c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2024-11-17 18:11:00 +03:00
Code Issues Packages Projects Releases Wiki Activity
1611bb7b4f
mirror_zfs/module/zfs/zfs_log.c

689 lines
18 KiB
C
Raw Normal View History

Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
* 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.
2010-08-27 01:24:34 +04:00
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
Add linux kernel disk support Native Linux vdev disk interfaces Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:45:02 +04:00
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
#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/vfs.h>
#include <sys/zfs_znode.h>
#include <sys/zfs_dir.h>
#include <sys/zil.h>
#include <sys/zil_impl.h>
#include <sys/byteorder.h>
#include <sys/policy.h>
#include <sys/stat.h>
#include <sys/mode.h>
#include <sys/acl.h>
#include <sys/dmu.h>
#include <sys/spa.h>
#include <sys/zfs_fuid.h>
#include <sys/ddi.h>
Rebase master to b105
2009-01-16 00:59:39 +03:00
#include <sys/dsl_dataset.h>
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
Rebase master to b105
2009-01-16 00:59:39 +03:00
* These zfs_log_* functions must be called within a dmu tx, in one
* of 2 contexts depending on zilog->z_replay:
*
* Non replay mode
* ---------------
* We need to record the transaction so that if it is committed to
* the Intent Log then it can be replayed. An intent log transaction
* structure (itx_t) is allocated and all the information necessary to
* possibly replay the transaction is saved in it. The itx is then assigned
* a sequence number and inserted in the in-memory list anchored in the zilog.
*
* Replay mode
* -----------
* We need to mark the intent log record as replayed in the log header.
* This is done in the same transaction as the replay so that they
* commit atomically.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
int
zfs_log_create_txtype(zil_create_t type, vsecattr_t *vsecp, vattr_t *vap)
{
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 23:24:09 +03:00
int isxvattr = (vap->va_mask & ATTR_XVATTR);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
switch (type) {
case Z_FILE:
if (vsecp == NULL && !isxvattr)
return (TX_CREATE);
if (vsecp && isxvattr)
return (TX_CREATE_ACL_ATTR);
if (vsecp)
return (TX_CREATE_ACL);
else
return (TX_CREATE_ATTR);
/*NOTREACHED*/
case Z_DIR:
if (vsecp == NULL && !isxvattr)
return (TX_MKDIR);
if (vsecp && isxvattr)
return (TX_MKDIR_ACL_ATTR);
if (vsecp)
return (TX_MKDIR_ACL);
else
return (TX_MKDIR_ATTR);
case Z_XATTRDIR:
return (TX_MKXATTR);
}
ASSERT(0);
return (TX_MAX_TYPE);
}
/*
* build up the log data necessary for logging xvattr_t
* First lr_attr_t is initialized. following the lr_attr_t
* is the mapsize and attribute bitmap copied from the xvattr_t.
* Following the bitmap and bitmapsize two 64 bit words are reserved
* for the create time which may be set. Following the create time
* records a single 64 bit integer which has the bits to set on
* replay for the xvattr.
*/
static void
zfs_log_xvattr(lr_attr_t *lrattr, xvattr_t *xvap)
{
uint32_t *bitmap;
uint64_t *attrs;
uint64_t *crtime;
xoptattr_t *xoap;
void *scanstamp;
int i;
xoap = xva_getxoptattr(xvap);
ASSERT(xoap);
lrattr->lr_attr_masksize = xvap->xva_mapsize;
bitmap = &lrattr->lr_attr_bitmap;
for (i = 0; i != xvap->xva_mapsize; i++, bitmap++) {
*bitmap = xvap->xva_reqattrmap[i];
}
/* Now pack the attributes up in a single uint64_t */
attrs = (uint64_t *)bitmap;
crtime = attrs + 1;
scanstamp = (caddr_t)(crtime + 2);
*attrs = 0;
if (XVA_ISSET_REQ(xvap, XAT_READONLY))
*attrs |= (xoap->xoa_readonly == 0) ? 0 :
XAT0_READONLY;
if (XVA_ISSET_REQ(xvap, XAT_HIDDEN))
*attrs |= (xoap->xoa_hidden == 0) ? 0 :
XAT0_HIDDEN;
if (XVA_ISSET_REQ(xvap, XAT_SYSTEM))
*attrs |= (xoap->xoa_system == 0) ? 0 :
XAT0_SYSTEM;
if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE))
*attrs |= (xoap->xoa_archive == 0) ? 0 :
XAT0_ARCHIVE;
if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE))
*attrs |= (xoap->xoa_immutable == 0) ? 0 :
XAT0_IMMUTABLE;
if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK))
*attrs |= (xoap->xoa_nounlink == 0) ? 0 :
XAT0_NOUNLINK;
if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY))
*attrs |= (xoap->xoa_appendonly == 0) ? 0 :
XAT0_APPENDONLY;
if (XVA_ISSET_REQ(xvap, XAT_OPAQUE))
*attrs |= (xoap->xoa_opaque == 0) ? 0 :
XAT0_APPENDONLY;
if (XVA_ISSET_REQ(xvap, XAT_NODUMP))
*attrs |= (xoap->xoa_nodump == 0) ? 0 :
XAT0_NODUMP;
if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED))
*attrs |= (xoap->xoa_av_quarantined == 0) ? 0 :
XAT0_AV_QUARANTINED;
if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED))
*attrs |= (xoap->xoa_av_modified == 0) ? 0 :
XAT0_AV_MODIFIED;
if (XVA_ISSET_REQ(xvap, XAT_CREATETIME))
ZFS_TIME_ENCODE(&xoap->xoa_createtime, crtime);
if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP))
bcopy(xoap->xoa_av_scanstamp, scanstamp, AV_SCANSTAMP_SZ);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (XVA_ISSET_REQ(xvap, XAT_REPARSE))
*attrs |= (xoap->xoa_reparse == 0) ? 0 :
XAT0_REPARSE;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
if (XVA_ISSET_REQ(xvap, XAT_OFFLINE))
*attrs |= (xoap->xoa_offline == 0) ? 0 :
XAT0_OFFLINE;
if (XVA_ISSET_REQ(xvap, XAT_SPARSE))
*attrs |= (xoap->xoa_sparse == 0) ? 0 :
XAT0_SPARSE;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
static void *
zfs_log_fuid_ids(zfs_fuid_info_t *fuidp, void *start)
{
zfs_fuid_t *zfuid;
uint64_t *fuidloc = start;
/* First copy in the ACE FUIDs */
for (zfuid = list_head(&fuidp->z_fuids); zfuid;
zfuid = list_next(&fuidp->z_fuids, zfuid)) {
*fuidloc++ = zfuid->z_logfuid;
}
return (fuidloc);
}
static void *
zfs_log_fuid_domains(zfs_fuid_info_t *fuidp, void *start)
{
zfs_fuid_domain_t *zdomain;
/* now copy in the domain info, if any */
if (fuidp->z_domain_str_sz != 0) {
for (zdomain = list_head(&fuidp->z_domains); zdomain;
zdomain = list_next(&fuidp->z_domains, zdomain)) {
bcopy((void *)zdomain->z_domain, start,
strlen(zdomain->z_domain) + 1);
start = (caddr_t)start +
strlen(zdomain->z_domain) + 1;
}
}
return (start);
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_CREATE, TX_CREATE_ATTR, TX_MKDIR, TX_MKDIR_ATTR and
* TK_MKXATTR transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*
* TX_CREATE and TX_MKDIR are standard creates, but they may have FUID
* domain information appended prior to the name. In this case the
* uid/gid in the log record will be a log centric FUID.
*
* TX_CREATE_ACL_ATTR and TX_MKDIR_ACL_ATTR handle special creates that
* may contain attributes, ACL and optional fuid information.
*
* TX_CREATE_ACL and TX_MKDIR_ACL handle special creates that specify
* and ACL and normal users/groups in the ACEs.
*
* There may be an optional xvattr attribute information similar
* to zfs_log_setattr.
*
* Also, after the file name "domain" strings may be appended.
*/
void
zfs_log_create(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
znode_t *dzp, znode_t *zp, char *name, vsecattr_t *vsecp,
zfs_fuid_info_t *fuidp, vattr_t *vap)
{
itx_t *itx;
lr_create_t *lr;
lr_acl_create_t *lracl;
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 23:24:09 +03:00
size_t aclsize = 0;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
size_t xvatsize = 0;
size_t txsize;
Illumos #3522 3522 zfs module should not allow uninitialized variables Reviewed by: Sebastien Roy <seb@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Matthew Ahrens <mahrens@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3522 illumos/illumos-gate@d5285cae913f4e01ffa0e6693a6d8ef1fbea30ba Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting notes: 1. ZFSOnLinux had already addressed many of these issues because of its use of -Wall. However, the manner in which they were addressed differed. The illumos fixes replace the ones previously made in ZFSOnLinux to reduce code differences. 2. Part of the upstream patch made a small change to arc.c that might address zfsonlinux/zfs#1334. 3. The initialization of aclsize in zfs_log_create() differs because vsecp is a NULL pointer on ZFSOnLinux. 4. The changes to zfs_register_callbacks() were dropped because it has diverged and needs to be resynced.
2013-02-11 10:21:05 +04:00
xvattr_t *xvap = (xvattr_t *)vap;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void *end;
size_t lrsize;
size_t namesize = strlen(name) + 1;
size_t fuidsz = 0;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
/*
* If we have FUIDs present then add in space for
* domains and ACE fuid's if any.
*/
if (fuidp) {
fuidsz += fuidp->z_domain_str_sz;
fuidsz += fuidp->z_fuid_cnt * sizeof (uint64_t);
}
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 23:24:09 +03:00
if (vap->va_mask & ATTR_XVATTR)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
xvatsize = ZIL_XVAT_SIZE(xvap->xva_mapsize);
if ((int)txtype == TX_CREATE_ATTR || (int)txtype == TX_MKDIR_ATTR ||
(int)txtype == TX_CREATE || (int)txtype == TX_MKDIR ||
(int)txtype == TX_MKXATTR) {
txsize = sizeof (*lr) + namesize + fuidsz + xvatsize;
lrsize = sizeof (*lr);
} else {
txsize =
sizeof (lr_acl_create_t) + namesize + fuidsz +
ZIL_ACE_LENGTH(aclsize) + xvatsize;
lrsize = sizeof (lr_acl_create_t);
}
itx = zil_itx_create(txtype, txsize);
lr = (lr_create_t *)&itx->itx_lr;
lr->lr_doid = dzp->z_id;
lr->lr_foid = zp->z_id;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
lr->lr_mode = zp->z_mode;
if (!IS_EPHEMERAL(zp->z_uid)) {
lr->lr_uid = (uint64_t)zp->z_uid;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
} else {
lr->lr_uid = fuidp->z_fuid_owner;
}
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (!IS_EPHEMERAL(zp->z_gid)) {
lr->lr_gid = (uint64_t)zp->z_gid;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
} else {
lr->lr_gid = fuidp->z_fuid_group;
}
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 22:33:08 +03:00
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(ZTOZSB(zp)), &lr->lr_gen,
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
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 22:33:08 +03:00
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_CRTIME(ZTOZSB(zp)),
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
lr->lr_crtime, sizeof (uint64_t) * 2);
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 22:33:08 +03:00
if (sa_lookup(zp->z_sa_hdl, SA_ZPL_RDEV(ZTOZSB(zp)), &lr->lr_rdev,
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
sizeof (lr->lr_rdev)) != 0)
lr->lr_rdev = 0;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
* Fill in xvattr info if any
*/
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 23:24:09 +03:00
if (vap->va_mask & ATTR_XVATTR) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
zfs_log_xvattr((lr_attr_t *)((caddr_t)lr + lrsize), xvap);
end = (caddr_t)lr + lrsize + xvatsize;
} else {
end = (caddr_t)lr + lrsize;
}
/* Now fill in any ACL info */
if (vsecp) {
lracl = (lr_acl_create_t *)&itx->itx_lr;
lracl->lr_aclcnt = vsecp->vsa_aclcnt;
lracl->lr_acl_bytes = aclsize;
lracl->lr_domcnt = fuidp ? fuidp->z_domain_cnt : 0;
lracl->lr_fuidcnt = fuidp ? fuidp->z_fuid_cnt : 0;
if (vsecp->vsa_aclflags & VSA_ACE_ACLFLAGS)
lracl->lr_acl_flags = (uint64_t)vsecp->vsa_aclflags;
else
lracl->lr_acl_flags = 0;
bcopy(vsecp->vsa_aclentp, end, aclsize);
end = (caddr_t)end + ZIL_ACE_LENGTH(aclsize);
}
/* drop in FUID info */
if (fuidp) {
end = zfs_log_fuid_ids(fuidp, end);
end = zfs_log_fuid_domains(fuidp, end);
}
/*
* Now place file name in log record
*/
bcopy(name, end, namesize);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles both TX_REMOVE and TX_RMDIR transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
void
zfs_log_remove(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
znode_t *dzp, char *name, uint64_t foid)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
itx_t *itx;
lr_remove_t *lr;
size_t namesize = strlen(name) + 1;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
itx = zil_itx_create(txtype, sizeof (*lr) + namesize);
lr = (lr_remove_t *)&itx->itx_lr;
lr->lr_doid = dzp->z_id;
bcopy(name, (char *)(lr + 1), namesize);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
itx->itx_oid = foid;
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_LINK transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
void
zfs_log_link(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
znode_t *dzp, znode_t *zp, char *name)
{
itx_t *itx;
lr_link_t *lr;
size_t namesize = strlen(name) + 1;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
itx = zil_itx_create(txtype, sizeof (*lr) + namesize);
lr = (lr_link_t *)&itx->itx_lr;
lr->lr_doid = dzp->z_id;
lr->lr_link_obj = zp->z_id;
bcopy(name, (char *)(lr + 1), namesize);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_SYMLINK transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
void
zfs_log_symlink(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
znode_t *dzp, znode_t *zp, char *name, char *link)
{
itx_t *itx;
lr_create_t *lr;
size_t namesize = strlen(name) + 1;
size_t linksize = strlen(link) + 1;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
itx = zil_itx_create(txtype, sizeof (*lr) + namesize + linksize);
lr = (lr_create_t *)&itx->itx_lr;
lr->lr_doid = dzp->z_id;
lr->lr_foid = zp->z_id;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
lr->lr_uid = zp->z_uid;
lr->lr_gid = zp->z_gid;
lr->lr_mode = zp->z_mode;
Prototype/structure update for Linux I appologize in advance why to many things ended up in this commit. When it could be seperated in to a whole series of commits teasing that all apart now would take considerable time and I'm not sure there's much merrit in it. As such I'll just summerize the intent of the changes which are all (or partly) in this commit. Broadly the intent is to remove as much Solaris specific code as possible and replace it with native Linux equivilants. More specifically: 1) Replace all instances of zfsvfs_t with zfs_sb_t. While the type is largely the same calling it private super block data rather than a zfsvfs is more consistent with how Linux names this. While non critical it makes the code easier to read when your thinking in Linux friendly VFS terms. 2) Replace vnode_t with struct inode. The Linux VFS doesn't have the notion of a vnode and there's absolutely no good reason to create one. There are in fact several good reasons to remove it. It just adds overhead on Linux if we were to manage one, it conplicates the code, and it likely will lead to bugs so there's a good change it will be out of date. The code has been updated to remove all need for this type. 3) Replace all vtype_t's with umode types. Along with this shift all uses of types to mode bits. The Solaris code would pass a vtype which is redundant with the Linux mode. Just update all the code to use the Linux mode macros and remove this redundancy. 4) Remove using of vn_* helpers and replace where needed with inode helpers. The big example here is creating iput_aync to replace vn_rele_async. Other vn helpers will be addressed as needed but they should be be emulated. They are a Solaris VFS'ism and should simply be replaced with Linux equivilants. 5) Update znode alloc/free code. Under Linux it's common to embed the inode specific data with the inode itself. This removes the need for an extra memory allocation. In zfs this information is called a znode and it now embeds the inode with it. Allocators have been updated accordingly. 6) Minimal integration with the vfs flags for setting up the super block and handling mount options has been added this code will need to be refined but functionally it's all there. This will be the first and last of these to large to review commits.
2011-02-08 22:16:06 +03:00
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(ZTOZSB(zp)), &lr->lr_gen,
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
sizeof (uint64_t));
Prototype/structure update for Linux I appologize in advance why to many things ended up in this commit. When it could be seperated in to a whole series of commits teasing that all apart now would take considerable time and I'm not sure there's much merrit in it. As such I'll just summerize the intent of the changes which are all (or partly) in this commit. Broadly the intent is to remove as much Solaris specific code as possible and replace it with native Linux equivilants. More specifically: 1) Replace all instances of zfsvfs_t with zfs_sb_t. While the type is largely the same calling it private super block data rather than a zfsvfs is more consistent with how Linux names this. While non critical it makes the code easier to read when your thinking in Linux friendly VFS terms. 2) Replace vnode_t with struct inode. The Linux VFS doesn't have the notion of a vnode and there's absolutely no good reason to create one. There are in fact several good reasons to remove it. It just adds overhead on Linux if we were to manage one, it conplicates the code, and it likely will lead to bugs so there's a good change it will be out of date. The code has been updated to remove all need for this type. 3) Replace all vtype_t's with umode types. Along with this shift all uses of types to mode bits. The Solaris code would pass a vtype which is redundant with the Linux mode. Just update all the code to use the Linux mode macros and remove this redundancy. 4) Remove using of vn_* helpers and replace where needed with inode helpers. The big example here is creating iput_aync to replace vn_rele_async. Other vn helpers will be addressed as needed but they should be be emulated. They are a Solaris VFS'ism and should simply be replaced with Linux equivilants. 5) Update znode alloc/free code. Under Linux it's common to embed the inode specific data with the inode itself. This removes the need for an extra memory allocation. In zfs this information is called a znode and it now embeds the inode with it. Allocators have been updated accordingly. 6) Minimal integration with the vfs flags for setting up the super block and handling mount options has been added this code will need to be refined but functionally it's all there. This will be the first and last of these to large to review commits.
2011-02-08 22:16:06 +03:00
(void) sa_lookup(zp->z_sa_hdl, SA_ZPL_CRTIME(ZTOZSB(zp)),
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
lr->lr_crtime, sizeof (uint64_t) * 2);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
bcopy(name, (char *)(lr + 1), namesize);
bcopy(link, (char *)(lr + 1) + namesize, linksize);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_RENAME transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
void
zfs_log_rename(zilog_t *zilog, dmu_tx_t *tx, uint64_t txtype,
znode_t *sdzp, char *sname, znode_t *tdzp, char *dname, znode_t *szp)
{
itx_t *itx;
lr_rename_t *lr;
size_t snamesize = strlen(sname) + 1;
size_t dnamesize = strlen(dname) + 1;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
itx = zil_itx_create(txtype, sizeof (*lr) + snamesize + dnamesize);
lr = (lr_rename_t *)&itx->itx_lr;
lr->lr_sdoid = sdzp->z_id;
lr->lr_tdoid = tdzp->z_id;
bcopy(sname, (char *)(lr + 1), snamesize);
bcopy(dname, (char *)(lr + 1) + snamesize, dnamesize);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
itx->itx_oid = szp->z_id;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Only commit the ZIL once in zpl_writepages() (msync() case). Currently, using msync() results in the following code path: sys_msync -> zpl_fsync -> filemap_write_and_wait_range -> zpl_writepages -> write_cache_pages -> zpl_putpage In such a code path, zil_commit() is called as part of zpl_putpage(). This means that for each page, the write is handed to the DMU, the ZIL is committed, and only then do we move on to the next page. As one might imagine, this results in atrocious performance where there is a large number of pages to write: instead of committing a batch of N writes, we do N commits containing one page each. In some extreme cases this can result in msync() being ~700 times slower than it should be, as well as very inefficient use of ZIL resources. This patch fixes this issue by making sure that the requested writes are batched and then committed only once. Unfortunately, the implementation is somewhat non-trivial because there is no way to run write_cache_pages in SYNC mode (so that we get all pages) without making it wait on the writeback tag for each page. The solution implemented here is composed of two parts: - I added a new callback system to the ZIL, which allows the caller to be notified when its ITX gets written to stable storage. One nice thing is that the callback is called not only in zil_commit() but in zil_sync() as well, which means that the caller doesn't have to care whether the write ended up in the ZIL or the DMU: it will get notified as soon as it's safe, period. This is an improvement over dmu_tx_callback_register() that was used previously, which only supports DMU writes. The rationale for this change is to allow zpl_putpage() to be notified when a ZIL commit is completed without having to block on zil_commit() itself. - zpl_writepages() now calls write_cache_pages in non-SYNC mode, which will prevent (1) write_cache_pages from blocking, and (2) zpl_putpage from issuing ZIL commits. zpl_writepages() will issue the commit itself instead of relying on zpl_putpage() to do it, thus nicely batching the writes. Note, however, that we still have to call write_cache_pages() again in SYNC mode because there is an edge case documented in the implementation of write_cache_pages() whereas it will not give us all dirty pages when running in non-SYNC mode. Thus we need to run it at least once in SYNC mode to make sure we honor persistency guarantees. This only happens when the pages are modified at the same time msync() is running, which should be rare. In most cases there won't be any additional pages and this second call will do nothing. Note that this change also fixes a bug related to #907 whereas calling msync() on pages that were already handed over to the DMU in a previous writepages() call would make msync() block until the next TXG sync instead of returning as soon as the ZIL commit is complete. The new callback system fixes that problem. Signed-off-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1849 Closes #907
2013-11-10 19:00:11 +04:00
* zfs_log_write() handles TX_WRITE transactions. The specified callback is
* called as soon as the write is on stable storage (be it via a DMU sync or a
* ZIL commit).
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
Make zfs_immediate_write_sz a module paramater zfs_immediate_write_sz variable is a tunable, but lacks proper module_param() instrumentation. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1032
2012-10-08 22:02:20 +04:00
long zfs_immediate_write_sz = 32768;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void
zfs_log_write(zilog_t *zilog, dmu_tx_t *tx, int txtype,
Only commit the ZIL once in zpl_writepages() (msync() case). Currently, using msync() results in the following code path: sys_msync -> zpl_fsync -> filemap_write_and_wait_range -> zpl_writepages -> write_cache_pages -> zpl_putpage In such a code path, zil_commit() is called as part of zpl_putpage(). This means that for each page, the write is handed to the DMU, the ZIL is committed, and only then do we move on to the next page. As one might imagine, this results in atrocious performance where there is a large number of pages to write: instead of committing a batch of N writes, we do N commits containing one page each. In some extreme cases this can result in msync() being ~700 times slower than it should be, as well as very inefficient use of ZIL resources. This patch fixes this issue by making sure that the requested writes are batched and then committed only once. Unfortunately, the implementation is somewhat non-trivial because there is no way to run write_cache_pages in SYNC mode (so that we get all pages) without making it wait on the writeback tag for each page. The solution implemented here is composed of two parts: - I added a new callback system to the ZIL, which allows the caller to be notified when its ITX gets written to stable storage. One nice thing is that the callback is called not only in zil_commit() but in zil_sync() as well, which means that the caller doesn't have to care whether the write ended up in the ZIL or the DMU: it will get notified as soon as it's safe, period. This is an improvement over dmu_tx_callback_register() that was used previously, which only supports DMU writes. The rationale for this change is to allow zpl_putpage() to be notified when a ZIL commit is completed without having to block on zil_commit() itself. - zpl_writepages() now calls write_cache_pages in non-SYNC mode, which will prevent (1) write_cache_pages from blocking, and (2) zpl_putpage from issuing ZIL commits. zpl_writepages() will issue the commit itself instead of relying on zpl_putpage() to do it, thus nicely batching the writes. Note, however, that we still have to call write_cache_pages() again in SYNC mode because there is an edge case documented in the implementation of write_cache_pages() whereas it will not give us all dirty pages when running in non-SYNC mode. Thus we need to run it at least once in SYNC mode to make sure we honor persistency guarantees. This only happens when the pages are modified at the same time msync() is running, which should be rare. In most cases there won't be any additional pages and this second call will do nothing. Note that this change also fixes a bug related to #907 whereas calling msync() on pages that were already handed over to the DMU in a previous writepages() call would make msync() block until the next TXG sync instead of returning as soon as the ZIL commit is complete. The new callback system fixes that problem. Signed-off-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1849 Closes #907
2013-11-10 19:00:11 +04:00
znode_t *zp, offset_t off, ssize_t resid, int ioflag,
zil_callback_t callback, void *callback_data)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
itx_wr_state_t write_state;
boolean_t slogging;
Revert "Remove TSD zfs_fsyncer_key" This reverts commit 31f2b5abdf95d8426d8bfd66ca7f62ec70215e3c back to the original code until the fsync(2) performance regression can be addressed. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-12-20 21:55:47 +04:00
uintptr_t fsync_cnt;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
ssize_t immediate_write_sz;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Only commit the ZIL once in zpl_writepages() (msync() case). Currently, using msync() results in the following code path: sys_msync -> zpl_fsync -> filemap_write_and_wait_range -> zpl_writepages -> write_cache_pages -> zpl_putpage In such a code path, zil_commit() is called as part of zpl_putpage(). This means that for each page, the write is handed to the DMU, the ZIL is committed, and only then do we move on to the next page. As one might imagine, this results in atrocious performance where there is a large number of pages to write: instead of committing a batch of N writes, we do N commits containing one page each. In some extreme cases this can result in msync() being ~700 times slower than it should be, as well as very inefficient use of ZIL resources. This patch fixes this issue by making sure that the requested writes are batched and then committed only once. Unfortunately, the implementation is somewhat non-trivial because there is no way to run write_cache_pages in SYNC mode (so that we get all pages) without making it wait on the writeback tag for each page. The solution implemented here is composed of two parts: - I added a new callback system to the ZIL, which allows the caller to be notified when its ITX gets written to stable storage. One nice thing is that the callback is called not only in zil_commit() but in zil_sync() as well, which means that the caller doesn't have to care whether the write ended up in the ZIL or the DMU: it will get notified as soon as it's safe, period. This is an improvement over dmu_tx_callback_register() that was used previously, which only supports DMU writes. The rationale for this change is to allow zpl_putpage() to be notified when a ZIL commit is completed without having to block on zil_commit() itself. - zpl_writepages() now calls write_cache_pages in non-SYNC mode, which will prevent (1) write_cache_pages from blocking, and (2) zpl_putpage from issuing ZIL commits. zpl_writepages() will issue the commit itself instead of relying on zpl_putpage() to do it, thus nicely batching the writes. Note, however, that we still have to call write_cache_pages() again in SYNC mode because there is an edge case documented in the implementation of write_cache_pages() whereas it will not give us all dirty pages when running in non-SYNC mode. Thus we need to run it at least once in SYNC mode to make sure we honor persistency guarantees. This only happens when the pages are modified at the same time msync() is running, which should be rare. In most cases there won't be any additional pages and this second call will do nothing. Note that this change also fixes a bug related to #907 whereas calling msync() on pages that were already handed over to the DMU in a previous writepages() call would make msync() block until the next TXG sync instead of returning as soon as the ZIL commit is complete. The new callback system fixes that problem. Signed-off-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1849 Closes #907
2013-11-10 19:00:11 +04:00
if (zil_replaying(zilog, tx) || zp->z_unlinked) {
if (callback != NULL)
callback(callback_data);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
Only commit the ZIL once in zpl_writepages() (msync() case). Currently, using msync() results in the following code path: sys_msync -> zpl_fsync -> filemap_write_and_wait_range -> zpl_writepages -> write_cache_pages -> zpl_putpage In such a code path, zil_commit() is called as part of zpl_putpage(). This means that for each page, the write is handed to the DMU, the ZIL is committed, and only then do we move on to the next page. As one might imagine, this results in atrocious performance where there is a large number of pages to write: instead of committing a batch of N writes, we do N commits containing one page each. In some extreme cases this can result in msync() being ~700 times slower than it should be, as well as very inefficient use of ZIL resources. This patch fixes this issue by making sure that the requested writes are batched and then committed only once. Unfortunately, the implementation is somewhat non-trivial because there is no way to run write_cache_pages in SYNC mode (so that we get all pages) without making it wait on the writeback tag for each page. The solution implemented here is composed of two parts: - I added a new callback system to the ZIL, which allows the caller to be notified when its ITX gets written to stable storage. One nice thing is that the callback is called not only in zil_commit() but in zil_sync() as well, which means that the caller doesn't have to care whether the write ended up in the ZIL or the DMU: it will get notified as soon as it's safe, period. This is an improvement over dmu_tx_callback_register() that was used previously, which only supports DMU writes. The rationale for this change is to allow zpl_putpage() to be notified when a ZIL commit is completed without having to block on zil_commit() itself. - zpl_writepages() now calls write_cache_pages in non-SYNC mode, which will prevent (1) write_cache_pages from blocking, and (2) zpl_putpage from issuing ZIL commits. zpl_writepages() will issue the commit itself instead of relying on zpl_putpage() to do it, thus nicely batching the writes. Note, however, that we still have to call write_cache_pages() again in SYNC mode because there is an edge case documented in the implementation of write_cache_pages() whereas it will not give us all dirty pages when running in non-SYNC mode. Thus we need to run it at least once in SYNC mode to make sure we honor persistency guarantees. This only happens when the pages are modified at the same time msync() is running, which should be rare. In most cases there won't be any additional pages and this second call will do nothing. Note that this change also fixes a bug related to #907 whereas calling msync() on pages that were already handed over to the DMU in a previous writepages() call would make msync() block until the next TXG sync instead of returning as soon as the ZIL commit is complete. The new callback system fixes that problem. Signed-off-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1849 Closes #907
2013-11-10 19:00:11 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
immediate_write_sz = (zilog->zl_logbias == ZFS_LOGBIAS_THROUGHPUT)
Make zfs_immediate_write_sz a module paramater zfs_immediate_write_sz variable is a tunable, but lacks proper module_param() instrumentation. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1032
2012-10-08 22:02:20 +04:00
? 0 : (ssize_t)zfs_immediate_write_sz;
Rebase master to b105
2009-01-16 00:59:39 +03:00
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
slogging = spa_has_slogs(zilog->zl_spa) &&
(zilog->zl_logbias == ZFS_LOGBIAS_LATENCY);
if (resid > immediate_write_sz && !slogging && resid <= zp->z_blksz)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
write_state = WR_INDIRECT;
else if (ioflag & (FSYNC | FDSYNC))
write_state = WR_COPIED;
else
write_state = WR_NEED_COPY;
Revert "Remove TSD zfs_fsyncer_key" This reverts commit 31f2b5abdf95d8426d8bfd66ca7f62ec70215e3c back to the original code until the fsync(2) performance regression can be addressed. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-12-20 21:55:47 +04:00
if ((fsync_cnt = (uintptr_t)tsd_get(zfs_fsyncer_key)) != 0) {
(void) tsd_set(zfs_fsyncer_key, (void *)(fsync_cnt - 1));
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
while (resid) {
itx_t *itx;
lr_write_t *lr;
ssize_t len;
/*
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 23:09:06 +03:00
* If the write would overflow the largest block then split it.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
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 23:09:06 +03:00
if (write_state != WR_INDIRECT && resid > ZIL_MAX_LOG_DATA)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
len = SPA_MAXBLOCKSIZE >> 1;
else
len = resid;
itx = zil_itx_create(txtype, sizeof (*lr) +
(write_state == WR_COPIED ? len : 0));
lr = (lr_write_t *)&itx->itx_lr;
Prototype/structure update for Linux I appologize in advance why to many things ended up in this commit. When it could be seperated in to a whole series of commits teasing that all apart now would take considerable time and I'm not sure there's much merrit in it. As such I'll just summerize the intent of the changes which are all (or partly) in this commit. Broadly the intent is to remove as much Solaris specific code as possible and replace it with native Linux equivilants. More specifically: 1) Replace all instances of zfsvfs_t with zfs_sb_t. While the type is largely the same calling it private super block data rather than a zfsvfs is more consistent with how Linux names this. While non critical it makes the code easier to read when your thinking in Linux friendly VFS terms. 2) Replace vnode_t with struct inode. The Linux VFS doesn't have the notion of a vnode and there's absolutely no good reason to create one. There are in fact several good reasons to remove it. It just adds overhead on Linux if we were to manage one, it conplicates the code, and it likely will lead to bugs so there's a good change it will be out of date. The code has been updated to remove all need for this type. 3) Replace all vtype_t's with umode types. Along with this shift all uses of types to mode bits. The Solaris code would pass a vtype which is redundant with the Linux mode. Just update all the code to use the Linux mode macros and remove this redundancy. 4) Remove using of vn_* helpers and replace where needed with inode helpers. The big example here is creating iput_aync to replace vn_rele_async. Other vn helpers will be addressed as needed but they should be be emulated. They are a Solaris VFS'ism and should simply be replaced with Linux equivilants. 5) Update znode alloc/free code. Under Linux it's common to embed the inode specific data with the inode itself. This removes the need for an extra memory allocation. In zfs this information is called a znode and it now embeds the inode with it. Allocators have been updated accordingly. 6) Minimal integration with the vfs flags for setting up the super block and handling mount options has been added this code will need to be refined but functionally it's all there. This will be the first and last of these to large to review commits.
2011-02-08 22:16:06 +03:00
if (write_state == WR_COPIED && dmu_read(ZTOZSB(zp)->z_os,
Rebase master to b117
2009-07-03 02:44:48 +04:00
zp->z_id, off, len, lr + 1, DMU_READ_NO_PREFETCH) != 0) {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
zil_itx_destroy(itx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
itx = zil_itx_create(txtype, sizeof (*lr));
lr = (lr_write_t *)&itx->itx_lr;
write_state = WR_NEED_COPY;
}
itx->itx_wr_state = write_state;
if (write_state == WR_NEED_COPY)
itx->itx_sod += len;
lr->lr_foid = zp->z_id;
lr->lr_offset = off;
lr->lr_length = len;
lr->lr_blkoff = 0;
BP_ZERO(&lr->lr_blkptr);
Prototype/structure update for Linux I appologize in advance why to many things ended up in this commit. When it could be seperated in to a whole series of commits teasing that all apart now would take considerable time and I'm not sure there's much merrit in it. As such I'll just summerize the intent of the changes which are all (or partly) in this commit. Broadly the intent is to remove as much Solaris specific code as possible and replace it with native Linux equivilants. More specifically: 1) Replace all instances of zfsvfs_t with zfs_sb_t. While the type is largely the same calling it private super block data rather than a zfsvfs is more consistent with how Linux names this. While non critical it makes the code easier to read when your thinking in Linux friendly VFS terms. 2) Replace vnode_t with struct inode. The Linux VFS doesn't have the notion of a vnode and there's absolutely no good reason to create one. There are in fact several good reasons to remove it. It just adds overhead on Linux if we were to manage one, it conplicates the code, and it likely will lead to bugs so there's a good change it will be out of date. The code has been updated to remove all need for this type. 3) Replace all vtype_t's with umode types. Along with this shift all uses of types to mode bits. The Solaris code would pass a vtype which is redundant with the Linux mode. Just update all the code to use the Linux mode macros and remove this redundancy. 4) Remove using of vn_* helpers and replace where needed with inode helpers. The big example here is creating iput_aync to replace vn_rele_async. Other vn helpers will be addressed as needed but they should be be emulated. They are a Solaris VFS'ism and should simply be replaced with Linux equivilants. 5) Update znode alloc/free code. Under Linux it's common to embed the inode specific data with the inode itself. This removes the need for an extra memory allocation. In zfs this information is called a znode and it now embeds the inode with it. Allocators have been updated accordingly. 6) Minimal integration with the vfs flags for setting up the super block and handling mount options has been added this code will need to be refined but functionally it's all there. This will be the first and last of these to large to review commits.
2011-02-08 22:16:06 +03:00
itx->itx_private = ZTOZSB(zp);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Revert "Remove TSD zfs_fsyncer_key" This reverts commit 31f2b5abdf95d8426d8bfd66ca7f62ec70215e3c back to the original code until the fsync(2) performance regression can be addressed. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-12-20 21:55:47 +04:00
if (!(ioflag & (FSYNC | FDSYNC)) && (zp->z_sync_cnt == 0) &&
(fsync_cnt == 0))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
itx->itx_sync = B_FALSE;
Only commit the ZIL once in zpl_writepages() (msync() case). Currently, using msync() results in the following code path: sys_msync -> zpl_fsync -> filemap_write_and_wait_range -> zpl_writepages -> write_cache_pages -> zpl_putpage In such a code path, zil_commit() is called as part of zpl_putpage(). This means that for each page, the write is handed to the DMU, the ZIL is committed, and only then do we move on to the next page. As one might imagine, this results in atrocious performance where there is a large number of pages to write: instead of committing a batch of N writes, we do N commits containing one page each. In some extreme cases this can result in msync() being ~700 times slower than it should be, as well as very inefficient use of ZIL resources. This patch fixes this issue by making sure that the requested writes are batched and then committed only once. Unfortunately, the implementation is somewhat non-trivial because there is no way to run write_cache_pages in SYNC mode (so that we get all pages) without making it wait on the writeback tag for each page. The solution implemented here is composed of two parts: - I added a new callback system to the ZIL, which allows the caller to be notified when its ITX gets written to stable storage. One nice thing is that the callback is called not only in zil_commit() but in zil_sync() as well, which means that the caller doesn't have to care whether the write ended up in the ZIL or the DMU: it will get notified as soon as it's safe, period. This is an improvement over dmu_tx_callback_register() that was used previously, which only supports DMU writes. The rationale for this change is to allow zpl_putpage() to be notified when a ZIL commit is completed without having to block on zil_commit() itself. - zpl_writepages() now calls write_cache_pages in non-SYNC mode, which will prevent (1) write_cache_pages from blocking, and (2) zpl_putpage from issuing ZIL commits. zpl_writepages() will issue the commit itself instead of relying on zpl_putpage() to do it, thus nicely batching the writes. Note, however, that we still have to call write_cache_pages() again in SYNC mode because there is an edge case documented in the implementation of write_cache_pages() whereas it will not give us all dirty pages when running in non-SYNC mode. Thus we need to run it at least once in SYNC mode to make sure we honor persistency guarantees. This only happens when the pages are modified at the same time msync() is running, which should be rare. In most cases there won't be any additional pages and this second call will do nothing. Note that this change also fixes a bug related to #907 whereas calling msync() on pages that were already handed over to the DMU in a previous writepages() call would make msync() block until the next TXG sync instead of returning as soon as the ZIL commit is complete. The new callback system fixes that problem. Signed-off-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1849 Closes #907
2013-11-10 19:00:11 +04:00
itx->itx_callback = callback;
itx->itx_callback_data = callback_data;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
off += len;
resid -= len;
}
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_TRUNCATE transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
void
zfs_log_truncate(zilog_t *zilog, dmu_tx_t *tx, int txtype,
znode_t *zp, uint64_t off, uint64_t len)
{
itx_t *itx;
lr_truncate_t *lr;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx) || zp->z_unlinked)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
itx = zil_itx_create(txtype, sizeof (*lr));
lr = (lr_truncate_t *)&itx->itx_lr;
lr->lr_foid = zp->z_id;
lr->lr_offset = off;
lr->lr_length = len;
itx->itx_sync = (zp->z_sync_cnt != 0);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_SETATTR transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
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 23:24:09 +03:00
zfs_log_setattr(zilog_t *zilog, dmu_tx_t *tx, int txtype,
znode_t *zp, vattr_t *vap, uint_t mask_applied, zfs_fuid_info_t *fuidp)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
itx_t *itx;
lr_setattr_t *lr;
xvattr_t *xvap = (xvattr_t *)vap;
size_t recsize = sizeof (lr_setattr_t);
void *start;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx) || zp->z_unlinked)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
/*
* If XVATTR set, then log record size needs to allow
* for lr_attr_t + xvattr mask, mapsize and create time
* plus actual attribute values
*/
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 23:24:09 +03:00
if (vap->va_mask & ATTR_XVATTR)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
recsize = sizeof (*lr) + ZIL_XVAT_SIZE(xvap->xva_mapsize);
if (fuidp)
recsize += fuidp->z_domain_str_sz;
itx = zil_itx_create(txtype, recsize);
lr = (lr_setattr_t *)&itx->itx_lr;
lr->lr_foid = zp->z_id;
lr->lr_mask = (uint64_t)mask_applied;
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 23:24:09 +03:00
lr->lr_mode = (uint64_t)vap->va_mode;
if ((mask_applied & ATTR_UID) && IS_EPHEMERAL(vap->va_uid))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
lr->lr_uid = fuidp->z_fuid_owner;
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 23:24:09 +03:00
lr->lr_uid = (uint64_t)vap->va_uid;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
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 23:24:09 +03:00
if ((mask_applied & ATTR_GID) && IS_EPHEMERAL(vap->va_gid))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
lr->lr_gid = fuidp->z_fuid_group;
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 23:24:09 +03:00
lr->lr_gid = (uint64_t)vap->va_gid;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
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 23:24:09 +03:00
lr->lr_size = (uint64_t)vap->va_size;
ZFS_TIME_ENCODE(&vap->va_atime, lr->lr_atime);
ZFS_TIME_ENCODE(&vap->va_mtime, lr->lr_mtime);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
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 23:24:09 +03:00
if (vap->va_mask & ATTR_XVATTR) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
zfs_log_xvattr((lr_attr_t *)start, xvap);
start = (caddr_t)start + ZIL_XVAT_SIZE(xvap->xva_mapsize);
}
/*
* Now stick on domain information if any on end
*/
if (fuidp)
(void) zfs_log_fuid_domains(fuidp, start);
itx->itx_sync = (zp->z_sync_cnt != 0);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
/*
Illumos #3742 3742 zfs comments need cleaner, more consistent style Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3742 illumos/illumos-gate@f7170741490edba9d1d9c697c177c887172bc741 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The change to zfs_vfsops.c was dropped because it involves zfs_mount_label_policy, which does not exist in the Linux port.
2013-06-11 21:12:34 +04:00
* Handles TX_ACL transactions.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
void
zfs_log_acl(zilog_t *zilog, dmu_tx_t *tx, znode_t *zp,
vsecattr_t *vsecp, zfs_fuid_info_t *fuidp)
{
itx_t *itx;
lr_acl_v0_t *lrv0;
lr_acl_t *lr;
int txtype;
int lrsize;
size_t txsize;
size_t aclbytes = vsecp->vsa_aclentsz;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (zil_replaying(zilog, tx) || zp->z_unlinked)
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 23:09:06 +03:00
return;
Prototype/structure update for Linux I appologize in advance why to many things ended up in this commit. When it could be seperated in to a whole series of commits teasing that all apart now would take considerable time and I'm not sure there's much merrit in it. As such I'll just summerize the intent of the changes which are all (or partly) in this commit. Broadly the intent is to remove as much Solaris specific code as possible and replace it with native Linux equivilants. More specifically: 1) Replace all instances of zfsvfs_t with zfs_sb_t. While the type is largely the same calling it private super block data rather than a zfsvfs is more consistent with how Linux names this. While non critical it makes the code easier to read when your thinking in Linux friendly VFS terms. 2) Replace vnode_t with struct inode. The Linux VFS doesn't have the notion of a vnode and there's absolutely no good reason to create one. There are in fact several good reasons to remove it. It just adds overhead on Linux if we were to manage one, it conplicates the code, and it likely will lead to bugs so there's a good change it will be out of date. The code has been updated to remove all need for this type. 3) Replace all vtype_t's with umode types. Along with this shift all uses of types to mode bits. The Solaris code would pass a vtype which is redundant with the Linux mode. Just update all the code to use the Linux mode macros and remove this redundancy. 4) Remove using of vn_* helpers and replace where needed with inode helpers. The big example here is creating iput_aync to replace vn_rele_async. Other vn helpers will be addressed as needed but they should be be emulated. They are a Solaris VFS'ism and should simply be replaced with Linux equivilants. 5) Update znode alloc/free code. Under Linux it's common to embed the inode specific data with the inode itself. This removes the need for an extra memory allocation. In zfs this information is called a znode and it now embeds the inode with it. Allocators have been updated accordingly. 6) Minimal integration with the vfs flags for setting up the super block and handling mount options has been added this code will need to be refined but functionally it's all there. This will be the first and last of these to large to review commits.
2011-02-08 22:16:06 +03:00
txtype = (ZTOZSB(zp)->z_version < ZPL_VERSION_FUID) ?
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
TX_ACL_V0 : TX_ACL;
if (txtype == TX_ACL)
lrsize = sizeof (*lr);
else
lrsize = sizeof (*lrv0);
txsize = lrsize +
((txtype == TX_ACL) ? ZIL_ACE_LENGTH(aclbytes) : aclbytes) +
(fuidp ? fuidp->z_domain_str_sz : 0) +
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 23:09:06 +03:00
sizeof (uint64_t) * (fuidp ? fuidp->z_fuid_cnt : 0);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
itx = zil_itx_create(txtype, txsize);
lr = (lr_acl_t *)&itx->itx_lr;
lr->lr_foid = zp->z_id;
if (txtype == TX_ACL) {
lr->lr_acl_bytes = aclbytes;
lr->lr_domcnt = fuidp ? fuidp->z_domain_cnt : 0;
lr->lr_fuidcnt = fuidp ? fuidp->z_fuid_cnt : 0;
if (vsecp->vsa_mask & VSA_ACE_ACLFLAGS)
lr->lr_acl_flags = (uint64_t)vsecp->vsa_aclflags;
else
lr->lr_acl_flags = 0;
}
lr->lr_aclcnt = (uint64_t)vsecp->vsa_aclcnt;
if (txtype == TX_ACL_V0) {
lrv0 = (lr_acl_v0_t *)lr;
bcopy(vsecp->vsa_aclentp, (ace_t *)(lrv0 + 1), aclbytes);
} else {
void *start = (ace_t *)(lr + 1);
bcopy(vsecp->vsa_aclentp, start, aclbytes);
start = (caddr_t)start + ZIL_ACE_LENGTH(aclbytes);
if (fuidp) {
start = zfs_log_fuid_ids(fuidp, start);
(void) zfs_log_fuid_domains(fuidp, start);
}
}
itx->itx_sync = (zp->z_sync_cnt != 0);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
zil_itx_assign(zilog, itx, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Make zfs_immediate_write_sz a module paramater zfs_immediate_write_sz variable is a tunable, but lacks proper module_param() instrumentation. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1032
2012-10-08 22:02:20 +04:00
#if defined(_KERNEL) && defined(HAVE_SPL)
module_param(zfs_immediate_write_sz, long, 0644);
MODULE_PARM_DESC(zfs_immediate_write_sz, "Largest data block to write to zil");
#endif
Reference in New Issue Copy Permalink