so that the ZFS code can update it like on Solaris while leaving the originals alone. This imposes some overhead. We could always revisit this should profiling show that the allocations are a problem. Signed-off-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #223 Closes #2373
2014-08-04 11:09:32 +00:00
written = zpl_write_common(ZTOI(zp), data, length, &offset,
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
UIO_SYSSPACE, 0, kcred);
ZFS replay transaction error 5 When zfs_replay_write() replays TX_WRITE records from ZIL it calls zpl_write_common() to perform the actual write. zpl_write_common() returns the number of bytes written (similar to write() system call) or an (negative) error. However, the code expects the positive return value to be a residual counter. Thus when zpl_write_common() successfully completes it is mistakenly considered to be a partial write and the error code delivered further. At this point the ZIL processing is aborted with famous "ZFS replay transaction error 5" error message given to the message buffer. The fix is to compare the zpl_write_commmon() return value with the buffer size and flag error only when they disagree. Signed-off-by: Cyril Plisko <cyril.plisko@mountall.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #933
2012-09-13 20:25:15 +00:00
if (written < 0)
error = -written;
else if (written < length)
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 18:41:28 +00:00
error = SET_ERROR(EIO); /* short write */
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(zp));
zsb->z_replay_eof = 0; /* safety */
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
return (error);
}
/*
* TX_WRITE2 are only generated when dmu_sync() returns EALREADY
* meaning the pool block is already being synced. So now that we always write
* out full blocks, all we have to do is expand the eof if
* the file is grown.
*/
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_write2(zfs_sb_t *zsb, lr_write_t *lr, boolean_t byteswap)
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
{
znode_t *zp;
int error;
uint64_t end;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_foid, &zp)) != 0)
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
return (error);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
top:
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
end = lr->lr_offset + lr->lr_length;
if (end > zp->z_size) {
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
dmu_tx_t *tx = dmu_tx_create(zsb->z_os);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
zp->z_size = end;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(zp));
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
if (error == ERESTART) {
dmu_tx_wait(tx);
dmu_tx_abort(tx);
goto top;
}
dmu_tx_abort(tx);
return (error);
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
(void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zsb),
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
(void *)&zp->z_size, sizeof (uint64_t), tx);
/* Ensure the replayed seq is updated */
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
(void) zil_replaying(zsb->z_log, tx);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-26 21:24:34 +00:00
dmu_tx_commit(tx);
Update core ZFS code from build 121 to build 141.
2010-05-28 20:45:14 +00:00
}
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(zp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_truncate(zfs_sb_t *zsb, lr_truncate_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
znode_t *zp;
flock64_t fl;
int error;
if (byteswap)
byteswap_uint64_array(lr, sizeof (*lr));
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_foid, &zp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
bzero(&fl, sizeof (fl));
fl.l_type = F_WRLCK;
fl.l_whence = 0;
fl.l_start = lr->lr_offset;
fl.l_len = lr->lr_length;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_space(ZTOI(zp), F_FREESP, &fl, FWRITE | FOFFMAX,
lr->lr_offset, kcred);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(zp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_setattr(zfs_sb_t *zsb, lr_setattr_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
znode_t *zp;
xvattr_t xva;
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
vattr_t *vap = &xva.xva_vattr;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
int error;
void *start;
xva_init(&xva);
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
if ((lr->lr_mask & ATTR_XVATTR) &&
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_version >= ZPL_VERSION_INITIAL)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_swap_attrs((lr_attr_t *)(lr + 1));
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_foid, &zp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
zfs_init_vattr(vap, lr->lr_mask, lr->lr_mode,
lr->lr_uid, lr->lr_gid, 0, lr->lr_foid);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
vap->va_size = lr->lr_size;
ZFS_TIME_DECODE(&vap->va_atime, lr->lr_atime);
ZFS_TIME_DECODE(&vap->va_mtime, lr->lr_mtime);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
/*
* Fill in xvattr_t portions if necessary.
*/
start = (lr_setattr_t *)(lr + 1);
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
if (vap->va_mask & ATTR_XVATTR) {
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_xvattr((lr_attr_t *)start, &xva);
start = (caddr_t)start +
ZIL_XVAT_SIZE(((lr_attr_t *)start)->lr_attr_masksize);
} else
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
xva.xva_vattr.va_mask &= ~ATTR_XVATTR;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay = zfs_replay_fuid_domain(start, &start,
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
lr->lr_uid, lr->lr_gid);
Drop HAVE_XVATTR macros When I began work on the Posix layer it immediately became clear to me that to integrate cleanly with the Linux VFS certain Solaris specific things would have to go. One of these things was to elimate as many Solaris specific types from the ZPL layer as possible. They would be replaced with their Linux equivalents. This would not only be good for performance, but for the general readability and health of the code. The Solaris and Linux VFS are different beasts and should be treated as such. Most of the code remains common for constructing transactions and such, but there are subtle and important differenced which need to be repsected. This policy went quite for for certain types such as the vnode_t, and it initially seemed to be working out well for the vattr_t. There was a relatively small amount of related xvattr_t code I was forced to comment out with HAVE_XVATTR. But it didn't look that hard to come back soon and replace it all with a native Linux type. However, after going doing this path with xvattr some distance it clear that this code was woven in the ZPL more deeply than I thought. In particular its hooks went very deep in to the ZPL replay code and replacing it would not be as easy as I originally thought. Rather than continue persuing replacing and removing this code I've taken a step back and reevaluted things. This commit reverts many of my previous commits which removed xvattr related code. It restores much of the code to its original upstream state and now relies on improved xvattr_t support in the zfs package itself. The result of this is that much of the code which I had commented out, which accidentally broke things like replay, is now back in place and working. However, there may be a small performance impact for getattr/setattr operations because they now require a translation from native Linux to Solaris types. For now that's a price I'm willing to pay. Once everything is completely functional we can revisting the issue of removing the vattr_t/xvattr_t types. Closes #111
2011-03-01 20:24:09 +00:00
error = zfs_setattr(ZTOI(zp), vap, 0, kcred);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_fuid_info_free(zsb->z_fuid_replay);
zsb->z_fuid_replay = NULL;
iput(ZTOI(zp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_acl_v0(zfs_sb_t *zsb, lr_acl_v0_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
ace_t *ace = (ace_t *)(lr + 1); /* ace array follows lr_acl_t */
vsecattr_t vsa;
znode_t *zp;
int error;
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
zfs_oldace_byteswap(ace, lr->lr_aclcnt);
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_foid, &zp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
bzero(&vsa, sizeof (vsa));
vsa.vsa_mask = VSA_ACE | VSA_ACECNT;
vsa.vsa_aclcnt = lr->lr_aclcnt;
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 20:09:06 +00:00
vsa.vsa_aclentsz = sizeof (ace_t) * vsa.vsa_aclcnt;
vsa.vsa_aclflags = 0;
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
vsa.vsa_aclentp = ace;
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_setsecattr(ZTOI(zp), &vsa, 0, kcred);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
iput(ZTOI(zp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
/*
* Replaying ACLs is complicated by FUID support.
* The log record may contain some optional data
* to be used for replaying FUID's. These pieces
* are the actual FUIDs that were created initially.
* The FUID table index may no longer be valid and
* during zfs_create() a new index may be assigned.
* Because of this the log will contain the original
* doman+rid in order to create a new FUID.
*
* The individual ACEs may contain an ephemeral uid/gid which is no
* longer valid and will need to be replaced with an actual FUID.
*
*/
static int
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zfs_replay_acl(zfs_sb_t *zsb, lr_acl_t *lr, boolean_t byteswap)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
{
ace_t *ace = (ace_t *)(lr + 1);
vsecattr_t vsa;
znode_t *zp;
int error;
if (byteswap) {
byteswap_uint64_array(lr, sizeof (*lr));
zfs_ace_byteswap(ace, lr->lr_acl_bytes, B_FALSE);
if (lr->lr_fuidcnt) {
byteswap_uint64_array((caddr_t)ace +
ZIL_ACE_LENGTH(lr->lr_acl_bytes),
lr->lr_fuidcnt * sizeof (uint64_t));
}
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if ((error = zfs_zget(zsb, lr->lr_foid, &zp)) != 0)
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
bzero(&vsa, sizeof (vsa));
vsa.vsa_mask = VSA_ACE | VSA_ACECNT | VSA_ACE_ACLFLAGS;
vsa.vsa_aclcnt = lr->lr_aclcnt;
vsa.vsa_aclentp = ace;
vsa.vsa_aclentsz = lr->lr_acl_bytes;
vsa.vsa_aclflags = lr->lr_acl_flags;
if (lr->lr_fuidcnt) {
void *fuidstart = (caddr_t)ace +
ZIL_ACE_LENGTH(lr->lr_acl_bytes);
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay =
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
zfs_replay_fuids(fuidstart, &fuidstart,
lr->lr_fuidcnt, lr->lr_domcnt, 0, 0);
}
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
error = zfs_setsecattr(ZTOI(zp), &vsa, 0, kcred);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
if (zsb->z_fuid_replay)
zfs_fuid_info_free(zsb->z_fuid_replay);
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
Wrap with HAVE_XVATTR For the moment exactly how to handle xvattr is not clear. This change largely consists of the code to comment out the offending bits until something reasonable can be done.
2011-02-08 19:33:08 +00:00
zsb->z_fuid_replay = NULL;
iput(ZTOI(zp));
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
return (error);
}
/*
* Callback vectors for replaying records
*/
Constify structures containing function pointers The PaX team modified the kernel's modpost to report writeable function pointers as section mismatches because they are potential exploit targets. We could ignore the warnings, but their presence can obscure actual issues. Proper const correctness can also catch programming mistakes. Building the kernel modules against a PaX/GrSecurity patched Linux 3.4.2 kernel reports 133 section mismatches prior to this patch. This patch eliminates 130 of them. The quantity of writeable function pointers eliminated by constifying each structure is as follows: vdev_opts_t 52 zil_replay_func_t 24 zio_compress_info_t 24 zio_checksum_info_t 9 space_map_ops_t 7 arc_byteswap_func_t 5 The remaining 3 writeable function pointers cannot be addressed by this patch. 2 of them are in zpl_fs_type. The kernel's sget function requires that this be non-const. The final writeable function pointer is created by SPL_SHRINKER_DECLARE. The kernel's set_shrinker() and remove_shrinker() functions also require that this be non-const. Signed-off-by: Richard Yao <ryao@cs.stonybrook.edu> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1300
2013-02-15 04:37:43 +00:00
zil_replay_func_t zfs_replay_vector[TX_MAX_TYPE] = {
(zil_replay_func_t)zfs_replay_error, /* no such type */
(zil_replay_func_t)zfs_replay_create, /* TX_CREATE */
(zil_replay_func_t)zfs_replay_create, /* TX_MKDIR */
(zil_replay_func_t)zfs_replay_create, /* TX_MKXATTR */
(zil_replay_func_t)zfs_replay_create, /* TX_SYMLINK */
(zil_replay_func_t)zfs_replay_remove, /* TX_REMOVE */
(zil_replay_func_t)zfs_replay_remove, /* TX_RMDIR */
(zil_replay_func_t)zfs_replay_link, /* TX_LINK */
(zil_replay_func_t)zfs_replay_rename, /* TX_RENAME */
(zil_replay_func_t)zfs_replay_write, /* TX_WRITE */
(zil_replay_func_t)zfs_replay_truncate, /* TX_TRUNCATE */
(zil_replay_func_t)zfs_replay_setattr, /* TX_SETATTR */
(zil_replay_func_t)zfs_replay_acl_v0, /* TX_ACL_V0 */
(zil_replay_func_t)zfs_replay_acl, /* TX_ACL */
(zil_replay_func_t)zfs_replay_create_acl, /* TX_CREATE_ACL */
(zil_replay_func_t)zfs_replay_create, /* TX_CREATE_ATTR */
(zil_replay_func_t)zfs_replay_create_acl, /* TX_CREATE_ACL_ATTR */
(zil_replay_func_t)zfs_replay_create_acl, /* TX_MKDIR_ACL */
(zil_replay_func_t)zfs_replay_create, /* TX_MKDIR_ATTR */
(zil_replay_func_t)zfs_replay_create_acl, /* TX_MKDIR_ACL_ATTR */
(zil_replay_func_t)zfs_replay_write2, /* TX_WRITE2 */
Initial Linux ZFS GIT Repo
2008-11-20 20:01:55 +00:00
};
Reference in New Issue Copy Permalink